Catalytic turnover of benzylamine by a model for the lysine tyrosylquinone (LTQ) cofactor of lysyl oxidase

Article abstract of DOI:10.1021/ja011141j

Lysyl oxidase differs from other copper amine oxidases in that its active quinone cofactor reflects cross-linking of a lysyl residue into the tyrosine-derived quinone nucleus found in the plasma and other copper amine oxidases. A model for the lysyl oxida

Full text of DOI:10.1021/ja011141j

9606
J. Am. Chem. Soc. 2001, 123, 9606-9611
Catalytic Turnover of Benzylamine by a Model for the Lysine
Tyrosylquinone (LTQ) Cofactor of Lysyl Oxidase
Ke-Qing Ling, Jisook Kim, and Lawrence M. Sayre*
Contribution from the Department of Chemistry, Case Western ReserVe UniVersity, CleVeland, Ohio 44106
ReceiVed May 8, 2001
Abstract: Lysyl oxidase differs from other copper amine oxidases in that its active quinone cofactor reflects
cross-linking of a lysyl residue into the tyrosine-derived quinone nucleus found in the plasma and other copper
amine oxidases. A model for the lysyl oxidase cofactor (LTQ), 3,3-dimethyl-2,3-dihydroindole-5,6-quinone
(4), was synthesized and found to be stable to both hydrolysis and oxidation events that prevent simpler models
from functioning as turnover catalysts. We show that 4 catalyzes the aerobic oxidative deamination of
benzylamine, though turnover eventually ceases on account of oxidation of the dihydrobenzoxazole tautomer
of the “product Schiff base” to form a benzoxazole, a reaction that may be physiologically relevant. The
mechanism of the overall reaction profile was elucidated by a combination of optical and NMR spectroscopy
and O2 uptake studies.
Introduction
resonance-stabilized anion whereas the latter is an amino-o-
quinone. There have been no studies to date reporting trans-
aminative activity of a viable LTQ model. Despite fairly
complete characterization of the TPQ-dependent reactions, aided
by the availability of several crystal structures and the behavior
of key mutants, the reaction catalyzed by the key mammalian
enzyme lysyl oxidase still awaits analogous investigation. Thus,
model studies aimed at identifying the intrinsic chemical
reactivity of the LTQ nucleus should provide important infor-
mation.
Interest in the mechanism of amine oxidation by the quinone-
dependent copper amine oxidases has resulted in model
systems1,2 that successfully mimic both the single turnover and
catalytic behavior of the 2,4,5-trihydroxyphenylalanine quinone
(TPQ) cofactor found in most of these enzymes.3 These reactions
involve condensation of primary amine with the C5 carbonyl
of TPQ to give a “substrate Schiff base”, followed by a rate-
limiting imine shift to give a “product Schiff base” that
hydrolyzes to release aldehyde product and a reductively
aminated cofactor, which in turn is reoxidized (O2 f H2O2) to
starting TPQ with release of NH3. One member of the copper
amine oxidase class, namely lysyl oxidase, exhibits different
spectroscopic features from the TPQ-dependent enzymes, now
understood in terms of it having a modified cofactor, termed
lysine tyrosylquinone (LTQ), where an active site lysine ꢀ-amino
group has added into the tyrosine-derived quinone nucleus.4 This
enzyme plays an integral role in the posttranslational maturation
of the mammalian connective-tissue proteins collagen and
elastin: conversion of protein-based lysines to R-aminoadipic
δ-semialdehydes initiates aldol and Schiff base condensation
chemistry which evolves into mature intra- and intermolecular
cross-links that stabilize the tissue matrix.
The first LTQ model (1) was synthesized to permit spectro-
scopic comparison with the newly identified enzyme cofactor,4,5
but molecules such as 1 are known to be unstable under aerobic
conditions6 and thus seem unsuitable as turnover catalysts. To
whatever extent molecules such as 1 might hydrolytically lose
the nitrogen substituent during catalytic oxidation of primary
amines, we considered that incorporation of this nitrogen into
a bicyclic skeleton might provide a more robust model. A
pertinent molecule in this regard (2) was previously described
Although TPQ and LTQ might be expected to share a similar
transamination capability utilizing the same electrophilic C5
carbonyl, they are quite different in that the former exists as a
* Address correspondence to this author.
(1) (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.
(2) (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.
(3) Brown, D. E.; McGuirl, M. A.; Dooley, D. M.; Janes, S. M.; Mu,
D.; Klinman, J. P. J. Biol. Chem. 1991, 266, 4049-4051.
(4) Wang, S. X.; Mure, M.; Medzihradszky, K. F.; Burlingame, A. L.;
Brown, D. E.; Dooley, D. M.; Smith, A. J.; Kagan, H. M.; Klinman, J. P.
Science 1996, 273, 1078-1084.
(5) Wang, S. X.; Nakamura, N.; Mure, M.; Klinman, J. P.; Sanders-
Loehr, J. J. Biol. Chem. 1997, 272, 28841-28844.
in studies on the oxidative cyclization of dopamine and
6-hydroxydopamine, but it is also an inadequate LTQ model
(6) Borchardt, R. T.; Smissman, E. E.; Nerland, D.; Reid, J. R. J. Med.
Chem. 1976, 19, 30-37.
10.1021/ja011141j CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/11/2001

Catalytic TurnoVer by a Lysyl Oxidase LTQ Model
J. Am. Chem. Soc., Vol. 123, No. 39, 2001 9607
Scheme 1a
a Conditions: (a) NaNH2/NH3(I), CH3I; (b) LiAlH4/Et2O/room temperature; (c) 48% HBr/100 °C; (d) Ag2O/CH3OH.
due to its susceptibility to be oxidized to 3.7 However, for the
gem-dimethyl analogue 4,8 such oxidation would be blocked,
and the alternative potential self-Diels-Alder reaction would
be sterically inhibited. Here we report the synthesis and catalytic
functioning of the LTQ model 4.
Results and Discussion
Synthesis and Properties of 4. Our synthesis of LTQ model
4 started from 3,4-dimethoxyphenylacetonitrile (5, Scheme 1).
Using sodium amide as a deprotonation reagent9 (phenylsodium
has also been used,10 but in our hands gave inferior yields), 5
was methylated to give 6 (optimal crude yields of 90% required
a two-step methylation), which was subsequently reduced with
LiAlH4 to afford amine 7.9 Deprotection of 7 was achieved by
refluxing with 48% HBr to give 8.10 The crude products 6 and
7 were both used without purification and the overall yield of
the first three steps (from 5 to 8) was 52%. The conversion of
8 to the target molecule 4, which involves oxidation to the
corresponding o-quinone 9, intramolecular amine conjugate
addition and tautomerization to give aminoquinol 10, and
oxidation of 10, can be achieved in one pot without isolation
of intermediates.11 We found that the ammonium bromide 8
could be directly oxidized by excess silver oxide, acting as the
base in generating the free amine 8 as well as the oxidant, to
give 4 in moderate yield (35%).
Compound 4 was found to be quite stable in buffered aqueous
acetonitrile over the pH range 7-10. At pH 7 it exhibits a λmax
of 485 nm (ꢀ ) 4110 M-1cm-1), 20 nm blue shifted from the
model 1 that closely mimics the absorption properties of lysyl
oxidase.5 The shifted absorption of model 4 likely arises from
a slight compromise in delocalization caused by the fused five-
membered ring. Evidence that the LTQ cofactor exists as the
neutral amino-o-quinone, contrasting the existence of TPQ as
a delocalized anion (the vinylogous carboxylic acid TPQ has a
pKa of 4.512), is that only the latter â-diketone/enolate system
exhibits rapid exchange of the C3-proton in D2O at pH 7.13 No
exchange is observed for LTQ,5 and likewise, 4 showed no
Figure 1. Progress of aerobic reaction of LTQ model compound 4
(1 mM) with benzylamine (50 mM) in 7:3 (v/v) pH 9 aqueous borate
buffer-CH3CN at 25 °C (0) plot for absorbance at 456 nm vs time;
(b) plot for catalytic turnover percentage ((S.D.) vs time.
exchange in 10 h in neutral CD3OD, though exchange could be
effected under basic conditions. On the basis of the C6 carbonyl
(13C NMR δ 164) being part of a vinylogous amide, it is the C5
carbonyl (13C NMR δ 184), as with TPQ, that is the electrophilic
center presumed to be the site of attack by amine substrates.
Catalytic Turnover. The ability of 4 to mediate the O2-
dependent oxidative deamination of benzylamine in buffered
aqueous CH3CN was studied in the pH range 7-9, as evaluated
by quantifying the 2,4-dinitrophenylhydrazone of PhCHdO. The
cumulative turnover yield plateaued over time, and the highest
yield (∼340%) was observed at pH 9 (borate buffer). Although
the reactions were faster at higher than lower pH, the factor
that governed the optimal yield at the point of plateau in every
case was a competing deterioration of the catalyst, as evidenced
by a gradual bleaching of the red color of 4, which also occurred
more rapidly at higher pH. Figure 1 shows the time course for
catalytic oxidative deamination of benzylamine at pH 9.0,
superimposed on the time course for deterioration of the
absorbance for 4 under the same conditions. The coincidence
of these plots implies that the plateau in turnoVer yield is due
to loss of the actiVe form of the LTQ model 4.
(7) Hikosaka, A.; Kumanotani, J. Bull. Chem. Soc. Jpn. 1970, 43, 2620-
2622.
(8) Compound 4 has been described (without characterization) to form
from oxidative cyclization of 5-(2-amino-1,1-dimethylethyl)benzene-1,2,4-
triol: Borchardt, R. T.; Reid, J. R.; Thakker, D. R.; Liang, Y. O.; Wightman,
R. W.; Adams, R. N. J. Med. Chem. 1976, 19, 1201-1209.
(9) Knabe, J.; Kubitz, J. Naturwissenschaften 1961, 48, 669.
(10) Finkelstein, J.; Chiang, E.; Brossi, A. J. Med. Chem. 1971, 14, 584-
588.
(11) Powell, W. S.; Heacock, R. A.; Smith, D. G.; McInnes, A. G. Can.
J. Chem. 1973, 51, 776-779.
(12) Mure, M.; Klinman, J. P. J. Am. Chem. Soc. 1993, 115, 7117-
7123.
Mechanism of Benzylamine Deamination. To learn about
the mechanism of the deamination reaction effected by 4, its
reaction with benzylamine was monitored under anaerobic
1
single-turnover conditions in degassed CD3CN by H NMR
spectroscopy.15 Using 1.5 equiv of benzylamine, suspended 4
was completely dissolved in 12 h (25 °C), yielding a spectrum
(14) Moenne-Luccoz, P.; Nakamura, N.; Steinebach, V.; Duine, J. A.;
Mure, M.; Klinman, J. P.; Sanders-Loehr, J. Biochemistry 1995, 34, 7020-
7026.
(13) This proton is readily exchangeable (in D2O) at pH 7 both for TPQ
in the E. coli amine oxidase14 and for TPQ models.14,1a

9608 J. Am. Chem. Soc., Vol. 123, No. 39, 2001
Ling et al.
Scheme 2
Scheme 3
dominated (∼92% from 4) by the product Schiff base 12 (see
Scheme 2). Also present to a lesser degree were the product of
interception of 12 by benzylamine, PhCHdNCH2Ph, and trace
amounts (∼8% from 4) of reduced cofactor 10. Both 10 and
12 were unstable with respect to routine chromatographic
isolation in air, but their structures were confirmed by isolation
and characterization of their acetylated derivatives, 13 (88%
yield) and 14 (trace), obtained by anaerobic quenching of the
reaction with Ac2O.
followed polarographically did not result in any increase in [O2],
indicating that H2O2 does not build up in the reaction mixture.
The finding that product Schiff base 12 is nearly the exclusive
fate of 4 in its anaerobic reaction with benzylamine provides
unambiguous evidence that model 4 reproduces the transami-
nation mechanism assumed to transpire for lysyl oxidase, as
opposed to an alternate addition-elimination mechanism (lead-
ing to PhCHdNH and quinol 10, Scheme 3, top). The
exclusivity of 12 is noteworthy in that the product Schiff base
for the analogous anaerobic reactions of benzylamine with TPQ
models was a minor component of a very complex equilibrium
product mixture reflecting redox and Schiff base interchanges.1c
Small amounts of the product Schiff base could alternatively
arise from the addition-elimination pathway by condensation
of PhCHdNH with the aminoresorcinol resulting from reduction
of quinone imine generated from the starting quinone and
released NH3. Thus, unambiguous confirmation of a transami-
nation mechanism in the TPQ model case required reliance on
more complex NMR kinetics arguments and the use of special
mechanistic probes.1c,16 In the present case, the predominance
of product Schiff base 12 obviates additional studies to confirm
a transamination mechanism. The high yield of 12 indicates that
its interception by benzylamine (giving PhCHdNCH2Ph) is
slower than its generation from 4 and benzylamine, unlike the
case for the TPQ-derived product Schiff base.15 The small
amount (∼8%) of the reduced LTQ (10) formed undoubtedly
arises not from a minor competing addition-elimination
pathway, but from reduction of 4 under anaerobic conditions
Both the single-turnover anaerobic and aerobic catalytic
turnover results can be understood in terms of Scheme 2.
Condensation of benzylamine at the most electrophilic C5
carbonyl of 4 and subsequent imine shift leads to product Schiff
base 12. Although 12 would undergo hydrolysis in enzyme-
catalyzed turnover to release PhCHdO, in the model turnover
reaction 12 is mainly trapped by the excess benzylamine to give
PhCHdNCH2Ph (which gives the same 2,4-dinitrophenylhy-
drazine derivative). In the enzyme reaction, the generated
aminophenol 15 undergoes oxidation to 16 accompanied by
reduction of O2 to H2O2. For the model aerobic turnover
reaction, 15 and any reduced cofactor 10 arising from reduction
of 4 by 15 undergoes O2-dependent autoxidation, as confirmed
by measuring O2 consumption with a Clark-type oxygen
electrode in a closed chamber. Consumption of O2 implicates
generation of H2O2 either directly or indirectly (via superoxide),
and since H2O2 is suspected to be a more reactive oxidant of
15 and/or 10 than O2, it is not expected to accumulate in
solution. Consistent with this suspicion, addition of catalase at
various times into the reaction of 4 with benzylamine being
(15) When the anaerobic reaction of 4 with benzylamine was conducted
in methanol, the reaction outcome was more complicated and was dominated
by formation of a phenoxazine, resulting from condensation of starting
quinone 4 with the aminophenol 15 generated according to Scheme 2.
Apparently in methanol but not acetonitrile, interception of product Schiff
base 12 by benzylamine is competitive with reaction of benzylamine with
4.
(16) Lee, Y.; Huang, H.; Sayre, L. M. J. Am. Chem. Soc. 1996, 118,
7241-7242.

Catalytic TurnoVer by a Lysyl Oxidase LTQ Model
J. Am. Chem. Soc., Vol. 123, No. 39, 2001 9609
by the low level of aminophenol 15 released when 12 is
intercepted by benzylamine (Scheme 3, bottom).
To find out the cause of catalyst destruction in the aerobic
turnover reactions, the reaction products were analyzed at the
first point of bleaching of the red color. The only isolable
product was the benzoxazole derivative 18, which presumably
arises from oxidation of the dihydrobenzoxazole 17 in equilib-
rium with the product Schiff base 12 (see Scheme 2). It was of
interest to determine what factors govern the partitioning of 12/
17 between productive turnover to 15 and irreversible oxidation
to 18, since the latter pathway must normally be suppressed in
the enzyme case. Such determination is closely tied to the
question of what serves as the oxidant of 17. Our finding that
running the aerobic catalytic turnover reaction in the presence
of catalase neither inhibited the formation of 18 nor extended
the lifetime of the catalyst indicates that oxidation of 17 is not
principally mediated by the H2O2 generated during turnover.
Another candidate for the oxidant is the starting quinone 4.
However, when the anaerobic model reaction (CD3CN) was
carried out using a deficiency (0.5 equiv) of benzylamine,
unreacted 4 was now present alongside the product Schiff base
12, and no 18 was detected even after several days. Thus, even
though the 12 / 17 equilibrium is unfavorable, as indicated by
Figure 2. Selected 13C-15N coupling data for 15N-labeled benzoxazole
18.
Scheme 4
1
nonobservance of 17 by H NMR spectroscopy, the failure of
12 + 4 to be converted to 18 over time indicates that the
responsible oxidant is not starting quinone 4.
constants are known to be large when the carbon atom is close
to the nitrogen lone pair.17,18 If the benzoxazole had structure
21 rather than 18, then the 15N should have been coupled to the
upfield rather than downfield benzoxazole aryl CH carbon.
Nonetheless, >90% formation of benzoxazole 18 was ob-
served when 12 (formed in situ in 92% yield from the reaction
of 4 and 1.5 equiv of benzylamine in CD3CN under argon) was
exposed to either O2 (slower) or H2O2 (faster). Thus the
oxidation of 17 is at least in part an O2-dependent autoxidation.
Although H2O2 is clearly a more competent oxidant, we believe
that the lack of effect of catalase on the formation of 18 must
reflect the fact that any H2O2 generated during turnover is
rapidly consumed in oxidizing 15, whereas 17 is always present
in very low concentration, but is eventually oxidized by the
plentiful O2.
Our assignment of the structure of substrate Schiff base 11
and all structures derived from it (e.g., 12 and 18) was predicated
on the assumption that benzylamine condenses with the more
electrophilic of the two possible quinone carbonyl groups of 4.
Nonetheless, our results could not unambiguously rule out the
possible alternative formation of substrate Schiff base 19,
Conclusions
We have shown that 4, a stable model for the LTQ cofactor
of lysyl oxidase, effects transaminative conversion of benzy-
lamine anaerobically in acetonitrile. Thus, despite the difference
in structures between the TPQ and LTQ nuclei, the two quinone
cofactors possess the same intrinsic capability of deaminating
amines by a transamination as opposed to the addition-
elimination mechanism. The LTQ model 4 also achieves a
catalytic aerobic deamination of benzylamine in buffered
aqueous acetonitrile. In the latter case, an optimal 3.4 turnovers
at pH 9 (borate buffer) is observed, on account of the
O2-dependent conversion of the catalyst to a benzylamine-
derived benzoxazole derivative. Although the enzyme must
normally avoid this side reaction, the same phenomenon was
observed previously for the reaction of benzylamine with TPQ
models.19 In this latter case, the benzoxazole-forming side
reaction is more than a chemical artifact in that it was shown
to be responsible for the benzylamine-dependent inactivation
of bovine plasma amine oxidase and other amine oxidases that
occurs when the byproduct H2O2 reaches significant concentra-
tions.19 It should be pointed out that syncatalytic inactivation
of lysyl oxidase has also been observed, and may represent a
regulatory mechanism for this enzyme in vivo.20 Although the
nature of the latter inactivation has not been determined,
benzoxazole formation may be a common physiologically
product Schiff base 20, and benzoxazole 21. To confirm the
structures as those shown in Scheme 2, we synthesized 15N-
labeled benzoxazole 18 by air exposure of Schiff base 12
generated anaerobically from [15N]benzylamine. The 15N in 18
was found to be coupled only to the downfield (113.1 ppm) of
the two CH carbons in the benzoxazole aryl ring, which is the
one meta to the aniline-like nitrogen (Figure 2). The upfield
CH carbon (91.3 ppm [91.4 ppm in [15N]-18]) is the one ortho
to the aniline-like nitrogen, as confirmed by the fact that this
CH is the one that undergoes base-induced solvent deuterium
exchange in the starting quinone 4, permitting synthesis of the
7-deuterio benzoxazole 18-d (Scheme 4), containing a weak
positive signal at 91.4 ppm in place of the negative APT 13C
NMR signal at 91.3 ppm in 18. Two bond 15N-13C coupling
(17) Levy, G. C.; Lichter, R. L. Nitrogen-15 nuclear magnetic resonance
spectroscopy; Wiley: New York, 1979.
(18) (a) Axenrod, T.; Mangiaracina, P.; Watnick, C. M.; Wieder, C. M.;
Bulusu, S. Org. Magn. Reson. 1980, 13, 197-199. (b) Boyd, D. R.; Stubbs,
M. E.; Thompson, N. J.; Yeh, H. J. C.; Jerina, D. M.; Wasylishen, R. E.
Org. Magn. Reson. 1980, 14, 528-533.
(19) Lee, Y.; Shepard, E.; Smith, J.; Dooley, D. M.; Sayre, L. M.
Biochemistry 2001, 40, 822-829.
(20) Kagan, H. M.; Soucy, D. M.; Zoski, C. G.; Resnick, R. J.; Tang,
S.-S. Arch. Biochem. Biophys. 1983, 221, 158-167.

9610 J. Am. Chem. Soc., Vol. 123, No. 39, 2001
Ling et al.
funnel packed with anhydrous Na2SO4 and NaCl. The funnel was further
washed with methanol (200 mL), and the combined dark red solution
was evaporated to dryness to yield a residue that was purified by silica
gel flash chromatography (EtOAc-acetone as eluent), affording model
compound 4 (6.22 g, 35%). The crude product was further purified by
crystallization in methanol-ethyl acetate to yield dark red microscopic
important mechanism that regulates the activity of the copper
amine oxidases.
Experimental Section
General Methods. Unless otherwise stated the solvents and reagents
were of commercially available analytical grade quality. Catalase
(20 000 units/mg) was from Sigma Chemical Co. (St. Louis, MO).
Reactions in aqueous CH3CN were carried out using Millipore purified
water, and all evaporations were carried out at reduced pressure with
a rotary evaporator. 1H NMR spectra (200 or 300 MHz) and 13C NMR
spectra (75 MHz) were recorded on Varian Gemini instruments. In all
cases, tetramethylsilane or the solvent peak served as an internal
standard for reporting chemical shifts, expressed on the δ scale;
Attached Proton Test (APT) designations for 13C NMR spectra are given
in parentheses. High-resolution mass spectra (HRMS) were obtained
at 20 eV on a Kratos MS-25A instrument. Optical spectra were obtained
with Perkin-Elmer model Lambda 3B or 20 spectrophotometers fitted
with a water-jacketed multiple cell holder for maintenance of constant
temperature.
3,3-Dimethyl-2,3-dihydroindole-5,6-quinone (4). (a) Methylation.
Liquid ammonia (500 mL) was collected in a 3 L three-neck round-
bottom flask cooled by a dry ice-acetone bath. To the mechanically
stirred solution was added ferric nitrate (0.1 g) and then 5.06 g (0.22
mol) of finely divided sodium, and the mixture was vigorously stirred
until the dark blue solution turned to gray. To this sodium amide
suspension was added within 30 min a clear solution of 3,4-
dimethoxyphenylacetonitrile (5, 35.4 g, 0.20 mol) in 1 L of ether. Some
of the nitrile 5 began to precipitate and the mixture was vigorously
stirred until all the nitrile 5 was dissolved. To the resultant orange
mixture was added a solution of 28.4 g (0.2 mol) of iodomethane in
50 mL of ether, and the reaction mixture was further stirred for 2 h.
After evaporation of ammonia, the solution was washed with water (3
× 500 mL) to remove NaI. The ether layer was separated, dried over
anhydrous Na2SO4, and evaporated to dryness to afford crude R-methyl-
3,4-dimethoxyphenylacetonitrile. The above methylation procedure was
repeated with 4.6 g (0.2 mol) of sodium and 28.4 g (0.2 mol) of
iodomethane, resulting in crude R,R-dimethyl-3,4-dimethoxyphenyl-
acetonitrile (6) as a light yellow oil which solidified on standing: 1H
NMR (CDCl3) δ 1.72 (s, 6H), 3.88 (s, 3H), 3.92 (s, 3H), 6.86 (d, 1H,
J ) 8.0 Hz), 6.98 (s, 1H), 6.99 (d, 1H, J ) 8.0 Hz).
(b) Reduction. To a solution of diethyl ether (1 L) in a 3 L three-
neck round-bottom flask fitted with a mechanical stirrer and a CaCl2-
drying-tube-protected condenser was added LiAlH4 (38 g, 1 mol). The
suspension was vigorously stirred while a solution of the crude nitrile
6 in 100 mL of ether was slowly added over 30 min, and the mixture
was further stirred for 2 h, then cooled in an ice bath, at which point
150 mL of ethyl acetate was carefully added to decompose any
unreacted LiAlH4. The mixture was then poured into a mixture of ice
and aqueous sodium tartrate with continuous stirring. The ether layer
was separated and the aqueous layer was further extracted with ether
(3 × 200 mL). The combined organic layer was dried (Na2SO4) and
evaporated to give crude 2-(3,4-dimethoxyphenyl)-2-methylpropylamine
(7) as a light yellow oil: 1H NMR (of HCl salt 7 in D2O) δ 1.31
(s, 6H), 3.14 (s, 2H), 3.73 (s, 3H), 3.77 (s, 3H), 6.89-6.94 (3H).
(c) Deprotection. To a 1 L flask were added the above prepared
crude amine 7 and 320 mL of 48% aqueous HBr. The mixture was
heated on a steam bath for 3 h and then evaporated to dryness in vacuo
at 60 °C. To the residue was added 500 mL of anhydrous ethanol, the
mixture was concentrated to 50 mL, and then 350 mL of ethyl acetate
was added. Cooling of the mixture resulted in deposition of the first
crop (23.24 g) of 2-(3,4-hydroxyphenyl)-2-methylpropylamine hydro-
bromide (8). The mother liquor was treated as above to give a second
crop of product 8 (3.68 g). The total product recovery corresponded to
an overall yield of 52% for the first three steps (from 5 to 8). Compound
8: mp 247-249 °C dec; 1H NMR (D2O) δ 1.26 (s, 6H), 3.08 (s, 2H),
6.81-6.87 (3H); 13C NMR (D2O) δ 40.6 (-), 51.0 (+), 65.5 (+), 128.9
(-), 131.1 (-), 133.4 (-), 151.2 (+), 157.4 (+), 158.7 (+).
1
needles: mp 160-161 °C; H NMR (DMSO-d6) δ 1.31 (s, 6H), 3.58
(s, 2H), 5.48 (br, s, 1H), 6.43 (s, 1H), 9.44 (br, s, 1H); 13C NMR (CD3-
OD) δ 27.4 (-), 41.3 (+), 63.9 (+), 93.8 (-), 124.7 (-), 164.3 (+),
164.7 (+), 184.7 (+); HRMS (EI) m/z calcd for C10H11NO2 177.0790,
found 177.0792 (relative intensity 82).
Catalytic Aerobic Deamination of Benzylamine Mediated by
LTQ Model 4. Quantitative analysis was conducted at pH 7-9 as
follows. A mixture of 7.5 mmol of benzylamine, 0.15 mmol of LTQ
model 4, and 0.75 mmol of sodium phosphate (pH 7 and 8) or sodium
borate (pH 9) buffer in 150 mL of 30% aqueous CH3CN was adjusted
to the desired pH with HCl. The solution was magnetically stirred
vigorously in an open 250 mL Erlenmeyer flask at 25 °C with moni-
toring of the pH, the reaction volume being maintained by periodic
addition of CH3CN, which evaporates somewhat at long reaction times.
Three 50 mL aliquots were worked up at different reaction times by
addition to each of 18 mL of standard 2,4-dinitrophenylhydrazine (150
mM) reagent (in 15 mL of H2SO4, 70 mL of EtOH, and 20 mL of
water). After being cooled to 0 °C for 1 h, the solution was filtered,
and the precipitate was dried to constant weight to obtain the yield of
the benzaldehyde 2,4-dinitrophenylhydrazone, the identity and purity
1
of which were confirmed by TLC and H NMR.
Spectral monitoring was conducted at pH 9 as follows. To a solution
of benzylamine (50 mM) in 25 mL of 5 mM sodium borate buffered
30% aqueous acetonitrile in an open 50 mL Erlenmeyer flask was added
4 (1 mM), and the mixture was stirred vigorously at 25 °C. At regular
time intervals, 0.3 mL aliquots of the reaction were transferred into 1
mm path length cuvettes and the absorbance was recorded at 456 nm.
Oxygen uptake was monitored in a closed chamber using a Yellow
Springs Instruments 5300 biological oxygen meter at 30 °C, equipped
with magnetic stirring and a Clark-type oxygen electrode. A solution
of benzylamine (10 mM final) in 10 mM pH 9 borate buffered 10%
aqueous acetonitrile (2.8 mL) was first equilibrated in the chamber
following insertion of the electrode. Reactions were initiated by injecting
200 µL of a 37.5 mM stock solution of 4 in CH3CN through the side
vent in the chamber, followed by commencement of monitoring of O2
consumption. When 50% of oxygen consumption was reached in the
reaction system, 50 µL of a solution of catalase (Sigma C-10 from
bovine liver, 2 mg/mL) was added to the reaction; no change in the O2
concentration was seen, indicating the absence of H2O2 at this point in
the reaction. Confirmation of the ability of catalase to release O2 from
H2O2 under these conditions was achieved by the addition of H2O2
(100 µM) following the catalase addition and noting the expected
increase in the O2 concentration. For these experiments the ambient
concentration of O2 was taken to be 223 µM.
Product Study for the Aerobic Deamination of Benzylamine
Catalyzed by LTQ Model 4. To a stirred solution of boric acid (124
mg, 2 mmol) in 380 mL of water were added 2.14 g (20 mmol) of
freshly distilled benzylamine and 120 mL of acetonitrile. The solution
was adjusted to pH 9.00 using aqueous HCl. To this buffered solution
was added 70.8 mg (0.4 mmol) of 4. The resultant dark red solution
was continuously stirred for 24 h in an open flask (the color turned
from red to orange after 5 h and then to yellow), at which time the pH
had dropped to 8.5. The acetonitrile was removed in vacuo at 30 °C
and the aqueous solution was extracted with dichloromethane (3 ×
100 mL). The combined organic layer was dried (Na2SO4) and
evaporated, and the residue was subjected to silica gel flash chromato-
graphic separation (eluant hexanes-EtOAc), affording as the only
isolable 4-derived product benzoxazole 18 (23.3 mg, 22%), which was
crystallized from hexanes-acetone as colorless plates: mp 133-135
1
°C; H NMR (CDCl3) δ 1.36 (s, 6H), 3.39 (s, 2H), 3.93 (br s, 1H,
NH), 6.76 (s, 1H), 7.36 (s, 1H), 7.40-7.60 (3H), 8.20 (m, 2H); 13C
NMR (CDCl3) δ 27.9 (-), 41.4 (+), 62.2 (+), 91.3 (-), 113.1 (-),
126.8 (-), 127.9 (+), 128.8 (-), 130.5 (-), 134.9 (+), 136.9 (+),
149.1 (+), 151.2 (+), 160.8 (+); HRMS (EI) calcd for C17H16N2O
264.1263, found 264.1267. The aqueous layer was divided into two
(d) Oxidation. To a vigorously stirred solution of 8 (26.03 g, 0.099
mol) in 1 L of methanol was added 93 g (0.4 mol) of Ag2O powder.
The mixture was stirred for 8 min and then passed through a Buchner

Catalytic TurnoVer by a Lysyl Oxidase LTQ Model
J. Am. Chem. Soc., Vol. 123, No. 39, 2001 9611
portions. One was adjusted to pH 11 and the other to pH 2. Both
solutions were extracted again with dichloromethane. TLC analysis
revealed that the alkaline extract mainly contained unreacted benzyl-
amine while the acidic extract contained a mixture of strongly polar
intractable materials. Benzoxazole 18 gradually decomposed in aerobic
solution to give intractable products, probably accounting in part for
the low yield isolated from the deaminative turnover reaction.
Anaerobic Reaction of LTQ Model 4 with Benzylamine in
Acetonitrile. A suspension of 4 (9 mg, 0.05 mmol) in CD3CN (0.07
mL) in an NMR tube was bubbled with dry argon for 30 min before
sealing the tube with a septum. Degassed freshly distilled benzylamine
(7.9 µL, 0.075 mmol) was added by syringe. The mixture was allowed
to stand at room temperature until the color turned from dark brown to
acetonitrile under argon followed by treatment with degassed Ac2O/
triethylamine gave 14 in 96% yield as colorless needles from hexanes-
ethyl acetate: mp 154-156 °C; 1H NMR (CDCl3) δ 1.36 (s, 6H), 2.20
(s, 3H), 2.26 (s, 3H), 2.28 (s, 3H), 3.80 (s, 2H), 6.93 (s, 1H), 8.03 (s,
1H); 13C NMR (CDCl3) δ 20.6 (-), 20.7 (-), 24.0 (-), 28.7 (-), 40.3
(+), 63.9 (+), 112.1 (-), 116.6 (-), 138.2 (+), 138.6 (+), 139.6 (+),
141.0 (+), 168.4 (+), 168.5 (+), 168.6 (+); HRMS (EI) calcd for
C16H19NO5 305.1263, found 305.1263 (relative intensity 20).
Preparation of 7-d-benzoxazole 18. To a clear solution of 44.3
mg (0.25 mmol) of 4 in 2 mL of CD3OD was added 0.2 mL of
triethylamine, and the mixture was stirred for 10 min with warming to
40 °C, when 1H NMR showed that the 7-H in 4 was completely
exchanged with deuterium. The solvent was removed in vacuo and the
residue was dried thoroughly. The resultant 7-d-4 was suspended in 4
mL of CD3CN-CD3OD (5:1, v/v), and dry argon was bubbled through
the mixture for 10 min while deuterium-exchanged benzylamine (C6H5-
CH2ND2, 39 µL, 0.38 mmol) was added by syringe. The mixture was
further bubbled with dry argon for 20 min and then sealed. After 8 h,
the mixture was purged with dry oxygen for 5 min, sealed, and allowed
to stand for 2 h. Evaporation of the solvent afforded a dark brown
residue that was subjected to silica gel flash chromatography (hexanes-
EtOAc as eluent) to afford 7-d-18 (40.3 mg, 61%) as colorless
needles: 1H NMR (CDCl3) δ 1.38 (s, 6H), 3.42 (s, 2H), 3.5 (br, s, 1H,
NH), 7.37 (s, 1H), 7.46-7.50 (m, 3H), 8.14-8.19 (m, 2H); 13C NMR
(CDCl3) δ 27.9 (-), 41.4 (+), 62.2 (+), 91.4 (+), 113.1 (-), 126.9
(-), 127.9 (+), 128.8 (-), 130.5 (-), 135.1 (+), 137.0 (+), 148.7
(+), 151.1 (+), 160.9 (+).
Preparation of [3-15N]-Benzoxazole 18. A suspension of 44.3 mg
(0.25 mmol) of 4 in 4 mL of CD3CN was bubbled with dry argon for
10 min while 15N-benzylamine (Isotec, >99% 15N, 39 µL, 0.38 mmol)
was added by syringe. The mixture was further bubbled with dry argon
for 20 min and then sealed. After 12 h, the mixture was purged with
dry oxygen for 5 min, sealed, and allowed to stand for 2 h. Workup as
for 7-d-18 above afforded [3-15N]-18 (54.1 mg, 83%) as colorless
needles: 1H NMR (CDCl3) δ 1.38 (s, 6H), 3.43 (s, 2H), 3.7 (br, s, 1H,
NH), 6.83 (s, 1H), 7.37 (s, 1H), 7.4-7.5 (m, 3H), 8.1-8.2 (m, 2H);
13C NMR (CDCl3) δ 27.9 (-), 41.4 (+), 62.2 (+), 91.4 (+), 113.1 (-)
(d, J ) 6.2 Hz), 126.8 (-) (d, J ) 1.9 Hz), 127.8 (+) (d, J ) 5.8 Hz),
128.8 (-), 130.5 (-), 135.0 (+), 137.0 (+) (d, J ) 2.0 Hz), 150.0
(+), 151.2 (+), 160.8 (+) (d, J ) 2.7 Hz).
1
bright yellow (12 h). H NMR showed that benzylamine was nearly
consumed and a mixture of Schiff base 12 (92% from 4 by integration)
and PhCHdNCH2Ph was obtained along with traces of 10 (8% from
1 by integration). Compound 12: 1H NMR (CD3CN) δ 1.28 (s, 6H),
3.28 (s, 2H), 6.16 (s, 1H), 7.26 (s, 1H), 7.2-7.5 (3H, overlapped with
those of PhCHdNCH2Ph and benzylamine), 7.94 (dd, 2H, J ) 7.68,
2.19 Hz), 8.70 (s, 1H). The appearance of only one singlet for CHdN
indicated that only one isomer of 12 (probably anti) was formed.
PhCHdNCH2Ph (syn and anti isomers): 1H NMR (CD3CN) δ 4.77
and 4.78 (2s, 2H total), 7.2-7.5 (8H total, overlapped with those of
12 and benzylamine), 7.75-7.79 (2H total), 8.46 and 8.47 (2s, 1H total).
Compound 10: 1H NMR (CD3CN) δ 1.18 (s, 6H), 3.12 (s, 2H), 6.14
(s, 1H), 6.51 (s, 1H).
1
To confirm the above H NMR assignment, a mixture of 44.3 mg
(0.25 mmol) of 4 in acetonitrile (0.35 mL) in an NMR tube was
degassed and sealed. Then 39.5 µL (0.375 mmol) of degassed freshly
distilled benzylamine was slowly added by syringe. After 12 h, 0.25
mL of degassed Ac2O and 0.5 mL of degassed triethylamine were
added, and the mixture was allowed to stand for 12 h. The resultant
mixture was partitioned between water and dichloromethane (50 mL
each). The organic layer was separated, washed with water (2 × 20
mL), dried (Na2SO4), and evaporated, and the residue was subjected
to silica gel flash chromatography (hexanes-EtOAc) to afford diacetyl
derivative 13 (77.2 mg, 88%) as yellow prisms from hexanes-ethyl
1
acetate: mp 218-220 °C; H NMR (CDCl3) δ 1.39 (s, 6H), 2.21 (s,
3H), 2.29 (s, 3H), 3.80 (s, 2H), 6.94 (s, 1H), 7.44-7.47 (3H), 7.86 (m,
2H), 8.01 (s, 1H), 8.45 (s, 1H, CHdN); 13C NMR (CDCl3) δ 20.7 (-),
24.1 (-), 28.7 (-), 40.3 (+), 63.9 (+), 111.7 (-), 112.5 (-), 128.8
(-), 131.5 (-), 136.3 (+), 138.8 (+), 139.7 (+), 140.3 (+), 143.9
(+), 159.9 (-), 168.6 (+), 169.5 (+); HRMS (EI) calcd for C21H22N2O3
350.1630, found, 350.1627 (rel intensity 49). A trace amount of the
triacetyl derivative 14 was also observed, as verified by synthesis of
an authentic sample. Thus, reduction of 4 with sodium borohydride in
Acknowledgment. We thank the NIH for support of this
work through grant GM 48812.
JA011141J