Oxidations of N-(3-indoleethyl) cyclic aliphatic amines by horseradish peroxidase: The indole ring binds to the enzyme and mediates electron-transfer amine oxidation

Article abstract of DOI:10.1021/ja075905s

Although oxidations of aromatic amines by horseradish peroxidase (HRP) are well-known, typical aliphatic amines are not substrates of HRP. In this study, the reactions of N-benzyl and N-methyl cyclic amines with HRP were found to be slow, but reactions of N-(3-indoleethyl) cyclic amines were 2-3 orders of magnitude faster. Analyses of pH-rate profiles revealed a dominant contribution to reaction by the amine-free base forms, the only species found to bind to the enzyme. A metabolic study on a family of congeneric N-(3-indoleethyl) cyclic amines indicated competition between amine and indole oxidation pathways. Amine oxidation dominated for the seven- and eight-membered azacycles, where ring size supports the change in hybridization from sp³ to sp² that occurs upon one-electron amine nitrogen oxidation, whereas only indole oxidation was observed for the six-membered ring congener. Optical difference spectroscopic binding data and computational docking simulations suggest that all the arylalkylamine substrates bind to the enzyme through their aromatic termini with similar binding modes and binding affinities. Kinetic saturation was observed for a particularly soluble substrate, consistent with an obligatory role of an enzyme-substrate complexation preceding electron transfer. The significant rate enhancements seen for the indoleethylamine substrates suggest the ability of the bound indole ring to mediate what amounts to medium long-range electron-transfer oxidation of the tertiary amine center by the HRP oxidants. This is the first systematic investigation to document aliphatic amine oxidation by HRP at rates consistent with normal metabolic turnover, and the demonstration that this is facilitated by an auxiliary electron-rich aromatic ring.

Full text of DOI:10.1021/ja075905s

Published on Web 12/29/2007
Oxidations of N-(3-Indoleethyl) Cyclic Aliphatic Amines by
Horseradish Peroxidase: The Indole Ring Binds to the
Enzyme and Mediates Electron-Transfer Amine Oxidation
Ke-Qing Ling,* Wen-Shan Li,† and Lawrence M. Sayre*
Department of Chemistry, Case Western ReserVe UniVersity, CleVeland, Ohio 44106
Received August 6, 2007; E-mail: kxl56@case.edu; lms3@case.edu
Abstract: Although oxidations of aromatic amines by horseradish peroxidase (HRP) are well-known, typical
aliphatic amines are not substrates of HRP. In this study, the reactions of N-benzyl and N-methyl cyclic
amines with HRP were found to be slow, but reactions of N-(3-indoleethyl) cyclic amines were 2-3 orders
of magnitude faster. Analyses of pH-rate profiles revealed a dominant contribution to reaction by the amine-
free base forms, the only species found to bind to the enzyme. A metabolic study on a family of congeneric
N-(3-indoleethyl) cyclic amines indicated competition between amine and indole oxidation pathways. Amine
oxidation dominated for the seven- and eight-membered azacycles, where ring size supports the change
in hybridization from sp3 to sp2 that occurs upon one-electron amine nitrogen oxidation, whereas only indole
oxidation was observed for the six-membered ring congener. Optical difference spectroscopic binding data
and computational docking simulations suggest that all the arylalkylamine substrates bind to the enzyme
through their aromatic termini with similar binding modes and binding affinities. Kinetic saturation was
observed for a particularly soluble substrate, consistent with an obligatory role of an enzyme-substrate
complexation preceding electron transfer. The significant rate enhancements seen for the indoleethylamine
substrates suggest the ability of the bound indole ring to mediate what amounts to medium long-range
electron-transfer oxidation of the tertiary amine center by the HRP oxidants. This is the first systematic
investigation to document aliphatic amine oxidation by HRP at rates consistent with normal metabolic
turnover, and the demonstration that this is facilitated by an auxiliary electron-rich aromatic ring.
Introduction
peroxidases utilize H2O2 (or ROOH) to produce the two-electron
oxidized species compound I, usually depicted as a Fe(IV)dO
porphyrin radical cation (abbreviated P+.Fe(IV)dO). Although
this is the same active oxidant generated in cytochrome P450
(usually from O2 and two-electron reduction), HRP is usually
considered incapable of either O-transfer or H‚ abstraction
chemistry on account of the inaccessibility of substrates to the
“buried” iron center.4 Reduction of compound I by 1e gives
compound II, PFe(IV)dO, which in turn is reduced by 1e to
the Fe(III) resting state of the enzyme.5 Although compound I
appears to be more reactive than compound II,4 the redox
potentials have been estimated to be quite similar,6 and observed
differences may be due to the different nature of the two oxi-
dants (the porphyrin highest occupied molecular orbital (HOMO)
“hole” for compound I and the Fe(IV) for compound II).
Recent computational, structural, and spectroscopic studies
are revealing details of the factors which control reactivity
properties of heme enzymes of the cytochrome P450 and
peroxidase superfamilies.1 In contrast to those P450 enzymes
that catalyze oxygen-transfer, electron-transfer, or H-atom
transfer mechanisms,2 oxidations by horseradish peroxidase
(HRP) uniformly involve electron-transfer, and its substrates
are most frequently electron-rich aromatics.3 HRP and other
† Current address: Institute of Chemistry, Academia Sinica, 128 Aca-
demia Road Sec. 2, Nankang Taipei 115 Taiwan, Republic of China.
(1) (a) Hersleth, H. P.; Ryde, U.; Rydberg, P.; Gorbitz, C. H.; Andersson, K.
K. J. Inorg. Biochem. 2006, 100, 460-76. (b) Behan, R. K.; Hoffart, L.
M.; Stone, K. L.; Krebs, C.; Green, M. T. J. Am. Chem. Soc. 2006, 128,
11471-4. (c) Harvey, J. N.; Bathelt, C. M.; Mulholland, A. J. J. Comput.
Chem. 2006, 27, 1352-62. (d) Green, M. T.; Dawson, J. H.; Gray, H. B.
Science 2004, 304, 1653-6.
(2) For recent reviews on the mechanisms of cytochrome P450 oxidations,
see: (a) Isin, E. M.; Guengerich, F. P. Biochim. Biophys. Acta 2007, 1770,
314-29. (b) Meunier, B.; de Visser, S. P.; Shaik, S. Chem. ReV. 2004,
104, 3947-80. (c) Makris, T. M.; Davydov, R.; Denisov, I. G.; Hoffman,
B. M.; Sligar, S. G. Drug Metab. ReV. 2002, 34, 691-708. For recent reports
on cytochrome P450 catalyzed amine oxidations, see: (d) Cerny, M. A.;
Hanzlik, R. P. J. Am. Chem. Soc. 2006, 128, 3346-54. (e) Jurva, U.; Bissel,
P.; Isin, E. M.; Igarashi, K.; Kuttab, S.; Castagnoli, N., Jr. J. Am. Chem.
Soc. 2005, 127, 12368-77. (f) Shaffer, C. L.; Harriman, S.; Koen, Y. M.;
Hanzlik, R. P. J. Am. Chem. Soc. 2002, 124, 8268-74.
On the basis that at least some amines are oxidized by P450
by electron transfer, as modeled by electrochemical oxidations
(4) (a) Ator, M. A.; Ortiz de Montellano, P. R.; J. Biol. Chem. 1987, 262,
1542-51. (b) Ortiz de Montellano, P. R.; Annu. ReV. Pharmacol. Toxicol.
1992, 32, 89-107.
(3) (a) Griffin, B. W.; Davis, D. K.; Bruno, G. V. Bioorg. Chem. 1981, 10,
342-55. (b) Van, der Zee, J.; Duling, D. R.; Mason, R. P.; Eling, T. E. J.
Biol. Chem. 1989, 264, 19828-36. (c) Shaffer, C. L.; Morton, M. D.;
Hanzlik, R. P. J. Am. Chem. Soc. 2001, 123, 8502-8. (d) Shaffer, C. L.;
Morton, M. D.; Hanzlik, R. P. J. Am. Chem. Soc. 2001, 123, 349-50.
(5) (a) Dunford, H. B. In Peroxidases in Chemistry and Biology; Everse, J.,
Everse, K. E., Grisham, M. B., Eds.; CRC Press: Boca Raton, FL, 1991;
Vol. 2, pp 1-24. (b) Dunford, H. B. Heme Peroxidases; John Wiley &
Sons, Inc.: New York, 1999.
(6) Hayashi, Y.; Yamazaki, I. J. Biol. Chem. 1979, 254, 9101-6.
9
10.1021/ja075905s CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 933-944
933

A R T I C L E S
Scheme 1
Ling et al.
of the Vinca alkaloid and studied their reactions with HRP
compounds I and II in comparison with the corresponding cyclic
benzylamines 5-8. Indoles 9 and 10 were included as possible
of amines,7 there has been interest in the plant enzyme HRP as
a relatively pure electron-transfer heme oxidant to learn about
one-electron oxidation mechanisms important in mammalian
metabolism. In this regard, a central question has been why
aliphatic amines appear to be unreactive toward HRP. One
consideration is that only amine “free bases” are subject to
oxidation and that typical aliphatic amines are protonated at
physiological pH. Thus, whereas a favorable Kd for binding of
free base amine to P450 can offset an unfavorable pKa
equilibrium, HRP is known to lack a “normal” substrate binding
domain and is thus limited to electron-transfer oxidation of its
typical aromatic donor substrates through their fairly weak 1:1
complex association at the “heme crevice.”4,8 Therefore, if HRP
is limited to oxidizing those aliphatic amines which have
sufficiently low pKa values to exist at least partly in the free
base form in solution at physiological pH, it must then be
recognized that structural factors that lower pKa at the same
time increase redox potential. Examples are the typical amine-
based “biological buffers” HEPES, PIPES, and MES, which
have pKa values near 7 but are not oxidized by HRP under
turnover conditions.9 Thus, the non-oxidation of aliphatic amines
by HRP may reflect the fact that amines with a sufficiently low
pKa have a redox potential too high for the “reach” of HRP.10,11
According to this pKa-E° dichotomy, the optimal amines to
test as candidate substrates would be those whose structures
confer a low redox potential without raising pKa. Examples
would be azacycles with five-, seven-, and eight-, but not six-
membered rings, where the geometry favors the transformation
of tetrahedral sp3 nitrogen to sp2 trigonal nitrogen that occurs
upon 1e oxidation to the aminium radical,12 but where basicity
is essentially constant. However, a preliminary study showed
that simple azacycles fail as turnover substrates of HRP.13
Nonetheless, on the basis that HRP can at least stoichiomet-
rically oxidize the tertiary amine center in certain indole (Vinca)
alkaloids to the corresponding iminium species (Scheme 1),14
we considered the possibility that aliphatic amine oxidation may
be facilitated in these molecules by interaction of the indole
ring with the usual aryl ring binding domain of HRP.
“controls” for the reactions of 1-4. Our results show that the
indole ring significantly enhances the HRP-mediated oxidation
of aliphatic amine centers with a favorable pKa-E° balance,
and we suggest that this reflects the action of the indole ring
not only in promoting binding to the HRP “heme crevice” but
also in mediating the electron-transfer oxidation of the amine
nitrogen. To our knowledge, this outcome is the first fulfillment
of the criteria needed for observation of aliphatic amine
oxidation by HRP at rates consistent with normal metabolic
turnover.
Results
Kinetics. Reductions of HRP compounds I and II by a given
substrate (1-10) were studied separately in aqueous buffer by
monitoring spectrophotometrically the formation of the corre-
sponding enzyme products, compound II and native HRP,
respectively, as described previously.15 Since no turnover was
involved, the reaction can be illustrated as
Kadpp
kcat
E + S w x ES 8 E′ + product
where E represents HRP compound I or compound II, E′
represents HRP compound II or native HRP, and ES represents
the enzyme substrate complex. The rate expression can be
deduced as
kcat
Kadpp + [S]
rate )
[E0][S]
(1)
where [E0] represents the total concentration of free and
substrate-bound HRP compound I or compound II, [S] is the
substrate concentration, kcat and Kadpp are the decomposition
rate constant and the apparent dissociation constant of the
substrate enzyme complex. Using an excess of substrate with
respect to the enzyme, the reaction can be treated according to
pseudo-first-order kinetics
To test this hypothesis, we synthesized a series of cyclic
3-indoleethylamines 1-4 bearing the key tryptamine fragment
(7) (a) Sasaki, J. C.; Fellers, R. S.; Colvin, M. E. Mutat. Res. 2002, 506-507,
79-89. (b) Guengerich, F. P.; Yun, C. H.; Macdonald, T. L. J. Biol. Chem.
1996, 271, 27321-9.
(8) (a) Fidy, J.; Paul, K.-G.; Vanderkooi, J. M. Biochemistry 1989, 28, 7531-
41. (b) Sukurada, J.; Takahashi, S.; Hosoya, T. J. Biol. Chem. 1986, 261,
9657-62.
rate ) kobsd[E0]
(2)
and the expression of observed rate constant is
(9) Sayre, L. M.; Naismith, R. T., II; Bada, M. A.; Li, W.-S.; Klein, M. E.;
Tennant, M. D. Biochim. Biophys. Acta 1996, 1296, 250-56.
(10) Guengerich, P.; MacDonald, T. L. FASEB J. 1990, 4, 2453-9.
(11) Wood, P. M. Biochem. Soc. Trans. 1992, 20, 349-52.
(12) (a) Lindsay Smith, J. R.; Mead, L. A. V. J. Chem. Soc., Perkin Trans. 2
1973, 206-10. (b) Lindsay Smith, J. R.; Masheder, D. J. Chem. Soc., Perkin
Trans. 2 1976, 47-51. (c) Lindsay Smith, J. R.; Masheder, D. J. Chem.
Soc., Perkin Trans. 2 1977, 1732-6.
kcat[S]
Kadpp + [S]
kobsd
)
(3)
Due to limited solubility of 1-10 in aqueous buffer, all the
kinetic studies were conducted using micromolar level of
substrates. Examples of pseudo-first-order plots are given in
Figures 1 and 2. The dependence of kobsd on substrate concen-
(13) Li, W.-S. Ph.D. Thesis, Case Western Reserve University, Cleveland, OH,
1997.
(14) (a) Elmarakby, S. A.; Duffel, M. W.; Goswami, A.; Sariaslani, F. S.;
Rosazza, J. P. N. J. Med. Chem. 1989, 32, 674-9. (b) Goswami, A.;
Macdonald, T. L.; Hubbard, C.; Duffel, M. W.; Rosazza, J. P. N. Chem.
Res. Toxicol. 1988, 1, 238-42.
(15) Ling, K.-Q.; Sayre, L. M. Bioorg. Med. Chem. 2005, 13, 3543-51.
9
934 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Oxidation of N-(3-Indoleethyl) Cyclic Aliphatic Amines
A R T I C L E S
Figure 1. Left panel: time-dependent ∆A420 (compound II) recorded during reactions of compound I with 1-4 in 0.1 M pH 9 borate buffer at 25 °C.
Right panel: pseudo-first-order ln[(A∞ - A0)/(A∞ - At)] vs t plots. [HRP] ) 0.7 µM; [H2O2] ) 7µM; [1], [2], [3], and [4] ) 3.5, 14, 28, and 28 µM,
respectively.
Figure 2. Left panel: time-dependent UV-vis spectra recorded during reaction of compound II with 2 in 0.1 M pH 9 borate buffer at 25 °C (time interval
45 s). Right panel: pseudo-first-order ln[(A∞ - A0)/(A∞ - At)] vs t plots for compound II oxidations of 1-4. [HRP] ) 2.1 µM; [H2O2] ) 2.1µM; [1-4]
) 21 µM.
tration in the micromolar range was examined with selected
organic substrates at different pH values. Consistent with the
millimolar Kadpp of these substrates with HRP as shown below,
no saturation behavior was observed under these conditions (an
example is given in Figure S2). Thus, eq 3 can be simplified as
for reduction of compound I by indole 9 remained constant (data
not shown), and the mean is calculated as the real second-order
rate constant (Table 1). For amine substrates, however, the amine
ionization plays a role in the net reaction, and one must consider
the possible reactions of compound I with both conjugate acid
and base:
kcat
kobsd
)
[S] ) kapp[S]
(4)
Ka
Kadpp
HS+ y
\
z S + H+
E(I) + S 9
k8 E(II) + product
where kapp is the apparent second-order rate constant. Therefore,
the kapp of each reaction can be extracted directly from kobsd
according to eq 4.
E(I) + HS+ 9+8 E(II) + product
k
pH-Rate Profiles. The pH dependence of the apparent
second-order rate constants (kapp) was examined for all the
substrates in the range of pH 8-10 to obtain real second-order
rate constants independent of the various ionization events in
the enzyme and in substrates. The limited pH range studied was
because at pH < 8, the rates for most amine substrates were
too slow to be measured accurately, whereas at pH > 10, the
enzyme became exceedingly unstable.
The kinetic expression of the apparent second-order rate constant
is deduced as
k
k+
Kapp
[Η+]
kapp
)
+
(5)
[Η+]
Kapp
1 +
1 +
Consistent with the lack of ionization for compound I in the
where Kapp is the apparent acid dissociation constant of amine
and k and k+ denote the real second-order rate constants of
reductions by amine-free base (S) and conjugate acid (HS+)
range of pH 8-10,16 the apparent second-order rate constants
(16) Ralston, I.; Dunford, H. B. Can. J. Biochem. 1978, 56, 1115-9.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 935

A R T I C L E S
Ling et al.
Table 1. Experimental Conjugate Acid Dissociation Constants,
Redox Potentials, Second-Order Enzyme Reduction Rate
Constants, and Apparent Conjugate Acid Dissociation Constants
for Tertiary Amine Substrates 1-8 and Indoles 9 and 10a
combinations of enzyme and amine species contribute to the
net reaction:
KE
\
HE(II)+ y
z E(II) + H+
HRP compound Id
HRP compound IIe
b
1
1
substrate
pKa
E
pa (V)c
k (M-1 s-
)
pKapp
k
1 (M-1 s-
)
Ka
HS+ y
\
z S + H+
1
2
3
4
5
6
7
8
9
9.60
9.59
9.22
9.18
9.12
9.19
9.21
9.39
0.54
0.55
0.74
0.66
0.70
0.72
0.89
0.79
1.26g
1.24g
4.7 × 105
1.2 × 105
2.5 × 104
1.8 × 104
1.3 × 103
3.0 × 102
9
9.7
9.8
9.5
9.5
9.1
9.2
9.3
9.3
1.0 × 104
2.2 × 103
2.0 × 102
1.5 × 102
14
k1
9
HE(II)+ + S
8 HE+ + Product
k2
24
HE(II)+ + HS+ 98 HE+ + Product
∼1.2f
69
7.1
3.3 × 103
4.6 × 102
k3
2.83 × 105 h
2.93 × 104 i
E(II) + S
98 E + Product
10
k4
a Derived rate constants of the free base forms of amines 1-8. b Measured
by potentiometric pH titration in 50% aqueous methanol. c Measured by
cyclic voltammetry vs SCE: 3 mM substrates, 100 mM NaClO4 in CH3CN,
glassy carbon, scan rate ) 100 mV/sec. d The k and pKapp values are
obtained from least-squares fits of pH-rate profiles according to eq 5. e The
k1 values were obtained from least-squares fits according to eq 6 for 9 or
according to eq 7 for 1-8. f Estimated using kapp at pH 9.00 according to
eq 7. g From ref 18, vs SCE under slightly different conditions. h From ref
15. i Estimated from the kapp at pH 10.0 (ref 15) according to eq 6.
E(II) + HS+ 98 E + Product
Rate constant k4 represents reaction of compound II with amine
substrate, both in their inactiVe forms (E(II) and HS+), which
is kinetically indistinguishable with k1 and therefore can be
omitted from the calculation. The kinetic expression of the
apparent second-order rate constant is then deduced as
k1
k2
forms, respectively. Nonlinear least-squares fits of pH-rate
profiles to eq 5 yielded apparent k+ values that were both
positive or negative but with absolute values that were at least
2 orders of magnitude smaller than k in all cases, suggesting
that these values are experimental artifacts of the fitting routine,
thus indicating that reductions of compound I by the conjugate
acid forms of amine substrates are negligible. The pH-rate
profiles were thus analyzed with the omission of the second
term in eq 5 as exemplified with amines 1-8 in Figure 3. The
real second-order rate constants (k) of the free base amine
substrates and the apparent conjugate acid amine dissociation
constants (pKapp) are listed in Table 1, along with the experi-
mental pKa values of these amine substrates measured in 50%
aqueous methanol by potentiometric pH titration. The good
agreement between calculated and experimentally determined
pKa values in MeOH-H2O probably reflects at least in part
that the operational dielectric of the enzyme active site is closer
to 50% aqueous methanol than to pure water.
For compound II reactions, it is known that dissociation of
the protonated distal His 42 affects the enzyme activity, with
only the conjugate acid form presumably being active.17 Indeed,
the reduction rate of compound II by neutral compound 9
decreased with increasing pH in the range of pH 8-10. A
nonlinear least-squares fit (Figure S3) according to eq 6 afforded
the real second-order rate constant (k1, Table 1), along with a
pKE value of 8.7, the same as the literature reported value.17
kapp
)
+
+
[Η+]
Kapp
KΕ
[Η+]
Kapp
[Η+]
KΕ
[Η+]
1 +
1 +
1 +
1 +
(
)(
)
(
)(
)
k3
(7)
[Η+]
Kapp
[Η+]
KΕ
1 +
1 +
(
)(
)
where k1 and k2 represent the second-order rate constants for
reactions of the protonated form of compound II with free base
and conjugated acid forms of amine substrate, whereas k3 is
the second-order rate constant for reaction of the deprotonated
form of compound II with the free base form of amine substrate.
To simplify data analysis, all the nonlinear least-squares fits
were conducted with the fixed pKE value of 8.7 and pKapp values
obtained from reductions of compound I by the same amines
(Table 1), the most relevant set of operational pKa values that
would reflect actual variations in microenvironment. Analysis
of pH-rate profiles according to eq 7 showed that k3 is negligible
for the most reactive 1, whereas for the slowly oxidizing 2-4,
k1, k2, and k3 all must be taken into account to achieve good
fits, though k1 is the major contributor in all cases (Supporting
Information). By contrast, for cyclic benzylamines 5, 6, and 8,
k2 and k3 are both negligible, so that reduction of the protonated
form of compound II by the free base amine form (k1) is the
only contributor. It should be noted that reduction of compound
II by 7 was too slow to allow for construction of a pH profile.
Assuming this reaction follows the same pH dependence as 5,
6, and 8, k1 for 7 was estimated from the kapp at pH 9.0 (0.13
M-1 s-1) according to eq 7. Typical pH-rate profiles and
nonlinear least-squares fits for compound II reductions are
shown in Figure 4, and the rate constants k1 are listed in Table
1. Values of k2 and k3 for 1-4 are listed in Table S1, along
with a proposed rationale for these minor rate constants.
Metabolism. Following the determination of rates of oxida-
tions of 1-4 by compounds I and II, we next verified that these
more reactive substrates could be metabolized under turnover
conditions. Since both the indole nucleus and the tertiary amine
k1
kapp
)
(6)
KΕ
1 +
[Η+]
For reduction of compound II by amine substrates 1-8, one
must take into account the ionization states of both enzyme
and substrates. Theoretically, considering a possible low activity
of the neutral His 42 form of compound II, four possible
(17) Dunford, H. B.; Adeniran, A. J. Arch. Biochem. Biophys. 1986, 251, 536-
42.
9
936 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Oxidation of N-(3-Indoleethyl) Cyclic Aliphatic Amines
A R T I C L E S
Figure 3. The pH-rate profiles for reduction of compound I by cyclic 3-indoleethylamines 1-4 (left panel) and cyclic benzylamines 5-8 (right panel) and
the corresponding nonlinear least-squares fit curves.
Figure 4. The pH-rate profiles for reductions of HRP compound II by cyclic 3-indoleethylamines 1-4 (left panel) and cyclic benzylamines 5, 6, and 8
(right panel) and the corresponding nonlinear least-squares fit curves.
moieties in 1-4 are susceptible to oxidation, it was expected
that products reflecting both oxidation pathways would be found,
and that such analysis could help deduce the regiochemistry of
the operational initial 1e oxidations. To ensure that metabolism
proceeds exclusively through the normal compound I/compound
II catalytic cycle (excluding participation by compound III), a
low ratio of H2O2/HRP was maintained in all reactions.19
Compound 2 was chosen for delineation of the metabolic
profile (Scheme 2). Incubation of 2 (800 µM) with HRP (40
µM) and H2O2 (200 µM) in pH 8.0 phosphate buffer resulted
in a fairly clean reaction mixture containing three major
metabolites as assessed by reverse phase HPLC-UV detection
at 290 nm (Figure 5). Fractions of each of these products were
collected and subjected to MS analyses. On the basis of their
ESI-MS and tandem MS fragmentation modes, the first and the
second fast-eluting metabolites were assigned as 3-hydroxyin-
dolenine 11 and aminoaldehyde/carbinolamine 12/13 (Scheme
2 and Supporting Information). The slow-eluting (less polar)
metabolite had poor sensitivity to ESI-MS, but could be isolated
in a preparative scale reaction and characterized as indole-3-
carboxaldehyde (14). The structure of 11 was confirmed by its
ready conversion to authentic oxindole 15 upon heating with
trifluoroacetic acid (Figure S6), a typical reaction of 3-hydroxy-
indolenines.20 The structure of 12/13 was confirmed by com-
parative HPLC and ESI-MS analyses with an authentic sample
obtained by deprotection of the diBoc-protected aldehyde 16,
in turn obtained by DIBAL reduction of the ester 17 (Supporting
Information).
The 3-hydroxyindolenine 11 represents the O2-dependent
indole oxidation (path A) previously observed in HRP oxidation
of 3-alkylindoles, where cleavage products like 18 and oxindoles
like 15 dominated (lower pH than in the current work).15 The
key intermediate for this pathway would be the hydroperoxide
19 formed following O2 interception of the initially generated
indole radical cation.15 The absence of 18 in the present reaction
is presumably due to the use of a high concentration of HRP
such that the native HRP would quickly reduce any hydroper-
oxide 19 to 11.15 The ability of HRP to reduce organic
hydroperoxides is well documented.21 Product 12/13 is the
hydrate form of the endo-iminum 21 generated following the
initial amine oxidation (path B in Scheme 2). Amine oxidation
(18) Job, D.; Dunford, H. B. Eur. J. Biochem. 1976, 66, 607-14.
(19) (a) Silva, S. O.; Ximenes, V. F.; Catalani, L. H.; Campa, A. Biochem.
Biophys. Res. Commun. 2000, 279, 657-62. (b) Ximenes, V. F.; Catalani,
L. H.; Campa, A. Biochem. Biophys. Res. Commun. 2001, 287, 130-4. (c)
Ximenes, V. F.; Campa, A.; Catalani, L. H. Arch. Biochem. Biophys. 2001,
387, 173-9.
(20) (a) Skordos, K. W.; Skiles, G. K.; Laycock, J. D.; Lanza, D. L.; Yost, G.
S. Chem. Res. Toxicol. 1998, 11, 741-9. (b) Fennell, R. C. G.; Plant, S.
G. P. J. Chem. Soc. 1932, 2872-6.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 937

A R T I C L E S
Ling et al.
Figure 5. HPLC profile (290 nm) recorded upon 10 min incubation of HRP (40 µM), 2 (800 µM), and H2O2 (200 µM) in pH 8.0 phosphate buffer at 25
°C (left panel) and the corresponding diode array UV spectra of the major metabolites (right panel). During HPLC separation, HRP decomposes to protein
(P*) and heme.
Scheme 2
would also be expected to generate exo-iminum 23, likely to
be the precursor of observed product 14, through the interme-
diacy of indole-3-acetaldehyde (22) (path C). Indeed, there is
an earlier report on HRP catalyzed oxidation of 22, where up
to a 50% yield of 14 could be observed at low pH.22 The yield
of 14 from 22 under our conditions was 45% by HPLC (Figure
S7). Our finding of 14 rather than 22 in the HRP oxidation of
2 presumably reflects the fact that 22 was oxidized more rapidly
than was the starting 2 by HRP/H2O2 in pH 8.0 phosphate buffer,
thereby establishing 14 as an end product representing 45% of
initial reaction along path C, Scheme 2. Although the mechanism
of this formal skeletal truncation is unclear, evidence that the
enol form 24 might be a key intermediate is that 22 underwent
autoxidation, albeit slowly, to give 14 in pH 8 phosphate buffer
(Figure S7 and Supporting Information). On the basis of
literature reports on electron-transfer oxygenation of enols,23
one can postulate that oxidation of 24 to 14 involves formation
and cleavage of a dioxetane intermediate 25.
Under the same conditions, HRP oxidation of 1 or 4 also
gave the corresponding amine oxidation products 14 and 29/32
or 31/34, and indole oxidation products 26 or 28 (Figure S9).
Surprisingly, HRP oxidation of 3 afforded only indole oxidation
product 27; no amine oxidation products 14 and 30/33 could
be detected by HPLC (Figure S9). By comparison with HRP
oxidation of 2, the major metabolites in these reactions were
tentatively identified by HPLC according to their diode array
UV spectra and confirmed by ESI-MS according to their tandem
MS fragmentation modes (Supporting Information). In particu-
lar, confirmation of the low yield metabolite 31/34 was aided
by independent synthesis of an authentic sample, which was
unstable and had to be generated in situ from diBoc-protected
(21) See for example: (a) Shirasaka, N.; Ohnishi, H.; Sato, K.; Miyamoto, R.;
Terashita, T.; Yoshizumi, H. J. Biosci. Bioeng. 2005, 100, 653-6. (b)
Adam, W.; Lazarus, M.; Hoch, U.; Korb, M. N.; Saha-Moeller, C. R.;
Schreier, P. J. Org. Chem. 1998, 63, 6123-7. (c) Adam, W.; Hoch, U.;
Lazarus, M.; Saha-Moeller, C. R.; Schreier, P. J. Am. Chem. Soc. 1995,
117, 11898-901.
(23) See for example, Wu, S.-P.; Liu, J.-F.; Jiang, Z.-Q. Chem. Commun. 1996,
493-4.
(22) Yeh, R.; Hemphill, D., Jr.; Sell, H. M. Biochemistry 1970, 9, 4229-32.
9
938 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Oxidation of N-(3-Indoleethyl) Cyclic Aliphatic Amines
A R T I C L E S
Figure 6. Difference spectra obtained by titration of native HRP (7.00 µM) with 2 (left panel) and 9 (right panel) in 0.1 M pH 10 borate buffer at 25 °C
and the corresponding 1/[S] vs [E]/∆A plots. Detailed conditions are given in the Experimental Section.
Table 2. Product Distributions in HRP Oxidations of 1-4a
conversionb
(%)
product distributionc
(%)
% indole
oxidation
% amine
oxidation
amine oxidation
endo:exo ratio
substrate
1
2
3
4
60
35
33
12
26 (4.0), 29/32 (86.5), 14 (9.5)
11 (29.0), 12/13 (60.7), 14 (10.3)
27 (100)d
4.0
29.0
100
96.0
71.0
0
9.1
5.9
28 (86.4), 31/34 (11.7), 14 (1.9)
86.4
13.6
6.2
a All reactions were conducted in 0.1 M pH 8.0 phosphate buffer at 25 °C with an indole/ H2O2/ HRP ratio of 20:5:1. b Determined by HPLC after 10
min (after which there is no further reaction). c Normalized yields based on corrected HPLC integrals of the three major metabolites. The yield for 14 was
corrected to reflect the % reaction along pathway C by dividing the observed yields of 14 by 45%, the observed yield of 14 obtained from a control HRP
oxidation of 22 (see Supporting Information). d No amine oxidation products 30/33 and 14 were found by HPLC within the detection limit.
aldehyde 35 prepared by DIBAL reduction of ester 36 (Sup-
porting Information). By normalizing the HPLC peak integrals
of the three major metabolites, which were adjusted according
to an estimation of their respective extinction coefficients at
the same wavelength (Supporting Information), the relative
product distributions could be estimated (Table 2), thereby
allowing for an appreciation of the partitioning not only between
indole and amine oxidation pathways, but also between en-
docyclic and exocyclic amine oxidation pathways.24
Binding Affinities. Binding of substrates to native HRP was
studied by optical difference spectroscopy.8 In 0.1 M pH 7.00
phosphate buffer, preliminary experiments with 2 and 7 showed
no detectable UV-vis difference spectral signals, suggesting
that the binding affinity of the conjugate acid form of the
aliphatic amines to HRP is low. Thus all the binding experiments
were conducted at pH 10, as the best compromise between (i)
ensuring high concentrations of the free base amine forms and
(ii) preventing loss of enzyme activity at high pH. Typical
difference spectra of complexes of native HRP with 2 and 9
are shown in Figure 6, and those with 7 are shown in Figure
S3. Analysis of these spectra (Experimental Section) yielded
the molar absorptivity differences ∆ꢀ of the free and bound
enzyme and the apparent dissociation constants Kdapp, which
were then converted to Kd values for the amine free base forms
(Table 3).
Assuming that only one-electron oxidation of substrates 1-4
by either compounds I or II is sufficient to result in final
products (e.g., completion of reaction would be mediated by
O2), the limiting maximal conversions that could be obtained
in the reactions reported in Table 2, based on the relative reactant
stoichiometry used, were then 10% based on HRP (if no
turnover) and 50% based on H2O2. Thus, in all cases, turnover
is evident, and the conversion of 1 in excess of 50% must reflect
at least some regeneration of HRP compound I by O2-derived
ROOH intermediates.
Two distinct spectral patterns were observed: (1) λmax and
min at 380-383 and 417-420 nm, respectively, and isosbestic
λ
points at 399-406 nm, seen for 1-4, 9, and 10, and (2) λmax
and λmin at ca. 402 and 420-423 nm, respectively, and isosbestic
point at ca. 413 nm, seen for 5-8. Although data could not be
obtained in some cases because of solubility problems, it is clear
that the binding affinities of the substrates are similar. Under
the same conditions, no UV-vis difference spectral signals
could be detected with N-methyl cyclic amines (Supporting
Information), indicating probable nonbinding of simple aliphatic
amines lacking an aryl substituent.
(24) If a large excess of H2O2 (2000-fold with respect to HRP) was used, the
oxidations became rather complicated and the 3-hydroxyindolenines became
the major products in all cases (see Figure S6 for an example). The reasons
are that (1) all the major amine oxidation products can be further
metabolized by HRP and (2), in the presence of a large excess of H2O2,
compound III is the major oxidant, which has been shown especially capable
of mediating indole oxygenation.19
Kinetic Saturation. Confirmation that observed binding of
amine substrates to native HRP provides reasonable evidence
that electron-transfer to compounds I and II occurs via similar
enzyme-substrate association complexes, requires a demonstra-
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 939

A R T I C L E S
Ling et al.
3 are reasonable estimates for binding affinities of substrates
to the oxidative forms of HRP. These results are important (i)
in that this is one of the first examples of kinetic saturation
behavior seen for HRP25 and (ii) in that because kapp ) kcat/
Kdapp (eq 4); despite the low binding affinities, we can conclude
that substrate binding contributes directly to the apparent second-
order rate constants.
Binding Mode. To investigate the basis of the experimentally
determined Kd values for the cyclic tertiary amine free base
forms with native HRP, possible binding modes of 1-8 were
examined by computational docking simulation, starting with
the published crystal structure of the enzyme and the energy-
minimized conformation of each amine. Docking modes to the
active site energy grid were investigated using the Discovery
Studio Software Package PC Version 1.7, which examines
possible rotation about each single bond in the substrate and
then allows assignment of energy scores according to the newly
developed scoring function LigScore2.26
Figure 7. Plot of pseudo-first-order rate constant (kobsd) vs substrate
concentration obtained in reduction of HRP compound II (21 µM) by 4 in
0.1 M of pH 9.00 phosphate buffer at 25 °C and the corresponding nonlinear
least-squares fit according to eq 3.
Table 3. Difference Spectral Parameters and Dissociation
Constants of HRP-substrate Complexesa
Crystallographic studies have shown that the conformations
of native HRP and compounds I and II are almost identical;27
thus we relied on the coordinates of the native enzyme to shed
light on structural features that might determine both compound
I and II oxidation rates. Of all the docking modes defined by
the program (14-22 modes for 1-4, and 8-9 modes for 5-8),
>95% of them involve orientation of the ligand molecules with
the aromatic rings pointing toward the heme edge and the
azacycles projecting out of the heme pocket (Figure S14). On
the basis of this statistical preference and the well-known ability
of simple aromatics (e.g., phenols and anilines) to bind to the
“heme edge” binding domain, as well as a recent computational
docking study of serotonin and melatonin with HRP,28 we
propose that substrates 1-8 all bind to the enzyme through their
aromatic nuclei.
isosbestic
λmax
(nm)
λmin
(nm)
point
(nm)
Kadpp
∆ꢀpeak
Kd
∼
trough
1
substrate
(mM)
(mM-1 cm-
)
(mM)c
1b
2
3
381
380
381
381
402
402
402
402
383
382
418
418
418
418
423
423
423
423
417
418
404
404
405
405
413
413
413
413
402
399
1.9
1.0
1.0
15
10
12
1.4
0.72
0.86
4
5b
6b
7
1.3
1.8
1.9
0.90
3.3
2.0
16
1.1
1.4
8
9
10
13
a Data obtained in 0.1 M pH 10 borate buffer at 25 °C. b Because of
poor solubility of substrate the peak and trough could be seen only at low
concentrations, no reasonable data analysis could be achieved. c Dissociation
constants of the free base amine forms.
For both N-(indole-3-ethyl) and N-benzyl series of substrates,
two universal and leading scored docking modes (DM I and II)
were identified (Figures 8 and S15). These two modes differed
in terms of the distance between substrate aromatic ring and
the heme edge, though the positioning of the azacycle was nearly
overlapping. Thus, whereas the amine nitrogen-heme distances
in both modes are 9-11 Å for all substrates, the minimum
aromatic terminus-heme distances in DM I are longer for 5-8
(5.5-6.5 Å) than for 1-4 (4-5 Å), and are even longer (7-8
Å) in DM II for all substrates (Table S6).
tion that the kinetics of substrate oxidation follows a saturation
phenomenon predicted by the preassociation equilibrium defined
in eq 3; that is, the kobsd should plateau at high substrate
concentrations. In practice, saturation for enzymes like HRP is
rarely observed because (i) most aromatic substrates of HRP
possess Kd values in the millimolar range and (ii) these substrates
are usually poorly soluble, confounding obtaining kinetic data
at high substrate concentration. Nonetheless, rate data for
compound II reduction by 4 could be obtained at high
concentration for this fortuitously more soluble substrate.
Although the maximum plateau rate was not reached experi-
mentally, the kobsd versus [4] plot clearly exhibits saturation
behavior (Figure 7).
For the N-(indole-3-ethyl) series 1-4, although ancillary van
der Waals and/or hydrophobic interactions between the substrate
indole ring and three aromatic residues (Phe-143, Phe-142, and
Phe-179) appear to be involved in both modes, the major
distinguishing feature of DM-I is an apparent stabilizing
interaction of the indole ring with the heme, whereas in DM II,
the major stabilizing interaction appears to be with Phe-68. In
the case of the N-benzyl series 5-8, whereas docking simula-
tions again resulted in two principal docking modes, the
distinction between the two modes is more subtle, reflecting
A nonlinear least-squares fit of this plot according to eq 3
app
afforded a Kd value of 1.79 ( 0.16 mM, and a kcat value of
0.0486 ( 0.0015 s-1, indicating that the highest concentration
of 4 used (6 mM) attained ∼80% saturation. According to eq
4, the kapp can be calculated as 27.2 s-1 M-1, which matches
well with the kapp (27.0 s-1 M-1) obtained at the same pH using
a low concentration of 4 (Figure 2). Moreover, correcting the
app
(25) For one example of kinetic saturation for HRP compound II reduction by
p-cresol see: Critchlow, J. E.; Dunford, H. B. J. Biol. Chem. 1972, 247,
3703-13.
kinetically determined Kd value by the amine pKa yields an
estimate of the binding constant Kd of 4 free base with HRP
compound II as 0.71 mM, which is close to the binding constant
of 4 free base with native HRP (0.86 mM) measured by optical
difference spectroscopy (Table 3). We can extrapolate this
finding and conclude that the remaining Kd values listed in Table
(26) Krammer, A.; Kirchhoff, P. D.; Jiang, X.; Venkatachalam, C. M.; Waldman,
M. J. Mol. Graphics Modell. 2005, 23, 395-407.
(27) Berglund, G. I.; Carlsson, G. H.; Smith, A. T.; Szoke, H.; Henriksen, A.;
Hajdu, J. Nature 2002, 417, 463-8.
(28) Hallingback, H. R.; Gabdoulline, R. R.; Wade, R. C. Biochemistry 2006,
45, 2940-50.
9
940 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Oxidation of N-(3-Indoleethyl) Cyclic Aliphatic Amines
A R T I C L E S
Figure 8. Best scored docking modes for 2 (left panel, DM I (red) and DM II (blue)) and 6 (right panel, DM I (green) and DM II (yellow)). The mouth
rim of HRP heme pocket is defined by four hydrophobic residues: Phe-143, Phe-142, Phe-179, and Phe-68.
Scheme 3
differences in the orientation of the phenyl ring relative to the
Phe residues lining the mouth of the “heme edge”. Thus, if it is
assumed that the main operational binding mode in both
substrate series is defined by DM-I, where the substrate aryl
ring is closer to the heme edge, then the better heme proximity
of the indole ring in 1-4 than of the phenyl ring in 5-8 may
explain the distinct optical difference spectra and larger extinc-
tion coefficients with HRP observed for amines 1-4 (and
indoles 9 and 10) relative to those for amines 5-8 (Table 3).
lism of 1-4 revealed that except for 3 (affording strictly indole
oxidation products), reactions of the five-, seven-, and eight-
membered azacycle compounds led to a mixture of indole and
amine oxidation products, with the latter predominating in the
case of 1 and 2 (Table 2). The product study also revealed that
endocyclic amine oxidation predominated over exocyclic amine
oxidation (Table 2), consistent with five-, seven-, and eight-
membered rings favoring an endocyclic rather than exocyclic
double bond.
When one examines the product distribution data reported
in Table 2, it is seen that the faster rates for compounds 1 and
2 (Table 1) correspond to those cases where there is more
product deriving from amine rather than indole oxidation.
Because k and k1 for oxidations by compounds I and II,
respectively, were measured using large excesses of substrates
at the initial stage of conversion, it is reasonable to assume that
these rates represent initial one-electron oxidations to generate
radical cations that would not compete with the starting
substrates in reducing compounds I and II. The radical cations
(shown for 2 in Scheme 3) would subsequently partition among
various possible pathways for loss of proton, ultimately reflect-
ing paths A-C described in Scheme 2.
Discussion
Structure-Reactivity and Mechanism. Previous studies
have established that simple aliphatic tertiary amines are poor
substrates of HRP under turnover conditions.13 Here we
confirmed that reductions of HRP compounds I and II by
N-benzyl cyclic amines 5-8 are slow (Table 1) and nearly as
slow as are the reductions by the corresponding N-methyl cyclic
amines (Table S4). In contrast, reductions of compounds I and
II by N-(indole-3-ethyl) cyclic amines 1-4 are 2-3 orders of
magnitude faster (Table 1), and preparative turnover reactions
could be observed, especially in the case of amine 1. Given
that indole (9) and especially 3-methylindole (10) also react
with HRP fairly quickly (Table 1), the fast reductions by 1-4
could simply reflect reactions with the indole nucleus in these
compounds. However, a full product analysis of HRP metabo-
For indoleethylamines 1-4, we suggest that there are two
high-lying filled molecular orbitals (MOs), one with the largest
coefficients in the indole π system and one with the largest
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 941

A R T I C L E S
Ling et al.
Figure 9. Linear free energy plots correlating the oxidation rates of cyclic 3-indoleethylamines 1-4 (left panel) and cyclic benzylamines 5-8 (right panel)
with redox potentials.
coefficients on the amine nitrogen lone pair. Thus, one electron
oxidation would result in two rapidly equilibrating valence
tautomeric radical cations, one where the half-filled orbital
density is distributed mainly on the indole ring (electronic state
37) and the other where it is distributed mainly on the tertiary
amine nitrogen (electronic state 38). Product partitioning would
then reflect the position of this equilibrium, governed by the
relative energy of these two MOs, in turn reflecting redox
potentials and the conformations of the substrate amines bound
to HRP.
where the process becomes thermodynamically activationless.
Our rates, however, are in the slower linear range of typical
Marcus plots. The absence of data in the fast regime prevents
a quantitative analysis according the Marcus equation, which
requires a curved plot to solve for the two unknown parameters
(the reorganization energy and the activationless rate).30
For our mechanistic needs, where it is assumed that the
electron-transfer distance is the same within each series of
amines, simple plots of log k vs E are more suitable,31 and are
shown in Figure 9 for the cyclic voltammetry data. Although
the plots for compounds 5-8 are roughly linear (there are only
four experimental points, so a statistical analysis of linearity
seems unwise), the plots for compounds 1-4 are markedly
curved. The leveling out of plots in going from compounds 1,
2, 4 to compound 3 correlates with the point where the oxidation
product profile changes from one reflecting a mixture of indole
and amine oxidation to one reflecting strictly indole oxidation.
Our interpretation is that whereas HRP-mediated oxidations of
5-8 involve strictly one-electron oxidations of the tertiary amine
center, oxidation rates for the indoleethylamines reflect a
changeover from mainly amine oxidation for low potential
amine centers, ultimately to mainly indole ring oxidation when
the amine potential becomes sufficiently high.32 Indeed, such a
change in mechanism would predict a curved LFER plot. As
an academic exercise, one could attempt to treat the partitioning
shown in Scheme 3 by partitioning the measured k values into
two components (amine oxidation vs indole oxidation) on the
basis of product yield distribution. The expectation would be
that the kindole values would be relatively constant and that the
12a
It is well-known that the rates of chemical (e.g., K3Fe(CN)6
or Cu(II)29) or electrochemical12b,12c oxidation of cyclic amines
depend critically on ring size, since the rehybridization from
sp3 to sp2 that occurs upon one-electron oxidation is confor-
mationally favored by five-, seven-, and eight-membered rings
but disfavored by six-membered rings on the basis of torsional
strain factors. Our observed electrochemical E trend (for both
series of compounds) 8 e 7 < 5 < 6 is congruent with the E
trend observed for the corresponding N-methyl cyclic amines.12c
Thus, the faster rates of oxidation of 1 and 2 reflect favorable
one-electron oxidations of the tertiary amine centers in the
seven- and eight-membered azacycles, associated with the
relatively greater stabilizing nature of aminium radical cation
forms like 38 for these ring sizes. In contrast, the less favorable
one-electron oxidation of the tertiary amine center in the six-
membered ring substrate 3 reduces the importance of 38 for
this ring size, in turn manifesting in exclusive indole oxidation.
Since compounds I and II were shown to have similar regiose-
lectivity in oxidizing a selected cyclic 3-indoleethylamine
substrate (Supporting Information), the conclusions above appear
to pertain to both sets of rate constants listed in Table 1.
Linear Free-Energy Relationships. For a structurally related
series of compounds such as 1-4 or 5-8, one might expect to
witness a linear free-energy relationship (LFER) between rates
and thermodynamic parameters, as is seen for rates of HRP
oxidation of substituted phenols and anilines as a function of
Hammett substituent constants σ.18 When the process involves
electron-transfer and the thermodynamic property is redox
potential, it has been more common recently to discuss rate
trends according to Marcus theory. The advantage of the latter
over a simple LFER treatment is that the Marcus equation
explains the leveling off of rates as they increase to the point
kamine values would exhibit a more linear free-energy dependence
on electrochemical potential. Our data appear to support this
expectation (Supporting Information).
How Does the Auxiliary Indole Ring Promote Amine
Oxidation? If the low reactivity toward HRP of simple aliphatic
amines (e.g., cyclic methylamines, Supporting Information) is
partly due to lack of binding, then the significantly improved
(30) Folkes, L. K.; Candeias, L. P. FEBS Lett. 1997, 412, 305-8.
(31) Colosi, L. M.; Huang, Q.; Weber, W. J., Jr. J. Am. Chem. Soc. 2006, 128,
4041-7.
(32) On the basis of the voltametrically measured redox potentials of 1-4 (Table
1) relative to the higher values of typical 3-alkylindoles (Waltman, R. J.;
Diaz, A. F.; Bargon, J. J. Phys. Chem. 1984, 88, 4343-6), it might be
predicted that oxidations of all four indoleethylamines would lead strictly
to amine oxidation products. However, oxidation outcome also reflects the
fact that it is the indole ring that is bound at the heme edge and is thus
closest to the HRP oxidants.
(29) Wang, F.; Sayre, L. M. J. Am. Chem. Soc. 1992, 114, 248-55.
9
942 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Oxidation of N-(3-Indoleethyl) Cyclic Aliphatic Amines
A R T I C L E S
Kinetics of Reductions of HRP Compounds I and II by 1-9.
HRP compound I was freshly prepared each time by mixing 3 mL of
HRP stock solution in buffer with a 10-fold excess of H2O2 in a 3-mL
quartz cuvette, to which was quickly added an appropriate amount of
substrate.15 The reaction was immediately followed by monitoring the
appearance of the compound II Soret band (419 nm for pH 8 and 420
nm for pH 9 and 10), and the data were then analyzed using Origin
7.5 (Figure 1). The substrate concentration (excess) was chosen so that
the rate in the presence of substrate was more than 10 times faster
than the background enzyme decay rate measured in the absence of
substrate. A fresh solution of compound II was prepared each time by
the addition of 1 equiv of H2O2 to the native HRP solution (3 mL, 2.1
µM in buffer) and then waiting for 45 min.15 After the addition of at
least 10-fold excess of substrate to the cuvette containing compound
II, the reaction was followed by monitoring the appearance of the native
HRP Soret band at 403 nm, and the data were analyzed using Origin
7.5 (Figure 2). The background decay of compound II was very slow
and could be neglected. Control experiments showed that the substrates
were all stable in buffer. Blank cuvets in these experiments contained
only the appropriate buffer.
Binding Experiments. Difference spectra (enzyme-substrate com-
plex vs enzyme) were recorded in the following manner using solutions
of 1-8 that were each pre-equilibrated to 25.0 °C. Initially, both the
sample and reference cuvets were filled with 3 mL of the HRP solution
(7 µM in buffer). Then small volumes of the substrate solution were
successively added to the sample cuvette, with concomitant addition
of the same volume of buffer into the reference cuvette. After each
addition, the contents of both cuvets were mixed well, and the difference
spectrum was recorded. In the meantime, the absorption spectra of
substrates at all the operational concentrations were also recorded using
similar procedures in the absence of HRP. All the difference spectra
were corrected by deduction of both the baseline and the corresponding
substrate absorption spectra using Origin 7.5 software. Volumes of
standard HRP and substrate stock solutions were mixed to attain 10
different final solutions with [HRP] ranging from 7.00 to 6.75 µM and
the [substrate] ranging from 0 to 2.04 mM. In some cases, severe
precipitation occurred during the titration and only the spectra for
solutions that remained homogeneous were used in the calculations
(Figures 6 and S11).
binding affinity of 1-4 and 5-8 would explain their enhanced
reactivity. Indoleethylamines 1-4 have lower potential and react
faster. However, although binding of these compounds occurs
through the indole ring, in three of the four cases, the product
mixtures derive partly or mainly from oxidation at the tertiary
amine center. It thus appears that binding of the indole ring
leads to substrate-HRP complexes that are conformationally
disposed for mediation of what effectively becomes medium
long-range SET oxidation of the tertiary amine centers by the
active oxidants of compounds I and II.
We are aware of recent studies suggesting that oxidation of
phenolic substrates by HRP involves a proton-coupled electron-
transfer (PCET) process, and that the involvement of an
porphyrin radical cation isomer of compound II (P+•Fe(III)-
OH rather than PFe(IV)-OH) would permit a convergence of
PCET details for both compound I and II reductions.33 Although
PCET would not apply to the oxidation of tertiary amine to
tertiary aminium radical cation, oxidations by HRP leading to
indole oxidation products could well be facilitated by a PCET
process, leading directly to neutral indolin-3-yl radicals like 39
(hashed arrow in Scheme 3).
Conclusion
Despite the recognition in some circles that aliphatic (as
opposed to aromatic) amines are poor substrates for HRP, this
realization has not been elaborated in the literature. To our
knowledge, the current study is the first to address this problem
systematically, and to identify one case that meets the criteria
needed for observing reaction. Thus, indoleethylamines that
contain tertiary amine centers that are unusually susceptible to
one-electron oxidation (due to geometrically enforced hybridiza-
tion factors) are excellent substrates for HRP and can undergo
turnover. It appears that the electron-rich indole ring not only
improves binding but, more importantly, is also able to mediate
the electron transfer between the tethered tertiary amine center
and the HRP oxidant centers. The latter factor is undoubtedly
under conformational control, and the successful oxidation of
Vinca alkaloids by HRP14 may further reflect the enforcement,
by the rigid ring system, of an appropriate orientation of the
tertiary amine lone pair relative to the indole ring. It is proposed
that although the indole ring is needed to “anchor” the alkaloid
molecule at the heme crevice, the lowest energy electron transfer
generates a radical cation localized mainly on the tertiary amine
nitrogen rather than indole ring.
From the difference spectra, the apparent dissociation constants
Kadpp, and the molar absorptivity differences ∆ꢀ of the free and bound
enzyme were obtained8 from a plot of 1/[S] vs [E]/∆A according to
∆ꢀ[E]
Kadpp∆A Kadpp
1
[S]
1
)
-
For amine substrates, the binding of conjugate acid forms were
neglected, and the binding constants Kd of free base were estimated
according to
On a broader perspective, our results confirm that a key
determinant of substrate activity for HRP is the availability of
a flat and rigid aromatic domain that can appropriately fit the
heme crevice. In this regard, although the physiological
relevance of aliphatic amine oxidation by peroxidases is unclear,
the ability of a redox-active aromatic moiety to facilitate
oxidation of an aliphatic electron donor may be important in
drug development and xenobiotic metabolism.
Kadpp
Kd )
[Η+]
Ka
1 +
Kinetic Saturation Phenomenon. A stock solution of 4 (40 mM)
was prepared by first dissolving the HCl salt form of 4 in water, and
then titrating the solution with aqueous NaOH to pH 8.2 (higher pH
resulted in precipitation). This was found adequate to permit dilution
into pH 9 buffer without causing a significant pH change (<0.02). To
a 3-mL quartz cuvette were added 21 µM of HRP and a prescribed
volume of 0.1 M of pH 9.00 borate buffer. After mixing, 21 µM of
H2O2 was added, and the mixture was allowed to stand at 25 °C for 45
min to give HRP compound II. Then a prescribed volume of the 40
mM stock solution of 4 was added to achieve a final volume of 3 mL.
After rapid mixing, the reaction was monitored at 403 nm at 25 °C.
Experimental Section
General experimental methods, synthesis and characterization of
amine substrates, determination of acid dissociation constants, and
details of product characterization for HRP oxidations, including
preparative oxidations, appear in Supporting Information.
(33) Derat, E.; Shaik, S. J. Am. Chem. Soc. 2006, 128, 13940-9.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 943

A R T I C L E S
Ling et al.
The collected data were analyzed as above to give the observed pseudo-
first-order rate constants kobsd, which were then plotted against substrate
concentrations (Figure 7). A nonlinear least-squares fit of the plot
according to eq 3 was conducted using Origin 7.5 to yield kcat and
A full conformational analysis for 3 was thus conducted and the energies
of various conformations were minimized by ab initio calculation using
Gaussian 03 to figure out the most stable one (Supporting Information).
All other congeners (1, 2, and 4) were then assigned the same
conformations as 3, which were energy minimized with MM2.
The crystal structure (PDB format) of native HRP (1H58)27 in the
Protein Data Bank was loaded on the Accelrys Discovery Studio
software package (PC version 1.6). All the crystal water molecules were
removed, and hydrogens were added by the program. The bond grades
of the heme residue and charges on the heme oxygen, the heme iron,
and its coordinating nitrogens were corrected manually. Protonation
states of all other ionizable amino acid residues were set as default.
The active site was then identified, and the energy grid was set
automatically by the program. The energy minimized; most stable ligand
structures (MOL format) were loaded. The “energy grid forcefield”
was set as “Drieding”, the “conformation search number of Monte Carlo
trials” was set as 30000, and “pose saving perform clustering” was
allowed, with “pose saving maximum cluster per molecule” being set
as 100. Each output pose represents the energy minimum of a cluster
of poses and is thus considered as a docking mode. The “scoring scores”
option was set as LigScore2. Other parameters were set as default. After
calculation, the output poses were ranked according to the LigScore2
scores, visualized, and analyzed. The results are shown in Figures 8,
S14, and S15 and Table S6.
Kadpp
.
General Procedures for HPLC and ESI-MS Analyses of HRP
Oxidations of Indoleethylamines. To a solution of an indoleethylamine
substrate (1-4, 800 µM) and HRP (40 µM) in 0.1 M pH 8.0 phosphate
buffer (100 µL) was added 2 µL of 10 mM of aqueous H2O2 at 25 °C.
After quick mixing, the brown solution immediately turned green, then
red, and finally brown. A 20 µL aliquot of the reaction mixture was
taken at the 10 min point and analyzed by HPLC using a 4.6 mm ×
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) and B (95% aqueous CH3CN), both containing 0.2%
(v/v) TFA, according to the following program: 100% A to 95% A
0-20 min, 95% A to 100% A 20-25 min. Addition of another 2 µL
H2O2 to the reaction mixture again resulted in an immediate color
change to green, then to red, and finally to brown, indicating that only
the normal compound I/compound II catalytic cycle was involved.
For ESI-MS analysis, the above incubation was repeated, but with
HRP, indoleethylamine substrate, and H2O2 all being 5-fold more
concentrated. After incubation for 20 min, the reaction mixture was
subjected to HPLC separation using the same conditions as above.
The fractions containing the major metabolites were collected and
subjected to ESI-MS and tandem MS analysis on a ThermoFinnigan
LCQ Advantage instrument operating in the positive mode using
nitrogen as auxiliary gas. The instrument was set in the syringe pump
mode, with the heated capillary temperature being maintained at 100-
150 °C. The source voltage was 4.57 kV and the capillary voltage was
8.27 V.
Acknowledgment. We thank the NIH for support of this work
through Grant GM 48812.
Supporting Information Available: Figures S1-S16, Tables
S1-S7, determination of minor rate constants for compound II
reduction, ESI-MS characterization of HRP metabolites of 1-4,
validation of determining product distributions by HPLC, HRP
oxidations and binding of cyclic N-methylamines, ab initio
calculations to determine the most stable conformation of 3,
partitioning rate constants for oxidation of 1-4 into kindole vs
kamine components, and detailed experimental methods. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Docking Simulation. Ligand structure construction and energy
minimization were performed on the Chem3D Pro 10.0 platform of
the ChemOffice software package (Cambridge). Only the free base form
of amine substrates was considered. In constructing the ligand models,
the conformations of azacycles were first optimized by MM2 and then
attached to the indoleethyl or benzyl groups in an equatorial manner,
and the resulting conformations were further energy minimized. For
5-8, only two bonds are rotatable, and the most stable conformation
can be readily determined. For 1-4, however, three bonds are rotatable.
JA075905S
9
944 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008