Enzymatic N-dealkylation of an N-cyclopropylamine: An unusual fate for the cyclopropyl group [10]

Full text of DOI:10.1021/ja003048l

J. Am. Chem. Soc. 2001, 123, 349-350
349
transfer (SET) to generate a radical-cation.11,15-18 In agreement
with their calculated instability,19 cyclopropylaminium radical-
cations (C3H5NH2+•) undergo ring-opening extraordinarily
rapidly.20-22 Ring-opening generates an acyclic, distonic radical-
cation that could, conceivably, covalently modify enzyme active
sites or go on to form stable acyclic metabolites. On the other
hand, N-dealkylation via the classical process of hydrogen atom
abstraction followed by hydroxyl recombination at C-1 of the cy-
clopropyl moiety should produce the ring-intact metabolite cyclo-
propanone (which undergoes extensive hydration in solution23).
To validate the use of cyclopropylamines as reporters for
differentiating between the SET vs classical pathways of N-
dealkylation, we sought to identify the metabolites derived from
the cyclopropyl group of a cyclopropylamine undergoing SET-
initiated N-dealkylation. We began our studies using HRP because
of its well-known action as an SET oxidant and we chose
N-cyclopropyl-N-methylaniline (2) as a relatively low-E1/2 sub-
strate because HRP oxidizes many aniline derivatives but does
not oxidize amines such as 1 due to their higher oxidation
potentials and pKa values.16,24 Incubation of 225 with HRP under
standard conditions26 resulted in its complete oxidation within
30 min. Product analysis by direct HPLC27 revealed a single UV-
absorbing peak identical with that of N-methylaniline (NMA),
and the absence of N-cyclopropylaniline (NCA). An aliquot of
reaction mixture was treated with 2,4-dinitrophenylhydrazine
(DNP) reagent28 and analyzed by HPLC and capillary GC/MS.29
These analyses revealed the absence of formaldehyde, acrolein,
and 3-hydroxypropionaldehyde (from N-demethylation vs N-
dealkylation of 2, respectively), although control experiments
showed these compounds to be readily detectable at levels
corresponding to as little as 5% conversion of substrate. Unfor-
tunately, despite considerable effort, we found that cyclopro-
panone hydrate in dilute aqueous solution (or even in more-
concentrated solution) does not form a DNP derivative. We could
rule out its formation, however, on the basis that amounts
corresponding to 50% conversion of 2 completely inactivated HRP
Enzymatic N-Dealkylation of an
N-Cyclopropylamine: An Unusual Fate for the
Cyclopropyl Group
Christopher L. Shaffer, Martha D. Morton, and
Robert P. Hanzlik*
Department of Medicinal Chemistry
UniVersity of Kansas, Lawrence, Kansas 66045-2506
ReceiVed August 16, 2000
The cyclopropylamine substructure is found in numerous drugs
and drug candidates, many of which undergo cytochrome P450-
catalyzed N-dealkylation with loss of the cyclopropyl group. Some
cyclopropylamines, such as N-cyclopropylbenzylamine (1), are
also suicide substrates for P450 enzymes.1-3 The latter property
is general among but unique to cyclopropylamines; acylating the
nitrogen, separating it from the cyclopropane moiety, or enlarging
the ring to cyclobutyl greatly reduces or eliminates suicide
substrate activity.4,5 Despite numerous studies of the P450-
catalyzed oxidation of cyclopropanes,6-9 the fate of the cyclo-
propyl groups lost during N-dealkylation and the nature of the
reactive intermediates involved in their inactivation of P450
enzymes remain obscure.
Monoamine oxidase, a flavoprotein, also catalyzes the N-
dealkylation of amines and is covalently inactivated by cyclo-
propylamines including 1.10,11 In this case product analysis showed
that the cyclopropyl group was converted to the acyclic metabolite
acrolein.12 Similarly, oxidation of cyclopropanone hydrate by
horseradish peroxidase (HRP), a heme enzyme whose redox cycle
involves a ferryl group similar to that in P450 enzymes, leads to
inactivation of the enzyme due to covalent modification of a meso
carbon of the porphyrin periphery by a 2-carboxyethyl substituent
(i.e. a net two-electron oxidation).13 While P450 can oxidize a
wide range of substrates by several mechanisms,14,15 extensive
data support the hypothesis that horseradish peroxidase, cyto-
chromes P450, and monoamine oxidase can each oxidize at least
some of their substrates by an initial step involving single electron
(16) Dunford, H. B. Heme Peroxidases; Wiley-VCH: New York, 1999.
(17) Castagnoli, N., Jr.; Rimoldi, J. M.; Bloomquist, J.; Castagnoli, K. P.
Chem. Res. Toxicol. 1997, 10, 924-940.
(18) Goto, Y.; Watanabe, Y.; Fukuzumi, S.; Jones, J. P.; Dinnocenzo, J.
P. J. Am. Chem. Soc. 1998, 120, 10762-10763.
(19) Bouchoux, G.; Alcaraz, C.; Dutuit, O.; Nguyen, M. T. J. Am. Chem.
Soc. 1998, 120, 152-160.
(20) Loeppky, R. N.; Elomari, S. J. Org. Chem. 2000, 65, 96-103.
(21) Ha, J. D.; Lee, J.; Blascktock, S. C.; Cha, J. K. J. Org. Chem. 1998,
63, 8510-8514.
(22) We could find no reported rate constant for ring-opening of a
cyclopropylaminium ion, but protonated dialkylaminium cation radicals are
(102-105)-fold more reactive than their neutral dialkylaminyl radical coun-
terparts, and the N-methyl-N-(trans-2-phenylcyclopropyl)aminyl radical ring-
opens with a rate constant of 7.2 × 1011 s-1. Musa, O. M.; Horner, J. H.;
Shahin, H.; Newcomb, M. J. Am. Chem. Soc. 1996, 118, 3862-3868.
(23) Wiseman, J. S.; Abeles, R. H. Biochemistry 1979, 18, 427-435.
(24) Sayre, L. M.; Naismith, R. T.; Bada, M. A.; Li, W. S.; Klein, M. E.;
Tennant, M. D. Biochim. Biophys. ActasProtein Struct. Mol. Enzymol. 1996,
1296, 250-256.
* Address correspondence to this author. Phone: 785-864-3750. Fax: 785-
864-5326. E-mail: rhanzlik@ukans.edu.
(1) Hanzlik, R. P.; Kishore, V.; Tullman, R. J. Med. Chem. 1979, 22, 760-
761.
(2) Hanzlik, R. P.; Tullman, R. H. J. Am. Chem. Soc. 1982, 104, 2048-
2050.
(3) Macdonald, T. L.; Zirvi, K.; Burka, L. T.; Peyman, P.; Guengerich, F.
P. J. Am. Chem. Soc. 1982, 104, 2050-2052.
(4) Bondon, A.; Macdonald, T. L.; Harris, T. M.; Guengerich, F. P. J. Biol.
Chem. 1989, 264, 1988-1997.
(25) Compound 2, a colorless oil, was synthesized in 31% yield by treating
N-methylformanilide with Ti(O-iPr)4 and EtMgBr as described by Chaplinski
and de Meijere (Chaplinski, V.; de Meijere, A. Angew. Chem., Int. Ed. Engl.
1996, 35, 413-414). Purity (>99% after preparative HPLC on silica gel) and
(5) Hall, L. R.; Hanzlik, R. P. Xenobiotica 1991, 21, 1127-1138.
(6) Riley, P.; Hanzlik, R. P. Xenobiotica 1994, 24, 1-16.
(7) Ortiz de Montellano, P. R.; Stearns, R. A. Drug Metab. ReV. 1989, 20,
183-191.
1
identity were confirmed by capillary GC/MS, H NMR, and 13C NMR.
(26) Incubations were conducted at room temperature under air and
contained 0.82 µg (0.021 nmol) of HRP (Sigma, RZ ) 0.5), 500 nmol of
substrate, and 1000 nmol of H2O2 in a final volume of 1.0 mL of potassium
phosphate buffer (0.4 M, pH 5.5). For HPLC analysis incubations were
quenched by addition of 0.67 mL of MeCN, which controls showed to stop
the oxidations completely.
(8) Toy, P. H.; Newcomb, M.; Hollenberg, P. F. J. Am. Chem. Soc. 1998,
120, 7719-7729.
(9) Newcomb, M.; Letadicbiadatti, F. H.; Chestney, D. L.; Roberts, E. S.;
Hollenberg, P. F. J. Am. Chem. Soc. 1995, 117, 12085-12091.
(10) Silverman, R. B.; Hoffman, S. J.; Catus, W., III B. J. Am. Chem. Soc.
1980, 102, 7126-7128.
(27) Aliquots of quenched incubations (20 µL) were injected onto a Vydac
C-18 column (5 µm, 4.6 × 250 mm) eluted at 1 mL/min with the following
two-step gradient: 0-12 min, 0-45% solvent B (MeCN) in solvent A (40%
MeCN in 50 mM NH4OAc); 12-17 min, hold at 45% B in A; 17-25 min,
45-100% B. Column effluent was passed through a UV detector (254 nm)
in series with a Ramona radioactivity flow detector with a solid scintillant
cell. Data were collected by a SRI chromatography data system and analyzed
using Peak Simple software.
(11) Silverman, R. B. Acc. Chem. Res. 1995, 28, 335-342.
(12) Vazquez, M. L.; Silverman, R. B. Biochemistry 1985, 24, 6538-6543.
(13) Wiseman, J. S.; Nichols, J. S.; Kolpak, M. X. J. Biol. Chem. 1982,
257, 6328-6332.
(14) White, R. E. Pharmacol. Ther. 1991, 49, 21-42.
(15) Testa, B. Biochemistry of Redox Reactions; Academic Press, Ltd.:
London, 1995.
10.1021/ja003048l CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/16/2000

350 J. Am. Chem. Soc., Vol. 123, No. 2, 2001
Communications to the Editor
almost immediately, whereas no loss of HRP activity was
observed after complete oxidation of 2.
Scheme 1
To focus more specifically on the fate of the 3-carbon
cyclopropyl moiety we resorted to radiolabeling. Incubation of
[1′-14C]-230 with HRP26 followed by HPLC analysis27 revealed a
single peak of radioactivity in the solvent front with little
associated UV absorbance (254 nm). Upon reducing or eliminating
the organic modifier from the HPLC mobile phase, the 14C
remained at the solvent front and was not retained on the C18
column, nor was it affected by (a) lyophilization and reconstitu-
tion, (b) refluxing in 1 M HCl for 24 h, (c) extraction with organic
solvents, and (d) treatment with DNP reagent.
Scheme 2
Oxidative N-dealkylation reactions often involve iminium ion
intermediates which can be trapped with cyanide ion.31-33 Thus,
we examined the oxidation of [1′-14C]-2 under standard condi-
tions26 with 1 mM KCN added. Since cyanide inhibits HRP, these
reactions took (3-4)-fold longer to reach completion. Examination
of product mixtures by direct HPLC (A254) revealed a sharp new
peak at tR ) 9.0 min in addition to NMA and traces of aniline,
but no NCA. The radiochromatogram showed a significantly
diminished solvent front peak and a sharp new peak at tR ) 9.0
min (metabolite M1). Throughout the entire incubation period
(0-100% consumption of 2) the 14C in the solvent front peak
plus that in the M1 peak equalled the amount of substrate con-
sumed, and the metabolite ratio (M1/solvent front) was constant
at ca. 60:40. Extraction of product mixtures with ether efficiently
removed M1, which capillary GC/MS showed to be a single
compound having m/z 172 (which corresponds to 2 - H + CN).
The properties of M1, together with the context of its formation,
suggested it might be either 3 or 4a, but after synthesizing 334 it
was found to be separable from M1 by both HPLC and capillary
GC/MS. We isolated M1 by preparative HPLC from larger scale
means of HRP oxidation of N-methyl-1,2,3,4-tetrahydroquinoline
(6) in the presence of cyanide. This generated a single product
indistinguishable from M1 by HPLC or capillary GC/MS. Thus,
SET oxidation of 2 gives rise to a reactive intermediate which
can be trapped efficiently by dilute cyanide ion to yield 5 as the
stable end product. In the absence of cyanide this intermediate is
converted to highly polar material which has thus far resisted
identification.
The formation of 5 from 2 requires a net two-electron oxidation.
A mechanism for this transformation is suggested in Scheme 1.
As reviewed by Chen et al.,37 the initial step is a one-electron
oxidation of 2 to 2+• by the oxidized (ferryl) form of HRP. The
next (and key) step is the unimolecular ring-opening of 2+• to
give the distonic cation radical 7. If ring-opening of the cyclo-
propane moiety were nucleophile-assisted, like the ring-opening
of arylcyclopropane cation radicals,38,39 products other than 5
would have been formed. This point is crucial to the application
of cyclopropylamines to probe for the involvement of SET
mechanisms in enzyme-catalyzed N-dealkylations because active
site nucleophiles (including solvent water) might not be readily
available, leading potentially to “false-negative” results.
Formation of 5 from 7 ultimately requires several additional
steps, the order of which is uncertain (and unimportant to the
intended application of cyclopropylamines for detecting SET
mechanisms). One plausible sequence is shown in Scheme 1; an
alternate, if perhaps less plausible, mechanism is shown in Scheme
2. Because cyanide failed to trap iminium ion precursors to 3,
4a, or 4b, we conclude that ring-opening of 2+• is substantially
faster than its deprotonation on either the methyl group or the
cyclopropyl C-1′ position. This again reinforces the suitability
of cyclopropylamines as reporters for detecting aminium ion
intermediates in oxidation reactions through the formation of ring-
opened products. The identity of the solvent front metabolite(s)
whose yield is reduced in the presence of cyanide is under
investigation and will be reported separately.
1
incubations and examined it by H and 13C NMR as well as
COSY, HMQC, and HMBC. These data clearly showed that M1
was not 4a,35 but was very likely 5,36 which we synthesized by
(28) DNP reagent (0.15 M 2,4-dinitrophenylhydrazine dissolved in a
mixture of concentrated H2SO4/water/absolute ethanol, 3:4:14, v/v/v) was
extracted thoroughly with hexanes (to remove numerous interfering contami-
nants) immediately before use. After quenching a 1.0 mL incubation with
0.67 mL of MeCN, hexanes-extracted DNP reagent (15 µL) was added. The
mixture was stirred at room temperature for 30 min and shaken vigorously
with EtOAc (300 µL). After phase separation a 1 µL aliquot of the EtOAc
layer was analyzed by GC/MS and/or a 10 µL aliquot was analyzed by HPLC
(Alltech cyano bonded phase column, 10 µm, 4.6 × 250 mm) eluting with
10% EtOAc in hexanes (solvent A) for 12 min followed by gradient elution
with solvent B (1% i-PrOH in CH2Cl2; 0-100% over 23 min) at a flow rate
of 1.5 mL/min. Standard solutions of expected carbonyl metabolites were
prepared and analyzed similarly.
(29) GC/MS analyses were performed on a HP 5890 Series II GC fitted
with a DB-5 capillary column coupled to a HP 5971A Mass Selective Detector.
Data were collected in the full spectrum mode (40-350 amu, 0.77 scans/s)
and analyzed using an HP 5970 ChemStation. Quenched incubations were
extracted with 300 µL of ether and a 1 µL aliquot of the organic phase was
analyzed by the temperature program: 60 °C for 3 min; 10 °C/min to 250
°C; hold at 250 °C for 5 min.
(30) For the synthesis of [1′-14C]-2, sodium [14C]-formate (1.0 mmol, 500
µCi, Moravek Biochemicals, Brea, CA) was combined with 2.2 mmol of 18-
crown-6 and 1.3 mmol of benzyl bromide in THF (5 mL) and heated in a
sealed screw-cap culture tube (70 °C, 60 h) maintained behind an adequate
safety shield. After cooling, concentration by Vigreux distillation and flash
chromatography (silica gel; 9% ether in pentane), the resulting benzyl [14C]-
formate (0.56 mmol, 282 µCi) was heated with N-methylaniline (500 µL, 4.6
mmol) in a screw-cap culture tube sealed under N2 (95 °C, 48 h). After cooling,
dilution with excess HCl, extraction with ether, concentration by Vigreux
distillation, and flash chromatography (silica gel; 33% ether in pentane), the
resulting N-methyl-[14C]-formanilide (0.39 mmol, 195 µCi) was converted to
[1′-14C]-2 (yield 0.11 mmol, 56 µCi) as described above.
Acknowledgment. Fellowship support for C.L.S. was provided by
the American Foundation for Pharmaceutical Education and an NIH
Predoctoral Training Grant (GM-07775).
(31) Nguyen, T.-L.; Gruenke, L. D.; Castagnoli, N., Jr. J. Med. Chem. 1976,
19, 1168-1169.
JA003048L
(32) Ho, B.; Castagnoli, N., Jr. J. Med. Chem. 1980, 23, 133-139.
(33) Murphy, P. J. J. Biol. Chem. 1973, 248, 2796-2800.
(37) Chen, H.; deGroot, M. J.; Vermeulen, N. P. E.; Hanzlik, R. P. J. Org.
Chem. 1997, 62, 8227-8230.
(34) Zhdankin, V. V.; Kuehl, C. J.; Krasutsky, A. P.; Bolz, J. T.; Mismash,
B.; Woodward, J. K.; Simonsen, A. Tetrahedron Lett. 1995, 36, 7975-7978.
(35) Ahlbrecht, H.; Ibe, M. Synthesis 1988, 210-214.
(38) Dinnocenzo, J. P.; Simpson, T. R.; Zuilhof, H.; Todd, W. P.; Heinrich,
T. J. Am. Chem. Soc. 1997, 119, 987-993.
(36) Le Gall, E.; Hurvois, J.-P.; Renaud, T.; Moinet, C.; Tallec, A.; Uriac,
P.; Sinbandhit, S.; Toupet, L. Liebigs Ann./Recl. 1997, 2089-2101.
(39) Dinnocenzo, J. P.; Zuilhof, H.; Lieberman, D. R.; Simpson, T. R.;
McKechney, M. W. J. Am. Chem. Soc. 1997, 119, 994-1004.