A class of 4-aza-lithocholic acid-derived haptens for the generation of catalytic antibodies with steroid synthase capabilities

Article abstract of DOI:10.1016/S0968-0896(00)00036-5

The syntheses of a class of three haptens derived from the same 4-aza- steroidal skeleton is described. The sequence begins with oxidative cleavage of ring A of commercially available, optically pure lithocholic acid. Insertion of nitrogen at position 4 and stereoselective hydrogenation of the resulting electron-rich enelactam under 600 psi H₂ yielded a system analogous to testosterone-5-alpha-reductase inhibitors. Upon exhaustive reduction of this compound with lithium aluminium hydride, a linker for bioconjugation was attached before the N-oxide key functionality is established in ring A. This functional group is believed to be a true transition-state mimic for the electronic nature of initiation of the cationic cyclization of 2,3-epoxy-squalene derivatives. In addition, it also holds promise for eliciting acidic residues as part of a bait-and-switch strategy. Remarkably, both N-oxide epimers obtained from mCPBA oxidation can be separated by column chromatography on a 60 mg scale and were used in enantiopure form for separate immunizations. Reliable configurative assignment was carried out by comparison studies with previously characterized and published systems. A catalytic antibody (HA8-25A10) was obtained from the immunization with the hapten bearing an aminoxide oxygen in the beta position. Surprisingly, an inhibition study showed that the isomer with the inverted configuration at the N-oxide bound more strongly to this catalytic antibody. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00036-5

Bioorganic & Medicinal Chemistry 8 (2000) 995±1003
A Class of 4-Aza-lithocholic Acid-Derived Haptens for the
Generation of Catalytic Antibodies with Steroid Synthase
Capabilities
Jens Hasserodt,* Kim D. Janda* and Richard A. Lerner*
Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA
Accepted 10 December 1999
AbstractÐThe syntheses of a class of three haptens derived from the same 4-aza-steroidal skeleton is described. The sequence
begins with oxidative cleavage of ring A of commercially available, optically pure lithocholic acid. Insertion of nitrogen at position
4 and stereoselective hydrogenation of the resulting electron-rich enelactam under 600 psi H2 yielded a system analogous to tes-
tosterone-5-alpha-reductase inhibitors. Upon exhaustive reduction of this compound with lithium aluminium hydride, a linker for
bioconjugation was attached before the N-oxide key functionality is established in ring A. This functional group is believed to be a
true transition-state mimic for the electronic nature of initiation of the cationic cyclization of 2,3-epoxy-squalene derivatives. In
addition, it also holds promise for eliciting acidic residues as part of a bait-and-switch strategy. Remarkably, both N-oxide epimers
obtained from mCPBA oxidation can be separated by column chromatography on a 60 mg scale and were used in enantiopure form
for separate immunizations. Reliable con®gurative assignment was carried out by comparison studies with previously characterized
and published systems. A catalytic antibody (HA8-25A10) was obtained from the immunization with the hapten bearing an amin-
oxide oxygen in the beta position. Surprisingly, an inhibition study showed that the isomer with the inverted con®guration at the N-
oxide bound more strongly to this catalytic antibody. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
cationic species along the reaction coordinate. The
dimensions of the catalytic cavity are larger than the
cyclized product hopene and the aromatic residues, that
are expected to stabilize the later cations, are distant
from the initiation site. This is in accord with the notion
that the linear substrate squalene, once coiled, needs a
greater volume of space than the cyclized product. It
was proposed that the substrate slides away from the
aspartate thus reaching aromatic residues that can sta-
bilize carbocations resulting from attack of the previous
cationic center.2 For termination, no basic amino acid
has been discovered that may likely be responsible for
deprotonation. Instead, a localized water molecule,
found at a decisive location within the active site, has
been proposed to play this role.
The enzymes responsible for the cyclization of squalene
or its epoxy derivative are stunningly sophisticated tools
developed by Nature for sterol biosynthesis. These iso-
merases not only generate tetracyclic (eucaryotic oxi-
dosqualene-lanosterol synthase, OSC) or pentacyclic
(bacterial squalene-hopene synthase, SHC) ring systems
from linear poly-ene precursors, but in the course of the
reaction cascade they also create six (OSC) to nine
(SHC) new stereocenters in one chemical transforma-
tion. With the help of mechanistic studies1 and a crystal
structure at high resolution,2 an increasingly re®ned
picture emerges of how these cyclases function. Muta-
genesis studies suggested a single aspartate (D456) to be
the triggering unit in yeast OSC;3 this was con®rmed by
the crystal structure of SHC that revealed the presence
of an aspartate (D376) most likely responsible for initi-
ating the cyclization. This three-dimensional structure
showed how aromatic residues could be positioned for
e€ective interaction of their quadrupoles with transient
Catalytic antibodies are ``currently the most successful
enzyme mimics''.4 It is of particular interest to examine
the possibility of catalyzing the multi-cyclization of a
squalene-derived substrate with an antibody. This for-
midable task involves creating a catalytic site that can
accomodate a large substrate while shielding it e€ec-
tively from solvent molecules. In principle, antibodies
should be able to achieve this, since they can provide a
combining site of close to 2000 A3.5 In addition, it
*Corresponding authors. Tel.: +1-858-784-2516; fax: +1-858-784-
2595; e-mail: kdjanda@scripps.edu; ralerner@scripps.edu; jensh@
scripps.edu
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00036-5

996
J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
should be possible to elicit an acidic residue at the
desired location within the active site by choosing a
haptenic structure containing a counter-charged moi-
ety.6 An N-oxide moiety not only has this property7
but it also displays a strong polarization as seen with
the heterolytic cleavage of an oxygen±carbon bond in
a potential epoxysqualene-derived substrate. N-oxide
derivatives of squalene have shown superior inhibitory
e€ects over their parent amines on natural OSCs.8 This
functional group may ful®ll both roles, acting as a bait-
and-switch tool9 as well as a transition-state analogue.
The extension of the cyclization cascade (propagation)
consists of the catalyst enforcing a tight coiled con-
formation of the otherwise linear substrate, thus bring-
ing double bonds and their respective carbocations into
close proximity for barrier-less collapse.10 This action
by the catalyst may be installed into the antibody com-
bining site by using a hapten where the cyclized struc-
ture has been preselected in form of a hydrophobic
steroidal skeleton.
Steroidal antigens (progesterone, digoxin and oestrone)
have been shown to elicit numerous aromatic residues
for binding during immunization.11,12 However, the
precise constellation of aromatic quadrupoles for opti-
mal stabilization and orientation of incipient carboca-
tions cannot be elicited by any current methodology.
Instead, the strategy here consists of the following: (a)
creating diversity in combining site architecture
(accomplished by action of the immune system itself)
with the help of proper pre-synthesis selection of a hap-
tenic structure; (b) Ecient post-synthesis selection from
the obtained pool of antibodies with pre-determined
speci®city.13
Scheme 1.
Results and Discussion
led to such widespread products as ®nasteride1 and
proscar1 for the treatment of benign prostatic hyper-
trophy, acne and hair loss in males.15 The long history
of azasteroids dates back to the 1930s when Bolt inves-
tigated the introduction of nitrogen at the 4-position in
the steroidal skeleton.16 In this investigation, the start-
ing material used was the oxocarboxylic acid obtained
from ozonation or oxidation with potassium perman-
ganate of cholestenone which resulted in the opening of
the A ring.17 Ring closure to form a piperidine system
within the steroid framework was achieved by sodium/
ethanol reduction of the corresponding oxime of this
carboxylic acid. This classic approach has somewhat
changed over the years to ensure selective formation of
the isomer bearing a trans fusion between ring A and B.
For this purpose, ring closure has to be enforced under
non-reducing conditions to obtain an enamide moiety.
This functionality is subsequently reduced stereo-
selectively to the corresponding 5-trans-lactam. More
recently, numerous e€orts have been made to accom-
plish complete diastereoselectivity. Here, the goal was to
reduce costs caused by puri®cation in industrial pro-
duction. Catalytic hydrogenation appears to be the
method of choice, yielding trans/cis ratios of up to 500:1
for speci®c substrates.18 However, others could not obtain
appreciable isomer ratios by catalytic hydrogenation
A dual-anchor design displaying functional groups at
the head (epoxide) and the tail (carboxylic acid/-amide
or 4-cyanobenzyl ether) of the otherwise hydrophobic
squalene-derived substrate was chosen in this investiga-
tion (Scheme 1). These functionalized ends should give
the antibody extra opportunities for speci®c molecular
recognition in the form of hydrogen bonds. The devel-
opment of this strategy has been supported by observa-
tions with an anti-oestrone antibody that particularly
recognizes both functionalized `ends' of the steroid
hapten.11 Thus, haptens 1a,b and 20a,b were targeted
for the above considerations. They di€er in the concept
of linker attachment, one being conjugated through its
head to carrier proteins (20) and the other through its
tail (1). The cyanobenzyl ether moiety in 20 was chosen
for convenient UV monitoring of substrate depletion
and any formed products in future assays while leaving
the opportunity to deprotect the hydroxyl group in any
formed products.
4-Aza-steroids, to which 1a,b and 20a,b belong, are
important targets as inhibitors of testosterone 5a-reduc-
tases and as antagonists of the androgen receptor.14 The
resulting 5a-dihydrotestosterone was shown to be the
more potent intracellular hormone. This research has

J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
997
and reported better results with acid-catalyzed reduc-
tion by complex hydrides.19
under 600 p.s.i. H2 for 5 days. An extended reaction
time of 5 days is necessary due to the electron-rich nat-
ure of the double bond and its steric congestion at the
ring junction. A variety of conditions reported in the
literature18 employing more expensive rhodium or pla-
tinum catalysts were applied to 7, however, they did not
appreciably lower the percentages of the cis-fusion. The
degree of stereoselectivity was determined by integra-
tion of the corresponding proton NMR signal for H5
(trans/cis: 10:1). This signal was also used to reliably
monitor the progress of reduction since the Rf values for
product and starting material were identical. The minor
isomer was carried through the synthesis alongside
the target compound and was chromatographically
removed at a later stage. Subsequent transformation of
the terminal carboxamide functionality into its methyl
ester (acidic amberlyst 15 in re¯uxing methanol22) furn-
ished 9. This was then exhaustively reduced with
LiALH4 to amino alcohol 10.23 Selective protection of
the nitrogen with either Cbz or Boc led into the two
separate synthetic pathways for both haptens.
The synthesis of haptens 1a,b and 20a,b starts from
commercially available lithocholic acid (Scheme 2). The
cis con®guration at C5 of this steroid is of no relevance
since it is eliminated during the ring opening/closure
procedure. The ®rst sequence of steps involves oxidative
bromination to yield the 4b-bromo ketone that is
directly treated with lithium carbonate in DMF to fur-
nish enone 5.20 Oxidative cleavage of this material with
a sodium periodate/potassium permanganate system in
re¯uxing tert-butanol a€ords ketodiacid 6 in high yields
only, if 5 has been thoroughly puri®ed.21 The key step
of nitrogen introduction involves bringing a solution of
6 in ethylene glycol close to its saturation point at 5 ꢀC,
quickly adding liquid ammonia, heating this mixture at
5 ꢀC/min to 180 ꢀC and ®nally cooling the solution.
Thus, a 9:1 mixture of primary amide 7 and acid 8 is
obtained in 83% yield.
This mixture is directly introduced into the next step,
catalytic stereoselective hydrogenation with 10%Pd/C
Completion of haptens 1a,b were as follows (Scheme 3).
Upon PDC oxidation of 11 to acid 13,24 1,1-dimethyl-
ethyl 6-amino-caproate (generated from Cbz-protected
1,1-dimethylethyl 6-amino-caproate by standard hydro-
genation) was attached by amide coupling to yield 14.
Selective deprotection of N4 led to a secondary amine
that was transformed into methylamine 15 by reductive
amination.25 At this stage, the 5-cis isomer was removed
by two rounds of ¯ash chromatography on a 160 mg
scale. Compound 15 was oxidized to an epimeric mixture
of aminoxides (16a,b) with mCPBA at 0 ꢀC. The mixture
Scheme 2. Conditions: (a) NBS, dioxane/H2O; (b) Li2CO3, DMF,
90 ꢀC; (c) KMnO4, NalO4, tBuOH re¯ux, 89%; (d) liquid NH3,
HOCH2CH2OH, 5±180 ꢀC, 83%; (e) 600 p.s.i. (40 bar H2, 10% Pd/C,
glacial HOAc, PARR BOMB, 5 d; (f) amberlyst 15, MeOH, re¯ux,
75% over 2 steps; (g) LiAlH4; (h) CbzCl (9->11), 60±68% over 2
steps.
Scheme 3. Conditions: (a) pyridinium dichromate, 71%; (b)
H2N(CH2)5COOtBu, EDC, NEt3, DMF, 91%; (c) H2, 10% Pd/C; (d)
40% formalin, NaBH4, 25 ꢀC, 70% over 2 steps; (e)mCPBA, 61%
separation of epimers; (f) TFA, 84%.

998
J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
was separated by ¯ash chromatography on a 60-mg
scale and both isomers were characterized before
deprotection to the target haptens 1a and 1b. The
assignment of their absolute stereochemistry was made
possible through the distinctive proton-NMR low-®eld
shift (0.35 ppm) of the C19-methyl group in 16a caused
by its through-space interaction with the axial amin-
oxide-oxygen. This behavior has been previously
observed in a comparable system involving a perhydro-
quinoline N-oxide system where it was con®rmed
through ROESY spectroscopy as well as an X-ray
structure.26,27
initiated the cyclization through epoxide opening and
controlled the formation of ring A of the steroid nucleus
(Scheme 5).29 However, propagation to the desired
multicyclic ring system with these particular substrates
(2 and 3) was not observed. HA8-25A10 regioselectively
produced two ole®nic regioisomers (21 and 22) with a
ratio of 93:7 while kinetically resolving racemic 2 and 3
(only (S)-isomer was consumed).
An inhibition study of IgG HA8-25A10 with the epi-
meric haptens 1a and 1b was performed to estimate the
impact of a single inversion at an asymmetric center.
Thus, we hoped to derive factors that may provide
insight for the successful elicitation of a catalyst. Since
both haptens were found to be tight-binding inhibitors,
the `classical' approach to Ki determination is not
applicable.30 Instead of studying the e€ect of inhibitor
concentration on enzyme activity under conditions of
varying substrate concentrations, a titration under a
®xed substrate concentration can be conducted.31 The
dose±response plot constructed from this data furnishes
the IC50 value for the corresponding inhibitor (Fig. 1).
With the knowledge of the Michaelis±Menten para-
meters, a Ki value for 1a of 9 mM has thus been deter-
mined. Surprisingly, the Ki for hapten 1b has been
found to be 0.5 mM! This result demonstrates that 1b
has the better ®t to the active site of HA8-25A10.
The synthesis of haptens 20a,b continued with cyano-
benzylation of 12 to form 17 (Scheme 4). This com-
pound was deprotected at N4 before chromatographic
removal of the 5-cis isomer. Alkylation of this iso-
merically pure secondary amine with 1,1-dimethylethyl-
6-bromohexanoate a€orded 18.28 Here too, N-oxidation
with mCPBA led to a mixture of epimers (19a,b) that
could be characterized separately after column chroma-
tography on a small scale before deprotection to the
target haptens 20a,b.
Three separate immunizations with haptens 1a, 1b and a
1:1 mixture of 20a,b, respectively, were carried out and
to date, the immune responses to 1a and 1b were har-
vested according to standard hybridoma protocol.
Screening of the obtained sets of monoclonal antibodies
with substrates 2 and 3 demonstrated that hapten 1a
did indeed elicit a catalyst. This antibody, HA8-25A10,
Scheme 5.
Scheme 4. Conditions: (a) 4-NC-C6H4CH2Br, NaH, 91%; (b) TFA,
CH2Cl2, 1 h; (c) Br(CH2)5COOtBu, NEtiPr2, DMF, 60 ꢀC, 40 h, 57%;
(d) mCPBA, CH2Cl2, 20 ꢀC->rt, 4 h, 99%, separation of epimers; (e)
TFA, CH2Cl2, 4 h, 99%.
Figure 1. Dose±response plot of inhibition by 1a (HA8) and 1b (HA9).
Conditions: 5 mM IgG HA8-25A10; 400 mM substrate 2; 0.2%
TWEEN 80; 50 mM phosphate bu€er; pH 7.0; 37 ꢀC. IC50 values
derived from plots: 18.9 mM (1a), 5.7 mM (1b).

J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
999
However, 1b did not yet lead to the discovery of any
antibody catalysts by conventional hybridoma technol-
ogy. As to why 1b did not lead to any catalytic anti-
bodies, ultimately only a crystal structure determination
of HA8-25A10 complexed with 1a and 1b will answer
this intriguing question.
stirring with an oversized stirring bar, resulting in some
precipitation. A heated solution of 28.5 g NaIO4 and
0.21 g KMnO4 in 74 mL H2O (that had to be boiled
1/2 h to dissolve all salts) was added dropwise under
vigorous stirring causing strong CO2 evolution, further
solid formation and a tan-like color. The mixture was
re¯uxed an additional 10 min and cooled to room tem-
perature. Filtration through Celite, extensive washing
with 1 M Na2CO3, water, evaporation of t-BuOH, and
acidi®cation with 1 N HCl a€orded a precipitate that
was ®ltered and washed with water. The obtained solid
was dissolved in CH2Cl2, extracted with H2O and brine
before being dried over MgSO4. Silicagel chromato-
graphy (column conditioned with hexanes:ethyl acetate
5:1 plus 0.1% glacial HOAc, gradient 5:1!1:1) yielded
6.713 g of 6 (17.10 mmol, 89%, oil). (The high purity of
the starting material is essential for this yield.) 1H NMR
(CDCl3): d 0.73 (3H, s), 0.93 (3H, d, J=6.5 Hz), 1.12
(3H, s), 2.27 (3H, m), 2.38 (2H, m), 2.54 (1H, dt,
J=14.5, 6.0 Hz). 13C NMR (CDCl3): d 11.9, 12.0, 18.1,
20.1, 21.4, 24.1, 27.9, 29.1, 30.6, 31.0, 31.3, 34.8, 35.2, 38.1,
39.2, 42.5, 47.8, 50.3, 55.6, 55.7, 179.9, 180.3. LRMS
(FAB, NBA/NaI) calcd for C23H36O5 (M+Na+) 415;
found 415.
Conclusion
This study has further paved the road towards ®nding
answers how nature controls poly-ene cyclization. Con-
cise hapten syntheses have been presented. A catalytic
antibody (HA8-25A10) forming a monocyclic product
has been elicited by hapten 1a. However, routine pro-
duction of monoclonal antibodies from immunization
with epimer 1b did not lead to any catalyst. A di€erence
in binding energies with the catalyst HA8-25A10 has
been observed between 1a and its epimer 1b. Surpris-
ingly, 1b, that was not present during immunization,
bound stronger. Future studies should provide a clearer
picture of how binding energy e€ects catalysis and how
an increase in apparent congruency of alternate sub-
strates to haptens 1a,b e€ects product outcome.
Enelactam-amide and -acid (7,8). Very pure 6 (5.417 g,
13.8 mmol) was dissolved in 25.4 mL of warm ethylene
glycol (1 L single-necked round-bottom ¯ask with con-
denser) before cooling to 5 ꢀC under stirring (Note:
some precipitation may occur). Next, liquid ammonia
(5.8 mL) were cautiously added in one portion causing
only limited losses through evaporation. The suspension
clears immediately and turns from colorless to light
yellow (enamine formation). The solution in the open
apparatus was heated slowly (3 ꢀC/min) to 180 ꢀC. At
160 ꢀC a ®nely powdered solid precipitates and the
mixture is kept at 180 ꢀC for 15 min while turning grey-
brown. Upon cooling to room temperature the mixture
was diluted with 500 mL H2O and acidi®ed to pH 1.0
with concd HCl. The solid was removed by ®ltration
and washed several times with water before being sus-
pended in THF. Filtration and washing a€ords a com-
pletely pure white powder (7) (2.665 g, sparingly soluble
in organic solvents except for DMSO). The recombined
washings yield 1.63 g of a crude mixture of 7 and 8
(Rf=0.18, CH2Cl2:MeOH 15:1). Combined yield: 4.29g,
83%.
Experimental
General procedures
If not stated otherwise, 1H NMR and 13C NMR spectra
were recorded on a Bruker AMX-500 instrument at 500
and 125 MHz, respectively. Only selected signals are
quoted for the proton spectra. Generally, these include
those for steroidal positions 3, 4, 5, 18, 19, 21, 23 and
24, and those caused by protons of the linkers, the N-
methyl group, and the cyanobenzyl group. Chemical
shifts (d) are given in ppm relative to CHCl3 in CDCl3
(7.27 ppm, H; 77.00 ppm, 13C). Signals are quoted as s
1
(singlet), d (doublet), t (triplet), q (quartet), p (pentet), o
(octet) and m (multiplet). High-resolution mass spectra
(HRMS) were recorded at The Scripps Research
Institute on a VG ZAB-ZSE mass spectrometer under
fast atom bombardment (FAB) or electrospray conditions.
All reactions were monitored by thin-layer chromato-
graphy (TLC), using 0.25 mM Merck Silicagel Glass
Plates (60F-254), fractions being visualized by UV light,
phosphomolybdic acid solution (or other staining solu-
tions as indicated) with subsequent heat application.
Column chromatography was carried out with Mal-
linckrodt SilicAR 60 silicagel (40±63 mM). Reagent
grade solvents for chromatography were obtained from
Fisher Scienti®c. Reagents and anhydrous solvents were
obtained from Aldrich Chemical Co. and used without
further puri®cation. All reactions were carried out
under anhydrous conditions and an atmosphere of
argon, unless otherwise noted. Reported yields were
determined after puri®cation to homogenous material.
Enamide-methylester (8a). The crude mixture of 7 and 8
(1.63 g) in 20 mL THF was treated with an excess of a
diazomethane solution and stirring was continued over-
night. Evaporation of all the volatile components and
silicagel chromatography on a short column (hex-
anes:ethyl acetate 3:1!1:2) a€orded 674 mg 8a as a
white solid. Further elution with MeOH can a€ord
additional enelactam-amide 7. 1H NMR (CDCl3): d
0.68 (3H, s), 0.91 (3H, d, J=6.5 Hz), 1.06 (3H, s), 2.00
(1H, dt, J=12.7, 3.5 Hz), 2.07 (1H, dt, J=16.9, 5.0 Hz),
2.20 (1H, m), 2.34 (1H, m), 2.44 (2H, m), 3.64 (3H, s),
4.89 (1H, q, J=2.5 Hz), 8.51 (1H, s). 13C NMR
(CDCl3): d 11.8, 18.2, 18.6, 20.8, 24.0, 28.0, 28.3, 29.6,
30.87, 30.94, 31.4, 31.5, 34.0, 35.2, 39.3, 42.4, 47.8, 51.4,
55.6, 56.3, 103.6, 139.8, 169.8, 174.6. HRMS (FAB,
Keto-diacid (6). Enone 520a (7.18 g, 19.27 mmol) in
107 mL t-BuOH was treated with 3.00 g Na2CO3 in
14 mL H2O. The mixture was brought to re¯ux under

1000
J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
NBA/NaI) calcd for C24H37NO3 (M+H+) 388.2852;
found 388.2863.
(1H, qd, J=13.2, 3.4 Hz), 2.64 (1H, td, J=13.0, 2.4 Hz),
2.83 (1H, dd, J=12.5, 3.0 Hz), 3.61 (2H, o, J=5.0 Hz),
4.32 (1H, dm, J=13.0 Hz), 5.07 (1H, d, J=13.0 Hz),
5.11 (1H, d, J=13.0 Hz), 7.31 (1H, m), 7.35 (4H, m).
13C NMR (CDCl3): d 12.0, 12.7, 18.6, 21.0, 22.7, 24.1,
27.5, 28.1, 29.3, 31.8, 32.0, 35.2, 35.5, 38.6, 39.1, 39.9,
42.5, 49.3, 53.8, 56.0, 56.1, 63.6, 66.3, 69.1, 127.7, 127.8,
128.4, 137.2, 155.5. HRMS (FAB, NBA/CsI) calcd for
C31H47NO3 (M+Cs+) 614.2610; found 614.2628.
Lactam-methylester (9). The mixture of 7 and 8 (3.088 g,
8.29 mmol) in 125 mL glacial HOAc was placed in the
glass liner of a 300 mL steel pressure vessel (PARR
Instr. Inc. Pt.-No. 4766/4316) equipped with a te¯on
stirring bar. Approximately 720 mg 10% Pd/C was
added and the mixture was stirred under hydrogen (600
p.s.i./40 bar) for 5 days before being checked for start-
1
ing material consumption by H NMR (formation of
BOC-protected alcohol (12). 568 mg crude aminoalcohol
10 in 4 mL dioxane was treated with (BOC)2O (375 mg,
1.716 mmol) in the presence of 0.57 mL Hunig's base
and stirred overnight. Work-up and chromatography as
for 11 a€orded 411 mg (0.918 mmol, 60% over 2 steps)
H5 signal/decrease of ole®nic H5 signal reveals 70%
conversion). The mixture was next hydrogenated at 600
p.s.i. for an additional day at 50 ꢀC (external oil bath) to
achieve complete conversion. Filtration and evapora-
tion a€ord a white solid (3.273 g, Rf=0.58, CH2Cl2:
MeOH 5:1) that is introduced into the next step,
directly; 45 mg Amberlyst 15 were washed 8 times with
MeOH until the ®ltrate reached pH 6.0. This washed
polymer is added to a solution of the above white solid
(3.000 g, 8.01 mmol) in 120 mL of anhydrous MeOH
and the mixture is re¯uxed for 14 h. After ®ltration, the
polymer is applied to a column and eluted with at least 4
times the volume of the resin of MeOH:NEt3 2:1. The
combined ®ltrates are evaporated and the residue is
subjected to silicagel chromatography (CH2Cl2: MeOH
30:1!10:1) to a€ord crystalline 9 (2.40 g, 6.193 mmol,
75% over 2 steps, Rf=0.42, CH2Cl2: MeOH 10:1),
contaminated with 10% of the 5-cis-con®gurated epi-
mer. 1H NMR (CDCl3): d 0.63 (3H, s), 0.86 (3H, s), 0.88
(3H, d, J=6.5 Hz), 2.16±2.21 (1H, m), 2.28±2.38 (3H,
m), 3.00 (1H, dd, J=12.5, 3.5 Hz), 3.63 (3H, s), 6.59
(1H, s). 13C NMR (CDCl3): d 11.2, 12.0, 18.1, 21.0,
23.9, 27.1, 27.9, 29.5, 30.8, 30.9, 33.2, 34.9, 35.2, 35.5,
39.5, 42.6, 51.1, 51.4, 55.7, 55.8, 60.6, 172.5, 174.6.
HRMS (FAB, NBA/NaI) calcd for C24H39NO3
(M+H+) 390.3008; found 390.3017.
1
crystalline 12. H NMR (CDCl3): d 0.66 (3H, s), 0.88
(3H, s), 0.92 (3H, d, J=6.5 Hz), 1.44 (9H, m), 2.33 (1H,
qd, J=13.2, 3.4 Hz), 2.57 (1H, td, J=13.0, 2.4 Hz), 2.78
(1H, dd, J=12.5, 3.0 Hz), 3.61 (2H, o, J=5.0 Hz), 4.23
(1H, dm, J=13.0 Hz). 13C NMR (CDCl3): d 12.0, 12.7,
18.6, 21.0, 21.1, 22.6, 24.1, 27.8, 28.0, 28.5, 29.3, 31.8,
32.2, 35.3, 35.5, 39.3, 39.9, 42.5, 49.3, 53.9, 56.0, 56.1,
63.6, 68.8, 79.0, 155.4.
Cbz-protected carboxylic acid (13). Compound 11
(491 mg, 1.019 mmol) in 6.5 mL DMF was treated with
1.58 g pyridinium dichromate (PDC, 4.19 mmol) in one
portion at room temperature and stirred overnight.
Addition of water (6.5 mL), exhaustive extraction with
CH2Cl2, re-extraction with water, brine and drying over
MgSO4 a€orded material that was puri®ed by silicagel
chromatography (hexanes:ethyl acetate 7:1!1:1, Rf=
0.45 (1:1)) to yield 361 mg (0.728 mmol, 71%) 13. It
should be noted that a signi®cant unidenti®ed side pro-
duct (44 mg, Rf=0.74, not an aldehyde!) was also iso-
1
lated. H NMR (CDCl3): d 0.66 (3H, s), 0.87 (3H, s),
0.92 (3H, d, J=6.5 Hz), 2.25 (1H, m), 2.39 (2H, m), 2.66
(1H, td, J=13.0, 2.4 Hz), 2.83 (1H, dd, J=12.5, 3.0 Hz),
4.32 (1H, dm, J=13.0 Hz), 5.08 (1H, d, J=12.5 Hz), 5.11
(1H, d, J=12.5 Hz), 7.31 (1H, m), 7.35 (4H, m). HRMS
(FAB, NBA/CsI) calcd for C31H45NO4 (M+Cs+)
628.2403; found 628.2378.
Aminoalcohol (10). Compound 9 (2.40 g, 6.193 mmol)
was dissolved in 40 mL anhydrous THF for enhanced
solubility before being diluted with 280 mL anhydrous
ethyl ether and treated cautiously with 5.5 g lithium
aluminum hydride. After 1 h of re¯ux an additional
280 mL of ether was added before re¯ux is continued for
2 days. The mixture is cooled to 0 ꢀC and 200 mL H2O
are added cautiously before extraction with CH2Cl2.
Filtration of both phases was necessary to achieve
complete phase separation. Washing with H2O and
brine followed by drying over MgSO4 a€orded 2.272 g
crude 10 as an oil (only one detectable spot on tlc,
Rf=0.50, CH2Cl2:MeOH 3:1) that was introduced into
the next step directly.
Cbz-protected tButyl ester (14). A mixture of 13 (97 mg,
0.196 mmol) and 1,1-dimethylethyl 6-amino-caproate
(generated from Cbz-protected 1,1-dimethylethyl 6-
amino-caproate by standard hydrogenation procedure)
in 0.5 mL CH2Cl2 was treated with bis(2-oxo-3oxazolidi-
nyl)phosphinic chloride (BOP-Cl) (51 mg, 0.200 mmol).
Then, ethyldiisopropylamine (Hunig's base) (52 mg,
0.400 mmol) in 0.2 mL CH2Cl2 was added over a period
of at least 30 min. After 1 h of stirring the mixture was
washed with 1 M citric acid, H2O, brine and dried over
MgSO4. Silicagel chromatography with hexanes:ethyl
acetate 7:1!1:1 yielded ®rst 35 mg recovered starting
material 13 and then 76 mg 14 (0.114 mmol, 91% based
on converted material, Rf=0.41 with hexanes:ethyl ace-
tate 1:1). 1H NMR (CDCl3): d 0.64 (3H, s), 0.86 (3H, s),
0.91 (3H, d, J=6.5 Hz), 1.44 (9H, s), 2.21 (2H, t,
J=7.0 Hz), 2.41 (1H, q, J=13.0 Hz), 2.65 (1H, t, J=
13.0 Hz), 2.80 (1H, dd, J=12.5, 3.0 Hz), 3.23 (2H, q),
4.31 (1H, dm, J=13.0 Hz), 5.09 (2H, q, J=13.0 Hz),
5.61 (1H, m), 7.30 (1H, m), 7.34 (4H, m). 13C NMR
Cbz-protected alcohol (11). 1.704 g crude amino alcohol
10 in 30 mL CH2Cl2 were treated with benzyl chloro-
formate (920 mg, 5.39 mmol) in the presence of 1.7 mL
Hunig's base and stirred overnight. Extraction with 1 M
citric acid, H2O and brine followed by drying over
MgSO4 a€orded crude 11 that was puri®ed by silicagel
chromatography (hexanes:ethyl acetate 10:1!1:1,
Rf=0.31 (3:1)). Yield: 1.522 g (3.16 mmol, 68% over 2
1
steps, Rf=0.21 (3:1), white solid). H NMR (CDCl3): d
0.65 (3H, s), 0.87 (3H, s), 0.92 (3H, d, J=6.5 Hz), 2.41

J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
1001
1
(CDCl3): d 12.0, 12.7, 18.3, 20.9, 22.7, 24.0, 24.5, 26.2,
27.5, 28.1, 29.2, 31.8, 32.0, 33.6, 35.1, 35.2, 35.5, 39.0,
39.1, 39.9, 42.5, 49.2, 53.7, 55.87, 55.93, 66.3, 69.1, 80.1,
127.7, 127.8, 128.3, 137.2, 155.5, 173.0, 173.5. HRMS
(FAB, NBA/CsI) calcd for C41H64N2O5 (M+Cs+)
797.3870; found 797.3846.
69.8, 79.8, 80.0, 173.0, 173.4. H NMR (16b, CDCl3): d
0.64 (3H, s), 0.92 (3H, d, J=6.5 Hz), 1.01 (3H, s), 1.45
(9H, s), 2.22 (2H, t, J=7.5 Hz), 2.77 (1H, dm, J=
11.0 Hz), 3.00 (1H, m), 3.01 (3H, s), 3.25 (2H, q,
J=7.0 Hz), 3.34 (1H, m), 3.62 (1H, m), 5.51 (1H, t,
J=4.0 Hz).
Methylamine, tBu-ester (15). A solution of 14 (168 mg,
0.253 mmol) in 3.5 mL EtOH containing 90 mg 10% Pd/
C was stirred under hydrogen for 4 h. After ®ltration
and evaporation the remaining free secondary amine
was dissolved in 2.4 mL THF and then treated with a
mixture of 0.11 mL 37% formaldehyde and 1.1 mL
CH3CN. Subsequent addition of 30 mg NaBH3CN, fol-
lowed by two drops of glacial acetic acid after 10 min.
and further stirring for 1 h resulted in complete
consumption of the starting material. Evaporation of
solvents, redissolution in CH2Cl2, extraction with 1 N
NaOH and brine precedes column chromatography
(2.5Â10 cm) with ethyl acetate:MeOH 20:1!5:1 (every
200 mL of solvent containing 0.5 mL NH4OH). Three
fractions were obtained: 6 mg pure 5-cis-con®gurated
isomer, 31 mg mixed fraction and 59 mg pure 5-trans-
isomer 15. Total yield: 96 mg (0.176 mmol, 70%). Water
solubility of 15 (!) accounts for losses in yield during
work-up. (Rf=0.53 for 5-cis isomer and 0.50 for 15 with
CH2Cl2:MeOH 5:1+4 drops of NH4OH, both stain
brown with ninhydrine). 1H NMR (CDCl3): d 0.63 (3H,
s), 0.90 (3H, d, J=6.5 Hz), 0.92 (3H, s), 1.43 (9H, s),
2.02 (1H, m), 2.14 (3H, s), 2.20 (2H, t, J=7.1 Hz), 2.21
(1H, m), 2.83 (1H, dm, J=11.0 Hz), 3.23 (2H, q,
J=7.0 Hz), 5.59 (1H, t, J=5.0 Hz). 13C NMR (CDCl3):
d 12.0, 18.3, 20.8, 22.1, 24.0, 24.5, 24.7, 26.2, 28.1, 29.2,
30.7, 31.8, 33.7, 34.5, 35.3, 35.5, 37.3, 39.1, 39.9, 42.4,
34.5, 52.7, 55.9, 56.3, 58.8, 73.2, 80.1, 173.0, 173.5. HRMS
(FAB, NBA/CsI) calcd for C34H60N2O3 (M+H+)
545.4682; found 545.4666.
Haptens HA8 (1a) and HA9 (1b). 16a or 16b. (4.5 mg,
8.03 mmol) in 1 mL CH2Cl2 was treated with 1 mL TFA
and stirred for 3.5 h at room temperature, respectively.
Evaporation of all volatile components and subsequent
rinsing of the remaining oily residue with 0.8 mL CHCl3
yielded very pure 1a (HA8) or 1b (HA9) (3.4 mg,
6.74 mmol, 84%), respectively (Rf=0.23 for 1a and 0.26
for 1b with acetonitrile:water 2:1, both stain brown with
1
ninhydrine). H NMR (1a, CD3OD, 400 MHz): d 0.70
(3H, s), 0.95 (3H, d, J=6.5 Hz), 1.20 (3H, s), 1.61 (2H, p
J=7.6 Hz) 2.28 (2H, t, J=7.4 Hz), 3.15 (2H, t, J=
7.0 Hz), 3.30 (1H, m), 3.44 (3H, s), 3.62 (1H, t, J=
1
10.0 Hz), 3.73 (1H, d, J=12.8 Hz). H NMR (1b, CD3
OD, 400 MHz): d 0.70 (3H, s), 0.95 (3H, d, J=6.5 Hz),
1.12 (3H, s), 1.61 (2H, p, J=7.6 Hz), 2.28 (2H, t,
J=7.4 Hz), 3.15 (t, J=7.6 Hz), 3.30 (3H, s), 3.40 (1H,
dd, J=12.4, 2.4 Hz), 3.59 (1H, tm, J=14.1 Hz), 3.90
(1H, d, J=12.1 Hz). HRMS (FAB, NBA/CsI) calcd for
C30H52N2O4 (M+H+) 505.4005; found 505.4023 (for
1b).
Cyanobenzylated BOC-alcohol (17). Compound 11
(103 mg, 0.223 mmol) in 2 mL anhydrous THF was
cooled to 0 ꢀC before NaH (35 mg 60% dispersion in
mineral oil) was cautiously added under stirring. The
mixture was stirred at room temperature for another
hour before the solution was decanted via syringe from
the settled excess NaH and introduced into a fresh
¯ame-dried ¯ask (the excess NaH is washed once with
1 mL anhydrous THF). Next, a-bromo-p-tolunitrile
(53 mg, 0.268 mmol) in 1 mL anhydrous THF was added
and the mixture was re¯uxed for 24 h under nitrogen.
The solvent was evaporated, the residue dissolved in
ethyl acetate, washed with H2O, brine, and dried over
MgSO4. Silicagel chromatography with hexanes:ethyl
acetate 9:1!4:1 ®rst yielded 76 mg 17 (Rf=0.49 with
hexanes:ethyl acetate 3:1, stains blue with anisaldehyde)
and then 37 mg recovered starting material 12. Yield,
N(O)Me, tBu-ester (16a,b). Compound 15 (59 mg,
0.108 mmol) in 1 mL CH2Cl2 at 20 ꢀC was treated with
mCPBA (33 mg, 0.189 mmol, >57% activity=>at
least 1 equiv) and the mixture is slowly warmed up
to room temperature overnight under stirring. Now,
the mixture was chromatographed on a short silicagel
column, conditioned with CHCl3:MeOH:NH4OH
125:25:1, yielding 12 mg of a ®rst fraction enriched in
the minor isomer 16b (ratio ca. 2:1) and 25 mg of pure
major isomer 16a (total yield 37 mg, 66 mmol, 61%). The
mixed fraction is reapplied to a new silicagel column
and eluted with CHCl3:MeOH 20:1!3:1 to obtain
4.5 mg pure minor isomer 16b (HRMS (FAB, NBA/
CsI) calcd for C34H60N2O4 (M+H+) 561.4631; found
561.4647), 1.7 mg slightly contaminated 16a and 2.2 mg
pure major isomer 16a (Rf=0.5 for minor isomer 16b
and 0.44 for major isomer 16a, stains brown with nin-
1
based on transformation: 91%. H NMR (CDCl3): d
0.66 (3H, s), 0.88 (3H, s), 0.92 (3H, d, J=6.5 Hz), 1.45
(9H, s), 1.95 (1H, dm, J=12.7 Hz), 2.34 (1H, qd,
J=13.1, 3.3 Hz), 2.58 (1H, td, J=13.0, 2.4 Hz), 2.79 1H,
dd, J=12.6, 3.3 Hz), 3.47 (2H, td, J=6.5, 2.7 Hz), 4.24
(1H, d, J=13.0 Hz), 4.55 (2H, s), 7.44 (2H, d, J=
8.0 Hz), 7.64 (2H, d, J=8.0 Hz). 13C NMR (CDCl3): d
12.0, 12.7, 18.5, 18.6, 21.0, 21.1, 22.6, 24.1, 26.2, 28.0,
28.2, 28.5, 28.6, 32.1, 32.2, 35.3, 35.5, 38.6, 39.3, 39.9,
42.5, 49.2, 53.9, 56.0, 68.7, 71.6, 71.8, 78.9, 111.1, 118.9,
127.6, 132.2, 144.4, 155.3. HRMS (FAB, NBA/CsI)
calcd for C36H54N2O3 (M+Cs+) 695.3189; found
695.3168.
1
hydrine). H NMR (16a, CDCl3): d 0.62 (3H, s), 0.88
(3H, d, J=6.5 Hz), 1.36 (3H, s), 1.41 (9H, s), 2.19 (3H,
t, J=7.5 Hz), 2.51 (1H, q, J=13.5 Hz), 2.58 (1H, dm,
J=9.5 Hz), 3.09 (3H, s), 3.17 (1H, t, J=12.0 Hz), 3.21
(2H, q, J=6.0 Hz), 3.37 (1H, d, J=10.5 Hz), 5.76 (1H,
t). 13C NMR (16a, CDCl3): d 12.0, 14.1, 17.1, 18.2, 20.0,
20.7, 23.9, 24.5, 26.2, 28.0, 29.2, 30.7, 31.7, 33.6, 34.2,
35.2, 35.4, 36.4, 39.1, 39.5, 42.4, 54.6, 55.7, 56.1, 59.5,
N-Alkylated cyanobenzylether (18). Compound 17
(124 mg, 0.220 mmol) in 2 mL CH2Cl2 was treated with
0.2 mL TFA for 1 h. The volatile components were eva-
porated and the residue in ethyl acetate was extracted

1002
J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
with 1 N NaOH, washed with H2O, brine, and dried
over MgSO4. The free secondary amine in 4.9 mL
anhydrous DMF was stirred with 1,1-dimethylethyl-6-
bromohexanoate (160 mg, 0.637 mmol, 2.9 equiv),
142 mL Hunig's base and a catalytic amount of tetra-
butylammonium iodide for 24 h at 50 ꢀC and subse-
quently for 16 h at 70 ꢀC. The cooled reaction mixture
was partitioned twice between ethyl acetate and water.
The combined organic phases were washed with brine,
dried over MgSO4 and the residue, upon evaporation,
was subjected to silicagel chromatography (CHCl3:
MeOH 50:1!25:1, 11 mg of mixture of A/B cis- and
trans-fused isomers (2:1) elute, 25:1!15:1, 68 mg of
major trans-con®gured isomer 18 elute; Rf=0.54
CH2Cl2:MeOH 6:1, stains brown with ninhydrine).
Yield: 79 mg (0.125 mmol, 57%). 1H NMR (CDCl3,
400 MHz): d 0.63 (3H, s), 0.91 (3H, d, J=6.5 Hz), 1.02
(3H, s), 1.43 (9H, s), 1.96 (1H, dm, J=12.5 Hz), 2.08
(1H, broad signal), 2.20 (2H, t, J=7.4 Hz), 2.39 (1H,
broad signal), 2.74 (2H, broad signal), 3.45 (2H, m),
4.53 (2H, s), 7.43 (2H, d, J=8.1 Hz), 7.62 (2H, d,
J=8.1 Hz). 13C NMR (CDCl3, 100 MHz): d 12.0, 13.7,
18.5, 20.7, 21.0 (broad), 22.5 (broad), 23.7 (broad), 23.9,
24.8, 26.2, 26.8, 28.0, 28.1, 30.6, 32.1, 34.3, 35.3, 35.4,
36.7 (broad), 37.1, 39.8, 42.4, 52.7, 53.0, 53.4, 55.9, 56.1,
69.5, 71.6, 71.8, 80.1, 111.0, 118.9, 127.6, 132.1, 144.3,
172.9. HRMS (FAB, NBA/CsI) calcd for C41H64N2O3
(M+H+) 633.4995; found 633.4974.
m, I+II), 2.45±2.5 (2H, m, one epimer), 2.83 (1H, d,
J=11.5 Hz, one epimer), 3.25 (1H, t, J=10.0 Hz, one
epimer), 3.34 (1H, m, one epimer), 3.47 (2H, t, J=
6.5 Hz, I+II), 3.72 (1H, m, one epimer), 3.81 (1H, m),
3.87 (1H, d, J=11.5 Hz), 3.94 (1H, m), 4.05 (1H, m),
4.55 (4H, s, I+II), 7.45 (4H, d, J=8.0 Hz, I+II), 7.64
(4H, d, J=8.0 Hz, I+II). HRMS (FAB, NBA/CsI) calcd
for C37H56N2O4 (M+H+) 593.4318; found 593.4334.
Acknowledgements
Financial support was provided by The Scripps
Research Institute, The National Institutes of Health
(GM-43858), and The Skaggs Institute for Chemical
Biology (K.D.J.). We thank Ping Fan and Alisa Moore
for help with the hybridoma work. Gary Siudzak and
Geo€ry Barker are gratefully acknowledged for their
expert assistance with the LC/MS analyses.
References and Notes
1. (a) Dang, T.; Abe, I.; Zheng, Y. F.; Prestwich, G. D. Chem.
Biol. 1999, 6, 333. (b) Pale-Grosdemange, C.; Feil, C.;
Rohmer, M.; Poralla, K. Angew. Chem. Int. Ed. Engl. 1998,
37, 2237. (c) Ceruti, M.; Rocco, F.; Viola, F.; Balliano, G.;
Milla, P.; Arpicco, S.; Cattel, L. J. Med. Chem. 1998, 41, 540.
(d) Sato, T.; Abe, T.; Hoshino, T. Chem. Commun. 1998, 2617.
(e) Corey, E. J.; Cheng, H., Baker, C. H.; Matsuda, S. P. T.;
Li, D.; Song, X. J. Am. Chem. Soc. 1997, 119, 1289. (f) Zheng,
Y. F.; Oehlschlager, A. C.; Georgopapadakou, N. H.; Hart-
mann, P. G.; Scheliga, P. J. Am. Chem. Soc. 1995, 117, 670. (g)
Ourisson, G.; Nakatani, Y. Chem. Biol. 1994, 1, 11. (h) Kyler,
K. S.; Novak, M. J. NATO ASI Ser., Ser. C 1992, 381, 3. (i)
Hoshino, T.; Williams, H. J.; Chung, Y.; Scott, A. I. Tetra-
hedron 1991, 47, 5925±5932.
Cyanobenzyl-N(O)COOtBu (19a,b). In a Dewar-ice
bath at 20 ꢀC, 18 (68 mg, 0.107 mmol) in 1.2 mL
CH2Cl2 was treated with mCPBA (32.5 mg, 0.188 mmol,
>57% quality). The stirred mixture was warmed to
room temperature overnight, concentrated and directly
®ltered through a short silicagel column that has pre-
viously been conditioned with CHCl3:MeOH:NH4OH
(100:5:0.5) to remove the reagent. Yield nearly quanti-
tative. In a second column chromatography using
CHCl3:EtOH (20:1!6:1) the two epimers 19b, 19a
can be partially separated. (Rf=0.44 (19b), 0.37 (19a)
CHCl3:EtOH 6:1, stains brown with ninhydrine).1H
NMR (19b, CDCl3, 500 MHz): d 0.64 (3H, s), 0.92 (3H,
d, J=6.5 Hz), 1.04 (3H, s), 1.45 (9H, s), 2.32 (2H, t,
J=7.5 Hz), 2.72 (1H, dm, J=9.5 Hz), 2.89 (1H, t,
J=8.5 Hz), 3.12 (1H, m), 3.24 (1H, td, J=17.0, 5.0 Hz),
3.47 (2H, t, J=6.5 Hz), 3.74 (1H, d, J=13.5 Hz), 4.55
(2H, s), 7.44 (2H, d, J=8.1 Hz), 7.64 (2H, d, J=8.1 Hz).
1H NMR (19a, CDCl3, 500 MHz): d 0.65 (3H, s), 0.92
(3H, d, J=6.5 Hz), 1.41 (3H, s), 1.44 (9H, s), 2.23 (2H,
t, J=7.5 Hz), 3.11 (1H, t, J=8.5 Hz), 3.19 (1H, t, J=
8.5 Hz), 3.28 (1H, d, J=14.0 Hz), 3.33 (1H, td, J=12.5,
4.5 Hz), 3.46 (2H, o, J=4.0 Hz), 4.55 (2H, s), 7.44 (2H,
d, J=8.1 Hz), 7.64 (2H, d, J=8.1 Hz).
2. Wendt, K. U.; Lenhart, A.; Schulz, G. E. J. Mol. Biol. 1999,
286, 175.
3. Corey, E. J.; Cheng, H.; Baker, C. H.; Matsuda, S. P. T.; Li,
D.; Song, X. J. Am. Chem. Soc. 1997, 119, 1277.
4. Kirby, A. J. Angew. Chem., Int. Ed. Engl. 1996, 35, 707.
5. Sinha, S. C.; Barbas, III, C. F.; Lerner, R. A. Proc. Natl.
Acad. Sci. USA 1998, 95, 14603.
6. Janda, K. D.; Weinhouse, M. I.; Danon, T.; Pacelli, K. A.;
Schloeder, D. M. J. Am. Chem. Soc. 1991, 113, 5427.
7. N-oxides are slightly less basic than their parent tertiary
amines: Kruger, T. L.; White, W. N.; White, H.; Hartzell, S.
L.; Kress, J. W.; Walter, N. J. Org. Chem. 1975, 40, 77.
8. Cerutti, M.; Delprino, L.; Cattel, L.; Bouvier-Nave, P.;
Duriatti, A.; Schuber, F.; Benveniste, P. J. Chem. Soc. Chem.
Commun. 1985, 1054.
9. Janda, K. D.; Shevlin, C. G.; Lerner, R. A. Science 1993,
259, 490.
10. Jenson, C.; Jorgensen, W. L. J. Am. Chem. Soc. 1997, 119,
10846.
11. Arevalo, J. H.; Hassig, C. A.; Stura, E. A.; Sims, M. J.;
Taussig, M. J.; Wilson, I. A. J. Mol. Biol. 1994, 241, 663.
12. He, M.; Gani, M.; Livnah, O.; Stura, E. A.; Beale, D.;
Coley, J.; Wilson, I. A.; Taussig, M. J. Immunology 1997, 90,
632.
13. Compare to: Eschenmoser, A. Angew. Chem. 1994, 106,
2455.
14. Li, X.; Singh, S. M.; Labrie, F. J. Med. Chem. 1995, 38,
1158.
Cyanobenzyl-N(O)COOH, 2 epimers (1:1) (20a,b). A
1:1 mixture of 19a and 19b (8.0 mg, 12.3 mmol) was
treated with 1 mL CH2Cl2 and 1 mL tri¯uoroacetic acid
for 4 h at room temperature. The volatile components
were removed and the residue is subjected to high
vacuum to yield 20a,b (7.5 mg, 99%). 1H NMR
(500 MHz, CDCl3): d 0.64 (3 H, s, one epimer), 0.66
(3H, s, other epimer), 0.92 (6H, d, both epimers), 1.08
(3H, s, epimer I), 1.24 (3H, s, epimer II), 2.35±2.45 (4H,
15. Rasmusson, G. H.; Toney, J. H. Annu. Rep. Med. Chem.
1994, 29, 225.

J. Hasserodt et al. / Bioorg. Med. Chem. 8 (2000) 995±1003
1003
16. Bolt, C. C. Recl. Trav. Chim. Pays-Bas. 1938, 57, 905.
17. Windaus, A., Ber., 1906, 39, 2008; Ber. 1917, 50, 133.
18. (a) Bakshi, R. K.; Patel, G. F.; Rasmussen, G. H.;
Baginsky, W. F.; Cimis, G. ; Ellsworth, K.; Chang, B.; Bull,
H.; Tolman, R. L.; Harris, G. S. J. Med. Chem. 1994, 37, 3871.
(b) Miller, R. A.; Humphrey, G. R.; Thompson, A. S. Tetra-
hedron Lett. 1995, 36, 7949.
(f) Singh, H.; Paul, D. J. Chem. Soc. Perkin Trans. 1 1974,
1475. (g) Shoppe, C. W.; Killick, R. W.; Kruger, G. J. Chem.
Soc. 1962, 2275. (h) Gut, M.; Uskokovic, M. J. Org. Chem.
1961, 26, 1943. (i) Gupta, R.; Pathak, D.; Jindal, D. P. Eur. J.
Med. Chem. 1996, 31, 241.
22. Greenlee, W. J.; Thorsett, E. D. J. Org. Chem. 1981, 46,
5351.
19. Morzycki, J. W.; Wilczewska, A. Z.; Zochowska, E.;
Lotowski, Z. Heterocycles 1995, 41, 2729.
23. Singh, H.; Sharma, P. P.; Mathur, R. B. Ind. J. Chem.
1973, 11, 1254.
20. (a) Miyamoto, K.; Kubodera, N.; Mursyama, E.; Ochi, K.;
Mori, T.; Matsunaga, I. Synth. Commun. 1986, 16, 513. (b)
Demir, A. S.; Sabol, M. R.; Jeganathan, A.; Dolence, E. K.;
Watt, D. S. Org. Prep. Proc. Int. 1987, 19, 197.
24. Corey, E. J.; Schmidt, G. Tetrahedron Lett. 1979, 399.
25. Borch, R. F.; Hassid, A. I. J. Org. Chem. 1972, 37, 1673.
26. Hasserodt, J.; Janda, K. D. Tetrahedron 1997, 53, 11237.
27. Paschall, C.; Hasserodt, J.; Jones, T.; Lerner, R. A.; Janda,
K. D.; Christianson, D. W. Angew. Chem., Int. Ed. Engl. 1999,
38, 1743.
28. Gomez-Parra, V.; Sanchez, F.; Torres, T. Synthesis 1985,
282.
29. Hasserodt, J.; Janda, K. D.; Lerner, R. A. J. Am. Chem.
Soc. 2000, 122, 40.
21. For application of this method to other steroidal systems:
(a) Uskokovic, M.; Gut, M. Helv. Chim. Acta 1959, 42, 2258.
(b) Rasmusson, G. H.; Reynolds, G. F.; Steinberg, N. G.;
Walton, E.; Patel, G. F.; Liang, T.; Cascieri, M. A.; Cheung,
A. H.; Brooks, J. R.; Berman, C. J. Med. Chem. 1986, 29,
2298. (c) Piniella, J. F.; Estape, J.; Lupon, P.; Merino, L.;
Puig, M.; Bonet, J.-J.; Brianso, J. L.; Germain, G. Bull. Chem.
Soc. Jpn. 1987 60, 3011. (d) Aboul-Enein, H. Y.; Doorenbos,
N. J. J. Heterocyclic Chem. 1974, 11, 557. (e) Singh, H.; Paul,
D.; Parashar, V. V. J. Chem. Soc. Perkin Trans. 1 1973, 1204.
30. Henderson, P. J. F. Biochem, J. 1972, 127, 321.
31. Copeland, R. A.; Lombardo, D.; Giannaras, J.; Decicco,
C. P. Bioorg. Med. Chem. Lett. 1995, 5, 1947.