Formation of bridge-methylated decalins by antibody-catalyzed tandem cationic cyclization

Article abstract of DOI:10.1021/ja970675a

We report the antibody catalysis of an electrophilic tandem ring forming process that yields a bicyclic ring system at neutral pH. Three closely related decalin systems that represent rings A and B of the steroid nucleus account for 50% of the overall pro

Full text of DOI:10.1021/ja970675a

VOLUME 119, NUMBER 26
J ULY 2, 1997
© Copyright 1997 by the
American Chemical Society
Formation of Bridge-Methylated Decalins by
Antibody-Catalyzed Tandem Cationic Cyclization
Jens Hasserodt, Kim D. Janda,* and Richard A. Lerner*
Contribution from the Department of Chemistry and The Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
ReceiVed March 3, 1997X
Abstract: We report the antibody catalysis of an electrophilic tandem ring forming process that yields a bicyclic
ring system at neutral pH. Three closely related decalin systems that represent rings A and B of the steroid nucleus
account for 50% of the overall products. The linear diene substrate has been designed to mimic the first two isoprene
units of 2,3-oxidosqualene, where the epoxide oxygen has been substituted by an arylsulfonate as leaving group.
The hapten is based on a decahydroquinoline system with an N-oxide functionality as the key structure to elicit a
combining site architecture capable of promoting leaving group release. The kcat for the formation of sulfonic acid
in the catalyzed reaction was determined to be 0.021 min-1. The efficiency of the antibody-catalyzed process is
underscored by the fact that the bicyclic products are not formed in the absence of the antibody catalyst under our
mild conditions.
Introduction
theses of cyclohexanol, cyclopentylmethylcarbinol, and, most
remarkably, a cis-fused cyclopropanocyclopentane.4 We sug-
gested that these reactions proceeded via a protonated cyclo-
propane intermediate that was selectively partitioned along
distinct reaction coordinates by the antibody catalyst. Subse-
quently we were able to cyclize terpenoid-like substrates to form
the cyclohexene core structures of the R- and γ-irones.5 We
now extend these studies to achieve tandem cationic cyclizations
to form bridge-methylated bicyclic systems that resemble rings
A and B of the steroid nucleus (Figure 1).
The conversion of 2,3-oxidosqualene to lanosterol or cyclo-
artenol as part of sterol biosynthesis is one of the most
demanding processes controlled by an enzyme (2,3-oxidosqualene-
lanosterol cyclase, OSC) because in a single step six new
stereocenters and four carbon-carbon bonds are formed.1 In
this regard, we have been interested in eliciting antibodies that
catalyze2 these complex processes. Initially, we have focused
on cationic monocyclization reactions where antibodies were
shown to control the reaction with remarkable efficiency.3 In
these experiments, antibodies catalyzed highly selective syn-
Results and Discussion
Polyene cyclization can be broken up into three events,
initiation, propagation, and termination.6 While each of these
processes is distinct, the success of a polyene cyclization reaction
still depends on the complex interplay between the method of
initiation, the nucleophilicity of the double bonds involved, and
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, June 15, 1997.
(1) (a) Abe, I.; Rohmer M.; Prestwich, G. D. Chem. ReV. 1993, 93, 2189-
2206. (b) Corey, E. J.; Virgil, S. C.; Cheng, H.; Baker, C. H.; Matsuda, S.
P.; Singh, V.; Sarshar, S. J. Am. Chem. Soc. 1995, 117, 11819-11820. (c)
Poralla, K. Bioorg. Med. Chem. Lett. 1994, 4, 285-290.
(2) (a) Schultz, P. G.; Lerner, R. A. Science 1995, 269, 1835-1842. (b)
Janda, K. D.; Shevlin, C. G.; Lo, C.-H. In ComprehensiVe Supramolecular
Chemistry; Murakami, Y., Ed.; Pergamon: Oxford, U.K., 1996; Vol. 4, pp
43-72.
(4) Li. T.; Janda, K. D.; Lerner, R. A. Nature 1996, 379, 326-327.
(5) Hasserodt, J.; Janda, K. D.; Lerner, R. A. J. Am. Chem. Soc. 1996,
118, 11654-11655.
(6) Sutherland, J. K. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Ed.; Pergamon: Oxford, U.K., 1991; Vol. 3, pp 341-377.
(3) Li, T.; Lerner, K. D.; Janda, K. D. Acc. Chem. Res. 1997, 30, 115-
121.
S0002-7863(97)00675-6 CCC: $14.00 © 1997 American Chemical Society

5994 J. Am. Chem. Soc., Vol. 119, No. 26, 1997
Hasserodt et al.
Figure 1. Antibody-catalyzed tandem cationic cyclization.
be essentially a relaxation process. Here, the role of the catalyst
is to enforce the appropriate conformation of the substrate 1
(Figure 1) and stabilize any developing cationic centers during
the cyclization cascade.9 The second six-membered ring in the
hapten design and its functionality at C5 and C6 were designed
to induce an antibody combining site that controls this stage in
the reaction. The half-chair conformation of the second ring
of structures I-III is expected to induce a binding pocket that
causes the otherwise extended conformation of substrate 1 to
adopt the pseudo-chair conformation in which the terminal
double bond is properly positioned to capture the developing
second carbocation as the cascade proceeds.
The termination step generally gives rise to either olefinic
(elimination) or hydroxyl (water addition) products and should
therefore be represented in the hapten in a comparable fashion
(I or III, Figure 2) to elicit hydrophobic or polar amino acid
residues in the binding pocket, respectively. Structure II may
provide a microenvironment of intermediate hydrophilicity
because the epoxide functionality can accept but not donate a
hydrogen bond. Geometrically, structure II still resembles I
because carbon atoms 4a and 5-7 and the 6-methyl substituent
are positioned only slightly out of plane compared to structure
I.10 Structure II has the advantage of having little similarity
with the possible products (alcohols or olefins), thereby
minimizing the chance of product inhibition. Moreover,
structure II might lead to an enhanced immune response by
covalent interaction in the antibody binding pocket (epoxide
opening by a nucleophile), thereby functioning as a reactive
immunogen.11 Finally, the linker, by which the hapten is
coupled to a protein for immunization, is positioned distal to
the second ring, thus favoring an antibody binding pocket in
which the second olefinic process might be placed at a deeply
buried site in the protein. In essence this would help to exclude
water and favor termination by an elimination event.
Figure 2. Hapten design and comparison with structures of known
inhibitors of natural squalene cyclases (LGM ) leaving group mimic).
the mechanism of termination. To properly address these points
in our hapten-design considerations, we started with structures
I-III (Figure 2). On the basis of inhibition studies of a variety
of cyclase enzymes1a (Figure 2), we reasoned that a key feature
of the initiation event is the stabilization of the first carbocation.
For this purpose, we relied on our concept of bait-and-switch
catalysis7 which we have previously shown to be successful in
eliciting antibodies that accomplish monocyclization processes.8
Thus, the N-oxide moiety displayed within structures I-III is
appropriately positioned to elicit amino acid residues in the
antibody combining site for the purpose of initiating the
cyclization reaction via solvolysis of the sulfonate functionality
found within 1. N-Oxides have also been used in inhibition
studies of natural oxidosqualene cyclases according to the
example given in Figure 2.1a
To test these ideas, five specific haptens that represent the
trans- and cis-fused isomers of structures I and II have been
utilized for antibody induction. Their synthesis will be presented
elsewhere.12 One of the five candidates, a diastereomeric
mixture (1:1) of 5a,b belonging to category II (Figure 2 and
Unlike cyclization processes in which a single ring is formed,
a tandem process also requires control over the propagation
event. Given that the activation energy for initiation (formation
of the first cationic center) has already been overcome by the
antibody, the process of ring cyclization can be envisaged to
(9) For two different explanations for carbocation stabilization, see: (a)
Johnson, W. S.; Telfer, S. J.; Cheng, S.; Schubert, U. J. Am. Chem. Soc.
1987, 109, 2517-2518. (b) Dougherty, D. A. Science 1996, 271, 163-
168.
(10) See X-ray structure of synthesis intermediate 14a in: Hasserodt,
J.; Janda, K. D. Tetrahedron 1997, in press.
(11) Wirsching, P.; Ashley, J. A.; Lo, C.-H.; Janda, K. D.; Lerner, R. A.
Science 1995, 270, 1775-1782.
(12) For a complete listing of hapten structures and their synthesis, see
ref 10.
(7) Janda K. D.; Weinhouse M. I.; Danon T.; Pacelli K. A.; Schloeder
D. M. J. Am. Chem. Soc. 1991, 113, 5427-5434.
(8) (a) Li T.; Janda K. D.; Ashley J. A.; Lerner R. A. Science 1994,
264, 1289-1293. (b) Li, T.; Hilton, S.; Janda, K. D. J. Am. Chem. Soc.
1995, 117, 3308-3309.

Formation of Bridge-Methylated Decalins
J. Am. Chem. Soc., Vol. 119, No. 26, 1997 5995
Scheme 1
Scheme 2
homologous compound. The key step involves the transforma-
tion of the tertiary allylic alcohol 10 to the primary allylic
bromide 11, creating a complete isoprene moiety (C5) that can
be coupled to a second, incomplete isoprene moiety (C4) by
reaction with methallyl magnesium bromide. The lack of a
methyl group on this terminal isoprene unit results in simplified
product analysis and should have no substantial electronic effects
on the finally evolving tertiary carbocation in the cyclization
process. The two methyl substituents of the first isoprene unit
of 1 were omitted in order to suppress background solvolysis
involving a stable tertiary carbocation. The finally obtained Z-
and E-mixture of 13 was separated as its mixture of sulfonates
(1 and 14) by preparative HPLC to 95% purity in one
purification step.
The antibody catalyst released sulfonic acid from 1 but not
from the Z-configurated 14, suggesting that only 1 is a
cyclization substrate for this antibody. Release of the leaving
group is a necessary but not sufficient requirement for the
formation of carbocycles from the hydrocarbon portion of 1.
Consequently, we analyzed the organic phase of our biphasic
assay.17 To resolve and reliably assign the major products of
the antibody-catalyzed reaction by gas capillary chromatography,
we produced 10 mg of organic products by antibody catalysis
using 1 as a substrate. These products were separated by silica
gel chromatography into an olefinic fraction (one spot by thin
layer chromatography, containing several unresolved species),
accounting for 70% of the product mixture and four pure
alcohols (30%).18 As was established by GC analysis of the
olefinic fraction, the closely related, regioisomeric decalins 2a-c
were the major antibody products as they represent 70% of the
olefinic mixture, therefore accounting for 50% of the overall
product mixture including hydroxylated products. The identity
of each olefin was determined by comparison to authenic
samples synthesized by an independent route (Scheme 2):
Decalone 16 was obtained through a known procedure19 and
then converted to a 57:43 mixture of 2a,b by a modified Shapiro
reaction.20 This unequal ratio enabled us to assign the two
corresponding GC signals of the authentic compounds 2a,b
unambiguously (see NMR spectra in Supporting Information).
Furthermore, transformation of 16 to 2c by means of a Wittig
Figure 4), successfully elicited several catalytic antibodies with
specificity for substrate 1. All monoclonal antibodies generated
against two haptens belonging to category I did not show any
catalytic activity toward solvolysis of substrate 1 but are
currently being checked for activity toward related substrate
structures. Attempts to synthesize hapten structures of category
III, that reflect the outcome of an antiperiplanar addition (as is
required by the Stork-Eschenmoser hypothesis) of a water
molecule to the terminal double bond, have been unsuccessful.
Diastereomers 5a,b, only differing in the configuration at the
tertiary N-oxide center, were coupled to the carrier protein KLH
(keyhole limpet hemocycanin) and coimmunized as racemates.13
The half-life of hapten 5a was determined under physiological
conditions (37 °C, PBS) to be approximately 100 h, which
strongly suggests that the immune system encountered the
epoxide structure and not one of the two possible hydrolysis
products.14 Twenty-two monoclonal antibodies were selected
from the immune response for their affinity to bind to 5a,b-
BSA conjugate (bovine serum albumin) by ELISA methods.15
When tested for their ability to enhance the solvolysis of
substrate 1 to form acetamidobenzenesulfonic acid, three of the
antibodies were found to be catalytic. HA5-19A4 was by far
the most active and was subsequently studied in more detail.
bromides (see: Ruzicka, L.; Firmenich, G. HelV. Chim. Acta 1939, 22, 392-
396). Instead they adapted the stereospecific Julia olefin synthesis to their
system (Julia, M.; Julia, S.; Guegan, R. Bull. Soc. Chim. Fr. 1960, 1072-
1079). We chose the former pathway to obtain the Z-isomer, too.
(17) 83% hexane/2% chloroform/15% phosphate buffer, 50 mM, pH 7.1;
shaken in a sealed Teflon tube for defined time units. See also: Ashley, J.
A.; Janda, K. D. J. Org. Chem. 1992, 57, 6691-6693.
(18) For a presentation of the TLC and the three alcohol structures, see
also the Supporting Information.
Substrates 1 and 14 were synthesized (Scheme 1) according
to a scheme outlined by Johnson and co-workers16 for a
(13) For a detailed procedure, see Supporting Information of ref 5.
(14) See Supporting Information for a plot of the decrease of 5a over
time.
(15) Ko¨hler, G.; Milstein, C. Nature 1975, 256, 495-497.
(16) Brady, S. F.; Ilton, A. I.; Johnson, W. S. J. Am. Chem. Soc. 1968,
90, 2882-2889. The authors never pursued that synthesis scheme further
because of the difficulty in separating the configurational isomers first
formed in the transformation of allylic tertiary alcohols to primary allylic
(19) (a) Nagata, W.; Kikkawa, I.; Fujimoto, M. Chem. Pharm. Bull.
(Tokyo) 1963, 11, 226-235. (b) Nagata, W.; Kikkawa, I. Ibid. 1963, 11,
289-293.
(20) Chamberlin, A. R.; Stemke, J. E.; Bond, F. T. J. Org. Chem. 1978,
43, 147-154.

5996 J. Am. Chem. Soc., Vol. 119, No. 26, 1997
Hasserodt et al.
Figure 3. Lineweaver-Burke plot of leaving group release from 1
catalyzed by IgG HA5-19A4 (2 µM). Biphasic conditions: 83% hexane,
15% phosphate buffer (50 mM, pH 7.0), 2% CHCl3, room temperature.
reaction led to the complete elucidation of the bicyclic products
formed by antibody HA5-19A4 as well as their ratio.21
1
Three of the four alcohols were identified by H NMR as
follows: They contain the unaltered terminal double bond but
lack the olefinic proton signal of the internal double bond in 1;
two of them are therefore assigned diastereomeric cyclohexyl
alcohols 4a,b (Figure 1), and one is tentatively assigned a six-
membered ring structure18 resulting from 1,2-hydrogen shift
prior to water addition. The fourth alcohol could not be
characterized due to immeasurable low amounts after isolation.
Compounds 2a,b,c were formed by the antibody in a ratio
of 2:3:1, which is in accord with the expected stabilities of these
regioisomers. This suggests that the antibody combining site
does not provide tight regiospecific control on the deprotonation
process as part of the termination. The formation of monocyclic
but not bicyclic alcohols may be explained by the partial
accessibility of the combining site to water. The epoxide
functionality of 5a may leave polar space in the binding pocket
close to the intermediate carbocation of the cyclization process
but not the final carbocation. Under the current reaction
conditions (pH 7.0, biphasic, 50 mM phosphate buffer), the
regioisomeric olefins 2a,b were formed with an average22
enantiomeric excess of 53%, determined by means of chiral gas
capillary chromatography; 2c yielded an ee of 80%.
The kinetic characterization of IgG HA5-19A4 was based
on sulfonic acid release. A Lineweaver-Burke plot (Figure 3)
yielded a kcat of 0.021 min-1 and a KM of 320 µM. The kuncat
for solvolysis was determined to 9.2 × 10-6 min-1, giving a
rate acceleration for leaving group release of 2.3 × 103. The
rate acceleration for formation of the decalin products must be
greater than the observed kcat/kuncat for solvolysis since 2a,b,c
cannot be detected in the uncatalyzed reaction under our assay
conditions. Initial inhibition studies were performed on both
hapten isomers originally used as a mixture for coimmunization
(Figure 4). A dose-response plot23 shows that hapten 5a is a
Figure 4. Inhibition study of IgG HA5-19A4 with haptens 5a,b.
potent inhibitor of catalysis (Ki ) 1.4 µM, IC50 ) 8.2 µM)
whereas 5b has no impact on IgG HA5-19A4 catalysis, strongly
suggesting that 5a induced the antibody we studied. The single
inversion at the tertiary N-oxide center causes the antibody to
distinguish between guest and nonguest, thereby demonstrating
the remarkable precision with which antibodies achieve comple-
mentarity and catalysis. Although the Stork-Eschenmoser
concept requires the leaving group to be expelled pseudo-
equatorially, our hapten design favors axial departure of the
leaving group. It is likely that the N(O)CH2CONH moiety of
the axial leaving group mimic leaves enough conformational
flexibility for the system to sufficiently resemble the transition
state with pseudo-equatorial extrusion of the sulfonate. Indeed,
kinetic analyses from our labs have shown that catalytic
antibodies may display complex binding modes including such
phenomena as induced fit.24
Finally, compound 15 was found not to be a substrate for
IgG HA5-19A4, even though it carries a number of common
structural features with 1. Instead, 15 functioned as a competi-
tive inhibitor (Ki ) 24 µM, IC50 ) 32 µM) when added to the
reaction with substrate 1, proving that the antibody binds to 15
but does not promote its cyclization. This suggests that the
antibody does not form an initial carbocation by mere leaving
group cleavage but rather argues for anchimeric assistance of
the internal double bond in 1. The trisubstituted double bond
in 15 is apparently not as nucleophilic; thus, the higher energy
of activation is not compensated by the antibody due to a lack
of electronic and steric complementarity to the transition state,
involving a six-membered rather than a five-membered ring
system.
(21) Johnson, W. S.; van der Gen, A.; Swoboda, J. J. J. Am. Chem. Soc.
1967, 89, 170-172. The authors prepared a mixture of 2a,b,c from 16 by
reaction with methyllithium and subsequent elimination.
(22) The major enantiomers of 2a,b as one group as well as their minor
enantiomers as the other could not be completely resolved either with
trifluoroacetylated γ-cyclodextrin nor with methylated â-cyclodextrin as
chiral immobile phase; therefore, an average ee was calculated from the
combined peak areas.
(23) Copeland R.; Lomabardo D.; Giannaras J.; Decicco C. Bioorg. Med.
Chem. Lett. 1995, 5, 1947-1952.
(24) Wirsching, P.; Ashley, J. A.; Benkovic, S. J.; Janda, K. D.; Lerner,
R. A. Science 1991, 252, 680-685.

Formation of Bridge-Methylated Decalins
J. Am. Chem. Soc., Vol. 119, No. 26, 1997 5997
Hz), 3.55 (2H, t, J ) 6.6 Hz), 3.62 (2H, t, J ) 5.8 Hz), 4.54 (2H, s),
7.31-7.39 (5H, m).
Conclusion
This paper describes an antibody catalyst accelerating the
transformation of a linear polyisoprenoid-like substrate (1) to
bicyclic olefins. This is to be contrasted to previous studies on
acid-catalyzed cyclizations of analoguous systems in organic
solvents in which the monocyclic alcohols were observed as
the major products.25 The success of an antibody catalyst can
be attributed to the ability of the experimenter to program a
binding pocket that can exert control over several features of
the transition state on the reaction coordinate, thereby favoring
the desired product distribution. In the present case, the major
differences between the catalyzed and the uncatalyzed reaction
would appear to involve both the ability of the antibody to
catalyze initiation and favor the second electrophilic addition.
These events may be concerted in that the presence of an
electron-rich double bond is required for efficient sulfonate
release.
The generation of catalytic antibodies with cyclase activity
and investigations into their mechanism can already be expected
to contribute answers to the question how proteins can direct
electrophilic cyclizations. In addition, such catalysts may
ultimately lead to artificial enzymes that process naturally
occurring polyisoprenoid substrates to various useful terpenoid
and steroid structures. Thus, our focus now turns to the
challenge of creating an artificial biocatalyst for steroid forma-
tion from oxidosqualene-type substrates.
5-(Benzyloxy)-2-(2-oxoethyl)pentanoic Acid, Ethyl Ester (8).
Anhydrous ethanol (46 mL) is placed in a 250 mL two-necked round-
bottom flask, equipped with reflux condenser, drying tube, magnetic
stirring bar and septum, and small pieces of sodium (2.05 g, 89.0 mmol)
are dissolved in it. Ethyl acetoacetate (11.58 g, 89.0 mmol) is added
in one portion, and the mixture is heated to reflux before 7 (22.50 g,
98.2 mmol, 1.1 equiv) is added gradually via syringe. Refluxing is
continued for 8 h followed by filtration from the precipitate, evaporation
in Vacuo, redissolution in ethyl acetate, and filtering through Celite.
The crude product is purified using silica gel chromatography (pentanes/
EtOAc, 20:1 f 2:1) to yield 8 as a colorless oil (17.74 g, 63.8 mmol,
72%, Rf ) 0.67 CH2Cl2/EtOAc (9:1), stains yellow-green with
anisaldehyde). 1H NMR (500 MHz, CDCl3): δ 1.27 (3H, t, J ) 7.1
Hz), 1.61 (2H, m), 1.96 (2H, ψ-q, J ) 7.6 Hz), 2.22 (3H, s), 3.47 (1H,
t, J ) 7.5 Hz), 3.49 (2H, td, J ) 6.3, 1.5 Hz), 4.19 (2H, qd, J ) 7.1,
2.1 Hz), 4.49 (2H, s), 7.28-7.36 (5H, m). 13C NMR (CDCl3): δ 14.0,
25.0, 27.4, 28.8, 59.4, 61.3, 69.6, 72.8, 127.5, 127.6, 128.3, 169.7, 203.1.
LRMS (FAB, NBA/NaI): found for C16H22O4 (M + H+) 279.
6-(Benzyloxy)-2-hexanone (9). A mixture of 8 (18.38 g, 66.1
mmol), 50 mL of ethanol (95%), 140 mL of water, and 45 g of Ba-
(OH)2‚8H2O is heated to reflux for 4 h. Then 350 mL of H2O, 90 mL
of 10% HCl, and 350 mL of Et2O are added, and the organic phase is
washed with brine, dried over MgSO4 to give, upon evaporation, 14.36
g of crude oil, that shows only one stainable spot on TLC (Rf ) 0.60
CH2Cl2/EtOAc (9:1), grey-brown with anisaldehyde). Distillation with
a microdistillation apparatus (120-130 °C, dynamic oil pump vacuum)
yields 9 as a colorless oil (11.11 g, 53.86 mmol, 82%). 1H NMR (500
MHz, CDCl3): δ 1.61-1.68 (4H, m), 2.13 (3H, s), 2.46 (2H, t, J )
7.4 Hz), 3.48 (2H, t, J ) 6.0 Hz), 4.50 (2H, s), 7.28-7.37 (5H, m).
13C NMR (CDCl3): δ 20.5, 29.1, 29.8, 43.3, 69.9, 72.9, 127.5, 127.6,
128.3, 138.4, 208.9. LRMS (FAB, NBA/NaI): found for C13H18O2
(M + H+) 207.
6-(Benzyloxy)-2-vinyl-2-hexanol (10). A stirred solution of vinyl-
magnesium bromide (23.27 mL, 1 M in THF, 28.6 mmol) is cooled to
0 °C and 9 (4.80 g, 23.3 mmol), dissolved in 6.4 mL of Et2O anhydrous,
and 3.21 mL of anhydrous THF is added dropwise. Stirring is continued
for 2 h at 0 °C and for 3.5 h at rt. The mixture is treated with 5.8 mL
of saturated NH4Cl containing 3 drops of NH4OH. The organic layer
is decanted from the solid precipitate, which is then dissolved in slightly
acidic water and extracted with Et2O. The combined organic phasesare
washed with brine and dried over MgSO4, and the solvent is evaporated.
Purification is carried out on a silica gel column (pentanes/EtOAc, 5:1
f 2:1) to yield 10 (4.05 g, 17.3 mmol, 74%, Rf ) 0.27 hexanes/EtOAc
3:1) as a colorless oil. 1H NMR (500 MHz, CDCl3): δ 1.28 (3H, s),
1.40-1.44 (2H, m), 1.53-1.57 (2H, m), 1.60-1.65 (2H, m), 3.48 (2H,
t, J ) 6.5 Hz), 4.51 (2H, s), 5.05 (1H, dd, J ) 10.8, 1.1 Hz), 5.21 (1H,
dd, J ) 17.4, 1.1 Hz), 5.91 (1H, dd, J ) 17.4, 10.8 Hz), 7.29-7.36
(5H, m). 13C NMR (CDCl3): δ 20.6, 27.6, 30.0, 42.0, 70.2, 72.8, 73.2,
111.6, 127.5, 127.6, 128.3, 138.5, 145.1. HRMS (FAB, NBA/NaI):
calcd for C15H22O2 (M + Na+) 257.1517, found 257.1528.
Experimental Section
General Procedures. The 500 MHz H NMR and 125 MHz 13C
1
NMR spectra were recorded on a Bruker AMX-500 instrument.
Chemical shifts (δ) are given in parts per million relative to CHCl3 in
CDCl3 (7.27 ppm, 1H; 77.00 ppm, 13C). Signals are quoted as s
(singlet), d (doublet), t (triplet), q (quartet), qnt (quintet), m (multiplet),
and ψ (pseudo-coupling pattern). High-resolution mass spectra (HRMS)
were recorded at The Scripps Research Institute on a VG ZAB-ZSE
mass spectrometer under fast atom bombardment (FAB) conditions.
All reactions were monitored by thin-layer chromatography (TLC),
using 0.25 mm Merck silicagel glass plates (60F-254), fractions being
visualized by UV light or staining with p-anisaldehyde or phospho-
molybdic acid solutions with subsequent heat application. Column
chromatography was carried out with Mallinckrodt SilicAR 60 silicagel
(40-63 µm). Reagent grade solvents for chromatography were
obtained from Fisher Scientific. Reagents and anhydrous solvents were
obtained from Aldrich Chemical Co. and used as is. All reactions were
carried out under anhydrous conditions and an atmosphere of argon,
unless otherwise noted. Reported yields were determined after purifica-
tion to homogenous material.
3-(Benzyloxy)propyl Bromide (7). NaH (60% disperson in mineral
oil, 6.70 g, 167.5 mmol, 1.29 equiv) is placed in a 1 L one-necked
round-bottom flask, equipped with a large stirring bar, washed three
times with 15 mL of pentane, dispersed in 550 mL of DMF anhydrous,
and kept under nitrogen. Benzyl bromide (23.03 g, 134.7 mmol, 1.03
equiv) is added, and the mixture is cooled to -78 °C under vigorous
stirring, followed by the dropwise addition of 3-bromo-1-propanol
(18.07 g, 130.0 mmol) over 2 h. The reaction flask is kept in the
cooling bath which gradually warms to room temperature (rt) over 4
h, and stirring is continued overnight. Then, 500 mL of water and
450 mL of hexanes are added. After phase separation, the aqueous/
DMF phase is reextracted with 100 mL and 50 mL of hexanes and the
combined organic phases are washed with 250 mL and 125 mL of
brine, dried over MgSO4, and evaporated to give 7 as a light yellow
oil containing, according to TLC (Rf ) 0.45 hexanes/EtOAc, 7:1) and
proton NMR spectrum, 10% benzyl bromide (25.98 g, 102.0 mmol of
7, 78%). This material is introduced into the next step without further
purification. 1H NMR (500 MHz, CDCl3): δ 2.15 (2H, qnt, J ) 6.0
Bromide 11. A stirred solution of 10 (2.42 g, 10.3 mmol) in 2.8
mL of hexanes anhydrous at -10 °C, containing 0.26 mL of anhydrous
pyridine, is treated with a solution of PBr3 (1.37 g, 0.48 mL, 5.05 mmol,
1.47 equiv) in 1.4 mL of hexanes. A white precipitate forms
immediately. The mixture is warmed to rt over 1.5 h before being
cooled to 0 °C, quenched with 0.6 mL of saturated NaHCO3, and
extracted with Et2O. The combined organic phases are washed with
brine, dried over MgSO4 to give, upon evaporation, crude 11 as a
colorless oil (2.83 g, 9.5 mmol, 92%, Rf ) 0.62 hexanes/EtOAc (3:1),
stains blue with anisaldehyde), suitable for the next step according to
TLC and proton NMR.
(5E,Z)-5,9-Dimethyldeca-5,9-dienyl Benzyl Ether (12). Magne-
sium turnings (2.43 g, 100.0 mmol) are suspended in 20 mL of
anhydrous THF, and 3-chloro-2-methylpropene (4.94 mL, 4.53 g, 50.0
mmol) in 15 mL of THF is added slowly while stirring at 0 °C.
Afterward the mixture is stirred at rt for 1 h, before the supernatant
solution is decanted by a syringe and added to a solution of crude 11
(2.83 g, 9.5 mmol) from the previous step in 4 mL of anhydrous THF.
After considerable heat evolution has ceased, the mixture is heated to
reflux for 0.5 h with a condenser attached to the reaction flask before
(25) Johnson, W. S.; Crandall, J. K. J. Org. Chem. 1965, 30, 1785-
1790.

5998 J. Am. Chem. Soc., Vol. 119, No. 26, 1997
Hasserodt et al.
being stirred overnight. The mixture is quenched at 0 °C with saturated
NH4Cl (made basic with a few drops of NH4OH), extracted with Et2O,
the combined organic phases washed with brine, and dried over MgSO4
to give, upon evaporation, 2.53 g of crude product (Rf ) 0.44 hexanes/
benzene (1:1), stains turquoise with anisaldehyde) as a colorless oil.
Silica gel chromatography (hexanes/Et2O, 20:1 f 3:1) gives 12 (2.15
g, 7.9 mmol, 77% over two steps) as a mixture of Z- and E-isomers
(31:69). 1H NMR (500 MHz, CDCl3): δ 1.46-1.49 (2H, m), 1.57-
1.66 (2H, m), 1.60 (3H, s, E-), 1.68 (3H, s, Z-), 1.726 (3H, s, Z-),
1.734 (3H, s, E-), 1.98-2.06 (4H, m), 2.10-2.16 (2H, m), 3.47-3.50
(2H, m), 4.51 (2H, s), 4.69 (1H, s, br), 4.71 (1H, s, br), 5.11-5.15
(1H, m), 7.27-7.31 (1H, m), 7.34-7.36 (4H, m). LRMS (FAB, NBA/
NaI): found for C19H28O (M + Na+) 295.
(5E,Z)-5,9-Dimethyldeca-5,9-dienol (13). A solution of 12 (100
mg, 0.37 mmol) in 1.7 mL of Et2O anhydrous is added dropwise to a
stirred solution of 50 mg of sodium in 5 mL of liquid ammonia and 2
mL of Et2O anhydrous at -78 °C. The cooling bath is removed and
stirring is continued for 0.5 h until 10 mL of saturated NH4Cl and 10
mL of Et2O are added cautiously. After evaporation of NH3, the phases
are separated, the aqueous phase is extracted with Et2O, and the
combined organic phases are washed with brine and dried over MgSO4
to give, upon evaporation, 13 as a colorless oil (71 mg, 0.37 mmol,
100%). TLC check shows no side product formation and no starting
material left (Rf ) 0.26 hexanes/EtOAc, 3:1, yellow + blue over one
another [Z+E] with anisaldehyde). LRMS (FAB, NBA/NaI): found
for C12H22O (M + H+) 183.
Sulfonates 1 and 14. A solution of 13 (32 mg, 0.18 mmol), in 0.4
mL of pyridine anhydrous is treated with N-acetylsulfanilyl chloride
(155 mg, 0.66 mmol, 3.8 equiv) and a few crystals of DMAP and stirred
for 5 h. The mixture is washed with 5% citric acid and extracted with
CH2Cl2. The combined organic phases are washed with brine and dried
over MgSO4, and the residue is purified, upon evaporation, using silica
gel chromatography (hexanes/EtOAc, 15:1 f 1:1) to give a mixture
of 1 and 14 as a slightly yellow oil (48 mg, 0.13 mmol, 72%, Rf )
0.28 hexanes/EtOAc (1:1), stains turquoise with anisaldehyde). HRMS
(FAB, NBA/NaI): calcd for C20H29O4S (M + Na+) 402.1715, found
402.1731. Separation of the Z- and the E-isomers (31:69) is carried
out on preparative HPLC [Vydac 201HS1022, 300 Å silica, 10 µm,
C18, 250 × 22 mm, isocratic elution of CH3CN/H2O (no TFA!) (60:
40), flow rate 20 mL/min].
4-(Acetylamino)phenylsulfonic Acid, (5E)-5,9-Dimethyldeca-5,9-
dienyl Ester (1). 1H NMR (500 MHz, CDCl3): δ 1.40 (2H, Ψ-qnt, J
) 7.8 Hz), 1.55 (3H, s), 1.61 (2H, m), 1.72 (3H, s), 1.93 (2H, t, J )
7.5 Hz), 2.01 (2H, t, J ) 7.1 Hz), 2.10 (2H, Ψ-q, J ) 7.8 Hz), 2.23
(3H, s), 4.02 (2H, t, J ) 6.5 Hz), 4.66 (1H, Ψ-q, br, J ) 0.6 Hz), 4.70
(1H, s, br), 5.06 (1H, t, J ) 6.9 Hz), 7.68 (1H, s, br), 7.72 (2H, d, J )
8.7 Hz), 7.83 (2H, d, J ) 8.7 Hz). 13C NMR (CDCl3): δ 15.7, 22.4,
23.4, 24.7, 26.1, 28.2, 37.7, 38.7, 70.8, 109.8, 119.3, 124.7, 129.2, 130.5,
134.2, 142.8, 145.8, 168.7.
4-(Acetylamino)phenylsulfonic Acid, (5Z)-5,9-Dimethyldeca-5,9-
dienyl Ester (14). 1H NMR (500 MHz, CDCl3): δ 1.40 (2H, m), 1.60-
1.66 (2H, m), 1.63 (3H, s), 1.71 (3H, s), 1.99 (4H, Ψ-t, J ) 7.9 Hz),
2.04-2.09 (2H, m), 2.24 (3H, s), 4.03 (2H, t, J ) 6.4 Hz), 4.66 (1H,
Ψ-q, br, J ) 0.8 Hz), 4.71 (1H, Ψ-q, br, J ) 0.6 Hz), 5.13 (1H, t, br,
J ) 6.7 Hz), 7.60 (1H, s, br), 7.72 (2H, d, J ) 8.7 Hz), 7.84 (2H, d,
J ) 8.7 Hz). 13C NMR (CDCl3): δ 22.5, 23.2, 23.6, 24.8, 26.0, 28.6,
30.9, 37.9, 70.7, 109.9, 119.2, 125.4, 129.2, 130.6, 134.4, 142.8, 145.7,
168.7.
Acknowledgment. This Research was supported by a NATO
stipend through the DAAD, Germany (J.H.), NIH Grant GM-
43858 (K.D.J.), and The Skaggs Institute for Chemical Biology
(K.D.J.). We thank D. M. Kubitz, M.-R. Benitez, and A. Neville
for technical assistance with the hybridoma work, Drs. D.-H.
Huang and G. Siudzak for their assistance with NMR and mass
spectroscopy, respectively, and Dr. D. Weiner for helpful
comments on the manuscript.
Supporting Information Available: A listing of the deter-
minations of Ki and half-life of 5a, the GC chromatograms of
antibody product distribution, the inhibition diagram of 15, the
characterization data of compounds 1-2c, 6-16, and a synthesis
scheme of 16 (38 pages). See any current masthead page for
ordering and Internet access instructions.
JA970675A