Synthesis of Thiamacrocycles and Conformational Studies on Their Precursors

Article abstract of DOI:10.1021/jo991139z

1-(4-Nitrophenyl)-7-phenylthioheptane (1) and -9-phenylthiononane (2) have been synthesized and their conformations studied in solution and in the solid state. MMX calculations suggest that the global energy minimum structures are bent in the gas phase, probably owing to edge-to-face intramolecular attractive interaction between the electron rich and the electron poor terminal aryl groups. These conformations were confirmed in solution using 2D NOESY NMR. In the solid state, 1 and 2 exist in the staggered, linear conformation, stacked head-to-tail, with the plane of the nitro group being tilted above the plane of the benzene ring. It appears that the crystal lattice forces overcome the weak edge-to-face intramolecular aromatic interactions that dominate in the gas phase and in solution. The corresponding azides were treated with trifluoromethanesulfonic acid to generate the nitrenium ions, which underwent intramolecular ring-closure to give the corresponding 17- and 19-membered ring thiamacrocycles in modest yields. These results support the suggestion that MMX calculations on appropriate model compounds may be useful in predicting which precursors will lead to macrocycles and which will not.

Full text of DOI:10.1021/jo991139z

J . Org. Chem. 2000, 65, 343-351
343
Syn th esis of Th ia m a cr ocycles a n d Con for m a tion a l Stu d ies on
Th eir P r ecu r sor s
Rudolph A. Abramovitch,* Xiaocong Ye, William T. Pennington, George Schimek, and
Dariuz Bogdal
Department of Chemistry, Clemson University, Clemson, South Carolina 29634-0973
Received J uly 16, 1999
1-(4-Nitrophenyl)-7-phenylthioheptane (1) and -9-phenylthiononane (2) have been synthesized and
their conformations studied in solution and in the solid state. MMX calculations suggest that the
global energy minimum structures are bent in the gas phase, probably owing to edge-to-face
intramolecular attractive interaction between the electron rich and the electron poor terminal aryl
groups. These conformations were confirmed in solution using 2D NOESY NMR. In the solid state,
1 and 2 exist in the staggered, linear conformation, stacked head-to-tail, with the plane of the
nitro group being tilted above the plane of the benzene ring. It appears that the crystal lattice
forces overcome the weak edge-to-face intramolecular aromatic interactions that dominate in the
gas phase and in solution. The corresponding azides were treated with trifluoromethanesulfonic
acid to generate the nitrenium ions, which underwent intramolecular ring-closure to give the
corresponding 17- and 19-membered ring thiamacrocycles in modest yields. These results support
the suggestion that MMX calculations on appropriate model compounds may be useful in predicting
which precursors will lead to macrocycles and which will not.
In tr od u ction
in very good yield and attributed this to a favorable
conformation of the cyclization precursor, indicated by
molecular modeling to have the two interacting sites
within 4.86 A of each other, resulting in a low activation
energy (-187.5 kJ /mol) and a favorable entropy for
cyclization.
In 1989 we reported the formation (in 30% isolated
yield) of a 16-membered ring involving intramolecular
aromatic amination by an arylnitrenium ion under
normal solution concentration (40 mM) conditions).8 It
was suggested that, in a long flexible chain bearing
electron-acceptor and electron-donor end groups, these
groups could “recognize” each other and get close enough
such that, if this resulted in a lower activation energy
for intramolecular cyclization compared with other pos-
sible intermolecular pathways (thus satisfying the Cur-
tin-Hammett hypothesis), intramolecular cyclization would
take place.
That this novel idea was plausible was first tested by
computing the docking of benzene and nitrobenzene in
several different approaches.9 The two lowest energy
orientations (nitrobenzene approaching benzene, and
benzene approaching nitrobenzene) from MM were car-
ried into MOPAC and minimized. The distance of nearest
approach between a hydrogen of one molecule and a
carbon of another was 2.8 Å. Nitrobenzene approaching
benzene (the plane of the C6H5NO2 is above, and orthogo-
nal to that of benzene) had a computed MOPAC energy
of 153.7 kJ /mol. The reverse approach (plane of C6H6
above, and orthogonal to the plane of C6H5NO2) had an
energy of 157.9 kJ /mol. This suggested that edge-to-face
aromatic interaction may account for the proposed mo-
lecular recognition.
The synthesis of macrocyclic compounds has attracted
extensive attention from synthetic chemists owing to the
existence of a number of macrocyclic natural products
that exhibit useful biological activities, e.g., vancomycin,1a
RA-VII,1b arnabinol,1c combretastatin D-2,1d myricanone,1e
sinuloriolide,1f to name but a few. Many effective methods
have been developed during the past 20 years, especially
for macrolide synthesis. Some involve high-dilution tech-
niques to enhance intramolecular interaction. There was
also some consideration of the effect of the conformation
of the precursor in the process leading to macrocycle
formation. For example, in the synthesis of macrolides,2-4
intermediates were formed that brought the reacting
sites close together, thus favoring lactonization. More
recently, Marshall and co-workers5 have synthesized a
series of 12-16-membered propargylic alcohols in good
yields through Lewis acid promoted electrophilic ring
closure. Post facto molecular modeling calculations (MMX)
on the corresponding formyl-O-protonated precursor6a
showed that the global MMXE minimum is the bent
structure in which the reacting alkene site is right under
the CdOH+ group. Roussi, Beugelmans, and co-workers7
formed a 15-membered ring biaryl ether intramolecularly
(1) (a) Evans, D. A.; Dinsmore, C. J . Tetrahedron Lett. 1993, 43,
6029. (b) Itokawa, H.; Suzuki, J .; Hitoyanagi, Y.; Kondo, K.; Takaya,
K. Chem. Lett. 1993, 695. (c) Yaa, X.-S.; Ehizukz, Y.; Noguchi, H.;
Kiuchi, F.; Shibya, M.; Itaka, Y.; Seto, H.; Sankawa, U. Chem. Pharm.
Bull. 1991, 39, 2956. (d) Boger, D. L.; Sakuya, S. M.; Yohannes, D. J .
Org. Chem. 1991, 56, 4204. (e) Whiting, D. A.; Wood, A. F. Tetrahedron
Lett. 1978, 26, 2335. (f) Tius, M. A. Chem. Rev. 1988, 88, 719.
(2) Mukaiyama, T.; Usui, M.; Saigo, K. Chem. Lett. 1976, 49.
(3) Corey, E. J .; Nicolaou, K. C. J . Am. Chem. Soc. 1974, 96, 5614.
(4) Kurihara, T.; Nakajima, Y.; Mitsunobu, O. Tetrahedron Lett.
1976, 28, 2455.
(5) (a) Marshall, J . A.; DeHoff, B. S.; Crooks, S. L. Tetrahedron Lett.
1987, 28, 527. (b) Marshall, J . A.; Anderson, M. W. J . Org. Chem. 1993,
58, 3912.
(6) (a) Abramovitch, R. A. Unpublished results, 1993. (b) Abramo-
vitch, R. A.; Shi, Q. Heterocycles 1994, 38, 2347.
(7) Roussi, G.; Zamora, E. G.; Carbonnelle, A.-C.; Beugelmans, R.
Tetrahedron Lett. 1997, 38, 4405.
(8) Abramovitch, R. A.; Chinnasamy, P.; Evertz, K.; Huttner, G. J .
Chem. Soc., Chem. Commun. 1989, 3.
(9) Abramovitch, R. A.; Boche, R. Unpublished results, 1993.
10.1021/jo991139z CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/29/1999

344 J . Org. Chem., Vol. 65, No. 2, 2000
Abramovitch et al.
There has been a great deal of interest lately in edge-
to-face aromatic interactions, particularly as regards
protein folding patterns10 and host-guest complexes.11
These weak, noncovalent interactions are thought to be
of fundamental importance in determining the three-
dimensional structures and functional properties of mo-
lecular systems in biology, chemistry, and material
sciences.10e It has been stated10e that direct measurement
of energies of weak noncovalent interactions at room
temperature is not possible because these interactions
are not strong enough to cause intermolecular complex-
ation in a well-defined orientation. The present paper
addresses this topic as well.
by a polymethylene chain, to examine further the concept
of intramolecular recognition in such systems.
Resu lts a n d Discu ssion
We started by using MMX force field calculations13 to
try and predict the global minimum energy conformation
of nitro compounds 1 and 2 in the gas phase. Changing
the dielectric constant in the computation from 1.5 to 30
had little effect if any on the calculated conformations:
the bent conformations (1B and 2B) were the global
energy minima, preferred by ca. 29 kJ /mol over the
“linear” ones (Figures 1 and 2). If the nitro group is
replaced by NH2 then the “linear” conformations are
calculated to be global energy minima, as expected on
the basis of the above hypothesis. In 1B the N-atom is
within 3.72-4.72 Å of the ortho- and para-positions of
ring B at the other end. Similarly, in 2B the N-atom is
within 4.77-5.49 Å of the ortho- and para-positions of
ring B, further away but still within possible bonding
distance of each other.
MMX calculations on 1-(4-nitrophenyl)-9-phenylnonane
as a model for 1-(4-phenylnitrenium)-9-phenylnonane
(the nitro group was used as an erzatz -NH+ since no
MM parameters are available for the latter) predicted
that the ground state (global energy minimum) confor-
mation was a bent structure in which the nitro group
was close to the para position of the 9-phenyl group (ring
B), and this was confirmed experimentally by 2D NOE-
SY, UV-vis and fluorescence spectroscopy.6b On the other
hand, and in agreement with the above hypothesis, both
1,9-diphenylnonane and 1-(4-aminophenyl)-9-phenyl-
nonane were predicted to have linear global energy
minimum conformations. Thus, replacing the electron-
withdrawing nitrophenyl group by the electron-rich ami-
nophenyl eliminates the intramolecular recognition. Sub-
sequently, the corresponding nitrenium ion was found
to cyclize to an 18-membered ring in 44% isolated yield.6b
(The protonated azide and the nitrenium are both highly
electron-withdrawing.) The end groups in the corre-
sponding butane were predicted (MMX, 2D NOESY) to
be further apart and, in agreement with this prediction,
macrocyclization took place in much lower yield.6b
More recently, the cyclization of 1-(3-benzyloxy)-2-(4-
azidophenyl)ethane (a simpler version of the precursor
to our original macrocycle8) was studied.12 MMX calcula-
tions predicted, and 2D NOESY confirmed, that, unlike
the above-mentioned precursor,8 this nitrophenyl one had
a bent conformation that, if carried over to the nitrenium
ion, would lead to six-membered ring formation rather
than to a macrocycle. Experimentally, this was indeed
found to be the case.12 These results, together with the
earlier ones, suggested that simple MMX calculations
could provide a rapid empirical indicator of which precur-
sors would have readily accessible conformations that
could then be predictors of whether macrocycles could
be formed from these or not.
The synthesis of compounds 1 and 2 is shown in
Scheme 1. Phenylthioalkanyl bromides were synthesized
Sch em e 1a
a
Reagents and conditions: bBr(CH2)nBr, 10% NaOH, TBAB, 46
c
°C, overnight, 47-57%. 1.5 equiv of ethyl p-nitrophenylacetate,
NaH, DMF, 2 days, 76-88%. d10% NaOH, EtOH, 2 h, then HCl,
e
100%. CuOCrO3, quinoline, 2 h, 72-73%.
in 47-57% yield by the slow addition of thiophenol in
benzene solution to a solution of the dibromoalkanes in
benzene in the presence of tetrabutylammonium bromide
and aqueous sodium hydroxide solution. Temperature
seemed to be the key factor for ensuring adequate
reaction rates and minimal double substitutions. The
bromides were then treated with the sodium enolate of
ethyl 4-nitrophenylacetate in DMF, giving the corre-
sponding alkylated products in very good yields. These
were hydrolyzed to yield the corresponding acids in
excellent yields. Decarboxylation of the acids in quino-
line-copper chromite gave the model compounds in good
yields.
We now report studies on the formation of the corre-
sponding thiamacrocycles, and conformational studies on
the nitrenium ion precursors, model compounds 1 and
2, which have a phenylthio as the electron-rich end and
a p-nitrophenyl as the electron-poor end group, separated
(10) (a) Burley, S. K.; Petsko, G. A. Science 1985, 229, 23-28. (b)
Hunter, C. A. Chem. Soc. Rev. 1994, 23, 101-109. (c) Kim, E.-I.;
Paliwal, S.; Wilcox, C. S. J . Am. Chem. Soc. 1998, 120, 11192-11193.
(d) Burley, S. K.; Petsko, G. A. Adv. Protein Chem. 1988, 39, 125. (e)
Adams, H.; Carver, F. J .; Hunter, C. A.; Morales, J . C.; Seward, E. M.
Angew. Chem., Int. Ed. Engl. 1996, 35, 1542-1544. (f) Wedermayer,
G. H.; Patten, P. A.; Wang, L. H.; Schultz, P. A.; Stevens, R. C. Science
1997, 276, 1665.
(11) (a) Rotello, V. M.; Viani, E. A.; Deslongchamps, G.; Murray, B.
A.; Rebek, J . J . Am. Chem. Soc. 1993, 115, 797-798. (b) Schladetzky,
K. D.; Haque, T. S.; Gellman, S. H. Science 1997, 276, 1665.
(12) Abramovitch, R. A.; Ye, Xiaocong. Paper presented at the 211th
National Meeting of the American Chemical Society, ORGN 446, New
Orleans, LA, March 1996; J . Org. Chem. 1999, 64, 5904.
(13) Using PC Model from Serena Software. To ensure that a global
minimum energy had been reached, the various local minimal struc-
tures were altered repeatedly until the lowest energy conformation
was reached consistently. Though these computations do not take into
account the influence of solvent, they do allow changes in the dielectric
constant. The latter has been changed from 1.5 to 30.0 D without much
change resulting in the global energy minimum conformations.

Thiamacrocycles
J . Org. Chem., Vol. 65, No. 2, 2000 345
F igu r e 1. Global minimum energy conformation of compound 1 calculated by MMX force field, compared with a stretched out
minimum energy conformation. Relative energies: 1A, 27.80 kJ /mol; 1B, 0.00 kJ /mol.
F igu r e 2. Global minimum energy conformation of compound 2 calculated by MMX force field, compared with a stretched out
minimum energy conformation. Relative energies: 2A, 30.73 kJ /mol; 2B, 0.00 kJ /mol.
The ground state conformations predicted by the MMX
calculations were confirmed using 2D NOESY NMR. In
properly executed NOESY experiments, cross-peaks be-
tween two hydrogen atoms can only be observed if these
are separated by a distance shorter than 5 Å.14 If the
model compounds are indeed in a bent conformation in
their ground state in solution, NOE signals between the
remote proton sets associated with the aromatic rings will
be observed.14 This was indeed found to be the case.
In the cases of 1 and 2, cross-peaks for the two
aromatic ring protons were not easily observed in CDCl3
solution (cf. refs 6b and 12) because the signals for the
protons meta to the nitro group in ring A overlap those
(14) Wuthrich, K. J . Biol. Chem. 1990, 265, 22059. Wagner, G.;
Braun, W.; Havel, T. F.; Schaumann, T.; Go, N.; Wuthrich, K. J . Mol.
Biol. 1987, 196, 611.

346 J . Org. Chem., Vol. 65, No. 2, 2000
Abramovitch et al.
F igu r e 5. Crystal packing in 1.
F igu r e 3. Thermal ellipsoid plot (50% probability) of 1 and
pair of “head-to-tail” stacked molecules.
F igu r e 6. Crystal packing in 2.
Unlike nitro compounds studied earlier,6b,12 1 and 2
were crystalline so that their crystal structures could be
determined. These are shown in Figures 3 and 4.16
Remarkably, it was found that both 1 and 2 exist in the
staggered (linear) conformation in the solid state, in
marked contrast to their conformations in solution in
CDCl3. The molecules are stacked head to tail, with the
nitrophenyl above, and make a dihedral angle of 21.9°
for 1 and 19.6° for 2 with the phenyl group of the adjacent
molecule. The plane of the nitrophenyl group is tilted
above the plane of the benzene ring (a compromise
between the desired orthogonal approach and steric
repulsion). Interestingly, despite very similar molecular
conformations and common formations of “head-to-tail”
stacked dimers, the crystal packing of the two compounds
(shown in Figures 5 and 6) is quite different. In 1, the
molecules assume a “herringbone” pattern in which the
dimers stack end-to-end in a zigzag fashion, while in 2
the end-to-end stacking of the dimers is linear. Both of
these packing motifs involve weak C-H‚‚‚O interactions
between oxygen atoms of the nitro group and hydrogen
F igu r e 4. Thermal ellipsoid plot (50% probability) of 2 and
pair of “head-to-tail” stacked molecules.
in ring B. On the other hand, cross-peaks for the R
methylene protons (δ 2.88-2.93 ppm) on the carbon
directly attached to the S atom and those directly
attached to the nitrophenyl benzylic carbon (δ 2.67-2.72
ppm) are clearly seen in both cases, providing strong
evidence for the bent conformations in solution, as
predicted by MMX for the gas phase. The cross-peaks
were unchanged when the solution was diluted, indicat-
ing that they are due to intra- rather than intermolecular
interaction (see also ref 6b in which the NOESY data
were also supported by UV-vis and fluorescence spec-
troscopy; see as well the edge to face interactions in the
X-ray structures of the linear molecules in the solid
phase: Figures 3 and 4). An acetonitrile solution of 2
showed the same NOE cross-peaks, but these were
weaker than in CDCl3. It seems that while hairpin
looping of the hydrocarbon chain should be favored by
the more polar solvent owing to hydrophilic-lipophilic
interactions,15 this was partly offset by the stronger
solvation of one (probably PhS) or both of the terminal
aryl groups. As expected, there should be no attraction
(electrostatic or edge-to-face aromatic interaction) be-
tween an aminophenyl and a phenylthio group, and this
was confirmed by both MMX calculations (stretched out
global energy minimum conformation) and the absence
of a n y NOE cross-peaks for the corresponding primary
amines (6) run under the same conditions as for the nitro
derivatives. This eliminates the possibility that the cross-
peaks may be attributed to baseline incorrectness as
suggested by a referee.
(16) X-r a y cr ysta llogr a p h ic a n a lysis. The crystal structures of
1 and 2 were determined from a crystal of dimension 0.10 × 0.30 ×
0.35 mm for 1 and 0.12 × 0.29 × 0.46 mm for 2. Intensity data for
both compounds were measured at 22 ( 1 °C by using ω/2θ scans (2θmax
) 50 °C) with graphite monochromated Mo Ka radiation (λ ) 0.710 73
Å); measurements for 1 were made on a rotating anode-based (18 kW)
Rigaku AFC7R diffractometer and those for 2 were made on a Nicolet
R3mV diffractomer. Periodic measurement of three intensity standards
indicated no need for a decay correction for either compound. Absorp-
tion corrections based on azimuthal scans (ψ-scans) of several mod-
erately intense reflections resulted in normalized transmission factors
of 0.94-1.00 for 1 and 0.93-1.00 for 2; Lorentz and polarization
corrections were also applied to the data for both compounds. The
structures were solved by using direct methods and refined by using
full-matrix least squares techniques. All non-hydrogen atoms were
refined anisotropically. Hydrogen atoms were placed in optimized
positions (dC-H ) 0.96 Å) and were allowed to ride on the atom to which
they were bonded; isotropic group thermal parameters were refined
for all of the hydrogen atoms in each compound (Uiso(H): 0.076(2) Å2
for 1, 0.068 (2) Å2 for 2). Crystal data for 1: C19H23NO2S, fw 329.44 u,
monoclinic space group, C2/c (No. 15), a ) 26.721(5), b ) 6.677(1), c )
20.352(4) Å, â ) 105.38(3)°, V ) 3501(2) Å3, Z ) 8, Dcalc ) 1.25 g cm-3
.
Final residual values of R ) 0.038 and Rw ) 0.046 for 1725 observed
(I > 2σ (I)) data. Crystal data for 2: C21H27NO2S, fw 357.50 u, triclinic
space group, P1h (No. 2), a ) 10.041(3), b ) 15.092(4), c ) 6.827(2) Å,
R ) 92.61(2)°, â ) 99.09 (2)°, γ ) 105.83 (2)°, V ) 978.6(4) Å3, Z ) 2,
Dcalc ) 1.21 g cm-3. Final residual values of R ) 0.041 and Rw ) 0.045
for 1774 observed (I > 2σ(I)) data. Crystallographic data (excluding
structure factors) for the structures reported in this paper have been
deposited with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC-113072 for compound 1 and
CCDC-113073 for compound 2. Copies of the data can be obtained free
of charge on application to CCDC, 12 Union Road, Cambridge CB2
1EZ, U.K. (fax: (+44) 1223-336-033; e-mail deposit@ccdc.cam.ac.uk).
(15) J iang, X. K.; Hui, Y. Z.; Fei, Z. X. J . Chem. Soc., Chem.
Commun. 1987, 689.

Thiamacrocycles
J . Org. Chem., Vol. 65, No. 2, 2000 347
atoms of the thiophenyl group (C‚‚‚O distances of 3.499-
(4)-3.520(4) Å in 1 and 3.397(4)-3.509(4) Å in 2).
Apparently, minor differences in these subtle forces and,
quite possibly, in crystallization conditions lead to the
very different crystal packing patterns observed. It is
expected that these two compounds could exhibit poly-
morphism and that each could form stable crystals of the
alternate form.
Thus, while in solution in CDCl3 (and undoubtedly in
the gas phase as well) weak attractive intramolecular
recognition of the electron-withdrawing and electron-
donating end groups leads to the bent structure for these
compounds (and hence the expectation that the corre-
sponding nitrenium ions will cyclize to give macrocycles),
crystal lattice forces overcome the intramolecular attrac-
tions and lead to intermolecular attraction favoring the
linear antiparallel stacking.
Compounds 1 and 2 were then reduced to the corre-
sponding amine using Fe/AcOH, conditions that did not
affect the C-S bond. The azides were then obtained as
usual (Scheme 2).
F igu r e 7. MMX force field calculated global minimum energy
conformation of 8.
Sch em e 2a
patterns as does amine 6a . The two components with
molecular weights of 297 and 393, respectively, were
cyclized products, one of which was N-trifluoroacetylated
(mass 393) and the other one was not (mass 297).
The mixture was resolved by preparative TLC yielding
two fractions. The major one (32% yield) had a parent
peak at m/z 297. Its IR spectrum indicated the presence
of a secondary aromatic amine (single peak at 3319 cm-1).
There was also a strong peak at 843 cm-1, which
indicated a 1,4-disubstituted aromatic ring,17 while the
two strong peaks at 731 and 689 cm-1 in the starting
azide, indicative of a monosubstituted benzene ring, were
no longer present, showing that intramolecular cycliza-
tion had occurred. The parent ion peak at m/z 297
corresponds to C19H23NS, as expected for the intramo-
lecular cyclization product 8.
a
Reagents and conditions: bFe, AcOH, EtOH, N2, reflux, 6 h,
c
98%. HCl, NaNO2, then NaN3, 65-71%.
Acid-catalyzed decomposition of 1-(4-azidophenyl)-7-
phenylthioheptane (7a ) (Scheme 3) was carried out in
carbon tetrachloride solution in the presence of trifluo-
roacetic acid (TFA) and trifluoromethanesulfonic acid
(TFSA) at 0 °C, and the solution was then allowed to
come to room temperature and stirred for 24 h. Neutral-
ization and workup gave a mixture whose GC-MS indi-
cated that it consisted of mainly three components: the
major one had a parent ion peak at m/z 297. Of the other
two, one had a parent ion peak at m/z 393 (minor com-
ponent), and the other at m/z 299. The one with a molec-
ular weight of 299 is obviously the hydrogen-abstraction
product because it gives exactly the same fragmentation
The structure of the major cyclized product was
determined by NMR spectroscopy. The most distinguish-
ing feature of this major product was that the two phenyl
rings were now both 1,4-disubstituted, as indicated by
the presence of three doublets in the aromatic regions:
two of these doublets [δ 7.29 (d, J ) 8.5 Hz); 6.93 (d, J )
8.3 Hz)] integrated for two protons each, and the other
[δ 7.09 (dd, J ) 8.3, 8.5 Hz)] represented four protons.
The last doublet was actually a doublet of doublets, as
seen from an expanded spectrum. An NH group was
1
indicated by the H NMR spectrum [δ 3.80 (bs, 1H), D2O
exchange]. Those features strongly suggest that the new
bond was formed between the nitrogen atom and the
para-position of the thiophenyl ring to give 8. In addition,
all the aliphatic protons absorbed at higher fields than
in the uncyclized compound. Eight protons were found
to have the chemical shifts below 1.0 ppm and, in
particular, two protons resonated at 0.17 ppm. This
phenomenon suggested that they were part of a strained
ring that caused the shift to higher field. The unusually
high chemical shifts for some protons could be explained
if these were located above the phenyl ring, thus being
strongly shielded by the aromatic ring current. This
seems to be confirmed by the global minimum energy
Sch em e 3
(17) Bellamy, L. J . The Infrared Spectra of Complex Molecules; J ohn
Wiley & Sons. New York, 1958; p 64.

348 J . Org. Chem., Vol. 65, No. 2, 2000
Abramovitch et al.
conformation of macrocycle 8 as calculated using MMX
(Figure 7). It can be seen that some of the protons are
indeed in the shielding cone of the aromatic rings. The
benzylic protons resonate at higher fields than those in
azide 7a , probably because the aryl groups are twisted
in a different orientation in the macrocycle.
The trifluoroacetylated macrocycle 9 could not be
obtained pure owing its formation in very low yield, while
the hydrogen-abstraction product was identical (GC-MS,
1H NMR) with amine 6a .
Acid-catalyzed decomposition of 1-(4-azidophenyl)-9-
phenylthiononane (7b) (Scheme 4) under the same condi-
Sch em e 4
F igu r e 8. MMX force field calculated global minimum energy
conformation of 11.
presence of a small amount of contaminant (possibly
inorganic since no peaks could be attributed to it in the
spectral data).
Assignment of structure 11 was also supported by its
1H NMR spectrum. Three aromatic proton peaks were
present: two 2H doublets [δ 7.27 (d, J ) 8.5 Hz), 6.93
(d, J ) 8.3 Hz)] and a 4H doublet of doublets [δ 7.03-
7.10 (dd, J ) 8.3, 8.5 Hz)] indicated that both aryl rings
were 1,4-disubstituted. Again, all the aliphatic protons
resonated at higher fields than those of the open chain
compounds. For example, the two R-methylene protons
of 11 resonated at 2.55-2.60 ppm, compared with 2.67-
2.93 ppm for amine 7b; also there are two protons present
at 0.66 ppm for 11, compared with 1.27 ppm for amine
7b. The aliphatic protons are not as shielded as those in
8, probably owing to the longer chain length allowing
greater flexibility. MMX computation of 11 (Figure 8)
confirms that these protons are further removed from the
face of the aromatic rings than the corresponding ones
in 8.
The third fraction (Rf ) 0.40) was a mixture whose
major component had a molecular weight of 421, and it
was first assumed that this was the N-trifluoroacetylated
macrocycle. This was not the case, however, since the
mass spectral base peak had m/z 110, corresponding to
C6H5SH. One possible structure is 12, which could be
formed as shown in Scheme 6.
tions as used for 7a gave mainly four components, as
indicated by GC-MS. One component had a parent ion
peak at m/z 325, and another had one at m/z 421. The
first is the desired macrocycle, and the second is a
trifluoroacetylated derivative thereof. The third was the
hydrogen-abstraction product (m/z 327, same fragmenta-
tion pattern as 6b), and the fourth had a parent ion peak
at m/z 312 with a base peak at m/z 110.
The mixture was resolved by preparative TLC into four
fractions. The first (Rf ) 0.85) (ca. 9.6% yield) had a
parent ion peak at m/z 312. The base peak at m/z 110
indicated that the unmodified phenylthio group was still
present, suggesting that deamination had occurred to
give 1-phenyl-9-phenylthiononane (10). The second frac-
tion (Rf ) 0.52) gave a parent ion (also the base peak) at
m/z 325, which corresponds to the desired macrocycle
(11), C21H27NS (16.5% yield). The second most abundant
peak was at m/z 213, which was also observed with
macrocycle 8 (Scheme 5). Microanalysis indicated the
The last fraction (Rf ) 0.32) was the hydrogen-
abstraction product (6b) (GC, GC-MS, NMR) formed in
19% yield.
In conclusion, simple MMX calculations have, once
again, been shown to predict the global minimum energy
conformations of compounds 1 and 2 in solution, and the
corresponding nitrenium ions have undergone intramo-
lecular cyclization to give macrocycles 8 and 11, albeit
in modest yields. The bent conformations of 1 and 2 are
probably owing to edge to face attractive aromatic
interactions between the electron-poor and the electron-
rich terminal aryl groups, which is overcome by crystal
packing forces in the solid state.
Sch em e 5
Exp er im en ta l Section
Gen er a l In for m a tion . Reactions were conducted either
under a dry N2 atmosphere or in air, as specified. THF was
distilled from sodium benzophenone ketyl under a nitrogen

Thiamacrocycles
J . Org. Chem., Vol. 65, No. 2, 2000 349
Sch em e 6
atmosphere. Super-dry ethanol was prepared according to the
procedure described in Vogel’s “Textbook of Practical Organic
Chemistry.”17 Benzene, toluene, and diethyl ether were dried
over sodium wire. For chromatography, distilled petroleum
ether (bp 40-60 °C), hexane, and EtOAc were used. All other
solvents and reagents were used directly without further
purification, unless otherwise noted. Organic solvents were
recovered using a rotary evaporator (water aspirator).
Column chromatography was carried out with Aldrich silica
gel (70-230 mesh) or Fisher Scientific basic alumina (80-200
mesh). Thin-layer chromatography was carried out using
commercially available aluminum plates coated with a 0.25
mm layer of silica gel containing a 254 nm fluorescent
indicator. Preparative TLC was carried out on glass plates (20
× 20 cm) coated with a 1 mm layer of Aldrich silica gel, Merck,
TLC grade 7749, with gypsum binder and fluorescent indica-
tor. Flash chromatography was carried out using Universal
Adsorbents Inc. silica gel 32-63 (40 µm).
MMX force field calculations were carried out by using
PCMODEL (Pi v.4.1) from Serena Software, Bloomington, IN.
The NMR spectra were run on a Bruker AC300 instrument.
2D NOESY spectra were run in either CDCl3 or CH3CN
solution, as specified in the Discussion or in the Experimental
Section, with an internal TMS standard. The spectra were
acquired in the phase-sensitive TPPI mode: 2K × 1024 FID’s
of 16 scans over sweep widths of 2262 Hz for compound 1, 1992
Hz for compound 6a , 2762 Hz for compound 2, and 2008 Hz
for compound 6b, with a recycle delay of 2.0 s, and collected
at room temperature. The mixing time (τm) was 1.0 s. Data
were processed in the F2 dimension with a Gauss function and
F1 dimension with a 45° shifted sine function, and zero filled
to 2k × 2k prior to Fourier transformation and phase correc-
tion. The spectra were symmetrized prior to plotting.
to 0 °C under a nitrogen atmosphere, ethyl 4-nitrophenylace-
tate (4.06 g, 19.4 mmol) in DMF (40 mL) was added dropwise
over a period of 1 h, and stirring continued for 2 h. Bromide
3a (1.76 g, 6.45 mmol) in DMF (40 mL) was then added
dropwise over a period of 1 h, and the resulting suspension
was stirred at 0 °C for 1 h, gradually brought to room
temperature (RT) over a period of 2 h, and kept there for 45
h. The reaction mixture was quenched with water, and the
mixture was evaporated to dryness. The residue was then
extracted with ethyl acetate, and the organic phase was
washed with water and dried (Na2SO4). The filtered solution
was decolorized with activated carbon for 1 h at room tem-
perature. Filtration and evaporation of the solvent afforded a
red-brown oily liquid (4.85 g). Unreacted starting material was
removed by bulb-bulb distillation at e160 °C/1.5 mmHg. The
residue was purified by chromatography on silica gel using
20% EtOAc in hexane as eluant, to give 1 as a pale yellow
solid (0.23 g, 11%), mp 61-63 °C (see synthesis of authentic
sample below) and the desired 4a an oil (1.99 g, 77%). IR
(KBr): 1740 (CO2Et) (s), 1528 (s), 1353 (NO2) cm-1 (s). 1H NMR
(CDCl3): δ 7.26-7.24 (m, 4H), 7.14-7.19 (m, 1H), 3.37-3.42
(t, 2H), 2.89-2.94 (t, 2H), 1.82-1.87 (m, 2H), 1.64-1.68 (m,
2H), 1.45 (bs, 4H). GC-MS (70 eV): m/z 401 (M•+). Anal. Calcd
for C22H27NO4S: C, 65.81; H, 6.78. Found: C, 65.71; H, 6.82.
2-(4-Nitr op h en yl)-8-p h en ylth ioocta n oic Acid (5a ). Es-
ter 4a (1.96 g, 4.89 mol) was dissolved in 10% NaOH solution
(8 mL) and ethanol (1 mL). The mixture was boiled under
reflux for 2 h. The ethanol was evaporated, and 10% HCl
solution (15 mL) was added. The precipitated acid was filtered,
washed with water, and dried at 70 °C to give 5a (1.8 g, 100%)
as a yellow-red solid, mp 97-98 °C (from hexane/CH2Cl2).
Anal. Calcd for C20H23NO4S: C, 64.32; H, 6.21. Found: C,
64.47; H, 6.29.
6-P h en ylth ioh exyl Br om id e (3a ). In a 250 mL three-
necked round-bottom flask connected to a condenser, a ther-
mometer, and a dropping funnel was placed sodium hydroxide
(10 g) in water (90 mL). To this solution was added 1,6-
dibromohexane (12 g, 49 mmol) in benzene (50 mL), and
tetrabutylammonium bromide (16.1 g, 48 mmol). The solution
was heated to 48 °C, and thiophenol (5.7 g, 52 mmol) in
benzene (100 mL) was then added dropwise over a period of
12 h. The mixture was stirred at that temperature for 6 h.
The reaction mixture was cooled and the organic layer
separated. The aqueous layer was extracted with ethyl acetate.
The organic layers were combined and washed with water and
dried (Na2SO4). Filtration and evaporation of the solvent gave
a semisolid crude product (14.1 g), which was purified by
1-(4-Nitr oph en yl)-7-ph en ylth ioh eptan e (1). Acid 5a (1.77
g, 4.75 mol) was dissolved in quinoline (7 mL), and copper
chromite (91.5 mg) was added. The mixture was heated under
reflux for 2.5 h in an oil bath kept at 210-220 °C. It was then
cooled, treated with 10% HCl (23 mL), and stirred for 2 h. The
aqueous solution was extracted with diethyl ether (3 × 50 mL),
and the combined organic layers were washed with 10% HCl
and water and then dried (Na2SO4). Silica gel (40-140 mesh)
column chromatography of the crude product [petroleum ether
(bp 40-60 °C)/EtOAc, 95:5 v/v] gave 1 (1.14 g, 73%), mp 61-
63 °C (from hexane/CH2Cl2), identical with the byproduct
obtained from 4a above. IR: (KBr) 1505 (s), 1338 (NO2) cm-1
(s). 1H NMR (CDCl3): δ 8.12-8.14 (d, 2H, J ) 8.47 Hz), 7.24-
7.31 (m, 6H), 7.16-7.18 (m, 1H), 2.88-2.93 (t, 2H), 2.67-2.72
(t, 2H), 1.62-1.64 (m, 4H), 1.18-1.42 (m, 6H). 13C NMR
(CDCl3): δ 150.62, 146.24, 136.88, 129.10, 128.85, 128.79,
125.67, 123.55, 35.78, 33.51, 30.82, 28.97, 28.87, 28.56. GC-
MS (70 eV): m/z 329 (M•+). Anal. Calcd for C19H23NO2S: C,
69.27; H, 7.04. Found: C, 69.31; H, 7.03.
chromatography on
a column of silica gel. Elution with
petroleum ether (bp 40-60 °C) afforded pure 3a as a colorless
liquid (56%). GC-MS (70 eV): m/z 272 (274) (M•+). Anal. Calcd
for C12H17BrS: C, 52.75; H, 6.27. Found: C, 52.87; H, 6.27.
Eth yl [2-(4-Nitr op h en yl)-8-p h en ylth io]octa n oa te (4a ).
Sodium hydride (0.78 g, 19.5 mmol, 60% suspension in mineral
oil) was suspended in DMF (40 mL), the suspension was cooled
1-(4-Am in op h en yl)-7-p h en ylth ioh ep ta n e (6a ). To com-
pound 1 (0.78 g, 2.37 mmol) in methanol (80 mL) was added

350 J . Org. Chem., Vol. 65, No. 2, 2000
Abramovitch et al.
hydrazine hydrate (2.4 mL) and Raney Ni (1.5 mL). The
suspension was heated under reflux for 3 h. The methanol was
evaporated, the residue was extracted with diethyl ether, and
the solution was washed with water and dried (Na2SO4).
Filtration and evaporation of the solvent afforded crude
product (0.70 g). GC-MS indicated it was the desired product
containing a small amount of 4-heptylaniline (produced by the
cleavage of the carbon sulfur bond). The pure product 6a (mp
49-51 °C) (85%) (from CH2Cl2) was obtained after silica gel
column chromatography using EtOAc/petroleum ether (bp 40-
60 °C) (1:4 v/v). IR (KBr): 3383 (NH2) (m), 3300 (NH2) cm-1
(w). 1H NMR (CDCl3): δ 7.23-7.32 (m, 4H), 7.11-7.16 (m, 1H),
6.93-6.96 (d, 2H, J ) 8.29 Hz), 6.58-6.61 (d, 2H, J ) 8.35),
2.86-2.91 (t, 2H), 2.44-2.49 (t, 2H), 1.51-1.67 (m, 4H),
1.35,1.42 (m, 2H), 1.26-1.32 (m, 4H). 13C NMR: δ 150.62,
146.24, 136.88, 129.10, 128.85, 128.79, 125.67, 123.55, 35.78,
33.51, 30.82, 28.97, 28.87, 28.56. GC-MS (70 eV): m/z 299
(M•+). Anal. Calcd for C19H25NS: C, 76.20; H, 8.41. Found: C,
76.15; H, 8.43.
1-(4-Azid op h en yl)-7-p h en ylth ioh ep ta n e (7a ). Amine 6a
(100 mg, 0.33 mmol) was dissolved in hot dilute HCl (0.68 mL
of concentrated HCl in 3.4 mL of H2O). The mixture was placed
in an ice-bath and cooled to <5 °C and treated with NaNO2
(45 mg) in H2O (5.6 mL). After stirring for 1.5 h, NaN3 (45
mg) in H2O (5.6 mL) was added and the solution was stirred
for 1.5 h. It was extracted with ethyl acetate, and the combined
organic layers were washed with water and dried (MgSO4).
Filtration and solvent evaporation, followed by silica gel
column chromatography (hexane) gave 7a (77 mg, 71%) as a
white solid, mp 44.5-45.5 °C. IR (KBr): 2095 (N3) cm-1 (s).
GC-MS (70 eV): m/z 299, 190, 146, 123, 106, 77 (no parent
ion observed). Anal. Calcd for C19H23N3S: C, 70.10; H, 7.14.
Found: C, 69.95; H, 7.05.
petroleum ether (bp 40-60 °C) afforded pure 3b as a colorless
liquid, bp 140-160 °C/1 mmHg (47%). GC-MS (70 eV): m/z
300(302) (M•+). Anal. Calcd for C14H21BrS: C, 55.81; H, 7.03.
Found: C, 56.01; H, 7.00.
Eth yl [2-(4-Nitr oph en yl)-10-ph en ylth io]decan oate (4b).
Sodium hydride (1.99 g, 49.8 mmol, 60% suspension in mineral
oil) was suspended in dry DMF (102 mL), the suspension was
cooled to 0 °C under
a nitrogen atmosphere, and ethyl
4-nitrophenylacetate (10.41 g, 49.8 mmol) in DMF (102 mL)
was added dropwise with stirring over a period of 1 h. Stirring
was continued for 2 h, and bromide 3b (5.0 g, 16.6 mmol) in
DMF (102 mL) was then added over a period of 1 h. The
resulting suspension was stirred at 0 °C and the temperature
gradually raised to room temperature over the course of 2 h
and kept there for 46 h. The reaction was quenched with water
and the solvent evaporated. The residue was extracted with
ethyl acetate, and the organic phase was washed with water
and dried (Na2SO4). The filtered solution was decolorized with
activated charcoal for 1 h at room temperature. Filtration and
evaporation of the solvent afforded a red-brown oily liquid (11.7
g). Most of the unreacted starting material was removed by
bulb-bulb distillation at e160 °C/1.5 mmHg. The residue was
purified by chromatography on silica gel using 20% EtOAc in
hexane as eluant, to give the desired ester 4b as an oil (5.39
g, 76%). IR (KBr): 1736 (CdO) cm-1 (s). GC-MS (70 eV): m/z
429 (M•+). Anal. Calcd for C24H31NO4S: C, 67.09; H, 7.29.
Found: C, 67.01; H, 7.22. Additionally, compound 2 was
isolated as a pale yellow solid (0.53 g, 9%), mp 53-54 °C. IR
1
(KBr): 1512 (NO2) (s), 1343 (NO2) cm-1 (s). H NMR (CDCl3):
δ 8.11-8.14 (d, 2H, J ) 8.7 Hz), 7.24-7.33 (m, 6H), 7.12-
7.18 (m, 1H), 2.88-2.93 (t, 2H), 2.67-2.72 (t, 2H), 1.60-1.69
(m, 4H), 1.36-1.43 (m, 2H), 1.29 (bs, 8H). 13C NMR (CDCl3):
δ 150.73, 146.21, 136.97, 129.10, 128.77, 125.60, 123.52, 35.81,
33.52, 30.90, 29.28, 29.05, 28.72. GC-MS (70 eV): m/z 357
(M•+). Anal. Calcd for C21H27NO2S: C, 70.54; H, 7.63. Found:
C, 70.52; H, 7.62.
2-(4-Nit r op h en yl)-10-p h en ylt h iod eca n oic Acid (5b ).
Ester 4b (4.7 g, 11 mol) was dissolved in aqueous 5% NaOH
solution (19.1 mL) and ethanol (2.73 mL) and boiled under
reflux for 2 h. The ethanol was then evaporated, and 10% HCl
solution (30 mL) was added to precipitate the carboxylic acid.
The product was filtered, washed with water, and dried at 70
°C to give 5b (4.4 g, 100%) as a yellow-red solid, mp 61-63
°C. IR (KBr): 1705 (CO2H) (s),1518 (NO2) (s), 1346 (NO2) cm-1
(s). Anal. Calcd for C22H27NO4S: C, 65.80; H, 6.79. Found: C,
65.90; H, 6.83.
1-(4-Nitr op h en yl)-9-p h en ylth ion on a n e (2). Acid (5b) (4.4
g) was dissolved in quinoline (17 mL), and copper chromite
(0.24 g) was added. The mixture was heated under reflux in
an oil bath kept at 210-220 °C for 3 h. The mixture was cooled,
treated with 10% HCl (50 mL), and stirred for 2 h. The solution
was extracted with diethyl ether and the combined organic
layers were washed with 10% HCl and then with water and
dried (Na2SO4). Silica gel (40-140 mesh) column chromatog-
raphy of the crude product [petroleum ether (bp 40-60 °C)/Et-
OAc, 95:5, v/v] gave 2 (2.82 g, 72%) as a pale yellow solid with
the same physical and spectroscopic data as reported above.
1-(4-Am in op h en yl)-9-p h en ylth ion on a n e (6b). To nitro
compound 2 (0.687 g, 1.92 mmol) in ethanol (10 mL) was added
glacial acetic acid (1.25 mL) and iron powder (0.54 g). The
suspension was heated under reflux for 6.5 h under a nitrogen
atmosphere (monitored by TLC). Water (40 mL) was added,
and the mixture was extracted with diethyl ether. The solution
was washed with saturated aqueous sodium bicarbonate
solution and water and then dried (Na2SO4). Filtration and
evaporation of the solvent afforded pure 7b (0.626 g, 99.5%),
mp 55-56 °C. IR (KBr): 3408 (m), 3340 cm-1 (w) (NH2). 1H
NMR (CDCl3): δ 7.24-7.33 (m, 4H), 7.13-7.18 (m, 1H), 6.95-
6.97 (d, 2H, J ) 8.22 Hz), 6.60-6.63 (d, 2H, J ) 8.25 Hz),
3.53 (bs, 2H), 2.88-2.93 (t, 2H), 2.46-2.51 (t, 2H), 1.61-1.66
(m, 2H), 1.57-1.59 (m 2H), 1.40 (bs, 2H), 1.27 (bs, 8H). GC-
MS (70 eV): m/z 327 (M•+). Anal. Calcd for C21H29NS: C, 76.99;
H, 8.94. Found: C, 76.88; H, 8.88.
Acid -Ca ta lyzed Decom p osition of 1-(4-Azid op h en yl)-
7-p h en ylth ioh ep ta n e (7a ). Azide 7a (76 mg, 0.23 mmol) was
dissolved in carbon tetrachloride (7 mL) at 0 °C under a
nitrogen atmosphere. Trifluoroacetic acid (1.5 mL) and trif-
luoromethanesulfonic acid (TFSA) (2 drops) were then added
dropwise to the solution, which was stirred at 0 °C for 0.5 h
and then at room temperature for 24 h. The solution was
basified with saturated aqueous sodium bicarbonate, and the
CCl4 layer was separated. The aqueous phase was extracted
with methylene chloride, the combined organic layers were
washed with saturated aqueous NaHCO3 and then with water
and dried (Na2SO4). The residue was resolved by preparative
TLC (hexane/EtOAc, 9:1 v/v) to give 4,4′-heptylenethiodiphen-
ylamine (8) as a white solid, mp 127-128 °C (22 mg, 32%). IR
1
(KBr): 3319 (NH) cm-1 (m). H NMR (CDCl3): δ 7.29 (d, 2H,
J ) 8.5 Hz), 7.09 (dd, 4H, J ) 8.3, 8.5 Hz), 6.93 (d, 2H, J )
8.3 Hz), 3.80 (bs, 1H, D2O exchange), 2.54 (t, 2H, J ) 7.2 Hz),
2.40 (t, 2H, J ) 7.3 Hz), 1.25-1.36 (m, 2H), 0.81-0.89 (m, 4H),
0.45-0.53 (m, 2H), 0.19-0.24 (m, 2H). 13C NMR (CDCl3): δ
150.90, 147.97, 139.54, 136.27, 129.76, 124.49, 124.33, 36.79,
34.69, 32.48, 29.68, 28.99, 28.59, 26.18. GC-MS (70 eV): m/z
297 (M•+). Anal. Calcd for C19H23NS: C, 76.70; H, 7.81.
Found: C, 76.54; H, 7.82. Two additional products were
detected by GC-MS: the hydrogen-abstraction product 6a
(12%), identical with the authentic sample, and a product
assumed to be the N-trifluoroacetylated macrocycle (9): GC-
MS (70 eV) m/z 393 (M•+), 354, 284, 226, 200, 150, 110, 77.
9-P h en ylth iooctyl Br om id e (3b). In a 250 mL three-
necked round-bottom flask connected to a condenser, a ther-
mometer, and a dropping funnel was placed sodium hydroxide
(6.9 g) in water (62 mL). To this solution was added 1,8-
dibromooctane (9.5 g, 34.9 mmol) in benzene (34 mL) and
tetrabutylammonium bromide (11.1 g, 34.4 mmol). The solu-
tion was heated to 48 °C, and thiophenol (4.1 g, 37.3 mmol) in
benzene (100 mL) was then added dropwise over a period of
21 h. The mixture was stirred at that temperature for 4.5 h.
The reaction mixture was cooled and the organic layer
separated. The aqueous layer was extracted with ethyl acetate.
The organic layers were combined, washed with water, and
dried (Na2SO4). Filtration and evaporation of the solvent gave
a semisolid crude product (11.1 g), which was purified by
1-(4-Azid op h en yl)-9-p h en ylth ion on a n e (7b). Amine 6b
(160 mg, 0.49 mmol) was dissolved in warm 20% HCl (10 mL).
chromatography on
a column of silica gel. Elution with

Thiamacrocycles
J . Org. Chem., Vol. 65, No. 2, 2000 351
The mixture was cooled to below 5 °C and treated with NaNO2
(100 mg) in H2O (5 mL). After the solution was stirred for 1.5
h, NaN3 (100 mg) in H2O (5 mL) was added dropwise and the
resulting solution stirred for 2 h. It was then extracted with
ethyl acetate, and the combined organic layers were washed
with water and dried (Na2SO4). The crude product was purified
by silica gel column chromatography (hexane) to give 7b (71
mg, 65%) as a white solid, mp 43-45 °C: IR (KBr): 2115 cm-1
(s) (N3). GC-MS (70 eV): m/z 327, 252, 207, 123, 106, 77. Anal.
Calcd for C21H27N3S: C, 71.33; H, 7.71. Found: C, 71.46; H,
7.73.
Acid -Ca ta lyzed Decom p osition of 1-(4-Azid op h en yl)-
9-p h en ylth ion on a n e (7b). Azide 7b (36 mg, 0.10 mmol) was
dissolved in carbon tetrachloride (4 mL) at 0 °C under a
nitrogen atmosphere. Trifluoroacetic acid (0.75 mL) and triflu-
oromethanesulfonic acid (TFSA) (3 drops) were then added to
the solution, which was stirred at 0 °C for 0.5 h and then at
room temperature for 24 h. It was basified with saturated
sodium bicarbonate, and the CCl4 layer was separated. The
aqueous phase was extracted with methylene chloride. The
combined organic layers were washed with saturated aqueous
NaHCO3 and then water and dried (Na2SO4). Filtration and
evaporation of the solvent afforded a mixture whose GC-MS
indicated the presence of mainly four components: one com-
ponent gave a parent ion peak at m/z 325, and one gave a
parent ion peak at m/z 421. One of the other two components
was definitely the hydrogen-abstraction product because of its
formula weight (with a parent ion peak at m/z 327) and the
fact that it exhibited the same fragmentation pattern as did
amine 6b. The last component had a parent ion peak at m/z
312, with a base peak at m/z 110. The mixture was resolved
by preparative TLC. Elution with ethyl acetate-hexane (1:9,
v/v) gave four fractions. The first fraction (Rf 0.85) was
compound 10 (9.6%): parent ion peak at m/z 312. The second
fraction afforded impure (as indicated by the C, H microanaly-
sis) macrocycle 11 (16.5%). 1H NMR (CDCl3): δ 7.27 (d, 2H, J
) 8.5 Hz), 7.08 (dd, 4H, J ) 8.3, 8.5 Hz), 6.93 (d, 2H, J ) 8.3
Hz), 2.55-2.60 (m, 4H), 1.45 (m, 2H), 1.25-1.28 (m, 4H), 1.09-
1.17 (m, 2H), 0.95-1.02 (m, 2H), 0.75-0.81 (m, 4H), 0.66 (m,
2H). 13C NMR (CDCl3): δ 134.44, 129.67, 124.86, 121.23, 35.58,
34.83, 30.72, 29.49, 29.32, 29.14, 28.59, 28.31, 26.50. GC-MS
(70 eV): m/z 325, 282, 227, 213, 181, 152, 91, 77. Anal. Calcd
for C21H27NS: C, 77.47; H, 8.38. Found: C, 73.56; H, 8.07. The
third fraction afforded a mixture that had a major component
12 in about 5.4% yield. GC-MS (70 eV): m/z 421 (M•+), 352,
312, 242, 200, 123, 110 (100%), 91, 77, 65. The last fraction
afforded the hydrogen-abstraction product 6b (19%). 1H NMR
(CDCl3): δ 7.24-7.33 (m, 4H), 7.13-7.18 (m, 1H), 6.95-6.97
(d, 2H, J ) 8.22 Hz), 6.60-6.63 (d, 2H, J ) 8.25 Hz), 3.53 (bs,
2H), 2.88-2.93 (t, 2H), 2.46-2.51 (t, 2H), 1.61-1.66 (m, 2H),
1.57-1.59 (m, 2H), 1.40 (bs, 2H), 1.27 (bs, 8H). GC-MS (70
eV): m/z 327 (M•+), 210, 164, 123, 106, 77, 55.
Ack n ow led gm en t. We are grateful for support from
the NSF-INT (Grant No. 9304094), and we thank NATO
(Grant 930536) for partial support of this work.
Su p p or tin g In for m a tion Ava ila ble: 2D NOESY spectra
of compounds 1, 2, 6a , and 6b in deuteriochloroform solution.
This material is available free of charge via the Internet at
http://pubs.acs.org.
J O991139Z