Studies on the SmI2-Promoted Pinacol-Type Cyclization: Synthesis of the Hexahydroazepine Ring of Balanol

Article abstract of DOI:10.1021/jo000538n

The efficiency of the samarium(II) iodide induced pinacol-type coupling for the construction of seven-membered cyclic amino alcohols has been investigated. With the acyclic carbonylhydrazones 6 and 16, good yields of the hexahydroazepines 22 and 23 were obtained (56-57%) with high trans-selectivity (= 10:1), which compares well with similar reactions generating the corresponding five-and six-membered carbocycles (Fallis, A. G.; Sturino, C. F. J. Am. Chem. Soc. 1994, 116, 7447). It is essential for ring formation that the strongly electron-donating ligand, hexamethylphosphoramide, be present, as in its absense intermolecular pinacol coupling forming the diols 27-30 is the dominant reaction. Hence, the role for HMPA appears not only to increase the rate of electron transfer but also to modulate rate constants for the subsequent reactions (cyclization and pinacol coupling) of the intermediate ketyl. This ring forming reaction has been applied to the construction of the fully functionalized hexahydroazepine ring of the PKC inhibitor, balanol. Initial attempts to develop an asymmetric version of this reaction indicate the use of chiral ligands based on the structure of HMPA.

Full text of DOI:10.1021/jo000538n

5382
J . Org. Chem. 2000, 65, 5382-5390
Stu d ies on th e Sm I2-P r om oted P in a col-Typ e Cycliza tion :
Syn th esis of th e Hexa h yd r oa zep in e Rin g of Ba la n ol
Ditte Riber, Rita Hazell, and Troels Skrydstrup*
Department of Chemistry, University of Aarhus, Langelandsgade 140, 8000 Aarhus C, Denmark
ts@kemi.aau.dk
Received April 11, 2000
The efficiency of the samarium(II) iodide induced pinacol-type coupling for the construction of seven-
membered cyclic amino alcohols has been investigated. With the acyclic carbonylhydrazones 6 and
16, good yields of the hexahydroazepines 22 and 23 were obtained (56-57%) with high trans-
selectivity () 10:1), which compares well with similar reactions generating the corresponding five-
and six-membered carbocycles (Fallis, A. G.; Sturino, C. F. J . Am. Chem. Soc. 1994, 116, 7447). It
is essential for ring formation that the strongly electron-donating ligand, hexamethylphosphoramide,
be present, as in its absense intermolecular pinacol coupling forming the diols 27-30 is the dominant
reaction. Hence, the role for HMPA appears not only to increase the rate of electron transfer but
also to modulate rate constants for the subsequent reactions (cyclization and pinacol coupling) of
the intermediate ketyl. This ring forming reaction has been applied to the construction of the fully
functionalized hexahydroazepine ring of the PKC inhibitor, balanol. Initial attempts to develop an
asymmetric version of this reaction indicate the use of chiral ligands based on the structure of
HMPA.
In tr od u ction
We have recently been interested in developing dia-
stereoselective and enantioselective pinacol coupling
The protein kinase C (PKC) inhibitor, balanol, has
attracted much interest owing to its high inhibitory
activity in the nanomolar range and novel structure,
composed of a disubstituted hexahydroazepine ring and
a highly functionalized benzophenone moiety (Figure
1).1,2 The family of PKC enzymes has been associated
with a variety of diseases, including cancer, cardiovas-
cular disorders, asthma, diabetes, central nervous system
dysfunction and AIDS.3 The selective inhibition of these
enzymes may therefore be of important therapeutic
value, with balanol representing a new lead compound
for achieving this goal. The synthetic community re-
sponded quickly to the challenge of preparing balanol and
analogues thereof, with six total syntheses having been
reported to date,4 as well as several syntheses of key
structural fragments.5-7
(5) For synthetic approaches to the hexahydroazepine fragment,
see: (a) Mu¨ller, A.; Takyar, D. K.; Wit, S.; Ko¨nig, W. A. Liebigs Ann.
Chem. 1993, 651. (b) Hughes, P. F.; Smith, S. H.; Olson, J . T. J . Org.
Chem. 1994, 59, 5799. (c) Hu, H.; J agdmann, G. E.; Hughes, P. F.;
Nichols, J . B. Tetrahedron Lett. 1995, 36, 3659. (d) Albertini, E.; Barco,
A.; Benetti, S.; De Risi, C.; Pollini, G. P.; Zanirato, V. Synlett, 1996,
29. (e) Tuch, A. T.; Sanie`re, M.; Le Merrer, Y.; Depezay, J .-C.
Tetrahedron: Asymmetry 1996, 7, 2901. (f) Albertini, E.; Barco, A.;
Benetti, S.; De Risi, C.; Pollini, G. P.; Zanirato, V. Tetrahedron 1997,
53, 17177. (g) Wu, M. H.; J acobsen, E. N. Tetrahedron Lett. 1997, 38,
1693. (h) Coulon, E.; Can˜o de Andrade, M. C.; Ratovelomanana-Vidal,
V.; Geneˆt, J .-P. Tetrahedron Lett. 1998, 39, 6467. (i) Herdeis, C.;
Mohareb, R. M.; Neder, R. B.; Schwabenla¨nder, F.; Telser, J . Tetra-
hedron: Asymmetry 1999, 10, 4521. (j) Cook, G. R.; Shanker, P. S.;
Peterson, S. L. Org. Lett. 1999, 1, 615. (k) Fu¨rstner, A.; Thiel, O. R. J .
Org. Chem. 2000, 65, 1738.
(6) For synthetic approaches to the benzophenone fragment, see:
(a) Hollinshead, S. P.; Nichols, J . B.; Wilson, J . W. J . Org. Chem. 1994,
59, 6703. (b) Nilsson, J . P.; Andersson, C.-M. Tetrahedron Lett. 1997,
38, 4635. (c) Denieul, M.-P.; Skrydstrup, T. Tetrahedron Lett. 1999,
40, 4901. (d) Storm, J . P.; Andersson, C.-M. Org. Lett. 1999, 1, 1451.
(7) For the synthesis and study of balanol analogues, see: (a)
Heerding, J . M.; Lampe, J . W.; Darges, J . W.; Stamper, M. L. Bioorg.
Med. Chem. Lett. 1995, 5, 1839. (b) J agdmann, G. E., J r.; Defauw, J .
M.; Lai, Y.-S.; Crane, H. M.; Hall, E. S.; Buben J . A.; Hu, H.; Gosnell,
P. A. Bioorg. Med. Chem. Lett. 1995, 5, 2015. (c) Lai, Y.-S.; Stamper,
M. Bioorg. Med. Chem. Lett. 1995, 5, 2147. (d) Lai, Y.-S.; Menaldino,
D. S.; Nichols, J . B.; J agdmann, G. E., J r.; Mylott, F.; Gillespie, J .;
Hall, S. E. Bioorg. Med. Chem. Lett. 1995, 5, 2151. (e) Lai, Y.-S.;
Mendoza, J . S.; Hubbard, F.; Kalter, K. Bioorg. Med. Chem. Lett. 1995,
5, 2155. (f) Mendoza, J . S.; J agdmann, G. E., J r.; Gosnell, P. A. Bioorg.
Med. Chem. Lett. 1995, 5, 2211. (g) Hu, H.; Hollinshead, S. P.; Hall, S.
E.; Kalter, K.; Ballas, L. M. Bioorg. Med. Chem. Lett. 1996, 6, 973. (h)
J agdmann, G. E., J r.; Defauw, J . M.; Lampe, J . W.; Darges, J . W.;
Kalter, K. Bioorg. Med. Chem. Lett. 1996, 6, 1759. (i) Crane, H. M.;
Menaldino, D. S.; J agdmann, G. E., J r.; Darges, J . W.; Buben, J . A.
Bioorg. Med. Chem. Lett. 1996, 5, 2133. (j) Nielsen, J .; Lyngsø, L. O.
Tetrahedron Lett. 1996, 37, 8439. (k) Defauw, J . M., Murphy, M. M.;
J agdmann, G. E., J r.; Hu, H.; Lampe, J . W.; Hollinshead, S. P.;
Mitchell, T. J .; Crane, H. M.; Heerding, J . M.; Mendoza, J . S.; Davis,
J . E.; Darges, J . W.; Hubbard, F. R.; Hall, S. E. J . Med. Chem. 1996,
39, 521. (l) Lai, Y.-S.; Mendoza, J . S.; J agdmann, G. E., J r.; Menaldino,
D. S.; Biggers, C.; Heerding, J . M.; Wilson, J . W.; Hall, S. E.; J iang, J .
B.; J anzen, W. P.; Ballas, L. M. J . Med. Chem. 1997, 40, 226. (m) Koide,
K.; Bunnage, M. E.; Paloma, L. G.; Kanter, J . R.; Taylor, S. S.; Brunton,
L. L.; Nicolaou, K. C. Chem. Biol. 1995, 2, 601.
(1) Kulanthaivel, P.; Hallock, Y. F.; Boros, C.; Hamilton, S. M.;
J anzen, W. P.; Ballas, L. M.; Loomis, C. R.; J iang, J . B. J . Am. Chem.
Soc. 1993, 115, 6452.
(2) Ohshima, S.; Yanagisawa, M.; Katoh, A.; Fujii, T.; Sano, T.;
Matsukuma, S.; Furumai, T.; Fujiu, M.; Watanabe, K.; Yokose, K.; M.
Arisawa, M.; Okuda, T. J . Antibiot. 1994, 47, 639.
(3) For a recent review of the application of PKC inhibitors, see:
Bradshaw, D.; Hill, C. H.; Nixon, J . S.; Wilkinson, S. E. Agents Actions
1993, 38, 135. See also: Ishii, H.; J irousek, M. R.; Koya, D.; Takagi,
C.; Xia, P.; Clermont, A.; Bursell, S.-E.; Kern, T. S.; Ballas, L. M.;
Heath, W. F.; Stramm, L. E.; Feener, E. P.; King, G. L. Science 1996,
272, 728.
(4) (a) Nicolaou, K. C.; Bunnage, M. E., Koide, K. J . Am. Chem. Soc.
1994, 116, 8402. (b) Nicolaou, K. C.; Koide, K.; Bunnage, M. E. Chem.
Eur. J . 1995, 1, 454. (c) Lampe, J . W.; Hughes, P. F.; Biggers, C. K.;
Smith, S. H.; Hu, H. J . Org. Chem. 1994, 59, 5147. (d) Lampe, J . W.;
Hughes, P. F.; Biggers, C. K.; Smith, S. H.; Hu, H. J . Org. Chem. 1996,
61, 4572. (e) Adams, C. P.; Fairway, S. M.; Hardy, C. J .; Hibbs, D. E.;
Hursthouse, M. B.; Morley, A. D.; Sharp, B. W.; Vicker, N.; Warner, I.
J . Chem. Soc., Perkin Trans. 1 1995, 2355. (f) Barbier, P.; Stadlwieser,
J . Chimia 1996, 50, 530. (g) Tanner, D.; Tadenborg, L.; Almario, A.;
Pettersson, I.; Csoeregh, I.; Kelly, N. M.; Andersson, P. G. Tetrahedron
1997, 53, 4857. (h) Miyabe, H.; Torieda, M.; Kiguchi, T.; Naito, T.
Synlett 1997, 580. (i) Miyabe, H.; Torieda, M.; Inoue, K.; Tajiri, K.;
Kiguchi, T.; Naito, T. J . Org. Chem. 1998, 63, 4397.
10.1021/jo000538n CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/29/2000

SmI2-Promoted Pinacol-Type Cyclization
J . Org. Chem., Vol. 65, No. 17, 2000 5383
to avoid arising gauche interactions in the product. For
successful cyclizations of these room-temperature reac-
tions, it was also noted that hexamethylphosphoramide
(HMPA) was necessary as a cosolvent; in its absence, no
reaction was observed under the same time scale. Whether
this reaction could be adapted for seven-membered ring
formation as required for the synthesis of the hexahydro-
azepine ring of balanol with the correct trans-selectivity
was not obvious to predict, considering that a radical
cyclization mechanism is invoked for seven-membered
ring formation with cyclization rates possibly too slow
to compete with the second electron transfer. In addition,
an 11-membered chelate intermediate would be required
for obtaining the correct relationship between the two
stereogenic centers. Nevertheless, Naito and co-workers
have recently applied this approach to the azapine ring
of balanol employing a carbonyloxime derivative (eq 2)
F igu r e 1.
reactions promoted by samarium diiodide.8,9 A retro-
synthetic analysis of balanol suggests that the hexa-
hydroazepine ring, containing a vicinal amino alcohol
with a trans-relationship, could possibly be constructed
from an achiral precursor employing an enantioselective
pinacol-type coupling between a monofunctionalized di-
aldehyde (Figure 1). Although SmI2-promoted pinacol
cyclizations of dialdehydes to five- and six-membered
rings display an overwhelming desire for cis-diol forma-
tion involving a chelated intermediate,10 the correspond-
ing carbonylhydrazones have been reported by Fallis and
Sturino to afford exclusively the trans-isomer (eq 1).11,12
The observed selectivity was explained in the case of the
cyclopentane formation by invoking a nine-membered
ring template where the large diphenylhydrazone sub-
stituent occupies a pseudoequatorial orientation and the
oxygen of the ketyl intermediate an axial position in order
and do indeed observe cyclization though with reduced
diastereoselectivity (6.6 to 1 in favor of the trans-iso-
mer).4h,i,13,14 As with the diphenylhydrazones, the addition
of HMPA was found indispensable, as in its absence no
reaction was observed even at room temperature.
To develop an enantioselective version of this reaction,
we initiated these cyclization studies with the diphenyl-
hydrazones, being stereochemically defined compounds
and not isomerical mixtures as with the corresponding
oximes,15 which is an important feature in both the
design and rationalization of the SmI2-promoted pinacol
(8) (a) Pedersen, H. L.; Christensen, T. B.; Enemærke, R. J .;
Daasbjerg, K.; Skrydstrup, T. Eur. J . Org. Chem. 1999, 565, 5. (b)
Christensen, T. B.; Riber, D.; Daasbjerg, K.; Skrydstrup, T. Chem.
Commun. 1999, 2051.
(9) For recent reviews on the application of SmI2 in organic
synthesis, see: (a) Krief, A.; Laval, A.-M. Chem. Rev. 1999, 99, 745.
(b) Molander, G. A.; Harris, C. R. Chem. Rev. 1996, 96, 307. (c)
Molander, G. A.; Harris, C. R. Tetrahedron 1998, 54, 3321. (d)
Skrydstrup, T. Angew. Chem., Int. Ed. Engl. 1997, 36, 345.
(10) (a) Molander, G. A.; Kenny, C. J . Org. Chem. 1988, 53, 2132.
(b) Molander, G. A.; Kenny, C. J . Am. Chem. Soc. 1989, 111, 8236.
(c) Chiara, J . L.; Cabri, W.; Hanessian, S. Tetrahedron Lett. 1991,
32, 1125. (d) Uenishi, J .; Matsuda, S.; Wakabayashi, S. Tetrahedron
Lett. 1991, 32, 5097. (e) Chiara, J . L.; Mart´ın-Lomas, M. Tetrahedron
Lett. 1994, 35, 2969. (f) Guidot, J . P.; Le Gall, T.; Mioskowski, C.
Tetrahedron Lett. 1994, 35, 6671. (g) Chiara, J . L.; Valle, N. Tetra-
hedron: Asymmetry 1995, 6, 1895. (h) Perrin, E.; Mallet, J .-M.;
Sinay¨, P. Carbohydr. Lett. 1995, 1, 215. (i) Sawada, T.; Shirai,
R.; Iwasaki, S. Tetrahedron Lett. 1996, 37, 885. (j) Kornienko, A.;
d′Alarcao, M. Tetrahedron Lett. 1997, 38, 6497. (k) Carpintero, M.;
Ferna´ndez-Mayoralas, A.; J aramillo, C. J . Org. Chem. 1997, 62, 1916.
(l) Adinolfi, M.; Barone, G.; Iadonisi, A.; Mangoni, L.; Manna, R.
Tetrahedron 1997, 53, 11767. (m) J enkins, D. J .; Potter, B. V. L. J .
Chem. Soc., Perkin Trans. 1 1998, 41. (n) Adinolfi, M.; Barone, G.;
Iadonisi, A.; Mangoni, L. Tetrahedron Lett. 1998, 39, 2021. (o) Storch
de Garcia, I.; Dietrich, H.; Bobo, S.; Chiara, J . L. J . Org. Chem. 1998,
63, 5883.
(13) In contrast to the diphenylhydrazones, which are stereochemi-
cally homogeneous compounds, the benzyloximes were prepared as a
mixture of cis-and trans-isomers (2:3).
(14) For other ketyl-oxime cyclizations, see: (a) Corey, E. J .; Pyne,
S. G. Tetrahedron Lett. 1983, 24, 2821. (b) Shono, T.; Kise, N.; Fujimoto,
T.; Yamanami, A.; Nomura, R. J . Org. Chem. 1994, 59, 1730. (c) Naito,
T.; Tajiri, K.; Harimoto, T.; Ninomiya, I.; Kiguchi, T. Tetrahedron Lett.
1994, 35, 2205. (d) Chiara, J . L.; Marco-Contelles, J .; Khiar, N.; Gallego,
P.; Destabel, C.; Bernabe´, M. J . Org. Chem. 1995, 60, 6010. (e) Kiguchi,
T.; Tajiri, K.; Ninomiya, I.; Naito, T.; Hiramatsu, H. Tetrahedron Lett.
1995, 36, 253. (f) Camps, P.; Font-Bard´ıa, M.; Mun˜oz-Torrero, D.;
Solans, X. Liebigs Ann. 1995, 523. (g) Naito, T.; Torieda, M.; Tajiri,
K.; Ninomiya, I.; Kiguchi, T. Chem. Pharm. Bull. 1996, 44, 624. (h)
Marco-Contelles, J .; Gallego, P.; Rodr´ıguez-Ferna´ndez, M.; Khiar, N.;
Destabel, C.; Bernabe´, M.; Mart´ınez-Grau, A.; Chiara, J . L. J . Org.
Chem. 1997, 62, 7397. (i) Tormo, J .; Hays, D. S.; Fu, G. C. J . Org.
Chem. 1998, 63, 201. (j) Bobo, S.; Storch de Gracia, I.; Chiara, J . L.
Synlett 1999, 1551.
(15) In the total synthesis of balanol by Naito and co-workers,4h,i
the enantiomerically pure hexahydroazepine fragment was obtained
by optical resolution of the racemic product prepared after the ketyl
cyclization step, followed by hydrogenolysis and acylation leading to
the complete seven-membered ring unit.
(11) (a) Fallis, A. G.; Sturino, C. F. J . Am. Chem. Soc. 1994, 116,
7447. (b) Fallis, A. G.; Sturino, C. F. J . Org. Chem. 1994, 59, 6514.
(12) For a recent review, see: Fallis, A. G.; Brinza, I. M. Tetrahedron
1997, 53, 17543.

5384 J . Org. Chem., Vol. 65, No. 17, 2000
Riber et al.
Sch em e 1
Sch em e 2
oxidation of the corresponding cyclic ketone followed by
DIBAL-H reduction, hydrazone formation and oxidation
with DMSO and SO3/Pyr.
The preparation of the precursor to balanol’s azepine
ring was accomplished via a simple five-step synthesis
(Scheme 1). Tosylamide 7 was N-alkylated with 4-butenyl
bromide followed by conversion of the obtained alkene
to its primary alcohol 9 by a hydroboration/oxidation
sequence. However, subsequent hydrolysis of the acetal
function to the corresponding aldehyde proved not to be
such a facile task as exemplified when the acetal was
subjected to standard hydrolysis conditions such as
AcOH/THF/H2O. Although the desired aldehyde was
obtained in 60% yield, it was accompanied by a less polar
aldehyde 12 (21% yield) in which the primary alcohol had
been acetylated even in the presence of H2O (Scheme 2).
A similar result was observed under conditions such as
CF3CO2H/CHCl3/H2O furnishing the trifluoroacetate 13
in a 53% yield and only 14% of the desired aldehyde. The
N-tosylamide function could be the possible culprit to this
unwanted side reaction with the intervention of a seven-
membered intermediate 14, as shown in Scheme 2, from
the acid-catalyzed expulsion of the hydroxyl group by the
N-tosylamide.17 Acetate attack and opening would then
lead to either 12 or 13.
cyclization with asymmetric ligands. Hence, a clear
picture of the mechanism of this reaction was required,
including a more indepth comprehension of the role the
presence of HMPA has for successful cyclizations. In this
paper, we report that cyclization of carbonylhydrazones
to seven-membered cyclic amino alcohols is also an
efficient process affording the same high trans-selectivity
as for the above smaller rings and provide evidence that
HMPA not only serves to increase the reducing power of
SmI2,16 but is indispensable for promoting the pinacol
cyclization itself. We also suggest that HMPA is respon-
sible for the trans-selectivity observed, without having
to include a chelated ketyl intermediate.11 Furthermore,
initial studies for the development of an enantioselective
pinacol cyclization induced by SmI2 are presented.
Whereas the use of PTSA in THF/water leads to some
decomposition, quantitative yields of the correct aldehyde
were obtained by refluxing a 5% aqueous acetone solution
of the diethyl acetal with 0.5 equiv of PPTS for 4 days.
Subsequent hydrazone formation to 15 and oxidation
afforded the carbonylhydrazone 16 as a crystalline solid
(mp 94-96 °C) with an overall yield of 52% for the five
steps.
The corresponding eight-membered ring precursor, the
carbonylhydrazone 18, and the benzyl oxime 20 were
prepared under identical conditions as shown in Scheme
1 in high yields. The latter compound was obtained as a
1:1.2 mixture of cis- and trans-isomers at the carbon-
nitrogen double bond.
Cycliza tion Stu d ies. The conditions reported by
Fallis and Sturino11a for promoting ring formation (SmI2/
HMPA) were first attempted with precursors 4 and 6.
Addition of a 0.1 M THF solution of SmI2 (4 equiv) to a
THF solution of 4 and HMPA (8-12 equiv) led to rapid
consumption of the carbonylhydrazone affording essen-
Resu lts a n d Discu ssion s
Syn th esis of th e Cyclic P r ecu r sor s. To study the
cyclization reaction for the construction of balanol’s
disubstituted hexahydroazepine ring, a series of six-,
seven- and eight-membered ring precursors were pre-
pared according to Scheme 1. The carbonylhydrazones 4
and 6 were synthesized employing the protocol reported
by Fallis and Sturino,11a involving initial Baeyer-Villiger
(16) (a) Otsubo, K.; Inanaga, J .; Yamaguchi, M. Tetrahedron Lett.
1986, 27, 5763. (b) Inanaga, J .; Ishigawa, M.; Yamaguchi, M. Chem.
Lett. 1987, 1485. (c) Otsubo, K.; Inanaga, J .; Yamaguchi, M. Chem.
Lett. 1987, 1487. (d) Handa, Y.; Inanaga, J .; Yamaguchi, M. J . Chem.
Soc., Chem. Commun. 1989, 298. (e) Ujikawa, O.; Inanaga, J .;
Yamaguchi, M. Tetrahedron Lett. 1989, 30, 2837.
(17) Denieul, M.-P.; Quiclet-Sire, B.; Zard, S. Z. Tetrahedron Lett.
1996, 37, 5495.

SmI2-Promoted Pinacol-Type Cyclization
J . Org. Chem., Vol. 65, No. 17, 2000 5385
Ta ble 1. Cycliza tion of Ald eh yd es 4, 6, 16, a n d 20 w ith
Sm I2/HMP A
Sch em e 3
CHN- protons are observed at approximately 3.6 and 2.8
ppm, respectively, in addition to possessing similar
coupling patterns. These observations also suggest that
22 possesses the same trans-relationship between the
vicinal substituents.
Attempts to cyclize the carbonyloxime 20 (R ) Ts), an
analogue of the Naito compound (R ) BOC),4h,i resulted
in a less clean reaction, and at best we could isolate the
azepine ring in only a 40% yield with a trans:cis-ratio of
approximately 5:1 (Table 1, entry 4). On the other hand,
extension of this reaction to the creation of the eight-
membered heterocyclic ring 25 from carbonylhydrazone
18 (entry 5) proved less rewarding with an array of
compounds formed according to TLC analysis. Appar-
ently, the limit in successful radical cyclizations has been
reached with this attempted 8-exo ring closure event.
Definite confirmation for the trans-stereochemistry
observed in the seven-membered ring closures was
obtained by carrying 23 through to the complete struc-
ture of balanol’s hexahydroazepine ring. Hence, a two-
step procedure involving reduction of the hydrazine with
18
Pearlman’s catalyst/H2 and acylation with p-benzyloxy-
benzoyl chloride (Scheme 3) afforded the corresponding
crystalline p-benzyloxybenzamide 26 (mp 166-167 °C),
of which its single crystal X-ray structure clearly reveals
the trans-relationship between the two vicinal ring
substituents (see Supporting Information).
tially only one cyclic product 21 with the trans-stereo-
chemistry in 57% yield, identical to that earlier reported
(Table 1, entry 1). Repetition of this experiment with the
homologue 6 proved equally rewarding, furnishing the
disubstituted cycloheptane 22 in the same yield and high
trans-selectivity (entry 2), the stereochemical proof of
which will be described below. Somewhat surprising is
the similarity of these two reactions in both yields and
stereoselectivity, considering the anticipated greater rate
constant for 6-exo cyclization in comparison to the 7-exo
ring formation, and that the latter cyclization would
require an 11-membered cyclic chelate compared to an
equally unfavored 10-membered intermediate.11a In any
event, this sole example suggests that the SmI2-promoted
pinacol cyclization of carbonylhydrazones also provides
an efficient entry to seven-membered carbocycles.
To identify other conditions for improving both the
cyclization yields and the trans:cis-selectivity, we next
investigated the same intramolecular pinacol reactions
employing SmI2 alone. This was founded under the
assumption that if a cyclic chelate indeed is the true
intermediate in these ketyl-hydrazone additions, an
increase in the trans-selectivity would be expected owing
to the better Lewis acid properties of the metal ion in
the absence of the electron-donating ligand, HMPA.
However, the groups of Fallis and Naito report no
reactivity without HMPA. This is somewhat puzzling
considering that alkyl aldehydes have previously been
known to undergo either intra- or intermolecular SmI2-
induced pinacol coupling in the absense of HMPA. More
puzzling was it, when carbonylhydrazone 16 was sub-
jected to SmI2 in THF at room temperature, that con-
sumption of the divalent lanthanide reagent was com-
pleted after only 10 min! However, the major compound
isolated after workup was surprisingly not the cyclic
product, of which none was detected, but the diol 29 (69%,
approximately 1:1 mixture of diastereomers) from inter-
molecular pinacol coupling (Table 2, entry 3). This chemi-
cal divergence was unanticipated under the simple pre-
sumption that radical cyclizations should be favored
When the same reaction conditions were applied to the
acyclic derivative 16 (Table 1, entry 3), we were pleased
to observe the formation of the desired hexahydroazepine
ring 23 in 59% yield though with a slightly reduced
1
trans:cis ratio of 10:1. The H NMR shifts and coupling
constants of the methine protons in the ring correspond
well with those for similar compounds possessing the
same trans-relationship between the vicinal substitu-
ents,4 whereas with the isomeric cis-compound, the H4
proton is downfield shifted by 0.5 ppm to δ 4.05 compared
to δ 3.52 for 23. This assignment was based on 2D NMR
techniques. This same trend was also observed with the
cis- and trans-isomers of 21, where the same protons are
observed at 4.02 and 3.68, respectively. In the case of the
cycloheptane derivative 22, a 1H NMR spectrum was
noted similar to that of the trans-21 where the CHO- and
(18) Miyata, O.; Muroya, K.; Koide, J .; Naito, T. Synlett 1998, 271.

5386 J . Org. Chem., Vol. 65, No. 17, 2000
Riber et al.
Ta ble 2. In ter m olecu la r P in a col Cou p lin g of Ald eh yd es
4, 6, 16, a n d 20 w ith Sm I2
Sch em e 4
Attempts to promote the cyclization with in situ
generated SmBr2 and SmCl2 were also made. Previous
electrochemical investigations show that these reagents,
prepared from SmI2 with the corresponding lithium
halide, are more potent reducing agents with oxidation
peaks (-1.78 V for SmBr2, -1.99 V for SmCl2 vs Fc+/Fc
in THF)21 similar or in the vicinity of SmI2 complexed
with four HMPA (-1.82 V) as determined by cyclic
voltammetry.20 It was therefore interesting to examine
the effect that would be exerted on the cyclization event
upon decreasing the oxidation potential of SmI2 in the
absence of the bulky hexamethylphosphoramide ligand.
As shown in Table 2 (entries 5 and 6), treatment of the
aldehyde 16 with SmI2, previously subjected to either
excess LiBr or LiCl, led either exclusively to the reduction
of the aldehyde function to the corresponding primary
alcohol as in 15 or to a 3:1 mixture of 15 and the cyclized
product 21, respectively, though in both cases without
any traces of the diol 29.
Finally, the recent findings that the low-valent tita-
nium complexes, Cp2TiCl22 and the more reactive Cp2-
TiPh,23 catalyze the inter- and intramolecular pinacol
couplings of alkyl aldehydes prompted us to investigate
these reagents as possible candidates for promoting the
cyclic amino alcohol formation. However, the catalytic
procedures reported employing Cp2TiCl2(5%)/Zn/TMSCl
or Cp2Ti(Ph)Cl(5%)/Zn/TMSCl were ineffective in cata-
lyzing the ring closure of carbonylhydrazone 16. In one
try, 16 was treated with both 3 equiv of Cp2TiPh
(prepared from Cp2TiCl2 with iPrMgCl and then Ph-
MgBr) and TMSCl, but even after 24 h no sign of any
reaction was observed. To test if Cp2TiPh had indeed
been generated in this reaction mixture, 1 equiv of
benzaldehyde was added, resulting in the rapid consump-
tion of aldehyde 16. The major product isolated was the
alcohol 31 (53% yield), in which phenyl group transfer
had occurred from the titanium metal center to aldehyde
16 (Scheme 4). This reaction is most likely promoted by
the corresponding TiIV pinacolate generated by the reac-
tion of Cp2TiPh with benzaldehyde or Cp2Ti(Ph)Cl formed
upon liberation of the TiIV metal ion from the pinacolate
with TMSCl. It is nevertheless noteworthy that such side
reactions were not reported in the previous ketyl cycliza-
tions using this reagent.
a
Reaction was performed with SmBr2 by pretreating the THF
b
solution of SmI2 with LiBr (10 equiv). See ref 21. Reaction was
performed as in entry 5 but with LiCl.
under these less reducing conditions.19,20 The same ten-
dency was observed with the carbonylhydrazones 4 and
6 (entries 1 and 2), as well as with the corresponding
benzyloxime 20 (entry 4). Whereas 6 and 20 led to the
formation of the vicinal diols 28 and 30 in 56% and 50%
yield, respectively, the treatment of the six-membered
ring precursor 4 with SmI2 was less clean, affording sev-
eral compounds. The major product was again identified
as the diol 27 (39%) from intermolecular pinacol coupling.
However, in this case the cyclic amino alcohol 21 could
be isolated, albeit in a low yield (10%) and a low selec-
tivity (trans:cis, 4:1). The other minor products were
identified as esters arising from Tischenko side reactions.
(19) (a) Hasegawa, E.; Curran, D. P. Tetrahedron Lett. 1993, 34,
1717. (b) Maze´as, D.; Skrydstrup, T.; Doumeix, O.; Beau, J .-M. Angew.
Chem., Int. Ed. Engl. 1994, 33, 1383. (c) Skrydstrup, T.; Maze´as, D.;
Elmouchir, M.; Doisneau, G.; Riche, C.; Chiaroni A.; Beau, J .-M. Chem.
Eur. J . 1997, 3, 1342.
(20) We have recently reported by cyclic voltammetry and con-
ductivity measurements the standard potential (E° ) for the redox
couples SmI2+/SmI2 and Sm(HMPA)43+ 2I-/Sm(HMPA)42+ 2I- in THF,
found to be -1.43 and -2.23 V vs Fc+/Fc, respectively. (a) Enemærke,
R. J .; Daasbjerg, K.; Skrydstrup, T. Chem. Commun. 1999, 343. (b)
Enemærke, R. J .; Hertz, T.; Skrydstrup, T.; Daasbjerg, K. Chem. Eur.
J ., in press. See also: (c) Shabangi, M.; Kuhlman, R. A.; Flowers, R.
A., II. Org. Lett. 1999, 1, 2133. (d) Shabangi, M.; Flowers, R. A., II.
Tetrahedron Lett. 1997, 38, 1137. (e) Shotwell, J . B.; Sealy, J . M.;
Flowers, R. A., II. J . Org. Chem. 1999, 64, 5251.
Mech a n istic Asp ects of th e Sm II-P r om oted Ketyl
Cycliza tion . A possible mechanistic scenario is depicted
(21) Fuchs, J . R.; Mitchell, M. L.; Shabangi, M.; Flowers, R. A., II.
Tetrahedron Lett. 1997, 38, 8157.
(22) Harao, T.; Hatano, B.; Asahara, M.; Muguruma Y.; Ogawa, A.
Tetrahedron Lett. 1998, 39, 5247.
(23) (a) Yamamoto, Y.; Matsumi, D.; Itoh, K. Chem. Commun. 1998,
875. (b) Yamamoto, Y.; Hattori, R.; Itoh, K. Chem. Commun. 1999,
825.

SmI2-Promoted Pinacol-Type Cyclization
J . Org. Chem., Vol. 65, No. 17, 2000 5387
Sch em e 5
with C may not be participating at all. The results
obtained upon the treatment of the carbonylhydrazone
5 with SmCl2 or SmBr2 (Table 2, entries 5 and 6) lend
support to this hypothesis. Under these high reducing
conditions, the ketyl intermediate A undergoes a second
reduction step to the corresponding anion C, which
apparently will not cyclize and therefore is subsequently
protonated, affording the primary alcohol.
Ketyl addition to hydrazones may be reversible reac-
tions (A T B, Scheme 5), in analogy to both the intramo-
lecular alkyl radical addition and opening of the 5-pen-
tanal and 6-hexanal radicals,27 as well as ketyl addition
to ketones.10b If so, it may be expected that the hydrazine
radical intermediate would be reduced more effectively
under the higher reducing conditions with SmI2/HMPA
and faster than the competing pinacol coupling of the
ketyl. On the other hand, with the poorer reducing agent,
SmI2 alone, where the standard potential E° for the redox
couple SmI2+/SmI2 is approximately 800 mV more posi-
tive than for equivalent redox couple with SmI2 com-
plexed with HMPA (4 equiv),20 the pinacol coupling step
dominates.28
Another plausible mechanism may be considered in
which the ratio kcycl/kpin is greatly influenced by the
addition of HMPA to the reaction mixtures, thus result-
ing in the observed divergence of product formation. In
the ketyl-alkene cyclizations, Molander suggests the
steric bulk around the ketyl oxygen is increased by
HMPA complexation with the metal center, hence in-
creasing the persistence of the ketyl and defavoring
intermolecular side reactions.26 This may also explain the
observed reduction product obtained when subjecting the
carbonylhydrazone 5 to SmBr2 or SmCl2. A second
explanation may lie in the reactivity difference between
the two ketyl radical anions. It has been well established
in electrochemical reductions of aromatic compounds
containing carbonyl groups that the rate of pinacol
dimerization is substantially increased upon hydrogen-
bonding with the ketyl radical anion.29 For example, in
the pinacol coupling of 3-chloroacetophenone, the rate
in Scheme 5. Upon reduction of the aldehyde to a SmIII
bound ketyl radical anion A,24 three options are avail-
able: (1) ketyl cyclization onto the hydrazone generating
an aminyl radical B followed by a reduction step; (2)
intermolecular pinacol coupling affording the dimeric
product; and (3) a two-step reduction of the aldehyde to
its dianion C followed by an ionic cyclization. Previous
kinetic studies by Fallis et al. have demonstrated that
secondary alkyl radical additions to the diphenylhydra-
zones are fast processes with rate constants as high as
1.1 × 108 s-1 (80 °C) for the 5-exo-cyclization process
(cis-product) and 9.4 × 105 s-1 for the corresponding
homologue (6-exo).11b,25 In competition experiments, 5-exo
and 6-exo ketyl-hydrazone cyclizations were found to
supersede the corresponding cyclizations onto alkenes
by a factor of 25 and 4.2, respectively.11b Both the
polarization of the CdN bond and the increased stability
of the developing nitrogen radical owing to a favorable
two-centered, three-electron interaction with the other
nitrogen explain these increased cyclization rates. It is
therefore plausible that the rate constant for the 7-exo
cyclization onto the diphenylhydrazones may be of an
appreciable size but nevertheless smaller than for both
the 5-exo and 6-exo radical cyclizations. This is indeed
confirmed by the isolation of seven-membered ring
products upon cyclization of the carbonylhydrazone with
SmI2/HMPA. On the other hand, the 8-exo cyclization
is probably just too slow to compete with other side
reactions.
constant in THF was measured to be 1.0 × 103 M-1 s-1
,
whereas with added water an increase to 1 × 104 M-1
s-1 (100 mM H2O) and >1 × 107 M-1 s-1 (1 M H2O) was
observed.29d,e,30 This rate enhancement is explained by
an increased spin density at the carbon center of the ketyl
upon protonation, as well as reduced electrostatic repul-
sions in the transition state of the dimerization. Hence,
the more anionic character of the ketyl, the slower the
dimerization process. A weaker Sm-O bond would prob-
ably be expected in the samariumIII ketyl complexed with
(27) (a) Tsang, R.; Fraser-Reid, B. J . Am. Chem. Soc. 1986, 108,
8102. (b) Tsang, R.; Dickson, J . K. J r.; Pak, H.; Walton, R.; Fraser-
Reid, B. J . Am. Chem. Soc. 1987, 109, 3484. (c) Beckwith, A. L. J .;
Hay, B. J . J . Am. Chem. Soc. 1989, 111, 230. (d) Beckwith, A. L. J .;
Hay, B. J . J . Am. Chem. Soc. 1989, 111, 2674.
(28) Of course, a direct comparison between the difference of the
The facts that successful ketyl-olefin ring closures of
up to eight-membered rings have been reported with
SmI2/HMPA26 and the radical cyclizations onto hydra-
zones and oximes are faster than for the corresponding
unactivated alkenes11b suggest that anionic cyclizations
E° of the redox couples SmI2+/SmI2 and Sm(HMPA)4
2I-/Sm-
3+
(HMPA)4 2I- with the difference in the rates of reduction of the
nitrogen radical intermediates cannot be made, as the latter process
most likely involves an inner-sphere electron transfer.
2+
(29) (a) Lamy, E.; Nadjo, L.; Save´ant, J .-M. J . Electroanal. Chem.
1974, 61, 141. (b) Save´ant, J .-M.; Tessier, D. J . Electroanal. Chem.
1975, 61, 251. (c) Fussing, I.; Gu¨llu¨, M.; Hammerich, O.; Husaain, A.;
Nielsen, M. F.; Utley, J . H. P. J . Chem. Soc., Perkin Trans 2 1996,
649. (d) J ensen, H.; Daasbjerg, K. Acta Chem. Scand. 1998, 52, 1151.
(e) Daasbjerg, K., personal communication.
(24) The hydrazone does not react with either SmI2 nor SmI2/HMPA
even after prolonged periods (several hours).
(25) Tauh, P.; Fallis, A. G. J . Org. Chem. 1999, 64, 6960.
(26) 8-endo ketyl-olefin cyclizations: (a) Molander, G. A.; McKie, J .
A. J . Org. Chem. 1994, 59, 3186. (b) Dinesh, C. U.; Reissig, H.-U.
Angew. Chem., Int. Ed. Engl. 1999, 38, 789.
(30) The rate constants for the dimerization of both the free and
protonated ketyl of benzaldehyde has been measured (PhCHO•-, 2.0
× 105 M-1 s-1; PhCHOH•, ∼109 M-1 s-1). Andrieux, C. P.; Grzeszczuk,
M.; Save´ant, J .-M. J . Am. Chem. Soc. 1991, 113, 8811.

5388 J . Org. Chem., Vol. 65, No. 17, 2000
Riber et al.
Ta ble 3. Cycliza tion Stu d ies w ith Ch ir a l Liga n d s
F igu r e 2. Single-crystal X-ray structure of the hexahydro-
azepine 26.
HMPA compared to the uncomplexed species, owing to
the strong electron-donating abilities of this ligand. The
former complex would therefore display more anionic
character at the oxygen, resulting in the generation of a
more persistent ketyl with kpin(SmI2/HMPA) < kpin(SmI2). In
addition, as the electrostatic interactions are less impor-
tant for ketyl-hydrazone addition reactions than for the
dimerization, the cyclization rates may not be influenced
as greatly upon addition of HMPA. With SmCl2, SmBr2
and SmI2, a tighter Sm-O bond would be anticipated,
resulting in a greater spin density at the ketyl carbon
center. The slow cyclization event, in particular for the
7-exo ring formation, cannot compete with the fast
intermolecular pinacol coupling when employing SmI2.
The high reducing conditions with SmCl2 and SmBr2,
however, would result in the quick reduction of the ketyl
before the pinacol dimerization.
Finally, the observation that cyclization of 4 with SmI2
leads to a low trans:cis ratio is inconsistent with a
chelated model (Scheme 1).11a The Lewis acid ability of
the samariumIII metal ion is anticipated to be greater in
the absence of HMPA, thus favoring a chelated inter-
mediate. Instead, the presence of the complexing agent
may simply just increase the sterical interaction between
the SmIII-ketyl26,31 with the large diphenylhydrazone
substitutent, forcing the two substituents to adopt pseudo-
axial orientations in the transition state rather than the
opposing equatorial (Figure 2). A similar explanation has
recently been put forth by Chiara and co-workers in their
work on the cyclization of carbohydrate-based carbonyl-
oximes to form five-membered ring systems,14j as well as
Itoh et al. in their trans-selective Cp2TiPh-promoted
pinacol cyclizations.23b
At t em p t s At Develop in g An E n a n t ioselect ive
Ketyl-Hyd r a zon e Cycliza tion . To develop an asym-
metric version of the pinacol-type cyclization, we inves-
tigated the possibility of adding chiral complexing agents
to SmI2 prior to the addition of the carbonylhydrazone.32
It was hypothesized that such ligands would screen the
one face of the ketyl intermediate, hence biasing the
direction of radical attack onto the hydrazone. With the
above results in hand, it became clear that for successful
cyclization to occur, such chiral ligands must display
strong electron-donating abilities, similar to HMPA.
Table 3 displays some of the ligands tested with the
carbonylhydrazone 4 and 16. In all cases, equimolar
amounts of ligands 32,32c 33,33 34, and 3534 were pre-
mixed with a THF solution of SmI2 for approximately 30
min before the ketyl precursor was added. Of the four
ligands examined, only 33 (entry 2) showed visible signs
of complexation to the divalent lanthanide ion with a
change of the solution color from blue to purple, essen-
tially as observed with HMPA. This was also manifested
in the subsequent cyclization studies where the ligand
33 promoted rapid ring formation affording 21 in the
highest yield yet obtained for this event (entry 2).
Nevertheless, this reaction was characterized by a sur-
prisingly poor diastereoselectivity, despite its closest
resemblance to HMPA, and essentially no asymmetric
induction. With ligands 32, 34, and 35 (entries 1, 3, and
4), formation of the cyclic amino alcohol 21 or 23 was
slow (approximately 15-30 min) and low yielding, in-
dicative of their poor electron-donating capacities. It is
not clear whether the ligands were coordinated at all to
either the divalent or trivalent metal ion, but both the
diastereo- and enantioselectivities were low or nonexist-
ent, respectively.
Con clu sion s
(31) (a) Molander, G. A.; McKie, J . A. J . Org. Chem. 1992, 57, 3132.
(b) Molander, G. A.; McKie, J . A. J . Org. Chem. 1995, 60, 872.
(32) For some recent examples on asymmetric synthesis with SmI2,
see: (a) Molander, G. A.; McWilliams, J . C.; Noll, B. C. J . Am. Chem.
Soc. 1997, 119, 1265. (b) Fukuzawa, S.; Seki, K.; Tatsuzawa, M.;
Mutoh, K. J . Am. Chem. Soc. 1997, 119, 1482. (c) Mikami, K.;
Yamaoka, M. Tetrahedron Lett. 1997, 38, 1137.
(33) Yang, W.-B.; Fang, J .-M. J . Org. Chem. 1998, 63, 1356.
(34) Ligands 24 and 25 were designed based on the known strong
binding interactions of amides with SmI2 in other systems.32a,b
The work described herein has shed some light onto
the role of HMPA in the samarium iodide induced
pinacol-type coupling of carbonylhydrazones. The electron-
donating ligand appears to promote the ketyl-hydrazone
cyclization for the formation of six- and seven-membered
cyclic amino alcohols by modulating the rate constants
for the ring formation and the competing intermolecular

SmI2-Promoted Pinacol-Type Cyclization
J . Org. Chem., Vol. 65, No. 17, 2000 5389
mmol) in THF (5 mL), 0.1 M SmI2 in THF (5.1 mL, 0.51 mmol)
and HMPA (220 µL, 1.26 mmol). Flash chromatography
(petroleum ether/EtOAc, 7:5), afforded 24 (25 mg, 40%) as a
colorless oil and as an inseparable 5:1 mixture of trans:cis
isomers: 1H NMR (200 MHz, CDCl3) δ 1.51-2.02 (m, 4H), 2.42
(s, 3H), 2.90 (dt, 1H, J ) 5.0, 8.6 Hz), 3.06 (ddd, 1H, J ) 5.0,
10.8, 12.6 Hz), 3.25-3.65 (m, 4H), 4.71 (s, 2H), 7.25-7.41 (m,
7H), 7.65 (d, 2H, J ) 8.0 Hz); 13C NMR (50 MHz, CDCl3) δ
21.7, 24.8, 31.5, 47.4, 49.7, 53.6, 67.7, 72.9, 127.2 (2C), 128.2,
128.5, 128.6 (2C), 128.8 (2C), 129.9 (2C), 136.1, 137.9; MS
(electrospray) m/z 413 (M + Na). HRMS: m/e calcd for
pinacol coupling, a side reaction that predominates in the
absense of HMPA. This reaction has been applied suc-
cessfully to the synthesis of the fully functionalized
hexahydroazepine ring of balanol in racemic form. Our
intial investigations using chiral complexing agents
suggest the use of strong donating ligands which re-
semble HMPA in structure. This study is currently
ongoing and will be reported in due course.
Exp er im en ta l P r oced u r e
C
20H26N2NaO4S (M + Na), 413.1511; found, 413.1515.
Ba la n ols Hexa h yd r oa zep in e F r a gm en t 26. A mixture
Gen er a l Con sid er a tion s. Unless otherwise stated, all
reactions were carried out under argon. THF was dried and
freshly distilled over sodium/benzophenone. Dichloromethane
was freshly distilled over P2O5. Reactions were monitored by
thin-layer chromatography (TLC) analysis. The enantiomeric
excess measured on compound 21 in the cyclization studies
employing chiral ligands was performed using chiral chroma-
tography (Chiral Pack AD column) eluting with hexane/2-
propanol (99/1, v/v). The following compounds were prepared
according to literature procedures: carbonylhydrazone 411a and
ligand 33.33
of the cyclization product 23 (95 mg, 0.21 mmol), camphor-
sulfonic acid (110 mg, 0.44 mmol) and Pearlman’s reagent (20%
Pd(OH)2/C) (100 mg) in methanol (6 mL) was left overnight
in an autoclave at 40 bar of H2. The reaction mixture was
filtered, diluted with CH2Cl2 and then washed with an aqueous
solution of NaOH (0.2 M) and brine. The organic phase was
dried over MgSO4 and evaporated to dryness. The resulting
oil was redissolved in dry CH2Cl2, whereafter Et3N (154 µL,
1.11 mmol), p-benzyloxybenzoyl chloride (52 mg, 0.21 mmol)
and DMAP (2 mg) were added. The reaction mixture was
allowed to stir overnight and then washed with water and
brine, dried over MgSO4 and evaporated to dryness. Purifica-
tion by flash chromatography (petroleum ether/EtOAc, 2:1)
yielded 50 mg (48%) of a solid, which could be recrystallized
from petroleum ether/ethyl acetate, affording 26 as colorless
needles: mp 166-167 °C; 1H NMR (200 MHz, CDCl3) δ 1.65-
2.05 (m, 4H), 2.45 (s, 3H), 2.54-2.70 (m, 1H), 3.15 (dd, 1H, J
) 2.6, 15.8 Hz), 3.70 (d, 1H, J ) 15.8 Hz), 3.84-4.06 (m, 3H),
4.88 (s, 1H), 5.13 (s, 2H), 7.04 (d, 2H, J ) 8.8 Hz), 7.33-7.46
(m, 7H), 7.69 (d, 2H, J ) 8.6 Hz), 7.82 (d, 2H, J ) 8.8 Hz),
8.03 (d, 1H, J ) 4.8 Hz); 13C NMR (50 MHz, CDCl3) δ 21.8,
226.1, 31.5, 50.2, 51.6, 58.2, 70.3, 77.6, 115.0 (2C), 125.9, 127.2
(2C), 127.7 (2C), 128.3, 128.9 (2C), 129.4 (2C), 130.2 (2C), 135.0,
136.5, 144.3, 161.9, 168.4; MS (electrospray) m/z 517 (M + Na).
HRMS: m/e calcd for C27H30N2NaO5S (M + Na), 517.1773;
found, 517.1776.
2-(N′,N′-Dip h en ylh yd r a zin o)cycloh ep ta n ol (22). Gen -
er a l P r oced u r e for th e Cycliza tion Rea ction s w ith Sm I2/
HMP A. The aldehyde 6 (89 mg, 0.30 mmol) in a dry flask
flushed with argon was dissolved in THF (5 mL) and HMPA
(500 µL, 2.86 mmol). To this solution was added a 0.1 M
solution of SmI2 in THF (12.0 mL, 1.20 mmol) at 20 °C. After
a few minutes of stirring the reaction mixture was quenched
with saturated aqueous NaHCO3. The resulting mixture was
diluted with ethyl acetate and the organic phase was washed
with water. The aqueous phase was extracted three times with
ethyl acetate and the combined organic layers were washed
with brine, dried over MgSO4 and concentrated in vacuo.
Further purification was affected by flash chromatography on
silica gel deactivated by washing first with a 1% solution of
Et3N added to the eluting solution (petroleum ether/EtOAc,
9:1). This afforded 51 mg (57%) of the cyclization product 22
as a colorless oil: 1H NMR (200 MHz, CDCl3) δ 1.23-1.90 (m,
10H), 1.95 (d, 1H, J ) 4.0 Hz), 2.83 (dt, 1H, J ) 3.5, 7.2 Hz),
3.68 (tt, 1H, J ) 4.0, 7.2 Hz) 6.95-7.06 (m, 2H), 7.14-7.37
(m, 8H); 13C NMR (50 MHz, CDCl3) δ 22.2, 23.1, 27.0, 27.7,
34.6, 63.5, 68.6, 119.9 (4C), 121.7 (2C), 128.5 (4C), 147.5 (2C);
MS (electrospray) m/z 319 (M + Na); HRMS: m/e calcd for
1,14-Bis(N ′,N ′-d ip h e n ylh yd r a zon o)t e t r a d e ca n e -7,8-
d iol (28). Gen er a l P r oced u r e for th e In ter m olecu la r
P in a col Cou p lin g. A 0.1 M solution of SmI2 in THF (8.5 mL,
0.85 mmol) was added to the aldehyde 6 (109 mg, 0.37 mmol)
in a dry flask flushed with argon. Workup was affected by
quenching with saturated aqueous NaHCO3 after TLC analysis
indicated the consumption of starting material (less than 15
min). The resulting mixture was diluted with ethyl acetate
and washed with water. The aqueous phase was extracted
three times with ethyl acetate and the combined organic layers
were then washed with brine, dried over magnesium sulfate
and concentrated in vacuo. Further purification was affected
by flash chromatography on silica gel deactivated by washing
first with a 1% solution of Et3N added to the eluting solution
(petroleum ether/EtOAc, 3:1). The pinacol product 28 was
obtained as a colorless syrup (61 mg, 56%): 1H NMR (200
MHz, CDCl3) δ 1.22-1.61 (m, 16H), 1.82 (m, 1H), 1.98 (m, 1H),
2.27 (dt, 4H, J ) 5.6, 7.2 Hz), 3.38 (m, 1H), 3.58 (m, 1H), 6.52
(t, 2H, J ) 5.6 Hz), 7.04-7.15 (m, 12H), 7.31-7.41 (m, 8H);
13C NMR (50 MHz, CDCl3) δ 25.4, 26.9, 28.9, 32.6 (4C), 62.9,
122.2 (8C), 123.7 (4C), 129.6 (8C), 140.0 (4C), 144.2; MS
(electrospray) m/z 613 (M + Na). HRMS: m/e calcd for
C
19H24N2NaO (M + Na), 319.1786; found, 319.1787.
3-(N′,N′-Dip h en ylh yd r a zin o)-1-(tolu en e-4-su lfon yl)a ze-
p in -4-ol (23). The hexahydroazepine 23 was prepared from
carbonylhydrazone 16 according to the general procedure
outlined for 22, with the following quantities: aldehyde 16 (90
mg, 0.20 mmol) in THF (2 mL), 0.1 M SmI2 in THF (8.0 mL,
0.80 mmol) and HMPA (278 µL, 1.60 mmol). Flash chroma-
tography (petroleum ether/EtOAc, 7:5), afforded 23 (50 mg,
1
56%) as a colorless oil. tr a n s-P r od u ct: H NMR (200 MHz,
CDCl3) δ 1.24-1.79 (m, 4H), 2.38 (s, 3H), 2.73 (dd, 1H, J )
9.4, 14.6 Hz), 2.99 (dt, 1H, J ) 3.2, 9.4 Hz), 3.01 (m, 1H), 3.46
(m, 2H), 3.91 (dd, 1H, J ) 3.2, 14.6 Hz), 4.61 (s, 1H), 6.98-
7.11 (m, 2H), 7.13-7.38 (m, 10H), 7.48-7.57 (m, 2H); 13C NMR
(50 MHz, CDCl3) δ 21.7, 22.1, 32.4, 45.0, 47.4, 64.7, 74.8, 121.1
(4C), 122.8 (2C), 127.2 (2C), 129.4 (4C), 129.8 (2C), 135.8, 143.3,
148.6 (2C); MS (electrospray) m/z 474 (M + Na). HRMS: m/e
calcd for C25H29N3NaO3S (M+Na), 474.1827; found, 474.1829.
cis-P r od u ct: 1H NMR (200 MHz, CDCl3) δ 1.52-1.78 (m, 2H),
1.94-2.08 (m, 2H), 2.40 (s, 3H), 2.49 (s, 1H), 3.01 (ddd, 1H, J
) 4.4, 6.2, 12.0 Hz), 3.05 (dd, 1H, J ) 9.4, 14.0 Hz), 3.25 (dt,
1H, J ) 2.6, 9.4 Hz), 3.52 (ddd, 1H, J ) 9.4, 6.2, 15 Hz), 3.63
(m, 1H), 3.96 (s, 1H), 4.05 (m, 1H); 13C NMR (50 MHz, CDCl3)
δ 19.6, 21.7, 29.0, 43.7, 47.9, 63.0, 67.8, 120.9 (4C), 123.4 (2C),
127.1 (2C), 129.6 (4C), 129.9 (2C), 136.1 (2C), 143.3 (2C), 148.5
(2C).
C
38H46N4NaO4 (M + Na), 613.3518; found, 613.3519.
1,12-Bis(N′,N′-d ip h en ylh yd r a zon o)d od eca n e-6,7-d iol
(27). The diol 27 was prepared from carbonylhydrazone 4
according to the general procedure outlined for 28, with the
following quantities: aldehyde 4 (74.4 mg, 0.27 mmol) and 0.1
M solution of SmI2 in THF (10.6 mL, 1.06 mmol). Flash
chromatography (petroleum ether/EtOAc, 2:1) afforded 27 (29
mg, 39%) as a colorless syrup along with the cyclized product
21 (7.2 mg, 10%) as a 4:1 mixture of trans:cis isomers. For 27:
1H NMR (200 MHz, CDCl3) δ 1.22-1.62 (m, 12H), 1.84 (m,
1H), 2.02 (m, 1H), 2.29 (dt, 4H, J ) 5.4, 7.2 Hz), 3.39 (m, 1H),
3.59 (m, 1H), 6.52 (t, 2H, J ) 5.4 Hz), 7.03-7.15 (m, 12H),
7.31-7.40 (m, 8H); 13C NMR (50 MHz, CDCl3) δ 25.4, 25.8,
27.2, 31.2, 32.8, 33.5, 74.6, 74.7, 122.5 (4C), 124.1 (2C), 129.9
3-(2-Ben zyloxya m in oeth yl)-1-(tolu en e-4-su lfon yl)a ze-
p in -4-ol (24). The hexahydroazepine 24 was prepared from
carbonyloxime 20 according to the general procedure outlined
for 22, with the following quantities: aldehyde 20 (49 mg, 0.13

5390 J . Org. Chem., Vol. 65, No. 17, 2000
Riber et al.
(4C), 140.0, 144.5 (2C); MS (electrospray) m/z 585 (M + Na).
HRMS: m/e calcd for C36H42N4NaO2 (M + Na), 585.3205;
found, 585.3217. For trans-21: 1H NMR (200 MHz, CDCl3) δ
1.18-1.38 (m, 4H), 1.57-1.76 (m, 2H), 1.86-2.01 (m, 2H), 2.40
(bs, 1H), 2.74 (dt, 1H, J ) 3.9, 9.7 Hz), 3.68 (dt, 1H, J ) 4.8,
9.7 Hz), 6.95-7.06 (m, 2H), 7.14-7.37 (m, 8H); 13C NMR (50
MHz, CDCl3) δ 23.5, 23.8, 28.9, 33.5, 60.6, 73.9, 120.0 (4C),
121.9 (2C), 128.5 (4C), 147.7 (2C); MS (electrospray) m/z 305
(M + Na). For cis-21: 1H NMR (200 MHz, CDCl3) δ 1.20-
1.81 (m, 7H), 1.96 (m, 1H), 2.55 (s, 1H), 2.96 (ddd, 1H, J )
2.6, 4.6, 10.4 Hz), 3.90 (s, 1H), 4.02 (m, 1H), 6.97-7.14 (m,
6H), 7.25-7.35 (m, 4H); MS (electrospray) m/z 305 (M + Na).
N-(2-(N′,N′-Dip h en ylh yd r a zon o)eth yl)-N-(8-((2-(N′,N′-
d ip h en ylh yd r a zon o)et h yl)b en zen esu lfon yla m in o)-4,5-
d ih yd r oxyoctyl)ben zen esu lfon a m id e (29). The diol 29 was
prepared from carbonylhydrazone 16 according to the general
procedure outlined for 28, with the following quantities:
aldehyde 16 (100 mg, 0.22 mmol) and 0.1 M solution of SmI2
in THF (8.9 mL, 0.89 mmol). Flash chromatography (petroleum
ether/EtOAc, 7:5) afforded 29 (69 mg, 69%) as a colorless
syrup: 1H NMR (200 MHz, CDCl3) δ 1.35-1.81, (m, 8H), 1.91
(d, 1H, J ) 3.8 Hz), 2.13 (d, 1H, J ) 3.8 Hz), 2.40 (s, 6H), 3.14
(t, 4H, J ) 7.2 Hz), 3.34 (m, 1H), 3.51 (m, 1H), 4.00 (d, 4H, J
) 5.6 Hz), 6.20 (t, 2H, J ) 5.6 Hz), 6.95-7.01 (m, 8H), 7.10-
7.39 (m, 16H), 7.59-7.64 (m, 4H); 13C NMR (50 MHz, CDCl3)
δ 21.0, 25.8, 29.5, 47.0, 49.2, 73.3, 121.5 (8C), 123.9 (4C), 126.4
(4C), 129.1 (4C), 129.2 (8C), 131.6, 135.9, 142.6, 142.6 (4C);
MS (electrospray) m/z 923 (M + Na). HRMS: m/e calcd for
analysis showed no consumption of 16, hence TMSCl (17 µL,
0.13 mmol) was injected. Again no consumption of 16 was
observed by TLC analysis, and therefore 3 equiv of benzalde-
hyde were added to test the reactivity of the Cp2TiPh formed.
After 30 min, 16 was completely consumed and the reaction
mixture was quenched with saturated aqueous NH4Cl. CH2-
Cl2 was added and the organic phase was washed once with
water, then dried over MgSO4 and evaporated to dryness.
Purification by flash chromatography (petroleum ether/EtOAc,
3:1) yielded 37 mg (53%) of 31 as a colorless oil: 1H NMR (200
MHz, CDCl3) δ 1.50-1.81 (m, 4H), 2.38 (s, 3H), 2.92 (broad s,
1H), 3.16 (m, 2H), 4.00 (d, 2H, J ) 5.6 Hz), 4.67 (m, 1H), 6.19
(t, 1H, J ) 5.6 Hz), 6.90-7.00 (m, 4H), 7.08-7.40 (m 13H),
7.55-6.64 (m, 2H); 13C NMR (50 MHz, CDCl3) δ 21.8, 24.8,
35.9, 47.9, 50.1, 62.5, 74.2, 79.3, 122.3, 124.7, 126.0, 127.1,
127.2, 127.8, 128.1, 128.3, 128.7, 129.9, 130.0, 132.4, 136.7,
140.1, 143.4, 144.6; MS (electrospray) m/z 550 (M + Na).
HRMS: m/e calcd for C31H33N3NaO3S (M + Na), 550.2131;
found, 550.2131.
Cycliza tion of Ca r bon ylh yd r a zon e 4 w ith Liga n d 33.
A 0.1 M solution of SmI2 in THF (6.3 mL, 0.63 mmol) was
added to the ligand 33 (300 mg, 0.63 mmol) in a dry flask
flushed with argon. After stirring for 10 min, a solution of the
carbonylhydrazone 4 (59 mg, 0.21 mmol) in THF (3 mL) was
added at 20 °C. After a few minutes of stirring, the reaction
mixture was quenched with saturated aqueous NaHCO3. The
resulting mixture was diluted with ethyl acetate and the
organic phase was washed with water. The aqueous phase was
extracted three times with ethyl acetate and the combined
organic layers were washed with brine, dried over MgSO4 and
concentrated in vacuo. Further purification was affected by
flash chromatography on silica gel deactivated by washing first
with a 1% solution of Et3N added to the eluting solution
(petroleum ether/EtOAc, 9:1). This afforded a 4:1 mixture of
trans-21 (37.2 mg) and cis-21 (9.4 mg) with a combined yield
of 77%.
C
50H56N6NaO6S2 (M + Na), 923.3600; found, 923.3599.
N-(2-Ben zyloxyim in oet h yl)-N-(8-((2-b en zyloxyim in o-
et h yl)b en zen esu lfon yla m in o)-4,5-d ih yd r oxyoct yl)b en -
zen esu lfon a m id e (30). The diol 30 was prepared from
carbonyloxime 20 according to the general procedure outlined
for 28, with the following quantities: aldehyde 20 (80 mg, 0.21
mmol) and 0.1 M solution of SmI2 in THF (8.2 mL, 0.82 mmol).
Flash chromatography (petroleum ether/EtOAc, 7:5) afforded
30 (44 mg, 50%) as a colorless syrup: 1H NMR (200 MHz,
CDCl3) δ 1.30-1.80 (m, 8H), 2.41 (s, 6H), 3.15 (m, 8H), 3.86
(d, 2H, J ) 6.2 Hz), 4.04 (d, 2H, J ) 4.2 Hz), 5.02 (s, 2H), 5.09
(s, 2H), 6.64 (t, 1H, J ) 4.3 Hz), 7.22-7.40 (m, 15H), 7.66 (dd,
4H, J ) 1.2, 8.2 Hz); MS (electrospray) m/z 801 (M + Na).
HRMS: m/e calcd for C40H50N4NaO8S2 (M + Na), 801.2968;
found, 801.3013.
N-(2-(N′,N′-Dip h en ylh yd r a zon o)eth yl)-N-(4-h yd r oxy-4-
p h en ylb u t yl)-4-m et h ylb en zen esu lfon a m id e (31). To a
solution of the Cp2TiCl2 (100 mg, 0.40 mmol) in THF (1.5 mL)
under argon was added a 2.0 M solution of isopropylmagne-
sium bromide in THF (170 µL, 0.40 mmol) at 20 °C. A 1.0 M
solution of phenylmagnesium bromide in THF (450 µL, 0.40
mmol) was added to the resulting green solution and the
reaction mixture was stirred for 30 min. The carbonylhydra-
zone 16 (60 mg, 0.13 mmol) in THF (0.5 mL) was then added
and the reaction mixture was left stirring overnight. TLC
Ack n ow led gm en t. We are indebted to the Danish
National Science Foundation (THOR Program), the
University of Aarhus, the Carlsberg Foundation, and
the Lundbeck Foundation for generous financial sup-
port. We thank Prof. Kim Daasbjerg for stimulating
discussions.
Su p p or tin g In for m a tion Ava ila ble: Single-crystal X-ray
structure and crystallographic data of the hexahydroazepine
26, experimental procedures and characterization of com-
pounds 2, 5, 6, 8-11, and 15-20, and 1H NMR spectra for
compounds 5, 6, 8-11, 15-24, and 26-31. This material is
available free of charge via the Internet at http//pubs.acs.org.
J O000538N