Synthesis of Medium- and Large-Size Rings by Intramolecular Nitrile Oxide Dimerization: An Efficient C-C Bond-Forming Ring-Closing Reaction

Full text of DOI:10.1021/jo991111n

8428
J . Org. Chem. 1999, 64, 8428-8431
as key reactions for the synthesis of complex natural
products having highly functionalized carbocyclic back-
bone. These recent achievements emphasize the synthetic
challenge of ring-closing reactions by using mild reaction
conditions and reagents compatible with a large panel
of functionality.
Syn th esis of Med iu m - a n d La r ge-Size Rin gs
by In tr a m olecu la r Nitr ile Oxid e
Dim er iza tion : An Efficien t C-C
Bon d -F or m in g Rin g-Closin g Rea ction
N. Maugein, A. Wagner, and C. Mioskowski*
Herein, we report a C-C bond ring-closing reaction for
the synthesis of carbocyclic structures using dialdehydes
as starting material. The ring closures proceed via
intramolecular dimerization of nitrile oxides resulting in
the formation of furoxan moieties that can easily be
converted into 1,2-dioximes, 1,2-diketones, 1,2-diols, or
1,2-diamines.10
Laboratoire de Synthe`se Bioorganique,
Universite´ Louis Pasteur de Strasbourg,
Unite´ associe´e au CNRS, UMR 7514,
Faculte´ de Pharmacie, 74 route du Rhin,
F-67400 Illkirch Graffenstaden, France
Received J uly 12, 1999
Nitrile oxides are reactive intermediates, generated in
situ, and are generally involved in cycloaddition reactions
in the presence of various dipolarophiles. Intramolecular
nitrile oxide cycloaddition (INOC) has proved to be an
efficient method to prepare small and large functionalized
carbocycles. This reaction finds interesting synthetic
applications but suffers from a few limitations. Under
unfavored conditions, e.g., with hindered dipolarophiles
or in the absence of dipolarophiles, dimerization of nitrile
oxides into furoxan was observed predominantly.11
The formation of medium- and large-size ring systems
remains a permanent challenge for organic chemists as
they are widespread, ranging from naturally occurring
compounds to macrocyclic synthetic receptors or ligands.
The synthesis depends essentially on the intrinsic carbo-
or heterocyclic nature of their framework.1 As for het-
erocyclic structures, the ring closures are usually per-
formed by carbon-heteroatom bond formation, e.g., ester,
amide, or ether bonds. These approaches are well docu-
mented, and the procedures are optimized especially for
macrocyclic natural products.2 In contrast, for carbocyclic
molecules, particularly for highly functionalized struc-
tures, synthesis involving intramolecular carbon-carbon
bond formation is often troublesome.
While cycloaddition reactions involving nitrile oxides
and dipolarophiles were extensively studied and used in
organic synthesis, nitrile oxide dimerization reactions
leading to furoxan formation remain a reaction of poor
synthetic interests. To our knowledge, intramolecular bis-
(nitrile oxide) dimerization has been reported only once
by Marx et al. in 1977 for the preparation of a five-
membered-ring carbocycle, an intermediate for biotin
synthesis.12 Despite this early result, the potential of this
reaction was never exploited to develop a synthetically
useful method for the preparation of carbocyclic com-
pounds.
A number of reactions such as Wurtz coupling,3 alkyl-
ation of malonic esters,4 acyloin condensation,5 and
samarium diiodide-promoted reaction6 were applied to
synthesize carbocyclic skeletons. However, low yields are
often obtained for strained medium-ring systems or for
highly functionalized structures mainly for compatibility
reasons between the reagent and the functional groups
present on the molecules.
Recently, the reductive coupling of aldehydes using
low-valent titanium,7 intramolecular olefin metathesis,8
and other metal-mediated coupling reactions9 were used
In our initial study, we evaluated the importance of
the stability of the nitrile oxide intermediates on the
outcome of the intramolecular dimerization reaction. The
stability of nitrile oxides is conditioned by their substitu-
tion at the R-position. Alkyl nitrile oxides are very
reactive while vinylic, arylic, or sterically hindered R,R-
disubstituted nitrile oxides are described to be moder-
ately stable at room temperature and reactive only when
heated.13 We prepared a set of three dialdehyde precur-
* To whom correspondence should be addressed. Phone: (33) 3 88
67 68 63. Fax: (33) 3 88 67 88 91. E-mail: mioskow@aspirine.u-
strasbg.fr; wagner@bioorga.u-strasbg.fr.
(1) Breitenbach, J .; Boosfeld, J .; Vo¨gtle F. In Comprehensive Su-
pramolecular Chemistry; Lehn, J . M., Ed.; Pergamon: New York, 1996;
Vol. 2, pp 29-67. Dietrich, B.; Viout, P.; Lehn, J . M. Macrocyclic
Chemistry; VCH: Weinheim, 1993.
(2) Anastassiou, A. G. In Comprehensive Heterocyclic Chemistry;
Katrizky, A. R., Rees, C. W., Ed.; Pergamon Press: New York, 1984;
Vol. 7, pp 709-730. Hamilton, A. D. In Comprehensive Heterocyclic
Chemistry; Katrizky, A. R., Rees, C. W., Ed.; Pergamon Press: New
York, 1984; Vol. 7, pp 731-762.
(3) Vo¨gtle, F.; Brodesser, G.; Nieger, M.; Rissanen, K. Rec. Trav.
Chim. Pays-Bas 1993, 112, 325-329. Billington, D. C. In Comprehen-
sive Organic Synthesis; Trost B. M. Ed.; Pergamon Press: New York,
1991; Vol. 3, pp 413-433.
(4) Deslongchamps, P.; Lamothe, S.; Lin, H. S. Can. J . Chem. 1987,
65, 1298-1307. Merz, T.; Wirtz, H.; Vo¨gtle, F. Angew. Chem., Int. Ed.
Engl. 1986, 25, 567-569.
(5) Cram, D. J .; Gaston, L. K. J . Am. Chem. Soc. 1960, 82, 6386-
6389.
(6) Takahashi, S.; Mori, N. J . Chem. Soc., Perkin Trans. 1 1991,
2029-2032.
(7) Lenoir, D. Synthesis 1989, 883-896. McMurry, J . E. Acc. Chem.
Res. 1983, 16, 405-411.
(9) Nicolaou, K. C.; He, Y.; Roschangar, F.; King, N. P.; Vourloumis,
D.; Li, T. Angew. Chem., Int. Ed. Engl. 1998, 37, 84-87. White, J . D.;
Tiller, T.; Ohba, Y.; Porter, W. J .; J ackson, R. W.; Wang, S.; Hansel-
mann, R. Chem. Commun. 1998, 79-80. Nicolaou, K. C.; Chu, X. J .;
Ramanjulu, J . M.; Natarajan, S.; Bra¨se, S.; Ru¨bsam, F.; Boddy, C. N.
C. Angew. Chem., Int. Ed. Engl. 1997, 36, 1539-1540.
(10) Armani, V.; Dell′Erba, C.; Marino, N.; Petrillo, G.; Tavani, C.
Tetrahedron, 1997, 53, 1751-1758. Gasco, A.; Boulton, A. J . Adv.
Heterocycl. Chem. 1981, 29, 251-339. Ackrell, J .; Alaf-ur-Rahman, M.;
Boulton, A. J .; Brown, R. C. J . Chem. Soc., Perkin Trans. 1 1972, 1587-
1594.
(11) Bianchi, G.; Gandolfi, R.; Gru¨nanger, P. In The Chemistry of
Functional Groups; Patai, S., Rappoport, Z., Eds.; J ohn Wiley & Sons
Ltd.: New York, 1983; pp 737-799. Grundmann, C. Synthesis 1970,
344-359. Huisgen, R. Proc. Chem. Soc. 1961, 357-369. Mukaiyama,
T.; Hoshino, T. J . Am. Chem. Soc. 1960, 82, 5339-5342.
(12) Marx, M.; Marli, F.; Reisdorff, J . H. Sandmeier, S.; Clark, S. J .
Am. Chem. Soc. 1977, 99, 6754
(13) Grundmann, C.; Dean, J . M. J . Org. Chem. 1965, 30, 2809-
2811. Grundmann, C.; Gru¨nanger, P. The Nitrile Oxides Springer-
Verlag: Berlin 1971. Barrow, S. J .; Easton, C. J .; Savage, G. P.;
Simpson, G. W. Tetrahedron Lett. 1997, 38, 2175-2178.
(8) Nicolaou, K. C.; Sarabia, F.; Ninkovick, S.; Ray, M.; Finlay, V.;
Boddy, C. N. C. Angew. Chem., Int. Ed. Engl. 1998, 37, 81-84.
Fu¨rstner, A.; Koch, D.; Langemann, K.; Leitner, W.; Six, C. Angew.
Chem., Int. Ed. Engl. 1997, 36, 2466-2469. Nicolaou, K. C.; He, Y.;
Vourloumis, D.; Vallberg, H.; Roschangar, F.; Sarabia, F.; Ninkovic,
S.; Yang, Z.; Trujillo, J . I. J . Am. Chem. Soc. 1997, 119, 7960-7973.
10.1021/jo991111n CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/29/1999

Notes
J . Org. Chem., Vol. 64, No. 22, 1999 8429
Sch em e 1
Sch em e 2
Sch em e 3
Sch em e 4
Alkyl oxime 3 was slowly added to a mixture of tert-
butyl nitrile oxide and sodium hypochlorite at room
temperature. Analysis of the reaction mixture showed the
presence of only two of the three possible combinations
of dimers. Furoxan, formed by dimerization of the reac-
tive nitrile oxide 2a , was predominant, while the one
arising from cross coupling was minor. As expected,
furoxan formed by dimerization of tert-butyl nitrile oxide
was not detectable in the crude. At low temperature, the
formation of the nitrile oxides by reaction of oximes and
sodium hypochlorite became very slow. Therefore, for
further studies of cross coupling at low temperature the
reactive nitrile oxide was generated in situ by treatment
of the primary nitro derivative 4 with DAST14 (Scheme
3).
sors of all possible combinations of stable/reactive nitrile
oxides (Scheme 1).
Dialdehyde 1a was quantitatively transformed into the
corresponding dioxime. This upon treatment with sodium
hypochlorite led to a stable bis(nitrile oxide) that could
be kept for hours under nitrogen at room temperature
or for an extended period of time at -20 °C. However, at
-20 °C, after several weeks we noticed that bis(nitrile
oxide) tended to oligomerize slowly into polyfuroxans.
Therefore, for each experiment the bis(nitrile oxide) was
freshly prepared.
The cyclization step was performed by the slow addi-
tion of a solution of bis(nitrile oxide) (sequence 1a ) to
refluxing toluene. After the addition, the mixture was
refluxed for a further 2 h, concentrated under vacuum,
and purified by silica gel chromatography. Thus, in three
steps/one purification process the eight-membered car-
bocyclic compound was isolated in 84% overall yield.
Two conditions were tested, (i) addition of DAST to a
mixture of tert-butyl nitrile oxide and primary nitroalkyl
at -30 °C and (ii) addition of the nitro compound to a
mixture of tert-butyl nitrile oxide and DAST. In both
cases, similar results were obtained. Symmetrical furox-
an 2a was formed predominantly in 90% yield by the
dimerization of reactive nitrile oxide. The presence of 5%
of furoxan 2b arising from the cross condensation of the
stable and the reactive nitrile oxide could also be
detected. These results show clearly that the formation
of unsymmetrical furoxans by intermolecular condensa-
tion of a reactive and a stable nitrile oxide is problematic
because hindered nitrile oxides appear to be poor dipo-
larophiles. At low temperature they are not reactive,
while at high temperature they undergo dimerization.
For substrates 1b and 1c, one or both aldehydes led
to reactive alkyl nitrile oxides preventing all attempts
from the isolation of the corresponding bis(nitrile oxide)
intermediates. To favor the cyclization over the oligo-
merization process, we tried to generate the bis(nitrile
oxides) under high dilution conditions, and cyclization
was carried out at high temperature. Only complex
mixtures could be obtained in which traces of cyclic
products could be detected by NMR. This observation
seems to prove that both substrates 1b and 1c react
intermolecularly as soon as they are formed. This limita-
tion reflects the basic problem with intramolecular
cyclizationsto avoid competing intermolecular reactionss
which is most difficult to achieve when the intermolecular
processes are favored by interaction of more reactive
groups.
In a last attempt, we prepared a bifunctional compound
5 bearing both the stable nitrile oxide and the primary
nitroalkyl function (Scheme 4).
Addition of this bifunctional compound 5 to a solution
of DAST at -30 °C or at 30 °C or addition of DAST to a
dilute solution of the bifunctional compound resulted in
However, to evaluate further the scope and limitations
of cross condensations between stable and reactive nitrile
oxides, we generated reactive, unstable nitrile oxides in
the presence of a stable one (Scheme 2).
(14) Maugein, N.; Wagner, A.; Mioskowski, C. Tetrahedron Lett.
1997, 38, 1547-11550.

8430 J . Org. Chem., Vol. 64, No. 22, 1999
Notes
Ta ble 1
However, the method revealed to be restricted to stable
bis(nitrile oxides). Considering the wide panel of substit-
uents that are able to stabilize or to temporarily stabilize
the nitrile oxides, e.g., dithioacetals, the reaction appears
nevertheless very useful. In addition, the furoxan moiety
introduced during the cyclization process allows an easy
access to interesting 1,2-functionalities present in many
macrocyclic structures. Therefore, this new C-C bond-
forming ring-closing reaction is of synthetic interest and
is complementary to other existing methodologies.
Exp er im en ta l P a r t
Gen er a l Meth od s. Thin-layer chromatograms (TLC) were
performed on precoated silica gel 60 F254 plates (Merck, 0.25
mm.). Flash chromatography separations were performed on
silica gel (Merck 60, 230-400 mesh). Reaction vessels were
flame-dried and were allowed to cool under an inert atmosphere.
Solvents were freshly distilled over appropriate desiccants prior
to use.
Tr a n sfor m a tion of Dia ld eh yd e in to Bis(n itr ile oxid e).
To a solution of dialdehyde (sequence 1a) (1.10 g, 5.55 mmol) in
30 mL of methylene chloride at 0 °C were added successively
hydroxylamine hydrochloride (1.91 g, 27.7 mmol), sodium car-
bonate (1.52 g, 14.42 mmol), and 10 mL of water. After the
mixture was stirred for 24 h at room temperature, the organic
layer was separated and the aqueous layer was extracted with
methylene chloride (3 × 30 mL). The combined organic extracts
were dried over sodium sulfate and concentrated under reduced
pressure to afford the dioxime (1.20 g, 5.26 mmol) in 95% yield.
the formation of complex mixtures in which mass spectral
analysis showed a mixture of dimers, trimers, and mainly
oligomers.
The crude dioxime (200 mg, 0.88 mmol) was then dissolved
in 8 mL of methylene chloride followed by the dropwise addition
of a solution of sodium hypochorite (2.5 M, 8 mL, 5 equiv) at 0
°C. After 2 h of vigorous stirring at 0 °C, the organic layer was
separated and the aqueous layer was extracted with methylene
chloride (3 × 20 mL). The combined organic phase was washed
with brine (40 mL), dried over sodium sulfate, and concentrated
at 0 °C under reduced pressure to afford the bis(nitrile oxide)
(196 mg, 0.88 mmol)) in quantitative yield.
Alkyl nitrile oxides revealed to be highly reactive
intermediates, both as dipoles and as dipolarophiles
favoring homo dimerization over cross-condensation with
stable nitrile oxides. Thus, our studies demonstrate
clearly that the prerequisite to favor cyclization over
oligomerization is to utilize dialdehydes that lead to
stable bis(nitrile oxides) as substrates. After isolation,
they can efficiently be subjected to cyclization. In Table
1 are the summarized results obtained by different
combinations of stable nitrile oxides leading to carbo- or
to heterocyclic structures of different sizes. In all cases,
the cyclic compounds were obtained as the major prod-
ucts. Dimerization of symmetrical (entries 1 and 2) as
well as nonsymmetrical dialdehydes (entries 3 and 4)
gave the cyclic products in good yields.
During the dimerization process, one of the nitrile
oxides acts as a dipole whereas the other acts as a
dipolarophile. Thus, in the case of nonsymmetrical dial-
dehydes (entries 1 and 2), a mixture of the two regioiso-
meric furoxans was obtained without significant selec-
tivity.
Interestingly, we observed that the ring size has little
influence on the yield of the reaction. Medium strained
ring structures such as 8-, 10-, and 12-membered rings
were consistently obtained in good yields, 82% (entry 1),
80% (entry 2), and 86% (entry 3), respectively. A larger
19-membered ring was also obtained in 85% yield (entry
4). It is worth noting that in contrary to many other
reactions, the steric hindrance in this reaction is not a
drawback.
In tr a m olecu la r Cycliza tion . The cyclization step was car-
ried out by adding a solution of bis(nitrile oxide) (196 mg, 0.88
mmol) in anhydrous toluene (20 mL) to refluxing anhydrous
toluene (40 mL) over a period of 4.5 h. After an additional 2 h
under reflux, the reaction mixture was concentrated under
reduced pressure and the product was purified by column
chromatography on silica gel using ether/hexane 9/1. The cyclic
product (sequence 1a) was obtained in 84% yield (164 mg, 0.64
mmol) as white crystals.
4,4,9,9-Tetr a m eth ylp er h yd r ocycloocta [c]fu r a za n 1-ox-
id e: Rf ) 0.40 (hexane/ether; 9/1); colorless oil; 1H NMR (300
MHz, CDCl3) δ 1.38 (6H, s), 1.51 (6H, s), 1,30-2.05 (8H, m); 13
C
NMR (75 MHz, CDCl3) δ 23.8, 23.6, 24.7, 32.1, 34.2, 37.0, 37.8,
39.6, 122.5, 163.0; IR 1562 (CdN-O) 1455 (CdN+(O-)-O, 1376
(N-O) cm-1; MS (CI/NH3) m/z 242 (M + NH4) +. Anal. Calcd for
C12H20N2O2 (224.1): C, 64.30; H, 8.91; Found: C, 64.72; H, 9.19.
9-Me t h oxy-4,4,9-t r im e t h ylp e r h yd r ocyclooct a [c]fu r a -
za n 1-oxid e: Rf ) 0.68 (hexane/ethyl acetate 7/3); colorless oil,
isolated as a 2/1 mixture of two regioisomers; 1H NMR (300 MHz,
CDCl3) δ 1.26, 1.32, 1.44, 1.49, 1.68, 1.77 (9H, 6xs), 1.20-2.10
(8H, m), 3.13, 3.18 (3H, 2 × s); 13C NMR (75 MHz, CDCl3) δ
20.0, 22.9, 23.5, 23.7, 23.8, 24.3, 24.6, 25.4, 27.6, 32.8, 34.0, 36.4,
36.6, 37.0, 41.3, 41.5, 51.0, 51.7, 76.7, 77.3, 115.9, 122.6, 158.7,
165.9; IR 1573 (CdN-O), 1458 (CdN+(O-)-O, 1364 (N-O)
cm-1; MS (CI/NH3) m/z 258 (M + NH4) +. Anal. Calcd for
C12H20N2O3(240.1): C, 60.02; H, 8.32. Found: C, 59.95; H, 8.22.
4,4-Dim et h yl-4,5,6,7-t et r a h yd r o-8H-9-oxa b en zo[b]fu r a -
za n cyclod ecen e 1-oxid e: Rf ) 0.64 (hexane/ethyl acetate 8/2);
white solid, isolated as a 1/1 mixture of two regioisomers; mp )
184 °C-186 °C; 1H NMR (300 MHz, CDCl3) δ 1.15-1.40 (8H, m),
1.50-1.65 (2H, m), 1.70-1.85 (2H, m), 4.29 (2H, t, J ) 6 Hz),
7.09-2.27 (2H, m), 7.29-7.32 (1H, m), 7.48-7.54 (1H, m); 13C
NMR (75 MHz, CDCl3) δ 19.5, 20.4, 26.1, 26.3, 36.9, 37.2, 39.5,
69.6, 69.8, 116.0, 116.8, 120.4, 121.9, 122.6, 131.3, 132.1, 155.5,
In summary, we describe a useful procedure for the
preparation of carbocyclic structures. Thermal dimeriza-
tion of stable bis(nitrile oxides) allow the ring closure of
a large range of sizes in fairly good yields. In addition,
the mild reaction conditions required for the reaction are
compatible with a large variety of functional groups.

Notes
J . Org. Chem., Vol. 64, No. 22, 1999 8431
158.7; IR 1577 (CdN-O), 1458 (CdN+(O-)-O, 1385 (N-O)
cm-1; MS (CI/NH3) m/z 292 (M + NH4) +. Anal. Calcd for
NMR (300 MHz, CDCl3) δ 3.53 (2H, t, J ) 5 Hz), 3.6-3.7 (8H,
m), 3.7-3.8 (4H, m), 4.17 (2H, t, J ) 5 Hz), 6.85-6.92 (2H, m),
6.95-7.10 (3H, m), 7.35-7.50 (3H, m); 13C NMR (75 MHz,
CDCl3) δ 68.4, 68.8, 69.1, 69.9, 70.6, 70.9, 71.0, 71.3, 112.7, 113.5,
114.0, 120.8, 121.1, 129.7, 130.6, 131.8, 132.4, 156.4, 157.1; IR
1601 (CdN-O), 1462 (CdN+(O-)-O cm-1; MS (CI/NH3) m/z 446
(M + NH4) +. Anal. Calcd for C22H24N2O7 (428.2): C, 61.70; H,
5.60. Found: C, 61.82; H, 5.71.
C
15H18N2O3(274.2): C, 65.71; H, 6.56. Found: C, 65.82; H, 6.50.
9,10,11,12-Tet r a h yd r o-8,13-d ioxa d ib en zo[gk ]fu r a za n o-
[3,4-i]cyclod od ecen e 1-oxid e: Rf ) 0.50 (hexane/ethyl acetate
8/2); white solid; mp ) 192 °C; 1H NMR (300 MHz, CDCl3) δ
1.90 (4H, m), 3.91 (2H, m), 4.23 (2H, m), 6.80-7.05 (5H, m) 7.30-
7.44 (3H, m); 13C NMR (75 MHz, CDCl3) δ 25.8, 26.6, 68.9, 69.9,
109.0, 109.2, 113.3, 113.5, 120.7, 120.8, 130.5, 130.9, 156.4, 156.9;
IR 1601 (CdN-O), 1462 (CdN+(O-)-O cm-1; MS (CI/NH3) m/z
342 (M + NH4) +. Anal. Calcd for C18H16N2O4(324.2): C, 66.68;
H, 4.93. Found: C, 66.56; H, 5.15.
Ack n ow led gm en t. We thank Rhoˆne-Poulenc Rorer
for financial support of this work through a BDI grant
to Nathalie Maugein and Alain Valleix for mass spec-
troscopic analysis.
9,10,12,13,15,16,18,19-Octa h yd r o-8,11,14,17,20-p en ta oxa -
d iben zo[n ,r ]fu r a za n o[4,3-p]cyclon on a d ecen e 1-oxid e: Rf )
0.64 (hexane/ethyl acetate 8/2); white solid; mp ) 112 °C; 1H
J O991111N