Camphor-derived, chelating auxiliaries for the highly diastereoselective intermolecular Pauson-Khand reaction: Experimental and computational studies

J . Org. Chem. 1998, 63, 7037-7052

7037

Ca m p h or -Der ived , Ch ela tin g Au xilia r ies for th e High ly

Dia ster eoselective In ter m olecu la r P a u son -Kh a n d Rea ction :

Exp er im en ta l a n d Com p u ta tion a l Stu d ies

Xavier Verdaguer,^†J ordi Va´zquez,^†Gerard Fuster,^†Vania Bernardes-Ge´nisson,^‡

Andrew E. Greene,^‡Albert Moyano,^†Miquel A. Perica`s,*^,†and Antoni Riera*^,†

Unitat de Recerca en S´ıntesi Asime`trica, Departament de Qu´ımica Orga`nica,

Universitat de Barcelona, Martı´ i Franque´s 1-11, 08028 Barcelona, Spain, and LEDSS,

Chimie Recherche, Universite´ J oseph Fourier, BP 53X, 38041 Grenoble, France

Received May 26, 1998

A family of enantiomerically pure (2R)-10-(alkylthio)isoborneols [methylthio (1), neopentylthio (2),

phenylthio (3)], specifically designed as chiral auxiliaries suitable for chirality transfer to cobalt in

Pauson-Khand reactions, has been synthesized. The dicobalt hexacarbonyl complexes of the

alkoxyacetylenes derived from these alcohols (10a -12a ) can be converted to the rather stable,

internally chelated, pentacarbonyl complexes 10b-12b by treatment with NMO. The intermo-

lecular Pauson-Khand reactions of 10b-12b with strained olefins take place with synthetically

useful rates at low temperatures (down to -20 °C), with high yields and diastereoselectivities:

norbornene (77%; 92:8), norbornadiene (82%; 96:4), bicyclo[3.2.0]hept-6-ene (91%; 93:7). The major

diastereomer of the adduct of 10b with norbornadiene, 14, has been used as the starting point for

a synthesis of (S)-(-)-4-alkyl-2-cyclopentenones through a sequence consisting of completely

diastereoselective conjugate addition, reductive cleavage with recovery (>95%) of the chiral auxiliary,

and retro Diels-Alder reaction. The stereochemical course of the reaction of 10b with norbornadiene

has been analyzed and rationalized by theoretical means by using a combined semiempirical [PM3-

(tm)]/density functional theory [VWN-Perdew-Wang 91] approach.

In tr od u ction

devoted to the development of practical enantioselective

versions of the reaction. Two main approaches have been

followed with this aim: The use of chiral auxiliaries

directly bound to either of the reacting fragments⁵and

the intermediate generation of complexes possessing a

The metal-mediated cyclocondensations¹between an

alkyne, an alkene, and a CO fragment probably represent

the most versatile methodology for the construction of

cyclopentenone skeletons (eq 1).²

(3) Reviews on the Pauson-Khand reaction: (a) Pauson, P. L.;

Khand, I. U. Ann. N. Y. Acad. Sci. 1977, 295, 2-14. (b) Pauson, P. L.

Tetrahedron 1985, 41, 5855-5860. (c) Pauson, P. L. In Organometallics

in Organic Synthesis. Aspects of a Modern Interdisciplinary Field; de

Meijere, A. tom Dieck, H. Eds.; Springer: Berlin, 1988; pp 233-246.

(d) Harrington, P. J . Transition Metals in Total Synthesis; Wiley: New

York, 1990; pp 259-301. (e) Schore, N. E. Org. React. 1991, 40, 1-90.

(f) Schore, N. E. In Comprehensive Organic Synthesis; Trost, B. M.,

Ed.; Pergamon Press: Oxford, 1991; Vol. 5, pp 1037-1064.

(4) For some recent examples, see: (a) J amison, T. F.; Shambayati,

S.; Crowe, W. E.; Schreiber, S. L. J . Am. Chem. Soc. 1997, 119, 4353-

4363. (b) Thommen, M.; Keese, R. Synlett 1997, 231-240. (c) Deol-

iveira, C. M.; Ferracini, V. L.; Foglio, M. A.; de Meijere, A.; Marsaioli,

A. J . Tetrahedron: Asymmetry 1997, 8, 1833-1839. (d) Donkervoort,

J . G.; Gordon, A. R.; J ohnstone, C.; Kerr, W. J .; Lange, U. Tetrahedron

1996, 52, 7391-7420. (e) Paquette, L. A.; Borrelly, S. J . Org. Chem.

1995, 60, 6912-6921. (f) Yoo, S. E.; Lee, S. H. J . Org. Chem. 1994, 59,

6968-6972. (g) Krafft, M. E.; Wright, C. Tetrahedron Lett. 1992, 33,

151-152. (h) Rowley, E. G.; Schore, N. E. J . Org. Chem. 1992, 57,

6853-6861. (i) Takano, S.; Inomata, K.; Ogasawara, K. Chem Lett.

1992, 443-446.

(5) (a) Poch, M.; Valent´ı, E.; Moyano, A.; Perica`s, M. A.; Castro

Tetrahedron Lett. 1990, 31, 7505-8. (b) Castro, J .; So¨rensen, H.; Riera,

A.; Morin, C.; Moyano, A.; Perica`s, M. A.; Greene, A. E. J . Am. Chem.

Soc. 1990, 112, 9388-9389. (c) Verdaguer, X.; Moyano, A.; Perica`s,

M. A.; Riera, A.; Greene, A. E.; Piniella, J . F.; Alvarez-Larena, A. J .

Organomet. Chem. 1992, 433, 305-310. (d) Castro, J .; Moyano, A.;

Perica`s, M. A.; Riera, A.; Greene, A. E. Tetrahedron: Asymmetry 1994,

5, 307-310. (e) Bernardes, V.; Verdaguer, X.; Kardos, N.; Riera, A.;

Moyano, A.; Perica`s, M. A.; Greene, A. E. Tetrahedron Lett. 1994, 35,

575-578. (f) Stolle, A.; Becker, H.; Salau¨n, J .; de Meijere, A. Tetra-

hedron Lett. 1994, 35, 3521-3524. (g) Fonquerna, S.; Moyano, A.;

Perica`s, M. A.; Riera, A. Tetrahedron 1995, 51, 4239-4254. (h) Balsells,

J .; Moyano, A.; Perica`s, M. A.; Riera, A. Tetrahedron: Asymmetry 1997,

8, 1575-1580. (i) Montenegro, E.; Poch, M.; Moyano, A.; Perica`s, M.

A.; Riera, A. Tetrahedron 1997, 53, 8651-8664. (j) Fonquerna, S.;

Moyano, A.; Perica`s, M. A.; Riera, A. J . Am. Chem. Soc. 1997, 119,

10225-10226.

Among these methods, the Pauson-Khand reaction

[M_x(CO)_n) Co₂(CO)₈]³stands out because of its experi-

mental simplicity, functional group compatibility, and

predictable regio- and stereochemical outcome and has

consequently found widespread use in total synthesis.⁴

It is thus not surprising that increasing efforts are being

^†Universitat de Barcelona.

^‡Universite´ J oseph Fourier.

(1) For representative examples of metal-mediated cyclocondensa-

tions leading to five-membered carbocycles, see the following. Ti: (a)

Hicks, F. A.; Buchwald, S. L. J . Am. Chem. Soc. 1996, 118, 11688-

11689. (b) Hicks, F. A.; Kablaoui, N. M.; Buchwald, S. L. J . Am. Chem.

Soc. 1996, 118, 9450-9451. Zr: (c) Agnel, G.; Negishi, E.-I. J . Am.

Chem. Soc. 1991, 113, 7424-7426. Mo: (d) J eong, N.; Lee, S. J .; Lee,

B. Y.; Chung, Y. K. Tetrahedron Lett. 1993, 34, 4027-4030. (e) Kent,

J . L.; Wan, H. H.; Brummond, K. M. Tetrahedron Lett. 1995, 36, 2407-

2410. W: (f) Hoye, T. R.; Suriano, J . A. J . Am. Chem. Soc. 1993, 115,

1154-1156. Fe: (g) Pearson, A. J .; Dubbert, R. A. Organometallics

1994, 13, 1656-1661. Ru: (h) Kondo, T.; Suzuki, N.; Okada, T.;

Mitsudo, T. J . Am. Chem. Soc. 1997, 119, 6187-6188. Ni: (i) Tamao,

K.; Kobayashi,. K.; Ito, Y. Synlett 1992, 539-546. (j) Page´s, L.; Llebaria,

A.; Camps, F.; Molins, E.; Miravitlles, C.; Moreto´, J . M. J . Am. Chem.

Soc. 1992, 114, 10449-10461. (k) Zhang, M. H.; Buchwald, S. L. J .

Org. Chem. 1996, 61, 4498-4499.

(2) For reviews on cyclopentannelation reactions, see: (a) Trost, B.

M. Chem. Soc. Rev. 1982, 11, 141-170. (b) Ramaiah, M. Synthesis

1984, 529-560. (c) Paquette, L. A.; Doherty, A. M. Polyquinane

Chemistry; Springer-Verlag: Berlin, 1987. (d) Hudlicky, T.; Price, J .

Chem. Rev. 1989, 89, 1467-1486.

Published on Web 09/11/1998

7038 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

disymmetric C₂Co₂core.⁶Despite the conceptual appeal

inherent to the second approach, only the first one has

found application in the synthesis of complex molecules.^5b,7

This is probably due to practical reasons; thus, the

preparation of diastereomerically pure (alkyne)Co₂-

(CO)₅L* complexes normally involves lengthy procedures

and/or tedious chromatographic separations.

During the past few years, we have systematically

explored the use of chiral alcohols as stereochemical

controllers in the Pauson-Khand reaction.^5a-eFor the

intramolecular process, the use of either trans-2-phenyl-

cyclohexanol (as an enol ether with trans configuration)^5b,d

or 3-(neopentyloxy)isoborneol (as an ynol ether)^5callows

the achievement of high levels of diastereoselectivity in

the process. On the other hand, a systematic scrutiny

of the usual arylcyclohexyl and camphor-based auxiliaries

has not allowed the identification of similarly efficient

controllers for the intermolecular process.^5e

F igu r e 1. Stereocontrol in Pauson-Khand reactions with

In an attempt to generate a satisfactory solution for

this more stereochemically demanding situation, and

bearing in mind the work by Krafft on the directed

Pauson-Khand reaction,⁸we started thinking about a

new type of chiral controller, specifically suited for the

Pauson-Khand cycloaddition, that could combine the

main advantages of the two approaches to stereocontrol

discussed earlier. A chiral auxiliary with chelating

capacity could display a dual role in the reaction: It could

allow an efficient transfer of chirality to the C₂Co₂cluster

and could secure the diastereoselectivity of the process

by directing the reaction toward one of the two diaste-

reotopic cobalt atoms (Figure 1).

Previous experience with alkoxyacetylenes⁹and the

ready availability of their dicobalt hexacarbonyl com-

plexes¹⁰suggested oxygen as the linking element (X in

Figure 1). On the other hand, the specification of sulfur

as the element planned for interaction with cobalt offered

significant advantages over other possible candidates

such as phosphorus. First, the ability of sulfur to

stabilize coordinatively unsaturated cobalt species had

been well established by Krafft.⁸Second, the rather

labile nature of the Co-S dative bond¹¹should greatly

chelating auxiliaries.

facilitate the dissociative equilibrium B (Figure 1),

thus contributing to a rate enhancement in the overall

process.

We report in the present paper full details of the

successful use in Pauson-Khand chemistry of a family

of camphor-derived chiral auxiliaries 1-3 designed ac-

cording to the principles we have just discussed.¹²The

origin of the diastereoselectivity observed when using this

new family of chelating auxiliaries is analyzed and

rationalized by means of theoretical (semiempirical and

DFT) calculations on one of the actual systems involved

in the study.

Resu lts a n d Discu ssion

Ch ir a l Con tr oller s a n d Alk oxya cetylen e P r ecu r -

sor s. Simple modeling studies, confirmed by preliminary

experiments, suggested that 3-(alkylthio) alcohols pre-

sented an optimal distance between functional groups for

the desired chelation.¹³Although these alkylthio alcohols

could be generally available from enantiomerically pure

R,â-unsaturated ketones by conjugate thiol addition and

carbonyl reduction,¹⁴the requirements of both enantio-

and diastereomeric purity in our target molecules along

with the desire for structural rigidity greatly reduced the

scope of possible candidates.

In this context, we thought that (2R)-10-mercap-

toisoborneol (4), ultimately available in enantiomerically

pure form from inexpensive ^D-camphorsulfonic acid,¹⁵was

an ideal template for the planned intermolecular Pau-

son-Khand reactions since, by simply tuning the elec-

tronic and steric nature of the sulfide substituent, we

should be able to maximize the diastereoselectivity of the

process. Along with the known (2R)-10-(methylthio)-

isoborneol (1),¹⁶we undertook the synthesis of two new

(6) (a) Bladon, P.; Pauson, P. L.; Brunner, H.; Eder, R. J . Organomet.

Chem. 1988, 355, 449-454. (b) Brunner, H.; Niedernhuber, A. Tetra-

hedron: Asymmetry 1990, 1, 711-714. (c) Park, H. J .; Lee, B. Y.; Kang,

Y. K.; Chung, Y. K. Organometallics 1995, 14, 3104-3107. (d) Kerr,

W. J .; Kirk, G. G.; Middlemiss, D. Synlett 1995, 1085-1086. (e) Hay,

A. M.; Kerr, W. J .; Kirk, G. G.; Middlemiss, D. Organometallics 1995,

14, 4986-4988.

(7) (a) Bernardes, V.; Kann, N.; Riera, A.; Moyano, A.; Perica`s, M.

A.; Greene, A. E. J . Org. Chem. 1995, 60, 6670-6671. (b) Tormo, J .;

Verdaguer, X.; Moyano, A.; Perica`s, M. A.; Riera, A. Tetrahedron 1996,

52, 14021-14040. (c) Castro, J .; Moyano, A.; Perica`s, M. A.; Riera, A.;

Greene, A. E.; Alvarez-Larena, A.; Piniella, J . F. J . Org. Chem. 1996,

61, 9016-9020. (d) Tormo, J .; Moyano, A.; Perica`s, M. A.; Riera, A. J .

Org. Chem. 1997, 62, 4851-4856.

(8) (a) Krafft, M. E. J . Am. Chem. Soc. 1988, 110, 968-970. (b)

Krafft, M. E.; J uliano, C. A.; Scott, I. L.; Wright, C. J . Am. Chem. Soc.

1991, 113, 1693-703. (c) Krafft, M. E.; J uliano, C. A. J . Org. Chem.

1992, 57, 5106-5115. (d) Krafft, M. E.; Scott, I. L.; Romero, R. H.

Tetrahedron Lett. 1992, 33, 3829-3832. (e) Krafft, M. E.; Scott, I. L.;

Romero, R. H.; Feibelmann, S.; Vanpelt, C. E. J . Am. Chem. Soc. 1993,

115, 7199-7207.

(9) (a) Perica`s, M. A.; Serratosa, F.; Valent´ı, E. Tetrahedron 1987,

43, 2311-2316. (b) Moyano, A.; Charbonnier, F.; Greene, A. E. J . Org.

Chem. 1987, 52, 2919-2922.

(12) A preliminary account of this work has appeared: Verdaguer,

X.; Moyano, A.; Perica`s, M. A.; Riera, A.; Bernardes, V.; Greene, A. E.;

Alvarez-Larena, A.; Piniella, J . F. J . Am. Chem. Soc. 1994, 116, 2153-

2154.

(13) Preliminary studies performed with the dicobalt hexacarbonyl

complexes of the alkoxyacetylenes derived from trans-2-phenylthiocy-

clohexanol and trans-2-(tert-butylthio)cyclohexanol did not provide any

evidence of chelation (see the following text in the paper).

(14) See, for instance: Eliel, E. L.; Lynch, J . E.; Kume, F.; Frye, S.

V. Org. Synth. 1987, 65, 215-223.

(15) (a) Eliel, E. L.; Frazee, W. J . J . Org. Chem. 1979, 44, 3598-

3599. (b) DeLucchi, O.; Lucchini, V.; Marchiorio, C. J . Org. Chem. 1986,

51, 1457-1466.

(10) Bernardes, V.; Verdaguer, X.; Moyano, A.; Perica`s, M. A.; Riera,

A.; Greene, A. E. J . Organomet. Chem. 1994, 470, C12-C14.

(11) See, for instance: Murray, S. G.; Hartley, F. R. Chem. Rev.

1981, 81, 365-414. For a general discussion on thioethers as ligands,

see: Mu¨ller, A.; Diemann, E. In Comprehensive Coordination Chem-

istry; Wilkinson, G. Gillard, R. D., McLeverky, J . A., Eds.; Pergamon

Press: Oxford, 1989; Vol. 2 (Ligands), pp 551-558.

Experimental and Computational Studies

J . Org. Chem., Vol. 63, No. 20, 1998 7039

Sch em e 1^a

Sch em e 2^a

a

Key: (a) CHCldCCl₂, KH, THF; (b) n-BuLi, Et₂O, TMSCl; (c)

a

Key: (a) MeI, MeONa/MeOH; (b) (1) NaH/THF, (2) BrCH₂C-

Co₂(CO)₈, hexane; (d) K₂CO₃, MeOH.

(CH₃)₃, HMPA, ∆; (c) PhSNa, HMPA/THF, ∆; (d) LiAlH₄, Et₂O.

Ta ble 1. Syn th esis of Dicoba lt Hexa ca r bon yl

Com p lexes of Alk oxya cetylen es 1-3

chelating auxiliaries, (2R)-10-(neopentylthio)isoborneol

(2) and (2R)-10-(phenylthio)isoborneol (3) displaying

larger and electronically different substituents on sulfur.

Thus, (2R)-10-mercaptoisoborneol (4), prepared from

camphorsulfonyl chloride by reduction with lithium

aluminum hydride,¹⁵was submitted to regioselective

alkylation with methyl iodide and neopentyl bromide to

afford 1 and 2 in high yield, respectively (Scheme 1). For

the preparation of 10-(phenylthio)isoborneol (3), 10-

iodocamphor (5)¹⁷was first reacted with sodium thiophe-

nolate in the presence of HMPA to afford 10-(phenylthio)-

camphor (6) in 87% chemical yield. Reduction of ketone

6 with LiAlH₄in ether afforded a 10/1 mixture of exo and

endo alcohols from which the target molecule 3 could

easily be separated in 85% yield by standard column

chromatography.

R*OH

olefin (yield, %)

complex (yield, %)

1

2

3

7 (78)

8 (74)

9 (77)

10 (62)

11 (72)

12 (71)

Sch em e 3

With the chiral auxiliaries in hand, we proceeded to

test their efficiency in the intermolecular Pauson-Khand

reaction. As a first step, the corresponding alkoxyacety-

lenes and their dicobalt hexacarbonyl complexes had to

be prepared. Alkoxyacetylenes derived from chiral sec-

ondary alcohols are usually prepared by a one-pot

sequence starting from trichoroethylene,^9bwhile the

formation in high yield of their dicobalt hexacarbonyl

complexes requires the intermediate silylation of the

acetylenic hydrogen.¹⁰In the present instance, however,

we found that much better results are obtained if the

isolation of the alkoxyacetylenes is obviated and their

precursor (E)-alkoxydichloroolefins are used as synthetic

intermediates. To do that, the original procedures were

slightly modified as follows: The starting chiral alcohols

1-3 were deprotonated with KH and reacted with

trichloroetylene in THF to afford the (E)-alkoxydichlo-

roalkenes 7-9 in high yield (Scheme 2 and Table 1).

These are stable intermediates and can be stored for

several weeks at room temperature without noticeable

decomposition. From the dichloroolefins, a four-reaction

transformation consisting of hydrogen chloride elimina-

tion/transmetalation with n-BuLi in diethyl ether, sily-

lation with TMSCl, complexation (Co₂(CO)₈/hexane), and

deprotection (K₂CO₃/ MeOH) was performed in a single

synthetic operation leading to the target dicobalt hexac-

arbonyl complexes 10-12 in high overall yields.

soon made evident. Thin-layer chromatography of the

original red hexacarbonyldicobalt complexes revealed the

presence of a minor component, characterized by a less

intense dark-brown spot at a lower R_f. In the case of the

complex containing the 10-(methylthio)isoborneol auxil-

iary (10a ), for instance, when a hexane solution of the

red hexacarbonyl complex was heated at 50 °C under a

nitrogen stream (CO purge), the intensity of the lower

brown spot increased while that of the red upper one

decreased. Conversely, when the resulting mixture of

complexes was cooled back to room temperature under

a CO atmosphere, the brown compound reverted to the

original red hexacarbonyl complex 10a . This is consis-

tent with an equilibrium between a hexacarbonyl 10a

and a pentacarbonyl chelated form 10b of the corre-

sponding (alkoxyacetylene)dicobalt complex, as shown in

Scheme 3.¹⁸

¹H NMR analysis (C₆D₆, 25 °C) of 10a /10b equilibrium

mixtures (Figure 2) disclosed some interesting data.

Upon chelation, the acetylenic proton experienced an

upfield shift from its original δ 5.25 ppm in the parent

hexacarbonyl complex to δ 4.94 ppm in the chelated

species. This is in accordance with a substitution of an

electroacceptor CO group for a more donating sulfide

ligand on cobalt. Similar shifts in the ¹H NMR spectrum

have been reported in the literature for alkyne-Co₂(CO)₅-

(PR₃) clusters.¹⁹Perhaps even more revealing are the

Beh a vior of th e Dicoba lt Hexa ca r bon yl Com -

p lexes: Gen er a tion of In ter n a lly Ch ela ted Sp ecies.

The chelating ability of this new kind of auxiliary was

(17) Oae, S.; Togo, H. Bull. Chem. Soc. J pn. 1983, 56, 3802-3812.

(18) Parallel behavior has been observed by Krafft and co-workers

in the dicobalt hexacarbonyl complex of a bis-homopropargylic thio-

ether (see ref 8e).

(16) Furukawa, N.; Sugihara, Y.; Fujihara, H. J . Org. Chem. 1989,

54, 4222-4224.

(19) Dunn, J . A.; Pauson, P. L. J . Organomet. Chem. 1991, 419, 383-

389.

7040 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

Sch em e 4

Ta ble 2. In ter m olecu la r P a u son -Kh a n d Rea ction s

Usin g Ch ela tin g Au xilia r ies

1

F igu r e 2. Portions of the H NMR spectra (300 MHz, C₆D₆)

of 10a and 10b showing the main changes associated with the

internal chelation process.

For reactivity studies, a complete conversion of 10a

into 10b was clearly required. We could finally succeed

at this goal by treating 10a with N-methylmorpholine

N-oxide (NMO); thus, addition of 6 equiv of NMO to a

CH₂Cl₂solution of the hexacarbonyl complex 10a under

nitrogen resulted in the total formation of the chelated

species (Scheme 3). In fact, it is known that amine

oxides, like NMO²⁰or TMAO,²¹are efficient promoters

of the Pauson-Khand reaction. Most probably, these

reagents act by oxidizing a coordinated CO ligand to

CO₂,²²thus providing an empty coordination site for the

incoming olefin. In the present case, the irreversible

character of the CO oxidation prevents the regeneration

of the starting hexacarbonyl complex 10a , thus allowing

its complete conversion.

A similar behavior was observed in the 10-(neopen-

tylthio)isoborneol-containing complex 11a and in its

phenylthio analogue 12a , thus confirming the importance

of molecular architecture for the internal chelation pro-

cess.

P a u son -Kh a n d Cycliza tion s. Once we were ac-

quainted with how to control the equilibrium between

the hexacarbonyl (a ) and the chelated pentacarbonyl (b)

forms of complexes 10-12, we proceeded to test their

efficiency in the intermolecular Pauson-Khand reaction

with strained alkenes (Scheme 4). The results of this

study are collected in Table 2.

variations in chemical shift displayed by the methyl

(-SCH₃) and methylene (-CH₂SR) groups directly at-

tached to sulfur. In the hexacarbonyl complex 10a , the

singlet signal corresponding to the methyl group (-SCH₃)

appears at δ 1.84 ppm and at δ 1.60 ppm in 10b. The

methylene (-CH₂SR) AB system in 10a displays reso-

nances at δ 2.24 and 2.80 ppm, while the chelated

complex 10b shows the same signals at δ 2.00 and 2.75

ppm, respectively.¹⁸Further evidence supporting a che-

lation process could be found in the mass spectrum (CI-

NH₃) of 10a . Along with the peak corresponding to the

molecular ion of the parent hexacarbonyl complex plus

ammonia at m/e ) 528 (M⁺+ 18, 25), a more intense

signal m/e ) 483 (M⁺- CO + 1, 80) due to the loss of

one CO unit is found. It is to be noted that the mass

spectra of dicobalt hexacarbonyl complexes of alkoxy-

acetylenes derived from nonchelating alcohols, when

recorded under similar conditions, do not exhibit any

significant peak corresponding to CO loss.

Although two diastereomers can exist for 10b, depend-

ing on which of the diastereotopic cobalt atoms is involved

in the chelation process, the NMR observations discussed

above tend to indicate the highly predominant existence

of a single diastereomer.

While we could shift the equilibrium between 10a and

10b toward the chelated form (vide supra) by thermal

activation, we were unable to do so in a complete fashion.

A maximum 85:15 ratio of chelated and unchelated

complexes was achieved on heating the solution of the

parent hexacarbonyl complex at 50 °C for 1 h with CO

removal. Longer heating times did not lead to any

improvement in the 10a /10b ratio and were deletereous

for complex recovery. Most likely, complex decomposition

affords free carbon monoxide, which reacts with 10b to

give back 10a , thus preventing complete conversion.

Our first goal was to identify the effect on diastereo-

selectivity of preforming the chelated pentacarbonyl

species. Reaction conditions A, consisting of working in

hexane solution under a CO atmosphere and performing

(20) Shambayati, S.; Crowe, W. E.; Schreiber, S. L. Tetrahedron Lett.

1990, 31, 5289-5292.

(21) J eong, N.; Chung, Y. K.; Lee, B. Y.; Lee, S. H.; Yoo, S. Synlett

1991, 6, 204-206.

(22) (a) Albers, M. O.; Coville, N. J . Coord. Chem. Rev. 1984, 53,

227-259. (b) Shen, J .-K.; Gao, Y.-C.; Shi, Q.-Z.; Basolo, F. Organome-

tallics 1989, 8, 2144-2147.

Experimental and Computational Studies

J . Org. Chem., Vol. 63, No. 20, 1998 7041

Sch em e 5

the reaction by thermal activation, would ensure a

maximum concentration of a -type complexes. This does

not necessarily mean that b-type complexes do not

participate as intermediates in the reactions but probably

ensures that b-type complexes, if formed, will immedi-

ately react with the partner olefin without accumulation

in the reaction medium or stereochemical equilibration

between diastereomers. According to our expectations,

when complex 10a was treated with norbornene under

conditions A at room temperature, no 10b could be

observed by TLC of the reaction mixture. The corre-

sponding cycloadduct 13 was isolated after 4 h in 60%

yield and with rather low (40%) diastereomeric excess

(Table 2, entry 1).

As we will see later, these results are of great significance

since 14 has been used as the starting point for an

enantioselective synthesis of 4-substituted 2-cyclopen-

tenones. Likewise, reaction with bicyclo[3.2.0]hept-6-ene,

a slightly less reactive olefin, provided 15 in 91% yield

and 86% de (Table 2, entry 9). On the other hand, chiral

auxiliaries with more sterically demanding sulfide-

chelating fragments did not induce higher stereoselec-

tivities in the reaction. Cycloaddition of 11, bearing a

(neopentylthio)methyl moiety, with norbornadiene (Table

2, entry 10) afforded adduct 16 in 92% de, exactly the

same diastereoselectivity displayed by 10, but in signifi-

cantly lower yield. The Pauson-Khand reaction of 12,

bearing a (phenylthio)methyl chelating handle, shows a

slightly lower (86% de) diastereoselectivity (Table 2, entry

11).

These results clearly show that neither the steric bulk

nor the electronic nature of the group attached to sulfur²³

exerts a significant influence on the diastereoselectivity

of the reactions. With respect to yield, it is clear that

the smaller methyl group displays optimal characteris-

tics, thus defining 1 as the chiral controller of choice for

this chemistry. A distinct characteristic of all of the

chelating controllers is that they undergo the Pauson-

Khand reactions with strained olefins with synthetically

useful rates at the low temperature of -20 °C.

The absolute configuration of the major diastereomer

of 13 could be established by X-ray crystallography,¹²

following isolation by flash chromatography and crystal-

lization from cold pentane, as the one depicted in Scheme

5. The crystal structure revealed that the absolute

configuration of the newly formed stereogenic centers in

13 is 1R,2R,6R,7R. It also confirmed an exo fusion

between the cyclopentenone and norbornene moieties as

expected on the basis of the normal face selectivity of

Pauson-Khand reactions. Moreover, we were able to

correlate the configurations of the major diastereomers

of 13 and 14. Thus, when the nonconjugated double bond

in the major diastereomer of 14 was selectively hydro-

genated over Pd/C in ethanol (Scheme 5), the product

obtained was spectroscopically (¹H and ¹³C NMR) identi-

cal with the major diastereomer of 13.

Alternatively, reaction conditions B, consisting in

reacting the starting mixture of complexes with 6 equiv

of NMO in dichloromethane at room temperature, cooling

the mixture to a given temperature and then adding the

partner olefin, would ensure that the reaction takes place

through b-type complexes. In this way, addition of 6

equiv of NMO to a solution of 10a in CH₂Cl₂at room

temperature resulted, after 10 min, in the complete

formation of 10b, as monitored by TLC. When the

resulting solution was cooled to 0 °C and norbornene was

added at that temperature, 13 was obtained in 66% yield

with 76% de (Table 2, entry 2). Most remarkably, by

using these conditions we were able to lower the reaction

temperature down to -20 °C. At this temperature, the

reaction proceeded to completion overnight to afford

cyclopentenone 13 in 77% yield and 84% de (Table 3,

entry 3). Interestingly enough, the diastereomers of 13

are separable by simple column chromatography. At-

tempts at further lowering the cyclization temperature

were unsuccessful, with no conversion to the desired

cyclopentenone being observed after 24 h at -40 °C.

In an attempt to simplify the experimental procedure

for the generation of b -type complexes, the purely

thermal dissociation of CO was also considered and used

in reaction conditions C. These consisted in displacing

the equilibrium toward b by heating a hexane solution

of the starting hexacarbonyl complex at 50 °C for 1 h.

This solution was then cooled to the desired temperature,

and the reacting alkene was added. Following this

protocol, norbornene addition to a hexane solution of 10b/

10a (ca. 85/15) at 50 °C yielded almost instantaneously

(10 min) cycloadduct 13, albeit in a disappointing 20%

de (Table 2, entry 4). When the cyclization temperature

was lowered to 0 °C, an almost quantitative yield of 13

was recorded, but the diastereomeric excess was only 60%

(Table 2, entry 5).

When the experiments performed under conditions B

and C at the same temperature (0 °C) are compared, it

is clear that higher diastereoselectivities are recorded

under conditions B. This can probably be attributed to

the fact that, under conditions C, some a -type complex

remains in the reaction medium, and as we have seen in

entry 1, its reaction with norbornene is poorly selective.

According to this, only conditions B were used for the

generation of b-type complexes in further experiments.

Reaction of 10 with other strained alkenes under

conditions B afforded the corresponding adducts in good

yield and excellent diastereoselectivities. In the most

favorable case, reaction with norbornadiene at -20 °C

(Table 2, entry 7) afforded 14 in 82% yield and with 92%

diastereomeric excess. Again, the diastereomers turned

out to be readily separable by column chromatography.

The high reactivity exhibited by pentacarbonyl complex

10b could not be satisfactorily exploited for reactions with

less reactive olefins. For example, when 10 was submit-

ted to cyclization with cyclopentene using NMO activa-

tion (reaction conditions B), no Pauson-Khand adduct

could be detected in the final reaction mixture. We

attribute this to a low stability of the chelated complex

10b or its olefin-coordinated derivative in the CH₂Cl₂/

NMO mixture: In the absence of a sufficiently reactive

alkene (i.e., one giving an olefin-coordinated complex that

readily evolves into a product-forming cobaltacycle),

complex 10b slowly decomposes leading to a mixture of

unidentified products. Nevertheless, cyclization with

(23) According to PM3 calculations, the electron density at sulfur

in alcohols 1-3 varies in the order 2 > 1 > 3.

7042 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

Sch em e 6

cyclopentene could be achieved, albeit in low yield (14%),

by thermal activation at 60 °C in a sealed tube under

nitrogen (Scheme 6). Unfortunately, H NMR analysis

of the resulting cyclopentenone 18 revealed a 1:1 mixture

of diastereomers.

1

F igu r e 3. Synthetic strategy for the synthesis of enantio-

merically pure (S)-4-substituted-2-cyclopentenones from ad-

duct 14.

Sch em e 7^a

Syn th esis of En a n tiop u r e 4-Su bstitu ted 2-Cyclo-

p en ten on es. Although the cyclopentane ring is wide-

spread among chiral natural products, and much effort

has been devoted to the enantioselective construction of

five-membered carbocycles,²⁴general methods for the

enantioselective synthesis of substituted cyclopentenones

are scarce,^24a,band the preparation of these compounds

still relies on enzymatic kinetic resolutions or involves

the use of chiral starting materials.²⁵In view of this,

we planned to use 14 as a versatile new entry to the

enantiopure cyclopentanoid core. Several features con-

vert 14 in an optimal starting point for this purpose:

First, the chiral auxiliary 1 used in its synthesis can be

easily prepared on a multigram scale; second, 14 is

formed in high yield in a highly diastereoselective fashion

and can be easily brought to diastereomeric purity;

finally, the norbornene core in 14 should exert valuable

substrate stereocontrol from which addition reactions to

the R-alkoxycyclopentenone should benefit and can easily

be removed by retro Diels-Alder methodology. Thus,

starting from 14, a three-step sequence consisting of

conjugate addition, chiral auxiliary removal, and retro

Diels-Alder reaction (Figure 3) should readily produce

enantiomerically pure 4-alkyl-2-cyclopentenones.

To demonstrate the feasibility of this protocol, we

examined each of these transformations. After a range

a

Key: (a) (CH₃)₂CuLi, TMSCl, ether, -10 °C; (b) (n-Bu)₂Cu-

(CN)Li₂, TMSCl, ether, -10 °C; (c) Ph₂CuLi, TMSCl, ether, rt; (d)

0.1 equiv of NBu₄F, THF, 0 °C.

of conditions for conjugate addition to 14 were explored,²⁶

the use of cuprates [Me₂CuLi, prepared at low-temper-

ature according to Bertz,²⁷Bu₂Cu(CN)Li₂,²⁸and Ph₂CuLi,

prepared from CuBr₂‚SMe₂] in the presence of trimeth-

ylsilyl chloride (TMSCl)²⁹were found to give the best

results. Under these conditions, the addition reactions

were completely stereoselective, leading in high yield,

after deprotection of the intermediate silyl enol ethers

with a catalytic amount of fluoride (0.1 equiv of NBu₄F,

THF, 0 °C),³⁰to configurationally pure cis-R-alkoxy-â-

alkylcyclopentanone derivatives 19-21 (Scheme 7). The

cis stereochemistry in these R-alkoxy ketones was de-

duced from the coupling constant between the R-alkoxy

and R-alkyl protons (J ) 11 Hz, CDCl₃). Conveniently,

if the conjugate additions are performed on the diaster-

eomeric mixtures directly arising from the Pauson-

Khand reactions, minor diastereomers of the addition

products can be chromatographically removed with great

ease; in practice, it is advisable to perform the separa-

tions at this point of the synthetic sequence.

(24) For some non-Pauson-Khand enantioselective cycloaddition

routes leading to five-membered rings, see: (a) Greene, A. E.; Char-

bonnier, F. Tetrahedron Lett. 1985, 26, 5525-5528. (b) Greene, A. E.;

Charbonnier, F.; Luche, M.-J .; Moyano, A. J . Am. Chem. Soc. 1987,

109, 4752-4753. (c) Taber, D. F.; Amedio, J . C.; Raman, K. J . Org.

Chem. 1988, 53, 2984-2990. (d) Yamamoto, A.; Ito, Y.; Hayashi, T.

Tetrahedron Lett. 1989, 30, 375-378. (e) Zhu, G.; Chen, Z.; J iang, Q.;

Xiao, D.; Cao, P.; Zhang, X. J . Am. Chem. Soc. 1997, 119, 3836-3837.

For some enantioselective cyclization routes to five-membered rings,

see: (f) Taber, D. F.; Saleh, S. A.; Korsmeyer, R. W. J . Org. Chem.

1980, 45, 4699-4702. (g) Taber, D. F.; Raman, K. J . Am. Chem. Soc.

1983, 105, 5935-5937. (h) Taber, D. F.; Petty, E. H.; Raman, K. J .

Am. Chem. Soc. 1985, 107, 196-199. (i) Wu, X.-M.; Funakoshi, K.;

Sakai, K. Tetrahedron Lett. 1993, 34, 5927-5930. (j) Barnhart, R. W.;

Wang, X.; Noheda, P.; Bergens, S. H.; Whelan, J .; Bosnich, B. J . Am.

Chem. Soc. 1994, 116, 1821-1830. (k) Nishida, M.; Ueyama, E.;

Hayashi, H.; Ohtake, Y.; Yamura, Y.; Yanaginuma, E.; Yonemitsu, O.;

Nishida, A.; Kawahara, N. J . Am. Chem. Soc. 1994, 116, 6455-6456.

(l) Nishida, M.; Hayashi, H.; Yamura, Y.; Yanaginuma, E.; Yonemitsu,

O.; Nishida, A.; Kawahara, N. Tetrahedron Lett. 1995, 36, 269-272.

(m) Watanabe, N.; Ogawa, T.; Ohtake, Y.; Ikegami, S.; Hashimoto, S.-

I. Synlett 1996, 85-86. (n) Bertrand, M. P.; Crich, D.; Nouguier, R.;

Samy, R.; Stien, D. J . Org. Chem. 1996, 61, 3588-3589. (o) Barnhart,

R. W.; McMorran, D. A.; Bosnich, B. J . Chem. Soc., Chem. Commun.

1997, 589-590. (p) Oppolzer, W.; Kuo, D. L.; Hutzinger, M. W.; Le´ger,

R.; Durand, J .-O.; Leslie, C. Tetrahedron Lett. 1997, 38, 6213-6216.

(25) (a) Klunder, A. J . H.; Huizinga, W. B.; Sessink, P. J . M.;

Zwanenburg, B. Tetrahedron Lett. 1987, 28, 357-360. (b) Dols, P. P.

M. A.; Verstappen, M. M. H.; Klunder, A. J . H.; Zwanenburg, B.

Tetrahedron 1993, 49, 11353-11372. (c) Dols, P. P. M. A.; Klunder, A.

J . H.; Zwanenburg, B. Tetrahedron 1994, 50, 8515-8538.

Subsequent auxiliary removal was effected by reduc-

tive cleavage with samarium diiodide³¹in THF/MeOH

(Scheme 8). An operational modification of the original

procedure described by Molander,³²consisting of the use

of a more concentrated (0.5 vs 0.1 M) solution of SmI₂at

(26) For general surveys, see: (a) Perlmutter, P. Conjugate Addition

Reactions in Organic Synthesis; Pergamon Press: Oxford, 1992. (b)

Rossiter, B. E.; Swingle, N. M. Chem. Rev. 1992, 92, 771-806.

(27) Bertz, S. H.; Gibson, C. P.; Dabbagh, G. Tetrahedron Lett. 1987,

28, 4251-4254.

(28) Cf.: Lipschutz, B. H.; Ellsworth, E. L.; Siahaan, I. J .; Shirazi,

A. Tetrahedron Lett. 1988, 29, 6677.

(29) Corey, E. J .; Boaz, N. W. Tetrahedron Lett. 1985, 26, 6019-

6022.

(30) Corey, E. J .; Snider, B. B. J . Am. Chem. Soc. 1972, 94, 2549-

2550.

(31) Best results were obtained with samarium diiodide prepared

in situ by Kende’s method: Kende, A. S.; Mendoza, J . S. Tetrahedron

Lett. 1991, 32, 1699-1702.

(32) Molander, G. A.; Hahn, G.; J . Org. Chem. 1986, 51, 1135-1138.

Experimental and Computational Studies

J . Org. Chem., Vol. 63, No. 20, 1998 7043

Sch em e 8

F igu r e 4. Commonly accepted mechanism of the Pauson-

Khand reaction.

37

proposed by Magnus (Figure 4), which accommodates

a much higher temperature (0 vs -78 °C) under argon,

was used. This resulted in the rapid cleavage of the

carbon-oxygen single bond (10 min) and gave rise to

cyclopentanones 22 and 23 in excellent yields and allowed

the recovery of auxiliary 2 in almost quantitative yield.

Finally, among the existing methodologies to accomplish

the last step in the sequence, we chose to explore the

Lewis acid-catalyzed retro Diels-Alder conditions re-

ported by Grieco,³³which allow the reaction to be

performed at low temperature. Thus, when the tricyclic

ketones 22 and 23 were submitted to reaction with 1

equiv of MeAlCl₂in the presence of maleic anhydride at

55 °C, they smoothly led in high yield to (-)-(S)-4-methyl-

2-cyclopentenone (24) and (-)-(S)-4-butyl-2-cyclopenten-

one (26), respectively. Due to its volatility, cyclopenten-

one 24 (R ) Me) was best isolated as the known (+)-(R)-

3-methylcyclopentanonesemicarbazone25.³⁴Theenantiomeric

purity of 25, determined by comparison of its optical

rotation with that of an enantiomerically pure sample,³⁵

turned out to be 93%. This suggested that some racem-

ization might have occurred, most likely in the final

cycloreversion step. However, when dioxolane 27 derived

from (-)-2,3-butanediol was prepared starting from opti-

cally active 26, its ¹³C NMR spectrum showed no peaks

of the corresponding minor diastereomer.³⁶This more

solid evidence indicated that no significant racemization

takes place in the retro Diels-Alder reaction and that

the optical purity of the final 4-alkyl-2-cyclopentenones

obtained according to the present synthetic scheme is

g98% ee.

the main experimental observations on the reaction, has

found general acceptation. According to this proposal,

the reaction involves four main steps from the initial

complex I: (a) coordination of the reacting olefin to cobalt

to give complex II, (b) insertion of the π-complexed olefin

into a Co-C bond leading to cobaltacycle III, (c) insertion

of carbon monoxide into the Co-C(sp³) bond with forma-

tion of acylcobaltacycle IV, and (d) reductive elimination

and loss of a dicobalt hexacarbonyl fragment to give the

final cyclopentenone V.

Within the conceptual frame of this mechanism, the

initial interaction between the dicobalt hexacarbonyl

complex of a chiral alkoxyacetylene and a prochiral

alkene, such as norbornadiene, can take place distinctly

at either of the diastereotopic cobalt atoms (pro-R ≡ Co_A;

pro-S ≡ Co_B) in the C₂Co₂cluster; this will produce the

diastereomeric complexes II_Aand II_B(Scheme 9) in

which, according to previous experience^3eand to our

experimental results, attack on the exo face of the olefin

is assumed. Evolution of these complexes into the

experimentally observed R-alkoxysubstituted cyclopen-

tenone regioisomers can only take place through cobal-

tacycles III.³⁸Up to four of these cobaltacycles can form

depending both on the cobalt atom (Co_A, series A, or Co_B,

series B) involved in the process and on the relative

stereochemistry (syn or anti) of the methano bridge and

the non-ring-forming Co(CO)₃moiety. These cobalta-

cycles should then evolve in a pairwise manner into the

diastereomers of the cyclopentenone adduct. Thus,

whereas syn-III_Aand anti-III_Bwould be transformed into

V_A, whose structure and stereochemistry are those of

14(minor) when the R*OH ≡ 1, syn-III_Band anti-III_A

would lead to V_B, which analogously corresponds to

14(major). According to this, stereoselectivity can arise

in the reactions as the result of different situations.

Thus, two pathways could lead to the observed stereoi-

Ra tion a liza tion of th e Obser ved Dia ster eoselec-

tivity. Although the mechanism of the Pauson-Khand

reaction has not been studied in detail, the pathway

(33) Grieco, P. A.; Abood, N. J . Org. Chem. 1989, 54, 6008-6010.

(34) Kokke, W. C. M.; Varkevisser, F. A. J . Org. Chem. 1974, 39,

1535-1539.

(35) Given the existence of E/Z isomers in 25, which could influence

the value of the optical rotation (cf. ref 34), a commercial sample of

enantiomerically pure (R)-(+)-3-methylcyclopentanone (Aldrich, no.

M3,970-9) was converted to the semicarbazone under exactly the same

experimental conditions. This semicarbazone was subsequently used

as a reference.

(37) (a) Magnus, P. C.; Exon, C.; Albaugh-Robertson, P. Tetrahedron

1985, 41, 5861-5859. (b) Magnus, P. C.; Principe, L. M. Tetrahedron

Lett. 1985, 26, 4851-4854. (c) Castro, J .; Moyano, A.; Perica`s, M. A.;

Riera, A. Tetrahedron 1995, 51, 6541-6556.

(38) The formation of these cobaltacycles has the stereoelectronic

requirement of eclipsing CdC and Co-C(H) bonds. A similar set of

regioisomeric cobaltacyles leading to the never detected â-alkoxy-

substituted cyclopentenones would form via eclipsing of the CdC and

the Co-C(OR*) bonds.

(36) The diastereomeric mixture of dioxolanes prepared from race-

mic 3-butylcyclopentanone and (-)-2,3-butanediol shows

a well-

resolved ¹³C NMR spectrum. For the original method, see: Hiemstra,

H.; Wynberg, H. Tetrahedron Lett. 1977, 18, 2183-2186.

7044 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

Sch em e 9

somer Vb ≡ 14(major): Coordination of the olefin to the

pro-R cobalt atom leading to complex II_Aand formation

of the anti-cobaltacycle anti-III_Aor, alternatively, coor-

dination to the pro-S cobalt atom (II_B) and formation of

the syn cobaltacycle syn-III_B.

In order to achieve a better understanding of the

stereochemical outcome of the reaction of 10 with nor-

bornadiene and other strained olefines, we wanted to

answer three fundamental questions implicit in the

mechanistic picture above:

(a) Does the chelating arm present in the chiral

auxiliary efficiently discriminate between the diaste-

reotopic cobalt atoms in the formation of the internally

chelated species 10b?

(b) Does a chelated complex preferentially react with

an olefin through the cobalt atom involved in the chela-

tion?

(c) Is there any preference for syn or anti stereochem-

istry in the putative cobaltacycle intermediates III?

To answer these questions, we decided to study the

different intermediates involved in the reaction of 10 with

norbornadiene by computational means. The geometries

of all the studied species were fully optimized with the

semiempirical procedure PM3(tm), which incorporates

parametrization for transition metals to the original PM3

procedure,³⁹as implemented in the package of programs

SPARTAN 4.1.1,⁴⁰after testing its performance in the

description of dicobalt hexacarbonyl complexes of acety-

lenes.⁴¹To compare different reaction options, accurate

energies were subsequently evaluated by single point

calculations performed with density functional theory

(DFT), as implemented in the ADF 2.0.1 code,⁴²using

the VWN local exchange correlation potential⁴³with

Perdew-Wang 91 nonlocal exchange and correlation

corrections.⁴⁴

(a ) Str u ctu r e a n d Con figu r a tion a l Sta bility of th e

Ch ela ted Com p lex 10b. As we have seen in the

preceding sections, the use in Pauson-Khand reactions

of acetylenic ethers derived from auxiliaries 1-3 intro-

duces a new element in the mechanistic pathway: the

generation of internally chelated complexes. In fact, the

generation of these complexes (10b-12b) is crucial to the

achievement of diastereoselectivity, so that it is impor-

tant to know their structures and configurational stabil-

ity. In principle, the (methylthio)methyl arm can occupy

either the axial or one of the equatorial coordination sites

(the other one is not available due to geometrical reasons)

at either of the two cobalt atoms. All four possible

structures were located and minimized with the theoreti-

cal procedure discussed above. For each of the com-

plexes, two structures of very similar energy, differing

only in the orientation of the methyl substituent, were

located; only the most stable ones have been considered.

The most stable complex (I_A-a x) involves the coordination

of the (methylthio)methyl arm at the axial site of the

pro-R cobalt atom. The preference of phosphine-⁴⁵and

(41) PM3(tm) reproduces well the structure and conformation of

dicobalt hexacarbonyl complexes of alkynes as determined by X-ray

crystallography. With respect to the C₂Co₂cluster, both the Co-Co

and the C-Co distances are accurately reproduced, while the C-C

distance is somewhat overestimated. If desired, this parameter can

be fixed to the mean crystallographic value of 1.336 Å without

introduction of distortion in the overall structure. With respect to

coordinatively unsaturated cobalt species, PM3(tm) shows an exag-

gerated tendency (with respect to higher level DFT methods) to

saturate cobalt either through bridging carbonyls or through agostic

interactions with hydrogen atoms. Structures presenting these char-

acteristics should be treated carefully until confirmation of these effects

by means of independent DFT optimization.

(42) (a) Baerends, E. J .; Ellis, D. E.; Ros, P. E. Chem. Phys. 1973,

2, 41-51. (b) Versluis, L.; Ziegler, T. J . Chem. Phys. 1988, 88, 322-

328. (c) Fan, L.; Ziegler, T. J . Chem. Phys. 1991, 95, 7401-7408. (d)

teVelde, G.; Baerends, E. J . J . Comput. Phys. 1992, 99, 84-98.

(43) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J . Phys. 1980, 58, 1200-

1211.

(39) Stewart, J . J . P. J . Comput. Chem. 1989, 10, 209-220.

(40) SPARTAN, version 4.1.1, Wavefunction, Inc., 18401 Von Kar-

man Ave., Suite 370, Irvine, CA, 92612.

(44) Perdew, J . P.; Chevary, J . A.; Vosko, S. H.; J ackson, K. A.;

Pederson, M. R.; Singh, D. J .; Fiolhais, C. Phys. Rev. B 1992, 46, 6671-

6687.

Experimental and Computational Studies

J . Org. Chem., Vol. 63, No. 20, 1998 7045

in which the Pauson-Khand reactions of 10b take place

with useful diastereoselectivity.⁵⁰

(b) Coor d in a tion of Nor bor n a d ien e to Coba lt:

Str u ctu r e of th e π-Com p lexes II. Once the chelated

complex I is formed, and its interaction with the reacting

olefin [norbornadiene as a typical example] is allowed to

take place, the initial stage of the Pauson-Khand reac-

tion could in principle occur according to routes a or b,

as represented in Scheme 11. In route a , the olefin would

coordinate to the nonchelated cobalt atom, whereas in

route b coordination of the olefin would occur at the site

formerly occupied by sulfur.

Although a route similar to a could be operating in the

Pauson-Khand reactions of phosphine-substituted cobalt

carbonyl complexes,⁶the lability of the cobalt-sulfur

bond, predicted by model PM3(tm) calculations,⁵¹sug-

gests that route b could be operating in this case. This

assumption, analogous to Krafft’s mechanistic hypothesis

for the intramolecular Pauson-Khand reactions acceler-

ated by coordinating ligands,^8d,eprovides a rationale for

understanding the greatly increased reactivity of 10b

relative to 10a .⁵²Accordingly, only route b has been

explored in the present theoretical study.

F igu r e 5. DFT calculated relative energies of the internally

chelated pentacarbonyl complexes I.

thioether-type⁴⁶ligands for axial positions in cobalt

carbonyl complexes of alkynes is well documented. How-

ever, the cyclic nature of 10b could somewhat affect the

normal pattern of stabilities,⁴⁷as shown by the fact that

the next most stable complex (I_B-eq) involves substitution

at an equatorial site of the pro-S cobalt. The DFT-

calculated relative energies of the located minima are

shown in Figure 5, whereas the PM3(tm)-optimized

structures of the most stable complexes (I_A-a x and I_B-

eq) are in Figure 6 (left). Thus, the performed theoretical

calculations predict that 10b will largely (g85%) exist

at room temperature as a single diastereomer (repre-

sented by I_A-a x) with the pro-R cobalt coordinated to

sulfur. This is compatible with the NMR results dis-

cussed earlier.

In a solvent with coordinating ability like dichlo-

romethane, in which the reactions are performed, it is

logical to expect that solvent molecules will contribute

to stabilize the coordinatively unsaturated intermediate

generated in the dissociative step, thus lowering its

energy barrier.⁵³We have represented in Scheme 12 the

plausible course of the ligand-substitution process. The

effect of solvent has been taken into account by means

of a single dichloromethane molecule.

An important concern that arises when the diastereo-

selectivity of the Pauson-Khand reactions of 10b is

considered is the configurational stability of the chelate.

It is known from NMR studies at variable temperature

that the dicobalt hexacarbonyl complexes of alkynes are

fluxional at ordinary temperatures.⁴⁸However, it is

believed that this fluxionality is due to carbonyl motions

at a single cobalt atom. In this respect, it is important

to recall that some diastereomerically pure phosphine-

substituted cobalt carbonyl complexes of alkynes show

good configurational stability at room and even higher

temperatures.^6a-c,45c

For simplicity, we have represented and evaluated the

substitution process as taking place at a single coordina-

tion site (axial, pro-R cobalt), although this is not

necessarily the case. The important finding, which is

rather independent of this assumption, is that ligand

substitution on I_A-a x leading to the first intermediate in

the Pauson-Khand reaction II_A-a x is predicted to be an

essentially thermoneutral process with low activation

energy when the presence of a coordinating solvent is

included in the calculation.⁵⁴For comparison, when the

molecule of dichoromethane is not included in the cal-

culation, the cleavage of the sulfur-cobalt bond in I_A-a x

In any case, the viability of a low-energy epimerization

pathway for 10b, as represented in Scheme 10, was

investigated. A structure with the characteristics of

I_{br id ge}was shown to be an energy minimum at the PM3-

(tm) level of theory. Then, a single-point DFT calcula-

tion⁴⁹provided for I_{br id ge}a relative energy of +20.2

kcal‚mol^-1(respect to I_A-a x). Since I_{br id ge}should be

connected with the chelated species through a higher

energy transition state, the present theoretical results

tend to indicate that complexes such as I_A-a x will be

configurationally stable over the range of temperatures

is calculated to be endothermic by 33.8 kcal‚mol^-1

.

(49) A geometry optimization of I_bridge, performed at the same DFT

level used for energy calculations in the present study, showed that

the bridged structure is a stationary point at that level of theory.

(50) Note, however, that this or a similar mechanism could be

involved in a hypothetical equilibration of 10b when generated at room

temperature by NMO oxidation of 10a in the absence of any olefin.

Under these reaction conditions (presence of olefin), if diastereoselec-

tivity depends on cobalt-specificity of the chelating arm/olefin substitu-

tion process, its level will be determined by the relative rates of

Pauson-Khand reaction and epimerization. This is in agreement with

the fact that the more reactive olefins provide the higher diastereomeric

ratios of ketone products.

(45) See, for instance: (a) Bonnet, J .-J .; Mathieu, R. Inorg. Chem.

1978, 17, 1973-1976. (b) Arewgoda, M.; Robinson, B. H.; Simpson, J .

J . Am. Chem. Soc. 1983, 105, 1893-1903. (c) Bradley, D. H.; Khan,

M. A.; Nicholas, K. M. Organometallics 1989, 8, 554-556. (d) J effery,

J . C.; Pereira, R. M. S.; Vargas, M. D.; Went, M. J . J . Chem. Soc.,

Dalton Trans. 1995, 1805-1811.

(51) A systematic PM3(tm) study of bond dissociation energies on

(acetylene)Co₂(CO)₅(SMe₂) reveals that the cobalt-sulfur bond is the

weakest in the molecule by ∼4 kcal‚mol^-1

.

(52) Phosphine-substituted dicobalt pentacarbonyl complexes of

alkynes are less reactive toward alkenes than the parent hexacarbonyl

complexes. See, for instance, ref 6e and: Bradley, D. H.; Khan, M. A.;

Nicholas, K. M. Organometallics 1992, 11, 2598-2607.

(46) Rumin, R.; Manojlovic-Muir, L.; Muir, K. W.; Pe´tillon, F. Y.

Organometallics 1988, 7, 375-383.

(47) For an example of a complex bis-equatorially substituted by

DPPM, see: Gelling, A.; Went, M. J .; Povey, D. C. J . Organomet. Chem.

1993, 455, 203-210.

(53) For a similar suggestion on solvent stabilization of (otherwise)

unsaturated intermediates in ligand-exchange reactions taking place

by a dissociative mechanism, see: Cowey, W. D.; Brown, T. L. Inorg.

Chem. 1973, 12, 2820-2825.

(48) (a) Aime, S.; Milone, L.; Rossetti, R.; Stanghellini, P. L. Inorg.

Chim. Acta 1977, 22, 135-139. (b) Band, E.; Muetterties, E. L. Chem.

Rev. 1978, 78, 639-658.

(54) The enthalpy change of the first step (+15.1 kcal‚mol^-1) should

be similar to the activation enthalpy of the substitution process.

7046 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

F igu r e 6. PM3(tm)-optimized structures and DFT-calculated relative energies of the most stable stereoisomers of intermediates

I (left), II (middle), and III (right) leading to the major and minor diastereomers of V.

Sch em e 10

Sch em e 11

In addition, to allow comparison with the nonselective,

purely thermal reactions of 10a , which should probably

take place through both cobalt atoms, we also studied

the set of complexes II containing an exo-coordinated

norbornadiene unit at the different sites of the pro-S

cobalt atom. The DFT-calculated relative energies of the

most stable conformational minima located for the three

possible coordination sites of the pro-R cobalt (II_A) and

of the pro-S cobalt atom (II_B) are shown in Figure 7.

As we have already mentioned, the stereoelectronic

requirement for the evolution of these complexes to

cobaltacycles en route to cyclopentenones is a synperipla-

nar arrangement of the CdC and the Co-C bonds

participating in the metathesis step.^3eAccordingly,

Since we have previously shown that epimerization of

I_Ato I_Binvolves a rather high-energy pathway and

because it is known that norbornadiene reacts only on

its exo face, we next explored the existence of different

energy minima of the π-olefin complexes II involving

coordination of the exo face of norbornadiene to the pro-R

cobalt atom in the C₂Co₂cluster. Due to previous

knowledge on fluxionality of similar complexes,^48ano site

specificity (i.e., axial/axial or equatorial/equatorial) was

considered for the substitution process.

Experimental and Computational Studies

J . Org. Chem., Vol. 63, No. 20, 1998 7047

F igu r e 8. DFT-calculated relative energies of the cobalta-

cycles III.

F igu r e 7. DFT-calculated relative energies of the norborna-

diene π-complexes II.

Sch em e 13

Sch em e 12

PM3 structures of the most stable anti-type cobaltacycles

are shown in Figure 6 (right).

The first important point relative to these structures

that deserves comment is the great difference in stability

between the syn and anti cobaltacycles. If we admit that

the transition states leading to these structures should

reflect their energy differences, it follows as a conse-

quence that syn cobaltacycles are not intermediates in

the Pauson-Khand reactions of 10b with olefins contain-

ing the bicyclo[2.2.1]heptane skeleton. Then, the sole

factor determining the diastereoselectivity of the process

is the cobalt atom in the C₂Co₂cluster participating in

the formation of the cobaltacycle.

The clear energetic preference for anti-type cobalta-

cycles (anti-III), together with the predicted configura-

tional stability of intermediates I and II, allows the

establishment of a genealogy for the diastereomers of the

reaction product V, back to the chelated complexes I

(Scheme 13).

Thus, the major stereoisomer of 14 (represented by V_A)

can only arise from cobaltacycle anti-III_A, in which all

the stereochemical information is already present. This

cobaltacycle, in turn, can only arise from the olefin

complex II_A-eq -R, while II_A-eq -R would likewise arise

from the most stable chelated complex with the (meth-

ylthio)methyl arm located at the pro-R cobalt, i.e., I_A.

Analogous arguments allow the tracing of the origin of

the minor estereoisomer of 14 (represented by V_B) back

to the most stable of the chelated complexes with the

sulfur side chain on the pro-S cobalt (I_B-eq). Very

interestingly, the present calculations predict that the

major stereoisomer of the cyclopentenone adduct is formed

through the sequence with the most stable intermediates.

Two additional aspects of these sequences deserve

complexes II_A-eq-â and II_B-eq-â could only lead to a

â-alkoxy-substituted cyclopentenone, which is never ob-

served, and, thus, does not require further consideration.

On the other hand, each of the other four complexes could

in principle lead to either of the two possible diastereo-

mers of V, depending on the stereochemistry (syn or anti)

of the intermediate cobaltacycle (see Scheme 9). It is

worth noting that, among the olefin complexes able to

lead to the observed products, the most stable one is II_A-

eq-R, in which the alkene is coordinated to the same

cobalt atom (Co_A) that is involved in the most stable

pentacarbonyl complex I_A. The PM3(tm)-optimized struc-

tures of the most stable π-complexes at each cobalt atom

(II_A-eq-R and II_B-a x) are shown in Figure 6 (middle).

(c) Str u ctu r e of th e Coba lta cycles III a n d Or igin

of th e Obser ved Ster eoselectivity. At this point of

the discussion, it is necessary to analyze the preferred

stereochemistry and relative energies of the four possible

cobaltacycles III leading to the observed final products.

While it is not clear whether the formation of these

species takes place in a concerted or stepwise manner, it

can be anticipated that the addition of CO (or a coordi-

nating solvent) to a hypothetical, coordinatively unsatur-

ated (16-electron) intermediate cobaltacycle should take

place without (or with very low) activation energy. We

have represented in Figure 8 the DFT-calculated relative

energies of the four possible cobaltacycles III, while the

7048 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

Sch em e 14

auxiliary has been effectively transmitted to cobalt. Both

the observed diastereoselectivities in the low-temperature

Pauson-Khand reactions of the chelated complexes and

the results of the theoretical study indicate that, at low

temperature, substitution of the alkylthio ligand by a

reacting olefin occurs with cobalt specificity. Given the

predicted high energetic preference for anti-type cobal-

tacycles, reaction (olefin coordination and irreversible

cobaltacycle formation) at a single cobalt atom is a

sufficient condition for stereocontrol. In this way, the

chiral information is transmitted from cobalt to the final

reaction product.

From an operational point of view, a unique advantage

offered by the present methodology is the possibility of

performing the normally difficult first dissociative step

of the Pauson-Khand reaction in a facile and stereocon-

trolled manner and in the absence of a reacting olefin.

The so-formed reactive species can then be brought to

the desired temperature for reaction. From a practical

perspective, the Pauson-Khand reaction of alkoxyacety-

lenes derived from chelating auxiliaries has already

found application as the key step in a highly enantiose-

lective synthesis of (+)-brefeldin A.^7a

If these auxiliaries had to be used in intramolecular

Pauson-Khand reactions, the generation of the inter-

nally chelated species in the absence of a reacting olefin

would be no longer possible. According to our arguments,

the achievement of stereocontrol in such reactions should

greatly depend on the relative rates of the generation

(and, probably, equilibration) of the chelated intermedi-

ate and of the subsequent Pauson-Khand reaction. This

is, in fact, what is experimentally observed. Thus, when

the cobalt complex 28 is submitted to either thermal or

chemical (NMO) activation, the formation of an inter-

mediate complex cannot be observed by TLC, and the

Pauson-Khand adduct is formed with low diastereose-

lectivity (e2:1).^7bConversely, in the case of complexes

29-31, the formation of chelated pentacarbonyl com-

plexes is readily promoted either thermally or chemically,

as evidenced by TLC, and the subsequent evolution of

these complexes into cyclopentenone adducts takes place

with high diastereoselectivity (up to 12:1).^7b,d

comment. First, it has been generally assumed that

cobaltacycle formation is the product-determining step

in the Pauson-Khand reaction. If one looks at the

energetics of this process in one particular case, i.e., the

conversion of II_A-eq-R into anti-III_A, as shown in Scheme

14, the calculated strong exothermic character indicates

that this step will be irreversible in practice, thus

confirming the qualitative predictions.

The second aspect refers to the stabilities of the olefin

complexes (II_A/II_B) (Figure 6). When these complexes

arise from the equilibrium mixture of chelated complexes

(I_A/I_B), generated either thermally or, preferentially,

oxidatively but always in the absence of the olefin

component (as we have considered until now), a clear

predominance of II_A, with the olefin located at the

(formerly) pro-R cobalt is to be expected. Conversely,

when the reaction is performed with thermal activation

and with the olefin present from the first moment, the

intermediacy of a chelated complex is not necessarily

involved, and the participation in the process of the set

of complexes II_Bcould be enhanced. As a consequence,

a decrease in diastereoselectivity should be expected; this

is exactly what is experimentally observed (Table 2).

In summary, although a complete mechanistic picture

of the Pauson-Khand reaction is still far away, the DFT//

PM3(tm) calculations performed on the different inter-

mediates in the reaction of 10b with norbornadiene

provide answers for the main questions underlying the

stereochemical outcome of the process and allow the

rationalization of the observed diastereoselectivity: The

formation of a chelated complex directs the reaction

toward a precise cobalt atom (the pro-R one in the initial

complex). A subsequent ligand substitution process

(sulfide/olefin) at that cobalt atom leads to an intermedi-

ate II_Aalready containing all of the structural elements

of the final product, and the irreversible evolution of this

intermediate into an anti cobaltacycle anti-III_Acreates

the new stereogenic centers with the configuration found

in the major diastereomer of the final product.

The concepts that led to the synthesis and use of 1-3

have also fostered the preparation of the thiol analogue

32.⁵⁵Quite gratifyingly, the Pauson-Khand chemistry

of 32 follows the same pattern as with 1-3, and its

cycloaddition with norbornadiene takes place with dias-

tereoselectivities of up to 95:5.⁵⁶

Con clu d in g Rem a r k s

The chiral auxiliaries with chelating ability represent

a new useful concept for diastereocontrol in Pauson-

Khand reactions. The oxidation with NMO of the dico-

balt hexacarbonyl complexes of the alkoxyacetylenes

derived from these auxiliaries selectively leads to pen-

tacarbonyl complexes, whose stabilization can be at-

tributed to an effective dative bond between the alkylthio

appendage in the auxiliary and the coordinatively un-

saturated cobalt atom. Spectroscopic data and theoreti-

cal calculations are consistent with the predominance of

a single diastereomer in these chelated complexes, thus

indicating that the chiral information contained in the

We are currently exploring new aspects and applica-

tions of the chemistry of coordinatively modified cobalt

(55) Montenegro, E.; Echarri, R.; Claver, C.; Castillo´n, S.; Moyano,

A.; Perica`s, M. A.; Riera, A. Tetrahedron: Asymmetry 1996, 7, 3553-

3558.

(56) Montenegro, E.; Poch, M.; Moyano, A.; Perica`s, M. A.; Riera,

A. Tetrahedron Lett. 1998, 39, 335-338.

Experimental and Computational Studies

J . Org. Chem., Vol. 63, No. 20, 1998 7049

was washed with brine, dried (MgSO₄), and concentrated in

vacuo. Flash column chromatography (SiO₂, hexane/AcOEt,

2.5%) of the residue afforded 160 mg (87% yield) of 10-

(phenylthio)camphor as a colorless oil: [R]_D) +7.0 (c 1.5,

carbonyl complexes of alkynes. Results will be reported

in due course.

Exp er im en ta l Section

CHCl₃); IR (film) ν_max) 2960, 1740, 1585, 1480, 1050 cm^-1

;

¹H NMR (200 MHz, CDCl₃) δ 0.94 (s, 3H), 1.09 (s, 3H), 1.15-

1.61 (m, 2H), 1.75-2.23 (m, 4H), 2.34-2.50 (m, 1H), 2.88-

3.37 (AB, J ) 11 Hz, 2H), 7.10-7.45 (m, 5H) ppm; ¹³C NMR

(50 MHZ, CDCl₃) δ 20.1 (CH₃), 20.3 (CH₃), 26.5(CH₂), 26.9

(CH₂), 30.8 (CH₂), 43.0 (CH₂), 43.4 (CH), 47.8 (C), 60.8 (C),

125.6 (CH), 128.5 (CH), 128.8 (CH), 138.2 (C) ppm; HRMS (EI)

calcd for C₁₆H₂₀OS 260.1235, found 260.1233.

Com p u ta tion a l Deta ils. Geometry optimizations at the

PM3(tm) level of theory were performed without geometrical

restrictions. For DFT calculations, a double-ú STO basis set

was used. The frozen core approximation^42awas used for the

1s²shell on carbon and oxygen, as well as the 1s², 2s², and

2p⁶shells on cobalt, chlorine, and sulfur.

Gen er a l Meth od s. Melting points were determined in a

open capillary tubes and are uncorrected. Infrared spectra

were recorded in Fourier transform mode, using film (NaCl)

(2R)-10-(P h en ylth io)isobor n eol (3). In a round-bottomed

flask under N₂were placed LiAlH₄(26 mg, 0.68 mmol) and

ether (1 mL). The resulting slurry was cooled at 0 °C, and a

solution of 10-(phenylthio)camphor (6) (149 mg, 0.57 mmol)

in ether (5 mL) was added dropwise via syringe. The mixture

was stirred at room temperature for 2 h, and the reaction was

quenched by sequential addition of 26 µL of water, 26 µL of

15% NaOH, and 80 µL of water. The resulting white precipi-

tate was filtered through Celite and washed with ether. The

filtrate was dried (MgSO₄) and concentrated in vacuo to

provide a mixture of exo/endo isomers in a 10/1 ratio (¹H NMR).

The exo isomer was purified by flash column chromatography

(SiO₂, hexane/AcOEt, 2%) to afford 127 mg (85% yield) of 10-

(phenylthio)isoborneol as a colorless oil: [R]_D) -35.5 (c 2.8,

CHCl₃); IR (film) ν_max) 3450 (broad), 2940, 1585, 1485, 1070

1

or KBr pellet techniques. The H NMR spectra were recorded

at 200 or 300 MHz in CDCl₃unless specified otherwise, with

tetramethylsilane as internal standard. The ¹³C NMR spectra

were recorded at 50.3 or 75.4 MHz in CDCl₃unless specified

otherwise. Signal multiplicities were established by DEPT

experiments. In all cases, chemical shifts are in ppm downfield

of TMS. Mass spectra were recorded at 70 eV ionizing voltage;

ammonia was used for chemical ionization (CI). Elemental

analyses were performed by the Servei d’Ana`lisis Elementals

del CSIC de Barcelona, and exact mass measurements (HRMS)

were performed by the Laboratori d'Espectrometria de Masses

del CSIC de Barcelona. THF and diethyl ether were distilled

from sodium benzophenone ketyl. Dichloromethane was

distilled from CaH₂. Hexamethyl phosphoramide (HMPA) was

purified by distillation from CaH₂at reduced pressure and

stored over 4 Å molecular sieves. All reactions were performed

in flame- or oven-dried glassware under an N₂or Ar atmo-

sphere. Reaction progress was followed by TLC (silica gel 60,

F254). Silica gel (70-230 mesh) was used for column chro-

matography. (2R)-10-Mercaptoisoborneol (4),¹⁵10-iodocam-

phor (5),¹⁷and (2R)-10-(methylthio)isoborneol (1)¹⁶were pre-

pared by literature procedures.

cm^-^{1 1}H NMR (200 MHz, CDCl₃) δ 0.87 (s, 3H), 1.09 (s, 3H),

:

1.00-1.90 (m, 7H), 2.95-3.23 (AB, J ) 12 Hz, 2H), 3.89-3.99

(m, 1H), 7.15-7.50 (m, 5H) ppm; ¹³C NMR (50 MHz, CDCl₃) δ

20.4 (CH₃), 21.7 (CH₃), 27.6 (CH₂), 31.5 (CH₂), 34.4 (CH₂), 39.7

(CH₂), 45.7 (CH), 48.3 (C), 52.7 (C), 77.3 (CH), 126.4 (CH),

129.4 (CH), 129.5 (CH), 137.9 (C) ppm; HRMS (EI) calcd for

C

₁₆H₂₂OS 262.1391, found 262.1388.

(2R-exo)-2-[(E)-Dich lor oeth en oxy]-10-(m eth ylth io)bor -

n a n e (7). In a round-bottomed flask under N₂, potassium

hydride in mineral oil (35%, 2.4 g, 21 mmol) was washed with

anhydrous hexane (3 × 10 mL). Anhydrous THF (2 mL) was

added, and over the resulting slurry was added dropwise a

solution of 10-(methylthio)isoborneol (1) (2.0 g, 10 mmol) in

THF (20 mL) via cannula. After the hydrogen evolution

ceased, the solution was cooled to -40 °C, and a solution of

trichloroethene (1.37 g, 10.5 mmol) in THF (15 mL) was added

via cannula. The reaction mixture was allowed to warm

gradually at -10 °C. At this point, TLC analysis revealed

complete conversion of the starting material. The reaction was

carefully quenched with MeOH and poured over saturated

NH₄Cl solution. The aqueous layer was extracted with hex-

ane, and the combined organic phases were washed with brine,

dried (MgSO₄), and concentrated in vacuo. Flash column

chromatography (2% v/v NEt₃pretreated SiO₂, hexane/AcOEt,

2.5%) afforded 2.3 g (78% yield) of exo-2-[(E)-dichloroethenoxy]-

10-(methylthio)bornane as a colorless oil: [R]_D) -72.3 (c 1.8,

CHCl₃); IR (film) ν_max) 3120, 2970, 1665, 1285, 1100 cm^-1; ¹H

NMR (200 MHz, CDCl₃) δ 0.89 (s, 3H), 1.12 (s, 3H), 1.00-

2.10 (m, 7H), 2.14 (s, 3H), 2.50, 2.92 (AB, J ) 12 Hz, 2H), 4.50

(dd, J ) 3, 7 Hz, 1H), 5.41 (s, 1H) ppm; ¹³C NMR (50 MHz,

CDCl₃) δ 17.4 (CH₃), 19.9 (CH₃), 20.4 (CH₂), 26.7 (CH₂), 30.7

(CH₂), 33.1 (CH₂), 38.0 (CH₂), 45.3 (CH), 47.8 (C), 53.2 (C),

85.6 (CH), 95.4 (CH), 142.5 (C) ppm. MS (DIP-CI-NH₃) m/e

) 183 (M⁺- C₂HCl₂O, 100), 295 (M⁺+ 1, 33), 312 (M⁺+ 18,

67), 327 (M⁺+ 35, 57).

(2R)-10-(Neop en tylth io)isobor n eol (2). In a round-bot-

tomed flask under N₂, sodium hydride (90 mg, 3.0 mmol, 80%

in oil) was washed twice with anhydrous hexane (2 mL). In

this order, anhydrous THF (3 mL) and a solution of 10-

mercaptoisoborneol (0.5 g, 2.7 mmol) in THF were added via

syringe. The resulting slurry was stirred at room temperature

for 30 min. Then, neopentyl bromide (210 µL, 3.0 mmol) and

HMPA (2 mL) were added via syringe. The reaction mixture

was heated at reflux overnight. The reaction was cautiously

quenched with water, the aqueous layer was extracted with

ether, and the combined organic layers washed with brine and

dried (MgSO₄). Solvent removal under vacuum afforded a

mixture of 10-(neopentylthio)isoborneol and starting material.

Further purification by column chromatography (SiO₂, hexane/

AcOEt, 2%) yielded 350 mg of 2 (70% yield) as a colorless oil:

[R]_D) -50.1 (c 1.6, CHCl₃); IR (film) ν_max) 3470 (broad), 2950,

1475, 1390, 1365, 1075 cm^-^{1 1}H NMR (300 MHz, CDCl₃) δ

;

0.82 (s, 3H), 0.99 (s, 9H), 1.05 (s, 3H), 0.93-1.11 (m, 1H), 1.19-

1.32 (m, 1H), 1.47-1.58 (m, 1H), 1.63-1.85 (m, 4H), 2.17 (s

broad, 1H), 2.44-2.54 (AB, J ) 12 Hz, 2H), 2.51-2.80 (AB, J

) 10 Hz, 2H), 3.88-3.92 (dd, J ) 3 and 8 Hz, 1H) ppm; ¹³C

NMR (50 MHz, CDCl₃) δ 20.3 (CH₃), 21.1 (CH₃), 27.6 (CH₂),

29.4 (3CH₃), 31.6 (CH₂), 33.9 (C), 34.9 (CH₂), 39.4 (CH₂), 45.6

(CH), 48.0 (C), 49.5 (CH₂), 52.9 (C), 77.6 (CH) ppm. Anal.

Calcd for C₁₅H₂₈OS: C, 70.25; H, 11.00; S, 12.50. Found: C,

70.20; H, 11.11; S, 12.24.

(+)-10-(P h en ylth io)ca m p h or (6). In a round-bottomed

flask under N₂, sodium hydride (120 mg, 4.0 mmol, 80% in

oil) was washed twice with anhydrous hexane (2 mL). THF

(2 mL) was added via syringe and the flask cooled at 0 °C.

Next, thiophenol (350 µL, 3.4 mmol) was added neat via

syringe, and the reaction mixture was stirred at room tem-

perature for 30 min. To this slurry, a solution of 10-iodocam-

phor (5)¹⁷(0.8 g, 2.8 mmol) in THF (2 mL) and HMPA (1.5

mL) were added via syringe. The reaction mixture was heated

at reflux, and the progress of the reaction was monitored by

TLC. After complete consumption of the starting material

(2-4 h), the reaction was cautiously quenched with water and

the aqueous layer extracted with ether. The organic phase

(2R-exo)-2-[(E)-Dich lor oeth en oxy]-10-(n eop en tylth io)-

bor n a n e (8). This compound was prepared following the

experimental procedure employed for the synthesis of 7. exo-

2-[(E)-Dichloroethenoxy]-10-(neopentylthio)bornane was ob-

tained in 74% yield as a colorless oil: [R]_D) -51.7 (c 1.3,

CHCl₃); IR (film) ν_max) 3110, 2970, 1670, 1630, 1280, 1080

1

cm^-1; H NMR (200 MHz, CDCl₃) δ 0.88 (s, 3H), 0.97 (s, 9H),

1.11 (s, 3H), 1.00-2.05 (m, 7H), 2.47 (s, 2H), 2.50-2.88 (AB,

J ) 12 Hz, 2H), 4.49-4.58 (dd, J ) 3, 7 Hz, 1H), 5.39 (s, 1H)

ppm; ¹³C NMR (50 MHz, CDCl₃) δ 20.0 (CH₃), 20.5 (CH₃), 26.8

(CH₂), 28.9 (3CH₃) 30.7 (CH₂), 32.9 (C), 33.0 (CH₂), 38.0 (CH₂),

45.5 (CH), 47.8 (C), 49.2 (CH₂), 52.7 (C), 85.5 (CH), 95.3 (CH),

143.0 (C) ppm.

7050 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

(2R-exo)-2-[(E)-Dich lor oeth en oxy]-10-(p h en ylth io)bor -

n a n e (9). The above compound was prepared following the

experimental procedure employed for the synthesis of 7. exo-

2-[(E)-Dichloroethenoxy]-10-(phenylthio)bornane was obtained

in 77% yield as a colorless oil: [R]_D) -76.6 (c 1.6, CHCl₃); IR

(film) ν_max) 2950, 1625, 1480, 1280 cm^-1; ¹H NMR (200 MHz,

CDCl₃) δ 0.89 (s, 3H), 1.16 (s, 3H), 1.00-2.10 (m, 7H), 2.92-

3.39 (AB, J ) 12 Hz, 2H), 4.49-4.58 (dd, J ) 3, 7 Hz, 1H),

5.36 (s, 1H) ppm; ¹³C NMR (75 MHz, CDCl₃) δ 20.1 (CH₃), 20.5

(CH₃), 26.8 (CH₂), 30.7 (CH₂), 32.9 (CH₂), 38.1 (CH₂), 45.4 (CH),

48.2 (C), 52.9 (C), 85.4 (CH), 95.8 (CH), 125.6 (CH), 128.6 (CH),

129.2 (CH), 137.8 (C), 142.5 (C) ppm; HRMS (EI) calcd for

the starting complex (20 h), the mixture was opened to the

air for 5 h to induce the formation of insoluble cobalt salts

and then filtered through a pad of Celite and the solvent

removed in vacuo. Flash column chromatography (2% v/v

NEt₃pretreated SiO₂, hexane/AcOEt, 2.5%) afforded 32 mg

(65%) of 13 in the form of a 70:30 (40% de) mixture of

1

diastereomers, as determined by H NMR.

Rea ction Con d ition s B. A solution of complex 10a (161

mg, 0.31 mmol) in methylene chloride (17 mL) under nitrogen

was treated with N-methylmorpholine N-oxide (221 mg, 1.8

mmol). When the reaction mixture was stirred at room

temperature for 10 min, TLC analysis showed complete

conversion of the starting red hexacarbonyl complex to the

brown pentacarbonyl complex 10b. The mixture was cooled

at -20 °C, and a solution of norbornene (280 g, 3.0 mmol) in

CH₂Cl₂(1 mL) was added via syringe. The reaction was stirred

at -20 °C overnight, the resulting purple mixture was filtered

through a pad of Celite, and the filtrate was evaporated in

vacuo. Flash column chromatography (2% v/v NEt₃pretreated

SiO₂, hexane/AcOEt, 5%) afforded 84 mg (77%) of 13 as a 92:8

(84% de) mixture of diastereomers (¹³C NMR). The major

diastereomer could be isolated by column chromatography as

a white crystalline solid. When the present procedure was

followed but the cyclization performed at 0 °C, cyclopentenone

13 was otained in 66% yield and 76% de.

C

₁₈H₂₂Cl₂OS 356.0768, found 356.0769.

Gen er a l P r oced u r e for th e P r ep a r a tion of th e Hexa c-

a r bon yl-/P en t a ca r b on ylDicob a lt Com p lexes: Dicob a lt

Com p lexes of (2R-exo)-2-(Eth yn yloxy)-10-(m eth ylth io)-

bor n a n e (10a /10b). To a solution of 7 (2.12 g, 7.8 mmol) in

ether (20 mL) under nitrogen at -65 °C was added dropwise

n-BuLi (10.4 mL, 16.3 mmol) via syringe. The reaction

mixture was stirred for 30 min at -65 °C and then allowed to

warm gradually to -10 °C; TMSCl (1.47 mL, 11.7 mmol) was

added via syringe. A white solid formed, and the reaction

mixture was stirred for 2 h at room temperature. At this point,

IR analysis showed complete conversion to the corresponding

silylated alkyne. Upon dilution with hexane, the crude was

filtered through Celite and the solvent removed in vacuo. The

residue was dissolved in hexane (50 mL), and solid Co₂(CO)₈

(90%, 3.2 g, 8.5 mmol) was added to this solution. When CO

evolution ceased (15 min), the solvent was partially evaporated

and the complex filtered through a short pad of alumina

(hexane/AcOEt, 2%). Solvent removal yielded a red oily

residue. The red complex was dissolved in MeOH (75 mL)

under CO atmosphere and treated with K₂CO₃(3.1 g) under

vigorous stirring. After 30 min at room temperature, TLC

analysis showed complete deprotection of the TMS group.⁵⁷

Methanol was then partially evaporated, and the solution was

extracted with hexane. The hexane layer was washed with

brine, dried (MgSO₄), and concentrated in vacuo to yield 2.3 g

of 10a (62% overall yield) as a viscous red oil. On standing,

10a slowly evolves into 10b with loss of carbon monoxide. This

evolution can be avoided by keeping 10a under a CO atmo-

sphere. When a hexane solution of 10a was heated at 50 °C,

a maximum concentration of 10b (10b/10a ≈ 85/15) was

achieved after 1 h. By treating a dichloromethane solution of

10a with N-methylmorpholine N-oxide (ca. 6 equiv) at room

temperature, a complete conversion to 10b was observed.

Except for the thermal or N-oxide-promoted conversion to 10b,

the hexacarbonyl complex 10a was not manipulated before use

in Pauson-Khand reactions. Hexacarbonyl complex 10a : red

spot on TLC (R_f) 0.75, SiO₂, hexanes/ethyl acetate 9/1); IR

(film) ν_max) 2086, 2046, 2016, 1957 cm^-1; ¹H NMR (300 MHz,

C₆D₆) δ 0.66 (s, 3H), 1.00 (s, 3H), 0.90-1.65 (m, 5H), 1.71-

1-78 (dd, J ) 14, 8 Hz, 1H), 1.84 (s, 3H), 1.94-2.04 (m, 1H),

2.24, 2.80 (AB, J ) 11 Hz, 2H), 4.05-4.08 (dd, J ) 3, 7 Hz,

1H), 5.25 (s, 1H) ppm; MS (DIP-CI-NH₃) m/e ) 483 (M⁺+ 1

- CO, 80), 528 (M⁺+ 18, 25), 545 (M⁺+ 35, 5). Internally

chelated pentacarbonyl complex 10b: brown spot on TLC (R_f

) 0.50, SiO₂, hexanes/ethyl acetate 9/1); IR (film) ν_max) 2086,

Rea ction Con d ition s C. A solution of complex 10a (82

mg, 0.16 mmol) in hexane (10 mL) was heated at 50 °C while

being purged with a nitrogen stream. After 1 h, no further

conversion to 10b was visible by TLC (ca. 15:85 ratio of 10a /

10b as determined by NMR). The solution was then cooled

at 0 °C, and a solution of norbornene (150 mg, 1.6 mmol) in

hexane (1 mL) was added via syringe. The reaction mixture

was stirred for 12 h at 0 °C and filtered through a pad of SiO₂

and the solvent removed in vacuo. This afforded 55 mg (99%)

of 13 in the form of an 80:20 (60% de) mixture of diastereomers,

as determined by ¹³C NMR. When the reaction with nor-

bornene was performed at 50 °C, the process took 10 min to

completion and afforded cyclopentenone 13 in 69% yield and

20% de. Major diastereomer: mp 112-113 °C; [R]_D) -106.6

(c 5, CHCl₃); IR (film) ν_max) 3060, 2950, 1710, 1610, 1330,

1300 cm^-1; ¹H NMR (200 MHz, CDCl₃) δ 0.88 (s, 3H), 1.05 (s,

3H), 0.90-2.20 (m, 15H), 2.06 (s, 3H), 2.35 (m, 1H), 2.50 (m,

1H), 2.50 (AB, J ) 11 Hz, 1H), 2.96 (AB, J ) 11 Hz, 1H), 4.12

(m, 1H), 6.17 (d, J ) 3 Hz, 1H) ppm; ¹³C NMR (50 MHz, CDCl₃)

δ 17.4 (CH₃), 20.2 (CH₃), 20.4 (CH₃), 27.0 (CH₂), 28.4 (CH₂),

28.8 (CH₂), 30.9 (CH₂), 33.2 (CH₂), 38.2 (CH₂), 38.5 (CH), 38.7

(CH), 43.9 (CH), 45.4 (CH), 47.9 (C), 52.3 (CH), 53.2 (C), 83.2

(CH), 130.3 (CH), 157.4 (C), 204.0 (CdO) ppm; EM (CI-NH₃)

m/e ) 347 (M⁺+ 1, 42), 364 (M⁺+ 18, 100). Anal. Calcd for

C

₂₁H₃₀O₂S: C, 72.78; H, 8.72; S, 9.25. Found: C, 72.50; H,

8.88; S, 9.14.

P a u son -Kh a n d Ad d u ct of 10a /10b w ith Nor bor n a -

d ien e (14). Rea ction Con d ition s A. Following the general

procedure described above, and starting from exo-2-(ethyny-

loxy)-10-(methylthio)bornane, 10a (58 mg, 0.11 mmol) in

hexane (7 mL), and norbornadiene (110 µL, 1.1 mmol), the

reaction was conducted at room temperature for 4 h and

yielded 37 mg (95% yield) of a 60:40 mixture of diastereomers

14 as determined by HPLC (Nucleosil C₁₈, 25 cm, H₂O/MeOH).

Rea ction Con d ition s B. Complex 10b was prepared from

10a (2.31 g, 4.5 mmol) and N-methylmorpholine N-oxide (3.17

g, 27 mmol). For the cyclization, 4.5 mL (45 mmol) of

norbornadiene was used, and the reaction temperature was

-20 °C (overnight). Flash column chromatography (SiO₂/NEt₃,

hexane/AcOEt, 5%) afforded 1.23 g (82%) of 14 as a 96:4 (92%

de) mixture of diastereomers (by HPLC). The major diaste-

reomer could be isolated by standard chromatography as a

white solid. Major diastereomer: mp 58-60 °C; [R]_D) -95.2

(c 7.3, CHCl₃); IR (film) ν_max) 3060, 2960, 1720, 1620, 1105

1

2056, 2002, 1950 cm^-1; H NMR (300 MHz, C₆D₆) δ 0.54 (s,

3H), 0.98 (s, 3H), 0.70-1.50 (m, 5H), 1.55-1.63 (dd, J ) 14, 8

Hz, 1H), 1.60 (s, 3H), 1.88-1.96 (m, 1H), 2.00, 2.75 (AB, J )

11 Hz, 2H), 4.05-4.09 (dd, J ) 8, 4 Hz, 1H), 4.94 (s, 1H) ppm.

Gen er a l P r oced u r es for P a u son -Kh a n d Rea ction :

Ad d u ct of 10a /10b w ith Nor bor n en e (13). Rea ction

Con d ition s A. A solution of the dicobalt hexacarbonyl

complex of exo-2-(ethynyloxy)-10-(methylthio)bornane (10a ) (73

mg, 0.18 mmol) and norbornene (169 mg, 1.8 mmol) in hexane

(10 mL) under CO atmosphere was stirred at room tempera-

ture. When TLC analysis showed complete consumption of

1

cm^-1; H NMR (200 MHz, CDCl₃) δ 0.88 (s, 3H), 1.05 (s, 3H),

1.00-1.85 (m, 7H), 2.07 (s, 3H), 2.19 (m, 1H), 2.48 (AB, J )

11 Hz, 1H), 2.95 (AB, J ) 11 Hz, 1H), 2.58 (m, 2H), 2.88 (s,

1H), 4.13 (m, 1H), 6.14-6.28 (m, 3H) ppm; ¹³C NMR (50 MHz,

CDCl₃) δ 17.5 (CH₃), 20.2 (CH₃), 20.5 (CH₃), 27.0 (CH₂), 30.9

(CH₂), 33.2 (CH₂), 38.3 (CH₂), 41.1 (CH₂), 42.5 (CH), 43.2 (CH),

(57) Longer deprotection times (K₂CO₃/MeOH) may result in the

formation of methyl 3-(10-methylthioisobornyloxy)acrylate as a byprod-

uct: ¹H NMR (200 MHz, CDCl₃) δ 0.88 (s, 3H), 0.98 (s, 3H), 1.10-1.90

(m, 7H), 2.08 (s, 3H), 2.45-2-75 (AB, J ) 11 Hz, 2H), 3.69 (s, 3H),

4.12 (m, 1H), 5.20, 7.60 (AB, J ) 12 Hz, 2H) ppm.

Experimental and Computational Studies

J . Org. Chem., Vol. 63, No. 20, 1998 7051

43.4 (CH), 45.5 (CH), 47.9 (C), 50.3 (CH), 53.3 (C), 83.0 (CH),

130.8 (CH), 136.4 (CH), 138.2 (CH), 157.6 (C), 203.3 (CdO)

ppm; HRMS (EI) calcd for C₂₁H₂₈O₂S 344.1810, found 344.1813.

P a u son -Kh a n d Ad d u ct of 10a /10b w ith Bicyclo[3.2.0]-

h ep t-6-en e (15). Rea ction Con d ition s A. Reaction tem-

perature: 50 °C (2h), 70% yield. Mixture of diastereomers (72:

28, 44% de) as determined by HPLC (Nucleosil C₁₈, 25 cm,

H₂O/MeOH).

Rea ction Con d ition s B. Reaction temperature: -20 °C

(2 days), 91% yield. Mixture of diastereomers (93:7, 86% de)

as determined by HPLC: colorless oil; [R]_D) -65.6 (c 0.4,

CHCl₃, 86% de mixture of diastereomers); IR (film) ν_max) 3060,

2950, 1715, 1615, 1125, 1035 cm^-1; ¹H NMR (200 MHz, CDCl₃)

mixture of diastereomers δ 0.89 (s, 3H), 1.06_minor, 1.09_major(s,

3H), 1.10-1.90 (m, 11H), 2.06_major, 2.09_minor(s, 3H), 2.30-2.62

was added via cannula. A yellow solid immediately formed,

and the reaction temperature was kept at -10 °C for 30 min

and then at 0 °C for 3 h. The reaction crude was poured over

a pH ) 8 (NH₃/NH₄Cl) buffer solution, phases were separated,

and the organic layer was washed with buffer solution until

the aqueous phase did not turn blue. The aqueous phase was

extracted with ether, and the combined organic extracts were

dried (K₂CO₃) and concentrated in vacuo to yield the corre-

ponding silyl enol ether as a colorless oil. Without further

purification, the crude silyl enol ether was dissolved in

anhydrous THF (40 mL) under nitrogen, the solution was

cooled at 0 °C, and a solution of tetrabutylamoniun fluoride

(37 mg, 0.15 mmol) in anhydrous THF (1 mL) was added

dropwise via syringe. When deprotection of the enol ether was

complete (TLC), the reaction mixture was diluted with hexane,

and saturated NH₄Cl solution was added until two clear

phases formed. The organic layer was separated, washed with

brine, dried (K₂CO₃), and concentrated in vacuo. Flash column

chromatography (2% v/v NEt₃pretreated SiO₂, hexane/AcOEt,

1-2.5%) allowed a convenient separation of diastereomers at

this stage and yielded 390 mg of 19 as a white solid (77%)

and 15 mg of the minor isomer (3%). Major diastereomer: mp

55-56 °C; [R]_D) -104.0 (c 1.0, CHCl₃); IR (film) ν_max) 3060,

2980, 1750, 1150, 1080, 720 cm^-1; ¹H NMR (200 MHz, CDCl₃)

δ 0.86 (s, 3H), 1.03 (s, 3H), 0.9 (m, 12H), 2.09 (s, 3H), 2,49,

3.12 (AB, J ) 11 Hz, 2H), 2.73 (m, 1H), 3.18 (m, 1H), 3.80 (dd,

J ) 3, 7 Hz, 1H), 4.01 (dd, J ) 2, 11 Hz, 1H), 6.13 (m, 2H)

ppm; ¹³C NMR (50 MHz, CDCl₃) δ 17.5 (CH₃), 19.1 (CH₃), 20.2

(CH₃), 20.8 (CH₃), 27.0 (CH₂), 30.6 (CH₂), 33.5 (CH₂), 39.1

(CH₂), 41.0 (CH), 43.4 (CH), 44.0 (CH), 44.6 (CH₂), 45.2 (CH),

45.9 (CH), 48.2 (CH), 85.3 (CH), 90.2 (CH), 136.2 (CH), 137.6

(m, 5H), 2.98 (d, J ) 10 Hz, 1H), 4.09-4.18 (m, 1H), 6.42_major

,

6.44_minor(d, J ) 3 Hz, 1H) ppm; ¹³C NMR (50 MHz, CDCl₃)

major diastereomer δ 17.6 (CH₃), 20.3 (CH₃), 20.5 (CH₃), 24.9

(CH₂), 27.1 (CH₂), 31.0 (CH₂), 32.3 (CH₂), 32.9 (CH₂), 33.3

(CH₂), 38.3 (CH₂), 38.3 (CH), 38.7 (CH), 44.2 (CH), 45.6 (CH),

46.9 (CH), 48.1 (C), 53.5 (C), 83.6 (CH), 131.0 (CH), 156.1 (C)

204.5 (CdO) ppm; HRMS(EI) calcd for C₂₁H₃₀O₂S 346.1966,

found 346.1963.

P a u son -Ka n d Ad d u ct of 11a /11b w ith Nor bor n a d ien e

(16). Starting from dichloroolefin 8 (0.86 g, 2.45 mmol) and

following the general procedure described for the preparation

of 10a /10b, complex 11a contamined by minor amounts of 11b

(1.00 g, 72% yield) was obtained as a red oil. The Pauson-

Khand reaction was then performed according to conditions

B: reaction temperature -20 °C (overnight), 61% yield.

Mixture of diastereomers (96:4, 92% de) as determined by

HPLC (Nucleosil C₁₈, 25 cm, MeOH/H₂O): colorless oil; [R]_D

) -83.8 (c 1.1, CHCl₃, 92% de mixture of diastereomers); IR

(film) ν_max) 3060, 2950, 1710, 1615, 1460, 1110 cm^-1; ¹H NMR

(200 MHz, CDCl₃) mixture of diastereomers δ 0.87 (s, 3H), 0.91

(s, 9H), 1.04 (s, 3H), 1.00-1.85 (m, 9H), 2.17 (d, J ) 5 Hz,

1H), 2.29-2.68 (m, 5H), 2.83-3.02 (m, 2H), 4.08-4.20 (m, 1H),

6.14-6.33 (m, 3H) ppm; ¹³C NMR (50 MHz, CDCl₃) mixture

of diastereomers δ 20.7 (CH₃), 20.9 (CH₃), 27.5 (CH₂), 29.3

(3CH₃), 31.3 (CH₂), 32.9 (C), 33.7 (CH₂), 38.8 (CH₂), 41.5 (CH₂),

43.0 (CH), 43.7 (CH), 43.9 (CH), 46.1 (CH), 48.3 (C), 49.7

(CH2), 50.8 (CH), 53.8 (C), 83.5 (CH), 130.6_minor, 131.0_major(CH),

136.9 (CH), 138.7 (CH), 158.5 (C) ppm; HRMS (EI) calcd for

(CH), 213.9 (CdO) ppm; EM (DIP-CI-NH₃) m/e ) 361 (M⁺

+

1, 82), 378 (M⁺+ 18, 100), 395 (M⁺+ 35, 29). Anal. Calcd

for C₂₂H₃₂O₂S: C, 73.28; H, 8.94; S, 8.89. Found: C, 73.68;

H, 9.28; S, 8.88. Minor diastereomer: ¹H NMR (200 MHz,

CDCl₃) δ 0.83 (s, 3H), 1.06 (s, 3H), 1.00-1.85 (m, 11H), 1.33

(d, J ) 6 Hz, 3H), 2.01-2.20 (m, 1H), 2.11 (s, 3H), 2.38-2.89

(AB, J ) 12 Hz, 2H), 2.76 (s, 1H), 3.15 (s, 1H), 3.58-3.68 (dd,

J ) 3, 7 Hz, 1H), 4.08-4.18 (dd, J ) 2, 11 Hz, 1H), 6.08-6.20

(m, 2H) ppm; ¹³C NMR (75 MHz, CDCl₃) δ 17.3 (CH₃), 19.2

(CH₃), 20.2 (CH₃), 20.6 (CH₃), 27.1 (CH₂), 31.1 (CH₂), 34.1

(CH₂), 40.1 (CH₂), 40.5 (CH), 43.3 (CH), 44.3 (CH), 44.6 (CH₂),

45.1 (CH), 45.8 (CH), 47.7 (C), 49.0 (CH), 53,5 (C), 85.3 (CH),

91.5 (CH), 136.9 (CH), 137.7 (CH), 215.2 (CdO) ppm.

C

₂₅H₃₆O₂S 400.2436, found 400.2438.

P a u son -Ka n d Ad d u ct of 12a /12b w ith Nor bor n a d ien e

(-)-(1R,2R,4R,5R,6R,7S)-4-[10-(Meth ylth io)isobor n eoxy]-

5-bu tyltr icyclo[5.2.1.0^2,6]d ec-8-en -3-on e (20) was prepared

according to the general procedure described above for conju-

gate additions: Copper(I) cyanide (0.42 g, 4.7 mol) in ether

(50 mL), n-BuLi 1.47 M in hexane (6.5 mL, 9.5 mmol), 14

(diasteromeric mixture of 92% de) (0.82 g, 2.4 mmol), and

TMSCl (1.2 mL, 9.5 mmol) were used. Deprotection of the silyl

enol ether was performed with tetrabutylamonium fluoride (63

mg, 0.24 mmol) in THF (1 mL). Flash column chromatography

(2% v/v NEt₃pretreated SiO₂, hexane/ether, 1%) afforded 67

mg of a mixture of diastereomers and 0.76 g of the major

diastereomer 20 as a colorless oil (81% overall yield). Major

(17). Starting from dichloroolefin 9 (0.35 g, 0.98 mmol) and

following the general procedure described for the preparation

of 10a /10b, complex 12a contamined with minor amounts of

12b (0.40 g, 71% yield) was obtained as a red oil. The Pauson-

Khand reaction was then performed according to conditions

B: reaction temperature -20 °C (overnight), 65% yield.

Mixture of diastereomers (93:7, 86% de) as determined by

HPLC (Nucleosil C₁₈, 25 cm, MeOH/H₂O): [R]_D) -58.3 (c 2.0,

CHCl₃, 86% de mixture of diastereomers); IR (film) ν_max

)

3050, 2950, 1715, 1615, 1130 cm^-1; ¹H NMR (200 MHz, CDCl₃)

mixture of diastereomers δ 0.97 (s, 3H), 1.08 (s, 3H), 1.00-

1.90 (m, 9H), 2.11 (d, J ) 5 Hz, 1H), 2.44 (m, 1H), 2.57 (s,

1H), 2.87 (s, 1H), 2.87-3.54 (AB, J ) 12 Hz, 2H), 4.09 (m,

1H), 6.10-6.30 (m, 3H), 7.05-7.50 (m, 5H) ppm; ¹³C NMR (50

MHz, CDCl₃) mixture of diastereomer δ 20.7 (CH₃), 21.0 (CH₃),

27.5 (CH₂), 31.4 (CH₂), 33.9 (CH₂), 38.8 (CH₂), 41.6 (CH₂), 43.0

(CH), 43.7 (CH), 48.8 (C), 46.0 (CH), 48.7 (C), 50.8 (CH), 53.6

(C), 82.9_major, 83.7_minor(CH), 126.2 (CH), 129.1 (CH), 130.6 (CH),

139.9 (CH), 136.9 (CH), 137.8 (C), 138.7 (CH), 158.0 (C) ppm;

HRMS (EI) calcd for C₂₆H₃₀O₂S 406.1966, found 406.1970.

Gen er a l P r oced u r e for Con ju ga te Ad d ition s: (-)-

(1R,2R,4R,5R,6R, 7S)-4-[10-(Meth ylth ioiso)bor n eoxy]-5-

m et h ylt r icyclo[5.2.1.0^2,6]d ec-8-en -3-on e (19). Copper(I)

iodide (0.55 g, 2.9 mmol) and 30 mL of ether were placed in a

round-bottomed flask under nitrogen. The resulting slurry

was cooled at -50 °C, and 1.6 M MeLi (3.4 mL, 5.4 mmol) was

added via syringe. The mixture was stirred at -50 °C for 30

min and then allowed to gradually warm to -10 °C. A solution

of 14 (diastereomeric mixture of 92% de) (0.5 g, 1.45 mmol)

and chlorotrimethylsilane (0.4 mL, 2.9 mmol) in ether (5 mL)

diastereomer: [R]_D) -107.1 (c 1.3, CHCl₃); IR (film) ν_max

)

3060, 2960, 1750, 1140, 1110, 1070 cm^-1; ¹H NMR (200 MHz,

CDCl₃) δ 0.86 (s, 3H), 0.95 (t, J ) 6 Hz, 3H), 1.03 (s, 3H), 1.10-

1.90 (m), 2.08 (s, 3H), 2.14 (d, J ) 10 Hz, 1H), 2.47-3.14 (AB,

J ) 12 Hz, 2H), 2.72 (s, 1H), 3.17 (s, 1H), 3.78-3.84 (dd, J )

3, 7 Hz, 1H), 4.01-4.08 (m, 1H), 6.08-6.20 (m, 2H) ppm; ¹³C

NMR (CDCl₃) δ 14.0 (CH₃), 17.5 (CH₃), 20.2 (CH₃), 20.8 (CH₃),

22.9 (CH₂), 29.4 (CH₂), 30.7 (CH₂), 33.5 (CH₂), 34.4 (CH₂), 39.0

(CH₂), 42.3 (CH), 43.5 (CH), 44.4 (CH₂), 45.3 (CH), 45.6 (CH),

47.4 (CH), 48.6 (C), 49.3 (CH), 53.3 (C), 85.4 (CH), 89.0 (CH),

136.8 (CH), 137.7 (CH), 214.3 (CdO) ppm; EM (DIP-CI-NH₃)

m/e ) 403 (M⁺+ 1, 100), 420 (M⁺+ 18, 25); HRMS (EI) calcd

for C₂₅H₃₈O₂S 402.2592, found 402.2587.

(-)-(1R,2R,4R,5R,6R,7S)-4-[10-(Meth ylth io)isobor n eoxy]-

5-p h en yltr icyclo[5.2.1.0^2,6]d ec-8-en -3-on e (18) was pre-

pared according to the general procedure described above for

conjugate additions: CuBr-SMe₂(0.25 g, 1.23 mol) in ether

(6 mL), PhLi 1.6 M in ether (1.5 mL, 2.46 mmol), 14 (dias-

teromeric mixture of 92%) (138 mg, 0.41 mmol), and TMSCl

7052 J . Org. Chem., Vol. 63, No. 20, 1998

Verdaguer et al.

(153 µL, 1.2 mmol) in ether (2 mL) were used. Cuprate

formation was run at 0 °C for 30 min. Substrate addition was

performed at -78 °C, and then the reaction mixture was

stirred at room temperature for 8 h. Addition of tetrabutyla-

monium fluoride (10 mg, 0.41 mmol) in THF (1 mL) at 0 °C

resulted in complete deprotection of the silyl group. Flash

column chromatography (2% v/v NEt₃pretreated SiO₂, hexane/

ether, 1-5%) afforded 5 mg of the minor diastereomer and 92

mg of the major diastereomer 18 as a colorless oil (66% global

yield). Major diastereomer: [R]_D) -57.6 (c 1.0, CHCl₃); IR

(250 mg, 2.5 mmol) in CH₂Cl₂(3 mL) was added MeAlCl₂1 M

in hexane (0.53 mL, 0.53 mmol) via syringe. Heating the

reaction mixture at 55 °C for 1.5 h resulted in complete

consumption of the starting material (TLC). The reaction

mixture was then cooled at room temperature, quenched with

saturated K₂CO₃solution, filtered through Celite, and dried

with MgSO₄. The CH₂Cl₂solution of (-)-(S)-4-methyl-2-

cyclopentenone (24) was diluted with MeOH (6 mL) and

hydrogenated using Pd/C (10%) at atmospheric pressure (H₂,

balloon). After 2 h, the complete disappearance of 24 could

be observed by TLC. The catalyst was then removed by

filtration through Celite, and the filtrate was treated with

AcONa (0.3 g), semicarbazide hydrochloride (0.2 g), and water

(2 mL). The resulting solution was heated at 80 °C for 1.5 h;

organic solvents were then evaporated in vacuo, and the

remaining aqueous phase was extracted with chloroform. The

chloroform extract was dried (Na₂CO₃) and evaporated in

vacuo. Flash column chromatography of the residue (SiO₂,

CHCl₃/MeOH) yielded 55 mg (72%) of the corresponding

semicarbazone 25 as a white solid: mp 170 °C (lit.³⁴mp 173-

174 °C); [R]_D) +38.5 (c 1.0, CHCl₃) (lit.³⁵[R]_D) +41.4 (c 0.7,

1

(film) ν_max) 3060, 2960, 1745, 1140, 690 cm^-1; H NMR (200

MHz, CDCl₃) δ 0.75 (s, 3H), 0.82 (s, 3H), 0.80-1.00 (m, 2H),

1.20-1.55 (m, 7H), 2.09 (s, 3H), 2.13-2.60 (m, 4H), 2.78 (s,

1H), 3.09 (d, J ) 11 Hz, 1H), 3.24-3.33 (m, 2H), 4.60-4.67

(dd, J ) 2, 12 Hz, 1H), 6.04-6.18 (m, 2H), 7.23-7.45 (m, 5H)

ppm; ¹³C NMR (50 MHz, CDCl₃) δ 17.4 (CH₃), 20.0 (CH₃), 20.6

(CH₃), 26.8 (CH₂), 30.5 (CH₂), 33.6 (CH₂), 39.0 (CH₂), 43.8 (CH),

44.3 (CH₂), 44.9 (2 CH), 45.9 (CH), 47.4 (C), 49.0 (CH), 53.2

(C), 53.6 (CH), 86.0 (CH), 91.6 (CH), 126.8 (CH), 128.0 (2 CH),

128.5 (2 CH), 137.0 (CH), 137.7 (CH), 142.9 (C), 212.6 (CdO)

ppm; EM (DIP-CI-NH₃) m/e ) 423 (M⁺+ 1, 100), 440 (M⁺

+

18, 6); HRMS (EI) calcd for

422.2282.

C

₂₇H₃₄O₂S 422.2279, found

CHCl₃)); IR (KBr) ν_max) 3460, 2960, 1690, 1220, 1080 cm^-1

;

¹H NMR (200 MHz, CDCl₃) δ 1.06 (t, J ) 6 Hz, 3H), 1.20-

2.60 (m, 7H), 5.60 (s broad, 2H), 8.0 (NH, 1H) ppm; ¹³C NMR

(50 MHz, CDCl₃) δ 19.5 (CH₃), 20.0 (CH₃), 27.2 (CH₂), 32.7

(CH₂), 32.8 (CH₂), 33.0 (CH₃), 33,4 (CH), 35.5 (CH₂), 41.3 (CH₂)

(-)-(1R,2S,5S,6S,7S)-5-Met h ylt r icyclo[5.2.1.0^2,6]d ec-8-

en -3-on e (22).^25aPowder samarium metal (0.39 g, 2.6 mmol)

and 1,2-diiodoetane (0.67 g, 2.4 mmol) were placed in an oven-

dried Schlenk flask under argon atmosphere. Anhydrous THF

(5 mL) was added via syringe, and the resulting slurry was

stirred at room temperature for 1-2 h. This resulted in a deep

blue solution of SmI₂. The reaction mixture was cooled at 0

°C, and a solution of 19 (290 mg, 0.8 mmol) in THF/MeOH 2/1

(2.5 mL) was added. The mixture was stirred at 0 °C for 5

min and then diluted with anhydrous THF first and finally

with wet THF in order to deactivate the excess samarium

diiodide.⁵⁸The mixture was diluted with hexane and treated

with saturated K₂CO₃. The aqueous phase was extracted with

ether, and the combined organic phases were dried (MgSO₄)

and evaporated in vacuo. Flash column chromatography (2%

v/v NEt₃pretreated SiO₂, hexane/AcOEt, 1.5%) afforded 153

mg (95% yield) of 10-(methylthio)isoborneol and 122 mg (93%)

of (-)-(1R,2S,5S,6S,7S)-5-methyltricyclo[5.2.1.0^2,6]dec-8-en-3-

one (22) as a colorless oil. [R]_D) -121.6 (c 3.6, CHCl₃); IR

ppm; EM (DIP-CI-NH₃) m/e ) 156 (M⁺+ 1, 50), 173 (M⁺

18, 100), 190 (M⁺+ 35, 12).

+

(-)-(S)-4-Bu tyl-2-cyclop en ten on e (26).⁵⁹In a round-

bottomed flask under nitrogen were placed 23 (130 mg, 0.63

mmol), maleic anhydride (0.3 g, 3.1 mmol), and CH₂Cl₂(4 mL).

To this solution was added MeAlCl₂(1 M in hexane, 0.7 mL,

0.7 mmol) via syringe. The reaction mixture was then heated

at 55 °C for 2 h, cooled to room temperature, quenched with

saturated K₂CO₃solution, and filtered through Celite. The

residue was purified by flash column chromatography (2% v/v

NEt₃pretreated SiO₂, hexane/ether) and afforded 85 mg (90%)

of (-)-(S)-4-butyl-2-cyclopentenone as a colorless oil: [R]_D

-160.8 (c 1.7, CHCl₃); IR (film) ν_max) 2960, 1710, 1585, 1180

cm^-^{1 1}H NMR (200 MHz, CDCl₃) δ 0.91 (t, J ) 7 Hz, 3H),

)

;

1.15-1.70 (m, 6H), 1.95-2.60 (ABX, J _ab) 19 Hz, J _ax) 2 Hz,

J _bx) 6 Hz, 2H), 2.85-3.00 (m, 1H), 6.14, 7.64 (AMX, J _am) 5

Hz, J _ax,mx) 2 Hz, 2H) ppm; ¹³C NMR (50 MHz, CDCl₃) δ 13.9

(CH₃), 22.6 (CH₂), 29.7 (CH₂), 34.4 (CH₂), 41.0 (CH₂), 41.4

(CH₂), 133.5 (CH), 168.8 (CH), 210.2 (CdO) ppm.

1

(film) ν_max) 3070, 2970, 1740, 700 cm^-1; H NMR (200 MHz,

CDCl₃) δ 1.19 (d, J ) 6 Hz, 3H), 1.15-1.25 (m, 1H), 1.38-

1.48 (m, 1H), 1.75-1.98 (m, 2H), 2.15-2.70 (ABX, J _ab) 18

Hz, J _ax) 7 Hz, J _bx) 10 Hz, 2H), 2.28-2.35 (m, 1H), 2.78 (s,

1H), 3.09 (s, 1H), 6.10-6.22 (m, 2H) ppm; ¹³C NMR (50 MHz,

CDCl₃) δ 22.3 (CH₃), 33.9 (CH), 44.5 (CH₂), 45.2 (CH), 47.3

(CH), 51.2 (CH), 51.3 (CH₂), 55.2 (CH), 137.4 (CH), 138.4 (CH),

(-)-2,3-Bu ta n ed iol Aceta l of th e (R)-3-Bu tylcyclop en -

ten on e (27). A solution of 26 (78 mg, 0.56 mmol) in MeOH

(8 mL) was hydrogenated over 10% Pd/C (13 mg) at atmo-

spheric pressure (H₂, balloon). The hydrogenated mixture was

filtered through Celite, and methanol was evaporated. The

resulting residue was placed in a micro Dean-Stark apparatus

along with (-)-2,3-butanediol (156 mg, 1.68 mmol), a catalytic

amount of camphorsulfonic acid, and toluene. The mixture

was heated at reflux for 6 h. The solvent was evaporated and

the crude purified by flash column chromatography (2% v/v

NEt₃pretreated SiO₂, hexane/ether, 2%) to give 87 mg of acetal

27 (73% yield) as a colorless oil: IR (film) ν_max) 2930, 1450,

218.7 (CdO) ppm; EM (DIP-CI-NH₃) m/e ) 114 (C₆H₈O⁺

+

18, 37), 163 (M⁺+ 1, 1), 180 (M⁺+ 18, 100), 197 (M⁺+ 35,

18).

(-)-(1R,2S,5S,6S,7S)-5-Bu tyltr icyclo[5.2.1.0^2,6]d ec-8-en -

3-on e (23).³³Reductive cleavage of 20 (0.3 g, 0.76 mmol) with

SmI₂was performed following the same procedure as for the

preparation of 22 and yielded 153 mg (99%) of 10-(methylthio)-

isoborneol and 154 mg of 23 (88%) as a colorless oil: [R]_D

)

1

1375, 1310, 1110 cm^-1; H NMR (200 MHz, CDCl₃) δ 0.87 (t,

-122.0 (c 1.3, CHCl₃); IR (film) ν_max) 3060, 2920, 1735, 1460,

3H), 1.00-1.55 (m, 14H), 1.65-2.08 (m, 5H), 3.47-3.62 (m,

2H) ppm; ¹³C NMR (75 MHz, CDCl₃) δ 14.0 (CH₃), 16.8 (CH₃),

19.9 (CH₃), 22.7 (CH₂), 30.4 (CH₂), 30.5 (CH₂), 35.5 (CH₂), 37.9

(CH), 38.0 (CH₂), 44.9 (CH₂), 78.0 (CH), 78.1 (CH), 117.0 (C)

ppm; EM (DIP-CI-NH₃) m/e ) 213 (M⁺+ 1, 91).

1

1190, 690 cm^-1; H NMR (200 MHz, CDCl₃) δ 0.92 (t, J ) 6

Hz, 3H), 1.10-1.80 (m, 9H), 1.94-2.01 (dd, J ) 5, 8 Hz, 1H),

2.15-2.26 (m, 1H), 2.29 (d, J ) 9 Hz), 2.54-2.67 (dd, J ) 18,

8 Hz, 1H), 2.76 (s, 3H), 3.09 (s, 1H), 6.10-6.27 (m, 2H) ppm;

¹³C NMR (50 MHz, CDCl₃) δ 14.0 (CH₃), 22.8 (CH₂), 30.0 (CH₂),

37.5 (CH₂), 39.3 (CH₂), 44.5 (CH₂), 45.2 (CH), 48.0 (CH), 49.6

(CH), 49.8 (CH₂), 54.9 (CH), 137.4 (CH), 138.4 (CH) ppm.

Ack n ow led gm en t. Financial support from DGICYT

(PB95-0265) and from CIRIT (1996SGR-00013) is grate-

fully acknowledged.

(+)-(R)-3-Meth ylcyclop en ta n on e Sem ica r ba zon e 25.³⁴

To a solution of 23 (80 mg, 0.49 mmol) and maleic anhydride

J O9809985

(58) Extended reduction times may result in overreduction of 22.

In this case, the corresponding alcohol can be reoxidized with PCC.

(59) Ackroyd, J .; Karpf, M.; Dreiding, A. S. Helv. Chim. Acta 1985,

68, 338-344.

Article Doi

Camphor-derived, chelating auxiliaries for the highly diastereoselective intermolecular Pauson-Khand reaction: Experimental and computational studies

DOI: 10.1021/jo9809985

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo9809985

Full text of DOI:10.1021/jo9809985

Products guided by the article

R&D Labs maybe for 214599-33-4

Relevant to this article

Hot Product