Carbohydrate Phosphinites as Practical Ligands in Asymmetric Catalysis: Electronic Effects and Dependence of Backbone Chirality in Rh-Catalyzed Asymmetric Hydrogenations. Synthesis of R- or S-Amino Acids Using Natural Sugars as Ligand Precursors

6012

J . Org. Chem. 1997, 62, 6012-6028

Ca r boh yd r a te P h osp h in ites a s P r a ctica l Liga n d s in Asym m etr ic

Ca ta lysis: Electr on ic Effects a n d Dep en d en ce of Ba ck bon e

Ch ir a lity in Rh -Ca ta lyzed Asym m etr ic Hyd r ogen a tion s. Syn th esis

of R- or S-Am in o Acid s Usin g Na tu r a l Su ga r s a s Liga n d P r ecu r sor s

T. V. RajanBabu,* Timothy A. Ayers, Gary A. Halliday, Kimberly K. You, and

J oseph C. Calabrese

Department of Chemistry, The Ohio State University, 100 W. 18th Avenue, Columbus, Ohio 43210, and

DuPont Central Research and Development, Experimental Station, Wilmington, Delaware 19880

Received May 19, 1997^X

Vicinal diarylphosphinites derived from carbohydrates are excellent ligands for the Rh(I)-catalyzed

enantioselective asymmetric hydrogenation of dehydroamino acid derivatives, producing the highest

enantioselectivity of any ligands directly prepared from natural products. The enantioselectivity

can be enhanced by the appropriate choice of substituents on the aromatic rings of the phosphinites.

For example, the use of phosphinites with electron-donating bis(3,5-dimethylphenyl) groups on

phosphorus provides ee’s up to 99% for a wide range of amino acids including some with easily

removable N-protecting groups. Electron-withdrawing aryl substituents, on the other hand, decrease

the enantioselectivity. Sense of chiral induction in the amino acid product depends on the relative

juxtaposition of the vicinal diphosphinites on a given sugar backbone. When readily available

D-glucopyranosides are used as the starting sugars, 2,3-phosphinites give the S-amino acids and

3,4-phosphinites give the R-amino acids. In the case of aromatic and heteroaromatic amino acids,

enantioselectivities > 95% are consistently obtained. Practical considerations such as the ease of

ligand synthesis, rates of reactions, catalyst turnover, and scope and limitations in terms of

substrates are discussed. A possible explanation for the enhancement of enantioselectivity by

electron-rich phosphinites is offered.

In tr od u ction

these instances model building can be a useful tool for

the design of better catalysts, provided there is sufficient

flexibility in the synthesis of the ligands. For other

reactions such as the Rh(I)-catalyzed hydrogenation of

acetamidoacrylates^9-11or the Ni(0)-catalyzed hydrocya-

nation of olefins,¹²where the relative reactivities of

transiently-formed, often undetectable intermediates

dictate the overall stereochemical outcome, the predictive

value of such models is tenuous at best. Recently,

increasing number of examples of electronic tuning of

asymmetric catalysts are beginning to appear in the

literature.^12-19A clear understanding of the origin of the

Catalytic asymmetric synthesis using organometallic

reagents is currently one of the most active areas of

research in organic chemistry.¹Despite the enormous

progress in this area, our inability to predict, a priori,

the type of ligands required to achieve useful levels of

enantioselectivity severely limits the rational develop-

ment of highly selective catalysts. With a few notable

exceptions, models which adequately predict the stereo-

chemical outcomes rely on spatial orientation of groups

in the chiral environment around the metal.²Such

models are especially useful for designing catalysts for

reactions where the precise structure of the catalytically

relevant intermediate is known and its absolute config-

uration can be related to that of the final product.^3-8In

(7) J ohnson, R. A.; Sharpless, K. B. Catalytic Asymmetric Epoxi-

dation of Allylic Alcohols. In Catalytic Asymmetric Synthesis; Ojima,

I., Ed.; VCH Publishers: New York, 1993; p 101.

(8) Togni, A.; Burckhardt, U.; Gramlich, V.; Pregosin, P. S.; Salz-

mann, R. J . Am. Chem. Soc. 1996, 118, 1031 and references cited

therein.

(9) Landis, C. R.; Halpern, J . J . Am. Chem. Soc. 1987, 109, 1746.

(10) Giovannetti, J . S.; Kelly, C. M.; Landis, C. R. J . Am. Chem.

Soc. 1993, 115, 4040.

(11) Takaya, H.; Ohta, T.; Noyori, R. Asymmetric Hydrogenation.

In Catalytic Asymmetric Synthesis; Ojima, I., Ed.; VCH Publishers:

New York, 1993; p 1.

^XAbstract published in Advance ACS Abstracts, August 1, 1997.

(1) For recent monographs and reviews see: (a) Noyori, R. Asym-

metric Catalysis in Organic Synthesis; J ohn Wiley: New York, 1994.

(b) Ojima, I. Asymmetric Synthesis; Verlag: New York, 1993. (c) Collins,

A. N.; Sheldrake, G. N.; Crosby, J . Chirality in Industry; J ohn Wiley:

New York, 1992. (d) Bosnich, B. Asymmetric Catalysis; Martinus

Nijhoff Publishers: Dordrecht, 1986. (e) Brunner, H. Synthesis 1988,

645. (f) Brown, J . M.; Davies, S. G. Nature 1989, 342, 631. (h) Pfaltz,

A. Chimia 1990, 44, 202. (i) Nugent, W. A.; RajanBabu, T. V.; Burk,

M. J . Science 1993, 259, 479.

(2) For some recent examples see: (a) Trost, B. M.; Van Vranken,

D. L. Chem. Rev. 1996, 96, 395 and references cited therein. (b) Bao,

J .; Wulff, W. D.; Rheingold, A. L. J . Am. Chem. Soc. 1993, 115, 3814.

(c) Morken, J . P.; Didiuk, M. T.; Hoveyda, A. H. J . Am. Chem. Soc.

1993, 115, 6997. (d) Seebach, D.; Devaquet, E.; Ernst, A.; Hayakawa,

M.; Ku¨hnle, F. N. M.; Schweizer, W. B.; Weber, B. Helv. Chim. Acta

1995, 78, 1636 and references cited therein.

(3) Mackenzie, P. B.; Whelan, J .; Bosnich, B. J . Am. Chem. Soc.

1985, 107, 2046.

(4) Consiglio, G.; Waymouth, R. M. Chem. Rev. 1989, 89, 257.

(5) (a) Frost, C. G.; Howarth, J .; Williams, J . M. J . Tetrahedron:

Asymmetry 1992, 3, 1089; (b) Sprinz, J .; Kiefer, M.; Helmchen, G.;

Reggelin, M.; Huttner, G.; Walter, O.; Zsolnai, L. Tetrahedron Lett.

1994, 35, 1523.

(12) Casalnuovo, A. L.; RajanBabu, T. V.; Ayers, T. A.; Warren, T.

H. J . Am. Chem. Soc. 1994, 116, 9869.

(13) See for example: (a) J acobsen, E. N.; Zhang, W.; Gu¨ler, M. L.

J . Am. Chem. Soc. 1991, 113, 6703. (b) Nishiyama, H.; Yamaguchi, S.;

Kondo, M.; Itoh, K. J . Org. Chem. 1992, 57, 4306. (c) Schnyder, A.;

Hintermann, L.; Togni, A. Angew. Chem., Int. Ed. Engl. 1995, 34, 931.

(d) Frost, C. G.; Howarth, J .; Williams, J . M. J . Tetrahedron: Asym-

metry 1992, 3, 1089. (e) Togni, A.; Breutel, C.; Schnyder, A.; Spidler,

F.; Landert, H.; Tijani, A. J . Am. Chem. Soc. 1994, 116, 4062. (f) Sakai,

N.; Mano, S.; Nozaki, K.; Takaya, H. J . Am. Chem. Soc. 1993, 115,

7033. (g) von Matt, P.; Lloyd-J ones, G. C.; Minidis, A. B. E.; Pfaltz, A.;

Macko, L.; Neuburger, M.; Zehnder, M.; Ru¨egger, H.; Pregosin, P. S.

Helv. Chim. Acta 1995, 78, 265. (h) Rieck, H.; Helmchen, G. Angew.

Chem., Int. Ed. Engl. 1995, 34, 2687.

(14) Inoguchi, K.; Sakuraba, S.; Achiwa, K. Synlett 1992, 169.

(15) RajanBabu, T. V.; Casalnuovo, A. L. J . Am. Chem. Soc. 1992,

114, 6265.

(6) Trost, B. M.; Van Vranken, D. L.; Bingel, C. J . Am. Chem. Soc.

1992, 114, 9327.

(16) RajanBabu, T. V.; Ayers, T. A.; Casalnuovo, A. L. J . Am. Chem.

Soc. 1994, 116, 4101.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6013

pronounced dependence of enantioselectivities on the

electronic properties of ligands is still lacking, and this

further impedes the de novo design of useful ligands.

Such inadequacies notwithstanding, steric and electronic

tuning of ligands has been successfully employed in

developing efficient asymmetric reactions.

F igu r e 1. Prototypical sugar phosphinite ligands giving

opposite enantioselectivities.

from phenyl â-^D-glucopyranoside (1) gave the (S)-2-

arylpropiononitrile as the major isomer whereas the 3,4-

bis-O-(diarylphosphinite)s from methyl R-^D-fructofura-

noside (2) gave the opposite enantiomer (Figure 1). The

enantioselectivity of the reaction increases dramatically

when ^D-glucose-derived 2,3-bis(diarylphosphinite)s (Fig-

ure 1) contain electron-withdrawing P-aryl substituents,

and enantiomeric excesses (ee) up to 91% were realized

for the hydrocyanation of 2-methoxy-6-vinylnaphthalene,

a precursor for the antiinflammatory drug, naproxen (eq

2, Ar ) 2-(6-methoxynaphthyl). In a detailed mechanistic

investigation that involved kinetic as well as deuterium

labeling studies we have provided convincing evidence

that the amplification of enantioselectivity by the electron-

withdrawing P-aryl sustituents is related to an increase

in the rate of reductive elimination of the product from

the penultimate (η³-benzyl)nickel cyanide intermediate

(eq 2).¹²At the time of our initial report¹⁵dealing with

the hydrocyanation of vinyl arenes, with a few notable

In light of the foregoing discussion, it is apparent that

an empirical approach to ligand design remains a power-

ful option, especially when dealing with an asymmetric

reaction for which the details of the mechanism and of

the origin of asymmetric induction are poorly understood.

In practical terms, such an approach is limited only by

the ability to make systematic structural changes in a

ligand scaffolding to produce the broadest possible elec-

tronic and stereochemical diversity. We were first con-

fronted with such a situation in our attempts to find an

asymmetric variant for the Ni(0)-catalyzed hydrocyana-

tion of vinylarenes, for which no synthetically useful

precedent existed.^20-22Initial experiments with ligands

(including BINAP, DIOP, CHIRAPHOS, BPPFA, DU-

PHOS)²³which were known to give excellent selectivities

in other asymmetric reactions resulted in disappointingly

low reactivity and/or selectivity in the hydrocyanation

reaction. A careful analysis of the various syntheses of

these and a number of other well-known ligand systems

quickly revealed that making structural changes even

within a given scaffolding could be prohibitively time

consuming for two reasons: (a) the number of synthetic

steps involved; (b) need of a resolution process which in

many cases has to be optimized for each substituted

derivative. Thus we embarked on a search for easily

tunable ligands which could be prepared in a few steps

from readily available natural products. In reviewing the

literature we were particularly attracted to chiral vicinal

diaryl and dialkyl phosphinites because of the inherent

simplicity in their synthesis (eq 1) from carbohydrates

and amino acids.²⁴In addition, the use of diarylchloro-

phosphine as the source of phosphorus in this modular

construction allows for easy manipulation of the elec-

tronic and steric properties of the chelating atom(s). In

the study of the Ni(0)-catalyzed asymmetric hydrocya-

nation of vinyl arenes (eq 2) using carbohydrate-derived

phosphinites, we have since demonstrated that ligand

optimization is indeed a powerful strategy to develop new

asymmetric processes.^12,19During this study, we discov-

ered two crucial control elements: (a) chirality of the

sugar backbone; (b) the electronic properties of the

chelating atoms. Thus, 2,3-bis-O-(diarylphosphinite)s

exceptions,^13a,14electronic tuning of ligands had seldom

been employed as a control element in asymmetric

catalysis. Since the sugar phosphinite system was

especially amenable to electronic tuning, we wondered

how a reaction where oxidative addition to the metal

center is a key step would respond to electronic effects

in the ligand. For this we chose the well-known Rh-

catalyzed asymmetric hydrogenation of R-(acylamino)acryl-

ic acid derivatives (eq 3), a reaction which has been

studied extensively because of its great synthetic poten-

tial for the synthesis of amino acids.^9,25-29In sharp

contrast to the Ni(0)-catalyzed asymmetric hydrocyana-

(17) RajanBabu, T. V.; Ayers, T. A. Tetrahedron Lett. 1994, 35, 4295.

(18) RajanBabu, T. V.; Casalnuovo, A. L. Pure Appl. Chem. 1994,

66, 1535.

(19) RajanBabu, T. V.; Casalnuovo, A. L. J . Am. Chem. Soc. 1996,

118, 6325.

(20) Elmes, P. S.; J ackson, W. R. Aust. J . Chem. 1982, 35, 2041.

(21) Hodgson, M.; Parker, D.; Taylor, R. J .; Ferguson, G. Organo-

metallics 1988, 7, 1761.

(22) Baker, M. J .; Pringle, P. G. J . Chem. Soc., Chem. Commun.

1991, 1292.

tion reaction, electron-donating substituents on the P-

aryl groups of vicinal phosphinites enhance the enanti-

oselectivity of the Rh(I)-catalyzed asymmetric hydrog-

(23) See reference 1b for the abbreviations and the structures of the

ligands.

(25) Knowles, W. S. Acc. Chem. Res. 1983, 16, 106.

(26) Koenig, K. E. The Applicability of Asymmetric Homogeneous

Catalytic Hydrogenation. In Aymmetric Synthesis; Morrison, J . D., Ed.;

Academic Press: Orlando, 1985; Vol. 5; p 71.

(27) Chan, A. S. C.; Pluth, J . J .; Halpern, J . J . Am. Chem. Soc. 1980,

102, 5952.

(28) Brown, J . M. Chem. Soc. Rev. 1993, 25.

(29) Burk, M. J .; Feaster, J . E.; Nugent, W. A.; Harlow, R. L. J .

Am. Chem. Soc. 1993, 115, 10125.

(24) The use of carbohydrate phosphinites for hydrogenation of

dehydroamino acids have been known since 1978: (a) Cullen, W. R.;

Sugi, Y. Tetrahedron Lett. 1978, 19, 1635. (b) Selke, R. React. Kinet.

Catal. Lett. 1979, 10, 135. (c) J ackson, R.; Thompson, D. J . J .

Organomet. Chem. 1978, 159, C29. (d) For the most recent contribu-

tions from the Selke group see: Berens, U.; Selke, R. Tetrahedron:

Asymmetry 1996, 7, 2055.

6014 J . Org. Chem., Vol. 62, No. 17, 1997

RajanBabu et al.

Sch em e 1. Syn th esis of “D-Glu co”

2,3-Dia r ylp h osp h in ite Rh od iu m Com p lexes

F igu r e 2. Pseudoenantiomeric relation between 2,3- and 3,4-

disubstituted glucopyranosides.

enation reaction.³⁰In this paper we provide the details

of this study.³¹

One of the perceived limitations of using natural

products as a source of chiral ligands is that often only

one of the enantiomeric series (in the case of carbohy-

drates and amino acids ^D- and L-series, respectively) is

readily available, and thus synthesis of only one enan-

tiomer of the product is economically viable. For ex-

ample, while ^D-glucose is available for less than $1.00/

Sch em e 2. P r ep a r a tion of

Dia r ylch lor op h osp h in es

kg, the corresponding ^L-sugar costs about $1.00/g.

A

solution to this problem can be found if one examines

the stereochemical relationships between the C-2, C-3,

and C-4 carbon atoms of the glucopyranoside ring system

in a pairwise fashion. The C-2:C-3-bis-phosphinite is

pseudoenantiomeric with respect to the C-3:C-4 phos-

phinite (Figure 2). Admittedly, there is no perfect

matching of all the groups in the periphery of the sugar

ring. Now if one makes the basic assumptions that the

enantioselectivity of the reaction depends only on the local

chirality, i. e., the chirality of the two vicinal carbons to

which the chelating phosphorus atoms are attached, and

that there is no large scale disruption of the equatorial

arrangement of the rest of the groups on the sugar

backbone, it should be possible to make both enantiomers

of a product using the ^D-sugar backbone. These assump-

tions have been borne out for the first time in the context

of asymmetric hydrogenation, making it possible to

develop a series of readily available ligands for the

synthesis of R- and S-amino acids by choosing the

appropriate ring carbons for the attachment of the

diarylphosphino groups to the ^D-sugar backbone. In the

hydrogenation of R-(N-acylamino)acrylic acid derivatives,

the cationic Rh-complexes of 2,3- and 3,4-bis(diarylphos-

phinite)s derived from sugars with the ^D-gluco- configu-

ration give S- and R-amino acid derivatives, respectively.

In nearly all cases, the electronic amplification of enan-

tioselectivity is key to obtaining practical levels of asym-

metric induction.

process for the manufacture of ^L-DOPA has been

developed.^35bDuring the course of our investigations two

limitations of the use of simple diphenylphosphinites

became apparent: (1) for a number of heteroaromatic and

aromatic amino acids with substituents on the ring,

diphenylphosphinites gave relatively low enantioselec-

tivities; (2) we (vide infra) and others³⁶have observed

low enantioselectivities in the reduction of dehydroamino

acids with protecting groups other than acetyl on the

nitrogen. The latter is an important limitation since it

has been found that removal of the acyl protecting group

may lead to low yields and occasionally significant

racemization at the R-carbon.³⁷It is desirable to have a

catalyst that gives high enantioselectivity for amino acids

carrying easily removable protecting groups such as

N-Cbz or N-Boc.³⁸

Among the chelating phosphines, arguably, the sugar

phosphinite ligand system appeared to be the most easily

accessible, and overcoming the above-mentioned limita-

tions using these ligands would be of practical signifi-

cance. Toward this goal, we decided to examine the role

of ligand electronic effects on the course of this reduction.

Accordingly, we prepared a variety of phenyl 2,3-bis-O-

(diarylphosphino)-â-^D-glucopyranosides with electroni-

cally different aryl groups by slight modifications of

procedures described in the literature (Scheme 1).^12,35a

The requisite chlorodiarylphosphines were prepared from

dichloro(diethylamino)phosphine³⁹and a Grignard re-

agent followed by treatment of the resulting diarylami-

nophosphine with anhydrous HCl (Scheme 2).⁴⁰The

chlorodiarylphosphines can also be prepared by the

Resu lts a n d Discu ssion

Su ga r P h osp h in ites a s Liga n d s in Rh -Ca ta lyzed

Asym m etr ic Hyd r ogen a tion . Phosphinites have been

used as ligands since 1975,³²and almost invariably their

use has been limited to Rh-catalyzed hydrogenations.^24,33,34

Based on the work of the Selke group,^24,35aa commercial

(33) For applications of phosphinites as ligands for metals other than

Rh, see: Pd: J ackson, W. R.; Lovel, C. G. Aust. J . Chem. 1982, 35,

2069. Trost, B. M.; Murphy, D. J . Organometallics 1985, 4, 1143. Ayers,

T. A.; RajanBabu, T. V.; Casalnuovo, A. L. Proceeding of Organome-

tallic Chemistry directed towards Organic Synthesis, IUPAC, 1995,

Santa Barbara, p 89. See also references 2d, 15, and 34.

(34) Nomura, N.; Mermet-Bouvier, Y. C.; RajanBabu, T. V. Synlett

1996, 745.

(35) (a) Selke, R.; Pracejus, H. J . Mol. Catal. 1986, 37, 213. (b) Vocke,

W.; Ha¨nel, R.; Flo¨ther, F.-U. Chem. Tech. (Leipzig) 1987, 39, 123.

(36) (a) Kreuzfeld, H.-J .; Do¨bler, C.; Krause, H. W.; Facklam, C.

Tetrahedron: Asymmetry 1993, 4, 2047. See also: (b) Do¨bler, C.;

Kreuzfeld, H.-J .; Krause, H. W.; Michalik, M. Tetrahedron: Asymmetry

1993, 4, 1833. (c) Ojima, I.; Yoda, N.; Yatabe, M.; Tanaka, T.; Kogure,

T. Tetrahedron 1984, 40, 1255. (d) Achiwa, K. Chem. Lett. 1977, 777.

(37) Taudien, S.; Schinkowski, K.; Krause, H. W. Tetrahedron:

Asymmetry 1993, 4, 73.

(30) Anecdotal evidence of electronic effects in Rh-catalyzed hydro-

genation dates back to some of the original studies done by Kagan.

See for example: (a) Dang, T. P.; Poulin, J . C.; Kagan, H. B. J .

Organomet. Chem. 1975, 91, 105. (b) Hengartner, U.; Valentine, D.,

J r.; J ohnson, K. K.; Larscheid, M. E.; Pigott, F.; Scheidl, F.; Scott, J .

W.; Sun, R. C.; Townsend, J . M.; Williams, T. H. J . Org. Chem. 1979,

44, 3741. (c) Yamagishi, T.; Yatagai, M.; Hatakeyama, H.; Hida, M.

Bull. Chem. Soc. J pn. 1984, 57, 1897. (d) Inoguchi, K.; Morimoto, T.;

Achiwa, K. J . Organomet. Chem. 1989, 370, C9. (e) Werz, U.; Brune,

H. A. J . Organomet. Chem. 1989, 365, 367. (f) Morimoto, T.; Chiba,

M.; Achiwa, K. Chem. Pharm. Bull. 1992, 40, 2894.

(31) A preliminary account of this work has been published; see

reference 16.

(32) Tanaka, M.; Ogata, I. J . Chem. Soc., Chem. Commun. 1975,

735.

(38) Use of DUPHOS ligands is another solution to this problem;

see reference 29.

(39) Perich, J . W.; J ohns, R. B. Synthesis 1988, 142.

(40) Unruh, J . D.; Christenson, J . R. J . Mol. Catal. 1982, 14, 19.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6015

reaction of 2 equiv the corresponding arylmagnesium

bromide with tri-n-butyl phosphite followed by treatment

of the resulting phosphinic acid with a halogenating

agent such as PCl₃.¹²Treatment of various diols with

the chlorodiarylphosphines gave high yields of the phos-

phinites. The corresponding cationic Rh-complexes were

prepared by treating the ligand with Rh⁺(L)₂X^-where

L is either 1,4-cyclooctadiene or norbornadiene and X is

SbF₆, OTf, or BF₄.⁴¹The structures of ligands used in

this study are shown in Figure 3.

1. (S)-Ar om a t ic a n d -H et er oa r om a t ic Am in o

Acid s. The best substrates for the sugar-derived phos-

phinites are the aromatic and heteroaromatic dehy-

droamino acids. The hydrogenation of various (Z)-N-acyl

dehydroamino acid substrates using cationic Rh-com-

plexes were carried out at 30-40 psi of hydrogen at room

temperature in a Fischer-Porter tube. The results of

hydrogenation using ligands 11a -h (Figure 3) are shown

in Table 1. Seven-membered Rh-chelates of electron-rich

vicinal phosphinites are very efficient catalysts,^35aand

typically, 1 mmol of a dehydroamino acid or the corre-

sponding methyl ester dissolved in 4 mL of THF is

quantitatively hydrogenated in less than 15 min using

0.004 mol of Rh⁺(11b)(NBD) SbF₆

. In general, the

-

scouting experiments were run for 2-3 h using 0.05-

0.1 mol% of the catalyst. The ee’s were determined by

capillary gas chromatographic analysis of the crude

product using Chirasil-^L-Val column under conditions

where a base-line separation of the enantiomers was

observed for the N-acetyl methyl esters of all the amino

acids reported here. In reactions where the acid was used

as the substrate, the methyl esters were prepared by

reaction of the hydrogenation product with diazomethane.

In several instances the enantiomeric purity were cor-

roborated with HPLC analysis using Chiralcel OJ or OB

column and optical rotation. The gas chromatographic

data is reproducible within (0.2%, and the HPLC data

within (0.5%. Several typical chromatograms are in-

cluded in the Supporting Information.

The most conspicuous result in the table is the re-

markable electronic effect of the aryl substituents of the

chelating phosphorus atom. In the case of methyl (Z)-

N-acetylcinnamate, the enantioselectivity varies from 2%

ee for the electron-deficient 3,5-difluorophenyl derivative

to 99% ee for the electron-rich 3,5-bis(trimethylsilyl)phen-

yl derivative⁴²(Table 1, entry 1). The same general trend

is seen for other substrates. Under otherwise identical

conditions, the acid substrates gave slightly better se-

lectivity as compared to the esters. Significant differ-

ences in ee’s between the unsubstituted (11e) and 3,5-

dimethylphenyl (11b )-substituted phosphinites were

noticed, and the difference is highly substrate dependent.

By some estimates the trimethylsilyl group is 1.25 times

as electron-donating as a methyl substituent⁴³and as

expected the corresponding phosphinites (11a ) gave

excellent ee’s. However, in practical terms, 3,5-dimethyl

derivatives are preferred because of the ease of prepara-

tion. Reduction of the Cbz-protected dehydroamino acid

ester (Scheme 3) in entry 6 (Table 1) is particularly

noteworthy since this substituent has been shown to give

low selectivities under other catalytic hydrogenation

conditions.³⁶Thus use of the diphenylphosphinite 11e

F igu r e 3. Carbohydrate-derived ligands for amino acid

synthesis.

gave only 62.0% ee where as the bis(dimethylphenyl)

derivative 11b gave 96% ee. The Cbz amino acid ester

did not give satisfactory separation on any of the chiral

columns, so the analysis of the product was conducted

on the N-acetyl derivative 15, which was prepared by

hydrogenolysis of the Cbz group in acetic anhydride

(Scheme 3). Alternatively, the ester can be reduced to a

primary alcohol 16 and the N-Cbz-protected alcohol(s)

can be analyzed by HPLC on Chiralcel OJ column.

Heterocyclic amino acids such as 2- and 3-thienylalanine

are also formed with excellent selectivity (entries 12 and

13) using the ligands 11a and 11b. The counterion

appears to have a small, yet discernible effect on the

enantioselectivity. For example, in THF, SbF₆^-generally

-

gave a slightly better selectivity than BF₄(Table 1,

entries 1, 2, 5, 7, and Table 2, entries 1, 3, 5, 11). By

simple recrystallization of the crude hydrogenation prod-

uct (entries 3, 7, and 12), optically pure derivatives of

phenylalanine (99.7% ee), 3,5-bis(trifluoromethyl)phenyl-

alanine (99.3% ee), and 3-thienylalanine (99.5% ee) were

prepared in excellent yields. Optical antipodes of two of

the corresponding amino acids are shown in Table 2,

entries 4 and 13. It should be noted that, in general, none

of the experimental protocols have been optimized.

2. Effect of th e Su ga r Ba ck bon e on th e Sen se of

Asym m etr ic In d u ction : (R)-Am in o Acid s. The high

selectivity and the ease of ligand synthesis notwithstand-

ing, a serious limitation of this approach for the synthesis

of (R)-amino acids is the scarcity of ^L-glucose. The

highest selectivity recorded for the preparation of an (R)-

amino acid using a carbohydrate-derived ligand prior to

this work is only 63%.^44,45A solution to this problem

might be the use of sugars such as arabinose which are

available in both enantiomeric forms. However, the

realization that in the context of the â-^D-glucopyranoside

(41) Schrock, R. R.; Osborn, J . A. J . Am. Chem. Soc. 1971, 93, 2397.

(42) For the preparation of bis[3,5-bis(trimethylsilyl)phenyl]dieth-

ylaminophosphine, see: Trost, B. M.; Murphy, D. J . Organometallics

1985, 4, 1143.

(43) Gassman, P. G.; Deck, P. A.; Winter, C. H.; Dobbs, D. A.; Cao,

D. H. Organometallics 1992, 11, 959.

(44) Habusˇ, I.; Raza, Z.; Sˇunjic´, V. J . Mol. Catal. 1987, 42, 173.

6016 J . Org. Chem., Vol. 62, No. 17, 1997

Ta ble 1. En a n tioselectivity in Hyd r ogen a tion s Usin g Ca ta lysts [11a -h ]Rh ⁺(L) X^-

RajanBabu et al.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6017

Sch em e 3. Hyd r ogen a tion of N-Cbz

Deh yd r oa m in o Acid s

For comparison, the syntheses of several amino acids

using the 2,6-pivaloyl complex [21b]Rh⁺[COD] BF₄^-were

carried out, and the results are shown in Table 2, entries

1, 3, 5 and 11 (the numbers in parentheses). In THF

solution, %ee’s of 92.4, 92.0, 93.1, and 93.0 were obtained

for these substrates. This should be compared to %ee’s

of 93, 96.2, 95.9, and 96.0 for the corresponding 2,6-di-

O-benzoyl ligands. The importance of electronic effect

is illustrated by the complexes of [21e] and [21g] which

gave ee’s of 84% and 11%, respectively (entry 1, Table

2), for example, for methyl (Z)-N-acetylcinnamate.

scaffolding, the 2,3-phosphinite is psuedoenantiomeric

with the 3,4-phosphinite (Figure 2), prompted us to

investigate this substitution pattern for ligand synthesis.

We initially chose R-methyl ^D-glucopyranoside because

of the availability of the starting sugar and ease of

synthesis of the corresponding 2,6-di-O-protected deriva-

tives. A series of 3,4-O-bis(diarylphosphinite)s of these

derivatives were synthesized from the corresponding

diols, which in turn were prepared using chemistry

reported by Ogawa et al.⁴⁶Thus R-methyl D-glucopyra-

noside 17 upon treatment with bis(tri-n-butyltin) ether

in refluxing toluene followed by reaction with either

benzoyl chloride or pivaloyl chloride selectively yielded

the 2,6-di-O-protected glucosides 18 and 19, respectively

(Scheme 4). The dibenzoyl derivatives are highly crystal-

line and easily prepared in large quantities.⁴⁷Prepara-

tion of bis-phosphinites 20 and 21 and the corresponding

Rh-complexes were accomplished as described earlier.

The results of hydrogenation of a variety of substituted

dehydroamino acids using the Rh complexes of 20 and

21 are shown in Table 2.

A careful inspection of the data reveals several impor-

tant points. The electronic effect of the phosphinites are

just as pronounced in this ligand system as in the 2,3-

bis-O-(diarylphosphinite) system 11. As compared to 11,

where the 4,6-oxygens form a conformationally locked

acetal, the manifestation of the electronic effect appears

to be more pronounced and systematic across the various

aryl substituents in the more flexible 3,4-bis(diarylphos-

phinite) 20.⁶⁴Electron-deficient phosphinites carrying

F or CF₃substituents uniformly gave very low selectivi-

ties (Table 2, entries 1-5). As expected methyl 2,6-

benzoyl-3,4-bis-O-[bis(3,5-dimethylphenyl)phosphino]-R-

D-glucopyranoside (20b) gave the best results. As before,

we observed that when the reaction is done in THF

solvent, use of SbF₆as a counteranion appears to be

slightly beneficial, especially if the dehydroamino acid

rather than the methyl ester is the substrate. Highly

enriched products can be recrystallized to get optically

pure (R)-amino acid derivatives. 3,5-Bis(trifluorometh-

yl)phenylalanine (99.4% ee) and 3-thienylalanine (99.5%

ee) (entries 4 and 13, Table 2) are illustrative.

Excellent enantioselectivity for (R)-amino acid was also

observed with ligands derived from 2-amino-2-deoxy-^D-

glucose, a sugar abundantly available from chitin. The

preparation of methyl N-acetyl-2-amino-2-deoxybis-3,4-

O-(diarylphosphino)-R-^D-glucopyranoside (26) is shown in

Scheme 5. Table 3 lists the results of hydrogenation of

selected dehydroamino acids using the ligands 26. Note

that the enantioselectivity in these reactions using the

vicinal (3,5-dimethylphenyl)phosphinites are among the

highest reported for these substrates,²⁹and even an

aliphatic amino acid such as alanine (entry 6) is produced

in high ee (vide infra).

Dependence of the chirality of the product on the

chirality of the vicinal carbons carrying the phosphinite

groups was examined with four other sugar backbones

(28, 29, 30, and 31). Synthesis of these ligands are shown

in Schemes 6, 7, 8, and 9, respectively. The relative

stereochemistry of the two phosphorus-carrying carbon

atoms, coincidentally C-3 and C-4 in both the pyranose

and furanose cases, is the same as that in 20, and the

major enantiomer is, as expected, the (R)-amino acid. The

enantioselectivities observed for the prototypical sub-

strate, (Z)-methyl N-acetylcinnamate and a correspond-

ing 4-fluoro derivative, are shown in Table 4. Even a

3,4-bis-O-(diarylphosphino)fructofuranoside (31) pro-

duces the (R)-amino acid. A similar switching of enan-

tioselectivity between 2,3-bis-O-gluco- and 3,4-bis-O-

fructo-diarylphosphinites has been observed before in the

asymmetric hydrocyanation of vinyl arenes.¹⁹The effect

of an axial substituent is examined with the mannose-

derived ligand 30. Note that enantioselectivity is con-

siderably lower than that observed when the correspond-

ing C₂-equatorial substituted ligands (e.g., 20 or 21) are

used. Removal of the axial methoxy group from 28

produces a ligand (29) which gives reasonably good

enantioselectivity. Comparing the selectivities observed

with ligands 28 and 29, it is conceivable that the axial

methoxy group in 28 is partly responsible for the lower

ee’s seen with this ligand. Thus, a pyranose system with

all equatorial substituents appears to be necessary for

high enantioselectivity in hydrogenation reactions.⁴⁸

3. Lim it a t ion s of t h e Ca r b oh yd r a t e-Der ived

Liga n d s: Alip h a tic Am in o Acid s. The sugar phos-

phinite ligand system thus appears to be among the most

practical ligands for the synthesis of (S)- and (R)-aromatic

and -heteroaromatic alanine derivatives. One of the

serious limitations of this ligand system that needs

further attention is in its utility for the synthesis of

aliphatic amino acids. The results with prototypical

ligands described in the previous sections are shown in

Table 5. The only amino acid that is amenable to the

currently available catalysts is alanine, the (S)-isomer

of which is produced in 96.9% ee with the Rh-complex of

The ligands derived from the 2,6-O-benzoyl pyranoside

20 are better than those from the pivaloyl derivative 21.

(45) After our initial communication¹⁶appeared, R. Selke informed

us of the use of conformationally restricted phosphinites i and ii for

the syntheis of D amino acids (presented at the International Confer-

ence on Circular Dichroism, Bochum, Germany, 1991, Abstracts p 305).

We thank Dr. Selke for this private communication. See also reference

24d.

(46) Ogawa, T.; Matsui, M. Tetrahedron 1981, 37, 2363.

(47) It has since been found that for large scale synthesis of 18 low

temperature benzoylation of R-methyl D-glucopyranose using 3-picoline

as a base can be employed, thereby avoiding the use of the toxic tin

derivatives (unpublished results with R. Shapiro).

(48) Selke, R.; Schwarze, M.; Baudisch, H.; Grassert, I.; Michalik,

M.; Oehme, G.; Stoll, N.; Costisella, B. J . Mol. Catal. 1993, 84, 223.

For related observations see: Selke, R. J . Prakt. Chem. 1987, 329, 717.

6018 J . Org. Chem., Vol. 62, No. 17, 1997

Ta ble 2. En a n tioselectivity of Hyd r ogen a tion s Usin g Ca ta lysts [20/21]Rh ⁺(L) X^-

RajanBabu et al.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6019

Sch em e 4. Syn th esis of “D-Glu co”

3,4-Dia r ylp h osp h in ites

Sch em e 7. Syn th esis of 3,4-Dia r ylp h osp h in ites

fr om 1,5-An h yd r o-2-d eoxyglu citol

Sch em e 8. Syn th esis of “D-Ma n n o”

3,4-Dia r ylp h osp h in ites

Sch em e 5. Syn th esis of Meth yl 2-Aceta m id o-2-

d eoxy-3,4-bis0O-(d ia r ylp h osp h in o)-6-O-(t-bu tyl-

d im eth ylsilyl)-r-D-glu cop yr a n osid e

Sch em e 9. Syn th esis of “D-F r u cto”

3,4-Dia r ylp h osp h in ites

for the (Z)-dehydrovaline whereas the corresponding (E)-

isomer gave only 73.3% ee. A mixture of (Z) and (E) (ratio

90/10) dehydro-precursors gave 86.5% ee. Likewise, â,â-

disubstituted R-acetamidoacrylates^49,50and a homophe-

nylalanine precursor gave unacceptably low selectivities

with both the 2,3- and the 3,4-bis(diarylphosphinite)s

(Table 5, under 11b and 20b, entries 4 and 5).

Ta ble 3. Hyd r ogen a tion s Usin g Ca ta lysts

[26]Rh ⁺(COD) X^-

4. Som e P r a ctica l Con sid er a tion s in Usin g th e

Su ga r -Der ived P h osp h in ites a s Liga n d s for Am in o

Acid Syn th esis. It is widely accepted that a seven-

membered chelate intermediate can undergo reorganiza-

tions within its coordination sphere faster than in a five-

membered one. For this reason catalytic processes that

involve seven-membered chelates are relatively faster.⁵¹

Such flexibility has to be balanced against the confor-

mational rigidity at critical stages in the catalytic cycle

if high enantioselectivity is desired. In the case of the

best sugar-derived vicinal diarylphosphinites both these

conditions are met, and the corresponding Rh chelates

are extremely active catalysts. Qualitative measure-

ments of the half-lives of the reactions done at room

temperature with 0.01 mmol of catalyst/mmol of sub-

strate in 5 mL of THF at 30-40 psi of hydrogen gives a

value of less than 10 min for the best catalyst 11b. For

a direct comparison, we chose one of the best five-

membered chelates viz., [Me-DUPHOS][COD]Rh⁺com-

plex,²⁹and the results are shown in Scheme 10. Using

THF as the solvent, at 0.001 equiv of the respective

precatalysts, the hydrogenation of (Z)-N-acetyl 3,5-bis-

(trifluoromethyl)phenylalanine methyl ester was carried

Sch em e 6. Syn th esis of Liga n d s fr om

2-Deoxyglu cop yr a n osid e

(49) For a solution to this problem see: Burk, M. J .; Gross, M. F.;

Harper, G. P.; Kalberg, C. S.; Lee, J . R.; Martinez, J . P. Pure Appl.

Chem. 1996, 68, 37.

(50) Sawamura, M.; Kuwano, R.; Ito, Y. J . Am. Chem. Soc. 1995,

117, 9602.

(51) Oliver, J . D.; Riley, D. P. Organometallics 1983, 2, 1032. For a

study of rates of hydrogenation using five-, six-, and seven-membered

rhodium chelates see also: Landis, C. R.; Halpern, J . J . Organomet.

Chem. 1983, 250, 485. Descotes, G.; Lafont, D.; Sinou, D.; Brown, J .

M.; Chaloner, P. A.; Parker, D. Nouv. J . Chim. 1981, 5, 167.

11b (Table 5, entry 1). The corresponding (R)-enanti-

omer is produced in a respectable 95.0% ee using the 3,4-

bis(phosphinite) derived from N-acetylglucosamine (27b).

It is interesting to note that (Z)-N-acetyl dehydroamino

acids are reduced with higher enantioselectivity as

compared to the corresponding (E)-isomers (entry 2).

Thus using Rh⁺(11b)(COD) SbF₆^-, 92.0% ee was observed

6020 J . Org. Chem., Vol. 62, No. 17, 1997

RajanBabu et al.

Ta ble 4. Dep en d en ce of Ch ir a lity of P h osp h in ooxy Ca r bon s a n d th e Su ga r Ba ck bon e in Rh -Ca ta lyzed Hyd r ogen a tion s

(a ll R-a m in o a cid s)

Ta ble 5. Syn th esis of Alip h a tic Am in o Acid s

Sch em e 10. Rela tive Ra tes of Hyd r ogen a tion of a

Deh yd r oa m in o Acid Ester Usin g [11b]Rh ⁺(NBD)

of preparation of the sugar-derived ligands. In addition,

as mentioned earlier, a single recrystallization usually

gave nearly optically pure (∼99.5% ee) product in cases

where we attempted such purifications. There is no

reason to doubt that this is the case with other solid

amino acids. For aliphatic amino acids the sugar-derived

ligands are still not competitive with the DUPHOS series,

except for alanine. Another important advantage of the

DUPHOS ligands is that they are able to reduce both

(Z)- and (E)-isomers of dehydroamino acids with equal

facility. Therefore, a mixture of the stereoisomers of

these substrates can be used for the production of

aliphatic amino acids. For the sugar-derived phosphin-

ites further optimizations are needed before this can be

accomplished. Such studies are in progress.

5. Or igin of En a n tioselectivity. Asymmetric hy-

drogenation of R-acetamidoacrylic acid derivatives is

arguably the most extensively studied catalytic asym-

metric reaction. Details of the kinetic and thermody-

namic aspects of the key steps under a variety of reaction

conditions are known^9,28and a number of intermediates

have been isolated.^27,52-55Several models have been

advanced to explain the sense, if not the level, of

-

SbF ₆a n d [(R,R)-Me-DUP HOS]Rh ⁺(COD) TfO^-

in THF

out for 15 min, and the reaction was terminated by

releasing the hydrogen and refilling the reaction vessel

with nitrogen. Conversion and ee’s were measured by

gas chromatography. While the phosphinite-catalyzed

reaction proceeded to completion giving 96.1% ee, the Me-

DUPHOS-mediated reaction was only 10.7% complete

(96.2% ee). However, in making this comparison, it

should be noted that in the original studies with the

DUPHOS catalysts, the reactions were carried out in

alcoholic solvents even though it has been reported that

there is little difference between various solvents.²⁹The

enantioselectivity for DUPHOS ligands are in general

1-3% higher for aromatic amino acids. This marginal

difference in selectivity is more than offset by the ease

(52) McCulloch, B.; Halpern, J .; Thompson, M. R.; Landis, C. R.

Organometallics 1990, 9, 1392.

(53) Brown, J . M.; Chaloner, P. A. J . Chem. Soc., Chem. Commun.

1978, 321.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6021

Sch em e 11. Con for m a tion of Seven -Mem ber ed

Ca tion ic Rh Ch ela tes a n d th e Sen se of Asym m etr ic

In d u ction in Hyd r ogen a tion

asymmetric induction.⁵⁶In the case of five- and seven-

membered chelates, there exists a good empirical cor-

relation between the λ- and δ-conformations (represented

in their limiting forms by 32 and 33 in Scheme 11) of

the starting Rh-catalyst. These conformations present

an asymmetric, alternating axial-equatorial array of

P-aryl groups around the metal to an approaching

substrate. Effect of such asymmetry can be seen even

on a coordinated cyclooctadiene,^51,57,58and the clockwise

or anticlockwise skewing of this ligand with respect to

the P-Rh-P plane can be related to the sense of

asymmetric induction. The same forces responsible for

the skewing of the diene are thought to be responsible

for the enantioselective recognition and the subsequent

selective hydrogenation of prochiral olefins.⁵⁹From a

large body of work it is well-known that the seven-

membered chelates (with a P-Rh-P bite angle of 96° vs

82° for five-membered chelates) are more flexible, and a

number of energetically similar conformations of the

chelate are available in this coordination geometry. This

makes modeling of intermediates difficult. Nonetheless

seven-membered chelates, by virtue of their flexibility

and the attendant ability to undergo facile reorganization

during the catalytic cycle, generally promote faster

reactions.⁵¹Occasional dramatic reversal of the sense

of asymmetric induction^60-62can also be attributed to

such flexibility in the seven-membered chelates. Yet

F igu r e 4. Chem3D version of the crystal structure of cationic

34. Hydrogens and the SbF₆counterion have been omitted

for clarity.

-

another consequence of such flexibility is a better mani-

festation of the electronic effects (see Tables 1 and 2).

Preliminary observations seem to suggest that a confor-

mationally labile system is capable of imparting better

substituent electronic effects.^63,64

After repeated attempts, we have been able to grow

single crystals of the complex 34 (Figures 4 and 5) derived

from (1S,2S)-1,2-bis[(diphenylphosphino)oxy]cyclohexane,

stereochemically related to the 2,3-bis-O-(diarylphos-

phino)-^D-glucopyranoside.⁶⁵The diphosphinites gave

high ee’s in the hydrogenation of acetamidocinnamic acid

derivatives (74.2 and 82.0, respectively for the diphen-

ylphosphinite and the bis[bis(3,5-dimethylphenyl)phos-

phinite].⁶⁶Two perspectives of a Chem3D rendition of

the crystal structure of the cationic Rh species of 34 are

shown in Figure 4.⁶⁶A schematic representation based

on the X-ray crystal structure is shown in Figure 5. This

corresponds approximately to a λ-twist-chair conforma-

tion, and as predicted by the correlations, (S)-amino acid

is the major product of hydrogenation with this catalyst.

Note that the COD ligand is skewed in a clockwise

direction as has been seen for other cationic Rh complexes

that give the (S)-enantiomer.^51,57,58Selected interatomic

distances and intramolecular angles are shown in Table

(54) Brown, J . M.; Chaloner, P. A. J . Chem. Soc., Chem. Commun.

1980, 344.

(55) Bircher, H.; Bender, B. R.; von Philipsborn, W. Magn. Res.

Chem. 1993, 31, 293.

(56) (a) Kagan, H. B. Asymmetric synthesis using organometallic

catalysts. In Comprehensive Organomeallic Chemistry; Wilkinson, G.,

Stone, F. G. A., Abel, E. W., Ed.; Pergamon Press: Oxford, 1982; Vol.

8, p 463. (b) Reference 25. (c) Reference 51. (d) Bogdan, P. L.; Irwin, J .

J .; Bosnich, B. Organometallics 1989, 8, 1450. (e) Brown, J . M.; Evans,

P. L. Tetrahedron 1988, 44, 4905. (f) Pavlov, V. A.; Klabunovskii, E.

I.; Struchkov, Y. T.; Voloboev, A. A.; Yanovsky, A. I. J . Mol. Catal.

1988, 44, 217. (g) Seebach, D.; Plattner, D. A.; Beck, A. K.; Wang, Y.

M.; Hunziker, D.; Petter, W. Helv. Chim. Acta 1992, 75, 2171. (h)

Sakuraba, S.; Morimoto, T.; Achiwa, K. Tetrahedron: Asymmetry 1991,

2, 597. (i) Michalik, M.; Freier, T.; Schwarze, M.; Selke, R. Magn. Reson.

Chem. 1995, 33, 835. For a highly pertinent, critical evaluation of these

various models, see reference 10.

(57) Kyba, E. P.; Davis, R. E.; J uri, P. N.; Shirley, K. R. Inorg. Chem.

1981, 20, 3616.

(58) Armstrong, S. K.; Brown, J . M.; Burk, M. J . Tetrahedron Lett.

1993, 34, 879.

(59) However, caution should be excercised in applying these

correlations because several cationic Rh complexes exist as different

polymorphs which can have widely different conformations. See

reference 51. We have obtained crystal structures of (1S,2S)-bis[(di-

phenylphosphino)oxy]cyclohexane-Rh⁺(L) SbF6^-with L as COD and

NBD. While the former has a well-defined λ-twist-chair conformation

(34), in the crystals of the latter, two kinds of conformations can be

identified, a λ-twist-chair and a λ-twist-boat (unpublished results with

T. A. Ayers and J . Calabrese). Both these conformations exhibit a

clockwise skewed rotation of the diene ligands as expected of the

catalysts that give the (S)-enantiomers.

(63) To´th, I.; Hanson, B. E. Organometallics 1993, 12, 1506.

(64) We have seen similar behavior in catalytic enantioselective

alkylation of π-allylpalladium compounds by stabilized carbanions.

Rigid backbones even with seven-membered chelates tend to diminish

the electronic effects. Conformationally flexible ligands show more

pronounced and generally predictable trends in electronic effects. At

present, more examples are needed before this hypothesis can be fully

verified (see reference 34).

(65) For the structure of

a related complex see: Kempe, R.;

(60) To´th, I.; Hanson, B. E.; Davis, M. E. Tetrahedron: Asymmetry

1990, 1, 913.

(61) Selke, R. J . Prakt. Chem. 1987, 329, 717.

Schwarze, M.; Selke, R. Zeit. Krystall. 1995, 210, 555.

(66) A more detailed NMR and modeling study of this simplified

system is forthcoming (unpublished results with C. Hadad, B. Radetich,

and K. You).

(62) Brown, J . M.; Murrer, B. A. Tetrahedron Lett. 1980, 21, 581.

6022 J . Org. Chem., Vol. 62, No. 17, 1997

RajanBabu et al.

Philipsborn.⁷³After an extensive study of the ¹⁰³Rh NMR

chemical shifts of the major and minor diastereomers of

the dehydroamino acid-Rh(I) complexes, these workers

concluded that there is considerable difference in the Rh-

shielding parameters for the two initially formed com-

plexes. On further analysis, the data seem to suggest

that a remote substituent on the olefin can create

substantial and disproportionately different ¹⁰³Rh shield-

ing between the major and minor complexes. For ex-

ample, the differences (∆δ) between the ¹⁰³Rh chemical

shifts of the major and minor enamide complexes from

the 4-nitro, 4-H, and 4-OH acetamidocinnamic acid are

247, 108, and 47, respectively. Enantioselectivity of

related enol acetate reductions seemed to be affected by

remote substituent, electron-rich double bond getting

reduced with higher selectivity.²⁶Taken together these

data suggest that the electron density at rhodium may

be related to the enantioselectivity of the reaction,

presumably brought about by the differences in relative

rates of oxidative addition of hydrogen to the major and

minor Rh⁺[enamide] complexes. It is thus conceivable

that one could also affect the electronic character of Rh

by making changes in the ligand, rather than the olefin

as in the above-mentioned cases. Would this result in

enhanced reactivity in the minor isomer resulting in

higher ee’s? Our results seem to suggest that this might

be the case with the vicinal bis(phosphinite)s. Thus, we

propose that the P-aryl substituents on our ligand system

change the reactivity patterns of the major and minor

complexes, presumably by increasing the electron density

on the central Rh atom, thereby affecting the relative

rates of oxidative additions.⁷⁴Admittedly, at this point

it is not clear whether this difference is due to an

increased reactivity of the minor isomer or a decreased

reactivity of the major isomer when the ligand is changed

from a vicinal bis(diphenylphosphinite) to a more electron-

rich system. Enhancement of rate of oxidative addition

of H₂to d⁸metals by electron-rich ligands as well as

stereoelectronic effects on such additions^70,75have been

documented and therefore it is possible that the latter

(i.e., increased rate of addition to the minor complex) is

the source of higher enantioselectivity. Since the enan-

tioselectivity is also affected, albeit to a lesser extent, by

the ratio of major to minor diastereomeric acetamidocin-

namate-Rh complexes, yet another scenario is that this

equilibrium is affected in favor of the kinetically impor-

tant minor isomer.⁷⁴Hydrogenation studies under non-

equilibrium, non-Curtin-Hammett conditions (at higher

pressures and low temperatures, for example) should

shed some light on this possibility. Further experiments

including pressure and temperature dependence on

enantioselectivity and ¹⁰³Rh NMR studies of the inter-

mediate complexes to probe this possibility in the vicinal

phosphinite system are underway.

F igu r e 5. λ-Twist-chair conformation of cationic Rh complex

34.

Ta ble 6. Selected In ter a tom ic Dista n ces a n d An gles for

34 (See F igu r es 4 a n d 5)

distances, Å

angles (degrees)

75.2 (rear)

Rh-P(1)

Rh-P(2)

P(1)-P(2)

Rh(1)-C(1)

Rh(1)-C(2)

Rh(1)-C(5)

Rh(1)-C(6)

2.250

2.239

3.179

2.279

2.290

2.266

2.286

R

ν

46.6 (front)

89.9 (rear)

31.1 (front)

90.2°

R′

â′

P₁-Rh-P₂

6. The dihedral angles are depicted using 34b, adapted

from Seebach et al.⁶⁷The front/rear designations refer

to the positioning of the P-aryl groups with respect to a

plane perpendicular to the P₁-Rh-P₂plane which passes

through P₁and P₂.

The origin of the enhancement of enantioselectivity as

a function of the electronic nature of the P-aryl groups

remains speculative at this point. Assuming the Halpern

mechanism⁹holds good for the catalysis by the vicinal

bis(phosphinite) Rh complexes, under ambient conditions,

the enantioselectivity arises primarily as a consequence

of the increased rate of oxidative addition of H₂to the

minor diastereomer that is formed between the dehy-

droamino acid (or ester) and the cationic catalyst. This

oxidative addition of hydrogen to d⁸metal complex, which

is the rate-determining step in the hydrogenation, has

been extensively studied. A pronounced role of electronic

effects on the course of H₂addition to a number of Rh

and Ir complexes including Vaska’s complex (trans-

IrCl(CO)(PPh₃)₂) has been documented.^68-71The major

interactions between the d⁸metal and the hydrogen

molecules involve (i) σ donation from the hydrogen

2

bonding orbital into the empty p_zor a p_z-d_zhybrid

orbital, (ii) transfer of electron density from one of the

filled d_πorbitals (d_xzor d_yz) into the hydrogen antibonding

orbitals, and (iii) repulsive interaction between the filled

2

σ bonding orbital of hydrogen and the d_zorbital of the

metal. Of these, the last one is likely to be the major

contributor for the activation barrier toward addition of

H₂.⁷²Any electronic effect that optimizes these orbital

interactions in the minor complex vis-a´-vis the major one

should affect the enantioselectivity. Such an approach

to catalyst optimization was suggested by a timely

publication that appeared from the laboratory of von

Exp er im en ta l Section

Gen er a l Meth od s. Manipulations of air and moisture

sensitive materials were conducted in a nitrogen atmosphere

by using either Schlenk techniques or a Vacuum Atmospheres

(73) Bender, B. R.; Koller, M.; Nanz, D.; von Philipsborn, W. J . Am.

Chem. Soc. 1993, 115, 5889.

(67) See reference 56g.

(74) At this time we cannot rule out the unlikely possibility that a

remote substituent might be bringing about conformational changes

in key intermediates. This possibility has been suggested by two

experiments involving a reversal of sense of induction under modified

reaction conditions: (i) when the hydrogenation reaction was done

under widely diffent solvents (ref 61); (ii) when 4-(dimethylamino)phen-

yl substituent was used in the DIOP series (ref 60). For related

observations in Pd-catalyzed allylation of stabilized carbanions, see

reference 34.

(68) Kunin, A. J .; J ohnson, C. E.; Maguire, J . A.; J ones, W. D.;

Eisenberg, R. J . Am. Chem. Soc. 1987, 109, 2963.

(69) J ohnson, C. E.; Eisenberg, R. J . Am. Chem. Soc. 1985, 107,

3148.

(70) Burk, M. J .; McGrath, M. P.; Wheeler, R.; Crabtree, R. H. J .

Am. Chem. Soc. 1987, 110, 5034.

(71) Sargent, A. L.; Hall, M. B.; Guest, M. F. J . Am. Chem. Soc.

1992, 114, 517.

(72) Dedieu, A.; Strich, A. Inorg. Chem. 1979, 18, 2940.

(75) Crabtree, R. H.; Uriarte, R. J . Inorg. Chem. 1983, 22, 4152.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6023

dry box. Tetrahydrofuran (THF), ether, dimethoxyethane

(DME), hexane, and pentane were distilled from sodium

benzophenone ketyl under nitrogen. Benzene and toluene

were distilled from lithium aluminum hydride and stored over

activated 4 Å molecular sieves. Methylene chloride, dimeth-

ylformamide (DMF), and acetonitrile were distilled from

calcium hydride and stored over activated 3 Å molecular

sieves. Flash column chromatography⁷⁶was carried out on

230-400 (0.040-0.063 mm) silica (EM Reagents). NMR

spectra were recorded on C₆D₆or CDCl₃solutions using a

General Electric QM-300 spectrometer (¹H at 300 MHz, ³¹P

at 121.7 MHz). Gas chromatographic (GC) analysis was

performed with a Hewlett-Packard (HP) Model 5890 GC fitted

with a HP3396A integrator. Conversions were determined

using an HP cross-linked methyl silicone capillary column (30

m × 0.530 mm). Chiral gas chromatographic (GC) separations

were accomplished using Chirasil ^L-Val on WCOT fused silica

(25 m × 0.25 mm, 0.12 µm film thickness) capillary GC column

purchased from Chrompack (1130 Route 202 South, Raritan,

NJ 08869). HPLC separations were carried out on Chiralcel

OJ or Chiralcel OB HPLC column (25 cm × 4.6 mm) from

Daicel (Chiral Technologies, Inc., 730 Springdale Drive, Drawer

1, Exton, PA 19341). Phenyl R-^D-glucopyranoside, tri-acetyl-

D-glucal, methyl R-D-glucopyranoside, methyl â-D-glucopyra-

noside, methyl R-^D-mannopyranoside, and the aryl bromide

precursors for the diarylchlorophosphines were purchased from

Aldrich. 2-Deoxy-^D-glucose, D-fructose, and N-acetyl-D-glu-

cosamine were purchased from Pfanstiehl Laboratories (1219

Glenrock Avenue, Waukegan, IL 60085-0439). Acetamidoacry-

the mixture was added sieve-dried cyclohexane (300 mL), the

solution was filtered through dry Celite under nitrogen using

a pressure funnel, and the filtrate was collected in a 500 mL

flask.

A

³¹P NMR (benzene-d₆) of the product at this stage

showed only a single peak at δ 63.3 corresponding to the

(diethylamino)diarylphosphine.

A Claisen adaptor was attached to the flask, and dry HCl

gas (dried by bubbling through concd H₂SO₄) was passed

through the solution for 45 min at a copious rate. The reaction

was mildly exothermic. The mixture was degassed by bubbling

nitrogen, and the solid precipitate (diethylamine hydrochlo-

ride) was filtered off using a pressure funnel. The precipitate

was washed with 100 mL of cyclohexane. The filtrate was

concentrated and distilled on a kugelrohr oven. ¹H NMR

(C₆D₆) 1.85 (4 s, 12 H), 6.62 (s, 2 H), 7.25 (m, 4 H); ³¹P NMR

(C₆D₆) 84.0. Yield 83-88%.

For the synthesis of diphosphinites which can be recrystal-

lized (for example 20b), there is no need to distill the

chlorophosphine. A crude NMR (¹H NMR and ³¹P NMR)

showed that the crude chlorophosphine is at least 90% pure.

This may be used in subsequent steps provided ca. 10% excess

is used for phosphinylation.

Diethylaminobis[3,5-bis(trimethylsilyl)phenyl]phosphine was

prepared according to the procedure of Trost et al.,⁴²and it

was converted into the chlorophosphine by treatment with

anhydrous HCl. ¹H NMR (C₆D₆) 0.10 (s, 36 H), 7.81 (m, 2 H),

8.05 (dd, J ) 1.1, 8.0 Hz, 4 H); ³¹P NMR (C₆D₆) 84.4.

P r epa r a tion of Ch lor op h osp h in es via th e Cor r esp on d -

in g Dia r ylp h osp h in ic Acid . An alternate preparation of

diarylchlorophosphines and their spectral data have been

described elsewhere.¹²

late derivatives,⁷⁷(COD)₂Rh⁺SbF₆^{- 41}, and (COD)₂Rh⁺

- 41

BF₄

were synthesized according to literature procedures.

E xa m p le of Syn t h esis of a Ca t ion ic Bis-(COD)-R h

Com p lex.⁴¹In a dry box 2.00 g (4.06 mmol) of [(COD)RhCl]₂

was dissolved in 40 mL of CH₂Cl₂and 4 mL of 1,5-cycloocta-

diene. With the room lights turned off, 2.79 g (8.11 mmol) of

AgSbF₆was added in one lot, and the mixture was stirred for

10 min. The solution was filtered through Celite, and the

Celite pad was washed with CH₂Cl₂. The filtrate was concen-

trated to 10 mL, and 12 mL of ether was added to precipitate

(COD)₂Rh SbF₆(1.09 g, 48%). Anal. C 34.69; H 4.42; F 19.80;

calcd for RhC₁₆H₂₄SbF₆C 34.63; H 4.36; F 20.54. This

analytically pure complex was used for subsequent steps for

the formation of the Rh-complexes. The compounds (COD)₂Rh⁺

OTf^-and (COD)₂Rh⁺BF₄^-were prepared in a similar manner.

Exam ple of Syn th esis of Diar ylch lor oph osph in es: Bis-

(3,5-d im eth ylp h en yl)ch lor op h osp h in e. P r ep a r a tion of

3,5-Dim eth ylp h en ylm a gn esiu m Br om id e. A three-necked

flask fitted with a mechanical stirrer, thermocouple adapter,

and an addition funnel was charged with 4.6 g (189 mmol) of

Syn th esis of th e Diol P r ecu r sor s. P h en yl 4,6-O-ben -

zylid en e-â-^D-glu cop yr a n osid e (10). A dry 2 L round-

bottomed flask was charged with 50 g (195 mmol) of phenyl

â-^D-glucopyranoside in 1 L of distilled acetonitrile. To the

heterogeneous mixture was added 1.37 g of p-toluenesulfonic

acid followed by 35.1 mL (234 mmol) of dimethoxytoluene.

After 2 h, a homogeneous solution was formed. Neat triethyl-

amine (1.2 mL) was added to neutralize the acid, and the

solvent was removed on the rotary evaporator. Recrystalli-

zation from neat ethyl acetate gave 58.9 g (88%) of product:

mp 154-160 °C; lit.⁷⁸162-165 °C.

Meth yl 2,6-Di-O-ben zoyl-r-^D-glu cop yr a n osid e (18). A

solution of 2.2 g (14 mmol) of methyl R-^D-glucopyranoside and

12.5 g (21 mmol) of (Bu₃Sn)₂O in 100 mL of toluene was heated

to reflux with azeotropic removal of water. When the volume

reached 10 mL of toluene, an additional 100 mL of toluene

was added. Distillation was continued until the volume

reached ca. 50 mL. After cooling to room temperature, 4.2 g

(30 mmol) of PhC(O)Cl in 5 mL of toluene was added dropwise.

After 2 h, 10 mL of a saturated KF solution and 100 mL of

ether were added. The resulting white solids were removed

by filtration through Celite. The organic phase was dried

(MgSO₄) and concentrated. Recrystallization from ethyl acetate/

hexane gave 3.6 g (64%) of methyl 2,6-di-O-benzoyl-R-^D-

glucopyranoside.⁴⁶

magnesium turnings in 60 mL each of THF and ether.

A

crystal of iodine was added to facilitate the initiation of the

reaction. To this suspension was added 35.44 g (180 mmol)

of bromo-3,5-dimethylbenzene in 35 mL of ether at such rate

that the temperature is kept around 22-28 °C. The reaction

was stirred overnight, and the contents were transferred into

dry addition funnel via a cannula. Additional 60 mL of THF

was used to facilitate quantitative transfer of the Grignard

reagent. Based on the unreacted Mg, the yield of the Grignard

reagent was estimated as 99.4%.

Mod ified P r oced u r e for th e P r ep a r a tion of Meth yl 2,6-

Di-O-ben zoyl-r-^D-glu cop yr a n osid e (18). In a Dean-Stark

set-up, a suspension of 40 g (0.21 mol) of dry methyl R-^D-

glucopyranoside and 184.0 g (157 mL, 0.31 mol) of di-n-butyltin

ether in 1 L toluene was brought to reflux. About 800 mL of

toluene and water was azeotopically removed, and the mixture

was cooled to 0 °C. To the above solution was added 87 g (72

mL, 0.62 mol) of benzoyl chloride, and the solution was brought

to room temperature. The mixture was stirred overnight, and

to the white solid formed was added 40 mL of glacial acetic

acid followed by 100 mL of toluene. The low boiling compo-

nents were removed on a rotary evaporator at 60 °C over 2 h.

The solid residue was taken up in an extraction thimble and

continuously extracted with boiling hexane (1500 mL) for 3 h.

The remaining white solid (55.52 g, 67%) in the extraction

thimble was dried under vacuum overnight at 0.1 mm. TLC

and NMR showed the material to be essentially pure. Further

A 500 mL three-necked round-bottomed flask fitted with a

mechanical stirrer and thermocouple adapter was charged

39

with 12.6 mL (15.0 g, 86.4 mmol) of Et₂NPCl₂in 80 mL of

THF. The addition funnel containing the Grignard solution

from the previous step was attached to the third neck. The

reaction flask was cooled to 0-5 °C, and the Grignard reagent

was added over 1 h. The mixture was stirred overnight at

5-10 °C.

The mixture was transferred into a 1 L dry flask, and it

was further concentrated to ∼150 mL on a vacuum pump. To

(76) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923.

(77) (a) Schmidt, U.; Griesser, H.; Leitenberger, V.; Lieberknecht,

A.; Mangold, R.; Meyer, R.; Riedl, B. Synthesis 1992, 487. (b) Schmidt,

U.; Lieberknecht, A.; Wild, J . Synthesis, 1988, 159. (c) Berti, F.; Ebert,

C.; Gardossi, L. Tetrahedron Lett. 1992, 33, 8145. (d) Cativiela, C.;

Diaz de Villegas, M. D.; Mayoral, J . A.; Melendez, E. Synthesis 1983,

899. (e) Herbst, R. M.; Shemin, D. In Organic Synthesis; Blatt, A. H.,

Ed.; J ohn Wiley: New York, 1943; Coll. Vol. 2; p 1.

(78) McGloskey, C. M.; Coleman, G. H. J . Org. Chem. 1945, 10, 184.

6024 J . Org. Chem., Vol. 62, No. 17, 1997

RajanBabu et al.

purification was achieved by recrystallizing from ethyl acetate

and hexane or diisopropyl ether: mp 141-144 °C (lit.⁴⁶140-

142 °C).

etonitrile. This compound was treated with NaBH₄and HCl⁸⁰

to provide a mixture of products from which the methyl 2,6-

di-O-benzyl-R-^D-mannopyranoside was isolated by flash chro-

matography. The assignment of this isomer was confirmed

by ¹H decoupling experiments on the corresponding bis(3,4-

O-(diphenylphosphino) derivative 30e. ¹H NMR 7.42-7.24 (m,

aromatic), 4.81 (d, J ) 1, 1 H), 4.75-4.54 (m, 4 H), 3.78-3.71

(m, 6 H), 3.36 (s, 3) H, 2.83 (s, br, 1 H), 2.43 (s, br, 1 H).

1,5-An h yd r o-2-d eoxy-6-O-(t r ip h en ylm et h yl)-^D-glu ci-

tol (Sch em e 7). To 1 g of tri-O-acetyl-^D-glucal dissolved in

10 mL of ethyl acetate and 5 mL of ethanol was added 100

mg of 10% Pd/C, and the solution was hydrogenated in a

Fischer-Porter tube at 30 psi of hydrogen. After vigorous

stirring overnight, the catalyst was filtered off with the aid of

Celite. The Celite bed was washed with excess ethyl acetate.

The combined filtrate was concentrated, and the reduced

triacetate was collected by column chromatography on silica

gel using 50% ethyl acetate/hexanes as solvent. ¹H NMR

2.00-2.10 (3 s, 9 H), 3.40-3.60 (m, 2 H), 4.00-4.30 (m, 5 H),

4.95-5.03 (m, 2 H). The crude product was used for the

subsequent reaction.

Meth yl 2,6-Di-O-pivaloyl-r-^D-glu copyr an oside (19). Pre-

pared in a route similar 18, except for substitution of pivaloyl

chloride in the place of benzoyl chloride: mp 91-94 °C. Anal.

C 56.34; H 8.65. Calcd for C₁₇H₃₀O₈C 56.34; H 8.34.

Meth yl 2-Deoxy-6-O-(ter t-bu tyld im eth ylsilyl)-r-^D-glu -

cop yr a n osid e (Sch em e 6). To a solution of 100 mg (0.56

mmol) of methyl 2-deoxy-R-^D-glucopyranoside, 41 mg (0.60

mmol) of imidazole, and 5 mg of DMAP in 1 mL of DMF was

added 90 mg (0.60 mmol) of tert-butyldimethylchlorosilane in

1 mL of DMF. After 18 h, the mixture was concentrated.

Flash chromatography using 50% hexane/ethyl acetate as

eluant gave 98 mg (60%) of the title compound: mp 49-50

1

°C; H NMR 0.07 (2 s, 6 H), 0.88 (s, 9 H), 1.61 (m, 1 H), 2.05

(m, 1 H), 3.29 (s, 3 H), 3.35 (m, 1 H), 3.46-3.55 (m, 2 H), 3.85-

3.78 (m, 4 H), 4.73 (d, J ) 3, 1 H). Anal. C 53.58; H 9.34.

Calcd for C₁₃H₂₈O₅Si C 53.49; H 10.04.

Meth yl 2-Aceta m id o-2-d eoxy-6-O-(ter t-bu tyld im eth yl-

silyl)-â-^D-glu cop yr a n osid e (25, Sch em e 5). Prepared by

silylation of methyl 2-acetamido-2-deoxy-â-^D-glucopyranoside,

which in turn was prepared from 2-amino-2-deoxyglucose via

2-acetamido-2-deoxy-3,4,6-tri-O-acetyl-â-^D-glucopyranosylchlo-

ride. (a ) 2-Aceta m id o-2-d eoxy-3,4,6-tr i-O-a cetyl-â-^D-glu -

cop yr a n osyl Ch lor id e (23). In a 500 mL flask, 25 g of

N-acetylglucosamine and 75 mL of distilled acetyl chloride was

combined, and the mixture was stirred for 24 h under nitrogen.

Methylene chloride (150 mL) was added, and the organic layer

was washed with water (4 × 50 mL) and then saturated

sodium bicarbonate (2 × 100 mL). After the solution was

concentrated to 100 mL, 300 mL of ether was added, and the

solution was allowed to stand overnight. The fine crystals

formed (58% yield) were filtered off with the aid of a pressure

funnel with exclusion of moisture, and the crystals were dried

in a vacuum desiccator overnight at <0.01 mm vacuum. (b)

Met h yl 2-Acet a m id o-2-d eoxy-3,4,6-t r i-O-a cet yl-â-^D-glu -

cop yr a n osid e. In a dry flask 3.067 g (8.38 mmol) of the

chloride was dissolved in 4 mL of CH₂Cl₂. A separate flask

was charged with 6.46 g (8.38 mmol) of silver carbonate and

0.8 g of 4 Å molecular sieves in 30 mL of 2:1 CH₂Cl₂/MeOH.

The chloride solution was carefully added to the methanol

solution under nitrogen, and the mixture was stirred over-

night. The insoluble precipitates were removed by filtration

and further washed with CH₂Cl₂. The pure product (2.423 g,

78%) was collected by column chromatography on silica gel

using neat ethyl acetate as a solvent. (c) Meth yl 2-Aceta -

m id o-2-d eoxy-â-^D-glu cop yr a n osid e (24). To a solution of

1 g of the triacetate from the previous experiment in 20 mL of

anhydrous methanol was added 100 mg of Biorad AGMP

(OH^-) resin, and the mixture was stirred at room temperature

overnight. Since the reaction was incomplete, an additional

200 mg of resin was added, and the stirring was continued

until all the starting material disappeared as judged by TLC

on silica gel using 10% methanol in CH₂Cl₂employing 5%

H₂SO₄for spot visualization. The resin was filtered off in a

pressure funnel, and it was then washed with hot methanol.

The methanol solution was evaporated to get 619 mg (95%) of

an off white solid which was used for the subsequent silylation

reaction. (d ) Meth yl 2-Aceta m id o-2-d eoxy-6-O-ter t-bu -

tyld im eth ylsilyl-â-^D-glu cop yr a n osid e (25). The triol from

the previous step was silylated using tert-butyldimethylchlo-

rosilane in DMF in the presence of imidazole, and the product

was purified by column chromatography on silica gel using

To a solution of the triacetate (4.56 g) in 50 mL of distilled

methanol was added 1.50 g of Biorad AGMP(OH^-) resin was

added. The reaction was stirred at room temperature until

all of the starting material disappeared (TLC, 10% methanol

in CH₂Cl₂, 18 h). The resin was filtered off, and the methanolic

filtrate was concentrated on a rotary evaporator. The last

traces of the solvent was removed high vacuum at 0.01 mm to

get quantitative yield of the expected triol. This analytically

pure crude product was used for the subsequent tritylation.

¹H NMR (D₂O) 1.60 (dddd, J ) 12, 12, 12, 6, 1 H), 1.97 (dd, br,

J ) 12, 6, 1 H), 3.21 (d m, 2 H), 3.50 (dt, J ) 12, 2, 1 H),

3.60-3.70 (m, 2 H), 3.90 (dd, J ) 10, 2, 1 H), 3.95 (dd, J ) 12,

5, 1 H). Anal. C 48.26; H 8.36. Calcd for C₆H₁₂O₄C 48.64; H

8.16.

The triol (148 mg) was converted into the corresponding 6-O-

trityl derivative by treatment with 307 mg (1.1 equiv) of trityl

chloride and 6 mg of 4-(dimethylamino)pyridine in 3 mL of

pyridine. After 5 days, 30 mL each of of cold water and

methylene chloride were added, and the organic layer was

separated. The aqueous layer was further extracted with

methylene chloride (30 mL × 2). The combined organic layer

was washed with ice-cold 1 N HCl followed by saturated

NaHCO₃. It was further dried and concentrated. Chroma-

tography on silica yielded 103 mg (26%) of desired product.

¹H NMR 1.70 (dddd, J ) 12, 12, 12, 6, 1 H), 1.90 (dm, 12, 1

H), 2.50 (s, br, exch OH), 2.90 (s, br, OH, exch 1 H), 3.30-

3.50 (m, 5 H), 3.59-3.70 (m, 1 H), 3.90 (ddd, J ) 12, 6, 2, 1

H), 7.20-7.50 (aromatic).

Meth yl 1,6-Bis-O-(tr ip h en ylm eth yl)-r- a n d â-^D-fr u cto-

fu r a n osid es (Sch em e 9). These compounds were prepared

by tritylation of a mixture of methyl R- and â-fructofuranosides

with trityl chloride in pyridine at 50 °C and separating these

isomers by column chromatography on silica gel using 10-

30% ethyl acetate in hexane as eluant. The R-anomer eluted

first. The R-anomer: mp 92-93 °C; ¹H NMR 3.05 (s, 3 H),

3.07 (d, J ) 10 , 1 H), 3.24, 3.48 (ABX, J _AB) 10, 2 H), 3.34,

3.50 (AB, J _AB) 10 , 2 H), 3.57 (d, J ) 10 Hz, 1 H), 3.87 (d, br,

J ) 11, 1 H) 4.05 (m, 1 H), 4.19 (d, J ) 10, 1 H), 7.20-7.60

(m, aromatic). Anal. C 78.85; H 6.37. Calcd C 79.62; H 6.24.

The â-anomer: mp 195-199 °C; ¹H NMR 2.15 (d, J ) 4, 1 H,

exch with D₂O), 2.65 (d, br, J ) 8, exch with D₂O, 1 H), 3.05

(s, 3 H), 3.07-3.27 (m, 4 H), 3.97 (m, 1 H), 4.17-4.22 (m, 2

H), 7.20-7.55 (m, aromatic). Anal. C 78.85; H 6.36. Calcd C

79.62; H 6.24.

1

ethyl acetate/hexanes as solvent: mp 151-155 °C; H NMR

The stereochemistry of the two fructosides were established

by conversion to the known^81,82methyl 3,4-anhydro-derivatives

by Mitsunobu reaction and examination of the ¹³C NMR

spectrum. For the R-anomer, the methoxy signal consistently

has the lower chemical shift. In the case of the triphenyl-

methyl fructofuranosides, the R-methoxy carbon appears at δ

49.20 and the â-derivative at 52.30 (CDCl₃).

0.00 (2 s, 6 H), 0.80 (2 s, 9 H), 1.98 (s, br, 3H), 3.20-3.32 (m,

1 H), 3.32-3.50 (s, superimposed on m, 5 H), 3.59 (dd, J ) 12,

8, 1 H), 3.76, 3.84 (ABX, J _AB) 18, 2 H), 4.28 (d, J ) 8, 1 H),

6.42 (d br J ) 4, 3 H). Anal. C 51.23; H 9.20. Calcd for

C₁₅H₃₁NO₆Si C 51.55; H 8.94.

Meth yl 2,6-Di-O-ben zyl-r-^D-m an n opyr an oside (Sch em e

8). A 2:1 mixture of exo- and endo-isomers of bis[(2,3-O),(4,6-

O)]benzylidene-R-^D-mannopyranoside⁷⁹was prepared by the

reaction of methyl R-^D-mannopyranoside with 2.2 equiv of R,R-

dimethoxytoluene and catalytic p-toluenesulfonic acid in ac-

(80) Horne, D. A.; J ordan, A. Tetrahedron Lett. 1978, 19, 1357.

(81) Guthrie, R. D.; J enkins, I. D.; Yamasaki, R. J . Chem. Soc.,

Perkin Trans. 1 1981, 2328.

(82) RajanBabu, T. V.; Nugent, W. A. J . Am. Chem. Soc. 1994, 116,

986.

(79) Lipta´k, A.; J oda´l, I.; Na´na´si, P. Carbohydr. Res. 1975, 44, 1.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6025

A Typ ica l P r oced u r e for t h e Syn t h esis of a Dia r yl-

p h osp h in ite.¹²To a solution of 300 mg (0.75 mmol) of methyl

2,6-O-dibenzoyl-R-^D-glucopyranoside and 10 mg of DMAP in

5 mL of pyridine was added 430 mg (1.56 mmol) of bis(3,5-

dimethylphenyl)chlorophosphine in 5 mL of CH₂Cl₂. After 16

h, the mixture was concentrated to dryness. Benzene was

added to the residue, and this mixture was filtered and

concentrated. Recrystallization from hexane/benzene gave 600

mg (91%) of 20b: ¹H NMR 1.73 (s, 6 H), 1.88 (s, 6 H), 1.91 (s,

6 H), 2.02 (s, 6 H), 2.75 (s, 3 H), 3.90 (dd, J ) 4, 12, 1 H), 4.04

(dd, J ) 4, 10, 1 H), 4.42 (d, J ) 12, 1 H), 4.89 (m, 1 H), 5.00

(d, J ) 3, 1 H), 5.23 (m, 1 H), 5.51 (dd, J ) 4, 10, 1 H), 6.03 (s,

1 H), 6.32 (s, 1 H), 6.46 (s, 1 H), 6.63 (s, 1 H), 6.70-7.30 (m,

aromatic), 7.80 (m, aromatic), 8.13 (m, aromatic); ³¹P NMR

124.7, 118.8. Anal. C 72.30; H 6.82; P 6.58. Calcd for

C₅₃H₅₆O₈P₂: C 72.10; H 6.39; P 7.02.

[(11g)Rh (COD)]⁺BF ₄^-. (C₆D₆) ν_A) 122.9, ν_B) 126.8, J _PP

) 39 , J _RhP) 179. Minor component (30%) ν_A) 123.9, ν_B

139.9, J _PP) 48, J _RhP1) 234, J _RhP2) 240.

)

[(11g)Rh (COD)]⁺OTf^-. (C₆D₆) ν_A) 122.4, ν_B) 125.9, J _PP

) 44 , J _RhP) 180.

-

[(11h )Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): ν_A) 125.0, ν_B

) 117.3, J _PP) 36 Hz, J _RhP) 173.

Met h yl 4,6-O-Ben zylid en e-2,3-b is-O-[b is(3,5-d im et h -

ylph en yl)ph osph in o]-r-^D-glu copyr an oside. ¹H NMR 7.34-

6.42 (m, aromatic), 5.13 (s, 1 H), 5.03 (m, 1 H), 4.49-4.42 (m,

2 H), 4.02 (dd, J ) 5, 10, 1 H), 3.88 (m, 1 H), 3.56 (t, J ) 9, 1

H), 3.37 (t, J ) 10, 1 H), 2.75 (s, 3 H), 1.95 (s, 6 H), 1.94 (s, 6

H), 1.91 (s, 6 H), 1.90 (s, 6 H); ³¹P NMR 119.6, 118.9.

Liga n d s a n d Ca ta lysts fr om Meth yl 2,6-Di-O-ben zoyl-

r-^D-glu cop yr a n osid e (18).

P r ep a r a tion of Rh -Com p lexes: Gen er a l P r oced u r e. In

a dry box under nitrogen, a solution of 0.50 mmol of phos-

phinite in 5 mL of CH₂Cl₂was added to 0.49 mmol of

+

Rh(COD)₂X^-(X ) SbF₆, BF₄, OSO₂CF₃) in 5 mL of CH₂Cl₂

at room temperature. The mixture was stirred for 30 min to

3 h and the solvent was carefully removed under vacuum. A

fine powder of the Rh-complex may be obtained by redissolving

the complex in 8 mL of benzene and freeze-drying the sample

under high vacuum.

Ligan ds an d Catalysts fr om P h en yl 4,6-O-Ben zyliden e-

â-^D-glu cop yr a n osid e (10).

Met h yl 2,6-Di-O-b en zoyl-3,4-b is-O-[b is(3,5-d im et h yl-

p h en yl)p h osp h in o]-r-^D-glu cop yr a n osid e (20b) (for prepa-

ration see Typical Procedure for the Synthesis of a Phosphin-

ite).

-

[(20b)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

ν_A) 129.0, ν_B) 130.4, J _PP) 10, J _AX) J _BX() J _RhP) ) 175.

-

[(20b)Rh (COD)]⁺SbF₆

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

ν_A) 132.8, ν_B) 134.2, J _PP) 30, J _AX) J _BX() J _RhP) ) 151.

[(20b)Rh (COD)]⁺OTf^-. ³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

ν_A) 130.5, ν_B) 131.8, J _PP) 34, J _AX) J _BX() J _RhP) ) 174.

Meth yl 2,6-Di-O-ben zoyl-3,4-bis-O-[bis(4-m eth oxyph en -

yl)p h osp h in o]-r-^D-glu cop yr a n osid e (20c). ¹H NMR 3.02

(s, 3 H), 3.19 (s, 3 H), 3.32 (s, 3 H), 3.34 (s, 3 H), 3.41 (s, 3 H),

4.19 (m, 1 H), 4.29 (m, 1 H), 4.65 (dd, J ) 2, 12, 1 H), 4.93 (m,

1 H), 5.27 (d, J ) 4, 1 H), 5.45 (m, 1 H), 5.69 (dd, J ) 4, 10, 1

H), 6.46-8.40 (m, aromatic); ³¹P NMR 120.5 (d, 1, J _PP) 5),

117.8 (d, 1, J _PP) 5). Anal. C, 65.33; H, 5.58; P 6.99. Calcd

P h en yl 4,6-O-ben zylid en e-2,3-bis-O-(d ia r ylp h osp h in o)-

â-^D-glu cop yr a n osid es (11a -11h ). The synthesis of these

ligands and their spectral properties have been described

elsewhere.¹²

-

[(11a )Rh (COD)]⁺SbF ₆

.

³¹P NMR (C₆D₆) ABX () P₁P₂Rh),

ν_A) 129.9, ν_B) 131.3, J _PP) 24, J _RhP) 176.

for C₄₉H₄₈O₁₂P₂C 66.06; H 5.43; P 6.95.

-

[(11a )Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆) ABX () P₁P₂Rh),

-

[(20c)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

ν_A) 125.6, ν_B) 128.3, J _PP) 38, J _AX) J _BX() J _RhP) ) 175.

ν_A) 134.4, ν_B) 136.1, J _PP) 28, J _AX) J _BX() J _RhP) ) 181 Hz.

Meth yl 2,6-Di-O-ben zoyl-3,4-bis-O-(diph en ylph osph in o)-

r-^D-glu cop yr a n osid e (20e). ¹H NMR 2.78 (s, 3 H), 3.91 (dd,

J ) 4, 12, 1 H), 4.04 (dd, J ) 4, 10, 1 H), 4.39 (d, J ) 12, 1 H),

4.70 (m, 1 H), 5.08 (d, J ) 3, 1 H), 5.22 (m, 1 H), 5.40 (dd, J )

4, 12, 1 H), 6.49-7.50 (m, aromatic), 7.85 (m, 2 H, aromatic),

8.12 (m, 2 H, aromatic); ³¹P NMR 120.0 (d, J _PP) 4), 116.0 (d,

-

[(11b)Rh (COD)]⁺SbF ₆

.

³¹P NMR (CDCl₃): ABX ()

P₁P₂Rh), ν_A) 136.6, ν_B) 136.8, J _PP) 27, J _AX) J _BX() J _RhP

)

) 177; in C₆D₆ABX () P₁P₂Rh), ν_A) 134.0, ν_B) 136.0, J _PP

)

29, J _AX) J _BX() J _RhP) ) 175.

-

[(11b )R h (COD)]⁺BF ₄

.

³¹P NMR (CDCl₃): ABX ()

P₁P₂Rh), ν_A) 133.6. ν_B) 129.3, J _PP) 31, J _AX) J _BX() J _RhP

)

J _PP) 4).

) 177.

-

[(20e)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

-

[(11c)Rh (COD)]⁺SbF ₆

.

³¹P NMR (C₆D₆) ABX () P₁P₂Rh),

ν_A) 130.8, ν_B) 133.7, J _PP) 32, J _AX) J _BX() J _RhP) ) 176.

ν_A) 139.5, ν_B) 140.1, J _PP) 24 , J _AX) J _BX() J _RhP) ) 182.

-

[(20e)Rh (COD)]⁺SbF₆

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

[[(11c)Rh (COD)]⁺OTf ^-

.

³¹P NMR (C₆D₆) ABX () P₁P₂Rh),

ν_A) 136.3, ν_B) 133.1, J _PP) 28 Hz, J _AX) J _BX() J _RhP) ) 177

Hz.

ν_A) 136.8, ν_B) 138.5, J _PP) 28 , J _AX) J _BX() J _RhP) ) 181.

-

[(11d )Rh (COD)]⁺SbF ₆

.

³¹P NMR (CDCl₃): ABX ()

Meth yl 2,6-Di-O-ben zoyl-3,4-bis-O-[bis(3,5-diflu or oph en -

yl)p h osp h in o]-r-^D-glu cop yr a n osid e (20f). ¹H NMR 2.81

(s, 3 H), 3.86 (m, 2 H), 4.23 (m, 1 H), 4.46 (m, 1 H), 4.82 (m, 1

H), 4.99 (d, 1, 4 H), 5.24 (dd, J ) 4, 10, 1 H), 5.71 (m, 1 H),

6.10-6.34 (m, 3 H), 6.58-8.20 (m, aromatic); ³¹P NMR 113.3,

P₁P₂Rh), ν_A) 134.7, ν_B) 137.9, J _PP) 28, J _AX) J _BX()J _RhP) )

182.

-

[(11d )Rh (COD)]⁺BF ₄

. (CDCl₃) ABX () P₁P₂Rh), ν_A)

131.4, ν_B) 135.9, J _PP) 30, J _AX) J _BX() J _RhP) ) 180.

-

[(11e)Rh (COD)]⁺SbF ₆

.

³¹P NMR (CDCl₃): ABX ()

109.8.

P₁P₂Rh), ν_A) 137.5, ν_B) 138.6, J _AB) 27, J _AX) J _BX()J _RhP

)

-

[(20f)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

) 176.

ν_A) 126.7, ν_B) 127.6, J _PP) 39, J _AX) J _BX() J _RhP) ) 179.

Meth yl 2,6-Di-O-ben zoyl-3,4-bis-O-[bis[3,5-bis(tr iflu o-

r om eth yl)p h en yl]p h osp h in o]-r-^D-glu cop yr a n osid e (20g).

¹H NMR 2.93 (s, 3 H), 3.98 (dd, J ) 5 12, 1 H), 4.11 (dd, J )

5, 10, 1 H), 4.28 (d, J ) 12, 1 H), 4.52 (m, 1 H), 5.11 (d, J ) 4,

1 H), 5.19 (m, 1 H), 5.45 (dd, J ) 4, 10, 1 H), 6.89-8.22 (m,

aromatic), ³¹P NMR 113.0, 107.5. Anal. C 48.38; H, 2.79; P

-

[(11e)Rh (COD)]⁺BF ₄

. )

³¹P NMR: ABX () P₁P₂Rh), ν_A

134.6,ν_B) 137.5, J _PP) 28, J _AX) J _BX() J _RhP) ) 179.

-

[(11f)Rh (COD)]⁺SbF ₆

.

³¹P NMR (CDCl₃): multiplet

superimposed on an ABX 8-line pattern with further small

coupling presumably due to long range interaction with

fluorines. δ 126.5, 126.8, 128.0, 128.3, 129.2, 129.5, 130.8,

131.1 (insoluble in benzene).

4.65. Calcd for C₅₃H₃₂F₂₄O₈P₂C, 48.42; H, 2.45; P 4.71.

-

[(11f)Rh (COD)]⁺BF ₄

. )

³¹P NMR: ABX () P₁P₂Rh), ν_A

-

[(20g)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆/CD₂Cl₂): 127.6 (d,

125.4, ν_B) 120.3, J _AB) 38, J _AX) J _BX() J _RhP) ) 179.

J _RhP) 181).

-

[(11g)Rh (COD)]⁺SbF₆

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

Met h yl 2,6-Di-O-b en zoyl-3,4-b is-O-[b is[4-(t r iflu or o-

m eth yl)p h en yl]p h osp h in o]-r-^D-glu cop yr a n osid e (20h ).

¹H NMR (C₆D₆) 2.80 (s, 3 H), 3.85 (dd, J ) 13, 4, 1 H), 4.06

(ddm, J ) 8, 4, 1 H), 4.28 (dd, J ) 13, 2, 1 H), 4.60 (dt, J ) 12,

12 1 H), 5.00 (m, 1 H), 5.03 (d, J ) 4, 1 H), 5.28 (dd, 12, 4, 1

H), 6.70-7.60 (m, aromatic). Anal. Calcd for C₄₉H₃₆F₁₂O₈P₂

C 56.44; H 3.48; P 5.94. Found C 56.36; H 3.75; P 5.92.

ν_A) 126.8, ν_B) 130.5, J _PP) 36, J _AX) J _BX() J _RhP) ) 182. In

addition this complex shows another minor component (10-

15%) with similar 8-line ³¹P spectrum. Seven of these lines

appear at δ 124.2, 124.6, 126.5, 138.4, 138.8, 140.3, and 140.7.

The 8th line could not be located. It is possibly under the peak

at 126.1.

6026 J . Org. Chem., Vol. 62, No. 17, 1997

RajanBabu et al.

-

[(20h )Rh (COD)]⁺BF₄

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

1,5-An h yd o-2-d eoxy-3,4-b is-O-[b is(3,5-d im et h ylp h en -

yl)p h osp h in o]-6-O-(tr ip h en ylm eth yl)-^D-glu citol (29b). ¹H

NMR (C₆D₆) 1.80-2.04 (m), 2.82-2.96 (m, 1 H), 3.22-3.58 (m,

4 H), 4.30-4.50 (m, 2 H), 6.30-7.60 (m, aromatic); ³¹P NMR

110.89 (d, J _PP) 4), 115.88 (d, J _PP) 4).

ν_A) 125.2, ν_B) 127.4, J _PP) 37, J _AX) J _BX() J _RhP) ) 177.

Also seen is a minor (<5%) isomer with an 8-line pattern in

the ³¹P NMR spectrum between δ 130 and 138.

Liga n d s a n d Ca t a lyst s fr om Met h yl 2,6-Bis-O-(t r i-

m eth yla cetyl)-r-^D-glu cop yr a n osid e (19).

-

[(29b)Rh (COD)]⁺SbF ₆

.

³¹P NMR (CD₃COCD₃) 129.3

(dd), 132.9 (dd), J _RhP) 176, J _PP) 30.

Liga n d s (30) a n d Ca ta lysts fr om Meth yl 2,6-Di-O-

ben zyl-r-^D-m a n n op yr a n osid e.

Meth yl 2,6-Bis-O-(tr im eth yla cetyl)-3,4-bis-O-[bis(3,5-

d im eth ylp h en yl)p h osp h in o]-r-^D-glu cop yr a n osid e (21b).

¹H NMR 0.93 (s, 9 H), 1.14 (s, 9 H), 1.90 (s, 6 H), 1.93 (s, 6 H),

1.98 (s, 6 H), 1.99 (s, 6 H), 2.88 (s, 3 H), 3.72 (m, 1 H), 3.95 (m,

1 H), 4.12 (dm, 1, J ) 12, 1 H), 4.50 (m, 1 H), 4.89 (m, 1 H),

5.08 (m, 1 H), 5.30 (m, 1 H), 6.33 (s, 1 H), 6.47 (s, 1 H), 6.53 (s,

1 H), 6.64 (s, 1 H), 6.85-6.95 (m, 2 H), 7.18-7.35 (m, aromatic);

³¹P NMR 122.1, 117.9.

-

[(21b)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

Meth yl 2,6-Di-O-ben zyl-3,4-bis-O-[bis(3,5-dim eth ylph en -

yl)p h osp h in o]-r-^D-m a n n op yr a n osid e (30b). ¹H NMR 1.87

(s, 6), 1.93 (s, 6 H), 1.94 (s, 6 H), 1.96 (s, 6 H), 3.00 (s, 3 H),

3.43-3.57 (m, 2 H), 3.74 (m, 1 H), 4.00-4.23 (m, 5 H), 4.56

(m, 1 H), 5.02-5.20 (m, 2 H), 6.40-7.50 (m, aromatic); ³¹P

ν_A) 129.0, ν_B) 135.2, J _PP) 30 Hz, J _AX) J _BX() J _RhP) ) 176.

Meth yl 2,6-Bis-O-(tr im eth yla cetyl)-3,4-bis-O-(d ip h en -

ylp h osp h in o)-r-^D-glu cop yr a n osid e (21e). ¹H NMR 0.93 (s,

9 H), 1.14 (s, 9 H), 2.97 (s, 3 H), 3.75 (dd, J ) 5, 12, 1 H), 3.94

(ddd, J ) 2, 5, 10, 1 H ), 4.17 (dd, J ) 2, 12, 1 H), 4.44 (m, 1

H), 5.00 (d, J ) 3, 1H), 5.05 (m, 1 H), 5.25 (dd, J ) 4, 10, 1 H),

6.78-7.50 (m, aromatic); ³¹P NMR 118.0 (d, 1, J _PP) 5), 114.8

NMR 120.5, 115.7.

-

[(30b)Rh (COD)]⁺BF ₄

. )

³¹P NMR (C₆D₆) ν_A) 124.8, ν_B

133.9, J _PP) 30; J _RhP) 176.

(d, 1, J _PP) 5).

Meth yl 2,6-Di-O-ben zyl-3,4-bis-O-(diph en ylph osph in o)-

r-^D-m a n n op yr a n osid e (30e). ¹H NMR 3.13 (s, 3 H), 3.55

(m, 2 H), 4.02 (m, 1 H), 4.11 (m, 1 H), 4.22 (m, 4 H), 4.72 (d,

-

[(21e)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): ABX () P₁P₂Rh),

ν_A) 134.0, ν_B) 136.5, J _PP) 30, J _AX) J _BX() J _RhP) ) 178.

Liga n d s a n d Ca ta lysts fr om Meth yl 2-Aceta m id o-2-

d e oxy-6-O-(t er t -b u t yld im e t h ylsilyl)-â-^D-glu cop yr a n o-

sid e (25).

J

) 2, 1 H), 4.95 (m, 1 H), 5.09 (m, 1H), 6.80-7.78 (m,

aromatic); ³¹P NMR 117.3, 110.4.

-

[(30e)Rh (COD)]⁺BF ₄

. )

³¹P NMR (C₆D₆) ν_A) 129.2, ν_B

137.2, J _PP) 27, J _RhP) 177.

Liga n d s (31) a n d Ca ta lysts fr om Meth yl 5,6-O-Bis-

(tr ip h en ylm eth yl)-r-^D-fr u ctofu r a n osid e.

Met h yl 2-Acet a m id o-6-O-(ter t-b u t yld im et h ylsilyl)-2-

d eoxy-3,4-O-bis[bis-(3,5-d im eth ylp h en yl)p h osp h in o]-â-^D-

glu cop yr a n osid e (26b). This compound was purified by

column chromatography inside the dry box using 40% ether/

hexanes on silica gel. ¹H NMR 0.15 (s, 3 H); 0.16 (s, 3 H),

1.12 (s, 9 H), 2.10-2.35 (4 s, total 27 H), 3.49 (s, 3 H), 3.80 (m,

4 H), 4.42 (m, 1 H), 4.73 (m, 1 H), 5.13 (d, J ) 8, 1 H), 5.26 (m,

1 H), 6.59-6.86 (m, aromatic), 7.22-7.55 (m, aromatic); ³¹P

Meth yl 1,6-Bis-O-(tr ip h en ylm eth yl)-3,4-bis-O-[bis(3,5-

d im eth ylp h en yl)p h osp h in o]-r-^D-fr u ctofu r a n osid e (31b).

¹H NMR 1.85, 1.91, 1.94, 2.05 (4 s, 3H each), 3.10 (s, 3H), 3.45-

3.60 (ABX, J _AB) 9, J _AX) J _BX) 5, 2H), 3.67, 3.80 (ABq, J _AB

) 10, 2H), 4.47 (q m, br, 1H), 5.63 (d, J ) 11 Hz, 1 H), 5.20

(m, 1 H), 6.50-7.80 (m, aromatic); ³¹P NMR (C₆D₆) 116.41 (d,

NMR 120.4 (d, J _PP) 4), 115.7 (d, J _PP) 4).

-

[(26b)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆) ABX () PPRh),ν_A

J _PP) 8), 118.53 (d, J _PP) 8).

) 118.9, ν_B) 126.6, J _PP) 34, J _RhP) 170.

-

[(31b)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆): 114.2 (d), 131.5

Met h yl 2-Acet a m id o-6-O-(ter t-b u t yld im et h ylsilyl)-2-

d eoxy-3,4-O-b is(d ip h en ylp h osp h in o)-â-^D-glu cop yr a n o-

sid e (26e). ¹H NMR (C₆D₆) -0.08 (s, 3 H), -0.02 (s, 3 H),

0.90 (s, 9 H), 2.11 (s, 3 H), 3.30 (s, 3 H), 3.35-3.55 (m, 4 H),

4.48 (q, J ) 8, 1 H), 4.72 (d, J ) 8, 1 H), 5.05 (d, J ) 8, 1 H),

5.15 (q, J ) 8, 1 H), 6.70-7.60 (m, aromatic); ³¹P NMR (C₆D₆)

(dd), J _RhP) 169, J _PP) 28.

Met h yl 1,6-Bis-O-(t r ip h en ylm et h yl)-3,4-b is-O-[b is(4-

m eth oxyp h en yl)p h osp h in o]-r-^D-fr u ctofu r a n osid e (31c).

¹H NMR (C₆D₆) 3.05-3.30 (4 s total 15 H), 3.40, 3.50 (ABX,

J _AB) 10, J _AX) 7, J _BX) 6, 2 H), 3.61, 3.79 (AB, J _AB) 10, 2

H), 4.58 (dd m, br, 1H), 4.90 (m, 1H), 5.05 (d, J ) 10, 1 H),

6.42-7.61 (m, aromatic); ³¹P NMR (C₆D₆) 115.0 (d), 115.2 (d),

112.70 (d, J _PP) 5), 117.17 (d, J _PP) 5).

-

[(26e)Rh (COD)]⁺SbF ₆

.

³¹P NMR (C₆D₆) ABX () PPRh),

J _PP) 7.

ν_A) 122.5, ν_B) 129.2, J _PP) 35, J _RhP) 173.

Liga n d s (28) a n d Ca ta lysts fr om Meth yl 6-O-(ter t-

Bu tyld im eth ylsilyl)-2-d eoxy-r-^D-glu cop yr a n osid e.

-

[(31c)Rh (COD)]⁺SbF ₆

.

³¹P NMR (C₆D₆) ABX () PPRh),

ν_A) 121.8, ν_B) 122.1, J _PP) 27, J _RhP) 167.

[(31c)Rh (COD)]⁺OTf^-. ³¹P NMR (C₆D₆) ABX () PPRh),

ν_A) 121.3, ν_B) 121.9, J _PP) 28, J _RhP) 166.

Meth yl 1,6-Bis-O-(tr ip h en ylm eth yl)-3,4-bis-O-(d ip h en -

ylp h osp h in o)-r-^D-fr u ctofu r a n osid e (31e). ¹H NMR (C₆D₆)

3.10 (s, 3H), 3.35, 3.45 (ABX, J _AB) 10, J _AX) 7, J _BX) 6, 2 H),

3.60, 3.78 (AB, J _AB) 10, 2 H), 4.50 (ddm, br, 1H), 4.88 (m,

1H), 5.00 (d, J ) 10, 1 H), 6.80-7.80 (m, aromatic); ³¹P NMR

Meth yl 2-Deoxy-3,4-bis-O-[bis(3,5-dim eth ylph en yl)ph os-

p h in o]-r-^D-glu cop yr a n osid e (28b). ¹H NMR inter alia

0.11-0.12 (2 s, 3 H each), 1.11 (s, 9 H), 2.12 (s, 6 H), 2.14 (s,

6 H), 2.16 (s, 6 H), 2.21 (s, 6 H), 3.14 (s, 3 H), 4.45 (d, J ) 3,

1 H), 4.64 (m, 1 H), 5.20 (m, 1 H), 6.60-7.61 (m, aromatic);

³¹P NMR 121.1 (d, J _PP) 2), 113.1 (d, J _PP) 2).

(C₆D₆) 114.2, 115.1, J _PP) 9.

-

[(31e)Rh (COD)]⁺SbF ₆

.

³¹P NMR (C₆D₆) ABX () PPRh),

ν_A) 119.7, ν_B) 122.8 J _PP) 29, J _RhP) 166.

Liga n d s a n d Ca ta lysts fr om (1S,2S)-tr a n s-1,2-cyclo-

h exa n ed iol. (1S,2S)-1,2-Bis-O-(d ip h en ylp h osp h in o)cy-

cloh exa n e. To 200 mg (1.72 mmol) of azeotropically dried

(1S,2S)-trans-1,2-cyclohexanediol, 11 mg (0.05 equiv) of 4-(di-

methylamino)pyridine, and 1 mL of pyridine in 2 mL of CH₂Cl₂

at 0 °C and under nitrogen was added 0.836 g (3.788 mmol) of

[(28b)Rh (COD)]⁺OTf^-. ³¹P NMR (C₆D₆) ABX () PPRh),ν_A

) 123.6, ν_B) 128.2, J _PP) 34, J _RhP) 173.

-

[(28b)Rh (COD)]⁺BF ₄

.

³¹P NMR (C₆D₆) ABX () PPRh)

ν_A) 124.9, ν_B) 127.4, J _PP) 33, J _RhP) 173.

Carbohydrate Phosphinites in Asymmetric Catalysis

J . Org. Chem., Vol. 62, No. 17, 1997 6027

chlorodiphenylphosphine dissolved in 2 mL of methylene

chloride. An additional 2 mL of methylene chloride was used

to facilitate transfer of the chlorophosphine quantitatively. The

mixture was slowly allowed to come to room temperature and

was stirred for 18 h. The precipitated solids were filtered off

with the aid of a cotton-plugged disposable pipette. The NMR

spectrum of this solid showed that it was pyridine hydrochlo-

ride. The vials were washed with 2 mL of 1:1 ether/hexanes,

and the solid in the disposable pipette was washed with this

solvent. The filtrate was concentrated to get a white solid.

The product was recrystallized from benzene. ¹H NMR (C₆D₆)

0.70-0.95 (m, 2 H), 1.10-1.45 (m, 4 H), 1.90-2.05 (m, 2 H),

4.05 (m, 2 H), 6.70-7.15 (m, aromatic), 7.45-7.70 9 (m,

aromatic); ³¹P NMR 108.7.

When a methyl ester of an N-acetylamino acid derivative of 3

was used, the crude mixture was analyzed directly by gas

chromatography (Chirasil S-Val, 170 °C/20 min, programmed

for 10 °C/min increase to 195 °C, 195 °C/30 min. (See Table 7

for retention times. See also the attached chromatograms in

the Supporting Information). However, when an N-acetyl-

amino acid acid was employed, the crude mixture was treated

with excess diazomethane (Caution: diazomethane is known

to be explosive, Aldrich mini-Diazald kit was used to prepare

the diazomethane) to convert the acid to the corresponding

methyl ester for GC analysis. See Tables 1-5 for results. The

assignments of structures for all compounds were confirmed

by comparison with authentic samples.²⁹

[(1S,2S)-1,2-[(Dip h en ylp h osp h in o)oxy]cycloh exa n e]

Ta ble 7. GC Reten tion Tim es (m in ) of Selected Am in o

Acid Der iva tives^a

-

Rh ⁺(COD) SbF ₆

. The complex was prepared by adding a

CH₂Cl₂solution of the bis(phosphinite) to a solution of (COD)₂

Rh⁺SbF₆in CH₂Cl₂at room temperature and stirring the

solution overnight. The low-boiling components were removed

and the NMR spectra were recorded in CDCl₃. ¹H NMR 0.70-

1.15 (m, 4 H), 1.35-1.60 (d, br, 2H), 1.78 (d, br, 2 H), 2.10-

2.30 (m, 4 H), 2.32-2.60 (m, 4 H), 3.82-4.02 (m, 2 H), 4.42 (q,

br, 2 H), 4.90 (s, br, 2 H), 7.00 -7.80 (m, aromatic); ³¹P NMR

131.0 (d, J ) 177). Minor peaks (together <2%) at 121.7 (d)

and 123.1 (d) were also observed. The compound was recrys-

tallized from benzene by slow evaporation. X-ray structure

analysis on this compound confirmed the structure.

X-r a y Cr ysta llogr a p h y. An orange prism of 34 grown

from benzene by slow evaporation with dimensions 0.35 × 0.19

× 0.37 mm was mounted in a glass capillary under argon and

placed on an Enarf-Nonius CAD4 diffractometer equipped with

a Mo KR source. The data was collected at -55 °C.

Crystal data: SbRhP₂F₆O₂C₅₀H₅₄; monoclinic-b, 12 a )

19.333(3), b ) 12.042(2), c ) 20.920(3) Å, â ) 94.94(1)°, T )

-55 °C, V ) 4852.3 Å³, Z ) 4, FW )1087.59, D_c) 1.489g/mL,

µ(Mo) ) 10.18 cm^-1

.

Data collection: Enarf-Nonius CAD4 diffractometer equipped

with a Mo KR source, 6577 data collected, 2.8^oe 2θ e 55.0°,

maximum h,k,l ) 25 15 27, data octants ) +++, -++, ω scan

method, scan width ) 1.20-2.20° ω, scan speed ) 1.70-5.00°/

min, typical half-height peak width ) 0.12° ω, 2 standards

collected 30 times, 1% fluctuation, 13.3% variation in azi-

muthal scan, no absorption correction, 156 duplicates, 1.6%

R-merge, 3629 unique reflections with I g 3.0σ(I).

Solution and refinement: Structure was solved by auto-

mated Patterson analysis (PHASE). The asymmetric unit

consists of one cation in a general position, with two half SbF₆

anions on special positions, and two partially occupied benzene

solvate molecules. Hydrogen atoms were idealized with C-H

) 0.95 Å, refinement by full-matrix least squares on F,

scattering factors from International Tables for X-ray Crystal-

lography, Vol. IV, including anomalous terms for Sb, Rh, P,

biweight

[σ²(I) + 0.00091²]^-1/2(excluded 5), refined aniso-

tropic: Sb, Rh, P, F, O, C, isotropic: F, C, fixed atoms: H,

498 parameters, data/parameter ratio ) 7.28, final R ) 0.051,

R_w) 0.044, error of fit ) 1.40, maximum ∆/σ ) 0.12. The

SbF₆disorder was modeled with partial F positions to yield a

correct formula unit, while the benzene occupancies were

adjusted to yield reasonable thermal parameters. The abso-

lute configuration corresponds to the enantiomorph with the

lowest R-value (5.09 vs 5.14%). Largest residual density )

0.58 e/Å³, near Rh1.

Exa m p les of La bor a tor y Sca le P r ep a r a tion s: (S)-N-

Acetylp h en yla la n in e Meth yl Ester . In the dry box, a 150

mL Fischer-Porter tube was loaded with 1.0 g (4.9 mmol) of

(Z)-acetamidocinnamic acid (Aldrich, recrystallized), 0.62 mg

of the catalyst (0.00048 mmol; 1/5000 equiv of catalyst,

measured by using 206 µL of a freshly prepared stock solution

Fractional coordinates, isotropic and anisotropic thermal

parameters and a complete listing of bond angles and bond

lenths are given in the Supporting Information.

Typ ica l Scou tin g P r oced u r e for th e Hyd r ogen a tion of

Aceta m id oa cr yla te Der iva tives (3, eq 3). In the dry box,

a 150 mL Fischer-Porter tube was loaded with 0.25 mmol of

methyl acetamidoacrylate, 1 mg (0.4 mol %) of 11b, and 2 mL

of THF or glyme. After sealing, the tube was removed from

the dry box and placed behind proper shielding. With ad-

equate stirring of the solution at room temperature, the tube

was first charged with 40 psi of H₂gas and was subsequently

evacuated. This procedure of filling and evacuation was

repeated twice more. The tube was charged with 30 to 40 psi

of H₂before stirring for 3 h. The conversion was quantitative

in most instances as judged by gas chromatographic analysis.

-

of 30 mg of [(11b)Rh (COD)]⁺SbF ₆in 5.0 mL of THF), and

8 mL of THF. After sealing, the tube was removed from the

dry box and placed behind proper shielding. With adequate

stirring of the solution at room temperature, the tube was

charged with 30 psi of H₂gas and vented. This procedure was

repeated twice. The tube was charged with 30 psi of H₂and

recharged as necessary to maintain 30 psi. After 16 h, the

tube was vented. A small sample of the crude mixture was

treated with diazomethane giving N-acetyl S-phenylalanine

methyl ester in 97.8% ee. The remainder was recrystallized

from 95% ethyl acetate/hexane to give 881 mg (88%) of (S)-

6028 J . Org. Chem., Vol. 62, No. 17, 1997

RajanBabu et al.

N-acetylphenylalanine with 99.7% ee (determined by GC on

the methyl ester).

column of silica using 40% ethyl acetate/hexane as eluant, and

the product was collected. GC retention time on an HP

methylsilicone column under conditions reported earlier at 180

°C was 11.43 min. Chiral GC on chirasil-^L-Val showed two

incompletely separated peaks at retention times 25.2 and 25.7

min, respectively, with the (R)-isomer appearing first. HPLC

analysis on the Chiralcel OJ column using 92.5% hexane/i-

PrOH at 1 mL/min as eluant showed 97% ee [retention time

(min): (S)-isomer 24.4 min, (R)-isomer 27.8 min]. 2-[N-(Ben-

zyloxycarbonyl)amino]-3-(4-fluorophenyl)-1-propanol: mp 95-

The reaction is sensitive to the purity of the starting

dehydroamino acid precursor. The same reaction repeated on

different batches of starting material (Interchem) with 1/10000

equiv of the catalyst (51 µL of the above solution) at room

temperature and 40 psi hydrogen on a scale of 0.500 g provided

(S)-N-acetylphenylalanine of 99.0% ee. Depending upon the

purity of the starting material, we have observed variations

as much as 81.0 to 99.0% ee for this and other hydrogenations.

Purification of the starting dehydroamino acid derivatives by

recrystallization/column chromatography is highly recom-

mended.

96 °C; [R]²⁵) -24.6 ( 0.8° (CHCl₃, c 1); ¹H NMR (CDCl₃)

D

2.10-2.30 (s, br, 1 H), 2.83 (d, 2, J ) 7, 2 H), 3.55, 3.68 (ABX,

J _AB) 13, J _AX) 4, J _BX) 5, 2 H), 3.90 (m, br 1 H), 5.01 (s, br,

1H), 5.16 (s, 2), 6.90-7.50 (m, aromatic); IR (CHCl₃) 3630,

3438, 1716 cm ^-1. Anal. C 67.10; H 6.04; N 4.64. Calcd for

C₁₇H₁₈FO₃N C 67.31; H 5.98; N 4.62.

The enantioselectivity was confirmed by converting the

N-Cbz derivative to the corresponding N-Ac derivative. For

this, the Cbz derivative was hydrogenated at 30 psi using 10%

Pd/C in 5:1 methanol/acetic anhydride. The solvents were

removed on the pump, and the product was analyzed by NMR,

GC, and HPLC. Chiral GC (Chirasil-^L-Val column, 160 °C)

and Chiral HPLC (Chiralcel OJ , 7.5% i-PrOH/hexane) gave

base-line separations of the N-Ac derivative, and the ee was

ascertained to be 96% by HPLC and 95.7% by GC.

P r ep a r a tion of Eith er An tip od e of N-Acetyl-3-(3-th ie-

n yl)a la n in e Meth yl Ester : Hyd r ogen a tion of Meth yl (Z)-

N-Acetyl-3-(3-th ien yl)-2-p r op en oa te. The (S)-isomer (pre-

pared using ligand 11b): crude ee 96.7% (GC) and 99%

(HPLC); [R]²⁰(CHCl₃, c 1) 99 ( 0.8°; HPLC (Chiralcel OJ ,

D

10% 2-propanol/hexanes, 1 mL/min) 13.00 min. Recrystallized

ee 99.5%; mp 114 °C; [R]²⁰100 ( 0.8° (c 1, CHCl₃). Anal. C

D

53.33; H 5.72; N 5.96; S 14.01. Calcd C 52.85; H 5.77; N 6.16;

S 14.11.

The (R)-isomer (prepared using ligand 20b): crude ee 97.0%;

Recrystallized ee 99.5%; mp 107-109 °C (reported^36b106-108

°C); [R]²⁰_D-15.6 ( 0.8° (EtOH, c 1; reported^36b-14.66); -95.0

( 0.8° (CHCl₃, c 1); HPLC (Chiralcel OJ , 10% 2-propanol/

hexanes, 1 mL/min) 10.6 min.

For this reduction, the following selectivities were observed

for various catalysts: [11b]Rh⁺(COD) SbF₆^-96% (S); [11c]Rh⁺-

(COD) SbF₆^-85% (S); [11e]Rh⁺(COD) SbF₆^-62% (S); [11f]Rh⁺-

H yd r ogen a t ion of Met h yl (Z)-N-Acet yl-3-[3,5-b is(t r i-

flu or om eth yl)p h en yl]-2-p r op en oa te: N-Acetyl-3-[3,5-bis-

(tr iflu or om eth yl)p h en yl]a la n in e Meth yl Ester . The (R)-

isomer (prepared using ligand 20b): crude ee 97.4%; ee of

recrystallized (10% ethyl acetate/hexane) sample: 99.4%; mp

-

(COD) SbF₆<3% (S); [11g]Rh⁺(COD) SbF₆<5% (S);

[20b]Rh⁺(COD) SbF₆^-90% (R); [27b]Rh⁺(COD) BF₄^-96% (R).

Rela tive Ra tes of Hyd r ogen a tion of a Deh yd r oa m in o

-

Acid Ester Usin g [11b]Rh ⁺(NBD) SbF ₆a n d [(R,R)-Me-

107-109 °C (reported²⁹111-112 °C); [R]²⁵-78.8 ( 0.8°

DUP HOS] Rh ⁺(COD) TfO^-in THF . In a dry box, two

Fischer-Porter tubes were charged with methyl (Z)-N-acetyl-

3-[3,5-bis(trifluoromethyl)phenyl]-2-propenoate (75 mg, 0.211

mmol) in 3 mL of THF. To one was added 0.25 mg (0.0002

mmol, as a THF solution of known concentration) of [11b]Rh⁺-

(NBD) SbF₆^-and to the other 0.14 mg (0.0002 mmol, as a THF

solution of known concentration) of [(R,R)-Me-Duphos]Rh⁺(COD)

TfO^-. The tubes were brought outside the box and were

connected to a hydrogen cylinder. The catalytic hydrogenation

was accomplished as described before at 40 psi of hydrogen

after three cycles of purging and evacuation. The tubes were

evacuated at 15 min and refilled with nitrogen. The contents

were analyzed by gas chromatography to determine the extent

of reaction and enantioselectivities. The results are shown in

Scheme 10.

D

(CHCl₃, c 1) reported²⁹-75.6°.

The (S)-isomer (prepared using ligand 11b): crude ee 95.8%;

recrystallized (from 10% ethyl acetate/hexane) 99.3% ee; [R]²⁵

D

-74.4 ( 0.8° (CHCl₃, c 1).

Exa m p le of Hyd r ogen a tion of a Deh yd r oa m in o Acid

w ith a Cbz P r otectin g Gr ou p on Nitr ogen . In the dry box,

a 150 mL Fischer-Porter tube was loaded with 50 mg (0.15

mmol) of methyl (Z)-2-[N-(benzyloxycarbonyl)amino]-3-(4-fluo-

rophenyl)prop-2-enoate, 1 mg of 11b, and 2 mL of THF. After

sealing, the tube was removed from the dry box and placed

behind proper shielding. With adequate stirring of the solu-

tion at room temperature, the tube was charged with 30 psi

of H₂gas and vented. This procedure was repeated twice. The

tube was charged with 30 psi of H₂and recharged as necessary

to maintain 30 psi. After 16 h, the tube was vented. The crude

mixture concentrated provided an oil in nearly quantitative

yield. HPLC columns Chiralcel OJ or OB did not separate

the two enantiomers using an 2-propanol/hexane solvent

system.

Ack n ow led gm en t. We thank W. M. Gray for tech-

nical assistance and W. A. Nugent and J . E. Feaster

for providing several of the starting dehydroamino acid

precursors. Financial support by The Dupont Company

(Aid to Education Fund) and Hoechst Marion Roussel

are gratefully acknowledged. Acknowledgment is also

made to the donors of The Petroleum Research Fund,

administered by the American Chemical Society, for

partial support of this research.

Purification by column chromatography on silica gel using

30% ethyl acetate/hexanes as solvent gave 44 mg (89%) of a

clear oil identified as the desired product from the following

data and its subsequent reactions. ¹H NMR (CDCl₃) 3.08 (d,

AB, J ) 6, J ) 13, 2 H), 3.72 (s, 3 H), 4.65 (ddd, J ) 7, 7, 6,

1 H), 5.10 (AB, J ) 12, 2 H), 5.21 (d, br, 1 H), 6.90-7.10 (m,

4 H), 7.20-7.42 (m, 5 H). [R]²⁵) +47.2 ( 0.8° (CHCl₃, c 1).

D

Su p p or tin g In for m a tion Ava ila ble: Conditions for chro-

matographic analysis, typical chromatograms of R and S

amino acid derivatives, and X-ray diffraction data for 34

including fractional coordinates, isotropic and anisotropic

thermal parameters, and a complete listing of bond angles and

bond lengths (19 pages). This material is contained in libraries

on microfiche, immediately follows this article in the microfilm

version of the journal, and can be ordered from ACS; see any

current masthead page for ordering information.

Enantiomeric purity was established on the corresponding

alcohol obtained via LiBH₄reduction or from the N-acetyl

derivative as described below.

To the crude product obtained from the above reaction using

0.1 g of starting dehydro derivative was added 0.25 mL of a

2.0 M solution of LiBH₄in THF. After 2 h, 5 mL of water was

added, and the mixture was stirred until gas evolution stopped.

The product was extracted into ether. The organic phase was

washed with water and dilute NaHCO₃, dried (MgSO₄), and

concentrated. The crude mixture was passed through a short

J O970884D

Article Doi

Carbohydrate Phosphinites as Practical Ligands in Asymmetric Catalysis: Electronic Effects and Dependence of Backbone Chirality in Rh-Catalyzed Asymmetric Hydrogenations. Synthesis of R- or S-Amino Acids Using Natural Sugars as Ligand Precursors

DOI: 10.1021/jo970884d

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo970884d

Full text of DOI:10.1021/jo970884d

Products guided by the article

R&D Labs maybe for 128804-06-8

Relevant to this article

Hot Product