Origins of enantioselectivity with nitrogen-sulfur chelate ligands in palladium-catalyzed allylic substitution

Article abstract of DOI:10.1021/jo9910177

The reaction of 1,3-diphenyl-2-propenyl acetate (9) with dimethylmalonate to give the substitution product 10 is effectively catalyzed by Pd complexes containing chiral imine-sulfide chelate ligands derived from amino acids. The ligand of choice, (S)-N-2′-chlorobenzylidene-2-amino-3-methyl-1-thiophenylbutane (6e), prepared in only two steps from (S)-valinol, gave an ee of 94%. Because the explanation of selectivity with the majority of other nitrogen-sulfur chelate ligands in this reaction assumes the selectivity to be controlled by an electronic bias, which contradicts our results, we characterized the Pd-allyl intermediate 14 by X-ray diffraction and solution NMR. The possible mechanism of chirality transfer is discussed. The site of nucleophilic attack on the allyl ligand is not trans to the perceived better π-acceptor ligand (sulfur), which would be analogous to chiral nitrogen-phosphorus systems. This reaction occurs trans to the imine donor, and the enantioselectivity is ultimately controlled by the subtle steric environment of the chiral imine-sulfur chelate ligand, which predisposes the allyl unit of the reaction intermediate to a preferred reaction trajectory. In light of results that emphasize the power of electronic desymmetrization for chiral recognition, these results suggest that electronically dissimilar ligands may not give rise to chiral recognition through electronic dissimilarity.

Full text of DOI:10.1021/jo9910177

8256
J . Org. Chem. 1999, 64, 8256-8262
Or igin s of En a n tioselectivity w ith Nitr ogen -Su lfu r Ch ela te
Liga n d s in P a lla d iu m -Ca ta lyzed Allylic Su bstitu tion
Harry Adams,† J ames C. Anderson,*,† Rachel Cubbon,† Daniel S. J ames,† and
J ohn P. Mathias‡
Department of Chemistry, University of Sheffield, Sheffield, S3 7HF U.K., and Pfizer Central Research,
Sandwich, Kent, CT13 9NJ U.K.
Received J une 25, 1999
The reaction of 1,3-diphenyl-2-propenyl acetate (9) with dimethylmalonate to give the substitution
product 10 is effectively catalyzed by Pd complexes containing chiral imine-sulfide chelate ligands
derived from amino acids. The ligand of choice, (S)-N-2′-chlorobenzylidene-2-amino-3-methyl-1-
thiophenylbutane (6e), prepared in only two steps from (S)-valinol, gave an ee of 94%. Because the
explanation of selectivity with the majority of other nitrogen-sulfur chelate ligands in this reaction
assumes the selectivity to be controlled by an electronic bias, which contradicts our results, we
characterized the Pd-allyl intermediate 14 by X-ray diffraction and solution NMR. The possible
mechanism of chirality transfer is discussed. The site of nucleophilic attack on the allyl ligand is
not trans to the perceived better π-acceptor ligand (sulfur), which would be analogous to chiral
nitrogen-phosphorus systems. This reaction occurs trans to the imine donor, and the enantiose-
lectivity is ultimately controlled by the subtle steric environment of the chiral imine-sulfur chelate
ligand, which predisposes the allyl unit of the reaction intermediate to a preferred reaction trajectory.
In light of results that emphasize the power of electronic desymmetrization for chiral recognition,
these results suggest that electronically dissimilar ligands may not give rise to chiral recognition
through electronic dissimilarity.
Since the first example of an enantioselective pal-
bias upon intermediate π-allyl complexes.8 Stereoelec-
tronically, the palladium-allyl terminus bond opposite
the more powerful acceptor atom (phosphorus) will be
longer and hence more susceptible to cleavage as a result
of nucleophilic attack.9 Such physical attributes have
been verified by NMR and X-ray studies and are claimed
to control enantioselection.8f,10 Heterobidentate nitrogen-
sulfur ligands can also control the enantioselection of this
process,11-15 but many researchers rationalize results by
ladium-catalyzed allylic substitution reaction with a
stabilized nucleophile,1 there have been numerous re-
ports of the design and synthesis of chiral bidentate
ligands to affect such a process.2 The reaction is charac-
terized by bond breaking and bond making on the face
of the π-allyl unit opposite palladium and its requisite
chiral environment, provided by judiciously chosen ligands.
The development of effective chiral ligands for this
process has been arduous, as the stereochemical informa-
tion inherent to the chiral ligand has to be transmitted
across the plane of the allyl fragment to the bond-
forming/-breaking event that is responsible for enantio-
discrimination. Homobidentate C2-symmetric ligands
have been extensively investigated and transition-state
models proposed.3-8 More recently, heterobidentate ni-
trogen-phosphorus chiral ligands have been reported
which capitalize upon the difference in electronic char-
acter of the two donor atoms to exert a stereoelectronic
(8) (a) Sprinz, J .; Helmchen, G. Tetrahedron Lett. 1993, 34, 1769-
72. (b) von Matt, P.; Pfaltz, A. Angew. Chem., Int. Ed. Engl. 1993, 32,
566-8. (c) Dawson, G. J .; Frost, C. G.; Williams, J . M. J .; Coote, S. J .
Tetrahedron Lett. 1993, 34, 3149-50. (d) Yang, H.; Khan, M. A.;
Nicholas, K. M. Organometallics 1993, 12, 3485-94. (e) von Matt, P.;
Loiseleur, O.; Koch, G.; Pfaltz, A.; Lefeber, C.; Fuecht, T.; Helmchen,
G. Tetrahedron: Asymmetry 1994, 5, 573-84. (f) Brown, J . M.; Hulmes,
D. I.; Guiry, P. J . Tetrahedron 1994, 50, 4493-506. (g) Burckhardt,
U.; Hintermann, L.; Schnyder, A.; Togni, A. Organometallics 1995, 14,
5415-25. (h) Saitoh, A.; Misawa, M.; Morimoto, T. Synlett 1999, 483-
5.
(9) (a) A° kermark, B.; Hansson, S.; Krakenberger, B.; Vitagliano, A.;
Zetterberg, K. Organometallics 1984, 3, 679-82. (b) A° kermark, B.;
Hansson, S.; Vitagliano, A. J . Am. Chem. Soc. 1990, 112, 4587-8. (c)
A° kermark, B.; Krakenberger, B.; Hansson, S.; Vitagliano, A. Organo-
metallics 1987, 6, 620-8.
(10) (a) Sprinz, J .; Kiefer, M.; Helmchen, G.; Reggelin, M.; Huttner,
G.; Walter, O.; Zsolnai, L. Tetrahedron lett. 1994, 35, 1523-6. (b) 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, 265-84. (c) Togni, A.; Burckhardt, U.; Gramlich, V.;
Pregosin, P. S.; Salzmann, R. J . Am. Chem. Soc. 1996, 118, 1031-7.
(d) Steinhagen, H.; Reggelin, M.; Helmchen, G. Angew. Chem., Int. Ed.
Engl. 1997, 36, 2108-10. (e) Selvakumar, K.; Valentini, M.; Wo¨rle,
M.; Pregosin, P. S. Organometallics 1999, 18, 1207-15.
(11) (a) Frost, C. G.; Williams, J . M. J . Tetrahedron: Asymmetry
1993, 4, 1785-8. (b) Frost, C. G.; Williams, J . M. J . Tetrahedron Lett.
1993, 34, 2015-8. (c) Dawson, G. J .; Frost, C. G.; Martin, C. J .;
Williams, J . M. J .; Coote, S. J . Tetrahedron Lett. 1993, 34, 7793-6.
(d) Allen, J . V.; Coote, S. J .; Dawson, G. J .; Frost, C. G.; Martin, C. J .;
Williams, J . M. J . J . Chem. Soc., Perkin Trans. 1 1994, 2065-72.
(12) (a) Chelucci, G.; Cabras, M. A. Tetrahedron: Asymmetry 1996,
7, 965-6. (b) Chelucci, G.; Berta, D.; Saba, A. Tetrahedron 1997, 53,
3843-8.
† University of Sheffield.
‡ Pfizer Central Research.
(1) Trost, B. M.; Strege, P. E. J . Am. Chem. Soc. 1977, 99, 1649-
51.
(2) Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96, 395-
422.
(3) (a) Hayashi, T. Pure Appl. Chem. 1988, 60, 7-12. (b) Hayashi,
T.; Yamamoto, A.; Ito, Y.; Nishioka, E.; Miura, H.; Yanagi, K. J . Am.
Chem. Soc. 1989, 111, 6301-11. (c) Hayashi, T.; Kishi, K.; Yamamoto,
A.; Ito, Y. Tetrahedron Lett. 1990, 31, 1743-6.
(4) Yamaguchi, M. Tetrahedron: Asymmetry 1991, 2, 663-6.
(5) Pfaltz, A. Acc. Chem. Res. 1993, 26, 339-45.
(6) (a) Trost, B. M.; Van Vranken, D. L.; Bingel, C. J . J . Am. Chem.
Soc. 1992, 114, 9327-43. (b) Trost, B. M.; Bunt, R. C. J . Am. Chem.
Soc. 1994, 116, 4089-90.
(7) (a) Dierkes, P.; Ramdeehul, S.; Barloy, L.; De Cian, A.; Fischer,
J .; Kamer, P. C. J .; van Leeuwen, P. W. N. M.; Osborn, J . A. Angew.
Chem., Int. Ed. 1998, 37, 3116-8. (b) Ramdeehul, S.; Dierkes, P.;
Aguado, R.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.; Osborn, J . A.
Angew. Chem., Int. Ed. 1998, 37, 3118-21.
10.1021/jo9910177 CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/14/1999

Enantioselectivity in Pd-Catalyzed Allylic Substitution
J . Org. Chem., Vol. 64, No. 22, 1999 8257
analogytonitrogen-phosphoruschelatingsystems.10a,11d,13a,14
As part of an ongoing research program, we have
developed a new series of heterodentate chelate ligands,
derived from amino acids, where the chirality inherent
to the backbone of the amino acid can be transmitted
closer to the reaction center due to the correct choice of
donor groups and substituents.16 The emphasis of this
work is on understanding the origins of chirality induc-
tion in our ligand systems to aid the design of simpler,
more efficient chiral ligands for asymmetric catalysis.
Recently we reported that the sulfur-imine mixed donor
chiral ligand 1 (R ) 2-ClC6H4, R′ ) iPr, R′′ ) Ph), derived
from (S)-valinol, gave 94% enantiomeric excess in the
palladium-catalyzed allylic substitution reaction between
dimethyl malonate and (()-(E)-diphenyl-2-propenyl ac-
etate.17 We proposed a transition-state model, 2, based
upon our results and the accepted notion that nucleo-
philic attack of the π-allyl complex occurs trans to the
better π-acceptor end of the chelate ligand from the most
stable Pd-allyl intermediate.8f,10,18 In the absence of any
data, we argued that in our ligand the imine group is
the better π-acceptor, as thioethers are considered poor
π-acceptors.19
derived from (S)-valinol via the intermediate sulfide 4.
This key compound can be synthesized by exchange of
the hydroxyl group for the phenylsulfide in a moderate
yield (58%) up to a 1 mmol scale using the Hata protocol,
but problems arose with the separation of 4 from the
excess of reagents and byproducts when the reaction was
scaled up. An alternative series of straightforward reac-
tions, starting from known 5,21 provided a similar overall
yield and was amenable to larger scale synthesis. Esch-
weiler-Clarke N,N-dimethylation22 furnished 3d in 93%
yield (eq 1).
Imines 6a -h were synthesized from 4 by condensation
with the requisite aromatic aldehyde in the presence of
magnesium sulfate. Imine 6i was synthesized by stirring
Our explanation contradicts those offered in the lit-
erature to account for the sense of enantioselection of
other chiral nitrogen-sulfur ligands.10a,11d,13a,14 Aside from
one report dealing with a NMR analysis of the mecha-
nism of enantioselection,15 there is scant structural
evidence reported for these types of ligands in this
reaction. We have investigated the steric and electronic
factors which account for the high enantioselection of this
ligand system using a combination of single-crystal X-ray
crystallography and NMR.
amine 4 with freshly distilled acetaldehyde in the pre-
sence of basic alumina. The imines were judged >95%
1
pure by H NMR and, as they were unstable to chroma-
tography, were used crude. Ligands 7a ,b were most
quickly prepared from (S)-tert-leucinol and (R)-phenyl-
glycinol using the Hata protocol (37% and 46% respec-
tively), followed by standard quantitative imine forma-
tion. Ligands 8a ,b were synthesized in good yield ac-
cording to eq 1 using the appropriate sodium thiolate.
Asym m etr ic Alk yla tion . Our studies used the popu-
lar test reaction involving the substitution of 1,3-diphe-
nyl-2-propenyl acetate 9 with the nucleophile derived
from the reaction of dimethylmalonate and N,O-bis-
(trimethylsilyl)acetamide (BSA) using potassium acetate
as catalyst (eq 2).23 This test reaction was chosen as the
Resu lts
Syn th esis of Liga n d s. Amine ligands 3a ,b were
synthesized according to our previously reported route.16b
Ligand 3c was synthesized by subjection of (S)-N-methyl-
N-phenyl-2-amino-3-methylbutan-1-ol16b to the Hata pro-
tocol20 in a sealed tube for 3 days (96%). Ligand 3d was
(13) (a) Chesney, A.; Bryce, M. R. Tetrahedron: Asymmetry 1996,
7, 3247-54. (b) Chesney, A.; Bryce, M. R.; Chubb, R. W. J .; Batsanov,
A. S.; Howard, J . A. K. Tetrahedron: Asymmetry 1997, 8, 2337-46.
(14) Morimoto, T.; Tachibana, K.; Achiwa, K. Synlett 1997, 783-5.
(15) Boog-Wick, K.; Pregosin, P. S.; Trabesinger, G. Organometallics
1998, 17, 3254-64.
(18) (a) Auburn, P. R.; Mackenzie, P. B.; Bosnich, B. J . Am. Chem.
Soc. 1985, 107, 2033-46. (b) Mackenzie, P. B.; Whelan, J .; Bosnich,
B. J . Am. Chem. Soc. 1985, 107, 2046-54.
(19) Mu¨ller, A.; Diemann, E. In Comprehensive Coordination Chem-
istry; Wilkinson, G., Ed.; Pergamon: Oxford, UK, 1987; Vol. 2, pp 515-
58.
(16) (a) Anderson, J . C.; Harding, M. Chem. Commun. 1998, 393-
4. (b) Anderson, J . C.; Cubbon, R.; Harding, M.; J ames, D. S.
Tetrahedron: Asymmetry 1998, 9, 3461-90.
(17) Anderson, J . C.; J ames, D. S.; Mathias, J . P. Tetrahedron:
Asymmetry 1998, 9, 753-6.

8258 J . Org. Chem., Vol. 64, No. 22, 1999
Adams et al.
Ta ble 1. Asym m etr ic Alk yla tion of 9 Ca ta lyzed by
P a lla d iu m -1 Com p lexes
yielda eeb 10 (%)
10 (%) (abs conf)
ligand
R
R1
R2
6a
6b
6c
6d
6e
6f
6g
6h
6i
7a
7b
8a
8b
Ph
iPr
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
86
77
85
90
87
78
19
27
88
30
35
86
88
89 (R)
82 (R)
88 (R)
84 (R)
94 (R)
89 (R)
90 (R)
94 (R)
84 (R)
65 (S)
78 (R)
90 (R)
96 (R)
4-O2NC6H4
4-MeOC6H4
iPr
iPr
1,2,6-Me3C6H2 iPr
2-ClC6H4
4-ClC6H4
2,6-Cl2C6H3
2-BrC6H4
Mec
2-ClC6H4
2-ClC6H4
2-ClC6H4
2-ClC6H4
iPr
iPr
iPr
iPr
iPr
(R)-Ph Ph
(S)-tBu Ph
iPr
iPr
Me
2,6-Me2C6H3
a
b
Isolated yield. Determined by 1H NMR using the chiral shift
reagent Eu(hfc)3. c Ligand contained ∼2% of the syn isomer by 1H
NMR.
F igu r e 1. Possible transition-state models leading to (R)-10.
(compare ligand 6i with 6b and 6d ). The increase in steric
bulk of the backbone chiral center was investigated by
preparation of the analogous ligands 7a and 7b from (R)-
phenylglycinol and (S)-tert-leucinol (entries 11 and 12).
The ee’s with these ligand systems were significantly
smaller (65% and 78%) and the chemical yields dramati-
cally diminished (30% and 35%). Finally, the steric bulk
of the thiol ether was investigated by substituting the
phenyl group for methyl (8a ) and 2,6-dimethylphenyl
(8b). It can be seen that the smaller substituent only
slightly reduced the ee (compare 6e with 8a ) and that
the larger group gave our highest ee (8b, 96%).
mechanism is fairly well understood and we could quickly
decipher the efficiency of our ligand system. Using
ligands 3a or 3b, no appreciable reaction was observed
after 5 days, so it was decided to modify 3a to contain
an N-methyl substituent (3c) in an attempt to make the
lone pair a stronger donor. In the test reaction, 3c gave
a poor yield (9%) of the product methyl-2-carbomethoxy-
3,5-diphenylpent-4-enoate (10) in essentially racemic
form after 5 days. The N,N-dimethyl derivative 3d , the
sulfur analogue of valphos,24 gave product (R)-10 in 93%
yield and 43% ee in only 1 day.
These disappointing results led us to speculate that
the lone pair of nitrogen needed to be as sterically
unencumbered as possible and “softer” for good coordina-
tion to the central metal atom of the reaction. It was
decided to transform the amine into an imine, 1. The
nitrogen atom of ligand 1 now resembled that present in
the very successful oxazoline-containing heterobidentate
ligands of Helmchen,8a Pfaltz,8b and Williams.8c
Discu ssion
The most influential part of this ligand system, to
ensure high enantiodiscrimination, appears to be the size
of the chiral backbone substituent. The transmission of
stereochemical information from this point to the reaction
sphere dominates the reaction and can be enhanced by
the correct choice of substituents on the mixed donor
atoms of the chelate system. There is a subtle balance
between sterics and electronics which effects the ef-
ficiency of this ligand system in this particular reaction.
We were interested in a transition-state model which
could explain the sense of enantioselection in this system.
In our original hypothesis, we considered all four
possible chiral ligand-Pd-π-allyl intermediates of this
reaction that could lead to the observed enantiomer of
the product (Figure 1). We assumed that the intermedi-
ates have a rapid mechanism of isomerization, and only
the syn,syn conformers of the allyl group were populated
to any significant extent. Although the importance of
syn,anti conformers has been recently highlighted in
other ligand systems,7b we have found no evidence of this
conformation in our system. Additionally, as the reaction
is exothermic, we assumed that the energy difference in
the ground state is reflected in the corresponding dia-
stereomeric transition states.25 Thus, as a starting point,
we may assume the major diastereoisomer of the inter-
mediate π-allyl complex leads to the major enantiomer
observed for the product.17 Based upon conformational
analysis and our original premise that the NdC π*
orbital may be a better π-acceptor than sulfur,19 we
arrived at transition state structure 2 (Figure 1), which
has the π-allyl group arranged in a “W” orientation of
We began our investigation by screening a series of
ligands which possessed different imine substituents 6
(Table 1). We found that it was unnecessary to preform
the Pd-L* complex and that simply stirring the ligand
and Pd dimer in dichloromethane for 15 min at room
temperature prior to the addition of the remaining
reagents was sufficient. Additionally, degassing the reac-
tion mixture in three freeze-evacuate-thaw cycles once
all the reagents had been added was found to increase
the yields, as did a nonaqueous workup. Each ligand
6a -i gave the (R)-isomer of 10, and overall the results
demonstrated that the choice of imine substituent did not
significantly affect the enantioselectivity. The ligand
system did not appear to be sensitive to electronic effects.
Steric bulk in the ortho position only had mild effects.
Substitution at both ortho positions reduced the ee (6d
and g), but mono o-chloro gave an enhanced ee (6e, 94%)
in good yield (87%). The corresponding bromo substituent
(6h ) gave an identical ee (94%) but a much reduced yield
(27%). Replacement of the aromatic imine with a less
sterically demanding methyl imine still resulted in ee’s
comparable to those of some aromatic substitutents
(20) Siedlecka, R.; Skarzewski, J . Synlett 1996, 757-8.
(21) Sakuraba, S.; Okada, T.; Morimoto, T.; Achiwa, K. Chem.
Pharm. Bull. 1995, 43, 927-34.
(22) Pine, S. H. J . Chem. Educ. 1968, 45, 118.
(23) Trost, B. M.; Murphy, D. J . Organometallics 1988, 4, 1143-5.
(24) Hayashi, T.; Konishi, M.; Fukushima, M.; Kanehira, K.; Hioki,
T.; Kumada, M. J . Org. Chem. 1983, 48, 2195-202.
(25) This is an interpretation of the Hammond postulate: Ham-
mond, G. S. J . Am. Chem. Soc. 1955, 77, 334-8.

Enantioselectivity in Pd-Catalyzed Allylic Substitution
J . Org. Chem., Vol. 64, No. 22, 1999 8259
F igu r e 3. (a) Bond lengths to Pd (pm). (b) Bond angles around
Pd. (c) Projection of the allyl ligand along the N-Pd-S plane.
the C(21) allylic terminus. As expected, there is a
significant difference in the bond length between N-Pd
(213.1 pm) and S-Pd (237.9 pm, Figure 3a). The allyl
ligand does not lie in an idealized orientation to the Pd-
N-S plane; it is rotated clockwise about the Pd-allyl axis
as shown (part structure) in Figure 3c. The idealized
geometry should possess two identical interplanar angles
defined by the Pd-N-S plane with the Pd-C(20)-C(21)
and Pd-C(19)-C(20) planes, respectively.10c Instead, we
note angles of 24.7° and 41.5°, respectively. This means
that the C(21)-C(20) (trans to nitrogen) is already
predisposed toward planarity with the Pd-N-S plane.
We elucidated the solution structure in CD2Cl2 (to
reflect the reaction medium) to help ascertain if this
structure was the same as the major intermediate in the
reaction mixture. We also conducted the identical cata-
lytic experiment (eq 2) but with 5 mol % of complex 14
as the source of chiral catalyst. This gave the substitution
product in a slightly lower yield of 73%, but with an
identical ee of 94%. This experiment verified that the
catalytically active species does not contain a chloride
substituent, that a 1:1 complex of Pd:chiral ligand gives
the same high ee and suggests that, although N,S-ligands
are, in genral, considered weakly coordinating when
compared to their phosphorus counterparts, this particu-
lar ligand system forms a catalytically active palladium
F igu r e 2. (a) Schematic representation of X-ray structure.
(b) Front view. (c) Left-hand-side view. (d) Right-hand-side
view.
the π-allyl group. Transition-state structure 2 is preferred
over 12, as the latter possesses a destabilizing 1,3-
interaction between the sulfide phenyl substituent and
the isopropyl group on the backbone of the chelate
ligand.17 Other reports of nitrogen-sulfur chiral chelate
ligands used in this reaction have assumed that the
sulfur atom is the better π-acceptor,10a,11d,13a,14 which
would mean transition-state models 11 and 13, having
the “M” orientation of the π-allyl unit, had to be consid-
ered. Initially, we could not categorically rule out 11 and
13, but they both contain destabilizing interactions with
respect to 2. Specifically, 11 has an unfavorable interac-
tion between the allyl phenyl group and the sulfide
phenyl group, while 13 possesses destabilizing interac-
tions similar to those described for 12. To provide firm
evidence for our transition-state model and to help us
understand the mechanism of chirality transfer for the
chiral ligand-Pd-π-allyl intermediate of this reaction,
we decided to obtain structural data in the solid state by
X-ray diffraction and in solution by NMR spectroscopy.
A high-quality crystal of the complex formed from the
trans-1,3-diphenylpropenyl palladium chloride dimer26
and ligand 6e in the presence of silver perchlorate was
obtained by slow evaporation from dichloromethane. The
crystal structure was found to be the π-allyl intermediate
14 drawn as part of transition-state structure 2 (Figure
2). The sulfur phenyl group is oriented trans to the
isopropyl group, thus avoiding an unfavorable 1,3-
interaction. The configuration of the imine is trans and
is accompanied by a twisting of the o-chlorophenyl ring
out of conjugation with the CdN bond by 40.9°. The plane
of the allyl ligand is tilted 15.5° from the perpendicular
to the Pd-N-S plane. In line with the structural
evidence for nitrogen-phosphorus ligands, we expected
that the two palladium-terminal allyl carbon bond dis-
tances would be different, with the bond trans to nitrogen
longer than the bond trans to sulfur.10,8f The actual bond
distances are 215.4 and 217.7 pm, respectively (Figure
3a), almost equal within experimental error (2.3 pm, ∼1%
difference). However, the N-Pd-C bond angle of 107.5°
is larger than the S-Pd-C bond angle of 101.8° (Figure
3b), suggesting some interaction between the ligand and
1
chelate complex. The H NMR spectrum of the complex
at the same temperature of the reaction (25 °C) showed
essentially one structure present (>95%), and its confor-
mation was assigned through a combination of COSY,
13C-1H correlation, and selected one-dimensional NOE
experiments. The 1H NMR showed only the syn,syn
arrangement of the allyl unit, as evidenced by the
coupling constants between the allyl protons H(19),
H(20), and H(21) of around 12 Hz. Key NOE enhance-
ments (Figure 4) suggest the complex is in the same “W”
conformation in solution as in solid-state 14 (Figure 2).
Irradiation of the signal at δ 8.39 (1H, dd, J ) 7.6, 1.8
Hz) induced an enhancement of 6.8% at δ 7.83 (1H, td, J
) 7.8, 1.6 Hz) and 5.1% at δ 4.54 (1H, d, J ) 11.3 Hz).
The former lower field signals are single proton reso-
nances, coupled to each other as evidenced by the COSY
spectrum, and originate from the unsymmetrical aro-
matic ring C of the benzylidene. Taking into account the
position and electronic effects of the substituents, the
signals at δ 8.39 and δ 7.83 can be assigned to H(12) and
H(11), respectively. By virtue of its multiplicity, the
magnitude of the coupling constant, and its NOE en-
hancement upon irradiation of H(12), the single proton
(26) Hayashi, T. Yamamoto, A.; Ito, Y.; Nishioka, E.; Miura, H.;
Yanagi, K. J . Am. Chem. Soc. 1989, 111, 6301-11.

8260 J . Org. Chem., Vol. 64, No. 22, 1999
Adams et al.
F igu r e 4. View of X-ray structure with selected protons to
show key NOE enhancements.
F igu r e 5. Front view of two possible Pd0(olefin) complexes
resulting from nucleophilic attack of the most stable Pd+(allyl)
intermediate.
resonance at δ 4.54 can be assigned to H(21). This
indicates that H(12) and H(21) are on the same face of
the palladium complex. Irradiation of H(21) gave a
corresponding enhancement of 5.0% at H(12), 10.8% at
the other allylic hydrogen H(19), and 6.3% at δ 6.77 (2H,
dd, J ) 8.2, 0.9 Hz). This latter signal (δ 6.77) we assign
to the ortho protons of aromatic ring A, H(23) and H(27).27
Irradiation of the isopropyl proton H(3) at δ 2.03-2.17
(1H, m) induced an enhancement of 1.7% at H(27).
Irradiation of one of the methyl signal at δ 0.98 (3H, d,
J ) 6.7) induced an enhancement at H(6) δ 8.73 (1H, s),
while irradiation of the other methyl signal at δ 0.84 (3H,
d, J ) 6.4 Hz) gave an enhancement of 7.4% at H(1â) δ
3.71 (1H, dd, J ) 12.5, 1.1 Hz). The methyl signal at δ
0.98 is assigned to the C(5) methyl, and irradiation of
this signal also gave a 3.2% enhancement of H(27). These
NOE contacts indicate that the isopropyl group and
aromatic ring A were on the same face of the complex in
solution.28 From the 13C-1H correlation spectrum, the
chemical shifts for C(19) and C(21) (83.0 and 86.0 ppm,
respectively) indicate little difference between the elec-
tronic environments of these atoms in solution, which is
in agreement with the crystal structure and much
smaller than the difference observed for oxazoline-
thioglucose ligands.15
Mech a n istic Im p lica tion s. The NMR experiments
verified that the intermediate in a solution of CD2Cl2 at
room temperature had the same orientation of the allyl
unit as in the solid state. This structure was the only
one clearly visible (g95%) by NMR. and as all our
reactions were performed in CH2Cl2 at room temperature,
we believe the X-ray structure 14 depicts the palladium
intermediate from which the major (R)-enantiomer of 10
is formed. This requires the nucleophile to attack the allyl
group trans to the imine function at C(19). This occurs
despite the possible electronic preference for attack trans
to sulfur, as suggested by the slightly longer Pd-C(21)
bond and lower field 13C chemical shift. To account for
the preferred trajectory, we believe that the steric influ-
ence of the chiral ligand predisposes the allyl group
toward nucleophilic attack at C(19). This is manifested
in the X-ray structure, where the C(20)-C(21) bond is
tending toward coplanarity with the N-Pd-S plane. This
type of distortion has been noted10c in a crystal structure
of a nitrogen-phosphorus chiral ligand and is evident
in others8f,10a but has always served to reinforce the
electronic preference set up by the trans effect of the
heterodentate ligand.29 In our case, this distortion is of
paramount importance, as reaction of the nucleophile
with the (η3-allyl)Pd complex to form the primary (η2-
olefin)Pd0 complex must be accompanied by rotation. This
concept was first advanced by Trost6a for ionization and
then for nucleophilic attack by the microscopic reverse.
The involvement of preferential rotation7 to help explain
enantioselection in asymmetric palladium-catalyzed al-
lylic substituion reactions has been promoted by Brown8f
and others.7,10c,15 In our case, the primary product result-
ing from nucleophilic attack at C(19) gives the thermo-
dynamically more stable (η2-olefin)Pd0 complex 15 (Fig-
ure 5). Formation of complex 16 from nucleophilic attack
at C(21) would involve a more strenuous rearrangement
in the transition state, causing unfavorable steric inter-
actions between the aromatic A ring and the benzylidene
aromatic ring C. We believe that this steric interaction,
as described by the NOE contacts, is dominant. The
desired and lowest energy pathway (2, Figure 1) obeys
the principle of least motion30 and gives the thermody-
namically most stable primary product 15, which upon
decomplexation gives the observed major enantiomer of
the product (R)-10.
Con clu sion
We can conclude that the catalyst provides a unique
chiral environment which orientates the allyl fragment
into a predisposed conformation around the Pd-allyl axis
and dictates reaction at one end of the allyl fragment. In
our particular system, sterics account for the nucleophile
approaching trans to the imine donor. This is contrary
to expectation from previously studied nitrogen-sulfur
ligands. This work serves to illustrate the principle that
the correct choice of substituents can reinforce the
stereoinducing power of the chirality inherent to a ligand
system. It would appear as so much emphasis has been
placed on the power of electronic desymmetrization for
chiral recognition that, in the absence of physical data
(27) For other NOE enhancements, see table in Supporting Informa-
tion.
(28) The existence of this NOE contact suggests that, when the steric
bulk of the backbone substituent was increased to Ph (7a ) or t-Bu (7b),
which resulted in decreased yield and enantioselectivity, the formation
of π-allyl intermediates similar to 14 is less favourable.
(29) For theoretical treatments, see: (a) Ward, T. R. Organometallics
1996, 15, 2836-8. (b) Blo¨chl, P. E.; Togni, A. Organometallics 1996,
15, 4125-32.
(30) For a review, see: Hine, J . Adv. Phys. Org. Chem. 1977, 15,
1-61.

Enantioselectivity in Pd-Catalyzed Allylic Substitution
J . Org. Chem., Vol. 64, No. 22, 1999 8261
for most other nitrogen-sulfur chelate ligand systems,31
it has been assumed that the trans effect dominates
enantioselection. We propose that the trans effect in
many of these systems is minimal and that enantiose-
lection arises through the predisposition of the allyl
fragment through the asymmetric steric influence of the
ligand system. The explanation for enantioselectivity
which we have structurally detailed is similar to expla-
nations promoted by Brown8f and others7,10c,15 to account
for the enantioselection of other ligand systems. These
effects may have important implications in the design
criteria of similar N,S-ligand systems, where attention
needs to be paid to both steric and electronic interactions.
mmol) in CH2Cl2 (1 mL) was added dropwise, followed by
potassium acetate (2.5 mg, 0.03 mmol). The reaction mixture
was degassed in three freeze-evacuate-thaw cycles and
stirred at room temperature for 24-96 h. Evaporation of the
volatile components and purification by column chromatogra-
phy (8% ethyl acetate/light petroleum) gave 10 as a clear oil.11d
{[(S)-N-2′-Ch lor oben zylid en e-2-a m in o-3-m eth yl-1-th io-
ph en ylbu tan e]-(η3-tr a n s-1,3-diph en ylallyl)palladiu m P er -
ch lor a te. To a stirred solution of 1,3-diphenylpalladium
chloride dimer26 (460 mg, 0.7 mmol) in MeOH/CH2Cl2 (1:1, 10
mL) was added AgClO4 (285 mg, 1.4 mmol) at room temper-
ature in the absence of light. After 45 min, the reaction mixture
was filtered through a 2-cm plug of Celite and washed with
MeOH/CH2Cl2 (1:1, 5 mL). A solution of ligand 6e (436 mg,
1.4 mmol) in 50% methanol/dichloromethane (5 mL) was added
to the clear yellow filtrate, which was stirred for a further 45
min. Concentration in vacuo gave an orange powder, which
was recrystallized from CH2Cl2/hexane to give the title com-
Exp er im en ta l Section
pound as orange crystals (927 mg, 94%): mp > 250 °C; [R]21
Gen er a l Meth od s. Our general experimental details have
been reported elsewhere.32 Syntheses of compounds 3c,d , 7a ,b,
and 8a ,b and spectroscopic data for 6a -i are contained in the
Supporting Information.
D
) +25.9° (c 0.19, CH2Cl2); IR (KBr) 2966, 1627 cm-1; 1H NMR
δ 0.84 (3H, d, J ) 6.4 Hz, C4H3), 0.98 (3H, d, J ) 6.7 Hz, C5H3),
2.03-2.17 (1H, m, C3H), 3.05 (1H, dd, J ) 12.8, 4.3 Hz,
C1HRHâ), 3.48-3.57 (1H, m, C2H), 3.71 (1H, dd, J ) 12.5, 1.1
Hz, C1HRHâ), 4.54 (1H, d, J ) 11.3 Hz, C21H), 5.18 (1H, d, J
) 12.2 Hz, C19H), 6.48 (1H, t, J ) 12.1 Hz, C20H), 6.77 (2H,
dd, J ) 8.2, 0.9 Hz, C23H + C27H), 7.09 (2H, t, J ) 7.8 Hz,
C24H + C26H), 7.18 (2H, t, J ) 7.5 Hz, C30H + C32H), 7.25-
7.48 (9H, m, ArH), 7.35 (1H, dd, J ) 7.0, 1.5 Hz, C9H), 7.72
(1H, td, J ) 7.8, 1.6 Hz, C10H), 7.83 (1H, dd, J ) 7.6, 0.9 Hz,
C11H), 8.39 (1H, dd, J ) 7.6, 1.8 Hz, C12H), 8.73 (1H, s, C6H);
13C NMR δ 18.7 (C4), 20.0 (C5), 32.1 (C3), 44.2 (C1), 81.0 (C2),
83.0 (C19), 86.0 (C21), 107.2 (C20), 127.2 (C23 + C27), 127.8 (Ar),
128.3 (C11), 128.7 (Ar), 129.2 (C24 + C26), 129.4 (Ar), 129.6 (C30
+ C32), 130.1 (Ar), 130.5 (Ar), 130.9 (Ar), 131.0 (C12), 133.1 (Ar),
133.5 (Ar), 134.9 (C10), 136.1 (Ar), 136.5 (Ar), 136.7 (Ar), 168.3
(C6); MS (FAB+) 617 (MH+). Anal. Calcd for C33H33NO4S2Cl2-
Pd: C, 55.28; H, 4.64; N, 1.95; S, 4.47; Cl, 9.89. Found: C,
55.56; H, 4.42; N, 1.85; S, 4.52; Cl, 9.96.
(S)-2-Am in o-3-m eth yl-1-th iop h en ylbu ta n e (4). Thiophe-
nol (1.67 mL, 16.3 mmol, 1.05 equiv) was added dropwise to
prewashed (hexane, 3 × 10 mL) NaH (0.65 g, 80% dispersion
in mineral oil, 16.3 mmol, 1.05 equiv) in DMF (30 mL) at 0
°C. After 15 min, a solution of (S)-N-tert-butoxycarbonyl-2-
amino-3-methyl-1-butyl-p-toluenesulfonate21 (5.53 g, 15.5 mmol)
in DMF (20 mL) was added, the mixture was stirred for 1 h
at 0 °C and 2 h at room temperature, and then aqueous NaOH
(1 M, 200 mL) was added. The mixture was extracted with
Et2O, and the combined organic layers were washed with
water and brine, dried (MgSO4), and concentrated in vacuo.
Purification by column chromatography (15% ethyl acetate/
light petroleum) gave the protected amine (S)-N-tert-butoxy-
carbonyl-2-amino-3-methyl-1-thiophenylbutane as
a white
solid (1.17 g, 71%): mp 74-75 °C; [R]21D ) +31.0° (c 1.00, CH2-
Cl2); IR (thin film) 3441, 2968, 2933, 2875, 1708, 1584, 1500,
1368, 1167 cm-1; 1H NMR δ 0.90 (6H, t, J ) 6.4 Hz), 1.41 (9H,
s), 1.85-1.98 (1H, m), 3.07 (2H, d, J ) 5.5 Hz), 3.62-3.68 (1H,
m), 4.56 (1H, d, J ) 8.2 Hz), 7.14-7.39 (5H, m); 13C NMR δ
17.9, 19.5, 28.4, 30.8, 37.6, 55.2, 83.9, 129.0, 129.7, 133.2, 155.4;
MS (EI+) 295 (M+). Anal. Calcd for C16H25NO2S: C, 65.05; H,
8.53; N, 4.74; S, 10.83. Found: C, 65.12; H, 8.76; N, 4.74; S,
11.05.
X-r a y Cr yst a llogr a p h ic St r u ct u r e Det er m in a t ion of
14. Crystallographic and data collection parameters are
provided as Supporting Information. Crystal data for C33H33
-
Cl2NO4PdS: M ) 716.96; crystallizes from dichloromethane
as orange blocks, crystal dimensions 0.66 × 0.34 × 0.34 mm;
monoclinic, a ) 9.347(4), b ) 18.920(8), and c ) 9.648(5) Å, â
) 108.71(2)°, U ) 1616.0(13) Å3, Z ) 2, Dc ) 1.473 Mg/m3,
space group P21 (C2 No.4), Mo KR radiation (λh ) 0.710 73 Å),
µ(Mo KR) ) 0.841 2mm-1, F(000) ) 732. Three-dimensional,
room-temperature X-ray data were collected in the range 3.5
< 2θ < 50° on a Siemens P4 diffractometer by the ω scan
method. Of the 3648 reflections measured, all of which were
corrected for Lorentz and polarization effects and for absorp-
tion by semiempirical methods based on symmetry-equivalent
and repeated reflections (minimum and maximum transmis-
sion coefficients 0.6067 and 0.7630), 2540 independent reflec-
tions exceeded the significance level |F|/σ(|F|) > 4.0. The
structure was solved by direct methods and refined by full-
matrix least-squares on F2. Hydrogen atoms were included in
calculated positions and refined in riding mode. Refinement
converged at a final R ) 0.0510 (wR2 ) 0.1842, for all 3087
data, 398 parameters, mean and maximum δ/σ 0.000, 0.000)
with allowance for the thermal anisotropy of all non-hydrogen
atoms. Minimum and maximum final electron density were
-1.790 and 0.828 e Å-3. A weighting scheme w ) 1/[σ2(Fo2) +
The protected amine (1.30 g, 4.40 mmol) was stirred with
trifluoroacetic acid (4.41 mL, 57.3 mmol, 13 equiv) and Et3-
SiH (1.76 mL, 11.0 mmol, 2.5 equiv) in CH2Cl2 (15 mL) for 1
h at room temperature. Aqueous NaOH (1 M, 100 mL) was
added, and the mixture was extracted with CH2Cl2. The
combined organic layers were dried (MgSO4) and concentrated
in vacuo to give 4 as a light brown oil (0.86 g, 100%): [R]21
)
D
+79.8° (c 0.99, CH2Cl2); IR (thin film) 3372, 3059, 2958, 1584,
1480, 1367 cm-1 1H NMR δ 0.91 (3H, d, J ) 5.2 Hz), 0.94
;
(3H, d, J ) 5.2 Hz), 1.38 (2H, br s), 1.65-1.78 (1H, m), 2.66-
2.76 (2H, m), 3.12-3.22 (1H, m), 7.14-7.37 (5H, m); 13C NMR
δ 17.6, 19.3, 33.0, 41.0, 55.4, 126.1, 128.9, 129.4, 136.3; MS
(EI+) 195 (M+). Anal. Calcd for C11H17NS: C, 67.58; H, 8.86;
N, 7.03; S, 16.65. Found: C, 67.58; H, 8.86; N, 7.03; S, 16.65.
Gen er al P r ocedu r e for th e For m ation of Im in e Ligan ds
6a -h . An equimolar mixture of 4 and the requisite aromatic
aldehyde with MgSO4 (0.54 g/mmol) was stirred in CH2Cl2 (2.5
mL/mmol) for 24 h. Filtration and concentration in vacuo gave
ligands 6a -i as oils which were unstable to chromatography
(0.0648P)2 + 0.6799P], where P ) (Fo + 2Fc2)/3 was used in
2
1
but judged to be >95% pure by H NMR.
the latter stages of refinement. Complex scattering factors
were taken from the program package SHELXL9733 as imple-
mented on the Viglen Pentium computer.
Gen er a l P r oced u r e for P a lla d iu m -Ca t a lyzed Allylic
Su bstitu tion . A solution of the acetate 9 (252 mg, 1.0 mmol),
allylpalladium chloride dimer (9.1 mg, 0.025 mmol), and ligand
(0.1 mmol) in CH2Cl2 (2 mL) was stirred at room temperature
for 15 min before a solution of dimethyl malonate (396 mg,
3.0 mmol) and N,O-bis(trimethylsilyl)acetamide (0.74 mL, 3.0
Ack n ow led gm en t. This work is part of the Ph.D.
Thesis of D.S.J ., University of Sheffield, 1998. We thank
(31) See ref 15 for NMR mechanistic study.
(32) Anderson, J . C.; Siddons, D. C.; Smith, S. C.; Swarbrick, M. E.
J . Org. Chem. 1996, 61, 4820-3.
(33) Sheldrick, G. M. SHELXL97, An integrated system for solving
and refining crystal structures from diffraction data (Revision 5.1);
University of Gottingen, Germany, 1993.

8262 J . Org. Chem., Vol. 64, No. 22, 1999
Adams et al.
and refinement, atomic coordinates, bond lengths and angles,
anisotropic thermal parameters; table of NOE data and
pertinent NOE, COSY, and 13C-1H correlation spectra for 14;
and 250-MHz 1H NMR spectra of compounds 3c,d , 6a -i, 7a ,b,
8a ,b, and 14. An X-ray crystallographic file, in CIF format,
for 14 is also available. This material is available free of charge
via the Internet at http://pubs.acs.org.
the University of Sheffield and Pfizer for financial
support, Dr. B. F. Taylor and Ms. S. Bradshaw for
running NMR experiments, Mr. A. J ones and Ms. J .
Stanbra for measuring microanalytical data, and Mr.
S. Thorpe and Mr. N. Lewus for providing mass spectra.
Su p p or tin g In for m a tion Ava ila ble: Syntheses of com-
pounds 3c,d , 7a ,b, and 8a ,b and spectroscopic data for 6a -i;
an ORTEP plot for 14; tables of crystal data, structure solution
J O9910177