Asymmetric molybdenum(0)-catalyzed allylic substitution

Article abstract of DOI:10.1016/S0040-4039(00)02007-4

Application of new ligands (R)-(-)-8,(S)-(+)-16, and (S)-(+)-17 to the title reaction (1 or 2 → 3) led to excellent regio- and enantioselectivities (> 30:1; ≤ 98% ee); although lacking the C₂-symmetry, the catalysts can be viewed as quasi-C₂-symmetrical since the single chiral center is sufficient to determine the sense of wrapping of the metal by the ligand.

Full text of DOI:10.1016/S0040-4039(00)02007-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 509–512
Asymmetric molybdenum(0)-catalyzed allylic substitution
Andrei V. Malkov,a,b,† Paul Spoor,b Victoria Vinaderc and Pavel Kocˇovsky´a,b,
*
aDepartment of Chemistry, University of Glasgow, Glasgow G12 8QQ, UK
bDepartment of Chemistry, University of Leicester, Leicester LE1 7RH, UK
cMedicines Research Centre, GlaxoWellcome, Research and Development, Stevenage, Herts. SG1 2NY, UK
Received 18 September 2000; revised 25 October 2000; accepted 8 November 2000
Abstract—Application of new ligands (R)-(−)-8, (S)-(+)-16, and (S)-(+)-17 to the title reaction (1 or 2“3) led to excellent regio-
and enantioselectivities (>30:1; 598% ee); although lacking the C2-symmetry, the catalysts can be viewed as quasi-C2-symmetrical
since the single chiral center is sufficient to determine the sense of wrapping of the metal by the ligand. © 2001 Published by
Elsevier Science Ltd.
Recently, Trost has reported on the first examples of
high asymmetric induction in Mo(0)-catalyzed allylic
substitution employing cinnamyl carbonates 1a and 2a
and their aromatic, heteroaromatic, and other conju-
gated analogues as model substrates (Scheme 1):1 with
malonate-type nucleophiles and bis-picolinic amide 72
as the chiral ligand (Fig. 1), both the regio- and enan-
tioselectivities attained were excellent (in favor of 3;
Table 1, entries 1 and 2).1 A similar strategy was
adopted by Pfaltz,3 who designed analogous ligands
with oxazoline units in place of Trost’s picolinic amide
moieties.4,5,‡ Application of microwave heating has now
been reported by Moberg to further improve the reac-
tivity of the Mo/7 catalyst.6
into bis-amide (R)-(−)-8 [a-picolinic acid, (PhO)3P,
pyridine, 60%].§
While Trost,1 Pfaltz,3 and Moberg6 have employed
C2-symmetrical ligands with trans-1,2-diaminocyclohex-
ane as the chiral scaffold, we felt that one chiral center
might be sufficient to determine the sense of wrapping
of the metal by the ligand, thereby creating a similar
chiral environment. In order to verify this hypothesis,
we focused on ligand (R)-(−)-8 that was readily pre-
pared (Scheme 2) via conversion of (R)-(−)-methyl
phenylglycinate into amide (R)-(−)-12 (aq. NH3, tolu-
ene, 70%),7 followed by reduction (LiAlH4, THF,
45%),8 and transformation of the resulting diamine 13
Scheme 1. a, R=Ph; b, 2-thienyl; c, 1-cyclohexenyl.
* Corresponding author. Present address: University of Glasgow.
Fax: +44-(0)141-3304888; e-mail: p.kocovsky@chem.gla.ac.uk
Present address: University of Glasgow.
For an earlier observation by us of an asymmetric induction with
†
‡
Figure 1.
Mo–bis-oxazoline complexes, see Ref. 4. Since this experiment
required a high catalyst loading, it was only cited in a footnote.
§ For the method, see Ref. 2b.
0040-4039/01/$ - see front matter © 2001 Published by Elsevier Science Ltd.
PII: S0040-4039(00)02007-4

510
A. V. Malko6 et al. / Tetrahedron Letters 42 (2001) 509–512
Cinnamyl-type carbonates 1a and 2a, in conjunction
with malonate nucleophiles, were employed to probe
the efficiency of ligand 8 (Scheme 1, Table 1). The
9
catalyst was generated in situ from (EtCN)3Mo(CO)3
or (C7H8)Mo(CO)3 and (R)-(−)-8 in THF;ꢀꢀ upon
10
addition of the ligand, the solution turned deep-red
(instantaneously with the former and within 5 min with
the latter complex). The reactions with NaCH(CO2-
Me)2, carried out in THF at 60°C, proved to be fairly
regio- and enantioselective in favor of the branched
product 3a (]8:1, ꢀ90% ee; entries 3, 4, and 6) with
good yields.¶ Little difference was observed between the
regioisomeric substrates 1a and 2a (Table 1, entries 3, 4,
and 6) and identical results were obtained with the
catalyst generated from (EtCN)3Mo(CO)3 and
(C7H8)Mo(CO)3 (compare entries 3 and 4).¶ With
NaCMe(CO2Me)2, the reaction proved to be much
slower and the selectivity lower (entry 5).**
Scheme 2. Py=a-pyridyl.
Thus, ligand 9 (a positional isomer of 8) failed to bring
about the reaction, while 11 (an ester/amide) was non-
selective (1:1 ratio of 3a:5a) and gave low conversion
rate (26%). With 10 (lacking one pyridine nitrogen
atom), the enantioselectivity was high (entry 7); how-
ever, the low conversion in this instance suggests an
almost stoichiometric process that can occur with a
different mechanism4,11 and/or mode of coordination.
Hence, these experiments have demonstrated that (1)
the original structural characteristics of the Trost–
Moberg ligand 71,6 and its Pfaltz analogue,3 namely the
two sp2-type nitrogen donors and two rigid amide
groups, are essential, and (2) one chiral center in the
scaffold is sufficient to induce high levels of enantiose-
lectivity. Noteworthy is the enhanced reactivity of these
Mo catalysts, as compared to the previously studied
bipyridine and phenanthroline complexes.12
Ligand 8, as well as its predecessor 7,1,3,6 can a priori
offer up to four ligating atoms, namely the pyridine
nitrogens and either the amidic carbonyls or nitrogen
atoms. To investigate the role of the individual struc-
tural features, we have synthesized ligands 9–11 (Fig.
1), all of which then turned out to be inferior to 8.
Although the mode of coordination of 7 and 8 to Mo is
unknown at present,†† it can be argued that the Ph
group in the quasi-C2-symmetrical ligand 8 acts as an
anchor, presumably occupying an ‘equatorial’ position
in the cyclic complex, thereby mimicking the rigid
scaffold of 7. We reasoned that the ligand performance
may be improved by implementing a bulkier anchor R
that would ensure more rigidity of the whole frame-
work. Therefore, we have synthesized ligands (S)-(+)-
16 (R=PhCH2) and (S)-(+)-17 (R=i-Pr) from
(S)-(+)-phenylalanine and (S)-(+)-valine amides, re-
spectively, in a similar fashion as in the case of (R)-(−)-
8. The benzyl derivative (S)-(+)-16 (note the lower A
value for PhCH2 than for Ph)13 turned out to exhibit
somewhat lower enantioselectivity (74–89% ee) than
the parent phenyl derivative (R)-(−)-8 (compare entries
3–6 with 8–10). By contrast, the isopropyl ligand (S)-
(+)-17 (higher A value for i-Pr) gave much improved
results that are in the same range as those reported by
Trost (98% ee, 32:1 regioselectivity; compare entries 11
and 12 with 1 and 2).
ꢀꢀ We favor (C7H8)Mo(CO)3 as it is more air-stable than
(EtCN)3Mo(CO)3 and, according to our experience, somewhat eas-
ier to prepare in a pure and defined form. Thus, while the prepara-
tion of the former complex is straightforward (reflux for 8 h,
followed by Soxhlet extraction),10a the latter complex is usually
contaminated by (EtCN)2Mo(CO)4 so that prolonged reflux (up to
3 days) and repeated crystallization is often required to obtain the
pure species. The complex generated in situ from the latter contam-
inant and ligand 8 is practically inert in the catalytic reaction, as
revealed by control experiments. This behavior seems to suggest that
a tridentate coordination of the metal by 8 is required to generate an
active catalyst. Our results, cited in Table 1, were obtained with pure
(EtCN)3Mo(CO)3.
¶ Typical experiment: A mixture of (EtCN)3Mo(CO)3(EtCN)3 (34 mg,
0.1 mmol) and a ligand (0.15 mmol) was dissolved in THF (3 mL).
The solution, which instantaneously turned deep red, was heated
with stirring at 60°C for 40 min. The solution was cooled to room
temperature and then a solution of the corresponding sodioma-
lonate (2.0 mmol) in THF (2 mL), generated from dimethyl mal-
onate (or dimethyl methylmalonate) and NaH, and a solution of
allylic carbonate (1.0–1.3 mmol) in THF (1 mL) were successively
added. Usually, the addition of the reactants was accompanied by a
change of color to orange or yellow-brown. The mixture was stirred
at 60°C until the reaction was complete (as evidenced by TLC), then
diluted with ether (20 mL), and washed successively with 5%
aqueous NaHCO3 and water. The organic phase was dried with
MgSO4 and the solvent was evaporated under reduced pressure. The
crude product was purified by flash chromatography on silica gel
(15×2 cm) with a 9:1 hexane–ethyl acetate mixture as an eluent.
Enantiomeric purity of the products 3a and 4a, respectively, was
determined by chiral HPLC using Chiralcel OD-H (3a) or Chiralpak
AD (4a) columns (equipped with a guarding silica gel column) and
a mixture of hexane and 2-propanol (99.5:0.5) as eluent; UV detec-
tion at 220 nm. For compound 3a, retention times were as follows:
(S)-enantiomer 17.4 min, (R)-enantiomer 18.7 min. For compound
4a, retention times were: (R)-isomer 6.4 min, (S)-isomer 10.7 min.
** Note the pseudo-inversion of configuration of the product due to
the change in the substituent priorities.
†† Trost has hypothesized on a bidentate mode in the complex of 7
with trans-configuration of the ligating nitrogens about the metal
center.1

A. V. Malko6 et al. / Tetrahedron Letters 42 (2001) 509–512
Table 1. Mo(0)-catalyzed allylic substitution (Scheme 1)a
511
Entry
Substrate
R%
Ligand
Time (h)
Ratio 3/4:5/6b
Yield (%)
ee of 3/4 (%)c
1
2
3
4
5
6
7
8
1a
2a
1a
1a
1a
2a
1a
1a
1a
2a
1a
2a
1b
1b
2b
1c
1c
1c
H
H
H
H
Me
H
H
H
Me
H
H
H
H
H
H
H
H
H
7d
7d
3
3
4
4
12
4
24
4
10
4
12
12
5
32:1
13:1
8:1
8:1
4:1
88
70
63
65
37
72
21
69
32
68
68
59
68
87
78
59
61
91
99 (S)f
92 (S)f
92 (S)
8d
8e
92 (S)
8d
73 (R)g
88 (S)
8d
12:1
12:1
13:1
10:1
13:1
32:1
38:1
10:1
18:1
19:1
5.5:1
6.2:1
11.5:1
10d
16d
16d
16d
17d
17d
8e
78 (S)
89 (R)h
57 (S)g,h
74 (R)h
98 (R)h
97 (R)h
90 (S)
9
10
11
12
13
14
15
16
17
18
17e
7d
5
2
24
24
3i
92 (R)h
88 (S)f
86 (S)
8e
17e
7d
87 (R)h
94 (S)j
a Conditions: THF, 60°C, cat. 7–10 mol%.
b Determined from the 1H NMR spectra of the product mixture.
c Determined by chiral HPLC.
d The catalyst was generated from (EtCN)3Mo(CO)3.
e The catalyst was generated from (C7H8)Mo(CO)3.
f Ref. 1a.
g The pseudo-inversion of configuration of the product is due to the change in the substituent priorities.
h Note that the ligand has the opposite absolute configuration to 8.
i In a toluene–THF mixture at 90°C.
j Ref. 1b.
Similarly high selectivities were observed for the 2-
thienyl and 1-cyclohexenyl substrates 1b and 1c (entries
13, 14, 16, and 17), demonstrating that the success of
our new ligands (especially 8 and 17) is not confined to
one substrate. Note that our results are comparable
with those reported by Trost1 (entries 15 and 18).
References
1. (a) Trost, B. M.; Hachiya, I. J. Am. Chem. Soc. 1998,
120, 1104. (b) Trost, B. M.; Hildbrand, S.; Dogra, K. J.
Am. Chem. Soc. 1999, 121, 10416.
2. For the preparation and other uses of ligand 7, see: (a)
Mulqi, M.; Stephens, F. S.; Vagg, R. S. Inorg. Chim. Acta
1981, 53, L91. (b) Adolfsson, H.; Moberg, C. Tetra-
hedron: Asymmetry 1995, 6, 2023. (c) Moberg, C.; Adolf-
sson, H.; Wa¨rnmark, K. Acta Chim. Scand. 1996, 50, 195
and references cited therein.
3. Glorius, F.; Pfaltz, A. Org. Lett. 1999, 1, 141.
4. Dvorˇa´k, D.; Stary´, I.; Kocˇovsky´, P. J. Am. Chem. Soc.
1995, 117, 6130.
5. For the first successful asymmetric induction in W(0)-cat-
alyzed allylic substitution and high level of enantioselec-
tion attained with a W(0)-phosphinooxazoline catalyst,
see: Lloyd-Jones, G. C.; Pfaltz, A. Angew. Chem., Int. Ed.
Engl. 1995, 34, 462.
In conclusion, we have synthesized a set of ligands 8,
16, and 17 to develop asymmetric, molybdenum(0)-cat-
alyzed allylic substitution. Optimization of the ligands
along the working hypothesis resulted in identifying the
isopropyl-substituted ligand 17 as that with highest
enantio- and regioselectivity in most cases (e.g. entry
11). For this champion ligand we propose the acronym
‘VALDY’ (valine+dipyridine). This study has demon-
strated that, in this instance, one chiral center in the
scaffold is sufficient to determine the sense of the chiral
environment at the metal, which may have further inter-
esting implications in asymmetric catalysis.
6. Belda, O.; Kaiser, N.-F.; Bremberg, U.; Larhed, M.;
Hallberg, A.; Moberg, C. J. Org. Chem. 2000, 65, 5868.
7. Ager, D. J.; Prakash, I. Synth. Commun. 1996, 26, 3865.
8. Belokon, Yu. N.; Pritula, L. K.; Tararov, V. I.; Bakhmu-
tov, V. I.; Struchkov, Yu. T. J. Chem. Soc., Dalton Trans.
1990, 1867.
9. Kubas, G. J.; Van der Sluys, L. S. Inorg. Synth. 1990, 28,
29.
10. C7H8=cycloheptatriene. (a) For the complex prepara-
tion, see: Cotton, F. A.; McCleverty, J.; White, J. E.
Acknowledgements
We thank Dr. Guy C. Lloyd-Jones and Professor
Christina Moberg for fruitful discussions and valuable
suggestions. We also thank the EPSRC for grants No.
GR/H 92067 and GR/K07140 and EPSRC and Glax-
oWellcome for a CASE award to P.S.

512
A. V. Malko6 et al. / Tetrahedron Letters 42 (2001) 509–512
Inorg. Synth. 1990, 28, 45. (b) For the numerous benefits
12. For selected examples of the application of a,a-bipyridine
and phenanthroline-type ligands in Mo(0)- and W(0)-cat-
alyzed allylic substitution, see the following: (a) Frisell,
of the analogous C7H8W(CO)3 complex as the catalyst
for allylic substitution, see: Lehmann, J.; Lloyd-Jones, G.
C. Tetrahedron 1995, 51, 8863.
,
H.; Akermark, B. Organometallics 1995, 14, 561. (b)
,
11. While catalytic reactions can occur via a ret–ret mecha-
nism,4 their stoichiometric counterparts proceed via a
ret–inv mechanism: (a) Faller, J. W.; Linebarrier, D.
Organometallics 1988, 7, 1670. (b) Ward, Y. D.; Vil-
lanueva, L. A.; Allred, G. D.; Liebeskind, L. S. J. Am.
Chem. Soc. 1996, 118, 897.
Eriksson, L.; Sjo¨gren, M. P. T.; Akermark, B. Acta
Crystallogr., Sect. C 1996, 52, 77. (c) Sjo¨gren, M. P. T.;
,
Frisell, H.; Akemark, B.; Norrby, P.-O.; Eriksson, L.;
Vitagliano, A. Organometallics 1997, 16, 942.
13. Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry
of Organic Compounds; J. Wiley: New York, 1994; p. 697.
.
.