Regioselective lithiation of silyl phosphine sulfides: asymmetric synthesis of p-stereogenic compounds

Article abstract of DOI:10.1021/ol901977p

From a trimethylsilyl-substituted phosphine sulfide of 99:1 er (generated by n-BuLi/(-)-sparteine-mediated asymmetric lithiation of a dimethylphosphine sulfide), a two-step process of regloselective lithiation-trapping and silyl group removal has been use

Full text of DOI:10.1021/ol901977p

ORGANIC
LETTERS
2009
Vol. 11, No. 21
5022-5025
Regioselective Lithiation of Silyl
Phosphine Sulfides: Asymmetric
Synthesis of P-Stereogenic Compounds
Jonathan J. Gammon,† Peter O’Brien,*,† and Brian Kelly‡
Department of Chemistry, UniVersity of York, Heslington, York YO10 5DD, U. K., and
Celtic Catalysts Ltd., NoVa Centre, Belfield InnoVation Park, Dublin 4, Ireland
paob1@york.ac.uk
Received August 26, 2009
ABSTRACT
Starting from a trimethylsilyl-substituted phosphine sulfide of 99:1 er (generated by n-BuLi/(-)-sparteine-mediated asymmetric lithiation of a
dimethylphosphine sulfide), a two-step process of regioselective lithiation-trapping and silyl group removal has been used to prepare a range
of P-stereogenic compounds, including precursors to diphosphine ligands (e.g., Mini-PHOS). This two-step protocol delivers products with the
opposite configuration to that obtained by direct asymmetric lithiation-trapping of a dimethylphosphine sulfide using n-BuLi/(-)-sparteine.
P-Stereogenic phosphines constitute one class of chiral
phosphines that have been utilized as ligands in a broad range
of transition-metal-catalyzed asymmetric hydrogenation pro-
cesses.1 Of the methods available for the synthesis of
P-stereogenic compounds,2 one of the most successful
involves the asymmetric lithiation-trapping of phosphine
boranes or sulfides using an organolithium reagent/(-)-
sparteine chiral base complex.3 Such an approach has been
used in the synthesis of P-stereogenic phosphine and
diphosphine ligands, and representative examples include
BisP*,4 Mini-PHOS,4b,5 and P,N-ligands6 (Figure 1). The
opposite configuration of the ligands depicted in Figure 1
can be accessed by employing the (+)-sparteine surrogates
Figure 1. P-Stereogenic mono- and diphosphine ligands.
(developed in our laboratory7), as first demonstrated by Kann
and co-workers8 and later by ourselves.9 However, to obviate
the need for synthesis of the (+)-sparteine surrogates, we
(4) (a) Imamoto, T.; Watanabe, J.; Wada, Y.; Masuda, H.; Yamada, H.;
Tsuruta, H.; Matsukawa, S.; Yamaguchi, K. J. Am. Chem. Soc. 1998, 120,
1635. (b) Gridnev, I. D.; Yamanoi, Y.; Higashi, N.; Tsuruta, H.; Yasutake,
M.; Imamoto, T. AdV. Synth. Catal. 2001, 343, 118.
† University of York.
‡ Celtic Catalysts Ltd.
(5) Yamanoi, Y.; Imamoto, T. J. Org. Chem. 1999, 64, 2988.
(6) Johansson, M. J.; Andersson, K. H. O.; Kann, N. J. Org. Chem.
2008, 73, 4458.
(1) (a) Cre´py, K. V. L.; Imamoto, T. AdV. Synth. Catal. 2003, 345, 79.
(b) Cre´py, K. V. L.; Imamoto, T. Top. Curr. Chem. 2003, 229, 1. (c) Tang,
W.; Zhang, X. Chem. ReV. 2003, 103, 3029. (d) Zhang, W.; Chi, Y.; Zhang,
X. Acc. Chem. Res. 2007, 40, 1278.
(7) (a) Dearden, M. J.; Firkin, C. R.; Hermet, J.-P. R.; O’Brien, P. J. Am.
Chem. Soc. 2002, 124, 11870. (b) O’Brien, P. Chem. Commun. 2008, 655.
(8) (a) Johansson, M. J.; Schwartz, L. O.; Amedjkouh, M.; Kann, N. C.
Eur. J. Org. Chem. 2004, 1894. (b) Johansson, M. J.; Schwartz, L. O.;
Amedjkouh, M.; Kann, N. Tetrahedron: Asymmetry 2004, 15, 3531.
(9) Genet, C.; Canipa, S. J.; O’Brien, P.; Taylor, S. J. Am. Chem. Soc.
2006, 128, 9336.
(2) (a) Pietrusiewicz, K. M.; Zablocka, M. Chem. ReV. 1994, 94, 1375.
(b) Grabulosa, A.; Granell, J.; Muller, G. Coord. Chem. ReV. 2007, 251,
25. (c) Glueck, D. S. Chem.sEur. J. 2008, 14, 7108.
(3) Muci, A. R.; Campos, K. R.; Evans, D. A. J. Am. Chem. Soc. 1995,
117, 9075.
10.1021/ol901977p CCC: $40.75
Published on Web 10/01/2009
 2009 American Chemical Society

have now developed an alternative approach to P-stereogenic
compounds with the configuration opposite to those shown
in Figure 1. Herein, we present our results.
phosphine sulfide (S)-8, which could be distinguished by 1H
NMR spectroscopy. Ultimately, we hoped to optimize
conditions for regioselective lithiation at the unsubstituted
methyl group, thus allowing access to achiral bis-silyl
phosphine sulfide 7 selectively.
Our proposed strategy for the asymmetric synthesis of
P-stereogenic compounds from dimethylphosphine sulfides
1 is outlined in Scheme 1 and comprises three key steps: (i)
n-BuLi/(-)-sparteine-mediated asymmetric lithiation-trap-
ping of phosphine sulfide 1 f 2 in which a suitable
functional group (X) would be introduced; (ii) regioselective,
kinetically controlled lithiation R to phosphorus on the less
sterically encumbered methyl group and subsequent elec-
trophilic trapping to give 3, and (iii) removal of the functional
group X in which the C-X bond is transformed into a C-H
bond to produce substituted phosphine sulfide (R)-4.10 In
contrast, direct asymmetric lithiation-trapping of phosphine
sulfide 1 using n-BuLi/(-)-sparteine would deliver (S)-4 (the
configuration present in the ligands in Figure 1). For our
strategy to be successful, the functional group X needed to
be easy to introduce/remove and had to be compatible with
the rather harsh lithiation-trapping conditions (2 f 3)
typically employed. With this in mind, we selected trialkyl-
silyl substituents (X ) SiR3) for our investigation. A related
strategy with phosphine boranes in which X ) OH (via
dianion chemistry) has been used by Imamoto to prepare
ent-BisP*, although a two-step method for converting
CH2OH into CH3 was employed.11
Scheme 2. Synthesis of Silyl Phosphine Sulfide (S)-6
Lithiation of silyl phosphine sulfide (S)-6 using n-BuLi
in Et2O at -78 °C for 3 h and subsequent electrophilic
trapping gave adduct (S)-8 exclusively (68% yield) (Table
1, entry 1). This clearly indicated preferential lithiation R to
the silyl substituent, which was opposite to that required for
our purposes (see Scheme 1). However, on switching to the
more sterically hindered base s-BuLi, lithiation-trapping gave
a 50:50 mixture of 7 and (S)-8 from which we isolated a
50% yield of adduct 7 (entry 2). A similar outcome was
obtained using n-BuLi and s-BuLi in THF (entries 3 and 4).
With a view to increasing the amount of lithiation at the
methyl group in (S)-6, we investigated the effect of diamines
(TMEDA 9 and bispidines 10 and 11) and a triamine
(PMDETA 12) (entries 5-13). Use of n-BuLi/TMEDA in
Et2O and s-BuLi/bispidine 10 in Et2O or THF all gave (S)-8
only, although the isolated yields were e44% (entries 5, 8
and 9). In contrast, use of s-BuLi/TMEDA 9 or PMDETA
12 in Et2O or THF and subsequent trapping generated the
highest proportions of adduct 7 (regioselectivity ranging from
70:30-80:20), entries 6, 7, 12, and 13). The highest overall
yields of 7 and (S)-8 were obtained in THF, and from these
regioselective lithiations it was possible to isolate 67-69%
yields of adduct 7 (entries 7 and 13). The highest level of
regioselectivity in the desired sense (i.e., to give 7) was
obtained using s-BuLi/PMDETA 12. We speculate that this
is due to the increased steric hindrance and basicity of the
s-BuLi/PMDETA reagent, which leads to preferential lithia-
tion at the methyl group in a kinetically controlled event (at
-78 °C).
Finally, we explored the use of different equivalents of
s-BuLi/PMDETA 12 and different lithiation times (Table 2).
Interestingly, use of s-BuLi/PMDETA 12 in THF and
increasing the lithiation time to 6.5 h led to poor regiose-
lectivity: a 60:40 mixture of 7 and (S)-8 was obtained (Table
2, entry 1). The usual 3 h lithiation time generated a 75:25
mixture of 7 and (S)-8 under comparable conditions (entry
2). To explain this, we suggest that preferential lithiation at
the methyl group occurs but during the extended lithiation
time, equilibration of the carbanion to the thermodynamically
preferred position (R to phosphorus and silicon) occurs. As
a result, we focused our attention on shorter lithiation times
(entries 3 and 4), and the best compromise between yield
and regioselectivity was observed using 1.5 equiv of s-BuLi/
Scheme 1
.
Strategy for the Asymmetric Synthesis of Substituted
Phosphine Sulfides (R)-4
To start with, PhMe2Si-substituted phosphine sulfide (S)-6
was prepared using our previously reported procedure.12
Thus, treatment of phosphine sulfide 5 with n-BuLi/(-)-
sparteine in Et2O at -78 °C and subsequent reaction with
PhMe2SiCl delivered silyl phosphine sulfide (S)-6 in 88%
yield (88:12 er by CSP-HPLC) (Scheme 2). Next, a wide
range of conditions was examined to explore the regiose-
lectivity of the lithiation of phosphine sulfide (S)-6 (Tables
1 and 2). Our plan was to trap the lithiated intermediate(s)
derived from (S)-6 with PhMe2SiCl as this would generate
achiral bis-silyl phosphine sulfide 7 or chiral bis-silyl
(10) We have assigned phosphine sulfide 4 in Scheme 1 with (R)-
configuration. This assumes that the CH2E substituent is of higher priority
in the Cahn-Ingold-Prelog rules than the R and CH3 groups.
(11) Cre´py, K. V. L.; Imamoto, T. Tetrahedron Lett. 2002, 43, 7735.
(12) Gammon, J. J.; Canipa, S. J.; O’Brien, P.; Kelly, B.; Taylor, S.
Chem. Commun. 2008, 3750.
Org. Lett., Vol. 11, No. 21, 2009
5023

to us ((S)-6 is not crystalline), the corresponding trimethyl-
silyl derivative, (S)-13, is crystalline and appeared suitable
for our purposes. Thus, silyl phosphine sulfide (S)-13 was
produced using our previously reported catalytic asymmetric
deprotonation protocol12 (1.1 equiv of n-BuLi, 0.05 equiv
(-)-sparteine in toluene). Recrystallization of the crude
product delivered a 56% yield of (S)-13 of 99:1 er (by CSP-
HPLC) (Scheme 3).
Table 1. Optimization of the Regioselective Lithiation of Silyl
Phosphine Sulfide (S)-6: Effect of Base, Solvent, and Ligand
Scheme 3
.
Synthesis and Regioselective Lithiation of Silyl
yield of 7c yield of
Phosphine Sulfide (S)-13
entry basea solvent ligand 7/(S)-8b
(%)
(S)-8c (%)
1
2
3
4
5
6
7
8
n-BuLi Et2O
s-BuLi Et2O
n-BuLi THF
s-BuLi THF
n-BuLi Et2O
0:100
50:50
0:100
60:40
0:100
80:20
70:30
0:100
0:100
f
0
50
0
37
0
50
67
0
0
0
26
60
69
68
32
60
15
44
8
19
9
40
0
9d
s-BuLi
s-BuLi
s-BuLi
s-BuLi
s-BuLi
s-BuLi
s-BuLi
s-BuLi
Et2O
THF
Et2O
THF
Et2O
THF
Et2O
THF
9d
9e
10d
10d
11d
11d
12g
12h
9
10
11
12
13
40:60
75:25
75:25
27
19
15
Regioselective lithiation-trapping (with Me3SiCl) of (S)-
13 using the lithiation conditions optimized above for (S)-6
(1.5 equiv s-BuLi/PMDETA 12 in THF at -78 °C for 1.5 h)
was rather disappointing as only a 42% yield of adduct 14
was obtained (Scheme 3). Surprisingly, we obtained a 20%
yield of (R)-16 in which two new trimethylsilyl groups had
been incorporated. In this case, use of excess s-BuLi has a
detrimental effect on the lithation. We suspect that the
initially formed carbanions trap to give 14 and (S)-15 and
these intermediates are then deprotonated by the excess
s-BuLi (which presumably reacts slowly with Me3SiCl at
-78 °C). Subsequent trapping then gives (R)-16. To cir-
cumvent this problem, the amount of s-BuLi/PMDETA 12
used was reduced to 1.0 equiv resulting in a 72% yield of
adduct 14 (with no formation of (R)-16) (Scheme 3).
Finally, it was necessary to demonstrate that silyl phos-
phine sulfide (S)-13 of 99:1 er could be used to access chiral
phosphines with the configuration opposite to that obtained
by direct lithiation-trapping with organolithium reagents and
(-)-sparteine. Thus, (S)-13 was subjected to regioselective
lithiation and trapped with DMF to generate aldehyde (S)-
17. Crude aldehyde (S)-17 was reduced (NaBH4) and then
desilylated (TBAF) to give a 50% yield (after chromatog-
raphy) of hydroxy phosphine sulfide (S)-19 (98:2 er) over
three steps (Scheme 4). As anticipated, the adduct generated
in this way had the configuration opposite to that obtained
from 5 by direct lithiation (n-BuLi/(-)-sparteine)-DMF
trapping-reduction (as shown by CSP-HPLC, see the Sup-
porting Information).13 The phosphine borane corresponding
to 19 (P-BH3 in place of PdS) has been used by Kann to
prepare P,N-ligands6 (Figure 1). In a similar way, lithiation-
a Reaction conditions: 1.1 equiv of RLi, 1.2 equiv of ligand (if used).
b Ratio determined by 1H NMR spectroscopy of the crude product. c Yield
after purification by column chromatography. d 1.0 equiv of RLi/ligand.
e 1.2 equiv of s-BuLi/9, 1.5 h lithiation time. f H NMR spectrum of crude
product indicated starting material only. g 1.1 equiv of s-BuLi/12. h 1.2 equiv
of s-BuLi/12.
Table 2. Optimization of the Regioselective Lithiation of Silyl
Phosphine Sulfide (S)-6: Effect of Lithiation Time
s-BuLia
(equiv)
yield of 7c
yield of
(S)-8c (%)
entry
time (h) 7/(S)-8b
(%)
1
2
3
4
1.2
1.1
1.5
1.1d
6.5
3
1.5
0.5
b
60:40
75:25
85:15
80:20
44
60
72
60
1
43
19
14
9
a s-BuLi/ligand (equiv). Ratio determined by H NMR spectroscopy
of the crude product. c Yield after purification by column chromatography.
d 1.1 equiv of s-BuLi/1.2 equiv 12.
PMDETA 12 in THF at -78 °C for 1.5 h. In this way, a
72% yield of adduct 7 was isolated (entry 3).
Although we had now identified optimum conditions for
the required regioselective lithiation of silyl phosphine sulfide
(S)-6, limitations to our proposed methodology remained:
(S)-6 was produced in only 88:12 er by a lithiation using
stoichiometric amounts of (-)-sparteine. Although improve-
ment of the er of (S)-6 via recrystallization was not available
(13) The sense of induction obtained by direct lithiation of phosphine
sulfide 5 using n-BuLi/(-)-sparteine is known (see ref 11).
5024
Org. Lett., Vol. 11, No. 21, 2009

(a protected form of MiniPHOS5) (Scheme 4). Lithiation of
(S)-13 was followed by reaction with Ph2PCl; subsequent
treatment with sulfur and then TBAF delivered a 69% yield
of (R)-21 (95:5 er) over the three steps. A similar sequence
but trapping with t-BuPCl2 and then reaction with MeMgBr
gave (R,R)-22 (99:1 er) in 46% yield. In the synthesis of
(R,R)-22, high stereoselectivity in the creation of the second
stereocenter at phosphorus (formed in the trapping with
t-BuPCl2) was observed: none of the meso diastereomer was
produced. This contrasts with the ∼1:1 diastereoselectivity
observed for the trapping of lithiated tert-butyldimethylphos-
phine borane in Imamoto’s synthesis of MiniPHOS.5 We
believe that this stereocontrol is due to the phosphine sulfide
and not due to the silyl substituent since lithiation of 5 using
n-BuLi/(-)-sparteine and subsequent reaction with t-BuPCl2,
MeMgBr, and then sulfur delivered a ∼3:1 mixture of (S,S)-
22 and its meso diastereomer.
Scheme 4
.
Synthesis of P-Stereogenic Mono- and Diphosphine
Sulfides
In summary, the successful implementation of a new
strategy for accessing P-stereogenic compounds with the
opposite configuration to that obtained by direct lithiation-
trapping using organolithium reagents and (-)-sparteine is
described. A study on the factors controlling the regioselec-
tivity of lithiation of silyl phosphine sulfides is also reported.
We believe that both these aspects will prove useful to those
involved in the synthesis of P-stereogenic phosphine ligands
via lithiation processes.
Acknowledgment. We thank the EPSRC for a DTA award
(to J.J.G.) and Celtic Catalysts for funding.
oxygenation of (S)-13, and then desilylation (using CsF and
18-crown-6) gave hydroxy phosphine sulfide (R)-20 (55%
yield, 97:3 er) (Scheme 4). Imamoto has used the phosphine
borane corresponding to 20 to prepare QuinoxP*.14
Supporting Information Available: Full experimental
1
procedures, characterization data, and copies of H NMR
and 13C NMR spectra. This material is available free of
charge via the Internet at http://pubs.acs.org.
Next, we prepared two diphosphine sulfides (R)-21
(a protected analogue of trichickenfootphos15) and (R,R)-22
OL901977P
(14) Imamoto, T.; Sugita, K.; Yoshida, K. J. Am. Chem. Soc. 2005, 127,
11934.
(15) Hoge, G.; Wu, H.-P.; Kissel, W. S.; Pflum, D. A.; Greene, D. J.;
Bao, J. J. Am. Chem. Soc. 2004, 126, 5966.
Org. Lett., Vol. 11, No. 21, 2009
5025