and 2/0.5 equiv (entries 4 and 8), though that in 2/2 equiv
was not improved (entry 2). Among other molar ratios of
ZnBr2 examined, 1 and 2 equiv were similarly effective (cf.,
footnote c of Table 1).8
p-BrC6H4CH2MgBr/CuBr·Me2S. The substrate (∼100% ee
by chiral HPLC) was synthesized by a method delineated in
Scheme 3 starting with (1R)-monoacetate 16, obtained by
lipase-catalyzed hydrolysis of the corresponding diacetate
followed by recrystallization.9 Allylic substitution of (1R)-
5a with the copper reagent derived from p-BrC6H4CH2MgBr
(2.0 equiv) and CuBr·Me2S (1.0 equiv) proceeded smoothly
under the above conditions (0 °C, 1 h) to afford 7a (R )
TBS) regioselectively, which was isolated as a mixture with
(p-XC6H4CH2)2 (X ) Br and H). The mixture was treated
with Bu4NF, and the alcohol separated by chromatography
was resilylated to 7a for the further reaction (92% from (1R)-
5a).
A copper reagent derived from BnMgCl (2.1 equiv) and
CuBr·Me2S (1.0 equiv) in the presence of ZnBr2 (3.0 equiv)
afforded 11a as well (entry 6; see entry 5 for the result
obtained in the absence of ZnBr2). These results with
BnMgCl are informative in a case where benzylic magnesium
bromides are hardly accessible from ArCH2Br and Mg due
to the rapid homocoupling reaction. In addition, ZnCl2 in
place of ZnBr2 retarded the reaction with the copper reagents
derived from BnMgCl/CuBr·Me2S in 2.1/1.0 and 2.0/0.5
equiv at 0 °C, and further reaction at rt produced a mixture
of 11a and 13a, though complete regioselectivity was
observed (data not shown).
Next, the reaction conditions of entry 4 of Table 1 were
applied to substrates rac-5b-f, which possess protective
groups other than the TBS group (Table 2). In all cases ZnBr2
assisted exclusive production of the anti SN2′ products
11b-f. Interestingly, the native selectivity obtained without
ZnBr2 (ratios in parentheses) was dependent on the protective
group: low selectivity with the electron-donating groups
(TBDPS and Bn in entries 1 and 2) as in the case of the
TBS group and high selectivity with the electron-withdrawing
group (Ac and Piv in entries 4 and 5). In addition, entries 4
and 5 show a chemoselectivity indicating the picolinoxy
group is the better leaving group compared to the AcO and
PivO groups (see ref 8).
Scheme 3
.
Synthesis of the Key Intermediate 7a through the
Picolinate (1R)-5a
Table 2. Effect of Substituent on the Regioselectivitya
Allylic picolinate 8, another key substrate, was prepared
by a method shown in Scheme 4. Wittig reaction of aldehyde
19, obtained from lactate 18 (natural form) in two steps, with
an ylide derived from [Ph3PEt]+Br- and NaN(TMS)2 af-
forded cis olefin 20 exclusively in 65% overall yield.
Removal of the THP group in MeOH was successful.
Unfortunately, removal of MeOH co-extracted with volatile
21 by evaporation resulted in substantial loss of 21. To avoid
the loss ethylene glycol was used as a solvent. The alcohol
21 extracted was free of the glycol for the next esterification
with 2-PyCO2H to afford picolinate 8 (96% ee (by chiral
HPLC)) in 66% yield from 20.
ratiob c d of
combined
yield, %e
,
,
entry
1
substrate
R1
11:12
rac-5b
TBDPS
99:1
(60:40)
99:1
(66:34)
100:0
(95:5)
100:0
(99:1)
100:0
(100:0)
nd
nd
nd
nd
95
83
82
76
90
82
2
3
4
5
rac-5c
rac-5d
rac-5e
rac-5f
Bn
PMB
Ac
Scheme 4. Synthesis of Picolinate 8
Piv
a Reactions with BnMgBr (2.1 equiv) and CuBr·Me2S (1.0 equiv) were
carried out in the presence of ZnBr2 (3.1 equiv) in THF/Et2O (3-7:1) at 0
°C for 1 h. b Determined by 1H NMR spectroscopy. Zero (0) indicates the
1
case that signals for 12 were not seen in the expanded H NMR spectra.
c The corresponding alcohol and the starting compound were not obtained.
d The ratios obtained without ZnBr2 are shown in parentheses. e nd, not
determined.
With the above results in mind, we chose (R)-5a (R )
TBS) as a substrate in the first allylic substitution with
Org. Lett., Vol. 11, No. 5, 2009
1105