COMMUNICATION
cleavage of the CF3CO moiety might render this amide an
ideal nucleophile for our purpose.
The asymmetric conversion of racemic allenyl acetate
(ꢀ)-3 into allenyl trifluoroacetamide 5 f was thus investigat-
ed (Table 1). Preliminary tests indicated no reactivity in the
Table 1. Search for the best chiral ligand.[a,b]
Entry
Ligand
t [h]
T [8C]
Yield [%]
e.r.[c]
1
2
3
4
5
6
7
8
9
L2
L3
L4
L5
L6
L7
L8
L9
L10
L11
6
4
24
16
4
25
25
55
55
25
25
25
25
25
25
87
79
8
43
96
92
64
68
49
75
21:79
47:53
50:50
32:68
51:49
50:50
14:86
86:14
87:13
17:83
3
24
24
24
24
Scheme 3. Scope of the hydroalkylation reaction (all reactions were run
on a 0.5 mmol scale. Yields refer to isolated yields). a) Only the trans
isomer was observed. b) 75:25 diastereoisomeric mixture.
10
[a] For the structures of the chiral ligands see the Supporting Informa-
tion. [b] No reaction was observed when using ligand L1. [c] e.r.’s deter-
mined by chiral HPLC SFC (AD-H, CO2/iPrOH 90:10, 2.5 mLminꢁ1
,
85 bar).
a moderate yield (58%). As the same lactam 2b had been
previously obtained by us by means of a Pd0-catalyzed ally-
lation process, the stereochemistry was unambiguously as-
signed as the trans configuration between the newly formed
stereocenters.[4a,7] The cyclohexyl-substituted precursor 1c
reacted smoothly to afford the desired lactam 2c in 71%
yield. The replacement of the acetyl group by the bulkier
pivaloyl group was also well tolerated, leading to pyrroli-
done 2d in 69% yield. Precursors bearing a stereogenic al-
lenyl moiety, such as 1e and 1 f afforded the (E)-alkene cyc-
lization products 2e and 2 f in good (72%) to excellent
(90%) yields, respectively.[8] The allyl-substituted allenyl b-
ketoamide 1aa could also be cyclized to the corresponding
pyrrolidone 2aa bearing a quaternary carbon atom, al-
though in a rather moderate yield (34%). Unfortunately, re-
lated substrates bearing electron-withdrawing groups other
than a ketone (MeO2C-, NC-, PhSO2-) failed to cyclize.
The observation that the chiral racemic allenic precursors
1e and 1 f led to single trans-(E)-diastereomers suggested
that the transformation might be stereospecific. This would
imply that starting from an enantioenriched substrate, axial-
to-central chirality transfer should be at work. Therefore,
we decided to test the cyclization of an enantioenriched
sample of 1 f. For this purpose, we planned to synthesize the
parent N-allenyl amine by an asymmetric palladium-cata-
lyzed amination of the corresponding racemic N-allenyl ace-
tate. Such type of asymmetric transformation has already
two precedents in the literature.[9] However, the reported
methods were mainly developed for the synthesis of tertiary
allenyl amines. As our synthetic plan required the use of
a secondary N-allenyl amine, we selected N-benzyl trifluor-
oacetamide (4) as the nucleophile. Indeed, we reasoned that
its easy deprotonation together with the smooth hydrolytic
presence of the Trost standard ligand L1,[10] whereas promis-
ing results were obtained with (R)-BINAP (L2; BINAP=
2,2’-bis(diphenylphosphino)-1,1’-binaphthyl).
A
careful
screening with this ligand involving reaction temperature,
palladium source, base, solvent, and additives as parameters
allowed to retain the following conditions: [NaH, [Pd2-
AHCTUNTGRENGUN(N dba)3]·CHCl3 (dba=dibenzylideneacetone; 2.5 mol%),
(R)-BINAP (7.5 mol%), LiCl (20 mol%), THF, RT.[11] Sub-
sequent use of these optimized reaction conditions in the
presence of (S,S)-DIOP (L3),[12] (S,S)-CHIRAPHOS (L4),[13]
(S)-tBu-PHOX (L5),[14] SL-W001-1 (L6),[15] SL-J005-
1 (L7),[16] (R)-SYNPHOS (L8),[17] (S)-MeOBIPHEP (L9),[18]
(S)-3,5-tBuMeOBIPHEP (L10),[19] and (R)-3,5-tBu-4-MeO-
MeOBIPHEP (L11)[19] indicated the atropoisomeric ligands
L2 (Table 1, entry 1) and L9 (Table 1, entry 8) as the ligands
of choice.
Further experiments[20] showed that the product isolated
after partial conversion possessed the same enantiomeric
ratio (e.r.)[21] (22:78) as that isolated after total substrate
conversion. Furthermore, the unreacted substrate recovered
from the above experiment was found to be enantioenriched
(e.r. 71:29) and its resubmission to the same reaction condi-
tions until full conversion, gave the final amide (ꢁ)-5 f,
again in 21:79 e.r.[21] (61% yield). This result clearly rules
out a static kinetic resolution phenomenon and points to
a dynamic kinetic asymmetric resolution (DYKAT)[22] with-
out the involvement of a memory effect.[23] Finally, resub-
mission of a racemic sample of the allenyl trifluoroaceta-
mide 5 to the asymmetric amination conditions, gave back,
as expected, the unchanged racemic material, thereby con-
ꢁ
firming the irreversibility of the N C bond-formation step.
Chem. Eur. J. 2012, 18, 3840 – 3844
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3841