Enantioselective Synthesis of Axially Chiral Natural Product Hybrids
FULL PAPER
this mixture (154 mg) with TBAF in THF (3 mL) gave, after flash chro-
matography (silica gel, heptanes/ethyl acetate 1:1), biaryl (S,aR)-9c as a
white solid (108 mg, 54% from (S)-10a). [a]2D2 =+53 (c=1.15, CHCl3);
m.p. 1798C; 1H NMR (300 MHz, [D6]DMSO) d=7.08 (s, 1H), 6.97 (s,
1H), 6.52 (s, 1H), 6.03 (s, 2H), 5.03 (t, J=5.1 Hz, 1H), 4.79 (d, J=
4.5 Hz, 1H), 4.30 (m, 1H), 4.11 (dd, J=13.4, 5.3 Hz, 1H), 3.92 (dd, J=
13.4, 5.3 Hz, 1H), 3.83 (s, 3H), 3.75 (s, 3H), 3.50 (s, 3H), 0.99 (d, J=
6.3 Hz, 3H) ppm; 13C NMR (75 MHz, [D6]DMSO) d=152.4, 149.9, 146.7,
145.3, 140.1, 139.6, 136.3, 125.8, 124.3, 109.8, 106.3, 105.3, 100.8, 65.2,
(1m) was added dropwise until the aqueous phase reached pH 12. The
aqueous layer was then extracted with dichloromethane and the com-
bined organic layers were washed with brine, dried over MgSO4, filtered,
and evaporated under vacuum. The residue was purified by flash chroma-
tography (silica gel, dichloromethane/methanol 95:5 then 9:1) to give di-
benzazepine (R,aR)-5c as an oil in 88% ee (7.7 mg, 95%). [a]2D5 =ꢁ47
(c=0.87, CHCl3); HPLC (Chiralpak AD, hexane/iPrOH 95:5 + 0.1%
Et3N, 1.0 mLminꢁ1) tR =27.5 min (major enantiomer), 34.4 min (minor
enantiomer); 1H NMR (300 MHz, CDCl3) d=7.06 (s, 1H), 7.00 (s, 1H),
6.75 (s, 1H), 6.05 (d, J=1.8 Hz, 1H), 6.02 (d, J=1.8 Hz, 1H), 4.36 (br s,
1H), 3.92 (s, 3H), 3.91 (s, 3H), 3.85–3.77 (m, 2H), 3.71 (s, 3H), 3.45 (d,
J=12.6 Hz, 1H), 1.63 (d, J=6.6 Hz, 3H) ppm; 13C NMR (75 MHz,
CDCl3) d=153.4, 150.8, 147.6, 147.3, 142.9, 130.4, 130.0, 129.2, 125.9,
110.3, 108.7, 105.7, 101.5, 61.2, 61.1, 56.2, 50.2, 48.0, 17.2 ppm; IR (neat):
n˜ =2929, 1484, 1457, 1409 cmꢁ1; HRMS (ESI) calcd for C19H22NO5 [M+
H+]: 344.1498; found: 344.1505.
60.6, 60.5, 60.4, 55.6, 25.2 ppm; IR (neat): n˜ =3392, 2935, 1479 cmꢁ1
;
HRMS (ESI) calcd for C19H22O7Na [M+Na+]: 385.1263; found:
385.1260.
Biaryl (S,aS)-9d (Scheme 7, Figure 1): From the precedingSuzuki cou-
plingof iodide ( S)-10a and boronate 8b, a small amount (30.5 mg) of a
mixture of the minor diastereomer (S,aS)-9b and by-products was isolat-
ed. This reaction mixture (27.5 mg) was treated with TBAF in THF
(1.5 mL) to give, after flash chromatography (heptanes/ethyl acetate 1:1),
Synthesis of dibenzazepine (R,aR)-5c by mesylation (Scheme 10): Tri-
ethylamine (7.2 mL, 0.052 mmol) and methanesulfonyl chloride (3 mL,
0.039 mmol) were added dropwise to a solution of (S,aR)-9 f (12 mg,
0.026 mmol) in dichloromethane (1 mL) at 08C. After stirringfor 1 h at
room temperature, water was added and the aqueous layer was extracted
with dichloromethane. The combined organic layers were washed with a
saturated aqueous NaHCO3 solution, brine, dried over MgSO4, filtered,
and evaporated under vacuum. The residue was purified by preparative
TLC (silica gel, heptanes/ethyl acetate 7:3) to give tBoc-protected diben-
1
biaryl (S,aS)-9d as an oil (15.8 mg, 9%). H NMR (300 MHz, CDCl3) d=
7.06 (s, 1H), 6.79 (s, 1H), 6.53 (s, 1H), 6.00 (s, 2H), 4.48 (q, 1H, J=
6.3 Hz), 4.23 (s, 2H), 3.91 (s, 3H), 3.89 (s, 3H), 3.65 (s, 3H), 2.90 (br s,
1H), 1.39 (d, J=6.3 Hz, 3H) ppm; 13C NMR (75 MHz, CDCl3) d=152.3,
151.6, 147.8, 146.7, 141.9, 137.8, 134.8, 128.1, 127.0, 109.8, 108.7, 105.8,
101.3, 66.2, 63.1, 61.0, 56.2, 22.9 ppm.
General cyclodehydration procedure (Schemes 6, 7, 10): A solution of
diol (1 equiv) in dichloromethane (c=0.02m) was cooled down to the ap-
propriate temperature, a solution of TFA (5 equiv) in dichloromethane
(0.35m) was then added dropwise and the reaction was run until com-
plete conversion of the startingmaterial (followed by TLC: an aliquot of
the reaction mixture was washed with a saturated aqueous NaHCO3 solu-
tion and extracted with ethyl acetate before beingspotted on the TLC
plate). A saturated aqueous NaHCO3 solution was added and the aque-
ous layer extracted with dichloromethane. The combined organic layers
were washed with brine, dried over MgSO4, filtered, and evaporated
under vacuum. The residue was purified by preparative TLC (silica gel,
heptanes/ethyl acetate).
1
zazepine (R,aR)-5d as an oil (8.9 mg, 77%). H NMR (300 MHz, CDCl3)
d=7.13 (s, 1H), 6.83–6.63 (m, 2H), 6.02 (d, J=1.4 Hz, 1H), 6.02 (d, J=
1.4 Hz, 1H), 5.07–4.66 (m, 2H), 3.92 (s, 3H), 3.91 (s, 3H), 3.64–3.50 (m,
4H), 1.51 (s, 9H), 0.89 (d, J=6.9 Hz, 3H) ppm; 13C NMR (75 MHz,
CDCl3) d=153.9, 153.1, 150.7, 146.9, 142.7, 133.0, 131.6, 128.7, 126.5,
111.7, 110.3, 108.5, 101.4, 79.9, 61.4, 60.6, 56.9, 56.3, 46.7, 28.8, 21.1 ppm;
IR (neat): n˜ =2930, 1681, 1395 cmꢁ1
C24H29NO7Na [M+Na+]: 466.1842; found: 466.1812.
;
HRMS (ESI) calcd for
mixture of
A
(R,aR)-5d (8.9 mg, 0.02 mmol), dichloromethane (1 mL), and TFA
(0.5 mL) was stirred for 45 min at 208C. An aqueous solution of NaOH
(1m) was added and the aqueous layer was extracted with dichlorome-
thane. The combined organic layers were dried over MgSO4, filtered, and
evaporated under vacuum. The residue was purified by flash chromatog-
raphy (silica gel, dichloromethane/methanol 95:5 then 9:1) to give diben-
zazepine (R,aR)-5c as an oil in 94% ee (6.7 mg, 97%).
Dibenzoxepine (R,aR)-5a (Scheme 7): The above general cyclodehydra-
tion procedure from diol (S,aR)-9c (44.0 mg, 0.12 mmol) in dichlorome-
thane (4 mL) at ꢁ508C gave, after preparative TLC (heptanes/ethyl ace-
tate 1:1), dibenzoxepine (R,aR)-5a as
a white solid with 96% ee
(35.5 mg, 86%, 96:4 mixture of interconverting atropisomers). [a]D24
=
ꢁ117 (c=1.09, CHCl3); HPLC (Chiralpak AD, hexane/ethanol 99:1,
1.0 mL.minꢁ1) tR =17.4 min (major enantiomer), 29.2 min (minor enan-
tiomer); m.p. 1048C ; 1H NMR (300 MHz, CDCl3) d=7.14 (s, 1H), 7.00
(s, 1H), 6.74 (s, 1H), 6.04 (d, J=1.5 Hz, 1H), 6.02 (d, J=1.5 Hz, 1H),
4.24 (d, J=11.3 Hz, 1H), 4.24 (q, J=6.6 Hz, 1H), 3.98 (d, J=11.3 Hz,
1H), 3.94 (s, 3H), 3.92 (s, 3H), 3.72 (s, 3H), 1.56 (d, J=6.6 Hz, 3H)
ppm; 13C NMR (75 MHz, CDCl3) d=153.1, 150.5, 147.3, 146.9, 142.7,
131.7, 131.4, 130.9, 126.3, 109.8, 108.3, 105.5, 101.3, 68.7, 68.1, 61.2, 61.0,
56.2, 18.2 ppm; IR (neat): n˜ =2936, 1483 cmꢁ1; HRMS (ESI) calcd for
C19H20O6Na [M+Na+]: 367.1158; found: 367.1140.
Calculations: Three-dimensional structures of 5a–c and 5 f (Scheme 6,
Figure 4): One thousand conformations of each compound were generat-
ed by random search Monte Carlo method and optimized by molecular
mechanics PRCG minimization method usingthe Macromodel (version
5.5) program with the MM2 force field.[48] The search was carried out on
blocks of 100 Monte Carlo steps until no additional conformation was
found to be of lower energy than the current minimum. From these con-
formational searches, all possible conformations within 3 kcalmolꢁ1 from
the global minimum were analyzed. For each compound, the geometries
of the most stable conformations were retained. These geometries were
used for buildingthe three-dimensional structures in the calculations of
the formation enthalpy usinga molecular-orbital semiempirical method.
Geometries were optimized by means of a gradient technique at RHF/
Synthesis of dibenzoxepine (S,aS)-5a by cyclodehydration with Deoxo-
fluor (Scheme 7): A solution of Deoxofluor (12.5 mL, 0.065 mmol) in di-
chloromethane (60 mL) was added dropwise to a stirred solution of diol
(S,aR)-9c (9.5 mg, 0.026 mmol) in dichloromethane (1 mL) at ꢁ788C.
The reaction mixture was stirred at ꢁ788C for 50 min, warmed to room
temperature and treated with a saturated aqueous NaHCO3 solution.
After extraction of the aqueous layer with dichloromethane, the com-
bined organic layers were washed with brine, dried over MgSO4, filtered,
and evaporated under vacuum. The residue was purified by preparative
TLC (silica gel, heptanes/ethyl acetate 3:2) to give (S,aS)-5a as a white
powder in 75% ee (4.7 mg, 52%). [a]2D3 =+119 (c=1.0, CHCl3); HPLC
(Chiralpak AD, hexane/ethanol 99:1, 1.0 mLminꢁ1) tR =15.2 min (minor
enantiomer), 24.1 min (major enantiomer).
[50]
AM1 level,[49] usingthe MOPAC program (version 5.0).
ꢁ
Cyclodehydration mechanism (Scheme 8): The energy barriers for C1 C2
ꢁ
and C3 C4 bond rotations were determined by rotatingin steps of 15 8.
Only the correspondingdihedral angles were fixed, all the other parame-
ters were optimized. RHF/AM1 transition structures TS1 and TS2 were
located usingthe procedures implemented in MOPAC 5.0. All variables
were optimized by minimizingthe sum of the squared scalar gradients
(NLLSQ and SIGMA).[51] Force calculations were carried out to ensure
that the transition structures located had one imaginary frequency. Final
ꢁ1
values of the gradient norms were <1 kcal and each transition struc-
Dibenzazepine (R,aR)-5c (Scheme 10): A solution of TFA (0.75 mL) in
dichloromethane (1 mL) was added dropwise over 1.5 h to a solution of
(S,aR)-9 f (11 mg, 0.024 mmol) in dichloromethane (1 mL) at ꢁ788C. The
reaction mixture was stirred for 2 h at ꢁ788C and then allowed to warm
up to room temperature over 30 min. An aqueous solution of NaOH
ture had one negative eigenvalue in the Hessian matrix as required. The
activation enthalpies were obtained by the difference between the forma-
tion enthalpy of the fully optimized reactant in the ground state with the
formation enthalpy of the correspondingtransition structures.
Chem. Eur. J. 2007, 13, 5450 – 5465
ꢀ 2007 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
5463