Synthesis, resolution and structure of axially chiral atropisomeric N- arylimides

Article abstract of DOI:10.1016/S0040-4039(00)00835-2

Reported is the synthesis and structures of a new series of axially chiral molecules based on restricted rotation. The ortho-substituted N- arylimides have enantiomeric atropisomers that are stable and separable at room temperature. The compounds are notable for their ease of synthesis as well as positioning of functionality (a carboxylic acid and an amine) in a highly asymmetric configuration. (C) 2000 Published by Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00835-2

Tetrahedron Letters 41 (2000) 5431±5434
Synthesis, resolution and structure of axially chiral
atropisomeric N-arylimides
Ken D. Shimizu,* Heather O. Freyer and Richard D. Adams
Department of Chemistry and Biochemistry, University of South Carolina, Columbia, SC 29208, USA
Received 15 April 2000; accepted 24 May 2000
Abstract
Reported is the synthesis and structures of a new series of axially chiral molecules based on restricted
rotation. The ortho-substituted N-arylimides have enantiomeric atropisomers that are stable and separable
at room temperature. The compounds are notable for their ease of synthesis as well as positioning of
functionality (a carboxylic acid and an amine) in a highly asymmetric con®guration. # 2000 Published by
Elsevier Science Ltd.
Keywords: atropisomers; resolution; X-ray crystal structures.
Molecules with axial chirality are among the most pro®cient in e€ecting enantioselective dif-
ferentiation. The most common examples are the chiral biphenyls and binaphthyls that are
ubiquitous components in asymmetric catalysts, resolving agents and chiral auxiliaries.1 Despite
their utility, the synthesis of new chiral biaryls remains challenging due to the diculties in
forming the Caryl±Caryl bond.2 Therefore, we have set out to develop the synthesis of a new class
of axially chiral molecules that retain the enantioselective properties of the biaryls but have
greater accessibility and variability in their structures. Our e€orts have focused on ortho-sub-
stituted chiral arylimides in which restricted rotation leads to the formation of two stable
enantiomeric atropisomers. The ortho-aryl substituent is easily varied and creates a highly asym-
metric environment as exempli®ed by carboxylic acid 1 and quinoline 2.
Restricted rotation in N-arylimides has long been recognized,3 however, only recently has the
utility of the resulting atropisomers been demonstrated. These have included applications in
* Corresponding author.
0040-4039/00/$ - see front matter # 2000 Published by Elsevier Science Ltd.
PII: S0040-4039(00)00835-2

5432
molecular recognition,4 stereoselective reactions,5 and as enzyme mimics.6 For the most part,
these applications have centered on the chemistry of the imide portions of the molecules, in par-
ticular as dieneophiles or electrophiles.5,6 Our strategy is to focus instead on the ortho-substituent
on the aryl ring as a handle for applications as resolving agents, transition metal catalysts, and
NMR shift reagents.7
The most attractive attribute of the new chiral compounds is the ease with which they can be
synthesized. The central Caryl±Nimide bond is eciently formed by simply heating an ortho-sub-
stituted aniline with a cyclic anhydride neat in vacuo at 150ꢀC.5a,8 To generate enantiomeric
atropisomers an unsymmetrical anhydride is required in which there are two di€erent types of
carbonyls. Diphenyl succinic anhydride 5 was selected because the phenyl groups project out of the
plane of the imide ring and block one face of the ortho-substituent.9 Molecular modeling suggested
that a single ortho-substituent would be insucient to inhibit rotation about the Caryl±Nimide
bond at ambient temperatures.10 Therefore, a second ortho-substituent (^Me) was inserted as a
blocking group. The resulting anilines 4 and 7 were readily synthesized from their corresponding
nitro arenes (Scheme 1). Condensations with succinic anhydride 5 cleanly yielded the desired aryl
imides 1 and 2.
Scheme 1.
The rotational barriers for 1 and 2 were determined indirectly by study of their monophenyl
analogs 8 and 9. Restricted rotation was immediately evident by the observation of two
diastereomeric rotamers (phenyl up and phenyl down) that were stable and separable at room
temperature. Equilibration experiments revealed a rotational barrier for carboxylic acid 8 of 28.0
kcal/mol (t1/2=40 h at 60ꢀC).11 The quinoline succinimide 9 was considerably more stable with a
barrier of 33.2 kcal/mol (t1/2=190 h at 106ꢀC). Molecular modeling gave the same ordering of
rotational barriers; however, the absolute magnitudes were ꢁ12% lower at 25 and 29 kcal/mol,
respectively.12

5433
The three-dimensional structures of the axially chiral acid and amines 1 and 2 were established
by X-ray crystallography (see Fig. 1).13 The crystals containing both enantiomers were obtained
by recrystallization of the racemates. The molecular structures are almost identical to those cal-
culated by molecular modeling. The cyclic imide and aryl planes are nearly perpendicular (85.4ꢀ
for 1 and 71.9ꢀ for 2), placing the carboxylic acid and quinoline amine in the desired chirally
enshrouded environment. Approach from one face of the functionality is hindered by the steric
bulk of the 3,3-diphenyl groups.
Figure 1.
Resolution of the enantiomeric conformations was achieved in the case of the acid 1 by ester-
i®cation with (S)-1-phenyldecanol (Scheme 2). The reaction of the highly hindered acid 1 was
done under Mitsunobu conditions that proceeds with absolute inversion of stereochemistry.14
The resulting diastereomeric esters were separable by ¯ash chromatography. Subsequent removal
of the chiral auxiliary under mild hydrogenolysis conditions a€orded the partially resolved acids
10a and 10b (ee >95%). Complete resolution was achieved by recrystallization of the partially
resolved acids in acetone to yield pure crystals of the predominant enantiomer.15 The quinoline
arylimide 2 has not been resolved in large scale. However, analytical chiral chromatography
(Covalent Pirkle HPLC column) clearly shows the presence of both isomers.16
We are currently exploring applications that utilize the chiral environments in carboxylic acid 1
and quinoline 2. For example, carboxylic acid 1 is an excellent NMR shift reagent for a-methyl-
benzyl amine. We have also synthesized axially chiral N-arylimides with higher rotational barriers.
For example, replacement of the ortho-methyl substituent with a t-butyl group lead to extremely
stable atropisomers (ÁEa>35 kcal/mol).

5434
References
1. Pu, L. H. Chem. Rev. 1998, 98, 2405±2494.
2. (a) Stanforth, S. Tetrahedron 1998, 54, 263±303. (b) Bringmann, G.; Walter, R.; Weirich, R. Angew. Chem., Int.
Ed. Engl. 1990, 29, 977±991.
3. (a) Verma, S. M.; Singh, N. D. Aust. J. Chem. 1976, 29, 295±300. (b) Kishikawa, K.; Yoshizaki, K.; Kohmoto, S.;
Yamamoto, M.; Yamaguchi, K.; Yamada, K. J. Chem. Soc., Perkin Trans. 1 1997, 1233±1239.
4. (a) Shimizu, K. D.; Dewey, T. M.; Rebek Jr., J. J. Am. Chem. Soc. 1994, 116, 5145±5149. (b) Rebek Jr., J. Angew.
Chem., Int. Ed. Engl. 1990, 29, 245±255.
5. (a) Curran, D. P.; Qi, H.; Geib, S. J.; DeMello, N. C. J. Am. Chem. Soc. 1994, 116, 3131±3132. (b) Osamu, K.;
Izawa, H.; Sato, K.; Dobashi, A.; Taguchi, T. J. Org. Chem. 1998, 63, 2634±2640.
6. Ohno, A.; Kunitomo, J.; Kawai, Y.; Kawamoto, T.; Tomishima, M.; Yoneda, F. J. Org. Chem. 1996, 61, 9344±
9355.
7. Dai, X.; Wong, A.; Virgil, S. C. J. Org. Chem. 1998, 63, 2597±2600.
8. Moormann, A. E.; Yen, C. H.; Yu, S. Synth. Commun. 1987, 17, 1695±1699.
9. Miller, C. A.; Long, L. M. J. Am. Chem. Soc. 1951, 73, 4895±4898.
10. The barriers of rotation were calculated using MM2 as implemented in the software program MacroModel v5.5
(Mohamadi, F.; Richards, N. G. J.; Cuida, W. C.; Liskamp, R.; Lipton, M.; Cau®eld, C.; Chang, G.;
Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990, 11, 440±467). The rotational barrier varied greatly with the
step angle of the dihedral driver. Accurate values were attained only for small step angles <1ꢀ. Larger step angles
would jump over the barrier and predict considerably smaller rotational barriers.
11. Calculated from the Arrhenius equation assuming an ideal value of A=2.08Â1010.
12. The discrepancy may be due to the severe distortions in the transition state that lead to signi®cant ÁS
contributions to the observed rotational barrier that are not accounted for in the molecular modeling force®elds.
13. Crystal data for acid 1: space group=P21/c, a=7.6765(7) A, b=19.431(3) A, c=13.378(1) A, ꢀ=90.00(0)ꢀ,
ꢁ=97.879(8)ꢀ, ꢂ=90.00(0)ꢀ, Z=4, 2920 re¯ections, R=0.034. Crystal data for amine 2: space group=P21/n,
a=8.667(2) A, b=9.316(3) A, c=24.669(4) A, ꢀ=90.00(0)ꢀ, ꢁ=93.48(2)ꢀ, ꢂ=90.00(0)ꢀ, Z=4, 3402 re¯ections,
R=0.04395. Crystallographic data for the structural analysis has been deposited with the Cambridge
Crystallographic Data Center, CCDC No. CCDC 139122 for 1 and CCDC 139123 for 2. Copies of this
information may be obtained free of charge. www:http://www.ccdc.cam.ac.uk.
14. Mitsunobu, O. Synthesis 1981, 1±28.
15. Characterization data for (+)-1 and (^)-1: 1H NMR (300 MHz, acetone-d6) ꢃ 8.04 (d, 1H, J=6.43 Hz), 7.5±7.2 (m,
12H), 3.74 (d, 1H, J=18.19 Hz), 3.45 (d, 1H, J=18.24 Hz), 1.90 (s, 3H); IR (KBr) 3056, 1781, 1714, 1591, 1474,
1442 cm^1; found (HRMS, FAB) m/z 385.1308; calcd for C24H19NO4: 385.1314; mp 225±228ꢀC, [ꢀ]2D5=^26.0 (c
1.350, 1:1 CH2Cl2/MeOH).
16. Compound (+/^)-2 1H NMR (300 MHz, CDCl3): ꢃ 8.67 (dd, 1H, J=1.65 Hz and 4.23 Hz), ꢃ 8.11 (d, 1H, J=1.63
Hz), 7.80 (d, 1H, J=8.46 Hz), 7.54±7.24 (m, 12H), 3.86 (d, 1H, J=19.4 Hz); 3.74 (d, 1H, J=18.28 Hz), 2.23 (s,
3H); found (HRMS, FAB): m/z 392.1521; calcd for C26H20N2O2: 392.1525. IR 3059, 2924, 2924, 1714, 1398, 1208
cm^1. Mp 194ꢀC.