Enantioselective Induction of Helical Chirality in Cyclooctapyrroles by Metal‐Complex Formation

Abstract

Fixed chirality: The treatment of cyclooctapyrroles (see picture) with a metal source with optically active carboxylate or amine ligands leads to enantioselective metalation to give stereochemically stable helical mononuclear and dinuclear complexes without a chiral auxiliary. The helicity of the dicopper complex was determined by the simulation of the CD spectrum on the basis of X-ray crystallographic data. Linear oligopyrroles generate helical structural motifs that are of great interest in supramolecular chemistry.1 Not only are helicates formed by the tetradentate coordination of bisdipyrrin ligands to a single metal, but two bisdipyrrin units assemble into dinuclear metal complexes of double helicates, such as 1 (Scheme 1).2 Homohelicity induction in these helicates has been the focus of a number of studies in view of their application to chirality sensing and asymmetric catalysis. Helical chirality is usually controlled by introducing a chiral auxiliary to cause the generation of a particular helical form in diastereomeric excess. For example, the axial coordination of chiral amines to bilinone–zinc complexes and the addition of a homochiral binol linker or homochiral amide substituents to a bisdipyrrin skeleton led to the successful induction of homohelicity.3, 4 However, at present it is difficult to generate stereochemically stable double-helicate metal complexes. Furthermore, the enantioselective induction of helicate formation with a chiral promoter molecule in the absence of chiral auxiliaries has not been reported. This point is important, because the binding of a chiral guest to a host in a mixture of diastereomers would be too much complicated. Helical structures of bisdipyrrin dinuclear complexes. Stereochemically stable helicate metal complexes closely related to 1 can be prepared by using cyclooctapyrroles, such as 2, in which π-conjugated bisdipyrrin units are linked by sp3-hybridized carbon atoms (Scheme 1).5 The figure-eight loop of cyclooctapyrroles is preorganized with metal-binding sites formed from four pyrrole nitrogen atoms. Mononuclear and dinuclear metal complexes of these macrocycles have been reported.6 However, studies on the helical chirality of cyclooctapyrroles are limited.7 Vogel et al. investigated the optical resolution of [36]octaphyrin(2.1.0.1.2.1.0.1) (3)8 and its dinuclear palladium(II) complex 3 Pd2 by HPLC on a chiral phase, and the helicity of each enantiomer of 3 Pd2 was determined unambiguously by X-ray crystallography.7a However, the circular dichroism (CD) spectra of 3 Pd2 have not been reported, and there is some complexity derived from rotation about the bond between the pyrrole α carbon atom and the ethylene carbon atom in the ethylene-bridged bipyrrole moiety at the crossing point upon the conversion of 3 into 3 Pd2. A conformational change due to rotation about this bond might be the reason why the order of HPLC elution of the two enantiomers was reversed after metal incorporation. Since the CD Cotton effect is very important in a discussion on chirality, it seemed necessary to confirm the relationship between the helical handedness of cyclooctapyrroles and their CD Cotton effect. The macrocycle 3 contains a helicene-like bisdipyrrin π system surrounding each metal site and an ethylene-bridged bisdipyrrin π system connecting the two metal sites. Thus, the application of the exciton chirality method to predict the helical handedness of 3 on the basis of the CD Cotton effect is not straightforward.9 The framework of 2, in which a pair of electronically isolated bisdipyrrin chromophores cross over, may be better suited for the exciton chirality method. We describe herein the synthesis, optical resolution, and chiroptical properties of the stereochemically stable cyclooctapyrrole metal complexes 2 M2. The helical chirality of 2 M2 was determined unambiguously on the basis of the theoretical simulation of the CD spectra by using the X-ray crystallographic data. Furthermore, we have found that metal sources with optically active carboxylate and amine ligands cause the enantioselective metalation of 2 to give helically asymmetric metal complexes with no chiral auxiliary. Inversion of the figure-eight loop of the cyclooctapyrrole-free base 2 occurs through unwinding of the molecular twist and then rewinding in the opposite direction. This molecular motion was detected in variable-temperature 1H NMR spectroscopic experiments in [D10]o-xylene (see the Supporting Information). Two signals due to the diastereotopic methylene hydrogen atoms of the isobutyl groups, at δ=3.0 and 1.9 ppm at 60 °C, coalesced at temperatures over 120 °C. The inversion barrier was estimated to be at least 77 kJ mol−1 on the basis of the frequency difference (Δν=435 Hz) and the coalescence temperature (Tc>393 K). As interconversion between the enantiomeric forms is prohibited in their metal complexes, optical resolution of the complexes is possible. The treatment of the free base 2 with one molar equivalent of Cu(OAc)2⋅H2O in CH2Cl2–MeOH at room temperature for 1 h afforded a mononuclear copper(II) complex, 2 Cu, in 43 % yield, whereas the treatment of 2 with 10 molar equivalents of Cu(OAc)2⋅H2O in the presence of triethylamine at room temperature for 5 h gave a dinuclear copper(II) complex, 2 Cu2, in 50 % yield. The absorption band in the visible region of the UV/Vis spectrum shifted from 541 to 570 and then 584 nm upon consecutive metalation of 2 with Cu (Figure 1 a). The mononuclear cobalt(II) complex 2 Co was also obtained in 85 % yield by the treatment of 2 with excess Co(OAc)2⋅4 H2O at room temperature for 2 h. The treatment of 2 Co with excess Cu(OAc)2⋅H2O at room temperature afforded a heterodinuclear complex, 2 CoCu, in 78 % yield. The absorption band in the visible region of the UV/Vis spectrum shifted from 541 to 569 and then 591 nm upon consecutive metalation of 2 with Co and Cu (Figure 1 b). a) UV/Vis spectra of 2, 2 Cu, and 2 Cu2 in CH2Cl2; b) UV/Vis spectra of 2, 2 Co, and 2 CoCu in CH2Cl2; c) CD spectra of optically resolved 2 Cu2 (first (1st) and second (2nd) HPLC eluate) in CH2Cl2; d) CD spectra of optically resolved 2 CoCu (first and second HPLC eluate) in CH2Cl2. X-ray crystallographic analysis of 2 Cu, 2 Cu2, 2 Co, and 2 CoCu indicated that 2,2′-bipyrrole units in the s-trans conformation are located at the crossing point of the figure-eight loop, and that the metal atom is disordered between two sites in the complexes 2 Cu, 2 Co, and 2 CoCu (Figure 2). The side view of these complexes shows the rhombic cavity enclosed by four panels of the planar dipyrrin unit connected with two hinges in the form of bipyrrole C2C2′ bonds and two gem-dimethyl-substituted bridging carbon atoms. ORTEP drawings of 2 Cu2 at the 50 % level with atom numbering; left: top view, right: side view. Four phenyl substituents are omitted at C5, C14, C24, and C33, and eight alkyl substituents are omitted at the bipyrrole β positions. The dinuclear complexes 2 Cu2 and 2 CoCu were resolved by HPLC with a chiral column (Daicel chiralpak IB) by eluting with n-hexane/CH2Cl2 (40:1). The first fraction of 2 Cu2 showed positive CD signals at 697 and 496 nm and a negative CD signal at 583 nm (Figure 1 c). Similarly, the first fraction of 2 CoCu showed positive CD signals at 711 and 492 nm and a negative CD signal at 590 nm (Figure 1 d). Theoretical UV/Vis and CD spectra (Figure 3 a,b) were calculated on the basis of the atom coordinates obtained from the X-ray crystallographic structure of 2 Cu2 in the M,M helical form; all peripheral substituents were omitted (see the Supporting Information). The calculated spectra are in good agreement with the observed spectra for the enantiomer of 2 Cu2 eluted second in terms of band position and intensity. Furthermore, calculations based on time-dependent density functional theory (TDDFT) revealed that the electronic transitions along the long axes of the π-conjugated bisdipyrrin chromophores connecting the two copper sites are responsible for the absorption band 18 in the theoretical UV/Vis spectra. Since this absorption band correlates with a negative sign in the theoretical CD spectrum, the absolute configuration of 2 Cu2 determined by calculation is in agreement with that determined on the basis of the CD exciton chirality method. In the figure-eight screw shown in Figure 3 c, the transition in the front bisdipyrrin chromophore couples with the transition in the back bisdipyrrin chromophore. When the coupling dipole direction is counterclockwise, as for the M,M enantiomer, a negative Cotton effect can be predicted according to the CD exciton chirality method.9 Thus, M,M helicity was assigned to the enantiomer of 2 Cu2 eluted second from the HPLC column. a) Theoretical and experimental UV/Vis spectra; b) theoretical and experimental CD spectra; c) coupling electronic transitions that give rise to the absorption band 18 and the sign of the corresponding Cotton effect in the left-handed cyclooctapyrrole M,M enantiomer. The theoretical spectra were calculated for the macrocyclic core of (M,M)-2 Cu2, and the experimental spectra were recorded for the enantiomer of 2 Cu2 that was eluted second by HPLC. The bars indicate the positions and oscillator strength (f) or rotary strength (Rr) of the transitions (18, 21, 22) calculated by TDDFT. We reported previously that porphyrinoids can be induced to adopt a particular unidirectional helical conformation preferentially by protonation with optically active carboxylic acids.7b, 10 In fact, the addition of 200 equivalents of (S)-(+)-mandelic acid to a solution of 2 in CH2Cl2 caused the appearance of a CD spectrum that is quite similar to that of (M,M)-2 Cu2 (see the Supporting Information). This result led us to investigate the enantioselective formation of stable cyclooctapyrrole metal complexes. The copper(II) salt prepared by the addition of (S)-(+)-mandelic acid sodium salt (4 equiv) in MeOH to CuCl2⋅2 H2O in water was used as the metal source. The treatment of copper(II) (S)-(+)-mandelate (30–40 equiv) with 2 in CH2Cl2 for 24 h resulted in the formation of a mixture of 2 Cu and 2 Cu2. The CD spectrum of the product 2 Cu2 showed a negative first Cotton effect at 697 nm. The ee value of 2 Cu2 based on the CD intensity (12 % ee) was in good agreement with that determined by HPLC analysis (14 % ee). Complex 2 Cu formed in this experiment showed negative CD signals at 669 and 484 nm and a positive CD signal at 564 nm. These wavelengths are only slightly blue-shifted (12–28 nm) relative to those of 2 Cu2, which suggests that the monocopper complex also has M,M helicity. This monocopper complex was converted into the dicopper complex by using excess Cu(OAc)2⋅H2O in the presence of triethylamine at room temperature. CD analysis of the product verified the ee value of 12 %. Thus, the ee value was not affected by the second metalation process to give 2 Cu2. The reaction of 2 with the copper(II) carboxylate derived from (R)-(−)-mandelic acid led to 2 Cu2 with 13 % ee and with a positive first Cotton effect at 697 nm. Optically active amines were used under homogeneous reaction conditions for the asymmetric synthesis of 2 Cu and 2 Cu2 in a much shorter reaction time than the 24 h required for the complete conversion of 2 under the heterogeneous reaction conditions with mandelic acid. When 2 was added to a solution of (R)-(+)-1-(1-phenyl)ethylamine (10 equiv) and CuCl2⋅2 H2O (5 equiv) in CH2Cl2, 2 reacted completely to give 2 Cu in 2 h. However, the ee value for helical chirality was very small (Table 1, entry 1). An increase in the concentration of (R)-(+)-1-(1-phenyl)ethylamine (80 equiv) retarded the metalation reaction, but the ee value of 2 Cu was improved to 13 % (Table 1, entry 2). The use of a large excess of a mixture of (R)-(+)-1-(1-phenyl)ethylamine and CuCl2⋅2 H2O in a 2:1 ratio at low temperature further improved the ee value of 2 Cu (Table 1, entry 4). Entry Molar ratio (amine/Cu/2) t [h] T [°C] CD sign (669 nm) ee [%][b] (2 Cu) 1 10:5:1 2 RT + 2 2 80:5:1 5 RT + 13 3 800:400:1 0.25 RT + 11 4 800:400:1 1.5 0 + 19 As noted above, the rates for the first metal-insertion process to give (M,M)-2 Cu and (P,P)-2 Cu (k1M and k1P, respectively) are dependent on the temperature and the amount of the optically active amine used (Table 1 and Scheme 2). To gain insight into the effect of the optically active amine on the second metalation step, a racemic mixture of 2 Cu was treated with (R)-(+)-1-(1-phenyl)ethylamine (10 equiv) and CuCl2⋅2 H2O (5 equiv) in CH2Cl2. Although these reaction conditions do not lead to a remarkable difference between k1M and k1P (Table 1, entry 1), an ee value of 33 % was observed for the P,P helical form of 2 Cu2 in the initial stages of the second metalation step (Table 2, entry 1). Thus, the rate (k2P) for the metal insertion into (P,P)-2 Cu is twice as high as the rate (k2M) for the metal insertion into (M,M)-2 Cu. As the metalation proceeds, (P,P)-2 Cu is consumed faster, which is consistent with the development of an excess of the M,M helical form of 2 Cu. Thus, 72 % ee was observed for the M,M helical form of 2 Cu in the final stage of the metalation (Table 2, entry 5). These results show that kinetic resolution may be a potential approach to the preparation of metal porphyrinoids with high optical purity. Metalation of 2 and 2 Cu with an optically active metal source. L*=(R)-(+)-1-(1-phenyl)ethylamine. Entry t [h] 2 Cu 2 Cu2 Molar ratio CD sign (669 nm) ee [%][b] CD sign (697 nm) ee [%][b] 2 Cu/2 Cu2 1 3 − 3 + 33 9:1 2 9 − 15 + 29 6:4 3 12 − 25 + 25 5:5 4 16 − 36 + 19 3:7 5 24 − 72 + 6 1:9 The unique figure-eight loop structure of cyclooctapyrroles has potential application to the synthesis of chiral functional materials. For this purpose, an understanding of the chiroptical properties of cyclooctapyrroles on the basis of their structure is of great importance. Mononuclear cobalt(II) and copper(II) complexes and homo and heterodinuclear complexes of a cyclooctapyrrole with two gem-dimethyl-substituted bridging carbon atoms were prepared in this study. The crossing of discrete bisdipyrrin chromophores in the figure-eight loop is an advantageous structural feature for the determination of the absolute configuration on the basis of the exciton chirality method. The absolute configuration was confirmed by the theoretical CD spectrum calculated for (M,M)-2 Cu2 on the basis of X-ray crystallographic data. The induction of excess unidirectional helicity in the metalation processes not only gives insight into the mechanism of chirality transfer but also provides a convenient method for the preparation of optically active metal cyclooctapyrroles for further applications. 2 Cu2 (50 % yield): UV/Vis (λmax (logε), CH2Cl2): 384 (4.63), 584 (4.92), 660 (shoulder, 4.61), 711 nm (shoulder, 4.38); ESIMS: m/z calcd for C86H88N8Cu2: 1360.58; found: 1360.30; magnetic moment (Evans method, CDCl3): 3.14; elemental analysis: calcd (%) for C86H88N8Cu⋅C6H14⋅H2O: C 75.43, H 7.16, N 7.65; found: C 75.51, H 6.94, N 7.65. Recrystallization from CH2Cl2/hexane gave crystals for X-ray crystallographic analysis. Crystal data: C92H102N8Cu2⋅2 C6H14, Mr=1533.07, triclinic, space group Pca21, a=22.2077(16), b=16.1092(12), c=22.2493(16) Å, α=90, β=90, γ=90°, V=7959.7(10) Å3, Z=4, ρcalcd=1.279 Mg m−3, μ(MoKα)=0.589 mm−1, T=296(2) K, crystal size: 0.50×0.40×0.10 mm. A total of 15 899 unique reflections were collected (3.62<2θ<54.72°) by using graphite-monochromated MoKα radiation. The structure was solved by the direct method by using the SHELX97 package. The positions of all non-hydrogen atoms were refined anisotropically (930 parameters). All hydrogen atoms were placed at ideal positions and included in the refinement. R1=0.0427, wR2=0.1051 for 13 648 reflections with I>2.00σ(I); R1=0.0546, wR2=0.1156 for all data; GOF (on F2)=1.041. CCDC 698457 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.