Synthesis and conformational study of a novel macrocyclic chiral(salen) ligand and its uranyl and Mn complexes.

A novel chiral macrocyclic ligand incorporating a chiral salen moiety into a framework containing two biphenyl units was synthesized. Structural properties and conformational aspects of the free ligand and an UO2 complex were studied by using NMR spectroscopy in solution and MM calculations. The Mn(III) complex was tested as catalyst in enantioselective oxidation of prochiral unfunctionalized olefins to the corresponding optically active epoxides under very mild conditions.


Introduction
The successful design, synthesis and use of molecules capable of the selective recognition of other species is of great interest in catalysis, separations, enzyme functions and other fields involving molecular recognition [1]. Chiral salen-metal complexes are currently recognized versatile, practical and efficient catalysts for a large number of asymmetric reactions [2][3][4][5][6][7][8][9][10][11][12]. Moreover, in recent years, much attention has been devoted to the synthesis of new neutral ditopic salen receptors able to simultaneously bind chiral ammonium cations and their counteranions [13]. Generally, in addition to electrostatic interactions, hydrogen bonding and dispersive non bonding π-π interactions cooperate for binding affinity of intimate contact ion pair in organic solvents.

OPEN ACCESS
In this paper we describe the synthesis of a novel 21-membered macrocyclic ligand incorporating a chiral salen moiety into a framework containing two biphenyl units. The salen moiety, due to the presence of two stereogenic carbon atoms in the diimine bridge, generates a chiral pocket which can coordinate metal cations (via imine nitrogen and oxygen phenolic atoms), such as uranyl or Mn. In particular the uranyl cation can be employed as a Lewis acidic site able to bind ion pairs [13] whereas the Mn metal center can act as a catalytic site in the enantioselective epoxidation of olefins [14].

Results and Discussion
Synthesis of the macrocyclic diimine ligand 5 is shown in Scheme 1. The bis(hydroxymethyl) dimer 1 was prepared in 40% yield as reported elsewhere [15]. The selective alkylation of both phenolic hydroxyl groups with (CH 3 ) 2 CH(CH 2 ) 3 Br in refluxing dry acetonitrile in the presence of one equivalent of potassium carbonate as a base, afforded 2, which was purified by column chromatography (67% yield Compound 2 was converted into 3 by acid-catalyzed condensation with a large excess of p-t-butylphenol in the presence of TsOH in benzene giving after column chromatography a white solid (76% yield). The subsequent formylation [14] of 3 afforded dialdehyde 4 (70% yield), which finally was cyclized with (1R,2R)-1,2-diphenylethylenediamine in refluxing ethanol under high diluting conditions (40% yield) to yield the macrocycle 5.
The structural characterization of all new compounds 2-4 was achieved by ESI-MS measurements and 1D and 2D-NMR investigations. The 1 H-and 13 C-NMR spectra of compounds 2, 3 and 4 in CDCl 3 consist of relative simple patterns of resonances showing only one set of signals for each pair of identical groups. The formation of the macrocyclic ligand 5 was fully supported by the ESI-MS measurements (m/z = 1,077 [MH] + ).
The 1 H-NMR spectrum of 5 in CDCl 3 exhibits two sharp singlet signals at 8.42 and 4.73 ppm assigned to the azomethine CH=N protons and the diimine bridge protons, respectively, while the broad singlet at 13.36 ppm arise from phenolic hydroxylic protons. The AB system centered at 4.24 ppm (Δδ = 0.05 ppm, J = 16.9 Hz) was assigned to the diasterotopic methylene protons located between the two p-phenyl-phenoxy groups. The AB system centered at 4.12 ppm (Δδ = 0.22 ppm, J = 15.8 Hz) integrating for four protons was easily attributed to the remaining two methylene groups. Aromatic protons resonate in the appropriate low field region and alkyl ether substituents gave one set of the expected multiplets (1-2 ppm). This simple spectrum, supported by the T-ROESY data (Figure 1a), indicates that the ligand assumes a C1 averaged symmetric cyclic structure in a non-interconvertible cone conformation ensured by the presence of sufficiently bulky 4-methylbutyl groups at the lower rim. In agreement with NMR spectroscopic data, MMFFs force field [16] calculation produced the computed lowest-energy structure of macrocycle 5 (Figure 1b). Ligand 5 was utilized to prepare both UO 2 (VI) and Mn(III) complexes 6-UO 2 and 6-Mn (Scheme 1). To a stirred solution of ligand 5 in EtOH, solid (AcO) 2 UO 2 •2H 2 O or Mn(AcO) 3 •2H 2 O respectively was added. The mixtures were allowed to stir overnight at room temperature and were monitored by TLC. Evaporation of the solvent gave a residue, which was dissolved in CH 2 Cl 2 , filtered and concentrated to produce the uranyl or Mn complex in a nearly quantitative yield. ESI-MS spectra confirmed the formation of the mono-metallic complexes.
Structural information on the uranyl(VI) complex (6-UO 2 ), in solution, were obtained from 1D and 2D-NMR studies. The 1 H-NMR spectrum in acetone-d 6 of 6-UO 2 complex showed an almost unvaried pattern of signals with respect to the free ligand, even considering the expected broadening and downfield shift of the resonances arising from the protons close to the coordination site. The ROE relationships measured by the phase-sensitive T-ROESY spectrum in acetone-d 6 are shown in Figure 2a.
The dipolar correlation contacts identified in the 2D map suggested that the symmetrical structure is maintained in solution upon metal complexation. The strong ROE correlations between aromatic protons signals and azomethine CH=N protons, diimine bridge protons and t-Bu groups allowed safe assignment of the partially overlapped signals of the aromatic protons system. Besides the characteristic bond-through ROE correlations in the aromatic moiety, no additional dipolar contacts were observed either between aromatics rings or between the aromatic protons and iso-hexyloxysubstituents. Interestingly, this finding ruled out any complex conformation where the diimine bridge phenyl rings were arranged in an almost anti-periplanar conformation. In agreement with MM calculated lowest energy structure, the complex adopts in solution a less hindered "disk-shaped" rather a "cup-like" shape ( Figure 2b).

(a) (b)
The broadening of the lines of the 1 H spectrum might envisage the flipping motion of the salicylaldehyde framework already observed for uranyl salophen and salen complexes [13,[17][18][19]. In order to enlighten on possible exchanging process occurring in solution, the proton signals of the uranyl complex were monitored carefully at lower temperature ( Figure 3). As the temperature was lowered all the signals progressively broadened. At 193 K the signals of CH=N and diimine bridge protons were no longer visible and methylene protons located between the macrocycle aromatic rings coalesced into an unresolved envelope. Further cooling to 183 K did not produce additional detectable changes in the spectral profile neither new signals could be revealed. Only trivial modifications were observed in the range 183-213 K in the aromatic spectra, easily attributed to slower bi-phenyl rotation.  Table 1. The reactions were allowed to proceed overnight (24 h) reaching 80%-90% completion (entries 2-4) and quantitative yield in the corresponding epoxides (entries 1-5). The observed effect of the coligand (compare entries 1 and 2) and the absolute configuration of the cis-epoxides were in agreement with literature data [14,[20][21][22].
Previously reported studies indicated that the dissymmetry of diimine bridge should favour the approaching of the si enantioface of the alkene towards the Mn-oxo site of chiral salen catalysts [14]. The observed ee values for the alkenes reported in Table 1 are in line with this finding. However the degree of enantioselectivity displayed is moderate. In order to try to understand the observed behaviour, molecular mechanics calculations were performed on the complexes. Figure 4 shows the attack to the catalyst of both re and si face of the dihydronaphthalene. Very small differences were found for calculated energies (<5 kJ mol −1 ) suggesting that the attack pathways by the re and si face are energetically equivalent. Similar results were obtained in all the modelled structures of the examined alkenes and the catalyst. This behaviour might be ascribed to the presence of the biphenyl rings, which can undergo atropoisomerization and then can adapt their spatial position to the approaching guest, favouring therefore π-π interactions with the aromatic framework of the alkene regardless of the alkene exposed face.

Conclusions
We have synthesized a new macrocyclic chiral salen ligand and its UO 2 and Mn(III) complexes. NMR studies of the (salen) UO 2 complex are in agreement with MM calculated structures which indicate that the shallow conformation of the receptor cavity and, probably, the mobility of biphenyl rings are responsible of the observed moderate enantioselectivity. At any rate, the catalyst synthesized in this work represents the first example of salen derivatives combined with biphenyl units and the work is in progress to avoid the atropoisomerization of biphenyl units, in order to control more efficiently molecular recognition.