A polypyrrolic molecular cage consisting of irreversible amide and C―N single bonds: Structure and anion binding properties

WEI Dehui，LI Xin，PU Junming，ZHANG Jun，CHEN Xi，ZHANG Zhan

（School of Chemistry and Materials Science， South-Central Minzu University， Wuhan 430074， China）

Abstract A chemically stable polypyrrolic cage constructed by irreversible amide and C—N bonds was synthesized through chemical reduction of its imine precursor. The structure of the molecular cage was characterized using NMR， MS and XRD. The anion binding properties were studied in deuterated chloroform or DMSO using NMR spectroscopy. The results indicated that this cage had a moderate affinity for anions such as Cl-， Br-， I-， HCO3-， AcO- and NO3- in CDCl3，and complexed with these anions in a 1∶1 mode through fast exchange， while the interaction with F-， SO42- and H2PO4-displayed slow anion-exchange kinetics.

Keywords molecular cage； anion binding； amide-amine hybridized cage； structure

The acidic NH group makes pyrrole a useful subunit for anion binding. Many polypyrrolic compounds，including linear， tripodal， macrocyclic and macrobicyclic ones （Fig.1）， have been discovered or developed as anion receptors or transporters［1-3］. For instances，prodigiosins （1）， a class of linear tripyrrolic compounds produced by mother nature， can cotransport HCl and trigger apoptosis［4-6］. Artificial polypyrrolic foldmers（2） bind sulfate or halides （especially fluoride and chloride） in water［7-8］. Polypyrrolic receptors with tripodal scaffolds （3）， on the other hand， bind rather large anions such as citrate， trimesic acid tricarboxylate and carbohydrates［9-11］. Compared with these open-chain receptors， polypyrrolic macrocycles （calix［n］pyrroles for example） have preorganized structures and are supposed to be superior receptors for anions. Indeed calix［4］pyrroles （4） show high affinity towards fluoride and chloride［12］. Great efforts have been paid to modify or functionalize calix［4］pyrroles to create anion receptors with enhanced performance［13-14］. One successful modification is the introduction of a “strap”to calix［4］pyrroles， resulting in the so-called strapped calix［4］pyrroles［15-16］. Following studies showed that this class of macrobicyclic receptors were not only excellent chloride receptors and transporters but also outstanding iron-pair receptors for KF and CsF［17-19］，depending on the structure of the strap introduced.Beside strapped calix［4］pyrroles， other macrobicyclic receptors such as C3symmetric polypyrrolic macrobicyclic receptors were also developed［20-21］， taking advantages of dynamic covalent bonds. An amine cage （5）， synthesized by reduction of a self-assembled imine cage［20］， was reported binding β-glucopyranosides. More recently an amide-imine hybridized polypyrrolic cage （6） was reported showing high affinity to tetrahedral oxyanion such as sulfate and pyrophosphate anions in chloroform［22］.

Fig.1 Representative polypyrrolic receptors图1 代表性多吡咯主体分子

Dynamic covalent bonds， especially imine bonds，have often been used to construct macrobicyclic molecules that are rather difficult to make through irreversible bonds［23-25］. However， molecules constructed through reversible imine bonds are readily to fall into parts under wet and acidic conditions. To improve the stability， imine bonds could be transformed to C—N single bonds by chemical reduction［20］. Since both amine cage 5 and amide-imine hybridized cage 6 showed interesting properties as supramolecular receptors， whether amideamine hybridized cage 9 with enhanced stability will function as a useful anion receptor needs to be explored.This paper herein reports the synthesis， characterization and anion binding studies of such a receptor.

1 Experimental

1.1 General

All reagents and solvents were purchased from commercial suppliers and used without further purification.Analytical thin-layer chromatography （TLC） was performed using commercial pre-coated silica gel plates containing a fluorescent indicator. Column chromatography was carried out using silica gel （0.030-0.040 mm）.1H NMR，13C NMR， and19F NMR spectra were recorded on Bruker AV400 and AV600 instruments.UV-Vis spectra were measured on a Varian Cary 5000 spectrophotometer. Mass spectra （MS） were taken on LTQ Orbitrap Elite （ESI）. X-ray crystallographic analyses were carried out on Bruker X8 APEX II instruments. Further details of the structures and their refinement are given in a later section.

1.2 Synthetic detail

Compound 11

To a 100 mL round bottom flask containing 10（2.00 g， 8.9 mmol） and NaOH （1.80 g， 45 mmol）were injected 80 mL ethanol and 16 mL water. The reaction mixture was stirred and heated at 90 ℃ for 6 h before it was cooled to r.t. After the removal of ethanol by distillation， 400 mL water was added to the mixture. The resulted mixture was subsequently treated with 3 mol‧L-1HCl so that the pH reached 3.The precipitate was collected by filtration and washed with water. After dried， the product was harvested as white solid （1.63 g， 93%）.

1H NMR （400 MHz， CDCl3）δ9.89 （s， 1H），9.78 （s， 1H）， 2.77 （dt，J= 13.0， 6.6 Hz， 4H），1.25 （t，J= 7.6 Hz， 3H）， 1.18 （t，J= 7.4 Hz， 3H）.13C NMR （101 MHz， DMSO-d6）δ181.7， 162.1，132.6， 131.9， 130.0， 124.2， 17.0， 16.9， 16.1.HRMS：m/z； ［M+H］+： calcd for C10H13NO3196.09682； found 196.09579.

Compound 7

To a 100 mL round bottom flask containing 1，3，5-tris（aminomethyl）-2，4，6-triethylbenzene trihydrochloride（0.61 g， 1.7 mmol） were injected 50 mL dry DMF and 0.7 mL （5.1 mmol） triethylamine. The reaction mixture was stirred at r.t. for 1 h before it was treated with 11（1.00 g， 5.1 mmol）， 1-（3-dimethylaminopropyl）-3-ethyl-carbodiimide hydrochloride （EDCI， 1.08 g， 5.6 mmol） and 1-hydroxybenzotrizole （HOBT， 0.75 g， 5.6 mmol）. After stirred at r.t for 3 days， the solvents were removed and the residue was extracted with DCM/water.A yellowish raw product was obtained after removal of solvent. The light yellow product （0.94 g， 71%） was purified through recrystallization out of THF/H2O （V（THF）/V（H2O）=1∶10）.

1H NMR （400 MHz， DMSO-d6）δ11.93 （s，3H）， 9.67 （s， 3H）， 8.12 （s， 3H）， 4.55 （s， 6H），2.83-2.66 （m， 18H）， 1.15-1.06 （m， 27H）.13C NMR（101 MHz， DMSO-d6）δ179.7， 159.8， 143.7， 135.6，132.1， 131.2， 128.6， 126.9， 37.1， 22.8， 17.5， 17.1，16.3， 16.1， 16.0. HRMS：m/z； ［M+Na］+： calcd for C45H60N6O6803.44665； found 803.44607.

Compound 8

The mixture of 1，3，5-tris（aminomethyl）-2，4，6-triethylbenzene trihydrochloride （0.43 g， 1.2 mmol），0.5 mL （3.6 mmol） triethylamine and 40 mL dry methanol was stirred at r.t. for 1h. To the mixture was charged 7 （0.94 g， 1.2 mmol） and then let the resulted mixture react at r.t. overnight. The white precipitate was filtered and washed with methanol.The product was dried and harvested as white solid（0.97 g， 83%）.

1H NMR （400 MHz， Chloroform-d）δ8.36 （s，3H）， 5.88 （t，J= 3.7 Hz， 3H）， 4.72-4.67 （m， 6H），4.50 （d，J= 3.7 Hz， 6H）， 2.84 （q，J= 7.4 Hz，6H）， 2.62 （q，J= 7.5 Hz， 18H）， 1.19 （qd，J= 7.5，3.1 Hz， 36H）；13C NMR （101 MHz， CDCl3）δ161.1，152.2， 144.4， 142.9， 133.6， 132.3， 132.3， 131.4，125.90， 124.0， 54.9， 38.3， 22.9， 22.5， 17.7， 17.6，16.8， 16.5， 16.0， 15.6. HRMS：m/z； ［M+Na］+： calcd for C60H81N9O3998.63546； found 998.63013.

Compound 9

To a 100 mL round bottom flask containing 8（1.00 g， 1.0 mmol），14 mL chloroform and 42 mL methanol were added NaBH4（0.14 g， 3.6 mmol）stepwise. The reaction mixture was stirred for 2 h before the removal of solvents using rotavap. The residue was then extracted with DCM/water （80 mL × 3）. The organic layer was dried and evaporated. The product（0.99 g， 99%） was harvested by recrystallization out of DCM/hexane （V（DCM）/V（hexane）=1∶20）.

1H NMR （400 MHz， Chloroform-d）δ8.95 （s，3H）， 5.54 （s， 3H）， 4.73 （s， 3H）， 4.49 （d，J= 3.9 Hz， 6H）， 3.96 （s， 6H）， 3.83 （s， 6H）， 2.84 （t，J=7.4 Hz， 6H）， 2.66-2.55 （m， 12H）， 2.43 （q，J= 7.5 Hz， 6H）， 1.27-1.18 （m， 27H）， 1.11 （t，J= 7.5 Hz，9H）.13C NMR （101 MHz， Chloroform-d）δ161.6，144.2， 141.7， 134.0， 133.0， 131.6， 127.9， 122.3，119.2， 47.3， 46.7， 38.1， 23.2， 22.9， 18.0， 16.9，16.8， 16.4， 16.4， 16.2. HRMS：m/z； ［M+H］+： calcd for C60H87N9O3982.70046； found 982.70435.

1.3 X-Ray crystallography

Single crystals suitable for X-ray diffraction were grown through slow diffusion of hexanes into an ethanol solution of 9. A needle crystal of C62H98N9O7having approximate dimensions of 0.2 mm × 0.2 mm ×0.15 mm was mounted on a glass fiber. All measurements were made on a Bruker APEX-II CCD diffractometer with graphite monochromated Mo-K radiation. The data were collected at a temperature of 200.0 K. Of the 10635 reflections that were collected，9845 were unique （Rint= 0.0456）； equivalent reflections were merged. Data were collected and integrated using the Bruker SAINT software package.The structure was solved by direct methods. All nonhydrogen atoms were refined anisotropically. All refinements were performed using the SHELXTL crystallographic software package Bruker-AXS. All CH hydrogen atoms were placed in calculated positions but were not refined.

Tab.1 Crystallographic data for compound 9表1 化合物9的晶体数据

2 Results and Discussion

The synthesis of target cage 9 is very straightforward（Fig.2）. The key precursor， tripodal 7， was prepared following reported procedure with a few modifications.The dynamic reaction of 7 with 1，3，5-tris（aminomethyl）-2，4，6-triethylbenzene in dry methanol smoothly produced amide-imine hybridized cage 8 as pale precipitate.Subsequent reduction of 8 with NaBH4provided amide-amine hybridized cage 9 almost quantitatively.The transformation from cage 8 to cage 9 could be feasibly confirmed by the changes of1H NMR spectra.Both the disappearance of imine CH signal at 8.36 and emergence of more peaks in the range of 3.5-9.0 supported the chemical conversion was successful.Deuterium exchange experiments suggested that the broad peaks at 8.95， 5.54 and 4.73 belonged to NH groups. These protons could be further attributed to pyrrole NH （Hb）， amide NH （Ha） and amine NH （Hc）respectively based on their chemical shifts and1H-1H COSY experiments. Other1H NMR signals also support the amide-amine hybridized structure. For instances，the protons （Hd） of CH2groups neighboring amide groups， similar to those of cage 8， resonated at 4.50.Meanwhile， the protons（ Heand H）fof CH2groups next to amine NHs resonated at 3.96 and 3.83. Compared with the same type protons of cage 8， the signals of these protons had significant upfield shifts due to the reduction of imine groups. It was also notable that the aliphatic protons were more split relative to those of cage 8， indicating cage 9 possesses a more flexible structure. As expected， receptor 9 displayed superior stability than 8 upon treatment of aqueous alkali or acid.Either alkali or acid didn′t to the decomposition of cage 9 while cage 8 degraded into 7 under such conditions.

Fig.2 Synthesis of polypyrrolic cage 9图2 多吡咯分子笼9的合成

Single crystals suitable for X-ray diffraction were grown through slow diffusion of an ethanol solution of 9. As shown in Fig.3， the X-ray crystal structure of polypyrrolic cage 9 was well consistent with the proposed structure （Solvent molecules were deleted for clarity）.The single bond nature of the amine C―N bonds could be identified unambiguously. The bond lengths of C―N bonds next to pyrrole rings had an average of 1.46 Å and the average bond length of C―N bonds next to benzene rings was 1.47 Å. Both types of bonds displayed single bond feature. As a comparison， the C=N of 8 was reported 1.26 Å long， displaying double bond character. In the case of amide groups， the average bond lengths of C=O was 1.24 Å and that of C―N was 1.35 Å. The bond lengths of C―N bonds between amide group and benzene ring had an average of 1.46 Å， again， displaying single bond feature. The distance between the two benzene rings is roughly 8.10 Å and the averaged distance between the two pyrrolic NH groups is 6.68 Å， forming a cavity with approximate radii of 3.8 Å.

The ability of cage 9 to bind anions in solution was initially examined in CDCl3using1H NMR spectroscopy. Upon treatment of anions in the form of（tetrabutylammonium） salts， significant chemical shifts of cage 9 were observed. As shown in Fig.4， anions such as Cl-， Br-， I-， HCO3-， AcO-， NO3-， SO42-and H2PO4-led to significant downfield shifts of pyrrole NH（Hb） and amide NH （Ha）， indicating cage 9 interacted with anions mainly through these acidic hydrogens.The interactions， however， were not so obvious by means of UV-Vis spectroscopy （cf. ESI）. Compared with 6［29］， the major absorption band of cage 9 had a blue shift due to the shorter conjugation of pyrrolic chromophores. Addition of an anion to its solution didn′t lead to significant changes of the absorption band. Therefore， the anion binding behaviors of cage 9 were essentially studied using NMR spectroscopy.

Fig.4 1H NMR titration of cage 9 with various anions in CDCl3图4 在氘代氯仿中进行的分子笼9的阴离子核磁滴定图谱

Upon titrating cage 9 with TBACl in CDCl3， the proton signal corresponding to pyrrole NH （Hb） was gradually shifted downfield from 8.4 to 10.6 while that of the amide NH （Ha） also experienced a similar downfield shift （cf. ESI）. The chemical shift changes（Δδ） caused by the titration were 2.2 for the pyrrole NH protons and 1.6 for the amide NH protons，respectively. When Br-or I-were used as guest for titration， similar downfield shifts were observed as well. The corresponding binding modes and constants in each case were determined and calculated using BindFit v5.0 （available from URL：http：//app.supramolecular.org/bindfit/） and the results are summarized in Tab.2.In the three cases， the 1∶1 binding was established and the binding constants were in the order of 103-104L‧mol-1，10-100 times less than those of the reported amideimine hybridized cage 6［29］. When the solvent was changed to DMSO， the 1∶1 binding mode of cage 9 with Cl-didn′t change， but the binding affinity dropped drastically， likely because of the solvation of guest anions. Besides Cl-， Br-and I-， anions such as HCO3-，AcO-and NO3-also bound to 9 in 1∶1 binding mode in CDCl3. In addition， cage 9 was found having very low affinity to anions such as ReO4-and BF4-. In all these cases， fast anion-exchange kinetics was observed during the titration.

Tab.2 Binding modes and constants （K） for cage 9 with various anions in CDCl3 or DMSO-d6表2 分子笼9在氘代氯仿和DMSO中与各阴离子的络合模式与络合常数（K）

In contrast to the gradual shifts seen in titrations with the above anions， when cage 9 was titrated with TBAF in CDCl3， a new set of signals emerged along with the original set of signals corresponding to anionfree form of cage 9（ cf. ESI）， displaying characters of slow anion-exchange kinetics. The original set of signals disappeared when approximate 1 equiv. of TBAF was added to the solution， indicating the emerged set of signals were corresponding to anionbound complex of 9. Fitting the titration data using BindFit v5.0 suggested there might be several plausible binding modes （1∶1 or 2∶1）. Although the score of 2∶1 was higher than 1∶1 binding， Job′s plot studies supported the 1∶1 binding mode. However，the corresponding binding constant was too small to be determined using BindFit v5.0.

Slow anion-exchange kinetics was also observed during titration when SO42-or H2PO4-were used as guest anion in either CDCl3or DMSO-d6. Job′s plots， again，supported 1∶1 binding mode， but simulation of the titrations using BindFit v5.0 didn′t provide any significant binding constants. When the solvent was switched to DMSO-d6， the binding constants were determined as 303.6 and 404.9， respectively.Attempts to grow single crystals of the complexes were unfortunately unsuccessful.

3 Conclusion

In summary， a novel amide-amine hybridized cage 9 was successfully synthesized. Its structure was characterized using NMR and MS spectroscopy and confirmed by X-ray single crystallography. Since the cage was constructed by irreversible amide and amine bonds， it displayed excellent stability towards acidic or basic conditions. This molecular cage， similar to the amide-imine hybridized cage 6， bind anions in 1∶1 mode. Fast anion-exchange kinetics was observed in the cases of halides， HCO3-and NO3-while slow anionexchange kinetics predominated when F-， SO42-and H2PO4-were used.