Varying the secondary coordination sphere: synthesis of cobalt and iron complexes of a tripodal ligand featuring two hydrogen-bond donors or acceptors

Abstract In order to further understand the role of the secondary coordination sphere in biomimetic systems, a series of cobalt(II) and iron(II) complexes bearing the non-heme bis(5-cyclohexyliminopyrrol-2-ylmethyl)-2-pyridylmethylamine (PyN(piCy)2) ligand were synthesized. This platform, based on the tris(5-cyclohexylimino-pyrrol-2-ylmethyl)amine (N(piCy)3) platform, reduces the number of possible hydrogen-bonding interactions from three (in N(piCy)3) to two, while maintaining the ability of the ligand to datively coordinate to a metal center. The secondary coordination spheres of the family of dative and anionic cobalt(II) and iron(II) complexes were characterized in the solid-state using X-ray crystallography and IR spectroscopy. Moreover, oxidation of an iron(II) complex was explored to compare its reactivity to that of the analogous iron(II) complex of the N(piCy)3 platform. GRAPHICAL ABSTRACT


Introduction
In biological systems, the function of a metalloenzyme is determined by both the primary and secondary coordination spheres of its active site [1,2]. The primary coordination sphere influences the electronic properties of the metal center, whereas the secondary coordination sphere aids with substrate selectivity [1]. Synthetic systems can readily mimic the primary coordination sphere of many metalloenzymes, but studies in which a secondary coordination sphere is incorporated in biomimetic systems are less common due to the difficulty in controlling intramolecular hydrogen-bonding interactions [2][3][4][5]. There are even fewer reports of the effects of varying the secondary coordination sphere on the reactivity of a metal center [6][7][8][9][10].
Pioneering work by Borovik and coworkers described the incorporation of hydrogen-bond donors into a rigid tripodal ligand framework, allowing for isolation of the first terminal iron(III)-oxo complex [2,11]. Additional reports by other groups have developed the utility of a secondary coordination sphere in rigid tripodal platforms [12]. The influence of the secondary coordination sphere on reactivity was further detailed in a later study by the Borovik group that systematically reduced the number of hydrogen-bond donors present in a series of cobalt complexes from three to zero [6]. The electronic structure of the cobalt(II) centers was maintained through substitution of urea functionalities with carboxamide moieties. Upon addition of dioxygen to generate cobalt(III)-hydroxo complexes, differences in the reactivity of the series were observed. When two or three hydrogen-bond donors were present, the complexes readily reacted with 0.5 equiv dioxygen to furnish the corresponding cobalt(III)hydroxo species. If only one hydrogen-bond donor was present, an excess of dioxygen was required to synthesize the cobalt(III) product, and when no hydrogen-bond donors were present, no reaction was observed. Thus, the number of hydrogen-bond donors present in the secondary coordination sphere had explicit involvement in the dioxygen activation at the cobalt(II) centers.
A similar systematic approach was taken in developing a series of urea/sulfonamido hybrid ligands, which contained a mixture of hydrogen-bond donors and acceptors in the secondary sphere [8]. Comparing the resulting series of cobalt(II)-hydroxo complexes with varied secondary coordination spheres revealed differences in the Co-O bond length and the strength of the hydrogen-bonding interaction between the bound hydroxo and the sulfonamido moiety of the ligand, as measured by IR spectroscopy. However, the variation in the ligand field strength of the urea versus the sulfonamide groups changed the electronic structure of the cobalt(II) centers across the series and made discerning the effects of the modulation of the secondary coordination sphere unattainable.
Because of the importance of the secondary coordination sphere in imparting functionality to metalloenzymes, our group also set out to delineate the influence of varying the primary and secondary coordination spheres on the reactivity of a metal center. The primary coordination sphere of our tripodal non-heme platform, tris(5cyclohexylimino-pyrrol-2-ylmethyl)amine (N(pi Cy ) 3 ), was brominated at the 3 0 position of the pyrrole backbone, resulting in an increased iron(II/III) redox couple [9,13]. In addition, exchanging the capping group from a cyclohexyl moiety to an electron-withdrawing phenyl moiety also increased the iron(II/III) redox couple, and the effects were additive, as observed in the case of the brominated ligand with a phenyl capping group (Figure 1). These studies, however, maintained the number of hydrogenbond donors or acceptors in the secondary coordination sphere.
The tripodal platform, bis(5-cyclohexyliminopyrrol-2-ylmethyl)-2-pyridylmethylamine ( Py N(pi Cy ) 2 ), was designed to vary the number of hydrogen-bond donors or acceptors present in the secondary coordination sphere. By installing an equatorial pyridine donor in place of one pyrrole-imine arm of the N(pi Cy ) 3 system, trigonal bipyramidal geometry would be maintained about a bound metal center and neutral coordination of the ligand platform could still be achieved. This ligand design would thus allow for direct comparisons between metal complexes of the Py N(pi Cy ) 2 and N(pi Cy ) 3 ligand platforms (Figure 2), providing insight into the structure-function relationship of the secondary coordination sphere in our system. This report details the synthesis of the Py N(pi Cy ) 2 ligand and a family of cobalt(II) and iron(II) complexes, featuring both anionic (pyrrole-imine, abbreviated pi) and neutral (azafulvene-amine, abbreviated afa) coordination to the ligand. The secondary coordination sphere of these complexes was characterized in the solid-state using Xray crystallography and IR spectroscopy. Finally, the oxidation of an iron(II) complex was explored for comparison to the analogous complex of the previously reported (N(pi Cy ) 3 ) platform [13][14][15].

Results and discussion
Synthesis of Py N(pi Cy ) 2 began with the reductive amination of 2 equiv of pyrrole-2-carboxaldehyde and 1 equiv of 2-picolylamine with sodium triacetoxyborohydride to form bis(pyrrol-2-ylmethyl)-2-pyridylmethylamine. We followed the modular approach previously established in our laboratory to generate the pyrrole-imine functionality; a Vilsmeier-Haack reaction resulted in 5 0 formylation of the pyrrolyl groups and subsequent condensation of cyclohexylamine furnished the ligand [9,13]. Analysis of the Py N(pi Cy ) 2    successful installation of the secondary imine and a C ¼ N stretching frequency of 1630 cm À1 was measured by IR spectroscopy. As with the N(pi Cy ) 3 ligand and its derivatives, the isolated Py N(pi Cy ) 2 ligand was hygroscopic, but could be made anhydrous by the dissolution of the ligand in diethyl ether over 4 Å molecular sieves [9,13].
Following the synthesis of Py N(pi Cy ) 2 , we explored its coordination to cobalt(II) salts (Scheme 1). Emulating the anionic coordination of N(pi Cy ) 3 [13], the two pyrrolyl moieties of the ligand were deprotonated by the addition of 2.1 equiv of KH to a THF solution of Py N(pi Cy ) 2 . After stirring for 2 h, the mixture was filtered into a slurry of CoCl 2 in acetonitrile, generating a deep-red solution. Solid-state structural characterization of the product revealed a cobalt(II)-aqua complex in trigonal bipyramidal geometry, Py N(afa Cy ) 2 Co(OH 2 ) ( Py Co-OH 2 ; Figure 3), with the aqua ligand bound trans to the apical nitrogen of the ligand platform. The electron density of the aqua hydrogen atoms was located in the difference map of the refined data, and their positions refined independently of their carrier atoms. Each was within hydrogen-bonding  We also sought to characterize cobalt(II) complexes that exhibited neutral coordination to the ligand platform. Upon addition of Co(OTf) 2 (MeCN) 2 to wet ligand (i.e. ligand that has not been dried over 4 Å molecular sieves), two species were formed: a cobalt(II)-triflate species, Py N(afa Cy ) 2 Co(OTf) 2 ( Py Co-OTf), and a cobalt(II)-hydroxo species, [ Py N(afa Cy ) 2 CoOH]OTf ( Py Co-OH). Crystals of Py Co-OH that were suitable for X-ray diffraction were grown from the mixture of products. Refinement of the data revealed a cobalt(II) center in trigonal bipyramidal geometry ( Figure 3). The hydroxo moiety was engaged in hydrogen-bonding with both amines in the secondary coordination sphere and the outer-sphere triflate anion, which was confirmed through the distances between the hydrogen-bond donors and acceptors and the nearly linear angles of the three atoms involved in each hydrogen-bond [16]. The O1-O2 distance was 2.917(2) Å and the O1-H1-O2 angle was 171(3) ; the N5-O1 and N6-O1 distances were 2.614(2) Å and 2.706(2) Å, respectively, and the N5-H5-O1 and N6-H6-O1 angles were 175(3) and 176(3) , respectively. The Co1-O1 bond distance, 1.9926(15) Å (Table 1), was within the range of other structurally characterized five-coordinate cobalt(II)-hydroxo complexes (1.938(2) À 2.053(3) Å) [18]. Analysis of Py Co-OH by solid-state IR spectroscopy revealed an O-H stretch at 3503 cm À1 and a C ¼ N stretch at 1659 cm À1 . The shortened Co1-O1 bond distance compared to Py Co-OH 2 and the blue-shift in the C ¼ N stretching frequency observed by IR spectroscopy were consistent with coordination of an anionic hydroxide ligand and neutral coordination of the ligand platform to the cobalt(II) center [13].
Both Py Co-OTf and Py Co-OH were independently synthesized (Scheme 1). When dried ligand was added to Co(OTf) 2 (MeCN) 2 , Py Co-OTf was the sole product formed. Crystals suitable for X-ray diffraction were unable to be grown; however, the analogous manganese(II)-triflate species, Py N(afa Cy ) 2 Mn(OTf) 2 ( Py Mn-OTf), was characterized crystallographically ( Figure 4). The refined data featured an octahedral manganese(II) center, with axial coordination of a triflate anion. A THF solvate was bound trans to the pyridyl nitrogen, leading to trans coordination of the azafulvene nitrogen atoms. Based on the elemental analysis of Py Co-OTf, which included a THF solvate, we proposed an analogous coordination environment about the cobalt(II) center. Py Co-OH was synthesized from the reaction of Py N(pi Cy ) 2 , Co(OTf) 2 (MeCN) 2 , and excess Li 2 O in THF overnight. The reaction initially turned deep red, indicative of the formation of Py Co-OTf, but gradually turned green as the red species was converted to Py Co-OH.
Py Fe-OH was synthesized from both the addition of wet ligand to Fe(OTf) 2 (MeCN) 2 and the reaction of dry ligand, Fe(OTf) 2 (MeCN) 2 , and excess Li 2 O. Crystals suitable for X-ray diffraction were grown from a saturated solution of THF layered with pentane. The structure contained a trigonal bipyramidal iron(II)-hydroxo with an outer-sphere triflate anion. The hydrogen atoms of the hydroxo and amine moieties were located in the difference map. The O-H bond was restrained to 0.84 Å (esd 0.01), and one N-H bond was restrained to be 0.88 Å (esd 0.01). Hydrogen-bonding interactions between the amines of the ligand, the hydroxo moiety, and the outer-sphere triflate were observed. The O1-O2 distance was 2.950(4) Å and the O1-H1-O2 angle was 170(4) . The N5-O1 and N6-O1 distances were 2.645(4) Å and 2.734(4) Å, respectively, and the N5-H5-O1 and N6-H6-O1 angles were 176(4) and 177(4) , respectively ( Table 1). The Fe1-O1 bond was 1.978(2) Å, which was in the range of reported iron(II)-hydroxo . This value is within error of the azafulvene-amine C ¼ N stretch of Fe II -OH (1655 cm À1 ) [13]. Based on the structural similarities between Py Fe-OH and Fe II -OH, the Py N(pi Cy ) 2 ligand appropriately modeled the coordination environment provided by the N(pi Cy ) 3 ligand while removing one hydrogen-bonding interaction from the secondary coordination sphere.
Py Fe-OTf was furnished from the addition of Fe(OTf) 2 (MeCN) 2 to a THF solution of dry ligand. Crystals suitable for X-ray diffraction were unable to be grown of the pale yellow product, precluding detailed structural comparisons between Py Fe-OTf and the analogous species of the N(pi Cy ) 3 system: N(afa Cy ) 3 Fe(OTf) 2 (Fe II -OTf). However, as was the case with Py Co-OTf, elemental analysis suggested the formulation of the iron(II)-triflate product contained a THF solvate. The binding of both arms of the ligand in the azafulvene-amine tautomeric form was confirmed in the solid-state IR spectrum, where a C ¼ N stretch was observed at 1638 cm À1 [13]. We were eager to examine the oxidative reactivity of the iron(II) complex Py Fe-OTf in order to compare it to the well-established reactivity of Fe II -OTf [14]. The addition of nitrite to a solution of Py Fe-OTf resulted in an immediate color change from yellow to brown.  Figure 5). The hydrogen atoms of the secondary coordination sphere were not located within the difference map; however, the distances between the amine nitrogen atoms of the ligand platform and the hydroxo oxygen atoms, which ranged from 2.672(7) Å to 2.937(11) Å, placed them within the range of hydrogen-bonding interactions [15]. In addition, the N-H-O angles ranged from 166. 3  The solid-state IR spectrum of the product contained an O-H stretch at 3554 cm À1 and a C ¼ N stretch at 1651 cm À1 , confirming the formulation of the product as containing a hydroxo moiety and possessing solely the azafulvene-amine tautomer of the ligand [13]. We proposed that adventitious water provided the additional oxygen atom needed for the formation of the oxo-bridged product. Moreover, the propensity of Py Fe-OTf to form the dimeric iron(III) product represents a significant limitation to this system. The decreased steric profile of the pyridyl moiety compared to that of the azafulvene-amine arms of the ligand could have contributed to the synthesis of the dimer instead of a monomeric iron complex. The reduction of the number of hydrogen-bond donors from three in the N(pi Cy ) 3 ligand to two in the Py N(pi Cy ) 2 ligand may have also contributed to the inability to observe a monomeric iron(III) product.

Conclusion
The Py N(pi Cy ) 2 ligand platform was synthesized, along with its neutral and anionic ligand coordination to cobalt(II) and iron(II) complexes. The secondary coordination sphere of each species was characterized through X-ray crystallography and IR spectroscopy and demonstrated the propensity of the ligand platform to engage in hydrogen-bonding interactions. Comparison of the iron(II)-hydroxo complex to that of the N(pi Cy ) 3 system revealed high structural fidelity between the two platforms while reducing the number of possible hydrogen-bonding interactions in the secondary coordination sphere from three to two. This reduction in hydrogen-bonding, however, coincided with a decrease in the steric profile of the ligand. The combination of these two effects manifested in the formation of a homodimeric product with a bridging oxo ligand upon oxidation of the iron(II)triflate species. Thus, the observed reactivity could not be assigned solely to the reduction in the number of hydrogen-bonding moieties present in the secondary coordination sphere or the reduction in the steric profile of the altered arm of the ligand platform.

General considerations
To avoid contact with oxygen and water, all air-and moisture-sensitive manipulations were carried out under an atmosphere of dinitrogen in an MBraun inert atmosphere drybox or using standard Schlenk techniques. Solvents for air-and moisture-sensitive manipulations were dried and deoxygenated using a Glass Contour System and stored over 4 Å molecular sieves prior to use. Celite 545 was heated to 150 C under dynamic vacuum for 24 h prior to use in the drybox. All reagents were purchased from commercial sources and used as received unless otherwise noted. Ferrous trifluoromethanesulfonate [22] and cobaltous trifluoromethanesulfonate [23] were synthesized according to literature procedure. NMR solvents (acetonitrile-d 3 , tetrahydrofuran-d 8 and chloroform-d 1 ) were degassed and stored over 4 Å molecular sieves prior to use. NMR spectra were recorded at ambient temperature on a Varian spectrometer operating at 500 MHz ( 1 H-NMR), 126 MHz ( 13 C-NMR), and 470 MHz ( 19 F-NMR) and referenced to the peak of the residual solvent (d parts per million and J in Hz). Solid-state infrared spectra were measured using a Perkin Elmer Frontier FTIR spectrophotometer equipped with a KRS5 thallium bromide/iodide universal attenuated total reflectance accessory. Electrospray ionization mass spectrometry (ESI) was performed using a Waters Q-TOF Ultima ESI instrument. Elemental analyses were performed by the University of Illinois at Urbana-Champaign School of Chemical Sciences Microanalysis Laboratory in Urbana, IL. Solid-state structural characterization was completed using a Bruker D8 Venture Duo or Bruker X8ApexII diffractometer at the George L. Clark X-Ray Facility and 3M Material Laboratory at the University of Illinois at Urbana-Champaign.

Preparation of dipyrrolylpicolylamine (dppa)
In a 500-mL round-bottom flask under an N 2 atmosphere, pyrrole-2-carboxaldehyde (4.00 g, 0.0526 mol) and picolylamine (2.84 g, 0.0263 mol) were combined in 50 mL 1,2dichloroethane and cooled to 0 C. Following dissolution of the pyrrole-2-carboxaldehyde, sodium triacetoxyborohydride (17.83 g, 0.0841 mol) was added slowly to the reaction mixture and stirring was continued for 1.5 h. The resulting bright orange solution was quenched with aqueous sodium bicarbonate and vigorously stirred until effervescence subsided. The product was extracted into diethyl ether, dried over Na 2 SO 4 , and evaporated to dryness. The pale orange solid was dissolved in dichloromethane and hexane was added until a red oil was formed. The colorless supernatant was decanted and dried in vacuo, yielding the product as a beige solid (4.30 g, 0.0162 mol, 62% yield). 1

Preparation of dppa CO
A 500-mL three-neck round-bottom flask was charged with dimethylformamide (3.5 mL, 0.0451 mol) and cooled to 0 C in an ice bath under N 2 . POCl 3 (3.5 mL, 0.0376 mol) was diluted in 15 mL 1,2-dichloroethane and added dropwise to the dimethylformamide via an addition funnel for 30 min with vigorous stirring. The solution was removed from the ice bath and stirred at room temperature for 15 min. Once the reaction mixture was cooled back to 0 C, the addition funnel was washed with 3 mL 1,2-dichloroethane. Dipyrrolylpicolylamine (4.00 g, 0.0150 mol) was dissolved in 20 mL 1,2-dichloroethane and 3 mL dimethylformamide, added to the addition funnel, and added dropwise for 30 min to the reaction mixture. The resulting dark red solution was stirred at room temperature for 1 h. A saturated aqueous solution of NaOAc (15.40 g, 0.188 mol) was added and the cloudy red-brown solution was heated to 45 C for 30 min before neutralization by slow addition of Na 2 CO 3 . The product was extracted into dichloromethane (3 Â 25 mL) and dried over Na 2 SO 4 . Removal of volatiles under reduced pressure yielded a red oil, which was treated with acetonitrile and diethyl ether (1:2) to precipitate the product dppa CO (1.58 g, 0.00490 mol, 32.7%). 1

Preparation of Py N(pi Cy ) 2
Dppa CO (1.10 g, 0.00334 mol) was added to 20 mL acetonitrile in a 250-mL round-bottom flask. Cyclohexylamine (0.84 g, 0.00845 mol) was added and stirring was continued for 8 h. The product precipitated from solution as a beige solid and was isolated by filtration. The filtrate was cooled to -10 C for 3 h, after which more product had crystallized from the solution. The isolated material was dissolved in 10 mL diethyl ether and dried over 4 Å molecular sieves under an atmosphere of dinitrogen. Following filtration and removal of volatiles under reduced pressure, the product was obtained as a white powder (1.19 g, 0.00246 mol, 74% yield). 1

Py N(pi Cy ) 2 Co(OH 2 ) ( Py Co-OH 2 )
The red-purple solution was dried in vacuo. Diethyl ether was added to the resulting residue and the red solution of the product was filtered over a pad of Celite and dried (0.0495 g, 0.0885 mmol, 86% yield). Crystals suitable for X-ray analysis were grown from a saturated solution of acetonitrile. Analysis for C 30    ] (0.0103 g, 0.0358 mmol) was weighed by difference and added to the stirring solution, which turned red-brown in color. The solution was stirred for 2 h and volatiles were removed under reduced pressure, yielding a brown solid. The product was recrystallized from the vapor diffusion of diethyl ether into a saturated THF solution (0.0142 g, 0.00993 mmol, 56% crystalline yield). Crystals suitable for X-ray diffraction were grown under similar conditions. Analysis for C 62 H 82 N 12 F 6 Fe 2 O 9 S 2 : Calcd C, 52.10; H, 5.78; N, 11.76. Found C, 52.08; H, 5.63; N, 11.81%. IR ¼ 3554 cm -1 (O-H); 1651 cm À1 (C ¼ N).

Disclosure statement
There are no conflicts to declare.