A CoII-Hydroxide Complex That Converts Directly to a CoII-Acetamide during Catalytic Nitrile Hydration

In exploring structural and functional mimics of nitrile hydratases, we report the synthesis of the pseudo-trigonal bipyramidal CoII complexes (K)[CoII(DMF)(LPh)] (1(DMF)), (NMe4)2[CoII(OAc)(LPh)] (1(OAc)), and (NMe4)2[CoII(OH)(LPh)] (1(OH)) (LPh = 2,2′,2’’-nitrilo-tris-(N-phenylacetamide; DMF = N,N-dimethylformamide; –OAc = acetate)). The complexes were characterized using NMR, FT-IR, ESI-MS, electronic absorption spectroscopy, and X-ray crystallography, showing the LPh ligand to bind in a tetradentate tripodal fashion alongside the respective ancillary donor. One of the complexes, 1(OH), is an unusual structural and functional mimic of the Co active site in Co nitrile hydratases. 1(OH) reacted with acetonitrile to yield the CoII-acetamide complex (NMe4)2[CoII(NHC(O)CH3)(LPh)], 2, which was also thoroughly characterized. In the presence of excess hydroxide, 1(OH) was found to catalyze quantitative conversion of the added hydroxide into acetamide. Despite the differences in Co oxidation state in nitrile hydratases and 1(OH) (CoIII versus CoII, respectively), 1(OH) was nonetheless an effective nitrile hydration catalyst, selectively producing acetamide over multiple turnovers.


■ INTRODUCTION
−3 Nitriles are generally difficult to hydrate in the absence of a catalyst and the hydration reaction is troublesome to control, conversion to carboxylic acid rather than amide being often observed.NHase, with their ability to selectively convert nitrile to acetamide, have found industrial applications for the preparation of nicotinamide (among others), one of very few examples of an industrial application of metalloenzymes. 4,5The active site of NHase contains either a mononuclear Fe or Co ion coordinated by two carboxamidate N-donors and multiple cysteinate or oxygenated cysteine (sulfinate or sulfenate) ligands. 3Interestingly, the oxidation state assigned to the Fe and Co ions in NHase was always +3, suggesting that nature has evolved to employ a relatively electron-poor metal catalyst to mediate nitrile activation.
Synthetic models that mimic both the structure and function of Co III NHase have been reported, 3,6−10 although they tend to be poor catalysts, demonstrating limited efficacy in multiturnover nitrile hydration. 11,12−16 For example, the direct conversion of a mononuclear Co IIhydroxide adduct to Co II -acetmide has not been reported, to the best of our understanding.The examples of mononuclear Co II catalysts capable of nitrile hydration provide limited mechanistic insight, nor do they demonstrate the role of metalbound ligands.
Two general mechanistic postulates defining the role of the metal ion in NHase have been made: 6 (a) the metal would act as a Lewis acid to activate a metal-bound nitrile to attack by hydroxide; (b) the metal would act to activate a metal-bound hydroxide ligand for nucleophilic attack on a nitrile group.Further nuance is likely: both hydroxide and nitrile could be metal-bound.Unfortunately, a consensus has not been reached, with synthetic models containing metal-bound nitrile and hydroxide ligands both demonstrating the ability to hydrate nitriles.Consideration of the oxidation state of the Co ion would favor mechanism (a), where a more Lewis acidic Co III may render a metal-bound nitrile highly active.In contrast, a Co II ion would be considered more Lewis basic and thus may favor nucleophilic attack by a metal-bound hydroxide ligand (thus mechanism (b)).Herein, we report the preparation of a Co II -hydroxide complex supported by a polycarboxamidate ligand that was capable of on-complex catalytic nitrile-to-acetamide conversion.The Co II −OH complex was not only a reasonable structural and functional mimic of NHase but also an effective catalyst.

■ RESULTS AND DISCUSSION
Addition of KH to 2,2′,2″-nitrilo-tris-(N-phenylacetamide) (L Ph ) in N,N-dimethylformamide (DMF) resulted in the evolution of a gas (Scheme 1, see the Supporting Information for details).Treatment of the resultant (presumed but not isolated) K 1(DMF), 1(OAc), and 1(OH) were crystallized via layering of diethyl ether (Et 2 O) into concentrated CH 3 CN or DMF solutions of the complexes and allowing for slow crystallization at room temperature (see the Supporting Information for details).Single-crystal X-ray diffraction analysis of 1(DMF) revealed a pseudo-trigonal bipyramidal (TBP) coordination geometry (Figure 1 and Table S1 and Figures S1−S4), with L Ph acting as a tripodal trianionic tetradentate ligand and a single DMF molecule coordinated via an O-atom in the fifth ligand site (Co−O = 2.11(7) Å, Table 1).K + counterions interacted with the amidate carbonyl groups of the L Ph backbone, bridging four units of 1(DMF) in the unit cell.Structural analysis of 1(OAc) revealed the same coordination of L Ph with the Co II ion as in 1(DMF).We noted the displacement of the ancillary DMF ligand with an − OAc ligand and the presence of two  18 An apical coordination of a hydroxide O-atom was observed in the Scheme 1. Preparation of 1(DMF), 1(OAc), 1(OH), and 2 Figure 1.ORTEPs of 1(DMF), 1(OAc), and 1(OH).Plotted at the 50% probability level.Hydrogen atoms, with the exception of the hydroxide in 1(OH), solvent molecules, and counterions have been omitted for clarity.The hydroxide hydrogen atom in 1(OH) was placed in an idealized position.
structure of 1(OH) (Co−O = 1.94(7)−24 A CCDC search of Co−OH complexes containing four N-donors resulted in 14 hits, where the average Co−O distance was 1.97(8) Å (Figure S5). 25 We concluded that 1(OH) was a hydroxide complex and not an aqua adduct (supported by the fact that two + [NMe 4 ] cations were observed for each [Co II (OH)(L Ph )] unit).−33 For the set of three complexes, a narrow range for the geometry index was noted (τ 5 = 0.76−0.82),where 1.0 presents a perfect TBP geometry and 0.0 represents square based pyramidal geometry. 34The discrepancies are indicative of the minor steric influences of the different ancillary (5th) ligands, leading to slight distortions in the overall structures.Finally, the more basic − OH ligand yielded a shorter Co−O bond (1.94 Å, Table 1) than the less basic DMF and − OAc ligands (2.11 and 2.03 Å, respectively), indicating a stronger Co/O interaction in moving through from weak field DMF, stronger field − OAc, to strongest field − OH ligands.Finally, the tris-amidate equatorial coordination provided by L Ph provides a reasonably good structural mimic for the bisamidate mono-cysteinate equatorial binding to Co in the NHase active site. 3SI-MS spectra measured on solutions of 1(DMF), 1(OAc), and 1(OH) displayed the most prominent peak at m/z = 472.09([Co II (L Ph )] − , expected m/z = 472.0946),confirming the presence of the mono anionic [Co II (L Ph )] − core in each complex.A low-intensity peak found at m/z = 531.1072,(assigned to [Co(OAc)(L Ph )] − , expected m/z = 531.1079)was observed for 1(OAc) (Figure S6).With the exception of 1(OAc), the ancillary fifth ligand could not be identified in ESI-MS studies, presumably due to the hard ionization technique, which caused fragmentation of the complexes.All other peaks were assigned to free solvent (DMF, δ = 2.72, 2.90, and 7.97 ppm).The appearance of these signals in the diamagnetic (0−12 ppm) range and at the natural frequency of the free solvent suggests that some free DMF was in solution in 1(DMF).For 1(OAc), we observed no resonance that could be assigned to the CH 3 of a metal-bound − OAc group.We have made similar observations for [Ni II (OAc)(L Ph )] 2− , 18 as have others for [Fe II (OAc)-(L Ph )] 2− , 17 where the − OAc group could not be identified by 1 H NMR. Free − OAc could be identified at δ = 1.44 ppm, suggesting that some free − OAc was present.Interestingly, the electronic absorption spectra of 1(DMF) and 1(OAc) in DMSO were not the same (Figure S10).Thus, despite the similarities in their 1 H NMR spectra, the ancillary anionic fifth ligand has not been displaced in these complexes.For 1(OH), a peak for the OH signal could not be identified; however, the CH resonances displayed significant shifts compared to those obtained for 1(DMF) and 1(OAc) suggesting that the OH ligand was Co-bound in 1(OH) in solution.
A solution state magnetic moment was determined via the Evans method at 20 °C. 35The three complexes displayed effective magnetic moments in the range of μ eff = 3.3−4.0(Table 1), corresponding to complexes with three unpaired electrons associated with the Co II ion.Assuming a pseudo-TBP geometry was maintained in solution, this value is consistent with a high spin S = 3/2 configuration of a d 7   absorption features are known for Co II −OH complexes (λ = 700−900 nm), and have been assigned to d-d transitions typically displaying extinction coefficients (ε) of <50 mol L −1 cm −1 . 19,23This feature has been predicted to stem from a 4 A 1 → 4 E transition in S = 3/2 Co II complexes in C 3v symmetry (as identified here for 1(OH)).Interestingly, for the weaker-field axial fifth donors in 1(DMF) and 1(OAc), this band appears to be blue-shifted (Figure S10), consistent with observations made previously where the ligand field was found to influence this transition. 19,23Taken together, the data suggest that the solution structure is in agreement with the solid-state structures.
Fourier transform infrared (FT-IR) spectra of 1(DMF), 1(OAc), and 1(OH) showed the loss of ν N−H of the amide in L Ph (ν = 3237 cm −1 , Figure S11), 17 indicating a deprotonation and complexation of the ligand had occurred.1(DMF) displayed a peak at ν = 1663 cm −1 , tentatively assigned to ν C�O of bound DMF.We were tentative in the assignment because L Ph contains three C�O groups that potentially display vibrational modes in the same window.1(OAc) showed features at ν = 1560 cm −1 and ν = 1375 cm −1 (Δν = 185 cm) that were (again tentatively) assigned to ν a/as modes of bound − OAc.The observed difference in ν a/as was larger than that observed in free − OAc (Δν = 163 cm −1 ), suggesting a monodentate coordination mode of the − OAc anion, in agreement with our XRD analysis. 36For 1(OH), a peak at ν = 3503 cm −1 was assigned to the ν O−H and was consistent with what is typically observed for Co−OH complexes. 37FT-IR thus confirmed the XRD and NMR assignments of 1(DMF), 1(OAc), and 1(OH).
Steady-state cyclic voltammetry of 1(DMF), 1(OAc), and 1(OH) was performed in DMF at room temperature using [ n Bu 4 N][PF 6 ] as supporting electrolyte (Figure S12 and Table 1).With the exception of 1(OAc), the complexes displayed either fully or quasi-reversible redox events at ca. −0.3 V versus the ferrocene/ferrocenium couple.We have assigned these peaks to the Co II /Co III redox couple, based on the range of peak anodic potentials measured between −0.32 and −0.21 V in DMF, for a series of similar tripodal Co II complexes supported by tris-anionic donor ligands. 37Little difference in the Co II /Co III redox couple was observed despite exchanging neutral DMF for more basic (anionic) − OAc and − OH donors.Mn and Fe complexes supported by L Ph have displayed similarly unchanged potentials, despite changes to the electrondonating properties of the supporting ligand. 38The highest E ox was observed for 1(OH) suggesting the most electron deficient Co II center, while 1(OAc) would appear to contain the most electron-rich Co II site.However, the differences among the sets of three are minimal, suggesting little difference in their redox properties.
Reaction of 1(OH) with CH 3 CN.We were interested in exploring the efficacy of 1(OH) as a mimic for Co NHase.We monitored the reaction of 1(OH) with CH 3 CN by electronic absorption spectroscopy (Figures 3 and S13).Stirring of a freshly prepared CH 3 CN solution of 1(OH) resulted in the blue-shifting of the near-IR band assigned to 1(OH) and a concomitant increase of a new absorption feature at λ = 593 nm assigned to a new species, defined as 2. The loss of the near-IR band typical of Co II −OH complexes and reversion to an electronic absorption spectrum similar to 1(OAc), suggested that the new species contained a weaker-field ancillary donor ligand.The reaction was complete within 5 h (Figure S14).We observed isosbestic points at λ = 630, 670, and 775 nm, consistent with a clean conversion of 1(OH) to 2. This indicated that the formation of 2 resulted from a reaction between 1(OH) and CH 3 CN, leading us to conclude that indeed 1(OH) was reacting with the solvent (CH 3 CN) to yield a presumed hydrated product.
Notably, when we attempted to crystallize the green complex 1(OH) in CH 3 CN we observed the formation of a purple crystalline material over the course of 1 day (see the Supporting Information for details).This product was identified as acetamide complex 2 and was obtained as a dark purple solid in 41% yield (Scheme 1 and Figure 4).XRD analysis on 2 revealed a five-coordinate structure with a striking resemblance to that of the earlier reported family of complexes.The 1 H NMR spectrum of 2 appeared different from 1(OH), confirming a reaction had occurred (Figures 2 and  S15).Interestingly, the 1 H NMR spectrum of 2 was more similar to those measured for 1(DMF) and 1(OAc), consistent with the three complexes displaying similar structures.New peaks that could be attributed to the NH and CH 3 resonances of the newly formed, Co-bound acetamide ligand, were not identified, consistent with previous observations for 1(DMF),  Inorganic Chemistry 1(OAc), and 1(OH), where ancillary ligand's OH and CH 3 resonances could not be identified.Importantly, there are shifts in the resonances associated with 2 when compared to 1(DMF) 1(OAc), and 1(OH), suggesting that the acetamide ligand induces small changes in the 1 H NMR data, consistent with it being bound in solution.FT-IR analysis of 2 showed the absence of the ν O−H = 3503 cm −1 peak attributed to the O−H stretch in 1(OH), and the formation of a higher-energy feature at ν = 3238 cm −1 for 2 (Figure S11).We assigned this new feature to the ν N−H of the newly formed acetamide group, which was consistent with a previous report on Co-acetamide complexes (ν N−H ∼ 3410 cm −1 ). 16Postreaction ESI-MS analysis from the reaction of 1(OH) with CH 3 CN showed the formation of acetamide (Figure S16), confirming the release of free acetamide from the reaction.Thus, although a molecular ion peak for 2 was not observed, there is evidence of the formation of acetamide through ESI-MS.In summary, our NMR, FT-IR, and ESI-MS analyses support the assignment (from XRD measurements) of 2 as a Co II -acetamide complex derived from acetonitrile.
To explore the catalytic efficacy of the hydroxide complex 1(OH) in nitrile hydration, we reacted 1(OH) with a large excess of [NMe 4 ][OH] (20 equiv) in CH 3 CN.The solution was then stirred under an inert atmosphere for 48 h at room temperature (20 °C).Following an aqueous workup, free acetamide was detected at the end of the reaction by 1 H NMR analysis (Figure S17).The yield of acetamide was found to be quantitative (thus a turnover number, TON = 20; after 24 h, the TON = 13).Furthermore, when 100 equiv.[NMe 4 ][OH] was added to a CH 3 CN solution of 1(OH), we observed quantitative conversion to acetamide within 65 h (TON = 100, Figure S18).The hydration reaction was thus slow but always quantitative.Importantly, we found that under aerobic conditions, the catalyst was less stable and stopped turning over after 6 h.Under anaerobic conditions, 2 was always the product of the catalyzed reaction, whereas 2 was observed to decay under aerobic conditions and stopped turning over.Our general observation was that as long as air was excluded from the reaction mixture, all [NMe 4 ][OH] that was added was turned over into acetamide.Finally, we found that 1(OH) also reacted with benzonitrile and butyronitrile to yield products that displayed spectral properties similar to those of 2 (thus both substrates were hydrated to the corresponding acetamide, Figure S19).
1(OH) was a reasonable structural mimic for the Co active site in NHase in that it provided a tris-anionic first coordination sphere (three amidates in 1(OH) versus two amidate and a cysteinate in NHase).Unlike Co NHase, which contains a Co III ion, 1(OH) contained a Co II ion.The Co ion in 1(OH) was thus likely to be less Lewis acidic.Despite this, 1(OH) was a reasonable functional mimic of Co NHase, efficiently converting nitriles into metal-bound acetamides.We postulate that the metal-bound hydroxide ligand is rendered highly nucleophilic by the Co II ion and the tris-anionic first coordination sphere.Given that isosbestic points were observed in the conversion of 1(OH) to 2, we tentatively assume that the nitrile group that is attacked was bound, although it is equally plausible that nitrile was not coordinated to the Co ion.We note there is a vacant "6th" site on the Co II ion, where nitrile could bind.The current example demonstrates that from a mechanistic perspective, rendering metal-bound hydroxide ligands highly nucleophilic through a strongly donating trianionic supporting ligand and lower metal oxidation state will generate an effective nitrile hydration catalyst.Overall, 1(OH) represents an unexpected example of a mononuclear Co NHase structural and functional mimic, where, despite the lower Co oxidation state, effective nitrile hydration to acetamide was achieved with multiple turnovers.

■ CONCLUSIONS
A series of mononuclear Co II complexes, 1(DMF), 1(OAc), and 1(OH), were synthesized and characterized using X-ray diffraction, NMR, FT-IR, and electronic absorption spectroscopies.The complexes were supported by the tris-carboxamidate ligand L Ph , which bonded in a tripodal fashion yielding Co II ions in a pseudo-trigonal bipyramidal coordination environment.One of the complexes, 1(OH), was found to be a reasonable structural and functional mimic of the Co active site in Co NHase.1(OH) reacted with acetonitrile to yield Co II -acetamide complex 2, which was also thoroughly characterized.In the presence of excess hydroxide, 1(OH) was found to catalyze the quantitative conversion of added hydroxide with acetonitrile into acetamide.Despite the differences in Co oxidation state in NHase and 1(OH) (Co III versus Co II , respectively), 1(OH) was nonetheless an effective nitrile hydration catalyst, selectively producing multiple turnovers of acetamide in acetonitrile hydration.

1 H
nuclear magnetic resonance (NMR) spectra of 1(DMF), 1(OAc), and 1(OH) displayed four broadly shifted peaks in the range of −20 to +70 ppm for each complex (Figures 2 and S7−S9).Integration of the peaks gave an approximate 2:2:2:1 ratio corresponding to the H atoms in L Ph assigned to the CH 2 , and ortho-, meta-, and para-CH arene positions, respectively.Sharp upfield resonances at δ = ca.−5 to −15 ppm were assigned to the para-CH of the phenyl ring based on their integration value.Broader signals at δ ∼ 0 to 25 ppm and sharper resonances at +10 to 20 ppm were assigned to either the ortho-CH or meta-CH positions.The most downfield shifted signals at δ ∼45 to 70 ppm were assigned to the CH 2 signal of the methylene bridge on L Ph .This assignment was made based on the assignments of CH signals in [Ni II (L Ph )-(OAc)] 2− and its deuterated analogue [Ni II (D 15 -L Ph )(OAc)] 2− (D 15 -L Ph = 2,2′,2″-nitrilo-tris(N-2,3,4,5,6-[D]-phenyl)acetamide) where a per-deuterated form of L Ph was used to identify the resonances.

a
Co II ion.The unpaired electrons are expected to be located in metal-type d z2 and degenerate d x2-y2 /d xy orbitals.1(OH) displayed an electronic absorption feature at λ = 790 nm.Such near-IR Table 1.Spectral, Magnetic, and Structural Properties of 1(DMF), 1(OAc), 1(OH), and 2 λ max, nm (ε, mol L −1 cm −1 ) μ eff (Bohr magneton) DMF was solvent.b CH 3 CN was solvent.c E pa = peak anodic potential, used because the CV for 1(OAc) appeared irreversible.Calibrated against the ferrocene/ferrocenium (Fc/Fc + ) couple measured under the same conditions.d As determined by X-ray diffraction measurements.Corresponding to O/N atom of ancillary 5th ligand, not L Ph .
The axial fifth ligand was identified as an N-bound acetamide ( − N ( H ) C ( O ) C H 3 ) , l e a d i n g u s t o d e fi n e 2 a s (NMe 4 ) 2 [Co II (N(H)C(O)CH 3 )(L Ph )] 2− .

Figure 3 .
Figure 3. Conversion of 1(OH) (green trace, 5.0 mM in CH 3 CN) to 2 (blue trace) which formed over the course of 5 h at 20 °C.Gray traces show incremental changes in the electronic absorption properties.

Figure 4 .
Figure 4. ORTEP of compound 2. Plotted at 50% probability level.Hydrogen atoms (apart from the acetamide NH), solvent molecules, and counterions have been omitted for clarity.The acetamide hydrogen atom was placed in an idealized position.