Evaluating Metal-ligand Interactions of Metal-binding Isosteres Using Model Complexes

Abstract

Bioisosteres are a useful approach to address pharmacokinetic liabilities and improve drug-like properties. Specific to developing metalloenzyme inhibitors, metal-binding pharmacophores (MBPs) have been combined with bioisosteres, to produce metal-binding isosteres (MBIs) as alternative scaffolds used in fragment-based drug discovery (FBDD). Picolinic acid MBIs have been reported and evaluated for their metal-binding ability, pharmacokinetic properties, and enzyme inhibitory activity. However, their structural, electronic, and spectroscopic properties with metal ions other than Zn(II) have not been reported, which might reveal similarities and differences between MBIs and the parent MBPs. To this end, [M(TPA)(MBI)] (M = Ni(II) and Co(II), TPA = tris(2-pyridylmethyl)amine) is presented as a bioinorganic model system for investigating picolinic acid, four heterocyclic MBIs, and 2,2’-bipyridine. These complexes were characterized by X-ray crystallography, NMR, IR, and UV−vis spectroscopy, and their magnetic moments were accessed. In addition, [(Tp^Ph,Me)Co(MBI)] (Tp^Ph,Me = hydrotris(3,5-phenylmethylpyrazolyl)borate) was used as a second model compound and the limitations and attributes of the two model systems are discussed. These results demonstrate that bioinorganic model complexes are versatile tools for metalloenzyme inhibitor design and can provide insights into the broader use of MBIs.

Graphical Abstract

INTRODUCTION

Metalloenzymes make up approximately 40% of all enzymes.¹ Their ubiquity makes metalloenzymes important targets for therapeutics. Despite their prevalence, fewer than 5% of small molecule drugs target a metalloenzyme.¹ A barrier for developing metalloenzyme inhibitors has been the longstanding reliance on a limited number of metal-binding pharmacophores (MBPs) to bind to the active site metal ion(s) in these targets. The hydroxamic acid functional group is an archetypal MBP and it has been extensively employed to inhibit a wide range of metalloenzymes.² Issues surrounding hydroxamic acids and other commonplace MBPs include pharmacokinetic concerns such as metabolic instability and poor membrane permeability, as well as limited chemical diversity and target selectivity.²

To expand the available chemical space for targeting metalloenzymes, a diversified MBP library can be utilized for fragment-based drug discovery (FBDD).³ A previously reported MBP library, consisting of ~300 low molecular weight fragments, has been used to generate active and selective hits against a variety of metalloenzyme targets.¹ This library has been gradually expanded to include a wider range of MBPs,¹ and recently metal-binding isosteres (MBIs) have been added as a new strategy to further address the pharmacokinetic liabilities of MBPs.⁴ MBIs combine the metal-binding features of MBPs with the principles of isosteres/bioisosteres to produce ligands with even greater chemical diversity and potentially improved biological and drug-like properties. Two prior studies have examined MBIs derived from the picolinic acid MBP, examining carboxylic acid isosteres⁴ and pyridine ring isosteres of this picolinic acid.⁵ These studies demonstrated the metal-binding ability of MBIs and showed that MBIs can have a pronounced influence on physicochemical properties of these ligands while retaining inhibitory activity against metalloenzymes. These prior results indicate that isosteric replacement combined with metal-binding functional groups can be a valuable approach to improving the drug-likeness of metalloenzyme small molecule inhibitors.

Bioinorganic model complexes have been used to validate the metal-binding properties of MBPs and MBIs.^{4, 6} Model complexes provide a facile strategy to elucidate detailed metal-ligand interactions at an atomic level that can be labor intensive to achieve using more advanced biophysical methods, such as protein crystallography or nuclear magnetic resonance (NMR) spectroscopy. Therefore, model complexes are advantageous in the early stages of metalloenzyme inhibitor development to elucidate the metal-binding capability of different MBPs/MBIs.⁴ Furthermore, they provide synthetic accessibility and flexibility that enables sophisticated structural modifications to mimic the target enzyme.^7, 8 For example, [(Tp^Ph,Me)Zn(L)] (Tp^Ph,Me = hydrotris(5,3-methylphenylpyrazolyl)borate, L = metal-binding ligand) and its derivatives have been used to model the active sites of several metalloproteins including carbonic anhydrase,⁹ liver alcohol dehydrogenases,¹⁰ and matrix metalloproteinases (Figure 1).^{11, 12} Despite the broad use of the [(Tp^Ph,Me)Zn(L)] model system, few studies have looked at other spectator ligands (i.e., ligands other than Tp^Ph,Me) or metal ions beyond Zn(II) to examine MBP or MBI binding and characteristics.¹³ Alternative bioinorganic model complexes could prove helpful to understand both the structural and electronic features of MBIs.

Figure 1.

a) Picolinic acid (pic), MBIs of picolinic acid, and 2,2’-bipyridine (bpy). b) Chemical structure of the [M(TPA)(MBI)]⁺ model system (TPA = tris(2-pyridylmethyl)amine). c) Chemical structure of the [(Tp^Ph,Me)Zn(MBI)] model system (Tp^Ph,Me = hydrotris(5,3-methylphenylpyrazolyl)borate).

Herein, the model complex, [M(TPA)(MBI)] (M = Ni²⁺ and Co²⁺, TPA = tris(2-pyridylmethyl)amine) has been employed to investigate the coordination chemistry and spectroscopic characteristics of MBIs (Figure 1). TPA has been used as a tetradentate ligand for biologically relevant first-row transition metal ions, while providing flexible structural modification via the pyridine moieties.^14–23 TPA complexes of Ni(II) have been used to study Ni(II)-dependent metalloenzymes such as acireductone dioxygenase,²³ urease,²⁰ and glyoxalase I.²² In the present study, Ni(II) and Co(II) complexes were used to enable structural and spectroscopic study of MBI-metal interactions. Picolinic acid (pic), several picolinic acid MBIs, and 2,2’-bipyridine (bpy) were examined (Figure 1). Bpy was included as an additional comparative ligand because it can also be regarded as an isostere of the heterocyclic MBIs studied here. The [M(TPA)(MBI)] complexes were prepared, and their reactivity, structural features, and spectroscopic properties were investigated by X-ray crystallography, NMR, IR, UV-Vis, and magnetic moment analysis. In addition, [(Tp^Ph,Me)Co(MBI)] complexes were similarly investigated as spectroscopically active analogues of the previously reported [(Tp^Ph,Me)Zn(MBI)] complexes.

EXPERIMENTAL METHODS

General Experimental Details.

All reagents and solvents were obtained from commercial sources (Sigma Aldrich, Alfa Aesar, TCI, Combi-Blocks etc.) and used without further purification. Tris((2-pyridylmethyl)amine) (TPA) was synthesized using a modified literature procedure.²⁴ [(Tp^Ph,Me)K] (Tp^Ph,Me = hydrotris(5,3-methylphenylpyrazolyl)borate) was prepared as previously reported.⁶ Picolinic acid and 2,2’-bipyridine were purchased from commercial suppliers. MBIs 1, 2, 3, and 4 have been previously reported and were prepared according to literature procedures.⁴ [Ni(TPA)(OAc)(H₂O)]PF₆ and [Co(TPA)(OAc)]BPh₄ were prepared using a modified literature procedure.^{19, 24} To obtain an X-ray structure of [Co(TPA)(pic)]⁺, it was necessary to use an alternative counterion, ClO₄⁻, because of its higher crystal quality and non-hygroscopic nature. [(Tp^Ph,Me)ZnOH] and [(Tp^Ph,Me)CoCl] was prepared according to literature methods,²⁵ and [(Tp^Ph,Me)Co(MBI)] and [(Tp^Ph,Me)Zn(bpy)] were prepared by a modification of the literature procedure.^{4, 13} Column chromatography was performed using a CombiFlash Rf automated system from Teledyne Isco using prepacked silica cartridges. All the solution-state characterization was carried out in DMSO to prevent any solubility issues. ¹H NMR spectra were recorded at ambient temperature on 300 MHz Bruker instrument and ¹³C NMR spectra were collected using 500 MHz Varian NMR instrument. Magnetic measurements were performed by solution ¹H NMR using the Evans’ method²⁶ on a 300 MHz Bruker NMR spectrometer for all complexes performed at room temperature in an open-air system. The measurements were performed in a standard 5 mm NMR tube containing the paramagnetic sample of 20 mM dissolved in 400 – 500 μL of DMSO-d₆ against a co-axial reference tube filled with the same solvent. Processing of the NMR data was performed using the MestReNova 14.2 program. UV-visible spectra were recorded with the samples of 5 mM dissolved in DMSO using a Perkin-Elmer Lambda 25 spectrophotometer. Absorbance maxima are given as λ_max (nm) with extinction coefficients reported (ε, M⁻¹ cm⁻¹). Infrared spectra were collected on a Bruker Alpha FT-IR instrument. Mass spectra were obtained at the Molecular Mass Spectrometry Facility (MMSF) in the Department of Chemistry and Biochemistry at the University of California, San Diego.

Synthesis and Characterization of Compounds.

2-(1H-tetrazol-5-yl)pyridine (1).

The compound was made following a previously reported procedure.⁴ Picolinonitrile (2.00 g, 19.2 mmol) was dissolved in 20 mL of DMF. Sodium azide (1.87 g, 28.8 mmol) and ammonium chloride (1.54 g, 28.8 mmol) were added to the DMF solution and the mixture was stirred under nitrogen at 110 °C for 3 h. The reaction mixture was concentrated down almost to dryness on a rotary evaporator. Water was added to the residue and the solution was acidified to pH 2.0 using 6 M HCl. The product precipitated as a solid and was collected by filtration. The collected white solid was washed with water and dried. Yield: 2.0 g (13.6 mmol, 71%). ¹H NMR (300 MHz, DMSO-d₆): δ 8.81 (ddd, J₁ = 4.8, J₂ = 1.7, J₃ = 0.9 Hz, 1H), 8.24 (dt, J₁ = 7.9, J₂ = 1.1 Hz, 1H), 8.09 (td, J₁ = 7.7, J₂ = 1.7 Hz, 1H), 7.64 (ddd, J₁ = 7.6, J₂ = 4.8, J₃= 1.2 Hz, 1H). ESI-MS (m/z): [M+H]⁺ calcd. for C₆H₆N₅, 148.06; found, 148.11. ATR-FTIR (neat, cm⁻¹): 1606, 1620, 1637.

3-(pyridin-2-yl)-1,2,4-oxadiazol-5(4H)-one (2).

The compound was made following a previously reported procedure.²⁷ N-Hydroxypicolinimidamide (1.37 g, 10.0 mmol) was dissolved in 20 mL of dry pyridine. Ethyl chloroformate was added to the pyridine solution and the mixture was stirred under nitrogen for 6 h at reflux. The reaction mixture was concentrated and diluted with water and a precipitate formed. The precipitate was washed with water and recrystallized from water producing beige needles, which were collected by filtration. Yield: 0.89 g (5.5 mmol, 55%). ¹H NMR (300 MHz, DMSO-d₆): δ 13.16 (s, 1H), 8.77 (ddd, J₁ = 4.8, J₂ = 1.6, J₃ =1.0 Hz, 1H), 8.09 – 8.02 (m, 1H), 8.00 (dt, J₁ = 7.9, J₂ = 1.4 Hz, 1H), 7.67 (ddd, J₁ = 7.3, J₂ = 4.8, J₃ = 1.5 Hz, 1H). ESI-MS (m/z): [M+H]⁺ calcd. for C₇H₆N₃O₂, 164.05; found, 164.12. ATR-FTIR (neat, cm⁻¹): 1565, 1582, 1779.

5-(pyridin-2-yl)-2,4-dihydro-3H-1,2,4-triazol-3-one (3).

The compound was made following a previously reported procedure.²⁸ Methyl picolinimidate (0.80 g, 5.9 mmol) and ethyl carbazate (0.61 g, 5.9 mmol) were melted together at 200 °C and stirred for 10 min. The resulting solid was recrystallized from EtOH to yield a white solid. Yield: 0.36 g (2.2 mmol, 37%). ¹H NMR (300 MHz, DMSO-d₆): δ 12.07 (s, 1H), 11.81 (s, 1H), 8.63 (dt, J₁ = 4.9, J₂ = 1.3 Hz, 1H), 7.93 (t, J₁ = 1.5 Hz, 1H), 7.92 (d, J₁ = 1.3 Hz, 1H), 7.52 – 7.42 (m, 1H). ESI-MS (m/z): [M+H]⁺ calcd. for C₇H₇N₄O, 163.06; found, 163.22. ATR-FTIR (neat, cm⁻¹): 1600, 1713.

3-(pyridin-2-yl)-1,2,4-oxadiazole-5(4H)-thione (4).

The compound was made following a previously reported procedure.⁴ N-Hydroxypicolinimidamide (1.0 g, 7.3 mmol) was dissolved in 60 mL of acetonitrile. 1,1’-Thiocarbonyldiimidazole (1.94 g, 10.9 mmol) and 1,8-diazabicyclo[5.4.0]undec-7-ene (4.4 mL, 29 mmol) were added to the acetonitrile solution and the mixture was stirred under nitrogen at room temperature overnight. The reaction mixture was diluted with water and acidified with 6 M HCl until a yellow precipitate formed. The precipitate was collected by filtration, washed with water, and dried resulting in a yellow solid. Yield: 0.93 g (5.2 mmol, 71%). ¹H NMR (300 MHz, DMSO-d₆): δ 8.80 (d, J₁= 4.7 Hz, 1H), 8.13 – 8.02 (m, 2H), 7.70 (ddd, J₁ = 6.8, J₂ = 4.8, J₃ = 2.1 Hz, 1H). ESI-MS (m/z): [M+H]⁺ calcd. for C₇H₆N₃OS, 180.02; found, 180.11. ATR-FTIR (neat, cm⁻¹): 1565, 1589.

(tris(2-pyridylmethyl)amine)) (TPA).

TPA was synthesized using a modified literature procedure.²⁴ In a 100 ml round bottom flask 2-(bromomethyl)pyridine hydrobromide (0.6 g, 2 mmol) and potassium carbonate (0.8 g, 6 mmol) were added to acetonitrile (30 mL), bis(pyridin-2-ylmethyl)amine (0.4 g, 0.4 mL, 2 mmol)was then added and the reaction was stirred overnight at room temperature. The reaction mixture was then filtered, and the filtrate was dried down and loaded onto silica. The product was purified via column chromatography using a CH₂Cl₂ : MeOH gradient of 0–10% (R_f = 0.30, 7% MeOH in CH₂Cl₂). The fraction containing the product was dried down yielding a light brown to white solid. Yield: 0.31 g (2 mmol, 50%). ¹H NMR (300 MHz, DMSO-d₆): δ 8.49 (d, J₁ = 4.4 Hz, 1H), 7.77 (td, J₁ = 7.7, J₂ = 1.9 Hz, 1H), 7.59 (d, J₁ = 7.8 Hz, 1H), 7.25 (ddd, J₁ = 7.5, J₂ = 4.8, J₃ = 1.2 Hz, 1H), 3.77 (s, 2H). ESI-MS (m/z): [M+H]⁺ calcd. for C₁₈H₁₈N₄, 291.15; found, 291.26. ATR-FTIR (neat, cm⁻¹): 1588, 1568, 1473, 1436, 1366, 1311.

[Ni(TPA)(pic)]PF₆.

Nickel(II) acetate tetrahydrate (85.7 mg, 344 μmol) and TPA (100 mg, 344 μmol) were dissolved in MeOH (5 mL) and stirred for 10 min at room temperature. Picolinic acid (42 mg, 344 μmol) was then added to the reaction vessel and stirred for another 10 min. Sodium hexafluorophosphate(V) (58 mg, 344 μmol) was then added and stirred overnight at room temperature. A light purple solid precipitate resulted, which was collected by filtration. The resulting solid was washed with MeOH and dried. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave crystals suitable for X-ray diffraction. Yield: 39%. ATR-FTIR (neat, cm⁻¹): 557, 834 (PF₆⁻), 1603, 1018, 1051, 1646, 1363. UV-vis: 531(21). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₄H₂₂N₅NiO₂, 470.11; found, 470.17. HR-MS (m/z): [M-PF₆]⁺ calcd. for C₂₄H₂₂N₅NiO₂, 470.1121; found, 470.1125.

[Ni(TPA)(1)]PF₆.

The same procedure was used as in the synthesis of [Ni(TPA)(pic)]PF₆. The product was isolated as a purple solid. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave purple crystals suitable for X-ray diffraction. Yield: 38%. ATR-FTIR (neat, cm⁻¹): 557, 833 (PF₆⁻), 1608, 1020, 1054. UV-vis: 534 (23). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₄H₂₂N₉Ni, 494.14; found, 494.14. HR-MS (m/z): [M-PF₆]⁺ calcd. for C₂₄ H₂₂ N₉ Ni, 494.1346; found, 494.1343.

[Ni(TPA)(2)]PF₆.

The same procedure was used as in the synthesis of [Ni(TPA)(pic)]PF₆. The product was isolated as a purple solid. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave purple crystals suitable for X-ray diffraction. Yield: 37%. ATR-FTIR (neat, cm⁻¹): 557, 833 (PF₆⁻), 1608, 1020, 1051, 1696, 1398. UV-vis: 548 (21). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₂N₇NiO₂, 510.12; found, 510.14. HR-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅ H₂₂ N₇ NiO₂, 510.1183; found, 510.1181.

[Ni(TPA)(3)]PF₆.

Nickel(II) acetate tetrahydrate (86 mg, 344 μmol) and TPA (100 mg, 344 μmol) were dissolved in iPrOH (5 mL) and stirred for 10 min at room temperature. 5-(Pyridin-2-yl)-2,4-dihydro-3H-1,2,4-triazol-3-one (3, 56 mg, 344 μmol) was then added to the reaction vessel and stirred for another 10 min. Tetrabutylammonium hexafluorophosphate(V) (133 mg, 344 μmol) was then added and stirred overnight. A light purple solid precipitate resulted, which was collected by filtration. The resulting solid was washed with iPrOH and dried. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave purple crystals suitable for X-ray diffraction. Yield: 41%. ATR-FTIR (neat, cm⁻¹): 557, 834 (PF₆⁻), 1606, 1023, 1052, 1625. UV-vis: 552 (19). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₃N₈NiO, 509.14; found, 509.14. HR-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₃N₈NiO, 509.1343; found, 509.1345.

[Ni(TPA)(4)]PF₆.

The same procedure was used as in the synthesis of [Ni(TPA)(pic)]PF₆. The product was a purple solid. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave purple crystals suitable for X-ray diffraction. Yield: 54%. ATR-FTIR (neat, cm⁻¹): 557, 834 (PF₆⁻), 1607, 1021, 1054, 1344. UV-vis: 555 (16). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₂N₇NiOS, 526.10; found, 526.14. HR-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₂N₇NiOS, 526.0955; found, 526.0950.

[Ni(TPA)(bpy)](PF₆)₂.

To a solution of [Ni(TPA)(OAc)(H₂O)]PF₆ (100 mg, 175 μmol) dissolved in MeOH (5 mL) was added 2,2’-bipyridine (27 mg, 175 μmol) and NaPF₆ (35 mg, 210 μmol) The reaction was stirred under ice bath for 2 h. The purple precipitates were collected by filtration and dried. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave purple crystals suitable for X-ray diffraction. Yield: 22%. ATR-FTIR (neat, cm⁻¹): 557, 827 (BPh₄⁻), 1607, 1021, 1054. UV-vis: 557 (12). ESI-MS (m/z): [M-PF₆]²⁺ calcd. for C₂₈H₂₆N₆Ni, 252.08; found, 252.16.

[Co(TPA)(pic)]PF₆.

Cobalt(II) acetate tetrahydrate (85.8 mg, 344 μmol) was dissolved in a mixture of MeOH (3 mL) and iPrOH (2 mL). TPA (100 mg, 344 μmol) was then dissolved to the solution, sonicated and stirred for 10 min until all pink cobalt solution was transformed into a bright green color. Picolinic acid (42 mg, 344 μmol) was added to the solution and stirred for 10 min giving brown color. Tetrabutylammonium hexafluorophosphate(V) (133 mg, 344 μmol) was added and stirred overnight. The pink precipitates were washed with a mixture of iPrOH and Ether. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave pink crystals suitable for X-ray diffraction. Yield: 70%. ATR-FTIR: 557, 834 (PF₆⁻), 1646, 1603, 1024, 1059. UV-vis: 493 (103). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₄H₂₂CoN₅O₂, 471.11; found, 471.11. HR-MS (m/z): [M-PF₆]⁺ calcd. for C₂₄H₂₂CoN₅O₂, 471.1100; found, 471.1089.

[Co(TPA)(1)]PF₆.

The same procedure was used as in the synthesis of [Co(TPA)(pic)]PF₆. The product was a pink solid. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave pink crystals suitable for X-ray diffraction. Yield: 59%. ATR-FTIR (neat, cm⁻¹): 557, 833 (PF₆⁻), 1606, 1020, 1052. UV-vis: 485 (117). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₄H₂₂CoN₉, 495.13; found, 495.09. HR-MS (m/z): [M-PF₆]⁺calcd. for C₂₄H₂₂CoN₉, 495.1325; found, 495.1328.

[Co(TPA)(2)]PF₆.

The same procedure was used as in the synthesis of [Co(TPA)(pic)]PF₆. The product was a pink solid. Vapor diffusion of Et₂O into a CH₃CN solution of the complex gave pink crystals suitable for X-ray diffraction. Yield: 61%. ATR-FTIR (neat, cm⁻¹): 557, 833 (PF₆⁻), 1606, 1023, 1052, 1713, 1731, 1398. UV-vis: 486 (64). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₂CoN₇O₂, 511.12; found, 511.09.

[Co(TPA)(3)]BPh₄.

5-(Pyridin-2-yl)-2,4-dihydro-3H-1,2,4-triazol-3-one (3, 8.7 mg, 54 μmol) was added into THF (5 mL) of [Co(TPA)(OAc)]BPh₄ (40 mg, 54 μmol). The solution changed color from green to brown. The solution was left to stir overnight at room temperature, resulting in pink precipitates, which were collected by filtration and dried. Vapor diffusion of cyclohexane into a CH₂Cl₂ solution of the complex gave pink crystals suitable for X-ray diffraction. Yield: 51%. ATR-FTIR (neat, cm⁻¹): 704, 732 (BPh₄⁻), 1602, 1021, 1058, 1633. UV-vis: 484 (88). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₃CoN₈O, 510.13; found, 510.19.

[Co(TPA)(4)]BPh₄.

3-(Pyridin-2-yl)-1,2,4-oxadiazole-5(4H)-thione (4, 9.7 mg, 54 μmol) was added into THF (5 mL) of [Co(TPA)(OAc)]BPh₄ (40 mg, 54 μmol). The solution changed color from green to brown. Pink crystals suitable for X-ray diffraction were grown by slow evaporation of the complex from THF. Yield: 56%. ATR-FTIR (neat, cm⁻¹): 704, 732 (BPh₄⁻), 1605, 1021, 1055, 1346. UV-vis: 483 (76). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₂₅H₂₂CoN₇OS, 527.09; found, 527.09.

[(Tp^Ph,Me)Co(pic)].

To a solution of [(Tp^Ph,Me)CoCl] (50 mg, 87 μmol) dissolved in CH₂Cl₂ (10 ml) was added 2–3 drops of triethylamine. 1 equiv of picolinic acid (11 mg, 87 μmol) dissolved in 2 mL of MeOH to this solution, resulting in a pink-colored solution. The mixture was stirred at room temperature overnight under a nitrogen atmosphere. After stirring, the solution was evaporated to dryness to give a purple solid. The resulting precipitates was purified via silica gel chromatography eluting a gradient of 0–3% MeOH in CH₂Cl₂ (R_f = 0.23, 5% MeOH in CH₂Cl₂). Pinkish-purple crystals suitable for X-ray diffraction were grown from a solution of the purified precipitates in benzene diffused with pentane. Yield: 63%. ATR-FTIR (neat, cm⁻¹): 1057, 1164, 1333, 1434, 1540, 1600, 1657, 1667, 2532 (B-H). UV-vis: 452 (41), 538 (49), 561 (51). ESI-MS (m/z): [M+Na]⁺ calcd. for C₃₆H₃₂BCoN₇O₂, 687.20; found, 687.28.

[(Tp^Ph,Me)Co(1)].

The same procedure was used as in the synthesis of [(Tp^Ph,Me)Co(pic)]. The product was a purple solid. (R_f = 0.18, 5% MeOH in CH₂Cl₂). Pinkish-purple crystals suitable for X-ray diffraction were grown from a solution of the purified precipitates in benzene diffused with pentane. Yield: 70%. ATR-FTIR (neat, cm⁻¹): 1061, 1169, 1341, 1434, 1538, 1613, 2537 (B-H). UV-vis: 456 (37), 540 (49), 563 (47). ESI-MS (m/z): [M+Na]⁺ calcd. for C₃₆H₃₂BCoN₁₁, 711.23; found, 711.33. HR-MS (m/z): [M+H]⁺ calcd. for C₃₆H₃₂BCoN₁₁, 689.2346; found, 689.2339.

[(Tp^Ph,Me)Co(2)].

The same procedure was used as in the synthesis of [(Tp^Ph,Me)Co(pic)]. The product was a purple solid. (R_f = 0.30, 5% MeOH in CH₂Cl₂). Pinkish-purple crystals suitable for X-ray diffraction were grown from a solution of the purified precipitates in benzene diffused with pentane. Yield: 89%. ATR-FTIR (neat, cm⁻¹): 1060, 1173, 1342, 1436, 1541, 1610, 1732, 1752, 2541 (B-H). UV-vis: 448 (47), 542 (44), 562 (42). ESI-MS (m/z): [M+Na]⁺ calcd. for C₃₆H₃₀BCoN₉O₂, 727.21; found, 727.25.

[(Tp^Ph,Me)Co(bpy)]PF₆.

To a solution of [(Tp^Ph,Me)CoCl] (50 mg, 87 μmol) dissolved in 10 mL CH₂Cl₂ was added 1 equiv of 2,2’-bipyridine (14 mg, 87 μmol) dissolved in 2 mL of MeOH. The mixture was stirred for 10 min. 2,2’-bipyridine was added and stirred at room temperature overnight under a nitrogen atmosphere. After stirring, the solution was evaporated to give a brownish orange solid. Brownish-orange crystals suitable for X-ray diffraction were grown from a solution of the precipitates in benzene diffused with pentane. Yield: 83%. ATR-FTIR (neat, cm⁻¹): 1060, 1175, 1340, 1435, 1541, 1603, 2552 (B-H). UV-vis: 417 (333), 520 (76), 676 (34). ESI-MS (m/z): [M-PF₆]⁺ calcd. for C₄₀H₃₆BCoN₈, 698.25; found, 698.21. HR-MS (m/z): [M-PF₆]⁺ calcd. for C₄₀H₃₆BCoN₈, 698.2489; found, 698.2484.

[(Tp^Ph,Me)Zn(bpy)]ClO₄.

[(Tp^Ph,Me)ZnOH] (40 mg, 71 μmol) was dissolved in 15 mL of CH₂Cl₂. 2,2’-bipyridine (11 mg, 71 μmol) and sodium perchlorate (9mg, 71 μmol) in 10 mL of MeOH was added to the solution, and the reaction mixture was stirred at room temperature overnight under a nitrogen atmosphere. The resulting solution was evaporated to give white solid. Colorless crystals suitable for X-ray diffraction were grown from a solution of the precipitates in benzene diffused with pentane. The crystals were characterized by X-ray crystallography. Yield: 80%. HR-MS (m/z): [M-ClO₄]⁺ calcd. for C₄₀H₃₆BZnN₈, 703.2449; found, 703.2443. Caution: Perchlorate salts of metal complexes with organic ligands are potentially explosive. Only small amounts of these materials should be prepared, and they should be handled with great care.

RESULTS AND DISCUSSION

Preparation of MBIs and model complexes.

Picolinic acid (pic), 2,2’-bipyridine (bpy), and four MBIs were examined in this study (Figure 1). The four heterocyclic carboxylic acid isosteres studied were 2-(1H-tetrazol-5-yl)pyridine (MBI 1), 3-(pyridin-2-yl)-1,2,4-oxadiazol-5(4H)-one (MBI 2), 5-(pyridin-2-yl)-2,4-dihydro-3H-1,2,4-triazol-3-one (MBI 3), and 3-(pyridin-2-yl)-1,2,4-oxadiazole-5(4H)-thione (MBI 4) (Figure 1). These heterocycles have been used as carboxylic acid isosteres in drug design to modulate key parameters, such as acidity, geometry (i.e., planarity), charge distribution, and lipophilicity.²⁹ 2,2’-Bipyridine was added as an alternative isostere of these heterocycles as it shares general structure of the other ring scaffolds, but is aprotic (neutral donor). Picolinic acid and 2,2’-bipyridine were purchased from commercial sources, while the MBIs were synthesized according to published procedures.^{4, 27, 28}

Each of the [Ni(TPA)(MBI)](PF₆) complexes including [Ni(TPA)(pic)](PF₆) were synthesized in a one-pot reaction. For complexes with pic, 1, 2, and 4, an equimolar solution of Ni(CH₃CO₂)₂·4H₂O, TPA, and the respective MBI was prepared in MeOH and then NaPF₆ was added to the mixture. The resulting purple precipitates were collected by filtration and washed with MeOH. To prepare [Ni(TPA)(3)](PF₆), the reaction was carried out in iPrOH to promote precipitation of the desired complex and NBu₄PF₆ was used as a counter ion source based on the solubility of the product in iPrOH. Different synthetic conditions were required to make [Ni(TPA)(bpy)](PF₆)₂, as under either of the aforementioned conditions bpy formed [Ni(bpy)₃]²⁺ as the major product. [Ni(TPA)(bpy)](PF₆)₂ was obtained by first isolating the intermediate, [Ni(TPA)(CH₃CO₂)(H₂O)]PF₆, followed by the addition of a MeOH solution of one equivalent of bpy and NaPF₆ with stirring in an ice bath for 2 h to generate a purple precipitate. All Ni(II) complexes were isolated as purple solids in yields between 22–54% and vapor diffusion of Et₂O into CH₃CN solutions of the complexes gave crystals suitable for X-ray diffraction.

The synthesis and crystallization of Co(II) complexes was slightly different from the Ni(II) complexes, depending on the MBIs. [Co(TPA)(pic)](PF₆), [Co(TPA)(1)](PF₆), and [Co(TPA)(2)](PF₆) were prepared in a one-pot reaction similar to the Ni(II) complexes. Equimolar amounts of Co(CH₃CO₂)₂·4H₂O, TPA, and the respective MBI were mixed in a 1:1 mixture of iPrOH and MeOH, and then NBu₄PF₆ was added to the solution. Pink solids were isolated in yields between 59–70% and vapor diffusion of Et₂O into CH₃CN yielded X-ray quality crystals. To make [Co(TPA)(3)](BPh₄) and [Co(TPA)(4)](BPh₄), it was necessary to first prepare an intermediate, [Co(TPA)(CH₃CO₂)]BPh₄.¹⁹ This intermediate was then combined with one equivalent of 3 or 4 in THF. [Co(TPA)(3)](BPh₄) was crystallized from cyclohexane diffusion into a CH₂Cl₂ solution of the complex, while [Co(II)(TPA)(4)](BPh₄) was directly obtained by slow evaporation from THF in yields of 56% and 51%, respectively. The use of the [Co(TPA)(CH₃CO₂)]BPh₄ intermediate allowed formation of the desired product and prevented formation of insoluble byproducts, including homoleptic complexes (i.e., [Co(MBI)₃]). Note that while the Co(II) complexes for pic and MBIs were readily synthesized, it was challenging to obtain [Co(TPA)(bpy)]²⁺ under any circumstance. Attempts to prepare [Co(TPA)(bpy)]²⁺ always produced [Co(bpy)₃]²⁺ as the major product. This result shows that the reactivity of bpy in these TPA model systems is significantly different from pic and the MBIs.

[(Tp^Ph,Me)Co(MBI)] was applied as an alternative model system with two objectives. The first objective was to compare the coordination geometry of [(Tp^Ph,Me)Co(pic)] and [(Tp^Ph,Me)Co(MBI)] complexes to each other and with those of the corresponding [(Tp^Ph,Me)Zn(MBI)] complexes.⁴ The second goal was to examine the spectroscopic properties of [(Tp^Ph,Me)Co(MBI)] with those of [(Tp^Ph,Me)Co(pic)] and [(Tp^Ph,Me)Co(bpy)]⁺, the latter of which could not be prepared using the TPA model system. [(Tp^Ph,Me)Co(MBI)] complexes were prepared via a [(Tp^Ph,Me)Co(Cl)] intermediate,^{13, 25} where 1 equiv. of each ligand in MeOH was added to the solution of [(Tp^Ph,Me)Co(Cl)] in CH₂Cl₂. To prepare Tp^Ph,Me complexes coordinated with pic and its MBIs, excess triethylamine (TEA) was required to prevent acid-promoted decomposition of the [(Tp^Ph,Me)Co(Cl)]. NaPF₆ was added as a counterion to produce [(Tp^Ph,Me)Co(bpy)]PF₆. Pinkish-purple solids were isolated for [(Tp^Ph,Me)Co(pic)], [(Tp^Ph,Me)Co(1)] and [(Tp^Ph,Me)Co(2)], while [(Tp^Ph,Me)Co(bpy)]PF₆ gave brownish-orange colored solid in good yields. The complexes were recrystallized by diffusing pentane into a solution of the complex in benzene. MBIs 3 and 4 generated [Co(pz^ph,Me)₂(MBI)₂] (pz^Ph,Me = 5,3-methylphenylpyrazole) as a major product (confirmed by X-ray crystallography, data not shown). This result is tentatively attributed to both the intrinsic instability of the starting material, [(Tp^Ph,Me)Co(Cl)],^{13, 30} and different metal-binding properties of MBI 3 and 4, as observed in [(Tp^Ph,Me)Zn(MBI)] system.⁴

Structural analysis.

[(Tp^Ph,Me)Zn(MBI)] complexes of pic, 1, 2, 3, and 4 have been previously reported and show bidentate binding to the Zn(II) ion in all cases.^{4, 31} In addition, the Tp^Ph,Me spectator ligand in [(Tp^Ph,Me)Zn(MBI)] complexes of pic, 1, and 2 show the expected tridentate binding, while with 3 and 4, an unexpected coordination mode where a water molecule displaces one of the pyrazole arms of the Tp^Ph,Me ligand was reported.⁴ Herein, the structural features of [M(TPA)(pic)]⁺, [M(TPA)(MBI)]⁺ and [M(TPA)(bpy)]²⁺ complexes were examined in the context of the previously reported [(Tp^Ph,Me)Zn(MBI)] complexes.

MBIs bound to the TPA complexes showed a bidentate mode of coordination, generating metal ion centres with distorted octahedral geometries (Figures 2 and 3). The bite angles of MBI ligands (N5-M-N6) decreased compared to that of the parent pic complexes (N5-M-O1) (Tables 1 and 2). This trend was expected because of the rigidity of the heterocyclic isosteres. [M(TPA)(1)]⁺ showed the most significant decrease in bite angle by ~2.1° and ~3.8° for the Ni(II) and Co(II) complexes, respectively (Table 1 and 2). All MBIs showed slightly longer bond distances (>0.1 Å) compared to the parent pic complexes. For all TPA complexes, the bond lengths between the metal ions and the tertiary nitrogen donor atom of TPA were similar (<0.1 Å) (Table S5 and S6) and were consistent with previously reported values.^{23, 32–35}

Figure 2.

Crystal structures of [Ni(TPA)(L)] complexes (ORTEP, 50% probability ellipsoids). a) [Ni(TPA)(pic)]⁺, b) [Ni(TPA)(1)]⁺, c) [Ni(TPA)(2)]⁺, d) [Ni(TPA)(3)]⁺, e) [Ni(TPA)(4)]⁺, f) [Ni(TPA)(bpy)]²⁺. Solvent molecules, counter ions, and hydrogen atoms have been omitted for clarity. Donor atoms are alphanumerically labelled. Color scheme: carbon = gray, oxygen = red, nitrogen = blue, nickel = green, and sulfur = yellow.

Figure 3.

Crystal structures of [Co(TPA)(L)] complexes (ORTEP, 50% probability ellipsoids). a) [Co(TPA)(pic)]⁺, b) [Co(TPA)(1)]⁺, c) [Co(TPA)(2)]⁺, d) [Co(TPA)(3)]⁺, e) [Co(TPA)(4)]⁺. Solvent molecules, counter ions, and hydrogen atoms have been omitted for clarity. Donor atoms are alphanumerically labelled. Color scheme: carbon = gray, oxygen = red, nitrogen = blue, cobalt = pink, and sulfur = yellow.

Table 1.

Selected bond lengths and angles for [Ni(TPA)(L)]. Structures with more than one complex in the asymmetric unit are marked with an asterisk.

Ligand	Ni-N5 (Å)	Ni-O1/N6 (Å)	Bite angle (°)	Sigma, Theta (Σ, Θ) (°)
Pic*	2.126(9)	1.984(8)	80.1(3)	89, 287
	2.129(9)	2.003(8)	80.0(3)	89, 294
1	2.231(4)	2.039(3)	77.99(15)	83, 258
2	2.111(2)	2.110(2)	79.70(9)	98, 257
3*	2.128(4)	2.123(4)	79.31(14)	100, 274
	2.116(4)	2.114(4)	79.52(14)	94, 240
	2.093(4)	2.104(4)	80.14(14)	99, 250
4	2.072(3)	2.155(3)	79.59(9)	101, 288
bpy	2.148(3)	2.054(3)	78.51(12)	104, 302

Table 2.

Selected bond lengths and angles for [Co(TPA)(L)]. Structures with more than one complex in the asymmetric unit are marked with an asterisks.

Ligand	Co-N5 (Å)	Co-O1/N6 (Å)	Bite angle (°)	Sigma, Theta (∑, Θ) (°)
pic	2.178(3)	2.012(2)	78.87(9)	113, 360
1*	2.317(5)	2.058(5)	75.2(2)	141, 490
	2.307(6)	2.058(6)	75.5(2)	132, 424
	2.320(6)	2.057(6)	75.3(2)	129, 416
	2.357(6)	2.062(5)	74.5(2)	142, 458
2	2.2743(18)	2.073(2)	77.22(7)	111, 302
3	2.267(3)	2.060(3)	76.94(12)	130, 386
4	2.130(3)	2.181(3)	77.48(11)	121, 401
	2.204(3)	2.161(3)	76.74(11)	122, 359

Some differences in metal binding orientation between Ni(II) and Co(II) complexes were observed. Pic, 1 and 4 showed the same binding orientation regardless of the metal ion, with [M(TPA)(pic)]⁺ and [M(TPA)(1)]⁺ positioning the pyridine ring nitrogen atom (N5) of the MBI cis to the bridgehead nitrogen (N1) of TPA while [M(TPA)(4)]⁺ had the N5 of the MBI trans to N1 of TPA. [Ni(TPA)(2)]⁺ and [Ni(TPA)(3)]⁺ bound such that N5 were trans to N1 of TPA while the corresponding Co(II) complexes showed the opposite (i.e., cis) orientation. The results suggest that subtle differences in metal ion coordination preferences influence how these MBIs bind to metals. Another observation is that the bite angles of the Co(II) complexes were smaller than the Ni(II) complexes (Tables 1 and 2). In addition, the Co(II) complexes are much more distorted than the Ni(II) complexes based on angular distortion parameter ∑ and Θ (Tables 1 and 2).^{36, 37} The higher degree of distortion observed in the Co(II) complexes can be explained by a subtle change in electronics and energetics of each complex including Jahn-Teller distortion which are commonly found in six-coordinate Co(II) complexes.^{38, 39} Overall, all MBIs showed slightly different, but not particularly notable, structural features when compared to the parent pic complex. The structure features of the [Ni(TPA)(MBI)]⁺ complexes were also generally similar to those found with [Ni(TPA)(bpy)]²⁺ (Figure 2, Table 1).

To have a better understanding of how the MBIs studied here are similar to pic or bpy, the structural features of [(Tp^Ph,Me)Co(1)] and [(Tp^Ph,Me)Co(2)] were compared with those of [(Tp^Ph,Me)Co(pic)] and [(Tp^Ph,Me)Co(bpy)]⁺ (Figure 4). In all cases, pic, 1, 2, and bpy showed bidentate coordination to the Co(II) center, as found for [(Tp^Ph,Me)Zn(MBI)] complexes as well (Figure S1).⁴ [(Tp^Ph,Me)Co(pic)] resulted in trigonal bipyramidal geometry (τ = 0.71), which is isomorphous with [(Tp^Ph,Me)Zn(pic)] (τ = 0.72). By contrast, [(Tp^Ph,Me)Co(1)] and [(Tp^Ph,Me)Co(2)] showed distorted square pyramidal geometries (τ = 0.32 and 0.41, respectively) when compared to the analogous Zn(II) complexes (τ = 0.52 and 0.55, respectively). Bpy formed a more square pyramidal geometry with both Co(II) and Zn(II) complexes (τ = 0.01 and 0.11, respectively). This pattern of Co(II) complexes trending toward square pyramidal geometry was expected, as the trigonal bipyramidal geometry is less likely to be preferred relative to square pyramidal in 5-coordinate d⁷ complexes due to ligand field stabilization energy.^{13, 40} The structural features of MBI 3 and 4 could not be compared as they could not be prepared under the same synthetic conditions used with pic, MBI 1 and 2. Interestingly, ligands 3 and 4 showed different coordination behaviour from pic, 1, and 2 in [(Tp^Ph,Me)Zn(MBI)] system,⁴ which suggests that MBI 3 and 4 might also have different metal-binding behavior in [(Tp^Ph,Me)Co(MBI)] system.

Figure 4.

Crystal structures of [(Tp^Ph,Me)Co(L)] complexes (ORTEP, 50% probability ellipsoids). a) [(Tp^Ph,Me)Co(pic)], b) [(Tp^Ph,Me)Co(1)], c) [(Tp^Ph,Me)Co(2)], d) [(Tp^Ph,Me)Co(bpy)]. Phenyl groups of Tp^Ph,Me have been omitted for clarity. Donor atoms are alphanumerically labelled. Color scheme: carbon = gray, oxygen = red, nitrogen = blue, cobalt = pink, and boron = pink.

Computational analysis.

In order to understand the energetics and in particular the crystallographic findings of the MBIs in the model complexes, density functional theory (DFT) calculations for MBIs 1 - 4 in the TPA model system complexes were performed. All calculations were carried out at the wB97x-D/LanL2DZ level of theory. The computations were performed under standard conditions (1 atm, 298 K) in the gas phase to exclude any solvent or crystal packing effects which might have been observed under experimental conditions.

The bond lengths and angles of the computationally optimized structures (Table S7) were compared with the crystallographically determined ones (Tables 1 and 2). The biggest difference in bond length (0.027 Å) and bite angle (1.1°) was observed for [Ni(TPA)(1)]⁺. Overall, a very good agreement between the experimentally determined and theoretically calculated structures was observed with the same trends as described between the MBIs and the corresponding metals. This indicates that the calculated structures accurately reproduce the experimentally characterized complexes.

To investigate the difference in binding orientation of the N5 of MBIs either cis or trans to the N1 of TPA, the difference in energy between these isomers was calculated. The crystallographically described binding orientation of the MBIs (Figure 2 and 3) were also found to possess the lowest calculated energy, confirming that the structures correspond to the inherent preferred orientation of the MBI (Figure 5). The direct comparison between coordination orientation isomers indicates that there are energy differences of between 2.3–4.2 kcal/mol. The biggest difference in energy was observed for MBI 1 (ΔG = 3.6 – 4.4 kcal/mol) and MBI 4 (ΔG = 3.8 – 4.2 kcal/mol), which showed the same orientation of the MBI pyridine N5 to N1 of TPA in the corresponding Ni(II) and Co(II) complexes.

Figure 5.

Computationally optimized structures and relative free Gibbs energy differences of the binding orientation isomers of the MBI 1 - 4 coordinated to Ni(II) (left) or Co(II) (right) TPA complexes at the wB97x-D/LanL2DZ level of theory. a) [Ni(TPA)(1)]⁺ and [Co(TPA)(1)]⁺, b) [Ni(TPA)(2)]⁺ and [Co(TPA)(2)]⁺, c) [Ni(TPA)(3)]⁺ and [Co(TPA)(3)]⁺, d) [Ni(TPA)(4)]⁺ and [Co(TPA)(4)]⁺. Color scheme: carbon = gray, oxygen = red, nitrogen = blue, nickel = green, cobalt = pink, and sulfur = yellow.

This result suggests that MBI 1 and 4 favors a specific coordination orientation independent from the central metal ion. By contrast, a somewhat smaller energy difference (ΔG = 2.3 – 3.1 kcal/mol) was observed for MBI 2 and 3, which show different binding orientations in the crystallographically characterized metal complexes. While the Ni(II) complexes of these MBIs favor a trans coordination of the N5 to the N1, a cis binding coordination preferred in the analogous Co(II) complexes. Note that the energy differences between these two different conformations is still quite small (<5.0 kcal/mol), suggesting that the binding orientation of the MBIs could be readily perturbed by the metalloenzyme active site.

Spectroscopic Analysis.

A distinct feature of both Ni(II) and Co(II) is that they are paramagnetic and have unpaired electrons – 2 for Ni(II) and either 3 or 1 for Co(II)) depending on the spin state. While the typical ¹H NMR spectra fall between 0 and 14 ppm, the ¹H resonances of Ni(II) and Co(II) complexes can often be found from −30 to 160 ppm. Paramagnetic ¹H NMR has been a useful tool for characterizing active site structural features of Co(II)-substituted Zn(II)-dependant enzymes including liver alcohol dehydrogenase⁴¹ as well as monitoring model reactions relevant to Ni(II)-containing metalloenzyme such as acireductone dioxygenase.^{23, 42} To evaluate the coordination complexes formed herein the ¹H NMR spectra of all complexes were obtained (Figures S2 and S3). The paramagnetically shifted resonances of the TPA complexes were observed, which was used for to confirm the stability and solution characteristics of the complexes. For the Ni(II) complexes, broad peaks between 30 – 70 ppm were assigned as the β-protons of TPA pyridyl rings based on ¹H NMR resonances of structurally related Ni(II) complexes (Figure S2).²³ It is interesting to note that all of the Ni(II) complexes, except [Ni(TPA)(bpy)]²⁺, showed more than two peaks in the range of 30 – 70 ppm, suggesting that the asymmetric scaffold of the MBIs in solution might cause magnetically inequivalence in the β-protons of the TPA spectator ligand. Paramagnetic ¹H NMR resonances of the Co(II) complexes showed sharper peaks compared to the Ni(II) complexes (Figure S3). Note that the ¹H NMR resonances of [Co(TPA)(4)]⁺ showed a distinct spectrum, having only one sharp peak between 35 and 75 ppm while the other complexes showed more than two peaks in this same region of the spectra. Typically, in this region, β-protons of TPA pyridyl rings are found in structurally related Co(II) complexes.^{20, 42} Interestingly, [Co(TPA)(4)]⁺ is the only compound that showed a different binding orientation in the solid state, where the N5 was trans to the N1 of TPA (in all the other [Co(TPA)(MBI)]⁺ complexes a cis orientation was observed). This suggests that 4 could have distinct binding characteristics both in solution and in the solid state when bound to Co(II) TPA complexes.

The magnetic properties of the Ni(II) and Co(II) complexes were examined to confirm the oxidation and spin state of each TPA complex. This measurement was especially meaningful for Co(II) complexes as they can adopt either high spin or low spin states while Ni(II) complexes should display two unpaired electrons in all cases.⁴³ The solution magnetic moment of all complexes were determined by the Evans method in DMSO-d₆ (Table 4).²⁶ The Ni(II) complexes were found to exhibit magnetic moments between 3.09 – 3.34 μ_B, which is higher than the spin-only magnetic moment for Ni(II) (2.8 μ_B) due to spin-orbit coupling.⁴⁴ These values confirm that Ni(II) complexes have two unpaired electrons and are well within the range predicted for high spin six-coordinate Ni(II).⁴⁴ Similarly, the Co(II) complexes had magnetic moments from 4.56 – 4.76 μ_B, which are higher than the spin-only value for Co(II) (3.9 μ_B).⁴⁴ These magnetic moments correspond to three unpaired electrons per Co(II) ion and are in good agreement with those observed for most high-spin octahedral Co(II) complexes.^{45, 46} This result is consistent with the distortion of the TPA Co(II) system observed in solid-state X-ray structural data. The ground state of an octahedral high spin d⁷ complex is ⁴T_1g that is Jahn-Teller active.^{39, 47}

Table 4.

UV-visible and magnetic moment data for TPA complexes.

	[Ni(TPA)(L)]n+		[Co(TPA)(L)]n+
Ligand	μeff , B.M.	λmax, nm (ε, M−1 cm−1)	μeff,, B.M.	λmax, nm (ε, M−1 cm−1)
pic	3.16	531 (21)	4.60	493 (103)
1	3.12	534 (23)	4.76	485 (117)
2	3.33	548 (21)	4.56	486 (64)
3	3.32	552 (19)	4.74	484 (88)
4	3.34	555 (16)	4.62	483 (76)
bpy	3.09	576 (21)	-	-

The UV-vis spectra of the TPA complexes were also measured, and all spectra showed a strong absorption at <400 nm. For the Ni(II) TPA complexes, all compounds had a single absorption in the region 400–700 nm, which was found in structurally related octahedral Ni(II) complexes (Table 4 and Figure S4).⁴⁸ The absorption maximum for [Ni(TPA)(1)](PF₆)] (531 nm) was similar to that for [Ni(TPA)(pic)](PF₆) (534 nm) while the other MBI complexes were clustered around 550 nm. This difference in absorption maxima is likely related to a combination of factors including the different binding orientations between MBI 1 versus 2, 3, and 4. Note that the absorption maxima of [Ni(TPA)(MBI)]⁺ complexes were more similar to that generated with pic than bpy in the Ni(II) systems. The UV-vis spectra of the Co(II) complexes showed one major absorption band in the range of 400 – 700 nm consistent with previously reported high-spin octahedral Co(II) complexes (Table 4 and Figure S5).⁴⁹ All the [Co(TPA)(MBI)]⁺ complexes showed a similar range of absorption wavelength maxima (483 – 486 nm), which was slightly shifted from the value for [Co(TPA)(pic)]⁺ (493 nm).

Finally, spectroscopic characterization of the [(Tp^Ph,Me)Co(MBI)] complexes was carried out to further understand the electronic difference between MBIs, pic, and bpy in the model systems. Regarding magnetic moments, pic, 1, and 2 showed a similar range of magnetic moments around 4.84 μ_B for each complex while bpy had a value of 4.66 μ_B. These values are in the range of those observed for other 5-coordinate, high-spin Co(II) complexes.⁴⁶ UV-vis data showed that there are three main absorption bands in the visible range. Interestingly, [(Tp^Ph,Me)Co(pic)], [(Tp^Ph,Me)Co(1)], and [(Tp^Ph,Me)Co(2)] showed similar UV-vis spectra, while the spectrum of [(Tp^Ph,Me)Co(bpy)] was clearly different from the others (Table 5 and Figure S6). These results suggest that MBIs 1 and 2 are electronically more similar to pic than bpy in these model systems.

Table 5.

UV-visible and magnetic moment data for Tp complexes.

		[(TpPh,Me)Co(L)]n+
Ligand	μeff,, B.M.	λmax, nm (ε, M−1 cm−1)
pic	4.84	452 (41), 538 (49), 561 (51)
1	4.84	456 (37), 540 (49), 563 (47)
2	4.85	448 (47), 542 (44), 562 (42)
bpy	4.66	417 (333), 520 (76), 676 (34)

CONCLUSIONS

A TPA model system was used to provide structural and electronic information on picolinic acid and its MBIs. Their coordination modes were determined from X-ray crystallography and DFT calculations were used to evaluate the energetic differences in binding conformations. From these data it was determined that all four heterocyclic MBIs retain bidentate coordination to the metal centers and small energetic disparities in binding geometries can explain differences in MBI binding orientation. NMR, UV-vis, and magnetic moment analysis indicate that each MBI displays slightly different electronic characteristics while retaining overall similarities in oxidation states and spin states. In addition, the comparison of MBIs to pic and bpy indicate that these MBIs are better described as isosteres of picolinic acid than 2,2’-bipyridine. Furthermore, the investigation of [(Tp^Ph,Me)Co(MBI)] as a secondary model complex underlines the need to use a different model systems to best implement isosteric replacement for metalloenzyme inhibition. This work demonstrates that the binding of MBIs can be influenced by several factors including structural and electronic nature of MBIs, the active site metal ion, the spectator ligand (TPA vs Tp^Ph,Me), and by analogy the metalloenzyme active site environment. When using model complexes other factors to be considered when translating these findings to the biological targets, such as steric constraints presented by the metalloenzyme, polar/nonpolar pockets, presence or absence of H-bond acceptors/donors, and potential competition from biological anions. Making broader use of model systems provides a better understanding of MBIs, and the results studied here will prove useful when optimizing important factors in the rational design of metalloenzyme inhibitors.

Table 3.

Selected bond lengths and angles for [(Tp^Ph,Me)Co(L)].

Ligand	Co-N7 (Å)	Co-O1/N8 (Å)	Bite angle (°)	Tau, τ (°)
pic	2.135(3)	1.965(2)	79.97(9)	0.71
1	2.145(4)	1.963(4)	78.48(17)	0.32
2	2.167(4)	2.023(4)	78.17(16)	0.41
bpy	2.080(3)	2.069(3)	78.09(13)	0.01

Synopsis.

Picolinic acid, several picolinic acid metal binding isosteres (MBIs), and 2,2’-bipyridine were examined using a bioinorganic model system of metalloenzyme inhibition [M(TPA)(MBI)] (M = Ni²⁺ and Co²⁺, TPA = tris(2-pyridylmethyl)amine). These complexes were characterized using synthetic, structural, and spectroscopic methods. [(Tp^Ph,Me)Co(MBI)] (Tp^Ph,Me = hydrotris(3,5-phenylmethylpyrazolyl)borate) was used as another model system and the limitations and attributes of the two model systems are discussed.