Mobility of Lewis Acids within the Secondary Coordination Sphere: Toward a Model for Cooperative Substrate Binding

Abstract

Distance dependence of appended Lewis acids in N₂H₄ binding and deprotonation was evaluated within a series of zinc complexes. Variation of spacer-length to a tethered trialkylborane Lewis acid revealed distinct preferences for binding and stabilization of the resulting deprotonated N₂H₃⁻ unit.

Graphical Abstract:

Acidic/basic residues within the secondary coordination sphere of metalloenzyme active sites are often critical for structure regulation and/or reactive-intermediate stabilization.¹ These interactions occur in a dynamic protein environment where acidic residues are highly flexible and mobile. For example, in the active site of type II β-carbonic anhydrase, an aspartate residue (Asp44) gates reactivity to a Zn-OH: initially binding to Zn, then migrating 1.5 Å to hydrogen bond with Zn-OH₂.² Similarly, a recent report of Mo-nitrogenase revealed a dynamic multi-metallic cofactor, where the nitrogen reduction sequence is proposed to involve dynamic rotation and substrate/H-bonding interactions.³

Synthetic models containing pre-arranged secondary sphere groups can provide insight into the roles through which acidic groups facilitate substrate binding.⁴ While such models often use rigid molecular scaffolds to provide critical snapshots of donor/acceptor adducts, they do not capture mobility-dependent reactivity. Modelling mobility of an acidic residue within the secondary coordination sphere is synthetically challenging.⁵

Our group is working to evaluate how the precise structural, electronic, and cooperative modes in the secondary coordination sphere can be used to regulate reactivity.⁶ Recently, our lab investigated the role of ligand-appended acidic groups in homolytic bond scission of hydrazine by transition metal complexes.⁷ A key design aspect in those studies was the flexibility of the tethered trialkylborane Lewis acid, which enabled acid/base interactions to occur in both the primary (cooperatively with the transition metal) and secondary (independent of metal) spheres. A first-generation ligand, ^3-BBNNN^tBu, containing a 3-carbon alkyl tether between the Lewis acid and the bidentate ligand was synthesized by hydroboration of 2-(1-allyl-5-(tert-butyl)-1H-pyrazol-3-yl)-6-methylpyridine with 9-borabicyclo[3.3.1]nonane (BBN).⁸ While this tether length stabilized monoatomic ligands (e.g. -NH₂) cooperatively with iron (Fig. 1), we obtained divergent results when attempting to sequester a diatomic substrate, cyanide.⁹ These disparate results highlight a Lewis acid dependence on substrate binding and illustrate a need to establish parameters (e.g. tether length) that maximize binding for a given substrate. Minimizing the energy requirement for cooperative acid-substrate-metal interaction will provide design cues that will enable the use of less-acidic Lewis acids, and ultimately facilitate product release—a challenge for catalytic turnover.

Fig. 1.

Left: Conceptual design. Right: specific substrates investigated in this study and their potential binding modes.

To probe distance relationships between the Lewis acid and a given substrate, we evaluated a set of compounds where: 1) the Lewis acid is held constant (9-BBN), 2) the alkyl tether length is systematically varied from 2 to 4 methylene (-CH₂-) units, and 3) the substrate contained variable bonding modes (Fig. 1). Zn was selected to ensure a consistent coordination geometry, and hydrazine (N₂H₄) was selected as the substrate (Fig. 1). Cooperative binding of N₂H₄ between the two Lewis acids, Zn²⁺ and trialkylborane, must involve coordinating each of the two lone pairs (i.e. μ−1,2-N₂H₄).

The two- and four-carbon length precursor ligands, ^vinylNN^tBu and ^butenylNN^tBu, were prepared by adapted literature procedures (see SI). Initial metalation strategies of the new ligands mirrored our synthesis of (^3-BBNNN^tBu)ZnBr₂ (A-3; 3 denotes tether length, A denotes compound series).⁸ Stirring a CH₂Cl₂ solution of ^vinylNN^tBu or ^butenylNN^tBu with ZnBr₂ furnished (^vinylNN^tBu)ZnBr₂ and (^butenylNN^tBu)ZnBr₂ as white powders (Fig. 2). Whereas late-stage hydroboration of (^butenylNN^tBu)ZnBr₂ with 9-BBN proceeded (RT, THF, 18 hr) to afford (^4-BBNNN^tBu)ZnBr₂ (A-4), (^vinylNN^tBu)ZnBr₂ was obstinate to hydroboration. The molecular structures of (^vinylNN^tBu)ZnBr₂ and (^butenylNN^tBu)ZnBr₂ (Fig. 2B) determined by single crystal X-ray diffraction (SC-XRD) revealed a potential origin of the difference in hydroboration reactivity. Although hydroboration of vinyl substituents is generally facile with 9-BBN,¹⁰ (^vinylNN^tBu)ZnBr₂ possesses two large steric moieties, the -C(CH₃)₃ and ZnBr₂, that render the vinyl group inaccessible.¹¹ To overcome this challenge, we pursued an early-stage hydroboration. Treating ^vinylNN^tBu with 9-BBN generated ^2-BBNNN^tBu in situ, which was metalated with ZnBr₂ in one-pot to afford (^2-BBNNN^tBu)ZnBr₂ (A-2). Spectroscopically and structurally, A-2 and A-4 are similar to A-3. The distance between the two acidic centres, Zn and the trialkylborane, increases by ca. 1 Å for each additional –CH₂– unit added to the tether length (Zn-B: A-2 = 4.76_ave; A-3 = 5.82; A-4 = 6.57 Å).¹² This trend suggests the system is well-suited for a distance-dependent cooperativity study.

Fig. 2.

A) Synthesis of complexes A. B) Molecular structures (50% probability ellipsoids) of A-2 and A-4 as well as their vinyl and butenyl precursors, respectively. H-atoms not attached to alkenyl moieties are omitted for clarity.

Our previous studies revealed the three-carbon tether was ideally suited for cooperative binding of a wide range of μ−1,1 substrates;^8–9 however, we hypothesized a distinct tether length would be needed for a μ−1,2 substrate. We used hydrazine (N₂H₄) to investigate this hypothesis. Previously, we demonstrated that the trialkylborane of A-3 could capture a single equivalent to N₂H₄ to afford (^3-BBNNN^tBu)ZnBr₂(N₂H₄) (B-3), where the terminal -NH₂ lone-pair is uncoordinated.¹³ The series of complexes B were synthesized by standard protocols and all share similar spectroscopic properties.⁸ Structurally, the coordination environment at Zn is unperturbed and the distances between the two acidic sites, Zn and boron, are variable and range from 4.80 (B-2) to 7.38 Å (B-4). All display weak intra- (B-2 and B-3) or intermolecular (B-4) NH…Br hydrogen bonding interactions¹⁴ that result in the N₂H₄ moieties being nearly equidistant (ave = 4.36 +/− 0.17 Å) to a Zn atom of the same, or an adjacent molecule.

Halide abstraction from complexes B with Tl⁺ afforded cationic complexes, [(^n-BBNNN^tBu)ZnBr(N₂H₄)][X], (C-2-C-4; X = OTf, PF₆) and were subjected to SC-XRD studies (Fig. 3). Each complex displays a C₁ symmetric tetrahedral bromido-Zn (τ₄ = 0.83–0.85) chelated by the ^n-BBNNN^tBu ligand. The fourth coordination site is occupied by a hydrazine ligand bridging to the appended trialkylborane (i.e. Zn-NH₂NH₂-BR₃). Across the series of compounds, the Zn-N₂H₄ bond distance elongates from 2.0256(11) – 2.0892(15) Å (C-2 < C-3 < C-4) as the tether length to the trialkylborane increases (Δ = 0.055 Å). This trend is also observed, though to a lesser degree, in the B-N₂H₄ distance where C-2 is shorter (1.6463(18) Å) in comparison to C-3 and C-4 (1.675(2) and 1.668(2) Å, respectively). Both the R₃B-N₂H₄¹⁵ and Zn-N₂H₄ distances¹⁶ are comparable to related species. The interactions of N₂H₄ with the two Lewis acids, Zn and boron, force a nearly fixed distance from one another with variation of only 0.33 Å across the C series (contrasting with the B series; ΔZn-B = 2.58 Å), highlighting the accordion-like flexibility of the acidic trialkylborane.

Fig.3.

A) Reversible formation of C from B. B) Molecular structures (50% probability ellipsoids, only H-atoms attached to heteroatoms are displayed). B-3 is previously reported.⁸

The solid-state data of C-2 revealed both the shortest Zn-N and B-N contacts suggesting that this binding pocket is best suited for favourable host/guest interactions with N₂H₄. Variable temperature NMR spectra provided additional support. At 25 °C, complexes C display C_s symmetric spectra that suggest a dynamic process. Upon cooling, each undergo broadening with a coalescence temperature of T_c = 276, 260, and 240 K for C-2, C-3, and C-4, respectively. For each complex, we propose this dynamic process is the same. From the coalescence temperature of C-2 in CDCl₃, we obtained an activation energy barrier for this process of 12.8 +/− 0.1 kcal/mol (see SI). This value is similar to a previously reported on/off binding event between Zn and a ligand-tethered amine (13 kcal/mol).¹⁷ Across the series of compounds C, this energy varies by ~ 1.5 kcal/mol. Complex C-4 displays both the lowest barrier for activation (11.5 kcal/mol), as well as the longest Zn-N₂H₄ bond distance.

Two limiting dynamic acid/base interactions are possible in complexes C: 1) Zn-N₂H₄ bond scission, or 2) R₃B-N₂H₄ bond scission. The latter was probed by attempting to form a borane- free analogue of complexes C. Treating (^butylNN^tBu)ZnBr₂, a borane-free surrogate where the alkyl-BBN portion of the ligand was replaced with n-butyl, N₂H₄ caused immediate demetalation of the ligand. These results suggest that boron-nitrogen bond scission in complexes C may result in decomposition, and highlights the requirement of an appended Lewis acid for stability. To ascertain the differences in Lewis acid strengths, we measured solution Gutmann-Beckett acidities.¹⁸ These data show that the Zn in C-2 is significantly more acidic than the appended borane (acceptor number = 66.3 vs. 25.0; see SI). The acceptor numbers suggest that a competitive base may promote dissociation of the weaker acid.¹⁹ Treating compounds C with 1.5 equiv. [Bu₄N][Br] rapidly regenerated compounds B, highlighting the lability of the Zn-N₂H₄ bond (Fig. 3A).²⁰ This result highlights the challenges of experimentally measuring and comparing Lewis acidities: due to hard/soft-acid/base mismatches and steric considerations, measured acidities are substrate specific and do not always correspond to accessible acidities.²¹

We propose the dynamic solution behaviour for complexes C is the result of dissociation of the Zn-N₂H₄ bond (Fig. 4, top). Thermodynamically, the tether length has minimal effect on the Zn-N₂H₄ bond dissociation energy.²² We computationally probed the electron density distribution of the Zn-N₂H₄ unit via density functional theory (DFT) methods. Complexes C were analysed at the B3LYP/6–311+G(2d,p) (CH₂Cl₂) level of theory for all atoms except Zn (6–311+G(2d) level). Natural bond orbital analyses are consistent with the extracted thermodynamic parameters from NMR spectroscopy: changing tether length results in minimal variation to the natural charges or bond indices. Kinetically, the tether length has a clear effect. Measured rates of Zn-N₂H₄ on/off binding at the coalescence temperature, k_c, for each complex follows the trend C-2 > C-3 > C-4.²³ This trend can be rationalized in terms of the distances between the two acidic residues, Zn and boron, in complexes B and C. As the tether length is increased, the rate of on/off binding is slowed because the distance between the two acids increases.

Fig. 4.

A) Top: Dynamic solution behaviour of complexes C. Left: reversible deprotonation of C-3 to form D-3 and molecular structure of D-3 (50% probability ellipsoids, only H-atoms attached to heteroatoms displayed). B) Calculated pKa values.

The acidification of hydrazine by Lewis acids was probed through DFT assessment of C-3, N₂H₄, and its adduct with 9-methyl-9-borabicyclo[3.3.1]nonane (9-Me-9-BBN).²⁴ Upon coordination of N₂H₄ to 9-Me-9-BBN, the proximal N-H protons are acidified by 20 pK_a units (Fig 4B). In C-3, the N-H protons are further acidified by ca. 10 pK_a units. Notably, the Lewis acidic Zn in C-3 acidifies the N-H protons to a greater extent (pK_a = 24.9 vs. 26.9; Fig. 4B, right) than the trialkylborane, supporting the measured acceptor numbers.²⁵ Overall, these data indicate that the addition of both Zn and boron Lewis acids serve additive roles to increase the acidity of coordinated N₂H₄.

We assessed the viability of deprotonating the cooperatively captured N₂H₄. Treating a THF solution of C-3 with KN(SiMe₃)₂ at low temperature resulted in formation of HN(SiMe₃)₂ and a C₁ symmetric complex by ¹H NMR spectroscopy—consistent with formation of a [N₂H₃]¹⁻ moiety and production of (^3-BBNNN^tBu)ZnBr(N₂H₃) (D-3).²⁶ In contrast, when similar reactions were attempted for the other tether lengths to form D-2 and D-4, we observed either demetalation or an intractable mixture (see SI), illustrating a unique stabilizing effect for the 3-carbon variant.

Due to the multiple bonding modes possible with [N₂H₃]¹⁻,²⁷ the solid-state structure was determined by SC-XRD. D-3 represents the first structurally characterized example of a Zn hydrazido¹⁻ complex. Both boron and Zn are attached to the same nitrogen of the hydrazido ligand (Fig. 4, bottom). Deprotonation results in a decrease in both the Zn-N (1.973(2); Δ = 0.074 Å) and B-N (1.630(4); Δ = 0.027 Å) bond lengths, compared to C-3. In D-3, the terminal -NH₂ does not interact with Lewis acidic residues and the N-N distance is identical to C-3. The tandem Lewis acid/metal stabilization of [N₂H₃]¹⁻ in D-3, is reminiscent of the vanadium Lewis acid/base triad that employed a weakly acidic, but rigid, tetramethyl-1,3,2-dioxaborolane Lewis acid (B-N₂H₃ = 1.623(4) Å).²⁸

The structure of D-3 is unique to the series because the 3-carbon tether can accommodate both μ−1,1 and μ−1,2 ligands (⁻N₂H₃ and N₂H₄). The ability to stabilize both types of substrates enables facile rearrangement upon deprotonation of C-3 to form D-3. Importantly, this process is reversible; treating D-3 with [Ph₂NH][OTf] quantitatively regenerates C-3 (Fig. 4, left). The mobility of the Lewis acid is highlighted for complexes B-3, C-3, and D-3. While operating independently, the boron atom is located 5.41 Å away from Zn (B-3). Upon forming a cooperative interaction with Zn in C-3 to capture a diatom, N₂H₄, the Zn-B distances decreases to 3.95 Å (Δ = 1.46 Å). Finally, following deprotonation, cooperative stabilization of the same N-atom in ⁻N₂H₃ decreases the Zn-B distance to 3.06 Å. Overall, the Lewis acid exhibits mobility of 2.35 Å, mirroring the distance traversed by amino acids in metalloenzymes during turnover.^{1a, 2}

We have described a system where it is possible to probe distant-dependent substrate-Lewis acid relationships. The trialkylborane in this system was tethered by -(CH₂)_n- units at defined distances, from a substrate. Of note, the three-carbon tether affords the most versatility in terms of substrate accommodation. Our results suggest that while a certain tether length may provide an ideal fit for a given substrate, the versatility of the three-carbon tether may be the most useful for stabilizing a variety of high-energy reduction products of a single substrate (e.g. N_xH_y from N₂). Further work from our lab is investigating both of these aspects.