Redox Tuning via Ligand-Induced Geometric Distortions at a YMn3O4 Cubane Model of the Biological Oxygen Evolving Complex.

The function of proteins involved in electron transfer is dependent on cofactors attaining the necessary reduction potentials. We establish a mode of cluster redox tuning through steric pressure on a synthetic model related to Photosystem II. Resembling the cuboidal [CaMn3O4] subsite of the biological oxygen evolving complex (OEC), [Mn4O4] and [YMn3O4] complexes featuring ligands of different basicity and chelating properties were characterized by cyclic voltammetry. In the absence of ligand-induced distortions, increasing the basicity of the ligands results in a decrease of cluster reduction potential. Contraction of Y-oxo/Y-Mn distances by 0.1/0.15 Å enforced by a chelating ligand results in an increase of cluster reduction potential even in the presence of strongly basic donors. Related protein-induced changes in Ca-oxo/Ca-Mn distances may have similar effects in tuning the redox potential of the OEC through entatic states and may explain the cation size dependence on the progression of the S-state cycle.

Through processes such as photosynthesis and respiration, electron transfer (ET) is fundamental to life.[1] In addition to controlling the rate of ET, tuning the redox potential of ET mediators can regulate biological reactions.[2−5] Factors that tune the redox potentials of metallocofactors include: (1) oxidation state and geometry of the metal center(s),[6,7] (2) ligands in the primary coordination sphere,[8,9] (3) secondary coordination sphere interactions such as hydrogen bonding and polarity of the medium,[10−14] and (4) binding of regulatory molecules.[15] Featuring a [CaMn4O5] core, the oxygen evolving complex (OEC) of Photosystem II catalyzes the 4 e–/4 H+ oxidation of H2O to O2.[16−18] The mechanism of O–O bond formation and the role of the redox-inactive Ca2+ ion have been the subject of numerous biochemical, spectroscopic, computational, and synthetic studies, but the role of Ca2+ remains unclear.[19−33] Removal of Ca2+ has a minimal effect on the [Mn4O5] core structure.[23] Substitution of Ca2+ with other metal ions has distinct outcomes. Incorporation of alkali metals reveals a cation size dependence in the S1 → S2 one e– oxidation step: Li+ and Na+ supplemented samples show the multiline EPR signal characteristic of the S2 state, while K+, Rb+, and Cs+ supplemented samples do not show formation of the S2 state, suggesting that the redox properties of the OEC are affected by the size of the redox-inactive metal.[34] Notably, turnover is inhibited by substitution of Ca2+ with other metal ions with the exception of Sr2+,[35] providing opportunities for mechanistic insight through systematic structure–function studies on model complexes.

Studies on synthetic heterometallic complexes featuring acetate-bridged [MMn3O4], [MMn3O(OH)], and [MFe3O(OH)] cores with redox-inactive metal ions M = Ca2+, Sr2+, Zn2+, Y3+, Ln3+, and Sc3+ have shown that cluster reduction potentials correlate linearly with the pKa of the metal aqua ion: the least acidic Ca2+- and Sr2+-containing clusters in the series have the lowest reduction potentials.[31,36−41] For the series of [Ln3+Mn3O4] complexes, redox potential is also found to correlate linearly with the ionic radii of the lanthanides with the larger, less acidic lanthanide-containing clusters having lower reduction potentials.[38] Theoretical studies on the cuboidal [MMn3O4] model complexes have validated the correlation between redox potential and the Lewis acidity of the redox-inactive metal ion; however, calculations also suggest that such correlation does not hold for the OEC, which is proposed to respond only to the charge of the redox-inactive metal ion.[42] Dinuclear examples have been reported in which redox inactive metal ions not only influence the redox potential of the transition metal, but also modulate the reactivity of oxo, peroxo, and other metal bound moieties.[43−50] In several cluster and bimetallic systems, correlations are observed between reduction potentials or rates of reaction and Lewis acidity of redox inactive metals.[31,36,38,39,51−55] Other systems show dependence between redox chemistry and the charge of the cation, not its Lewis acidity;[42,46] notably, such systems have constrained binding environments such as the protein cavity for the OEC or pendant crown ethers. Finally, correlations involving the Lewis acidity or the charge of the redox-inactive metal both fail to address the cation size dependence experimentally observed in the OEC: larger, less acidic alkali metals inhibit the S1 → S2 oxidation.[34]

Herein, we report the synthesis, crystal structure, electrochemical characterization, and comparison of [Mn4O4] and [YMn3O4] complexes featuring bridging ligands of different basicity and chelating properties. We demonstrate that geometric pressure imposed by a chelating ligand results in contraction of metal-oxo distances, leading to an increase in cluster reduction potential despite the presence of more electron-rich ligands. We propose that related changes in Ca-oxo/Ca–Mn distances driven by the protein active site cavity may have a similar effect in tuning the redox potential of the OEC. A similar effect may explain the cation size dependence in the S1 → S2 oxidation, whereby the cavity surrounding the OEC may enforce nonequilibrium metal-oxo distances that impact the reduction potential of the OEC.

To investigate the effect of ligand basicity in modulating cluster reduction potential, [Mn4O4] complexes featuring carboxylate and amidate bridging ligands were synthesized (Scheme ). Treatment of the previously reported LMn4O4(OAc)3 complex 1-Mn with a tethered diamidate ligand results in the formation of LMn4O4(diam)(OAc) (3-Mn).[56] Subsequent treatment of 3-Mn with p-CF3-benzoic acid results in the formation of LMn4O4(diam)(OBzCF3) (2-Mn).[56] Complexes 1-Mn ∼ 3-Mn are nearly isostructural with respect to the Mn-oxo distances; the [Mn2IIIMn2IV]/[MnIIIMn3IV] couple is observed at +250, −15, and −150 mV vs Fc/Fc+, respectively.[37,56] Toward further decreasing the potential of this redox couple, a triamidate-supported [Mn4O4] cluster was targeted (Scheme ). Deprotonation of H3triam with KH followed by treatment with 1-Mn results in the formation of 4-Mn. The crystal structure of 4-Mn is consistent with the LMn4O4(triam) formulation (Figure a). The reversible [Mn2IIIMn2IV]/[MnIIIMn3IV] couple is observed at −465 mV vs Fc/Fc+, representing a shift of 600 mV relative to 1-Mn (Figure S13). Treatment of 4-Mn with Ag(OTf) affords the one electron oxidized species [LMn4O4(triam)][OTf] (4-Mn-ox). In a related series of [Co4O4] cuboidal systems, cluster reduction potentials were found to be inversely proportional to the weighted sum of ligand pKa’s (effective basicity) in H2O.[57] A similar correlation can be obtained for 1-Mn ∼ 4-Mn using the pKa of HOAc (12.6), p-CF3–C6H4CO2H (9.6), and N-methylacetamide (25.9) in DMSO with a slope of −70 mV/pKa (Figure , Table S3), establishing a linear trend between ligand basicity and cluster potential in [Mn4O4] complexes.[58−61]

Scheme 1

Synthesis of [Mn4O4] and [YMn3O4] Complexes Studied in This Work

Figure 2

Truncated crystal structures of (a) 4-Mn, (b) 2-Y, and (c) 3-Y. Bolded bonds highlight metal-oxo bonds. Mn (green), O (red), N (blue), Y (purple).

Figure 1

Linear correlation between redox potential and effective ligand basicity in [Mn4O4] complexes 1-Mn ∼ 4-Mn. Similar trend based on ligand basicity observed for [YMn3O4] complexes 1-Y and 2-Y. Deviation from the trend in 3-Y attributed to a geometric effect described in this study.

To investigate the effect of ligand basicity in modulating the reduction potential of clusters featuring redox-inactive metals, [YMn3O4] complexes supported by different bridging ligands were targeted (Scheme ). For [LYMn3O4(OAc)3]+ (1-Y), the [YMn3IV]/[YMnIIIMn2IV] couple is observed at −430 mV vs Fc/Fc+.[39] Treatment of 1-Y with Cp*2Fe results in the formation of the one electron reduced complex [LYMn3O4(OAc)3] (1-Y-red).[62] Treatment of 1-Y with 1 equiv of a chelating bis-oxime proligand (H2N4O2) results in the formation of 2-Y. The crystal structure of 2-Y is consistent with the [LYMn3O4(N4O2)(OAc)(DMF)][OTf] formulation (Figure b).[63] Despite the pKa difference of 13 units between acetic acid and acetoxime (pKa = 25.2)[64] moieties, the reaction is thought to be driven by a kinetic chelate effect. For 2-Y, the reversible [YMn3IV]/[YMnIIIMn2IV] couple is observed at −860 mV vs Fc/Fc+ (Figure S15). The bis-oximate ligand decreases the reduction potential of 2-Y by 430 mV relative to that of 1-Y, consistent with the increased basicity of the oximate donors compared to acetates. The difference in redox potential between 1-Y and 2-Y is similar to that between 1-Mn and 3-Mn, suggesting that a similar trend based on effective ligand basicity may be operative in [YMn3O4] complexes.

On the basis of the effective basicity trend, a triamidate-supported [YMn3O4] complex was targeted to further decrease the potential of the [YMn3IV]/[YMnIIIMn2IV] couple. Due to the larger size of Y compared to Mn, triam3– was not suitable as a supporting ligand. However, treatment of 1-Y with a tren-based triacetamide proligand (H3Ntriam) and 3 equiv of NaOBu results in the formation of the amidate-supported, one electron reduced complex 3-Y-red (Scheme ). The ESI-MS peak at m/z = 1443 is consistent with the mass of [LYMn3O4(Ntriam)]+ (Figure S10). For 3-Y-red, the reversible [YMn3IV]/[YMnIIIMn2IV] couple is observed at −290 mV vs Fc/Fc+ (Figure S17). Accordingly, treatment of 3-Y-red with (Fc)(OTf) leads to the formation of the one electron oxidized complex 3-Y. The crystal structure of 3-Y is consistent with the [LYMn3O4(Ntriam)][OTf] formulation (Figure c). Despite the similarity in pKa for acetoxime and acetamide moieties and the increased effective ligand basicity in 3-Y, the tris-amidate ligand increases the reduction potential of 3-Y by 140 mV relative to that of 1-Y, inconsistent with the increased basicity of the amidate donors compared to acetates.

To obtain a rationale for the shifts in redox potential observed in 1-Y ∼ 3-Y, metal-oxo and metal–metal distances were compared among the series of oxidized and reduced complexes (Table ). Comparing the reported crystal structures of 1-Y-red and 1-Y, a slight contraction of Y-oxo and Y–Mn distances is observed in 1-Y-red.[39,62] This contraction can be rationalized in terms of the increased basicity of the bridging oxos in the reduced cluster, resulting in the observed Y-oxo/Y–Mn contraction. Comparing the structures of 1-Y and 2-Y, the average Y-oxo and Y–Mn distances differ only by about 0.01 and 0.03 Å, respectively. Therefore, binding of the bis-oximate ligand does not significantly change the geometry of the [YMn3O4] core, and the decrease in reduction potential of 2-Y relative to that of 1-Y can be attributed to the increased effective basicity of the ligand framework.

Table 1

Y-oxo and Y–Mn Distances (Å) in Complexes 1-Y-red, 1-Y, 2-Y, and 3-Ya

	1-Y-red	1-Y	2-Y	3-Y
Y(1)–O(1)	2.297(3)	2.432(2)	2.308(2)	2.289(4)
Y(1)–O(2)	2.344(3)	2.335(2)	2.396(2)	2.278(4)
Y(1)–O(3)	2.306(3)	2.389(2)	2.422(3)	2.289(4)
Y–O average	2.316(3)	2.385(2)	2.375(2)	2.285(4)
Y(1)–Mn(1)	3.212(1)	3.239(1)	3.181(1)	3.106(1)
Y(1)–Mn(2)	3.144(1)	3.298(1)	3.193(1)	3.119(1)
Y(1)–Mn(3)	3.192(1)	3.213(1)	3.295(1)	3.100(1)
Y–Mn average	3.183	3.250(1)	3.223(1)	3.108(1)

Average distances underlined for emphasis.

Comparing the structures of 1-Y and 3-Y, a more significant contraction in the average Y-oxo and Y–Mn distances is observed in 3-Y by about 0.1 and 0.15 Å, respectively. This contraction in 3-Y can be attributed to the geometric pressure exerted by the chelating tripodal tris-amidate ligand framework, pulling the Y center closer to the Mn3 core than the thermodynamic distances observed in 1-Y and 2-Y. Despite the increase in ligand basicity going from acetates to amidates, the shorter Y-oxo interactions enforced by the chelating ligand framework increase the reduction potential of 3-Y, potentially by decreasing the electron density available for the Mn centers as a consequence of the shorter, stronger Y-oxo interactions. The Y center in 3-Y is effectively more Lewis acidic because of its closer proximity to the redox sites (Mn3O4) enforced by the ligand. In the series of [Mn4O4] complexes 1-Mn ∼ 4-Mn, noticeable changes in the [Mn4O4] core enforced by the chelating ligand have not been observed (Table S2).[56] By taking into account only the basicity of the triamidate ligand (Ntriam3–), the reduction potential of 3-Y would be expected to be close to −1000 mV vs Fc/Fc+, implying that small geometrical changes (0.1/0.15 Å) in the Y-oxo/Y–Mn distances may shift the cluster redox potential by ∼700 mV. Finally, compared with 1-Y-red and 1-Y which display nonchelating ligands on Y, the structure of 3-Y has Y-oxo/Y–Mn distances more similar to the reduced cluster, 1-Y-red, suggesting that the [Ntriam]3– ligand enforces a geometry closer to the preferred thermodynamic structure of the reduced [YMn3O4] core. Therefore, the geometric pressure imposed by the ligand favors the reduced form, as highlighted by the more positive potential, despite the significantly more electron-rich ligand set.

In summary, [Mn4O4] and [YMn3O4] complexes featuring bridging ligands of different basicity and chelating properties were synthesized and characterized by X-ray crystallography and cyclic voltammetry. In agreement with previous studies of [Co4O4] clusters, increasing the effective basicity of the ligand framework of [Mn4O4] results in a decrease of cluster reduction potential.[57] Ligand-induced distortion of cluster geometry is demonstrated as a mode of tuning cluster reduction potential. A significant contraction of Y-oxo/Y–Mn distances by 0.1/0.15 Å enforced by the chelating ligand results in a positive shift of the cluster reduction potential even in the presence of electron donating tris-amidate donors. We propose that within the rigid cavity surrounding the OEC,[23,65] structural changes that affect Ca-oxo/Ca–Mn distances may have a similar effect in tuning the redox potential of the OEC. Furthermore, our model studies suggest that the cation size dependence in the S1 → S2 one e– oxidation in the OEC is the result of redox tuning through a similar geometric effect: the rigid cavity surrounding the OEC may enforce shorter, nonequilibrium metal-oxo distances for cations with ionic radii larger than that of Ca2+, resulting in an increase in the reduction potential of the OEC and inhibiting the S1 → S2 transition. While other factors such as the pKa of the water bound to the redox-inactive metal may contribute to the slower turnover frequency of the Sr-substituted OEC, a similar size effect on redox chemistry may also be in place.[66,67] Related geometric constraints in synthetic systems may result in nonlinear changes of reduction potentials and reactivity.