The first high oxidation state manganese–calcium cluster: relevance to the water oxidizing complex of photosynthesis

Abstract

Synthetic entry has been achieved into high oxidation state Mn–Ca cluster chemistry with the preparation of [Mn₁₃Ca₂O₁₀(OH)₂(OMe)₂(O₂CPh)₁₈(H₂O)₄]; the structure contains [Mn₄CaO₄] sub-units similar to that found in the photosynthetic water oxidizing complex.

---

The water oxidizing complex (WOC) near photosystem II (PS II) catalyses the oxidation of H₂O to O₂ gas in green plants and cyanobacteria.^1,2 This four-electron process involves various oxidation levels of the WOC (the so-called S_n states, n = 0 to 4),³ and is the source of essentially all the O₂ on this planet. The WOC has been studied for many years by a variety of spectroscopic and physicochemical techniques, and has long been known to require Ca²⁺ for activity,⁴ but it was the recent crystal structure resolution of the PS II of the cyanobacterium Thermosynechococcus elongatus at 3.5 Å that finally revealed the details of a heterometallic Mn₄CaO₄ cluster.⁵ This comprises a Mn₃CaO₄ cubane, with the fourth Mn atom attached to one of its bridging O²⁻ ions.

The synthesis and detailed study of synthetic analogues (models) of this cluster would greatly enhance our understanding of the spectroscopic, physical and redox properties of the WOC, as well as its reactivity characteristics. However, not only is this heterometallic Mn₄CaO₄ cluster currently unknown in inorganic chemistry, there are no structurally characterized molecular species to our knowledge containing both Ca²⁺ and higher oxidation state Mn³⁺ and/or Mn⁴⁺ ions. Currently known is a single Mn²⁺–Ca²⁺ species, which is polymeric,⁶ and a salt where the Ca²⁺ is the cation of a Mn₄-containing anion (shortest Mn⋯Ca separation is 4.895 Å).⁷ This is in stark contrast to the many Mn₄ complexes currently known, such as those with butterfly, cubane, and dimer-of-dimer structures.⁸ None of these has ever been reported to incorporate calcium. However, we can now report that a synthetic route has been developed into this area of inorganic chemistry.

Various reactions between Mn and Ca sources were explored under many different conditions before the following was developed. The reaction of (NBuⁿ₄)[Mn₄O₂(O₂CPh)₉(H₂O)]⁹ with 0.25 equiv. of Ca(NO₃)₂·4H₂O in MeCN/MeOH (20 ∶ 1 v/v) gave a dark brown solution. This was filtered, and the filtrate slowly concentrated by evaporation to give brown crystals of [Mn₁₃Ca₂O₁₀(OH)₂(OMe)₂(O₂CPh)₁₈(H₂O)₄]·10MeCN (1·10MeCN) in 40% yield. The same product was obtained when Ca(ClO₄)₂ or Ca(O₂CPh)₂ was used. Complex 1·10MeCN crystallizes‡ in the triclinic space group P1̄ with the cluster on an inversion centre. The structure (Fig. 1) consists of a [Mn₁₃Ca₂]⁴²⁺ core held together by 4 μ₃-O²⁻, 6 μ₄-O²⁻, 2 μ₃-HO⁻ and 2 μ₃-MeO⁻ ions. It can be described as two Mn₄O₄ cubes attached to a central, planar Mn₃O₄ unit, to which are also attached two Mn–Ca pairs, one above and one below the plane. The cluster is mixed-valent (MnIV, 10 MnIII, 2 MnII) with the MnIV in the centre (Mn2) and the MnII next to the Ca²⁺ ions (Mn3, Mn3a). Peripheral ligation is provided by 14 μ-, 2 μ₃- and 2 unusual η³, η¹, μ₄- bridging benzoate groups, as well as terminal H₂O molecules on Mn6, Mn6a, Mn7, and Mn7a. The metal oxidation states and the protonation levels of O²⁻, MeO⁻ and HO⁻ ions were established by bond-valence sum calculations, charge considerations, inspection of metric parameters and the identification of Mn^III Jahn–Teller (JT) elongation axes on all Mn except Mn2, Mn3 and Mn3a. All Mn and the Ca²⁺ ions are six- and eight-coordinate, respectively; seven of the eight Ca–O bonds are in the range 2.297–2.770 Å, but the eighth is longer (Ca1–O2 = 3.039 Å). The Mn3–O5 bond is also very weak (2.998 Å), presumably due to strain in the molecule.

Fig. 1 The centrosymmetric structure of 1 (with the benzoate rings omitted for clarity, except for the ipso C atoms) and its core; the dashed lines are the weak Mn–O contacts. Colour code: Ca cyan; MnIV blue; MnIII violet; MnII yellow; O red; C grey.

The solid state dc magnetic susceptibility (χ_M) of 1 was measured in the 5.00–300 K range in a 1 kG (0.1 T) field and plotted as χ_MTvs.T in Fig. 2. The χ_MT value at 300 K of 30.96 cm³ mol⁻¹ K is much lower than the 40.63 cm³ mol⁻¹ K value expected for a cluster of MnIV, 10 MnIII, 2 Mn^II non-interacting ions, indicating the presence of strong antiferromagnetic exchange interactions within complex 1. The χ_MT value steadily decreases with decreasing temperature to 3.47 cm³ mol⁻¹ K at 5.00 K, consistent with a small ground-state spin (S) in the S = 5/2 ± 1 region. In order to identify the ground state, magnetization (M) data were collected in the 0.5–5 T and 1.8–8 K ranges. Fitting of the data gave S = 5/2, g = 1.86(5) and D = −1.63(25) cm⁻¹, where D is the axial zero-field splitting parameter.† This ground state S value is supported by ac susceptibility measurements, whose real (in-phase) χ_M′T signal of ∼3.3 cm³ mol⁻¹ K is consistent with S = 5/2 and g < 2, as expected for Mn.

Fig. 2 χ _M T vs. T plot for complex 1.

The primary purpose of this communication is to report the development of higher oxidation state Mn–Ca cluster chemistry. The Mn₁₃Ca₂ nuclearity of 1 is, of course, much higher than that found in the biological site (Mn₄Ca).⁵ Nevertheless, it hasn’t escaped our notice that within the structure of 1 is a sub-unit whose structure is intriguing with respect to the WOC, in that it possesses the Mn₃CaO₄ distorted-cubane topology of the native site. In fact there are two types of cubane-containing Mn₄Ca moieties within 1, and they are shown in Fig. 3. One has the Ca²⁺ ion within the cube, and the other has it outside. The former, on the left of Fig. 3, is the one relevant to the WOC site. The Mn3⋯O5 contact is weak (2.998 Å), undoubtedly because Mn3 is also attached to O8 from a different unit, but the distorted, heteronuclear Mn₃CaO₄ cubane arrangement is nevertheless clear, its first observation outside the WOC. The Ca–Mn separations are in the range 3.505–3.961 Å, and these would undoubtedly decrease in a discrete version of this unit; for comparison, calcium EXAFS (extended X-ray absorption fine structure) data of the latter indicate the Mn⋯Ca separation in the WOC to be ∼3.4 Å.¹⁰ On the right is a related structure, but with the Ca²⁺ ion attached to an O atom of a Mn₄O₄ cube. Such cubes were never serious structural candidates for the WOC site, based on EXAFS data.¹

Fig. 3 Two cubane sub-units within 1. On the left is the Mn₄CaO₄ unit, as found in the WOC; the dashed line indicates the weak Mn–O bond. On the right is a Ca²⁺ ion attached to a Mn₄O₄ cubane.

In summary, the initial example of a higher oxidation state Mn–Ca cluster containing intimately associated Mn and Ca ions has been synthesized. From a bioinorganic perspective, the pentadecanuclearity of 1 is too high to model the WOC, but we nevertheless believe this work provides a foundation from which to tackle the synthesis of the native Mn₄Ca unit responsible for the binding and oxidative coupling of two water molecules to O₂ gas. In this regard, the occurrence within 1 of sub-units with a recognizable structural relationship to the WOC site is encouraging, and suggests that modification of the reaction system to foster lower nuclearity products might possibly yield Mn₃CaO₄ cubane-containing products. An obvious modification is the addition of chelates to the reaction that yields 1, or alternatively to treat pre-formed 1 with chelates to break it up into smaller nuclearity products. Note that the reaction system for 1 contained no such chelate to limit nuclearity growth, and that it is thus not surprising that a large nuclearity product was obtained. Both the described strategies involving chelates, and others, to obtain lower nuclearity Mn–Ca products are currently under investigation.

This work was supported by the National Science Foundation.

Notes and references

1. V. K. Yachandra, K. Sauer and M. P. Klein, Chem. Rev., 1996, 96, 2927.
2. V. K. Yachandra and K. Sauer, Biochim. Biophys. Acta, 2004, 1655, 140.
3. B. Kok, B. Forbrush and M. McGloin, Photochem. Photobiol., 1970, 11, 457.
4. R. J. Debus, Biochim. Biophys. Acta, 1992, 1102, 269; W. Ruttinger and G. C. Dismukes, Chem. Rev., 1997, 97, 1.
5. K. N. Ferreira, T. M. Iverson, K. Maghlaoui, J. Barber and S. Iwata, Science, 2004, 303, 1831; S. Iwata and J. Barber, Curr. Opin. Struct. Biol., 2004, 14, 447.
6. I. Gil de Muro, M. Insausti, L. Lezama, M. K. Urtiaga, M. I. Arriortua and T. Rojo, J. Chem. Soc., Dalton Trans., 2000, 3360.
7. S. M. Gorun, R. T. Stibrany and A. Lillo, Inorg. Chem., 1998, 37, 836.
8. G. Christou, Acc. Chem. Res., 1989, 22, 328 and references therein.
9. M. W. Wemple, H.-L. Tsai, S. Wang, J.-P. Claude, W. E. Streib, J. C. Huffman, D. N. Hendrickson and G. Christou, Inorg. Chem., 1996, 35, 6450.
10. R. M. Cinco, K. L. M. Holman, J. H. Robblee, J. Yano, S. A. Pizarro, E. Bellacchio, K. Sauer and V. K. Yachandra, Biochemistry, 2002, 41, 12928.
11. PLATON: L. Spek, Acta. Crystallogr., Sect. A, 1990, 46, 1–34.

---

Footnotes

† Electronic supplementary information (ESI) available: fits of magnetization data. See http://www.rsc.org/suppdata/cc/b4/b413680b/

‡ Vacuum-dried solid analyzed (C, H, N) as solvent-free. Crystal structure data for complex 1·10MeCN: C₁₄₈H₁₃₆Ca₂Mn₁₃N₁₀O₅₄, M_r = 3713.05, triclinic, P1̄, a = 15.0839(17), b = 16.3794(19), c = 17.959(2) Å, α = 112.343(2), β = 103.301(2), γ = 92.272(2)°, Z = 1, V = 3953.9(8) ų, d_calc = 1.559 g cm⁻³, T = 173 K. The structure was solved by direct methods in SHELXTL6 and refined on F² using full-matrix least squares. The 10 MeCN molecules were disordered and could not be modelled properly, hence the program SQUEEZE¹¹ was used to calculate the solvent disorder area and remove its contribution to the overall intensity data. The final refinement parameters R1 and wR2 were 9.06 and 21.68%, respectively. CCDC 249919. See http://www.rsc.org/suppdata/cc/b4/b413680b/ for crystallographic data in .cif or other electronic format.

---