Oxygen uptake in complexes related to [NiFeS]- and [NiFeSe]-hydrogenase active sites

Abstract

A biomimetic study for S/Se oxygenation in Ni(μ-EPh)(μ-SN₂)Fe, (E = S or Se; SN₂ = Me-diazacycloheptane-CH₂CH₂S); Fe = (η⁵-C₅H₅)Fe^II(CO) complexes related to the oxygen-damaged active sites of [NiFeS]/[NiFeSe]-H₂ases is described. Mono- and di-oxygenates (major and minor species, respectively) of the chalcogens result from exposure of the heterobimetallics to O₂; one was isolated and structurally characterized to have Ni–O–Se_Ph–Fe–S connectivity within a 5-membered ring. A compositionally analogous mono-oxy species was implicated by ν(CO) IR spectroscopy to be the corresponding Ni–O–S_Ph–Fe–S complex; treatment with O-abstraction agents such as P(o-tolyl)₃ or PMe₃ remediated the O damage. Computational studies (DFT) found that the lowest energy isomers of mono-oxygen derivatives of Ni(μ-EPh)(μ-SN₂)Fe complexes were those with O attachment to Ni rather than Fe, a result consonant with experimental findings, but at odds with oxygenates found in oxygen-damaged [NiFeS]/[NiFeSe]-H₂ase structures. A computer-generated model based on substituting ⁻SMe for the N-CH₂CH₂S⁻ sulfur donor of the N₂S suggested that constraint within the chelate hindered O-atom uptake at that sulfur site.

---

Introduction

Hydrogenases are metalloenzymes that catalyze reversible H₂ production from protons and electrons. Likely originating in pre-biotic ages and under a reducing atmosphere, the active sites contain iron, nickel and sulfur, along with simple diatomic ligands, CO and CN, in optimal arrangements that produce superb biocatalysts found throughout nature.¹ These structures offer guidance for design of molecular catalysts comprising earth abundant metals for application in electrolyzers and fuel cells.^2,3 While the protein superstructures protect the biological redox centers from O₂ as an oxidant that competes with protons, there is ample evidence, including protein crystal structures, of the detrimental effects of O₂ invasion into the active site. Indeed the initial understanding of the [NiFeS]-H₂ase was plagued with Ni-based EPR signals from various deactivated enzyme states damaged by partial oxidation or oxygenation. Vague terms such as “ready” and “unready” states, with recovery from oxygen damage on the order of seconds vs. hours, respectively, were adopted in attempts to express reductive (added H₂ or H⁺ + e⁻) reactivation requirements.^4,5

As shown in Fig. 1 two subclasses exist for the [NiFe]-H₂ase enzymes. The predominant form, [NiFeS]-H₂ase, contains a set of four cysteines around Ni; two Cys-S bridge Ni to an Fe^II(CO)(CN)₂ unit. A second form, [NiFeSe]-H₂ase is identical to the [NiFeS]-H₂ase except a terminal cysteine is replaced by selenocysteine. As the incorporation of selenocysteine requires an intricate dedicated biosynthetic machinery, as well as a high energetic cost, this form is found in few microorganisms; an example is Methanococcus voltae.^6,7

Fig. 1 Reduced (middle) active sites of [NiFe]-H₂ases and selected oxidized active sites of [NiFeS] (left) and [NiFeSe]-H₂ases (right).^8,9,15,16

Despite sharing similar active site structures, [NiFeSe]-H₂ase is deemed superior to the all-sulfur, [NiFeS]-H₂ase, with respect to H₂ production and O₂ tolerance, the latter defined as resistance to O₂ and recovery of full activity on appropriate reductive treatment.^4,8,9 Fundamental differences between sulfur and selenium include the higher acidity of Se-Cys, which might account for the proton reduction disparities, while the low redox potential and high nucleophilicity of selenium in selenolates relative to sulfur in thiolates might account for the higher oxygen-tolerance of [NiFeSe]-H₂ase.⁷

Both the ready and unready states of [NiFeS]-H₂ase are identified as containing Ni^III; the ready state contains a hydroxide bridge between Ni and Fe, (Ni–B), while the unready state is proposed to contain a κ₁-peroxide bridge or a sulfenate within a 5-membered Ni–O–S_cys–Fe–S_cys ring (Ni–A), as depicted in Fig. 1.^10,11 However, in [NiFeSe]-H₂ase, there are thus far revealed only rapid recovery or ready states comprised of oxygenates of Se- and S-cysteines; selected structures are shown in Fig. 1.⁹ Notably, multiple levels of oxidized chalcogenide sites, both S and Se, have been uncovered in [NiFeSe]-H₂ase; nevertheless the basic NiFe core structure is maintained and reductive repair is possible. As these oxygenated chalcogenides play vital roles in the reversibility of oxygen damage, small molecule analogues are needed to explore aerobic damage and repair, both for insight into the enzyme as well as the design of robust synthetic catalysts. Examples are limited.

Ogo, et al. isolated a high valent iron(IV) peroxo complex on reacting solvent-coordinated complexes [Ni^IILFe^II(RCN)(η⁵-C₅Me₅)]⁺ with O₂.^12–14 However, no oxygenated sulfur species were reported from their system. The first S-oxygenated [NiFe] complex was reported by Driess's lab as a model for sulfenate intermediates in O₂-tolerant hydrogenase; it was synthesized from FeBr₂ and the pre-formed sulfenato nickel complex¹⁷ rather than direct oxygenation.¹⁸

Herein, we describe a biomimetic study for S/Se oxygenation in complexes Ni(μ-EPh)(μ-SR′)Fe, (E = S or Se; R′ = N₂S) with certain features of the [NiFeS]/[NiFeSe]-H₂ases' active sites. We report the first XRD structure of an oxygenate of selenium within a Ni–Fe complex and use it as a reference point for DFT computational analyses and predictions.

Results and discussion

Synthesis and characterizations

Nickel dithiolate complexes such as NiN₂S₂ are well known to react as metalloligands by formation of Ni-(μ₂-SR)₂-M′ bridges.¹⁹ Adopting a [NiN₂S]₂²⁺ bimetallic nickel dimer²⁰ as platform for dimer cleavage reactions,^19,21,22 we have derived monomeric Ni(N₂S)(EPh) (E = S and Se, complexes A and B, respectively). These cis-dichalcogenides are subsequently used as metalloligands to an iron receiver unit, (η⁵-C₅H₅)Fe^II(CO)⁺, in analogy to well-known NiN₂S₂ derivatives.²³ The synthetic scheme shown in Fig. 2 is further detailed (mass spectra, elemental analyses, ¹H and ¹³C NMR spectra, and XRD crystal structures) in the ESI.† The lower value of ν(CO) (1934 cm⁻¹) in the NiSe_PhFe⁺, complex 2, as contrasted to NiS_PhFe⁺, complex 1 (1939 cm⁻¹), is attributed to the better electron donor properties of Se over S,⁷ resulting in better pi-backbonding from Fe^II to the CO reporter ligand.

Fig. 2 Synthetic scheme for NiE_PhFe⁺ complexes 1 and 2. The ν(CO) IR values of the products recorded in CH₂Cl₂.

X-ray quality crystals of monomeric Ni complex A were obtained by diethyl ether vapor transfer into a solution of CH₃CN. Complex B and the Ni–Fe complexes (1, 2 and 2′), were crystallized from a pentane-layered CH₂Cl₂ solution at −35 °C. Polymorphs, 2 and 2′, are distinguished by the orientation of the SePh planes in the NiSe_PhFe⁺ complexes. Dark red needle crystals of 2 exist in a pbca space group; 2′, as dark red blocks, is in the P2₁/c space group.

Molecular structures determined by XRD are unexceptional, Fig. 3. The >3.0 Å distances between Ni and Fe are beyond bonding. The NiE_PhFe⁺ complex structures find a typical piano-stool geometry exists about the [(η⁵-C₅H₅)Fe(CO)]⁺ unit, and, as usual for bridging dithiolates, butterfly-like cores are due to the lone pair-imposed steric requirement of the chalcogenide bridges. Hinge angles, defined as the intersection of the best N₂SE plane with the SEFe plane, are ca. 140–145°. The selenium-containing complexes, 2 and 2′ contain hinge angles slightly larger than found in 1.

Fig. 3 Molecular structures of A, B and 1 (NiS_PhFe), 2 (NiSe_PhFe), determined by single-crystal XRD, with the BF₄⁻ ions and H atoms omitted. E in A and 1 is sulfur; E in B and 2 is selenium.

Reactivity

The chemical responsiveness of the heterobimetallic complexes was illustrated by reaction with CO(g) which, in both the NiS_PhFe⁺ and NiSe_PhFe⁺ complexes, resulted in Ni–Fe splitting by transferring the EPh to Fe, generating (η⁵-C₅H₅)Fe(CO)₂EPh, and returned the Ni to its [NiN₂S]₂²⁺ dimeric form. On mixing complex 1 with an excess of Na⁺SePh⁻ in 5 : 1 CH₂Cl₂/MeOH, complex 2 is formed over the course of hours. The opposite, i.e., an attempt to replace the bridging SePh⁻ in complex 2 by SPh⁻, was unsuccessful. These reactions emphasize the mobility of the EPh⁻ unit in the NiE_PhFe⁺ complex.

At 22 °C, O₂ gas was bubbled through CH₂Cl₂ solutions of heterobimetallic complexes 1 and 2/2′; the reactions were monitored by FTIR spectroscopy. The clear reddish black solution of NiSe_PhFe⁺ developed into a slightly cloudy, dark red solution after 2 h, with ν(CO) shifting from a sharp band at 1934 cm⁻¹ to a broader band at 1954 cm⁻¹, designated as complex 2+O (Fig. 4). On filtration, the mother liquor mainly contained complex 2+O and, on removal of solvent and redissolving in CH₂Cl₂, a high resolution ESI-MS⁺ showed the parent molecular ion to have m/z 552.9634 (calc. for [2+O]⁺, 552.9661). This mono-oxygenate was isolated in ca. 50% yield (Fig. S20†). The uptake of two oxygen atoms was additionally indicated by a minor signal at m/z 568.9419 (calc. for [2+2O]⁺, 568.9611). The residue, which was soluble in CH₃CN, was confirmed as the [NiN₂S]₂²⁺ dimer by XRD with crystals obtained by ether diffusion; it was isolated in 35% yield. A remaining insoluble yellow solid was found to be an iron(III) oxide complex (see ESI†). Reactions performed in air instead of pure O₂ yielded (more slowly) the same products.

Fig. 4 Reactions of 1 and 2 in the presence of 1 atm O₂ in CH₂Cl₂ at 22 °C; conversions determined by ν(CO) analysis.

The O₂-bubbled CH₂Cl₂ solution of the all-sulfur analogue NiS_PhFe⁺, 1, required 7 h to completion, indicated by the decrease in IR absorption at ν(CO) 1939 cm⁻¹ as the oxidized species appeared as a broad band at ν(CO) 1964 cm⁻¹ (Fig. 4). The reactivity (time) difference is consistent with the fact that S²⁻ is oxidized with more difficulty than Se²⁻.⁷ The high resolution ESI-MS⁺ of products from 1 and O₂ in CH₃CN indicated the uptake of one, two, and three oxygens (Fig. S19†). Based on DFT calculations (vide infra), complex 1+O is of a similar structure as 2+O. Both 1+O and 2+O show sharp signals in their NMR spectra and are EPR-silent, indicating that Ni and Fe are low spin, +2 species (Fig. S30, S31 and S38†). To confirm that the 1+O and 2+O were derived from molecular O₂, the same reactions were carried out with added ¹⁸O₂. Mass spectral analysis indicated the uptake of labelled O-18, Fig. S43 and S44.†

Crystals of complex 2+O were obtained by layering pentane on DCM solutions of 2+O. As seen in Fig. 5, complex 2+O contains a 5-membered, puckered NiOSeFeS ring; the Ni(II) is the center of a square planar N₂SO binding site. From XRD the bond distances of Ni–S and Fe–S in 2+O are largely the same as in the reduced form 2 (or 2′); the oxygen insertion into the Ni–Se bond results in an Ni–O bond at 1.870(2) Å. The Fe–Se bond length becomes slightly shorter in 2+O, 2.333(1) Å, as compared to 2, 2.406(1) Å. The distance between Ni and Fe in 2+O is 3.568 Å, ca. 0.3 Å longer than in the reduced form, reflecting the expansion of the ring. Note that the Ni–A form of the [NiFe]-H₂ase enzyme active site with sulfenate also has a 5-membered NiOSFeS ring, Fig. 1.²⁴ The Ni⋯O distance in Ni–A is 1.824 Å, while in 2+O it is 1.870(2) Å. In contrast, a minor fraction from [NiFeSe]-H₂ase oxidation reported by Volbeda et al. has a different connectivity, NiSeOFe, Fig. 1.⁹ Complex 2+O is, to our knowledge, the first model of an oxygenated [NiFeSe]-H₂ase synthetic analogue derived from direct reaction.

Fig. 5 Overlay of experimental and computational structures of 2 and 2+O along with selected geometric parameters. Hydrogens deleted for clarity.

As a comparison to the NiE_PhFe complexes 1 and 2, constrained NiN₂S₂ derivatives, [NiN₂S₂·Fe(η⁵-C₅H₅)(CO)]⁺[BF₄]⁻ and [NiN₂S₂·Fe(η⁵-C₅Me₅)(CO)]⁺[BF₄]⁻ N₂S₂ = bismercaptoethanediazacycloheptane) were tested for reactivity with oxygen.^25,26 After stirring their solutions in an O₂ atmosphere for several days, they remain intact with no indication of reaction. We surmise that the CO which is bound to Fe prevents O₂ activation at the Fe center and both sulfur and nickel are deactivated towards O₂ reactivity in the rigid chelating N₂S₂ ligands. However, in complexes 1 and 2, the mobility at EPh (E = S or Se) provides a potential site for O₂ attack on the Ni or Fe, consistent with the splitting of 1 and 2 by CO(g) (Fig. S33†). For comparison, elevated temperature (70 °C) and 11 bar CO(g) are required to break one Ni–S bond in [NiN₂S₂·Fe(η⁵-C₅H₅)(CO)]⁺[BF₄]⁻.²⁵

Computational section

The crystal structures of 2 and 2+O were used as geometric starting points for DFT calculations (TPSSTPSS functional with the 6-311++G(d,p) basis set on all nonmetal atoms and the 6-311+G basis set for Ni and Fe atoms) and then optimized using the Gaussian 09 program using the SMD implicit solvation model with acetonitrile as the solvent.²⁸ Cartesian coordinates of calculated structures are given in the ESI.† The calculated structures for 2 and 2+O showed excellent agreement with the corresponding experimental structures, Fig. 5. Similar structures were found for the sulfur analogues, 1 and 1+O.

The structure of 2+O was investigated to determine contributing factors to the thermodynamically favored oxygenation product. A less rigid model of 2+O was made by breaking the carbon–carbon bond linking the sulfur donor to the neighbouring nitrogen donor, changing the resulting methylene fragments into methyl groups, and replacing the phenyl group on Se-Ph with a methyl group. This bidentate N₂S_MeSe_Me model, with unconnected SMe, is designated as 3 and is designed to create similar electronic, steric environments and constraints around the sulfur and selenium atoms.

Four oxygenated isomers, 3+O_a–d, as well as four corresponding isomers, 2+O_a–d, were calculated in which the location of the oxygen atom was varied according to the scheme described in Fig. 6. The isomer 2+O_a, which corresponds to the experimental structure, is indeed the lowest energy isomer of 2+O; the next lowest isomer, 2+O_b, is 8.6 kcal higher in energy. This contrasts with the N₂S_MeSe_Me model in which the lowest energy isomer, 3_d, has the oxygen atom inserted between the nickel and the sulfur atoms rather than the nickel and selenium. The corresponding isomer 2+O_d is the highest energy isomer for the tridentate model and the oxygen atom appears in a position bridging the nickel sulfur bond rather than full insertion. This is likely due to the geometric restraints imposed by the tridentate N₂S ligand that disfavor the addition of the oxygen atom into its rigid binding arrangement. This constraint is not found in the more open bidentate model with separated SMe donors. There it is found that formation of a sulfur–oxygen bond is thermodynamically favored, by 3.2 kcal mol⁻¹, as compared to the formation of a selenium–oxygen bond.

Fig. 6 Relative energies of oxygenated isomers of N₂SSe_Ph model, 2+O, and the bidentate, separated donors in the N₂S_MeSe_Me model, 3+O, in kcal mol⁻¹. A range of functionals and basis sets were explored and found to elicit minor differences on energies of complexes but do not change the ordering of isomers' energies (Fig. S41†).

O-atom removal results

As arbiters between O₂ and the [NiFe] or [NiFeSe]-H₂ase active sites, chalcogenides prevent complete degradation of the organometallic active sites of hydrogenases, and perform this role reversably.^24,27 To explore the possibility of reversal of oxygenated selenate/sulfenate, O-abstracting agents, PR₃ (R = Me or o-tolyl), were employed for O-atom removal (Fig. 7). The ν(CO) monitor of the reaction of 2+O and 1 equiv. of PMe₃ in DCM showed a band shift from 1954 to 1934 cm⁻¹. The latter absorption is made up from a mixture of complex 2 and the PMe₃ species (η⁵-C₅H₅)Fe(SePh)(PMe₃)(CO), which is confirmed by the ⁺ESI-MS data (Fig. S37†). The reaction of a mixture of 1+O and 1+2O with 1 equiv. of PMe₃ resulted in partial conversion of 1+O to 1, finding that 1+2O was unchanged, even in presence of excess PMe₃. The ⁺ESI-MS (Fig. S36†) indicates the formation of reduced complex 2, unreacted 1+2O and the PMe₃ derivative (η⁵-C₅H₅)Fe(SPh)(PMe₃)(CO). Removal of the O-atom from 1+O and 2+O by reducing agents containing H⁻, as well as electrochemically (e⁻ + H⁺), have thus far been unsuccessful.

Fig. 7 Reaction of complex 2+O (E = Se) or 1+O (E = S) with O-abstracting agents, PR₃ (R = Me or o-tolyl).

Conclusion

In summary, we have observed oxygenation and O-atom removal from two biomimetic complexes with features of the [NiFeS]/[NiFeSe]-H₂ase active sites. Our study highlights the usefulness of the [NiN₂S]₂²⁺ synthetic platform for generating NiFe biomimetic complexes. Clearly the NiE_PhFe⁺ complex differs from the active site of [NiFeSe]-H₂ase as the selenium in the model is in a bridging position, rather than terminal as nature has adopted. The relative reactivities however are consistent with what is found in nature. As compared to complex 1, the selenium-bridged complex 2 required a shorter time to generate oxygenated selenium, 2+O. The partial conversion of oxygenated chalcogenides (Se or S oxidation states of 0) back to reduced S/Se (oxidation state −2) in this study provides a foundation for the mechanism of reactivation of S/Se-oxygenated [NiFeE]-H₂ases. Oxygen-uptake by S/Se in [NiFeS]- and [NiFeSe]-H₂ases results in modification of the active site, but less severe than would result in irreparable degradation.²⁷ During the oxygenation and O-atom removal from the chalcogen atom, we note that the mobility in the Ni–E bond plays a vital role.

The variety of oxygenates thus far found in nature, Fig. 1, speak to the ability of the chalcogens to maintain a close attachment to the NiFe site. Detailed O₂/H₂ase active site reaction mechanisms remain obscure but of great interest.^29,30 Our studies of model complexes point to the possibility of opening the Ni–E bond in the Ni–(μE_Ph)–Fe unit. Such bond cleavage exposes a reactive lone pair on sulfur or selenium that attracts the electrophilic O₂ molecule concomitant with providing an open site on Ni²⁺ for assisted activation, leading to the product. A related strategy for protection of the active site from oxidative damage is seen in the MBH (membrane-bound respiratory [NiFeS]-hydrogenase) wherein the proximal [4Fe-3S] cluster donates two electrons and one proton for reduction of adventitious O₂.³¹

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was funded by the National Science Foundation (CHE-1300787, CHE-1664866 to M. B. H. and CHE-1266097, CHE-1665258 to M. Y. D.) and the Robert A. Welch Foundation (A-0648 to M. B. H. and A-0924 to M. Y. D.).

Notes and references

1. W. Lubitz, H. Ogata, O. Rudiger and E. Reijerse, Chem. Rev., 2014, 114, 4081.
2. L. L. Beer, E. S. Boyd, J. W. Peters and M. C. Posewitz, Curr. Opin. Biotechnol., 2009, 20, 264.
3. M. Hambourger, M. Gervaldo, D. Svedruzic, P. W. King, D. Gust, M. Ghirardi, A. L. Moore and T. A. Moore, J. Am. Chem. Soc., 2008, 130, 2015.
4. C. S. A. Baltazar, M. C. Marques, C. M. Soares, A. M. DeLacey, I. A. C. Pereira and P. M. Matias, Eur. J. Inorg. Chem., 2011, 948.
5. D. W. Wakerley and E. Reisner, Energy Environ. Sci., 2015, 8, 2283.
6. P. M. Vignais and B. Billoud, Chem. Rev., 2007, 107, 4206.
7. H. J. Reich and R. J. Hondal, ACS Chem. Biol., 2016, 11, 821.
8. M. C. Marques, R. Coelho, A. L. De Lacey, I. A. Pereira and P. M. Matias, J. Mol. Biol., 2010, 396, 893.
9. A. Volbeda, P. Amara, M. Iannello, A. L. De Lacey, C. Cavazza and J. C. Fontecilla-Camps, Chem. Commun., 2013, 49, 7061.
10. M. Horch, L. Lauterbach, M. A. Mroginski, P. Hildebrandt, O. Lenz and I. Zebger, J. Am. Chem. Soc., 2015, 137, 2555.
11. H. Ogata, W. Lubitz and Y. Higuchi, Dalton Trans., 2009, 7577.
12. M. Isegawa, A. K. Sharma, S. Ogo and K. Morokuma, Organometallics, 2018, 37, 1534.
13. T. Kishima, T. Matsumoto, H. Nakai, S. Hayami, T. Ohta and S. Ogo, Angew. Chem., Int. Ed., 2016, 55, 724.
14. T. Matsumoto, T. Kishima, T. Yatabe, K.-S. Yoon and S. Ogo, Organometallics, 2017, 36, 3883.
15. E. Garcin, X. Vernede, E. Hatchikian, A. Volbeda, M. Frey and J. Fontecilla-Camps, Structure, 1999, 7, 557.
16. A. Volbeda, M.-H. Charon, C. Piras, E. C. Hatchikian, M. Frey and J. C. Fontecilla-Camps, Nature, 1995, 373, 580.
17. C. A. Grapperhaus and M. Y. Darensbourg, Acc. Chem. Res., 1998, 31, 451.
18. N. J. Lindenmaier, S. Wahlefeld, E. Bill, T. Szilvasi, C. Eberle, S. Yao, P. Hildebrandt, M. Horch, I. Zebger and M. Driess, Angew. Chem., Int. Ed., 2017, 56, 2208.
19. J. A. Denny and M. Y. Darensbourg, Chem. Rev., 2015, 115, 5248.
20. R. M. Jenkins, M. L. Singleton, L. A. Leamer, J. H. Reibenspies and M. Y. Darensbourg, Inorg. Chem., 2010, 49, 5503.
21. E. M. Gale, D. M. Cowart, R. A. Scott and T. C. Harrop, Inorg. Chem., 2011, 50, 10460.
22. D. Huang, L. Deng, J. Sun and R. H. Holm, Inorg. Chem., 2009, 48, 6159.
23. D. Brazzolotto, M. Gennari, N. Queyriaux, T. R. Simmons, J. Pecaut, S. Demeshko, F. Meyer, M. Orio, V. Artero and C. Duboc, Nat. Chem., 2016, 8, 1054.
24. A. Volbeda, L. Martin, E. Barbier, O. Gutiérrez-Sanz, A. L. D. Lacey, P.-P. Liebgott, S. b. Dementin, M. Rousset and J. C. Fontecilla-Camps, J. Biol. Inorg Chem., 2014, 20, 11.
25. S. Ding, P. Ghosh, A. M. Lunsford, N. Wang, N. Bhuvanesh, M. B. Hall and M. Y. Darensbourg, J. Am. Chem. Soc., 2016, 138, 12920.
26. P. Ghosh, M. Quiroz, N. Wang, N. Bhuvanesh and M. Y. Darensbourg, Dalton Trans., 2017, 46, 5617.
27. K. A. Vincent, A. Parkin, O. Lenz, S. P. J. Albracht, J. C. Fontecilla-Camps, R. Cammack, B. Friedrich and F. A. Armstrong, J. Am. Chem. Soc., 2005, 127, 18179.
28. M. J. Frisch; G. W. Trucks; H. B. Schlegel; G. E. Scuseria; M. A. Robb; J. R. Cheeseman; G. Scalmani;V. Barone; B. Mennucci; G. A. Petersson; H. Nakatsuji; M. Caricato; X. Li; H. P. Hratchian; A. F. Izmaylov; J. Bloino; G. Zheng; J. L. Sonnenberg; M. Hada; M. Ehara; K. Toyota; R. Fukuda; J. Hasegawa; M. Ishida; T. Nakajima; Y. Honda; O. Kitao; H. Nakai; T. Vreven; J. A. Montgomery Jr; J. E. Peralta; F. Ogliaro; M. Bearpark; J. J. Heyd; E. Brothers; K. N. Kudin; V. N. Staroverov; R. Kobayashi; J. Normand; K. Raghavachari; A. Rendell; J. C. Burant; S. S. Iyengar; J. Tomasi; M. Cossi; N. Rega; J. M. Millam; M. Klene; J. E. Knox; J. B. Cross; V. Bakken; C. Adamo; J. Jaramillo; R. Gomperts; R. E. Stratmann; O. Yazyev; A. J. Austin; R. Cammi; C. Pomelli; J. W. Ochterski; R. L. Martin; K. Morokuma; V. G. Zakrzewski; G. A. Voth; P. Salvador; J. J. Dannenberg; S. Dapprich; A. D. Daniels; O. Farkas; J. B. Foresman; J. V. Ortiz; J. Cioslowski; D. J. Fox, Gaussian 09, revision D.01, Gaussian, Inc., Wallingford, CT, 2009.
29. S. Ogo, Coord. Chem. Rev., 2017, 334, 43.
30. S. Ogo, Chem. Rec., 2014, 14, 397.
31. Y. Shomura, K. S. Yoon, H. Nishihara and Y. Higuchi, Nature, 2011, 479, 253.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 1861574, 1861577–1861580 and 1861890. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c8sc04436h

---