Aza- and Oxadithiolates Are Proton Relays in Functional Models for the [FeFe]-Hydrogenases

Abstract

The dithiolate cofactor for the [FeFe]-hydrogenase models, Fe₂(xdt)(CO)₂(dppv)₂ (where xdt = 1,3-propanedithiolate (pdt), azadithiolate (adt), (SCH₂)₂NH, and oxadithiolate (odt), (SCH₂)₂O; dppv = cis-1,2-bis-(diphenylphosphino)ethylene) have been probed for their functionality as proton relays enabling formation and deprotonation of terminal hydrides. Compared to the propanedithiolate derivative, the azadithiolate and oxaditiholate show enhanced rates of proton transfer between solution and the terminal site on one Fe center. The results are consistent with the heteroatom of the dithiolate serving a gating role for both protonation and deprotonation. The pK_a of the transiently formed ammonium (pK ^CD₂Cl₂ 5.7–8.2) or oxonium (pK ^CD₂Cl₂ −4.7–1.6) regulates the proton transfer. As consequence, only the azadithiolate is capable of yielding the terminal hydride from weak acids. The aza- and oxadithiolates manifested the advantages of proton relays: the odt derivative proved to be a faster catalyst for hydrogen evolution than the pdt derivative as indicated from cyclic voltammetry plots of i_c/i_p vs. [H⁺]. The adt derivative was capable of proton reduction from the weak acid [HPMe₂Ph]BF₄ (pK ^CD₂Cl₂ = 5.7). The proton relay function does not apply to the isomeric bridged-hydrides [Fe₂(xdt)(μ-H)(CO)₂(dppv)₂]⁺, where the hydride is too distant and too basic to interact to be affected by the heteroatomic relay site. None of these μ-H species can be deprotonated.

The [FeFe]-hydrogenases are among the very best catalysts known for the reduction of protons to dihydrogen, with turnover frequencies estimated to be ~6000 mol H₂/mol enzyme per second operating at nearly Nerstian potentials.¹ The question about why the [FeFe]-hydrogenases are so efficient is topical,² and the answer is likely related to the incompletely characterized dithiolate cofactor that bridges the diiron subunit. In 2001, Nicolet et al. proposed that this dithiolate is the azadithiolate (adt, (SCH₂)₂NH), wherein the amine functionality could relay protons to and from the apical site on the distal Fe center.³ It is known that, unlike typical amine bases, transition metals can be slow to protonate.⁴ The adt hypothesis is attractive because it potentially shows how to couple the kinetic facility of amine protonation with the redox abilities of iron hydrides. Indeed, DuBois has demonstrated that amine bases constrained within diphosphine ligands greatly accelerate both H₂ uptake and production for mononuclear iron and nickel phosphine complexes.⁵ A recent DFT investigation suggests that the dithiolate cofactor is the oxadithiolate (odt, (SCH₂)₂O), which also merits evaluation since protein crystallography cannot distinguish between C, N, and O.⁶

Efforts to understand and exploit the possible presence of the adt cofactor have been underway for several years. The amine can be protonated independently of the diiron site.^7,8 However, N-protonation has little effect on proton reduction by the catalysts of the general type Fe₂(μ-H)(SR)₂L₆, where μ-H indicates that the hydride bridges the two Fe centers.^9,10

The recent discovery that diiron(I) dithiolates initially protonate to give terminal, not bridging, hydrides opens a new and potentially significant phase in elucidating the role of the dithiolate cofactor in the catalysis.¹¹ Terminal hydride ligands¹² would be directly adjacent to the heteroatom in the dithiolate, which is therefore well positioned as a site for proton-relay. We recently demonstrated that protonations of the electronically symmetrical Fe₂(pdt)(CO)₂(dppv)₂ (1) and Fe₂(adt)(CO)₂(dppv)₂ (2) yield relatively stable (t_1/2 ~ minutes at 25 °C), terminal hydride derivatives (dppv = cis-1,2-bis(diphenylphosphino)ethylene; pdt = 1,3-propanedithiolate; adt = 2-azapropane-1,3-dithiolate). These terminal hydrides undergo reduction at a milder potential than the isomeric bridging hydride species, are catalytically competent, and are sufficiently robust to study in detail.¹³ The crucial unanswered question is whether a heteroatom in the dithiolate participates in proton transfer to and from the terminal hydride. To address this question, we report results that demonstrate a functional role of oxa- and azadithiolates as proton relays. These experiments are benchmarked relative to the third crystallographically feasible dithiolate, propanedithiolate. We prepared Fe₂(odt)(CO)₂(dppv)₂ from the hexacarbonyl¹⁴ and confirmed spectroscopically (odt = 2-oxopropane-1,3-dithiolate).

Protonation of Fe₂(odt)(CO)₂(dppv)₂ (3) at −78 °C with the strong acid [H(Et₂O)₂]BAr ^F₄ afforded the terminal hydride [3(t -H)]BAr^F₄. ¹H and ³¹P NMR analysis confirmed that protonation occurred at a single Fe center, characteristic of related derivatives.¹⁵ This terminal hydride was found to isomerize upon warming to give the μ-hydride complex, [3(μ-H)]BAr^F₄ (2.6 × 10⁻⁴ s⁻¹, −10 °C), a process following unimolecular kinetics. The isomerization rate is similar to that for [2(t-H)]BAr^F₄ (1.4 × 10⁻⁴ s⁻¹) but is faster than [1(t-H)]BAr^F₄ (2.5 × 10⁻⁵ s⁻¹).

Odt and adt Accelerate Deprotonation of [HFe₂(xdt)(CO)₂(diphosphine)₂]⁺

We first compared the facility with which a CD₂Cl₂ solution of [3(t-H)]BAr^F₄ deprotonates. At −78 °C, [3(t-H)]BAr^F₄ is unreactive toward base, but upon warming to ~ 0 °C two products form, [3(μ-H)]BAr^F₄ and 3, as assayed by ¹H and ³¹P NMR spectroscopy. The ratio of these two products was unaffected by concentration of the base as well as its pK_a, as indicated by deprotonations with both strong and weak bases, respectively tetramethylguanidine (TMGH⁺, pK_a ~23) and PPh₃ ([HPPh₃] BF₄, pK ^CD₂Cl₂ = 1.6).¹⁶ In contrast, [1(t-H)]BAr^F₄ cannot be deprotonated by any base, even at room temperature, where isomerization to [1(μ-H)]⁺ eventually occurs. Deprotonation of [2(t-H)]BAr^F₄ is however immediate with PBu₃ ([HPBu₃]BF₄, pK ^CD₂Cl₂ = 8.2) even at −90 °C, exclusively providing 2. The close similarity of the IR spectra in the ν_CO region for [1(t-H)]BAr^F₄, [2(t-H)]BAr ^F₄, and [3(t-H)]BAr ^F₄ suggests that these terminal hydrides should have similar thermodynamic acidities.¹⁷ The similar thermodynamic acidities of these three hydrides indicate that the rate of deprotonation is strongly influenced by the presence of a heteroatom in the dithiolate (Scheme 1). Not only is deprotonation of the terminal hydrides strongly affected by the identity of the dithiolate ligand, the stereochemistry of the hydride also has a profound effect. The three bridging hydrides, [1(μ-H)]⁺, [2(μ-H)]⁺, and [3(μ-H)]⁺ are not deprotonated by NEt₃ at room temperature.

Scheme 1.

Acid-base reactions of 1–3.

Heteroatom in the Dithiolate Strongly Affects the Protonation of Fe₂(xdt)(CO)₂(diphosphine)₂

The presence of a heteroatom was found to strongly affect the rate of protonation at iron. The strong acid [H(Et₂O)₂]BAr ^F₄ protonated 1, 2, and 3 quickly at −90 °C, but the billion fold weaker acid [HPMe₂Ph]BF₄ (pK^CD₂Cl₂ = 5.7) protonated only 2 (−90 °C), not 1 or 3.¹⁶ The pK_a of [2H]⁺ is bracketed by the finding that 2 is not protonated by [HPBu₃]BF₄. The implication that the acidity of the ammonium and terminal hydride tautomers of [2H]⁺ are comparable is supported by the previously reported finding that the ratio of the ammonium and hydride tautomers can be shifted by the solvent: MeOH favors the ammonium tautomer, CH₂Cl₂ the hydride tautomer.¹⁸

These results are consistent with a mechanism whereby hydride formation is regulated by the basicity of the heteroatom in the dithiolate: the ammonium center in 2 is easily protonated and then quickly relays protons to Fe. In contrast for complexes with weakly basic oxadithiolate (pK^CD₂Cl₂(R₂OH⁺) ~−4.7 to 1.6) or nonbasic propanedithiolate, the Fe site can only be protonated by strong acids, even though the basicities of these diiron centers are very similar.

Heteroatom-Containing Dithiolates Enhance Proton Reduction Catalysis

As the azadithiolate exhibits enhanced rates of protonation, this enhancement could be manifested in catalysis by accelerating the rate of proton reduction. At −40 °C, where these terminal hydrides are stable, the hydrides [1(t-H)]BF₄, [2(t-H)]BF₄, and [3(t-H)]BF₄ all catalyze hydrogen evolution at about the same potentials, ~−1.5 V vs Fc/Fc⁺ (~−0.8 V vs NHE). Using [HPMe₂Ph]BF₄ (pK^CD₂Cl₂ = 5.7), however, [2(t-H)]BF₄ is catalytically active, but [1(t-H)]BF₄ and [3(t-H)]BF₄ are not. The pK^CD₂Cl₂ of [HPMe₂Ph]BF₄ has been estimated to correspond to an aqueous pK_a of 6.8.¹⁶ Catalysis by the amine [2(t-H)]BF₄ with strong acids is complicated because protonation occurs at both the amine and terminally at Fe.^7,10,18 Interestingly, for strong acids, [3(t-H)]BF₄ is a significantly faster catalyst for hydrogen evolution than is [1(t-H)]BF₄, which suggests that even the weakly basic ether group assists in proton relay (Figure 1).

Figure 1.

Dependence of current (i_c/i_p) vs of [HBF₄·Et₂O] for [1(t-H)]BF₄ and [3(t-H)]BF₄ (−40 °C, 1 mM catalyst), where i_c is peak catalytic current and i_p is the peak current in the absence of acid.

The results presented in this paper indicate that the presence of a heteroatom in the dithiolate bridge strongly facilitates proton transfer to and from the apical site on Fe, but only to the extent that the acid can protonate the bridgehead atom. Although both azadithiolate (in 2) and oxadithiolate (in 3) exhibit relay-like behavior, indicated by enhanced rates of proton reduction catalysis by 2 and 3, only the azadithiolate 2 enables hydride formation from weak acids, which is relevant to catalysis at low overpotentials.¹⁹

Supplementary Material

1_si_001. Supporting Information Available.

Preparative and spectroscopic details. This material is available free of charge via the Internet at http://pubs.acs.org.