The active form of Chlamydia trachomatis ribonucleotide reductase R2 protein contains a heterodinuclear Mn(IV)/Fe(III) cluster with S = 1 ground state

Abstract

The class I ribonucleotide reductase from Chlamydia trachomatis uses a stable Mn^IV/Fe^III cofactor to initiate nucleotide reduction by a free-radical mechanism. The enzyme provides the first example both of a Mn-dependent ribonucleotide reductase and of a Mn/Fe redox cofactor. In this work, we have used variable-field Mössbauer spectroscopy to demonstrate that the active cofactor has an S = 1 ground state due to antiferromagnetic coupling between the Mn^IV (S_Mn = 3/2) and high-spin Fe^III (S_Fe = 5/2) sites.

The reduction of nucleotides to 2′-deoxynucleotides by ribonucleotide reductases (RNRs)¹ begins with the abstraction of the 3′ hydrogen atom of the substrate by a transient cysteine thiyl radical (C•).^2–4 Class I, II, and III RNRs differ in subunit composition and the nature of the cofactor used to generate the 3′-H-abstracting C•.² A conventional class I RNR uses a stable tyrosyl radical (Y•), which is introduced into the enzyme’s homodimeric R2 subunit (also denoted β₂) by reaction of O₂ with an adjacent carboxylate-bridged non-heme diiron cluster,⁵ to generate the C• in its R1 subunit (also denoted α₂) via long-range (~ 35 Å) proton-coupled electron transfer (PCET).^6–8 The identification in Chlamydia trachomatis (Ct) and other species of pathogenic bacteria of RNRs having the class I subunit architecture but the Y•-harboring tyrosine replaced by phenylalanine raised the question of how such an RNR might function without the initiating Y•.^9,10 We recently showed that Ct RNR uses a stable Mn^IV/Fe^III cofactor, generated by reaction of the Mn^II/Fe^II-R2 complex with O₂, for radical initiation.¹¹ Although heterobinuclear Mn/Fe complexes of various oxidation states,^12–14 including one example of an inorganic Mn^IV/Fe^III complex,¹⁵ have been reported, Ct RNR is, to our knowledge, the first case in which an enzyme has been shown to use a Mn/Fe cluster as a redox cofactor. Here we characterize this novel cofactor by Mössbauer spectroscopy, showing that it has a triplet (S_Total = 1) ground state resulting from antiferromagnetic coupling of its Mn^IV (S_Mn = 3/2) and high-spin Fe^III (S_Fe = 5/2) constituents and providing parameters to calibrate calculation of its geometric and electronic structure.

The 4.2-K/zero-field Mössbauer spectrum of a sample prepared by reaction of a solution of R2, Mn^II, and Fe^II with O₂¹⁶ (Figure 1, hashed marks) comprises two quadrupole doublets with similar isomer shifts (δ) but different quadrupole splittings (ΔE_Q). The minor doublet (25% of the total intensity) matches the spectrum of the Fe₂^III/III cluster, which is characterized by δ = 0.50 mm/s and ΔE_Q = 0.79 mm/s (Figure 1, solid line). The major doublet (75% intensity) arises from the functional Mn^IV/Fe^III cofactor.¹¹ For in-depth characterization of this species, we collected Mössbauer spectra at 4.2 K in variable applied magnetic fields. To analyze these spectra, it was necessary first to remove the 25% contribution attributed to the Fe₂^III/III complex (Figure S1). The derived spectra of the Mn^IV/Fe^III complex are shown in Figure 2.

Figure 1.

4.2-K/zero-field Mössbauer spectrum of the final product of the reaction of Mn^II/Fe^II Ct-R2 with O₂ (hashed marks). The solid line is the spectrum of the final product of the reaction of Fe^II/Fe^II Ct-R2 with O₂ (25% of the total intensity).

Figure 2.

4.2-K Mössbauer spectra of Mn^IV/Fe^III-R2 derived by component analysis of the experimental spectra. The solid lines are simulations according to the spin Hamiltonian given in the Supporting Information and the following parameters: S_Total = 1, D_S=1 = −1.9 cm⁻¹, (E/D)_S=1 = 0.33, δ = 0.52 mm/s, ΔE_Q = 1.32 mm/s, η = −2.6, (A/g_Nβ_N)_Fe = (−40.2, −38.9, −38.0) T.

The zero-field spectrum is a sharp quadrupole doublet with parameters (δ = 0.52 mm/s and ΔE_Q = 1.32 mm/s) typical of high-spin Fe^III. Spectra recorded in applied fields were analyzed with the assumption of slow relaxation and according to the spin Hamiltonian with respect to the total-spin ground state, S_Total = 1, given in the Supporting Information. For this analysis, the parameters determined from the zero-field spectrum were fixed. The other parameters in the simulations are the asymmetry parameter, η, the axial and rhombic zero-field splitting (ZFS) parameters of the S_Total = 1 ground state, D_S=1 and (E/D)_S=1, respectively, and the hyperfine tensor for the Fe^III ion, (A/g_Nβ_N)_Fe.

The fact that the overall splitting of the spectrum is larger at 4 T than at 8 T reveals that the electronic Zeeman term dominates the ZFS interaction at these magnetic field strengths. As a consequence, the spin expectation value of the ground state (〈S〉) and the internal magnetic field [B_int = −〈S〈·(A/g_Nβ_N)_Fe] approach their limiting values (corresponding to 〈S〉 = −1) at these applied fields, and the hyperfine tensor can be determined accurately from the spectra. In a spin-coupled cluster, the A-tensor for the iron ion with respect to the S_Total = 1 ground state (A_Fe) is given by equation 1, in which c_Fe and a_Fe represent the vector coupling coefficient and the intrinsic hyperfine tensor, respectively.¹⁷

$${\mathbf{A}}_{\text{Fe}}={\text{c}}_{\text{Fe}}{\mathbf{a}}_{\text{Fe}}$$ (1)

For high-spin Fe^III sites, (a/g_Nβ_N)_Fe is dominated by the Fermi contact term and exhibits nearly isotropic values of ~ −21 T.¹⁸ For the S_Total = 1 state, c_Fe = +7/4. Thus, (A/g_Nβ_N)_Fe is expected to be nearly isotropic with values of ~ −37 T. Indeed, analysis of the spectra yielded a nearly isotropic hyperfine tensor, (A/g_Nβ_N)_Fe = (−40.2, −38.9, −38.0) T, similar to the expected value.

With the A_Fe determined from the 4-T and 8-T spectra, the ZFS-parameters were then determined accurately from the spectra with lesser applied magnetic fields, under which conditions the ZFS is significant compared to the electronic Zeeman term. Simulation of the spectra provided for determination of the magnetic-field dependence of 〈S〉, which determines the magnitude of the internal magnetic field and the resultant splitting in each spectrum. From the field-dependent spectra, D_S=1 = −1.9 cm⁻¹ and (E/D)_S=1 = 0.33 were found. D_S=1 is related to the intrinsic D-values of the Mn^IV and Fe^III ions via equation 2, with the assumption that the total spin states are well separated in energy (i.e., J ≫ D).¹⁷

$$D_{S=1}=14/5D_{\text{Fe}}+3/10D_{\text{Mn}}$$ (2)

D_S=1 is dominated by the contribution from the Fe^III site¹⁹ and can be rationalized with a value typical for high-spin Fe^III (|D_Fe| of ~ 0.7 cm⁻¹).¹⁸

The Mn^IV and Fe^III ions could conceivably couple ferromagnetically to yield an S_Total = 4 ground state. However, the inherent anisotropy of 〈S〉 of the S_Total = 4 system precludes simulation of all the experimental spectra with a single set of spin-Hamiltonian parameters. Specifically, comparison of the 4-T and 8-T spectra establishes that 〈S〉 along all three molecular axes nearly reach saturation in an applied field of ~4 T. |D_S=4| would therefore have to be small (< 0.5 cm⁻¹, see Figure S2). Conversely, with a D_S=4 of this small magnitude, 〈S_z〉 (for D_S=4 < 0) saturates at ~30 mT (Figure S2), resulting in much greater magnetic hyperfine splitting in the weak-field spectra than is observed (Figure S3). Thus, ferromagnetic coupling to yield S_Total = 4 can be ruled out.

The demonstration by Mössbauer spectroscopy that the Mn^IV/Fe^III cluster in the active form of Ct R2 has a S_Total = 1 ground state arising from antiferromagnetic coupling between the S_Mn = 3/2 and S_Fe = 5/2 ions represents an important first step toward defining its electronic structure. Moreover, parameters obtained in the analysis can now be used to calibrate density functional theory calculations to extract deeper insight.^20,21 This insight will permit an important unresolved aspect of class I RNR function – how electron transfer to the oxidized cofactor in R2 is “gated” so as to occur only in the active holoenzyme complex – to be addressed. Direct interrogation of the radical initiation step in conventional class I RNRs has proven to be extremely challenging.⁷ The R2 product of radical transfer does not accumulate during turnover because of unfavorable kinetics.^7,22 Accumulation of the reduced cofactor can be promoted by use of radical-trapping substrate analogues or R1 variants,^23–26 but extracting atomic-level insight into the changes to the cofactor upon reduction is not straightforward. The reduced Y lacks a useful spectroscopic signature. The Fe₂^III/III cluster is either unchanged by the radical-transfer step or, at most, may transfer a proton to the Y from a coordinated water ligand, a change not obviously demonstrable by the spectroscopic methods to which the EPR-silent cluster is amenable. The use of the novel cofactor by Ct RNR affords a unique opportunity to dissect the initiation step. The metal cluster undergoes reduction from Mn^IV/Fe^III to Mn^III/Fe^III, which is, by contrast to the reduced cofactor of a conventional class I RNR, EPR active.¹¹ Moreover, although the reduced form might, as in E. coli RNR, not accumulate during turnover, it can be prepared in stable form and its structure is affected by binding of R1 and nucleotides, as demonstrated by marked changes to its EPR spectrum.¹¹ The structural changes caused by formation of the complex are almost certainly elements of the gating mechanism. By defining what they are and extending the description of the geometric and electronic structure of the Mn^IV/Fe^III cofactor initiated by this work, insight into the mechanisms of radical transfer and conformational gating thereof may be obtained.

Scheme 1.

Radical-generating cofactor of Ct RNR (left) and conventional class I RNR (right).