An active dimanganese(III)-tyrosyl radical cofactor in Escherichia coli class Ib ribonucleotide reductase

Abstract

Escherichia coli class Ib ribonucleotide reductase (RNR) converts nucleoside 5'-diphosphates to deoxynucleoside 5'-diphosphates and is expressed under iron-limited and oxidative stress conditions. This RNR is composed of two homodimeric subunits: alpha2 (NrdE), where nucleotide reduction occurs, and beta2 (NrdF), which contains an unidentified metallocofactor that initiates nucleotide reduction. nrdE and nrdF are found in an operon with nrdI, which encodes an unusual flavodoxin proposed to be involved in metallocofactor biosynthesis and/or maintenance. Ni affinity chromatography of a mixture of E. coli (His)(6)-NrdI and NrdF demonstrated tight association between these proteins. To explore the function of NrdI and identify the metallocofactor, apoNrdF was loaded with Mn(II) and incubated with fully reduced NrdI (NrdI(hq)) and O(2). Active RNR was rapidly produced with 0.25 +/- 0.03 tyrosyl radical (Y*) per beta2 and a specific activity of 600 units/mg. EPR and biochemical studies of the reconstituted cofactor suggest it is Mn(III)(2)-Y*, which we propose is generated by Mn(II)(2)-NrdF reacting with two equivalents of HO(2)(-), produced by reduction of O(2) by NrdF-bound NrdI(hq). In the absence of NrdI(hq), with a variety of oxidants, no active RNR was generated. By contrast, a similar experiment with apoNrdF loaded with Fe(II) and incubated with O(2) in the presence or absence of NrdI(hq) gave 0.2 and 0.7 Y*/beta2 with specific activities of 80 and 300 units/mg, respectively. Thus NrdI(hq) hinders Fe(III)(2)-Y* cofactor assembly in vitro. We propose that NrdI is an essential player in E. coli class Ib RNR cluster assembly and that the Mn(III)(2)-Y* cofactor, not the diferric-Y* one, is the active metallocofactor in vivo.

FIGURE 1

EPR spectrum at 20 K of Mn^II₂-NrdF (40 μM). ApoNrdF was incubated with 4 Mn^II/β2 and mononuclear Mn^II was removed by Sephadex G25. The resulting protein contained 3.4 ± 0.2 Mn/β2 by atomic absorption spectroscopy.

FIGURE 2

Mn^II₂-NrdF interacts strongly with NrdI. Lanes 1–5: Mn^II₂-NrdF was incubated with 2 NrdI_ox/β2 and loaded onto a Ni affinity column. Lane 1: flowthrough; lanes 2–5: washes with Buffer B containing 0, 10, 50, and 250 mM imidazole, respectively. Equal volumes of each sample were loaded onto the gel. Lanes 6–8: Mn^II₂-NrdF in the absence of NrdI does not bind to the Ni column. Flowthrough (lane 6), wash with Buffer B (lane 7), wash with Buffer B containing 10 mM imidazole (lane 8).

FIGURE 3

Spectra of the ox (solid lines), sq (dotted lines), and hq (dashed lines) forms of NrdI in the presence (black) and absence (red) of apoNrdF, in Buffer C. The spectra of the neutral and anionic sq forms were estimated as described in Materials and Methods.

FIGURE 4

Visible spectra of dimanganese-Y• NrdF. A) Visible spectra of 50 μM Mn^II₂-NrdF reconstituted with 100 μM NrdI_hq and 1 mM O₂ in Buffer B (solid line); 50 μM Mn^II₂-NrdF with 100 μM NrdI_ox (dashed line); and dimanganese-Y• NrdF after incubation with 50 mM HU for 8 min (dotted line). The arrow indicates the characteristic feature of Y• at 408 nm. Inset: Spectrum of Y•, obtained by subtraction of the spectrum of HU-treated NrdF from that of dimanganese-Y• NrdF. The presence of features at 500–700 nm in this difference spectrum suggests partial reduction of the Mn cluster by HU. B) Spectrum of the dimanganese-Y• cofactor, obtained by subtraction of the spectrum of Mn^II₂-NrdF in the presence of NrdI_ox from that of dimanganese-Y• NrdF.

FIGURE 5

Specific activity, Y•/β2, and specific activity/Y• of dimanganese-Y• NrdF assembled with increasing concentrations of NrdI_hq. A) SA (empty squares) and Y•/β2 (filled squares) are dependent on NrdI_hq concentration in the assembly reaction. Mn^II₂-NrdF was preincubated with 0, 0.4, 0.8, 1.2, 1.6, 2, or 4 NrdI_hq/β2, in Buffer B and exposed to excess O₂. Y• was determined by EPR spin quantitation as described in Materials and Methods. Error bars indicate standard deviations of at least 2 independent experiments. B) SA/Y• plotted against Y•/β2 from data in Figure 5A.

FIGURE 6

EPR spectra of dimanganese-Y• NrdF. A) Comparison of the EPR spectra at 20 K of dimanganese-Y• NrdF and Mn^II₂-NrdF in the presence of NrdI_ox. In black, Mn^II₂-NrdF (50 μM) was reconstituted with 2 NrdI_hq/β2 (100 μM) and 1 mM O₂. In red, an identical sample, except NrdI_hq was oxidized prior to addition of Mn^II₂-NrdF (control). A small amount of mononuclear Mn^II is visible at g = 2.0054 (3345 G). Inset: Expansion of the 2500–3100 G region to show the decrease in Mn^II₂ hyperfine intensity upon cofactor assembly. The arrows indicate the peak-to-trough intensity used to compare Mn^II₂ cluster concentrations. B) EPR spectrum at 20 K of dimanganese-Y• NrdF (50 μM) after EDTA and Sephadex G25 treatment, and after subtraction of a buffer sample. C) Comparison of the 77 K EPR spectra of EDTA-treated Mn^III₂-Y• NrdF (black, acquired at 1 mW power) and Fe^III₂-Y• NrdF (red, 50 μW power), with the vertical scales normalized for sample concentration and spectrometer settings except for power. D) EPR spectrum at 3.6 K of EDTA-treated Mn^III₂-NrdF, after subtraction of a buffer sample.

FIGURE 7

Specific activity, Y•/β2, and SA/Y• for Fe^III₂-Y• NrdF. A) Correlation of specific activity and Y•/β2. ApoNrdF was preincubated anaerobically with 0, 0.6, 1, 2, 3, 4, or 5 Fe^II/β2 followed by addition of 3.5 O₂/β2. Data is shown for two sets of independent experiments (filled and open circles). SAs were determined using the radioactive assay. Y•/β2 was determined by EPR spin quantitation. Errors in the SA and Y• determinations are estimated at <10%. B) SA/Y• plotted against Y•/β2.

SCHEME 1

Proposed mechanism for formation of Mn^III₂-Y• NrdF by NrdI_hq and O₂.