The role of exogenous iron in the activation of ribonucleotide reductase from Escherichia coli

 Protein R2, the small component of ribonucleotide reductase from Escherichia coli, contains a diferric center and a catalytically essential tyrosyl radical. In vitro, this radical can be produced in the protein from two inactive forms, metR2, containing an intact diiron center and lacking the tyrosyl radical, and apoR2, lacking both iron and the radical. While activation of apoR2 requires only a source of ferrous iron and exposure to O₂, activation of metR2 was achieved using a multienzymatic system consisting of an NAD(P)H:flavin oxidoreductase, superoxide dismutase and a poorly defined protein fraction, named fraction b (Fontecave M, Eliasson R, Reichard P (1987) J Biol Chem 262 : 12325–12331). In both reactions, reduced R2, containing a diferrous center, is a key intermediate which is subsequently converted to active R2 during reaction with O₂. By in vivo labeling of E. coli with radioactive ⁵⁹Fe, we show that fraction b contains iron. Depletion of the iron in fraction b inactivates it, and fraction b can be substituted for by ferric citrate solutions. Furthermore, aqueous Fe²⁺ in the presence of dithiothreitol is able to convert metR2 into reduced R2. Therefore we propose that the function of fraction b is to provide, in association with the flavin reductase, ferrous iron for reduction of the endogenous diiron center. Since fraction b is not a single well-defined protein, it remains to be shown whether, in vivo, that function resides in a specific protein. Exogenous iron can thus participate in activation of both apoR2 and metR2, but it is incorporated into R2 only in the former case. A unifying mechanism is proposed. Received: 13 November 1996 / Accepted: 3 April 1997     

The Zn center of the anaerobic ribonucleotide reductase from E. coli

Strict and facultative anaerobes depend on a class III ribonucleotide reductase for their growth. These enzymes are the sole cellular catalysts for de novo biosynthesis of the deoxyribonucleotides needed for DNA chain elongation and repair. In its active form, the class III ribonucleotide reductase from Escherichia coli contains a free radical located on the G681 residue which is essential for the activation of the ribonucleotide substrate toward its reduction. The 3D structure of the homologous enzyme from bacteriophage T4 has revealed the presence of a metal center bound to four conserved cysteine residues. In this report we identify the metal of the E. coli enzyme as Zn. We show that the presence of Zn in this site protects the protein from proteolysis and prevents the formation of disulfide bridges within it. Finally, we show with the fully Zn-loaded reductase that thioredoxin or small thiols are dispensable for the formation of the glycyl radical. However, they are necessary for obtaining high turnover numbers, suggesting that they intervene in radical transfer steps subsequent to the formation of the glycyl radical.     

Characterization of enzymatic properties of human ribonucleotide reductase holoenzyme reconstituted in vitro from hRRM1, hRRM2, and p53R2 subunits

Ribonucleotide reductase (RR) is a highly regulated enzyme in the deoxyribonucleotide synthesis pathway. RR is responsible for the de novo conversion of ribonucleoside diphosphates to deoxyribonucleoside diphosphates, which are essential for DNA synthesis and repair. Besides two subunits, hRRM1 and hRRM2, p53R2 is a newly identified member of RR family that is induced by ultraviolet light in a p53-dependent manner. To understand the molecular interaction of RR subunits, we employed a eukaryotic expression system to express and purify all three subunits. After in vitro reconstitution, the results of [(3)H]CDP reduction assay showed that both eukaryotic recombinant hRRM2 and p53R2 proteins could interact with hRRM1 to form functional RR holoenzyme. The reconstituted RR activity was time-dependent and the reaction rate reached the plateau phase after 40min incubation. No matter the concentration, RR holoenzyme reconstituted from p53R2 and hRRM1 could only achieve about 40-75% kinetic activity of that from hRRM2 and hRRM1. The synthetic C-terminal heptapeptide competition assays confirmed that hRRM2 and p53R2 share the same binding site on hRRM1, but the binding site on hRRM1 demonstrated higher affinity for hRRM2 than for p53R2. In allosteric regulation assay, the effect of activation or inhibition of hRRM1 with ATP or dATP suggested that these effectors could regulate RR activity independent of different RR small subunits. Taken together, the eukaryotic expression system RR holoenzyme will provide a very useful tool to understand the molecular mechanisms of RR activity and the interactions of its subunits.     

Iron-sulfur interconversions in the anaerobic ribonucleotide reductase from Escherichia coli

The anaerobic ribonucleotide reductase from Escherichia coli contains an iron-sulfur cluster which, in the reduced [4Fe-4S]⁺ form, serves to reduce S-adenosylmethionine and to generate a catalytically essential glycyl radical. The reaction of the reduced cluster with oxygen was studied by UV-visible, EPR, NMR, and Mössbauer spectroscopies. The [4Fe-4S]⁺ form is shown to be extremely sensitive to oxygen and converted to [4Fe-4S]²⁺, [3Fe-4S]^+/0, and to the stable [2Fe-2S]²⁺ form. It is remarkable that the oxidized protein retains full activity. This is probably due to the fact that during reduction, required for activity, the iron atoms, from 2Fe and 3Fe clusters, readily reassemble to generate an active [4Fe-4S] center. This property is discussed as a possible protective mechanism of the enzyme during transient exposure to air. Futhermore, the [2Fe-2S] form of the protein can be converted into a [3Fe-4S] form during chromatography on dATP-Sepharose, explaining why previous preparations of the enzyme were shown to contain large amounts of such a 3Fe cluster. This is the first report of a 2Fe to 3Fe cluster conversion.     

Effect on ribonucleotide reductase of novel lipophilic iron chelators: the desferri-exochelins

Desferri-exochelins are siderophores secreted by Mycobacterium tuberculosis that are both lipid- and water-soluble and have a high binding affinity for iron. Desferri-exochelin 772SM inhibits DNA replication and ribonucleotide reductase activity at 10-fold less concentration than the lipid-insoluble iron chelator deferoxamine, which is currently in clinical use. Neither chelator can extract iron directly from ribonucleotide reductase. However, because of its lipid-solubility and high binding affinity, desferri-exochelin is able to enter cells rapidly and access intracellular iron, while deferoxamine has limited capacity to cross the cell membrane.     

EPR evidence of two structurally different diferric sites in Mycobacterium tuberculosis R2-2 ribonucleotide reductase protein

Mixed-valent species were generated in the diiron site of active (with tyrosyl free radical) and met (without radical) forms of protein R2-2 in a class Ib ribonucleotide reductase from Mycobacterium tuberculosis by low temperature reduction (γ-irradiation) at 77 K. The primary mixed-valent EPR signal is a mixture of two components with axial symmetry and g_av<2.0, observable at temperatures up to 77 K, and assigned to antiferromagnetically coupled high spin ferric/ferrous sites. The two components in the primary EPR signal can be explained by the existence of two structurally distinct μ-oxo-bridged diferric centers, possibly related to structural heterogeneity around the iron site, and/or different properties of the two polypeptide chains in the homodimeric protein after the radical reconstitution reaction. Annealing of the irradiated R2-2 samples to 143 K transforms the primary EPR signal into a rhombic spectrum characterized by g_av<1.8 and observable only below 25 K. This spectrum is assigned to a partially relaxed form with a μ-hydroxo-bridge. Further annealing at 228 K produces a new complex rhombic EPR spectrum composed of at least two components. An identical EPR spectrum was observed and found to be stable upon chemical reduction of Mycobacterium tuberculosis RNR R2-2 at 293 K by dithionite.     

Haemonchus contortus: molecular cloning, sequencing, and expression analysis of the gene coding for the small subunit of ribonucleotide reductase

Gastro-intestinal (GI) parasites are of great agricultural importance, annually costing the livestock industry vast amounts in resources to control parasitism. One such GI parasite, Haemonchus contortus, is principally pathogenic to sheep; with the parasite's blood-feeding behaviour causing effects ranging from mild anaemia to mortality in young animals. Current means of control, which are dependent on repeated treatment with anthelmintic chemicals, have led to increasing drug resistance. Together with the growing concern over residual chemicals in the environment and food chain, this has led to attempts to better understand the biology of the parasite with the aim to develop alternate or supplementary means of control, including the development of molecular vaccines. As a first step towards the understanding of the synthesis of deoxyribonucleotides in H. contortus, and its potential application to therapeutic control of this economically important parasite, we report the cloning, sequencing, and mRNA expression analysis of the ribonucleotide reductase R2 gene.     

EPR evidence of two structurally different diferric sites in Mycobacterium tuberculosis R2-2 ribonucleotide reductase protein

Mixed-valent species were generated in the diiron site of active (with tyrosyl free radical) and met (without radical) forms of protein R2-2 in a class Ib ribonucleotide reductase from Mycobacterium tuberculosis by low temperature reduction (γ-irradiation) at 77 K. The primary mixed-valent EPR signal is a mixture of two components with axial symmetry and g_av<2.0, observable at temperatures up to 77 K, and assigned to antiferromagnetically coupled high spin ferric/ferrous sites. The two components in the primary EPR signal can be explained by the existence of two structurally distinct μ-oxo-bridged diferric centers, possibly related to structural heterogeneity around the iron site, and/or different properties of the two polypeptide chains in the homodimeric protein after the radical reconstitution reaction. Annealing of the irradiated R2-2 samples to 143 K transforms the primary EPR signal into a rhombic spectrum characterized by g_av<1.8 and observable only below 25 K. This spectrum is assigned to a partially relaxed form with a μ-hydroxo-bridge. Further annealing at 228 K produces a new complex rhombic EPR spectrum composed of at least two components. An identical EPR spectrum was observed and found to be stable upon chemical reduction of Mycobacterium tuberculosis RNR R2-2 at 293 K by dithionite.     

Chemical reduction of the diferric/radical center in protein R2 from mouse ribonucleotide reductase is independent of the proposed radical transfer pathway

The rates of reduction of the diferric/radical center in mouse ribonucleotide reductase protein R2 were studied by light absorption and EPR in the native protein and in three point mutants of conserved residues involved in the proposed radical transfer pathway (D266A, W103Y) or in the unstructured C terminal domain (Y370W). The pseudo-first order rate constants for chemical reduction of the tyrosyl radical and diferric center by hydroxyurea, sodium dithionite or the dihydro form of flavin adenine dinucleotide, were comparable with or higher (particularly D266A, by dithionite) than in native R2. Molecular modeling of the D266A mutant showed that the iron/radical site should be more accessible for external reductants in the mutant than in native R2. The results indicate that no specific pathway is required for the reduction. The dihydro form of flavin adenine dinucleotide was found to be a very efficient reductant in the studied proteins compared to dithionite alone. The EPR spectra of the mixed-valent Fe(II)Fe(III) sites formed by chemical reduction in the D266A and W103Y mutants were clearly different from the spectrum observed in the native protein, indicating that the structure of the diferric site was affected by the mutations, as also suggested by the modeling study. No difference was observed between the mixed-valent EPR spectra generated by chemical reduction in Y370W mutant and native mouse R2 protein.     

Biochemical modulation of Ara-C effects by amidox, an inhibitor of ribonucleotide reductase in HL-60 promyelocytic human leukemia cells

Amidox, a new polyhydroxy-substituted benzoic acid derivative, is a potent inhibitor of the enzyme ribonucleotide reductase (RR), which catalyses the de novo synthesis of DNA. RR is considered to be an excellent target for anti cancer chemotherapy. We investigated the biochemical and antineoplastic effects of amidox as a single agent and in combination with Ara-C in human HL-60 promyelocytic leukemia cells. Amidox inhibited the growth of HL-60 cells in a growth inhibition assay with an IC50 of 25 microM. In a soft agar colony forming assay, amidox yielded a 50% inhibition of colony formation at 13 microM. We also investigated the effects of amidox treatment on the formation of deoxynucleosidetriphosphates. Amidox (50 and 75 microM for 24 hours) could significantly decrease intracellular concentrations of dCTP, dATP and dGTP pools, whereas dTTP levels increased. We then tested the combination effects of amidox with Ara-C; this combination yielded additive cytotoxic effects both in growth inhibition and in soft agar colony formation assays. This effect was due to the increased formation of Ara-CTP, the active metabolite of Ara-C, after preincubation with amidox. Preincubation of HL-60 cells with 75 and 100 microM amidox for 24 hours caused an increase in the intracellular Ara-CTP concentrations by 576% and 1143%, respectively. Therefore amidox might offer an additional option for the treatment of leukemia and thus be further investigated in in vivo studies as a single agent and in combination with Ara-C.     

Structure of the large subunit of class Ib ribonucleotide reductase from Salmonella typhimurium and its complexes with allosteric effectors

The three-dimensional structure of the large subunit of the first member of a class Ib ribonucleotide reductase, R1E of Salmonella typhimurium, has been determined in its native form and together with three allosteric effectors. The enzyme contains the characteristic ten-stranded alpha/beta-barrel with catalytic residues at a finger loop in its center and with redox-active cysteine residues at two adjacent barrel strands. Structures where the redox-active cysteine residues are in reduced thiol form and in oxidized disulfide form have been determined revealing local structural changes. The R1E enzyme differs from the class Ia enzyme, Escherichia coli R1, by not having an overall allosteric regulation. This is explained from the structure by differences in the N-terminal domain, which is about 50 residues shorter and lacks the overall allosteric binding site. R1E has an allosteric substrate specificity regulation site and the binding site for the nucleotide effectors is located at the dimer interface similarly as for the class Ia enzymes. We have determined the structures of R1E in the absence of effectors and with dTTP, dATP and dCTP bound. The low affinity for ATP at the specificity site is explained by a tyrosine, which hinders nucleotides containing a 2'-OH group to bind.     

Kinetic studies on the reduction of the R2 subunit of mouse ribonucleotide reductase with hydroxyurea, hydrazine, phenylhydrazine and hydroxylamine

 Four reductions of the R2 subunit of mouse ribonucleotide reductase have been studied and found to exhibit different behaviour from that of Escherichia coli R2. An important difference is that there is no stable met-R2 (Fe₂ ^II I) form of mouse R2. With hydroxyurea, hydrazine and hydroxylamine uniphasic kinetics are observed for the combined reduction of radical Tyr ^˙ and Fe₂ ^II I components to tyrosine and Fe₂ ^II respectively. The rate constants, determined at 370 nm (emphasising Fe^III decay) and 417 nm (emphasising Tyr ^˙ decay), differ by factors of 2–3, allowing some mechanistic features to be defined. The studies with hydrazine are particularly important. In the case of E. coli R2, a first phase corresponding to two-equivalent reduction of the met-R2 component has been observed [18]. It is likely that the four times slower second phase reaction of active E. coli R2 also corresponds to the Fe₂ ^II I → Fe₂ ^II change and is followed by fast intramolecular Fe₂ ^II reduction of the higher potential Tyr ^˙. The latter changes are believed to hold also for (active) mouse R2. The Fe^IIFe^III semi forms have been detected at low levels by EPR for mouse R2 (9%) and E. coli (∼5%) in previous studies. Further substrate reduction of Fe^IIFe^III occurs at a comparable rate to account for the transient behaviour of Fe^IIFe^III. For mouse R2 the combined Fe^III decay processes (which we are unable to separate) give smaller uniphasic rate constants at 370 nm than at 417 nm. A fitted-base-line (FBL) treatment of absorbance changes at 417 nm targets more closely the Tyr ^˙ decay as a means of monitoring the rate-determining step. The FBL method gives rate constants k (M^–1 s^–1) at 25  °C and pH 7.5 for hydroxyurea (1.46), hydrazine (0.163) and hydroxylamine (4.4). Surprisingly, phenylhydrazine, with a less strong reduction potential (0.25 V), gives a substantially faster reduction of the Tyr ^˙ as the only redox step (rate constant 27 M^–1 s^–1). In this case a slower second phase at 370 nm is independent of reductant and corresponds to rate-controlling release of Fe^III. Overall the results indicate a more reactive redox centre for mouse R2 and help develop further an understanding of factors affecting the reactivity of R2. Received: 11 October 1996 / Accepted: 11 February 1997     

Site-specific incorporation of fluorotyrosines into the R2 subunit of E. coli ribonucleotide reductase by expressed protein ligation

Expressed protein ligation (EPL) allows semisynthesis of a target protein with site-specific incorporation of probes or unnatural amino acids at its N or C termini. Here, we describe the protocol that our lab has developed for incorporating fluorotyrosines (F(n)Ys) at residue 356 of the small subunit of Escherichia coli ribonucleotide reductase using EPL. In this procedure, the majority of the protein (residues 1-353 out of 375) is fused to an intein domain and prepared by recombinant expression, yielding the protein in a thioester-activated, truncated form. The remainder of the protein, a 22-mer peptide, is prepared by solid-phase peptide synthesis and contains the F(n)Y at the desired position. Ligation of the 22-mer peptide to the thioester-activated R2 and subsequent purification yield full-length R2 with the F(n)Y at residue 356. The procedure to generate 100 mg quantities of Y356F(n)Y-R2 takes 3-4 months.     

A mixed valence form of the iron cluster in the B2 protein of ribonucleotide reductase from Escherichia coli.

A mixed valent form of the iron cluster (Fe(II)Fe(III) in the B2 protein of ribonucleotide reductase has been isolated and characterized. The irons in this state of the protein are ferromagnetically coupled as indicated by the observation of a novel S = 9/2 EPR spectrum. This is the first ferromagnetically coupled Fe(II)Fe(III) cluster reported for a protein and the first observation of the mixed valence form of ribonucleotide reductase.     

Characterization of NO adducts of the diiron center in protein R2 of<Emphasis Type="Italic"> Escherichia coli</Emphasis> ribonucleotide reductase and site-directed variants; implications for the O<Subscript>2 </Subscript>activation mechanism*

The R2 subunit of Escherichia coli ribonucleotide reductase contains a diiron site that reacts with O₂ to produce a tyrosine radical (Y122·). In wild-type R2 (R2-wt), the first observable reaction intermediate is a high-valent [Fe^III-Fe^IV] state called compound X, but in related diiron proteins such as methane monooxygenase, ⁹-desaturase, and ferritin, peroxodiiron(III) complexes have been characterized. Substitution of iron ligand D84 by E within the active site of R2 allows an intermediate (-1,2-peroxo)diiron species to accumulate. To investigate the possible involvement of a bridging peroxo species within the O₂ activation sequence of R2-wt, we have characterized the iron-nitrosyl species that form at the diiron sites in R2-wt, R2-D84E, and R2-W48F/D84E by using vibrational spectroscopy. Previous work has shown that the diiron center in R2-wt binds one NO per iron to form an antiferromagnetically coupled [{FeNO}⁷]₂ center. In the wt and variant proteins, we also observe that both irons bind one NO to form a {FeNO}⁷ dimer where both Fe–N–O units share a common vibrational signature. In the wt protein, (Fe–NO), (Fe–N–O), and (N–O) bands are observed at 445, 434 and 1742 cm^–1, respectively, while in the variant proteins the (Fe–NO) and (Fe–N–O) bands are observed ~10 cm^–1 higher and the (N–O) ~10 cm^–1 lower at 1735 cm^–1. These results demonstrate that all three proteins accommodate fully symmetric [{FeNO}⁷]₂ species with two identical Fe–N–O units. The formation of equivalent NO adducts in the wt and variant proteins strongly favors the formation of a symmetric bridging peroxo intermediate during the O₂ activation process in R2-wt.     

Reversible interconversion of the functional state of the gene regulator FNR from Escherichia coli in vivo by O2 and iron availability

FNR, the gene regulator of anaerobic respiratory genes of Escherichia coli is converted in vivo by O₂ and by chelating agents to an inactive state. The interconversion process was studied in vivo in a strain with temperature controlled synthesis of FNR by measuring the expression of the frd (fumarate reductase) operon and the reactivity of FNR with the alkylating agent iodoacetic acid. FNR from aerobic bacteria is, after arresting FNR synthesis and shifting to anaerobic conditions, able to activate frd expression and behaves in the alkylation assay like anaerobic FNR. After shift from anaerobic to aerobic conditions, FNR no longer activates the expression of frd and reacts similar to aerobic FNR in the alkylation assay. The conversion of aerobic (inactive) to anaerobic (active) FNR occurs in the presence of chloramphenicol, an inhibitor of protein synthesis. Anaerobic FNR can also be converted post-translationally to inactive, metal-depleted FNR by growing the bacteria in the presence of chelating agents. The reverse is also possible by incubating metal-depleted bacteria with Fe²⁺. From the experiments it is concluded that the aerobic and the metal-depleted form of FNR can be transferred post-translationally and reversibly to the anaerobic (active) form. The response of FNR to changes in O₂ supply therefore occurs at the FNR protein level in a reversible mode.Abbreviation BV_red = reduced benzyl viologen     

Cloning of the vaccinia virus ribonucleotide reductase small subunit gene. Characterization of the gene product expressed in Escherichia coli.

During its infectious cycle, vaccinia virus expresses a virus-encoded ribonucleotide reductase which is distinct from the host cellular enzyme (Slabaugh, M.B., and Mathews, C.K. (1984) J. Virol. 52, 501-506; Slabaugh, M.B., Johnson, T.L., and Mathews, C.K. (1984) J. Virol. 52, 507-514). We have cloned the gene for the small subunit of vaccinia virus ribonucleotide reductase (designated VVR2) into Escherichia coli and expressed the protein using a T7 RNA polymerase plasmid expression system. After isopropyl beta-D-thiogalactopyranoside induction, accumulation of a 37-kDa peptide was detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and this peptide reacted with polyclonal antiserum raised against a TrpE-VVR2 fusion protein. The 37-kDa protein was purified to homogeneity, and gel filtration of the purified protein revealed that the recombinant protein existed as a dimer in solution. Purified recombinant VVR2 protein was shown to complement the activity of purified recombinant ribonucleotide reductase large subunit, with a specific activity that was similar to native VVR2 from a virus-infected cell extract. A CD spectrum of the recombinant viral protein showed that like the mouse protein, the vaccinia virus protein has 50% alpha-helical structure. Like other iron-containing ribonucleotide reductase small subunits, recombinant VVR2 protein contained a stable organic free radical that was detectable by EPR spectroscopy. The EPR spectrum of purified recombinant VVR2 was identical to that of vaccinia virus-infected mammalian cells. Both the hyperfine splitting character and microwave saturation behavior of VVR2 were similar to those of mouse R2 and distinct from E. coli R2. By using amino acid analysis to determine the concentration of VVR2, we determined that approximately 0.6 radicals were present per R2 dimer. Our results indicate that vaccinia virus small subunit is similar to mammalian ribonucleotide reductases.     

Generation and electron paramagnetic resonance spin trapping detection of thiyl radicals in model proteins and in the R1 subunit of Escherichia coli ribonucleotide reductase.

In the Escherichia coli class Ia ribonucleotide reductase (RNR), the best characterized RNR, there is no spectroscopic evidence for the existence of the postulated catalytically essential thiyl radical (R-S(*)) in the substrate binding subunit R1. We report first results on artificially generated thiyl radicals in R1 using two different methods: chemical oxidation by Ce(IV)/nitrilotriacetate (NTA) and laser photolysis of nitric oxide from nitrosylated cysteines. In both cases, EPR spin trapping at room temperature using phenyl-N-t-butylnitrone, and controls with chemically blocked cysteines, has shown that the observed spin adduct originates from thiyl radicals. The EPR line shape of the protein-bound spin adduct is typical for slow motion of the nitroxide moiety, which indicates that the majority of trapped thiyl radicals are localized in a folded region of R1. In aerobic R1 samples without spin trap that were frozen after treatment with Ce(IV)/NTA or laser photolysis, we observed sulfinyl radicals (R-S(*)=O) assigned via their g-tensor components 2.0213, 2.0094, and 2.0018 and the hyperfine tensor components 1.0, 1.1, and 0.9 mT of one beta-proton. Sulfinyl radicals are the reaction products of thiyl radicals and oxygen and give additional evidence for generation of thiyl radicals in R1 by the procedures used.