Use of 3 aminotyrosine to examine pathway dependence of radical propagation in Escherichia coli ribonucleotide reductase

Abstract

Escherichia coli ribonucleotide reductase (RNR), an α2β2 complex, catalyzes the conversion of nucleoside 5′-diphosphate substrates (S) to 2′-deoxynucleoside 5′-diphosphates. α2 houses the active site for nucleotide reduction and the binding site for allosteric effectors (E). β2 contains the essential diferric tyrosyl radical (Y₁₂₂•) cofactor which, in the presence of S and E, oxidizes C₄₃₉ in α to a thiyl radical, C₄₃₉•, to initiate nucleotide reduction. This oxidation occurs over 35 Å and is proposed to involve a specific pathway: Y₁₂₂•→W₄₈→Y₃₅₆ in β2 to Y₇₃₁→Y₇₃₀→C₄₃₉ in α2. 3-Aminotyrosine (NH₂Y) has been site specifically incorporated at residues 730 and 731, and formation of the aminotyrosyl radical (NH₂Y•) has been examined by stopped-flow (SF) UV-vis and EPR spectroscopies. To examine the pathway dependence of radical propagation, the double-mutant complexes Y₃₅₆F–β2:Y₇₃₁NH₂Y-α2, Y₃₅₆F–β2:Y₇₃₀NH₂Y–α2, and wt-β2:Y₇₃₁F/Y₇₃₀NH₂Y–α2, in which the non-oxidizable F acts as a pathway block, were studied by SF and EPR spectroscopies. In all cases, no NH₂Y• was detected. To study off-pathway oxidation, Y₄₁₃, located 5 Å from Y₇₃₀ and Y₇₃₁ but not implicated in long-range oxidation, was examined. Evidence for NH₂Y₄₁₃• was sought in three complexes: wt-β2:Y₄₁₃NH₂Y-α2 (a), wt-β2:Y₇₃₁F/Y₄₁₃NH₂Y–α2 (b), and Y₃₅₆F-β2:Y₄₁₃NH₂Y–α2 (c). With (a), NH₂Y• was formed with a rate constant of 25-30% and an amplitude of 25% that observed for its formation at residues 731 and 730. With (b), the rate constant for NH₂Y• formation was 0.2-0.3% that observed at 731 and 730, and with (c), no NH₂Y• was observed. These studies suggest the evolution of an optimized pathway of conserved Ys in the oxidation of C₄₃₉.

Long-range electron transfer (ET) (1) is prevalent in biology and plays a central role in many processes including respiration (2), nitrogen fixation (3), and the photosynthetic reaction centers (4). Studies over many decades have established that ET occurs rapidly and efficiently between metal clusters that are spaced 10 to 15 Å apart (5) and that the process involves electron tunneling as described by the Marcus-Levich equation (6). The importance of the medium, the intervening polypeptide structural motifs, has been studied extensively in model proteins and biological systems (1, 7). The protein fold clearly plays a central role in lowering the reorganization energy of the biological ET reaction. It is also clear that ET kinetics can be regulated by the dynamics of conformational changes, especially across protein-protein interfaces (8, 9).

The discovery of the central role of a tyrosyl radical (Y•) in the class Ia ribonucleotide reductase (10) and the O₂-evolving complex of photosystem II (11) led to the realization that ET processes over long distances are not limited to metal clusters. Long-range ET can also involve the aromatic amino acids tryptophan and tyrosine, the oxidations of which require loss of both an electron and a proton (12). Ribonucleotide reductase (RNR), which catalyzes the conversion of nucleotides to deoxynucleotides (13), has served as the paradigm for long-range proton-coupled electron transfer (PCET) with the tyrosyl radical (Y•) in the β subunit mediating the oxidation of a cysteine in the α subunit 35 Å removed (Figure 1) (12, 14). The present paper describes the use of 3-aminotyrosine (NH₂Y), site specifically incorporated into the α subunit of RNR (15), to examine the pathway dependence, and the importance of transient intermediates, proposed for this oxidation.

FIGURE 1.

The proposed radical propagation pathway within an α:β pair of E. coli RNR. Y₃₅₆, Y₇₃₁, and Y₇₃₀ have been shown to be redox-active during radical transfer and their positions within the complex have been determined structurally (14) and/or experimentally (35). Y₃₅₆ is located in the structurally disordered C-terminal tail of β (44), and thus its location in the complex is not known.

Tyrosyl radicals (Y•) and tryptophan radicals (W•) have been studied in three different contexts indirectly related to radical propagation in RNRs. In the area of model proteins and peptides, oxidizable amino acids have been positioned between electron donor/acceptor pairs. Studies of the ET process in the resulting systems have demonstrated rate accelerations associated with a hopping model requiring formation of a transient intermediate (16, 17). These “simple” systems have allowed detailed mechanistic analyses and provide precedent for the multiple hopping steps proposed for RNR.

Y and W radicals have been observed in heme and non-heme iron dependent proteins, in which a reactive metal-based intermediate generated during catalysis is, in the absence of substrate, reduced by ET from an aromatic residue(s) in the vicinity of the metallocofactor (18-21). In general, low amounts of multiple radical species are generated on a slow (second) time scale. These systems are starting to define the mechanisms and consequences of aberrant oxidations and provide a model for the phenotype of an off-pathway oxidation in RNR.

Finally, transient aromatic amino acid radicals have been studied in enzymes such as cytochrome c peroxidase and prostaglandin synthase, in which metallocofactors are responsible for the oxidation of aromatic amino acids over short distances in the presence of substrate (22, 23). Only ribonucleotide reductase and photolyase (PL) however are believed to utilize multiple, transient aromatic amino acid radical intermediates in long-range ET. PL, as isolated, contains a flavin cofactor in the semiquinone form (FADH•) that needs to be reduced to its active form (FADH⁻) prior to catalysis (24). Time-resolved kinetic studies suggest that the reduction occurs by a hopping mechanism involving three conserved Ws over a 15 Å distance (25). To obtain insight into the pathway dependence of reduction of FADH•, the non oxidizable amino acid phenylalanine (F) was incorporated in place of each of the Ws and re-analyzed by ultrafast kinetics (26-28). The measurements on the wild-type and mutant PLs support a model in which 3 Ws act as a wire to the flavin cofactor.

The class Ia ribonucleotide reductases are the only known enzymes whose physiological function requires long-range oxidation, thought to occur in a pathway-dependent fashion through a series of aromatic amino acids (Figure 1). This provocative mechanism was first proposed by Uhlin and Eklund on the basis of a docking model of the α2:β2 complex and on the strict conservation of the residues in primary sequence alignments (14). Mutagenesis studies in which each of the residues was replaced individually by F demonstrated their importance in catalysis, structure, or both (29-32). However, only with the advent of technology for the site-specific incorporation of unnatural amino acids into α and β has the involvement of the three tyrosines (356 in β2, 731 and 730 in α2, Figure 1) and the distance of the individual oxidation “steps” been examined (15, 33-36). We have demonstrated recently the site-specific incorporation of 3-aminotyrosine (NH₂Y) into the α and β subunits at each of these positions. SF UV-vis and EPR studies revealed that NH₂Y• is generated in a nucleotide-dependent manner with multiphasic kinetics, and that the rates of formation are maximized when substrate and effectors are bound (15, 37). In addition, our studies indicated that NH₂Y-αs are capable of supporting nucleotide reduction, suggesting that the observed radical may be an intermediate on the proposed reaction pathway. From these collective studies, we have proposed that radical propagation occurs via orthogonal PCET in the β2 subunit, with the proton transferring off-pathway to an amino acid acceptor or to solvent, and via co-linear PCET in the α2 subunit, with both proton and electron transferring between the same donor/acceptor pair (12, 15).

Despite the appeal of a conserved PCET pathway, evidence in support of transient radical intermediates in RNR has been challenging to obtain because the reaction is rate-limited by conformational changes gated by binding of substrates and effectors to α (38). In previous efforts to induce light-mediated turnover on a photopeptide:α2 RNR complex, we have observed deoxynucleotide formation and transient formation of a Y• within the peptide, but due to conformational gating, we have thus far been unable to detect transient intermediates within α (39, 40).

NH₂Y is easier to oxidize than Y by 190 mV at pH 7 (15). From a thermodynamic perspective, the detection of NH₂Y• in NH₂Y-αs may not be entirely surprising. Thus the pathway dependence of NH₂Y• formation is important to establish (Figure 1). To provide further support for a defined pathway, F has been incorporated in place of each Y in the pathway to function as a block of NH₂Y• formation. Incubation of Y₃₅₆F-β2:NH₂Y₇₃₁-α2, Y₃₅₆F-β2:NH₂Y₇₃₀-α2, or wt-β2:Y₇₃₁F/NH₂Y₇₃₀-α2 with CDP/ATP completely blocks NH₂Y oxidation, supporting the importance of a specific pathway. As a further test of the pathway dependence, the residue Y₄₁₃ was selected for site specific “off-pathway” NH₂Y incorporation due to its proximity to Y₇₃₁ and Y₇₃₀ in α. Analogous studies were conducted with wt-β2:Y₄₁₃NH₂Y-α2, wt-β2:Y₇₃₁F/Y₄₁₃NH₂Y-α2, and Y₃₅₆F-β2:Y₄₁₃NH₂Y-α2 in which evidence for NH₂Y• formation was sought by SF and EPR spectroscopy. The results of these experiments establish that functional long-range PCET in RNR occurs via an optimized pathway of amino acids spanning >35 Å across the α:β interface.

MATERIALS AND METHODS

Materials

The expression and purification of wt-α2 (2500 nmol/min/mg), wt-β2 (1.2 Y₁₂₂•/dimer, 7600 nmol/min/mg), Y₇₃₁NH₂Y-α2 (150 nmol/min/mg), Y₇₃₀NH₂Y-α2 (110 nmol/min/mg), His-Y₃₅₆F-β2 (1.0-1.2 Y₁₂₂•/dimer, <1 nmol/min/mg), and Y₇₃₀F-α2 (<25 nmol/min/mg) were conducted as previously described (15, 41). His-Y₃₅₆F-β2 will be referred to as “Y₃₅₆F-β2” henceforth. E. coli thioredoxin (TR, 40 units/mg) and thioredoxin reductase (TRR, 1400 units/mg) were isolated as previously described (41). [5-³H]-CDP was purchased from ViTrax (Placentia, CA).

Generation of pTrc-nrdA-TAG₇₃₀TTT₇₃₁, pTrc-nrdA-TAG₄₁₃ and pTrc-nrdA-TAG₄₁₃TTT₇₃₁

The QuikChange Kit (Stratagene) was used according to manufacturer’s instructions to generate pTrc nrdA-TAG₇₃₀TTT₇₃₁, pTrc-nrdA-TAG₄₁₃ and pTrc-nrdA-TAG₄₁₃TTT₇₃₁. To generate pTrc-nrdA-TAG₇₃₀TTT₇₃₁, the template, pTrc-nrdA-TAG₇₃₀, (15) was amplified in the presence of forward (5′-TTC GGG GTC AAA ACA CTG TAG TTT CAG AAC ACC CGT GAC GGC GCT-3′) and reverse (5′-AGC GCC GTC ACG GGT GTT CTG AAA CTA CAG TGT TTT GAC CCC GAA-3′) primers to insert a TTT codon (Phe) at residue 731. To generate pTrc-nrdA-TAG₄₁₃, the template, pTrc-nrdA-wt, was amplified in the presence of forward (5′-GCG TCT ACC GGT CGT ATC TAG ATT CAG AAC GTT GAC CAC TGC-3′) and reverse (5′-GCA GTG GTC AAC GTT CTG AAT CTA GAT ACG ACC GGT AGA CGC-3′) primers to insert the amber stop codon (TAG) at residue 413. pTrc-nrdA-TAG₄₁₃TTT₇₃₁ was then generated from pTrc-nrdA-TAG₄₁₃, with forward (5′-C GGG GTC AAA ACA CTG TAT TTT CAG AAC ACC CGT GAC G-3′) and reverse (5′-C GTC ACG GGT GTT CTG AAA ATA CAG TGT TTT GAC CCC G-3′) primers to insert a TTT codon (Phe) at residue 731. For each plasmid, the mutation(s) was confirmed by sequencing the entire gene at the MIT Biopolymers Laboratory.

Expression, purification, and activity determination of Y₇₃₁F/Y₇₃₀NH₂Y-α2

Y₇₃₁F/Y₇₃₀NH₂Y-α2 was expressed, purified, and pre-reduced as described for Y₇₃₀NH₂Y-α2 and Y₇₃₁NH₂Y-α2 (15). Specific activity was determined by the radioactive RNR assay (33). Assays were conducted with mutant α2 and wt β2 at concentrations of 0.5 μM and 2.5 μM, respectively, in 50 mM HEPES, 15 mM MgSO₄, 1 mM EDTA, pH 7.6 (assay buffer) at 25 °C. [5-³H]-CDP had a specific activity of 1450 cpm/nmol. The reaction product was dephosphorylated by treatment with alkaline phosphatase and the resulting dC was quantitated by the method of Steeper and Steuart (33).

Expression, purification, and activity determination of Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2

Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2 were expressed and purified as described for Y₇₃₀NH₂Y-α2 and Y₇₃₁NH₂Y-α2 (15). However, an additional step using a Poros HQ/20 FPLC anion exchange column (Applied Biosystems, 1.6 × 10 cm, 20 mL) was required to obtain protein of >95% purity judged by SDS-PAGE analysis. The column was equilibrated in 50 mM Tris, 1 mM EDTA, and 5 mM DTT, pH 7.6 and was loaded with 15-20 mg of Y₄₁₃NH₂Y–β2 or Y₇₃₁F/Y₇₃₀NH₂Y-α2. The column was washed with one column volume of equilibration buffer at a flow rate of 4 mL/min, then eluted with a linear gradient of 100 to 500 mM NaCl (60 mL × 60 mL) in the same buffer at the same flow rate. In the case of Y₇₃₁F/Y₇₃₀NH₂Y-α2, the protein was chromatographed twice. Subsequent to the FPLC purification step, the protein was pre-reduced according to the standard procedure (15).

The activities of Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2 were determined using the spectrophotometric (41) and/or radioactive assays (33), with minor modifications. Mutant α2s (0.1 or 0.2 μM) were assayed in the presence of a 5-fold excess of β2 (0.5 μM or 1.0 μM) in assay buffer at 25 °C. [5-³H]-CDP had activity between 2700 and 4800 cpm/nmol.

Determination of the K_d for Y₄₁₃NH₂Y-α2 and β2 interaction

The K_d between Y₄₁₃NH₂Y-α2 and wt β2 was determined using Y₇₃₀F-α2 as a competitive inhibitor of nucleotide reduction (29, 42). Y₄₁₃NH₂Y-α2 (0.7 μM) and β2 (0.35 μM) were incubated with CDP (1 mM), ATP (1.6 mM), TR (50 μM), TRR (1 μM), and NADPH (0.2 mM) in assay buffer. Inhibition of RNR activity (as measured by decrease in NADPH consumption) was determined in the presence of increasing Y₇₃₀F-α2 (0.1 μM to 3 μM). The data analysis for this procedure typically relies on the mutant α2 of unknown K_d acting as the inhibitor of nucleotide reduction (36). The analysis was modified, as the K_d of the active Y₄₁₃NH₂Y-α2 is not known, whereas the K_d of the inhibitor (Y₇₃₀F-α2) had been previously determined under conditions in which the K_i and K_d of the inhibitor are equivalent (30). The K_d for the Y₄₁₃NH₂Y-α2:wt-β2 complex was approximated and the experimentally determined relative activities were then used to extrapolate the K_d for the Y₇₃₀F-α2:wt-β2 complex. The K_d for Y₄₁₃NH₂Y-α2 was refined by iterative fitting of the experimental data until the fit yielded a K_d for Y₇₃₀F-α2 that was in good agreement with the literature value.

Reaction of NH₂Y-α2s with Y₃₅₆F-β2 monitored by EPR spectroscopy

Pre reduced Y₇₃₀NH₂Y-α2 or Y₇₃₁NH₂Y-α2 (15 μM, final concentration) was mixed with Y₃₅₆F-β2 (15 μM, 1 Y•/β2), and CDP/ATP (1 mM/3 mM, respectively) in assay buffer at 25 °C. The reactions were quenched at 20 s or 2 min by hand in liquid N₂. The reaction of Y₄₁₃NH₂Y-α2 and Y₃₅₆F-β2 was conducted similarly, except the final protein concentration was 50 μM. EPR spectra were recorded at 77 K on a Brüker ESP-300 X-band spectrometer equipped with a quartz finger dewar containing liquid N₂ in the Department of Chemistry Instrumentation Facility. EPR parameters were as follows: microwave frequency=9.34 GHz, power=30 μW, modulation amplitude=1.5 G, modulation frequency=100 kHz, time constant=5.12 ms, scan time=41.9 s. EPR spin quantitation was carried out in WinEPR (Brüker) using Cu^II as a standard (43). EPR subtractions were conducted with an in-house, Excel-based program using the spectrum of Y₁₂₂• from either wt β2 or a wt-β2:mutant-α2 complex as a reference.

Reaction of Y₃₅₆F-β2 with NH₂Y-α2s monitored by SF UV-vis spectroscopy

SF kinetics were performed on an Applied Photophysics DX 17MV instrument equipped with the Pro-Data upgrade. Pre-reduced Y₇₃₀NH₂Y-α2 (or Y₇₃₁NH₂Y-α2) and ATP were mixed with Y₃₅₆F-β2 and CDP to yield final concentrations of 8 μM, 3 mM, 8 μM and 1 mM, respectively, in assay buffer at 25 °C. The concentrations of Y₁₂₂•, NH₂Y₇₃₁• and NH₂Y₇₃₀• were monitored at 410 nm (ε = 3,700 M⁻¹ cm⁻¹), 320 nm (ε = 11,000 M⁻¹cm⁻¹) and 325 nm (ε = 10,500 M⁻¹ cm⁻¹), respectively, using PMT detection. In each experiment, 5–7 traces were averaged and kinetic parameters obtained by curve fitting using OriginPro or Kaleidagraph software. Iterative rounds of fitting were carried out until both the residual plot and the r² correlation value were optimized.

Reaction of Y₇₃₁F/Y₇₃₀NH₂Y-α2, Y₄₁₃NH₂Y-α2, and Y₇₃₁F/Y₄₁₃NH₂Y-α2 with wt β2 monitored by EPR and SF UV-vis spectroscopies

Y₇₃₁F/Y₇₃₀NH₂Y-α2, Y₄₁₃NH₂Y-α2, or Y₇₃₁F/Y₄₁₃NH₂Y-α2 was reacted with wt β2 in the presence of CDP/ATP at 25 °C as described for NH₂Y-α2s above. For the EPR and SF experiments, α2 and β2 concentrations were 50 μM and 5 μM, respectively. The EPR reactions were hand quenched at 40 s. The EPR spectrum of a putative NH₂Y₄₁₃• was obtained subsequent to subtraction of the spectrum of Y₁₂₂• (in the wt-β2:mutant-α2 complex) as described above.

RESULTS

Reaction of Y₇₃₀ NH₂Y-α2 (or Y₇₃₁NH₂Y–α2) with Y₃₅₆F-β2 monitored by EPR and SF spectroscopies

It was previously demonstrated that reaction of either Y₇₃₀NH₂Y-α2 or Y₇₃₁NH₂Y-α2 with wt-β2, CDP, and ATP resulted in the loss of Y₁₂₂• (λ_max = 410 nm, g_av of 2.0047) concomitant with formation of a new radical species, assigned as NH₂Y• on the basis of its absorbance and EPR spectroscopic features (λ_max = 320-325 nm, g_av of 2.0043) (15). In this paper, experiments with Y₃₅₆F-β2, in which the radical propagation pathway is disrupted with a Y to F mutation, were conducted to establish that trapping of NH₂Y• at residues 730 and 731 is the result of the participation of these residues in radical transfer, rather than the consequence of introducing a pathway-independent thermodynamic trap into α2. If formation of NH₂Y₇₃₀• or NH₂Y₇₃₁• requires a redox-active Y₃₅₆ residue, then no NH₂Y• should be observed. Reactions were carried out with CDP/ATP, Y₃₅₆F-β2 and Y₇₃₀ NH₂Y-α2 (or Y₇₃₁NH₂Y–α2) and quenched at 77 K after 20 s and 2 min. Analysis of the EPR spectra, subsequent to subtraction of the spectrum of the resting Y₁₂₂• at time zero gave no evidence of NH₂Y• at either time point for reactions with Y₇₃₀NH₂Y-α2 or with Y₇₃₁NH₂Y–α2 (Figure S1). In addition, spin quantitation of the signal at time zero in comparison to times 20 s and 2 min revealed no significant loss of spin. A control with non tagged Y₃₅₆F-β2 gave the same result as that observed with His-Y₃₅₆F-β2, demonstrating that the His-tag does not interfere with radical formation. Oxidation of Y₃₅₆ is thus a prerequisite for hole migration into the α2 subunit.

To determine if NH₂Y• was formed transiently, the reactions of NH₂Y-α2s with Y₃₅₆F-β2, CDP, and ATP were monitored by SF UV-vis spectroscopy at 325 nm (λ_max of NH₂Y₇₃₀•) or 320 nm (λ_max of NH₂Y₇₃₁•) and at 410 nm (λ_max of Y₁₂₂•). The results for Y₇₃₀NH₂Y-α2 with Y₃₅₆F-β2 are shown in Figure 2, and for Y₇₃₁NH₂Y-α2 with Y₃₅₆F-β2 in Figure S2. A small increase at 325 (320 nm) corresponding to ~2 % of total initial Y₁₂₂• was observed. However, these spectral changes were not correlated with the expected decrease in absorbance at 410 nm due to loss of the Y•. In fact, a small increase in this wavelength was observed (Figures 2 & Figure S2, inset). The lack of correspondence between the spectral changes at 325 (or 320) nm and 410 nm indicates these features are not related to NH₂Y• formation. The nature of these changes is not understood at present, but similar changes have been observed previously under different reaction conditions (37) and may be related to minor structural changes of the Y₁₂₂• diferric cluster upon α2:β2 complex formation (Seyedsayamdost, Minnihan, and Stubbe, unpublished results). The SF UV-vis and EPR spectroscopic data together indicate that no NH₂Y• is formed at residues 730 or 731 with Y₃₅₆F-β2, with our lower limit of detection approximated as less than 2% (by SF UV-vis) and less than 3% (by EPR) of the total initial Y₁₂₂•. These data support the conclusion that a redox-active residue at 356 is essential for radical transfer into α2.

FIGURE 2.

SF UV-vis spectroscopy of the reaction of Y₃₅₆F-β2 with Y₇₃₀NH₂Y-α2 in the presence of CDP/ATP. The reaction was monitored at 325 nm (trace A, λ_max of NH₂Y₇₃₀•) and at 410 nm (trace B, λ_max of Y₁₂₂•). Insets show magnified views of the initial 0.5 s of each trace.

Reaction of Y₇₃₁F/Y₇₃₀NH₂Y-α2 with wt-β2 monitored by EPR and SF spectroscopies

To assess the role of residue Y₇₃₁ in the radical transfer chain, the double mutant Y₇₃₁F/Y₇₃₀NH₂Y-α2 was isolated, characterized, and examined for NH₂Y• formation. The protein was purified to homogeneity and its specific activity measured to be 36 nmol/min/mg. This low activity, approximately 1.4% that of the wild-type enzyme, is likely associated with co-purifying endogenous α2, rather than activity inherent to the mutant protein. This result is in agreement with the activity of Y₇₃₁F-α2 (26 nmol/min/mg) and contrasts to activities measured in Y₇₃₀NH₂Y-α2 or Y₇₃₁NH₂Y-α2 (110-150 nmol/min/mg) (15).

Y₇₃₁F/Y₇₃₀NH₂Y-α2 and ATP were mixed with wt-β2 and CDP and the reaction examined by EPR and SF spectroscopy. With EPR, the reactions were quenched at 40 s or 2 min and analyzed as described above. No NH₂Y• formation was detected in either case and the total spin relative to starting Y₁₂₂• remained unchanged (Figure S3). The SF UV-vis experiments gave results similar to those observed with Y₃₅₆F-β2 at both 325 nm and 410 nm (Figure S3). Thus formation of NH₂Y₇₃₀• also requires a redox active Y₇₃₁ for radical propagation.

Expression, purification, and characterization of Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y -α2

As an additional test of the pathway dependence of NH₂Y• formation at Y₇₃₀ and Y₇₃₁, NH₂Y was site-specifically incorporated at a Y in α2 that is thought not to participate in the radical transfer pathway (Figure 1). Residue Y₄₁₃ was selected, as the X-ray structure of E. coli α2 (14) reveals distances of 5.0 Å or 5.2 Å between the phenolic oxygen of Y₄₁₃ and that of Y₇₃₀ or Y₇₃₁, respectively (Figure 3). Thus, Y₄₁₃ is a distance from Y₇₃₀ and Y₇₃₁ that is reasonable for ET between these residues. Y₇₃₁ and Y₇₃₀ are separated by 3.3 Å in the same structure. Alignment of 142 primary sequences of class Ia/b α2s reveals that while 413 is not conserved, it is always an aromatic amino acid. Y, F, and W occupy this position 65%, 21%, and 14% of the time, respectively. Thus, the proteins Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2 were expressed and purified to be examined for NH₂Y• formation.

FIGURE 3.

Location of Y₄₁₃ (magenta) in relation to Y₇₃₁ (blue), Y₇₃₀ (blue), and C₄₃₉ (yellow) of the radical propagation pathway within the α subunit. Y₄₁₃ is located 5.0 and 5.2 Å from Y₇₃₀ and Y₇₃₁, respectively, with distances measured between phenolic oxygens (red). The figure was generated as a PyMol (www.pymol.org) rendition of PDB ID: 1RLR (14).

Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2 were expressed in a fashion analogous to other NH₂Y-substituted α2s. Their purification, however, proved more difficult than for previous NH₂Y-sustituted α2s. Protein of high purity (as judged by SDS-PAGE) was eventually isolated via anion-exchange FPLC with an optimized salt gradient, but at the expense of overall protein yield (Figure S4).

The activities of the purified mutants were determined by the spectrophotometric and/or radioactive RNR assays. The specific activity of Y₄₁₃NH₂Y-α2 was 46% of wt α2 (~1200 nmol/min/mg), while that of Y₇₃₁F/Y₄₁₃NH₂Y-α2 was 1.9% of wt α2 (~48 nmol/min/mg). The activity of Y₇₃₁F/Y₄₁₃NH₂Y-α2 is similar to that of Y₇₃₁F/Y₇₃₀NH₂Y-α2, and is likely associated with endogenous E. coli α2 that co-purified with the recombinantly expressed mutant protein.

The reduction in activity of Y₄₁₃NH₂Y-α2 relative to wt α2, the proximity of Y₄₁₃ to the putative α2:β2 interface, and the difficulty encountered in purification of the related mutants provided the impetus to determine the K_d for the interaction between this mutant and β2 using a modified version of the competitive inhibition assay developed by Climent et al. (29, 36, 42). Data analysis was modified relative to the previously published method (36) to accommodate for the condition that the active species, rather than the inhibitor, was also the species of unknown K_d (Figure S5). Analysis by this method afforded a K_d of 0.05 μM, similar to the value of 0.06 μM for wt-α2:β2 in the presence of CDP and ATP (Hassan and Stubbe, unpublished results). This finding, in conjunction with the activity of Y₄₁₃NH₂Y-α2, suggests that Y₄₁₃NH₂Y-α2 is folded and that binding between subunits is minimally perturbed. However, attempts to crystallize Y₄₁₃NH₂Y-α2 under conditions optimized for wt-α2 and successfully used to crystallize other NH₂Y-α2 mutants failed to provide crystals of Y₄₁₃NH₂Y-α2, suggesting that mutation of Y₄₁₃ produces some effect on protein stability or solubility (Uhlin and Stubbe, unpublished results).

Reactions of Y₄₁₃NH₂Y-α2 or Y₇₃₁F/Y₄₁₃NH₂Y-α2 with wt-β2, and Y₄₁₃NH₂Y-α2 with Y₃₅₆F-β2 monitored by EPR spectroscopy

EPR experiments were carried out as described above by mixing Y₄₁₃NH₂Y-α2 (or Y₇₃₁F/Y₄₁₃NH₂Y-α2) with CDP, ATP and wt-β2, or Y₄₁₃NH₂Y-α2 with CDP, ATP, and Y₃₅₆F-β2. Each reaction was quenched at 40 s. The resulting spectra, subsequent to subtraction of the resting Y₁₂₂• in the mutant α2:β2 complex, are shown in Figure 4, and the spin quantitation of the resulting NH₂Y• is given in Table 1. In the reactions of Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2 with wt-β2 (Figure 4a and 4b, respectively), the difference EPR spectrum reveals features characteristic of an NH₂Y•. For the reaction with Y₄₁₃NH₂Y-α2, the signal attributed to NH₂Y• accounts for 12% of the total spin. Differences in the spectral features relative to those previously reported NH₂Y₇₃₀• and NH₂Y₇₃₁• cannot be assigned at this stage, but perhaps report on differences in the dihedral angles of the C_β-protons on NH₂Y at position 413 versus those at positions 730 and 731. For the reaction with Y₇₃₁F/Y₄₁₃NH₂Y-α2, the signal attributed to NH₂Y• accounts for 7% of the total spin. The magnitude of the signal in both cases is substantially lower than the “on pathway” radicals, which have been shown by EPR to constitute 53% and 45% of the total spin at the same time point when NH₂Y is incorporated at 730 or 731, respectively (15). The difference EPR spectrum from the reaction of Y₄₁₃NH₂Y-α2 with Y₃₅₆F-β2 with CDP/ATP quenched at 40 s reveals no detectable NH₂Y• (Figure 4c).

FIGURE 4.

Formation of NH₂Y₄₁₃• monitored by EPR spectroscopy. In each case, CDP and ATP were mixed at 25 °C with (A) wt β2 and Y₄₁₃NH₂Y-α2, (B) wt β2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2, and (C) Y₃₅₆F-β2 and Y₄₁₃NH₂Y-α2 and quenched at 77 K after 40 s. For each reaction, the residual Y₁₂₂• signal (blue) was subtracted from the composite reaction spectrum (black) to give a putative NH₂Y₄₁₃• (red). The inset gives an expanded view of the difference spectrum, which in the case of (A) and (B) corresponds to the putative NH₂Y₄₁₃•.

TABLE 1.

Characterization of NH₂Y• formation in α:β complexes with various on- and off-pathway mutations

β:α complex	% NH2Y• (EPR)a	% NH2Y• (SF UV-vis)a	first phase kobs (s−1)	second phase kobs (s−1)
wt-β2:Y730NH2Y-α2b	53	39	12.8	2.5
wt-β2:Y731NH2Y-α2b	45	35	19.2	2.7
Y356F-β2:Y730NH2Y-α2	<3% c	<2% c	--	--
Y356F-β2:Y731NH2Y-α2	<3%	<2%	--	--
wt-β2:Y731F/Y730NH2Y-α2	<3%	<2%	--	--
wt-β2:Y413NH2Y-α2	12	8	4.8d	0.5
wt-β2:Y731F/Y413NH2Y-α2	7	12	0.04	--
Y356F-β2:Y413NH2Y-α2	<3%	NDe	--	--

^aY₁₂₂• converted to NH₂Y•, as a percent of the starting [Y₁₂₂•].

^bOriginally reported in reference 15.

^cNo conversion detected within our lower limit of detection, approximated to be 3% of the total initial Y₁₂₂• by EPR and 2% by SF UV-vis spectroscopy.

^dTaken as the average of the rate constants measured at 325 nm and 410 nm.

^eNot determined.

Reactions of Y₄₁₃NH₂Y-α2 or Y₇₃₁F/Y₄₁₃NH₂Y-α2 with wt-β2 monitored by SF UV-vis spectroscopy

Given the results of the EPR experiments above, the kinetics of formation of the NH₂Y₄₁₃• were assessed for both Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2 using SF UV-vis spectroscopy. The results with Y₄₁₃NH₂Y-α2 and Y₇₃₁F/Y₄₁₃NH₂Y-α2 are shown in Figure 5a and 5b, respectively, and are summarized in Table 1. For the reaction with Y₄₁₃NH₂Y-α2 monitored at 325 nm, the observed absorbance increase may be fit to biphasic kinetics with rate constants of 26.4 s⁻¹ and 4.2 s⁻¹. These phases correspond to 30% and 70%, respectively, of the total absorbance increase observed at 325 nm over the reaction course. The faster phase is not correlated, however, with a decrease at 410 nm, and thus is not associated with NH₂Y• formation.

FIGURE 5.

Reaction of wt β2, CDP, ATP and (A) Y₄₁₃NH₂Y-α2 or (B) Y₇₃₁F/Y₄₁₃NH₂Y-α2 monitored by SF UV-vis spectroscopy. The reaction was monitored for an increase at 325 nm (λ_max of NH₂Y₄₁₃•) and a decrease at 410 nm (λ_max of Y₁₂₂•). The reaction with Y₄₁₃NH₂Y-α2 (A) was complete within 2 seconds. The reaction with Y₇₃₁F/Y₄₁₃NH₂Y-α2 was monitored over 200 s (B), with the first 30 s expanded in the inset. Exponential fits to the data are indicated by solid black lines.

The A_{410 nm} decrease can be fit to biphasic kinetics with rate constants of 5.3 s⁻¹ 410 nm and 0.5 s⁻¹, corresponding to 47% and 53% of the total absorbance change at 410 nm during the reaction. The rate constant of 5.3 s⁻¹ agrees well with the rate constant of 4.2 s⁻¹ at 325 nm, and together suggest the formation of NH₂Y₄₁₃•. Using the extinction coefficients of NH₂Y• and Y•, the extent of formation of NH₂Y₄₁₃• was determined to be 8% of the starting Y₁₂₂•. This is substantially lower than the amount of NH₂Y• observed in SF UV-vis studies with NH₂Y at 730 (39%) and 731 (35%) under analogous reaction conditions (15). While a rate constant of 4-5 s⁻¹ is fast enough to be involved in production of dNDPs, the collective data on the three Y₄₁₃ mutant complexes suggest NH₂Y₄₁₃• formation is off-pathway and we propose that it is not functionally relevant to catalysis.

For the reaction of Y₇₃₁F/Y₄₁₃NH₂Y-α2, a slow increase in A_{325 nm} was observed concomitant with a loss of absorbance at 410 nm (Figure 5b). The first 30 s of data collected at both wavelengths were fit to a monoexponential function with a rate constant of 0.04 s⁻¹. The SF data suggest that 12% of the starting Y₁₂₂• is converted to NH₂Y₄₁₃• over the course of 30 s. The slow rate of radical formation, 320 to 475-fold slower than NH₂Y• formation at 730 (13 s⁻¹) or 731 (19 s⁻¹), suggest off-pathway oxidation.

DISCUSSION

The rate and extent of NH₂Y• formation in the presence of the β subunit and the S/E pair CDP/ATP (when NH₂Y has been site-specifically incorporated into α at residue 730, 731, or 413) may be used as indicators of the participation of a specific residue in a defined pathway for C₄₃₉ oxidation in RNR (Figure 1). We have previously established that formation of NH₂Y• is biphasic, giving rise to rate constants and amplitudes (as a percent of initial Y₁₂₂•) for NH₂Y• formation of 19 s⁻¹(24%) and 2.7 s⁻¹ (11%) at 731, and 13 s⁻¹ (20%) and 2.5 s⁻¹ (19%) at 730 (15). We have analyzed the kinetics of NH₂Y• formation with all S/E pairs and argue that the slower rate constant is kinetically competent for deoxynucleotide formation (Seyedsayamdost, Minnihan, and Stubbe, unpublished results). An unexpected observation of our initial studies with the 730- and 731-substituted proteins was the ability of these NH₂Y-α2s to catalyze deoxynucleotide formation, suggesting that NH₂Y• at either position may be an intermediate on the pathway or may replace Y₁₂₂• as a radical initiator.

In our current studies, replacement of Y with the non-oxidizable F in the pathway at positions 356 in β or in 731 in β, and incorporation of NH₂Y within the pathway after the block, resulted in no detectable NH₂Y• by EPR analysis at 20 s or 2 min (Table 1). Since the EPR signal is always a composite of Y• and NH₂Y•, a subtraction of the Y• signal is required for spin quantitation of the two species. The broadness of the Y• signal relative to the NH₂Y• makes this subtraction straightforward, with the lower limit of detection of an NH₂Y• estimated to be 3% of the total spin. It should be noted that cleaner subtractions are achieved when Y₁₂₂• in an α:β complex is used as the reference, as the hyperfine interactions of Y₁₂₂• in β appear to be subtly affected by complex formation with α.

To establish that NH₂Y• is not generated transiently in the presence of a Y to F mutation, SF UV-vis was conducted and the reaction was monitored at 325 (320) nm and 410 nm. An increase in the former absorption is associated with NH₂Y• formation at 730 (731) and a loss of the latter with Y• loss. Monitoring a reaction at 325 nm is complicated by features of the diferric cluster which also absorb in this region. Thus, we have assumed transient features observed by SF are related to NH₂Y• formation only if the rate constant and amplitude increase of the 325 nm feature parallels the rate constant and amplitude decrease of the 410 nm feature. As indicated in Figure 2, in all experiments described herein in which the pathway has been “blocked,” a small rise phase (approximately 2% of the Y•) has been observed at 325 nm. This rise is rapid (corresponding to a rate constant >20 s⁻¹), has no corresponding decrease at 410 nm, and is attributed to a small change associated with the diferric Y• cofactor. In the current analysis, this feature is not considered further. Thus, the SF data are in agreement with the EPR results, indicating that a redox-active Y at 356 and 731 is a prerequisite for NH₂Y• formation.

As noted above, at pH 7.0 NH₂Y is easier to oxidize than Y by 190 mV and thus could potentially be oxidized in an off-pathway process. To test this proposal, studies were conducted in which NH₂Y was incorporated in place of a residue proximal to the proposed pathway. Y₄₁₃ (Figure 3) is located within 5 Å of both Y₇₃₁ and Y₇₃₀, a distance feasible for a single oxidation “hop,” and thus became the target of our off-pathway studies. Three experiments were carried out using mutants in which NH₂Y was site-specifically incorporated at position 413.

In the first, NH₂Y was incorporated at residue 413 while keeping all other on-pathway residues in tact. In two additional experiments, a pathway block was introduced in the form of a Y to F mutation at either 356 or 731. When the single mutant, Y₄₁₃NH₂Y-α2, was reacted with wt-β2, CDP, and ATP, SF experiments revealed formation of NH₂Y₄₁₃• with a rate constant of 4-5 s⁻¹, corresponding to an 8% conversion of the initial Y₁₂₂• to NH₂Y₄₁₃• (Table 1). EPR analysis of the same reaction after 40 s was in agreement with the SF results, indicating the formation of a NH₂Y• species accounting for 12% of the total spin (Figure 4a). The differences in the percent conversion of Y₁₂₂• to NH₂Y₄₁₃• as measured by SF and EPR may be indicative of the error inherent to the two methods in quantitation of low levels of radical species. Alternately, the differences may arise from the 10-fold difference in protein concentration used in the two experiments (38). As argued above, the rate constant and accumulation of NH₂Y• is indicative of the relevance of that specific residue to catalysis in RNR. The rate constant of 4.2 s⁻¹ for NH₂Y₄₁₃• formation is slower by a factor of 4 and the amplitude is reduced by 75% relative to NH₂Y• formation at 731. This result suggests that NH₂Y at position 413 can be oxidized in an off-pathway fashion by either Y₇₃₁ or by Y₃₅₆. With F incorporated at 731 and NH₂Y at 413, NH₂Y• is still observed. However, a fit of the SF data for the first 30 s gave a rate constant of 0.04 s⁻¹ and an amplitude of 12%. EPR spin quantitation of the same reaction after 40 s indicated 7% of the total spin was associated with the putative NH₂Y• (Figure 4b). Finally, with F at 356 and NH₂Y at 413, no NH₂Y• was detected (Figure 4c). Thus the oxidation at 413 appears to occur predominantly through Y₃₅₆. The very slow rate constant for NH₂Y₄₁₃• formation with F at 731 (~0.2% that of NH₂Y• formation at 731, and ~2% of the steady state-rate constant for dCDP formation) and the complete absence of NH₂Y₄₁₃• with F at 356 suggest that radical formation at 413 in the NH₂Y-mutants results from off-pathway ET.

It is experimentally challenging to determine directly whether NH₂Y₄₁₃• can participate as a chemically- and kinetically-competent intermediate in C₄₃₉ oxidation. However, strong evidence that ET through 413 does not constitute a viable alternate radical pathway comes from the observation that no NH₂Y• is generated in the reaction with Y₇₃₁F/Y₇₃₀NH₂Y-α2. Y₄₁₃ cannot serve as a bypass to a block at 731, and thus cannot support the oxidation of Y at 730 required for C₄₃₉• formation and subsequent nucleotide reduction. The NH₂Y₄₁₃• observed in experiments with Y₄₁₃NH₂Y-α2 or Y₇₃₁F/Y₄₁₃NH₂Y-α2 thus results from an off-pathway oxidation and is nonfunctional.

Many examples of off-pathway oxidation exist in the literature for heme- and non-heme iron-dependent systems. Likewise, evidence for off-pathway oxidation has been reported previously in the class I E. coli RNR (19-21, 41). The phenotypes are usually slow rate constants for radical formation and low amplitudes of new radical species. Thus the experiments presented herein using F and NH₂Y as an oxidation block and trap, respectively, support the original proposal of Uhlin and Eklund for long-range oxidation of C₄₃₉ in α via a conserved pathway of defined aromatic amino acids.

ABBREVIATIONS

α

ribonucleotide reductase large subunit

ATP

adenosine 5′-triphosphate

β

ribonucleotide reductase small subunit

C•

thiyl radical

CDP

cytidine 5′-diphosphate

dC

2′-deoxycytidine

DTT

dithiothreitol

E

effector

EPR

electron paramagnetic resonance

ET

electron transfer

FPLC

fast protein liquid chromatography

NH₂Y•

3-aminotyrosyl radical

PCET

proton-coupled electron transfer

PL

photolyase

RNR

ribonucleotide reductase

S

substrate

SF

stopped-flow

TR

thioredoxin

TRR

thioredoxin reductase

W•

tryptophan radical

wt

wild-type

Y•

tyrosyl radical