The Involvement of Arg265 of Mouse Ribonucleotide Reductase R2 Protein in Proton Transfer and Catalysis*

Ribonucleotide reductase class I enzymes consist of two non-identical subunits, R1 and R2, the latter containing a diiron carboxylate center and a stable tyrosyl radical (Tyr·), both essential for catalysis. Catalysis is known to involve highly conserved amino acid residues covering a range of ∼35Å and a concerted mechanism involving long range electron transfer, probably coupled to proton transfer. A number of residues involved in electron transfer in both the R1 and R2 proteins have been identified, but no direct model has been presented regarding the proton transfer side of the process. Arg265 is conserved in all known sequences of class Ia R2. In this study we have used site-directed mutagenesis to gain insight into the role of this residue, which lies close to the catalytically essential Asp266 and Trp103. Mutants to Arg265 included replacement by Ala, Glu, Gln, and Tyr. All mutants of Arg265 were found to have no or low catalytic activity with the exception of Arg265 to Glu, which shows ∼40% of the activity of native R2. We also found that the Arg mutants were capable of stable tyrosyl radical generation, with similar kinetics of radical formation and R1 binding as native R2. Our results, supported by molecular modeling, strongly suggest that Arg265 is involved in the proton-coupled electron transfer pathway and may act as a proton mediator during catalysis.

Ribonucleotide reductase (RNR) 2 of the class I family catalyzes the reduction of nucleoside diphosphates to deoxynucleoside diphosphates and is essential for de novo synthesis of deoxyribonucleotides (dNTPs) required for DNA replication and repair (1)(2)(3)(4). The RNR enzymes isolated so far have been grouped into three main classes (I, II, and III). The classification is based upon the different oxygen dependence and metal cofactors involved in the generation of the essential catalytically active stable free radicals as well as structural differences among them (4,5). The class Ia is found in eukaryotes and some bacteria and viruses, whereas class Ib, II, and III are only found in microorganisms. The class Ia mouse enzyme is composed of two non-identical subunits called R1 and R2 proteins (3)(4)(5)(6)(7)(8). R1 is a homodimer (␣ 2 ) with a molecular mass of 2 ϫ 90 kDa, whereas R2, also a homodimer (␤ 2 ), has a molecular mass of 2 ϫ 45 kDa. The three-dimensional structures of proteins R1 and R2 are known for Escherichia coli and for a number of other species, whereas only one study concerns the structure of the active R1/R2 holoenzyme complex (9).
The activity of RNR is the first dedicated step in DNA synthesis. At the protein level the activity is determined by 1) R1/R2 interactions, 2) formation and maintenance in protein R2 of a tyrosyl radical neighboring a diiron carboxylate site, which contains two high spin ferric ions antiferromagnetically coupled by one -oxo bridge and one -carboxylate bridge, and 3) positive and negative nucleotide allosteric effectors acting in protein R1.
The overall enzyme activity is controlled by the limiting levels of protein R2 that are undetectable in the G 1 phase and increase in the S phase (10,11). In vitro, the tyrosyl radical is formed by loss of an electron and a proton in a reconstitution reaction between apoprotein R2, ferrous iron, and molecular oxygen. The R1 protein binds the substrate and allosteric effectors, whereas R2 contains the radical harboring tyrosyl residue, Tyr 177 , in mouse R2.
During catalysis it is proposed that the oxidation equivalent stored in the tyrosyl side chain in R2 is transferred through an ϳ35-Å-long pathway to a cysteine residue at the active site, Cys 429 , in the R1 protein (12)(13)(14)(15). Using site-directed mutagenesis, a pathway at the mammalian RNR has been identified as the route for the electron transfer that leads from the cysteinyl radical in R1 through the residues Cys 429 -Tyr 738 -Tyr 737 across the R1/R2 interface and then in R2 through the residues Tyr 370 -Trp 103 -Asp 266 toward the iron site (16 -19). The electron transfer process has been proposed to take place in a concerted manner, where the transfer of the electron is coupled to the transfer of a proton (20 -25). This transfer pathway has, therefore, been termed proton-coupled electron transfer (PCET) pathway or radical transfer pathway. The catalytically active cysteinyl radical on Cys 429 is proposed to act directly on the substrate. The x-ray structures of proteins R1 and R2 are in agreement with a PCET mechanism mediated by hydrogen bonds connecting the residues (15,23,25). However, so far no specific residue(s) directly involved in the transfer of the proton has been identified. In the amino acid sequence of protein R2 the arginine residue, Arg 265 , is conserved in all R2 sequences known so far. In fact, it seems to be part of the fingerprint that characterizes R2 proteins. It has also been hypothesized that it participates in the PCET pathway based on the crystal structures of E. coli R2 and Salmonella typhimurium R2F proteins (3,26,28). However, so far no data have been provided to support these hypotheses. Upon examination of several crystal structures including a high resolution E. coli R2 and three mouse crystal structures, Arg 265 was found to be consistently close to two crucial residues in the PCET pathway, Asp 266 and Trp 103 ( Fig. 1 and Refs. 28 -31). In fact, in all presently solved crystal structures of R2 protein from six different organisms the arginine residue is connected to the PCET pathway residues either by a direct hydrogen bond to the Asp 266 or in most cases via a water molecule that is located between Asp 266 and Trp 103 .
The position of Arg 265 , the fact that it is highly conserved and that it has been found to be coordinated to a water molecule residing in the vicinity of Asp 266 and Trp 103 , suggest that it may play a role in the PCET transfer process. Taking into consideration the biochemical characteristics of arginine, this residue could most likely be involved in the proton transfer part of the PCET pathway rather than in electron transfer part. It has also been suggested to be involved in R1 binding (3,23).
Arginine residues have been implicated in various systems as key members in proton-dependent transfer mechanisms. An arginine residue has been identified as the proton donor in the reduction of fumarate. In fact, in the Shewanella fumarate reductase enzyme it has been shown that arginine is capable of working as a Lewis base stabilizing the build-up of negative charge upon hydride transfer from FAD to fumarate and as a proton donor to the substrate (32,33). In the rabbit renal cortical proximal tubule protein (PepT1) Arg 282 has been found to be necessary for the proton coupling in the peptide-gated cation channel (34). Moreover, studies in the bacteriorhodopsin light-driven proton pump suggest that an arginine residue may act as a proton release group and stabilize the active site (35)(36)(37)(38)(39). An arginine residue has also been known to modulate the proton flux in the E. coli enzyme lactose permease (40). These examples establish the ability of arginine residues to act or modulate enzymatic proton transfers in biological systems. We have investigated the enzymatic activity in mouse RNR containing either native protein R2 or selected point mutants designed to probe the role of Arg 265 as a proton mediator. Point mutations include replacement of Arg 265 either by alanine, glutamate, glutamine, or tyrosine. In addition we have studied the formation of the tyrosyl radical in the reconstitution as well as the kinetics of radical formation in the mutants. Results indicate the Arg 265 is crucial to catalysis but is not directly involved in tyrosyl radical formation.

EXPERIMENTAL PROCEDURES
Expression and Purification of R2 Proteins-Site-directed mutagenesis was achieved using the QuikChange ® II site-directed mutagenesis kit from Stratagene. The mouse proteins were overexpressed in logarithmically growing Rosetta 2(DE3)pLysS bacteria containing pETR2 plasmids encoding native mouse R2 protein or the mutant mouse R2 proteins Arg to Ala (R265A), Arg to Glu (R265E), Arg to Gln (R265Q), or Arg to Tyr (R265Y). For expression of native and R2 mutants, an overnight culture of Rosetta 2(DE3)pLysS inoculated with native or mutant R2 was diluted 200 times in 4.8 liters of Luria-Bertani medium containing 34 g/ml chloramphenicol and 100 g/ml carbenicillin. The bacteria were grown at 37°C until they reached an A 595 of 0.6 and then induced with 400 M isopropyl-1-thio-␤-D-galactopyranoside and grown for an additional 3 h before harvesting. The native and mutant proteins were purified as previously described in Mann et al. (41) with a small modification. Grinding with aluminum oxide was omitted due to the presence of pLys in the vector. This preparation resulted in iron-containing R2 protein. To obtain R2 apoprotein, the same procedure was employed except that 0.2 mM EDTA was added 20 min before induction, and all purification buffers contained 0.2 mM EDTA.
Preparation of the Anaerobic Fe 2ϩ Solution-Buffer containing 50 mM Tris-HCl was thoroughly degassed in a septumsealed 250-ml bulb for several hours by flushing oxygen-free argon and evacuation. A second bulb containing Mohr's salt crystals, (NH 4 ) 2 Fe(SO 4 ) 2 , was sealed and evacuated using the same procedure. A plastic syringe was made anaerobic by repeated washing with oxygen-free argon gas. Then the appropriate volume of degassed buffer was used to prepare the desired iron concentration by transferring the buffer into the bulb containing Mohr's salt.
Reconstitution Reaction of Apoprotein R2 with Fe 2ϩ and Oxygen-Reconstitution of the iron site in the R2 proteins were carried out by mixing oxygen saturated apoprotein R2 with anaerobic Fe 2ϩ solution. Rapid freeze quench was performed with a System 1000 apparatus from Update Instruments to obtain reaction times from 20 ms to 1 s. The EPR tubes were submerged in isopentane at a temperature of Ϫ110°C. Anaerobic iron solution and apoprotein were rapidly mixed and then quenched by spraying them onto the EPR tube. The crystals were tightly packed into the tube using a packing rod made from Teflon. Isopentane from the EPR tube was completely removed by placing the tubes under vacuum for 15 min. For reaction times above 1 s, slow-freeze quench was achieved by hand mixing. A gas-tight Hamilton syringe washed several times with anaerobic Fe 2ϩ was used to reconstitute the aerobic apoprotein previously placed in an EPR tube. Reactions were quenched by immersion of the EPR tubes into cold isopentane at Ϫ110°C. To assess the reconstitution efficiency after EPR measurements, the protein was passed through a NAP TM 5 column to remove excess iron. The total iron content was deter- mined using an iron/TIBC (ferrozine) reagent set from Eagle Diagnostics.
Ribonucleotide Reductase Activity Assay-The ability of native and mutant R2 proteins to catalyze the reduction of [ 3 H]CDP in the presence of R1 protein was determined as described earlier (42). One unit is defined as the amount of protein that, in excess of R1 protein, catalyzes the formation of 1 nmol of dCDP/min at 37°C. The following reagents were incubated for 20 min at 37°C in a total volume of 50 l: 0.5 mM [ 3 H]CDP, 3 mM ATP, 10 mM dithiothreitol (DTT), 100 mM KCl, 6.4 mM MgCl 2 , 0.02 mM FeCl 3 , and 40 mM HEPES buffer, pH 7.5.
Surface Plasmon Resonance Analysis of R1-R2 Complex Formation-All surface plasmon resonance measurements were performed using Biacore 3000 equipped with research grade CM5 sensor chips. The R2 proteins were immobilized using the amino-coupling chemistry method as described earlier (11). After activation of the dextran layer, 30 ml of R2 proteins at concentrations of 40 M were injected at a flow rate of 5 l/min. The R2 proteins were allowed to interact with R1 protein at different concentrations previously equilibrated with HBS-EP buffer (10 mM HEPES, pH 7.4, 0.15 M NaCl, 3 mM EDTA, 0.005% surfactant P20) plus 2 mM DTT. Five different R1 concentrations were used to determine kinetics for the R1-R2 interactions. Increasing concentrations of pure R1 in the HBS-EP buffer containing 2 mM DTT and 0.125 mM dTTP were used to obtain kinetic data. A constant flow of 5 l/min of the same running buffer was used as eluent during the subunit interaction. The surface was regenerated between each protein R1 with 10 l of 0.5 M KCl in HBS-EP buffer. Data were collected and analyzed using the Bioeval 3.2 program to obtain relative unit values of bound R1 protein for each injection.
Electron Paramagnetic Resonance Spectroscopy-X-band EPR spectra were measured on a Bruker ESP 300 spectrometer as described earlier (43). All spectra were recorded at 20 K.
Modeling of the Arg Site in Mouse R2-To model the accessibility of the surface of the protein, the NACESS program was used (44). Insight II was also used to measure distances and minimize the structures containing the different amino acid replacements.

Activity of Apo and Reactivated R2 Proteins-Because
Arg 265 is a highly conserved residue, we were interested in studying the enzymatic activity as well as tyrosyl radical formation in several selected point mutants. The amino acids chosen to replace Arg 265 were chosen either to be small and hydrophobic (Ala), negatively charged (Glu), polar and non-hydrogen bonding (Gln), or aromatic (Tyr). The catalytic activities of the R2 point mutants were studied and compared with native R2 protein.
The specific activities found for the native and mutant R2 proteins in the mouse RNR enzyme reaction at pH 7.5 are summarized in Table 1. All proteins were assayed directly after preparation and as described in Mann et al. (41). Proteins R2 or R1 alone were used as controls. Each assay was repeated at least three times, and the reported values are the average with an approximate uncertainty of Ϯ5% as can be observed in Table 1. The R265Q mutant showed insignificant activity, similar to those reported for R2 protein alone in the absence of R1, which were employed as controls. Therefore, we considered it to be inactive. A similar case was observed with the R265Y mutant, for which an activity of about 4% was detected. For the R265A mutant, an activity of ϳ10% when compared with the native protein was observed.
A different result was obtained for the R265E mutant protein. R265E was found to have ϳ40% of the activity compared with the native R2. This is a significant activity, which indicates that this mutant is capable of catalysis.
Formation of the Stable Tyrosyl Radical in Native and Mutant R2 Proteins-The tyrosyl radical is formed in vitro in a reaction between apoprotein R2, ferrous iron, and molecular oxygen. The formation of a tyrosyl radical was studied by EPR spectroscopy. The extent of tyrosyl formation was determined for the mutants. As can be seen in Fig. 2, R265E and R265Q mutants were both capable of forming stable Tyr ⅐ in similar amounts (difference of 10%) as the native R2 protein after reactivation with iron and oxygen. A similar result was found for the R265A and R265Y proteins (data not shown). The tyrosyl radical EPR spectra did not show any qualitative differences between the mutants and the native R2.
Rate of Stable Tyrosyl Radical Formation-Previous studies showed that mutations to the hydrogen-bonded PCET pathway (e.g. D266A or W103Y) resulted in kinetic differences in tyrosyl radical formation (11,13). These mutants were also catalytically inactive (11). The present study shows that the Arg 265 mutants FIGURE 2. X-band EPR spectra of the stable tyrosyl radical of the R265E, R265Q, and native R2 proteins. The spectra were recorded at non-saturating conditions at 20 K. a.u., absorbance units. had substantially less enzymatic activity than the native enzyme, with the exception of the R265E mutant. These differences could be due to mechanistic or kinetic differences in the formation of the stable Tyr ⅐ . To assess the rate of formation of the Tyr ⅐ , we recorded the time courses of the reconstitution reaction at room temperature with a ratio [Fe(II)]/[polypeptide of R2] of 3 using apo forms of either native mouse R2 or mutant R265E proteins. The apo forms of protein R2 were obtained by adding 0.2 mM EDTA before induction and purifying them in the presence of 0.2 mM EDTA. Before reconstitution, native apoR2 had an average iron content of 1-2% and less than 5% of Tyr (Table 2). After reconsti-tution, 68% of Tyr ⅐ formation was observed with a k form of 2.15 s Ϫ1 . Both quantitatively and kinetically these values are well in accordance with previous studies (45,46). Formation of stable Tyr ⅐ was also assessed for the R265E mutant in the same manner as above. The k form of the R265E mutant R2 was 3.11 s Ϫ1 ( Table 2). The amount of Tyr ⅐ per polypeptide after reconstitution was 63%. These results show that R265E and native R2 have virtually the same quantitative and kinetic characteristics regarding Tyr ⅐ formation. It should be noted that attempts to reconstitute and increase radical concentrations in iron containing R2 proteins grown in LB media also resulted in increased radical concentrations. However, the extent of radical formation was less in those proteins, and the kinetics were slower.

R1-R2 Interaction Studies by Surface Plasmon Resonance-
The role of Arg 265 in the interaction between R2 and R1 was studied by immobilizing the native and mutant R2 proteins and allowing them to interact with the increasing concentrations of R1 protein containing DTT and the allosteric effector dTTP. Previous studies have shown that dTTP increases the affinity between the R1 and R2 proteins (47). Sensorgrams showed an increase in bound R1 with increasing concentrations or R1 protein consistent with previous studies (data not shown). The kinetic parameters for R1-R2 binding are summarized in Table  3. As can be observed, no significant differences were observed between the native R2 protein and the mutants R265E and R265Q.  (29)), cobalt-substituted mouse (light gray, PDB 1H0O (30)), and reduced mouse R2 at pH 6.0 (gray, PDB code 1W69 (31)). Stars represent water molecules.  the backbone is observed, resulting in a further 0.9 Å separation between the carboxylate of Asp 266 and the guanidino group of Arg 265 . Yet the water molecule retains its position between both residues (Fig. 3C). An analysis of the accessible area of the protein surface of the E. coli and the mouse R2 proteins reveals that the Arg 265 is much more exposed to the surface than the Asp 266 and Trp 103 (Table 4 and Fig. 4). The program calculates the solvent-accessible surface of a protein defined as a locus of the center of a probe sphere (representing the solvent molecule) as it rolls over the van der Waals surface of the protein (44). A larger surface area indicates a residue more exposed to the exterior. Hence, a residue such as the His 173 , which is one of the iron ligands, is found to be completely buried and with a calculated surface area of zero, whereas Arg 79 in a loop region fully exposed to water has an estimated area of 231.3 Å 2 (Table 4). From this it is evident that Arg 265 with a surface area of 95.8 Å 2 is fairly close to the surface compared with Trp 103 and Asp 266 with surface areas around 5 Å 2 . Arg 265 is, therefore, a viable candidate to act as a proton mediator from the solvent or even to be hydrogenbonded to and mediate proton transfer from the catalytically essential Tyr 370 . The Tyr 370 residue is not seen in the crystal structures since it is located in the flexible C terminus of R2, but it has been postulated to bridge the PCET pathway between the R2 and R1 proteins (17,19,24,27,48). Crystal structures of the R265E and R265Q mutants are under way to elucidate more on this question.

DISCUSSION
PCET reactions have been extensively studied in various systems including respiration and photosynthesis (for review, see Refs. 20 and 22). In RNR a key feature of the enzyme involves an unusual long PCET chain that connects the essential tyrosyl radical in the R2 subunit to the putative thiyl radical at the catalytic site in the R1 subunit. This dedicated transfer has been established by a combination of three-dimensional structure determinations, site-directed mutagenesis, and phylogenic sequence comparisons. Extensive site-directed mutagenesis data support the hypothesis that the PCET process occurs via a series of hydrogen-bonded amino acid residues in the R1 and R2 proteins (for review, see Refs. 3 and 25). Mutant proteins designed to alter the PCET pathway in this enzyme have resulted in catalytically inert or substantially diminished proteins (18,43,45).
The present study shows the importance of Arg 265 during catalysis and explores the role of this residue as a proton mediator in the PCET pathway. This highly conserved residue is critical for catalysis and in conjunction with other residues may regulate the proton transfer process in the R2 subunit of RNR.
The Role of Arg 265 in Proton Transfer-Our findings show that changes in the Arg 265 site via point mutations result in differences in enzymatic activity. However, the kinetics of tyrosyl radical formation and the ability to bind R1 protein remained mostly unaltered. Therefore, it can be argued that Arg 265 residue is most likely involved in the proton transfer side of the radical transfer pathway that is necessary for catalysis to occur.
Upon substitution of Arg 265 by alanine, the specific activity decreased considerably when compared with native R2 ( Table 1). The substitution by alanine should confer the inability to perform direct proton transfer due to the nonpolar characteristics of alanine. The low but significant activity observed in this mutant may indicate that some proton transfer during catalysis is still possible. One explanation may be that replacement with alanine, which is a much smaller residue than arginine, leaves the possibility for the water molecule to position itself close to the Asp 266 and that this water molecule mediates proton transfer. Similarly, such an interpretation may also explain the lack of activity in the R265Q mutant, in which the bulky residue may hinder the water molecule to take up its required space. The low but significant activity of the R265Y mutant may be that, despite the bulky tyrosine side chain which may prevent the binding of an adjacent water molecule, the OH group of tyrosine itself may participate to a small extent in the proton transfer. Histidine ligands to the iron are considered as non-exposed to the surface whereas Arg 79 of the mouse structure located in a loop was used as a reference depicting a fully exposed residue.   Table 4).

Residue numbering
Surprisingly, when Arg 265 was substituted by Glu, the mutant had a significant activity, about 40% that of the native R2. At pH 7.5 it may be expected that Glu may be deprotonated given the intrinsically low pK a of ϳ4 in solution. It can be argued that the overall reduced activity may result from the presence of two negatively charged residues, Asp 266 and Glu 265 , in close proximity. These electrostatic alterations may result in a change of the pK a most likely of the Asp 266 , obstructing the proton transfer process and resulting in the observed activity.
Substitution of arginine with glutamine results in total loss of activity. The loss of activity may be due to the fact that glutamine is a non-proton donor/acceptor and, hence, unable to participate in the proton transfer process. Glutamine is also devoid of a charge, and its size may prevent interaction of the water molecule and Asp 266 .
Previous studies have suggested Arg 265 involvement in R1 binding (3,23). Our results showed no evidence that Arg 265 participates in R1 binding. Hence, we propose that the crucial role of Arg 265 in catalysis is due to its participation in PCET.
Kinetics of Tyrosyl Radical Formation-All studied Arg 265 mutants were capable of forming tyrosyl radicals to the same extent and with similar kinetics as in native R2 protein. This indicates that Arg 265 is not directly involved in the tyrosyl radical generation. This in contrast to earlier observations on mutants to Asp-266 and Trp-103, residues that are proposed participants in the long range radical transfer chain. For these mutants dramatic differences in the kinetics of the radical formation were observed (43,45).
The Arg 265 Environment-Examination of the crystal structures showed a highly conserved water molecule that appears to bridge the Asp 266 and Arg 265 . This water molecule may play a role during deprotonation of Asp 266 in the PCET pathway. Although Asp 266 may be the major proton acceptor/donor in the PCET pathway, the Arg 265 residue is crucial for enzymatic activity. Similar observations have been reported in the bacteriorhodopsin photocycle in which an arginine residue interacts with an aspartate regulating proton release. The aspartate is directly involved in deprotonation processes, whereas the positively charged arginine is thought to modulate the aspartate pK a (36,39).
Our results suggest that Arg 265 is a good candidate to play a similar role. The proximity of Arg 265 to the known catalytically active Trp 103 and Asp 266 may also suggest another role for the Arg 265 residue in which it stabilizes the structure of the pair, most likely and, particularly, the Asp 266 residue. Further analysis is on its way including structures of these R2 mutants and other double mutants to fully assess the function of the Arg 265 and gain more insight into proton-coupled electron transfer processes in the R2 subunit of RNR.