Evidence by Mutagenesis that Tyr370 of the Mouse Ribonucleotide Reductase R2 Protein Is the Connecting Link in the Intersubunit Radical Transfer Pathway*

Ribonucleotide reductase catalyzes all de novo synthesis of deoxyribonucleotides. The mammalian enzyme consists of two non-identical subunits, the R1 and R2 proteins, each inactive alone. The R1 subunit contains the active site, whereas the R2 protein harbors a binuclear iron center and a tyrosyl free radical essential for catalysis. It has been proposed that the radical properties of the R2 subunit are transferred ∼35 Å to the active site of the R1 protein, through a coupled electron/proton transfer along a conserved hydrogen-bonded chain, i.e. a radical transfer pathway (RTP). To gain a better insight into the properties and requirements of the proposed RTP, we have used site-directed mutagenesis to replace the conserved tyrosine 370 in the mouse R2 protein with tryptophan or phenylalanine. This residue is located close to the flexible C terminus, known to be essential for binding to the R1 protein. Our results strongly indicate that Tyr370 links the RTP between the R1 and R2 proteins. Interruption of the hydrogen-bonded chain in Y370F inactivates the enzyme complex. Alteration of the same chain in Y370W slows down the RTP, resulting in a 58 times lower specific activity compared with the native R2 protein and a loss of the free radical during catalysis.

The crystal structure of the R2 protein from E. coli and later also from mouse (8,9) showed the radical to be buried in a hydrophobic pocket 10 Å from the surface of the protein. It has not yet been possible to co-crystallize the R1-R2 protein complex, but computer modeling localized the active site of the R1 protein to a deep cleft 25 Å away from the proposed interaction area of the R2 protein (10). To explain the involvement of the R2 tyrosyl radical in catalysis, a long-range electron transfer pathway between the two subunits was suggested involving the following conserved residues (mouse numbering): (Fe 1 ), His 173 ,Asp 266 ,Trp 103 in the R2 protein and Tyr 738 ,Tyr 737 , Cys 429 in the R1 protein (10,11). In the following, we will use the mouse numbering system, if not otherwise specified.
We have earlier presented strong experimental support for the R2 part of this hydrogen-bonded electron transfer pathway using site-directed mutagenesis to replace the conserved Trp 103 in mouse R2 with tyrosine or phenylalanine and the conserved Asp 266 with alanine (12,13). Similar observations have been made in the E. coli system (14,15). Recently, we could demonstrate that the hydrogen-bonded pathway in R2 is used not only during catalysis, but also during the generation of the mouse R2 protein iron/tyrosyl free radical center (16). In the radical generation reaction, we propose that a reducing equivalent is delivered from ferrous iron in the solution as a coupled electron/proton pair, i.e. a hydrogen radical. Together with two electrons from the binuclear iron center and one electron from Tyr 177 , this gives the four electrons required to reduce one O 2 to water (17). Theoretical calculations support our model of a radical transfer pathway (RTP) 1 since it minimizes charge separation and is more energetically favorable than a pure electron transfer between weakly coupled redox centers (18).
The C-terminal residues 344 -375 in the E. coli R2 protein and 354 -390 in the mouse R2 protein cannot be visualized in the x-ray crystal structure (8,9). Therefore, it has not been possible to determine which residue in the R2 protein links the R2 RTP to the R1 RTP, although a conserved tyrosine (Tyr 370 ) has been suggested (19). NMR studies showed that the mouse R2 protein in solution contains ϳ30 highly flexible residues, most of which can be attributed to the C terminus of the polypeptide chain (20). The C terminus is required for complex formation between the R1 and R2 proteins (21)(22)(23), and upon binding to the R1 protein, the mobile R2 protein tail becomes rigid (20). Peptides and peptidomimetics corresponding to the R2 protein C terminus specifically inhibit ribonucleotide reductase (24 -27). They may have a dual function by both interfering with R1-R2 complex formation and the hydrogen radical transfer between the subunits.
To identify the linking amino acid residue in the radical transfer pathway between the R1 and R2 subunits, we substituted the conserved Tyr 370 in the mouse R2 protein with tryptophan or phenylalanine. Tryptophan has the potential to form a hydrogen-bonded linkage, whereas phenylalanine cannot form such a linkage. We here report data strongly supporting Tyr 370 as the connecting link in the intersubunit RTP.

EXPERIMENTAL PROCEDURES
Plasmids-The T7 RNA polymerase-responsive vector pET3a (28,29) containing the cDNA encoding the mouse R2 protein (7) was used as a source for site-directed mutagenesis. The ClaI/EcoRI sites in the original pET3a vector were removed (12), resulting in plasmid pETR2-Cla/R1, containing unique EcoRI and ClaI sites.
Oligonucleotide-directed Mutagenesis-Tyr 370 was mutagenized to phenylalanine or tryptophan (Y370F and Y370W) by the polymerase chain reaction technique, using the primers indicated below and pETR2-Cla/R1 as a template. All oligonucleotides used were synthesized on an Applied Biosystems Model 392 DNA/RNA synthesizer. The primers were 5Ј-CTGTCGAATTCGACATGACTCCCATCCTCTGAAA-CTCGCC (Y370F) and 5Ј-CTGTCGAATTCGACATGACTCCCATCCTC-TGACCCTCGCC (Y370W). Boldface letters denote the mutated codon, and the recognition site for EcoRI used in the subcloning is underlined. The same upstream primer was used for both of the mutants: 5Ј-GGA-CAAAGAGGCTACGTATGGAGAACGCGT-3Ј. The underlined nucleotides denote an MluI recognition site. A polymerase chain reaction mixture containing 1 ng of pETR2-Cla/R1, 330 ng of each primer, and 0.2 mM dNTPs (Amersham Pharmacia Biotech) in a buffer recommended for Pfu polymerase (Stratagene) was preheated for 5 min at 97°C before the addition of Pfu polymerase. After 30 cycles (94°C for 1 min, 55°C for 30 s, and 72°C for 1 min), EDTA was added to a final concentration of 10 mM, and the two polymerase chain reaction products were purified by phenol/chloroform extraction. The mutated fragments were recovered by digestion with MluI and EcoRI, resulting in a fragment of 459 base pairs. This fragment was finally cloned into pETR2-Cla/R1 digested with the same enzymes.
To amplify the mutated constructs, they were transfected into E. coli strain MC1061 to permit large-scale plasmid preparations. To ensure that no unwanted mismatches had been incorporated during the polymerase chain reaction, both constructs were verified by doublestranded DNA sequencing of the area covering the subcloned fragments. Finally, the constructs pETR2Y370F and pETR2Y370W were transfected into E. coli strain BL21(DE3) pLysS (29), which enables isopropyl-1-thio-␤-D-galactopyranoside-induced overexpression of the R2 proteins.
Expression and Purification of the Mutated and Native R2 Proteins-The native or mutated R2 proteins were overexpressed by induction of logarithmically growing BL21(DE3) pLysS cells containing the native or mutated pETR2 plasmid. Three 6-liter flasks, each with 1.7 liters of LB medium including carbenicillin (50 g/ml) and chloramphenicol (25 g/ml), were inoculated with 5 ml of an overnight culture of the desired bacterial construct. The cells were then grown and induced with isopropyl-1-thio-␤-D-galactopyranoside as described earlier (7). Bacterial cells, resuspended in 100 ml of 50 mM Tris-HCl (pH 7.6), were lysed by freezing at Ϫ70°C and gentle thawing at 25°C in a water bath. The lysed cells were then centrifuged in a Beckman Ti-70 rotor at 45,000 rpm for 45 min at 4°C. The subsequent streptomycin sulfate precipitation, ammonium sulfate precipitation, and ion-exchange chromatography on a DEAE-cellulose column (DE52, Whatman) were carried out according to Mann et al. (7) with one exception: no EDTA or protease inhibitor (phenylmethylsulfonyl fluoride) was used in the buffer solutions.
The purity of the proteins was assessed by SDS-polyacrylamide gel electrophoresis using 12% gels stained with Coomassie Brilliant Blue. The concentration of pure protein fractions was determined using the molar extinction coefficient ⑀ 280 -310 ϭ 62,000 M Ϫ1 cm Ϫ1 (7).
Preparation of the R1 Protein-The recombinant mouse R1 protein was obtained as described previously (30). Pure R1 protein fractions from the dATP-Sepharose chromatography were precipitated with solid ammonium sulfate (0.523 g/ml, nominally 80%). After centrifugation at 12,000 ϫ g for 10 min at 4°C, the protein precipitate was dissolved in 50 mM Tris-HCl (pH 7.6) and equilibrated with 50 mM Tris-HCl (pH 7.6) and 0.1 M KCl on a Sephadex G-50 column.
Subunit Interaction-The interaction of the mutated R2 proteins with the R1 protein was studied at 22°C by biosensor analysis using the Biacore AB (BIAcore) method as described earlier (12,31). Pure fractions of the native R2 protein or mutated proteins were immobilized on a CM5 sensor chip (Biacore AB, Uppsala, Sweden) as described previously (31). Approximately 0.17 nmol of dimeric native R2 protein or mutated R2 protein was injected during the amino coupling procedure. During the immobilization, there was a constant flow (5 l/min) of HBS running buffer (10 mM Hepes, 0.15 M NaCl, 5 mM MgCl 2 , 1 mM EDTA, 0.05% (v/v) Surfactant P20 (Biacore AB) (pH 7.4)).
Subunit interaction was studied by allowing the immobilized R2 protein to interact with two different concentrations of R1 protein (200 and 800 nM dimeric protein) in a solution of HBS running buffer supplemented with dithiothreitol and dTTP (2 and 0.1 mM final concentrations, respectively). The same buffer was used as eluent during the subunit interaction. After the first R1 injection, the surface was regenerated with 10 l of 0.5 M KCl in HBS running buffer prior to the next injection. Increasing subunit interaction was monitored as increasing surface plasmon resonance, detected as resonance units (RU).
Reconstitution of the Iron/Radical Center-Before use, the native and mutated R2 apoproteins were reactivated with iron. An anaerobic ferrous ammonium sulfate solution was freshly prepared as described previously (16). The aerobic R2 protein was incubated for 30 s at 25°C with a 3-fold molar excess over the R2 polypeptide of the anaerobic ferrous ammonium sulfate. Excess iron was removed by equilibrating the protein solution with 50 mM Tris-HCl (pH 7.6) and 100 mM KCl on a Sephadex G-25 column.
EPR Analyses-EPR spectra at 9.3 GHz were measured on a Bruker ESP 300 spectrometer using a standard rectangular Bruker EPR cavity (ER4102T) equipped with an Oxford helium flow cryostat or a cold finger Dewar flask for measurements at 77 K. Spin concentrations were determined by comparison of double integrals of EPR spectra with that of the tyrosyl radical in the mouse wild-type R2 protein that had been calibrated with a 1 mM Cu 2ϩ standard. Rapid freeze-quench experiments were performed as described (16).
Iron Assay and Light Absorption Spectra-The amount of iron/R2 polypeptide was determined by a colorimetric method (32). Light absorption spectra of the reactivated native R2 protein (20 M polypeptide) and mutated R2 proteins (44 M polypeptide) were recorded between 280 and 500 nm using a Perkin-Elmer Lamda 40 UV-visible spectrometer.
Ribonucleotide Reductase Activity-The ability of the mutated R2 proteins to catalyze reduction of [ 3 H]CDP in the presence of the R1 protein was determined as described previously (33), where 1 unit corresponds to the amount of enzyme that catalyzes the formation of 1 nmol of dCDP/min at 37°C. In the assay, the following reagents were incubated in a final volume of 50 l: 25 nmol of [ 3 H]CDP (specific activity, 20,000 cpm/nmol), 2 mol of 4-(2-hydroxyethyl)-1-piperazinesulfonic acid buffer (pH 7.6), 0.15 mol of ATP, 0.32 mol of MgCl 2 , 0.5 mol of dithiothreitol, 5 mol of KCl, 1 nmol of FeCl 3 , and enzyme as indicated.
Kinetics of the Tyrosyl Radical during CDP Reduction-A mixture containing 50 mM Tris-HCl (pH 7.6), 100 mM KCl, 10 mM MgCl 2 , 10 mM dithiothreitol, 5 mM ATP, 50 M R1 protein, and 25 M reactivated native or mutated R2 protein was freshly prepared. The reaction was started by the addition of CDP to a final concentration of 0.5 mM. The mixture was transferred to an EPR tube and incubated at 27°C for a certain time and then quenched by immersing the EPR tube into liquid nitrogen (77 K). Formation of transient species and quantification of tyrosyl radical content were made by EPR spectroscopy.
The iron chelator desferrioxamine was used to inhibit reconstitution of the tyrosyl free radical during the CDP reduction (34). The reactivated Y370W R2 protein was incubated at 37°C in the same reaction mixture as described above supplemented with 100 M desferrioxamine. Control samples were incubated in the same solution, but in the absence of the R1 protein and CDP.

Expression and Purification of the Mouse Y370F and Y370W
R2 Proteins-The construction of the pETR2Y370F and pETR2Y370W R2 expression vectors and the expression and purification of the mutated R2 proteins were described under "Experimental Procedures." The procedures resulted in a yield of 10 -15 mg of R2 apoprotein/liter of bacterial culture. SDSpolyacrylamide gel electrophoresis of the purified mutated R2 proteins showed them to be as pure as the native R2 protein (Fig. 1). All preparations showed the presence of full-length R2 protein (45 kDa) as well as variable amounts of N-terminally degraded forms (43.5 and 41 kDa). However, as shown earlier (7), this truncation does not affect the catalytic activity of the R2 protein; indeed, deletion of up to 60 amino acid residues from the N-terminal end does not affect the specific activity of the recombinant mouse R2 protein. 2 Subunit Interaction-It is known that the C-terminal part of the R2 protein is essential for binding to the R1 protein. We therefore investigated if the Y370W and Y370F mutations affected R1 binding, using biosensor analyses as described previously (12,31). The native and mutated R2 proteins were immobilized to the sensor chips, yielding 1500 -2200 immobilized RU, where 1 RU corresponds to a surface concentration of 1 pg of protein/mm 2 of a 100-nm-thick dextran layer (35). Fig. 2 shows the sensorgram for the interaction between the R1 protein and the native or mutated R2 protein. The association and dissociation of the R1-R2 complex can be followed as changes in surface plasmon resonance, which are detected as RU in the sensorgram. After injection of the R1 protein in the running buffer, the association phase in Fig. 2 was detected as an increase in RU. The association reached a steady-state equilibrium level, and this occurred faster with increasing concentrations of R1 protein. When the injection of the R1 protein was stopped and only running buffer with dTTP was used, the R1-R2 complex started to dissociate as seen by the decrease in RU. As shown in Fig. 2, there was no significant difference in shape between the response curves for the native and mutated R2 proteins. Control experiments in which the same concentrations of R1 protein were injected on sensor chips without the immobilized R2 proteins gave only an insignificant response of ϳ50 RU (data not shown). Therefore, we conclude that the R2 protein point mutations did not influence R1 protein binding in a significant way.
Properties of the Iron/Radical Center in the Y370F and Y370W R2 Proteins-The iron/radical center in the mutated proteins was reconstituted by the addition of a 3-fold molar excess of an anaerobic ferrous iron solution to an aerobic protein solution (see "Experimental Procedures"). EPR spectroscopy of the reactivated proteins, shown in Fig. 3 (traces B and  C), displayed a normal EPR signal with the same hyperfine pattern of the tyrosyl radical as for the native R2 protein (Fig.   3, trace A). The comparison of the microwave saturation behavior of the tyrosyl free radical in the two R2 protein mutants with that of the wild-type R2 protein at 20 and 30 K did not show any significant difference in the P 50 value. Therefore, the mutation did not influence the magnetic interaction of the tyrosyl free radical with the metal site, indicating that the radical structure is not disturbed. Determination of radical and iron content showed that the Y370F and Y370W R2 proteins are similar to what was previously reported for the native R2 protein (7), with 0.6 and 0.8 nmol of radical and 1.8 and 2.2 nmol of iron/nmol of R2 polypeptide, respectively.
The diferric iron center and the tyrosyl radical have a very characteristic light absorption spectrum (5-7). The mouse R2 protein iron center light absorption band is clearly visible at 2 L. Thelander, unpublished results. 329 nm, and the absorption of the tyrosyl radical has a sharp peak at 417 nm (Fig. 4A). Fig. 4 shows the spectra of the reactivated native and mutated R2 proteins recorded between 280 and 500 nm. The arrows in Fig. 4A (the native protein) denote the iron center and radical center absorption bands. Clearly, the spectra of the two mutated proteins (Fig. 4, B and  C) show the same characteristics. Together, these results show that the point mutations did not affect the ability to form a normal diferric iron/tyrosyl radical center.
Enzyme Activity-The enzyme activity of the reactivated Y370W and Y370F R2 proteins was measured in the presence of the R1 protein using [ 3 H]CDP as a substrate and ATP as a positive effector. In Table I, the results are compared with the activity of the reactivated native R2 protein. No activity was detected for the Y370F R2 protein. In contrast, the Y370W R2 protein showed a low but still significant activity, which increased in a linear way with increasing amounts of added mutated protein. The reaction was linear during the 30-min incubation (data not shown). The reactivated native R2 protein assayed in parallel had a specific activity of 228 units/mg. The specific activity of the Y370W R2 protein was ϳ3.9 units/mg. A control experiment in which the highest amount of Y370W protein was assayed in the absence of added mouse R1 protein showed only column background values (0.04 -0.05 nmol). We therefore conclude that the Y370W protein has ϳ1.7% of normal R2 protein activity, and this activity is not due to any contamination of E. coli ribonucleotide reductase.
Substrate-dependent Inactivation of the Tyrosyl Radical of the Y370W R2 Protein during Catalysis-One reason for the low specific activity of the Y370W R2 protein compared with the native R2 protein might be a slower rate of radical transfer. This might allow the detection of transient EPR reaction intermediates. We therefore used EPR spectroscopy to study the kinetics of the CDP reduction catalyzed by the native and mutated R2 proteins in the presence of a 2-fold molar excess of the R1 protein, but in the absence of added iron. Under these conditions, where the incubation mixture was incubated at 27°C in an EPR tube, we were not able to detect any visible transient EPR intermediates. Instead, an unexpected time-dependent decay of the tyrosyl radical was observed with the Y370W R2 protein (Fig. 5A, f). Neither the native R2 protein nor the Y370F R2 mutant showed such a decay under similar incubation conditions (Fig. 5A, q and E, respectively). Furthermore, control samples with the native or mutated R2 proteins in the presence of the R1 protein, but without the addition of CDP, showed no significant decrease in radical content (data not shown). Therefore, the Y370W-specific tyrosyl radical decay is absolutely dependent on the presence of the CDP substrate.
The iron/radical center of the mouse R2 protein is labile under physiological conditions. After 30 min at 37°C, the protein loses ϳ50% of its iron and radical content. Continuous loss of ferric iron in the presence of the reductant dithiothreitol in the assay mixture results in a simultaneous decay and reconstitution of the tyrosyl radical (34). This may explain the observed rapid initial decay of the radical, followed by a steadystate plateau at ϳ60% of the initial value (Fig. 5A). To support this hypothesis, we repeated the CDP reduction at 37°C in the presence of the previously characterized iron chelator desferrioxamine (34). When the reactivated Y370W R2 protein was incubated with desferrioxamine in the absence of the R1 protein and CDP, the same spontaneous loss of the tyrosyl radical was seen as observed earlier (Fig. 5B, OE). However, in the presence of the R1 protein and CDP, the decay of the radical was much more pronounced (Fig. 5B, f). This further strengthens our hypothesis that the decay of the tyrosyl radical in the case of the Y370W protein is dependent on ongoing catalysis.
Reconstitution Kinetics of the Native and Mutated R2 Proteins-To investigate if Tyr 370 is involved in the reconstitution of the diferric/tyrosyl free radical center of the R2 protein, we used rapid freeze-quench EPR with 3 nmol of Fe(II)/nmol of R2 polypeptide and a reconstitution temperature of 5°C. With  6.0 0.65 a The activity was measured in the presence of 10 g of recombinant mouse R1 protein (33). The Dowex column background was ϳ0.04 -0.05 nmol and has not been subtracted.
native R2 apoprotein, we could confirm the earlier reported k form for the tyrosyl radical of 0.14 s Ϫ1 (16). The values for the Y370W and Y370F R2 proteins (0.19 and 0.16 s Ϫ1 , respectively) did not differ significantly from the value of the native protein. Furthermore, the final yield of the tyrosyl free radical was the same for all three proteins (0.5-0.6 nmol of radical/nmol of polypeptide chain). DISCUSSION The key to understanding the mechanism of the reaction catalyzed by the enzyme ribonucleotide reductase is a better knowledge of how the radical properties of the R2 protein are transferred to the active site of the R1 protein. After the threedimensional structures of the R1 and R2 proteins were reported (8,10), many studies have been made to address this question.
Biological electron transfer has been extensively studied for redox reactions involved in the respiratory and photosynthetic chains. The two commonly discussed models for electron transfer over distances of up to 20 Å argue whether or not the intervening protein structure is an important factor for the efficiency of the electron transfer (36,37). In contrast to these models, the electron transfer in ribonucleotide reductase has to propagate quite a long distance of ϳ35 Å between the oxidized tyrosine in the R2 protein and a neutral cysteine at the active site of the R1 protein. It is unlikely that the current electron transfer theories can be applied to this system since too much energy would be required to overcome the charge separation arising from a pure electron transfer. Instead, a more recent theoretical model explains this transfer as a coupled electron/ proton transfer, i.e. a hydrogen radical transfer (H ⅐ ) (18). The advantage of this model is that it minimizes charge separation as an uncharged hydrogen radical is moved along a conserved hydrogen-bonded chain, i.e. an RTP. This model also accounts for the reaction that initially takes place during ribonucleotide reduction, i.e. the abstraction of an H ⅐ from the 3Ј-position of the substrate.
We have previously tried to obtain experimental evidence for the proposed RTP of the mouse R2 protein during the catalytic reaction as well as during the reconstitution of the iron/tyrosyl radical center (12,16). Our results strongly support the RTP model and stress the importance of a proper hydrogen-bonded chain for the formation of such a pathway, both during catalysis and the generation of the iron/radical center. Similar results were obtained for the E. coli R1 and R2 proteins (14,15).
The connection between the R1 protein RTP with the R2 protein RTP is unknown since no co-crystallization of these proteins has been obtained. In the E. coli R2 protein, the conserved tyrosine 356 has been suggested as the connecting link since substitution of this residue with alanine and phenylalanine inhibits enzyme activity (19,38). Tyr 356 is part of the E. coli R2 protein C-terminal tail, which is required for binding to the R1 protein (26). In the mouse R2 protein, this residue corresponds to tyrosine 370, which is not involved in the 7-amino acid-long C-terminal region required for R1-R2 complex formation (22). By substituting the mouse R2 protein Tyr 370 with either of 2 aromatic residues, tryptophan or phenylalanine, with completely different hydrogen bond-forming properties, we aimed to probe the importance of an intact hydrogenbonded chain for the RTP.
Using the biosensor technique, we could not detect any major difference between the native and mutated R2 proteins regarding their ability to bind to the R1 protein. This was expected since Tyr 370 does not belong to the C-terminal amino acids involved in R1 binding. Furthermore, the results argue against any major collapse of the R2 protein structure. The ability to form a normal iron/tyrosyl free radical center in the mutated R2 proteins also indicates that the overall protein conformation was retained.
However, the Y370F R2 protein was completely inactive, whereas the Y370W R2 protein had a low but significant activity of ϳ1.7% compared with the native R2 protein. Since the R1-R2 complex formation and the iron/tyrosyl radical center are unaffected in these mutated R2 proteins, the results strongly indicate the conserved Tyr 370 to be involved in the RTP. According to the model of Siegbahn et al. (18), the radical transfer is dependent on a proper hydrogen-bonded chain. Our results are in good agreement with this prediction since interruption of the RTP totally inhibits the catalytic reaction (Y370F), whereas an alternation allows partial catalysis (Y370W). In contrast to the catalytic reaction, Tyr 370 appears not to be involved in the reconstitution of the iron/tyrosyl free radical center in the isolated R2 protein. This is different from the conserved Trp 103 and Asp 266 (16) and may be a consequence of its position in the flexible C-terminal tail of the protein.
The importance of proper hydrogen bonds as linkages in the FIG. 5. Time-dependent decay of the Y370W tyrosyl radical during CDP reduction. A, the content of the tyrosyl radical in the R2 proteins incubated at 27°C with the R1 protein and CDP under assay conditions, but in the absence of added iron, was quantified by EPR spectroscopy. f, Y370W R2 protein; q, native R2 protein; E, Y370F R2 protein. B, time-dependent decay of the Y370W tyrosyl radical during CDP reduction in the presence of 100 M desferrioxamine at 37°C. f, Y370W R2 protein incubated with the R1 protein and CDP; OE, Y370W R2 protein incubated in the absence of the R1 protein and CDP. RTP was also addressed in studies in which the conserved aspartic acid 237 in the E. coli R2 protein, proposed to be involved in the radical transfer, was substituted with asparagine or glutamate (15). The D237N R2 protein was catalytically inactive, whereas the D237E R2 protein retained ϳ7% of the activity of the native R2 protein. The three-dimensional structure of the D237E protein showed that this residue could indeed form the same hydrogen bonds to Trp 48 and His 118 as in the native R2 protein. However, the reduced activity might be explained by the slightly longer distance of the bond to Trp 48 .
One unexpected observation in our study is the time-dependent decay of the tyrosyl free radical when the Y370W R2 protein was incubated with an excess of the R1 protein and CDP under assay conditions, but in the absence of added iron. EPR measurements showed a significant time-dependent loss of the tyrosyl radical. As this loss was detected neither in the native R2 protein nor in the Y370F R2 protein under the same conditions, we propose that this reaction is specific for the Y370W R2 protein. The recorded low formation of dCDP with time excludes a rapid initial reaction resulting in a fast inactivation of the R2 protein. Instead, our results show that the entire catalysis is slowed down in the Y370W mutant, and this slow rate of radical transfer may explain the loss of radical probably as a consequence of structural changes.
Together with our earlier results (12,16), our observations strongly support a radical transfer mechanism along the previously suggested array of conserved residues, where an intact hydrogen-bonded chain is a prerequisite for an efficient transfer. Even if the C-terminal 37 amino acids of the mouse R2 protein are too flexible to be mapped in the crystal structure, they contain at least one catalytically essential residue, Tyr 370 . Upon R1-R2 complex formation, this residue may be placed in a well defined structural position, bridging the RTP of the R1 protein with the RTP of the R2 protein.