The crystal structure of class II ribonucleotide reductase reveals how an allosterically regulated monomer mimics a dimer

In contrast to most enzymes central to metabolism, ribonucleotide reductases do not seem to be evolutionarily conserved with respect to primary sequence, quaternary structure, substrate preference or metallocofactor usage. The diversity of their metallocofactors has provided the basis for the subdivision of RNRs into three classes. Class I RNRs, found in eukaryotes, bacteria, bacteriophage and viruses, use a diiron-tyrosyl radical (Y). Class II RNRs, found in bacteria, bacteriophage, algae and archaea, use coenzyme B₁₂ (adenosylcobalamin, AdoCbl). Class III RNRs, found in anaerobic bacteria and bacteriophage, use an FeS cluster and S-adenosylmethionine to generate a glycyl radical (G). Many organisms have more than one class of RNR present in their genomes⁸. Despite this apparent diversity, extensive chemical and biochemical studies have provided a common mechanistic model for all three classes of RNR⁹. In this model, the function of each metallocofactor is to generate an active site thiyl radical (S). This thiyl radical then initiates the nucleotide reduction process by hydrogen atom abstraction from the ribonucleotide (Fig. 1). In class I and III RNRs, the active site and cofactor binding sites are located on separate polypeptide chains ( and , respectively). In contrast, class II RNRs are less complex, using the small molecule B₁₂ in place of the ₂ subunits of class I RNR or the ₂ activating enzyme of class III RNR¹⁰.

Figure 1. General reaction catalyzed by ribonucleotide reductases. Each of the three well-characterized RNR classes uses a different metallocofactor to generate the thiyl radical (S). For class II RNRs, the S (Cys 408 in L. leichmannii) is generated by AdoCbl carbon-cobalt (C−Co) bond homolysis. Hydrogen atom (red) abstraction from the substrate by the thiyl radical and the subsequent multiple step radical rearrangements result in the loss of the 2' hydroxyl group in the form of water. In class I and II RNRs, reducing equivalents for the reaction are provided by the oxidation of two Cys residues to a disulfide (Cys 119−Cys 419 in L. leichmannii)35. In contrast, class III RNR obtains reducing equivalents by the oxidation of formate36. Full Figure and legend (15K)

We describe here the crystal structure of class II ribonucleoside triphosphate reductase (RTPR, E.C. 1.17.4.2) from Lactobacillus leichmannii in the apo enzyme form and in complex with the B₁₂ analog adeninylpentylcobalamin at 1.75 and 2.0 Å resolution, respectively. Although crystal structures exist for the ₂ (R1) and ₂ (R2) subunits of the Escherichia coli class I RNR^{11, 12} and for the ₂ subunit of class III RNR from anaerobic bacteriophage T4 (ref. 13), there is currently no structural information for the functionally intact ₂₂ heterodimeric RNRs. The class II RNR structure we report here completes the trilogy of structures defining the catalytic subunits of the three well-characterized classes of RNR and provides the first opportunity to observe the juxtaposition of the active site, cofactor and allosteric-specificity sites in a ribonucleotide reductase. The first monomeric RNR structure also allows us to examine how allosteric regulation of specificity occurs in the absence of a dimer interface.

The structure of RTPR from L. leichmannii in the apo enzyme form was solved using multiwavelength anomalous dispersion (MAD) techniques after incorporation of selenomethionine (SeMet) into the protein (Table 1). RTPR has a global fold composed of a core 10-stranded /-barrel (Fig. 2a,b), similar to that seen in the RNR class I ₂ subunit (R1) and the class III ₂ subunit structures. This barrel contains two parallel five-stranded -sheets oriented in an antiparallel fashion and, therefore, differs significantly from the more common (/)₈ TIM barrel in which all eight -strands are parallel. The center of the RTPR barrel contains the prototypical RNR 'finger loop' consisting of two antiparallel strands, with the active site S being generated on Cys 408 at the 'fingertip'. A superposition of the core barrel residues from all three classes of RNR (Fig. 2c) shows the striking similarity in global fold despite the low sequence identity between the classes (<10% based on structural alignment). The superimposed root mean square (r.m.s.) deviations for 70 -strand C atoms taken from the (/)₁₀ core are as follows: class I versus II = 1.0 Å, class I versus III = 1.7 Å and class II versus III = 1.8 Å. Note that these values differ from those reported earlier¹³, where the identity of the atoms used for superposition was not provided. This superposition indicates that the class I and II RNR structures are more similar to one another than either is to the class III structure.

Figure 2. Global fold conservation: the RNR structural family. a, Stereo view of RTPR from L. leichmannii. Ribbon representation with -helices colored red and -strands colored yellow. The structure has overall dimensions of 55 60 75 Å3. b, Topology diagram for class II RTPR. The central -strands from the core (/)10-barrel are labeled A−J. The 130-amino acid insert comprising the effector-binding region (green) lies between strands B and C. The B12-binding region (red) consists of the insert between strands H and I, in combination with the C-terminal portion following strand J. c, Comparison of the core 10-stranded /-barrels for class I (left), class II (center) and class III (right) RNRs. Only -helices and -strands that are structurally and topologically similar have been included. At the center of each barrel lies the 'finger loop' that contains the active site Cys residue (yellow sphere) at the 'fingertip.' Figure prepared using Ribbons37. Full Figure and legend (125K)

Table 1. Data collection and refinement statistics Full Table

In addition to global fold similarities, key active site residues in class I and II RNR are also structurally conserved (Fig. 3). These residues include the S Cys 408 (L. leichmannii numbering), the general acid/base catalyst Glu 410 and its hydrogen-bonding partner Asn 406, and the redox-active disulfide Cys 119-Cys 419 (corresponding to residues Cys 439, Glu 441, Asn 437 and Cys 225-Cys 462 in E. coli R1, respectively). Active site residues that are structurally conserved between class II and III RNRs include Cys 290, the putative precursor to the S at the barrel 'fingertip' (Fig. 2c), as well as Cys 79, the putative redox-active equivalent of Cys 119 from L. leichmannii class II RNR and Cys 225 from E. coli class I RNR. There are also common features in the putative substrate phosphate-binding regions of the class I and II active sites, most notably the location of two short -helices (₁₂ and ₂₄) that bind the -phosphate of GDP in the E. coli class I structure (Fig. 3). The high level of structural similarity in the substrate phosphate-binding region of E. coli class I RNR and L. leichmannii class II RNR is puzzling because these enzymes have different substrate preferences; class I RNR from E. coli binds nucleoside diphosphates (NDPs), whereas class II RNR from L. leichmannii binds nucleoside triphosphates (NTPs). In the case of the T4 bacteriophage class III RNR, an enzyme which also uses NTPs, substrate preference has been explained by the presence of two His residues near the putative -phosphate-binding site¹³. These His residues are not conserved in the L. Leichmannii class II RNR structure. In fact, there are no obvious amino acid substitutions in the vicinity of ₁₂ or ₂₄ that explain the preference of the class II enzyme for tri- instead of diphosphate substrates. The only significant difference between class I and II enzymes in the substrate phosphate binding region is that B₁₂ binds next to helix ₂₄ in the class II structure (Fig. 3). Whether B₁₂ plays a role in controlling the preference for di-versus triphosphate substrates remains to be investigated. An important caveat in the analysis of substrate binding is that the only structure with substrate bound is of the class I enzyme¹⁴. Even in that case, the substrate was not bound at full occupancy. From an evolutionary perspective, it is interesting to note that the structures of L. leichmannii class II and E. coli class I RNRs resemble each other very closely, even in the substrate phosphate-binding region where differences were expected.

Figure 3. Active site conservation between class I and II RNRs. Stereo view with C atom traces displayed for class I (thin lines, PDB code 1RLR) and class II (thick lines), along with side chain atoms for key active site residues that are spatially conserved (see text). The superposition was carried out as described (see text). The substrate position has been deduced from a superposition of class I RNR apo enzyme in the oxidized form (PDB code 1RLR) with class I RNR in the reduced state with GDP bound (PDB code 4R1R)14. Two short -helices (12 and 24) whose N-termini (and helix dipoles) are positioned to stabilize phosphate anions are labeled. Atom colors are as follows: carbon in green, nitrogen in blue, oxygen in red, sulfur in yellow and phosphate in magenta. Figure prepared using Ribbons37. Full Figure and legend (40K)

Figure 4. B12 bound to RNR. a, Chemical structures of adenosylcobalamin (AdoCbl, left panel) and adeninylpentylcobalamin (AdPentCbl, right panel). b, Difference (Fo - Fc) electron density at 2.0 Å resolution (2 contour) for AdPentCbl bound to L. leichmannii RTPR, calculated before the inclusion of any AdPentCbl atoms in the refinement. The orientation of AdPentCbl in (b) is the same as in (a). Figure prepared using Ribbons37. Full Figure and legend (62K)

Superposition of the AdPentCbl cocrystal structure with the apo enzyme structure reveals the concerted movement of 100 amino acids (Fig. 5a). The majority of the structural changes are concentrated in three -strands (residues 570−578, 592−596 and 602−610), whereas two -helices in the C-terminal portion of this domain (residues 685−724) seem to make small adjustments in response to the -strand movement. The largest movement observed is >3.5 Å for corresponding C atoms at the edge of this region. On the basis of this conformational change, we define the B₁₂-binding region as residues 565−626 and 685−724 (Figs 2b, 5b). Even though the majority of the interactions between B₁₂ and RTPR involve residues in this region, there are also contacts between Arg 33 and bound cofactor (data not shown). The B₁₂-binding region of RTPR shares no structural similarity with other known B₁₂-binding domains, including that of diol dehydratase, an enzyme that uses similar chemistry and has the same DMB configuration as RTPR¹⁶. Thus, RTPR uses a newly discovered fold for B₁₂ binding and is unique compared to other B₁₂-dependent enzyme families. Unexpectedly, two key -strands of this binding region (residues 570−578 and 602−610) have structural counterparts (Fig. 6a) in class I E. coli RNR (residues 641−647 and 650−655, respectively). The B₁₂-binding region of RTPR is surprisingly more similar to related structural elements in class I RNR than it is to the cofactor-binding elements of other B₁₂-dependent enzymes. This observation, indicating that only subtle changes occurred in this region during the course of evolution, differs from an earlier model of the class II structure²⁰, which predicted a large cleft for B₁₂ binding.

Figure 5. RTPR conformational changes, domain architecture and thiyl radical formation. a, B12-induced conformational changes, based on a superposition of the apo enzyme structure (blue) with the complex containing AdPentCbl bound (red). All C atoms were used for superposition to provide an unbiased representation of potential conformational changes. The B12-binding region closes down (orange arrow) in the direction of the active site Cys residue upon binding of B12. A putative sulfate ion (black circle and arrow) is observed at the potential hinge region centered near His 565. The active site Cys 408 is shown as a yellow sphere. The B12 is shown in green stick representation, and the putative sulfate is shown in ball-and-stick representation with atoms colored as follows: oxygen in red and sulfur in yellow. b, Ribbon drawing showing the domain architecture of L. leichmannii RTPR. The B12-binding region is colored red; the catalytic core (/)10-barrel, blue; and the insert comprising the effector-binding region, green. AdPentCbl atoms are depicted as an orange ball-and-stick, with the cobalt atom in red. The active site Cys 408 is shown as a yellow sphere. c, Conservation of the radical pathway in class I, II and III RNRs, based on a superposition of the catalytic Cys residues. Comparison of the B12 position from class II (ball and stick) with the Gly 580 position from class III (cyan sphere, based on the C position of Ala 580 in PDB entry 1B8B) and Tyr 730−Tyr 731 from class I (ball and stick). The corresponding active site Cys residues and disulfides are also shown. Ribbon drawing in light orange is for class II RNR. This superposition was based solely on backbone atoms N, C, C and side chain atom C for residues corresponding to Cys 408 and Cys 119 from RTPR — that is, Cys 439 and 225 from 4R1R and Cys 290 and 79 from 1B8B. Atom colors are as follows: carbon in green, nitrogen in blue, oxygen in red, sulfur in yellow, phosphorous in purple and cobalt as magenta sphere. Figure prepared using Ribbons37. Full Figure and legend (93K)

Figure 6. Allosteric regulation of specificity in dimeric versus monomeric RNRs. a, Ribbon drawing of the class I RNR 2 domains (PDB entry 4R1R), the class II RNR domain, and the class III RNR 2 domains (PDB entry 1B8B). Specificity, activity, and substrate binding sites are labeled. For dimeric class I and class III RNRs, molecule 1 is shown in red, and molecule 2 is shown in blue. For the single polypeptide chain of class II RNR, an insert in the barrel (residues 168-298) is shown in blue. Two key -strands from class II RNR (residues 570−578, 602−610) involved in B12-induced conformational changes are colored yellow. The analogous -strands in class I RNR are also shown in yellow. b, Topology of the effector binding regions in class I (dimer), class II (monomer), and class III (dimer) structures. Loops in the effector binding region of class I and class III RNR are labeled A through D (A' through D' in molecule 2). Loops in the putative effector-binding region of class II RNR are labeled to reflect structural and sequence similarity with class I RNR. The site of effector binding in the class I RNR structure is shown as green and cyan asterisks and the proposed effector binding site in class II RNR is marked with a green asterisk. Colors are the same as in (a), except that differences in connectivity for the class II RNR are colored magenta. c, Close-up view of the relationship between the effector binding site (black arrow), specificity site effector loops A−D (green), the active site Cys (yellow sphere), and B12 (gold ball-and-stick, with central cobalt atom colored pink) in L. leichmannii class II RNR. Substrate (ball-and-stick, colored by atom type) has been modeled into the active site based on alignment with class I RNR pdb entry 4R1R. Note that in an earlier publication14 loop C was referred to as loop1, loop D was referred to as loop2, and loops A and B were not named. Figure prepared using Ribbons37. Full Figure and legend (281K)

The net result of the B₁₂-induced structural change (Fig. 5a) is a conformation that is slightly more 'closed' relative to the apo enzyme form. Closure is essential to protect the highly reactive nucleotide radical intermediates from oxygen and solvent, and for the proper positioning of B₁₂ relative to Cys 408 for carbon−cobalt (C−Co) bond homolysis and S formation. Our structure probably does not represent the fully closed form of the enzyme because the active site Cys 408 is still solvent exposed and the distance between the cobalt and the Cys 408 sulfur is 10 Å, longer than the 5.5−7.5 Å distance determined from EPR experiments²¹. This long distance is not a surprise because the structure does not have effector bound and effector binding is known to enhance the rate of C−Co bond homolysis¹⁰. Presumably, effector binding enhances C−Co bond cleavage by triggering a conformational change to the closed state of the enzyme. The structure suggests that movement toward this closed state could follow the same trajectory as the conformational change observed between apo and B₁₂-bound forms of RTPR. The apparent hinge for this motion is near a bound anion (Fig. 5a) at the base of two -strands of the B₁₂-binding region (residues 570−578 and 602−610) (Figs 5a,b, 6a). Based on sample preparation and crystallization conditions (see Methods), we have chosen to model this anion as sulfate rather than phosphate. The observation of a sulfate molecule is interesting given that anions have been shown to activate RTPR from L. leichmannii²² and to eliminate the requirement of an allosteric effector in class I and II RNRs^{22, 23, 24}. The E. coli class I RNR structure has an analogous hinge region that includes two -strands (residues 641−647 and 650−655) (Fig. 6a), but no bound anion was reported. We are conducting further studies to determine if this sulfate ion is catalytically relevant or merely an artifact of crystallization in high ammonium sulfate concentrations. Another interesting issue to investigate is whether these -strands in class I RNR undergo a conformational change comparable to the B₁₂-induced change seen in the class II structure, or if their presence in class I RNR is merely a vestige of B₁₂-binding ability that has been lost during the course of evolution.

A mechanistically informative observation, resulting from a comparison of the structures of the three -subunits of the RNRs, relates to the mechanism by which the active site S is generated on each finger loop (Fig. 2c). Specifically, both the class II and III RNRs generate the S directly by hydrogen atom abstraction via the putative 5'-deoxyadenosyl radical or the glycyl radical, respectively. The mechanism by which the class I RNR S is generated is open to debate. Both long-range electron-coupled proton transfer and hydrogen atom transfer have been proposed²⁵. A comparison of the structures reveals that the axial ligand of B₁₂ in class II RNR, the G in class III RNR and the two Tyr residues (730 and 731) implicated in cysteine oxidation in E. coli class I RNR²⁶ are all similarly situated with respect to the site of S formation (Fig. 5c). Despite low overall sequence identity in this portion of the active site, the strict conservation of the spatial arrangement of the residues and/or cofactors involved in radical initiation in all three RNRs highlights S chemistry as a major driving force in the divergent evolution of RNRs.

One notable feature of ribonucleotide reductases is that a single enzyme catalyzes the reduction of all four ribonucleotides². The specificity for each of the four substrates is determined by the binding of an allosteric effector (ATP, dNTP) at a location distant (>15 Å) from the catalytic site. A comparison of the structure of RTPR from L. leichmannii with the cocrystal structure of the class I ₂ subunit from E. coli with the effector molecule dTTP bound¹⁴ can be used to localize the specificity effector site in RTPR to the top of a four-helix bundle (Fig. 6a). The four loops (A, B, C and D) at the top of this helical bundle (Fig. 6b,c) are similar in three-dimensional position and sequence to the corresponding loops of the effector-binding region in the E. coli class I structure. Based on these structural comparisons we define the effector-binding region of class II RTPR from L. leichmannii as two helices from the (/)₁₀-barrel (residues 147−158 and 298−313) in combination with a single insertion into the /-barrel from residues 168−298 (Figs 2b, 5b, 6a,b). Note that class II RNR from L. leichmannii contains only the effector specificity site and lacks the effector activity site found near the N-terminus of E. coli class I RNR (Fig 6a).

Perhaps the most unanticipated result from the structure of class II RTPR is that the dimer interface responsible for effector binding in the class I RNR has been preserved in a monomer. This explains how a monomeric enzyme can control substrate specificity as efficiently as a multimeric enzyme. Specifically, the four-helix bundle comprising the effector-binding region is common to both structures (Fig. 6a,b). In addition, three abutting -strands from the (/)₁₀-barrel of the second subunit are also conserved. Thus, whereas RTPR from L. leichmannii is monomeric, the 130-amino acid insert maintains the 'essence' of the dimer interface present in the class I enzyme. An r.m.s. deviation of 1.69 Å, calculated using corresponding C atoms for the six -strands and four -helices comprising the dimer interfaces of the class I and II RNR structures (Fig. 6b), demonstrates the strong structural similarity between the effector-binding regions of class I and II RNR. However, a comparison of the topological connections in this region reveals that they are not identical (Fig. 6b). Thus, a single genetic 'cut-and-paste' event cannot convert the class II monomer into the class I dimer, or vice versa.

An unresolved issue is the mechanism(s) by which substrate specificity can be controlled at a distant allosteric effector site. In class I RNR, it is not clear if this communication is intermolecular (between different -subunits in a dimer) or intramolecular (within a single -subunit) (Fig. 6a). The distances between the specificity sites and active sites are: 15 Å intermolecular and 25 Å intramolecular in class I RNR; 15 Å in class II RNR; 25 Å in class III RNR. In terms of the communication in class I RNR, the class II RNR structure provides insight and suggests that allosteric regulation of specificity occurs between -subunits in a dimer, because this effector site-active site pair is conserved in RTPR (Fig. 6a). The significant structural homology between E. coli class I RNR and L. leichmannii class II RNR has allowed us to re-examine the roles of the four loops involved in the allosteric regulation of the class I and II enzymes. Residues from loops A and C (Fig. 6b,c) are within hydrogen bonding distance of the effector in the E. coli class I RNR structure¹⁴. Only two residues, Asp 232 and Arg 262 from loops A and C, respectively, are believed to make side chain−effector hydrogen bonds in E. coli class I RNR; these are conserved in L. leichmannii class II RNR (Asp 223 and Arg 258), emphasizing the importance of these loops in effector binding. In addition to binding effector, loop C may also be involved in communicating between the effector site and the active site via loops B and D (Fig. 6c). The tip of putative effector-binding loop C in RTPR is Gly-rich (Gly 266-Phe-Gly-Gly 270) and fairly mobile, as indicated by elevated B-factors (20 Ų above average), and could communicate the presence of effector through conformational changes. In both structures, loops B and D lie between the effector site and active site, suggesting that conformational changes in these loops can pass information from loop C and/or the effector to substrate^{14, 27}. On the basis of this comparison, we believe that the mechanism for information transfer from the specificity site to the active site will be the same for class I and II RNRs. In contrast, class III RNR differs from class I and class II RNR in terms of the topology of the effector binding region (Fig. 6a,b), the distance between the specificity and active sites (Fig. 6a), and the residues involved in effector binding²⁷. Thus, it is unlikely that the mechanism by which the signal is transferred between specificity and active sites will be identical in all three classes.

Using crystal structures for the catalytic subunits of all three well-characterized RNRs, it is possible to look for common patterns, as well as differences, among the classes. In the structure of RTPR, we find conservation of the (/)₁₀-barrel fold and key active site residues. In addition, there is intriguing structural conservation in terms of S generation. The residues involved in its formation in the class I and III enzymes are in the same location as the B₁₂ in class II RNR. These similarities reinforce the idea that chemistry has been a major driving force in the divergent evolution of RNRs. In terms of allosteric regulation, one of the most surprising results is the preservation, in the class II monomer, of the class I dimer interface responsible for effector binding. Thus, despite a different framework, the key structural features of the allosteric effector specificity region are conserved between class I and II RNRs. Therefore, the class II RTPR seems to provide a relatively simple paradigm for understanding the regulation of specificity in class I RNRs, including human RNR. However, the relationship of the class III allosteric regulation to the class I and II RNRs is still an open question.

In many regards, the class II RNR from L. leichmannii is far less complex than the class I and III enzymes. Instead of requiring the protein for radical generation, class II RNR uses a small molecule cofactor. The binding site for this cofactor represents only a minor modification from the architecture of the class I enzyme and does not involve additional binding motifs or domains. Instead of requiring two -subunits to generate a specificity allosteric site, L. leichmannii class II RNR creates this site in a monomeric framework. Proposing that such a simple member of the RNR family may have evolved early, and that the added complexity of oligomeric RNRs came later, is tempting. However, the close relationship between the structures of class I and II RNRs contradicts this idea, because class I RNRs have probably evolved late in the evolutionary time line. Regardless of its place on the evolutionary path, the simplicity of class II RTPR provides us with the opportunity to identify all the basic features required for the reduction of ribonucleotides, a reaction that is essential for DNA synthesis, and one that links RNA and DNA worlds.

RTPR from L. leichmannii (738 amino acids and 82 kDa molecular weight) was cloned into E. coli, overexpressed and purified as described²⁸. Crystals of apo RTPR were grown using standard hanging drop vapor diffusion at room temperature. The hanging drops consisted of 2 l of protein solution (RTPR at 20 mg ml^-1 in 50 mM sodium citrate, pH 5.6, 20% (v/v) glycerol, 1 mM dithiothreitol (DTT) and 1 mM EDTA) mixed with 2 l of reservoir solution (1.6 M ammonium sulfate, 2% (w/v) PEG 8000, 20% (v/v) glycerol and 100 mM sodium citrate, pH 5.6). Because we observed superior crystallization behavior for the double mutant C731S/C736S, this protein was used for all structural studies. For data collection, crystals were flash-cooled directly in liquid nitrogen or a nitrogen gas stream. Cocrystals of RTPR in complex with AdPentCbl were grown under identical conditions, except for first incubating the protein with 400 M AdPentCbl for 60 min before crystallization. AdPentCbl was chosen for cocrystallization because it is a competitive inhibitor of RTPR with respect to AdoCbl (K_is = 0.2 M)²⁹. All manipulations in the presence of AdPentCbl, including data collection, were done under dim red light. Crystals belong to space group P2₁, with four molecules per asymmetric unit (apo enzyme: a = 120.5, b = 114.7, c = 121.3 Å and = 110.2°, and AdPentCbl complex: a = 120.0, b = 113.5, c = 118.7 Å and =110.7°). A three-wavelength inverse beam MAD dataset was collected at ALS beamline 5.0.2. Native datasets were collected at NSLS beamline X25. All data were processed and scaled using DENZO and SCALEPACK³⁰.

	Top

1. Zhou, B.B. & Elledge, S.J. The DNA damage response: putting checkpoints in perspective. Nature 408, 433−439 (2000). | Article | PubMed | ISI | ChemPort |
2. Jordan, A. & Reichard, P. Ribonucleotide reductases. Annu. Rev. Biochem. 67, 71−98 (1998). | Article | PubMed | ISI | ChemPort |
3. Kunz, B.A. et al. International Commission for Protection Against Environmental Mutagens and Carcinogens. Deoxyribonucleoside triphosphate levels: a critical factor in the maintenance of genetic stability. Mutat. Res. 318, 1−64 (1994). | Article | PubMed | ISI | ChemPort |
4. Plunkett, W., Huang, P., Searcy, C.E. & Gandhi, V. Gemcitabine: preclinical pharmacology and mechanisms of action. Semin. Oncol. 23, 3−15 (1996). | PubMed | ISI | ChemPort |
5. Duan, J. et al. Antiviral activity of a selective ribonucleotide reductase inhibitor against acyclovir-resistant herpes simplex virus type 1 in vivo. Antimicrob. Agents Chemother. 42, 1629−1635 (1998). | PubMed | ISI | ChemPort |
6. Stubbe, J. Ribonucleotide reductases: the link between an RNA and a DNA world? Curr. Opin. Struct. Biol. 10, 731−736 (2000). | Article | PubMed | ISI | ChemPort |
7. Reichard, P. From RNA to DNA, why so many ribonucleotide reductases? Science 260, 1773−1777 (1993). | PubMed | ISI | ChemPort |
8. Jordan, A. et al. Ribonucleotide reduction in Pseudomonas species: simultaneous presence of active enzymes from different classes. J. Bacteriol. 181, 3974−3980 (1999). | PubMed | ISI | ChemPort |
9. Stubbe, J. & van der Donk, W.A. Protein radicals in enzyme catalysis. Chem. Rev. 98, 705−762 (1998). | Article | PubMed | ISI | ChemPort |
10. Tamao, Y. & Blakley, R.L. Direct spectrophotometric observation of an intermediate formed from deoxyadenosylcobalamin in ribonucleotide reduction. Biochemistry 12, 24−34 (1973). | Article | PubMed | ISI | ChemPort |
11. Nordlund, P., Sjöberg, B.M. & Eklund, H. Three-dimensional structure of the free radical protein of ribonucleotide reductase. Nature 345, 593−598 (1990). | Article | PubMed | ISI | ChemPort |
12. Uhlin, U. & Eklund, H. Structure of ribonucleotide reductase protein R1. Nature 370, 533−539 (1994). | Article | PubMed | ISI | ChemPort |
13. Logan, D.T., Andersson, J., Sjöberg, B.M. & Nordlund, P. A glycyl radical site in the crystal structure of a class III ribonucleotide reductase. Science 283, 1499−1504 (1999). | Article | PubMed | ISI | ChemPort |
14. Eriksson, M. et al. Binding of allosteric effectors to ribonucleotide reductase protein R1: reduction of active-site cysteines promotes substrate binding. Structure 5, 1077−1092 (1997). | Article | PubMed | ISI | ChemPort |
15. Lawrence, C.C. et al. Binding of Cob(II)alamin to the adenosylcobalamin-dependent ribonucleotide reductase from Lactobacillus leichmannii. Identification of dimethylbenzimidazole as the axial ligand. J. Biol. Chem. 274, 7039−7042. (1999). | Article | PubMed | ISI | ChemPort |
16. Shibata, N. et al. A new mode of B₁₂ binding and the direct participation of a potassium ion in enzyme catalysis: X-ray structure of diol dehydratase. Structure Fold. Des. 7, 997−1008 (1999). | Article | PubMed | ISI | ChemPort |
17. Drennan, C.L., Huang, S., Drummond, J.T., Matthews, R.G. & Ludwig, M.L. How a protein binds B₁₂: a 3.0 Å X-ray structure of B₁₂-binding domains of methionine synthase. Science 266, 1669−1674 (1994). | PubMed | ISI | ChemPort |
18. Mancia, F. et al. How coenzyme B₁₂ radicals are generated: the crystal structure of methylmalonyl-coenzyme A mutase at 2 Å resolution. Structure 4, 339−350 (1996). | Article | PubMed | ISI | ChemPort |
19. Reitzer, R. et al. Glutamate mutase from Clostridium cochlearium: the structure of a coenzyme B₁₂-dependent enzyme provides new mechanistic insights. Structure Fold. Des. 7, 891−902 (1999). | Article | PubMed | ISI | ChemPort |
20. Mulliez, E. & Fontecave, M. Ribonucleotide reductases: metal and free radical interplay. Coordin. Chem. Rev. 186, 775−793 (1999). | Article | ISI |
21. Gerfen, G.J., Licht, S., Willems, J.P., Hoffman, B.M. & Stubbe, J. Electron paramagnetic resonance investigations of a kinetically competent intermediate formed in ribonucleotide reduction: evidence for a thiyl radical-Cob(II)alamin interaction. J. Am. Chem. Soc. 118, 8192−8197 (1996). | Article | ISI | ChemPort |
22. Jacobsen, D.W. & Huennekens, F.M. Ion-dependent activation and inhibition of ribonucleotide reductase from Lactobacillus leichmannii. Biochem. Biophys. Res. Commun. 37, 793−800 (1969). | Article | PubMed | ISI | ChemPort |
23. Brown, N.C., Canellakis, Z.N., Lundin, B., Reichard, P. & Thelander, L. Ribonucleoside diphosphate reductase. Purification of the two subunits, proteins B1 and B2. Eur. J. Biochem. 9, 561−573 (1969). | Article | PubMed | ISI | ChemPort |
24. Eliasson, R., Pontis, E., Jordan, A. & Reichard, P. Allosteric control of three B₁₂-dependent (class II) ribonucleotide reductases. Implications for the evolution of ribonucleotide reduction. J. Biol. Chem. 274, 7182−7189. (1999). | Article | PubMed | ISI | ChemPort |
25. Ekberg, M. et al. Preserved catalytic activity in an engineered ribonucleotide reductase R2 protein with a nonphysiological radical transfer pathway. The importance of hydrogen bond connections between the participating residues. J. Biol. Chem. 273, 21003−21008 (1998). | Article | PubMed | ISI | ChemPort |
26. Ekberg, M., Sahlin, M., Eriksson, M. & Sjöberg, B.M. Two conserved tyrosine residues in protein R1 participate in an intermolecular electron transfer in ribonucleotide reductase. J. Biol. Chem. 271, 20655−20659 (1996). | Article | PubMed | ISI | ChemPort |
27. Larsson, K., Andersson, J., Sjöberg, B., Nordlund, P. & Logan, D.T. Structural basis for allosteric substrate specificity regulation in anaerobic ribonucleotide reductases. Structure 9, 739−750 (2001). | Article | PubMed | ISI | ChemPort |
28. Booker, S. & Stubbe, J. Cloning, sequencing, and expression of the adenosylcobalamin-dependent ribonucleotide reductase from Lactobacillus leichmannii. Proc. Natl. Acad. Sci. USA 90, 8352−8356 (1993). | PubMed | ChemPort |
29. Sando, G.N., Grant, M.E. & Hogenkamp, H.P. The interaction of adeninylalkylcobalamins with ribonucleotide reductase. Biochim. Biophys. Acta 428, 228−232 (1976). | PubMed | ISI | ChemPort |
30. Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307−326 (1997). | Article | PubMed | ISI | ChemPort |
31. Doublié, S. Preparation of selenomethionyl proteins for phase determination. Methods Enzymol. 276, 523−530 (1997). | Article | PubMed | ISI | ChemPort |
32. Brünger, A.T. et al. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D 54, 905−921 (1998). | Article | PubMed | ISI |
33. Bailey, S. The CCP4 Suite — programs for protein crystallography. Acta Crystallogr. D 50, 760−763 (1994). | Article | PubMed | ISI |
34. Cowtan, K.D. & Zhang, K.Y. Density modification for macromolecular phase improvement. Prog. Biophys. Mol. Biol. 72, 245−270 (1999). | Article | PubMed | ISI | ChemPort |
35. Booker, S., Licht, S., Broderick, J. & Stubbe, J. Coenzyme B₁₂-dependent ribonucleotide reductase: evidence for the participation of five cysteine residues in ribonucleotide reduction. Biochemistry 33, 12676−12685 (1994). | Article | PubMed | ISI | ChemPort |
36. Mulliez, E., Ollagnier, S., Fontecave, M., Eliasson, R. & Reichard, P. Formate is the hydrogen donor for the anaerobic ribonucleotide reductase from Escherichia coli. Proc. Natl. Acad. Sci. USA 92, 8759−8762 (1995). | PubMed | ChemPort |
37. Carson, M. Ribbons. Methods Enzymol. 277, 493−505 (1997). | Article | ISI | ChemPort |

	Top

Competing interests statement:  The authors declare that they have no competing financial interests.