EDUARD TORRENTS, MARIANN WESTMAN, MARGARETA SAHLIN AND BRITT-MARIE SJÖBERG1
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
JBC Papers in Press. Published on July 7, 2006 as Manuscript M601794200 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M601794200 RIBONUCLEOTIDE REDUCTASE MODULARITY: ATYPICAL DUPLICATION OF THE ATP-CONE DOMAIN IN PSEUDOMONAS AERUGINOSA Eduard Torrents, MariAnn Westman, Margareta Sahlin and Britt-Marie Sjöberg1 Department of Molecular Biology and Functional Genomics, Arrhenius Laboratories for Natural Sciences, Stockholm University, SE-10691 Stockholm, Sweden. Running title: Pseudomonas aeruginosa class Ia ribonucleotide reductase. Keywords: protein evolution, allosteric regulation, protein modularity, ATP-cone, tyrosyl radical, diiron-oxo center 1 Corresponding author. Mailing address: Department of Molecular Biology and Functional Genomics, Stockholm University, Svante Arrhenius väg 16-18 F3, SE-106 91 Stockholm, Sweden. E-mail: britt-marie.sjoberg@molbio.su.se The opportunistic pathogen that protein modularity built on the common Pseudomonas aeruginosa causing serious catalytic core of all RNRs plays an important nosocomial infections is a γ-proteobacterium role in class diversification within the RNR Downloaded from http://www.jbc.org/ by guest on March 3, 2015 that can live in many different environments. family. Interestingly, P. aeruginosa encodes three ribonucleotide reductases (RNRs) that all Ribonucleotide reductase (RNR) is an differ from other well-known RNRs. The essential constituent of all living cells and RNR enzymes are central for de novo organisms. It catalyzes the production of synthesis of deoxyribonucleotides and deoxyribonucleotides for DNA synthesis from essential to all living cells. The RNR of this their corresponding ribonucleotides and thereby study (class Ia) is a complex of the NrdA controls a key step for cell growth and protein harboring the active site and the proliferation. In recent years it has also been allosteric sites, and the NrdB protein recognized that RNR played a key role in the harboring a tyrosyl radical necessary to transition from an RNA-world to a DNA-world. initiate catalysis. P. aeruginosa NrdA Surprisingly, there are at least three major contains an atypical duplication of the N- classes of RNRs in contemporary organisms, terminal ATP-cone, an allosteric domain differing primarily in cofactor requirements and that can bind either ATP or dATP and quaternary structures (1-4). Yet, a common regulates the overall enzyme activity. Here theme among all three classes is a radical-based we characterized the wild type NrdA and reaction mechanism. Recent structural studies two truncated NrdA variants with precise N- have shown that the catalytic core regions of terminal deletions. The N-terminal ATP- the three RNR classes are essentially identical cone (ATP-c1) is allosterically functional, (5-8), strongly suggesting that all RNR classes whereas the internal ATP-cone lacks have a common evolutionary origin. allosteric activity. The P. aeruginosa NrdB is One signature of all RNRs is an also atypical with an unusually short-lived elaborate allosteric regulation controlling the tyrosyl radical, that is efficiently regenerated enzymes’ substrate specificities. The net result in presence of oxygen as the iron ions remain of the allosteric regulation is a balanced tightly bound to the protein. The P. production of the four different aeruginosa wild type NrdA and NrdB deoxyribonucleotide products. Carefully proteins form an extraordinary tight equilibrated concentrations of the four dNTPs complex with a suggested α 4β 4 composition. are required by all cells to avoid errors during An α 2 β 2 composition is suggested for the DNA replication and repair (9). Binding sites complex of truncated NrdA (lacking ATP-c1) for the allosteric effectors which are ATP and and wild type NrdB. Duplication or several dNTPs have been resolved in a number triplication of the ATP-cone is found in some of RNR crystal structures (10-12) and are other bacterial class Ia RNRs. We suggest comparatively close to the active site regions. Specific side chain and backbone movements 1 Copyright 2006 by The American Society for Biochemistry and Molecular Biology, Inc.
may explain the general molecular basis for the of the allosteric activity site is present in P. allosteric regulation of substrate specificity. aeruginosa NrdA (Fig. 1a) and to study its Another allosteric fine-tuning of RNRs possible role in the allosteric and biochemical concerns the overall activity of ribonucleotide properties of the pure enzyme system. reduction and presence of this type of control differs between the different RNR classes and Material and Methods also within the classes. The overall activity control was first localized to the N-terminal Bacterial strains, plasmids and materials - E. region of the E. coli R1 (or NrdA) component coli DH5α (Clontech) was used for cloning and and forms a separate globular domain propagation of plasmids. E. coli Rosetta (DE3) consisting of four α−helices and three (Novagen) was used for protein over- β−strands (10). This domain binds ATP as a expression. Wild type P. aeruginosa PAO1 positive allosteric effector and dATP as a strain (ATCC15692/CECT4122) obtained from negative effector. Recently, similar domains the Spanish National Collection of Type were found in several different proteins and the Cultures was used for genomic DNA motif was termed ATP-cone (13). Most of the purification. All strains were grown at 37ºC in conserved residues of the ATP-cone are LB medium. Plasmid vectors used were pGEM- involved in stabilizing the secondary structure T Easy (Promega) for cloning of PCR amplified elements, which indicates that all sequences fragments and pET22b (Novagen) for protein adopt the same structure and presumably bind over-expression. Oligonucleotide primers were Downloaded from http://www.jbc.org/ by guest on March 3, 2015 ATP. from Invitrogen (England). Restriction The presence or absence of an ATP- endonucleases and other enzymes were from cone in the N-terminal region of RNR is one Fermentas (Germany). basis for dividing the class I RNRs into subclasses Ia (with the ATP-cone) and Ib General DNA and protein techniques - (lacking the ATP-cone). Many class II RNRs Molecular cloning techniques were by standard lack the ATP-cone, whereas most class III procedures (17). Genomic DNA of P. RNRs include an ATP-cone. In addition, groups aeruginosa PAO1 was isolated using the of organisms have RNRs that diverge from this DNeasy Tissue Kit (Quiagen) according to the general presence or absence of ATP-cone, e.g. manufacturer’s specifications. the herpesviridae carry an ATP-cone that has lost its capacity to bind nucleotides. Among the Construction of overproduction strains - All eubacterial species there are a few examples of constructs were cloned into the over-expression organisms that carry more than one ATP-cone. plasmid pET22b. The entire wild type nrdA The most striking example is that all known gene was amplified from genomic DNA by Chlamydiaceae have an N-terminal triplication PCR using the following primers; forward OP- of their ATP-cone (14), and some NrdA1up, 5’- proteobacteria have a duplication of the ATP- ACATATGCATACCGACACCACACG-3’ cone. and reverse OP-NrdAlw, 5’- One example of an ATP-cone TAAGCTTGCCCGGCCGCGACACGGAC-3’. duplication is found in Pseudomonas The primers were designed to generate an NdeI aeruginosa, a γ−proteobacterium that encodes and HindIII restriction site (underlined), three different classes of RNR, Ia, II and III respectively, at the start and end of the resulting (15). We have earlier shown that all three RNR amplified fragment. The 2937-bp PCR product classes in P. aeruginosa are expressed, albeit at was cloned in the pGEM-T Easy vector. After different growth conditions, and are digestion with NdeI and HindIII, the nrdA enzymatically functional (15,16). All class Ia fragment was ligated into pET22b generating RNRs consist of two different components, plasmid pETS141. The following primers NrdA and NrdB. The active site region where (underlining shows the NdeI restriction site) the substrates bind is found in the NrdA were used for PCR amplification of the component, whereas the NrdB component different nrdA fragments and the nrdB gene; for harbors a stable tyrosyl radical that is needed to nrdAΔ147 forward OP-NrdA2up, 5’- initiate the radical-based catalysis. The aim of ACATATGCGCATCACCCGCGCCG-3’ and this work was to understand why a duplication reverse OP-NrdAlw, for nrdAΔ182 forward 2
OP-NrdA3up, 5’- DuoFlow System fast protein liquid ACATATGATCGAGCGCGAGACCC-3’ and chromatography instrument (BioRad) reverse OP-NrdA-lw, and for nrdB forward OP- previously equilibrated with 10 volumes of NrdBup, 5’- buffer A. The different RNR components (wild GAGCCATATGCTGAGCTGGGACG-3’ and type NrdA and NrdB, truncated NrdA proteins) reverse OP-NrdBlw, 5’- were eluted with a linear gradient of ammonium AGGATCCTAAATCGAGCATCTCAATCC- sulfate (0.75 to 0 M, 1 ml/min) in buffer B (50 3’ (this underlining shows a BamHI restriction mM Tris-HCl, pH 7.5, 2 mM DTT). After SDS- site). All amplified fragments were cloned into PAGE fractions with the highest purity were the pGEM-T Easy vector as described concentrated using Centricon (Millipore). previously. The NdeI-HindIII digested DNA NrdB protein was further purified by fragments nrdAΔ147 (2499 bp) and nrdAΔ182 ion exchange chromatography. Pooled fractions (2385 bp) were cloned into the pET22b were diluted to 5 mg/ml protein concentration generating pETS142 and pETS143, and loaded on a HiLoad™ 16/10 Q-Sepharose® respectively. The NdeI-BamHI digested nrdB High Performance (GE Healthcare) pre- DNA fragment (1268 bp) was cloned into equilibrated with buffer B plus 30 mM KCl. pET22b generating plasmid pETS144. All four NrdB protein was eluted with a linear gradient pETS clones were sequenced in both directions of KCl (30 to 300 mM, 3 ml/min) in buffer B. to assure the lack of mutations during the PCR Fractions containing the NrdB protein were amplification and cloning steps. pooled, concentrated using Centricon 30 Downloaded from http://www.jbc.org/ by guest on March 3, 2015 (Millipore) and step-by-step the buffer was Overproduction and purification - The E. coli exchanged to buffer B. Rosetta (DE3) cells containing any of the NrdAΔ182 protein was purified as pETS141, pETS142, pETS143 and pETS144 described for the NrdB protein followed by a were grown in LB medium (Difco) with 100 gel filtration chromatography. Pooled fractions µg/ml ampicillin and 17 µg/ml from the ion exchange purification were loaded chloramphenicol. The cultures were grown with at 0.5 ml/min on a 24 ml Superdex-75 vigorous shaking (250 rpm) at 30ºC. When cells column equilibrated and eluted with buffer B reached an OD550≈0.5 expression of the cloned plus 200 mM KCl. Each fraction was analyzed gene was induced by the addition of IPTG to a by Phastgel electrophoresis (GE Healthcare) final concentration of 0.5 mM. After four hours and fractions with the highest purity were of induction the cells were harvested by pooled, concentrated in a Centriprep-50 centrifugation and cell pellets were frozen on (Millipore) and finally freed from KCl by dry ice and stored at –80ºC. washing with buffer B. All the protein purification steps were carried out at 4ºC. All RNR proteins used in Iron reconstitution of P. aeruginosa NrdB for this study were purified using the same enzymatic assays - NrdB protein was procedure. Frozen cells were disintegrated by reconstituted at 25ºC by reaction with anaerobic X-press (Biox AB), homogenized and extracted ferrous iron ascorbate solution under aerobic with 4 volumes (v/w) of buffer containing 50 conditions as previously described (18) with a mM Tris-HCl, pH 7.5, 30 mM KCl, 10 mM molar ratio of 3 irons per NrdB polypeptide. DTT and Complete protease cocktail inhibitor After 5 minutes the reaction mixture was (Roche), and centrifuged for one hour at desalted on a NAP-10 column (GE-Healthcare) 40 000xg. Nucleic acids were precipitated by equilibrated with buffer B and concentrated addition of streptomycin sulfate to a final with Centricon 30 (Millipore). concentration of 1% and removed by centrifugation. The supernatant was precipitated Gel filtration in presence of effector nucleotides by slow addition of ammonium sulfate to 45% - A 2 ml column of Superdex-200 (GE- final saturation. After centrifugation the Healthcare) was used to estimate the molecular precipitate was dissolved in 5 ml of buffer A mass of the NrdA, NrdAΔ147 and NrdB (0.75 M ammonium sulfate, 5 mM DTT). proteins alone or in combination. Proteins at Around 200 mg of total protein was applied to a 1.25, 2.5 and 5 mg/ml were chromatographed at HiLoad16/10 PhenylSepharose High 0.1 ml/min in buffer containing 50 mM Tris- Performance (GE Healthcare) on a BioLogic HCl, pH 7.5, 10 mM MgCl2, 10 mM DTT, 100 3
mM NaCl, and 5 mM ATP. The column had addition of Fe2+ and the formation and decay of been standardized with thyroglobulin (669 absorption bands were observed between 300- kDa), apoferritin (443 kDa), β−amylase (200 500 nm in cycles of 90 s with a scanning rate of kDa) and alcohol dehydrogenase (150 kDa), all 240 nm/min. The tyrosyl radical concentration supplied from Sigma. at different time points was determined by To estimate the relative proportions of subtracting an “endpoint” spectrum at ca 30 NrdA versus NrdB components in the min after iron addition from the spectrum holoenzyme complexes, eluted protein peaks immediately after iron addition, giving a tyrosyl were analyzed by SDS-PAGE followed by radical absorption at 413 nm (Supplement, Fig. densitometric analyses. 1) that was quantified using the extinction coefficient 3.25 mM–1cm–1 (23). Enzymatic assay and other analytical For experiments measuring iron content techniques - Standard conditions implied the during radical decay 1.2 ml of apoNrdB was use of 0.7 mM 3H-CDP (10 775 cpm/pmol), 5 reconstituted with 1.5 Fe per polypeptide, since mM ATP, 30 mM DTT, 30 mM magnesium higher iron-to-protein ratio gave less well- acetate, 1 mg/ml BSA, 30 mM Tris-HCl, pH resolved spectra. After 1, 2, 5, 9 and 15 scans 7.5 and 30 nmoles of NrdA and 60 nmoles of (i.e. after 70, 190, 560, 1040 and 1500 s) 110 µl NrdB in a final volume of 50 µl. Incubation was was withdrawn, mixed with 2 µl 15 mM EDTA, at 25ºC for 20 min and the amount of dCDP desalted on a NAP-5 column equilibrated with formed was determined by the standard method 50 mM Tris-HCl, pH 7.6, and used for analyses Downloaded from http://www.jbc.org/ by guest on March 3, 2015 as previously described (19). One unit of of iron and protein concentrations. The enzyme corresponds to 1 nmol of product registered spectra (Fig. 3 inset) were used to formed per min, and the specific activity is calculate the tyrosyl radical after subtraction as expressed as units per mg total protein. KL, KD described above. and Vmax values were obtained by direct curve fitting as described before (20,21) using the EPR Spectroscopy - EPR spectra were KaleidaGraph software (Synergy Software). registered on an ESP 300 X-band spectrometer Standard errors of mean shown in tables 3-4 (Bruker) equipped with a liquid nitrogen flow and in figures 5-6 are not repeated in the text. system for temperatures ≥90K. Radical Protein concentrations were determined concentrations were determined using a B. by the Bradford assay, with crystalline bovine anthracis NrdF sample and an E. coli NrdB serum albumin as a standard. Iron analyses sample with known tyrosyl radical were determined as described by Sahlin & al. concentrations as secondary standards. The (22). error limits for radical quantitations are ≤±10%. UV-vis absorption spectroscopy and formation Nucleotide binding assay - Nucleotide binding and decay of radical-diiron site - Stopped flow was performed according to the method of spectroscopy was applied to study formation of direct partition through ultrafiltration (24). The the radical-diiron site. Kinetic traces were experiments were carried out at 25ºC, using a registered every 5th nm in the region 440-350 range of 0.5 to 100 µM NrdA and NrdAΔ147 nm using a DX 18MV Biosequential stopped polypeptide and 0.5 to 100 µM dATP/ATP. flow ASVD spectrometer (Applied Photophysics) and a split time base of 2/20s. RESULTS Equal volumes of 80 µM apo-NrdB polypeptide (aerobic) and 80 µM Fe2+ (anaerobic) were Atypical class Ia RNR in P. aeruginosa mixed at 25°C, and 55 µl of each reactant were The nrdA and nrdB genes code for the used per shot. class Ia RNR that is active only during aerobic Scanning spectra were registered on a conditions (15,16). The nrdA gene (2892 bp) Perkin Elmer Lambda2 spectrometer at 25° C. encodes a protein of 963 amino acid residues Typically 20 µM NrdB polypeptide in 50 mM with an expected molecular mass of 107 kDa. Tris-HCl, pH 7.6 was used. For reconstitution The nrdB gene (1248 bp) encodes a protein of 1-6 µl of 5 mM Fe2+ in 0.1 M HCl was added to 415 amino acid residues with an expected 500 µl protein solution giving 0.5-3 Fe/NrdB molecular mass of 47 kDa. Whereas the NrdB polypeptide. Spectra were registered before protein is overall similar to other NrdB 4
proteins, the NrdA protein revealed interesting were carried out for each protein (see Material differences compared to other well-known class and Methods) yielding approximately 8.5 mg of Ia RNRs. The deduced amino-acid sequence of pure protein per liter of culture for the NrdA P. aeruginosa NrdA is 202 residues longer than wild type, 17.6 mg per liter for the NrdAΔ147 the 761-residues E. coli class Ia prototype. and 22.5 mg per liter for the NrdB purification. Alignment of these two proteins shows that After several purification steps for the most of the extra amino acids (around 155 NrdAΔ182 protein we obtained only a partially residues) are located at the N-terminus of the purified fraction (around 60% purity) with a protein (Fig. 1a). The remaining 47 residues are yield of 2.5 mg per liter of culture. SDS-PAGE found in the C-terminal part of the protein. All of the purified proteins gave only one band for the essential cysteine residues known to be NrdA wild type, NrdAΔ147 and NrdB with involved in catalysis and enzyme turnover and apparent polypeptide sizes similar to the other residues that are functionally important in theoretical molecular masses of the cloned class I proteins are present in the deduced P. genes (107, 91 and 47 kDa respectively; Fig 2). aeruginosa NrdA protein (data not shown). One of the two prominent bands in the Comparison of the NrdA against the Pfam NrdA∆182 preparation also corresponds to its database (25) revealed high similarity to the theoretical molecular mass (88 kDa, Fig. 2, lane RNR central core (residues 293-910, 3). Western-blot analysis using polyclonal Pfam02867; E value 5*e–187) and to an ATP- antibodies against the wild type P. aeruginosa cone domain (residues 32-132 (denoted ATP- NrdA protein confirmed that only the 88-kDa Downloaded from http://www.jbc.org/ by guest on March 3, 2015 c1), Pfam03477; E value 3*e–28) previously band is a NrdA protein product (data not described as the allosteric site regulating the shown). overall enzyme activity of the enzyme (10). With a lower similarity (E value 5*e–3) we General biochemical properties of the purified found another ATP-cone structure between the NrdA proteins two structures previously identified (residues None of the purified proteins had any 151-250, denoted ATP-c2). RNR activity on their own (data not shown). The amino acid sequence of the N- Combinations of either NrdA wild type or terminal region of P. aeruginosa NrdA (Fig. NrdAΔ147 proteins with reconstituted NrdB 1b) implies a duplication of the residues for a protein (see below and Table 1) exhibited putative activity site domain. As shown in specific class Ia RNR enzyme activity. Using figure 1b, several residues responsible for the ATP as positive allosteric effector and DTT as nucleotide binding (10) are present in both an artificial electron donor for this system we putative allosteric domains (ATP-c1 and ATP- obtained a specific activity of 88 U/mg for the c2). To study whether both domains can bind NrdA wild type and 96 U/mg for the nucleotides and regulate the overall activity of NrdAΔ147. The preferred substrate for this the enzyme two different constructions were system is nucleoside diphosphates with a KM made (Fig. 1a) and their biochemical and for CDP of 27±8 µM for both NrdA wild type allosteric properties were studied. and NrdAΔ147. A calculated half maximal activity was obtained at 7.9 mM DTT for the Purification of wild type P. aeruginosa proteins NrdA wild type protein and at 3.6 mM for and truncated NrdA proteins NrdAΔ147. The full-length P. aeruginosa nrdA No activity was detected when gene product and two engineered variants of NrdAΔ182 was assayed together with the truncated nrdA, designated NrdAΔ147 and reconstituted NrdB protein. Competition assays NrdAΔ182, were overproduced in E. coli. The of the wild type NrdA protein with different P. aeruginosa NrdAΔ147 protein (816 residues) amounts of NrdAΔ182 showed no inhibition of lacks the first activity site (ATP-c1). The the wild type protein activity (data not shown) NrdAΔ182 (781 residues) lacks the first and suggesting that the NrdA∆182 protein cannot part of the second activity site domain (ATP- bind to the NrdB protein. Due to the lack of c2) and corresponds in size and protein enzymatic activity with the NrdAΔ182 protein sequence to the class Ib NrdE proteins (Fig. 1a). no further experiments were carried out with The P. aeruginosa nrdB gene was also this protein. expressed in E. coli. Different purification steps 5
The tyrosyl radical in P. aeruginosa NrdB components (Fig. 4b). However, the radical protein forms and decays rapidly content is lower than normally observed when The purified NrdB protein contained no recombinant NrdB proteins are expressed and tyrosyl radical and no diiron site as determined analyzed under similar conditions (e.g. Bacillus from EPR, light absorption and iron analysis. anthracis NrdF with ca 45 µM radical content, Reconstitution of purified NrdB apoprotein Supplement, Fig. 3), suggesting that the P. with Fe2+/O2 gives a spectrum with aeruginosa tyrosyl radical decays also under characteristic diiron-oxo charge transfer bands these conditions. at 325 and 370 nm and a prominent tyrosyl radical band at 413 nm (top spectrum, inset Fig. Recombinant NrdB protein requires continuous 3). Stopped flow experiments showed that the oxygen-dependent reactivation similar to the major fraction of the radical and the diiron site native P. aeruginosa system formed with a rapid rate of ca 4 s-1 and a minor As expected, the enzyme activity of the fraction (relative amplitude 0.13-0.17 of the reconstituted P. aeruginosa NrdB protein faster rate) with a rate of ca 0.15 s-1 followed the radical content when assayed (Supplement, Fig. 2). This means that the immediately after addition of iron (Table 1), but radical and diiron sites were essentially formed what happens to the diiron site with time and after 2 s. Increasing amounts of added iron during assay conditions? We have shown that resulted in increasing amount of radical up to a the radical and diiron-oxo features have ratio of 2 Fe/polypeptide when it leveled off (cf. disappeared after long incubation times (lowest Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Table 1). Addition of more iron did not increase spectrum, inset Fig. 3), but that iron still the amount of formed radical, instead it remains bound to the protein (Fig. 3). Notably, increased background absorbance and hence a reconstituted NrdB sample that was desalted disturbed the resolution of diiron-oxo charge and concentrated as the samples used for the transfer bands in the spectra. activity assays described above, displayed no Surprisingly, the formed spectra radical band but had almost full enzyme activity immediately began to decay, losing both radical (Table 1, lowest line). This experiment and diiron-oxo bands concomitantly (Fig. 3). convincingly shows that the radical is The decay of the tyrosyl radical can be fitted efficiently regenerated during the normal with two rate constants of 0.016±0.005 s-1 and enzyme activity measurements. 0.0011±0.0001 s-1 (relative amplitude 2.3 of the Class Ia RNR purified from growing P. faster decay). Tyrosyl radical decay was not aeruginosa cells was previously found to prevented by reconstitution in the presence of 1 require continuous supply of oxygen during mg/ml BSA (as used in activity measurements) enzyme activity measurements (15). To and/or an excess of NrdA (data not shown). establish if this was also the case for the Instead, we observed that the iron ions recombinant P. aeruginosa class Ia RNR remained tightly bound to the protein (Fig. 3) system used in this study we performed despite the fact that the typical diiron-oxo experiments where the enzyme mixtures were charge transfer bands decayed concomitantly preincubated anaerobically or aerobically for 40 with the decay of the radical band (Fig. 3 inset). min and then assayed for enzyme activity Rapid freezing of a sample 4 s after anaerobically or aerobically. The recombinant addition of 2.5 Fe per NrdB polypeptide yielded P. aeruginosa class Ia RNR system behaves as 0.29 radical/monomer. The tyrosyl radical in P. the native system and has a prominent aeruginosa NrdB as observed by EPR is a requirement for continuous oxygen supply typical class Ia radical in line shape (Fig. 4a). It during enzyme activity measurements (Table differs from the E. coli radical (Fig. 4c) in the 2). In a parallel set of assays we added EDTA resolution of the spectrum, and the 0.7 mT to the samples to chelate any iron that was not hyperfine splittings (26) from the 3,5 ring bound to protein. The results in Table 2 show protons of the P. aeruginosa NrdB tyrosyl that the enzyme activity is the same in presence radical are well resolved in the 93K spectrum. as in absence of EDTA, again demonstrating The P. aeruginosa NrdB tyrosyl radical content that the iron ions are tightly bound to the of ca 5 µM in overproducing E. coli cells shows protein and that continuous regeneration of the that the P. aeruginosa radical can be formed in radical requires oxygen. the absence of any additional P. aeruginosa 6
Only one allosteric overall activity site in P. the wild type protein and 16±3 µM for the aeruginosa NrdA is fully functional. NrdAΔ147 protein. In general the Vmax values As shown in figure 1b two putative for both proteins were considerably lower with allosteric activity sites (ATP-c1 and ATP-c2) dTTP as effector (Fig. 6b, starting values) were found in the wild type NrdA protein. To compared to those obtained with ATP (Fig. 5a) test whether these sites are fully functional we or dATP. Binding of dTTP to the allosteric studied the behavior of the wild type protein specificity site was efficiently displaced by and the truncated protein NrdAΔ147 against increasing concentrations of dATP, but with different concentrations of ATP and dATP. In diametrically opposite results for NrdA∆147 an ATP curve both proteins showed similar and NrdA wild type (Fig. 6b). In NrdA∆147, behavior (Fig. 5a) and KL values (2.9 mM for dATP amplified the dTTP-stimulated reaction the NrdA wild type and 2.7 mM for more than 5-fold (Fig. 6b), as expected when NrdAΔ147). The NrdAΔ147 protein had a two positive effector nucleotides compete for slightly higher calculated Vmax (118 U/mg) the specificity site. The apparent KL of about 33 compared to the wild type (95 U/mg). µM for the dATP stimulation in this At low concentrations dATP works as a competition experiment compares well with the positive effector for class Ia RNR by binding to KL of 17.3 µM for dATP’s allosteric effect on the allosteric specificity site, but at high the NrdA∆147 protein. In stark contrast, dATP concentrations it inhibits the enzyme activity by inhibited the dTTP-stimulated reaction in the binding to the allosteric overall activity site. wild type NrdA protein approximately 4-fold, Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Titration curves of the P. aeruginosa NrdA as expected if dATP binds both competitively wild type and NrdA∆147 proteins with different with dTTP to the specificity site and with concentrations of dATP highlighted drastic negative allosteric effect to the overall activity differences between the two proteins (Fig 5b). site. This experiment shows conclusively that As expected for a typical NrdA protein the P. the truncated NrdA∆147 protein retains only aeruginosa wild type protein was activated by the allosteric specificity site and lacks the low concentrations of dATP with a KL1 allosteric overall activity site, confirming that (specificity site binding) of 3.2 µM and the ATP-c2 in P. aeruginosa NrdA has lost its inhibited by higher dATP concentration with a function as an allosteric site. KL2 (activity site binding) of 17.3 µM. Contrarywise, the NrdAΔ147 protein was The wild type P. aeruginosa class Ia RNR is a completely resistant to high concentrations of high molecular mass complex dATP and remained fully active even at 1 mM To estimate the apparent molecular dATP (Fig. 5b) with a KL1 value of 17.1 µM. A weight of the complex formed by the wild type likely explanation for the observed difference in NrdA or NrdAΔ147 with the NrdB, mixtures of the dATP response of the two proteins could be proteins at different concentrations (1.25, 2.5, 5 in the ability to bind the nucleotide. Binding mg/ml) were chromatographed on a assays confirmed that the NrdA wild type Superdex-200 column. The Ve/Vo (elution protein binds 2.1 molecules of dATP per volume/void volume) was calculated and polypeptide chain (KD=15 µM, Fig 6a); translated to molecular mass for all the samples plausibly one dATP bound at the allosteric (Table 3). NrdB was resolved in a single peak specificity site and one at the allosteric overall with an estimated molecular mass of 184 kDa, activity site, as is typically found for other class which corresponds to around four NrdB Ia NrdAs (4). On the other hand the NrdAΔ147 subunits (β4). The NrdA wild type protein protein binds only half as much dATP, 0.9 eluted in a major peak (estimated molecular molecules per polypeptide chain (KD=26 µM, masses of 528 kDa) and a minor peak Fig 6a), in this case only to the specificity site. (estimated molecular mass ca. 120 kDa) These experiments suggest that the second corresponding to oligomerization of four NrdA putative nucleotide-binding site (ATP-c2) is not wild type subunits (α4-6) and to a single NrdA allosterically functional. wild type protein (α). NrdAΔ147 eluted as a The allosteric effector dTTP that is single peak corresponding an estimated known to bind only to the specificity site homodimeric (α2) molecular mass of 202 kDa. stimulated in our case the reduction of CDP in Equimolar mixtures (at 1.25, 2.5 and 5 both proteins with a KL value of 19±3 µM for mg/ml) of NrdA wild type or NrdAΔ147 with 7
NrdB protein in the presence of ATP as a emphasize that P. aeruginosa class Ia RNR positive effector were used to detect oligomeric differs in several aspects from the well- forms of active complexes (Table 3). For the characterized class Ia RNRs of E. coli, first time we could observe distinct complexes bacteriophage T4 and mouse (29). in gel filtration chromatographies compared to Firstly, the NrdA component of P. other RNRs studied so far (27,28). Wild type aeruginosa class Ia RNR contains in its N- complexes eluted in two peaks with molecular terminus a duplication of the allosteric activity masses estimated to be 634 and 217. Analysis domain, which is a member of the ATP-cone of the high molecular mass peak showed an family (13). We engineered a precise truncation equimolar amount of NrdA and NrdB of the N-terminal ATP-cone (NrdA∆147) and polypeptides and the molecular mass indicated measured the allosteric regulation of the that four subunits each of NrdA wild type and NrdA∆147 enzyme activity as well as its NrdB proteins (α4β4) were present in the number of binding sites and dissociation complex (Table 3). For the second peak the constants for effector nucleotides. The wild densitometric analyses indicated two NrdB type P. aeruginosa NrdA displayed the polypeptides per each NrdA, and the observed canonical allosteric response with an initial molecular mass indicated an αβ2 complex. activation by dATP followed by an inhibition of Surprisingly, also the truncated protein enzyme activity at higher dATP concentrations, NrdAΔ147 can form a tight complex to the wild whereas the truncated NrdA∆147 responded to type NrdB protein. The complex eluted in a dATP activation but lacked the dATP inhibitory Downloaded from http://www.jbc.org/ by guest on March 3, 2015 single peak with an estimated molecular weight response. Competition experiments with dATP of 250 kDa corresponding to a heterodimeric in presence of dTTP convincingly showed that structure (α2β2) of the active complex (Table 3), the remaining ATP-cone (ATP-c2) in the which was corroborated by the densitometric truncated NrdA∆147 protein lacked nucleotide- analysis. binding properties. ATP-c2 lacks some of the side chains known to be involved in nucleotide P. aeruginosa NrdA and NrdB form a tight binding in E. coli NrdA (Fig. 1b), e.g. the complex in presence of ATP equivalent of Lys-21 (10), and also His-59 In our gel filtration experiments we recently shown to be important for allosteric observed distinct complexes between the P. regulation (21), offering a plausible explanation aeruginosa NrdA and NrdB components for the lack of allosteric response in ATP-c2. indicating that the proteins formed strong Secondly, the P. aeruginosa NrdAB complexes. We titrated the NrdB protein over holoenzyme forms at least a 7-fold stronger the NrdA wild type or NrdAΔ147 proteins in complex than e.g. the E. coli NrdAB the presence of ATP and estimated a KD of ca. holoenzyme complex (30) and appears to have 60 nM (based on NrdA and NrdB polypeptide a different quaternary structure. The truncated concentrations) for the NrdA wild type and the P. aeruginosa NrdA∆147/NrdB complex also NrdAΔ147 complexes (Table 4, and has a strong KD, whereas the additionally truncated NrdA∆182 does not interact with Supplement Fig. 4). The KDs were also NrdB, suggesting that residues between measured in the presence of dTTP as a positive positions 148 and 182 in NrdA may influence effector. The results showed a considerably weaker complex with a KD of 0.41 µM the strength of the complex. The strong interaction of P. aeruginosa NrdAB was (polypeptide concentrations) for a wild type reaffirmed with gel filtration experiments in complex and a KD of 0.79 µM (polypeptide which the active complex chromatographed concentrations) for a NrdAΔ147 complex with an apparent heterotetrameric structure (Table 4). (α4β4) in contrast to the α2β2 structure found for DISCUSSION E. coli NrdAB (29,31,32), and the hexameric form (α6β6) of mouse NrdAB in presence of In this study we have characterized the ATP (33). Interestingly, the oligomeric P. aeruginosa class Ia RNR, one of the three structure of the NrdA∆147/NrdB complex was physiologically relevant RNRs in this heterodimeric (α2β2) suggesting that the N- opportunistic pathogen (15,16). Three terminal ATP-cone (ATP-c1) in P. aeruginosa unexpected features of the class Ia RNR NrdA is crucial for the wild type quaternary 8
structure of the NrdAB complex as well as for of a class Ia is widespread. Figure 7 clearly the allosteric regulation of overall enzyme shows that P. aeruginosa NrdA clusters activity. together with NrdAs from some β- and γ- Thirdly, the P. aeruginosa NrdB diiron- proteobacteria and the Chlamydiaceae. tyrosyl radical center formed with about the Interestingly, all these proteins have a same rate as the corresponding center in E. coli duplication (or triplication) of the ATP-cone (34,35), but the P. aeruginosa radical-diiron domain. Other NrdAs in the β- and γ- center is short lived and the radical and diiron proteobacteria are clustered in a different sites decomposed together. This could reflect a subgroup, and all these sequences have a single more open structure and have functional ATP-cone domain. The observed tree topology reasons. Interestingly, the iron ions remained (a tree based on only the catalytic part of the bound to the protein and the radical was protein has the same topology) suggests that the efficiently reactivated during assay conditions. alteration in the N-terminal region has occurred We also observed that the functional enzyme recently and is not a rare event. Why is this needs continuous supply of oxygen, as expected duplication maintained and widely distributed? if the diiron-radical center has to be reformed Our results show that only one ATP-cone continuously for reduction of ribonucleotides. (ATP-c1) is allosterically functional, whereas The oxygen requirement of the native class Ia the role of the other ATP-cone (ATP-c2) RNR activity was previously observed in crude remains unclear. We suggest that part of the extracts from P. aeruginosa (15). As also ATP-c2 domain might be important for NrdB Downloaded from http://www.jbc.org/ by guest on March 3, 2015 shown earlier expression of class Ia RNR in P. interaction, since the deletion of only 35 aeruginosa drops 6-fold in early stationary additional residues (NrdA∆182) mimicking an phase, when the expression of class II RNR NrdE protein knocked out enzyme activity and increases 6-7-fold (16). The atypical continuous interaction with the NrdB protein. oxygen-requirement of the P. aeruginosa class The widespread duplication of the Ia RNR could imply that enzyme activity is ATP-cone in RNRs is an illustration of the inactivated in microaerophilic conditions and plasticity of this domain. Interestingly, in may explain why this organism also encodes a Chlamydia the catalytic domain and the class II RNR. The presence of functional genes adjacent ATP-cone seem to have eukaryotic for all three classes of RNRs may thus allow P. origin whereas the two proximal copies are aeruginosa to proliferate at any given oxygen more similar to prokaryotic counterparts availability that it encounters. (13,14). Hence it was suggested that a Interestingly, analysis of the flanking recombination between eukaryotic and bacterial regions of the P. aeruginosa nrdAB operon (16) genes might have resulted in a protein with revealed that it is preceded by a transposase multiple ATP-cones. We suggest that class gene (PA1153, 177 residues) and followed by diversification within the RNR family is an an open reading frame of phage origin example of protein modularity. The modularity (PA1154, 184 residues). These two proteins are concept is an attractive hypothesis that has probably reminiscent footprints of a phage and implications for how proteins may have support the idea that the class Ia genes might evolved in Nature. It has been suggested that have been introduced in the P. aeruginosa construction of new types of proteins with genome via a lateral gene transfer event (3). diverse functions could occur by use of simple Crucial for the long-term persistence of newly building blocks or modules (36,37). In RNR the acquired DNA is that it confers a selectable central catalytic core of the enzyme (the 10- function. As the different RNRs have different stranded β/α-barrel structure) has been highly oxygen dependencies it is expected that constrained during evolution, as seen in the multiple RNRs will present selective conserved three-dimensional structure (5-8), advantages for organisms that explore several whereas the different domains (modules) different ecosystems, e.g. the Pseudomonas connected to the core region are more labile. genus. Subtle shuffling and recombination involving A comparison of the P. aeruginosa the central catalytic domain, the allosteric ATP- NrdA protein with other bacterial class Ia cone domain(s) and docking sites for other proteins gives surprising results because the ligands may have formed the three well-defined presence of a duplication in the N-terminal part contemporary RNR classes. Without these extra 9
domains, the enzyme core could not sustain the prerequisites of the environments. Therefore, required fine-tuning of the dNTP pools needed contemporary RNRs are a functional and for accurate DNA synthesis and repair. Neither structural mosaic family of proteins as obvious could it achieve the radical-based reaction for instance with the availability of different mechanism, which is initiated in additional RNR protein sequences (38). domains and/or components as specified by the ACKNOWLEDGEMENTS We are grateful to Patrick Young and Anthony Poole for valuable discussions and critical comments on the manuscript. This study was supported by grants from the Swedish Cancer Foundation and the Swedish Research Council (to BMS) and from Carl Trygger Foundation (to MS). ET was supported by a postdoctoral fellowship from the Spanish Ministerio de Educación y Ciencia. ABBREVIATIONS BSA, bovine serum albumin; DTT, dithiothreitol; EPR, electron paramagnetic resonance; IPTG, isopropyl-1-thio-β−D-galactopyranoside; PCR, polymerase chain reaction; RNR, ribonucleotide Downloaded from http://www.jbc.org/ by guest on March 3, 2015 reductase; SDS-PAGE, sodium dodecyl sulphate polyacrylamide gel electrophoresis REFERENCES 1. Stubbe, J. (2000) Curr Opin Struct Biol 10, 731-736 2. Poole, A. M., Logan, D. T., and Sjöberg, B.-M. (2002) J Mol Evol 55, 180-196 3. Torrents, E., Aloy, P., Gibert, I., and Rodriguez-Trelles, F. (2002) J Mol Evol 55, 138-152 4. Nordlund, P., and Reichard, P. (2006) Annu Rev Biochem 75, in press 5. Uhlin, U., and Eklund, H. (1994) Nature 370, 533-539 6. Logan, D. T., Andersson, J., Sjöberg, B.-M., and Nordlund, P. (1999) Science 283, 1499-1504 7. Sintchak, M. D., Arjara, G., Kellogg, B. A., Stubbe, J., and Drennan, C. L. (2002) Nat Struct Biol 9, 293-300 8. Uppsten, M., Färnegårdh, M., Jordan, A., Eliasson, R., Eklund, H., and Uhlin, U. (2003) J Mol Biol 330, 87-97 9. Kunz, B. A., Kohalmi, S. E., Kunkel, T. A., Mathews, C. K., McIntosh, E. M., and Reidy, J. A. (1994) Mutat Res 318, 1-64 10. Eriksson, M., Uhlin, U., Ramaswamy, S., Ekberg, M., Regnström, K., Sjöberg, B.-M., and Eklund, H. (1997) Structure 5, 1077-1092 11. Larsson, K. M., Andersson, J., Sjöberg, B.-M., Nordlund, P., and Logan, D. T. (2001) Structure 9, 739-750 12. Larsson, K. M., Jordan, A., Eliasson, R., Reichard, P., Logan, D. T., and Nordlund, P. (2004) Nat Struct Mol Biol 11, 1142-1149 13. Aravind, L., Wolf, Y. I., and Koonin, E. V. (2000) J Mol Microbiol Biotechnol 2, 191-194 14. Roshick, C., Iliffe-Lee, E. R., and McClarty, G. (2000) J Biol Chem 275, 38111-38119 15. Jordan, A., Torrents, E., Sala, I., Hellman, U., Gibert, I., and Reichard, P. (1999) J Bacteriol 181, 3974-3980 16. Torrents, E., Poplawski, A., and Sjöberg, B.-M. (2005) J Biol Chem 280, 16571-16578 17. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory Press 18. Atkin, C. L., Thelander, L., Reichard, P., and Lang, G. (1973) J Biol Chem 248, 7464-7472 19. Thelander, L., Sjöberg, B.-M., and Eriksson, S. (1978) Methods Enzymol 51, 227-237 20. Climent, I., Sjöberg, B.-M., and Huang, C. Y. (1991) Biochemistry 30, 5164-5171 21. Birgander, P. L., Kasrayan, A., and Sjöberg, B.-M. (2004) J Biol Chem 279, 14496-14501 10
22. Sahlin, M., Sjöberg, B.-M., Backes, G., Loehr, T., and Sanders-Loehr, J. (1990) Biochem Biophys Res Commun 167, 813-818 23. Petersson, L., Gräslund, A., Ehrenberg, A., Sjöberg, B.-M., and Reichard, P. (1980) J Biol Chem 255, 6706-6712 24. Ormö, M., and Sjöberg, B.-M. (1990) Anal Biochem 189, 138-141 25. Finn, R. D., Mistry, J., Schuster-Bockler, B., Griffiths-Jones, S., Hollich, V., Lassmann, T., Moxon, S., Marshall, M., Khanna, A., Durbin, R., Eddy, S. R., Sonnhammer, E. L., and Bateman, A. (2006) Nucleic Acids Res 34(Database issue), D247-251 26. Lendzian, F. (2005) Biochim Biophys Acta 1707, 67-90 27. Birgander, P. L., Bug, S., Kasrayan, A., Dahlroth, S. L., Westman, M., Gordon, E., and Sjöberg, B.-M. (2005) J Biol Chem 280, 14997-15003 28. Gleason, F. K., and Olszewski, N. E. (2002) J Bacteriol 184, 6544-6550 29. Eklund, H., Uhlin, U., Färnegårdh, M., Logan, D. T., and Nordlund, P. (2001) Prog Biophys Mol Biol 77, 177-268 30. Climent, I., Sjöberg, B.-M., and Huang, C. Y. (1992) Biochemistry 31, 4801-4807 31. Brown, N. C., and Reichard, P. (1969) J Mol Biol 46, 25-38 32. Thelander, L. (1973) J Biol Chem 248, 4591-4601 33. Kashlan, O. B., and Cooperman, B. S. (2003) Biochemistry 42, 1696-1706 34. Bollinger, J. M. J., Edmondson, D. E., Huynh, B. H., Filley, J., Norton, J. R., and Stubbe, J. (1991) Science 253, 292-298 Downloaded from http://www.jbc.org/ by guest on March 3, 2015 35. Sahlin, M., Lassmann, G., Pötsch, S., Sjöberg, B.-M., and Gräslund, A. (1995) J Biol Chem 270, 12361-12372 36. Fraser, H. B. (2005) Nat Genet 37, 351-352 37. Nikitin, F., and Lisacek, F. (2003) Comput Biol Chem 27, 481-495 38. Lundin, D., Torrents, E., Furrer, E., Larsson Birgander, P., Sahlin, M., Poole, A. M., and Sjöberg, B.-M. (2005) RNRdb: Ribonucleotide Reductase Database. In., RNRdb: Ribonucleotide Reductase Database (http://rnrdb.molbio.su.se) Dept. of Molecular Biology, Stockholm University, Stockholm, Sweden 11
Table 1. Radical content and specific activity of P. aeruginosa NrdB as a function of added iron during reconstitution. Iron added per NrdBa (mol/mol) Radical content (µM) Enzyme activityd (mU) 0b 0 ≤26d 0.5 2.1 185 1 4.2 280 1.5 5.2 350 2 5.5 280 2.5 5.7 374 3 5.3 403 3 (after ≈90 min, plus desalting)c
Table 3. Complexes formed with P. aeruginosa RNR wild type proteins and with NrdA∆147 NrdA wild type NrdA∆147 NrdB Additions a a kDa Composition kDa Composition kDa Compositiona 528 ±81 α4-6b none 202 ±11 α2 184 ±10 β4 ≈120 αb 634 ±40 α4β4 NrdB 250 ±19 α2β2 n.a.c 217 ±11 αβ2 a closest integer(s) b the major component is α4-6, whereas α is a minor component c n.a., not applicable Table 4. P. aeruginosa NrdAB RNR forms a strong complex in presence of ATP.a Downloaded from http://www.jbc.org/ by guest on March 3, 2015 KD(NrdAB) (µM)b NrdA protein in presence of ATP in presence of dTTP NrdA wt 0.059 ±0.010 0.41 ±0.07 NrdA∆147 0.057 ±0.015 0.79 ±0.43 a All protein concentrations are given for monomers; KD values were calculated using the equation published by Climent & al. (20). b Concentrations of NrdA proteins were 0.10 µM when ATP (5 mM) was used as allosteric effector and 0.6 µM when dTTP (0.1 mM) was used. 13
FIGURE LEGENDS Figure 1. Schematic alignment of P. aeruginosa and E. coli NrdA proteins. A, Wild type and mutated P. aeruginosa NrdA with the E. coli NrdA protein. Nucleotides binding to the allosteric sites in E. coli NrdA (4) are shown. The cysteines corresponding to those in the active site of E. coli (C225, C439 and C462) as well as the C-terminal cysteines involved in transthiolation are shown. In addition, a Pfam domain architecture (25) is represented. B, Sequence of the E. coli allosteric activity site (ATP-cone) and the two activity sites sequences in P. aeruginosa (ATP-c1 and ATP-c2). Protein fragments were aligned using ClustalX as previously described (16). Black arrows indicate E. coli residues at the activity site involved in nucleotide binding (10). Black shadow denotes identity in all three sequences, dark grey similarity in the three sequences and light grey identity in only two sequences. Figure 2. SDS-PAGE analysis of final purified Nrd proteins. The proteins were run on a 10% polyacrylamide gel and Coomassie-stained after the run. 4 µg of protein for each sample were loaded onto the gel. Lane 1, NrdA wild type; lane 2, NrdAΔ147; lane 3, NrdAΔ182 and lane 4, NrdB protein. Figure 3. Tyrosyl radical decay and iron content in P. aeruginosa NrdB as a function of time. A 20 µM NrdB monomer solution was reconstituted in the cuvette at pH 7.6 by addition of 30 µM Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Fe2+, and spectra were recorded at 90 s intervals. The top spectrum in the inset shows that a radical concentration of 4.7 µM was reached 40 s after addition of iron. The arrows indicate iron charge transfer bands and tyrosyl radical absorption. The lowest spectrum (filled triangles in inset) was registered 1 hour and 54 minutes after iron addition. The radical decay (open circles) was determined after subtraction of a spectrum at 1600 s after iron addition. Filled squares show iron bound to NrdB at different time points during radical decay. Figure 4. EPR spectra at 93 K. a) Purified reconstituted NrdB from P. aeruginosa frozen 4 s after addition of 1 Fe2+/polypeptide, 37 µM radical, b) NrdB radical from P. aeruginosa in cell suspension of overproducing E. coli cells was determined to 5 µM from subtraction of a fraction of the spectrum in a) leaving a background signal of 4 µM; 9 scans, spectrum divided by 2 for layout reasons, c) NrdB tyrosyl radical from purified E. coli protein normalized to 37 µM radical. Recording parameters: microwave power 1 mW; modulation amplitude 0.2 mT. The hyperfine coupling of the doublets was measured to 1.9 mT in a) and 2.1 mT in c). Figure 5. Stimulation of the CDP reduction by ATP and dATP. Pure proteins, 30 nmoles NrdA () or NrdA∆147 () and 60 nmoles NrdB were assayed under standard conditions except for the concentrations of ATP (in A) or dATP (in B) indicated on the abscissa. In B. the maximum activity for the NrdA wild type was 130 mU and for the NrdA∆147 190 mU. Figure 6. Effects of dATP in absence and presence of dTTP. A, binding assays with dATP. B, competition assays with dTTP (1 mM) and increasing concentrations of dATP. Binding and activity assays were under standard conditions using 0.6 µM of wild type NrdA () or NrdA∆147 () protein. Figure 7. Unrooted phylogenetic tree of representative NrdA proteins. NrdA sequences with duplication or triplication in the N-terminal region are marked in dark grey shadow. The β− and γ-proteobacteria without domain duplication in the NrdA protein are shown in light grey. All sequences were from the Ribonucleotide Reductase Database (38). The alignment and the tree were performed in the same way as previously described (16). Only bootstrap values below 950 are shown. 14
Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Figure 1
Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Figure 2
Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Figure 3
Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Figure 4
Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Figure 5
Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Figure 6
Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Figure 7
Enzyme Catalysis and Regulation: Ribonucleotide reductase modularity: atypical duplication of the ATP-cone domain in Pseudomonas aeruginosa Eduard Torrents, MariAnn Westman, Margareta Sahlin and Britt-Marie Sjöberg J. Biol. Chem. published online July 7, 2006 Access the most updated version of this article at doi: 10.1074/jbc.M601794200 Downloaded from http://www.jbc.org/ by guest on March 3, 2015 Find articles, minireviews, Reflections and Classics on similar topics on the JBC Affinity Sites. Alerts: • When this article is cited • When a correction for this article is posted Click here to choose from all of JBC's e-mail alerts Supplemental material: http://www.jbc.org/content/suppl/2006/07/10/M601794200.DC1.html This article cites 0 references, 0 of which can be accessed free at http://www.jbc.org/content/early/2006/07/07/jbc.M601794200.citation.full.html#ref-list-1
You can also read