ADPRibase-Mn (Mn2+-dependent ADP-ribose/CDP-alcohol pyrophosphatase) was earlier isolated from rat liver supernatants after separation from ADPRibase-I and ADPRibase-II (Mg2+-activated ADP-ribose pyrophosphatases devoid of CDP-alcohol pyrophosphatase activity). The last mentioned are putative Nudix hydrolases, whereas the molecular identity of ADPRibase-Mn is unknown. MALDI (matrix-assisted laser-desorption ionization) MS data from rat ADPRibase-Mn pointed to a hypothetical protein that was cloned and expressed and showed the expected specificity. It is encoded by the RGD1309906 rat gene, which so far has been annotated simply as ‘hydrolase’. ADPRibase-Mn is not a Nudix hydrolase, but it shows the sequence and structural features typical of the metallophosphoesterase superfamily. It may constitute a protein family of its own, the members of which appear to be specific to vertebrates, plants and algae. ADP-ribose was successfully docked to a model of rat ADPRibase-Mn, revealing its putative active centre. Microarray data from the GEO (Gene Expression Omnibus) database indicated that the mouse gene 2310004I24Rik, an orthologue of RGD1309906, is preferentially expressed in immune cells. This was confirmed by Northern-blot and activity assay of ADPRibase-Mn in rat tissues. A possible role of ADPRibase-Mn in immune cell signalling is suggested by the second-messenger role of ADP-ribose, which activates TRPM2 (transient receptor potential melastatin channel-2) ion channels as a mediator of oxidative/nitrosative stress, and by the signalling function assigned to many of the microarray profile neighbours of 2310004I24Rik. Furthermore, the influence of ADPRibase-Mn on the CDP-choline or CDP-ethanolamine pathways of phospholipid biosynthesis cannot be discounted.

INTRODUCTION

ADP-ribose is part of the NAD(P)+-derived network of regulators. It is a primary product of NAD+, formed by hydrolysis of the N-glycosidic linkage between nicotinamide and ribose, and also a secondary product formed by turnover of NAD primary metabolites [1,2]. Its reducing ribose moiety is prone to react non-enzymically, which explains its protein-glycating character and its reputation of being cytotoxic [35]. On the other hand, intracellular ADP-ribose is also a regulator of ion channels. Very interesting is the role of ADP-ribose as an activator of the TRPM2 (transient receptor potential melastatin channel-2) that participates in Ca2+-mediated cell death. The channel opens in response to ADP-ribose as a mediator of oxidative/nitrosative stress in the immune system – a major site for TRPM2 expression [611].

ADP-ribose levels in mammals are probably controlled by a low-Km Mg2+-dependent specific ADP-ribose pyrophosphatase [6,1215], that belongs to the Nudix superfamily of NDP-X hydrolases [16,17]. However, there are other mammalian ADP-ribose hydrolases with diverse preferences for NDP-X substrates and for Mg2+ or Mn2+ as activators. In rat liver, ADP-ribose can be hydrolysed by the broad-specificity nucleotide pyrophosphatase or phosphodiesterase I ectoenzymes [18] and four intracellular hydrolases [19,20]. ADPRibase-I (ADP-ribose pyrophosphatase 1; cytosolic) and ADPRibase-m (mitochondrial) are specific ADP-ribose pyrophosphatases that hydrolyse only ADP-ribose and non-physiological IDP-ribose in the presence of Mg2+. Their specificity is less strict in the presence of Mn2+, but Mg2+ is likely to be the physiological activator. ADPRibase-II (cytosolic) is an Mg2+-dependent ADP-sugar pyrophosphatase that is also less strict in its NDP-X specificity when assayed with Mn2+. Finally, ADPRibase-Mn is not activated by Mg2+, but by low-micromolar concentrations of Mn2+. Its specificity is unusual, as it hydrolyses ADP-ribose, IDP-ribose, CDP-glycerol, CDP-choline and CDP-ethanolamine, but not other non-reducing ADP-sugars or CDP-glucose [19].

None of these rat ADP-ribose hydrolases has been cloned, but ADPRibase-I and ADPRibase-m are nearly identical with human Nudix enzymes NUDT9β (NUDT≡Nudix) and NUDT9α [6,1315,1921], both of which are encoded by the NUDT9 gene, whereas ADPRibase-II is related to human NUDT5 [21,22], which is also a Nudix enzyme encoded by the NUDT5 gene. In contrast, ADPRibase-Mn has not been molecularly identified. In the present study we show that it forms a novel metallophosphoesterase family preferentially expressed in rodent immune tissues and cells.

EXPERIMENTAL

Chemicals and biochemicals

1,10-Phenanthroline monohydrate and the sodium salts of ADP-ribose, ADP-glucose, CDP-choline, CDP-ethanolamine, CDP-glycerol, CDP-glucose, UDP-glucose, CDP, AMP and CMP were from Sigma. ADP (potassium salt), Tris base and alkaline phosphatase (grade I from calf intestine) were from Roche. MnCl2, MgCl2, KCl, NaCl, and EDTA (Titriplex™ III) were high-purity preparations from Merck.

Purification of rat liver ADPRibase-Mn and MALDI (matrix-assisted laser-desorption ionization) MS of trypsin digests

Rat ADPRibase-Mn was purified from 100000 g liver supernatants by a six-step chromatographic procedure. The first five steps were as previously described [19] and, for the final step, the enzyme obtained from 30 g of liver was adsorbed to a 2.5-cminternal-diameter×1.5-cm-long Reactive Red 120–agarose column (Sigma) equilibrated in 20 mM Tris/HCl (pH 7.5)/0.5 mM EDTA, washed with 5 ml of 0.5 M KCl in the same buffer, followed by 10 ml of equilibrating buffer. The enzyme (3.5 ml) was recovered with a wash with water and concentrated to 0.1 ml in a centrifugal filter with a 3000-Da-cut-off poly(ether sulfone) membrane (Filtron Technology Corp, Northborough, MA, U.S.A.). This preparation, on SDS/PAGE, showed a single protein band of ≈38 kDa. After staining for 1 h with Coomassie Blue G-250 [1 g/litre in 50% (v/v) methanol] and destaining in 40% methanol, the band was cut out and processed for trypsin digestion and MALDI MS as described [23] [Servicio de Proteómica, UAM (Universidad Autónoma de Madrid), Madrid, Spain]. The peptide-mass fingerprint (MS mode) and the ion spectra (tandem MS mode) of two fragments of 1746.88 and 2126.98 Da were used for database querying.

PCR cloning and sequencing of rat ADPRibase-Mn cDNA

The 1014-nt ORF (open reading frame) coding for rat ADPRibase-Mn (mRNA accession number BC088174) was amplified by PCR from a rat liver Marathon Ready cDNA preparation with the Advantage™ cDNA polymerase mix (both from Clontech). The forward primer, GTCTGGAATTCCATGGCTGATAAACCGGAC, contained an EcoRI site (underlined), followed by C and the first 18 nucleotides of the ORF, with a designed G12A substitution; the reverse primer, GCTCCGTCGACTTATTCCCTTTTGTAATTC, contained an SalI site (underlined), followed by the reverse complement of the last 19 nucleotides of the ORF. A product of the expected size was obtained, cut with EcoRI and SalI, and inserted into the corresponding sites of the pGEX-6P-3 plasmid (Amersham) to obtain plasmid pGEX-6P-3-rADPRMn. Both strands of the insert were sequenced in the Servicio de Secuenciación Automática [IIB, CSIC (Consejo Superior de Investigaciones Científicas)–UAM, Madrid, Spain].

Recombinant ADPRibase-Mn expression and purification

Cultures of BL21 cells (100 ml) transformed with pGEX-6P-3-rADPRMn, were grown at 30 °C and induced with isopropyl β-D-thiogalactoside. The GST (glutathione transferase)–ADPRibase-Mn fusion recovered in post-sonication supernatants [in 10 ml of a solution containing 20 mM Tris/HCl, pH 7.5, 0.5 mM EDTA, 1 mM DTT (dithiothreitol), 50 mM KCl and one dissolved tablet of protease-inhibitor cocktail (Complete™ Mini; Roche)] was adsorbed on to a glutathione–Sepharose column (1.5 ml; Amersham) equilibrated in the same buffer without protease inhibitors. The unbound proteins were washed out with 15 ml of a solution containing 50 mM Tris/HCl, pH 7.5, 1 mM EDTA, 1 mM DTT and 150 mM NaCl. ADPRibase-Mn was separated from the GST tag by in-column proteolysis with PreScission™ protease, followed by elution with wash buffer. The eluted enzyme (1.8 ml) was frozen at −80 °C in 50 μl aliquots. This was a high-purity enzyme preparation (see Supplementary Figure S1, lane 6, at http://www.BiochemJ.org/bj/413/bj4130103add.htm). Upon thawing, the enzyme was stable at least for 1 week. The expression and purification procedure was carried out three times with the same results.

Determination of the native molecular mass of ADPRibase-Mn by gel-filtration chromatography

Samples of ADPRibase-Mn and molecular-mass markers were chromatographed in a Protein-Pak 125 column (7.8 mm internal diameter×300 mm long) from Waters, equilibrated in 50 mM Tris (pH 7.5)/1 mM EDTA/150 mM NaCl, and connected to a Hewlett–Packard HP1090 chromatograph. The elutions were performed with equilibrating buffer at a flow rate of 1 ml/min, and the chromatographic profiles were recorded at 280 nm. For ADPRibase-Mn activity assay, 0.25 ml fractions were collected.

Enzyme activity assays

The hydrolytic activities of ADPRibase-Mn on NDP-X substrates were assayed at 37 °C by coupling the enzyme to alkaline phosphatase and measuring colorimetrically the amount of Pi formed with an SDS/ascorbate/molybdate reagent [21]. The assay mixtures contained, except when otherwise stated, 50 mM Tris-HCl (pH 7.5 at 25 °C), 500 μM NDP-X, 100 μM MnCl2, 1.5 μg/ml (5.5 units/ml) alkaline phosphatase and the appropriate amount of enzyme sample. For the assay of the chromatographic profiles of ADPRibase-I, ADPRibase-Mn and ADPRibase-II in fractionated rat tissue extracts, 1 mg/ml BSA was included in the reaction mixtures. Whenever the substrate was NDP or NMP, alkaline phosphatase was omitted. For the pH–activity profiles, different buffers were used, MnCl2 was increased to 1 mM, and the concentration of alkaline phosphatase was doubled. In every case, the reaction was stopped and colour developed by addition of either 0.7 ml of standard Pi reagent to 0.05–0.1 ml or, when higher sensitivity was needed, 0.5 ml of a concentrated reagent to 0.4 ml of reaction mixture [21]. For studies of enzyme saturation, initial rates were measured at diverse substrate concentrations and the data were adjusted by non-linear regression to the Michaelis–Menten equation [24]. For inactivation studies with o-phenanthroline, the activity was assayed in standard reaction mixtures without alkaline phosphatase, by measuring the conversion of ADP-ribose into AMP by HPLC after 15 min of reaction. Enzyme activities were linear with time and amount of enzyme. Blanks without enzyme and/or substrate were run in parallel. One unit enzyme activity is defined as 1 μmol of substrate hydrolysed/min.

HPLC detection of NMP as reaction product of ADPRibase-Mn

Reaction mixtures with all the substrates hydrolysed by ADPRibase-Mn were analysed by ion-pair reverse-phase HPLC in a 15 cm×0.4 cm octadecylsilica column (Kromasil 100; Teknokroma, Sant Cugat del Vallès, Barcelona, Spain) with a 1 cm×0.4 cm pre-column of the same material, using a Hewlett–Packard HP1100 chromatograph. Two elution buffers were used: buffer A comprised 5 mM sodium phosphate, pH 7.0, 20 mM tetrabutylammonium and 10% (v/v) methanol; buffer B was the same as buffer A, but contained 100 mM sodium phosphate. Before each run, the column was equilibrated in buffer A. Samples (20 μl) of reaction mixture were injected into the chromatograph, and the elution was accomplished at a 1 ml/min flow rate with a linear 10 min 5–100 mM phosphate gradient, followed by a 10 min isocratic wash with buffer B. Eluates were monitored at 260 nm. The retention times (min) of compounds relevant to the reactions catalysed by ADPRibase-Mn were as follows: adenosine, 2.6; AMP, 6.6; ADP, 11.7; ADP-ribose, 8.5; cytidine, 1.5; CMP, 4.4; CDP, 7.7; CDP-ethanolamine, 2.3; CDP-choline, 2.2; CDP-glycerol, 5.0. This allowed the identification of either AMP or CMP as the nucleotide product of hydrolysis for every substrate of ADPRibase-Mn.

Multiple-tissue Northern blot

A multiple-rat-tissue Northern blot, containing 2 μg of mRNA from each tissue, prestained with Blot Stain Blue (see Supplementary Figure S2 at http://www.BiochemJ.org/bj/413/bj4130103add.htm; Sigma), was used. The blot was probed with a digoxigenin-labelled cDNA containing the full ORF of rat ADPRibase-Mn. The labelling was performed with the DIG High Prime DNA Labeling kit (Roche) according to the manufacturer's instructions. Hybridization and detection were performed by the optimized protocol described elsewhere [25], except that the probe concentration was 10 ng of cDNA/ml, the hybridization steps were carried out at 55 °C, and ready-to-use CSPD™ [disodium 3-(4-methoxyspiro{1,2,-dioxyetane-3,2-(5-chloro)tricyclo-[3.3.1.1.]decan}-4-yl)phenyl phosphate] (Roche) was the substrate for the chemiluminescent reaction catalysed by alkaline-phosphatase-labelled anti-digoxigenin (Roche). The results were recorded, first, by quantitative luminescence detection (Kodak Image Station 2000R) under conditions in which signal accumulation was linear with time and with the amount of labelled probe and, secondly, by exposure of X-ray film to the chemiluminescent blot.

Rat tissue or cell extract fractionation to assay ADPRibase-I, ADPRibase-Mn and ADPRibase-II

Female Wistar rats weighing 200–220 g, fed ad libitum, were provided by the Servicio de Animalario, Universidad de Extremadura, Badajoz, Spain. The animals were killed by rapid decapitation, and the organs of interest (liver, skeletal muscle, spleen and thymus) were immediately dissected. After washing in cold saline (0.9% NaCl), the tissue was blotted dry and weighed. Depending on the organ, 0.35–1.3 g of tissue was processed.

The protocol to obtain rat splenocytes and to remove red blood cells was as described in [26], except that spleens were crushed in 10 ml of cold saline. The red-cell-lysing buffer was purchased from Roche, and the final washes of splenocytes were performed with 14.5 mM Tris, pH 7.6, 5 μM CaCl2, 98 μM MgCl2, 0.54 mM KCl, 126 mM NaCl and 0.1 g/l glucose.

The separation of ADPRibase-I, ADPRibase-Mn and ADPRibase-II was carried out, with some simplification, as described in [19]. The tissue samples were homogenized on ice, in buffer containing 50 mM Tris, pH 7.5, 0.5 mM EDTA, 1 mM PMSF, using 6 ml/g of tissue (liver and muscle), 9 ml/g of tissue (spleen and thymus) or 4.5 ml per washed splenocyte pellet obtained from one spleen. A hand-driven Potter-type homogenizer was used, except in the case of muscle, which was homogenized first in a motor-driven blender and then finished in the Potter-type homogenizer. For every tissue, 4–5 ml of homogenate was centrifuged for 60 min at 4 °C and 130000 g in a Beckman SW55 rotor. The supernatant was chromatographed in a Sephadex G-100 column (1 cm internal diameter×90 cm long) equilibrated and eluted with 20 mM Tris (pH 7.5)/0.5 mM EDTA at 20 ml/h. Fractions of volume about 2.5 ml each were collected and the Mg2+- and the Mn2+-dependent ADP-ribose hydrolase activities were assayed. A small peak of activity eluted in the void volume was discarded, and the fractions of a large included peak of ADP-ribose hydrolase, active with both Mg2+ or Mn2+, were pooled. This peak contained ADPRibase-I, ADPRibase-Mn and ADPRibase-II, and it was applied to a DEAE-cellulose column (1.5 cm internal diameter×12 cm long) equilibrated in 20 mM Tris, pH 7.5, and washed with 2–3 vol. of 20 mM Tris (pH 7.5)/0.5 mM EDTA/50 mM KCl, followed by a 130 ml linear gradient of 50–400 mM KCl in the same buffer. The three enzymes were resolved within the gradient, allowing their individual quantification. A careful accounting was kept such that activities quantified after the DEAE-cellulose step could be extrapolated back and referred to total tissue protein measured in the homogenate with BSA as a standard [27].

Docking of ADP-ribose to the homology model of rat ADPRibase-Mn

The theoretical co-ordinates of rat ADPRibase-Mn were taken from the SwissProt Repository [28]. This homology model is based on the crystal structure of the zebrafish (Danio rerio) protein [PDB (Protein Data Base) entry 2NXF], which contains two Zn2+ ions in a typical metallophosphoesterase dinuclear centre. Since there was an excellent structural similarity between the zebrafish and the rat proteins concerning the residues located within 5 Å (1 Å=0.1 nm) of the Zn2+ ions [rmsd (root mean square deviation) 0.082 Å], the metal co-ordinates of the crystal structure were used to place them in the homology model. The protonation state of histidine side chains was assigned with the program Reduce [29]. Atom types and Kollman partial charges were assigned with ADT (AutoDockTools [28a]). The zinc ions were modelled with charge 2+ and default AutoDock 4 radius and well depth. ADP-ribose was prepared with Marvin 4.1.13 Java-based chemistry software (updated 2007; ChemAxon, Budapest, Hungary; http://www.chemaxon.com), its Gasteiger charges were calculated with ADT (total charge: 2–), and all its torsionable groups were made flexible except 6-NH2, and 2′-, 3′-, 2″- and 3″-OH. Docking of ADP-ribose was run with AutoDock 4 [30] within a scoring grid (sized 33.8×29.2×29.2 Å with a resolution of 0.375 Å) centred around the putative active site. AutoDock was run with the Lamarckian genetic algorithm with default parameters, except that 100 independent runs were performed, the maximum number of energy evaluations per run was set to 5×107, and the step sizes were adjusted to 0.2 Å for translations and 5° for orientations and torsions.

RESULTS

Tandem MS of rat liver ADPRibase-Mn pointed to a hypothetical rat protein that was cloned and expressed

For identification, a peptide-mass fingerprint of the 38 kDa band of rat liver ADPRibase-Mn was obtained. In a Mascot search [31] against the NCBInr (National Center for Biotechnology Information non-redundant) database, it did not give any significant candidate. Therefore, two peptides of 1746.88 and 2126.98 Da were analysed by tandem MS. Searches with these data pointed to two tryptic peptides formed by amino acids 47–61 (HSLVHLQGAIEDWNK) and 102–118 (VPVHHTWGNHEFYNFSR) of a 337-amino-acid hypothetical rat protein [NCBI (National Center for Biotechnology Information) accession no. AAH88174]. Its theoretical mass of 38710 Da accorded with the 38 kDa protein band. It was recorded as a product of the rat gene RGD1309906 (GeneID 287406), which had been annotated simply as ‘hydrolase’. The coding sequence (NCBI accession no. BC088174) of the hypothetical protein was used to design the PCR cloning described in the Experimental section.

Plasmid pGEX-6P-3-rADPRMn contained, in-frame with a GST affinity label, the PCR-amplified 1014-nt ORF coding for ADPRibase-Mn. Relative to the ORF used as reference, it displayed three nucleotide changes: G12A (due to the design of the forward PCR primer), A131G and A636G, only one causing an amino acid change (Y44C). This sequence was deposited with GenBank®.

The GST–ADPRibase-Mn fusion protein of theoretically 65080 Da was expressed from pGEX-6P-3-rADPRMn as an approx. 65 kDa protein (Supplementary Figure S1, lane 2), and it was adsorbed to glutathione–Sepharose. The ADPRibase-Mn moiety was recovered by in-column proteolysis as an approx. 40 kDa protein (Supplementary Figure S1, lane 6). This agrees with an expected mass of 39 360 Da – the sum of the mass of ADPRibase-Mn and that of the N-terminal extension (GPLGSPNS) left by proteolysis.

Characterization of the recombinant protein as a Mn2+-dependent ADP-ribose/CDP-alcohol pyrophosphatase identical with ADPRibase-Mn

Like rat liver ADPRibase-Mn [19], the recombinant enzyme hydrolysed ADP-ribose and CDP-alcohols. At 500 μM substrate and 100 μM MnCl2, CDP-choline gave the highest rate, followed by CDP-glycerol, CDP-ethanolamine, ADP-ribose and ADP (Table 1). However, kinetic parameters derived from hyperbolic saturation curves (see Supplementary Figure S3 at http://www.BiochemJ.org/bj/413/bj4130103add.htm) showed ADP-ribose as the best substrate (kcat/Km=328000 M−1·s−1), whereas CDP-alcohols and ADP were hydrolysed with 3.5–15-fold lower efficiencies (Table 1). ADP-glucose, UDP-glucose, CDP-glucose, CDP, CMP and AMP were not hydrolysed.

Table 1
Substrate specificity and kinetic parameters of ADP-ribose/CDP-alcohol pyrophosphatase
 Enzyme purified from rat liver Enzyme expressed from cloned DNA (the present study) 
Substrate* Activity at 500 μM substrate† (%) Activity at 500 μM substrate‡ (%) kcat§ (s−1Km§ (μM) kcat/Km (M−1·s−1
ADP-ribose 100 100±8 12.8±0.2 39±2 328200 
CDP-choline 140 156±23 28.5±0.6 300±22 95000 
CDP-ethanolamine 89 84±18 41.3±1.1 1422±99 29000 
CDP-glycerol 89 103±10 28.1±1.4 304±59 92400 
ADP 27 26±4 4.4±0.2 201±30 21400 
 Enzyme purified from rat liver Enzyme expressed from cloned DNA (the present study) 
Substrate* Activity at 500 μM substrate† (%) Activity at 500 μM substrate‡ (%) kcat§ (s−1Km§ (μM) kcat/Km (M−1·s−1
ADP-ribose 100 100±8 12.8±0.2 39±2 328200 
CDP-choline 140 156±23 28.5±0.6 300±22 95000 
CDP-ethanolamine 89 84±18 41.3±1.1 1422±99 29000 
CDP-glycerol 89 103±10 28.1±1.4 304±59 92400 
ADP 27 26±4 4.4±0.2 201±30 21400 
*

Neither the purified nor the expressed enzymes displayed significant activity towards ADP-glucose, UDP-glucose, CDP-glucose, CDP, CMP and AMP (<3%).

Taken from [19].

Means±S.D. for three experiments; 100% was 23 μmol/min per mg; protein was assayed as described in [50].

§

Kinetic parameters±S.E. [24] were obtained from the pooled data from two or three experiments.

All the reactions involved the hydrolysis of the phosphoanhydride linkage, since (i) Pi was formed by alkaline phosphatase from the NDP-X reaction products, but not from the substrates, (ii) Pi was directly formed from ADP, and (iii) AMP or CMP were identified as products by HPLC (results not shown).

ADPRibase-Mn was inactive without added Mn2+, and it showed a high affinity for the cation, since saturation was observed at low-micromolar concentrations of Mn2+ well below the substrate concentration (500 μM; Supplementary Figure S3). The concentration of MnCl2 giving the half-maximal rate was 1–4 μM. In no case could Mg2+ replace Mn2+.

Incubation with 2.5 mM EDTA did not affect ADPRibase-Mn activity. This was in agreement with the stability of the enzyme frozen for months in a buffer with 1 mM EDTA. However, o-phenanthroline caused near full inactivation in 1.5 h (Figure 1).

Inactivation of recombinant ADPRibase-Mn by o-phenanthroline

Figure 1
Inactivation of recombinant ADPRibase-Mn by o-phenanthroline

Enzyme samples were diluted 1:1 in 5 mM (●) o-phenanthroline or (○) EDTA, each dissolved in 50 mM Tris/HCl, pH 7.5, and were incubated for the indicated length of time at 37 °C. Aliquots of enzyme were taken at the end of the incubation, diluted 100-fold in 50 mM Tris/HCl, pH 7.5, and immediately used for ADP-ribose hydrolase activity assay by HPLC.

Figure 1
Inactivation of recombinant ADPRibase-Mn by o-phenanthroline

Enzyme samples were diluted 1:1 in 5 mM (●) o-phenanthroline or (○) EDTA, each dissolved in 50 mM Tris/HCl, pH 7.5, and were incubated for the indicated length of time at 37 °C. Aliquots of enzyme were taken at the end of the incubation, diluted 100-fold in 50 mM Tris/HCl, pH 7.5, and immediately used for ADP-ribose hydrolase activity assay by HPLC.

The effect of pH on ADPRibase-Mn activities was studied over the range pH 6–10. The profile of ADP-ribose hydrolase showed an optimum at pH 8.0–8.5, whereas the activities towards ADP and CDP-alcohols were highest at pH 6–8 and decreased abruptly at higher pH (Figure 2).

pH profiles of the activities of recombinant ADPRibase-Mn

Figure 2
pH profiles of the activities of recombinant ADPRibase-Mn

Initial rates of hydrolysis were measured with the indicated substrate in the presence of 1 mM MnCl2. The buffers used were: 100 mM Tris/acetate (pH 6–7), 100 mM Tris/HCl (pH 7–9) and 100 mM Caps [3-(cyclohexylamino)propane-1-sulfonic acid]/NaOH (pH 9-10.1). Reaction mixtures of volume 0.4 ml were prepared and, at the end of the incubation period, 0.1 ml was used for the assay of activity and 0.3 ml for pH measurement with a glass electrode.

Figure 2
pH profiles of the activities of recombinant ADPRibase-Mn

Initial rates of hydrolysis were measured with the indicated substrate in the presence of 1 mM MnCl2. The buffers used were: 100 mM Tris/acetate (pH 6–7), 100 mM Tris/HCl (pH 7–9) and 100 mM Caps [3-(cyclohexylamino)propane-1-sulfonic acid]/NaOH (pH 9-10.1). Reaction mixtures of volume 0.4 ml were prepared and, at the end of the incubation period, 0.1 ml was used for the assay of activity and 0.3 ml for pH measurement with a glass electrode.

ADPRibase-Mn behaved as a monomer, with an apparent mass of ~30 kDa by gel filtration (see Supplementary Figure S4 at http://www.BiochemJ.org/bj/413/bj4130103add.htm), in good agreement with the native mass of 32 kDa observed for the enzyme purified from rat liver [19].

Characterization of ADPRibase-Mn as a protein belonging to the metallophosphoesterase, not to the Nudix, superfamily

In contrast with other ADP-ribose pyrophophatases, ADPRibase-Mn lacked the amino acid signature of the Nudix superfamily. Searches in the InterPro site (http://www.ebi.ac.uk/InterProScan) and Superfamily database (http://supfam.org/SUPERFAMILY) identified a metallophosphoesterase domain (InterPro accession no. IPR004843), indicating that ADPRibase-Mn belongs to the Pfam (protein family database) metallophos family (ID: PF00149) and is a member of the SCOP (Structural Classification of Proteins) metallo-dependent phosphatase superfamily (ID: SSF56300). These proteins contain (1) a dimetal centre with diverse ion pairs or combinations, (2) a βαβαβ secondary structure signature within (3) a four-layered fold with two β-sheets flanked by α-helices (α/β/β/α fold) and (4) a disperse sequence signature that includes, in five conserved regions, the amino acids co-ordinated with the metal ions: DX(H/X)XnGDXX(D/X)Xn-GNH(D/E)Xn(G/X)HXnGHX(H/X) [3235].

The alignment of ADPRibase-Mn orthologues found by BlastP searches (Table 2) revealed these five regions [Figure 3, (I)–(V)].

Table 2
ADPRibase-Mn orthologues in representative vertebrates and plants, and absence of orthologues in invertebrates and fungi

BlastP searches were run in the NCBI site using the conceptual translation of the ORF cloned from rat liver cDNA (GenBank® accession no. EU037900) as the query against the NCBInr protein database. Only representative hits are shown for vertebrates and plants. When the BlastP program failed to find a putative orthologous protein, TBlastN searches were run against the NCBInr nucleotide database and the assembled genomes of relevant taxonomic groups. The results shown were obtained on 8 September 2007.

 Orthologous protein  
Species or group of organisms Name BlastP E-value Identity (%) Similarity (%) Query coverage (%) No. of amino acids Accession no. Gene ID 
H. sapiens Hypothetical protein LOC56985 6×10−176 85 93 100 342 NP_064618 56985 
M. musculus 2310004I24Rik protein <10−256 95 97 100 340 AAH04029 66358 
Gallus gallus (chicken) Similar to putative protein product of Nbla03831 7×10−30 52 65 38 156 XP_001231601 768578 
X. laevis x 006 protein 2×10−100 53 71 95 356 NP_001080428 380120 
D. rerio Hypothetical protein LOC393393 6×10−84 50 68 94 322 NP_956715 393393 
Insects − >0.1* − − − − − − 
Nematodes − >0.1† − − − − − − 
Echinoderms − >0.1‡ − − − − − − 
Fungi − >0.1§ − − − − − − 
O. sativa Os07g0688000 2×10−63 41 59 96 321 NP_001060698 4344352 
 Os08g0557200 3×10−58 40 54 94 339 NP_001062482 4346295 
A. thaliana Calcineurin-like phosphoesterase family protein 6×10−52 37 57 92 311 NP_194204 828575 
O. lucimarinus Predicted protein 7×10−32 31 51 93 331 XP_001419532 5003684 
 Orthologous protein  
Species or group of organisms Name BlastP E-value Identity (%) Similarity (%) Query coverage (%) No. of amino acids Accession no. Gene ID 
H. sapiens Hypothetical protein LOC56985 6×10−176 85 93 100 342 NP_064618 56985 
M. musculus 2310004I24Rik protein <10−256 95 97 100 340 AAH04029 66358 
Gallus gallus (chicken) Similar to putative protein product of Nbla03831 7×10−30 52 65 38 156 XP_001231601 768578 
X. laevis x 006 protein 2×10−100 53 71 95 356 NP_001080428 380120 
D. rerio Hypothetical protein LOC393393 6×10−84 50 68 94 322 NP_956715 393393 
Insects − >0.1* − − − − − − 
Nematodes − >0.1† − − − − − − 
Echinoderms − >0.1‡ − − − − − − 
Fungi − >0.1§ − − − − − − 
O. sativa Os07g0688000 2×10−63 41 59 96 321 NP_001060698 4344352 
 Os08g0557200 3×10−58 40 54 94 339 NP_001062482 4346295 
A. thaliana Calcineurin-like phosphoesterase family protein 6×10−52 37 57 92 311 NP_194204 828575 
O. lucimarinus Predicted protein 7×10−32 31 51 93 331 XP_001419532 5003684 
*

Best TBlastN E-value=0.004.

Best TBlastN E-value=0.35.

Best TBlastN E-value=4.3.

§

Best TBlastN E-value=0.033.

Protein alignment of rat ADPRibase-Mn orthologues showing the conservation of a metallophosphoesterase sequence signature

Figure 3
Protein alignment of rat ADPRibase-Mn orthologues showing the conservation of a metallophosphoesterase sequence signature

The sequences were aligned using the Clustal-W program. Orthologous proteins (with Genbank accession numbers): Rnor, Rattus norvegicus (conceptual translation of EU037900); Hsap, Homo sapiens (human; NP_064618); Mmus, Mus musculus (mouse; AAH04029); Xlae, Xenopus laevis (South African clawed frog; NP_001080428); Drer, Danio rerio (NP_956715); Atha, Arabidopsis thaliana (thale cress; NP_194204); Osat1 and Osat2, Oryza sativa (rice; NP_001060698 and NP_001062482); Oluc, Ostreococcus lucimarinus (a marine phytoplanktonic green alga; XP_001419532). The asterisks and the dots below the sequences mark residues conserved as identities or similarities respectively. The white letters on dark background represent conserved amino acids that are part of the metallophosphoesterase sequence signature divided in five regions (I–V). This signature is shown in the text. The vertical marks at the top of the sequences are positioned at intervals of ten amino acids relative to the Rattus norvegicus sequence.

Figure 3
Protein alignment of rat ADPRibase-Mn orthologues showing the conservation of a metallophosphoesterase sequence signature

The sequences were aligned using the Clustal-W program. Orthologous proteins (with Genbank accession numbers): Rnor, Rattus norvegicus (conceptual translation of EU037900); Hsap, Homo sapiens (human; NP_064618); Mmus, Mus musculus (mouse; AAH04029); Xlae, Xenopus laevis (South African clawed frog; NP_001080428); Drer, Danio rerio (NP_956715); Atha, Arabidopsis thaliana (thale cress; NP_194204); Osat1 and Osat2, Oryza sativa (rice; NP_001060698 and NP_001062482); Oluc, Ostreococcus lucimarinus (a marine phytoplanktonic green alga; XP_001419532). The asterisks and the dots below the sequences mark residues conserved as identities or similarities respectively. The white letters on dark background represent conserved amino acids that are part of the metallophosphoesterase sequence signature divided in five regions (I–V). This signature is shown in the text. The vertical marks at the top of the sequences are positioned at intervals of ten amino acids relative to the Rattus norvegicus sequence.

The crystal structure of the zebrafish orthologue complexed with Pi and two Zn2+ ions [36], and a homology model of the rat protein [28], were found in public databases. They display the α/β/β/α metallophosphoesterase fold, with the five regions of the sequence signature brought together each at the C-end of a different β-strand: regions I–III belong to one of the β-sheets, and regions IV–V to the other. Seven of the conserved amino acids, which show an excellent structural similarity between the zebrafish and the rat proteins, form a dimetal centre located at the bottom of a groove. ADP-ribose was succesfully docked to the rat protein model into this site, indicating that it could be the active centre (Figures 4A and 4B). The amino acids that form the dimetallic centre and/or might interact with the substrate are shown in Figure 4(C).

Theoretical model of rat ADPRibase-Mn with docked ADP-ribose

Figure 4
Theoretical model of rat ADPRibase-Mn with docked ADP-ribose

The Figure shows different views of the lowest-energy pose found after docking as described in the Experimental section. The two Zn2+ ions are shown as yellow spheres and ADP-ribose is shown in blue (adenine), red (pyrophosphate) and green (ribose-1 and ribose-2, respectively proximal and distal to the adenine moiety). Oxygen atoms of the β-phosphate group interact with Zn1 and Zn2. (A) Full homology model showing the α/β/β/α fold, typical of metallophosphoesterases, and the docked ADP-ribose. (B) Molecular surface of the active centre with ADP-ribose, as viewed from above in (A). (C) Stereo view of a selection of amino acids that form the dimetallic centre and/or might interact with the substrate, as viewed from the right-hand side in (A). Possible interactions of amino acid side chains in (C) are as follows: (i) co-ordination with Zn1: Asp25, Gln27, Asp74 and His280; (ii) co-ordination with Zn2: Asp74, Asn110, His241 and His278; (iii) H-bridge with O–Pβ: Gln27, His278 and His280; (iv) H-bridge with O–Pα: Arg43 and His111; (v) H-bridge with the phosphoanhydride O atom: Asn110 and His111; (vi) hydrophobic contact with the adenine rings: Leu196; (vii) contacts with ribose-1 or ribose-2 rings: Phe210 and Cys253 respectively. All the abovementioned amino acids are sequentially conserved (see Figure 3) and superimposable on those of the zebrafish protein (not shown), except Leu196, which is conserved in, but not superimposable on, both proteins.

Figure 4
Theoretical model of rat ADPRibase-Mn with docked ADP-ribose

The Figure shows different views of the lowest-energy pose found after docking as described in the Experimental section. The two Zn2+ ions are shown as yellow spheres and ADP-ribose is shown in blue (adenine), red (pyrophosphate) and green (ribose-1 and ribose-2, respectively proximal and distal to the adenine moiety). Oxygen atoms of the β-phosphate group interact with Zn1 and Zn2. (A) Full homology model showing the α/β/β/α fold, typical of metallophosphoesterases, and the docked ADP-ribose. (B) Molecular surface of the active centre with ADP-ribose, as viewed from above in (A). (C) Stereo view of a selection of amino acids that form the dimetallic centre and/or might interact with the substrate, as viewed from the right-hand side in (A). Possible interactions of amino acid side chains in (C) are as follows: (i) co-ordination with Zn1: Asp25, Gln27, Asp74 and His280; (ii) co-ordination with Zn2: Asp74, Asn110, His241 and His278; (iii) H-bridge with O–Pβ: Gln27, His278 and His280; (iv) H-bridge with O–Pα: Arg43 and His111; (v) H-bridge with the phosphoanhydride O atom: Asn110 and His111; (vi) hydrophobic contact with the adenine rings: Leu196; (vii) contacts with ribose-1 or ribose-2 rings: Phe210 and Cys253 respectively. All the abovementioned amino acids are sequentially conserved (see Figure 3) and superimposable on those of the zebrafish protein (not shown), except Leu196, which is conserved in, but not superimposable on, both proteins.

Gene encoding ADPRibase-Mn in Rattus norvegicus (brown or Norwegian rat) and mammalian orthologues

A BlastN query to the R. norvegicus genome using the ORF sequence of rat ADPRibase-Mn revealed that the gene RGD1309906 (Gene ID 287406; also known as MDS006 and MGC108803) is the only significantly related locus. It is a predicted gene located in rat chromosome 10 (10q24) that contains five exons, of which only the three last ones are translated and were hit by the BlastN search. The ADPRibase-Mn ORF spans from the 5′-end of exon 3 to within exon 5 (Figure 5). The reference rat genome contains the coding sequence of hypothetical protein LOC287406 (NCBI accession no. BC088174), identical with that coding for recombinant ADPRibase-Mn, except for the nucleotide differences mentioned above. Figure 5 also shows the organization of the mouse and human genes and the ORFs coding for ADPRibase-Mn orthologues in these organisms. TBlastN and BlastN searches pointed to an inverse 427 bp repetition, including most of exon 5 and located 130 kb downstream of the rat gene, a feature not observed in the mouse or human genomes.

Rat ADPRibase-Mn gene and orthologues in mouse and human genomes

Figure 5
Rat ADPRibase-Mn gene and orthologues in mouse and human genomes

The exon (black boxes)–intron (thin lines) structures are shown as they currently appear in the NCBI Gene database, including the ORFs coding for rat ADPRibase-Mn and its mouse and human orthologues after splicing. The exons of the rat gene are currently supported by a 1.3 kb mRNA (BC088174), which agrees with the size of the most abundant transcript seen in a Northern blot (see Figure 6). The white prolongations of the first translated exon show how the mRNAs would remain after operation of the alternative 5′-splice in the next intron (see the text). The alternative splicing would give rise to proteins shorter than ADPRibase-Mn and its orthologues, due to an in-frame stop codon either in the portion of intron retained as exon (mouse and human) or in the next exon (rat).

Figure 5
Rat ADPRibase-Mn gene and orthologues in mouse and human genomes

The exon (black boxes)–intron (thin lines) structures are shown as they currently appear in the NCBI Gene database, including the ORFs coding for rat ADPRibase-Mn and its mouse and human orthologues after splicing. The exons of the rat gene are currently supported by a 1.3 kb mRNA (BC088174), which agrees with the size of the most abundant transcript seen in a Northern blot (see Figure 6). The white prolongations of the first translated exon show how the mRNAs would remain after operation of the alternative 5′-splice in the next intron (see the text). The alternative splicing would give rise to proteins shorter than ADPRibase-Mn and its orthologues, due to an in-frame stop codon either in the portion of intron retained as exon (mouse and human) or in the next exon (rat).

In human, mouse and rat, in addition to ADPRibase-Mn orthologues of 337–342 amino acids (Figure 3), BlastP searches found shorter protein variants attributable to an alternative 5′-splice site of the intron following the first translated exon. They were either conceptual translations from alternatively spliced transcripts (e.g. accession nos. BC001294 and BC070155 for human and AK009142 for mouse mRNAs) or genomic predictions (accession nos. EDL10423 for mouse and EDM04778 for rat proteins). The alternative 5′-splice leaves 86 nt of the intron 5′-end attached to the preceding exon after splicing (Figure 5). The alternative transcription pattern contains an in-frame stop codon either within the retained intron segment (human and mouse) or in the next exon (rat). At the present time, some databases point to the shorter version of the protein as the gene product.

Restricted taxonomic distribution of ADPRibase-Mn orthologues

Orthologous genes of RGD1309906 are recorded only among vertebrates, higher plants and algae. As this seemed an important phylogenetic feature, we ran BlastP and TBlastN searches that found ADPRibase-Mn orthologous in those phyla, but not in invertebrates and fungi (Table 2). In the latter cases, TBlastN searches of complete genomes confirmed that the absence of orthologous proteins was not due simply to their not having been recorded yet.

ADPRibase-Mn gene expression: differential expression in immune and non-immune rodent tissues

Upon inspection of the GEO (Gene Expression Omnibus) database for ADPRibase-Mn-related genes, a genomics-based investigation aimed to identify mouse immune genes came to our attention (GEO dataset GDS2068 [37]). Out of the 8734 genomic features explored in this microarray experiment, 360 showed preferential expression in thymus, spleen, peripheral-blood mononuclear cells, lymph nodes (unstimulated or stimulated) or in vitro activated T-cells. These features were linked to 333 unique genes, one of them being 2310004I24Rik, the orthologue of ADPRibase-Mn. The probe representing this gene in the microarray is a cDNA of the alternatively spliced mRNA (AK009142) that encodes the short protein variant (see above and Figure 5). It is likely to hybridize with both types of transcripts, as they differ only in an 86-nt insert within a length of ~1500 nt. The expression profile of 2310004I24Rik is shown in Supplementary Figure S5 at http://www.BiochemJ.org/bj/413/bj4130103add.htm, together with the heat plot of its profile neighbours (listed in Table 3 with their assigned functions [37]). To confirm the preferential expression of ADPRibase-Mn in rodent immune tissues, the UniGene database was inspected, and a Northern-blot analysis and a fractionation study of the ADP-ribose pyrophosphatases in soluble extracts of selected rat tissues were run.

Table 3
ADPRibase-Mn profile neighbours of expression among the genes preferentially expressed in mouse immune cells and tissues

The list was found by automatically seaching for 2310004I24Rik gene profile neighbours in the GEO dataset GDS2068 within the GEO Profiles database at the NCBI site. Out of the 31 records found, only those classified as ‘immune genes’ are listed, together with their assigned functions [37]. Expression profiles are shown in Supplementary Figure S5.

Gene symbol Full name Assigned function 
Cd53 CD53 antigen Signalling 
Coro1a Coronin, actin-binding protein 1A Signalling 
1200013B08Rik (Sly1) RIKEN cDNA 1200013B08 gene Signalling 
Adcy7 Adenylate cyclase 7 Signalling 
Ptprcap Protein tyrosine phosphatase, receptor type, C polypeptide-associated protein Signalling 
Syk Spleen tyrosine kinase Signalling 
Cd37 CD37 antigen Signalling 
Ptprc Protein tyrosine phosphatase, receptor type, C Signalling 
Il2rg Interleukin 2 receptor, γ chain Signalling 
H2-Eb1 Histocompatibility 2, class II antigen E β Immune 
H2-Oa Histocompatibility 2, O region α locus Immune 
H2-Ob Histocompatibility 2, O region β locus Immune 
Cd79b CD79B antigen Immune 
Ms4a1 Membrane-spanning 4-domains, subfamily A, member 1 Immune 
Ms4a4b Membrane-spanning 4-domains, subfamily A, member 4B Immune 
Ncf4 Neutrophil cytosolic factor 4 Defence 
Sp100 Nuclear antigen Sp100 Cell cycle 
Ifi203 Interferon-activated gene 203 Apoptosis 
Itgb2 Integrin β 2 Adhesion 
Pou2af1 POU* domain, class 2, associating factor 1 Transcription 
Tap1 Transporter 1, ATP-binding cassette, subfamily B (MDR†/TAP) Transport 
Nup210 Nucleoporin 210 Nuclear pore 
Gene symbol Full name Assigned function 
Cd53 CD53 antigen Signalling 
Coro1a Coronin, actin-binding protein 1A Signalling 
1200013B08Rik (Sly1) RIKEN cDNA 1200013B08 gene Signalling 
Adcy7 Adenylate cyclase 7 Signalling 
Ptprcap Protein tyrosine phosphatase, receptor type, C polypeptide-associated protein Signalling 
Syk Spleen tyrosine kinase Signalling 
Cd37 CD37 antigen Signalling 
Ptprc Protein tyrosine phosphatase, receptor type, C Signalling 
Il2rg Interleukin 2 receptor, γ chain Signalling 
H2-Eb1 Histocompatibility 2, class II antigen E β Immune 
H2-Oa Histocompatibility 2, O region α locus Immune 
H2-Ob Histocompatibility 2, O region β locus Immune 
Cd79b CD79B antigen Immune 
Ms4a1 Membrane-spanning 4-domains, subfamily A, member 1 Immune 
Ms4a4b Membrane-spanning 4-domains, subfamily A, member 4B Immune 
Ncf4 Neutrophil cytosolic factor 4 Defence 
Sp100 Nuclear antigen Sp100 Cell cycle 
Ifi203 Interferon-activated gene 203 Apoptosis 
Itgb2 Integrin β 2 Adhesion 
Pou2af1 POU* domain, class 2, associating factor 1 Transcription 
Tap1 Transporter 1, ATP-binding cassette, subfamily B (MDR†/TAP) Transport 
Nup210 Nucleoporin 210 Nuclear pore 
*

Pit-1/Oct/Unc-86.

Multidrug resistance transporter.

Using the UniGene's EST Profile Viewer to analyse mouse ESTs (expressed sequence tags), one can see that ADPRibase-Mn is most strongly expressed in thymus (see Supplementary Table S1 at http://www.BiochemJ.org/bj/413/bj4130103add.htm). Expression in spleen was also observed, but not as prominently as in the microarray.

The Northern blot was hybridized with a rat cDNA probe that revealed a major transcript of ~1.4 kb in all tissues analysed (Figure 6), in agreement with the structure of RGD1309906 gene (Figure 5). Longer transcripts of ~3 kb and ~5 kb were also observed in some tissues. Quantification of the RNA signals indicated a higher expression in thymus and spleen than in the other rat tissues.

Expression of RGD1309906 transcripts in rat tissues

Figure 6
Expression of RGD1309906 transcripts in rat tissues

A multiple rat-tissue Northern blot is shown after hybridization with digoxinenin-labelled ADPRibase-Mn cDNA. The Figure shows a 3 h exposure of X-ray film to the chemiluminescent membrane made 19 h after allowing the detection reaction to proceed. Numbers below the lanes represent the quantification (as percentages of the spleen signal) of the 1.4 kb band by image analysis during the first 3 h of detection. Abbreviation: Sk., skeletal.

Figure 6
Expression of RGD1309906 transcripts in rat tissues

A multiple rat-tissue Northern blot is shown after hybridization with digoxinenin-labelled ADPRibase-Mn cDNA. The Figure shows a 3 h exposure of X-ray film to the chemiluminescent membrane made 19 h after allowing the detection reaction to proceed. Numbers below the lanes represent the quantification (as percentages of the spleen signal) of the 1.4 kb band by image analysis during the first 3 h of detection. Abbreviation: Sk., skeletal.

As stated in the Introduction, besides ADPRibase-Mn, rat liver supernatants contain two other ADP-ribose pyrophosphatases, named ADPRibase-I and ADPRibase-II, which work with either Mg2+ or Mn2+ as the activating cation [19]. To distinguish the ADP-ribose pyrophosphatase activity of ADPRibase-Mn from the other hydrolases, fractionation of tissue supernatants was necessary (see Supplementary Figure S6 at http://www.BiochemJ.org/bj/413/bj4130103add.htm). This allowed the quantification of the three enzymes in five different rat sources: thymus, spleen, splenocytes, liver and skeletal muscle (Table 4). ADPRibase-Mn activity in thymus and spleen was 2.5–5-fold higher than in liver and muscle. The ratio was even higher when ADPRibase-Mn was quantified in splenocytes, where the activity was 4–8-fold higher than in the non-immune tissues (Table 4).

Table 4
Preferential expression of ADPRibase-Mn activity in rat immune tissues

The activities were quantified after separation of the three ADPRibases by gel filtration of tissue supernatants, followed by ion-exchange chromatography (see Supplementary Figure S6). The results are referred to an amount of tissue containing 1 g of protein (assayed in the crude homogenates) and are means±S.D. for three independent experiments, each performed on one animal.

 Activity (m-units/g of tissue protein) 
 ADPRibase-I ADPRibase-Mn ADPRibase-II 
Thymus 470±140 330±80 170±30 
Spleen 500±80 300±50 160±60 
Splenocytes 410±40 550±190 190±30 
Liver 590±140 130±30 270±70 
Skeletal muscle 180±20 70±20 60±10 
 Activity (m-units/g of tissue protein) 
 ADPRibase-I ADPRibase-Mn ADPRibase-II 
Thymus 470±140 330±80 170±30 
Spleen 500±80 300±50 160±60 
Splenocytes 410±40 550±190 190±30 
Liver 590±140 130±30 270±70 
Skeletal muscle 180±20 70±20 60±10 

DISCUSSION

Our results provide a molecular identity for ADPRibase-Mn and a biochemical basis for the functional annotation of the rat gene RGD1309906, probably extrapolatable to other mammalian genes coding for strongly similar hypothetical metallophosphoesterases. This includes the mouse gene 2310004I24Rik and the human gene c17orf48, which respectively encode proteins 95 and 85% identical with rat ADPRibase-Mn.

Not every known NDP-X hydrolase is a member of the Nudix superfamily [38,39] but, in fact, all those with a clear or relative preference for ADP-ribose [1315,22,40,41] or CDP-alcohol [41,42] are. It has been pointed out that the Mn2+-dependent Nudix hydrolase YZGD, from Paenibacillus thiaminolyticus, displays a specificity similar to that of ADPRibase-Mn [41]. Therefore the fact that the latter is a metallophosphoesterase was relatively unexpected. In the SCOP classification, the metallo-dependent phosphatase superfamily is currently divided into several families, including purple acid phosphatase, DNA double-strand-break repair nuclease, 5′-nucleotidase (synonym UDP-sugar hydrolase) N-terminal domain protein serine/threonine phosphatase and others. Searches in the Superfamily server (http://supfam.org/SUPERFAMILY) failed to assign ADPRibase-Mn to any established SCOP family (no E-value <0.01 was obtained; the E-value of a query-to-family match is the P value multiplied by the number of families in the database), suggesting that it may constitute a novel protein family that can be provisionally described as ADP-ribose/CDP-alcohol pyrophosphatase.

ADPRibase-Mn is inactive in the absence of added metals, and is activated by low-micromolar concentrations of Mn2+, but not of Mg2+. Activation by Mn2+ reflects metal binding to the enzyme, not just to the free substrate, as half saturation was obtained at 1–4 μM MnCl2, a value more than 100-fold lower than that of the substrate. This is compatible with Mn2+ being a natural activator of the enzyme, as free Mn2+ is estimated to be present at 0.7 μM in hepatocytes from fed rats and at 0.25 μM in hepatocytes from food-deprived rats [43].

The metal centre of metallophosphoesterases can contain a couple of the same or different metals in various combinations. However, it is unclear whether activation by Mn2+ reflects binding to the dinuclear centre or otherwise. In fact, the structure of the zebrafish orthologue of ADPRibase-Mn was determined to contain two Zn2+ ions in the dinuclear centre, but it also contained two Zn2+ ions at other locations [36]. The time-dependent inactivation of rat ADPRibase-Mn by phenanthroline, not by EDTA, suggested that a metal other than Mn (perhaps Fe or Zn, that form more stable complexes with the former chelator) could be part of the enzyme. However, phenanthroline could also be acting as an inhibitor by binding to the active centre. In summary, further evidence is needed to elucidate the nature of the metals bound to ADPRibase-Mn and their structural and/or functional role.

The putative active centre of ADPRibase-Mn is in the groove where Pi lies bound to the Zn2+ ions of the crystal structure of the zebrafish protein [36]. In the homology model of the rat protein, a docked ADP-ribose molecule fits into that groove, with the β-phosphate of ADP occupying a position similar to Pi in the X-ray structure. All the ADP-ribose interactions with the Zn2+ ions, and most of those established with the protein, involve the pyrophosphate group (Figure 4 and its legend), and it is likely that some of them are important for catalysis. For instance, His111 (part of the conserved GNHE region III in Figure 3) interacts with the oxygen atom of the phosphoanhydride linkage hydrolysed by ADPRibase-Mn, as in other metallophosphoesterases, where this interaction is catalytically relevant [44]. On the other hand, the specificity of ADPRibase-Mn for NDP(-X) compounds (Table 1) should be accounted for by interactions of the N and the X portions of the substrate. In the ADP-ribose docking model, although the pyrophosphate group is in a rather closed environment, adenine and the distal ribose protrude towards more open areas of the protein surface (Figure 4B). On the N side of ADP-ribose, a hydrophobic contact of adenine with Leu196 is observed, but not any hydrogen bond. On the X side, two possible interactions of the 1″-OH group with backbone atoms of His278 and His280, and a favourable contact of the ribose ring with Cys253 are observed. It is unclear whether these interactions alone account for the preferences of ADPRibase-Mn for ADP as against CDP, for NDP-X as against the corresponding NDPs, and for ADP-ribose as against ADP-glucose, as summarized in Table 1. Anyhow, it must be stressed that the theoretical structure of rat ADPRibase-Mn is derived by homology with a zebrafish protein complexed with Pi, a ligand much smaller than ADP-ribose. In a true enzyme–substrate complex, specific interactions may depend on an induced fit of the enzyme which Pi would not possibly evoke.

ADPRibase-Mn may have a function in vertebrate immune systems. Mining of expression databases identified a mouse microarray experiment in which gene 2310004I24Rik, which codes for the mouse orthologue of ADPRibase-Mn, is called an ‘immune gene’, preferentially expressed in immune as against non-immune cells and tissues [37]. As a complement to this, rat ADPRibase-Mn mRNAs were found to be more abundant (Figure 6) and enzyme activity was higher (Table 4) in thymus and spleen than in non-immune rat tissues. ADPRibase-Mn activity was even higher in splenocytes, including spleen immunocytes (Table 4). Such a pattern was not observed for the Nudix-type ADPRibase-I (Nudt9) and ADPRibase-II (Nudt5), either when assaying enzyme activities in rat tissues (Table 4) or in the RNA expression of the mouse orthologue of the rat Nudt5 gene that codes for ADPRibase-II (a probe for the Nudt9 orthologue, which codes for ADPRibase-I, was not included in the microarray [37]). Also in agreement with a role in the vertebrate immune system, a restriction was found in the taxonomic distribution of ADPRibase-Mn orthologues. BlastP and TblastN searches revealed that they occur in all vertebrates, with the possible exception of birds, but not in invertebrates and most lower eukaryotes (Table 2). This contrasts, for example, with genes orthologous to Nudt9 (coding for ADPRibase-I), which, as well as in vertebrates, is represented in the fruitfly Drosophila and the nematode worm Caenorhabditis elegans [14] or to Nudt5 (coding for ADPRibase-II), which is represented in C. elegans and the budding yeast Saccharomyces [22]. The immune role of ADPRibase-Mn is unknown. In the abovementioned microarray study [37], a cellular function was assigned to 298 of the 333 mouse immune genes identified. These genes were grouped in about 20 functional sets classified as controlling functions such as, for example, defence, apoptosis, immunity, signalling, metabolism etc. (see, e.g., Table 3). Immune gene 2310004I24Rik, here shown to code for the mouse orthologue of ADPRibase-Mn, was annotated simply as ‘metallophosphoesterase’, and could not be assigned to either of those functional sets [37]. The role of ADP-ribose as a second messenger of stress in immune cells [9,10,4547] points to a signalling role for ADPRibase-Mn. This agrees with the roles of the immune genes that are profile neighbours of 2310004I24Rik in the microarray study. Although 16% of the 298 immune genes with a known role are classified as having a ‘signalling’ function [37], this increased to 43% among the profile neighbours of 2310004I24Rik (Table 3). ADPRibase-Mn in the immune system could be related to ADP-ribose turnover and to the termination of its effects on TRPM2 channels. Since the immune-gene microarray does not contain a TRPM2 probe, in order to compare 2310004I24Rik and TRPM2 expression in mouse tissues, we inspected the UniGene database using the EST Profile Viewer. Both genes are expressed at low levels (e.g. as compared with β-actin), which makes comparisons difficult in terms of the numbers of ESTs counted. Anyhow, considering only body sites for which a high number of total ESTs is on record (>40000), one can see that TRPM2 expression is accompanied by ADPRibase-Mn. This represents only a partial correlation, because several sites express ADPRibase-Mn, but not TRPM2 (Supplementary Table S1).

CDP-choline and CDP-ethanolamine are key intermediates in biosynthetic pathways of, respectively, phosphatidylcholine and phosphatidylethanolamine [48,49]. The formation of CDP-choline by CTP-phosphocholine cytidylyltransferase is rate-limiting and important for the regulation of phosphatidylcholine biosynthesis. ADPRibase-Mn is the only mammalian NDP-X hydrolase showing a relative preference for CDP-choline (Table 1). Working simultaneously with the cytidylyltransferase, ADPRibase-Mn would form a futile substrate cycle that could contribute also to regulation:

 
formula

A similar reasoning can be applied to the CDP-ethanolamine pathway. In conclusion, a role of ADPRibase-Mn related to its CDP-alcohol substrates cannot be discounted.

We are grateful to Professor Antonio Sillero and Dr María Antonia Günther Sillero (Departamento de Bioquímica Instituto de Investigaciones Biomédicas Alberto Sols, UAM/CSIC, Facultad de Medicina, Madrid, Spain) for their generous encouragement and support of this research. J. R. R. is indebted to the Escola Superior de Tecnologia e Gestão, Instituto Politécnico de Leiria, Leiria, Portugal, for permission to be on study leave at the Universidad de Extremadura. We thank the Ministerio de Educación y Ciencia, Spain, for Grant BFU2006-00510, co-financed by FEDER (Fondo Europeo de Desarrollo Regional), to the Consejería de Infraestructuras y Desarrollo Tecnológico, Junta de Extremadura, Spain, for grants GRU06031 and GRU07066, co-financed by FSE (Fondo Social Europeo) and FEDER, and to the Vicerrectorado de Investigación, Desarrollo e Innovación, Universidad de Extremadura, for grant A7-09.

Abbreviations

     
  • ADPRibase

    ADP-ribose pyrophosphatase (used only for rat enzymes)

  •  
  • ADT

    AutoDockTools

  •  
  • CSIC

    Consejo Superior de Investigaciones Científicas

  •  
  • DTT

    dithiothreitol

  •  
  • EST

    expressed sequence tag

  •  
  • FEDER

    Fondo Europeo de Desarrollo Regional

  •  
  • GEO

    Gene Expression Omnibus

  •  
  • GST

    glutathione transferase

  •  
  • MALDI

    matrix-assisted laser-desorption ionization

  •  
  • NCBI

    National Center for Biotechnology Information

  •  
  • NDP-X

    nucleoside 5′-diphosphate–X

  •  
  • ORF

    open reading frame

  •  
  • TRPM2

    transient receptor potential melastatin channel-2

  •  
  • UAM

    Universidad Autónoma de Madrid

References

References
1
Hilz
H.
ADP-ribose. A historical overview
Adv. Exp. Med. Biol.
1997
, vol. 
419
 (pg. 
15
-
24
)
2
Berger
F.
Ramirez-Hernandez
M. H.
Ziegler
M.
The new life of a centenarian: signalling functions of NAD(P)
Trends Biochem. Sci.
2004
, vol. 
29
 (pg. 
111
-
118
)
3
Jacobson
E. L.
Cervantes-Laurean
D.
Jacobson
M. K.
ADP-ribose in glycation and glycoxidation reactions
Adv. Exp. Med. Biol.
1997
, vol. 
419
 (pg. 
371
-
379
)
4
Just
I.
Wollenberg
P.
Moss
J.
Aktories
K.
Cysteine-specific ADP-ribosylation of actin
Eur. J. Biochem.
1994
, vol. 
221
 (pg. 
1047
-
1054
)
5
Wondrak
G. T.
Cervantes-Laurean
D.
Jacobson
E. L.
Jacobson
M. K.
Histone carbonylation in vivo and in vitro
Biochem. J.
2000
, vol. 
351
 (pg. 
769
-
777
)
6
Perraud
A. L.
Fleig
A.
Dunn
C. A.
Bagley
L. A.
Launay
P.
Schmitz
C.
Stokes
A. J.
Zhu
Q.
Bessman
M. J.
Penner
R.
, et al. 
ADP-ribose gating of the calcium-permeable LTRPC2 channel revealed by Nudix motif homology
Nature
2001
, vol. 
411
 (pg. 
595
-
599
)
7
Sano
Y.
Inamura
K.
Miyake
A.
Mochizuki
S.
Yokoi
H.
Matsushime
H.
Furuichi
K.
Immunocyte Ca2+ influx system mediated by LTRPC2
Science
2001
, vol. 
293
 (pg. 
1327
-
1330
)
8
Gerth
A.
Nieber
K.
Oppenheimer
N. J.
Hauschildt
S.
Extracellular NAD+ regulates intracellular free calcium concentration in human monocytes
Biochem. J.
2004
, vol. 
382
 (pg. 
849
-
856
)
9
Kühn
F. J.
Heiner
I.
Lückhoff
A.
TRPM2: a calcium influx pathway regulated by oxidative stress and the novel second messenger ADP-ribose
Pflugers Arch.
2005
, vol. 
451
 (pg. 
212
-
219
)
10
Perraud
A. L.
Takanishi
C. L.
Shen
B.
Kang
S.
Smith
M. K.
Schmitz
C.
Knowles
H. M.
Ferraris
D.
Li
W.
Zhang
J.
, et al. 
Accumulation of free ADP-ribose from mitochondria mediates oxidative stress-induced gating of TRPM2 cation channels
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
6138
-
6148
)
11
Gasser
A.
Glassmeier
G.
Fliegert
R.
Langhorst
M. F.
Meinke
S.
Hein
D.
Krüger
S.
Weber
K.
Heiner
I.
Oppenheimer
N.
, et al. 
Activation of T cell calcium influx by the second messenger ADP-ribose
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
2489
-
2496
)
12
Miró
A.
Costas
M. J.
García-Díaz
M.
Hernández
M. T.
Cameselle
J. C.
A specific, low Km ADP-ribose pyrophosphatase from rat liver
FEBS Lett.
1989
, vol. 
244
 (pg. 
123
-
126
)
13
Lin
S.
Gasmi
L.
Xie
Y.
Ying
K.
Gu
S.
Wang
Z.
Jin
H.
Chao
Y.
Wu
C.
Zhou
Z.
, et al. 
Cloning, expression and characterisation of a human Nudix hydrolase specific for adenosine 5′-diphosphoribose (ADP-ribose)
Biochim. Biophys. Acta
2002
, vol. 
1594
 (pg. 
127
-
135
)
14
Perraud
A. L.
Shen
B.
Dunn
C. A.
Rippe
K.
Smith
M. K.
Bessman
M. J.
Stoddard
B. L.
Scharenberg
A. M.
NUDT9, a member of the Nudix hydrolase family, is an evolutionarily conserved mitochondrial ADP-ribose pyrophosphatase
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
1794
-
1801
)
15
Carloto
A.
Costas
M. J.
Cameselle
J. C.
McLennan
A. G.
Ribeiro
J. M.
The specific, submicromolar Km ADP-ribose pyrophosphatase purified from human placenta is enzymically indistinguishable from recombinant NUDT9 protein, including a selectivity for Mn2+ as activating cation and increase in Km for ADP-ribose, both elicited by H2O2
Biochim. Biophys. Acta
2006
, vol. 
1760
 (pg. 
1545
-
1551
)
16
Bessman
M. J.
Frick
D. N.
O'Handley
S. F.
The MutT proteins or “Nudix” hydrolases, a family of versatile, widely distributed, “housecleaning” enzymes
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
25059
-
25062
)
17
McLennan
A. G.
The Nudix hydrolase superfamily
Cell. Mol. Life Sci.
2006
, vol. 
63
 (pg. 
123
-
143
)
18
Cameselle
J. C.
Costas
M. J.
Günther Sillero
M. A.
Sillero
A.
Two low Km hydrolytic activities on dinucleoside 5,5′-P1,P4-tetraphosphates in rat liver. Characterization as the specific dinucleoside tetraphosphatase and a phosphodiesterase I-like enzyme
J. Biol. Chem.
1984
, vol. 
259
 (pg. 
2879
-
2885
)
19
Canales
J.
Pinto
R. M.
Costas
M. J.
Hernández
M. T.
Miró
A.
Bernet
D.
Fernández
A.
Cameselle
J. C.
Rat liver nucleoside diphosphosugar or diphosphoalcohol pyrophosphatases different from nucleotide pyrophosphatase or phosphodiesterase I: substrate specificities of Mg2+-and/or Mn2+-dependent hydrolases acting on ADP-ribose
Biochim. Biophys. Acta
1995
, vol. 
1246
 (pg. 
167
-
177
)
20
Bernet
D.
Pinto
R. M.
Costas
M. J.
Canales
J.
Cameselle
J. C.
Rat liver mitochondrial ADP-ribose pyrophosphatase in the matrix space with low Km for free ADP-ribose
Biochem. J.
1994
, vol. 
299
 (pg. 
679
-
682
)
21
Ribeiro
J. M.
Carloto
A.
Costas
M. J.
Cameselle
J. C.
Human placenta hydrolases active on free ADP-ribose: an ADP-sugar pyrophosphatase and a specific ADP-ribose pyrophosphatase
Biochim. Biophys. Acta
2001
, vol. 
1526
 (pg. 
86
-
94
)
22
Gasmi
L.
Cartwright
J. L.
McLennan
A. G.
Cloning, expression and characterization of YSA1H, a human adenosine 5′-diphosphosugar pyrophosphatase possessing a MutT motif
Biochem. J.
1999
, vol. 
344
 (pg. 
331
-
337
)
23
Cabezas
A.
Costas
M. J.
Pinto
R. M.
Couto
A.
Cameselle
J. C.
Identification of human and rat FAD-AMP lyase (cyclic FMN forming) as ATP-dependent dihydroxyacetone kinases
Biochem. Biophys. Res. Commun.
2005
, vol. 
338
 (pg. 
1682
-
1689
)
24
Cleland
W. W.
The statistical analysis of enzyme kinetic data
Adv. Enzymol. Relat. Areas Mol. Biol.
1967
, vol. 
29
 (pg. 
1
-
32
)
25
Engler-Blum
G.
Meier
M.
Frank
J.
Muller
G. A.
Reduction of background problems in nonradioactive Northern and Southern blot analyses enables higher sensitivity than 32P-based hybridizations
Anal. Biochem.
1993
, vol. 
210
 (pg. 
235
-
244
)
26
Kruisbeek
A. M.
Coligan
J. E.
Bierer
B. E.
Margulies
D. H.
Shevach
E. M.
Strober
W.
Isolation and fractionation of mononuclear cell populations
Current Protocols in Immunology
2000
New York
John Wiley and Sons
(pg. 
3.1.1
-
3.1.5
)
27
Lowry
O. H.
Rosebrough
N. J.
Farr
A. L.
Randall
R. J.
Protein measurement with the Folin phenol reagent
J. Biol. Chem.
1951
, vol. 
193
 (pg. 
265
-
275
)
28
Kopp
J.
Schwede
T.
UniProt AC Q5M886, model for similar to RIKEN cDNA 2310004I24 gene from Rattus norvegicus (rat)
2007
 
28a
Sanner
M. F.
A component-based software environment for visualizing large macromolecular assemblies
Structure
2005
, vol. 
13
 (pg. 
447
-
462
)
29
Word
J. M.
Lovell
S. C.
Richardson
J. S.
Richardson
D. C.
Asparagine and glutamine: using hydrogen atom contacts in the choice of side-chain amide orientation
J. Mol. Biol.
1999
, vol. 
285
 (pg. 
1735
-
1747
)
30
Morris
G. M.
Goodsell
D. S.
Halliday
R. S.
Huey
R.
Hart
W. E.
Belew
R. K.
Olson
A. J.
Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function
J. Comput. Chem.
1998
, vol. 
19
 (pg. 
1639
-
1662
)
31
Perkins
D. N.
Pappin
D. J.
Creasy
D. M.
Cottrell
J. S.
Probability-based protein identification by searching sequence databases using mass spectrometry data
Electrophoresis
1999
, vol. 
20
 (pg. 
3551
-
3567
)
32
Lohse
D. L.
Denu
J. M.
Dixon
J. E.
Insights derived from the structures of the Ser/Thr phosphatases calcineurin and protein phosphatase 1
Structure
1995
, vol. 
3
 (pg. 
987
-
990
)
33
Klabunde
T.
Strater
N.
Frohlich
R.
Witzel
H.
Krebs
B.
Mechanism of Fe(III)-Zn(II) purple acid phosphatase based on crystal structures
J. Mol. Biol.
1996
, vol. 
259
 (pg. 
737
-
748
)
34
Richter
W.
3,5′-Cyclic nucleotide phosphodiesterases class III: members, structure, and catalytic mechanism
Proteins
2002
, vol. 
46
 (pg. 
278
-
286
)
35
Shenoy
A. R.
Capuder
M.
Draskovic
P.
Lamba
D.
Visweswariah
S. S.
Podobnik
M.
Structural and biochemical analysis of the Rv0805 cyclic nucleotide phosphodiesterase from Mycobacterium tuberculosis
J. Mol. Biol.
2007
, vol. 
365
 (pg. 
211
-
225
)
36
Bitto
E.
Wesenberg
G. E.
Phillips
G. N.
Jr
McCoy
J. G.
Bingman
C. A.
Crystal structure of a dimetal phosphatase from Danio rerio LOC 393393
2006
 
37
Hutton
J. J.
Jegga
A. G.
Kong
S.
Gupta
A.
Ebert
C.
Williams
S.
Katz
J. D.
Aronow
B. J.
Microarray and comparative genomics-based identification of genes and gene regulatory regions of the mouse immune system
BMC Genomics
2004
, vol. 
5
 pg. 
82
 
38
Stefan
C.
Jansen
S.
Bollen
M.
NPP-type ectophosphodiesterases: unity in diversity
Trends Biochem. Sci.
2005
, vol. 
30
 (pg. 
542
-
550
)
39
Galperin
M. Y.
Moroz
O. V.
Wilson
K. S.
Murzin
A. G.
House cleaning, a part of good housekeeping
Mol. Microbiol.
2006
, vol. 
59
 (pg. 
5
-
19
)
40
Yagi
T.
Baroja-Fernández
E.
Yamamoto
R.
Muñoz
F. J.
Akazawa
T.
Hong
K. S.
Pozueta-Romero
J.
Cloning, expression and characterization of a mammalian Nudix hydrolase-like enzyme that cleaves the pyrophosphate bond of UDP-glucose
Biochem. J.
2003
, vol. 
370
 (pg. 
409
-
415
)
41
Tirrell
I. M.
Wall
J. L.
Daley
C. J.
Denial
S. J.
Tennis
F. G.
Galens
K. G.
O'Handley
S. F.
YZGD from Paenibacillus thiaminolyticus, a pyridoxal phosphatase of the HAD (haloacid dehalogenase) superfamily and a versatile member of the Nudix (nucleoside diphosphate X) hydrolase superfamily
Biochem. J.
2006
, vol. 
394
 (pg. 
665
-
674
)
42
Xu
W.
Dunn
C. A.
Jones
C. R.
D'souza
G.
Bessman
M. J.
The 26 Nudix hydrolases of Bacillus cereus, a close relative of Bacillus anthracis
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
24861
-
24865
)
43
Ash
D. E.
Schramm
V. L.
Determination of free and bound manganese(II) in hepatocytes from fed and fasted rats
J. Biol. Chem.
1982
, vol. 
257
 (pg. 
9261
-
9264
)
44
Mitic
N.
Smith
S. J.
Neves
A.
Guddat
L. W.
Gahan
L. R.
Schenk
G.
The catalytic mechanisms of binuclear metallohydrolases
Chem. Rev.
2006
, vol. 
106
 (pg. 
3338
-
3363
)
45
Vu
C. Q.
Coyle
D. L.
Tai
H. H.
Jacobson
E. L.
Jacobson
M. K.
Intramolecular ADP-ribose transfer reactions and calcium signalling. Potential role of 2′-phospho-cyclic ADP-ribose in oxidative stress
Adv. Exp. Med. Biol.
1997
, vol. 
419
 (pg. 
381
-
388
)
46
Fonfria
E.
Marshall
I. C.
Benham
C. D.
Boyfield
I.
Brown
J. D.
Hill
K.
Hughes
J. P.
Skaper
S. D.
McNulty
S.
TRPM2 channel opening in response to oxidative stress is dependent on activation of poly(ADP-ribose) polymerase
Br. J. Pharmacol.
2004
, vol. 
143
 (pg. 
186
-
192
)
47
Naziroglu
M.
New molecular mechanisms on the activation of TRPM2 channels by oxidative stress and ADP-Ribose
Neurochem. Res.
2007
, vol. 
32
 (pg. 
1990
-
2001
)
48
Vance
J. E.
Vance
D. E.
Phospholipid biosynthesis in mammalian cells
Biochem. Cell Biol.
2004
, vol. 
82
 (pg. 
113
-
128
)
49
Kent
C.
Regulatory enzymes of phosphatidylcholine biosynthesis: a personal perspective
Biochim. Biophys. Acta
2005
, vol. 
1733
 (pg. 
53
-
66
)
50
Bradford
M. M.
A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding
Anal. Biochem.
1976
, vol. 
72
 (pg. 
248
-
254
)

Author notes

1

This paper is dedicated to Professor Antonio Sillero, our teacher, on his 70th Birthday.

2

On leave from the Escola Superior de Tecnologia e Gestão, Instituto Politécnico de Leiria, Leiria, Portugal.

The nucleotide sequence data reported will appear in GenBank®, EMBL, DDBJ and GSDB Nucleotide Sequence Databases under the accession number EU037900.

Supplementary data