Methylthioadenosine phosphorylase from the archaeon Pyrococcus furiosus Mechanism of the reaction and assignment of disulfide bonds Giovanna Cacciapuoti 1 , Maria Angela Moretti 2 , Sabrina Forte 1 , Assunta Brio 1 , Laura Camardella 3 , Vincenzo Zappia 1 and Marina Porcelli 1 1 Dipartimento di Biochimica e Biofisica ‘F. Cedrangolo’, Seconda Universita ` di Napoli, Naples, Italy; 2 Centro Regionale di Competenza in Biotecnologie Industriali (BioTekNet), Seconda Universita ` di Napoli, Naples, Italy; 3 Istituto di Biochimica delle Proteine, CNR, Naples, Italy The extremely heat-stable 5¢-methylthioadenosine phos- phorylase from the hyperthermophilic archaeon Pyrococcus furiosus was cloned, expressed to high levels in Escherichia coli, and purified to homogeneity by heat precipitation and affinity chromatography. The recombinant enzyme was subjected to a kinetic analysis including initial velocity and product inhibition studies. The reaction follows an ordered Bi–Bi mechanism and phosphate binding precedes nucleo- side binding in the phosphorolytic direction. 5¢-Methyl- thioadenosine phosphorylase from Pyrococcus furiosus is a hexameric protei n with fi ve cysteine residues per subunit. Analysis of the f ragments obtained after digestion of the protein a lkylated without previous reduction identified two intrasubunit disulfide bridges. The enzyme is very resistant to chemical denaturation and the transition midpoint for guanidinium c hloride-induced unfolding was determined to be 3.0 M after 2 2 h incubation. This value d ecreases t o 2 .0 M in the presence of 3 0 m M dithiothreitol, f urnishing evidence that disulfide bonds are needed for protein s tability. The guanidinium chloride-induced unfolding is completely reversible as demonstrated by the analysis of the refolding process by activity assays, fluore scence measurements and SDS/PAGE. The finding of multiple disulfide bridges in 5¢-methylthioadenosine phosphorylase from Pyrococcus furiosus argues strongly that disulfide bond formation may be a significant molecular strategy for stabilizing i ntra- cellular h yperthermophilic proteins. Keywords: d isulfide bonds; hyperthermostability; 5¢-methyl thioadenosine phosphorylase; purine nucleoside phos- phorylase; Pyroco ccus fur iosus. Hyperthermophilic enzymes which retain their s tructure and function near the boiling point of water have been, over the past decade, the object of extensive studies on protein stabilization, folding a nd e volutionary aspec ts [1–4]. More- over, their unique structure–function properties of high thermostability are potentially significant for d eveloping biotechnological applications [5,6]. Thus, there is a great deal of interest in studies on the biochemical adaptation of hyperthermophiles whose enzymes provide unique models for t he study and understanding of the evolution of en zymes in terms of structure, specificity a nd catalytic properties. Much work has been done to identify the structural determinants of the enhanced stability o f h yperthermophilic proteins. Several mechanisms of thermal stabilization have been proposed, among which additional networks of salt bridges and hydrogen bonds, improved packing density and enhanced secondary structure are the most cited [2–4,7–9]. In spite of this, no general rules have been established to date, and it has been concluded that each protein evolves individually through a limited number of factors that occur at different levels, also involving the amino acid sequence and the quaternary structure of the p roteins. In recent years, growing attention has been paid to the presence of disulfide bonds in intracellular hyperthermo- philic proteins where these covalent links may play a key role in protein s tabilization in the extreme t hermal environ- ment [10–14]. 5¢-Methylthioadenosine phosphorylase (MTAP) cata- lyzes the reversible phosphorolysis of 5¢-methylthioadeno- sine (MTA), a sulfur-containing nucleoside formed from S-adenosylmethionine (AdoMet) via several independent pathways of which the polyamine biosynthesis is quantita- tively the most important [15]. The products of the M TA cleavage r eaction are adenine and 5-methylthioribose- 1-phosphate. M TA phosphorylase was first characterized in rat ventral prostate [16]. The enzyme was purified to homogeneity from mammalian tissues [17–19] and from the Correspondence to G. Cacciapuoti, Dipartimeno di Biochimica e Biofisica ÔF. CedrangoloÕ, Seconda Universita ` di Napoli, Via Costantinopoli 16, 80138, Napoli, Italy. Fax: +39 081 441 688; Tel.: +39 081 566 7519; E-mail: giovanna.cacciapuoti@unina2.it Abbreviations: AdoHcy, S-adenosyl- L -homocysteine; AdoMet, S-adenosylmethionine; GdmCl, guanidinium chloride; hMTAP, human MTAP; IPTG, isopropyl thio-b- D -galactoside; MTA, 5¢-methylthioadenosine; MTAP, 5¢-methylthioadenosine phosphory- lase; PfMTAP, 5¢-methylthioadenosine phosphorylase from Pyro- coccus furiosus; PNP, purine nucleoside phosphorylase; SsMTAP, MTAP from Sulfolobus solfataricus; TFA, trifluoroacetic acid. (Received 2 6 July 2 004, revised 1 2 October 200 4, accepted 22 October 2004) Eur. J. Biochem. 271, 4834–4844 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04449.x Archaea Sulfolobus solfataricus [20] and Pyrococcus furiosus [21]. Moreover, crystal structures have been obtained for human MTAP (hMTAP) [ 22] and for MTAP fr om S. solfataric us (SsMTAP) [10]. 5¢-Methylthioadenosine phosphorylase from Pyrococcus furiosus (PfMTAP) is a member of the purine nucleoside phosphorylase (PNP) family of en zymes, which function in the purine salvage pathway of cells [23]. PNP are classified into two main categories: Ôlow- molecularmassPNPÕ, homotrimers, specific for catalysis of 6-oxopurines and their nucleosides [23], and Ôhigh- molecular mass PNPÕ, homohexamers, with broad substrate specificity in that they accept both 6-oxo- and/or 6-amino- purines and their nucleosides as substrates [23]. The two classes do not have sequence homology but the analysis of the three-dimensional structure of their monomers showed significant similarity [24,25]. PfMTAP can b e considered a P NP with unique features [21]. In fact, because o f its hexameric quaternary s tructure, this enzyme belongs to the high-molecular mass class of PNP. By contrast, because of its a mino acid sequence, PfMTAP is more similar to hMTAP, a trimeric enz yme with high substrate specificity for MTA [ 22]. PfMTAP is highly thermoactive with an optimum temperature of 125 °C and is extremely thermostable, retaining 98% residual activity after 5 h at 100 °Cand showing a half-life of 43 m in at 130 °C [21]. The enzyme is also extremely stable to proteolytic cleavage and after incubation with protein d enaturants, detergents, organic solvents, and salts even at high t emperature [21]. PfMTAP contains 30 cysteine residues (five p er subunit). These residues, on the basis of biochemical evidence such as decrease of the thermal stability in the presence of dithio- threitol and d ifferent mobility levels of t he enzyme on SDS/ PAGE run under reducing and nonreducing conditions, are thought to be involved in intrasubunit disulfide bonds [21]. We describe here the in vitro expression, purification and characterization of the hyperthermostable PfMTAP. We carried out a detailed k inetic investigation in orde r to clarify the mechanism of the reaction and the sequence of binding of substrates. Moreover, we determined the pattern of disulfide bridges for the first time and demonstrated, o n the basis of e quilibrium studies of guanidinium chloride (GdmCl)-induced denaturation in the presence and absence of reducing a gents, that disulfide bonds are needed for PfMTAP stability. Materials and methods Bacterial strains, plasmid, enzymes, and chemicals Plasmid pET-22b(+) and the NucleoSpin Plasmid kit for plasmid DNA preparation were obtained f rom Genenco (Duren, Germany). Escherichia coli strain BL21(kDE3) was purchased from N ovagen ( Darmstadt, German y). P. furio - sus ch romosomal DNA was kindly provided by C . B ertoldo (Technical University Hamburg-Harburg, Germany). Specifically synthesized oligodeoxyribonucleotides were obtained f rom P rimm (Naples, I taly). R estriction endonuc- leases and DNA-modifying enzymes were obtained from Takara Bio, Inc. (Otsu, Shiga, Japan). Pfu DNA poly- merase was purchased from Stratagene (La J olla, CA, USA). [methyl- 14 C]AdoMet (50–60 mCi Æmmol )1 was sup- plied by the Radiochemical Centre (Amersham Bioscience, Buckinghamshire, UK). MTA and 5 ¢-[methyl- 14 C]MTA were prepared from unlabeled and labeled AdoMet [26] and purified by HPLC [27]. Sephacryl S-300, AH-Sep harose 4B, S-adenosyl- L -homocysteine (AdoHcy), adenosine, adenine, guanosine, guanine, inosine, hypoxanthine, O-br omoacetyl- N-hydroxysuccinimide and standard proteins used in molecular mass studies were obtained from Sigma (St. Louis, MO, USA). GdmCl a nd dithiothreitol were from Applichem ( Darmstadt, Germany). 4 -Vinylpyridine and CNBr were purchased from A ldrich (Steinheim, Germany). PD-10 columns were from Amersham Pharmacia Biotech. All r eagents were of the purest commercial grade. Enzyme assay MTA phosphorylas e activity w as determined by measuring the formation of [methyl- 14 C]5-methylthioribose-1-phos- phate from 5¢-[methyl- 14 C]MTA [20]. Unless otherwise stated, the standard incubation mixture contained the following: 20 lmol potassium phosphate buffer, pH 7.4, 80 nmol of [methyl- 14 C]MTA (6.5 · 10 5 cpmÆlmol )1 ), and the enzyme p rotein in a final volume of 200 lL. The incubation was p erformed in sealed glass vials for 5 min a t 80 °C, except where indicated otherwise. The vials were rapidly cooled in ice, and the reaction was stopped by the addition of 100 lL of 10% (v/v) trichloroacetic acid. The mixture was then applied to a Dowex 50-H + column (0.6 · 2 cm) equilibrated in H 2 O. 5-[methyl- 14 C]Methylth- ioribose-1-phosphate produced was eluted with 2.5 mL of 0.01 M HCl directly into scintillation vials and counted for radioactivity. Control experiments in the absence of the enzyme were performed in order to correct for MTA hydrolysis. When the assays were carried out at tempera- tures above 80 °C, the reaction mixture was preincubated for 2 min without the enzyme, which was added immedi- ately before starting th e reaction. When inosine, guanosine, and adenosine were used as substrates, the formation of purine base was measured by HPLC using a Beckman system Gold a pparatus. The amount of purine b ase formed is d etermined by m easuring the percentage of the absorbance integrated peak area of purine base formed w ith respect to the total (nucleo- side + purine base) absorbance integrated peak areas. An Ultrasil-CX column (Beckman) eluted with 0.05 M ammo- nium formate, pH 3 at a flow rate of 1 mLÆmin )1 was used when adenosine and/or guanosine w ere the substrates o f the reaction. In these experimental conditions the retention times of adenosine and adenine, guanosine and guanine were 7 .3 and 12.4 min, and 4.2 and 6 min, respectively. When the assays were carried out in the presence of inosine as substrate, an Ultrasphere O DS R P-18 column was employed and t he elution was carried out with 5 : 95 (v/v) mixture of 95% methanol and 0 .1% trifluoroacetic acid (TFA) in H 2 O. The retention times of inosine a nd hypoxan thine were 10.5 a nd 4.7 min, respectively. The s ame H PLC assay h as been c arried out with unlabeled MTA as s ubstrate. In this case an Ultrasphere ODS RP-18 c olumn was equilibrated and eluted with 20 : 80 (v/v) mixture of 95% methanol and 0.1% TFA in H 2 O. The retention times of MTA and adenine were 10 and 4.2 min, respectively. Ó FEBS 2004 Methylthioadenosine phosphorylase from P. furiosus (Eur. J. Biochem. 271) 4835 In product inhibition studies, 0.4 lg of enzyme protein in a final v olume of 200 lL w ere e mployed. The reaction was carried out in the p resence o f 2 lmol of potassium phosphate buffer pH 7.4 when MTA and adenosine were the variable substrates, and in the p resence of 5 lmol Hepes buffer pH 7.4 when phosphate was the variable substrate. In all o f t he kinetic a nd pu rification s tudies the a mount of the protein was adjusted so that no more than 10% of the substrate was converted to product and the reaction rate was strictly linear as a function of time and protein concentration. All enzyme reactions were performed in triplicate at 80 °C. K m and V max values were obtained f rom linear regression analysis of data fitted to the Michaelis–Menten equation. Cloning and expression of the PfMTAP-encoding gene The previously obtained N -terminal amino acid s equence of PfMTAP [21] was u sed for BLAST search of the c omplete genome sequence of P. furiosus. (http://comb5-156.umbi. umd.edu/). The coding region of PfMTAP was cloned into t he pET- 22b(+) expression vector via two engineered restriction sites (NdeIandBamHI) introduced by PCR with the following primers 5¢-GACGGTGATA CATATGCCCAAGATAG GG-3¢,sense,and5¢-G CAGCTACAA GGATCCAAAG TAAATAGG-3¢, antisense (the introduced restriction sites are underlined). Isolated genomic P. furiosus DNA (20 ng), hydrolyzed using BamHIwasusedasatemplate.PCR amplification was performed with P. furiosus DNA polym- erase and a Minicycler (Genenco) programmed for 29 cycles, each cycle consisting of denaturation at 92 °Cfor 1 min, annealing at 55 °C for 2 min and e xtension at 72 °C for 2 min plus 5 sÆcycle )1 , followed by a n extension final step of 15 min at 72 °C. The amplified ge ne (25 ng), hydrolyzed using NdeIandBamHI was inserted into pET22b(+) (150 ng) cut with the same restriction enzymes. The recombinant plasmid was named pET-MTAP. The nucleotide s equence of t he inserted gene was d etermined by MWG BIOTECH t o ensure t hat no m utations were present in the gene. For the expression of recombinant P fMTAP, an over- night culture of E. coli BL21 (kDE3) tr ansformed with the plasmid pET-MTAP w as used as 0.5% inoculum in 1 L of Luria–Bertani medium [28] containing 100 lgÆmL )1 ampi- cillin at 37 °C. At a late stage of cellular growth (when the culture r eached an optical density of 3.0) isopropyl thio- b- D -galactoside (IPTG) was added to 1 m M final concen- tration a nd the induction was p rolonged for 16 h. Cells were harvested by cen trifugation and lyse d a s described by Sambrook et al. [28]. The cell debris was removed by centrifugation at 20 000 g for 60 min at 4 °Candthe supernatant was used as a c ell-free extract. Purification of recombinant PfMTAP Recombinant P fMTAP w as purified in two steps. T he cell- free extract of BL21 E. coli cells expressing PfMTAP was heated at 100 °C for 10 min and centrifuged at 20 000 g for 60 min. After dialysis overnight against 10 m M Tris/HCl pH 7.4, the enzyme was applied to an affinity column of AdoHcy Sepharose (2 · 12 cm) prepared as described by Porcelli et al. [29] equilibrated with 2 0 m M Tris/HCl pH 7.4. The column was washed stepwise with 50 mL of the equilibration buffer and then with the same b uffer containing 0.5 M NaCl until the absorbance at 280 nm reached the baseline. MTA phosphorylase a ctivity was then eluted with 20 m M Tris/HCl pH 7.4 containing 0.5 M NaCl and 3 m M MTA. Active fractions were pooled, concentrated and dialyzed extensively against 10 m M Tris/HCl pH 7.4. Protein analysis Proteins were assayed by the Bradford method [30] using bovine s erum alb umin as standard. P rotein eluting from the columns during purification was monitored as absorbance at 280 n m. The concentration of purified PfMTAP was estimated spectrophotometrically using e 280 ¼ 23 500 M )1 Æcm )1 . The molecular mass of the native protein was determined by gel fi ltration and nondenaturating PAGE. Gel filtration was performed on a calibrated Sephacryl S-300 column (2.2 · 95 cm) equilibrated with 1 0 m M Tris/HCl pH 7.4 containing 0.3 M NaCl at a flow rate of 4 mLÆh )1 .The column was calibrated by using standard proteins of known molecular mass. Nondenatur ating PAGE was carried out at pH 7.5 as reported by Cacciapuoti et al.[31].Thegelswere either staine d with C oomassie Blue or cut in to thin slices and assayed for MTA phosphorylase activity by incubating in the assay mixture at 80 °C for 10 min. The subunit molecular mass was d etermined by S DS/PAGE, as described b y Weber et al. [32], using 1 2 o r 15% acrylamide resolving gel and 5% acrylamide stackin g gel. Samples were heated at 100 °Cfor 5 m in in 2% SDS, 5% 2-mercaptoethanol and run in comparison with molecular m ass standards. Enzyme thermostability was tested by incubating the protein in sealed glass vials at temperatures between 100 a nd 145 °C. Samples (2 lg) were taken at time intervals and residual activity was determined by the standard assay at 80 °C. Activity values are expressed as a percentage of the zero-time control (100%). Fluorescence spectroscopy Fluorescence emission spectra of tryptophan 69 and 208 of PfMTAP were used to monitor any changes i n t he environment o f these residues upon the unfolding of the protein. Intrinsic fluorescence emission measurements were carried out on a Perkin–Elmer (Norwalk, CT, USA) MMF-44 spectrofluorometer in the range of fluorescence linearity using a 1-cm path length quartz c uvette and 5-nm slit width. The absorbance of all solutions was 0.05– 0.10 at the excitation wavelength. Fluorescence emission spectra were recorded at 300–450 nm at the controlled temperature of 25 °C with the excitation wavelength set at 290 n m. Experiments were corrected for background signal. Equilibrium experiments on GdmCl-induced unfolding and refolding For unfolding, PfMTAP (final concentration 0.125 mgÆmL )1 ) was incubated for 22 h at 25 °CinGdmClat 4836 G. Cacciapuoti et al. (Eur. J. Biochem. 271) Ó FEBS 2004 various concentrations (0–6 M )in20m M Tris/HCl pH 7.4 in the presence and in the absence of 30 m M dithiothre- itol. Unfolding was probed by recording the intrinsic fluorescence emission. After 22 h, refolding was started by 20-fold dilution of the unfolding mixture in Tris/HCl 20 m M pH 7.4 at 25 °C. The final concentration of GdmCl in the renaturation mixture was 0.3 M ,whereas the protein concentr ation w as about 6 lgÆmL )1 .The refolded enzyme, after extensive dialysis against Tris/HCl 20 m M pH 7.4 until complete removal of GdmCl, was analyzed by intrinsic fluorescence emission, catalytic activity measurements under s tandard conditions, and SDS/PAGE analysis. Protein fragmentation with CNBr and peptide mapping Purified recombinant PfMTAP was alkylated w ith 4 -vinyl- pyridine under d enaturing conditions with and without previous reduction by the f ollowing procedure. The e nzyme (0.4 mg, 13.3 nmoles) was dissolved in denaturing buffer containing 0.5 M Tris/HCl, pH 7 .8, 2 m M EDTA, 6 M GdmCl in the presence and in the absence o f dithiothreitol at a 150-fold molar excess over cysteine residues and the solution was incubated at 40 °C under n itrogen for 2 h (this step was omitted in the sample alkylated without pre vious reduction). 4-Vinylpyridine (fivefold molar excess over all thiol g roups) was added to the reduced and nonreduced samples and the r eaction proceeded at room temperature i n the dark under nitrogen for 45 min. T he resulting a lkylated samples were immediately desalted b y gel filtration on prepacked PD-10 column equilibrated with 0.1% (v/v) TFA, and dried under v acuum. Cleavage at methionyl residues was achieved by dissolving t he samples in GdmCl 6 M /HCl 0.2 M followed b y a ddition of 160-fold molar excess ( over methionine) of CNBr. The samples were kept at 25 °C in the dark for 24 h and then dried under vacuum. The peptide mixture was separated by reverse-phase HPLC on a 4.6 · 250 mm Vydac C 18 column using a Beckman system Gold apparatus. The elution was accomplished by a linear gradient from 5 to 60% in 60 min o f solvent B ( 0.08% TFA in aceton itrile) in solvent A (0.1% aqueous TFA) at a flow rate of 1 mLÆmin )1 . The eluate was monitored at 220 and 280 nm. Individual peptide fractions were manually collected, dried under vacuum, and sequenced. N-Terminal sequencing and mass spectrometric analysis Peptides were analyzed by automated Edman degradation using a protein sequencer model Procise 492 from Applied Biosystem with i n line phenylthiohydantoine analysis. Mass spectrometry analysis of individual peptide fractions were performed by MALDI-MS mass spectrometry on a Voyager DE Pro mass spectrometer (Applied Biosystem, Foster City, CA, USA) operating in positive-ion linear mode. S amples were mixed with saturated solution of a-cyano-4-hydroxycinnaminic acid (10 mgÆmL )1 ) in aceto- nitrile/0.2% TFA 70 : 30 (v/v) and applied to the metallic sample plate b efore air-drying. Mass calibration was performed with the ions from ACTH (fragment 18–39) at 246 600 Da and cytochrome c at 618 100 Da (MH 2 ) 2+ as internal standard. Average mass values were measured in this analysis. Results and Discussion Analysis of the gene, cloning, expression and purification From the complete genome sequence of P. furiosus the gene PF0016, encoding PfMTAP was identified as a 774- bp fragment that, when translated, contained an N-terminus that matched exactly the one determined from the purified enzyme [21]. The structural gene of PfMTAP encodes a protein of 257 residues with a predicted molecular mass of 29 219 Da, which is in good agreement with that of 30 ± 1 k Da estimated by biochemical a nalyses for the native enzyme [21]. The coding region starts with an ATG triplet, at the position 14 581 of the P. furiosus genome, in agreement with data from protein amino acid sequence determination, which i ndicates that the N-terminal methionine is not post- translationally removed. The first s top codon , T GA, is encountered at the position 15 355. Upstream from the coding region, 13 bp before the starting codon, there is a stretch of purine-rich nucleosides (GACGG) that may function as the ribosome-binding site [33]. Putative promo- ter elements, which are in good agreement with t he archaeal consensus [33], designated box A and box B, were found close to the transcription start site. A hexanucleotide with the sequence TAAATA similar to the box A is located 27 bp upstream from the start codon and resembles the TATA box which is involved in binding the archaeal R NA polymerase [ 33]. A putative box B ( ATGC) overlaps the ATG codon. Finally, a pyrimidine-rich region (TTT TTTAT), strictly r esembling the archaeal terminator signal [33], was localized 2 bp downstream from the translation stop codon. All these sequences were identified on the bas is of their similarity with those reported in nearby r egions of other g enes of proteins isolated from P. furiosus [34,35] or from other Archaea [33]. As reported fo r P. furiosus and other Pyrococcus genomes [36] a strong bias against the CG dinucleotide is observed in t he gene encoding PfMTAP which is reflected i n all codons except one proline and one threonine codon. In contrast, the CG dinucleotide-containing codons are fre- quently used in E. coli [37]. The significance of this bias in P. furiosus and other hyperthermophiles may be that, at the optimal growth temperature approaching 1 00 °C, cytosine deamination can occur which causes the formation of uracil in the DNA. The subsequent C t o T trans ition will produce a damage of protein function. The PCR-amplified fragment o f PfMTAP was cloned i nto pET-22b(+). The sequence of the gene was found to be identical with the published PfMTAP sequence (GenBank identifier PF0016). The recombinant PfMTAP was expressed in soluble form in E. coli BL21 cells harboring the p ET-MTAP plasmid at 3 7 °C i n the presence of IPTG. The most favorable conditions for the expression of the enzyme were found to be when IPTG was added at a late stage of cellular growth and when the induction was prolonged for 1 6 h. Therefore, these co nditions were chosen for the large scale production of recombinant PfMTAP and about 12 g of w et cell paste was obtained from 1 L o f culture. Recombinant PfMTAP was easily purified to homogen- eity 12.6-f old by t wo-step pur ification procedure ( Table 1). The first step in the purification of the protein from crude Ó FEBS 2004 Methylthioadenosine phosphorylase from P. furiosus (Eur. J. Biochem. 271) 4837 cell lysate was an optimized heat precipitation, made possible by the extreme thermostability of the enz yme. As shown in F ig. 1, which reports the analysis by SDS/PAGE of recombinant PfMTAP at different stages o f purification, most E. coli thermolabile proteins can be denaturated and precipitated by heating and only minor contaminants of thermostable recombina nt PfMTAP are detectable. The remaining impurities were removed by an affinity chroma- tography on AdoHcy-Sepharose. A bout 9.2 mg of enzyme preparation w ith a 43% yield was easily obtained from 1 L of culture. IPTG-induced E. coli cells transformed with p ET-MTAP produced  1.77 mg of recombinant protein per gram of cells: thus the expression is about 15-fold higher than for MTAP from P. furiosus [21]. The purified recombinant PfMTAP was biochemically analyzed with respect to molecular properties and compared with the n ative enzyme purified from P. furiosus.The apparent molecular masses of the en zyme (180 kDa, as determined by gel filtration) and its subunit (30 kDa, as judged by SDS/PAGE) were indistinguishable from those of thenativeenzymefromP. furiosus. When compared with the native PfMTAP, the recombinant enzyme shows the same features of thermoactivity (optimum temperature 125 °C) thermoresistance to reversible denaturation (appar- ent T m 137 °C after 10 min preincubation as a f unction of temperature) and stability in the presence of protein denaturants, and detergents. All these data indicate that in vitro reconstitution of PfMTAP yielded a recombinant hexameric enzyme with properties identical to those of the native enzyme isolated from P. furiosus including proper folding. As observed f or native PfMTAP [21], i n the recombinant enzyme thiol groups are not involved in the c atalytic process, whereas disulfide bond(s) are present because incubation with 0.8 M dithiothreitol signific antly reduces the thermostability of the enzyme. Furthermore, we can hypo- thesize that the disulfide linkage(s) a re positioned i ntrasub- unit because, when subjected to SDS/PAGE, the reduced and nonreduced form of the enzyme migrates as a protein band at  30 kDa, which corresponds to the monomer of the enzyme (data not shown). It is interesting to note that PfMTAP is one of the few disulfide bonds-co ntaining proteins functionally over- expressed in E. coli where t he foldin g of proteins with postbiosynthetic modifications as disulfides, could represent a limiting step in their pr oduction [38]. Mechanism of the reaction The reaction catalyzed by PNP is reversible. Thermo- dynamically, the equilibrium of the reaction is shifted in favor of nucleoside synthesis. However, under physiological conditions, the reaction proceeds in the phosphorolytic direction owing to the rapid removal and metabolism of the phosphorolysis products, i.e. purine bases and pentose- 1-phosphate [23]. In analogy, PfMTAP is able to catalyze the reverse synthetic reaction. The incubation for 5 min at 80 °C of reco mbinant PfMTAP in the pre sence of adenine or guan ine or hypoxanthine and ribose-1-phosphate resul- ted in the synthesis of the corresponding nucleosides. Like native PfMTAP [21], the recombinant enzyme i s characterized by broad substrate specificity toward purine nucleosides. In fact, it shows a similar 10-fold higher affinity for MTA (K m 147 l M ) and adenosine (K m 109 l M )with respect to inosine (K m 963 l M ) and guanosine (K m 916 l M ). As previously demonstrated, the K cat /K m values for MTA (1.66 · 10 5 M )1 Æs )1 ) and adenosine ( 2.09 · 10 5 M )1 Æs )1 ), are  20-fold higher than for inosine (9.74 · 10 3 M )1 Æs )1 )and guanosine (7.98 · 10 3 M )1 Æs )1 ), and indicate that 6-amino purine nucleosides are the preferred a nd p robably t he physiological substrates of the e nzyme [21]. The broad substrate specificity of P fMTAP towards purine nucleosides is of interest for potential biotechnological applications. It is well known that nonspecific bacterial phosphorylases are useful tools f or enzymatic synthesis of nucleoside analogues with chemotherapeutic activity [39]. Moreover, a gene therapy for human tumors profits b y the differences in substrate specificity of human and E. coli PNPs [40]. In order to define the m echanism of t he reaction catalyzed by PfMTAP and the sequence of binding of the substrates, a detailed kinetic investigation has been carried out. The double reciprocal plot of the initial velocities at variable concentrations of phosphate and five fixed Table 1. Purification of recombinant 5¢-methylthioadenosine phos- phorylase from P. furiosus. A typical purification from 12 g of wet cells is shown. Specific a ctivity is expressed as lmol MTA cleaved p er min per mg of protein at 80 °C. Sample Total protein mg Total activity units Specific activity unitsÆmg )1 Yield % Purifi- cation n-fold Crude extract 268 139.36 0.52 100 – Heat treatment 27.3 125.58 4.6 90.1 8.85 AdoHcy-Sepharose 9.2 60.44 6.57 43.3 12.63 Fig. 1. SDS/PAGE of recombinant PfMTAP at different stages of purification. Lane A, molecular mass markers; lane B, E. coli BL- 21 transformed with pET-MTAP, crude extract (20 lg); lane C, E. c oli BL-21 transformed with pET-MTAP af ter induction w ith IPTG, crude extract (20 lg); lane D, the same sample as lane C heated at 100 °C f or 10 min a nd cleared by c entrifugation at 20 0 00 g (10 lg); laneE,thesamesampleaslaneDafteraffinitychromatography (2 lg). 4838 G. Cacciapuoti et al. (Eur. J. Biochem. 271) Ó FEBS 2004 concentrations of MTA yielded a series of lines intersect- ing t o the left of the vertical axis ( Fig. 2A). A similar pattern was observed when MTA was varied at five fixed concentrations of inorganic orthophosphate (Fig. 2B). The K m values, graphically extrapolated by replotting the slopes and the intercepts of the primary double- reciprocal plots vs. the reciprocal concentrations of the nonvariable substrates (insets in Fig. 2A,B) are 107 ± 6. 4 l M for MTA and 280 ± 14 l M for phosphate. The obtained results permit us, according to the consid- erations of Cleland [ 41], to rule out a ping-pong mechanism and are consistent with a sequential mechan- ism. A sequential mechanism has been proposed for PNP from both high- and low-molecular mass class [23], whereas no evidence for a ping-pong mechanism has been reported. In addition, the existence of ternary complexes have been revealed by X-ray studies for E. coli [42], calf spleen [43], and human erythrocyte PNP [44] and for MTAP from Sulfolobus solfataricus [10]. On the basis of steady-state kinetic data, several different kinetic mechanisms have been identified for PNPs isolated from a variety of tissues and species. A lthough a sequential Bi–Bi mechanism has been pr oposed most often, there is n o consensus on whether it is ordered or random, and on the order of substrate binding and product release [23]. Similarly, different kinetic mechanisms have been proposed for MTAP, i.e. a random-sequential mechanism has been shown f or mammalian e nzyme [45] and an ordered- sequential mechanism, with MTA as the first substrate to bind and 5-methylthioribose-1 phosphate as the first prod- uct to leave, has been demonstrated for rat lung MTAP [46] and for Drosophila melanogaster MTAP [47]. To verify whether the reaction catalyzed by PfMTAP in the p hosphorolytic direction proceeds via an ordered binding of substrates or via a r andom mechanism, product inhibition studies have been designed. From the secondary plots o f slopes and intercepts vs. t he concentrations of inhibitors, shown a s i nsets o f F ig. 3, an inhibition con stant Fig. 2. Two-substrate s teady-state kinetics. (A) Plot o f the reciprocal of initial velocity ( V) vs. the reciprocal of phosphate concentration (l M )at 36.7 l M (r), 72.7 l M (m), 126. 7 l M (s), 252.7 l M (h) a nd 504.7 l M (d) M TA concentration. In the insets a re reported linear replots of the slopes and of the intercept s of plot (A) vs. the r eciprocal of con centr atio ns of MTA. (B) Plo t of the reciproc al of initial velocity ( V)vs.thereciprocalof MTA concentration (l M )at100l M (j), 250 l M (s), 500 l M (r), 2000 l M (m), and 10 000 l M (d) p hosphate c oncentration. In the insets a re reported linear replots of the slopes and of the intercepts of plot (B) vs. the reciprocal of concentrations of phosphate. Purified enzyme (0.125 lg) was employed. The values of K m extrapolated from the r eplots are 107 l M for MTA and 200 l M for phosphate. Ó FEBS 2004 Methylthioadenosine phosphorylase from P. furiosus (Eur. J. Biochem. 271) 4839 can b e d etermined from the slope replot (K is )ortheintercept replot (K ii ) and is the h orizontal intercept on t he plot [41]. When MTA or adenosine w as varied at fixed concen- trations of phosphate, both adenine (Fig. 3A) and ribose- 1-phosphate (Fig. 3B) products exert a noncompetitive inhibition in a similar experimental protocol. By contrast, when phosphate was varied with MTA or adenosine as a fixed substrate, adenine acts as a noncompetitive inhibitor (Fig. 3 C), whereas a pattern of competitive inhibition was observed for ribose-1-phosphate (Fig. 3D) suggesting that phosphate and ribose-1-phosphate compete f or the same site or the same enzyme form. These product inhibition studies are consistent with an ordered Bi–Bi mechanism [41] in which phosphate is the first substrate to add to the enzyme and ribose-1-phosphate is the last product to dissociate from the enzyme surface. This m echanism proposal is strengthened by the protection exerted by phosphate against thermal inactivation, previously dem- onstrated for native PfMTAP [21], suggesting that phos- phate forms a binary complex with the enzyme. Equilibrium studies of GdmCl-induced unfolding and refolding To analyze quantitatively the stability of PfMTAP and to point out the presence of disulfide bond(s) we performed equilibrium transition studies by incubating the enzyme at increasing GdmCl concentrations (0–6 M )in20m M Tris/ HCl, pH 7.4 for 22 h at 25 °C in the presence and in the absence of 30 m M dithiothreitol. In Fig. 4 are reported the denaturation curves determined by monitoring the shift in fluorescence maximum w avelength upon excitation at 290 n m, where only tryptophanyl residues are specifically excited. In the n ative state, PfMTAP exhibits a fluorescence emission maximum at 330 nm typical of a protein with partially buried tryptophanyl residues. The denaturation process induced by 6 M GdmCl brought about a large red shift and a decrease in fluorescence intensity. Both changes are expected for a n i ncreased exposure o f t he tryptophanyl residues to the more polar aqueous solvent. The different values of the maximum fluorescence emission wavelength Fig. 3. Product inhibition studies. (A) Plot of the reciprocal of the initial velocity with respect to the re ciprocal of MTA concentration in the absence (r) and presence of 20 l M (j)and50l M (m) adenine. The inh ibition co nstants K is and K ii estimated from the rep lots o f the slopes and t he intercepts vs. the concentration of t he in hibitor, shown in the insets, are 12.5 and 33.2 l M , respectively. (B) Plot of t he rec iprocal of the i nitial velocity with respect t o t he rec i procal of adenosine concentration in the absence ( r)andinthepresenceof50l M (j)and200l M (m)ribose 1-phospate. In the insets are shown the replot s o f the slopes and the intercep ts vs. the con centration of t he inhibitor. The c alculated i nhibition constants K is and K ii are 436 .2 and 366.9 l M , respectively. (C) Plo t of the re cipro cal of the in itial velocity with respect t o the reciprocal of phosphate concentration i n t he abse nce ( r) and in t he presence 50 l M (j)and100l M (m) adenine. In the insets are shown the replots of the slopes and the intercepts vs. the concentration of the inhibitor. The c alculated K is and K ii are 47.3 a nd 107.7 l M , respectively. (D) P lot of the reciprocal of the initial velocity with respect to the reciprocal of phosphate concentration in the absence (r) and in the presence o f 50 l M (j)and200l M (m)ribose- 1-phospate. The inhibition constant (Ki) of ribose-1-ph osphate calculated from the replot shown in the in se t is 113.5 l M . 4840 G. Cacciapuoti et al. (Eur. J. Biochem. 271) Ó FEBS 2004 observed at t he end of t ransition, i.e. 349.3 and 339.8 nm in the absence and presence of reducing agents, respectively, suggest significant modifications of the enzyme due to the reduction of disulfide bond(s). T he observed GdmCl- induced denaturation curves show a single sigmoidal transition indicating an apparent two-state transition from the native to t he unfolded state without any detectable intermediate. The 3 M GdmCl value of the midpoint transition observed under nonreducing conditions indicates that the enzyme is not only extremely thermostable, but also very resistant to chemical denaturation. The addition of 30 m M dithiothreitol to a GdmCl-induced denaturation experi- ment shifted the apparent midpoint of the transition for GdmCl to  2 M indicating a significant decrease of protein s tability i n t he pre sence of reducing agents. This result and the already observed loss of activity after incubation of the enzyme at high temperature in the presence of 0.8 M dithiothreitol offer convincing evidence of th e presence of d isulfide(s) bonds and suggest the crucial role played by these co valent links in the stabilization of the protein. To examine whether the GdmCl-induced unfolding of PfMTAP is reversible, the refolding reaction was induced by 20-fold dilution of the sample. Extensive dialysis was then carried out until the complete removal of the denaturant. The refolding p rocess w as monitored b y fluorescence measurements, SDS/PAGE and enzymatic assays. As observed in Fig. 5, the intrinsic fluorescence emission intensity of PfMTAP unfolded in 6 M GdmCl in the absence of reducing agents (curve c) was decreased about twofold a s compared w ith t hat of t he native enzyme (curve a), indicating that one or both tryptophanyl residues e nvironment i s structurally perturbed b y the denaturant. By contrast, the presence of dithiothreitol (curve d) induces a further decrease in fluorescence emission inten sity, suggesting that the reduction of disulfide bond(s) represents the structural modification producing the observed spectral changes. The observation that after the complete removal of the denaturant and of the reducing agent (curve b) the p rotein exhibits a fluorescence spectrum with the same features of the native enzyme, i.e. a fluorescence maximum centered at 330 nm and a sim ilar value of relative fluorescence intensity, indicates that the denaturation process is reversible. Only the denaturation of the enzyme carried out in the p resence of reducing agents proved to be reversible indicating that the presence o f intact d isulfide bonds interferes with the refolding process of PfMTAP. Owing to the stability of PfMTAP towards 2% SDS at room temperature [21], it has been possible to monitor the native hexameric s tructure r ecovery of the protein by SDS/ PAGE. The inset in F ig. 5 compares the SDS/PAGE pattern of the native and refolded enzyme. The samples were subjected to SDS/PAGE without boiling and under nonreducing conditions to obtain a picture of the protein species present. Afte r r efolding (lane 2), the e nzyme migrates as a single band at 180 kDa, corresponding to the m olecular mass of the hexameric PfMTAP. Further- more, w hen assayed f or catalytic a ctivity, the r efolded enzyme shows t he same specific activity of t he native form. On the basis of the r eported d ata we concluded t hat PfMTAP represents a rare example, if not the only, reported in the literature so far, o f a oligomeric hyperther- mophilic protein with disulfide bonds able to undergo a reversible unfolding. Studies are in progress to quantita- tively evaluate the equilibrium and k inetic stability of PfMTAP and t o determine the m ain thermodynamic parameters of the protein. Fig. 5. Fluorescence emission spectra of PfMTAP. The fluoresce nce emission spec tra we re rec orded a fter 22 h incubation at 25 °C. (A) PfMTAP in 20 m M Tris/HCl, pH 7.4; (B) refolded PfMTAP after unfolding in the presence of 30 m M dithiothreitol; ( C) PfMTAP in 6 M GdmCl in the abse nce of reducing agents; (D) P fMT AP in 6 M GdmCl in the presence of 30 m M dithiothreitol. The inset shows the SDS/ PAGE pattern of PfMTAP and refolded PfMTAP. The samples were subjected to SDS/PAGE without boiling and under nonreducing conditions. Lane 1, molecular mass markers; lane 2, PfMTAP (5 lg); lane 3, refolded PfMTAP (5 lg). Fig. 4. GdmCl-induced fluorescence changes of PfMTAP. Fluores- cence changes are r eported a s k max by monitoring the shift in flu or- escence maximum wavelength, in 2 0 m M Tris/HCl, pH 7.4 in the presence (m) and absence (r)of30m M dithiothreitol. Th e spectra were recorded at 25 °C after 22 h incubation. Ó FEBS 2004 Methylthioadenosine phosphorylase from P. furiosus (Eur. J. Biochem. 271) 4841 Assignment of disulfide bridges In order to determine the arrangement of disulfide bridges in PfMTAP, the protein was alkylated with 4-vinylpyri- dine under denaturing conditions, with and without previous reduction with dithiotreitol, and then subjected to CNBr cleavage. This acidic cleavage was chosen in order to m inimize d isulfide interchan ge, which could occur at alkaline pH. Figure 6 reports the a mino acid sequence of PfMTAP and shows the position of cysteine residues and the CNBr peptides. The peptide elution patterns are shown in Fig. 7. Peaks corresponding to CB5 (with Cys195) and CB2 (with Cys130 and Cys156) present in the protein alkylated after reduction (A), are c ompletely absent in the nonreduced protein (B), whereas the peak corresponding to CB6 (with Cys246 and Cys248) is still present although in lower amount. In the reduced sample (A), the large peak CB1 contained the N-terminal peptide, which d oes not contain Cys residues. In the nonreduced sample (B), peptides CB2, CB5, and CB6 were found in the same peak, coeluting with the large N-terminal peptide. The a mino acid sequence indicated that CB2 and CB5 were present in equimolar amounts, and that Cys130 and Cys195 were not alkylated. This result suggested t hat the two peptides are connected by a disulfide bridge between Cys130 and Cys195. Peptide CB6 containing Cys246 and Cys248 was also foun d in the late large peak, but not in stoichiometric amounts with the previous peptides. This finding was most probably due to a carry over by the other large size peptides. Under nonreducing conditions ( B) CB6 w as found at the same position as in the digestion of the reduced protein (A). To investigate the oxidation state of the two cysteine residues in positions 246 and 248, CB6 w as analyzed by MALDI-TOF/MS. Following direct alkylation, the molecular mass of the peptide did not reveal the incorpor- ation o f vinylpiridine residues (3777.17 Da), whereas, when the peptide was alkylated after the reducing step, the molecular mass of the peptide increased by 212.2 to 3989.11 Da, indicating the addition of two vinylpiridine groups and demonstrating t hat Cys246 and Cys248 present in the peptide are involved in a disulfide bridge. It h as to be noted that these two cysteines are separated by a sin gle amino acid residue, however, t his residue is a glycine which, due to its s mall size and i ts conformational freedom, c ould allow the formation of a disulfide bridge between the two adjacent cysteines. Oxidized CXC sequences are rare in nature. The few examples reported in the literature include CSC in Mengo virus coat protein [48], CDC in Bac illus Ak.1 protease [49] and CTC in chaperone Hsp33 from E. coli [50]. More recently it has been reported that a d isulfide with the same CGC sequence found in PfMTAP is present near t he C-terminus of a yeast thiol oxidase, and it has been postulated that the enzyme could take a dvantage of a relatively strained CXC disulfide to perform efficient oxidation [51]. I t has to be pointed out that i n a ll the listed CXC disulfides, the X residue is small and therefore, the Gly residue found in PfM TAP certainly fits this pattern. One m ay ask what i s the structural function in PfM TAP of the disulfide bridge Cys246–Cys248 that links two cysteine residues so near each other in the sequence? The observations that (a) this disulfide is localized in the C-terminal region of PfMTAP; (b) the C-terminal as well as the N-terminal r egion of mesop hilic proteins are u sually highly flexible and disordered and are tho ught to be the first Fig. 6. Amino acid sequence alignment of PfMTAP and hMTAP. Asterisks indicate the cysteine res idues. Lines above the sequence indicate the expected CNBr peptides. Fig. 7. HPLC of peptides derived from CNBr cleavage of PfMTAP alkylated with (A) and w ithout (B) previous reduction with dithiothreitol. CNBr peptides were id ent ified by amino acid se quenci ng and discussed in the t ext. 4842 G. Cacciapuoti et al. (Eur. J. Biochem. 271) Ó FEBS 2004 portion of the protein which undergoes denaturation at high temperature [52]; and (c) two recently discovered h ypothet- ical MTAPs from the hype rthermophilic Archaea S. solfa- taricus (SS02343) a nd P. furiosus (PF0853) contain cysteine residues localized at the same positions of PfMTAP, suggest the hypothesis that a disulfide bond in the C-terminus of PfMTAP migh t increase the stability of this protein region. It is also interesting to note that the five cysteine residues of PfMTAP are well conserved i n both hypothetical M TAPs, suggesting that a s imilar disulfide bridge pattern may be present in these proteins and that t hese covalent links could represent the molecular strategy of thermal stabilization adopted by MTAP from hypertherm ophilic so urces. The high sequence identity (52%) between human and P. furiosus MTAP (Fig. 6) a llowed us t o make a sequence– structure mapping of the P. furiosus enzyme utilizing as a template the known three-dimensional structure of the human enzyme [44]. The obtained data by Swiss Pdb V iewer ( SPDBD 37 B 2000 program ) give a good support to the validity of our experimental results showing that, in the modeled P fMTAP structure, only the pairs C ys130-Cys195 and Cys246-Cys248 are at a distance compatible with a disulfide bond (Ca atoms) (5.88 A ˚ and 5.65 A ˚ , r espectively), while Cys156 is too far from all other cysteines to form t his type of link. Moreover, i n t he modeled structure, the disulfide Cys13-Cys195 appears buried. Disulfide bonds are a typical feature of secr etory proteins and a re considered to contr ibute signifi cantly to their overall stability [53]. In contrast, in intracellular p roteins from well- known organisms, because of the reductive chem ical environment inside the cells [54], the presence of these covalent links is limited to prote ins involved in the mechanism of response to redox stress [55] or to proteins catalyzing oxidation–reduction processes [56]. The availab- ility of m any completely sequenced hyperthermophilic genomes has indicated that cysteine residues, in spite of their high sensitivity to oxidation at high temperature [57], are present in remarkable amounts in hyperthermophilic proteins. In these proteins, t hese thermolabile residues are probably protected against thermal inactivation by being buried in the protein interior o r b y their involvement in specific stabilizing interactions such as metal liganding or disulfide bridges [ 4]. T he presence of d isulfide bonds within several archaeal and thermophilic genomes has been postulated, taking into account the results of a recent computational s tudy based on t he combination of g enomic data with protein structure [13]. Moreover, the increasing number o f solved c rystallographic structures h as highlighted the presence of disulfide bonds in several h yperthermophilic proteins [10–12,14]. The results reported here on multiple disulfide bonds in PfMTAP add a new example, to the few present in t he literature, of an intracellular hyperthermo- philic protein with disulfide bonds and argue strongly that intracellular disulfides could represent a significant mech- anism to ac hieve superior le vels of the rmostability. References 1. Kumar, S. & Nussinov, R . (2001) How do thermophilic proteins deal with heat? Cell Mol. Life Sci. 58, 1216–1233. 2. Ladenstein, R. & An tranikian, G . (1998) Prote ins from hyperthermophiles: stability and enzymatic catalysis close to the boiling point of water. In Advances in Biochemical Engineering/ Biotechnology (Sheper, T., ed.), Vol. 612. S pringer-Verlag, Berlin. 3. Sterner, R. & Lieb l, W. ( 2001) T hermophilic a daptatio n of p ro- teins. Crit. Rev. B iochem. Mol. B iol. 36, 39–106. 4. Vieille, C. & Zeikus, G.J. (2001) Hyperthermophilic enzymes: sources, uses and molecular me chanisms for thermostability. Microbiol. Mol. Biol. Rev. 65, 1–43. 5. Niehaus, F., Bertoldo, C., K ahler, M. & Antranikian, G. (1999) Extremophiles a s a source of n ovel e nzymes for i ndustrial appli- cation. Appl. M icrobiol. Biotechnol. 51, 711–729. 6. Zeikus, J.G., Vieille, C. & Savchenko, A. (1998) Thermozymes: biotechnology and structure–fun ction relationships. Extremo- philes 2, 179–183. 7. Kumar, S., Tsai, C.J. & Nussinov, R. (2000) Facto r s enhancing protein t hermostability. Protein Eng. 13 , 179–191. 8. Vogt, G., Woell, S. & Argos, P . (1977) Protein t hermal sta bility, hydrogen bonds and ion p airs. J. Mol . Biol. 269, 631– 643. 9. Xiao, L. & Honig, B. (1999) Electrostatic contributions to the stability of hyperthermophilic proteins. J. Mol. Biol. 289, 1435– 1444. 10. Appleby, T.C., Mathews, I.I., Porcelli, M., Cacciapuoti, G. & Ealick, S.E. (2001) Three-dimensional structure of a hyper- thermophilic 5 ¢-deoxy-5¢-methylthioadenosine phosphorylase from Sulfolobus solfataricus. J. Biol. Chem. 42, 392 32–39242. 11. Choi, I G., Bang, W G., Kim, S H. & Yu, Y.G. (1999) Extremely thermostable s erine-type protease from Aquifex pyro- philus. M olecular cloning, ex pression an d characterization. J. Biol. Chem. 274, 881–888. 12. Cort,J.R.,Mariappan,S.V.,Kim,C Y.,Park,M.S.,Peat,T.S., Waldo, G. S., Terwillger, T.C. & Kennedy, M.A. ( 2001) Solution structure of Pyrobaculum aerophilum DsrC,anarchaealhomo- logue of the c-subunit of dissimilatory sulfite reductase. Eur. J. Biochem. 26 8, 5842–5849. 13. Mallick, P., Boutz, D.R., Eisenberg, D. & Yeates, T.O. (2002) Genomic evidence that the intrace llular proteins of archaeal microbes contain d isulfide bon ds. Proc. Natl Acad. Sci. 99, 9679– 9684. 14. Toth, E.A., Worby, C., Dixon, J.E., Goedken, E.R., Marqusee, S. & Yeates, T.O. (2000) The crystal structure o f adenylosuccinat e lyase from Pyrobaculum aerophilum re veals an intracellular protein with three disulfide bonds. J. Mol. Biol. 30 1, 433–450. 15. Williams-Ashman, H .G., Seidenfeld, J. & Galletti, P. ( 1982) Trends in the biochemical pharmacology of 5 ¢-d eoxy-5¢-methyl- thioadenosine. Biochem. Ph armacol. 31, 277–288. 16. Pegg, A.E. & Williams-Ashman, H.G. ( 1969) Phosphate-stimu- latedbreakdownof5¢-methylthioadeno sine by rat ventral pros- tate. Biochem. J. 115, 241–247. 17. Della R agione, F., Cartenı ` -Farina, M., Gragnianiello, V., Schet- tino, M.I. & Zappia, V. (1986) Purification a nd characterization of 5¢-deoxy-5¢-methylthioadenosine phosphorylase from human placenta. J. Biol. Chem. 261, 12324–12329. 18. Della Ragione, F., Oliva, A., Gragnianiello, V., Russo, G.L., Palumbo, R. & Zappia, V. (1990) Physicochemical and immunological studies on mammalian 5¢-deoxy-5¢-methylthio- adenosine phosphorylase. J. Bi ol. Chem. 265 , 6241–6246. 19. Toorchen, D. & Miller, R.L. (1991) Purification and character- ization o f 5 ¢-deoxy-5¢-methylthioadenosine (MTA) phosphorylase from human l iver. Biochem. Pharmacol. 41, 2 023–2030. 20. Cacciapuoti, G., Porcelli, M., Bertoldo, C., De Rosa, M. & Zappia, V. (1994) Purification and chara cterization of e xtremely thermophilic and thermostable 5¢-m ethylthioadenosine p hos- phorylase f rom the archaeon Sulfolobus solfataricus.Purine nucleodide phosphorylase activity a nd evidence for inte rsubunit disulfide bonds. J. Biol. Chem. 26 9, 24762–24769. 21. Cacciapuoti, G., Bertoldo, C. , B rio, A., Zappia, V. & Porcelli, M. (2003) Purification and characterization of 5¢-methylthioadenosine Ó FEBS 2004 Methylthioadenosine phosphorylase from P. furiosus (Eur. J. Biochem. 271) 4843 [...]... Cacciapuoti et al (Eur J Biochem 271) 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 phosphorylase from the hyperthermophilic archaeon Pyrococcus furiosus Substrate specificity and primary structure analysis Extremophiles 7, 159–168 Appleby, T.C., Erion, M.D & Ealick, S.E (1999) The structure of ˚ human 5¢-deoxy-5¢ -methylthioadenosine phosphorylase at 1.7 A resolution provides insights into substrate binding... Vos, W.M (1996) Isolation and characterization of hyperthermostable serine-protease, pyrolysin, and its gene from hyperthermophilic archaeon Pyrococcus furiosus J Biol Chem 271, 20426–20431 Robb, F., Maeder, D.L., Brown, J.R., Di Ruggiero, J., Stump, M.D., Yeh, R.K., Weiss, R.B & Dunn, D.M (2001) Genomic sequence of hyperthermophile Pyrococcus furiosus: implications for physiology and enzymology Methods... expression of the extremely thermostable glutamate dehydrogenase from the hyperthermophilic archaeon ES4 J Biol Chem 268, 17767–17774 Ó FEBS 2004 38 Hockney, R.C (1994) Recent developments in heterologous protein production in Escherichia coli Trends Biotechnol 12, 456–463 39 Chae, W.-G., Chan, D.C.K & Chang, C (1998) Facile synthesis of 5¢-deoxy- and 2¢,5¢-dideoxy-6-thiopurine nucleosides by nucleoside phosphorylases... Moracci, M., Pisani, F.M., Rossi, M & de Vos, W.M (1998) Molecular biology of hyperthermophilic Archaea Adv Biochem Eng Biotechnol 61, 87–115 Jorgensen, S., Vorgias, C.E & Antranikian, G (1997) Cloning, sequencing, characterization and expression of an extracellular a-amilase from the hyperthermophilic archaeon Pyrococcus furiosus, E coli and Bacillus subtilis J Biol Chem 272, 16335– 16342 Voorhorst,... A., De Rosa, M & Zappia, V (1993) S-Adenosylhomocysteine hydrolase from the thermophilic archaeon Sulfolobus solfataricus: purification, physico-chemical and immunological properties Biochim Biophys Acta 1164, 179–188 Bradford, M.M (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding Anal Biochem 72, 248–254 Cacciapuoti,... expressing purine nucleoside phosphorylase as a vector for tumour-directed gene theraphy Hum Gene Ther 10, 649–657 41 Cleland, W.W (1963) The kinetics of enzyme-catalyzed reactions with two or more substrates or products II Inhibition: nomenclature and theory Biochim Biophys Acta 67, 173–187 42 Koellner, G.L., Luic, M., Shugar, D., Saenger, W & Bzowska, A (1998) Crystal structure of the ternary complex of... Press, New York 46 Garbers, D.L (1978) Demonstration of 5¢ -methylthioadenosine phosphorylase activity in various rat tissues: some properties of the enzyme from rat lung Biochim Biophys Acta 523, 82–93 47 Shugart, L., Mahoney, L & Chastain, B (1981) Kinetic studies of Drosophila melanogaster methylthioadenosine nucleoside phosphorylase Int J Biochem 1, 559–564 48 Krishnaswamy, S & Rossmann, M.G (1990)... FAD-binding fold and intersubunit disulfide schuttle in the thiol oxidase Erv2p Nat Struct Biol 9, 61–67 52 Adams, M.W.W & Kelly, R.M (1994) Thermostability and thermoactivity of enzymes from hyperthermophilic Archaea Bioorg Med Chem 2, 659–667 53 Matsumura, M., Signor, G & Mathews, B.W (1989) Substantial increase of protein stability by multiple disulfide bonds Nature 342, 291–293 54 Gilbert, H.F (1990) Molecular... nucleoside phosphorylase ˚ at 3.2 A resolution J Biol Chem 265, 1812–1820 45 Zappia, V., Della Ragione, F., Pontoni, G., Gragnianiello, V & Cartenı` -Farina, M (1988) Human 5¢-deoxy-5¢ -methylthioadenosine phosphorylase: kinetic studies and catalytic mechanism In Progress in Polyamine Research (Zappia, V & Pegg, A.E., eds), pp 165–177 Plenum Press, New York 46 Garbers, D.L (1978) Demonstration of 5¢ -methylthioadenosine. .. Purine nucleoside phosphorylases: properties, functions and clinical aspects Pharmacol Ther 88, 349–425 Bzowska, A (2002) Calf spleen purine nucleoside phosphorylase: complex kinetic mechanism, hydrolysis of 7-methylguanosine and oligomeric state in solution Biochim Biophys Acta 1596, 293–317 Pugmire, M.J & Ealick, S.E (2002) Structural analysis reveals two distinct families of nucleoside phosphorylases . Methylthioadenosine phosphorylase from the archaeon Pyrococcus furiosus Mechanism of the reaction and assignment of disulfide. CNR, Naples, Italy The extremely heat-stable 5¢ -methylthioadenosine phos- phorylase from the hyperthermophilic archaeon Pyrococcus furiosus was cloned,