Crystal structure of the purine nucleoside phosphorylase (PNP) from Cellulomonas sp. and its implication for the mechanism of trimeric PNPs

Perspectives and challenges in developing small molecules targeting purine nucleoside phosphorylase

As a cytosolic enzyme involved in the purine salvage pathway metabolism, purine nucleoside phosphorylase (PNP) plays an important role in a variety of cellular functions but also in immune system, including cell growth, apoptosis and cancer development and progression. Based on its T-cell targeting profile, PNP is a potential target for the treatment of some malignant T-cell proliferative cancers including lymphoma and leukemia, and some specific immunological diseases. Numerous small-molecule PNP inhibitors have been developed so far. However, only Peldesine, Forodesine and Ulodesine have entered clinical trials and exhibited some potential for the treatment of T-cell leukemia and gout. The most recent direction in PNP inhibitor development has been focused on PNP small-molecule inhibitors with better potency, selectivity, and pharmacokinetic property. In this perspective, considering the structure, biological functions, and disease relevance of PNP, we highlight the recent research progress in PNP small-molecule inhibitor development and discuss prospective strategies for designing additional PNP therapeutic agents.

The molecular structure of Schistosoma mansoni PNP isoform 2 provides insights into the nucleoside selectivity of PNPs

Purine nucleoside phosphorylases (PNPs) play an important role in the blood fluke parasite Schistosoma mansoni as a key enzyme of the purine salvage pathway. Here we present the structural and kinetic characterization of a new PNP isoform from S. mansoni, SmPNP2. Thermofluorescence screening of different ligands suggested cytidine and cytosine are potential ligands. The binding of cytosine and cytidine were confirmed by isothermal titration calorimetry, with a K_D of 27 μM for cytosine, and a K_M of 76.3 μM for cytidine. SmPNP2 also displays catalytic activity against inosine and adenosine, making it the first described PNP with robust catalytic activity towards both pyrimidines and purines. Crystal structures of SmPNP2 with different ligands were obtained and comparison of these structures with the previously described S. mansoni PNP (SmPNP1) provided clues for the unique capacity of SmPNP2 to bind pyrimidines. When compared with the structure of SmPNP1, substitutions in the vicinity of SmPNP2 active site alter the architecture of the nucleoside base binding site thus permitting an alternative binding mode for nucleosides, with a 180° rotation from the canonical binding mode. The remarkable plasticity of this binding site enhances our understanding of the correlation between structure and nucleotide selectivity, thus suggesting new ways to analyse PNP activity.

Towards understanding the E. coli PNP binding mechanism and FRET absence between E. coli PNP and formycin A

The aim of this study is threefold: (1) augmentation of the knowledge of the E. coli PNP binding mechanism; (2) explanation of the previously observed 'lack of FRET' phenomenon and (3) an introduction of the correction (modified method) for FRET efficiency calculation in the PNP-FA complexes. We present fluorescence studies of the two E. coli PNP mutants (F159Y and F159A) with formycin A (FA), that indicate that the aromatic amino acid is indispensable in the nucleotide binding, additional hydroxyl group at position 159 probably enhances the strength of binding and that the amino acids pair 159-160 has a great impact on the spectroscopic properties of the enzyme. The experiments were carried out in hepes and phosphate buffers, at pH7 and 8.3. Two methods, a conventional and a modified one, that utilizes the dissociation constant, for calculations of the energy transfer efficiency (E) and the acceptor-to-donor distance (r) between FA and the Tyr (energy donor) were employed. Total difference spectra were calculated for emission spectra (λ_ex 280nm, 295nm, 305nm and 313nm) for all studied systems. Time-resolved techniques allowed to conclude the existence of a specific structure formed by amino acids at positions 159 and 160. The results showed an unexpected pattern change of FRET in the mutants, when compared to the wild type enzyme and a probable presence of a structure created between 159 and 160 residue, that might influence the binding efficiency. Additionally, we confirmed the indispensable role of the modification of the FRET efficiency (E) calculation on the fraction of enzyme saturation in PNP-FA systems.

Computer Simulations Reveal Substrate Specificity of Glycosidic Bond Cleavage in Native and Mutant Human Purine Nucleoside Phosphorylase

Purine nucleoside phosphorylase (PNP) catalyzes the reversible phosphorolysis of purine ribonucleosides and 2'-deoxyribonucleosides, yielding the purine base and (2'-deoxy)ribose 1-phosphate as products. While this enzyme has been extensively studied, several questions with respect to the catalytic mechanism have remained largely unanswered. The role of the phosphate and key amino acid residues in the catalytic reaction as well as the purine ring protonation state is elucidated using density functional theory calculations and extensive empirical valence bond (EVB) simulations. Free energy surfaces for adenosine, inosine, and guanosine are fitted to ab initio data and yield quantitative agreement with experimental data when the surfaces are used to model the corresponding enzymatic reactions. The cognate substrates 6-aminopurines (inosine and guanosine) interact with PNP through extensive hydrogen bonding, but the substrate specificity is found to be a direct result of the electrostatic preorganization energy along the reaction coordinate. Asn243 has previously been identified as a key residue providing substrate specificity. Mutation of Asn243 to Asp has dramatic effects on the substrate specificity, making 6-amino- and 6-oxopurines equally good as substrates. The principal effect of this particular mutation is the change in the electrostatic preorganization energy between the native enzyme and the Asn243Asp mutant, clearly favoring adenosine over inosine and guanosine. Thus, the EVB simulations show that this particular mutation affects the electrostatic preorganization of the active site, which in turn can explain the substrate specificity.

Targeting the protein backbone with aryl halides: systematic comparison of halogen bonding and π⋯π interactions using N-methylacetamide

Systematic plane scans reveal a seamless transition from σ-hole interactions with the carbonyl oxygen to interactions with the amide π-electrons at increasing distances.

Development of a new HPLC method using fluorescence detection without derivatization for determining purine nucleoside phosphorylase activity in human plasma

Purine nucleoside phosphorylase (PNP) activity is involved in cell survival and function, since PNP is a key enzyme in the purine metabolic pathway where it catalyzes the phosphorolysis of the nucleosides to the corresponding nucleobases. Its dysfunction has been found in relevant pathological conditions (such as inflammation and cancer), so the detection of PNP activity in plasma could represent an attractive marker for early diagnosis or assessment of disease progression. Thus the aim of this study was to develop a simple, fast and sensitive HPLC method for the determination of PNP activity in plasma. The separation was achieved on a Phenomenex Kinetex PFP column using 0.1% formic acid in water and methanol as mobile phases in gradient elution mode at a flow rate of 1ml/min and purine compounds were detected using UV absorption and fluorescence. The analysis was fast since the run was achieved within 13min. This method improved the separation of the different purines, allowing the UV-based quantification of the natural PNP substrates (inosine and guanosine) or products (hypoxanthine and guanine) and its subsequent metabolic products (xanthine and uric acid) with a good precision and accuracy. The most interesting innovation is the simultaneous use of a fluorescence detector (excitation/emission wavelength of 260/375nm) that allowed the quantification of guanosine and guanine without derivatization. Compared with UV, the fluorescence detection improved the sensitivity for guanine detection by about 10-fold and abolished almost completely the baseline noise due to the presence of plasma in the enzymatic reaction mixture. Thus, the validated method allowed an excellent evaluation of PNP activity in plasma which could be useful as an indicator of several pathological conditions.

8-Azapurines as isosteric purine fluorescent probes for nucleic acid and enzymatic research

The 8-azapurines, and their 7-deaza and 9-deaza congeners, represent a unique class of isosteric (isomorphic) analogues of the natural purines, frequently capable of substituting for the latter in many biochemical processes. Particularly interesting is their propensity to exhibit pH-dependent room-temperature fluorescence in aqueous medium, and in non-polar media. We herein review the physico-chemical properties of this class of compounds, with particular emphasis on the fluorescence emission properties of their neutral and/or ionic species, which has led to their widespread use as fluorescent probes in enzymology, including enzymes involved in purine metabolism, agonists/antagonists of adenosine receptors, mechanisms of catalytic RNAs, RNA editing, etc. They are also exceptionally useful fluorescent probes for analytical and clinical applications in crude cell homogenates.

Different expression profiles of Interleukin 11 (IL-11), Intelectin (ITLN) and Purine nucleoside phosphorylase 5a (PNP 5a) in crucian carp (Carassius auratus gibelio) in response to Cyprinid herpesvirus 2 and Aeromonas hydrophila

Interleukin 11 (IL-11), Intelectin (ITLN) and Purine nucleoside phosphorylase 5a (PNP5a) play important roles in innate immunity. In a previous study to identify differentially expressed immune-related genes, suppression subtractive hybridization (SSH) assay was used to characterize differentially expressed genes in crucian carp (Carassius auratus gibelio) infected with Cyprinid herpesvirus 2 (CyHV-2) in which IL-11, ITLN and PNP5a were identified to be the three most significantly up-regulated genes (Xu et al., Archives of Virology, 2014, http://dx.doi.org/10.1007/s00705-014-2011-9). In this study, the complete open reading frames (ORF) of IL-11, ITLN and PNP5a genes were cloned and sequenced. The full-length cDNAs of the three genes contained an ORF of 597, 945 and 882 bp, encoding a polypeptide of 198, 314 and 293 amino acids, respectively. Phylogenetic analysis indicated that the three genes shared high homology to other bony fish species including Zebrafish. Interestingly, the ITLN gene of crucian carp lacked a 10 aa peptide that was found in the C-terminal of other fish species. A real-time RT-PCR assay was developed to quantitatively examine their tissue distribution. We found that IL-11, ITLN and PNP5a were expressed at low levels in all of the tissues examined. To monitor the response of these genes to CyHV- 2 or Aeromonas hydrophila (A. hydrophila) infection, we determined the expression level of IL-11, ITLN and PNP5a at different time points after infection in kidney. Significant up-regulation of IL-11, ITLN and PNP5a was only observed 72 h post-CyHV-2 injection (P < 0.01), whereas significant up-regulation was observed as early as 6 h after infection with A. hydrophila (P < 0.01). Our results demonstrated that host innate immune response to CyHV-2, at least in which IL-11, ITLN and PNP5a were involved, was slow in comparison to that induced by A. hydrophila. It suggested that CyHV-2 might suppress host innate response during early infection. The lack of a C-terminal peptide of crucian carp ITLN gene implied a possible functional difference of this gene during evolution, which merit further investigation.

Structural basis of the substrate specificity of Bacillus cereus adenosine phosphorylase

Purine nucleoside phosphorylases catalyze the phosphorolytic cleavage of the glycosidic bond of purine (2'-deoxy)nucleosides, generating the corresponding free base and (2'-deoxy)-ribose 1-phosphate. Two classes of PNPs have been identified: homotrimers specific for 6-oxopurines and homohexamers that accept both 6-oxopurines and 6-aminopurines. Bacillus cereus adenosine phosphorylase (AdoP) is a hexameric PNP; however, it is highly specific for 6-aminopurines. To investigate the structural basis for the unique substrate specificity of AdoP, the active-site mutant D204N was prepared and kinetically characterized and the structures of the wild-type protein and the D204N mutant complexed with adenosine and sulfate or with inosine and sulfate were determined at high resolution (1.2-1.4 Å). AdoP interacts directly with the preferred substrate through a hydrogen-bond donation from the catalytically important residue Asp204 to N7 of the purine base. Comparison with Escherichia coli PNP revealed a more optimal orientation of Asp204 towards N7 of adenosine and a more closed active site. When inosine is bound, two water molecules are interposed between Asp204 and the N7 and O6 atoms of the nucleoside, thus allowing the enzyme to find alternative but less efficient ways to stabilize the transition state. The mutation of Asp204 to asparagine led to a significant decrease in catalytic efficiency for adenosine without affecting the efficiency of inosine cleavage.

Crystal structure and molecular dynamics studies of purine nucleoside phosphorylase from Mycobacterium tuberculosis associated with acyclovir

Consumption has been a scourge of mankind since ancient times. This illness has charged a high price to human lives. Many efforts have been made to defeat Mycobacterium tuberculosis (Mt). The M. tuberculosis purine nucleoside phosphorylase (MtPNP) is considered an interesting target to pursuit new potential inhibitors, inasmuch it belongs to the purine salvage pathway and its activity might be involved in the mycobacterial latency process. Here we present the MtPNP crystallographic structure associated with acyclovir and phosphate (MtPNP:ACY:PO(4)) at 2.10 Å resolution. Molecular dynamics simulations were carried out in order to dissect MtPNP:ACY:PO(4) structural features, and the influence of the ligand in the binding pocket stability. Our results revealed that the ligand leads to active site lost of stability, in agreement with experimental results, which demonstrate a considerable inhibitory activity against MtPNP (K(i) = 150 nM). Furthermore, we observed that some residues which are important in the proper ligand's anchor into the human homologous enzyme do not present the same importance to MtPNP. Therewithal, these findings contribute to the search of new specific inhibitors for MtPNP, since peculiarities between the mycobacterial and human enzyme binding sites have been identified, making a structural-based drug design feasible.

Forodesine: review of preclinical and clinical data

Purine nucleoside phosphorylase (PNP) is an important catalytic enzyme in the purine salvage pathway; its deficiency is associated with T-cell lymphopenia and with humoral deficiency. This clinical observation led to the investigation of PNP inhibitors and their possible clinical application in the management of hematologic malignancies, notably those of T-cell lineage. Forodesine is the most potent of the PNP inhibitors. Its effect appears to be linked to increased 2 -deoxyguanosine levels in plasma, which in turn is converted to 2 -deoxyguanosine triphosphate in target cells and disrupts DNA synthesis. Several preclinical studies have shown forodesine's effect against lymphocytes in vitro and in vivo, and these findings have led to several Phase I/II studies in patients with lymphoid neoplasms. Early clinical trials show that forodesine has promise as a single agent for the treatment of relapsed/refractory hematologic malignancies, and combination therapies might be warranted to improve clinical results.

Structure of grouper iridovirus purine nucleoside phosphorylase

Purine nucleoside phosphorylase (PNP) catalyzes the reversible phosphorolysis of purine ribonucleosides to the corresponding free bases and ribose 1-phosphate. The crystal structure of grouper iridovirus PNP (givPNP), corresponding to the first PNP gene to be found in a virus, was determined at 2.4 A resolution. The crystals belonged to space group R3, with unit-cell parameters a = 193.0, c = 105.6 A, and contained four protomers per asymmetric unit. The overall structure of givPNP shows high similarity to mammalian PNPs, having an alpha/beta structure with a nine-stranded mixed beta-barrel flanked by a total of nine alpha-helices. The predicted phosphate-binding and ribose-binding sites are occupied by a phosphate ion and a Tris molecule, respectively. The geometrical arrangement and hydrogen-bonding patterns of the phosphate-binding site are similar to those found in the human and bovine PNP structures. The enzymatic activity assay of givPNP on various substrates revealed that givPNP can only accept 6-oxopurine nucleosides as substrates, which is also suggested by its amino-acid composition and active-site architecture. All these results suggest that givPNP is a homologue of mammalian PNPs in terms of amino-acid sequence, molecular mass, substrate specificity and overall structure, as well as in the composition of the active site.

Adenosine binding to low-molecular-weight purine nucleoside phosphorylase: the structural basis for recognition based on its complex with the enzyme fromSchistosoma mansoni

Schistosomes are unable to synthesize purinesde novoand depend exclusively on the salvage pathway for their purine requirements. It has been suggested that blockage of this pathway could lead to parasite death. The enzyme purine nucleoside phosphorylase (PNP) is one of its key components and molecules designed to inhibit the low-molecular-weight (LMW) PNPs, which include both the human and schistosome enzymes, are typically analogues of the natural substrates inosine and guanosine. Here, it is shown that adenosine both binds toSchistosoma mansoniPNP and behaves as a weak micromolar inhibitor of inosine phosphorolysis. Furthermore, the first crystal structures of complexes of an LMW PNP with adenosine and adenine are reported, together with those with inosine and hypoxanthine. These are used to propose a structural explanation for the selective binding of adenosine to some LMW PNPs but not to others. The results indicate that transition-state analogues based on adenosine or other 6-amino nucleosides should not be discounted as potential starting points for alternative inhibitors.

Preliminary crystallographic studies of purine nucleoside phosphorylase from the cariogenic pathogen Streptococcus mutans

The punA gene of the cariogenic pathogen Streptococcus mutans encodes purine nucleoside phosphorylase (PNP), which is a pivotal enzyme in the nucleotide-salvage pathway, catalyzing the phosphorolysis of purine nucleosides to generate purine bases and alpha-ribose 1-phosphate. In the present work, the PNP protein was expressed in Escherichia coli strain BL21 (DE3) in a soluble form at a high level. After purification of the PNP enzyme, the protein was crystallized using the sitting-drop vapour-diffusion technique; the crystals diffracted to 1.6 A resolution at best. The crystals belonged to space group H3, with unit-cell parameters a = b = 113.0, c = 60.1 A.

Mechanism of thiol-supported arsenate reduction mediated by phosphorolytic-arsenolytic enzymes: I. The role of arsenolysis

Several mammalian enzymes catalyzing the phosphorolytic-arsenolytic cleavage of their substrates (thus yielding arsenylated metabolites) have been shown to facilitate reduction of arsenate (AsV) to the more toxic arsenite (AsIII) in presence of their substrate and a thiol. These include purine nucleoside phosphorylase (PNP), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and glycogen phosphorylase-a (GPa). In this work, we tested further enzymes, the bacterial phosphotransacetylases (PTAs) and PNP, for AsV reduction. The PTAs, which arsenolytically cleave acetyl-CoA producing acetyl-arsenate, were compared with GAPDH, which can also form acetyl-arsenate by arsenolysis of its nonphysiological substrate, acetyl-phosphate. As these enzymes also mediated AsV reduction, we can assert that facilitation of thiol-dependent AsV reduction may be a general property of enzymes that catalyze phosphorolytic-arsenolytic reactions. Because with all such enzymes arsenolysis is obligatory for AsV reduction, we analyzed the relationship between these two processes in presence of various thiol compounds, using PNP. Although no thiol influenced the rate of PNP-catalyzed arsenolysis, all enhanced the PNP-mediated AsV reduction, albeit differentially. Furthermore, the relative capacity of thiols to support AsV reduction mediated by PNP, GPa, PTA, and GAPDH apparently depended on the type of arsenylated metabolites (i.e., arsenate ester or anhydride) produced by these enzymes. Importantly, AsV reduction by both acetyl-arsenate-producing enzymes (i.e., PTA and GAPDH) exhibited striking similarities in responsiveness to various thiols, thus highlighting the role of arsenylated metabolite formation. This observation, together with the finding that PNP-mediated AsV reduction lags behind the PNP-catalyzed arsenolysis lead to the hypothesis that arsenolytic enzymes promote reduction of AsV by forming arsenylated metabolites which are more reducible to AsIII by thiols than inorganic AsV. This hypothesis is evaluated in the adjoining paper.

Tryptophan-free human PNP reveals catalytic site interactions

Human purine nucleoside phosphorylase (PNP) is a homotrimer, containing three nonconserved tryptophan residues at positions 16, 94, and 178, all remote from the catalytic site. The Trp residues were replaced with Tyr to produce Trp-free PNP (Leuko-PNP). Leuko-PNP showed near-normal kinetic properties. It was used (1) to determine the tautomeric form of guanine that produces strong fluorescence when bound to PNP, (2) for thermodynamic binding analysis of binary and ternary complexes with substrates, (3) in temperature-jump perturbation of complexes for evidence of multiple conformational complexes, and (4) to establish the ionization state of a catalytic site tyrosine involved in phosphate nucleophile activation. The (13)C NMR spectrum of guanine bound to Leuko-PNP, its fluorescent properties, and molecular orbital electronic transition analysis establish that its fluorescence originates from the lowest singlet excited state of the N1H, 6-keto, N7H guanine tautomer. Binding of guanine and phosphate to PNP and Leuko-PNP are random, with decreased affinity for formation of ternary complexes. Pre-steady-state kinetics and temperature-jump studies indicate that the ternary complex (enzyme-substrate-phosphate) forms in single binding steps without kinetically significant protein conformational changes as monitored by guanine fluorescence. Spectral changes of Leuko-PNP upon phosphate binding establish that the hydroxyl of Tyr88 is not ionized to the phenolate anion when phosphate is bound. A loop region (residues 243-266) near the purine base becomes highly ordered upon substrate/inhibitor binding. A single Trp residue was introduced into the catalytic loop of Leuko-PNP (Y249W-Leuko-PNP) to determine effects on catalysis and to introduce a fluorescence catalytic site probe. Although Y249W-Leuko-PNP is highly fluorescent and catalytically active, substrate binding did not perturb the fluorescence. Thermodynamic boxes, constructed to characterize the binding of phosphate, guanine, and hypoxanthine to native, Leuko-, and Y249W-Leuko-PNPs, establish that Leuko-PNP provides a versatile protein scaffold for introduction of specific Trp catalytic site probes.

Fluorescence studies of calf spleen purine nucleoside phosphorylase (PNP) complexes with guanine and 9-deazaguanine

Interactions of trimeric calf spleen purine nucleoside phosphorylase (PNP) with guanine (Gua) and its analogue, 9-deazaguanine (9-deaza-Gua), were studied by means of the steady-state fluorescence. The aim was to test the hypothesis that the enzyme stabilizes the anionic form of purine, inferred previously from the unusual increase of fluorescence observed after binding of guanine by calf spleen PNP. We have found that the dissociation constants obtained form titration experiments are in fact pH-independent in the range 7.0-10.25 for both PNP/Gua and PNP/9-deaza-Gua complexes. In particular, at pH 7.0 we found Kd = 0.12 +/- 0.02 micro M for Gua and 0.16 +/- 0.01 micro M for 9-deaza-Gua, while at the conditions where there is more than 40% of the anionic form the respective values were Kd = 0.15 +/- 0.01 micro M for Gua (pH 9.0) and 0.25 +/- 0.02 micro M for 9-deaza-Gua (pH 10.25). Hence, the enzyme does not prefer binding of anionic forms of these ligands in respect to the neutral ones. This result questions the involvement of the anionic forms in the reaction catalyzed by trimeric PNPs, and contradicts the hypothesis of a strong hydrogen bond formation between the enzyme Asn 243 residue and the purine N7 position.

Pharmacology and mechanism of action of forodesine, a T-cell targeted agent

Purine nucleoside phosphorylase (PNP) was recognized more than 30 years ago as a potential target for the treatment of patients with T-cell malignancies when an inherited deficiency of PNP was reported to be associated with a profound T-cell lymphopenia. The biochemical basis for this T-cell deficiency was subsequently shown to be related to the accumulation of plasma 2'-deoxyguanosine (dGuo) and intracellular dGuo triphosphate (dGTP). These observations have led to a search for PNP inhibitors that would be useful clinically in the management of T cell-derived malignancies. The most potent inhibitor of PNP described to date is forodesine, a rationally designed, transition-state analogue inhibitor. The preclinical and clinical pharmacology of forodesine showed its effectiveness in inhibiting PNP and augmenting dGuo levels in plasma. Increased dGTP concentrations in leukemia cells of different lineages provides strong support for the potential use of this agent in the treatment of patients with hematologic malignancies of both T- and B-cell origin.

Biochemical and structural characterization of mammalian-like purine nucleoside phosphorylase from the Archaeon Pyrococcus furiosus

We report here the characterization of the first mammalian-like purine nucleoside phosphorylase from the hyperthermophilic archaeon Pyrococcus furiosus (PfPNP). The gene PF0853 encoding PfPNP was cloned and expressed in Escherichia coli and the recombinant protein was purified to homogeneity. PfPNP is a homohexamer of 180 kDa which shows a much higher similarity with 5'-deoxy-5'-methylthioadenosine phosphorylase (MTAP) than with purine nucleoside phosphorylase (PNP) family members. Like human PNP, PfPNP shows an absolute specificity for inosine and guanosine. PfPNP shares 50% identity with MTAP from P. furiosus (PfMTAP). The alignment of the protein sequences of PfPNP and PfMTAP indicates that only four residue changes are able to switch the specificity of PfPNP from a 6-oxo to a 6-amino purine nucleoside phosphorylase still maintaining the same overall active site organization. PfPNP is highly thermophilic with an optimum temperature of 120 degrees C and is characterized by extreme thermodynamic stability (T(m), 110 degrees C that increases to 120 degrees C in the presence of 100 mm phosphate), kinetic stability (100% residual activity after 4 h incubation at 100 degrees C), and remarkable SDS-resistance. Limited proteolysis indicated that the only proteolytic cleavage site is localized in the C-terminal region and that the C-terminal peptide is not necessary for the integrity of the active site. By integrating biochemical methodologies with mass spectrometry we assigned three pairs of intrasubunit disulfide bridges that play a role in the stability of the enzyme against thermal inactivation. The characterization of the thermal properties of the C254S/C256S mutant suggests that the CXC motif in the C-terminal region may also account for the extreme enzyme thermostability.

Novel purine nucleoside analogues for T-cell-lineage acute lymphoblastic leukaemia and lymphoma

Purine nucleoside phosphorylase (PNP) deficiency is a rare, inherited immunodeficiency disorder in which the specific molecular defect was identified. Clinically, a lack of PNP manifests as profound T-cell deficiency with minor or variable changes in the humoral system. Biochemically, the absence of PNP results in an increase in plasma deoxyguanosine (dGuo) and a T-cell-specific increase in intracellular deoxyguanosine triphosphate (dGTP). This observation has been the impetus for the search for either inhibitors of the enzyme or PNP-resistant dGuo analogues as potential anti-T-cell-lineage agents over the past 30 years. Forodesine (an inhibitor of PNP) and nelarabine (a PNP-resistant dGuo analogue) proved to be T-cell selective when tested in clinic. This review summarises the preclinical, clinical and pharmacokinetic investigations with these novel agents.

Towards the mechanism of trimeric purine nucleoside phosphorylases: Stopped-flow studies of binding of multisubstrate analogue inhibitor — 2-amino-9-[2-(phosphonomethoxy)ethyl]-6-sulfanylpurine

The binding of multisubstrate analogue inhibitor - 2-amino-9-[2-(phosphonomethoxy)ethyl]-6-sulfanylpurine (PME-6-thio-Gua) to purine nucleoside phosphorylase from Cellulomonas sp. at 20 degrees C, in 20 mM Hepes buffer with ionic strength adjusted to 50 mM using KCl, at several pH values between 6.5 and 8.2, was investigated using a stopped-flow spectrofluorimeter. The kinetic transients registered after mixing a protein solution with ligand solutions of different concentrations were simultaneously fitted by several association reaction models using nonlinear least-squares procedure based on numerical integration of the chemical kinetic equations appropriate for given model. It is concluded that binding of a PME-6-thio-Gua molecule by each of the binding sites is sufficiently well described by one-step process, with a model assuming interacting binding sites being more probable than a model assuming independent sites. The association rate constants derived from experimental data, assuming one step binding and independent sites, are decreasing with an increase in pH, changing from 30 to 6 microM(-1)s(-1) per binding site. The dissociation rate constants are in the range of 1-3 s(-1), and they are rather insensitive of changes in pH. Interestingly, for each pH value, the one-step binding model with interacting sites results in the association rate constant per site 1.5-4 times smaller for the binding of the first ligand molecule than that for the binding of the second one. Decrease of association constants with pH indicate that the enzyme does not prefer binding of the naturally occurring anionic form of the 6-thioguanine ring (pK(a) 8.7) resulting from a dissociation of N(1)-H. This finding supports the mechanism in which hydrogen bond interaction of N(1)-H with Glu204 (Glu 201 in mammalian PNPs) is crucial in the catalytic process. Results obtained also indicate that, in contrast to transition-state analogues, for which binding is followed by a conformational change, binding of multisubstrate analogue inhibitors to trimeric PNPs is a one-step process.

Probing the mechanism of purine nucleoside phosphorylase by steady-state kinetic studies and ligand binding characterization determined by fluorimetric titrations

Reversible reaction catalyzed by trimeric purine nucleoside phosphorylase (PNP) from Cellulomonas sp. with typical and non-typical substrates, including product inhibition patterns of both reaction directions, and interactions of the enzyme with bisubstrate analogue inhibitors, were investigated by the steady-state kinetic methods and fluorimetric titrations. The ligand chromophores exist most probably as neutral species, and not N(1)-H monoanions, in the complex with PNP, as shown by determination of inhibition constants vs. pH. This supports the mechanism in which hydrogen bond interaction of N(1)-H with Glu204 is crucial in the catalytic process. Stoichiometry of ligand binding, with possible exception of hypoxanthine, is three molecules per enzyme trimer. Kinetic experiments show that in principle the Michaelis-Menten model could not properly describe the reaction. However, this model seems to hold for certain experimental conditions. Data presented here are supported by earlier findings obtained by means of fluorimetric titrations and protective effects of ligands on thermal inactivation of the enzyme. All results are consistent with the following mechanism for trimeric PNPs: (i) random binding of substrates, (ii) potent binding and slow release of some reaction products leading to the circumstances that the chemical step is not the slowest one and that rapid-equilibrium assumptions do not hold, (iii) a dual role of phosphate--a substrate and also a reaction modifier.

Kinetic properties of Cellulomonas sp. purine nucleoside phosphorylase with typical and non-typical substrates: implications for the reaction mechanism

Phosphorolysis catalyzed by Cellulomonas sp. PNP with typical nucleoside substrate, inosine (Ino), and non-typical 7-methylguanosine (m7Guo), with either nucleoside or phosphate (Pd) as the varied substrate, kinetics of the reverse synthetic reaction with guanine (Gua) and ribose-1-phosphate (R1P) as the varied substrates, and product inhibition patterns of synthetic and phosphorolytic reaction pathways were studied by steady-state kinetic methods. It is concluded that, like for mammalian trimeric PNP, complex kinetic characteristics observed for Cellulomonas enzyme results from simultaneous occurrence of three phenomena. These are sequential but random, not ordered binding of substrates, tight binding of one substrate purine bases, leading to the circumstances that for such substrates (products) rapid-equilibrium assumptions do not hold, and a dual role of Pi, a substrate, and also a reaction modifier that helps to release a tightly bound purine base.

Spectroscopic and kinetic studies of interactions of calf spleen purine nucleoside phosphorylase with 8-azaguanine, and its 9-(2-phosphonylmethoxyethyl) derivative

Spectroscopic and kinetic studies of interactions of calf spleen purine nucleoside phosphorylase with 8-azaguanine, an excellent fluorescent/fluorogenic substrate for the synthetic pathway of the reaction, and its 9-(2-phosphonylmethoxyethyl) derivative, a bisubstrate analogue inhibitor, were carried out. The goal was to clarify the catalytic mechanism of the enzymatic reaction by identification of ionic/tautomeric forms of these ligands in the complex with PNP.

Kinetics and crystal structure of human purine nucleoside phosphorylase in complex with 7-methyl-6-thio-guanosine

Purine nucleoside phosphorylase (PNP) catalyzes the reversible phosphorolysis of nucleosides and deoxynucleosides, generating ribose 1-phosphate and the purine base, which is an important step of purine catabolism pathway. The lack of such an activity in humans, owing to a genetic disorder, causes T-cell impairment, and drugs that inhibit this enzyme may have the potential of being utilized as modulators of the immunological system to treat leukemia, autoimmune diseases, and rejection in organ transplantation. Here, we describe kinetics and crystal structure of human PNP in complex with 7-methyl-6-thio-guanosine, a synthetic substrate, which is largely used in activity assays. Analysis of the structure identifies different protein conformational changes upon ligand binding, and comparison of kinetic and structural data permits an understanding of the effects of atomic substitution on key positions of the synthetic substrate and their consequences to enzyme binding and catalysis. Such knowledge may be helpful in designing new PNP inhibitors.

Structures for the potential drug target purine nucleoside phosphorylase from Schistosoma mansoni causal agent of schistosomiasis

Despite the availability of effective chemotherapy, schistosomiasis continues to be one of the major parasitic infections to affect the human population worldwide. Currently, little is known of the structural biology of the parasites that are responsible for the disease and few attempts have been made to develop second generation drugs, which may become essential if resistance to those currently available becomes an issue. Here, we describe three crystal structures for the enzyme purine nucleoside phosphorylase (PNP) from Schistosoma mansoni, a component of the purine salvage pathway. PNP is known to be essential for the recovery of purine bases and nucleosides in schistosomes, due to an absence of the enzymes for de novo synthesis, making it a sensitive point in the parasite's metabolism. In all three structures reported here, acetate occupies part of the base-binding site and is directly bound to the conserved glutamic acid at position 203. One of the structures presents the crystallization additive sulfobetaine 195 (NDSB195) occupying simultaneously the ribose and phosphate binding sites, whilst a second presents only phosphate in the latter. The observation of sulfobetaine specifically bound to the protein active site was unexpected and is unique to this structure as far as we are aware. Considerable flexibility is observed in the active site, principally due to variable structural disorder in the regions centered on residues 64 and 260. This conformational plasticity extends to the way in which both NDSB195 and phosphate bind to the individual monomers of the trimeric structure reported here. Differences between the parasite and human enzymes are limited principally to the base-binding site, where the substitution of V245 in the mammalian enzymes by S247 introduces additional hydrogen bonding potential to the site. This is satisfied in the structures described here by a water molecule whose presence is normally observed only in complexes with 6-oxopurines. Residue Y202, which replaces F200 in human PNP, is able to reach over the ribose-binding site to interact with H259 and is predicted to form an additional hydrogen bond with the 5' hydroxyl of nucleoside substrates.

A proof-of-principle pharmacokinetic, pharmacodynamic, and clinical study with purine nucleoside phosphorylase inhibitor immucillin-H (BCX-1777, forodesine)

The discovery of purine nucleoside phosphorylase (PNP) deficiency and T lymphocytopenia suggested that inhibition of this enzyme could serve as a therapeutic target. Inhibitors of PNP failed until structure-based synthesis of immucillin-H (BCX-1777, forodesine), a transition-state analog of PNP. The picomolar potency for PNP, T cell-selective cytotoxicity, and animal studies provided the rationale for use of forodesine in T-cell malignancies. Five patients were treated with an intravenous infusion of forodesine (40 mg/m2) on day 1; treatment continued on day 2; forodesine was administered every 12 hours for an additional 8 doses. Plasma and cellular pharmacokinetics and pharmaco-dynamics were investigated. Median peak level of forodesine (5.4 microM) was achieved at the end of infusion. This level was sufficient to increase plasma 2'-deoxyguanosine (dGuo) concentrations in all patients. Intracellular deoxyguanosine triphosphate (dGTP) increased by 2- to 40-fold in 4 of 5 patients (8 of 9 courses) and correlated with antileukemia activity in 4 patients. However, objective responses were not observed. This was the first clinical study in humans to demonstrate the plasma pharmacokinetics and the pharmacodynamic effectiveness of the PNP inhibitor, forodesine; however, regrowth of leukemia cells in the blood and marrow after course 1 suggested that a different therapeutic schedule should be considered for future studies.

Stopped-flow studies of guanine binding by calf spleen purine nucleoside phosphorylase

The binding of guanine to calf spleen purine nucleoside phosphorylase at 20 degrees C, in 20 mM Hepes-NaOH buffer, pH 7.0, at several ionic strength between 5 and 150 mM was investigated using a stopped-flow spectrofluorimeter. The kinetic transients registered after mixing a protein solution with ligand solutions of different concentrations were simultaneously fitted by several association reaction models using nonlinear least-squares procedure based on numerical integration of the chemical kinetic equations appropriate for given model. It is concluded that binding of a guanine molecule by each of the binding sites is a two-step process and that symmetrical trimeric calf spleen purine nucleoside phosphorylase represents a system of (identical) interacting binding sites. The interaction is visible through relations between the rate constants and non-additivity of changes in "molar" fluorescence for different forms of PNP-guanine complexes. It is also probable that electrostatic effects in guanine binding are weak, which indicates that it is the neutral form of the ligand which is bound and dissociated by PNP molecule.

Preliminary investigation of the three-dimensional structure of Salmonella typhimurium uridine phosphorylase in the crystalline state

Uridine phosphorylase (UPh) catalyzes the phosphorolytic cleavage of the C-N glycosidic bond of uridine to ribose 1-phosphate and uracil in the pyrimidine-salvage pathway. The crystal structure of the Salmonella typhimurium uridine phosphorylase (StUPh) has been determined at 2.5 A resolution and refined to an R factor of 22.1% and an Rfree of 27.9%. The hexameric StUPh displays 32 point-group symmetry and utilizes both twofold and threefold non-crystallographic axes. A phosphate is bound at the active site and forms hydrogen bonds to Arg91, Arg30, Thr94 and Gly26 of one monomer and Arg48 of an adjacent monomer. The hexameric StUPh model reveals a close structural relationship to Escherichia coli uridine phosphorylase (EcUPh).

New catalytic mechanism for human purine nucleoside phosphorylase

Human purine nucleoside phosphorylase has been submitted to intensive structure-based design of inhibitors, most of them using low-resolution structures of human PNP. Recently, several structures of human PNP have been reported, which allowed redefinition of the active site and understanding of the structural basis for inhibition of PNP by acyclovir and immucillin-H. Based on previously solved human PNP structures, we proposed here a new catalytic mechanism for human PNP, which is supported by crystallographic studies and explains previously determined kinetic data.

Crystal Structure of Calf Spleen Purine Nucleoside Phosphorylase with Two Full Trimers in the Asymmetric Unit: Important Implications for the Mechanism of Catalysis

The crystal structure of the binary complex of trimeric purine nucleoside phosphorylase (PNP) from calf spleen with the acyclic nucleoside phosphonate inhibitor 2,6-diamino-(S)-9-[2-(phosphonomethoxy)propyl]purine ((S)-PMPDAP) is determined at 2.3A resolution in space group P2(1)2(1)2(1). Crystallization in this space group, which is observed for the first time with a calf spleen PNP crystal structure, is obtained in the presence of calcium atoms. In contrast to the previously described cubic space group P2(1)3, two independent trimers are observed in the asymmetric unit, hence possible differences between monomers forming the biologically active trimer could be detected, if present. Such differences would be expected due to third-of-the-sites binding documented for transition-state events and inhibitors. However, no differences are noted, and binding stoichiometry of three inhibitor molecules per enzyme trimer is observed in the crystal structure, and in the parallel solution studies using isothermal titration calorimetry and spectrofluorimetric titrations. Presence of phosphate was shown to modify binding stoichiometry of hypoxanthine. Therefore, the enzyme was also crystallized in space group P2(1)2(1)2(1) in the presence of (S)-PMPDAP and phosphate, and the resulting structure of the binary PNP/(S)-PMPDAP complex was refined at 2.05A resolution. No qualitative differences between complexes obtained with and without the presence of phosphate were detected, except for the hydrogen bond contact of Arg84 and a phosphonate group, which is observed only in the former complex in three out of six independent monomers. Possible hydrogen bonds observed in the enzyme complexed with (S)-PMPDAP, in particular a putative hydrogen bonding contact N(1)-H cdots, three dots, centered Glu201, indicate that the inhibitor binds in a tautomeric or ionic form in which position N(1) acts as a hydrogen bond donor. This points to a crucial role of this hydrogen bond in defining specificity of trimeric PNPs and is in line with the proposed mechanism of catalysis in which this contact helps to stabilize the negative charge that accumulates on O(6) of the purine base in the transition state. In the present crystal structure the loop between Thr60 and Ala65 was found in a different conformation than that observed in crystal structures of trimeric PNPs up to now. Due to this change a new wide entrance is opened into the active site pocket, which is otherwise buried in the interior of the protein. Hence, our present crystal structure provides no obvious indication for obligatory binding of one of the substrates before binding of a second one; it is rather consistent with random binding of substrates. All these results provide new data for clarifying the mechanism of catalysis and give reasons for the non-Michaelis kinetics of trimeric PNPs.

Crystal Structure of Purine Nucleoside Phosphorylase from Thermus thermophilus

The purine nucleoside phosphorylase from Thermus thermophilus crystallized in space group P4(3)2(1)2 with the unit cell dimensions a = 131.9 A and c = 169.9 A and one biologically active hexamer in the asymmetric unit. The structure was solved by the molecular replacement method and refined at a 1.9A resolution to an r(free) value of 20.8%. The crystals of the binary complex with sulfate ion and ternary complexes with sulfate and adenosine or guanosine were also prepared and their crystal structures were refined at 2.1A, 2.4A and 2.4A, respectively. The overall structure of the T.thermophilus enzyme is similar to the structures of hexameric enzymes from Escherichia coli and Sulfolobus solfataricus, but significant differences are observed in the purine base recognition site. A base recognizing aspartic acid, which is conserved among the hexameric purine nucleoside phosphorylases, is Asn204 in the T.thermophilus enzyme, which is reminiscent of the base recognizing asparagine in trimeric purine nucleoside phosphorylases. Isothermal titration calorimetry measurements indicate that both adenosine and guanosine bind the enzyme with nearly similar affinity. However, the functional assays show that as in trimeric PNPs, only the guanosine is a true substrate of the T.thermophilus enzyme. In the case of adenosine recognition, the Asn204 forms hydrogen bonds with N6 and N7 of the base. While in the case of guanosine recognition, the Asn204 is slightly shifted together with the beta(9)alpha(7) loop and predisposed to hydrogen bond formation with O6 of the base in the transition state. The obtained experimental data suggest that the catalytic properties of the T.thermophilus enzyme are reminiscent of the trimeric rather than hexameric purine nucleoside phosphorylases.

Crystal Structures of Escherichia coli Uridine Phosphorylase in Two Native and Three Complexed Forms Reveal Basis of Substrate Specificity, Induced Conformational Changes and Influence of Potassium

Uridine phosphorylase (UP) is a key enzyme in the pyrimidine salvage pathway that catalyses the reversible phosphorolysis of uridine to uracil and ribose 1-phosphate. Inhibiting liver UP in humans raises blood uridine levels and produces a protective effect ("uridine rescue") against the toxicity of the chemotherapeutic agent 5-fluorouracil without reducing its antitumour activity. We have investigated UP-substrate interactions by determining the crystal structures of native Escherichia coli UP (two forms), and complexes with 5-fluorouracil/ribose 1-phosphate, 2-deoxyuridine/phosphate and thymidine/phosphate. These hexameric structures confirm the overall structural similarity of UP to E.coli purine nucleoside phosphorylase (PNP) whereby, in the presence of substrate, each displays a closed conformation resulting from a concerted movement that closes the active site cleft. However, in contrast to PNP where helix segmentation is the major conformational change between the open and closed forms, in UP more extensive changes are observed. In particular a swinging movement of a flap region consisting of residues 224-234 seals the active site. This overall change in conformation results in compression of the active site cleft. Gln166 and Arg168, part of an inserted segment not seen in PNP, are key residues in the uracil binding pocket and together with a tightly bound water molecule are seen to be involved in the substrate specificity of UP. Enzyme activity shows a twofold dependence on potassium ion concentration. The presence of a potassium ion at the monomer/monomer interface induces some local rearrangement, which results in dimer stabilisation. The conservation of key residues and interactions with substrate in the phosphate and ribose binding pockets suggest that ribooxocarbenium ion formation during catalysis of UP may be similar to that proposed for E.coli PNP.

Crystal structure of calf spleen purine nucleoside phosphorylase in a complex with multisubstrate analogue inhibitor with 2,6-diaminopurine aglycone

The crystal structure at 2.05 A resolution of calf spleen PNP complexed with stoichiometric concentration of acyclic nucleoside phosphonate inhibitor, 2,6-diamino-(S)-9-[2-(phosphonomethoxy)propyl]purine, in a new space group P2(1)2(1)2(1) which contains two full trimers in the asymmetric crystal unit is described.

Structural basis for substrate specificity of Escherichia coli purine nucleoside phosphorylase

Purine nucleoside phosphorylase catalyzes reversible phosphorolysis of purine nucleosides and 2'-deoxypurine nucleosides to the free base and ribose (or 2'-deoxyribose) 1-phosphate. Whereas the human enzyme is specific for 6-oxopurine ribonucleosides, the Escherichia coli enzyme accepts additional substrates including 6-oxopurine ribonucleosides, 6-aminopurine ribonucleosides, and to a lesser extent purine arabinosides. These differences have been exploited in a potential suicide gene therapy treatment for solid tumors. In an effort to optimize this suicide gene therapy approach, we have determined the three-dimensional structure of the E. coli enzyme in complex with 10 nucleoside analogs and correlated the structures with kinetic measurements and computer modeling. These studies explain the preference of the enzyme for ribose sugars, show increased flexibility for active site residues Asp204 and Arg24, and suggest that interactions involving the 1- and 6-positions of the purine and the 4'- and 5'-positions of the ribose provide the best opportunities to increase prodrug specificity and enzyme efficiency.

The purine nucleoside phosphorylase from Trichomonas vaginalis is a homologue of the bacterial enzyme

Trichomonas vaginalis is a parasitic protozoan and the causative agent of trichomoniasis. Its primary purine salvage system, consisting of a purine nucleoside phosphorylase (PNP) and a purine nucleoside kinase, presents potential targets for designing selective inhibitors as antitrichomonial drugs because of lack of de novo synthesis of purine nucleotides in this organism. cDNA encoding T. vaginalis PNP was isolated by complementation of an Escherichia coli strain deficient in PNP and expressed, and the recombinant enzyme was purified to apparent homogeneity. It bears only 28% sequence identity with that of human PNP but 57% identity with the E. coli enzyme. Gel filtration showed the enzyme in a hexameric form, similar to the bacterial PNPs. Steady-state kinetic analysis of T. vaginalis PNP-catalyzed reactions gave K(m)'s of 31.5, 59.7, and 6.1 microM for inosine, guanosine, and adenosine in the nucleosidase reaction and 45.6, 35.9, and 12.3 microM for hypoxanthine, guanine, and adenine in the direction of nucleoside synthesis. This substrate specificity appears to be similar to that of bacterial PNPs. The catalytic efficiency of this enzyme with adenine as substrate is 58-fold higher than that with either hypoxanthine or guanine, representing a distinct disparity with the mammalian PNPs, which have negligible activity with either adenine or adenosine. The kinetic mechanism of T. vaginalis PNP-catalyzed reactions, determined by product inhibition and equilibrium isotope exchange, was by random binding of substrates (purine base and ribose 1-phosphate) with ordered release of the purine nucleoside first, followed by inorganic phosphate. Formycin A, an analogue of adenosine known as an inhibitor of E. coli PNP without any effect on mammalian PNPs, was shown to inhibit T. vaginalis PNP with a K(is) of 2.3 microM by competing with adenosine. T. vaginalis PNP thus belongs to the family of bacterial PNPs and constitutes a target for antitrichomonial chemotherapy.

Xanthosine and xanthine. Substrate properties with purine nucleoside phosphorylases, and relevance to other enzyme systems

Substrate properties of xanthine (Xan) and xanthosine (Xao) for purine nucleoside phosphorylases (PNP) of mammalian origin have been reported previously, but only at a single arbitrarily selected pH and with no kinetic constants. Additionally, studies have not taken into account the fact that, at physiological pH, Xao (pKa = 5.7) is a monoanion, while Xan (pKa = 7.7) is an equilibrium mixture of the neutral and monoanionic forms. Furthermore the monoanionic forms, unlike those of guanosine (Guo) and inosine (Ino), and guanine (Gua) and hypoxanthine (Hx), are still 6-oxopurines. The optimum pH for PNP from human erythrocytes and calf spleen with both Xao and Xan is in the range 5-6, whereas those with Guo and Gua, and Ino and Hx, are in the range 7-8. The pH-dependence of substrate properties of Xao and Xan points to both neutral and anionic forms as substrates, with a marked preference for the neutral species. Both neutral and anionic forms of 6-thioxanthine (pKa = 6.5 +/- 0.1), but not of 2-thioxanthine (pKa = 5.9 +/- 0.1), are weaker substrates. Phosphorolysis of Xao to Xan by calf spleen PNP at pH 5.7 levels off at 83% conversion, due to equilibrium with the reverse synthetic pathway (equilibrium constant 0.05), and not by product inhibition. Replacement of Pi by arsenate led to complete arsenolysis of Xao. Kinetic parameters are reported for the phosphorolytic and reverse synthetic pathways at several selected pH values. Phosphorolysis of 200 micro m Xao by the human enzyme at pH 5.7 is inhibited by Guo (IC50 = 10 +/- 2 micro m), Hx (IC50 = 7 +/- 1 micro m) and Gua (IC50 = 4.0 +/- 0.2 micro m). With Gua, inhibition was shown to be competitive, with Ki = 2.0 +/- 0.3 micro m. By contrast, Xao and its products of phosphorolysis (Xan and R1P), were poor inhibitors of phosphorolysis of Guo, and Xan did not inhibit the reverse reaction with Gua. Possible modes of binding of the neutral and anionic forms of Xan and Xao by mammalian PNPs are proposed. Attention is directed to the fact that the structural properties of the neutral and ionic forms of XMP, Xao and Xan are also of key importance in many other enzyme systems, such as IMP dehydrogenase, some nucleic acid polymerases, biosynthesis of caffeine and phosphoribosyltransferases.

Purine nucleoside phosphorylase from Cellulomonas sp.: physicochemical properties and binding of substrates determined by ligand-dependent enhancement of enzyme intrinsic fluorescence, and by protective effects of ligands on thermal inactivation of the enzyme

Purine nucleoside phosphorylase (PNP) from Cellulomonas sp., homotrimeric in the crystalline state, is also a trimer in solution. Other features of the enzyme are typical for "low molecular mass" PNPs, except for its unusual stability at pH 11. Purine bases, alpha-D-ribose-1-phosphate (R1P) and phosphate enhance the intrinsic fluorescence of Cellulomonas PNP, and hence form binary complexes and induce conformational changes of the protein that alter the microenvironment of tryptophan residue(s). The effect due to guanine (Gua) binding is much higher than those caused by other ligands, suggesting that the enzyme preferentially binds a fluorescent, most probably rare tautomeric anionic form of Gua, further shown by comparison of emission properties of the PNP/Gua complex with that of Gua anion and its N-methyl derivatives. Guanosine (Guo) and inosine (Ino) at 100 microM concentration show little and no effect, respectively, on enzyme intrinsic fluorescence, but their protective effect against thermal inactivation of the enzyme points to their forming weak binary complexes with PNP. Binding of Gua, hypoxanthine (Hx) and R1P to the trimeric enzyme is described by one dissociation constant, K(d)=0.46 microM for Gua, 3.0 microM for Hx, and 60 microM for R1P. The binding stoichiometry for Gua (and probably Hx) is three ligand molecules per enzyme trimer. Effects of phosphate on the enzyme intrinsic fluorescence are due not only to binding, but also to an increase in ionic strength, as shown by titration with KCl. When corrected for effects of ionic strength, titration data with phosphate are most consistent with one dissociation constant, K(d)=270 microM, but existence of a very weak binding site with K(d)>50 mM could not be unequivocally ruled out. Binding of Gua to the PNP/phosphate binary complex is weaker (K(d)=1.7 microM) than to the free enzyme (K(d)=0.46 microM), suggesting that phosphate helps release the purine base in the catalytic process of phosphorolysis. The results indicate that nonlinear kinetic plots of initial velocity, typical for PNPs, including Cellulomonas PNP, are not, as generally assumed, due to cooperative interaction between monomers forming the trimer, but to a more complex kinetic mechanism than hitherto considered.

Calf spleen purine nucleoside phosphorylase: complex kinetic mechanism, hydrolysis of 7-methylguanosine, and oligomeric state in solution

The active enzyme form was found to be a homotrimer, no active monomers were observed. Only in the presence of an extremely high orthophosphate concentration (0.5 M) or at a low enzyme concentration (0.2 microg/ml) with no ligands present a small fraction of the enzyme is probably in a dissociated and/or non-active form. The specific activity is invariant over a broad enzyme concentration range (0.017 microg/ml-0.29 mg/ml). At concentrations below 0.9 microg/ml and in the absence of ligands the enzyme tends to loose its catalytic activity, while in the presence of any substrate or at higher concentrations it was found to be active as a trimer. In the absence of phosphate the enzyme catalyses the hydrolysis of 7-methylguanosine (m7Guo) with a catalytic rate constant 1.3x10(-3) x s(-1) as compared with the rate of 38 s(-1) for the phosphorolysis of this nucleoside. The initial pre-steady-state phase of the phosphorolysis of m7Guo, 70 s(-1), is almost twice faster than the steady-state rate and indicates that the rate-limiting step is subsequent to the glycosidic bond cleavage. Complex kinetic behaviour with substrates of phosphorolytic direction (various nucleosides and orthophosphate) was observed; data for phosphate as the variable substrate with inosine and guanosine, but not with their 7-methyl counterparts, might be interpreted as two binding sites with different affinities, or as a negative cooperativity. However, the titration of the enzyme intrinsic fluorescence with 0.2 microM-30 mM phosphate is consistent with only one dissociation constant for phosphate, K(d)=220+/-120 microM. Protective effects of ligands on the thermal inactivation of the enzyme indicate that all substrates of the phosphorolytic and the synthetic reactions are able to form binary complexes with the calf spleen purine nucleoside phosphorylase. The purine bases, guanine and hypoxanthine, bind strongly with dissociation constants of about 0.1 microM, while all other ligands studied, including 7-methylguanine and 7-methylhypoxanthine, bind at least 3 orders of magnitude less potently. Binding of guanine and hypoxanthine is about 10-fold weakened by the presence of phosphate. These observations are best interpretable by the complex kinetic mechanism of the phosphorolytic reaction involving (i) random substrate binding, (ii) unusually slow, hence strongly rate-limiting, dissociation of the products guanine and hypoxanthine, but not 7-methylguanine and 7-methylhypoxanthine, and (iii) dual function of the phosphate binding site with phosphate acting as a substrate and as a modifier helping in the release of a purine base after glycosidic bond cleavage.

Transition state analogue inhibitors of purine nucleoside phosphorylase from Plasmodium falciparum

Immucillins are logically designed transition-state analogue inhibitors of mammalian purine nucleoside phosphorylase (PNP) that induce purine-less death of Plasmodium falciparum in cultured erythrocytes (Kicska, G. A., Tyler, P. C., Evans, G. B., Furneaux, R. H., Schramm, V. L., and Kim, K. (2002) J. Biol. Chem. 277, 3226-3231). PNP is present at high levels in human erythrocytes and in P. falciparum, but the Plasmodium enzyme has not been characterized. A search of the P. falciparum genome data base yielded an open reading frame similar to the PNP from Escherichia coli. PNP from P. falciparum (P. falciparum PNP) was cloned, overexpressed in E. coli, purified, and characterized. The primary amino acid sequence has 26% identity with E. coli PNP, has 20% identity with human PNP, and is phylogenetically unique among known PNPs with equal genetic distance between PNPs and uridine phosphorylases. Recombinant P. falciparum PNP is catalytically active for inosine and guanosine but is less active for uridine. The immucillins are powerful inhibitors of P. falciparum PNP. Immucillin-H is a slow onset tight binding inhibitor with a K(i)* value of 0.6 nm. Eight related immucillins are also powerful inhibitors with dissociation constants from 0.9 to 20 nm. The K(m)/K(i)* value for immucillin-H is 9000, making this inhibitor the most powerful yet reported for P. falciparum PNP. The PNP from P. falciparum differs from the human enzyme by a lower K(m) for inosine, decreased preference for deoxyguanosine, and reduced affinity for the immucillins, with the exception of 5'-deoxy-immucillin-H. These properties of P. falciparum PNP are consistent with a metabolic role in purine salvage and provide an explanation for the antibiotic effect of the immucillins on P. falciparum cultured in human erythrocytes.

Open and closed conformation of the E. coli purine nucleoside phosphorylase active center and implications for the catalytic mechanism

The crystal structure of the ternary complex of hexameric purine nucleoside phosphorylase (PNP) from Escherichia coli with formycin A derivatives and phosphate or sulphate ions is determined at 2.0 A resolution. The hexamer is found as a trimer of unsymmetric dimers, which are formed by pairs of monomers with active sites in different conformations. The conformational difference stems from a flexible helix (H8: 214-236), which is continuous in one conformer, and segmented in the other. With the continuous helix, the entry into the active site pocket is wide open, and the ligands are bound only loosely ("open" or "loose binding" conformation). By segmentation of the helix (H8: 214-219 and H8': 223-236, separated by a gamma-turn), the entry into the active site is partially closed, the pocket is narrowed and the ligands are bound much more tightly ("closed" or "tight binding" conformation). Furthermore, the side-chain of Arg217 is carried by the moving helix into the active site. This residue, conserved in all homologous PNPs, plays an important role in the proposed catalytic mechanism. In this mechanism, substrate binding takes place in the open, and and the catalytic action occurs in the closed conformation. Catalytic action involves protonation of the purine base at position N7 by the side-chain of Asp204, which is initially in the acid form. The proton transfer is triggered by the Arg217 side-chain which is moved by the conformation change into hydrogen bond distance to Asp204. The mechanism explains the broad specificity of E. coli PNP, which allows 6-amino as well as 6-oxo-nucleosides as substrates. The observation of two kinds of binding sites is fully in line with solution experiments which independently observe strong and weak binding sites for phosphate as well as for the nucleoside inhibitor.

Structural analyses reveal two distinct families of nucleoside phosphorylases

The reversible phosphorolysis of purine and pyrimidine nucleosides is an important biochemical reaction in the salvage pathway, which provides an alternative to the de novo purine and pyrimidine biosynthetic pathways. Structural studies in our laboratory and by others have revealed that only two folds exist that catalyse the phosphorolysis of all nucleosides, and provide the basis for defining two families of nucleoside phosphorylases. The first family (nucleoside phosphorylase-I) includes enzymes that share a common single-domain subunit, with either a trimeric or a hexameric quaternary structure, and accept a range of both purine and pyrimidine nucleoside substrates. Despite differences in substrate specificity, amino acid sequence and quaternary structure, all members of this family share a characteristic subunit topology. We have also carried out a sequence motif study that identified regions of the common subunit fold that are functionally significant in differentiating the various members of the nucleoside phosphorylase-I family. Although the substrate-binding sites are arranged similarly for all members of the nucleoside phosphorylase-I family, a comparison of the active sites from the known structures of this family indicates significant differences between the trimeric and hexameric family members. Sequence comparisons also suggest structural identity between the nucleoside phosphorylase-I family and both 5'-methylthioadenosine/S-adenosylhomocysteine nucleosidase and AMP nucleosidase. Members of the second family of nucleoside phosphorylases (nucleoside phosphorylase-II) share a common two-domain subunit fold and a dimeric quaternary structure, share a significant level of sequence identity (>30%) and are specific for pyrimidine nucleosides. Members of this second family accept both thymidine and uridine substrates in lower organisms, but are specific for thymidine in mammals and other higher organisms. A possible relationship between nucleoside phosphorylase-II and anthranilate phosphoribosyltransferase has been identified through sequence comparisons. Initial studies in our laboratory suggested that members of the nucleoside phosphorylase-II family require significant domain movements in order for catalysis to proceed. A series of recent structures has confirmed our hypothesis and provided details of these conformational changes. Structural studies of the nucleoside phosphorylases have resulted in a wealth of information that begins to address fundamental biological questions, such as how Nature makes use of the intricate relationships between structure and function, and how biological processes have evolved over time. In addition, the therapeutic potential of suppressing the nucleoside phosphorylase activity in either family of enzymes has motivated efforts to design potent inhibitors. Several research groups have synthesized a variety of nucleoside phosphorylase inhibitors that are at various stages of preclinical and clinical evaluation.

Three-dimensional structure of a hyperthermophilic 5'-deoxy-5'-methylthioadenosine phosphorylase from Sulfolobus solfataricus

The structure of 5'-deoxy-5'-methylthioadenosine phosphorylase from Sulfolobus solfataricus (SsMTAP) has been determined alone, as ternary complexes with sulfate plus substrates 5'-deoxy-5'-methylthioadenosine, adenosine, or guanosine, or with the noncleavable substrate analog Formycin B and as binary complexes with phosphate or sulfate alone. The structure of unliganded SsMTAP was refined at 2.5-A resolution and the structures of the complexes were refined at resolutions ranging from 1.6 to 2.0 A. SsMTAP is unusual both for its broad substrate specificity and for its extreme thermal stability. The hexameric structure of SsMTAP is similar to that of purine-nucleoside phosphorylase (PNP) from Escherichia coli, however, only SsMTAP accepts 5'-deoxy-5'-methylthioadenosine as a substrate. The active site of SsMTAP is similar to that of E. coli PNP with 13 of 18 nearest residues being identical. The main differences are at Thr(89), which corresponds to serine in E. coli PNP, and Glu(163), which corresponds to proline in E. coli PNP. In addition, a water molecule is found near the purine N-7 position in the guanosine complex of SsMTAP. Thr(89) is near the 5'-position of the nucleoside and may account for the ability of SsMTAP to accept either hydrophobic or hydrophilic substituents in that position. Unlike E. coli PNP, the structures of SsMTAP reveal a substrate-induced conformational change involving Glu(163). This residue is located at the interface between subunits and swings in toward the active site upon nucleoside binding. The high-resolution structures of SsMTAP suggest that the transition state is stabilized in different ways for 6-amino versus 6-oxo substrates. SsMTAP has optimal activity at 120 degrees C and retains full activity after 2 h at 100 degrees C. Examination of the three-dimensional structure of SsMTAP suggests that unlike most thermophilic enzymes, disulfide linkages play a key in role in its thermal stability.

Structure of E. coli 5'-methylthioadenosine/S-adenosylhomocysteine nucleosidase reveals similarity to the purine nucleoside phosphorylases

BACKGROUND

5'-methylthioadenosine/S-adenosyl-homocysteine (MTA/AdoHcy) nucleosidase catalyzes the irreversible cleavage of 5'-methylthioadenosine and S-adenosylhomocysteine to adenine and the corresponding thioribose, 5'-methylthioribose and S-ribosylhomocysteine, respectively. While this enzyme is crucial for the metabolism of AdoHcy and MTA nucleosides in many prokaryotic and lower eukaryotic organisms, it is absent in mammalian cells. This metabolic difference represents an exploitable target for rational drug design.

RESULTS

The crystal structure of E. coli MTA/AdoHcy nucleosidase was determined at 1.90 A resolution with the multiwavelength anomalous diffraction (MAD) technique. Each monomer of the MTA/AdoHcy nucleosidase dimer consists of a mixed alpha/beta domain with a nine-stranded mixed beta sheet, flanked by six alpha helices and a small 3(10) helix. Intersubunit contacts between the two monomers present in the asymmetric unit are mediated primarily by helix-helix and helix-loop hydrophobic interactions. The unexpected presence of an adenine molecule in the active site of the enzyme has allowed the identification of both substrate binding and potential catalytic amino acid residues.

CONCLUSIONS

Although the sequence of E. coli MTA/AdoHcy nucleosidase has almost no identity with any known enzyme, its tertiary structure is similar to both the mammalian (trimeric) and prokaryotic (hexameric) purine nucleoside phosphorylases. The structure provides evidence that this protein is functional as a dimer and that the dual specificity for MTA and AdoHcy results from the truncation of a helix. The structure of MTA/AdoHcy nucleosidase is the first structure of a prokaryotic nucleoside N-ribohydrolase specific for 6-aminopurines.

Structures of purine nucleoside phosphorylase from Mycobacterium tuberculosis in complexes with immucillin-H and its pieces

A structural genomics comparison of purine nucleoside phosphorylases (PNPs) indicated that the enzyme encoded by Mycobacterium tuberculosis (TB-PNP) resembles the mammalian trimeric structure rather than the bacterial hexameric PNPs. The crystal structure of M. tuberculosis PNP in complex with the transition-state analogue immucillin-H (ImmH) and inorganic phosphate was solved at 1.75 A resolution and confirms the trimeric structure. Binding of the inhibitor occurs independently at the three catalytic sites, unlike mammalian PNPs which demonstrate negative cooperativity in ImmH binding. Reduced subunit interface contacts for TB-PNP, compared to the mammalian enzymes, correlate with the loss of the cooperative inhibitor binding. Mammalian and TB-PNPs both exhibit slow-onset inhibition and picomolar dissociation constants for ImmH. The structure supports a catalytic mechanism of reactant destabilization by neighboring group electrostatic interactions, transition-state stabilization, and leaving group activation. Despite an overall amino acid sequence identity of 33% between bovine and TB-PNPs and almost complete conservation in active site residues, one catalytic site difference suggests a strategy for the design of transition-state analogues with specificity for TB-PNP. The structure of TB-PNP was also solved to 2.0 A with 9-deazahypoxanthine (9dHX), iminoribitol (IR), and PO(4) to reconstruct the ImmH complex with its separate components. One subunit of the trimer has 9dHX, IR, and PO(4) bound, while the remaining two subunits contain only 9dHX. In the filled subunit, 9dHX retains the contacts found in the ImmH complex. However, the region of IR that corresponds to the oxocarbenium ion is translocated in the direction of the reaction coordinate, and the nucleophilic phosphate rotates away from the IR group. Loose packing of the pieces of ImmH in the catalytic site establishes that covalent connectivity in ImmH is required to achieve the tightly bound complex.