The binding of 3′- N-piperidine-4-carboxyl-3′-deoxy- ara-uridine to ribonuclease A in the crystal

The binding of 3′- N-piperidine-4-carboxyl-3′-deoxy- ara-uridine to ribonuclease A in the crystal

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 2 | Pages: 10

Extension: PDF | Download: 0

  The binding of 3 0 - N  -piperidine-4-carboxyl-3 0 -deoxy- ara -uridineto ribonuclease A in the crystal Demetres D. Leonidas, a,* Tushar Kanti Maiti, c Anirban Samanta, c Swagata Dasgupta, c Tanmaya Pathak, c Spyros E. Zographos a andNikos G. Oikonomakos a,b a Institute of Organic and Pharmaceutical Chemistry, The National Hellenic Research Foundation,48 Vas. Constantinou Avenue, 11635 Athens, Greece b Institute of Biological Research and Biotechnology, The National Hellenic Research Foundation,48 Vas. Constantinou Avenue, 11635 Athens, Greece c Department of Chemistry, Indian Institute of Technology, Kharagpur 721302, India Received 15 March 2006; revised 28 April 2006; accepted 3 May 2006Available online 30 May 2006 Abstract—  The binding of a moderate inhibitor, 3 0 - N  -piperidine-4-carboxyl-3 0 -deoxy- ara -uridine, to ribonuclease A has been studiedby X-ray crystallography at 1.7 A˚resolution. Two inhibitor molecules are bound in the central RNA binding cavity of RNase Aexploiting interactions with residues from peripheral binding sites rather than from the active site of the enzyme. The uracyl moietyof the first inhibitor molecule occupies the purine-preferring site of RNase A, while the rest of the molecule projects to the solvent.The second inhibitor molecule binds with the carboxyl group at the pyrimidine recognition site and the uridine moiety exploitsinteractions with RNase A residues Lys66, His119 and Asp121. Comparative structural analysis of the 3 0 - N  -piperidine-4-carbox-yl-3 0 -deoxy- ara -uridine complex with other RNase A–ligand complexes provides a structural explanation of its potency. The crystalstructure of the RNase A–3 0 - N  -piperidine-4-carboxyl-3 0 -deoxy- ara -uridine complex provides evidence of a novel ligand-bindingpattern in RNase A for 3 0 - N  -aminonucleosides that was not anticipated by modelling studies, while it also suggests ways to improvetheefficiencyandselectivityofsuchcompoundstodeveloppharmaceuticalsagainstpathologiesassociatedwithRNaseAhomologues.   2006 Elsevier Ltd. All rights reserved. 1. Introduction In the human genome 13 distinct vertebrate-specificribonuclease (RNase) genes have been identified, alllocalized in chromosome 14. 1 Ribonucleases (RNases)are enzymes that control, post-transcriptionally, theRNA population in cells. The RNase A superf amily isthe only enzyme family restricted to vertebrates. 2 Recentstudies indicate that there has been a rapid divergenceunder an unusual evolutionary pressure and suggest thatthe lineage might have started as host defence proteins. 1 Several homologues of the mammalian pancreaticRNase A (EC superfamily are also involvedin various human pathologies and RNase activity inserum and cell extracts is elevated in a variety of cancersand infectious diseases. 3 Angiogenin, a potent inducer of neovascularization, displays pathological side effectsduring cancer. 4 It has been proposed that the RNaseA superfamily started off from a progenitor with struc-tural similarities to Angiogenin. 2 Two other homo-logues, eosinophil cationic protein (ECP) andeosinophil derived neurotoxin (EDN), are both involvedin the immune response system and inflammatory disor-ders. 5–7 The pathological functions of these RNase Ahomologues are linked to their enzymatic activity, a factthat renders them as attractive targets for rational liganddesign of potent and selective inhibitors, that could beuseful as potential pharmaceutics to combat cancerand inflammatory disorders. Several rational inhibitordesign efforts 8–10 that target these enzymes are currentlyin progress.The RNase A superfamily comprises pyrimidine-specificsecreted endonucleases that degrade RNA through a 0968-0896/$ - see front matter    2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.bmc.2006.05.011 Keywords : Ribonuclease A; 3 0 - N  -Piperidine-4-carboxyl-3 0 -deoxy- ara -uridine; inhibition; X-ray crystallography; Structure-assistedinhibitor design.*Corresponding author. Tel.: +30 210 7273841; fax : +30 2107273831; e-mail: ddl@eie.grBioorganic & Medicinal Chemistry 14 (2006) 6055–6064  two-step transphosphorolytic–hydrolytic reaction. 11 Several subsites exist within the central catalytic grooveof RNase A, where substrate RNA binds, that aredefined as P o  . . .  P n  R o  . . .  R n , and B o  . . .  B n  accordingto the phosphate, ribose and base of RNA that bind,respectively, ( n  indicates the position of the group withrespect to the cleaved phosphate phosphodiester bondwhere  n  = 1). 12 The central region of the active site(B 1 R 1 P 1 R 2 B 2 ) is conserved in all RNases and therefore,structure-assisted inhibitor design studies have focusedmainly on the parental protein, RNase A, since inhibi-tors developed against this enzyme could also inhibitother members of the superfamily. 8 Today several inhib-itors, mainly substrate analogues, mono and diphos-phate (di)nucleotides with adenine at the 5 0 position,and cytosine or uridine at the 3 0 position of the scissilebond, have been studied. 8,13,14 All these compoundsare rather marginal inhibitors with dissociation con-stants in the mid-to-upper micromolar range. The bestinhibitor so far is pdUppA-3 0 p with  K  i  values of 27 nM, 180 nM and 360  l M for RNase A, EDN andAngiogenin, respectively, 8,15,16 whereas transition statetheory predicts picomolar values for genuine transitionstates. 3 The majority of small molecule RNase A inhibitorsstudied thus far have acidic groups such as phosphate,carboxylate, or sulfate. 10,14 Aminonucleosides like3 0 - N  -piperidine-4-carboxyl-3 0 -deoxy- ara -uridine ( 3e )have been selected for inhibition studies with the viewthat uridine derivatives with amino groups that havebasicities comparable to those of the imidazole of His12 and His119 might be able to perturb the proton-ating/deprotonating environment of the P 1  subsite andhence inhibit the enzymatic activity of RNase A andAngiogenin. 17 Thus, biochemical and biological studieshave led to the identification of 3 0 - N  -alkylamino-3 0 -de-oxy- ara -uridines as a new class of inhibitors of the enzy-matic activity of RNase A and the Angiogenin inducedangiogenesis. 17 These compounds are the first that donot have a phosphate or a sulfate group which arereported to inhibit the enzymatic activity of RNase Aand Angiogenin induced angiogenesis. Here we presentthe high resolution (1.7 A˚) crystal structure of theRNase A-complexed with  3e , the most potent RNaseA inhibitor from the series of the 3 0 - N  -alkylamino-3 0 -de-oxy- ara -uridines that were tested, 17 which reveals thestructural basis for its potency and indicates ways forimproving its selectivity and efficiency. In the crystalstructure two molecules of compound  3e  bind at theperipheral binding sites of RNase A rather than the cen-tral active site in a mode that differs strikingly from theone anticipated by modeling studies based on previousRNase A complex structures. 17 2. Results and discussion2.1. Overall structures The crystallographic asymmetric unit of the monoclinicRNase A crystals used in this study contains two pro-tein molecules (A and B). 18 For comparison reasonsand due to the lack of a crystal structure of free RNaseA from monoclinic crystals at cryogenic conditions(100 K) [the previous crystal structure of free RNaseA from monoclinic crystals (pdb entry 1AFU) wasdetermined at 2.0 A˚, at room temperature 18 ], we havedetermined the structure of free RNase A at 1.5 A˚res-olution, at 100 K. The 100 K free RNase A structure issimilar to that reported previously at 2.0 A˚at roomtemperature. 18 The rms distances between the earlierand the present free RNase A structures are 0.60/0.27, 0.55/0.28 and 0.83/0.70 A˚(molecule A/moleculeB of the RNase A non-crystallographic dimer) forC a , main chain atoms and all atoms of 124 equivalentresidues, respectively. Most of the amino acid residuesin the new structure are very well defined in the elec-tron density map with the exception of the loop regionbetween residues 16 and 24, which has higher temper-ature factors (average  B   factor = 23.0 A˚ 2 ) comparedto the rest of the protein residues (average B factorexcluding 16–24 loop = 12.1 A˚ 2 ). This loop appears tobe flexible, as in the room temperature structure, 18 evenat 100 K. The side chains of Ser59, Gln69 and Asp83are found in two alternative conformations. The cata-lytic site in the free RNase structure is occupied by18 water molecules which participate in a network of hydrogen bonds involving residues Gln11, His12,Lys41, Asn44, Gln69, Asn71, Val118, His119 andPhe120.Two inhibitor molecules (I and II) were bound at theperipheral binding sites of molecule A (Fig. 1) of thenon-crystallographic RNase A dimer but none in mole-cule B. This partial binding has also been observed inprevious binding studies with monoclinic crystals of RNase A 13,18,19 and might be attributed to the latticecontacts that limit access to the RNA binding sites of  Figure 1.  A schematic diagram of the RNase A molecule with the twoinhibitor molecules bound.6056  D. D. Leonidas et al. / Bioorg. Med. Chem. 14 (2006) 6055–6064  molecule B in the asymmetric unit. In all free RNase Astructures reported so far the side chain of the catalyticresidue His119 adopts two conformations denoted asproductive ( v 1 =   160  ) and non-productive( v 1 =   80  ), which are related by a 100   rotationabout the C a  –C b  bond and a 180   rotation about theC b  –C c  bond. 20–23 These conformations are dependenton the pH 24 and the ionic strength of the crystallizationsolution. 25 In the cryogenic free RNase A structurereported here His119 adopts the non-productive confor-mation in protein molecule A ( v 1 =   51  ) but the pro-ductive conformation ( v 1 = 173  ) in protein moleculeB of the non-crystallographic dimer indicating the diver-sity of this side chain in the unliganded enzyme struc-ture. Previous studies 26 have shown that binding of ligand groups in P 1  induces the productive conformationof the side chain of His119 and even though in  3e  liganddoes not bind at P 1  the side chain of His119 adopts thisconformation ( v 1 = 159  ).Upon binding to RNase A, each of the inhibitor mole-cules I and II displaces six water molecules from theperipheral binding sites of the free enzyme. With theexception of the side chain of His119 that was discussedabove, there are no other significant conformationalchanges in the catalytic site of RNase A upon ligandbinding. The rms distances between the structures of freeRNase A and the RNase A–  3e  complex are 0.29 A˚,0.31 A˚and 0.77 A˚for C a , main chain and all atoms of 124 equivalent residues in RNase A molecule A,respectively. 2.2. The binding of 3e to RNase A All atoms of the two  3e  molecules (I and II) are welldefined within the sigmaA weighted 2 F  o    F  c  electrondensity map of the RNase A–  3e  complex (Fig. 2).Upon binding to RNase A, each inhibitor moleculeadopts a different conformation. The glycosyl torsionangle  v  in both molecules I and II adopts the frequentlyobserved 27 anti   conformation (Table 1). The riboseadopts the quite rare  O 4- endo  puckering in  3e  moleculeI, while in molecule II it is found in the preferred, forfree and protein bound nucleotides, 27 C2 - endo  confor-mation. The rest of the backbone torsion angles are inthe common range for protein bound pyrimidines 27 withthe exception of the torsion angle  c  in ligand molecule IIwhich is at the unusual + ac  range (Table 1). The 4-car-boxypiperidine moiety in molecule I adopts the chairconformation, whereas in molecule II it is found in thehalf-chair conformation (Fig. 2). The numbering schemeused for compound  3e  is shown in Scheme 1.Inhibitor molecule I binds to RNase A by anchoring itsuracyl group to subsite B 2  where it is involved in hydro-gen-bond interactions with the side chains of Asn67,Asn71 and Glu111 (Fig. 3A and Table 2). Although the B 2  subsite has been shown to exhibit a strong basepreference in the order A > G > C > U, 28 only the inter-actions of purines in the B 2  site have been examined bycrystallography or NMR [complexes with d(Ap) 4 , 29 d(CpA), 30,31 UpcA, 32,33 2 0 ,5 0 , CpA, 31,34 d(ApTpApA), 35 ppA-3 0 -p, ppA-2 0 -p 18 , 3 0 ,5 0 ADP, 2 0 ,5 0 ADP, 5 0 ADP, 13 dUppA-3 0 -p, 36 pdUppA-3 0 -p, 15 and IMP 19 ], thus far.The RNase A–  3e  complex provides the first structuralinsights of uracyl binding to B 2 . The ribose binds awayfrom subsite R 2  towards the N-terminus of the proteinand it is held in place by participating in an extendedwater-mediated hydrogen-bonding network along withthe protein. The 4-carboxypiperidine group is pointingtowards the solvent in a location that is probably im-posed by the stereochemistry of the ligand (Fig. 3A).Inhibitor molecule II binds, with the 4-carboxypiperi-dine moiety, at the subsite B 1  where it is involved inhydrogen-bond interactions through its carboxyl groupwith Thr45. The uracyl group binds close to subsite P 0 almost parallel to the side chain of Lys66 involved inhydrogen-bond interactions with the side chains of His119 and Asp121 (Fig. 3B and Table 2). RNase A res- idues and inhibitor atoms are involved in a total of 55van der Waals contacts (Table 3).The structural mode of binding of   3e  to RNase A doesnot resemble any previous binding patterns for otherRNase A inhibitors. However, it seems that in  3e  mole-cule II the carboxyl group imitates the carbonyl groupsof uracyl and binds to subsite B 1 , whereas the uracylgroup binds close to P 1  engaging in hydrogen-bondinteractions with His119 and Asp121 and van der Waalsinteractions with Lys66 (the sole component of subsiteP 0 ). The interaction with Asp121 is also of importancefrom a mechanistic view since it has been shown thatAsp121 serves primarily to orient His119 properly tofulfill its catalytic function, 37 while replacement of Asp121 by Ala in mutation studies 37 diminishes k  cat / K  m  values for transphosphorylation by about  100-fold. Thus, it seems that the binding of   3e  to theperipheral binding sites stabilizes the productive confor-mation of the side chain of His119 through this interac-tion with Asp121 even though it does not bind directlyto subsite P 1 . This also provides an explanation to thefact that the side chain of His119 is in this conformationin the ligand complex despite the fact there is no groupbound at subsite P 1 . The binding of   3e  molecule I seemsto be dictated by the docking of uracyl to subsite B 2 ,while the rest of the molecule projects to the solvent.Upon binding to RNase A,  3e  molecules I and IIbecome buried (Fig. 4). The solvent accessibilities of the free ligand molecules I and II are 509 A˚ 2 and506 A˚ 2 , respectively. When bound these accessiblemolecular surfaces shrink to 254 A˚ 2 and 218 A˚ 2 , respec-tively. This indicates that approximately 50% and 57%of the ligand molecule I and II surfaces become buried.The greatest contribution for  3e  molecule I comes fromthe polar groups that contribute 159 A˚ 2 (62%) of thesurface, which becomes inaccessible, whereas for  3e molecule II, the greatest contribution comes from thenon-polar groups that contribute 198 A˚ 2 (69%). On theprotein surface, a total of 392 A˚ 2 solvent accessiblesurface area becomes inaccessible on binding of thetwo inhibitor molecules. The total buried surface area(protein plus two ligand molecules) for the RNase A–  3e  complex is 909 A˚ 2 . The shape correlation statistic D. D. Leonidas et al. / Bioorg. Med. Chem. 14 (2006) 6055–6064  6057  Sc, which is used to quantify the shape complementarityof interfaces and gives an idea of the ‘goodness of fit’ be-tween two surfaces, 38 is 0.72, and 0.69, for the associa-tion to the enzyme of   3e  molecules I and II,respectively, and 0.71 for the combined molecular sur-face of the two inhibitor molecules.Although the structure presented here is based on asoaking experiment, data from RNase A co-crystallizedwith 10 mM  3e  were also available at 2.0 A˚resolution.Preliminary analysis of this structure showed that theinhibitor is bound in exactly the same way as in thesoaked crystal.Kinetic studies showed that  3e  is a competitive inhibitorof the enzyme with a  K  i  = 103  l M at pH 6.0. 17 An elec-tron density map calculated from X-ray data of RNaseA crystals, soaked with 0.7 mM of compound  3e  (thehighest concentration used for the kinetic experiments 17 )in the crystallization medium for 2 h, showed only  3e molecule II bound at the peripheral binding site of theenzyme. It seems that this site has a higher affinity forthe ligand molecule than the site where ligand moleculeI binds, and therefore the inhibition profile observed inthe kinetic experiments for compound  3e  correspondsonly to the binding of   3e  molecule II to RNase A. 2.3. Comparative structural analysis Structural superposition of the RNase A–  3e  complexonto the RNase A–pdUppA-3 0 -p complex reveals thatthe uridine part of   3e  molecule I superimposes ontothe adenosine part, while the 4-carboxypiperidine andthe ribose of   3e  molecule II are close to the positionsoccupied by the uracyl and the 5 0 phospate group of pdUppA-3 0 -p (Fig. 5A). The superposition of the RNase Figure 2.  Stereo diagrams of the sigmaA 2| F  o |    | F  c | electron density maps calculated from the RNase A model before incorporating the coordinatesof the ligand are contoured at 1.0  r  level. The refined structures of the inhibitor are shown for  3e  molecules I (A), and II (B), respectively.6058  D. D. Leonidas et al. / Bioorg. Med. Chem. 14 (2006) 6055–6064  A–pdUppA-3 0 -p complex onto the RNase A–  3e  com-plex indicates also ways for improving the potency of  3e . Thus, following the molecular architecture betweenatom C4 0 of the ribose in the adenosine part to the 5 0 phosphate group in the uridine part of pdUppA-3 0 -p(Fig. 5A) we could propose a suitable linker betweenthe 5 0 hydroxyl group of   3e  molecule I to the 2 0 hydroxylgroup of   3e  molecule II. Such a linker might exploitadditional interactions with subsite P 1  of the RNase Acatalytic site which are not utilised by  3e .Superposition of the  3e  complex onto the 3 0 CMP com-plex 30 shows that only the carboxyl group of   3e  mole-cule II superimposes onto the cytidine of 3 0 CMP(Fig. 5B). Similarly, structural comparison of the bind-ing of   3e  to the binding of araUMP to RNase A, anoth-er class of non-natural 3 0 nucleotide RNase inhibitorsidentified recently, 14 shows that only the carboxyl moie-ty of   3e  molecule II binds closely to position of the ura-cyl ring in the araUMP complex. Both 3 0 CMP andaraUMP exploit interactions with subsite P 1  throughtheir phosphate group which cannot be utilised by com-pound  3e  since it is lacking such a group. However, theaffinity of   3e  for RNase A is comparable to that of 3 0 CMP ( K  i  = 103  l M) 39 but much lower than that of araUMP ( K  i  = 6  l M). 14 The weaker binding of   3e  com-pared to araUMP could be attributed to the phosphateinteractions of araUMP with residues in subsite P 1 which  3e  does not exploit. Based on the similar bindingaffinities of   3e  and 3 0 CMP it seems that the binding of the carboxyl group to subsite B 1  imitates well enoughthe cytidine binding at this subsite, while the hydro-gen-bond interactions between the uracyl of   3e  and theside chains of His119 and Asp121 compensate for thelack of interactions with other RNase A residues in sub-site P 1  that 3 0 CMP utilises upon binding to RNase A. 2.4. Modelling Although the present structure provides a promisingstarting point for the rational design of tight-bindingRNase inhibitors, it also reveals that this process may Figure 3.  Stereo diagrams of the interactions between RNase A and  3e molecules I (A), and II (B). The side chains of protein residues involvedin ligand binding are shown as ball-and-stick models. Bound watersare shown as black spheres. Hydrogen-bond interactions are repre-sented as dashed lines. NHOONOHONHOOHO N1C2N3O4C4C5C6O2N7C13C12C9C8C10C11C12 C13C1'C2'C3'C4'C5'O5'O4'O2' Scheme 1.  The chemical structure of   3e  with the numbering schemeused. Table 1.  Torsion angles for  3e  when bound to RNase AInhibitor molecule I II Backbone torsion angles O5 0  –C5 0  –C4 0  –C3 0 ( c ) 68 (+ sc ) 138 (+ ac )C5 0  –C4 0  –C3 0  –N7 ( d ) 113 ( +ac ) 137 ( +ac )C5 0  –C4 0  –C3 0  –C2 0  128   109C4 0  –C3 0  –C2 0  –O2 0 114 104 Glycosyl torsion angle O4 0  –C1 0  –N1–C2 ( v )   133 ( anti  )   166 ( anti  ) Pseudorotation angles C4 0  –O4 0  –C1 0  –C2 0 ( v 0 )   18   6O4 0  –C1 0  –C2 0  –C3 0 ( v 1 ) 13 14C1 0  –C2 0  –C3 0  –C4 0 ( v 2 )   4   16C2 0  –C3 0  –C4 0  –O4 0 ( v 3 )   6 13C3 0  –C4 0  –O4 0  –C1 0 ( v 4 ) 15   4Phase 103 ( O 4 0 - endo ) 162 ( C2 0 - endo )Definitions of the torsion angles are according to the current IUPAC-IUB nomenclature, 51 and the phase angle of the ribose ring is calcu-lated as described previously. 52 For atom definitions, see Scheme 1. D. D. Leonidas et al. / Bioorg. Med. Chem. 14 (2006) 6055–6064  6059
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks