No description available.

No description available.

Graduation date: 1997

Not available / text

This thesis presents the crystal structures of 2', 5', cyclic and modified nucleosides / nucleotides. Chapter I gives a brief account of the structural studies on 2', 5' and modified nucleotides. It also presents a short, summary of unusual nucleic acids structures studies on hydration patterns and metal ion interactions Nomenclature and conventions used for describing the conformatioNa1 features are presented. FiNa1ly, the crystallographic suite of programs used for processing the intensity data, structure solution, refinement and generating various diagrams are mentioned. Chapter II describes the crystal structures of anhydrous and hydrated sodium salt of N6-methyl adenosine-S'-monophosphate. N6-AMP-A (anhydrous form) belongs to the trigoNa1 space group P3221 with unit cell dimensions a = b = 10.30 A and c= 25.03 A while N6-AMP-H (hydrated form) belongs to orthorhombic space group C222X with a= 6.910 A, b= 19.318 A, and c= 41.070 A. CuKα intensity data consisting of 1740 and 2740 observed reflections were collected on a CAD4 diffractometer. Both structures were solved using SHELXS97 and refined to R factors of 0.0336 and 0 0381 for anhydrous and hydrated forms respectively. In both structures the adenine bases are in the ant% conformation with respect to the ribose but their torsion angles XCN differ significantly by 78° The ribose moiety shows CS-endo puckering and the conformation about the C4/-C5/ bond is g+ and t in the anhydrous and hydrated structures respectively. The two Na+ ions, present m the hydrated form, coordinate with water oxygen atoms only. A notable feature of the Na+ ion coordination in the anhydrous form is the participation of N3 and N7 of the base besides macrochelation between base-ribose and base-phosphate moieties. Adenine bases in both forms stack at a separation of about 3.4 A between them N6-AMP molecules pack as if one set of bases intercalate between the other set in the hydrated structure while they form helix like pattern m the anhydrous structure Molecular dynamics calculations were carried out for both structures with a view to obtain greater insight into the effect of hydration on the conformation of the molecule. Stereochemically permissible models for poly-A using the N6-AMP-H coordinates were generated using the method developed by Srinivasan and Olson. Its features and possible biological relevance are discussed. Chapter III deals with the structure of sodium adenosine-5'-monosulfate trihydrate (5'-AMS). Intensity data for this modified nucleotide were collected at the Brookhaven NatioNa1 Laboratory, Synchrotron facility, USA. 5'-AMS belonged to the orthorhom bic space group P2!2!2i with unit cell parameters a= 20.698 A, b= 24.621 A and c= 25.925 A and eight molecules, eight Na+ ions and 23 water molecules in the asymmetric unit of the lattice. Never before a nucleotide structure having eight molecules in the asymmetric unit has been reported. Out of 84177 reflections collected using a radiation of A =0.92 A, 9108 independent reflections having Io>2a(Io) were considered observed. The structure was solved using the program Shake and Bake (SnB) and refined by, SHELXL97. The fiNa1 R factor for 1971 parameters was 0.0397. Adenine bases of all the eight 5'-AMS molecules are in anti conformation with respect to the ribose moiety with XCN angles varying from -150 to -177°. But the conformations of the ribose moieties and the sulfate groups about the C4/-C5/ bond are not the same for all the molecules. 5'-AMS molecules A, B and D show C2-exo-C3-endo mixed puckering while C has C£-exo puckering. The remaining four molecules E, F, G and H have C3-endo conformation. The conformation about the C4/-C5/ bond for molecules A, B, C and D is g~ while for E, F and G it is g+. Molecule H shows both g+ and g~ since the 05' atom is disordered. An important feature of the metal ion coordination is the bidentate formation by sodium ions Na3 and Na7 with the sulfate group of molecule C and ribose hydroxyl oxygen atoms of molecule D respectively. Another feature which deserves mention is the participation of Nl and N7 of the adenine base m metal coordination Adenine bases of molecules A, B, C and D form self pairs with those of H, G, F and E respectively through N6...N7 and N6...N1 hydrogen bonds. The 5'-AMS molecules pack as duplexes in the unit cell. A Stereochemically permissible model for poly-A with sugar sulfate backbone using the 5'-AMS coordinates were generated using the method developed by Srinivasan and Olson and its features are discussed. Crystal structures of two polymorphs of mixed sodium and potassium salts of cytidine-5'-monophosphate hexahydrate are discussed in Chapter IV. The two polymorphs of 5'-CMP were grown using methanol and DMF respectively m the crystallization experiments. MoKα intensity data for CMP-I were collected on a Rigaku AFC image plate system while that for CMP-II were collected on a Bruker CCD Smart system. Both belong to the monoclinic space group P2X with a= 8.869 A, b= 20 580 A, c= 23.179 A, β= 105.79° and a= 8.929 A, b= 22.257 A and c= 20.545 A, β= 90.02° for CMP-I and II respectively. The the unit cell volume of the two polymorphs differ by just 12 A3 as the unit cell parameters are same, although the b and c axes are interchanged m CMP-II and their β value differs by 16°. Both polymorphs of CMP have four nucleotide molecules in the asymmetric unit of their orthorhombic lattices. But the number of metal ions and solvents are not the same in the two structures. CMP-I has five sodium ions, three potassium ions, 23 water and two methanol molecules while CMP-II has two sodium ions, four potassium ions, 22 water and an unknown solvent molecule (assigned as dimethyl ether) in the asymmetric unit. This is the first nucleotide structure having two different alkali metal ions (Na+ and K4") in the crystal structure. Out of 36946 and 31293 reflections collected 12247 and 15476 independent reflections having IO>2<J(I0) were considered observed for CMP-I and II respectively. Both structures were solved by combination of heavy-atom and direct methods using DIRDIF96 and refined using SHELXL97 to R factors of 0.0819 and 0 0867 for CMP-I and II respectively In both forms all the four molecules have anti conformation about the glycosidic bond, CS-endo conformation for the ribose moiety and g+ conformation about the C4'-C5' bond but their metal coordination patterns are significantly different. K1 ion in CMP-I forms an intra molecular macrochelate between the ribose and adenine base while K2 and K3 ions form bidentates with the cytosine and phosphate group of molecules A and D respectively. Na1, Na3 and Na5 are all involved in bidentate interactions with the ribose of molecule C, ribose of molecule A and phosphate of molecule D respectively. In contrast, Na2 and Na4 coordinates with solvent atoms only and do not interact with the nucleotide atoms at all. K1 and K2 ions of CMP-II form bidentates with the cytosines of molecules C and D respectively while K2 and K4 form intra molecular macrochela-tion between the base and ribose of molecules C and B respectively. Na1 and Na2 form bidentates with the ribose of molecules C and D respectively Comparison of the two polymorphs of CMP reveals that despite several striking conformatioNa1 similarities there are also significant differences between them. It was noticed that molecules A, B, C and D of CMP-I corresponds to C, B, D and A of CMP-II. Out of eight metal ions (five Na+ and three K+ ions) present in CMP-I four of them (Kl, K2, K3 and Na3) are found to have partners (K4, Kl, K3 and Na1) in CMP-II within a distance of 0.75 A. One of the water molecules OW8 of CMP-I is replaced by a potassium ion K2 in CMP-II within a distance of 0.92 A. Out of 23 water molecules present in the structure 14 are common to both of them and only 8 are different while one is replaced by an ion. The four crystallographically independent 5'-CMP molecules are linked by metal ions Kl, K3, Na1, Na3, Na5 and Kl, K2, K4, Na1 ions forming a tetramers in CMP-I and CMP-II respectively. An interesting feature of CMP-I and CMP-II is the simultaneous display of base-base and base-ribose stacking patterns. The four nucleotide molecules in the asymmetric unit are related by several pseudo two-fold axis and the r.m.s. Deviations between them after applying the pseudo symmetry are 0.21 and 0.17 A for CMP-I and II respectively. The nucleotide molecules in CMP-I and II pack as infinite linear chains parallel to the b and c axis respectively which repeat along the c and b axis respectively. In between these nucleotide columns metal ions and water molecules are located forming channels between them. Chapter V deals with anhydrous cytidine-2/-phosphate and potassium uridine-5'-phosphate hexahydrate structures. 2'-CMP crystallizes m the orthorhombic space group P212121 with a= 6.698 A, b= 7.436 A and c= 25.291 A with one molecule in the asymmetric unit. MoKα intensity data were collected on a CCD SMART system consisting of 7647 reflection of which 1456 independent reflections having lo>2a(lo) were considered observed. The structure was solved and refined to an R factor of 0.0385 for 186 parameters using SHELXL97. The cytosine base is in the anti conformation with respect to the ribose with XCN = -141.1° similar to that in the hydrated structure. But it differs significantly from the syn conformation observed in several 2'-purine and 2'-5' dinucleotide structures containing purine-pyrimidine sequences. The ribose moiety shows Ctf-endo and the conformation about the C4/-C5/ bond is t with (f)α = 169.3° and <pO( = -72 7° The t conformation in the anhydrous form is different from the g+ conformation m the hydrated form of 2'-CMP 5'-UMP.K crystallizes in the monoclinic space group P2;i with a= 13 034 A, b= 8 916 A, c= 16 205 A and β=98 64° with two nucleotides, four K ions and ten water molecules in the asymmetric unit MoKα intensity data of 19261 were measured on a Bruker CCD system of which 6891 independent reflections having lo>2a(lo) were accepted as observed. The structure was solved and refined by full matrix least square methods to an R factor of 0.0324 for 609 parameters. Uracil bases of both nucleotide molecules are in the anti conformation with respect to the ribose with XCN= -129.4° and -132 7°. Uracil bases of both nucleotide molecules are protonated at N3 Both ribose moieties show C2’-endo puckering with C2' atom displaced by 0.57 and 0.59 A from the best plane constituted by the remaining atoms The phosphate group is in a staggered orientation and the conformation about the C4/-C5/ bond is g* with <j>00 = -67.1 and -62.7 and Øoc = 54.6 and 59.5 for molecules A and B respectively. Potassium ion K2 forms a bidentate by coordinating with ribose 02' and 03' atoms of molecule B and a macrochelate between the uracil base and ribose of molecule A by coordinating with 02 and 02' atoms. K4 also forms a bidentate by coordinating with ribose O2' and 03' atoms of molecule A. The two 5'-UMP molecules form a dimer by coordinating with K2 and K3 ions. They are related by a pseudo two-fold axis and the r.m.s. deviation between the coordinates is 0 12 A. Crystal structures of 8-Benzylamino cychc-3'-5'-monophosphate (8-Benz-cAMP) and 8-mercaptoguanosine (8-MERG) are presented in Chapter VI. 8-Benz-cAMP crystallizes in the monoclinic space group P2x with unit cell dimensions a= 7.989 A, b= 12 589 A, c= 11.773 A and β= 93.82°. MoKα data were collected on a CCD system yielded 4331 independent observed reflection with Io2cr(Io) out of 9733 reflections collected. The structure was solved and refined to a R factor of 0 0451 with 367 parameters. The adenine base is in the syn conformation with XCN= 84.7° as in few other 8-substituted cyclic purine nucleotides but different from the simple cyclic purine nucleotides. The phenyl moiety is in the trans conformation with respect to the base. The ribose moiety shows rare C4’-exo puckering with a deviation of 0.70 A from the best plane constituted by the remaining four atoms. The 05' atom is m the t conformation with respect to the ribose with cpα = -174.8° and <pα = -59.6° since only in this conformation 3' and 5' cyclization is possible. Hydrogen bonds Nl. .O1P and N6...O5' link two nucleotide molecules. Adenine bases stack on the phenyl ring from above and below. The only water molecule present in the structure form hydrogen bonds with the nucleotide atoms. 8-mercaptoguanosine crystallizes in the monoclinic space group C2 with unit cell dimensions a= 23.246 A, b=9.751 A, c= 6.406 A and b= 90.91°. MoKα intensity data collected on CAD diffractometer yielded 2683 independent observed reflections having I0>2<r(I0). The structure was solved using SHBLXS 97 and refined using SHELXL97 to a R factor of 0.0565. The guanine base is in the syn conformation with XCN= 64.1°. The ribose ring shows C2-endo puckering with C2' atom deviating by 0.62 A from the best plane. An interesting feature of this structure is the intra-molecular hydrogen bond between the base N3 and the ribose 05' atoms. The last chapter (VII) describes the crystal structures of three modified adenine nucleosides N6-benzyl adenosine (N6-BA), N6-cyclohexyl adenosine (N6-CA) and 5'-trityl adenosine (5'-TA). N6-BA belongs to the triclinic space group PI with a= 5.008 A, b= 8.921 A, c= 9.762 A and a = 111.73°, β= 90.37°, 7 = 91.42° while N6-CA and 5'-TA belong to the monoclinic space group P2i with a= 12.205 A, b=15.265 A, c= 15.095 A, P = 110.64° and a= 8.823 A, b= 15.613 A, c= 10.078 A and β = 115.01° with three and one molecules in their asymmetric units respectively. The three structures of N6-BA, N6-CA and 5'-TA were solved and refined to R factors of 0.0355, 0.0655 and 0.0262 using 1656, 7549, 2473 independent reflections and 244, 677, 360 parameters respectively using SHELX97. The adenine base of N6-BA is in the anti conformation with XCN= 168.9°. The benzyl moiety is in the distal geometry with respect to the imidazole ring. The furanose ring shows CSI-exo-CS'-endo mixed puckering. There are several 7r-7r interactions observed in this structure. In contrast to N6-BA all three molecules of N6-CA show syn conformation about the glycosidic bond with XCN= 47.0°, 54 8°, 49 V for molecules A, B and C respectively. The cyclohexyl moiety of all three molecules are in the chair conformation. The ribose moieties of all three molecules show C2-endo puckering with C2' atom deviating by 0 59, 0 54, 0.57 A for molecules A, B and C respectively. The adenine base of the 5'-TA is in the anti conformation with \Cs= 168.4° and the ribose moiety shows C2-endo puckering The three phenyl rings of the trityl group are in staggered orientation. Interesting tape formation via N6…02' and N7...O3' hydrogen bonds is observed in all three nuclosides.

Les séquences riches en guanines, qu’elles soient ADN ou ARN, peuvent former des structures non canoniques à quatre brins appelées quadruplexes de guanines ou G-quadruplexes. Ces structures reposent sur la formation et l’empilement de quartets de guanines ; elles peuvent être trouvées dans de nombreuses régions du génome. Des motifs G-quadruplexes apparaissent fréquemment lors de la sélection d'aptamères par SELEX : on constate un biais dans la proportion de guanines par rapport aux autres nucléotides dans les bases de données regroupant les séquences d'aptamères connus. Nous avons entrepris une analyse systématique, in silico puis in vitro, de motifs aptamères décrits dans la littérature. Nous avons utilisé un algorithme de prédiction actuellement développé au sein du laboratoire dans le but de déterminer quelles sont les séquences susceptibles de former des G-quadruplexes in silico. Afin de vérifier ces prédictions, un test biophysique permettant de cribler rapidement de nombreuses séquences a été mis en place. Nos résultats démontrent que de nombreux aptamères publiés sont susceptibles de se replier en G-quadruplexe. Un autre volet de ce travail concernait la mise en pratique du SELEX avec deux objectifs distincts. Nous avons d'abord effectué une sélection in vitro contre un ligand de G-quadruplexes, afin de préciser quels motifs nucléiques peuvent interagir avec cette molécule. Nous avons réalisé ensuite des expérience de SELEX contre plusieurs G-quadruplexes ADN ou ARN biologiquement pertinents (motif télomérique humain, répétitions minisatellites et séquences présentes dans les UTR de différents ARNm), dans le but d'obtenir des sondes spécifiques de ces conformations. / Guanine-rich DNA or RNA sequences can adopt non-canonical structures composed of four strands called guanines quadruplexes or G-quadruplexes. These structures are made by the formation and stacking of guanine quartets; they can be found in various region of the genome. G-quadruplexes motifs are frequently found during the selection of aptamers by the so-called SELEX method: a bias exists in the proportion of guanines in comparison with other nucleotides in a database gathering together known aptamers sequences. We did an in silico then in vitro systematic analysis of aptameric motifs described in the literature. We used a prediction algorithm currently developed in the laboratory to determine which sequences car form G-quadruplexes in silico. In order to check those predictions, a biophysical test allowing to quickly assay a lot of sequences was set up. Our results demonstrate that a lot of published aptamers are able to fold into G-quadruplexes. Another part of this work is related to the use of the SELEX with two different goals. First of all we did an in vitro selection against a G-quadruplexes ligand, in order to tell exactly which nucleic motifs can interact with this molecule. We then carried out SELEX experiments against several DNA or RNA biologically relevant G-quadruplexes (human telomeric motifs, minisatellite repetitions and sequences from UTR of mRNA), with the aim of getting specific probes for these conformations.