ALBERT

Notes: Chemistry of α-Aminonitriles. Aldomerisation of Glycolaldehyde Phosphate to rac-Hexose 2,4,6-Triphosphates and (in Presence of Formaldehyde) rac-Pentose 2,4-Diphosphates: rac-Allose 2,4,6-Triphosphate and rac-Ribose 2,4-Diphosphate Are the Main Reaction ProductsGlyclaldehyde phosphate aldomerizes in aqueous NaOH solution to a product mixture containing the racemates of the two diastereoisomeric tetrose 2,4-diphosphates and eight hexose 2,4,6-triphosphates. At room temperature in the absence of air and after 7 days, a solution 0.08Min glycolaldehyde phosphate (=formylmethyl dihydrogenphosphate)and 2M in NaOH gives products, in up to 80% yield, with a tetrose/hexose derivative ratio of ca 1:10 and with rac-allose, 2,4,6-triphosphate comprising up to 50% of the mixture of sugar phosphates. When the reaction is run under the same conditions but in the presence of 0.5 mol-equiv. of formaldehyde, sugar phosphates are formed in up to 45% yield, with pentose 2,4-diphosphates now predominating over hexose triphosphates by a ration f 3:1 rac-Ribose 2,4-diphosphate is found to be the major component, the ratios am ribose, arabinose, lyxose, and xylose 2,4-diphosphates being 52:14:23:11 in a representative experiment. The pentose diphosphates are constitutionally stable under the reaction conditions (observed for 23 weeks), but the diastereoisomeric ratios slowly change with time (tc 22:34:30:14 after 23 weeks), showing that ribose 2,4-diphosphate is not the thermodynamically favored diastereoisomer. The observed product distributions in both the pentose and the hexose series (after 1 week) reveal an aldolization mode that is preferentially erythro in the product-determining step (the reaction of glycolaldehyde phosphate as its enolate with glycerinaldehyde 2-phosphate and tetrose 2,4-diphosphate, respectively). An attempt is made to rationalize both this fact and the kinetic predominance of ribose 2,4-diphosphate in the pentose series and allose 2,4,6-triphosphate in the hexose series. Their configuration along the C-chain can be interpreted as corresponding to a minimum number of 1-5 repulsions in the packing of phosphate and OH substituents and minimal steric interaction between substituents at the reaction centers in the transition state of the product-determining step. The aldomerization of glycolaldehyde phosphate in the presence of formaldehyde is a variant of the formose reaction, It avoids the formation of complex formose product mixtures, largely as a consequence of the fact that aldoses which are phosphorylated at the C(2)

Notes: Qualitative conformational analysis of the entirety of conceivable hexo- and pentopyranosyl oligonucleotide systems derived from the diastereoisomeric aldohexoses (CH2O)6 and aldopentoses (CH2O)5 predicts the existence of a variety of pairing systems which have not been experimentally investigated so far. In particular, the analysis foresees the existence of a ribopyranosyl isomer of RNA (‘p-RNA’), containing the phosphodiester linkage between the positions C(4′) and C(2′) of neighboring ribopyranosyl units. Double strands of p-RNA oligonucleotides are expected to have a linear structure and to show purine-pyrimidine and purine-purine (Watson-Crick) pairing comparable in strength to that observed in homo-DNA. Experimentally, synthetic β-D-ribopyranosyl (4′→2′)-oligonucleotides derived from adenine and uracil confirm this prognosis: adenine-uracil pairing in p-RNA duplexes is stronger than in the corresponding RNA duplexes. Importantly, adenine in p-Ribo(A8) does not show (reverse-Hoogsteen) self-pairing, in sharp contrast to its behavior in the homo-DNA series. The sheer existence of strong and selective pairing in a system that is constitutionally isomeric to RNA and can be predicted to have a linear structure has implications for the problem of RNA's origin. In this context, a comprehensive experimental study of the pairing properties of p-RNA, of its potential for constitutional assembly, self-replication, and intra-duplex isomerization to RNA seems mandatory.

Notes: Base pairing in p-RNA (β-D-ribopyranosyl-(4′ → 2′)-oligonucleotides) is not only stronger than in DNA and RNA, but also more selective in the sense that it is strictly confined to the Watson-Crick mode. Homopurine sequences (tested up to decamers) exist as single strands under conditions where they undergo reverse-Hoogsteen self-pairing in homo-DNA or Hoogsteen self-pairing in DNA. This exceptional pairing selectivity is rationalized as hinging on two structural features of p-RNA: the large inclination between backbone axis and base-pair axes in p-RNA duplexes, and the higher rigidity of the p-RNA backbone compared with RNA, DNA, and homo-DNA. The most important consequence of the pairing selectivity refers to the potential of p-RNA to replicate. Replicative copying of sequence information by nonenzymatic template-controlled ligation is not hampered by self-pairing of guanine-rich templates, as it is known to be the case in the RNA series. We have demonstrated two replicative cycles in which G-rich p-RNA-octamer templates induce sequence-selective ligation of tetramer-2′-phosphate derivatives to complementary C-rich octamer sequences, and in which the latter, with comparable efficiency, induce corresponding ligation reactions back to the original G-rich octamers. Ligation is most satisfactorily achieved after pre-activation of the 2′-phosphate groups as 2′,3′-cyclophosphate derivatives; in this version, the process does not proceed as oligocondensation, but as a genuine oligomerization. This is of considerable promise for the search for potentially natural conditions under which homochiral p-RNA strands might self-assemble and self-replicate.

Notes: Replication (single-turnover) of pyranosyl-RNA (= p-RNA) sequences can be accomplished reliably by template-directed ligation of 2′, 3′-cyclophosphates of short oligomers. The ligation process was studied using (mostly) octamers as templates and tetramers as ligands. The transcription of the sequence pr(GGGCGGGC) into the (antiparallel) complementary sequence pr(GCCCGCCC) by ligation of two molecules of pr(GCCC)-2′, 3′-cp was investigated in detail; In aqueous 1.5M LiCI solution of pH 8.5 at room temperature (0.45 mM ligand, 0.15 m,M template), the reaction proceeds in up to 60% yield within a week. It is limited by concomitant hydrolysis of cyclophosphate groups of both reactand and ligation product as the only efficient side reaction, the latter occurring ca. three times more slowly than ligation. No ligation at all is observed in the absence of template. The reaction is highly regioselective: the (4′ → 2′) phosphodiester junction is formed exclusively; no isomeric (4′ → 3′) junctions are found. For ligation to occur, template and ligand must be homochiral and must have the same sense of chirality; with chiro-diastereoisomeric tetramer-2′, 3′-cyclophosphates containing a single enantio-ribopyra-nosyl unit, no ligation is observed, except to a minor extent in the case of the diastereoisomer that has that unit at the 4′-end. Observations made in experiments involving six different octamer templates containing isomeric base sequences indicate that the ligation process does not tolerate a mismatch at ligation sites. However, ligation still takes place when a mismatch occurs at either end of the (octamer) template. Ligation efficiencies differ widely, depending on the nature, as well as the sequence, of participating bases. These differences can be understood qualitatively by considering the relative stability of ternary pre-ligation complexes, together with the differences in interstrand base stacking at ligation sites, Dominance of the latter over intrastrand base stacking is the feature of the p-RNA structure that appears to determine most of the characteristic properties of p-RNA.As regards the etiological context of our work on nucleic-acid alternatives, it is essential that the chemical properties found for p-RNA be compared with the corresponding properties of RNA. In the RNA series, the two ligations of the replicative cycle r(GGGCGGGC) ↔ r(GCCCGCCC) using the corresponding ribofuranosyl-te-tramer 2′, 3′-cyclophosphates as ligands are found to proceed also, though somewhat less efficiently than in the p-RNA series; however, the ligation step produces exclusively the unnatural (5′ → 2′) phosphodiester junctions instead of the natural (5′→ 3′) junctions. This is in sharp contrast to p-RNA, where template-controlled 2′, 3′-cyclophosphate ligations produce the ‘correct’ phosphodiester junctions.

Notes: Chemistry of α-Aminonitriles. Formation of 2-Oxoethyl Phosphates (“Glycolaldehyde Phosphates”) from rac-Oxiranecarbonitrile and on (Formal) Constitutional Relationships between 2-Oxoethyl Phosphates and Oligo(hexo- and pentopyranosyl)nucleotide BackbonesOxiranecarbonitrile in basic acqueous solution at room temperature reacts regioselectively with inorganic phosphate to give the cyanohydrin of 2-oxoethyl phosphate (“glycolaldehyde phosphate”), a source of (the hydrate of) the free aldehyde, preferably in the presence of formaldehyde. In aqueous phosphate solution buffered to nearly neutral pH, oxiranecarbonitrile produces the phosphodiester of glycoladehyde as its bis-cyanohydrin in good yield. In contrast to mono- and dialkylation, trialkylation of phosphate with oxiranecarbonitrile is difficult, and the triester derivative is highly sensitive to hydrolysis. Glycolaldehyde phosphate per se is of prebiotic interest, since it had been shown [5] to aldomerize in basic aqueous solution regioselectively to rac-hexose 2, 4, 6-triphosphates and - in the presence of formaldehyde - mainly to rac-pentose 2, 4-diphosphates with, under appropriate conditions, rac-pentose 2, 4-diphosphates as the major reaction product. However, the question as to whether oxiranecarbonitrile itself has the potential of having been a prebiological natural constituent remains unanswered.Backbone structures of hexopyranosyl-oligonucleotides with phosphodiester linkages specifically between the positions 6′ → 4′, 6′ → 2′, or 4′ → 2′ of the sugar residues can formally be derived via the (hypothetical) aldomerization pathway, a combinatorial intermolecular aldomerization of glycoladehyde phosphate and bis(glycolaldehyde)-phosphodiester in a 1: 1 ratio. The constitutional relationships revealed by this synthetic analysis has played a decisive role as a selection criterion in the pursuit of our experimental studies toward a chemical etiology of the natural nucleic acids' structure. The Discussion in this paper delineates how the analysis contributed to the conception of the structure of p-RNA.The English Footnotes to Schemes 1-11 provide an extension of this summary.

Notes: The solution structure of the duplex formed by self-pairing of the p-RNA octamer β-D-ribopyranosyl-(2′→4′)-(CGAATTCG) was studied by NMR techniques and, independently, by molecular-dynamics calculations. The resonances of all non-exchanging protons, H-bearing C-atoms, P-atoms, and of most NH protons were assigned. Dihedral angle and distance constraints derived from coupling constants and NOESY spectra are consistent with a single dominant conformer and corroborate the main structural features predicted by qualitative conformational analysis. The duplex displays Watson-Crick pairing with antiparallel strand orientation. The dihedral angles β and ∊ in the phosphodiester linkages differ considerably from the idealized values. Model considerations indicate that these deviations from the idealized model allow better interstrand stacking and lessen unfavorable interactions in the backbone. The average base-pair axis forms an angle of ca. 40° with the backbone. The resulting interstrand π-π stacking between either two purines, or a purine and a pyrimidine, but not between two pyrimidines, constitutes a characteristic structural feature of the p-RNA duplex. A 1000-ps molecular-dynamics (MD) calculation with the AMBER force field resulted in an average structure of the same conformation type as derived by NMR. For the backbone torsion angle ∊, dynamically averaged coupling constants from the MD calculation agree well with the experimental values, but for the angle β, a systematic difference of ca. 25° remains. The two base pairs at the ends of the duplex are calculated to be highly labile, which is consistent with the high exchange rate of the corresponding imino protons found by NMR.

Notes: Enantiomeric oligoribonucleotides (= ent-RNA) up to a sequence length of thirty-five and consisting of the (L-configurated) nucleosides ent-adenosine, ent-guanosine, ent-cytidine, ent-uridine, and 1-(β-L-ribofuranosyl)thymine were prepared by automated synthesis from appropriate building blocks, carrying a known photo-labile 2′-O-protecting group. A simple large-scale synthesis of the new, prefunctionalized L-ribose derivative 5 from D-glucose (Scheme 1) and its straightforward conversion into the five phosphoramidites 28-32 and five solid supports 38-42, respectively, were elaborated (Scheme 4). Within this project, a novel, superior strategy for the synthesis of the 2′-O-{[(2-nitrobenzyl)oxy]methyl}-substituted key intermediates 18-22 by regioselective alkylation of their 5′-O-dimethoxytritylated precursors 13-17 was developed. Furthermore, an improved set-up for the final light-induced cleavage of the 2′-O-protecting groups from the oligonucleotide sequences was designed (Scheme 5 and Fig. 1). The correct composition of all ent-oligoribonucleotides prepared was established by their MALDI-TOF mass spectra. The 1H-NMR-spectroscopic data of a dodecameric ent-RNA sequence was in excellent agreement with the published data of its natural counterpart, synthesized by conventional methods. The known specific cleavage of a tetradecamer sequence by a 35mer ribozyme structure could be reproduced by ent-oligoribonucleotides, synthesized by the presented methods (Fig. 4).