A new 2'-hydroxyl protecting group for the automated synthesis of oligoribonucleotides.

We have developed a new type of 2'-hydroxyl protecting group for the automated machine synthesis of RNA oligomers: a 2-hydroxyisophthalate formaldehyde acetal (HIFA). The unique feature of this protecting group is that, as the bis ester, it is relatively stable to the acidic conditions that are used for repeated removal of dimethoxytrityl groups during chain elongation, but the final deprotection step in alkali, which cleaves the chain from the support and removes the base and phosphate protecting groups, converts it to the bis carboxylate and this can be removed relatively rapidly by treatment with mild acid. Conversion of the bis ester to the bis carboxylic acid increases the rate of acid-catalyzed hydrolysis of the acetal by 42-fold at pH 1, and, possibly, by 1320-fold at pH 3. The bis ester is 112 times more stable than the 1-(2-fluorophenyl)-4-methoxypiperidin-4-yl group (Fpmp) towards hydrolysis at pH 1, while the bis acid is only 2.35 times more stable than Fpmp at pH 3. In synthesis of the dimers UpU and UpG, with a coupling time of 5 min, the dimethoxytrityl cation assay indicated coupling yields of > 98%.

The evolution of nucleotides.

There has been much speculation about the types of molecules that were present in the first living forms. Recent discoveries show that RNA is a more versatile molecule than was previously believed, but whether it was ever able, for example, to synthesize its own monomers from available precursors is not yet known. If amino acids coexisted with nucleosides on the prebiotic Earth, then it seems likely that these two classes of molecules would have interacted with each other. We have been studying oligonucleotide-directed peptide bond formation, and during this work we discovered that aminoacylation of the internal 2'-hydroxyl groups of RNA occurred stereoselectively. Investigation of the mechanism of this reaction has been aided by the use of 3'-inosine methyl phosphate (as a simplified model for a dinucleoside monophosphate) and proton nmr spectroscopy of t-butoxycarbonyl-alanyl esters of nucleosides as models for the transition state of the aminoacylation reaction itself.

On the stereoselective aminoacylation of RNA.

Aminoacylation of the "internal" 2'hydroxyl groups of poly(A) with the imidazolide of racemic DL-leucine resulted in the formation of an enantiomeric excess of the D-ester, whereas aminoacylation with the imidazolide of racemic N-3,5-dinitrobenzoyl-DL-leucine gave an enantiomeric excess of the L-ester. Comparison of these results with those obtained earlier for alanine shows that the larger side chain tends to favor the formation of the D-ester. High resolution proton magnetic resonance spectroscopy was performed on the D- and L-alanyl "internal" 2'-esters of 3'5'-ApA. The resonance of the side chain methyl group of the D-ester is upfield from that of the L-ester, and the separation of the peaks increases with a decrease in temperature.

Stereoselective aminoacylation of a dinucleoside monophosphate by the imidazolides of DL-alanine and N-(tert-butoxycarbonyl)-DL-alanine.

The aminoacylation of diinosine monophosphate (IpI) was studied. When the acylating agent was the imidazolide of N-(tert-butoxycarbonyl)-DL-alanine, a 40% enantiomeric excess of the L isomer was incorporated at the internal 2' site and the positions of equilibrium for the 2' in equilibrium with 3' migration reaction differed for the D and L enantiomers. The reactivity of the nucleoside hydroxyl groups decreased in the order 2'(3') greater than internal 2' greater than 5', and the extent of reaction was affected by the concentration of the imidazole buffer (pH 7.1). In contrast, reaction of IpI with the imidazolide of unprotected DL-alanine led to an excess of the D isomer at the internal 2' site, while reaction with the N-carboxy anhydride of DL-alanine proceeded without detectable stereoselection. The relevance of these results to the evolution of optical activity and the origin of genetically directed protein synthesis is discussed.

Synthesis of 2'(3')-O-DL-alanyl hexainosinic acid using T4 RNA ligase: suppression of the enzymic reverse transfer reaction by alkaline phosphatase.

2'(3')-O-DL-Alanyl (Ip)5I was synthesized by a new method. An alanine ortho ester of inosine 5'-phosphate was added to (Ip)4I using the ATP-independent reaction of T4 RNA ligase, and the product was converted smoothly to the desired ester. The enzymic reverse transfer reaction was conveniently suppressed by the dephosphorylation of the adenosine 5'-phosphate coproduct, catalyzed in situ by alkaline phosphatase.

Geometry of the dry-state oligomerization of 2',3'-cyclic phosphates.

Evaporation of a solution of thymidine plus either the exo or the endo diastereomer of uridine cyclic 2',3'-O, O-phosphorothioate (U greater than p(S) in 1,2-diaminoethane hydrochloride buffer gave the 2',5' and 3',5' isomers of (P-thio) uridylylthymidine (Up(S)dT) in a ratio of 1:2 with a combined yield of about 20%. These isomers were re-converted to U greater than p(S) and dT by a reaction that is known to proceed by an in-line mechanism. Both the 2',5' and 3',5' isomers gave as product the same diasteromer of U greater than p(S) that had been used originally in their formation. These dry-state 'prebiotic' reactions (Verlander, Lohrmann, and Orgel 1973) are thus shown to be stereospecific, and both the 2',5' and 3',5' internucleotide bonds are formed by an in-line mechanism.

Reverse-phase HPLC of DNA restriction fragments and ribooligonucleotides on uncoated Kel-F powder.

Uncoated Kel-F powder offers some unique features as a support for reverse-phase HPLC of oligonucleotides and DNA restriction fragments. Compounds are eluted from the column by a gradient of acetonitrile (0 tto 18% v/v) in 0.1 M aqueous triethylammonium acetate. In contrast to RPC-5 chromatography, oligonucleotides are not eluted by aqueous salt solutions alone, and the separation of restriction fragments depends only on the chainlength. The packing material is cheap, easy to pack, chemically inert, and does not bleed, so that separations are highly reproducible. The DNA loading capacity for Kel-F is presently inferior to RPC-5, but recovery of microgram amounts of material is typically better than 50%.

Early chemical evolution of nucleic acids: a theoretical model.

Recent experimental work suggests a possible cyclical pathway for early prebiotic oligonucleotide formation that involves (i) dry-state (nontemplate) synthesis of random copolymers with mixed 2',5' and 3',5' bonds, (ii) passage of these oligomers into solution at low temperatures, and (iii) a preferential hydrolysis of the 2',5' bond in any short helices that have formed. This early system could have selected for complementary sequences that were largely 3',5'-linked, but may not have selected efficiently for a single enantiomer of ribose.

Hydrolytic stability of helical RNA: a selective advantage for the natural 3',5'-bond.

Dodecaadenylic acid containing a single 2', 5'-linkage at a defined position was formed by the coupling of two hexamers on a poly(U) template at 2 degrees. The rate of hydrolysis of this dodecamer was compared with that of a dodecamer that contained only the natural 3', 5'-linkages. At 40 degrees, in 1 M aqueous ethylenediamine at pH 8 in the absence of poly(U), both dodecamers hydrolyzed at comparable rates, but the addition of two equivalents of poly(U) caused a 7-fold increase in the initial rate of hydrolysis of the oligomer containing the 2', 5'-bond, and a 5-fold decrease in the initial rate of hydrolysis of the natural oligomer. When the oligomers are fully constrained in helical form, the ratio of the rates of cleavage of one 2', 5'-bond to one 3', 5'-bond under these conditions is probably about 900:1. The use of the 3', 5'-bond, in combination with a right-handed helix, appears to have had a large selective advantage over the use of the 2', 5'-bond for the storage of genetic information.

Nonenzymic joining of oligoadenylates on a polyuridylic acid template.

3',5'-Linked hexa-adenylic acid with a 2',3'-cyclic phosphate terminus [(A5A less than p] couples on a polyuridylic acid template in the presence of ethylenediamine to form the dodecamer (24 percent) and octadecamer (5 percent). The bond produced is largely that of the 2',5' isomer.

Geometry of the first step in the action of ribonuclease-A (in-line geometry-uridine2',3'-cyclic thiophosphate- 31 P NMR).

Advantage was taken of the reversibility of the first step of ribonuclease-A action to synthesize the dinucleoside phosphorothioate Up(S)C from the crystalline isomer of uridine 2',3'-cyclic phosphorothioate [Up(S)] and cytidine. Cyclic phosphorothioate was then reformed from Up(S)C by a nonenzymic reaction known to proceed by an in-line mechanism. The geometry of the enzymic reaction was determined to be in-line by a comparison of the Up(S) product with the Up(S) originally used. By the principle of microscopic reversibility, the geometry of the first step in the action of ribonuclease-A is shown to be in-line.

The hydrolysis of uridine cyclic phosphonate catalyzed by ribonuclease-A: implications for the mechanism of action of the enzyme.

Uridine 2',3'-cyclic phosphonate (I) is slowly hydrolyzed by ribonuclease-A with k(2) and K(m) values at pH6 that are respectively 1900 and 15 times smaller than the same parameters at the same pH for the related phosphate (II). Since the ratio of rate constants for hydroxide ion catalyzed hydrolysis is about 4, this result is consistent with, but does not prove a mechanism for the enzymic reaction that requires a pseudorotation.

On the mechanism of ribonuclease action.

All the suggestions that have been made for the mechanism of action of ribonuclease can be classified into one or another of two classes, depending on the geometry of the displacement in the first step. Studies of related model compounds indicate that one of these two classes should include as a part of the mechanism a pentacoordinate intermediate, which undergoes pseudorotation before product can be formed. It is unlikely that the pseudorotation would be rate-limiting in the reaction of the normal substrate.