Syntheses of Oligonucleotide-Amlno Acid Conjugates: Using TentaGel and CPG Matrices for the Synthesis of 3'-Phosphoryityrosine-Terminated Oligonucleotides

Conservative site-specific recombinases and topoisomerases cleave DNA via the nucleophilic attack of a hydroxy group of a hydroxylated amino acid (usually tyrosine) upon an intemucleotide phosphodiester bond. t'2 As a result, two shorter DNA strands are formed, one with a free 5'-OH and the other bound to the enzyme through a 3'-phosphodiester linkage involving the hydroxyl of the amino acid. The amino acid acts as the leaving group in the nucleophilic substitution by a primary hydroxy group of another oligonucleotide. As a result, the intemucleotide phosphodiester bond is resealed (Scheme 1).

Oligonucleotides that are terminated in the 3'-phosphoryltyrosine have been used as substrates for the study of the strand joining reaction catalyzed by conservative site-specific recombinases such as the Flp recombinase and by mammalian topoisomerase I. l-3 We previously reported the synthesis of 3'- phosphoryltyrosine-termL'tated oligonucleotides using a Teflon fibre-hased solid support (Scheme 2; upper portion). When this support was no longer available, we were obliged to explore the use of other matrices. As shown in Scheme 2, the original synthesis began with tyrosine whose hydroxyl was protected by O-tbutyl. I~  After linking the protected tyrosine to the Teflon support, the t-butyl group was removed by thetreatment with trifluoroacetic acid. The oligonucleotide synthesis started on the hydroxyl group of tyros/he, and it used the standard phosphoramidite chemistry with an ABI automatic oligonucleotide synthesizer. At the completion of the oligonucieotide synthesis, Fmoc was removed by the treatment with 30% aqueous NH 3 The 3'-phosphoryltyrosine terminated oligonucleotide was released from the Teflon support at the same time through the ammonolysis of the ester bond between the support and tyrosine resulting in the formation of

a new carboxylic amide

Since the ftmctiol~,'tliTed Teflon support recently became unavailable, we sought an alternative support. We
first investigated the use of a TentaGel support (Rapp Polymere, Tfibingen, Germany). TentaGels are polymers of ethylene glycol grafted onto polystyrene. While they are used primarily for polypeptide syntheses, + they were recently utilized for oligonucleotide syntheses as well 3 and their high loading capacity was cited 5 as their primary advantage over CPG. Tentagels seemed to be sufficiently stable under the conditions of the Teflon-based synthesis. The synthetic protocol with which the automatic synthesizer operates on both CI'G and Teflon supports has been modified for oligonucleotide synthesis on TentaGels. 6 Although the higher loading capacity of TentaGels over CPG was noticeable in the synthesis of shorter (< 20 nucleotides), the lower yields per cycle resulted in generally lower yields of tyrosine-oligonucleotide conjugates when synthesis of longer oLigonucleotides was attempted (see below). Consequently we have turned our attention to synthesis with CPG. 7 Since the CPG solid support was not sufficiently acid stable we had to avoid treatment with trifluoroacetic acid. We sought a protective group which could be removed from the tyrosine hydroxyl under non-acidic conditions, s The levulinoyl 9 ester group satisfied this requirement as NaBI-~ is the most suitable deprotecting reagent. 9,m Hydrazine 9b cannot be used as a deprotection agent in the presence of ester groups.

The complete synthesis is portrayed in Scheme 2. Briefly, the starting Fmoc tyrosine was treated with levulinic anhydride in pyridine at room temperature, tt The O-levulinoyl tyrosine which was formed in excellent yield was subsequently linked with a functionalized alkylamine CPG column nb using dicyclobexyl carbodiimide, t,2 This heterogeneous condensation as well as the subsequent removal of the levulJnoyl group with sodium borohydride n proceeded smoothly. The free hydroxyl of tyrosine then assumed the role of the first 5'-hydroxyl in the standard oligonueleotide synthetic procedure. Comparing efficiency of the oligonucleotide synthesis on all three supports, we noted that the yields per cycle were comparable when CPG and Teflon were used (above 98%). The yields per cycle on Tentagel were lower (90-91%) so that we were not able to synthesize a 36-mer on Tentagel. The products also differed in purity. After cartridge purification, the 3'-phosphoryl tyrosine-terminated 18-oligonucleotide conjugates were labelled with 321' at their 5'-termini with polynuclenfide kinase t3, annealed with the two unlabelled oligonuclentides to form a synthetic ligation substrate analyzed by electrophoresis (Fig.l; eddnumbered lanes)) 18nt conjugates made on Teflon and CPG both contained approximately 35 % of impurities as measured by phosphoimage analysis (see Fig. 1, Panel A, lanes 1 and 5). The same conjugate made on Tentagel was free of these impurities (Panel A, lane 3), while the impurities could be removed from the conjugates by electrophoresis n (cf.Panel B, lane 1). We further compared the effectiveness of the three oligonucleotide conjugates in the Flp-mediated ligation assay. The products formed after incubation with Flp were analyzed on a denaturing polyacrylamide gel (Fig.l; even-numbered lanes). As seen in Fig.l, t3 each of the synthetic oligonucleotides gave equally efficient ligation as evidenced by the appearance of a rip_labelled, 40-nucleotide product. The gels show that the impurities do not inhibit the llgation reaction.