Universal linker phosphoramidite

paper


Introduction
Standard oligonucleotide synthesis uses a solid support that contains the first nucleoside covalently bound to the solid phase by a linker that is hydrolyzed during the cleavage step following solid-phase assembly.This support-bound nucleoside becomes the 3'-terminal residue of the final oligonucleotide after the cleavage and deprotection steps.Clearly, this approach requires the use of at least four solid supports for general DNA synthesis along with four additional supports for RNA synthesis.Various solid supports containing unusual nucleosides for specific applications are also required.
A universal support has the intended 3'-nucleoside added in the first cycle, generating an undesired phosphate linkage between this nucleoside and the universal support.The approach requires that this phosphate linkage be removed during the cleavage and/or deprotection steps.However, the universal support strategy offers the following clear advantages: eliminates the possibility of errors in parallel synthesis applications where up to 384 wells may contain different supports; eliminates the need for at least four supports for DNA synthesis and four for RNA synthesis.
A universal linker phosphoramidite that may be employed for conversion of conventional nucleoside bound solid phase into the universal support directly on the oligonucleotide synthesizer will be of certain value in research laboratories, performing syntheses of modified oligonucleotides, incorporating modified or unusual nucleosides at the 3'-terminus.

Results and Discussion
[3]

Scheme 1
The usii support was thoroughly compared to a number of commercially available Universal solid phases and was demonstrated to be superior in terms of mildest conditions and shortest time of target oligonucleotide cleavage from the solid phase. 4The main advantage of the usii support consists of the fact that processes of cleavage and 3'-dephosphorylation appear to be, in essence the same extremely fast reaction that releases only the desired product -3'dephosphorylated oligonucleotide.
In our previous studies 1,2 we have found that basically three acyl groups, protecting the 2hydroxyl function of 3-acylamido-1,2-propanediol tether may be nearly completely cleaved with ammonia prior to phosphate deprotection (β-elimination of the cyanoethyl group).These groups were: formyl-, dichloroacetyl and 2,4-dichlorophenoxyacetyl, listed in accordance with increasing stability towards ammonolisis.
Yields of oligonucleotides, synthesized on versions of the usii, bearing the formyl (usii f ) and dichloroacetyl (usii dca ) were either equal or even exceeded yields of oligomers, prepared using conventional nucleoside bound solid supports.The usii f version failed to meet the criteria of prolonged storage -this support proved to be not stable enough as a potential commercial product.The usii dca version appeared satisfactory, allowing synthesizing oligonucleotides with yields equal to oligomers generated on nucleoside bound supports and having a reasonable shelf life -1 year at +4 ºC.The versions of the usii, bearing the 2,4-dichlorophenoxyacetyl group (usii dcpa ) proved to be the most stable -more than 1 year at r.t.However, yields of oligonucleotides assembled on the usii dcpa and cleaved as described earlier 1,2,3 were always about 15-25 % lower than yields of the same oligomers, generated on either nucleoside bound supports or the usii dca .
The main disadvantage of the usii is that its preparation takes 5 steps of modification of solid phase -construction of linker and capping steps.Evidently, this lengthy and laborious method does not allow to precisely loading the solid phase with the Universal Linker.Moreover, this fact makes synthesis of the usii support on the industrial scale questionable.
Recently we developed a new simple and very efficient method of carbomoylation of amines. 62][3] Scheme 2 shows structures of the usiii supports, where the urea function bridges the solid phase and the universal linker.As in the case with the usii, we have tested two variants of usiiithe one bearing the dichloroacetyl group (usiii dca ) and another, bearing the 2,4-dichlorophenoxyacetyl group (usiii dcpa ).Both supports proved to be as stable upon storage as their usii counterparts.The yields of oligonucleotides, synthesized on the usiii dca normally exceeded yields of oligomers, prepared using conventional nucleoside bound supports by about 5%.However, yields of oligonucleotides assembled on the usiii dcpa and cleaved as described earlier [1][2][3][4] were again as with usii dcpa about 15-25 % lower than yields of the same oligomers, generated on either nucleoside bound supports or the usiii dca .
We believe that mechanistically the cleavage/dephosphorylation reaction proceeds as shown in Scheme 3. The fastest reaction upon treatment with NH 3 /MeOH cleaves the acyl group of a and gives rise to intermediate b.Amide group of either succinyl linker (usii) or urea linker (usiii), bridging the solid phase and the 3-amino-1,2-propanediol tether assists an instant attack of the 2-hydroxyl group on the phosphorous atom (b) prior to phosphate deprotection (βelimination of cyanoethyl function that gives rise to undesired phosphodiester c).This attack leads to the elimination of the 3'-dephosphorylated target ologonucleotide N-OH along with the formation of the cyclic phosphotriester d.The competing reaction, if any leads to the formation of a solid phase bound oligonucleotide c with the phosphodiester group at the 3'-terminus.This intermediate ultimately converts into a stable phosphodiester e linked to the solid phase.We assume that in the case of usiii dca , bearing dichloroacetyl group a reaction, leading to intermediate b appears much faster than that leading to c. Therefore, practically no stable polymer bound phosphodiester c is being formed in this case and this fact may explain high yields of oligonucleotides N-OH, when dichloroacetyl group is used for protection of the 2hydroxyl function of 3-acylamido-1,2-propanediol tether.

Scheme 3
On the other hand, in the case of usiii dcpa reaction rates, leading to intermediates b and c may appear comparable due to the higher stability of 2,4-dichlorophenoxyacetyl group towards ammonolisis.This phenomena brings about formation of certain fraction of stable polymer bound phosphodiester c and hence the somewhat decreased yields of desired oligonucleotides N-OH (15-25 % lower than yields of the same oligomers, generated on either nucleoside bound supports or the usiii dca ).
Obviously, mechanistic considerations discussed above look reasonable.However, no experimental evidence seems feasible due to the fact that in case of usiii dcpa , the formation of diester c and ultimately polymer bound tethered oligonucleotide e appear difficult to support by conventional physico-chemical methods.
Our new carbomoylation procedure proved applicable for the preparation of various types of urea derivatives. 5Among other compounds we have synthesized urea 1 6 (Scheme 4).This compound bears a free hydroxyl that allows for one step transformation 7 into the universal linker phosphoramidite 2 in about 80% yield (Scheme 4).The structure of phosphoramidite 2 was supported by NMR spectroscopy.The further characterization was achieved by hydrolyzing 2 into H-phosphonate 3 and subsequent oxidation of 3 into a mixture of phosphodiester 4 and phosphomonoester 5 (Scheme 4 and Experimental Section).Compound 2 may be employed to convert any kind of conventional nucleoside bound solid support for oligonucleotide synthesis into the universal phase, employing standard steps on a DNA synthesizer (Scheme 5).

Scheme 5
We performed the following experiment: the protected modified oligonucleotide 5'-ATACCGATTAAGCGAAGTTTXTTA-3'(protected nucleoside units are given in uppercase italics) was assembled on a commercial solid support (Bz-dA-CPG), employing DMTr-off mode.No changes of standard synthesizer protocol were made to introduce the unit X, employing phosphoramidite 2 (Scheme 6).After the assembly, the solid phase bound protected oligomer was treated with 3.5 M ammonia in methanol for 40 min.at r.t.(methanolic ammonia treatment converts unit X into the de-acylated unit Y, Scheme 7).Methanolic solution was removed from the solid phase, diluted (3 times) with concentrated aqueous ammonia and heated at 55 ºC for additional 8 h.The sample was evaporated to dryness and the resulting oligonucleotide material was dissolved in water and analyzed, using IE HPLC (Figure 1).

Scheme 6
The peak with τ R 21.5 min corresponds to the 20mer deprotected oligonucleotide 5'-ATACCGATTAAGCGAAGTTT-3', N-OH (deprotected nucleoside units are given in uppercase plain), as was proved by co-injection with the standard 20mer (data not shown).The structure of oligonucleotide with τ R 22.5 min was tentatively assigned as the 24mer 5'-ATACCGATTAAGCGAAGTTTYTTA-3', f.In order to precisely assign structure of -YTTA-3' tether we have assembled two shorter oligonucleotides -protected 5'-TTTTTTYTTA-3', and 5'-TTTTTTYTTT-3', on the corresponding dA-bound and dT-bound supports and employing the Universal Phosphoramidite 2. After the assembly, the solid phase bound oligomers were treated with 3.5 M ammonia in methanol for 40 min.at r.t.Concentrated aqueous ammonia (2 volumes) was then added to each mixture and the resulting suspensions were left for additional 24 h at r.t.Solid material was removed, both solutions were evaporated to dryness and the resulting oligonucleotide material was dissolved in water and analyzed, using IE HPLC (data not shown; IE HPLC traces were similar to that in Figure 1).Two major eluting peaks were collected in each case and their sequences were established.In both cases, the earlier eluting peak corresponded to the oligo-T 6 as was proved by co-injection with the standard oligo-T 6 (data not shown).The later eluting peaks from both mixtures (5'-TTTTTTYTTA-3', and 5'-TTTTTTYTTT-3') were collected and characterized with ESI MS.Found molecular masses, being 2940.0 (calcd.2938.8) and 2931.0 (calcd.2929.8),univocally supported their sequences to be 5'-TTTTTTYTTA-3' and 5'-TTTTTTYTTT-3' correspondingly.
The syntheses of oligo-T 6 and 20mer N-OH, employing Universal Phosphoramidite 2 appeared to be reasonably good, when the final cleavage/deprotection was performed with 3.5 M ammonia in methanol within maximum 40 min.at r.t and subsequent immediate removal of the methanolic solution from the solid phase.In that case, additional treatment of methanolic solution with concentrated aqueous ammonia (55ºC for additional 8h) facilitated only the cleavage of nucleic bases' protecting groups and the ratio of target 20mer N-OH to the 24mer f varied in several similar experiments from about 10:1 to 12:1.
We believe that the 40 min.treatment of solid support bound protected oligonucleotide with ammonia in methanol at r.t.leads to the release of three products (N-OH, k and f, Scheme 7).The fastest reaction -removal of 2,4-dichloroacetyl group proceeds completely and instantly leads to the release of the desired 20mer oligonucleotide N-OH as a major component in accordance with mechanism, shown in Scheme 7 (formation of triester g leads to N-OH plus cyclic triester j, followed by conversion of j into cyclic diester k) .Two competing slower reactions -β-elimination of cyanoethyl function and ammonolisis of the succinate linker (bridging the controlled pore glass and the nucleoside unit) bring about the release of two minor phosphodiester products -h and I, both subsequently giving rise to tethered oligonucleotide f.However, when the assembled solid support bound protected oligonucleotide was subjected to a treatment with ammonia in methanol followed by aqueous ammonia (24 h at r.t) without the removal of the solid phase, all three competing reactions appear complete.For example, in case with the synthesis of standard 20mer N-OH, this procedure leads to a lower content of desired oligonucleotide along with substantial amount of 24mer f in a ratio of about 7:3.

Conclusions
Phosphoramidite 2 helps to clarify why the yields of oligonucleotides assembled on the usii dcpa or usiii dcpa (bearing 2,4-dichlorophenoxyacetyl group) are giving rise to oligonucleotides with yields about 15-25% lower than yields of the same oligomers, generated on either nucleoside bound supports or universal supports usii dca or usiii dca (bearing dichloroacetyl group).In addition, universal linker phosphoramidite 2 may be successfully employed for conversion of conventional nucleoside bound solid phase into the universal support.
Scheme 7 Waters chromatograph on a xBridge C18, 5 µm (4.6 x 150mm, Waters) column, using a gradient elution of acetonitrile (from70 to 95%) in 0.1M triethylammonium acetate, pH 8.5 over 20 min.; flow rate 1 ml/min (system B).IR spectra were obtained with Thermo Nicolet Nexus 470 FT-IR spectrometer.NMR spectra were taken with Avance 500 spectrometer (Bruker) in CD 3 CN.Elemental analysis was performed using Thermo Quest CE Instruments EA 1110 analyser.Electrospray Ionization Mass spectra (ESI MS) were obtained with Finnigan LTQ MS spectrometer in negative mode.Deconvolution of raw files obtained was performed with BioworksBrowser Rev 3.1 SR1 program.Mongo Oligo Mass Calculator v2.06 and ChemDraw Ultra 8.0 programs were used for calculation of theoretical molecular masses.Oligonucleotide syntheses were performed on an ASM-800 synthesizer (Biosset) on a 0.1 µmolar scale, employing a universal solid support usiii of Metkinen Chemistry (for synthesis of control 20mer oligonucleotide N-OH) or a nucleoside bound supports of Glen Research (for testing of universal phosphoramidite 2) and using standard protocol for DNA synthesis.Monomer concentrations were about 0.1M in all cases.After completion of assembly, oligonucleotides were cleaved from the solid phase with 3.5 N ammonia in methanol (7 N ammonia in methanol of Aldrich was diluted two times with dry cold methanol) and deprotected as described above.