Advanced method for oligonucleotide deprotection

Abstract

A new procedure for rapid deprotection of synthetic oligodeoxynucleotides has been developed. While all known deprotection methods require purification to remove the residual protective groups (e.g. benzamide) and insoluble silicates, the new procedure based on the use of an ammonia-free reagent mixture allows one to avoid the additional purification steps. The method can be applied to deprotect the oligodeoxynucleotides synthesized by using the standard protected nucleoside phosphoramidites dG^iBu, dC^Bz and dA^Bz.

INTRODUCTION

Chemical synthesis of oligonucleotides has been nearly completely automated during the last decade. It involves the sequential assembling of nucleotides on an insoluble solid support, cleavage of the synthesized oligonucleotides from the support, removal of protecting groups, and purification. However, the two final procedures, deprotection and purification, require some manual handling.

Deprotection is usually a relatively long process. Two different strategies have been applied to reduce the deprotection time. One uses labile protecting groups such as phenoxyacetyl (PAC) (1,2), tert-butylphenoxyacetyl (3) or dimethylaminomethylene (4). The other involves more efficient deprotecting reagents such as ammonia-methylamine (AMA) (5) or a mixture of hydrazine, ethanolamine and methanol (6). Although some deprotection systems, like AMA, seem to be quick and cost effective, all existing methods require evaporation, solubilization, centrifugation, precipitation, and/or desalting steps to remove the residual protective groups and insoluble silicates. These steps become a bottleneck in massive oligonucleotide synthesis, especially in construction of large oligonucleotide libraries. For example, the manufacturing of high-density gel microarrays containing thousands of immobilized oligonucleotides requires the low content of non-nucleotide materials in stock solutions (7,8). Such purification of thousands of crude oligonucleotides is time consuming and cannot be completely automated.

We report here a new method for deprotection of synthetic oligodeoxynucleotides that eliminates the post-synthetic operations. The method is based on the use of organic alkaline solution to cleave oligonucleotides from a support and remove internucleotide and base-protecting groups. This procedure can be applied to remove both the standard [N⁶-benzoyl-2′-deoxyadenosine (dA^Bz), N⁴-benzoyl-2′-deoxycytidine (dC^Bz) and N²-isobutiryl-2′-deoxyguanosine (dG^iBu)] and labile protective groups from the synthesized oligonucleotides. The method is especially useful for deprotecting the short oligonucleotides that require no further purification from uncompleted sequences. In this case, one can directly obtain the ready-to-use oligonucleotides dissolved either in pure water or in a buffer. The method proposed is suitable for deprotecting the dimethoxytrityl-containing oligonucleotides and allows the subsequent purification by cartridges or reverse-phase HPLC.

We have found that the treatment of the support-bound protected oligonucleotides with a mixture of triethylamine and lithium hydroxide in methanol for 1 h at 75°C (or for 40 min for the PAC-protected oligodeoxynucleotides) cleaves the oligonucleotides from the support, precipitates them, and completely washes off the released protective groups. We tested this method on the oligonucleotides ranging in length from 6 to 20 bases.

MATERIALS AND METHODS

Lithium hydroxide monohydrate and triethylamine were purchased from Aldrich Chemicals Co., 3′-Amino-Modifier C7 controlled pore glass (CPG), 2 M triethylammonium acetate (TEAA) and all phosphoramidites were from Glen Research, Texas Red“ sulfonyl chloride dye was from Molecular Probes and CPG-10-700 Å supports (120–200 mesh) were from Fluka Chemie AG. Mass spectrum analysis was carried out with a matrix assistant laser desorbtion ionization time of flight (MALDI-TOF) mass spectrometer COMPACT MALDI 4 (Kratos analytical). IR spectra were registered in a KBr pellet with a Bruker 500 spectrophotometer.

Oligonucleotide synthesis

Oligonucleotides were synthesized either on columns on a 0.2–1 µmol scale with an ABI 394 DNA/RNA synthesizer (Applied Biosystems) or in 96-well plates with a liquid chemical dispensing robot (LCDR) (Avantech Automation) by using the CPG-based solid supports according to the standard phosphoramidite procedure.

Oligonucleotide microarray manufacturing and hybridization

Microchips were manufactured as described earlier (7–10). Microarrays consisting of 100 × 100 × 20 µm gel pads were prepared by photopolymerization of 5% acrylamide–bisacrylamide as previously described (9). Each oligonucleotide (0.2 pmol) was transferred by a robot onto the gel pad with a pin (7) and immobilized by means of Schiff reaction between the 3′-amino group of the oligonucleotide and the aldehyde group of the activated gel pad (10). Hybridization of the microarray with complementary fluorescently labeled oligodeoxynucleotides was carried out in 6× SSPE containing 1% Tween 20 at 20°C for 12 h. Hybridization patterns were obtained with a fluorescent microscope equipped with a charge couple device camera, Peltier thermotable and a computer interface (7).

Deprotection of oligonucleotides

The oligonucleotides synthesized on columns were treated according to the new deprotection protocol as follows. The oligonucleotides bound to the CPG support were dried in argon flow. The CPG support was transferred into a 1.5 ml centrifuge tube and treated with a mixture of 0.5 M aqueous lithium hydroxide (30 µl) and 3.5 M triethylamine in methanol (300 µl) at 75°C for 60 min. Then the reaction mixture was chilled to –20°C for 5 min and 75 µl of glacial acetic acid was added. After that the resulting solution of triethylammonium acetate was carefully removed from the glass support. The support was successively washed twice with 90% aqueous acetonitrile (400 µl) and twice with acetonitrile (400 µl). Finally, either pure water or a buffer (200 µl) was added twice to elute the oligonucleotides. Alternatively, after deprotection with LiOH–Et₃N solution, the reaction mixture was cooled to 0°C for 5 min and the deprotection solution was carefully removed from the glass support. The support was successively washed with 90% aqueous acetonitrile (400 µl) three times, with 0.1 M triethylamine formate in 90% acetonitrile (400 µl) four times, and with acetonitrile (400 µl) three times. Pure oligodeoxynucleotides were eluted from the CPG support either by water or a buffer.

In the case of the LCDR-synthesized oligonucleotides treated according to the new deprotection protocol, each well of the 96-well filter plate was filled with 300 µl of deprotection solution and the plate was incubated at 75°C for 60 min. The plate was then cooled to room temperature and each well was successively washed with 90% aqueous acetonitrile (400 µl) three times, with 0.1 M triethylamine formate in 90% acetonitrile (400 µl) four times, and with acetonitrile (400 µl) three times, allowing the solution to pass through the wells freely each time. In all cases, the pH of the oligonucleotide solution was neutral allowing the oligonucleotides to be directly immobilized on microchips (7). The deprotected oligonucleotides were analyzed by both ion-exchange and reverse-phase HPLC or by PAGE under denaturing conditions.

To compare the new method and the standard one (32% aqueous ammonia, 55°C, 8 h), two identical oligonucleotide sets were created either by parallel synthesis or by splitting the support-bound protected oligonucleotides into two equal parts after their synthesis.

RESULTS AND DISCUSSION

The new deprotection procedure described here was originally developed to create the oligonucleotide libraries used in the manufacturing of oligonucleotide microarrays (7,11), such as the generic microchip containing a complete set of hexadeoxynucleotides (8,12).

We applied the new protocol to deprotect a set of oligodeoxynucleotides ranging in length from 6 to 20 bases, which were automatically synthesized using the standard phosphoramidite procedure.

The similar total amounts of 10mers (16–20 OD) were obtained by 0.2 µmol synthesis with the use of both deprotection procedures. Both protocols were generally equivalent in terms of the deprotection efficiency and quality/quantity of oligonucleotide material, as revealed by the comparison of the analytical reverse-phase HPLC profiles (Fig. 1A) and the ion-exchange HPLC profiles (Fig. 1B and C) of the oligonucleotides synthesized using an ABI 394 DNA synthesizer or LCDR. However, the traditional procedure does not allow one to remove the residual protective groups and soluble silicates, which requires some additional purification steps. Such contaminating material is usually abundant and is easily detected either visually or by HPLC analysis.

Figure 1.

Comparison of oligonucleotides deprotected by the standard and the new procedures. (A) Coinjection profile of the oligonucleotide 5′-d(AC)₅dA, synthesized with an ABI 394 machine and deprotected by the standard and new protocols. Analytical reverse-phase HPLC analysis was carried out using a Hypersil ODS column (5 µm, 4.6 × 250 mm) at 40°C. The mobile phases used were as follows: buffer A, 0.1 M TEAA (pH 7); buffer B, 0.1 M TEAA in 50% aqueous acetonitrile (pH 7). The gradient system used was 100% buffer A for 2 min, then 0–50% buffer B for 36 min. The flow rate was 1 ml/min. (B) Ion-exchange profile of the oligonucleotide 5′-d(TTACGCA)-NH₂ synthesized with the LCDR machine and deprotected by the standard protocol. NH₂ represents C7 3′-aminomodifier from Glen Research. (C) Ion-exchange profile of the same oligonucleotide synthesized with the LCDR machine and deprotected by the new method. Ion-exchange HPLC analysis was carried out using a Nucleosil NH₂ column (5 µm, 4.6 × 250 mm) at 40°C. The mobile phases used were as follows: buffer A, 0.1 M TEAA (pH 7); buffer B, 0.8 M TEAA (pH 7). The gradient system used was 0–50% buffer B for 30 min. The flow rate was 1 ml/min.

The new method is based on the fact that the synthesized oligonucleotides are insoluble in organic media and precipitate on the CPG support, while the contaminating material remains in solution and can be easily washed off (demonstrated in Fig. 2). The standard (Fig. 2A) and new (Fig. 2B) deprotection schemes were applied to the DMT-protected oligomer d(AC)₂A. In the case of the new protocol, no peak corresponding to benzamide was revealed, while the oligonucleotides deprotected by the standard method were contaminated with benzamide. However, in the presence of lithium hydroxide, the deprotection of the benzoyl-protected oligonucleotides yielded lithium benzoate. Nevertheless, we found that the solubility of lithium benzoate, isobutirate and acetate was sufficient for these substances to be completely washed off by 90% aqueous acetonitrile during the washing procedure.

Figure 2.

HPLC profiles of the DMT-protected oligonucleotide treated by two different deprotection protocols. (A) DMT-d(AC)₅dA treated according to the standard protocol. (B) The same oligonucleotide deprotected according to the new protocol. The HPLC conditions were the same as in Figure 1A, except the gradient used (20% buffer B for 2 min and 20–100% buffer B for 36 min).

One of the essential steps of the new deprotection scheme is the neutralizing of lithium silicate that is chemically or physically bound to the glass surface. The oligonucleotide solution eluted from the CPG support, which was not neutralized, was characterized by the high pH value and contained dissolved silicates. Later, upon neutralization, large amounts of the silicates precipitated from the oligonucleotide solution became insoluble. We tested several reagents in terms of their applicability for the neutralization of alkaline components in reaction mixture, and only a few of them turned out to be suitable. We found that 0.1 M solutions of triethylamine formate or acetate in 90% acetonitrile were quite efficient as neutralizing agents, but such treatment included numerous washing steps to obtain the final neutralized oligonucleotide solution. The problem has been overcome by adding acetic acid to the cold reaction mixture. Due to the low temperature, heterogeneous conditions of neutralization and reagent stoichiometry, there was no detrytilation observed for the DMT-protected oligonucleotides as evidenced by reverse-phase HPLC.

Despite the fact that the insoluble silicates were undetectable visually when the advanced protocol was applied, a trace amount of dissolved SiO₂ might still be present in oligonucleotide eluates. To compare two deprotection methods, the standard and the new one, in terms of the silicate content, we quantitatively analyzed by IR spectroscopy the dry residues obtained after the evaporation of either ammonia solution or the neutral eluate from the underivatized CPG. It is known that the prominent band in the IR spectra in the range 1050–1100 cm^–1 corresponds to the Si–O–Si stretch vibrations (13,14). Taking into account that the amount of the CPG support was the same with both procedures, we can directly compare the band intensities from two samples to estimate quantitatively the silicate contamination. Figure 3 shows the IR spectra of the samples treated by different deprotection methods. In the case of the new protocol (Fig. 3, spectrum B), the intensity of the band corresponding to the Si–O–Si stretch vibrations was considerably less than that observed for the sample obtained by the standard protocol (Fig. 3, spectrum A).

Figure 3.

IR spectra of solid residues obtained by the standard (A) and advanced (B) protocols.

The identity of the oligodeoxynucleotides obtained by the new and standard deprotection procedures was additionally proved by the MALDI-TOF analysis. We recorded the MALDI-TOF mass spectra of the oligodeoxynucleotides synthesized in parallel and deprotected by two different methods to confirm their size and the absence of modified products. This follows from Figure 4 displaying two spectra of the oligonucleotide d(AGC)₂AC deprotected by the standard (Fig. 4A) and advanced (Fig. 4B) protocols. The spectra obtained for both products testified to the identity of their molecular weights.

Figure 4.

Mass spectrum profiles of the oligonucleotide d(AGC)₂AC obtained by the standard (A) and advanced (B) protocols.

In order to confirm the identity by performance of the oligonucleotides treated by two different deprotection protocols, they were hybridized with complementary fluorescently-labeled oligonucleotide in microarray format (15). The samples with the same concentration of the oligodeoxynucleotide 5′-d(TTACGCA)-NH₂ obtained after deprotection by two different procedures were loaded on gel pads of a small microarray. The immobilization was performed as described earlier (9). The hybridization pattern is demonstrated in Figure 5. Both parts of the microchip (Fig. 5A and B) corresponding to the different deprotection schemes displayed the same signals of hybridization, which proves the full functionality of the oligonucleotides treated by the advanced protocol.

Figure 5.

Hybridization of the oligonucleotide 5′-d(TTACGCA)-NH₂ obtained by the standard (A) and new (B) protocols with the complementary oligonucleotide 5′-d(TGCGTAA)-NH-TexRed.

To summarize, the described deprotection scheme eliminates the conventional evaporation, centrifugation and desalting steps and provides purified oligonucleotides in a relatively short time. It is based on the use of inexpensive nucleoside phosphoramidites containing the standard base protection groups and seems to be promising in the large-scale synthesis of short oligomers, especially with the LCDR machine in the 96-well format. The new deprotection method considerably facilitates handling of both in-column and in-plate synthesized oligonucleotides, which mainly benefits in the case of parallel synthesis with the LCDR machine. It provides the possibility of directly transferring the synthesized oligonucleotides from the 96-well plate to the stock plate avoiding the collection of ammonia or AMA solutions, their evaporation and other transfer and purification steps.