A Simple and Efficient Microfluidic System for Reverse Chemical Synthesis (5′-3′) of a Short-Chain Oligonucleotide Without Inert Atmosphere

Abstract

Reverse DNA synthesis (5′-3′) plays diverse functional roles in cellular biology, biotechnology, and nanotechnology. However, current microfluidic systems for synthesizing single-stranded DNAs at a laboratory scale are limited. In this work, we develop a simple and efficient polydimethylsiloxane- (PDMS-) based microfluidic system for the reverse chemical synthesis of short-chain oligonucleotides (in the 5′-3′ direction) under ambient conditions. The use of a microfluidics device and anhydrous conditions effectively surpass the problem of moisture sensitivity during oligonucleotide synthesis. With optimized microfluidic synthesis conditions, the system is able to synthesize up to 21 bases-long oligonucleotides in air atmosphere. The as-synthesized oligonucleotides, without further purification, are characterized using matrix-assisted laser desorption ionization–time of flight (MALDI-TOF/TOF) mass spectroscopy (MS) supported by the denatured polyacrylamide gel electrophoresis (PAGE) analysis. This developed system is highly promising for producing the desired sequence at the nanomolar scale on-chip and on-demand in the near future.

1. Introduction

The past two decades have witnessed significant advancement in research efforts to incorporate synthetic short-chain ssDNAs into a variety of applications and miniaturized devices, such as oligonucleotide-Chip technology [1,2,3,4,5]. In the miniaturization of devices made of nucleic acids, single-stranded DNAs are the preferred choice, which are largely synthesized autonomously or non-autonomously via a staged process that uses conventional phosphoramidite chemistry [6]. The rapid growth in the use of synthetic oligonucleotides made chemical synthesis (3′-5′ direction) the standard protocol for preparing oligos using a commercial DNA synthesizer. This occurred because of the easy accessibility of the corresponding protected monomeric building blocks [7]. On the other hand, reverse chemical synthesis (5′-3′ direction) has been limited to nuclease resistance and hairpin loop applications, where the terminal linkage of natural 5′-3′ antisense oligonucleotides was modified via the construction of alternating 3′-3′ or 5′-5′ internucleotide linkages (terminal capping), and hence has not been matured much yet [8,9,10,11]. However, addressed micro-arrays of surface-bound ssDNAs (5′-3′) with a free 3′ terminal group recently found multiple biotechnology applications in genetics, for example, in enzymatic extension via DNA or RNA polymerase at the 3′ hydroxyl group, and transcriptomics/genomic analysis of cells using 3′-poly T tail-based primers (5′-3′) containing short-chain oligos as a barcode address-tag (<12 bases) [12,13]. In later applications, researchers have been preparing random sequences for barcode-tags using mix and pool synthesis, which limits the application. Direct reverse synthesis of a specific sequence as an address-tag is attractive to overcome this limitation. Our work presents an application of reverse synthesis which centers upon on-chip and on-demand synthesis of a barcode as an address-tag which can be used for applications like positional information-based transcriptomics and proteomics analysis of single cells.

To develop a system for the reverse synthesis of oligonucleotides, microfluidic devices were the favored choice because of their advantages like precise and reproducibly actuate fluids, low reagent consumption, and plentiful quantity production [14]. Most of the microfluidic systems have been limited to continuous flow reactors fabricated in glass or silicon because a variety of organic reagents are needed for chemical syntheses, which limits the workable complexity and reagent savings [15]. In some cases where solvents are mild, it has been possible to use integrated elastomeric material to perform batch syntheses at the nanogram scale [16,17]. Hua et al. developed a chemically resistant microreactor made of a hybrid silicon–elastomer microfluidic system, but with limited chemical characterizations of the products [18], whereas Quake et al. reported a photocurable perfluoropolyether elastomer for oligonucleotide synthesis (5′-3′) [19,20]. In addition, Quake et al. succeeded in developing mild reagents, which made DNA synthesis feasible on a polydimethylsiloxane- (PDMS-) based microfluidic device [21]. Despite significant advancements, the necessity of an inert atmosphere, skilled operators, and expensive controller units to perform the synthesis is still a significant challenge. Therefore, the development of an efficient system with a simple design and ample product yield under ambient conditions is necessary; this will not only reduce the cost of the process, but also facilitate on-chip and on-demand processing at a laboratory scale.

In this paper, we present a simple and low-cost PDMS microfluidic system that enables the synthesis of reverse oligonucleotides (5′-3′) under ambient conditions suitable for on-chip and on-demand applications. The single-layered microchannel fabrication makes the microfluidic system simple and easy to use at common laboratory scales. We have also optimized the microfluidic conditions, like flow rates and flow times of the synthesis step, to achieve a high yield of the desired sequence.

2. Materials and Methods

2.1. Materials

The DNA phosphoramidite monomers and controlled pore glass (CPG) (nominal particle size 110 μm, average pore size = 1000 Å) were purchased from Link Technologies Ltd. (Scotland, UK). Superdehydrated acetonitrile (water < 0.001%, 10 ppm) was purchased from Wako Pure Chemical Industries Ltd. (Osaka, Japan). A 25–30% NH₄OH aqueous solution was purchased from Fluka Analytical (Munich, Germany). The (1S)-(+)-(10-camphorsulfonyl) oxaziridine (CSO) oxidizer, 5-(ethylthio)-1H-tetrazole (ETT activator, 0.25 mol/L solution in anhydrous acetonitrile), 5-(benzylthio)-1H-tetrazole (BTT activator, 0.3 mol/L solution in anhydrous acetonitrile), and anhydrous acetonitrile (MeCN) were purchased from Glen Research (Sterling, VA, USA). The SYBR gold nucleic acid gel stain was purchased from Invitrogen (Carlsbad, CA, USA). Tris-EDTA buffer (pH = 7.5) was purchased from Integrated DNA Technologies (IDT; Coralville, IA, USA). All the DNA oligonucleotides, except the ones that we synthesized from the microfluidic chips, were ordered from IDT and were purified using high-performance liquid chromatography (HPLC). Other chemicals, including tris (hydroxymethyl) aminomethane (Tris), Tris-borate-EDTA (TBE) buffer 10x, poly (vinylpyrrolidone), and TFA, were purchased from Sigma-Aldrich (Tokyo, Japan).

2.2. Fabrication of the PDMS Microfluidic Chip

The chip fabrication steps were conducted in a clean room facility at JAIST (see supplementary information, Figure S1) [22]. The PDMS chip with a microchannel was developed from a defined mold structure, made of SU-8 over silicon wafers using photolithography techniques. PTFE tubes (ID 1mm, New England Small Tube Co., Litchfield, NH, USA) were inserted into the inlet and outlet holes of the chip, and sealed the gap with PDMS. Another end of the inlet PTFE tube was bonded with a needle to introduce reagents from rubber capped glass vials (see supplementary information, Figure S2a). The PDMS chip consisted of a long chamber (W = 2 mm, L = 10 mm, and H = 220 µm), micropillars (W = 120 µm, L = 500 µm, and H = 220 µm), a zigzag channel (W = 400 µm, L = 60 mm, and H = 220 µm) near the outlet, and a through-hole (2 mm) close to inlet (Figure 1). These micropillars are designed (the gap between two pillars was set to be 100 µm) to trap the CPG particles of a 110 μm size. For CPG loading, 20 µL of freshly prepared CPGs slurry (60 mg, loading capacity 25–40 µmol/g) in 1 mL anhydrous acetonitrile was introduced slowly from the through-hole using a micropipette, and once the micro-chamber was filled, CPGs were washed with fresh anhydrous acetonitrile. A column of porous CPG particles was therefore packed inside the reaction chamber in a defined manner of particles monolayer (see supplementary information, Figure S2b). The PDMS chip filled with CPGs was then placed at a high temperature, 120 °C, overnight to ensure that particles were completely dry prior to experiments.

2.3. Operation of the Microfluidic Chip

A micro syringe pump ESP-64 from EiCOM Corp. (Kyoto, Japan) in reverse flow mode was connected to the outlet, and controlled the flow of reagents within the microfluidic channel. At the inlet, reagent vials were exchanged manually in accordance with the synthesis steps. The flow rate and flow-time of each synthesis step were optimized for the PDMS microfluidic system according to the desired length of the oligonucleotides. It was easy to change these variables precisely with the micro syringe pump, prior to the next synthesis step. After completing the synthesis, a forward flow with fresh anhydrous acetonitrile was applied using a syringe pump, to recollect the CPG particles from the inlet in a small Eppendorf tube (200 µL). We did not notice any damage to the channel of the PDMS chip due to the exposure of synthesis reagents, even after repeating multiple steps of the oligonucleotide synthesis. Therefore, the chips were operated on a multiple-use basis. In the case of oligonucleotide synthesis in inert atmosphere, the Glove box from Miwa mfg. Co. Ltd. (Osaka, Japan) filled with Argon (Ar) atmosphere and O₂ (<1 ppm level) was used. A micro syringe pump, PDMS chip, and all the synthesis reagents were shifted inside the glove box one day before the experiments.

2.4. On-Chip Chemical Syntheses of Reverse Oligonucleotides (5′-3′)

Solid-phase organic synthesis on a microchip has the advantages of a high reaction efficiency, low reagent consumption, and high product purity [14,22]. Phosphoramidite chemistry was adopted for on-chip reverse chemical synthesis of oligos [23,24]. Traditionally, oligonucleotide elongation is induced by repeating the four steps (deprotection, coupling, capping, oxidation, and four in-between washing steps) in a single cycle. A detailed scheme of the solid-phase reverse oligonucleotide synthesis on CPG as a solid support is shown in Figure 2. Before synthesis, glass vials (10 mL) were filled with 2 g, and molecular sieves were dried at 180 °C in a vacuum oven overnight to remove the absorbed moisture and other materials, if any, and were allowed to cool down to room temperature 15 min prior to use. All the used solutions were kept under a molecular sieve (3 Å) for 24 h to make the solution completely moisture-free. Phosphoramidite compounds were dissolved in superdehydrated MeCN to form 0.1 mol/L solutions, and the CSO was dissolved in anhydrous MeCN (0.1 g/mL) and filtered through a 0.45 µm filter. The porous CPGs were modified with the first nucleoside attached.

First, the CPG particles were rinsed with anhydrous acetonitrile at a flow rate of 10 µL min⁻¹ for 5 min to ensure continuous flow of the solvent through the chamber. The flow rate was optimized and maintained at 10 µL min⁻¹ for all solutions. The first 3′-DMTr-protected nucleoside attached to the CPG particles was deprotected by introducing a mild deblocking reagent (5% trifluoroacetic acid (TFA) in anhydrous acetonitrile) at a flow rate of 10 µL min⁻¹ for 2–4 min (the 5% TFA in acetonitrile shows high solvent compatibility with PDMS compared to conventional deblocking solvent dichloromethane (DCM), which swells the PDMS and destroys the microstructure by clogging the channel). The spectrophotometric assay of the resultant solution (containing trityl cations) was collected and measured to monitor the efficiency of the synthesis step. Stepwise coupling yields were estimated on the basis of the characteristic trityl absorbance assay in CH₂Cl₂ (ε = 76000 LM⁻¹cm⁻¹) (see supplementary information, 9) [25]. The residual deblocking reagent in the chamber was washed out through the outlet via a 2–3 min superdehydrated acetonitrile washing step performed at the same flow rate. Since the nitrogenous bases of the growing DNA chain are susceptible to acid-catalyzed depurination, the deblocking step needs optimization, and an acetonitrile rinse thoroughly removes the deblocking agent from the support. Also, coupling efficiency and accuracy are increased by this wash, since premature detritylation of the incoming phosphoramidite monomer is prevented. To add the incoming monomer to the freed 3′-OH, the next phosphoramidite was first activated with the ETT activator for 30 s and then introduced into the chip from the inlet for 4 min. After the completion of the coupling reaction, the washing step was repeated for 2 min. Even with a high coupling efficiency, a small percentage of the 3′-OH group remained unreacted, which in subsequent coupling, led to deletion sequences. These active 3′-OH groups were inactivated with capping steps using cap mix A and cap mix B for 1 min, even though, in our study, we did not find any noticeable difference in the yield of the desired sequence with or without the capping step (confirmed with PAGE analysis, results are not shown here). A similar finding has been mentioned in a previous study [19]. Therefore, we eliminated the capping step from our synthesis cycle. The resultant dimer from the first coupling produced an unstable trivalent phosphite triester, which was further oxidized to a stable pentavalent phosphotriester by introducing a 0.1 M CSO reagent into the anhydrous acetonitrile for 2 min. Then, another deprotection procedure was performed to initiate a new cycle. When the desired oligonucleotide elongation was achieved, a final 3′-DMTr deprotection step was performed with a 5% TFA solution into the chip, followed by thorough washing with the superdehydrated acetonitrile solvent for 5 min. Detailed microfluidic conditions with optimized parameters for on-chip oligo synthesis are described in Table S1 (see supplementary information, Table S1).

2.5. Characterization Using Mass Spectroscopy (MS)

After synthesis, the CPG particles were collected from the inlet by reversing the flow with fresh anhydrous acetonitrile in an Eppendorf tube. The solvent was evaporated using CentriVap. Then, the oligonucleotides were cleaved from the solid support (CPG particles) and deprotected. This is a two-step process that starts with cleavage and deprotection. However, it is not uncommon to perform it in one step. We added 200 µL of 25–30% ammonium hydroxide into an Eppendorf tube containing CPGs and then incubated the tube at 55 °C for 4 h to remove the base-protection groups from the oligonucleotides. Finally, the tube was lyophilized to yield solid-form oligonucleotides (with salts). In some cases, long incubation (>6 h) led to an insoluble white powder, which resulted from the CPG particles being dissolved with concentrated ammonium hydroxide.

The synthesized oligonucleotides, without any further HPLC purification or desalting, were characterized using matrix-assisted laser desorption ionization–time of flight (MALDI-TOF/TOF) spectroscopy. 3-Hydroxypicolinic acid (3-HPA) was found to be the most suitable matrix for the soft ionization of the as-synthesized oligonucleotides. 10 µl of sample: matrix (1:1) solution was dropped over the MALDI substrate and dried in air atmosphere at room temperature before the characterization.

2.6. Electrophoresis

The synthesized oligonucleotides were resuspended in Milli-Q water and normalized to a stock concentration of 100 pmol/µL using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Japan). All gels were cast prior to the experiments in 5x concentrations of TBE buffer by precisely diluting the stock buffer in 56x TBE (450 mM Tris base, 450 mM boric acid, and 10 mM EDTA, pH = 8.9) with the required quantity of deionized water. All electrophoreses were carried out with a 16% w/v polyacrylamide gel composed of Acrylamide/Bis 19:1 in the TBE buffer (pH = 8.3) at 25 °C and 25 V/cm. Before gel loading, the oligonucleotides were mixed with the TBE-urea sample loading buffer (Invitrogen) and heated to 90 °C for 2 min. Samples of 2 µL were loaded. Samples with different lengths were run at room temperature at 175 V for different time intervals. The gels were stained with SYBR gold dye (1x) and visualized using a Fluoroimager, Quantity One (Bio-Rad, Hercules, CA, USA), with 550 nm for the excitation and 570 nm for the emission.

3. Results and Discussions

On-Chip Reverse Synthesis of Oligonucleotides

Due to the fact that the coupling reaction is highly moisture-sensitive, the reverse oligonucleotide synthesis is challenging under ambient conditions [26]. To overcome the problem associated with the moisture sensitivity, we are taking advantage of a closed microfluidic system by minimizing the exposure to the atmosphere to some extent.

In standard phosphoramidite chemistry, the primary 5′-OH group of the phosphoramidite monomer is significantly more reactive than the secondary 3′-OH group (or 2′-OH group). Contrary to standard synthesis (3′-5′ direction), in (5′-3′) reverse synthesis, the 3′-OH group of the phosphoramidite monomer (dC^Bz, Figure 2) is protected with DMT, which makes the 5′-OH group of the monomer (dC^Bz, Figure 2) available to form a dimer with the less reactive secondary 3′-OH group of dG^iBu-CPG in order to grow the sequence in the 3′-5′ direction. Due to this chemistry, the coupling in reverse synthesis takes place at a slower rate than standard chemistry [8]. Therefore, to achieve a higher coupling efficiency for the reverse synthesis, optimization of the coupling and deprotection step is required in the first part. To optimize the conditions for a high coupling efficiency, oligonucleotides were synthesized by varying the synthesis conditions, such as the coupling time, deprotection time, and washing time (especially washing after the deprotection step as coupling efficiency and accuracy are increased by this wash, since premature detritylation of the incoming phosphoramidite monomer is prevented) (see Supplementary Information Figure S3). Figure S3d shows that the single intense peak at 2986.89 (m/z) corresponds to the 10-mer oligonucleotide, which was obtained as a major product with negligible small peaks of n-mer oligonucleotides (n < 10, truncated products), when the synthesis was performed with 4 min of coupling and 3 min washing. While the oligo synthesis was performed at a lower coupling time (see Supplementary Information, Figure S3a–c), 10-mer oligonucleotides were obtained as minor or negligible products, whereas truncated products were obtained as major products. The MS spectra obtained from the above optimization with coupling and washing time variation between 2 min and 4 min, depict that reaction propagation takes place in a stepwise manner. The longer coupling time enhances the chances of efficient coupling of the active site of the oligonucleotide chain with the incoming phosphoramidite more effectively, and therefore results in a high yield of the desired sequence (oligos yield calculation was made by trityl cation monitoring during the synthesis) with a single intense peak. The improved coupling efficiency and yield are attributed to the optimized flow conditions of the enhanced coupling and thorough washing. During the synthesis of short chain oligos, the capping step was omitted from the synthesis scheme; therefore, the effect of the capping step was also confirmed by performing synthesis of the oligonucleotide with and without the capping step (see supplementary information, Figure S4). It is clear from the results that the capping step is not as significant in the case of short chain oligos synthesis.

In the next step, a comparative study of the oligonucleotide synthesis with and without inert atmosphere was performed. Figure 3 shows the MALDI-TOF/TOF MS spectra (Bruker Daltonics UltrafleXtreme, Bremen, Germany) of 10-mer oligonucleotides (TTCGAGACTG, m/z = 3041) in an inert (N₂) atmosphere, and (TATGTACGAC, m/z = 3026) in an air atmosphere. Specific synthesis conditions are mentioned in the graph. The MS data of oligonucleotides synthesized in the inert atmosphere [Figure 3a] and in the air atmosphere [Figure 3b] show the single intense peak of the desired product at m/z = 3043.8 and m/z = 3026.2, respectively, corresponding to 10-mer oligonucleotides. Some other peaks are also observed in Figure 3a,b with extremely low intensities and can be assigned to truncated products (products with one or more missing nucleotides from the desired sequence) with a negligibly low amount compare to the target sequence.

The MS results (Figure 3) confirm the successful synthesis of 10-mer oligonucleotides as major products in both cases. The overall yield of the 10-mer oligonucleotides was calculated to be 95% in the N₂ atmosphere and >91% in the air atmosphere from the stepwise yield using trityl cation absorbance. From the MS spectra and calculated yield, it is explicit that reverse syntheses of oligonucleotides in inert atmosphere and in air atmosphere obtained comparable results. This achievement was made possible by applying anhydrous conditions to all the freshly prepared solutions using molecular sieves, and optimized synthesis conditions. Figure S5 (see supplementary information) confirms this explanation as we synthesized the same length of oligonucleotides with anhydrous reagents without molecular sieve treatment. Before the experiment, reagents were kept in closed vials for at least 24 h, and identical synthesis conditions were applied. The result clearly shows that the major products obtained from this synthesis were truncated products with an unacceptable low amount of the desired sequence. The above experimental results clearly show that this microfluidic system and anhydrous conditions allow the reverse synthesis of oligonucleotides without inert atmosphere by providing sufficient protection from atmospheric moisture.

The on-chip reverse synthesis conditions for oligonucleotides (>10 bases) were optimized further with an increased flow time of the deprotection step. Figure 4 shows the MALDI-TOF/TOF MS spectra of the 12-mer oligonucleotides synthesized in an air atmosphere before and after changing the flow time of the deprotection step. Detailed microfluidic syntheses conditions are summarized in Table S1, unless otherwise specified (see Supplementary Information Table S1). Figure 4a shows the MS spectra of the 12-mer oligonucleotide (CTGGCTTTAGTA, m/z = 3649) with a deprotection time of 2 min, and the peak of the desired sequence corresponds to 3651.62 (m/z), which was obtained with a <70% yield (calculated). It also shows the peaks of truncated sequences (n-1 mer, n = 12) with a peak intensity comparable to the demanded sequence. It is clear from the MS spectra that optimized conditions are not acceptable for synthesizing 12-mer oligos with a high quality. To improve the quality of oligo sequences (>10-mer), the time of the deprotection step was varied from 2 min to 2.5 min, and 3 min, which resulted in an improved quality, as shown in Figure 4b (data with deprotection time of 2.5 min not shown here). Figure 4b shows that the MS-spectra of the synthesized 12-mer oligonucleotide (CCAGTCGACCGA, m/z = 3613) with a deprotection time of 3 min corresponds to the intense peak at 3615.13 (m/z). With increased deprotection time, we become successful at achieving an improved quality of 12-mer oligonucleotides with a 85% yield (synthesized in an air atmosphere). The MS spectra (Figure 4b) also demonstrate peaks at 2972.96 and 1752.64 (m/z), which correspond to (n-2) and (n-6) truncated products, respectively, but these sub-sequences with a minor amount can be ignored compared to the amount of desired sequence. From the MS spectra (Figure 4), it is clear that oligo synthesis (<10 bases) with 2 min is insufficient, which leads to incomplete deprotection. It interferes in the next synthesis cycle, which results in the production of truncated sequences with remaining protected sites from the previous cycle. The improved result of the 12-mer oligonucleotides using PDMS chips in an air atmosphere confirms that the extended exposure of the deprotection reagents along with optimized coupling and washing times leads to a pure product with a high yield.

In this work, we extended the oligonucleotide chain up to 18 mer and 21 mer using optimized conditions for the on-chip reverse synthesis. Figure 5a,b show the MALDI-TOF/TOF MS spectra of the 18-mer (CAGTCTGAGTCAGCTGAT, m/z = 5511) and 21-mer (GCGGCTGAAGACGGCCTATGT, m/z = 6480) oligo sequences, respectively. Both spectra show the intense peak of the corresponding sequences with a maximum yield of 75% in each case (average yield of >98%). The MS spectra (Figure 5) also demonstrate multiple peaks of sequences other than the target sequence. Some of these correspond to the truncated products of the respective target sequence, but there are still plenty number of peaks which correspond to unknown sequences. We are not able to analyze these peaks through m/z with respect to the synthesized sequence. The exact reasons for these peaks are not clear yet, but it is expected that the peaks are obtained from the sequences with the deletion of nucleotides during the synthesis. The most probable reason for those unexpected peaks could be because of omitting the capping steps from the synthesis cycle. Although we proved in Figure S5 (see supplementary information) that the capping step is not significant, long-chain oligo synthesis (>12 bases) is expected to have a large impact from the omitting of the capping step as capping blocks the remaining active site from the previous cycle. Another reason could be the effect of long exposure of the deprotection reagent on the growing oligo chain, which results in the depurination of bases and thus creates another site for a side reaction. Therefore, a study on the depurination of bases with an optimized deprotection time is required in the future to understand the origin of these peaks. Finally, the random fragmentation of the oligo sequence during ionization, most probably related to the high laser intensity, could also not be neglected, which led to multiple peaks at random m/z that are difficult to analyze (see supplementary information also, Figure S6).

For analyzing the quality of synthesized oligos (21-mer), we also performed a denatured polyacrylamide gel electrophoresis (PAGE) analysis with a known marker to confirm the length of the synthesized product. Three different oligonucleotide sequences with the same length of 21 bases were synthesized separately using the same conditions and checked using gel electrophoresis. The oligonucleotides were dissolved in Milli-Q water, and sample solutions of 10 pmol/µL were prepared using the NanoDrop system for the PAGE analysis. Figure 5b (inset) shows a denatured gel image with all four samples of 21 mer (lanes 2–5) and a 5 bp ladder (lane 1) for a standard marker. The gel image shows that the dominant sharp bands in sample lanes 2–5 have a slightly lower mobility than the band of 20 mer from the ladder, which confirms the presence of synthesized 21-mer oligonucleotides. The band intensities of the 21-mer product in all the lanes are higher, which is directly related to the amount of 21-mer sequence in the synthesized product, including the truncated products. The PAGE image also reveals multiple bands of lower intensities compared to the target band, which directly correspond to the truncated products from the same synthesis with a negligible amount. However, the sum of these negligible amounts of truncated sequences is highly comparable to the amount of target sequence. The gel electrophoresis results of 21-mer oligos support the MALDI-TOF/TOF MS analysis and it is clear from both analyses that 21-mer oligos were synthesized with a large number of truncated products. It is also clear that the it is difficult to provide an accurate yield and quality of the synthesized target sequence, even though it was calculated from the trityl cation absorbance during each cycle. Without capping step inclusion, it will not be wise to predict the yield accurately only on the basis of trityl cation absorbance. Therefore, it is highly recommended that the synthesis conditions are optimized further for synthesizing the long chain oligos (>15 bases) with high purity. From the oligo synthesis (<15 bases) results, it is explicit that such negligibly low amounts of truncated products are expected to not interfere with the application of the synthesized oligos. We found that, at this length, the resulting oligonucleotides were of a sufficiently high quality, but we still required further analysis and optimization in the case of long chain oligo synthesis on-chip. Fortunately, however, many biochemical applications do not have stringent purity requirements, and if the coupling efficiency is high enough, a mixture of produced products can often be used with either minimal (desalting) or no purification [27].

4. Conclusions

A simple PDMS microfluidic device is demonstrated for on-chip reverse chemical syntheses of oligonucleotides (5′-3′) without inert atmosphere. In this work, we have synthesized oligonucleotides up to 21 bases with optimized microfluidic synthesis conditions. The results confirm that the developed system, to some extent, is suitable for overcoming the limitation of inert atmosphere for short-chain oligo synthesis. The sealed microchamber and anhydrous conditions were most significant to achieve the adequate protection from atmospheric moisture. The design of the microchamber provides a systematic orientation to the CPG particles, which results in synthesizing oligos with more than a 75% yield by uniform interaction with synthesis reagents. The MS results also demonstrate that reverse synthesis with the developed microfluidic system provides a sufficient purity of short-chain oligonucleotides (<15-mer) with an acceptably low amount of truncated sequences. The reproducibility of synthesis by the developed system opens ups the possibility of designing an automated programmable microfluidic system with multichannel synthesis in the future. However, the optimization of synthesis conditions is still required for long-chain oligos. In future work, we will also investigate methods for the analysis of synthesized oligos for precise quality and accurate yield determination with the introduction of a capping step in the synthesis cycle. We are also developing a computer controlled programmable system for on-demand syntheses of oligonucleotides and oligopeptides, which can further be used for various biosensing, biotechnology, and molecular biology applications.