Express photolithographic DNA microarray synthesis with optimized chemistry and high-efficiency photolabile groups

Abstract

Background: DNA microarrays are a core element of modern genomics research and medical diagnostics, allowing the simple and simultaneous determination of the relative abundances of hundreds of thousands to millions of genomic DNA or RNA sequences in a sample. Photolithographic in situ synthesis, using light projection from a digitally-controlled array of micromirrors, has been successful at both commercial and laboratory scales. The advantages of this synthesis method are its ability to reliably produce high-quality custom microarrays with a very high spatial density of DNA features using a compact device with few moving parts. The phosphoramidite chemistry used in photolithographic synthesis is similar to that used in conventional solid-phase synthesis of oligonucleotides, but some unique differences require an independent optimization of the synthesis chemistry to achieve fast and low-cost synthesis without compromising microarray quality.

Results: High microarray quality could be maintained while reducing coupling time to a few seconds using DCI activator. Five coupling activators were compared, which resulted in microarray hybridization signals following the order ETT > Activator 42 > DCI ≫ BTT ≫ pyridinium chloride, but only the use of DCI led to both high signal and highly uniform feature intensities. The photodeprotection time was also reduced to a few seconds by replacing the NPPOC photolabile group with the new thiophenyl-NPPOC group. Other chemical parameters, such as oxidation and washing steps were also optimized.

Conclusions: Highly optimized and microarray-specific phosphoramidite chemistry, along with the use of the very photosensitive thiophenyl-NPPOC protecting group allow for the synthesis of high-complexity DNA arrays using coupling times of 15 s and deprotection times of 9 s. The resulting overall cycle time (coupling to coupling) of about 50 s, results in a three-fold reduction in synthesis time.

Fig. 1

The optimal coupling times for Activator 42, DCI, ETT and BTT were determined in microarray synthesis and hybridization experiments. DCI and ETT activators result in maximum hybridization signal at very short coupling times whereas the hybridization signal from arrays synthesized with BTT and 42 increases with coupling time. Error bars are the SEM

Fig. 2

Direct comparison activators by synthesizing sets of mixed base 25-mers on a single microarray surface. Each oligonucleotide set was synthesized using a different activator. To avoid order-of-synthesis effects, all sets were synthesized approximately in parallel using the scheme shown in A. The results for Activator 42, DCI and ETT  (Graph B) were evaluated using the intensity values obtained by hybridizing with a Cy3-labeled complementary oligonucleotide

Fig. 3

Scan image details of microarrays synthesized using four different activators. Only the use of DCI activator resulted in microarrays with highly homogenous features. The small features are 14 × 14 µm and the large rectangles are made up of an array of 5 × 5 of the smaller features

Fig. 4

The effect of DCI and ETT activator concentration on hybridization intensity. Higher ETT concentration results in higher hybridization signals, but for DCI, 0.25 M is sufficient to achieve the highest hybridization signal value

Fig. 5

Oxidation optimization. Graph a shows the hybridization intensity features on a microarray synthesized with a long oxidation step in each cycle (A), a long oxidation step in each cycle but no helium drying step (B), a short oxidation step in each cycle (C), and a single final oxidation (F). Values below the expected trend (dotted line) indicate worse results. In Graph b, bars A, B and C are short oxidations steps in each cycle and F is a single final oxidation

Fig. 6

Microarray drying step optimization. Two microarrays synthesized with a variety of helium drying times (5, 15, 30 and 0 s in Graph a; 10, 20, 30, and 0 s in Graph b) indicate that the drying step between coupling and light exposure significantly increases hybridization intensity, but that even short drying times are effective

Fig. 7

Hybridization intensity values for microarrays of 25-mers synthesized using SPh-NPPOC relative to the equivalent arrays synthesized using NPPOC. Synthesis was carried out using an exposure solvent consisting of either 1 or 4 % imidazole in DMSO, and at radiant power values of 7, 34, or 70 mW/cm²

Fig. 8

Left Details of 2.5 mm resolution scan images from gene expression microarrays synthesized with the Legacy method (top) and the Express method (bottom) and hybridized with Cy3-labeled cDNA and synthetic spike in controls. The size of each square is ~14 × 14 µm. Right Scatterplots of the RMA-processed expression data from the gene expression microarrays synthesized with the legacy method (top) and the express method (bottom)

Fig. 9

Synthesis time for high-density gene expression microarrays. The optimizations presented here are labeled as “Express”