An important goal for many areas of biological research and clinical testing is the ability to detect and quantitate thousands of DNA segments in an automated and inexpensive fashion. We are trying to combine the power of DNA amplification with the ability to detect thousands of different sequences at one time using DNA chip technology. A major technical challenge is to develop a method for automatically subdividing a small liquid sample into thousands of smaller droplets that can be repeatedly heated to near boiling temperatures for DNA amplification.We continued to investigate methods to form sub-microliter aqueous droplets in an array format on microscope slides and prevent evaporation during thermocycling. We began by polymerizing thin layers of acrylate on microscope slides by exposing a uniform film of acrylate pre-polymer to ultraviolet light through a mask. Unpolymerized acrylate was washed off and the reverse of the acrylate pattern was transferred to a wax mold by pouring molten wax (melting temperature around 100?C) on the acrylate. The reverse of the pattern in the wax was then transferred to agarose by overlaying with molten agarose (melting temperature around 60?C) and cooling to room temperature. Finally, the reverse of the agarose pattern was transferred to polydimethylsiloxane (PDMS) by overlaying the agarose with a PDMS prepolymer solution covered by a microscope slide. The wax and agarose intermediate molds were necessary because PDMS sticks too tightly to acrylate and wax to be cleanly removed, and each wax mold could be used to form many agarose molds, each of which is destroyed in forming a PDMS-slide. The product at this stage was a microscope slide covered with a film of PDMS about 100-300 um thick with an array of 0.5 - 1mm diameter wells in the PDMS. Oligonucleotide primers for pcr and detection of pcr product were then spotted in the wells by pipeting small droplets of oligonucleotide in a low melting temperature wax-detergent mixture. The purpose of this mixture was to stick the oligonucleotides to the bottom of the well during sample loading in a way that they would be released during thermocycling. Several features of this process had to be optimized including the concentration of detergent, the fraction of water in the emulsion, and the nature of the side chains in the siloxane polymer, which affect the affinity of the wax-oligonucleotide mixture for the well. The siloxane-oligonucleotide-patterned slide was then covered with a second microscope slide spaced about 100 um above the siloxane layer by thin supports at the edges. An aqueous sample containing pcr solution and sample DNA was loaded into the space between the two slides via capillary action. This was followed by about 200 ul of a PDMS pre-polymer mixture which displaced the aqueous sample everywhere except in the region of the wells, thereby creating an array of tiny aqueous samples. The PDMS displacing fluid was allowed to cure for about 30 minutes at room temperature. The device was then heated and cooled to amplify target DNA by pcr and examined with a confocal microscope to determine the fluorescence of each well, which correlated with the amount of DNA amplified according to a standard fluorescence resonance energy transfer assay.While pcr worked in these devices in a portion of the wells, many sample droplets were lost during thermocycling. This apparently occurred via two evaporative mechanisms: (i) separation of siloxane layers from each other or from the surrounding glass, which was detectable by the appearance of a silvery patch in the region of dehiscence, or (ii) escape of water vapor through the siloxane. We tried several methods to strengthen the bonds between layers, including mixing PDMS with silanes, silanizing the glass surfaces with acrylates and adding acrylates to the PDMS, and using PDMS mixtures with excess hydrogen groups in one layer and excess of vinyl groups in the other. The permeability of siloxanes to water vapor is related to their glass transition temperatures. Above the glass transition temperature, permeability increases significantly. The siloxane elastomers we tried have glass transition temperatures below the DNA melting temperature used in pcr. We tried to reduce the porosity of the PDMS by increasing the amount of cross-linker and introducing bulky side-chain groups with octylmethylsiloxane, diethylsiloxane, and diphenylsiloxane. However, these modifications did not adequately solve the problem of evaporation.
| Silver, J; Mi, Z; Takamoto, K et al. (1999) Controlled Formation of Low-Volume Liquid Pillars between Plates with a Lattice of Wetting Patches by Use of a Second Immiscible Fluid. J Colloid Interface Sci 219:81-89 |
