High-throughput binding characterization of RNA aptamer selections using a microplate-based multiplex microcolumn device

We describe a versatile 96-well microplate-based device that utilizes affinity microcolumn chromatography to complement downstream plate-based processing in aptamer selections. This device is reconfigurable and is able to operate in serial and/or parallel mode with up to 96 microcolumns. We demonstrate the utility of this device by simultaneously performing characterizations of target binding using five RNA aptamers and a random library. This was accomplished through 96 total selection tests. Three sets of selections tested the effects of target concentration on aptamer binding compared to the random RNA library using aptamers to the proteins green fluorescent protein (GFP), human heat shock factor 1 (hHSF1), and negative elongation factor E (NELF-E). For all three targets, we found significant effects consistent with steric hindrance with optimum enrichments at predictable target concentrations. In a fourth selection set, we tested the partitioning efficiency and binding specificity of our three proteins’ aptamers, as well as two suspected background binding sequences, to eight targets running serially. The targets included an empty microcolumn, three affinity resins, three specific proteins, and a non-specific protein control. The aptamers showed significant enrichments only on their intended targets. Specifically, the hHSF1 and NELF-E aptamers enriched over 200-fold on their protein targets, and the GFP aptamer enriched 750-fold. By utilizing our device’s plate-based format with other complementary plate-based systems for all downstream biochemical processes and analysis, high-throughput selections, characterizations, and optimization were performed to significantly reduce the time and cost for completing large-scale aptamer selections. Figure Schematic breakdown of a microplate-based enrichment device for the selection of aptamers (MEDUSA), which can be customized and assembled in both parallel and serial configurations. Up to 96 selections can be performed simultaneously. Electronic supplementary material The online version of this article (doi:10.1007/s00216-014-7661-7) contains supplementary material, which is available to authorized users.

All of the oligos used in this work were obtained from Integrated DNA Technologies.

Preparation of protein-and background-binding aptamers
Sequence verified DNA templates for each one of the specific aptamers used in this study were transcribed using T7 RNA Polymerase. After transcription, the samples were treated with DNase I (Ambion), PAGE-purified, phenol:chloroform and chloroform extracted, isopropanol precipitated, and then re-suspended in DEPC-treated H 2 O.

RNA selections and quantification
The RNA pools were injected at a rate of 33 μL/min for 30 min with a 10 μL aliquot of each pool set aside and used as a standard for quantitative polymerase chain reaction (qPCR) analysis. All buffers and solutions were degassed prior to use and introduced into the microcolumns via programmable multichannel syringe pumps (Harvard Apparatus) with MEDUSA placed onto a 96-well format liquid waste reservoir. The microcolumns were reconfigured to run in parallel by removing the caps and silicone layers permitting the connectivity of microcolumns, and reassembling the device with the appropriate caps for a parallel configuration, and washed with 3 mL of binding buffer at a rate of 300 μL/min. The RNA/RNA-protein complexes were eluted directly into a 96-well microplate from the individual microcolumns by flowing elution buffer [binding buffer + 50 mM ethylenediaminetetraacetic acid (EDTA pH 8.0) for selections with Ni-NTA resin; binding buffer + 10 mM glutathione for selections with GSH resin; binding buffer + 10 mM maltose for selections with amylose-resin] at a rate of 50 μL/min for 12 min. Samples and standards were phenol/chloroform-extracted and ethanol-precipitated together with 1 μL of GlycoBlue (Ambion) and 40 μg of yeast tRNA (Invitrogen), and the resulting pellet was resuspended in 20 µL of RNase-free water, and reverse transcribed with Moloney Murine Leukemia Virus Reverse Transcriptase (MMLV-RT) in two 96-well microplates. The N70 library, HSFapt, NELFapt, BBS1, and BBS2 all contain the same 3' constant region and were reverse transcribed using Lib-REV primer complementary to the 3' constant region in the RNA.
For the experiments containing GFPapt, 4 µL of the resuspended pools and the standards were reverse transcribed using the GFPapt-REV primer specific to GFPapt. A 10-µL volume of each of the cDNA products was used for qPCR analysis using 384-well plates on a LightCycler 480 instrument (Roche) to determine the amount of RNA library and of each specific aptamer that was recovered from each microcolumn. Different sets of oligonucleotides (see above) were used to independently evaluate the amount of N70 library and specific aptamers in each pool.

Descriptions of MEDUSA's Components
Each layer of MEDUSA was fabricated from either transparent biocompatible poly(methyl methacrylate) (PMMA) plastic or silicone. As seen in Figure 1A (lower boxed inset), for parallelized microcolumns, there are 5 layers of plastic and 2 layers of silicone as well as NanoPorts (IDEX Health and Science) for inputs and outputs on each side. The center most plastic layer (number "1" in Figure 1A) is 1/2" thick and contains 96 microcolumns that each hold 10 µL of total volume. The next pair of layers (numbered "5" in Figure 1A above and below the microcolumns) are 1/16" silicone layers for making a liquid tight seal across all 96 microcolumns. These layers contain 2 mm diameter holes for inserting porous polyethylene frits above and below each microcolumn to retain target-bound affinity resins, and have adhesive on one side for bonding to the microcolumn layer. The next pair of layers (numbered "2" in Figure  1A) are 1/4" plastic capping layers which have small holes and NanoPorts (numbered "4") bonded around them to allow solutions to flow in and out of the microcolumns. The outer most plastic layers (numbered "3"in Figure 1A) are 1 mm thick and designed to simultaneously aid the alignment of the NanoPorts to the capping layers, as well as to bear and distribute forces from the assembly of all the layers by acting as a washer. All of the layers contain 35 evenly-spaced holes, with the middle microcolumn layer being threaded, for sealing the device together with screws ( Figure 1A and 1B upper inset photographs). For serialized microcolumns ( Figure 1B, lower boxed inset), the design and assembly is similar. However, there are 2 additional layers of silicone (numbered "6" in Figure 1B). These layers are fabricated in 1/32" silicone (no adhesive) and are programed to allow for the connectivity of microcolumns through small interconnecting channels.

Fig. S1
Layout for the 96 targets on MEDUSA according to its analogous microplate position given by the rows A-H, and the columns 1-12. In section I, the 8 indicated targets were connected in series from A to H to test the specificity and partitioning efficiency of various RNA aptamers. This was tested in triplicate in columns 1 to 3. Sections II, III, and IV tested the effects of target surface concentration on aptamer enrichments. The colored triangles indicate decreasing concentrations of each protein from 10 µg/µL (row A) to 0.016 µg/µL (row H) in 2.5-fold dilutions. Section II (green triangle) aimed to confirm previous enrichment behaviors shown with GFP. Sections III and IV tested the same concentrations of the proteins hHSF1 (blue triangle) and NELF-E (red triangle) to assess the prevalence of target surface concentration effects on binding due to steric hindrance or other effects in other aptamer selections Empty (no Resin) 0 2