Highly Multiplexed RNA Aptamer Selection using a Microplate-based Microcolumn Device

Aptamers are short, often 30 to 100 nucleotides (nt) long, single-stranded nucleic acids with structures determined by their specific nucleotide sequence. These molecules are capable of binding with high affinity and specificity to a variety of targets ranging from small metal ions to membrane proteins on the surface of living cells1. Aptamers are usually discovered using a process called Systematic Evolution of Ligands by EXponential enrichment (SELEX) in which they are selected from a very large, sequence-diverse library of nucleic acids (1012–1016 unique sequences)2,3,4. SELEX is an iterative process, involving cycles of i) binding, where the target is incubated with the library; ii) partitioning, where the target-bound sequences are separated from the free sequences; and iii) amplification, where the target-bound sequences are amplified to create an enriched pool for the next round of selection. This process results in a final pool that is dominated by the strongest binding aptamer candidates. Aptamers are used in applications in a wide range of fields, including diagnostics, therapeutics, biotechnology, chemical analysis and separations5,6,7,8. The standard SELEX process tends to be time consuming and uses a large amount of reagents making aptamer selections costly9,10. To reduce time and reagent consumption, we developed a modified version of SELEX we called RNA APtamer Isolation via Dual-cycles (RAPID)11. This method incorporates non-amplification cycles in which the eluted RNA is purified and used in a second binding cycle without prior amplification of the material. To compare the techniques, RAPID was performed in parallel with traditional SELEX using the same targets, and many of the same top aptamer candidates (~10% identical aptamers in the top 10,000 enriched sequences) emerged. Both selections using RAPID and traditional SELEX started with a library of 5 × 1015 unique sequences, yet the RAPID selection took approximately one third the time as our previously-optimized SELEX process and significantly reduced the amount of reagents used11. In this approach, only one non-amplification cycle was performed between amplification cycles, so it remained unknown whether multiple consecutive non-amplification cycles could be successful and further increase the number of total binding cycles possible within a given time. Previously, we also developed a microcolumn technology designed for aptamer selections using an affinity chromatography approach. The target is immobilized on a resin packed within a microcolumn and the nucleic acid library is pumped through the column12. The scale of these devices (5–30 μL resin capacity) enabled a significant reduction in reagents used for a selection. Flow conditions were carefully optimized to yield the highest enrichment12. The microcolumns can also be connected in series allowing multiplex selections to be performed using a single aliquot of the valuable random library. In later cycles, selections are performed in parallel to avoid potential cross-contamination of enriched libraries12. One drawback of these devices is they are tedious to use in large-scale multiplex experiments. Consequently, we designed and fabricated a scaled-up version of the microcolumns for higher throughput. This device, named MEDUSA (Microplate-based Enrichment Device Used for the Selection of Aptamers), has all of the same benefits as the individual microcolumns, including reduced reagent consumption and the ability to be reconfigured between serial and parallel connections13. MEDUSA was also designed with specific dimensions and spacing that allow it to directly couple to a 96-well microplate. This enabled the continued use of high-throughput techniques in any downstream processing, which could be automated using existing liquid-handling robots. Similarly to individual microcolumn devices, we have optimized selection conditions, including protein density on the resin and flow-rates for aptamer library binding, washing, and elution, to yield the highest enrichment of previously-selected aptamers on their corresponding targets in a high-throughput manner13. In this article, we describe the combination of the multi-well format MEDUSA device with the RAPID approach to simultaneously perform multiplexed aptamer selections to 19 different targets, including 4 background targets. The background targets, including maltose-binding protein (MBP), and amylose, nickel-nitrilotriacetic acid (Ni-NTA), and anti-FLAG resins, served as negative selections in the first round with the microcolumns connected in series. This allowed us to identify non-specific sequences that bound to these resins or the MBP-tag. The primary protein targets are associated with regulation of transcription, RNA stability, or localization. Mouse DXO and its yeast homolog (Ag Rai1) are decapping exoribonucleases that possess pyrophosphohydrolase, decapping and 5′-to-3′ exoribonuclease activities, and are thus implicated in RNA surveillance14,15. PNPase is a mitochondrial polynucleotide phosphorylase with 3′-to-5′ exoribonuclease and poly-A polymerase activities, and has recently been shown to be involved in the import of nuclear encoded RNAs into the mitochondrial matrix, thus indicating its involvement in mitochondrial homeostasis16,17. p23 is a molecular chaperone that initiates the disassembly of protein-DNA complexes, which impacts transcription factor activation potential and response time to environmental cues18. RTF1 is a subunit of the Paf1 complex (PAF1C), a multifunctional protein complex that associates with RNA Polymerase II and is implicated in histone modification, as well as transcriptional and posttranscriptional gene regulation19. NELF-E is a component of the NELF complex that negatively regulates elongation of transcription by RNA polymerase II20. The remaining 8 protein targets are various enzymes and enzymatic subunits involved in histone modification that are key features in epigenetic transcription regulation. These proteins are: TIP60-Chromo, GCN5-GNAT, MOF-Chromo, UTX-JMJC, JMJD2-Clav., JMJD2-JMJC, ASH1-BAH, and Trx-ZnFinger. We exploited the reconfigurable design of the MEDUSA device to enable multiple selections to be performed using just 2 aliquots of library, and its microplate-based dimensions to perform downstream processes such as reverse transcription, PCR, quantitative PCR (qPCR), and transcription in a 96-well plate format. Encouraged by our earlier results with RAPID where amplification and non-amplification cycles are alternating, we decided to take our RAPID approach a step further by increasing the number of non-amplification cycles with later rounds of selection to drastically reduce the time and reagents needed for the selection. We performed the equivalent of a 10-cycle SELEX (with only 4 of these cycles requiring amplification) to 19 targets in about the time it would take to perform 4 cycles of conventional SELEX to one target. Analysis of high-throughput sequencing results allowed us to identify potential aptamer candidates, and many of these candidate aptamers were verified to bind the corresponding target proteins with a dissociation constant (KD) in the nanomolar range. In addition, further inspection of aptamer sequences enriched for DXO led us to identify aptamers that not only bind, but also resist DXO’s exoribonuclease activity. Furthermore, in studies monitoring the degradation of a 30-nt RNA substrate by DXO, the aptamers demonstrated significant inhibition of DXO activity. These and the aptamers to other proteins could be used for inhibiting target protein function, and the sequencing data provides a rich source of information to understand the biology of these target proteins and perhaps uncover novel functions (e.g. RNA interaction) and specificity.