SIZE-BASED RNA FRACTIONATION USING ISOTACHOPHORESIS Charbel Eid 1 , Juan G. Santiago 1 , and Robert J. Meagher 2 1 Stanford University, USA 2 Sandia National Laboratory, USA ABSTRACT We have developed a rapid microfluidic technique for size-based RNA fractionation. We aim to address the need for effective size-selection in RNA sequencing library preparation workflows. Our assay leverages isotachophoresis (ITP) to preconcentrate RNA fragments and an ionic spacer and sieving matrix to create two RNA fractions based on a user-defined cutoff size. We demonstrate this approach using synthetic DNA as well as an RNA ladder. Our assay is completed in under 10 min, and easily tunable for user-specified cutoff sizes. KEYWORDS: isotachophoresis, small RNA, mRNA, sequencing, separation INTRODUCTION RNA transcripts circulating in blood and plasma have widespread utility as biomarkers of infection, cancer, and numerous other pathophysiological states. Profiling of circulating RNAs by next-generation sequencing (NGS) techniques like RNA-Seq has immense potential to identify meaningful biomarkers for many diseases.[1] However, circulating RNA are typically <200nt, and include both “small” RNAs (such as microRNA and others) as well as “large” degraded mRNA and other species. Library preparation, sequencing, and bioinformatics analysis of “small” (<50 nt) and “large” (>100 nt) RNA differs significantly. Therefore, though both fractions provide useful information, it is imperative to analyze them separately. Though there are several commercial size-selection kits for DNA fractionation, we only know of Pippin Prep kit to perform small RNA size-selection.[2] Pippin Prep retains RNA molecules less than 150-200 nt, a cutoff that is significantly larger than our intended application. Furthermore, larger RNA fragments are not preserved in this process. The assay presented here addresses this need by integrating ITP-based nucleic acid fractionation in a large- channel microfluidic device capable of processing 10 μL of sample. ITP is an electrophoretic technique that uses two buffer systems, a high-mobility leading electrolyte (LE) and a low-mobility trailing electrolyte (TE). When an electric field is applied, analytes which have mobilities intermediate to those of the LE and TE, focus at a sharp interface between the two.[3] Recently, the Santiago group has demonstrated the use of high-concentration spacer ions, which have intermediate mobility and form a plateau zone, to separate reaction products following ITP- based reaction.[4, 5] Here, we extend the spacer concept to perform size-based RNA fractionation in large PMMA chips, as shown in Figure 1. Our channel consists of two regions; the first containing sample mixed in with TE and spacer ions, the second containing LE and sieving matrix. We use a sieving matrix to increase the difference in effective mobility between the small and large RNA molecules.[6] In the first region, all RNA, independent of size, focus in ITP ahead of both the spacer and TE ions. This is due to the largely size-independent electrophoretic mobility of nucleic acids in free-solution. Upon entering the second region, which contains the sieving matrix, spacer ions gradually overtake the larger RNA molecules but not the smaller RNA, and form a zone separating the two RNA fractions based on size. We design the spacer ion and sieving matrix concentration for a size cutoff of approximately 50-60 nt, though we note that this approach is highly tunable for different size cutoffs. RNA molecules smaller than 50 nt focus at the LE-spacer interface, whereas those larger than 50 nt focus at the spacer- TE interface. EXPERIMENTAL APPROACH We designed a microfluidic device capable of processing 10 μL of sample. We designed the channel layers using AutoCAD software, and fabricated the poly(methyl methacrylate) (PMMA) device using a laser cutter. We used pressure-sensitive double-sided adhesive tape to bond the three PMMA layers together and form leak-proof channels. We added a slit in the top layer above the channel to facilitate creation of a distinct sample zone, and an extraction reservoir to facilitate sample extraction following ITP. The chip layout is shown in Figure 2. We designed buffering reservoirs in order to ensure sufficient buffering capacity and prevent bubbles formed in the reservoir from entering the channel. These reservoirs consisted 1 mL pipette tips filled with buffering TE and LE, a Pd electrode connected to an external voltage source, and a layer of solidified agarose, as shown in Figure 2. We filled the channel in 3 steps: we first loaded the LE by slowly pipetting from the LE reservoir until buffer reached the loading slit. We then filled the rest of the channel by loading the TE, spacer, and sample from the TE reservoir. Finally, we loaded “clean” TE into the TE reservoir. We then initiated ITP separation, which was completed in approximately 10 min. Our LE consisted of 30 mM HCl, 60 mM Tris, 1% PVP, 4M Urea, and 2.5% HEC. PVP was used to suppress electroosmotic flow. Urea is a denaturant, which helps reduce secondary structure in the longer RNA fragments.