Circular RNA migration in agarose gel electrophoresis

Abstract

Circular RNAs are garnering increasing interest as potential regulatory RNAs and a format for gene expression. The characterization of circular RNA using analytical techniques commonly employed in the literature, such as gel electrophoresis, can, under differing conditions, yield different results when attempting to distinguish circular RNA from linear RNA of similar molecular weights. Here, we describe circular RNA migration in different conditions, analyzed by gel electrophoresis and high-performance liquid chromatography (HPLC). We characterize key parameters that affect the migration pattern of circular RNA in gel electrophoresis systems, which include gel type, electrophoresis time, sample buffer composition, and voltage. Finally, we demonstrate the utility of orthogonal analytical tests for circular RNA that take advantage of its covalently closed structure to further distinguish circular RNA from linear RNA following in vitro synthesis.

Keywords: E-gel; RNase H; RNase R; circRNA; circular RNA; electrophoresis.

Figure 1.. Circular RNA migration using standard agarose systems.

(A) Schematic and expected sizes of precursor linear RNA, linear control RNA, circular RNA, and excised introns. Linear control RNA is splicing incompetent as it lacks the full upstream and downstream introns necessary for splicing but retains remnant introns expected after splicing. Probes targeting the indicated segments of the RNA are numbered 1–6 in the indicated colors. (B) Northern blot correctly identifies circular RNA using standard agarose electrophoresis. Circular RNA and linear control RNA were treated with RNase R for 15 minutes and 60 minutes. Samples were incubated in NorthernMax-Gly Sample Loading Dye containing glyoxal, glycerol and DMSO, and run at 50V for 80 minutes at room temperature using traditional agarose gels cast in our lab. Size markers are indicated on the left; biotinylated probes used for each Northern blot are indicated on the left and correspond with panel A. RNA species are labeled on the right. Gel and Northern blot images were cropped above the 2000nt marker for clarity.

Figure 2.. Circular RNA apparent size changes with electrophoresis conditions.

(A) Circular RNA migration on E-Gel EX systems using sample buffer containing formamide only. Circular RNA migrates slower than precursor RNA when run on E-Gel EX 2% gels using program “EX1–2%” for 10 minutes at room temperature. Extending electrophoresis duration to 18 minutes enhances separation of circular and linear splicing reaction species and maintains the same migration pattern as 10 minutes. Gel artifacts are visible as bubbles in the gel, increasing in severity with longer electrophoresis. RNA species are labeled on the right. (B) Circular RNA migration can overlap precursor RNA when using commercially available sample buffer. Circular RNA was denatured in GLB II sample buffer and electrophoresis performed as in (A). The circular RNA band is clearly separated from the precursor band after 13 minutes of electrophoresis, however overlaps the precursor after 18 minutes. The mass of RNA loaded into the gel does not affect this migration pattern. L, Ladder; R, Circularized RNA not treated with RNaseR. (C) Changes in circular RNA apparent size is specific to the E-Gel EX system. Top panels: E-Gel EX; bottom panels: non-EX E-Gel. RNaseR treated circularized RNA samples were denatured in GLB II buffer and subjected to the same electrophoresis conditions as in (A). Time points are indicated above gels. Circular RNA migration relative to the 2000nt RNA marker is shown by colored arrowheads (red: above 2000nt marker, yellow: with 2000nt marker, green: below 2000nt marker).

Figure 3.. Sample buffer conditions affect migration pattern of circular RNA.

(A) EDTA decreases apparent size of circular RNA and protects circular RNA from nicking and degradation. Circularized RNA (treated with and without RNaseR) were denatured in sample buffer containing reagents depicted in the figure and run on either E-Gel EX 2% or non-EX E-Gel 2% precast gels for the indicated time using program “EX1–2%” at room temperature. Concentrations of each reagent prior to 1:1 mixing with RNA is as follows: Formamide: 95%, EDTA: 18mM, SDS: 0.025%, Bromophenol Blue (BB) and Xylene Cyanol (XC) 0.025% each. Ladder is shown on the left and was prepared in formamide only sample buffer. RNA species are labeled on the right. (B) Buffer carryover in circular RNA preparations alters circular RNA mobility during electrophoresis in the presence of 50mM EDTA. The distance between circular RNA and linear precursor RNA narrows when 1x T4 RNA Ligase I buffer and EDTA are added to samples prior to gel loading, but not when buffer alone is added (red dotted line). A proportionally smaller shift towards lower molecular weight is visible for precursor RNA. RNA species are labeled on the right. (C) Buffer carryover in circular RNA preparations dramatically alters precursor RNA mobility during size exclusion chromatography through the emergence of a new peak, but minimally affects circular RNA mobility. 1x T4 RNA ligase buffer I was added to samples prior to HPLC sample injection with or without 50mM EDTA. Note the lack of introns in all IVT conditions indicating that splicing is not occurring. (D) Lower voltage causes overlap of circular RNA with precursor linear RNA. RNA was prepared in sample buffers as in (A), loaded onto two E-Gel EX 2% gels and subjected to either a high voltage program “EX1–2%” (upper) or low voltage program “SizeSelect2%” (lower) at room temperature for the time periods indicated on the left. These times are comparable with each voltage setting based on the distance of xylene cyanol migration depicted as the white areas on the right side of each gel. Ladder is shown on the left and was prepared in formamide only sample buffer. RNA species are labeled on the right. (E) Denaturation in the presence of SYBR-GOLD I in the sample buffer abrogates the aberrant migration pattern of circular RNA seen on E-Gel EX systems. Ladder and RNaseR-treated circular RNA were denatured in the sample buffers described on top at a 1:1 ratio of sample to buffer, loaded onto E-Gel EX 2% gels, and run for 10 minutes (top panel) or 18 minutes (bottom panel) at room temperature using program “EX1–2%”. RNA species are labeled on the right. Sample buffer conditions prior to 1:1 dilution: Form only: 95% formamide; Form +SyGOLD: 95% formamide with 0.025% SYBR-GOLD I; Form +EtBr: 95% formamide with 0.025% ethidium bromide; GLB II: 95% formamide with 18mM EDTA and 0.025% each of SDS, bromophenol blue, and xylene cyanol; RLD: 95% formamide with 0.5mM EDTA and 0.025% each of SDS, ethidium bromide, bromophenol blue, and xylene cyanol.

Figure 4.. Orthogonal methods to confirm circular RNA identity.

(A) Northern blot analysis of E-Gel EX system. Samples were denatured 1:1 in RLD (2xRNA Loading Dye, ThermoFisher) or GLB II (Gel Loading Buffer II, ThermoFisher), run on E-Gel EX 2% gels and subjected to different voltage programs indicated at the top. Agarose image is shown in top panel, Northern blots shown in lower panels. The probes are depicted on the right of each blot and correspond to the schematic shown in figure 1A. (B) Size exclusion chromatography of circular RNA. Circular RNA elutes off the column with a slightly smaller molecular weight than predicated. Larger and smaller linear RNA species are on the left and right, respectively. Nicked circular RNA is visible as a minor peak at 9.75 minutes, directly before the major circular RNA peak. RNase R digestion identifies circular RNA as a single enriched peak. (C) RNA circularity can be validated by nicking the RNA randomly using heat and divalent cations such as Mg²⁺. 1x T4 RNA ligase I buffer, containing magnesium, was added to RNA sample and RNA in buffer was heated at 70C for the indicated duration. IVT material, which is mostly linear, produces a smear of variable molecular weight species extending from the full-length band when randomly nicked. Circularization reaction material digested with RNase R, which is mostly circular, enriches the lower ‘nicked’ linear RNA band before producing a smear that extends from this band when randomly nicked. No smear extends from the circular RNA band. Note that the top and bottom bands are of equal molecular weight. (D) Samples in (C) were analyzed by HPLC. Degradation of precursor RNA by heat and magnesium shows typical tailing of RNA towards lower molecular weights, consistent with the smear seen by gel electrophoresis. In contrast, degradation of circular RNA results in enrichment of a new peak corresponding to nicked RNA of equivalent molecular weight, which is followed by tailing as degradation continues. Unlike gel electrophoresis wherein intact circular RNA appears at much higher molecular weight than nicked RNA, circular RNA appears smaller than its nicked counterpart when analyzed using HPLC. (E) Site-specific degradation using oligonucleotide-guided RNase H digestion is another method for validating RNA circularity. Site specific degradation will yield two linear products when cutting linear RNA species, and one linear product when cutting circular RNA. Predicted degradation products for precursor RNA in this experiment are 1371nt and 399nt in length, visible as the two major product bands in lane 4. Degraded circular RNA yields a single product at 1455nt. Several off-target digestion products are visible as fainter bands. (F) Samples in (E) were analyzed by HPLC. Degradation of precursor RNA using RNase H yields two smaller fragments. Degradation of circular RNA using RNase H results in enrichment of a single fragment corresponding to nicked RNA of equivalent molecular weight, similar to the product generated by heat/magnesium nicking except that the starts and ends of the nicked molecules are expected to be homogenous.