Chemical crosslinking and ligation methods for in vivo analysis of RNA structures and interactions

Abstract

RNA structures and interactions in living cells drive a variety of biological processes and play critical roles in physiology and disease states. However, studies of RNA structures and interactions have been challenging due to limitations in available technologies. Direct determination of structures in vitro has been only possible to a small number of RNAs with limited sizes and conformations. We recently introduced two chemical crosslink-ligation techniques that enabled studies of transcriptome-wide secondary and tertiary structures and their dynamics. In a dramatically improved version of the psoralen analysis of RNA interactions and structures (PARIS2) method, we detailed the synthesis and use of amotosalen, a highly soluble psoralen analogue, and enhanced enzymology for higher efficiency duplex capture. We also introduced spatial 2′-hydroxyl acylation reversible crosslinking (SHARC) with exonuclease (exo) trimming, a method which utilizes a novel crosslinker class that targets the 2′-OH to capture three-dimensional (3D) structures. Both are powerful orthogonal approaches for solving in vivo RNA structure and interactions, integrating crosslinking, exo trimming, proximity ligation, and high throughput sequencing. In this chapter, we present a detailed protocol for the methods and highlight steps that outperform existing crosslink-ligation approaches.

1. Introduction

RNA has been widely recognized as mediators of biological functions in all living organisms and viruses (Alberts et al., 2002). The versatile functions of RNA drive processes such as biochemical reactions and regulation of gene expression (Cech, 2012). Given the complexity of the various RNA classes and their respective roles in cells, it is essential to explore the biological activity and impact on physiology and pathology (Castello, Fischer, Hentze, & Preiss, 2013; Lu & Matera, 2014a; Maenner et al., 2010). One facet of RNA studies that has garnered much attention is RNA structure, as its dynamics in secondary and tertiary structures result in different types of interactions and biological activities (Byeon et al., 2021; Gluick & Draper, 1992; Kruger et al., 1982; Lu et al., 2020; Wang & Chang, 2011; Wu & Bartel, 2017; Yao, Wang, & Chen, 2019). A myriad of different methods have been developed to characterize RNA structures at the atomic level, such as cryo-electron microscopy, x-ray crystallography, nuclear magnetic resonance, and other related technologies (Faruqi & Henderson, 2007; Ferré-D’Amaré, Zhou, & Doudna, 1998; Ma, Jia, Zhang, & Su, 2022). While there are many recent advancements of these methods, such studies remain impractical for the vast majority of cellular RNAs. They require purified and “well-behaving” RNA, while structures in living cells are highly dynamic and heterogeneous in nature, which may be easily disrupted upon purification (Larsen, Choi, Prabhakar, Puglisi, & Puglisi, 2019; Lu & Chang, 2018; Zhang & Keane, 2019; Zhang & Ferré-D’Amaré, 2014).

In parallel with in vitro experimental methods, various chemical and enzymatic methods have been developed to understand the reactivity properties of RNA in cells (Kudla, Granneman, Hahn, Beggs, & Tollervey, 2011; Rouskin, Zubradt, Washietl, Kellis, & Weissman, 2014; Spitale et al., 2013; Spitale et al., 2015; Velema & Lu, 2022; Wang & Padgett, 1989). Probing data obtained from these methods are useful for improving structure modeling (Lu & Chang, 2016). These methods have enabled high-throughput analysis on the structures and interactions of various complex RNA, providing insights into previously unknown biological functions (Cai et al., 2020; Lu et al., 2016; Sharma, Sterne-Weiler, O’Hanlon, & Blencowe, 2016; Spitale et al., 2015). Probing methods such as dimethyl sulfate alkylation (DMS) and selective 2′-hydroxyl acylation primer extension (SHAPE) generate data on nucleotide flexibility/accessibility, which in turn improve secondary structure predictions (Spitale et al., 2013; Wang & Padgett, 1989). Other analogous chemical methods such as mutate-and-map (M²), multiplexed •OH cleavage analysis (MOHCA), and RNA interacting group mutational profiling (RING-MaP) reveal spatial proximity of nucleotides, but provide limited insight for larger and more complex RNAs (Cheng et al., 2015; Homan et al., 2014; Kladwang, VanLang, Cordero, & Das, 2011; Tian & Das, 2016). While markedly improved over initially described methods, they remain narrow in application. For instance, some of these methods are only applicable to synthetic RNAs in vitro or yield poor signal to noise ratios.

Rather than capturing nucleotide flexibility/accessibility, crosslink-ligation methods use chemical crosslinkers to directly capture base pairing between or within RNA. Methods including PARIS, COMRADES, LIGR-seq, and SPLASH have analyzed RNA interactions in vivo, achieving near base pair resolution with single-molecule accuracy in high throughput (Aw et al., 2016; Lu et al., 2016; Lu, Gong, & Zhang, 2018; Sharma et al., 2016; Zhang et al., 2021; Ziv et al., 2018). Chemical crosslinking has evolved into a powerful set of tools that enable the study of RNA structures at high resolution when paired with high throughput sequencing.

One common attribute of the crosslink-ligate methods described above is their use of photo-reversible, psoralen-based crosslinkers. The psoralen family of chemicals was first discovered in the 1970s and has since been widely used to specifically capture interactions between staggered pyrimidines on opposite helical regions of nucleotides (Cole, 1970; Moazed, Stern, & Noller, 1986; Thompson & Hearst, 1983). Although commonplace, psoralen analogues are not flawless, as crosslinking occurs slowly in cells, requires UV irradiation at phototoxic wavelengths, and has limited reactivity, biasing nucleotide capture to pairs of pyrimidine (Calvet & Pederson, 1979; Lu et al., 2016; Zhang et al., 2021). To further drive RNA structural studies, we developed the next generation psoralen analysis of RNA interactions and structures (PARIS2) method, optimizing every step and systematically improving the efficiency of the crosslink-ligation principle (Zhang et al., 2021). More specifically, we reported the simplified synthesis and application of psoralen derivative amotosalen (Lin et al., 1997). In PARIS2, we described the crosslinker-agnostic total nucleic acid (TNA) extraction method, overcoming issues with RNA isolation using the classic acid guanidine thiocyanate phenol chloroform (AGPC) approach. The denatured-denatured 2D (DD2D) gel system robustly isolates pure crosslinked RNA. Moreover, we introduced enzymatic improvements to prevent and bypass photochemical damage to RNA in the crosslink reversal step. Altogether, PARIS2 has resulted in a massively improved (> 4000-fold efficiency) process for studying RNA in living cells (Zhang et al., 2021).

In addition to improvements to the psoralen-based crosslinkers, we used the same crosslink-ligation principle and optimized new crosslinking chemistry to create an orthogonal approach. Bifunctional acylating reagents are an alternative class of crosslinkers that react with the 2′-OH of all four nucleotides, mitigating nucleotide bias as captured by psoralens. The bis-nicotinic azide reversible interaction (BINARI) method has been shown as a potential approach for capturing nucleotide pairs in spatial proximity, detailing 3D structure information (Velema, Park, Kadina, Orbai, & Kool, 2020). However, despite the advantages of BINARI reagents, the compound requires a complex synthesis procedure, and the crosslinks are difficult to reverse, resulting in poor performance in measuring RNA tertiary contacts in cellular applications. To address this issue, we introduced the spatial 2′-hydroxyl acylation reversible crosslinking (SHARC) method to target the 2′-OH of RNA (Van Damme et al., 2022). it overcomes technical challenges in the synthesis of BINARI-like reagents and simplifies the acylation and hydrolysis reversal reactions. The SHARC method further improves on the PARIS2 method, increasing the precision of distance measurements between crosslinked atoms using an optimized exonuclease trimming reaction. Treating crosslinked fragments with RNase R reduces the nucleotides from the 3′-end until blocked by the crosslink, and upon ligation, the two arm ends joint to form a continuous RNA molecule, thus representing a pair of nucleotides near each other (Van Damme et al., 2022). The SHARC-exo method has been demonstrated to capture distance information, suitable for constraining 3D RNA modeling in living cells. Moreover, the high-throughput method measures distances between nucleotides either within or between RNA molecules.

In this chapter, we describe the optimized experimental protocol using SHARC and PARIS reagents for RNA 2D and 3D structure crosslinking studies (Fig. 1). This approach takes advantage of two distinct crosslinking mechanisms and enzyme-based structure reactivity to effectively increase the number of duplexes captured. We detail the TNA method for efficient RNA recovery and show examples of the DD2D gel system isolating crosslinked RNA. The protocol can be applied to study a myriad of biological systems and reveal new RNA functions and mechanisms. After obtaining the gapped/chimeric reads from high throughput sequencing, the data can be analyzed to build complex structure models using several published computational tools, which are described elsewhere (Lu & Matera, 2014b; Travis, Moody, Helwak, Tollervey, & Kudla, 2014; Zhang et al., 2022; Zhou et al., 2020).

Fig. 1.

Overview of the crosslink-ligate workflow with improvements. Specific optimization details have been previously described (Van Damme et al., 2022; Zhang et al., 2021). (A) Cells are treated with crosslinking reagent and subject to irradiation (PARIS2) or simple incubation (SHARC). (B) The TNA method is used to isolate RNA and DNA, which is then treated with DNAse I to isolate pure RNA. (C) RNase III is used to fragment double- and single- stranded RNA. (D) The DD2D gel system is used to select crosslinked RNA. (E) Exonuclease trimming is achieved by RNase R and the arm ends ligated. (F) Crosslink reversal is completed with acridine-orange protected irradiation (PARIS2) or mild hydrolysis (SHARC). (G) Standard reverse transcription (SHARC) or reverse transcription with modified buffer to bypass photochemical damage (PARIS2). (H) cDNA library preparation and sequencing following a standardized protocol, using custom barcodes.

2. Materials

2.1. Before you begin

All materials used for RNA work should be RNase free. To mitigate sample degradation, all steps involving RNA are carried out with tubes on ice. Since completion of the protocol takes multiple days, safe pause points are denoted in the protocol as (*). At pause points, samples should be kept frozen (−20 °C) unless otherwise stated. Expected yields for the various steps are detailed in Table 1.

Table 1.

Overview of start and end amount for the major steps.

Step	Starting	End
TNA extraction	—	200–300 μg
DNase I digestion	100 μg	50–75 μg
RNase III digestion	10 μg	7–9 μg
DD2D gel	10 μg	50–200 ng
cDNA library	~100 ng	5–20 ng

Note that the values shown refer to HEK293T cells grown to 70–80% confluence (~8 × 10⁶ cells) in a 10 cm dish. Experiments can be scaled to the number of cells available.

2.2. Cell culture

2.2.1. HEK293T cells

Cells can be purchased from ATCC and maintained at 37 °C, 5% CO₂ in Dulbecco’s Modified Eagle’s Medium (DMEM) with l-glutamine, supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin.

Note: While this protocol is broadly applicable to different cell lines, it assumes that cells are adherent and prepared in a 10 cm dish. Non-adherent cells need to be spread in the dish to allow efficient UV crosslinking for psoralens (the PARIS method).

2.3. Reagent setup

2.3.1. APS (ammonium persulphate)

Dissolve 1 g of APS in a final volume of 10 mL of NFW. Store at 4 °C for up to 6 months.

2.3.2. Gel elution buffer

Add 1 mL Tris-HCl (1 M, pH 7.5), 5 mL sodium acetate (3 M, pH 5.2), 100 μL EDTA (0.5 M, pH 8.0), 1.25 mL SDS (10% w/v), and make to 50 mL volume using NFW. If SDS precipitates from the solution, warm the elution buffer to 37 °C and invert to mix prior to use. Store at room temperature for up to 3 months.

2.3.3. SHARC crosslinker (DPI)

SHARC reagents are made by dissolving 1-part SHARC reactant (e.g., 2,6-pyridinedicarboxylic acid) and 2-parts CDI in anhydrous DMSO (in terms of molar ratio). Pipette SHARC reactant solution into the CDI solution. After a brief vortex and spin down, insert a needle through the tube cap to allow CO₂ ventilation. Let the mixed solution react at room temperature for 30 min prior to crosslinking. Aspirate only the solubilized part of the reaction for crosslinking and do not disturb the precipitate, if there remains any. Note that an indicator of successful SHARC reagent activation is production of CO₂ bubbles.

2.3.4. PARIS crosslinker (amotosalen)

PARIS reagent is prepared by transferring 0.010 g amotosalen to a 1.5 mL tube and dissolving in 1 mL NFW. Upon dissolution, dilute 1:10 in NFW and store aliquots of 1 mg/mL at −20 °C. 4′-aminomethyltrioxsalen hydrochloride (AMT) can also be used, but it is less efficient.

2.4. Key resources table

Reagent or resource	Source	Identifier
Chemicals and Reagents
1,1’-Carbonyldiimidazole (CDI)	Sigma Aldrich	115533–10G
2,6-Pyridinedicarboxylic acid	Thermo Scientific	131880250
Acridine orange	Invitrogen	A1301
Boric acid	Thermo Scientific	33253.36
dNTP Mix (10 mM)	Thermo Scientific	R0192
Guanidine thiocyanate	Invitrogen	AM9422
Water saturated phenol (pH 6.6)	Invitrogen	AM9712
Commercial Assays
DNA Clean & Concentrator-5	Zymo	D4014
Gel DNA Recovery Kit	Zymo	D4008
Qubit RNA High Sensitivity Assay Kit	Fisher Scientific	Q32852
RNA Clean & Concentrator-5	Zymo	R1016
SPRI magnetic beads (DNA)	BioDynami	40051S
Enzymes
10X 5’ DNA Adenylation reaction buffer	New England Biolabs	B2610S
10X CircLigase II buffer	Lucigen	SS000881-D2
10X RecJf buffer	New England Biolabs	B7002S
10X RNase R buffer	BioVision	M1228
10X ShortCut buffer	Invitrogen	B0245S
10X T4 RNA ligase buffer	New England Biolabs	B0216SVIAL
10X TURBO DNase buffer	Invitrogen	4022G
5’ deadenylase	New England Biolabs	M0331S
ATP (100 mM)	New England Biolabs	N0437AVIAL
CircLigase II ssDNA Ligase	Lucigen	E0129–100D4
High Concentration T4 RNA ligase 1	New England Biolabs	M0437MVIAL
Mth RNA Ligase	New England Biolabs	M2611
Phusion HF	New England Biolabs	M0530L
Phusion HF Buffer (5X)	New England Biolabs	B0518S
Proteinase K	New England Biolabs	P8107S
RecJf	New England Biolabs	M0264L
RNase H	New England Biolabs	M0297S
RNase R	BioVision	M1228
ShortCut RNase III	New England Biolabs	M0245L
SUPERase In	Invitrogen	AM2694
SuperScript IV Reverse Transcriptase	Invitrogen	18090010
TURBO DNase I	Invitrogen	AM2222
Library Oligonucleotides
ddcRNA adapter	/5rApp/AGATCGGAAGAGCGGTTCAG/3ddC/
P3-Tall	GGCAT-TCCTG-CTGAA-CCGCT-CTTCC-GATCT
P6-Tall	CTCTT-TCCCC-TTGTG-TGTGA-AGCGA-AGGGT
P3-Solexa	5’-CAAGC-AGAAG-ACGGC-ATACG-AGATC-GGTCT-CGGCA-TTCCT-GCTGA-ACCGC-TCTTC-CGATC-T-3’
P6-Solexa	5’-AATGA-TACGG-CGACC-ACCGA-GATCT-ACACT-CTTTC-CCTAC-ACGAC-GCTCT-TCCGA-TCT-3’
P6_custom_SeqPrimer	CACTC-TTTCC-CCTTG-TGTGT-GAAGC-GAAGG-GTA
RT_primers	/5phos/WWW-NNN-XXXXXX-NNNNN-TACCC-TTCGC-TTCAC-ACACA-AG/iSp18/GGATCC/iSp18/TACTG-AACCGC W = A/T, N = A/T/G/C; discriminates PCR duplicates “XXXXXX” = 6-mer experimental barcode /iSp18/ = spacer for preventing PCR concatemers
6-mer barcodes	1. ATCACG, 2. CGATGT, 3. TTAGGC, 4. TGACCA, 5. ACAGTG, 6. GCCAAT, 7. CAGATC, 8. ACTTGA, 9. GATCAG, 10. TAGCTT, 11. GGCTAC, 12. CTTGTA, 13. AGTCAA, 14. AGTTCC, 15. ATGTCA, 16. CCGTCC, 17. GTCCGC, 18. GTGAAA, 19. GTGGCC, 20. GTTTCG, 21. CGTACG, 22. GAGTGG, 23. ACTGAT, 24. ATTCCT

3. Materials and equipment

Source and identifiers for common equipment, materials, and reagents are omitted from the lists.

3.1. Equipment (general)

1. Blue light transilluminator

2. CO₂ incubator

3. Fluorometer

4. Magnetic separation rack

5. PCR System

6. Refrigerated microcentrifuge

7. Rotator

8. Spectrophotometer

9. ThermoMixer

10. UV crosslinker

11. Vortexer

3.2. Equipment (electrophoresis)

1. 100 bp DNA ladder

2. 10X MOPS (0.2 M MOPS pH 7.0, 20 mM sodium acetate, 10 mM EDTA pH 8.0)

3. 10X TBE (1 M Tris Base, 1 M Boric acid, 0.02 M EDTA)

4. 50 bp DNA ladder

5. 50X TAE (2 M Tris base, 1 M acetic acid, 50 mM EDTA)

6. 6X Loading Dye (10 mM Tris-HCl (pH 7.6) 0.03% bromophenol blue, 0.03% xylene cyanol FF, 60% glycerol 60 mM EDTA.)

7. Agarose gel running tank

8. Ammonium persulfate (APS)

9. Denaturing gel loading buffer (95% Formamide, 18 mM EDTA, and 0.025% SDS, Xylene Cyanol, and Bromophenol Blue)

10. dsRNA Ladder

11. Gel imager

12. High-voltage power supply

13. Mini-PROTEAN Comb, 15-well 1.0 mm, 26 μL

14. Mini-PROTEAN Glass Plates

15. Mini-PROTEAN Casting Frame

16. Mini-PROTEAN Casting Stand

17. Mini-PROTEAN Short Plates

18. Noble agar

19. SYBR Gold (10,000X)

20. Tetramethylethylenediamine (TEMED)

21. UreaGel Buffer (0.89 M Tris-Borate-20 mM EDTA buffer pH 8.3 (10X TBE) and 7.5 M urea)

22. UreaGel Concentrate (237.5 g of acrylamide, 12.5 g of methylene bisacrylamide per liter, and 7.5 M urea)

23. UreaGel Diluent (7.5 M urea in deionized water)

3.3. Materials

1. 0.45 μm Spin-X centrifuge tube filters

2. 0.65-mL microcentrifuge tubes

3. 1.5-mL microcentrifuge tubes

4. 10 cm cell culture dish

5. 2.0-mL microcentrifuge tubes

6. 25 G needle

7. 8-strip PCR tubes

8. Cell lifter

3.4. Reagents

1. 10X PBS (phosphate buffered saline, KH₂PO₄ 1.44 g/L, NaCl 90 g/L, Na₂HPO₄-7 H₂O 7.95 g/L)

2. Dithiothreitol (DTT, 1 M)

3. DMSO (100%)

4. EDTA (0.5 M, pH 8.0)

5. GlycoBlue coprecipitant (15 mg/mL)

6. MnCl₂ solution (0.1 M)

7. Nuclease free water (NFW)

8. PEG8000 (50% w/v)

9. Potassium acetate solution (3 M, pH 5.5)

10. Reagent grade ethanol (100%)

11. Reagent grade ethanol (70%)

12. Reagent grade isopropanol (100%)

13. Sodium acetate buffer (3 M, pH 5.2)

14. Sodium dodecyl sulfate (SDS, 10% w/v)

15. Tris-HCl (1 M, pH 8.0)

4. Step-by-step method details

The following is a detailed description of the crosslink-ligate method using two crosslinker classes. Examples of successful implementations of specific steps are given in the expected outcomes section. Standard procedures (e.g., gel casting) are omitted and readers should refer to manufacturer instructions. Crosslinking can be achieved with SHARC or PARIS reagents, based on experimental design. Note that the stepwise procedures differ as noted in the respective subsections for these two types of crosslinkers.

4.1. Crosslinking mammalian cells

4.1.1. Crosslinking with SHARC reagents (DPI)

Timing: 1 h 30 min

1. Culture adherent cells in a 10 cm dish to 70–80% confluence. Remove media from the cell culture dish and wash with 1 mL PBS (1X), aspirating away any liquid. Repeat wash step once more.

2. Use a cell lifter to scrape cells from the dish and transfer to a fresh 1.5-mL centrifuge tube. Add 500 μL PBS (1X) to the dish and scrape again, transferring remaining cells and PBS to the tube.

3. Centrifuge (3600 × g, 5 min, 4 °C) and remove the supernatant without disturbing the pellet. Resuspend in PBS (1X) to a volume of 900 μL.

4. Prepare a 10X crosslinking solution (e.g., 250 mM DPI) and let react at room temperature (30 min).

5. Add 100 μL of the DPI solution to each tube of resuspended cells and incubate on a rotator (30 min, room temperature).

6. Centrifuge (3600 × g, 5 min, 4 °C) and remove the supernatant without disturbing the pellet.

7. Wash cell pellet twice with 1 mL PBS (1X). Remove any remaining liquid and store pellets at − 80 °C.^(*)

Note:

• Preparation of DPI will produce CO₂ bubbles throughout the 30-minute activation period. If no bubbles are present, the reaction has likely failed.

• For a low number of samples, cells can be washed, transferred, and resuspended during the 30 min SHARC crosslinker activation period.

4.1.2. Crosslinking with PARIS reagents (amotosalen)

Timing: 1 h 15 min

1. Culture adherent cells in a 10 cm dish to 70–80% confluence. Remove media from the cell culture dish and wash with 1 mL PBS (1X), carefully aspirating away any liquid. Repeat wash step twice more.

2. Prepare a 500 μL crosslinking solution (e.g., 250 μL PBS (2X) + 250 μL amotosalen (1 mg/mL)) for each dish and add to the dish without disturbing the cells. Gently swirl the solution across the cells by rocking the dish back and forth. Incubate (20 min, room temperature).

3. Fill trays with ice and pack tightly to ensure a flat surface. Irradiate (365 nm, 30 min) 15 cm away from the UV bulbs, swirling the cross-linking solution over the cells every 10 min and keeping the surface of the ice flat.

4. Use a cell lifter to scrape cells from the dish and transfer by pipette to a fresh 1.5-mL tube. Add 500 μL PBS (1X) to the dish and scrape again, transferring remaining cells and PBS to the tube.

5. Centrifuge (3600 × g, 5 min, 4 °C) and remove the supernatant without disturbing the pellet.

6. Wash cell pellet twice with 1 mL PBS (1X). Remove any liquid and store pellets at −80 °C.^(*)

Note:

• Crosslinking in the presence of psoralens will cause certain cell lines to detach from the dish, therefore care should be taken when handling cell dishes.

4.2. Total nucleic acid (TNA) extraction

Timing: 1 h 30 min

1. For each 1.5-mL tube containing a crosslinked cell pellet, add 100 μL GuSCN (6 M). Lyse cells vigorously for 1 min by vortex or by running the tube across a rack.

2. To the lysate, add 12 μL EDTA (500 mM), 60 μL PBS (10X), and bring the total volume to 600 μL with NFW (~350 μL). Mix well by vortex and briefly spin down.

3. Add proteinase K to a final concentration of ~1 mg/mL (30 μL of 20 mg/mL), pipette mix, and incubate on a thermomixer (900 RPM, 60 min, 37 °C). Manually invert the tube every 15 min during the incubation to facilitate thorough mixing.

4. After proteinase K digestion, briefly spin down, add 60 μL sodium acetate (3 M, pH 5.3) and 600 μL water-saturated phenol (pH 6.7). Mix well by vortex and briefly spin down.

5. Divide the sample into two 1.5-mL tubes and add 600 μL of pure isopropanol to each tube. Mix well.

6. Centrifuge (17,000 × g, 20 min, 4 °C) and remove the supernatant. Wash the pellet twice with ethanol (70%). Combine pellets from the two tubes, resuspend in 300 μL NFW, and quantify by spectrophotometer.^(*)

Note:

• Fully dislodge the cell pellet from the bottom of the tube after adding GuSCN. After mixing, the lysate should be homogenous, but may not be entirely clear.

• It is optional, but highly recommended to re-precipitate the TNA suspension for a more accurate quantification. Briefly, use a P1000 pipette tip to agitate the insoluble pellet. Incubate on a thermomixer (900 RPM, 10 min, 37 °C) and centrifuge (17,000 × g, 5 min, 4 °C) before transferring the supernatant to a fresh tube. Use the standard isopropanol precipitation method to pellet TNA, resuspend in 300 μL NFW, and quantify by spectrophotometer using dsDNA measurement.

4.3. DNase I digestion

Timing: 45 min

1. Transfer 100 μg TNA to a fresh 1.5-mL tube. Add 20 μL 10x TURBO^™ DNase Buffer, 25 μL TURBO^™ DNase I. Bring the final reaction volume to 200 μL using NFW. Incubate on a thermomixer (500 RPM, 20 min, 37 °C).

2. After DNAse I digestion, add 20 μL sodium acetate (3 M, pH 5.3), 220 μL water-saturated phenol (pH 6.7), 3 μL GlycoBlue, and 450 μL pure isopropanol. Mix well by vortex.

3. Centrifuge (17,000 × g, 20 min, 4 °C) and remove the supernatant. Wash pellet twice with ethanol (70%). Resuspend RNA in 20 μL NFW and quantify by spectrophotometer.^(*)

4.4. RNase III digestion

Timing: 30 min

1. Transfer 10 μg RNA to a fresh 1.5 mL tube. Add 4 μL 10X ShortCut Buffer, 4 μL 50 mM MnCl₂, 10 μL ShortCut RNase III. Bring the final reaction volume to 40 μL using NFW. Incubate (at exactly 5 min, 37 °C).

2. After RNase III digestion, immediately add 60 μL of water-saturated phenol (pH 6.7) and mix well. Briefly spin down, add 4 μL sodium acetate (3 M, pH 5.3), 3 μL GlycoBlue, and 360 μL pure ethanol. Mix well by vortex.

3. Centrifuge (17,000 × g, 20 min, 4 °C) and remove the supernatant. Wash pellet twice with ethanol (70%). Resuspend RNA in 10 μL NFW and quantify by spectrophotometer.^(*)

Note:

• Immediately quench the RNAse III digestion after the specified time. Extended incubation with enzyme will degrade RNA fragments and decrease both the quality and quantity of starting material for library preparation.

4.5. DD2D gel purification

Timing: 2 h + Overnight.

The electrophoresis buffers and temperatures differ for SHARC and PARIS samples. For casting and running SHARC gels, use 1X MOPS. Keep the running buffer at a maximum temperature of 60 °C to prevent premature decrosslinking. For PARIS gels, use the UreaGel system buffer according to manufacturer specifications. Running buffer is 0.5X TBE, maintained between 60 and 70 °C to promote complete denaturation.

First-dimension gel.

• 1Prepare an 8%, 1.5 mm thick denaturing gel using the UreaGel system with a suitable buffer (~8 cm × 8 cm). Use 15-well combs to give the second dimension gel a higher resolution.
• 2Add 10 μL GLB II loading dye (2X) to 10 μL of RNA. Use 200 ng dsRNA ladder as a molecular weight marker.
• 3Run the first-dimension gel (30 W, 10 min PARIS or 200 V, 12 min SHARC).
• 4After electrophoresis, stain the gel by rocking with 2 μL SYBR Gold in 20 mL running buffer (> 5 min). Image the gel using a 300 nm transilluminator and excise each lane from 50 nt to the top of the gel lane (Fig. 2A).

Fig. 2.

DD2D gel system overview. (A) Depiction of 1D gel slice with size for excision (> 50 bp) boxed. (B) The glass top plate is used to align and sandwich the gel slice(s) prior to clasping the entire assembly together with tape or binder clips. Pouring the 2D gel solution from the top of the glass plates mitigates bubble formation and ensures complete 1D slice encapsulation. (C) Fully assembled DD2D gel system with control and crosslinked RNA slices oriented in head-to-toe manner, with sufficient spacing around each slice of 1D gel.

Second dimension gel

• 5Place the notched bottom plate onto a pipette tip box and arrange the gel slices in a head-to-toe manner with a small gap between them. Use the short glass plate as an aligning guide, such that when pouring the gel solution, the first-dimension gel slice will be completely encapsulated. Apply 20 μL running buffer to each gel slice to avoid air bubbles later in the assembly (Fig. 2B).
• 6With the 1D slices correctly oriented, assemble the cassette by aligning the bottoms of the plates to be flush. Hold the plates together using binder clips or tape and remove any excess running buffer.
• 7Prepare a 16% denaturing gel solution using the UreaGel system with a suitable buffer. Use a 5 mL pipette to dispense the 2D gel solution from the top of the notched plate (side closest to gel slices). If there are bubbles, tilt the plates and force them to the top. Use thin loading tips to draw out any remaining air bubbles (Fig. 2B).
• 8Run the second-dimension gel (30 W, 50 min PARIS or 200 V, 55 min SHARC) (Fig. 2C).
• 9After electrophoresis, stain the gel by rocking with 2 μL of SYBR Gold in 20 mL running buffer (> 5 min). Image the gel using a 300 nm transilluminator and excise the region of the second-dimension gel containing crosslinked RNA, which typically appears as a distinct arc above the diagonal.

Gel purification.

1. Use a syringe to punch a hole in the bottom of a fresh 0.65-mL centrifuge tube and insert into a 2-mL collection tube. Transfer the excised gel slices to the stacked tubes.

2. Centrifuge (17,000 × g, 6 min, room temperature). Gel slices should be sheared into slurry by passing through the apparatus.

3. Add 3x gel volume of gel elution buffer and invert to mix. Freeze-thaw the slurry three times and rotate overnight at room temperature.^(*)

4. Transfer 600 μL gel slurry to Spin-X 0.45 μm column. Centrifuge (3400 × g, 1 min, room temperature) and transfer the supernatant to a new 2 mL tube. Repeat until all the gel slurry is filtered.

5. Precipitate RNA using standard isopropanol precipitation with glycogen as a carrier.

6. Resuspend RNA in 11 μL NFW and dilute 1 μL RNA sample for Qubit fluorometer analysis. The expected yield is typically 0.1–0.5% from 10 μg input RNA.^(*)

Note:

• Flush wells with a P1000 prior to 1D gel loading to agitate any urea settled in the lane.

• To increase the working time prior to gel polymerization, keep the gel solution on ice and mix in APS (10%) and TEMED immediately before pouring.

• Whenever possible, separate the samples with blank wells in the 1D gel to simplify excision and prevent cross-contamination due to spillover.

• To maintain a consistent 1D gel slice size between gel runs, while the gel is still in the buffer tank apparatus, visually inspect the lower dye front and mark the outside of the buffer tank.

• Trim the 1D gel whiskers close to the top of the lane to ensure proper encapsulation of the 1D slice without trapping air bubbles during the 2D gel casting.

• Gel electrophoresis at 30 W generates heat and will curl the 2D gel. Let cool to room temperature before removing the gel from the cassette and staining.

• The Spin-X filter columns will sometimes get clogged with crushed slurry. Remove any supernatant from the collection tube, cap the column and invert. Sharply tap the filter column onto a hard surface to dislodge the slurry.

4.6. Exonuclease trimming

Timing: 2 h (or overnight) + 30 min

1. Prepare an RNase R exonuclease reaction. To 10 μL RNA, add 1 μL ATP (100 mM), 2 μL 10 X RNase R Buffer, and 2 μL RNase R. Bring the final reaction volume to 20 μL using NFW. Incubate (12 h, 45 °C).^(*)

2. After trimming, add 2 μL sodium acetate (3 M, pH 5.3), 22 μL water-saturated phenol (pH 6.7), 3 μL GlycoBlue, and 141 μL pure ethanol. Mix well by vortex.

3. Centrifuge (17,000 × g, 20 min, 4 °C) and remove the supernatant. Wash pellet twice with ethanol (70%). Resuspend RNA in 10 μL NFW.^(*)

Note:

• The RNase R exonuclease reaction can be used if necessary to refine the resolution. Overnight (12 h) trimming leads to significant sample loss. In experiments where crosslinker concentration is low, a shorter trimming time (2 h) may be a better choice.

4.7. Proximity ligation

Timing: 1 h 30 min

1. Prepare a proximity ligation mixture. To 10 μL RNA, add 2 μL 10X 5′ DNA Adenylation Reaction Buffer, 2 μL Mth RNA Ligase, 1 μL SUPERase In, and 5 μL NFW. Incubate (20 min, 65 °C) and inactivate afterwards (5 min, 85 °C).

2. Add 2 μL proteinase K (20 mg/mL) and incubate (30 min, 37 °C).

3. After proteinase K digestion, add 2 μL sodium acetate (3 M, pH 5.3), 2 μL GlycoBlue, 25 μL water-saturated phenol (pH 6.7), and 60 μL pure isopropanol. Mix well by vortex.

4. Centrifuge (17,000 × g, 20 min, 4 °C) and remove the supernatant. Wash pellet twice with ethanol (70%). Resuspend RNA in 10 μL NFW.^(*)

4.8. Crosslinking reversal

4.8.1. Decrosslinking SHARC reagents

Timing: 2 h 30 min

1. Prepare a decrosslinking reaction mixture. To 10 μL ligated RNA, add 10 μL of 5X decrosslinking buffer (500 mM Boric Acid, pH 11) and 30 μL NFW. Mix well by vortex and incubate in a PCR block (120 min, 45 °C).

2. Transfer reverse-crosslinked sample to a fresh 1.5-mL tube. Add 5 μL sodium acetate (3 M, pH 5.3), 2 μL GlycoBlue, and 165 μL pure ethanol. Mix well by vortex.

3. Centrifuge (17,000 × g, 20 min, 4 °C) and remove the supernatant. Wash pellet twice with ethanol (70%). Resuspend RNA in 6 μL of NFW.(*)

4.8.2. Decrosslinking PARIS reagents

Timing: 1 h 30 min

1. Prepare a decrosslinking reaction mixture. To 10 μL ligated RNA, add 2 μL of acridine orange (2.5 mM) and 8 μL NFW. Mix well by pipette and transfer to a clean surface (e.g., a fresh 6-cm dish lid). Place the clean surface on an ice filled tray and irradiate (254 nm, 30 min).

2. Transfer reverse-crosslinked sample to a fresh 1.5-mL tube. Add 2 μL sodium acetate (3 M, pH 5.3), 2 μL GlycoBlue, and 60 μL pure ethanol. Mix well by vortex.

3. Centrifuge (17,000 × g, 20 min, 4 °C) and remove the supernatant. Wash pellet twice with ethanol (70%). Resuspend RNA in 6 μL of NFW.^(*)

4.9. Adapter ligation

Timing: 4 h 15 min (or overnight + 1 h).

1. Heat crosslink reversed RNA (90 s, 80 °C) and snap cool on ice (> 2 min).

2. To 6 μL RNA, add the adapter ligation mixture: 2 μL 10X T4 RNA ligase buffer, 1 μL DTT (0.1 M), 5 μL PEG8000 (50% v/v), 2 μL DMSO, 3 μL ddc RNA adapter (10 μM), 1 μL High Concentration T4 RNA ligase 1. Incubate (3 h or overnight, room temperature).*

3. After adapter ligation, remove the free adapters. Add 3 μL 10X RecJf buffer, 2 μL RecJf, 1 μL 5′ deadenylase, 1 μL Superase In, and 3 μL NFW. Incubate in a PCR block (30 min, 30 °C; 20 min, 37 °C).

4. Add 20 μL NFW to each sample, increasing total volume to 50 μL. Purify using the Zymo RNA clean & concentrator-5 kit standard protocol. Elute with 12 μL NFW (use 6 μL twice on the column).^(*)

4.10. Reverse transcription

Timing: Overnight + 45 min

1. To the purified RNA, add 2 μL custom RT primer with barcode and 1 μL dNTPs (10 mM). Heat the samples to 65 °C for 5 min in a PCR block and snap cool on ice.

2. Prepare the 5x SSIV buffer modified with Mn²⁺: 4 μL Tris-HCl (1 M, pH 8.3), 1.2 μL CH₃COOK (5 M), 2.4 μL MnCl₂ (50 mM), and 8.4 μL NFW.

3. Add reverse transcriptase mix to the RNA. 4.6 μL 5x SSIV Mn²⁺ buffer, 2 μL DTT (100 mM), 1 μL SUPERase In, 1 μL SuperScript IV. Set up the qPCR program: 25 °C for 15 min, 42 °C for 10 h, 80 °C for 10 min, 10 °C hold.^(*)

4. After the reverse transcription, add 1 μL RNase H and incubate (30 min, 37 °C).

5. Purify using the Zymo DNA clean & concentrator-5 kit. Add 7x binding buffer and 8x pure ethanol to bind. Wash according to the standard protocol. Elute with 16 μL NFW (use 8 μL twice on the column).

4.11. cDNA circularization, library PCR, sequencing

Timing: 4 h 30 min

• 1Prepare the circularization reaction mixture. To the cDNA, add 2 μL 10x CircLigase II Buffer, 1 μL MnCl₂ (50 mM), 1 μL CircLigase II Enzyme. Incubate in a PCR block (100 min, 60 °C; 10 min, 80 °C).
• 2Prepare the Phusion HF 2X mix: 40 μL 5X HF buffer, 4 μL 10 mM dNTP, 2 μL Phusion, 54 μL NFW.
• 3To the circularized cDNA, add 1 μL P3/P6 Tall (20 μM) and 20 μL Phusion HF 2X. Set up the qPCR program: 98 °C, 2 min; 10 cycles: 98 °C, 15 s; 60 °C, 30 s; 72 °C, 45 s; 10 °C hold.
• 4Transfer PCR product to a fresh 1.5-mL tube and purify using SPRI DNA beads. Add 2x volume of SPRI DNA beads and mix well by pipetting. Incubate (5 min, room temperature) and place on the magnetic rack (5 min, room temperature). Remove the supernatant and wash pellet once with 80% ethanol (200 μL) with the tube still on the magnetic rack. Air dry for 2 min. Elute twice with 10.5 μL NFW (recover ~20 μL).
• 5To the purified DNA product, add 21 μL 2X PCR Solexa mix: 1 μL P3/P6 Solexa (20 μM) and 20 μL Phusion HF 2X. Set up the qPCR program: (98 °C, 2 min; 3 cycles of 98 °C, 15 s; 65 °C, 30 s; 72 °C, 45 s; and 10 °C hold).
• 6Purify the cDNA library using the Zymo DNA clean & concentrator-5 kit standard protocol. Elute with 12 μL NFW (use 6 μL twice on the column) and add 3 μL Orange G loading dye.

cDNA library purification and sequencing

• 7Prepare an agarose gel (2%) and load samples, using 50 bp and 100 bp ladders on the ends as a molecular weight marker. Run gel (120 V, 40 min) in 1X TAE or until the 100–200 bp region is clearly separated.
• 8Excise the library band from 200 bp to 400 bp (corresponding to a > 60 bp insert) and recover DNA using the Zymo agarose gel extraction kit. Elute twice with 6 μL of elution buffer.^(*)
• 9Sequence the libraries on an Illumina sequencer using standard conditions and the P6_Custom_seqPrimer. For detailed data analyses, refer to (Zhang, M. et al., 2022) and other related publications from the Lu lab.

Note:

1. In cases where there is an undesirable band, repeat the agarose gel and recover the DNA as with before and precipitate using standard isopropanol method. A second gel purification may cause too much sample loss, so it is not recommended.

2. Barcoded libraries can be pooled together for sequencing if necessary.

3. A 70 nt single ended sequencing reaction is enough for PARIS. The multiplexing and random barcodes are sequenced together with the insert.

5. Expected outcomes

Successful completion of all reaction and extraction steps will yield information on RNA secondary structures and interactions, including results as follows.

1. Efficient crosslinking in live cells using either amotosalen or SHARC reagents, indicated by obvious signal above the main diagonal in the DD2D gel system.

2. Consistent quantity and quality of RNA extracted via TNA method, indicated by absorbance ratios that fall into the expected ranges.

3. Isolation of XLRNA from non-crosslinked RNA using DD2D gel electrophoresis allows for robust recovery of crosslinked RNAs, regardless of the crosslinker used (Fig. 3).

4. cDNA library generation with size selection for gapped RNA duplexes (Fig. 4).

5. Consistent and high-quality sequencing data generated to study RNA structures and interactions in high throughput.

Fig. 3.

DD2D gel examples for various crosslinker types are provided. (A) Non- and DPI-crosslinked RNA samples. XLRNA region is demarcated by dashed line. (B) Non- and psoralen-crosslinked RNA samples.

Fig. 4.

cDNA library examples. (A) Ideal library size with expected signal above 150 bp and faint empty backbone band. (B) Strong empty backbone (~137 bp) bank and significant signal for only short inserts (~150 bp).

6. Advantages

1. Increased solubility of psoralen-based crosslinking by using amotosalen. Widely used, psoralen-based AMT is soluble up to 1 mg/mL, impacting in vivo crosslinking. Amotosalen is more soluble, up to 230 mg/mL in water.

2. Alternative crosslinking using SHARC reagents targets the unconstrained 2′-OH of RNA, reducing the nucleotide bias as reported in psoralen-based photo-crosslinking.

3. Complete extraction of XLRNA using crosslinker agnostic TNA method. Psoralen-crosslinked RNA was found to repartition into the interphase during standard AGPC (TRIzol) extraction. TNA extraction enables the full recovery of intact crosslinked RNAs.

4. Simplified RNA isolation and fragmentation using DNase I and RNase III enzymes in succession yield pure and short RNAs for isolation and ligation.

5. DD2D isolation of XLRNA enables complete isolation of XLRNA from non-crosslinked RNA (Fig. 2). This method yields high quality XLRNA fragments and avoids complex exonuclease digestion or biotin-tagging approaches.

7. Limitations

1. Psoralen-based analogues would benefit from further optimizations, such as increased solubility and faster reaction time. Moreover, the uridine bias of AMT and amotosalen could be addressed to capture previously uncrosslinkable duplexes.

2. The kinetics of SHARC reactive species are unknown, despite parallel in vitro and in vivo studies. The two-step acylation and hydrolysis steps are difficult to characterize. Orthogonal studies would benefit the understanding of crosslinking chemistry.

3. Acylation-based crosslinkers are applicable to any nucleotide base, albeit limited to flexible ones. In highly structured species, the number of flexible sites is reduced, and thus not capturable by SHARC reagents. With such constraints, a different functional group of RNAs could be a prime target, both reducing nucleotide bias and circumventing the flexibility limitations.

4. RNase R trimming in the exonuclease step is still relatively inefficient due to monoadduct products generated during incomplete crosslinking. As a result, only low concentrations of the SHARC reagent can be used for structure analysis. Addressing the trimming stop sites experimentally would further increase the accuracy and resolution of the structure modeling.

5. The proximity ligation step is still inefficient, producing ~10% total gapped reads. Overcoming this limitation would significantly increase the number of usable reads and improve the sensitivity of crosslink-ligate methods.

8. Optimization and troubleshooting

Problem	Solution
Abnormal spectrophotometer absorbance ratios	A260/A230 < 1.9: Phenol or other chemical contamination. Purify samples again using standard alcohol precipitation.
	A260/A280 < 1.7: Low quantity or quality of nucleic acid, carefully check starting material concentration before each step. Purify samples again using standard alcohol precipitation.
Small RNA pellet after RNase fragmentation or purification	The insoluble material formed as a byproduct of fragmentation will sometimes stick to the inner wall of the tubes. Scrape down with a pipette tip and re-centrifuge to ensure consistent sized pellets.
Abnormal signals in the DD2D gel system	Weak above 80 nt, strong below 80 nt in 1D gel: Excessive RNase III fragmentation yields shorter fragments. Re-measure RNA concentrations after DNase I digestion and ensure 10 μg RNA is used for RNase III treatment. Less than 10 μg RNA may lead to over-digestion (Fig. 5A). Weak below 80 nt, very strong above 80 nt in 1D gel: Insufficient RNase III fragmentation yields longer fragments. Re-measure RNA concentrations after DNase I digestion and ensure 10 μg RNA is used for RNase III treatment. More than 10 μg RNA may lead to underdigestion (Fig. 5B).
	No or poor separation from upper diagonal in 2D gel: Ensure that the sample is correctly crosslinked with psoralen + UV or SHARC reagent (Fig. 5C).
	Fault line in the middle of the 2D gel (esp. for SHARC experiments): Ensure the running buffer is pre-heated at the correct temperature range. Check buffer temperature after a half hour of running. For too high or low temperatures, change out the buffer with fresh pre-heated buffer before continuing the gel running (Fig. 5D).
Abnormal concentrations after DD2D recovery	High: Include a control, non-crosslinked sample in the 2D gel step to distinguish and correctly excise only the crosslinked region.
	Low: Repeat the DD2D gel and pool samples before quantification.
Small RNA pellet after DD2D purification and recovery	Add 3 μL GlycoBlue to RNA solution/emulsion, vortex, and centrifuge again. Using 1.5-mL tubes may enhance visibility.
Poor cDNA library quality (poor yield, incorrect sizes, etc.)	Poor final yield: Having enough starting material is crucial. If there is an insufficient amount of RNA after DD2D purification, collect more prior to library preparation.
	Strong band at ~137 bp: The ~137 bp empty PCR backbone band is expected for normal libraries, but a prominent signal at ~137 bp is indicative of low library starting material (Fig. 4). Start with > 100 ng for library preparation, only use less if working with limited starting material.
	Strong band ~150 bp: The ~150 bp band is expected in normal libraries, but this size range contains short, fragmented inserts and may be contaminated by the empty backbone. Run the agarose gel for longer to ensure better separation. Choose to use a longer or higher concentration agarose gel. Refer to manufacturer manuals to determine the highest concentration of gel that can be processed.
	Incorrect size fragments: SPRI bead size selection may be incorrect. Test beads with a 10 or 50 bp DNA marker prior to use with library samples.
	Incorrect sizes (too small): Pay close attention to enzymatic digestion steps (i.e., RNase III and RNase R). Once the reaction is complete, immediately quench.

Fig. 5.

Failed DD2D gel examples. (A) Over-digestion with RNase III leads to degradation and fragmentation below expected sizes. (B) Under-digestion with RNase III yields insufficiently fragmented samples, leading to poor separation of crosslinked samples. (C) Poor crosslinking prior to the 2D gel system results in poor or no distinct crosslink arc signal above the main diagonal. (D) Excessive heating during the 2D gel running causes the gel to melt and creates a fault line.

9. Summary

In this chapter, we have described crosslink-ligation methods that capture base-paired and/or spatially interacting RNA regions in complex biological systems. Using either the psoralen analogue amotosalen, or acylation based SHARC reagent DPI, this method is an efficient approach to understanding the structures and interactions of RNA. Amotosalen has markedly improved solubility and reactivity, while SHARC DPI bypasses the nucleotide bias of psoralens by targeting the 2′-OH in in RNA. By incorporating further optimizations for the extraction, isolation, and capture of crosslinked RNA duplexes, this method has become a standardized and robust approach in our studies of RNA structure and interactions.

Abbreviations

CDI

1,1′-carbonyldiimidazole

DD2D

denatured-denatured 2D ge

DPI

dipicolinic acid imidazolide

NFW

nuclease free water

PARIS

psoralen analysis of RNA interactions and structures

SHARC

spatial 2′-hydroxyl acylation reversible crosslinker/crosslinking

TNA

total nucleic acid

XLRNA

crosslinked ribonucleic acid