RIPiT-Seq: A tandem immunoprecipitation approach to reveal global binding landscape of multisubunit ribonucleoproteins

Abstract

RNA-binding proteins (RBPs) regulate all aspects of RNA metabolism. The ability to identify RNA targets bound by RBPs is critical for understanding RBP function. While powerful techniques are available to identify binding sites of individual RBPs at high resolution, it remains challenging to unravel binding sites of multicomponent ribonucleoproteins (RNPs) where multiple RBPs or proteins function cooperatively to bind to target RNAs. To fill this gap, we have previously developed RNA Immunoprecipitation in Tandem followed by high-throughput sequencing (RIPiT-seq) to characterize RNA targets of compositionally distinct RNP complexes by sequentially immunoprecipitating two proteins from the same RNP and sequencing the co-purifying RNA footprints. Here, we provide an updated and improved protocol for RIPiT-seq. In this protocol, we have used CRISPR-Cas9 to introduce affinity tag to endogenous protein of interest to capture a more representative state of an RNP complex. We present a modified protocol for library preparation for high-throughput sequencing so that it exclusively uses equipment and reagents available in a standard molecular biology lab. This updated custom library preparation protocol is compatible with commercial PCR multiplexing systems for Illumina sequencing platform for simultaneous and cost-effective analysis of large number of samples.

1. Introduction

All steps in RNA metabolism—from RNA synthesis to RNA degradation—are mediated by proteins that bind RNA to control its biogenesis and function (reviewed in Hentze, Castello, Schwarzl, & Preiss, 2018; Müller-McNicoll & Neugebauer, 2013; Singh, Pratt, Yeo, & Moore, 2015). Such proteins include RNA-binding proteins (RBPs) and their interacting proteins, which often act in concert within multisubunit ribonucleoprotein (RNP) complexes. Identification of RNAs bound to an individual RBP or within a multisubunit RNP provides insight into functions of RBPs, RNPs, and RNAs themselves. Thus, methods for identifying RBP/RNP cargo RNAs serve as an important tool to understand gene regulation.

The past decade has seen tremendous progress in identifying RNA targets of individual RBPs via coupling of ultra-violet (UV) light Cross-Linking and Immunoprecipitation (CLIP) with high-throughput sequencing (CLIP-seq) (reviewed in Lee & Ule, 2018). In CLIP-seq, RNAs are UV-crosslinked to their bound RBPs and RNA segments bound by a specific RBP are recovered after stringent immunoprecipitation (IP) ofthe crosslinked RBP-RNA complex. These RNA segments are then converted into DNA libraries for identification via high-throughput sequencing. Depending on the library preparation strategy employed, some CLIP-seq variants can identify the RBP crosslinking site on RNA, and hence reveal RBP binding site at nucleotide resolution. However, CLIP-seq is disadvantageous when used for certain RBPs or RBP-associated factors which poorly crosslink to RNA with UV-light (Patton et al., 2020; reviewed in Wheeler, Nostrand, & Yeo, 2018). In addition, when studying multisubunit RNP complexes (e.g., cleavage and polyadenylation specificity factor (CPSF) complex involved in mRNA 3′ processing, multifactor mRNA decapping complex), CLIP-seq can only identify targets of individual RBP but not the RNP complex, thus might lose the valuable information about the sites where two or more proteins are potentially acting in a synergistic manner (Patton et al., 2020).

Our lab studies exon junction complex (EJC), an RNA-binding protein complex that is deposited on mRNA exon-exon junctions during pre-mRNA splicing by the spliceosome (reviewed in Boehm & Gehring, 2016; Hir, Saulière, & Wang, 2016; Woodward, Mabin, Gangras, & Singh, 2017). EJC is composed of a trimeric core and more than a dozen peripheral factors. The EJC composition is heterogeneous due to incorporation of distinct peripheral factors that carry out different functions (Mabin et al., 2018; Wang, Ballut, Barbosa, & Le Hir, 2018). To study the RNA substrates of these individual EJC compositions, we have previously developed and employed an approach termed RNA immunoprecipitation in tandem (RIPiT) followed by high-throughput sequencing (RIPiT-seq) to identify the RNA targets of such multisubunit RNP complexes (Mabin et al., 2018; Singh et al., 2012). In the RIPiT procedure, one subunit of an RNP complex is first immunoprecipitated via an affinity tag, which enables non-denaturing elution of the RNPs containing this protein by preserving protein-protein associations. During the first IP, the RNPs are also treated with RNases to digest away RNAs that are not directly bound within RNPs. The RNase-digested RNPs thus obtained are then subjected to second IP of another subunit of the RNP complex. Following the second IP, the RNA footprints are isolated and converted into cDNA libraries for high-throughput sequencing. RIPiT-seq enables identification of RNA targets of compositionally distinct EJCs and it should be applicable for identification of RNA targets of other multisubunit RNP complexes in general (Fig. 1). While RIPiT-seq is less technically challenging as compared to CLIP-seq and can identify RNA targets of an RNP complex, it is important to note that it likely has reduced specificity as compared to CLIP-seq due to substantial carryover of contaminating RNAs (e.g., ribosomal RNA) during the nondenaturing immunoprecipitations.

Fig. 1.

An overview of the RIPiT scheme. The main steps in the RIPiT procedure following cell lysis are shown. From total extract, RNP complex is first subjected to immunoprecipitation of the tagged protein (shown here to be FLAG-tagged). The immunoprecipitated complexes are then subject to RNase I digestion while still bound to FLAG-affinity resin. The RNP complexes are then eluted under nondenaturing condition with the 3 × FLAG peptide and subject to a second immunoprecipitation using primary antibody targeting second protein of the RNP complex. RNAs are then eluted from the protein complex and converted into cDNA libraries for high-throughput sequencing. RNA bound by a specific composition in an otherwise heterogeneous RNP complex can be determined by choosing different pairs of proteins for the two immunoprecipitations.

Here, we describe an updated RIPiT-seq method, where we use CRISPR-Cas9 to introduce an affinity tag at the endogenous gene locus thereby reflecting the RNA cargoes of an RNP complex more accurately than via overexpression of a tagged protein. We also present an updated kit-free RNA library preparation method that, compared to its previous versions (Gangras, Dayeh, Mabin, Nakanishi, & Singh, 2018; Heyer, Ozadam, Ricci, Cenik, & Moore, 2015; Sterling, Veksler-Lublinsky, & Ambros, 2015; Woodward, Gangras, & Singh, 2019), is more generally accessible as it uses common lab equipment and reagents, and is compatible with commercial Illumina indexing primers, which enables high-level multiplexing at lower cost.

2. Before you begin

The RIPiT-seq procedure described here uses HCT 116 cells, a near-diploid colorectal cell line that is amenable to CRISPR-Cas9-mediated engineering of an affinity tag at endogenous gene locus and can be grown in medium-scale (1–3 × 10⁹ cells) cultures in standard tissue culture dishes. Other cell lines satisfying both these criteria can also be used for RIPiT-seq. Alternatively, affinity-tagged proteins can be expressed at near-endogenous levels from exogenously introduced DNA copies (e.g., we have successfully used TRex 293 cells and the Flp-In system for Tetracycline-inducible expression of affinity-tagged proteins for RIPiT-seq (Mabin et al., 2018)).

For the first IP in RIPiT, we have used FLAG tag, a short eight amino-acid tag for which a high-affinity antibody resin and competitor peptide for gentle elution are commercially available for scalable affinity purifications. The short nature of this tag allows it to be easily introduced into an endogenous gene locus via genome engineering to tag a polypeptide at either terminus and minimizes any chance of interference of protein function by the tag. Numerous other small affinity tags (e.g., HA tag, 3 × FLAG, Strep tag II) can also be similarly used in the procedure with minor modifications.

All home-made buffers can be prepared ahead of time. Components that should be added fresh just before use are noted. All buffers should be prepared with stocks that have been sterilized by autoclaving unless otherwise noted. One exception is Milli-Q water used to make large volume of isotonic wash buffer (IsoWB) used during RIPiT. While we have successfully carried out RIPiT using unautoclaved Milli-Q water to make IsoWB; to minimize concerns of RNase contamination autoclaved (and DEPC-treated, if necessary) water can be used.

For all library preparation steps, we recommend using pre-sterilized filter tips.

3. Materials and equipment

3.1. Equipment

• Gene Pulser Xcell Electroporation Systems (or another electroporation system for mammalian cells) (Bio-Rad #1652660)

• Sonifier Cell Disruptor (e.g., Branson benchtop sonifier SFX250)

• Blue Light Transilluminator

• Automated Cell Counter (e.g., Countess II Automated Cell Counters) or a Hemocytometer

• Magnetic Rack Separator (Sergi Lab Supplies)

• Gene Pulser Electroporation Cuvettes, 0.2cm gap (Bio-Rad #1652082) or similar products

3.1.1. Cell line

• HCT 116, Human Colorectal Carcinoma (ATCC CCL-247)

3.1.2. Reagents

• 5′ DNA Adenylation Kit (NEB #E2610S)

• T7 Endonuclease I (NEB #M0302S)

• 3 × FLAG peptide (APExBIO #A6001)

• Protease Inhibitor Cocktail (APExBIO #K1007)

• Phosphatase Inhibitor Cocktail (APExBIO #K1015)

• Recombinant Cas9 (QB3 MacroLab, UC Berkeley)

• Aphidicolin (Fisher Scientific #AC611970010)

• ANTI-FLAG M2 Affinity Gel (Sigma #A2220)

• RNase I (Lucigen #N6901K)

• EIF4A3 Antibody (Bethyl Laboratories #A302–980A)

• SYBR Gold Nucleic Acid Gel Stain (Thermo Fisher Scientific #S11494)

• 10% TBE-Urea gel (Bio-Rad #4566036) (or equivalent product from other vendors)

• Maxima H Minus Reverse Transcriptase (Thermo Fisher Scientific #EP0753)

• McCoy’s 5A (Modified) Medium (Fisher Scientific #16-600-108)

• Fetal Bovine Serum (Sigma-Aldrich #F2442)

• Penicillin-Streptomycin (10,000U/mL) (Fisher Scientific #15-140-122)

• Trypsin-EDTA (0.05%), phenol red (Fisher #25300–062)

• Ingenio Electroporation Solution, or another mammalian electroporation solution (Mirus Bio #MIR50114)

• Dynabeads Protein A (Thermo Fisher Scientific #10002D)

• Dynabeads Protein G (Thermo Fisher Scientific #10004D)

• 5PRIME Phase Lock Gel (QuantaBio #2302820) or similar product such as Corning high-vacuum silicone grease (Sigma-Aldrich #Z273554)

• Quick CIP (NEB #M0525)

• T4 RNA Ligase 2, truncated K227Q (NEB #M0351L)

• T4 RNA Ligase Buffer (NEB #B0216L)

• Biotin-11-dATP (AAT Bioquest #17014 or PerkinElmer #NEL540001EA)

• Biotin-16-Aminoallyl-2′-dCTP (TriLink BioTechnologies #N-5002-1)

• Low Molecular Weight DNA Ladder (NEB #N3233S) or similar DNA ladder

• RNasin Plus Ribonuclease Inhibitor (Promega #N2615)

• Hydrophilic Streptavidin Magnetic Beads (NEB #S1421S)

• CircLigase ssDNA Ligase and CircLigase 10 × Reaction Buffer (Lucigen #CL4115K)

• HighPrep PCR beads (MagBio #AC-60005) or similar product such as NucleoMag (Macherey-Nagel #744970.5) and AMPure XP (Beckman Coulter #A63880)

• NEBNext Q5 Master Mix (NEB #M0544S)

• 5M Betaine solution (Sigma-Aldrich #B0300)

• SYBR Green I Nucleic Acid Gel Stain, 10,000 × concentrate (Thermo Fisher Scientific #S7563)

3.1.3. Chemicals

• Phenol/Chloroform/Isoamyl Alcohol 125:24:1, pH 4.3 (PCIA pH 4.3; Fisher Scientific #BP1754I-400)

• Phenol/Chloroform/Isoamyl Alcohol 25:24:1, pH 8.0 (PCIA pH 8.0; Fisher Scientific #BP1752I-100)

• 50% PEG8000

• Deionized Formamide (Sigma-Aldrich #4650-500ML)

• Cycloheximide (Sigma-Aldrich #C7698-1G)

3.1.4. Home-made buffers

• 1 × Phosphate Buffered Saline (PBS)

  137mM NaCl; 2.7mM KCl; 10mM Na₂HPO₄; 1.8mM KH₂PO₄

• Hypotonic Lysis buffer

  20mM Tris-HCl pH 7.5; 15mM NaCl; 10mM EDTA (Substitute with 1mM MgCl₂ if interactions are Mg dependent); 0.1% Triton X-100; 1 × Protease Inhibitor Cocktails*; (optional) 1 × Phosphatase Inhibitor Cocktail* (* add fresh every time)

• Isotonic Wash Buffer (IsoWB)

  20mM Tris-HCl pH 7.5; 150mM NaCl; 0.1% IGEPAL CA-630

• 2 × Dilution Buffer

  20mM Tris-HCl pH 7.5; 150mM NaCl; 0.2% Triton X-100; 10mM EDTA; 0.2mg/ml BSA*

  2 × Protease Inhibitor Cocktail*; (optional) 2 × Phosphatase Inhibitor Cocktail* (* add fresh every time)

• Conjugation Buffer

  1 × PBS; 0.01% Tween-20

• Clear Sample Buffer

  100mM Tris-HCl pH 6.8; 4% SDS; 10mM EDTA

• 5 × First strand (FS) w/o MgCl₂ Buffer

  250mM Tris-HCl pH 8.3; 375mM KCl

• 1 × Urea Load Buffer

  1 × TBE; 12% ficoll; 7M Urea; 0.01% bromophenol blue; 0.02% xylene cyanole FF

• 100mg/mL cycloheximide stock

• DNA Elution Buffer:

  300mM NaCl; 1mM EDTA

• 50 × Denhardt’s solution:

  1% Ficoll 400; 1% Polyvinylpyrrolidone (PVP); 1% Bovine serum albumin (Fraction V)

• Strep Bead Wash Buffer:

  10mM Tris-HCl pH 7.5; 1mM EDTA; 0.3M NaCl; 0.1% Tween-20

• Strep Bead Resuspension Buffer:

  10mM Tris-HCl pH 7.5; 0.1mM EDTA; 0.3M NaCl

• 1mM biotin-dNTPs

  0.25mM dGTP; 0.25mM dTTP; 0.175mM dATP; 0.075mM biotin-dATP; 0.1625mM dCTP; 0.0875mM biotin-dCTP

3.1.5. Oligos

• sgRNA or cr:tracrRNA: Guide RNA can be designed using various platforms (e.g., https://benchling.com/). Guide RNA cut site should be as close to the tag insertion site as possible. Synthetic guides can be ordered from multiple vendors such as Synthego and IDT. If necessary, editing efficiency of individual guide can be evaluated from T7 Endonuclease I assay according to the manufacturer protocol.

• Single-stranded oligo donor (ssODN): ssODN contains an affinity tag sequence flanked by 35–50nt homology arm (Fig. 2A). We include a short GGGS linker following the tag sequence that also contains BamHI (encodes GS) site for the screening purpose.

• mirCat33 adapter: /rApp/TGGAATTCTCGGGTGCCAAGG/ddC/. If adenylated adapter is not commercially available, adenylation can be performed using 5′ DNA Adenylation Kit following the manufacturer protocol. We have reduced RNA:enzyme ratio down to 4:1 without any obvious decrease in adenylation efficiency. The adenylated adapter can be purified with silica column or with Urea-PAGE gel.

• RT Primer: /5Phos/GGNNNNNNNNAGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT/iSp18/GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCTCCTTGGCACCCGAGAATTCCA: Underlined is an 8-mer unique molecular identifier (UMI) used for removing PCR duplicates. RT primer is ordered via IDT as DNA oligo with HPLC purification.

• I5_uni: AATGATACGGCGACCACCGAGATCTACACACACTCTTTCCCTACACGACGCTCTTCCGAT

• C*T (* Phosphorothioate Bond). Order from IDT as desalted Ultramer.

• I7_uni: CAAGCAGAAGACGGCATACGAGATGTGACTGGAGTTCAGACGTGTGCTCTTCCGATC*T (* Phosphorothioate Bond). Order from IDT as desalted Ultramer.

• Commercial Illumina I5/I7 Indexing Primers: I5/I7_uni primers do not contain barcodes and are used in qPCR for the cycle number determination. For final PCR to prepare library for high-throughput sequencing, commercial Illumina indexing primer set can be ordered from vendors such as Lexogen or IDT. Alternatively, individual indexing primers can be ordered as desalted Ultramers from IDT.

Fig. 2.

CRISPR-Cas9-mediated insertion of affinity tag and bulk assessment of tag insertion. (A) Scheme of CRISPR-Cas9-mediated insertion of affinity tag (FLAG) into endogenous locus of a gene. Single-stranded oligo donor (ssODN) for homology-directed repair is on the top where FLAG tag sequence and a flexible linker (amino acid sequence GGGS) is flanked by 35–50nt DNA sequence with perfect homology with the DNA sequence surrounding the cut site (triangle) induced by guide RNA-Cas9 RNP complex. Primers (labeled Fwd and Rev) flanking the insertion site are used in the PCR reaction with genomic DNA as a template to identify the insertion event. (B) A bulk-level (i.e., in a pool of cells) analysis of CRISPR-mediated insertion efficiency by amplifying the insertion site with PCR 48–72h after guide RNA-Cas9 RNP complex electroporation. With a 20–30% insertion efficiency and relatively small amplicon, there should be a clear shift of amplicon containing inserted tag (two replicates shown in lanes 2 and 3, labeled tagged locus) as compared to wild-type endogenous locus (single band in lane 1).

4. Step-by-step method details

4.1. Affinity tag knock-in using CRISPR-Cas9

1. 1 day before CRISPR-Cas9 ribonucleoprotein (RNP) electroporation, seed approximately 2.5 × 10⁶ HCT 116 cells in a 10-cm plate in growth media (McCoy’s 5A (Modified) Medium, 10% FBS, 1% Penicillin-Streptomycin) supplemented with 2μg/mL aphidicolin for cell cycle synchronization (Lin, Staahl, Alla, & Doudna, 2014; Rivera-Torres & Kmiec, 2017; Rivera-Torres, Strouse, Bialk, Niamat, & Kmiec, 2014). Grow cells overnight under standard growth conditions (5% CO₂ and 37°C). Note: Cell synchronization, as show in the references indicated earlier as well as in our experience, strongly enhances the HDR efficiency.

2. 4h before electroporation, replace media with fresh growth media without aphidicolin to release the cells from aphidicolin-block and enter S phase (Rivera-Torres & Kmiec, 2017).

3. Prepare RNP complex as described below. RNP complex should be at least 1μM in the 50μL final electroporation reaction.

   Cas9	50pmol
   sgRNA	50–100pmol (Cas9:guide RNA ratio can be between 1:1 and 1:2)

   Add Ingenio electroporation solution to a total volume of 20μL. Mix well and incubate at room temperature (RT) for ~20min.

4. To harvest cells for electroporation, wash cells with 10mL of 1 × PBS. Add 1mL of Trypsin-EDTA and incubate at 37°C for 5min to dissociate cells. Resuspend cells in 9mL of growth media. Use a small aliquot to count cells using a cell counter to estimate cell concentration. Aliquot into 1.5mL centrifuge tubes appropriate volume of cell suspension to get 2.5 × 10⁵ cells per reaction. Harvest cells at 500 × g for 3min.

5. To wash cells with 1 × PBS, resuspend cell pellet in 1mL 1 × PBS and centrifuge again at 500 × g for 3min.

6. Resuspend cells in 28.5μL Ingenio electroporation solution. Add 1.5μL (150pmol) of ssODN. Mix resuspended cells with RNP complex from step 3 to a final volume of 50μL.

7. Transfer the 50μL cell-RNP mix to a 0.2cm electroporation cuvette. For HCT 116 cells, use the following setting for electroporation: 120V, 13ms/per pulse; 2 pulses with 1s interval.

8. Transfer cells to a 6-well plate containing 2mL growth media and incubate under standard growth conditions.

9. 48–72h later, dissociate cells by trypsinization as in step 4 above, and resuspend in 2mL growth media. Count cells, dilute cells to either achieve 5–10 cell/mL to seed in 96-well plates or seed diluted cells (5–15 cells/mL) in 10-cm plates to isolate single colonies. Pellet remaining cells to isolate genomic DNA and analyze homology-directed repair (HDR) efficiency via PCR (Fig. 2). Note: To confirm efficient insertion of affinity tag into an endogenous gene locus, a bulk-level screening of CRISPR-Cas9 knock-in efficiency can be performed on the RNP-transfected cells. This can be achieved using a pair of primers flanking the insertion site, as shown in Fig. 2. Template DNA for this genotyping PCR can be obtained via a direct lysis protocol described here (Ramlee, Yan, Cheung, Chuah, & Li, 2015), skipping the need to isolate genomic DNA. After 2–3 weeks, pick single colonies (number of colonies to pick will depend on the HDR efficiency) and screen cells for correct insertion. Individual colonies are screened using the same PCR primers in the previous step to identify the cell clones that harbors the tag insertion. Sanger sequencing is required to validate the correct insertion of the tags.

4.2. RNA immunoprecipitation in tandem

1. Grow HCT 116 cells expressing affinity-tagged protein of interest in 15-cm plates under standard growth conditions. (To ensure enough RNA is recovered, at least four 15-cm plates with around 70% cell confluency are needed per RIPiT for an abundant RNP like EJC.)

2. 1h prior to harvesting cells, add cycloheximide (CHX) to 100μg/mL final concentration in the media (prepare fresh 100mg/mL CHX stock each time). Note: This step allows accumulation of untranslated mRNAs and hence the EJC-bound pool of mRNAs over the 1-h period. It can be omitted if RNP of interest is not impacted by the translating ribosome.

3. Rinse cells once with 25mL of 1 × PBS (with 100μg/mL CHX) and harvest cells in 10mL 1 × PBS (with 100μg/mL CHX) using a cell scrapper. Pellet cells at 1000 × g for 5min at 4°C.

4. Resuspend cells in 4mL ice-cold gentle hypotonic lysis buffer. To solubilize chromatin associated RNPs, sonicate lysate using Branson tabletop sonifier at 15% amplitude for 30s in burst of 2s with 3-s intervals. Note: This step is optional and can be omitted if RNPs of interest are primarily in soluble nucleoplasm or cytoplasm.

5. Add NaCl to 150mM. Keep on ice for 5min. Spin at 15000 × g, 4°C, 10min. Note: To increase stringency, higher salt concentrations can be used as long as the RNP of interest remains intact at these elevated ionic conditions.

6. Save 40μL lysate supernatant as input. Store at −20°C. Rest of the lysate will be used for FLAG immunoprecipitation.

7. Wash anti-FLAG Affinity Gel three times with 1mL IsoWB. Note: 40μL of the affinity gel (20μL beads volume) is usually enough to deplete endogenous FLAG-tagged protein from total extract prepared from one 15-cm plate. 200μL affinity gel is usually sufficient for 4–6 15-cm plates.

8. Add cleared lysate from step 6 to the washed FLAG affinity gel. Nutate at 4°C for 1h.

9. Pellet FLAG affinity beads at 1000 × g, 1min at 4°C. Save 40μL of supernatant to later check efficiency of depletion of the target protein. Store at −20°C.

10. Wash beads four times with 1mL IsoWB (ice-cold).

11. After fourth wash, add 200–300μL of IsoWB supplemented with RNase I at concentration of 0.01–0.1 unit/μL. Gently vortex at 4°C for 15min.

12. Wash beads four times with 1mL ice-cold IsoWB.

13. Elute FLAG-tagged protein containing RNPs with at least one bed volume (100μL) of elution buffer containing 3 × FLAG peptide at 500μg/mL in IsoWB. Gently shake tube contents at 4°C for 30min. Repeat the elution twice and combine both the elutions. Save 20μL of elution to test for efficiency of FLAG IP. Note: 4°C elution is important to preserve protein-protein association.

14. During incubation period for the FLAG elution, prepare antibody-protein A/G DynaBeads conjugates for the second IP. Pre-wash 25μL protein A or protein G DynaBeads three times with conjugation buffer and resuspend the beads in 1mL conjugation buffer. Add appropriate amount of the primary antibody against the second protein of the RNP. For EIF4A3 (EJC protein), 5μg of antibody is enough for one RIPiT from 4 to 6 15cm plates. Gently vortex for 30min at RT. Wash antibody conjugated beads three times with 1mL IsoWB.

15. Add 1 volume of 2× Dilution Buffer to the FLAG elution and mix the elution with protein A/G conjugated primary antibody for the second protein target.

16. Nutate at 4°C for 1h.

17. Wash beads six times with 1mL ice-cold IsoWB.

18. To elute RNPs, add 20μL Clear Sample Buffer and incubate at 50°C for 5min. Save 2μL and store at −20°C for western blot (Fig. 3).

19. To extract RNA from the rest of the elution, add 80μL water and 100μL of Phenol/Chloroform/Isoamyl Alcohol (PCIA) (125:24:1; pH 4.3), mix vigorously. Centrifuge at 15000 × g for 5min at RT.

20. Transfer 100μL aqueous phase to a new tube, add 1/10 volume of 3M sodium acetate pH 5.2, 1/100 volume of 1M MgCl₂, 2μL of 5mg/mL glycogen, and 3 volumes of cold 100% ethanol. Mix well and incubate at −80°C for 1h or −20°C overnight. Note: Using vacuum grease or commercial phase locking tubes is helpful in physically separating the organic phase to facilitate the transfer of aqueous phase for every phenol chloroform cleanup step hereafter.

21. Centrifuge for 30min at 15000 × g at 4°C. Discard supernatant.

22. Wash pellet with ice old 70% ethanol and centrifuge for 5min at 15000 × g at 4°C. Discard supernatant.

23. Air dry RNA pellet and resuspend RNA in 17μL RNase-free water.

24. Dephosphorylate resuspended RNA by adding 2μL NEB CutSmart Buffer (10×) and 1μL Quick CIP enzyme. Mix well and incubate at 37°C for 30min.

25. To the dephosphorylated RNA, add 80μL water and 100μL of Phenol/Chloroform/Isoamyl Alcohol (PCIA) (125:24:1; pH 4.3), mix vigorously. Centrifuge at 15000 × g for 5min.

26. Transfer 100μL aqueous phase to a new tube, add 1/10 volume of 3M sodium acetate pH 5.2, 1/100 volume of 1M MgCl₂, 2μL of 5mg/mL glycogen, and 3 volumes of cold 100% ethanol. Mix well and incubate at −80°C for 1h or −20°C overnight.

27. Centrifuge for 30min at 15000 × g at 4°C. Discard supernatant.

28. Wash pellet with ice old 70% ethanol and centrifuge for 5min at 15000 × g at 4°C. Discard supernatant.

29. Air dry RNA pellet and resuspend RNA in 5μL water.

30. Optional: To visualize length distribution of immunoprecipitated RNA fragments, and to quantify their amount, 1μL of RNA can be end labeled with [γ–32P] ATP using T4 polynucleotide kinase and visualized on high resolution (20–26% polyacrylamide) gel (Fig. 3).

Fig. 3.

Visualization of proteins and RNAs enriched via RIPiT. (A) An example of RIPiT validation via western blots. Western blots showing levels of the proteins (on the right) in various fractions labeled on the top (EL, elution; FT, flowthrough). Cell lines expressing the endogenously tagged protein used in RIPiT are indicated above each lane. As seen in the figure, different complexes have different composition, as reflected by different levels of UPF2 protein in the three complexes. (B) An example of RNA footprints from FLAG-CASC3:EIF4A3 containing complexes following 5′-end-labeling with [γ–32P] ATP and resolution on a 26% Urea-PAGE gel. Shown image was spliced at the vertical dotted lines. Size distribution of the RNA footprints can be estimated via comparison with radio-labeled base-hydrolyzed poly-U₄₅ RNA ladder (lane 2) and DNA ladders (lane 4 and 5). A known amount (0.1pmole) of single-stranded RNA oligo (lane 1) is radio-labeled and run on the gel to estimate RIPiT RNA concentration.

4.3. RNA footprints library preparation

1. Set up following reaction in a PCR tube:

   mirCat33 adapter	7 pmol
   RNA (from RIPiT)	4 μL
   Water	to 4.8 μL

2. Incubate in a PCR machine at 65°C 10min; 16°C 5min; 4°C hold.

3. Transfer tubes to ice and add:

   T4 RNA Ligase Buffer	1.5 μL
   50% PEG8000	7.5 μL
   20 mM DTT	0.75 μL
   T4 RNA Ligase 2 (K227Q)	0.45 μL
   Total Volume	15 μL

   Incubate at 30°C for 6h, 65°C for 20min and 4°C hold.
4. Add to the same tube containing 15μL ligation (on ice):

   10 μM RT Primer	1 μL
   1 mM biotin-dNTPs	10 μL

5. Incubate at 65°C 5min followed by 4°C for at least a minute.

6. Add the following reagents to the same tube:

   5 × FS w/o MgCl2 Buffer	9 μL
   100 mM DTT	2.25 μL
   Maxima H-RTase	1.2 μL
   RNasin Plus	1.25 μL
   Water	5.3 μL
   Total volume	45 μL

7. Incubate at 55°C for 45min; inactivate RTase at 70 °C 15min; hold reaction at 4°C if needed.

8. Transfer the reaction to 1.5mL Eppendorf tube. Add 155μL water and 200μL PCIA (25:24:1; pH 8.0) mix vigorously. Centrifuge at 15000 × g for 5min.

9. Transfer 200μL aqueous phase to a new tube, add 1/10 volume of 3M sodium acetate pH 5.2, 1/100 volume of 1M MgCl₂, 2μL of 5mg/mL glycogen, and 3 volumes of cold 100% ethanol. Mix well and incubate at −20°C for 15min.

10. Centrifuge for 15min at 15000 × g at 4°C. Discard supernatant.

11. Wash pellet with ice old 70% ethanol and centrifuge for 5min at 15000 × g at 4°C. Discard supernatant.

12. Air dry and resuspend cDNA pellet in 10μL 1× Urea Loading Dye. Denature at 65°C for 5min before loading on a 10% TBE-Urea gel. Run at 200V for 40min. For size markers, load Low Molecular Weight DNA Ladder and 10pmol of RT primer along with the cDNA samples.

13. Stain gel with SYBR gold following the manufacturer guidelines.

14. Visualize gel on a blue light illuminator. Excise gel segments corresponding to 130–200nt (or application-dependent) long reverse transcribed products (Fig. 4). Elute with the PAGE elution buffer at 30°C overnight (300–600μL). To maximize recovery, elution from gel pieces can be done twice and the two elutions can be combined.

15. Aliquot 10μL of streptavidin beads for each sample. Wash three times with 0.1M NaOH.

16. Wash twice with Strep Bead Wash Buffer.

17. Wash once with Strep Bead Resuspension Buffer.

18. Block beads with 30μL of Strep Bead Resuspension Buffer + 3μL 50 × Denhardt’s solution. Incubate at 30°C for 20min with gentle shaking.

19. Wash beads twice with Strep Bead Wash Buffer.

20. Add cDNA elution to the beads, incubate at 30°C for 30min with gentle shaking.

21. Wash beads twice with Strep Bead Wash Buffer and once with Strep Bead Resuspension Buffer.

22. Add 5μL water to resuspend the beads and proceed with cDNA circularization.

23. To resuspended streptavidin beads (volume will be ~5.5μL), add the following:

    CircLigase 10 × Reaction Buffer	1 μL
    1 mM ATP	0.5 μL
    50 mM MnCl2	0.5 μL
    5M betaine	2 μL
    CircLigase	0.5 μL
    Total Volume	10 μL

24. Incubate 60°C for 3h followed by head inactivation at 80°C for 10min.

25. Wash beads with twice with Strep Bead Wash Buffer and twice with Strep Bead Resuspension Buffer.

26. Elute circularized cDNA in 15μL 95% deionized formamide + 10mM EDTA at 65°C for 5min and 90°C for 2min.

27. Add 85μL water and 100μL PCIA (25:24:1; pH 8.0), mix vigorously. Centrifuge at 15000 × g for 5min.

28. Transfer 100μL aqueous phase to a new tube, add 1/10 volume of 3M sodium acetate pH 5.2, 1/100 volume of 1M MgCl₂, 2μL of 5mg/mL glycogen, and 3 volumes of cold 100% ethanol. Mix well and incubate at −20°C for 15min.

29. Centrifuge for 15min at 15000 × g at 4°C. Discard supernatant.

30. Wash pellet with ice old 70% ethanol and centrifuge for 5min at 15000 × g at 4°C. Discard supernatant.

31. Air dry and resuspend cDNA pellet in 11μL water.

32. Use 1μL of cDNA for 30μL qPCR. Set up a parallel reaction using 1μL water as control:

    NEBNext Q5 Master Mix	15 μL
    I5_uni (20 μM)	1.5 μL
    I7_uni (20 μM)	1.5 μL
    2.5 × SYBR Green I	1.2 μL
    cDNA	1 μL
    Water	9.8 μL

33. Run the following qPCR program:

    Initial Denaturation	98 °C	30s
    Denaturation	98 °C	10s
    Annealing/Extension	65 °C	75s × 35 cycles
    Final Extension	65 °C	5 min

34. To determine cycle number for final PCR, identify cycle number required for qPCR to reach 50% of the maximum fluorescence value, then subtract 3 cycles and this will be used for final PCR using the rest 10μL cDNA.

35. Set up 30μL final PCR reaction:

    NEBNext Q5 Master Mix	15 μL
    I5 Indexing Primers (20 μM)	1.5 μL
    I7 Indexing Primers (20 μM)	1.5 μL
    cDNA	10 μL
    Water	2 μL

36. Run the following qPCR program:

    Initial denaturation	98 °C	30s
    Denaturation	98 °C	10s
    Annealing/extension	65 °C	75s × Cycle number in Step 34
    Final extension	65 °C	5 min
    Hold	4 °C	∞

37. Add 36μL (1.2×) HighPrep PCR beads or similar product. Mix thoroughly and incubate at RT for 5min.

38. Place the tube on magnetic rack for ~2min or until beads are separated. Carefully discard supernatant.

39. While still on magnetic rack, wash beads twice with freshly prepared 80% ethanol for 30s each.

40. Discard ethanol and air dry for 2–5min.

41. Add 20μL water to the dried beads. Mix and incubate for 2min.

42. Place the tube on magnetic rack until beads are separated, transfer elution to a new tube.

43. 2μL of the library DNA can be diluted with 2μL water for High Sensitivity-DNA TapeStation Bioanalyzer assay. As PCR products resulting from unextended circularized RT primer is expected to be 167bp (with I5/I7 dual 8bp indices), insert size can be deduced from the Bioanalyzer results. Library molarity of the region containing desired size PCR products as reported by the 2200 TapeStation Software (Agilent) will be used for library pooling.

Fig. 4.

An overview of the main steps in cDNA library prep for high-throughput sequencing. (A) The first step consists of ligation of a pre-adenylated mirCat33 adapter to the dephosphorylated 3′ end of RNA footprints with T4 RNA Ligase 2, truncated K227Q. (B) Ligation product from A is then used for reverse transcription reaction (in the same tube as ligation) using the special RT primer that includes a region complementary to mirCat33 adapter, binding sites for read two primer binding sequence linked via polyethylene glycol linker (Sp18) to read one primer binding sequence, the 8-nucleotide random sequence that serves as unique molecular identifier (UMI) and two Gs at the 5′-end that conform to the preferred base for the CircLigase used in a subsequent step. The RT reaction includes biotinylated dATP and dCTP to specifically label the extended RT products for enrichment in the next step. (C) Left: A schematic of a gel image where RT products (lane 3) are resolved on UREA-PAGE and visualized via SYBR–gold staining along with a low range DNA ladder (lane 1) and unextended RT primer (lane 2). Size-selected cDNAs (marked by dotted box) are eluted from gel, pulled down using streptavidin beads and eluted from beads. Such cDNAs are next subject to circularization. Circularized cDNAs are then ready for PCR amplification with indexing primers. (D) Schematic of the final library product with Illumina I5/I7 dual indexes and a bioanalyzer trace example. Rd1 and Rd2 represent read1 and read2 sequencing primers.

5. Expected outcomes

5.1. Affinity tag knock-in using CRISPR-Cas9

RNP electroporation and ssODN-mediated HDR regularly yields 20–30% knock-in efficiency in our hands. Certain gene loci can present exceptions and show lower knock-in efficiency. In such cases, a selectable marker-based knock-in strategy can be adopted (Park, Won, Pentecost, Bartkowski, & Lee, 2014). Small scale FLAG IP should be performed to validate that the knock-in protein can be successfully immunoprecipitated and that it can co-IP its associated proteins.

5.2. RNA immunoprecipitation in tandem

After RIPiT of the RNP complex, small fractions saved during the procedure (total extract, depleted extract after first IP, immunoprecipitates from first IP, unbound fraction from the second IP and immunoprecipitates from the second IP) should be analyzed by western blotting to validate the immunoprecipitation of the complex (Fig. 3A). Proteins immunoprecipitated during first and second IP should be of highest abundance on western blots (Fig. 3A). Other factors of the same RNP complex should also be readily observed on the western blots along with the immunoprecipitated proteins. A comparison of total extract with depleted extract after the first IP will indicate the efficiency of the first IP step. Similarly, the unbound fraction after the second IP will contain proteins that associate with the target of the first IP but not with the protein immunoprecipitated second.

RNA footprints isolated from RIPiT should contain a complex mixture of RNAs that should appear as a smear on the RNA gel (Fig. 3B) or BioAnalyzer profile. The presence of prominent distinct-sized bands in the RNA profiles will be indicative of high levels of RNAs of distinct size and sequences, often originating from abundant RNAs such as transfer RNAs or ribosomal RNAs. In such cases, specificity of IPs can be increased by more stringent washing (e.g., higher salt in IsoWB) or by formaldehyde crosslinking RNPs prior to cell lysis followed by denaturing FLAG IP (see below). EJC RIPiTs from typical input material noted in the procedure above regularly yield sufficient amount of RNA for subsequent library preparation. Recommend amount of RNA for input into the downstream library preparation is 1pmol, although the library preparation method described above works for inputs as low as 0.1pmol (Gangras et al., 2018).

5.3. RNA footprint library preparation

Following reverse transcription (RT), the extended RT product will appear as a smear above the prominent band at 98nt, which is the unextended RT primer (Fig. 4C). To ensure that final cDNA library has inserts >30nt, which is the lower limit for robust mapping of fragment sequences to the human genome, RT products of lengths >130nt should be size-selected. When input RNA is >1pmol, the final PCR reaction should require 7–12 cycles. However, if RNA input is higher or lower, PCR cycles outside this range should also be fine for the downstream high-throughput sequencing. It is highly recommended to carefully estimate the number of PCR cycles for the final PCR to be in the linear amplification range. However, if over-amplification occurs, the unique molecular identifiers (UMIs) in the RT primer (Fig. 4C) enable removal of PCR duplicates during downstream processing of sequencing reads. Finally, when the above in-house RNA library preparation procedure is used for the first time, we recommend cloning a small amount of the final PCR product into a plasmid vector (e.g., TA cloning vector provided PCR products carry A overhangs) and identifying insert sequences of 6–12 clones via traditional Sanger sequencing to make sure the library products contain expected sequences.

6. Quantification and statistical analysis

Following Illumina high-throughput sequencing, the first step is to extract the UMI sequences. The UMI-tools package (Smith, Heger, & Sudbery, 2017) offers such an extraction command. Alternatively, first eight nucleotides of sequencing reads can be manually extracted using program such as Awk. UMI sequences should be moved to the end of identifier line in the fastQ file for downstream deduplication purpose as indicated in the manual (Smith et al., 2017). Next, fastQ files are subject to adapter trimming. We select only those reads that start with a CC and end with the mirCat33 adapter sequence. Inclusion of reads missing the mirCat33 adapter might mislead the binding site information of RBP as the 3′ end of the RNA fragment will be unknown.

Clean reads thus obtained after adapter trimming can then be aligned to the reference genome. We highly recommend to first map reads to rRNA or other highly abundant non-coding RNAs that are not of interest, and then align the remaining reads to the genome. Following genome-mapping, PCR duplicates can be removed using UMI-tools dedup function (Smith et al., 2017). The genome-mapped reads should reflect binding site of the RNP complex and it will vary based on the complex being pulled down. For the EJC, we observe strong enrichment on the exon-exon junctions when aligned reads are visualized using tools such as Integrated Genome Viewer or UCSC Genome Browser. Similar quality control should be performed based on the RNP of interest that has been immunoprecipitated. As a first step, read densities can be quantified in the expected binding regions of the targeted RNP complex and compared to average read densities across all regions. For further identification and quantification of specific RNP binding sites, peak calling algorithms can be used (e.g., Kucukural, Özadam, Singh, Moore, & Cenik, 2013; Uren et al., 2012).

7. Advantages

RIPiT can be used to identify footprints of RNP complex as well as RBPs that are inefficiently UV-crosslinked to RNA (Patton et al., 2020). The procedure is less technically demanding as compared to CLIP, which also requires the antibody-target protein interaction to withstand high salt and denaturing detergents. The newly updated RNA footprint library preparation method described here allows a much more streamlined, accessible and cost-effective method to prepare cDNA libraries from small RNA fragments. First of all, most of the reagents and equipment needed for the procedures are readily available in a standard molecular biology lab. Only one step in the procedure calls for size selection of DNA, and this can be accomplished using commercially available precast Urea gels that can be run in regular SDS-PAGE gel running tanks. Estimation of PCR cycle numbers via qPCR offers a rapid and precise method to determine appropriate PCR amplification. Finally, the compatibility of the single RT primer used in the procedure with the commercial Illumina indexing primers allows sample multiplexing for cost-effectiveness.

8. Limitations

One major challenge with RIPiT is that tandem IP requires a relatively large number of input cells. We usually start with at least four 150-mm plates of 70–80% confluent cells to ensure enough RNA is recovered at the end. This is especially true if targeted RBP is substoichiometric to the RNP complex. Thus, RIPiT can be applied when biological starting material is not limiting. As compared to CLIP-seq where crosslinked RNPs are under stringent wash and gel purification, RIPiT-seq only uses the gentle wash conditions for preserving protein-protein and protein-RNA interactions. Due to this reason, RIPiT-seq reads contain higher level of rRNA carryover (and snRNA in the case for EJC due to its spliceosome association). Also, as RIPiT-seq will enrich RNAs bound directly to a protein as well as those bound via other proteins, it cannot differentiate between direct versus indirect RNA binding. Due to lack of any crosslinks between the interacting RNAs and proteins, RIPiT-seq also cannot reveal the binding site information whereas some variants of CLIP-seq can reveal RBP crosslinking sites at nucleotide resolution.

9. Alternative methods/procedures

In cases where RNA-protein interactions within RNPs are labile and/or short-lived in cell extracts, chemical crosslinking of cells prior to cell lysis with formaldehyde (Mabin et al., 2018; Patton et al., 2020) or other crosslinking agents (Obrdlik, Lin, Haberman, Ule, & Ephrussi, 2019) can stabilize such RNPs. Once RNPs are crosslinked, the first IP can be carried out under more stringent conditions in the presence of high salt and/or stronger detergents. Such alternative strategies can also alleviate non-specific enrichment of abundant RNAs such as rRNAs. As the underlying principle of RIPiT is to perform double purification of a compositionally defined RNP, other methods that can achieve two step purifications can also be applied to an RNP of interest, if suitable. Such examples include tandem affinity purification (e.g., Chen et al., 2018) or density gradient fractionation combined with IP (Bohlen, Fenzl, Kramer, Bukau, & Teleman, 2020; Wagner et al., 2020). In case of certain stable RNPs such as spliceosomal complexes, single IP may be sufficient to enrich an RNP and its bound RNAs (e.g., Burke et al., 2018). Nonetheless, the RIPiT-seq procedure described here is an accessible, scalable and facile approach to interrogate genome-wide binding landscapes of multisubunit RNPs, which operate at almost all steps in regulation of gene expression.