Protocol for analyzing intact mRNA poly(A) tail length using nanopore direct RNA sequencing

Summary

Poly(A) tail metabolism contributes to post-transcriptional regulation of gene expression. Here, we present a protocol for analyzing intact mRNA poly(A) tail length using nanopore direct RNA sequencing, which excludes truncated RNAs from the measurement. We describe steps for preparing recombinant eIF4E mutant protein, purifying m7G- capped RNAs, library preparation, and sequencing. Resulting data can be used not only for expression profiling and poly(A) tail length estimation but also for detecting alternative splicing and polyadenylation events and RNA base modification.

For complete details on the use and execution of this protocol, please refer to Ogami et al. (2022).¹

Graphical abstract

Highlights

• •Preparation of recombinant eIF4E mutant protein from bacteria
• •Purification of m7G-capped RNA using in-house recombinant protein
• •Genome-wide poly(A) tail length analysis of the intact RNAs

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Poly(A) tail metabolism contributes to post-transcriptional regulation of gene expression. Here, we present a protocol for analyzing intact mRNA poly(A) tail length using nanopore direct RNA sequencing, which excludes truncated RNAs from the measurement. We describe steps for preparing recombinant eIF4E mutant protein, purifying m7G- capped RNAs, library preparation, and sequencing. Resulting data can be used not only for expression profiling and poly(A) tail length estimation but also for detecting alternative splicing and polyadenylation events and RNA base modification.

Before you begin

Poly(A) tail metabolism affects post-transcriptional regulation of gene expression,² and Various high-throughput methods have been developed to measure poly(A) tail length.^{1,3,4,5,6,7,8,9,10,11} Each method has its advantages and disadvantages (e.g., requirements of the customized sequencing recipes, minimum amount of input RNAs, levels of computational expertise and cost) (see limitations). Therefore, it is recommended to carefully consider which strategy is accessible and suitable for your research during experiment planning.

Our protocol is more straightforward compared to other methods because it does not require RNA fragmentation, RNA size selection, 5′ adapter ligation, and PCR. These steps are necessary in next generation sequencing-based techniques. Our protocol, on the other hand, relies on nanopore dRNA-seq of RNAs purified using eIF4E mutant protein. This method only analyzes intact RNAs harboring both 5′ cap and 3′ poly(A) tail, avoiding sequencing of degradation intermediates. (Figure 1). Furthermore, because of PCR-free library preparation in nanopore dRNA-seq, there is no risk of PCR bias, which often observed in homopolymeric sequences. Most importantly, nanopore dRNA-seq generates long reads, preserving information on RNA processing and modification, including poly(A) tail length, alternative splicing and polyadenylation, and base modification (e.g., m6A and m5C).¹² By combining with polysome fractionation, these characteristics can be directly linked to translation at per RNA molecule level.¹ The drawbacks are summarized in limitations.

Figure 1.

Schematic of the procedure for direct intact mRNA sequencing

Combination of cap-purification and splint ligation of oligo(dT)₁₀ adapter guarantees only the intact mRNAs are sequenced.

The protocol begins with preparation of GST-fused eIF4E mutant protein (eIF4E_K119A)¹³ from bacteria, which is used to purify 5′ capped RNAs. eIF4E_K119A bind to the m7G capped mRNA with a 2.5-fold higher affinity than wild-type (estimated Kd was 60 pM in eIF4E_K119A and 150 pM in wild-type).¹³ Notably, we have tried several other methods to isolate 5′ capped RNAs, for instance immunoprecipitation (IP) using anti-m7G Cap antibodies or 5′ exonucleolytic degradation of un-capped RNAs. However, we found these approaches ineffective and unpractical in terms of the cost to obtain required yield of input RNA (data not shown). Besides, IP using anti-m7G Cap co-purifies abundant RNA harboring m7G modification (e.g., tRNA, sn/snoRNA and 18S rRNA). Wild-type eIF4E protein can be an alternative for the K119A mutant, although it is possible more protein and Glutathione particles may be required in RNA purification steps. GST- eIF4E protein is available in some vendors such as Sigma Aldrich (cat #: SRP0255). When using these commercially available recombinant proteins, the amount of recombinant protein and Glutathione particles should be optimized. Otherwise, RNA yield can potentially be significantly lower than required. The use of in-house prepared recombinant eIF4E_K119A was the most cost-effective in our hands.

We then describe how to prepare input RNA. The procedure includes LiCl-precipitation, which aims to remove abundant small RNAs (some types of them are capped). This will maximize the yield of capped mRNAs. If small RNAs are in your scope, consider performing ethanol- or isopropanol- precipitation or using RNA purification columns.

Lastly, we describe library preparation and sequencing. As mentioned below, we strongly recommend performing reverse transcription after adapter ligation to maximize the throughput. Also, it is good practice to perform small-scale pilot experiments using a low-cost Flongle flow cell to assess the number of reads required for your transcripts of interest. We typically obtain 1.2–2.2 million reads per a MinION R9.4 flowcell and 20–60 k reads per Flongle R9.4 flowcell. Spike-in controls (such as RCS supplied in Oxford Nanopore’s direct RNA-sequencing kit) should be added to monitor the efficacy of your sequencing.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Bacterial and virus strains
BL21-CodonPlus(DE3)-RIPL Competent Cells	Agilent Technologies	230280
Chemicals, peptides, and recombinant proteins
Isopropyl-β-D(-)-thiogalactopyranoside (IPTG]	FujiFilm	094-05144
Lysozyme, from egg white	FujiFilm	127-06724
100X Protease inhibitor cocktail (EDTA-free)	Nacalai Tesque	03969-34
Glutathione Sepharose 4B	Cytiva	17-0756-01
Glutathione	FujiFilm	071-02014
Amicon Ultra-0.5 Ultracel 30 kDa	Merck Millipore	UFC503008
MagneGST Glutathione Particles	Promega	V8611
Yeast tRNA (10 mg/mL)	Thermo Fisher Scientific	AM7119
Turbo3C protease, (HRV3C Protease)	Nacalai Tesque	H0101S
Glycogen	Sigma-Aldrich	10901393001
Polyvinylalcohol (polymerization degree about 500)	FujiFilm	163-03045
Recombinant RNase inhibitor	TaKaRa	2313B
TRIzol reagent	Thermo Fisher Scientific	15596026
SYBR Gold	Thermo Fisher Scientific	S11494
RNA Clean XP	Beckman Coulter	A63987
SuperScriptIII	Thermo Fisher Scientific	18080093
MinION Flow Cell R9.4.1 (see Note below)	Oxford Nanopore Technologies	FLO-MIN106
Flongle Flow Cell R9.4.1 (see Note below)	Oxford Nanopore Technologies	FLO-FLG001
Direct RNA sequencing kit (see Note below)	Oxford Nanopore Technologies	SQK-RNA002
Critical commercial assays
QuantiFluor RNA System	Promega	E3310
Deposited data
dRNA-seq	Ogami et al., 2022	GEO: GSE137807
Experimental models: Cell lines
HEK293T		Ogami et al., 2022
Recombinant DNA
pGEX6P-1/eIF4E K119A	Addgene	Addgene: 199440
Other
Quantus Fluormeter	Promega	E6150
Sonicator	TOMY	UD-211
MinION	Oxford Nanopore Technologies	MIN-101B
Software and algorithms
MinKNOW	Oxford Nanopore Technologies	https://community.nanoporetech.com
Guppy	Oxford Nanopore Technologies	https://community.nanoporetech.com
MinIONQC	Lanfear et al. (2019)14	https://github.com/roblanf/minion_qc
NanoPlot	Coster et al. (2018)15	https://github.com/wdecoster/NanoPlot
minimap2	Li et al. (2018)16	https://github.com/lh3/minimap2
nanopolish	Workman et al.6	https://github.com/jts/nanopolish
Subread/FeatureCounts	Liao et al. (2014)17	http://subread.sourceforge.net
Tailfindr	Kraus et al. (2019)18	https://github.com/adnaniazi/tailfindr
Samtools	Li et al. (2009)19	https://samtools.sourceforge.net

Note: Although SQK-RNA002/R9.4.1 combination is the only choice for dRNA-seq at this point, the R.9.4.1 is progressively replaced with the newer pore chemistry R10.4.1. According to the Oxford Nanopore Technologies, the R9.4.1 will not be retired until all the kits are upgraded for the R10.4.1 pores. However, it is possible that SQK-RNA002 kit will be discontinued in the future. The upgrade may cause trouble in poly(A) tail length estimation. For more detail, see Troubleshooting Problem 4.

Materials and equipment

Lysis buffer

Reagent	Final concentration	Amount
1 M Tris-HCl (pH 7.5)	50 mM	5 mL
5 M NaCl	150 mM	3 mL
0.5 M EDTA (pH 8.0)	1 mM	200 μL
10% Triton X-100	1%	10 mL
1M DTT	1 mM	100 μL
MilliQ	N/A	81.7 mL
Total	N/A	100 mL

Lysis buffer without DTT can be stored at 4°C for several months. DTT should be added just before use, and the DTT-containing Lysis buffer should be kept on ice throughout the experiments.

Protein Elution buffer

Reagent	Final concentration	Amount
1 M Tris-HCl (pH 8.6)	50 mM	500 μL
Glutathione	10 mM	30.733 mg
MilliQ	N/A	To 10 mL
Total	N/A	10 mL

Protein Elution buffer should be freshly prepared just before use. Make sure that the pH of Elution buffer is near neutral by spotting the buffer onto pH test paper.

eIF4E Storage buffer

Reagent	Final concentration	Amount
1 M Tris-HCl (pH 7.5)	50 mM	500 μL
5 M NaCl	150 mM	300 μL
0.5 M EDTA (pH 8.0)	0.1 mM	2 μL
0.5 M EGTA	0.1 mM	2 μL
1 M DTT	1 mM	10 μL
MilliQ	N/A	9.186 mL
Total	N/A	10 mL

eIF4E Storage buffer without DTT can be stored at 4°C for several months. DTT should be added just before use. Purified recombinant proteins in eIF4E Storage buffer can be safely stored at -30°C for several months.

2X 4E-PD buffer

Reagent	Final concentration	Amount
1 M Tris-HCl (pH 8.0)	100 mM	500 μL
1 M KCl	200 mM	1 mL
0.5M EDTA (pH 8.0)	4 mM	40 μL
10% Triton X-100	0.01%	5 μL
50% Glycerol	5%	500 μL
10% Polyvinylalcohol	2.6%	1.3 mL
1 M DTT	2 mM	10 μL
MilliQ	N/A	1.645 mL
Total	N/A	5 mL

4E-PD buffer without DTT can be stored at 4°C for several months. DTT should be added just before use, and the DTT-containing Lysis buffer should be kept on ice throughout the experiments.

RNA Elution buffer

Reagent	Final concentration	Amount
1 M Tris-HCl (pH 8.6)	100 mM	200 μL
5 M NaCl	150 mM	60 μL
0.5 M EDTA (pH 8.0)	1 mM	4 μL
10% Triton X-100	0.005%	1 μL
Glutathione	40 mM	24.6 mg
1 M DTT	1 mM	2 μL
2 U/μL Turbo3C protease	5U/mL	2.5 μL
MilliQ	N/A	To 2 mL
Total	N/A	2 mL

RNA Elution buffer should be prepared just before use. Make sure that the pH of RNA Elution buffer is near neutral by spotting the buffer onto pH test paper. RNA Elution buffer should be equilibrated to room temperature before use.

Step-by-step method details

Purification of GST-eIF4E_K119A from E.coli BL21 DE3 strains

Timing: 2 days

This section details how to obtain GST-eIF4E_K119A protein produced in BL21 DE3 strains.

• 1.Transform E.coli BL21 DE3 strain with the GST-eIF4E_K119A construct and plate onto a LB plate containing an appropriate antibiotics.

Note: Although the vendor’s transformation protocol suggests performing heat pulse and SOC pre-culture before plating to maximize the efficiency, we typically omit these steps since obtaining a few colonies is enough for the purpose. We simply place the mixture of 10–50ng plasmid (0.5 μL) and E.coli (5 μL) on ice for 30 minutes, and then spread the mixture directly on the plate pre-incubated at 37°C.

• 2.Allow the transformed E.coli to form colonies by incubating at 37°C overnight.
• 3.Pick up a colony and inoculate it in 1.5 mL of LB medium containing 0.2% glucose and appropriate antibiotics.
• 4.Culture at 37°C for 4 h in a shaking incubator.
• 5.Transfer the whole amount of the 1.5 mL culture into a flask containing 100 mL of LB medium containing 0.2% glucose and appropriate antibiotics.
• 6.Culture at 37°C in a shaking incubator until an OD₆₀₀ reaches ∼0.4.
• 7.Culture at 25°C for 2 h in a shaking incubator.
• 8.Allow E.coli to express GST-eIF4E_K119A by adding 500 μL of 100 mM IPTG (the final concentration is 0.5 mM).
• 9.Culture at 25°C in a shaking incubator overnight.

Note: The conditions above work the best in our laboratory in terms of the protein yield in the soluble fraction, although there is a significant decrease in GST-eIF4E_K119A in soluble fraction (Figure 2A). Adding higher concentration of IPTG and/or culturing at higher temperature up to 37°C may improve the yield, however, it is advised to check the solubility of GST-eIF4E_K119A before proceeding to the purification steps. Higher level of induction may cause more incorporation of GST-eIF4E_K119A into the inclusion bodies.

• 10.Transfer the E.coli culture in a centrifuge tube, and centrifuge at 5,000 g at 4°C for 10 min.
• 11.Discard the supernatant completely.

Pause point: The E.coli pellet can be stored at −80°C for several months.

• 12.Resuspend the pellet in 10 mL of Lysis buffer, and add 100 μL of 100XProtease inhibitor cocktail
• 13.Vortex vigorously.
• 14.Add 0.5 mg/mL lysozyme and mix well by inverting the tube.
• 15.Incubate on ice for 20 min.
• 16.Sonicate the lysate (level = 2, duty = 50, 30 s, 10 times, place the tube on ice at least for 30 s at every interval).
• 17.Centrifuge at 18,000 g at 4°C for 10 min. Transfer the supernatant to a new tube.
• 18.Pre-equilibrate 100 μL bed volume of Glutathione Sepharose 4B by washing 3 times with 500 μL of Lysis buffer.
• 19.Add 100 μL bed volume of Glutathione Sepharose 4B to the supernatant and rotate it at 10°C for 1.5 h.
• 20.Centrifuge at 500 g at 4°C for 5 min.
• 21.Discard the supernatant.
• 22.Resuspend the beads in 10 mL of Lysis buffer.
• 23.Centrifuge at 500 g at 4°C for 5 min.
• 24.Discard the supernatant.
• 25.Repeat steps 21–23, 4 times.
• 26.Resuspend the beads in 1 mL of Lysis buffer.
• 27.Transfer the beads to a clean 2 mL tube.
• 28.Centrifuge at 500 g at 4°C for 5 min.
• 29.Discard the supernatant.
• 30.Resuspend the beads in 1 mL of Lysis buffer.
• 31.Centrifuge at 500 g at 4°C for 5 min.
• 32.Discard the supernatant.
• 33.Repeat steps 29–31, 3 times.
• 34.Resuspend the beads in 500 μL of Protein Elution buffer and incubate on ice for 5 min.
• 35.Pre-rinse an Amicon Ultra-0.5 mL 30 kDa device with 500 μL of Lysis buffer. Discard flow through.
• 36.Apply the eluate to the device and centrifuge at 14,000 g at 4°C for 5 min. Discard flow through.
• 37.Resuspend the beads in 500 μL of Protein Elution buffer again and incubate on ice for 5 min.
• 38.Apply the eluate to the top unit of the same device again and centrifuge at 14,000 g at 4°C for 5 min. Discard flow through.
• 39.Repeat steps 36–37 at least 5 times.

Note: Typically, around 100 μL of concentrated eluates will remain in the device.

• 40.Add 400 μL eIF4E storage buffer to the same device.
• 41.Centrifuge at 14,000 g at 4°C for 5 min. Discard flow through.
• 42.Repeat the steps of 39–40.
• 43.Recover the concentrated, buffer-exchanged eluates in a clean 1.5mL tube by reverse centrifugation at 1,000 g at 4°C for 2 min.
• 44.Add 80% glycerol to the final concentration of 50%.
• 45.Measure protein concentration by standard Bradford assay or any other alternative methods.
• 46.Make aliquots and store at −30°C until use.

Note: We confirmed that purified GST-eIF4E_K119A can be used at least several months. Before using the protein samples for RNA purification, check the purity by SDS-PAGE and CBB staining. We typically make aliquots of 10–25 μg/μL GST-eIF4E_K119A.

Figure 2.

Preparation of GST-eIF4E_K119A, input RNA and cap-purified RNA

(A) Small fraction of samples in the indicated steps are resolved by SDS-PAGE and proteins were visualized with CBB. IPTG(-): before induction, IPTG(+): after induction, Sonicated: sonicated sample before centrifugation, Cleared sup: sonicated sample after centrifugation, Flow-through: proteins unbound to glutathione 4B Sepharose, Eluted1/2/3: proteins eluted using glutathione, Beads: proteins remained on the beads after three consecutive elution.

(B) EtBr staining showing efficient removal of small RNAs by two consecutive LiCl precipitation. ppt: precipitated, sup: unprecipitated fraction was ethanol-precipitated and then analyzed.

(C) CBB staining of samples after glutathione/3C protease elution of capped RNAs.

(D) RNA samples in the indicated steps are resolved by agarose gel electrophoresis and stained with SYBR Gold. Recombinant GST protein was used in the control pull down experiments.

Preparation of input RNAs

Timing: 2.5 h

The procedure below ensures removal of abundant small RNAs harboring a m7G cap such as snRNA and snoRNA. If relatively smaller capped and polyadenylated RNAs (< 200 nt) such as some types of long non-coding RNAs are in your scope of the analysis, consider performing ethanol- or isopropanol- precipitation or using an appropriate RNA purification column instead of LiCl precipitation.

• 47.Perform total RNA extraction using TRIzol reagent according to the manufacturer’s instruction (Protocol can be found here).
• 48.Dissolve isopropanol-precipitated total RNA in an 180 μL of nuclease-free water.
• 49.Add 100 μL of 7.5 M LiCl and 20 μL of 500 mM EDTA.
• 50.Incubate at −30°C for 30 min.
• 51.Centrifuge at the max speed at 4°C for 20 min.
• 52.Discard the supernatant and dissolve the pellet in the same volume of nuclease-free water as in step 2.

Optional: If desired, transfer the supernatant in a new tube and perform ethanol precipitation to monitor the removal of small RNAs is successful.

• 53.Repeat the steps 3–6.
• 54.Wash the pellet in 300 μL of 70% ethanol.

Pause point: The RNA pellet can be stored in 70% ethanol at −80°C.

• 55.Remove the supernatant completely and briefly dry the pellet.

Note: Do not dry the RNA pellet completely, since the overdried pellet is hard to solubilize.

• 56.Dissolve the pellet in 150 μL of nuclease-free water.
• 57.Check the integrity of RNA samples by applying small portion of RNAs in agarose gel electrophoresis/SYBR Gold staining or Bioanalyzer (example results of SYBR Gold staining are shown in Figure 2B).

Pause point: Although we recommend using the RNA solution as soon as possible to avoid the risk of degradation, RNA in nuclease-free water can be safely stored at −80°C for several days.

Purification of capped RNAs

Timing: 3 h

The following protocol outlines the steps for capped RNA purification using GST-eIF4E_K119A protein and small RNA-removed total RNA obtained in the above section.

• 58.Pre-equilibrate 50 μL bed volume of MagneGST Glutathione Particles by washing 3 times with 200 μL of 1X 4E-PD buffer in a 1.5 mL tube.
• 59.Resuspend the beads in 200 μL of 1X 4E-PD buffer, and add 100 μg of GST-eIF4E_K119A.
• 60.Rotate the tube at 10°C for 2 h.
• 61.Place the tube on a magnet stand for at least 1 min.
• 62.Discard the supernatant, and resuspend the beads in 200 μL of 1X 4E-PD buffer.
• 63.Rotate the tube at room temperature for 1 min.
• 64.Repeat the steps 4–6 twice.
• 65.Place the tube on a magnet stand for at least 1 min.
• 66.Discard the supernatant from the tube on the magnetic stand, and resuspend in 200 μL of 2X 4E-PD buffer supplemented with 2.5 μL of 10 mg /mL yeast tRNA and 1 μL of 40 U/μL Recombinant RNase inhibitor.
• 67.Dilute >30 μg of the LiCl-precipitated RNA in nuclease-free water to make 196.5 μL solution, and heat-denature at 65°C for 10 min, and then chill on ice more than 1 min.

Note: Typically, about 1.2 μg of capped RNA is purified from 30 μg of total RNA extracted from HEK293T cells.

• 68.Add the heat-denatured RNA to the tube containing GST-eIF4E_K119A-bound beads.
• 69.Rotate at 25°C for 2 h.
• 70.Place the tube on a magnet stand for at least 1 min.
• 71.Discard the supernatant, and resuspend the beads in 500 μL of 4E-PD buffer.
• 72.Rotate the tube at room temperature for 1 min.
• 73.Repeat the steps 12–14 three times.
• 74.Place the tube on a magnet stand for at least 1 min, and discard the supernatant completely.
• 75.Resuspend the beads in 100 μL of RNA Elution buffer, and incubate on ice for 5 min.
• 76.Place the tube on a magnet stand for at least 1 min, and transfer the eluates to a new 2 mL tube.
• 77.Repeat the steps 17–19 three more times.

Note: We routinely check the cleavage and elution of GST-eIF4E_K119A by sampling a small amount of samples for SDS-PAGE/CBB-staining (Figure 2C).

• 78.Combine the eluate (final volume should be ∼400 μL).
• 79.Add 1.2 mL of TRIzol to the tube, and vortex vigorously for at least 1 min.
• 80.Add 240 μL of chloroform, and vortex vigorously for at least 1 min.
• 81.Incubate at room temperature for 5 min.
• 82.Centrifuge at 12,000 g at 4°C for 10 min.
• 83.Transfer the aqueous phase (∼1 mL) to a new 15 mL tube.
• 84.Add 1 μL of glycogen, mix well.
• 85.Add 2.5 mL of ethanol and 100 μL of 3 M sodium acetate.
• 86.Mix well, and incubate at -30°C for at least 1 h. Overnight incubation is recommended.

Pause point: The solution can be stored at -30°C or -80°C for several weeks.

• 87.Centrifuge at the maximum speed at 4°C for 15 min
• 88.Remove the supernatant completely, and briefly dry the pellet.

Note: Do not dry the RNA pellet completely, since the overdried pellet is hard to solubilize.

• 89.Dissolve the RNA pellet in 10 μL of nuclease-free water.
• 90.Use 1 μL of RNA to measure the concentration using Fluorometer such as QuantiFluor RNA system.

Note: We recommend using more than 1 μg of capped RNA to proceed the following steps. Using the lower amount of RNA might cause significant reduction in the final amount of library RNAs and the throughput.

Optional: Quality of cap-purified RNA can be checked by agarose gel electrophoresis/SYBR Gold staining or Bioanalyzer (example results are shown in Figure 2D).

Library preparation and nanopore direct RNA sequencing

Timing: 3 h (library preparation), up to 72 h (sequencing run)

Finally, the recovered capped RNA is then used for library preparation for nanopore dRNA-seq.

• 91.Add 1.8 volume of RNAClean XP to >1 μg of capped RNA.
• 92.Purify capped RNA according to the manufacturer’s protocol (Protocol can be found here)
• 93.Elute RNA in 11 μL of nuclease-free water.
• 94.Use 1 μL of RNA to measure the concentration using Fluorometer such as QuantiFluor RNA system.
• 95.Preparate library using 9 μL of RNA eluates according to “Direct RNA sequencing” protocol from Oxford Nanopore for use with the kit (Protocol can be found here, note: registration and sign in required).

CRITICAL: Although it is optional, we strongly recommend performing reverse transcription after the adapter ligation. When using reverse transcriptase other than SuperScriptIII (Thermo Scientific), it is advised to compare the throughput and the quality of the sequencing data with the library prepared using SuperScript III. This can be achieved at low cost by applying a small fraction of libraries to Flongle flow cells. In our experience, SuperScript III gives better results in nanopore direct RNA-sequencing than SuperScript IV (data not shown). The other RT enzymes such as Maxima H Minus (Thermo Fisher Scientific, EP0751) and Induro Reverse Transcriptase (New England BioLabs, M0681S) may give even better results. Induro is a recently released RT enzyme with high processivity (for more information, visit here).

Optional: It is a good practice to always add spike-in controls such as RNA CS (RCS, provided in the Oxford nanopore’s kit) in the beginning of the library preparation.

• 96.Load the RNA library to the MinION flowcell and run using MinKNOW.

Note: We routinely measure library concentration using Fluorometer such as QuantiFluor RNA system before applying to a flowcell. When starting amount of capped RNA is 1 μg, we typically obtain 500–750 ng RNA library.

An example pipeline for base calling, alignment and downstream analysis

Timing: days to months

When the required number of reads are obtained, or when almost all the pores are inactivated, you may stop sequencing and proceed to computational analysis. The following is an example for the downstream analysis.

• 97.MinKNOW software generates HDF5 format files with an extension “.fast5”, which contains the electrical signal data per RNA molecule sequenced. The signal can be converted into nucleic acid sequence using basecaller such as Oxford Nanopore’s Guppy, which produces FASTQ files and a text summary file. Basecalling can be performed during or after sequence run. We routinely perform quality control analysis of sequence using software such as MinIONQC,¹⁴ NanoPlot¹⁵ and pycoQC.²⁰ Long read alignment to a reference genome can be performed using minimap2,¹⁶ which generates SAM files when running with an “-a” option. Minimap2 is one of the least resource-demanding software in terms of CPU hours and peak memory usage.²¹ BAM and BAM index files can be obtained using samtools.¹⁹ The results of read alignment can be visualized by loading the BAM files to IGV. When using Oxford Nanopore’s RCS (yeast Enolase II, YHR174W) as a spike-in control, align them to the yeast genome. Read assignment to genes and read counting per gene can be achieved using featureCounts¹⁷ with a “-L” option or LIQA.²² LIQA is designed for isoform-specific quantification, and thus suitable for alternative splicing aware analysis. There are several poly(A) tail length estimation tools such as nanopolish polya⁶ and tailfindr.^18,23

Expected outcomes

This protocol allows the user to generate high-throughput sequencing data of the intact RNAs. Typically, we obtain ∼1 mg of GST-eIF4E_K119A protein from a 100 mL culture of BL21 (DE3) strain, and > 1 μg of 5′ capped RNA from 30 μg of LiCl-precipitated total RNA isolated from HEK293T cells using 100 μg of GST-eIF4E_K119A. Successful sample preparation and sequencing operation typically generates 1.2–2.5 million reads per MinION flow cell with mean and median RNA quality scores of 8.5–10.5, but shorter reads (< 100 nt) tend to have a low quality score (Figure 3A). These values can be variable depending on the source of BL21 strains, cell types and the quality of flowcells.

Figure 3.

Direct RNA-sequencing

(A) A scatter plot showing the relationship between read length and quality score.

(B) An example of IGV snapshot of direct RNA-sequencing result. ILF3 gene locus is shown. Each read has various information on the processing of mRNA (e.g., alternative polyadenylation (APA) sites and alternative splicing including retained intron).

(C) Poly(A) tail length analysis was performed using nanopolish polya. Poly(A) tail length distribution of HEK293T transcriptome and Oxford nanopore’s RCS strand is shown.

Despite using the intact RNA for sequencing, reads acquired by dRNA-seq are not always full-length (Figure 3B). This is presumably due to the clogging of the nanopores by contaminants or pore blockage by RNA secondary structures. Reverse transcription during library preparation helps resolve RNA secondary structures, but incomplete cDNA synthesis could cause the blockage.

Poly(A) tail length distribution of Oxford Nanopore’s RCS should show a sharp peak around 30 nt (Figure 3C). The distribution pattern of cap-purified RNA is dependent on cell types and culture conditions (e.g., gene knockdown/knockout, drug treatment and metabolic stress).

Limitations

While nanopore dRNA-seq generates high-throughput data, sequencing depth is still limited compared to Illumina or PacBio-based methodologies. Performing low-cost preliminary experiments using Flongle (generating ∼50,000 reads) is strongly advised to estimate how much depth required to analyze your gene of interest. Depending on the aims of the study, there are two alternatives to improve sequencing depth. First, direct cDNA sequencing generates significantly higher number of reads, however, RNA base modification information will be lost. Second, PCR-cDNA sequencing outputs the highest throughput, but instead, information on poly(A) tail length and RNA base modification will be lost. In addition, the number of effective reads is expected to be lower in research focusing on RNA 5′ end, since many reads do not reach to the 5′ most end presumably due to premature pore clogging or blockage. Thus, it is important to choose carefully which methodology is suited for your own aims.

Troubleshooting

Problem 1

No or low yield of GST-eIF4E_K119A (step 46).

Potential solution

To find out which step in Purification of GST-eIF4E_K119A from E.coli BL21 DE3 Strains causes the problem, transfer small amount of samples to a new 1.5 mL tube at the steps 7 (-IPTG), 9 (+IPTG), 17(soluble) and 33 (eluate). If no induction was observed in “+IPTG”, check if correct pGEX plasmid and BL21 strain were used. If no or low GST-eIF4E_K119A is present in “soluble” fraction, first examine if lysozyme treatment and sonication are efficient. If the problem lies in GST-eIF4E_K119A solubility, perform small scale experiments to determine the induction condition by modulating parameters (e.g., induction temperature 16°C–37°C, IPTG concentration 0.05–1 mM, OD₆₀₀ 0.2–0.6). If no or low protein was detected in “eluate”, check if pH of protein elution buffer is neutral and if the beads are well mixed with the buffer.

Problem 2

No or low yield of capped RNAs (step 90).

Potential solution

Check the integrity of your input RNA by gel electrophoresis/SYBR Gold staining or Bioanalyzer (Figure 2D). Perform all the RNA works at a clean place and free from RNases. If necessary, wipe the surface of working place, equipment, pipettes using RNase decontamination solution such as RNaseZap (Thermo Fisher Scientific, cat#: AM9780) or RNase Knockout (FujiFilm, cat#: 181-03381). Make sure concentration of your recombinant protein is correct and your protein is properly stored. Check the protein cleavage and elution by CBB staining (Figure 2C).

Problem 3

dRNA-seq generates low throughput and/or low quality data. Reads are shorter than expected (step 97).

Potential solution

Always check if the flow cell to be used has enough active pores. Measure the final library concentration using Fluorometer such as QuantiFluor RNA system. Typically, we obtain 500–750 ng of library from 1 μg of starting capped RNA, and we found that lower amount of the library produces lower throughput data. Also, monitor output over run time in the MinKNOW. Impurity of RNA library or poor handling of flowcell (e.g., introducing air bubbles during priming and sample load) causes faster pore inactivation. The quality of reverse transcription is also crucial for high-quality data. Inspect MinKNOW-reported end reason. In a library suffering from poor reverse transcription, many of the reads show “Read became blocked” or “Unblock voltage reversal”. cDNA content can be potentially evaluated by fluorometer using DNA assay system. If the problem persists, consider using other enzymes such as Maxima H minus (Thermo Fisher Scientifc, EP0751) and Induro Reverse Transcriptase (New England BioLabs, M0681S). If successful, runtime using around 500–750 ng library is 48–72 h, and mean and median RNA quality scores exceeds 7.0 (typically 8.5–10.5) (Figure 3A).

Problem 4

Failure in poly(A) tail length estimation after upgrading direct RNA sequencing chemistry (step 97).

Potential solution

Although SQK-RNA002/R9.4.1 combination is the only choice for dRNA-seq at this point, the R.9.4.1 is progressively replaced with the newer pore chemistry R10.4.1. According to Oxford Nanopore Technologies, the R9.4.1 will not be retired until all the kits are upgraded for the R10.4.1 pores. However, it is possible that SQK-RNA002 kit will be discontinued in the future. Since the newer kit is likely to provide faster nucleotide translocation rate, poly(A) tail length estimation software will not work for data produced using the new sequencing chemistry. We have experienced the case before, when the kit was switched from SQK-RNA001 to SQK-RNA002. Re-training of the software using new chemistry data will be required. Developers of the software should take care of this upon request.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Shin-ichi Hoshino (hoshino@phar.nagoya-cu.ac.jp).

Materials availability

Plasmids generated in this study have been deposited to Addgene.

Data and code availability

This study did not generate code. dRNA-seq data are available in GEO database (GSE137807). All additional technical information required to perform the experiments and dRNA-seq data analysis are available from the technical contact, Koichi Ogami (koichi.ogami@med.nagoya-u.ac.jp) upon request.