Determining mRNA stability by metabolic RNA labeling and chemical nucleoside conversion

Abstract

The varying rates at which mRNAs decay are tightly coordinated with transcriptional changes to shape gene expression during development and disease. But currently available RNA sequencing approaches lack the temporal information to determine the relative contribution of RNA biogenesis, processing and turnover to the establishment of steady-state gene expression profiles.

Here, we describe a protocol that combines metabolic RNA labeling with chemical nucleoside conversion by thiol-linked alkylation of 4-thiouridine to determine RNA stability in cultured cells (SLAMseq). When coupled to cost-effective mRNA 3' end sequencing approaches, SLAMseq determines the half-life of polyadenylated transcripts in a global and transcript-specific manner using untargeted or targeted cDNA library preparation protocols.

We provide a step-by-step instruction for time-resolved mRNA 3' end sequencing, which augments traditional RNA-seq approaches to acquire the temporal resolution necessary to study the molecular principles that control gene expression.

1. Introduction

Gene expression comprises a series of tightly controlled molecular events that involve the biogenesis, processing and turnover of RNA molecules. Among these, the regulated decay of messenger RNA (mRNA) acts as a key step to control the quality and quantity of cellular transcripts, which in turn ensures faithful protein expression [1]. To this end, a multitude of cis- and trans-acting factors adjust the life-span of mRNAs according to its encoded function [2], and in response to environmental conditions, including stress [3], metabolic adjustments [4], or developmental cues [5]. Therefore, robust transcriptomics approaches that measure gene expression kinetics in an accessible, cost-effective, and scalable manner are key to systematically dissect the biological processes that modulate RNA decay. Early attempts to measure transcript stabilities mostly employed highly invasive and cytotoxic approaches that globally interfere with RNA synthesis using transcriptional inhibitors or conditional mutants [6, 7]. But such measurements remain imprecise and frequently induce cellular stress responses that may activate alternative routes for transcript-specific decay or interfere with the intrinsic coupling of transcription and decay [8]. In contrast, metabolic labeling of RNA in intact cells represents an attractive alternative to follow the fate of transcripts under unperturbed cellular conditions. This was initially achieved by low-throughput protocols that utilized radiolabeled nucleotides and probe-based filter-binding [9]. Later refinements to this approach included biologically inert ribonucleotide analogs, such as 4-thiouridine (s⁴U), which exhibit unique physicochemical properties that enabled the reversible conjugation of labeled transcripts, for example to mercury agarose [10]. Numerous updates to this protocol meanwhile combine different nucleotide analogs (e.g. 4-thiouridine, 5-ethyniluridine, or 5'-bromouridine) with diverse biochemical enrichment methods (e.g. reversible biotinylation or classical immunoprecipitation) and gene expression profiling approaches (e.g. qRT-PCR, microarray, or RNAseq) to assess transcript-specific and global RNA stabilities [11]. However, such approaches face significant technical challenges, including the labor-intensive and semi-quantitative nature of biochemical enrichment techniques, the need for ample starting material, and the complex and laborious kinetic modeling steps that integrate sub-fraction-derived datasets (i.e. input, bound, and supernatant) into gene expression dynamics. Many of these limitations were recently overcome by simple chemical conversion of s⁴U into cytosine analogs [12–14]. Here, total RNA prepared from cells after s⁴U metabolic labeling is subjected to chemical treatment that changes the base-pairing capacity of s⁴U top prompt specific T to C conversions in metabolically labeled transcripts upon cDNA library preparation. By detecting metabolically labeled transcripts in the context of total RNA, time-resolved RNA sequencing provides insights into gene expression dynamics at unprecedented accessibility, efficiency and scalability. Among the available conversion chemistries, SLAMseq employs a rapid (<15 minutes reaction time) and efficient (>94% reaction efficiency) nucleophilic substitution reaction that transfers a carboxyamidomethyl-group from iodoacetamide to the thiol group of 4-thiouridine, which causes the mis-incorporation of guanine across alkylated 4-thiouridine during reverse-transcription (Figure 1). When coupled to metabolic RNA labeling protocols, SLAMseq revealed RNA decay rates for thousands of transcripts in mouse embryonic stem cells, and reliably portrayed the regulatory effects of cellular microRNAs and post-transcriptional RNA modifications, such as N6-methyladenosine [13]. Beyond the quantitation of RNA decay, SLAMseq has also been successfully applied to measure (1) rapid changes in transcription for the delineation of primary target genes of transcription factors [15], (2) tissue- and cell-type-specific gene expression profiling in mammals [16], and (3) RNA transport between tissues in vivo in mammals [17].

Figure 1.

Schematic overview of thiol(SH)-linked alkylation for the metabolic sequencing of RNA (SLAMseq). Cultured cells are subjected to metabolic RNA labeling by the addition of 4-thiouridine (s⁴U) to the culture medium. Following total RNA extraction at several time-points after uridine chase, treatment with iodoacetamide results in thiol-specific alkylation, which modulates the base-pairing capacity of s⁴U, prompting the site-specific mis-incorporation of guanine across chemically modified s⁴U during cDNA synthesis by reverse transcription. As a consequence, SLAMseq specifically identifies s⁴U incorporations at single-nucleotide resolution by sequencing and enables to quantify transcript decay kinetics in a global and transcript-specific manner by standard high-throughput sequencing.

Here, we describe a step-by-step protocol for the global and transcript-specific determination of mRNA half-lives using SLAMseq. While this metabolic RNA sequencing is in principle compatible with any standard RNA sequencing approach that involves the preparation of cDNA libraries, we describe a cost-effective way to assess the stability of polyadenylated RNA polymerase II transcripts from low input material by employing mRNA 3'end sequencing. This sequencing approach enables to independently address mRNA 3' end variants, which frequently modulate the fate and function of transcripts [18]. Finally, we report a simple targeted version of this sequencing protocol that enables to address the decay kinetics even of poorly expressed transcripts, which frequently fail to report robust conversion data in untargeted library preparation protocols.

2. Material

All reagents and labware must be free of nucleases and nucleic acid contamination. Precautions need to be taken at all times during the protocol to ensure nuclease-free conditions. Solutions and reagents should be stored at room temperature unless otherwise specified by the manufacturer.

2.1. General Equipment

• -Humidified CO₂-controlled tissue culture incubator
• -Laminar flow hood
• -Benchtop refrigerated centrifuge (up to 20,000 × g)
• -Microvolume Spectrophotometer (e.g. Nanodrop)
• -Heat block
• -Benchtop thermal cycler
• -Real-Time PCR detection system
• -Magnetic Stand for PCR tubes
• -pH meter
• -Capillary electrophoresis instrument (e.g. AATI Fragment Analyzer)
• -Illumina high-throughput sequencing machine (e.g. Illumina HiSeq 2500)
• -Luminescence Plate Reader (e.g. Synergy, BioTek)
• -Orbital shaker
• -Water bath (set to 37°C)
• -Sterile tissue culture plasticware
• -500 ml Nalgene® Rapid-Flow™ Filter Units and Bottle Top Filters (0.2 μm)
• -Volumetric pipettes (5, 10, 25 mL)
• -Filter tips (1-1000 μL)
• -Safe-lock 1.5 mL tubes
• -15mL and 50 mL Falcon tubes
• -0.2 mL 8-tube PCR strips and ultraclear flat cap strips
• -Plate-reader compatible 96-well plates (e.g. Cell culture microplate, 96 well, PS, F-bottom, white, cellstar® TC, Greiner Bio-One)

2.2. Reagents and Buffers

• -DMEM High Glucose medium
• -certified (ES cell pre-tested) FBS (Gibco)
• -100 x penicillin- streptomycin solution (100 U/ml penicillin, 0.1 mg/ml streptomycin)
• -200 mM L-glutamine
• -100 x MEM Non-essential amino acid solution (Sigma or other)
• -100 mM sodium pyruvate
• -50 mM 2-Mercaptoethanol
• -2 mg/ml LIF
• -0.5 % Trypsin-EDTA (10x)
• -1× sterile Phosphate buffered saline (PBS)
• -4-Thiouridine (Sigma-Aldrich)
• -Uridine (Sigma-Aldrich)
• -Cell viability assay, e.g. CellTiterGlo® Luminescent Cell Viability assay (Promega)
• -TRIzol™ Reagent (Ambion, Life Technologies)
• -Chloroform:Isoamyl Alcohol 24:1
• -3 M NaOAc (pH 5.2)
• -2-Propanol
• -Ethanol absolute
• -Glycogen (20 mg/ml)
• -OmniPur® DTT (Merck)
• -Nuclease-free water
• -Iodoacetamide (Sigma-Aldrich)
• -DMSO
• -NaH₂PO₄ (monobasic, Sigma Aldrich)
• -Na₂HPO₄ (dibasic, Merck)
• -DNF-471 Standard Sensitivity RNA Analysis Kit (AATI)
• -TURBO DNA-free^™ Kit (Thermo Fisher Scientifi cor other)
• -QuantSeq 3' mRNA-Seq Library Kit for Illumina (Lexogen) (optional)
• -QuantSeq-Flex Targeted RNA-Seq Library Prep Kit for Illumina (Lexogen) (optional)
• -PCR Add-on Kit for Illumina (Lexogen)
• -RNaseOUT™ Recombinant Ribonuclease Inhibitor (Thermo Fisher Scientific)
• -SuperScript III Reverse Transcriptase kit (Thermo Fisher Scientific)
• -10 mM dNTPs
• -RNase H (New England Biolabs)
• -2× Taq Master Mix, e.g. 2x GoTaq® Green Master Mix (Promega)
• -AMPure XP Beads (Beckman Coulter)
• -KAPA Real-time Library Amplification kit (Kapa Biosystems)
• -DNF-474 High Sensitivity NGS Fragment Analysis Kit (1 bp - 6,000 bp) (AATI)

2.3. Cell lines

Mouse embryonic stem cells (mESCs) (e.g. clone AN3-12, obtained from IMBA Haplobank [19])

3. Methods

General Considerations:

Time-resolved RNA sequencing is based on metabolic RNA labeling that can be applied in principle to any cultured metazoan cell type. To this, thiol(SH)-linked alkylation for the metabolic sequencing of RNA (SLAMseq) employs the nucleotide analog 4-thiouridine (s⁴U), which is biologically inert when applied at appropriate doses. When added as a supplement to cell culture medium, metazoan cells readily take up s⁴U via nucleoside equilibrate transporters and incorporate it into newly synthesized RNA. Notably, the efficiency of s⁴U incorporation into nascent RNA depends on cell type-specific variables, such as cellular uptake kinetics and overall transcriptional activity. Thus, assessing the sensitivity to s⁴U and the efficiency of s⁴U incorporation into RNA prior to a SLAMseq experiment is key to obtain meaningful results and experimental parameters, such as s⁴U concentration, s⁴U labeling time or cDNA library sequencing depth, need to be adjusted accordingly. Here, we describe how optimal s⁴U labeling conditions can be achieved in mouse embryonic stem cells (mESCs). Cell-type-specific adjustments that enable to apply the protocol to different cell types of various organismal origin will be discussed in the notes section.

3.1. Mouse embryonic stem cell (mESC) culture procedures

General maintenance: Operate under sterile conditions. Maintain mESCs (clone AN3-12) as described previously [19]. Keep cells in a humidified cell culture incubator at 37°C with 5% CO₂. Passage cells 1:10 every second day using Trypsin-EDTA for a maximum number of 10-15 passages.

1. Mix the following reagents to prepare mESC culturing media:

   • 450 ml DMEM High Glucose medium
   • 75 ml FBS
   • 5 ml Penicillin-streptomycin solution (100x)
   • 5 ml L-glutamine (200 mM)
   • 5 ml MEM Non-essential amino acid solution (100x)
   • 5 ml Sodium pyruvate (100 mM)
   • 550 μl 2-Mercaptoethanol (50 mM)
   • 5 μl LIF (2 mg/ml)

2. Sterile filtrate mESC media using a 500 ml 0.2 μm bottle top filter and store media at 4°C for a maximum of 2 months.

3.2. Determining optimal s⁴U concentration for metabolic RNA labeling

In order to achieve optimal metabolic RNA labeling in the absence of cytotoxic side-effects, the concentration of s⁴U needs to be carefully adjusted for any given cell type.

Testing s⁴U-induced cytotoxicity:

Here, we describe how to use CellTiterGlo® to assess cell viability after s⁴U-labeling of mESCs. Alternatively, other assays can be used to assess cell viability (e.g. viability dyes for flow cytometry).

1. Prepare a 0.5 M stock solution of s⁴U in H₂O, aliquot the stock solution and store the aliquots at -20°C for several weeks. Avoid exposure to unnecessary long exposures to UV light throughout the experiment (Note 1).

2. Seed 5000 mESCs per well on a 96-well plate in a volume of 200 μl mESC culturing medium the day before the labeling experiment. Assess the cell viability by Tryptan blue staining in the process of cell counting. To guarantee reproducible results, ensure that the cell viability exceeds 90% prior to seeding

3. Incubate cells at 37°C overnight.

4. Prepare serial dilutions of s⁴U in mESC culturing medium in a concentration range of 0 – 1 mM.

5. Remove old media from the cells and add 150 μl of s⁴U-containing media per well (Note 2). Use each s⁴U-concentration in triplicates and include three wells of ‘no s⁴U’ control. Additionally, prepare three wells without cells to measure the background luminescence of the culturing media.

6. Replace media every three hours with fresh media containing the respective s⁴U concentration (Note 3).

7. Prepare CellTiterGlo® reagent by equilibrating the CellTiter-Glo® Buffer and lyophilized CellTiter-Glo® Substrate to room temperature. Transfer the appropriate volume of CellTiter-Glo® Buffer to lyophilized CellTiter-Glo® Substrate to form the CellTiterGlo® reagent and mix gently by vortexing.

8. After 12 h or 24 h of s⁴U labeling, add 150 μl CellTiterGlo® reagent per well and mix the contents for two minutes on an orbital shaker. Protect the plate from light.

9. Incubate the plate at room temperature for 10 min.

10. Transfer the lysate to a luminescent reader compatible 96-well plate.

11. Measure luminescence, e.g. on Synergy (BioTek) using Gen5 Software.

12. Assess cellular viability under different s⁴U-labeling conditions: Subtract the background luminescence signal of wells without cells, normalize luminescence to fit a sigmoidal dose response curve and calculate the EC₅₀.

13. For pulse-chase experiments, use a s⁴U concentration that maintains a cell viability > 90% after 12h or 24h of s⁴U labeling. For mESCs, we use a s⁴U concentration of 100 μM for metabolic labeling up to 24h with a s⁴U-containing media exchange every 3 hours (Note 4).

3.3. Measuring global and specific mRNA stability in mESCs

The following set-up describes a pulse-chase experiment to measure mRNA stabilities in mESCs. The experimental set-up is designed for technical triplicate measurements over 7 time-points after a 24h pulse labeling (0h, 0.5h, 1h, 3h, 6h, 12h, 24h chase) and ‘no labeling’ controls. The pulse labeling is performed in a 6-well format and includes media exchange every three hours during 24 hours of the s⁴U pulse. In order to adjust RNA yields to downstream applications, scale the experiment accordingly.

3.3.1. Pulse-chase labeling experiment

1. Seed 2 x10⁵ cells on 24 wells of 6-well plates the day before the labeling experiment (Note 5, Note 6). Assess the cell viability by Tryptan blue staining in the process of cell counting. To guarantee reproducible results, ensure that the cell viability exceeds 90% prior to seeding.

2. Incubate cells at 37°C overnight.

3. On the following day, prepare 350 ml mESC culturing medium supplemented with 100 μM s⁴U (Note 7). Aliquot the s⁴U-containing media into seven 50 ml flasks for media exchanges and store at 4°C protected from light (Note 8).

4. Prepare 50 ml mESC culturing medium containing 10mM of uridine and store at 4°C (Note 9).

5. Equilibrate the first aliquot of s⁴U-medium to 37°C.

6. Remove the media from the cells. Add 2 ml culturing media without s⁴U to three wells (‘unlabeled’). To all other wells, add 2 ml s⁴U-containing media to start the pulse labeling.

7. Exchange the s⁴U containing media with fresh s⁴U-containing media every three hours. Always equilibrate 50 mL medium aliquots to 37°C prior to media exchange.

8. After 24h, remove s⁴U-containing media from the cells. Add 500 μl of TRIzol® to the three wells without s⁴U (‘unlabeled’) and to three wells with s⁴U (‘0h chase’) and transfer the lysate to 1.5 ml tubes (Note 10). Freeze these samples at -80°C. For all the other wells, wash cells twice with 2 ml 1× PBS. Start the chase by the addition of 2 ml uridine-containing mESC culturing medium per well (Note 11).

9. Collect time-points in triplicates at 0.5h, 1h, 3h, 6h, 12h and 24h after chase onset by lysing the cells in 500 μL TRIzol® (Note 12). Transfer lysate to 1.5 ml tubes and freeze samples at -80°C.

10. Store samples at -80°C or directly proceed to RNA isolation (3.3.2.).

3.3.2. RNA isolation

Protect s⁴U-labeled RNA from light. Work under fume-hood at least until the EtOH washing step.

1. Thaw TRIzol® lysate and incubate 5 min at room temperature.

2. In the meantime, prepare 0.2 mM DTT in 2-propanol (14 μl 0.1M DTT in 7 mL 2-propanol) and 0.1 mM DTT in 75% EtOH (13 μl 0.1 M DTT in 13 mL 75% EtOH) (Note 13).

3. Add 100 μl Chloroform:Isoamyl Alcohol 24:1 per 0.5 ml of TRIzol®.

4. Vortex tube for 15 sec.

5. Incubate at room temperature for 2-3 min.

6. Spin down at 16,000 x g for 15 min at 4°C.

7. Carefully transfer aqueous phase (~ 200 μL) to a new tube. Do not disturb the interphase.

8. Add 200 μl DTT-containing 2-propanol (0.1 mM final concentration) and 1 μl glycogen (20 mg/ml) (optional) and vortex well.

9. Incubate 10 min at room temperature.

10. Spin down at 16,000 x g for 20 min at 4°C.

11. Carefully take off the supernatant with a pipette and discard.

12. Add 500 μl DTT-containing 75% EtOH (0.1 mM final concentration) and vortex well.

13. Spin down at 7,500 x g for 5 min at room temperature.

14. Completely remove supernatant with a pipette and let the pellet dry for 5-10 min. Do not overdry as it might become difficult to resuspend the RNA pellet.

15. Resuspend in 20-30 μL of 1 mM DTT.

16. Incubate for 10 min at 55°C to completely dissolve the RNA pellet.

17. Measure concentration by Nanodrop. Typically, we obtain from one well of cultured mESCs 10-15 μg of total RNA. Store RNA at -80°C (Note 14) or proceed to s⁴U alkylation.

18. Control the quality of the RNA, e.g. on a capillary electrophoresis system like Fragment Analyzer (AATI, kit DNF-471). See Figure 2 (top) for typical results.

Figure 2. Electropherograms of total RNA prepared from mouse embryonic stem cells before (left panels) and after metabolic RNA labeling for 12 h (right panels) obtained on a Fragment Analyzer.

Treatment of total RNA with iodoacetamide for s⁴U-alkylation (bottom panels) as part of the SLAMseq protocol does not influence total RNA quality or integrity when compared to untreated conditions (top panels), as judged by comparing RNA quality value (RQN). Blue and red peaks correspond to the 28S and 18S region, respectively. Fragment size is indicated in nucleotides (nt). RFU, relative fluorescence unit; LM, lower marker.

3.3.3. s⁴U-Alkylation (Iodoacetamide treatment)

1. Prepare the Sodium Phosphate buffer (500 mM NaPO₄, pH 8) as following: Make 1 M stocks solutions of monobasic NaH₂PO₄ (138 g in 1 L H₂O) and dibasic Na₂HPO₄ (142 g in 1L H₂O). To prepare 200 ml of 0.5 M sodium phosphate buffer, mix 93.2 mL of 1 M Na₂HPO₄ and 6.8 ml of 1 M NaH₂PO₄ and add 100 ml of H₂O. Adjust to pH 8 after buffer preparation using a pH meter.

2. Freshly prepare a 100 mM IAA stock in EtOH 100%. Keep iodoacetamide protected from light.

3. Dilute 2 μg of total RNA in 15 μL nuclease-free water (Note 15). Keep RNA on ice.

4. Prepare the following mastermix (Note 16):

   	Volume	Final concentration
   IAA (100 mM) in EtOH 100%	5 μl	10 mM
   NaPO4, pH 8 (500 mM)	5 μl	50 mM
   DMSO	25 μl	50% (v/v)
   Final volume	35 μl

5. Add 35 μl of the mastermix to 15 μl RNA, mix well.

6. Incubate the reaction at 50°C for 15 min.

7. Transfer samples on ice and immediately quench the reaction by adding 1 μl 1M DTT (Note 17). Vortex briefly.

8. Precipitate RNA by adding 1 μl glycogen (20 mg/ml), 5 μl NaOAc (3M, pH 5.2) and 125 μl EtOH 100%. Vortex briefly and incubate for 30 min at -80°C or for several hours at -20°C.

9. Spin down at 16,000 x g for 30 min at 4°C.

10. Remove supernatant with a pipette and add 1 ml 75% EtOH, vortex.

11. Spin down at 16,000 x g for 10 min at 4°C.

12. Completely remove supernatant with a pipette and let the pellet dry for 5-10 min. Do not overdry as it might become difficult to resuspend the RNA pellet.

13. Resuspend RNA pellet in 10 μl H₂O.

14. Measure RNA concentration on Nanodrop.

15. Assess RNA quality e.g. on a capillary electrophoresis system like Fragment Analyzer (AATI, kit DNF-471). Note, that the conditions of the s⁴U alkylation reaction do not interfere with RNA integrity (see Figure 2, bottom).

3.3.4. Targeted and untargeted mRNA 3' end cDNA library preparation

SLAMseq is in principle compatible with any RNA sequencing library preparation protocol that converts RNA into cDNA. To determine global and transcript-specific stability of poly-adenylated RNA polymerase II transcripts, we commonly employ mRNA 3' end sequencing approaches (e.g. Lexogen’s QuantSeq 3' mRNA-Seq Library Prep Kit for Illumina), as previously described [13]. Using this approach, an average sequencing depth of 10-20 mio reads per time-point provides robust insights into the stability of thousands of transcripts [13].

To further reduce bioinformatic workload and/or sequencing costs, or to increase the sensitivity towards lowly-expressed transcripts, we recommend a targeted mRNA 3' end sequencing approach (Figure 3). To this end, a set of gene-specific forward primers for second strand synthesis is designed taking the following considerations into account (for general considerations follow the manufacturer’s instructions of Lexogen’s QuantSeq-Flex Targeted RNA-Seq Library Prep Kit):

1. Define the primer binding site of your custom primers within 150-400 bp upstream of the 3' end of the transcripts of interest. Avoid amplicon regions with low T-content in order to be able to robustly evaluate T>C conversions across multiple T positions within the sequenced region.

2. Blast primer binding sites against the genome to avoid off-target priming.

3. If possible, check cell line-specific genomic DNA sequencing data for potential SNPs in the primer binding region.

4. Design DNA oligonucleotides that contain the following sequence information in 5 ’-to-3' direction: 5' - adapter sequence (CACGACGCTCTTCCGATCT); [optional: include random hexamer (NNNNNN) as unique molecular identifier]; specific transcript-binding region - 3' (for a list of example custom primers targeting a set of transcripts in mESC see Table 1).

5. Prepare 10x Custom Targeted Primer Mix solution by mixing 2 μl of each gene-specific forward primer (100μM stock). For second strand synthesis dilute the 10x Custom Targeted Primer Mix solution 1:10 to a final working concentration of 10 μM

Figure 3. Schematic overview of cDNA library preparation protocol for targeted (right) or untargeted (left) mRNA 3' end sequencing on Illumina platforms.

See text (section 3.3.4) for details.

Table 1. Gene-specific forward primer sequences for the targeted amplification of mRNAs after oligo(dT)-primed reverse transcription.

Gene	Ensembl ID	Primer sequence (DNA, 5'-3')
ActB	ENSMUSG00000029580	CACGACGCTCTTCCGATCTNNNNNNCCCACTCCTAA GAGGAGGATG
GusB	ENSMUSG00000025534	CACGACGCTCTTCCGATCTNNNNNNGTGAGAGGCT GGAGTGAAGG
Hspa1b	ENSMUSG00000090877	CACGACGCTCTTCCGATCTNNNNNNCAAACGTCTTG GCACTGTGT
Nat10	ENSMUSG00000027185	CACGACGCTCTTCCGATCTNNNNNNCTCTCCTGCTC CTCCCTTCT
Sox2	ENSMUSG00000074637	CACGACGCTCTTCCGATCTNNNNNNTGCAGGTTGAT ATCGTTGGT
Tnrc6b	ENSMUSG00000047888	CACGACGCTCTTCCGATCTNNNNNNTGAAGGGTGG TGTTTTCTCA

Targeted mRNA 3’ end libraries are prepared according to the following protocol (as a commercially available alternative use ‘QuantSeq-Flex Targeted RNA-Seq Library Prep Kit’ from Lexogen) :

• 6.Prior to reverse transcription, remove possible gDNA contaminants by DNase digest, e.g. using TURBO DNA-free^™ Kit according to manufacturer’s instructions. Ensure that the RNA is free of divalent cations and compatible with downstream RT-PCR, otherwise perform an additional clean-up step before setting up the reverse transcription reaction.
• 7.
  Prepare the reverse transcription reaction:

  • 5μl RNA (500 ng)
  • 4μl 5x First-Strand buffer (part of SuperScript III Reverse Transcriptase kit)
  • 0.5 μl 10 μM oligo(dT) primer (final concentration 0.25 μM)
  • 0.5 μl RNase OUT
• 8.Incubate for 5 min at 65°C and quickly cool samples on ice.
• 9.
  To each reaction add:

  • 1 μl 0.1M DTT
  • 1.25 μl dNTPs (10 mM)
  • 0.5 μl SuperScript III
  • 7.25 μl H₂O.
• 8.Incubate the RT-reaction in a thermal cycler for 10 min at 25°C, 50 min at 50°C and 15 min at 70°C.
• 9.Perform an RNase H digest according to manufacturer’s instructions.
• 10.
  Set up the gene specific second strand synthesis PCR in a final volume of 25 μl (primer sequences see Table 1 and 2):

  • 10 μl cDNA (after RNase H digest)
  • 12.5 μl 2x Taq MM (e.g. GoTaq® Green Master Mix)
  • 1 μl Custom Targeted Primer Mix solution (10 μM)
  • 1 μl oligo(dT) primer (10 μM)
  • 0.5 μl H₂O
• 11.
  Perform the PCR reaction with the following cycling conditions (note: cycle numbers may need to be optimized depending on transcript expression levels):

  95°C	2:00 min
  12 cycles of:		95°C	0:30 min
  		58 °C	0:30 min
  		72°C	0:30 min
  72°C	5:00 min
  4°C	Hold

• 12.Remove DNA fragments below 200 bp from the cDNA by purifying with AMPure XP beads according to manufacturer’s instructions. Elute cDNA in 20μL H₂O.
• 13.
  Prepare the following library amplification reaction in single ultraclear flat cap PCR tubes (primer sequences see Table 2):

  • 10 μl cDNA
  • 2.5 25 μl KAPA HiFi HotStart Real-time PCR Master Mix (2x)
  • 2.5 μl Illum. FWD primer (10 μM)
  • 2.5 μl Illum. REV index primer (10 μM)*
  • 10 μl H₂O
  • 50 μl final volume
    * Use a different Illum. REV index primer for each sample to uniquely barcode each library.
• 14.Thaw the KAPA Fluorescence Standards 3 and 4 for at least 15 min at room temperature, mix well by pipetting and transfer 50 μl per Standard into a ultraclear flat cap PCR tube.
• 15.
  Place the Fluorescence Standards and the library samples into a real-time PCR machine and amplify the libraries using the following cycling conditions:

  98°C	0:45 min
  98°C	0:15 min
  65 °C	0:30 min (annealing)	cycling*
  72°C	0:20 min (extension 1)
  Signal detection
  72°C	0:15 min (extension 2)

  ^*

  Monitor the PCR amplification curves in real-time. Interrupt the cycling during extension 2 as soon as the amplification curve of a sample has reached the exponential phase in a signal range between KAPA Fluorescence Standard 3 and 4, roughly corresponding to 3500 - 4500 RFU (Note 18). Take out the sample and transfer the tube to ice, then continue with the cycling program and repeat this step for the remaining samples.
• 16.Remove DNA fragments below 200 bp from the cDNA library by purifying with AMPure XP beads according to manufacturer’s instructions. Elute cDNA in 17μL H₂O.
• 17.Assess the library quality, e.g. on a capillary electrophoresis system like Fragment Analyzer (AATI, kit DNF-474).

Table 2. Primer sequences for the targeted mRNA 3' end library preparation.

Primer	Primer sequence (DNA, 5'-3')
oligo(dT)	GTTCAGACGTGTGCTCTTCCGATCT-(T)n-V
Illum. FWD	AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGC TCTTCCGATCT
Illum. REV index	CAAGCAGAAGACGGCATACGAGATNNNNNNGTGACTGGAGTTCAGA CGTGTGCTCTTCCGATCT

3.3.5. High-throughput sequencing

As a standard approach for SLAMseq datasets generated via mRNA 3' end sequencing, we typically recommend 10-20 million reads per QuantSeq library for an efficient quantification of s⁴U-containing transcripts of the whole transcriptome (>5000 transcripts). We recommend sequencing with single read 100 (SR 100) mode, which typically enables to recover the vast majority of labeled transcripts (>70% in mESC at a labeling efficiency of ~2.3 %)[13]. For targeted libraries, sequencing depth can be scaled down substantially depending on the number of transcripts amplified and the read mode depends on the T content downstream of the custom primer-binding site.

3.3.6. Data analysis

To pre-process the raw sequencing reads, demultiplex and adaptor trim (e.g. using Cutadapt) the reads. In addition, we recommend trimming the first 12 nt from the 5 ’ end of each read in QuantSeq and target mRNA 3' end sequencing datasets.

To align SLAMseq data generated by QuantSeq or targeted mRNA 3' end libraries, and to quantify T>C conversions in these data sets, we recommend using SLAMDUNK after pre-processing of the sequencing data. SLAMDUNK is a fully automated analysis pipeline based on the NGM mapper algorithm but is adapted to robustly quantify nucleotide-conversion-containing high-throughput datasets (Neumann et al., submitted, http://t-neumann.github.io/slamdunk/) [21]. Briefly, SLAMDUNK includes an alignment scoring scheme that does not penalize T>C conversions during the alignment step. In addition, SLAMDUNK is optimized to process 3' end sequencing data sets as it includes an optional multimapper reassignment strategy, which is specifically optimized for low-complexity regions such as the 3' UTRs of genes.

We recommend aligning the trimmed reads with SLAMDUNK using local alignment scoring and retaining alignments with a minimum identity of 95% and a minimum of 50% of the read bases mapped during the filtering step. To align QuantSeq libraries, a multimapping reassignment strategy can be applied to recover multimapping reads that map simultaneously to other low-complexity non-UTR regions in the genome [13]. SNPs can be filtered from the sequencing reads using a coverage cutoff of 10x and a variant fraction of 0.8 for sequencing data from haploid mouse embryonic stem cells. Adjust the variant fraction according to the ploidy of your cell line of interest. To call non-SNP overlapping T>C conversions with high confidence, apply a base quality threshold of Phred score > 26. SLAMDUNK determines the T>C conversion rate for each position along the custom defined amplicon intervals by normalizing to genomic T content and coverage of each position and averaged per position. For targeted RNAseq approaches, custom defined amplicon intervals can be defined by extracting the expected amplicon regions from the genome.

For QuantSeq data sets, tissue-specific 3' end annotations can be obtained by specifically extracting sequencing reads at the 3' end extending into the polyA tail and thereby assessing the polyadenylation sites (https://github.com/AmeresLab/UTRannotation) [13]. From these 3' end annotations, amplicon intervals extending 250 nt upstream from the custom-defined or annotation-based 3' end can be used to analyze T>C conversions in a mRNA 3' end-specific manner, as described previously [13].

To assess mRNA half-lives from T>C conversions, first subtract the background T>C conversions observed in the ‘unlabeled’ sample. Next, normalize the measured T>C conversion rate at a given time to the T>C conversion rate measured in the ‘0h chase’ sample. A single exponential decay model can be used to fit the normalized T>C conversions over time in order to obtain half-life information for each transcript. For representative examples of a SLAMseq mRNA half-life measurement in mESC using targeted or untargeted mRNA 3' end sequencing strategies, as described in this protocol, see Figure 4.

Figure 4. Transcript stability for the indicated genes in mESCs as determined by SLAMseq in untargeted (left, black) or targeted (right, red) mRNA 3' end cDNA libraries.

T>C conversion rates ± standard deviation of experimental duplicates were determined for each time-point of a s⁴U-pulse/chase experiment and fit to a single-exponential decay model to derive half-life (t½) in hours (h). Comparisons include each two examples for highly (>100 counts per million, cpm), medium (>10 cpm) or lowly (<10 cpm) expressed genes with each one stable (left) and one unstable (right) transcript. Sequencing data for untargeted (GSE99978) [13] and targeted (GSE118121) mRNA 3' end sequencing for the indicated genes is available at GEO.