Application of high-throughput 5′P sequencing for the study of co-translational mRNA decay

Summary mRNA degradation is connected to the translation process up to the degree that 5′-3′ mRNA degradation follows the last translating ribosome. To study 5′-3′co-translational mRNA decay and the associated ribosome dynamics, here we present an improved high-throughput 5′P degradome RNA sequencing protocol (HT-5Pseq). We exemplify its application in Saccharomyces cerevisiae, but in principle, it could be applied to any other eukaryotic organism. HT-5Pseq is easy, scalable, and uses affordable duplex-specific nuclease-based rRNA depletion. For complete details on the use and execution of this protocol, please refer to Zhang and Pelechano (2021).


MATERIALS AND EQUIPMENT
Filter sterilize the LET lysis buffer. Use RNAse-free water. Possible for long term storage at 20 C-25 C.

OPEN ACCESS
v. Vortex in vortex mixer for additional 2 min followed by centrifugation at 4 C for 2 min at 14,000 3 g. vi. Remove aqueous phase (approx. 450 mL) and add to Tube #1. Vortex 30 sec and spin 1 min at 14,000 3 g. vii. Remove aqueous phase and add to Tube #2. Vortex 30 sec and spin 1 min at 14,000 rpm. viii. Remove aqueous phase (approx. 400 mL) and add to Tube #3 mix well. Add 1 mL 95% (vol/vol) ethanol, mix and place at À20 C/À80 C, 30 min. ix. Collect RNA by centrifugation at 4 C for 20 min at 14,000 3 g.
CRITICAL: Phenol and Chloroform are acute toxic. Be careful when handling phenol and chloroform, always wear gloves, lab coat and perform manipulations following local safety regulations in fume hood.
CRITICAL: To avoid RNA degradation, perform RNA extraction with phenol:chloroform as fast as possible.
2. Check RNA quality by running a Bioanalyzer RNA gel or an agarose gel.
Note: Any alternative approach producing high-quality RNA may be used instead.

Timing: 45 min
Any contaminant DNA is removed from the sample.
3. Starting with 6 mg of total RNA, prepare the following mix and incubate the samples for 20 min at 37 C.
Note: It is possible to lower starting material to 500 ng in total. However, low input material usually decreases library complexity and increase the PCR duplicates.
4. Add 2 mL of TURBO DNAse inactivation reagent and incubate 5 min at 20 C-25 C (tapping once in a while). 5. Centrifugate at 14,000 3 g for 2 min at 20 C-25 C and transfer the supernatant to a clean precooled tube.
Note: This pellet is normally quite loose, repeat the centrifugation if it is resuspended and avoid the carryover of any DNase inactivation reagent. 6. Ethanol precipitate the DNA-free RNA by adding 2.5 volumes (with respect to the sample volume) of 95% (vol/vol) ethanol, a 1/10 volume of 3 M sodium acetate, 1 mL of glycoblue. Mix sample by gently inverting and incubate it for minimum 30 min at À20 C/À80 C.
Pause point: The ethanol precipitation can be left 16-18 h at À20 C/À80 C.
7. Centrifugate at 14,000 3 g for 30 min at 4 C to precipitate the RNA. 8. Wash the pellet with 500 mL of cold 70% (vol/vol). 9. Centrifugate 14,000 3 g for 10 min at 4 C. 10. Remove the remaining ethanol, air-dry pellet for 3 min and resuspend it in 1.8 mL of RNAse-free water.
Note: If the next step is the single-strand RNA ligation, RNA can be directly resuspended in ligation mix (step 11) and top up RNAse-free water to 10 mL .
CRITICAL: When handling with RNA samples, always keep them in RNase-free environment and place samples on ice.
11. Prepare a 10 mL reaction mix with the components listed below: CRITICAL: Add PEG provided by T4 RNA ligase 1 in the end of the reaction mix, as it is sticky at high concentration. Note: avoid over drying the beads as that might result in sample loss.

Reverse transcription
Timing: 1.5 h cDNA library is transcribed by the defined ratio of oligo-dT and random hexamer.
17. Prepare a reaction mix as the components listed below: CRITICAL: This optimized ratio of oligo-dT and random hexamer increase coverage in the 3 0 region of the gene. Altering the ratio of oligo-dT and random hexamer will lead to differences in the relative 5 0 /3 0 coverage.
18. Denature the sample at 65 C for 5 min. Then place on ice directly. 19. To each tube, add 6.2 mL of mixture containing the following components: 27. Set up the following 16 mL reaction: CRITICAL: The concentration of rRNA depletion probe mix is optimized for using 6 mg of total RNA as starting material. If the input RNA increases, the rRNA depletion probes should increase accordingly. The ratio between depletion probe mix and targeted molecules (cDNA) is around 2:1.
28. Denature sample for 2 min at 98 C using thermocycler. 29. Incubate the sample for 5 min at 68 C. 30. Add pre-warmed (2 min at 68 C) mix containing the following components:

Reagent Final Concentration Amount
Yeast rRNA depletion Probes (2 mM  Note: For pair-end 75 cycles sequencing in this case, using 60 bp for read1 and 15 bp for read2. Read2 will identify the molecule primed by either oligo-dT or random hexamer. In general, we recommend at least 6 million raw reads per yeast sample. Any alternative Illumina platform could be used instead of a NextSeq 500. Read sequencing length can be altered depending on the complexity of the genome of interest and the ability to uniquely map reads to the genome.

PCR Cycling Conditions
Steps

EXPECTED OUTCOMES
This protocol will generate sequencing libraries of 5 0 P mRNA degradation intermediates, detailed workflow is shown in Figure 1. In the final HT-5Pseq library, the average size is expected to be around 450 bp, including 150 bp Illumina adapter sequences ( Figure 2). The expected concentration of library can be 0.5-2 ng/mL. The sequencing depth required will depend on the library complexity and analysis requirement. By mapping HT-5Pseq reads to the reference genome, the expected results are as following: 1) reads coverage is distributed along the whole mRNA regions ( Figure 3); 2) rRNA contamination of HT-5Pseq library is less than 12%-20% ( Figure 4); 3) A clear 3-nt pattern can be observed with respect to specific codons, including start and stop codon at metagene level ( Figure 5); 4) Codon-specific/amino acid specific pausing can be extracted with respect to specific codons.

QUANTIFICATION AND STATISTICAL ANALYSIS
Here we provide a potential bioinformatic pipeline for 5 0 Pseq data.
De-multiplex raw data using the indexing information: Using bcl2fastq (v2.20.0) for base-calling. We recommend allowing 1 mismatch in index 1 and 1 mismatch in index2. Trim sequencing adaptor: Use cutadapt V1.16 to trim sequencing adapter (-a AGATCGGAAGAG CACACGTCTGAACTCCAGTC). Extract UMI: Use UMI-tools (v0.5.4) to extract 8-nt random barcodes on the 5 0 ends of reads. These UMI information will be used to remove PCR duplicates. Align sequencing reads: Use star/2.7.0 (Dobin et al., 2013) to align 5 0 -end reads to reference genome (SGD R64-1-1 for S. cerevisiae genome). For mapping the 5 0 -ends reads to the genome, we recommend using the parameter -alignEndsType Extend5pOfRead1 to exclude soft-clipped bases on the 5 0 end. Remove PCR duplicates: Use UMI-tools (v0.5.4) to remove duplicated 5 0 ends of read introduced by PCR during library preparation.

OPEN ACCESS
Quantify transcripts: Use Subread package (featureCounts) (Liao et al., 2014) to count mRNA, tRNA, rRNA and snRNA and snoRNA transcripts. Use DESeq2 packages from R and Bioconductor (Love et al., 2014) to perform differential gene expression analysis. Analysis 5 0 ends positions: Use Fivepseq package to map 5 0 ends with respect to start, stop codon and codons at metagene level (Nersisyan et al., 2020).

LIMITATIONS
Although HT-5PSeq offers high quality degradome information at a fraction of the costs and with significantly decreased hands-on time in comparison with standard 5PSeq (Pelechano et al., 2015), this approach has several limitations that need to be accounted for.
Firstly, the main limitation is that HT-5Pseq approach focus on the subpopulation of 5 0 end of mRNA undergoing decay. Therefore, any exposed 5 0 P end of molecule can be captured independent of their relationship with co-translation degradation process.
Next, HT-5Pseq measures the kinetics competition of 5 0 -3 0 degradation machinery and ribosome, therefore we do not recommend this approach to directly measure absolute translation rates.  However, we have shown that the last translating ribosome during co-translational decay can infer the general ribosome dynamics (Pelechano et al., 2015).
Thirdly, as the abundance of 5 0 P end molecule depends on both translation and mRNA stability, any factor involved in those process can affect the observed 5 0 P seq profile. For example, when investigating mRNA degradation profiles in xrn1D the ribosome associated 3-nt pattern is greatly decreased (Pelechano et al., 2015). In xrn1D the observed 5 0 P profiles reflects a combination of transcription start site mapping (as expected from the exposed 5 0 P after decapping) complemented by other endonucleolitica cleavages events. In addition, HT-5Pseq libraries may vary in library complexity as a result of the variation of fractions on mRNA degradation intermediates present in a sample in respect to the total RNA. For example, HT-5PSeq libraries from xrn1D cells are in general more complex, as 5 0 P mRNA degradation intermediates are not efficiently removed and thus represent a higher proportion of the total RNA population. To control for this, we add UMI during the RNA ligation step. If a lower fraction of mRNA degradation intermediates is expected, we recommend increasing the amount of total RNA starting material.

TROUBLESHOOTING
Problem 1

RNA degradation
Potential solution 1) When handling with RNA samples, always keep RNA on the ice. 2) Check the RNA integrity of RNA extraction.
3) Use RNase inhibitor during the protocol (steps 11 and 19) and aliquot RNase-free reagents. 4) Perform RNA extraction with phenol-chloroform (step 1) as fast as possible.
Problem 2 Low yield DNA library (step 45)

Potential solution
1) If starting RNA is less, increase the starting RNA amount.
2) Inefficient removal of RNA template (step 22), this will loss the cDNA library after DSN treatment. Optimize RNA removal steps (step 22). 3) Increase few PCR cycles in PCR amplification steps (step 38). Final PCR cycles should be less than 20 cycles. 4) For bead cleanup, do not over dry the beads. That might lead to sample loss.

Problem 3 Large number of rRNA reads
Potential solution 1) DNA leftover in RNA sample can potential saturate DSN enzymatic activity (step 27). This may decrease the rRNA depletion efficiency. Perform DNase treatment to RNA samples (step 3). 2) Optimize DSN for rRNA depletion if using custom depletion probes (step 27-36). Optimize mix annealing temperature (steps 29-30) based on the melting temperature of newly designed probes. Mix samples with probes by pipetting and keep them on the thermocycle at the selected temperature.

Problem 4
Large number of PCR duplicates

Potential solution
3) Increase the input RNA material to increase the complexity. 4) Decrease the final PCR cycles (from step 38) that will increase the useful reads.

Problem 5
Biased 5 0 P reads to 3 0 end Potential solution 5) If the 5 0 P reads biased towards to 3 0 , decrease the usage of oligo-dT in reverse-transcription (step 17) to get more homogenous distribution profile.

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Vicent Pelechano (vicente.pelechano.garcia@ki.se) .

Materials availability
This study did not generate new unique materials nor reagents.

Data and code availability
The raw and processed sequencing data are deposited at GEO with accession number GSE152375.