Generation of Sequencing Libraries for Structural Analysis of Bacterial 5′ UTRs

Summary The structure of 5′ untranslated regions (5′ UTRs) of bacterial mRNAs often determines the fate of the transcripts. Using a dimethyl sulfate mutational profiling with sequencing (DMS-MaPseq) approach, we developed a protocol to generate sequence libraries to determine the base-pairing status of adenines and cytosines in the 5′ UTRs of bacterial mRNAs. Our method increases the sequencing depth of the 5′ UTRs and allows detection of changes in their structures by sequencing libraries of moderate sizes. For complete details on the use and execution of this protocol, please refer to Ignatov et al. (2020).


SUMMARY
The structure of 5 0 untranslated regions (5 0 UTRs) of bacterial mRNAs often determines the fate of the transcripts. Using a dimethyl sulfate mutational profiling with sequencing (DMS-MaPseq) approach, we developed a protocol to generate sequence libraries to determine the base-pairing status of adenines and cytosines in the 5 0 UTRs of bacterial mRNAs. Our method increases the sequencing depth of the 5 0 UTRs and allows detection of changes in their structures by sequencing libraries of moderate sizes. For complete details on the use and execution of this protocol, please refer to Ignatov et al. (2020).

BEFORE YOU BEGIN
A number of approaches, collectively known as Structure-seq, combine chemical probing of RNA structure with high-throughput sequencing. These methods allow study of RNA structures inside living organisms at the whole-transcriptome level to characterize their ''RNA structurome'' (Bevilacqua et al., 2016). DMS (dimethyl sulphate) reagent can be used to probe RNA structures and protein:RNA interactions inside living cells (Wells et al., 2000). DMS selectively methylates N1 of adenine and N3 of cytosine if these nucleotides are not base-pairing or interacting with proteins (Inoue and Cech, 1985;Lempereur et al., 1985). The modified bases can be detected by incorporation of non-complementary nucleotides during cDNA synthesis. These approaches, termed MaP for ʹMutational profilingʹ, calculate DMS reactivity rates for each nucleotide as the percentage of mutations in the cDNA library (Smola et al., 2015). After normalization, the DMS reactivity rates can serve as a measure of the base-pairing status of individual nucleotides (Homan et al., 2014;Siegfried et al., 2014;Zubradt et al., 2017). In this protocol, we described details of using DMS-MaPseq to generate sequence libraries for global structural analysis of bacterial mRNAs.
Gel Purification of 5 0 RNA Adapter (3 0 RA) Timing: 8 h Note: The 5 0 RA is an RNA adapter. To prevent self-ligation, it has 3 0 -OH and 5 0 -OH ends. However, due to incomplete synthesis or degradation, truncated variants can be observed. These truncated versions are prone to self-ligation and formation of long RNA molecules that can 4. Add 300 mL of 300 mM NaCl solution to the crushed gel in the 2 mL tube and incubate for 2 h at 50 C and shaking at 300 rpm. 5. Transfer the gel slurry to the Costar Spin-X column (Merck) and centrifuge for 3 min at 10,000 3 g to remove the gel pieces. 6. Precipitate RNA by adding 2 ml GlycoBlue coprecipitant (ThermoFisher) and 900 mL 100% EtOH, incubating at 20 C for 1 h and spinning at 20 000 g for 30 min at 4 C. 7. Resuspend the purified RNA adaptor in 20 mL RNase-free water, measure its concentration on Nanodrop spectrophotometer (ThermoFisher) and adjust the concentration to 15 mM.
Note: For the 1 nmol input of RNA adaptor expect to recover 0.4 to 0.6 nmol of the purified RNA adaptor (25-40 mL of 15 mM purified RNA adaptor).

RESOURCE AVAILABILITY
Lead Contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Jö rgen Johansson (Jorgen.johansson@umu.se).

Materials Availability
This study did not generate new unique reagents.

Data and Code Availability
This study did not generate any unique datasets or code. CRITICAL: Acrylamide/Bis-acrylamide and b-mercaptoethanol pose a significant health risk. Refer to your country guidelines on working with these chemicals.

STEP-BY-STEP METHOD DETAILS
Growth of L. monocytogenes Bacteria -Days 1-3 Timing: 30 min (day 1) Note: The protocol was developed to determine the structure of 5 0 UTRs in the human bacterial pathogen Listeria monocytogenes, but could in principle be applied to any other bacterial species.
1. Using an inoculation loop, plate L. monocytogenes cells on a Brain Heart Infusion (BHI) agar plate. Grow for 16 h (overnight) in a 37 C incubator.
Timing: 30 min (day 2) 2. The next day pick a colony and transfer it to 20 mL of BHI broth. Grow for 16 h (overnight) in a water incubator at 37 C and 150 rpm shaking.

Timing: 3-5 h (day 3)
3. The next day transfer 0.25 mL of the overnight culture to 25 mL of prewarmed BHI medium (dilution 1:100) and grow until the mid-logarithmic growth phase at 37 C and 150 rpm shaking.
Note: Not only BHI, but also LB broth and chemically defined media (Whiteley et al., 2017) can be used to grow L. monocytogenes. Depending on the medium type and source, the genetic background of bacteria and equipment used to measure OD, the growth dynamics and OD values can vary considerably. Therefore, prior to the main experiment it is necessary to grow bacteria and plot the growth curves. Based on these data, optimal OD to perform dimethyl sulphate (DMS) treatment can be determined. Also, it is important to have the necessary volume of bacteria for DMS treatment to get sufficient RNA for further experiment. In this protocol we perform DMS treatment of L. monocytogenes in the mid-logarithmic growth phase (Ignatov et al, 2020).

DMS Treatment -Day 3
Timing: 1 h (day 3) 4. Prewarm 15 mL tube to 37 C on the incubator. Using a prewarmed Pasteur pipette rapidly transfer 5 mL bacterial culture to the 15 mL tube.
CRITICAL: Do not allow the culture to cool down. Temperature is one of the main parameters affecting RNA structure. Furthermore, many bacteria possess very sensitive systems to detect temperature decrease and activate the cold shock response, which might affect the experiment results.
5. Using a 1 mL tip transfer 150 mL DMS to 5 mL of the culture (final DMS concentration 3%). Adjust the pipette sampling volume to 500 mL and using the same tip, pipette the culture up and down 10 times. Incubate the mixture for 3 min. Every 30 s pipette the mixture up and down.

OPEN ACCESS
STAR Protocols 1, 100046, September 18, 2020 CRITICAL: DMS and b-mercaptoethanol are highly toxic and a special protective equipment should be used when working with these chemicals. Therefore, refer to your country safety regulations when working with DMS and b-mercaptoethanol.
Note: DMS does not dissolve in water solutions completely and without thorough mixing, DMS accumulates at the bottom of the tube.
6. Inactivate DMS by adding 6 mL Quenching solution and vigorously shaking the tube. 7. Collect bacteria by centrifugation at 5,000 3 g for 10 min at +4 C. Wash the pellet with 5 mL ice cold 30% b-mercaptoethanol solution. Pulse centrifuge and completely discard the supernatant. Immediately proceed to RNA isolation [Problem 1].

RNA Isolation -Days 3-4
Timing: 2-3 h (day 3) Note: L. monocytogenes is a Gram-positive bacterium with a thick cell wall. To isolate RNA the cell wall is disrupted by mechanical grinding followed by acid guanidinium thiocyanatephenol-chloroform extraction. For other bacterial species the cell lysis protocols can be different.
8. Dissolve the bacterial pellet in 400 mL ice-cold Cell Disruption solution. 9. Transfer the cell suspension to the pre-cooled 2 mL screw-cap tube containing 500 mL of 0.1 mm zirconium beads (Biospec) and 500 mL of acid phenol (pH 4.5). Keep the tubes on ice.
CRITICAL: Phenol, TRI reagent and Chloroform are highly toxic and a special protective equipment should be used when working with these chemicals. Therefore, refer to your country safety regulations when working with Phenol, TRI reagent and Chloroform.
Note: It is important to keep cells at low temperature during all preparations before addition of TRI reagent. This helps to minimize the potential activity of cellular RNases.
10. Disrupt the cells on a BeadBeater 8 (Biospec) cell disruptor at maximum speed for 30 s.
Alternatives: other cell disruptors can be used instead of BeadBeater. As an example, we successfully tested Fastprep-24 (MP Biomedicals).
11. Centrifuge the disrupted cells at 12,000 3 g for 5 min at +4 C. Collect the upper aqueous phase and transfer to a new 2 mL tube. 12. Add 1 mL of TRI reagent solution (ThermoFisher) and vigorously shake the tube by hands. Incubate the tube on the bench for 5 min. 13. Add 200 ml of Chloroform, vigorously shake the tube by hands and incubate on ice for 5 min. 14. Centrifuge the tube at 12,000 3 g for 5 min at +4 C. Collect the upper aqueous phase and transfer to a new 2 mL tube. Add 500 mL of Chloroform and vigorously shake the tube by hands. 15. Centrifuge the tube at 12,000 3 g for 5 min at +4 C. Collect the upper aqueous phase and transfer to a new 1.5 mL tube. Add isopropanol at 0.73 volume of the aqueous phase. Mix by vortexing and incubate at À20 C for 1 h or longer.
Pause Point: at this step the sample can be stored for up to one week at À20 C.
Timing: 2 h (day 4) ll OPEN ACCESS 16. Precipitate RNA by centrifuging the tube at 20,000 3 g for 30 min at +4 C. Carefully discard the supernatant and wash the pellet with 500 mL of 80% ethanol. 17. Centrifuge the tube at 20,000 3 g for 5 min at +4 C and completely discard the supernatant. Dry the RNA pellet for 10 min on the bench and dissolve in 50 ml of RNase free water. 18. Measure the RNA concentration on nanodrop and check its quality on agarose gel [Problem 2] [Problem 3].
Pause Point: The isolated DMS modified RNA can be stored at À20 C for a month or at À80 C for up to 6 months.
RNA Treatment with DNase I, 5 0 Polyphosphatase and Ribosomal RNA Depletion -Day 4 Timing: 6-8 h (day 4) Note: Ribosomal RNA constitutes up to 90% of the RNA content in bacterial cells. Therefore, to obtain a sufficient coverage of other transcripts it is necessary to remove it.
In our original work we used Ribo-Zero rRNA Removal Kit (Gram-Positive Bacteria) (Illumina). However, since then the kit has been discontinued by Illumina. As a substitute we suggest riboPOOL kit (siTOOLs Biotech); RiboMinus depletion kit (K155014, Thermofisher) or Ribo-Zero Plus rRNA depletion Kit (Illumina) but none of these kits have been tested by us.
Note: To get enough RNA for subsequent steps of library preparation we started with 10 mg of DMS-treated total RNA. 19. Treat 10 mg of total RNA sample with DNase I (Roche) for 30 min at 37 C. 20. Purify RNA on RNeasy MinElute (Qiagene) columns according to the protocol from the manufacturer, elute from the column with 15 mL of RNase free water. 21. Remove g and b phosphates from 5 0 -triphosphorylated RNAs by treatment with RNA 5 0 polyphosphatase (Epicentre): 22. Purify RNA on RNeasy MinElute columns according to the protocol from the manufacturer, elute with 15 mL of RNase free water. 23. Deplete ribosomal RNA using appropriate kit (see above). Adjust the volume of RNA solution to a final volume 20 mL with RNase free water.
Pause Point: The sample depleted of ribosomal RNA can be stored at À20 C for up to one week or at À80 C for up to one month.
Ligation of the 5 0 Adaptor, Fragmentation and Size Selection -Day 5 Note: From this step of the protocol we suggest introducing a negative control sample (K-): the 5 0 RA ligation reaction set up without bacterial RNA. All subsequent steps should be the same for the samples and the negative control. After purification of Illumina cDNA library, the negative control should not have any products visible on electrophoresis. This guarantees that the cDNA libraries represent the sequences of bacterial RNAs and not any artefacts.
26. Ethanol precipitate RNA with ligated 5 0 RA a. Add 70 mL of RNase free water, 10 mL of 3 M sodium acetate (pH 5.0), 2 mL of GlycoBlue coprecipitant and 300 mL of 100% ethanol. Vortex and incubate for 1 h at À20 C. b. Precipitate RNA by centrifuging the tube at 20,000 3 g for 30 min at +4 C.
Carefully discard the supernatant and wash the pellet with 500 mL of 80% ethanol. c. Centrifuge the tube at 20,000 3 g for 5 min at +4 C and completely discard the supernatant. Dry the RNA pellet for 10 min on the bench and dissolve in 9 mL of RNase free water.
Note: After depletion of ribosomal RNA, the quantity of RNA decreases significantly and for ethanol precipitation it is necessary to use a coprecipitant.
27. Perform RNA fragmentation with RNA fragmentation reagents (ThermoFisher): a. Add 1 mL of the 103 Fragmentation Buffer, mix and spin briefly. b. Incubate at 70 C for 3.5 min in a thermocycler with the heating lid. c. Add 1 mL of the Stop Solution, mix, spin briefly and place the sample on ice.
Note: The fragmentation parameters were empirically chosen to fragment L. monocytogenes RNAs to lengths of approximately 70-350 nucleotides (nts). If using another fragmentation method or bacterial species, it is advisable to calibrate the fragmentation conditions.
28. Resolve RNA on 6% denaturing polyacrylamide gel and isolate the fragments having lengths from 125 to 400 nts ( Figure 1): b. Mix the fragmented RNA with 12 mL of 23 RNA loading dye (ThermoFisher), heat the mixture at 70 C for 10 min and chill on ice. c. Apply the RNA sample to the gel in parallel with the RiboRuler Low Range RNA ladder (ThermoFisher) and run until the bromophenol blue dye migrates for 3 cm.
Note: There is no need to extensively separate the RNA-fragments on the gel. Even after a short migration time, the non-ligated adapters and abundant tRNA species will be separated from the mRNA fragments. The short migration also decreases the size of the excised fragment, thereby decreasing the gel volume and increasing the recovery efficiency.

d. Stain the gel with SYBR gold and visualize on a blue light transilluminator.
Excise the gel fragments with the lengths between 125 and 400 nucleotides.
Note: the 6% PAA gel is very fragile and should be treated with care.
e. Disintegrate the gel fragments. With a hot needle punch $0.5 mm hole in the bottom of a 0.5 mL tube. Place the gel fragments in the 0.5 mL tube and insert it into a 2 mL tube. Centrifuge for 10 min at > 10 000 g. f. Add 300 mL of 300 mM NaCl solution to the crushed gel and incubate for 2 h at 50 C and shaking at 300 rpm. g. Transfer the gel slurry to the Costar Spin-X column and centrifuge for 3 min at 10,000 3 g to remove the gel pieces. h. Precipitate RNA by adding 2 mL GlycoBlue co-precipitant and 900 mL 100% EtOH, incubating at 20 C for 1 to 16 h and spinning at 20 000 g for 30 min at 4 C. i. Resuspend the purified RNA fragments in 10 mL RNase-free water.
Pause Point: This RNA sample can be stored at À20 C for a week.  c. Add 70 mL of RNase free water, 10 mL of 3 M sodium acetate (pH 5.0), 2 mL of GlycoBlue coprecipitant and 300 mL of 100% ethanol. Vortex and incubate for 1 h at À20 C. d. Precipitate RNA by centrifuging the tube at 20,000 3 g for 30 min at +4 C. Care fully discard the supernatant and wash the pellet with 500 mL of 80% ethanol. e. Centrifuge the tube at 20,000 3 g for 5 min at +4 C and completely discard the supernatant. Dry the RNA pellet for 10 min on the bench and dissolve in 20 mL of RNase free water. 34. Remove non-ligated 3 0 DA using RNAClean XP Kit. Follow the Protocol provided by the manufacturer but with the following modifications: 1) use the 1.6 3 volume of magnetic beads to sample ratio (i.e. add 32 mL of beads to 20 mL of sample) 2) at the final step, elute the purified RNA from beads with 30 mL of RNase free water.

Ligation of the 3 0 Adaptor -Day 6
Pause Point: This RNA sample can be stored at À20 C for a week.

Reverse Transcription Using TGIRTIII Enzyme -Day 7
Timing: 5 h (day 7) 35. Concentrate RNA solution on a SpeedVac Concentrator to 4.5 mL. 36. Perform reverse transcription with TGIRTIII enzyme (InGex) (Mohr et al., 2013;Zubradt et al., 2017): a. Mix 4.5 mL of the RNA sample with 1 mL of 1 mM RT primer and 2 mL of 53 Reverse transcription buffer. Heat for 2 min at 80 C using the thermocycler with the heated lid. Transfer the tube to the room temperature ($23 C) and allow RT primer to anneal for 5 min. b. Add the other components to set up the final reaction: Note: Use freshly prepared 0.1M dithiothreitol (DTT). c. Incubate reaction at 57 C for 2 h in a thermocycler with a heated lid. d. Add 1 mL 5 M NaOH, pipette up and down to mix, and incubate at 95 C for 3 min. 37. Ethanol precipitate the synthesized first strands of cDNA: a. Add 160 mL of RNase free water, 20 mL of 3 M sodium acetate (pH 5.0), 2 mL of GlycoBlue co-precipitant and 600 mL of 100% ethanol. Vortex and incubate for 1 h at À20 C. b. Precipitate the cDNA first strands by centrifuging the tube at 20,000 3 g for 30 min at +4 C. Carefully discard the supernatant and wash the pellet with 500 mL of 80% ethanol. c. Centrifuge the tube at 20,000 3 g for 5 min at +4 C and completely discard the supernatant. Dry the pellet for 10 min on the bench and dissolve in 20 mL of nuclease free water. Pause Point: The first strands of cDNA can be stored at À20 C for a week or at À80 C for longer period.

Amplification of cDNA Libraries -Days 7-8
Note: The amplification is performed by two rounds of PCR with several clean up steps to deplete the short DNA fragments, mostly represented by primers and adaptor dimers.
Note: If you plan to sequence several DMS-MaPseq libraries with a single run on Illumina sequencing machine, carefully pick the LibAmp_RPIXX_R primers for amplification. The variable parts of these primers correspond to reverse complement of Illumina index sequences. To check the compatibility of different index adapters with each other, refer to Index Adapters Pooling Guide from Illumina.

Timing: 4 h (day 7)
38. Run the first round of PCR to amplify cDNA libraries.
a. Set up the PCR reaction: b. Run the PCR reaction on a thermocycler with the heated lid: 39. Purify the PCR product using AMPure XP beads. Use 1:1 volume ratio of AMPure XP beads and PCR reaction (i.e. mix 25 mL of beads with 25 mL of PCR reaction). Elute from the beads with 40 mL of nuclease-free water.
Note: the decreased ratio of beads to samples allows to deplete not only the PCR primers, but also very short cDNA fragments and adaptor dimers. b. Run the PCR reaction on a thermocycler with the heated lid: Note: It is advisable to first calibrate the number of PCR cycles (Figure 2A). The product of the second PCR should be collected at the exponential phase of amplification to minimize amplification biases. Therefore, first perform a calibrating PCR by collecting 5 mL of PCR products from different cycles (e.g. 4, 6, 8, 10 and 12) and resolve them on 1.5% agarose gel along with GeneRuler 50 bp DNA ladder (ThermoFisher). Find the optimal cycles generating sufficient product for further work, but where the concentration has not yet reached saturation. In our experience, 4 to 6 cycles of the second PCR round is sufficient to get enough cDNA for Illumina sequencing.
Timing: 3-5 h (day 8) 41. Purify the PCR product using the RNeasy MinElute kit. Follow the protocol (i.e. do not add increased volume of ethanol to the mixture of PCR reaction and RLT buffer). At the final step elute from the column with 20 mL of nuclease-free water. 42. Purify the PCR product using AMPure XP beads. Use 1:1 volume ratio of AMPure XP beads and PCR reaction (i.e. mix 20 mL of beads with 20 mL of PCR product). Elute with 20 mL of nucleasefree water. 43. Measure the concentration of PCR product (library) on nanodrop. Resolve 5 mL of the library on 1.5% agarose or native PAA gel ( Figure 2B) to study the distribution of the lengths of the fragments. For more precise determination of the lengths of the fragments use Agilent bioanalyzer with Agilent DNA 1000 Kit ( Figure 2C).
Note: At this step the negative control should not have any visible DNA sequences.
Note: We recommend to clone the cDNA library (e.g. using the CloneJET PCR Cloning Kit from ThermoFisher) and sequence the inserts from 5 to 10 clones. This will ensure that both the 5 0 and 3 0 sequencing adapters are present in the fragments. Also, it will be possible to estimate the percentage of modified adenines and cytosines. Normally, 4%-6% of adenines and cytosines should be DMS modified. After mapping to the genome sequence, the modified bases can be identified as mismatches in cDNA sequences [Problem 4].
Note: To obtain sequencing depth sufficient for comparison of 5 0 UTR structures of highly expressed L. monocytogenes mRNAs we sequenced the libraries comprising 10 million reads. However, to compare the structures of moderately and low expressed transcripts it would be necessary to increase the sequencing volume to 50-100 million of reads per replicate.

EXPECTED OUTCOMES
Accurate prediction of RNA secondary structures by DMS-MaPSeq approach (Zubradt et al., 2017) requires generation of cDNA libraries of good quality. This protocol presents a fast and reliable method to generate cDNA libraries with an increased sequencing depth for bacterial 5 0 UTRs. Such libraries can be used for sequencing and subsequent structural determination and are identical The depletion of ribosomal RNAs decreases their representation in the library to less than 25% and the size selection step (transcripts shorter than 100 nt are removed) depletes tRNAs. The majority of reads thus represent mRNAs and small RNAs, including highly abundant RNase P and 4.5S RNA (Figure 3B). Since the 5 0 adaptors are ligated to RNA fragments having the g and b phosphates removed prior to RNA fragmentation, the 5 0 ends of bacterial mRNAs have increased sequencing coverage ( Figure 3C).

QUANTIFICATION AND STATISTICAL ANALYSIS
Guideline for the Analysis of 5 0 UTR Targeted DMS-MaPseq Data Note: the base pairing status of an adenine or cytosine nucleotide can be calculated as a ratio between the number of mismatches of cDNA sequence at that nucleotide and its transcriptional coverage. For DMS-MaPseq this ratiometric measure of base pairing does not require background correction (Zubradt et al., 2017). However, the parallel sequencing of DMS untreated sample can be useful to identify endogenous mRNA modifications (Zubradt et al., 2017).

OPEN ACCESS
Note: The analysis of DMS-MaPseq data requires writing custom scripts in Python programming language. For that we suggest using the HTseq framework that contains multiple tools for processing of RNA-seq data (Anders et al., 2015).
1. Map the Illumina reads to Listeria monocytogenes EGD-e genome (NC_003210) with Bowtie 2 aligner (Langmead and Salzberg, 2012) using the -end-to-end -very-sensitive mode. This results in generation of SAM files recording the coordinates of aligned reads and mismatches encoded in cDNAs ( Figure 4).
2. Using the SAMtools software (Li et al., 2009) convert SAM files to BAM files. Visualize BAM files with a genome browser, for example Integrative Genomics Viewer (Thorvaldsdottir et al., 2013). An alternative way to visualize transcriptional profile is to calculate the transcriptional coverage for each nucleotide programmatically and visualize the generated profile. 3. For further analysis the coordinates of the 5 0 ends of mRNAs are necessary. For some species the coordinates of transcriptional start sites (TSS) can be obtained from literature: for L. monocytogenes EGDe the TSS coordinates were retrieved from (Wurtzel et al., 2012). In principle, the coordinates of TSSs can be determined using the data from the 5 0 UTR targeted DMS-MaPseq experiment itself. This can be done programmatically by searching for the steep increase of transcriptional coverage upstream of each start codon. Afterwards the identified TSSs should be manually checked using the transcriptional profile visualized in the genome browser. Please note however that our approach does not discriminate between the primary TSSs and the processed 5 0 ends generated by the action of cellular RNases. To obtain the primary TSSs we suggest using the differential RNA-seq approach (Borries et al., 2011). 4. Define the mRNA regions for further analysis. In our case the genome regions corresponding to non-coding RNAs, 5 0 UTRs and the first 30 nucleotides of coding sequences were selected for further analysis. 5. Use the HTseq framework to extract the coordinates of mismatches from the CIGAR string of the SAM files. 6. For each adenine and cytosine nucleotide of the selected regions, calculate the coverage as the number of reads mapped at that position and the number of mismatches in the mapped cDNAs. 7. For each nucleotide calculate the rate of mismatches by dividing the number of mismatches by the transcriptional coverage.
Note: This ratiometric measure of base-pairing status depends on the degree of DMS modification of RNA. Therefore, the conditions of DMS treatment should be very reproducible. However, in some cases, for example when DMS treatment is performed at different temperatures, this is not achievable. In this case, we suggest normalization of the data. For that the average rate of mismatches in a sample should be calculated. The division of a nucleotide mismatch rate by the average rate in the sample will generate the normalized measure of base pairing status that we termed ''DMS values'' (Ignatov et al., 2020).
Note: The obtained data on base-pairing status of adenines and cytosines can further be used to model RNA secondary structure and perform comparison of RNA structures in different samples. These further steps can be variable and we consider their description to be beyond the topic of our protocol. As a primer for further analysis we refer the reader to the review by Choudhary et al. (Choudhary et al., 2017).

LIMITATIONS
The protocol has been designed to increase the sequence coverage of 5 0 UTRs. This decreases the proportion of reads from other parts of the transcript. If planning to obtain structural information of whole transcripts, rather use a whole-transcriptome DMS-MaPseq protocol.
Note: It is possible to use Cap-Clip Acid Pyrophosphatase (Cambio #C-CC15011H) to hydrolyze the pyrophosphate bonds of the 5 0 -terminal m7GpppG "cap" of eukaryotic messenger RNAs. This decapping method allows to adapt our protocol for eukaryotic cells. However, the protocol was developed for bacteria and has not been comprehensively tested in eukaryotes.

Problem 1
The bacterial pellet is lost after DMS quenching.

Potential Solution
The Listeria pellet is rather loose after DMS quenching and this also might be observed if using other bacterial species. It is therefore critical to run a pilot experiment and if necessary, increase the volume of bacterial culture to facilitate the precipitation of bacteria.

Problem 2
Low RNA yield.

Potential Solution
The library preparation requires at least 10 mg of high quality DMS-treated RNA. We obtained this quantity from 5 mL of mid-exponential L. monocytogenes culture. However, for other bacteria or growth conditions it may be necessary to increase the volume of the culture for DMS treatment. Another reason for low yield might be the inefficient cell lysis. When cells are disrupted on a Bead beater we noticed that the geometry of tubes for disruption can be important. Therefore, never use 1.5 mL screw-capped tubes, but only the 2 mL screw-capped tubes.

Problem 3
Low RNA quality on agarose gel.

Potential Solution
There can be several reasons for the compromised RNA quality. When DMS decomposes in water, sulfuric acid is generated and in the absence of a proper buffer, the pH can drop significantly. In this protocol, we used broth buffered with 50 mM MOPS pH 7.3. However, depending on DMS concentration and the time of treatment, it might be necessary to increase MOPS concentration to 100 mM. Another reason for RNA degradation might be the activity of bacterial or external RNases. Therefore, it is critical to use RNase free reagents and labware and always keep the samples on ice.

Problem 4
Low amount of modified bases.

Potential Solution
The fraction of modified adenines and cytosines should generally lie between 4% and 6%.
DMS activity depends on temperature, time of incubation and mixing regime. Therefore, for a new experimental setup it is advisable to first perform a series of pilot experiments with different incubation times and/or DMS concentrations in the range of 2%-5%.