Structure-seq of tRNAs and other short RNAs in droplets and in vivo

Abstract

There is a multitude of small (<100 nt) RNAs that serve diverse functional roles in biology. Key amongst these is transfer RNA (tRNA), which is among the most ancient RNAs and is part of the translational apparatus in every domain of life. Transfer RNAs are also the most heavily modified class of RNAs. They are essential and their misregulation, due to mutated sequences or loss of modification, can lead to disease. Because of the severe phenotypes associated with mitochondrial tRNA defects in particular, the desire to deliver repaired tRNAs via droplets such as lipid nanoparticles or other compartments is an active area of research. Here we describe how to use our tRNA Structure-seq method to study tRNAs and other small RNAs in two different biologically relevant contexts, peptide-rich droplets and in vivo.

1. Introduction

RNA serves diverse structural and functional roles in extant biology and likely did so on early Earth. Many such functional RNAs are relatively small (<100 nt), and folding domains of larger RNAs are often on this scale as well. Such short RNAs include tRNA, ribozymes, and riboswitches [1]. These RNAs adopt compact tertiary structures that respond to changes in cellular conditions including different concentrations of Mg²⁺ and metabolites [2], [3]. In addition, confinement and covalent modification of RNA can lead to different folding outcomes [4], [5]. This is especially true in the case of transfer RNA (tRNA), which is the most heavily modified RNA on a per nucleotide basis. Transfer RNAs are essential to all domains of life, and their misregulation due to misfolding, mutated sequences, and change in modification levels can lead to disease [6]–[8]. This is particularly manifested in mitochondrial tRNAs which are hypomodified to start with [9], [10]. Transfer RNAs are also used as therapeutics in the delivery of repaired tRNAs to cells via droplets such as lipid nanoparticle or other compartments [11], [12].

We and others have developed techniques to probe RNAs using next generation sequencing (NGS). In this approach, multiple RNAs are probed either in vivo or in vitro with a diverse array of chemicals, which are read out by either reverse transcription (RT) stops or via mutational profiling (MaP) with select reverse transcriptases [11], [12]. Here we describe how to use our tRNA Structure-seq method [13] to study tRNAs and other small RNAs in two different biologically relevant contexts of peptide-rich droplets and in vivo.

Our lab and the Assmann lab developed Structure-seq to obtain single-nucleotide resolution of RNA structure in vivo [14]. This was originally developed with DMS-treated poly(A)-purified plant mRNAs that were reverse transcribed with random hexamers and RT stops were recorded [15]. We also applied the method to DMS-treated rRNA-depleted mRNAs in bacteria where RT stops were recorded [16], [17]. Most recently, we applied Structure-seq to DMS-treated tRNA from bacteria where the modification was read out by MaP [13], [18]. Common to these approaches is the avoidance of the highly abundant rRNA, by isolating the RNAs of interest via poly(A) pulldown (mRNA), by depletion of the rRNA (mRNA), or by PAGE purification (tRNA). In the case of tRNA, there were additional challenges including the potential for loss of information at the 3′-end with conventional RT, high intrinsic RNA structure, and abundant RNA modifications. These were overcome by ligating an adapter to the 3′-end of deacylated tRNAs using T4 RNA ligase II truncated and by using an ultra-processive RT, Marathon RT with the addition of 2 mM Mn²⁺ that can read through RNA structure and modifications while inserting mutations. A general approach moving forward is to combine these strategies of depleting nuisance RNAs and enriching for the RNAs of interest by physical isolation methods such as pulldowns, PAGE, and liquid chromatography. Depending on the application, an adapter can be added to the 3′-end of the RNA to preserve information in the very 3′-end. Reverse transcription can be from the adapter, a common sequence in the RNAs of interest such as poly(A), or a random hexamer.

Herein, we describe the experimental protocols that we used to examine the structure of short structured RNAs in droplets and in vivo. We focus on tRNA, but the ideas can readily be applied to other short RNAs, under a myriad of conditions including purified individual RNAs or complex libraries of sequences, in vitro or in vivo, and in compartments or free in solution. We give an overview of tRNA Structure-seq and then go on to discuss handling reagents, preparing and end-labeling RNA with radioactive and fluorescent probes, performing RNA accumulation studies, and conducting tRNA Structure-seq in droplets and in vivo. We also include a detailed pipeline on analyzing the tRNA Structure-seq output.

2. Materials

2.1. General equipment

1. pH probe

2. Micropipettors and tips

3. 0.2 mL and 1.7 mL polypropylene tubes

4. Vortex mixer

5. Microfuge

6. Thermocycler

7. Dry bath with PCR tube inserts

8. Glass plates for casting polyacrylamide gels (6.5 × 9.5 × 0.125 in)

9. Apparatus for PAGE

10. Power source for PAGE, with the ability to vary voltage (0–3000 V), amperage (0–100 mA) and wattage (0–120 W)

11. 50 mL conical tubes

12. Hand-held UV light, shortwave (254 nm)

13. Fluor-coated thin layer chromatography (TLC) plate for UV-shadowing

14. Tube revolver for 1.7 mL and 50 mL tubes

15. Benchtop centrifuge for 1.7 mL tubes with the capability to spin at 17,000 x g

16. SpeedVac vacuum concentrator

17. Nanodrop or other small volume UV-vis spectrophotometer

18. Heat block with 1.7 mL insert

19. Floor centrifuge for 50 mL conical tubes

20. Geiger counter(s)

21. Liquid scintillation counter

22. Gel dryer

23. Phosphor storage cassettes and screens

24. Phosphorimager

25. Confocal microscope

26. Bioshaker

2.2. General materials

1. RNase-free deionized water

2. DNA templates for tRNAs with T7 promotor site (see [13])

3. Forward PCR primer (see [13])

4. Reverse PCR primer (see [13])

5. Phusion HF DNA polymerase

6. 5x Phusion HF buffer

7. 10 X Deoxyribonucleoside 5′ triphosphate (dNTP) mix (10 mM each of dATP, dCTP, dGTP and dTTP)

8. TriTrack 6 X DNA loading dye

9. Low molecular weight DNA ladder

10. 100 bp DNA ladder (Invitrogen, 15628019)

11. Non-denaturing polyacrylamide solution (40% acrylamide solution (acrylamide:bis-acrylamide = 29:1))

12. 10X Tris-Borate-EDTA (TBE) (900 mM Tris, 900 mM borate, and 100 mM EDTA pH 8.0)

13. 10% w/v Ammonium persulfate (APS) (prepare fresh from the solid or store at 4 °C up to 3 months)

14. Tetramethylethylenediamine (TEMED)

15. Teflon spacers (1.0–1.5 mm thick) (we cut these with a razor blade from sheets)

16. 1X gel running buffer (1X TBE)

17. Vacuum grease

18. Plastic wrap

19. Permanent marker for marking up gel

20. SYBR-GOLD (10,000X concentrate in DMSO)

21. Crush and soak buffer (10mM Tris pH 8.0, 1 mM EDTA, and 250 mM sodium chloride)

22. Sodium Acetate (3 M pH 5.2)

23. GlycoBlue coprecipitant

24. 100% ethanol (chilled to −20 ºC)

25. BsaI HFv2 Restriction Endonuclease

26. 10X rCutSmart Buffer

27. 10X Transcription buffer (400 mM Tris, pH 7.5, 40 mM dithiothreitol, 20 mM Spermidine)

28. 1.0 M MgCl₂

29. Ribonucleoside 5′ triphosphate (rNTP) mix: 25 mM each of rATP, rCTP, rGTP, and rUTP purchased as solids and prepared in-house in deionized water, neutralized to pH 7 with Tris-base

30. T7 RNA polymerase, either commercial or prepared in-house

31. Denaturing gel solution: 8.3 M urea, 20 % acrylamide (29:1 acrylamide : bis-acrylamide) in 1X TBE

32. Denaturing gel diluent: 8.3 M urea in 1X TBE

33. 2X Formamide Loading Dye (2X FLD): 95% formamide, 0.1X TBE, 20 mM EDTA, 0.025% Bromophenol blue)

34. 0.2 μm filters

35. 10 mL syringes

36. 18 and 20 gauge needles

37. Phosphatase enzyme (rSAP-recombinant Shrimp Alkaline Phosphatase)

38. T4 Polynucleotide Kinase

39. 10X T4 Polynucleotide Kinase Buffer

40. [γ−³²P] ATP: 9000 Ci/mmol total ATP) (Perkin Elmer, NEG035C005MC)

41. Room temperature lead pig for holding source material

42. Plexiglass shields

43. Light box

44. Beta-blocking 0.2 and 1.7 mL tube racks

45. Glogos II Autorad Markers (Agilent, 420201)

46. X-ray film

47. Developer/Fixer solution

48. Staining trays

49. Scintillation fluid (Research Products International, 111198)

50. Drawing paper to lift gels off of plates and prepare for drying (14 × 17 in)

51. Polycation: Oligolysine (10mer, 30mer solids purchased from Alamanda Polymers)

52. 10 X Polycation solution (10 mM for 10mer and 3.3 mM for 30mer)

53. Polyanion: Oligoaspartic acid (10mer, 30mer solids purchased from Alamanda Polymers)

54. 10 X Polycation solution (10 mM for 10mer and 3.3 mM for 30mer)

55. 10 X Salt buffer (150 mM KCl and 5–100 mM MgCl₂)

56. Dimethyl Sulfate (5 M solution diluted in ethanol) (very hazardous) See cautions below

57. 10X DMS Reaction buffer (100 mM Tris pH 7.5, 150 mM KCl, and either 5 mM or 100 mM MgCl₂)

58. Dithiothreitol powder (DTT). Stored at −20 °C, with desiccant

59. Quenching solution (2.5 M dithiothreitol)

60. Guanidinium Hydrochloride (6 M)

61. Coacervate Loading dye (66 mM EDTA, 8.3 mM poly(acrylic acid) 1.8 kDa, 30% formamide, 0.05% (w/v) bromophenol blue sodium salt)

62. Mth RNA ligase

63. 10X DNA adenylation buffer

64. Adenosine triphosphate (1 mM)

65. 5′-p-DNA Ligation adapter (see [13])

66. T4 RNA Ligase 2 truncated

67. 10X T4 RNA Ligase Reaction Buffer

68. Polyethylene glycol (PEG) 8000 (50% w/v)

69. Reverse Transcriptase enzyme (SuperScript III (Invitrogen, 18080093), SuperScript II (Invitrogen, 18064014), TGIRT-III (InGex, TGIRT50), or Marathon (Kerafast, EYU007)

70. MnCl₂ (20 mM, prepared in-house, and 50 mM purchased)

71. 2X Reverse transcription buffer (100 mM Tris (pH 8.3), 400 mM KCl, 10 mM dithiothreitol, 40% w/v glycerol)

72. Reverse transcription primer (see [13])

73. CircLigase II ssDNA Ligase

74. Betaine (5 M)

75. 10X CircLigase buffer

76. Forward indexing primer (see [13])

77. Unique reverse indexing primers (see [13])

78. Nucleospin Gel and PCR Clean-up kit

79. RNA Clean and Concentrator Kit

80. LB medium

81. Escherichia coli BW25113 strain

82. TRIzol reagent

83. Isopropyl alcohol

84. Indexing primers (see [13])

85. Sodium chloride (1 M)

86. Microscope cover slips (22 × 22 mm, VWR)

87. Sodium periodate (100 mM, freshly prepared)

88. Sodium acetate (1 M, pH 4.0)

89. Hydrazide-functionalized fluorophores (CF488A, or CF647)

90. Sodium cyanoborohydride (200 mM, freshly prepared)

91. Amicon Ultra-0.5 mL centrifugal filter unit (10 kDa cutoff)

92. Linux PC

93. Image J (https://imagej.nih.gov/ij/download.html)

94. miniconda (https://docs.conda.io/en/latest/miniconda.html)

95. Cutadapt (https://cutadapt.readthedocs.io/en/stable/)

96. ShapeMapper2 (https://github.com/Weeks-UNC/shapemapper2)

97. RNAstructure (https://www.urmc.rochester.edu/rna/)

98. tRNA_structure-seq (https://github.com/Ryota-Yamagami/tRNA_structure-seq)

3. Methods

3.1. Overview of tRNA Structure-seq

In this chapter, we provide detailed protocols for tRNA Structure-seq in droplets (Figure 1A) and in vivo (Figure 1B). The tRNA Structure-seq protocol consists of in vitro transcription of tRNA or growth of cells, DMS reaction in droplets and in vivo, library preparation, and NGS. In our previous reports, we performed the Structure-seq experiments in vitro and in vivo (in E. coli), and in complex coacervate conditions to evaluate tRNA structure [13], [18]. The protocol is robust and can be applied to other small RNA species and other organisms.

Figure 1: Workflows for tRNA Structure-seq experiments A) in droplets and B) in vivo.

A) In-droplet experiments start by (1) forming the coacervate droplets. Then (2) the tRNA is added to the droplets and incubated for 10 min to allow it to accumulate in the droplets. (3) The DMS is added and (4) allowed to react for 1 min before being centrifuged for 2 min. (5) The DMS is allowed to react for 2 more min before (6) the dilute phase is carefully pipetted off and placed into a waste container and quenched with DTT. Next, DTT, guanidinium hydrochloride (GuHCl) and 2X formamide loading dye (2X FLD) are added to the droplets (7) before being PAGE purified. (8) Finally the DMS-modified tRNA is ready to be used in library preparation for NGS. B) In vivo tRNA Structure-Seq begins by (1) growing E. coli to an OD_600nm of 0.8 at 37 °C. (2) Then, the cells are probed with DMS for 5 min at 37 °C. (3) The DMS is then quenched with DTT and translation is halted with chloramphenicol (CHL). (4) The cells are centrifuged to pellet them, the supernatant is decanted, and (5) then a TRIzol extraction is performed. (6) Then, the tRNA fraction is purified by PAGE. (7) Finally, the DMS-treated tRNA is used in library preparation for NGS.

3.2. Preparation of reagents

1. To generate RNA(s) for DMS probing experiments in droplets and in vitro, we suggest preparing them by in vitro transcription. DNA templates for in vitro transcription can be purchased from Integrated DNA Technologies (IDT) as single templates with a T7 promoter site, or as oligopools containing tens to thousands of sequences with a T7 promoter site. Here we used them to prepare about 50 tRNAs at once. In our experience, oligopools require PCR amplification to produce enough material for transcriptions. To amplify an entire oligopool, we design common forward and reverse PCR primer sites into each oligonucleotide at the 5′ and 3′-ends (Figure 2A), see [13] for the sequences. Because the forward primer is upstream of the T7 promoter site, it is not incorporated into the RNA. However, the reverse primer site will be incorporated into the RNA, so we typically design a BsaI restriction site at the 5′-end of the reverse primer site such that the reverse primer can be removed after PCR amplification without a scar. This is important because conventional restriction enzymes leave extra nucleotides at the 3′-end of the RNA that have the potential to base pair with the desired short RNA and interfere with its fold and therefore function. This is especially problematic with tRNA, where one desires the authentic CCA 3′-end. We suggest gel purifying templates after each step in Figure 2 to ensure that only the desired material you want is carried on into the sequencing experiments. Before attempting tRNA Structure-seq experiments in droplets, RNAs should be either 5′ or 3′-end radio- or fluorescent-labeled to determine some of the physical properties of the droplets [19]–[22]. Once these controls have been established, tRNA Structure-seq experiments can be performed.

Figure 2: Oligopool design and workflow.

A) Single-stranded DNA oligopools for tRNAs that contain common forward and reverse PCR primer sites, a T7 promoter site, and a BsaI cutsite. B) After performing PCR, dsDNA oligopools are produced that are then be cleaved by BsaI such that the reverse PCR primer is removed before transcription. C) After the BsaI cleavage reaction, the dsDNA oligopool is in vitro transcribed with T7 RNA Polymerase. D) After transcription, the tRNAs are ready to be used in DMS chemical probing experiments.

3.3. PCR amplification of single-stranded DNA oligopools

1. DNA templates for the RNA(s) of interest should be ordered as either single-stranded DNA oligopools with forward and reverse PCR primer sites and a T7 promoter site (convenient when multiple RNAs will be studied simultaneously as libraries of up to thousands of sequences can be ordered in this way) or as an individual single-stranded DNA oligonucleotide with a T7 promoter site from a vendor such as Integrated DNA Technologies (IDT) or Twist Bioscience. Single-stranded DNA templates in oligopools must be PCR amplified to have enough material for in vitro transcription (Figure 2B).

2. Set up the following PCR reaction with single-stranded-DNA oligopools.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	66	-
   5X Phusion HF Buffer	5X	20	1X
   Forward Primer	10 μM	5	0.5 μM
   Reverse Primer	10 μM	5	0.5 μM
   dNTPs	10 mM	2	0.2 mM
   Template DNA	100 nM	1	1 nM
   Phusion HF DNA Polymerase	2 U/μL	1	2 U
   Total	-	100	-

   In brief, add reagents in the order listed above to a 0.2 mL tube. Run the following thermocycler program: Initial denaturation at 98 °C for 5 sec. Anneal at 50 °C for 10 sec. Elongate at 72 °C for 20 sec, then repeat steps 1–3 for 20–30 cycles. Final Elongation 72 °C for 3 min (See Note 4.1).

3. While the PCR is running, pour a 10% non-denaturing polyacrylamide gel. Prepare 6.5 × 9.5 × 0.125 in glass plates with 1.5 mm Teflon spacers, and use wide-toothed combs so that the wells are big enough for 120 μL of sample. Then, prepare 150 mL of non-denaturing gel solution by mixing 37.5 mL of 40% acrylamide solution, 15 mL of 10X TBE and 97.5 mL of water. To cast the gel, add 1.5 mL of 10% (w/v) Ammonium Persulfate solution and 150 μL of TEMED to the acrylamide solution. Swirl after each addition to prevent rapid local polymerization. Pour the gel and let it polymerize for 10–20 min.

4. Add 20 μL of TriTrack 6X DNA loading dye to the PCR products.

5. Load the PCR products, two lanes of Low Molecular Weight DNA Ladder, and one lane of 100 bp ladder onto the 10% non-denaturing gel (see Note 4.2). Run the gel with limiting voltage at 180 V for 4–6 h. Low voltage is used to keep the gel to prevent denaturation of the dsDNA because when there are multiple DNA species the pairs may anneal improperly. Periodically inspect the electrophoresis apparatus for buffer leaks. Apply vacuum grease if needed to the leaking areas, being sure to turn off the power supply before touching the apparatus.

6. In the dark, prepare 50 mL of SYBR-GOLD staining solution by adding 5 μL of 10,000 X SYBR-GOLD to 50 mL of 1X TBE in a 50 mL conical tube. Limit exposure to light as this can bleach the fluorophore.

7. Take the gel down, carefully separate the glass plates, cut off any excess gel with a razorblade, place a piece of plastic wrap over the gel, and label the lanes with a permanent marker. Then, remove the other plate so that the gel is exposed with the plastic wrap underneath it.

8. In the dark, pour an appropriate volume of SYBR-GOLD stain on top of the exposed gel to completely cover it and allow it to stain for 5 min. Carefully absorb the residual liquid with some paper towels, and dispose of them in solid hazardous chemical waste. Place another piece of plastic wrap on top of the stained gel.

9. Wearing UV-blocking eye protection and a lab coat, use a handheld UV-wand to visualize the bands on the gel and mark them with a permanent marker. Bands should appear bright yellow if the gel is on a dark background.

10. Excise the full-length PCR product band (~ 130 bp for tRNAs) with a new razorblade and place in a 1.7mL tube with 1 mL of crush and soak buffer. To extract the DNA from the gel slice, spin it in a tube revolver overnight at 4 °C.

11. Briefly centrifuge to collect the liquid, then carefully pipette ~300 μL of the liquid into each of three 1.7 mL tubes. To each tube, add 1 μL of GlycoBlue coprecipitant to help visualize the pellet, 40 μL of 3 M sodium acetate (pH 5.2), and 1 mL of −20 °C 100% ethanol. Place the tubes in either powdered dry ice or a −80 °C freezer for 1.5 h. Centrifuge the tubes at 17,000 x g for 30 min, and decant the liquid into new 1.7 mL tubes. Dry in a SpeedVac for ~20 min to evaporate any remaining ethanol.

12. Resuspend in 50 μL of water and use a microvolume UV-vis spectrophotometer monitoring at 230, 260 and 280 nm to determine the DNA concentration and its purity.

3.4. BsaI Restriction endonuclease treatment

1. To facilitate amplification of DNA oligopools, we include common forward and reverse PCR primers in each oligonucleotide. The forward primer site is positioned before the T7 promoter site, and therefore is not transcribed, but the reverse primer is transcribed. As mentioned above, to remove the reverse primer and leave the template with a “scarless” end, we use BsaI, a type IIS restriction endonuclease that cleaves outside of its binding site (Figure 2).

2. Set up the following BsaI reaction.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Purified dsDNA	6,000–8,000 ng	49	100–160 ng/μL
   rCutSmart buffer	10X	6	1X
   BsaI HFv2	20 U/μL	4	80 U
   Total	-	~60	-

   Add reagents in the order listed above to a 0.2 mL tube. Incubate for 16 h at 37 °C.

3. The next day, prepare an 8% non-denaturing polyacrylamide gel as per Section 3.3 step 3.

4. Add 12 μL of TriTrack 6X DNA loading dye to the BsaI reaction.

5. Refer to Section 3.3 steps 5–9 to run and stain the gel with SYBR-GOLD. Outline the full-length and the BsaI-cut fragment with a permanent marker. The BsaI-cut fragment should be below the full-length dsDNA, because it is ~20 nt shorter.

6. Excise only the BsaI-cut fragment and avoid cutting out any remaining full-length dsDNA. Full-length dsDNA will appear in the downstream sequencing data wasting reads.

7. Follow steps 10–11 in Section 3.3 to crush and soak, and dry down the sample. Resuspend in 100 μL of RNase-free water and measure the concentration and purity of the BsaI-cut dsDNA with a microvolume UV-vis spectrophotometer. This material is now prepared for in vitro transcription in the next section.

3.5. In vitro transcription of RNA using hemi-duplexed DNA

1. In our lab, there are two types of transcriptions that we typically perform: one from a hemi-duplexed DNA template (a template strand with an ~20 nucleotide T7 promoter oligonucleotide annealed to it) that we discuss in this section, or from dsDNA templates generated by PCR that we discuss in the next section. Here we provide a protocol for the production of a single RNA species by in vitro transcription from hemi-duplexed DNA. Anneal a T7 promoter (see [23]) to the template before the start of transcription. To do that, set up the following annealing solution.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	45.5	-
   Sodium Chloride	100 mM	5.5	10 mM
   Template DNA	100 μM	2	4 μM
   T7 Promoter oligo	100 μM	2	4 μM
   Total	-	~55	-

   Add reagents in the order listed above to a 1.7 mL tube. Place this tube in a heat block at 95 °C for 3 min. Then let it cool to room temperature for 10 min. Quick spin the tube in a microfuge to collect the solution at the bottom of the tube. This can then be added to the transcription mixture in the next step.

2. Set up the following transcription.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	277.5–324.5	-
   Transcription Buffer	10X	50	1X
   rNTP mix	25 mM	60	3 mM
   MgCl2	1 M	12.5	25 mM
   Annealed Template DNA (from previous step)	4 μM	50	400 nM
   T7 RNA Polymerase	-	5–50	1–10% (w/v) (see Note 4.3)
   Total	-	500	-

   Add the reagents in the order listed above to a 1.7 mL tube. Mix thoroughly by pipetting or gently vortexing. Incubate the transcription for 2–6 h at 37 °C. To increase RNA yield, additional volumes of T7 RNA polymerase can be added every 1.5 h.

3.6. In vitro transcription of RNA using BsaI-cleaved dsDNA oligopools

1. Here we provide a protocol for the production of ~50 different RNAs from an oligopool by in vitro transcription from dsDNA templates. BsaI-cleaved dsDNA oligopools should be transcribed without reannealing.

2. Set up the following transcription.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	227.5–272.5	-
   Transcription Buffer	10X	50	1X
   rNTP mix	25 mM	60	3 mM
   MgCl2	1 M	12.5	25 mM
   BsaI-cleaved dsDNA oligopool	-	100	-
   T7 RNA Polymerase	-	5–50	1–10% (w/v)
   Total	-	500	-

   Add the reagents in the order listed above to a 1.7 mL tube. Mix thoroughly by pipetting or gently vortexing. Incubate the transcription for 2–6 h at 37 °C. To increase RNA yield, additional volumes of T7 RNA polymerase can be added every 1.5 h.

3.7. Purification of in vitro transcribed RNA

1. Prepare gel plates as in Section 3.3 step 3. Then, prepare 150 mL of gel solution by mixing 75 mL of denaturing gel solution with 75 mL of denaturing gel diluent to give a 10% denaturing gel. To cast the gel, add 1.5 mL of 10% (w/v) ammonium persulfate solution and 150 μL of TEMED to the acrylamide solution. Swirl after each addition to prevent rapid local polymerization. Pour the gel and let it polymerize for 10–20 min. Polymerization of leftover acrylamide solution can be used to estimate when the gel is fully polymerized.

2. After the gel has polymerized fully, set up the gel apparatus, and pre-run the gel for 30–60 min at 30 W.

3. Add an equivalent volume of 2X Formamide loading dye to the transcription reaction. Load the gel and run it with limiting watts at 30 W, until the bromophenol blue dye is ~2.5 cm from the bottom of the glass plates.

4. Disassemble the gel as in Section 3.3 step 7 and place a second piece of plastic wrap on the exposed side of the gel.

5. Then either turn the lights off or go to a darkroom to UV shadow the gel. Place the plastic wrap sandwiched gel onto a fluor-coated thin-layer chromatography plate, and while wearing UV protective eye protection and a lab coat, use a handheld UV wand to shadow the gel. The RNA band should appear as a thick dark shadow, which can be marked with a permanent marker. Minimize the time of UV light exposure of the gel to prevent UV crosslinking of the RNA.

6. Excise the RNA band with a new razor blade, cut it into 1 × 1 mm squares, and place them into a 50 mL conical tube.

7. Add crush and soak buffer until the gel slices are completely covered by solution and there is 1–2 mL of clear liquid on top of the gel slices. This varies with volume of the gel slices but is typically 7–15 mL.

8. Spin the 50 mL conical tube in a tube revolver overnight at 4 °C.

9. Briefly centrifuge to collect all of the gel slices at the bottom of the tube and then carefully draw up the liquid with a 10 mL syringe equipped with an 18 gauge needle. Once all of the liquid has been drawn into the syringe, discard the needle into sharps waste and use a lure locking 0.2 μm syringe filter to filter the liquid into a new 50 mL conical tube.

10. Add 5 μL of GlycoBlue coprecipitant, 1/10 of the recovered liquid′s volume of 3 M sodium acetate, pH 5.2, and 3 volumes of −20 °C 100% ethanol. Carefully mark a spot on the bottom of the tube with a permanent marker, and line that up with a mark on the lid to know where to look for the pellet. Because of the large volume of liquid used to extract the RNA from the gel slices, we suggest performing two ethanol precipitations the first to reduce the volume to a more manageable and more concentrated volume and the second to recover the RNA.

11. Place the tube in either dry ice or a −80 °C freezer for 1.5 h.

12. Orient the tubes such that the mark on the lid is on the outside of the rotor, and centrifuge at 10,000 x g for 1 h at 4 °C. Decant the supernatant into a fresh 50 mL conical tube and save it until the RNA has been successfully recovered.

13. Invert the tubes onto a paper towel for 1–5 min to allow any residual liquid to drain off/dry, being careful to not dislodge the pellet.

14. Resuspend the pellet in ~500 μL of crush and soak buffer, and pipette it into a new 1.7 mL tube.

15. Add 1 μL of GlycoBlue, 50 μL of 3 M sodium acetate pH 5.2, and 1 mL of −20 °C 100% ethanol. Place the tube in either dry ice or a −80 °C freezer for 1.5 h.

16. Spin at 17,000 x g for 30 min at 4 °C. Decant the supernatant into a new 1.7 mL tube being careful not to dislodge the pellet.

17. Dry the pellet in a SpeedVac for ~20 min.

18. Resuspend the RNA in 100 μL of RNase-free water and check its concentration and purity on a microscale UV-vis spectrophotometer.

3.8. Removal of 5′ triphosphates for radiolabeling

1. After the preparation of RNA by in vitro transcription, there is a 5′-triphosphate. This is incompatible with 5′-end labeling reactions, which require a 5′-hydroxyl group. Therefore, the 5′-triphosphate must be removed before the RNA can be 5′-end labeled. We typically use rSAP (recombinant Shrimp Alkaline Phosphatase) because it is heat inactivatable and inactivation of the phosphatase is critical for success of downstream kinase reactions. However, other phosphatases such as Calf Intestinal Phosphatase (CIP) are also suitable, although these reactions generally require phenol chloroform purification as phosphatases like CIP are not easily inactivated.

2. Set up the following phosphatase reaction.

   Reactant	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	12	-
   RNA	50 μM	3	7.5 μM
   rCutSmart Buffer	10X	2	1X
   rSAP	1 U/μL	3	3 U
   Total	-	20	-

   Add water and RNA to a 0.2 mL tube. Heat to 95 ºC for 1 min, then cool to room temperature for 10 min. Add rCutSmart buffer and rSAP and pipette the mixture thoroughly. Incubate at 37 ºC for 1 h. Heat to 65 ºC for 5 min to inactivate rSAP. This RNA can now be used in 5′-end labeling experiments without further purification.

3.9. Radioactive ³²P 5′-end labeling

1. To judge whether an RNA accumulates in a droplet, it is useful to have a detectable label such as a fluorophore or radioisotope like ³²P. It is important to consider which label is most appropriate. For example, radiolabeling avoids changing the chemical properties of the RNA, can be detected at pM or lower concentrations, and is not sensitive to the environment (e.g., viscosity, pH, polarity, etc.) unlike many fluorescent dyes. Fluorophores, on the other hand, can directly be used to estimate the concentration of an RNA in a droplet, simultaneously imaging the morphology of coacervate droplets, and allow for spatio-temporal experiments such as Fluorescence Recovery After Photobleaching (FRAP).

2. Whenever working with radiation it is essential to take all necessary safety precautions like wearing dosimetry, lab coats, double gloves, and eye protection, making sure that there is proper shielding for the type of radiation, and have a Geiger counter nearby to check for contamination. While working with radiation remember the acronym ALARA As Low as Reasonably Achievable. Radioactive material should always be behind appropriate shielding, workers should keep as much distance as possible from the radiation source (radiation decreases at a rate of 1/r² where r is the distance between the worker and the source), and limit the time that the radioactive material is being worked with. Don′t rush, but also be mindful to work as quickly and as safely as possible. Never work with radiation alone. Be sure to refer to and follow the comprehensive health and safety guidelines from your institution.

3. Before working with radioactive material, scan the work area with a Geiger counter to check for contamination, and place a new piece of bench paper onto that workspace. Acquire a glass radiation waste jar, place a plastic waste bag in it, and cover it with a plexiglass lid. Obtain 3–4 plexiglass shields, placing three of them on the bench. Place a beta-blocking 0.2 mL tube rack behind one shield. Cover a light box with plastic wrap, and place the fourth shield on top of it. Place a beta-blocking 1.7 mL tube rack on top of the light box with a labeled 1.7 mL tube for each RNA you are labeling. Place a room temperature plastic shelled lead storage container (lead pig) behind one of the shields to hold the source vial while it thaws. Keep a log of when radioactive material is used that has a sign-in, sign-out procedure and use it before starting radioactive work. This allows regulators to follow whenever radioactive material was used and helps to ensure that material is not lost or stolen.

4. Using a plexiglass shield, take the lead pig containing the [γ−³²P] ATP out of the −20 °C freezer, and behind one of your shields carefully transfer the source vial from the cold pig to the room temperature pig. Before closing the room temperature lead pig, turn the lid of the source vial ¼ turn to break the seal and equalize pressure with ambient to prevent any expulsion of radioactive droplets after thawing. Allow the source material to thaw for 30 min. Using a shield, put the cold pig back in the −20 °C radiation freezer.

5. Prepare an appropriate percentage acrylamide gel for your RNA. We used a 10% denaturing polyacrylamide gel (see Note 4.4). Pre-run the gel at 20 W for 30 min.

6. Set up the following 5′-end radiolabeling reaction.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   rSAP-treated RNA	~7 μM	6	~5 μM
   10X T4 PNK buffer	10X	1	1X
   T4 Polynucleotide Kinase (PNK)	10 U/μL	1	10 U
   γ−32P ATP	150 μCi/μL	2	300 μCi
   Total	-	10	-

   Add the first three reagents as listed to a 0.2 mL tube. Once the γ−³²P ATP has thawed, add it to the kinase reaction, and pipette mix it thoroughly. Incubate the kinase reaction for 30–60 min at 37 °C.

7. Add an equivalent volume of 2X formamide loading dye to the kinase reaction. Load the gel and run it with limiting watts at 20 W for 30–90 min (See Note 4.4).

8. Prepare a phosphorimager cassette to transport the gel by placing plastic wrap over it.

9. After the gel has run for an appropriate amount of time, turn off the power source and disassemble the gel apparatus. Remove one of the glass plates, carefully place the gel in the transportation cassette cover the gel with the plastic wrap and use a shield to transport the cassette to a darkroom.

10. Before turning off the lights in the darkroom, place three Glogos II Autorad Markers onto the periphery of the gel and expose them light for 2–3 min to allow them to luminesce.

11. Turn off the lights, take out a piece of X-ray film, open the cassette, place the x-ray film on top of the gel, close the cassette, and expose for 3–5 min.

12. While this is exposing, set out three staining trays, one for developer solution, one for water, and one for fixer solution. Add each solution to each staining tray.

13. After the X-ray film has been exposed, remove it from the cassette, and dip it into the developer for 10–15 sec, carefully agitating it to evenly coat the X-ray film. Then let it drip dry briefly before dipping it into the water for several seconds. Once any remaining developer is washed off, carefully hold it up to the red light to see if the bands and Autorad markers are visible. If the bands are visible, dip the film into the fixer solution for several seconds to stop the development, then rinse it off with the water and leave it to dry. If the bands were not visible, place the film back into the developer for 1–2 more min.

14. Put away all light sensitive equipment such as X-ray film and developer and fixer solutions then turn on the light and transport the gel to the lightbox prepared in step 3.

15. Place the dry X-ray film on the light box and then standing behind the shield, carefully place the gel on top of the X-ray film, aligning the Autorad markers to their spots on the x-ray film.

16. Once the gel is aligned, take a new razorblade, excise the labeled RNA band, and place it in a 1.7 mL tube in a beta-blocking tube rack.

17. Add 1 mL of TEN250 crush and soak buffer to the gel slice and spin it in a tube revolver behind shielding overnight at 4 °C.

18. Follow steps 10–11 in Section 3.3 to crush and soak and dry down the sample. Resuspend the RNA pellet in 50 μL nuclease-free water.

19. To measure the labeled RNA concentration, add 1 μL of the labeled RNA to a scintillation vial containing 10 mL of liquid scintillation fluid. Place this vial and a blank vial into a liquid scintillation counter (LSC) and measure the counts per minute. The cpm/μL of the RNA can be used to estimate its concentration. See [23] for a thorough explanation of how to calculate the labeled RNA concentration.

3.10. 3′-end fluorescent labeling

1. The gold standard for determining whether an RNA is inside of a droplet is a combination of fluorescent and radiometric methods. In our hands, the easiest way to fluorescently label an RNA is by 3′end ring opening and reacting the terminal cis-diol with a hydrazide functionalized fluorophore (Figure 3).

2. Set up the following ring opening reaction.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	250	-
   RNA	-	100	-
   Sodium periodate	100 mM	100	20 mM
   Sodium acetate, pH 4.0	1 M	50	100 mM
   Total	-	500	-

   To a 1.7 mL tube, add reagents as listed above. Let the reaction proceed at room temperature for 1 h.

3. Pipette the reaction into a 0.5 mL 10 kDa Amicon Ultra centrifugal filter unit to remove sodium periodate.

4. Centrifuge at 17,000 x g for 5 min, or until there is roughly 100 μL of solution remaining in the filter then decant the filtrate from the collection tube. The RNA should be in the centrifugal filter unit.

5. Add 400 μL of water and centrifuge again at 17,000 x g for 5 min. Repeat four more times for a total of five 5-fold dilutions.

6. Determine the concentration of the oxidized RNA by microvolume UV-vis spectrophotometer. The RNA is now prepared for the fluorophore coupling reaction.

7. To label the RNA, set up the following fluorophore coupling reaction.

   Reactant	Initial Concentration	Volume (μL)	Final Concentration
   Oxidized RNA	-	35	-
   Sodium acetate pH 4.0	1 M	5	100 mM
   CF488A hydrazide	4 mM	10	800 μM
   Total	-	50	-

   Add reagents as listed above to a 1.7 mL tube and mix by vortexing. To prevent photobleaching of the fluorophore, cover the tube with aluminum foil or perform the reaction in the dark. Incubate at 37 °C for 30 min.

8. Add 10 μL of freshly prepared 200 mM sodium cyanoborohydride and incubate for another 30 min at 37 °C.

9. Centrifuge at 17,000 x g for 5 min, or until there is roughly 100 μL of solution remaining in the filter then decant the filtrate from the collection tube. The RNA should be in the centrifugal filter unit still.

10. Add 400 μL of water and centrifuge again at 17,000 x g for 5 min. Repeat nine more times for a total of ten dilutions.

11. Determine the RNA concentration by microvolume UV-vis spectrometry and take an absorbance spectrum to detect the fluorophore.

12. To determine the labeling efficiency, calculate the label concentration over the total RNA concentration. Use the Abs at 488 nm to solve A= ebc where e is the molar absorptivity, A is the Abs at 488 nm, and c is the concentration in M. Typical yields are between 25–50% labeling efficiency.

Figure 3: General Scheme of periodate ring opening of the 2′−3′ cis diol.

(1) followed by hydrazide dye (green ball) coupling (2) and reduction with Na[BH₃(CN)] (3) to stabilize the covalent linkage between the dye and RNA.

3.11. Radiation-detected accumulation studies

1. It is imperative to know whether the RNA of interest is inside of the compartment or droplet of interest because it cannot be assumed that the structural profile of the RNA inside of the droplets is the same as that outside of the droplets. To do this, our lab has used a combination of radioactive and fluorescent accumulation experiments as orthogonal and complementary techniques (Figure 4). Radioactive experiments are more sensitive and less perturbing, as described above, while fluorescent ones allow visualization of the RNA. Accumulation into the droplets depends on several factors such as RNA length, polyion length, charge density, and ionic strength; therefore, it is important to establish whether under the conditions being investigated the RNA remains strongly partitioned into the coacervate phase. Often, the partitioning into the droplets is sufficiently strong that any RNA outside of the droplets can typically be ignored if it represents less than 10% of the total RNA.

2. The 5′-end radiolabeled RNAs from above (See Section 3.9) can be used in RNA accumulation experiments. To do this, first form the compartments of interest. Here, this is done with polycation and polyanion solutions, but for other droplet systems such as those composed of Intrinsically Disordered Proteins (IDPs) like Ddx4, instead add the protein at the appropriate concentration to phase separate. Set up the following reaction.

   Reactant	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	25	-
   10 X DMS reaction buffer	10X	5	1X
   10 X Polyanion solution	10X	5	1X
   10 X Polycation solution	10X	5	1X
   unlabeled RNA	5,000 nM	5	500 nM
   5′−32P-labeled RNA	40 kcpm/μL	5	200 kcpm total
   Total		50

   To a 0.65 mL tube, add the reagents in the order listed above. Mix thoroughly by pipetting. Allow this to incubate at the reaction temperature in our case 37 °C for 5–60 min (see Note 4.5).

3. After it has partitioned, centrifuge at 17,000 x g for 2 min to collect the coacervate phase at the bottom of the tube. (The sample can be centrifuged longer if desired, this may be necessary for some droplets.).

4. Immediately pipette one-half (25 μL) of the dilute phase into a separate 0.65 mL tube labeled TL (Top Layer) being careful not to disturb the pelleted droplets.

5. Then place the tube containing TL and the other tube labeled BL (Bottom Layer) into separate 10 mL liquid scintillation vials.

6. Scintillation count the samples. To calculate the percentage of RNA in each phase, see Frankel et al [24], Choi et al [20].

Figure 4: Radioactive and fluorescent RNA accumulation experiments.

A) (1) Radioactive RNA accumulation experiments where droplets (white) are formed. (2) The RNA of interest is added and incubated to allow it to accumulate into the droplets, then centrifuged (3–4) to allow for separation of the dilute phase from the coacervate phase, and then (5) scintillation counted. B) (1) Fluorescent RNA accumulation experiments where a calibration curve of the fluorescent RNA is made without droplets. (2) Then, separately droplets are formed, (3) the RNA of interest is added and incubated to allow it to accumulate into the droplets, (4) the fluorescence intensity of the droplets is measured by fluorescent microscopy and (5) using the calibration curve, the RNA concentration in the droplets is estimated.

3.12. Fluorescence-detected accumulation studies

Reactant	Initial Concentration	Volume (μL)	Final Concentration
Water	-	32.8	-
10 X Salt buffer (KCl, MgCl2)	10X	5.0	1X
10 X Polyanion solution	10X	5.0	1X
10 X Polycation solution	10X	5.0	1X
unlabeled RNA	16.85 μM	1.2	0.4 μM
3′-Fluor-labeled RNA	5 μM	1.0	0.1 μM
Total		50

1. The 3′-end fluorescently labeled RNAs from above can be used directly to corroborate the radioactive accumulation experiments and establish baselines for FRAP experiments. The experiments should be performed in the dark so that the fluorophore is exposed the least amount of light that can cause photobleaching (see Note 4.6).

2. Prepare the fluorescent RNA stock solution in nuclease-free water (100 μM). For RNA accumulation experiments, dilute the fluorescent RNA stock solution to 5 μM. Store these stock solutions at –80 ℃ when not being used. Before use, incubate stock solutions at between 90–95 ℃ for 2 min to renature the structure of the RNA and then incubate at room temperature for 30 min.

3. To a 0.65 mL tube, add reagents as listed above. Mix thoroughly by pipetting. Note that we use a limiting amount of RNA to prevent alteration of the physical properties of the coacervates. Allow this to incubate at the reaction temperature (or at room temperature) avoiding any lights for 5–60 min (see Notes 4.5-4.7.)

4. Once the RNA has been incubated for the desired amount of time, pipette the sample onto a silanized-glass coverslip with a silicone spacer and cover it with another coverslip. Wait for 5–10 min to allow coacervate droplets to settle on the silanized coverslip (see Note 4.8.)

5. Image coacervate samples using a confocal microscope. As this is a fluorescent intensity-based measurement, it is important to use the same confocal microscope settings, such as objective and zoom factors for the field size, laser power, detector gain, scanning rate, collected fluorescence wavelength range, and the dichroic filters, for all samples and the calibration curve so that they can be compared to each other. Image at least three different locations on each slide that contain at least ten coacervate droplets per coacervate sample for statistical sampling.

6. Image two negative control samples with the same confocal microscope settings from step 5 above. The first negative control is the coacervate samples without fluorescent RNA added and the second is just water. Make sure the coacervate droplets without fluorescent RNA do not have significant variation in morphology from the samples with fluorescent RNA, for example gelation, which can be indicative of the material properties of the coacervates being changed by the addition of RNAs. These negative controls will also be used to correct for the background fluorescent signal and guide the endpoints in calibration in next steps.

7. To generate a calibration curve of fluorescent intensity as a function of known RNA concentrations, place 5–10 μL of 100 μM fluorescent RNA stock solution on a new coverslip without a cover slip on top, and image at least three different locations using the same confocal setting before performing a dilution. Make sure that the imaging occurs on the right focal plane and work quickly to minimize evaporation. Add a known volume of nuclease-free water or salt buffer solution (150 mM KCl or NaCl) directly to the droplet of 100 μM fluorescent RNA stock solution on the coverslip followed by gentle pipette-mixing for the first dilution. Image another three locations at this more dilute concentration. (Record the volume of buffer solution added to calculate the RNA concentration in the drop of sample on the coverslip.) Keep performing dilutions and imaging until the fluorescent intensity is similar to the negative control intensities. Aim to get at least five different RNA concentrations per curve for the fitting later.

8. Using Image J, measure the fluorescent intensities in coacervate droplets (I_coa) (draw the Regions Of Interest (ROIs) for each coacervate droplet using oval ROIs to define the area in each coacervate droplet). Measure the intensity of the entire image field from calibrations (I_cali(y), where y is known RNA concentration), and negative controls (I_nc). I_coa and I_cali should be background corrected by subtracting I_nc from those individually. Calibration curves will be plotted as RNA concentrations versus intensity and fit with a linear curve, where y = ax + b (a and b are fitting parameters, y is the RNA concentration and x is the average value of the background-corrected intensity). Aim to get b close to be 0 and R² for the calibration curve fitting greater than 0.9. If b is not close to 0, examine your background correction. If R² is lower than 0.9, increasing the number of data points and the calibration sample volume can be helpful. Using background-corrected I_coa, estimate RNA concentration in the coacervate from the equation given the assumption that the fluorophore is not significantly influenced by the material properties of coacervates.

9. If desired, to understand the fluidity of the droplets, Fluorescent Recovery After Photobleaching (FRAP) can be performed on these samples as well, which is out of scope of this chapter but well described by Alshareedah et al [25].

3.13. Preliminary DMS reaction for tRNAs in droplets

1. Before performing tRNA Structure-seq experiments in droplets, it is important to determine what concentration of DMS is required to obtain a sufficient level of modification of the RNA. This can vary depending on the ability of DMS to get into the droplets and whether the droplets will react with some of the DMS. We suggest checking this by performing a DMS chemical probing experiment in the same way that it would be performed in tRNA Structure-seq (Figure 1A) and then use SuperScript III Reverse Transcriptase with a 5′-end-radiolabeled reverse transcription primer to observe DMS induced reverse transcription stops.

2. Set up the following DMS reaction. Note that all DMS chemical probing experiments must be performed inside of a fume hood wearing proper PPE.

   Reactant	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	275	-
   10 X DMS reaction buffer	10X	40	1X
   10 X Polyanion solution	10X	40	1X
   10 X Polycation solution	10X	40	1X
   unlabeled RNA	1,000 ng/μL	5	5,000 ng
   Total	-	400	-

   To a 1.7 mL tube, add the reagents in the order listed above. Mix thoroughly by pipetting. Incubate the mixture at 37 ºC for 5–10 min. Once the sample has equilibrated, add an appropriate amount of either DMS solution (8 μL of a 5 M DMS solution diluted in ethanol) or -DMS control solution (in our case 8 μL of 100% ethanol). Allow the DMS to react for 1 min before centrifuging at 17,000 x g for 2 min.

3. Immediately pipette off the supernatant being careful to avoid disturbing the pellet. (Collect the supernatant in a 50 mL conical tube and add 100 μL of 2.5 M dithiothreitol to quench the dilute phase).

4. After 5 min of total reaction time add 100 μL of 2.5 M dithiothreitol to quench the reaction.

5. Add guanidinium chloride to a final concentration of 600 mM to help dissolve the droplets. Then add a 300 μL of Coacervate loading dye.

6. Fractionate the DMS-reacted tRNA solution on a 10% denaturing (8.3 M urea, 1X TBE) polyacrylamide gel for 2–3 h at 40 Watts.

7. Once the bromophenol blue dye is ~ 2.5 cm from the bottom of the glass plate, disassemble the gel apparatus, stain the gel with SYBR-GOLD, visualize the RNA under UV light, excise the RNA, cut it into 1 × 1 mm squares, and place it in a 50 mL conical tube.

8. Follow steps 7–18 from section 3.7 to recover the RNA.

9. Once the RNA has been recovered, set up the following SuperScript III reverse transcription reaction.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Nuclease-free water	-	3.5	-
   10 X SuperScript III reaction buffer	10 X	1	1 X
   5’−32P-labeled DNA primer	~200 kcpm/μL	1	~20 kcpm/μL
   DMS-treated RNA	-	1	-
   10 X dNTP solution	10X	1	1 X
   DTT	100 mM	1	10 mM
   MgCl2	30 mM	1	3 mM
   SuperScript III RT	200 U/μL	0.5	10 U/μL
   Total	-	10	-

   Add water,10 X SuperScript III reaction buffer, ³²P-labeled DNA primer, and DMS-treated RNA. Pipette mix thoroughly and briefly centrifuge. Place the tube into a thermocycler and run at 90 ºC for 1 min and then place the tube on ice for 10 min to anneal the primer to the RNA.

10. Add the dNTP solution, DTT, MgCl₂, and SuperScript III RT. Perform the RT reaction at 55ºC for 60 min.

11. Add 0.5 μL 2 M NaOH and incubate at 95 ºC for 5 min to hydrolyze the template RNA.

12. Add 10 μL 2x FLD.

13. Run a denaturing gel at 40 W for ~ 2 h (This is dependent on your gel size and gel apparatus).

14. Dry the gel and expose it to a phosphor screen overnight.

15. Visualize the gel using Phosphorimager.

3.14. tRNA Structure-seq in droplets (Library preparation)

After the tRNAs have been probed by DMS in the droplets, NGS libraries must be made (Figure 1A). This involves reverse transcription of the tRNAs by Marathon RT in the presence of Mn²⁺. Marathon RT is an ultra-processive RT that can read through the strong structure and abundant chemical modifications of tRNAs, and Mn²⁺ increases its ability to insert mutations in the cDNA opposite chemical or natural modification sites. These mutations will be read out in the data analysis step below after NGS. To prepare tRNA Structure-seq libraries, first ligate a reverse transcription primer binding site to the 3’ end of the tRNAs. Then, reverse transcribe with a RT primer that has a C18 spacer to generate cDNA, circularize the cDNA with Circligase II and then finally perform indexing PCR with standard Illumina sequencing primers.

1. After it has been have established that the RNA of interest can be sufficiently modified by DMS in the droplets of interest, tRNA Structure-seq experiments can be performed. These experiments consist of several steps: DMS chemical probing, gel purification, and library preparation. Here we describe a procedure to use Marathon Reverse Transcriptase but other ultra-processive reverse transcriptases such as TGIRT can be adapted to this technique.

2. Perform the DMS reaction as determined in the previous section.

3. Perform the 5′-end Adenylation reaction of ligation adapter for reverse transcription. This step serves to prepare the DNA ligation adapter to be ligated onto the 3′ end of the RNA.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   5′p-DNA ligation adapter	200 μM	2	40 μM
   10X DNA adenylation buffer	10X	1	1X
   ATP	1 mM	2	200 μM
   Mth RNA ligase	50 μM	4	20 μM
   Water	-	1	-
   Total	-	10	-

   Add the reactants in the order listed above to a 0.2 mL tube, mix thoroughly by flicking, briefly centrifuge to collect the liquid at the bottom and incubate at 65 °C for 1 h. Then, heat inactivate the enzyme at 85 °C for 5 min. The adenylated ligation adapter can then be used without further purification in the ligation reaction.

4. Perform the following adapter ligation.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   PEG 8000	50 % w/v	10	25 % w/v
   Purified DMS-treated RNA	-	4	-
   5′-Adenylated DNA oligonucleotide (from previous step)	40 μM	3	6 μM
   T4 RNA ligase buffer	10X	1	1X
   T4 RNA ligase 2 truncated	200 U/μL	1	200 U
   Water	-	1	-
   Total	-	20	-

   Add the reagents in the order listed above to a 0.2 mL tube, mix thoroughly by flicking, briefly centrifuge to collect the liquid at the bottom of the tube, and incubate at 16 °C for 16 h. Then, add 20 μL 2X FLD. Prepare a no-ligation sample with 4 μL of purified DMS-treated RNA and 4 μL of 2X FLD (This will be enough volume to run two no-ligation lanes on the gel one on either side of the ligated samples. The two no-ligation samples make it apparent that the ligated samples have a supershift on the gel), fractionate the sample on a 10 % denaturing (8.3 M urea, 1X TBE) polyacrylamide gel at 30 W for 2 h. Next, stain the gel with SYBR-GOLD, visualize the ligated product with UV light, excise the band, cut it into 1 × 1 mm squares, incubate the gel slices in TEN250 crush and soak buffer at 4 °C overnight. The next day, recover the ligated product by ethanol precipitation. Resuspend the ligated product in 10 μL water and determine its concentration with a microvolume UV-vis spectrophotometer.

5. Perform the following reverse transcription.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   2X RT buffer (-Mn)	2X	5	1X
   Renatured ligated product/primer	-	2	-
   MnCl2	20 mM	1	2 mM
   dNTP mix	10 mM	1	1 mM
   Marathon Reverse transcriptase	20 U/μL	1	20 U
   Total	-	10	-

   Add the reagents in the order listed above, mix thoroughly by flicking, and briefly centrifuge to collect the liquid at the bottom of the tube, incubate the reaction at 42 °C for 60 min. After 60 min, add 1 μL 2M NaOH and incubate at 95 °C for 5 min. This heat inactivates the enzyme and degrades the RNA template. Then add 11 μL 2X FLD, fractionate on a 10% denaturing (8.3 M urea, 1X TBE) polyacrylamide gel, stain with SYBR-GOLD, visualize bands under UV-light, excise the RT product band, cut the gel slice into 1 × 1 mm slices, put the slices in a 1.7 mL tube with 500 μL TEN250 crush and soak buffer and spin at 4 °C overnight. The next day, recover the cDNA via ethanol precipitation and resuspend it in 10 μL water.

6. Perform the following circularization reaction.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	3.5	-
   Betaine	5 M	3	1M
   cDNA	-	5	-
   10 X CircLigase Buffer	10X	1.5	1X
   MnCl2	50 mM	1	3.3 mM
   CircLigaseII	100 U/μL	1	100 U
   Total	-	15	-

   Add all of the reagents to a 0.2 mL tube, mix thoroughly by flicking, briefly centrifuge to collect the liquid at the bottom of the tube, and incubate the reaction at 60 °C for 2 h. After 2 h, the circularized product can be used in an indexing PCR without further purification.

7. Perform the following indexing PCR to prepare the sequencing library.

   Reagent	Initial Concentration	Volume (μL)	Final Concentration
   Water	-	41	-
   Phusion HF buffer	5X	15	1X
   Circligated cDNA (from the previous step)	-	9	-
   Forward indexing primer (see [13])	10 μM	3.8	0.5 μM
   Reverse primer (unique indexing primer see [13])	10 μM	3.8	0.5 μM
   dNTP mix	10 mM	1.5	200 μM
   Phusion HF DNA Polymerase	2 U/μL	1	2 U
   Total	-	~75	-

   Add the reagents in the order listed above to a 0.2 mL tube, mix thoroughly by flicking and briefly centrifuge to collect the liquid at the bottom of the tube. Put the tube into the thermocycler and run the following program: Step 1 initial denaturation at 98 °C for 30 sec, Step 2 denaturation at 98 ºC for 10 sec, Step 3 annealing at 60 ºC for 10 sec, Step 4 extension at 72 ºC for 5 sec, Step 5 final extension at 72 ºC for 5 min. Repeat Steps 2–4 for 17 cycles (see Note 4.9).

8. After the indexing PCR, fractionate the library on an 8% non-denaturing (1X TBE) polyacrylamide gel at 180 V for 2–3 h, stain the gel with SYBR-GOLD, visualize the product bands under UV light, excise the library, cut the gel slice into 1 × 1 mm squares, and incubate at 4 °C overnight in a 1.7 mL tube with 500 μL of TEN250.

9. The next day, recover the library with a Nucleospin Gel and PCR Clean up kit (Macherey-Nagel, 740609.50). The final volume should be 25 μL.

10. Measure the concentration of the library with a microvolume UV-vis spectrophotometer. A good library will have a concentration between 5–20 ng/μL at this step.

11. We typically have the libraries quality controlled by running them on a Tapestation and determine the library concentration before sequencing by qPCR. A sample Tapestation output is provided in Figure 5.

12. Once the libraries have been found to be good by Tapestation and the concentration of each library measured by qPCR, perform next generation sequencing with NextSeq 500/550 using Illumina′s custom sequencing primer (see Note 4.10).

Figure 5: Examples of good and poor quality libraries as judged by Tapestation.

A) Example Tapestation with 4 tRNA Structure-Seq libraries shown (samples A-D) where the library size should be ~230 bp. B) Example of a good library that is ready to be sequenced that has a single peak at ~230 bp. C) Example of a poor quality library that needs to be remade, where there is an off-target peak that is ~305 bp.

3.15. tRNA Structure-seq in vivo (Library preparation)

tRNA Structure-seq can also be performed for in vivo conditions. The following procedures were established for Escherichia coli tRNAs (Figure 1B).

1. Prepare 120 mL LB medium and autoclave it. In a laminar flow cabinet, dispense 30 mL the LB medium into each of 4 sterile conical tubes. Culture E. coli BW25113 strain in one of the 30 mL conical tubes of LB medium at 37 ºC overnight in a bioshaker. The next day, inoculate 150 μL cells into the three remaining conical tubes, each containing 30 mL of LB medium. Of three tubes, two tubes are for plus/minus DMS condition, and the other tube is for measuring the optimal density at 600 nm (OD_600nm) to monitor E. coli growth. Grow at 37 ºC for several hours in a bioshaker, and measure the OD_600nm every 1 h. Once the OD_600nm reaches 0.8, bring the two tubes into a chemical hood and add 300 μL of 5 M (1% w/v) DMS for +DMS condition and add 300 μL ethanol for – DMS condition before bringing the tubes back to the bioshaker and incubating at 37 ºC for 5 min. Then bring the tubes back to the hood and add 0.6 g DTT powder and 30 μL of 30 mg/mL chloramphenicol to quench the reaction and halt growth. Incubate the cells at 37 ºC for 2 min, then place the culture tubes on ice-water to stop growth (Figure 1B)

2. Centrifuge the cells at 10,000 x g for 15 min at 4 ºC and discard the supernatant to a hazardous waste tube.

3. To extract total RNA, add 1 mL TRIzol into the cultured cells and homogenize the cells as follows. Have a disposable syringe (use 5 mL syringe) with a 20 gauge needle and draw the TRIzol-cells mixture into the cylinder and then strongly push the plunger down to vigorously eject the TRIzol solution and cells. Repeat this twenty times. In this way, the cells are well homogenized. Incubate the mixture at room temperature for 5 min. Add 0.2 mL of chloroform and incubate at room temperature for 2 min before centrifuging at 12,000 x g for 15 min. After that, transfer the aqueous phase (upper) to a new tube and add 0.5 mL isopropyl alcohol. Incubate the tube at 4 ºC for 10 min to precipitate the RNA and centrifuge it at 12,000 x g for 10 min. Discard the supernatant and dry the RNA pellet. Dissolve the pellet in 50 μL water

4. To purify tRNA fraction, fractionate the RNA on a 10% denaturing polyacrylamide gel (8.3 M urea) and follow Section 3.7 Steps 6–18 to recover the RNA.

5. To prepare the libraries, follow Section 3.14 Steps 3–12.

3.16. tRNA Structure-seq analysis

We provide a brief workflow of tRNA Structure-seq NGS sequencing data analysis (Figure 6). First, the adapter sequence in the fastq.gz files from the NGS experiment is removed by Cutadapt (Figure 6A). Also, reference files need to be prepared for ShapeMapper2 and the tRNA Structure-seq analysis. The gtRNA_setup.py script converts tRNAscan-SE results to those reference files (Figure 6B). tRNAscan-SE is a software that searches tRNA genes encoded in a genome and provides a list of tRNA genes, loci, sequences, and predicted secondary structures [26]. tRNAscan-SE results are provided in the genomic tRNA database [27]. For example, in the case of E. coli, download the tRNAscan-SE results at http://gtrnadb.ucsc.edu/genomes/bacteria/Esch_coli_K_12_MG1655/, and run the gtRNA_setup.py script to prepare the reference files. Second, DMS reactivity profiles are calculated by ShapeMapper2 (Figure 6C). DMS profiles from at least two independent experiments are checked with tRNA_correlation.py to see if those replicates have similar DMS modification patterns. If the replicates show a high correlation, those fastq files from the replicates can be combined. The combined fastq file is again analyzed by ShapeMapper2 (Figure 6C) [28]. Then, the DMS reactivity profiles, reference fasta, and reference CT files are used for tRNA Structure-seq analysis (Figure 6D). The tRNA_structure_seq.py first filters the DMS reactivities from the natural tRNA modifications and the reactivities at G and U nucleotides. Then, it utilizes the Fold algorithm in RNAstructure to predict the minimal free energy structure with restraints from the DMS experiment (Figure 6D). The Fold results are then used to calculate the prediction accuracy. Finally, a summarized csv of the data and a bar graph of prediction accuracies are generated. The python scripts needed for tRNA Structure-seq are distributed at https://github.com/Ryota-Yamagami/tRNA_structure-seq. ShapeMapper2 that requires a Linux system, therefore, a Linux PC or Linux-booted Windows OS system is required to perform the analysis.

Figure 6: tRNA Structure-seq computational analysis.

A) Cutadapt removes the adapter sequences from input fastq files. The adapter trimmed reads are used for mutational profiling analysis by ShapeMapper2. B) the gtRNA_setup.py script generates three outputs. One for a reference fasta file for ShapeMapper2. The others are reference fasta and CT files for each tRNA gene for tRNA Structure-seq analysis. C) DMS reactivity is calculated by alignment-based variant calling. A correlation plot of DMS reactivities from 2–3 biological replicates can be made by tRNA_correlation.py. Then, the fastq files are merged and re-analyzed by ShapeMapper 2. D) tRNA_structure_seq.py executes RNA folding with the experimental restraints, calculates prediction accuracy, and then visualizes the data.

1. To perform tRNA Structure-seq analysis, first set up a miniconda environment. Open a command line and type the following:

   # Specify the URL for miniconda installer.

   URL = https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh

   # Download the installer.

   curl $URL > miniconda-installer.sh

   # Run the miniconda installation script.

   bash miniconda-installer.sh -b

2. Create a new conda environment by typing the following code:

   # Create a new conda environment for tRNA-MaP.

   conda create -y --name tRNAstructureseq python=3.7

   # Activate the environment.

   conda activate tRNAstructureseq

   # Configure environment channels

   conda config --add channels conda-forge

   conda config --add channels bioconda

3. Install the software by typing the following code:

   # Install tools required for the analysis.

   conda install -c anaconda pandas

   conda install -c conda-forge matplotlib

   conda install -c anaconda seaborn

   conda install -c bioconda Cutadapt

   conda install -c bioconda rnastructure

   # After installing rnastructure, you will need to add a DATAPATH to the data_tables in rnastructure to .bashrc file.

   # For the installation of shapemapper2 (See https://github.com/Weeks-UNC/shapemapper2), download the file

   wget https://github.com/Weeks-UNC/ShapeMapper2/releases/download/2.1.5/shapemapper-2.1.5.tar.gz

   # Extract release tarball with

   tar -xvf shapemapper-2.1.5.tar.gz

   #Add shapemapper executable to PATH

4. Download the example sequencing data from (https://github.com/Ryota-Yamagami/tRNA_structure-seq). Make a directory named “Example_Data” and put the fastq.gz files into there. Run the following command to remove the adapter sequence from the reads with Cutadapt [29]. The adapter trimmed fastq files are saved in a folder named “MaP_analysis”

   # Adapter trimming (Do this for all fastq files)

   # Create folders “unmodified_rep1”, “unmodified_rep2”, “modified_rep1”, and “modified_rep2” if these are not in the MaP_analysis folder. mkdir MaP_analysis/unmodified_rep1 MaP_analysis/unmodified_rep2 MaP_analysis/modified_rep1 MaP_analysis/modified_rep2

   cutadapt -a CTGTAGGCACCATCAAT -o MaP_analysis/unmodified_rep1/minus_DMS_rep1_cut.fastq.gz Example_Data/minus_DMS_rep1.fastq.gz

   cutadapt -a CTGTAGGCACCATCAAT -o MaP_analysis/unmodified_rep2/minus_DMS_rep2_cut.fastq.gz Example_Data/minus_DMS_rep2.fastq.gz

   cutadapt -a CTGTAGGCACCATCAAT -o MaP_analysis/modified_rep1/plus_DMS_rep1_cut.fastq.gz Example_Data/plus_DMS_rep1.fastq.gz

   cutadapt -a CTGTAGGCACCATCAAT -o MaP_analysis/modified_rep2/plus_DMS_rep2_cut.fastq.gz Example_Data/plus_DMS_rep2.fastq.gz

5. To setup reference files for ShapeMapper2 and the tRNA Structure-seq analysis, download tRNAScan-SE results at genomic tRNA database (http://gtrnadb.ucsc.edu/index.html). The tRNAScan-SE results contain information about all tRNA genes encoded in a genome. For example, the E. coli K-12 strain genome (http://gtrnadb.ucsc.edu/genomes/bacteria/Esch_coli_K_12_MG1655/) has 89 different tRNA genes. Of these tRNA genes, some tRNA sequences are perfectly matched, and we can combine those tRNA genes that share the sequence. The gtRNA_setup.py does this and generates a refence fasta for ShapeMapper2. Also, the tRNAScan-SE results contain structural information of the canonical secondary structure predicted by a covariation folding model for each tRNA gene. The gtRNA_setup.py retrieves the secondary structure of the tRNA and converts it into a CT file to prepare it for tRNA Structure-seq analysis. The reference fasta for ShapeMapper2 and reference fasta and CT files for the tRNA Structure-seq analysis are saved in folders named “MaP_analysis” and “Structure_seq_analysis”, respectively. Run the following command.

   # If “MaP_analysis” and “Structure_seq_analysis” folders does not exist in a working directory, create them by typing

   # mkdir MaP_analysis Structure_seq_analysis

   python gtRNA_setup.py

6. Run ShapeMapper2. The following command performs mutational profiling via read alignment, variant calling, and data normalization. Use the following eight parameters:

   --name	Experiment name
   --target	Path to a reference fasta file
   --out	Output folder name
   --overwrite	Overwrites the file if it already exists
   --indiv-norm	Individual normalization of reactivity data
   --mim-mutation-separation	Two mutations must be separated by at least this many unchanged reference sequence nucleotides
   --modified –unpaired-folder	Path to the folder that stores DMS-treated data
   --untreated –unpaired-folder	Path to the folder that stores DMS-untreated data

   Here is an example command for ShapeMapper2. Note that the example data provide a limited amount of ~500,000 reads in order to reduce the data size. This sequence size is not good enough for the MaP analysis of ~50 different tRNAs.

   # ShapeMapper2

   # Replicate 1

   shapemapper --name Eco_invivo_rep1 --target MaP_analysis/reference_fasta_ShapeMapper/reference_shapemapper.fa --out MaP_analysis/MaP_data_rep1 --overwrite --indiv-norm --min-mutation-separation 1 --min-depth 1000 --modified --unpaired-folder MaP_analysis/modified_rep1 --untreated --unpaired-folder MaP_analysis/unmodified_rep1

   # Replicate 2

   shapemapper --name Eco_invivo_rep2 --target MaP_analysis/reference_fasta_ShapeMapper/reference_shapemapper.fa --out MaP_analysis/MaP_data_rep2 --overwrite --indiv-norm --min-mutation-separation 1 --min-depth 1000 --modified --unpaired-folder MaP_analysis/modified_rep2 --untreated --unpaired-folder MaP_analysis/unmodified_rep2

7. Make correlation plots of the mutation rates between pairs of replicates(we normally perform 2 or 3 individual experiments) (Figure 7). The tRNA_correlationl.py script (available at https://github.com/Ryota-Yamagami/tRNA_structure-seq.) makes these plots. To execute this script, specify the shapemapper output folders. The python script reads the mutation rates from each profile.txt file, compiles them, and generates a correlation plot. In the script, we use the following three parameters:

   --input1 (-i1)	Specify shapemapper output folder 1
   --input2 (-i2)	Specify shapemapper output folder 2

   # correlation plot

   python tRNA_correlation.py -i1 MaP_analysis/MaP_data_rep1 -i2 MaP_analysis/MaP_data_rep2

   If the mutation rates between the replicates are highly correlated to each other (r² ≥ 0.85), you can combine all the fastq data files together and run Shapemapper2 with the combined fastq data.

8. Fold RNA using the experimental restraints from ShapeMapper output using the Fold software in RNAstructure developed by the Mathews lab [30]. The software predicts an RNA structure that has the minimal folding energy and can accept RNA structural information such as DMS and SHAPE reactivities as experimental restraints. For example, DMS reactive nucleotides indicate that the nucleotides are solvent accessible (i.e. they tend to be located in single stranded regions). The Fold algorithm gives a pseudo free energy [slope (1.8 kcal/mol) and intercept (–0.6 kcal/mol)] as a penalty when the DMS reactive nucleotides against base pairing. The tRNA_structure_seq.py script predicts the structure of a tRNA by first removing all of the DMS reactivities for Gs and Us. Second, the Fold algorithm performs the structure prediction and provides a CT file, which is a file containing information on the secondary structure of the RNA. Refer to the CT File Format on the Mathews Lab web page (https://www.urmc.rochester.edu/rna/#CT). In parallel, the CT file is converted to a dot bracket file by ct2dot and scorer calculates the sensitivity and PPV. These values indicate the structural similarity between a predicted structure and reference structure. Thus, for this calculation, the reference tRNA sequences and structures formatted into fasta and CT are required, and can be generated by gtRNA_setup.py (See Note 4.11). Finally, the python script approximates the Matthew’s Correlation Coefficient by taking the geometric mean of sensitivity and PPV [31] and summarizes the Fold results and compiles the data of the prediction accuracy in an output folder (Fig. 6D). The tRNA_structure_seq.py script can be downloaded from Github at https://github.com/Ryota-Yamagami/tRNA_structure-seq. Note that the tRNA_structure_seq.py script was not utilized in our PNAS paper but is built from the same analysis done there. In the script, we use the following two parameters:

   --shapemapper (-s)	Specify shapemapper output folder
   --output (-o)	Output folder name

   Have a shapemapper output folder, the accepted_ct folder, reference_fasta folder in the same directory named “Structure_seq_analysis” and run the script with the following command.

   # analysis

   python tRNA_structure_seq.py -s Structure_seq_analysis/MaP_combined/ -o Structure_seq_output

Figure 7: Correlation plot of mutational profiling data between two replicate experiments.

DMS reactivity profiles are compared to see a correlation between two replicates. if they are highly correlated, then the two fastq files can be merged.