Native purification and labeling of RNA for single molecule fluorescence studies

Abstract

The recent discovery that non-coding RNAs are considerably more abundant and serve a much wider range of critical cellular functions than recognized over previous decades of research into molecular biology has sparked a renewed interest in the study of structure-function relationships of RNA. To perform their functions in the cell, RNAs must dominantly adopt their native conformations, avoiding deep, non-productive kinetic traps that may exist along a frustrated (rugged) folding free energy landscape. Intracellularly, RNAs are synthesized by RNA polymerase and fold co-transcriptionally starting from the 5’ end, sometimes with the aid of protein chaperones. By contrast, in the laboratory RNAs are commonly generated by in vitro transcription or chemical synthesis, followed by purification in a manner that includes the use of high concentrations of urea, heat and UV light (for detection), resulting in the denaturation and subsequent refolding of the entire RNA. Recent studies into the nature of heterogeneous RNA populations resulting from this process have underscored the need for non-denaturing (native) purification methods that maintain the co-transcriptional fold of an RNA. Here, we present protocols for the native purification of an RNA after its in vitro transcription and for fluorophore and biotin labeling methods designed to preserve its native conformation for use in single molecule fluorescence resonance energy transfer (smFRET) inquiries into its structure and function. Finally, we present methods for taking smFRET data and for analyzing them, as well as a description of plausible overall preparation schemes for the plethora of non-coding RNAs.

1. Introduction

Over the last decade, the classical role of RNA as the messenger between DNA and protein, the legislative and executive branches of molecular biology, has been greatly expanded with the discovery of a multitude of non-protein coding (or non-coding) RNAs (1). These RNAs have been shown to play a central role in a variety of cellular processes including the regulation of gene expression (2, 3), splicing (4), translation (5) and maintenance of chromosome ends (6). Furthermore, it has been shown that the effect of secondary structure inherent to the untranslated regions (UTRs) of messenger RNAs (mRNAs) on gene expression has been largely underestimated (7). As a result, much renewed interest has shifted towards studies of structure and function of RNAs. For example, the discovery of riboswitches has added a new dimension to the role of RNA structure in gene regulation. Riboswitches are structural elements found most commonly in the 5’ untranslated regions (UTRs) of bacterial mRNAs that respond to an intracellular metabolite signal (8). Binding of this metabolite to a highly conserved aptamer domain induces a structural change in the downstream expression platform. The molecular communication between the aptamer domain and expression platform results in a genetic “off” (or sometimes “on”) signal, in which expression of the downstream gene is down regulated, usually through attenuation of transcription or inhibition of translation initiation. A plethora of studies focused on the conformational disparity between the ligand-bound and ligand-free states (9–14) are beginning to yield a detailed picture of the mechanisms by which riboswitches can regulate gene expression. However, the question of how a ligand binding event can cause a downstream conformational change remains largely shrouded. To better understand the complex function of RNA secondary and tertiary structure motifs in gene regulation, as exemplified by riboswitches, it is of great importance to study the conformational changes and folding dynamics inherent in non-coding RNAs.

Ideally, RNA folding and function would be investigated in vivo; however, due to the complexity of the cellular environment and our limited ability to manipulate it, it has become accepted practice to study RNA structure and folding in vitro. Typically, the RNA required to perform these studies is produced either by chemical synthesis or by run-off transcription in vitro from a linearized DNA template using T7 RNA polymerase (15). Following synthesis or transcription, the most common means of RNA purification is through the use of denaturing polyacrylamide gel electrophoresis (PAGE), followed by gel elution and refolding. This method relies heavily on the assumption that in vitro synthesized RNA will refold into its native and functional structure, and therefore, will give proper insight into its properties in vivo. However, this process requires a variety of harsh denaturants including urea and heat, as well as the exposure to UV light (for detection). In contrast, RNAs generated in vivo fold in the crowded environment of the cell on their own co-transcriptionally and segmentally beginning from the 5’ end (16), or with the aid of a multitude of protein chaperones that aid in navigating an often complex folding free energy landscape (7, 17, 18). Recently, an array of studies has shown that the commonly accepted methods of denaturing RNA purification can result in severe conformational heterogeneity where populations of active (properly folded) and inactive (misfolded) RNA species coexist (19–21), highlighting the need for suitable non-denaturing (native) RNA purification techniques. Such techniques better preserve the co-transcriptional fold of an RNA and thus have the potential to simplify the complexity of an RNA population for further study. This feature is of particular importance for riboswitches that are thought to often switch in a co-transcriptional fashion. However, the challenge then becomes how to attach detectable labels to the RNA as needed for many biophysical in vitro studies.

The purpose of this chapter is to highlight how native RNA purification may be coupled with the attachment of labels that enable single molecule fluorescence resonance energy transfer (smFRET) measurements in particular and label-dependent biophysical studies in general while maintaining the RNA’s co-transcriptional fold. smFRET is a highly sensitive technique that can reveal even subtle conformational and dynamic changes in single RNA molecules, thus often capturing transient and/or rare events that are averaged out in more conventional ensemble (bulk) experiments (22–24). Recently, the capabilities of smFRET have been widely established on a variety of non-coding RNAs, such as ribozymes (25–33), riboswitches (34–36), spliced pre-mRNAs (37) and telomerase RNA (38, 39), underpinning its usefulness in the study of RNA (un)folding by directly measuring unperturbed kinetic rate constants at equilibrium. A particular challenge of such smFRET studies is the need to site-specifically attach both a biotin molecule for surface immobilization (required for extended observation) and a pair of fluorophores for FRET, requiring three orthogonal labeling strategies. In addition to native purification and smFRET measurement techniques, we therefore provide here a selection of proven protocols and examples for the labeling of in vitro transcribed RNAs for smFRET measurements of co-transcriptionally folded samples. We anticipate that combining these protocols in various ways will become increasingly important for studying folding and function of an ever-expanding non-coding RNA universe.

1.1 Combining Native Purification with Labeling of RNA: Overview of Methods

Native purification methods can offer a solution to the formation of heterogeneous RNA populations by avoiding the need for harsh denaturants. Different strategies are used to natively purify RNA that can be broadly classified into two categories. The first set of methods isolates the full-length transcript utilizing either a protein (40–43) or a complementary oligonucleotide (20, 44) that specifically binds an incorporated affinity tag in the RNA, followed by ribozyme or DNAzyme mediated cleavage to obtain the correct length product. Alternatively, RNA is primed during in vitro transcription with a photocleavable, biotinylated GMP tag that is removed subsequent to affinity purification (45). A second set of methods directly purifies the RNA from the transcription reaction mix using HPLC or a combination of weak anion exchange and size exclusion chromatographies (46, 47). Although the affinity based methods generally require more steps and construction of an altered DNA template, they more easily generate the correct length product with precise, homogeneous 5’ and/or 3’ end.

We have developed a simple and rapid RNA native purification technique that uses streptavidin coated magnetic beads and glmS ribozyme mediated cleavage to generate homogeneous RNA in a short time (20) (Fig. 1). It removes the requirement for large quantities of protein and expensive FPLC systems associated with an earlier version of protein based affinity purification (41). The DNA template (Fig. 1a) encodes a full length transcript containing the RNA of interest at the 5’ end, followed by a glmS ribozyme sequence and a 3’ binding sequence that is complementary to a 30-nt biotinylated DNA oligonucleotide (capture strand). The in vitro transcription reaction is performed in the presence of the capture strand and, after a phenol-chloroform extraction to remove protein, streptavidin-coated magnetic beads are used to affinity purify the full length transcript via the hybridized biotinylated capture strand. Any terminated or shorter transcripts and free NTPs are washed away after this binding step. Addition of glucosamine-6-phosphate (GlcN6P) containing buffer to the bead-bound RNA activates the glmS ribozyme, resulting in its self-cleavage and release of the RNA segment of interest (Fig. 1a). The magnetic beads can be easily regenerated by heating to disrupt the biotin-streptavidin interaction and be reused for multiple rounds of purification with little decrease in overall yield. Using this method, sufficient quantities of conformationally homogeneous RNA can be made (Figs. 1b and 1c), which can be labeled for smFRET studies.

Fig. 1.

Native purification of RNA, adapted from reference 20. Panel (a): The purification scheme begins with T7 RNA polymerase transcription of a conjugated RNA composed of target RNA-glmS ribozyme-binding sequence. The binding sequence binds to a biotinylated capture strand included in the transcription reaction, and is subsequently incubated with streptavidin coated beads. Following cleavage by addition of glucosamine-6-phosphate (GlcN6P), the RNA of interest is eluted from the beads without any harsh denaturing steps. Panel (b): Representative denaturing PAGE gel showing that the hepatitis delta virus (HDV) ribozyme elutes cleanly from the beads after glmS-mediated cleavage. Panel (c): Non-denaturing PAGE gel showing that the homogeneously folded VS ribozyme obtained from non-denaturing purification redistributes into multiple bands when heat denatured and refolded.

Studying RNA folding and dynamics using smFRET by total internal reflection microscopy (TIRFM) methods requires labeling with donor and acceptor fluorophores and biotin for surface immobilization to increase the time of observation (22, 24). Although different chemical and enzymatic methods are available for labeling RNA at the 5’ or 3’ end (48), not all of them are compatible with native purification as they involve harsh conditions that could denature the RNA. Here we describe some of the labeling methods that may be used to generate a triply labeled, co-transcriptionally folded RNA sample. Labeling of the 5’ end of RNA can be accomplished by reacting the phosphate group with the carbodiimide cross linker EDC (1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride) and imidazole (49), generating a reactive phosphorylimidazolide group that can be coupled to any primary amine containing molecule such as biotin hydrazide. Alternatively, fluorophore labeling is achieved by reacting the phosphorylimidazolide group with ethylenediamine to generate a free primary amine that can be coupled to any NHS ester modified molecule (50, 51). Other methods are available to directly generate RNA with 5’ reactive groups or labels through in vitro transcription with T7 RNA polymerase, since the enzyme can incorporate modifications at the 5’ end of the transcript (52–54). Periodate chemistry oxidizes the ribose 2’,3’-diol in RNA to aldehyde groups that are reactive with a hydrazide modified fluorophore or biotin to generate 3’ end labeled RNA (49). In addition, enzymatic labeling of the RNA 3’ end can be achieved using T4 RNA ligase (55) or poly(A) polymerase (56). Poly(A) polymerase is generally used for polyadenylation of RNAs but can also be used to incorporate azide(N₃)-modified NTPs at the 3’ end of RNA (56). Under optimized conditions, yeast and E. coli poly(A) polymerases add a single 2’-N₃-2’-dNTP that can be reacted with an alkyl-derivative of any fluorophore label using ‘click’ chemistry. However, since it is hard to control the number of added nucleotides, 3’-N₃-2’,3’-ddNTPs can be used to incorporate just a single nucleotide at the 3’ end of RNA (56). Both copper-catalyzed and copper free ‘click’ chemistry can then be used for labeling (56), but the presence of copper in the reaction leads to degradation of RNA and could alter its native conformation. Recent studies have focused on optimizing solvent conditions for copper-catalyzed click chemistry for labeling RNA to obtain maximum yield with minimal degradation (54, 57). However, whether these methods preserve the native fold of RNA still needs to be tested. Strain promoted azide-alkyne cycloaddition (SPAAC) or copper-free click chemistry is a mild and potentially more efficient way of labeling azide-modified RNA with fluorophore conjugated cyclooctynes. Since the reaction does not require metal ions and also occurs at low temperatures, it is compatible with native labeling of RNA. A variety of fluorophores conjugated to different cyclooctynes are now commercially available for labeling using SPAAC (Clickchemistrytools.com).

In addition to end labeling, in many cases RNA needs to be labeled with one of the fluorophores internally at a specific site for smFRET studies to be most sensitive to a particular conformational change. Site-specific internal labeling of RNA is difficult to accomplish. One relatively easy way is to use fluorophore labeled short DNA or RNA oligonucleotides that are complementary to a single stranded region in RNA. This method has been used to study large RNAs and RNA-protein complexes such as RNase P, the group II intron, the ribosome and the spliceosome (26, 37, 58, 59). In such cases, evolutionarily less conserved and functionally less relevant hairpin loops are extended to facilitate binding of the labeled oligonucleotides. A more commonly used method for internal labeling of smaller RNAs is to end label one of two fragments (the 3’ end of the upstream fragment or the 5’ end of the downstream fragment) and ligate them together so that one of the end labels becomes an internal label (60). The fragments are individually transcribed and need to fold into the same conformation as adopted in the full length RNA. Some studies have used this strategy to label RNAs for smFRET studies and showed it to work for a two-step ligation with three fragments (61). DNA or RNA splints are commonly used to bring together two or more RNA fragments for efficient ligation. Among the available ligases, T4 RNA ligase 2 (T4 Rnl-2) is the most efficient enzyme that ligates nicks in double stranded RNA and works equally well with both DNA and RNA splints (62). Truncated versions of T4 Rnl-2 are also available that generate less by-products, but require a pre-adenylated substrate (62). The most important step in generating natively purified and internally labeled RNA is to make sure that the individual RNA fragments used for ligation do not fold into alternative structures. RNA structure probing should be performed on them or secondary structure prediction programs such as MFold (63) may be used to check whether the fragments are likely to fold into any undesired alternate conformations.

1.2. smFRET Measurement and Analysis: General Considerations

Within the last twenty years, the development of single molecule microscopy has revolutionized the field of biosciences (23, 64). Classically, questions regarding complex biological systems were investigated using bulk ensemble assays that average out rare but often significant idiosyncrasies within a given population of molecules. By contrast, single molecule experiments have the unique capability of revealing the diversity that exists from molecule to molecule. Observation of one molecule for an extended period of time and/or of many single molecules for limited periods of time will yield information related to that of the entire population as in a bulk ensemble experiment. However, single molecule observation will additionally access the interconversion kinetics of different species at equilibrium as well as reveal transient species, both features difficult to detect otherwise. Moreover, single molecule experiments are advantageous due to their use of very small quantities of sample, eliminating the multimerization effects that bulk experiments may suffer (64). To this end, a variety of techniques have been developed utilizing the single molecule regime, including single molecule fluorescence microscopy (65, 66), super-resolution imaging and super-accuracy co-localization techniques (67–70), and force microscopy (71–73). The vast array of applications available today equips scientists with the ability to obtain mechanistic information regarding biological systems that was not possible only a few years ago.

In this chapter, we describe total internal reflection fluorescence (TIRF) based smFRET microscopy. smFRET has been applied to study the folding of many biologically-relevant molecules (22, 74), including RNA (24–39, 75, 76). It constitutes a non-radiative energy transfer process in which a single dye molecule is excited by a laser. The distance between this “donor” fluorophore and a second “acceptor” fluorophore is reported in real-time through transfer of excitation energy to the acceptor based on an overlap of the donor emission with the acceptor excitation spectrum. As such, the smaller the donor-acceptor distance, the lower (more quenched) the emission of the donor dye and the greater the emission of the acceptor dye. FRET is generally quantified as a FRET efficiency (E), which is given by the Förster equation, $\text{E}={\left(1+{\left(\frac{\text{R}}{{\text{R}}_0}\right)}^6\right)}^{-1}$, where R₀ is the Förster radius (or distance) for the specified donor-acceptor dye pair at which their FRET efficiency is 50%, and R is the donor-acceptor distance. Since every donor and acceptor dye pair has its own specific Förster radius, based on the given overlap of their spectra and the assumption of isotropic relative spatial orientation of their transition dipole moments, they must be chosen carefully and in a manner that will report on the specific distance to be studied. Conversely, the molecule of interest must be labeled in such a way that the distance between the fluorophores is within ~2-fold of their Förster radius, differences in smFRET efficiency are expected between conformational states of interest, and the biological function of the RNA is retained. The process of choosing appropriate labeling sites in an RNA is ameliorated when a crystal or NMR structure is available and hypotheses exist on conformational changes of interest. If no such information is available, chemical probing methods under varying conditions may reveal residues that are accessible to solvent as suitable for labeling, as well as changes in base-pairing and tertiary interactions that lead to hypotheses on biologically relevant shape changes amenable to smFRET probing.

smFRET in general requires doubly fluorophore labeled RNA molecules. In addition, for TIRF-based smFRET RNA molecules must additionally be tethered to the surface of a quartz slide, usually through a streptavidin-biotin interaction, leading to the overall requirement of three RNA modifications. TIRF microscopy utilizes an evanescent field created by the total internal reflection of a laser beam at the quartz-water interface of the slide surface (22). This results in illumination only of molecules that are directly tethered to the slide surface, as the field penetrates only ~100 nm into the solution, thus reducing background signal. This method of reducing the illuminated volume is advantageous to others, such as diffusion smFRET in which the molecules are freely diffusing through a focused laser beam (24), as TIRF-based smFRET allows for the detection of molecules for extended periods of time and is not limited by the time a molecule spends in a confocal volume. Instead, TIRF-based smFRET observation is limited by the longevity of the fluorophores before they photobleach, which can be extended by the addition of an oxygen scavenging system. We have had particular success with the protocatechuic acid/protocatechuate-3,4-dioxygenase system (77) supplemented with Trolox to decrease photoblinking of the acceptor dye (78), even in complex reaction mixtures (37). Taken together, this TIRF-based smFRET approach allows for the detection of hundreds of single molecules in one field of view for several tens of seconds at a time. From these data, FRET efficiency histograms are constructed to reveal conformational changes, as well as the dwell times before interconversion of states (see smFRET Data Processing Section and (79)).

2. Materials

Make all buffers using autoclaved double-deionized water and store at room temperature unless otherwise specified. Exercise caution to avoid RNase contamination by using Latex gloves at all times and dispose of chemical waste following regulations.

2.1. In vitro Transcription Using T7 RNA Polymerase

1. 100 μg plasmid DNA coding for the RNA of interest followed by a cassette comprising the glmS ribozyme and biotin capture strand (see Note 1, Figure 1a and reference (20) for details). To obtain only the RNA including the cassette, the plasmid must be linearized downstream of the cassette with the appropriate restriction endonuclease (following manufacturer’s instructions) prior to the transcription reaction. Once linearized, the plasmid DNA is extracted with a phenol:chloroform:isoamyl alcohol mix (25:24:1 v/v), ethanol precipitated and resuspended in double-deionized water. Store linearized plasmid at −20°C.

2. Transcription buffer (10×): 400 mM HEPES-KOH (pH 8.0), 2 M NaCl, 250 mM MgCl₂, 50 mM DTT, stored at −20°C.

3. 5’ biotinylated DNA capture strand complementary to the binding sequence encoded in the plasmid (see Fig. 1a). Biotinylation can be performed in house or during synthesis.

4. 100 mM ATP, stored at −20°C (see Note 2).

5. 100 mM UTP, stored at −20°C (see Note 2).

6. 100 mM CTP, stored at −20°C (see Note 2).

7. 100 mM GTP, stored at −20°C (see Note 2).

8. T7 RNA Polymerase, stored at −20 °C (see Note 3).

9. Inorganic pyrophosphatase (100 U/mL, New England Biolabs), stored at −20 °C.

2.2. Native Purification of RNA

1. In vitro transcription reaction.

2. Streptavidin-coated magnetic beads (Dynabeads MyOne Streptavidin C1, 1 μm diameter, Invitrogen). Stored at 4 °C.

3. Magnetic particle collector (MPC, DynaMag -2 Magnet, Invitrogen). This MPC can handle 16 standard 1.5 – 2 mL micro-centrifuge tubes. Larger MPCs are also available that can handle tubes with volumes up to 50 mL, stored at 4°C.

4. 5 M NaCl in double-deionized water.

5. Glucosamine-6-Phosphate (GlcN6P) stock (≥ 98 % pure, Sigma-Aldrich). Store at −20°C.

6. Wash buffer (WB, 1×): 40 mM HEPES-KOH (pH 7.4), 1 M NaCl.

7. Cleavage buffer (CB, 1×): 40 mM HEPES-KOH (pH 7.4), 1 M NaCl, 10 mM MgCl₂, 200 μM GlcN6P.

8. Elution buffer (EB, 1×): 10 mM Tris-HCl (pH 8.0).

9. Phenol/chloroform/isoamyl alcohol (25:24:1 v/v, pH 5.2, Fisher Scientific).

10. Chloroform saturated in TE buffer (10 mM Tris-HCl, pH 6.0, 1 mM EDTA).

11. SYBR Gold nucleic acid stain (Invitrogen), stored at −20 °C protected from light.

12. TBE buffer (1×): 89 mM Tris-Borate, 2 mM EDTA (see Note 4).

13. Typhoon Imager (GE Healthcare).

14. Imaging Software (ImageQuant or ImageJ).

2.3. Biotinylation of the 5’ End of RNA

1. 10 μg in vitro transcribed, natively purified RNA, unlabeled on the 5’ end.

2. Antarctic phosphatase buffer (10×): 500 mM Bis-Tris-Propane-HCl (pH 6.0), 10 mM MgCl₂, 1 mM ZnCl₂, stored at −20°C(New England Biolabs).

3. T4 DNA ligase buffer (10×): 500 mM Tris-HCl (pH 7.5), 100 mM MgCl₂, 10 mM ATP, 100 mM dithiothreitol (DTT), stored at −20°C (New England Biolabs).

4. Phenol/chloroform/isoamyl alcohol (25:24:1 v/v, pH 5.2, Fisher Scientific).

5. Chloroform saturated in TE buffer (10 mM Tris-HCl, pH 6.0, 1 mM EDTA).

6. Non-denatured EtOH (>99.5% pure, Acros), stored at −20 °C.

7. 3 M NaOAc, pH 5.2.

8. 70% (v/v) EtOH in dH₂O, stored at −20 °C (see Note 5).

9. EDC reaction buffer (2×): 20 mM K-phosphate (pH 7.0), 300 mM NaCl, 20 mM EDTA.

10. Double-deionized water.

11. 12.5 mg 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC, Piercenet), stored at −20°C and kept from moisture.

12. 44 mg biotin hydrazide (Sigma B7639), stored at 4°C.

13. 0.1 M imidazole (pH 6.0).

14. Illustra NAP-5 column (GE Healthcare).

15. T4 polynucleotide kinase (10 U/μL, New England Biolabs), stored at −20°C.

16. Antarctic phosphatase (5 U/μL, New England Biolabs), stored at −20°C.

17. Refrigerated centrifuge (4 °C) with rotor for 1.5 mL tubes (Eppendorf).

18. Vacufuge vacuum concentrator (Eppendorf, optional).

2.4. Fluorophore Labeling 5’ End of RNA with NHS Ester

1. 10 μg in vitro transcribed, natively purified RNA, unlabeled on the 5’ end.

2. Antarctic phosphatase buffer (10×): 500 mM Bis-Tris-Propane-HCl (pH 6.0), 10 mM MgCl₂, 1 mM ZnCl₂, stored at −20°C (New England Biolabs).

3. T4 DNA ligase buffer (10×): 500 mM Tris-HCl (pH 7.5), 100 mM MgCl₂, 10 mM ATP, 100 mM dithiothreitol (DTT), stored at −20°C (New England Biolabs).

4. Phenol/chloroform/isoamyl alcohol (25:24:1 v/v, pH 5.2, Fisher Scientific).

5. Chloroform saturated in TE buffer (pH 6.0).

6. Non-denatured EtOH (>99.5% pure, Acros) stored at −20 °C.

7. 3 M NaOAc, pH 5.2.

8. 70% (v/v) EtOH, stored at −20 °C (see Note 5).

9. EDC reaction buffer (2×): 20 mM K-phosphate (pH 7.0), 300 mM NaCl, 20 mM EDTA.

10. Double-deionized water.

11. 12.5 mg 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC, Piercenet), stored at −20°C and kept from moisture.

12. 0.1 M imidazole (pH 6.0).

13. Illustra NAP-5 column (GE Healthcare).

14. Ethylenediamine (Piercenet).

15. 0.1 M sodium carbonate buffer (pH 8.7), stored at −20°C.

16. DMSO (Sigma-Aldrich).

17. NHS-ester derivative of the desired fluorophore, such as Cy3 Mono-Reactive Dye Pack (GE Healthcare), stored at 4°C (see Note 6).

18. T4 polynucleotide kinase (10 U/mL, New England Biolabs), stored at −20°C.

19. Antarctic phosphatase (5 U/mL, New England Biolabs), stored at −20°C.

20. Refrigerated centrifuge (4 °C) with rotor for 1.5 mL tubes.

21. Vacufuge vacuum concentrator (Eppendorf, optional).

2.5. 3’ End Labeling Using Periodate Chemistry

1. 2 nmol in vitro transcribed, natively purified RNA, unlabeled on the 3’ end.

2. 400 mM NaIO_4.

3. 3 M NaOAc, pH 5.2.

4. Non-denatured EtOH (>99.5% pure, Acros), stored at −20 °C.

5. 70% (v/v) EtOH, stored at −20 °C (see Note 5).

6. Hydrazide derivative of the desired fluorophore, such as Alexa Fluor 488 Hydrazide (Invitrogen), stored at 4°C (see Note 6).

7. Illustra NAP-5 column (GE Healthcare).

8. Double-deionized water.

9. Refrigerated centrifuge (4 °C) with rotor for 1.5 mL tubes (Eppendorf).

10. Vacufuge vacuum concentrator (Eppendorf, or other vendor, optional).

2.6. 3’ End Labeling Using Poly(A) Polymerase and ‘Click’ Chemistry

1. Yeast poly(A) polymerase (600 U/μL, Affymetrix), stored at −20°C.

2. Yeast poly(A) polymerase (PAP) reaction buffer (5×): 500 mM Tris-HCl (pH 7.0), 400 mM KCl, 3 mM MnCl₂, 0.1 mM EDTA, 1 mM DTT, 500 μg/mL acetylated BSA, 50% (v/v) glycerol (Affymetrix), stored at −20°C.

3. E. coli poly(A) polymerase (5 U/μL, New England Biolabs), stored at −20°C (see Note 7).

4. E. coli poly(A) polymerase reaction buffer (10x): 500 mM Tris-HCl (pH 7.9), 2.5 M NaCl, 100 mM MgCl₂ (New England Biolabs), stored at −20°C.

5. 10 mM ATP (New England Biolabs), stored at −20°C.

6. 50 mM phosphate buffer (pH 7.0) at 25 °C, 0.1 M KCl, 1 mM Mg²⁺.

7. 1 μmol 3’-N₃-2’,3’-ddATP (Trilink Biotechnologies).

8. DBCO (Dibenzylcyclooctyne) containing fluorophores, such as DBCO-Fluor 488 (Jena Biosciences), stored at −20°C (see Note 6).

9. Phenol/chloroform/isoamyl alcohol (25:24:1 v/v, pH 5.2, Fisher Scientific).

10. Illustra NAP-5 column (GE Healthcare).

11. 3 M NaOAc, pH 5.2.

12. Non-denatured EtOH (>99.5% pure, Acros), stored at −20 °C.

13. 70% (v/v) EtOH, stored at −20 °C (see Note 5).

14. Refrigerated centrifuge (4 °C) with rotor for 1.5 mL tubes (Eppendorf).

15. Vacufuge vacuum concentrator (Eppendorf, or other vendor, optional).

2.7. Internal Labeling of Large RNA Using Short Fluorophore Labeled Oligonucleotides

1. DNA oligonucleotides (14–16 bases long) with 5’ or 3’ amino modification complementary to unstructured regions of the RNA of interest (see Note 8).

2. NHS-ester derivative of the desired fluorophore, such as Cy3 Mono-Reactive Dye Pack (GE Healthcare), stored at 4°C (see Note 5).

3. Hybridization buffer: 50 mM Tris-HCl (pH 7.5), 100 mM KCl, 1 mM MgCl₂ (see Note 9).

2.8. RNA Ligation

1. T4 RNA ligase 2 (T4 Rnl-2, 10 U/μL, New England Biolabs), stored at −20°C.

2. T4 RNA ligase 2 reaction buffer (10×, Rnl-2 buffer): 500 mM Tris-HCl (pH 7.5), 20 mM MgCl₂, 10 mM DTT, 4 mM ATP (New England Biolabs), stored at −20°C.

3. DNA splint oligonucleotide, stored at −20°C (see Note 10).

4. DNase I (RNase-free, 2 U/μL, New England Biolabs), stored at −20°C.

2.9. Surface Treatment of Slides

1. Quartz slides (3 inch x 1 inch x 1 mm thick, G. Finkenbeiner Inc.) with holes drilled on either side to allow for input and output flow of material (see Note 11).

2. Glass cover slips (24 x 30 mm, VWR International).

3. Alconox detergent (Alconox, Inc.).

4. Double sided tape (1/2” wide, ~100 μm thick, 3 M).

5. Polypropylene coplin jars (Fisher Scientific).

6. Sonicator.

7. 70% (v/v) EtOH.

8. Methanol (Fisher).

9. 1 M KOH (see Note 12).

10. Hydrogen peroxide, 35 % (w/w) solution in water (Fisher), stored at −20 °C.

11. Ammonium hydroxide, NH₄OH, or 30 % (w/w) solution of NH₃ in water (Fisher), stored at −20 °C.

12. Propane torch.

13. Immunopure biotinylated bovine serum albumin (bBSA) (Fisher), stored at 4°C (see Note 13).

14. T50 buffer (1×): 10 mM Tris-HCl (pH 7.0), 50 mM NaCl.

15. Immunopure streptavidin (Invitrogen) (see Note 14).

16. Aminosilane reagent, APTES (N-(2-aminoethyl)-3-aminopropyltrimethoxysilane, Sigma).

17. Biotin-PEG (MW 5000, Laysan Bio Inc.), stored at −20 °C.

18. Methoxypolyethylene glycol (mPEG succinimidyl valerate, MW 5,000, Laysan Bio Inc.), stored at −20 °C in the dark.

19. PEGylation buffer (1×): 10 mM sodium bicarbonate (pH 8.5) (see Note 15).

20. PEGylation reaction mix: Dissolve 8 mg biotin-PEG, 80 mg mPEG in 320 μL PEGylation buffer in an microcentrifuge. Mix the solution thoroughly and vortex to dissolve the PEG. Centrifuge the tube at 10,000 rpm for 1 min to remove bubbles.

21. DST buffer: Prepare fresh by dissolving 10 mg DST (Sulfodisuccinimidyl tartrate?) in 350 μL PEGylation buffer. Prepare this solution immediately before use.

22. N₂ gas.

23. smFRET buffer (1×): 50 mM Tr_is-HCl (pH 7.5), 100 mM KCl, 1 mM MgCl₂.

24. OSS buffer (1×): 50 mM Tris-HCl (pH 7.5), 100 mM KCl, 1 mM MgCl₂, 5 mM protocatechuic acid (PCA), 50 nM protocatechuate-3,4-dioxygenase (PCD), 2 mM Trolox (see Note 16).

3. Methods

The following methods are used for native purification of RNA and labeling with an aim to preserve its co-transcriptionally folded conformation. At each step, care should be taken to maintain physiologically relevant conditions (pH 5 – 8, 10 – 100 mM monovalent cations and/or 1–10 mM Mg²⁺) to prevent denaturation of RNA and preserve its tertiary structure. In addition, use of high laser intensity in smFRET experiments may lead to photodamage to RNA and therefore, the signal-to-noise ratio should be optimized at the lowest laser power possible.

3.1. In vitro Transcription Using T7 RNA Polymerase

The following protocol is intended for the transcription of unlabeled RNA molecules for the study of RNA folding by smFRET.

1. This protocol for a 2 mL transcription reaction requires the use of 100 μg of linearized plasmid DNA as described in Materials section (see Note 17).

2. Prepare the following reaction in a conical 15 mL reaction tube. All reactants should be thawed completely and T7 RNA polymerase should be added to the reaction last.

   10× Transcription Buffer	200 μL
   100 mM ATP	80 μL
   100 mM UTP	80 μL
   100 mM CTP	80 μL
   100 mM GTP	80 μL
   5’-biotinylated capture strand (see Note 18)	3 μM
   Linearized dsDNA Template	50 nM
   PPiase	1.2 U
   T7 RNA Polymerase	0.1 mg/mL
   Double-deionized H2O	adjust to bring final volume to 2 mL

3. Incubate the reaction at 37 °C for 4 h.

3.2. Native Purification of RNA

The following protocol is used to purify in vitro transcribed RNAs in their native conformation. This protocol was developed for the VS ribozyme that folds into a heterogeneous population upon heat annealing but can be applied to any RNA of interest.

1. The presence of protein (RNAP, PP_iase) in the in vitro transcription reaction results in beads sticking to the sides of the microcentrifuge tubes during purification and decreases the yield of RNA. Remove protein using phenol/chloroform extraction as follows:

   1. Add an equal volume of phenol/chloroform/isoamyl alcohol to the transcription reaction and mix gently by inverting the microcentrifuge tube few times.

   2. Centrifuge the tube at 13,000 rpm for 2 min.

   3. Pipet the top aqueous layer into a fresh tube. Avoid collecting the protein at the interphase between the aqueous and organic phase.

   4. Centrifuge at 13,000 rpm for 2 min and collect the top aqueous layer. This removes any residual protein contaminants.

   5. Measure the volume and add equal volume of chloroform to the aqueous layer.

   6. Mix gently by inverting the tube a few times and centrifuge at 13,000 rpm for 2 min.

   7. Collect the top aqueous layer into a clean microcentrifuge tube.

2. Measure the volume and using 5 M NaCl stock, bring the final concentration of NaCl to 1 M.

3. Pipet 100 μL (1 mg) of streptavidin coated beads and remove the storage buffer using the MPC (see Note 19). Wash with beads with WB at least three times and remove the WB.

4. Add the phenol/chloroform extracted transcription reaction to the bead pellet and mix gently with a pipette. Tumble the tube at room temperature for 30 min. This step results in binding of RNA:capture strand complexes to the beads.

5. Remove the supernatant (S) by placing the tube in the MPC (see Note 19) and wash the beads with WB four times to remove any non-specifically bound RNA. Save the supernatant (S) and the washes (W1–W4) for later analysis using denaturing PAGE.

6. Add 100 μL of CB that contains glucosamine-6-phosphate (GlcN6P) to the bead pellet, mix a few times with a pipette and incubate at room temperature for 2 min (see Note 20).

7. Collect the supernatant which contains the natively purified RNA. The cleavage step can be repeated to increase the yield slightly.

8. Wash the beads three times with EB and save for later analysis using denaturing PAGE. On the fourth wash with EB, incubate at 70 °C for 5 min and save the supernatant (R1).

9. Wash the beads with EB one more time, incubate at −70 °C for 5 min and collect the supernatant (R2).

10. Wash the beads two times with EB and save the washes (R3–R4). These washes remove the bound RNA and regenerate the beads for more rounds of purification.

11. Place all the collected supernatants on the MPC and remove any residual beads.

12. To 20 μL of each supernatant fraction, add an equal amount of denaturing loading buffer and run a denaturing gel at 20 W for 2 h.

13. Stain the gel in 1:100,000 dilution of SYBR Gold in 1× TBE for 5 min.

14. Image the gel using a Typhoon Imager with an excitation of 488 nm, an emission bandpass of 530±10 nm and a PMT setting of 500 V. Quantification of bands can be accomplished using ImageQuant or ImageJ software.

3.3. Biotinylation of 5’ End of RNA

The following protocol is used to biotinylate unlabeled RNA molecules at the 5’ end for subsequent attachment to a solid surface for TIRF smFRET measurements.

1. In order to obtain a monophosphate at the 5’ end of the RNA for biotin attachment via EDC coupling, prepare the following reaction: 5 μg RNA, 10 μL Antarctic phosphatase buffer, and 5 units Antarctic phosphatase in a volume of 50 μL (see Note 21).

2. Incubate this reaction for 30 min at 37°C.

3. Remove the enzyme by a phenol-chloroform extraction as described in step 1 in section 3.2. If needed, the resulting stock can be stored at −20°C.

4. For monophosphate addition, prepare the following reaction: 10 μL (1 μg) of dephosphorylated RNA (from above reaction), 5 μL of 10 x T4 DNA Ligase buffer, and 20 U of polynucleotide kinase in a total volume of 50 μL (see Note 21).

5. Incubate this reaction at 37°C for 1 h.

6. To remove the enzyme, phenol-chloroform extract the RNA.

7. Precipitate the RNA with EtOH as follows:

   1. Add 2.5 volumes of non-denaturing EtOH and 1/10 volume of 3 M NaOAc (pH 5.2) to the RNA mixture.

   2. Incubate the mixture for 30 min at −20°C.

   3. Spin the mixture at 4°C for 30 min in a centrifuge at 13,000 rpm.

   4. Discard the supernatant and wash the resulting pellet with 500 μL of cold 70% (v/v) EtOH.

   5. Spin the mixture at 4°C in a centrifuge at 13,000 rpm for 30 min.

   6. Discard the supernatant and dry the resulting pellet by incubating at 37°C for ~10 min, or through the use of a Vacufuge vacuum concentrator.

8. Resuspend the RNA pellet in 37.5 μL 2× EDC reaction buffer and 37.5 μL autoclaved double-deionized water.

9. Add 12.5 mg of EDC to the RNA solution from step 8. Vortex to dissolve and centrifuge to collect contents (see Note 22).

10. Add 44 mg of biotin hydrazide to 1 mL of autoclaved water. Biotin hydrazide will not completely dissolve. This solution can be stored at −20°C for future use.

11. To the RNA/EDC solution, add 10 μL of the biotin hydrazide slurry, taking care to mix well before pipetting to ensure sufficient transfer.

12. Add 315 μL of 0.1 M imidazole, pH 6.0 (see Note 23).

13. Incubate the reaction at room temperature overnight (12–16 h) (see Note 24).

14. To remove any non-reacted EDC and its byproducts, use a NAP-5 column (see Note 25).

15. Precipitate the RNA by repeating step 7. Resuspend the RNA in dH₂O and store at −20°C.

3.4. Fluorophore Labeling of 5’ End with NHS Ester

This protocol is intended for the purpose of labeling RNA molecules at the 5’ end via NHS-ester dye.

1. Follow steps 1–10 exactly as written in section 3.3.

2. Dissolve ethylenediamine to final concentration of 0.25 M in a volume of 100 μL in 0.1 M imidazole, pH 6.0.

3. Add 12.5 mg of EDC to the RNA solution from step 1. Vortex to dissolve and centrifuge to gather contents (see Note 22).

4. Add 50 μL of the ethylenediamine solution to the EDC/RNA solution. Vortex to dissolve and centrifuge to gather contents.

5. Add an additional 200 μL of 0.1 M imidazole, pH 6.0 (see Note 23).

6. Incubate the reaction at 37°C for 3 h.

7. To remove any non-reacted EDC and its byproducts, use a NAP-5 column (see Note 25).

8. Precipitate the RNA with EtOH following step 7 from section 3.3.

9. Dissolve the resulting pellet in 20 μL of 0.1 M sodium carbonate buffer pH 8.7.

10. Dissolve one dye pack of desired fluorophore, such as Cy3 Mono-Reactive Dye Pack (GE Healthcare) in 30 μL DMSO. Add dissolved Cy3 dye to RNA mix. Wrap in aluminum foil to protect from light (see Note 26).

11. Tumble reaction mixture at room temperature for 4 h in the dark.

12. Bring total reaction volume up to 0.5 mL with autoclaved H₂O.

13. Desalt RNA and remove unreacted dye using a NAP-5 column (see Note 25).

14. Precipitate the RNA with EtOH.

15. Resuspend the pellet in autoclaved double-deionized water. Store at −20°C and keep wrapped in aluminum foil to protect the dye from light.

3.5. RNA 3’ End Labeling with Fluorophore Using Periodate Chemistry

This protocol is intended for fluorophore labeling of natively transcribed RNA at the 3’ end.

1. To oxidize the 2’-3’ diols of the RNA to aldehydes, combine the following reaction: 5 μM RNA, 2.5 μL of 400 mM NaIO₄, 13.33 μL of 3 M NaOAc, pH 5.2 in a volume of 400 μL.

2. Incubate the reaction on ice for 50 min.

3. Precipitate the RNA following step 7 from section 3.3.

4. Dissolve the resulting pellet in 13.33 μL of 3 M NaOAc, pH 5.2 and 1 mM of the hydrazide derivative of the desired fluorophore, such as Alexa Fluor 488 Hydrazide (Invitrogen) in a total volume of 400 μL.

5. Tumble the reaction overnight at 4°C in the dark.

6. Precipitate the RNA with ethanol.

7. Remove unreacted dye from labeled RNA by using a NAP-5 column (see Note 25).

8. Precipitate the RNA with ethanol.

9. Resuspend the pellet in double-deionized water. Store at −20°C and keep wrapped in aluminum foil to protect the dye from light.

3.6. Labeling 3’ End of RNA Using Poly(A) Polymerase and ‘Click’ Chemistry

Use the following protocol to label the 3’ end of RNA with a single 2’-N₃-2’-dNTP.

1. For 3’ end labeling with 2’-N₃-2’-dNTP, where N = C, G, U (as the nucleotide substrates that yeast PAP best accepts in 2’-azide modified form) use the following protocol:

   1. Assemble the reaction in 1× yeast PAP reaction buffer at 5 μM final RNA concentration, 500 μM 2’-N₃-2’-dNTP (N = C, G, U) and 24 U/μL yeast PAP.

   2. Incubate the reaction at 37°C for 20 min for 2’-N₃-2’-dCTP, 2 h for 2’-N₃-2’-dGTP or 5 min for 2’-N₃-2’-dUTP. The incubation time may vary for different RNAs and needs to be optimized.

2. For labeling with 2’-N₃-2’-dATP (as the nucleotide substrate that E. coli PAP best accepts in 2’-azide modified form), use the following protocol:

   1. Prepare the following reaction: 0.25 U/μL of E. coli PAP in 1x E.coli PAP reaction buffer, 1 mM ATP, 2.5 mM MnCl₂.

   2. Incubate the reaction at 37°C for 20 min.

   3. Remove free unincorporated 2’-N₃-2’-dNTPs using a NAP-5 column (see Note 25).

3. Perform a phenol-chloroform extraction following step 1 of section 3.2 to remove the enzyme and stop the reaction.

4. Ethanol precipitate the labeled RNA following step 7 in section 3.3.

5. Label ~1 μM of the 2’-N₃ modified RNA with 50 μM of any DBCO-label (biotin or fluorophore) in 50 mM phosphate buffer (pH 7.0), 0.1 M KCl, 1 mM Mg²⁺ for 2 h at 37°C.

6. Ethanol precipitate the labeled RNA and resuspend in a suitable buffer (see Note 27).

7. Labeling efficiency can be checked using a denaturing PAGE gel and, for fluorophore labeling, scanning with a Typhoon scanner.

3.7. Internal Labeling of Large RNA Using Short Fluorophore Labeled Oligonucleotides

Follow the protocol below to hybridize a short fluorophore labeled oligonucleotide to unstructured regions in larger RNAs that may not be amenable to the techniques listed above.

1. Label the 3’ amino-labeled oligonucleotide with the Mono-functional NHS ester dye following steps 9 – 15 of section 3.4. Use 3 – 5 nmol of oligonucleotide in 0.1 M sodium bicarbonate buffer.

2. Dissolve the fluorophore labeled oligonucleotide and the RNA of interest in hybridization buffer at 1:5 molar ratio and incubate at 37°C for 1 h (see Note 9).

3. For better hybridization efficiency, use at least 100 nM of the labeled oligonucleotide (see Note 28).

3.8. RNA Ligation

The following protocol is used for ligating two RNA molecules using a DNA splint (see Note 29).

1. Aliquot 10–20 pmol of each RNA strand to be ligated in a microcentrifuge tube.

2. Add equimolar concentration of DNA splint oligonucleotide.

3. Incubate the reaction at 37 °C for 30 min in final 1× Rnl-2 buffer concentration. This leads to formation of RNA-DNA hybrid with the 5’ phosphate and 3’ hydroxyl of both RNA strands adjacent to each other, forming a nick.

4. Add 1 U T4 Rnl-2 and incubate the reaction at 37°C for 30 min.

5. Add CaCl₂ to a final concentration of 0.5 mM to the reaction.

6. Add 1 U of DNase I and incubate at 37°C for 30 min. DNase I will degrade the DNA splint allowing the RNA to fold fully into its native conformation.

7. Remove the protein by phenol-chloroform extraction as described in step 1 in section 3.2 to stop the reaction.

8. Ligation efficiency can be checked by running a denaturing PAGE gel.

3.9. Slide Preparation Using PEG Coating and bBSA-Streptavidin Coupling

PEG coated slides are used for single molecule studies on RNA-protein complexes. PEG prevents non-specific binding of proteins to the glass surface.

1. Clean the slides using the following protocol (see Note 30):

   1. If slides have been used previously, remove the tape and epoxy resin on the slides by boiling them in water on a hotplate for ~20 min. The epoxy turns yellow when soft and can be scraped off using a razor blade.

   2. Scrub (see Note 31) the slides with a thick paste of Alconox and rinse with tap water.

   3. Sonicate the slides in a 10% (w/v) Alconox solution for 30 min in polypropyl coplin containers. Rinse the slides thoroughly with double-deionized water.

   4. Scrub and rinse the slides with methanol/ethanol and rinse with double-deionized water. Optionally, slides can be sonicated in double-deionized water for 10 min.

   5. Boil the slides in 100 mL double-deionized water, 20 mL NH₄OH, 20 mL 30% H₂O₂ (5:1:1 v/v) for 30 min. Rinse the slides thoroughly with double-deionized water. This step removes any organic material stuck to the surface and restores silanol groups on the surface. It is optional but helpful in cleaning very dirty slides.

   6. Sonicate the slides in 1 M KOH in polypropyl containers for 30 min and rinse them with double-deionized water.

   7. Sonicate the slides in double-deionized water for 10 min.

   8. Flame the slides using a propane torch slowly spending 1–2 min on each side. Cool the slides by blowing nitrogen over them or let them sit at room temperature for 10 min on clean steel rails while keeping the center of the slides free from contact (see Note 32).

2. Clean the polypropylene coplin jars using the following protocol:

   1. Sonicate in 1 M KOH for 20 min.

   2. Dispose of KOH and rinse thoroughly with double-deionized water.

   3. Rinse with methanol and sonicate in methanol for 20 min.

   4. Dispose of methanol, rinse with methanol and dry using N₂ gas.

3. Perform aminosilanzation of the slide surface using the following protocol:

   1. Allow aminosilane to reach room temperature for ~2 h in the dark.

   2. To 98 mL acetone, add 2 mL of the aminosilane reagent (APTES) in a clean beaker or flask. Approximately 70 mL is required for each reaction container with 5 slides.

   3. Pour the aminosilanization mix into the polypropylene coplin jar covering the slides.

   4. Incubate at room temperature for 10 min.

   5. Sonicate for 1 min and incubate at room temperature for 10 min.

   6. Dispose of the reaction mix and rinse thoroughly with double-deionized water.

   7. Dry the slides with N₂ gas.

4. Perform PEGylatioon using the following protocol:

   1. Clean empty pipette tip boxes to hold the slides and place water in the bottom to maintain a humid environment for the PEGylation reaction.

   2. Take out PEG bottle from the freezer and let it sit at room temperature in dark.

   3. Place slides in pipette tip boxes and carefully pipette 70 μL of PEG solution onto the middle of the slide.

   4. Place a coverslip carefully over the slide on top of the PEG solution avoiding air bubbles.

   5. Incubate for 2–3 h (or overnight) at room temperature in dark using the tip boxes.

   6. Rinse thoroughly with double-deionized water and dry in N₂ gas. Carefully pipette 70 μL of DST buffer onto each slide.

   7. Place coverslip on top of the slide and DST buffer avoiding air bubbles.

   8. Incubate at room temperature for 30 min in the dark in the pipette tip boxes.

   9. Rinse thoroughly with double-deionized water and dry in nitrogen.

   10. Assemble the slides according to your requirements.

   11. Store the slides in a dark and dry place. They could be stored in corning tubes covered by black tape or in polypropylene containers under N₂ gas.

To immobilize RNA through biotinylated-BSA coating on the quartz slides, the following protocol can be used. This protocol can be used when proteins will not be used in an smFRET experiment.

1. Clean slides as in Step 1 described previously.

2. Prepare microfluidic channel on the surface of each slide as shown in (ref) (see Note 33).

3. Flow 80 μL of biotinylated-BSA through the microfluidic channel. Allow it to bind to the glass surface non-specifically at room temperature for 10 min.

4. Wash unbound bBSA out of the channel with 100 μL of T50 buffer.

5. Flow 80 μL of streptavidin through the microfluidic channel. Allow to equilibrate at room temperature for 5 min.

6. Wash unbound streptavidin out of the channel with 100 μL smFRET buffer.

7. Flow 100 μL of ~20–50 pM biotinylated and fluorophore labeled RNA through the prepared microfluidic channel. Allow to bind for 5 min in the dark to avoid photobleaching of fluorophores.

8. Check the surface density of the immobilized RNA. If the density of molecules is low, increase the concentration of labeled RNA (see Note 34).

9. Wash unbound RNA out of the channel with 100 μL of smFRET buffer.

10. Flow 100 μL of OSS buffer through the channel to reduce photobleaching and blinking of the fluorophores. Allow to react in the dark for 5 min (see Note 35).

11. Record smFRET movies at the required time resolution (generally 10 – 100 ms). Custom made programs written in Visual C++, WinView, or ImageJ are normally used for recording smFRET movies.

3.10. Prism-Based TIRF smFRET Data Processing

Following acquisition of the raw data, they must now be processed in such a way that FRET values and kinetic parameters can be extracted (79). Generally, scripts most commonly written for the programs IDL or MATLAB are used to calculate the FRET data collected from the raw image data. Once the data are accumulated from multiple fields of view and across different channels, individual trajectories must be selected as appropriate for smFRET analysis. Before selecting traces, it is beneficial to establish selection criteria for all trajectories, as this process is often done by eye and can be subject to bias. Single molecules displaying suitable characteristics for smFRET analysis display a variety of characteristics, including, but not limited to (Fig. 2):

Fig. 2.

Characteristics of an smFRET trajectory suitable for analysis. A suitable trace will display two photoactive fluorophores (dark red, acceptor; light green, donor) that exhibit single-step photobleaching, as well as clear anti-correlation during FRET.

• Molecules must have photoactive fluorophores – establish an intensity threshold.

• If applicable, must be doubly labeled – for molecules exhibiting little-to-no FRET, a short excitation of the acceptor fluorophore at the end or intermittently throughout the trajectory can be used to probe for the presence of acceptor dye.

• If FRET is observed, molecules must display clear anticorrelation between the donor and acceptor intensities.

• Single-step photobleaching of each fluorophore towards the end of the trajectory.

After all unsuitable trajectories are discarded, the simplest way to visualize single molecule data is through the creation of FRET population histograms by sampling a small subset of all the available traces in a data set (Fig. 3). For example, FRET values within the first 100 frames of all single molecule trajectories are binned together and plotted as a histogram. When the data are fit with a sum of Gaussian functions, a variety of characteristics inherent in the data set are brought to light, including the mean FRET values sampled by the molecules, as given by the centers of the peaks of each Gaussian distribution. In addition, the relative likelihood of time spent in each state is given by the area under each peak. Finally, the relative widths of each peak can yield insight into underlying ensembles of unresolved conformations. Further analysis is needed to bore out the existence of these states, however.

Fig. 3.

smFRET data analysis (79). From a large collection of smFRET traces (top right panel), FRET population histograms (top left panel) are constructed by binning the first 10 s of FRET data (grey box around black line in FRET trace). Rate constants describing the interconversion between FRET states are derived from the cumulative dwell times spent in each FRET state (purple and orange lines in FRET trace). Finally, TDPs and TODPs are created to describe the number of transitions as well as the frequency of transitions between FRET states (purple and orange arrows in the FRET trace).

FRET population histograms are useful for the above reasons; however, they do not reveal any kinetic information for a given data set. To this end, statistical algorithm Hidden Markov Modeling (HMM) has been widely utilized to bore out the FRET states sampled in complex trajectories and the path a molecule takes in order to transition from one state to the next (79, 80). The most commonly used programs for HMM analysis on smFRET trajectories are HaMMy (80), vbFRET (81) and QuB (82). All three of these programs are publicly available and can be used to extract the dwell times of molecules in a number of FRET states (79). By plotting the cumulative population of a data set’s dwell times and fitting with a single- or multi-exponential curve, then correcting for the limits imposed by the observation time window and photobleaching (29), one can furthermore obtain the rate constant for interconversion between two FRET states (Fig. 3). Finally, transition density plots (TDPs) are a useful tool for further visualization of the transitions between FRET states, as they illustrate the frequency of transitions in a three-dimensional heat map (Fig. 3). To remove the bias towards fast, frequently occurring transitions inherent in TDPs, transition occupancy density plots (TODPs) can be used to plot the percentage of molecules displaying a certain transition. For a more in-depth discussion of data processing of single molecule trajectories, we refer to (79).

4.1. An Example: smFRET of the Thermoanaerobacter tengcongensis preQ₁ Riboswitch

Here we provide an example of an smFRET study of the translationally-operating Thermoanaerobacter tengcongensis (Tte) preQ₁ riboswitch. Binding of the preQ₁ ligand by the highly conserved Tte aptamer domain in the 5’ UTR of mRNA induces a downstream conformational change in which the Shine-Dalgarno sequence becomes occluded by formation of a pseudoknot, thus inhibiting translation initiation (13, 83). Our RNA for smFRET probing was designed with the prospect of visualizing loop L3 docking and helix P2 formation as a function of ligand binding; therefore, the RNA was labeled with the donor (DY547) dye on the 3’ end and the acceptor (Cy5) dye in the L2 loop (84). The riboswitch RNA was immobilized to the quartz slide via a 5’ biotin modification. Donor and acceptor intensities of individual molecules with the appropriate smFRET qualities were collected and representative traces are shown in Fig. 3. FRET population histograms were constructed as a function of ligand concentration, demonstrating two subpopulations consisting of a mid and high FRET conformation (Fig. 3). As ligand concentration is increased, the population of the high FRET state increases at the expense of the mid FRET state, indicating that the high FRET state represents the fully folded, ligand bound conformation seen in the crystal structure (84). Coarse-grained simulations of corresponding RNA distance distributions confirmed this hypothesis. From the same data set, dwell time analyses can be performed, in which the rate constants of interconversion between the mid and high FRET states at various ligand concentrations are compared. Additionally, these transitions can be visualized by transition density plots (TDPs) and transition occupancy density plots (TODPs) (Fig. 3).

4.2. Envisioned Application of Native Labeling Methods to smFRET of Heterogeneously Folding RNAs

The labeling methods presented in this article were chosen to complement the native co-transcriptional folding maintained by the non-denaturing RNA purification methods developed in our laboratory and others. From these protocols, one can begin with plasmid DNA, which encodes for the RNA of interest, and design various avenues to producing a homogenously folded, triply labeled RNA suitable for smFRET analysis (Fig. 4). For example, it is plausible that small RNAs (<200 nt) with hierarchal tertiary structures can be labeled through the RNA ligation method depicted in Fig. 4a. Since RNA structures fold in a hierarchal manner such that secondary structure precedes tertiary structure formation, it is expected that for many RNAs 5’ and 3’ segments can be transcribed separately from one another. These shorter RNA fragments can be modified using the protocols presented here to incorporate biotin and fluorophore at the 5’ and 3’ ends, preserving the 3’ end of the 5’ RNA, and vice versa. Alternatively, the central portion can be chemically synthesized with an internal amino-, azido- or sulfhydryl-modification for subsequent conjugation with a fluorophore following ligation via a DNA splint. By contrast, longer RNAs may not be amenable for labeling using the conjugation methods presented. They can, alternatively, be internally labeled by hybridizing a fluorophore or biotin labeled synthetic oligonucleotide to extended and/or unstructured regions of the RNA (Fig. 4b). To obtain high melting temperatures between the oligonucleotide and RNA, locked nucleic acids (LNAs) may be used. Although this method does not result in covalent modification of the RNA with the fluorophore or biotin, a high melting temperature of the hybridizing oligonucleotide ensures that the interaction will be sufficiently stable under most experimental temperatures. It is envisioned that the methods presented in this article will be increasingly used to study RNA dynamics, kinetics of RNA-RNA or RNA-protein interactions particularly for heterogeneously folding RNA to better understand the molecular mechanisms of non-coding RNA function in vivo.

Fig. 4.

Creating a triply-labeled RNA for smFRET analysis. Panel (a): For shorter RNAs that may exhibit hierarchal folding behavior, the RNA can be segmented into three pieces. The 5’ and 3’ pieces may be transcribed, natively purified and 5’ and 3’ end labeled, whereas the middle piece can be synthetically produced with an amino, sulfhydryl or azide label. Following a two-step splinted ligation, the synthetic modification is used to conjugate a fluorophore using the protocols presented herein. Panel (b): For longer RNAs, hybridization of a complementary labeled oligonucleotide is used to create an internal fluorophore (or biotin) label.