RNAs are key players in the central dogma of molecular biology. Messenger RNA (mRNA), ribosomal RNA (rRNA) and transfer RNA (tRNA) participate in protein synthesis, while the group of non-coding RNAs (ncRNAs) are crucial for gene regulation as interfering RNAs (miRNAs) and guide RNAs. As RNAs fulfill so many important and diverse functions, cells need more than four building blocks and thus a large chemical diversity of ribonucleosides is found in all domains of life (Boccaletto et al., 2018). Studies on RNA and its modifications have evolved in the last decades and the sensitivity and depths of analysis have improved. One key aspect of this development is the use of stable isotopes as probes, labels or standards. These stable isotopes comprise deuterium (hydrogen-2), carbon-13, nitrogen-15, oxygen-18, fluorine-19 and sulfur-34. Especially methods like nuclear magnetic resonance (NMR) and mass spectrometry (MS) improved significantly through modern isotope labeling techniques. Here, we want to review the current approaches in isotope labeling techniques. These include metabolic labeling, enzymatic approaches like in vitro transcription and of course solid phase synthesis. In the second part of this review, we will focus on the methods that utilize the labeled RNAs or nucleosides and how our understanding of RNA structure, interactions and modifications improved.
Methods for incorporation of stable isotopes into RNA
The production of RNA is well established and possible by many techniques. Only recently, a guide to large-scale RNA sample preparation was presented (Baronti et al., 2018). RNA isotope labeling can be divided into radioactive labeling (e.g. with phosphorus-32 or tritium) and stable isotope labeling. While radioactive labeling is still a valuable technique, no technological improvements were made in the last decades. This is in stark contrast to stable isotope labeling, which has boosted the methodological possibilities of many techniques. Stable isotopes became quantitatively available for basic research in the 1940s as recently reviewed (Wilkinson, 2018). The first stable isotope labeling of RNA was done metabolically. Bacteria were cultured in the presence of carbon-13 and nitrogen-15 containing nutrients and the isotopes were found to be incorporated into the nascent RNAs by the bacterial metabolism (Nikonowicz et al., 1992). From this RNA, stable isotope labeled nucleotides were isolated and used for enzymatic production of an RNA transcript of interest. Here, the isotope label is distributed equally throughout the transcript. Later techniques, such as position selective labeling of RNA (PLOR), allowed site-specific incorporation of stable isotope labeled nucleosides (Liu et al., 2015, 2016a,b). The site-specific incorporation is often crucial to the success of the subsequent application and solid-phase synthesis is a key tool in that respect. In summary, stable isotope labeling techniques can be divided into (i) metabolic approaches, (ii) enzymatic techniques and (iii) solid phase synthesis.
Metabolic labeling, also referred to as biosynthetic labeling, relies on simple organisms, which are able to incorporate stable isotope labeled nutrients into their biomolecules. The first studies were conducted in bacteria such as Escherichia coli, where growth media with a single carbon or nitrogen source were used. With isotopically labeled glucose or ammonium sulfate as nutrients a complete labeling of the nucleic acids with carbon-13 and nitrogen-15 was achieved (Nikonowicz et al., 1992). Later, stable isotope labeling was achieved in Saccharomyces cerevisiae (Kellner et al., 2014a,b; Heiss et al., 2017), in the algae Chlamydomonas reinhardtii (Sarin et al., 2018) and also in the multicellular organism Caenorhabditis elegans (van Delft et al., 2017). These approaches are uniform labeling approaches, as all carbon and nitrogen atoms become labeled. By combining bacterial knockout strains with differentially labeled nutrients such as 1,3-13C2-glycerol, an atom-specific isotope labeling becomes possible and, e.g. all carbons, except C4′ become carbon-13 labeled (Thakur et al., 2012). Both NMR and MS studies commonly use metabolic labeling, but the in vivo production of labeled RNAs is a key tool to NMR and thus described in more detail. Metabolic labeling has also proven to be a fundamental tool for the identification of bacterial communities that share similar metabolic pathways using RNA-stable isotope probing (RNA-SIP).
In 1992, Pardi and coworkers established the field by feeding E. coli with 13C-glucose and 15N ammonium sulfate (Nikonowicz and Pardi, 1992; Nikonowicz et al., 1992). The bacterial RNAs were isolated and enzymatically digested to the 5′-nucleotide-monophosphates before conversion into 5′-nucleotide-triphosphates (NTPs). The stable isotope labeled NTPs were used in DNA templated T7 in vitro transcription, which is discussed in more detail in the next section. These early techniques used a single carbon or nitrogen source in the bacterial growth media. From these nutrients, the cell formed the amino acids glycine, aspartic acid and glutamine, which are in combination with the carbon donor tetrahydrofolate (THF), carbon dioxide (CO2) and the phosphoribosylpyrophosphate (PRPP) building block the main components during nucleotide biosynthesis (Figure 1A). Carbon-13 labeled PRPP, THF, CO2, aspartic acid and glycine are formed from 13C6-glucose metabolism which leads to the incorporation of carbon-13 into the nucleotides. 15N-ammonium sulfate as a nutrient leads to 15N-labeled aspartic acid, glutamine and glycine, which are the nitrogen donors of nucleotide biosynthesis. In pyrimidines the N1, C4, C5 and C6 positions derive from aspartic acid (Asp), the C2 from CO2 and the N3 from glutamine (Gln). The N4 of cytidine is incorporated from a glutamine (Gln). Purine position N1 and the exocyclic N6 of adenine are from aspartic acid (Asp). Position C4, C5 and N7 are from glycine (Gly). Positions N3, N9 and the exocyclic N2 position of guanine are from glutamine (Gln). Positions C2 and C8 are from THF and the remaining position C6 derives from CO2 incorporation. A color-coded overview is given in Figure 1A. After de novo nucleotide synthesis, dNTPs and rNTPs are used for DNA synthesis and RNA transcription and stable isotope labeled nucleic acids become available.
Uniform labeling with stable isotopes is commonly used in MS to identify and study modified nucleosides. The complete substitution of all, e.g. carbon or nitrogen atoms by stable isotope labeling has become a key tool for sum formula generation and is very helpful for structure prediction and verification by MS (Dumelin et al., 2012; Kellner et al., 2014a,b; Thiaville et al., 2016; Dal Magro et al., 2018). Quantification of modified nucleosides requires the availability of stable isotope labeled internal standards. Due to the high chemical diversity, metabolic labeling is an optimal tool for their production (Waghmare and Dickman, 2011; Popova and Williamson, 2014; Kellner et al., 2014a,b; Miranda-Santos et al., 2015).
In addition to discovery and quantification purposes, isotope labeling has been found to be a valuable tool to study the dynamics of RNA modifications in vivo. The new concept is abbreviated NAIL-MS (nucleic acid isotope labeling coupled mass spectrometry) and was introduced by our lab in 2017 (Heiss et al., 2017; Reichle et al., 2018a, b). Here, the benefits of uniform metabolic labeling are exploited, as described later in this review.
Atom-specific nucleoside labeling
An early study suggested acetate based metabolic labeling for atom-specific labeling of the nucleobases (Hoffman and Holland, 1995). For instance, when grown on minimal media supplemented with 1-13C-acetate, the wild-type K12 E. coli incorporates 13C at the C2 and C4 base carbons of the pyrimidines, at the C4 and C6 base carbons of the purines, and at the C3′ position of the ribose ring (3Hoffman and Holland, 1995). Later genetic variants of E. coli with mutations in the citric acid (TCA) or the oxidative pentose phosphate pathways were fed with 13C-glucose or D8-glycerol to achieve the desired isotope labeling patterns in nucleotides (Johnson et al., 2006). An E. coli mutant strain that lacks succinate dehydrogenases and grown on [3-13C]-pyruvate, affords ribonucleotides with atom-specific labeling at C5′ (~95%) and C1′ (~42%) and minimal enrichment elsewhere in the ribose ring. Enrichment is also achieved at purine C2 and C8 (~95%) and pyrimidine C5 (~100%) positions with minimal labeling at pyrimidine C6 and purine C5 positions. These labeling patterns contrast with those obtained with a malate dehydrogenase knockout E. coli grown on [1, 3-13C]-glycerol for which the ribose ring is labeled in all but the C4′ carbon position (Thakur et al., 2012). An E. coli strain deficient in the transketolase gene (tktA) and grown on glucose that is labeled at different carbon sites is shown to facilitate cost-effective and large-scale production of useful nucleotides. These nucleotides are site specifically labeled in C1′ and C5′ with minimal scrambling within the ribose ring as shown in Figure 1B (Thakur et al., 2012). A summary of all atom-specific labeled nucleosides is given in Table 1. A clear benefit of atom-specific carbon-13 labeling is that unwanted one bond 13C-13C scalar and dipolar couplings are completely removed. Together with uniformly labeled nucleotides, these atom-specifically labeled nucleotides are mostly used for the production of labeled RNAs with in vitro transcription, as described in the section ‘In vitro transcription (IVT) and other enzymatic approaches’, but the metabolic approaches may also be used to produce specifically labeled RNAs in vivo using the tRNA- or 5S-scaffold technologies as described later.
In vivo production of labeled RNAs
Producing recombinant RNA in vivo is hampered by difficulties, such as degradation by RNases, heterogeneity of the products and problems in purifying the RNA from cell extracts. In vivo production of labeled RNAs were first motivated by the need to produce isotopically labeled ribosomal RNA (rRNA) or tRNAs with post-transcriptional modifications required for their folding and function. For instance, the Vibrio proteolyticus 5S rRNA (Moore et al., 1988) and the human tRNALys3 (Tisne et al., 2000) have been successfully expressed and uniformly 15N or 15N/13C-labeled in E. coli. The success of these productions relies on the fact that these RNAs are recognized by the cellular machinery, processed precisely, post-transcriptionally modified and are not subjected to 3′-polyadenylation, which triggers RNA degradation. Therefore, the recombinant RNA is well-folded and protected from degradation by nucleases, it can be isotopically labeled, and produced in milligram quantities for NMR by growing cells in commercially available isotopically enriched bacterial growth medium (Moore et al., 1988; Wallis et al., 1995; Tisne et al., 2000). The recombinant RNAs are then recovered by phenol extraction and purified by anion-exchange chromatography. Generic approaches for expressing and purifying any structured RNA in E. coli were subsequently developed using tRNA (Ponchon and Dardel, 2007) or 5S rRNA as scaffolds (Pitulle et al., 1995; Stepanov and Fox, 2015). The coding sequence of the RNA of interest is introduced in the anticodon loop of tRNA or replaces a stem-loop of the Vibrio proteolyticus 5S rRNA in order to hijack the host machinery and escape cellular RNases. In those systems, the RNA of interest is produced as a chimera of the RNA of interest embedded in the tRNA or the 5S rRNA. It is often desirable to study the RNA of interest without the RNA scaffold. Different strategies for selectively cleaving the RNA of interest from the scaffold were developed, either using RNaseH (Ponchon et al., 2009), DNAzyme (Liu et al., 2010) or cis-acting hammerhead (HH) ribozymes flanking the termini of the RNA of interest in the plasmid expressing the construct (Nelissen et al., 2012, 2015). With these approaches, one can only produce RNAs uniformly labeled with 15N and 13C isotopes. Large, uniformly labeled RNAs generated using these methods still have 13C-13C coupling, leading to NMR signal loss and low-resolution NMR spectra. A method to produce milligram quantities of uniform 15N- and atom-specific 13C-labeled RNAs using wild-type K12 and mutant tktA E. coli in combination with a tRNA-scaffold approach was developed to overcome this problem (Aoyagi et al., 2015). Different atom-specific labeling possibilities have been described in the section ‘Atom-specific nucleoside labeling’), possibilities that can readily be exploited to produce atom-specifically labeled RNAs in vivo. Finally, the tRNA-scaffold methodology has been extended to allow RNA/protein co-expression in E. coli (Ponchon et al., 2013). This last improvement opens many applications, including the production of nuclease-sensitive RNAs encapsulated in viral protein pseudo-particles, the co-production of non-coding RNAs with chaperone proteins, the incorporation of post-transcriptional RNA modifications by co-production with appropriate modifying enzymes and finally the production and purification of an RNA-His-tagged protein complex by nickel affinity chromatography. These tools pave the way to large-scale structural investigations of RNA function and interactions with proteins. They allow the production of milligram quantities of uniformly or atom-specifically-labeled RNA and offer the advantage of benefitting from the cellular machinery to fold the RNA of interest properly.
Undefined labeling for stable isotope probing (SIP)
The concept of SIP was introduced in 1998 by Boschker et al. Here, a collection of microbes is fed with stable isotope labeled carbon sources, e.g. 13C-methane. Only bacteria capable of utilizing methane will incorporate the carbon-13 into their DNA and by density-gradient centrifugation, the DNA of methane-metabolizers and non-metabolizers is separated. The identity of the methanophilic bacteria is subsequently revealed by analysis of the stable isotope labeled DNA (Boschker et al., 1998). Later, DNA-SIP was the basis for the development of rRNA-SIP. Instead of DNA isolation, ribosomal RNA of the bacteria is isolated and used for species identification by 16S sequencing. With this approach, methylotrophic (Lueders et al., 2004), butyrate-degrading (Hatamoto et al., 2008), perchloroethane dehalorespiring (Kittelmann and Friedrich, 2008), sialic acid consuming (Young et al., 2015) or glucose consuming (Herrmann et al., 2017) members of microbial communities were identified. A comprehensive review can be found by Whitley et al. (Whiteley et al., 2007).
In vivo transcription (IVT) and other enzymatic approaches
In vivo transcription
RNAs larger than 10 nucleotides can be produced by in vitro transcription using the T7 RNA polymerase enzyme and isotopically labeled NTPs can easily be incorporated in these enzymatic reactions. The use of different types of NTP labeling enables the production of a large variety of RNA samples with a designed isotopic labeling scheme relevant for the downstream NMR or MS studies. For instance, some isotope labeled nucleotides can be obtained commercially, produced by metabolic labeling (Nikonowicz et al., 1992) or from chemical synthesis (Alvarado et al., 2014a,b). It is also possible to use stable isotope labeled nucleobases, e.g. purines labeled at position 8 with carbon-13 and to convert these enzymatically to the corresponding NTP (Longhini et al., 2016a,b). Commonly used NTPs comprise 15N-GTP and 15N-UTP (Sochor et al., 2016), 15N/13C-GTP (Paulines and Limbach, 2017), ATP with a fluor-19 label at position 2 (Sochor et al., 2016) or 13C2′ and 13C4′ ribose labeling (Johnson and Hoogstraten, 2008). An overview of some building blocks can be found in Figure 2A .
In addition to uniform labeling, nucleotide-specific labeling is a very common labeling scheme used for NMR studies. In this type of labeled RNA samples, each of the four different nucleotides may harbor a different labeling scheme, such as 13C, 15N-labeled for one or two nucleotide types, the remaining being unlabeled or perdeuterated. Although this type of labeling greatly simplifies NMR spectra, multiple samples with complementary labeling are often required for a complete structural characterization of the RNA of interest (Kang et al., 2014; Bonneau et al., 2015; Barnwal et al., 2016; Bourbigot et al., 2016; Imai et al., 2016; Wolter et al., 2017). Atom-specific labeling adds further to the diversity of the RNA labeling schemes and consists in the use of NTPs with defined isotope labeling at specific positions in the enzymatic in vitro production of the RNA of interest. Atom-specific labeling may be implemented with a uniform labeling type or strategy, in which every nucleotide of the RNA carries the same type of position specific labeling, but is mainly used in combination with a nucleotide-specific labeling scheme (D)’Souza et al., 2004; Bullock et al., 2010; Barnwal et al., 2016; Imai et al., 2016; Beusch et al., 2017; Wolter et al., 2017). Atom-specific labeled NTPs may be obtained from commercial sources or prepared in the laboratory from a long series of enzymatic reactions (Batey et al., 1992, 1995; Scott et al., 2000), or from a combination of chemical and enzymatic syntheses (Thakur et al., 2010, 2012; Thakur and Dayie, 2012; Alvarado et al., 2014a,b; Le et al., 2015). The types of atom-specific labeled NTPs that can be obtained from all possible sources are highly diverse and virtually countless, the choice of labeling being driven by the considered applications, scientific questions and the particularities of the system under investigation. Several examples from the literature will be given in section NMR spectroscopy, but options for site-specific labeling are not limited to those. In general, a combination of several samples with well-chosen complementary labeling schemes is necessary for complete characterization of the RNA of interest (D)’Souza et al., 2004; Bullock et al., 2010; Barnwal et al., 2016; Imai et al., 2016; Longhini et al., 2016a,b; Beusch et al., 2017; LeBlanc et al., 2017; Wolter et al., 2017). In addition, unlabeled NTPs may be deuterated via 1H/2H exchange at specific positions of the bases following simple procedures (Huang et al., 1997; Nikonowicz, 2001; Lu et al., 2010; Duss et al., 2012; Keane et al., 2015). For instance, H8 protons of purines can easily be exchanged with deuterium. H5 and H6 protons of pyrimidines can also specifically be exchanged with deuterium, but protocols are less straightforward and effective than 1H/2H exchange at the C8 position of purines. Conversely, commercially available perdeuterated NTPs may be utilized to re-introduce protons at specific positions to facilitate assignment procedures and structural studies (D)’Souza and Summers, 2004; D’Souza et al., 2004; Miyazaki et al., 2010; Keane et al., 2016). Overall, the different labeling possibilities of each individual NTP combined with the intrinsic flexibility of RNA production by in vitro transcription provides the means to produce a multitude of RNA samples with different labeling patterns and complementary information. A thorough structural investigation of RNAs often requires several of these complementary samples, which renders RNA studies with isotope labeling considerably more expensive than studies of unlabeled RNA.
Segmental isotope labeling
When the size of the RNA increases, the labeling schemes described already are not sufficient to reduce the number of overlapping resonances in NMR. In such cases, segmental isotope labeling provides the means to simplify the NMR spectra while performing the study in the context of the full-length RNA. Several protocols to ligate two or more fragments using T4 DNA ligase or T4 RNA ligase have been reported (see Figure 2B) (Xu et al., 1996; Kim et al., 2002; Tzakos et al., 2006, 2007; Nelissen et al., 2008; Duss et al., 2010, 2012; Miyazaki et al., 2010; Lu et al., 2011). Any kind of differential isotope labeling patterns for the different fragments to be ligated can be conceived, the fragments being most of the time produced by in vitro transcription using different sources of NTPs as described. It is worth noting that such ligation procedures can be employed to introduce labels for paramagnetic relaxation enhancement or electron paramagnetic resonance studies (Buttner et al., 2013; Duss et al., 2014a,b,c, 2015; Lebars et al., 2014). To avoid heterogeneity in the segmental isotope labeled samples, RNA fragments must be produced with defined 5′ and 3′ extremities. For instance, 3′ heterogeneity in the upstream fragment would indeed lead to heterogeneity and non-natural nucleotide incorporation at the site of ligation. Different strategies employing the HH ribozyme, the Varkud satellite (VS) ribozyme, the hepatitis delta virus (HDV) ribozyme or RNase H cleavage have been employed towards this goal (Xu et al., 1996; Kim et al., 2002; Duss et al., 2010; Lu et al., 2011; Miyazaki et al., 2010). Segmental isotope labeled RNAs are mostly used in downstream NMR applications but may also be used in MS to localize protein-RNA interactions simultaneously at amino acid and nucleotide resolution (Dorn et al., 2017).
Position-selective labeling of RNAs (PLOR)
The production of long RNA with position-selective stable isotope labeling is challenging. A solution to the problem is PLOR, a technology that combines aspects of solid phase chemical synthesis and liquid-phase enzymatic synthesis of RNAs (Liu et al., 2015). The concept is briefly summarized in Figure 2C. A detailed description of the technique was presented by the Wang laboratory (Liu et al., 2016a,b). In principle, the DNA template is fixed to a solid support and thus enzyme, buffers and NTPs can be exchanged by washing steps. Initially, only three of the conventionally used NTPs are added upon T7 transcription initiation. The transcript is elongated to the point, where the missing NTP is supposed to be incorporated. Now, the reaction chamber is washed and all NTPs are removed before the missing NTP, containing the label of choice, is added and transcription continues. After washing, the next set of NTPs is added and transcription continues and pauses as designed by the scientist. Thus, long RNA with labeled nucleotides at desired positions can be produced.
Chemical synthesis and solid phase synthesis
The ideal way to control the position and type of stable isotope labeling in a nucleoside or oligonucleotide is chemical synthesis. Today, many chemicals with stable isotopes like deuterium, carbon-13 or nitrogen-15 are available and these are utilized during the synthesis of RNA building blocks. The Carrel lab pioneered the synthesis of stable isotope labeled modified nucleosides in both RNA and DNA (Bruckl et al., 2009; Globisch et al., 2010; Brandmayr et al., 2012; Iwan et al., 2018). The ribonucleosides were used for absolute quantification of modified nucleosides in tRNA from tissue samples (Brandmayr et al., 2012).
The chemical production of RNA strands is commonly done by solid phase synthesis. The process starts with a nucleotide fixed through its 3′ OH to a solid support. The 5′ end is DMT (dimethoxytrityl) protected and deblocked before addition of the next phosphoramidite building block. The free 5′OH and the deprotected phosphoramidite react to form a phosphite linkage, which is subsequently oxidized to the phosphate. From here on, the cycle can be repeated, and the nucleic acid chain grows from the 3′ to the 5′ end (and thus inverse to the natural formation of nucleic acids). A detailed description of the method can be found from the Petzold lab (Baronti et al., 2018). With this technique, RNA of up to 100 nucleotides can be synthesized according to a pre-programed sequence. Especially NMR studies have benefitted from the use of stable isotope labeled RNA phosphoramidites, some of which became very recently commercially available. The used building blocks allow atom-specific labeling of, e.g. pyrimidines labeled at position 6 with 13C (Wunderlich et al., 2012, 2015), purines labeled at position 8 with 13C, 1,3-15N-uridine, 2,8-13C-inosine and in addition a 2′-cyanoethoxymethyl for protection of the 2′OH of the ribose (Kremser et al., 2017). Furthermore, completely 13C5 labeled ribose (Wenter et al., 2006), deuterated building blocks (Huang et al., 1997) or 13C615N1-labeled inosine (Dallmann et al., 2016) were presented. A phosphoramidite building block and its potential stable isotope labeling sites can be found in Figure 3 .
It is also possible to use non-labeled building blocks and incorporate heavy oxygen-18 during solid phase synthesis by using 18O-H2O for the production of 18O-phosphate labeled siRNA (Hamasaki et al., 2013). The labeled siRNA was used to determine the blood concentration of 18O-labeled RNA after administration to mice by MS. In addition, the localization of the labeled RNA was visualized by isotope microscopy.
Use of isotopically labeled RNAs
Stable isotope labeled RNA and nucleosides have boosted our understanding of RNA structure and RNA modifications. Structure determination by NMR spectroscopy of RNA longer than 40 nucleotides has become possible due to the reduced and controllable chemical shift overlaps of resonances. Thus, spectra interpretation is simpler and signal intensities could be improved. Especially segmental and atom-specific isotope labeling have changed the possibilities of RNA NMR spectroscopy to the point that folding and dynamics of continuously larger RNAs become accessible. Similar advances have been made in the mass spectrometric analysis of modified RNAs. Stable isotope labeled RNAs are crucial for absolute quantification of modified nucleosides by all mass spectrometric techniques. In addition, new approaches, which allow the assessment of dynamic modification and demodification processes, have been made possible by metabolic stable isotope labeling. The recent progress in NMR spectroscopy and MS is summarized in the following two chapters.
NMR spectroscopy is a highly powerful method for structural studies of RNAs. Due to their rather uniform negatively charged surface, their flexibility and propensity to conformational heterogeneity, RNAs are in general less amenable to structural studies with X-ray crystallography than proteins. NMR in contrast can very well deal with more dynamic molecules and this intrinsic property of NMR places it at a central position for structural studies of RNAs. Indeed, whereas NMR accounts for only 10% of the protein structures deposited at the PDB, around 40% of the deposited RNA structures were solved with NMR spectroscopy. However, NMR studies of RNAs suffer from two major problems, which become even more salient when the size of the RNA increases, namely chemical shift overlap of resonances and line broadening leading to complete signal loss. The different strategies involving isotope labeling, which have been used to solve these two major issues, have been presented from a technical perspective in the previous section (for recent reviews see also Lu et al., 2010; Duss et al., 2012; Longhini et al., 2016a,b; Barnwal et al., 2017). It is worth noting that in contrast to liquid-state NMR, solid-state NMR does not suffer from increased line broadening when the size of the system increases. Although RNA structure determination by solid-state NMR is still in its infancy, the remarkable advances of this field in recent years suggest that spectacular progresses are to be expected in the coming years (Marchanka et al., 2015, 2018). Regardless of the NMR technique used in fine, liquid-state NMR or solid-state NMR, the elaborated techniques of RNA isotope labeling presented in the previous section enabled the study of systems with increasing size and complexity. To illustrate these points, recently reported systems for which advanced isotope labeling techniques were central to the NMR investigations will be presented in the next paragraphs.
Whereas small RNAs can be studied using uniform 13C/15N-labeling, the study of RNAs larger than 30–40 nucleotides is extremely difficult without specific labeling. One of the most common strategies to overcome the size problem in NMR of RNAs is the use of deuteration at specific sites to simplify the spectra of the RNA, and to improve the intensity of the signals from the remaining protonated sites. For instance, in combination with a traditional ‘divide and conquer’ strategy, Varani and coworkers recently solved the structure of the 68 nucleotide-long CssA thermometer from Neisseria meningitidis with the use of partially deuterated (H6/H8/H2, H1′, H2′, D3′, D4′, D5′/D5′′ and D5) rNTPs (Barnwal et al., 2016). Validation of the initial assignments and additional structural restraints were obtained using these site-specifically deuterated NTPs on ribose positions and on the pyrimidine C5 position. In combination with other biophysical methods, this elegant NMR structural work showed that the RNA thermometer acts as a rheostat, the stability of which is optimized to respond to small temperature variations (Barnwal et al., 2016). Similarly, Wagner and coworkers recently solved the structure of the 108 nucleotide-long J-K region of the encephalomyocarditis virus (EMCV) IRES RNA and studied its interaction with the HEAT-1 domain of eukaryotic initiation factor 4G (eIF4G) using different types of partially deuterated samples (Imai et al., 2016). An RNA sample in which adenosine and cytidine residues are protonated with an otherwise deuterated background (u-2H, 1H-Ade, 1H-Cyt) was essential to define important long-range interaction in the central region of the IRES. In addition, an RNA sample in which only H1′, H2′, H2 and H8 of adenosines are protonated within a deuterated background [u-2H, (H1′, H2′, H2, H8)-Ade] was crucial for the delimitation of the interaction surface between the IRES RNA and the HEAT-1 domain of eIF4G (Imai et al., 2016). In order to determine the structure of the 44 nucleotide-long Drosophila K10 transport and localization signal RNA (K10 TLS), Lukavsky and coworkers used site-specific deuteration of pyrimidine C5 positions which were crucial to resolve resonance overlap in the H1′/H6/H8 region by suppressing the strong H5-H6 crosspeaks in homonuclear NOESY spectra (Bullock et al., 2010). This atom-specific labeling was essential in the structure determination process of the K10 TLS, which consist of two segments of double-stranded RNA helix adopting an unusual A′-form conformation with widened major-grooves that act as critical recognition sites for the transport machinery (Bullock et al., 2010). Similarly, for studying the interaction of an intronic splicing silencer (ISS) sequence of 21 nucleotides with the RNA-binding domains of hnRNP A1, Allain and coworkers used atom-specific deuteration of a protein/RNA sample with the RNA deuterated on specific positions, namely C8 of purines and C5 of pyrimidines (Beusch et al., 2017). This simplified homonuclear NOESY spectra in the H1′/H6/H8 region and helped identifying H2 of adenines in the long single-stranded ISS RNA. This enabled protein/RNA inter-molecular contacts to be identified, which demonstrated that both RNA recognition domains of hnRNP A1 can bind simultaneously to the single bipartite motif found in the ISS RNA (Beusch et al., 2017). Another example of the usefulness of nucleotide site-specific deuteration is provided by the structure determination of the 59 nucleotide-long Streptococcus pneumoniae preQ1-II riboswitch bound to prequeuosine (preQ1), for which Feigon and coworkers used a partially deuterated (H6/H8/H2, H1′, H2′, D3′, D4′, D5′/D5′′ and D5) RNA sample for the classical H1′-base proton sequential assignments (Kang et al., 2014). This structure, in combination with dynamic and mechanistic investigations, revealed the complex molecular mechanism by which the preQ1-II riboswitch regulates gene expression (Kang et al., 2014).
Segmental isotope labeling
Another strategy to further simplify the NMR spectra and resolve ambiguities from overlapping resonances relies on segmental isotope labeling of the RNA under investigation. Contrary to the traditional ‘divide and conquer’ strategy for RNA structure determination by NMR, segmental isotope labeling has the great advantage to simplify the NMR spectra while performing the study in the context of the full-length RNA. However, the preparation of segmentally labeled RNA samples remains highly demanding and the ‘divide and conquer’ strategy remains a method of choice in the many cases where no interaction exist between the different sub-domains of the larger RNA assembly. In the case of the 127 nucleotide-long non-coding RNA RsmZ that binds three RsmE dimer proteins at six distinct GGA binding-sites, Allain and coworkers used several 13C/15N segmental isotope labeled RNAs to decipher the order of the sequential binding events of each RsmE dimers on the RsmZ RNA (Duss et al., 2014b,c). In addition, segmentally labeled RNAs were used within an original structure determination protocol to find the exact parts of the isolated GGA binding-site complexes, which were solved in a previous study (Duss et al., 2014b,c), that retain their structure in the full complex. This required the segmental labeling of each single GGA binding-site at a time within the full RsmZ RNA. Overall, this study proposes that RsmZ is well tuned to sequester, store and release RsmE and serves the role of a protein sponge (Duss et al., 2014b,c). For the structure determination of the 57 nucleotide-long 5′ hairpin of the 7SK RNA, Lebars and coworkers used a segmentally labeled RNA sample in combination with the traditional ‘divide and conquer’ strategy (Bourbigot et al., 2016). This segmentally labeled sample, which was also selectively labeled on adenosines (13C/15N-Ade), was used to validate and extend the initial assignments obtained for the full-length construct. Overall, the RNA structure obtained with this strategy suggested that binding of its protein partners to the 7SK RNA may originate from a conformational selection mechanism involving a particular motif at the top of the 7SK RNA (Bourbigot et al., 2016).
An ultimate strategy to further simplify NMR spectra to an extreme level relies in the atom-specific labeling of RNAs. This type of labeling may be implemented in a uniform background using atom-specifically labeled NTPs in in vitro transcription, meaning that identical nucleotides will carry the same type of atom-specific labeling. But, atom-specific labeling best reveals its impressive potential when implemented with RNA chemical synthesis, which gives the means to control the exact positions at which isotopes are introduced. With a few well-chosen isotope-labeled atoms in an RNA construct, NMR signal ambiguities are easily resolved, which can be decisive for the assignment and structure determination of RNAs. This approach also shows a great potential for mechanistic studies of RNAs, such as the investigation of their dynamic properties, including folding processes and structural rearrangements (Wunderlich et al., 2012, 2015; Dallmann et al., 2016; Strebitzer et al., 2018). For instance, in the structure determination process of the 34 nucleotide-long GTP-binding RNA aptamer, Wöhnert and coworkers used 10 different atom-specifically labeled RNA samples with 13C-C8 labeling of purines and 13C-C6 labeling of pyrimidines at one or two controlled positions within the full-length RNA construct (Wolter et al., 2017). In combination with uniformly labeled and nucleotide-specifically labeled RNAs, these atom-specifically labeled RNA samples were essential to verify and extend sequence specific assignments for aromatic CH-groups and to obtain a precise NMR structure of this RNA adopting a complicated fold with two base triplets and a base quartet with an original geometry (Wolter et al., 2016, 2017). To decipher whether the destabilizing effects of N6-methyladenosine (m6A) post-transcriptional RNA modification in A-U base pairing could induce changes in RNA secondary structure, Al-Hashimi and coworkers used RNA constructs with two atom-specific 15N-labeling. A guanosine 5′ to the m6A site was labeled on its N1 position and a uracil on the opposite strand that may base-pair either with the m6A site or with the aforementioned guanosine 5′ to the m6A site was labeled on its N3 position (Liu et al., 2018). In this study, atom-specifically labeled RNAs provided direct and simple probes to unambiguously distinguish the different RNA conformations around the m6A site and establish the role of Mg2+ in stabilizing one of the conformations (Liu et al., 2018). Similarly, to characterize in detail the participation of an important adenosine (A32) in the mechanism of phosphodiester cleavage of pistol ribozymes, Patel and coworkers used a 47 nucleotide-long RNA construct with two atom-specifically labeled nucleotides (Ren et al., 2016). A 15N-labeled uracil on N1/N3 positions was used as a reporter probe that directly senses the proper folding and formation of the catalytically active form of the ribozyme. In addition, a 13C-labeled adenosine on C2 position was used to precisely determine the pKa value of the catalytic residue A32, which revealed a peculiar pKa value shifted by one pH unit compared with regular adenosines (Ren et al., 2016). With such extremely parsimonious atom-specific labeling, the tedious chemical shift assignment of the RNA is no longer required, as the identification of the unique signal becomes self-evident.
Use of isotopic labeling strategies of RNA to study HIV-1 replication
Over the last 30 years, HIV research efforts have greatly advanced the characterization of RNA structures and interactions using NMR. In this paragraph, we will sum up how advanced methods of isotopic labeling of RNAs have benefited the field of retroviruses with special focus on HIV-1 genomic RNA structure and dimerization and on the initiation of reverse transcription (RT).
The 5′-UTR is the most conserved part of the HIV-1 RNA genome. Its secondary structure is composed of the following domains: TAR stem-loop (SL), the poly(A) SL, the U5 domain, the tRNALys3 primer binding site (PBS) for RT followed by four SL (SL1 to SL4). SL1 contains the dimerization initiation site (DIS) of the viral RNA, SL2 contains the splice donor site (SD), SL3 is one of the main determinants of viral RNA encapsidation and, SL4 contains the gag AUG initiation codon of translation (reviewed in Sleiman et al. (2012). A 356-nucleotide HIV-1 5′-leader RNA that includes the entire 5′-UTR and the first 21 nucleotides of gag gene was prepared by means of enzymatic ligation of non-labeled 5′-RNA (nucleotides 1–327) and 13C-enriched AUG fragments (nucleotides 328–356) (Lu et al., 2011). The 5′-leader exists as an equilibrium of two conformers, one in which dimer-promoting residues and HIV-1 nucleocapsid (NC) binding sites are sequestered, packaging is attenuated and translation is promoted; and one in which these sites are exposed and packaging is promoted. The residues spanning the gag start codon (AUG) form a hairpin in the monomeric leader and base pair with residues of the U5 domain in the dimeric form. To directly probe for U5:AUG base pairing, Lu et al. developed an NMR approach, called long-range probing by adenosine interaction detection (lr-AID) that involves replacement of a short stretch of adjacent base pairs, here the U5:AUG base pairing, by A-U base pairs. The substituting element affords an upfield-shifted of one of the adenosine H2 NMR signal (~6.5 ppm), enabling direct detection of cross-strand adenosine-H2 and -H1′ NOEs. They could measure cross-strand NOEs demonstrating the existence of this interaction on a mutated 5′-leader RNA sample in which only ribose protons and H2 of adenosines are protonated within a deuterated background [u-2H, (H1′-H5′s, H2)-Ade]. These findings support a packaging mechanism in which translation, dimerization, NC binding and packaging are regulated by an RNA structural switch. Next, Keane et al. determined by NMR the structure of a minimal HIV-1 packaging element (Keane et al., 2015). Using a segmental based 2H-edited NMR approach, they solved a high-resolution structure of the 155-nt RNA, which is the largest RNA structure determined by NMR to date. The structure adopted an unexpected tandem three-way junction architecture, exposing guanosines essential for packaging and high-affinity binding to the NC. The AUG start codon and the SD are both sequestered through long-range pairing, providing a mechanism by which dimerization can suppress both translation and splicing. Despite these advances, important details were missing regarding the dimeric interface. Keane et al. (2016) adapted the lr-AID strategy to make it possible to distinguish intra- from intermolecular base pairing. In this approach, a 1:1 mixture is prepared, consisting of two leader RNA constructs containing distinct NMR labeling patterns. In one construct, all residues except adenosine are ‘NMR-invisible’ via deuteration, whereas in the second construct, all residues except guanosine are unobservable. 1H-1H NOEs between guanosine and adenosine residues in the dimeric leader are only observed if the two molecules form intermolecular base pairing. The dimeric interface is extensive and includes DIS:DIS base pairing in an extended duplex state as well as intermolecular pairing between elements of the U5 sequence and those near the AUG sequence. These studies illustrate the utility of deuteration for determining the structures and folding kinetics of large RNAs.
The second focus deals with the initiation of HIV-1 reverse transcription that requires the annealing of the human tRNALys3 to the PBS, the 18-nucleotides of the 3′-end of tRNALys3 being complementary to the PBS sequence. This annealing requires the action of the HIV-1 NC protein that acts as an RNA chaperone (Sleiman et al., 2012). The major structural RNA rearrangements that occur when tRNALys3 is annealed onto the viral genome was addressed by NMR (Tisne et al., 2004; Barraud et al., 2007; Puglisi and Puglisi, 2011; Sleiman et al., 2013; Coey et al., 2016). A novel approach to analyze dynamically such RNA refolding events using NMR of mixtures of 15N-labeled and unlabeled large RNA fragments (up to 50 kDa) in the presence of the NC was developed. As the annealing of tRNALys3 requires the opening of its structure, Tisne et al. designed an in vivo production of 15N-labeled tRNALys3 in E. coli in order to work with a tRNA that bears post-transcriptional modifications known to stabilize the tRNA structure (Tisne et al., 2000). Using this recombinant 15N-labeled tRNALys3, they observed the progressive formation of the HIV-1 RT initiation complex. In particular, they identified the nucleation sites where the viral RNA starts invading the tRNALys3 structure and characterized previously unexpected intermediates and differentially annealed states (Tisne et al., 2004; Barraud et al., 2007). The NC protein plays a key role by ‘unlocking’ stable tertiary interactions within the primer tRNA. In addition to the nucleotide complementarity to the PBS, a sequence of the viral RNA genome called PAS for ‘primer activation signal’ was proposed to interact with the T-arm of tRNALys3, this interaction stimulating the initiation of reverse transcription. Using 15N-labeling of tRNALys3, NMR provided the first direct observation at a single base-pair resolution of the PAS/tRNALys3 association (Sleiman et al., 2013; Coey et al., 2016).
Mass spectrometry (MS)
RNAs contain a vast variety of chemical modifications, which derive from the four canonical nucleosides adenosine, guanosine, uridine and cytidine. Their detection is possible by chemical means (Heiss and Kellner, 2017; Heiss et al., 2017), by sequencing and MS. Even with the ever-rising number of sequencing techniques, which detect modified nucleosides in whole transcriptomes, MS remains the key technique for characterization of modified nucleosides. RNA MS analytics can be subdivided into two principles. The first uses enzymes, which only partially digest the RNA into smaller oligonucleotides. Here, some of the sequence context surrounding a modified nucleoside remains and the technique is used to place modified nucleosides in known and unknown RNA sequences. The second relies on complete enzymatic digestion of the RNA into the nucleoside building block and is highly sensitive. This technique is commonly used for detection, quantification or discovery of modified nucleosides. Especially for quantification, stable isotope labeled compounds are necessary to overcome the limitations of MS. For quantification, the signal intensity of an analyte must correlate with its concentration or amount. In MS, the signal intensity depends of course on the amount of analyte, but in addition on a multitude of other parameters such as salt load, ionization properties of the analyte, instrument parameters and so on. These detection fluctuations make quantification by MS a challenging task, which can only be done by using stable isotope labeled internal standards (SILIS) of the analyte of interest. Knowledge on the absolute abundance of, e.g. modified nucleosides in an RNA of interest is crucial to the field and allows studies of RNA modification function and impact.
For a long time, no stable isotope labeled internal standards for oligonucleotide MS (oligo-MS) were available and the technique was limited to non-quantitative statements about the localization of modified nucleosides. The main challenge is that it is not possible to predict and synthesize all possible oligonucleotides as stable isotope labeled standards. Here, metabolic labeling has proven to be a useful tool. By culturing E. coli in 15N-media, the 16S rRNA was stable isotope labeled and successfully analyzed by oglio-MS (Waghmare and Dickman, 2011). Another elegant way to solve the problem was presented by the Limbach lab (Li and Limbach, 2012). They performed the enzymatic digest of the unknown RNA in the presence of H218O, which results in oxygen-18 incorporation into the oligonucleotide 3′-phosphate. The unknown, 18O-labeled RNA is mixed with a non-labeled digest of a known RNA. The comparative analysis of RNA digests (CARD) allows to rapidly screen total tRNAs from gene deletion mutants or comparatively sequence total tRNA from two related bacterial organisms. De-isotoping, the metabolic removal of natural 13C and 15N isotopes, reduces the isotope pattern and thus aides the MS spectra interpretation (Wetzel et al., 2014). The technique was further improved in 2017, when they used RNA as a reference that was in vitro transcribed in the presence of 13C-15N-labeled GTP (Paulines and Limbach, 2017). With this SIL-CARD approach, quantitative assessment of tRNA abundances is possible. Similarly, SILNAS allowed the determination and quantification of all modified nucleosides in Schizosaccharomyces pombe ribosomal (r)RNAs and generated the first complete modification maps of eukaryotic rRNAs at single-nucleotide resolution (SILNAS, Stable Isotope-Labeled riboNucleic Acid as an internal Standard) (Taoka et al., 2015). Quantitative analysis of dsRNA is also possible by using metabolically prepared RNA form E. coli (Kung et al., 2018).
With the development of thermospray and electrospray ionization, the analysis of nucleosides became possible by liquid chromatography coupled-mass spectrometry (LC-MS) (Pomerantz and McCloskey, 1990). After enzymatic digestion, the nucleoside mixture is directly separated by LC followed by sensitive detection in the mass spectrometer.
Complete metabolic labeling of, e.g. all carbons or nitrogens, has become a key tool for the identification of novel RNA and DNA modifications. A uniform labeling of, e.g. all carbon atoms with 13C or all nitrogen atoms with 15N helps in sum formula generation and structure prediction of novel compounds. With this technique, new modified nucleoside candidates were revealed (Kellner et al., 2014a,b), geranylated nucleosides (Dumelin et al., 2012) and msms2i6A (Dal Magro et al., 2018) were found in bacterial tRNA. By combination with genetic tools, a deazaguanosine derivative was identified and quantified in bacterial DNA using metabolic labeling (Thiaville et al., 2016).
For those novel and most known modified nucleosides, no stable isotope labeled internal standards are available for quantification. Thus, the absolute abundance of modified nucleosides within their RNA is very difficult to assess. The first analysis of absolute quantities of modified nucleosides was made by the synthetic effort of the Carrel lab. Here, six stable isotope labeled adenosine modifications were produced and used for quantification (Bruckl et al., 2009). The effort was later repeated by synthesis of another six modified nucleosides and their stable isotope labeled standards (Brandmayr et al., 2012).
To date, over 170 modified nucleosides are known in all domains of life (Boccaletto et al., 2018). The chemical synthesis of stable isotope labeled derivatives of these is a major challenge. A solution is metabolic labeling, where all modified nucleosides of an organism are produced by feeding isotopically labeled nutrients to a microorganism (Kellner et al., 2014a,b). With these, absolute quantification of many modified nucleosides becomes easily accessible. A similar approach was presented for GC coupled MS (Miranda-Santos et al., 2015).
In DNA, methylation was found to be dynamic. The addition or removal of a methylation on carbon C5 of cytosine can switch genes off or on. As this chemical code is additional information on top of the sequence, it is referred to as the epigenetic code.
While epigenetics is an intensively studied area, the analogue process in RNA, termed epitranscriptomics, is far less studied. This is mainly caused by our limited number of tools to study the dynamics of RNA modifications and, in addition, the complex process of finding biological consequences to RNA modifications. The key difference in the studies of dynamic processes is that DNA is a storage molecule while RNA is transient, and modifications may be removed by RNA degradation and transcription of new RNA molecules. In contrast, modifications of DNA must be removed by an enzymatic/chemical process, which leaves the DNA sequence untouched and the DNA intact. Otherwise, mutations would occur and harm the organism. While it should be possible for the cell to use enzymatic/chemical removal mechanisms of DNA in RNA, RNA has a second option for removal of an unwanted modification – the RNA itself is degraded and a new RNA is transcribed. This potential competition between these two processes makes it difficult to study the dynamics of RNA modifications. Although we have seen many studies in the last few years, which claim that mRNA is enzymatically demodified, no solid in vivo proof has yet been presented. Currently, quantitative MS of RNA modification profiles was used to claim an active demodification in vivo (Meyer and Jaffrey, 2014). However, the absolute number of a modification within an RNA does not reflect the origin of the modification. For example, the decrease in modification density can be explained by enzymatic demodification processes but also by increased degradation of modified RNA or even by increased transcription of the RNA, without it being modified. Vice versa, an increase in modification density can be explained by additional modification events in the original RNA or by increased degradation of non-modified RNAs.
We have presented a technique, which overcomes these current limitations by utilizing metabolic stable isotope labeling. The technique is termed NAIL-MS, nucleic acid isotope labeling coupled MS.
The principle of NAIL-MS for analysis of dynamic (de-)modification processes relies on a pulse-chase type experimental set-up. The cells are grown in a particularly labeled growth media and upon exposure to a pulse, e.g. RNA damaging agent, the media is exchanged to a differently labeled media. With this principle, it is possible to study the fate of the RNA present during the pulse and chase their behavior (Figure 4). It is also possible to study the kinetics and behavior of the newly transcribed RNAs and observe their modification status over time. For further details, we have recently summarized the pre-requisites, labeling techniques and necessary validation experiments of NAIL-MS studies in a recent methods paper (Reichle et al., 2018a,b).
Our first steps to overcome the limitation of static modification analysis was in 2017, when we utilized stable isotope labeling of DNA to observe damage and repair of a DNA modification in bacteria. The combination of different labeling media allowed the creation of a pulse-chase experiment, which was used to observe repair of phosphorothioates in bacterial DNA (Kellner et al., 2017). In the same year, we adapted the approach to RNA modification analysis in yeast (Heiss et al., 2017). We followed the modification density of tRNA in dependence of the growth phase and we identified the underlying mechanisms for several modified nucleosides (Figure 4A).
Recently, we could clearly observe repair of tRNA damage methylation in vivo using NAIL-MS (Reichle et al., 2018a),b). In this study, we focused on the damage introduced by the methylating agent methyl-methanesulfonate (MMS) and the subsequent fate of the damaged tRNA in E. coli. During exposure to MMS, adenosine is methylated at position 1 and 6, guanosine at position 7 and cytidine at position 3 (Figure 4B). Using NAIL-MS we could further distinguish 7-methylguanosine from enzymatic formation from 7-methylguanosine formed by direct methylation though MMS. The exposure was carried out in non-labeled growth medium and the non-labeled tRNAs were damaged by MMS. In the next step, we replaced the media with a heavy labeled variant. Newly transcribed tRNAs received a heavy isotope labeling, which we could clearly distinguish from the damaged, non-labeled tRNAs by MS. We analyzed the abundance of the damaged nucleosides in the damaged tRNAs and we indeed found a slow decrease of 1-methyladenosine and 3-methylcytidine, and even potentially 6-methyladenosine. Compared to non-stressed cells, we did not see evidence for increased tRNA degradation. In combination, our data clearly confirmed demethylation as a tRNA repair mechanism and excluded tRNA degradation (Reichle et al., 2018a,b).
With NAIL-MS we can assess RNA turnover, RNA biosynthesis and dilution effects and determine the modification content of RNAs in response to growth or stress.
Summary and conclusion
Stable isotope labeling of nucleic acids and especially RNA has opened the door to new analytical possibilities, which simplify and push our understanding of RNA structure and modification. We have presented a vast overview of how stable isotope labeled RNA can be produced in vivo, in vitro and by chemical synthesis. Metabolic labeling of full-length RNA has the advantage of correct post-transcriptional modification and folding. Longer RNAs of increasing size and complexity are best produced by in vitro transcription using nucleotide-specific or even atom-specific labeling techniques. Thus, the chemical shift overlap of resonances is reduced and the problem of line broadening is less pronounced. Considering the dynamic nature of RNAs, NMR spectroscopy can sometimes achieve better than crystallography in the determination of RNA structures.
In RNAs, massive NMR chemical shift overlaps, especially of the ribose signals, make spectra interpretation difficult to impossible. For that reason, deuteration has been found to be ideal as it simplifies the spectra and in addition improves the signal intensity of the remaining protons. Uniform or atom-specific deuteration became therefore relatively popular in the structural studies of large RNAs. The techniques of segmental isotope labeling further expand the applicability of RNA analysis by NMR as now key structures of a short RNA stretch within its full-length RNA context become accessible. The ultimate strategy for simplifying NMR spectra with isotope labeling consists of the atom-specific labeling. With this technique, the dynamics properties of RNA such as folding or structural re-arrangements can be monitored by NMR (Wolter et al., 2017). 6-Methyladenosine (m6A), the first dynamic RNA modification found in mRNA, is currently in the focus of research and the field of epitranscriptomics has been coined by this modification. To understand its function and impact within an RNA, NMR spectroscopy of isotopically labeled RNA was used to investigate the effect of m6A on the RNA conformation (Liu et al., 2018). Thus, NMR spectroscopy is at the very heart of current epitranscriptome research.
So is MS. Here, the quantification of modifications within an RNA of interest is facilitated by usage of stable isotope labeled internal standards (SILIS). Without SILIS, no absolute quantification would be possible and a comparison of sample to sample would be impossible. Quantification of modified nucleosides is possible within its sequence context by Oligo-MS (Taoka et al., 2015; Paulines and Limbach, 2017). The production of SILIS depends on metabolic isotope labeling or in vitro transcription. With these techniques, the abundance of modified nucleosides at certain positions of rRNA or tRNA can be determined and compared to, e.g. different species or genetic mutants. In the field of nucleoside analytics, the production of SILIS nucleosides is possible by chemical synthesis, which is fairly time-consuming (Brandmayr et al., 2012), or by metabolic labeling (Kellner et al., 2014a,b). With these standards, the quantities of modified nucleosides can be assessed and whole RNA modification profiles can be determined. The key benefit of nucleoside MS over oligo-MS is its superb sensitivity, where all modified nucleosides are quantifiable in the double-digit femtomol range and some even in the attomol range. Thus, it is possible to assess the modification profile of very low amounts of RNA sample.
The metabolic isotope labeling for SILIS production has initiated the development of dynamic RNA modification analysis by NAIL-MS. Current quantification reports the abundance of a modification at a certain moment, and by analysis of consecutive time points, changes in the abundance can be assessed. In these studies, it remains unclear how these changes were introduced into the RNA. For example; a drop in a modification’s abundance can be caused by degradation of modified RNA, but also by active demethylation. Vice versa, an increase in modification density might be caused by increased modification of existing transcripts, new transcripts or both. In addition, degradation of unmodified RNA might also be causative for increased abundance of a modification. With the goal to unravel these competing processes, we combined different metabolic RNA labeling techniques to create the option of pulse-chase like experiments (NAIL-MS). The differential labeling allows discrimination of, e.g. stress-exposed RNA from new transcripts and the behavior of modified nucleosides within the stress-exposed RNA can be analyzed. With NAIL-MS, it is now possible to analyze the dynamic processes of post-transcriptional RNA modification and demodification. In addition, demodification processes can be finally distinguished from the competing processes of RNA degradation and turnover in vivo.
In conclusion, we present a vast overview of current RNA isotope labeling techniques and their current applications in RNA analysis by both NMR and MS. The trend is without doubt shifting from the analysis of snapshot moments towards the development and application of tools capable of analyzing the dynamics of RNA structure and modification profiles. These developments are rendered possible by stable isotope labeling and the vast toolbox for generation of such RNA.
We thank Matthias Heiß for providing graphical support in figure generation.
Alvarado, L.J., LeBlanc, R.M., Longhini, A.P., Keane, S.C., Jain, N., Yildiz, Z.F., Tolbert, B.S., D’Souza, V.M., Summers, M.F., Kreutz, C., et al. (2014a). Regio-selective chemical-enzymatic synthesis of pyrimidine nucleotides facilitates RNA structure and dynamics studies. ChemBioChem 15, 1573–1577. CrossrefGoogle Scholar
Alvarado, L.J., Longhini, A.P., LeBlanc, R.M., Chen, B., Kreutz, C., and Dayie, T.K. (2014b). Chemo-enzymatic synthesis of selectively 13C/15N-labeled RNA for NMR structural and dynamics studies. Methods Enzymol. 549, 133–162. CrossrefGoogle Scholar
Aoyagi, T., Hanada, S., Itoh, H., Sato, Y., Ogata, A., Friedrich, M.W., Kikuchi, Y., and Hori, T. (2015). Ultra-high-sensitivity stable-isotope probing of rRNA by high-throughput sequencing of isopycnic centrifugation gradients. Environ. Microbiol. Rep. 7, 282–287. CrossrefPubMedGoogle Scholar
Barnwal, R.P., Loh, E., Godin, K.S., Yip, J., Lavender, H., Tang, C.M., and Varani, G. (2016). Structure and mechanism of a molecular rheostat, an RNA thermometer that modulates immune evasion by Neisseria meningitidis. Nucleic Acids Res. 44, 9426–9437. PubMedGoogle Scholar
Batey, R.T., Inada, M., Kujawinski, E., Puglisi, J.D., and Williamson, J.R. (1992). Preparation of isotopically labeled ribonucleotides for multidimensional NMR spectroscopy of RNA. Nucleic Acids Res. 20, 4515–4523. PubMedCrossrefGoogle Scholar
Beusch, I., Barraud, P., Moursy, A., Clery, A., and Allain, F.H. (2017). Tandem hnRNP A1 RNA recognition motifs act in concert to repress the splicing of survival motor neuron exon 7. Elife.6, 1–35. Google Scholar
Boccaletto, P., Machnicka, M.A., Purta, E., Piatkowski, P., Baginski, B., Wirecki, T.K., de Crecy-Lagard, V., Ross, R., Limbach, P.A., Kotter, A., et al. (2018). MODOMICS: a database of RNA modification pathways. 2017 update. Nucleic Acids Res. 46, D303–D307. CrossrefGoogle Scholar
Bonneau, E., Girard, N., Lemieux, S., and Legault, P. (2015). The NMR structure of the II–III–VI three-way junction from the Neurospora VS ribozyme reveals a critical tertiary interaction and provides new insights into the global ribozyme structure. RNA 21, 1621–1632. PubMedCrossrefGoogle Scholar
Boschker, H.T.S., Nold, S.C., Wellsbury, P., Bos, D., de Graaf, W., Pel, R., Parkes, R.J., and Cappenberg, T.E. (1998). Direct linking of microbial populations to specific biogeochemical processes by 13C-labelling of biomarkers. Nature 392, 801. CrossrefGoogle Scholar
Bourbigot, S., Dock-Bregeon, A.C., Eberling, P., Coutant, J., Kieffer, B., and Lebars, I. (2016). Solution structure of the 5′-terminal hairpin of the 7SK small nuclear RNA. RNA 22, 1844–1858. CrossrefPubMedGoogle Scholar
Brandmayr, C., Wagner, M., Bruckl, T., Globisch, D., Pearson, D., Kneuttinger, A.C., Reiter, V., Hienzsch, A., Koch, S., Thoma, I., et al. (2012). Isotope-based analysis of modified tRNA nucleosides correlates modification density with translational efficiency. Angew. Chem. Int. Ed. 51, 11162–11165. CrossrefGoogle Scholar
Bruckl, T., Globisch, D., Wagner, M., Muller, M., and Carell, T. (2009). Parallel isotope-based quantification of modified tRNA nucleosides. Angew. Chem. Int. Ed. 48, 7932–7934. CrossrefGoogle Scholar
Bullock, S.L., Ringel, I., Ish-Horowicz, D., and Lukavsky, P.J. (2010). A′-form RNA helices are required for cytoplasmic mRNA transport in Drosophila. Nat. Struct. Mol. Biol. 17, 703–709. CrossrefGoogle Scholar
Buttner, L., Seikowski, J., Wawrzyniak, K., Ochmann, A., and Hobartner, C. (2013). Synthesis of spin-labeled riboswitch RNAs using convertible nucleosides and DNA-catalyzed RNA ligation. Bioorg. Med. Chem. 21, 6171–6180. CrossrefPubMedGoogle Scholar
Coey, A., Larsen, K., Puglisi, J.D., and Viani Puglisi, E. (2016). Heterogeneous structures formed by conserved RNA sequences within the HIV reverse transcription initiation site. RNA 22, 1689–1698. PubMedCrossrefGoogle Scholar
D’Souza, V., Dey, A., Habib, D., and Summers, M.F. (2004). NMR structure of the 101-nucleotide core encapsidation signal of the Moloney murine leukemia virus. J. Mol. Biol. 337, 427–442. CrossrefPubMedGoogle Scholar
Dal Magro, C., Keller, P., Kotter, A., Werner, S., Duarte, V., Marchand, V., Ignarski, M., Freiwald, A., Muller, R.U., Dieterich, C., et al. (2018). A vastly increased chemical variety of RNA modifications containing a thioacetal structure. Angew. Chem. Int. Ed. 57, 7893–7897. CrossrefGoogle Scholar
Dallmann, A., Beribisky, A.V., Gnerlich, F., Rubbelke, M., Schiesser, S., Carell, T., and Sattler, M. (2016). Site-specific isotope-labeling of inosine phosphoramidites and NMR analysis of an inosine-containing RNA duplex. Chemistry 22, 15350–15359. CrossrefPubMedGoogle Scholar
Dorn, G., Leitner, A., Boudet, J., Campagne, S., von Schroetter, C., Moursy, A., Aebersold, R., and Allain, F.H. (2017). Structural modeling of protein-RNA complexes using crosslinking of segmentally isotope-labeled RNA and MS/MS. Nat. Methods 14, 487–490. PubMedCrossrefGoogle Scholar
Dumelin, C.E., Chen, Y., Leconte, A.M., Chen, Y.G., and Liu, D.R. (2012). Discovery and biological characterization of geranylated RNA in bacteria. Nat. Chem. Biol. 8, 913–919. CrossrefPubMedGoogle Scholar
Duss, O., Maris, C., von Schroetter, C., and Allain, F.H. (2010). A fast, efficient and sequence-independent method for flexible multiple segmental isotope labeling of RNA using ribozyme and RNase H cleavage. Nucleic Acids Res. 38, e188. PubMedCrossrefGoogle Scholar
Duss, O., Michel, E., Diarra dit Konte, N., Schubert, M., and Allain, F.H. (2014a). Molecular basis for the wide range of affinity found in Csr/Rsm protein-RNA recognition. Nucleic Acids Res. 42, 5332–5346. CrossrefGoogle Scholar
Duss, O., Michel, E., Yulikov, M., Schubert, M., Jeschke, G., and Allain, F.H. (2014b). Structural basis of the non-coding RNA RsmZ acting as a protein sponge. Nature 509, 588–592. CrossrefGoogle Scholar
Duss, O., Yulikov, M., Jeschke, G., and Allain, F.H. (2014c). EPR-aided approach for solution structure determination of large RNAs or protein-RNA complexes. Nat. Commun. 5, 3669. CrossrefGoogle Scholar
Duss, O., Diarra Dit Konte, N., and Allain, F.H. (2015). Cut and paste RNA for nuclear magnetic resonance, paramagnetic resonance enhancement, and electron paramagnetic resonance structural studies. Methods Enzymol. 565, 537–562. CrossrefPubMedGoogle Scholar
Globisch, D., Munzel, M., Muller, M., Michalakis, S., Wagner, M., Koch, S., Bruckl, T., Biel, M., et al. (2010). Tissue distribution of 5-hydroxymethylcytosine and search for active demethylation intermediates. PLoS One 5, e15367. CrossrefPubMedGoogle Scholar
Hamasaki, T., Matsumoto, T., Sakamoto, N., Shimahara, A., Kato, S., Yoshitake, A., Utsunomiya, A., Yurimoto, H., Gabazza, E.C., and Ohgi, T. (2013). Synthesis of 18O-labeled RNA for application to kinetic studies and imaging. Nucleic Acids Res. 41, e126. CrossrefGoogle Scholar
Hatamoto, M., Imachi, H., Yashiro, Y., Ohashi, A., and Harada, H. (2008). Detection of active butyrate-degrading microorganisms in methanogenic sludges by RNA-based stable isotope probing. Appl. Environ. Microbiol. 74, 3610–3614. CrossrefPubMedGoogle Scholar
Heiss, M., Reichle, V.F., and Kellner, S. (2017). Observing the fate of tRNA and its modifications by nucleic acid isotope labeling mass spectrometry: NAIL-MS. RNA Biol. 14, 1260–1268. PubMedCrossrefGoogle Scholar
Herrmann, E., Young, W., Rosendale, D., Reichert-Grimm, V., Riedel, C.U., Conrad, R., and Egert, M. (2017). RNA-based stable isotope probing suggests Allobaculum spp. as particularly active glucose assimilators in a complex murine microbiota cultured in vitro. Biomed Res. Int. 2017, 1829685. Google Scholar
Huang, X., Yu, P., LeProust, E., and Gao, X. (1997). An efficient and economic site-specific deuteration strategy for NMR studies of homologous oligonucleotide repeat sequences. Nucleic Acids Res. 25, 4758–4763. PubMedCrossrefGoogle Scholar
Imai, S., Kumar, P., Hellen, C.U., D’Souza, V.M., and Wagner, G. (2016). An accurately preorganized IRES RNA structure enables eIF4G capture for initiation of viral translation. Nat. Struct. Mol. Biol. 23, 859–864. PubMedCrossrefGoogle Scholar
Iwan, K., Rahimoff, R., Kirchner, A., Spada, F., Schroder, A.S., Kosmatchev, O., Ferizaj, S., Steinbacher, J., Parsa, E., Muller, M., et al. (2018). 5-Formylcytosine to cytosine conversion by C–C bond cleavage in vivo. Nat. Chem. Biol. 14, 72–78. CrossrefGoogle Scholar
Johnson Jr., J.E. and Hoogstraten, C.G. (2008). Extensive backbone dynamics in the GCAA RNA tetraloop analyzed using 13C NMR spin relaxation and specific isotope labeling. J. Am. Chem. Soc. 130, 16757–16769. CrossrefPubMedGoogle Scholar
Johnson Jr., J.E., Julien, K.R., and Hoogstraten, C.G. (2006). Alternate-site isotopic labeling of ribonucleotides for NMR studies of ribose conformational dynamics in RNA. J. Biomol. NMR 35, 261–274. PubMedCrossrefGoogle Scholar
Kang, M., Eichhorn, C.D., and Feigon, J. (2014). Structural determinants for ligand capture by a class II preQ1 riboswitch. Proc. Natl. Acad Sci. USA 111, E663–E671. Google Scholar
Keane, S.C., Heng, X., Lu, K., Kharytonchyk, S., Ramakrishnan, V., Carter, G., Barton, S., Hosic, A., Florwick, A., Santos, J., et al. (2015). RNA structure. Structure of the HIV-1 RNA packaging signal. Science 348, 917–921. Google Scholar
Keane, S.C., Van, V., Frank, H.M., Sciandra, C.A., McCowin, S., Santos, J., Heng, X., and Summers, M.F. (2016). NMR detection of intermolecular interaction sites in the dimeric 5′-leader of the HIV-1 genome. Proc. Natl. Acad. Sci. USA 113, 13033–13038. CrossrefGoogle Scholar
Kellner, S., Neumann, J., Rosenkranz, D., Lebedeva, S., Ketting, R.F., Zischler, H., Schneider, D., and Helm, M. (2014a). Profiling of RNA modifications by multiplexed stable isotope labelling. Chem. Commun. (Cambr.) 50, 3516–3518. CrossrefGoogle Scholar
Kellner, S., Ochel, A., Thuring, K., Spenkuch, F., Neumann, J., Sharma, S., Entian, K.D., Schneider, D., and Helm, M. (2014b). Absolute and relative quantification of RNA modifications via biosynthetic isotopomers. Nucleic Acids Res. 42, e142. CrossrefGoogle Scholar
Kellner, S., DeMott, M.S., Cheng, C.P., Russell, B.S., Cao, B., You, D., and Dedon, P.C. (2017). Oxidation of phosphorothioate DNA modifications leads to lethal genomic instability. Nat. Chem. Biol. 13, 888–894. CrossrefPubMedGoogle Scholar
Kittelmann, S. and Friedrich, M.W. (2008). Identification of novel perchloroethene-respiring microorganisms in anoxic river sediment by RNA-based stable isotope probing. Environ. Microbiol. 10, 31–46. PubMedGoogle Scholar
Kremser, J., Strebitzer, E., Plangger, R., Juen, M.A., Nussbaumer, F., Glasner, H., Breuker, K., and Kreutz, C. (2017). Chemical synthesis and NMR spectroscopy of long stable isotope labelled RNA. Chem. Commun. (Camb). 53, 12938–12941. CrossrefPubMedGoogle Scholar
Kung, A.W., Kilby, P.M., Portwood, D.E., and Dickman, M.J. (2018). Quantification of dsRNA using stable isotope labeling dilution liquid chromatography/mass spectrometry. Rapid Commun. Mass Spectrom. 32, 590–596. CrossrefPubMedGoogle Scholar
Le, M.T., Brown, R.E., Simon, A.E., and Dayie, T.K. (2015). In vivo, large-scale preparation of uniformly 15N- and site-specifically 13C-labeled homogeneous, recombinant RNA for NMR studies. Methods Enzymol. 565, 495–535. CrossrefGoogle Scholar
Lebars, I., Vileno, B., Bourbigot, S., Turek, P., Wolff, P., and Kieffer, B. (2014). A fully enzymatic method for site-directed spin labeling of long RNA. Nucleic Acids Res. 42, e117. PubMedCrossrefGoogle Scholar
LeBlanc, R.M., Longhini, A.P., Le Grice, S.F.J., Johnson, B.A., and Dayie, T.K. (2017). Combining asymmetric 13C-labeling and isotopic filter/edit NOESY: a novel strategy for rapid and logical RNA resonance assignment. Nucleic Acids Res. 45, e146. PubMedCrossrefGoogle Scholar
Liu, Y., Stepanov, V.G., Strych, U., Willson, R.C., Jackson, G.W., and Fox, G.E. (2010). DNAzyme-mediated recovery of small recombinant RNAs from a 5S rRNA-derived chimera expressed in Escherichia coli. BMC Biotechnol. 10, 85. CrossrefPubMedGoogle Scholar
Liu, Y., Holmstrom, E., Zhang, J., Yu, P., Wang, J., Dyba, M.A., Chen, D., Ying, J., Lockett, S., Nesbitt, D.J., et al. (2015). Synthesis and applications of RNAs with position-selective labelling and mosaic composition. Nature 522, 368–372. PubMedCrossrefGoogle Scholar
Liu, B., Merriman, D.K., Choi, S.H., Schumacher, M.A., Plangger, R., Kreutz, C., Horner, S.M., Meyer, K.D., and Al-Hashimi, H.M. (2018). A potentially abundant junctional RNA motif stabilized by m6A and Mg2+. Nat. Commun. 9, 2761. PubMedCrossrefGoogle Scholar
Longhini, A.P., LeBlanc, R.M., Becette, O., Salguero, C., Wunderlich, C.H., Johnson, B.A., D’Souza, V.M., Kreutz, C., and Dayie, T.K. (2016a). Chemo-enzymatic synthesis of site-specific isotopically labeled nucleotides for use in NMR resonance assignment, dynamics and structural characterizations. Nucleic Acids Res. 44, e52. CrossrefGoogle Scholar
Lu, K., Heng, X., Garyu, L., Monti, S., Garcia, E.L., Kharytonchyk, S., Dorjsuren, B., Kulandaivel, G., Jones, S., Hiremath, A., et al. (2011). NMR detection of structures in the HIV-1 5′-leader RNA that regulate genome packaging. Science 334, 242–245. PubMedCrossrefGoogle Scholar
Lueders, T., Wagner, B., Claus, P., and Friedrich, M.W. (2004). Stable isotope probing of rRNA and DNA reveals a dynamic methylotroph community and trophic interactions with fungi and protozoa in oxic rice field soil. Environ. Microbiol. 6, 60–72. PubMedGoogle Scholar
Miranda-Santos, I., Gramacho, S., Pineiro, M., Martinez-Gomez, K., Fritz, M., Hollemeyer, K., Salvador, A., and Heinzle, E. (2015). Mass isotopomer analysis of nucleosides isolated from RNA and DNA using GC/MS. Anal. Chem. 87, 617–623. CrossrefPubMedGoogle Scholar
Miyazaki, Y., Irobalieva, R.N., Tolbert, B.S., Smalls-Mantey, A., Iyalla, K., Loeliger, K., D’Souza, V., Khant, H., Schmid, M.F., Garcia, E.L., et al. (2010). Structure of a conserved retroviral RNA packaging element by NMR spectroscopy and cryo-electron tomography. J. Mol. Biol. 404, 751–772. CrossrefPubMedGoogle Scholar
Moore, P.B., Abo, S., Freeborn, B., Gewirth, D.T., Leontis, N.B., and Sun, G. (1988). Preparation of 5S RNA-related materials for nuclear magnetic resonance and crystallography studies. Methods Enzymol. 164, 158–174. CrossrefPubMedGoogle Scholar
Nelissen, F.H., van Gammeren, A.J., Tessari, M., Girard, F.C., Heus, H.A., and Wijmenga, S.S. (2008). Multiple segmental and selective isotope labeling of large RNA for NMR structural studies. Nucleic Acids Res. 36, e89. PubMedCrossrefGoogle Scholar
Nelissen, F.H., Leunissen, E.H., van de Laar, L., Tessari, M., Heus, H.A., and Wijmenga, S.S. (2012). Fast production of homogeneous recombinant RNA – towards large-scale production of RNA. Nucleic Acids Res. 40, e102. CrossrefPubMedGoogle Scholar
Nelissen, F.H., Heus, H.A., and Wijmenga, S.S. (2015). Production of homogeneous recombinant RNA using a tRNA scaffold and hammerhead ribozymes. Methods Mol. Biol. 1316, 33–44. CrossrefPubMedGoogle Scholar
Nikonowicz, E.P., Sirr, A., Legault, P., Jucker, F.M., Baer, L.M., and Pardi, A. (1992). Preparation of 13C and 15N labelled RNAs for heteronuclear multi-dimensional NMR studies. Nucleic Acids Res. 20, 4507–4513. CrossrefPubMedGoogle Scholar
Paulines, M.J. and Limbach, P.A. (2017). Stable isotope labeling for improved comparative analysis of RNA digests by mass spectrometry. J. Am. Soc. Mass Spectrom. 28, 551–561. PubMedCrossrefGoogle Scholar
Pitulle, C., Hedenstierna, K.O., and Fox, G.E. (1995). A novel approach for monitoring genetically engineered microorganisms by using artificial, stable RNAs. Appl. Environ. Microbiol. 61, 3661–3666. PubMedGoogle Scholar
Ponchon, L. Beauvais, G., Nonin-Lecomte, S., and Dardel, F. (2009). A generic protocol for the expression and purification of recombinant RNA in Escherichia coli using a tRNA scaffold. Nat. Protoc. 4, 947–959. CrossrefPubMedGoogle Scholar
Ponchon, L., Catala, M., Seijo, B., El Khouri, M., Dardel, F., Nonin-Lecomte, S., and Tisne, C. (2013). Co-expression of RNA-protein complexes in Escherichia coli and applications to RNA biology. Nucleic Acids Res. 41, e150. PubMedCrossrefGoogle Scholar
Popova, A.M. and Williamson, J.R. (2014). Quantitative analysis of rRNA modifications using stable isotope labeling and mass spectrometry. J. Am. Chem. Soc. 136, 2058–2069. PubMedCrossrefGoogle Scholar
Reichle, V.F., Kaiser, S., Heiss, M., Hagelskamp, F., Borland, K., and Kellner, S. (2018a). Surpassing limits of static RNA modification analysis with dynamic NAIL-MS. Methods 156, 91–101. Google Scholar
Reichle, V.F., Weber, V., and Kellner, S. (2018b). NAIL-MS in E. coli determines the source and fate of methylation in tRNA. ChemBioChem 19, 2575–2583. Google Scholar
Ren, A., Vusurovic, N., Gebetsberger, J., Gao, P., Juen, M., Kreutz, C., Micura, R., and Patel, D.J. (2016). Pistol ribozyme adopts a pseudoknot fold facilitating site-specific in-line cleavage. Nat. Chem. Biol. 12, 702–708. PubMedCrossrefGoogle Scholar
Sarin, L.P., Kienast, S.D., Leufken, J., Ross, R.L., Dziergowska, A., Debiec, K., Sochacka, E., Limbach, P.A., Fufezan, C., Drexler, H.C.A., et al. (2018). Nano LC-MS using capillary columns enables accurate quantification of modified ribonucleosides at low femtomol levels. RNA 24, 1403–1417. PubMedCrossrefGoogle Scholar
Sleiman, D., Goldschmidt, V., Barraud, P., Marquet, R., Paillart, J.C., and Tisne, C. (2012). Initiation of HIV-1 reverse transcription and functional role of nucleocapsid-mediated tRNA/viral genome interactions. Virus Res. 169, 324–339. PubMedCrossrefGoogle Scholar
Sleiman, D., Barraud, P., Brachet, F., and Tisne, C. (2013). The Interaction between tRNA(Lys) 3 and the primer activation signal deciphered by NMR spectroscopy. PLoS One 8, e64700. CrossrefPubMedGoogle Scholar
Sochor, F., Silvers, R., Muller, D., Richter, C., Furtig, B., and Schwalbe, H. (2016). 19F-labeling of the adenine H2-site to study large RNAs by NMR spectroscopy. J. Biomol. NMR 64, 63–74. CrossrefGoogle Scholar
Strebitzer, E., Nussbaumer, F., Kremser, J., Tollinger, M., and Kreutz, C. (2018). Studying sparsely populated conformational states in RNA combining chemical synthesis and solution NMR spectroscopy. Methods 148, 39–47. PubMedCrossrefGoogle Scholar
Taoka, M., Nobe, Y., Hori, M., Takeuchi, A., Masaki, S., Yamauchi, Y., Nakayama, H., Takahashi, N., and Isobe, T. (2015). A mass spectrometry-based method for comprehensive quantitative determination of post-transcriptional RNA modifications: the complete chemical structure of Schizosaccharomyces pombe ribosomal RNAs. Nucleic Acids Res. 43, e115. CrossrefPubMedGoogle Scholar
Thakur, C.S., Sama, J.N., Jackson, M.E., Chen, B., and Dayie, T.K. (2010). Selective 13C labeling of nucleotides for large RNA NMR spectroscopy using an E. coli strain disabled in the TCA cycle. J. Biomol. NMR 48, 179–192. CrossrefGoogle Scholar
Thakur, C.S., Luo, Y., Chen, B., Eldho, N.V., and Dayie, T.K. (2012). Biomass production of site selective 13C/15N nucleotides using wild type and a transketolase E. coli mutant for labeling RNA for high resolution NMR. J. Biomol. NMR 52, 103–114. CrossrefGoogle Scholar
Thiaville, J.J., Kellner, S.M., Yuan, Y., Hutinet, G., Thiaville, P.C., Jumpathong, W., Mohapatra, S., Brochier-Armanet, C., Letarov, A.V., Hillebrand, R., et al. (2016). Novel genomic island modifies DNA with 7-deazaguanine derivatives. Proc. Natl. Acad. Sci. USA 113, E1452–E1459. Google Scholar
Tisne, C., Rigourd, M., Marquet, R., Ehresmann, C., and Dardel, F. (2000). NMR and biochemical characterization of recombinant human tRNA(Lys)3 expressed in Escherichia coli: identification of posttranscriptional nucleotide modifications required for efficient initiation of HIV-1 reverse transcription. RNA 6, 1403–1412. CrossrefPubMedGoogle Scholar
Tzakos, A.G., Easton, L.E., and Lukavsky, P.J. (2006). Complementary segmental labeling of large RNAs: economic preparation and simplified NMR spectra for measurement of more RDCs. J. Am. Chem. Soc. 128, 13344–13345. CrossrefPubMedGoogle Scholar
Tzakos, A.G., Easton, L.E., and Lukavsky, P.J. (2007). Preparation of large RNA oligonucleotides with complementary isotope-labeled segments for NMR structural studies. Nat. Protoc. 2, 2139–2147. CrossrefPubMedGoogle Scholar
van Delft, P., Akay, A., Huber, S.M., Bueschl, C., Rudolph, K.L.M., Di Domenico, T., Schuhmacher, R., Miska, E.A., and Balasubramanian, S. (2017). The profile and dynamics of RNA modifications in animals. ChemBioChem 18, 979–984. CrossrefPubMedGoogle Scholar
Waghmare, S.P. and Dickman, M.J. (2011). Characterization and quantification of RNA post-transcriptional modifications using stable isotope labeling of RNA in conjunction with mass spectrometry analysis. Anal. Chem. 83, 4894–4901. CrossrefPubMedGoogle Scholar
Wallis, N.G., Dardel, F., and Blanquet, S. (1995). Heteronuclear NMR studies of the interactions of 15N-labeled methionine-specific transfer RNAs with methionyl-tRNA transformylase. Biochemistry 34, 7668–7677. CrossrefPubMedGoogle Scholar
Wenter, P., Reymond, L., Auweter, S.D., Allain, F.H., and Pitsch, S. (2006). Short, synthetic and selectively 13C-labeled RNA sequences for the NMR structure determination of protein-RNA complexes. Nucleic Acids Res. 34, e79. PubMedCrossrefGoogle Scholar
Wolter, A.C., Duchardt-Ferner, E., Nasiri, A.H., Hantke, K., Wunderlich, C.H., Kreutz, C., and Wohnert, J. (2016). NMR resonance assignments for the class II GTP binding RNA aptamer in complex with GTP. Biomol. NMR Assign. 10, 101–105. PubMedCrossrefGoogle Scholar
Wolter, A.C., Weickhmann, A.K., Nasiri, A.H., Hantke, K., Ohlenschlager, O., Wunderlich, C.H., Kreutz, C., Duchardt-Ferner, E., and Wohnert, J. (2017). A stably protonated adenine nucleotide with a highly shifted pKa value stabilizes the tertiary structure of a GTP-binding RNA aptamer. Angew. Chem. Int. Ed. 56, 401–404. CrossrefGoogle Scholar
Wunderlich, C.H., Spitzer, R., Santner, T., Fauster, K., Tollinger, M., and Kreutz, C. (2012). Synthesis of (6-13C)pyrimidine nucleotides as spin-labels for RNA dynamics. J. Am. Chem. Soc. 134, 7558–7569. CrossrefGoogle Scholar
Wunderlich, C.H., Juen, M.A., LeBlanc, R.M., Longhini, A.P., Dayie, T.K., and Kreutz, C. (2015). Stable isotope-labeled RNA phosphoramidites to facilitate dynamics by NMR. Methods Enzymol. 565, 461–494. CrossrefPubMedGoogle Scholar
Xu, J., Lapham, J., and Crothers, D.M. (1996). Determining RNA solution structure by segmental isotopic labeling and NMR: application to Caenorhabditis elegans spliced leader RNA 1. Proc. Natl. Acad. Sci. USA 93, 44–48. CrossrefGoogle Scholar
Young, W., Egert, M., Bassett, S.A., and Bibiloni, R. (2015). Detection of sialic acid-utilising bacteria in a caecal community batch culture using RNA-based stable isotope probing. Nutrients 7, 2109–2124. CrossrefGoogle Scholar
About the article
Paria Asadi Atoi studied Analytical Chemistry at the Alzahra University, Tehran, Iran. During her MSc, she worked on an experimental project in both fields of Electrochemistry and Separation. She started her PhD under the supervision of Stefanie Kellner at the Ludwig-Maximilians University Munich in September 2018. Since then, she is studying the dynamic nature of RNA modifications and their influence in neurological diseases.
Pierre Barraud studied Physical Chemistry at the École normale supérieure de Lyon (France). He then turned to structural biology and joined the group of Frédéric Dardel at the University Paris Descartes where he received his PhD in 2008. He then carried out postdoctoral research in the group of Frédéric Allain at the ETH Zurich (Switzerland), where he studied RNA-binding domains involved in mRNA processing and protein localization with NMR spectroscopy as a major tool. In 2014 he was recruited as a CNRS researcher and returned to France. He currently works at the Institut de biologie physico-chimique (IBPC) in Paris. His research focuses on structural aspects of RNA maturation in yeast and bacteria, with a special interest for chemical modifications in tRNAs.
Carine Tisne studied Physics and Biophysics at University Paris Diderot (Paris, France). For her PhD she joined Muriel Delepierre’s lab at the Pasteur Institute where she received her PhD in 1998. She then worked with Frédéric Dardel at the Faculty of Pharmacy in Paris (University Paris Descartes) to create a new laboratory of structural biology where she was appointed researcher at CNRS in 2002. In 2008, Dr. Tisne received the Maurice Nicloux prize from the French Society of Molecular Biochemistry and Biology for her work on the initiation of HIV-1 reverse transcription. From 2010 to 2016, she led a research group in this laboratory focused on RNA structure, interactions and anti-infectives, using NMR spectroscopy. She was deputy director of this lab from 2012 to 2016. In 2016, she moved to the IBPC to create a new group focused on biogenesis, architecture and interactions of RNA.
Stefanie Kellner received her state exam in Pharmacy in 2009 from the University of Heidelberg. For her PhD she joined Mark Helms lab at the Johannes Gutenberg University of Mainz where she received her PhD in 2012. After a 1-year Postdoc in the Helm lab, she moved 2014 to Boston for a second Postdoc. Here, she worked with Peter Dedon on the DNA phosphorthioate modification of bacteria at the Massachusetts Institute of Technology. In 2016 she returned to Germany where she started her own lab at the Ludwig Maximilians University in Munich funded by the Liebig stipend of the Fonds der Chemischen Industrie. Her main research focus is the analytical chemistry of nucleosides, especially modified RNA nucleosides. Since 2017 she is funded by the Emmy Noether Program of the DFG.
Published Online: 2019-04-08
Published in Print: 2019-06-26
Funding Source: DFG
Award identifier / Grant number: KE 1943/3-1
S.K. acknowledges funding from the Emmy-Noether program of the DFG (Funder Id: http://dx.doi.org/10.13039/501100001659, KE 1943/3-1). P.B. and C.T. acknowledge financial support from the CNRS, the ANR TriggeRNA (ANR-08-PCVI-0025), the ANR NMR-VitAmin (ANR-14-CE09-0012-01) and the Labex DYNAMO.