paper type EAXTR VIEW Nucleus 5:1, 1–4; January/February 2014; © 2014 Landes Bioscience

The timing of pre-mRNA splicing visualized in real-time

Maria Carmo-Fonseca1,* and Tomas Kirchhausen2 1Instituto de Medicina Molecular; Faculdade de Medicina; Universidade de Lisboa; Lisboa, Portugal; 2Departments of Cell and Pediatrics; Harvard Medical School and Program in Molecular and Cellular Medicine at Boston Children’s Hospital; Boston, MA USA

ince it became clear that intervening isolated RNA molecules can hybridize Ssequences or are spliced out to double-stranded DNA by displacing from precursor pre-mRNA molecules one of the DNA strands.1 Using this

in the nucleus before mature mRNAs approach, the Roberts and Sharp teams distribute. are exported to the cytoplasm, independently realized that adenoviral questions were raised about the timing polysomal do not hybridize in a of splicing. Does splicing start while continuous manner with the DNA from not RNA polymerase II is still transcribing? which they are transcribed. Rather, the

Is splicing a slow or a fast process? Is mRNA sequence was complementary to Do timing important to control the splicing multiple noncontiguous regions of the reaction? Although our understanding DNA.2,3 They concluded that adenoviral on the mechanism and function of were split into several pieces, and splicing is largely based on data obtained because adenovirus is a mammalian using biochemical and large-scale virus, it seemed likely that at least some “omic” approaches, microscopy has been mammalian genes could have a similar instrumental to address questions related structure. This was soon found to be the to timing. Experiments done with the case. Although different mechanisms

electron microscope paved the way to could result in joining together certain Bioscience. the discovery of splicing and provided segments of genomic sequence into mature unequivocal evidence that splicing mRNA, evidence quickly accumulated can occur co-transcriptionally. More supporting a model for mRNA synthesis recently, live-cell microscopy introduced that involves splicing of a larger precursor a technical breakthrough that allows molecule or pre-mRNA.

real-time visualization of splicing Understanding the splicing mechanism Landes dynamics. We discuss here some of the required the development of cell-free microscopy advances that provided the systems where conditions could be Keywords: splicing, , basis for the current conceptual view of controlled and the components purified RNA, lifetime; live-cell imaging, single- the splicing process, and we outline a and reconstituted. In vitro splicing assays molecule visualization most recent development that permits based on extracts obtained by moderate *Correspondence to: Maria Carmo-Fonseca; direct measurement, in living cells, of salt solubilization of lysed cells or crude ©2014 Email: [email protected] the time it takes to synthesize and excise nuclei were able to accurately splice Submitted: 01/05/2014; Revised: an from individual pre-mRNA exogenously added pre-mRNA, indicating 01/26/2014; Accepted: 01/30/2014; molecules. that splicing and transcription are Published Online: 01/30/2014 In the summer of 1977, a set of electron independent processes (for a review see ref. http://dx.doi.org/10.4161/nucl.28056 micrographs shown at the annual Cold 4). Yet, visualizing dispersed chromatin by Spring Harbor Symposium revolutionized electron microscopy showed that splicing Extra View to: Martin RM, Rino J, Carvalho C, Kirchhausen T, Carmo-Fonseca M. Live-cell the way molecular biologists perceived was coupled to transcription in vivo. Back visualization of pre-mRNA splicing with single- the structure of genes. Following the in 1969, Miller and Beatty reported that molecule sensitivity. Cell Rep 2013; 4:1144-55; development of methodologies to observe hypotonic lysis of cells released chromatin PMID:24035393; http://dx.doi.org/10.1016/j. isolated RNA and DNA molecules by from the nucleus.5 The released chromatin celrep.2013.08.013 electron microscopy, it was found that was then centrifuged onto a grid for

www.landesbioscience.com Nucleus 1 observation in the electron microscope. The pioneer method relied on synthesis waiting for the opportunity to assemble This procedure resulted in a loosened of radioactive RNA complementary to a new . Since actively spliced two-dimensional array of chromatin the DNA of interest. Cells were made genes are intimately associated with fibers that maintained the nucleosomal permeable and incubated with the the periphery of nuclear speckles, one structure. Nascent transcripts remained radioactive probe, resulting in formation of possibility is that clustering of spliceosomal attached to the chromatin and appeared RNA-DNA hybrids. The location of these components in dedicated compartments as ribonucleoprotein fibers extending hybrids was then detected at the cytological increases their local concentration in the from the chromatin backbone. The level by autoradiography.11,12 Over the next neighborhood of nascent transcripts, thus length of nascent transcripts, which could 30 years, this technology progressed from enhancing spliceosome assembly on newly be precisely measured, increased with a laborious, time-consuming method synthesized introns.21 This could be a increasing distance from the transcription limited to the detection of highly abundant reason why splicing is so much faster in start site. Using this approach, Beyer and species with low cytological vivo than in vitro. Osheim showed that the spliceosome resolution, to a relatively fast, highly An imaging breakthrough was formed shortly after synthesis of the 3′ precise, and sensitive imaging of single introduced by the development of splice site and that splicing of pre-mRNA genes and RNA molecules in the cell. An genetically encoded fluorescent tags often occurred on the nascent transcript.6 important advance was the optimization that combined with fluorescence-based This visual demonstration that introns of the in situ hybridization procedure microscopic approaches of increasingly could be removed prior to transcript release for fluorescent detection of nuclear RNA higher spatial and temporal resolution, from the template and thus presumably using biotin-labeled DNA probes and making it possible to analyze distribute. prior to challenged the fluorophore-conjugated avidin.13 Next, movement in living cells. Time-series prevailing textbook view that splicing the method was further optimized for recordings of cells expressing the green took place after polyadenylation. Beyer the simultaneous detection of DNA, fluorescent protein (GFP) fused to not and Osheim assumed an elongation rate unspliced precursors, and spliced mRNA. an essential splicing protein provided

of 1500 nucleotides per minute to then Using these tools, several laboratories direct evidence that nuclear speckles Do deduce that splicing of introns removed showed that spliced mRNA localized in supply spliceosomal components to cotranscriptionally occurred within close proximity to the DNA from which nearby activated sites of transcription.22 3 min after synthesis of the 3′ splice it was transcribed, consistent with the Subsequent developments of methods site. Similar conclusions were obtained view that many introns are excised while such as FRAP (fluorescence recovery after when studying actively transcribed genes the pre-mRNA is still tethered to the photobleaching) unravelled the kinetic in Chironomus tentans salivary gland locus via RNA polymerase.14-16 There properties of splicing in the polytene chromosomes, which can be are, however, situations where splicing is nucleus of live cells.23 FRAP revealed that isolated by microdissection for direct uncoupled from transcription: in the case spliceosomal proteins are continuously 7

observation in the electron microscope. In of regulated events, moving in the nucleus, shuttling in and Bioscience. vivo splicing rates in the range of 5 min or intron removal can be delayed until after out of nuclear speckles within seconds. less were also measured biochemically for release of the pre-mRNA from the site Surprisingly, trafficking kinetics was adenovirus and β-globin transcripts.8-10 of transcription.17 In situ hybridization independent of ongoing transcription Such rates differed substantially from the procedures were also optimized to localize and splicing, challenging the view that results obtained in cell-free assays, where a the small nuclear RNAs that form the spliceosomal components are stored in 18 lag period of 15–45 min was observed at building blocks of the spliceosome. nuclear speckles until a signal triggers their Landes the start of the reaction before the onset By combining in situ hybridization and recruitment to nascent introns; rather, it of splicing, following which spliced RNA immuno-labeling for fluorescent and is more likely that the building blocks of accumulated for 2–3 h.4 Thus, splicing electron microscopic visualization, it the spliceosome are constantly roaming in vivo is at least 10-fold more rapid than became evident that spliceosomal small the nuclear space until they collide and in vitro. This discrepancy highlights nuclear ribonucleoproteins and splicing transiently interact with either a nascent the importance of nuclear organization: protein factors accumulated in dedicated pre-mRNA, to form a spliceosome, or with ©2014 most likely, spliceosome components in sub-nuclear compartments termed Cajal other splicing proteins localized in nuclear the nucleus are organized in a way that bodies and nuclear speckles (for a review speckles.24 FRAP analysis further revealed favors efficient recognition of splice site see ref. 19). Unlike other RNA-based that core spliceosomal snRNP proteins sequences as soon as they emerge from macromolecular machines in the cell such have a residence time of 15–30 s in the RNA polymerase. as the , form anew nucleoplasm, where spliceosomal Imaging the spatial organization of on each nascent intron and disassemble are thought to interact predominantly genes, nascent RNAs, and spliceosome after each splicing event.20 As most cells with pre-mRNA.25 Based on these components in the intact contain a surplus of long-lived spliceosomal results it was suggested that splicing can was made possible through successive components in the nucleus, these molecules be accomplished within 30 s, which is optimizations of another microscopic are dynamically oscillating between two significantly more rapid than previously technique termed in situ hybridization. states: being part of a spliceosome or reported.25

2 Nucleus Volume 5 Issue 1 Following the advent of genetically these significant limitations and the close to the 5′ splice site, the duration of encoded fluorescent protein tags, Belmont potential problems introduced into the these fluorescence cycles provided, for the and colleagues pioneered a method to interpretation of the data, we directly first time, a direct measure of the intron image chromatin dynamics in vivo. tracked in time the fluorescence emitted lifetime. We found that transcription They introduced bacterial lac operator by single pre-mRNAs labeled with MS2- and excision of short introns (1.3–1.4 kb) repeats into the genome of and GFP in the intron.32 We integrated a single occurs in 20–30 s, which implies a splicing mammalian cells that expressed a GFP-lac reporter gene in the genome of human rate much faster than previously reported repressor .26 Binding of the cells and used a spinning-disk confocal for the adenovirus-derived short intron fluorescent repressor to its target sequence microscope. Spinning disk confocal based on FRAP experiments.30 We believe makes that particular region of chromatin microscopy combines high sensitivity the reason for this discrepancy derives visible in the nucleus. A similar approach with high speed optical sectioning with from the higher resolution of our single- was later developed to visualize RNAs in minimal photobleaching, making it molecule analysis: the rapid fluorescence living cells by genetically inserting the one of the methods of choice to analyze fluctuations that we observed for a single binding sites for the MS2 bacteriophage intracellular dynamics of single molecules pre-mRNA molecule are probably hidden coat protein in the RNA of interest.27 with high spatial and temporal resolution. in the bulk measurement of fluorescence The resulting reporter gene was then The fluorescence associated with a recovery from a multitude of molecules at integrated in the genome of cells that single transcription site in the nucleus different stages of the splicing cycle. We expressed the MS2 coat protein fused to was detected as a diffraction-limited then addressed the longstanding question

GFP. Insertion of the MS2 binding sites object and its intensity was determined of whether some introns are spliced faster distribute. in the terminal of reporter genes as a function of time. Increments in than others. First, we compared the revealed kinetic properties of the entire fluorescence intensity were due to de lifetimes of the first and second introns mRNA cycle, from transcription to novo transcription of binding sites for of the β-globin gene and detected a 1.5- not transport in the nucleus and export to the MS2-GFP, while reductions reflected fold difference: the most frequent lifetime 28,29

cytoplasm. Based on these results, a intron excision followed by degradation value was 20 s for the first intron and Do prediction was that an intronic insertion or diffusion. These fluorescence intensity 30 s for the second; to rule out the influence of MS2 binding sites should suffice to fluctuations were used to determine the of transcription, the second intron was track splicing in real-time. Bertrand and intron lifetime, defined as the time it shortened by an internal deletion so that colleagues combined such an approach takes for an intron to be transcribed, both introns have approximately the same with FRAP to measure co-transcriptional spliced, and degraded (or released from length.32 Next, we compared the lifetimes splicing kinetics.30 They analyzed an the transcription site). The fluorescence of two introns derived from the mouse ensemble population of pre-mRNAs intensities at any given time point varied immunoglobulin µ gene. The two introns synthesized from a gene cluster comprising from cell to cell, presumably reflecting had exactly the same length but differed

~20 copies of a reporter gene that contains variation in the rate of transcription in the sequence. We Bioscience. MS2 binding sites in an artificial short initiation: the higher the transcription found that the lifetime of an intron with intron derived from the adenovirus rate the more nascent pre-mRNAs are a weak splice site sequence was 1.4-fold genome. Upon bleaching the MS2-GFP simultaneously present at the transgene longer than the lifetime of an intron with fluorescence associated with introns at locus, resulting in higher fluorescence a strong splice site sequence.32 Finally, we the transcription site, they measured a intensity. To determine the fluorescence analyzed the influence of transcription on

half-life of 105 s for fluorescence recovery intensity emitted by a single pre- intron lifetime by comparing two introns Landes and estimated a mean splicing time of mRNA molecule, we treated cells with with the same splice site sequences but 162 s. This value was in good agreement spliceostatin A (SSA), a potent splicing differing in length by ~1000 nucleotides with the splicing rates previously inhibitor that causes release of unspliced inserted downstream of the MS2 binding deduced from electron microscopy6,7 and pre-mRNAs to the nucleoplasm.33 After sites. We found that the cycles of biochemical pulse-chase experiments.31 release from the transcription site, pre- fluorescence gain and loss were ~18 s longer However, inferring the splicing rate mRNAs diffused throughout the nucleus for the longer intron. We also estimated ©2014 from FRAP curves required the use of making it possible to resolve individual that the array of MS2 binding sites is numerical models because fluorescence transcripts. Having determined the transcribed at a rate of approximately recovery is influenced by many variables fluorescence intensity of a single pre- 4 kb per minute. At this rate, transcription in addition to splicing. These include mRNA, we then searched for cells that of the extra 1000 nucleotides present rate of transcription by RNA polymerase synthesized one reporter transcript at in the long intron is expected to take II, diffusion and binding of MS2-GFP, the time, i.e., cells with fluorescence ~15 s, an interval very close to the and diffusion or degradation of excised fluctuations that started at background difference in duration of the fluorescence introns. Modeling of these results is further level, increased to a value in the range cycles for the long and short introns complicated by the simultaneous presence, corresponding to a single pre-mRNA, and (~18 s). This suggests that longer introns in the region analyzed, of multiple then returned again to background. As take longer to be spliced not because asynchronous transcripts. To circumvent the MS2 binding sites were inserted very splicing is slower but rather because they

www.landesbioscience.com Nucleus 3 take longer to be transcribed. We could of splicing, particularly in the context of appearance of distinct mRNAs in also infer that splicing is a much more alternative splicing decisions. This system response to TNF induction.34 rapid process than transcription.32 To could also be useful to study in real-time conclude, we developed an experimental how delays in splicing may regulate the Disclosure of Potential Conflicts of Interest system that we hope will prove valuable timing of expression of specific genes, No potential conflict of interest was to further reveal the kinetic properties as recently proposed for the sequential disclosed.

References 13. Lawrence JB, Singer RH, Marselle LM. Highly 25. Huranová M, Ivani I, Benda A, Poser I, Brody Y, localized tracks of specific transcripts within Hof M, Shav-Tal Y, Neugebauer KM, Stanek D. The 1. Thomas M, White RL, Davis RW. Hybridization interphase nuclei visualized by in situ hybridization. differential interaction of snRNPs with pre-mRNA of RNA to double-stranded DNA: formation of Cell 1989; 57:493-502; PMID:2541917; http:// reveals splicing kinetics in living cells. J Cell Biol R-loops. Proc Natl Acad Sci U S A 1976; 73:2294- dx.doi.org/10.1016/0092-8674(89)90924-0 2010; 191:75-86; PMID:20921136; http://dx.doi. 8; PMID:781674; http://dx.doi.org/10.1073/ 14. Xing Y, Johnson CV, Dobner PR, Lawrence JB. org/10.1083/jcb.201004030 pnas.73.7.2294 Higher level organization of individual gene 26. Robinett CC, Straight A, Li G, Willhelm C, Sudlow 2. Chow LT, Gelinas RE, Broker TR, Roberts transcription and RNA splicing. Science 1993; G, Murray A, Belmont AS. In vivo localization RJ. An amazing sequence arrangement at the 259:1326-30; PMID:8446901; http://dx.doi. of DNA sequences and visualization of large- 5′ ends of adenovirus 2 messenger RNA. Cell org/10.1126/science.8446901 scale chromatin organization using lac operator/ 1977; 12:1-8; PMID:902310; http://dx.doi. 15. Zhang G, Taneja KL, Singer RH, Green MR. repressor recognition. J Cell Biol 1996; 135:1685- org/10.1016/0092-8674(77)90180-5 Localization of pre-mRNA splicing in mammalian 700; PMID:8991083; http://dx.doi.org/10.1083/ 3. Berget SM, Moore C, Sharp PA. Spliced segments nuclei. Nature 1994; 372:809-12; PMID:7997273; jcb.135.6.1685 at the 5′ terminus of adenovirus 2 late mRNA. Proc http://dx.doi.org/10.1038/372809a0 27. Bertrand E, Chartrand P, Schaefer M, Shenoy SM, Natl Acad Sci U S A 1977; 74:3171-5; PMID:269380; 16. Custódio N, Carmo-Fonseca M, Geraghty F, Pereira Singer RH, Long RM. Localization of ASH1 mRNA http://dx.doi.org/10.1073/pnas.74.8.3171 HS, Grosveld F, Antoniou M. Inefficient processing particles in living yeast. Mol Cell 1998; 2:437- 4. Padgett RA, Grabowski PJ, Konarska MM, impairs release of RNA from the site of transcription. 45; PMID:9809065; http://dx.doi.org/10.1016/ distribute. Seiler S, Sharp PA. Splicing of messenger RNA EMBO J 1999; 18:2855-66; PMID:10329631; S1097-2765(00)80143-4 precursors. Annu Rev Biochem 1986; 55:1119-50; http://dx.doi.org/10.1093/emboj/18.10.2855 28. Janicki SM, Tsukamoto T, Salghetti SE, Tansey WP, PMID:2943217; http://dx.doi.org/10.1146/annurev. 17. Vargas DY, Shah K, Batish M, Levandoski M, Sinha Sachidanandam R, Prasanth KV, Ried T, Shav-Tal bi.55.070186.005351 S, Marras SA, Schedl P, Tyagi S. Single-molecule Y, Bertrand E, Singer RH, et al. From silencing to not 5. Miller OL Jr., Beatty BR. Visualization of nucleolar imaging of transcriptionally coupled and uncoupled : real-time analysis in single cells. Cell genes. Science 1969; 164:955-7; PMID:5813982; splicing. Cell 2011; 147:1054-65; PMID:22118462; 2004; 116:683-98; PMID:15006351; http://dx.doi. http://dx.doi.org/10.1126/science.164.3882.955 org/10.1016/S0092-8674(04)00171-0 http://dx.doi.org/10.1016/j.cell.2011.10.024 Do 6. Beyer AL, Osheim YN. Splice site selection, 18. Carmo-Fonseca M, Tollervey D, Pepperkok R, 29. Shav-Tal Y, Darzacq X, Shenoy SM, Fusco D, Janicki rate of splicing, and alternative splicing on Barabino SM, Merdes A, Brunner C, Zamore PD, SM, Spector DL, Singer RH. Dynamics of single nascent transcripts. Genes Dev 1988; 2:754- Green MR, Hurt E, Lamond AI. Mammalian nuclei mRNPs in nuclei of living cells. Science 2004; 65; PMID:3138163; http://dx.doi.org/10.1101/ contain foci which are highly enriched in components 304:1797-800; PMID:15205532; http://dx.doi. gad.2.6.754 of the pre-mRNA splicing machinery. EMBO J 1991; org/10.1126/science.1099754 7. Baurén G, Wieslander L. Splicing of Balbiani ring 10:195-206; PMID:1824936 30. Schmidt U, Basyuk E, Robert MC, Yoshida M, 1 gene pre-mRNA occurs simultaneously with 19. Handwerger KE, Gall JG. Subnuclear organelles: Villemin JP, Auboeuf D, Aitken S, Bertrand E. Real- transcription. Cell 1994; 76:183-92; PMID:8287477; new insights into form and function. Trends Cell Biol time imaging of cotranscriptional splicing reveals http://dx.doi.org/10.1016/0092-8674(94)90182-1 2006; 16:19-26; PMID:16325406; http://dx.doi. a kinetic model that reduces noise: implications 8. Curtis PJ, Mantei N, Weissmann C. Characterization org/10.1016/j.tcb.2005.11.005 for alternative splicing regulation. J Cell Biol 2011; and kinetics of synthesis of 15S β-globin RNA, 20. Rino J, Carmo-Fonseca M. The spliceosome: a self- 193:819-29; PMID:21624952; http://dx.doi. a putative precursor of β-globin mRNA. Cold organized macromolecular machine in the nucleus? org/10.1083/jcb.201009012 Spring Harb Symp Quant Biol 1978; 42:971- Trends Cell Biol 2009; 19:375-84; PMID:19616950; 31. Singh J, Padgett RA. Rates of in situ transcription and Bioscience. 84; PMID:277329; http://dx.doi.org/10.1101/ http://dx.doi.org/10.1016/j.tcb.2009.05.004 splicing in large human genes. Nat Struct Mol Biol SQB.1978.042.01.098 21. Misteli T. Cell biology of transcription and pre- 2009; 16:1128-33; PMID:19820712; http://dx.doi. 9. Gattoni R, Keohavong P, Stévenin J. Splicing mRNA splicing: nuclear architecture meets org/10.1038/nsmb.1666 of the E2A premessenger RNA of adenovirus nuclear function. J Cell Sci 2000; 113:1841-9; 32. Martin RM, Rino J, Carvalho C, Kirchhausen T, serotype 2. Multiple pathways in spite of excision PMID:10806095 Carmo-Fonseca M. Live-cell visualization of pre- of the entire large intron. J Mol Biol 1986; 22. Misteli T, Cáceres JF, Spector DL. The dynamics of mRNA splicing with single-molecule sensitivity. 187:379-97; PMID:3009832; http://dx.doi. a pre-mRNA splicing factor in living cells. Nature Cell Rep 2013; 4:1144-55; PMID:24035393; http:// org/10.1016/0022-2836(86)90440-7 1997; 387:523-7; PMID:9168118; http://dx.doi. dx.doi.org/10.1016/j.celrep.2013.08.013 Landes 10. Audibert A, Weil D, Dautry F. In vivo kinetics org/10.1038/387523a0 33. Martins SB, Rino J, Carvalho T, Carvalho C, Yoshida of mRNA splicing and transport in mammalian 23. Phair RD, Misteli T. High mobility of proteins M, Klose JM, de Almeida SF, Carmo-Fonseca M. cells. Mol Cell Biol 2002; 22:6706-18; in the mammalian cell nucleus. Nature 2000; Spliceosome assembly is coupled to RNA polymerase PMID:12215528; http://dx.doi.org/10.1128/ 404:604-9; PMID:10766243; http://dx.doi. II dynamics at the 3′ end of human genes. Nat Struct MCB.22.19.6706-6718.2002 org/10.1038/35007077 Mol Biol 2011; 18:1115-23; PMID:21892168; http:// 11. John HA, Birnstiel ML, Jones KW. RNA-DNA 24. Rino J, Carvalho T, Braga J, Desterro JM, dx.doi.org/10.1038/nsmb.2124 hybrids at the cytological level. Nature 1969; 34. Hao S, Baltimore D. RNA splicing regulates the Lührmann R, Carmo-Fonseca M. A stochastic ©2014 223:582-7; PMID:5799530; http://dx.doi. view of spliceosome assembly and recycling in temporal order of TNF-induced gene expression. org/10.1038/223582a0 the nucleus. PLoS Comput Biol 2007; 3:2019-31; Proc Natl Acad Sci U S A 2013; 110:11934-9; 12. Gall JG, Pardue ML. Formation and detection PMID:17967051; http://dx.doi.org/10.1371/journal. PMID:23812748; http://dx.doi.org/10.1073/ of RNA-DNA hybrid molecules in cytological pcbi.0030201 pnas.1309990110 preparations. Proc Natl Acad Sci U S A 1969; 63:378- 83; PMID:4895535; http://dx.doi.org/10.1073/ pnas.63.2.378

4 Nucleus Volume 5 Issue 1