Measuring Transcription Dynamics of Individual Genes Inside Living Cells

Abstract

Transcription is a highly dynamic process, which, for many genes, occurs in stochastic bursts. Studying what regulates these stochastic bursts requires visualization and quantification of transcription dynamics in single living cells. Such measurements of bursting can be accomplished by labeling nascent transcripts of single genes fluorescently with the MS2 and PP7 RNA labeling techniques. Live-cell single-molecule microscopy of transcription in real time allows for the extraction of transcriptional bursting kinetics inside single cells. This chapter describes how to set up the MS2 or PP7 RNA labeling system of endogenous genes in both budding yeast (Saccharomyces cerevisiae) and mammalian cells (mouse embryonic stem cells). We include how to genetically engineer the cells with the MS2 and PP7 system, describe how to perform the live-microscopy experiments and discuss how to extract transcriptional bursting parameters of the genes of interest.

1 Introduction

Live-cell imaging has revealed that genes are often transcribed in bursts, where active periods are interspersed by periods of no transcriptional activity [1,2,3] (Fig. 1a). Understanding the biological steps that underlie transcription requires an understanding of how each regulatory step controls the dynamics of transcriptional bursting [4,5,6]. However, what regulates the start, end and intensity of bursts is largely unknown. Studying the molecular mechanisms regulating transcriptional bursting requires precise measurements of the transcription kinetics inside single cells. Since bursting is stochastic and asynchronous between individual cells, transcription dynamics cannot be captured by population-based assays such as reverse-transcriptase quantitative polymerase chain reaction (RT-qPCR) or RNA sequencing (RNA-seq), as the averaging over all cells within a population masks the dynamics that occur in individual cells. Single-cell approaches such as single-cell RNA-seq or single-molecule RNA fluorescence in situ hybridization (FISH) [7, 8] reveal instantaneous single-cell transcription levels, but these methods only access a single time point and do not directly reveal how transcription fluctuates over time. To measure dynamics, time-resolved measurements are needed. Although protein fluctuations have been used to infer dynamic information on transcription, these indirect measurements are limited by the half-lives of the protein reporters used [9]. Currently, the most sensitive techniques to directly measure transcription dynamics at individual genes in real time are the PP7 and MS2 systems. With the correct imaging settings, the MS2 and PP7 approaches provide single-molecule sensitivity.

Fig. 1

The PP7 and MS2 systems for real-time visualization of transcription dynamics in individual living cells. (a) Schematic showing transcriptional bursting: short periods of time during which several RNAs are produced. The high and low yellow line indicates active (ON) and inactive (OFF) periods of transcription. Blue vertical lines represent initiation events of single RNAs. (b) A sequence encoding for the PP7 or MS2 stem loops is integrated into the gene of interest (GOI). Upon transcription, the resulting mRNA molecule contains stem loops. These loops are bound by GFP-tagged MS2/PP7 coat protein (PCP/MCP) that is expressed in the same cell. Accumulation of multiple coat protein molecules on an RNA molecule and multiple RNA molecules at the TS results in a bright nuclear spot at the TS of the GOI. (c) Example image (maximum intensity projection) of budding yeast cells showing TSs of the GAL10 gene. Here, GAL10 is tagged with 14 × PP7 at the 5′end and the cell expresses PCP-GFPEnvy. Image taken on a Widefield AxioObserver microscope. Scalebar: 10 μm. (d) Example images (maximum intensity projections) of mouse embryonic stem cells (mESCs) showing TSs of Sox2-MS2v7 in cells expressing MCP-EGFP, imaged on LSM980 confocal Airyscan 2 microscope (left) and on Axio-Observer Widefield microscope (right). Scalebar: 10 μm

The MS2 and PP7 systems (Fig. 1b) are two orthogonal, bacteriophage-derived systems that each allow monitoring of transcription of an individual gene in real time [10,11,12]. For these systems, repeats of the MS2 or PP7 sequences are introduced into a gene of interest (GOI). Once transcribed, the resulting nascent RNA will form a series of stem loops, which are bound by a fluorescently tagged coat protein that is expressed in the same cell. Accumulation of multiple coat protein molecules on an RNA causes the RNA at the transcription site (TS) to appear as a bright fluorescent spot above the background fluorescence of the nuclear coat protein (Fig. 1c, d). Although in principle all cellular RNAs are fluorescently labeled, the difference in mobility between nascent and mature single RNAs allows one to use longer exposure times and lower excitation powers to motion-blur single-RNAs and to only visualize the site of nascent transcription. Fluctuations in transcriptional activity can then be monitored by measuring the intensity of the transcription site over time.

The MS2 and PP7 stem loop sequences are orthogonal, such that, by using two spectrally separable fluorophores for labeling of MS2-coat protein (MCP) and PP7-coat protein (PCP), the two methods can be used simultaneously [13]. The MS2 and PP7 techniques have been used to study bursting in a range of model organisms, including bacteria [14], Dictyostelium [15], Drosophila [16], budding yeast [17], mouse [9, 18, 19], and human [20, 21]. We describe the methodology for both Saccharomyces cerevisiae (budding yeast) and mammalian cells systems, where we focus on mouse embryonic stem cells (mESC). Because bursting is highly conserved, budding yeast allows for studying the basic principles of transcription regulation using fast and easy genetics [22]. Moreover, the clustered regularly interspaced short palindromic repeats (CRISPR) technology has enabled genetic editing of mammalian cells [23] for live-cell transcription imaging, providing insight into more complex layers of transcriptional regulation in higher eukaryotic systems.

2 Materials

2.1 Plasmids and Plasmid Amplification (For Budding Yeast)

• 1. Plasmid with MS2 (Addgene 104392) or PP7 (Addgene 189939) loops and LoxP-KanMX-LoxP resistance (see Note 1).

• 2. Plasmid expressing Cre-Recombinase, inducible with galactose (pSH47, from Euroscarf) and with URA selection marker (see Note 2).

• 3. Plasmid with MS2-coat protein (MCP, Addgene 189940) or with PP7 coat protein (PCP, Addgene 189941) fused to Green Fluorescent Protein Envy (GPFEnvy) for single genomic integration at the URA locus (see Note 3).

• 4. Restriction enzyme PacI (see Note 4).

• 5. Competent bacteria for plasmid amplification.

2.2 Plasmids and Plasmid Amplification (Mammalian Cells)

• 1. Plasmid construct with MS2 or PP7 loops (MS2v7: Addgene 140705).

• 2. pX330 plasmid with Cas9 and guideRNA (gRNA) (Addgene 42230),

• 3. Plasmid containing MCP or PCP with fluorophore of choice, such as Enhanced Green Fluorescent Protein (EGFP).

  (MCP-EGFP flanked by inverted terminal repeats for PiggyBac integration: Addgene 190003) + PiggyBac plasmid (from System Biosciences for example)

  (Addgene 61764: MCP-EGFP, lentiviral)

  (Addgene 103832: MCP-mCherry)

• 4. Cre recombinase plasmid.

  (Addgene 13775)

• 5. Restriction enzymes.

  BbsI for pX330 cloning.

• 6. T4 ligation buffer.

• 7. T4 PNK.

• 8. Quick Ligase and Quick Ligation Buffer.

• 9. Competent bacteria for plasmid amplification.

• 10. OneShot Stbl3 to be used for repetitive sequence cloning, such as the MS2 or PP7 loops.

2.3 Yeast Culture and Loop Integration (for Budding Yeast)

• 1. Incubator with rotating wheel (such as Brunswick Scientific TC-7 Tissue Culture Roller Drum Rotator) at 30 °C.

• 2. Yeast strain of choice with URA auxotrophy in which stem loops and coat protein are to be integrated.

• 3. Synthetic complete (SC) medium. The protocol of how to make this medium is provided in [24]: In a 500 mL bottle, add 450 mL of Milli-Q water, 3.35 g Yeast Nitrogen Base Without Amino Acids and carbohydrate and with Ammonium Sulfate (YNB w/o AA, carbohydrate & w/ AS) and 1 g Drop-out Mix Complete w/o YNB (alternatively, if the yeast strains used require this, a Drop-out Mix lacking specific amino acids should be used). Mix until the powder is dissolved. Sterilize by autoclaving 20 min at 121 °C. If using the media immediately, wait for media to cool down to approximately 55–65 °C before adding the carbon source. This can be done by placing the medium in a waterbath or incubator set within this temperature range. Alternatively, you can store medium at 18–22 °C for future use, preferably in the dark and without carbon source added. It then remains stable for months.

• 4. Yeast Peptone Dextrose (YPD) media: In a 500 mL bottle, add 450 mL of Milli-Q water, 5 g yeast extract powder and 10 g peptone in a 500 mL bottle.

  Sterilize by autoclaving 20 min at 121 °C. If using the media immediately, wait for media to cool down to approximately 55–65 °C before adding the carbon source. This can be done by placing the medium in a waterbath or incubator set within this temperature range. Alternatively, you can store medium at 18–22 °C for future use, preferably in the dark and without carbon source added. It then remains stable for months.

• 5. YPD plates. The protocol of how to make these plates are provided in [24]: Put 5 g yeast extract powder and 10 g peptone in a 250 mL bottle and put 10 g bacto-agar in a 500 mL bottle. Divide 450 mL Milli-Q water over both bottles. The two bottles should contain roughly equal volumes, but they are combined later on so the exact division over the bottles does not matter. Sterilize by autoclaving 20 min at 121 °C. Add YEP medium to agar solution and mix by inversion. Wait for media to cool down to approximately 55–65 °C. This can be done by placing the medium in a waterbath or incubator set within this temperature range. Add 50 mL 20% (W/V) D-glucose and mix by inversion. Pour plates of roughly 25 mL each. Pour slowly to prevent the formation of bubbles. Leave the plates to dry for 3 days at 18–22 °C. Put the plates upside down in a plastic bag and store at 4 °C. Store plates upside down to prevent infections and to prevent waterdrops from the lid from falling on the plates. Plates can be stored for months.

• 6. YPD + G418 plates: Plates can be made similar to YPD plates without selection marker, but after autoclaving, add 2 mL G418 (of stock solution with a concentration of 50 mg/mL) in a total volume of 500 mL (see Note 5).

• 7. SC-URA plates: In a 250 mL bottle add 3.35 g YNB w/o AA, carbohydrate and w/ AS, 1 g Drop-out Mix (minus URA) w/o YNB. Put 10 g bacto-agar in a 500 mL bottle. Divide 450 mL H₂O over both bottles. Sterilize by autoclaving 20 min at 121 °C. Add SC medium to agar solution and mix. Add 50 mL 20% (W/V) D-glucose (Sigma Aldrich, 8270-10KG). This will be enough to pour 20 plates. Allow the plates to dry for 3 days at RT, then store at 4 °C (see Note 6).

• 8. SC+5FOA plates: In a 250 mL bottle, add 3.35 g YNB w/o AA, carbohydrate and w/ AS, 1 g Drop-out Mix (minus URA) w/o YNB and 25 mg Uracil. Put 10 g bacto-agar in a 500 mL bottle. Divide 450 mL H₂O over both bottles. Sterilize by autoclaving 20 min at 121 °C. Add SC medium to agar solution and mix. Add 50 mL 20% (W/V) D-glucose and 0.5 g 5-fluoroorotic acid (5-FOA).Mix and put in waterbath at 55 °C. Swirl medium every now and then, until 5-FOA has been completely dissolved. Pour 20 plates. Allow the plates to dry for 3 days at RT, then store at 4 °C (see Note 7).

• 9. SC-URA medium: In a 500 mL bottle, add 450 mL H₂O, 3.35 g YNB w/o AA, carbohydrate and with AS and 1 g Drop-out Mix (minus URA) w/o YNB. Stir until the powder is dissolved and adjust the volume to 450 mL with H₂O. Sterilize by autoclaving for 20 min at 121 °C. Add 50 mL 20% (W/V) D-glucose or other appropriate carbon source before use.

• 10. Materials for yeast transformation (1 M lithium acetate, Tris–EDTA buffer (TE), PEG-3350, salmon sperm DNA, DMSO solution).

• 11. Thermocycler (see Note 8).

• 12. Materials for PCR amplification and gel electrophoresis, including high fidelity polymerase.

• 13. PCR Cleanup Kit.

• 14. PCR primers for generation of PCR template for loop integration and for colony PCR to check clones after integration of the loops, excision of the KanMX marker, and coat protein integration (see Note 9).

  • Primer pair for PCR template PP7 loop integration:

    Forward: 60-bp homology + CAAAGTGGGAGCGAGGAGATCC.

    Reverse: 60-bp homology + GCATAGGCCACTAGTGGATCTG; gives a 2611 bp band

  • Primer pair for PCR template MS2 loop integration:

    Forward: 60-bp homology + CCGCTCTAGAACTAGTGGATCC.

    Reverse: 60-bp homology + GCATAGGCCACTAGTGGATCTG; gives a 2571 bp band

  • Primer pairs for colony PCR to check clones for correct loop integration.

    Primer pair #1: spanning complete insert, design outside of homology arms based on sequence of gene of interest.

    Primer pair #2: spanning left-hand insertion site; design forward primer outside of left homology arm based on sequence of gene of interest, reverse primer: CGCGATCGCTGTTAAAAGGACAA.

    Primer pair #3: spanning right-hand insertion site;

    Forward primer: TTGTCCTTTTAACAGCGATCGCG; design reverse primer outside right homology arm based on sequence of gene of interest

  • Primer pairs for colony PCR to check clones for correct excision of the KanMX marker after Cre-Recombinase expression; design based on sequence of gene of interest.

  • Primer pairs for colony PCR to check clones for correct integration of PCP or MCP:

    Primer pair #1: CATCCGCTCTAACCGAAAAG and CCCTACACGTTCGCTATGCT; gives a 702 bp band if the integration was successful and does not give a band if the integration was not successful.

    Primer pair #2: GAGAAGGGCAACGGTTCAT and CCCTACACGTTCGCTATGCT; gives a 4515 bp band (for PCP) or 4518 bp (for MCP) if the integration was successful and a 1770 bp band if the integration was not successful.

2.4 Cell Culture and Loop Integration (Mammalian Cells)

• 1. Incubator at 37 °C, 5% (V/V) CO₂.

• 2. Cell line of choice.

• 3. Culture dishes (e.g., from Corning).

• 4. mESC cell culture media N2B27+2i+LIF. For 100 mL: 50 mL Neurobasal medium (phenol-red free), 50 mL DMEM/F12 (phenol-red free), 0.5 mL N2 supplement, 1 mL B27 (+Retinoic acid) supplement, 0.666 mL 7.5% (W/V) BSA, 1 mL Glutamine, 1-thioglycerol (1.5 × 10^–4 M), 1 mL 100× LIF (20 U/mL, Millipore), GSK-3β inhibitor CHIR99021 (3 μM), MEK inhibitor PD0325901 (1 μM) .

• 5. mESC culture media serum+LIF+2i. For 100 mL: 85 mL GMEM, 15 mL ES grade FBS, 1 mL glutamine, 1 mL sodium pyruvate, 1 mL non-essential amino acids, 100 μL of 0.1 M beta-mercaptoethanol, 1 mL 100× LIF (20 U/mL, Millipore), GSK-3β inhibitor CHIR99021 (3 μM), MEK inhibitor PD0325901 (1 μM) .

• 6. 0.1% (W/V) gelatin for plate coating

• 7. Trypsin.

• 8. Phosphate-buffered saline (PBS) solution.

• 9. Materials for method of integration: mESC electroporation kit and access to an electroporation device (Nucleofector^®).

• 10. Access to Fluorescence-Activated Cell Sorting (FACS) machine to sort cells based on fluorescent protein expression.

• 11. Antibiotics, if using antibiotic resistance for selection.

• 12. 96-well plate

• 13. Direct lysis PCR solution (Viagen Biotech).

• 14. Proteinase K.

• 15. Thermocycler.

• 16. Materials for PCR amplification and gel electrophoresis.

• 17. Access to Sanger sequencing.

2.5 Microscopy Sample Preparation (For Budding Yeast)

• 1. Clean coverslips (#1.5, thickness 0.17 mm) stored in 70% (V/V) ethanol (see Note 10).

• 2. Sonicator.

• 3. Spectrophotometer with capability to measure OD at 600 nm in cuvettes with a light path of 10 mm.

• 4. Attofluor™ Cell Chamber.

• 5. SC with appropriate carbon source (see Note 11).

• 6. Microscope stage incubation chamber, equipped with temperature control.

2.6 Microscopy Sample Preparation (Mammalian Cells)

• 1. Glass bottom dishes compatible with your stage insert.

• 2. BioLaminin 511 LN (BioLamina).

• 3. Parafilm.

• 4. ProLong™ Live Antifade (Thermo Fisher).

• 5. Microscope stage incubation chamber, equipped with temperature and CO₂ control.

2.7 Microscope (Both Yeast and Mammalian Cells)

• 1. Inverted widefield microscope (e.g., Zeiss AxioObserver Z1).

  • 60–100× Objective, NA >1.3 (see Note 12).

  • Light excitation source (at suitable wavelength depending on your fluorophore of choice). Either a LED source with narrow bandpass excitation filters or a laser (see Note 13).

  • Dichroic mirror to split excitation and emission light (see Note 14).

  • Suitable emission filters (depending on fluorophore of choice) (see Note 15).

  • Sensitive camera for fluorescence detection, for example, sCMOS or EMCCD (see Note 16).

  • Focusing strategy (see Note 17).

• 2. Confocal (use a fast system, described here is an LSM 980 Airyscan 2 setup).

  • Microscope: Zeiss Axio Observer 7 SP stand for LSM 980.

  • Scanning stage (see Note 18).

  • 60–100× Obective, NA >1.3 (see Note 19).

  • Laser excitation source (at suitable wavelength depending on your fluorophore of choice) (see Note 20).

  • Sensitive detectors/camera (see Note 21).

  • Emission filters depending on your fluorophore of choice (see Note 22).

  • Focusing strategy (e.g., DefiniteFocus 2, Zeiss).

3 Methods

3.1 Strategic Decisions Before Cloning and Integration of Stem Loops and Coat Protein (for Both Yeast and Mammalian Cells)

• 1. Consider where the loops are to be integrated into your GOIs (Fig. 2) (see Note 23).

• 2. Consider whether to use the PP7 or MS2 system. The two systems are orthogonal so they can be used to measure transcription dynamics at two locations simultaneously using dual-color microscopy (see Note 24).

• 3. Consider which version of the MS2 loops will be used (see Note 25).

• 4. Consider how many loops are optimal for your experiment (see Note 26).

• 5. Strategy for integration of the stem loops (for budding yeast) (see Note 27).

• 6. It is important to take into account that insertion of loops into a gene may affect transcription of the GOI. Moreover, integration may alter the protein levels by affecting the stability of the produced RNAs or translation (see Note 28).

• 7. Optimize the expression level of the coat protein (see Note 29).

Fig. 2

Distinct locations of PP7 or MS2 stem loop integration into the gene of interest provides distinct readouts of various aspects of transcription kinetics. (a) A single integration of either PP7 or MS2 stem loops at either 5′ or 3′ end of the gene of interest (GOI) provides readout of the transcription kinetics of the GOI. Insertion into an intron within the GOI provides a readout of transcription and splicing dynamics. (b) One can use both the PP7 and the MS2 stem loops in different genomic locations to obtain the dynamics of both signals in the same cell. Integration at two separate genes provides insight into how transcription of two genes is correlated. Integration at both ends of the same gene allows to determine the elongation rate of the gene. Integration in both an intron and an exon of the same gene allows the monitor splicing kinetics. Integration in both orientations within the same gene provides a readout of the sense and antisense transcription

3.2 Creating Yeast Strain with Stem Loops

• 1. Create repair template for yeast transformation by PCR from the plasmid containing the sequences of the stem loops and the KanMX marker. Homology arms for the transformation should be included as part of the PCR primers (Fig. 3a) (see Note 30).

• 2. To integrate loops into the yeast strain, use a standard transformation protocol to transform with 1–3 μg of PCR product generated above. Plate on selection plate (YPD+KanMX). After 3 days, restreak colonies on selection plates (YPD+KanMX). After 3 more days, check single colonies by colony PCR using the PCR strategy described in Fig. 3. Grow correct clones in 700 μL YPD overnight. In the morning, add 300 μL 50% (V/V) glycerol and store in −80 °C in a cryotube for long-term storage (see Note 31).

• 3. For positive clones, perform a second transformation with a plasmid encoding for the inducible Cre-Recombinase (pSH47) and plate on selection plate (SC-URA). Induce single colonies by inoculation in 1 mL SC-URA with galactose and grow overnight. Plate on SC+FOA to eliminate Cre-recombinase plasmid. Restreak each colony both on YPD and on YPD+G418. Clones that grow on YPD but not on YPD+G418 are likely to have successfully excised the KanMX marker. To verify, check these colonies for successful recombination by colony PCR (Fig. 3). Grow correct clones in 700 μL YPD overnight. In the morning, add 300 μL 50% (V/V) glycerol and store in −80 °C in a cryotube for long-term storage (see Note 32).

• 4. Further downstream controls should be performed to check for effects of integration on transcription, RNA and protein expression levels (see Notes 28 and 33).

Fig. 3

Schematic overview of genetic steps for integration of PP7/MS2 stem loops and PP7/MS2 coat protein in budding yeast. (a) PP7 or MS2 stem loops are integrated in yeast strain using transformation of PCR product containing PP7 or MS2 stem loops (yellow) and a KanMX marker (orange) flanked by LoxP sites (yellow) and homology to the gene of interest (GOI) (blue). Integration is verified by colony PCR using indicated PCR primers (red, purple, cyan). Subsequently, the KanMX marker can be excised using recombination of the LoxP sites by Cre recombinase. Recombination is verified by colony PCR using indicated PCR primers. (b) PP7/MS2 coat protein (PCP/MCP) is integrated at the ura3 locus in yeast strain using transformation of PacI-digested PCP or MCP plasmid, containing PCP/MCP tagged with GFPEnvy (green), URA3 selection marker (orange) flanked by LoxP sites (yellow) and homology to ura3Δ0 (blue). Integration is verified using colony PCR using indicated PCR primers (red and cyan)

3.3 Coat Protein Integration (Yeast)

To visualize the RNA stem loops integrated in the procedure above, fluorescently labeled MS2 coat-protein (MCP) or PP7 coat protein (PCP) (depending on the choice of PP7 or MS2 loops) must be integrated into the yeast strain (Fig. 3b). This can be achieved using a single-integration vector (SIV) [25] that integrates at the URA locus. These SIVs contain homology arms to the URA3 locus, such that the ura3Δ0 is replaced by the cassette from the plasmid (see Note 34).

The coat protein can be integrated according to the following procedure:

• 1. Digest the coat protein SIV plasmid using PacI and purify with a PCR Cleanup Kit.

• 2. To integrate this digested plasmid, use a standard transformation protocol to transform with 1 μg PacI-digested vector. Plate on selection plate (SC-URA). After 3 days, restreak colonies on selection plates (SC-URA). After 3 more days, check single colonies by colony PCR using the PCR strategy depicted in Fig. 3. Grow correct clones in 700 μL YPD overnight. In the morning, add 300 μL 50% (V/V) glycerol and store in −80 °C in a cryotube for long-term storage (see Note 35).

3.4 Creating Cell Line with Stem Loops (Mammalian)

• 1. Clone the donor template on a plasmid, containing the stem-loops, flanked by homology arms for your GOI, consisting of 500–800 bp upstream and downstream from your target site (Fig. 4a). In addition to the loop-sequence, the cassette must contain a selection marker that is in-frame with your gene of interest. For selection, an antibiotic resistance cassette can be used to select for knock-in cells using antibiotics, or a fluorescent protein marker can be used to sort for knock-in clones by FACS. To facilitate knock-in clone screening, add two primer sequences, which amplify the endogenous locus into the donor construct, next to the homology arms. A small sequence from the 5′ side of the homology arm is added to the 3′ end of the construct and vice versa. Knock-in clones will show amplification of two smaller amplicons, versus just one in the wildtype situation [21] (Fig. 4a, b) (see Note 36).

• 2. A second plasmid is required containing the gene encoding the Cas9 protein, along with a scaffold for gRNA expression, such as the pX330 plasmid. Design the gRNA to target a protospacer adjacent motif (PAM) sequence (NGG) near the insertion site in the target gene. Clone the pX330 plasmid to contain the gRNA targeting your GOI according to the guide from the Zhang lab [23] (see Note 37).

• 3. Design primers that contain the sequence of the designed gRNA, with nucleotide overhangs required for integration into the BbsI site in pX330, according to the guide from the Zhang lab [23]:

  • 5′ – CACCGNNNNNNNNNNNNNNNNNNN – 3′

  • 3′ – CNNNNNNNNNNNNNNNNNNNCAAA – 5′

    • Phosphorylate and anneal the oligos to form double-stranded DNA, by setting up the following reaction: 1 μL of oligo 1 and oligo 2 (100 μM), 1 μL of 10× T4 Ligation Buffer, 6.5 μL H₂O, 0.5 μL T4 PNK. Incubate in a thermocycler at 37 °C for 30 min, at 95 °C for 5 min and then ramp down to 25 °C at 5 °C/min. Dilute the annealed oligos 200×.

    • Digest the pX330 plasmid using BbsI with the following reaction: 1–2 μg of pX330, 2 μL 10× NEBuffer 2.1, 1 μL BbsI H₂O to a final volume of 20 μL. Incubate the reaction at 37 °C for at least 1 h. Run the digestion on an agarose gel to confirm efficient cutting. Heat inactivate the reaction at 65 °C for 20 min and column purify the backbone.

    • Set up the ligation reaction and incubate at RT for 10 min: 50 ng of digested pX330, 1 μL diluted phosphorylated and annealed oligos, 5 μL 2X QuickLigation Buffer H₂O up to 10 μL. Mix the reaction and add 1 μL of Quick Ligase.

    • Transform competent cells with 1–2 μL of ligation product. Pick colonies the following day and verify insertion of the gRNA.

• 4. The pX330 plasmid with gRNA and donor template plasmid are brought into the cells using electroporation. Seed your mESCs on gelatin-coated plates in serum-containing medium without antibiotics 1–2 days prior to electroporation. Make sure mESCs are in early growth-phase, showing small round colonies. Refresh the medium ~3 h before electroporation. Trypsinize mESCs to get a single-cell suspension and spin down, wash with PBS and spin down again. Prepare Eppendorfs containing 2–20 μg of plasmid DNA from the plasmids containing CRISPR-Cas9 and gRNA and homology template in 10 μL Mouse ES Cell Nucleofector Solution. Resuspend 1 × 10⁶ cells in 90 μL Mouse ES Cell Nucleofector Solution and transfer to a cuvette, then close the cuvette. Insert the cuvette into electroporation device and activate program appropriate for mESCs (e.g., A-30 on Nucleofector^® I device, or A-030 on Nucleofector^® II device). Following electroporation, add pre-warmed Serum LIF medium without antibiotics to the electroporated cells immediately and transfer to gelatin-coated plates. Refresh medium daily and expand electroporated cells for knock-in selection (see Note 38).

• 5. Grow out the electroporated cells for at least 1 week to dilute out plasmids, after which the cells can be sorted for expression of the fluorescence marker. Take along a negative control population of non-electroporated parental cells for comparison. Sort cells that express the fluorescence marker into single cells in a 96-well plate (see Note 39).

• 6. Once single colonies are growing out in the 96-well plate, pick them for expansion and screening. When a clone is passaged, take 5 μL of single cell suspension to be added to 25 μL of direct-PCR lysis buffer containing proteinase K. The lysate is incubated in a thermocycler (incubated at 55 °C for 30 min, and at 95 °C for 45 min) and can then be stored at −20 °C. Use this lysate for PCR-mediated integration screening (Fig. 4b). Run the PCR reactions on a 2% (W/V) agarose gel to identify positive knock-in clones.

• 7. Once positive knock-in clones have been identified, isolate genomic DNA from the clones to confirm correct integration by PCR integration checks (Fig. 4c). Sanger sequence the locus to see if any point mutations occurred during the knock-in process (see Note 40).

• 8. Further downstream controls should be performed to check for the effects of integration on transcription, RNA, and protein expression levels (see Notes 28, 33 and 41).

Fig. 4

Schematic overview of genetic steps to integrate PP7 or MS2 loops in mouse embryonic stem cells. (a) Donor template for PP7 or MS2 loops contains a fluorescent protein (FP) marker (orange) flanked by LoxP sites (orange), the MS2/PP7 stem loop repeats (yellow), screening primer sequences (green) and homology arms (HA) to the gene of interest (GOI) (blue). Together with the CRISPR-Cas9 and gRNA plasmid pX330, the donor plasmid is inserted in mouse embryonic stem cells (mESCs). (b) After insertion, clones are verified using PCR screening strategy, by using primers that bind to the screening primer sequences from the knock-in cassette (forward primer in magenta, reverse primer in purple). This PCR reaction will give differently sized amplicons based on the integration. With successful knock-in two regions will be amplified, whereas in a situation with no knock-in only one PCR product will be amplified. Three products represent heterozygous insertion. (c) Additional PCR checks on clone genomic DNA are required in knock-in clones to verify correct integration in the GOI, by amplifying the region that covers the end of the HA into the gene body or 3′ UTR into the knock-in cassette (primers in blue (5′) and dark green (3′)). Check that the loop repeats have not recombined by amplifying the loop region (primers in orange). (d) The plasmid containing the CMV-promoter (orange) and PP7/MS2 coat protein (PCP/MCP) fused to EGFP (green) insert, flanked by inverted internal repeats (ITRs) (yellow), together with the PiggyBac transposase plasmid are inserted into the mESCs that contain MS2 or PP7 loops

3.5 Stable Coat Protein Integration (Mammalian)

• 1. To visualize gene transcription in live cells, fluorescently labeled MS2 coat-protein (MCP) or PP7 coat protein (PCP) must be integrated into your cell line (Fig. 4d) (see Note 42).

• 2. Co-transfect cells with coat-protein plasmid, containing Piggybac recognition sites and the PiggyBac-transposase plasmid using electroporation (see Subheading 3.4, step 3).

• 3. Following the introduction of the fluorescently labeled MCP/PCP into the cell line, keep the cells in culture for at least 1 week to dilute out the plasmid. Use FACS for sorting out the cell population with stable coat protein integration. Sort cells into three populations of MCP/PCP expression levels, ranging from low, medium to high (see Note 43).

• 4. Take live cell movies with the three different levels of MCP/PCP to see which expression level is optimal for visualizing transcription in your cell line and to optimize imaging conditions (see Subheading 3.2) (see Note 44).

3.6 Sample Preparation for Microscopy (yeast)

The full protocol for culturing yeast for imaging is published in [24].

• 1. Four days before imaging, streak yeast cells onto a YPD plate.

• 2. In the morning of the day before imaging, start a culture by inoculating one colony of yeast into 1 mL of SC medium (see protocol above). Grow at 30 °C while rotating.

• 3. In the evening, dilute cultures into prewarmed medium (30 °C) to a density so that the cells have an optical density (OD_600nm) of ~0.2–0.4 in the morning (see Note 45).

• 4. Prepare an agar pad for imaging:

  • Add 0.4 g of agarose to 18 mL of SC medium (see protocol above; use medium without added carbon source), yielding a final agarose concentration of 2% (W/V) (see Note 46).

  • Boil the mixture in the microwave. Make sure the agar is completely melted.

  • Cool down to approximately 55–65 °C, but do not let the mixture solidify.

  • Add an appropriate carbon source, such as 2% (W/V) D-glucose (see Note 47).

  • Mix gently by inversion.

  • Pour into a petri dish.

  • Wait for agar pad to solidify (this takes at least 20 min).

  • Keep at 30 °C until use. Can be stored during one day, but always prepare fresh on the day of imaging.

• 5. On the morning of the imaging experiment, measure the OD_600nm of the culture using a spectrophotometer. If the OD_600nm is below 0.2, let the culture grow until the OD_600nm reaches 0.2. If the OD_600nm is above 0.8, dilute the culture to OD_600nm 0.1–0.2 and let it grow for two cell-cycles (180 min for wildtype yeast growing in glucose). If the OD_600nm is between 0.2 and 0.8, cells are ready for imaging. Then, proceed to the next steps.

• 6. When the cells are ready for imaging, spin 1 mL of yeast culture at 3824× g for 1 min.

• 7. Remove supernatant.

• 8. Resuspend cells in 4 μL of prewarmed SC medium with appropriate carbon source by vortexing briefly.

• 9. Take one cleaned coverslip (stored in 70% (V/V), see above) and rinse it with 70% (V/V) ethanol.

• 10. Dry the coverslip over a flame. Do not put the coverslip in the flame, as it is likely to break, and flaming may lead to cloudy coverslips. Just hold the coverslip close to the flame such that it dries quickly but the ethanol is not set on fire. Alternatively, you can air dry the coverslips on the bench or in a small oven (set to max 50 °C).

• 11. Place coverslip in the Attofluor™ Cell Chamber.

• 12. Pipet the entire volume (usually 4–10 μL) onto the coverslip.

• 13. Cut agar pad. This is easily done by stabbing a circle out of the agar imaging pad using a 14 mL growth tube.

• 14. Put agar pad on top of the yeast cells. Press down gently with your thumb. Be careful not to break the coverslip (see Note 48).

3.7 Sample Preparation for Microscopy (Mammalian)

• 1. Two to three passages leading up to the imaging experiment, switch your cells to phenol-red free N2B27 2i medium (see Note 49).

• 2. mESCs grow in dome-shaped colonies, which makes imaging of individual cells challenging. mESCs therefore require laminin coating of the imaging dish to promote growth in 2D structures. Dilute Biolaminin 511 in PBS to get a final concentration of 2 μg/cm², where the volume depends on the surface of your imaging dish that is compatible with your microscope insert. Make sure the solution is spread evenly across the surface (see Note 50).

• 3. Seal the plate with parafilm to prevent evaporation and leave the dish in the fridge overnight, or for more rapid coating, incubate the dish at 37 °C for 2 h.

• 4. Seed your cells at least 1 day before your imaging experiment onto a glass-bottom dish. Before seeding, take off the Biolaminin solution and wash once with PBS (see Note 51).

• 5. On the day of your imaging experiment, refresh the phenol red-free medium and supplement with ProLong™ Antifade.

3.8 Live Cell Microscopy

3.8.1 Widefield or Confocal

Two imaging modalities that are used for imaging of transcription dynamics are widefield fluorescence microscopy and confocal fluorescent microscopy. The main difference between these two modalities is the fluorescence excitation: in widefield fluorescence imaging, the entire field of view is illuminated with excitation light, whereas confocal imaging utilizes point-based illumination in combination with scanning to build up an image. Confocal excitation uses a higher intensity, but since the light is highly localized, each pixel is only illuminated briefly. In addition, confocal imaging reduces the amount of out-of-focus light, yielding a higher signal-to-noise ratio. Conversely, the advantage of a widefield system is its imaging speed and the use of cameras, which are generally more sensitive than confocal detectors. However, certain confocal systems, such as Zeiss Airyscan 2, also allow for extremely sensitive detection and fast point scanning. To reduce the excitation power during confocal imaging, we recommend opening the pinhole to collect as much light as possible (see Note 52).

In this section, we outline a typical experiment to monitor transcription of an individual PP7- or MS2-tagged gene in cells expressing PCP or MCP. This protocol is meant as a starting point, where the specific imaging parameters must be optimized depending on many experimental factors.

• 1. Turn on the microscope, camera, and fluorescence excitation light source.

• 2. Turn on the stage incubator and set it to the correct temperature. Let the temperature in the incubator stabilize; this takes at least 30 min (see Note 53).

• 3. Turn on the stage controller.

• 4. Start the microscopy software.

• 5. Focus on the cells using transmitted light.

• 6. Switch the light source to fluorescence excitation.

• 7. Move the stage to find a suitable field of view (see Note 54).

• 8. Set acquisition parameters in imaging software:

  • Excitation power. The power used depends on the excitation source and the expression of the GOI (see Note 55).

  • Set exposure time/scanning speed (see Note 56).

  • Pinhole: when using a confocal setup, use the most open setting for the pinhole (see Note 57).

  • Set an appropriate number of z-slices and z-step size to cover the entire cell volume (see Note 58).

  • Choose the number of frames (see Note 59).

  • Frame interval time; depending on the dynamics of your GOI (see Note 60).

  • Activate autofocus (see Note 61).

• 9. Turn on fluorescence excitation and focus on the center of the cells.

• 10. Start acquisition (see Note 62).

3.9 Data Analysis/Processing

To extract quantitative parameters describing transcriptional bursting from the obtained fluorescence images requires processing and analysis of these images. The precise steps of such analysis procedures depend on the biological question and the microscopy system used to acquire the images. In general, the following steps are needed to analyze and quantify the experiments described in this chapter (Fig. 5):

• 1. Making a maximum intensity projection (see Note 63).

• 2. Segmentation of cells and nuclei (see Note 64).

• 3. Detection and tracking of the TS over time (see Note 65).

• 4. From the TS intensity over time, a thresholding approach can be used to determine when the gene is ON and when the gene is OFF to extract the burst duration, time between bursts, and burst intensity (Fig. 1a) (see Note 66).

Fig. 5

Schematic overview of data analysis including analysis steps and sample images of mouse embryonic stem cells. Live image analysis pipeline overview, with mouse embryonic stem cells containing 24 × MS2v7 and MS2 coat protein fused to EGFP. (1) Making maximum intensity projections, showing sample images of z-slices of two timepoints of a live cell movie. (2) Next, nuclear segmentation is performed on the max projections to acquire nuclear masks. (3) Spot fitting is performed to detect active TSs, which are subsequently tracked throughout the movie. (4) Shows an example fluorescence trace of a TS in a single cell (magenta), and a binarized signal (green) showing that the TS is “ON” or “OFF”

4 Notes

• 1. This plasmid is used as a template for creating the PCR product needed to integrate the stem loops into the yeast strain of interest.

• 2. This plasmid is needed to excise the KanMX selection marker after integration of the loops.

• 3. Depending on the choice of loops (PP7 or MS2), one of these plasmids is needed to express the relevant coat protein fused to GFPEnvy.

• 4. PacI is needed to digest the SIV coat protein plasmid to create the template needed for integration in the yeast strain of interest.

• 5. These plates are needed for selection upon integration of the MS2 or PP7 stem loops.

• 6. These plates are needed for selection upon integration of the plasmid expressing Cre-recombinase and for selection upon integration of PCP or MCP.

• 7. These plates are needed to eliminate the plasmid expressing Cre-recombinase after excision of the KanMX marker from the yeast strain of interest.

• 8. Needed for generation of PCR template for loop integration and for colony PCR to check clones after integration of the loops, excision of the KanMX marker, and coat protein integration.

• 9. Colony PCR is performed by dissolving half of a yeast colony in 20 μL 20 mM NaOH. Incubate 5 min at 95 °C and spin for 5 min at 2000 rpm. Use 1 μL of supernatant as PCR template for PCR with MyTaq polymerase.

• 10. Coverslips are cleaned by 20 min sonication in 1 M KOH, rinsing in H₂O and subsequent 20 min sonication in 100% (V/V) ethanol, as described in [24].

• 11. SC media is used for microscopy rather than YPD because YPD is opaque and contains autofluorescent components, increasing the background signal in microscopy images.

• 12. For yeast, a magnification of at least 100× is recommended due to the small cell size. We use a Zeiss 4-alpha Plan-Apochromat 100×/1.46 Oil, 420792-9800-000 objective.

• 13. We used a Lumencor SPECTRA X light engine WL:360–680 nm with 470/24 excitation filter for GFPEnvy excitation.

• 14. We used a Chroma 59012bs dichroic for GFPEnvy.

• 15. We used a Semrock FF01-515/30–25 emission filter for GPFEnvy.

• 16. We used a sCMOS (ORCA Flash 4.0 V3, 000000-1370-927).

• 17. Focusing system is needed for experiments longer than approximately 10 min. We use the Zeiss definite focus 2 system.

• 18. We used a Zeiss PIEZO Scanning stage with multi-well plate insert.

• 19. We used a 63×/1.40 Oil, 420782-900-799 objective with 1.7× zoom. mESCs are quite small, so lower magnification could work better for different cell lines, enabling imaging of more cells at once.

• 20. We used a Zeiss 488 diode laser for EGFP excitation.

• 21. We used a Zeiss Spectral detector 32 channels GaAsP PMT plus 2 channels MA-PMT.

• 22. We used LSM 980 twin gate 488/639 main beam splitter. We used no secondary beam splitter to maximize the amount of emission light to the Airyscan 2 detection area.

• 23. The sites of integration of the loops depend on the research question and experimental limitations. Examples of different experimental setups using different loop integration strategies are outlined in Fig. 2. For measuring transcription dynamics at an individual gene, loops can either be integrated in the 5′untranslated region (UTR) or in the 3′UTR (Fig. 2a). In general, integration in the 5′UTR may have larger effects on the expression of the GOI. In yeast, owing to the relatively short 3′UTRs [26] and fast elongation rates [12, 13], 3′UTR integration may lead to a very brief fluorescence signal at the transcription site (TS) that may be difficult to detect. Therefore, we generally resort to integration in the 5′UTR for yeast genes, where the RNA is observed at the TS for a longer period of time. However, integration in the 5′ UTR can lead to nonsense-mediated decay through translational stop codons in the cassette, which can be circumvented by integration in the 3′UTR or use of loops that are optimized for translation [27, 28]. For measuring the dynamics of both transcription and splicing, loops can be integrated into an intron of the GOI [29]. In addition, the PP7 and MS2 system can be used simultaneously to label two genomic locations (Fig. 2b). For example, by introducing PP7 loops in one GOI and MS2 loops in another GOI, transcription at both genes can be monitored, revealing correlations between the GOIs [30]. Also, two sets of loops can be introduced at the 5′ and 3′ end of the same GOI to monitor transcription elongation rates at the single-cell level [13, 31]. Alternatively, inserting PP7 loops into the intron of a GOI and MS2 loops in the exon of the same GOI allows measurement of the kinetics of co-transcriptional splicing events [32]. Finally, by placing the loops in both sense and antisense direction, the temporal correlation between sense and antisense transcription can be monitored [17].

• 24. In principle, both the PP7 and MS2 system work in mammalian and yeast cells. However, in mammalian cells, we have noticed that, for some genes, the PP7 system can result in nuclear retention of the tagged RNAs. The MS2-system is therefore preferred when measuring transcription at a single location, and PP7 is used if the research question requires loops in a second location. For yeast cells, we find that both systems work well, but that the PP7 system has better signal-to-noise.

• 25. Multiple different versions of the MS2 loops are available, differing in their DNA sequence and affinity for the coat protein [33, 34]. In yeast, MS2v5 loops result in aggregate formation, and therefore the use of the lower-affinity MS2v6 is recommended [33]. For mammalian cells, the MS2v7 loops are preferred, as MS2v7 shows a brighter and more photostable fluorescence signal for single RNAs when compared to MS2v6 [34].

• 26. The number of loops may be optimized for a particular experiment. Increasing the number of loops increases the number of coat protein molecules accumulating on the nascent RNA at the TS and thereby increases the signal-to-noise ratio. In yeast, we generally use 12× MS2 and 14× PP7, which yields good results for genes with varying expression levels. We have noticed that addition of more loops may result in RNA-coat protein aggregates at highly transcribed genes. In mammalian cells, 24× MS2 gives a good signal over background. Experiments have been done with up to 128× MS2 [35]. However, a larger number of loops may result in larger perturbations in the expression of the tagged locus. If the loops are not placed in an intron, loops may change RNA stability, the efficiency of RNA nuclear export or translation. The optimal number of loops is a balance between obtaining sufficient signal versus minimizing the effect on expression of the GOI.

• 27. Loops can either be integrated using a standard yeast transformation procedure based on antibiotic selection, or by a CRISPR-Cas9 approach [36]. We have obtained good results with high integration efficiencies using antibiotic selection. We have also attempted to integrate PP7 loops using a CRISPR-based approach, but we have found that often fewer loops than expected are integrated into the GOI. We suspect this is due to the efficient homologous recombination in budding yeast, leading to recombination within the repetitive loop sequence and thus resulting in loss of loops. Therefore, we recommend and describe the strategy based on selection by antibiotic resistance.

• 28. To which extent the transcription or protein expression of your GOI is altered, can differ per gene or per cell line. Because the transcriptional activity is the main readout, it is important to ensure that overall transcription kinetics are not affected by the loop integration. The easiest way to do measure effects on transcription is to perform smFISH in untagged and tagged cells with probes against the GOI and to check whether the fraction of cells with active TSs in the nucleus and the intensity of the TSs is unchanged by the tagging. In addition, dual-color smFISH of heterozygously tagged cells with both probes against the GOI and the MS2/PP7 loops allows for comparison of the intensity of tagged and untagged alleles in the same cells. In addition to effects on transcription, there may be downstream effects on mRNA stability or translation. If the gene is essential or important for cell fitness, consider heterozygous tagging to ensure the availability of a wildtype copy of the gene. To test to what extent mRNA or protein levels are changed, perform RT-qPCR or smFISH to measure changes in RNA levels or Western blot to measure changes in protein levels upon insertion of the loops into the GOI.

• 29. For monitoring transcription at different genes with different expression levels, the nuclear coat protein expression levels may need to be optimized. This can be done by integration of the coat protein at multiple genomic loci or by changing the promoter driving coat protein expression. Increasing the coat protein expression level increases the background fluorescence intensity. If the coat protein expression level is too high, the TS is thus not visible above the background signal. If the coat protein expression level is too low, the loops may not be saturated with coat proteins. Moreover, in yeast the pool of nuclear coat protein is limited due to its small nuclear size. Nuclear export of RNA that contains coat protein-bound loops causes export of the coat protein, thereby depleting the nuclear concentration of coat protein. Imaging of highly expressed genes for extended time periods thus requires higher coat protein expression.

• 30. On our plasmids with loops, the sequence of the KanMX selection marker is flanked by LoxP sites so that it can be removed by Cre recombinase. We recommend using a high-fidelity DNA polymerase, such as Herculase or Phusion, to prevent mutations.

• 31. The step where clones are restreaked eliminates potential background colonies and ensures that single clones are selected. For these check-PCRs, one can use MyTaq DNA polymerase.

• 32. Removal of the selection marker reduces the effect of the loop integration on expression of the GOI. If you are integrating both the PP7 and MS2 loops in the same yeast strain, be aware that Cre-recombination leaves a LoxP “scar” in the genome after recombination. Therefore, when excising the marker after introduction of the second set of loops, unwanted recombinations between the two GOIs can occur. Make sure to check by PCR that this has not occurred. Alternatively, a CRISPR-based strategy can be used for insertion of the second set of loops [36].

• 33. For verification of transcription, we recommend comparing the TS intensity and fraction of active cells using smFISH. RNA levels can be compared to wildtype using smFISH or RT-qPCR (we do not recommend using qPCR on the repetitive loop sequence), and protein levels can be compared using Western blot or IF.

• 34. Integration at a single genomic locus ensures homogeneous coat protein expression levels between individual cells compared to plasmid-based expression. Although it is possible to express PCP or MCP from a plasmid, this will lead to more variability in coat protein levels between cells, which complicates downstream analysis.

• 35. The URA marker is flanked by LoxP sites so it can be removed using Cre-Recombinase as described above. However, the presence of the URA marker does not affect coat protein expression so removal is not necessary, unless further genetic modifications are to be performed where the URA marker is needed for selection.

  The excision of the KanMX marker after loop integration using Cre-recombinase (Subheading 3.2, step 3 of the protocol) results in a LoxP site that remains in the genome after recombination. If the URA marker is also removed using Cre-recombinase, be aware that this may lead to unwanted recombinations between the loops at the GOI and the ura3 locus. Additional PCRs should be performed to check for these recombinations.

• 36. For antibiotic selection, it is advisable to place a P2A protein cleavage sequence in front of your marker, for separate expression of your gene of interest and the antibiotic marker. We recommend that the sequence of the fluorescent or antibiotic marker is flanked by loxP sites, so that it can be removed by Cre-recombination following clone selection. We have experience using the HaloTag for selection, which can be visualized upon the addition of a dye to medium, which offers more flexibility when combining different fluorophores within a cell line [37].

  In case your GOI is not expressed when you are creating the knock-in (for example your GOI is only expressed upon cellular differentiation), you can choose to insert a selection marker under its own promoter, so the marker is expressed in knock-in cells irrespective of whether your GOI is expressed.

  By adding the sequences of screening primers to your knock-in construct, you can distinguish no knock-in, heterozygous or homozygous knock-in clones by PCR bands of different sizes, using only one primer pair and one PCR reaction per clone. By adding two separate sequences at both the 5′ and 3′ end of the construct, you confirm integration of the entire construct into your GOI. Make sure you place the sequences in such a way that they result in differently sized PCR products that can be distinguished on a 2% (W/V) agarose gel, and that the products are <500 base pairs.

  For cloning constructs with loops, it is important to use bacterial strains that have less risk of recombination of repetitive sequences, such as OneShot Stbl3 bacteria.

• 37. We recommend using web tools for designing your gRNA, to get optimal targeting and minimal off-target effects. If you are tagging the 3′ end of your gene heterozygously, do not target your gRNA exactly at the stop codon of the gene, but ~15–20 nt downstream in the 3′ UTR. Targeting the stop codon may result in loss of the codon through mutation and a new stop codon downstream, which may perturb the protein. Make sure the PAM sequence is not included in the gRNA sequence.

• 38. Depending on your cell line of choice, using Lipofectamine transfection may be more efficient in order to bring the two plasmids into your cells.

  For antibiotic selection, the medium should be replaced the day after transfection with medium containing the selection drug. The medium should be changed every 2–3 days until single colonies develop (this process could take a couple of weeks).

• 39. Certain mammalian cell lines may not be able to grow out from single cells. An alternative to single cell sorting for such cell lines is to sort the positive population in bulk and to seed the cells very sparsely onto a plate. Once single colonies grow out, these can be picked and separately expanded and genotyped.

• 40. We recommend checking by PCR that the loops have not recombined, resulting in a loss or increase in the number of loops in your clone. In addition, we recommend performing a PCR that amplifies a region that spans outside the homology arms, to ensure that integration occurred specifically in the endogenous GOI.

• 41. Removal of the selection marker using Cre recombination can reduce the effects of integration on your gene of interest. We note that if a separate promoter is used for the marker, removal of the promoter and the selection marker are essential.

• 42. Different coat protein constructs are available on Addgene (see materials). Your choice may depend on which fluorophore is most suited for the experiment and on how the MCP construct will be integrated. Depending on your cell line, the MCP could be randomly integrated into the genome by viral transduction or by transfection using a PiggyBac transposase. Alternatively, the MCP could be knocked into a safe harbor locus.

  Often, the coat protein is expressed from a highly active promoter, such as viral promoters (CMV), or the UbC promoter.

  In our experiments, we have used a UbC-promoter driven MCP fused with EGFP. The insert is flanked by inverted internal repeats, which will be recognized by PiggyBac transposase. Upon co-transfection of the MCP and PiggyBac transposase plasmids, the MCP will be randomly inserted into the genome.

• 43. We recommend using low to medium expression in mammalian cells, to obtain optimal signal over background when performing live cell imaging.

  For downstream analysis, it is important to obtain highly similar background levels among cells, enabling better quantification of the transcriptional burst signal.

• 44. If the coat protein is too variable between cells, consider sorting again using a narrower expression range, or using single-cell clones.

• 45. The exact dilution depends on the growth rate of the yeast strain and the carbon source. If the growth rate in your conditions is unknown, we recommend making a dilution series. Wild-type growth rates are 90 min doubling time in glucose, 180 min doubling time in raffinose, and 150 min doubling time in galactose.

  The optical density (OD_600nm) must be kept below 0.8, to keep the cells in mid-log phase. At higher densities, cells will enter diauxic shift, where the carbon source is exhausted, and cells grow based on aerobic respiration. During this phase, cells become autofluorescent and may exhibit lower gene expression.

• 46. Agarose is used rather than bacto-agar because it is more homogeneous and more transparent.

• 47. For a 2% (W/V) final concentration, add 2 mL of 20% (W/V) D-glucose or other appropriate carbon source.

• 48. If imaging for several hours or more on a microscope that does not control the humidity in the sample chamber, place an additional coverslip on top of the agar pad to prevent sample drying during the experiment.

  An alternative to the use of an agar pads to create a monolayer of cells for imaging is to adhere the cells to a glass coverslip using Concanavilin-A [24]. Using Concanavilin-A keeps the cells in liquid, such that media can be switched or drugs can be added during the experiment. However, cells are more mobile in this setup, so we prefer to use agar pads if possible.

• 49. Phenol-red increases the levels of background fluorescence in your cells.

• 50. As a reference, for ibidi 8 μ-Slide inserts, we use 300 μL per well.

• 51. Note that the laminin matrix will inactivate if let dry for too long, it is important to work quickly and to always cover the matrix with either PBS or medium when waiting to seed the cells. The density of seeding depends on how much time the cells need to expand before the imaging experiment. It is important to ensure that the cells are still in active growing phase (~70% confluent, colonies are round and separate) when imaging.

• 52. For yeast, widefield microscopy is generally sufficient. For mESCs, using Airyscan 2 confocal imaging enabled us to prolong our total imaging time from every 2 min for 2 h on a widefield fluorescence microscope, to every 2 min for over 10 h before the cells showed significant signs of phototoxicity. If you have a fast and sensitive system available, such as an Airyscan 2 or spinning disk confocal, we therefore recommend confocal imaging for mammalian cells (Fig. 1d).

• 53. Typically yeast cells are imaged at 30 °C and mammalian cells at 37 °C and 5% (V/V) CO₂. Warming up the incubator longer may reduce drift, due to the heating up of the optics.

• 54. In general, we aim for a field of view with 50–100 yeast cells in an area of 100 × 100 μm. Avoid cells with high levels of autofluorescence; these cells are likely dead or dying. mESCs grow in colonies, and we usually center our field of view around a colony. Check by focusing up and down to confirm that the colony is indeed growing in 2D and most nuclei lie within the imaging volume.

• 55. Typically, we use 0.1–1% power of the Lumencor source (50–500 mW output power) to monitor transcription dynamics over tens of minutes. If the power light source cannot be reduced further by software control, place neutral density filters in your light path to reduce excitation. Using confocal imaging, we use 0.25–2% laser power (30 mW) to image cells for hours.

• 56. For imaging gene transcription, we typically use around 150 ms exposure time with widefield fluorescence microscopy. With confocal imaging on LSM 980 Airyscan 2, we use the fastest scan speed (Zeiss ZEN blue setting 11, frame time of 235.93 ms and pixel time of 0.77 μs) and bidirectional scanning.

• 57. This allows you to collect as much light as possible, increasing sensitivity and minimizing the amount of excitation light and limiting phototoxicity.

• 58. We typically use 9–15 slices with 0.5 μm increments for yeast, or 15–20 slices with 0.5 μm increments for mESCs. This covers the entire nucleus in our microscopy setup. With the 1.3–1.46 NA objective that we typically use for yeast and mESCs, we slightly undersample the imaging volume, but we retain the ability to localize and track the TS. It is possible to increase the number of slices or decrease the increment, but this may lead to increased effects of phototoxicity and photobleaching.

• 59. The number of frames that can be acquired in a single experiment depends on the stability of the fluorophore, the light sensitivity of the cells, the excitation power, and exposure time. We can typically monitor transcription dynamics for up to 240 frames.

• 60. For imaging transcription of a highly active gene in yeast cells, we use 10–30 s interval between consecutive frames. In mammalian cells, larger intervals will catch most events, ranging from every 1–5 min or more for less active genes.

• 61. For experiments longer than 10 min, we recommend using an autofocus system to prevent z-drift during the experiment.

• 62. For each microscopy system and each cell line or yeast strain, the exact imaging parameters and microscope settings must be optimized. In general, this optimization is a trade-off between increasing the signal-to-noise by increasing the amount of excitation light and limiting or delaying the effects of phototoxicity and photobleaching. Phototoxicity can be recognized by a sudden termination of the fluorescence signal, termination of transcription, or termination of cell division. In mammalian cells, you often see cells decreasing in size, abnormal nuclear shapes, moving away from one another and rounding up of the cytoplasm.

  Examples of experiments performed at optimal imaging settings are provided in Video 1 for yeast and Video 2 for mESCs.

• 63. Making a maximum intensity projection captures the TS in all z-slices in a single image. This allows visualization of transcription of an individual cell in a 1D representation over time. Importantly, we recommend checking in the 3D stack that the TS does not move out of the 3D image stack in the z-direction.

• 64. This can be performed using CellProfiler [38], which provides an intuitive, GUI-based pipeline for image segmentation.

• 65. Because the TS is not always visible (it is only fluorescent during a transcriptional burst), it cannot be readily tracked with standard tracking packages such as TrackMate [39] or MatTrack [40]. To obtain a continuous trace at all timepoints (also during the OFF periods), we fit spot intensity (during the ON periods) and background intensity (during the OFF periods).The analysis pipeline that we use for analysis of transcription dynamics in budding yeast is available from https://github.com/Lenstralab/livecell_analysis. This script includes all of the steps above to quantify transcription dynamics of up to two genes (labeled with PP7 and MS2) simultaneously, and the key steps are described in [41].

• 66. Typically, we use a threshold based on the variation in background intensity levels.