Protocol to assess RNA-RNA interactions in situ using an RNA-proximity ligation assay

Summary

Here, we describe a protocol to assess RNA-RNA interactions in situ using an adapted proximity ligation assay (PLA). We detail steps to perform RNA-probe hybridization, in situ rolling circle amplification, and immunofluorescence confocal microscopy. With these tools, it is possible to detect and characterize the intracellular localization of interacting RNA pairs using small cell numbers. This protocol provides a targeted approach to understanding RNA-RNA interactions in intact cells that can complement other established deep-sequencing-based approaches.

For complete details on the use and execution of this protocol, please refer to Basavappa et al. (2022).¹

Graphical abstract

Highlights

• •Generate specific DNA probes to detect interacting RNAs in intact, fixed cells
• •Assess low-abundance RNAs using rolling circle amplification (RCA)
• •Visualize the localization of RNA-RNA interactions with fluorescence microscopy
• •Evaluate RNA-RNA interactions with minimal cell numbers

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Here, we describe a protocol to assess RNA-RNA interactions in situ using an adapted proximity ligation assay (PLA). We detail steps to perform RNA-probe hybridization, in situ rolling circle amplification, and immunofluorescence confocal microscopy. With these tools, it is possible to detect and characterize the intracellular localization of interacting RNA pairs using small cell numbers. This protocol provides a targeted approach to understanding RNA-RNA interactions in intact cells that can complement other established deep-sequencing-based approaches.

Before you begin

Design PLA probes

Timing: 2 weeks

To assess potential interactions between two RNAs, one probe is required for hybridization to each RNA. The PLA probes contain a tripartite structure: 1.) a 40–50 mer sequence complementary to your target RNA 2.) a polyA linker and 3.) an assay-specific, non-targeting PLA sequence (Figure 1).^2,3,4 One probe serves as a PCR primer (Priming) and the other probe serves as a non-priming partner (Non-Priming) containing three 2′-O-methylated-uridines at the 3′ end. This ensures distinct directionality of the resulting amplicon, allowing for maximal accumulation of a single product that can be detected downstream using an antisense oligonucleotide probe.

• 1.
  Design the complementary region targeting your RNAs of interest.

  • a.
    Navigate to the Stellaris Probe Designer tool (https://www.biosearchtech.com/support/tools/design-software/stellaris-probe-designer).
  • b.
    Enter your RNA sequence.

    • i.
      If the specific interacting regions are known a priori, enter only the relevant sequence range (if desired).
  • c.
    Adjust masking and sequence specificity requirements to the highest possible stringency.

    • i.
      If your sequence(s) of interest are short and/or highly repetitive, you may receive an error indicating that an insufficient number of probes can be designed. This is not a concern for this protocol. Only one probe region is required for RNA-PLA.
    • ii.
      Reduce the masking and sequence specificity stringencies ONLY if no probes can be designed for your sequence at the maximum level.
  • d.
    Using the sequence alignment view, look for regions where there are multiple probes in tandem. These represent more open regions of RNA that are more likely to be accessible for the longer sequence length required for the PLA probes.
  • e.
    Take the reverse complement of 40–50 nucleotides of the RNA sequence underlying the probes to derive a continuous sequence capable of binding your target RNA.
• 2.
  Append either the “Priming” or ‘Non-Priming” PLA sequence and the polyA linker to the 5′ end of the complementary region (see Figure 1).

  • a.
    Priming: TATGACAGAACTAGACACTCTT.
  • b.
    Non-Priming: GACGCTAATAGTTAAGACGCTT.
• 3.Append 3 methylated-uridines to the 3′ end of your designated Non-Priming PLA probe.
• 4.Purchase probes as 4 nM ULTRAmers with normal desalting from IDT.
• 5.Resuspend PLA probes in 80 μL of RNAse-Free H₂O to generate a 50 μM stock and store at −20°C.

Note: The PLA Connectors can also be ordered at this time in the same manner.

Figure 1.

PLA probe design

From the 5′ to 3′ direction, the PLA probes contain: i.) a sequence region of 40–50 nucleotides complementary to an RNA of interest, ii.) a polyA linker of 17–20 nucleotides, and iii.) an assay-specific, non-targeting PLA sequence.

Prepare coverslips

Timing: 2–3 days

• 6.
  Plate your cell type of interest on a sterile 12 mm² coverslip placed in a 24-well tissue culture plate to achieve 50%–70% confluence at the time of fixation.

  • a.
    For human brain microvascular endothelial cells (HBMEC), plate 2 × 10⁴ cells/coverslip.

CRITICAL: For the best quality images, cells should not be >70% confluent at the time of fixation. Depending on the growth rate of the cell type of interest and/or potential proliferative effects caused by downstream treatment conditions, modulate the number of initial cells plated to achieve a final confluency of 50%–70%.

• 7.Allow cells to adhere for 16–24 h using the appropriate incubation conditions for your cell type.
• 8.Treat/stimulate/infect cells as needed for your experiments.
• 9.Fix coverslips using 4% formaldehyde/1X DPBS for 10 min at 20°C–25°C.
• 10.Discard formaldehyde in appropriate secondary containment.
• 11.Wash coverslips 3× with 1× DPBS.
• 12.Store coverslips in 500 μL 1× DPBS at 4°C for up to 1 month.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Bacterial and virus strains
CHIKV	Dr. David Weiner, University of Pennsylvania	N/A
Chemicals, peptides, and recombinant proteins
Formaldehyde	Sigma	252549-500ML
1× DPBS (w/out Mg2+, Ca2+)	Invitrogen	Cat#14190136
TRITON X-100	Fisher Scientific	Cat#BP151-500
Tris-base	Fisher Scientific	Cat#BP152-5
Glacial acetic acid	Fisher Scientific	Cat#A38-212
Magnesium acetate	Fisher Scientific	Cat#BP215-500
Potassium acetate	Sigma	Cat#P1190-500G
Sodium chloride	Fisher Scientific	Cat#S271-10
Bovine serum albumin (BSA)	Fisher Scientific	Cat#BP1600-1
Tween-20	Fisher Scientific	Cat#BP337-500
Nuclease-free water	Sigma	Cat#W4502
T4 DNA ligase (400 U/μL)	New England Biotechnologies	Cat#M0202
Phi29 Polymerase	New England Biotechnologies	Cat#M0269
bisBenzamide H 33342 trihydrochloride (Hoechst)	Sigma	Cat#B2261-25MG
Vectashield	Fisher Scientific	Cat#NC9265087
Experimental models: Cell lines
Human brain microvascular endothelial cells (HBMEC)	Dr. Carolyn Coyne, University of Pittsburgh	N/A
Oligonucleotides
RNA-PLA: CHIKV-NonPriming:AGAGACATAGCTGTG TCACGCGTCTCCGCTGTTTCTTGTAAAAAAAAAAAA AAAAAGACGCTAATAGTTAAGACGCTT [UUU]	Basavappa et al.1	N/A
RNA-PLA: CHIKV50bpdown-Priming:TTGGTGCACCGA AGGAGATCGGCGGGTGACTCAGTTCCGTAAAAAAA AAAAAAAAAAAAATATGACAGAACTAGACACTCTT	Basavappa et al.1	N/A
RNA-PLA: ALPHA-Priming:TCACAGCAGGACACACTA TGTAATTCATATCAACATTTGGGAAAAAAAAAAAAAA AAAAAATATGACAGAACTAGACACTCTT	Basavappa et al.1	N/A
RNA-PLA: GAPDH-Priming:GCTGGCGACGCAAAAGA AGATGCGGCTGACTGTCGAACAGAAAAAAAAAAAA AAAAAAAATATGACAGAACTAGACACTCTT	Basavappa et al1	N/A
PLA Connector-1:/5Phos/CTATTAGCGTCCAGTGAAT GCGAGTCCGTCTAAGAGAGTAGTACAGCAGCCGT CAAGAGTGTCTA	Söderberg et al.3	N/A
PLA Connector-2:/5Phos/GTTCTGTCATATTTAAGCGTCTTAA	Söderberg et al.3	N/A
RNA-PLA: Amplicon Probe:/5Cy5/CAGTGAATGCGAGTCCGTCT	Söderberg et al.3	N/A
Other
Leica DM5500 confocal microscope	Leica	N/A
Hybridization oven	N/A	N/A
24-well tissue-culture-treated plates	Corning	Cat#3524
10 cm2 tissue-culture-treated plates	Corning	Cat#
12 mm2 coverslips	Fisher Scientific	Cat#1254581
1.5 mL microcentrifuge tubes	Fisher Scientific	Cat#05-408-129
Parafilm	Fisher Scientific	Cat#S37440
Paper towels	N/A	N/A
Pointed forceps	N/A	N/A
Microscope slides	Azer Scientific	Cat#2752511

Materials and equipment

PLA Permeabilization Buffer:

Reagent	Final concentration	Amount
TRITON X-100	0.2%	200 μL
1× DPBS		100 mL

Store at 20°C–25°C indefinitely.

2 M Tris-Acetate:

Reagent	Final concentration	Amount
Tris-Base	2 M	121.14 g
Glacial Acetic Acid	pH to 8.0
Milli-Q H2O	Fill to 500 mL

Store at 4°C for up to 1 year.

PLA Blocking Buffer:

Reagent	Final concentration	Amount
2 M Tris-Acetate, pH 8.0	10 mM	2.5 mL
Magnesium-Acetate	10 mM	777 mg
Potassium Acetate	50 mM	2.45 g
Sodium Chloride	250 mM	7.3 g
BSA	0.25 μg/mL	125 mg
Tween-20	0.05%	250 μL
Milli-Q H2O	Fill to 500 mL

Store at 4°C for up to 6 months.

Step-by-step method details

PLA probe hybridization

Timing: 13.5–19.5 h

Fixed cells are permeabilized and incubated for 12–18 h with PLA probes to allow for binding to putative interacting RNAs (Figure 2).

• 1.To permeabilize cells, add 500 μL of PLA Permeabilization Buffer to each well containing a coverslip and incubate for 10 min at 20°C–25°C.
• 2.Aspirate PLA Permeabilization Buffer.
• 3.Block cells using 500 μL of PLA Blocking Buffer per well containing a coverslip at 37°C for 1 h.
• 4.
  Prepare PLA probe mixture:

  • a.
    Prepare probe working solutions:

    • i.
      In an RNase-free, 1.5 mL microcentrifuge tube, dilute an aliquot of the probe stock solution (50 μM) to 5 μM using RNase-free H₂O.
  • b.
    For each coverslip, add 1 μL of Priming probe and 1 μL Non-Priming probe to 48 μL of PLA Blocking Buffer.
  • c.
    Mix the probe mixture well by pipetting gently up and down.

Note: Working solutions of 5 μM can be stored at −20°C and freeze-thawed multiple times without loss of PLA reactivity.

• 5.
  Build a hybridization chamber.

  • a.
    Fold a clean paper towel into a square such that it fits within a 10 cm² tissue culture dish.
  • b.
    Soak the paper towel thoroughly with dH₂O and place inside a 10 cm² tissue culture dish.
  • c.
    Cut a piece of parafilm such that it fits within the paper towel square.

Note: The parafilm square should be smaller than the paper towel such that the edges of the paper towel are exposed and form a border around the parafilm.

• 6.Pipette the prepared 50 μL PLA probe mixture onto the parafilm (it will form a contained droplet).
• 7.Using a 200 μL pipette tip, lift the coverslip off the bottom of the well.
• 8.Gently grasp the coverslip at the edges with forceps.

Note: Flat rather than pointed forceps are better for this step.

• 9.Place the coverslip onto the PLA probe mixture such that the cells are in contact with the liquid.
• 10.Seal the chamber with parafilm.
• 11.Incubate at 37°C for 12–18 h.

Figure 2.

RNA-PLA

Fixed, permeabilized cells plated on coverslips are subjected to in situ hybridization with PLA probes possessing a tripartite structure: i.) a 40–50 nt sequence complementary to a target RNA of interest ii.) a polyA linker iii.) and an assay-specific, non-targeting oligonucleotide sequence designated “Priming” or “Non-priming” by the addition of three methylated-uridines at the 3′ end (marked by a black circle). Following hybridization, two PLA Connectors are introduced which can complement with the PLA sequence of the PLA probes (region (iii)). If the PLA Connectors are in close proximity to each other, the addition of T4 DNA ligase will result in circularization of the PLA Connectors. This circle then serves as a template for rolling circle amplification (RCA) driven by the priming PLA probe which generates a single-stranded, contiguous PCR product which can be detected using a Cy5-conjugated antisense DNA oligonucleotide probe coupled with confocal microscopy.

PLA connector and linker ligation

Timing: 30 m

Two PLA Connectors are introduced into the cells after PLA probe hybridization. These Connectors can complement and bind the PLA probes. After binding, if the PLA Connectors are in close proximity to each other (suggesting that the PLA probe-bound target RNAs are also in close proximity) they can be circularized by T4 DNA ligase (Figure 2).

• 12.Using forceps, remove the coverslip from the hybridization chamber and place it in a 24-well plate with the cells facing up.

Note: The same 24-well plate used for storage, permeabilization, and blocking, can also be used for the remainder of this protocol. The previous contact with buffer can help in the following steps to spread the ligation mixture evenly across the coverslip. Wash the wells 3× with 1X DPBS prior to replacing the coverslip.

Note: Add 500 μL 1X DPBS to the wells prior to adding the coverslip.

Note: Make sure the coverslip is securely placed on the bottom of the well and is not floating in the 1X DPBS.

• 13.Wash 3× with 1X DPBS with gentle rocking agitation for 5 min for each wash.
• 14.Prepare ligation master mix in the following order:

Reagent	Amount
Nuclease-free H2O	178 μL
T4 DNA Ligation Buffer (10×, NEB)	20 μL
PLA Connector-1 (50 μM)	0.5 μL
PLA Connector-2 (50 μM)	0.5 μL
T4 DNA Ligase (NEB)	1 μL

CRITICAL: Prepare the ligation master mix IMMEDIATELY before placing it onto the coverslip. If prepared too far in advance, the enzyme will begin ligating the PLA Connectors in the tube, resulting in non-specific PLA signal downstream.

• 15.Mix ligation master mix quickly by inverting the tube 3–4 times and/or pipetting up and down.
• 16.Add 200 μL of ligation master mix to each coverslip.

CRITICAL: Pipette the master mix in a dropwise fashion across the coverslip to help the liquid spread. The master mix may pool in certain areas within the well, leaving regions of the coverslip dry. Draw up the mixture and pipette it along dry areas until you achieve full coverage of the coverslip. This may take a few repetitions.

• 17.Replace plate lid and incubate at 37°C for 30 min.

Rolling circle amplification (RCA)

Timing: 2 h

After PLA Connector ligation, the resultant circle serves as a template for phi29 polymerase-mediated rolling circle amplification (RCA) driven by the Priming PLA probe. This will generate a single-stranded, processive PCR amplicon which can be detected downstream.

• 18.Aspirate ligation master mix and discard.
• 19.Wash coverslips 3× with 1X DPBS with gentle rocking agitation for 5 min for each wash.
• 20.Prepare rolling circle amplification (RCA) mix in the following order:

Reagent	Amount
Nuclease-free H2O	218.5 μL
Phi29 Buffer (10×, NEB)	25 μL
10 mM dNTPs	2.5 μL
BSA (20 mg/mL, NEB)	2.5 μL
Anti-Amplicon-Cy5 Probe (5 μM)	0.5 μL
Phi29 (NEB)	1 μL

• 21.Mix master mix by inverting 3–4 times and/or pipetting up and down.
• 22.Add 250 μL of RCA master mix to each coverslip.
• 23.Replace plate lid and cover with aluminum foil to protect from light.
• 24.Incubate at 37°C for 1 h 40 min.

PLA signal detection

Timing: 2–6 h

In the previous step, a Cy5-conjugated, antisense DNA oligonucleotide probe complementary to the RCA amplicon was introduced as part of the phi29 master mix. Here, we visualize the Cy5-conjugated probe bound to the RCA amplicon using confocal microscopy (Figure 2).

Note: All incubations should be done protected from light from this point forward.

• 25.Aspirate RCA master mix and discard.
• 26.
  Prepare Hoechst working solution:

  • a.
    Dilute 5 mg/mL Hoechst stock 1:1000 in 500 μL 1X DPBS/coverslip.
• 27.Add 500 μL Hoechst working solution to each coverslip.
• 28.Incubate at 20°C–25°C with gentle rocking agitation for 15 min.
• 29.Aspirate Hoechst and discard.
• 30.Wash 3× with 1X DPBS with gentle rocking agitation at 20°C–25°C for 5 min each wash.
• 31.
  Mount the coverslip onto a microscope slide.

  • a.
    Pipette approximately 10 μL of Vectashield mounting media per coverslip onto the slide(s).
  • b.
    Using a 200 μL pipette tip, lift the coverslip from the bottom of the well.
  • c.
    Grasp the coverslip at the edges using forceps.
  • d.
    Gently dab the bottom edge of the coverslip onto a Kim-Wipe to remove excess 1X DPBS.
  • e.
    Place the coverslip at a 45° angle such that the cells are facing towards the Vectashield.
  • f.
    With a swift motion, allow the coverslip to drop onto the Vectashield.
  • g.
    (Optional) Seal coverslips with clear nail polish.
• 32.Let slides cure at 20°C–25°C for 2 h.

Note: Slides can be imaged after as little as 15 min however, we have observed better quality images following a longer cure.

Pause point: Slides can also be stored at −20°C for one day. We do not recommend storing slides for longer than three days as the Cy5 signal will begin to fade.

• 33.Image using a Leica DM5500 or equivalent confocal microscope.

Note: We use a 63× oil immersion objective to image RNA-PLA samples.

Note: We recommend collecting images along the outer third of the coverslip circumference where cells are better distributed, rather than in the center. Furthermore, the ligation and RCA mixes tend to pool around the edges due to surface tension, resulting in more robust signal. We collect images for 40–50 cells throughout all four quadrants of the coverslip to prevent bias as much as possible.

Expected outcomes

We performed RNA-PLA in HBMEC infected with chikungunya virus (CHIKV) for 24 h. Infection with an RNA virus such as CHIKV provides an abundant, well-defined RNA pool to detect and optimize this approach. As a positive control, we used two PLA probes (one Priming and one Non-Priming) both targeting the CHIKV genome but residing 50 nucleotides apart from each other. We hypothesized that the close proximity of these probes should yield robust signal. As a negative control, we paired the CHIKV-Non-Priming probe with a Priming probe designed to bind the cytoplasmic host mRNA GAPDH. In parallel we also paired the CHIKV-Non-Priming probe with a priming probe targeting the antiviral lncRNA, ALPHA. In previous work, we have shown that ALPHA binds CHIKV RNA by orthogonal methods.¹ Of note, ALPHA is a low abundance transcript providing a means to assess the sensitivity of the RNA-PLA assay.

As anticipated, CHIKV + CHIKV yielded abundant PLA signal, presenting as distinct puncta in the cytoplasm (Figure 3). Importantly, pairing GAPDH + CHIKV, produced very few PLA puncta (Figure 3). Strikingly, we observed a visibly higher, intermediate number of PLA puncta in the ALPHA + CHIKV condition relative to GAPDH + CHIKV, even though ALPHA is expressed at much lower levels than GAPDH (Figure 3). This suggests that lower abundance RNAs can be effectively detected by RNA-PLA with a relatively high level of specificity. Furthermore, the localization of potential RNA-RNA interactions can be easily visualized within cells and could potentially be coupled with co-immunostaining for organelle markers.

Figure 3.

RNA-PLA detects distinct RNA-RNA interactions in HBMEC

A PLA probe designed to detect CHIKV genomic RNA was paired with one of three partner probes: 1.) a second probe targeting CHIKV which binds 50 nucleotides downstream of the first CHIKV probe (Virus + Virus), 2.) a probe targeting the host mRNA, GAPDH (GAPDH + Virus), or 3.) a probe targeting the host antiviral lncRNA, ALPHA (ALPHA + Virus), which has been shown to bind CHIKV RNA by orthogonal methods. RNA-PLA was performed and images of approximately 50 cells/condition were collected by confocal microscopy. Representative merged images are shown. Scale bars represent 10 μm.

Limitations

While RNA-PLA has many potential advantages, it should not be used as the sole tool for identifying RNA-RNA interactions. As the technique’s name suggests, it is possible that two RNAs that are in close enough proximity to each other within the three-dimensional space of a cell can generate positive PLA signal without necessarily directly interacting. Furthermore, while the RCA step allows for detection of lower abundance RNAs, there is likely an expression threshold below which RNA-PLA will no longer be effective. Along the same lines, conclusions about relative RNA abundance should not be made from RNA-PLA. Unlike single-molecule RNA FISH and other similar techniques, the addition of an amplification step in the RNA-PLA protocol may cause the transcript number to PLA signal ratio to be non-linear, preventing concrete conclusions about either absolute or relative RNA abundance.

Troubleshooting

Problem 1

Unexpected signal in negative controls (steps 3, 4, and 14).

Potential solution

High background signal in negative controls may be a result of inappropriate ligation of the PLA Connectors during preparation of the ligation master mix prior to being added to the cells. As indicated within the protocol, be sure to only prepare the ligation master mix right before addition to the coverslip to avoid this problem. Alternatively, the inclusion of sheared salmon sperm DNA (sssDNA) to the PLA Blocking Buffer during both the block and hybridization steps, may help to eliminate non-specific signal. We recommend performing a titration with your positive control condition to find the optimal sssDNA concentration for your experiments.

Problem 2

No signal in positive controls (step 4).

Potential solution

It is possible that multiple probe combinations will have to be empirically tested in order to generate positive signal. Furthermore, the ideal parameters for your treatment conditions of interest may need to be specifically optimized for RNA-PLA. For example, in the context of viral infection, we recommend performing both a time course and testing a variety of multiplicity of infections (MOIs) using your positive control probes to maximize signal.

Problem 3

Weak or sparse signal across conditions (step 20).

Potential solution

Weak signal may result from suboptimal cell density. Sparse cell confluency may indicate improper fixation. Ensure that the 4% formaldehyde solution is fresh and stored away from intense light. Overly confluent cells can also prevent the optimal distribution of PLA probes and reagents, resulting in weak signal. Finally, the Cy5 probe working solution should be reprepared regularly as multiple freeze thaws will cause signal degradation over time.

Problem 4

Cells with unusual morphology (steps 8, 12, and 31).

Potential solution

Improper handling of the coverslip can result in scratched cells which appear folded, producing low quality images. Ensure that the coverslip is only being grasped at the very edges to avoid this problem. Unexpected morphology may also be indicative of suboptimal treatment conditions causing unintended stress to the cells. For example, very high MOIs of CHIKV can cause cellular damage. We recommend testing a matrix of treatment variables using the same plating conditions required for RNA-PLA i.e., cells plated on coverslips, to find the ideal parameters required for the phenotype of interest while maintaining optimal cell viability.

Problem 5

Blurred and/or low quality images (steps 31–33).

Potential solution

Excess Vectashield can cause the coverslip to shift while imaging, making it very difficult to acquire high quality images. This will be evident if the Vectashield is seeping out around the edges of the coverslip. If this is observed, reduce the Vectashield volume to mount future coverslips and thoroughly dab the edges of the coverslip on a clean paper towel to remove as much excess liquid as possible. Clear nail polish can also be used to seal the coverslips to help prevent movement. Increasing the cure time may also be useful. Bubbles in the Vectashield will also result in poor images. Using your non-dominant hand to stabilize the edge of the coverslip will help ensure a swift motion while mounting that should successfully prevent bubbles. Finally, confirm that the microscope is placed on a stable and level tabletop devoid of extraneous vibrations as even small disturbances in the environment can dramatically affect image acquisition quality.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Sara Cherry (cherrys@pennmedicine.upenn.edu).

Materials availability

This protocol did not generate unique reagents.

Data and code availability

This protocol does not report new data or code.