FISHing RNA for beginners

Fluorescent in situ hybridization is a very powerful method to optically detect specific nucleic acid sequences in fixed but otherwise intact cells. FISH can be used to detect specific DNA or RNA sequences. DNA FISH can yield interesting things (for instance to determine the number of gene copies, the location of the gene in the nucleus when it is turned on/off, DNA repair and replication studies). However, I will discuss here about RNA FISH, which has become a very powerful tool to study gene expression and RNA biology. I described an RNA FISH application in a previous post.

RNA FISH is a quantifiable method that can determine the localization of single mRNA molecules in the cell, their number, their transcription and degradation rates and more. I will dicuss these applications in other posts.

Recently, I started FISHing myself. Obviously, I ran into some difficulties (but also success). So I’ll start by describing the method and protocol, some tips and problems to avoid, how I analyze the data, different types of dyes etc… Hopefully, I’ll also learn something from all this writing.

The idea of FISH (or rather, ISH; patented by my PI almost quarter of a century ago) is that nucleic acids with complementary sequences tend to form a double helix. It can be DNA:DNA, RNA:RNA or RNA:DNA. Therefore, one can synthesize a short DNA probe that carries some sort of tag. In the olden days, it was radioactivity, or biotin – that would be bound by avidin. Today, the F in FISH stands for a fluorescent dye that is chemically bound to the DNA probe.

The length of the probe is significant. The longer it is, the more specific it will become. However, we will need harsher conditions for hybridizations and also it is more difficult to synthesize. Commonly, people use either 50 bases or 20 bases (20-mer) probes.

Tagging of the probe is also quite variable. The probe can be tagged at one or both ends or one or more times in the middle. The chemistry of end-labeling and middle labeling is different, and so is the synthesis process.

The most common fluorophores that people use are the cyanine dyes (known as the Cy dyes). But there are other types such as the Alexa series, Quasar dyes and more. The choice depends on required ex/em wavelengths, availability and funds. Today, many companies design and produce the probes for the researcher, but some companies use their own dye brand.

A major problem is that of signal intensity. We need multiple dye molecules bound to the same mRNA molecule in order to be able to detect it above the background. For this reason, it is recommended to have either multiple end-labeled probes (preferably >20 probes) or at least a few long multi-labeled probes (to get a minimum of twenty dye molecules or more per mRNA).

Here is a short version of the protocol (for adherent mammalian cells):

First, we need to grow the cells on a thin glass coverslip. This is important – the cells cannot be transferred from culture dish to the glass once the cells are fixed.

Next, we fix the cells (kill cells and immobilize all biomolecules) using 4% paraformaldehyde in buffer. This chemical creates multiple cross-linking bonds between all bio-molecules, thus “fixing” them in place relative to other biomolecules. This process keeps the shape of the cell and all the sub-cellular structures.

The fixation step needs to be calibrated – too much or too little can affect the hybridization efficiency. The protocol I got uses 10 min for fixation.

The cross linker is then removed, and the cells are permeabilized. There are two major protocols for permeabilization. One uses 70% ethanol at 4°C over-night. The other uses a low % solution of detergent (I use 0.1% triton X100) for a few minutes. From my brief experience, the triton works better (I’ll show images later).

The cells are then washed again, and placed in pre-hybridization buffer (2xSSC) that contains 10% formamide (for 20-mer probes) or higher (50% for 50-mer probes). The formamide denatures DNA and RNA secondary structures and assists in hybridization. SSC stands for saline-sodium citrate. The salts increase the stringency of the hybridization.

Hybridization is performed using a similar buffer, but with the addition of competitor RNA (usually tRNA) and competitor protein (BSA) to reduce background. The protocol also requires dextran sulfate (accelerates hybridization rate) and vanadyl ribonucleoside complex, a potent ribonuclease inhibitor (because we don’t want any rogue RNAses digesting the RNA in the cells).

Usually, a small drop (40-50µl) of the hyb buffer + probe is deposited on a clean surface in a humid chamber. The glass coverslip is them placed face down on the drop (i.e. cells facing the liquid).

The chamber is closed and placed at 37°C for at least 3 hours up to over-night.

We continue with several rounds of washing (which also includes an optional DAPI staining step) and finishing with mounting the coverslip onto a microscope slide using an anti-fade mounting medium.
Then, we go to the microscope.

FISH for MEF cells expressing beta-actin-MBS, with MBS 20mer probe mix labeled with Cy3. Image is maximum projection merge of three channels using ImageJ. Blue- DAPI stain; green – GFP (cells are GFP negative); red – Cy3.

It is important to have a negative control (cells not expressing the RNA in question) for further analysis.

Here’s an image of negative cells:

FISH for wild type MEFs expressing GFP but not expressing beta-actin-MBS. Probe details and image processing as the previous figure.

I will discuss spot analysis at the next post.

Sources for protocols:

Singer lab website (these are protocols for 50-mer or longer probes).

A more recent published protocol of 50-mer (for yeast)

Stellaris protocols (Biosearch technology company)  for 20mer probes. [ I do not use their mounting method, but use Prolong gold from Invitrogen. It works very well].

Protocols from Arjun Raj’s lab (includes their “Turbo FISH”  – FISH in 5 minutes. Never tried it myself, yet It works, but requires a large amount of probes compared to regular FISH (10-20 fold) which makes it very expensive).

Addition 11/2018: My bio-protocol article on smFISH in adherent animal cells.

27 responses to “FISHing RNA for beginners

  1. Pingback: Fluorescence in situ Hybridization (FISH): Protocols and Applications (Methods in Molecular Biology) · WWW.MYHEALRHOCARE.COM

  2. This Blog Post is regarded FISHing RNA This is very helpful for beginners of this field .


  3. How does one start with designing of the probes? Which areas should I select in the mRNA? Are there any websites doing that?


  4. Thanks a lot. What is better, a DNA:RNA probe or RNA:RNA probe for FISHing mRNA?


  5. Thanks for your information. The blog is very useful. While making internal modifications for a 50-mer by adding fluors, how many positions recommended?


  6. How about the washes? Can you be more detailed? . I usually lost a lot of cells during the washes (leishmania).


    • Hi Rodrigo,
      I assume these are not adherent cells. With adherent cells one usually do not loose cells. For non-adherent, I suggest bean more carful with the washes (e.g. do not use strong vaccum suction, do not touch the cell pellet etc…
      I currently can’t offer any more help, other that trying doing lots of empirical tries with different buffers and check for cell recovery after “washes”.


  7. Dear galicolagfb, the answers were useful and I got some success with your suggestions. Any idea where one can buy hybridization buffer? I have a one odd experiment and I don’t want to buy all individual components to make it up. .


  8. Dear galicolagfb,
    Thank you so much for writing a blog about RNA FISH. It is really helpful. Do yo have standard protocol that you can share which would give more information about the amounts and timing of each steps. It would be really helpful.


  9. Hi galicolagfb. How sure can one be that the FISH spots seen in the nucleus represent nuclear RNA and not DNA? In other words, which is the basis of specificity for RNA (and not for DNA) of RNA-FISH? Is there a passage during the protocol that guarantees you are getting rid of DNA, or at least that DNA will not hybridize with your probe?


  10. Hi Silvia, we do not get rid of DNA. But for DNA FISH you usually require high temperatures to melt the DNA double helix (~70C), compared to the RNA (we usually use 37C). You can easily test the specificity by comparing to a negative control with e.g. transcription inhibition.


  11. Hello galicolagfb,
    Very good explanation on the RNA-FISH.
    I am doing DNA-FISH and I was wondering if I can combine both methods on one slide or if you think that the conditions for the DNA-FISH (boiling in the microwave, 80°C heating plate for 3minutes) are too harsh and would destroy the RNA.


  12. Hi, I don`t know how to make RNA probe. could you tell me how to make it ?


    • Hi Ben
      You can use the tools from Biosearch to design Stellaris probes. You can either order from them, or use their tool and make the probes yourself. It involves ordering oligos with amine groups at the 5′ &/or 3′ ends and labeling them with the dye of your choice (e.g. Cy3, Cy5).


  13. Hi galicolagfb,
    Very interested on the RNA-FISH blog. For prepare samples for RNA-FISH, which of the methods: FFPE or Frozen section is best. Do you have any protocols for RNA-FISH with enteroids ?


  14. Hi,
    This blog helps me a lot, thx!
    I’m working on 20mers RNA FISH. When I labeled with cy3, every thing is OK, but when i changed the dye into either cy5 or Alexa488, the background becomes high and i even lost my single! I don’t know what happen. Please help me!


    • Hi,
      I’m glad you like my blog 🙂
      I think that A488 is known to have high bckgrnd in FISH and i rarely see it being used.
      Cy5 should be as good as Cy3 and I really don’t know why you get a high background.
      A few Q’s:
      1. did you try imaging the cells without the Cy5 labeling to check the backgrnd of the cells?
      2. is it specific to probes for a specific mRNA, or do you see that high background for other mRNAs (specific bckgrnd in Cy5).
      3. is this background specific to a single cell-line or do you see that in any cell?
      4. if you see high bckgrnd in Cy5 channel in all cases, maybe the problem is with the filter. I had such a problem (an area in the middle of the field was very bright, even without a sample) and had the filter replaced.



    I really like your blog.
    Maybe you can help me, we’re trying to make a protocol where we use flow cytometry instead microscope.
    I’m working with a 20nt probe label 5 ‘6-FAM in a bacteria model, the problem is that, I have two conditions one should have the mRNA target express and other should not have, but in both conditions the fluorescence is high .
    I do not know if my probes are Hybridizing my target or not.


    • Is this a single probe, or multiple 20-mer probes? multiple probes will give higher specificity.
      did you check with RT-PCR to verify that the mRNA is absent in that condition?
      do you have a negative control (gene deletion, different strain/species that lacks the seq corresponding to the probe) to make sure there’s no non-specific staining?
      did you try a control without the probe, to check for background fluorescence?
      if you did all the controls and their fine, then i would start playing with the FISH conditions (e.g. fixation time, permeabilization, % formamide, incubation time for hybridization, washes)
      good luck


  16. Pingback: Intercellular mRNA transfer through membrane nanotubes – behind the scenes. | greenfluorescentblog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s