Jeffrey Perkel has been a scientific writer and editor since 2000. He holds a PhD in Cell and Molecular Biology from the University of Pennsylvania, and did postdoctoral work at the University of Pennsylvania and at Harvard Medical School.
Want to know which genes a cell is expressing? You have several options.
The most popular—qPCR, DNA microarrays and RNA-Seq—are biochemical, what Marc Beal, director of corporate development and licensing at Biosearch Technologies, calls “grind and find” methods. “You extract all the guts out of the cells and try to rescue that RNA that was originally localized in the cell.”
These methods provide a good overall view of gene-expression patterns. But what they cannot provide is spatial information—where specific RNAs were located. Nor, for that matter, can they help identify cell-to-cell variation in expression levels, as that information is lost in the averaging of the cell population under study.
Enter in situ hybridization (ISH), a microscopic method that addresses both shortcomings. ISH uses labeled nucleic-acid probes to detail the spatial location and abundance of DNA or RNA in fixed cultured cells and tissue sections. The method comes in two flavors, fluorescent (FISH) and colorimetric (CISH), the latter underlying the transcriptional profiling data compiled by the Allen Brain Atlas project.
In their traditional incarnations, FISH and CISH are qualitative: Cells are either positive or negative for a given nucleic acid. But single-molecule quantitative versions have also been developed; these provide absolute transcript abundance data by literally counting transcripts one by one, one cell at a time.
A view of the flu
Timothee Lionnet, a project scientist in the Transcription Imaging Consortium at the Howard Hughes Medical Institute’s Janelia Farm Research Campus in Ashburn, Va., recently used single-molecule FISH (smFISH) to study influenza-virus RNA packaging in cells.
The influenza-virus genome comprises eight RNA molecules, and their trafficking, says Lionnet, is complicated. “From the moment the particle enters, how do these eight strands travel to the nucleus to get replicated and rearrange [in the cytoplasm] to form functional particles? To answer that, we needed spatial information.”
In smFISH, RNAs are probed not with a single molecule but with a collection of short 20-mer oligonucleotides, each labeled with a single fluorophore. These probes hybridize along the length of the RNA, producing an aggregate signal (without amplification) bright enough to detect above the background. (According to a Biosearch Technologies video explaining the process of smFISH using the company’s Stellaris® RNA FISH probes: “Each individual transcript is observed as a discrete spot, like a bright star in the night sky, by using standard fluorescence microscopy.” Check out the company's image gallery here.)
The process is multiplexable, meaning multiple transcripts can be observed simultaneously in the same cell by using different fluorescent dyes. Lionnet limited his multiplexing to just two probe sets at a time, labeling each viral RNA with either Cy3 or Cy5 and using signal overlap to determine whether they were colocalizing [1].
But up to five-color FISH is achievable “in a ‘vanilla’ manner,” says Arjun Raj, assistant professor of bioengineering at the University of Pennsylvania School of Engineering and Applied Science, and higher plexing is possible too. Using spectral color coding and a technique called iceFISH (intron chromosomal expression FISH), Raj simultaneously measured the gene expression and chromosomal structure of 20 genes on human chromosome 19 [2]. CalTech chemist Long Cai used a more sophisticated barcoding strategy and super-resolution microscopy to image 32 transcripts in single yeast cells [3].
“[Cai] is the rock star for [smFISH] multiplexing,” says Michael Mancini, scientific director of the Integrated Microscopy Core at the Baylor College of Medicine.
Still, 32 simultaneous transcripts, or even 10 times that many, is a far cry from the genome-scale analyses reported using single-cell RNA-Seq studies, so depending on the number of genes you're interested in, other approaches may better suit your needs. (That said, some researchers do question the ability of single-cell RNA-Seq to accurately quantify low-abundance transcripts.) But for Lionnet, smFISH provided all the information he needed. His results suggest the flu genome segments travel to the nucleus en masse, but leave individually and come together to form complete genomes later [1].
smFISH requirements
According to Robert Singer, the Harold and Muriel Block Chair of Anatomy and Structural Biology at the Albert Einstein College of Medicine and the first to implement smFISH, the technique has relatively few requirements. The first, of course, is a good set of probes. Singer’s lab used to make and QC its own, but now the lab buys Stellaris probe sets from Biosearch Technologies (for whom Singer is a paid consultant).
According to Raj (who also consults for Biosearch), a probe set can contain 30 to 50 probes per transcript, but “the sweet spot is probably around 32 20-mers per RNA.”
An alternative smFISH strategy uses “branched DNA technology” for signal amplification. The new method dramatically improves signal strength over the original nonamplified method by building a branching oligonucleotide arbor at the site of probe hybridization, each of which can contain huge numbers of probe “leaves.”
According to one recent report, the branched method produced spots 100-fold brighter than the older method, with “at least 2-3 times higher” signal-to-noise ratios [4]. The technique is also more amenable to automation, such that the researchers were able to analyze nearly 1,000 distinct transcripts in parallel. Branched DNA FISH probes are available from Affymetrix and Advanced Cell Diagnostics (ACD). (According to Stephen Oldfield, senior market development manager at Life Technologies, the company will begin distributing ACD’s smFISH reagents starting in November.)
Another requirement of successful smFISH is good reagents. Mancini recommends using electron microscopy-grade formaldehyde for fixation, for instance. In fact, use the best-quality reagents you can, he says. “Make sure you’re using good disposables, buy high-quality water—all of that is absolutely required,” he says.
Hybridization typically takes from four hours to overnight, but Raj developed a new technique called Turbo FISH that works far faster [5]. By switching from formaldehyde fixation to methanol and increasing probe concentration, he reduced hybridization time to five minutes, on average, and to as little as 30 seconds. “We can go end to end, from taking cells out of the incubator to imaging RNA FISH on the scope in about five minutes, including fixation and everything,” he says.
According to Beal, this acceleration makes it possible to use FISH as a point-of-care diagnostic in doctors’ offices and surgical suites, for instance to assess tumor malignancy or diagnose a viral infection. “It now almost becomes a real-time event,” he says.
smFISH also requires a good microscope, typically one with 60x or 100x oil-immersion high-numerical aperture objective. Recently, though, Biosearch acquired a company called LightSpeed Genomics, which developed a technology called Synthetic Aperture Optics (SAO). Using SAO, Beal explains, researchers can obtain the resolution of 100x oil-immersion objectives using 20x non-oil-immersion objectives—an advance that both dramatically speeds imaging, because of the wider field of view, and is more compatible with automation. SAO also enables researchers to get better statistics than they might otherwise obtain by averaging transcript counts across, say, 500 cells.
According to Beal, SAO is currently in development, with beta instruments anticipated shortly.
The final item required for smFISH is analysis software. Several free applications have been developed, including Lionnet’s Airlocalize (for MatLab; available on request), StarSearch and FISH-quant (for MatLab). Commercial tools also can handle smFISH, including Molecular Devices’ MetaMorph®.
Considerations
According to Singer, when it comes to smFISH, “the biggest mistake people make is not having a control probe. If they don’t see a signal, they don’t know if it’s something they did or if the RNA isn’t there.”
Be sure to have good positive and negative controls, he says—transcripts that definitely should or should not be present (such as a housekeeping gene and a prokaryotic sequence in eukaryotic cells, respectively). “Those two are important, because they bracket your technique or your level of skill,” he says.
Start simple, adds Raj, who developed a website dedicated to smFISH, including FAQ lists and protocols. Try widefield imaging before confocal, for instance. Follow the protocols. And, if you’re going to use multiplexing, make sure you have good biological controls.
“Start with what works and expand from there,” Raj advises.
References
[1] Chou, YY, et al., “Colocalization of Different Influenza Viral RNA Segments in the Cytoplasm before Viral Budding as Shown by Single-molecule Sensitivity FISH Analysis,” PLoS Pathog, 9(5):e1003358, 2013. [PubMed]
[2] Levesque, MJ, Raj, A, “Single-chromosome transcriptional profiling reveals chromosomal gene expression regulation,” Nat Meth, 10:246-8, 2013. [PubMed]
[3] Lubeck, E, Cai, L, “Single-cell systems biology by super-resolution imaging and combinatorial labeling,” Nat Meth, 9:743-8, 2012. [PubMed]
[4] Battich, N, et al., “Image-based transcriptomics in thousands of single human cells at single-molecule resolution,” Nat Meth, 2013. DOI:10.1038/nmeth.2657. [PubMed]
[5] Shaffer SM, et al., "Turbo FISH: A method for rapid single molecule RNA FISH," PLoS ONE, 8[9]:e75120, 2013. [PubMed]
Image: Cells labeled for EGFR (yellow), TOP1 (green) and POLR2A (red) mRNAs using Stellaris RNA FISH probes. Courtesy of Biosearch Technologies Inc. (image link)