“So you say, what the heck is going on here? Here you have a gene that's absolutely required for the cell to survive and some cells have none of them, and some cells are just loaded,” Singer says. But that difference in abundance can lead to tremendous differences in a cell's protein output, enabling a considerable level of robustness.
Take the genes in yeast that mitigate copper toxicity, for example. Although yeast are generally unlikely to encounter toxic levels of the metal, if small numbers of cells in a population carry these stress-response genes, at least a small fraction of a population can survive a toxic insult. “The cell has evolved a way to keep a lot of its baggage, a lot of its DNA, under wraps,” Singer says.Extending FISH
Such variability could also underlie the variability in disease phenotypes seen in certain genetic disorders, says Arjun Raj, an assistant professor of Bioengineering at the University of Pennsylvania.
As a graduate student in Sanjay Tyagi 's lab at the Public Health Research Institute at the University of Medicine and Dentistry of New Jersey, Raj modified Singer's FISH method. Instead of using a small number of multiply labeled oligos, he performed his RNA-FISH experiments with as many as 52 distinct, singly labeled oligos. He published the updated procedure in 2008. Two years later, as a postdoc with Alexander van Oudenaarden at MIT, Raj applied the method to C. elegans, demonstrating that the stochastic nature of transcription can explain the “incomplete penetrance” of mutant genes— that is, why different individuals with the same mutation display different phenotypes. In effect, Raj explains, it's as if for some individuals, the cell rolled the dice and came up snake-eyes.
More recently, Raj modified FISH yet again to yield intron chromosomal expression FISH (iceFISH), a method combining the single-cell resolution of RNA FISH with the spatial information inherent in microscopy. The goal was to determine if chromosome structure somehow influences gene expression. By simultaneously lighting up each of 20 genes on human chromosome 19 in fibroblasts and HeLa cells, Raj and his graduate student Marshall Levesque were able to count how often each gene was on or off in each cell. Importantly, they could also see the spatial distribution of each chromosome and trace the chromosomal locations of each gene. As a result, they were able to ask whether chromosomal translocations influence gene expression patterns by directly comparing a translocated chromosome with its normal partner.
“What we found is that for the 13;19 [translocation], those genes on the translocated portion of chromosome 19 were much more frequently transcribed than the same gene on the normal copy of chromosome 19,” Raj explains.
Long Cai, Assistant Professor of Chemistry at the California Institute of Technology, has also extended the original FISH methodology through multiplexing. In 2012, Cai showed that the technique worked with 32 genes per cell, and theoretically, he says, “high-hundred” multiplexing is possible. “The overarching goal … is to turn single cells into microarrays,” Cai said in a 2012 BioTechniques news story (1). To do that, he is combining superresolution microscopy and fluorescent barcoding.
Typically, the smallest point visible using FISH is a diffraction-limited spot, about 300–400 nm in diameter. While that's far larger than a single RNA molecule, researchers could potentially zoom in further to see more detail. “If you assume a cell is something like 10 microns cubed, and if your resolution is 10 nanometers cubed, then effectively you have something like a billion pixels in one cell,” Cai says. In theory, this is more than enough to spatially resolve the million or so mRNAs present in any single cell.
Working with his graduate student, Eric Lubeck, Cai combined 7 fluorescent colors into a 3-color barcode to study 32 yeast genes controlled by the transcription factor Crz1 in each of 62 cells, using STORM superresolution microscopy. The results again underscore the random nature of gene expression. “Even though these genes are controlled by a single transcription factor, they're not turned on at the same time.”Looking at live cells
Even though FISH has proven a popular technique for single-molecule analyses, it has a significant drawback: it only works in dead, fixed cells. “What's missing in FISH is the time dimension,” explains Singer.