NEW YORK (GenomeWeb News) – Using a revamped fluorescent in situ hybridization-based approach, researchers from the University of Pennsylvania were able to detect and quantify single-nucleotide variants in RNA transcripts in both populations of cells and in single cells. As they reported in Nature Methods yesterday, they also used their technique to distinguish between transcripts from maternally and paternally inherited chromosomes.
By combining a toe-hold probe strategy with guide probes, Arjun Raj, an assistant professor of bioengineering at UPenn, and his colleagues were able to detect SNVs with specificity and with limited false positives. Other methods to detect SNVs in situ, the researchers said, have a tendency to be complex and have a low efficiency.
"Development of such a method with general applicability would be of great utility for studying genetics and gene regulation, particularly for measuring allele-specific gene expression in single cells and single molecules," Raj and his colleagues noted.
RNA FISH, they added, is also susceptible to mismatches in which the 20-base oligonucleotide probes align with RNA despite one-nucleotide differences. To find an SNV, the researchers said, the probes would have to be more stringent.
To increase the selectivity of the probes in their FISH screen, the researchers hybridized a 28-base single-stranded DNA SNV detection probe to a smaller mask oligonucleotide, making a probe that is partially double-stranded. The single-stranded region of the detection probe binds to the RNA target, and the mask oligonucleotide — which is bound to the other part of the detection probe — dissociates, allowing the rest of the detection probe to bind to the RNA target.
But relying on just one probe also opens up the possibility of false positive signals, the researchers added. To avoid that issue, Raj and his colleagues used multiple oligonucleotide probes —dubbed 'guide' probes — that bind to the target RNA, which they said increased the robustness of the identification.
To test their method, the researchers designed a probe targeting a known mutation in the BRAF oncogene as well as a probe aimed at the wild-type gene. By testing these probes on a series of human melanoma cell lines known to be either homozygous or heterozygous for the T to A mutation, they determined the overall co-localization efficiency to be in the range of 65 percent.
The researchers found that they could detect imbalances in the abundance levels of maternal and paternal transcript levels in the GM12878 melanoma cell line using this approach. For example, at the cell population level, they noted that the EBF1 and SUZ12 genes had more mRNA derived from the paternal chromosome, while the DNMT1 gene had no such transcript imbalance.
Interestingly, when they focused on DNMT1 transcript expression in single cells from the GM12878 line, they found that a small portion of the cells deviated greatly from the population-level average, while few cells ventured far from the population average for EBF1 and SUZ12 transcript expression.
Raj and his group previously developed probes that target various introns on chromosome 19, producing what they called an RNA-based chromosome 'paint.' In this study, they applied that method with a twist: They developed detection probes for the introns of the genes with SNVs in the GM12878 line, painting 15 introns from the paternal chromosome a certain 'color.'
"In this manner, we visualized and classified chromosomes as maternal or paternal in situ," the investigators said. "These results demonstrate that our method is applicable to introns, enabling us to measure allele- specific transcriptional activity directly."
Further, they noted that being able to home in on certain chromosomes would allow them to determine if a new SNV is found on a maternal or paternal chromosome, or from which chromosome a transcript with no SNV is derived.
"Aside from diagnostic applications, particularly in genotyping single cells in situ, our method has the potential to provide insights into allele-specific effects in gene expression," Raj and his colleagues added.