NEW YORK (GenomeWeb Daily News) – Based on a global analysis of allelic expression in single cells, Ludwig Institute for Cancer Research investigators reported that which allele is expressed in a given cell is random.
As they described in a paper appearing in Science today, Rickard Sandberg and his colleagues examined 269 individual cells from embryos made by crossing two different mouse cell lines. Through transcriptome sequencing, the researchers followed how maternally versus paternally inherited alleles were expressed in those individual cells. From this, they found between 12 percent and 24 percent monoallelic expression of autosomal genes and noted that which allele is expressed varies between cells and can change over time.
"We find that for those genes that are not imprinted, roughly one in five alleles is randomly and dynamically expressed only one at a time," Sandberg said in a statement. "And if one allele is being expressed, the other doesn't know about it. There's no coordination between two."
That different alleles of the same cell are expressed fairly randomly may partially account for why organisms with similar genetic backgrounds have different traits and varying propensities for disease.
While it generally had been thought that autosomal genes are transcribed from both parental alleles, certain genes had previously been found to express only one allele, an effect called allelic exclusion. Additionally, about 1 percent of genes are imprinted, allowing for parental-specific expression, and on a broader scale, one copy of the X chromosome is inactivated in female cells. Still, the researchers said, not much was known in general about allelic expression patterns in single cells.
To look at such expression patterns, Sandberg and his colleagues isolated single cells from embryos resulting from a cross between CAST/EiJ and C57BL/6J mice during the oocyte through blastocyst stages of pre-implantation development. Using Smart-seq or Smart-seq2 — methods previously developed by Sandberg and his colleagues in conjunction with Illumina — they generated transcriptome profiles of each cell.
Using strain-specific SNPs — about 82 percent of genes expressed in this stage included at least one informative SNP — they could distinguish between maternal and paternal chromosomes.
By categorizing allelic expression as biallelic, maternal monoallelic, or paternal monoalleic, Sandberg and his colleagues reported that an average of 54 percent of genes had monoallelic expression across preimplantation development stages.
To account for the effects of RNA species loss due to the single-cell transcriptomic methods, the researchers also lysed individual cells and divvied the lysate into two volumes that were processed separately. Based on the allele calls for those pairs, the researchers modeled the stochastic losses as well as inferred the levels of biallelic and monoallelic expression in gene sets binned by expression level.
By focusing on abundant transcript — those that were likely less affected by random sampling — the researchers found an average of between 12 percent and 24 percent monoallelic expression for mRNAs. Just a few genes, typically ones with housekeeping roles, had biallelic expression, the researchers reported.
Pooling the cells eliminated signals of monoallelic expression, leading the researchers to conclude that there is cell-specific randomness in monoallelic expression.
"We therefore concluded that a dynamic type of random monoallelic expression is abundant in blastomeres," Sandberg and his colleagues wrote.
Turning beyond blastomeres, the researchers characterized the single-cell transcriptomes from liver cells from mature mice as well as adult mouse fibroblasts and found similar random monoallelic expression — about 24 percent, on average, of expressed genes in each cell.
Additionally, researchers examined both the pattern of expression as well as expression levels to find that the alleles are independently transcribed.
"We have captured a fundamental randomness at the level of gene expression that has never before been described — one that persists throughout development and into adulthood," Sandberg said.
These patterns, the researchers added, are consistent with models of transcriptional bursting in which independent transcription occurs from both alleles over time, but only one allele is transcribed at any given time.
They noted that such stochastic transcription of heterozygous alleles could contribute to phenotypic variation among cells and organisms from the same genotypic background.
"It is likely that stochastic transcription of heterozygous alleles contributes to variable expressivity — phenotypic variation among cells and individuals of identical genotypes — which may have fundamental implications for variable disease penetrance and severity," Sandberg and his colleagues said.