Fifty years ago, Nobel laureate Ed Lewis observed that two mutant alleles of a gene, one lacking a functional regulatory region and one lacking a functional protein-coding region, can interact in trans to rescue gene expression. This process, which he called transvection, requires homologous chromosome pairing and is disrupted by chromosomal rearrangements. With only a handful of DNA elements linked to transvection and chromosome pairing since Ed Lewis’s initial discovery, many questions remain in the field. What role do these phenomena play in regulating gene expression in the wild? Does crosstalk between alleles always require chromosome-wide pairing? What DNA elements are necessary for gene copies to locate each other in the genome, and are these elements similar between genes? Our work seeks to answer these fundamental questions using stochastic gene expression in the Drosophila retina as a model.
In the retina, the gene spineless (ss) is expressed in a random on/off manner in 67% of R7 photoreceptors. A transvection-like mechanism known as Interchromosomal Communication (InterCom) allows cross-regulation of expression states between ss alleles within each R7. InterCom is one of the only known examples of transvection in a biological context, acting to average the expression frequencies of functional, naturally derived ss alleles. Using a recently described DNA FISH technique called DNA Oligopaints, we have developed a system to study ss pairing and InterCom in a quantitative manner. With this system, we have determined that InterCom requires copies of ss to pair. We have also found that, unlike classical transvection, pairing and InterCom can occur between copies of ss located in heterologous genomic locations, independent of homologous chromosome pairing. Pairing and InterCom of ss loci requires surprisingly minimal regions of homology: a ss transgene of only 110 kb can drive pairing between non-homologous chromosomes. Additionally, fragments of ss as small as 46 kb can pair and perform InterCom with endogenous ss when inserted 0.4 or 4.6 Mb away on the same chromosome arm, indicating that even small portions of the gene can mediate interactions. Our data suggest that strongly interacting loci such as ss may hold chromosomes together at specific nodes, rather than chromosome-wide homology driving tight pairing along the entire chromosome.