How transcription sets the odds of a cell fate
Some genes are expressed reliably and uniformly while others are more variable. We want to know what that variability does to development, and we study it where a single gene’s behavior decides a cell’s fate.
In the Drosophila retina, each R7 photoreceptor makes an independent, probabilistic decision to express the transcription factor Spineless. About 70% turn it on and 30% do not, and that ratio sets the proportions of two ommatidial types across the eye. We found that the ratio is encoded in the dynamics of spineless transcription (Ainsworth and Zhao et al., 2025). Replacing roughly 120 bp of core promoter at the endogenous locus, with the expression pattern left intact, shifts the adult ratio by as much as 35%. The promoter that produces longer gaps between transcriptional bursts lowers the fraction of cells that turn Spineless on, and the operative variable is the continuity of transcription.
That result reframes what a regulatory element specifies: not only where a gene is expressed, but with what continuity, and therefore with what probability a downstream decision is made. It connects to earlier work showing that reliability in the embryo is actively built, by redundant “shadow” enhancers (Perry et al., 2010; 2011) and by feedback at hunchback (Perry et al., 2012), rather than being a passive consequence of morphogen gradients.
Current directions
Mechanism at spineless. We image transcription live at the endogenous locus with MS2, combine CRISPR-edited core promoter series with single-cell readouts of nuclear position and locus compaction, and ask how burst structure is converted into a stable ON or OFF state. Quantitative analysis of the imaging data is a collaboration with Hernan Garcia (UC Berkeley).
Generality in neuroblasts. Visual system neural stem cells express a series of temporal transcription factors, each defining a window in which particular neuron types are born. Using Dichaete as a test case, we are asking whether the same dynamic rules set the duration of a competence window and the reliability of the neurons produced in it.
Human retinal differentiation. Engineered expression cassettes fall silent during the directed differentiation of human cells, a persistent problem for cell engineering and gene therapy. With Karl Wahlin (UC San Diego), we are testing whether the logic learned at spineless, continuity of transcription rather than its amount, explains why some loci silence and others do not, in iPSC-derived retinal cells. Live imaging of transcription in human cells is part of this effort.
People: Julia Ainsworth, Tim Zhao. Collaborators: Karl Wahlin (UC San Diego), Hernan Garcia (UC Berkeley).
Key papers: Ainsworth and Zhao et al., Current Biology 2025 (cell fate ratios are encoded by transcriptional dynamics); Perry et al., Current Biology 2010; Perry et al., PNAS 2011 (shadow enhancers); Perry et al., Current Biology 2012 (hunchback boundary); Agarwal and Dash et al., npj Regenerative Medicine 2023 (human retinal ganglion cells, with the Wahlin lab).
Funding: NIH National Eye Institute EY027016 (R00); Hellman Fellows Program.