Ashwin Gopinath

"Reproducible" Stochasticity

Cellular behavior and environment are extremely stochastic. Even within an isogenic population, every cell is unique in the way they respond to chemical signals, the precise size/shape of their organells as well as their gene expression levels. This stochasticity arises because all bio-chemical reactions, at its core, is driven by random intermolecular collisions. Over the last 15 years, two forms of stochasticity and the associated `noise' has been very well characterized in biochemical networks within prokaryotes and eukaryotes. They include intrinsic stochasticity, generated by the random timing of individual reactions. And, this is enhanced by low numbers of molecules because low numbers make individual reaction events more significant. Extrinsic stochasticity is generated by the system interacting with other stochastic systems in the cell or its environment. These studies have revealed curious properties like stochastic bursts of both mRNA and protein synthesis in many different types of organisms. Typically, in these experiments fluorescent protein reporters are used to detect fluctuations in protein concentrations while mRNA levels have been followed using fluorescently tagged mRNA-binding proteins or RNA aptamers. Various models have also been proposed to explain these different experimental results. Although, curiously stochasticity has not been experimentally explored in completely artificial biochemical networks. For instance, there is no experiment, as far as I know (June 17) that systematically attempt to understand how a protein or mRNA expression occurs from a single copy of gene in a given small volume, do these system display transcriptional or translational bursts? Or, how do simple biochemical networks, like a repressilator or toggle switch, behave in with low  copy numbers regime within defined volumes. Such model experiments can elucidate fundamental properties about stochasticity as well as potentially enable unique network behavior that functiona only within stochastic regime. However, there is no technology currently available can reliably create defined volumes with discrete countable number of interacting molecules. It is in this context that DOP and its compatibility with standard microfabrication can be invaluable.

Fig (Stochasticity) details my basic experimental setup which involves creating femto- or atto-liter chambers of well defined volumes within which a defined number of biomolecules (DNA, RNA or protein) can be loaded using DNA origami placement. Further, the chambers are going to filled with reaction mixture which could be cell free expression system or other synthetic expression system like pure Express followed by capping with a lipid bilayer to isolate the reaction volume. The figure also shows some results where I have positioned discrete number of DNA origami, each carrying a single dye molecule, within chambers of well defined volumes that is carrying Cell-Free TX-TL mixture and capped with lipid bilayer. These preliminary experiments prove the technical feasibility of setting up the experiment but the actual experiments have yet to be started. Using this basic framework I am performing two major classes of experiments,

(a) The first class of experiments aims to merely characterize transcription and translation in low copy number regime. For instance, this would involve loading each chamber with GFP and RFP expressing genes and observing the signal. I are particularly interested in characterizing variations between chambers of identical volumes, with identical number of active genes. How close are their behavior and will this value change with the volume of the chamber, which is proportional to the transcription and translation machinery. Further, will there translational bursts in these systems, like the ones that are observed in natural systems? And, it would be interesting to observe the point where the systems move from stochastic behavior to something more deterministic.

(b) The second class of experiments aim to characterize and study simple bio-chemical networks like repressilator or toggle switch behave in low-copy number regime. Even here the big questions are going to be, at the early stages, how much does the "stochasticity" depend on the volume of chamber (proportional to the transcription/translation machinery) and the number of network elements. Further, once these initial characterization are done I will move towards developing networks that perform robustly due to the stochasticity

On a more higher level this platform represents a more rigourous approach to studying stochasticity by enabling a precise initial conditions for the experiments. One of the paramountly important experiment will be studying how two set of experiments performed on different days would behave. Do they statistically have the same characteristics? I will also be attempting to improve the models of stochasticity as informed from these experiments and at the end of the day the hope is to start understanding if the insights gained here can inform how low-copy number molecules behave in natural cells. To that end, all these experiments would also need to be compared against similar experiments done on live cells. Even in this case DNA origami placement can be useful for deterministically loading live cells with defined number of molecules (Description is currently not on this website since my collaborator does not share my love for open science :) ).