Robust Synthetic Circuits for Two-Dimensional Control of Gene Expression in Yeast

Abstract

Cellular phenotypes are the result of complex interactions between many components. Understanding and predicting the system level properties of the resulting networks requires the development of perturbation tools that can simultaneously and independently modulate multiple cellular variables. Here, we develop synthetic modules that use different arrangements of two transcriptional regulators to achieve either concurrent and independent control of the expression of two genes, or decoupled control of the mean and variance of a single gene. These modules constitute powerful tools to probe the quantitative attributes of network wiring and function.

Keywords: control; mating pathway; noise; orthogonal; sic1.

Figure 1

Simultaneous and independent control of gene expression using two chimeric transcriptional regulators (TRs) in parallel. (a) Scheme of two chimeric TRs constitutively expressed under the control of a crippled ADH1 promoter (pADH1(cr)). The TRs (GEM and ZPM) are a fusion of a DNA binding domain (Gal4p-DBD or Zif268-DBD, respectively), a human hormone receptor lipid binding domain (from the estradiol or progesterone receptors, hER-LBD and hPR-LBD) and a transcriptional activation domain (Msn2p-AD). (b) Representation of the action of the TRs GEM (left) and ZPM (right) in the presence of estradiol (E) or progesterone (P), respectively. (c) Diagram of the experimental setup. Two strains were cocultured, both expressing the GEM and ZPM constructs but alternatively containing the transcriptional reporters pGAL1-YFP (top) or pZ-YFP (bottom). A red-fluorescent protein was used to label the pGAL1-YFP-containing strain (top). Fluorescence, volume and cell counts were measured by automated flow cytometry. (d) Time-dependent distributions of pGAL1-YFP (red) and pZ-YFP (blue) fluorescence for different combinations of logarithmically spaced doses of estradiol and progesterone. Drug addition time is marked with black arrows. Fluorescence values are volume-corrected. (e) Steady-state dose response of the pGAL1-YFP reporter as a function of estradiol for different progesterone concentrations (shades of blue). (f) Steady-state dose response of the pZ-YFP reporter as a function of progesterone for different doses of estradiol (shades of red). In (e) and (f) pGAL1-YFP and pZ-YFP expression rates are averaged over progesterone or estradiol concentrations, respectively. The expression rates were also averaged from 3 to 12.3 h after induction.

Figure 2

A 2D-parallel circuit controlling the expression of positive and negative regulators modulates the transcriptional output of the yeast mating pathway. (a) Mating pathway signaling cascade regulators relevant to this study. A positive regulator (Ste4p) is under the control of ZPM and three negative regulators (Gpa1p, Msg5p, Dig1p) are under the control of GEM. The activity of the pathway is measured with a transcriptional reporter consisting on the AGA1 promoter driving the expression of YFP (pAGA1-YFP). (b–d) Upper panels: Schemes of the strains used in each experiment. Middle panels: Steady-state dose response of pAGA1-YFP as a function of estradiol modulating the expression of Gpa1p (b), Msg5p (c) or Dig1p (d). Different doses of progesterone (shades of blue) generate different levels of Ste4p. Fluorescence was measured after 4 h of induction and was volume-corrected. Lower panels: Mean steady-state growth rate as a function of estradiol and progesterone (shades of blue). Insets: same data as in line graphs, shown as a heat map.

Figure 3

Two TRs connected in series act as a noise rheostat circuit that decouples mean and variance. (a) Schematic of the noise rheostat. An estradiol-responsive transcriptional regulator (GEM) is used to dial the abundance of a progesterone-responsive regulator (ZPM), which in turns dials the expression of yellow- and red-fluorescent proteins (YFP and mCherry, respectively) under the control of pZ. (b) Mean and CV² of pZ-YFP distribution for different combinations of estradiol and progesterone. (c) Steady-state pZ-YFP distributions centered at (i) 750, (ii) 1500, (iii) 3000 and (iv) 6000 arbitrary fluorescence units. For each one of these cases, the same mean of the distribution but different variance around the mean could be achieved using different combinations of estradiol and progesterone (inset). (d) Time-dependent pZ-YFP distributions for conditions in (b) that generate the minimum (upper) and maximum (lower) spread over the mean. (e) Coefficient of variation of the intrinsic (green), extrinsic (red) and total (blue) noise as a function mean fluorescence, Only conditions for which the fluorescence is greater than autoffuorescence are plotted.

Figure 4

Noise rheostat circuit control of cell cycle regulator SIC1 shows effects of variability in gene expression at the same mean. (a) Cell cycle regulator SIC1 expression under the control of the noise rheostat circuit. (b) Bright field microscopy of cells harboring a SIC1 copy driven by the noise rheostat with no or high concentration of estradiol and progesterone. Size bars: 10 μm. (c) Coefficient of variation as a function of mean of the single-cell YFP fluorescence distribution for different combinations of estradiol and progesterone in a strain containing the noise rheostat regulating the expression of SIC1 (left) and a control strain containing the noise rheostat without SIC1 construct (right). Cytometry cell counts (normalized to noninduced control) are represented in colors of dots. Data for SIC1 noise rheostat experiment represent mean of three replicates. Dotted lines show increasing cell counts with increasing CV² for data points with similar mean fluorescence.