Figures and data

The MCP/MS2-mediated live imaging system enables the visualization of transcriptional activity at the single-cell level in living Drosophila embryos.
(A) Representative snapshot of MS2-based imaging, where endogenous eve is tagged with MS2 to visualize the spatial expression pattern in embryos. Regions of the embryo (ventral, dorsal, anterior and posterior) are provided. (B) A single cell in the embryo showing GFP fluorescence with intensity assumed to be proportional to the number of nascent mRNAs. (C) Schematic showing how the nascent transcripts are tagged with fluorescent proteins using the MS2/MCP system. (D) (Top:) Representative snapshots of a single nucleus producing transcripts (green puncta) over time. (Bottom:) Fluorescence dynamics over time (green line) and the inferred transcriptional state (red line). From the inferred state, we estimate the duration of active (ON) states (τON ) and the duration of inactive (OFF) states (τOFF ). The black dashed line shows the burst threshold: the level where any sudden increase from the base level is considered as a burst (see Methods). Bracket in the top of the trajectory represents the activity time: the time between the beginning of the first burst and the end of the last one.

Heterogeneous spatial expression patterns for enhancer rho and Kr.
(A,G) Gene expression driven by the rho (A) and Kr (G) enhancers which exhibit a dorsoventral and anterior-posterior gradient, respectively. The yellow box indicates, approximately, the region imaged by live imaging. (B,H) Representative fluorescence trajectories (green) and the inferred transcriptional state (red) for rho (B) and Kr (G). Fluorescence signals are normalized to their maximum. Activity time is represented by black brackets. Dashed horizontal line represents the burst threshold. (C,I) Average fluorescence intensity, (D,J) cell averaged τON , (E,K) average τOFF , and (F,L) the activity time. The position of each point represents one cell in the beginning of NC14 and its color represents the mentioned average burst property. Cells presented in (B,H) are highlighted in red in (C-F) and (I-L) respectively. Marginal plots show the mean variables calculated with a spatial binning. The error bars show the 5% and 95% quantiles, while the length of the boxes show the 25% and 75% quantiles for each bin. Black dots represent the mean value. The means are connected by lines to show the spatial trends. Marginal plots also present the color scale at their edges used in the central scatter plot.

Spatial patterns of sna gene expression for different enhancers: shadow (SE) and proximal (PE)
(A) Image of the sna expression across the embryo. The yellow box indicates, approximately, the region imaged. (B, G) Representative gene expression trajectories normalized to the maximum (green) and the inferred transcriptional state (red) for snaPE (B) and snaSE (G). Activity time is represented by black brackets. Dashed horizontal line represents the burst threshold. These cells are highlighted in red in (C-F) and (H-K). (C, H) Spatial trends of average fluorescence, (D, I) average τON , (E, J) average τOFF , and (F, L) activity time. Marginal plots show how the mean changes in different spatial bins of the expression domain. Bar length represents the 25-75% interquantile distance and error bars, the 5-95% interquantile distance.

Statistical properties of transcriptional bursting timing across different constructs.
(A) Mean duration of τON and τOFF for the four studied enhancers, each with three replicas. (B) Random variability of τON and τOFF measured by the squared coefficient of variation

Statistical properties of the transcriptional bursts with respect to the mean fluorescence for different enhancers.
We plot the trend of different signal statistics such as (A) activity time, (B) mean ON state duration, (C) mean OFF state duration, and (D) mean loading rate (the rate of signal increase) relative to the mean fluorescence. Each dot in the plots represents the mean value of a single cell obtained from its fluorescence trajectory during NC14. These trends suggest that the activity time aligns better with mean fluorescence.

Spatial patterns of endogenous eve gene expression.
(A) Expression pattern of the eve gene. The yellow box indicates, approximately, the region imaged. (B) Expression trajectories normalized to the maximum (green) and the state of the inferred gene (red). Activity time is represented by black brackets. The horizontal dashed line represents the burst threshold which is the same as used in the constructs. These cells are highlighted in red in (C-F). (C) Spatial trends of average fluorescence, (D) average τON , (E) average τOFF , and (F) the activity time in the studied embryo domain. Bar length represents the 25-75% interquantile distance and error bars, the 5-95% interquantile distance.

Properties of transcriptional burst timing for the expression of endogenous eve gene.
Plots of average burst properties with respect to single-cell averaged fluorescence level (A) activity time. (B) Mean τON , (C) Mean τOFF and (D) Mean loading rate. Each dot in these plots corresponds to a cell and the statistics are calculated over the NC14 time-span. Error bars represent the binned mean trend with 95% confidence. We also present the statistics of the burst timing estimated using all bursts and all cells for 6 different biological replicas: (E) mean burst duration ⟨τON ⟩, (F) variability on burst duration