Transcriptional activators in the early Drosophila embryo perform different kinetic roles

Abstract

Combinatorial regulation of gene expression by transcription factors (TFs) may in part arise from kinetic synergy-wherein TFs regulate different steps in the transcription cycle. Kinetic synergy requires that TFs play distinguishable kinetic roles. Here, we used live imaging to determine the kinetic roles of three TFs that activate transcription in the Drosophila embryo-Zelda, Bicoid, and Stat92E-by introducing their binding sites into the even-skipped stripe 2 enhancer. These TFs influence different sets of kinetic parameters, and their influence can change over time. All three TFs increased the fraction of transcriptionally active nuclei; Zelda also shortened the first-passage time into transcription and regulated the interval between transcription events. Stat92E also increased the lifetimes of active transcription. Different TFs can therefore play distinct kinetic roles in activating the transcription. This has consequences for understanding the composition and flexibility of regulatory DNA sequences and the biochemical function of TFs. A record of this paper's transparent peer review process is included in the supplemental information.

Keywords: D. melanogaster embryo; MS2/MCP live imaging; diSPIM; kinetic modeling; transcription factors; transcription regulation.

Figure 1.. Measuring the activity of individual TFs against benchmark regulatory sequences.

(A) Schematics of the minimal even-skipped stripe two enhancer (eve2) transcription reporter constructs. Each contains the even-skipped promoter driving expression of 24 repeats of the MS2 stem loop sequence followed by a partial sequence of the bacterial lacZ operon. eve2:neutral contains a spacer sequence with no predicted transcription factor binding sites (dashed line). eve2:wt is eve2 with a spacer containing the wild type locus sequence between the enhancer and the promoter. The three transcription factor reporters—eve2[Zld]:neutral, eve2[Bcd]:neutral, and eve2[Dst]:neutral—are identical to eve2:neutral but contain two mutations to add predicted binding motifs for a single transcription activator (dashed box), either Zld, Bcd, or Dstat, respectively. Colored bars are transcription factor binding sites predicted by the software SiteOut . Hb, Kr, and Gt stand for Hunchback, Kruppel, and Giant, respectively. (B) Left: image of a 2D maximum projection of the diSPIM microscope field of view with histone-red fluorescent protein (magenta) and GFP-MS2 coat protein (green). Gallery images: magnified view of the marked region over time showing a detected active transcription locus. t = 0 corresponds to the beginning of nuclear cycle 14. Scale bar 10 μm. (C) Example MCP-GFP fluorescence emission record from a single nucleus during NC 14. Green marks detected active transcription signal; gray marks intervals during which no fluorescent signal was detected. (D) Dynamic transcription profiles during NC 14 for the constructs in A. Binary detection of the number of detected active transcription loci within the microscope field of view from two replicate experiments for each construct. There are 4535 loci detected across 114 active nuclei for eve2:neutral, 5265 loci from 238 nuclei for eve2:wt, 3631 loci from 289 nuclei for eve2[Zld]:neutral, 1512 loci from 113 nuclei for eve2[Bcd]:neutral, and 4535 loci from 207 nuclei for eve2[Dst]:neutral.

Figure 2.. First passage activation kinetics.

(A) Graphical depiction of a first passage into transcription measurement. Emission record as in Fig. 1C. The arrow denotes the first passage time for this nucleus. Cartoon: during this interval the transcription reporter transitions from a state incapable of activating transcription (left), through a number of rate limiting steps (dashed arrow ), to a transcriptionally competent one that activates transcription (right). The mathematical model (Eq. 1) is shown at the bottom. (B) Cumulative first passage distributions (solid curves) overlaid with a model (Eq. 1, dashed curves) with the characteristic number of rate limiting steps, $k$, a characteristic time constant for each of those steps, ${\tau }_0$, and the fraction of active nuclei within the center of the stripe, $A_f$; see Table 1 for parameter values. The curves are normalized to the total number of nuclei in the center of the stripe (see SFig. 3). There were 74 active nuclei and 166 nuclei total in the center of the stripe for eve2:neutral; 79/88 nuclei were active in the center of the strip for eve2:wt; 90/103 for eve2[Zld]:neutral; 67/91 for eve2[Bcd]:neutral; 86/93 for eve2[Dst]:neutral. Shaded regions represent the 90% confidence intervals from bootstrapping methods (see Methods).

Figure 3.. Active transcription and idle period kinetics.

(A) Graphical depiction of active transcription lifetime measurements. The double headed arrows denote two example intervals of active transcription. Cartoon: during these intervals the reporter locus transitions from a state containing many RNA polymerase molecules (gray bean) undergoing RNA synthesis (green line) to one lacking detectable active transcription (green stars). The model (Eq. 2) is shown at the bottom. (B) Cumulative lifetime distributions of active transcription (solid curves). $n=\mathrm{190,\; 507,\; 363,\; 236}$, and 444 for eve2:neutral, eve2:wt, eve2[Zld]:neutral, eve2[Bcd]:neutral, and eve2[Dst]:neutral, respectively. Data is overlaid with a model (Eq. 2; dashed curves). Shaded regions represent 90% confidence intervals from bootstrapping methods using 10,000 simulated datasets (Methods). The mean and standard error of the model parameters, time constants ${\tau }_1$ and ${\tau }_2$ and relative amplitude $A$, are to the right. These plots also show 1,000 randomly selected parameter sets from fitting Eq. 2 to simulated data from bootstrapping (points). The bi-modal nature of the eve2:neutral parameter values is due to simulated data frequently lacking a substantial long-lived population, making those distributions best characterized solely by the ${\tau }_1$ parameter (with $A=1$). (C) As in B, but the data have been partitioned into the 20% of active transcription lifetimes that first appear during NC 14 (see Methods). n = 37, 104, 75, 51, and 89 for eve2:neutral, eve2:wt, eve2[Zld]:neutral, eve2[Bcd]:neutral, and eve2[Dst]:neutral, respectively. (D–F). Idle transcription period. (D) Graphical depiction of idle transcription period measurements. The double headed arrows denote two idle periods. Cartoon: during these intervals the transcription reporter locus transitions from a state lacking active transcription to one containing many RNA polymerase molecules undergoing RNA synthesis. The model (Eq. 3) is shown at the bottom. (E) Cumulative frequency distributions of the idle periods (left). The inverse of the mean idle period, in units of s⁻¹, can be read directly from these distributions via the vertical axis intercept (inset); these, along with their standard error, are depicted on the right. See Table 3. n = 116, 428, 273, 169, and 358 for eve2:neutral, eve2:wt, eve2[Zld]:neutral, eve2[Bcd]:neutral, and eve2[Dst]:neutral, respectively. (F) As in E, but for the 20% of idle periods that first occur during NC 14.

Figure 4.. Kinetic roles of three transcriptional activators.

(A) A graphical summary of the impact of each TF on each kinetic parameter. The arrows denote if a parameter is increased or decreased by a TF. Parameter changes consistent with activation (i.e. lead to more transcription) are shown in green and changes consistent with repression in red. Beige and peach mark the first passage transcription model parameters (Eq. 1), blue marks the mean idle period (Eq. 3), and purple the active transcription model parameters (Eq. 2). The gray shaded region denotes active and idle parameters early in NC 14. (B) Model of transcription and its regulation during NC 14. Colored regions are as in A. All three TFs increase the fraction of active nuclei (arrow in beige region), and they decrease the number of steps to first passage transcription (arrow in peach region). At different times during NC 14, the TFs promote active transcription by suppressing the transition from the active to the idle state (T-bars in purple regions). Early in NC 14, Zld increases transcription by increasing the rate into the active state from the idle state, while Dst does so later (arrows in blue regions). Zld decreases this same rate later in the NC 14 (T-bar in blue region). The ellipsis represents a variable number of rate limiting steps on the pathway to first passage transcription.