Modulation of Phase Shift between Wnt and Notch Signaling Oscillations Controls Mesoderm Segmentation

Abstract

How signaling dynamics encode information is a central question in biology. During vertebrate development, dynamic Notch signaling oscillations control segmentation of the presomitic mesoderm (PSM). In mouse embryos, this molecular clock comprises signaling oscillations of several pathways, i.e., Notch, Wnt, and FGF signaling. Here, we directly address the role of the relative timing between Wnt and Notch signaling oscillations during PSM patterning. To this end, we developed a new experimental strategy using microfluidics-based entrainment that enables specific control of the rhythm of segmentation clock oscillations. Using this approach, we find that Wnt and Notch signaling are coupled at the level of their oscillation dynamics. Furthermore, we provide functional evidence that the oscillation phase shift between Wnt and Notch signaling is critical for PSM segmentation. Our work hence reveals that dynamic signaling, i.e., the relative timing between oscillatory signals, encodes essential information during multicellular development.

Keywords: Notch; Wnt; entrainment; mesoderm segmentation; oscillations; phase shift; presomitic mesoderm; relative timing; signaling dynamics.

Graphical abstract

Figure S1

Real-Time Visualization of Wnt Signaling Oscillations during Mesodermal Patterning, Related to Figure 1 (A) Schematic depiction of Axin2T2A reporter strategy. Note that Axin2 coding sequences and endogenous untranslated regions (UTR) remain unaffected. Exons coding for untranslated regions are indicated in blue, exons corresponding to the protein coding sequence are shown as gray boxes. (B) Quantification of Axin2T2A oscillation periods in in vivo experiments and ex vivo cell culture assays reveals no significant difference (p = 0.38). Error bars indicate SEM. (C) In situ hybridization analysis against Axin2 transcripts reveals comparable expression patterns and transcript abundances in control embryos and embryos homozygous for the targeted Axin2T2A allele. In addition, Venus expression in Axin2T2A embryos corresponds to Axin2 expression patterns. Scale bar, 500 μm (whole mount embryos); scale bar, 200 μm (PSM). (D) Time series of real-time imaging experiments showing ex vivo cultures treated with the Wnt agonist CHIR99021 (Chiron) or the Wnt antagonist IWP-2 (n = 9/9 for Chiron and n = 5/5 for IWP-2). (E) Fluorescence intensity kymograph of Chiron-treated sample shown in (D) illustrates an increase in Axin2T2A fluorescence (left panel). Fluorescence intensity kymograph of IWP-2-treated sample shown in (D) reveals persistent decrease of Axin2T2A signal (right panel). Fluorescence intensity is color-coded.

Figure 1

Real-Time Visualization of Wnt Signaling Oscillations during Mesodermal Patterning (A) Snapshot of a representative Axin2T2A in vivo real-time imaging experiment (also shown in Movie S1). Dorsal view of the posterior PSM and recently formed somites. White arrow indicates stripe of Axin2T2A expression in the posterior half of newly forming somite. Scale bar, 100μm. (B) Axin2T2A reporter signal in region of interest in (A) (ROI, red circle) is plotted over time. (C) Intensity kymograph established from real-time imaging of an Axin2T2A and Mesp2-GFP double-positive PSM. Axin2T2A fluorescence (left panel), Mesp2-GFP fluorescence (middle panel), and merge of the two channels (right panel) are shown. After registration of real-time movie, kymograph was generated by averaging signal intensity over width of a 70 μm wide line. Note that in newly forming segments, Axin2T2A expression alternates with Mesp2-GFP expression. (D) Snapshot of a representative ex vivo Axin2T2A reporter cell culture assay (also shown in Movie S2). White arrowhead indicates stripe of expression in forming segment. Scale bar, 100μm. (E) Red box in (D) is magnified (green) and overlaid with the corresponding brightfield image (gray) illustrating Axin2T2A reporter expression stripe in the posterior half of newly forming segment. Red line highlights formed segment. Scale bar, 100μm. (F) Axin2T2A reporter signal in ROI (red circle in [D]) is plotted over time. Note long-lasting Axin2T2A reporter oscillations. (G) Intensity kymograph established from real-time imaging of an Axin2T2A and Mesp2-GFP double-positive ex vivo culture. Axin2T2A fluorescence (left panel), Mesp2-GFP fluorescence (middle panel), and merge of the two channels (right panel) are shown. To generate kymograph, signal intensity was averaged over width of a 70 μm wide line.

Figure 2

Quantification of Wnt and Notch Signaling Reporter Oscillations Reveals Changing Phase Shift from Posterior to Anterior mPSM (A and B) Fluorescence intensity kymographs of ex vivo cell culture assays using Wnt signaling reporter Axin2T2A (A) or Notch signaling reporter LuVeLu (B) (the same kymographs and brightfield images of the cultures are shown in Figure S2). Fluorescence intensity is color-coded independently for each kymograph. (C) Quantification of posterior oscillation periods in Axin2T2A or LuVeLu reveals no significant difference (p = 0.2). (D) Inverse velocity was quantified based on phase kymographs. LuVeLu waves were significantly slower than Axin2T2A waves (p = 0.0016). For comparison, quantification of the fourth wave during the segmentation cycle of the 2D ex vivo culture (Lauschke et al., 2013) is depicted (see Figure S2G for further analysis). Error bars in (C) and (D) indicate SEM. (E) Quantification (detrending, normalization) of Axin2T2A-Luci (magenta) and LuVeLu (cyan) reporter activity in ex vivo mPSM assay at beginning of culture. Note that only posterior PSM cells are used to generate ex vivo cultures so that oscillations detected at beginning of cultivation period represent exclusively posterior PSM cells (Lauschke et al., 2013). (Reporter signal was quantified in region of interest [yellow circle] shown in Figure S2H.) (F) Quantification of detrended, normalized Axin2T2A-Luci and LuVeLu reporter signal in anterior region of ex vivo assay using Axin2T2A-Luci (magenta) and LuVeLu (cyan) double-positive samples. (Oscillations were obtained by determining reporter signal along the line in the kymograph of anterior mPSM in Figure S2I.) This reveals that preceding the abrupt decrease in Axin2T2A wave velocity, which marks future segment boundaries (see Figure 1C and 1G for onset of Mesp2), Axin2T2A and LuVeLu oscillate in phase.

Figure S2

Phase Shift between Wnt and Notch Signaling Oscillations Differs between Posterior and Anterior mPSM, Related to Figure 2 (A-G) Comparison of Wnt and Notch signaling reporter oscillations. Ex vivo cell culture assays using Wnt reporter Axin2T2A (A-C, boxed in red) or Notch reporter LuVeLu (D-F, boxed in blue) were analyzed. Dashed red arrows in A and D depict lines along which the fluorescence intensity kymographs (B,E) were generated (the same kymographs are also shown in Figures 2A and 2B). Fluorescence intensity is color-coded individually for both kymographs. Oscillation phases were calculated and plotted in phase kymographs (C,F). Oscillation phase values from –π rad to +π rad are color-coded. Note the differences in oscillation kinetics between the biphasic Axin2T2A (C) and the LuVeLu (F) profile. (G) Based on LuVeLu and Axin2T2A phase kymographs, respectively, inverse velocity of signaling waves was measured and plotted against corresponding oscillation number (waves were numbered relative to the onset of segment formation in 2D ex vivo assay (Lauschke et al., 2013)). Error bars indicate s.e.m, ^∗∗p < 0.01, ^∗∗∗p < 0.001. (H) In posterior PSM cells Axin2T2A-Luci and LuVeLu oscillate out-of-phase. Representative brightfield image of ex vivo mPSM culture using Axin2T2A-Luci/LuVeLu double-positive embryo at beginning of culture. Note that only posterior PSM cells are used to generate ex vivo cultures, so that oscillations detected at beginning of cultivation period represent exclusively posterior PSM cells (Lauschke et al., 2013). Scale bar, 100 μm. Quantification of reporter activity in region of interest (yellow circle) is shown in Figure 2E. (I) Detrended, normalized intensity kymograph of simultaneous imaging of anterior region of ex vivo assay using Axin2T2A-Luci (magenta) and LuVeLu (cyan) double-positive samples. Quantification of reporter oscillations along dashed line is shown in Figure 2F.

Figure 3

Microfluidic System Enables Entrainment of Signaling Oscillations to a Periodic External Force (A) Schematic representation of the microfluidic setup consisting of a PDMS chip, perfused using several pumps, enabling on-chip mPSM ex vivo cultures combined with simultaneous real-time imaging. (B) General chip design showing one of the two cultivation chambers present on a microfluidic chip (top view). The depth of the chip is approximately 300 μm. Inset depicts brightfield image of mPSM culture within microfluidic chip (scale bar of inset, 100 μm). (C) Scheme of experimental setup. Periodic pulses of signaling pathway modulator were applied to mPSM ex vivo cultures, and endogenous signaling pathway oscillations were detected. Dashed lines in representative fluorescence intensity kymographs of LuVeLu and Axin2T2A mPSM cultures depict the region corresponding to anterior mPSM, in which oscillations were measured for further analysis. (D–H) Entrainment of LuVeLu oscillations to periodic pulses of the Notch inhibitor DAPT (2μM): (D and F) Quantification of (detrended, normalized) LuVeLu signal in anterior mPSM reveals oscillations in control (D) and DAPT-treated samples (F). Experiments (N = independent experiments, n = individual samples) were combined using the external force (gray bars) as an objective time reference. (E and G) Phase-phase plots of the phase relation between endogenous rhythm (LuVeLu) and external periodic force (control: DMSO pulses [E], treatment: DAPT pulses [G]) reveals stable phase relationship for entrained DAPT samples. Density of points within the plots is color-coded. (H) Mean period of oscillations in (D) and (F) (see full timeseries data in Data S1A) was quantified. Error bars indicate SEM.

Figure S3

Entrainment of the Segmentation Clock to Pulses of 5 μM Chiron, Related to Figures 3 and 4 (A–D) Entrainment of Axin2T2A oscillations to periodic pulses of the canonical Wnt signaling agonist CHIR99021 (Chiron, 5 μM): (A and C) Quantification of (detrended, normalized) Axin2T2A signal in anterior mPSM reveals oscillations in control (A) and Chiron-treated samples (C). Experiments (N = independent experiments, n = individual samples) were combined using the external force (gray bars) as an objective time reference. (B and D) Phase-phase plots of the phase-relation between endogenous rhythm (Axin2T2A) and external periodic force (control: DMSO pulses (B), treatment: Chiron pulses (D)). Density of points within the plots is color-coded. (E–H) Entrainment of LuVeLu oscillations to periodic pulses of the Wnt signaling activator Chiron (5 μM): (E and G) Quantification of (detrended, normalized) LuVeLu signal in anterior mPSM reveals oscillations in control (E) and Chiron-treated samples (G). Experiments (N = independent experiments, n = individual samples) were combined using the external force (gray bars) as an objective time reference. (F and H) Phase-phase plots of the phase-relation between endogenous rhythm (LuVeLu) and external periodic force (control: DMSO pulses (F), treatment: Chiron pulses (H)). Density of points within the plots is color-coded. (I) Mean reporter activity (black line) and s.d. (gray shading) of Axin2T2A oscillations shown in (C) and LuVeLu oscillations shown in (G) reveal in-phase oscillations in anterior mPSM upon entrainment with Chiron pulses (N = independent experiments, n = individual samples). (J) Representative Axin2T2A fluorescence intensity kymographs of DMSO control- (left panel) and Chiron-treated mPSM ex vivo cultures (right panel). Fluorescence intensity is color-coded.

Figure S4

Single-Drug Pulses of Either DAPT or Chiron Do Not Significantly Alter Absolute Intensity or Mean Amplitude of LuVeLu and Axin2T2A Oscillations in Anterior mPSM, Related to Figures 3, 4, and S3 Quantification of LuVeLu and Axin2T2A signals in anterior mPSM of ex vivo assays cultured with pulses of 2 μM DAPT (A-D) or 5 μM Chiron (E-H). Mean intensity of LuVeLu (A,E) or Axin2T2A (C,G) reporter signal and mean amplitude of LuVeLu (B,F) or Axin2T2A (D,H) reporter oscillations were determined. DAPT or Chiron-treated samples, respectively, were quantified relative to DMSO-treated samples. Quantifications are based on time points 50 to 120 to exclude beginning of ex vivo culture, when anterior mPSM region has not re-established yet (Lauschke et al., 2013). Error bars denote SEM.

Figure 4

Wnt Signaling Oscillations Are Linked to Notch Signaling Oscillations (A) Scheme of experimental setup. Periodic pulses of Notch signaling inhibitor DAPT (2 μM) were applied to mPSM ex vivo cultures and endogenous Axin2T2A oscillations were detected. (B–E) Entrainment of Axin2T2A oscillations to periodic pulses of the Notch inhibitor DAPT (2μM): (B and D) Quantification of (detrended, normalized) Axin2T2A signal in anterior mPSM reveals oscillations in control (B) and DAPT-treated samples (D). Experiments (N = independent experiments, n = individual samples) were combined using the external force (gray bars) as an objective time reference. (C and E) Phase-phase plots of the phase relation between endogenous rhythm (Axin2T2A) and external periodic force (control: DMSO pulses [C], treatment: DAPT pulses [E]). Density of points within the plots is color-coded. (F) Mean reporter activity (black line) and SD (gray shading) of Axin2T2A oscillations shown in (D) and LuVeLu oscillations shown in Figure 3F reveal in-phase oscillations in anterior mPSM upon entrainment with DAPT pulses (N = independent experiments, n = individual samples). (G) Representative Axin2T2A fluorescence intensity kymographs of DMSO control- (left panel) and DAPT-treated mPSM ex vivo cultures (right panel). Fluorescence intensity is color-coded. (See Data S1B for full timeseries data.)

Figure 5

Experimental Modulation of Phase Shift between Wnt and Notch Signaling Oscillations Using Microfluidics (A–H) Quantification (detrending, normalization) of LuVeLu (A and E) or Axin2T2A (B and F) signals in mPSM samples cultured either with alternating (A–C) or simultaneous pulses (E–G) of 2 μM DAPT and 5 μM Chiron. All measurements were done in anterior mPSM (see Figure 3C). Individual experiments were aligned to each other using external perturbations as objective time reference. Black line depicts mean, gray shading depicts SD of combined samples (N = independent experiments, n = individual samples). (C and G) Phase-phase plots of LuVeLu versus Axin2T2A oscillations upon alternating (C) or simultaneous DAPT/Chiron pulses (G). Density of points is color-coded. (D and H) Schematic representation of phase relationship between LuVeLu and Axin2T2A oscillations in anterior mPSM upon alternating (D) or simultaneous (H) DAPT/Chiron pulses. (See Data S2 for full timeseries of treatment with both simultaneous and alternating drug pulses.)

Figure S5

Alternating or Simultaneous Pulses of DAPT and Chiron Do Not Significantly Change Absolute Intensity or Mean Amplitude of LuVeLu and Axin2T2A Oscillations in Anterior mPSM, Related to Figure 5 (A–H) Quantification of LuVeLu and Axin2T2A signals in anterior mPSM of ex vivo assays cultured with alternating (A–D) or simultaneous pulses (E–H) of 2 μM DAPT and 5 μM Chiron. Mean intensity of LuVeLu (A and E) or Axin2T2A (C and G) reporter signal and mean amplitude of LuVeLu (B and F) or Axin2T2A (D and H) reporter oscillations were determined. DAPT/Chiron-treated samples were quantified relative to DMSO-treated samples. Quantifications are based on time points 50 to 120 to exclude beginning of ex vivo culture, when anterior mPSM region has not re-established yet (Lauschke et al., 2013). Error bars denote SEM.

Figure 6

Modulation of Phase Shift between Wnt and Notch Signaling Oscillations Delays Oscillation Arrest and Impairs Proper mPSM Segmentation (A) Representative fluorescence intensity kymographs of ex vivo cultures using the Notch reporter LuVeLu are depicted for samples treated either with DMSO pulses (left panel), alternating (middle panel, “alt.”), or simultaneous pulses (right panel, “sim.”) of 2 μM DAPT and 5 μM Chiron. Fluorescence intensity is color-coded. White arrows mark time point of oscillation arrest. (B) Quantification of LuVeLu fluorescence intensity in a particular region in anterior mPSM (white lines in [A]). Individual samples were normalized to the fluorescence intensity of the first oscillation peak within each time-series. Asterisks indicate additional cycles of oscillations in the sample treated with simultaneous pulses of DAPT/Chiron. (C) Quantification of time until oscillation arrest in experiments with either alternating (left panel) or simultaneous external force pulses (right panel). N = independent experiments, n = individual samples. Error bars denote SEM. (D) Representative brightfield images of ex vivo cultures showing segment formation in samples treated either with DMSO (left panel) or alternating DAPT/Chiron pulses (middle panel). In contrast, physical boundary formation was absent in samples entrained with simultaneous DAPT/Chiron pulses (right panel). (E and F) Quantification of ex vivo cultures forming physical boundaries in experiments treated with alternating (E) or simultaneous (F) DAPT/Chiron pulses compared to DMSO control. Error bars denote SEM.

Figure S6

Effect of Anti-phase Wnt and Notch Signaling Oscillations on Segmentation, Related to Figure 6 (A and B) Molecular analysis of segmentation marker expression upon altered Wnt/Notch phase shift. (A) Ex vivo cultures were treated with either alternating or simultaneous pulses of DAPT and Chiron. Samples were fixed after 18 h of culture and subjected to in situ hybridization using probes against the indicated mRNAs. Representative images are shown. (B) Percentage of ex vivo cultures expressing Mesp2, Ripply2 and Uncx4.1 was quantified. (C and D) Quantification of Mesp2-GFP reporter activity in ex vivo mPSM cultures either in the presence or absence of the transcriptional repressor Hes7. (C) Representative fluorescence intensity kymographs of samples generated from Hes7-expressing (Hes7^+/+ or Hes7^+/−, left panel) or Hes7-knockout embryos (Hes7^−/−, right panel). (D) Quantification of time until onset of Mesp2-GFP expression is depicted. (E and F) Quantification of Mesp2-GFP reporter activity in ex vivo mPSM cultures treated with alternating pulses of DMSO or pulses of 2 μM DAPT and 5 μM Chiron. (E) Representative fluorescence intensity kymographs for DMSO- (left panel) or DAPT/Chiron-treated samples (right panel). (F) Quantification of time until onset of Mesp2-GFP expression is shown. (G and H) Quantification of Mesp2-GFP reporter activity in ex vivo mPSM cultures treated with either alternating or simultaneous pulses of 2 μM DAPT and 5 μM Chiron. (G) Representative fluorescence intensity kymographs for samples treated with alternating (left panel) or simultaneous DAPT/Chiron pulses (right panel). (H) Quantification of time until onset of Mesp2-GFP expression is shown. (I–L) Induction of Axin2T2A expression in posterior segment halves in anterior mPSM is prevented by anti-phase Wnt/Notch oscillations. Representative fluorescence intensity kymographs of ex vivo cultures treated either with DMSO (I) or alternating DAPT and Chiron pulses (J) are shown. In contrast, the anterior Axin2 expression domain was absent in samples entrained with simultaneous pulses of DAPT and Chiron (K). (L) Quantification of ex vivo cultures showing Axin2 stripe in anterior mPSM. Fluorescence intensity in kymographs is color-coded. Error bars denote SEM.