• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Dev Cell. Author manuscript; available in PMC Nov 13, 2013.
Published in final edited form as:
Dev Cell. Nov 13, 2012; 23(5): 995–1005.
doi:  10.1016/j.devcel.2012.09.009
PMCID: PMC3500528
NIHMSID: NIHMS417414

Single cell resolution imaging of the impact of Notch signaling and mitosis on segmentation clock dynamics

Summary

Vertebrate body segmentation is controlled by the segmentation clock, a molecular oscillator involving transcriptional oscillations of cyclic genes in presomitic mesoderm cells. The rapid and highly dynamic nature of this oscillating system has proved challenging for study at the single cell level. We achieved visualization of clock activity with a cellular level of resolution in living embryos, allowing direct comparison of oscillations in neighbor cells. We provide direct evidence that presomitic mesoderm cells oscillate asynchronously in zebrafish Notch pathway mutants. By tracking oscillations in mitotic cells, we reveal that a robust cell-autonomous, Notch-independent mechanism resumes oscillations after mitosis. Finally, we find that cells preferentially divide at a certain oscillation phase, likely reducing the noise generated by cell division in cell synchrony and suggesting an intriguing relationship between the mitotic cycle and clock oscillation.

Introduction

In vertebrates, the metameric vertebrae and axial muscles are derived from repeated mesodermal segments, called somites, which form from the presomitic mesoderm (PSM) during embryogenesis. As a result of gastrulation and tail elongation, the PSM progressively extends by entry of new cells in its caudal part. At the same time, somites are sequentially pinched off from its anterior part and deposited along the anterior-posterior axis. A striking feature of PSM segmentation is its spatial and temporal periodicity. Somitogenesis is controlled by a molecular oscillator, called the segmentation clock, that cycles within the PSM with the same period as somite formation (Oates et al., 2012; Pourquié, 2011). According to the ‘Clock and Wavefront’ model and its modern variations, somite periodicity and total somite number are determined by the clock interacting with a positional signal called the wavefront. The position of the wavefront is set by global gradients across the anterior-posterior axis of the embryo and moves posteriorly as the tailbud extends. Future somite boundaries become specified in the PSM when a group of neighboring cells in permissive phase of the clock encounters the wavefront (Cooke and Zeeman, 1976; Oates et al., 2012; Pourquié, 2011).

Much progress has been made for understanding the molecular nature of the segmentation clock. The activity of a molecular oscillator in the PSM was revealed by the striking discovery of the first “cyclic” gene cHairy1, a member of the hairy and Enhancer-of-split related family (Palmeirim et al., 1997). Since then, it has been established that other vertebrate Hes/her genes also cycle, along with additional genes of the Notch, FGF and Wnt signaling pathways (Oates et al., 2012; Pourquié, 2011). Typically, the expression of cyclic genes in fixed embryos presents as stripes in the anterior PSM and homogenous staining, of variable intensity from one embryo to another, in the posterior PSM. Careful analysis of multiple fixed embryos and, more recently, real-time live imaging, have revealed that waves of cyclic gene expression originate in the posterior PSM with the same pace as somite formation and move anteriorly across the PSM (Aulehla et al., 2008; Masamizu et al., 2006; Oates et al., 2012; Takashima et al., 2011). The expression pattern of cyclic genes within individual PSM cells over time has been inferred from these observations, assuming that very little cell movement takes place once cells internalize into the PSM. First, cells entering the posterior PSM oscillate with the same period as somite formation, and do so in synchrony with their neighbors. Then, the oscillations slow down as the cells reach the anterior PSM, and stop upon somite formation. The slowing of oscillations creates small delays between cells and results in stripes of cyclic gene expression in the anterior PSM of fixed embryos. Thus, the clock is composed of a multitude of elementary oscillators, the PSM cells, which are finely coordinated with each other.

Oscillation dynamics and coordination appear to be controlled by a complex genetic and biochemical network, though its exact nature remains unknown and likely varies from one organism to another (Oates et al., 2012; Pourquié, 2011). A conserved feature in vertebrates is the involvement of cycling Hes/her transcriptional repressors (Krol et al., 2011). It has been suggested that a negative feedback loop created by Hes/Her proteins down-regulating their own transcription generates alternating oscillations of proteins and transcripts, constituting a core cell-autonomous mechanism essential for clock oscillations (Bessho et al., 2003; Hirata et al., 2002; Giudicelli et al., 2007; Lewis, 2003; Takashima et al., 2011). Notch signaling is also largely implicated in the clock regulation, as Hes/her genes are Notch targets and somites are disrupted when Notch signaling is impaired (Oates et al., 2012; Pourquié, 2011). Because Notch receptors are activated in one cell by ligands of the Delta/Jagged/Serrate family on adjacent cells, Notch was proposed to be essential for coupling PSM cell oscillations, although its precise role as an initiator or synchronizer of the clock is still debated (Holley, 2007; Lewis et al., 2009; Oates et al., 2012; Pourquié, 2011). In zebrafish, embryos with impaired Notch pathway display defects of somite boundary formation and “salt and pepper” expression of cyclic genes, although the first anterior-most somites do form normally (Holley, 2007). These defects have been interpreted as evidence that PSM cells cycle but progressively fall out of synchrony in absence of Notch signaling, suggesting a jmrole of Notch signaling for synchronizing oscillations in PSM cells (Horikawa et al., 2006; Jiang et al., 2000; Mara et al., 2007; Ozbudak and Lewis, 2008; Riedel-Kruse et al., 2007).

Many insights into segmentation clock regulation have been described using transcript detection in fixed embryos (Giudicelli et al., 2007; Horikawa et al., 2006; Mara et al., 2007; Soroldoni and Oates, 2011). Because the oscillation period is short, ranging from about 30 minutes in zebrafish to 2 hours in mouse, the real-time reporters required to investigate clock dynamics have proved an extreme technical challenge (Soroldoni and Oates, 2011). The current mouse reporter strategies allow visualization in real-time with a tissue level of resolution (Aulehla et al., 2008; Masamizu et al., 2006; Takashima et al., 2011). Imaging the clock in vivo at a single cell level of resolution is crucial for understanding how the clock activity is precisely related to cell oscillations and how oscillations are coordinated between PSM cells. Here, we present a real-time reporter of zebrafish segmentation clock dynamics and a semi-automated 3-D cell tracking and analysis program that allowed us to image clock dynamics with single cell resolution in the developing PSM. We describe how PSM cells oscillate over time in wild-type embryos. We show that Notch pathway mutant cells oscillate, yet are largely out of phase with neighboring cells, providing direct evidence for the role of Notch signaling in clock synchronization. Strikingly, we show that after mitosis, sibling cells oscillate in tight synchrony in wild-type and Notch pathway mutant embryos, highlighting the cell autonomous and Notch-independent nature of segmentation clock oscillation resumption after mitosis. Finally, we show that mitosis, a source of biological noise in this system (Horikawa et al., 2006; Zhang et al., 2008), occurs most frequently during the “off phase” of the Her1 oscillation wave, suggesting that regulation of mitosis and clock expression are mechanistically linked.

Results

Live imaging of the segmentation clock activity with single cell resolution

To investigate the dynamic mechanism of the clock, we developed tools for measuring oscillations in individual PSM cells in living zebrafish embryos. To develop a dynamic reporter compatible with the short periodicity of zebrafish segmentation, we fused the fast-folding Yellow Fluorescent Protein, Venus(Nagai et al., 2002), to the C-terminus of the Her1 protein, anticipating that destabilization sequences within the Her1 protein would similarly destabilize the fusion protein (Hirata et al., 2002; Hirata et al., 2004) (Figures 1A, S1A). We used a previously characterized 8.6 kb upstream her1 regulatory region (Gajewski et al.,2003) to drive cyclic expression of Her1-Venus fusion protein, and 1.1 kb of downstream sequence (including the her1 3′UTR and her1 polyadenylation siteto mimic endogenous her1 transcript dynamics as closely as possible (Chen et al., 2005, Ozbudak and Lewis, 2008)) to generate a stable Tg(her1:her1-Venus) bk15 line (Figure S1A). In situ hybridization and immunochemistry in heterozygous transgenic embryos revealed that reporter transcript and protein are cyclically expressed (Figure 1B–D) and oscillate out of phase with each other (Figure 1E), consistent with the simple negative feedback loop proposed for Her cyclic regulation (Giudicelli et al., 2007; Lewis, 2003). As expected for a transcription factor, the Her1-Venus reporter protein localizes to the nucleus. Live timelapse confocal imaging (Figure 1F; Movie S1) reveals that waves of cyclic expression emanate posteriorly and travel anteriorly, ceasing upon reaching the forming somite boundary.

Figure 1
The zebrafish transgenic her1:her1-Venus line recapitulates dynamic her1 expression

Expression of other tested reporters was either too stable or non-detectable (Figure S1). Fusion of Venus to the N-terminus of Her1 protein produced a reporter protein that persisted in the newly formed somites even though the reporter was no longer transcribed, suggesting that Her1 N-terminus is essential for instability (Figure S1B, E, E′). A Venus-PEST protein identical to the reporter used by Aulehla et al. (2008) for real-time imaging of mouse segmentation clock was also too stable for imaging zebrafish clock oscillations (Figure S1C, F, F′). On the other hand, although appending ubiquitin moieties to the N-terminus of Venus, in a strategy similar to the one used by Masamizu et al. (2006) for imaging the mouse clock, produced a reporter protein with a striped pattern comparable to that of her1 expression in fixed embryos (Figure S1D, G, G′), we were unable to detect reporter signal by confocal microscopy (data not shown). Compared to other constructs, the Her1-Venus reporter thus constituted an ideal combination of stability and signal intensity.

Because Hes/Her proteins are thought to negatively regulate their own expression (Brend and Holley, 2009; Giudicelli et al., 2007; Hirata et al., 2002; Lewis, 2003), we tested whether expression of the Her1-Venus fusion protein influenced the endogenous clock. Expression of her7 cycles similarly in wild-type and heterozygous Tg(her1:her1-Venus) embryos, although her7 stripes are slightly more diffuse in the latter (Figure 1G–G′, J–J′). Expression of deltaC (dlc), another cyclic gene, and mespa, a clock readout, are indistinguishable between wild-type and heterozygous Tg(her1:her1-Venus) embryos (Figure 1H–I, K–L), except that the angle between PSM stripes and notochord is broader in transgenic embryos. The slight differences in gene expression do not impact somite periodicity, which is the same in wild-type and heterozygous Tg(her1:her1-Venus) embryos (Figure 1M).

Homozygosity of the reporter transgene does appear to have some impact on the clock. Somite periodicity is slowed (Figure 1M) and total segment number is decreased (Figure 1N,O), consistent with a direct relationship between segmentation speed and somite number. In addition, her7 gene expression is noticeably dampened in the anterior (but not posterior) PSM (Figure S2A, A′, D, D′). Despite these differences, dlc and mespa expression are almost normal in homozygous Tg(her1:her1-Venus) embryos (Figure S2B–C, E–F). To minimize any impact on endogenous clock function, we performed all subsequent analyses on heterozygous Tg(her1:her1-Venus) embryos.

To analyze reporter expression in individual cells, we injected transgenic embryos with membrane-mCherry and histone H2A-cerulean encoding mRNAs to serve as membrane and nuclear landmarks, respectively (Figure 2A,B), and imaged embryos for 4–6 hours beginning at the 8–12 somite stage by confocal microscopy (Movie S2, top left corner). Z-stacks of about 30 images were acquired every 4 minutes. To efficiently process and analyze the large volumes of imaging data, we developed a MATLAB program to automatically track individual presomitic cells. Individual cell contours were predicted across three dimensions based on shape and fluorescence patterns (Keller et al., 2008) and linked across time points (Sbalzarini and Kounoutsakos, 2005) (Figure 2C). The reporter fluorescence of each cell was then quantified based on intensity within the predicted nuclear contour (Figure 2D). Because of the high nucleus to cytoplasm ratio of PSM cells, some errors in the prediction of cell contours occurred, especially along the z-axis for which spatial resolution is lower. We thus created a graphical user interface to manually validate and, if needed, correct each cell. It also allowed us to label tracked cells with specific properties, such as mitotic cells, for use in subsequent analyses (Figure S3). Cells were only validated if it was possible to track them throughout the entire movie. The phase at each time point of each “validated cell” oscillation was calculated using a smoothing heuristic (see Methods and Figure 2E,F) and used in subsequent analyses.

Figure 2
Detection and analysis of clock oscillations at single cell resolution

Last oscillation occurs in S-1 and lasts about twice the period of somite formation

We tracked and validated fluorescence intensity of 50–100 PSM cells per timelapse movie for 3 embryos that were wild type apart from the presence of the reporter transgene. The somitogenesis period was lengthened by lowering the temperature to 22–23°C (Schröter et al., 2008), which allowed us to obtain enough time for z-stack image acquisition between consecutive time points and to potentially increase reporter lifetime. Because each embryo was imaged separately and temperature might vary slightly among imaging experiments, we did not compare cells between movies or calculate exact periods; instead, we made observations and comparisons within a given embryo.

Most of the robust oscillations we detected occurred in the S-3 to S0 region of the PSM (encompassing 4 anterior-most presumptive somites, Figure 3A). Although weaker in intensity, robust oscillations were also detected in a number of cells located as far as S-6. Our analyses reveal that cells oscillate in the posterior PSM with a period equivalent to that of somite formation (Figure 3A′) as expected from analysis of cyclic gene expression patterns in fixed embryos (Giudicelli et al., 2007;Sawada et al., 2000). The clock period lengthens during the second to last oscillation, at the level of S-2, coinciding with expression of the first markers of anterior-posterior somite patterning (Sawada et al., 2000). During the last oscillation, the signal peaks in S-1, and decreases in S0, with a pseudo-period at the peak that is almost twice the period of somite formation (Figure 3A″). The higher fluorescence intensity in anterior relative to posterior PSM cells possibly results from stronger expression, reduced degradation rate, and/or mechanisms increasing the length of the clock pseudo-period in the anterior PSM and giving the reporter more time for maturation and accumulation.

Figure 3
Notch signaling is required for synchronous oscillation in neighboring PSM cells

PSM cells oscillate in Notch pathway mutants and do so asynchronously

When Notch signaling is disrupted in zebrafish, in situ hybridization analyses reveal that cyclic genes are expressed in the PSM in a salt-and-pepper pattern instead of clear stripes (Holley, 2007); the interpretation is that PSM cells still oscillate, but asynchronously, in the absence of Notch signaling (Jiang et al., 2000; Lewis, 2003; Mara and Holley, 2007). To directly address whether Notch signaling is required to maintain synchronous clock expression among neighboring cells, we crossed the segmentation clock reporter into the beamter (bea/deltaC), deadly seven (des/notch1a), and after eight (aei/deltaD) mutant backgrounds. We confirmed that the Venus reporter is expressed in a speckled pattern in fixed embryos, similar to expression of cyclic genes in Notch pathway mutants (Figure S4). Timelapse analysis and cell tracking revealed that cells do oscillate in the absence of Notch signaling (Figure 3B–D; Movie S2).

By pseudo-coloring PSM cells using a color map (Figure 3E) indicating phase of oscillation, we obtained snapshots of cell oscillation dynamics relative to position (Figure 3F–I) at any given time point. Notch pathway mutant embryos lack the smooth transitions indicative of neighbor cell synchrony normally observed in wild-type embryos (Figure 3E–I and Movie S3). To analyze synchrony on a global scale, we compared the phase of Venus expression for each cell relative to its direct neighbors. For this, we automatically sorted all possible pairs of validated cells separated by less than 10 microns (about 1 cell diameter) and calculated the phase difference (“phase shift”) between cells for each pair of direct neighbors. By computing the phase shift for all validated cell pairs at all time points, we obtained a total of 2,935–19,180 comparisons per embryo, all of which were plotted onto a histogram (Figures 3J–3M). In wild-type embryos (Figure 3J), there is a strong bias towards little or no phase shift (phase shift close to zero), with very few neighbors being in anti-phase (phase shift close to π); a small proportion of desynchronized (anti-phase) cells are expected, for example at segment borders in the anterior PSM. By contrast, Notch pathway mutants were desynchronized (Figures 3K–M). Together, these data directly demonstrate that PSM cells cycle asynchronously in the absence of Notch signaling, providing critical support for the role of Notch signaling in the maintenance of neighbor cell oscillation synchrony.

Most dividing cells undergo temporary disruptions in oscillation synchrony

Synchrony maintenance in a group of molecular oscillators is proposed to be essential to counteract biological noise. Mitosis has been proposed to be a significant source of noise in oscillating systems (Horikawa et al., 2006; Zhang et al., 2008). Using our segmentation clock reporter, we examined oscillations in sibling cells after mitosis, relative to each other and their neighbors in real time (Figure 4A,B). During division, some cells maintain surprisingly synchronous oscillations with neighbors throughout mitosis (Figure 4A,C; 20% of mitotic events). However, most cells become delayed relative to neighbors following mitosis (Figure 4B,C; 60% of mitotic events), or temporarily display erratic oscillations or no cycling (Figure 4C; 20% of mitotic events), entirely consistent with the idea of mitosis-induced noise. Measuring synchrony of cells with their neighbors at different times after mitosis shows that the proportion of daughter cells oscillating in synchrony with their environment increases over time, as anticipated from the existence of mechanism for synchrony maintenance. The large majority of recently divided cells resynchronize with their respective neighbors within 2 oscillation cycles (Figures 4C and S5), consistent with previous estimates (Horikawa et al., 2006).

Figure 4
Mitosis produces highly synchronized sibling cells that gradually resynchronize with neighbors

After mitosis, sibling cells oscillate in tight synchrony with each other in wild-type and Notch mutant embryos

Although synchrony between a dividing cell and its neighbors can be variable, we noticed that siblings are strikingly synchronous over time (Figure 4A,B). By collectively examining mitotic events, comparing the phase of a recently divided cell at every time point post-division with either the phase of its sibling or its neighbors (and between non-dividing neighbors as controls), we observed a clear difference between sibling-sibling synchrony and sibling-neighbor synchrony (Figure 4D). This global phase shift analysis confirms that siblings are significantly more synchronized with each other than their neighbors (two-tailed t-test, α=0.05, p<10−5). To analyze if sibling cells might be synchronized by signals received from their shared neighbors, we followed sibling oscillations in Notch pathway mutants, where cell divisions occur in a largely asynchronous background (Figure 3). Like in wild-type embryos (Fig. 4D), Notch pathway mutant sibling cells are significantly more synchronized with each other than with their neighbors in all genotypes tested (two-tailed t-test, α=0.05, p<10−5) (Fig. 5A–F). Thus although blocking mitosis in Notch pathway mutants can delay the onset of global asynchrony (Zhang et al., 2008), mitotic events in Notch pathway mutants occurring once global asynchrony has occurred actually generate a pair of tightly synchronous cells. Comparing sibling oscillation phases at different times after mitosis revealed that most siblings remain highly synchronous with each other over one full oscillation and at least the beginning of a second oscillation in all genotypes (Figures S6, 4A,B, 5A,C,E). These findings suggest that clock components equally segregated to sibling cells during mitosis are sufficient to govern timing of at least two subsequent protein oscillations and highlight that a robust Notch-independent, cell-autonomous mechanism drives clock oscillations in the PSM regions we analyzed (S-III to S0).

Figure 5
After mitosis, sibling cells oscillate in tight synchrony in a Notch-independent manner

Noise induced by cell division is likely reduced because mitosis preferentially occurs during the off-phase of the oscillation wave

Another striking aspect of oscillations in dividing cells emerged when we examined the oscillation phase of mitotic cells and neighbors upon cytokinesis. As anticipated, mitosis disturbs cyclic expression, and clock reporter levels are at their lowest levels in the large majority of dividing cells at cytokinesis in wild-type and Notch pathway mutant embryos (Figure 6A, left, and not shown). Importantly, we discovered in wild-type embryos, where the phase of each dividing cell could be compared to the global collective phase of its neighbors, that in the majority of cases, not only were sibling cells in the trough of an oscillation at cytokinesis, but so were many of their neighbors (Figure 6A, right). This observation suggests that mitosis tends to occur during the “off phase” of the Her1 oscillation wave at the level of the entire tissue. An hour post-mitosis, the phase differences between siblings and their neighbors were generally smaller for cells for which division had occurred at the trough of the oscillation cycle, relative to mitosis at other phases of the cycle (Figure 6B). These data suggest an intriguing relationship between clock oscillation and mitosis that may serve to limit mitosis-induced noise.

Figure 6
Mitosis preferentially occurs during the off-phase of the Her1 oscillation wave

Discussion

Essential tools for analyzing PSM cell oscillations

For many years, the dynamics of the segmentation clock has been deduced from expression patterns of cyclic genes in fixed embryos. As cyclic gene expression patterns varied between different embryos with identical somite number, large collections of embryos were required to estimate oscillations dynamics. More recently, high resolution in situ hybridization provided further insight into the clock dynamics, by allowing discrimination between cells actively transcribing cyclic genes (with nuclear nascent transcripts) and cells more advanced in the oscillation cycle (with cytoplasmic mature transcripts). However, comparison of oscillation dynamics in neighbor cells, or in cells after clock perturbation, was still limited. The real-time reporters of the clock that were recently developed beautifully revealed the propagation dynamics of the cyclic gene expression wave across the PSM, but did not allow analysis of the clock at the single cell level. Reaching such a level of resolution is crucial for understanding the mechanism of the segmentation clock, as PSM cells constitute its elementary oscillators. In this paper, we introduce two tools, essential for analysis of oscillations in individual PSM cells in vivo.

First, we developed a highly dynamic reporter of the clock, the nuclear localization of which largely facilitated the detection of oscillations at a cellular level of resolution. Among the various reporters we generated, only Her1-Venus displayed instability compatible with the very short period of the zebrafish segmentation clock while maintaining detectable levels of expression in a transgenic line using confocal microscopy. A good reporter should have minimal impact on the oscillations. Because overexpression of her1 causes somite defects (Takke et al., 1999), the presence of the entire Her1 sequence in the Her1-Venus reporter could be an issue. However, heterozygous zebrafish embryos from the Tg(her1:her1-Venus) line showed no effect of the reporter on somite formation and very little impact on cyclic gene expression. Importantly, the Her1-Venus reporter displayed clear expression differences between wild-type and Notch pathway mutant backgrounds, validating its use for analyzing the impact of Notch signaling on the segmentation clock. Although the last 3 to 5 oscillations undergone by PSM cells before somite formation were efficiently revealed by the Her1-Venus reporter, analysis of cyclic gene expression in the very posterior regions, including “progenitor zone”, “initiation zone” and posterior-most PSM (Mara et al., 2007), will require further development of clock reporters and imaging techniques. Second, we generated a semi-automated program which performs 3D segmentation of confocal images into individual cells, tracking of cell positions across time, measurement of reporter nuclear signal and computing of oscillation phase at any given time point and for any given cell. Using these tools in zebrafish allowed us to compare oscillations in neighbor PSM cells in living embryos. With the rapid progress of in vivo imaging techniques, similar approaches will undoubtedly soon be possible in the mouse.

Oscillations are controlled by a robust cell autonomous mechanism

Dissociated mouse PSM cells display autonomous oscillations that are asynchronous and irregular, suggesting the presence of an unstable oscillator within PSM cells (Maroto et al., 2005; Masamizu et al., 2006). We show that after mitosis, sibling cells are strikingly synchronized with each other, usually for initiation of at least 2 cycles, yet most are delayed relative to their direct neighbors. Although newly generated sibling cells progressively resynchronize with their neighbors in wild-type embryos, the average synchrony remains higher for sibling cells than for random neighbor cells. It was recently shown that cytoplasmic bridges persist between sibling cells for several hours after mitosis in epiblast cells of zebrafish gastrula (Caneparo et al., 2011). These intercellular bridges were not detected in the hypoblast at gastrula stage (Caneparo et al., 2011), nor have we detected them in PSM cells scatter-labeled with membrane tdTomato (data not shown). Thus, mitosis in the PSM likely generates sibling cells that are physically independent from each other and nevertheless cycle in tight synchrony. This suggests that the biochemical material inherited by sibling cells is sufficient for precise timing of oscillation start for at least 2 cycles. Thus, the cell autonomous mechanism generating oscillations in zebrafish appears more robust than previously anticipated from mouse PSM cell dissociation experiments.

In zebrafish, Her1 and Her7 have been proposed to play an essential cell-autonomous role for generating oscillations. Mathematical modeling reveals that a mechanism of repression of her1 and her7 genes by their own products, involving transcriptional and translational delays, could generate transcript oscillations, alternating with protein oscillations (Lewis, 2003). Although our primary goal was to develop a reporter with minimal impact on the endogenous clock as in heterozygous tg(her1:her1-Venus) embryos, we noticed that embryos homozygous for the her1:her-Venus transgene displayed longer segmentation period and fewer somites than wild-type embryos. This observation is consistent with previous work showing that the total number of somites is controlled by a balance between the speed of segmentation and the rate of PSM elongation and wavefront regression (Gomez et al., 2008, Schröter and Oates, 2010). It is not clear if transgene homozygosity impacts the clock because Her1-Venus fusion protein interferes with normal Her1 function or because Her1 activity is too high. At high doses, Her1-Venus may disrupt Hes6 function, a protein shown to heterodimerize with Her1 and to control segmentation speed and somite number in zebrafish (Kawamura et al., 2005, Sieger et al., 2006, Schröter and Oates, 2010, Schröter et al., 2012). Altogether, these observations support a role of the Her/Hes machinery as a pacemaker of the zebrafish segmentation clock. Whether the robust cell autonomous mechanism generating synchronous oscillations in sibling cells relies mainly on the her/hes negative feedback loop or involves additional complexity remains to be understood.

Role of Notch signaling

Although a role for Notch signaling synchronizing the oscillations has been proposed in zebrafish for over a decade (Horikawa et al., 2006; Jiang et al., 2000; Lewis, 2003; Mara et al., 2007; Ozbudak and Lewis, 2008; Riedel-Kruse et al., 2007), it is possible that “salt and pepper” expression of cyclic genes in Notch pathway mutants results from stochastic and/or stable expression in a subset of cells, rather than asynchronous cycling. Here we provide direct evidence that, as anticipated, cells oscillate out of synchrony in the intermediate and anterior PSM of Notch pathway mutants.

Because aei/deltaD mutants display no cyclic gene expression in the posterior PSM, DeltaD was proposed to be involved in oscillation initiation (Mara et al., 2007). We found that after mitosis, sibling cells in the PSM resumed oscillations with similar delay and synchrony in wild-type and aei/deltaD, bea/deltaC and des/notch1a mutant embryos, suggesting that oscillations in the intermediate and anterior PSM are generated independently of Notch signaling. Next-generation reporters will be required for analyzing the importance of the different Notch pathway components for oscillation initiation in the posterior PSM.

Impact of mitosis on cell synchrony

Using time lapse microscopy, Horikawa and colleagues measured that 10–15% of cells undergo mitosis during one cycle of oscillation in the posterior PSM, and that the M phase, during which transcription is largely switched off, lasts at least half the period of a cycle (Horikawa et al., 2006). This suggested that mitosis could be a significant source of noise for oscillation synchrony. The disruption of synchrony between neighbor cells in Notch pathway mutants seems slightly less severe when the cell cycle is disrupted in emi1 (early mitotic inhibitor 1) mutants (Zhang et al., 2008), consistent with mitosis creating noise in the system. Indeed, we measured in Tg(her1:her1-Venus) embryos that 80% of the cells undergoing division are impacted by this event, most of them being delayed relative to their neighbors. In the remaining 20% of cases for which oscillations were unaffected by mitosis, the off phase of her1 transcript oscillation may have coincided with the general transcriptional depression caused by mitosis. In mice, two groups of cyclic genes (one enriched with genes of the HES family and of Notch and FGF signaling pathways, the other one enriched with genes of the Wnt signaling pathway) oscillate in opposite phase (Dequéant et al., 2006; Krol et al., 2011). Although no such groups of genes cycling in phase opposition were found in zebrafish, the off phase of her1 transcript may not coincide with off phase of other important cyclic transcripts (Krol et al., 2011). However, zebrafish her genes broadly oscillate in phase with each other (Krol et al., 2011; Oates and Ho, 2002) and likely represent crucial genes for oscillation genesis. Thus, linking mitosis to her genes oscillation dynamics could help reduce the impact of mitosis on the clock. Strikingly, we observed that an unexpected high number of cell divisions occur during the “off phase” of the Her1 oscillation wave, and generate siblings that are on average less desynchronized with their environment. This suggests an intriguing hypothesis that mitosis and the clock are linked in such a way that cell division creates less noise than previously thought.

The in vivo reporter we describe, and “next-generation” versions that will undoubtedly be made, open many new doors of opportunity for understanding somitogenesis. We can now study clock attributes in single cells across space and time, which will continue to reveal a deeper understanding of this dynamic process.

Experimental Procedures

Fish stocks

Adult fish strains (AB wild type, beab663 [Henry et al., 2005], desb638 [Gray et al., 2001] aeitr233 [van Eeden et al., 1996]) were kept at 28.5°C on a 14-hour light/10-hour dark cycle. Embryos were obtained by natural crosses or in vitro fertilization and staged as previously described (Kimmel et al., 1995).

Plasmid construction and transgenesis

The her1:her1-Venus plasmid was assembled using the following sequences. The 8.6 kb PstI-NcoIher1 upstream region was isolated from Construct I (Gajewski et al., 2003). her1 coding sequence was amplified from pCS2+her1 plasmid (Takke et al., 1999) using 5′-ACCTGCCAGCCATGGTTACTCCAAAAATG-3′ forward and 5′-GCTAGCAGTCGACCCTCCACTACCTCCCCAGGGTCTCCACAAAGG-3′ reverse primers, Venus coding sequence from Venus/pCS2 plasmid (Nagai et al., 2002) using 5′-GCTAGCGGTGGAATGGTGAGCAAGGGCGAGGA-3′ forward and 5′-CTTAAGACGCGTTACTTGTACAGCTCGTCCATGCCG-3′ reverse primers, and a fragment containing 1.1 kb of her13′ non -coding sequence from AB genomic DNA using 5′-ACCCTCTTAAGCAAAACTGAAGACACTTAGCATGAGAATAACCAGCG-3′ forward and 5′-AAACAGCGGCCGCCGTCATTATTTACTCTTAAACCTGTTTGAACACC-3′ reverse primers. Fragments containing the 8.6 kb PstI-NcoIher1 upstream region, her1 coding sequence (digested with BfuAI [to create an NcoI-compatible end] and NheI), Venus coding sequence (NheI/AflII-digested), and 1.1kb her13′ noncoding sequence (AflII/NotI-digested) were inserted in that order into the pBSKI2 plasmid (Thermes et al., 2002) between PstI and NotI sites. Transgenic lines were generated as previously described using I-SceI-based transgenesis (Thermes et al., 2002). Reporter transcripts from both founders analyzed gave striped expression by in situ hybridization. The Tg(her1:her1-Venus)bk15 stable line was generated from founder m7, displays reproducibly strong oscillating expression, and transmits as a single Mendelian locus.

Wholemount in situ hybridization and immunohistochemistry

Digoxygenin-labelled anti-sense RNA probes were synthesized from the following templates: deltaC (Jiang et al., 2000), her7 (Gajewski et al., 2003), mespa (Sawada et al., 2000) and Venus (Nagai et al., 2002). In situ hybridization was performed as previously described (Thisse and Thisse, 2008). For Venus immunohistochemistry, standard protocols were followed, using 4% PFA fixation, 2% Triton-X100/PBS permeabilization, 2% BSA/2 % goat serum/1% DMSO/0.1% Tween 20/PBS blocking, anti-GFP rabbit antibody (Molecular Probes, diluted 1:1000), peroxidase-conjugated anti-rabbit goat secondary antibody (Molecular Probes, diluted 1:200), and diaminobenzidine staining. For double Venus transcript and protein staining, immunohistochemistry was performed before in situ hybridization. Head and yolk were removed in 70% ethanol and embryos were flat mounted in 80% glycerol. Images were captured using a Zeiss AxioPlan upright microscope and AxioCam camera.

Live imaging

Zebrafish embryos were injected at the one-cell stage with about 40 pg H2B-cerulean and 20 pg membrane-mCherry mRNA (Megason, 2009), raised at 28°C to 30°C until 11 hpf, then held at 23°C for several hours prior to imaging. At the 8–12 somite stage, embryos were mounted laterally, with no coverslip, in embryo arrays (Megason, 2009) in Embryo Medium plus 0.01% Tricaine. Confocal sections were performed every 1.34 μm, with stacks taken every 4 minutes, using an upright Olympus FV1000 confocal microscope, a XLUMPXFL 20x water objective (NA 0.95), and temperature controlling ring set to 23°C. Image resolution is 2.092 pixels/micron. Images were converted to 8-bit before processing.

3D cell contour detection and tracking

MATLAB script was developed for the sole purpose of automatically and accurately detecting PSM cells in our experiments. Our cell-tracking and data analysis tools have not been optimized or validated for any other purpose. After initial conversion of the membrane and nuclear channels fluorescence images into compatible MATLAB files, noise in both the membrane and nuclear channels were removed using a low-pass filter, treating for noise and non-uniform illumination. The separate optical slices were then merged into a single matrix and each continuous three-dimensional cluster of fluorescent pixels was indexed. Cell contours were predicted and connected across time and space based on previous work (Keller et al., 2008; Sbalzarini and Kounoutsakos, 2005). The potential problems in tracking due to mitosis were circumvented by tracking the cells in reverse, starting from the last time point. For each cell, the optical section with the highest average brightness was recorded as the cell fluorescence for each given time point. The cell’s reporter fluorescence, position in the embryo, Cartesian coordinates, catalog number, and mitotic activity were recorded in a data matrix for future analyses. To validate the automated trackings and efficiently make necessary edits to our data matrices, we developed a graphical user interface (GUI) that displayed images for a given time point and z-slice as well as for the previous time point at the same z-slice (Fig. S3). The interface allowed for rapid navigation across space and time, simple tagging and manual correction of 3D contour of individual cells, and filters to display different fluorescence channels, labels, and contours. This configuration allowed us to properly link cells across time frames that were previously left unconnected by the automatic analysis. We also used the GUI to manually tag cells for complex qualitative properties that cannot be easily computed automatically (mitotic activity, relative position within somites, cell type).

Fluorescence smoothing and phase calculation

Due to high variability of fluorescence with time (increasing amplitude of oscillations along the PSM, noise), simplifying assumptions on the form of the signal have to be made to smoothen the fluorescence signal and define a phase for the oscillators. We assume fluorescence F(t) behaves as a harmonic oscillator where values of amplitude A(t), basal fluorescence B(t) and angular velocity ω(t) slowly change with time (left hand side of Equation 1). Our smoothing heuristic (Equation 1) removed the average fluorescence in a given time window T comparable to the period, estimated the new amplitude over the same time window, and rescaled the sine wave to that amplitude. The raw fluorescence was treated twice by this method to isolate a pseudo sine wave for each cell. Assuming this readout behaved like a harmonic oscillator, angular velocity and consequently phase (ϕ(t)=ω(t)t, taken modulo 2π), can be simply computed (Fig. 2E,F, Fig. 3E). This heuristic is extremely simple to implement and we checked many examples to confirm that it gives robust and realistic results. Phase shift between two oscillations is the absolute value of the difference of the two computed phases, taken between 0 and π. MATLAB scripts were developed to automatize all these calculations. One obstacle with this calculation is that the period calculation assumes a period exists. An option was included in the GUI to crop out ranges of calculated periods in cells that were not visibly oscillating.

F(t)=A(t)sinω(t)t+B(t)Tsinω(t)t(F(t)-<F(t)>T)2<(F(t)-<F(t)>T)2>T
Equation 1

with

<F(t)>T=1Tt-T/2t+T/2F(u)du

and T a fixed time window of the order of the period.

Statistical Analysis

2D histograms were generated by plotting the calculated periods of cells after mitosis as a function of their sibling or their neighbor at each time point. Statistical comparisons of synchrony levels between these two groups were determined by comparing the overall phase differences using a two-sample t-test. Other details included in Figure legends.

Highlights

  • segmentation clock dynamics are imaged with cellular level of resolution
  • presomitic mesoderm cells oscillate asynchronously in the absence of Notch signaling
  • after mitosis, sibling cells resume oscillations in tight synchrony with each other
  • mitosis occurs preferentially at the trough of the oscillation phase

Supplementary Material

01

04

Acknowledgments

We thank Martin Gajewski for providing the her1 promoter plasmid, Jasmine McCammon for constructing an early version of the reporter, Olivier Pourquié for advice and providing the Venus-PEST construct, and Ryoichiro Kageyama for sharing the HES1-Ub2luc construct. We thank Sean Megason, Frederique Ruf, and Scott Fraser for imaging consultation, for sharing membrane-mCherry and H2B-cerulean constructs and for hands-on confocal training and Holly Aaron and the UC Berkeley Molecular Imaging Center for confocal access and advice. EAD thanks Laurent Schaeffer and Véronique Morel for advice. This work was funded by Association Française contre les Myopathies (EAD), a Marie-Curie Outgoing International Fellowship (EAD), a Pew Scholar Award (SLA), and an NIH grant and ARRA supplement (1-R01-GM061952) (SLA). PF is supported by Natural Science and Engineering Research Council of Canada (NSERC), Discovery Grant program, and Regroupement Québécois pour les matériaux de pointe (RQMP).

Footnotes

The authors declare no competing financial interests.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Aulehla A, Wiegraebe W, Baubet V, Wahl MB, Deng C, Taketo M, Lewandoski M, Pourquié O. A beta-catenin gradient links the clock and wavefront systems in mouse embryo segmentation. Nat Cell Biol. 2008;10:196–193. [PubMed]
  • Bessho Y, Hirata H, Masamizu Y, Kageyama R. Periodic repression by the bHLH factor Hes7 is an essential mechanism for the somite segmentation clock. Genes Dev. 2003;17:1451–1456. [PMC free article] [PubMed]
  • Bird NB, Mabee PM. Developmental morphology of the axial skeleton of the zebrafish, Danio rerio (Ostariophysi:Cyprinidae) Dev Dyn. 2003;228:337–357. [PubMed]
  • Brend T, Holley SA. Expression of the oscillating gene her1 is directly regulated by Hairy/Enhancer of Split, T-box, and Suppressor of Hairless proteins in the zebrafish segmentation clock. Dev Dyn. 2009;238:2745–2759. [PMC free article] [PubMed]
  • Caneparo L, Pantazis P, Dempsey W, Fraser SE. Intercellular bridges in vertebrate gastrulation. PLoS One. 2011;6:e20230. [PMC free article] [PubMed]
  • Chen J, Kang L, Zhang N. Negative feedback loop formed by Lunatic fringe and Hes7 controls their oscillatory expression during somitogenesis. Genesis. 2005;43:196–204. [PubMed]
  • Cooke J, Zeeman EC. A clock and wavefront model for control of the number of repeated structures during animal morphogenesis. J Theor Biol. 1976;58:455–476. [PubMed]
  • Dequéant ML, Glynn E, Gaudenz K, Wahl M, Chen J, Mushegian A, Pourquié O. A complex oscillating network of signaling genes underlies the mouse segmentation clock. Science. 2006;314:1595–1598. [PubMed]
  • Du SJ, Frenkel V, Kindschi G, Zohar Y. Visualizing normal and defective bone development in zebrafish embryos using the fluorescent chromophore calcein. Dev Biol. 2001;238:239–246. [PubMed]
  • Gajewski M, Sieger D, Alt B, Leve C, Hans S, Wolff C, Rohr KB, Tautz D. Anterior and posterior waves of cyclic her1 gene expression are differentially regulated in the presomitic mesoderm of zebrafish. Development. 2003;130:4269–4278. [PubMed]
  • Giudicelli F, Ozbudak EM, Wright GJ, Lewis J. Setting the tempo in development: an investigation of the zebrafish somite clock mechanism. PLoS Biol. 2007;5:e150. [PMC free article] [PubMed]
  • Gomez C, Ozbudak EM, Wunderlich J, Baumann D, Lewis J, Pourquie O. Control of segment number in vertebrate embryos. Nature. 2008;454:335–339. [PubMed]
  • Gray M, Moens CB, Amacher SL, Eisen JS, Beattie CE. Zebrafish deadly seven functions in neurogenesis. Dev Biol. 2001;237:306–323. [PubMed]
  • Henry CA, McNulty IM, Durst WA, Munchel SE, Amacher SL. Interactions between muscle fibers and segment boundaries in zebrafish. Dev Biol. 2005;287:346–360. [PubMed]
  • Hirata H, Bessho Y, Kokubu H, Masamizu Y, Yamada S, Lewis J, Kageyama R. Instability of Hes7 protein is crucial for the somite segmentation clock. Nat Genet. 2004;36:750–754. [PubMed]
  • Hirata H, Yoshiura S, Ohtsuka T, Bessho Y, Harada T, Yoshikawa K, Kageyama R. Oscillatory expression of the bHLH factor Hes1 regulated by a negative feedback loop. Science. 2002;298:840–843. [PubMed]
  • Holley SA. The genetics and embryology of zebrafish metamerism. Dev Dyn. 2007;236:1422–1449. [PubMed]
  • Horikawa K, Ishimatsu K, Yoshimoto E, Kondo S, Takeda H. Noise-resistant and synchronized oscillation of the segmentation clock. Nature. 2006;441:719–723. [PubMed]
  • Jiang YJ, Aerne BL, Smithers L, Haddon C, Ish-Horowicz D, Lewis J. Notch signalling and the synchronization of the somite segmentation clock. Nature. 2000;408:475–479. [PubMed]
  • Kawamura A, Koshida S, Hijikata H, Sakaguchi T, Kondoh H, Takada S. Zebrafish Hairy/Enhancer of split protein links FGF signaling to cyclic gene expression in the periodic segmentation of somites. Genes Dev. 2005;19:1156–1161. [PMC free article] [PubMed]
  • Keller PJ, Schmidt AD, Wittbrodt J, Stelzer EH. Reconstruction of zebrafish early embryonic development by scanned light sheet microscopy. Science. 2008;322:1065–1069. [PubMed]
  • Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF. Stages of embryonic development of the zebrafish. Dev Dyn. 1995;203:253–310. [PubMed]
  • Krol AJ, Roellig D, Dequéant ML, Tassy O, Glynn E, Hattem G, Mushegian A, Oates AC, Pourquié O. Evolutionary plasticity of segmentation clock networks. Development. 2011;138:2783–2792. [PMC free article] [PubMed]
  • Lewis J. Autoinhibition with transcriptional delay: A simple mechanism for the zebrafish somitogenesis oscillator. Curr Biol. 2003;13:1398–1408. [PubMed]
  • Lewis J, Hanisch A, Holder M. Notch signaling, the segmentation clock, and the patterning of vertebrate somites. J Biol. 2009;8:44. [PMC free article] [PubMed]
  • Mara A, Holley SA. Oscillators and the emergence of tissue organization during zebrafish somitogenesis. Trends Cell Biol. 2007;17:593–599. [PubMed]
  • Mara A, Schroeder J, Chalouni C, Holley SA. Priming, initiation and synchronization of the segmentation clock by deltaD and deltaC. Nat Cell Biol. 2007;9:523–530. [PubMed]
  • Maroto M, Dale JK, Dequéant ML, Petit AC, Pourquié O. Synchronised cycling gene oscillations in presomitic mesoderm cells require cell-cell contact. Int J Dev Biol. 2005;49:309–315. [PubMed]
  • Masamizu Y, Ohtsuka T, Takashima Y, Nagahara H, Takenaka Y, Yoshikawa K, Okamura H, Kageyama R. Real-time imaging of the somite segmentation clock: revelation of unstable oscillators in the individual presomitic mesoderm cells. Proc Natl Acad Sci USA. 2006;103:1313–1318. [PMC free article] [PubMed]
  • Megason SG. In toto imaging of embryogenesis with confocal time-lapse microscopy. Methods Mol Biol. 2009;546:317–332. [PMC free article] [PubMed]
  • Nagai T, Ibata K, Park ES, Kubota M, Mikoshiba K, Miyawaki A. A variant of yellow fluorescent protein with fast and efficient maturation for cell-biological applications. Nat Biotechnol. 2002;20:87–90. [PubMed]
  • Oates AC, Ho RK. Hairy/E(spl)-related (Her) genes are central components of the segmentation oscillator and display redundancy with the Delta/Notch signaling pathway in the formation of anterior segmental boundaries in the zebrafish. Development. 2002;129:2929–2946. [PubMed]
  • Oates AC, Morelli LG, Ares S. Patterning embryos with oscillations: structure, function and dynamics of the vertebrate segmentation clock. Development. 2012;139:625–639. [PubMed]
  • Ozbudak EM, Lewis J. Notch signalling synchronizes the zebrafish segmentation clock but is not needed to create somite boundaries. PLoS Genet. 2008;4:e15. [PMC free article] [PubMed]
  • Palmeirim I, Henrique D, Ish-Horowicz D, Pourquié O. Avian hairy gene expression identifies a molecular clock linked to vertebrate segmentation and somitogenesis. Cell. 1997;91:639–648. [PubMed]
  • Pourquié O. Vertebrate segmentation: from cyclic gene networks to scoliosis. Cell. 2011;145:650–663. [PMC free article] [PubMed]
  • Riedel-Kruse IH, Müller C, Oates AC. Synchrony dynamics during initiation, failure, and rescue of the segmentation clock. Science. 2007;317:1911–1915. [PubMed]
  • Sawada A, Fritz A, Jiang YJ, Yamamoto A, Yamasu K, Kuroiwa A, Saga Y, Takeda H. Zebrafish Mesp family genes, mesp-a and mesp-b are segmentally expressed in the presomitic mesoderm, and Mesp-b confers the anterior identity to the developing somites. Development. 2000;127:1691–1702. [PubMed]
  • Sbalzarini IF, Koumoutsakos P. Feature point tracking and trajectory analysis for video imaging in cell biology. J Struct Biol. 2005;151:182–195. [PubMed]
  • Schröter C, Herrgen L, Cardona A, Brouhard GJ, Feldman B, Oates AC. Dynamics of zebrafish somitogenesis. Dev Dyn. 2008;237:545–553. [PubMed]
  • Schröter C, Ares S, Morelli LG, Isakova A, Hens K, et al. Topology and dynamics of the zebrafish segmentation clock core circuit. PLoS Biol. 2012;10:e1001364. [PMC free article] [PubMed]
  • Schröter C, Oates AC. Segment number and axial identity in a segmentation clock period mutant. Curr Biol. 2010;20:1254–1258. [PubMed]
  • Sieger D, Ackermann B, Winkler C, Tautz D, Gajewski M. her1 and her13.2 are jointly required for somitic border specification along the entire axis of the fish embryo. Dev Biol. 2006;293:242–251. [PubMed]
  • Soroldoni D, Oates AC. Live transgenic reporters of the vertebrate embryo’s Segmentation Clock. Curr Opin Genet Dev. 2011;21:600–605. [PubMed]
  • Takashima Y, Ohtsuka T, González A, Miyachi H, Kageyama R. Intronic delay is essential for oscillatory expression in the segmentation clock. Proc Natl Acad Sci USA. 2011;108:3300–3305. [PMC free article] [PubMed]
  • Takke C, Campos-Ortega JA. her1, a zebrafish pair-rule like gene, acts downstream of notch signalling to control somite development. Development. 1999;126:3005–3014. [PubMed]
  • Thermes V, Grabher C, Ristoratore F, Bourrat F, Choulika A, Wittbrodt J, Joly JS. I-SceI meganuclease mediates highly efficient transgenesis in fish. Mech Dev. 2002;118:91–98. [PubMed]
  • Thisse C, Thisse B. High-resolution in situ hybridization to whole-mount zebrafish embryos. Nat Prot. 2008;3:59–69. [PubMed]
  • van Eeden FJ, Granato M, Schach U, Brand M, Furutani-Seiki M, Haffter P, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, et al. Mutations affecting somite formation and patterning in the zebrafish, Danio rerio. Development. 1996;123:153–164. [PubMed]
  • Zhang L, Kendrick C, Jülich D, Holley SA. Cell cycle progression is required for zebrafish somite morphogenesis but not segmentation clock function. Development. 2008;135:2065–2070. [PMC free article] [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...