Foxj1a is necessary and sufficient to specify motile cilia. Using transcriptional studies and slow-scan two-photon live imaging capable of identifying the number of motile and immotile cilia, we now established that the final number of motile cilia depends on Notch signalling (NS). We found that despite all left-right organizer (LRO) cells express foxj1a and the ciliary axonemes of these cells have dynein arms, some cilia remain immotile. We identified that this decision is taken early in development in the Kupffer’s Vesicle (KV) precursors the readout being her12 transcription. We demonstrate that overexpression of either her12 or Notch intracellular domain (NICD) increases the number of immotile cilia at the expense of motile cilia, and leads to an accumulation of immotile cilia at the anterior half of the KV. This disrupts the normal fluid flow intensity and pattern, with consequent impact on dand5 expression pattern and left-right (L-R) axis establishment.https://doi.org/10.7554/eLife.25165.001
Embryonic motile cilia play an essential role in body laterality patterning by generating a directional fluid flow inside the vertebrate left-right organizer (LRO) (Nonaka et al., 1998; Takeda et al., 1999; Nonaka et al., 2002). The mouse LRO (or node) is lined with monociliated cells, presenting motile and immotile cilia (McGrath et al., 2003). While motile cilia in the pit region generate a directional fluid flow (the nodal flow), the immotile cilia, present mostly in the perinodal crown cells, are thought to be able to sense it (McGrath et al., 2003). This sensing mechanism is still unknown and continues to deserve tentative updates in the field both using the mouse model (Delling et al., 2016) and the zebrafish model (Ferreira et al., 2017). Whether it involves mechano- or chemosensation, or both, is still not demonstrated but it is known to lead to the activation of the channel Polycystin-2 (PKD2) (Yoshiba et al., 2012). This channel was elegantly shown to be necessary for the asymmetric expression of dand5 (DAN domain family, member 5) on the right side of the mouse node (Yoshiba et al., 2012) and consequently allowing the propagation of Nodal into the left Lateral plate mesoderm (LPM) (Marques et al., 2004). This signal is amplified via a self-enhanced lateral-inhibition system (SELI) (Nakamura et al., 2006) at the left LPM, which consists in the activation of the genetic cascade Nodal-Pitx2-Lefty2 and ends with the correct formation and asymmetric positioning of the visceral organs (Nonaka et al., 2002).
In zebrafish, the left-right (L-R) axis establishment starts in a fluid-filled organ designated Kupffer’s vesicle (KV) (Essner et al., 2005; Kramer-Zucker et al., 2005). Functionally, this organ is the homologue of other vertebrate LROs like the mouse node (Nonaka et al., 2002) and the gastrocoel roof plate in Xenopus (Schweickert et al., 2007). The KV originates from a cluster of cells, the dorsal forerunner cells (DFCs), which migrate in the forefront of the shield during gastrulation (Cooper and D'Amico, 1996). At the end of gastrulation, the DFCs form an ellipsoid fluid filled vesicle. While KV lumen inflates each cell extends one cilium towards the lumen (Amack et al., 2007; Oteíza et al., 2008). As in the mouse node, the KV cilia also produce a directional fluid flow that leads to an asymmetric dand5 gene expression (Lopes et al., 2010; Sampaio et al., 2014). Our previous work determined that KV cilia can also be divided into two populations according to whether they are functionally motile or immotile (Sampaio et al., 2014). We also showed that the DeltaD zebrafish mutant (dld−/−) for Notch signalling (NS) presents an increase in the number of motile cilia in the LRO (Sampaio et al., 2014), suggesting that NS may modulate the number of motile cilia in this organ. Equivalently, Boskovski et al. reported that GALNT11, an N-acetylgalactosamine type O-glycosylation enzyme needed to activate NS, regulates the ratio between motile and immotile cilia in the Xenopus’ LRO, where less NS also increased the number of motile cilia (Boskovski et al., 2013). The authors showed that changing the ratio between motile and immotile cilia caused downstream defects in L-R patterning of the laterality marker pitx2c (paired-like homeodomain 2), and in the correct heart looping of Xenopus embryos (Boskovski et al., 2013).
The transcription factor Forkhead box J1a (Foxj1a) has been established as the motile cilia master regulator in the KV cells (Stubbs et al., 2008; Yu et al., 2008). Without it cilia do not form, altering the expression of L-R markers and randomizing organ situs (Tian et al., 2009). Its transcription initiates during gastrulation in the DFCs, and Foxj1a is responsible for the transcriptional activation of several motility genes, such as dnah7 (Choksi et al., 2014) and dnah9 (axonemal heavy chain dyneins that mediate the movement of cilia by hydrolysing ATP) (Yu et al., 2008; Choksi et al., 2014). This suggests that in wild type (WT) embryos, where motile and immotile cilia are present in neighbouring cells (Sampaio et al., 2014), Foxj1a function may be antagonized by other factors, explaining why cilia remain immotile in some cells.
In order to understand the mechanisms behind the choice of motile versus immotile cilia, we manipulated NS and Foxj1a levels and evaluated their impact in the ratio of motile and immotile cilia in the zebrafish LRO. We concluded that, independently from variations in foxj1a mRNA levels, all cilia seem to acquire a motile ultrastructure. However, NS modulates the final number of functionally moving cilia early in the KV precursors, via a mechanism that involves the activity of Her12 (hairy-related 12), a transcription repressor so far only involved in somitogenesis (Shankaran et al., 2007). We then experimentally showed that Her12 mediated motile/immotile cilia ratio imbalance impacts not only on fluid flow intensity but also on its pattern. This supports our previous data showing that for the KV to develop a robust fluid flow, capable of promoting normal organ situs, it requires a minimum of 30 motile cilia and an anterior-dorsal cluster of motile cilia (Sampaio et al., 2014; Smith et al., 2014). Overall, the evidence presented demonstrates the importance of regulating the motile/immotile cilia ratio in the generation of a robust and functional fluid flow in the zebrafish LRO.
As the KV starts to form and inflate, cilia begin to emerge and some start to beat, generating a fluid flow (Essner et al., 2005; Kramer-Zucker et al., 2005), while others remain static (Sampaio et al., 2014). How this dichotomy evolves over time is unknown. We thus questioned whether the percentage of immotile cilia at the 8-somites stage (ss) was the same since the beginning of the LRO formation. To better understand this phenomenon, we injected WT zebrafish embryos at the 1 cell stage with a titrated, non-toxic concentration of arl13b-GFP mRNA, and performed live imaging in a two-photon microscope. By scanning the whole KV at a low speed (0.16 frames per second, fps) in a slow scanning mode by using a high pixel dwell time (22.4 microseconds) (Video 1; Figure 1A) it was possible to accurately distinguish the motile cilia (Figure 1B) from the immotile cilia (Figure 1C), quantify them, and track them throughout the KV development. An higher pixel dwell time allows to better identify motile from immotile cilia, thus explaining some differences in quantification of immotile cilia in comparison with another recent work (Ferreira et al., 2017), We then imaged the same embryos from 3 to 8 ss (5 embryos; 294 cilia); we identified every motile and immotile cilium at each time point using the 3D stacks acquired, and then tracked them throughout time in 3D projections (Video 2).
The results showed that at 3 ss, the vast majority of cilia were immotile (Figure 1D) and cilia distribution and KV cell shape were homogeneous (Figure 1F and G). Between 3 and 5 ss, we observed the highest rate of increase in the percentage of motile cilia (Figure 1D, F–K), which then tended to stabilize at 6–8 ss (Figure 1D, L–O). This is accompanied with a change in KV cell shape and anterior-dorsal cluster formation as previously reported (Wang et al., 2012). While tracking cilia, we observed two main different types of cilia behaviour (Figure 1E; 4 embryos; 231 cilia). We found that 79 ± 7% of cilia transitioned from immotile at 3 ss to motile at 8 ss (Figure 1E), while 16 ± 8% of cilia never became motile during the time-window of the assay (Figure 1E). We reasoned that if immotile cilia are always immotile from the beginning to the end of the KV lifetime then something should be determining the fate of such cilia very early in development.
In agreement with the increase in number of motile cilia from 3 to 8 ss, we also observed changes in the pattern of the KV fluid flow throughout development (Figure 2). As we had previously reported (Sampaio et al., 2014) we can calculate the speed and direction of the KV fluid flow by tracking native particles present in the KV by bright field microscopy (Figure 2—videos 1–3). Therefore, at 3–4 ss we observed a complete absence of directional fluid flow (Figure 2A and C; 5 embryos; 258 tracks), with the native particles presenting what seemed to be Brownian motion (2.32 ± 0.95 μm s−1; Figure 2—video 1; Figure 2B). This observation is in agreement with the low number of beating cilia found at 3 ss (6 ± 5%; Figure 1D). Again, here our results contrast with those from Ferreira et al. (2017), which are built on numerical predictions of flow supported by a much lower number of immotile cilia. So, flow forces were calculated in different ways in the two studies, while we measured flow velocity by following native particles in live embryos averaged over a number of embryos, Ferreira et al. (2017) mapped cilia position and tilting in live embryos and then predicted the flow forces based on numerical simulations. As development progressed and the number of beating cilia increased (5 ss – 63 ± 23%; 6 ss – 76 ± 8%; Figure 1D), we began to observe a weak directional fluid flow at 5–6 ss (6.61 ± 3.26 μm s−1; Figure 2G), which at this time was still homogeneous throughout the KV (Figure 2—video 2; Figure 2E – G; 5 embryos; 309 tracks). Finally, at 7–8 ss, when the number of beating cilia began to stabilize (8 ss – 83 ± 6%; Figure 1D), the directional fluid flow acquired its characteristic velocity (9.81 ± 5.36 μm s−1) and heterogeneous pattern (Figure 2—video 3) (Sampaio et al., 2014), presenting a faster flow at the anterior-left quadrant of the KV (Figures 2I, K and 3 embryos, 128 tracks). While tracking native particles we could also account for cilia motility and thus by an independent method we confirmed our previous quantifications on the ratio between motile vs immotile cilia. This method was based on a high frame rate acquisitions (500 fps) using a high-speed video camera.. Again, we observed an increase in the percentage of motile cilia throughout development, ranging from 27% at 3–4 ss (4 embryos; 48 cilia; Figure 2D), 70% at 5–6 ss (4 embryos; 63 cilia; Figure 2H) to 84% at 7–8 ss (4 embryos; 89 cilia; Figure 2L). We also determined the cilia beat frequency (CBF) at these three different stages and we found that cilia beat faster as development progressed (Figure 2D – L, right panels). The CBFs calculated were 33 ± 4 Hz (4 embryos; 8 cilia), 37 ± 5 Hz (4 embryos; 29 cilia), and 40 ± 4 Hz (4 embryos; 34 cilia) for the 3–4 ss, 5–6 ss and 7–8 ss, respectively. The ciliary frequency at 7–8 ss was significantly higher than at 3–4 ss. Overall the values found for the percentage of motile cilia together with the CBF may account for the differences found in the velocity of the fluid flow throughout development.
We have previously reported that DeltaD (dld−/−) mutant embryos have more cilia moving in the KV than the WT (Sampaio et al., 2014), indicating that Notch signalling (NS) could be responsible for the decision between motility and immotility of cilia in the KV. To understand why dld−/− had more motile cilia than control embryos we performed a comparative transcriptomic analysis of DFCs from WT and dld−/− mutant zebrafish embryos (Figure 3A–D). After curation, we obtained a list of 706 genes with a linear fold change (FC) in transcription higher than 2 (Table S1a in Supplementary file 1).
Our intention was to determine which genes in dld−/− mutants were implicated specifically in cilia motility. Therefore, we decided to use the Cildbv2 reference base (Arnaiz et al., 2009) and Genevenn algorithm (Pirooznia et al., 2007) to determine the hits in our list that had orthologues already associated with cilia motility in other model organisms. We chose Caenorhabditis elegans, a primary cilia model; and Chlamydomonas reinhardtii, a motile cilia/flagella model (Table S1b in Supplementary file 1). The intersection analysis between these models permitted us to distinguish the ciliary genes specific for motility in our list. Because foxj1a was up-regulated (FC = 1.62, p value = 0.029) in the microarray, we also compared our gene list with the list of Foxj1a-induced genes published by Choksi et al. (2014). The result of these intersections showed that 18% (129 genes) of the genes in our original list were related to ciliogenesis and of those, 67% (86 genes) were specifically related to motility (Table S1b in Supplementary file 1). Motility specific column). Among these were the known motility axonemal genes dnah7 and dnah10 (Dynein, Axonemal, Heavy Chain 7 and 10), rsph3 and rsph9 (Radial Spoke Head 3 and 9) together with dld, foxj1a and rfx2. The axonemal genes were up-regulated in dld−/− mutants and were therefore good candidates to explain the increased number of motile cilia. To validate some of the motility-associated genes obtained in the microarray, we focused on the following genes: the ciliogenesis master regulator, rfx2 (regulator master x 2) (Swoboda et al., 2000); the motile cilia master regulator foxj1a, the notch ligand, dld; and two motile cilia axonemal components, dnah7 and rsph3. We performed quantitative PCR (qPCR) using mRNA extracted from fluorescently activated cell sorted DFCs. Of these 5 tested genes, only 3 (foxj1a, dld, and dnah7) showed consistent changes in transcription regulation between the two methods (Figure 3E and F), the microarray and qPCR analyses, also confirmed by in situ hybridization in bud stage and 8 somite stage (Figure 3—figure supplement 1).
In order to check if Foxj1a, the master regulator of motility, was expressed in all KV cells or KV precursor cells, we decided to look at a zebrafish line that expresses GFP under a foxj1a minimal promoter (Tg:foxj1a:GFP) (Caron et al., 2012) and determined whether all WT KV cells were GFP positive. Our reasoning was that if we found that Foxj1a was absent in around 16–20% of WT DFCs, this alone could account for the lack of motility observed at 8 ss (Figure 1D). We fixed embryos at 8 ss and by immunofluorescence with an antibody against GFP and another against acetylated alpha-tubulin (Figure 4—figure supplement 1C–E; Video 3), we confirmed that the number of foxj1a-positive cells and the number of cilia present in each KV were concordant (Figure 4—figure supplement 1F; p=0.9479, 7 embryos, 514 cells, 513 cilia). This experiment showed that foxj1a:GFP was present in all KV cells despite the fact that some cells have immotile cilia throughout development. So, we then questioned how this foxj1a promoter related to the actual mRNA expression. We confirmed that, as reported before by Yu et al. (2008) foxj1a mRNA is strongly expressed in the DFCs at bud stage (Figure 4—figure supplement 1A) (Yu et al., 2008). However, this early expression is not yet seen with the foxj1a:GFP minimum promoter described by Caron et al. (2012) (Figure 4—figure supplement 1) (Caron et al., 2012). On the other hand, at 8 ss the mRNA expression of foxj1a is no longer observed in the KV, while it is detectable in the pronephros and neural tube, but the reporter shows GFP in all KV cells (Figure 4—figure supplement 1B and C–E, respectively). So, we must conclude that this promoter is not faithfully representing the foxj1a gene expression along time. However, by comparing the foxj1a in situs with another reporter line for DFCs, the sox17:GFP reporter, we showed that the foxj1a mRNA staining closely matches the sox17:GFP labelling (Figure 4A–C). Altogether these experiments suggest that foxj1a is expressed in all DFCs and that the foxj1a:GFP observed at 8 ss is due to a delayed reporter and/or GFP perdurance.
Moreover, the information that foxj1a was up-regulated in dld−/− mutant DFC cells (Table S1a in Supplementary file 1 suggested a regulation of foxj1a expression by NS. Therefore, in order to understand the crosstalk between NS and foxj1a in regulating cilia motility, we monitored the transcription levels of her12 (hairy-related 12), dnah7 and foxj1a itself while manipulating NS and Foxj1a protein levels. These genes were chosen because: i) her12 is a NS direct transcription target in the tail bud and somites (Shankaran et al., 2007) and in this study we found that it is highly expressed in DFCs at bud stage, and down-regulated in dld−/− mutants (Table S1a in Supplementary file 1) as previously described (Thisse and Thisse, 2004); and ii) dnah7 was reported to be a transcriptional target of Foxj1a (Choksi et al., 2014) and an essential structural component of motile cilia, without which zebrafish KV cilia become static (Sampaio et al., 2014). Similar to the expression levels of foxj1a, we found dnah7 significantly up-regulated in dld−/− mutants (Figure 4—figure supplement 1H) and expressed in DFCs at bud stage (Table S1a in Supplementary file 1). The restricted expression patterns of her12, foxj1a and dnah7 at bud stage indicated that it was possible to use mRNA from whole embryos to determine their expression levels by qPCR (Neugebauer et al., 2013; Jurisch-Yaksi et al., 2013). As a control of this procedure, we tested dnah9 and rfx4, two genes that, despite being transcribed in DFCs were not influenced by NS, according to our microarray and qPCR data (Figure 4—figure supplement 1H).
We then set out to manipulate the levels of NS by using the single mutants dld−/− and delta like c (dlc−/−), and the double mutant dlc−/−;dld−/−. In order to increase NS we overexpressed Notch intracellular domain (NICD OE). Showing the efficiency of NS manipulations, the her12 transcription levels behaved accordingly, being significantly lower in the dld−/−;dlc−/− double mutant (Figure 4D), and significantly higher in the NICD OE (Figure 4D).
For each NS manipulation we also interfered with the levels of Foxj1a. This was assured either by knocking down the foxj1a expression with a morpholino against foxj1a translation start site (Foxj1a KD) or by overexpressing it via the injection of Danio rerio foxj1a mRNA (Foxj1a OE). As expected, the dnah7 expression served as readout of the Foxj1a manipulation efficiency, as its expression significantly correlated both with the endogenous levels of Foxj1a (Foxj1a Not Manipulated; NM) and upon its overexpression (Foxj1a OE) (Figure 4E; Pearson’s correlation coefficient: r = 0.8799, p value = 0.0208; and r = 0.9523, p value = 0.0034, respectively). Moreover, when compared to the WT situation, the dnah7 mRNA levels were significantly higher upon the Foxj1a OE (Figure 4E). This observation is in agreement with Choksi et al. (2014) who previously showed that dnah7 is induced by Foxj1a. Concordantly, the transcription levels of dnah7 were lower in the Foxj1a KD than in the control situation (Figure 4E), demonstrating the efficiency of the morpholino in blocking the foxj1a mRNA translation. Interestingly, we observed that the changes in dnah7 expression were dependent only on the variations in Foxj1a, irrespectively of the NS background (Figure 4D and E; Pearson’s correlation coefficient: r = −0.1859, p value = 0.4198).
The quantification of the expression data led us to conclude that the regulation of foxj1a transcription by NS, uncovered in our microarray, was dependent on the NS context (Figure 4—figure supplement 1I). In short, the qPCR analysis showed that NS impacts indirectly on the transcription of dnah7 by affecting foxj1a transcription levels. This data confirms the results obtained in the comparative transcriptome analysis between dld−/− mutants and WT for the genes in question: her12, foxj1a and dnah7.
Next, to independently evaluate the impact on the combined manipulation of NS and Foxj1a levels on KV cilia motility we performed live imaging of KV cilia at 8 ss. We selected three NS conditions: dld−/−;dlc−/− double mutants (low NS); WT embryos (normal NS); and embryos injected with NICD mRNA at 1 cell stage (high NS), while simultaneously manipulating the levels of Foxj1a by overexpressing it in the three experimental groups. We chose not to knockdown Foxj1a because this would lead to the loss of cilia altogether (Yu et al., 2008). We decided to use the double mutants dld−/−;dlc−/− because we had previously observed laterality defects in both dld−/− and dlc−/− single mutants (Lopes et al., 2010) and more importantly, we wanted to avoid redundancy among the genes coding for the two Delta ligands. We filmed ciliary motility at 8 ss because by this stage we know that the anterior and left flow hotspots are established (Figure 2I), which we demonstrated before to be crucial for the L-R patterning (Sampaio et al., 2014; Smith et al., 2014; Montenegro-Johnson et al., 2016).
The live imaging analysis showed that when NS is reduced, the number of immotile cilia in the KV decreased significantly from 20 ± 5% in controls to 12 ± 4% in dld−/−;dlc−/− mutants (Figure 4F; 24 embryos, 1047 total cilia in control; 7 embryos, 304 total cilia in dld−/−;dlc−/− mutants). Conversely, when NS is increased, the number of immotile cilia increased from 20 ± 5% in controls to 31 ± 10% (Figure 4F; 8 embryos, 353 total cilia in NICD OE;). Importantly, NS levels interfered with the motile/immotile cilia ratio and not with the total number of cilia (24 embryos, 44 ± 12 cilia in Control versus 8 embryos, 44 ± 11 cilia in NICD OE,; p=0.9154) or with cilia length (Figure 4—figure supplement 2A). Regarding cilia length, we evaluated the cilia length from all treatments done in Figure 4F. These experiments were done using a 50 pg of arl13b-GFP so that we could observe the cilia motion live. Using the same movies we have also measured cilia length in 3D. We found no differences between any treatment (Figure 4—figure supplement 2A). Using the arl13b-GFP overexpression we have normalized cilia length. In this way, we could overcome the cilia length differences that Notch signaling manipulations produce and successfully uncouple cilia length from cilia motility.
So, despite the qPCR data showing a significant increase in both dnah7 and foxj1a transcription levels after foxj1a was over-expressed (Figure 4E and Figure 4—figure supplement 1I, respectively), live imaging experiments showed that there were no changes in the actual number of motile cilia when foxj1a was over-expressed, irrespective of the NS background (Figure 4F; nine embryos, 380 total cilia in Foxj1a OE; eight embryos, 368 total cilia in dld−/−;dlc−/− mutants + Foxj1 a OE; eight embryos, 347 total cilia in NICD OE +Foxj1 a OE). We conclude that the failure of Foxj1a OE to increase the number of motile cilia, and specifically the failure in rescuing the loss of motile cilia in the NICD OE embryos (Figure 4F) strongly suggests that the KV cells are already committed to a certain motility fate. We propose NS modulates such functional decision downstream of Foxj1a function.
Next, we investigated whether a higher level of foxj1a transcription, as the one found in the dld−/− mutants, was causing increased CBFs on motile cilia. However, by overexpressing Foxj1a, we observed no significant change in motile cilia CBF (WT = 37.27 ± 7.202 Hz, Foxj1a OE = 36.30 ± 6.877 Hz; p=0.4990; Figure 4—figure supplement 1G). This experiment also agreed with the fact that in dld−/− mutants, we did not observe an increase in CBF (Sampaio et al., 2014). Our results suggest that this level of Foxj1a overexpression, despite eliciting the transcription of the downstream target dnah7 (Figure 4—figure supplement 1I) has no consequences in KV cilia motility, contrary to our initial assumptions.
Next, we investigated whether the Her12 transcription factor was directly involved in the regulation of the ratio between motile and immotile cilia. In order to test this hypothesis we overexpressed Her12 (Her12 OE) by injecting its mRNA at 1 cell stage and imaging the embryos at 8 ss. We observed a significant increase in the percentage of immotile cilia from 20% in controls to 27% (Figures 4F, 9 embryos, 305 total cilia), recapitulating the results observed in the NICD OE assay (Figure 4F).
Overall these observations indicate that, in parallel to the activation of the motile cilia program by foxj1a transcription, which should occur in all DFCs, NS decides which cilia will be stopped. Since both foxj1a and her12 are expressed at bud stage in the DFCs (Thisse and Thisse, 2004; Neugebauer et al., 2013; Jurisch-Yaksi et al., 2013), we reason that on one hand, Foxj1a specifies cilia capable of moving, while on the other hand, NS prevents the fulfilment of the motile cilia Foxj1a-activated program. This could be achieved either by structurally changing cilia (e.g. preventing the assembly of dynein arms), or by somehow inhibiting the motility of structurally motile cilia (e.g. switching off the dynein motors). To tackle these different scenarios in the absence of appropriate zebrafish antibodies, we cloned the coding sequence of dnal1, fused it to the mCherry fluorescent tag and injected this construct in arl13b-GFP transgenic embryos that enable live imaging of motile and immotile cilia. As Dnal1 is a light chain outer dynein arm axonemal dynein motor required for cilia movement (Horváth et al., 2005; Mazor et al., 2011), we reasoned that if only motile cilia expressed this construct then we could favour the hypothesis that motile cilia were likely to be structurally different. On the other hand if immotile KV cilia also expressed the Dnal1 construct then we could predict that all KV cilia might be structurally similar. In this last scenario, perhaps some other factor capable of switching off the dynein motors ATPase activity could be occurring.
Our results showed that Dnal1 was present in both motile and immotile KV cilia (Figure 4G – R; six embryos; compare K, N, Q with L, O, R). The quantification for the injection of the construct dnal1-mCherry was done in a sample of 56 cilia in a total of 4 embryos. In the sample of cilia positive for dnal1-mCherry, we saw 72% of motile cilia and 28% of immotile cilia. This ratio is not statistically different from the one for the controls used in Figure 4F (p=0.17, Fisher test) meaning that the dnal1-mCherry construct did not affect the cilia motility status and that both motile and immotile cilia can express dnal1-mCherry. Importantly, primary cilia from the tail region around the KV were all negative for dnal1-mCherry, which shows that in a mild overexpression scenario, this construct cannot enter a primary cilium (Figure 4J,M,P). Overall this experiment suggests that most KV cilia have dynein arms, i.e., may have the necessary machinery to move.
In order to have a definitive answer as to whether all KV cilia structurally belong to the sub-type of motile cilia, we performed transmitted electron microscopy by sampling the full KV of 3 embryos. We used 10 ss embryos and sectioned their KVs with 5–7 micron intervals (Figure 4—figure supplement 4S). We imaged a total of 101 cilia, providing an average of 34 cilia per KV, which according to Figure 4—figure supplement 2B, represents on average a random coverage of 77% of the cilia population of each KV. Our findings demonstrated that all cilia imaged had an ultra-structure characteristic of motile cilia. Some cilia had a clearly visible central pair while others showed an unfocused central pair and all cilia had visible dynein arms (Figure 4U - W; Table 1). So, we conclude that it is very likely that all KV cilia are equipped with motility apparatus.
From the microarray and qPCR validations we discovered that her12, a homologue of the mammalian hes5 that belongs to a family of transcription repressors (reviewed in (Kageyama et al., 2007)), was a likely candidate to mediate the cilia immotility regulation. Furthermore, we showed that overexpressing her12 induced a significant increase in immotile cilia, and thus recapitulated the NICD overexpression phenotype. Next we asked how many cells in the cluster of DFCs were expressing her12 at bud stage, because we reasoned that if Her12 was mediating the establishment of immotile cilia, then the number of immotile cilia should be similar to the number of her12 positive cells. For this assay we used the transgenic line sox17:GFP, which allowed us to identify the DFCs while detecting her12 expression pattern by fluorescent in situ hybridization (Figure 5—videos 1 and 2; Figure 5A–F).
Importantly, we showed that in the control situation 29 ± 8% of the DFCs were her12 positive (Figure 5M; 9 embryos) which represent 10% more than the percentage of immotile cilia detected at 8 ss in the motility assay (Figure 4F). We also checked how was her12 in NICD OE embryos (Figure 5—videos 3 and 4; Figure 5G–L). We found a significant increase of 14% in the percentage of her12 positive DFCs (Figure 5M; from 30 to 44 ± 11%; 9 embryos). This represents a 14% more than the percentage of immotile cilia detected at 8 ss for this treatment (Figure 4F). This differences between her12 positive cells and immotile cilia percentages could be explained by the existence of a fine-tuning mechanism (Basch et al., 2016).
Additionally, we noticed that these cells were preferentially located at the posterior half of the DFC 3-dimensional cluster (Figure 5N; 40 ± 5% and 60 ± 5% for the anterior and the posterior halves, respectively; 9 embryos). When NICD is over-expressed, regarding cell distribution, we noticed that her12 positive cells became homogeneously distributed in both the anterior and posterior halves of the DFC cluster in contrast to the control situation (Figure 5N; 9 embryos).
It is well established that KV maturation involves changes in cell shape to form an anterior-dorsal cluster (Roxo-Rosa et al., 2015) crucial for the anterior-left fluid flow hotspots (Sampaio et al., 2014; Smith et al., 2014; Wang et al., 2012; Montenegro-Johnson et al., 2016). However, in light of the bias of her12 positive DFCs at the posterior part of the DFC cluster in control embryos (at bud stage Figure 5C,D,F,N), this led us to test a new hypothesis. Knowing that at 8 ss there is no anterior-posterior bias in the position of immotile cilia (Sampaio et al., 2014) we postulated that in a WT embryo with 3 ss, more cells bearing immotile cilia would be present in the posterior part of the KV, judging from the expression pattern of her12. We thus predict that some of those cells will be pushed to the anterior part of the KV, as previously reported (Wang et al., 2012) and also confirmed here, leading to a homogeneous final distribution of immotile cilia (Figure 5O; immotile cilia in blue) at 8 ss. On the other hand, when NICD was overexpressed, we found that the posterior bias of her12 positive DFCs was lost (Figure 5N), which should lead to a homogenous distribution of immotile cilia in the KV at 3 ss, prior to cell shape changes. Therefore, upon KV maturation, we predict an accumulation of immotile cilia in the anterior half (Figure 5P). To test this hypothesis, we looked again to the KVs scanned at 8 ss, and used them to create 3D reconstructions and to map the position of each immotile cilia for all treatments. Using the principles of the Cartesian referential and, establishing the centre of the KV as the origin (0, 0, 0) of the three coordinate axes, it was possible to attribute Cartesian coordinates (x, y, z) to each immotile cilium. Using this information, we were able to plot the distribution of the immotile cilia along the anterior-posterior, the dorsal–ventral, and the left–right axes (Figure 5Q–V and Figure 5—figure supplement 1A–F). Our results showed that the position of immotile cilia was homogeneous across all 3 axes in the control (Figure 5Q–S; 17 embryos; 159 cilia), when foxj1a is overexpressed (Figure 5—figure supplement 1A–C; 9 embryos; 89 cilia), and in the double mutant dld−/−;dlc−/− (Figure 5—figure supplement 1D–F; 7 embryos; 38 cilia). Only when NICD was overexpressed, was there a significant accumulation of immotile cilia at the anterior part of the KV (Figure 5T–V; 8 embryos; 107 cilia), thus supporting our hypothesis. To quantify which transitions in cell position were occurring (anterior to posterior, left to right, dorsal to ventral, or vice-versa) we revisited our time-lapse results and tracked the immotile cilia through development for the control situation. Overall, our results showed that in WT embryos, posterior to anterior transitions are the most common (Figure 5—figure supplement 2A – B, four embryos, nine cilia transitions, 19 immotile cilia tracked), further confirming our model. Here we show one such immotile cilium transitioning from the posterior part of the KV (at three ss), to an anterior position (at eight ss). We found a total of 4 immotile cilia transitioning from a posterior position to a more anterior position in as many embryos. If we assume an average KV with 44 ± 12 cilia (Figure 4—figure supplement 2B), 20% of these cilia will be immotile, giving an approximate number of 8 immotile cilia. Since at bud stage, 60% of the her12 positive cells are located at the posterior part of the DFCs cluster, five immotile cilia (60% of total immotile cilia) start at a more posterior position, while 3 (40% of total immotile cilia) will be positioned more anteriorly. Therefore, one immotile cilia transition from posterior to anterior would suffice in order to evenly distribute immotile cilia at eight ss. This analysis is concordant with the observed number of transitions.
Finally, the accumulation of her12 positive cells in the posterior half of the DFC cluster of control embryos led us to question how NS was occurring, i.e. which cells were DeltaD positive. Using a DeltaD specific antibody, we looked for the expression of DeltaD at bud stage in the vicinity of DFCs (Figure 5—video 2; Figure 5—figure supplement 2C–G). This assay showed that DeltaD protein is mainly present in the cells surrounding posteriorly, laterally and dorsally to the cluster of DFCs (Figure 5—video 5; Figure 5—figure supplement 2E,G and F, respectively). We thus conclude that NS occurs between the DeltaD positive cells, forming a crescent posteriorly to the DFCs, depicted in our model in Figure 5—figure supplement 2H. notch1b positive DFCs (FC = 3.3 in Table S1a, Supplementary file 1) together with notch1a, notch2 and notch3 are also expressed in the DFCs at bud stage. Some DFCs, preferentially at the posterior boundary, will transcribe her12 (Figure 5—figure supplement 2H) and we postulate that those are the ones where ciliary motility will be inhibited.
To determine whether the 7% increase in immotile cilia observed upon her12 overexpression is relevant for the determination of laterality (Figure 4F) we measured fluid flow and scored dand5 expression in the same embryos, this time without co-injecting arl13b-GFP. Furthermore, we scored the heart position by observing it in the live embryo at 30 hpf and the liver position at 53 hpf by ISH with a foxa3 probe. Fluid flow analysis showed that her12 overexpression leads to, not only a loss of the dorsal-anterior flow hotspot (compare Figure 6A–B with Figure 6C-D) but also a 12.5% reduction in flow velocity (Figure 6B,D; from 9.94 ± 4.47 in WT to 8.70 ± 4.26 µm s−1 in Her12 OE; 9 and 12 embryos, respectively; 533 and 521 tracks, respectively). Regarding dand5 expression pattern, it correlated with the difference in the fluid flow: in contrast to 100% of asymmetric on the right in control embryos (Figures 6E, 10 embryos), 42% of embryos overexpressing Her12 showed a symmetric expression pattern (Figures 6E, 12 embryos). In terms of organ situs, 41% of Her12 overexpressing embryos displayed heterotaxia (incorrect positioning of either the heart or liver; Figures 6F, 75 embryos) while 100% of the control embryos showed left heart and left liver (situs solitus; Figures 6F, 99 embryos). To test if the observed decrease in flow speed and pattern, as well as the change in dand5 expression pattern and organ situs were not due to a change in cilia length we measured the cilia length from 7 KVs upon Her12 OE (without arl13b-GFP). The results showed that there was no significant difference between treated and control experiments (Figure 6G). In order to check if the Her12 OE really recapitulated the NICD effect leading to an increase in the anterior immotile cilia, we then counted the motile and immotile cilia that were present on the Anterior-Posterior halves of each KV. These results confirmed an increase in the number of immotile cilia at the anterior KV halves upon Her12 overexpression (Figure 6H–I). Altogether, we demonstrated that the change in location of immotile cilia and its accumulation at the anterior half of the KV had a significant effect in fluid flow dynamics, impacting on dand5 expression pattern and consequently in organ situs.
In this manuscript we show the progressive increase of cilia motility in the KV from 3 ss to 8 ss using two independent methods: slow scanning two-photon microscopy allowing for cilia beat pattern trajectories to be revealed and fast acquisition transmitted light imaging by high speed videomicroscopy allowing for slow-motion video analysis (Figures 1 and 2). This increase in motility occurs as immotile cilia start acquiring motility throughout development as previously reported (Yuan et al., 2015). We report here an increase in the intensity of the fluid flow velocity together with a significant increase in the CBF from 3 to 8 ss (Figure 2). The changes in fluid flow velocity as the number of motile cilia increased are in line with our previous predictions (Sampaio et al., 2014), which dictated that a minimum of 30 cilia is necessary for a robust fluid flow showing anterior-left hotspots at 8 ss (Smith et al., 2014; Montenegro-Johnson et al., 2016). By performing live imaging in a two-photon microscope we could establish that the average number of cilia in a KV is 44 ± 12 (Figure 4—figure supplement 2B; 1047 cilia in 24 embryos), which according with our previous predictions (Sampaio et al., 2014)is more than sufficient to create a robust fluid flow at 8 ss, even accounting for the observed 20% of immotile cilia in WT embryos.
In order to understand why dld−/− mutants had more motile cilia in the KV we performed a tissue specific comparative transcriptomic analysis that showed differential expression of genes known to be Foxj1a-induced (Choksi et al., 2014) (Table S1b in Supplementary file 1) We also found that the dld−/− mutant presented a significant up-regulation of foxj1a (FC = 1.62), which indicated a possible regulation in transcription of this gene and its downstream targets by NS. To determine if that was in fact the case we observed the KV cilia in action, using fluorescent live imaging at 0.16 frames per second (fps) and a resolution of 512 × 512 pixels. Contrary to our expectations, we showed that the number of motile cilia was not affected by overexpressing foxj1a. Furthermore, we demonstrated that the balance between motile and immotile cilia is regulated by NS, similarly to what was reported in the Xenopus LRO (Boskovski et al., 2013). However we did not detect the same interaction between NS and foxj1a as suggested by Boskovski et al. (2013). Interactions between NS and Foxj1a have also been suggested by Jurisch-Yaksi et al. (2013) who showed that decreased levels of Rer1 (retention in endoplasmic reticulum sorting receptor 1), lead to increased γ-secretase activity and consequently increased NS, which lead to decreased foxj1a levels in zebrafish embryos (Jurisch-Yaksi et al., 2013). They showed this interaction occurs in several ciliated organs, such as neuromasts, pronephros, olfactory pits, and the sensory patch of the inner ear. In KV, Rer1 knockdown causes the decrease of foxj1a expression, but unlike in other organs, when NS is increased, foxj1a levels remain unchanged (Jurisch-Yaksi et al., 2013). This last observation concurs with our qPCR results that show that when NICD is over expressed, the expression of foxj1a is unchanged (Figure 4—figure supplement 1I).
It is also known that NS affects Gemc1/MCIDAS, which are required for Foxj1 expression and multiciliated cell fate determination (Kyrousi et al., 2015). In this context NS indirectly regulates the pattern of multicilliated cells in the epidermis of Xenopus laevis embryos and in the human airway epithelium, by the process of lateral inhibition (Deblandre et al., 1999; Marcet et al., 2011a). Additionally, the epithelia of the pronephric ducts from zebrafish embryos also present two types of cells, a multiciliated cell and a monociliated principal cell. Their differentiation and patterning is also determined by a NS-dependent lateral inhibition mechanism (Ma and Jiang, 2007; Liu et al., 2007). In this organ, multiciliogenesis is also inhibited by NS, with multiciliated cells expressing rfx2 and jagged2, and principal cells expressing notch3 and her9 (hairy-related 9, a NS downstream target) (Ma and Jiang, 2007; Liu et al., 2007). In the three multicilliogenesis models discussed, it was suggested that NS inhibited multiciliogenesis by blocking the early expression of either foxj1 (Marcet et al., 2011a; Marcet et al., 2011b) or rfx2 (Ma and Jiang, 2007; Liu et al., 2007). There was an exclusion of expression patterns for foxj1/rfx2 and notch, and it was this exclusion that triggered cell fate specification and differentiation. Boskovski et al. (2013) also suggest that NS functions upstream of Foxj1 and Rfx2 in determining the cilia type at the Xenopus LRO. Nevertheless, they never tested whether overexpressing foxj1a would rescue the immotility phenotype observed when NS was up-regulated (Boskovski et al., 2013). We were the first to show in the present study that foxj1a overexpression fails to rescue the immotile cilia number caused by the up-regulation of NS provided by NICD injection (Figure 4F). We propose here that Foxj1a acts early to specify all KV cilia as motile cilia and then Notch signalling triggers an immotility switch. Our strongest evidence comes from the random sampling by transmitted electron microscopy that clearly points to all KV cilia being of the motile sub-type. As far as we are aware we were the first to sample the whole KV by transmitted electron microscopy and to provide a high number of analysed cilia. Collectively, our results show that NS is acting downstream of Foxj1a to establish immotility.
Additionally, this study provides a new player involved in the motility decision: Her12, a paralogue of the mouse Hes5, known as a bona fide NS transcriptional target (Shankaran et al., 2007). In our hands her12 transcription proved a faithful NS readout in the KV cell precursors. We thus suggest that DeltaD binds to Notch1a [as suggested by Shankaran et al. (2007)] to activate transcription of her12. Subsequently, Her12 perhaps together with other factors, restricts the number of motile cilia, most likely by inhibiting the early transcription of a yet unknown crucial motility switch. Here we determined that overexpressing her12 resulted in a significant increase of immotile cilia in the anterior half of the KV, thereby softening the difference between anterior and posterior flow speed (Figure 6A–D). We demonstrate that this flow pattern/intensity disruption impacted on the expression pattern of dand5 (making it symmetric) and on organ situs generating heterotaxia (Figure 6E–F). We would like to emphasize that the results presented here on motile/immotile cilia ratio were uncoupled from the cilia length defects that Notch signalling causes (Figure 4—figure supplement 2A and Figure 6G) (Lopes et al., 2010).
A recent work by Ferreira et al. (2017) determined a considerably smaller percentage of immotile cilia in the zebrafish LRO. This discrepancy can be explained by the different pixel dwell times used during acquisition in the two different studies. The pixel dwell time was ten times higher in our study, meaning that while we spent 22.4 microseconds per pixel, Ferreira et al. spent 2.4 microseconds. This parameter measures the time that is spent scanning on each pixel, so the longer the dwell time on a particular pixel, the more signal will be detected and the less it will be distorted. In opposition, the faster the scanning the more distorted becomes the imaged object and any cilia movement, even when caused passively by the KV flow bending immotile cilia, will produce a cilium blur that may be inaccurately scored as a motile cilium. Therefore, a longer pixel dwell time is crucial because it will produce a very sharp image of the trajectory of each cilium, reconstituting the beating pattern of a motile cilium (Figure 1B), or in the case of an immotile cilium it will show a very bright stiff cilium (Figure 1C). Additionally, the number of discarded cilia by Ferreira et al. may have biased the ratio between motile and immotile cilia. We are aware that immotile cilia can be difficult to detect in 3D reconstructions due to cilia passive motion. In our own study to accurately identify immotile cilia we compared the 3D stacks side by side with the corresponding 3D reconstruction. For some immotile cilia this was crucial due to the smoother blur these may show in a 3D reconstruction. According to Ferreira et al. some embryos at 8 somites show 0% of immotile cilia which contrasts with the 20% we have detected by two independent methods.
Finally, the reason why immotile cilia detection method is important to be clarified is because Ferreira et al. (2017) refute the immotile cilia mechanosensory hypothesis based on their results. We therefore cannot agree with such conclusion. In fact, according to Ferreira et al. own theory for mechanosensation to work, we would need at least 3 immotile cilia on each LR side to discern noise from productive flow. Therefore based on our results and their theory, we can conclude that from 5 to 6 somite stage, we have enough immotile cilia per KV to be able to sense flow (13 cilia at 5 ss and 9 cilia at 8 ss, on average, which is more than 3 cilia on each side). So, overall we propose that the mechanosensing hypothesis cannot be excluded until demonstration of a chemosensing or mixed mechanism emerges.
Taken together it is possible to conclude that the balance between motile and immotile LRO cilia in zebrafish needs to be tightly regulated for the proper establishment of the L-R axis, by producing a robust, and most importantly, a heterogeneous patterned fluid flow. We would like to stress that in opposition to the mouse embryo (Shinohara et al., 2012) the fish LRO is very sensitive to motile cilia number and localization, most likely because of the different topographies of these two LROs. Whereas in mouse all motile cilia are on the floor of the node contributing to the effective flow, in zebrafish ventral pole cilia are antagonistic to the desired flow (Smith et al., 2014).
For the microarray and respective qPCR assays we used transgenic sox17:GFP zebrafish on WT AB background (gift from Carl-Philipp Heisenberg) and on deltaD/aeitr233 homozygous mutant AB background. The later line was generated by crossing dld−/− mutants with the sox17:GFP line, growing the progeny to adulthood and then incrossing and selecting the GFP fluorescent dld−/− mutant fish by their somite phenotype (Tübingen 2000 Screen Consortium et al., 2005). For the motility assays (live imaging and qPCR) we used WT zebrafish line, and the homozygous double mutant line dld−/−;dlc−/− line, both from AB background. This last line was obtained by crossing the homozygous mutant lines dld−/− and dlc-/-, growing the progeny to adulthood and then incrossing and selecting the dld−/−;dlc−/− mutant fish by their dlc−/− somite phenotype, and genotyping for the presence of the aeitr233 mutation. To genotype the deltaD/aeitr233 mutation in dld−/−;dlc−/− double mutants, we isolated genomic DNA from individual adults’ tails, amplified it and sequenced it. Homozygous mutant DeltaC/beatm98/tm98 fish and the transgenic zebrafish line Tg(Foxj1a:GFP) were gifts from Leonor Saúde lab. All these zebrafish lines were maintained and used as described in Westerfield, 2000. Embryos were kept at either 32°C (for Fluorescence Activated Cell Sorting, qPCR, and for CBFs measurements from 3 to 8 ss), or overnight (ON) at 25°C (for live imaging), in the dark and in E3 embryo medium, and were staged according to Kimmel et al. (1995). The procedures performed to zebrafish were approved by the Portuguese Veterinary General Administration (DGAV - Direcção Geral de Alimentação e Veterinária).
Embryos at bud stage (10hpf) were dechorionated with pronase (2 mg ml−1) and washed extensively in Danieu’s buffer. Cells were dissociated by manual pipetting in CO2 independent medium (Gibco) complemented with 0.5 mM EDTA. Cells were centrifuged at 700 g and re-suspended in 4 ml of the same medium (step performed 3 times). The cells were then re-suspended in 1 ml and filtered with a 30 µm filter (CellTrics) directly into a round bottom tube. FACS was performed with a FACSAria bench top High Speed Cell Sorter (Becton Dickinson) at 140 kPa (20 psi) and with a 70 µm nozzle, using a 488 nm laser and a 530/40 nm bandpass filter to excite and detect GFP, respectively. GFP positive cells present both in WT and in dld−/− mutant embryos were selected, collected into RTL buffer from the RNeasy micro Kit (Qiagen Inc., Valencia, CA), and immediately frozen at −80°C. We collected three independent samples for WT and mutant embryos, each with around 20 000 cells.
Total RNA was isolated with the RNeasy micro Kit following the manufacturer’s instructions. RNA quantification was done by using a nanophotometer P-class (Implen) and integrity was confirmed using an Agilent 2100 Bioanalyzer for an Eukaryote total RNA pico assay (Agilent Technologies). Three biological replicates were produced with an equivalent number off cells and used in the microarray.
2 ng total RNA were used to produce amplified single stranded (ss) cDNA with the Ovation Pico WTA System (NuGEN Technologies). 3 µg of the ss cDNA were used to produce double stranded (ds) cDNA with the WT-Ovation Exon Module (NuGEN Technologies). Target labelling was performed using the One-Color DNA Labelling Kit (Roche NimbleGen, Inc.) with 1 µg ds cDNA as input material. All procedures were performed according to the manufacturer’s recommendations.
6 µg Cy3-labeled cDNA target was hybridized to Zebrafish Gene Expression 385K Arrays (Roche NimbleGen, Inc., Design ID 090506_Zv7_EXPR) following the instructions in the ‘NimbleGen Arrays User’s Guide: Gene Expression Analysis v3.2.
Normalized datasets (Gene Calls: _RMA.calls files) were created during data extraction in the software NimbleScan according to the ‘NimbleGen Arrays User’s Guide. Statistical data analysis was performed using the GeneSpring GX software (Agilent Technologies, Version 11.0.2). An unpaired T-test without multiple testing corrections was applied to the data and differences in expression were considered significant when p<0.05. A list of sequence IDs was created with sequences that had a fold change in transcription >2, and expression values > 8 in at least 3 out of the 6 datasets. The list of Ensembl gene IDs was later updated to a more current Ensembl release and genome build (Ensembl release-80, ZV10) by remapping the original probes against the updated transcriptome. The list of genes was then curated to remove those with repetitive probes, unspecific probes with more than one gene target or genes with insufficient probes mapping to them. The resulting gene list was then analysed with clusterProfiler, an R package for comparing biological themes among gene clusters (http://bioconductor.org/packages/release/bioc/html/clusterProfiler.html [Yu et al., 2012]). The data was also analysed with Cildbv2 (http://cildb.cgm.cnrs-gif.fr/, [Arnaiz et al., 2009]) and with Genevenn (http://genevenn.sourceforge.net/, [Pirooznia et al., 2007]) in order to filter gene targets associated with motility. These lists are in Table S1a and Table S1b in Supplementary file 1.
We performed quantitative PCR (qPCR) using cDNA from KV’s cells at bud stage (selected as in the microarray experiment). Total RNA from WT and dld−/− mutant embryos was isolated as previously described and reverse-transcribed using iScript cDNA Synthesis Kit (Bio-Rad) according to the manufacturer’s instruction. qPCR was performed on the CFX96 Real-Time PCR Detection System (BIO-RAD, Hercules, CA) using the SsoFast EvaGreen Supermix (BIO-RAD, Hercules, CA). Primers for amplification of foxj1a (forkhead box J1a), dnah7 (Dynein, Axonemal, Heavy Chain 7), rsph3 (Radial Spoke 3 Homolog), rfx2 (Regulatory Factor X, 2), and dld (delta D) were designed with Primer-BLAST (NCBI). Two reference genes sox17 (Sex Determining Region Y-Box 17) and eef1al1 (eukaryotic elongation factor 1 alpha 1 like 1) were used. Primers can be found in Table S1c in Supplementary file 1. All reactions were performed with two biological replicates and three technical replicates. Results were evaluated with the Bio-Rad CFX Manager 2.0 software (BIO-RAD, Hercules, CA). Significant differences in the transcription level were determined using either the Welsh t-test or the Mann–Whitney U-test (p<0.05), depending on the normality of the populations as established by the KS normality test.
Morpholino blocking translation of Foxj1a was used as previously described (Tian et al., 2009; Echeverri and Oates, 2007). Danio rerio her12 coding sequence was cloned into a pCS2 +vector with the primers 5’ TCAAGCTTCGAAATGGCACCCCACTCAGC (forward) and 5’ CTGGAGACCCTGGTAGTCTAGAAGCGGC (reverse). Danio rerio dnal1 coding sequence was cloned into a pCS2 +mCherry vector, with mCherry at the N-terminus of Dnal1, with the primers 5’ TCAAGCTTCGAAATGGCAAAAGCAACAACTATTAAAGAGGC (forward) and 5’ CGCTGGATCCTTAACTCTCCCCTTCAGTTTCC (reverse). her12, mCherry-Dnal1, Notch-intracellular domain (NICD) (Takke and Campos-Ortega, 1999), full-length foxj1a (Lopes et al., 2010), and Arl13b-GFP (a gift from Helena Soares) constructs were used to produced mRNA with the mMESSAGE mMACHINE kit (Ambion). The RNA was purified with the kit RNA Clean and Concentrator-5 (Zymo) and injected at one-cell stage at a concentration of 50 pg, 200 pg, 100 pg, 100 pg, and 400 pg respectively. Embryos were left to develop at 32°C or 25°C until the desired stage.
Total RNA was extracted from groups of zebrafish embryos at bud stage using the Qiagen RNeasy Mini Kit and reverse transcribed using both oligo(dT)18 and random hexamer primers with the RevertAid First Strand cDNA Synthesis Kit following the manufacturers' instructions. This procedure was repeated for the 3 biological replicates. Expression was quantified using Roche SYBR Green I Master and the PCR was run in a Roche LightCycler 96 Real-Time PCR System with 3 technical replicates for each biological replicate. Results were analysed and depicted as fold-change of transcript levels in injected embryos relative to transcript levels in control embryos. foxj1a, rfx4 (Regulatory factor X, 4), dnah7, dnah9 (dynein, axonemal, heavy chain 9) and her12 (Hairy-related 12) levels were normalized in relation to eef1al1 and rpl13a (ribosomal protein L13a) expression. Significant differences in the transcription level were determined using either the Welsh t-test or the Mann–Whitney U-test (p<0.05), depending on the normality of the populations as established by the Shapiro-Wilks normality test. Significant differences between different Foxj1a treatments in the same NS assay were established with a Kruskal–Wallis one-way analysis of variance (p<0.05). Primer sequences summarized in Table S1c in Supplementary file 1. A biological replicate results from the cDNA produced from the total RNA extracted from 25 zebrafish embryos at bud stage. The technical replicates use the same cDNA (of each biological replicate) in the qPCR reactions. All valid replicates are used in the statistical analysis.
Embryos were mounted live in 1% (w/v) low-melting agarose at either 2 or 7 somites stage, and covered with E3 medium. Live imaging was performed in a Prairie Multiphoton fluorescence microscope with Olympus 40x water immersion lens (NA 0.8) at 28°C. To assess motility, whole KVs were scanned with z sections of 0.5 μm, with an acquisition rate of 9.6 slices per minute (6.25 s per slice), which provided a pixel dwell time of 22,4 μs. Each experiment was repeated either two (WT vs Foxj1a OE; dld−/−;dlc−/− vs dld−/−;dlc−/− + Foxj1 a OE; WT vs NICD OE vs NICD OE +Foxj1 OE) or three times (WT vs Her12 OE). Each time an experiment was repeated, both control (WT) and tests were assayed in equal numbers until enough n was obtained. Time lapses were performed starting at the third somite stage and stacks were acquired every 30 min up until the sixth somite stage and 1 hr later at the eight somite stage. Since only two embryos could be imaged in this manner, this experiment was repeated 4 times in order to gather enough n.
Stacks were reconstructed and surfaces of the KVs were obtained by image segmentation in Imaris software (Bitplane, UK). Cilia were labelled with a red or blue dot based on whether they were motile or immotile, respectively. This was done for 4 embryos throughout time. Certain conditions were tracked through all time-points: immotile cilia that remained so from the first time point to the last, cilia that were motile from the beginning of the time-lapse (3 somites stage), cilia that started immotile, became motile and then stopped moving again, and cilia that started immotile and became motile as development progressed. In order to determine the 3D coordinates of the immotile cilia in the KV, we used the stacks obtained at 8 ss in the different treatments. These were reconstructed as described, the cilia were labelled and the Cartesian coordinates (x, y, z) of each cilium were calculated by establishing the centre of the Cartesian referential as the centre of the KV.
3D cilia length were measured using the ‘Simple Neurite Tracer’ plugin (Longair et al., 2011). In live imaged KV’s only immotile cilia were sampled for each condition from 8 ss embryos expressing arl13b-GFP (400 pg). In total we analysed: WT – 36 cilia, 11 embryos; Foxj1a OE – 18 cilia, 5 embryos; dld−/−;dlc−/− – 24 cilia, 6 embryos; dld−/−;dlc−/− + Foxj1a OE – 25 cilia, 7 embryos; NICD OE – 29 cilia, 8 embryos; NICD OE +Foxj1 a OE – 24 cilia, 7 embryos; Her12 OE – 18 cilia, 7 embryos. Specifically to compare WT and Her12 OE cilia length, 8 somite-stage embryos were fixed and the axonemal skeleton of the cilia was labeled with acetylated alpha tubulin. In this way we analyzed: WT - 138 cilia, 7 embryos; Her12 OE - 138 cilia, 7 embryos.
We followed the methods described previously (Sampaio et al., 2014) for mounting and filming embryos, and calculating the fluid flow and the cilia beat frequency. We tracked native particles and calculated the respective flow velocity with an R script (Supplementary file 2) in WT embryos from 3 to 8 ss. Embryos were kept at 32°C until desired stage and then filmed at room temperature. We imaged ciliary movement for beating frequency analysis in embryos overexpressing Foxj1a at 8 ss and the corresponding non-injected WT siblings (embryos kept at 25°C until desired stage and then filmed at room temperature). Each fluid flow and CBF measurement experiments were repeated once.
Zebrafish embryos with 14 hpf (10 ss) were fixed for 16 hr at 4°C in 0.1 M sodium cacodylate buffer, pH 7.3, containing 2,5% gluteraldehyde (v/v) enriched with sucrose and calcium chloride. After washings with sucrose enriched buffer the fragments were post-fixed for 1 hr (on ice) in 1% (aq.) osmium tetroxide and contrasted in block in 1% (aq.) uranyl acetate for 30 min. Dehydration was made using ethanol gradient (50-70-95–100%). Samples were embedded in bottle neck beem capsules (Ted Pella) using EPON resin (Electron microscopy sciences) and hardened at 60°C for 72 hr. After polymerization the tip containing the embryo was sawed and the embryo was re-embed in flat silicon mold for better orientation (KV parallel to the section plan). Resin blocks were sectioned using an ultramicrotome UC7 (Leica microsystems), semi-thin sections (300 nm) were stained with toluidine blue for optic light microscopy, semi-thin were collected until the KV start appearing. Ultra-thin sections (80 nm), were obtained systematically, 2 grids with 2 sections every 5–8 µm until the end of the vesicle. Sections were collected into formvar coated copper slot grids (AGAR scientific), and counter-stained with uranyl acetate and lead citrate (Reynold recipe), the whole KV was screened in a Hitachi H-7650 transmission electron microscope at 100kV acceleration, cilia were tilted in order to check for ultrastructural features, pictures were taken using a XR41M mid mount AMT digital camera.
We followed the methods described previously (Neugebauer et al., 2009) for immunostaining. The mixed immunofluorescence and in situ hybridization technique was adapted (Thisse and Thisse, 2008) as follows: on the second day, we added the antibody anti-GFP together with the antibody Anti-Digoxigenin-AP Fab Fragments (Roche) and incubated over night at 4°C in a horizontal rotator; on the third day, we added the secondary antibody anti-rabbit Alexa Fluor 488 and incubated over night at 4°C in a horizontal rotator; on the fourth day, we developed the RNA probe with Fast-Red substrate (Roche) until a red deposit was observed and the reaction was stopped with 4% PFA (in PBS) for 5 min. her12 RNA probe was a gift from Leonor Saude. Antibodies used were anti-acetylated α-tubulin (1:300; T7451 from Sigma), anti-GFP (1:500; ab290 from Abcam), anti-DlD (1:50; zdd2 monoclonal antibody [Itoh et al., 2003]), anti-mouse Alexa Fluor 564, and anti-rabbit Alexa Fluor 488 (1:500; Invitrogen). Nuclei were stained with DAPI (1:500). Flat-mounted embryos were examined with a Zeiss LSM 710 Meta confocal microscope and a Zeiss 40x water immersion C-Apochromat lens (1.2 NA). Three-colour confocal z-series images were acquired using sequential laser excitation, and analysed using Fiji software (LSM Reader) (Schindelin et al., 2012).
Whole-mount in situ hybridization was performed as described previously (Thisse and Thisse, 2008). Digoxigenin RNA probes were synthesized from DNA templates of dnah7 (Sampaio et al., 2014), dand5 (Hashimoto et al., 2004), foxa3 (Monteiro et al., 2008) and foxj1a (Yu et al., 2008). Images of flat-mounted embryos were acquired in a Zeiss Z2 Widefield Microscope with Zeiss air EC Plan-Neofluar 5x (0.16 NA) and 10x (0.3 NA) lenses. Whole-mount in situ hybridization to detect the expression of dand5 and foxa3.
Statistical analysis was performed with Prism 6 and R (Wilcoxon test and Fisher's Exact Test). Data populations were tested for normality with the Shapiro-Wilks or the KS normality tests and the different statistical tests were used accordingly. These are specified in the materials and methods section and on the Figures’ legend. The values presented are Means ± SD, unless stated otherwise.
Identification and analysis of axonemal dynein light chain 1 in primary ciliary dyskinesia patientsAmerican Journal of Respiratory Cell and Molecular Biology 33:41–47.https://doi.org/10.1165/rcmb.2004-0335OC
Rer1p maintains ciliary length and signaling by regulating γ-secretase activity and Foxj1a levelsThe Journal of Cell Biology 200:709–720.https://doi.org/10.1083/jcb.201208175
Primary ciliary dyskinesia caused by homozygous mutation in DNAL1, encoding dynein light chain 1The American Journal of Human Genetics 88:599–607.https://doi.org/10.1016/j.ajhg.2011.03.018
Origin and shaping of the laterality organ in zebrafishDevelopment 135:2807–2813.https://doi.org/10.1242/dev.022228
Fiji: an open-source platform for biological-image analysisNature Methods 9:676–682.https://doi.org/10.1038/nmeth.2019
ZFIN Direct Data Submission (Http://zfin.org) [InternetFast release clones: A high throughput expression analysis, ZFIN Direct Data Submission (Http://zfin.org) [Internet.
Both foxj1a and foxj1b are implicated in left-right asymmetric development in zebrafish embryosBiochemical and Biophysical Research Communications 380:537–542.https://doi.org/10.1016/j.bbrc.2009.01.111
A Guide for the Laboratory Use of Zebrafish (Danio Rerio)The Zebrafish Book, A Guide for the Laboratory Use of Zebrafish (Danio Rerio), 4th Edn, Eugene, University of Oregon Press.
clusterProfiler: an R package for comparing biological themes among gene clustersOMICS: A Journal of Integrative Biology 16:284–287.https://doi.org/10.1089/omi.2011.0118
Foxj1 transcription factors are master regulators of the motile ciliogenic programNature Genetics 40:1445–1453.https://doi.org/10.1038/ng.263
Jeremy F ReiterReviewing Editor; University of California, San Francisco, United States
In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.
Thank you for submitting your article "Notch/Her12 signalling sets the ratio of motile/immotile cilia independently of Foxj1a in zebrafish left-right organizer" for consideration by eLife. Your article has been favorably evaluated by Marianne Bronner (Senior Editor) and four reviewers, one of whom is a member of our Board of Reviewing Editors. The reviewers have opted to remain anonymous.
The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.
The manuscript by Taveres et al. addresses the role of Notch signaling in determining the ratio of motile/immotile cilia in Kupffer's Vesicle (KV) through its target, her12. The left-right organizers (LROs) in mouse, Xenopus and zebrafish are known to contain both motile and immotile cilia which has led to the two-cilia hypothesis, where two types of cilia are required to produce and sense flow during left-right patterning. Experiments in the Xenopus LRO by Boskovski (2013) put forth evidence that Notch signaling affects the ratio of motile to immotile cilia and this was proposed to occur by suppressing Foxj1 and RFX gene expression, critical regulators required for motile cilia to form. However, work, mainly by the Lopes lab, has shown that Notch signaling promotes longer cilia in zebrafish LRO, a finding seemingly at odds with those in Xenopus, since motile cilia tend to be longer than sensory cilia. The current manuscript attempts to resolve some of these issues by examining genes downstream of Notch signaling in the zebrafish LRO. The results presented are extensive. The overall impression is that the conclusions are not suitable for eLife but that upon revision the work may be.
The first two figures document when immotile and motile cilia arise developmentally as the LRO forms in Zebrafish and how this correlates with the appearance of flow. These experiments ground the subsequent analysis. The authors then present a transcriptomic analysis of Delta mutants, finding that Foxj1 and motile cilia components appear to be Notch regulated, but then the story gets complicated. First, in Figure 3A-D they argue that Foxj1 is not differentially regulated in the LRO, despite the differential expression in the microarray. The clear limitation with the experiment is that they use a minimal promoter driving GFP to carry out this analysis, and one wonders whether the regulation via Notch is simply not capture by this reporter. There is no attempt to show that the minimal promoter is a faithful proxy for endogenous foxj1a expression in the KV. Ideally Foxj1a protein would be tracked through an antibody staining, but if that is not possible then an in situ for foxj1a expression (similar to what was done for her12 in Figure 4) is necessary to show that foxj1a expression is uniform throughout the KV. Along the same lines, from the image shown in Figure 2A, there may be differing levels of foxj1a::GFP expression in some cells – could all cells be green, but there be significant differences in expression that could account for the reasonably subtle effects on motility that are observed?
In addition, the authors do not ask whether the expression of this reporter changes when Notch signaling is altered. They show qPCR results in Figure 4—figure supplement 1 but the results are also complicated: Foxj1a goes up in single Delta mutants but is largely unchanged in the double mutants, and NICD has no affect. Finally, the ratio between motile and immotile cilia in the LRO when Notch signaling or Foxj1 activity is manipulated barely changes in Figure 3G, even when Notch ICD, or her12, overexpression is used. This result is not compatible with the idea that Notch/Her12 signaling has a major impact on motile cilia formation. The simplest interpretation of these results is that neither Notch nor Fox and may suggest that the rather limited input seen with Notch is unlikely to be direct. The data shown in Figure 4 also do not provide much in the way of additional support for the author's conclusion that Notch/her12 is a major regulator of the motile/immotile decision. The number of her12 expressing cells goes up to 70% in a NICD experiment, but the number of immotile cilia only changes from the wild type by maybe 5% on average. The authors also focus on the number of anterior versus posterior immotile cilia in NICD injected embryo but again these results document subtle differences. Some comment should be made on the relative importance of Notch/her12 pathway compared to other regulators of ciliary motility.
Figure 3G, Imaging at 0.16 fps seems to slow to quantify reliably the rotational speed of motile cilia. This might be the reason that the authors did not detect any difference in H. The reference the authors provide (Sampiao et al) images at 500fps, which is a 10 fold oversampling in relation to CBF and sufficient to resolve cilia movement. The authors should repeat the experiment using faster imaging, as, if it turns out that the CBF is reduced after NICD or Her12 overexpression, the interpretation of the data will change. It would also be helpful to know the frame rates used in the other figures. For example, what frame rates for imaging motile cilia 4Q-V?
The authors present data in Figure 6, suggesting that a subtle change in the motile/immotile cilia ratio causes heterotaxy. However, the author's own work in Development (2010) shows that Notch signaling also affects the length of cilia, which one might predict would affect ciliary flow in KV. The authors have not examined cilia length in any of the experiments reported here, making it unclear how they dissociate length versus number of motile/nonmotile cilia in terms of generating the phenotype. The authors need to address the confounding issue of effects on ciliary length versus motility/immotility.
[Editors' note: further revisions were requested prior to acceptance, as described below.]
Thank you for resubmitting your work entitled "Notch/Her12 signalling modulates motile/immotile cilia ratio downstream of Foxj1a in zebrafish left-right organizer" for further consideration at eLife. Your revised article has been favorably evaluated by Marianne Bronner (Senior Editor), a Reviewing Editor, and one reviewer.
The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below:
One reviewer still has substantive concerns about this work. As you will read, most of the points can be addressed with textual changes. One point will require analysis of data that I suspect that you may have in hand. The final point, which the reviewer thinks is important, may require an additional loss of function experiment.
The revised manuscript by Lopes and colleagues has improved over the previous version, although technical difficulties prevent them from addressing several critical issues. The main strength of the manuscript is that it raises some interesting ideas about the nature of immotile cilia in the KV, and how they might arise. It further fuels the controversy about whether flow is sensed by immotile sensory cilia, or by another mechanism during left-right patterning. It also raises the possibility that Notch is influencing immotile cilia differentiation, in a mechanism that is independent of Foxj1, in contrast to what has been reported in Xenopus by the Khoka lab. The authors have included some additional data, like the TEM analysis, which adds further to the considerable amount of data included in the manuscript. At the same time, there are several issues I think still could be addressed to make this a manuscript suitable for eLife.
1) The first two figures document how the ratio of motile to immotile cilia changes over time during the formation of the KV. The recent eLife paper from Ferreira et al. (2017), which carried out a similar analysis, reaches significantly different conclusions. I am sympathetic to the author's plight, since their work is under review, as a competing manuscript is published. Nonetheless, I feel that there needs to be a more thorough discussion of the marked discrepancy between the two papers (now only briefly mentioned in passing in the Discussion), beginning in the Results section. Otherwise, we have two eLife papers reaching different conclusions about the number of motile versus immotile cilia in the KV, leaving the general reader with the impression that scoring these numbers is very unreliable. This discrepancy, for example, is critical when assessing the fact that manipulating Notch signaling only causes a very small number of cilia to switch from immotile to motile, or vice versa.
2) The previous review asked the authors to validate their conclusion based on the microarray analysis that decreasing Notch signaling in the KV (which leads to a decrease in immotile cilia) increases the expression of Foxj1 and some key targets of Foxj1 required for cilia motility. Apparently this is technically difficult to do since the promoter reporter doesn't recapitulate the normal expression of Foxj1 and the in situ protocol they are using does not have the resolution they need to examine expression in the small number of cells with immotile cilium, where this expression would be predicted to change. And so in the end, whether or not Foxj1 or motile gene expression is affected by Notch in the KV remains unknown. The expression data on Her12 seems clear. However that on foxj1 and dnah7 is not: microarray data reports a change, but then the authors seem to want to conclude they are not changed based on low-resolution detection methods that wouldn't necessarily detect the critical changes in a small percentage of cells. In addition, the quantitative PCR does not support a consistent change as predicted by microarray results. The authors could do a much better job summarizing the expression data than the current one written in the fifth paragraph of the subsection “Motile cilia fate decision is regulated by Notch signalling independently of Foxj1a”. In addition, the two supplemental figures (Figure 3—figure supplements 1 and 2) describing G0 term analysis of the microarray data seems not to present information that is informative. I would highly recommend removing these data from the manuscript, especially since the microarray data have not been validated in a meaningful way.
3) Figure 5 and 6 relate to the expression of Her12 as a Notch target gene that the authors conclude is required for setting the ratio of motile to immotile cilia. It is rather surprising that this conclusion is reached without a loss-of-function analysis, showing that Her12 normally plays a role in cilia motility and left-right patterning. This seems to be a major omission for a critical point made in the manuscript.
4) The DnaI1-GFP data in Figure 4G-I are potentially valuable since they support the author's suggestion that the immotile cilia in the KV are not simply sensory cilia, but have motile features. However, the analysis of Dnal1-GFP in Figure 4 cannot be published as is, and needs to be quantified by examining localization over a large number of cilia.
5) The authors previously showed that manipulating Notch in the KV affects overall cilia length and left-right patterning. In this paper, they argue that they bypass this effect by overexpressing Arl13b (which causes cilia lengthening) so that they can then assess how left-right patterning is affected by a change in the ratio of motile and immotile cilia upon alteration by Notch. This might be kosher, but how Arl13b promotes cilia lengthening is not well understood (maybe more membrane than axoneme?). By using Arl13b to correct cilia length, have they really corrected the effects of Notch on cilia length and function they previously reported? This technical limitation is significant and really needs to be discussed explicitly in a way that makes this caveat clear, so that the reader understands the limitations in the data in Figure 6.https://doi.org/10.7554/eLife.25165.037
- Susana Santos Lopes
- Sara Pestana
- Andreia Vaz
- Raquel Jacinto
- Barbara Tavares
- Pedro Sampaio
- Susana Santos Lopes
The funders had no role in study design, data collection and interpretation, or the on the decision to submit the work for publication.
The authors wish to thank Julian Lewis for the DeltaD antibody, Carl-Philipp Heisenberg for the gift of the transgenic zebrafish line Tg(sox17:GFP), Leonor Saúde for the gift of the transgenic zebrafish line Tg(foxj1a:GFP), the homozygous mutant DeltaC/beatm98/tm98, and the her12 in situ probe, and Helena Soares for the Arl13b-GFP construct. We would like to thank Idan Tuval, Domingos Henrique, Rita Teodoro and Adán Guerrero for insightful discussion of the manuscript. We also want to thank the IGC Advanced Image Facility, mainly to Gabriel Martins for advice and the use of Imaris software, and the IGC and CEDOC Fish Facilities. R.J. was supported by the Mechanisms of Disease and Regenerative Medicine (ProRegeM) PhD programme.
- Jeremy F Reiter, Reviewing Editor, University of California, San Francisco, United States
© 2017, Tavares et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.