The initial morphogenetic change that occurs during the budding of the tube of the salivary glands in the Drosophila embryos is the apical constriction of cells at the point of the future site of invagination, the pit (Figure 1A and A’; Girdler and Röper, 2014; Myat and Andrew, 2000a; Myat and Andrew, 2000b; Sanchez-Corrales et al., 2018). We recently provided the first quantitative morphometric analysis of the very early stages of this process, and revealed that, first, isotropic apical constriction near the future pit commences long before changes at the tissue level are apparent, and that, second, the apical constriction is indicative of a 3D cell behavior of wedging of cells at the pit (Figure 1A’). Importantly, we uncovered a strong regionalization of cell behaviors during the early stages of tube budding: cells near the invagination point predominantly showed apical constriction behavior while cells at a distance from the pit predominantly showed intercalation behavior (Sanchez-Corrales et al., 2018; Figure 1A). What is unclear is how the early tube initiation via isotropic apical constriction of cells, combined with directional intercalation of the cells still positioned away from the forming pit, evolves over time. The process of tube formation is characterized by the continuous invagination of cells and the formation of a symmetrical narrow-lumen tube from an asymmetric placode primordium. We can imagine two scenarios: in one, cell behavior is fixed at early stages of budding, that is, apical cell constriction is restricted to the pit and only apparent during early stages. In a second scenario, cell behavior depends on a cell’s position relative to the invagination pit and thus cell behaviors change as cells move toward the invagination point.

To distinguish between these scenarios, we segmented and tracked placodal cells from time-lapse movies covering a –40 min to +70 min time interval of the invagination process, with t=0 min defined as the first occurrence of tissue bending (see Figure 1—figure supplement 1 for time intervals of analyzed movies and cell numbers analyzed) and performed quantitative analyses (Figure 1B and C). At the latest time point covered, the majority of secretory cells had completely internalized and formed the internal tube. As calculated previously (Sanchez-Corrales et al., 2018), we employ a radial coordinate system centered at the invagination point at t=0 min. We specify the ‘near to the pit’ region as the region located between the invagination point and up to 33% of the distance between the invagination point and the far edge of the placode (with this region usually being within 13.7±3 µm of the pit), and with the rest of the cells defined as ‘far from the pit.’

Cells positioned near the pit at t=0 min (Figure 1B, pink) constricted their apices isotropically and rapidly and were internalized by about t=+ 20 min (Figure 1B and C and dashed curve in Figure 1E). Cells positioned far from the pit at t=0 min (Figure 1B, gray) and moving into a position closer to the invagination point began constricting (Figure 1B and D and solid curve in Figure 1E; Figure 1—video 1), suggesting that cells changed their behavior depending on where they were positioned in the placode with respect to the invagination point. Cells near the pit displayed isotropic constriction and thus a negative cumulative apical area change from the start (Figure 1E, pink highlighted), whereas cells initially positioned far from the pit did not change their apical area over the first ~20 min (Figure 1C, gray highlighted), but then began to display isotropic apical constriction (Figure 1E, blue highlighted). The cell behaviors and changes of cell behaviors were very apparent when we analyzed individual cells: cells located in the near pit region (Figure 1F, yellow and green cells) constricted progressively, whereas cells located in the region far from the pit at t=0 min (Figure 1F, blue and magenta cells) retained their apical area initially and only began constricting from t>+20 min onwards. As described previously (Sanchez-Corrales et al., 2018), cells far from the pit initially showed predominantly a cell intercalation behavior that constricts the tissue primordium circumferentially whilst expanding it radially, thereby moving cells closer to the invagination pit (Figure 1G). As they approached the invagination pit, these cells also started to apically constrict (Figure 1E solid line, highlighted in blue and Figure 1F).

These data strongly suggest that an initially pre-patterned behavior, isotropic constriction at the pit and intercalation further away, is dynamically adjusted during tube budding from a flat epithelial sheet, so that cells within a similar distance to the invagination pit display the same behavior.

Early apical constriction at the pit is driven by a highly dynamic apical-medial actomyosin network (Booth et al., 2014; Sanchez-Corrales et al., 2018). We next sought to determine whether the apical-medial actomyosin dynamics would differ among regions in the placode. We analyzed apical-medial actomyosin levels in cells across the placode, and firstly illustrate our findings in two snapshots: cells near the future invagination point at t=0 min showed strong accumulation of apical-medial myosin (Figure 2A, A’’ and C; see also Figure 2—video 1), whereas cells far from the pit and close to the anterior placode boundary (Figure 2A and A’) showed mainly junctional myosin II accumulation. At t=30 min, the cells now located right next to the invagination pit continued to display strong apical-medial myosin accumulation (Figure 2B and B’’), but in addition cells that were initially ‘far from the pit,’ but were now in a closer proximity to the pit though still near the anterior boundary, now also started to display apical-medial actomyosin (Figure 2B and B’). Thus, being moved into a position near the invagination pit appeared to determine levels of apical-medial myosin accumulation. Looking placode-wide over the whole time period analyzed, the peak of the apical-medial myosin fluorescence intensity tracked in a near-fixed position to the edge of the pit over time (Figure 2D, white dotted line). The location of the highest medial myosin intensity was also where cells showed the shortest cycle lengths of medial myosin II fluctuations (Figure 2E, white dotted line). Interestingly, cells near the pit at t=0 min showed an overall shorter myosin II cycle length (the time period elapsed between two peaks of highest myosin II intensity during the oscillation, see schematic in Figure 2E) compared to cells located far from the pit (Figure 2F). As longer cycle length correlates with unproductive myosin cycles (Booth et al., 2014), this was in line with cells further away from the pit showing less apical constriction. Furthermore, the strength of myosin oscillations can be expressed as the product of the amplitude and frequency of the oscillation (Booth et al., 2014). This myosin strength across the placode at t=0 min was much higher in cells near the invagination point when plotted against the distance to the pit, with a gradual decrease in myosin strength toward the placode boundary (Figure 2G). However, cells originally far from the pit at t=0 min increased their myosin strength over time (and concomitantly decreased their cycle length; see Figure 2—figure supplement 1) whilst moving into a position closer to the pit, so that by t=55 min their myosin strength had increased to levels previously shown by cells near the pit at t=0 min (Figure 2H).

Figure 2 with 2 supplements see all

Download asset Open asset

Thus, our data show that the early asymmetric setup of high apical-medial myosin II intensity and activity that was clustered near the pit was maintained during the continued invagination of the tube, in turn driving the continued apical constriction near the pit.

In order to uncover the role of the initial asymmetric patterning of cell behaviors within the placode, we turned our attention to mutant situations: we already knew that the transcription factor Fkh was important for invagination, as fkh^−/− mutants do not show localized apical constriction or any invagination (Myat and Andrew, 2000a; Sanchez-Corrales et al., 2018). A possibly more interesting situation is present in a previously published mutant of the transcription factor Huckebein (Hkb) that was reported to show a central invagination pit, combined with malformed invaginated salivary glands at later stages (Myat and Andrew, 2000b; Myat and Andrew, 2002).

We collected and segmented time-lapse movies of hkb^−/− mutant embryos over the same time period as the control wild-type embryos (Figure 3 and Figure 3—figure supplement 1; Figure 3—videos 1–3). Whereas in wild-type embryos, at t ~ +40 min, most secretory cells of the salivary gland placode have already invaginated (Figure 3A’’’, B’’’ and C’’’), in the hkb^−/− mutants no invagination had occurred yet (Figure 3D’’’, E’’’ and F’’’). Eventually, these embryos invaginated a dilated tube through an enlarged pit (see Figure 3—figure supplement 2). We analyzed the spatial distribution of apical area changes over time in wild-type and hkb^−/− placodes (Figure 3G and H). In the wild-type, the area of strongest apical area constriction was always located in a position near the pit, similar to the distribution of medial myosin intensity and activity (Figure 2D and E). In contrast, in hkb^−/− mutant placodes cells began to constrict in an aberrant central position from about t=+20–30 min (Figure 3G and H). Thus, hkb^−/− mutants show a delayed central invagination pit compared to the eccentric pit observed in wild-type placodes (Figure 3(I)1’). For our further quantitative analyses, we now took this change to a central pit position in the mutants into account. We plotted cumulative apical area change for cells in hkb^−/− mutant placodes, split into cells near the (now central) pit and cells far from the pit (at the placode periphery). This analysis showed that cells in hkb^−/− mutant placodes located near the central pit started to constrict at a much slower rate than cells in wild-type placodes, and only from about t=+20 min on (Figure 3J, wild-type data reproduced from Figure 1E). In fact, the rate of the constriction of the central cells in hkb^−/− mutants was very similar to that of the cells located at a distance to the pit in the wild-type (compare blue dashed and magenta solid curves), suggesting that the central cells in hkb^−/− mutants continue to behave identical to the wild-type cells occupying the same position. Cells far from the central pit in hkb^−/− mutants showed a slight apical area expansion up to t=+20 min followed by a slight constriction, but did not invaginate over the time interval shown in the plots. In agreement with the delayed central constriction of cells in hkb^−/− mutant placodes, we found no accumulation of apical-medial actomyosin at the eccentric position where a pit forms in the wild-type (compare Figure 3E and K to Figure 2D). Instead, only at t~+40 min did a central group of cells in hkb^−/− mutant placodes display increased accumulation of apical-medial myosin II (Figure 3K), inducing a shallow tissue bending in this region at this point (see also Figure 3—figure supplement 2).

Figure 3 with 6 supplements see all

Download asset Open asset

Thus, in comparison to wild-type placode with an eccentric invagination, in hkb^−/− mutant placodes a central invagination formed with a significant delay in time.

The above analyses strongly suggest that the initial asymmetrical setup of cell behaviors in the placode might be disrupted in hkb^−/− embryos. We therefore performed a strain (deformation) rate analysis of placodes in hkb^−/− mutants in comparison to wild-type embryos to assess this quantitatively. The strain rate analysis is based on the assumption that any change in tissue shape can be accounted for by changes in cell shape and cell intercalations (Blanchard et al., 2009; Sanchez-Corrales et al., 2018). In many tissues, cell division or cell death and delamination also contribute to strain rates, but these are absent in the salivary gland placode during invagination and can thus be neglected. The origin of the radial coordinate system used for this analysis was specified at t=0 min and positioned at the center of of the placode in the hkb^−/− mutant embryos (where the central constriction eventually occurred) and in wild-type embryos at the dorsal-posterior corner of the placode (where the wild-type pit will form). In wild-type placodes, cells that were located near the eccentric invagination pit (Figure 4A; with the ‘near’ position defined at t=0 min; see also Figure 4—figure supplement 1) constricted isotropically, with near equal change in strain at the tissue and cell shape level, and very little intercalation strain (Figure 4A’ and A’’, solid curves). These cells in wild-type placodes could only be tracked up to t=20 min as at this point they had all invaginated into the embryo to form the initial part of the tube. In contrast, in the hkb^−/− mutants, cells near the forming central pit (Figure 4A; defined to be in a ‘near’ position at t=0 min) showed a much reduced tissue and cell shape strain rate, with near identical radial and circumferential contributions, and constriction only commenced at about half the rate of the wild-type placodes from t=+20 min onwards (Figure 4A’ and A’’, dashed curves; see also Figure 4—figure supplement 1). In wild-type placodes, cells far from the eccentric pit (Figure 4B; defined to be ‘far’ at t=0 min), showed cell intercalations leading to circumferential convergence and radial extension of the tissue over the first approximately 20 min after invagination commenced, as we also reported previously (Figure 4B’ and B’’, solid curves; tissue and intercalation strain rate showing expansion radially and contraction circumferentially; Sanchez-Corrales et al., 2018). Interestingly, we then observed a clear switch in cell behavior of these cells to a second phase after t=+20 min. Once these cells were located closer to the pit position, they now also displayed apical contraction (beyond t=+20 min; Figure 4B’ and B’’, solid green curves). In the hkb^−/− mutants, for cells at the placode periphery and hence far from the central constriction (Figure 4B), the tissue expanded slightly in both radial and circumferential direction (Figure 4B’ and B’’, dashed gray curves), with cell shapes narrowing radially (dashed green curve in Figure 4B’) and extending circumferentially (dashed green curve in Figure 4B’’). We could observe a similar rate of intercalation in these cells as in cells far from the pit in wild-type placodes, with some radial elongation and circumferential contraction (dashed orange curves in Figure 4B’ and B’’). As the tissue strain rate in the hkb^−/− mutants showed no overall signature of directional changes with only a slight expansion both radially and circumferentially, we were curious to understand better what these tissue-scale signatures could represent at the individual cell level.

Figure 4 with 1 supplement see all

Download asset Open asset

We therefore analyzed individual events of neighbor gains across the tissue in wild-type and hkb^−/− mutant embryos. We compared two types of intercalation events, those leading to circumferential neighbor gains that in wild-type placodes explain the observed intercalation strain rates, and those leading to radial neighbor gains, and thus potentially opposing the changes seen at the tissue and intercalation strain rate level (Figure 4C). In wild-type placodes, circumferential neighbor gains outweigh radial ones by a large margin throughout (Figure 4D, dashed curves), leading to a steady increase in cumulative productive neighbor gains for cells across the placode (Figure 4E, yellow curves). In contrast, in hkb^−/− mutant embryos circumferential and radial neighbor gains across the placode were overall much reduced in number and occurred with equal frequency (Figure 4D, solid curves). Furthermore, cells near the delayed central pit in hkb^−/− mutants showed no productive neighbor gains (defined as the difference between circumferential and radial neighbor gains, corrected for the proportion of interfaces that could successfully be tracked; Figure 4E, purple dashed curve), whereas cells at the periphery away from the central pit showed productive neighbor gains (Figure 4E, solid purple curve). Active neighbor gains in the placode and other morphogenetic events are initiated and driven by junctional actomyosin (Sanchez-Corrales et al., 2018; Tetley et al., 2016), prompting us to analyze junctional myosin polarisation in hkb^−/− mutant embryos. We previously reported on the unipolar and bipolar enrichment of junctional myosin in the wild-type at early stages of placode morphogenesis (Sanchez-Corrales et al., 2018) and expand it here to the whole process of secretory cell invagination in comparison to hkb^−/− mutants (Figure 4F and G). Whereas in control placodes junctional myosin is clearly enriched in circumferential junctions by both measures, up to the point of most secretory cells having internalized by t=+40 min (Figure 4F and G; dashed green curve vs. dashed purple curve), in hkb^−/− mutant placodes myosin unipolarity and bipolarity are strongly reduced, even though some bipolarity remained (Figure 4F and G, solid curves). Thus, the major driver of active intercalations appeared to be strongly diminished in the mutants, posing the question of what could drive the observed intercalations here. We also analyzed cell elongation in cells far from the pit, with cells in wild-type placodes being initially elongated circumferentially at t<+20 min, while early intercalation events are initiated by circumferential junctional myosin (Figure 4H, gray curve; Sanchez-Corrales et al., 2018). These cells then became increasingly elongated radially as they moved into a position nearer to the pit and were being actively funneled toward the pit. In hkb^−/− mutant embryos, cells far from the pit at the periphery of the placode very quickly became circumferentially elongated and remained so throughout the time period analyzed. This aberrant circumferential elongation together with the mild bipolarity of junctional myosin polarisation that we observed prompted us to analyze individual events of cell intercalation in time-lapse movies (Figure 4I–L). In cells far from the central pit in hkb^−/− mutants, we could observe groups of cells clearly undergoing intercalations (Figure 4I and J) with individual cells making circumferential neighbor gains and elongating (cells 1 and 2 in Figure 4K), whilst the overall shape of the local group of cells undergoing the intercalation event was not changing (Figure 4L). Thus, although cells far from the aberrant central pit in hkb^−/− mutants still intercalated and gained circumferential neighbors, this behavior lacked several key signatures of wild-type intercalations, and overall in the hkb^−/− mutants intercalations did not lead to convergence and extension of the tissue toward the misplaced pit.

These analyses show that in the absence of Hkb, the initial radial patterning of cell behaviors across the placode is highly aberrant. Cells located at the position corresponding to where the eccentric pit would be located in the wild-type clearly failed to constrict apically, suggesting that Hkb could be involved in the specification of this pit behavior observed in wild-type.

The loss of correct patterning of cell behaviors in the early placode in hkb^−/− mutant embryos suggested that Hkb was involved in establishing these in wild-type embryos. Furthermore, hkb mRNA expression has been reported to show a dynamic pattern by in situ hybridization and β-gal reporter labeling (Myat and Andrew, 2000b). In order to be able to analyze Hkb protein levels dynamically and in comparison to cell shapes, we generated a Venus-tagged version of Hkb using CRISPR/Cas9 to tag the endogenous protein. Using this in vivo reporter, we detected Venus-Hkb at high levels concentrated in the dorsal-posterior region of the just forming placode, in the area where the invagination pit will form, already at t=–30 min prior to the first tissue bending (Figure 5A, A’, D and M and Figure 5—figure supplement 1B,B'' ), and even as early as t=–63 min prior to invagination (Figure 5—figure supplement 1A). The expression then expanded more broadly and anteriorly across the whole placode over time (Figure 5B–C’, E, F and M and Figure 5—video 1). At the moment of first tissue bending (t=0 min) Venus-Hkb was found across the placode, with elevated levels still at the invagination pit position, and a second increase toward the anterior. At late stages (t>>0 min) Venus-Hkb levels close to the invagination point were lower than further anterior (Figure 5M).

Figure 5 with 3 supplements see all

Download asset Open asset

As mentioned above, embryos mutant in fkh also show a lack of early apical constriction at the dorsal-posterior corner of the placode (Myat and Andrew, 2000a; Sanchez-Corrales et al., 2018). Similar to Hkb, analysis of fkh mRNA and protein levels in fixed embryos indicated an early enrichment near the forming invagination pit (Myat and Andrew, 2000a; Sanchez-Corrales et al., 2018). Using an in vivo genomic GFP-Fkh reporter (Spokony and White, 2013), we followed Fkh protein levels and patterning over the same time period whilst also assessing apical cell shapes (Figure 5G–L). Already at t=–60 min prior to tissue bending, Fkh-GFP was visible in the cells near the future invagination pit (Figure 5G, G’, J and N and Figure 5—figure supplement 1C–C''), from where it spread anteriorly and ventrally across the whole placode (Figure 5—video 2). At t=0 min, Fkh-GFP was already detected in all cells of the placode, though with levels still highest near the pit and decreasing toward the anterior edge of the placode (Figure 5H, H’, K and N). At t=+30 min and beyond, levels of Fkh-GFP were more uniform across the placode (Figure 5I, I', L, N).

Thus, both Hkb and Fkh transcription factors showed a clear temporally and spatially graded pattern of expression within the early placode that developed dynamically. Changes were especially pronounced at the dorsal-posterior corner, with highest levels here already an hour before tissue bending commenced. Combined Hkb and Fkh action could thereby define the future position of the invagination pit. We therefore compared Venus-Hkb as well as Fkh-GFP fluorescence intensity levels from time-lapse movies that in parallel allowed us to determine the corresponding apical area of cells (Figure 5O–P’). Venus-Hkb intensity levels peaked in cells located at the site of the forming eccentric invagination pit at about t ~ –30 min (Figure 5O), with an appearance of smaller apical areas in this position following on about 30–40 min later (Figure 5O’). Similarly, Fkh-GFP intensity began to peak at about the same time, t ~ –30 min, and then tracked at high level at a fixed distance to the pit over time (Figure 5P). Again also here, the increased reduction in apical area appeared with a delay of about 30–40 min (Figure 5P’). Such a delay would be consistent with the estimated time required for the transcription of the fog locus, followed by translation, protein folding, secretion, binding to the receptor, and signaling leading to myosin phosphorylation (Garcia et al., 2013; Shamir et al., 2016).

Thus, Hkb and Fkh dynamic expression at the future eccentric pit could prefigure the start of apical constriction at this site. The continuous high levels of Fkh tracking at a fixed distance to the pit suggest that Fkh is key to the continuous apical constriction of cells moving into a near pit position. These dynamic changes might trigger the graded and patterned expression of targets of both transcription factors, leading to the observed patterned cell behaviors we uncovered, as well as the switch between intercalation and constriction.

With the transcription factors Fkh and Hkb both displaying intriguing protein expression patterns across the early placode pre-morphogenesis, targets of both factors are likely to play key roles in instructing cell behaviors such as wedging and cell intercalation. Interestingly, an upstream activator of apical-medial actomyosin activity, the GPCR-ligand Folded gastrulation (Fog) was found to be dependent on Fkh (Chung et al., 2017; Dawes-Hoang et al., 2005; Figure 6) and suggested to be downstream of Hkb (Myat and Andrew, 2000b; though in this publication as ‘data not shown’). Fog is an apically secreted ligand that acts in an autocrine fashion, and thus the cells that switch on Fog expression will apically constrict (Dawes-Hoang et al., 2005). Previously published in situ hybridization of fog mRNA suggested increased levels near the forming pit, similar to the early Hkb and Fkh enrichment (Nikolaidou and Barrett, 2004). Fog expression and signaling in particular could not only be upstream of the isotropic constriction of cells near the forming pit, but could also be involved in maintaining this behavior in new coronas of cells moving into a position near the pit over time.

Figure 6 with 1 supplement see all

Download asset Open asset

We therefore decided to carefully analyze Fog protein levels in comparison to apical area and apical-medial myosin intensity in salivary gland placodes from early to late invagination stages (Figure 6A–I). Fog levels were low at mid-stage 10 concomitant with placode specification (Figure 6A, A’ and C). Just before tissue bending commenced at t=0 min, Fog was already strongly enriched in cells near the forming invagination pit, and these cells were the ones already showing the decreased apical area (Figure 6D–F). Fog protein levels remained enriched near the invagination point once the tube had started to internalize (Figure 6G–I). In fact, Fog protein levels in cells were positively correlated with both smaller apical areas (Figure 6K) as well as higher average apical-medial myosin II intensity (Figure 6L; Figure 6—figure supplement 1).

hkb^−/− mutant embryos, as detailed above, showed a very clear disruption of the early asymmetric placodal patterning. At late stage 11, Fog protein levels in hkb^−/− mutant embryos appeared to be similar to wild-type embryos and enriched in the placode compared to the surrounding epidermis, though the pattern of Fog enrichment was altered. Cells at the posterior corner, where the invagination pit would have formed in wild-type placodes, lacked high Fog protein levels, whereas only cells in the center of the placode now showed high levels of Fog protein (Figure 6M–O). When quantitatively analyzing many hkb^−/− mutant embryos in comparison to wild-type, the peak intensity of Fog protein was clearly shifted to a central, much more symmetrical, position (Figure 6P). In contrast, although we could detect some Fog protein in fkh^−/− embryos at late stage 11, it was no longer enriched at the position where the invagination pit forms in the wild-type. The levels also did not appear raised anywhere within the placode compared to the surrounding epidermis (Figure 6Q-S and Figure 6—figure supplement 1 D-D’’), as the quantification of many placodes confirmed (Figure 6T). In both mutants, altered levels and patterning of Fog protein still correlated with the now altered levels and patterns of residual apical-medial myosin (Figure 6M and Q).

In order to understand the cause of the central and delayed constriction in hkb^−/− mutant embryos upstream of the also only centrally localized Fog, we examined Fkh protein localization in hkb^−/− mutants (Figure 6—figure supplement 1E-J). At late stage 10, when the Fkh protein expression pattern is just expanding from the eccentric pit position to the central part of the placode in the control, the pattern was indistinguishable in the hkb^−/− mutants (Figure 6—figure supplement 1 E-F’ and H-I’). The pattern remained identical to wild-type placodes throughout stage 11 and beyond (Figure 6—figure supplement 1 G,G’, J, J’). Hence the natural expansion of Fkh protein expression to the cells in the center of the placode that is also observed in the hkb^−/− mutant placodes might be the cause of the delayed constriction of the cells in this region, whereas the earlier initial constriction at the site of the eccentric pit was absent in the mutant because it requires both Fkh and Hkb. Both transcription factors appeared in fact to be independently regulated, as also Hkb levels and protein expression pattern were unchanged in fkh^−/− mutant embryos (Figure 6—figure supplement 1 K-N’).

Thus, the asymmetrical setup of the salivary gland placode prior to initiation of tube budding relies on a correct spatial and temporal patterning of the key transcription factors, Hkb and Fkh. These transcription factors are in turn upstream of GPCR/Fog signaling that leads to patterned and sustained apical-medial myosin activity and hence contractile cell behaviors.

Our data and previously published results show that the asymmetrical invagination point depends on patterning by transcription factors and downstream GPCRs/Fog signaling at the posterior corner, leading to a focussed and point-like invagination pit (Chung et al., 2017; Myat and Andrew, 2000a; Myat and Andrew, 2000b; Myat and Andrew, 2002; Sanchez-Corrales et al., 2018). We next sought to test the importance of this asymmetric setup for successful morphogenesis of a narrow-lumen and symmetrical tube. To do so, we overexpressed Fog in the salivary gland placode only, using UAS-Fog x fkhGal4 which led to placode-wide increased levels of Fog from early stages on (Figure 7 and Figure 7—figure supplement 1). Compared to the focused apical constriction near the forming invagination pit in wild-type placodes at early stage 11 (Figure 7A and A’), when Fog was overexpressed across the placode most secretory cells of the placode constricted simultaneously (Figure 7B, B’, C and C’). At late stage 11, when in wild-type placodes cells had begun to invaginate and a short symmetrical narrow-lumen tube had formed (Figure 7D and D’), in UAS-Fog x fkhGal4 embryos the apically constricting cells had formed a large shallow depression across the whole dorsal part of the placode, with the beginning of a deeper invagination at the dorsal-posterior corner (Figure 7F and F’). Following on from this, at stage 12, in contrast to the narrow opening to the invagination in wild-type placodes (Figure 7E), in UAS-Fog x fkhGal4 placodes this whole section usually formed a deepened very wide pit (Figure 7G). This widened tube in Fog-overexpressing embryos, at stages 15–16 when all cells had invaginated, displayed a significantly enlarged and flattened lumen compared to wild-type tubes (Figure 7H–J). Interestingly, overexpression of Fkh using UAS-Fkh x fkhGal4 led to a similarly enlarged area of apical constriction early on and invaginated tubes with aberrant and inflated lumens at late stages (Figure 7—figure supplement 2).

Figure 7 with 2 supplements see all

Download asset Open asset

We set out to explain the link between an enlarged or widened invagination pit in the placode early during tube formation and the expanded lumen tube observed at late stages by considering why the invagination pit forms in its eccentric position in wild-type placodes. We formulated a simplistic theoretical consideration that assumes that following the specification of an initial group of cells that apically constrict and form the initial invagination (Figure 7K–M), neighboring coronae of cells will follow and invaginate (Figure 7K’–M’). Using an idealized version of apices of a wild-type placode with an eccentric pit, the number of cells invaginating, going from corona to corona, remains very steady and low (Figure 7K’ and green curve in Figure 7N). In contrast, a centrally located pit (Figure 7L) would lead to an ever-increasing number of cells per corona that would need to invaginate (Figure 7L’ and magenta curve in Figure 7N). An enlarged pit, a combination of the two previous scenarios and resembling aspects of the enlarged pit and invagination seen in UAS-Fog x fkhGal4 embryos (Figure 7M and M’), would lead to the invagination of a large number of cells initially, followed by similarly large numbers in each corona (Figure 7M’ and gray curve in Figure 7N). In fact, when counting invaginating coronae of cells in frames of time-lapse movies of wild-type (Figure 7O, green curve), hkb^−/− (Figure 7O, magenta curve) curve and UAS-Fog x fkhGal4 (Figure 7O, gray curve) embryos, the actual numbers observed were in very good agreement with the theoretical considerations outlined above (Figure 7K–N).

In the hkb^−/− mutant embryos, the delayed central area of constriction turned into a widened invagination and the enlarged pit at late stages often collapsed into an enlarged crescent of subduction at the boundary (Figure 3—figure supplement 2 and Figure 3—video 3; Figure 3—video 4), leading to malformed and enlarged lumen glands at the end of embryogenesis, as has been observed previously (Myat and Andrew, 2002).

Thus, the eccentrically placed invagination pit in the dorsal-posterior corner of the salivary gland placode is a key factor in ensuring the formation of a symmetrical narrow lumen tube during invagination, in part likely due to the control over the number of invaginating cells in each corona destined to internalize.

The following transgenic fly lines were used: Armadillo-YFP (CPTI collection described in Lye et al., 2014), sqhAX3; sqh::sqhGFP42 (Royou et al., 2004) and fkhGal4 (Henderson and Andrew, 2000; Zhou et al., 2001) (kind gift of Debbie Andrew); UAS-Fog (kind gift of Thomas Lecuit); y¹ w^* cv¹ sqh^AX3; P[w^+mC = sqh-GFP.RLC]C-42 M[w^+mC = Ubi-TagRFP-T-CAAX]ZH-22A (Kyoto DGRC number 109822, referred to as sqh^AX3;sqhGFP; UbiRFP); [sqhGFP42,UbiRFP; fkh[6]/TM3 Sb Twi-Gal4::UAS-GFP] (fkh[6] allele from Bloomington); [hkb², p², Scribble-GFP]; [sqhGFP42,UbiRFP; hkb², p²/ TM3 Sb Twi-Gal4::UAS-GFP] hkb² allele from Bloomington BDSC: 5457, [UbiRFP, Venus-HKB] (this study), FkhGFP (Bloomington BDSC:43951) and [FkhGFP; wgGal4, UAS-palmYFP] (generated from membrane Brainbow; Förster and Luschnig, 2012; Hampel et al., 2011). See Table 1 for details of genotypes used for individual figure panels.

• The CRISPR Venus-HKB was created using the following gRNA targets flanking the N-terminus of Hkb:

• CACCGCAAACCTACTCGCGACTT.

• A linker sequence between the inserted fluorescent protein and the protein was also used:

• GGAGGCGGAGGCTCGGGAGGCGGAGGCTCG.

Embryos were collected on apple juice-agar plates and processed for immunofluorescence using standard procedures. Briefly, embryos were dechorionated in 50% bleach, fixed in 4% formaldehyde, and stained with primary and secondary antibodies in PBT (PBS plus 0.5% bovine serum albumin and 0.3% Triton X-100). Anti-Hkb was a gift from Chris Doe (McDonald and Doe, 1997) anti-Fkh was a gift from Herbert Jäckle (Weigel et al., 1989) and anti-Fog was a gift from Naoyuki Fuse (Fuse et al., 2013), PY-20, (*P11120, Transduction Laboratories); anti-CrebA (DSHB). Secondary antibodies used were Alexa Fluor 488/Fluor 549/Fluor 649 coupled (Molecular Probes) and Cy3 coupled (Jackson ImmunoResearch Laboratories). Samples were embedded in Vectashield (Vector Labs).

Images of fixed samples were acquired on an Olympus FluoView 1200 or a Zeiss 780 Confocal Laser scanning system as z-stacks to cover the whole apical surface of cells in the placode. Z-stack projections were assembled in ImageJ or Imaris (Bitplane), 3D rendering was performed in Imaris.

For live time-lapse experiments (see Table 1), embryos were dechorionated in 50% bleach and extensively rinsed in water. Stage 10 embryos were manually aligned and attached to heptane-glue coated coverslips and mounted on custom-made metal slides; embryos were covered using halocarbon oil 27 (Sigma-Aldrich) and viability after imaging after 24 hr was controlled prior to further data analysis. Time-lapse sequences were imaged under a 40×/1.3NA oil objective on an inverted Zeiss 780 Laser scanning system, acquiring z-stacks every 0.58–3 min with a typical voxel xyz size of 0.22×0.22×1 μm³. Z-stack projections to generate movies in Supplementary Material were assembled in ImageJ or Imaris. The absence of fluorescent Twi-Gal4::UAS-GFP was used to identify homozygous hkb[2] and fkh[6] mutant embryos. The membrane channel images from time-lapse experiments were denoised using nd-safir software (Boulanger et al., 2010).

Cell segmentation and tracking were performed using custom software written in Blanchard et al., 2009. First, the curved surface of the embryonic epithelium was located by draping a ‘blanket’ over all image volumes over time, where the pixel-detailed blanket was caught by, and remained on top of binarised cortical fluorescence signal. Quasi-2D image layers were then extracted from image volumes at specified depths from the surface blanket. This step permitted to account for the curvature of the embryos in these quasi-2D projections. We took maximum intensity projections of a small number of near-surface image layers typically at 0–4 μm from membrane channels, for cell-apical analyses, apicomedial myosin and anti-Fog fluorescence channels and projections of image layers typically from 5 to 16 μm encompassing nuclear fluorescence for Venus-HKB and Fkh-GFP analyses. Membrane channels were filtered with median, top-hat, or high/low frequency filters as necessary to optimize subsequent cell tracking.

Cells in membrane channels were segmented using an adaptive watershedding algorithm as they were simultaneously linked in time. Manual correction of segmented cell outlines was performed for all fixed and time-lapse data. The segmentation of all the movies used in this study was manually corrected to ensure at least 90% tracking coverage of the placode at all times. Tracked cells were subjected to various quality filters (lineage length, area, aspect ratio, and relative velocity) so that incorrectly tracked cells were eliminated prior to further analysis. The number of embryos analyzed and the number of cells can be found in Figure 1—figure supplement 1, Figure 3—figure supplement 1.

Wild-type movies were aligned in time using as t=0 min, the frame just before the first sign of invagination at the future tube pit was evident. hkb^−/− mutants were aligned using as a reference of embryo development the level of invagination of the tracheal pits that are not affected in the hkb^−/− mutant as well as other morphological markers such as appearance and depth of segmental grooves in the embryo. Cells belonging to the salivary placode (without the future duct cells that comprise the two most ventral rows of cells in the primordium) were then manually outlined at t=0 min using the surrounding myosin II cable as a guide and ramified forwards and backwards in time. Only cells of the salivary gland placode were included in subsequent analyses.

At t = 0 min, the center of the future tube pit was specified manually as the origin of a radial coordinate system, with radial distance (in µm) increasing away from the pit. Circumferential angle was set to zero toward posterior, proceeding anti-clockwise for the placode on the left-hand side of the embryo, and clockwise for the placode on the right so that data collected from different sides could be overlaid.

The radial coordinate system was ‘mobile,’ in the sense that its origin tracked the center of the pit, forwards and backwards in time, as the placode translated within the field of view due to embryo movement or to on-going morphogenesis.

As calculated previously (Sanchez-Corrales et al., 2018), we employ a radial coordinate system centered at the invagination point at t=0 min. We specify the ‘near to the pit’ region as the region located between the invagination point and up to 33% of the distance between the invagination point and the far edge of the placode (with this region usually being within 13.7±3 µm of the pit), and with the rest of the cells defined as ‘far from the pit.’

Detailed spatial patterns of the rates of deformation across the placode and over time quantify the net outcome of active stresses, viscoelastic material properties, and frictions both from within and outside the placode. We quantified strain (deformation) rates over small spatio-temporal domains composed of a focal cell and one corona of immediate neighbors over a ~3-min interval (Blanchard et al., 2009) and reviewed in Blanchard, 2017. On such 2D domains, strain rates are captured elliptically, as the strain rate in the orientation of greatest absolute strain rate, with a second strain rate perpendicular to this.

For the early morphogenesis of the salivary gland placode, in which there is no cell division or gain/loss of cells from the epithelium, three types of strain rates can be calculated. First, total tissue strain rates are calculated for all local domains using the relative movements of cell centroids, extracted from automated cell tracking. This captures the net effect of cell shape changes and cell rearrangements within the tissue, but these can also be separated out. Second, domain cell shape strain rates are calculated by approximating each cell with its best-fit ellipse and then finding the best mapping of a cell’s elliptical shape to its shape in the subsequent time point, and averaging over the cells of the domain. Third, intercalation strain rates that capture the continuous process of cells in a domain sliding past each other in a particular orientation, are calculated as the difference between the total tissue strain rates and the cell shape strain rates of cells. Strain rates were calculated using custom software written in L3Harris Geospatial IDL (code provided in Blanchard et al., 2009 or by email from G.B.B.).

The three types of elliptical strain rates were projected onto our radial coordinate system (Sanchez-Corrales et al., 2018) using the eccentric pit location in wild-type (dorsal-posterior corner) or the center of the misplaced invagination point in hkb^−/− mutants. The radial coordinate system permitted us to analyze radial and circumferential contributions. Strain rates in units of proportional size change per minute can easily be averaged across space or accumulated over time. We present instantaneous strain rates over time for spatial subsets of cells in the placode (Figure 4—figure supplement 1), and cumulative strain ratios for the same regions over time. These plots were made from exported data as done previously in Sanchez-Corrales et al., 2018. To test for differences in instantaneous strain rates of wt and hkb^−/− time-lapse movies we used a mixed-effects model implemented in R (lmer4 package as in Butler et al., 2009; Lye et al., 2015; Sanchez-Corrales et al., 2018) with a significance threshold of p<0.05. The phenotype (wt or hkb^−/−) was considered a fixed effect while the variation between embryos from the same phenotype was considered a random effect.

Cell shape elongation (Figure 4H) was calculated by fitting an ellipse into each cell. The two axes of each ellipse were projected into radial coordinates and the ratio of them was calculated.

We used changes in neighbor connectivity in our tracked cell data to identify neighbor exchange events (T1 processes). Neighbor exchange events were defined by the identity of the pair of cells that lost connectivity in t and the pair that gained connectivity at t+1. The orientation of gain we defined as the orientation of the centroid-centroid line of the gaining pair at t+1. We further classified gains as either radially or circumferentially oriented, depending on which the gain axis was most closely aligned to locally. We did not distinguish between solitary T1s and T1s involved in rosette-like structures.

From visual inspection, we knew that some T1s were subsequently reversed, so we characterized not only the total number of gains in each orientation but also the net gain in the circumferential axis, by subtracting the number of radial gains. Furthermore, when comparing embryos and genotypes, we controlled for differences in numbers of tracked cells by expressing the net circumferential gain per time step as a proportion of half of the total number of tracked cell-cell interfaces in that time step. We accumulated numbers of gains, net gains, and proportional rate of gain over time for WT and hkb embryos (Figure 4D and E). Two sample Kolmogorov-Smirnov tests were used to determine significance at p<0.05 for data in Figure 3.

We extracted a quasi-2D layer image from the myosin II channel at a depth typically from 0 to 3 μm that maximized the capture of apical-medial myosin. We background-subtracted the myosin images and quantified the average intensity of apical-medial myosin as the fluorescence inside the segmented cell, excluding the fluorescence along cell-cell interfaces encompassing the cell edge pixels plus two pixels in a perpendicular direction on either side (considered junctional myosin).

The analysis of apical-medial fluorescence fluctuations was performed as described before in Blanchard et al., 2010 and Booth et al., 2014. Briefly, the apical-medial fluorescence values in the time series were detrended by subtracting a boxcar average with a window of 6 min (larger than the maximum expected fluctuation cycle length). All peaks and troughs were then identified in the fluorescence signal. Peaks or troughs that were associated with very small amplitude cycles were skipped. The remaining peaks and troughs were used to calculate the amplitude and the cycle length (peak to peak) and the amplitude (average peak to trough) of fluctuations.

The analysis of junctional myosin polarity was performed as described before in Sanchez-Corrales et al., 2018; Tetley et al., 2016. Briefly, the average junctional fluorescence intensity is considered a periodic signal around the cell and decomposed into individual components (modes) using Fourier analysis. The strength of myosin bipolarity corresponds to the amplitude of the second mode while the strength of unipolarity corresponds to the amplitude of the first mode. Similarly, the phase of modes 1 and 2 (uni-/bipolarity) represents the orientation of cell polarity. The uni- and bipolarity amplitudes are expressed as a proportion of the mean cell perimeter fluorescence. Cells at the border of the placode that are part of the supra-cellular actomyosin cable at the boundary were excluded from the analysis.

Average fluorescence was calculated after background subtraction as the mean fluorescence inside a segmented cell excluding the cell edge pixels and two pixels in a perpendicular direction on either side. Fixed samples of Venus-Hkb and FkhGFP embryos of mid to late stage 10 (t<<0 min), early stage 11 (t+0 min) and late stage 11 (t>>0 min) were pooled together with relevant time points from live-imaging as shown in Table 1 and figure legends. Fluorescence was normalized in each image by dividing the average fluorescence of each cell by the 98th percentile value. Calculations were performed in R.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

The authors would like to thank the following people; for help with an experiment: Ghislain Gillard; for reagents and fly stocks: Debbie Andrew, Thomas Lecuit, Stefan Luschnig, Herbert Jäckle, and Naoyuki Fuse. The authors thank members of the lab for input on the manuscript. KR and YSC were supported by the Medical Research Council (file reference number U105178780). GB was supported by a Wellcome Trust Investigator Award (207553/Z/17/Z) to Bénédicte Sanson, Department of PDN, Cambridge, UK.

© 2021, Sánchez-Corrales 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.