1. Cell Biology
Download icon

Prickle isoforms determine handedness of helical morphogenesis

  1. Bomsoo Cho
  2. Song Song
  3. Jeffrey D Axelrod  Is a corresponding author
  1. Stanford University School of Medicine, United States
Research Article
  • Cited 0
  • Views 712
  • Annotations
Cite this article as: eLife 2020;9:e51456 doi: 10.7554/eLife.51456

Abstract

Subcellular asymmetry directed by the planar cell polarity (PCP) signaling pathway orients numerous morphogenetic events in both invertebrates and vertebrates. Here, we describe a morphogenetic movement in which the intertwined socket and shaft cells of the Drosophila anterior wing margin mechanosensory bristles undergo PCP-directed apical rotation, inducing twisting that results in a helical structure of defined chirality. We show that the Frizzled/Vang PCP signaling module coordinates polarity among and between bristles and surrounding cells to direct this rotation. Furthermore, we show that dynamic interplay between two isoforms of the Prickle protein determines right- or left-handed bristle morphogenesis. We provide evidence that, Frizzled/Vang signaling couples to the Fat/Dachsous PCP directional signal in opposite directions depending on whether Pkpk or Pksple predominates. Dynamic interplay between Pk isoforms is likely to be an important determinant of PCP outcomes in diverse contexts. Similar mechanisms may orient other lateralizing morphogenetic processes.

eLife digest

Our right and left hands are mirror images of each other and cannot be precisely superimposed. This property, known as chirality, is vital for many tissues and organs to form correctly in humans and other animals. For example, fruit flies have hair-like sensory organs on the edges of their wings known as bristles. One of the cells in each bristle forms a shaft that generally tilts away from the main body of the fly and is anchored in place by another cell known as the socket.

A signaling pathway known as PCP signaling controls the directions in which many chiral tissues and organs in animals form. The pathway contains two signaling modules: the global module collects “directional” information about the orientation of the body and sends it to the core module, which interprets this information to control how the tissue or organ grows.

Fruit flies have two different versions of one of the core module components – known as Prickle and Spiny legs – that are thought to alter the direction the core module responds to the information it receives. Mutant flies known as pkpk mutants are unable to make Prickle and their wing bristles tilt in the opposite direction compared to those in normal flies, but it was not clear exactly why this happens.

To address this question, Cho et al. studied PCP signaling in the wings of normal and pkpk mutant flies. The experiments showed that Prickle directed the bristles on the right wing of a normal fly to grow in left-handed corkscrew-like patterns in which the emerging shaft and socket of each bristle twisted around each other. As a result, the bristles tilted away from the bodies of the flies. In the pkpk mutants, however, Spiny legs substituted for Prickle, causing the equivalent bristles to grow in a right-handed corkscrew pattern and tilt towards the body.

The findings of Cho et al. show that PCP signaling controls the direction fly bristles grow by selectively using Prickle and Spiny legs. In the future, this work may also aid efforts to develop effective screening and treatments for birth defects that result from the failure of chiral tissues and organs to form properly.

Introduction

PCP signaling controls the polarization of cells within the plane of an epithelium, orienting asymmetric cellular structures, cell divisions and cell migration. In flies, PCP signaling controls the orientation of trichomes (hairs) on the adult cuticle, orientation of ommatidia in the eye, and orientation of cell divisions, though the full range of phenotypic outputs has not been explored. While much focus has been placed on mechanistic studies in flies, medically important developmental defects and physiological processes in vertebrates are also under control of PCP signaling, motivating mechanistic studies in flies that might inform similar studies in vertebrates. Defects in the core PCP mechanism result in open neural tube defects, conotruncal heart defects, deafness, situs inversus and heterotaxy (reviewed in Butler and Wallingford, 2017; Henderson et al., 2018; Blum and Ott, 2018). PCP is also believed to participate in both early and late stages of cancer progression and in wound healing. PCP polarizes skin and hair, the ependyma and renal tubules. Paralogs of the PCP component Prickle are mutated in an epilepsy-ataxia syndrome (Tao et al., 2011; Mei et al., 2013; Bassuk et al., 2008; Ehaideb et al., 2014; Paemka et al., 2015). Mutations in ‘global’ PCP components have been associated with a human disorder of neuronal migration and proliferation (Zakaria et al., 2014) and in developmental renal disorders (Zhang et al., 2019).

Work in Drosophila indicates that at least two molecular modules contribute to PCP signaling. The core module acts both to amplify molecular asymmetry, and to coordinate polarization between neighboring cells, producing a local alignment of polarity. Proteins in the core module, including the serpentine protein Frizzled (Fz), the seven-pass atypical cadherin Flamingo (Fmi; a.k.a. Starry night), the 4-pass protein Van Gogh (Vang; a.k.a. Strabismus), and the cytosolic/peripheral membrane proteins Dishevelled (Dsh), Diego (Dgo), and the PET/Lim domain protein Prickle (Pk) adopt asymmetric subcellular localizations that predict the morphological polarity pattern such as hairs in the fly wing (reviewed in Zallen, 2007; Butler and Wallingford, 2017). These proteins communicate at cell boundaries, recruiting one group to the distal side of cells, and the other to the proximal side, through the function of an incompletely understood feedback mechanism, thereby aligning the polarity of adjacent cells. A global module serves to provide directional information to the core module by converting tissue level expression gradients to asymmetric subcellular Fat (Ft) - Dachsous (Ds) heterodimer localization (reviewed in Matis and Axelrod, 2013; Butler and Wallingford, 2017; Zallen, 2007). The atypical cadherins Ft and Ds form heterodimers which may orient in either of two directions at cell-cell junctions. The Golgi resident protein Four-jointed (Fj) acts on both Ft and Ds as an ectokinase to make Ft a stronger ligand, and Ds a weaker ligand, for the other. Graded Fj and Ds expression therefore result in the conversion of transcriptional gradients to subcellular gradients, producing a larger fraction of Ft-Ds heterodimers in one orientation relative to the other. Other less well defined sources of global directional information appear to act in partially overlapping, tissue dependent ways (Wu et al., 2013; Sharp and Axelrod, 2016).

Various Drosophila tissues depend primarily on either of two isoforms of Pk, Prickleprickle (Pkpk) and Pricklespiny-legs (Pksple) (Gubb et al., 1999). These isoforms have been proposed to determine the direction in which core PCP signaling responds to directional information provided by the Ft/Ds/Fj system (Olofsson and Axelrod, 2014; Hogan et al., 2011; Ambegaonkar and Irvine, 2015; Ayukawa et al., 2014). Pksple binds directly to Ds, orienting Pksple-dependent core signaling with respect to the Ds and Fj gradients (Ayukawa et al., 2014; Ambegaonkar and Irvine, 2015), while Pkpk-dependent core signaling has been proposed to couple less directly through a mechanism in which the Ft/Ds/Fj module directs the polarity of an apical microtubule cytoskeleton on which vesicles containing core proteins Fz, Dsh and Fmi undergo directionally biased trafficking (Matis et al., 2014; Olofsson and Axelrod, 2014; Shimada et al., 2006; Harumoto et al., 2010). Others, however, have argued that Pkpk-dependent core signaling is instead uncoupled from the Ft/Ds/Fj signal (Merkel et al., 2014; Ambegaonkar and Irvine, 2015; Casal et al., 2006; Lawrence et al., 2007; Brittle et al., 2012).

The most intensively studied morphogenetic responses to PCP signaling in Drosophila occur in epithelia such as wing and abdomen, in which cellular projections called trichomes (hairs) grow in a polarized fashion from the apical surface, and in ommatidia of the eye, in which photoreceptor clusters achieve chirality via directional cell fate signaling. Chaete (bristles), which serve as sensory organs, are also polarized by PCP signaling (Schweisguth, 2015; Gubb and García-Bellido, 1982). Bristles comprise the 4–5 progeny of a sensory mother cell (SMC), one of which, the shaft, extends a process above the epithelium such that it tilts in a defined direction with respect to the tissue. In mechanosensory microchaete on the notum, one daughter of the SMC divides to produce the shaft and a socket cell that surrounds the shaft where it emerges from the epithelial surface; the other SOP daughter divides to produce a glial cell, a sheath cell and a neuron. Studies of microchaete polarity have shown that the initial division of the SMC is polarized by PCP in the epithelium from which it derives, such that the two daughters are born in defined positions with respect to each other (Gho and Schweisguth, 1998; Bellaïche et al., 2001). However, subsequent events that ultimately determine the direction of shaft polarity have not been described.

We chose to study bristle polarity on the anterior margin of the wing (AWM). A row of stout mechanosensory bristles and a row of curved chemosensory bristles are on the dorsal surface, and a mixed row of mechano and chemosensory bristles is on the ventral side (Figure 1A, Figure 1—figure supplement 1A). All of these bristles tilt toward the distal end of the wing in wildtype. In pkpk mutant wings, and in wings overexpressing Pksple, a large fraction of the AWM bristles point proximally rather than distally; the pkpk phenotype is suppressed by mutation of dsh, implicating the core PCP signaling mechanism in this process (Gubb et al., 1999). However, the morphogenetic process resulting in polarity and the genetics of its apparent control by PCP signaling have not been explored. Here, focusing on the dorsal mechanosensory bristles, we report our analysis of the underlying morphogenesis leading to AWM bristle polarization, and show that polarization results from a corkscrew-like helical morphogenetic process involving the shaft and socket cells. Furthermore, our results reveal how interplay between Pkpk and Pksple control the handedness of the helical growth, and how the Ft/Ds/Fj system directs it in opposite orientations depending on whether the core PCP mechanism operates in a Pkpk- or Pksple-dependent mode.

Figure 1 with 2 supplements see all
Morphology of wildtype dorsal mechanosensory bristles.

(A) Dorsal and ventral views showing adult dorsal mechanosensory bristles (red arrowheads), ventral chemo- and mechanosensory bristles (blue arrowhead) and dorsal chemosensory bristles (black arrowhead). Bristles are separated by the exoskeleton secreted by cells displaying trichomes (hairs), similar to those in the majority of the wing blade. See also Figure 1—figure supplement 1A for a schematic view. (B) 3D reconstruction of a section of a 36 hr control (w1118) AWM containing two clones expressing cytoplasmic RFP, each labeling two adjacent dorsal mechanosensory bristles. Several clones labeling hair cells are also present in this sample. Costaining with Su(H) marks all socket cells, and Vang::EYFP is present at apical cell junctions. The boxed region is displayed from several angles in panel D). (C–C’) Wildtype adult wing and equivalent region of a 36 hr pupal wing stained for Su(H) and actin. (D–D’) 3D views from different angles of the RFP clone(s) shown in panel B). (D’’) Cartoon interpretation of the sibling shaft-socket pairs from D’). All images throughout are of right wings and are displayed proximal to the left and distal to the right. Scale bars: 20 μm.

Results

Proximo-distal relationship of wing mechanosensory bristle shaft-socket cell pairs is reversed in pkpk mutants

To begin to characterize the determinants of AWM mechanosensory bristle polarity, we labeled the externally exposed socket and shaft cells in wildtype (wt) and pkpk mutants with anti-Su(H) (Suppressor of hairless) antibody and phalloidin, respectively. In wildtype control wings at 36 hr apf, the apical ends of socket cells are tilted toward the distal end of the wing and are interspersed with the actin bundles of the shafts (Figure 1C,C’, Figure 1—figure supplement 1B–B’’). The shafts are interspersed between socket cells, and appear to ascend along the proximal side of the adjacent socket cell and pass through an opening at its apical surface. Consistent with polarity patterns of adult bristles, actin bundles of pkpk mutant shafts near the distal end of the wing show a reversed, proximal, tilt, whereas shafts in the proximal and the very most distal regions show the normal distal tilt (Figure 2A,A’, Figure 1—figure supplement 1C–C’’). Between these regions, shafts show a smooth transition between proximal and distal tilt, with some shafts pointing straight up (neutral tilt). The socket cells tilt at angles that correlate with the polarity of neighboring actin bundles in pkpk mutant wings: P-D tilt at the proximal region and D-P tilt at the distal region, with smooth transition between those regions (Figure 1—figure supplement 1C–C’’). Furthermore, shaft actin bundles with reversed polarity appear to be positioned on the distal side of the socket cell through which they pass, opposite to their relationship in wildtype and to their relationship in the proximal region in pkpk mutant wings where their tilt shows the normal, distal, direction (Figure 1—figure supplement 1B–B’’, compare with Figure 1—figure supplement 1C–C’’).

Figure 2 with 1 supplement see all
Morphology of pkpk mutant dorsal mechanosensory bristles.

(A–A’) pkpk30 adult wing and equivalent region of a pkpk30 36 hr pupal wing stained for Su(H) and actin. (B–B’) 3D views of RFP clones in a pkpk30 36 hr pupal wing revealing reversed orientation of sibling shaft-socket pairs. Su(H) stains socket cells. (B’’) Cartoon interpretation of shaft-socket pair from B’). Scale bars: 20 μm.

Previous studies of AWM bristle ultrastructure have been insufficiently detailed to appreciate the determinants of polarity (Hartenstein and Posakony, 1989; Palka et al., 1979). To better understand the structures of shaft-socket pairs, and to unambiguously determine the relationship between sibling shaft-socket cell pairs, random individual bristle lineages were labeled by clonal expression of RFP, filling the cell bodies of labeled shaft and socket cells. Simultaneous staining of the socket cells (identified by Su(H) expression) allows one to identify the sibling shaft-socket cell pairs. 3D confocal reconstructions facilitate examination from a variety of viewpoints and enable the positions and shapes of cell bodies, nuclei, shaft, and apical opening of the socket cells to be visualized (Figure 1B–B’’’, Figure 1—video 1). These views allow us to see that the wildtype nucleus and main portion of the shaft cell is dorsal and extends slightly posterior to that of the sibling socket cell. The base of the shaft rises from the cell body, wraps clockwise along its socket sibling (as viewed from the apical side in a right wing), and then rises through a groove in the socket that extends apically along the proximal side of the socket, finally emerging through the donut shaped apical surface of the socket cell (Figure 1D–D’’, Figure 1—video 1). In contrast, in the distal bristles of pkpk mutant wings where bristles are reversed, the shaft is positioned within an oppositely oriented groove on the distal side of the sibling socket cell. The shaft nucleus is dorsal and posterior to the socket nucleus, similar to their arrangement in wildtype, though their alignment is not as regular. Therefore, the structure of pkpk mutant bristles is roughly a mirror image of that in wildtype, with nuclei in similar positions, but the shaft bending in a counterclockwise direction and rising through a groove on the opposite side of the socket cell compared to wildtype (Figure 2B–B’’). In the proximal region of the pkpk mutant wing, the relationship of shaft-socket siblings resembles that in wildtype, reflecting the normal bristle polarity in that region (Figure 1—figure supplement 1C–C’’ and Figure 3—figure supplement 1B). Therefore, the P or D position of shaft relative to the sibling socket cell correlates to the normal or reversed bristle polarity in proximal and distal regions of pkpk mutants, suggesting that the shaft position relative to the socket determines bristle polarity (tilt).

Helical growth of the shaft-socket pair positions the shaft relative to the socket

Since the relative position of the shaft to the socket appears to be important for bristle polarity, we wished to identify the developmental process by which this relationship is achieved. In microchaete of the notum, the orientation of the initial division of the sensory organ precursor cell is specified by PCP signaling (Schweisguth, 2015). Assuming the orientations of subsequent daughter cell divisions are similarly regulated, the shaft-socket relationship may be determined by their relative positions at their birth. A similar process might occur in AWM mechanosensory bristles. Alternatively, a post-division morphogenetic process may determine their final configurations.

To examine this process, we analyzed 3D structures of shaft-socket sibling pairs at earlier developmental stages. In wildtype bristles at 24 hr apf, the shaft cell nucleus is posterior and just slightly dorsal to the socket cell nucleus, similar to their positions at 36 hr apf (Figure 3A,B). The extending shaft is just reaching the apical surface, and is positioned in a groove on the dorsal side of the socket cell. At the apical surface, the shaft, sits in a shallow indentation in the apical surface of the socket cell, which adopts a crescent shape with the opening of the crescent pointing in the dorsal direction. Over time, from 28 to 32 to 36 hr apf, as the shaft continues to extend above the apical surface, the socket cell crescent rotates clockwise, and gradually closes to form a ring around the shaft (Figure 3A,B, Table 2, Table 1, Figure 3—videos 1, 2, 3). Because the apical surface rotates while the nuclei remain relatively stationary, the shaft and socket twist to form a left-handed helical shape. As the shaft grows out above the socket cell, it points distally. The cell bodies, are initially relatively flat in the dorsal-ventral direction, but as they grow, extend dorsally, becoming flatter in the anterior-posterior direction. Quantification of rotation was performed by measuring rotation angles, as diagrammed in Figure 3B, from 3D images of socket cells captured at different time points (c.f. Figure 3C), and data are displayed in rose plots (Figure 3F, Figure 2—figure supplement 1A, Table 2). Rotation at 32 hr averages 55° in the clockwise direction. Note that throughout, we describe analyses of right wings. In all cases, left wings develop as mirror images of right wings.

Figure 3 with 4 supplements see all
3D images of shaft-socket pairs at different times reveal a clockwise helical growth in control and counterclockwise growth in pkpk30 mutant bristles.

(A) Reconstructed 3D images from varying angles of 24 hr, 28 hr, 32 hr and 36 hr shaft-socket clones marked with RFP and stained with Su(H). (B) Cartoon interpretation of images from panel A), showing the clockwise rotation of the apical aspects of the shaft and socket cells. Views from dorsal (red dots) and proximal (blue dots). (C) Diagrams illustrating scoring of rotation angles. (D–E) Control (w1118) and pkpk30 adult AWMs and corresponding regions from 32 hr pupal wings showing 3D reconstructed images of socket cells stained with Su(H). Orientation angles for these socket cells are indicated by yellow arrows. (F) Quantification of rotation angles for 24 hr and 32 hr control and pkpk30 socket cells. Each rose plot represents an individual wing, with scoring limited to the distal AWM anterior to vein L2 unless otherwise indicated (for complete set of rose plots and description of sample sizes, see the legend for Figure 2—figure supplement 1A). Most variation between individual wings is likely attributable to variation in developmental timing. For detailed description of quantification, see Materials and methods. Statistical analyses for all genotypes are in Table 2.

Table 1
p values for comparison of rotation angles for control W1118 socket cells at different times apf.
W111824 hr
5.0°
28 hr
28.4°
24 hr
5.0°
28 hr
28.4°
<0.0001
32 hr
54.588°
<0.0001<0.0001
Table 2
Summary statistics for rotational angles.

CSD = Circular Standard Deviation.

GenotypeTimeGrand Mean Vector (GM)Length of Grand Mean Vector (r)Number of
means (wings)
Mean CSD
W111824 hr5.04°0.99664.14
2828.368°0.99574.075
3254.588°0.98162.940333
pkpk3024 hr0.63°0.98837.781333
32 hr348.482°0.946715.09
pkpk-sple1324 hr1.061°0.99733.687
32 hr47.59°0.99336.582
pksple124 hr17.8°0.97937.244667
32 hr59.288°0.9944.24425
MS1096 >> sple24 hr358.057°0.99435.796
32 hr333.296°0.964313.31267
fzR5224 hr3.351°0.99833.923333
32 hr20.471°0.905623.5668
dsh124 hr5.684°0.99625.023
32 hr20.144°0.944614.316
fmi RNAi24 hr1.15°0.99634.78
32 hr19.291°0.978411.741
vangstbm624 hr1.51°0.967311.52167
32 hr25.447°0.94614.21117
w1118 MS1096 > dsRNAi32 hr16.593°0.79433.10867
pkpk MS1096 > dsRNAi32 hr29.323°0.961314.37067
pksple MS1096 > dsRNAi32 hr6.596°0.97312.63867
pkpk-sple MS1096 > dsRNAi32 hr40.297°0.9936.219667
  1. Table 2—source datas 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, containing measurement of rotation angles of socket cells in control and various PCP mutant wings, with associated statistics, are provided.

Based on stereotypical patterns from images of timed, fixed samples of the apical surface, we infer that a margin cell at the dorsal side of the shaft-socket pair rotates together with the pair toward the proximal side, generating new junctions with the neighboring socket and margin cells and widening the gap between shaft-socket pairs (Figure 2—figure supplement 1B). During these events, the number of margin cells surrounding the shaft-socket pair is maintained and cell-cell junctions are remodeled.

Counterclockwise helical growth of shaft-socket pairs causes bristle reversal in pkpk mutants

To characterize the events leading to reversed bristle polarity in pkpk mutants, rotation angles of shaft-socket pairs during development were analyzed as above. At 24 hr apf, shaft cells were positioned dorsal to their socket sibling cells, as in wildtype. At later times, apical rotation proceeded counterclockwise, opposite to the wildtype direction, in the distal bristles that adopt a reversed polarity, giving rise to a right-handed helical shape in contrast to the left-handed helical shape of wildtype pairs (Figure 3E, Figure 2—figure supplement 1A, Table 3). Shaft-socket pairs in the proximal region rotated clockwise, corresponding to their normal polarity, and bristles in the region between normal and reversed bristles rotated very little, corresponding to their neutral, upright, polarity (Figure 3—figure supplement 1B). Therefore, bristle polarity does not depend on the birth positions of shaft and socket cells, which are born by 14 hr apf (Hartenstein and Posakony, 1989), but rather on the direction of apical rotation of the developing shaft-socket pair, beginning shortly after 24 hr apf.

Table 3
p values for comparison of rotation angles for control and pk-related genotypes at 32 hr apf. ns = not significant.
32 hr anglesW1118
54.6°
pksple1
59.3°
pkpk-sple13
47.6°
pkpk30
348.5°
W1118
54.6°
pksple1
59.3°
(ns) 0.0702
pkpk-sple13
47.6°
0.01430.0343
pkpk30
348.5°
<0.0001<0.00010.0004
>>Pksple
333.2°
0.0012<0.00010.001(ns) 0.1826

Pkpk versus Pksple isoform expression determines the direction of rotation

Since pkpk mutants show counterclockwise rotation leading to reversed shaft-socket positioning (D-P) and bristle reversal, and Pksple overexpression similarly reverses bristle polarity as previously reported (Figure 3—figure supplement 1E,E’ and Gubb et al., 1999), we surmised that Pksple induces the counterclockwise helical growth that leads to D-P orientation of shaft-socket pairs and reversed bristle polarity. Consistent with this, counterclockwise rotation of shaft-socket pairs and proximal polarity in pkpk mutants is suppressed by removing pksple (in pkpk-sple13/pk-sple13; Figure 3—figure supplement 1A–D’, Table 3, Gubb et al., 1999), and pksple overexpression induces counterclockwise rotation similar to that in pkpk mutants (Figure 3—figure supplement 1E,E’ Table 2Table 3). Pksple is therefore needed to reverse shaft-socket rotation in pkpk mutants. Overexpression of Pksple induced counterclockwise rotation and reversed polarity in a wider region than in pkpk mutants (compare Figure 3—figure supplement 1B to E). Thus, endogenous Pksple is only poised to act at the distal margin, but exogenous Pksple can reverse most, if not all, bristles. Though the potential for Pksple to reverse bristle polarity is unmasked in the absence of Pkpk, it plays no essential role in wildtype polarization, as pksple bristles fully rotate, or perhaps marginally over-rotate (59.3°±4.2° vs 54.6°±2.9°, p=0.0702; Table 3, Figure 3—figure supplement 1C,C’).

Surprisingly, pkpk-sple13 mutant bristles rotate only moderately less than wildtype bristles (47.6°±6.6° vs. 54.6°±2.9°, p=0.0143; Table 3), suggesting that Pkpk might play only a modest role in controlling the magnitude of rotation in wildtype. We propose that this is due to residual core PCP signaling activity observed in the absence of Pk (Strutt and Strutt, 2007; Lawrence et al., 2004; Adler et al., 2000), and the implications of this result are considered more fully in the Discussion.

Localization of Pkpk and/or Pksple correlates with handedness of helical rotation

We have shown that in wildtype bristles, Pkpk antagonizes Pksple to direct clockwise rotation of shaft-socket pairs, and that Pksple, when overexpressed, outcompetes Pkpk to direct counterclockwise rotation. Although the idea that Pkpk and Pksple antagonize each other has been previously proposed (Gubb et al., 1999), how this occurs has been obscured in part by the inability to specifically visualize the endogenous expression of each isoform. We therefore modified the endogenous genomic sequence encoding Pkpk or Pksple by appending a V5 tag to the N-terminus, facilitating the tissue and cellular level evaluation of their native expression patterns at various developmental stages. Both tagged isoforms support wildtype polarity development in all tissues and various controls suggest that expression of these genomically tagged versions reflect that of the native loci (Figure 4, Figure 4—figure supplements 1 and 2). Here, we describe their expression in developing wings.

Figure 4 with 2 supplements see all
Expression of Pkpk and Pksple isoforms in pupal wings.

(A–B’’) V5::Pk (A–A’’) and V5::Sple (B–B’’) in pupal wings of ages indicated. Red arrowheads mark AWM and yellow arrowheads mark veins L3 and L4. (C–D’) Surface views of V5::Pk at 24 hr (C–C’) or V5::Sple at 32 hr (D–D’) counterstained for Su(H) to locate socket cells. Some socket cell locations are indicated by yellow dots. (E–G’’) Planar sections of socket cells (Su(H)) of 28 hr pupal wings with apical at the top. V5::Pk is at all junctions between socket and margin cells (E–E’’). Pksple (detected with V5::Sple) appears to be localized to the proximal side of control (w1118) socket cells (F–F’’) but to the distal side of pkpk mutant socket cells (detected with anti-Pk[C]; G–G’’). (H–I’) Mosaic expression of V5::Pk (H–H’’) or V5::Sple (I–I’’) in otherwise wildtype wings (28 hr). Cells lacking expression are marked with RFP (blue). Both V5::Pk and V5::Sple localize to the proximal side of expressing (orange arrowheads) but not non-expressing (yellow arrowheads) socket cells, demonstrating their proximal localization. V5::Pk localizes to the proximal side of expressing (orange arrow) but not non-expressing (yellow arrow) margin wing cells. V5::Sple is also proximal in margin wing cells, though no informative clones were captured in this image. Several relevant cells are outlined in (H and I) for clarity. Scale bars: 20 μm (A,B) and 5 μm (C–I).

Notably, V5::Pk and V5::Sple reveal that throughout wing development, expression is spatiotemporally dynamic. Early in wing development, Pkpk protein is strongly expressed and is present in most or all cells. In discs, Pkpk is relatively elevated in AWM proneural cells. Following a slight dip in levels around 8 hr apf, expression levels climb, peaking around 32 hr apf, when wing hairs emerge, and then decline with little detectable expression remaining by 40 hr apf (Figure 4A–A’’, Figure 4—figure supplement 1B–H,K–K’’). Expression of Pksple is below detection in discs (Figure 4—figure supplement 1B,I,J), makes a small peak at around 8 hr apf, and is then not detectable between 16 and 28 hr apf. Beginning around 28 hr apf, Pksple expression is specifically detected in dorsal AWM cells and in vein L3. This expression persists through 32 hr apf, when wing hairs emerge, and weaker expression in other veins becomes apparent (Figure 4B–B”, Figure 4—figure supplement 1B,L,L’). At these times, Pksple is below detection levels on Western blots. Beginning sometime after 32 hr apf, Pksple expression increases in most cells, with its level equaling that of the declining Pkpk by 36 hr apf and reaching its highest level at around 40 hr apf (the latest time we examined), when Pkpk is no longer detected. At 40 hr apf, when Pksple is at peak expression in most of the wing blade, expression has disappeared from vein cells (Figure 4—figure supplement 1B,L–L”).

Most pertinent to bristle development, at the 24 hr apf AWM, Pkpk is expressed at apical junctions at similar levels in anterior margin and socket cells (Figure 4C,C’, Figure 4—figure supplement 2A). By 28 hr apf, Pkpk expression begins to decrease during bristle-socket rotation, first in socket cells, and later in all cells at and near the margin (Figure 4—figure supplement 2A). At the same time, Pksple expression becomes evident and increases over time, first uniformly in cells near the margin, and gradually becoming strongest in the socket and shaft cells (Figure 4D,D’,F–F’’, Figure 4—figure supplement 2B, compare with Pkpk (V5::Pk) in Figure 4E). The timing of the shift from Pkpk to Pksple expression is accelerated at the margin relative to the interior of the wing. Mosaic experiments demonstrate that Pkpk localizes proximally in socket and other margin cells, and that, importantly, Pksple colocalizes with Pkpk to the proximal side of socket cells. Proximal Pksple localization is an unexpected observation based on previous studies in which overexpressed Pksple localized at the distal junctions of hair cells (Ayukawa et al., 2014; Ambegaonkar and Irvine, 2015). Because Pksple expression is stronger in socket cells compared to margin cells by 32 hr apf, it has the useful property of effectively being expressed as a mosaic, allowing its localization to be scored without inducing clones (Figure 4F–F’’,H–H’’,I–I’’).

As Pksple activity reverses rotation direction of shaft-socket pairs in pkpk mutant wings, Pksple protein localization was analyzed in pkpk mutant wings. Because anti-Pk[C] antibody recognizes the common region of Pkpk and Pksple, the antibody reveals Pksple isoform localization in pkpk mutants. In the region of pkpk mutant wings where P-D reversal occurs, Pksple protein localized at the distal side of socket cells, whereas in the proximal region where polarity is not reversed, Pksple protein shows minimal asymmetry (Figure 4G–G’’, Figure 4—figure supplement 2C–C’’’). Distal localization of Pksple in the region of polarity reversal was verified by clonal knockdown of pksple in pkpk mutant wings. Notably, the socket-shaft pairs that lacked both Pksple and Pkpk failed to rotate (Figure 4—figure supplement 2D–E’’’).

These results suggest that Pkpk normally inhibits Pksple from localizing distally by recruiting it to the proximal junction of socket cells (and likely also in nearby margin cells, although this is hard to visualize due to lower expression in those cells). Furthermore, the distal localization of Pksple in pkpk mutants correlates with its ability to determine counter-clockwise rotation of shaft-socket pairs on a cell-by-cell basis, suggesting that this localization is likely the determinant of counter-clockwise rotation.

Core PCP components control rotation of shaft-socket pairs

Our results thus far show that the direction of Pk polarization, whether Pkpk, Pksple or both, corresponds to the direction of bristle polarization. Suppression of polarity reversal in pkpk mutants by dsh implicates the core PCP signaling mechanism in this process Gubb et al. (1999). We therefore asked whether the remaining components of the core PCP signaling mechanism contribute to AWM bristle polarization. As with dsh mutation, knock down of fz or vang in pkpk mutants suppressed reversal of bristle tilt, shaft-socket orientation, and rotation direction (Figure 5A–D). Core PCP signaling is therefore required for reversed, Pksple-dependent polarity.

Figure 5 with 1 supplement see all
Core PCP components control shaft-socket rotation.

(A–D) Reversed bristle polarity in pkpk mutant compared to control is abrogated in pk fz and pk vang double mutants, demonstrating requirement for core PCP activity for polarity reversal in pkpk mutants (reversed bristle polarity is suppressed in all double mutant wings analyzed (n = 20 of each genotype)).(E–H) While bristle tilt is only mildly disturbed, socket cell rotation is strongly impaired in fmi knockdown, fz, vang and dsh mutant wings. Adult wings, representative socked cell images (32 hr) and quantification of three individual wings (32 hr) for each genotype are shown (fmi knockdown - 97 sockets from four wings; fz mutant - 120 sockets from six wings; vang muant - 114 sockets from six wings; and dsh mutant - 110 sockets from six wings). Statistical analyses for all genotypes are in Table 2. (I–J’) Expression of Fz::EGFP (I–I’) and Vang::EYFP (J–J’) in margin cells. (K–L) Mosaic expression demonstrates distal localization of Fz::EGFP (K) and proximal localization of Vang::EYFP (L) in both socket (yellow) and margin (white arrowheads) cells. Some informative expressing cells (yellow dots) next to non-expressing cells (red dots) are marked. Scale bars: 10 μm. Statistical analyses for all genotypes are in Table 2.

To assess a potential contribution of core PCP signaling to Pkpk-dependent bristle polarization, the anterior region of adult wings from fmi RNAi, fz, vang, and dsh mutants were analyzed, and the rotation of shaft-socket pairs was evaluated for each genotype (Figure 5E–H). Adult mechanosensory bristles of core PCP mutants are less tilted toward the distal direction than those of wildtype and the tilting angles are somewhat irregular, with some bristles tilting out of the plane of the wing. Consistent with the adult wing defects, in pupal wings the shaft position relative to the socket varies, sometimes abruptly, in the same mutant wing, showing less local correlation than in wildtype. Quantification reveals substantial under-rotation of shaft-socket pairs, and a broader distribution of rotation angles than in wild type (Figure 5E–H, compare with Figure 3D,F, Table 2). Thus, careful morphological analysis reveals that core PCP signaling is required for normal, Pkpk-dependent polarity as well as reversed, Pksple-dependent polarity.

Consistent with a role for core PCP components in mediating rotation of shaft-socket pairs, junctional asymmetry of Fz::EGFP and Vang::EYFP is well preserved between socket and margin cells and between adjacent margin cells (Figure 5I–J’). Little accumulation was observed at junctions between shaft and socket cells. Mosaic analyses demonstrated the expected distal localization of Fz and proximal localization of Vang at both margin cell-margin cell and margin cell-socket cell junctions, suggesting that PCP signaling likely occurs between margin cells and between margin and socket cells (Figure 5K,L). Similarly, the reversed bristle polarity observed in pkpk mutants was accompanied by reversed Vang localization in pkpk mutant socket and margin cells, consistent with the idea that the direction of core PCP polarization is reversed in pkpk mutants (Figure 5—figure supplement 1). To functionally test polarity propagation between margin and socket cells by core PCP components, fz or vang knock-down clones were generated, and clones at the AWM were analyzed (Figure 6, Figure 6—figure supplement 1). fz or vang knock-down clones, whether in just margin cells, just shaft-socket pairs, or both, showed non-autonomy as assessed by sequestration of Fz (fz clones) or Vang (vang clones) at the clone borders. Near fz RNAi clones, distal cells, including both bristle and hair cells, were re-oriented: sockets on the distal side of the clones showed counter-clockwise rotation, and hairs on the distal side grew toward the clones. Similarly, sockets on the proximal side of vang RNAi clones rotated counter-clockwise (Figure 6—figure supplement 1). These results indicate that core PCP signaling propagates between bristle and margin cells to control the rotation direction of shaft-socket pairs.

Figure 6 with 1 supplement see all
Polarity propagates between bristle and margin cells.

3D reconstructed views of a Fz::EGFP wing (32 hr) with fz knockdown clones, stained for actin to mark wing hairs and bristle shafts, Su(H) to mark socket cells and RFP to indicate knockdown clones. A clone involving three bristles (red arrowheads, false-colored in top image) shows domineering non-autonomy, reversing the polarity of nearby bristles and hairs on the distal side of the clone (bristles marked a-f). The effect on both hair cells and bristles diminishes with distance. A small clone affecting margin cells (red open arrowheads) disrupts the polarity of bristle cells on the distal side of the mutant cells. 20 fz RNAi clones from eight wings were analyzed for cell autonomous and non-autonomous effects. All clones showed cell autonomous polarity disruption and non-autonomous reversal of varying extent depending on the clone size.

The core PCP module differentially interprets directional signals from ds when operating in Pkpk- or Pksple-dependent modes

We have previously proposed that a signal from the Ft/Ds/Fj system provides a directional cue to orient core PCP signaling in some tissues (Ma et al., 2003; Yang et al., 2002; Matis et al., 2014; Olofsson et al., 2014), although others have argued that this system operates in parallel with core PCP signaling (Casal et al., 2006; Lawrence et al., 2007; Brittle et al., 2012). An asymmetry of Ft-Ds heterodimers, with a small excess of Ds displayed on the distal side of the cell, and Ft on the proximal side, has been observed, and is proposed to provide this signal (Ambegaonkar et al., 2012; Bosveld et al., 2012; Brittle et al., 2012). We have also proposed that the core PCP module differentially interprets directional signals from Ds when operating in Pkpk- or Pksple-dependent modes, with Pksple directing localization of the Fmi-Vang complex to the side where Ds is in excess, while the Fmi-Vang complex localizes to the opposite side when functioning in a Pkpk-dependent manner [(Olofsson and Axelrod, 2014); see also Lawrence et al. (2004). Pksple has been shown to bind to Ds and the associated Dachs protein, providing a mechanism for orienting Pksple-dependent core function to the Ft/Ds/Fj signal (Ayukawa et al., 2014; Ambegaonkar and Irvine, 2015), whereas a less direct, microtubule-dependent mechanism was proposed to mediate this response when Pkpk is predominant (Shimada et al., 2006; Harumoto et al., 2010; Olofsson and Axelrod, 2014; Matis et al., 2014). In contrast, some have proposed that Pkpk-dependent core signaling is instead uncoupled from the Ft/Ds/Fj signal (Merkel et al., 2014; Ambegaonkar and Irvine, 2015).

We reasoned that our reagents would allow us to analyze effects of the Ft/Ds/Fj signal on each Pk isoform. We first examined the phenotype resulting from knockdown of Ds (Figure 7C). As previously observed (Adler et al., 1998), normal bristle tilt was substantially disturbed. The pattern of disturbance consistently showed regions of coordinated polarity that smoothly transition through neutral polarity to adjacent regions of opposite polarity, although the number and position of those domains varied. The effectively mosaic expression of V5::Sple localization allows one to observe precisely correlated regions of distal and proximal Sple localization corresponding to the regions of reversed and normal polarity, respectively, with relatively symmetric localization in the intervening transitions (Figure 7C). These results suggest that local core PCP signaling maintains polarity correlation among immediate neighbors but that alignment to the tissue axis is eliminated in the absence of Ds.

Figure 7 with 1 supplement see all
Pkpk and Pksple activity responds to Ds.

(A) A control (w1118) wing showing normally oriented bristles and rotated socket cells, and proximal V5::Sple. (B) A pkpk mutant wing with reversal of distal bristles, counter-rotated socket cells and distal Pksple. (C) Wings knocked down for ds (MS1096-GAL4, UAS-dsRNAi) show variable regions of locally correlated but either reversed or normal polarity. V5::Sple localization varies, corresponding to the local reversed (blue) or normal (yellow) polarity. (D) In pkpk mutant wings in which ds is knocked down, socket cells are minimally rotated, and Pksple, probed with anti-Pk[C] antibody, shows minimal apical localization. (E) pksple wing with ds knocked down. (F) pkpk-sple wing with ds knocked down. Quantification of socket cell rotation for the genotypes shown in C–F are given in Figure 7—figure supplement 1. (G–K) The normal proximal-high and distal-low Ds gradient was reversed in the distal wing by driving two copies of UAS-ds by dll-Gal4 in control (w1118), fzR52/R52, pkpk, pkpk-sple and pksple wings. The approximate gradient of ectopic ds expression is shown in green. Polarity reversal in the distal part of a control wing (G) was reversed (arrows, and images representing boxed regions) and depended on the presence of Fz (H) and Pkpk but not on Pksple (I–K). Wing hair images (G–K) are of the dorsal side. Subjectively similar hair polarity patterns were obtained from 13 of 15 wings for each genotype in G-K).

We then asked if the Pksple-dependent reversed polarity in pkpk mutants depends on Ds. When ds was knocked down in pkpk mutants, bristle reversal was blocked, producing a phenotype similar to that of core mutants, and the distal localization of Pksple was no longer observed (Figure 7A–D). Therefore, the reversed, Pksple-dependent polarity in pkpk mutants requires Ds, and we interpret this to indicate that the Ds global signal recruits Pksple to sites of enriched Ds (distal) in the absence of Pkpk (Figure 7B; compare with 7D), which drives reversal of shaft-socket orientation and therefore reversal of bristle polarity.

In ds knockdowns, we are unable to readily interpret the localization pattern of V5::Pk because levels are similar in socket and margin cells. Nonetheless, other results suggest that the local polarity correlation in the absence of Ds is mediated primarily by asymmetric localization of Pkpk rather than the asymmetric localization of Pksple that we can observe. First, recall that in wildtype, Pksple is recruited to colocalize with Pkpk at proximal sites, so Pkpk is likely to similarly recruit the colocalization of Pksple in the absence of Ds. Consistent with this idea, when ds was knocked down in pkpk mutants, neither proximal nor distal localization of Pksple was observed in socket cells, and local correlation of bristle polarity was weak (Figure 7D). Thus, the local domains of correlated asymmetric Pksple localization in ds knock-down socket cells depend on the presence of Pkpk; Pksple alone is insufficient to facilitate local signaling between neighbors. Finally, removing Pksple in ds knock-down wings failed to significantly modify ds knock-down effects on the polarity of bristles (and also hairs) while removing both Pkpk and Pksple does (Figure 7E,F, Figure 7—figure supplement 1A–D), confirming that ds knock-down affects the Pkpk-mediated, rather than the Pksple-mediated, PCP signal for wing bristle (and hair) polarity. Taken together, these observations suggest that the locally correlated domains of polarity observed in ds knock-down wings depend on Pkpk activity.

These and previous results indicate that Pkpk is the principal isoform functioning in core PCP signaling during bristle polarization. They are most consistent with, though do not definitively show, that in wildtype, the Ds global signal directs orientation of core signaling such that Vang and Pkpk localize to the proximal side (and incidentally colocalizing Pksple to the proximal side) to establish normal polarity. The alternative possibility is that Ds activity is permissive, and some other signal directs this orientation of Pkpk-dependent core polarization. The proposal that Ds is instructive for orienting Pkpk-dependent core signaling, while consistent with our prior interpretation of coupling between the Ft/Ds/Fj signal and Pkpk-dependent core PCP signaling in wing hair polarization (Ma et al., 2003; Olofsson et al., 2014; Sharp and Axelrod, 2016; Yang et al., 2002), is at odds with other reports asserting that while under Pkpk control, core PCP directionality is uncoupled from the Ft/Ds/Fj signal (Merkel et al., 2014; Ambegaonkar and Irvine, 2015). Rigorous testing of this hypothesis requires reorienting the Ft/Ds/Fj signal and assessing the isoform dependence of the response.

It was previously shown that reversing the gradient of Ds expression near the distal part of the wing under control of distal-less-GAL4 (dll >2 x-ds) reverses wing hair polarity (Harumoto et al., 2010). Assuming that hair polarity is determined by Pkpk, for which ample evidence exists, and that it depends on core signaling, this result would demonstrate coupling of Pkpk-dependent core PCP signaling to the Ds signal. We rigorously tested this assumption by testing the core signaling and Pk isoform dependence of this response (Figure 7G–K). dll >2 x-ds reverses polarity of a substantial swath of wing hairs, precisely in the region where the Dll expression gradient is expected to be steepest (Figure 7—figure supplement 1E). dll >2 x-ds, however does not reverse AWM bristle polarity, as it does not produce a proximal-to-distal expression gradient at the AWM. Because our results show that AWM bristle and wing hair polarization show indistinguishable responses to Ft/Ds and to core PCP manipulations, we propose that wing hairs are a suitable readout for this assay. We first asked whether dll >2 x-ds-driven hair polarity reversal depends on core module activity by removing Fz, and found that ectopic Ds-dependent reversal is blocked in a fz mutant background (Figure 7G,H). Furthermore, ectopic Ds re-orients the core PCP protein orientation (Figure 7—figure supplement 1F–I). These results rule out the possibility that ectopic Ds reverses polarity through a pathway that does not include the core PCP module. We then tested the Pk isoform dependence of reversal, and found that it is almost entirely abolished upon removal of Pkpk (pkpk or pkpk-sple), but is largely unchanged upon removal of Pksple (pksple) (Figure 7G,I–K). This result decisively demonstrates that Pkpk-dependent core PCP signaling in wing hair polarization is oriented by the Ft/Ds/Fj signal, and strongly suggests that the same coupling occurs during Pkpk-dependent AWM bristle polarization.

Discussion

Direction of helical morphogenesis determines bristle polarity

Producing structures of defined chirality requires directional information on three Cartesian axes. Our results indicate that in determining bristle chirality, PCP provides directional information along the proximal-distal axis. The apical-basal axis is defined by the epithelium, while the dorsal-ventral axis is likely defined by the dorsal-ventral compartment boundary.

We have shown that the polarity of wing margin bristles (proximal or distal tilt) is determined by controlling the handedness of helical growth. The entwined shaft and socket cells undergo an apical clockwise or counterclockwise rotation that results in a left-handed or right-handed helical structure, placing the shaft to the proximal (wildtype) or distal (pkpk mutant) side of the socket cell. The direction of rotation depends on PCP signaling among and between margin and socket cells. Helical cellular structures of defined handedness, such as the bristles resulting from properly directed rotation, have been noted in bacterial and plant species, but few examples have been described in animals.

The entwined twisting of the shaft and socket is a coordinated morphogenetic event, and the apparent stereotyped junctional rearrangement of additional margin cells suggests that at least some other cells are involved as well. We do not know in which cell or cells mechanics are regulated to drive this morphogenesis. One possibility is that an internal cytoskeletal mechanism induces the helical growth of the socket and/or shaft cells. Another possibility is that the side of the socket cell crescent marked by Pk at 24 hr is anchored, while the other side grows to wrap around the shaft, inducing junctional rearrangements and propelling the rotation of the apical portion of the shaft relative to the socket cell. Apical rotation could then cause twisting of the more basal portions of the socket cell, and could in turn direct the shaft to the corresponding side of the socket cell.

The precise location at which the PCP signal is required to determine rotation direction is unclear. Because we observe asymmetric core complexes at margin-socket cell junctions, but very little at shaft-margin or shaft-socket junctions, we hypothesize that interaction between the socket and surrounding margin cells is the essential determinant of rotation. PCP proteins at these junctions could control junctional dynamics, as is known to occur in other systems (Huebner and Wallingford, 2018). This will require further investigation.

Role of core PCP signaling in socket-shaft rotation

Core PCP signaling participates in regulating rotation, as the magnitude of rotation is substantially impaired in the core mutants fz, dsh, vang and fmi. Nonetheless, we note that a small amount of clockwise rotation still occurs in these mutants. We hypothesize that tissue scale mechanical forces may drive this rotation, though we do not rule out the possibility that some other signaling activity may also be involved. Compared to other core proteins, the impact of removing Pkpk on the magnitude of rotation is subtle. The clockwise rotation in pkpk-sple mutants is only slightly less than in wildtype. This result is reminiscent of Pkpk function in polarizing wing hairs: polarity in pkpk mutants is strongly perturbed due to the presence of Pksple, while hair polarity in pkpk-sple mutants is only weakly perturbed (Gubb et al., 1999). These findings are consistent with previous proposals that the core PCP mechanism retains a residual capacity to propagate some asymmetry in the absence of Pk (Strutt and Strutt, 2007; Lawrence et al., 2004; Adler et al., 2000).

The core PCP signal, in addition to executing directed rotation in response to Pkpk or Pksple, coordinates polarity between neighboring bristles. Local correlation between rotation angles is strong when core signaling is intact, even in the absence of the Ft/Ds signal, but is weak when core signaling is disrupted. This is analogous to the proposed mechanism for locally coordinating polarity between adjacent wing hairs. It is important to note that the local polarity signal must pass through intervening margin cells to signal from bristle to bristle.

Spatiotemporal dynamics and selection of Pkpk versus Pksple for polarity determination

Our data suggest that whether the Pkpk or Pksple isoform dominates to control the direction of PCP signaling depends not only on the relative amounts of each isoform, but also on the dynamics of expression and its effect on competition for participation in the core complex. During rotation of AWM bristle shaft-socket pairs, both Pkpk and Pksple isoforms are detected at the apical junction of the socket with an inverse temporal relationship; high expression of Pkpk decreases during rotation, while the initially undetectable level of Pksple protein increases. In these conditions, the system is controlled by Pkpk, and both Pkpk and Pksple localize proximally, thus orienting the core complex in its wildtype configuration. We hypothesize that Pksple is recruited by Pkpk through their known ability to interact heterotypically (Ayukawa et al., 2014; Ambegaonkar and Irvine, 2015). This ability of Pkpk and Pksple to colocalize has not been previously observed in wildtype conditions. Notably, however, in the wing, ectopic Pksple localization follows the expected position of Pkpk when Ds and Dachs cues are removed, though each were not independently visualized (Ambegaonkar and Irvine, 2015). Conversely, Pksple overexpression was seen to recruit Pkpk to the distal side of wing cells (Ayukawa et al., 2014), reversing hair polarity as it does bristle polarity. We suggest that the temporal expression pattern in the AWM allows the system to initiate polarization under Pkpk control, and that the gradually accumulating Pksple colocalizes with Pkpk rather than outcompeting established proximal localization. Because bristles in pksple mutant wings fully polarize, the proximal Pksple is inconsequential for normal bristle polarization.

In contrast, overexpression of Pksple, producing early and sustained high level expression, enables it to outcompete endogenous Pkpk and reverse polarity by driving localization to the distal side through its interaction with Ds and Dachs, likely recruiting Pkpk along with it. Similarly, in pkpk mutants, endogenous Pksple, free from recruitment to the proximal side, localizes distally and reverses polarity. We infer that during the critical period for determining bristle rotation direction in wildtype, Pksple does not reach a sufficient level to outcompete Pkpk and reverse the rotation.

That pkpk mutation only reverses polarity of a region of distal bristles, whereas Pksple overexpression can reverse polarity of most or all AWM bristles, indicates that endogenous Pksple is only poised to act in a limited region of the margin. This may reflect subtle differences in the timing of its expression increase across the margin. Alternatively, it may reflect differences in the strength of the Ft/Ds/Fj signal across the margin. Our analyses do not have sufficient resolution to distinguish these possibilities.

Dynamic isoform expression appears to have important consequences for other aspects of wing development. Hair polarity is determined by Pkpk (at around 32 hr apf), but it can be inferred that some Pksple is already present, as is evident from the difference between the hair polarity patterns of pkpk and pkpk-sple mutants (Gubb et al., 1999), and as confirmed by our expression analyses. We suggest that hair polarity does not fully reverse in pkpk mutants either because levels of Pksple are not yet high enough or because expression is primarily in veins and at the AWM at the time hair polarity is fixed. In contrast, the polarity of ridges, established later in wing development [(Merkel et al., 2014); (Doyle et al., 2008) notwithstanding], depends on Pksple. We propose that by the time ridge polarity is determined, the amount of Pksple has increased and the amount of Pkpk has decreased sufficiently to allow Pksple to exert control of ridge polarization. Though likely unimportant for normal development, the somewhat earlier expression of Pksple in veins relative to the intervein regions may contribute to polarity discontinuities observed in pkpk mutant wings, especially around L3 (Gubb and García-Bellido, 1982; Hogan et al., 2011; Merkel et al., 2014). Ft-Ds polarity appears to also be distorted around veins (Merkel et al., 2014). Pkpk and Pksple expression dynamics are likely at play in determining the PCP response in other tissues as well.

The ft/Ds/Fj global signal orients both Pkpk-dependent and Pksple-dependent core PCP signaling in the wing

The idea that Ds controls the direction of core PCP signaling was first proposed by Adler based on wing hair polarity phenotypes (Adler et al., 1998). We subsequently studied the Ft/Ds/Fj system and similarly concluded that it directs core PCP protein localization in the wing (Ma et al., 2003), a Pkpk-dependent process, and polarization of ommatidia in the eye (Yang et al., 2002), a Pksple-dependent process. We proposed that coupling in wing hair polarization is necessarily weak (Ma et al., 2003), and the more recently proposed model in which Ft/Ds/Fj orient microtubules to orient directional trafficking of Fz, Dsh and Fmi-containing vesicles (Matis et al., 2014; Olofsson and Axelrod, 2014; Shimada et al., 2006; Harumoto et al., 2010) is consistent with a weak coupling mechanism in Pkpk-dependent processes. The finding of direct binding of Pksple to Ds and Dachs (Ayukawa et al., 2014; Ambegaonkar and Irvine, 2015) suggests a model for more direct and potentially stronger coupling of Pksple-dependent processes to the Ft/Ds/Fj system. The idea of coupling in Pk-dependent signaling has been controversial, and based largely on correlation, subsequent studies have led to the argument that Pksple-dependent core signaling is coupled, but Pkpk-dependent signaling is uncoupled from the the Ft/Ds/Fj system (Merkel et al., 2014; Ambegaonkar and Irvine, 2015). Yet others have suggested that the Ft/Ds/Fj and core PCP systems always function in parallel rather than being coupled (Lawrence et al., 2007; Casal et al., 2006). Here, we report strong evidence that Pkpk-dependent core PCP signaling is responsive to the Ds signal, at least in polarizing wing hairs. We propose that the same is the case in polarizing bristles that are controlled by essentially similar responses to Pk isoforms and to the Ft/Ds/Fj system.

In bristles, we directly observe the requirement for Ds to distally localize Pksple when Pkpk is absent, confirming Pksple coupling. The evidence that Pkpk-dependent core signaling is coupled to the upstream Ft/Ds/Fj signal is less apparent. In wildtype, correct bristle polarization requires Pkpk to prevent reversal by recruiting Pksple to the proximal side, though as noted above, proximal Pksple plays no essential role. But absent the need to antagonize Pksple coupling, is there evidence that core signaling in the presence of just Pkpk (pksple mutant) is coupled to Ft/Ds/Fj? When the Ft/Ds/Fj system is intact, Pkpk localizes proximally, but without Ds or Ft, Pkpk localizes proximally and distally in random domains, driving domains of correct and reversed rotation analogous to the random but locally correlated domains of hair polarity in ft or ds mutant wing tissue (Adler et al., 1998; Ma et al., 2003). The same random domains of bristle polarity are seen when only Pkpk is available (ds knockdown in a pksple mutant). This result demonstrates that the Ft/Ds/Fj system is required for correct polarization of the core PCP system while solely under Pkpk control, though it cannot distinguish a permissive from an instructive role. An instructive role is, however, concordant with its instructive role in directing Pkpk-dependent wing hair polarity.

The proposal that Pkpk-dependent core signaling is coupled to and responds to the Ft/Ds/Fj signal in bristle polarization might at first appear to conflict with the observation that properly oriented rotation proceeds to a significant extent in the absence of both Pkpk and Pksple. This is explained by pointing out that our model for coupling invokes Ft/Ds/Fj directed microtubule-based transport of Fz and Dsh, but that the involvement of Pkpk is indirect (Matis et al., 2014; Olofsson and Axelrod, 2014; Shimada et al., 2006; Harumoto et al., 2010). As have others, we propose that Pkpk functions to amplify the asymmetry introduced by this transport, but that some asymmetry, and communication of polarity information between cells, can still occur in its absence (Strutt and Strutt, 2007; Lawrence et al., 2004; Adler et al., 2000). In other words, Ft/Ds/Fj coupling to Pkpk-dependent core PCP signaling is not directed by Pkpk, but rather, is permitted to occur because the Pksple-dependent mechanism is not operating to override it. Because core module function is required, this activity does not result from Pk isoform action influencing Ft/Ds/Fj output independent of core signaling, as has been recently suggested in another context (Casal et al., 2018). It is important to caution that the model for the relationship between Ft/Ds/Fj and core signaling presented here does not necessarily extend to their relationship in other tissues where their interactions may well be different, and that experiments done in other tissues may not be directly relevant to wing hair and AWM bristles.

The results presented here indicate that mapping the spatiotemporal dynamics of Pk isoform expression is essential to understanding how various developmental events can be differentially coupled to upstream global directional signals in a given tissue.

A hypothesis for translating PCP into organ rotation

Chirality, or left-right laterality, is a key feature of many organs in invertebrates and vertebrates. In Drosophila, rotation of the gut and of the male genitalia occurs in a defined direction to produce such laterality. In vertebrates, rotation of the gut and heart tube also leads to left-right asymmetry in these organs. In many cases, PCP has been implicated in control of this lateralization (Blum and Ott, 2018). In the Drosophila hindgut, both core PCP and the Ft/Ds system play essential roles in directing normal dextral rotation (González-Morales et al., 2015). Though the forces that drive these organ rotations are not well understood, left-right asymmetries in actomyosin distribution, cell shape, and localization of other cellular structures, together with PCP dependence (Harris, 2018; Blum and Ott, 2018), indicate that chirality at the cellular level is an important determinant of rotational direction. Indeed, chirality of isolated cells from looping chick heart has been directly demonstrated (Ray et al., 2018). We therefore propose that regulation of chiral shaft-socket cell pair rotation may share much in common with the mechanisms that determine larger organ laterality, and its investigation could therefore yield insights that will enlighten understanding of organ rotation and laterality.

Materials and methods

Key resources table
Reagent type
(species) or resource
DesignationSource or referenceIdentifiersAdditional
information
Genetic reagent (Drosophila melanogaster)pkpk-sple13Gubb et al., 1999, PMID: 10485852BDSC:41790;
FLYB:FBal0060943;
RRID:BDSC_41790
FlyBase symbol: pkpk-sple-13
Genetic reagent (Drosophila melanogaster)pkpk-sple14Gubb et al., 1999, PMID: 10485852FLYB:FBal0035401FlyBase symbol: pkpk-sple-14
Genetic reagent (Drosophila melanogaster)pkpk30Gubb et al., 1999, PMID: 10485852BDSC:44229;
FLYB:FBal0101223;
RRID:BDSC_44229
FlyBase symbol: pk30
Genetic reagent (Drosophila melanogaster)pksple1Gubb et al., 1999, PMID: 10485852BDSC:422;
FLYB:FBal0016024;
RRID:BDSC_422
FlyBase symbol: pksple-1
Genetic reagent (Drosophila melanogaster)vangA3Taylor et al., 1998, PMID: 9725839FLYB:FBal0093183FlyBase symbol:
VangA3
Genetic reagent (Drosophila melanogaster)vangstbm6Wolff and Rubin, 1998, PMID: 9463361BDSC:6918;
FLYB:FBal0062424;
RRID:BDSC_6918
FlyBase symbol:
Vangstbm-6
Genetic reagent (Drosophila melanogaster)fzR52Krasnow and Adler, 1994, PMID: 7924994FLYB:FBal0004939FlyBase symbol: fz23
Genetic reagent (Drosophila melanogaster)dsh1Bloomington Drosophila Stock CenterBDSC:5298;
FLYB:FBal0003138;
RRID:BDSC_5298
FlyBase symbol: dsh1
Genetic reagent (Drosophila melanogaster)UAS-pkspleBloomington Drosophila Stock CenterBDSC:41780;
FLYB:FBti0148928;
RRID:BDSC_41780
FlyBase symbol:
P{UAS-sple+}3
Genetic reagent (Drosophila melanogaster)UAS-pkRNAiVienna Drosophila Resource CenterVDRC:v101480; FLYB:FBst0473353; RRID:FlyBase_FBst0473353FlyBase symbol:
P{KK109294}VIE-260B
Genetic reagent (Drosophila melanogaster)UAS-fmiRNAiBloomington Drosophila Stock CenterBDSC:26022;
FLYB:FBti0114752;
RRID:BDSC_26022
Flybase symbol:
P{TRiP.JF02047}attP2
Genetic reagent (Drosophila melanogaster)UAS-fzRNAiBloomington Drosophila Stock CenterBDSC:34321;
FLYB:FBti0140932;
RRID:BDSC_34321
Flybase symbol:
P{TRiP.HMS01308}attP2
Genetic reagent (Drosophila melanogaster)UAS-vangRNAiBloomington Drosophila Stock CenterBDSC:34354;
FLYB:FBti0140967;
RRID:BDSC_34354
Flybase symbol:
P{TRiP.HMS01343}attP2
Genetic reagent (Drosophila melanogaster)UAS-dsRNAiBloomington Drosophila Stock CenterBDSC:32964;
FLYB:FBti0140473;
RRID:BDSC_32964
Flybase symbol:
P{TRiP.HMS00759}attP2
Genetic reagent (Drosophila melanogaster)UAS-dsMatakatsu and Blair, 2004, PMID: 15240556FLYB:FBtp0019964Flybase symbol:
P{UAS-ds.T}
Genetic reagent (Drosophila melanogaster)dll-GAL4Bloomington Drosophila Stock CenterBDSC:3038;
FLYB:FBti0002783;
RRID:BDSC_3038
Flybase symbol:
P{GawB}Dllmd23
Genetic reagent (Drosophila melanogaster)MS1096-GAL4Bloomington Drosophila Stock CenterBDSC:8860;
FLYB:FBti0002374;
RRID:BDSC_8860
Flybase symbol:
P{GawB}BxMS1096
Genetic reagent (Drosophila melanogaster)armP-fz::EGFPStrutt, 2001, PMID:11239465FLYB:FBtp0014592Flybase symbol:
P{arm-fz.GFP}
Genetic reagent (Drosophila melanogaster)actP-vang::EYFPStrutt, 2002, PMID: 12137731FLYB:FBtp0015854Flybase symbol:
P{Act5C(-FRT)stbm-EYFP}
Genetic reagent (Drosophila melanogaster)actP > CD2>vang::EYFPStrutt, 2002, PMID: 12137731FLYB:FBtp0084387Flybase symbol:
P{Act5C(FRT.polyA)stbm-EYFP}
Genetic reagent (Drosophila melanogaster)ci-GAL4Croker et al., 2006, PMID: 16413529FLYB:FBtp0057188Flybase symbol:
P{ci-GAL4.U}
Genetic reagent (Drosophila melanogaster)UAS-mCherryBloomington Drosophila Stock CenterBDSC:38424;
FLYB:FBti0147460;
RRID:BDSC_38424
Flybase symbol:
P{UAS-mCherry.NLS}3
Genetic reagent (Drosophila melanogaster)actP > CD2>Gal4Bloomington Drosophila Stock CenterBDSC:30558;
FLYB:FBti0012408;
RRID:BDSC_30558
Flybase symbol:
P{GAL4-Act5C(FRT.CD2).P}S
Genetic reagent (Drosophila melanogaster)UAS-RFPBloomington Drosophila Stock CenterBDSC:30558;
FLYB:FBti0129814;
RRID:BDSC_30558
Flybase symbol:
P{UAS-RFP.W}3
Antibodygoat polyclonal anti-Su(H)Santa CruzSanta Cruz:sc-15183
RRID:AB_672840
1/200 (immunolabelling)
AntibodyMouse monoclonal anti-V5Thermo-FisherThermo_Fisher:R960-25, RRID:AB_25565641/200 (immunolabelling)
1/1000 (Western blotting)
AntibodyGuinea pig polyclonal anti-Pk[C]Olofsson et al., 2014,
PMID: 25005476
N/A1/800 (immunolabelling)
1/1000 (Western blotting)
AntibodyRat monoclonal anti-dEcadDSHBRRID:AB_5281201/200 (immunolabelling)
AntibodyMouse monoclonal anti-γ-TubulinSigma-AldrichSigma-Aldrich: T6557
RRID:AB_477584
1/1000 (Western blotting)
Recombinant DNA reagentpCFD4AddgeneRRID:Addgene_49411CRISPR gRNA backbone
Recombinant DNA reagentpDsRedattpAddgeneRRID:Addgene_51019Donor recombinant DNA backbone
Recombinant DNA reagentpCR-Blunt-II-TOPOThermo-FisherRRID:Addgene_29705Backbone for sub-cloning
Sequence-based reagentpkpk gRNA 1This papergRNA sequence in PCR primersATGGCTCAGGCCCGATCTAG
Sequence-based reagentpkpk gRNA 2This papergRNA sequence in PCR primersGTGGATCAACCCCTGGAAAC
Sequence-based reagentpksple gRNA 1This papergRNA sequence in PCR primersCTCGTAAATTTAGCTTCGAG
Sequence-based reagentpksple gRNA 2This papergRNA sequence in PCR primersAGATGCAATTTGGCCGCCCT

Drosophila genetics

Request a detailed protocol

Drosophila melanogaster flies were grown on standard cornmeal/agar/molasses media at 25°C. FLP-on (using the actP >CD2>GAL4 construct for trans-gene expression) and FLP/FRT mitotic clones were generated by incubating third-instar larvae at 37°C for 1 hr. 36 to 48 hr later, white prepupae were collected and aged to the desired developmental time point prior to dissection and fixation.

Drosophila mutant alleles and transgenic stocks are described in the Key resources table and detailed chromosomes and genotypes are provided below. pkpk-sple13 (FBst0044230), pkpk-sple14 (Gubb, 1993), pkpk30 (FBst0044229), pksple1 (FBst0000422), vangA3 (Taylor et al., 1998), vangstbm6 (FBst0006918), fzR52 (Krasnow and Adler, 1994), dsh1 (FBst0005298), UAS-pksple (FBst0041780), UAS-pkRNAi (VDRC ID: 101480), UAS-fmiRNAi (FBst0026022), UAS-fzRNAi (FBst0034321), UAS-vangRNAi (FBst0034354), UAS-dsRNAi (FBst0032964), UAS-ds (Matakatsu and Blair, 2004), dll-Gal4 (FBst0030558), MS1096-Gal4 (FBst0008860), armP-fz::EGFP (Strutt, 2001), actP-vang::EYFP (Strutt, 2002), actP >CD2>vang::EYFP (Strutt, 2002), ci-Gal4 (Croker et al., 2006), UAS-mCherry (FBst0038424), actP >CD2>Gal4 UAS-RFP (FBst0030558).

Genotypes of experimental models

Figure 1

Request a detailed protocol
  • (A, C) w1118

  • (B, D) y w hsflp/+;; actP >CD2>GAL4 UAS-RFP/+

Figure 2

Request a detailed protocol
  • (A) pkpk30/pkpk30

  • (B) y w hsflp/+; pkpk30/pkpk30; actP >CD2>GAL4 UAS-RFP/+

Figure 3

Request a detailed protocol
  • (A) y w hsflp/+;; actP >CD2>GAL4 UAS-RFP/+

  • (D) w1118

  • (E) pkpk30/pkpk30

Figure 4

Request a detailed protocol
  • (A, C, E) V5::3Xmyc::APEX2::pkpk/V5::3Xmyc::APEX2::pkpk

  • (B, D, F) V5::3Xmyc::APEX2::pksple/V5::3Xmyc::APEX2::pksple

  • (G) pkpk30/pkpk30

  • (H) y w hsflp/+; V5::3Xmyc::APEX2::pkpk/UAS-pkRNAi; actP >CD2>GAL4 UAS-RFP/+

  • (I) y w hsflp/+; V5::3Xmyc::APEX2::pksple/UAS-pkRNAi; actP >CD2>GAL4 UAS-RFP/+

Figure 5

Request a detailed protocol
  • (A) w1118

  • (B) pkpk30/pkpk30

  • (C) MS1096-GAL4/+; pkpk30/pkpk30;UAS-fzRNAi/+

  • (D) MS1096-GAL4/+; pkpk30/pkpk30;UAS-vangRNAi/+

  • (E) MS1096-GAL4/+;;UAS-fmiRNAi/+

  • (F) fzR52/fzR52

  • (G) vangstbm6/vangstbm6

  • (H) dsh1/dsh1

  • (I) armP-fz::EGFP/armP-fz::EGFP

  • (J) actP-vang::EYFP/actP-vang::EYFP

  • (K) y w hsflp/+; FRT42D armP-fz::EGFP/FRT42D

  • (L) y w hsflp/+; FRT42D actP-vang::EYFP/FRT42D

Figure 6

Request a detailed protocol
  • y w hsflp/+; armP-fz::EGFP/+; actP >CD2>GAL4 UAS-RFP/UAS-fzRNAi

Figure 7

Request a detailed protocol
  • (A) w1118

  • (B) pkpk30/pkpk30

  • (C) MS1096-GAL4/+; V5::3Xmyc::APEX2::pksple/+; UAS-dsRNAi/+

  • (D) MS1096-GAL4/+; pkpk30/pkpk30; UAS-dsRNAi/+

  • (E) MS1096-GAL4/+; pksple1/pksple1; UAS-dsRNAi/+

  • (F) MS1096-GAL4/+; pkpk-sple13/pkpk-sple13; UAS-dsRNAi/+

  • (G) w1118 and dll-GAL4/+; UAS-ds/UAS-ds

  • (H) fzR52/fzR52 and dll-GAL4/+; fzR52 UAS-ds/fzR52 UAS-ds

  • (I) pkpk30/pkpk30 and pkpk30 dll-GAL4/pkpk30; UAS-ds/UAS-ds

  • (J) pkpk-sple13/pkpk-sple14 and dll-GAL4 pkpk-sple13/pkpk-sple14; UAS-ds/UAS-ds

  • (K) pksple1/pksple1 and pksple1 dll-GAL4/pksple1; UAS-ds/UAS-ds

Figure 1—figure supplement 1

Request a detailed protocol
  • (B) w1118

  • (C) pkpk30/pkpk30

Figure 2—figure supplement 1

Request a detailed protocol
  • (A2) w1118 and pkpk30/pkpk30

  • (B) actP-vang::EYFP/+

Figure 3—figure supplement 1

Request a detailed protocol
  • (A) w1118

  • (B) pkpk30/pkpk30

  • (C) pksple1/pksple1

  • (D) pkpk-sple13/pkpk-sple13

  • (E) MS1096-GAL4/+;;UAS-pksple/+

Figure 4—figure supplement 1

Request a detailed protocol
  • (A) V5::3Xmyc::APEX2::pkpk/V5::3Xmyc::APEX2::pkpk,

  • V5::3Xmyc::APEX2::pksple/V5::3Xmyc::APEX2::pksple

  • (B)Lane 1: pkpk-sple13/pkpk-sple13, other lanes: w1118

  • (C-H) V5::3Xmyc::APEX2::pkpk/V5::3Xmyc::APEX2::pkpk; ci-GAL4 UAS-mCherry/+

  • (K)V5::3Xmyc::APEX2::pkpk/V5::3Xmyc::APEX2::pkpk

  • (I-J, L) V5::3Xmyc::APEX2::pksple/V5::3Xmyc::APEX2::pksple

Figure 4—figure supplement 2

Request a detailed protocol
  • (A) V5::3Xmyc::APEX2::pkpk/V5::3Xmyc::APEX2::pkpk,

  • (B) Left: V5::3Xmyc::APEX2::pksple/V5::3Xmyc::APEX2::pksple; actP-vang::EYFP/+

  • Middle and right: V5::3Xmyc::APEX2::pksple/V5::3Xmyc::APEX2::pksple

  • (C) pkpk30/pkpk30

  • (D) y w hsflp/+; pkpk30/pkpk30 UAS-pkRNAi; actP >CD2>GAL4 UAS-RFP/+

Figure 5—figure supplement 1

Request a detailed protocol
  • y w hsflp/+; pkpk30/pkpk30; actP >CD2>vang::EYFP/+

Figure 6—figure supplement 1

Request a detailed protocol
  • y w hsflp/+; actP-vang::EYFP/+; actP >CD2>GAL4 UAS-RFP/UAS-vangRNAi

Figure 7—figure supplement 1

Request a detailed protocol
  • (A) MS1096-GAL4/+;; UAS-dsRNAi/+

  • (B) MS1096-GAL4/+; pkpk30/pkpk30; UAS-dsRNAi/+

  • (C) MS1096-GAL4/+; pksple1/pksple1; UAS-dsRNAi/+

  • (D) MS1096-GAL4/+; pkpk-sple13/pkpk-sple13; UAS-dsRNAi/+

  • (E) dll-GAL4/+; UAS-ds/UAS-ds

  • (F–I) dll-GAL4/+; UAS-ds/UAS-mCherry

Immunohistochemistry

Request a detailed protocol

Pupal wings were dissected at indicated developmental time points after puparium formation (apf). Pupae were removed from their pupal cases and fixed for 60–90 min in 4% paraformaldehyde in PBS at 4°C. Wings were then dissected and extracted from the cuticle, and were washed two times in PBST (PBS with 0.1% Triton X-100). After blocking for 1 hr in 5% Bovine serum Albumin in PBST at 4°C, wings were incubated with primary antibodies overnight at 4°C in the blocking solution. Incubations with secondary antibodies were done for 90 min at room temperature in PBST. Washes in PBST were performed three times after primary and secondary antibody incubation, and incubations in phalloidin (1:200 dilution) in PBST were done for 15 min followed by wash at room temperature before mounting if required. Stained wings were mounted in 15 μl Vectashield mounting medium (Vector Laboratories). Primary antibodies were as follows: goat polyclonal anti-Su(H) (1:200 dilution, Santa Cruz, sc-15183), mouse monoclonal anti-V5 (1:200 dilution, Thermo-fisher, R960-25), guinea pig polyclonal anti-Pk[C] (1:800 dilution, Olofsson et al., 2014), rat monoclonal anti-dEcad (1:200 dilution, DSHB). Secondary antibodies from Thermo Fisher Scientific were as follows: 488-donkey anti-mouse, 488-goat anti-guinea pig, 546-donkey anti-goat, 633-goat anti-guinea pig, 633-goat anti-rat, 647-donkey anti-mouse. Alexa 635 and Alexa 350 conjugated phalloidin were from Thermo Fisher Scientific.

Imaging and quantification

Request a detailed protocol

Adult wings were dissected and washed with 70% EtOH and mounted in DPX (Sigma) solution. All adult wings were imaged on a Nikon Eclipse E1000M equipped with a Spot Flex camera (Model 15.2 64 MP). All immunofluorescence images were taken with a Leica TCS SP8 AOBS confocal microscope and processed with LAS X (Leica) and Adobe Photoshop. For three dimensional wing margin images, 50 to 100 z-stacks, each with 0.2 μm thickness, were collected and combined using 3D reconstitution software (Leica). Scale bars are not provided for three dimensional images due to errors introduced by perspective, but approximate scale can be inferred from related two- dimensional images. To measure the apical rotation angles of socket cells, a horizontal line linking centers of circles around apical surfaces of socket cells was drawn, and perpendicular lines intersecting the center of each socket cell apex were drawn (black lines in Figure 3C). Vectors from each socket cell center passing through the center of the apical opening of the socket circles (blue vectors in Figure 3C) were drawn and angles between the vertical lines and the vectors were measured with Image J software. Statistical analysis was performed and rose plots generated using Oriana four software. Comparisons were made using Student’s t-test and p values are reported. Summary statistics are provided in Table 2. For qualitative results such as expression patterns, a minimum of 20 biological replicates from at least two independent experiments were examined and representative images are shown.

CRISPR/Cas9 homology directed recombination for tagging V5 sequence to pkpk and pksple genomic locus

Construction of gRNA containing plasmids

Request a detailed protocol

pkpk gRNA1 (5’-ATGGCTCAGGCCCGATCTAG-3’) and pkpk gRNA2 (5’- GTGGATCAACCCCTGGAAAC-3’) were assembled into pCFD4 plasmid (Addgene, 49411) digested by the BbsI restriction enzyme using Gibson Assembly (NEB) to express two gRNAs under the U6 promoter. The same procedure was carried out to assemble two pksple gRNAs, pksple gRNA1 (5’-CTCGTAAATTTAGCTTCGAG-3’) and pksple gRNA2 (5’- AGATGCAATTTGGCCGCCCT-3’), into pCFD4. Stable transgenic flies expressing two pkpk gRNAs or two pksple gRNAs were generated by BestGene using the PhiC31 standard injection method.

Construction of donor plasmids containing two homology arms and V5::3Xmyc::APEX2 tag sequences

Request a detailed protocol

To generate the donor plasmid with homology arms of the pkpk genomic sequence and the V5::3Xmyc::APEX2 tag sequence, a 1.5 kb 3’ homology arm (HR2) flanking the pkpk gRNA2 cleavage site was amplified and assembled into the pDsRed-attp (Addgene, 51019) plasmid cut with SapI to make the pDsREd-attP-pkpkHR2 plasmid.

To generate the donor plasmid for tagging pkpk, three DNA fragments including 5’ 1.5 kb homology arm (HR1; containing a 1.2 kb homology arm flanking the pkpk gRNA1 cleavage site and a 5’ 0.3 kb sequence of the start codon), V5::3Xmyc::APEX2 tag with a linker sequence, and the fragment starting from the start codon of pkpk to the cleavage site targeted by the pkpk gRNA2, were assembled into the pDsREd-attP-pkpkHR2. To prevent the donor sequence from being cleaved by Cas9, a point mutation was introduced in the PAM sequence of the HR1 using the NEB point-mutagenesis kit after sub-cloning the fragment into the pCR-Blunt-II-TOPO vector (Thermo-Fisher, K280002). The HR1 fragment bearing the mutant PAM sequence was then amplified for the assembly process. All three fragments were assembled into the pDsREd-attP-pkpkHR2 plasmid cut with EcoRI and NheI.

To generate the donor plasmid for tagging pksple, similar strategies were applied. Briefly, 1.2 kb 5’ homology arm containing the mutant PAM sequence, the V5::3Xmyc::APEX2 tag with a linker sequence, and the fragment from the start codon of pksple to the cleavage site of pksple gRNA2 were assembled into the pDsREd-attP-pkspleHR2 (bearing the 1.25 kb 3’ homology arm, HR2) plasmid.

The donor plasmids containing the tag sequence and pkpk homology, or pksple homology, arms, were sequenced and then injected into the stable transgenic flies expressing two pkpk gRNAs, or pksple gRNAs, and nosCas9, respectively, to generate recombinants. DsRed signal in the fly eyes was monitored for selecting the recombinants by BestGene, and dsRed and flanking sequences were removed by the Cre-Lox site-specific recombination method. The resulting modified alleles are referred to in the text as V5::Pk and V5::Sple for simplicity.

Western blots

Request a detailed protocol

Third-instar larval wing discs and pupal wings at appropriate developmental stages were dissected and lysed in protein loading buffer. Lysates from eight discs or wings were loaded per lane for SDS-PAGE analysis and western blots were performed using standard procedures. Antibodies: Guinea pig polyclonal anti-Pk[C] (1:1000 dilution, the same antibody used for immunostaining), mouse monoclonal anti-V5 (1:2000 dilution, the same antibody used for immunostaining), mouse monoclonal anti-γ-Tubulin (1:1000 dilution, Sigma-Aldrich, T6557). Secondary antibodies were Peroxidase-conjugated goat anti-guinea pig (1:10000) and goat anti-mouse (1:10000) antibodies (both from Jackson Immuno Research), and detection used SuperSignal West Pico Chemiluminescent Substrate (Thermo-Fisher, 34080)

References

  1. 1
    Mutations in the cadherin superfamily member gene dachsous cause a tissue polarity phenotype by altering frizzled signaling
    1. PN Adler
    2. J Charlton
    3. J Liu
    (1998)
    Development 125:959–968.
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
    Genes controlling cellular polarity in Drosophila
    1. D Gubb
    (1993)
    Development 119:269–277.
  20. 20
  21. 21
    A genetic analysis of the determination of cuticular polarity during development in Drosophila Melanogaster
    1. D Gubb
    2. A García-Bellido
    (1982)
    Journal of Embryology and Experimental Morphology 68:37–57.
  22. 22
  23. 23
    Development of adult sensilla on the wing and notum of Drosophila Melanogaster
    1. V Hartenstein
    2. JW Posakony
    (1989)
    Development 107:389–405.
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
    A single frizzled protein has a dual function in tissue polarity
    1. RE Krasnow
    2. PN Adler
    (1994)
    Development 120:1883–1893.
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
    Van Gogh: a new Drosophila tissue polarity gene
    1. J Taylor
    2. N Abramova
    3. J Charlton
    4. PN Adler
    (1998)
    Genetics 150:199–210.
  50. 50
    Strabismus, a novel gene that regulates tissue polarity and cell fate decisions in Drosophila
    1. T Wolff
    2. GM Rubin
    (1998)
    Development 125:1149–1159.
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55

Decision letter

  1. Danelle Devenport
    Reviewing Editor; Princeton University, United States
  2. Utpal Banerjee
    Senior Editor; University of California, Los Angeles, United States
  3. Danelle Devenport
    Reviewer; Princeton University, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

Using the mechanosensory bristles of the Drosophila wing, Cho et al. describe a new mode of planar polarized morphogenesis. The authors show that bristle orientation arises when the two major cell types that comprise each bristle – the socket and shaft cell – rotate with a defined chirality like a corkscrew and become intertwined. Rotation is dependent on the core PCP pathway and, strikingly, the chirality of rotation depends on which of two Prickle isoforms (Pk or Sple) predominates. By investigating how endogenously-tagged Prickle isoforms respond to changes in the Fat/Dachsous PCP module, they suggest that the two Pk isoforms orient the core PCP module relative to the Fat/Dachsous directional input in opposite ways. Thus, this paper defines a new PCP-dependent morphogenetic process and provides further insights into the intriguing interplay between Prickle isoforms in PCP. Readers broadly interested in cell polarity and morphogenesis will benefit from reading these new contributions to the field.

Decision letter after peer review:

Thank you for submitting your article "Prickle isoforms determine handedness of helical morphogenesis" for consideration by eLife. Your article has been reviewed by three peer reviewers, including Danelle Devenport as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Utpal Banerjee as the Senior Editor.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

Cho et al. investigate how mechanosensory bristles of the Drosophila wing margin are planar polarized, finding that PCP-dependent rotation of socket and shaft cells, rather than asymmetric positioning of daughter cells, determines bristle orientation. Prickle (Pk) isoform expression appears to determine the chirality of this rotation, and the authors argue that the two Pk isoforms orient the core PCP module relative to the Fat/Dachsous directional input in opposite ways. Overall, the reviewers felt that the documentation of PCP-mediated helical morphogenesis together with further insights into the intriguing interplay between Pk isoforms makes an important contribution to the field. They raised some issues about data presentation, quantification and statistical analysis that should be addressed prior to publication.

The reviewers were also in agreement that the section on Pkpk coupling to the Fat/Ds system was weak and detracted from the rest of the manuscript. The simplest solution would be to remove this section from the revised manuscript so it can be elaborated in a separate study. However, in case the authors do want to address the issues with this section, those comments can be found at the end and separated from the rest of reviewer points.

Reviewer 1:

1) Figure 4H-I. Mosaic knockdown of pk by RNAi is used to distinguish which side of the cell junction Pk and Sple isoforms reside. The authors conclude that both isoforms are proximal. This is an important result as it is different from previously published studies overexpressing Sple and should be shown more clearly. It is very difficult to distinguish the boundaries between expressing and non-expressing cells, especially considering that margin cells may contribute to the signal.

2) Figure 4—figure supplement 1A-B. The loading control levels are highly variable making it difficult to interpret how Pk and Sple isoform levels change over time. Perhaps tubulin is not an appropriate loading control if it levels too change over time. Can the authors plot Pk and Sple levels relative to Tub loading control? Or use V5 fluorescent intensities from confocal images to compare levels?

3) Figure 4. The authors conclude from the localization of Pk and Sple isoforms "that distal localization of Pksple in Pkpk mutants drives counter-clockwise rotation of shaft-socket pairs." In addition, the title of this section is "Localization of Pkpk and/or Pksple determines handedness of helical rotation". I think these are overstated conclusions. The localization certainly correlates with the direction of rotation, but this seems insufficient to conclude this drives rotation, especially since elimination of both isoforms has very little effect on rotation.

4) Figure 7D. This experiment shows that reversed Sple localization in Pkpk mutants depends on Ds, and "we interpret this to indicate that the Ds global signal recruits Pksple to sites of enriched Ds (distal) in the absence of Pkpk". What about in the proximal part of the wing where bristle polarity is normal? Is Sple localized proximally, and independent of Ds? How do the authors interpret the distally restricted phenotype of Pkpk mutants? Is the Ds signal stronger distally?

5) Coupling between the Fz-Vang and Ft-Ds pathways by Pk isoforms. "we interpret this to indicate that the Ds global signal recruits Pksple to sites of enriched Ds (distal) in the absence of Pkpk (Figure 7B; compare with 7D), which drives reversal of shaft-socket orientation and therefore reversal of bristle polarity." Are Fz and Vang localizations flipped in socket cells with reversed polarity in Pkpk mutants? Although this has been shown previously for wing hairs overexpressing Sple, I feel it is important to show here under endogenous Sple levels to definitely show the core PCP module is flipped, not just a downstream component.

Reviewer 2:

1) A "helical structure of defined chirality" may not be appropriate. It seems to be an anisotropic cell rearrangement directed by distal-proximal and dorsal-ventral polarity. Such rearrangement should change the interface between shaft and socket as shown in Figure 3B and C.

2) It is difficult to evaluate the phenotypic analyses without statistical information. What are the percentages of wings showing the representative phenotypes? How may socket cells were scored from how many wings? How many clones were analyzed and what was the percentages showing the described phenotypes? Some of them can be found in source data files. However, it should be also stated in the text.

3) Figure 4H-I', Figure 2—figure supplement 1B, and Figure 4—figure supplement 2 were unclear. These data should be shown more clearly.

Reviewer 3:

1) As a general comment, it might be helpful to readers to explain that in processes such as left-right asymmetry determination (analogous to the chiral morphogenesis being discussed here) there is symmetry breaking on three axes: apical-basal, and two axes in the plane of the tissue. Hence, I presume PCP specifies clockwise or anticlockwise morphogenesis by specifying polarity in one axis of the plane, in concert with a second polarity cue in the orthogonal axis in the plane (which I take to be an AP oriented cue, ultimately set up by the Hedgehog pathway?). Without this second cue, the PCP cue alone wouldn't be sufficient to specify chirality?

2) Introduction paragraph three: I think consistent nomenclature for proteins in flies is capitalized Roman so should be "Prickle" and "Spiny-legs" (and probably "PkPk" and "PkSple")? (Also some genotypes change to all caps italics on figure panels e.g. compare Figure 5E-F to 5G-H).

3) Results paragraph one: typo "Supressor"

4) Figure 4—figure supplement 1A is mislabeled? The blue asterisks ("unidentified") and the purple arrow ("Pk") should be swapped? This would then be consistent with panel B?

5) Subsection “Pkpk and Pksple differentially interpret directional signals from Ds” imply an "indirect" mechanism for PkPk coupling to Ft-Ds, but the Introduction (I think more in line with previous work, including from the authors) suggest an indirect mechanism for core pathway coupling to Ft-Ds via Fz/Dsh/Fmi transport with (presumably) PkPk being required for magnitude of core pathway activity but not for orientation? This is a complex area, but I think it would be easier for readers if the authors stuck to one view. At least based on the data so far, we don't see any evidence that PkPk couples, only that it enhances core pathway activity?

6) Paragraph three of subsection “Pkpk and Pksple differentially interpret directional signals from Ds”: the authors deduce that in the absence of PkPk, PkSple is recruited to distal cell edges by distally localised Ds, but I find this confusing (at least without further clarification) as Merkel et al. seemed to show pretty clearly that – at least in the wing blade – Ds is oriented primarily on the AP axis from 16h APF onwards, and trichomes orient primarily on the AP axis in pkpk as well (e.g. Merkel et al. and Gubb et al.). So I would expect Sple to be recruited largely to AP cell edges (maybe with a slight distal bias?) not to distal cell edges. But maybe something happens on the edge of the wing blade to tilt the Ds polarity axis more PD? What do the authors think?

7) Paragraph four subsection “Pkpk and Pksple differentially interpret directional signals from Ds”: this appears to be a key sentence, as here the authors appear to argue that ds knockdown affects a PkPk-mediated signal for bristle polarity, however I completely failed to follow the logic that indicates PkPk is providing a polarity cue rather than contributing to polarity magnitude (see paragraph two of subsection “Pkpk versus Pksple isoform expression determines the direction of rotation” and Table 3) and suppressing effects of "ectopic" Sple.

– The key piece of data is (I think) that ds- gives reversals of bristle polarity, and (perhaps surprisingly) these are suppressed by loss of Pk and loss of Pk and Sple, but not by loss of Sple alone. I may be missing something important here, but these results seem to say that having Pk activity causes a disruption in polarity in the absence of Ds, but does nothing to support the view that Pk orients polarity?

– Having given it some thought, I think the key to understanding these bristle phenotypes is the original observation of Gubb and Garcia-Bellido, 1982 that "Trichomes and chaetae express the same polarity" i.e. the bristles do what the adjacent trichomes do. In ds we know trichomes reverse in the mid-anterior wing margin (Adler et al., 1998) and this seems to be due to altered core PCP in turn produced by disrupted cell flows, oriented cell divisions etc. (Aigouy et al., Merkel et al.) [other explanations for the swirling phenotype are possible but don't alter the argument I think]. In ds double mutants with the core pathway, trichomes show a fz/in pattern (Adler et al., 1998) and point broadly distal on the anterior wing margin, and we see bristles doing the same in ds pk-sple i.e. the phenotype of core loss-of-function is epistatic to that of ft-ds for trichome polarity. In ds pkpk we also see trichomes pointing distally as do the bristles (e.g. Hogan et al., 2011), presumably because Sple isn't being mislocalized by Ds (as also suggested in the text by the authors here) and we're seeing either a wild-type or fz/in loss-of-function polarity phenotype?

Comments related to trichome polarity:

1) The authors also revisit the issue of whether Ft-Ds instruct trichome polarity in Figure 7G-K. However, these results just seem to agree with the view that the core loss-of-function trichome polarity phenotypes are epistatic to loss- or gain-of-function ft-ds trichome polarity phenotypes. There is no dispute that dll>2x-ds makes trichome polarity swirl (and as they normally point distally, any swirling can be interpreted as "reversal"), but I think it is hard to prove or disprove whether this is direct effect on the core pathway or an indirect effect on the complex process of wing morphogenesis (c.f. work from Eaton lab) that alters core polarity. I don't think it's a surprise that loss of fz or pksple suppresses this phenotype (epistasis noted by Adler et al., 1998) or that pksple does not (this leaves core pathway intact).

The interesting question is why the dll>2x-ds; pkpk phenotype looks like the pkpk phenotype. The authors describe this as a "suppression", and seem to conclude that this means PkPk couples to Ft-Ds, but I don't follow the leap of logic. I think the accepted view is that in pkpk the residual population of PkSple is oriented by endogenous Ft-Ds generating the familiar pkpk AP oriented trichome pattern, and it appears that this AP orientation persists in a dll>2x-ds background. This persistence is interesting, but seems to be more about how and when PkSple couples to Ft-Ds, and doesn't seem to reveal anything about whether PkPk can couple to Ft-Ds. Have I missed something?

2) Overall I feel it's a pretty shaky case that Ft-Ds is directly orienting the core in these experiments or that Pk is acting to couple the core to Ft-Ds. Of course if you did want to argue that Ft-Ds orient the core in the relevant region of the wing blade to affect either distal trichome polarity or the triple row bristles, I think you would have to carefully reconcile the following published observations:

i) Ft-Ds polarity seems to be AP while core is PD, so they don't point the same way? (A timing issue maybe?)

ii) flattening Ds and Fj gradients seems to give normal polarity? (Also of bristles?)

iii) ft dachs and ft UAS-ft-intra and ft UAS-wts give normal trichome polarity? (Triple row bristles also point distally in these genotypes?)

However, this all seems like a different argument, not pertinent to helical morphogenesis.

3) Nevertheless, in light of the above, I worry about the statement in the Abstract "Pkpk and Pksple, determines right- or left-handed bristle morphogenesis, each by coupling Frizzled/Vang signaling to the Fat/Dachsous PCP directional signal in opposite directions". On one hand I don't follow the argument showing PkPk is coupling to Ft-Ds or that Ft-Ds is determining bristle polarity. The fact that bristle polarity is disrupted in ds mutants doesn't necessarily imply Ds is instructive. I agree that in the absence of PkPk, PkSple is probably reversing bristle polarity by coupling to Ds – but this doesn't happen in normal development.

https://doi.org/10.7554/eLife.51456.sa1

Author response

Reviewer 1:

1) Figure 4H-I. Mosaic knockdown of pk by RNAi is used to distinguish which side of the cell junction Pk and Sple isoforms reside. The authors conclude that both isoforms are proximal. This is an important result as it is different from previously published studies overexpressing Sple and should be shown more clearly. It is very difficult to distinguish the boundaries between expressing and non-expressing cells, especially considering that margin cells may contribute to the signal.

We recognize that these images require some effort to interpret. To assist readers, we have drawn outlines just inside the cell boundaries of the relevant cells in the initial panels. We prefer to not also place those outlines in the second and third panels so as to not obscure the data. We think that a motivated reader will be able to make their own interpretation and that it will agree with our description.

2) Figure 4—figure supplement 1A-B. The loading control levels are highly variable making it difficult to interpret how Pk and Sple isoform levels change over time. Perhaps tubulin is not an appropriate loading control if it levels too change over time. Can the authors plot Pk and Sple levels relative to Tub loading control? Or use V5 fluorescent intensities from confocal images to compare levels?

We have provided the requested quantification of Pk and Sple relative to γ-tubulin and presented the results as a bar chart.

3) Figure 4. The authors conclude from the localization of Pk and Sple isoforms "that distal localization of Pksple in Pkpk mutants drives counter-clockwise rotation of shaft-socket pairs." In addition, the title of this section is "Localization of Pkpk and/or Pksple determines handedness of helical rotation". I think these are overstated conclusions. The localization certainly correlates with the direction of rotation, but this seems insufficient to conclude this drives rotation, especially since elimination of both isoforms has very little effect on rotation.

We agree, and the wording has been changed to reflect correlation. In the conclusion, we then “suggest” that because Pksple determines counter-clockwise rotation, its localization is likely the determinant of that rotation. As discussed more extensively below, we have made revisions to the text to more clearly emphasize that Pkpk does not actively determine the direction of rotation, but instead allows the core module to interpret the direction of rotation.

4) Figure 7D. This experiment shows that reversed Sple localization in Pkpk mutants depends on Ds, and "we interpret this to indicate that the Ds global signal recruits Pksple to sites of enriched Ds (distal) in the absence of Pkpk". What about in the proximal part of the wing where bristle polarity is normal? Is Sple localized proximally, and independent of Ds? How do the authors interpret the distally restricted phenotype of Pkpk mutants? Is the Ds signal stronger distally?

The localization of Pksple is discussed in the manuscript (paragraph four subsection “Localization of Pkpk and/or Pksple correlates with handedness of helical rotation”; Figure 4—figure supplement 2 C-C’’’; and in the Discussion, paragraph four of subsection “Spatiotemporal dynamics and selection of Pkpk versus Pksple 514 for polarity determination”). Our analyses do not have sufficient resolution to fully understand the differences between the distal wing where polarity is reversed and the proximal wing, where it is not, but we refer the reviewer to these passages for some possible explanations.

5) Coupling between the Fz-Vang and Fat-Ds pathways by Pk isoforms. "we interpret this to indicate that the Ds global signal recruits Pksple to sites of enriched Ds (distal) in the absence of Pkpk (Figure 7B; compare with 7D), which drives reversal of shaft-socket orientation and therefore reversal of bristle polarity." Are Fz and Vang localizations flipped in socket cells with reversed polarity in Pkpk mutants? Although this has been shown previously for wing hairs overexpressing Sple, I feel it is important to show here under endogenous Sple levels to definitely show the core PCP module is flipped, not just a downstream component.

We have now done this analysis for Vang, and indeed the core module is flipped. The data are provided in Figure 5—figure supplement 1.

Reviewer 2:

1) A "helical structure of defined chirality" may not be appropriate. It seems to be an anisotropic cell rearrangement directed by distal-proximal and dorsal-ventral polarity. Such rearrangement should change the interface between shaft and socket as shown in Figure 3B and C.

Here, we presume that in saying “anisotropic cell rearrangement,” the reviewer refers to junctional rearrangement. If instead they intend it more generically then it is just a vaguer statement of generating a structure of defined chirality. With that in mind, we think that both generating a structure of helical chirality and anisotropic cell rearrangements are both valid observational descriptions of what we see, each conveying different aspects of the event. Adding that these are “directed by distal-proximal and dorsal-ventral polarity” is an interpretation. Regarding the description, we prefer emphasizing the structure, as that is better documented that the junctional rearrangement, for which we lack live imaging that would be needed to solidify that inference.

2) It is difficult to evaluate the phenotypic analyses without statistical information. What are the percentages of wings showing the representative phenotypes? How may socket cells were scored from how many wings? How many clones were analyzed and what was the percentages showing the described phenotypes? Some of them can be found in source data files. However, it should be also stated in the text.

We have provided additional statistical information in the legends for main Figure 5, 6, and 7, in the legends for Figure 2, 3, 6, and 7 figure supplements, and in several additional source data files. In addition, we have added several comments about consistency of qualitative results where appropriate. In some cases, we have not addressed “What are the percentages of wings showing the representative phenotypes?” In all cases the phenotypes we show are very consistent unless otherwise indicated (ie dsRNAi). In the Materials and methods section, we have added the sentence “For qualitative results such as expression patterns, a minimum of 20 biological replicates from at least two independent experiments were examined and representative images are shown.” We would hope that it is not necessary to state that the qualitative results we describe are consistent and that we have chosen representative images.

3) Figure 4H-I', Figure 2—figure supplement 1B, and Figure 4—figure supplement 2 were unclear. These data should be shown more clearly.

To aid in interpreting Figure 4H-I’, we have added some outlines of relevant cells that we hope will be helpful. For Figure 2—figure supplement 1B, we have added some additional colored dots and explanation. We reiterate that the junctional rearrangements are presumed, since we do not have live imaging. If the reviewer is concerned that the entirety of Figure 4—figure supplement 2 is unclear, we’re at a loss as to how to respond. In every instance we have attempted to select representative images that show as clearly as possible the feature of interest.

Reviewer 3:

1) As a general comment, it might be helpful to readers to explain that in processes such as left-right asymmetry determination (analogous to the chiral morphogenesis being discussed here) there is symmetry breaking on three axes: apical-basal, and two axes in the plane of the tissue. Hence, I presume PCP specifies clockwise or anticlockwise morphogenesis by specifying polarity in one axis of the plane, in concert with a second polarity cue in the orthogonal axis in the plane (which I take to be an AP oriented cue, ultimately set up by the Hedgehog pathway?). Without this second cue, the PCP cue alone wouldn't be sufficient to specify chirality?

We agree, and a statement articulating this idea has been added.

2) Introduction paragraph three: I think consistent nomenclature for proteins in flies is capitalized Roman so should be "Prickle" and "Spiny-legs" (and probably "PkPk" and "PkSple")? (Also some genotypes change to all caps italics on figure panels e.g. compare Figure 5E-F to 5G-H).

Since these are isoforms of one protein, we think the correct designation is “Prickleprickle (Pkpk) and Pricklespiny-legs (Pksple)”.

3) Results paragraph one: typo "Supressor"

Thank you.

4) Figure 4—figure supplement 1A is mislabeled? The blue asterisks ("unidentified") and the purple arrow ("Pk") should be swapped? This would then be consistent with panel B?

The arrows are correct, although the confusion may have arisen from incorrect designation of molecular weight markers (now corrected) in panel B. The unidentified band co-migrates with V5::Pk and occurs in variable intensities.

5) Subsection “Pkpk and Pksple differentially interpret directional signals from Ds” imply an "indirect" mechanism for PkPk coupling to Ft-Ds, but the Introduction (I think more in line with previous work, including from the authors) suggest an indirect mechanism for core pathway coupling to Ft-Ds via Fz/Dsh/Fmi transport with (presumably) PkPk being required for magnitude of core pathway activity but *not* for orientation? This is a complex area, but I think it would be easier for readers if the authors stuck to one view. At least based on the data so far, we don't see any evidence that PkPk couples, only that it enhances core pathway activity?

Both sections are intended to refer to the same model, in which Ft/Ds couples to the core pathway indirectly via microtubule directed transport when Pkpk is participating. We have adjusted the language to better capture this idea throughout the manuscript so as not to imply more than what we mean to say. Specifically, we do not wish to suggest an active role for Pkpk in coupling, but rather the microtubule mechanism couples when Pkpk is used for core signaling (and evidently to some extent in the absence of either Pk isoform).

6) Paragraph three of subsection “Pkpk and Pksple differentially interpret directional signals from Ds”: the authors deduce that in the absence of PkPk, PkSple is recruited to distal cell edges by distally localised Ds, but I find this confusing (at least without further clarification) as Merkel et al. seemed to show pretty clearly that – at least in the wing blade – Ds is oriented primarily on the AP axis from 16h APF onwards, and trichomes orient primarily on the AP axis in pkpk as well (e.g. Merkel et al. and Gubb et al.). So I would expect Sple to be recruited largely to AP cell edges (maybe with a slight distal bias?) not to distal cell edges. But maybe something happens on the edge of the wing blade to tilt the Ds polarity axis more PD? What do the authors think?

The Merkel analysis of Ds orientation didn’t extend to the wing margin, so it is not directly pertinent. We furthermore suggest that the Merkel et al. analysis doesn’t have sufficient resolution or precision to make this assessment. And finally, the reviewer is correct that only a subtle PD bias would be sufficient. Of note, in the region where bristles are reversed in pkpk mutants, the hairs nearest to the margin are similarly reversed, rather than being primarily AP as the reviewer suggests, and indeed notes themself in the following point 7. Please see Author response image 1.

7) Paragraph four subsection “Pkpk and Pksple differentially interpret directional signals from Ds”: this appears to be a key sentence, as here the authors appear to argue that ds knockdown affects a PkPk-mediated signal for bristle polarity, however I completely failed to follow the logic that indicates PkPk is providing a polarity cue rather than contributing to polarity magnitude (see paragraph two of subsection “Pkpk versus Pksple isoform expression determines the direction of rotation” and Table 3) and suppressing effects of "ectopic" Sple.

– The key piece of data is (I think) that ds- gives reversals of bristle polarity, and (perhaps surprisingly) these are suppressed by loss of Pk and loss of Pk and Sple, but not by loss of Sple alone. I may be missing something important here, but these results seem to say that having Pk activity causes a disruption in polarity in the absence of Ds, but does nothing to support the view that Pk orients polarity?

The reviewer is overinterpreting our statement in the cited lines. Here, we’re only trying to explain the random local domains of P and D polarity in ds knockdown wings. These depend on Pkpk, though we cannot see the Pkpk localization for technical reasons. We’re not saying Pkpk orients polarity; we’re simply arguing that it allows the randomly oriented, locally correlated domains to form. Language clarifying this claim has been added to the manuscript.

– Having given it some thought, I think the key to understanding these bristle phenotypes is the original observation of Gubb and Garcia-Bellido, 1982 that "Trichomes and chaetae express the same polarity" i.e. the bristles do what the adjacent trichomes do. In ds we know trichomes reverse in the mid-anterior wing margin (Adler et al., 1998) and this seems to be due to altered core PCP in turn produced by disrupted cell flows, oriented cell divisions etc (Aigouy et al., Merkel et al.) [other explanations for the swirling phenotype are possible but don't alter the argument I think]. In ds double mutants with the core pathway, trichomes show a fz/in pattern (Adler et al., 1998) and point broadly distal on the anterior wing margin, and we see bristles doing the same in ds pkpk-sple i.e. the phenotype of core loss-of-function is epistatic to that of ft-ds for trichome polarity. In ds pkpk we also see trichomes pointing distally as do the bristles (e.g. Hogan et al., 2011), presumably because Sple isn't being mislocalized by Ds (as also suggested in the text by the authors here) and we're seeing either a wild-type or fz/in loss-of-function polarity phenotype?

We agree that bristles and hairs are under the same control at the AWM, and that polarity is transmitted between them, as discussed above. But just saying that bristles and hairs share the same polarity is hardly an explanation for the direction we see in various mutants, since it doesn’t explain why either the hairs or bristles behave as noted. As an aside, we’re unconvinced that cell flows are the answer, but that is another discussion. The essence of our finding is that in this region, Pksple expression comes up earlier than the rest of the wing, accounting for the cleaner reversal when Pkpk is removed.

Comments related to trichome polarity:

1) The authors also revisit the issue of whether Ft-Ds instruct trichome polarity in Figure 7G-K. However, these results just seem to agree with the view that the core loss-of-function trichome polarity phenotypes are epistatic to loss- or gain-of-function ft-ds trichome polarity phenotypes. There is no dispute that dll>2x-ds makes trichome polarity swirl (and as they normally point distally, any swirling can be interpreted as "reversal"), but I think it is hard to prove or disprove whether this is direct effect on the core pathway or an indirect effect on the complex process of wing morphogenesis (c.f. work from Eaton lab) that alters core polarity. I don't think it's a surprise that loss of fz or pk-sple suppresses this phenotype (epistasis noted by Adler et al., 1998) or that pksple does not (this leaves core pathway intact).

We agree entirely that our data are consistent with epistasis between Ft-Ds and core PCP signaling. Epistasis, done properly, implies an upstream-downstream relationship which we indeed believe to be the case between these two systems when pkpk is participating in core function.

However, we wish to go one step further, and address whether Ft-Ds is instructive during Pkpk function, and loss-of-function epistasis is not sufficient to make this assessment. Demonstrating an instructive role requires reorganizing the Ft-Ds signal in a defined way and asking if the core response follows the predicted input. We therefore identified a condition, dll>2x-ds, in which a reorganized Ds gradient redirects polarity. The reviewer correctly notes that there are some swirls in this pattern, and that by definition if there are some swirls there must be areas if reversed polarity. But one way to create swirls is to induce areas of reversed polarity, and we argue that dll>2x-ds does exactly that. To try to make the point more convincingly, we now provide an image showing a much larger portion of such a wing, and highlight the region where the reversed gradient is most pronounced (new Figure 7—figure supplement 1E). Note that in the majority of this region, hairs are reoriented to point away from the margin, as expected based on the shape of the dll expression domain, whereas swirls appear primarily at the borders of this domain where the reversed polarity confronts regions that are not reversed because ds is either expressed at too low a level or too high a level to produce a meaningful gradient. In addition, we’ve moved the location of the close-up images in Figure 7G-K away from the L3-L4 intervein space where we sometimes saw distortions of the reversal, as in this example, perhaps due to the veins, or possibly due to the geometry of the dll gradient, to the L4-L5 intervein space where the effect is more cleanly seen.

The reviewer also focuses on “whether this is direct effect… or an indirect effect…” We hope the reviewer will agree that the questions of epistasis and of instructive function imply nothing at all about how direct or indirect the signal is between the two systems; how direct or indirect is instead a quite distinct question. If, as the reviewer suggests, Ft-Ds acts on core PCP by modulating wing morphogenesis (c.f. work from Eaton lab), this would simply provide an explanation for the epistasis and instructive function. In this manuscript, we make no specific argument about the mechanism – we simply argue that the Ft-Ds system instructs directionality of the core system when Pkpk is participating. However, for this reviewer, we note that we do not favor an Eaton mechanism, such as perturbation of cell flow and shear, because these physical processes would be expected to alter the mutant polarity patterns as well as the wildtype pattern, yet this is not observed. The fzfz, pkpk-sple and pkpk patterns are essentially unperturbed by dll>2x-ds.

The interesting question is why the dll>2x-ds; pkpk phenotype looks like the pkpk phenotype. The authors describe this as a "suppression", and seem to conclude that this means PkPk couples to Ft-Ds, but I don't follow the leap of logic. I think the accepted view is that in pkpk the residual population of PkSple is oriented by endogenous Ft-Ds generating the familiar pkpk AP oriented trichome pattern, and it appears that this AP orientation persists in a dll>2x-ds background. This persistence is interesting, but seems to be more about how and when PkSple couples to Ft-Ds, and doesn't seem to reveal anything about whether PkPk can couple to Ft-Ds. Have I missed something?

As discussed above, for the purposes of our argument, we suggest the important observation is that dll>2x-ds does redirect the polarity pattern when Pkpk is present – if the systems are uncoupled, as has been argued by others, this should not occur.

There is, however, an interesting point to be made regarding the question the reviewer brings up: why the dll>2x-ds; pkpk phenotype looks like the pkpk phenotype. Indeed, it has been argued that in the absence of Pkpk, Pksple contributes to the polarity pattern, and this can be seen in the difference between the polarity patterns of pkpk and pkpk-sple mutants. We see relatively subtle differences between the pkpk pattern and the dll>2x-ds; pkpk pattern, consistent with our data that there is little Pksple expression when hair polarity is established. One exception is near the anterior wing margin, where Pksple expression does become significant at this time. In the pkpk pattern, hairs (and some bristles) point proximally, and in the dll>2x-ds; pkpk pattern, they mostly point distally (as would be predicted with reversed Ds gradient and Pksple control of core response).

Author response image 1
pkpkand dll>ds;pkpkanterior wing margins.

2) Overall I feel it's a pretty shaky case that Ft-Ds is directly orienting the core in these experiments or that Pk is acting to couple the core to Ft-Ds. Of course if you did want to argue that Ft-Ds orient the core in the relevant region of the wing blade to affect either distal trichome polarity or the triple row bristles, I think you would have to carefully reconcile the following published observations:

i) Ft-Ds polarity seems to be AP while core is PD, so they don't point the same way? (A timing issue maybe?)

ii) flattening Ds and Fj gradients seems to give normal polarity? (Also of bristles?)

iii) ft dachs and ft UAS-ft-intra and ft UAS-wts give normal trichome polarity? (Triple row bristles also point distally in these genotypes?)

However, this all seems like a different argument, not pertinent to helical morphogenesis.

Again, we reiterate that we are not claiming that Ft-Ds is “directly orienting the core.” We argue that Ft-Ds orients the core system while under Pkpk control but make no statement about how direct or indirect that signal is. Nor do we argue that “Pk is acting to couple the core to Ft-Ds.” Indeed, we have previously proposed a rather indirect model in which Ft-Ds orient polarized microtubules on which vesicles containing core proteins traffic. In this model, coupling is indirect and is not mediated by Pkpk.

Nonetheless, we will address the points raised by the reviewer.

(i) Orientation of the Ft-Ds system has not been examined at the anterior wing margin. Published data only measure to a distance of several cells away from the margin. Still, looking at the Eaton data (Merkel et al.), both the Ft-Ds and core systems evolve from a radial pattern, but the core system becomes parallel while the Ft-Ds system does not. As we suggested originally in Ma et al., 2003, the tendency of the core system toward local alignment is stronger than the coupling to the Ft-Ds system. Thus, the early Ft-Ds orientation and perhaps a subtle remaining P-D orientation at later times could be sufficient to produce the effects we see.

(ii-iii) While these genotype-phenotype relationships are interesting, we believe they address a different question, specifically, whether there is some other information that is revealed when Ft-Ds is partially or wholly disabled. There is little doubt that perturbing the Ft-Ds signal in a variety of ways and in several tissues impinges on core PCP directionality. The listed genotypes attempt to completely remove this information. If that is successful (we’d argue that’s not entirely clear), what remains is indicative of what other information might orient core PCP, but does not bear on how Ft-Ds, when active, interacts with the core PCP system. Yes, we’d like to better understand these observations, but that is not necessary to interpret the experiments in this manuscript.

3) Nevertheless, in light of the above, I worry about the statement in the Abstract "Pkpk and Pksple, determines right- or left-handed bristle morphogenesis, each by coupling Frizzled/Vang signaling to the Fat/Dachsous PCP directional signal in opposite directions". On one hand I don't follow the argument showing PkPk is coupling to Ft-Ds or that Ft-Ds is determining bristle polarity. The fact that bristle polarity is disrupted in ds mutants doesn't necessarily imply Ds is instructive. I agree that in the absence of PkPk, PkSple is probably reversing bristle polarity by coupling to Ds – but this doesn't happen in normal development.

We agree the wording in the Abstract was problematic, although for a somewhat different reason, and it has been changed. We do not aim to imply an active role for Pkpk in coupling, but rather that Ft-Ds couples to the core PCP system when the core system utilizes Pkpk. It is also correct that “The fact that bristle polarity is disrupted in ds mutants doesn't necessarily imply Ds is instructive.” As discussed above, the reorienting response of the core PCP system, and of cellular polarity, upon introduction of an alternately produced Ds gradient is the basis for claiming Ft-Ds is instructive. We hope that this discussion, and the improved language in the revised manuscript, is persuasive.

https://doi.org/10.7554/eLife.51456.sa2

Article and author information

Author details

  1. Bomsoo Cho

    Department of Pathology, Stanford University School of Medicine, Stanford, United States
    Contribution
    Conceptualization, Investigation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8970-9160
  2. Song Song

    Department of Pathology, Stanford University School of Medicine, Stanford, United States
    Contribution
    Conceptualization, Data curation, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4587-8592
  3. Jeffrey D Axelrod

    Department of Pathology, Stanford University School of Medicine, Stanford, United States
    Contribution
    Conceptualization, Supervision, Funding acquisition, Project administration
    For correspondence
    jaxelrod@stanford.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6094-7392

Funding

National Institute of General Medical Sciences (R01 GM097081)

  • Jeffrey D Axelrod

National Institute of General Medical Sciences (R37 GM059823)

  • Jeffrey D Axelrod

National Institute of General Medical Sciences (R35 GM131914)

  • Jeffrey D Axelrod

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

Acknowledgements

We thank Helen McNeill, Lucy O’Brien and members of the Axelrod lab for constructive comments.

Ethics

Animal experimentation: This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocol (#11840) of Stanford University.

Senior Editor

  1. Utpal Banerjee, University of California, Los Angeles, United States

Reviewing Editor

  1. Danelle Devenport, Princeton University, United States

Reviewer

  1. Danelle Devenport, Princeton University, United States

Publication history

  1. Received: August 29, 2019
  2. Accepted: January 10, 2020
  3. Accepted Manuscript published: January 14, 2020 (version 1)
  4. Version of Record published: February 6, 2020 (version 2)

Copyright

© 2020, Cho 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.

Metrics

  • 712
    Page views
  • 246
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Developmental Biology
    Moran Cohen-Berkman et al.
    Research Article
    1. Cell Biology
    2. Neuroscience
    Eduardo Javier López Soto, Diane Lipscombe
    Research Article