Two conserved regions within OR463 are fundamental for wing development.

A. Illustration of ap upstream intergenic region. In blue: enhancer sequences relevant during wing development (Bieli, Kanca, Requena, et al., 2015). In black: OR463 sequence conservation among related insects. B. Different regions in which OR463 was divided for mutation analysis. In dark grey, sequences with highly conserved TF binding sites. C. Percentage of wild-type wings when each of the fragments was deleted in homozygosis and (D) hemizygosis (over the apDG3 allele). E. Control wing (Re-integrated WT OR463 in apR2 landing site). F. Loss of wing displayed by ΔOR463 mutants. G. Loss of wing displayed by Δm3 mutants. H. Representative wing phenotype derived from Δm1 deletion in homozygosis. Anterior compartment presents correct venation, while P compartment presents an outgrowth. Venation pattern in this outgrowth is disturbed. Inset 1: Detail of the A compartment bristles. Inset 2: bristles of A identity in the P compartment. I. Representative wing phenotypes derived from Δm1 deletion in hemizygosis. P compartment venation pattern is disturbed. No outgrowth is formed but the P compartment present severe venation problems, with some veins position perpendicularly to its normal direction. Inset 1: Detail of A bristles. Inset 2: Detail of A bristles in the P compartment. Note that transformation of the P margin into A is not complete. I’. Wing phenotype also present in Δm1 hemizygous flies. In these, the P compartment is severely reduced and mostly A bristles are present in the margin.

Mirror image duplications arise due to changes in DV and AP boundary position.

A. Anti-Ap and anti-Wg immunostaining of control wing discs (Re-integrated WT OR463). A’. Anti-Ap and anti-Wg staining of wing discs in homozygous Δm1 mutants. The DV boundary is distorted in the P compartment, where it is extended into the presumptive wing hinge (arrowheads). A’’. Anti-Ap staining of Δm1 hemizygous wing discs. The DV boundary is further deformed in the P compartment (arrowhead). B. Quantification of the relative P size (P Area/Total Area) and relative PD size (PD Area/Total Area) in different mutants (p value<0.0005, control: n=13, N1/DG3: n=16, Δm1/DG3: n=10, Δm1m2m4/DG3: n=9). C, C’ and C’’. Relative position of AP and DV boundaries as revealed by immunostaining with anti-Ap and anti-Ptc in wing discs of control, Δm1/DG3 and 1′N1/DG3. Different quadrant maps are then subtracted and the AP and DV boundaries represented with dashed lines (red and blue, respectively). Asterisk depict the intersections of AP and DV boundaries. Scale bars: 100μm. D. Schematic of wild-type wing development from larval to pupal stages. In the larval disc, the AP boundary (green) intersects the DV boundary (purple) at a single point (asterisk). During pupal eversion, the wing pouch unfolds along the DV boundary, which becomes the future wing margin, and veins are patterned in parallel to the AP boundary. The arrow marks the region where AP and DV boundaries overlap within the posterior compartment. D’. Model in which PD growth is strongly reduced or nearly absent. After eversion, the anterior compartment will present anterior identity on its dorsal side and posterior identity on its ventral side. This mixed orientation accounts for the partial A identity observed at the posterior border. D’’. Model in which PD size is reduced but still occupies a considerable pouch area. Here, the AP and DV boundaries meet at two distinct positions within the pouch, spaced far enough apart to support the specification of two wing organizing centers. After eversion, the secondary center generates a wing structure with dorsal A identity and ventral P identity, producing characteristic venation defects patterned along the ectopic DV boundary.

Spatiotemporal characterization of OR463 via localized dCas9 expression.

A. Schematic representation of the method. Upon localized expression, dCas9 would bind to DNA displacing or interfering with the binding of TF. B. Target sites of the gRNAs present in the U6-OR463.gRNAx4 transgene. C. Wings of animals expressing dCas9 under the en-Gal4 driver in the presence of the control gRNA at 23° and 25° degree. D. Wings of animals expressing dCas9 under the en-Gal4 driver in the presence of U6-OR463.gRNAx4 at 23° and 25° degree. Arrowhead indicates the P outgrowth. Notice the presence of A bristles in the P edge (dashed box). E. Control wing discs, expressing dCas9 under the en-Gal4 driver in the presence of control gRNAs at 23°. E’. Anti-Ap immunostaining showing the normal Ap expression pattern in late wing discs. E’’. Quadrant map subtracted from panel E. Notice the single perpendicular intersection between AP and DV boundaries. F. Anti-Ptc and Anti-Ap immunostaining of wing discs expressing dCas9 under the en-Gal4 driver in the presence of U6-OR463.gRNAx4 at 23°. Notice the P outgrowth (arrowhead). F’. Anti-Ap immunostaining exhibiting the extended posterior pattern in the posterior compartment. F’’. Quadrant map subtracted from panel F. Notice reduce size of PD quadrant. AP and DV contact twice, once perpendicularly as in control discs and once tangentially in the P outgrowth. G. Anti-Ptc and Anti-Ap immunostaining of wing discs expressing dCas9 under the ptc-Gal4 driver in the presence of U6-OR463.gRNAx4 at 23°. Notice lack of Ap signal along the central wing disc (asterisk). G’. Anti-Ap immunostaining of the disc in G. G’’. Quadrant map subtracted from panel F. Notice the lack of contact between AD and PD quadrants. Scale bars: 100μm. H. Schematic of the temperature-shift experiment used to define the developmental window during which the apE enhancer is sensitive to repression. en-Gal4, tub-Gal80ts, UAS-dCas9; U6-OR463.gRNA(4×) animals were kept at 18 °C to suppress dCas9 and shifted to 29°C at successive developmental intervals (0–168 h AEL) to induce dCas9 expression at distinct stages. H′. Timeline summarizing the temperature regime and the percentage of adult wings showing strong apterous-like phenotypes following induction at each time point. Early induction (0–48 h AEL) produced the highest penetrance, with progressively weaker effects after 72 h AEL.

Pnt and Hth are required for apE activity via m1.

A. Schematic representation of m1 and m4 predicted binding sites and the generated deletions within m1. B. Phenotypic penetrance of the different deletions within m1 in homozygosis, and hemizygosis (B’). C. Projected area of the different quadrants of control (WT/apDG3, Δm1/ apDG3, Δm1.2/ apDG3 and Δm1.3/ apDG3 wing discs, based on the immunostaining against Ap and Ptc in real wing discs. D. Effect on the Ap expression domain by the expression of hthRNAi and pntRNAi in the P domain using en-Gal. In both cases the PD domain is reduced. In green, the UAS-CD8:GFP reporter marks the P compartment. E. Adult wing phenotypes upon pntRNAi expression via en-Gal4. Left wing: Example of a wing outgrowth resembling Δm1 phenotype. Middle wing: Example of a partial A mirror-image transformation. Inset 1: campaniform sensillae of the A compartment. Inset 2: ectopic campaniform sensillae formed in the P compartment upon pntRNAi expression. Notice the presence of A bristles within the P compartment. Right wing: Example in which P compartment is reduced. F. Effect of the expression of hthRNAi and pntRNAi in the P domain on anti-Ap localization and apE-LacZ reporter (visualized with anti-ßGal) during L2 stage. In control discs, anti-Ap can be detected in the nucleus of P cells (marked with UAS-CD8:GFP) (arrowheads). anti-ßGal (in white) is detected in the dorsal compartment in both A and P compartments. Upon pntRNAi or hthRNAi, no anti-Ap nor anti-ßGal signal was detected in P cells (arrowheads). Scale bars: panels C and D, 100μm; panel E, 25μm

High resolution genetic analysis suggests a HOX-GATA complex important for m3 activity.

A. anti-Ap and anti-Wg immunostaing in control and Δm3 third instar wing discs. In contrast to control wing discs, Ap can only be detected in a small group of cells in the anterior hinge in Δm3 mutants (arrowhead). The territory within Wg inner ring is totally missing (asterisk). No Wg stripe is detected in the pouch. B. X-Gal staining of control apE-LacZ and apE1′m3-LacZ in third instar wing disc. apE1′m3-LacZ only showing minimal X-Gal staining in the P hinge. C. Summary of the base pair substitutions generated within m3.1. Each row corresponds to a different allele containing the changes labelled in red. D. Scoring of wild-type wings across the library of Δm3 mutants. In black, the percentage of WT wings. Between 80 and 250 wings were scored in each case. Asterisks denote mutants that gave rise to no wings with different penetrance. E-E’’’. Wing phenotypes of the control, and three of the mutants of the library. F. Phenotypic penetrance scoring of control animals and individuals in which the linker between GATA and HOX binding sites was extended or contracted. G. Percentage of flies with wings upon deletion of m3, or deletion of m3 and simultaneous contraction of m2 linker. Scale bars: 100μm

the GATA TF Grain is fundamental for wing development.

A. Phenotypes produced upon UAS-grnRNAi expression driven by en-Gal4 in the presence of UAS-dicer. Reduction of the P size and partial P to A transformation (evidenced by the campaniform sensilla in the P territory (arrowhead)). A’. Mirror image duplication of an anterior proximal rudimentary wing structure. A’’. Thorax defects observed in some of the adults. B. Anti-Ap and anti-Wg immunostaining of control wing discs and P knockdown of grn (grnRNAi driven by en-Gal4). In both cases, UAS-GFP was used to mark P cells and UAS-dicer was included to increase knockdown efficiency. B’. Example of wing disc in which the P compartment (arrowhead) was located close to the DV boundary. In these cases, the pouch was specified and grew to some extent, as revealed by the space within the inner ring of Wg (asterisk). B’’. Example of wing disc in which the remanent of the P compartment (arrowhead) is located close to the notum. Here, the notum primordium grew until a considerable size and presented the characteristic Wg band, indicating, to some extent, correct patterning. The pouch (asterisk) is completely absent. C. anti-Ap immunostaining upon tissue-specific knock-out of grn in the P cells using en-Gal4, UAS-Cas9 in the presence of U6-grn.gRNAs. UAS-CD8:GFP labelled posterior cells. Notice the total loss of P compartment in mid L3 wing discs (illustrated by the complete absence of GFP signal). D. L2 wing disc of the same genotype as in C. Immunostaining of apE-LacZ using anti-ßGal4, as well as anti-Ap reveals total lack of signal in the P cells. Scale bars: panels B and C, 100μm; panel D, 25μm

The Hox gene Antp is fundamental for early wing development and ap expression.

A. Scheme of the experimental setup to delete Antp during early stages of wing development. sna1.7-Gal4 driver is used to express Cas9 in the embryonic precursors of wing and haltere. Cas9 is targeted to the Antp locus by three gRNAs (labelled with an arrow). Position of the homeodomain within Antp sequence is also indicated. B. Anti-Antp and anti-Ap immunostaining in control third instar wing discs. Notice the distribution of Antp throughout the disc, with its highest levels in the A compartment and hinge, and the faint band along the DV boundary. C. Example of wing disc derived from sna1.7-Gal4 driven Cas9 in the presence of U6-grn.gRNA. In this case Ap presented a rather normal distribution in the tissue. Antp could be still detected in the wing discs, amid a reduction of its levels. C’. Wing disc of the same genotype as C presenting severe problems in the Ap expression pattern. Ap was detected in the pouch only in two groups of cells (marked with an arrowhead). C’’. Example of a wing disc of the same genotype as C and C’’ in which no immunoreactivity against Antp is detected. Note the complete lack of the anti-Ap signal. In grey, DAPI. These discs lack all recognizable structures (no notum, and no pouch). D. Adult phenotype arising from the same genotype as B (control). The haltere is marked by an asterisk. E. Example of a severe case in which Antp was knocked out from the wing primordia (same genotype as in C, C’ and C’’). In this case, only a small portion of the right notum in still present, with the total loss of the right wing (arrowhead). Left wing presents severe morphological defects and forms a balloon-like structure. Further analysis of this wing revealed the presence of campaniform sensillae and A bristles in the P compartment (data not shown). Notice that the halteres were unaffected (asterisk). Scale bars: 100μm

Working model for OR463 regulatory inputs.

During late embryonic or early larval stages, the HOX input in m3, possibly Antp, would be responsible of OR463 activation. During this early phase the enhancer is not yet functional and the HOX could be priming the enhancer, permitting the later action of the other factors. Grain would participate in this process but its requirement is less critical than that of the HOX. Pnt and Hth would then, during L2 larval stage, activate apE first in the proximal area of the wing disc. Grn could also play a role in this early activity, confining the activity of ap to the dorsal hinge.

C2 and C5 activity throughout larval development.

A. illustration of ap locus. ap coding region is highlighted in dark grey, C2 in red and C5 in green. Within them, apE, OR463 and apDV are labelled in black. B-H. Wing discs of C2-mKate2 (red) and C5-eGFP (green) animals dissected at the indicated stages and immunostained for anti-Wg (blue). In early stages, only C2 is active in the central part of the tissue (B). C2 activity peaks in mid L2 (C) to later decrease and eventually disappear (G-H). C5 activity starts in few cells at the DV boundary (C) and later expands to all the dorsal pouch and part of the dorsal hinge (D-H). B’-H’. anti-Wg signal of the panels B-H. Scale bars: B-F: 20μm G-H: 50μm.

Identification of the minimal apE enhancer within apC2.

A. Eight DNA fragments from the apC2 region are shown. Their names are given on the left of the panel and their length in bp is indicated. These fragments were tested with two in vivo assays. (1) The fragments were introduced into reentry plasmid DB59 where they are combined with apC5 (contains apDV). These plasmids were brought back into landing site apattPΔEnh (Bieli et al 2015). Stocks were established and their rescue potential was scored in hemizygous flies (over apDG3). (2) The fragments were cloned into a LacZ reporter plasmid and inserted into the zh-86Fb landing site. Their LacZ activity was assayed by standard procedures. B. The rescue activity of each allele is shown. +: normal wings are formed. Nd = not done. C. The LacZ activity of each reporter construct is shown. +: normal apterous expression pattern in the wing imaginal disc. -: no LacZ activity. nd = not done. Based on these experiments, fragment OR463 was chosen as the minimal apE fragment.

Generation and validation of the OR463 landing site.

A. Schematic representation of CRISPR-mediated targeting of OR463. Embryos were injected with plasmids encoding for two gRNAs (red) laying within the OR463 fragment (blue box) to direct Cas9 to the locus. A donor plasmid was co-injected to serve as template during HDR. The donor plasmid contained: 1) the attP site (dark blue), 2) Dad13 minimal enhancer (517bp in green), flanked by FRT sites (red) in identical orientation, 3) homology arms (each 536bp long) to trigger integration via HDR, 4) flanking gRNA targets to linearize the donor plasmid inside the embryos. A’. Dominant wing phenotype of animals in which dad13 was integrated in ap locus. These animals presented variable phenotypes despite all having the same insertion. Phenotypes ranged from almost wild-type wings with small border defects in the distal tip (Lower wing, with detail of the tip) to a partial phenocopy of the classical Xasta indentations (upper wing). B. Dad13 was then excised by the action of Flp. B’. Homozygous wing phenotype of flies in which OR463 has been substituted by the landing site. C. Scheme of the re-integration of wild-type OR463 into the landing site. Re-integration constructs contained: 1) an attB site, 2) the OR463 sequence, 3) an FRT oriented as that in the landing site, 4) a mini-yellow marker. After insertion via φ C31 integrase, y+ transformants were used to establish stocks. In a second step, the y+ marker was deleted by Flp tratment. C’. Hemizygous wing phenotype of flies containing the reintegrated WT OR463 sequence before removal of the mini-yellow marker. These wings often presented aberrant phenotype, which included mirror-image A wing duplications. D. Final arrangement of the locus after Flp-mediated excision of mini-yellow. D’. Wing phenotype when wild-type OR463 sequence was re-inserted in the OR463 landing site after mini-yellow excision (in hemizygosis).

Ap localization in various genotypes.

Anti-Ap immunostaining in different OR463 mutants. Notice the deformation of the DV boundary in the P compartment in ΔN1/ apDG3, Δm1 and Δm1/apDG3. In Δm1m2m4 (note that m3 is still present in this genotype), the Ap expression pattern is severely reduced, apparently missing from the P compartment. Scalebars: 100μm

Loss of en expression cannot explain the mirror-image duplications of apE mutants.

A. Anti-En staining in wild-type and ΔapE mutants in hemizigosis (ΔapE/ apDG3). While the wing disc size of ΔapE mutants is affected, En localization is still constrained to the P side and there is not dramatic change in its pattern. B. Control and C. Δm1 hemizygous wing discs in which the current expression of en (Red) and its history of expression (Green) are overlayed. No big differences are found between the two signals apart from small clones of current expression present in both control and mutant. Δm1 hemizygous wing discs present a severe deformation of AP boundary and the P size is severely affected. Scalebars: 100μm

Vestigial expression in OR463 mutants.

In magenta, anti-Wg immunostaining, in green anti-Vg. A. Control (R2-WT/apDG3) L3 wing disc. Wg is detected in its stereotypic stripe along the wing pouch and is also detected in the hinge and notum. Vg is detected in the wing pouch and in the laterals of the proximal hinge.B. Hemizygous Δm1 mutant (R2-Δm1/apDG3) L3 wing disc. Wg stripe is extended in the anterior compartment (arrow) coinciding with the outgrowth. Vg localizes throughout the outgrowth in a pattern highly reminiscent of the pouch staining, with its highest levels in along the DV boundary. C. Hemizygous Δm3 mutant (R2-Δm3/apDG3) L3 wing disc. Wg stripe is missing and only the outer Wg ring of the hinge is detected, the inner ring being reduce to a group of cells. Vg is not detected at levels comparable to the WT discs.

Model for the development of mirror image duplications in ap OR463 mutants.

A. Model of Wild type wing development across larval, pupal and adult stages. The AP boundary (green) intersect with the DV boundary (purple) in only one spot (asterisk) during larval stages. In pupal stages, the wing pouch everts along the DV boundary, that become the wing margin. As wing veins are patterned by the AP border, their location is relatively parallel to it. A’. Example of wild-type wing disc and wing extracted from figure 1 and 2. The morphology of the posterior border is highlighted. It consists of a single row of fine bristles and looks distinctively different from the double-row bristles near of the tip of the. Wing or the triple-row bristles at the anterior margin. B. Wing disc model in which the PD size is affected but still occupies part of the pouch. In these cases, the AP border intersect with the DV boundary in two points which are far enough to specify two independent pouches. The “secondary” pouch does not present a real intersection, but both AP and DV boundaries coincide for a stretch after meeting. During pupal stages, after wing disc eversion, the outgrowth will present a dorsal side with A identity and a ventral side with P identity, resulting in venation defects. These veins, moreover, will be patterned parallel to the margin, as Hh and Dpp will be expressed along the DV boundary. B’. Example of a Δm1 wing disc and wing extracted from figure 1 and 2 that might be following this model. Note the presence of tip-margin bristles in the posterior outgrowth C. Wing disc in which PD is almost missing. In these cases, the whole P compartment will behave as does the outgrowth in panel B. C’. Example of a Δm1/apDG3 wing disc and wing extracted from figure 1 and 2 that might be following this case. In this case, A-specific dorsal triple-row bristles can be detected at the posterior margin. D. Model of wing development in which the DV is totally missing. In these cases, DV and AP boundaries are no longer crossing each other. If in close proximity, as it is the case here, a small wing pouch will still be specified. During wing disc eversion, the DV boundary will only present A identity, P compartment being reduced to one of the two sides of the wing depending on the P size, it might form a small “bag-like” structure. D’ Example of a Δm1/DG3 wing disc and wing extracted from figure 1 and 2 that might be following this model. Note that the margin of this wing is only composed of anterior and tip-specific margin bristles E. Δm1 homozygous mutant wing and detail of the campaniform sensillae in the P wing.

Posterior hinge and Upd expression patterns are affected in ap mutants.

A. Locali-zation of Ptc (Magenta) and Hth (Gray) in control wing discs. Hth is expressed along the hinge. (Posterior dorsal hinge is indicated with an arrowhead) A’. Localization of Ptc and Hth in Δm1/apDG3 mutants. In these cases, P hinge is depleted (arrowhead) and the Ptc stripe deviated. B. UAS-CD8:GFP driven by upd-Gal4 in wild-type wing discs. In the control discs, GFP signal could be detected in three main spots in the dorsal hinge (arrowheads) and two weak areas of expression in the ventral body wall (asterisks). B’. UAS-CD8:GFP driven by upd-Gal4 in ΔOR463/ apDG3 hemizygous background. Notice that the P source of Upd in the hinge is totally lost in this condition. Scale bars: 100μm

Deletion analysis within m3 region.

A. Schematic representation of the different deletions and representative phenotypes. Examples of the loss of the GATA binding site and the most distal m3 region are shown. Note that Δm3.3 flies do not develop wing structures. B. Scoring of the penetrance of mutant phenotypes in the different deletions. C. anti-Ap and anti-Wg immunostaining in both Δm3 deletion and Δm3.3. In both cases Ap localization is only residual within the anterior hinge.

Correct spacing between GATA and HOX binding sites is essential.

A. Adult wing phenotype of control flies. B. Loss of wing produced upon extension of the spacing between the GATA and HOX binding sites by 6pb Notice that this genotype never resulted in normal wings (See scoring of Figure 5D). Wings were either completely missing (35%) or showed sever phenotype. C. Example wing phenotype resulting from the contraction of the linker between the GATA and HOX binding sites. In this case, observed phenotypes were clearly weaker than for m3+6bp. Apart from normal wings (30%), most of the wings had an outgrowth from the posterior compartment or looked as depicted. D. Adult wing phenotype of control flies (Same specimen as in A). E. Loss of wing produced upon Δm3 deletion (same as Figure 1G). F. Example of the wing phenotypes obtained upon contraction of the m2’s GATA-HOX linking sequence in the absence of m3. Note that the majority (68%) of wings were missing for this genotype. Among the rest, most wings looked as shown in F’. A few cases as shown in F’’ were also present.

Pattern of Grn:GFP localization during wing development.

grn:GFP wing disc stained with anti-Ap and anti-Wg. (56h) At this stage Grn:GFP localizes to the presumptive hinge and seems to be highly down-regulated in the future notum and incipient pouch (marked by Wg). Grn:GFP was detected at higher levels in the anterior compartment. It was also detected broadly in the peripodial membrane. (62h) Similar Grn:GFP localization to that of 56h. (78h) Grn:GFP localization is higher in the proximal hinge, but persist at low levels in some lateral areas. Peripodial expression is higher in the lateral sides. (94h) Grn:GFP localizes sharply in the proximal hinge, totally excluded from the notum. Proximal hinge levels are very low. Only the cubic cells of the peripodial membrane seem to express grn. Arrowheads indicate the intersection between trachea and the disc proper. Scalebars: 50μm

Antp is required for clone survival in early stages and does not affect ap expression at later stages.

A. Example of wing disc in which Antp MARCM clones were generated during embryonic stages (0-24h) (See materials and methods for full genotype). In this case, many clones could be retrieved in other tissues (data not shown), but never in the disc proper of the wing disc. Clones, marked by the presence of GFP, could only be retrieved in the peripodial membrane (detail of the peripodial membrane in the right panel). In magenta, anti-Ap immunostaining reveals normal Ap pattern (left panel). B. Example of a wing disc of the same genotype as in A in which the clones were generated at 48h in development. In this case, small Antp clones (GFP positive) could be detected in the disc proper. Anti-Ap immunostaining revealed no change in the levels inside the clone (arrowheads).