|Select image to enlarge|
Weird wings that reveal the logic of A-P patterning. Gene abbreviations: dpp (decapentaplegic), en (engrailed), hh (hedgehog), inv (invected), ptc (patched), smo (smoothened). All panels show the dorsal side of a right wing (most bristles omitted for clarity), light vs. dark shading denote en-OFF vs. en-ON states, thick hatching (omitted in some panels) means dpp-ON, and arrows mark clone limits.
a. Wild-type wing. Landmarks: tegula (Te), proximal, medial, and distal costa (Cp, Cm, Cd), veins 1-5, posterior margin (m), alula (Al), and axillary cord (AC). Inset at left shows the base of a 30A-Gal4:UAS-en wing where en is ectopically expressed in the Te-Cm region . This region has transformed to AC, and a mirror-image copy of Cd has grown out -- possibly due to Dpp elicited at the en-ON/OFF interface. Evidently, this interface does not enforce a vein 3/4 identity immediately, though the dpp-ON cells may bootstrap themselves to such a level eventually if they have sufficient time to do so (e.g., see g).
b. How en controls dpp. Cells whose en gene is ON secrete Hh (not shown), and diffusion of Hh into en-OFF territory turns on dpp (cf. Fig. 6.4). 'Xenotopic' clones that express en in the A region (above) thus acquire a 'halo' of dpp-ON cells, while en-OFF clones in the P region (below) acquire a dpp-ON 'lining'. Only clones that cross the D/V boundary reorganize pattern or grossly alter growth, and they are elongated (e.g., e and f) vs. round .
c, d. Symmetries that arise when inv is misexpressed. c. Phenotype of an invGOF (dominant) mutant that expresses inv throughout the wing, but which still expresses en only in the P region . Unlike en, inv affects how cells respond to Dpp (i.e., A- vs. P-type competence) but does not affect dpp or its regulators (hh, ptc, ci, etc.). Thus, inv acts downstream of en (cf. Fig. 6.4). P/P wings are also seen when a Gal4 transgene forces en -- and hence inv -- ON in the A part of the blade (not shown) , but in that case vein 1 stays intact. Likewise, fat-headLOF wings manifest a P/P phenotype because fat-head normally represses en in the A compartment . d. Partial A/A symmetry (akin to en1 [1369, 1837]). Surprisingly, en-Gal4:UAS-en depresses en and inv in response to initial excess of En  -- implying negative autoregulation (En | en) . A/A wings are also seen with LOF changes in EGFR signaling: dRafLOF , pointedLOF , Ras1LOF , and veinLOF . The etiology of the EGFR effects is obscure.
e, f. Effects of inducing ptcnull (e) vs. ptcnull ennull invnull (f) clones in the vein 1 area. Disparity in clone breadth (3-3 distance) may depend on activation (e) vs. failure to activate (f) en within the clone, which dictates whether dpp turns ON outside or inside (zone in f is conjectural) . In both cases, the ectopic dpp-ON tissue induces a symmetric outgrowth, and the R-L-R handedness obeys Bateson's rule (cf. Fig. 5.3). Weaker ('1221' vs. '123321') outgrowths arise with dppGOF clones (not shown) .
g. Effect of an ennull invnull clone at the P margin. This A-type clone evidently formed a new dpp-ON zone (not shown) at its interface with P-type tissue, and that zone fostered a whole new A and P compartment. The R-L mirror plane (cf. e, f) illustrates an 'edge effect' predicted by Meinhardt .
h. Effect of substituting a membrane-tethered Hh (which cannot diffuse) for the endogenous Hh. The wing is small, and veins 3 and 4 are fused, but other features are relatively normal (cf. Fig. 6.13c).
i. Effect of a smonull clone of A-type cells at the A/P border (cf. Fig. 6.5). Hh diffuses farther over smonull cells but is not transduced by them, so Hh only turns ON dpp when it reaches competent tissue beyond the clone. Evidently, this journey diminishes Hh's strength because the amount of Dpp fails to foster (1) normal growth, (2) patterning of the P region, or (3) full '123321' duplication of the A region (cf. e, f).
Wings (all at same scale) were traced from pictures in  (inset in a),  (c),  (d-g),  (h), and [754, 786] (i). Schematic (b) is based on [354, 1647, 4136, 4229, 4848]. Another 'brainteaser' phenotype (not shown) whose etiology was figured out using known circuits is dpp-Gal4:UAS-tkv  (their Fig. 6b). How duplications arise in EgfrLOF groLOF wings (not shown) is unknown .