Insights into Digit Evolution from a Fate Map Study of the Forearm

11 The cellular and genetic networks which contribute to the development of the zeugopod, 12 (radius and ulna of the forearm, tibia and fibula of the leg) are not well understood, although 13 these bones are susceptible to loss in congenital human syndromes and to the action of 14 teratogens such as thalidomide. Using a new fate mapping approach in transgenic chickens, 15 we show that there is a small contribution of SHH expressing cells to the posterior ulna, 16 posterior carpals and digit 3. We establish that while the majority of the ulna develops in 17 response to paracrine SHH signaling in both the chicken and mouse, there are differences in 18 the contribution of SHH expressing cells to other tissues of the zeugopod between these two 19 species as well as between the chicken ulna and fibula. This is evidence that although 20 zeugopod bones are clearly homologous according to the fossil record, the zeugopod bones of 21 the wing and leg are formed by subtly different signalling and patterning events during 22 embryonic development, which can be used to understand the shaping of the bird wing 23 skeleton during the evolution of powered flight.


25
Limbs first form as small paired forelimb or hindlimb buds growing from the flank of a 26 developing embryo (Tickle, 2015). The mesodermal cell component of the limb bud, derived 27 from the lateral plate mesoderm (Gros and Tabin, 2014), forms the majority of the limb 28 skeleton from the proximal shoulder/pelvic girdle to the digit tips. The cells that make up the 29 early limb look homogenous but fate maps of the early chicken wing bud show that at stage The zeugopod is, however, subject to many of the same patterning mechanisms as the 48 autopod and parallels between these parts of the limb can be drawn, specifically between the 49 antero-posterior axis patterning by SHH and FGF pathways (Chiang et   The evolution of the bird wing, in particular understanding which two digits were 'lost' and 74 which three remain in the modern tridactyly wing, is studied both to understand the context of  Xu et al., 2014). In these studies, evolutionary anatomical changes in the 88 zeugopod bones, have been overlooked as homology of the radius and ulna is easily assigned 89 and both are clearly present throughout the fossil record. Rather, the emphasis has been that 90 morphology of the carpals and digits has evolved distal to the 'unchanging' bony anatomy of 91 the forearm, the radius and ulna (Fig. 1D). This is embodied in the a foundation principal, the  The 'frame-shift' model ( Fig. 1E), primarily based on embryological evidence such as the 99 development of SOX9+ digit primordia, proposes that the primary axis is maintained and the 100 ulna-digit 4 articulation remains unchanged, but that a modified digit 4 takes on a 101 morphological identity of a digit III through a homeotic transformation, thereby concluding 102 that digit 1 and 5 are lost (de Bakker et al., 2013). Alternatively based on both fossil and 103 embryological data, specifically the contribution of SHH expressing cells to the digits as a 104 indicator of lineage, the 'axis-shift' model ( Fig. 1F) suggests that the articulation between the 105 primary axis/ulna shifts from digit 4 to digit 3, but does not account for how the change in 106 this relationship might have occurred (Towers et al., 2011). A limitation of all these studies 107 has been a lack of analysis of the bones proximal to the digits although analysis of the carpals 108 suggests that these bones, articulating the zeugopod with the autopod, have been even more 109 radically altered than the digits (Botelho et al., 2014). We propose that understanding

134
The ulna arises from a discrete area within the chick limb bud 135 To locate the area from which the ulna is specified in the stage 20HH chick wing bud, we

146
As the ulna is known to be dependent on SHH signaling, we first examined the fate of the 147 presumptive zeugopod forming region at 20HH as identified by Saunders (1948; Fig. 1A, B).  Fig. 2A, B). This region of the limb lies above the SHH expressing ZPA 152 cells but expresses PTCH1, a hedgehog receptor whose expression is induced by the ligand,    Supplementary Fig. 2). 254 These results demonstrate that the ulnar primordium is spatially defined, consistent with the 255 original chicken limb fate maps of Saunders (1948) and Summerbell (1974), but SHH     Saunders (1948) and Summerbell (1974) 285 suggest that the proximal chicken ulna is specified before stage 20HH, so we therefore sought 286 to establish if the difference between mouse and chicken data could be resolved by 287 undertaking ZPA grafts earlier in development. We implanted TAT-Cre beads into the distal 288 limb mesenchyme of stage 18HH limbs at the axial level of anterior somite 19 (Fig. 4F) and 289 more posteriorly into the ZPA (Fig. 4I). Localisation of fluorescent clones were substantially grafts to stage 18HH eGFP embryos (Fig. 4L). In this instance we did find that dtTom ZPA  which limb-specific SHH signaling is lost (Ros et al., 2003). We mapped the chick ulna in the 376 stage HH20 limb bud, showing that it consistently arises from a highly discrete area that is 377 adjacent to anterior somite 19 and predominantly outside of the ZPA (Fig. 5A, B). Its stage 378 HH20 primordium is SHHbut PTCH1+, suggesting that the ulna is primarily patterned 379 through paracrine SHH activity. Overall, the mouse and chicken ulna both appear to be 380 largely subject to paracrine SHH signaling but have a varied distribution of SHH descendants.

381
In the OZD chicken, the fibula is also lost (Ros et al., 2003). However, unlike the ulna, we 382 found that the majority of fibular cartilage is derived from SHH expressing cells, suggesting 383 mostly autocrine SHH signaling. Digit identity as determined by ZPA contribution is often 384 used as a fixed preliminary for proposing hypotheses for limb variations across species but 385 also interchangeably between the fore and hind limbs. However, we show that even structures  In conclusion, we show although the mouse and chicken ulna are predominantly SHH-, 413 suggesting paracrine patterning. Unlike the ulna, chick fibular cartilage is mostly descended 414 from the ZPA and thus, although the postaxial zeugopod is seen as fixed and often considered 415 as analogous, we demonstrate that these actually have different constituents of SHH lineage.

416
The ulna and fibula may be more evolutionarily diverse than supposed and therefore, their   The graft was manoeuvred into the host site and, when necessary, secured with a piece of 458 0.02mm oxidised nickel chrome wire. Care was taken to ensure ectoderm orientation was 459 maintained between donor and host. Embryos for wholemount analysis were culled and 460 dissected at around stage 33HH, fixed and cleared with CUBIC reagent 1 before being 461 imaged on a Zeiss Axiozoom V16 microscope. Embryos for HCR in situ hybridisation were 462 allowed to incubate for 3 hours after graft insertion, then culled and dissected in cold DEPC 463 PBS before being fixed with 4% PFA at 4C overnight. 465 Fertilised eggs were windowed, prepared for manipulation as per Tiecke and Tickle (2007) 466 and staged (Hamburger and Hamilton, 1951). Once bead manipulations were complete, the 467 window was sealed with tape and incubated at 38C in a humidified and light-free    The authors state that they have no financial and non-financial competing interests.