The Drosophila Neurogenin, Tap, controls axonal growth through the Wnt adaptor protein Dishevelled

The Neurogenin (Ngn) transcription factors control early neurogenesis and neurite outgrowth in mammalian cortex. In contrast to their proneural activity, their function in neurite growth is poorly understood. Drosophila has a single predicted Ngn homologue called Tap, whose function is completely unknown. Here we show that Tap is not a proneural protein in Drosophila but is required for proper axonal growth and guidance of neurons of the mushroom body (MB), a neuropile required for associative learning and memory. Genetic and expression analyses suggest that Tap inhibits excessive axonal growth by fine regulation of the levels of the Wnt signaling adaptor protein, Dishevelled. Summary The Drosophila Neurogenin homologue, Target of Pox neuro (Tap), prevents axonal overgrowth by regulating the Wnt Planar Cell Polarity pathway adaptor protein Dishevelled.


41
Proper function of the nervous system is based on the production of a diversity of 42 neuronal and glial cells, as well as the precise targeting of their axons and dendrites. 43 A key transcription factor (TF) family, which controls the commitment of neuronal 44 cell fate and neurite guidance, is the Neurogenin (Ngn) family. Ngn proteins belong to 45 a structurally and functionally conserved basic helix-loop-helix (bHLH) superfamily. 46 Ngns in vertebrates are sufficient to initiate the neuronal cell fate in the central 47 nervous system (CNS) (reviewed in Bertrand et al., 2002;Ma et al., 1998). In 48 contrast, very little is known about the function of Ngn proteins in invertebrate 49 systems. Gain of function of analyses in Drosophila and vertebrate models suggested 50 that during evolution a switch in proneural activity occurred between the Ngns and a 51 highly related family of bHLH TFs called the Atonal family. Specifically, whereas 52 Ngns are necessary and sufficient for the induction of neurogenesis in vertebrates, 53 they cannot do so in flies. Conversely, in flies Atonal type proteins can induce 54 -3 -neurogenesis, but fail to do so in vertebrates (Quan et al., 2004). A crucial test of this 55 "evolutionary proneural switch hypothesis" is whether or not Drosophila Neurogenins 56 act as proneural genes in flies and vertebrates. 57 58 In addition to their proneural function, Ngns play various critical roles in the 59 development of vertebrate nervous system, including regulating the outgrowth and 60 targeting of both axons and dendrites (Hand et al., 2005;Hand and Polleux, 2011;61 reviewed in Yuan and Hassan, 2014). The regulation of such neurite growth is 62 independent from the proneural activity, as the phosphorylation of a single tyrosine in 63 mouse Ngn2 is necessary to specify the dendritic morphology without interfering the 64 cell fate commitment (Hand et al., 2005). However, the mechanism of how Ngn 65 proteins regulate neurite guidance is poorly understood. 66

67
The Drosophila genome encodes a single member of Ngn family based on the 68 conservation of family-defining residues in the bHLH domains (Hassan and Bellen, 69 2000), named Target of Pox neuro (Tap). Previous work suggests that Tap is 70 expressed in secondary progenitors of both neurons and glia, called ganglion mother 71 cells, in CNS (Bush et al., 1996), as well as the putative support cells in the peripheral 72 nervous system (PNS) during embryogenesis (Gautier et al., 1997). A previously 73 reported putative tap mutant allele (Ledent et al., 1998) was later shown to be a 74 mutation in a different gene called blot (Johnson et al., 1999). Therefore, the function 75 of the Drosophila Ngn Tap remains uncharacterized.  76   77 Here we generated a null mutant allele of tap by replacing the single coding exon of 78 the tap gene with Gal4. Using expression, gain of function and loss of function 79 analyses we show that, whereas ectopic expression of Tap in vertebrates can induce 80 neurogenesis, tap is not a proneural gene in flies, consistent with the evolutionary 81 proneural switch hypothesis. Instead, Tap is required to prevent overgrowth of axons 82 during brain development, at least in part through the activity of the axonal Wnt 83 planar cell polarity (Wnt-PCP) pathway by fine tuning the levels of the Wnt signalling 84 adaptor protein Dishevelled (Dsh). 85

87
Tap is the only Ngn homolog in Drosophila 88 Drosophila Tap shares significant identity (~70%) in the bHLH domain with mouse 89 and human Ngns (Fig. 1A), compared to Drosophila Atonal or Scute. To test if Tap, 90 like its vertebrate courter parts, has proneural activity, we injected Tap mRNA into 91 one cell of 2-cell stage Xenopus embryos and assessed neurogenesis by staining for 92 neuronal markers. Like mouse Ngn1, but unlike fly Atonal, Tap efficiently induces 93 neurogenesis in this system ( Fig. 1B-E). Conversely, when ectopically expressed in a 94 classic Drosophila proneural assay Tap -like Ngn1, and in contrast to Atonal -fails 95 to induce neurogenesis ( Fig. 1H-I). These data suggest that Tap may not be a 96 proneural gene in flies, even though it has neural induction potential in vertebrates, as 97 expect for a bona fide Ngn protein. 98 Tap is widely expressed in the nervous system during development

99
To investigate the function of Tap, we generated a mutant allele using ends-in 100 homologous recombination (Rong and Golic, 2000) to replace the open reading form 101 of tap with an external driver, Gal4 ( Fig. 2A). The tap Gal4 allele, in homozygosity, 102 serves as a tap null mutant; while in heterozygosity, it was used as a driver to reveal 103 the expression pattern of Tap. We made several attempts to generate a Tap antibody, 104 but this was not successful. Tap  To circumvent embryonic lethality we began by exploiting the heterozygous tap Gal4 133 allele alone or in combination with two independent RNAi strains targeting distinct 134 regions of tap. In wild-type Drosophila, axons of the medially projecting β lobes 135 terminate near the midline but do not cross it (Strausfeld et al., 2003). In contrast to 136 the wild-type MB morphology (Fig. 3D), tap loss-of-function brains exhibit that β 137 lobe fibers extend across the midline (Fig. 3E), sometimes sufficient to cause fusion 138 of the two contralateral β lobes (Fig. 3F). The phenotype is variable in severity, thus 139 we classified it as "normal", "mild", or "severe" based on the thickness and density of 140 the β lobe fibers crossing the midline. Tap heterozygotes display an increase of the 141 penetrance of mild defects, while both tap RNAi strains raise the incidence and 142 severity of β lobe midline crossing defect. When Tap is re-expressed, the defect can 143 be rescued to control levels (Fig. 3J). This β axon midline-crossing defect is 144 developmental in origin as it can be observed at early pupal stage (Fig. S3). 145

146
In addition to the β axon midline-crossing defect, an "α lobe missing" phenotype was 147 observed in Tap loss of function MBs (Fig. 3H, I). In some brains with a missing α 148 lobe, the β lobe appears to branch into two bundles (Fig. 3I), suggesting this α lobe 149 defect may be an axonal targeting defect rather than a growth defect. Unlike the β 150 -6 -lobe defect, the penetrance of α lobe missing phenotype varies considerably (Fig. 3J). 151 Considering the frequency is still higher in the tap mutant flies compared to wild-type 152 and rescue strains, we conclude that Tap is essential for the growth and correct 153 targeting of both lobes of α /β neurons. 154 Tap is required cell-autonomously for the growth and targeting of the any defects (Fig. 3K). In contrast, in 32% of tap null mutant clones, β axons project 165 beyond the β lobe domains (Fig. 3L, N). In severe cases, axons were observed to 166 cross the midline and project to the contralateral β lobes. This defect can be rescued 167 by re-introduction of tap specifically in the mutant clones (Fig. 3M, N). This suggests 168 Tap is required cell-autonomously for the development of the β axon branch. 169 However, none of the mutant clones showed loss of α axon growth, suggesting Tap 170 plays a non-cell-autonomous role during α lobe development. 171

172
In order to investigate the molecular mechanism of how Tap regulates axonal growth, 173 we performed dominant interaction tests between Tap and well-established MB 174 axonal guidance factors, particularly those whose loss of function causes β lobe 175 overgrowth. Specifically, we asked if heterozygosity for any of these genes strongly 176 enhances the very mild phenotypes observed in tap heterozygotes over the sum of the 177 phenotypes observed in each allele. Among the candidate genes, drl, Sli and Dscam1 178 did not show any obvious synergism with Tap (Fig. 4A). However, loss of one copy 179 of dsh strongly enhances both α lobe guidance and β lobe overgrowth defects in the 180 tap heterozygous background (Fig. 4A, B)  of dsh, dsh 6 , but does not affect the α lobe missing defect (Fig. 4B). These data 199 suggest that the role of Tap in the regulation of axonal growth and guidance in α lobe 200 and β lobe are independent. Furthermore, this is consistent with previous findings that 201 Wnt-PCP signaling β lobe growth cell autonomously, but α lobe growth cell non-202 autonomously (Soldano et al., 2013). In addition to Dsh, three other key Wnt-PCP 203 components regulate α β lobe growth namely, Appl and Vang and Wnt5a. We find that 204 heterozygosity for all three mildly enhances the tap defect, but much less than Dsh 205 does, suggesting that Tap regulates PCP signaling specifically through Dsh (Fig. S4). 206 207 Our data suggest that tap regulates neuronal extension and guidance through the Wnt-208 PCP pathway. Cellular polarity created by the PCP pathway is essential to direct cell 209 movement, which is analogous to guide the movement of growth cones, thus navigate    The mRNAs of ato, tap and mouse Ngn1 were injected in a single blastomere of 314 Xenopus embryos at the two-cell stage. The whole embryos were in situ hybridized 315 with anti-N-tubulin as described in (Quan et al., 2004). 316 317 Cloning and gene targeting