Hedgehog signaling acts with the temporal cascade to promote neuroblast cell cycle exit

Abstract

In Drosophila postembryonic neuroblasts, transition in gene expression programs of a cascade of transcription factors (also known as the temporal series) acts together with the asymmetric division machinery to generate diverse neurons with distinct identities and regulate the end of neuroblast proliferation. However, the underlying mechanism of how this "temporal series" acts during development remains unclear. Here, we show that Hh signaling in the postembryonic brain is temporally regulated; excess (earlier onset of) Hh signaling causes premature neuroblast cell cycle exit and under-proliferation, whereas loss of Hh signaling causes delayed cell cycle exit and excess proliferation. Moreover, the Hh pathway functions downstream of Castor but upstream of Grainyhead, two components of the temporal series, to schedule neuroblast cell cycle exit. Interestingly, hh is likely a target of Castor. Hence, Hh signaling provides a link between the temporal series and the asymmetric division machinery in scheduling the end of neurogenesis.

Figure 1. Hedgehog signaling affects the proliferation of NBs.

(A–C) Third instar larval brains harbouring wt (A), ptc^S2 (B), and smo^IA3 (C) MARCM clones were immunostained to show the clone size (green, GFP channel), and Mira (red) localization. In wt clones (D–D′), and smo^IA3 (F–F′) clones, the mitotic NBs showed a strong Mira crescent; but Mira was highly cytoplasmic in ptc^S2 NBs (E–E′). In interphase ptc^S2 NBs (G–G′, GFP clones), the cortical Mira was weakened or absent (yellow arrowheads) compared to their wt counterparts (yellow arrows). DNA was stained with either PH3 (D and F), or ToPRO3 (A–C, E, and G). (H) Quantification of NB mitotic index in wt, ptc^S2, smo^IA3 and hh^AC clones based on the percentage of the NBs that expressed PH3 at 96 h ALH. (I–K) BrdU (red) incorporation in wt (I), ptc^S2 (J), and smo^IA3 (K) clones labeled by CD8:GFP (green) after 4 h of continuous feeding with BrdU. Scale bar = 10 µm.

Figure 2. A canonical Hh signaling pathway is required to control NB proliferation but not neuronal differentiation.

(A–I′) NB clones of different genotypes: wt (A–A′, D–D′, and G–G′), ptc^S2 (B–B′, E–E′, and H–H′) and smo^IA3 (C–C′, F–F′, and I–I′) were marked by CD8:GFP (green). Dpn was expressed in the nuclei of wt NBs (A–A′), and its expression remained unchanged in smo^IA3 NB (C–C′). However ptc^S2 NB had decreased levels of Dpn expression (B–B′) compared to wt NBs located outside of the clone (arrows). wt and smo^IA3 NBs showed a Pros crescent during mitosis (D–D′ for wt; unpublished data for smo^IA3) and did not show visible nuclear Pros expression during interphase (F–F′ for smo^IA3; unpublished data for wt), but ptc^S2 interphase NBs showed nuclear-localized Pros (E–E′). (G–G′) wt NB clone showing three undifferentiated GMCs (arrowheads) which lacked Elav expression. (H–H′) In a ptc^S2 NB clone all of the cells surrounding the NB expressed Elav. (I–I′) In smo^IA3 NB clones an enlarged cluster of cells surrounding the NB did not express Elav. (J–M′) Homozygous ci⁹⁴ clones as marked by the absence of GFP. (J–L′) Three consecutive z-sections (6 µm apart from each other) of a ci⁹⁴ clone exhibited areas that are both Dpn and Elav negative, occupied by GMC-like cells (arrows). The pink dotted line marked the outline of the NB (J–K). (M–M′) Pros was localized to the GMCs and neurons of ci⁹⁴. (N–N″) Over-expression of constitutively active form of ci using act-GAL4 flip-out system showed a single undifferentiated GMC (arrow) within the clone. (D–L, N) DNA was stained with ToPRO3. Scale bar = 10 µm. (O) Quantification of three different cell fates based on the absence or presence of Dpn and Elav expressions for both type I and type II NB clones in wt, ptc^S2, and smo^IA3 backgrounds. Error bars showed standard error of the mean (SEM). Statistical significance was determined using Student's t test; *p<0.05; **p<0.003.

Figure 3. The developmentally regulated cell cycle length of NB is affected by Hh signaling.

(A–C) Live-imaging video stills of a 72-h ALH wt NB (A) marked by histone-red fluorescent protein (RFP) (red) and G147 (green), as well as those of 72 h ALH ptc^S2 (B) and smo^IA3 (C) NBs marked by histone-RFP (red) and CD8:GFP (green). Cell cycle lengths were determined by counting the time taken from the appearance of cleavage furrow from one division to the next, on the basis of the appearance of GMC bud (A and B) or nuclear membrane elongation when GMC budded off (C). (D) Quantification of cell cycle length in wt, ptc^S2, and smo^IA3 NBs, at 48 h, 72 h, and 96 h ALH. ptc^S2 NBs failed to divide at 96 h ALH under our live imaging conditions. Error bars represent standard deviation (SD). Statistical significance was determined using Student's t test; *p<0.003; **p<0.05.

Figure 4. Hh signaling interacts with NB temporal cascade component.

(A–E′) About 50% of smo^IA3 (B–B′), cas²⁴ (C–C′), and svp¹ (D–D′) clones had a single NB that continues to express Mira (red) and PH3 (blue, unpublished data for svp¹) at 48 h APF. Meanwhile, none of the cells within any wt clone (A–A′) express Mira or PH3 at that time point. Such delay in cell cycle exit could be largely reverted by expressing ptc^RNAi in the cas²⁴ mutant background (E–E′). (F–G′) cas²⁴ and svp¹ clone had increased numbers of Dpn-negative, Elav-negative, GMC-like cells at 96 h ALH (Dpn in red, Elav in blue; arrows, in focus; and arrowhead, out of focal plane). (H–I′) This phenotype of cas²⁴ clones can be largely suppressed by the introduction of ptc^RNAi (H–H′) or PP4-19C^RNAi (I–I′). Note the number of GMC-like cells (arrows and arrowhead). Scale bar = 10 µm. (J) Quantification of 48 h APF clones harboring Mira positive cells in various backgrounds. (K) Quantification of GMC numbers in various mutant backgrounds. A typical wt NB clone contained three to five GMC-like cells which are Dpn- and Elav-negative. In ptc^S2 clones, there was a decrease in the number of GMC-like cells while smo^IA3, svp¹, and cas²⁴ clones had ectopic GMC-like cells. The number of GMC-like cells in cas²⁴ clones could be reduced to a level close to that of wt with the expression of ptc^RNAi or PP4-19C^RNAi.

Figure 5. An early transient pulse of Cas is required for later hh expression.

(A–B) hh transcript (red) was detected mainly in the GMCs of 72 h ALH brain lobe (B), but not in that of 48 h ALH brain lobe (A). CD8:GFP (green) was driven with elav-Gal4 to mark the outlines of the cells. (C–C″) In situ hybridization of hh (red) in a wt clone showed that the transcript was expressed in the GMC adjacent to the Dpn-expressing NB. (D–D′) In situ hybridization using an intronic probe that detects hh pre-mRNA (red) in a wt clone showing hh expression in the mitotic NB (note the nuclear morphology, arrow), as well as in the GMC next to it (arrowhead). (E–E′) The mature form of hh mRNA (red) was detected in the cytoplasm of the NB (arrow) and to a lesser extent, the adjacent GMC (arrowhead). (F–H) Immuno-staining against Cas (red) showed its expression in cells (NBs and intermediate neural precursors [INPs]) that were also co-expressing Dpn (blue, arrowhead), as well as in some other neurons at different developmental time points. The outlines of the cells were marked with membrane GFP (green). The brain lobes were position such that the dorsal side is facing up and the dotted line indicated the midline of the brain. (I–J′) In situ hybridization of hh mRNA showed that embryonic Cas was required for normal hh expression. cas²⁴ clones induced at 24 h ALH did not affect hh expression (I–I′) while cas²⁴ clones induced during late embryonic stage could reduce or abolish hh expression (J–J′). Pon (blue) showed the outline of the newly born GMCs that typically expressed hh mRNA. (A–F, I–J) Scale bar = 10 µm for all panels. (K) ChIP for Flag-tagged Cas transfected into S2 cells showed that Cas was heavily enriched within the 6-kb region at the 5′ UTR of hh gene (orange box). There are 19 putative Cas binding sites within that region. The enrichment of any region of the chromatin was counted as the multiple of specific binding (anti-Flag) against non-specific binding (anti-IgG), and normalized to the enrichment at the actin promoter site (negative control). A known target of Cas, pdm-1 was used as a positive control (grey box) for comparison purposes. The value of the blue bars was the average enrichment (times) from three independent transfections and five independent ChIPs. Error bar correspond to standard error of the mean (SEM).

Figure 6. Hh signaling affects the maintenance of grh expression.

(A–D′) Grh (red) was expressed in wt NBs (marked by Dpn, blue) and some GMCs at 72 h ALH (A–A′) and 96 h ALH (B–B′). Grh expression was down regulated in the NBs and abolished in the GMCs by 12 h APF (C–C′) and its expression was barely visible at 24 h APF (D–D′). (E–G′) ptc^S2 NB showed normal expression of Grh at 72 h ALH (E–E′) but its level quickly decreased by 96 h ALH (F–F′) and was completely abolished at 12 h APF (G–G′). (H–L′) Grh expression was detected in smo^IA3 NB and GMCs at 72 h ALH (H–H′), and 96 h ALH (I–I′) but persisted until 12 h APF (J–J′) and failed to be down-regulated at 24 h APF (K–L′). (L) A smo^IA3 NB at 24 h APF with persistent Grh expression amidst the wt background (non-green) in which all the NBs had down-regulated their Grh expression. Scale bar = 10 µm.

Figure 7. Hh signaling interacts genetically with grh to orchestrate NB cell cycle exit.

(A–C′) smo^IA3 clones (A–A′) marked with CD8:GFP (green) often contained a single NB that expressed Mira (blue) but was devoid of nuclear Pros (red) at 24 h APF. However, no Mira expressing cell could be found in smo^IA3 clone that had Grh level reduced by RNA interference (B–B′). As a control, clones expressing grh^RNAi transgene (C–C′) alone did not contain any Mira expressing cell either. RNA mediated knock-down of grh in smo^IA3 background at 96 h ALH (D–D′) significantly rescued the ectopic GMC-like cells phenotype as the number of cells negative for Dpn (red) and Elav (blue) plummeted to wt level (arrowheads). (E–E′) showed a control grh^RNAi clone with its GMCs pointed out by the arrowheads, at 96 h ALH. (F–I″) Over-expression of grh in ptc^S2 clones substantially rescued the Mira delocalization and Dpn-expression defects in the interphase NBs at late L3 stage. More than 60% of the NB within ptc^S2 clones (marked by CD8:GFP in green) displayed weak cortical Mira (red) and low intensity of nuclear Dpn (blue) as compared to the surrounding wt NBs (F–F″), while the rest of the interphase ptc^S2 NBs had largely normal Mira localization and nuclear Dpn intensity (G–G″). ptc^S2 clones that over-expressed grh had a higher frequency NBs with normal Mira localization and nuclear Dpn intensity (H–H″), whereas control NB that over-expressed grh was indistinguishable from wt NB in terms of Mira localization and nuclear Dpn intensity (I–I″). Scale bar = 10 µm.

Figure 8. Hh signaling acts as a functional link between the temporal cascade and the asymmetric division machinery.

(A–B′) The excessive cytoplasmic Mira and weak Mira crescent seen in ptc^S2 NB during mitosis (A–A′) can be rescued by removing one copy of pros (B–B′). (C–D) Similarly, the excessive cytoplasmic Mira and weak Mira crescent seen in flfl⁷⁹⁵/flfl^N42 transheterozygous NB (C) can be rescued by removing a copy of pros (D). (E–F′) A flfl⁷⁹⁵ NB showing weak Mira crescent and cytoplasmic Mira (E–E′). Such Mira localization defects can be rescued via the introduction of ci^RNAi. (G) Quantification of Mira localization phenotypes in various mutant backgrounds.

Figure 9. The model.

(A) An early pulse of Cas at early larval stage primes the expression of hh in both NB and GMCs. The Hh ligand (unpublished data) acts in an autocrine and/or paracrine fashion to activate Hh signalling transduction in the NB. Among the outcomes of Hh signaling pathway activation in the context of postembryonic NB, grh expression is down-regulated and Pros moves into the nucleus, eventually leading to NB cell cycle exit at early pupal stage. The strength of Hh signaling activation is likely to be regulated PP4, which dampen Hh pathway activity by dephosphorylating Smo. svp probably constitutes a parallel pathway, which may or may not converge with cas→hh pathway at a point prior to the eventual NB cell cycle exit. Dotted lines and double question marks denote uncertainties while solid lines represent availability of experimental evidences. Triangular and diamond-shaped arrowheads imply positive and negative regulations respectively. This illustration is not drawn to scale. (B) Type I NB lineage trees at three different developmental stages, during which the expression of Grh (red), Cas (blue), Svp (yellow), Hh (green), and Elav (dark blue) are shown. The NB, GMC, and neuron are represented by circles with black, tan, and brown outlines, respectively. The green arrows show the autocrine/paracrine mode of Hh signaling during lineage progression.