Signal strength and signal duration define two distinct aspects of JNK-regulated axon stability

Abstract

Signaling proteins often control multiple aspects of cell morphogenesis. Yet the mechanisms that govern their pleiotropic behavior are often unclear. Here we show activity levels and timing mechanisms determine distinct aspects of Jun N-terminal kinase (JNK) pathway dependent axonal morphogenesis in Drosophila mushroom body (MB) neurons. In the complete absence of Drosophila JNK (Basket), MB axons fail to stabilize, leading to their subsequent degeneration. However, with a partial loss of Basket (Bsk), or of one of the upstream JNK kinases, Hemipterous or Mkk4, these axons overextend. This suggests that Bsk activity prevents axons from destabilizing, resulting in degeneration and overextension beyond their terminal targets. These distinct phenotypes require different threshold activities involving the convergent action of two distinct JNK kinases. We show that sustained Bsk signals are essential throughout development and act additively but are dispensable at adulthood. We also suggest that graded Bsk inputs are translated into AP-1 transcriptional outputs consisting of Fos and Jun proteins.

Fig. 1

JNK is highly expressed in MB axons and dendrites. (A) MB axons labeled by CD8-GFP expression, using the OK107-Gal4 driver. These adult MB axonal projections are wild type. Axons terminate close to the midline (to the right of all images, unless indicated with a dashed white line), or close to the anterior dorsal cortex (dashed orange line). The different axon lobes (γ, α'β', αβ) are indicated, as previously (Lee et al., 1999). (A') A schematic of these MB neuron subtypes (γ, α'β', αβ) and the relative location of cell bodies, dendrite (‘calyx’) and axon projections. (B) The same brain immunostained with anti-JNK1, the overlap with the GFP marker (magenta in B') shows axonal localization. (C) The same brain also immunostained with anti-phospho JNK (P-JNK). (C') The overlap shows high JNK activity in MB axons. (D–I) P-JNK brain staining at various developmental stages, from wandering larvae L3 and pupae at different time-points after puparium formation (APF), indicated in hours (h). The additional panels (D'–I') show the corresponding overlap between P-JNK and the GFP labeled MB axons at these stages. Unless indicated otherwise, these and subsequent images are z-stack of serial confocal images taken at 1-μm thickness. In some images (such as in A), cell body sections have been omitted to clearly reveal axonal projections. Scale bar: 20 μm.

Fig. 2

JNK loss results in axon destabilization. (A, A') Adult MB bsk^147eneuroblast clones show axon thinning (yellow arrow) and termination defects (open white arrowheads). (B–I) Images of CD8-GFP labeled wild-type (B–E) and bsk^147e (F–I) neuroblast clones analyzed at developmental stages: 0 h (B, F), 24 h (C, G), 48 h (D, H) and 72 h APF (E, I). At the onset of puparium formation (0 h APF), the majority of wild-type MB neurons consist of γ and α'β' neurons (B). A phase of neurogenesis occurs at this period and MB neuroblasts give rise to αβ neurons, which are visible at 24 h APF (C). As axon continue to grow, dorsal α and medial β lobes become more prominent. At 48 h APF, these lobes are similar to adult MB projections (D compare with Fig. 1A, respectively). Note bsk axonal defects at later stages of development (quantified in N). Axon defects were also more pronounced at later stages (compare H to I). The cell body section has been omitted from I. (J–L) Images of bsk^147e single cell clones (γ-neurons) showing breaks (yellow arrowheads) along different regions of the axon. (J', K and L) Higher magnification of MB axons shows breaks and axon thinning in the proximal and mid-axonal shaft (J',K, respectively) and in the distal section (L). Scale bars: 20 μm. Unless indicated otherwise, CD8-GFP labeled neurons are shown in green or grayscale and Fas2 immunostaining (magenta) labels a subset (γ weakly and αβ strongly) of all MB axons. (M–N) Quantification of β-axon termination defects in adult MB bsk^147e neuroblast clones (M), and at specific time points in development (N). n, number of neuroblast clones analyzed. While all MB axons displayed degeneration phenotypes to some extent, to simplify this study, we decided to focus on the medially projecting β-lobe. For bsk additional images and quantifications of other MB projections, see Supplementary Fig. 2.

Fig. 3

JNK phosphorylation is essential for axonal morphogenesis. (A, B) bsk^147e neuroblast clones in the presence of Bsk mTPY (A), or wild-type Bsk (B). Wild-type Bsk, but not Bsk mTPY, expression rescues the axon phenotype. Cell body sections have been omitted in both panels. Scale bars: 20 μm. Green, CD8-GFP. Magenta, Fas2.

Fig. 4

Expression study of JNK kinases Hep and MKK4. (A–F ') Single confocal sections of MB neurons labeled with CD8-GFP and immunostained with Hep (A–C') or Mkk4 (D–F') antibodies. The corresponding panels (A'–F') show overlap between Hep and Mkk4 signals and CD8-GFP labeling. Single sections show γ (A, D), α/β and α'/β' axons (B, E) and MB cell bodies (cb) (C, F), as indicated in A'–C'. (G–G) CD8-GFP labeled MB neurons (green) expressing ectopic Hep (shown in magenta in G and G' and grayscale in G). Dorsal (y) projection views (G' and G) show Hep is mainly localized to axons (ax). (H–H) Representative image of MB neurons expressing MKK4YFP (green) and stained with anti-MKK4 (magenta in H and H' and white in H). Dorsal views (H' and H) show MKK4 is localized to axons and cell bodies (cb). Scale bar: 20 μm (x-only).

Fig. 5

Loss of function of Hep and MKK4 in MB neurons. Representative images of hep^R39 (A), hep^R75 (B) and MKK4^e01458 (C, D) neuroblast clones exhibiting β lobe axon degeneration (A), axon overextension (B, C) and cell proliferation (D) phenotypes. Neuroblast proliferation phenotypes are characterized by the presence of early-born γ neurons and absent later-born α'β' and αβ neurons. (E–H) hep^R39(E, F) and MKK4^e01458(G, H) clones in the presence of ectopic MKK4YFP (E, H) or Hep (F, G). Scale bars: 20 μm. Green, CD8-GFP. Magenta, Fas2. (I) Quantification of the β axon phenotypes. As Mkk4 leads to proliferation defects in early-born neuroblast clones, later-born αβ neuroblast clones were generated for axon studies. n, number of neuroblast clones analyzed.

Fig. 6

Partial inactivation of Bsk leads to axon overextension. (A, B) MB neurons expressing Bsk RNAi. High level of Bsk RNAi knockdown leads to a bias in axon degeneration phenotypes (open arrowheads in A), whereas medium RNAi activity levels lead to dorsal axon overgrowth and medial overextensions (open arrows in B). (C) Dominant-negative Bsk (DN Bsk) misexpression resulted in similar phenotypes. We found that pan-MB inactivation of Bsk also resulted in defasciculation defects, characterized by wider, splayed-out axon lobes (for example, the dorsal projection in A, indicated by the open arrow). (D) Quantification of these phenotypes. n, number of hemispheres analyzed. Given that, in many instances, the β-lobe overextension from one hemisphere overlaps against either a similarly overextending, or otherwise wild-type, axon from the contralateral side, we also present a different analysis of the medial projections for all (pan-MB) overexpression genotypes as the number of brains that have β-lobe overextensions over the total number of brains analyzed (in italics next to the relevant bar). Additional analysis and quantifications for dorsal projections are in Supplementary Fig. 6. (E) Representative image of a hep^R75, MKK4^e01458 mutant single-cell αβ clone. Note axon breaks in the dorsal and medial branch (open arrowheads). Similar breaks were also observed in bsk^147E single-cell αβ clones (data not shown; quantified in H). (F, G) Representative image of bsk^H15 neuroblast clones showing axon degeneration (F) and overextension (G) phenotypes. Cell body sections were removed from A, C and F to clearly reveal axon projections. Scale bars: 20 μm. (H) Quantification of axon breaks in hep^R75, MKK4^e01458 double, bsk^147E, hep^R75 or MKK4^e01458 or babo⁵² mutants, as indicated. Ten single-cell αβ clones were analyzed for each genotype. Statistical analysis show a significant difference between hep^R75, MKK4^e01458 double or bsk^147E clones when compared to hep^R75 or MKK4^e01458 or babo⁵² mutants (P < 0.05) but no significant difference between hep^R75, MKK4^e01458 double compared to bsk^147E clones, or between hep^R75 and MKK4^e01458 or babo⁵² mutants (P > 0.05). The only exception is in the distal axon section of hep^R75 axons, where a small proportion of degeneration was observed, as reflected in the P-value (0.014). babo⁵² clones were used as the control in the statistical study. (I) Quantification of the bsk^H15 axon phenotypes in neuroblast clones, with a comparison to the null (147E) clones.

Fig. 7

Sustained Bsk levels are essential for axon stability. (A) Bsk activity (as measured by the P-Bsk signal) is detected throughout development and adulthood (black arrow). Yet the axonal phenotypes are observed only at later stages (∼30 h after puparium formation, APF) (blue arrow). To determine the temporal requirements of Bsk-dependent phenotypes, we used the TARGET system to determine whether the induction (+) or suppression (−) of TARGET expression (see schematic in Supplementary Fig. 7A), by transgenic rescue analysis or RNAi at specific stages (larval, pupal or adult) alters the extent of the observed phenotypes. The dashed arrows indicate the extent of the TARGET expression under different temporal and developmental settings. (B–E) Images of bsk^147eneuroblast clones with the Bsk-myc expression under TARGET control. As a control, Bsk TARGET flies raised at 29 °C (GAL4-permissive) throughout exhibited > 90% wild type projections (B). bsk^147eclones with Bsk-myc expression induced at developmental stages L3 (C), and 0h APF (D). (E) bsk^147eclones with Bsk-myc expression restricted only to the adult stage for 10 days post-eclosion. (F) Quantification of these bsk rescue phenotypes. Induced at shorter periods, many flies exhibited β-axon degeneration. n, number of neuroblast clones analyzed. (G–J) Bsk RNAi expressed under TARGET control. Bsk RNAi expression induced at stages L3 (G), 0 h (H), or 96 h APF (I) resulted in axon degeneration (G), and axon overextension phenotypes (H, I). (J) Bsk RNAi restricted to adult stages for 10–20 days post-eclosion showed wild-type projections. With the exception of adult-stage induced flies, all earlier induced flies were dissected as 3-day adults. Scale bars: 20 μm. Green, CD8-GFP. Magenta, Fas2. (K) Quantification of the Bsk TARGET RNAi phenotypes. n, number of brain hemispheres analyzed. In K and L, the β-lobe overextension analysis is also presented as the number of overextended brains over the total number of brains analyzed (italics). Note both protocols do not reflect a strict ‘on’ time at the indicated developmental stages but rely on the ‘on’ kinetics of the TARGET system. In our manipulations, we detected expression from 24 h and robustly at 72 h post-induction (Supplementary Fig. 7B–E; data not shown). Therefore, a period of RNAi and Bsk-myc accumulation is required for effective downregulation of Bsk mRNA transcripts and suppression of bsk-null phenotypes (respectively). (L) Quantification of the Bsk RNAi phenotype using a reverse TARGET protocol, with an early induction followed by a suppression of RNAi transgene expression at L3 or 0 h APF, as indicated. Note the increased representation of axon degeneration over overextension phenotypes at the early phase of Bsk inactivity.

Fig. 8

A graded AP-1 signal mediates Bsk responses. (A–D) MB neurons expressing kay RNAi (A), kay RNAi plus Dcr2 (B), Fbz (C), or Fbz plus Jbz (D) showed axon overextension (A, C) and degeneration (B, D) phenotypes. Copy numbers of expressed transgenes are indicated in parenthesis. Progeny derived from Fbz and Jbz crosses were raised at 29 °C to increase the possibility of detecting any axonal phenotypes due to misexpression. Scale bars: 20 μm. Green, CD8-GFP. Magenta, Fas2. (E) Quantification of these phenotypes. In E and F, also shown is the β-lobe overextension analysis expressed as the number of overextended brains over the total number of brains analyzed (italics). (F) Genetic interaction assay using Bsk RNAi (line 34138) with dominant-negative AP-1. Bsk RNAi expressing flies were grown at 29 °C, in presence of one copy of Fbz or two copies of Jbz, showed an enhancement in axonal defects. As controls, Bsk RNAi flies were expressed with CD8GFP alone. Single copy expression of Fbz (line 5 or 7) or two copies of Jbz (lines 1 and 10) (E) did not result in gross axon defects. n, number of hemispheres analyzed.