Drosophila Hrp48 Is Required for Mushroom Body Axon Growth, Branching and Guidance

RNA binding proteins assemble on mRNAs to control every single step of their life cycle, from nuclear splicing to cytoplasmic localization, stabilization or translation. Consistent with an essential role of RNA binding proteins in neuronal maturation and function, mutations in this class of proteins, in particular in members of the hnRNP family, have been associated with neurological diseases. To date, however, the physiological function of hnRNPs during in vivo neuronal development has remained poorly explored. Here, we have investigated the role of Drosophila Hrp48, a fly homologue of mammalian hnRNP A2/B1, during central nervous system development. Using a combination of mutant conditions, we showed that hrp48 is required for the formation, growth and guidance of axonal branches in Mushroom Body neurons. Furthermore, our results revealed that hrp48 inactivation induces an overextension of Mushroom Body dorsal axonal branches, with a significantly higher penetrance in females than in males. Finally, as demonstrated by immunolocalization studies, Hrp48 is confined to Mushroom Body neuron cell bodies, where it accumulates in the cytoplasm from larval stages to adulthood. Altogether, our data provide evidence for a crucial in vivo role of the hnRNP Hrp48 in multiple aspects of axon guidance and branching during nervous system development. They also indicate cryptic sex differences in the development of sexually non-dimorphic neuronal structures.


Introduction
Developing neurons extend neuronal processes-dendrites and axons-that have to navigate through a complex environment to find their targets, and establish synaptic connections.As revealed recently, neuronal cells heavily rely on post-transcriptional regulatory mechanisms such as alternative splicing, mRNA transport or precise translational control to finely tune gene expression in space and time, and regulate critical processes underlying the assembly of neuronal networks, from neurite guidance and branching to synaptogenesis [1].Key players of these regulatory mechanisms are RNA binding proteins, which recognize specific signatures present in subsets of target mRNAs, and control their processing and cytoplasmic fate [2][3][4].
Consistent with a crucial role of RNA binding proteins in neural development and function, mutations in a number of these proteins have been linked to neurological disorders [5,6].Heterogeneous nuclear ribonucleoproteins (hnRNPs) constitute the largest family of RNA binding proteins that comprises 20 major conserved proteins designated hnRNP A to hnRNP U [7,8].Recently, pathogenic mutations in hnRNP A2/B1 and hnRNP A1 were found in families with inherited neurodegeneration syndroms [9].Furthermore, members of the hnRNP A2/B1 family of proteins were identified as genetic modifiers in a fly model of Fragile X associated tremor ataxia syndrome (FXTAS) [10,11].Surprisingly, the physiological function of hnRNP A/B proteins during central nervous system (CNS) maturation has so far remained unclear.
To characterize the in vivo role of hnRNP A/B proteins during CNS maturation, we have explored the function of Drosophila Hrp48 (also known as Hrb27C), a fly homologue of mammalian hnRNP A2/B1 [12], using Mushroom Bodies (MBs) as a model system.MBs are symmetrical structures located in the central brain.They are essential for higher order function including olfactory learning and memory [13], and have a precisely described and stereotypic development [14].Each MB is composed in adult of about 2,000 neurons projecting their axons ventrally toward the anterior surface of the brain, where axonal branches segregate into dorsal and medial terminal lobes.MB neurons have been subdivided into three main populations (αβ, α'β' and γ neurons) based on their birth order, the markers they express, and the characteristic morphology of their axonal projections in the lobes [14][15][16].While each αβ and α'β' neuron bifurcates and sends one axon branch to the dorsal lobe and one to the medial lobe, adult γ neurons only project to the medial lobe.
Here, we show that Hrp48 controls multiple aspects of MB axonal development, including axon growth, branching and guidance.We have also discovered that Hrp48 is required to prevent the overextension of dorsal αβ and α'β' axonal branches, and that the penetrance of this phenotype is much stronger in females than in males.Finally, our results indicate that Hrp48 is restricted to MB cell bodies in vivo, and that it is thus unlikely to mediate axonal transport and local translation of target mRNAs in this system.Together, these results highlight the biological importance of post-transcriptional regulation during CNS development.Furthermore, they reveal cryptic sex-specific differences in the regulation of MB neuron development.

Image analysis
Phenotypic classes (symmetric, asymmetric, truncated and ectopic projections) were defined based on FasciclinII staining.αβ projections were classified as asymmetric when the ratio of α lobe to β lobe thicknesses was significantly different from control conditions (ie when its gap to the mean control ratio was larger than 2 times the standard deviation).Lobe thicknesses were measured on z projections of stacks of confocal images using the Zeiss LSM Image Browser or the Leica LAS AF Lite softwares.As α lobe diameter varies along the dorsoventral axis, we focused on α lobe central most-region and defined α lobe width as the largest diameter in this region.α or β lobes were classified as truncated when they did not fully extend or did not develop at all.Ectopic projection defects were counted separately, as they cannot be quantified on truncated lobes and can be detected on asymmetric lobes.
The number of MB neurons was counted on single confocal sections of mCD8-GFP-expressing neurons.Sections to be analysed were chosen such that the cell bodies of all peripheral-most MB neurons were visible.Non-properly oriented brains were not considered for the analysis.

Results
Hrp48 is required for the growth, guidance and branching of MB αβ and α'β' axons Hrp48 is an essential gene whose function is required for embryonic development [21].To investigate the function of hrp48 in MB morphogenesis, we thus first used two semilethal transheterozygous combinations of hrp48 mutant alleles-hrp48 k10413 /hrp48 02647 and hrp48 k10413 / hrp48 k16203 [21]-and two inducible RNAi lines targeting different regions of hrp48 coding sequence-#16041 and #101555 -(Fig 1A).These mutant conditions were associated with a partial, but significant decrease in Hrp48 levels, as visualized by Western-Blot analysis on hrp48 k10413 /hrp48 02647 and hrp48 k10413 /hrp48 k16203 adult brain lysates ( To visualize the morphology of adult MB lobes, we used FasciclinII (FasII) antibodies that weakly label γ axons and strongly label αβ axons.Furthermore, a membrane-tagged form of GFP (mCD8-GFP) was introduced in RNAi contexts to visualize the axonal projections of the entire population of MB neurons.As shown in Most strikingly, defects in MB αβ axon projection patterns were observed in hrp48 mutants.As revealed by anti-FasII staining, αβ neurons normally produce two branches projecting  While the truncated lobe phenotype reflects a role of Hrp48 in axon growth, the uneven distribution of αβ processes between the dorsal and the medial lobes may either result from a failure of αβ neurons to form two independent branches (branching defects), or from defects in segregating axonal branches in divergent directions (guidance defects).To discriminate between these hypotheses and precisely visualize the morphology of both αβ and α'β' hrp48 mutant neurons, we generated single αβ or α'β' neurons homozygous for the lethal mutation 7E7-18 (Fig 1A, [18]) using the MARCM technique [14,20].data not shown).Furthermore, hrp48 7E7-18 αβ or α'β' mutant neurons lacking a branch were observed (Fig 3C and 3D and data not shown), indicating that hrp48 function is required both for the formation of main axonal branches, and for their guidance and growth in their respective lobes.The penetrance of these defects was however low (9/63 for α'β' neurons and 2/26 for αβ neurons; see S1 Table for a detailed distribution of phenotypic classes), which may reflect the fact that neurons do not regulate branch formation and guidance cell-autonomously, bur rather collectively, at the population level [22].
Altogether, these results indicate that hrp48 has pleiotropic functions and is required for axonal growth, guidance and branching of MB α'β' and αβ neurons.Notably, hrp48 function appears to be cell-specific, as hrp48 mutant MB γ lobes were indistinguishable from wild-type γ lobes (Fig 2).

Hrp48 prevents overextension of MB dorsal axonal branches
Our analysis of hrp48 mutant phenotypes revealed that hrp48 MBs show an additional phenotype unrelated to defective branching or growth inhibition, and characterized by an abnormal overextension of dorsal lobes.This phenotype was observed in hrp48 semilethal transheterozygous conditions ( To confirm these phenotypes using independent lethal mutations, we generated MB neuroblast clones labelled by mCD8-GFP and composed of neurons homozygous for the hrp48 7E7-18 or hrp48 k02814 mutant alleles.As shown in Fig 5A -5C, ectopic extension of dorsal bundles was observed in about 55% and 70% of hrp48 7E7-18 or hrp48 k02814 mutant clones respectively.These defects were significantly suppressed when expressing a wild-type copy of hrp48 in mutant neurons (Fig 5C).However, they were rarely observed in single cell-clones (2/63 hrp48 7E7-18 α'β' clones and 1/26 hrp48 7E7-18 αβ clones; see S1 Table ), suggesting that they reflect a population-autonomous rather than a strictly cell-autonomous function of hrp48.
Altogether, these results indicate that Hrp48 is required to prevent the overextension of MB dorsal axonal branches.

The penetrance of hrp48 axon overextension phenotypes is significantly higher in females than in males
Inactivation of hrp48 function in Drosophila imaginal wing discs was shown to induce aberrant wing phenotypes, with a higher penetrance in females than in males [23].To test whether the penetrance of hrp48 MB axon projection defects was sex-dependent, we compared the proportion of MBs with defective axonal projection patterns in males and females expressing hrp48 RNAi construct under the control of the OK107-Gal4 driver.While the proportion of MBs with asymmetric or truncated α or β lobes was only slightly higher in hrp48 RNAi females than in hrp48 RNAi males (S3 Fig), the proportion of MBs with axon overextension defects was much higher in hrp48 RNAi mutant females than in hrp48 RNAi mutant males (Fig 6A).Notably, such a difference between males and females was observed for both UAS-RNAi lines, and upon generation of neuroblast clones mutant for the lethal hrp48 k02814 allele (Figs 1A and 6B).
To test whether the observed sex-specific response was due to differences in basal levels of Hrp48 in males vs females, we quantitatively compared Hrp48 protein levels in male and female brains.As shown in Fig 6C, no significant difference could be observed between sexes.Thus, these results reveal a differential sensitivity of male and female MB neurons to the inactivation of hrp48, and cryptic differences between male and female MB development.

Hrp48 accumulates in the cytoplasm of neuronal cell bodies at steady state
Members of the hnRNP family are multi-functional proteins that are dynamically shuttling between the nucleus and the cytoplasm and regulate both nuclear and cytoplasmic events, ranging from RNA transcription and splicing to RNA stability, transport or translational control [7].In cultured mammalian oligodendrocytes and neurons, hnRNP A2 localizes to cellular processes and regulates the transport and the translation of target mRNAs important for different aspects of neural development [24][25][26][27].In the polarized Drosophila oocyte, Hrp48 was shown to associate with the asymmetrically localized gurken and oskar mRNAs to promote their targeting to the antero-dorsal and posterior poles respectively, and to prevent their premature translation [17,18,28].
To determine the subcellular distribution of Hrp48 in MB neurons, we stained Drosophila brains with anti-Hrp48 antibodies.A cytoplasmic signal, whose intensity significantly decreased upon hrp48 downregulation, was observed in adult and larval MB cell bodies (Figs 1C and 7A).Such a cytoplasmic accumulation was also observed when expressing a Flagtagged version of Hrp48 in MB neurons using the OK107-Gal4 line (Fig 7B  Thus, these results indicate that Hrp48 exhibits a predominantly cytoplasmic localization at steady-state.Furthermore, Hrp48 is confined to the cell bodies of MB neurons, and appears not to be targeted to axons.

Hrp48 controls multiple aspects of axonal development
Members of the hnRNP A/B family of proteins have been shown to regulate the spatio-temporal expression pattern of transcripts important for neural development [25][26][27], yet the role of these proteins during nervous system maturation in vivo has remained elusive.Here, we have shown that the hnRNP A/B family member Hrp48 controls multiple aspects of axon morphogenesis in Drosophila brains.First, MB α'β' and αβ neurons mutant for hrp48 frequently exhibit shorter axons, suggesting that Hrp48 promotes the growth of axonal branches in these populations.Second, as revealed by our clonal analysis, α'β' and αβ neurons lacking a dorsal or a Such a pleiotropic phenotype likely results from the misregulation of various mRNA targets.Although hnRNP A/B family members such as hnRNP A2 were shown to regulate the transport and the translation of target mRNAs to neurites [24][25][26][27], our immunolocalization analysis rather points to a role of Hrp48 in MB cell bodies.Given the known function of hnRNP proteins [7], this role may include translational control of target mRNAs and/or regulation of their alternative splicing.Consistent with this latter hypothesis, a systematic microarray-based analysis has revealed that alternative splicing of more than 300 transcripts is under the control of Hrp48 in Drosophila S2 cells [29].Remarkably, genes involved in neuronal development and cell morphogenesis were shown to be significantly overrepresented in this population of hrp48affected genes.Among interesting hrp48 target genes identified in this study are enable, trio and short-stop, which encode regulators of the cell cytoskeleton known for their role in axon growth [30,31], but also slit, frazzled and sema-2a, which encode proteins shown to regulate axon guidance in various systems [32,33].nrg, which encodes an L1-type cell adhesion molecule that regulates MB αβ axon branching and outgrowth [34], and dscam, which encodes a transmembrane protein promoting the segregation of MB αβ sister branches [35], are two other hrp48 targets whose misregulation may explain some of the axonal defects observed in hrp48 mutant contexts.Notably, hrp48 loss-of-function does not affect the development of MB γ neurons, revealing that it is probably not regulating core components of the axon growth and branching machineries, but rather cell-specific regulators of axon morphogenesis.
Sex-specific differences in MB development MB neurons are not sexually dimorphic as γ, α'β' and αβ neurons exhibit similar axonal projection patterns in males and females.Our study, however, revealed cryptic differences between male and female MB development as downregulation of hrp48 induces much stronger axon overextension defects in females than in males.Such a sex-dependent sensitivity to the loss of hrp48 function has already been reported during wing development, where Hrp48 is required to repress the expression of the female determinant sex-lethal (sxl) [23].As shown in this study, Sxl accumulates specifically in female hrp48 mutant imaginal disc cells and inhibits Notch signalling, thus interfering with wing growth and patterning in a sex-dependent manner.In the proposed model, Hrp48 would prevent Sxl, the female-specific determinant underlying sexual dimorphism, from disrupting the development of monomorphic tissues by limiting its expression.In MBs, however, hrp48 mutant phenotypes do not seem to result from a misregulation of sxl, as i) no increase in Sxl levels was observed in hrp48 mutant neurons, and ii) axonal projection defects were not phenocopied upon overexpression of Sxl (data not shown).We thus propose an alternative model in which the observed sex-specific penetrance of axonal projection defects would result from differential expression levels of Hrp48 target mRNAs.
Fig 2A-2F, a general decrease in MB lobe volume associated with a decrease in the total number of MB neurons (S1A Fig) was observed upon hrp48 downregulation, suggesting that Hrp48 may play a role in the regulation of MB neuroblast proliferation.