Deciphering the roles of subcellular distribution and interactions involving the MEF2 binding region, the ankyrin repeat binding motif and the catalytic site of HDAC4 in Drosophila neuronal morphogenesis

Background Dysregulation of nucleocytoplasmic shuttling of histone deacetylase 4 (HDAC4) is associated with several neurodevelopmental and neurodegenerative disorders. Consequently, understanding the roles of nuclear and cytoplasmic HDAC4 along with the mechanisms that regulate nuclear entry and exit is an area of concerted effort. Efficient nuclear entry is dependent on binding of the transcription factor MEF2, as mutations in the MEF2 binding region result in cytoplasmic accumulation of HDAC4. It is well established that nuclear exit and cytoplasmic retention are dependent on 14–3-3-binding, and mutations that affect binding are widely used to induce nuclear accumulation of HDAC4. While regulation of HDAC4 shuttling is clearly important, there is a gap in understanding of how the nuclear and cytoplasmic distribution of HDAC4 impacts its function. Furthermore, it is unclear whether other features of the protein including the catalytic site, the MEF2-binding region and/or the ankyrin repeat binding motif influence the distribution and/or activity of HDAC4 in neurons. Since HDAC4 functions are conserved in Drosophila, and increased nuclear accumulation of HDAC4 also results in impaired neurodevelopment, we used Drosophila as a genetic model for investigation of HDAC4 function. Results Here we have generated a series of mutants for functional dissection of HDAC4 via in-depth examination of the resulting subcellular distribution and nuclear aggregation, and correlate these with developmental phenotypes resulting from their expression in well-established models of neuronal morphogenesis of the Drosophila mushroom body and eye. We found that in the mushroom body, forced sequestration of HDAC4 in the nucleus or the cytoplasm resulted in defects in axon morphogenesis. The actions of HDAC4 that resulted in impaired development were dependent on the MEF2 binding region, modulated by the ankyrin repeat binding motif, and largely independent of an intact catalytic site. In contrast, disruption to eye development was largely independent of MEF2 binding but mutation of the catalytic site significantly reduced the phenotype, indicating that HDAC4 acts in a neuronal-subtype-specific manner. Conclusions We found that the impairments to mushroom body and eye development resulting from nuclear accumulation of HDAC4 were exacerbated by mutation of the ankyrin repeat binding motif, whereas there was a differing requirement for the MEF2 binding site and an intact catalytic site. It will be of importance to determine the binding partners of HDAC4 in nuclear aggregates and in the cytoplasm of these tissues to further understand its mechanisms of action. Supplementary Information The online version contains supplementary material available at 10.1186/s12915-023-01800-1.

, second row).These data demonstrate that reduced expression of HDAC4 also result in defects, thus close to wild-type levels are required for correct mushroom body development.This was confirmed with an independent protein trap (HDAC4::EYFP, in which EYFP flanked by a splice acceptor and donor is inserted into the second intron of HDAC4) (64), which we have previously characterized (40) (Table S3).

Figure S2. Optimization of deGradFP experimental conditions for replacement of endogenous HDAC4
with HDAC4 WT .elav-GAL4;HDAC4:EGFP males were crossed to females of the following genotypes: UAS-tubPGAL80ts;OK107-GAL4 (control), UAS-tubPGAL80ts;UAS-deGradFP;OK107-GAL4 (HDAC4 KD ) and UAS-tubPGAL80ts;UAS-deGradFP/UAS-HDAC4 WT ;OK107-GAL4 (HDAC4 KD + HDAC4 WT ).Flies of each set of genotypes were raised at the four indicated temperatures to modulate the level of expression of HDAC4 WT and deGradFP to determine an expression level in which HDAC4 WT is expressed at a similar level to endogenous HDAC4, which would rescue the β lobe fusion phenotype.The percentage of brains of progeny displaying severe β lobe fusion is shown.
When HDAC4 WT was co-expressed with deGradFP at 30°C, 89% of brains displayed severe b lobe fusion similarly to that of HDAC4 WT without deGradFP, indicating that the overall level of HDAC4 was still well above endogenous HDAC4 and thus displaying overexpressioninduced phenotypes (Table S2A, third row).We therefore modulated the level of GAL80ts via raising the flies at several different temperatures to determine the optimal level of expression which endogenous HDAC4 was knocked down and replaced with the appropriate level of expression of HDAC4 WT as to not induce a significant overexpression phenotype (Fig S2, Table S2B-D).At 27°C, knockdown of HDAC4 was sufficient to induce severe b lobe fusion in 21% of brains, but co-expression of HDAC4 WT reduced severe b lobe fusion to 8% of brains, therefore the overall level of expression was close that of endogenous HDAC4.The percentage of brains displaying a single phenotype of severe, moderate or minor β lobe fusion, or thin/absent lobe(s) is shown.As a measure of severity of the phenotype, the percentage of brains displaying both β lobe fusion and thin or absent lobes is also shown.The total percentage of brains displaying β lobe fusion is also calculated by combining minor, moderate and severe β lobe fusion and the brains displaying both β-lobe fusion and thinner lobe(s).
Efficient knockdown and expression of HDAC4 WT and mutant transgene expression was confirmed by immunohistochemistry (Fig S3).

Figure S3. Verification of HDAC4::EGFP knockdown and HDAC4 transgene expression.
A-F. GFP and Myc immunohistochemistry on confocal stacks through the Kenyon cell layer of flies expressing deGradFP and the indicated Myc-tagged HDAC4 transgene.Flies of the indicated genotypes were raised at varying temperatures for gradual linear increase in transgene expression.Reduced expression of HDAC4::EGFP on increase in temperature is observed as it is knocked down with deGradFP, while HDAC4-Myc is induced.The gain was kept the same between all samples to enable direct comparison between temperatures.The relative intensity of staining between Myc and GFP cannot be compared directly, and the GFP epitope is internal within endogenous HDAC4 and not as accessible as the Myc epitope tag on HDAC4-Myc, thus the staining is fainter.The graph displays the relative EGFP intensity of HDAC4::EGFP raised at 18°C (no induction of deGradFP) vs 30°C (full induction), confirming the a significant reduction in endogenous HDAC4 expression by deGradFP.n=4 to 6 brains/genotype.A. HDAC4

Table S1 . Frequency of mushroom body defects resulting from expression of HDAC4 WT and mutants with OK107-GAL4;tub-GAL80ts
Flies were raised at 30°C throughout development.The percentage of brains displaying a single phenotype of severe, moderate or minor β lobe fusion, or thin/absent lobe(s) is shown.No brains displayed both β lobe fusion and thin or absent lobes.The total percentage of brains displaying β lobe fusion is also shown.Knockdown of endogenous HDAC4 was induced via the deGradFP genetic tool; flies were generated that carried one copy each of UAS-Nslmb-vhhGFP4 (hereafter referred to as deGradFP), tubP-GAL80ts, OK107-GAL4 as well as HDAC4::EGFP.When raised at 30°C to induce expression of deGradFP, this resulted in defects in 45% of brains, with 28% displaying severe b lobe fusion Table

Table S2 . Frequency of mushroom body defects resulting from expression of HDAC4 WT in the mushroom body in an HDAC4-depleted background.
HDAC4 WT .The percentage of brains displaying each of the defects is shown.In the absence of deGradFP, expression of HDAC4 WT and the mutants at 27°C resulted in the same overall pattern as previously seen in Fig 3C, with severe mushroom body defects in all except HDAC4 DMEF2 expressing brains (Fig 3D, left graph, Table 2).However in the reduced endogenous HDAC4 background (with HDAC4 knockdown and transgene expression confirmed in Fig S2B-F), only HDAC4 3SA and HDAC4 DANK induced a significant increase in mushroom body defects compared to HDAC4 WT (Fig 3D right graph, Table3).