Mxc, a Drosophila homolog of mental retardation-associated gene NPAT, maintains neural stem cell fate

Background Mental retardation is a complex neurodevelopmental disorder. NPAT, a component of the histone locus body (HLB), has been implicated as a candidate gene for mental retardation, with a mechanism yet to be elucidated. Results We identified that mxc, the Drosophila ortholog of NPAT, is required for the development of nervous system. Knockdown of mxc resulted in a massive loss of neurons and locomotion dysfunction in adult flies. In the mxc mutant or RNAi knockdown larval brains, the neuroblast (NB, also known as neural stem cell) cell fate is prematurely terminated and its proliferation potential is impeded concurrent with the blocking of the differentiation process of ganglion mother cells (GMCs). A reduction of transcription levels of histone genes was shown in mxc knockdown larval brains, accompanied by DNA double-strand breaks (DSBs). The subsidence of histone transcription levels leads to prematurely termination of NB cell fate and blockage of the GMC differentiation process. Our data also show that the increase in autophagy induced by mxc knockdown in NBs could be a defense mechanism in response to abnormal HLB assembly and premature termination of NB cell fate. Conclusions Our study demonstrate that Mxc plays a critical role in maintaining neural stem cell fate and GMC differentiation in the Drosophila larval brain. This discovery may shed light on the understanding of the pathogenesis of NPAT-related mental retardation in humans. Supplementary information The online version contains supplementary material available at 10.1186/s13578-022-00820-8.


Background
Mental retardation, also known as intellectual disability (ID), is a common neurodevelopmental disorder which has been estimated to be present in 1% of the global population [1]. The etiology of mental retardation is complicated with both external and genetic factors often contributing to this neurological disorder [2]. Chromosomal microarray analysis (CMA) studies, where the copy number variants (CNVs) containing the NPAT locus have been categorized as the pathogenic events for mental retardation, implicate NPAT as a candidate disease gene [3,4]. However, no point mutations within NPAT gene have yet been reported in mental retardation cases and any potentially related pathogenic mechanisms remain to be elucidated.
Drosophila is an ideal model for studying central nervous system (CNS) development and deciphering causes of related diseases. In Drosophila, multi sex combs (mxc) is the ortholog of NPAT [5]. NPAT/ Mxc is a key component of histone locus body (HLB) [6][7][8]. Studies have shown that NPAT/Mxc acts as the scaffold protein of the HLB required for HLB assembly and transcription of canonical histone genes [5]. The N-terminal of NPAT/Mxc is crucial for self-interaction, while the C-terminal is required for recruiting other members of the HLB [9,10]. NPAT/Mxc is phosphorylated by cyclin E/Cdk2, such phosphorylation being a key factor in the formation of phase separation allowing the formation of HLB to be precisely and dynamically regulated [11,12]. Several studies on NPAT/mxc mutations have been published with the mutations of NPAT/mxc leads to the failure of HLB formation and in the block ageing of histone transcription in cultured cells [10,11]. The chromosome instability caused by mxc mutations affects Drosophila male meiosis [13]. In Drosophila larval hematopoietic tissue, depletion of mxc leads to tumor formation [14,15]. Based on these reports, NPAT /mxc appears to prevent chromosome instability and acts as a tumor suppressor gene. How a "tumor suppressor" NPAT, when its function is defective, leads to human mental retardation remains a key question. Our study focuses on the specific role of NPAT/Mxc in the development of Drosophila CNS, and reveals its possible mechanisms related to mental retardation.
Drosophila CNS is derived from neuroblasts (NBs), also known as neural stem cells [16]. The NB divide asymmetrically producing a self-renewing NB and a smaller ganglion mother cell (GMC). The GMCs further undergo differentiation and divide terminally, giving rise to two neurons [17,18]. At the pupal stage, all NBs terminate their cell fate and exit the cell cycle [19,20]. No mitotic NBs remain detectable in the central brain or ventral nerve cord of adult flies [21]. Deadpan (Dpn) and Prospero (Pros) are two important transcription factors that regulate the stemness and differentiation of NBs, and are commonly used as markers to identify NBs and their progeny [22][23][24]. Dpn is expressed in all NBs and mature INPs, but not in immature INPs or GMCs [25]. Ectopic expression of Dpn in immature INPs leads to dedifferentiation and tumor formation [23,26]. Pros, as a cell fate determinant, is expressed in all NBs but only localized at the basal cortex during mitosis in the larval central brain. At telophase, Pros is exclusively segregated into the GMCs [27,28]. Soon after the formation of GMCs, Pros enters the nucleus and acts as a transcription factor that inhibits stemness and promotes differentiation [24,29]. At the early pupal stage, during which all CNS neurons are developed, Pros starts to accumulate in the nucleus of NBs. These NBs with nuclear Pros divide symmetrically to generate two equal sized daughters and terminate the NB cell fate [30,31].
In the present study, we identify that Mxc/NPAT functions to maintain the proliferation of NBs and their progeny in Drosophila larval brains. The lack of Mxc leads to decreased transcription levels of histones and increased DNA double strand breaks (DSBs), resulting in premature termination of NB cell fate and impaired NB selfrenew potential. Knocking down mxc also inhibits the differentiation of GMCs. These abnormalities drastically reduce the numbers of neurons in the central brain, eventually leading to neurological defects. Our data also show that in the absence of Mxc, autophagy is elevated and seems to protect against the formation of Dpn + (differentiation-halted) GMCs. These data shed a light on the possible pathogenesis of human mental retardation disease caused by NPAT mutations.

Mxc deficiency leads to a massive loss of neurons and locomotion dysfunction in adult flies
To explore the potential functions of Mxc in CNS development, we employed a NB specific RNAi knockdown approach. The mxc RNAi lines were driven by wor-GAL4. mxc knockdown flies survived till the early adult stage, but all stayed on the bottom of the vials after eclosion without climbing ability and exhibited akinesia. These flies then died within the first one or two days of adulthood (n = 107; Fig. 1B). Overexpression of Mxc in mxc knockdown flies partially rescued this phenotype, in which about one third of the flies survived beyond two days (34.58%, n = 106; Fig. 1B).
We dissected brains from the adult flies within one day of eclosion and stained with anti-Elav, a pan-neuronal marker. Compared to the wt (Fig. 1A). The mxc knockdown flies showed a massive loss of neurons in the central brain (arrowheads, Fig. 1D). The exogenous overexpression of Mxc in larval NBs largely rescued this phenotype (Fig. 1E). It is interesting to note that no obvious abnormality of brain development was observed when Mxc was overexpressed in larval brain NBs of wt flies. These observations indicate that Mxc plays a vital role in the CNS development. mxc knockdown in larval NBs causes the loss of neurons in the adult brains, leading to locomotion dysfunction and early lethality of adult flies.
To further evaluate the important role of Mxc in CNS development, we examined the genetic mutant of mxc. mxc 16a − 1 is a mutant line of mxc (Obtained from BDSC), containing a deletion of 4 bp (TTCG) at residues F1823 of Mxc and resulting in a frameshift at the C-terminal [5]. Hemizygotes of mxc 16a − 1 die at late larval or early pupal stage [32]. NB specific overexpression of exogenous Mxc in hemizygotes of mxc 16a − 1 prolonged the survival time of the mutant animals to the late pupal stage (80.26%, n = 101) where a small percentage (10.82%, n = 101) of the flies even survived to adulthood without obvious mobility problems. This rescue data agrees well with the mxc RNAi result and suggest that Mxc is critical for CNS development.

Knockdown of mxc results in premature localization of Pros in the nucleus of NBs
The massive loss of neurons in adult central brains could be due to defects in NBs and GMCs in larval brains. This prompted us to investigate the NB cell fate maintenance in mxc knockdown larval brains. In NBs, nuclear localization of Pros is recognized as a signal for the termination of stem cell fate [20,30,31].We observed that NB specific mxc knockdown with wor-GAL4 led to a premature accumulation of Pros in nuclei of the NBs of the third-instar larval central brains (arrowheads; Fig. 2B). To ensure the observed phenotype was not an off-target effect, a second mxc-RNAi line (THU5844, obtained from Tsinghua Drosophila Center) was used and the rescue experiment was conducted as well with a mxc overexpressed line ( Fig. 2A-F). The efficiency of RNAi knockdown was confirmed using Western blotting analysis with MPM-2, an antibody recognizing cyclin E/Cdk2dependent phosphorylation sites within Mxc [5] (Fig. 2I). The brain samples were double-labeled with anti-Dpn, a stem cell marker [23,24] and anti-Pros. In wt third-instar larval brain NBs, Pros was not detected in the nuclei of NBs ( Fig. 2A). In contrast, we observed nuclear Pros in NBs in both mxc RNAi lines (THU0893: 27.51%, n = 740; THU5844: 24.84%, n = 69; Fig. 2B, C, F). Overexpression of exogenous Mxc in mxc knockdown background partially rescued the phenotype (16.67%, n = 108; Fig. 2D, F). The overexpression of Mxc alone in wt NBs did not result in any phenotypes (Fig. 2E).

Mxc maintains the NB number and cell fate in larval central brains
We quantified the NB numbers in the brain lobes at the 3rd instar larval stage (120 h AEL) for both wt and mxc knockdown animals (Fig. 3A). In the wt brain lobes, the average number of total Dpn + NBs per lobe at 120 h AEL was 72.0 (n = 5 lobes), while in mxc knockdown lines, the average number of NBs was reduced to 38.5 (n = 6 lobes; Fig. 3A). In this way it appears that about half of the NBs in mxc knockdown brains were missing at the late thirdinstar larval stage. To exclude the possibility that apoptosis was the cause for NB number changes, wt and mxc knockdown larval brains were stained with anti-Caspase 3. No obvious signal was identified in either wt or mxc knockdown NBs (Additional file 1: Fig. S1). The cell sizes of the NBs in the mxc RNAi lines had also remained unchanged from those of the wt (Additional file 1: Fig.  S2).
To determine the time window when the observed premature NB cell fate termination had occurred, we used nuclear Pros as a marker to quantify the ratio of NBs with nuclear Pros at different developmental stages (Fig. 3B). At the second instar larval stage (48 and 72 h AEL) nuclear Pros remained undetected (48 h AEL: 0.71%, n = 239; 72 h AEL: 0.94%, n = 112; Fig. 3B). The nuclear Pros phenotype was first observed at the early third-instar larval stage (96 h AEL; 11.00%, n = 310; Fig. 3B). The average numbers of total NBs per lobe at 96 h AEL was also quantified. The number of NBs in mxc RNAi brains was slightly lower than those in the controls (wt:71.6, n = 5 lobes; mxc RNAi:65.8, n = 5 lobes; Fig. 3A). These data indicate that the NBs in mxc knockdown brains begin premature termination at the early third-instar larval stage and shows about half of the NBs missing by the late third-instar larval stage (120 h AEL).
We next performed FLP-out mosaic analysis on mxc RNAi lines [34], to evaluate the effects of mxc knockdown on the proliferation of NBs. We focused on Type I NB lineages. As shown in Fig. 3C-E, GFP-labeled clonal lineages containing mxc knockdown NBs were significantly smaller than those of the wt (wt: 42.40 cells, n = 5 clones; mxc RNAi: 7.18 cells, n = 11 clones). This indicates that, over a given time, cell cycling becomes limited in mxc knockdown NBs. In addition, the six-fold cell number difference between the wt and the mxc knockdown NB lineages also suggests that the cell cycles of GMCs have either become elongated or arrested. To further confirm this, wt and mxc knockdown larval brains were stained with anti-PH3 EdU signals in mxc knockdown NBs (Fig. 3I), indicating that the lack of Mxc had led to slower rate of cell cycling. These data indicate that Mxc is required for the maintenance of NB cell fate and proliferation in Drosophila larval brains.

Absence of Mxc results in the ectopic expression of Dpn in GMCs
In the central brains of the wt larvae, the stem cell marker Dpn is expressed in NBs and mature INPs, but not in GMCs [22,23,25]. In addition to normal large sized NBs, we observed many smaller Dpn positive (Dpn + ) cells in mxc knockdown larval brains (Fig. 4A,  B). The sizes of these extra Dpn + cells were similar to those of GMCs. To estimate total numbers of Dpn + cells in the central brain, we quantified all Dpn + cells in each brain lobe. An average of 129.0 Dpn + cells/ lobe (n = 6 lobes) in wt brains and 276.5 Dpn + cells/ lobe (n = 4 lobes) in mxc knockdown brains were observed (Fig. 4C). This more than two-fold change in Dpn + cell number suggest that more cells are in an undifferentiated state in the absence of Mxc. Since the total number of NBs decreased in mxc knockdown brains (Fig. 3A), it is most likely that GMC-like neural precursor cells with Dpn expression contribute to the increased Dpn + populations. As the total number of GMCs is much larger than that of INPs, we focused on the possibility that in the absence of Mxc, large numbers of GMC-like cells, if not all, displayed altered cell fates and ectopically expressed Dpn. We then established mxc RNAi FLP-out clones to analyze the pattern of Dpn in Type I NB lineages ( Fig. 4D-E'). In the wt Type I NB lineages, the NB was the only Dpn + cell (Fig. 4D). However, as judged by cell sizes, in mxc knockdown lineages not just NBs but also GMCs were Dpn positive (Fig. 4E). Similar experiments were carried with genetic mutants (Fig. 4F-G'). In mxc 22a − 6 Type I NB MARCM clones, the GMCs were also Dpn + (Fig. 4G). This was identical to the RNAi data (Fig. 4E). These Dpn + GMC-like cells became the major populations of Dpn + cells in the mxc knockdown larval brains. We concluded that Mxc acts to suppress the expression of Dpn in GMCs and promotes GMC differentiation.

The nuclear localization of Pros in NBs and the differentiation of GMCs depend on Mxc-mediated HLB assembly and histone gene transcription
Mxc is a key component of the HLB (Additional file 1: Fig. S3). Using MPM-2 and anti-Mute antibodies, we observed that Mxc was colocalized with another component of HLB, Mute [5,10,35,36], in the nucleus of wt NBs, forming unambiguous nuclear foci (arrowhead; Additional file 1: Fig. S4A). The localization of Mute in HLB was Mxc-dependent [5]. In mxc knockdown NBs no MPM-2 puncta were detected and the Mute staining pattern was less condensed and with increased numbers of multiple smaller puncta observed compared to the wt (Additional file 1: Fig. S4B-B"). Overexpression of Mxc in mxc RNAi NBs reversed this phenotype; MPM-2 and Mute restored colocalization at nuclear foci (arrowheads; Additional file 1: Fig. S4C-C"). These data indicate that, in the absence of Mxc, Mute is less condensed and exhibits multiple dot-shaped distributions. Additional experiments were carried out with mxc 16a − 1 mutants and mxc 22a − 6 MARCM clones. Similar results were observed in both two mutant lines (Additional file 1: Fig. S4D-G"). These results confirm that knockdown mxc in larval brain NBs leads to less organized Mute distribution which ultimately impairs HLB assembly.
We considered whether the absence of Mxc would affect the transcription of histone genes in third-instar larval brains. An elav-GAL4 line was used to drive mxc RNAi in the larval brains. qPCR analyses showed that the transcription levels of five canonical histone genes were all significantly decreased in both mxc knockdown (Fig. 5A) and mxc 16a − 1 mutant larval brains (Fig. 5B). Western blot also showed that His3 protein levels had decreased in the mxc knockdown larval brains (Fig. 5C), which was consistent with our qPCR results.
To clarify that the reduced histone gene expression is responsible for the mxc knockdown-induced NB phenotypes in larval brain, we performed knockdown of either His3 or His4 in larval brains and characterized the (See figure on next page.) Fig. 3 Mxc is required for NB proliferation and cell fate maintenance.A Statistical analysis of the numbers of NBs at different developmental stages (96 and 120 hours AEL) in wt and mxc RNAi brain lobes. Dpn + cells larger than 8µm in diameter were considered NBs [25]. The data are plotted as mean ± SD. ****p < 0.0001 using a Student's t test, p = 2.555E-05 and 7.502E-06, respectively. Numbers of brain lobes counted N = 5, N = 5, N = 5, N = 6, respectively. B Statistics on nuclear Pros in the NBs in mxc RNAi larval central brain at various developmental stages of 48, 72, 96 and 120 hours AEL (After Eggs Laying). The data are plotted as mean ± SD. *p < 0.05 and ****p < 0.0001 using a Student's t test, p = 0. RNAi and His4 RNAi FLP-out clones in Type I NB also confirmed that Dpn was ectopically expressed in the GMCs in both His3 (Fig. 5H-) and His4 knockdown lineages ( Fig. 5J-J'). These data indicate that reduced transcription of histone genes contribute to the blockage of GMC differentiation. As canonical histones make up the nucleosomes that package the genome to form chromatin [37], a reduction in canonical histone gene transcription of may cause DNA damage and genomic instability. We used  anti-phosphorylated H2Av, a marker of DNA doublestrand breaks (DSBs) [38][39][40], to visualize DNA damages in mxc knockdown larval brains (Additional file 1: Fig. S5). Obvious phosphorylated H2Av signals were observed (arrowheads; Additional file 1: Fig. S5B), suggesting occurrences of DSBs in the NBs of mxc RNAi lines. We also stained the larval brains of His3 and His4 RNAi lines with anti-phosphorylated H2Av and observed similarly phosphorylated H2Av signals in these NBs (arrowheads; Additional file 1: Fig. S5C, D). This indicates that knockdown of His3 or His4 recapitulates the mxc phenotypes. Taken together, these results demonstrate that Mxc-mediated HLB assembly and histone gene transcription are critical for the maintenance of NBs cell fate and their progeny in larval brains.

The increase in autophagy induced by mxc knockdown as a defense mechanism in NBs
Persistent DNA DSBs can represent genotoxic stress on a cellular level, in which the DNA damage response (DDR) pathway is activated to combat this, which induces a variety of biological processes [41]. Autophagy, a lysosomedependent degradation pathway, is one of the processes activated via the DDR pathway [42,43]. According to previous reports, autophagy is thought to have dual roles in responding to DNA damage [44][45][46][47][48]. To explore whether autophagy is activated in mxc knockdown NBs, we stained wt and mxc knockdown larval brains with anti-Atg8a, a marker of autophagosomes [49]. In wt NBs, only a basal Atg8a signal was detected (Fig. 6A), suggesting autophagy only occurred at a low level. In contrast, an increased signal of Atg8a puncta was observed in the cytoplasm of mxc knockdown NBs (arrowheads; Fig. 6B), indicating an accumulation of autophagosomes. We quantified the Atg8a puncta in wt and mxc knockdown NBs (Fig. 6C). The number of dots per NB in mxc knockdown lines (2.30; n = 87) was significantly higher than that in the wt NBs (0.52; n = 60). To further verify the activation of autophagy, exogenous ATG8a-GFP was expressed as a visual reporter of autophagosome in NBs (Additional file 1: Fig. S6D, E). ATG8a-GFP puncta were clearly observed as accumulated in the cytoplasm of mxc knockdown NBs (arrowheads; Additional file 1: Fig. S6E), but with almost no obvious corresponding signals in wt NBs (Additional file 1: Fig. S6D). Similar experiments were carried out with mxc 16a − 1 mutants (Additional file 1: Fig. S6A, B). The number of ATG8a-GFP dots per NB in mxc 16a − 1 mutants (3.48; n = 50) were also significantly higher than that those of wt NBs (0.33; n = 57; Additional file 1: Fig. S6C). These results were consistent with mxc knockdown data, indicating that mxc deficiency triggers an increased autophagy in NBs.
We performed knockdown of Atg1, a key gene of the autophagy pathway [50,51], in the NBs of both wt and mxc knockdown lines (Fig. 6D-H). No obvious phenotype was observed in Atg1 knockdown wt larval brains (Fig. 6E, H), which was consistent with the low levels of autophagy detected above (Fig. 6A). In Atg1 and mxc double knockdown larval brains, more GMC-like Dpn + cells were detected, as compared with that in only mxc knockdown larval brains (Fig. 6F, G). Quantification showed a higher number with about 368 Dpn + cells/lobe (n = 4 lobes) in double knockdown brain lobes, as compared with that in mxc knockdown brains, with about 276.5 Dpn + cells/lobe (n = 4 lobes; Fig. 6H). These data suggest that elevated autophagy induced by Mxc deficiency in NBs might be a defense mechanism responding to abnormal HLB assembly and consequently pre-terminated NB cell fate.

Discussion
NPAT/Mxc is a component of the HLB, participating in transcription of canonical histone genes and in pre-mRNA processing [5,10]. Although NPAT gene has been reported as a candidate gene for mental retardation, no genetic mutations of NPAT gene have been found in clinic cases. Our study shows that knockdown mxc in larval brain NBs causes the reduction of transcription levels of histone genes and increased DNA DSBs. These defects lead to premature termination of NB cell fate and blockage of GMC differentiation processes, resulting in neuronal and locomotion defects in adult flies.
In mxc knockdown flies, total NB numbers decrease and its proliferation potential is disrupted, leading to severe neuronal loss in the adult central brain. The mutant flies show locomotion disorder and survive no more than two days after eclosion. It is notable that the hemizygotes of mxc 16a − 1 die as late larvae or early pupae. NB specific overexpression of Mxc in mxc 16a − 1 mutants rescues these animals, at least to the late pupal stage. Some of these flies then go on to survive to adulthood without obvious mobility problems. This indicates that mxc mutant lethality at larval or early pupal stage is mainly due to the disruption of CNS development. Such results support the hypothesis that NPAT/mxc deficiency could act as a pathogenic factor for neurodevelopmental disorders such as mental retardation.
Notably the hemizygotes of mxc 16a − 1 exhibits weaker phenotype of nuclear Pros in NBs and die as late larvae or early pupae. Since mxc 16a − 1 mutation only occurred from the 1824 aa at the very C-terminal end with the rest aa sequence unchanged, it is likely that the very C-terminal sequence is required for colocalization of MPM2 and Mute in NBs while some other domain(s) resided within unchanged N-terminal region maintain(s) function to prevent nuclear Pros formation.
It has been reported that the ectopic expression of Dpn in immature INP cells can transform these cells into NBs-like cells that divide uncontrollably, causing tumor formation [23]. Although Dpn + GMCs might have proliferation potential, no tumor over-growth phenotype is observed in our study. This may reflect the fact that in addition to NBs, the cell-cycles of GMCs in mxc knockdown lines are also slowed down or even arrested. The slowed down cell cycle seems to protect NBs and GMCs from tumor over-growth. We propose that in wt NBs, Mxc acts to suppress Dpn expression in GMCs and promotes GMC differentiation. This conclusion is consistent with notion that in the absence of Mxc functions, large populations of GMC-like cells continue to express Dpn, and their differentiation process is blocked. The results from His3 and His4 RNAi lines also show Dpn + GMCs phenotypes consistent to the mxc data.
DSBs are often considered to be a factor for tumor formation [52]. A previous report demonstrated that mxc knockdown and mxc 16a − 1 led to larval lymph gland hyperplasia [14]. Interestingly, in our study, although knockdown of mxc or histone genes cause severe DSBs in the larval brains, there was no evidence suggesting the over-growth of NBs or GMCs. Such a discrepancy of Mxc function between lymph gland cells [14] and NBs could possibly be due to the cell fate determinant, Pros. The premature accumulation of nuclear Pros in larval brain NBs is likely to be a mechanism to prevent NBs from over-proliferation when under a state of DSBs stress or perhaps under other instances of genomic instability. Similar results have been reported in that ionizing radiation treatments are noted to induce both DSBs and prematurely accumulated nuclear Pros in Drosophila larval brain NBs [53], with no NB over-proliferation observed. Premature nuclear Pros, that early terminates the NB cell fate and diminishes their proliferative properties, may play a role as a protective mechanism to avoid NB over-growth.
Autophagy is a conserved cellular process which can be induced under stress conditions such as nutrient starvation or DNA damage [43,54]. At present, no data has been presented describing the role of autophagy in the larval brain NB development of Drosophila. Based on our data, autophagy shows a protective role against Dpn + GMC formation. In the wt larval brain NBs, autophagy is kept at a low level. However, the autophagy levels are elevated in mxc knockdown NBs. Blocking the autophagy process by knocking down Atg1 enhances the Dpn + GMC population. Although autophagy does not appear to be required in wt larval brain NBs under normal conditions, it represents a protective function under the DSBs stress induced by mxc knockdown.

Conclusions
Our study revealed the functions of Mxc in the CNS development (Fig. 7). The lack of canonical histone genes caused by mxc knockdown leads to lower NB proliferation potential and premature termination of NB cell fate, as well as blockage of differentiation process of GMCs, which ultimately results in neuronal defects in adult flies. Our study provides the first important clue that NPAT/mxc may act as a pathogenic factor for neurological disease, potentially those related to mental retardation.
UAS-Mxc was generated in this study. To generate pUAST-mxc construct, the full-length mxc cDNA was amplified with the primers 5′-CGG AAT TCA TGG AGT CGA TTG TCC TGC ATT C-3′ and 5′-TAA AGC GGC CGC TCA GGT GCC GTG CAG ATG TGA-3′. The sequence was cloned into a pUAST vector. The construct was injected into the VK33 lines at the Core Facility of Drosophila Resource and Technology, CEMCS, CAS, following standard methods.
For FLP-out mosaic analysis, the fly line with the genotype of hsFLP; act > y + >GAL4, UAS-CD8-GFP was used to cross with RNAi lines [34]. The FLP-out clones were labeled with GFP. Both MARCM clones and RNAi FLPout clones were induced using a 45 min heat-shock treatment of 37 °C at 32-36 h after egg laying.

Antibody generation
The sequence of Mute for the antibody, aa 3-164, was selected according to Bulchand's work [36]. The sequence was cloned into a pGEX-4T-1 expression vector. The antigen was bacterially produced and the antibody was raised in guinea pigs.

Immunofluorescence staining and antibodies
Different stages of larval brains and adult brains were dissected in ice-cold Schneider's Drosophila Medium (Gibco). Samples were fixed for 18 min in PBS (10 mM NaH2PO4/Na2HPO4, 175 mM NaCl, pH7.4) with 4% paraformaldehyde at room temperature [55,56]. The samples were incubated with primary antibodies at 4℃ overnight, and then with secondary antibodies at room temperature for 1-2 h. Antifade Mounting Medium (P0126, Beyotime, China) was used to protect the fluorescent signals of the samples. The images were obtained using an Olympus FV1000 confocal microscope and processed using Adobe Photoshop CC 2018.

Quantitative RT-PCR
RNA was extracted by the TRIzol (Invitrogen) method. About 30 brains of third-instar larvae were dissected for each sample. HiScript III 1st Strand cDNA Synthesis Kit (Vazyme, China) was used for reverse transcription according to the manufacturer's protocol. Note that it was necessary to select Random hexamers rather than Oligo (dT) for reverse transcription of histone mRNA.

Statistical analysis
All data are expressed as the mean ± SD. All statistical data are processed using unpaired two-tailed Student's t-test in GraphPad Prism.