E2f1 transcription factor is required for neurodevelopment through regulation of conserved miR-33 and RhoGEF trio

Abstract

Background

The E2f1 transcription factor exhibits diverse and complex functions beyond its canonical roles in cell cycle regulation, DNA damage response, and apoptosis. Aberrant expression and activity of E2f1 in neurons have been observed in neurodegenerative disorders such as Parkinson’s disease (PD) and Alzheimer’s disease (AD). However, it remains unclear whether the E2f1 is required for neurodevelopment.

Methods

Functions of the E2f1 in neurodevelopment were explored using mutant flies and over-expression flies. The CNS axons in embryos and synaptic growth at larval neuromuscular junctions were assessed by morphological and immunofluorescence analysis. Reverse transcription-quantitative PCR (RT-qPCR) was performed to quantify gene expression of miR-33, hmiR-33, e2f1, he2f1, and dp. Western blot analysis was performed to assess Trio protein level. Chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR) was employed to detect direct regulation of miR-33 by the E2f1. The dual-luciferase reporter assay was utilized to investigate the post-transcriptional regulation of trio by the miR-33. Grid crossing and peristalsis contraction assays were performed to analyze the locomotor behavior.

Results

Flies with dysfunctional E2f1, reduced expression or overexpression of E2f1 in neurons exhibit disrupted axons and abnormal synaptic growth at neuromuscular junctions. The miR-33, an important regulator of lipid and glucose metabolisms with uncharacterized roles in neurodevelopment, is directly regulated by the E2f1 and works downstream of the E2f1 to regulate neurodevelopment. The RhoGEF Trio, a central regulator of neurodevelopment and synaptic plasticity, acts as a target of the miR-33 and the downstream effector of the E2f1-miR-33 axis to modulate neurodevelopment. Notably, flies with dysfunctional e2f1 or loss of miR-33 display aberrant locomotor behaviors. Furthermore, the E2f1-miR-33-Trio signaling and its functions in neurons are conserved in Drosophila and human.

Conclusions

This study uncovers a novel role of the E2f1 transcription factor in the central nervous system, identifies upstream regulators of the RhoGEF Trio, and reveals an evolutionarily conserved E2f1-miR-33-Trio signaling cascade critical for neurodevelopment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-025-02612-2.

Background

The E2f1 transcription factor is the first identified member of the E2f family. It is highly conserved and plays critical roles to regulate cell cycle, DNA damage response, and apoptosis. In mammalian cells, the E2f family encompasses nine members, including transcriptional activators E2f1, E2f2 and E2f3a, repressors E2f3b, E2f4, E2f5 and E2f6, and atypical members E2f7 and E2f8 lacking a DP-binding domain. The Drosophila E2f family is a simpler system with only two members, the activator E2f1 and the repressor E2f2 [1, 2]. Expression and activity of the E2f1 are tightly regulated at multiple levels, including transcription, mRNA and protein stability, post-translational modifications, and interactions with other regulatory proteins [3]. microRNAs (miRNAs) are important components in the E2f1 signaling network. miRNAs can be directly regulated by the E2f1 transcription factor. On the other hand, many miRNAs regulate E2f1, its canonical interacting-partner retinoblastoma protein pRB, or genes downstream of the E2f1, thus forming mutual regulatory networks to modify E2f1 expression and activities [4, 5]. Systemic investigations analyzing miRNAs in the E2f1 signaling network have been performed in multiple studies. Expression of miRNAs in response to mitogenic stimulation was analyzed in primary fibroblasts, and four miRNA clusters, including let-7a-d, let-7i, miR-15b-16-2 and miR-106b-25, were found downregulated in e2f1-knockout or e2f3-knockdown cells [6]. Differentially expressed miRNAs in e2f1 mutant flies and e2f1 p53 double mutant flies were identified during development and in response to X-ray irradiation, and the E2f1 plays a more prominent role to regulate miRNAs compared with the p53 [7]. Moreover, many miRNAs related with the pRB were analyzed in retinoblastoma patients [8–10].

Beyond its canonical roles in cell cycle regulation, DNA damage response, and apoptosis, the E2f1 has diverse and complex functions. Depending on tumor types and context, the E2f1 has both tumor suppressive and oncogenic functions, and dysregulation of the E2f1 is closely associated with cancers [3, 11]. It is suggested that the E2f1 is an important regulator of metabolic homeostasis at both physiological and pathological conditions [12]. The e2f1 null mice exhibits smaller pancreas, dysfunctional islets, and reduced insulin secretion in response to glucose challenge [13]. In muscle and brown adipose, the E2f1 represses expression of key genes required for energy homeostasis and mitochondrial functions, and mice loss of the e2f1 are resistant to diet-induced obesity [14]. Studies in Drosophila find that e2f1-deficient fat body has reduced storage of fat and a lower level of circulating trehalose [15]. And the E2f1 regulates expression of the Phosphoglycerate kinase Pgk, and loss of E2f1 regulation on pgk decreases glycolytic flux and causes an abnormal mitochondrial morphology [16].

In neurodegenerative disorders, such as Parkinson´s disease (PD) and Alzheimer disease (AD), aberrant expression and activities of the E2f1 and pRB in neurons were observed [17–19]. Aberrant E2f1 activation was also observed in a polyglutamine (polyQ) Drosophila model [20]. While neurons are highly differentiated cells and normally do not enter a cell cycle, E2f1 works as cell cycle suppressor in mature neurons [21]. Studies in mice indicates increased expression of the e2f1 gene in the aging process, and e2f1 mutant mice had age-dependent synaptic perturbations and olfactory and memory-related deficits, however, brain development was unaffected [22]. Another study using a different strain of e2f1-deficient mice also indicates that brains development is normal, however, division of stem cells and progenitors in the proliferative zones of the lateral ventricle wall and the hippocampus is significantly decreased [23]. Until now, what the role E2f1 might play in the central nervous system (CNS), beyond its canonical roles in cell cycle control and apoptosis, is still not clear.

In this study, we demonstrate that dysfunctional E2f1 or dysregulated expression of E2f1 in neurons leads to disrupted axons in embryos and aberrant synaptic growth at larval neuromuscular junctions in Drosophila. The miR-33 acts downstream of the E2f1 to mediate the E2f1 signaling in neurons, and miR-33 knockout flies have disrupted axons and aberrant synaptic growth at neural muscular junctions similar as e2f1 mutant flies. The RhoGEF Trio works as a target of the miR-33 and the effector of the E2f1 signaling to regulate neurodevelopment. Moreover, the E2f1-miR-33-Trio signaling and its functions in neurons are conserved in Drosophila and human. Furthermore, our study discloses a novel conserved regulatory mechanism of the Trio in neurons by the E2f1/miR-33 signaling and provides new knowledge to understand how the RhoGEF Trio is regulated.

Methods

Fly genetics

All flies were maintained on standard corn meal with a 12 h light-dark cycle at required temperatures (18 °C, 25°C or 29 °C). The following fly lines were used: e2f1⁰⁷¹⁷² (BL11717), e2f1ⁱ² (BL7274), UAS-e2f1-RNAi (BL27564), UAS-dp-RNAi (BL33372), UAS-e2f1 (BL4770), UAS-miR-33-sponge (BL61385), UAS-miR-33 (BL59871), trio¹ (BL9129), trio^s137203 (BL8594) and UAS-trio (BL9513) were obtained from the Bloomington Drosophila Stock Center; miR-33^KO (DGRC116361) was obtained from the Kyoto Stock Center; UAS-trio-RNAi (VDRC40138) was obtained from the Vienna Drosophila Resource Center, and w¹¹¹⁸; elavGal4; hhGal4; elavGal4-tubGal80^ts were maintained in the lab.

RNA extraction and RT-qPCR

The Drosophila total RNA was extracted from 50 L3 larval brains, and human total RNA was extracted from 1 × 10 ⁷ HEK-293T cells, using the Trizol reagent (Invitrogen, #15596026) following the manufacturer’s instructions. One microgram of total RNA was reverse-transcribed using the Evo M-MLV RTase Enzyme Kit (Accurate Biotechnology, AG11705) to synthase cDNA for mRNA quantification. To quantify the miR-33, 2 µg of total RNA was treated with Poly(A) Polymerase, then reverse transcribed with a universal reverse transcription primer in the kit, which consists of an Oligo(dT) and a specific tag at its 5’ terminus, using the miRNA 1 st strand cDNA synthesis kit (Accurate Biotechnology, AG11716). Quantitative PCR was performed using the SYBR mix (Accurate Biotechnology, AG11701) with 3 independent biological replicates. The ribosomal gene rp49 and hGAPDH were used as the normalization control for a Drosophila gene and a human gene, respectively, and the U6 gene and hU6 gene were used as the normalization control for a Drosophila miRNA and a human miRNA, respectively. Primers were listed in Supplementary Table S1.

Plasmids construction

A 117–1099 nt fragment of trio 3’ UTR was PCR amplified using genomic DNA (gDNA) from wild-type w ¹¹¹⁸ flies as the template, and cloned it into BamHI/EcoRI sites of pBluescript Ⅱ-KS (+) plasmid. Binding-sites of the miR-33 seed-sequence at positions 736–742 nt (mut 1) and 740–747 nt (mut 2) in the trio 3’ UTR were mutated from “AUGCA” to “GACAG” using FAST Site-Directed Mutagenesis Kit (Tiangen Biotechnology, KM101), respectively. The trio 3’ UTR fragments of wild-type and mutants were subcloned into XbaI site of the Actin5C-firefly luciferase plasmid [24]. A gDNA fragment containing the 463 bp-upstream to 101 bp-downstream pre-miR-33 gene locus was PCR amplified and cloned into XhoI/EcoRI sites of the pAc5.1/V5/His B plasmid vector (Sigma-Aldrich) for miR-33 expression. A fragment containing 182 bp-upstream to 196 bp-downstream of the miRNA lin4 from c. elegans was inserted into XhoI/SacII sites of the pAc5.1/V5/His B plasmid vector as a control. A 659–870 nt fragment of htrio 3’ UTR was provided by the GenePharma. The binding-site of the hmiR-33 seed-sequence at 805–811 nt was mutated from “AAUGCA” to “UUACGU” to generate the htrio 3’ UTR ^mut. The htrio 3’ UTR and the htrio 3’ UTR ^mut fragments were cloned them into SacI/XhoI sites of the GP-miRGLO luciferase reporter vector (GenePharma), respectively. To generate transgenic flies of the trio 3’ UTR, the trio 3’ UTR 117–1099 nt fragment was cloned it into XhoI/NotI sites of the pCaspeR-tub-DsRed-attB plasmid (gift from Peng Zhang, University of Utah) to construct pCaspeR-tub-DsRed-attB-trio 3’ UTR plasmid. To generate transgenic flies of the pre-miR-33, a gDNA fragment containing the 463 bp-upstream to 101 bp-downstream pre-miR-33 gene locus was cloned into XbaI site of the pW-moe-GFP-GPI plasmid (gift from Peng Zhang, University of Utah) to construct pW-moe-GFP-GPI-miR-33 plasmid. To generate transgenic flies of the pre-hmiR-33, a gDNA fragment containing the 140 bp-upstream to 122 bp-downstream pre-hmiR-33 gene locus was cloned it into EcoRI/XbaI sites of the pUAST-attB plasmid to construct pUAST-attB-hmiR-33 plasmid. The he2f1 cDNA was cloned into EcoRI/BamHI sites of the pCDH-CMV plasmid (Genecreate Biotechnology) for he2f1 overexpression in HEK-293T cells and was provided by Genecreate Biotechnology. The lentiviral shRNA for he2f1 (sh-he2f1) and the negative control shRNA (sh-NC) were provided by Wzbio Technology. The sh-he2f1 sequence is: 5’-CGTGGACTCTTCGGAGAACTTCTCGAGAAGTTCTCCGAAGAGTCCACGTTTTTT-3’. The sh-NC sequence is:5’-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGTGCTCTTCATCTTGTTGTTTTTG-3’.All constructed plasmids were verified by sequencing. Primers were listed in Supplementary Table S1.

Generation of transgenic flies

The pCaspeR-tub-DsRed-attB-trio 3’ UTR plasmid DNA was microinjected into the Drosophila melanogaster genome at the attP40 site on chromosome 2 (BL25709). The pW-moe-GFP-GPI-miR-33 plasmid DNA, pUAST-attB-hmiR-33 plasmid DNA, and pBID2 20xUAS_V5 hTrio plasmid DNA (gift from Brian D. McCabe, Brain Mind Institute, Swiss Federal Institute of Technology Lausanne) [25] were microinjected into the Drosophila melanogaster genome at the attP2 site on chromosome 3 (BL24749), respectively. All microinjections were performed at the Core Facility of Drosophila Resource and Technology, Chinese Academy of Sciences, Shanghai.

Immunostaining and imaging

Embryos laid by 7–14 days-old flies were collected at stage 15–16, dechorionated in 3% sodium hypochlorite for 3 min, and washed extensively with distilled water. Dechorionated embryos were fixed in 4% paraformaldehyde (Thermo Fisher) diluted in 1 × PBS with an equal volume of heptane added for 20 min. Sample vials were rotated on an agitator for 20 min. Third instar larvae were dissected in 1 × PBS along the dorsal midline to expose the ventral nerve cord (VNC) and abdominal muscles. To dissect brains from adults, heads were gently removed from bodies, then brains were dissected from adult head cuticles and surrounding tracheas were carefully removed. After dissection, larvae and brain samples were fixed in 4% paraformaldehyde (Thermo Fisher). All samples were washed in PBT (1 × PBS/0.1% Triton X-100) 3 times, blocked in PBTB (PBT containing 5% normal goat serum) for 2 h, and incubated with the primary antibody overnight. The following primary antibodies were used: mouse anti-CNS axons antibody (DSHB, BP102, 1:50), mouse anti-Fasciclin II antibody (DSHB, 1D4, 1:50), rabbit anti-Horseradish Peroxidase (HRP) antibody (Jackson ImmunoResearch, 323-001-021, 1:100), and mouse anti-Discs Large 1 (DLG) antibody (DSHB, 4F3, 1:30). After 3 washes with PBT, anti-rabbit or anti-mouse fluorescence antibodies, including Alexa 488 and 555 (Cell Signaling Technology, 1:500) were used. All immunostained samples were mounted in antifade mountant (Beyotime), and analyzed on an Olympus confocal laser scanning microscope (FV3000). Images were processed using Adobe Photoshop, Illustrator, and ImageJ.

Quantitation of bouton number, branch number and length in NMJs

Quantification of bouton number, branch number and length in Neuromuscular Junctions (NMJs) were performed on muscles 6/7 of abdominal segment 3 (A3) by anti-HRP and anti-DLG staining. At least 12 samples of each genotype were analyzed. The Z-stack images were acquired with an Olympus confocal laser scanning microscope (FV3000) with a 60 × oil immersion objective lens. A bouton was defined as a synaptic varicosity (swelling) compared with adjacent axonal segments in NMJs labeled with the presynaptic marker anti-HRP antibody and the postsynaptic marker anti-DLG antibody [26]. Branch numbers were defined as any branch with three or more boutons from the primary nerve terminal and any subsequent nerve terminals. Branch length was measured from the first bouton of primary nerve terminal to subsequent branches. The bouton number, branch number and branch length were determined using ImageJ. The bouton density for each NMJ was then calculated as the total number of boutons divided by the total branch length, expressed as boutons per micrometer (boutons/µm).

Western blotting

To measure the endogenous Trio protein level, brains of adult flies were homogenized in RIPA buffer (140mM NaCl, 1mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 10mM Tris-Cl PH8.0) supplemented with freshly-added protease and phosphatase inhibitors (Beyotime) on ice. Lysates were denatured in sample buffer, subjected to SDS-PAGE (polyacrylamide gel electrophoresis), and transferred to polyvinylidene fluoride membranes. Blots were blocked in 5% milk for 1 h, followed by overnight incubation with mouse anti-Trio antibody (DSHB, AB528494, 1:200), and mouse anti-Alpha-Tubulin antibody (Cell Signaling, #3873, 1:2000). Blots were washed in 0.1% TBST, incubated with a horseradish peroxidase–conjugated goat anti-mouse secondary antibody (Proteintech, SA0001) and processed for chemiluminescence using Femto-sig ECL Western Blotting Substrate (Tanon, #180–506). The gray value of the bands was calculated using ImageJ after subtracting background density.

Chromatin immunoprecipitation assay

The Chromatin immunoprecipitation (ChIP) assay was performed using the Chromatin Immunoprecipitation (ChIP) Kit (Genecreate, JKR23002A). Three hundred adult fly brains were dissected in cold 1 × PBS, fixed in 0.5% formaldehyde/PBS for 10 min, then treated with 125 mM glycine for 5 min. Brain samples were homogenized and lysed in SDS lysis-buffer on ice for 30 min, and sonicated on a 10 s “on”/10 s “off” cycle for 25 cycles to obtain 200–500 bp chromatin fragments by using a Sonicator (Scientz, UP-250). Soluble chromatin from fly brains was immunoprecipitated using the Drosophila anti-E2F1 antibody [27]. Rabbit IgG antibody was used as a negative control. Ten million of HEK-293T cells were crosslinked with 0.5% formaldehyde/PBS for 10 min, then treated with 125 mM glycine for 5 min. Cells were lysed in SDS lysis buffer on ice for 30 min, and sonicated on a 10 s “on”/10 s “off” cycle for 45 cycles to obtain 200–500 bp chromatin fragments by using a Sonicator. Soluble chromatin from HEK-293T cells was immunoprecipitated using the human anti-E2F1 antibody (Cell Signaling, #3742). Human IgG antibody (Beyotime, A7001) was used as a negative control. After an overnight incubation, immunocomplexes from fly brains or cells were precipitated by adding Protein A Dynabeads (Invitrogen), respectively, washed four times with wash buffer. The DNA samples were eluted with 150 µl of elution buffer. After elution, samples were incubated at 65℃ for 6 h to reverse the cross-link. DNA samples were purified using a QIA quick PCR purification kit (QIAGEN, #28104) and resuspended in 50 µl distilled water. A 2 µl eluate was used for quantification by real-time PCR (qPCR). qPCR primers were listed in Table S1. The percentage of input (Percent input) represents the amount of DNA pulled down by the designated antibody in the ChIP assay relative to the amount of starting material (input).

Luciferase reporter assay

Fly S2 cells were cultured in InsectPro Sf9 Medium (Basal, H832KJ) with 5% heat-inactivated Fetal Bovine Serum (FBS) (Gibco, #26010074) and 1% Pen-Strep Amphotericin B (Beyotime, C0224) at 25℃ until 80–90% confluency. Cells were transferred to a 24-well plate with 6 × 10 ⁴ cells per well, and cultured for 12 h to 70–80% confluency. Transfection was performed using Effectene transfection reagent (QIAGEN, #301425) following the manufacturer’s instructions. Three microliter of plasmids DNA including 1 µL wild-type trio 3’ UTR (100 ng) or mutant trio 3’ UTR construct, 1 µL pAC5.1-HisB-miR-33 or pAC5.1-HisB-lin-4, and 1 µL pAC5.1-HisB-Renilla luciferase plasmid DNA (50 ng) were mixed with 1 µL H₂O, 60 µL Buffer EC and 1.6 µL Enhancer solution. After 3 min, 5 µL Effectene was added into the mixture and left it for 8 min to obtain the transfection cocktail mix. The transfection cocktail mix was added into cells in the 24-well plate. Transfected cells were cultured in a 25℃ incubator, and medium was changed with fresh medium after 12 h. The HEK-293T cells were inoculated in a 24-well plate at 1 × 10 ⁵/well and left grow until 60% confluency. Plasmid DNA of GP-miRGLO-htrio 3’ UTR or GP-miRGLO-htrio 3’ UTR mutant were co-transfected with hmiR-33 mimics or NC mimics, respectively. Cells were transfected using the Effectene transfection reagent (QIAGEN, #301425). Plasmid DNA consisting 1µL of wild-type htrio 3’ UTR (100 ng) or mutant htrio 3’ UTR, 1 µL of hmiR-33 mimics or NC mimics (50 ng), 2 µL H2O, 60 µL Buffer EC solution, and 1.6 µL Enhancer solution were mixed and left for 3 min. Five microliter Effectene was added into the DNA mixture and left for 8 min to obtain the transfection cocktail mix. The transfection cocktail mix was added into cells in the 24-well plate. Transfected cells were cultured in a 37℃ incubator, and medium was changed with fresh medium after 12 h. The relative luciferase activity was measured at 48 h post-transfection using the Dual-Luciferase reporter Assay System (Promega, #E1910). One hundred and twenty microliter 1 × cell lysis buffer was added into each well, incubated for 20 min, transferred to a microtube and centrifuged at 3,000 rpm for 3 min, and supernatant was transferred to a 96-well plate. One hundred microliter of Dual&Glo working solution was added in to the well, and activity of the Firefly luciferase was measured at 560 nm absorbance. Then 100 µL Stop&Glo working solution was added to each well and the activity of Renilla luciferase was measured at 460 nm absorbance. The Firefly luciferase activity relative to the Renilla luciferase activity in each comparison group was calculated, and was set as 1 unit in the control.

Larval locomotion assay

L3 larvae were collected and washed in distilled water to remove any traces of food. One larva was let stay on a 100 mm-diameter Petri dish covered with 2% non-nutritious agar (Yeasen, 10208ES) for 1 min before testing. For the grid crossing assay, individual larva was placed at the dish center above a graph paper with 0.5 cm ² marked grids. The number of grid-line crossed by a larva within 30 s was recorded. For the peristalsis contraction assay, full body peristalsis contractions (full posterior to anterior movement = 1 contraction) in 1 min were counted for each larva under a dissection microscope. Five trials per larva were conducted, and the total number of larvae analyzed per genotype was 30.

Sequence alignment and analysis

The binding-site of the miR-33 seed-sequence in the trio 3’ UTR and hmiR-33 seed-sequence in the htrio 3’ UTR was analyzed using TargetScan (http://www.targetscan.org/). The trio 3’ UTR complementary to the miR-33 seed-sequences from different species was aligned with Blast. The E2f-site in ~ 2 kb upstream genomic locus of the miR-33/hmiR-33 was analyzed using Promo (https://alggen.lsi.upc.es/cgi- bin/promo_v3/promo/promoinit.cgi? dirDB = TF_8.3).

Statistical analysis

All statistical comparisons were performed using Prism 9.0. For NMJs quantification and larval locomotion assay, One-way analysis of variance (ANOVA) was used for comparisons between groups with the Tukey’s post hoc test. For qPCR and luciferase reporter assay, a Two-tailed, two-sample t test was used for pairwise comparisons of two samples. The significance level was indicated as * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. All experiments were repeated for at least three times. The error bar in graphs denotes the standard error of the mean value from the replicated experiments.

Results

Flies with dysfunctional E2f1 or abnormal e2f1 expression have impaired axons and abnormal growth of synapses

Although aberrant expression and activation of the E2f1 transcription factor was observed in patients with neurodegenerative disorders [17–19], whether the E2f1 is required for the central nervous system (CNS) development is still not clear. Brains in third instar larvae (L3) and adults of e2f1 mutant flies displayed normal morphology (Figure S1), which were similar as the e2f1 mutant mice [22, 23]. However, we found that axons in e2f1 ^07172/i2 transheterozygous embryos were disrupted when embryos at stage 15–16 were stained with mAb BP102, a marker of the axon scaffold. The commissures and connectives in e2f1 mutant embryos were broken (n = 23, Fig. 1A, Table S2). In addition, knockdown of the e2f1 gene expression by a pan-neural elav-Gal4 (elav-e2f1 RNAi) caused similar defects as in e2f1 mutant embryos (n = 20, Fig. 1A and S2, Table S2). And overexpression of the e2f1 gene by elav-Gal4 ^ts completely rescued defective commissure segments in e2f1 mutant flies (elav ^ts > e2f1, e2f1 ^07172/i2, n = 15, Fig. 1A and S2, Table S2). Moreover, overexpression of the e2f1 gene by elav-Gal4 ^ts (elav ^ts > e2f1) caused disrupted axons similar as that in e2f1 mutant flies (n = 20, Fig. 1A, S2 Table S2), suggesting that the dosage of the e2f1 gene is critical for axons development.

Fig. 1.

Drosophila E2f1 is required for axon patterning and NMJs growth. A CNSaxons in stage 15–16 embryos of indicated genotypes were visualized with mAb BP102 staining. Scale bar = 10 μm. Broken commissures (arrowhead) and connectives (arrows) were observed in e2f1 ^07172/i2 (n = 23), elav > e2f1 RNAi (n = 20), elav ^ts > e2f1 (n = 20), and elav > dp RNAi (n = 18) flies, whereas axons are intact in wild-type w ¹¹¹⁸ (n = 15) and elav > + (n = 15) and elav ^ts > + (n = 15) control flies. Overexpression of the e2f1 by elav ^ts-Gal4 fully rescued disrupted axons in e2f1 ^07172/i2 mutant flies (elav ^ts > e2f1, e2f1 ^07172/i2, n = 15). B Confocal images of NMJ boutons at muscles 6/7 in L3 larvae of indicated genotypes were visualized with anti-HRP (red) and anti-DLG (green) antibodies staining. Scale bar = 20 μm. The bouton number (C), branch length (D), and branch number (E) per M6/7 NMJ in (B) were quantified, they were all significantly increased in e2f1 ^07172/i2 mutant flies (n = 18) compared with w ¹¹¹⁸ flies (n = 15), and elav > e2f1 RNAi flies (n = 16) and elav > dp RNAi flies (n = 18) compared with elav > + flies (n = 15), and decreased in elav ^ts > e2f1 flies (n = 25) compared with elav ^ts > + flies (n = 12). Overexpression of the e2f1 by elav ^ts-Gal4 fully rescued abnormal growth of NMJs in e2f1 ^07172/i2 mutant flies (elav ^ts > e2f1, e2f1 ^07172/i2, n = 16). Data are mean ± SEM, **** p < 0.0001, n.s., no significance. One-way ANOVA test with Tukey’s post hoc test

Next, we examined synaptic growth at L3 neuromuscular junctions (NMJs) by immunostaining of the presynaptic marker HRP and postsynaptic marker Dlg [26]. Compared with wild-type w ¹¹¹⁸ flies and elav-Gal4 control flies, the bouton number, branch length and branch number were all significantly increased in e2f1 mutant flies (n = 18, p < 0.0001, Fig. 1B-E) and elav-e2f1 RNAi flies (n = 16, p < 0.0001, Fig. 1B-E), indicating that dysfunctional E2f1 or reduced e2f1 expression promotes synaptic overgrowth. Increased expression of the e2f1 gene by elav-Gal4 ^ts fully rescued the synaptic overgrowth in e2f1 mutant flies (n = 16, p < 0.0001, Fig. 1B-E), while overexpression of the e2f1 gene alone decreased the bouton number, branch length and branch number (n = 25, p < 0.0001, Fig. 1B-E). Knockdown of the dp, a core component of the DREAM complex and essential E2f1 co-factor [1], via elav-Gal4 (elav-dp RNAi) caused synaptic overgrowth similar as that of e2f1 mutant flies and elav-e2f1 RNAi flies (n = 18, p < 0.0001, Fig. 1B-E and S2), further validating the requirement of appropriate E2f1 activity for NMJs development. However, the mushroom bodies in e2f1 mutant adult showed normal morphology (Figure S3). Together, these data suggested that the E2f1 transcription factor is required for neurodevelopment in Drosophila. The bouton density was calculated and revealed no significant difference between the e2f1 mutant and wild-type w ¹¹¹⁸ flies (Figure S4). This result indicates that the increase in bouton number is proportional to the overall increase in NMJ branch length.

miR-33 acts as a direct target of E2f1 to regulate neuron development

MicroRNAs (miRNAs) constitute important components of the E2f1 signaling [4]. It is known that miRNAs play important roles in all aspects of neuronal development, functions and plasticity, and dysregulation of miRNAs has been implicated in neurodegeneration disorders [28, 29]. We explored whether miRNA(s) mediates the E2f1 functions in neurodevelopment. Based on our previous work that differentially expressed miRNAs were identified in e2f1 mutant flies [7], we analyzed the evolutionary conservation of miRNAs that were downregulated in the e2f1 mutant. We then examined the presence of the E2F-binding site in these conserved miRNAs. Our analysis identified the miR-33 as a candidate target of the E2f1, as it is evolutionarily conserved and contains an E2F-binding site (Figure S5). Quantitative RT-PCR confirmed that miR-33 expression was downregulated in e2f1-mutant brains and upregulated in elav ^ts-e2f1 brains (p < 0.0001, Fig. 2A-B). To verify that the miR-33 is a functional target downstream of the E2f1 transcription factor to regulate neurodevelopment, the miR-33 expression was induced by elav-Gal4 (elav-miR-33, Figure S6). We found that the elav-miR-33 fully rescued disrupted axons in embryos (n = 17, Fig. 2C, Table S2) and synaptic overgrowth in L3 larvae of e2f1 mutant flies (n = 17, p < 0.0001, Fig. 2D-G).

Fig. 2.

miR-33 works downstream of E2f1 to regulate axon patterning and NMJs growth. Relative expression of the miR-33 in L3 larvae brains of w ¹¹¹⁸ and e2f1 ^07172/i2 mutant flies (A), and elav ^ts > + flies and elav ^ts > e2f1 flies (B) was detected via RT-qPCR. Data were normalized with U6 as an internal control. Expression of the miR-33 was decreased in e2f1 mutant flies compared with w ¹¹¹⁸ flies (A), and increased in elav ^ts > e2f1 flies compared with elav ^ts > + flies (B). **** p < 0.0001, two-tailed two-sample t test. C CNS axons in stage 15–16 embryos of indicated genotypes were visualized with mAb BP102 staining. Scale bar = 10 μm. Overexpression of the miR-33 by elav-Gal4 fully rescued disrupted axons in e2f1 ^07172/i2 mutant flies (elav > miR-33, e2f1 ^07172/i2, n = 17). D Confocal images of NMJ boutons at muscles 6/7 in L3 larvae of indicated genotypes were visualized with anti-HRP (red) and anti-DLG (green) antibodies staining. Scale bar = 20 μm. The bouton number (E), branch length (F), and branch number (G) per M6/7 NMJ in (D) were quantified. Overexpression of the miR-33 by elav-Gal4 fully rescued abnormal growth of NMJs in e2f1 ^07172/i2 mutant flies (elav > miR-33, e2f1 ^07172/i2, n = 17). Data are mean ± SEM, **** p < 0.0001, n.s., no significance. One-way ANOVA test with Tukey’s post hoc test. H Scheme of the miR-33 promoter region. Putative E2f-sites are depicted, and nucleotide positions of E2f-sites are provided with respect to the miR-33 transcription start site. I ChIP-qPCR analysis revealed significant binding of the E2f1 protein to site #1 and #2 in the miR-33 promoter region, and the PCNA promoter region as a positive control, and no binding to site #3 in the miR-33 promoter region and the negative control rp49 promoter region. * p < 0.05, *** p < 0.001, **** p < 0.0001, n.s., no significance. Two-tailed two-sample t test

Next, we investigated whether the miR-33 is a direct transcriptional target of the E2f1 transcription factor. In silico promoter analysis using PROMO identified three putative E2f-binding sites located ~ 2.0 kb upstream of the miR-33 locus (Fig. 2H). By performing a chromatin immunoprecipitation (ChIP) assay using an anti-E2f1 antibody, we found that the E2f1 protein was highly enriched at site #2, weakly enriched at site #1, and no enrichment at site #3 (Fig. 2I). As a positive control, the E2f1 protein was highly enriched at the promoter region of the PCNA gene, but not enriched at the promoter region of the negative control rp49 gene (Fig. 2I). Together, these data indicated that the miR-33 is directly regulated by the E2f1 transcription factor, and mediates the E2f1 functions in neurodevelopment.

miR-33 is required for neuron development

Previous studies have shown that the miR-33, an intronic miRNA in the gene encoding sterol-regulatory element-binding protein (SREBP) 2 [30], is an important regulator of lipid and glucose metabolisms. Dysregulation of the miR-33 is closely related with cardiovascular, metabolic, and inflammatory diseases [31]. The miR-33 is inferred to neurological disorders due to its critical roles on lipid metabolisms [32]. However, its involvement in neurodevelopment remains uncharacterized.

Given that overexpression of the miR-33 by elav-Gal4 fully rescued disrupted axons and synaptic overgrowth in e2f1 mutant flies (Fig. 2C-G), we investigated whether the miR-33 is required for neurodevelopment. RT-qPCR analysis of miR-33 expression in different organs revealed high-level miR-33 expression in adult brains and moderate-level expression in L3 larval brains (Figure S7). The BP102 staining indicated broken axons in miR-33 knockout embryos (miR-33 ^KO, n = 25, Fig. 3A and S6, Table S2) and elav-miR-33 RNAi embryos (n = 22, Fig. 3A, Table S2), and synaptic overgrowth in L3 NMJs of miR-33 ^KO flies (n = 18, p < 0.0001, Fig. 3B-E) and elav-miR-33 RNAi flies (n = 15, p < 0.0001, Fig. 3B-E), which were similar as that of e2f1 mutant flies (Fig. 2C-G). And overexpression of the miR-33 by elav-Gal4 fully rescued broken axons (n = 17, Fig. 3A, Table S2) and synaptic overgrowth in miR-33 ^KO flies (n = 18, p < 0.0001, Fig. 3B-E), suggesting that the miR-33 is essential for neurodevelopment. Moreover, axons were disrupted (n = 29, Fig. 3A, Table S2) and synapses were undergrowth in elav-miR-33 flies (n = 31, p < 0.0001, Fig. 3B-E), suggesting that the dosage of the miR-33 is important for normal growth of neurons. Similar to e2f1 mutants, miR-33 ^KO adult flies exhibited normal mushroom body morphology (Figure S8).

Fig. 3.

miR-33 is required for axon patterning and NMJs growth. A CNS axons in stage 15–16 embryos of indicated genotypes were visualized with mAb BP102 staining. Scale bar = 10 μm. Broken commissures (arrowhead) and connectives (arrows) were observed in miR-33 ^KO flies (n = 25) and elav > miR-33 RNAi flies (n = 22). Overexpression of the miR-33 by elav-Gal4 fully rescued disrupted axons in miR-33 ^KO flies (elav > miR-33, miR-33 ^KO, n = 17), and overexpression of the miR-33 by elav-Gal4 caused broken commissures (arrowhead) and connectives (arrows) (elav > miR-33, n = 29). B Confocal images of NMJ boutons at muscles 6/7 in L3 larvae of indicated genotypes were visualized with anti-HRP (red) and anti-DLG (green) antibodies staining. Scale bar = 20 μm. The bouton number (C), branch length (D), and branch number (E) per M6/7 NMJ in (B) were quantified, they were all significantly increased in miR-33 ^KO flies (n = 18) compared with w ¹¹¹⁸ (n = 15), and in elav > miR-33 RNAi flies (n = 15) compared with elav > + control (n = 15), and decreased in elav > miR-33 flies (n = 31) compared with elav > + control. Overexpression of the miR-33 by elav-Gal4 fully rescued abnormal growth of NMJs in miR-33 ^KO flies (elav > miR-33, miR-33 ^KO, n = 18). Data are mean ± SEM, **** p < 0.0001, n.s., no significance. One-way ANOVA test with Tukey’s post hoc test

miR-33 regulates neuron development by targeting RhoGEF Trio

miRNAs execute their functional roles through post-transcriptional regulation of protein-coding genes. The potential target(s) of the miR-33 was analyzed using the Targetscan, and the trio gene was identified as a candidate target of the miR-33. The Ras homologous (Rho) guanine nucleotide exchange factor (GEF) Trio is a central regulator of neuron development and synaptic plasticity, and aberrant activity of the Trio is implicated in neurodevelopmental and neurodegenerative disorders [33]. The trio 3’ UTR harbors two overlapping sites matching the seed-sequence of the miR-33 (Fig. 4A). Dual-luciferase assays confirmed the miR-33-dependent repression of a wild-type trio 3’ UTR reporter (p < 0.001, Fig. 4B), which was abolished by mutagenesis of either matching-site of the miR-33 seed-sequence (Fig. 4B). In vivo validation in wing discs revealed reduced DsRed reporter expression in the posterior region of wing discs upon GFP-miR-33 overexpression driven by hh-Gal4 (Fig. 4C), and endogenous Trio protein level was greatly reduced in the posterior region of wing discs when a different miR-33 transgene expression was induced by hh-Gal4 (Fig. 4D). Moreover, the Trio protein level was decreased in elav-miR-33 flies (Fig. 4E) and increased in brains of miR-33 ^KO flies (Fig. 4F). Together, these data proved that the RhoGEF Trio is a direct miR-33 target.

Fig. 4.

miR-33 mediates neurodevelopment via targeting Trio. A Predicted binding-sites for the miR-33 seed-sequence in the trio 3’ UTR. The miR-33 seed-sequence and mutagenesis of binding-sites for the miR-33 seed-sequence in the trio 3’ UTR are shown in red and blue, respectively. B Luciferase assay showing activity of a firefly luciferase reporter containing trio 3’ UTR (trio 3’ UTR ^wt) or mutated trio 3’ UTR (trio 3’ UTR ^mut1 and trio 3’ UTR ^mut2) treated with a miR-33 expression, the miRNA lin4 from c. elegans was used as a negative control. Data are mean ± SEM from three independent experiments. *** p < 0.001, n.s., no significance. Two-tailed two-sample t test. C A DsRed reporter expression of the tub-DsRed-trio 3’ UTR in wing discs showing reduced expression in the posterior region when a GFP-miR-33 overexpression was driven by a hh-Gal4. D Reduced Trio protein level was detected by an anti-Trio antibody in the posterior region of wing discs when a miR-33 overexpression was driven by a hh-Gal4. E-F Western blot detection of the Trio protein level in adult brains showing decreased Trio level in elav > miR-33 flies compared with elav > + flies (E), increased Trio level in miR-33 ^KO flies compared with w ¹¹¹⁸ flies (F). Alpha Tubulin was used as a loading control. Protein level was analyzed using the Image J, data are from three independent experiments. G CNS axons in stage 15–16 embryos of indicated genotypes were visualized with mAb BP102 staining. Scale bar = 10 μm. Broken commissures (arrowhead) and connectives (arrows) in miR-33 ^KO flies were greatly rescued by heterozygous trio ^1/+ (trio ^1/+, miR-33 ^KO, n = 16) or fully rescued by decreased expression of the trio gene by elav-trio RNAi (elav-trio RNAi, miR-33 ^KO, n = 16). Broken commissures (arrowhead) and connectives (arrows) in elav > miR-33 flies were fully rescued by overexpression of the trio (elav > miR-33/trio, n = 16). Commissures (arrowhead) and connectives (arrows) were broken in heterozygous trio ^1/+ flies (n = 18), elav-trio RNAi flies (n = 23) and elav > trio flies (n = 27). (H) Confocal images of NMJ boutons at muscles 6/7 in L3 larvae of indicated genotypes were visualized with anti-HRP (red) and anti-DLG (green) antibodies staining. Scale bar = 20 μm. The bouton number (I), branch length (J), and branch number (K) per M6/7 NMJ in (H) were quantified. Increased bouton number, branch length and branch number in miR-33 ^KO flies were fully rescued by heterozygous trio ^1/+ (trio ^1/+, miR-33 ^KO, n = 18) or by elav-trio RNAi (elav-trio RNAi, miR-33 ^KO, n = 18). Decreased bouton number, branch length and branch number in elav > miR-33 flies were fully rescued by overexpression of the trio gene (elav > miR-33/trio, n = 18). Heterozygous trio ^1/+ flies (n = 15) and elav-trio RNAi flies (n = 25) had decreased bouton number, branch length and branch number, and elav > trio flies (n = 15) had increased bouton number, branch length and branch number. Data are mean ± SEM, **** p < 0.0001, n.s., no significance. One-way ANOVA test with Tukey’s post hoc test

Next, we investigated whether altered expression of the Trio can rescue disrupted axons and abnormal synaptic growth in miR-33 ^KO flies and elav-miR-33 flies. Our results indicated that disrupted axons in miR-33 ^KO flies were greatly rescued by haploinsufficiency of the trio gene (trio ^1/+, miR-33 ^KO, n = 16, Fig. 4G, Table S2), or fully rescued by knockdown of trio expression induced by elav-Gal4 (elav-trio RNAi, miR-33 ^KO, n = 16, Fig. 4G, Table S2), and overexpression of the trio gene by elav-Gal4 also fully rescued disrupted axons in the elav-miR-33 flies (elav > miR-33/trio, n = 16, Fig. 4G, Table S2). Moreover, synaptic overgrowth in miR-33 ^KO flies was fully rescued by haploinsufficiency of the trio gene (n = 18, p < 0.0001, Fig. 4H-K) or elav-trio RNAi (n = 18, p < 0.0001, Fig. 4H-K), and synaptic undergrowth in elav-miR-33 flies was rescued by elav-trio (n = 18, p < 0.0001, Fig. 4H-K). The disrupted axons and synaptic undergrowth were also detected in elav-trio RNAi flies (axons, n = 23, Fig. 4G, Table S2; synapses, n = 25, p < 0.0001, Fig. 4H-K), trio ^1/+ heterozygous flies (axons, n = 18, Fig. 4G, Table S2; synapses, n = 15, p < 0.0001, Fig. 4H-K), and trio ^1/s137203 transheterozygous mutant flies (Figure S9), which were consistent with previous studies [34–36]. Existence of disrupted axons and abnormal growth of synapses in trio ^1/+ heterozygous flies, elav-trio RNAi flies and elav-trio flies (Fig. 4G-K) suggested that the dosage of the Trio is important for neurodevelopment, and it is also consistent with a previous study showing that the Trio and Abl have a dosage-sensitive genetic interaction [37]. In addition, disrupted mushroom bodies were detected in trio mutant adult flies (Figure S9), which is different from that of e2f1 mutant flies and miR-33 ^KO flies. Altogether, the RhoGEF Trio is a downstream target of the miR-33 to regulate neurodevelopment.

Trio is the downstream effector of E2f1 signaling to regulate neuron development

As the miR-33 is directly regulated by the E2f1 transcription factor and the trio gene is a target of the miR-33 to regulate neuron development, one critical question is whether the RhoGEF Trio is the downstream effector of the E2f1 signaling in neuron development. We investigated whether disrupted axons and abnormal growth of synapses in e2f1 mutant flies and elav ^ts-e2f1 flies can be rescued by altered expression of the Trio. Genetic epistasis experiments revealed that elav-driven trio knockdown (elav > trio RNAi, e2f1 ^07172/i2, n = 17, Fig. 5A, Table S2) rescued disrupted axons in e2f1 mutants, while disrupted axons in elav ^ts-e2f1 flies were greatly rescued by elav-driven trio overexpression (elav ^ts > e2f1/trio, n = 17, Fig. 5A, Table S2). Moreover, increased bouton number, branch length and branch number in e2f1 mutant flies were fully rescued by elav-trio RNAi (n = 16, p < 0.0001, Fig. 5B-E), and decreased bouton number, branch length and branch number in elav ^ts-e2f1 flies were fully rescued by elav ^ts-trio (n = 18, p < 0.0001, Fig. 5B-E). To provide more evidence that the E2f1 transcription factor modulates the expression level of the Trio protein, E2f1 was overexpressed in the posterior region of wing discs by hh-Gal4, and an anti-Trio antibody staining indicated that the Trio protein level was reduced in the posterior region (Figure S10). Altogether, the RhoGEF Trio is the downstream effector of the E2f1 signaling to regulate neuron development.

Fig. 5.

Trio works as the effector of E2f1 signaling in neuron development. A CNS axons in stage 15–16 embryos of indicated genotypes were visualized with mAb BP102 staining. Scale bar = 10 μm. Knockdown of the trio expression by elav > trio RNAi fully rescued disrupted axons in e2f1 ^07172/i2 mutant flies (elav > trio RNAi, e2f1 ^07172/i2, n = 17), and overexpression of the trio gene by elav ^ts-Gal4 greatly rescued disrupted axons in elav ^ts > e2f1 flies (elav ^ts > e2f1/trio, n = 17). B Confocal images of NMJ boutons at muscles 6/7 in L3 larvae of indicated genotypes were visualized with anti-HRP (red) and anti-DLG (green) antibodies staining. Scale bar = 20 μm. The bouton number (C), branch length (D), and branch number (E) per M6/7 NMJ in (B) were quantified. Knockdown of the trio expression by elav > trio RNAi fully rescued abnormal growth of NMJs in e2f1 ^07172/i2 mutant flies (elav > trio RNAi, e2f1 ^07172/i2, n = 16), and overexpression of the trio gene by elav ^ts-Gal4 greatly rescued abnormal growth of NMJs in elav ^ts > e2f1 flies (elav ^ts > e2f1/trio, n = 18). Data are mean ± SEM, * p < 0.05, **** p < 0.0001. One-way ANOVA test with Tukey’s post hoc test

E2f1-miR-33-Trio signaling is conserved in Drosophila and human

The E2f1 transcription factor and the RhoGEF Trio are highly conserved. Cross-species analysis of the miR-33 seed-sequence revealed that the intronic miR-33 within the SREBP locus is conserved from Drosophila to human (Fig. 6A). To explore whether the E2f1-miR-33-Trio regulatory axis is conserved in human, we analyzed human miR-33 (hmiR-33) expression in HEK-293T cells. Quantitative RT-PCR showed that hmiR-33 was significantly downregulated upon human E2f1 (hE2f1) knockdown and upregulated upon hE2f1 overexpression (Fig. 6B-C and S11). In silico promoter analysis using PROMO identified four putative E2f-binding sites located ~ 2 kb upstream of the hmiR-33 locus (Fig. 6D). Chromatin immunoprecipitation (ChIP) with an anti-hE2f1 antibody confirmed specific enrichment of hE2f1 protein at site #4, with no binding detected at other sites (Fig. 6E). As a positive control, the hE2f1 protein was highly enriched at the promoter region of the PCNA gene, but not enriched at the promoter region of the negative control GAPDH gene (Fig. 6E). Cross-species alignment of the trio 3’ UTR revealed conservation of the binding-site of the miR-33 seed-sequence (Fig. 6F). TargetScan analysis predicted a single hmiR-33 binding-site in the htrio 3’ UTR (Fig. 6G). Luciferase reporter assays showed that hmiR-33 mimics significantly reduced wild-type htrio 3’ UTR reporter activity (P < 0.001, Fig. 6H). When the binding-site of the hmiR-33 seed-sequence in the htrio 3’ UTR was mutated, the luciferase activity of the reporter containing mutated htrio 3’ UTR showed no significant difference compared with the control (Fig. 6H). Collectively, these data suggested that the E2f1-miR-33-Trio regulatory axis is conserved in human.

Fig. 6.

Conserved regulations among human homologs of E2f1, miR-33 and Trio. A Scheme of the miR-33 sequence embedded in its host-gene SREBP and conservation in multiple species during evolution. B-C Relative expression of the hmiR-33 in HEK-293T cells transfected with shRNA sh-he2f1 or the control sh-NC (B), and transfected with pCDH-CMV-MCS-EF1-he2f1 or with the control pCDH-CMV-MCS-EF1 vector (C), was detected via RT-qPCR. Data were normalized with hU6 as an internal control. Expression of the hmiR-33 was decreased in sh-he2f1 treated cells compared with the control sh-NC (B), and increased in pCDH-CMV-MCS-EF1-he2f1 transfected cells compared with the control pCDH-CMV-MCS-EF1 vector transfected cells (C). ** p < 0.01, two-tailed two-sample t test. D Scheme of the hmiR-33 promoter region. Putative E2f-sites are depicted, and nucleotide positions of E2f-sites are provided with respect to the hmiR-33 transcription start site. E ChIP-qPCR analysis revealed significant binding of the hE2f1 protein to site #4 in the hmiR-33 promoter region, and the hPCNA promoter region as a positive control, and no binding to site #1, #2 or #3 in the hmiR-33 promoter region and the negative control hGAPDH promoter region. *** p < 0.001, **** p < 0.0001, n.s., no significance. Two-tailed two-sample t test. F Predicted binding-site for the miR-33 seed-sequence in the trio 3’ UTR in multiple species. G The binding-site of the hmiR-33 seed-sequence in the htrio 3’ UTR and mutagenesis of the binding-site are shown in red and blue, respectively. H Luciferase assay showing activity of a luciferase reporter containing htrio 3’ UTR (htrio 3’ UTR ^wt) or mutated htrio 3’ UTR (htrio 3’ UTR ^mut) treated with hmiR-33 mimics, or a negative control NC mimics. Data are mean ± SEM from three independent experiments. *** p < 0.001, n.s., no significance. Two-tailed two-sample t test

Defective neurodevelopment in flies can be rescued by human homologs of miR-33 and Trio

We further investigated whether disrupted axons and abnormal growth of synapses in e2f1 mutant flies, miR-33 ^KO flies, elav-miR-33 flies, and trio mutant flies could be rescued by human homologs of the miR-33 or Trio. We established “humanized” transgenic flies of the precursor-hmiR-33 and htrio cDNA, respectively. By performing rescue experiments, we found that elav-driven hmiR-33 expression greatly rescued disrupted axons in e2f1 mutant flies (elav > hmiR-33, e2f1 ^07172/i2, n = 15, Fig. 7A, Table S2), and fully rescued disrupted axons in miR-33 ^KO flies (elav > hmiR-33, miR-33 ^KO, n = 15, Fig. 7A, Table S2). Concomitantly, synaptic overgrowth in e2f1 mutant flies (elav > hmiR-33, e2f1 ^07172/i2, n = 16, p < 0.0001, Fig. 7B-E) and miR-33 ^KO flies (elav > hmiR-33, miR-33 ^KO, n = 15, p < 0.0001, Fig. 7B-E) was fully rescued by the elav-driven overexpression of the hmiR-33. Moreover, elav-driven htrio overexpression greatly rescued disrupted axons in elav-miR-33 flies (elav > miR-33/htrio, n = 15, Fig. 7F, Table S2), and fully rescued synaptic undergrowth in elav-miR-33 flies (n = 17, p < 0.0001, Fig. 7G-J). Furthermore, htrio overexpression greatly rescued disrupted axons (elav > htrio, trio ^1/s137203, n = 15, Fig. 7F, Table S2) and fully rescued synaptic undergrowth (n = 16, p < 0.0001, Fig. 7G-J) in trio mutants. These results are aligned with a recent study showing that htrio expression restores changed bouton size and fragmentation in aged trio knock-down flies [25], reinforcing Trio’s conserved roles in neurodevelopment. Together, these findings demonstrate that not only the E2f1-miR-33-Trio signaling axis is evolutionary conserved, but also their functions to preserve neurodevelopment across the Drosophila and human.

Fig. 7.

Conserved functions of human miR-33 and Trio in neuron development. A and F CNS axons in stage 15–16 embryos of indicated genotypes were visualized with mAb BP102 staining. Scale bar = 10 μm. B and G Confocal images of NMJ boutons at muscles 6/7 in L3 larvae of indicated genotypes were visualized with anti-HRP (red) and anti-DLG (green) antibodies staining. Scale bar = 20 μm. A Broken commissures (arrowhead) and connectives (arrows) in e2f1 ^07172/i2 flies were greatly rescued by overexpression of a human homolog of the miR-33 (elav > hmiR-33, e2f1 ^07172/i2, n = 15), and broken commissures (arrowhead) and connectives (arrows) in miR-33 ^KO flies were fully rescued by overexpression of the hmiR-33 (elav > hmiR-33, miR-33 ^KO, n = 15). Overexpression of the hmiR-33 (elav > hmiR-33, n = 20) caused broken commissures and connectives in flies. The bouton number (C), branch length (D), and branch number (E) per M6/7 NMJ in (B) were quantified. Increased bouton number, branch length and branch number in e2f1 ^07172/i2 mutant flies were fully rescued by overexpression of the hmiR-33 (elav > hmiR-33, e2f1 ^07172/i2, n = 16), and increased bouton number, branch length and branch number in miR-33 ^KO flies were fully rescued by overexpression of the hmiR-33 (elav > hmiR-33, miR-33 ^KO, n = 15). Overexpression of the hmiR-33 (elav > hmiR-33, n = 24) decreased bouton number, branch length and branch number compared with elav > + control flies. Data are mean ± SEM, **** p < 0.0001, n.s., no significance. One-way ANOVA test with Tukey’s post hoc test. F Broken commissures (arrowhead) and connectives (arrows) in elav > miR-33 flies were greatly rescued by overexpression of a human homolog of the trio gene (elav > miR-33/htrio, n = 15), and broken commissures (arrowhead) and connectives (arrows) in trio mutant flies were greatly rescued by overexpression of the htrio gene (elav > htrio, trio ^1/s137203, n = 15). Overexpression of the htrio gene (elav > htrio, n = 24) caused broken commissures and connectives in flies. The bouton number (H), branch length (I), and branch number (J) per M6/7 NMJ in (G) were quantified. Decreased bouton number, branch length and branch number in elav > miR-33 flies were fully rescued by overexpression of the htrio gene (elav > miR-33/htrio, n = 17), and decreased bouton number, branch length and branch number in trio mutant flies were fully rescued by overexpression of the htrio gene (elav > htrio, trio ^1/s137203, n = 16). Overexpression of the htrio gene in flies (elav > htrio, n = 17) increased bouton number, branch length and branch number compared with elav > + control flies. Data are mean ± SEM, **** p < 0.0001, n.s., no significance. One-way ANOVA test with Tukey’s post hoc test

Aberrant locomotor behaviors in e2f1 and miR-33 mutant flies

Given disrupted axons and abnormal synaptic growth in e2f1 mutant flies and miR-33 ^KO flies, we further studied whether these defects impact locomotor behavior. We performed behavior assays to examine locomotion activity, patterns and rhythms at third instar larval stage. Using a grid crossing assay, we found that e2f1 mutant flies showed significantly decreased number of grids crossed (Fig. 8A) and significantly reduced peristaltic contractions compared with w ¹¹¹⁸ flies (Fig. 8B), while elav ^ts-e2f1 rescued decreased number of grids crossed (Fig. 8A) and reduced peristaltic contractions of e2f1 mutant flies (Fig. 8B). Similarly, miR-33 ^KO flies showed impaired grid crossing (Fig. 8C) and reduced peristalsis (Fig. 8D), which were rescued by elav-driven miR-33 expression (Fig. 8C, D). Furthermore, overexpression of the trio gene by elav-Gal4 rescued locomotor deficits in trio mutants, including reduced grid crossings (Fig. 8E) and peristaltic activity (Fig. 8F), consistent with the previous study showing trio haploinsufficiency mice have impaired motor coordination [38]. Collectively, these data demonstrate that E2f1 dysfunction or miR-33 loss causes aberrant locomotor behaviors, linking the E2f1-miR-33-Trio axis to motor control in Drosophila.

Fig. 8.

e2f1, miR-33 and trio mutant flies exhibit aberrant locomotor behaviors. Locomotor behaviors of L3 larvae of indicated genotypes were assayed by measuring numbers of 0.5 cm² grids crossed in 30 s (A, C and E), and numbers of full-body peristaltic contractions in 1 min (B, D and F). Data were acquired from 30 samples. A Numbers of grids crossed were reduced of e2f1^07172/i2 mutant flies compared with w¹¹¹⁸flies, and elav^ts> e2f1 flies compared with elav^ts> + control flies, and overexpression of e2f1 by elav^ts-Gal4 rescued motor deficitsin e2f1^07172/i2 mutant flies (elav^ts> e2f1, e2f1^07172/i2). B Numbers of full-body peristaltic contractions were reduced of e2f1^07172/i2 mutant flies compared with w¹¹¹⁸ flies, and elav^ts> e2f1 flies compared with ela ^ts> + control flies, and overexpression of e2f1 by elav^ts-Gal4 rescued motor deficits in e2f^07172/i2 mutant flies (elav^ts> e2f1, e2f1^07172/i2). C Numbers of grids crossed were reduced of miR-33^KO flies compared with w¹¹¹⁸ flies, and elav > miR-33 flies compared with elav > + control flies, and overexpression of miR-33 by elav-Gal4 rescued motor deficits in miR-33^KO flies (elav > miR-33, miR-33^KO). D Numbers of full-body peristaltic contractions were reduced of miR-33^KO flies compared with w¹¹¹⁸ flies, and elav > miR-33 flies compared with elav > + control flies, and overexpression of miR-33 by elav-Gal4 rescued motor deficits in miR-33^KO flies (elav > miR-33, miR-33^KO). E Numbers of grids crossed were reduced of trio^1/s137203 mutant flies compared with w¹¹¹⁸ flies, and elav > trio flies compared with elav > + control flies, and overexpression of trio by elav-Gal4 rescued motor deficits in trio^1/s137203 mutant flies (elav > trio, trio^1/s137203). F Numbers of full-body peristaltic peristaltic contractions were reduced of trio^1/s137203 mutant flies compared with w¹¹¹⁸ flies, and elav > trio flies compared with elav > + control flies, and overexpression of trio by elav-Gal4 rescued motor deficits in trio^1/s137203 mutant flies (elav > trio, trio^1/s137203). Data are mean ± SEM. **** p < 0.0001, n.s., no significance. One-way ANOVA test with Tukey’s post hoc test

Discussion

In this study, we uncover a previously unrecognized role for the E2f1 transcription factor in Drosophila neurodevelopment, extending beyond its canonical functions in cell cycle control, DNA damage response, and apoptosis. The e2f1 mutant flies and flies with e2f1-reduced expression or -overexpression in neurons have disrupted axons and abnormal synaptic growth. E2f1 directly regulates miR-33, a conserved metabolic regulator with hitherto unknown roles in the central nervous system. The RhoGEF Trio, a critical neurodevelopmental regulator, serves as a downstream target of the miR-33 and the effector of E2f1 signaling in neurons. Notably, the E2f1-miR-33-Trio signaling axis and its functions in the CNS are evolutionarily conserved between Drosophila and human (Fig. 9).

Fig. 9.

A model. The E2f1/miR-33/Trio signaling is critical for neurodevelopment and conserved between Drosophila and human. Loss or dysregulated expression of any member in the signaling disrupts axon patterning, causes abnormal synaptic growth at neuromuscular junctions (NMJs), and eventually leads to aberrant locomotor behaviors. Keeping the activity and well-balanced dosage of the E2f1/miR-33/Trio signaling is essential for neuron development and synaptic plasticity 

We demonstrate that e2f1-mutant flies exhibit neurodevelopmental defects that can be observed as early as embryogenesis and throughout the third instar larval stage, whereas the gross morphology of brains in L3 larvae or adults and mushroom bodies in adults looks normal. In accordance with disrupted neurons at embryonic stage and abnormal growth of synapses in L3 larvae, e2f1 mutant flies have aberrant locomotor behaviors. Studies in mice have indicated that e2f1 mutant mice have normal brains development at the adult stage, although numbers of newly generated granule cells were mildly decreased in the dentate gyrus and the olfactory bulb [23], as well as age-dependent olfactory and memory-related deficits [22]. Knockout of the e2f4 gene in mice causes loss of ventral telencephalic structures and impaired self-renewal of neural precursor cells [39]. The two isoforms of the e2f3 gene, E2f3a and E2f3b, work antagonistically to regulate expression of the pluripotency factor Sox2 in neurodevelopment, and loss of the e2f3a in mice results in defects in hippocampal neurogenesis and memory formation [40]. The mammalian E2f family has multiple members with overlapping roles, complexed regulatory networks, and diversified functions in different organs. These complexed contexts in mammals hinder the study to clarify the role of a single E2f factor in the central nervous system. With a simpler E2f family, only two members of E2f1 and E2f2, the Drosophila melanogaster will bring a lot of benefits to explore functions of the E2f factors, and gain a deeper understanding towards their relations with neurodevelopment or neurodegenerative disorders. We have analyzed publicly available RNA-SEQ data from patients with neurodegenerative disorders, but we failed to locate the e2f1 mutation. Whether mutations of the E2f1 or other members of the E2f family are related with neurodevelopmental or neurodegenerative disorders, beyond their dysregulated expression or activities, needs further study.

The miR-33 has been proved to play important regulatory roles in lipid and glucose metabolisms [31], and is inferred to neurological disorders due to its roles on lipid metabolisms [32]. Our study reveals previously unrecognized functions of the miR-33 in neurodevelopment, including a dosage-dependent regulatory role, an expected feature of a miRNA. Our findings expand the known functions of the miR-33 to the central nervous system (CNS) and suggest a potential link between its metabolic and neurodevelopmental roles. Drosophila offers a powerful system to decipher interrelations between neurological disorders and cardiovascular, metabolic, and inflammatory diseases related with the dysregulated miR-33.

The RhoGEF Trio serves as an evolutionarily conserved master regulator of the neurodevelopment. Its functions have been extensively characterized in C. elegans, Drosophila, mice, and cultured primary neurons. The Trio protein is expressed throughout development, and displays a dynamic expression pattern from early developmental stages to adulthood [33]. The trio gene has many isoforms due to alternative splicing, and different Trio isoforms might play distinct roles to regulate Rac1 and RhoA activity [41]. A recent study found that the Drosophila RNA-binding protein Nab2 and Mettl3 methyltransferase regulate Trio splicing and protein level to support nervous system development [42]. And tyrosine phosphorylation of the Trio protein by the Src family kinase Fyn is required for Rac1 activation [43]. Moreover, the Trio activity is enhanced by a molecular chaperone heat shock cognate protein 70-Hsc70 [44]. Despite these studies, the upstream regulatory signaling governing Trio activation and expression remain poorly understood.

The trio knockout mice, neural-specific knockout mice or haploinsufficiency mice develop abnormal brains [38, 45–47]. In Drosophila, trio-mutant neurons display defects from embryogenesis to adulthood, affecting CNS structures including axons, neuromuscular junctions, and mushroom bodies [34, 35, 37]. The trio gene interacts genetically with the rac gene or the abl gene in a dosage-sensitive manner [35, 37], and the Trio level in motor synapses declines with age [25]. Our data also indicated that manipulation of the Trio dosage change the severity of mushroom body defects (Figure S12). As the E2f1 transcription factor shows a dosage effect towards cell cycle, checkpoint and apoptosis [48], and miRNAs shows a global dosage control to achieve accurate post-transcriptional regulation [49], the tempo-spatial regulation among the E2f1-miR-33-Trio might be critical for plasticity of the neuron functions. Our study about E2f1/miR-33 functions in the CNS and their regulatory roles upstream of the RhoGEF Trio will provide more clues to understand aberrant expression and activity of E2f1 and Trio in neurodegenerative disorders.

Conclusions

In this study, we demonstrate that dysfunctional E2f1 or dysregulated expression of E2f1 in neurons leads to disrupted axons in embryos and aberrant synaptic growth at larval neuromuscular junctions in Drosophila. The miR-33 acts downstream of the E2f1 to mediate the E2f1 signaling in neurons. Furthermore, the RhoGEF Trio works as a target of the miR-33 and the effector of the E2f1 signaling to regulate neurodevelopment. Moreover, the E2f1-miR-33-Trio signaling and its functions in neurons are conserved in Drosophila and human. Together, our study discloses a novel conserved regulatory mechanism of the Trio in neurons by the E2f1/miR-33 signaling and provides new knowledge to understand how the RhoGEF Trio is regulated.

Abbreviations

miRNAs

MicroRNAs

PD

Parkinson′ s disease

AD

Alzheimer disease

polyQ

Polyglutamine

WT

Wild-type

CNS

Central nervous system

L3

Third instar larvae

NMJs

Neuromuscular junctions

HRP

Horseradish Peroxidase

ts

Temperature sensitivity

ChIP

Chromatin immunoprecipitation

rp49

Ribosomal protein 49

SREBP

Sterol-regulatory element-binding protein

Rho

Ras homologous

GEF

Guanine nucleotide exchange factor

3′ UTR

3′ Untranslated region

hmiR-33

Human miR-33

hE2f1

Human E2f1

PCNA

Proliferating cell nuclear antigen

GAPDH

Glyceraldehyde-3-phosphate dehydrogenase

RT-qPCR

Real-time Quantitative PCR

gDNA

Genomic DNA

pre-miRNA

Precursor miRNA

shRNA

Short hairpin RNA

NC

Negative control

VNC

Ventral nerve cord

A3

Abdominal segment 3

Authors’ contributions

Conceptualization: L.W. and X.B. Methodology: L.W., B.Z., D.L., and X.B. Investigation: L.W., B.Z., Y.Y., X.Y., L.Z., L.B., and Y.S. Formal analysis: L.W. and B.Z. Visualization: L.W. Supervision: X.B. Funding acquisition: X.B. Resources: X.B. Project administration: X.B. Writing—original draft: L.W. and X.B. Writing—revision: L.W. and X.B.

Funding

This work was supported by grants from National Natural Science Foundation of China Grant No. 32270596 and 31970605 to X.B.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.