ASCL1 protein domains with distinct functions in neuronal differentiation and subtype specification

Abstract

ASCL1 is a neural basic helix-loop-helix (bHLH) transcription factor that plays essential roles during neural development, including neural differentiation and neuronal subtype specification. bHLH factors are defined by their motifs, including a basic region interacting with DNA and an HLH domain involved in protein-protein interactions. We previously defined specific regions within the bHLH domain of ASCL1 as important for its specific functions directing neuronal differentiation in the chick neural tube. Here, we build upon these findings to show how specific mutations within the basic region block DNA binding but not heterodimer formation with E-protein partners TCF3 (E12/E47) and TCF12 (HEB) yet have differential abilities to show dominant negative phenotypes. Additionally, truncating domains outside the bHLH define a nuclear localization signal, a requirement for the C-terminal acidic residues, and the non-essentiality of the N-terminal glutamine/alanine repeats. This structure/function analysis identifies functional domains for ASCL1 activity.

Keywords: ASCL1; Neurogenesis; Neuronal specification; Spinal cord development; bHLH transcription factors.

Fig. 1.

Diagrams of the chick in ovo electroporation assay and phenotypes with overexpression of ASCL1. ASCL1 (from Rattus norvegicus) or ASCL1 mutant constructs were overexpressed in chick neural tubes using an in ovo electroporation assay to assess the effect on neuronal proliferation and subtype specification. (A) DNA is injected into the neural tube of chick embryos, and electrodes placed on either side of the embryo pass a current such that the cells within the ventricular zone (VZ) on the side of the positive electrode (+) take up the DNA (Timmer et al., 2001). (B) Overexpression of ASCL1 in the neural tube drives the cells to differentiate as detected by a loss of BrdU incorporation, migration lateral to the mantle zone (MZ), and induction of the neuronal marker RBFOX3. In addition, ASCL1 alters the composition of dorsal interneuron (dI) cell populations by increasing dI3 and dI5 while decreasing dI1, dI2, dI4, and dI6 (Nakada et al., 2004). (C) ASCL1 amino acid sequence (from Rattus norvegicus) containing an N-terminal Q/A-rich domain (in gray), a nuclear localization domain (NLS, in cyan), the basic domain for DNA binding, the helix-loop-helix domain for dimerization with other bHLH proteins, and a short acidic domain on the C-terminus (in red). The numbers represent amino acid residues and indicate truncation boundaries further shown in panel E. (D) Heterodimer protein structure ASCL1 (in green) and TCF12 (in blue) binding a short DNA oligo with an E-box binding motif generated in AlphaFold3 (Abramson et al., 2024). Disordered regions of TCF12 are hidden for a clearer view. ASCL1 protein domains and basic region mutation residues are labeled in the complex. Residues R127 and E124 face toward the DNA groove, while N126 faces away from the DNA groove. (E) Diagram of basic domain site mutations and protein truncations of ASCL1 used in this study. The basic region mutations are indicated by magenta letters: ASCL1^NR–AQ (designated ASCL1^N126A;R127Q in the text); ASCL1^E–G (ASCL1^E124G); ASCL1^N–A (ASCL1^N126A); ASCL1^R–Q (ASCL1^R127Q); ASCL1^{E–G/NR–AQ} (ASCL1^{E124-G/N126A;R127Q}). For the truncations, three mutants were made with either the N-terminal Q/A-rich region deleted (dN1, Δ aa 2–72), the Q/A-rich region and the NLS deleted (dN2, Δ aa 2–111), or the C-terminal acidic domain deleted (dC1, Δ aa 221–233). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2.

Requirement for specific amino acids in the basic domain for ASCL1 activity in neuronal differentiation. HH14–16 chick neural tubes were electroporated in ovo with constructs expressing MYC-tagged vectors: control empty vector (A, A′), ASCL1 (B, B′), ASCL1^NR–AQ (C, C′), ASCL1^E–G (D, D′), ASCL1^N–A (E, E′), ASCL1^R–Q (F, F′), or ASCL1^{E–G/NR–AQ} (G, G′), and harvested 24 h later (HH23–24). A cartoon of a chick embryo and a transverse section of the neural tube is shown in the top middle, with the boxed region representing that imaged in A-G’ (+ is the electroporated side). (A-G′) Immunofluorescence with anti-MYC antibody indicating transgene expression in the electroporated cells (red). BrdU incorporation was used to detect proliferating cells (A-G, green), and RBFOX3 identifies differentiating neurons (A′-G′, green). (H–I) The percentage of electroporated cells (red) co-expressing BrdU (H) and RBFOX3 (I) is shown. ASCL1 induces cells to move laterally out of the ventricular zone, express the neuronal marker RBFOX3, and exit the cell cycle. Black asterisks indicate those cases significantly different from wild-type ASCL1 overexpression. Red asterisks highlight the mutants that decrease co-expression with RBFOX3 relative to control (I). Each data point represents a biological replicate, and the error bars are SEM around the mean, unpaired t-test, **p ≤ 0.001, *p ≤ 0.01, ns = not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3.

Requirement for specific amino acids in the basic domain for ASCL1 activity in neuronal subtype specification of dorsal interneurons. HH14–16 chick neural tubes were electroporated in ovo with constructs expressing MYC-tagged vectors: control empty vector (A-A″), ASCL1 (B-B″), Ascl1^NR–AQ (C-C″), ASCL1^E–G (D-D″), ASCL1^N–A (E-E″), ASCL1^R–Q (F-F″), or ASCL1^{E–G/NR–AQ} (G-G″), and harvested 24 h later (HH23–24). Dorsal neural tube diagrams above the images indicate the dorsal interneuron populations assayed with each set of markers, with the boxed region representing that imaged in A-G″. Immunofluorescence for LHX2/9 (red) identifies dI1 and ISL1 (green) identifies dI3 (A–G), LHX1/5 (green) identifies dI2 and PAX2; LHX1/5 (yellow) identifies dI4/6 (A′-G′), and LMX1B (green) identifies dI5 (A″-G″). In each case, the electroporated side (+) is on the right and is compared with the control side (−) on the left. The white line indicates the midline. (H–L) The quantification of each population by comparing the electroporated side to the control side of the embryo is shown. Black asterisks indicate those cases significantly different from wild-type ASCL1 overexpression. Red asterisks highlight the mutants that decrease ISL1 expression (dI3) relative to control (H). Each data point represents a biological replicate, and the error bars are SEM around the mean, unpaired t-test, **p ≤ 0.001, *p ≤ 0.01, ns = not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4.

Co-immunoprecipitation of ASCL1 or basic region mutants with the E-protein TCF12. MYC-tagged ASCL1 (MYC-ASCL1) and its mutants (Fig. 1E) were co-transfected with MYC-tagged TCF12 (MYC-TCF12) into COS-7 cells. Immunoprecipitation (IP) with antibodies to ASCL1 and western blotting (WB) with antibodies to MYC indicate all ASCL1 mutants can heterodimerize with the E-protein TCF12. The sizes indicated are in kilodaltons (kDa).

Fig. 5.

DNA binding ability of ASCL1 and the basic region mutants. EMSA (electrophoretic mobility shift assays) were performed using in vitro transcribed and translated ASCL1, ASCL1 mutants (defined in Fig. 1E), and the E-proteins TCF3 and TCF12 as indicated. Two sequences containing the bHLH binding motif (E-box), CAGATG (E5) or CACCTG (E2), derived from an ASCL1 target gene Dll3, were used (Henke et al., 2009). Specificity was assessed with cold competitor sequence with and without the E-box intact (E2 cold, E2M cold). All ASCL1 basic region mutants except ASCL1^N–A have dramatically decreased DNA binding activity relative to wild-type ASCL1 (WT) (lanes 3, 13, and 23). The bands for E-protein homodimers (TCF3 and TCF12) and the ASCL1/E-protein heterodimers (ASCL1/TCF3 and ASCL1/TCF12) are indicated on the right.

Fig. 6.

Domains outside the bHLH are required for ASCL1 activity in cell-cycle exit and neuronal differentiation. HH14–16 chick neural tubes were electroporated in ovo with constructs expressing MYC-tagged vectors: ASCL1 (A, A′), ASCL1^dN1 (B, B′), ASCL1^dN2 (C, C′), or ASCL1^dC1 (D, D′), and harvested 24 h later (HH23–24). Cross sections through the neural tube are as shown in the diagram in Fig. 2. Immunofluorescence with anti-MYC antibody indicating transgene expression in the electroporated cells (red). BrdU incorporation identifies proliferating cells (A-D, green), and RBFOX3 detects differentiating neurons (A′-D′, green). (E–F) The percentage of electroporated cells (red) co-expressing BrdU (E) and RBFOX3 (F) is shown. ASCL1 induces cells to move laterally out of the ventricular zone, express the neuronal marker RBFOX3, and exit the cell cycle. Asterisks indicate those cases significantly different from wild-type ASCL1 overexpression. Each data point represents a biological replicate, and the error bars are SEM around the mean, unpaired t-test, **p ≤ 0.001, *p ≤ 0.01, ns = not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 7.

Domains outside the bHLH have variable importance to ASCL1 in neuronal subtype specification activity. HH14–16 chick neural tubes were electroporated in ovo with constructs expressing MYC-tagged vectors: ASCL1 (A, A″), ASCL1^dN1 (B, B″), ASCL1^dN2 (C, C″), or ASCL1^dC1 (D, D″), and harvested 24 h later (HH23–24). Dorsal neural tube diagrams above the images indicate the dorsal interneuron populations assayed with each set of markers, with the boxed region representing that imaged in A-D″. Immunofluorescence for LHX2/9 (red) identifies dI1 and ISL1 (green) identifies dI3 (A–D), LHX1/5 (green) identifies dI2 and PAX2; LHX1/5 (yellow) identifies dI4/6 (A′-D′), and LMX1B (green) identifies dI5 (A″-D″). In each case, the electroporated side (+) is on the right and is compared with the control side (−) on the left. The white line indicates the midline. (E–I) The quantification of each population by comparing the electroporated side to the control side of the embryo is shown Black asterisks indicate those cases significantly different from wild-type ASCL1 overexpression. Each data point represents a biological replicate, and the error bars are SEM around the mean, unpaired t-test, **p ≤ 0.001, *p ≤ 0.01, ns = not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)