Enhancer function is dependent on bHLH factors. (A) Two evolutionarily conserved blocks, M50 and M100, are found in the M250 enhancer. Mouse, rat and human sequences of M50 and M100 are aligned, and nucleotide coordinates of the mouse and human sequences are indicated. Two canonical E box sites are marked with asterisks (^*). (B) Mutations were introduced into the E boxes as shown, and functional activity was tested in the context of the M250+ΔNheI:GFP construct shown in Fig. 1B,F. (C,D) GFP expression in HH stage 24 chick embryos following electroporation of each reporter construct. (E,F) GFP expression in E11.5 transgenic mouse embryos containing wild-type or E box mutant constructs.

NeuroD contributes to the activation of Hb9. (A) P19 cells were transfected with a luciferase reporter containing reiterated E boxes and increasing amounts of expression constructs encoding bHLH factors. Ngn2 strongly activated transcription, whereas NeuroM (NeuM) and NeuroD (NeuD) exhibited markedly less inherent trans-activating function. (B) NeuroD activates M250 enhancer-mediated transcription in a synergistic manner with Isl1 and Lhx3 in P19 cells. (C) NeuroD/E47 dimer binding to the M250 enhancer was examined using gel retardation assays. Myc-tagged NeuroD [NrD(Myc)] and HA-tagged E47 were translated in vitro separately and incubated with the M50 oligonucleotide in the presence or absence of antibodies directed against each epitope. This translation condition favored the formation of E47:E47 complexes. Similar results were seen with oligos containing the M100 E box (not shown). (D) The binding of co-translated NeuroD/E47 on M50 was challenged by 20- or 100-fold molar excess of wild type (self), E-box mutated (E-mt) or unrelated (non-self) DNA. Co-translation of proteins favored the formation of NeuroD:E47 heterodimers, which bind with high specificity to the E box sites within the enhancer. (E-H) The co-expression of NeuroD with Isl1 and Lhx3 triggers the differentiation of Hb9⁺ motoneurons in >70% of the transfected P19 cells. This activity is comparable with that seen with NeuroM (Lee and Pfaff, 2003).

Identification of multiple positive-acting elements within the M250 enhancer for motoneuron expression. (A) Point mutations were introduced into nine sites (M1-M9) as indicated. (B-K) GFP expression in HH stage 24 chick embryos with each mutant reporter tested in the context of the M250+ΔNheI:GFP construct (see Fig. 1B,F). (L) E11.5 transgenic mouse embryo generated with the M250 reporter containing M1 and M9 mutations and stained with anti-GFP antibody. (M) Location of oligonucleotides (indicated by rectangles) in the Hb9 promoter used for gel retardation assays. The sequence of these regions is highly conserved between mouse and human. (N) The binding of Isl1/Lhx3 to the M50 DNA segment was challenged by 20-, 100-fold molar excess of the unlabeled probes indicated above the lanes. M50 and M100C competed efficiently for Isl1/Lhx3 DNA binding, while the M3-oligo and 3′CD-oligos failed to do so. This suggests the M3 element is not a binding site for the LIM-HD factors Isl1 and Lhx3.

General-activators of Hb9 in non-motoneurons. (A) Fragments of the mouse Hb9 promoter were linked to the reporter nuclear lacZ or GFP. (B-H) DNA constructs were electroporated into chick embryos and reporter expression was detected using immunocytochemistry at HH stage 24. (B) The 9.2 kb promoter of Hb9 drives motoneuron-specific gene expression. (C) Deletion of the distal 5.6 kb BamHI fragment impairs the motoneuron-specific activity leading to widespread ectopic reporter expression. (D,E) Likewise, large 5′ deletions to the NheI and HindIII sites leaving only 1386 and 550 nucleotides of Hb9 sequences, respectively, continued to promote widespread ectopic reporter expression. (F-H) The 2.5 kb distal element (–8129 to –5575) was linked to three different proximal fragments of Hb9 as indicated in A. 2.5 distal + ΔNheI or ΔHindIIII drove high levels of GFP expression in motoneurons, but 2.5 distal + NI-HIII was less active, suggesting the –550 proximal segment of Hb9 is necessary for enhancer function. (I) General activators are predicted to interact with the proximal (prox.) region of Hb9 and promote widespread transcription of the gene. The enhancer of Hb9 (MN^E) in combination with general-activators drive motoneuron-specific expression. Deletion of the proximal segment of Hb9 disrupts enhancer function. (J) The nucleotide sequence of the proximal 550 nucleotide region of Hb9 from mouse and human are aligned. The conserved sequences in both species are marked in bold, and potential binding sites for activators are boxed and labeled. This sequence lacks an obvious TATA box motif. The longest cDNA extends to –195 in mouse Hb9 (–181 in human). Because the precise start of transcription has not been mapped, we use the start of translation as the reference point for the coordinates shown with the sequences (see Materials and methods). (K) CV1 cells were transfected with increasing concentrations of plasmids encoding E2F1 and/or Sp1 together with a luciferase reporter linked to Hb9 (2.5 distal+ΔNheI).

Repressors interacting with dispersed elements suppress ectopic Hb9 expression. (A) Distal fragments were linked to the proximal– 1386 segment (ΔNheI) of Hb9 to map elements involved in repressing ectopic gene expression. (B-E) DNA clones were electroporated into chick embryos, and immunocytochemistry was used to detect lacZ (green) and Hb9 (red) at HH stage 24. White bar indicates dorsal limit of Hb9⁺ motoneuron population. (B) The proximal segment of Hb9 drives widespread reporter expression. (C) The 1 kb distal fragment does not restrict the ectopic expression of Hb9. (D) The 4.6 kb distal fragment limited reporter expression to motoneurons. (E) The 2.5 kb distal fragment expressed nlacZ in motoneurons and a few cells in the V2 interneuron region. (F) The proximal 550 nucleotide segment of Hb9 activates transcription throughout the spinal cord. In non-motoneurons (non-MNs), this activation is suppressed by factors that require elements in the 4.6 kb distal region of the promoter.

Repression of Hb9 by homeodomain transcription factors. (A) Plasmids encoding progenitor factors, Nkx2.2, Nkx6.1, Pax6 and Irx3, were co-transfected with the 2.5 kb distal+TK:luciferase reporter into 293 cells and luciferase activity was used to monitor the fold repression relative to vector-only transfections. (B) Hb9 and EnR-Hb9 (Hb9 homeodomain linked to eh1 engrailed repressor domain) repressed the synergistic activation of M250 by Isl1/Lhx3/NeuroM in transfected P19 cells. By contrast, Hb9-HD and Hb9-HD fused VP16 activation domain (VP16-Hb9) did not. (C,D) Gel retardation analysis reveals that Hb9 binds to the M50 and M100 (not shown) portions of its enhancer. Antibodies against Hb9 supershifted or disrupted the DNA:protein complex, whereas IgG control serum had no effect on DNA binding. Mutation of the ATTA sequences within M50 (ATTAm) disrupted Hb9 binding. The protein complex formed by Isl1/Lhx3 relies on the same ATTA elements for binding (Lee and Pfaff, 2003), suggesting Hb9 may compete with the LIM factors for access to the enhancer.