Cloning and functional analysis of hypothalamic homeobox gene Bsx1a and its isoform, Bsx1b

Abstract

The hypothalamus is a key regulatory unit of the neuroendocrine system and plays an essential role in energy balance and reproduction. Despite its important role, the molecular mechanisms underlying hypothalamic development are not fully understood. Here, we report molecular analyses of a newly identified murine homeobox gene, Bsx/Bsx1a, that is expressed in the developing and postnatal hypothalamus. We demonstrate that BSX1A is a DNA binding protein and a transcriptional activator. Transcriptional reporter assays identified the C-terminal region of BSX1A as an activation domain. We have isolated an alternative splice form of Bsx1a, designated Bsx1b, which retains the N-terminal region but lacks the homeodomain. Analyses of subcellular localization using transfected cell lines revealed that BSX1A and BSX1B localize in the nuclei and cytoplasm, respectively. Immunohistochemical analyses suggested that both BSX1A and BSX1B are expressed in the neonatal hypothalamus. Taking these data together, we propose that alternative RNA splicing is involved in hypothalamic development/function.

FIG. 1.

Gene structure, schematic representation, deduced amino acid sequences, and expression of BSX1A and BSX1B. (A) Both BSX1A and BSX1B are encoded by three coding exons (squares) on mouse chromosome 9. The horizontal lines indicate noncoding genomic DNA. HD, homeodomain (black squares); N, N-terminal region; C, C-terminal region. The BSX1B-specific region is shaded. (B) Deduced amino acid sequences of BSX1A and BSX1B. The homeodomain sequence is underlined. (C) Immunohistochemistry on neonatal mouse hypothalamic sections using anti-BSX1A and anti-BSX1B sera. On the right are high-magnification images. Immunoreactive cells were detected around the third ventricle (arrows). III, third ventricle; ME, median eminence; NC, negative control.

FIG. 2.

Subcellular localization of BSX1A and BSX1B. (A) Schematic representation of myc-tagged proteins expressed in COS7 cells. HD, homeodomain. (B) Immunocytological analyses of myc-tagged protein-transfected COS7 cells. The arrows indicate BSX1A- or BSX1B-immunoreactive cells. The right column contains higher magnifications of parts of the merged images (the dotted squares). Note that BSX1A and BSX1B were detected in the nuclei and cytoplasm, respectively.

FIG. 3.

Subcellular localization of BSX1AΔNLS. (A) Schematic representation of myc-tagged proteins expressed in Hs683 cells. HD, homeodomain. (B) Immunocytological analyses of myc-tagged protein-transfected Hs683cells. The images on the left (green) are myc-immunoreactive signals, and those on the right show BSX1A -or BSX1B-immunoreactive cells (red). Note that BSX1A was detected in the nuclei and BSX1AΔNLS and BSX1B in the cytoplasm.

FIG. 4.

DNA binding ability of BSX1A. (A) Schematic representation of GST fusion proteins used for EMSA. HD, homeodomain. (B) EMSA using BBS. Wild-type and mutated (mut) BBSs were used as competitors. Mutated residues are shown in boldface. Below are Coomassie-stained GST fusion proteins used for EMSA.

FIG. 5.

Transcriptional activity of BSX1A in yeasts. (A) Schematic representation of full-length or truncated (ΔC or ΔN) BSX1A fused to GAL4DBD. HD, homeodomain. (B) (Top) β-Gal activity measured by a liquid assay using o-nitrophenol-β-d-galactoside (ONPG) as a substrate. The values of the ONPG assay are indicated in Miller units. (Bottom) Results of a β-Gal filter assay using X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) as a substrate. Vec, vector. (C) Western analysis using anti-GAL4DBD antibody to confirm GAL4DBD fusion protein expression.

FIG. 6.

Transcriptional activity of BSX1A in mammalian cells. (A) Schematic representation of myc-tagged full-length or truncated (ΔC) BSX1A and a luciferase reporter construct. HD, homeodomain. (B) Luciferase activities of myc-tagged protein-transfected Hs683 (above) and Neuro2a (below) cells; 0.2 μg of reporter plasmid was used for each transfection. All luciferase activities were normalized for transfection efficiency by measuring the Renilla luciferase activity of cotransfected pRL-SV40 vector (0.05 μg). The error bars indicate standard deviations.

FIG. 7.

Transcriptional activity of mutant BSX1A in mammalian cells. (A) Schematic representation of myc-tagged full-length, truncated (ΔAD), and mutated (ΔNLS) BSX1A and a luciferase reporter construct. A putative NLS (₁₀₉RRRKAR₁₁₄) was mutated to ₁₀₉GRGKAR₁₁₄ (*). HD, homeodomain. (B) Luciferase activities of myc-tagged protein-transfected Hs683 cells; 0.2 μg of reporter plasmid was used for each transfection. All luciferase activities were normalized for transfection efficiency by measuring the Renilla luciferase activity of cotransfected pRL-SV40 vector (0.05 μg). The error bars indicate standard deviations. (C) Western analysis of myc-tagged wild-type and mutant BSX1A and BSX1B. The expression vectors were transfected 293T cells, and cell lysates were blotted with α-myc antibody (Sigma). (D) EMSA. GST-BSX1AΔAD, but not GST-BSX1AΔNLS, retains the DNA binding ability to BBS. (Below) Coomassie-stained GST fusion proteins used for EMSA.

FIG. 8.

Functional-domain mapping of BSX1A and BSX1B. (A) The N-terminal regions of BSX1A and BSX1B possibly serve as protein interaction domains. The C-terminal region of BSX1A contains a transcriptional activation domain. The functions of BSX1B-specifc sequences (shaded square) are unknown. HD, homeodomain (black square); N, N-terminal region; C, C-terminal region. (B) An acidic residue cluster in the C-terminal region conserved among vertebrates. Acidic residues are underlined.