Loss of function but no gain of function caused by amino acid substitutions in the hexapeptide of Hoxa1 in vivo

Abstract

Homeodomain containing transcription factors of the Hox family play critical roles in patterning the anteroposterior embryonic body axis, as well as in controlling several steps of organogenesis. Several Hox proteins have been shown to cooperate with members of the Pbx family for the recognition and activation of identified target enhancers. Hox proteins contact Pbx via a conserved hexapeptide motif. Previous biochemical studies provided evidence that critical amino acid substitutions in the hexapeptide sequence of Hoxa1 abolish its interaction with Pbx. As a result, these substitutions also abolish Hoxa1 activity on known target enhancers in cellular models, suggesting that Hoxa1 activity relies on its capacity to interact with Pbx. Here, we show that mice with mutations in the Hoxa1 hexapeptide display hindbrain, cranial nerve, and skeletal defects highly reminiscent of those reported for the Hoxa1 loss of function. Since similar hexapeptide mutations in the mouse Hoxb8 and the Drosophila AbdA proteins result in activity modulation and gain of function, our data demonstrate that the functional importance of the hexapeptide in vivo differs according to the Hox proteins.

FIG. 1.

Targeted mutagenesis of the Hoxa1 hexapeptide. (A) Point mutations introduced in the sequence of the Hoxa1 hexapeptide and the resulting WM-to-AA amino acid substitutions. The mutagenic primers used are indicated by half-arrows with asterisks. (B) Genomic organization of the Hoxa2-Hoxa1 genomic locus and structure of the targeting construct pGIH415 (top). The PGK-neomycin resistance cassette flanked by loxP sites was introduced into the unique XbaI site located in the intron of Hoxa1. Details of the mutant allele after homologous recombination and after Cre-mediated recombination are shown below. Sequences corresponding to the primers Test1, Test7, RRA110, and SRA12 used for animal genotyping are represented by half-arrows. Sequences corresponding to primers RRA110 and RRA11 used for RT-PCR analysis are also represented. The asterisk indicates the mutated hexapeptide. Shaded boxes, Hoxa2 and Hoxa1 coding sequences; solid boxes, PGK-neomycin (neo) resistance cassette; solid triangles, loxP sites; thick solid lines, additional nonhomologous sequences. B, BamHI; E, EcoRI; X, XbaI. (C) PCR analysis of DNAs isolated from wild-type (WT) and heterozygous (WM-AA/+) and homozygous (WM-AA/WM-AA) mutant embryos. PCR amplification with primers RRA110 and SRA12 shown in panel B gives rise to a 390-bp product for the wild-type allele and a 470-bp product for the mutated allele, due to the presence of the loxP tag. (D) RT-PCR analysis of mRNAs extracted from wild-type and homozygous mutant 8-dpc embryos. Amplification products of the expected size (508 bp) corresponding to the correctly spliced Hoxa1 mRNA are obtained at similar levels from homozygous mutant and wild-type embryos.

FIG. 2.

Analysis of hindbrain patterning by whole-mount in situ hybridization. (A) Wild-type (top) and mutant (bottom) embryos between 7.5 and 8.75 dpc hybridized with a probe for Hoxa1, Krox-20, Hoxb1, or kreisler. The arrowheads indicate the anterior limits of expression of Hoxa1, corresponding to the presumptive r3-r4 boundary. In the mutant, the r3 expression domain of Krox-20 is enlarged, while its r5 expression domain is drastically reduced. Moreover, the r4 expression domain of Hoxb1 is reduced, and the kreisler r5-r6 domain of expression is only one rhombomere long. (B) Wild-type (top) and mutant (bottom) 9.5-dpc embryos hybridized with a probe for Krox-20, Hoxb1, or Hoxb2. In the mutant, the r5 expression domain of Krox-20 is almost absent (arrowhead), the r4 expression domain of Hoxb1 is clearly reduced, and the highly stained portion of the Hoxb2 expression domain, corresponding to r3-r6, is shortened (bracket). Scale bars, 200 μm.

FIG. 3.

Analysis of hindbrain patterning by in situ hybridization on serial coronal sections. Expression of Krox-20, Hoxb1, and Hoxb2 on serial coronal sections of 9.5-dpc wild-type (top) and mutant (bottom) embryos. Bright-field views after toluidine blue staining are presented on the left, and dark-field views after in situ hybridization are shown on the right. The arrowhead points to the Krox-20 r5 expression domain, drastically reduced in the mutant. ov, otic vesicle. Scale bar, 200 μm.

FIG. 4.

Analysis of neural crest derivatives and skeletons. (A) Whole-mount in situ hybridization on wild-type (top) and homozygous mutant (bottom) embryos at 9.25 and 9.5 dpc. The embryos were hybridized with a CrabpI or a Sox10 probe. The open arrow indicates the reduction of the ncc population expressing CrabpI and migrating into the second branchial arch (ba2). The solid arrows indicate the reduction of neurogenic ncc at the level of nerve VII-VIII and the absence of Sox10 expression at the level of nerves IX and X. ba, branchial arch; ov, otic vesicle. (B) Immunostaining of cranial nerves of a 10.5-dpc wild-type embryo (top) and two Hoxa1^WM-AA/WW-AA mutant embryos (bottom) (caudal is on the left). In the mutants, the open arrows indicate the absence of nerve IX dorsal roots, the solid arrowhead indicates the fusion of the inferior ganglion of nerve IX (IXg) with the inferior ganglion of nerve X (Xg), and the yellow arrowhead points to the lack of the interganglionic portion of nerve X. (C) (Top) Dorsal view of the skull of wild-type (left) and Hoxa1^WM-AA/WW-AA mutant (right) newborns. In the mutant, the shape of the basioccipital bone (bo) is altered, the otic capsules (oc) are reduced, and the tympanic ring (t) is displaced. (Bottom) Dissected otic capsules and tympanic rings from a wild-type skull (left) and three Hoxa1^WM-AA/WW-AA mutant skulls (right). co, cochlea; i, incus; ma, malleus; ow, oval window; rw, round window; s, stapes; scc, semicircular canals; sty, styloid process; t, tympanic ring. Scale bars, 200 (A and B) and 500 (C) μm.