Reciprocal gene replacements reveal unique functions for Phox2 genes during neural differentiation

Abstract

The paralogous paired-like homeobox genes Phox2a and Phox2b are involved in the development of specific neural subtypes in the central and peripheral nervous systems. The different phenotypes of Phox2 knockout mutants, together with their asynchronous onset of expression, prompted us to generate two knock-in mutant mice, in which Phox2a is replaced by the Phox2b coding sequence, and vice versa. Our results indicate that Phox2a and Phox2b are not functionally equivalent, as only Phox2b can fulfill the role of Phox2a in the structures that depend on both genes. Furthermore, we demonstrate unique roles of Phox2 genes in the differentiation of specific motor neurons. Whereas the oculomotor and the trochlear neurons require Phox2a for their proper development, the migration of the facial branchiomotor neurons depends on Phox2b. Therefore, our analysis strongly indicates that biochemical differences between the proteins rather than temporal regulation of their expression account for the specific function of each paralogue.

Figure 1

Targeting strategy to introduce the Phox2b cDNA into the Phox2a locus. (A) Schematic diagram of Phox2a genomic locus (top), the targeting vector (middle) and the expected recombinant allele (bottom). The restriction enzymes, the probes and the size of the resulting wild-type and recombinant restriction fragments are indicated, together with the primers (red arrows). (B) Southern blot analysis of DNA derived from mouse ES cells. (C) PCR genotyping of the recombinant mouse line using primers recognizing specifically the KI and wild-type alleles. (D–F) Immunofluorescence on transverse sections through the midbrain of E10.5 wild-type (Phox2a^+/+) and homozygous (Phox2a^KI2b) embryos using antibodies directed against Phox2a and Phox2b proteins. The asterisk in panel F indicates the appropriate misexpression of Phox2a^KI2b allele according to the endogenous Phox2a gene in homozygous mutants. IIIv: third ventricle. Dashed lines indicate the ventricular (next to IIIv) and pial surfaces of the neuroectoderm.

Figure 2

Phox2b fully rescues the development of the LC and parasympathetic and cranial sensory ganglia. (A–D′) Sagittal and (E–H′) parasagittal sections of Phox2a^+/+ (A–H) and Phox2a^KI2b (A′–H′) embryos at E13.5 (A, A′) and E18.5 (B–H′), labeled with the antibodies or RNA probes shown on the left. (A–B′) DBH expression shows that differentiation of the LC is completely restored in Phox2a^KI2b embryos both at E13.5 and E18.5. (C′–D′) In Phox2a^KI2b embryos, staining for Phox2a is lost, but is replaced by Phox2b staining (arrow in D′), which is normally extinguished at this stage (D). The sphenopalatine ganglia (spg) are normal, as shown by the expression of peripherin (E–E′). The nodose-petrose (IX–X) cranial sensory ganglia are fully rescued, as assessed by the presence of Phox2b (F–F′) and by the expression of peripherin and Ret (G–H′). (I) Quantification of the ganglionic size visualized by peripherin staining and expressed as a percentage of the area occupied by the wild-type ganglion (100±15%, n=4 for wild type and 88.4±9.5%, n=6 for Phox2a^KI2b ganglia). Note that the difference is not statistically significant; P=0.19. LSC: locus subcoeruleus.

Figure 3

Phox2b partially rescues the development of oculomotor and trochlear neurons. Coronal (A–H′) and transverse (I–J′) sections of Phox2a^+/+ (A–J) and Phox2a^KI2b (A′–J′) embryos at E18.5 (A–D′), E13.5 (E–H′) and E10.5 (I–J′). Strong reduction in the size of nIII and nIV in the KI mutant embryos as shown by the decrease of Phox2b- (A–B′, E–F′), peripherin- (C–D′) and Isl1- (G–H′) positive neurons. The asterisks in panels E, E′, G and G′ indicate the contralaterally migrating nIII neurons. (I–J′) Double immunofluorescence with anti-Phox2b (green) and anti-Isl1 (red) at E10.5 shows a marked reduction of double-labeled cells (arrows in I′, J′) in Phox2a^KI2b embryos. (K) Quantification of the number of Phox2b-positive cells in nIII and nIV nuclei. In control embryos, oculomotor neurons were 99.2±3.5 per section (n=9), whereas in Phox2a^KI2b mutants, they were 23.8±3.1 (n=9). Control trochlear neurons were 51.5±7 per section (n=4), whereas Phox2a^KI2b neurons were 26.7±1 (n=4). (L, L′) Whole-mount immunohistochemistry with anti-neurofilament antibody shows a reduced trochlear (III) nerve and absence of the oculomotor (IV) nerve. IIIv: third ventricle.

Figure 4

Targeting strategy to introduce Phox2a cDNA into the Phox2b locus. (A) Schematic diagram of the Phox2b genomic locus (top), the targeting vector (middle) and the expected recombinant allele (bottom). The restriction enzymes, the probes and the sizes of the resulting wild-type and recombinant restriction fragments are indicated, together with the primers (red arrows). (B) Southern blot analysis of genomic DNA derived from mouse ES cells. (C) PCR genotyping of the recombinant mouse line using primers recognizing specifically the KI and wild-type allele. (D) PCR strategy to recognize the neomycin excision event following germline transmission. (E–G) Ventricular views of flat-mount wild-type (Phox2b^+/+) and homozygous (Phox2b^KI2a) E10.5 hindbrains hybridized with RNA probes for Phox2b (E) and Phox2a (F, G). In the Phox2b^+/+ embryos, the lateral stripes are identified by the expression of Phox2b (asterisk in E); however, they do not express Phox2a (asterisk in F). In Phox2b^KI2a embryos, the lateral stripes are now positive for Phox2a expression (asterisk in G), indicating that in Phox2b^KI2a, Phox2a is expressed according to Phox2b.

Figure 5

Phox2a partially rescues the differentiation of the LC and of cranial sensory ganglia. (A–B′) Sagittal sections of heterozygous and homozygous Phox2b^KI2a embryos at E13.5, stained with the markers indicated on the left. The asterisks in panels A′ and B′ indicate lack of LC cells. (C–D′) Flat-mounted preparations of the met-mesencephalic region of E10.5 heterozygous and homozygous Phox2b^KI2a embryos hybridized with Phox2a (C, C′) or DBH (D, D′) probes. The arrows indicate an abnormal distribution of Phox2a-positive cells (C′) and partial rescue of DBH expression (D′). (E, E′) Transverse sections through the isthmus of E10.5 heterozygous and homozygous Phox2b^KI2a embryos stained by combined DBH in situ hybridization (dark blue) and anti-Phox2a immunohistochemistry (brown). The arrowheads in panel E′ indicate that only a subpopulation of LC neurons is double positive for Phox2a and DBH in Phox2b^KI2a embryos; in control embryos, all the cells are double labeled (E). Outlined areas in panels E and E′ are magnified in the insets. (F, F′) Lateral views of Phox2b^+/+ and Phox2b^KI2a embryos hybridized with Phox2a showing normal appearance of the cranial ganglia at E10.5. (G–I′) Parasagittal sections of E16.5 Phox2b^+/+ and Phox2b^KI2a embryos immunostained with anti-Phox2a and hybridized with peripherin and Ret probes. (J) Quantification of the ganglionic size visualized by peripherin staining and expressed as a percentage of the area occupied by the wild-type ganglion (100±8.8%, n=9 for wild type and 37.8±5.1%, n=4 for mutant ganglia).

Figure 6

Phox2a can partially substitute for Phox2b during differentiation of hindbrain branchio-visceral-motor neurons, but cannot drive proper fbm neuron migration. (A–B′) In situ hybridization with Phox2a and Isl1 probes on transverse sections through r4 of Phox2b^+/+ (A, B) and Phox2b^KI2a (A′, B′) E10.5 embryos. The asterisk in panel A′ indicates Phox2a misexpression in the r4 ventricular zone in Phox2b^KI2a embryos. (B, B′) In situ hybridization with Isl1 probe reveals a mild reduction of Isl1-expressing bm neurons in r4 in Phox2b^KI2a embryos. Ventricular (C, C′) and pial (D, D′) views of flat-mounted hindbrains hybridized with Tbx20, a specific marker of bm/vm neurons (Kraus et al, 2001; unpublished results), in Phox2b^+/+ and Phox2b^KI2a embryos at E11.5 and E13. (C′) In Phox2b^KI2a embryos, trigeminal neurons migrate to their appropriate position (asterisk in C′), but fbm neurons fail to undergo a caudal migration (arrowhead in C′). At E13, in Phox2b^KI2a mutants, most fbm neurons are still stalling in r4 (arrowhead in D′) and a facial-like nucleus is forming in dorsal r4 (arrow in D′). (E, E′) Lateral views of E11.5 Phox2b^+/+/Nkx6.2^tauLacZ (E) and Phox2b^KI2a/Nkx6.2^tauLacZ (E′) embryos stained for β-galactosidase to label the bm/vm projections. Note that in Phox2b^KI2a embryos, all the axonal tracts are thinner compared to control embryos (arrowheads in E′). (F–J′) Transverse sections through the brainstem of E16.5 Phox2b^+/+ and Phox2b^KI2a embryos hybridized with a peripherin probe. Sections are shown in an anterior (upwards) to posterior (downwards) direction. The mutant facial nucleus is located more anteriorly, whereas in wild-type embryos, it extends from the level of the abducens nucleus (G) to more posterior locations (H). nVI and nXII: nuclei of the abducens and hypoglossal nerves, respectively; MesV; mesencephalic nucleus of the trigeminal nerve.