XB-ART-555Development. May 1, 2006; 133 (9): 1703-14.
The relative contributions of different FGF ligands and spliceforms to mesodermal and neural patterning in Xenopus have not been determined, and alternative splicing, though common, is a relatively unexplored area in development. We present evidence that FGF8 performs a dual role in X. laevis and X. tropicalis early development. There are two FGF8 spliceforms, FGF8a and FGF8b, which have very different activities. FGF8b is a potent mesoderm inducer, while FGF8a has little effect on the development of mesoderm. When mammalian FGF8 spliceforms are analyzed in X. laevis, the contrast in activity is conserved. Using a loss-of-function approach, we demonstrate that FGF8 is necessary for proper gastrulation and formation of mesoderm and that FGF8b is the predominant FGF8 spliceform involved in early mesoderm development in Xenopus. Furthermore, FGF8 signaling is necessary for proper posterior neural formation; loss of either FGF8a or a reduction in both FGF8a and FGF8b causes a reduction in the hindbrain and spinal cord domains.
PubMed ID: 16554360
Article link: Development.
Grant support: GM42341 NIGMS NIH HHS
Genes referenced: act3 actl6a cdx4 dbx1 egr2 en2 epha4 fgf4 fgf8 gal.2 hoxb9 hoxd1 myod1 otx2 rax sox2 t tbx2
Morpholinos referenced: fgf8 MO1 fgf8 MO2
Article Images: [+] show captions
|Fig. 1. FGF8 expression. (A) Alignment of Xenopus FGF8a and FGF8b amino acid sequences (ClustalW). The signal sequence cleavage position is predicted to be after the 22nd amino acid residue (alanine) (arrowhead), determined by using the SignalP 3.0 program (Bendtsen et al., 2004). Underlining indicates the N-linked glycosylation site (NFT) (reviewed by Dempski and Imperiali, 2002). (B) Xenopus FGF8a and FGF8b result from alternative splicing of the third exon; FGF8a uses an alternative splice acceptor (3′) site. The ATG indicates the translational start; black arrows indicate primers used for PCR amplification in the RTPCR panel in C,D. (C) X. laevis and (D) X. tropicalis RTPCR analysis of whole embryos using primers to amplify both FGF8a and FGF8b simultaneously. The FGF8a product is 253 bp; FGF8b is 286 bp. (E-N) In situ hybridization profile for FGF8 in X. tropicalis at the indicated stages. (E,G,I,L,N) Dorsal views with anterior towards the left; (F) a blastorporal view; (H,J) frontal views, dorsal upwards; (K,M) lateral views, anterior towards the left. FGF8 is expressed circumferentially around the blastopore (E,F) but is restricted to the dorsal posterior mesoderm as gastrulation proceeds; expression in the posterior mesoderm strengthens during neurulation and MHB expression begins (G,H). DM, dorsal mesoderm; MHB, midbrain-hindbrain boundary; ANR, anterior neural ridge; P, placode region; FB, forebrain; PA, pharyngeal arches; S, somite; OV, otic vesicle; Pr, pronephric anlage; TB, tail bud.|
|Fig. 2. FGF8b is a robust mesoderm inducer. (A) Diagram of explant assay. (B) RT-PCR analysis of explants injected as indicated at the one-cell stage, excised at stage 9 and cultured to stage 11; EF1α was used as a loading control. Lane 1, whole embryo (WE) positive control; lane 2 negative control minus reverse transcriptase (RT-); lane 3, uninjected explant lane. FGF8b and FGF8f robustly induce xbra expression (lanes 6-10). (C-H) Animal caps injected as indicated and cultured to stage 19. (I-T) Overexpression of Xenopus FGF8b expands xbra in whole embryos. Embryos were injected with mRNA, as indicated, into the marginal zone of one cell at the two-cell stage and cultured until stage 10.5. Embryos in the top row are shown from blastoporal views; the bottom row of images show the same embryos as the respective one above but from a lateral view with the blastopore down.β -galactosidase mRNA was injected as a lineage tracer and detected using Red-Gal substrate. (I,J) Control uninjected embryos. Neither (K,L) XlFGF8a (15/15 embryos) or (M,N) MmFGF8a (14/15 embryos) affects xbra expression; (O,P) XlFGF8b (15/15 embryos), (Q,R) HsFGF8b (20/20) and (S,T) MmFGF8f (18/18 embryos) robustly expand the xbra expression domain in a non-cell-autonomous manner. (U-FF) FGF8a and FGF8b have separable activities. Embryos were injected as indicated into one cell at the two-cell stage and processed by in situ hybridization for expression of myoD (top row) or neuronal β-tubulin (ntub) (bottom row) at stage 20. All are dorsal views with anterior towards the left. Effects on mesoderm and production of ectopic neurons is scored below the images; - indicates no effect. Overexpression of XlFGF8a and MmFGF8a results in massive ectopic ntub expression without affecting mesodermal development (W-Z). A minimum of eight embryos were examined and they showed consistent phenotypes for each injection.|
|Fig. 4. Effects of FGF8 reduction on mesoderm formation. (A-I) X. laevis embryos; all embryos injected into the marginal zone of one cell at the two-cell stage; (F-H,K,L,N,O) pink staining for the fluorescein-conjugated control MO indicates injected side; (B-E,I) red-gal indicates injected side. (A) Control xbra expression; (B) XlMOF8 (40 ng) strongly reduces xbra expression (19/20 embryos), and this loss can be rescued by human FGF8b (13/17), mouse FGF8f (14/20) and Xenopus FGF4 (17/23) (C-E). (F) MOSDF8 (85 ng) reduces xbra expression (27/31 embryos); (G) FGF8b rescues MOSDF8 phenotype (33/40); but FGF8a does not rescue (32/42) (H). (I) MOSAF8a (FGF8a-specific) (60 ng) does not affect xbra expression (15/16). (J-L) X. tropicalis embryos. MOSDF8 (42 ng) reduces xbra expression (22/28) (K). FGF8b rescues MOSDF8 (30/37) (L). (M-O) X. laevis embryos, dorsal views with anterior to the left. (M) Control stage 13 myoD expression. (N) MOSDF8 85 ng (39/39 reduced); (O) XlMOF8 (30 ng) (22/22 reduced).|
|Fig. 5. Effect of FGF8a overexpression on neural patterning. (A) FGF8a induces posterior neural genes in ectodermal explants. Embryos were injected into the animal hemisphere at the one-cell stage. Explants were excised at stage 9 and cultured until stage 20. Whole embryo (WE) stage-control embryo sample; whole embryo lysate processed without reverse transcriptase (RT-); explants from uninjected embryo (UI). Embryos were injected as indicated along the top. EF1α, loading control; muscle actin (MA) is an indicator of dorsal mesoderm. Endogenous neural gene expression domains are as follows: sox2, general neural tissue; otx2, forebrain and midbrain; en2, MHB; krox20, hindbrain r3 and r5; hoxD1, posterior hindbrain and spinal cord; hoxB9 and cad3, spinal cord. (B) Uninjected X. laevis tadpole; (C) FGF8a-injected tadpole. (D-V) Embryos displayed dorsoanteriorly; red staining indicates the lineage tracer. Embryos injected into a dorsal blastomere at the four-cell stage with 50 pg of FGF8a mRNA and cultured until neural tube stage 19/20. (D-I) FGF8a does not expand expression of the mesodermal genes myoD (32/32 embryos) or coll II (6/6), but sox2 expression domains are mispatterned and expanded (25/25). (M) Fluorescein-conjugated control MO was injected with the FGF8a mRNA as a lineage tracer (pink). (J-V) FGF8a expands posterior neural domains and reduces anterior neural domains. The displayed phenotypes are representative of the effects of FGF8a mRNA quantitated as follows: (K) otx2, 13/14; (M) rx1, 17/17; (O,P) ephA4, 33/35; (R,S) en2, 25/32; (U,V) krox20, 45/52; hoxB9, 43/44.|
|Fig. 6. Lower level reduction of FGF8a and FGF8b or strong reduction of FGF8a alone with XlMOF8, MOSDF8 and MOSAF8a prevents proper formation of posterior neural tissue at the early neurula stage.X. laevis embryos displayed dorsoanteriorly. Pink staining indicates the lineage tracer; embryos injected into one cell at the two- or four-cell stage. (A,E,I,M) Control MO (40 ng); (B,F,J,N) MOF8 20 ng; (C,G,K,O) MOSDF8 43 ng; (D,H,L,P) MOSAF8a 60 ng. (A-D) XlMOF8 (44/45), MOSDF8 (18/19) and MOSAF8a (38/39) cause a mispatterning of sox2 expression. (E-H) otx2 expression is expanded toward the posterior after XlMOF8 (23/25), MOSDF8 (20/20) and MOSAF8a (26/29) injection, while the posterior neural gene dbx is absent (MOF8, 25/25; MOSDF8, 20/20; MOSAF8a, 35/38). (I-L) en2 expression is diminished and sometimes completely absent on the injected side: XlMOF8 (13/13), MOSDF8 (17/17) and MOSAF8a (32/33). (M-P) both the spinal cord domain (hoxB9) and the hindbrain domain (krox20) is strongly reduced and shifted toward the posterior of the embryo on the XlMOF8 (42/42), MOSDF8 (20/20) and MOSAF8a (40/40) injected side of the embryo. Effects on the uninjected side are present but much weaker. (Q-S) Neural tube stage 20 embryos treated as indicated; (T-V) FGF8a mRNA (50 pg) rescued the reduction of hoxB9 caused by XlMOF8 (32/44), MOSAF8a (19/35) and MOSDF8 (19/20).|
|Fig. 7. Lowering FGF8 levels causes a reduction in placode formation and neural differentiation. (A-E) X. laevis embryos displayed dorsoanteriorly. Embryos were cultured with XlMOF8 (10 ng) or MOSAF8a (60 ng) until the neural tube stage 20; lineage tracer (pink). (B,C) XlMOF8 caused sox2 mispatterning (40/42), so did MOSAF8a but to a lesser degree (20/22). (E) MOSAF8a resulted in a slight expansion of the rx1 domain toward the posterior (16/20). (F-M) The effect of FGF8 reduction on early neuronal differentiation in X. tropicalis. XtMOF8 (8 ng), MOSDF8 (17 ng) and MOSAF8a (16 ng) were injected into one cell at the two-cell stage; injected side is oriented downwards. (F,H,J,L) Embryos were cultured until neurula stages; injected embryos demonstrate an early strong reduction in neuronal differentiation [XtMOF8 (20/20), MOSDF8 (9/9), MOSAF8a (16/16)]. (G,I,K,M) Embryos were cultured until the early tadpole stage; injected embryos demonstrate posterior truncations and continued reduction in differentiated neurons|