XB-ART-4751Development October 1, 2003; 130 (20): 4943-53.
We analyzed the endogenous requirement for Cerberus in Xenopus head development. ''Knockdown'' of Cerberus function by antisense morpholino oligonucleotides did not impair head formation in the embryo. In contrast, targeted increase of BMP, Nodal and Wnt signaling in the anterior dorsal-endoderm (ADE) resulted in synergistic loss of anterior head structures, without affecting more posterior axial ones. Remarkably, those head phenotypes were aggravated by simultaneous depletion of Cerberus. These experiments demonstrated for the first time that endogenous Cerberus protein can inhibit BMP, Nodal and Wnt factors in vivo. Conjugates of dorsal ectoderm (DE) and ADE explants in which Cerberus function was ''knocked down'' revealed the requirement of Cerberus in the ADE for the proper induction of anterior neural markers and repression of more posterior ones. This data supports the view that Cerberus function is required in the leading edge of the ADE for correct induction and patterning of the neuroectoderm.
PubMed ID: 12952900
Article link: Development
Genes referenced: ag1 bmp4 cer1 egr2 emx1l en2 eomes foxg1 hhex myc nkx2-1 nodal nodal1 not odc1 otx2 shh wnt8a
Morpholinos: cer1 MO1
Article Images: [+] show captions
|Fig. 5. Molecular markers analysis after microinjection of McerP-BMP4, -Xnr1 and -Xwnt8 in frog embryos. (A) RT-PCR analysis at stages 12, 13 and 15 show that the cement gland and anterior neural markers Xag, XBF1, Xemx1 and Nkx2.1 are downregulated in embryos coinjected with McerP-BMP4, McerP-Xnr1 and McerP-Xwnt8 (20 pg each), when compared with uninjected controls, while the levels of more posterior neural markers, like En2 and Krox20, are not changed. ODC was used as a loading control. RNA extracts used for the RT-PCRs were made from pools of 5 randomly picked embryos. (B-K) In situ hybridization analysis for different molecular markers at stages 22/24. The injection of McerP-BMP4, McerP-Xnr1 and McerP-Xwnt8 (20 pg each) leads to the suppression of the anterior domains of expression of XBF1 (B,C), Xotx2 (D,E) and Xnot2 (F,G). Expression of the hindbrain marker, Krox20 (H,I), was not significantly changed in the injected embryos as well as in the controls. (J,K) Xshh expression in injected embryos does not extend as anteriorly as it does with the uninjected sibling embryos.|
|Fig. 4. Misexpression of BMP4, Xnr1 and Xwnt8 does not interfere with anterior endomesoderm patterning. (A-A′ and C-C′) Stage 10 and (B-B′ and D-D′) 12 embryos halves from uninjected (A-B′) or injected twice dorsally at the four-to-eight cell stage embryos, with a mixture of 20 pg each of McerP-BMP4, McerP-Xnr1 and McerP-Xwnt8 (C-D′), were sagitally sectioned and each half was hybridized with a Xhex or a Xcer probe. The expression of these endomesodermal markers was unchanged in the injected embryos (C,C′,D,D′) when compared to the uninjected embryos (A,A′,B,B′).|
|Fig. 1. Cer morpholino inhibits translation of cerberus mRNA. (A) Scheme of the HA-tagged Cerberus expression construct (5′ UTR cer-HA; top), the HA-tagged Cer rescue construct (cer-HA; bottom) and the morpholino sequence targeting the cerberus 5′ UTR sequences. (B,C) Western blot analysis of HA-tagged Cerberus and myc-tagged ΔN-moesin proteins. (B) In vitro transcription/translation of Cerberus protein in reticulocytes from 220 ng of plasmid was blocked by 20 pmol of the CerMo (lane 2) but not by control morpholino (CoMo) (lane 1). Transcription/translation from an equal amount of rescue plasmid was not blocked by the CerMo (lane 3). (C) Four-cell stage embryos injected in the animal pole with a total of 120 pg of 5′ UTR cer-HA construct were grown till stage 10.5 and one embryo equivalent protein extracts were used per lane in western blots. Translation of 5′ UTR cer-HA was blocked by coinjection with 1.6 pmol of CerMo (lane 2), but not with 2.0 pmol CoMo (lane 1). Coinjection of 80 pmol of the rescue construct was able to overcome the CerMo effect (lane 3). (D-F) Ectopic head-like structures induced by the injection of 700 pg of 5′ UTR cer-HA capped mRNA in the ventral side of four-cell stage embryos (E) are suppressed by coinjection of 3.2 pmol of CerMo (F). White arrowhead indicates the cement gland of the primary axis while the black arrowhead points to the ectopic cement gland. (G-I) No significant anterior defects are visible in embryos microinjected in the two dorsal-vegetal blastomeres at four-to-eight cell stage either with a total of 16 pmol of CoMo (G) or with 16 pmol (H) and 3.2 pmol (I) of CerMo.|
|Fig. 2. lacZ expression driven by the Mcer promoter mimics endogenous cerberus expression domain in the early frog embryo. (A) Schematic of the McerP-lacZ and CMV-lacZ constructs. (B-E) β-gal staining of embryos injected at the four- to eight-cell stage either in the two dorsal-vegetal (B,D) or ventral-vegetal (C,E) blastomeres with McerP-lacZ or CMV-lacZ constructs. Embryos were injected at the four- to eight-cell stage in both dorsal-vegetal blastomeres with McerP-lacZ, grown to stage 10+ (F-F′) or 11 (G-G′), sagitally sectioned and each half was hybridized with a lacZ (F,G) or a Xcer (F′,G′) probe.|
|Fig. 3. Head defects induced by McerP-BMP4, Xnr1 and Xwnt8 microinjection in the frog embryo. Embryos were injected at the four-to-eight-cell stage in the two dorsal-vegetal blastomeres. (A-C) Injection of 80 pg of CMV.BMP4 (A; 100%, n=22), CMV.Xnr1(B; 94%, n=17) and CMV.Xwnt8 (C; 20%, n=25) led to very severe phenotypes. (D-F) Injection of 20 pg (F; 60%, n=20), 40 pg (E; 62%, n=34) and 80 pg (D; 50%, n=24) of McerP-BMP4 showed a concentration dependent increase in head truncation. (G-I) A progressive head reduction and loss of eye were observed when 20 pg (I; 100%, n=26), 40 pg (H; 24%, n=21) and 80 pg (G; 59%, n=34) of McerP-Xnr1. (J-L) Increasing the amount of McerP-Xwnt8 from 20 pg (L; 60%, n=25) to 40 pg (K; 62%, n=28) to 80 pg (J; 74%, n=19) resulted in loss of cement gland and cyclopia and ultimately in complete truncation of the head. (M,N) Synergistic effect of McerP-BMP4, Xnr1 and Xwnt8 is shown by the coinjection of 8 pg (N; 65%, n=23) and 20 pg (M; 63%, n=46) of each construct which resulted in more severe defects than the ones observed in embryos injected with equal amounts of the individual constructs (F,I and L). 20 pg of McerP-lacZ was coinjected to access the correct targeting of the promoters to the ADE (yellow arrowheads in E, H and K).|
|Fig 6. Knockdown of endogenous Cerberus enhances the head phenotypes induced by microinjection of McerP-BMP4, -Xnr1 and -Xwnt8. (A-D) The head defects observed by the coinjection of McerP-BMP4, -Xnr1 and -Xwnt8 (8 pg each) together with the CoMo (B; 64%, n=62) can be aggravated when endogenous cerberus is knock-down by 1.6 pmol of CerMo (C; 65%, n=46). The specificity of this sensitization was verified by the coinjection of Cer-Long plasmid, which could rescue the phenotype (D; 58%, n=42). (E-J) The mild phenotypes obtained by individually injecting McerP-BMP4 (E; 30%, n=44), McerP-Xnr1 (G; 40%, n=40) or McerP-Xwnt8 (I; 66%, n=47) were also aggravated by the coinjection of 1.6 pmol CerMo (F; 25%, n=41. H; 60%, n=39. J; 58%, n=44 respectively). (K-L) CerMo dependent aggravation of the McerP-Xwnt8 phenotype could be completely rescued by Cer-Long construct (K; 66%, n=48), but only partially rescued if Cer-Short plasmid (L; 55%, n=44) is used instead. These results were observed in two independent experiments. (M) Graphical representation of the range of phenotypes observed by increasing amounts of BMP, Wnt and Nodal misexpressed in the ADE, showing the requirement of lower amounts of these signals to generate the same phenotypes when endogenous Cerberus is depleted.|
|Fig. 7. Endogenous Cerberus activity is required for correct expression of neural markers in a tissue recombination induction assay. (A) ADE (yellow) explanted from either stage 10.5 uninjected embryos or from ones injected dorsal-vegetally with 3.2 pmol of CerMo, were conjugated with isochronic dorsal ectoderms (blue). Conjugates were grown until sibling embryos reached stage 30/31. (B) RT-PCR analysis of telencephalic markers in DE/ADE conjugates at stage 30/31. DE explants show expression of XBF1 and Xemx1. When DE is conjugated with control ADE, dorsal telencephalic markers are up-regulated (lane 3). CerMo injected ADEs are no longer able to up-regulate neither XBF1 nor Xemx1 in the DE/ADE conjugates (line 4). Cer-HA DNA construct, which lacks the 5′UTR sequences complementary to CerMo, con rescue the induction of both telencephalic markers in the DE/ADE CerMo conjugates. (C) RT-PCR analysis of neural markers from the DE/ADE conjugates at stage 30/31. DE show expression of Xemx1 and eomes but, when conjugated with control ADE, marked up-regulation of these dorsal telencephalic markers (lane 3). ADE CerMo conjugated with dorsal ectoderm suppresses expression of Xemx1 and eomes (line 4). Expression levels of the mid-hindbrain marker En2 are unchanged both in unconjugated DE (lane 2) as well as in DE /ADE (lane 3) or DE/ADE CerMo conjugates (lane 4). Krox20 is downregulated in the DE/ADE conjugates (lane 3) but its expression levels are partially rescued in DE conjugates with CerMo injected ADE (lane 4).|