Fig. 1. The initial Hox expression is localised in the non-organiser mesoderm. Whole mount in situ hybridisation of Hox genes and mesodermal marker genes. (A–D) Vegetal view (organiser is up) of midgastrula stage embryos stained for Hoxd-1 (A), which shows colocalisation with the mesodermal marker Xbra (B) with the exception
of a gap in the organiser region [indicated by chordin (chd) expression (C)]. The expression domain of the secreted antiorganiser signal BMP-4 (i.e. the region of highest levels of secreted protein) is localised in similar embryonic regions (D). (E– J) Dissections were made across the initial Hox expression domain close to the organiser (O) (as indicated in the schematic drawings). In each case, one-half of an embryo shows the early expression of either Hoxd-1 at stage 10.5 (E), Hoxc-6 at stage 11.5 (G) or Hoxa-7 at stage 12.5 (I), whilst the corresponding second half is stained for the mesodermal marker Xbra (F, H, J). The early expression of the different Hox genes is located within the Xbra domain. The arrowheads point to corresponding positions in the two half embryos.
Fig. 2. The effects of mesoderm-inducing factors on early Hox expression. (A) Levels of Hoxd-1 in explanted animal cap sandwiches from stage 11.5 noninjected embryos, embryos injected with Activin and embryos injected with Activin and the dominant interfering construct Xbra-EnR. Lightcycler PCR was used to quantitatively measure the levels of Hoxd-1, which were normalised to ODC levels and are shown as a percentage of the endogenous levels in whole embryos (WE). (B, C) The growth factor Activin was injected into the blastocoel of stage 8 embryos. In situ hybridisations (lateral views, organiser to the right) are shown for Hoxd-1 at stage 11 in noninduced control (B) and Activin-injected (C) embryos. In induced embryos, the Hoxd-1 expression is expanded in the animal direction. Arrowheads point to the animal border of Hoxd-1 expression. (D–K) Xbra was ectopically expressed in the animal region. Hox expression was analysed by in situ hybridisation. Lateral views (organiser is to the right) of noninjected control embryos (ni) (D, F, H, J) and Xbra-injected embryos (E, G, I, K) stained for Hoxd-1 at stage 11 (D, E), Hoxb-4 at stage 11.5 (F, G), Hoxc-6 at stage 12 (H, I) and Hoxb-9 at stage 12.5 (J, K). Compared to corresponding controls, the expression of all analysed Hox genes in the Xbra-injected embryos is expanded in the animal direction. (L) The ability of Xbra to induce Hox genes in stage 11.5 explanted animal cap sandwiches (AC) was analysed by RT-PCR. All the Hox genes examined (Hoxd-1, Hoxb-4, Hoxc-6 and Hoxb-9) were induced by Xbra. The endogenous expression in whole embryos (WE) is also shown.
Fig. 3. Repression of Xbra results in a repression of Hox genes. Nonorganiser
site injection of factors that lead to a repression of Xbra. (A–F)
Vegetal views (organiser is up) of noninjected (ni) control embryo (A) and
embryos injected with otx-2 (B), mix.1 (C), gsc (D), Sia (E) and Xbra-EnR
(F) mRNAs. All injected embryos show downregulation of Hoxd-1 on the
site of injection (arrowheads). (G–L) Non-organiser side views of
noninjected embryos (G, I, K) and Xbra-EnR-injected embryos (H, J, L).
Staining for Hoxb-4 (G, H), Hoxc-6 (I, J) and Hoxb-9 (K, L) show that these
Hox genes are also downregulated by repression of Xbra function.
Arrowheads point to the side of injection.
Fig. 4. The constitutively active BMP receptor, ALK-6, ventralises and posteriorises embryos. (A, B) Embryos were injected with the constitutively active human BMP receptor (Alk-6) at stage 1. Phenotypic analysis shows the expected effects of head reduction and shortened trunks in Alk-6-injected embryos (B) compared to noninjected (ni) controls (A). (C–N) Marker analysis of Alk-6-injected embryos. In situ hybridisations were performed on noninjected embryos (C, F, I, L) and embryos injected with 600 pg Alk-6 (D, G, J, M) or 1.2 ng Alk-6 (E, H, K, N) using probes for the anterior gene, otx-2 (C, D, E), the organiser gene, chordin (F, G, H), the posterior gene, Hoxb-9 (I, J, K) and the ventral gene, Xvent-2 (L– N). Expression of otx-2 and chordin was reduced, whereas expression of Hoxb-9 and Xvent-2 was expanded. Embryos are shown from the anterior (C– E), from the dorsal site with anterior to the right (F–K) or from the lateral side with anterior to the right (L – N).
Fig. 5. Ectopic activation of the BMP pathway expands Hox expression to the organiser side. (A–H) Vegetal views (organiser is up) of noninjected (ni) and Alk-6-injected embryos stained for Hoxd-1 (A, B), Hoxc-6 (C, D), Hoxa-7 (E, F) and Hoxb-8 (G, H). Ectopic Alk-6 expression results in an expansion of the Hox expression on the organiser side. (I) A cross section (as indicated in the schematic drawing) of the marginal zone of an Alk-6-injected embryo (stage 11.5, organiser side indicated by O). Ectopic expression of Hoxd-1 on the organiser side is present in the mesoderm, but not in the overlying ectoderm (arrowheads). The dashed lines indicate Brachet’s cleft, which separates involuted mesoderm and the non-involuted tissue (i.e. neuroectoderm on the organiser side and preinvoluted mesoderm on the non-organiser side).
Fig. 6. Knock down of the BMP pathway results in repression of Hox
expression. We used a dominant negative BMP receptor (tBR), a BMP
antagonist [chordin (chd)] and two different morpholinos against BMP-4
(BMP4MO1, BMP4MO2), to repress the BMP-4 pathway. Arrowheads
point to the side of injection. (A–F) Vegetal views (organiser is up) of
embryos stained for Hoxd-1. Noninjected (ni) embryos (A) and embryos
injected with a control morpholino (conMO, D) show the characteristic
horseshoe-shaped expression domain. Injection of tBR or chordin mRNAs
on the non-organiser side resulted in a repression of Hoxd-1 (B, C), as did
the injection of the two different BMP-4 morpholinos (E, F). (G–J)
Experiments with the BMP4MO1 demonstrate that the expression of Hoxc-
6 (H) and Hoxa-7 (J) are repressed compared to control morpholino
injection (G, I). (K–N) To ensure that the BMP-4 morpholino really affects
the BMP-4 pathway, the known downstream target Xvent-2 was analysed
for changes in its expression after injection of BMP4MO1 and was seen to
be downregulated (K, L). The organiser gene Xlim-1 is upregulated (M, N).
Fig. 8. BMP-4 and Xbra affect Hox expression independently but cooperatively. Arrowheads in all panels point to the site of injection. (A–D) Vegetal views
(organiser is up) of gastrula stage embryos after injection of BMP-4 morpholino (BMP4MO1) (A, B) or control morpholino (C, D). Whilst Hoxd-1 was
downregulated in BMP4MO1-injected embryos (A), Xbra expression was still present (B). (E–H) Vegetal views (organiser is up) of gastrula stage embryos
after injection of noggin (nog) or chordin (chd) mRNA. Both of these BMP inhibitors downregulated Hoxd-1 expression (E, G) but not Xbra expression (F, H).
(I, J) To see whether Xbra had an effect on the BMP pathway, the BMP-4 target gene Xvent-2 was analysed after animal injection of Xbra. The expression
pattern was unchanged (I, J) (lateral views, organiser to the right). (K–N) Views of an uninjected embryo from the non-organiser side (K), an embryo injected
with Xbra alone (L), an embryo injected with BMP4MO1 (M) and an embryo injected with BMP4MO1 and Xbra (N). The Xbra coinjection did not rescue the
Hoxd-1 downregulation by BMP4MO1. (O) Lightcycler PCR was performed to demonstrate that the upregulation of Hoxd-1 expression by Xbra in animal cap
sandwiches (Xbra + conMO AC) can be reduced by coinjection of a BMP-4 morpholino (Xbra + BMP4MO1 AC). The graph shows Hoxd-1 levels normalised
to odc levels and expressed as a percentage of endogenous expression in whole embryos (WE). Noninjected cap sandwiches (ni AC) are also shown. Error bars
indicate standard deviation (n = 3).
Fig. 9. Cooperation of Xbra and BMP-4. (A–D) In situ hybridisation of embryos dissected along the midline from stage 11 noninjected (ni) embryos (A),
embryos injected with Xbra (B), embryos injected with BMP-4 mRNA (C) and embryos injected with both Xbra and BMP-4 (D). Pictograms indicate the
localisation of Hoxd-1 expression in the half embryos (blue colour) and projections of the expression onto the exterior of whole embryos (light blue line). In
noninjected embryos (A), the normal expression in the non-organiser portion of the marginal zone is shown. No expression is present in the organiser. In
embryos injected with Xbra (B), the Hoxd-1 expression is expanded in the animal direction, but not to the organiser side. Expansion of the BMP-4 function by
BMP-4 mRNA injection (C) leads to ectopic Hoxd-1 expression in organiser mesoderm, but not in animal parts of the embryo. Combination of both Xbra
mRNA and BMP-4 mRNA injection resulted in ectopic expression of Hoxd-1 all over the mesoderm and ectoderm (D). O– organiser side; NO—non-organiser
side; AN—animal; VG—vegetal. (E) Projection of the embryo into a Cartesian coordinate system: Xbra and BMP-4 restrict the Hox expression domain. The
Xbra expression domain (dotted) overlaps with the functional domain of BMP-4 (grey gradient). In the overlapping region (blue), Hox genes are initially
expressed. An actual embryo stained for Hoxd-1 is shown in the same orientation. The expression of the Xbra repressor mix.1 is also indicated, as is the
presence of the organiser (org).
Fig. 7. Rescue of BMP-4 morpholino effects on Hox gene expression by BMP-4 protein treatment. (A) In situ hybridisation of stage 11 embryos for Hoxd-1.
Shown are ventral views of a noninjected embryo (A), an embryo ventrally injected with 30 ng of BMP-4 morpholino (B) and an embryo ventrally coinjected
with 30 ng of BMP-4 morpholino and 3 ng of BMP-4 protein. (D) Table showing the numbers of the rescue experiment. Using the v2 test to compare BMP-4
morpholino injection to BMP-4 morpholino + BMP-4 protein injection, the rescue is significant at a significance level of a V 0.01.