February 5, 2002;
Action range of BMP is defined by its N-terminal basic amino acid core.
During early development, cells receive positional information from neighboring cells to form tissue
patterns in initially uniform germ layers. Ligands of the transforming growth factor (TGF-beta) superfamily are known to participate in this pattern formation. In particular, activin has been shown to act as a long-range dorsalizing signal to establish a concentration gradient in Xenopus. In contrast, BMP-2 and BMP-4, other members of the family, appear to influence and induce ventral
fates only where they are expressed. This raises a question as to how the action of BMPs is tightly restricted to the region within and around the cells that produce them. Here, we have demonstrated that a basic core of only three amino acids in the N-terminal region of BMP-4 is required for its restriction to the non-neural ectoderm
as its expression domain. Our results also suggest that heparan sulfate proteoglycans bind to this basic core and thus play a role in trapping BMP-4. The present study is the first to identify the critical domain of BMP that is responsible for its interaction with the extracellular environment that restricts its diffusion in vivo.
[+] show captions
N-Terminal Structure of BMP and Its Variants and Their Biological Activities
(A) The N-terminal amino acid sequences of mature BMP subfamily members have a core of basic amino acids, which makes them unique in the TGF-β superfamily. Basic amino acids (H, K, and R) are indicated in red. Comparison of the BMP-2/4 subfamily ligands among several species. The motif is highly conserved among several species from flies to humans. Comparison of other ligands of the TGF-β superfamily. Amino-acid sequence of Xenopus wild-type (WT) BMP-4 and the BMP-4 variants used in this study. δ1BMP-4 lacking the eight amino acids KQQRPRKK and δ2BMP-4 lacking the three amino acids RKK.
(B) Analysis of Xmsx-1 and Xvent-1 expression in the animal cap by RT-PCR. Animal caps were either uninjected (lane 1) or injected with 1 pg (lanes 2, 5, and 8), 10 pg (lanes 3, 6, and 9), or 100 pg (lanes 4, 7, and 10) of WT, δ1, or δ2BMP-4. Lane 11 shows the expression of each marker in whole embryos (stage 11), and lane 12 shows the control reactions with no RT step. Xmsx-1 and Xvent-1 appeared to be slightly less efficiently induced at each dose of δ1BMP-4 mRNA and similarly induced at each dose of δ2BMP-4 mRNA compared with the same dose of WTBMP-4.
(C) Binding assay of WT and δ2BMP-4 to sBMPR by BIACORE. Equal amounts of WT or δ2BMP-4 immunoreactivity measured by Western blot analysis (inset) were injected to flow over the sensor chips as an analyte. Arrowheads represent the initiation and termination of injections. The same binding profiles were obtained wFigure 1
BMP-4 Variants Acquired a More Long-Range Effect Compared with Wild-Type
(A) Schematic representation of the animal cap conjugate assay. (B) The processed mature form (∼20 kDa) of WTBMP-4 or the BMP-4 variants in the culture medium and cell lysate from animal caps was detected by Western blot analysis using an anti-BMP-4 antibody (Ab97). For the negative controls, an uninjected animal cap was used (Uninj.). The levels based on the immunoreactivity of the processed mature form of the BMP-4 variants were lower than or similar to that of WTBMP-4 (∼20 kDa). (C and F–I) Immunostaining with PS1 antibody (anti-BMP-driven pSmads: green) of conjugated animal caps. Donor animal caps were either uninjected (C) or injected with 300 pg mRNA of wild-type BMP-4 ([G]; WTB4), δ1BMP-4 ([H]; δ1B4), or δ2BMP-4 ([F and I]; δ2B4). Recipient animal caps were injected with either the lineage tracer RLDx (red) only ([C and G–I]; R) or RLDx with 500 pg mRNA for a dominant-negative BMPIA receptor ([F]; R + dn-BR). (D and E) Immunostaining with PS2 antibody (anti-activin-driven pSmads: green). Donor animal caps were injected with 50 pg of activin βB mRNA (WTAct). Recipient animal caps were injected with either RLDx only ([D]; R) or RLDx with 1.25 ng mRNA for a dominant-negative activin type II receptor ([E]; R + dn-ActR). Animal caps that were uninjected or injected with dominant-negative receptors were not stained with PS1 or PS2 (C, E, and F). (J–L) In situ hybridization of conjugated animal caps with an Xbra antisense probe (purple). Recipient animal caps were injected and marked with lineage tracer RLDx (data not shown). Donor animal caps (above the dotted line) were injected with 300 pg of mRNA for WTBMP-4 (J) or the BMP-4 variants (K and L). Uninjected donor animal caps did not express Xbra (data not shown). When the mRNAs for the BMP-4 variants and activin βB were injected into the donor animal cap, the domain marked by the nuclear staining of pSmads (D, H, and I) and the Xbra expression domain ([K], [L], and data not shown) were markedly enlarged in the recipient animal caps.
The BMP-4 Variants Have Enhanced Ventralizing Activity in the Xenopus Embryo Compared with WTBMP-4, When Ventrally Overexpressed
Embryos were either uninjected or injected in the two ventral blastomeres at the four cell stage with 100 pg of mRNA encoding WTBMP-4 or the BMP variants (A). Ventralized embryos injected with 100 pg mRNA were scored by DAI, and its number of embryos is represented as percentages. Similar results were observed when these mRNAs were injected into two ventral blastomeres at the 16 cell stage (data not shown). (B) In situ hybridization with an Xvent-1 antisense probe (purple) of an injected or an uninjected embryo at stage 11. Dorsal is to the right, and the animal pole is at the top. (C) Analysis of Xmsx-1 expression in the VMZ or DMZ of uninjected (uninj.) or injected embryos (WT, δ1, and δ2) by RT-PCR. Lane 9 shows the expression of Xmsx-1 in whole embryo (stage 11), and lane 10 shows the control reactions with no RT reaction. The Xmsx-1-inducing activity of both the BMP-4 variants in the DMZ was markedly enhanced.
WTBMP-4 Is Trapped by the ECM but the BMP-4 Variants Are Not
(A) Mature (∼20 kDa) and degraded (∼19 kDa) forms of WT or δ2BMP-4 in the culture medium from animal caps (AC) or dissociated animal cap cells (Diss.) were detected by Western blot analysis (lanes 2, 3, 5, and 6). For the negative controls, uninjected dissociated or undissociated animal caps were used (lane 1 or 4). For the positive controls, WT or δ2BMP-4 in the culture medium from HEK 293T cells was used (lane 7 or 8). Although WT and δ2BMP-4 were almost equally produced and secreted into the medium from dissociated animal cap cells (lanes 2 and 3), more δ2BMP-4 diffused into the medium from the ECM of the cell surface of the intact (undissociated) animal cap (lane 6) than WTBMP-4 (lane 5).
(B) WT or δ2BMP-4 obtained from HEK293T cells (lane 7 or 8) was incubated with heparin-sepharose. The level of the mature form of the BMPs in the bound (lanes 1 and 2) or passed (3 and 4) fraction was detected by Western blot analysis. For the negative controls (lanes 5 and 6), sepharose beads with no immobilized protein were used. WTBMP-4 was detected in the heparin bead binding fraction (lane 1) but not in the passed fraction (lane 3), while a large portion of δ2BMP-4 was detected in the passed fraction (lane 4) and less in the heparin bead binding fraction (lane 2).
(C) WTBMP-4 acquired a more long-range effect in animal caps treated with heparitinase I than in intact animal caps. Donor animal caps were injected with 300 pg of mRNA for WTBMP-4 (above the red line), and recipients were injected with lineage tracer RLDx. The left upper panel shows an example of RLDx lineage labeling and its fluorescence image, and the red lines are drawn along the boundary between the donor and recipient animal caps. Neither the donor nor recipient animal caps were treated with heparitinase I (d−/r−). The donor animal cap (d+/), recipient animal cap (/r+), or both were from embryos treated with heparitinase I. In situ hybridization of a conjugated animal cap with an Xbra antisense probe (purple). When either the donor or the recipient animal cap was dissected from an untreated (buffer-injected) embryo, WTBMP-4 failed to induce Xbra expression in the entire recipient animal cap (d−/ and/or /r−). When both caps were dissected from heparitinase I-treated embryos, even WTBMP-4 induced Xbra expression that extended to the opposite end of the donor cap at a higher frequency (d+/r+).
Figure S1. Substitution of the Three Basic Amino Acids RKK, which
Are Deleted in 2BMP-4, to ALA Residues Affects BMP-4’s Action
Range and Binding Affinity to Heparin
(A) The N-terminal amino acid sequences of WTBMP-4 and substituted
(B) Immunostaining of conjugated animal caps with PS1 antibody
(green) or in situ hybridization with Xbra antisense probe (purple).
Donor animal caps were injected with WT or alaBMP4 mRNA. (See
legend to Figure 2 in the main text and the section Conjugated
Animal Cap Assay) When the mRNA of alaBMP-4 was injected into
the donor animal cap, the domain marked by the nuclear staining
of pSmads (upper panels) and the Xbra expression domain (lower
panels) were markedly enlarged and expanded over the recipient
(C) Mature form of WTBMP4 or alaBMP4 in the bound or passed
fraction was analyzed by Western blot (see legend to Figure 4B and
the section Heparin Binding Assay). We detected WTBMP-4 in the
heparin bead binding fraction but not in the passed fraction, while
a large portion of alaBMP-4 was detected in the passed fraction
and less was in the heparin-bead binding fraction.