Chordin Affects Pronephros Development in Xenopus Embryos by Anteriorizing Presomitic Mesoderm

Abstract

Spemann’s organizer emits signals that pattern the mesodermal germ layer during Xenopus embryogenesis. In a previous study, we demonstrated that FGFR1 activity within the organizer is required for the production of both the somitic muscle- and pronephros-patterning signals by the organizer and the expression of chordin, an organizer-specific secreted protein (Mitchell and Sheets [2001] Dev. Biol. 237:295-305). Studies from others in both chicken and Xenopus embryos provide compelling evidence that pronephros forms by means of secondary induction signals emitted from anterior somites (Seufert et al. [1999] Dev. Biol. 215:233-242; Mauch et al. [2000] Dev. Biol. 220:62-75). Here we provide several lines of evidence in support of the hypothesis that chordin influences pronephros development by directing the formation of anterior somites. Chordin mRNA was absent in ultraviolet (UV)-irradiated embryos lacking pronepheros (average DAI<2) but was always found in UV-irradiated embryos that retain pronepheros (average DAI>2). Furthermore, ectopic expression of chordin in embryos and in tissue explants leads to the formation of anterior somites and pronephros. In these experiments, pronephros was only observed in association with muscle. Chordin diverted somatic muscle cells to more anterior positions within the somite file in chordin-induced secondary trunks and induced the expression of the anterior myogenic gene myf5. Finally, depletion of chordin mRNA with DEED antisense oligonucleotides substantially reduced somitic muscle and pronephric tubule and duct formation in whole embryos. These data and previous studies on ectoderm and endoderm (Sasai et al. [1995] Nature 377:757) support the idea that chordin functions as an anteriorizing signal in patterning the germ layers during vertebrate embryogenesis. Our data support the hypothesis that chordin directs the formation of anterior somites that in turn are necessary for pronephros development.

INTRODUCTION

Cells of Spemann’s organizer secrete signaling proteins, such as chordin and noggin, that function nonautonomously to influence the fate of neighboring cells (De Robertis and Kuroda, 2004). The formation of specific tissues such as the pronephros and somitic muscle from the mesoderm germ layer depends upon these organizer signals. The existence of these signals and their importance for inducing specific tissues have been deduced from a variety of embryological and molecular studies. However, in many cases the specific tissues induced by the individual organizer signaling proteins and whether this induction is direct or indirect remain to be elucidated.

Several lines of evidence indicate that pronephros development depends upon signals from the anterior somites. The pronephros consists of three major structures, the pronephric tubules, duct, and glomus. These structures are derived from the intermediate mesoderm and arise in the embryo at a position ventral to somites 3 to 5—the anterior somites. The formation of the pronephros in close proximity to the anterior somites is functionally significant—the anterior somites produce pronephros-inducing signals. The existence of these signals was demonstrated using tissue manipulations in frog and chick embryos. Separation of pronephros progenitor cells from the nearby somitic tissue in chick embryos inhibited pronephros differentiation (Mauch et al., 2000). Culture of chick pronephros precursor cells with quail somites induced pronephros in the chick mesoderm. Similarly, culture of isolated, presumptive pronephric tissue with anterior somites from frog embryos induced pronephric tubule formation, whereas incubation of this tissue with notochord or neural tissue did not (Seufert et al., 1999). These results establish the anterior somites as an important source of pronephros inducing signals. However, very little in known about the signals and mechanisms that direct somitic muscle precursors to form anterior somites.

Organizer signaling affects pronephros development indirectly by means of its affect on the formation of the pronephros-inducing anterior somites. Cortical rotation and, therefore, organizer signaling can be blocked to different extents by exposing fertilized eggs to varying doses of ultraviolet (UV) light. Treatment with high doses of UV light abolishes cortical rotation, and eliminates organizer signaling and the formation of tissues that depend upon this signaling (Gerhart et al., 1984, 1989; Gerhart and Keller, 1986). The resulting embryos lack most mesodermal cell types, including pronephroi, somitic mesoderm, and notochord, and lack neural tissues. Embryos treated with intermediate doses of UV light develop with a subset of mesodermal tissues—pronephroi and somitic muscle are still present in these embryos, but notochord and anterior neural structures are absent (Seufert et al., 1999). These results demonstrate that tissues have different requirements for organizer signaling. The similar sensitivities of the pronephros and somitic muscle to disruptions in organizer function supports the idea that pronephroi are induced by signals from the anterior somites (Seufert et al., 1999). The organizer produces several different signaling molecules, and it is not known which of these are involved in the formation of the anterior somites to influence pronephros formation (De Robertis and Kuroda, 2004).

Clues to the identity or the organizer signals important for both anterior somite formation and pronephros development have come from studies that specifically block FGFR1 function in the organizer cells. Organizer cells expressing DN-FGFR1 are severely disrupted in their ability to emit signals for mesodermal patterning (Mitchell and Sheets, 2001). This disruption in signaling eliminates pronephric tubule and duct formation and reduces somatic muscle formation, indicating that an FGFR1-dependent signal from the organizer is required for pronephros formation and somitogenesis. Intriguingly, disrupting FGFR1 function also significantly reduces the expression of chordin mRNA (Mitchell and Sheets, 2001), but not the expression of other mRNAs such as noggin or Xlim1. These observations raise the possibility that chordin is an important FGFR1-dependent component of organizer signaling that is necessary for anterior somite formation that in turn affects pronephros development.

In this study, we investigated chordin’s role as an organizer-specific signaling protein important for anterior somite formation and pronephros development. Our results strongly suggest that chordin is important for the formation of anterior somites that in turn generate secondary signals that affect pronephros development. These results also suggest that chordin’s role in pronephros development is indirect and occurs by promoting anterior somite formation and the pronephros-inducing signals generated from these cells. The data provide additional support for the hypothesis that chordin and other organizer-specific bone morphogenetic protein (BMP) antagonists, such as noggin, function as anteriorizing factors that modify and pattern the mesodermal germ layers (Lane and Sheets, 2000, 2002; Lane et al., 2004).

RESULTS

Pronephros Formation and Chordin Expression Are Sensitive to Similar Levels of Reduction in Organizer Function

One of the key observations connecting the anterior somites with pronephros formation, is the result that both tissues are sensitive to similar reductions in organizer function (Seufert et al., 1999). We used the same approach to examine the relationship of pronephros and chordin expression. Organizer function can be gradually reduced in Xenopus embryos by incrementally disrupting cortical rotation with increasing amounts of UV irradiation (Seufert et al., 1999). Fertilized eggs were UV irradiated for different times before cortical rotation, and the treated embryos were analyzed for chordin expression by in situ hybridization at the gastrula stage (stage 10.5). Sibling embryos cultured to stage 35 were scored for defects according to the DAI series (Kao and Elinson, 1988) and also analyzed for expression of the pronephric tubule-specific mRNA XSMP-30, using in situ hybridization, and for muscle using immunocytochemistry with the muscle-specific antibody 12/101. Chordin mRNA was highly expressed in embryo batches with average DAI scores of 2 and above (Fig. 1A,B; Table 1), and expression was significantly reduced in embryos with DAI scores below 2 (Fig. 1C; Table 1). In contrast, goosecoid mRNA was absent from embryos with DAI scores 3.3 and below (Fig. 1D,E; Table 1), indicating that goosecoid expression was more sensitive to reductions in cortical rotation compared with chordin. In addition, the vast majority of embryos with average DAI scores of 2 and above developed pronephric tubules and somitic muscle (Fig. 1F-I; Table 1) as previously shown (Seufert et al., 1999), whereas embryos with DAI scores below 2 (Fig. 1J,K; Table 1) contained limited amounts of muscle and pronephric tubules (Fig. 1J,K) (Seufert et al., 1999). Therefore, chordin expression and pronephros formation were sensitive to very similar levels of reduction in organizer function.

Fig. 1.

Chordin expression correlated with the presence of pronephros in ultraviolet (UV)-irradiated embryos. A-E: Fertilized eggs were irradiated with different amounts of UV light before cortical rotation. At stage 10.5, 1/3 of the embryos at each UV dose were analyzed for chordin or goosecoid expression by in situ hybridization. Chordin expression in stage 10.5: DAI5 untreated embryos (A), embryos with an average DAI of 3.0 (B), embryos with an average DAI of 0.5 (C). Goosecoid expression in stage 10.5: DAI5 untreated embryos (D), embryos with an average DAI of 3.3 (E). F-K: Stage 35 embryos representative of each DAI level (5-0). Sibling embryos from each level of UV treatment were scored at stage 35 using the DAI scale and analyzed for pronephros and somitic muscle formation using in situ hybridization for XSMP-30 (blue, black arrowhead) and immunocytochemistry with the 12/101 antibody (brown, red arrowhead).

TABLE 1.

Chordin Expression Correlates With the Presence of Pronephros in Ultraviolet-Irradiated Embryos^a

Average DAI score	% expressing chordin	% with pronephros
0.5 (n = 79)	8	9
1.3 (n = 96)	39	48
2.0 (n = 97)	90	89
2.5 (n = 37)	88	82
3.0 (n = 31)	93	86
3.3 (n = 38)	100	83
3.7 (n = 36)	86	97
DAI	% expressing goosecoid	% with pronephros
2.2 (n = 151)	0	86
3.3 (n = 75)	25	90

^aSee Figure 1.

Chordin-Induced Secondary Trunks Contain Pronephros

Ectopic chordin expression in Xenopus embryos induces secondary trunks containing muscle and neural tissue (Sasai et al., 1994, 1995). A previous study noted the presence of pronephroi in secondary axes (Carroll and Vize, 1999). To more extensively examine chordin-generated secondary trunks for pronephros, a single posterior blastomere of four-cell embryos was injected with chordin mRNA, and the resulting secondary trunks were analyzed for the presence of pronephros using in situ hybridization to detect the pronephros-specific mRNA Xsmp30 or using immunocytochemistry to detect pronephric tubules or ducts with the 3G8 or 4A6 antibodies (Fig. 2C-G; Table 2). Embryos were also analyzed for somitic muscle (stage 32) using immunocytochemistry (12/101 antibodies). The majority of the chordin-directed secondary trunks contained pronephros (Xsmp30 expression; Fig. 2C; Table 2), pronephric ducts and tubules (Fig. 2E,G; Table 2). As previously described, chordin-derived secondary trunks contained somitic muscle, but not notochord (Fig. 2A,B; Table 2; Sasai et al., 1994, 1995). Thus, secondary trunks that result from ectopic chordin expression contained ectopic pronephros.

Fig. 2.

Secondary trunks resulting from ectopic chordin expression contained pronephros. One posterior blastomere of four- to eight-cell stage embryos was injected with chordin mRNA. A,B: Embryos analyzed form muscle and notochord using immunocytochemistry 12/101 and Tor70 antibodies. C,D: At stage 33/34, some embryos were analyzed for pronephros (blue, yellow arrowhead) using in situ hybridization to detect XSMP-30 expression and analyzed for somitic muscle (brown, red arrowhead) formation immunocytochemistry with the 12/101 antibody. E,F: Some embryos were analyzed at stage 38-40 for pronephric duct formation using the 4A6 antibodies. The black arrows indicate pronephric ducts in the secondary and primary axis. G: Injected embryos analyzed at stage 38 for pronephric tubule formation using the 3G8 antibodies. The black arrows indicate pronephric tubules in the secondary and primary axis. H: One posterior blastomere of four- to eight-cell stage embryos was injected with noggin mRNA. Embryos were analyzed at stage 38-40 for pronephric duct formation using the 4A6 antibodies. The black arrows indicate pronephric ducts in the secondary and primary axis.

TABLE 2.

Secondary Trunks Resulting From Ectopic Chordin Expression Contain Pronephros^a

mRNA injected	% secondary trunks	Tissues in secondary trunks
		% pronephros	% muscle	% notochord
Chordin	43 (n = 244)	47 (n = 72)	80 (n = 35)	0 (n = 19)
None	0 (n = 358)	Not applicable

^aSee Figure 2.

Secondary trunks produced as the result of ectopic expression of other BMP antagonists, noggin and the DN-ALK3 receptor, also contained ectopic pronephros (Fig. 2H, and data not shown). We have focused on chordin because our previous results indicated that specifically blocking FGFR1 function in the organizer cells disrupted pronephros and somitic muscle formation and the expression of chordin mRNA (Mitchell and Sheets, 2001), but not the expression of other mRNAs such as noggin or Xlim1. These observations raise the possibility that chordin is an important FGFR1-dependent component of organizer signaling that is necessary for anterior somite formation that, in turn, affects pronephros development.

Depletion of Chordin mRNA Causes Defects in Pronephros Development

The experiments described above provided data consistent with a role for chordin in pronephros formation. However, to test this requirement definitively required perturbing chordin function in the embryo. Toward this end, we analyzed embryos depleted of chordin mRNA by means of injection of an oligonucleotide containing internucleoside phosphate linkages modified with the cation N,N-diethyl-ethylenediamine (DEED)—a DEED-modified antisense oligonucleotide (Gururajan et al., 1991; Bailey et al., 1998; Bailey and Weeks, 2000; Dagle et al., 2000; Dagle and Weeks, 2000; Jallow et al., 2004). The modified internucleoside phosphate linkages in these oligonucleotides promote stability, allowing them to persist in embryonic cells and degrade specific zygotic mRNAs by means of the endogenous RNaseH activity (Bailey et al., 1998; Bailey and Weeks, 2000; Dagle et al., 2000; Dagle and Weeks, 2000). A DEED oligonucleotide complimentary to the 5′ end of chordin mRNA was injected into the two anterior cells in two- to four-cell embryos—the cells that will give rise to Spemann’s organizer. At the gastrula stage, RNAs isolated from injected embryos were examined for chordin mRNA by RNA blot hybridization. Chordin mRNA was specifically depleted in a dose-dependent manner from the chordin DEED antisense oligonucleotide-injected embryos (Fig. 3A, lanes 5-7), but not from controls (Fig. 3A, lanes 1-4). Goosecoid mRNA levels were unchanged by injection of the chordin antisense oligonucleotide, consistent with the DEED oligonucleotide showing specificity for chordin mRNA.

Fig. 3.

Depletion of chordin mRNA in Xenopus embryos caused defects in pronephros and muscle formation. A: Blot hybridization of RNA isolated from uninjected control embryos (lane 1), embryos injected with increasing amounts of the control N,N-diethyl-ethylenediamine (DEED) oligonucleotide (lanes 2-4), and embryos injected with increasing amounts of the chordin DEED-antisense oligonucleotide (lanes 5-7). The blot was hybridized with probes to detect chordin and goosecoid mRNA. B-E: Morphology of chordin mRNA-depleted embryos. B: Embryos injected with chordin-antisense DEED oligo. C: Embryo injected with control DEED oligonucleotide. D: Uninjected control embryo. E: Embryo injected with chordin mRNA + chordin-antisense DEED oligo. F-H: Analysis of pronephric duct formation in chordin-depleted embryos using immunocytochemistry with the 4A6 antibody. F: Injected with chordin-antisense DEED oligo. G: Injected with control DEED oligo. H: Uninjected control. I-K: Analysis of pronephric tubule formation in chordin-depleted embryos using immunocytochemistry with the 3G8 antibody. I: Injected with chordin-antisense DEED oligo. J: Injected with control DEED oligo. K: Uninjected control. L-N: Analysis of somitic muscle formation in chordin-depleted embryos using immunocytochemistry with the 12-101 antibody. L: Uninjected control. M: Injected with control DEED oligo. N: Injected with chordin-antisense DEED oligo.

Most chordin mRNA-depleted embryos exhibited anterior defects such as the reduction or absence of eyes (Fig. 3B; Table 3) and developed with reduced elongation of the trunk and tail (Fig. 3B). Embryos injected with the control oligonucleotide developed with normal morphology, similar to uninjected controls (Fig. 3C,D).

TABLE 3.

Chordin mRNA Depletion Causes Pronephros Defects^a

Sample injected	% Normal	% Head defects	% Pronephros defects
None (n = 62)	100	0	0
Control DEED oligonucleotide (n = 61)	92	0	0
Chordin antisense DEED oligonucleotide (n = 136)	23	62	83
Chordin antisense DEED oligonucleotide+chordin mRNA (n = 40)	75	22

^aSee Figure 3.

To demonstrate that the biological effects were specifically due to chordin mRNA depletion, we performed a rescue experiment. In this experiment, a modified chordin mRNA that lacked complimentarity to the DEED antisense oligonucleotide was injected along with the oligonucleotide into embryos. Significantly, the vast majority of these embryos exhibited no morphological abnormalities (Fig. 3E; Table 3), indicating that the co-injected chordin mRNA was sufficient to rescue defects due to depletion of the endogenous chordin mRNA. These data provided compelling evidence that the defects observed in DEED antisense-injected embryos were due to depletion of the endogenous chordin mRNA. Importantly, the phenotype we observe with the DEED antisense oligonucleotide was similar to that produced by injecting chordin morpholinos (Oelgeschlager et al., 2003).

Having established that we could specifically deplete chordin mRNA from embryos, we tested the effects of this depletion on pronephros formation. Analysis of embryos (stage 38) injected with the chordin antisense oligonucleotide revealed that depletion of chordin mRNA resulted in significant reductions in formation of the pronephric ducts (4A6 staining cells) and pronephric tubules (3G8 staining cells; compare Fig. 3F and I with 3G, H, J and K and see Table 3). These observations demonstrate that Xenopus embryos depleted of the endogenous chordin mRNA develop with defects in pronephros formation.

In addition, analysis of embryos injected with the chordin antisense oligonucleotide revealed that depletion of chordin mRNA resulted in significant reductions in somitic muscle (compare Fig. 3L with 3N and see Table 4). Therefore, Xenopus embryos depleted of the endogenous chordin mRNA develop with somitic muscle defects. Muscle defects were also observed in embryos injected with chordin morpholinos (Oelgeschlager et al., 2003).

TABLE 4.

Chordin mRNA Depletion Causes Somitic Muscle Defects^a

Sample injected:	% Muscle defects
None (n = 82)	0
Control DEED oligonucleotide (n = 66)	0
Chordin antisense DEED oligonucleotide (n = 73)	80

^aSee Figure 3.

Chordin Anteriorizes Somitic Mesoderm

The results presented above indicate that secondary trunks due to ectopic chordin expression contained pronephros (Fig. 2; Table 2) and that depletion of the endogenous chordin mRNA caused significant reductions in pronephros and muscle formation (Fig. 3; Table 3). We next wanted to examine what step in pronephros development chordin influenced.

Pertinently, studies in frog and chick demonstrate that pronephroi forms from the intermediate mesoderm in response to inducing signals emitted from anterior somitic muscle (Seufert et al., 1999; Mauch et al., 2000). In addition, chordin influences embryonic tissues to form anterior derivatives of the ectodermal and endodermal germ layers (Sasai et al., 1995). For example, chordin induces animal cap cells to form anterior neurectoderm and express the anterior neural mRNAs xanf1, otx2, and xif3, as well as the anterior ectodermal mRNA cg-13 (Sasai et al., 1995). Based on these observations, a reasonable postulate was that chordin’s role in pronephros formation was indirect and due to its affect on anterior somite formation. To test this idea, we lineage-traced chordin-injected cells and analyzed their anterior-posterior contribution within the ectopic somite file. Chordin mRNA was co-injected with a lineage tracer (biotin-labeled mini-ruby dextran) into one C4 blastomere of 32-cell embryos. The biotin on the lineage tracer was detected with horseradish peroxidase (HRP)-coupled streptavidin (black staining) and somitic muscle was detected by immunocytochemistry using the 12/101 antibody (brown staining). The anterior-most contribution of the lineage-labeled cells was analyzed in the presence and absence of chordin mRNA. From this analysis, we found that the anterior-most contribution of chordin-injected cells corresponds to somite 3 of the secondary axis (see red arrow Fig. 4B, average of 19 embryos), whereas in the control embryos, C4 cells injected with the lineage tracer alone, the anterior-most labeled somite is somite 9 (see red arrow Fig. 4A, average of 16 embryos). In addition, significantly more lineage labeled cells occupied more anterior positions when chordin mRNA was co-injected as compared with the cells injected with lineage label alone. Therefore, ectopic chordin expression redirected cells from the somitic muscle of the primary axis into a secondary trunk, and the chordin-injected cells now contributed to a more anterior position within the somite file compared with the cells injected with lineage label alone (Fig. 4A,B).

Fig. 4.

Ectopic chordin expression promoted anterior somite formation and ectopic myf-5 expression. A,B: A C4 blastomere of 32-cell stage embryos was injected with biotin-labeled mini-ruby dextran ± chordin mRNA, and embryos were analyzed at stage 32 to detect the injected cells (black, horseradish peroxidase/streptavidin) and somitic muscle (brown, immunocytochemistry with the 12/101antibody). The red arrow indicates the anterior-most dextran-labeled muscle and, therefore, the anterior-most somite contribution from C4 progeny ± chordin mRNA. C,D: A single C4 blastomere of 32-cell stage embryos was injected with green fluorescent protein (GFP) mRNA ± chordin mRNA. Gastrulae exhibiting GFP fluorescence opposite the blastopore lip were split into two groups: half was analyzed at stage 10.5 for myf5 expression by in situ hybridization; the other half was analyzed at stage 33 for ectopic trunk formation and by in situ hybridization for pronephros and somitic muscle using XSMP30 and muscle actin probes (Table 5). The red arrow indicates the blastopore lip. The blue arrowhead indicates the normal myf5 expression (C). The black arrowhead indicates the ectopic myf5 expression (D).

To further examine chordin’s ability to direct somites to more anterior fates, we analyzed the formation of anterior somite progenitor cells following chordin mRNA-injection. Although mRNAs exclusively expressed by the anterior somites have not been identified, myf5 mRNA is an early myogenic marker that is specifically expressed in the anterior somite progenitor cells that flank the blastopore lip (the B2 & C2 derivatives) in stage 10.5 gastrulae (Dale and Slack, 1987; Moody, 1987; Bauer et al., 1994; Lane and Smith, 1999; Lane and Sheets, 2000, 2002). Therefore, myf5 expression at stage 10.5 is an indicator of whether a cell has the potential to form anterior somites.

To examine whether ectopic chordin expression affected myf5 expression, chordin and green fluorescent protein (GFP) mRNAs were injected into one C4 blastomere of 32-cell embryos. At stage 10.5, injected embryos exhibiting GFP fluorescence opposite the blastopore lip were analyzed for myf5 mRNA expression by in situ hybridization. Sibling embryos were analyzed at stage 33 for pronephros and somitic muscle formation. Ectopic myf5 expression was observed in all of the chordin-injected embryos (Fig. 4D; Table 5), and the vast majority of their stage 32 siblings developed a secondary trunk containing both pronephros and somitic muscle (Table 5). Similar results with myf5 expression have been observed with other BMP inhibitors (Dosch et al., 1997; Wardle et al., 1999; Blitz et al., 2000). In addition, injection of chordin morpholinos disrupted endogenous myf5 expression (Oelgeschlager et al., 2003). Together these results supported the postulate that chordin promotes anterior somitic muscle formation.

TABLE 5.

myf-5 Is Ectopically Expressed in Embryos Injected With Chordin mRNA^a

		Tissues in secondary trunks of sibling embryos
mRNA injected	% Gastrulae expressing ectopic myf5	% ectopic pronephros	% ectopic muscle
Chordin	100 (n = 43)	71	100
none	0 (n = 19)	0	0

^aSee Figure 4.

Chordin Directed Pronephroi Are Only Observed in the Presence of Somitic Muscle

Our hypothesis is that chordin affects pronephros formation by directing the formation of anterior somites that, in turn, produce a pronephros inducing signal. Alternatively, other evidence suggests that chordin directly influences BMP signaling that causes the mesodermal germ layer to differentiate as pronephros (Dosch et al., 1997). A critical distinction between these possibilities is the association of muscle and pronephros. If chordin’s affect on pronephros development occurs as a secondary consequence of its affect on somitic muscle, then pronephros and muscle tissues should always form in close association. However, if chordin influences BMP signaling that induces the mesoderm to differentiate as pronephros then these tissues should form relatively independent of one another; we should see pronephroi form without muscle.

To distinguish between these possibilities, the presence of both pronephros and somitic muscle was analyzed in tissue explants expressing chordin. Different amounts of chordin mRNA were injected into the two posterior cells of four-cell Xenopus embryos. Tissue pieces containing the injected cells were excised at the gastrula stage and cultured until sibling embryos reached stage 33. Tissues were analyzed for the presence of pronephros using in situ hybridization to monitor XSMP-30 expression and analyzed for muscle using immunocytochemistry (the 12/101 antibody). Tissue explants receiving the lowest amounts of chordin mRNA contained muscle, but not pronephros (Fig. 5B; Table 6). As more chordin mRNA was injected, tissue explants formed both muscle and pronephros (Fig. 5C-E; Table 6). Pronephroi were never observed in the absence of muscle, and the pronephros in the explants always formed in very close proximity to and at one end of the elongated muscle tissue (Fig. 5C-E). Analysis of pronephric- and muscle-specific mRNA expression by reverse transcriptase-polymerase chain reaction (RT-PCR) also indicated that these tissues formed together in response to similar amounts of chordin mRNA (Fig. 5F). In addition, secondary trunks from chordin-injected embryos also revealed that pronephros was only detected when somitic muscle was present (Fig. 2). Thus although somitic muscle could form without pronephros, pronephros formation was never observed without the close association with muscle. These results support the idea that chordin functions in pronephros development by promoting the formation of anterior somitic muscle and not by directly inducing the mesodermal germ layer to form pronephros.

Fig. 5.

Pronephros formation occurred in association with somitic muscle in explants expressing different amounts of chordin. A-E: Both posterior blastomeres of four-cell embryos were injected with different amounts of chordin mRNA (0-500 pg). A 90-degree wedge of posterior tissue was removed from injected embryos at stage 10⁺, cultured to stage 33/34, and analyzed for both pronephros using in situ hybridization for XSMP-30 (blue stain, black arrowheads) and somitic muscle (brown stain, red arrowhead) using immunocytochemistry. Some of the explants expressing chordin mRNA occasionally developed a cement gland, and these are marked with an by an asterisk (*). None of the control explants from uninjected embryos ever formed a cement gland. The dashed box shows a magnified view of a single explant from each panel. F: Chordin-expressing explants were analyzed for pronephros (XSMP-30) and muscle (muscle actin) specific gene expression using reverse transcriptase-polymerase chain reaction (RT-PCR).

TABLE 6.

Pronephros Formation Occurs in Association With Somitic Muscle in Explants Expressing Chordin^a

	Tissues in explants
Chordin mRNA injected (pgs)	% pronephros + muscle	% muscle only	% pronephros only
0 (n = 59)	0	14b	0
50 (n = 36)	8	47	0
125 (n = 37)	35	49	0
250 (n = 33)	36	55	0
500 (n = 40)	52	48	0

^aSee Figure 5.

^bMost of these explants only expressed scattered 12-101-positive cells and not blocks of muscle tissue.

DISCUSSION

In this report, we demonstrate that chordin is an organizer signaling protein important for pronephros formation in Xenopus embryos. These results provide new insights into the mechanisms of organizer-directed mesodermal patterning and pronephros differentiation, and provide additional evidence that chordin functions as an anteriorizing factor for the germ layers (Sasai et al., 1995; Lane and Sheets, 2000, 2002; Lane et al., 2004).

Chordin Function Is Important for Pronephros Development in Frog Embryos

Patterning of the mesodermal germ layer into specific tissue types is controlled by secreted signals from the organizer (De Robertis and Kuroda, 2004). We previously demonstrated that FGFR1 regulates the organizer signals required for the pronephros formation. Additionally, we demonstrated that FGFR1 was required for chordin mRNA expression but not the expression of other organizer-specific genes (Mitchell and Sheets, 2001). Based on these observations, we hypothesized that chordin is an organizer secreted signal important for pronephros development. The results presented here support this hypothesis. Ectopic chordin expression in whole embryos produced secondary trunks that contained pronephric duct and tubules in close association with the somitic muscle. Significant reductions in pronephric development were observed upon depletion of chordin mRNA with antisense DEED oligonucleotides. In tissue explants intermediate to high doses of chordin directed pronephros always associated with somitic muscle, whereas low doses form predominantly muscle with very little pronephros. Finally, analysis of embryos in which organizer signaling was reduced incrementally by exposure to different doses of UV light revealed that chordin expression and pronephros formation were reduced by the same amounts of UV treatment.

Chordin Affects Pronephros Development by Promoting the Formation of Anterior Somites

Previous studies in frog and chick provided evidence that anterior somitic muscle generates inducing signals necessary for the formation of the pronephros in both frog and chick (Seufert et al., 1999; Mauch et al., 2000). Several pieces of evidence support this idea. Pronephros development could be prevented by surgically eliminating anterior somites, by placing a physical barrier between the anterior somites and intermediate mesoderm, or by reducing the organizer function with UV light treatments. Similarly, combining competent progenitor cells with anterior somites, but not notochord or neural tube, leads to pronephros formation. In addition, these results indicate that the pronephros is the result of inductive interactions between organizer-induced anterior somitic muscle and intermediate mesoderm and not the direct result of inducing signals from the organizer. However, the events responsible for anterior somite formation and the role of organizer signaling molecules in the formation of anterior somites were not addressed in previous studies.

The importance of anterior somites for pronephros formation led us to investigate whether chordin affected pronephros development by means of its effect on anterior somites. Ectopic chordin expression directed posterior cells to significantly more anterior positions within the ectopic somite file and also induced ectopic myf5 expression at stage 10.5—the stage at which myf5 is specifically expressed in the cells that give rise to anterior somites (Dale and Slack, 1987; Moody, 1987; Hopwood et al., 1991; Bauer et al., 1994). Also, reductions in chordin translation by morpholino injection eliminate myf5 expression (Oelgeschlager et al., 2003) and cause reductions in somitic muscle formation. Collectively our data and the results from previous studies lead us to suggest that chordin is an organizer signaling protein that patterns the mesoderm into anterior somites, and this tissue then produces an unknown inducing signal(s) that direct(s) intermediate mesoderm to form pronephros (Seufert et al., 1999; Mauch et al., 2000). These results lead us to propose a model where chordin directly affects anterior muscle development and indirectly affects pronephros development.

Our data are inconsistent with the idea that chordin directly affects pronephros development (Dosch et al., 1997). In this model differentiation of notochord and somitic muscle tissues results from low levels of BMP signaling (high chordin expression), whereas differentiation of pronephros and blood tissues results from high levels of BMP signaling (low chordin expression). However, we observed the opposite in tissue explants. Specifically, pronephros formation required higher or at least equivalent amounts of chordin compared with the amounts required for muscle differentiation. In addition, pronephros formation was always observed in association with muscle and never in isolation, supporting the interdependence of these tissues. These results argue against the idea that pronephros formation is determined by the position of the pronephric precursor cells within a BMP gradient.

BMP Antagonists; Regulators of Anterior Cell Fates

Ectopic chordin induces myf5, a marker of anterior somite progenitors at stage 10.5, suggesting that chordin patterns the mesoderm into anterior somitic fates. Similar results for myf5 have been observed with noggin (Dosch et al., 1997) and dominant-negative BMPR1 (Wardle et al., 1999; Blitz et al., 2000). Additionally, reductions in chordin translation by morpholino injection eliminate myf5 expression (Oelgeschlager et al., 2003), indicating that myf5 expression depends on chordin. In this report, we lineage traced chordin-injected cells and observed that chordin diverts cells that normally contribute to posterior somites to populate more anterior somites in the ectopic axis. Similar results have been obtained with the BMP antagonist noggin (Lane et al., 2004). Previous studies suggested that chordin is a positive regulator of anterior cell fate and anterior genes. Expression of chordin in animal cap cells results in the expression of the anterior neural genes xanf1, otx2, and xif3, as well as the anterior ectodermal gene cg-13 (Sasai et al., 1995). Analysis of the zebrafish mutants chordino and swirl reveals that expression of the anterior endoderm gene her1 is reduced in the absence of chordin and expanded in the absence of bmp2b (Tiso et al., 2002), again suggesting a role for chordin and BMPs in anterior-posterior patterning. These observations support the revised fate map proposed for Xenopus embryos in which cells of the organizer contribute to anterior tissues of the embryo, and support the view that BMP antagonists promote anterior cell fates in the embryo (Lane and Smith, 1999; Lane and Sheets, 2000, 2002; Lane et al., 2004).

Chordin Expression Is Governed by Unique Modes of Regulation

Studies in Xenopus, zebrafish, and ascidians indicate that FGFR1 controls chordin expression, but not the expression of several other organizer-specific mRNAs (Mitchell and Sheets, 2001; Imai et al., 2004; Londin et al., 2005). Our current results demonstrate that chordin expression is minimally affected by intermediate doses of UV light, conditions that totally eliminate goosecoid expression. Furthermore, chordin, noggin, follistatin, and cerberus all exhibit two distinct temporal phases of expression: initiation and maintenance (Wessely et al., 2001). Therefore, it is clear that the chordin gene is subject to distinct modes of regulation that may be unique or shared with only a small subset of organizer-specific genes. This unique regulation may be critical for precisely controlling chordin levels that affect BMP dependent events in the embryo.

EXPERIMENTAL PROCEDURES

Embryos and Injections

Embryos were obtained by standard methods and staged according to Nieuwkoop and Faber. mRNAs encoding chordin and GFP were generated as previously described (Sasai et al., 1994; Zernicka-Goetz et al., 1996). Four-cell embryos received two injections in one posterior blastomere just above and below where the medial cleavage furrow was expected to form. The 32-cell embryos were injected into one C4 blastomere.

For surgical experiments, both posterior blastomeres of four-cell embryos were injected in the marginal zone with 100 pg of either GFP and 25, 50, 125, 250, or 500 pg of chordin mRNA. For experiments examining myf5 expression, one C4 blastomere of 32-cell embryos was injected with 260 pg of chordin mRNA. For chordin depletion experiments, 2, 4, or 6 nL (0.4-1.2 ng) of DEED oligonucleotides were injected into the marginal zones of two- to four-cell embryos.

UV Irradiation

Embryos were dejellied 15 min after fertilization, irradiated with 1,000 - 2,000 μJ of ultraviolet light in quartz cuvettes, and left undisturbed for 3 hr. One third of the irradiated embryos from each UV dose were fixed with MEMFA at stage 10.5. The remaining embryos were scored for defects at stage 33-40 according to the DAI scale (Kao and Elinson, 1988) and fixed.

Tissue Dissections

Cleavage stage embryos were injected with various amounts of chordin and GFP mRNAs. Only embryos exhibiting a GFP signal opposite the upper blastopore lip were used. Posterior tissue was excised from injected stage 10⁺ embryos in agarose-coated dishes containing DFA/BSA solution (Sater et al., 1993) using a hairknife and an eyepiece protractor to remove a 90-degree wedge of tissue. The next day tissues were examined for elongation and fixed with MEMFA when uninjected siblings reached stage 33-34. Pronephros formation was determined by in situ hybridization for the tubule-specific mRNA, XSMP-30 (Sato et al., 2000); somitic muscle differentiation was determined by either immunocytochemistry with the muscle-specific antibody 12/101 (Kintner and Brockes, 1984) or in situ hybridization for muscle actin.

Immunocytochemistry

Stage 30-33 injected embryos were evaluated for somitic muscle using the muscle-specific antibody 12/101 (Kintner and Brockes, 1984) and goat anti-IgG secondary antibody coupled to either HRP or alkaline phosphatase (AP). Notochord was detected in stage 30-32 embryos using the notochord-specific antibody Tor70 and an HRP-coupled anti-IgM antibody. Pronephric tubules were detected in stage 38-40 embryos using the tubule-specific antibody 3G8 (Vize et al., 1995), and an anti-IgG AP-coupled secondary antibody. Pronephric ducts were detected in stage 38-40 embryos using the duct-specific antibody 4A6 (Vize et al., 1995), and an anti-IgG AP-coupled secondary antibody. HRP-coupled antibodies were detected with diaminobenzidene substrate; 5-bromo-4-chloro-3-indolyl phosphate (BCIP) or BM purple substrates were used to detect the AP-coupled antibody. Pronephros defects were assessed by visual comparison of tubule and/or duct mass to control embryos. Pronephros formation was considered defective if the tubules and/or ducts were either entirely absent or greatly reduced in mass, as indicated by the absence of or strong reduction in 3G8- or 4A6-positive cells.

In Situ Hybridization

In situ hybridization was performed as described (Harland, 1991). Myf 5 probe was synthesized by linearizing the Xmyf-5-2 template (Hopwood et al., 1991) with HindIII and transcribing with SP6 RNA polymerase.

RT-PCR

RNA isolation and cDNA synthesis was performed using standard protocols. PCR amplification was performed with the following primer pairs ODC (F:GCCATTGTGAAGACTCTCTCCATTC, R:TTCGGGTGATTCCTTGCCAC [Heasman et al., 2000]), muscle actin (F:TCCCTGTACGCTTCTGGTCGTA, R:TCTCAAAGTCCAAAGCCACATA[Stutz and Spohr, 1986]), and XSMP-30 (F:GGCAAAAGCTCAAATCGAAGG, R:ATTGAGGCTGGCGTTTCTTCC).

Oligonucleotide Synthesis

Partially modified DEED oligonucleotides were synthesized as described (Dagle et al., 2000; Dagle and Weeks, 2000). The chordin antisense sequence used was as follows: T+T+C+T+A+TGGACCAT+G+A+G+C+C+ where “+” indicates a modification of the phosphate linkage with DEED. The control oligonucleotide sequence was identical to the chordin sequence but contained additional DEED modifications that sterically prevent recognition of these oligonucleotides by RNase-H.