The Notch-effector HRT1 gene plays a role in glomerular development and patterning of the Xenopus pronephros anlagen

The pronephros, which is the functional embryonic kidney of amphibian and fish embryos, is used as a model to study human kidney development and disease. In amphibians, the pronephros is a paired organ that consists of a single nephron composed of three basic components: (1) the glomus, which is the site of blood filtration; (2) the tubules, where filtrate resorption occurs; and (3) the duct, which carries the urine to the cloaca(Brändli, 1999; Vize et al., 2003; Ryffel, 2003; Jones, 2005).

All three components of the pronephros develop within the intermediate mesoderm right posterior to the head. In Xenopus, at late neurula(around stage 21), cells in the lateral layer of the intermediate mesoderm below somites 3-7 start to condense. Cells from the dorsoanterior region of the pronephric field will form the tubules, while those from the ventroposterior region migrate posteriorly out of the original kidney primordium to give rise to the majority of the duct. Concomitantly, cells in the adjacent medial layer undergo morphogenesis to form the glomus. The molecular mechanisms that control the early specification of the pronephros have been well studied in frog and chicken(Brennan et al., 1998; Brennan et al., 1999; Seufert et al., 1999; Obara-Ishihara et al., 1999; Carroll and Vize, 1999; Mauch et al., 2000; Chan et al., 2000; James and Schulteiss, 2005). By contrast, much less is known about the gene products that pattern the pronephric anlagen. The Wilms' tumor xWT1 gene encoding a zinc finger transcription factor, which is expressed around the dorsal and anterior border of the future pronephros, is thought to have a role in the specification of the glomus by suppressing tubule and duct gene expression(Carroll and Vize, 1996; Wallingford et al., 1998). Evi1 is another gene encoding a zinc finger transcription factor that may play a role in the partitioning of the pronephros; it is selectively expressed in the ventroposterior part of the pronephros anlagen, giving rise to the distal tubule and duct compartments(Van Campenhout et al., 2006). Notch signaling has also been shown to play an important role in the partitioning of the pronephros, inhibiting duct and distal tubule differentiation in the dorsoanterior region of the anlagen, where cells are normally fated to form proximal tubules and to increase the expression of the xWT1 gene (McLaughlin et al., 2000; Van Campenhout et al., 2006). Studies in mice have demonstrated that Notch signaling is similarly required during metanephros development for glomerular podocyte and proximal tubule fates (McCright et al., 2001; Wang et al., 2003a; Cheng et al., 2003; Cheng and Kopan, 2005). However, the stages of nephron morphogenesis that are dependent upon the activation of Notch remain unidentified.

In Xenopus, the XHRT1 gene (also named Hey1/HERP2/Hesr-1/CHF2), encoding a downstream basic helix-loop-helix Orange (bHLH-O) mediator of Notch signaling,has been shown to be expressed in numerous tissues during development,including the pronephros, and to be responsive to Notch signaling(Rones et al., 2002; Pichon et al., 2002). XHRT1 is a member of the HRT subfamily of bHLH-O proteins that forms heterodimers with hairy proteins through the bHLH-O and downstream sequences, and represses transcription in a groucho-independent manner(Iso et al., 2003; Taelman et al., 2004; Pichon et al., 2004). In the embryonic mouse metanephros, several intracellular Notch effectors have been found to be expressed in a segment-specific manner in early nephrons, but nothing is known as yet about their role in patterning cell fate decisions(Leimeister et al., 2003; Piscione et al., 2004; Chen and Al-Awqati, 2005).

Here, we show that XHRT1, when compared with the other bHLH-O factors expressed in the developing kidney, plays a predominant role in the pronephros as a Notch effector, being required for glomus formation and for proximodistal patterning of the pronephric primordium. We show that this is due not only to its earlier temporal expression pattern, but also to intrinsic properties of the protein that the HES proteins lack.

The hGR/X-Su(H)/Ank, X-Su(H)DBM, Xhairy2b-MT-hGR, hGR-ESR9,hGR-ESR10 and XHRT1-MT-hGR expression plasmids have been described (Chitnis et al.,1995; Wettstein et al.,1997; Taelman et al.,2004). The following plasmids were constructed: pCS2+MT XHes2-hGR, encoding an inducible, Myc-tagged XHes2 protein; pCS2+MT XHRT1-XHes2-hGR, encoding an inducible, truncated form of XHRT1 (amino acids 1 to 160) to which has been attached the carboxy terminus of XHes2(amino acids 129 to 191); and pCS2+MT Hes2-XHRT1-hGR, encoding an inducible, truncated form of XHes2 (amino acids 1 to 127) with the carboxy terminus of XHRT1 (amino acids 161 to 294). In addition, pCS2+XHRT1a-mut-MT-hGR, derived from pCS2+XHRT1a-MT-hGR,incorporates several mismatches (small letters) in the MO target sequence:5′-ATGAAaaGaGGcCAtGAtTA (predicted start codon is underlined). pCS2+XHRT1-eGFP encodes the full-length XHRT1 protein fused with eGFP at its carboxy terminus. pCS2+xWT1-eGFP encodes a similar fusion with the N-terminal part of xWT1 (amino acids 1-91).

Antisense morpholinos for xWT1, XHRT1, esr9, esr10 and Xhairy2b (GeneTools)consist of the following sequences (sequence complementary to the predicted start codon is underlined):

XHRT1 MO, 5′-TAGTCGTGTCCCCGCTTCATGGCTG-3′;

xWT1a MO, 5′-CATATCCCGGACATCAGACCCCATC-3′;

xWT1b MO, 5′-CATATCCCGCACATCAGATCCCATC-3′;

esr9 MO, 5′-CTGTCTGGTAATGGGATGTGATGGA-3′;

esr10 MO, 5′-CTGTTTAGTAAGTGGATATGATGGA-3′;

Xhairy2a MO, 5′-ATGGTATCTGCGGGCATGTTCAGTT-3′;

Xhairy2b MO, 5′-GGCATGTTCAGATGTTGTATCCGGA-3′.

Individual MOs, or a mixture of both MOs for xWT1, were injected at 15 ng/blastomere.

Xenopus eggs were obtained from hormone-induced (chorionic gonadotropin, Sigma) adult female frogs and fertilized using standard methods. Capped mRNAs were transcribed using the mMessage mMachine Kit (Ambion). For targeting the pronephros, synthetic RNA (500 pg) was injected into one blastomere in the lateral marginal zone of 8-cell stage embryos. nuc-lacZ mRNA (100-250 pg/blastomere) was used as a lineage tracer. Induction of hGR constructs in embryos was performed by addition of dexamethasone (Dex; 10 μM; Sigma). Injected embryos were fixed in MEMFA,stained for β-galactosidase activity with 5-bromo-4-chloro-3-indolyl-β-galactopyranoside (X-Gal, Bioline) or 6-chloro-3-indolyl-β-D-galactoside (Red-Gal, Research Organics) and stored in ethanol at -20°C. Only embryos that were phenotypically normal and show lacZ staining in the pronephic region were scored.

Whole-mount in situ hybridization was carried out as previously described(Harland et al., 1991). X-Serrate-2 was identified through a search of the EST database (NIBB #XL054p21). The X-Serrate-2 plasmid linearized with EcoR1 was transcribed with T7 polymerase. Plasmids used for generating the other in situ hybridization probes are: XHRT1(Pichon et al., 2002), XHairy2b, Xhairy1 (Tsuji et al.,2003; Taelman et al.,2004), Evi1, xSat1, xPDZK1(Van Campenhout et al., 2006), xCLC-K (Vize, 2003), ESR-4, ESR-5 (Jen et al.,1999), ESR-6e(Deblandre et al., 1999), esr9, esr10 (Li et al.,2003), XHes2 (M.S., unpublished), Hes6(Koyano-Nakagawa et al.,2000), XSMP-30 (Sato et al., 2000), xWT1(Carroll and Vize, 1996), Pax8 (Heller and Brandli,1999), X-Delta-1, N-tubulin(Chitnis et al., 1995), Ep. keratin (Jonas et al.,1985), X-Serrate-1(Kiyota et al., 2001) and Nephrin (Gerth et al.,2005). For sections, embryos were gelatine embedded and vibratome sectioned.

Western blot analysis was performed as described(Taelman et al., 2004) using the 9E10 anti-Myc and anti-β-tubulin monoclonal antibodies (Sigma).

In a search for downstream effectors of Notch signaling in the early pronephros anlagen, we surveyed by whole-mount in situ hybridization the expression pattern of the previously identified Xenopus bHLH-O genes XHRT1 (Pichon et al.,2002), Xhairy1(Taelman et al., 2004), Xhairy2b (Tsuji et al.,2003), ESR-1(Wettstein et al., 1997), ESR-4, ESR-5 (Jen et al.,1999), ESR-6e(Deblandre et al., 1999), esr9, esr10 (Li et al.,2003), XHes2 (M.S., unpublished) and Hes6(Koyano-Nakagawa et al.,2000), and compared their expression with that of the ligands X-Delta-1 (Chitnis et al.,1995), X-Serrate-1(Kiyota et al., 2001) and X-Serrate-2, identified by EST mining. ESR-1, ESR-4, ESR-5,ESR-6e, XHes2 and Hes6 mRNA were not detected in the pronephric anlagen between stages 20 and 36. XHRT1, Xhairy1, Xhairy2b, esr9 and esr10 are all expressed from the early tailbud stage in the dorsoanterior region of the developing pronephros, XHRT1 expression being detectable slightly earlier than the others.

During early tailbud stages, XHRT1, esr9 and esr10expression appears localized to the most dorsoanterior portion of the pronephric anlagen, whereas Xhairy1 and Xhairy2b are more broadly expressed within the pronephric mesoderm(Fig. 1A,E,G,I,K,M). In transverse sections of stage 20-23 embryos, XHRT1, esr9 and esr10 transcripts are predominantly found, similarly to xWT1transcripts (Carroll and Vize,1996), around the dorsoanterior border of the pronephros anlagen(Fig. 1B,D,F,H,J). By contrast, Xhairy1 and Xhairy2b expression is detected both around and inside the developing pronephros, strong Xhairy2b staining being also observed in the sensorial layer of the ectoderm covering the pronephros anlagen (Fig. 1L,N). X-Delta-1 is the only Notch ligand to be expressed in the developing pronephros at early tailbud stage. Its expression is detected in the lateral mesodermal layer in cells surrounded by the XHRT1-positive cells(Fig. 1C).

During late tailbud to early tadpole stages, XHRT1 expression demarcates the most dorsoanterior portion of the pronephros, whereas esr9,esr10, Xhairy1 and Xhairy2b occupy more ventral regions. Within this dorsoanterior portion of the pronephros, high levels of XHRT1expression progressively become restricted to the tip of the forming tubules,while expression of the other bHLH repressors remains broader(Fig. 2A,C,E,G,I,K,R). Sectioning of those embryos revealed that during this period, expression of XHRT1, esr9 and esr10 disappear in the medial layer and that they are now actively transcribed in the dorsoanterior portion of the pronephros anlagen itself. Xhairy1 and Xhairy2b, which were initially transcribed in both layers, are now predominantly expressed in the lateral layer too (Fig. 2B,D,F,H,J,L). During this period, all Notch ligand genes are expressed in the dorsoanterior portion of the pronephros anlagen in a region slightly ventral to XHRT1 expression. Whereas X-Delta-1expression is restricted to a band just below XHRT1, Serrate2expression is broader (Fig. 2M-Q). X-Serrate-1 is expressed similarly to X-Serrate2 at that stage (data not shown). Although pronephric expression of XHRT1, esr9, esr10 and Xhairy1 is no longer detectable at late tadpole stages (stage 35), Xhairy2b expression is maintained in the proximal and distal tubules. X-Serrate-1 is the only Notch ligand to remain expressed in the dorsoanterior portion of the pronephros at that time, and its expression appears to be similar to that of Xhairy2b (Fig. 2S,T). The spatially and temporally distinct expression patterns of those bHLH-O repressor-encoding genes suggest that they may have non-identical functions during pronephros development.

The XHRT1 gene has previously been shown to be responsive to Notch signaling in the pronephros (Rones et al.,2002). We investigated whether Notch signaling also affects the expression of the other bHLH-O genes. To study the consequence of activation of Notch signaling in the developing pronephros without affecting earlier developmental steps, we used an hormone-inducible form of the transcription factor Su(H) that mediates Notch signaling(Wettstein et al., 1997). Injected embryos were induced with dexamethasone at stage 18 and assayed for expression of the different bHLH-O genes between stages 25-30. We observed that, as in the case of XHRT1, activation of Notch signaling using an inducible Notch ICD-ankyrin fusion of Su(H) increased the pronephric expression of esr9 (n=24), esr10 (n=35), Xhairy1 (n=26) and Xhairy2b (n=25) in all injected embryos (Fig. 3A,parts a-j). In many embryos, expression of those bHLH-O genes expands within the posterior part of the intermediate mesoderm, the strongest staining being detected in the case of esr9, esr10 and XHRT1 in the lateral part of the intermediate mesoderm, while Xhairy1 and Xhairy2b expression is found in both layers (arrowheads). Xhairy2b, which is expressed in the ectoderm overlying the pronephros anlagen, was also strongly activated in the ectoderm (Fig. 3B, parts a-e). As reported in the case of XHRT1, suppression of Notch signaling using a dominant-negative form of Su(H) decreased the expression in the pronephros of esr9, esr10, Xhairy1 and Xhairy2b(n=8 for esr9, 13 for esr10, 29 for Xhairy1 and 31 for Xhairy2b; Fig. 3A, parts k-t). Thus, esr9, esr10, Xhairy1 and Xhairy2b may function together with XHRT1 in the developing pronephros as downstream mediators of Notch signaling.

As our results indicate that XHRT1 is expressed in a dynamic manner in the pronephros, we wanted to know whether this reflects a difference in the temporal responsiveness of the medial and lateral layers to activation of Notch signaling. Therefore, we analysed XHRT1 expression in earlier embryos (stage 23). As observed in stage 25-30 embryos, activation of Notch induced XHRT1 expression in both layers(Fig. 3B, part f). By contrast, xWT1 is only activated in the medial layer at all stages analysed(Fig. 3B, part g). Thus, the successive expression of XHRT1 in the medial and lateral mesodermal layers is not a consequence of a difference in the temporal competence of the two layers to respond to Notch activation.

The xWT1 gene is thought to play an important role in the development of the pronephros by repressing lateral-specific gene expression in the portion of the pronephric mesoderm fated to form the glomus(Wallingford et al., 1998; Van Campenhout et al., 2006). To determine whether xWT1, which is activated at about the same time as XHRT1 in the pronephros, is required for the expression of bHLH-O repressors in the forming glomus, we generated antisense MOs that block the translation of both xWT1 pseudoalleles. Injection of those MOs specifically blocked the translation in vitro and in vivo of its target mRNA(Fig. 4A; data not shown). Embryos injected with those xWT1 MOs were analysed by in situ hybridization with XHRT1 and other pronephric markers. As expected, knockdown of xWT1 abolished the expression of nephrin, a marker of glomerular podocytes, which is directly activated in mice by WT1(Wagner et al., 2004; Guo et al., 2004; Gerth et al., 2005) (100%inhibited, n=199; Fig. 4B,C). Expression of the XSMP-30 proximal tubule (82%unaffected, n=72) and the Pax8 (83% unaffected, n=24) and Evi1 (90% unaffected, n=20) early pronephric markers was not affected (Fig. 4D-K). Interestingly, knockdown of xWT1 decreased the early glomus-specific expression of XHRT1 (48% inhibited, n=125)but did not perturb its late expression in the pronephros anlagen (80%unaffected, n=35) (Fig. 4L-O). Together, these experiments are consistent with the idea that xWT1, which promotes glomus formation, may play a role in XHRT1early expression. They also suggest that there may be factors other than xWT1 that repress the expression of lateral-specific genes in the developing glomus.

To determine if XHRT1, esr9, esr10 and Xhairy2b are functioning as mediators of Notch signaling in the pronephros anlagen, we used previously described hGR-inducible constructs(Taelman et al., 2004). Embryos were injected with the different hGR constructs mixed withβ-galactosidase mRNA as a lineage tracer. In the injected embryos, we examined how the expression of those bHLH-O genes affects pronephric gene markers. We first examined the ability of XHRT1, which is the earliest and strongest bHLH-O gene expressed in the pronephric primordium, to inhibit distal tubule and duct cell fates. Addition of dexamethasone at the end of neurulation (stage 18) resulted in a hormone-dependent inhibition at early tadpole stage of the formation of the distal tubule and duct as revealed by the Evi1 marker (94% inhibited, n=32)(Van Campenhout et al., 2006). This phenotype is similar to that previously described upon activation of Notch signaling using a hGR/Su(H)/Ank construct (McLaughlin et al., 2000)(Fig. 5A-F). Similar results were observed using the xClC-K marker (data not shown). The same experiments were repeated for esr9, esr10 and Xhairy2b. However, unlike XHRT1,although effective in inhibiting primary neurogenesis, esr9, esr10 and Xhairy2b did not inhibit Evi1 expression (unaffected: 72%, n=21; 76%, n=33; 88%, n=40; for esr9, esr10 and Xhairy2b, respectively; Fig. 5G-O). XHes2, another bHLH-O repressor, which is not found in the pronephros (M.S., unpublished), was also inefficient at repressing Evi1 (80% unaffected, n=30)(Fig. 5P-R). It has been suggested that XHRT1 may participate in regulating aspects of gene expression that are linked to cell-cycle control and apoptosis(Wang et al., 2003b; Huang et al., 2004). The inhibition of distal tubule and duct formation could thus be explained by a decrease of proliferation or the apoptotic elimination of pronephric cells. However, immunostaining using an antibody recognizing phosphorylated histone H3, and TUNEL analysis of injected embryos, revealed no change in the pattern of mitotic and apoptotic cells (data not shown). We therefore conclude that activation of XHRT1, but not that of esr9, esr10 or Xhairy2b, may play a crucial role during pronephros formation in the inhibition of distal tubule and duct cell fates.

Activation of Notch signaling in the pronephros anlagen has been shown to perturb the differentiation of the tubule network and to increase xWT1 expression (McLaughlin et al., 2000). We therefore compared the ability of XHRT1 with that of Su(H)Ank to modulate the expression of the XSMP-30 proximal tubule markers and the xWT1 and nephrin glomus markers. We first observed that the responsiveness of these markers to activation of Notch signaling was temporally specific. Activation of Notch signaling before stage 25, when Notch bHLH-O effectors are expressed in the glomus, efficiently induced xWT1 expression and inhibited XSMP-30 expression. Later activation, when they are expressed in the dorsal part of the pronephros anlagen, had no effect on xWT1 expression. In a few embryos activated between stage 22 and stage 27, XSMP-30 expression was expanded(Fig. 6A). In contrast to the effects observed upon injection of hGR/Su(H)/Ank mRNA, early or late activation of XHRT1 has no effect on xWT1 (77% unaffected, n=39) and nephrin (85% unaffected, n=20). Although early activation decreases the expression of XSMP-30 in some embryos(58% inhibited, n=33), late activation (stage 20-27) did not affect its expression (n=130; Fig. 6B-M). Similar results were obtained with other proximal tubule markers (xSat1, xPDZK1, 3G8; data not shown). We conclude that activation of Notch in the pronephric mesoderm is essential first for glomus and then for proximal tubule fates, and that XHRT1 only mediates part of its effects.

To further determine the importance of XHRT1 as a mediator of Notch signaling in the pronephros, we co-injected embryos with mRNA encoding Su(H)DBM together with XHRT1-MT-hGR mRNA. Injected embryos were induced at stage 22 and assayed for the expression of Evi1 and XSMP-30. We observed that early inhibition of Notch signaling in untreated embryos reduces the expression of the proximal tubule marker XSMP-30 and, as previously reported, elevates the expression of the distal tubule and duct marker Evi1(Fig. 7A-D,I). By contrast, in dexamethasone-treated embryos, we observed a reduction of both XSMP-30 and Evi1 expression(Fig. 7E-H,I). Thus, XHRT1 could reverse the effect of Su(H)DBM on Evi1but is not sufficient to restore the expression of XSMP-30. Together,these results indicate that XHRT1 functions as an important downstream effector of Notch signaling. Compared with the other bHLH-O repressors, it appears to play a specific role in early pronephros development, contributing to the inhibition of distal tubule and duct cell fates in cells that form the glomus and the proximal tubules.

To determine whether XHRT1 is required for glomus and proximal tubule development, we generated a MO that targets a 100% conserved 25 bp stretch,including the AUG initiation codon, in the two XHRT1 pseudoalleles. In vivo and in vitro, the XHRT1-MO specifically and efficiently blocks the translation of the corresponding mRNA (Fig. 8A). XHRT1-MO-injected embryos displayed a decrease in the expression of all proximal tubule-specific markers tested [including XSMP-30 (72%, n=59), xPDZK1 (88%, n=17)and xSat1 (68%, n=15)] and downregulated the expression of xWT1 (77%, n=40) and nephrin (88% inhibited, n=17; Fig. 8B-I; data not shown). Transverse sections revealed that the glomus was present and that in many injected embryos, the proximal tubules were reduced in size(Fig. 8J,K). These effects are not due to a change in the pattern of mitotic and apoptotic cells, as revealed by analysis of the injected embryos by phosphorylated histone H3 immunostaining and TUNEL (data not shown). Expression of the distal tubule and duct markers Evi1 and xClC-K appears unaffected(Evi1 64%, n=37; xClC-K 72%, n=11; Fig. 8L,M; data not shown). Expression of other markers, such as Ep. keratin, and N-tubulin, was also unaltered (Ep. Keratin, none inhibited, n=27; N-tubulin, none inhibited, n=25; Fig. 8N,O). Injection of an MO designated against esr9, esr10 or xHairy2b that efficiently inhibits their target mRNA did not affect the expression of any of the pronephric genes tested (see Fig. S1 and S2 in the supplementary material),which further supports the idea that XHRT1 has a specific function in the partitioning of the pronephros anlagen.

In order to determine whether the phenotype caused by the injection of XHRT1-MO can be rescued by co-injection of XHRT1 mRNA, we generated an inducible XHRT1a-mut construct (XHRT1a-mut-MT-hGR). Co-injection with XHRT1a-mut-MT-hGR was sufficient to restore normal expression of XSMP-30 in dexamethasone-treated XHRT1-MO-injected embryos, indicating that the XHRT1-MO knockdown phenotype is specific(Fig. 8P). We next investigated whether blocking XHRT1 activity could block the effect of hGR/Su(H)/Ank. As shown in Fig. 8Q, we observed that activation of Notch signaling at stage 22-25 has no effect, or in some embryos, increases the expression of XSMP-30. A reduction of XSMP-30 expression was observed in embryos co-injected with the XHRT1-MO. Thus, injection of XHRT1-MO is sufficient to impede the effect of overexpression of hGR/Su(H)/Ank on XSMP-30 further supporting the idea that XHRT1 is an important component of the Notch signaling pathway that leads to glomus and proximal tubule formation.

Our results support the idea that, compared with the other bHLH-O repressors tested, XHRT1 has a specific function in the developing pronephros. To identify the region(s) that are required for its activity, we performed domain-swapping experiments between XHRT1 and one of the bHLH-O repressors inactive in the pronephros. We chose the novel bHLH-O gene XHes2because both genes are expressed in the retina where XHes2 but not XHRT1 promotes Müller glial development(Satow et al., 2001) (M.S.,unpublished), which may provide another assay to identify the regions required for their distinct regulatory functions. As we and others have previously shown that the bHLH and the Orange domains of XHRT1 are required for efficient DNA-binding and homo- and heterodimerization(Taelman et al., 2004), we decided to keep the bHLH-O regions of both proteins intact and swap their C-terminal sequences (Fig. 9A). Both chimeric molecules were linked to glucocorticoid receptor ligand-binding domain sequences to generate inducible constructs. Injected embryos were induced at stage 18 and assayed for the expression of Evi1. All constructs were effectively translated when expressed in embryos(Fig. 9B). As shown in Fig. 9C, only the hybrid containing the XHRT1 C-terminal domain fused to the XHes2 bHLH-O domain reduced xEvi1 expression to a similar extent as the wild-type XHRT1 protein. Thus, the specific properties of XHRT1 in the pronephros appear to be linked to its C-terminal region.

In this work, we analysed the expression and function of several potential downstream effectors of Notch signaling during pronephric development. Our results indicate that XHRT1, compared with the other bHLH-O factors expressed in the pronephros, plays a predominant role in the patterning of the kidney, favoring first glomus and later proximal tubule formation. We also provide evidence that its distinct specific function in the developing kidney is not only due to its earlier temporal expression pattern but also to the intrinsic properties of the protein.

Notch-1, X-Delta-1 and X-Serrate-1 have distinct expression patterns in the developing pronephros, suggesting that spatiotemporal control of Notch activity is an important determinant of the patterning of the early kidney anlage (McLaughlin et al., 2000). In agreement with those observations, our results indicate that five bHLH-O genes, namely XHRT1, Xhairy1 and Xhairy2b (related to murine HES1), esr9 and esr10 (related to HES5)have distinct dynamic restricted expression patterns within the developing pronephros and are regulated by Notch. In most cases, their expression pattern resembles that of their murine orthologs. In mouse, HRT1 mRNA is detected during the earliest stage of development of the nephron. At the comma and S-shaped body stages, HRT1 transcripts are localized to the more proximal regions of the developing nephron that will form the loop of Henle to the developing podocytes (Leimeister et al., 2003; Chen and Al-Awqati,2005). We found that in Xenopus, XHRT1 is expressed early in the developing pronephros region in the pronephric mesoderm that will give rise to the glomus. X-Delta-1 is the only Notch ligand to be expressed at those early stages in the pronephric region. It is expressed in cells of the lateral layer in close contact with XHRT1-expressing cells, suggesting that it may trigger XHRT1 activation in the surrounding cells. xWT1 is another potential regulator of XHRT1, as both genes are activated at about the same time in the forming glomus and its translational inhibition decreased XHRT1expression. Interestingly, we observed that the expression of lateral specific genes is not expanded in the xWT1 morphants, suggesting that xWT1 is not the only factor that controls mediolateral patterning of the pronephros. Later on,expression of XHRT1 in the medial layer is downregulated and strong staining is detected transiently in the dorsoanterior proximal-tubule forming part of the pronephros anlagen itself. Thus, both in mouse and in Xenopus,HRT1 is early and selectively expressed in proximal compartments of the developing nephron, suggesting an early role in glomus and proximal tubule formation. At early tadpole stage, we observed that XHRT1 expression is restricted to the tip of the forming tubules, suggesting another later role in the subcompartmentalization of the tubules.

HES1 in the mouse is expressed in most cells during early stages of kidney development. Its expression becomes more restricted later during development(Leimeister et al., 2003; Chen and Al-Awqati, 2005). Similarly in Xenopus, Xhairy1 and Xhairy2b are initially expressed in both pronephric mesodermal layers, suggesting that they may have early a broader role than XHRT1. In contrast to the other bHLH-O genes, Xhairy2b expression is maintained in late tadpoles. Its expression at that stage resembles that of the Notch ligand X-Serrate-1. As recent investigations have demonstrated that the Notch ligands not only deliver their signal by binding to Notch receptors, but also by playing a functional role in the cells in which they are expressed(Ascano et al., 2003; LaVoie and Selkoe, 2003; Ikeuchi and Sisodia, 2003), it is tempting to speculate that X-Serrate-1 may play a role in Xhairy2bactivation.

Hes5 expression in the mouse is restricted to the middle segment of S-shaped bodies that gives rise to the loop of Henle(Piscione et al., 2004; Chen and Al-Awqati, 2005). Similarly, in Xenopus, esr9 and esr10 expression are detected in the pronephros anlagen of late tailbud embryos in a region slightly more ventral to XHRT1 expression. At that stage, expression of esr9 and esr10 resembles that of the Notch ligands X-Serrate-1 and X-Serrate-2, suggesting that they may be involved in their regulation. Further experiments are needed to analyse the contribution of the distinct Notch ligands to the spatiotemporal regulation of bHLH-O genes in the pronephros.

In the Xenopus pronephros and the mouse metanephroi, Notch activation has been shown to be essential for proximal tubule and glomus formation (McLaughlin et al., 2000; McCright et al., 2001; Wang et al., 2003a; Cheng et al., 2003; Cheng and Kopan, 2005; Van Campenhout et al., 2006). Here, we show that activation of Notch in the pronephric primordium favors first glomus and later proximal tubule fates, which correlates with the temporal expression pattern of XHRT1, esr9 and esr10. To the best of our knowledge, this is the first time that the precise temporal dependence on Notch activation of those two processes has been investigated. We observed that overexpression and the morpholino knockdown of XHRT1, but not that of esr9, esr10 or Xhairy2b, phenocopy the defects observed upon activation and inhibition of Notch signaling. These results indicate that XHRT1 may act to repress distal tubule and duct cell fates in the portion of pronephric mesoderm fated to form the glomus and the proximal tubules. They also suggest that XHRT1 has a role distinct to that of the other bHLH-O repressors in the earliest stage of pronephros development. However, we cannot exclude that these finding may arise from a reduced efficiency of the corresponding MOs. Kidneys from mice where the Hes1 or Hes5 genes were deleted show no defects, whereas compound homozygotes for both Hes1 and Hes5 die before kidney development(Chen and Al-Awqati, 2005). Further evaluation of the targeted disruption of these genes in conditional knockout mice is required to determine their contribution in nephron patterning.

In contrast to Notch activation using an inducible form of an activated Su(H) construct, XHRT1 overexpression does not increase proximal tubule/glomus formation. We also observed that XHRT1 does not restore the expression of proximal tubule markers in embryos where Notch has been inhibited by injection of a Su(H)DBM construct, suggesting that it mediates only part of the effects executed by Notch. xWT1 is another transcriptional repressor that has also been suggested to have a role in the repression of tubule and duct specific genes in the forming glomus(Wallingford et al., 1998; Van Campenhout et al., 2006). Further investigations are required to elucidate the hierarchical relationship that links XHRT1 and xWT1, and to identify other factors that may contribute to glomus/proximal tubule cell fate decisions.

In HRT1 single mutant or in HRT1/HRT2 double mutant mice, no kidney defects have been reported(Fisher et al., 2004; Kokubo et al., 2005). The difference in phenotype between Xenopus and mouse may be due to differential evolution or expression of the HRT genes in both species. Differential evolution of the HRT2 gene has been recently reported in fish (Winkler et al., 2003). In the mouse, HRT1 and HRT3, but not HRT2, have overlapping expression in the developing nephrons(Leimeister et al., 2003). The identification of the functional role of HRT1 in mouse nephrogenesis will require the analysis of the phenotype through nephrogenesis of HRT1/HRT3 double knockout mice.

Swapping experiments between XHRT1 and the related bHLH-O XHes2 that is inactive in the pronephros have shown that the specific properties of XHRT1 are dependent on their divergent C-terminal sequences. Whereas the XHes2 protein has a C-terminal domain of 62 amino acids terminated by a WRPW motif,the XHRT1 C-terminal domain is much longer (133 amino acids) and does not contain the WRPW motif; this is replaced by a related sequence (YRPW) near its C terminus. In E(spl) in Drosophila, this region has also been shown to be important, as mutants that lack the sequences C terminal to the Orange domain act as dominant-negative variants(Giebel and Campos-Ortega,1997). In zebrafish Her4, the Orange domain-WRPW interval is also essential for its ability to block neurogenesis(Takke et al., 1999). At present, the functional role of the bHLH-O C-terminal sequences is unclear. In XHRT1, those sequences are involved, together with the bHLH and Orange domains, in dimerization and selection of the bHLH partners, and they possess intrinsic repression activity (Taelman et al., 2004). In HES1, the C-terminal domain allows interaction with the Runt-related protein CbFa1 (McLarren et al., 2000). Further studies are needed to clarify the role of the C-terminal sequences in XHRT1-specific function.

We thank Makoto Asashima, Andre Brändli, Elizabeth Jones, Jonathan Licht, Chris Kintner, Peter Vize and David Wilkinson for providing us with various constructs and antibodies, and Sadia Kricha for technical assistance. This work was supported by the Belgian program Interuniversitary attraction poles (grant P5/35 to E.J.B.), by the Communauté Française de Belgique (grant number ARC 00/05-250 to E.J.B.), by the Fund for Scientific Medical Research (FRSM 3.4555.01 and 3.4556.01), and by a grant from the DFG in the context of the CMPB to T.P.