Notch activates Wnt-4 signalling to control medio-lateral patterning of the pronephros

Recent research in the field of nephrology has determined the Notch signalling pathway as an important regulator of renal development(Dressler, 2008; Kopan et al., 2007; Liu et al., 2007; Taelman et al., 2006). Aberrant Notch signalling during development or in adulthood is a major cause of renal disease, specifically affecting glomerulus function(Barisoni, 2008; McCright et al., 2001; Mertens et al., 2008; Niranjan et al., 2008). Despite these studies, the mechanism by which the Notch signalling pathway regulates nephrogenesis remains unknown.

The Notch signalling pathway is a paracrine signalling pathway composed of the large transmembrane proteins Notch, Delta and Serrate (or Jagged in mammals and zebrafish). In the core pathway, Notch receptors bind Delta and Serrate ligands on opposing cells, which causes cleavage of the Notch receptor(Brou et al., 2000; Kopan and Goate, 2000),liberating its intracellular domain (Notch-ICD)(Schroeter et al., 1998; Taniguchi et al., 2002). Notch-ICD then translocates to the nucleus where it binds to the CSL transcription factor, switching it from a transcriptional repressor to a transcriptional activator (Jarriault et al., 1995). Consequently, altered gene expression directs the cell towards a specified cell fate. In addition to Notch-ICD canonical signalling,Notch-ICD independent and ligand-dependent signalling have also been reported to be important in the overall signal transduction of this pathway(Ascano et al., 2003; Ikeuchi and Sisodia, 2003; Kolev et al., 2005; LaVoie and Selkoe, 2003).

The pronephros, the embryonic kidney, develops from nephrogenic mesenchyme within the intermediate mesoderm lateral to the anterior somites(Brändli, 1999; Dressler, 2006; Jones, 2005). The pronephros is a paired organ, consisting of four distinct compartments; the glomus,coelomic cavity, nephrostomes and tubules (which are further characterized as proximal, intermediate, distal and connecting tubules), together these components form a non-integrated nephron(Nieuwkoop and Faber, 1994; Reggiani et al., 2007; Saxén, 1987; Vize et al., 1995). The pronephros is a simple organ to study, showing both morphological and physiological similarities to more complex kidney forms, the meso- and metanephros, which make it an ideal model to study with reference to kidney development (Vize et al.,1995).

McLaughlin and co-workers (McLaughlin et al., 2000) were the first to highlight the role of the Notch signalling pathway in Xenopus laevis pronephrogenesis. This study showed that expression of components of the Notch signalling pathway in the pronephros appears in two phases. The early phase is characterized by notch1, serrate1 and delta1 expression in the dorso-anterior region of the pronephric anlagen between stages 21 and 32. This is then followed by late phase expression of notch1 and serrate1more ventrally, in the developing proximal tubules. Knowledge of how the Notch signalling pathway regulates pronephrogenesis was further advanced by Taelman and colleagues (Taelman et al.,2006), who provided experimental evidence that early-phase pronephric Notch signalling promotes glomus formation, whereas late-phase pronephric Notch signalling promotes proximal tubule development.

In this study, we show that early-phase pronephric Notch signalling is required for medio-lateral patterning, a function that we suggest indirectly mediates proximal-distal patterning and permits the glomus and proximal tubules to develop in isolation. We provide evidence highlighting a novel role for Wnt-4 in establishing the glomus and nephrostomes, a function that is probably regulated by the Notch effector gene hrt1 (also known as hey1). Furthermore, we report pronephric expression of X. laevis homologues of the Drosophila fringe gene, lunatic fringe and radical fringe, and propose a role for their gene products in proximal pronephrogenesis.

Whole-mount in situ hybridisation was carried out as described elsewhere(Harland, 1991). The embryos were fixed in MEMFA (0.5 M MOPS, pH 7.4, 100 mM EGTA, 1 mM MgSO₄,4% formaldehyde) and linearized plasmid from Na⁺/K⁺-ATPase (SmaI/T7), slc5a2(EcoRI/T7), odf3 (EcoRI/T7), nephrin (SmaI/T7),Wnt-4 (XhoI/T7), Notch-1 (ClaI/SP6), Delta-1(NdeI/T7), Serrate-1 (HindIII/T7), lunatic fringe(NcoI/T7), radical fringe (BamHI/T7), ClC-K(NotI/T7), GATA-3 (SmaI/T7), Xbrachyury (SacII/T3), MHC(NcoI/SP6), HRT-1 (Not-1/T7) and Lim-1 (XhoI/T7)was used to generate digoxigenin-11-UTP-labelled (Boehringer Mannheim)antisense RNA probes from the polymerases indicated. Probes were visualised using anti-DIG-alkaline phosphatase secondary and NBT/BCIP for the colour reaction according to the manufacturer's recommendations (Boehringer Mannheim).

Embryos were obtained by in vitro fertilisation of hormonally stimulated Xenopus laevis and staged according to Nieuwkoop and Faber(Nieuwkoop and Faber, 1994). Standard embryological procedures and culturing conditions were used as described by Jones and Woodland (Jones and Woodland, 1986).

Capped RNAs were synthesised in vitro from Notch-ICD, Delta^STU,Su(H)^DBM (Chitnis et al.,1995), HRT-1-hGR(Taelman et al., 2006), Wnt-4,Lfng and Rfng using the SP6 Message Machine Kit (Ambion). Typically 600 pg notch-ICD mRNA, 1 ng delta^STU mRNA, 1 ng su(H)^DBM mRNA, 1 ng hrt1 mRNA, 1 ng wnt4mRNA, 500 pg lfng mRNA, 2 ng rfng mRNA and 400 pg beta-galactosidase mRNA was injected. The lfng, rfng, wnt4and hrt1 antisense morpholino oligonucleotides used were: lfng MO1 5′-gcttcttcccccagttcttcagcat-3′; lfng MO2 5′-gcgctaccttctctcctcctctata-3′; rfngMO 5′-tgaggccaacgtaagtgactcttcat-3′; wnt4 MO 5′-gtactctggggtcatcctgctgctg-3′; and hrt1 MO 5′-tagtcgtgtccccgcttcatggctg-3′; (Gene Tools LLC). The control morpholino is a random sequence of the same length. Typically 20 ng of rfng MO, hrt1 MO and control MO, 10 ng of wnt4 MO and 5 ng lfng MO1 or lfng MO2 was injected. Embryos were dejellied and injected with mRNA alone or in combination with MO (as specified in the text) into a V2 blastomere of an eight-cell-stage embryo to target the future pronephros (Dale and Slack,1987; Moody and Kline,1990), under 5% Ficoll in BarthX.

Whole-mount immunohistochemistry was performed using standard methods on MEMFA-fixed embryos. The primary antibodies used were monoclonal antibody 3G8,which detects the proximal tubules, and 4A6, which detects the intermediate and distal tubules (Vize et al.,1995). The secondary antibody was alkaline phosphatase-conjugated goat anti-mouse (Sigma). BCIP/NBT (Boehringer) or Fast Red TR/Napthol AS/MX(Sigma) was used for the colour reaction, according to the manufacturer's recommendations.

Either mRNA alone, or mRNA with a MO, was translated in vitro in the Rabbit Reticulocyte Lysate System (Promega) with ³⁵S-Met according to the manufacturer's protocol. Reactions (25 μl) were denatured at 95°C in 2×SDS loading buffer (Harlow and Lane, 1988) and run on a 10% (w/v) SDS-PAGE resolving gel using a vertical mini-gel apparatus for 2 hours at 100 V. The gel was exposed to Kodak X-ray film overnight at room temperature, before being developed.

In order to mis-activate the Notch signalling pathway we injected notch-ICD mRNA, a truncated form of the Notch receptor that is constitutively active, or a dominant-negative form of the Delta ligand,Delta^STU (Chitnis et al.,1995). Both constructs were targeted to the lateral mesoderm including the future pronephros by injection into a V2 blastomere of embryos at the eight-cell stage, with beta-galactosidase mRNA to act as a lineage tracer. Embryos were cultured to various stages of development and effects on different markers of distinct regions of the pronephros were observed. Pronephric phenotypes were determined by comparing the injected and uninjected sides. All data presented here that produced pronephric phenotypes were analysed statistically, using the Chi-squared test, and were significantly different to embryos injected with control mRNA within 95%confidence limits. Experiments were repeated several times, yielding similar data.

Overexpression of notch-ICD mRNA in embryos harvested at stage 41 and whole-mount antibody stained with 3G8, which detects nephrostomes and proximal pronephric tubules, and 4A6, which detects intermediate and distal pronephric tubules (Vize et al.,1995), showed an increase in size of the proximal region of the pronephros at the expense of the distal pronephric region. A total of 76% of embryos had ectopic 3G8 staining, whereas 4A6 staining was reduced in 97% of embryos (n=34; Fig. 1A). Injection of delta^STU mRNA inhibited proximal pronephros formation but had no effect on distal pronephros development. 33% of embryos had reduced 3G8 staining, with only 5% having reduced 4A6 staining (n=36; Fig. 1B), suggesting that the Notch signalling pathway has no role in distal pronephros cell fate determination.

To investigate this phenotype in more detail, we injected notch-ICD mRNA and delta^STU mRNA and cultured embryos to stage 34. In situ hybridisations for two markers of pronephric tubules; Na⁺/K⁺-ATPase, an ion transporter expressed in the proximal, intermediate and distal tubules, and slc5a2, a solute carrier expressed in the proximal tubules(Raciti et al., 2008; Reggiani et al., 2007), was then performed. Injection of notch-ICD mRNA reduced Na⁺K⁺ATPase (also known as atp1a1)expression in 96% of embryos (n=24; Fig. 1C), and similarly reduced slc5a2 expression in 95% of embryos analysed (n=38; Fig. 1E), thus tubulogenesis was completely inhibited. Injection of delta^STU mRNA reduced slc5a2 expression in 76% of embryos (n=34; Fig. 1F), but had a localised effect on Na⁺K⁺ATPase expression. In 23% of embryos Na⁺K⁺ATPase expression was completely reduced, however the most prominent phenotype, observed in 47% of embryos, was a reduction in only the proximal tubule region of the Na⁺K⁺ATPase expression pattern (n=43; Fig. 1D).

Since injection of notch-ICD mRNA inhibited proximal tubule formation (slc5a2 reduction), but caused ectopic 3G8 staining, we concluded that this 3G8 staining must be completely nephrostomal, because 3G8 only stains nephrostome and proximal tubule domains. To confirm this, embryos were injected with notch-ICD mRNA and delta^STUmRNA as previously described, cultured to stage 34 and in situ hybridised for expression of outer dense fibre 3 (odf3), a marker of multi-ciliated cells in the epidermis and nephrostomes. Overexpression of delta^STU inhibited nephrostomal odf3 expression in 89% of embryos (n=28; Fig. 1H). However, as expected, we observed ectopic nephrostomal odf3 expression in 74% of embryos injected with notch-ICDmRNA (n=27; Fig. 1G). As the percentage of embryos showing an increase in odf3 expression was equivalent to those with ectopic 3G8 staining, we conclude that notch-ICD overexpression resulted in excess nephrostomal tissue.

McLaughlin and co-workers (McLaughlin et al., 2000) and Taelman and colleagues(Taelman et al., 2006) showed that early misactivation of the Notch signalling pathway induced ectopic glomus formation, a result we have confirmed; 81% of embryos injected with notch-ICD mRNA had increased staining of the glomus marker nephrin, as observed by in situ hybridisation (n=27)(Fig. 1I). delta^STU mRNA inhibited nephrin expression in 88%of embryos (n=34; Fig. 1J). In conclusion, early phase Notch signalling inhibits tubulogenesis and promotes glomus and nephrostome formation.

Delta^STU is presumed to act as an anti-morph to inhibit Notch signalling, but its exact mechanism of action is unknown. To confirm that the phenotypes observed after delta^STU overexpression resulted from inhibited Notch-ICD signalling we targeted a dominant-negative Suppressor of Hairless construct, Su(H)^DBM, to the pronephros. Overexpressing su(H)^DBM produced exactly the same pronephric phenotypes as delta^STU overexpression (see Fig. S1 in the supplementary material), confirming that the pronephric phenotypes were a consequence of reduced Notch-ICD signalling. To further control for the specificity of these results, we have shown identical phenotypes using hormone inducible Notch constructs (see Fig. S2 in the supplementary material), which are consistent with previous publications(McLaughlin et al., 2000; Taelman et al., 2006).

During development, the Notch and Wnt signalling pathways frequently, if not always, act in an integrated manner to promote specific cell fates(Hayward et al., 2008). The Wnt gene wnt4, is expressed in the proximal tubules and nephrostomes during stages of development when the Notch signalling pathway is inducing nephrostome and glomus cell fate decisions(Saulnier et al., 2002). We hypothesised that these two pathways could be integrated and thus aid cell fate decisions in the dorso-anterior region of the pronephric anlagen. In order to investigate this possibility, wnt4 mRNA and beta-galactosidase mRNA were co-injected as previously described, and embryos were cultured to various stages of development to determine the effect of wnt4 overexpression on different markers of distinct regions of the pronephros.

wnt4 mRNA induced ectopic 3G8 staining in 69% of embryos, with 4A6 staining reduced in 77% of embryos (n=26; Fig. 2A). Expression of the tubule markers Na⁺K⁺ATPase and slc5a2was reduced in 83% (n=23; Fig. 2B) and 92% (n=36; Fig. 2C) of embryos,respectively. wnt4 mRNA induced ectopic nephrostomal odf3expression in 76% of embryos (n=34; Fig. 2D) and increased the domain of nephrin expression in 75% of embryos (n=32; Fig. 2E). Thus, wnt4overexpression induced ectopic nephrostome and glomus formation at the expense of tubules, an identical phenotype to notch-ICD overexpression.

To observe whether Notch signalling affected wnt4 expression in the pronephros, embryos were injected with notch-ICD mRNA or delta^STU mRNA into the V2 blastomere, cultured to stage 28, and analysed for wnt4 expression by in situ hybridisation. Of embryos injected with notch-ICD mRNA, 64% had ectopic wnt4expression (n=69; Fig. 3A). Injection of delta^STU mRNA inhibited wnt4 expression on the injected side in 70% of embryos(n=71; Fig. 3B). In conclusion, activation of the Notch-ICD dependent pathway induced wnt4 expression. Conversely, suppression of Notch signalling inhibited wnt4 expression.

To establish whether pronephric wnt4 overexpression could upregulate components of the Notch signalling pathway, wnt4 mRNA was injected into the V2 blastomere and embryos were cultured to stage 28 where whole mount in situ hybridisation for expression of notch1, delta1and serrate1 was performed (Fig. 3C,D). wnt4 overexpression had no significant effect on notch1 (data not shown), delta1 or serrate1expression levels. delta1 expression was reduced in only 6% of embryos (n=32; Fig. 3C). Expression of serrate1 was reduced in 14% of embryos(n=36; Fig. 3D). However, we did observe a change in the anterior-posterior register of gene expression of both delta1 (84%) and serrate1 (61%). Frequently, elements of the expression pattern were more distal (see Fig. 3B, marked with bars), and different from delta1 or serrate1 expression on the uninjected side. In summary, wnt4 overexpression alters the expression pattern of delta1 and serrate1, but does not seemingly inhibit or upregulate the expression levels of these genes.

Given that notch-ICD overexpression induced wnt4expression, but wnt4 overexpression had little effect on delta1 and serrate1 expression levels, we hypothesised that Wnt-4 could be acting downstream of the Notch signalling pathway. To investigate this possibility we simultaneously overexpressed notch-ICD and depleted endogenous wnt4 translation using a morpholino oligonucleotide (MO) (Saulnier et al., 2002). Embryos injected with notch-ICD mRNA and wnt4 MO were cultured to stage 32, where they were analysed for nephrin expression.

Single injections of notch-ICD mRNA induced ectopic expression of nephrin in 94% of embryos (n=47, data not shown). 74% of embryos injected with the wnt4 MO had reduced nephrinexpression (n=27; Fig. 4A). Co-injection of notch-ICD mRNA and the wnt4MO reduced expression of nephrin in 50% of embryos, with no embryos showing ectopic nephrin expression (n=28; Fig. 4B). We also repeated this experiment to determine the effect on 3G8 and 4A6 staining, and observed a 71%reduction in both 3G8 and 4A6 staining when notch-ICD mRNA and the wnt4 MO are co-injected (n=24, data not shown). Again, none of these embryos had ectopic 3G8 staining. Consequently, ectopic proximal pronephrogenesis normally caused by overexpression of notch-ICD mRNA was completely prevented by the wnt4 MO. In conclusion, this result suggests that Wnt-4 acts downstream of the Notch signalling pathway in the proximal pronephros to influence medio-lateral patterning.

We next attempted the reciprocal experiment; overexpressing wnt4with concomitant inhibition of Notch-ICD signalling. Co-injection of wnt4 mRNA with either su(H)^DBM mRNA (57%increased, n=21; Fig. 4C) or delta^STU mRNA (67% increased, n=36; see Fig. S3A in the supplementary material) caused ectopic nephrin expression, again suggesting Wnt-4 functions downstream of Notch signalling. We also observed the effects these co-injections had on proximal tubulogenesis. Co-injections of wnt4 mRNA and su(H)^DBM mRNA caused ectopic slc5a2 expression in 43% of embryos (21% of embryos had reduced slc5a2 expression; n=28; Fig. 4D). This result is surprising because single injections of either wnt4 mRNA or su(H)^DBM mRNA reduced slc5a2 expression(Fig. 2A; see Fig. S2C in the supplementary material). Similarly, co-injection of wnt4 mRNA and delta^STU caused ectopic proximal tubulogenesis in 15% of embryos, with only 8% of embryos having reduced slc5a2 expression(n=26; see Fig. S3B in the supplementary material).

The Hairy-related transcription factor gene hrt1, encodes a downstream mediator of Notch signalling that has been shown to be responsive to Notch signalling in numerous tissues(Pichon et al., 2002; Rones et al., 2002). HRT-1 has been shown to regulate glomus formation and patterning of the pronephros anlagen (Taelman et al.,2006). We aimed to determine whether hrt1 overexpression and MO knockdown affected wnt4 expression in the pronephros in order to position the activity of HRT-1 upstream or downstream of Wnt-4 during proximal pronephrogenesis.

We targeted mRNA encoding a hormone-inducible hrt1-hGR construct to the future pronephros as described previously. Embryos were left to develop to stage 18, where half of the injected embryos were switched into a dexamethasone containing culture medium to activate the message. Embryos were fixed at stage 28, stained for lineage label and analysed for wnt4expression. Embryos injected with hrt1-hGR mRNA alone displayed no phenotype (3% reduced, n=38; Fig. 5A). Of embryos incubated in dexamethasone, 29% displayed ectopic wnt4 expression (Fig. 5B), 10% had reduced wnt4 expression, and the remainder had no phenotype (n=41). We also knocked down hrt1translation using a MO previously shown to specifically deplete endogenous HRT-1 (Taelman et al., 2006). Injection of this hrt1 MO to the pronephros caused a reduction in wnt4 expression in 69% of embryos scored (n=26; Fig. 5C).

To place HRT-1 in the genetic hierarchy, we examined the effect of notch-ICD and wnt4 overexpression on hrt1expression. Mis-activation of the Notch signalling pathway induces hrt1 expression (Taelman et al.,2006). We reproduced this phenotype; 80% of embryos injected with notch-ICD mRNA had ectopic hrt1 expression (n=70; Fig. 5D). However, following wnt4 overexpression, no embryos showed regions of ectopic hrt1 expression. A total of 92% of embryos showed a slight phenotype,of which 73% had disrupted hrt1 expression, with the remaining embryos having reduced hrt1 expression (n=28; Fig. 5E). Thus, overexpression of wnt4 did not significantly affect hrt1 expression. We suggest the disrupted pronephric expression profile of hrt1 after wnt4 overexpression is a product of the overall effect Wnt-4 has on pronephrogenesis (as observed in Fig. 2A). Thus HRT-1 acts upstream of Wnt-4 and is required for wnt4 expression.

Overexpression of notch-ICD caused highly significant pronephric phenotypes, increasing the size of the glomus and nephrostomes at the expense of the proximal tubules. To establish the exact spatial relationships between these two differentiated tissues and other internal structures, the injected embryos were analysed in section.

Embryos injected with notch-ICD mRNA and stained for nephrin expression by in situ hybridisation at stage 32, marking the medial pronephric region, were sectioned(Fig. 6A). In these embryos,the medial pronephric mesoderm extended into the normal region of lateral pronephric mesoderm after notch-ICD mRNA injection. The only other gene whose expression is increased after overexpression of notch-ICDis odf3, a marker specific for the nephrostomes in the pronephros. Despite their position on the lateral side of the coelom, nephrostomes form from medial pronephric mesoderm (Howland,1916). Thus it is perhaps unsurprising ectopic glomus formation is accompanied by ectopic nephrostome formation. In conclusion, we suggest that lateral pronephric mesoderm does not form in embryos overexpressing notch-ICD, the entire anlagen switches to medial pronephric fates.

Embryos, injected with notch-ICD mRNA and co-injected with gfp mRNA lineage label, were sorted for left- or right-injected sides at stage 26. At stage 41, embryos were fixed, wax-embedded, sectioned and analysed histologically with haematoxylin and eosin. Sections through the proximal pronephric region showed an abnormal mass of cells without tubular structure on the injected side (Fig. 6B). Normal pronephric development of the proximal pronephric region could be clearly seen on the uninjected side of this embryo. The glomus, indicated by dispersed podocytes, the coelomic cavity, and the proximal tubules all formed in their correct positions. In sections through the distal pronephric region ectopic glomus and nephrostomal tubule formation was clearly visible on the injected side(Fig. 6C), whereas a single distal tubule was visible on the uninjected side.

Fringe proteins are glycosyl transferases that extend carbohydrate chains on the EGF repeats of the extracellular domain of the Notch receptor. In Drosophila, such post-translational modifications promote Notch interactions with its Delta ligand(Bruckner et al., 2000; Moloney et al., 2000). Previous studies have highlighted roles for downstream Notch effector genes during proximal pronephrogenesis (Rones et al., 2002; Taelman et al.,2006), yet an understanding for how Notch signalling is regulated,and how this regulation affects pronephros development, has not been investigated. If X. laevis homologues of the Drosophila fringe gene, lunatic fringe (lfng) and radical fringe (rfng), were temporally and spatially expressed appropriately in the pronephros, this finding would further understanding of the mechanism by which Notch signalling is regulated in the proximal pronephros.

We performed whole-mount in situ hybridisations on uninjected embryos using 3′UTR-specific DIG-labelled antisense RNA probes to detect the temporal and spatial expression of lfng and rfng in X. laevis (Fig. 7). A previous study investigated the expression pattern of lfng in X. laevis (Wu et al., 1996);however, pronephric expression was not detected. Using an improved in situ hybridisation protocol, we observed pronephric expression of both lfng and rfng. lfng is expressed in the dorso-anterior region of the proximal pronephros between stages 22 and 32. Pronephric rfng expression was detected in the same region between stages 26 and 32. The temporal and spatial expression of lfng and rfngtherefore suggests they might have a role in regulation of the Notch signalling pathway because notch1, serrate1, serrate2, delta1 and wnt4 are all expressed in this pronephric region at a similar time during development (McLaughlin et al.,2000; Rones et al.,2002; Saulnier et al.,2002; Taelman et al.,2006).

In order to observe whether mis-expressing lfng and rfngcaused pronephric phenotypes, morpholino oligonucleotides (MOs) were designed to specifically knockdown translation of endogenous lfng and rfng transcripts (for MO sequences see Fig. S4A-C in the supplementary material), and overexpression experiments were performed. In vitro translation of lfng and rfng mRNA, in a rabbit reticulocyte lysate system using ³⁵S-methionine, could be knocked down by each specific MO (see Fig. S4D-F in the supplementary material).

We observed gross developmental defects upon mis-expression of lfng (see Figs S5-S7 in the supplementary material). Consequently, we were unable to perform functional analysis of Lfng in the pronephros. Furthermore, targeted injection of the rfng MO to knockdown endogenous rfng expression had no statistically significant effect on development of any region of the pronephros. Na⁺K⁺ATPase, slc5a2 and nephrinexpression were normal in these embryos (data not shown). Antibody staining of the mature pronephros with 3G8 and 4A6 caused a statistically insignificant 6%decrease in 3G8 staining and 2% decrease in 4A6 staining (n=47; see Fig. S5F in the supplementary material). Normal levels of rfng are therefore not required for normal development of the pronephros during embryogenesis.

By contrast, overexpression of rfng had a dramatic effect on pronephros development: rfng mRNA overexpression caused ectopic expression of the early pronephric marker lim1 in 94% of embryos at stage 26 (n=17; Fig. 8A). Expression of hrt1 was also increased by rfng overexpression in 77% of embryos (n=35; Fig. 8B). Unlike injection of notch-ICD mRNA, rfng overexpression caused ectopic 4A6 staining (63%, n=41), in addition to ectopic 3G8 staining (93%, n=41; Fig. 8C). Frequently, the proximal pronephric domain appeared to be completely duplicated in more distal regions (white arrow in Fig. 8C). Expression of tubule markers Na⁺K⁺ATPase (n=23; Fig. 8D) and slc5a2(n=22; Fig. 8E) was increased in 67% and 59% of embryos, respectively. Ectopic formation of the glomus was observed in 81% of embryos examined by in situ hybridisation for nephrin expression (n=21; Fig. 8F). Ectopic nephrostome formation was also observed, with the domain of odf3 expression increased in 62% of embryos (n=21, data not shown). In situ hybridisation detecting expression of gata3, a marker of the distal tubules, was also observed. gata3 expression was reduced in 84% of embryos (n=112; Fig. 8G). In conclusion, Rfng promotes formation of the proximal pronephros and suppresses distal pronephrogenesis.

In vertebrates, Fringe proteins promote Notch-ligand interactions(Haines and Irvine, 2003; Hicks et al., 2000; Tsukumo et al., 2006). We have shown that Wnt-4 acts downstream of Notch signalling (see above), thus we wished to confirm that Rfng also acted upstream of Wnt-4. Co-injection of rfng mRNA with the wnt4 MO reduced nephrinexpression in 65% of embryos (n=23; Fig. 8H) and also inhibited slc5a2 expression in 82% of embryos (n=22; Fig. 8I). This result indicates Rfng acts upstream of Wnt-4, most likely by mediating Notch-ligand interactions in the proximal pronephros.

Fringe proteins are enzymes whose activity is dependent on available substrate (Haines and Irvine,2003). To understand the mechanism by which Rfng induces ectopic pronephrogenesis, we injected rfng mRNA as previously described and performed in situ hybridisation for expression of wnt4, delta1 and serrate1 at stage 32 (Fig. 9). Injection of rfng mRNA caused ectopic wnt4expression in 73% of embryos (n=62; Fig. 8A). Expression of delta1 and serrate1 was ectopic in 84% and 80% of embryos,respectively, when rfng was overexpressed (n=51, Fig. 9B; n=65, Fig. 9C). We suggest that rfng overexpression caused ectopic pronephrogenesis by promoting Notch signalling, as has been described in other systems(Hicks et al., 2000; Tsukumo et al., 2006). Promotion of pronephric Notch signalling activates wnt4 expression,the protein product can be secreted in more distal positions than those found normally, thus producing ectopic proximal pronephros cell fates in distal regions.

We have established three novel findings that extend our knowledge of the mechanism by which the Notch signalling pathway regulates proximal pronephrogenesis in X. laevis. We show that early-phase Notch signalling patterns the medio-lateral axis of the dorso-anterior region of the pronephros anlagen. This activity is mediated by regulation of a downstream Wnt-4 function. We also identify lfng and rfng expression in this pronephric region, which most likely act as regulators of Notch signalling in the proximal pronephros. These findings further the mechanistic understanding for how the Notch signalling pathway regulates proximal pronephrogenesis.

Early mis-activation of the Notch signalling pathway in the pronephros, by injection of notch-ICD mRNA, caused increased nephrostomal and glomus formation but inhibited tubulogenesis completely. This novel identification of a nephrostomal phenotype has not been observed previously because pronephric markers used to indicate the effects of mis-expressing components of the Notch signalling pathway failed to distinguish between nephrostomes and proximal tubules. McLaughlin and co-workers(McLaughlin et al., 2000)reasoned that the increase in 3G8 staining and lim1 and pax2expression they observed after mis-activation of the Notch signalling pathway was due to an increase in proximal pronephric development (glomus,nephrostomes and proximal tubules). However, complete inhibition of tubulogenesis, indicated by loss of Na⁺K⁺ATPaseexpression, suggested that the ectopic 3G8 staining observed after notch-ICD overexpression was completely nephrostomal. This theory was confirmed by notch-ICD mRNA injection inducing ectopic odf3expression. We therefore extend the results currently in the literature to show mis-activation of early phase pronephric Notch signalling induces ectopic glomus and nephrostome development, at the expense of tubules. Indeed,mis-activation of early phase pronephric Notch signalling caused the entire anlagen to acquire a medial pronephric fate.

Overexpression of wnt4 highlighted a novel role for the Wnt signalling pathway in pronephric anlagen patterning. Consistent with this, wnt4 is expressed in the dorso-anterior pronephros during early tail-bud stages of development (Saulnier et al., 2002). We overexpressed wnt4 and showed that it caused identical phenotypes to notch-ICD overexpression: ectopic nephrostome and glomus formation, accompanied with inhibited tubulogenesis. Intriguingly, Saulnier and colleagues showed that overexpression of wnt4 disrupted nephrostome formation and proximal tubulogenesis without affecting more distal tubule development(Saulnier et al., 2002). These findings are similar to our own, although they failed to observe ectopic nephrostome formation or inhibition of distal tubule development. We suggest that this difference is due to the different amounts of exogenous message injected in these two experimental series. We injected 1 ng of wnt4mRNA compared with 0.25 ng wnt4 mRNA injected by the other researchers (Saulnier et al.,2002); hence, the phenotypes they observed were less pronounced. They also knocked down translation of endogenous wnt4 transcripts using a MO and showed this inhibited proximal tubule development, but left distal tubulogenesis largely unaffected(Saulnier et al., 2002). We have confirmed this result and extended these observations to show a reduction in glomus size. Thus, wnt4 MO knockdown produces an identical phenotype to delta^STU and su(H)^DBMoverexpression. The Wnt signalling pathway is therefore required for proximal pronephrogenesis and the correlation between the phenotypes observed following overexpression of both wnt4 and notch-ICD, and inhibition of these pathways using delta^STU mRNA and wnt4 MO injection, suggest these two pathways might be integrated.

We aimed to see whether the Notch and Wnt signalling pathways were integrated by observing the effect mis-activation of either pathway had on expression of components of the other pathway. We found that injection of notch-ICD mRNA caused ectopic wnt4 expression on the injected side. Suppression of the Notch signalling pathway by injection of delta^STU mRNA inhibited wnt4 expression on the injected side, a phenotype that correlates well with the inhibitory effects caused by overexpressing delta^STU on proximal pronephrogenesis.

Overexpression of wnt4 did not have an obvious effect on overall levels of notch1, delta1 or serrate1 expression, indicating that Wnt-4 does not either induce or suppress expression of components of the Notch signalling pathway. This result is informative because it suggests there is no positive-feedback loop between the Notch and Wnt signalling pathways in the proximal pronephros and the Notch-ICD-dependent pathway is required upstream for wnt4 expression. We have provided evidence that Wnt-4 acts downstream of pronephric Notch signalling since co-injection of a wnt4 MO with notch-ICD mRNA prevented the ectopic glomus and nephrostome phenotype observed when notch-ICD overexpression is performed alone. Additionally, overexpression of wnt4 and concomitant suppression of Notch signalling still caused ectopic glomus formation. This result suggests that Wnt-4 acts downstream of Notch signalling, and also illustrates that the major role of Notch signalling in the proximal pronephros is to regulate Wnt-4 function, therefore suggesting that Wnt-4 is the integral protein that patterns the proximal pronephric anlagen.

We also investigated the role of HRT-1 in regulation of wnt4expression. HRT-1 has previously been shown to regulate glomus formation and patterning of the proximal pronephros(Taelman et al., 2006). We show that HRT-1 is capable of inducing wnt4 expression, and MO depletion of HRT-1 prevented wnt4 expression, although we do not know whether these effects are either direct or indirect. By contrast, we observed a very slight disruption of hrt1 expression following wnt4overexpression, which we believe is due to the overall effect of wnt4overexpression on pronephrogenesis. Thus, we conclude that Notch-ligand interactions promote hrt1 expression, which then promotes wnt4 expression, and Wnt-4 functions to pattern the proximal pronephros.

lfng and rfng are expressed in the dorso-anterior pronephros, suggesting that Fringe proteins are active in regulation of the Notch signalling pathway in the pronephros. Interestingly, lfng, rfngand manic fringe are expressed in the proximal compartment of the mammalian nephron, although their functional role is unknown(Leimeister et al., 2003). Misexpression of lfng caused gross developmental defects, thus we were unable to confirm a role for Lfng in pronephrogenesis. However, rfng mRNA was able to rescue the effects of lfng depletion(see Fig. S8 in the supplementary material), supporting previous studies that suggest it has functional homology with Lfng(Rampal et al., 2005). Overexpression of rfng increased only the size of the proximal pronephros region at the expense of the distal pronephros, a result that correlates with the predicted proximal pronephric phenotypes we expected to observe given the expression profile of this gene in the pronephros. rfng overexpression also caused ectopic expression of delta1,serrate1 and wnt4, suggesting that Fringe proteins mediate Notch signalling in this pronephric region, which then regulates wnt4expression and ultimately proximal pronephrogenesis.

We initially expected an asymmetrical Notch-ligand interaction mechanism for Notch-mediated boundary formation - as has been described in the Drosophila imaginal wing disc(Johnston et al., 1997) - to be conserved in the pronephros anlagen. This hypothesis was strengthened by our novel observations that early phase pronephric Notch signalling regulates the medio-lateral patterning of the pronephros anlagen and that both lfng and rfng are expressed in this dorso-anterior pronephric compartment. However, discrete spatial expression of Delta and Serrate genes across the medial-lateral axis of the proximal pronephros anlagen is not observed. Previous studies have outlined expression profiles for wnt4 (Saulnier et al.,2002), delta1, serrate1, notch1(McLaughlin et al., 2000), hrt1 and serrate2(Taelman et al., 2006). All these genes are expressed in the dorso-anterior region of the pronephros anlagen during the stages of lateral and medial pronephric mesoderm separation. Surprisingly, all these genes (apart from hrt1, which is initially expressed in both the lateral and medial pronephric mesoderms,before localising to the lateral pronephric mesoderm around stage 25) are expressed on the lateral side of the dorso-anterior pronephric mesoderm. None of these genes is expressed in the medial pronephric mesoderm where the future glomus develops. This lack of medial expression of genes involved in Notch signalling suggests that differential Notch-ligand interactions do not mediate separation of the lateral and medial pronephric mesoderms in X. laevis. Therefore, medio-lateral patterning of the dorso-anterior pronephros must occur by a novel Notch-mediated mechanism.

Taelman and co-workers (Taelman et al.,2006) showed that early phase pronephric Notch signalling promoted glomus development and later proximal tubulogenesis. Overexpression of rfng seemingly overcame these temporal effects and promoted formation of the entire proximal pronephros. We suggest one explanation for this phenotype could be the new-found complexity of the Notch signalling pathway(Bray, 2006; Fiuza and Arias, 2007; Kadesch, 2004). Misactivation of Notch signalling by overexpression of notch-ICD will only affect the Notch-ICD-dependent pathway; theoretically all cells in the anlagen will undergo gene expression activated by the CSL transcription complex. However,Notch-ICD-independent pathways are known to exist and are being realised as an important part of the overall signal transmitted by Notch-ligand interactions(Brennan and Gardner, 2002; Bush et al., 2001; Karsan, 2008). Unlike injection of notch-ICD mRNA, rfng overexpression would probably have a more global effect on Notch signalling because vertebrate Fringe proteins promote Notch-ligand interactions rather than one specific aspect of Notch signal transduction (Hicks et al., 2000; Tsukumo et al.,2006). Thus, we suggest that Notch-ICD-independent pathways have a role in development of the proximal pronephros. In addition, upon inhibition of Notch-ICD dependent signalling by overexpression of su(H)^DBM, we have shown that Wnt-4 was able to promote proximal tubulogenesis. This surprising result is important because it indicates that Wnt-4 is capable of promoting formation of all proximal pronephric domains. Since we have shown that Wnt-4 is the integral gene regulated by Notch signalling that controls proximal pronephrogenesis, it is likely that the temporal effects of Notch signalling identified by Taelman and colleagues (Taelman et al.,2006) mediate Wnt-4 function such that Wnt-4 promotes glomus formation early, and proximal tubulogenesis later.

Collating the data we have presented here with published results, we propose the following model for Notch-mediated pronephric patterning(Fig. 9D). Notch signalling begins in a pool of cells on the lateral side of the dorso-anterior pronephros anlagen around stage 22, with Notch-ligand interactions promoted by fringe proteins. The Notch-ICD signal induces expression of wnt4, which is then secreted by these cells, perhaps acting as a morphogen, generating glomus cell fates in the medial pronephros. This signal is maintained throughout tail-bud stages of development until the lateral and medial pronephric mesoderms physically separate, around stage 28, after which medio-lateral patterning is not required and proximal pronephric Notch signalling then promotes proximal tubulogenesis.

In summary, the mechanism for Notch-mediated boundary formation observed in the Drosophila imaginal wing disc does not separate the medial and lateral pronephric mesoderms in X. laevis. Instead, we propose a pool of cells, under the control of the Notch signalling pathway regulates medio-lateral patterning in a temporal fashion. We suggest that future experimentation should focus on the function of Fringe proteins, whether Wnt-4 acts temporally, and the effect Notch-ICD-independent pathways could have on renal patterning. Given the apparent conservation of mechanisms associated with renal development, such investigations are likely to uncover novel aspects of nephrogenesis that could explain the causes of nephropathy associated with aberrant Notch signalling in humans.