Dlgh1 (discs large homolog 1) is a mammalian homolog of the Drosophila tumor suppressor Discs large 1, and is a member of the membrane-associated guanylate kinase (MAGUK) scaffolding proteins that contain three PSD-95/Dlg/ZO-1 (PDZ) domains. Discs large 1 is involved in epithelial polarization and cell-cell adhesion complex formation during Drosophila development. However, the functions of Dlgh1 during mammalian development remain to be elucidated. We generated Dlgh1-knockout mice and found that homozygous Dlgh1-knockout mice developed various abnormalities in their renal and urogenital organs. The kidneys and ureters were hypoplastic and the lower ends of the ureters were ectopic. In addition, the vagina and seminal vesicle, which are derived from the lower part of the Müllerian and Wolffian duct, respectively, were absent. Unexpectedly, loss of Dlgh1 function in the developing ureters did not disrupt cell-cell junctional complexes, but did impair cellular proliferation in the epithelium. These results suggest a novel role for Dlgh1 in regulating epithelial duct formation and morphogenesis during mammalian development. Although congenital absence of the vagina associated with other variable Müllerian duct abnormalities has been reported in humans, its mechanism has not yet been clarified. Our findings might contribute to a better understanding of such abnormalities.

Dlgh1 (discs large homolog 1, also known as Dlg1 and SAP97) has been identified as a mammalian homolog of the Drosophila tumor suppressor Discs large 1, and is a member of the membrane-associated guanylate kinase(MAGUK) proteins. Discs large 1 is important for cellular polarity establishment, cell-cell adhesion integrity and regulation of cellular proliferation in Drosophila, as these functions are all disrupted in discs large 1 mutant animals(Woods and Bryant, 1991). Dlgh1 is expressed in mammalian epithelial and neuronal cells, and has been reported to be involved in cell-cell adhesion of intestinal epithelial cells(Laprise et al., 2004), cell cycle inhibition in cervical cancer and cultured cells(Ishidate et al., 2000; Watson et al., 2002), cellular polarization in migrating astrocytes(Etienne-Manneville et al.,2005), and localization and clustering of glutamate receptors at the synaptic membrane (Leonard et al.,1998). Furthermore, Naim et al. showed that Dlgh1 is involved in nephrogenesis (Naim et al.,2005).

Dlgh1 is considered to function as a scaffold protein and is known to interact with various proteins. The L27 domain in the N-terminal region of Dlgh1 interacts with calmodulin-associated serine/threonine kinase (CASK),another member of the MAGUK proteins, and this interaction recruits Dlgh1 to regions of cell-cell contact (Lee et al.,2002). Furthermore, Dlgh1 contains three PDZ domains, an SH3 domain and a GK-like domain, which are involved in protein-protein interactions.

During ontogeny, epithelial tissues with functional characteristics appropriate for their developmental stage and environmental circumstances are organized. In this process, the cells acquire two important epithelial features: cell-cell adhesion and apicobasal polarity. In addition, epithelial tissues change their shape during early developmental stages in order to form organs. During these processes, many functional proteins must be localized at precisely regulated subcellular locations, and scaffolding proteins appear to play important roles in these protein localizations. Since many binding partners of Dlgh1 have been identified and as Dlgh1 is widely distributed in the body (Caruana and Bernstein,2001), we hypothesized that Dlgh1 might be involved in the development of epithelial tissues.

The roles of Dlgh1 in mammalian development have previously been analyzed using gene-targeted mice lacking the C-terminal half of the Dlgh1 protein(Caruana and Bernstein, 2001). In these mice, a lacZ gene is inserted into Dlgh1, such that the gene product contains the N-terminal three PDZ domains, but lacks the SH3,protein 4.1 and GK-like domains. The homozygous mice exhibit growth retardation in utero, hypoplasia of the premaxilla and mandible and a cleft secondary palate, and die perinatally. In addition, the nephron number is decreased by 30% (Naim et al.,2005). In that study, however, the functions of Dlgh1 that are attributed to intermolecular interactions via its PDZ domains could not be analyzed in the gene-targeted mice. Therefore, the physiological significance of Dlgh1 functions in epithelial cells, which have been analyzed in invertebrate and vertebrate cultured cells, has not yet been confirmed.

We generated mutant mice null for the Dlgh1 gene. In these mice, a neo cassette sequence is inserted into exon 3 of Dlgh1, such that the protein product does not contain any PDZ domains, even if it is translated. In the present study, we analyzed the phenotypes of these targeted mice to elucidate the functions of full-length Dlgh1 protein in developing epithelial tissues. We focused on the development of the urogenital tracts, in which the correct organization of epithelial tissues is of particular importance for morphogenesis. In normal development of the urinary and genital tracts, a ureteric bud sprouts from the Wolffian duct and then penetrates the metanephric mesenchyme. After such penetration, reciprocal interactions between the ureteric bud tips and the metanephric mesenchyme induce a series of ureteric branches that form the collecting duct system, and differentiation of the condensed mesenchymal cells into epithelial cells that compose the urinary tubules. The lower part of the ureteric bud elongates out of the metanephric mesenchyme and becomes the ureter. Beside the Wolffian duct, the Müllerian duct (also called the paramesonephric duct) is formed by invagination of the coelomic epithelium. The Wolffian and Müllerian ducts differentiate into male and female reproductive tracts, respectively. Developmental analyses of the urogenital organs in homozygous(Dlgh1-/-) mice revealed structural abnormalities of the kidneys and urogenital tracts derived from the Wolffian and Müllerian ducts, indicating the involvement of Dlgh1 in the morphogenesis of these organs. In order to identify the factors responsible for the abnormal development of these organs, we compared the proliferative indexes and localizations of cell-cell junctional proteins in the ureteric epithelium of Dlgh1+/+ and Dlgh1-/- mice. Contrary to previous hypotheses, the results revealed that epithelial cell proliferation was decreased in Dlgh1-/- mice, and that the subcellular distribution of the junctional proteins was not affected by the loss of Dlgh1 protein. These results provide novel insights into the functions of Dlgh1 and suggest that it plays crucial roles in the development of urogenital organs.


An anti-Dlgh1 antibody raised against the N-terminal 100 amino acids was described previously (Matsumine et al.,1996). The other primary antibodies used in this study were as follows: anti-glyceraldehyde 3-phosphate dehydrogenase (G3PDH) antibody(MAB374; Chemicon), anti-pancytokeratin antibody (C2562; Sigma), anti-Pax2 antibody (71-600; Zymed), anti-E-cadherin antibody (C20820; BD Transduction),anti-ZO-1 antibody (61-7300; Zymed) and anti-occludin antibody (71-1500;Zymed). All Alexa Fluor-conjugated secondary antibodies were purchased from Molecular Probes and used at 1:200 dilution. For western blotting,peroxidase-conjugated secondary antibodies were purchased from Jackson ImmunoResearch Laboratories.

Gene targeting

A genomic Dlgh1 DNA clone was isolated from a TT2 genomic library,and used to generate a Dlgh1-targeting vector containing a short homologous arm (0.85 kb) on the 5′ side of exon 3, and a long homologous arm (8.5 kb) on the 3′ side of exon 3. A neomycin resistance (neo)cassette without a promoter or polyadenylation signal was inserted in-frame between the two homologous arms at the ClaI site in exon 3(Fig. 1A). In this mutant allele, the DNA sequence encoding the N-terminal 18 amino acids of Dlgh1 was followed by the neo cassette. Therefore, the protein encoded by the mutant allele would only contain a small part of the L27 domain, and none of the other domains of Dlgh1, such as the PDZ or SH3 domains. TT2 embryonic stem(ES) cells were electroporated and clones selected by standard procedures. Surviving clones were screened for homologous recombination by PCR using the following primers: 5′ Dlg-forward primer (F),5′-GCTGTCAGTCCACAGCTAACACAGGCTACT-3′; 3′ Dlg-reverse primer(R), 5′-TGTCCTAAGTTAAGGACCATCTAGAGAGCC-3′; neo gene primer (N),5′-TCGTGCTTTACGGTATCGCCGCTCCCGATT-3′. Correctly targeted clones were used for aggregation with 8-cell embryos and chimeric males were mated with C57BL/6 females to generate heterozygous offspring. In this study, two strains of mice derived from different recombinant ES clones were used and the phenotypes were similar between the two strains. Both strains were maintained through strict brother-sister mating for more than ten generations. To minimize any possible impact of the undefined genetic background, we used littermates as controls in all experiments. For western blotting, tissue lysates were prepared from the brain and kidney of E15.5 mice. Aliquots of total protein (5 μg) were electrophoresed in individual lanes and then transferred onto a PVDF membrane. Dglh1 protein was immunodetected using the SuperSignal WestFemto Maximum Sensitivity Substrate (Pierce). RT-PCR was performed on RNA samples from the kidney of E15.5 mice using Omniscript reverse transcriptase (Qiagen) and ExTaq (Takara), following the manufacturers' protocols. All animal protocols were carried out according to the Guidelines for the Management of Laboratory Animals in Fujita Health University. The numbers of mice used in this study are indicated in Table 1.

Embryo dissections and measurements

At the indicated stages, the fetal urogenital organs were dissected in PBS under a dissecting microscope (MZ FLIII; Leica). To measure the ureteric and metanephric length of E18.5 mice, microscopic photo-images were taken and each length was measured using LSM Image Browser software (Carl Zeiss). For earlier developmental stages, tissues were visualized by staining with an anti-Pax2 antibody and fluorescence images were used for the measurements.


For histological analyses, either the kidneys or urogenital organs were dissected from E18.5 embryos and fixed overnight in 4% paraformaldehyde (PFA)in PBS. Next, the tissues were embedded in paraffin and serially sectioned at 5 μm. The sections were stained with Hematoxylin and Eosin. For measurement of the renal area, longitudinal sections were prepared and microscopic images were captured using a BX50 (Olympus) and DC200 (Leica) digital camera system. The renal area, except for the cavity of the renal pelvis, was measured in each section using an imaging application software package (Adobe Photoshop version 7), and the largest area among serial sections was used for comparisons among genotypes.

Whole-mount immunofluorescence staining

To visualize the embryonic urogenital organs, dissected organs were immunofluorescently stained with an anti-Pax2 antibody. Urogenital tissues were dissected from fetuses in PBS and fixed in 4% PFA in PBS for 10-30 minutes. The tissues were incubated with an anti-Pax2 antibody diluted in PBS containing 0.1% Tween 20 (PBST) and 2% normal goat serum (2% NGS/PBST) at 4°C overnight. In some experiments, an anti-pancytokeratin antibody was added to this solution. After the incubation, the tissues were washed with PBST (1×3 minutes, 2×30 minutes, then overnight) and then incubated with Alexa Fluor 488-conjugated anti-rabbit IgG diluted in 2%NGS/PBST. For detection of the anti-pan-cytokeratin antibody, Alexa Fluor 568-conjugated anti-mouse IgG was used. Finally, the tissues were washed as described above, immersed in glycerol and observed under a fluorescence microscope (SteREO Lumar V12; Carl Zeiss).


E11.5 embryos from pregnant wild-type mice were fixed overnight in 4% PFA in PBS at 4°C. For immunohistochemistry of occludin and ZO-1, urogenital organs dissected from E15.5 embryos from pregnant heterozygous mice were fixed for 30 minutes in 4% PFA in PBS at room temperature and frozen in OCT compound(Sakura). The frozen tissues were then sectioned and immunostained as described previously (Iizuka-Kogo et al.,2005). For detection of Dlgh1 and E-cadherin in E14.5 embryos,embryos were fixed overnight in 4% PFA in PBS at 4°C. After embedding in paraffin, the tissues were sectioned at 5 μm, rehydrated and stained as for the frozen sections. The stained sections were observed under a laser scanning confocal microscope (LSM 510; Carl Zeiss).

Detection of cell proliferation by 5-bromo-2′-deoxyuridine(BrdU) labeling

E12.5 pregnant heterozygous mice were injected with 50 μg of BrdU per g body weight. 2 hours after the injection, the embryos were collected, fixed overnight in 4% PFA in PBS and paraffin-embedded. Horizontal serial cross-sections of the embryos were cut at 5 μm thickness and 20 μm intervals from the ureteropelvic junction to the lower end of the Wolffian duct at the urogenital sinus. Thereafter, BrdU incorporated into the ureteric epithelium was detected using an anti-BrdU antibody(5-Bromo-2′-deoxyuridine Labeling and Detection Kit; Roche) in combination with a Histofine Mouse Stain Kit (Nichirei), according to the manufacturers' protocols. The number of immunopositive nuclei per unit area of each ureteric cross-section was calculated in every section and the mean values were compared between Dlgh1+/+ and Dlgh1-/- mice.

Statistical analysis

F-tests and t-tests were used for statistical evaluation of the data. All statistical analyses were performed using Microsoft Excel 2003(Microsoft Corporation). Values of P<0.05 were considered to indicate statistical significance.

Generation of Dlgh1 mutant mice

To clarify the functions of Dlgh1 during development, Dlgh1-null mutant mice were generated by homologous recombination in ES cells. The Dlgh1 gene was disrupted by inserting a neo cassette into exon 3(Fig. 1A). Among 45 recombinant clones screened by PCR, we identified three clones that showed site-specific recombination (Fig. 1B,C). Two of the three clones gave rise to chimeras that transmitted the mutated allele to their offspring. Western blotting using an anti-Dlgh1 antibody confirmed the absence of Dlgh1 protein in Dlgh1-/- mice(Fig. 1D). Furthermore, we performed RTPCR to exclude the possibility of exon 3 skipping in the mutated allele, which would leads to the expression of a nearly full-length protein but lacking the 44 amino acids encoded by exon 3. An exon 3-skipped mRNA was not amplified in Dlgh1-/- mice (see Fig. S1 in the supplementary material). These results confirmed that the mutant was a Dlgh1-null mutant.

Abnormal structures of urogenital organs in prenatal Dlgh1-/- mice

Since Dlgh1-/- mice die soon after birth, prenatal E18.5 embryos were used for the following observations. The mean body length(crown-to-rump length, CRL) was significantly shorter in Dlgh1-/- mice than in Dlgh1+/+ or Dlgh1+/- mice (see Fig. S2A in the supplementary material). In addition, a cleft palate was observed in Dlgh1-/- mice, as reported previously(Caruana and Bernstein, 2001). Furthermore, the whole-intestine length and whole-lung weight of Dlgh1-/- mice were significantly reduced compared with those of Dlgh1+/+ or Dlgh1+/- mice,whereas the length of the left-right and sagittal axes of the cerebrum and the whole-liver weight were not affected in Dlgh1-/- mice(A.I.-K., T.A. and T.S., unpublished). In the present study, we focused on the morphology of the urogenital organs. In contrast to Dlgh1+/+ mice (Fig. 2A-C), hypoplasia of the kidney and ureter(Fig. 2D), megaureter(Fig. 2E), duplicated ureter(Fig. 2F), hydronephrosis(Fig. 2G), malposition of the gonads and vaginal aplasia (Fig. 2H) were found in Dlgh1-/- mice at various incidences. These abnormalities were not seen in Dlgh1+/+mice (Fig. 2A-C,I,J).

First, in order to clarify whether the kidney hypoplasia was associated with the impaired growth of the whole body, relative kidney length (ratio of the longitudinal kidney length to the CRL) was compared among the genotypes. As shown in Fig. 2I, the relative kidney length was significantly shorter in Dlgh1-/- mice than in Dlgh1+/+ mice. Given these results, the kidney hypoplasia does not appear to be a result of the decrease in body size. Histological observation did not reveal any significant differences in the number of glomeruli per unit area or histological dysplasia of the renal tissues. However, hydronephrosis was seen in 15% of Dlgh1-/- mouse kidneys, but not in Dlgh1+/+ mice (Fig. 2B,G).

Next, the ureteric lengths were compared among the genotypes. As shown in Fig. 2J, the mean ureteric length relative to the CRL was significantly decreased in Dlgh1-/- mice, and all eight Dlgh1-/-mice used for the calculation had unilateral or bilateral hypoplastic ureter(s). The other ureteric abnormalities, namely bilateral megaureter(Fig. 2E) and unilateral duplicated ureter (Fig. 2F),were each observed in one of 11 Dlgh1-/- mice.

Regarding the reproductive organs, malposition of the ovaries and testes were often seen in Dlgh1-/- mice (five of eight Dlgh1-/- mice). In these mice, the uteri were located on the abdominal side of the kidneys with the ovaries positioned between the adrenal glands and the kidneys in the females, and the testes were found in the abdominal cavity at irregular positions in the males (data not shown). In Dlgh1+/+ and Dlgh1+/- female mice, the vaginal cavity formed downward behind the urethra(Fig. 2C). By contrast, no such downward formation of the vaginal cavity was seen in Dlgh1-/- female mice(Fig. 2H).

Development of hypoplastic kidneys and ureters in Dlgh1-/- mice

To clarify the nature of the abnormalities in the urinary organs in prenatal Dlgh1-/- mice, we followed their development. To visualize the structures of the urinary organs, we performed whole-mount immunostaining for cytokeratin (an epithelial marker) and Pax2 (a transcription factor). Since Pax2 is expressed in the Wolffian and Müllerian ducts, metanephric mesenchyme and ureters(Torres et al., 1995), their shapes can be observed by means of whole-mount immunofluorescence staining. Immunofluorescence staining of the urinary organs revealed that the Pax2-positive mesenchyme, denoting the metanephric mesenchyme, was contracted in Dlgh1-/- mice during early development of the kidney(Fig. 3A-C; see Fig. S2B-E in the supplementary material). Regarding the ureteric structure, budding and invasion of the ureter from the duct into the metanephric mesenchyme were seen in both Dlgh1+/+ and Dlgh1-/- mice at E11.2 (Fig. 3A). In normal embryos, the ureteric bud divides for the first time and forms a T-shaped structure by E11.5 (Fig. 3B,upper, inset). However, among the E11.5 Dlgh1-/- embryos,bilateral ectopic caudal branching of the ureters was observed in one of seven Dlgh1-/- mice (Fig. 3B, lower, inset). This ectopic branching of the ureter might cause the ureteric duplication seen at E18.5. At E12.5, the numbers of ureteric bud tips in the metanephric mesenchyme, which are formed through ureteric branching, were statistically equivalent among the different Dlgh1 genotypes. In addition, the normal ureteric branching pattern was observed in all Dlgh1+/+ and Dlgh1+/- kidneys and in some of the Dlgh1-/- kidneys (Fig. 3C, upper, inset). However, in 31.8% of cases, the Dlgh1-/- ureters showed abnormal branching structures(Fig. 3C, lower, inset). The ureteric length from the lower end of the ureter to the upper end of the T-shaped ureteric bud tip did not differ significantly between Dlgh1-/- mice and the other genotypes at E11.5. By contrast, by E12.5, the ureteric length relative to the CRL was significantly shorter in Dlgh1-/- mice than in the other genotypes (see Fig. S2F in the supplementary material).

Development of ectopic ureters in Dlgh1-/-mice

Megaureter and hydronephrosis were found in Dlgh1-/-mice at E18.5. These disorders are caused by urinary obstruction, which is induced by either constriction of the ureter, or by ectopic opening of the ureter. Since the ureter sprouts from the Wolffian duct, the lower end of the ureter does not open into the bladder immediately after budding. At this stage, the lower part of the Wolffian duct beneath the budding site of the ureter is known as the common nephric duct(Fig. 3A-D, the region between the two arrowheads). The common nephric duct is thought to become incorporated into the bladder wall by E13.5, when the lower end of the ureter descends from the Wolffian duct to the bladder wall. After E14.5, the lower ends of the ureter and the Wolffian duct become separated from each other by proliferation of the epithelial cells between the two ducts(Batourina et al., 2002). These successive changes were observed in Dlgh1+/+ mice(Fig. 3, upper panels, Fig. 5B). However, in all Dlgh1-/- mice at E12.5, the lower end of the ureter was connected to the Wolffian duct and did not descend to the bladder wall(Fig. 3C, lower). Furthermore,no Dlgh1-/- mice showed complete ureteric descent by E13.5 and they retained the common nephric duct(Fig. 3D, lower). The common nephric duct in normal development is equal to the lower part of the Wolffian duct. We compared tissue sections around the remaining common nephric duct in E13.5 Dlgh1-/- mice with the equivalent region containing the base of the ureter and the Wolffian duct in Dlgh1+/+mice at the same stage (Fig. 4). In Dlgh1+/+ mice, the ureter was surrounded by sparse mesenchymal cells(Fig. 4A), whereas the Wolffian duct was surrounded by dense mesenchymal cells that showed no definite orientation (Fig. 4B). In Dlgh1-/- mice, the mesenchymal tissue on the ventral side of the remaining common nephric duct (Fig. 4C, lower right) was dense and similar to that around the normal Wolffian duct (Fig. 4B). By contrast, the dorsal side of the duct was surrounded by a rather sparse mesenchymal tissue in which the cells were orientated toward the duct(Fig. 4C, upper left). This sparse mesenchymal tissue was similar to that around the normal ureter(Fig. 4A). Taken together, the remaining common nephric duct in Dlgh1-/- mice was surrounded by distinct mesenchymal tissues different from that around the Wolffian duct in Dlgh1+/+ mice. Subsequently, at E15.5 or later, some Dlgh1-/- ureters seemed to connect directly to the bladder wall (Fig. 5H, left side), even though the connecting structures among the ureter, bladder and Wolffian duct were very different from those in Dlgh1+/+mice, and these mice appeared to develop a passable urinary tract. However,some Dlgh1-/- ureters were dilated owing to urinary obstruction at E16.5 (Fig. 5I),indicating their failure to connect to the bladder. Therefore, such ectopic connection of the ureters seems to cause the megaureter and hydronephrosis found in Dlgh1-/- mice at E18.5.

The occurrence of megaureter or hydronephrosis caused by descent failure of the ureteric opening was previously reported in Bmp4+/-mice (Miyazaki et al., 2000),in which the lower end of the ureter did not reach the bladder wall because the budding site of the ureter was too rostral. By contrast, a comparison of the lengths of the common nephric duct at E11.2, immediately after penetration of the ureteric bud into the metanephric mesenchyme, indicated no significant difference between Dlgh1+/+ and Dlgh1-/- mice. As late as E11.5, the common nephric duct length in Dlgh1-/- mice was 1.71-fold longer than that in Dlgh1+/+ mice, a significant difference(P<0.01). This indicates that the descent failure of the ureteric end is not caused by ectopic budding of the ureter, but occurs during the descent process itself.

Development of the genital organs derived from the Wolffian and Müllerian ducts

Vaginal aplasia was observed in prenatal Dlgh1-/- mice. To clarify the mechanism of this vaginal aplasia, we followed the development of the Müllerian ducts, from which the upper vagina is thought to be derived.

In Dlgh1+/+ mice, the lower part of the Müllerian duct became identifiable beside the Wolffian duct at E13.5, and was close to the bladder wall (Fig. 3D,upper). At E14.5, the lower ends of the Wolffian and ureteric ducts separated from each other (Fig. 5A,B). The lower ends of the Müllerian ducts were laterally fused in both sexes at E15.5 (Fig. 5C,D). At this stage, attenuation of the Wolffian and Müllerian ducts, indicating their degeneration, was observed in the female and male embryos, respectively. By E16.5, the female Wolffian ducts and male Müllerian ducts had almost completely disappeared (Fig. 5E,F). In addition, in the female embryos, the fusion part of the Müllerian ducts, which should form the uterine cervix and upper vagina,was seen to expand laterally and downward(Fig. 5E). In the male embryos,the primordium for the seminal vesicle emerged from the dorsal side of the lowest part of the Wolffian duct (Fig. 5F).

In contrast to Dlgh1+/+ mice, the lower end of the Müllerian duct in Dlgh1-/- mice had not elongated to the bladder wall by E13.5 (Fig. 3D, lower). Instead, in most embryos, it only reached to just above the branching point of the ureter from the Wolffian duct. At E14.5, the ureter and Wolffian duct remained connected with each other. In some embryos,the Müllerian duct was connected to the common nephric duct(Fig. 5G). Furthermore, the Müllerian ducts had not fused laterally by E15.5, and the lower end of the ureter did not separate from the Wolffian duct(Fig. 5H). As described above,some female embryos appeared to have a dilated ureter with a constriction(Fig. 5I), which might have been caused by the protracted ureteric connection to the degenerating Wolffian duct. In male embryos, primordia for the seminal vesicles could not be identified (Fig. 5J). The failure of lateral fusion and the obstruction of the Müllerian ducts in female Dlgh1-/- mice were confirmed by observation of serial cross-sections of the urogenital organs at E18.5. Fig. 6 shows cross-sections of Dlgh1+/+ and Dlgh1-/- mice at the level of the vesicoureteral junction. At this level, the uteri were fused laterally and a cervical lumen was present in Dlgh1+/+mice (Fig. 6A). By contrast,the uterine lumens were separated laterally in Dlgh1-/-mice (Fig. 6B), and were seen to disappear in serial sections (data not shown).

Characterization of the ureteric epithelium in Dlgh1-/- mice

Immunofluorescence staining of the embryonic urogenital organs with an anti-Dlgh1 antibody revealed intense labeling of the epithelial cells and weaker labeling of the surrounding mesenchymal cells(Fig. 7A). The distribution of Dlgh1 was broader than that of Pax2 (Fig. 7B), a key transcription factor essential for urogenital development, and included the intestinal epithelia. Dlgh1 accumulated at the basolateral region of the epithelial cells, as previously described for other epithelial cells (Iizuka-Kogo et al.,2005; Matsumine et al.,1996; Muller et al.,1995). Knockdown of Dlgh1 expression in Caco-2/15 cells has been reported to decrease the accumulation of E-cadherin (also known as cadherin 1)at cellcell contact sites, thereby inhibiting the structural integration of adherens junctions (Laprise et al.,2004). To examine whether the abnormal structure of the ureters in Dlgh1-/- mice was caused by changes in the structure of cell-cell junctional complexes, we determined the distribution of E-cadherin,occludin and ZO-1 (also known as Tjp1) in the embryonic ureteric epithelium. Contrary to our expectations, strong accumulation of E-cadherin at cell-cell contact sites and accumulation of occludin and ZO-1 at the apicalmost portion of the lateral membrane were also observed in Dlgh1-/-mice (Fig. 7C-J).

The ureteric or renal hypoplasia seen in Dlgh1-/- mice could be caused by defective cell proliferation. We therefore compared the proliferative index in the ureteric epithelium of Dlgh1-/-mice with that of Dlgh1+/+ mice(Table 2). The proliferative index of Dlgh1-/- mice at E12.5, when prominent elongation of the ureter occurs in normal mice, was significantly lower than that of Dlgh1+/+ mice. This finding suggests that decreased proliferation in the ureteric epithelium causes the short ureter of Dlgh1-/- mice.

Molecular region of the Dlgh1 protein responsible for development of the urogenital tract

We performed gene targeting of Dlgh1. Functional Dlgh1 protein appeared to be absent from the generated mice. In the present study, we have described various abnormalities in the urogenital organs of Dlgh1-/- mice. These defects occurred at various stages of development, indicating that Dlgh1 continues to perform important functions in these tissues. Recently, Naim et al. reported renal hypoplasia and decreased numbers of ureteric branches in the kidneys of gene-targeted mice lacking the C-terminus of Dlgh1 protein (Naim et al.,2005). In that study, however, no other morphological abnormalities of the ureter and genital tract that resembled those observed in our gene-targeted mice were described. If developmental abnormalities in the lower urogenital tract occur only in our mutant mice, and not in mutant mice lacking the Cterminus of Dlgh1, this would suggest that the N-terminal half of Dlgh1, including the PDZ domains, is essential for development of the lower urogenital tract. On the other hand, as renal hypoplasia was present in both types of mutant mice, this phenotype appears to be due to the loss of the C-terminus of Dlgh1.

Mechanism of the occurrence of renal hypoplasia

Pax2 is a transcription factor that plays important roles in the early development of urogenital organs (Torres et al., 1995). The Pax2-positive metanephric mesenchyme was clearly contracted in Dlgh1-/- mice. Pax2-positive mesenchymal cells have the ability to differentiate into convoluted tubules(Rothenpieler and Dressler,1993). Therefore, their number could contribute directly to the size of the mature kidneys. Furthermore, a decrease in the total Pax2 expression level in the metanephric mesenchyme might repress the expression of Gdnf, which is a target gene of Pax2 and essential for ureteric branching. Consequently, both the differentiation of the convoluted tubules and the ureteric branching might have been affected, which would lead to hypoplasticity of the kidneys in Dlgh1-/- mice. The mechanism regulating the expression of Pax2 in the metanephric primordium has not yet been clarified. Elucidation of the mechanism by which Dlgh1 affects this process would be of great interest.

Cause and effect of the disturbed development of the ureter

Prenatal hydronephrosis and megaureter are considered to arise owing to failure of ureteric opening into the bladder wall, and are referred to as CAKUT (congenital abnormalities of the kidney and urinary tract). In Foxc1-/- (Kume et al.,2000) and Bmp4+/-(Miyazaki et al., 2000) mice,the ureter buds from an ectopically anterior site, and therefore cannot descend sufficiently or open into the bladder wall, thus resulting in hydronephrosis and other correlated defects similar to those in Dlgh1-/- mice. In addition, mice homozygous for the retinoic acid receptor show hydronephrosis and megaureter(Batourina et al., 2002). In this case, the budding site and descent process of the ureter are normal, but the lower end of the ureter does not separate from the Wolffian duct on the bladder wall. By contrast, the descent process itself does not progress completely in Dlgh1-/- mice, suggesting that Dlgh1 functions in a different context from the other genes described above. The common nephric duct is thought to become incorporated into the bladder wall during the descent of the ureteric end. BrdU analysis revealed that cellular proliferation is decreased in the ureteric epithelium of Dlgh1-/- mice at E12.5. Therefore, it might be possible that elongation of the ureter caused by epithelial proliferation is required for incorporation of the common nephric duct into the bladder wall. If this is the case, then the loss of Dlgh1 and consequent decreased cellular proliferation would result in impairment of the descent process of the ureter.

Mechanism of the defects in Müllerian duct development

In addition to the abnormalities in the urinary organs, aplasia of the uterine cervix and vagina was observed in Dlgh1-/- mice. This is caused by fusion failure and obstruction of the Müllerian ducts. Why were the Müllerian ducts obstructed in Dlgh1-/-mice? Since lateral fusion of the Müllerian ducts can be impaired without obstruction in humans (Gell,2003), fusion failure does not seem to be a direct cause of the obstruction. Rather, the impaired downgrowth of the Müllerian ducts appears to cause the obstruction in Dlgh1-/- mice. Kobayashi et al. confirmed that the Wolffian duct is required for Müllerian duct formation by demonstrating that Wolffian ductspecific knockout of the Lim1 gene (Lhx1 - Mouse Genome Informatics),which is required for Wolffian duct development, inhibited the formation of the Müllerian duct as well as that of the Wolffian duct itself(Kobayashi et al., 2005). As mentioned above, the mesenchymal tissues surrounding the Wolffian duct and ureter are histologically very different. This indicates that the Wolffian duct and ureter develop different characteristics through interactions with their surrounding mesenchymal tissues, even though they are derived from the same origin. By contrast, the common nephric duct in Dlgh1-/- mice was surrounded by two types of mesenchymal tissues, ventral Wolffian-type mesenchyme and dorsal ureter-type mesenchyme. The persistent common nephric duct, unlike the normal Wolffian duct, might not induce the formation of the Müllerian duct. Determination of the functional difference between the Wolffian duct and common nephric duct in Dlgh1-/- mice might help to elucidate the induction mechanism of the Müllerian duct by the Wolffian duct.

The mechanism for the lateral fusion of the Müllerian ducts is presently unclear. Because downward elongation of the Müllerian ducts was impaired in Dlgh1-/- mice, the Müllerian ducts might be unable to access each other adequately. Alternatively, there might be defects in the cell-cell recognition process that appears to be necessary during the integration of the two epithelial cell layers of the left and right Müllerian ducts.

Congenital absence of the uterus and vagina is known as Rokitansky-Kuster-Hauser Syndrome in humans. In addition, and as mentioned above, failure of lateral fusion of the Müllerian ducts with complete formation of the vagina has been reported(Gell, 2003). These disorders are associated with abnormalities of the renal and urinary organs with various incidences, similar to those in Dlgh1-/- mice, suggesting that the functions of Dlgh1 are impaired at the onset of these congenital abnormalities.

A possible function of Dlgh1 in promoting cellular proliferation in the ureter

In mature epithelial cells, Dlgh1 is localized at cell-cell contact sites. Dlgh1 has previously been considered to facilitate cell-cell adhesion, thereby inhibiting cellular proliferation. Regarding adhesion, Dlgh1 has been reported to be involved in cellular polarization and/or integration of epithelial cell-cell adhesion complexes in Nematoda and Drosophila and in human Caco-2/15 cells (Firestein and Rongo,2001; Laprise et al.,2004; Woods and Bryant,1991). Regarding proliferation, Dlgh1 has been reported to cooperate with APC and negatively regulate the cell cycle in the mouse fibroblast cell line NIH3T3 (Ishidate et al., 2000). In addition, the onset of cervical cancer mediated by the human papilloma virus is associated with virus-induced degradation of Dlgh1 protein at cell-cell contact sites(Watson et al., 2002). However, our present results indicate that Dlgh1 is not required for the proper subcellular localization and maintenance of E-cadherin, occludin and ZO-1 in the developing ureter. Furthermore, the BrdU analysis revealed that cellular proliferation was decreased in the ureteric epithelium of Dlgh1-/- mice, and this probably underlies the shortening of the ureters in Dlgh1-/- mice. Likewise, the hypoplasia of the metanephros and maxillary process and the decreased body size of Dlgh1-/- mice might also be due to the decreased cellular proliferation. From the present results, we speculate that, under some specific conditions during development, Dlgh1 might be involved in the correct localization of intracellular signal transduction components and contribute to efficient signal reception and transduction, resulting in positive regulation of cellular proliferation. Recently, Frese et al. reported that Dlgh1 is positively involved in human adenovirus type-9 E4-mediated oncogenesis, and demonstrated a novel mechanism in which Dlgh1 changes its subcellular distribution through an interaction with E4-ORF1 oncoprotein and subsequently induces cell-autonomous proliferation(Frese et al., 2006). These results indicate that, according to the cellular conditions and through interactions with other functional proteins, Dlgh1 can carry out various functions that might sometimes be opposing, such as promoting and inhibiting cellular proliferation, and are consistent with our present results and hypothesis.

It will be important to clarify the developmental mechanisms underlying the urogenital disorders in Dlgh1-/- mice in order to fully understand the ontogenic functions of Dlgh1. The present results indicate that Dlgh1 is involved in regulating the structural organization of the epithelial ducts during ontogeny. Identification and characterization of the interactions between Dlgh1 and its binding partners in the urogenital organs and examination of the roles of Dlgh1 in other developmental processes will help to further elucidate its developmental functions.

We thank Drs Ryuji Nomura, Yoshimi Hasegawa, Atsushi Shimomura and Hiroshi Kogo for helpful discussions. We also thank Hiroe Muramatsu, Mikiko Ichikawa,Mikako Ohno, Emiko Kodera, Kazuhiro Yanagisawa, Yohei Takeuchi and Kazuko Hikita for technical support. This research was supported by the following:Grant-in-Aid for Scientific Research on Priority Areas 17052025 to A.I.-K.;Grant-in-Aid for Young Scientists (B) 17790147 to A.I.-K.; Grant-in-Aid for Scientific Research (B) 16390051 to T.S.; grants from the Special Fund for Science and High Promotion in Private Universities to T.S. from the Japanese Ministry of Education, Culture, Sports, Science and Technology; grants from the Science Research Promotion Fund to T.S.; and grants from the Fujita Health University Research Fund to A.I.-K. and T.S.

Batourina, E., Choi, C., Paragas, N., Bello, N., Hensle, T.,Costantini, F. D., Schuchardt, A., Bacallao, R. L. and Mendelsohn, C. L.(
). Distal ureter morphogenesis depends on epithelial cell remodeling mediated by vitamin A and Ret.
Nat. Genet.
Caruana, G. and Bernstein, A. (
). Craniofacial dysmorphogenesis including cleft palate in mice with an insertional mutation in the discs large gene.
Mol. Cell. Biol.
Etienne-Manneville, S., Manneville, J. B., Nicholls, S.,Ferenczi, M. A. and Hall, A. (
). Cdc42 and Par6-PKCzeta regulate the spatially localized association of Dlg1 and APC to control cell polarization.
J. Cell Biol.
Firestein, B. L. and Rongo, C. (
). DLG-1 is a MAGUK similar to SAP97 and is required for adherens junction formation.
Mol. Biol. Cell
Frese, K. K., Latorre, I. J., Chung, S. H., Caruana, G.,Bernstein, A., Jones, S. N., Donehower, L. A., Justice, M. J., Garner, C. C. and Javier, R. T. (
). Oncogenic function for the Dlg1 mammalian homolog of the Drosophila discslarge tumor suppressor.
Gell, J. S. (
). Mullerian anomalies.
Semin. Reprod. Med.
Iizuka-Kogo, A., Shimomura, A. and Senda, T.(
). Colocalization of APC and DLG at the tips of cellular protrusions in cultured epithelial cells and its dependency on cytoskeletons.
Histochem. Cell Biol.
Ishidate, T., Matsumine, A., Toyoshima, K. and Akiyama, T.(
). The APChDLG complex negatively regulates cell cycle progression from the G0/G1 to S phase.
Kobayashi, A., Kwan, K. M., Carroll, T. J., McMahon, A. P.,Mendelsohn, C. L. and Behringer, R. R. (
). Distinct and sequential tissue-specific activities of the LIM-class homeobox gene Lim1 for tubular morphogenesis during kidney development.
Kume, T., Deng, K. and Hogan, B. L. (
). Murine forkhead/winged helix genes Foxc1 (Mf1) and Foxc2 (Mfh1) are required for the early organogenesis of the kidney and urinary tract.
Laprise, P., Viel, A. and Rivard, N. (
). Human homolog of disc-large is required for adherens junction assembly and differentiation of human intestinal epithelial cells.
J. Biol. Chem.
Lee, S., Fan, S., Makarova, O., Straight, S. and Margolis,B. (
). A novel and conserved protein-protein interaction domain of mammalian Lin-2/CASK binds and recruits SAP97 to the lateral surface of epithelia.
Mol. Cell. Biol.
Leonard, A. S., Davare, M. A., Horne, M. C., Garner, C. C. and Hell, J. W. (
). SAP97 is associated with the alpha-amino-3-hydroxy-5-methylisoxazole-4-propionic acid receptor GluR1 subunit.
J. Biol. Chem.
Matsumine, A., Ogai, A., Senda, T., Okumura, N., Satoh, K.,Baeg, G. H., Kawahara, T., Kobayashi, S., Okada, M., Toyoshima, K. et al.(
). Binding of APC to the human homolog of the Drosophila discs large tumor suppressor protein.
Miyazaki, Y., Oshima, K., Fogo, A., Hogan, B. L. and Ichikawa,I. (
). Bone morphogenetic protein 4 regulates the budding site and elongation of the mouse ureter.
J. Clin. Invest.
Muller, B. M., Kistner, U., Veh, R. W., Cases-Langhoff, C.,Becker, B., Gundelfinger, E. D. and Garner, C. C. (
). Molecular characterization and spatial distribution of SAP97, a novel presynaptic protein homologous to SAP90 and the Drosophila discs-large tumor suppressor protein.
J. Neurosci.
Naim, E., Bernstein, A., Bertram, J. F. and Caruana, G.(
). Mutagenesis of the epithelial polarity gene, discs large 1, perturbs nephrogenesis in the developing mouse kidney.
Kidney Int.
Rothenpieler, U. W. and Dressler, G. R. (
). Pax-2 is required for mesenchyme-to-epithelium conversion during kidney development.
Torres, M., Gomez-Pardo, E., Dressler, G. R. and Gruss, P.(
). Pax-2 controls multiple steps of urogenital development.
Watson, R. A., Rollason, T. P., Reynolds, G. M., Murray, P. G.,Banks, L. and Roberts, S. (
). Changes in expression of the human homologue of the Drosophila discs large tumour suppressor protein in high-grade premalignant cervical neoplasias.
Woods, D. F. and Bryant, P. J. (
). The discs-large tumor suppressor gene of Drosophila encodes a guanylate kinase homolog localized at septate junctions.