Faculty Bibliography

The apical surfaces of urothelial cells are almost entirely covered with plaques consisting of crystalline, hexagonal arrays of 16 nm uroplakin particles. Although all four uroplakins, when SDS-denatured, can be digested by chymotrypsin, most uroplakin domains in native urothelial plaques are resistant to the enzyme, suggesting a tightly packed structure. The only exception is the C-terminal, cytoplasmic tail of UPIII (UPIII) which is highly susceptible to proteolysis, suggesting a loose configuration. When uroplakins are solubilized with 2% octylglucoside and fractionated with ion exchangers, UPIa and UPII were bound as a complex by a cation exchanger, whereas UPIb and UPIII were bound by an anion exchanger. This result is consistent with the fact that UPIa and UPIb are cross-linked to UPII and UPIII, respectively, and suggests that the four uroplakins form two pairs consisting of UPIa/II and UPIb/III. Immunogold labelling using a new mouse monoclonal antibody, AU1, revealed that UPIII is present in all urothelial plaques, indicating that the two uroplakin pairs are not segregated into two different types of urothelial plaque and that all plaques must have a similar uroplakin composition. Taken together, these results indicate that uroplakins form a tightly packed structure, that the four uroplakins interact specifically forming two pairs, and that both uroplakin pairs are required for normal urothelial plaque formation

Urothelium synthesizes a group of integral membrane proteins called uroplakins, which form two-dimensional crystals (urothelial plaques) covering >90% of the apical urothelial surface. We show that the ablation of the mouse uroplakin III (UPIII) gene leads to overexpression, defective glycosylation, and abnormal targeting of uroplakin Ib, the presumed partner of UPIII. The UPIII-depleted urothelium features small plaques, becomes leaky, and has enlarged ureteral orifices resulting in the back flow of urine, hydronephrosis, and altered renal function indicators. Thus, UPIII is an integral subunit of the urothelial plaque and contributes to the permeability barrier function of the urothelium, and UPIII deficiency can lead to global anomalies in the urinary tract. The ablation of a single urothelial-specific gene can therefore cause primary vesicoureteral reflux (VUR), a hereditary disease affecting approximately 1% of pregnancies and representing a leading cause of renal failure in infants. The fact that VUR caused by UPIII deletion seems distinct from that caused by the deletion of angiotensin receptor II gene suggests the existence of VUR subtypes. Mutations in multiple gene, including some that are urothelial specific, may therefore cause different subtypes of primary reflux. Studies of VUR in animal models caused by well-defined genetic defects should lead to improved molecular classification, prenatal diagnosis, and therapy of this important hereditary problem

Urothelial surface is covered by numerous plaques (consisting of asymmetric unit membranes or AUM) that are interconnected by ordinary looking hinge membranes. We describe an improved method for purifying bovine urothelial plaques using 2% sarkosyl and 25 mM NaOH to remove contaminating membrane and peripheral proteins selectively. Highly purified plaques interconnected by intact hinge areas were obtained, indicating that the hinges are as detergent-insoluble as the plaques. These plaque/hinge preparations contained uroplakins, an as yet uncharacterized 18-kDa plaque-associated protein, plus an 85-kDa glycoprotein that is known to be hinge-associated in situ. Examination of the isolated, in vitro-resealed bovine AUM vesicles by quick-freeze deep-etch showed that each AUM particle consists of a 16-nm, luminally exposed 'head' anchored to the lipid bilayer via a 9-mm transmembranous 'tail', and that an AUM plaque can break forming several smaller plaques separated by newly formed particle-free, hinge-like areas. These data lend support to our recently proposed three-dimensional model of mouse urothelial plaques. In addition, our findings suggest that urothelial plaques are dynamic structures that can rearrange giving rise to new plaques with intervening hinges; that the entire urothelial apical surface (both plaque and hinge areas) is highly specialized; and that these two membrane domains may be equally important in fulfilling some of the urothelial functions

The luminal surface of mouse urothelium in contact with the urine is almost entirely covered with plaques consisting of uroplakin-containing particles that form p6 hexagonal crystals with a center-to-center distance of 16 nm. A combination of quick-freeze/deep-etch images and our previous negative staining data indicate that the head domain of the uroplakin particle, which is exposed without an extensive glycocalyx shield, interacts closely with the head domains of the neighboring particles, while the membrane-embedded tail domains are farther apart; and that urothelial particles and plaques are not rigid structures as they can change their configuration in response to mechanical perturbations. Based on these data, we have constructed three-dimensional models depicting the structural organization of urothelial particles and plaques. Our models suggest that the head-to-head interaction may play a key role in determining the shape and size of the urothelial plaques. These models can explain many properties of urothelial plaques including their unique shape, detergent-insolubility, and morphological changes during vesicle maturation.

Uroplakins (UPs) are integral membrane proteins that are synthesized as the major differentiation products of mammalian urothelium. We have cloned the human UP-II gene and localized it on chromosome 11q23. A survey of 50 transitional cell carcinomas (TCCs) revealed a UP-II polymorphism but no tumor-specific mutations. Immunohistochemical staining using rabbit antisera against a synthetic peptide of UP-II and against total UPs showed UP reactivity in 39.5% (17 of 43 cases) of conventional TCCs, 12.8% (5 of 39) of bilharzial-related TCCs, and 2.7% (1 of 36) of bilharzial-related squamous cell carcinomas (SCCs). The finding that fewer bilharzial TCCs express UPs than conventional TCCs (12.8 versus 40%) raised the possibility that the former are heterogeneous, expressing SCC features to varying degrees. Our data strongly support the hypothesis that urothelium can undergo at least three pathways of differentiation: (a) urothelium-type pathway; (b) epidermis-type pathway; and (c) glandular-type pathway, characterized by the production of UPs, K1/K10 keratins, and secreted glycoproteins, respectively. Vitamin A deficiency and mesenchymal factors may play a role in determining the relative contributions of these pathways to urothelial differentiation as well as to the formation of TCC, SCC, and adenocarcinoma, or a mixture thereof

Uroplakin genes are expressed in a bladder-specific and differentiation-dependent fashion. Using a 3.6-kb promoter of mouse uroplakin II gene, we have generated transgenic mice that express human growth hormone (hGH) in their bladder epithelium, resulting in its secretion into the urine at 100-500 ng/ml. The levels of urine hGH concentration remain constant for longer than 8 months. hGH is present as aggregates mostly in the uroplakin-delivering cytoplasmic vesicles that are targeted to fuse with the apical surface. Using the bladder as a bioreactor offers unique advantages, including the utility of all animals throughout their lives. Using urine, which contains little protein and lipid, as a starting material facilitates recombinant protein purification

TLS (FUS) and the related gene EWS encode the N-terminal portion of many fusion oncoproteins involved in human sarcomas and leukemia. TLS is an RNA-binding nuclear protein that is identical to hnRNP P2 and may be implicated in mRNA metabolism. When RNA polymerase II is inhibited, TLS immunostaining in the nucleus is dramatically altered, from its normal diffuse nucleoplasmic pattern to accumulation in dense nuclease-resistant aggregates. Co-immunostaining with antibodies to fibrillarin or p80 coilin and immunoelectron microscopy revealed that the TLS aggregates are associated with the nucleolus and are distinct from other known structures such as the coiled body or the interchromatin granule. Injection of cells with an oligodeoxynucleotide that disrupts splicing does not result in redistribution of TLS, indicating that the event is specific to inhibition of transcription. Oncoproteins that contain the N-terminal domain from either TLS, EWS or their Drosophila homologue, SARFH (CAZ), are also targeted to the same structure. These findings suggest a correlation between the topogenic and transforming activities of TLS and EWS N-termini and imply the existence of cellular targets that are shared by the germ-line encoded proteins and their oncogenic derivatives