Galectins and other endogenous carbohydrate-binding proteins of animal bladder.

Defects in the glycocalyx of the bladder epithelium may be related to the development of bladder diseases including interstitial cystitis which is a chronic bladder disease of unknown etiology. Indirect evidence has implicated alterations in the bladder epithelial glycoconjugates in interstitial cystitis and vesicaler instillation of glycosaminoglycans is promoted as treatments. However, information on the nature of the glycoconjugates of the bladder epithelium and lectins that may interact with the exogenous instilled glycoconjugates is very limited. We have examined the endogenous lectin associated with bladder epithelium by immunohistochemistry using biotinylated neoglycoconjugates. The strong calcium-independent binding of beta-D-galactose probe suggested the presence of galectins in rabbit and human bladder. Extracts of rabbit bladder organ cultures metabolically labeled with [14C]-amino acids were subjected to affinity chromatography on immobilized lactose and the specifically bound material eluted with 0.2 M lactose. SDS-PAGE of the recovered proteins revealed a major band of approximately 30 kDa and a minor band of 21 kDa. Polymerase chain reaction and northern blot analysis showed that both galectin-3 and galectin-4 are expressed in rabbit bladder. Since galectin-3 from rabbit had been previously cloned, we cloned and sequenced galectin-4 from rabbit bladder. The deduced full length sequence of 328 amino acids revealed four distinct regions: a N-terminal peptide of 19 residues, two carbohydrate recognition domains of 130 residues each, and a linker region of 49 residues. Comparison of the rabbit galectin-4 sequence with those of human, pig, rat, and mouse revealed two invariant peptide motifs that are proposed as signature sequences for identifying related galectins.

A new direct test of bladder permeability.

PURPOSE: A proposed cause of interstitial cystitis is increased bladder permeability but to our knowledge this theory has not been proved by direct testing. We developed a safe, relatively painless, direct test of bladder permeability. MATERIALS AND METHODS: The original permeability test involved placing 4% lactulose and 1% rhamnose intravesically, then drawing blood to assay for these sugars. Initial feasibility studies were performed in rabbits with bladder epithelium that was intact or disrupted by a 50% acetone rinse. In humans the initial goal was to distinguish intact bladders from those known to have increased permeability. Since distention is known to increase permeability temporarily, we studied patients with interstitial cystitis immediately after distention. RESULTS: Neither sugar was absorbed from intact rabbit bladders, while each was absorbed from acetone rinsed bladders at 10, 20 and 30 minutes. We used 100 ml. of solution in the initial 8 humans, including 6 with interstitial cystitis and 2 controls. At 30 minutes each sugar was absorbed from interstitial cystitis bladders but neither was absorbed from intact bladders. The test solution was then changed to 5% rhamnose. Mean rhamnose absorption plus or minus standard deviation was much greater in the 6 patients with interstitial cystitis than in 8 controls (26.3 + or - 26.1 versus 0.78 + or - 0.87 nmol. /ml. serum, p = 0.008). With 1 exception interstitial cystitis serum levels were at least 4-fold higher than the highest control level. CONCLUSIONS: This new permeability test clearly distinguishes intact versus distended bladders. It may be performed to test whether bladder permeability is increased in interstitial cystitis.

Characterization of the rabbit homolog of human MUC1 glycoprotein isolated from bladder by affinity chromatography on immobilized jacalin.

The urinary bladder is lined by transitional epithelium, the glycocalyx on the luminal surface has interesting properties and is implicated in protective functions. Glycoconjugates are major components of the glycocalyx, but their biochemical nature is not well understood. Previous studies on rabbit bladder indicated the presence of significant levels of sialoglycoproteins compared to glycosaminoglycans in the epithelium. In this study, rabbit explant cultures were radiolabeled by precursor sugars or amino acids and a major lectin-reactive glycoprotein of rabbit bladder mucosa was isolated by affinity chromatography on jacalin-agarose. The radiolabeled glycoprotein was purified to homogeneity by a second cycle on the lectin column, followed by gel filtration and density gradient centrifugation. The average molecular mass of the glycoprotein was estimated to be 245 kDa and 210 kDa by gel filtration and SDS-PAGE, respectively. Its buoyant density was 1.40 g/ml, suggesting a carbohydrate content of approximately 50%. The percent distribution of glucosamine-derived tritium label in sialic acid, galactosamine, and glucosamine was 30, 52, and 18, respectively. The glycoprotein consisted entirely of small sialylated and neutral oligosaccharides O-glycosidically linked to serine and threonine residues. The same glycoprotein could be immunoprecipitated with an antibody against the carboxy terminal 17 amino acid peptide of human MUC1 mucin glycoprotein. This suggests that this mucin glycoprotein is the rabbit homolog of MUC1 glycoprotein, which has been previously established to be a component of human bladder urothelium and has been purified from human urine and biochemically characterized.

The central domain of bovine submaxillary mucin consists of over 50 tandem repeats of 329 amino acids. Chromosomal localization of the BSM1 gene and relations to ovine and porcine counterparts.

We previously elucidated five distinct protein domains (I-V) for bovine submaxillary mucin, which is encoded by two genes, BSM1 and BSM2. Using Southern blot analysis, genomic cloning and sequencing of the BSM1 gene, we now show that the central domain (V) consists of approximately 55 tandem repeats of 329 amino acids and that domains III-V are encoded by a 58.4-kb exon, the largest exon known for all genes to date. The BSM1 gene was mapped by fluorescence in situ hybridization to the proximal half of chromosome 5 at bands q2. 2-q2.3. The amino-acid sequence of six tandem repeats (two full and four partial) were found to have only 92-94% identities. We propose that the variability in the amino-acid sequences of the mucin tandem repeat is important for generating the combinatorial library of saccharides that are necessary for the protective function of mucins. The deduced peptide sequences of the central domain match those determined from the purified bovine submaxillary mucin and also show 68-94% identity to published peptide sequences of ovine submaxillary mucin. This indicates that the core protein of ovine submaxillary mucin is closely related to that of bovine submaxillary mucin and contains similar tandem repeats in the central domain. In contrast, the central domain of porcine submaxillary mucin is reported to consist of 81-amino-acid tandem repeats. However, both bovine submaxillary mucin and porcine submaxillary mucin contain similar N-terminal and C-terminal domains and the corresponding genes are in the conserved linkage regions of the respective genomes.

Lectin histochemical examination of rabbit bladder glycoproteins and characterization of a mucin isolated from the bladder mucosa.

The glycocalyx of the mucosal surface of urinary bladder acts as an effective barrier against invasion by pathogenic microorganisms and injury from toxic substances in the urine. Defects in these bladder mucosal components could thus be important factors in the development of diseases such as interstitial cystitis and lower urinary tract infections. However, information on the nature of glycoconjugates of mammalian bladder mucosa is very limited. In this study, the glycoconjugates of rabbit bladder were examined histochemically using biotinylated lectins with specificities for a variety of carbohydrate moieties. Three [Artocarpus integrifolia (Jacalin), Datura stramonium (DSL), and Maackia amurensis II (MAL-II)] of the lectins bound predominantly to the luminal cell layer, with decreased binding to the basal layers of the epithelium. In contrast, Ricinus communis I and Sambucus nigra lectins did not bind to the cells in the epithelium but strongly interacted with the subepithelial layers, especially the lamina propria. The intensity of the staining by Jacalin and MAL-II was significantly reduced by prior treatment of the bladder sections with O-sialoglycoprotein endopeptidase, indicating that the ligands of these lectins are primarily mucin glycoproteins. In parallel biochemical studies, a high-molecular-weight glycoprotein with characteristics typical of epithelial mucins was purified from the mucosa of rabbit bladder explant cultures metabolically labeled with [(3)H]glucosamine. Quantitative analysis of the sialic acid, uronic acid, and hexosamine contents of delipidated rabbit bladder mucosa revealed a larger proportion of sialoglycoproteins compared with glycosaminoglycans. Taken together, the results of histochemical and biochemical analyses indicate that glycoproteins rather than glycosaminoglycans are the major components of the bladder epithelium, and that the former include a mucin.

Molecular basis of the polydispersity of mucins: implications for the generation of saccharide diversity.

Secreted epithelial mucins are large macromolecules which exhibit extreme polydispersity, the molecular basis of which is not fully understood. We have obtained partial sequences of two genes (BSM1 and BSM2) coding for two distinct molecules. This is the first time that such closely-related genes have been identified for any mucin from an animal. We propose that a combination of multiple homologous genes, alternative splicing, differential glycosylation, and additional post-translational processing all contribute to the extreme polydispersity of mucins. The multiple domain structure and non-identical tandem repeats are also very important for the generation of the saccharide diversities of mucins.

Endogenous carbohydrate-binding proteins of rabbit and human bladder.

OBJECTIVES: To identify the endogenous lectins of the human bladder with the long-term goal of developing improved strategies for the treatment of interstitial cystitis and other bladder disorders. METHODS: Rabbit and human bladder sections were examined histochemically using biotinylated neoglycoconjugates. Affinity chromatography of extracts of rabbit bladder was performed on immobilized lactose to purify the galactose-binding protein. RESULTS: Biotinylated beta-D-galactose neoglycoconjugate showed the strongest specific staining of the rabbit and human bladder sections. The beta-D-N-acetylglucosamine neoglycoconjugate also showed significant staining; the alpha-L-fucose, alpha-D-mannose, alpha-D-N-acetylneuraminic acid, and alpha-D-N-acetylgalactosamine neoglycoconjugates showed either very weak or no reaction. The strong Ca2+ -independent binding of beta-D-galactose neoglycoconjugates suggested the presence of galectins in rabbit and human bladder. Affinity chromatography of rabbit bladder extract on lactose gel yielded a galectin of about 30 kDa, consistent with the molecular biological data confirming the expression of galectin-3 in bladder. CONCLUSIONS: Beta-D-galactose binds strongly and specifically to rabbit and human bladder tissue sections. This information would be useful for the purpose of modifying drugs used for the treatment of bladder disorders with ligands of galactose-binding lectins to improve their retention in the bladder.

Signature sequences for the galectin-4 subfamily.

Galectins are a distinct family of animal lectins that have a cation-independent affinity for beta-galactoside sugars and share characteristic amino acid sequences. The cDNA encoding rabbit bladder galectin-4 has been cloned and sequenced (GenBank accession no. AF091738). The deduced 328 amino acid sequence predicts a multidomain structure consisting of an N-terminal peptide (19 residues) and two carbohydrate recognition domains (130 residues each) connected by a linker region (49 residues). Comparison of rabbit galectin-4 with related proteins reveals that two peptide motifs, M-A-F/Y-V-P-A-P-G-Y-Q-P-T-Y-N-P-T-L-P-Y in the N terminus and A-F-H-F-N-P-R-F-D-G-W-D-K-V-V-F in the first carbohydrate recognition domain are highly conserved in human, pig, rat, and mouse galectin-4 as well as in mouse galectin-6. The two peptide motifs are proposed here as the signature sequences to identify new members of the galectin-4 subfamily.

Four distinct regions in the auxiliary domain of heterogeneous nuclear ribonucleoprotein C-related proteins.

A 306 amino acid sequence deduced from a rabbit bladder cDNA is 98% identical to the human heterogeneous nuclear ribonucleoprotein (hnRNP) C2. The sequence comparison of the hnRNP C-related proteins reveals four distinct regions in the C-terminal auxiliary domain. The region next to the N-terminal RNA-binding domain is variable in length. Following the variable region, a basic region and a leucine zipper are conserved in all hnRNP C-related proteins including mouse and human Raly. Several Lys-Ser-Gly repeats are present in the basic region of the hnRNP C proteins. The C-terminal region is more divergent between hnRNP C and Raly. Signature sequences and possible functions are proposed for the different regions of the hnRNP C proteins.

Sequence of a second gene encoding bovine submaxillary mucin: implication for mucin heterogeneity and cloning.

Secreted epithelial mucins are extremely large and heterogeneous glycoproteins. We report the 5 kilobase DNA sequence of a second gene, BSM2, which encodes bovine submaxillary mucin. The determined nucleotide and deduced amino acid sequences of BSM2 are 95.2% and 92. 2% identical, respectively, to those of the previously described BSM1 gene isolated from the same cow. Further, the five predicted protein domains of the two genes are 100%, 94%, 93%, 77%, and 88% identical. Based on the above results, we propose that expression of multiple homologous core proteins from a single animal is a factor in generating diversity of saccharides in mucins and in providing resistance of the molecules to proteolysis. In addition, this work raises several important issues in mucin cloning such as assembling sequences from seemingly overlapping clones and deducing consensus sequences for nearly identical tandem repeats.

Increased urinary hyaluronic acid and interstitial cystitis.

PURPOSE: We compared urinary levels of hyaluronic acid in patients who met the National Institute for Diabetes, and Digestive and Kidney Diseases criteria for interstitial cystitis and in age matched healthy female controls. MATERIALS AND METHODS: Urinary hyaluronic acid was measured by solid phase radiometric assay using hyaluronic acid binding protein. Hyaluronic acid and symptom scores were compared in interstitial cystitis patients who gave multiple urine samples during treatment. Since hyaluronic acid changed with treatment in some patients, 17 samples from untreated interstitial cystitis patients were selected and compared with 17 control samples. RESULTS: Mean plus or minus standard deviation urinary hyaluronic acid concentrations were similar in the 2 groups (interstitial cystitis group 574 +/- 496, controls 512 +/- 324 ng./ml., p = 0.77). When normalized to creatinine urinary hyaluronic acid was significantly higher in interstitial cystitis patients (interstitial cystitis group 674 +/- 220, controls 446 +/- 220 ng./mg. creatinine, p = 0.0019). Urinary creatinine concentrations did not differ significantly (interstitial cystitis group 842 +/- 715, controls 1,162 +/- 516 mg./l., p = 0.12). CONCLUSIONS: Urinary hyaluronic acid was higher in interstitial cystitis patients than healthy controls. Since bladder hyaluronic acid is below the epithelium, this finding may indicate leakage across the epithelium into the urine in interstitial cystitis patients.

Identification of the glycosidically bound sialic acid in mucin glycoproteins that reacts as "free sialic acid" in the Warren assay.

A widely employed colorimetric assay for sialic acids based on periodate oxidation followed by reaction with thiobarbituric acid depends on the formation of a hexos-5-uluronic acid product, the pre-chromogen, by the periodate cleavage of the C6-C7, C7-C8, and C8-C9 bonds in free sialic acid. Glycosidically bound sialic acids are not expected to react in the assay since cleavage cannot occur between C6-C7 to yield the pre-chromogen. However, several investigators have reported the detection of a positive reaction by certain sialoglycoconjugates. In this study, it was found that various mucins but not other classes of sialoglycoconjugates or asialomucins exhibited this phenomenon. Of the mucins tested, ovine submaxillary mucin showed the maximum reactivity followed by the bovine and porcine counterparts. The disaccharide Neu5Acalpha2-->6 GalNAc(OH) released from mucins by alkaline borohydride treatment also reacted, albeit weakly compared to the native mucins, but other sialyl saccharides including 6'-sialyllactose and 6'-sialyl N -acetyllactosamine did not react. The positive reaction of the submaxillary mucins is not due to the presence of 3-deoxy-d-glycero-d-galacto-2-nonulosonic acid (KDN), a minor component in submaxillary mucins, or the release of sialic acid by the acidic condition of the assay. It is demonstrated that sialyl residues linked alpha2-->6 to unsubstituted N -acetylgalactosamine (sialyl Tn antigen structure) in mucin glycoproteins is responsible for the positive reaction. Apparently, periodate oxidation of the N -acetylgalactosamine residue leads to the release of sialic acid from the Neu5Acalpha2-->6 GalNAc linked to serine/threonine by an acid-catalyzed beta-elimination reaction. The findings provide a basis for the development of a chemical method to estimate sialyl Tn epitopes associated with cancer cells.

Histone H2A.F/Z subfamily: the smallest member and the signature sequence.

The nucleotide sequence of a 700 basepair cDNA obtained from rabbit bladder was determined. It encodes a 123 amino acid protein, which is the smallest member of histone H2A.F/Z subfamily. The known H2A.F/Z variants are highly conserved in their central core regions of about 100 amino acids but more divergent in their N- and C-terminal ends. In addition to the seven amino acid signature sequence previously known for the H2A proteins, all the cloned H2A.F/Z variants contain an identical peptide sequence, L-E-Y-L-T-A-E-V-L-E-L-A-G-N-A. This 15 amino acid motif is proposed here as the signature sequence to identify new members of the H2A.F/Z subfamily.

Bovine submaxillary mucin contains multiple domains and tandemly repeated non-identical sequences.

A number of cDNA fragments coding for bovine submaxillary mucin (BSM) were cloned, and the nucleotide sequence of the largest clone, BSM421, was determined. Two peptide sequences determined from the purified apoBSM were found near the N-terminus of the mucin-coding region of BSM421. This clone does not contain a start or stop codon, but its 3' end overlaps with the 5' end of a previously isolated clone, lambdaBSM10. The composite sequence of 1589 amino acid residues consists of five distinct protein domains, which are numbered from the C-terminus. The cysteine-rich domain I can be further divided into a von Willebrand factor type C repeat and a cystine knot. Domains III and V consist of similar repeated peptide sequences with an average of 47 residues. Domains II and IV do not contain such sequences but are similar to domains III and V in being rich in serine and threonine, many of which are predicted to be potential O-glycosylation sites. Domain III also contains two sequences that match the ATP/GTP-binding site motif A (P-loop). Only beta-strands and no alpha-helices are predicted for the partial deduced amino acid sequence. Northern analysis of submaxillary gland RNA with the BSM421 probe detected multiple messages of BSM with sizes from 1.1 to over 10 kb. The tandemly repeated, non-identical peptide sequences of approx. 47 residues in domains III and V of BSM differ from the tandemly repeated, identical 81-residue sequences of pig submaxillary mucin (PSM), although both BSM and PSM contain similar C-terminal domains. In contrast, two peptide sequences of ovine submaxillary mucin are highly similar (86% and 65% identical respectively) to the corresponding sequences in domain V of BSM.

Purification and characterization of the MUC1 mucin-type glycoprotein, epitectin, from human urine: structures of the major oligosaccharide alditols.

The MUC1 glycoprotein, epitectin, a component of the human bladder epithelium, was purified from human urine. Sedimentation equilibrium analysis and gel filtration using polysaccharide or protein standards revealed a polydisperse preparation with molecular weights ranging from about 0.9 to 1.3 x 10(6). This suggests that in the native state epitectin exists as aggregates of three or four monomer units of 350-400 kDa. Epitectin was found to have significant affinity to hexyl-, octyl- or phenyl agarose indicating that hydrophobic interactions and possibly carbohydrate-carbohydrate interactions may be responsible for the self-association. Chemical and enzymic deglycosylation of [125I]-labeled urine epitectin and metabolically labeled H.Ep.2 epitectin resulted in extremely polydisperse products. The buoyant densities of epitectin purified from urine and H.Ep.2 cells were found to be 1.39-1.40 g ml(-1), suggesting that the total carbohydrate content of these preparations is not significantly different. The O-linked saccharides of epitectin were fractionated by HPLC and analyzed by permethylation and FAB-MS. The neutral saccharides from both sources contain three common structures, namely Gal1 --> 3GalNAc, GlcNAc1 --> 6 (Gal1 --> 3) GalNAc and Gal1 --> 4GlcNAc --> 6 (Gal1 --> 3)GalNAc. The sialic acid of urine epitectin consisted entirely of N-acetylneuraminic acid. The two sources of epitectin, in vitro labeled on sialic acid, were found to have the same sialyl oligosaccharides but in different proportions. Metabolic labeling and N-glycanase susceptibility experiments firmly established the presence of N-linked saccharides in epitectin as minor components. The remarkable similarities in the total carbohydrate content, the carbohydrate composition and structures of saccharides between epitectin from urine, a non-malignant source, and H.Ep.2 cells is surprising in view of the prevailing view that MUC1 glycoproteins of cancer cells are underglycosylated compared to those produced by non-malignant cells.

An investigation of the nature of bladder mucosal glycoconjugates and their role in interstitial cystitis.

The long-term objective of this study is to elucidate the role of bladder mucosal glycosaminoglycans and mucin glycoproteins in the development of interstitial cystitis and other bladder diseases. Bladder biopsies and urine samples from patients and healthy controls were analyzed for glycoconjugates by biochemical and immunochemical methods. Due to the limited availability of human bladders for research purposes, detailed analysis of rabbit bladders glycoconjugates were also carried out. Biochemical analysis of rabbit bladders indicate that while the major portion of the glycoconjugates in the urothelium is sialoglycoprotein, low levels of heparan sulfate and chondroitin sulfate are also present. The correlating immunohistochemical data show very weak staining of the rabbit bladder epithelium by antiglycosaminoglycan antibodies. In contrast, the lamina propria and muscle layers stained intensely for chondroitin sulfate and hyaluronic acid. Thus, the quantity of glycosaminoglycans associated with the bladder epithelial layer, particularly as extracellular matrix components on the luminal surface of the bladder, appears insignificant. On the other hand, several lectins and anti-epitectin (a MUC1 sialoglycoprotein) antibodies showed strong staining of the luminal surface of rabbit and normal human bladders. Further, preliminary results with anti-epitectin antibodies reveal a weaker and patchy staining of biopsy specimens from interstitial cystitis patients compared to controls. The urinary levels of glycosaminoglycans and epitectin, in interstitial cystitis patients and healthy controls were determined by chemical or immunoassays. Urinary epitectin, but not glycosaminoglycans, was decreased in interstitial cystitis patients.

Urinary chondroitin sulfates, heparan sulfate and total sulfated glycosaminoglycans in interstitial cystitis.

PURPOSE: We compared urinary glycosaminoglycan levels in patients with interstitial cystitis and healthy controls. MATERIALS AND METHODS: Total sulfated glycosaminoglycans assayed by dimethylmethylene blue binding and individual glycosaminoglycans analyzed by cellulose acetate electrophoresis were compared in patients with interstitial cystitis and healthy controls. Also, multiple urine samples were obtained from healthy female controls for 2 months to assess the relationship of urinary glycosaminoglycan and creatinine concentrations, and to determine whether glycosaminoglycan excretion changes during the menstrual cycle. RESULTS: Total sulfated glycosaminoglycan and creatinine concentrations correlated well in random voided samples. Menstrual cycle day did not affect total sulfated glycosaminoglycan levels. Cellulose acetate electrophoresis revealed 3 bands corresponding to chondroitin sulfates, heparan sulfate and acidic glycoprotein. Patients with interstitial cystitis had decreased urinary concentrations of each of these individual components and total sulfated glycosaminoglycans. However, glycosaminoglycan-to-creatinine ratios were similar in interstitial cystitis and control urine. CONCLUSIONS: Using these assays total and individual urinary glycosaminoglycan levels normalized to creatinine were not altered in interstitial cystitis.

Urinary epitectin (MUC-1 glycoprotein) in the menstrual cycle and interstitial cystitis.

PURPOSE: We compared interstitial cystitis and control urine specimens for epitectin (MUC-1 glycoprotein), an epithelial mucin. MATERIALS AND METHODS: Urinary epitectin was measured in 28 patients with interstitial cystitis and 26 healthy controls. Ten controls provided multiple urine samples to determine whether urinary epitectin changes with the menstrual cycle. RESULTS: Epitectin levels were stable throughout the menstrual cycle. Interstitial cystitis cases had decreased urinary epitectin-to-creatinine ratios (mean 3.89 versus 6.38 micrograms./mg. creatinine for controls, p = 0.0035) and epitectin concentrations (mean 1.96 versus 4.30 micrograms./ml., respectively, p = 0.0005). CONCLUSIONS: Decreased mean urinary epitectin levels may reflect a cause (epithelial mucin deficiency) or a consequence of interstitial cystitis.

A novel monoclonal antibody specific for sialylated MUC1 mucin.

Development of a new monoclonal antibody (mAb) MY.1E12 which reacts with sialylated MUC1 mucins is described. The mAb did not react with any component in the lysates of COS-1 cells, whereas it bound to sialylated MUC1 mucins produced by COS-1 cells transiently transfected with MUC1 mucin cDNA, strongly suggesting that the expression of the epitope of mAb MY.1E12 depends on the presence of the MUC1 mucin core peptide. The requirement of sialyl residues for antibody recognition was established by Western blotting analysis of extracts of various carcinoma cells and in situ desialylation. In all cases, the mAb binding of electrophoretically separated MUC1 mucin diminished after desialylation by mild acid hydrolysis. When Capan-1 pancreatic carcinoma cells were pretreated with benzyl-N-acetylgalactosaminide in culture, the MUC1 mucins produced under these conditions, which were detected by core peptide-specific mAbs, did not react with mAb MY.1E12. These results suggest that O-linked carbohydrate chains are important for the mAb binding.