Attachment of newly formed crystals to renal tubular epithelial cells appears to be a critical step in the development of kidney stones. The present study was undertaken to identify autocrine factors released from renal epithelial cells into the culture medium that inhibit adhesion of calcium oxalate crystals to the cell surface. A 39-kDa glycoprotein that is constitutively secreted by renal cells was purified by gel filtration chromatography. Amino acid microsequencing revealed that it is novel and not structurally related to known inhibitors of calcium oxalate crystallization. Hence, it was named crystal adhesion inhibitor, or CAI. Immunoreactive CAI was detected in diverse rat tissues, including kidney, heart, pancreas, liver, and testis. Immunohistochemistry revealed that CAI is present in the renal cell cytosol and is also on the plasma membrane. Importantly, CAI is present in normal human urine, from which it can be purified using calcium oxalate monohydrate crystal affinity chromatography. CAI could be an important defense against crystal attachment to tubular cells and the subsequent development of renal stones in vivo.
- calcium oxalate monohydrate
- cell-crystal interaction
- DING protein
nephrolithiasis is an extremely common condition in the United States, affecting up to 10% of the population at some point during their lives (26, 32). Although many affected individuals have identifiable urinary metabolic risk factors, such as excessively concentrated urine that may contain too much calcium, uric acid, or oxalate, or perhaps too little citrate, many do not (12). Therefore, the concentration of these urinary constituents does not appear to fully explain the formation of renal stones. In addition, nucleation and growth of individual crystals appear unlikely to produce particles large enough to occlude the nephron lumen in vivo (15). Recent evidence suggests, that in many calcium oxalate stone formers, the earliest changes may be depositions of calcium phosphate in the medullary interstitium, which then serve as a nidus for a calcium oxalate stone (14). The processes that mediate calcium phosphate deposition and its evolution into calcium oxalate stones remain to be determined. In more marked hyperoxaluric states (e.g., enteric or primary hyperoxaluria), direct adhesion of calcium oxalate crystals to renal epithelial cells may predominate (14). Therefore, our laboratory (19–24) and others (29, 30, 37, 38, 43) have sought to identify regulatory mechanism(s) by which urinary calcium oxalate and calcium phosphate crystals in tubular fluid bind to renal epithelial cells, are retained in the kidney, and become a nidus for stone formation.
Previous studies have demonstrated that anionic molecules in tubular fluid can coat calcium oxalate monohydrate (COM) and hydroxyapatite (HA) crystals, making them less likely to bind to renal cells in culture (20). Effective anions include the small molecule citrate as well as the glycoprotein osteopontin (20). During our studies, we unexpectedly observed that renal epithelial cells in culture produce a factor that markedly inhibits adhesion of exogenous COM crystals to their surface. In this report, we identify and characterize this novel factor.
MATERIALS AND METHODS
Nontransformed African green monkey renal epithelial cells of the BSC-1 line (36) were grown in Dulbecco-Vogt modified Eagle's medium containing 25 mM glucose (DMEM) at 38°C in a CO2 incubator, as previously described (24). To prepare high-density, quiescent BSC-1 cultures, 1 × 106 cells/35-mm plastic plate (9.62 cm2, Nunc, Naperville, IL) were plated in DMEM containing 1% calf serum and 1.6 μM biotin. Three days later, when they were confluent, the medium was aspirated and replaced with fresh medium containing 1% calf serum and 1.6 μM biotin. Cells were plated in 1% calf serum, and 3 days later (day 6) the medium was changed to DMEM containing 0.01% calf serum. BSC-1 cells were routinely used for study on day 9 after being plated. Madin-Darby canine kidney (MDCK) type I cells were a generous gift of Carl Verkoelen (Erasmus University, Rotterdam, The Netherlands). To prepare high-density, quiescent MDCKI cultures, 1 × 106 cells/35-mm plastic plate (9.62 cm2, Nunc) were plated in DMEM containing 10% calf serum and 1.6 μM biotin. Two days later, when they were confluent, the medium was aspirated and replaced with fresh medium containing 5% calf serum and 1.6 μM biotin. The monolayer was used for study the next day. Rat inner medullary collecting duct (cIMCD) cells were a generous gift of Jack Kleinman (Medical College of Wisconsin, Milwaukee, WI). To prepare confluent cIMCD cultures, 1 × 106 cells/35-mm plastic plate (9.62 cm2, Nunc) were plated in DMEM containing 1% calf serum and 1.6 μM biotin. Two days later, when they were confluent, the medium was aspirated and replaced with fresh medium containing 1% calf serum and 1.6 μM biotin. The monolayer was used for study the next day.
Crystal binding assay.
To quantitate crystal binding, 200 μg/ml of [14C]COM crystals were added to confluent monolayers of cells whose medium had been replaced by medium conditioned (CM) by the cells for different amounts of time, or with fresh DMEM or PBS. Two minutes later, monolayers were washed three times with PBS, and the cell-associated radioactivity was counted, as described previously (19).
Preparation of antibodies against crystal adhesion inhibitor.
After its isolation, crystal adhesion inhibitor (CAI) was subjected to amino acid microsequencing. The amino acid sequence information was employed to prepare synthetic peptides for use as antigens to generate polyclonal antibodies to CAI. The potential antigenicity (hydrophilicity) of the different peptide fragments was compared by using computer software programs based on known algorithms to choose the optimal antigenic sequence for synthesis (MacVector, Acelrys, San Diego, CA). The sequence of the NH2-terminal 16 amino acids and that of a 13-amino acid internal fragment were each used to prepare a synthetic peptide at the University of Chicago Protein-Peptide Core Facility (NH2-terminal 16 amino acids: NH2-Lys-Ile-Asn-Gly-Gly-Gly-Ala-Thr-Leu-Pro-Gln-Pro-Leu-Tyr-Gln-Thr-COOH; 13 amino acid-peptide internal fragment: NH2-Leu-Asn-Asn-Asp-Tyr-Ser-Gln-Phe-Gly-Thr-Gly-Thr-Lys-COOH). Each peptide was linked to a branched polylysine backbone, thereby increasing its multiplicity in a “multiple antigenic peptide system” (MAPS) (27). These “MAPS-peptides” were used as distinct antigens to immunize rabbits (Pocono Rabbit Farm and Laboratory, Canadenesis, PA). Antibody 1 was prepared against the 16-mer peptide, antibody 2 against the 13-mer. Each antibody reacted strongly against the MAPS-protein antigen that elicited it when evaluated in an ELISA.
The MAPS-peptides were coupled to ECH Sepharose 4B beads (Pharmacia) in carbodiimide at pH 4.5 during an overnight incubation (13). MAPS-antigen columns were prepared and equilibrated with 10 mM Tris, pH 7.5. A solution of the immune serum diluted 1:10 in 10 mM Tris (pH 7.5) was passed over the column three times. The columns were washed with 10 mM Tris (pH 7.5), then 10 mM Tris (pH 7.5) plus 0.5 M NaCl, and finally eluted with 100 mM glycine (pH 2.5) (10).
MDCKI cells or cIMCD cells grown to confluence on coverslips were each rinsed twice with PBS and fixed using 3.7% formaldehyde in PBS for 10 min. Fixed cells were permeabilized with 0.1% Triton X-100, or were not permeabilized, and then processed. Cells were then rinsed three times with PBS (5 min each) and incubated with 1% goat serum for 1 h to block nonspecific binding sites. The goat serum was aspirated, and the cells were incubated overnight with polyclonal anti-CAI antibody and monoclonal anti-actin (BD Transduction Labs, Franklin Lakes, NJ) or anti-nucleolin antibodies (MBL International, Woburn, MA) at 4°C in a humidified chamber. Next, cells were washed with PBS (5 min × 3) and then incubated for 1 h with an anti-mouse IgG conjugated with Alexa-488 and anti-rabbit IgG conjugated with Texas red (BD Transduction Labs). The cells were then washed with PBS (10 min × 3), and the coverslips were mounted onto glass slides using Slow-Fade (Molecular Probes, Eugene, OR). The XY and XZ sections were scanned using a LSM 510 confocal microscope (Carl Zeiss, Oberkochen, Germany) equipped with an Axiovert 100M microscope and a c-Apochromat 63/1.2 numerical aperture water-immersion lens. Alexa 488 and Texas red were excited with the 488- and 568-nm line, respectively, from an argon-krypton laser. Emissions for Alexa 488 and Texas red were collected at 505–550 nm and >585 nm, respectively.
Preparation of cellular extracts.
Cell protein extracts were prepared by rinsing a monolayer of BSC-1 or MDCKI cells with PBS containing PMSF (1 mM) and aprotinin (0.15 U/ml) and then lysing the cells in 2% SDS containing 5% β-mercaptoethanol and 1 mM PMSF. Tissue blocks (∼1 mm3) were also obtained from specified organs of normal male Sprague-Dawley rats (∼250 g) at the time of death, placed in 1 ml of the cell lysis buffer, and sheared with an 18-gauge needle (20 times) after being placed in buffer.
Preparation of cellular fractions.
Monolayers of BSC-1 cells were rinsed with ice-cold PBS and scraped into a centrifuge tube containing 10 ml of homogenization buffer (0.3 M sucrose, 0.05 M Tris, pH 7.4) with Complete protease inhibitor cocktail (Boehringer Mannheim). Cells were homogenized by passage through an 18-gauge needle (20 times) and centrifuged twice in succession (600 g × 10 min) to obtain a nuclear fraction. The supernatants were combined and centrifuged (10,000 g × 30 min) to remove the mitochondrial and lysosomal fractions. The supernatants were then centrifuged (100,000 g × 120 min) to prepare microsomal (pellet) and cytosolic fractions (supernatant).
SDS-PAGE and immunoblotting.
Concentrated fractions containing CAI activity were electrophoresed on a 1.5-mm-thick, 10% polyacrylamide gel at 100-V constant current using a Mini-Protean II apparatus (Bio-Rad, Hercules, CA). After electrophoresis, the gel was either stained using 0.1% Coomassie blue or transferred onto a Sequi-Blot membrane (Bio-Rad) in CAPS buffer (10 mM CAPS, 40% methanol, pH 11.0) at a 250-mA constant current for 60 min using a Trans-Blot apparatus (Bio-Rad). Blots were stained using Coomassie blue or probed with the monospecific antibodies raised against CAI MAPS-peptides, and the protein of interest was detected using a horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Piscataway, NJ).
Microsequencing of COM-binding protein.
In the University of Chicago Amino Acid and Protein Core, Rockefeller Institute Protein Core, and Mayo Clinic Protein Core, CAI was subjected to in situ enzymatic digestion with Asp N protease, Lys C protease, or trypsin. The resulting peptide fragments were separated by reverse-phase HPLC, collected directly onto a polyvinylidine difluoride strip, treated with biobrene in methanol, and then subjected to microsequencing (Applied Biosystems Procise 492 HT, Applied Biosystems, Foster City, CA). The data were analyzed with ABI model 610A data-analysis software (Applied Biosystems).
Cultured renal cells constitutively release CAI.
Binding of COM crystals to cells was markedly reduced in the presence of CM. [14C]COM crystal adhesion (2-min assay) was reduced by 95% when radioactive crystals were added to cultures containing medium that had been exposed to the cells for 3 days compared with monolayers in which the medium was replaced with fresh DMEM containing 0.01% calf serum or PBS (Fig. 1). To investigate this observation further, 3-day quiescent CM was replaced with PBS to which increasing quantities of CM were added. A concentration-dependent decline in crystal adhesion was observed as the amount of CM was increased (Fig. 2). Inhibition was maximal when >2.5 ml of CM were added. These initial observations suggested that the cells release a factor (CAI) into CM that blocks adhesion of COM crystals to cells.
To define the kinetics on the appearance of CAI activity, BSC-1 cells were grown to high density and made quiescent in fresh medium containing 0.01% calf serum. The cells were allowed to condition the medium for different amounts of time. Then, the CM was assayed for its capacity to inhibit adhesion of [14C]COM crystals to BSC-1 cells during the next 4 days. Significant CAI activity was detected on day 2 and was greater on day 3 (Fig. 3). Minimal additional release of CAI from the cells into the medium was apparent after 3 days.
Characterization of CAI.
CAI activity in CM was retained by ultrafiltration with a YM 30 membrane (Amicon) that allows passage of molecules with an Mr of <30,000 and is resistant to freezing, heating to 56°C for 30 min, or boiling for 10 min (41). Treatment with heparinase I (5 U/ml, 37°C, pH 7.4, 4 h) (25), heparinase III (0.5 U/ml, 43°C, pH 7.4, 4 h) (25), and chondroitinase ABC (0.5 U/ml, 37°C, pH 7.4, 4 h) (44) failed to abolish activity, as did treatment with DNAse (10 ng/ml, 23°C, pH 7.4, 1 h) or nitrous acid (0.25 M, 23°C, pH 7.4, 2 h) (11). Activity was abolished by treatment with neuraminidase (1 U/ml, pH 5.5, 37°C, 1 h) (40) or sodium hydroxide (0.2 M, 23°C, 18 h) (11). Utilization of the DIG glycan detection kit (Boehringer Mannheim), as directed by the manufacturer, suggested that the factor contained carbohydrate, as was inferred by a positive carbazole reaction (6). When applied to a DEAE-Sepharose anion-exchange column, CAI eluted with 0.4 M NaCl. Treatment with proteinase K did not abolish activity but did change the elution profile from the Biogel A 0.5M sizing column (see below). Together, these observations suggest that CAI is anionic, contains carbohydrate and possibly sialic acid residues, and has an Mr >30,000.
Purification of CAI from CM.
Utilizing the above information, a strategy was developed to purify CAI from CM. Medium from 3-day quiescent BSC-1 cultures containing 0.01% calf serum was collected and then passed through a 0.22-μm filter to remove any cells or debris. CM was then ultrafiltered through a YM 30 membrane (Amicon) to eliminate molecules with an apparent Mr of <30,000. The fraction containing molecules with Mr > 30,000 was loaded onto a DEAE-Sepharose anion-exchange column and was eluted with 0.4 M NaCl. The eluate was then applied to a Biogel A 0.5 M sizing column. Activity was detected in fractions consistent with an apparent Mr between 60,000 and 100,000, with peak activity at an apparent Mr of 80,000 (Fig. 4A). However, when the DEAE eluate was incubated with EDTA for 2 days at 4°C, CAI eluted from the Biogel column in a sharp peak with an apparent Mr of ∼15,000 (Fig. 4B). Therefore, it appears that CAI released from cells forms an aggregate when CM is concentrated, and the aggregate can be dissociated by chelation of its associated cations (e.g., calcium and magnesium) with EDTA. For its subsequent purification, CAI was chelated with EDTA before separation on the sizing column because the activity eluted in a sharper peak in the dissociated form (Fig. 4B).
To purify CAI further, we took advantage of our previous observations which indicated that molecules inhibiting COM crystal adhesion to cells act by coating the crystals (20). Therefore, an EDTA-treated DEAE eluate was loaded onto a Biogel column, and the fractions demonstrating greatest activity (Fig. 4B) were concentrated with a Centricon microconcentrator (10-kDa cutoff, Millipore, Bedford, MA) and incubated with COM crystals (100 mg) in PBS overnight at 4°C in a 2-ml tube subjected to end-over-end rotation. The next day, crystals were washed, in succession, with supersaturated calcium oxalate solution (2 ml) containing no NaCl, then 1 M NaCl, and finally 4 M NaCl (twice), each for 1 h. Antiadhesion activity was absent in the PBS and all wash fractions, suggesting that CAI was tightly associated with the crystals. COM crystals were then dissolved by incubation in 10 mM Tris (pH 8) containing 250 mM Na-EDTA. This solution was replaced daily and saved until the crystals were completely dissolved. After dissolution of the crystals, the pooled material was placed in Spectrapor dialysis tubing (12- to 14-kDa cutoff) and dialyzed at 4°C against 10 mM sodium phosphate buffer to remove EDTA, calcium, and oxalate. The volume of the dialyzed material was reduced with a Centricon microconcentrator (10-kDa cutoff). Material recovered after dissolution of the crystals demonstrated CAI activity in the crystal binding assay, produced a single band on an SDS-PAGE (20%) with an apparent Mr of 39 kDa (Fig. 5A), and was eluted from a reverse-phase HPLC C4 column as a single sharp peak at 51% acetonitrile (utilizing a 1:80% acetonitrile gradient containing 0.1% trifluoroacetic acid) (Fig. 5B). Material produced by dissolution of control crystals (not coated with CAI) did not exhibit CAI activity.
Therefore, CAI appears to be a hydrophobic glycoprotein having marked affinity for COM crystals and an apparent Mr of 39 kDa. As yet unidentified physical/chemical properties of CAI apparently cause it to migrate differently through a sizing column composed of agarose beads (apparent Mr ∼15 kDa) than in 20% polyacrylamide (apparent Mr ∼39 kDa).
Amino acid composition and sequence of CAI.
Pooled fractions containing CAI activity from the Biogel sizing column were combined, concentrated, and rinsed with 10 mM Tris (pH 7.4) using a Centricon-10 microconcentrator, and subjected to SDS-PAGE (20%) followed by electrophoretic transfer to a polyvinylidine difluoride membrane (Imobilon-PSQ, Millipore) in a transverse electric field of 20 V overnight in 10 mM CAPS buffer containing 10% methanol, pH 11. Blots were stained with Coomassie blue (0.1%), destained with 50% methanol and 10% acetic acid, and the predominant band at 39 kDa was excised. Amino acid compositional analysis, performed at the University of Chicago Amino Acid and Protein Core Lab, is shown in Table 1, in which CAI is compared with three other urinary crystallization inhibitors, uropontin (UP) (31), human urinary prothrombin fragment 1 (UPTF1) (33), and bikunin (34). Whereas the estimated total net charges on the protein core of UP (−25.5), UPTF1 (−14.3), and bikunin (−7.4) are each negative, CAI (−2.1) was nearly neutral. Therefore, unlike these three known anionic glycoproteins (UP, UPTF1, bikunin), carbohydrate (possibly sialic acid residues) rather than acidic amino acids appears to contribute importantly to the strongly anionic character of CAI glycoprotein. The role of sialic acid is inferred from the observation that treatment of partially purified CAI with neuraminidase abolished its CAI activity.
Electroblots of CAI were then submitted for amino acid microsequencing. Results for the NH2 terminus are shown in Table 2. To achieve more complete characterization of CAI, additional electroblots of CAI were prepared. The band of interest (Mr 39,000) was excised, the eluted protein was subjected to proteolytic cleavage, and the resulting fragments were separated by reverse-phase HPLC and then subjected to microsequencing at the Rockefeller Institute. LysC protease cleavage of electroblotted CAI resulted in the six fragments shown in Table 2. Microsequencing of additional fragments of CAI produced by cleavage with AspN protease, or by trypsin cleavage of CAI obtained from human urine processed identically to CM (see below), were performed in the Mayo Clinic Protein Core. Where overlap occurred, the sequences from human urine and monkey renal cells were identical (Table 2). In summary, the sequences for 283 amino acids of CAI fragments are now known. Because the complete amino acid sequences of UP, UPTF1, and bikunin are all known, it is clear that CAI differs from these three glycoprotein inhibitors of COM crystal growth. A search of peptide and nucleic acid sequence databases (2) revealed that CAI is novel but exhibits a homology of its NH2 terminus to the DING family of proteins (4) (Table 3).
Monospecific antibodies against fragments of CAI.
The amino acid sequence information in Table 2 was employed to prepare two synthetic MAPS-peptides for use as antigens to generate polyclonal antibodies to CAI in rabbits. Each antibody reacted strongly against the MAPS-peptide antigen that elicited it when evaluated in an ELISA. Each cognate MAPS-peptide was used as an antigen to prepare an affinity-purified antiserum from its respective serum as described in materials and methods. Each of these monospecific antibodies recognized a single protein band whose apparent size was Mr 39 kDa when proteins in partially purified CM from BSC-1 cells were separated by SDS-PAGE, and the gels were blotted and probed.
Detection of CAI in renal epithelial cells.
Cell protein extracts were prepared by rinsing a monolayer of BSC-1 cells with PBS containing PMSF (1 mM) and aprotinin (0.15 U/ml) and then lysing the cells in 2% SDS containing 5% β-mercaptoethanol and 1 mM PMSF. Total cell proteins were separated by SDS-PAGE, electroblotted, and probed with each of the monospecific antibodies against peptide fragments of CAI. Each antiserum recognized a band of Mr 39 kDa in the total BSC-1 cell extract, which appears to represent a cell-associated form of the protein (Fig. 6). CAI was also detected in a cell extract of a second renal epithelial cell line, MDCKI (not shown). As MDCKI cells are derived from the distal tubule, this observation suggests that the factor is produced in a nephron segment in which stone formation seems likely to begin.
BSC-1 cells were then fractionated to determine the organelle in which CAI originates or is stored. Antibody 1 detected CAI mainly in microsomes, whereas antibody 2 detected it in the cytosol (Fig. 6). Both antibodies also detected small amounts of CAI in the nuclear fraction. MDCKI cells were also examined immunohistochemically using antibody 2. In nonpermeabilized cells, small amounts of CAI were detected on the plasma membrane, perhaps representing ongoing secretion (not shown). In permeabilized cells, CAI appeared to be diffusely distributed in the cytosol, where it partially colocalized with actin (Fig. 7, A–C). Minimal nuclear staining was observed, whereby CAI partially colocalized with nucleolin (Fig. 7, D–F). Taken together, these observations suggest that CAI is a cytosolic protein that is secreted by the cells.
To determine the systemic distribution of CAI, the proteins in cellular extracts of diverse rat organs were probed with CAI antibody 1. A 39-kDa band was recognized in all tissues examined, including brain, testis, ureter, spleen, kidney, heart, and pancreas (Fig. 8). In addition, smaller bands (∼ 34 kDa) as well as larger bands (∼55 and 80 kDa) were recognized in certain tissues, including BSC-1 cell lysates, suggesting that CAI might exist in various isoforms. Note that no 15-kDa bands were detected.
CAI detected in human urine.
The results summarized above suggest that CAI is a cytosolic glycoprotein that is secreted by renal epithelial cells into the extracellular fluid, e.g., CM of cultured cells. To determine whether it is released into human urine by the intact kidney, we used the same protocol that was employed to purify CAI from the CM of BSC-1 cells.
One liter of human urine from a non-stone-forming adult volunteer was collected using thymol as a preservative to prevent bacterial growth in a protocol approved by the Mayo Clinic IRB. The urine was concentrated with a YM-30 membrane, applied to a DEAE-Sepharose anion-exchange column, and then eluted with 0.4 M NaCl. Sodium-EDTA (50 mM) was added to the eluate, and 2 days later it was applied to a Biogel A 0.5 M column. Fractions 40–45, which contained CAI activity were combined, and the proteins were separated by SDS-PAGE and electroblotted. When the blot was stained with Coomassie blue, a single prominent band was detected at 39 kDa (Fig. 9A). Each of the two monospecific antibodies generated against MAPS-peptides of CAI recognized this band (Fig. 9, B and C).
Because CAI has a strong affinity for COM crystals, they were employed to purify the factor from urine. A 10-fold concentrate of urinary proteins was prepared using an Amicon stir cell and YM-10 membrane. This concentrate (2 ml) was incubated with COM crystals (10 mg) overnight. The crystals were washed twice with PBS and dissolved in 0.5 M EDTA, pH 8. The crystal extract was concentrated and washed using a Centricon microconcentrator (10-kDa cutoff), and its proteins were resolved by SDS-PAGE and electroblotted. A 39-kDa band was recognized by CAI antibody 1 in the crystal extract. The band was then excised from the blot and subjected to NH2-terminal microsequencing in the Mayo Clinic Proteomics Core, which confirmed that it was CAI (see Table 3). Therefore, CAI in human urine, like that in BSC-1 cell CM, has a strong affinity for COM crystals.
The experiments described above demonstrate that monkey renal epithelial cells release a glycoprotein that blocks adhesion of COM crystals to the apical cell surface. This molecule has been named CAI, and it is also found in normal human urine. Renal epithelial cells in culture constitutively produce CAI, suggesting that cells lining the nephron may do so similarly in vivo. Molecular characterization suggests that CAI is a strongly anionic glycoprotein with an apparent Mr of 39 kDa. Microsequencing of the NH2 terminus and 10 fragments has identified a total of 283 amino acids, indicating that CAI is novel.
The NH2 terminus of CAI bears strong homology to four human proteins in the database (Table 3). These are members of the DING family of proteins, named for the characteristic DINGGG amino acid sequence at the NH2 terminus (4). Members of this family have been isolated from diverse tissues and were identified by their unique biological functions: affinity for the intracellular signaling molecule genestein (breast cancer cells) (3); cell surface cotinine receptor (brain cells) (28); affinity for hirudin with proteinase and growth factor activity isolated form fibroblast culture medium (1, 9); and as a T cell stimulator in synovial fluid (GenBank accession no. 1711419) (7). To date, no gene encoding any of these proteins has been reported (3). The NH2 terminus of CAI has limited homology to a peptide fragment obtained by dissolving a human kidney stone and separating the resulting matrix proteins by two-dimensional gel electrophoresis (GenBank accession no.1082909) (5). This “urinary stone matrix protein” (Table 4) may be a common constituent of human kidney stones because similar peptides were visualized by these investigators on gels prepared from the matrix of three additional urinary stones (5).
The NH2 termini of CAI and the DING family of proteins also exhibit significant homology to the predicted amino acid sequence of an ABC phosphate transport protein (17) in the recently sequenced Pseudomonas aeruginosa genome (GenBank accession no. 32040700; Table 3). Each of the 11 internal fragments of CAI also bear homology to this predicted protein, some more strongly than others (Table 3). A weaker homology of the CAI NH2 terminus to the predicted amino acid sequence of a membrane-bound form of alkaline phosphatase from P. aeruginosa (GenBank accession no. 15595885) (35) and a membrane-bound form of [Fe]-hydrogenase from Desulfovibrio desulficans (GenBank accession no. 1218062) (16) is also apparent in Table 4. Eight of the 10 internal fragments of CAI also bear weak homology to the predicted amino acid sequence of Pseudomonal alkaline phosphatase, one of a family of bacterial membrane-associated phosphate-binding proteins (18, 35). Together, the homologies in Tables 3 and 4 hint that CAI is a membrane-associated or cytosolic protein constitutively released by renal epithelial cells and that it, or a homologous DING protein, may also be released by other types of cells, such as breast epithelial cells, brain cells, fibroblasts, or synoviocytes. Although each fragment of CAI has some homology to the predicted sequence of the Pseudomonal ABC-type transporter, certain of the homologies are weaker (e.g., fragments 1, 4, 8, 10, and 11). We do not believe that CAI is a Pseudomonal contaminant for several reasons: CAI immunoreactivity was detected in both sterile culture medium and sterile human urine, and CAI was found in sterile lysates of cultured renal cells, in which bacterial contamination is highly unlikely. Importantly, nonrenal proteins homologous to CAI have been isolated by independent investigators from diverse tissues using different purification techniques. We believe that as a member of the DING family of proteins, CAI bears significant nucleotide and amino acid sequence homology to the Pseudomonas ABC phosphate transporter protein that couples ATP hydrolysis to ionic transport (8). The functional significance of these homologies remains to be determined.
As CAI is a constituent of human urine, it is likely that the glycoprotein is present in sufficient quantities in tubular fluid in vivo to coat crystals and prevent their adhesion to renal epithelial cells. In susceptible individuals, CAI could become incorporated into the matrix of kidney stones (5), a hypothesis that can now be tested using our antisera. Further study of CAI, and the factors that regulate its production, could increase our understanding of the mechanisms that underlie kidney stone formation. Because crystals adhere avidly to proliferating renal cells in culture (39), it is possible that CAI production in vivo could increase to protect cells against crystal binding during repair after tubular injury. In addition, in certain individuals, production of CAI glycoprotein that is either quantitatively or functionally defective could predispose them to kidney stone formation. Should CAI-deficient patients be detected, adhesion of crystals to tubular cells would be expected to occur more readily, and once retained in the nephron, these crystals could grow into kidney stones. Studies are ongoing to test this hypothesis and clone the gene that encodes CAI.
This work was supported by grants to J. C. Lieske from the National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK-53399 and R21-DK-60707), the Oxalosis and Hyperoxaluria Foundation, and the Mayo Foundation.
We thank G. Reddy of the University of Chicago Core, J. Fernandez of the Rockefellar Institute Protein Core, B. Madden of the Mayo Clinic Protein Core, and J. Tarara of the Mayo Clinic Optical Morphology Core Facility for technical assistance and valuable advice. We also thank R. Norris, H. Jiang, A. Zago, and E. Huang for technical assistance, M. Walsh-Reitz for valuable discussions, and Y. Nakagawa for preparation of reagents.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society