Kidney androgen-regulated protein interacts with cyclophilin B and reduces cyclosporine A-mediated toxicity in proximal tubule cells.

The gene for kidney androgen-regulated protein (KAP) is the most abundant and specific gene expressed in mouse kidney proximal tubule cells, where it is tightly regulated by steroid and thyroid hormones in different tubule segments. Despite the cell-specific expression, strict regulatory mechanisms, and relative abundance, nothing is known of the function of its encoded protein, which does not exhibit known structural or functional domains, or homologies with other sequences in the data bases. We raised monoclonal antibodies against KAP, which specifically recognize a protein with an apparent molecular mass of 20 kDa in crude kidney homogenates, the distribution and regulation of which parallel that of its mRNA. To gain insight into its function, we performed a yeast two hybrid screen and determined that KAP specifically interacts with cyclophilin B. Furthermore, cyclosporine A (CsA)-treated mice exhibited a significant decrease in KAP levels, and tetracycline-controlled overexpression of KAP in stably transfected proximal tubule cells significantly decreased the toxic effects of CsA. Taken together, these results indicate a functional relationship among KAP-, cyclophilin B-, and CsA-mediated nephrotoxicity and suggest an important role of KAP in renal physiology, providing new data on the molecular mechanisms implied in the toxic effects of CsA.

Control mice received vehicle alone (95% olive oil, 5% ethanol). After treatment, animals were euthanized by cervical dislocation. Several tissues were collected and immediately frozen in liquid N 2 or embedded in OCT compound and frozen in cold 2-methylbutane.
RNA Extraction and Northern Blot Analysis-Total RNA was extracted from different tissues using guanidium thiocyanate-acid phenol (13) followed by single-round mRNA extraction (Amersham Pharmacia Biotech). mRNA was fractionated by electrophoresis through agaroseformaldehyde gels and transferred to ZetaProbe membranes (Bio-Rad). Membranes were hybridized at 42°C overnight with random primed [ 32 P]dCTP (Amersham Pharmacia Biotech)-labeled cDNA probes, washed following the membrane manufacturer's instructions, and exposed to Hyperfilm (Amersham Pharmacia Biotech). Where noted, band intensity was measured by densitometric scanning of the resultant autoradiograph using the Bio-Rad GS700 image densitometer and the Molecular Analyst 1.40 program.
RT-PCR and Southern Blotting-RT-PCRs were performed under linear conditions with respect to RNA input and the number of amplification cycles. PCRs were determined as linear for 18 -24 cycles. Primers for detecting KAP transcripts were as follows: upper primer, 5Ј-TTG CCT TAA CCC TAC TAA AGC-3Ј; lower primer, 5Ј-GGA AGT AGG GGA GAC TGG-3Ј. ␤-Actin was amplified as a control of RNA amount and integrity. Amplification products were separated on 2% agarose gel and transferred to ZetaProbe membranes (Bio-Rad). The blots were probed with specific [␣-32 P]ATP 5Ј end-labeled primer corresponding to an internal sequence of the amplified product. Hybridization, washes, and exposure were performed as above.
pKAP1 and pKAP2 ELISA-For hybridoma screening, pKAP1 and pKAP2 were conjugated with BSA using N-succinimidyl-3-(2-pyridyldithio)propionate (Pierce) as linker agent. Maxisorp Nunc immunoplates were coated with 2% BSA in PBS, pH 8.3, and incubated for 4 h at room temperature. Plates were washed three times. 100 l of SPDP diluted at 10 g/ml PBS, pH 8.3, were incubated at room temperature for 1 h. Plates were again washed three times. 50 l of freshly diluted peptides were incubated overnight at room temperature. After three washing cycles, plates were blocked 2 h with 5% BSA in PBS, pH 8.3. Culture supernatants were incubated for 2 h, and bound antibodies were detected with rabbit anti-mouse Igs alkaline phosphatase conjugate (DAKO A/S, Copenhagen, Denmark). Immune sera and culture media were used as positive and negative controls, respectively. Hybridomas showing a positive reaction in pKAP1 or pKAP2-SPDP-BSA ELISA and a negative reaction against PBS-SPDP-BSA ELISA were selected. Twenty-three hybridomas were found to be specific for pKAP1 and 65 for pKAP2. Anti-pKAP1 hybridomas 377-1B8 and 377-7D10 and anti-pKAP2 hybridomas 388-1D7, 3881E4, 388-1F2, 388-6F8, 388-7H9, and 388-10D1 showed high reactivity against these peptides and were selected for cloning. Once cloned, hybridomas were cultured in 225-cm 2 flasks and left to grow to saturation. Antibodies were purified from spent culture media with HiTrap protein a columns (Amersham Pharmacia Biotech) following the manufacturer's instructions.
Monoclonal Antibody Characterization-Antibody isotyping was performed by ELISA with monoclonal antibody-based Ig isotyping kit (PharMingen, San Diego, CA). 388-10D1 was found to be of IgM,k isotype. All the other hybridomas secreted IgG1,k. Hybridoma specificity was analyzed by testing its reactivity against four synthetic peptides of unrelated proteins: Sequence 759-770 of ␣ v ␤ 5 integrin and sequences 1-15, 344-367, and 369-385 of hementin. These four peptides were coupled to BSA using SPDP as linker and following the methodology as described for pKAP1 and pKAP2 ELISA. 377-1B8 and 377-7D10 were positive against pKAP1 and negative against pKAP2 and control peptides. Anti-KAP2 hybridomas did not react with pKAP1 and control peptides. The monoclonal antibodies produced by 377-1B8 and 388-1D7 were named abKAP1 and abKAP2, respectively, and were used to identify KAP in mouse tissues.
Western Blot Analysis-Tissues were homogenized by N 2 cavitation in RIPA buffer (DOC 0.5%, Nonidet P-40 1%, SDS 0.1%, and protease inhibitors in 1ϫ PBS). For Western blot analysis, samples were normalized for protein concentration using the Bradford assay (Bio-Rad), adjusted for equal protein levels, and separated on 15% SDS-polyacrylamide gel electrophoresis under reducing conditions. Proteins were transferred to a PVDF (Schleicher & Schuell) membrane, and blots were blocked overnight at 4°C in 5% nonfat dried milk in PBS. Primary monoclonal antibodies (abKAP1 and abKAP2) were diluted at 5 g/ml in blocking buffer, and membranes were probed for 2 h. Washes were performed following the membrane manufacturer's instructions and secondary antibody (horseradish peroxidase-conjugated rabbit antimouse, DAKO A/S) diluted 1:5000 and incubated for 1 h at room temperature. After washing, bands were detected using the ECLϩ chemiluminescence detection method (Amersham Pharmacia Biotech) and exposed to Hyperfilm.
Immunohistochemistry-Thin kidney sections (5 m) were obtained in a cryostat (Bright Instrument Company Ltd., Huntingdon, United Kingdom). Samples were fixed in cold acetone for 10 min and washed twice in PBS. Peroxidase activity was blocked for 10 min in 0.3% H 2 O 2 and washed twice in PBS. Sections were blocked in normal rabbit serum 1:10 in PBS for 10 min, and primary antibody was diluted in PBS at 20 g/ml. After 1 h of primary antibody incubation, 2-min washes in PBS were performed three times. Secondary antibody (horseradish peroxidase-conjugated rabbit anti-mouse, DAKO A/S) diluted 1:200 in PBS was incubated for 30 min followed by extensive washing in PBS. Finally, the sections were incubated for 5 min with the diaminobenzidine substrate solution. Nuclei were stained with hematoxylin-eosin solution, and slides were dehydrated and mounted in DPX mounting medium (Agar Scientific Ltd., Essex, United Kingdom).
Yeast Two Hybrid Library Screening-The pBD-KAP yeast two hybrid expression vector was constructed by PCR amplification of KAP cDNA with specific primers (KUP2Hyb, 5Ј-TAT GTC GAC GCA TGA TGC TTT TCA AGG-3Ј, and KLOW2Hyb, 5Ј-TAG TCG ACG TCA GGA AGT AGG GGA GAC TGG-3Ј) and ligation into the SalI restriction endonuclease site of pBD-GAL4 (Stratagene, La Jolla, CA). Male mouse kidney cDNA library (HybriZap TM two hybrid library, Stratagene) was cotransformed with pBD-KAP into YRG-2-competent yeast cells following the manufacturer's instructions. A total of 6 ϫ 10 3 putative interacting clones were identified by growth in selective media (Leu Ϫ , Trp Ϫ , His Ϫ ). From them, 10 clones were positive when screened for ␤-gal expression. Plasmids from these clones were purified and cotransformed again with pBD-KAP and with control plasmids in order to confirm the interaction. pBD-P53 and pAD-SV40 expression vectors coding for proteins P53 and large T antigen of SV40 were used as a positive control for interaction.
Interacting Domains Identification-Fragments of the KAP coding sequence KAP1, KAP2, and KAP3 were generated by PCR amplification and ligation into SalI-digested pBD-GAL4. CyPB1, CyPB2, CyPB3 and CyPB4 fragments were also generated by PCR amplification and ligation into EcoRI-XhoI-digested pAD-GAL4. The primers used in the amplification are shown in Table I. Cloned fragments were in frame and without mutations as determined by automatic sequencing (ABI Prism 310, Applied Biosystems, Foster City, CA). Interaction assays were performed as described above. Liquid ␤-gal assays were performed as previously reported.
Expression and Purification of Recombinant Proteins-JM109 bacterial cells were transformed with constructs expressing GST and GST-KAP. Luria broth (100 ml) containing 50 g/ml ampicillin was inoculated with 1 ml of bacteria and incubated in an orbital shaker at 37°C. Bacteria were grown to A 600 ϭ 0.5, induced with 0.1 mM isopropyl-1thio-␤-D-galactopyranoside, and shaken for 3 h at 37°C. The GSTsoluble protein was purified with B-PER TM reagent (Pierce). The recombinant GST-KAP aggregates into inclusion bodies. Therefore, the insoluble protein was solubilized with 6 M guanidine HCl and refolded with DSB 201 as previously reported (57).
Soluble lysate was conjugated with glutathione-Sepharose 4B (Amersham Pharmacia Biotech) overnight at 4°C. Beads were washed twice with PBS, and protein concentrations were assessed after electrophoresis on 10% polyacrylamide denaturing gels by Coomassie Blue.
GST Pull-down Assays-For pull-down experiments, ϳ1 g of GST alone or 1 g of GST-KAP fusion protein bound to Sepharose 4B (Amersham Pharmacia Biotech) was used in each sample incubation. Total volume was adjusted to 150 l with GST buffer (5 mM MgCl 2 , 150 mM KCl, 50 mM Tris-Cl, pH 7.8, 0.5 mM EDTA, 10% glycerol, 0.1% Triton X-100, 0.1% Nonidet P-40, and protease inhibitors). Beads and in vitro translated proteins were incubated overnight at 4°C and then centrifuged and washed four times in GST buffer. Beads were resuspended in SDS loading buffer to a final volume of 150 l and boiled for 5 min. Proteins were electrophoresed on SDS-containing 15% polyacrylamide gels at 40 mA, along with standard protein molecular weight markers. Gels containing [ 35 S]methionine-labeled proteins were dried and exposed to Hyperfilm (Amersham Pharmacia Biotech) overnight.
Cell Culture and Tetracycline-regulated KAP-expressing System Establishment-Proximal convoluted tubule cells PKSV-PCT were cultured as described previously (14 -16). Transfections were performed using LipofectAMINE (Life Technologies Ltd.) according to the manufacturer's instructions. Expression of GFP and GFP-KAP was achieved by transient transfection of pEGFP-3 and pEGFP-3-KAP (CLONTECH, Palo Alto, CA) constructs into PKSV-PCT cells. After transfection, cells were trypsinized and seeded for immunocytochemistry assays. In order to obtain a regulated expression system for KAP, we used the Tet-Off TM gene expression system (CLONTECH). PKSV-PCT cells were stably transfected with the regulator plasmid pTet-Off and selected according to their low background and high induction levels. The more suitable clone was selected for a second transfection with pTRE-KAP and pTK-Hyg expression plasmids, and clones resistant to hygromycin (400 g/ ml) were selected. All of them were assayed for KAP mRNA and protein expression, and clones presenting the highest induction levels, tight tetracycline regulation, and low background were chosen for further studies.
Immunocytochemistry-Cells were grown on glass slides, and 24 h later they were washed in cold PBS twice and fixed in cold 4% paraformaldehyde for 1 h followed by three 10-min PBS washes. Aldehyde groups were blocked with 50 mM NH 2 Cl for 30 min and washed for 5 min in PBS four times. Cells were permeabilized with 0.1% saponin, 1% BSA in PBS for 30 min. Slides were incubated for 90 min at room temperature with the primary antibody diluted to 20 g/ml in permeabilization buffer. Upon washing, cells were incubated for 1 h at room temperature with the secondary antibody diluted 1:200. After four rinses in PBS, nuclei were stained with TO-PRO3 (Molecular Bioprobes, Eugene, OR) following the manufacturer's instructions, and fluorescence labeling was visualized in a Leica DM IRBE confocal microscope.
Cytotoxicity Assay-Tet-Off TM selected clones were grown with or without tetracycline on 96-well plates and treated with increasing doses of CsA (0, 6.25, 9.375, 12.5, 18.75, 25, 37.5, and 50 g/ml) for 24 h. Twelve independent wells were assayed for each CsA concentration. Cell death was quantified using a cytotoxicity detection kit (lactate dehydrogenase) (Roche Diagnostics GmbH).
Statistical Analysis-Two-way analysis of variance was applied to investigate whether the cell death induced by CsA treatment was affected by the expression of KAP. Planned contrasts were used to test for differences between Tcϩ and Tc-for each clone, at different CsA concentrations. The analysis was carried out with the SAS 6.11 statistical package (SAS Institute, Cary, NC) using a general linear model (PROC GLM), which allows for the occurrence of unbalanced designs.

RESULTS
Tissue-specific Distribution of KAP mRNA and Identification of Its Encoded Protein-Although tissue specificity of the KAP mRNA has been previously reported (17), the more sensitive RT-PCR/Southern technique was used, and distribution of KAP mRNA was performed in a wider panel of tissues. Results shown in Fig. 1A confirm that KAP mRNA is exclusively expressed in the kidney and uterus of pregnant female mice from day 13 of gestation. Although not included in the panel, and confirming previous results (7), we also observed that KAP mRNA becomes undetectable in mouse uterus right after birth.
Computational analysis of the KAP (GenBank TM accession number M22810) deduced peptide sequence defines a protein of 121 amino acids in length, hydrophilic and negatively charged (2). IgG1 monoclonal antibodies abKAP1 and abKAP2 were raised against two different epitopes of KAP. They both recognized a protein with an apparent molecular mass of 20 kDa in SDS-PAGE from mouse crude kidney extracts, which was undetectable in liver and lung (Fig. 1B) or in kidney when peptide-specific preabsorbed antibodies were used (not shown). The antibodies were also able to recognize the in vitro translated KAP that migrates with an apparent size of 19 kDa in SDS-PAGE (Fig. 1B, panel 1) and the expected 48-kDa recombinant GST-KAP, heterologously produced in Escherichia coli (Fig. 1B, panel 2).
Immunohistochemistry assays in frozen kidney sections identified KAP in epithelial cells of proximal convoluted tubules, which correlates with KAP mRNA site synthesis (Fig.  1C). Antibodies preabsorbed with their corresponding antigenic peptides failed to recognize the KAP target protein in kidney, thereby proving the specificity of the immune reaction (Fig. 1C).
The Kidney Androgen-regulated Protein Resembles Its mRNA Cell Distribution and Androgen Regulation-To test whether the cell distribution and abundance of KAP are also androgen-dependent, Western blot and immunohistochemistry assays were performed in kidneys of intact male, female, and castrated male mice. Results in Fig. 2A demonstrate that the relative amounts of KAP correlate with previously reported relative mRNA levels (2), being much more abundant in males than in females and exhibiting the lowest levels in castrated males. Like its mRNA, KAP is expressed in S3 cells, but it is only expressed in the S1/S2 segments of intact males (Fig. 2B).
The Kidney Androgen-regulated Protein Specifically Interacts with Cyclophilin B (CyPB)-The full-length KAP cDNA was screened against a male mouse kidney cDNA library in a yeast two hybrid assay. One of the clones was effectively  interacting with KAP to support the permissive growth of yeast cells in selective medium and to produce positive results in ␤-gal assays (Fig. 3A, section 1). This clone was identified as the mouse peptidyl-prolyl-cis-trans-isomerase B (CyPB, GenBank TM accession number M60456). Positive and negative interacting control proteins were also included (  Fig. 3B demonstrate that the binding capacity of CyPB for GST-KAP depends on the KAP moiety, because GST alone is unable to bind CyPB. Similarly, KAP binds specifically to CyPB because the Oatp1-translated product does not get bound to KAP, thereby demonstrating the specificity of the interaction. When CsA (10 and 25 M) was added to the interacting proteins, no competitive effect of the immunosuppressor was observed when CyPB was preincubated with CsA or when the immunosuppressor was added to the KAP-CyPB complex (data not shown). Finally, co-immunoprecipitation assays were performed on kidney crude extracts as an in vivo demonstration of the interaction (Fig. 3C). Although we were able to recover CyPB from the abKAP1 and abKAP2 immunoprecipitated products (Fig.  3C, lanes 1 and 3, respectively), the reverse experiment, i.e. immunoprecipitating with the antibody against CyPB and de- FIG. 1. mRNA tissue distribution and protein detection of KAP. A, RT-PCR/Southern blotting was performed on a wide tissue panel, including prostate, bone, muscle, heart, lung, cerebellum, brain, liver, submaxillary gland, testis, and kidney. ␤-Actin was used as an internal control. KAP mRNA was only present in kidney. Similarly, RT-PCR/Southern blotting was performed on uterus and placenta at different days of pregnancy. np, nonpregnant. B, the monoclonal antibodies abKAP1 and abKAP2 specifically recognized KAP in Western blot assays from male mouse tissue extracts (kidney, liver, and lung), in vitro translation product (panel 1) and recombinant GST-KAP (panel 2). C, immunohistochemistry assays in male mouse frozen kidney, 5-m sections, using antibodies abKAP1 and abKAP2. Preabsorbing the antibodies with a 10-fold excess of each antigenic peptide proved the specificity of the reactions. Original magnification, ϫ 200. Within the kidney, the protein was found in the S3 segment in male, female, and castrated male mice. Only intact males presented expression in the S1/S2 segment. tecting with abKAP1 or abKAP2, gave negative results (not shown).
KAP and CyPB Co-location in Kidney Culture Cells-Since its identification, CyPB has been known to be part of the secretory pathway, being present in the membrane fraction (plasma membrane, endoplasmic reticulum, Golgi, and microsomes) (18). GFP alone and GFP-KAP fusion construct were transfected into PKSV-PCT cells. The GFP-KAP fusion protein was distributed around the nucleus of proximal tubule cells following a reticular pattern consistent with an endoplasmic reticulum location, clearly different from the distribution observed for GFP alone (Fig. 4A). Immunocytochemistry using anti-CyPB antibody performed on GFP-KAP-transfected cells showed co-location of both proteins in intracellular organelles, whereas the plasma membrane remained positive for CyPB staining only (Fig. 4B). Cells transfected with a GFP-KAPexpressing plasmid presented reduced endogenous CyPB staining at the plasma membrane level compared with nontransfected cells (see arrows in Fig. 4B).
Interacting Domains between KAP and CyPB Proteins-To determine the domains that participate in KAP and CyPB interaction, several KAP cDNA truncations encompassing nucleotides 1-194, 1-321, and 1-437 of the KAP open reading frame were screened against the entire coding sequence of the CyPB cDNA (Fig. 5A). Co-transformed yeast cells carrying the KAP 1-321 and KAP 1-437 constructs were able to grow in selective medium, whereas the KAP 1-194 construct was not (Fig. 5B). These results indicate that the CyPB binding capacity exhibited by KAP must be located between base pairs 194 and 321 of the cDNA, i.e. in a region of ϳ40 amino acids in length, in the middle of the protein. The reverse experiment was also performed and different CyPB overlapping fragments, such as those from base pairs 28 -231, 28 -515, 231-664, and 515-664 of the cDNA, tested against the full KAP open reading frame. As shown in Fig. 4B, only fragments CyPB 231-664 and CyPB 515-664 were positive. Because CyPB 28 -515 was negative, the region encompassing base pairs 551-664, which codes for 37 amino acids of the C-terminal domain, is the one responsible for KAP interaction. These results were further confirmed by the ␤-gal liquid assays (Fig. 5C).
Effects of CsA Treatment on Kidney KAP Expression: Correlation between the mRNA and the Protein Levels-CsA binds to CyPB and modulates the expression of several genes (19), some of which, but not all, are implicated in the immune response. Because treatment with CsA increases CyPB plasma levels (20) and CyPB has been found to interact with KAP, we aimed to observe the effects of acute CsA treatment on KAP kidney expression under different hormonal conditions. Female, male, and castrated male mice were treated for 3 days with CsA at 15 mg/kg/day by intramuscular injection and compared with vehicle-treated mice. Results shown in Fig. 5A demonstrate a significant 2.2-and 3.6-fold increase in KAP/ glyceraldehyde-3-phosphate dehydrogenase ratios in CsAtreated males and females, respectively, and no effect on KAP mRNA levels in castrated males (Fig. 6A). Western blot analysis performed in crude extracts of counterpart kidneys revealed a significant decrease in KAP in castrated males and females that was not apparent in males (Fig. 6B). The hypothesis that CsA treatment would affect KAP expression preferentially in the S3 segment was proved by immunohistochemistry analysis in frozen kidney sections of CsA-treated and nontreated mice. Fig. 6C demonstrates the lower and almost nonexistent KAP levels in females and castrated males, respectively, and the higher decrease in the protein in S3 cells than in S2 and S1 cells in males.

Production of a Stably Transfected and Tetracycline-controlled KAP Expression System in PKSV-PCT Cells as a Model for Studying the Effects of KAP Expression on CsA-mediated
Toxicity-KAP expression vectors were stably transfected in the PKSV-PCT cell line, originally derived from proximal tubules of transgenic mice carrying the large T and little t SV40 antigens, under the control of 5Ј regulatory sequences of the L-type pyruvate kinase gene that had proved to conserve the main features of the parental cells from which it was derived (14 -16). This cell line was used to express KAP in a controlled and inducible manner using the Tet-Off TM system from CLONTECH.
Results in Fig. 7A are representative of KAP mRNA levels in pTet-off regulator plasmid stably transfected PKSV-PCT cells (clone 3-26) and in three independent stably expressing tet-regulated pTRE-KAP clones (3-26-37, 3-26-71, and 3-26-310). Endogenous cyclophilin A (CyPA) expression levels were determined as a control for the amount and integrity of mRNA. Whereas results in Fig. 7A indicate that selected clones have the ability to induce KAP mRNA in a tetracyclinedependent controlled manner, those shown in Fig. 7B demonstrate their ability to express KAP in around 3-4 h upon tetracycline removal. Indirect immunofluorescence analysis of expressed KAP shows a cytoplasmic reticular distribution in these cells.
The above-described pTRE-KAP clones were used to assess the effects of KAP expression on CsA-mediated toxicity in proximal convoluted tubule cells, using clone 3-26 as a control. Cells were exposed to increasing doses of CsA (from 0 to 50 g/ml) for a period of 24 h, and culture medium from each situation was assayed for lactate dehydrogenase activity, as a marker of cell death. Results depicted in Fig. 7D for clone 3-26 show the cytotoxicity levels exhibited by these cells at increasing CsA concentrations, which were not modified by tetracycline. Contrarily, in the three independently generated pTRE-KAP clones, tetracycline removal significantly reduced the toxic effects exerted by CsA (see Fig. 7D). These data are representative of results obtained in three independent experiments and clearly demonstrate that KAP expression protein reduces CsA toxicity in proximal tubule cells. DISCUSSION Since the first description of KAP mRNA in 1979 (1), the identification and function of its encoded protein have remained elusive. This paper reports, for the first time, KAP identification and distribution and makes a direct attempt to gain insight into its function.
The tissue and cell specificity previously reported for KAP mRNA (3-6, 9, 17), were further analyzed by means of the more sensitive RT-PCR/Southern technique in a wider panel of tissues. Kidney specificity and unique expression of KAP in the uterus during a short period of time prior to birth, described earlier by Kasik et al. (7), were definitively assessed in our experiments. No functional data are available to explain the appearance of KAP mRNA in the uterus before delivery, which, for the time being, is associated with the estrogen sensitivity of the gene (6) and the estrogenic peak that occurs at that time in this tissue (7).
Monoclonal antibodies raised against KAP synthetic peptides specifically identified a protein with an apparent molecular mass of 20 kDa in SDS-PAGE that follows the same distribution and androgenic regulation as its encoding mRNA in epithelial cells of proximal convoluted tubules. Although the expected molecular mass for KAP (121 amino acids in length) is around 13 kDa, computational analysis of the KAP-deduced peptide sequence defined a hydrophilic and negatively charged protein (3), which might explain the delay observed in SDS-PAGE.
As our data show KAP to be a highly specific and tightly regulated kidney protein, we hypothesized that it might play an important role in the homeostatic and metabolic events of proximal tubule cells. This notion was further supported by the finding that KAP expression is up-regulated in nephrectomized mouse kidney (21) and that its mRNA levels significantly decrease in a mouse nephrolithiasis model (22). Because no functional or structural domains were identified on the KAP-deduced peptide sequence, we focused our efforts on the identification of proteins able to interact with KAP in vivo, which might in turn be informative of KAP function.
The major finding of the present study was that KAP interacts with cyclophilin B, a member of the immunophilin family that exhibits petidyl-prolyl-cis-trans-isomerase activity and the ability to bind the potent immunosuppressor CsA (23,24). The interaction, initially observed by means of the two hybrid assay, was further confirmed by GST pull-down assays and by co-immunoprecipitation of the KAP-CyPB complex from crude kidney extracts. We found these data very interesting because the great clinical benefits of CsA on the improvement of graft survival rates in solid organ transplantation are concomitant with important undesirable nephrotoxic effects (25). The immunosuppressive effects of CsA are known to be related to binding of the immunophilins-CsA complex to the calcium/ calmodulin-dependent serine/threonine phosphatase calcineurin, which blocks, in turn, its intrinsic phosphatase activity in vitro (26). As a result, activation and translocation of the nuclear factor of activated T cells are compromised, nuclear factor of activated T cells-supported transcription of the interleukin 2 gene and other cytokine genes is abolished, and T-cell activation is inhibited. FK-506 (tacrolimus), a second T-cell immunosuppressive drug, binds to a different cellular protein, FKBP12, and the FKBP12-FK 506 complex also binds to and inhibits calcineurin (27).
Whereas the molecular mechanisms involved in immunosuppression are well known, those involved in nephrotoxicity are less understood. CsA exerts its nephrotoxic effects through (i) acute changes in renal hemodynamics (28 -32) followed by irreversible striped interstitial fibrosis (33)(34)(35)(36), and (ii) cytotoxicity in proximal tubule cells (37)(38)(39)(40). In vitro model systems have revealed a site-selective action of the cells of the proximal tubular region of the nephron (41, 42) preferentially in epithe-lial cells of the S3 segment (43). Recent reports have concluded that CsA exerts a direct toxic effect on proximal cells by reducing DNA synthesis and cell cycle blockade (44), which are coincidental with elevated p53 levels (45). In an experimental model of chronic CsA nephrotoxicity, an increase in apoptotic specific genes has been found, and it has been proposed that increased apoptosis could explain the tubular dropout and loss of cellularity with fibrosis (46).
Apart from these studies, the molecular mechanisms of CsAinduced toxicity in the kidney have not been completely established, nor has it been determined why CsA exerts its deleterious effects on the proximal tubule and preferentially on cells of the S3 segment. The cloning of a third mammalian cyclophilin, CyPC, which in contrast to the widely distributed CyPA and CyPB family members was restricted to the kidney (47) and preferentially expressed in proximal and straight tubules of the kidney (48), suggested the possibility that CyPC could be a mediator of the immunosuppressive and nephrotoxic actions of CsA, but no functional data have emerged supporting the initial association between CyPC restricted expression and CsA nephrotoxicity.
Identification of the biologically active cyclophilins to promote T cell activation was performed in Jurkat T cells by overexpression of different cyclophilin family members cotransfected with a phosphatase reporter plasmid containing multiple copies of the nuclear factor of activated T cells or nuclear factor-interleukin 2A binding sites from the interleukin 2 enhancer (49). Both CyPA and CyPB but not CyPC increase T cell sensitivity to CsA, demonstrating that CyPA and CyPB are the active immunophilins able to mediate the inhibitory effects of CsA in vivo (49). In addition to their ability to bind calcineurin, the subcellular location of each immunophilin was shown to be important for mediating inhibition of signal transduction by CsA. In this regard, deletions of CyPB and CyPC domains, which change their intracellular locations, altered their activities significantly. Although both immunophilins are located in the ER, CyPB is associated with a specialized part of the ER related to calcium activation events, which permits its activity (50). It is interesting to note that the carboxyl-terminal 37 residues of CyPB were able to confer a significant gain in function on CyPC and that the converse construction lowered CyPB activity (49). This region of CyPB contains the previously characterized location signal, which could direct a foreign protein to the specialized CyPB retention site (the calcium-containing ER substructure or calciosome) (50). CyPC exclusion from the calciosome may prevent it from participating in inhibition of T-cell activation. The finding that CyPB interacts with a protein involved in calcium signaling in T cells named CAML (51) reinforces the notion that CyPB might be involved in calcineurin activation through calcium mobilization events.
Involvement of CyPA and/or CyPB in the nephrotoxic effects of CsA has not been assessed. Although a cyclophilin A knockout mouse has recently been produced, nothing is known on toxicity in these animals, 3 and the question as to which cyclophilin is relevant in nephrotoxicity remains open. Because FK506 exhibits the same toxic effects on kidney as CsA and the only common functional mechanism between both immunosuppressors is calcineurin inhibition, it is accepted that nephrotoxicity will also occur through the inhibition of this calcium calmodulin-dependent phosphatase. Furthermore, chemical modifications of the CsA molecule aiming at eliminating its toxicity while preserving immunosuppressive effects have failed, suggesting that both actions occur through common 3 J. Luban, personal communication. mechanisms (52). Taking into account the wide tissue and cell distribution of cyclophilins, the finding that a kidney-specific protein of the proximal tubule interacts with CyPB (the CyP that shows the highest affinity for CsA (53) and the highest expression levels in kidney (54)) was suggestive of a putative new molecular pathway of CsA-mediated toxicity in this tissue. Co-immunoprecipitation experiments and co-location within the cell demonstrate that both proteins reside in the same subcellular compartment. Definition of KAP and CyPB interacting domains by means of two hybrid assays determined that A, RT-PCR assay with total RNA from selected clones. Cyclophilin A (CypA) was used as an internal control. Cells were seeded in the presence of tetracycline (0.01 g/ml). After extensive washing, cells were left to grow on media with or without Tc for 24 h; RNA was then obtained, and RT-PCR was performed. B, induction timing for KAP expression. Cells were grown in 0.01 g/ml of Tc. After complete Tc removal, cultures were assayed for protein expression at different time points (from 0 to 24 h) by Western blot assays with abKAP1. The protein was detectable ϳ3-4 h post-derepression. C, immunocytochemistry performed on clone PCT 3-26-310 with antibody abKAP1. Staining was present in the cytoplasm, displaying a reticular pattern. Original magnification, ϫ 630. D, cytotoxicity assays. Lactate dehydrogenase activity in culture media from 24 h CsA-treated control Tet-Off TM clone PCT 3-26 and KAP-expressing clones PCT 3-26-37, PCT 3-26-71, and PCT 3-26-310 was measured to evaluate cell death. *, p Ͻ 0.001. the C-terminal domain of CyPB is responsible for interaction. This result was further confirmed in the co-immunoprecipitation assays in which we failed to detect KAP when the immunoprecipitation was performed with an antibody that specifically recognizes the C-terminal end of CyPB. As mentioned previously, this is the domain responsible for CyPB sublocation to the calciosome in the ER and the domain that is highly conserved in CyPBs from different species and not present in other cyclophilins. Interaction of KAP and CyPB was not prevented by CsA, which concurs with the fact that domains for CsA binding and peptidyl-prolyl-cis-trans-isomerase (PPiase) activity are in the middle of the protein (55), and that KAP-CyPB interaction takes place at the C-terminal domain.
Apart from being located in the endoplasmic reticulum, CyPB has also been found in the plasma membrane and secreted to the medium (56). Confocal microscopy performed in cultured proximal tubule cells in this study demonstrated that endogenous CyPB in these cells is located in the expected cellular compartments, i.e. the ER and the plasma membrane. Interestingly, we observed that for cells transiently transfected with a GFP-KAP fusion protein expression vector, the overlay of KAP and CyPB shows a perfect co-location for both proteins, and that in the presence of higher KAP levels in these transfected cells, endogenous CyPB is located intracellularly rather than in the plasma membrane. This observation suggests that by interacting with CyPB, KAP might contribute to retaining this cyclophilin in the ER compartment with possible implications on CyPB function, particularly if we take into consideration that CsA promotes the secretion of CyPB out of the cell (56). In a different set of experiments, we clearly demonstrated that CsA significantly diminishes KAP levels in crude kidney extracts, preferentially in the most sensitive site of CsA toxicity, the S3 segment of the proximal tubules. We do not yet know whether CsA promotes KAP degradation or its secretion out of the cell as it does for CyPB, but the fact is that KAP decreases in the kidney in the presence of CsA. If we hypothesize that in physiological conditions KAP contributes to CyPB being retained in the ER, suboptimal KAP levels upon CsA treatment might prompt the secretion of CyPB out of the cell, thereby lowering the CyPB reservoir in the reticulum, which might impair calcium mobilization and, finally, calcineurin activation.
With this in mind, we aimed at observing the effect of KAP overexpression on CsA-mediated toxicity in PKSV-PCT cells. It was remarkable to observe that in three independently stably transfected clones, selected by their ability to induce KAP expression in a tetracycline-dependent manner, CsA-induced toxicity was significantly diminished when KAP expression was permitted. It is important to point out that because the cells were stably rather than transiently transfected, the levels of KAP expression were not much higher than the basal levels, and consequently, the effect of KAP expression was highly relevant in diminishing toxicity. Although overexpression and knocking out of the KAP gene in whole animals will be required to demonstrate that KAP may protect against CsA toxicity in vivo, the experiments presented in this paper are significant because they relate the presence of KAP to CsA toxicity in a system isolated from the multiple and complex interactions that occur in the kidney upon treatment with the immunosuppressive drug. This cell system will be a valuable model for gaining further insight into the physiological role of KAP in the kidney and its relevance in CsA-mediated toxicity in proximal tubule cells. The concept that KAP could be involved in processes that end in calcium mobilization will be studied further, and it will be of great interest to determine whether the upregulation that KAP mRNA exhibits in the uterus of pregnant females is involved in calcium mobilization, as this is an important step in the excitability of uterine smooth muscle cells during delivery.