Biochemical Evidence for the Presence of Two Vitamin D-dependent Calcium-binding Proteins in Mouse Kidney*

Mouse kidney, a vitamin D target organ, was inves- tigated for the presence of vitamin D-dependent cal-cium-binding proteins (CaBP). Mouse kidney cytosol was fractionated by several biochemical methods including gel filtration chromatography, gel permeation high performance liquid chromatography, and chromatofocusing. Mouse kidney was found to possess two CaBPs which completely differed biochemically and exhibited no cross-immunoreactivity. One had a mblec- ular weight of 25,000 and a PI of 5.9. The other, with a molecular weight of 10,000 and a PI of 4.9, was biochem- ically identical with mouse duodenal 10,000 CaBP. In addition, mouse renal and duodenal 10,000 CaBPs were immunologically identical. Moreover, the 10,000 CaBP was the predominant CaBP in mouse kidney since the latter contained about twice as much 10,000 CaBP as 25,000 CaBP (in mol/mg of renal cytosolic protein). In vitro incorporation of [3H]leucine into renal 10,000 CaBP demonstrated that it is synthesized in situ by mouse kidney. Renal 10,000 CaBP was already present during fetal life, and reached its adult level during the first week after birth. The vitamin D dependency of both mouse renal 10,000 and 25,000 CaBPs was assessed by their decrease in vitamin D-deficient mice and sub-sequent rise after 1,25-dihydroxyvitamin DS injection. The concomitant

' The abbreviations used are: CaBP, calcium-binding protein; high performance liquid chromatography; RIA, radioimmunoassay. duodenum contains a CaBP with a molecular weight of Purification of duodenal 10,000 CaBP from rat and mouse allowed us to develop two specific RIAs for each of these two species (14,16). When studying the tissue distribution of 10,000 CaBP in these species, we found a CaBP identical with the duodenal one in the uterus, chorioallantoic placenta, and yolk sac of both rat and mouse (16,17). When we tested mouse kidney cytosol by RIA, we detected substantial amounts of immunoreactive material. This was surprising in view of the fact that rat and other mammalian kidneys contain little or no immunoreactive 10,000 CaBP (10,14,18,19). This prompted us to define biochemically the immunoreactive material from mouse kidney and to investigate its dependency on vitamin D. 9-10,000 (10)(11)(12)(13)(14)(15)(16).

Animals and Diets
For biochemical and immunological studies, normal Swiss mice, 6-8 weeks old, were obtained from CERJ, France and fed on a normal diet (UAR, France).
For vitamin D dependency studies, 69 normal weanling Swiss mice (CERJ) were randomized into 3 groups. All were raised in the dark on a vitamin D-free diet (containing 0.50% Ca and 0.36% P) for 5 weeks and then on a low-calcium vitamin D-free diet (0.03% Ca and 0.36% P) for 1 week. In one group, ZOO0 IU of vitamin Ddkg dry diet were added to the daily feed (vitamin D-repleted group). In another group, each mouse received a subcutaneous injection of 1 ng of 1,25(OH)2Da/g of body weight in 10 p1 of 95% ethanol 48 and 24 h before killing (vitamin D-deficient + 1,25(OH)2D3 group). A third group received one subcutaneous injection of 10 pl of solvent alone, 48 and 24 hours before killing (vitamin D-deficient group). Animals were fed ad libitum and were not fasted before killing. In each group of 23 mice, tissue samples from 16 were pooled, and samples from the remaining 7 were treated individually.
For developmental studies, normal pregnant Swiss mice were obtained at 15-16 days of gestation. In this strain, parturition occurred after 19 days of gestation. Litter size was reduced to 8-10 pups within 48 h of birth. Mothers were removed from cages 20 days after birth. On given days, pups were taken a t random from the litters and killed.

Preparation of Tissues
Mice were bled by decapitation, and the kidneys were removed, decapsulated, and weighed. They were minced and suspended (l:4, w/v) in ice-cold Tris-buffer (13.7 mM Tris-HC1, pH 7.4, containing 120 mM NaCl and 3 mM KCl), to which 4 mM phenylmethanesulfonyl fluoride and 1% aprotinin (Sigma) were added. They were homogenized in a Potter Elvehjem homogenizer and centrifuged a t 1 O O , x g for 1 h a t 4 "C. The supernatant without the fluffy lipid layer was taken as the cytosol fraction and kept frozen at -30 "C. Other mouse or chick tissues studied were similarly processed.
For the vitamin D dependency studies, the proximal 5 cm of the small intestine were excised in addition to the kidneys, rinsed in icecold Tris-buffer, and everted. The duodenal mucosa was scraped and treated as described above.
For the developmental studies, in addition to the kidneys, the entire intestine, from the pylorus to the ileocaecal junction, was dissected from the pancreas and peritoneum, opened along its whole length, and thoroughly rinsed. Kidneys were excised, and all the samples were frozen.

Biochemical a n d Immunochemical Methods
Gel Chromatography-Gel chromatography was performed with Sephadex G-75 (Pharmacia, France) in 0.1 M ammonium acetate buffer, pH 7.2, with 1 mM mercaptoethanol. The column (2.6 X 100 cm) was calibrated with 4 proteins of known molecular weight: bovine serum albumin (67,000), ovalbumin (45,000), chymotrypsinogen (25,000), and cytochrome (12,500) (all from Boehringer) detected by their UV absorbance at 280 nm. The void volume of the column was determined by blue dextran exclusion. The partition coefficient between the liquid and the gel phases (&") was plotted against the log (molecular weight) of the protein and yielded a straight line.
Gel Permeation-High Performance Liquid Chromatography-GP-HPLC was performed through a Waters 1125 gel-permeation column equipped with a Waters 6000 A solvent pump and a Waters U6 K injector. The elution buffer was 0.1 M Tris-acetate buffer, pH 7.0, containing 0.15 M NaCl and was run at 0.4 ml/min. Chromutofocuszng-The protein solution was fmst equilibrated with start buffer by gel filtration with Sephadex G-25 (Pharmacia). Chromatofocusing was then carried out using Polybuffer Exchanger 94 (Pharmacia), packed in a column (0.9 X 30 cm). The two pH gradients (7-4 and 8-5) were respectively produced with Polybuffer 74 a t pH 4.0 and with a mixture of Polybuffer 96 (30%) and Polybuffer 74 (70%) at pH 5. Collected fractions were monitored for pH with a Tacussel pH meter.
4'Ca-Chelex Method-The calcium-binding activities eluted from the different columns were detected by the 45Ca-Chelex method (12), using Chelex 100 (200-400 mesh, Bio-Rad). When chromat.ofocusing was performed in the presence of 1 mM CaC12, the eluted fractions had to be freed of calcium prior to the Chelex assay. Therefore 0.1 ml of the Chelex 100/Tris HCl buffer, pH 7.4 (4:l) suspension was added to 0.3 ml of aliquots of each fraction. Samples were shaken for 30 min and centrifuged for 5 min at 5000 X g. Calcium-binding activities were then assayed on 0.25 ml of aliquots of the supernatant incubated with Ca and Chelex 100 suspension according to the Chelex method. Radioactivity was counted in 7 ml of Picofluor 15 (Packard Instrument Co.) in an SL 40 Intertechnique scintillation counter (Kontron, France).
Radioimmunoassay of CuBP-CaBP concentrations in cytosol samples were directly measured by an RIA using immunserum directed against the duodenal 10,000 CaBP from rats, with pure mouse 10,000 CaBP as standard and '2sI-labeled pure mouse 10,000 CaBP as tracer, as previously described (16). Results were expressed as pg of CaBP/mg of cytosolic proteins. Immunoreactivity was also tested on fractions eluted from the different columns, by incubating aliquots with the immunserum and '251-labeled pure mouse 10,000 CaBP. The results are expressed as 1251-labeled CaBP bound to the immunserum in the presence of aliquots ( B ) uersus '2511-labeled CaBP bound to the immunserum in the absence of CaBP (Bo). Radioactivity was counted in a CG 2,000 Intertechnique y counter (Kontron, France).
The flask was gassed with 95% 02, 5% CO:! and incubated for 5 h in a shaking water bath a t 37 "C. The cytosol was prepared, and the total protein synthesis was estimated as described by Bruns et al. (15). Since, in earlier preliminary experiments, we observed nonspecific absorption by Protein-A (Pansorbin, Calbiochem) of high molecular weight "H-labeled proteins, ["H]leucine-labeled cytosol was precipitated with 300 mg of solid (NH4)2S04/ml for 20 rnin at 4 "C before immunoprecipitation. After centrifugation, the supernatant was desalted through a short column of Sephadex G, eluted with 0.1 M ammonium acetate, pH 7.2. After lyophilization, the proteins were dissolved in 0.4 ml of a buffer containing 0.01 M Tris-HCI, pH 7.4, with 0.25 M sucrose and 1.5% Triton X-100. Following exactly the procedure described by Bruns et al. (15), 0.2 ml of aliquots were then immunoprepipitated with 20 pl of the anti-duodenal rat 10,000 CaBP 45 immunserum. 20 pg of pure unlabeled rat 10,000 CaBP were added to the remaining 0.2 ml before immunoprecipitation, to assess the latter's specificity. After three washings, the immunoprecipitate was dissociated by heating in buffer containing 0.1 M Tris-HC1, 2% sodium dodecyl sulfate, 10% glycerol, and 1.5% dithiothreitol. "-labeled 10,000 CaBP synthetized by mouse kidney was characterized by both sodium dodecyl sulfate-15% (w/v)-polyacrylamide gel electrophoresis and by GP-HPLC.
Other Methods-Plasma calcium was measured by atomic absorption spectroscopy (Perkin Elmer, France). Protein contents were determined by a modified Lowry method (20).

Biochemical Evidence for Two CaBPs in the Mouse Kidney
To investigate the calcium-binding activity, mouse-kidney cytosol was fractionated by several biochemical techniques including gel fitration chromatography, GP-HPLC, and chromatofocusing. The eluted fractions were analyzed for calciumbinding activity and immunoreactivity using the anti-duodenal rat 10,000 CaBP immunserum and "'I-CaBP purified from mouse duodenum as described under "Materials and Methods." Sephadex G-75 Chromatography-Chromatography of mouse kidney cytosol showed that 45Ca-binding activity was distributed into two major 45Ca-binding peaks (Fig. 1). The two peaks were well separated and of roughly similar size. Immunoreactive material was present as a single peak, exactly superimposable on the low molecular weight 45Ca-binding activity. The estimated molecular weight of the protein which exhibited immunoreactivity simultaneously with calciumbinding activity was approximately 10,000. Duodenal mouse 10,000 CaBP exhibited the same property (15,16). The other 45Ca-binding peak corresponded to a CaBP with molecular weight of 25-27,000.
GP-PHLC-The recent progress achieved in the field of protein GP-HPLC (21) led us to use this powerful tool to further compare these two calcium-binding proteins. Thus, each of the two calcium-binding activities, 25,000 and 10,000 obtained by gel fitration of the mouse kidney cytosol, was submitted to GP-HPLC (Fig. 2). GP-HPLC performed at a flow rate of 0.4 ml/min was capable of discriminating between these two calcium-binding activities in 30 min. The 25,000 calcium-binding activity eluted a t a volume of 8. Cytosol (300 mg of protein) from mouse kidney was precipitated by 40% ammonium sulfate. The supernatant (5000 X g) was chromatographed through a Sephadex G-75 column (2.6 X 100 cm) in 0.1 M ammonium acetate, pH 7.0. The flow rate was 9 cm/h. Fractions (4.9 m l ) were collected and assayed for "Ca-binding activity (0) and for immunoreactivity (0). The immunserum was anti-duodenal rat 10,000 CaBP. Arrows indicate the elution volume of molecular weight markers. Blue D., blue dextran; B.S.A., bovine serum albumin; OV., ovalbumin; Chym., chymotrypsinogen; Rib., ribonuclease.
by guest on March 23, 2020 http://www.jbc.org/ Downloaded from t i d y t h e same as the elution volume obtained for the chick CaBP we prepared from duodenal mucosa. The 10,000 calcium-binding activity eluted at a volume of 9.2 mI and was also associated with immunoreactivity. The latter elution volume was exactly the one we found for the 10,000 CaBP from mouse duodenal mucosa. These results for GP-HPLC are in complete agreement with those for Sephadex G-75 filtration, but GP-HPLC has the marked advantage of being faster and more reproducible.
The respective effects of calcium and EDTA on the elution of renal calcium-binding activities from the GP-HPLC column were investigated. The 10,000 CaBP was chosen for this study because we could easily detect it by its immunoreactivity. As shown in Fig. 3, the mobility of the 10,000 CaBP depended on whether calcium or EDTA was present in the buffers. In the presence of 1 mM CaC12, the elution volume was 8.8 ml , but in the presence of 1 mM EDTA, this was always decreased to 8.4 ml, and the apparent molecular weight therefore increased. Similar changes in the apparent molecular weight were reported for a few other calcium-binding proteins such as calmodulin using electrophoresis under denaturing conditions (22,23). GP-HPLC thus provides a satisfactory new approach to detection of the particular changes induced by calcium binding and might prove useful in identifying and purifying CaBPs.
Chromatofocusing-Further characterization of both the 25,000 and 10,000 CaBPs from mouse kidney required determination of their isoelectric point. For this purpose, we chose chromatofocusing, a very convenient method for estimating PI of partially purified proteins (Fig. 4). The two calciumbinding activities (25,000 and 10,000) were applied together to

Two Vitamin D-dependent Ca-binding Proteins in Mouse Kidney
a column of Polybuffer Exchanger 94 with a pH gradient descending from 7 to 4 formed by Polybuffer 74. Chromatofocusing of mouse kidney cytosol displayed two well separated peaks of specific calcium-binding activity. One calcium-binding activity eluted a t a PI of 4.9 and was associated with immunoreactivity. Furthermore, this acid PI was identical with that of duodenal mouse 10,000 CaBP as determined by chromatofocusing and isoelectric focusing (16). The other calcium-binding component peaked at a PI of 5.9 and corresponded to the 25,000 calcium-binding activity. In addition, we found that mouse cerebellum cytosol contained a calciumbinding activity which eluted at 25,000 on a Sephadex G-75 column (not shown) and which also peaked at a PI of 5.9 on the same chromatofocusing column. Lastly, after fractionation on a Sephadex G-75 column, cytosol from chick duodenal mucosa exhibited a calcium-binding activity with a PI of 4.1. The latter value is in complete agreement with the one previously obtained by other methods (24), thus underlining the validity of the PI values we determined for mouse CaBPs. In addition, this result for chick CaBP indicates that despite the immunological cross-reactivity reported between chick and mammalian renal CaBPs (2,25) there are biochemical differences between the 25,000 CnBPs from one species to another.
T o study an influence of calcium on the apparent PI values of mouse renal calcium-binding activities, we performed chromatofocusing in the presence of 1 mM CaC12 in all the buffers, with a pH gradient descending from 8 to 5. Two calciumbinding activity peaks were again obvious. In the presence of calcium, the apparent PI of the calcium-binding activity associated with immunoreactivity was 6.8. This PI of 6.8 is identical with what we found for the calcium-bound form of duodenal mouse 10,000 CaBP (16). The 25,000 calcium-binding activity had an apparent PI of 7.0 in the presence of calcium. Therefore, the apparent PI of the two calcium-binding activities from mouse kidney changes in the presence or absence of calcium. However, the variations in PI for each renal CaBP were not the same (2 pH units for the 10,000 CaBP and 1 pH unit for the 25,000 CaBP), but the reason for this is not known.
Note that the 45Ca baseline in chromatofocusing experiments rose as pH decreased since the action of Chelex resin is known to be very dependent on pH (12).
Ouchterlony Double Immunodiffusion-The two CaBPs present in mouse kidney were compared from an immunochemical point of view using Ouchterlony double immunodiffusion (Fig. 5). With anti-duodenal rat 10,000 CaBP immunserum ( A ) a single precipitating immunocomplex was formed with mouse duodenal cytosol, in agreement with our previous report (16). A single precipitation line was also observed when this immunserum was incubated with a similar amount of mouse kidney cytosolic proteins. In addition, both precipitation lines from mouse kidney and mouse duodenum fused without spur, indicating the presence of apparently identical antigenic determinants in both mouse organs. Furthermore, in RIA, the immunodilution curve for mouse kidney cytosol was parallel to that of mouse duodenum cytosol (results not shown), thus confirming the complete immunological identity of renal and duodenal 10,000 CaBP. This immunoreactive material corresponded to the 10,000 calcium-binding activity eluted from the Sephadex G-75 column. Unlike mouse kidney, rat kidney cytosol gave no reaction with this anti-duodenal rat 10, OOO CaBP immunserum.
When anti-human cerebellar 25,000 CaBP immunserum2 was used (Fig.  5B), a precipitating immunocomplex was * The anti-human cerebellar 25,000 CaBP immunserum was a gift from Dr. 0. Parkes, Vancouver, Canada. formed with mouse kidney cytosol. This immunoreactive material was only associated with the 25,000 calcium-binding activity and not with the 10, OOO one. In addition, a precipitating immunocomplex was formed with rat kidney cytosol. The precipitating line from this cytosol completely fused with that from mouse kidney cytosol. Lastly, no cross-reacting material was detected in mouse duodenum cytosol by this anti-25,000 CaBP immunserum. All these results demonstrate that two CaBPs, which have completely different biochemical and immunological properties, are present in mouse kidney cytosol. One is identical with the cerebellar mouse CaBP, has a molecular weight of 25-27,000 and a PI of 5.9, and will be referred to as mouse renal 25,000 CaBP. The other is identical with the duodenal mouse CaBP, has a molecular weight of 9-10,000 and a PI of 4.9, and will be termed the mouse renal 10,000 CaBP.

I n Vitro Synthesis of Mouse Kidney l0, OOO 'H-CaBP
T o show that the 10, OOO CaBP we found in mouse kidney was an intrinsic protein, we tested mouse kidney homogenates for their ability to incorporate [''Hlleucine into CaBP. Radioactive 10,000 CaBP was detected by immunoprecipitation using the anti-duodenal rat 10,000 CaBP immunserum. Fig. 6 shows the GP-HPLC pattern of the immunoprecipitate from newly synthesized cytosolic protein in mouse kidney. It exhibited a single radioactive peak which eluted a t exactly the same place as mouse duodenal and renal 10,000 CaBP. This radioactive peak was no longer detected when an excess of unlabeled pure 10, OOO CaBP was added to the cytosol prior to incubation with the anti-duodenal rat 10,000 CaBP immunserum. This radioactive peak was also absent when the immunserum was omitted or replaced by non immuneserum. The same results were obtained when the immunoprecipitates were analyzed by sodium dodecyl sulfate-15% (w/v)-polyacrylamide electrophoresis (not shown). The above experiments suggested that the 10,000 CaBP present in mouse kidney is in fact synthesized in situ.

Vitamin D Dependency of CaBPs in Mouse Kidney
We investig%ted the influence of vitamir D status on both CaBPs from mouse kidney. For this purpose, native kidney cytosols from the 3 groups of mice were prepared, were assessed to contain the same amount of proteins, and were successively applied on the same Sephadex G-75 column (Fig.  7). In the cytosol from the vitamin D-repleted group (Fig. 7A), both calcium-binding activities were of similar importance. In the vitamin D-deficient group, both calcium-binding activities diminished but there was no change in their relative importance (Fig. 7B). In the vitamin D-deficient+l,25(0H)~D:j group (Fig. 7B), both activities rose. However, the increase in 10,000 calcium-binding activity was seen to be larger than that observed for 25,000 activity. The variations noted for the renal 10,000 calcium-binding activity led us to measure the mouse renal 10,000 CaBP directly by RIA, to obtain individual and more precise data (Table I). In kidneys from vitamin D-deficient mice, 10,000 CaBP concentrations (0.7 f 0.07 pg of CaBP/mg of cytosolic proteins) were significantly lower than those found in the vitamin D-repleted group (2.9 -t 0.17 pg of CaBP/mg of cytosolic proteins). Furthermore, renal 10,000 CaBP followed almost the same pattern as duodenal CaBP. Injection of 1,25(OH)2D3 in vitamin D-deficient mice completely restored the 10,000 CaBP contents in the duodenum and kidney. However, renal 10,000 CaBP concentrations rose to 4.5 k 0.34 pg of CaBP/mg of cytosolic proteins, a value higher than that for the vitamin D-repleted group, thus confirming the results obtained by the Chelex method.
Changes in Renal 10,000 CaBP during Normal Development in Mice Measurements of 10,000 CaBP in baby mouse kidney were made simultaneously with those of 10,000 CaBP in the entire small intestine, between 1 day before birth and day 23 after birth (Fig. 8). Before birth, substantial amounts of 10,000 CaBP were not only present in mouse fetal small intestine, but also in mouse fetal kidney. After birth, the developmental pattern of 10,000 CaBP in mouse kidney completely differed from that of intestinal 10,000 CaBP. In the kidney, the 10,000 CaBP concentration increased in the first week of life, since it was 1.96 f 0.07 of pg/mg of cytosolic proteins on day 2 and rose to 4.16 * 0.11 of pg/mg of cytosolic proteins on day 8.
Thereafter, the 10,000 CaBP concentration in mouse kidney remained unchanged until the postweaning period. In mouse small intestine, the 10,000 CaBP concentration remained unchanged from birth until day 18, when it increased to 3.94 * 0.42 of pg/mg of cytosolic proteins. The developmental pattern of 10,000 in mouse intestine was similar to that described in rat intestine (17,26).

DISCUSSION
Our data clearly demonstrate that two different vitamin Ddependent CaBPs are simultaneously present in the mouse kidney. The existence of the 25,000 CaBP that we partially characterized was to be expected since the kidneys of several species possess a 25-28,000 CaBP. By contrast, the synthesis of substantial amounts of a 10,000 CaBP by mouse kidney was a surprising result. Biochemical characterization of this renal 10,000 CaBP revealed its complete identity with the 10,000 CaBP from mouse duodenum. Both have the same molecular size in gel-fitration chromatography and GP-HPLC; they have the same apparent isoelectric point which exhibits similar calcium-dependent changes, and they are immunologically identical in immunodiffusion and RIA. It was tempting to estimate the molar ratio of the two CaBPs in mouse kidney. After cytosol gel filtration, the two peaks of calcium-binding activity were of similar size (Fig. 1). Although a precise calcium-binding comparison of the two CaBPs could not be made without a saturation analysis, the result suggests that in mouse kidney the molar content of 10,000 CaBP is about twice that of 25,000 CaBP since 1 mol of 10,000 CaBP binds 2 mol of calcium (12,13,27), whereas 1 mol of 25,000 CaBP binds 4 mol of calcium (28). In addition, the 10,000 CaBP concentration in mouse kidney cytosol, as determined by RIA, was about 4 pg/mg of cytosolic proteins, i.e. 0.4 nmol of CaBP/mg of cytosolic proteins. The 25,000 CaBP concentration in mouse k-idney cytosol3 was also 4 pg/mg of cytosolic proteins, i.e. 0.2 nmol of CaBP/mg of cytosolic proteins. These quantitative results thus confirm the molar ratio of 21 in favor of the 10,000 CaBP in mouse-kidney cytosol. Therefore, the predominance in mouse kidney of a 10,000 CaBP, both immunologically and biochemically identical with duodenal mouse 10,000 CaBP is a new finding in kidney physiology.
The intrarenal distribution of mouse 10,000 CaBP is also of interest. We showed that 10,000 CaBP is present in both the cortex and medulla of mouse kidney, whereas 25,000 CaBP is probably present in the cortex only, as in rats and chicks (9,25,29,30). Further immunolocalization studies will specify 0. Parkes, personal communication.
which mouse renal cells contain each type of CaBP.
Although the renal and intestinal 10,OOO CaBPs are biochemically identical and vitamin D-dependent, their respective variations may be different in certain physiological states. The present study shows that during the postnatal development, the evolution of mouse renal 10,000 CaBP was completely different from that of intestinal 10,000 CaBP. Renal 10,000 CaBP increased during the first week after birth, which is a period of intense nephrogenesis (31,32). Intestinal CaBP, on the other hand, reached its adult level only at weaning when the intestine becomes mature (33). It seems clear, therefore, that the 10,000 CaBP synthesis in kidney and intestine during mouse development is related to the maturation specific to each organ.
The vitamin D dependency of both mouse renal 25,000 and 10,000 CaBPs was illustrated by the fact that they decreased in vitamin D-deficient mice and rose after 1,25(OH)2D3 injection. This lends support to the idea that the cytosolic and nuclear 1,25(OH)2D3 receptors in mouse kidney are functional (34,35). Therefore, mouse kidney appears as a peculiar organ in which the hormonal action of 1,25(OH)2D3 results in two different molecular expressions. Consequently, it could provide a unique model for studying the genomic action of 1,25(OH)zD3.