1,25-Dihydroxyvitamin D3-mediated Intestinal Calcium Transport

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY by The American Society of Biological Chemists, Inc Vol. 261, No. 34, Issue of December 5, PP ,1986 Printed in U.S.A. 1,25Dihydroxyvitamin D3mediated Intestinal Calcium Transport BIOCHEMICAL IDENTIFICATION OF LYSOSOMES CONTAINING CALCIUM PROTEIN (CALBINDIND,sK)* Ilka Nemere, Valerie Leathers, and Anthony W. Norman AND CALCIUMBINDING (Received for publication, May 1, 1986) From the Department of Biochemistry and Division of Biomedical Sciences, University of California, Riverside, California A variety of intestinal cell organelles and proteins epithelium continues to be a debatable subject (1). Historihave been proposed to mediate 1,225dihydroxyvitamin cally, two of the earliest candidates proposed for calcium D3 (1,25(OH)2D3)stimulated calcium absorption. In sequestration include the vitamin Dinduced calciumbinding the present study biochemical analyses were under protein (CaBP or calbindin) reported by Wasserman and taken to determine the subcellular localization of Ca Taylor (2) and mitochondria (3). Lysosomes were implicated after calcium transport in vivo in ligated duodenal in the effect of vitamin D on the intestine in the pioneering loops of vitamin Ddeficient chicks injected with 1.3 work of Jande and Brewer (4) and later by other laboratories nmol of 1,25(OH)2D3 or vehicle 15 h prior to experimentation. Separation of Golgi, mitochondria, basal (59). Since the studies of Warner and Coleman (10) on lateral membrane, and lysosome fractions in the epi localization of absorbed calcium by electronprobe analysis thelial homogenates was achieved by differential sedi have alternately been interpreted to implicate lysosomes (1) mentation followed by centrifugation in PercollO gra or Golgi, the latter organelle has also been extensively studied dients and evaluation of appropriate marker enzyme and found to be affected by vitamin D status (1113). More activities. Both vitamin Ddeficient and 1,25(OH)zD3 recently, microsomes have been suggested as a site of intratreated chicks had the highest levels of Caspecific cellular calcium sequestration (14). activity in lysosomal fractions. The lysosomes were The conclusions of several of the reports cited above are also the only organelles to exhibit a 1,25(OH)2D3 based on the result of electron microscopic analyses that have mediated difference in calcium content, increasing to been criticized because of potential redistribution of calcium 138% of controls. Lysosomes prepared from 1,225 during processing of the tissue. Many of the other observa (OH)zD3treated chicks also contained the greatest lev tions arise from studies on the uptake of 45Ca by isolated els of immunoreactive ~a1bindind~~~ (calciumbinding organelles in uitro. The current report utilizes biochemical protein). Chloroquine, a drug known to interfere with methods to determine subcellular 45Ca localization after a lysosomal function, was tested and found to inhibit 1,25(OH)2D3stimulated intestinal calcium absorpperiod of linear transport in uiuo. The results indicate that tion. Neither l,25(oh)2d3 nor chloroquine affected with careful processing of the tissue and inclusion of inhibitors [3H]20 transport. In additional experiments, micro of Ca2+ redistribution in the isolation media, intestinal lysosomal membranes (105,000% gpellets) were subjected somes are implicated in 1,25(OH)2D3mediated calcium to gradient centrifugation. The highest levels of 4SCa transport as vesicular carriers of the cation. specific activity and calciumbinding protein in material from 1,25(OH)2D3treated chicks were found in MATERIALS AND METHODS fractions denser than endoplasmic reticulum and may AnimakWhite Leghorn cockerels (Lakeview Farms, Lakeview, represent endocytic vesicles. In studies on intestinal CA) were obtained on the day of hatch and raised for 34 weeks on a mucosa of 1,225(OH)zD3treated birds fractionated standard rachitogenic diet (15). Chicks, four per group, were injected after 30 min of exposure to lumenal Ca2+ or Ca2+ plus intramuscularly with 1.3 nmol of 1,25(OH)*Da or an equivalent chloroquine, 4SCa was found to accumulate in lyso volume of vehicle (ethanol:propanediol, 1:1, v/v) 15 h prior to expersomes and putative endocytic vesicles, relative to conimentation and fasted overnight. On the day of experimentation, chicks were anesthetized with ether, the duodenal loop surgically trols. A mechanism involving vesicular flow is proexposed, and ligated at the proximal and distal ends. The lumen of posed for 1,25(OH)2D3mediated intestinal calcium the loop was then injected with 50 pci of 45Ca and 4 mg of 40Ca in a transport. Endocytic internalization of Ca2+, fusion of total volume of 0.4 ml, unless otherwise indicated. The loopwas the vesicles with lysosomes, and exocytosis at the basal replaced in the abdominal cavity, and calcium transport allowed to lateral membrane complete the transport process. proceed for 30 min in vivo. At the end of this time, the chicks were reanesthetized, and the duodenal loop excised and removed to icecold physiological saline, prior to decapitation of the animals. This procedure circumvented the extreme sensitivity of intestinal tissue to The subcellular site of calcium during 1,25dihydroxyvita anoxia and concomitant release of 5Ca from its site of sequestration min DB (1,25(OH),D3) stimulated transport across intestinal prior to the introduction of redistribution inhibitors. After thorough chilling, the duodena were slit longitudinally and the mucosa collected * This work was supported by United States Public Health Service by scraping into 20 ml of icecold homogenization medium (0.25 M Grant AM The costs of publication of this article were sucrose, 5 mm histidine:imidazole, ph 7.0) supplemented with 10 pg/ defrayed in part by the payment of page charges. This article must ml of ruthenium red and 2 mm EGTA to prevent redistribution of therefore be hereby marked advertisement in accordance with 18 the radionuclide (16). All subsequent steps were performed at 04 C. U.S.C. Section 1734 solely to indicate this fact. Subcellular FractionationFollowing disruption of the intestinal The abbreviations used are: 1,25(OH)2D3, 1,25dihydroxyvitamin epithelium in 40 ml of homogenization medium (70 strokes by hand using a Dounce homogenizer and Teflon pestle) differential centrif D3; CaBP, calciumbinding protein; EGTA, [ethylenebis(oxyethylenenitri1o)ltetraacetic acid; HEPES, 4(2hydroxyethyl)lpiperazineethanesulfonic acid; ELISA, enzymelinked immunoadsorption assay ugation fractions were obtained by the procedure of Weiser et al. (17). A pellet (PI) containing nuclei and brush borders was obtained after centrifugation at 1,000 X g for 20 min. The resulting supernatant

2 fraction was centrifuged again for 10 min at 20,000 x g to yield Pp, enriched in intracellular organelles and basal lateral membranes, and the supernatant fraction Sl, containing microsomes and cytosol. In other experiments, microsomes (Pa) were sedimented at 105,000 X g for 1 h. Separation of the components in P, or P, was achieved by centrifugation in Percoll (Pharmacia, Piscataway, NJ). The pellets were resuspended in 35 ml of Percoll (starting density, 1.06 g/ml), 0.25 M sucrose, 5 mm histidine:imidazole, 10 pg/ml ruthenium red, 2 mm EGTA, ph 7.0. Twentythree ml of each suspension were transferred to polycarbonate tubes containing 2ml cushions consisting of 2.4 M sucrose. The suspensions were then centrifuged for 45 min at 30,000 X g, in a 60 Ti rotor (Beckman Instruments). It should be noted that although Percoll gradients are rapidly formed, they are also altered during the course of centrifugation. Thus, the gradient centrifugation conditions described pertain to a particular Beckman L550 ultracentrifuge. Use of a different L550 ultracentrifuge equipped with more efficient acceleration has been known to result in a downward shift of lysosomal and mitochondrial fractions (see Fig. 2, top panel). Fractions of 40 drops were collected from the gradients by gravity flow: a 20cm length of metal tubing (10 gauge), the top end of which was connected to an additional length of polyethylene tubing, was lowered into the gradient. Fraction collection was initiated by the application of a mild vacuum to the free end of the polyethylene tubing. Duplicate aliquots were taken from each fraction for liquid scintillation spectrophotometry and analytical determinations. Analytical DeterminutiomProtein was determined by the method of Lowry et al. (18). Blank gradients were analyzed in parallel to correct for significant interference by Percoll. Procedures for the determination of ouabainsensitive Na+,K+ATPase (19), galactosyltransferase (19), succinate dehydrogenase (20) glucose6phosphatase (21), cathepsin B (7), and alkaline phosphatase (22) were as previously described. Acid phosphatase activity was assayed in 0.1 M citrate buffer, ph 5.2, using pnitrophenyl phosphate as substrate (Sigma). Calciumbinding protein was determined immunologically (23). Intestinal Calcium AbsorptionThe duodenal loops of ether anesthetized chicks were ligated at the proximal and distal ends and 0.4 ml of a solution containing 5 pci of "Ca2+ and 4 mg of "Ca" injected. Where indicated, the designated concentrations of chloroquine or its analogues, quinacrine andprimaquine (all from Sigma), were included in only the lumenal solution. When necessary, the ph of the control calcium solutions was adjusted to correspond to that of the drug 1,25(0H)2D3 and Calcium Transport in Lysosomes ity Na+,K+ATPase. An apparent difference in galactosyltransferase activity as a function of vitamin D status is also indicated in Fig. lb, in that basal lateral membranes prepared from vitamin Ddeficient chicks (open circles) were found to have higher enzyme activity levels than corresponding frac tions from 1,25(0H),D3treated chicks (closed circles). The mitochondrial marker enzyme, succinate dehydrogenase, occupied a rather broad area in the middle, shallow region of the gradient, spanning fractions 413 (Fig. 1D). Acid phosphatase activity, a marker enzyme for lysosomes, was predominantly enriched in fractions 13 at the bottom of the gradient (Fig. 1E). Lysosomal fractions obtained from the intestinal epithelium of 1,25(OH),D3treated chicks tended to have greater levels of acid phosphatase activity than corresponding preparations from vitamin Ddeficient controls, although the differences shown in Fig. 1D were not significant, as judged by Student's t test for unpaired observations. Fig. 1F depicts the distribution of another lysosomal marker enzyme, the sulfhydrylrequiring protease cathepsin B (cf. Ref. 7). Whole homogenate values were found to be 2.25 and 2.45 nmol/min/ mg protein for vitamin Ddeficient and treated chicks, respectively. Fig. 1G confirms the absence of gross contamina tion of Pp by brushborder membrane fragments. Although the majority of the alkaline phosphatase activity in the gradient is associated with basal lateral membranes, the levels are not enriched over whole homogenate values (291 f 34 and for vitamin Ddeficient and treated preparations, respectively). The smaller peak associated with the lysosomal fractions confirms earlier observations on the internalization of this plasma membrane marker enzyme (8,24,25). Table I provides a comparison of the distribution of the activities analyzed in whole homogenates and differential centrifugation fractions as a function of vitamin D status. Evaluation of the data for whole homogenate levels of protein and marker enzyme activities indicates that 1,25(OH)2D3 treatment for 15 h did not significantly alter these values containing solution. For [3H]20 transport studies, the lumenal solution consisted of 25 pci in water, plus or minus chloroquine. After 30 relative to vitamin Ddeficient controls. In contrast, 45Camin of transport in uiuo, the chicks were decapitated and the blood specific activity was substantially lower in all fractions precollected. pared from 1,25(OHlzD3treated chicks, relative to controls, CalculationsCaBP levels were corrected for protein interference no doubt due to the secosteroidinduced increase in transcelat low dilutions by subtracting the values determined in fractions lular calcium transport. prepared from vitamin Ddeficient chicks from the corresponding The final enrichment factors and recoveries of the activities fraction prepared from 1,25(OH),D3treated birds. Despite nonspecific protein interference, the anticabp antibody did not recognize analyzed are presented in Table 11. The apparent differences calmodulin at concentrations corresponding to the linear dilution in enrichment factors for the marker enzyme activities can range of the CaBP standards. Statistical analyses were performed be attributed to a slight decrease of the corresponding value using Student's t test for unpaired observations. in whole homogenates prepared from 1,25(OH),D3treated chicks, relative to controls (see Table I), rather than differ RESULTS ences in gradient profiles (see Fig. 1). The calculated recov Distribution of Marker Enzyme ActivitiesSeparation of eries for the parameters studied were mainly in the range of the components in the 20,000 X g pellet fraction (P,) of chick 85% or more with the exception of succinate dehydrogenase intestinal mucosa was achieved by centrifugation through a and acid phosphatase activities which showed recovery levels Percoll gradient. Fig. 1 illustrates the distribution of protein of 6369%. and the marker enzyme activities analyzed. Fractions 16 and Subcellular Distribution of 45Ca during Transport in Vivo 17, near the top of the gradient, were found to contain basal Pilot studies were undertaken to determine optimal homogelateral membranes, as judged by the enrichment in Na+,K+ nization conditions that minimized the redistribution of the ATPase activity (Fig. 1B). A smaller peak of Na+,K+ATPase radionuclide into subcellular organelles such as mitochondria. activity was also observed in fractions 13 and may represent Transport of 45Ca (5&i/0.4 ml) in ligated duodenal loops of that portion of the basal lateral membrane that undergone has vitamin Ddeficient chicks was allowed to proceed for 30 min endocytosis and fusion with lysosomes (see below; Refs ). Galactosyltransferase, an enzyme activity known to be in situ. The mucosa collected from such duodena were then homogenized in isotonic sucrose in the presence or absence enriched in both the Golgi apparatus and plasma membranes of ruthenium red and EGTA (16). In the absence of any (17), was found in fractions 1418 (Fig. 1C). Since fractions additions to the sucrose homogenization and Percoll media 16 and 17 contained basal lateral membranes, position the of (Fig. 2, top panel), peaks of 45Caspecific activity (cpm/mg Golgi membranes was assigned to fractions 14 and 15, as this protein) were observed in those regions o'f the gradient cormaterial exhibited enriched galactosyltransferase activity, but responding to lysosomes, mitochondria, and Golgi memrelatively low levels of the basal lateral marker enzyme activ branes. The values to the right of the panels represent the

3 ,25(0H)~D3 Calcium Transport and in Lysosomes BOTTOM FRACTlONUMBER w E BOTTO t? 500 t; LI Y F300 i e 200 o n ra a " 4 0 I E 9 10 II IE EOTTOH NUMBER FRICTION FIG. 1. Distribution of protein and marker enzyme activities in Percoll gradients. Vitamin Ddeficient chicks (four per group) were injected with 1.3 nmol of 1,25(OH)2D3 or vehicle 15 h prior to experimentation. On the day of use, the ligated duodenal loops of anesthetized chicks were injected with 50 pci of 45Ca and 4 mg of 40Ca. After 30 min of transport in uiuo, the chilled duodenal mucosa was collected by scraping, homogenized, and the 20,000 X g pellet fractions (PJ prepared for centrifugation in Percoll gradients. Fortydrop fractions were collected for determination of radioactivity (see Figs. 4 and 5), protein, and enzyme activities. For A, B, D, and E, values represent the mean f S.E. for four gradients prepared from tissue of vehicle controls (0) and three for intestinal epithelium of 1,25(OH)2D3treated chicks (0). Values in C represent average f range for duplicate gradients. Values in F and G represent single gradient profiles. 45Caspecific activity in whole homogenates and sequentially prepared pellet and supernatant fractions. In the absence of status, the highest levels of the radionuclide comigrate with lysosomal fractions during Percoll gradient centrifugation. It the ruthenium red and EGTA additions, fraction Pz was is also evident that impaired calcium transport in the rachitic enriched in radionuclidespecific activity, whereas only low chick resulted in much higher levels of the radionuclide in all levels of 45Ca were found in Sp, the supernatant fraction fractions, relative to corresponding fractions obtained from containing microsomes and cytosol. Inclusion of 10 pg/ml of 1,25(OH)2D3treated chicks. In order to facilitate a direct ruthenium red, an inhibitor of mitochondrial calcium uptake, comparison, the specific activity in each fraction was exabolished the peaks of radioactivity in the mitochondrial and pressed as the percent of total gradient specific radioactivity, Golgi regions of the gradient, but not the lysosomal peak (Fig. 2, middle panel). Inclusion of both ruthenium red and 2 mm EGTA resulted in a sharper lysosomal peak, as well as a large increase in "soluble" radioactivity (Fig. 2, bottom panel). After ascertaining that the subcellular calcium was predomas shown in Fig. 4. Virtually no difference in 45Ca content was observed in basal lateral membrane and Golgi fractions prepared from 1,25(OH)zD3treated deficient or chicks, nor was there a significant difference in the mitochondrial fractions. However, fraction 3 of the Percoll gradient, which contained inantly contained within the lysosomal compartment of vita the peak of lysosomal acid phosphatase activity, exhibited min Ddeficient chick intestinal epithelium, studies were un 45Ca levels that were 138% of corresponding fractions from dertaken to determine whether the divalent cation content of vitamin Ddeficient controls (p < 0.01). Fraction 4 prepared this organelle exhibited a vitamin Dmediated difference. Fig. from 1,25(OH)zD3treated chicks also contained greater 3 illustrates the results of analyses for 45Ca in the same radionuclide levels than corresponding controls (p < 0.05), fractions prepared to determine marker enzyme distribution but this may be due to greater mitochondrial contamination (shown in Fig. 1). When the data are expressed as 45Ca in the gradient fraction prepared from vitamin Ddeficient specific activity, it is evident that regardless of vitamin D chicks (see Fig. IC).

4 +I +I +I +I +I +I +I +I +I +I +I 1,25(0H)2D3 and Calcium Transport in Lysosomes Subcellular Distribution of CaBPThe biochemical identification of calciumcontaining lysosomes generated during transport of the divalent cation in vivo, prompted an analysis of the same gradient fractions for the vitamin Dinduced CaBP by an enzymelinked immunoadsorbtion assay (ELISA). The results of these determinations, shown in Fig. 5, indicate that the highest immunoreactive CaBP levels were also localized in the lysosomes, but in gradient fractions 1 and 2, rather than 3 (p < 0.02 for fractions 1 and 2, relative to other gradient fractions). Evaluation of the Role of Lysosomes during Ca2+ Transport in VivoIn order to assess the relationship between calciumcontaining lysosomes and absorption of the divalent cation, bioassays were undertaken with the drug chloroquine, a known inhibitor of lysosomal function (28). Fig. 6A illustrates that increasing concentrations of lumenal chloroquine resulted in progressive inhibition of 1,25(OH)*D3induced intestinal calcium absorption. Significant suppression was observed at 60 and 100 mm chloroquine (both, p < 0.001). Although this latter concentration may appear to be high it can be calculated that under typical tissue culture conditions in which chloroquine is a test substance, each cell receives 1 10 nmol of the drug (eg. Ref. 29). The chick duodenal loop contains approximately 1.4 x lo9 cells, based on analyses for DNA,' and it can thus be estimated that each cell received 60 pmol of chloroquine during studies using the highest drug concentration. Lumenal exposure of vitamin Ddeficient chicks to 100 mm chloroquine reduced intestinal calcium absorption to 194 & 14 cpm/o.l ml of serum (mean f S.E. for seven birds), a value somewhat lower than that observed in 1,25(OH)2D3treated animals exposed to an equivalent concentration of drug (Fig. 6A). These observations were subsequently extended to include a range of compounds structurally related to chloroquine, each of which was tested at a concentration of 60 mm. Fig. 6B indicates that 60 mm quinacrine reduced 45Ca transport to 171% of values obtained in vitamin Ddeficient chicks, and an equivalent concentration of primaquine resulted in intestinal calcium transport levels that were 106% of those in vitamin Ddeficient birds. Chloroquine, which inhibited transport to 157% of levels in rachitic chicks, thus fell between the efficacy of quinacrine and primaquine. Additional studies were conducted to determine whether chloroquine acted through general disruption of transport. For this purpose, [3H]20 (25 &i/0.4 ml) was introduced to the duodenal lumen of vitamin Ddeficient chicks and 1,25 (OH)2D3treated birds, half of which also received 100 mm lumenal chloroquine. Mean serum tritium/o.l ml f S.E. was found to be 7875 rt 308, , and 7980 rt 130, respectively, for each group of five birds. Microsomal AnalysesThe recent report by Rubinoff and Nellans (14) suggesting that intestinal microsomes are involved in calcium sequestration, prompted an investigation of the 105,000 X g pellet fraction (P3). Fig. 7 depicts the distribution of two marker enzyme activities following Percoll gradient centrifugation. Glucose6phosphatase activity, an indicator of endoplasmic reticulum, was observed in fractions 1418 (Fig. 7A). Since the light membrane fractions could also contain endocytic vesicles originating from the brushborder membrane, the same gradient fractions were analyzed for alkaline phosphatase activity (Fig. 7B). It is evident that treatment of chicks for 15 h with 1.3 nmol of 1,25(OH)zD3 substantially enhanced the level of alkaline phosphatase activity in nearly every fraction of the gradient, but particularly in fractions 45 and 18. The marked enrichment of alkaline J. E. Bishop and G. Theofan, personal communication.

5 ~ ~~ ,25(OHj2D3 Calcium and Transport in Lysosomes TABLE I1 Enrichment and recovery of activities analvzed Final enrichment factof % Recovery: differential centrifugationb D +D D +D Protein 85 f 3 86 f 6 Wa f 1 90 f 3 Na+,K+ATPase f f 19 Galactosyl transferase f f 5 Succinate dehydrogenase f 7 69 f 11 DhosDhatase Acid f 1 66 f 2 Peak gradient fraction relative to whole homogenate value. *Recovery of activity in fractions PI, P,, and Sz relative to activity in whole homogenates. Recovery of total activity in Percoll gradients (fractions 118) relative to initial activity in P,. x WH 1428 P, 2010 P S2 238 NO ADDITIONS % Recovery: gradient' D +D 94 f 2 98 f 5 86 f 5 76 f 3 79 f f f f f f f c WH 1153 PI 1053 P S RUTHEN I UM RED PI 420 P2 294 S RUTHENIUM RED EGTA I I BOTTOM FRACTION NO. FIG. 2. Comparison of homogenization conditions on distribution of 46Ca in Pz fractions from vitamin Ddeficient chicks given 5 pci of 46Ca2+ intralumenally for 30 min. The sucrosebased homogenization media either contained no additions, 10 pg/ml of ruthenium red, or an equivalent concentration of ruthenium red and 2 mm EGTA. Gradient values represent single experiments. Values to the right of each panel represent cpm/mg protein in whole homogenates (WH), 1,000 X g pellets (Pd, 20,000 X g pellets (Pd, and supernatant fractions (S2). I IO I I BOTTOM FIG. 3. Subcellular distribution of 46Caspecific activity in 1,25(OH)zD3treated and deficient chicks. See Fig. 1 for experimental protocols. Values represent the mean f S.E. for four control (0) and three treated preparations (0). I I I I I I I I I I I I l l I I I I * LYSOSOMES MITOCHONDRIA GA BLM I phosphatase activity in fractions 1 and 2 relative to whole homogenate levels (see above) suggests the presence of dense fragments of brushborder and associated core material at the bottom of the Ps gradients. Determination of 45Ca distribution in microsomal membranes was conducted at two different time points. The initial pilot study, conducted after 30 min of calcium transport in uiuo, revealed peaks of radioactivity in gradient fractions 35 (Fig. 8A). However, the 45Ca peak in membranes prepared from 1,25(OH)2D3treated chicks was somewhat denser and contained much less radioactivity than similar membranes prepared from vitamin Ddeficient controls. If this region of ' I IO II BOTTOM *_PcO.OI. EP"0.05 RELATIVE TO LEVELS IN CORRESPONDING FRACTIONS FROM CONTROLS FIG. 4. The data in Fig. 3 are presented for each fraction as percent of total gradient specific activity.

6 I the gradients does indeed represent endocytic vesicles, the increased density of the peak from 1,25(OH)2D3 chicks could be due to the presence of specific calciumtransport protein(s). The decrease in radioactivity, relative to vitamin Ddeficient controls, could be due to completion of fusion of endocytic vesicles with lysosomes after 30 min of calcium transport in uiuo. Experiments were thus conducted in which calcium transport was allowed to proceed for only 3 min (Fig. 8B). Under such brief transport conditions, it is evident that the major peak of radioactivity in membranes prepared from 1,25 (OH),D,treated chicks retains its greater density, but also contains greater levels of the radionuclide. A second, smaller peak was also observed in fraction 18 from 1,25(OH)zD3 treated birds (Fig. 8B). The same gradient fractions (prepared after 3 min of calcium transport in vivo) were also analyzed for the presence of CaBP. Fig. 8C illustrates these results. Although several regions of CaBP immunoreactivity are evident within the gradients prepared from microsomes of 1,25(OH)2D3treated chicks, the content was highest and most consistent in fraction 4, thereby paralleling 45Ca distribution. Effect of Chloroquine on the Subcellular Distribution of 4sC~Following the identification of two 1,25(OH)2D3sensitive, subcellular compartments enriched in 45Caspecific activity, studies were undertaken to determine the effect of chloroquine on radionuclide distribution. Fifteen h after a single dose of 1,25(OH)2D3, the duodenal lumen of each chick was injected with the calcium solution plus or minus 100 mm chloroquine buffered to ph 6.8 with 100 mm HEPES. After 0 1 ' " ~ " " ' ~ " ' " ~ ' 1 1 I IO II I FIG. 5. Distribution of vitamin Dinduced CaBP immunoreactivity in Percoll gradients. Samples described in Fig. 1 were analyzed for CaBP by ELISA procedures. Due to nonspecific protein interference at the low dilutions used, values determined for fractions from vitamin Ddeficient birds were subtracted from corresponding values in gradient fractions from 1,25(OHI2D3treated chicks. 1,25(0H)2D3 and Calcium Transport in Lysosomes min of absorption, the tissue was processed as described for gradient centrifugation of fractions Pp and P1. Chloroquine exposure severely labilized the calcium sequestering organelles upon homogenization as revealed by reduced specific activity in both the P2 and P3 fractions, and concomitant rise in S3 soluble specific activity (Fig. 9A, inset). In order to compare the relative distribution of 45Ca within the gradients, the data for each fraction were again expressed as percent of gradient specific activity. As illustrated in Fig. 9, lumenal chloroquine increased lysosomal 45Ca levels to % of corresponding fractions isolated from 1,25(OH)2D3treated birds that were not exposed to the drug (Fig. 9A). A significant decrease in Golgiassociated 45Ca (fraction 15) was also noted in material prepared from chloroquinetreated chicks. A 30 min exposure to the drug chloroquine also resulted in an increase in 45Caspecific activity in the denser microsomal fractions, to % of levels in corresponding fractions prepared from 1,25(OH),D3treated controls (Fig. 9B). These increases may, however, be an underestimate, since homogenization of chloroquinetreated tissue resulted in a large in crease in soluble 45Caspecific activity (see inset, Fig. 9A). The finding that chloroquine increased 45Ca levels in both lysosomes and putative pinocytic vesicles is very likely due to initial fusion of the chloroquinecontaining vesicles with lysosomes which resulted in a blockage of both subcellular compartments DISCUSSION The present work demonstrates that differential sedimentation combined with Percoll gradient centrifugation allows resolution of several subcellular organelles and basal lateral membranes of chick intestinal epithelium. When combined with 4sCa transport in uiuo, the subcellular fractionation protocols allow identification of lysosomes as the organelles responsible for sequestering the highest levels of calcium radionuclidespecific activity. Observation of the comigration of 45Ca with the lysosomal fraction under the three homogenization conditions tested lends credence to the proposition that this organelle is intimately involved in calcium seques tration and transport in vivo. The peak is present in the absence of redistribution inhibitors and therefore is not artifactually generated by the addition of ruthenium red and EGTA (see Fig. 2). The calciumcontaining lysosome peak is also found in the presence of ruthenium red, with or without EGTA, suggesting that the divalent cation is sequestered within the organelle. In contrast, 45Caspecific activity is profoundly diminished by these agents in the mitochondrial I600 FIG. 6. Effect of chloroquine and analogues on 1,25(OH)zD,mediated intestinal calcium absorption. The ligated duodenal loops of anesthetized chicks were injected with control solution (5 pci of 45Ca and 4 mg %a in 0.4 ml) or the same solution containing the indicated concentration of chloroquine (A). In B all concentrations were 60 mm analogue. After 30 min of calcium transport in uiuo, the chicks were decapitated, the blood collected, and serum prepared for liquid scintillation counting. Values represent mean * S.E. for between 3 and 11 birds per group. 1400kA a w E AJ 'loo ZOO 0 I.ZSlOH1~0~ CHLOROOUINE 20mM 60mM IOO~M u v) 1000 ti ' 800 k Y) Ilr,

7 16112 l,25(0hj2d3 and Calcium Transport in Lysosomes FIG. 7. Distribution of protein, glucose6phosphatase, and alkaline phosphatase activities in microsomes. Procedures were as described in the legend to Fig. 1 except that microsomes sedimenting at 105,000 X g (P3) were subjected to Percoll gradient centrifugation. A and C, values represent mean? S.E. for three independent experiments; B, values represent average? range for two independent experiments. io I E 9 10 I I5 16 I7 IS 19 FR4CTlON NUMBER FIG. 8. Distribution of 45Ca and immunoreactive CaBP in Percoll gradients of microsomes prepared from 1,25(OH)zD~treated and deficient chicks. Procedures were as described in the legend to Fig. 7. A, microsomes prepared after 30 min of calcium transport in uiuo; B, microsomes prepared after 3 min of calcium transport in uiuo; values represent mean? S.E. for three independent experiments, each using four chicks treated with 1.3 nmol of 1,25(OH)2D3 (0) and four with vehicle (0) 15 h prior to experimentation. C, microsomes prepared as in B were analyzed for CaBP by ELISA. Values represent average & range for two independent experiments. I I, I, 1, I,,,,,,,,,, CPY,rnQ PROTEIN x IO" ; 12 8 I fo3 246f37 2s I t57 l6zfob 53fDD ,flB lalfip FRACTlON t I E 9 10 I 1 I I I I5 I6 I7 10 NUMBER FRACTlON.f<OO5,RELATlVE TO 1,25lOU)pD, ~~~OOO~.REL~TIVE TO 1.zs(on)203 :&o 005 *Ee0,035 FIG. 9. Effect of lumenal chloroquine on the subcellular distribution of 4aCa2'. Procedures were as described in the legend to Fig. 1 except that the lumenal calcium solution also contained 100 mm HEPES plus or minus 100 mm chloroquine, ph 6.8. Values represent means & S.E. for gradients prepared from intestinal epithelium of 1,25(OH)*D3treated chicks, without lumenal chloroquine (0, n = 4), and with lumenal chloroquine (m, n = 3). A depicts results obtained after gradient centrifugation of fraction P, (containing lysosomes) whereas results from microsomes, prepared during the same experiments are illustrated in B. and Golgi fractions, indicating that calcium binding by the vitamin Ddependent difference in radionuclide content is the latter two organelles is due to redistribution of the radio lysosome. The observed increase in 45Caspecific activity in nuclide attendant upon homogenization (16). lysosomes prepared from 1,25(OH),D3treated chicks, rela Moreover, when 45Ca distribution is analyzed as "percent tive to controls, is in agreement with earlier findings (5, 6, of gradient," a procedure that compensates for 1,25(OH)2D3 10). Although in the present work the calculated increase in stimulated calcium transport, the only organelle to exhibit a calcium content of lysosomes after 1,25(OH)*D3 treatment

8 appears small (138% of controls) when compared with the magnitude of absorption, as determined by 45Ca in the serum, the lack of excessive cation accumulation by lysosomes supports the concept of their function as a transport organelle. Constant delivery of the divalent cation to the contralumenal surface through fusion of the lysosomes with the basal lateral membrane, and exocytosis of the calcium, would prevent a massive increase of the divalent cation in the subcellular organelles. A number of earlier reports by Feher and Wasserman (30 32) have documented the existence of a membraneassociated CaBP, although the authors were unable to conclusively identify the source of the membrane. In the current work, it is evident that one source of membraneassociated CaBP is the lysosome. Indeed, extraction of the Percoll gradient fractions with Triton X100 resulted in detection of somewhat higher levels (approximately 1.4fold) of lysosomal CaBP than in the absence of the detergent, as judged by ELISA procedures, but not to the extent that would be expected if the protein were sequestered within the organelle in purely soluble form.3 However, neither does CaBP appear to be an integral lysosomal membrane protein. In epithelial tissue that is not adequately chilled prior to homogenization in the presence of redistribution inhibitors, lysosomal labilization occurs as judged by decreased organelleassociated acid phosphatase. Such labilization is accompanied by a parallel redistribution of CaBP and 45Ca into the mitochondrial fractions of the Percoll gradients. The origin of the CaBP in the lysosomes is currently unknown. However, in keeping with what is currently known about receptormediated endocytosis, it is conceivable that CaBP arising in the brushborder membrane (33) could act as a receptor for calcium, and upon binding the divalent cation, become internalized through formation of endocytic vesicles. The most usual fate of such vesicles is fusion with lysosomes (cf. Ref. 34). A similar endocytic mechanism was proposed by Jande and Brewer in 1974 (4), on the basis of electron microscopic investigations. Evidence for the postulated endocytic vesicles has been previously published on the basis of fine ultrastructural studies (4) and electron probe analysis (5). Moreover, in the present study, fractionation of microsomal membranes in Percoll gradients reveals a subpopulation of membranes, distinct from endoplasmic reticulum that fulfills a number of anticipated criteria for identification as endocytic vesicles (4): i.e. enrichment in 45Caspecific activity, CaBP, and a brushborder marker enzyme, alkaline phosphatase. In this regard, it is interesting to note that photoaffinity studies with CaBP and brushborder membranes have indicated an interaction between vitamin Dinduced calbindin and alkaline phosphatase (35). Further support for the concept of the tentatively identified endocytic vesicles comes from the initial observations made after 30 min of calcium transport in uiuo: at this time, the carrier vesicle fractions obtained from 1,25(OH)2D3treated chicks exhibit a depleted 45Ca content. This observation can be explained in terms of near completion of the lysosome vesicle fusion phase. Moreover, chloroquineinduced accumulation of 45Ca in these fractions also suggests that these vesicles are an acidic compartment, in analogy to endosomes which are known to undergo acidification prior to fusion with lysosomes (36, 37). The observation that peak lysosomal CaBP and 45Ca levels migrate in separate fractions lacks a ready explanation. It 1,25(0HIzD3 and Calcium Transport in Lysosomes may be that CaBP is not the actual calcium binding/recognition moiety, but rather functions in another aspect of transepithelial calcium transport in the intestine. Alternatively, the subpopulation of lysosomes found in fraction 3 may have a greater capacity to degrade CaBP. A third possibility is that CaBP is acquired during transport and/or fusion of the lysosomes at the basal lateral membrane. Further studies will be required to resolve this apparent discrepancy. Nevertheless, the postulated mechanism of calcium uptake across the brushborder membrane by endocytic vesicles, fusion of the vesicles with lysosomes, and ultimate delivery of the divalent cation to the basal lateral membrane by exocytosis is capable of explaining a number of earlier observations. For example, Nemere and Szego (7, 8) reported that low levels of 1,25 (OH)2D3 added to isolated intestinal cells of normal rats in uitro, resulted in enhanced lysosomal enzyme release, without loss of cell viability. These observations would be expected if calcium is delivered to the contralumenal membrane through exocytosis of lysosomal contents. Indeed, this mechanism of delivery could also account for the origins of serum CaBP (38, 39). The proposed mechanism of calcium internalization, endocytosis, also explains earlier observations on 1,25(OH)& mediated alterations in brushborder membrane lipids and stimulation of calcium uptake by brushborder membrane vesicles (4042). A number of workers have reported that endocytosis (and the closely coupled process of exocytosis) results in measurable alterations in phospholipid composition of the plasma membrane (4347). Thus it is conceivable that 1,25(OH)2D3induced changes in brushborder membrane phospholipids, whether in composition of fatty acyl side chains or polar head group moiety, is attributable to calcium transport by vesicular flow. Shultz et al. (48), have demonstrated that glucocorticoids at concentrations known to suppress 1,25(OH),D3mediated calcium absorption, fail to alter calcium uptake in brushborder membrane vesicles prepared from similarly treated animals. It is also well known that pharmacological levels of glucocorticoids stabilize lysosomal membranes (49, 50), further supporting the role of vesicular flow in vitamin D mediated intestinal calcium absorption. In the present study, evidence for the role of lysosomes in calcium transport in vivo was obtained under bioassay conditions with the lysosomeperturbing agent chloroquine, and structurally related analogues. Whereas 100 mm chloroquine completely abolished 1,25(OH)2D3mediated calcium transport, an equivalent concentration of the drug failed to suppress [3H]z0 absorption. Since the first stage of fluidphase pinocytosis (~60 min) largely bypasses the lysosomal compartment (51), this result is not surprising. Of equal interest is the finding that 60 mm primaquine also completely abolishes calcium absorption stimulated by the secosteroid, since recent work suggests that this drug analogue has little effect on the Golgi complex (52). Despite earlier electron micro scopic evidence and the current biochemical findings, it is still possible that calcium uptake does not occur through endocytosis. An alternative mechanism could involve transport of the cation through a brushborder channel with subsequent binding to a cytosolic carrier protein and ultimate delivery to lysosomes and microsomal membranes. Proof of this pathway will require isolation of the yet to be found brushborder calcium channel and a demonstration that the proposed cy tosolic carrier protects calciumsensitive cytoskeletal elements (1) from disruption. V. L. Leathers, unpublished observations. I. Nemere and V. L. Leathers, unpublished observations. AcknowledgmentsWe wish to thank Jungwhan Choi for technical assistance and Lean Gill for secretarial assistance. I. N. also wishes

9 ,25(0H)2D3 Calcium Transport and in Lysosomes to thank A. W. N. for his generous provision of laboratory space, 25. Bainton, D. F. (1973) J. Cell Biol. 58, equipment, and moral support that allowed her to pursue these ideas. 26. Muller, W. A., Steinman, R. M., and Cohn, Z. A. (1980) J. Cell Biol. 86, REFERENCES 27. Widnell, C.C., Schneider, Y. J., Pierre, B., Baudhuin, P., and 1. Nemere, I., and Norman, A. W. (1982) Biochim. Biophys. Acta Trouet, A. (1982) Cell 28, , de Duve, C., de Barsy, T., Poole, B., Trouet, A., Tulkens, P., and 2. Wasserman, R. H., and Taylor, A. N. (1966) Science 152, 791 Van Hoof, F. (1974) Biochem. Pharmacal. 23, Gorin, E., Gichting, G., and Goodman, H. M. (1984) Endocrinol 793 ogy 115, Sampson, H. W., Matthews, J. L., Martin, J. H., and Kunin, A. 30. Feher, J. J., and Wasserman, R. H. (1978) Biochim. Biophys. Acta S. (1970) Cakif. Tissue Res. 5, , Jande, S. S., and Brewer, L. M. (1974) 2. Anut. Entwicklungs 31. Feher, J. J., and Wasserman, R. H. (1979) Biochim. Biophys. Acta Gesch. 144, , Davis, W. L., Jones, R. G., and Hagler, H. K. (1979) Tissue & 32. Feher, J. J., and Wasserman, R. H. (1979) Endocrinology 104, Cell 11, Davis, W. L., and Jones, R. G. (1981) Tissue & Cell 13, Shimura, F., and Wasserman, R. H. (1984) Endocrinology 115, 7. Nemere, I., and Szego, C. M. (1981) Endocrinology 108, Szego, C. M., and Pietras, R. J. (1984) Int. Reu. Cytol. 88, Nemere, I., and Szego, C. M. (1981) Endocrinology 109, Norman, A. W., and Leathers, V. (1982) Biochem. Biophys. Res Commun. 108, Nemere, I., and Norman, A. W. (1985) in Vitamin D: Chemical, 36. Brown, M. S., Anderson, R. G. W., and Goldstein, J. L. (1983) Biochemical, and Clinical Update (Norman, A. W., Schaefer, Cell 32, K., Grigoleit, H.G., and Herrath, D. V., eds) pp , 37. Steinman, R. M., Mellman, I. S., Muller, W. A,, and Cohn, Z. A. Walter de Gruyter, Berlin (1983) J. Cell Biol. 96, Warner, R R., and Coleman, J. R. (1975) J. Cell Biol. 64, Christakos, S., and Norman, A. W. (1978) Science 202, Freedman, R. A., Weiser, M. M., and Isselbacher, K. J. (1977) 39. Bar, A., Maoz, A., and Hurwitz, S. (1979) FEBS Lett. 102, 79 Proc. Natl. Acad. Sci. U. S. A. 74, Freedman, R. A., MacLaughlin, J. A., and Weiser, M. M. (1981) 40. Rasmussen, H., Fontaine, O., Max, E. E., and Goodman, D. B. P. Arch. Biochem. Biophys. 206, (1979) J. Biol. Chem. 254, MacLaughlin, J. A., Weiser, M. M., and Freedman, R. A. (1980) 41. Matsumoto, T., Fontaine, O., and Rasmussen, H. (1981) J. Biol. Gastroenterology 78, Chem. 256, Rubinoff, M. J., and Nellans, H. N. (1985) J. Biol. Chem. 260, 42. Fontaine, O., Matsumoto, T., Goodman, D. B. P., and Rasmussen, H. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, Norman, A. W., and Wong, R.G. (1972) J. Nutr. 102, Sastry, P. S., and Hokin, L. E. (1966) J. Biol. Chem. 241, Severson, D. L., Denton, R. M., Bridges, B. J., and Randle, P. J. 44. Schroeder, F. (1981) Biochim. Biophys. Acta 649, (1976) Biochem. J. 154, Hook, V. Y. H., Heisler, S., and Axelrod, J. (1982) Proc. Natl. 17. Weiser, M. M., Neumeier, M. M., Quaroni, A., and Kirsch, K. Acad. Sci. U. S. A. 79, (1978) J. Cell Biol. 77, Bormann, B. J., Huang, C.K., Mackin, W. M., and Becker, E. L. (1984) Proc. Natl. Acad. Sci. U. S. A. 81, Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. 47. Camoratto, A. M., and Grandison, L.L. (1985) Endocrinology (1951) J. Biol. Chem. 193, , Putkey, J. A., Spielvogel, A. M., Sauerheber, R. D., Dunlap, C. 48. Shultz, T. D., Bollman, S., and Kumar, R. (1982) Proc. Natl. S., and Norman, A. W. (1982) Biochim. Biophys. Acta 688, Acad. Sci. U. S. A. 79, Ringrose, P. S., Parr, M. A., and McLaren, M. (1975) Biochem. 20. Pennington, R. J. (1961) Biochem. J. 80, Phurmacol. 24, Hubscher, G., and West, G. R. (1965) Nature 205, Lewis, D. A., and Krygier, H. E. (1977) J. Pharm. Sci. 66, Nemere, I., Putkey, J. A., and Norman, A. W. (1983) Arch Biochem. Biophys. 222, Swanson, J. A., Yirinec, B. D., and Silverstein, S. C. (1985) J. 23. Miller, B. E., and Norman, A. W. (1983) Methods Enzymol. 102, Cell Biol. 100, Strous, G. J., Du Maine, A., ZijderhandBleekemolen, J. E., Slot, 24. Hugon, J., and Borgers, M. (1967) J. Cell Biol. 33, J. W., and Schwartz, A. L. (1985) J. Cell Biol. 101,