Calcium transport systems in the LLC-PK1 renal epithelial established cell line

70

Biochimica et Biophysica Acta 888 (1986) 70-81

Elsevier BBA 11805

Calcium transport systems in the LLC-PK 1 renal epithelial established cell line

Jan B. Parys, Humbert De Smedt and Roger Borghgraef Laboratorium voor Fysiologie, Campus Gasthuisberg, Katholieke Universiteit Leuven, B-3000 Leuven (Belgium)

(Received April 18th, 1986)

Key words: Ca2+ transport; Inositol trisphosphate; (Epithelial cell) ATP-dependent calcium uptake was measured in membrane vesicles prepared from the renal epithelial LLC-PK t established cell line. The relative contribution of the nonmitochondrial versus the mitochondrial calcium uptake is larger in LLC-PK I cell homogenates than in homogenates from renal cortex. Two types of calcium pump, characterized by the formation of calcium-dependent phosphointermediates of 135 kDa and 115 kDa, were found in membrane fractions from LLC-PK t cells. The 135 kDa calcium pump was also detected by 12Sl-labelled calmodulin overlay. Although the subcellular localization in LLC-PKt cell membranes could not be unambiguously determined, it is conceivable that the 135 kDa and the 115 kDa molecules represent the plasma membrane calcium pump and the endoplasmic reticulum calcium pump respectively, in agreement with what was found for renal cortex preparations. Extravesicular sodium partially inhibits ATP-driven calcium uptake in a plasma-membrane-enriched fraction of the LLC-PK ! cells. The effect is potentiated by a vesicle inside-negative membrane potential. Although the effect is less pronounced than in renal cortex basal-lateral membranes, this observation suggests that an Na+-Ca 2+ exchange mechanism is also present in LLC-PKt cells. ATP-dependent calcium uptake in nonmitochondrial intraceilular stores was investigated, using saponin-permeabilized cells. Permeabilized LLC-PK t cells lowered the free calcium concentration in the medium to less than 0.4 p M. More than 60% of the accumulated calcium can be released by addition of inositoi 1,4,5-trisphosphate. Our data indicate that the LLC-PK t cell line can be successfully used as model system for the study of renal calcium handling.

Introduction There is growing interest in cell cultures as experimental model systems for the study of epithelial transport [1]. The renal epithelial established cell line LLC-PK 1, originally derived from pig kidney by Hull et al. [2], has been extensively studied. Several transport systems were described in LLC-PK1 cells: ( N a + + K+)-ATPase [3], the sodium-dependent hexose carrier [4,5], sodium-deAbbreviations: EGTA, ethylene glycol bis(fl-aminoethyl ether)-N,N,N',N'-tetraacetic acid; IP3, inositol 1,4,5-trisphosphate; PMSF, phenylmethylsulfonyl fluoride; SDS, sodium dodecyl sulfate.

pendent as well as sodium-independent amino-acid carriers [6-8], the sodium-dependent phosphate carrier [9,10], an N a + / H + exchanger [11], an H + / tetraethylammonium exchanger [12], an Na ÷, K +, 2C1--cotransport system [13] and a C 1 - / H C O 3 exchanger [14]. Calcium reabsorption is an important feature of the kidney [15]. Moreover, intracellular calcium regulates a variety of mechanisms, including transepithelial ion and water transport [16]. Recently Bonventre and Cheung [17] determined the cytosolic free calcium concentration in LLC-PK I cells. A value of 0.1 /~M was found. Measurements in the absence of Mg 2÷ and ATP indicated that mitochondria have a high calcium

0167-4889/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

71 sequestering capability, but that in basal conditions most of the intracellular calcium is located in non-mitochondrial stores. These stores were not further investigated [17]. Borle and co-workers have studied calcium fluxes in the LLC-MK 2 established cell line, originating from rhesus monkey kidney. Kinetic analysis of the fluxes revealed the importance of mitochondria in the calcium homeostasis of these cells [18]. An intracellular calcium concentration of 0.06 /xM was estimated, using aequorin [19]. Recently Snowdowne and Borle [20] have demonstrated the presence of an N a + / C a 2+ exchanger in the plasma membrane of LLC-MK 2 cells. An ATP-driven calcium pump was postulated as major calcium efflux pathway, but no evidence for its presence was given [20]. From experiments with membrane vesicles, isolated from renal cortex, evidence was obtained that both an ATP-driven calcium pump and an N a + / C a 2+ exchanger are present in the basallateral plasma membrane [21-26]. In this study we have investigated the mechanisms for calcium transport in the LLC-PK 1 cells. We have used membrane preparations from these cells to study ATP-driven calcium uptake, the c a l c i u m - i n d u c e d f o r m a t i o n of p h o s p h o i n termediates and calmodulin binding. It is shown that the LLC-PK 1 cells reflect very well the properties of renal cortical preparations with respect to the transport systems for calcium. Uptake of calcium in non-mitochondrial intracellular stores was studied using saponinpermeabilized cells. It is shown that calcium can be released from intracellular stores of LLC-PK~ cells by inositol 1,4,5-trisphosphate. Materials and Methods

Cell culture LLC-PK 1 cells (American Type Culture Collection CRL 1392/CL101, obtained through Flow laboratories Ltd., Irvine, Scotland) were used between passages 194 and 204. At regular intervals new cultures were initiated with cells stored in liquid nitrogen. Serial cultures were routinely maintained in 75 cm 2 plastic culture flasks at 37°C, in air, in following medium: Dulbecco's modified Eagle's medium with 20 mM Hepes (Cat.

No. 12-334, Flow Laboratories) supplemented with 6.4 mM N a H C O 3, 3.8 mM L-glutamine, 0.9% (v/v) non-essential amino acids, 85 I U / m l penicillin, 85 ~ g / m l streptomycin and 10% (v/v) fetal calf serum. The monolayers were subcultured every 7 days by trypsinization. Cell density at plating was (1-1.5)- 10 4 cells/cm 2 (split ratio 1/10). The cultures were fed with fresh medium three times a week. The cultures reached confluency on day 4 and 5 after subculturing. From the 5th day on, numerous domes were observed, indicating a high activity of transepithelial transport of solutes and water. Membrane vesicle preparations from L L C - P K 1 cells The cells were grown in Multitray Units (Nunc, Roskilde, Denmark) with an area of 6000 cm 2. The growth conditions were as described under 'Cell culture'. The 14th day after subculturing the cell monolayer was rinsed with 400 ml 0.02% (w/v) EDTA in normal saline (Cat. No. 23-203, Flow Laboratories) and the cells (109 cells) were harvested in 300 ml t r y p s i n / E D T A mixture (Cat. No. 16-891, Flow Laboratories). The trypsinization was stopped after 20-25 min with 450 ml complete growth medium containing 1 m g / m l turkey egg white trypsin inhibitor (Cat. No. T4385, Sigma Chemical Co.). By this procedure 97-99% of the cells were intact and viable, as judged from Trypan blue exclusion. The cells were sedimented (Sorvall GSA-rotor, 30 rain, 13000 rpm, 27600 X gmax) and resuspended in 20 ml homogenization medium containing 250 mM sucrose, 10 mM triethanolamine-HC1 (pH 7.6), 0.1 mM PMSF, 1 m g / m l soybean trypsin inhibitor (Cat. No. 109894, Boehringer). The cells were resuspended in the medium with a 18G-needle, followed by three strokes of a tight-fitting glassTeflon homogenizer at 1000 rpm. Homogenization was routinely performed by a brief sonication step (M.S.E. sonicator, equipped with an exponential probe, 20 kHz, amplitude 15/~m, 2 x 25 s). At this stage no intact cells could be detected with phasecontrast microscopy at 312 × magnification. The suspension was diluted to 70 ml with homogenization medium. Differential centrifugation was used for the separation of different membrane vesicle fractions. At each centrifugation step the pellet was saved and the supernatant was used for fur-

72 ther fractionation. All centrifugation steps were done with a Sorvall SS-34 rotor, except the 105000 X gmax centrifugation which was carried out with a Kontron T F T 45 rotor. Successively the fractions P1 (15 min, 4500 rpm, 2400 × gmax), P2 (15 rain, 9000 rpm, 9800×gma~) , P3 (20 min, 12000 rpm, 17400×gmax) , P4 (20 rain, 14000 rpm, 23600×gmax) , P5 (40 min, 17500 rpm, 36900 ×gmax) and P6 (60 min, 30000 rpm, 105 000 × gmax) were sedimented. The last supernatant (S) was also saved. The fractions P1-P6 were resuspended with a 23G-needle at a protein concentration of 5-10 m g / m l in homogenization buffer from which trypsin inhibitor was omitted. Homogenization of pig renal cortex with a glass-Teflon homogenizer and preparation of basal-lateral plasma membrafies was as previously described [26]. For marker enzyme analysis the homogenates were further treated with an UltraTurrax TP18-10 homogenizer (Janke & Kunkel KG. Ika Werk, F.R.G.) at 18000 rpm (3 × 30 s).

Assays The samples were assayed for following marker enzymes: (Na ÷ + K+)ATPase (EC 3.6.1.3): basallateral plasma membranes; -f-glutamyltransferase (EC 2.3.2.2), alkaline phosphatase (EC 3.1.3.1), aminopeptidase M (EC 3.4.11.2): apical plasma membranes; rotenone-insensitive N A D H : cytochrome c oxidoreductase (EC 1.6.99.3), glucose6-phosphatase (EC 3.1.3.9): endoplasmic reticulum; succinate dehydrogenase (EC 1.3.99.1), cytochrome c oxidase (EC 1.9.3.1): mitochondria. Succinate dehydrogenase was mesured according to Pennington [29]. ,&Glutamyltransferase was tested using an optimized kinetic test (Merckotest 14302, Merck A.G.). The other assays were done as previously described [26,30]. All enzyme activities were determined at 37°C except for alkaline phosphatase and y-glutamyltransferase (25°C). The proteins were determined by the method of Lowry et al. [31] after precipitation of the proteins in 10% (w/v) ice-cold trichloroacetic acid.

Calcium uptake in oesicles Uptake of calcium-45 was measured with a rapid filtration technique as described elsewhere [30]. The uptake conditions were optimized by including in the media 40 mM phosphate and 2

m g / m l albumin, except as stated otherwise. The media were buffered with 30 mM imidazole-HCl (pH 7.0 at 25°C), and contained 1 mM EGTA. Other components of the incubation media are noted in the legend to the figures. The concentration of ionized calcium was 2.7/~M and the concentration of ionized magnesium was 1 mM. The calculations of the calcium and magnesium concentrations were done, using stability constants corrected for temperature (37 ° C) and determined at an ionic strength of 0.1-0.15 [32]. The decimal logarithms of the association constants are: HEGTA, 9.23; H-HEGTA, 8.61; Ca-EGTA, 10.64; Ca-HEGTA, 5.29; Mg-EGTA, 5.41; Mg-HEGTA, 3.36; H-ATP, 6.54; H-HATP, 3.95; Ca-ATP, 3.74; Ca-HATP, 1.95; Mg-ATP, 4.23; Mg-HATP, 2.11; H-C204, 3.86; H-C204 H, 1.07; Ca-6204, 3.00; Mg-C204, 2.76; H-HPO4, 6.75; H-H2PO4, 2.05; Ca-HPO 4, 1.30; Mg-HPO 4, 1.80. The intravesicular medium consisted of 250 mM sucrose/10 mM triethanolamine-HCl (pH 7.6)/0.1 mM PMSF, except for the uptake experiments where sodium was used. In these experiments, vesicles were preloaded with 30 mM imidazole-HC1 (pH 7.0 at 25 o C ) / 1 0 0 mM KC1/100 mM sucrose/5 mM Tris-azide and in addition the vesicles were preincubated in this medium for 40 min on ice with 1.5 mM ouabain. Uptake was initiated by dilution of the membrane vesicles in the appropriate incubation medium.

Cell permeabilization experiments The LLC-PK 1 cells were maintained in culture in Singletray Units (Nunc) with an area of 600 cm 2. A week after subculturing, the cells were detached from the tray, using EDTA and trypsin and the trypsinization was stopped as described for the preparation of membrane vesicles. To remove the calcium, the procedure described by Streb and Schulz [27] was followed. The ceils were washed twice in a buffer comprising 135 mM KC1/1 mM MgCI2/1.2 mM K H 2 P O 4 / 1 0 mM Hepes (pH 7.4). Subsequently the cells were incubated for 30 min in the same buffer containing in addition the mitochondrial inhibitors Trisazide (10 mM) and oligomycin (1/~g/ml). During the incubation the cell suspension was gently agitated in an end-over-end rotator. Finally the cells were pelleted (Sorvall HS-4 rotor, 7 min,

73

1000 rpm, 190x gmax) and resuspended in the same medium at a final concentration of 9-10 7 cells/ml. About 20% of the cells became permeable for Trypan blue after the washing procedure. Calcium flux studies were carried out at 25°C. The medium is given in the legend to Fig. 8. It was verified that 98% permeabilization was obtained by incubation with saponin (50 btg/ml) for 10 min at 25°C. The free calcium concentration was recorded with a calcium-sensitive electrode. The preparation of the calcium electrode was according to Affolter and Sigel [28], the calibration was as described by Streb and Schulz [27]. Each electrode was used for 1 week. Inositol 1,4,5-trisphosphate was obtained from Sigma Chemical Co.

Phosphorylation experiments The formation of phosphorylated intermediates in the presence of ['y-32p]ATP, subsequent analysis by SDS-polyacrylamide gradient gel electrophoresis at low pH and autoradiography were done as previously described [33].

A

B

C

/

/

% %

~ s

/ ~o--o--o--o 2

/ o~O--O--O

4

/

/

~ j l J A~A

~ o

/

2

/

o ~o~o-o

4.

T~me

D

2

/

/

.o~O_O_ ¢

/ o

2

:ram J

Fig. 1. Calcium uptake in LLC-PK~ cell homogenates. Homogenization was performed with a Polytron homogenizer type PT 10/35, setting 4, 2 × 2 min (A) or with an M.S.E. probe sonicator, 20 kHz, 15 /~m amplitude for 30 s (B), 2 x 30 s (C), or 5 × 30 s (D). 45Ca uptake was assayed by a rapid filtration technique. Incubation temperature was 37°C. All experiments were done in duplicate. The basic incubation medium contained 30 m M imidazole-HCl (pH 7.0 at 25°C), 100 m M KC1, 100 m M sucrose, 5 m M Tris-azide, 1 m M K-EGTA, 1 m M ionized Mg 2÷ and 2.7 /~M ionized Ca 2÷ (O). The basic medium was supplemented with 5 m M Tris-ATP, 40 m M phosphate and 2 m g / m l albumin (A).

Calmodufin binding ~25I-labelled calmodulin overlay and densitometric scanning of the autoradiograms were effectuated as described previously [30]. Results

Calcium uptake in membrane vesicles from LLCPK/ cells Calcium uptake in homogenates of LLC-PK~ cells was strikingly dependent on the homogenization procedure (Fig. 1). ATP-dependent uptake was measured in the presence of 5 mM azide, in order to avoid uptake in mitochondria. Optimal uptake conditions were obtained by inclusion in the medium of 2 m g / m l albumin and 40 mM phosphate [30]. Homogenates were obtained from the same lot of cells by Polytron homogenization (A) or by sonication for different time periods (B-D). Calcium uptake was much lower in the Polytron-treated homogenate. Sonication for long time periods (more than 1 min) also tended to decrease the calcium pump activity. The activity of marker enzymes was very similar in all fractions, except for a slight decrease of (Na + + K + ) -

ATPase and alkaline phosphatase activity after sonication for longer than 1 min (data not shown). Since by microscopic analysis cell disruption was observed in the Polytron-treated sample as well as in the sonicator-treated samples, differences in uptake properties are probably due to differences in the degree of vesiculation or in leakiness of the vesicles. Calcium uptake in a LLC-PK~ homogenate was compared with uptake in an homogenate of pig renal cortex (Fig. 2). The cortical tissue was disrupted with a glass-Teflon homogenizer. The ATP-independent calcium uptake amounted to 0.5 nmol calcium/rag protein in the renal cortex homogenate, and to 0.7 nmol calcium/mg protein in the LLC-PK1 homogenate. These values were not affected by the presence of phosphate and albumin. ATP induces a time-dependent calcium accumulation. In the presence of mitochondrial inhibitors (azide or azide plus oligomycin) the calcium accumulation, measured after 9 min, was 15- and 30-times higher than the uptake in the absence of ATP, for the renal cortex homogenate and the LLC-PK 1 homogenate, respectively. The

74 A

B

TABLE I ACTIVITIES OF MARKER ENZYMES AND OF ATP-DEPENDENT, AZIDE-INSENSITIVE CALCIUM UPTAKE

20

•

/

0

.

i .

, lI

z L/ S , ,,. -

= ~ - -

5 Tithe

( mm

i

Fig. 2. Calcium uptake in homogenates from pig renal cortex (A) and LLC-PKl cells (B). Calcium uptake was measured as described in the legend to Fig. 1. Calcium uptake was measured under different medium conditions. ATP-independent uptake was measured in the basic medium (O) and in the basic medium supplemented with 40 mM phosphate and 2 mg/ml albumin (13).Calcium uptake remained at the same level in the presence of 5 mM Tris-ATP, 40 mM phosphate and 2 mg/ml albumin if 5/zg/ml calcium ionophore (A23187) was included (v). ATP-dependent calcium accumulation is observed in the basic medium supplemented with 5 mM Tris-ATP, 40 mM phosphate and 2 mg/ml albumin (t). In one experiment A23187 was added in course of the experiment (arrow) and uptake value ,after 9 min was measured (Q). Uptake was also measured with 5 mM Tris-ATP, 40 mM phosphate, 2 mg/ml albumin and 4 ~g/ml oligomycin added to the basic medium (+), or in the same medium from which Tris-azide and oligomycin were omitted ( x ).

calcium ionophore A23187 abolished the ATP-dep e n d e n t uptake. The calcium accomulated in the vesicles is rapidly released by addition of A23187. W h e n azide a n d oligomycin are omitted from the i n c u b a t i o n m e d i u m , the A T P - d e p e n d e n t calcium uptake is m u c h less stimulated in the LLC-PK1 cell h o m o g e n a t e (150% stimulation) than in the renal cortex h o m o g e n a t e (650%). The activity of different marker enzymes a n d the initial rate of calcium uptake in the h o m o genate of L L C - P K 1 cells a n d the h o m o g e n a t e of renal cortex are listed in Table I. N A D H : c y tochrome c oxidoreductase a n d ~{-glutamyltransferase are a b o u t equally active in L L C - P K 1 cell h o m o g e n a t e s as in renal cortex homogenates. The other marker enzymes exhibited a lower activity in the LLC-PK~ homogenate: the activities of the m i t o c h o n d r i a l enzymes are a b o u t 2-times lower; ( N a ÷ + K + ) - A T P a s e a n d alkaline phosphatase are a b o u t 4-times lower; a m i n o p e p t i d a s e M a n d glu-

Renal cortex homogenates are compared with LLC-PK1 homogenates. The enzyme activities were determined as described in Materials and Methods and expressed in mU/mg protein. The initial uptake rate of calcium was measured as previously described [30] and is expressed in nmol Ca2+/mg protein per min. All values are mean activities+ S.E. (number of preparations). Pig renal cortex LLC-PK1 (Na+ + K+ )-ATPase 99 +10(9) y-Glutamyltransferase 436 + 48 (3) Alkaline phosphatase 761 +43 (10) Aminopeptidase M 153 + 7 (9) NADH : cytochrome c oxidoreductase 223 + 18 (9) Glucose-6-phosphatase 186 +22 (5) Succinate dehydrogenase 44 + 5 (3) Cytochrome c oxidase 281 + 25 (8) ATP-dependent, azideinsensitive calcium uptake 0.52 + 0.04 (4)

24 499 210 14

+ 2(5) + 65 (6) + 34 (3) + 2 (5)

176 8 16 147

+ 15 (7) _+ 1 (4) + 1 (5) + 39 (5)

2.05_+ 0.03 (3)

cose-6-phosphatase are more than 10-times lower. The A T P - d e p e n d e n t , azide-insensitive calcium uptake activity in contrast was 4-times higher in the L L C - P K 1 cell h o m o g e n a t e than in the r e n a l cortex homogenate.

Characterization of the calcium transport systems The cell h o m o g e n a t e was fractionated by differential centrifugation as described u n d e r Materials a n d Methods. The fractions were analysed for m a r k e r enzyme activities a n d calcium uptake rate. T h e results are s u m m a r i z e d in Fig. 3. The m e a n e n r i c h m e n t factor is plotted against the m e a n relative protein c o n t e n t [34]. The activity of the marker enzymes is u n e v e n l y distributed over all fractions. The basal-lateral p l a s m a m e m b r a n e marker, (Na ÷ + K + ) - A T P a s e , is f o u n d preferentially in the P 1 - P 3 fractions, whereas the e n d o p l a s m i c reticul u m marker glucose-6-phosphatase is more enriched in the P 4 - P 6 fractions. The calcium uptake activity, measured in the presence of A T P and azide, clearly d e m o n s t r a t e d two maxima, one in P1, the other in the P 5 - P 6 fraction. I n Fig. 4, the azide-insensitive calcium uptake in the P1 a n d P5 fractions is shown for different

75

p

i ' I z is#isl~

conditions. The ATP-independent calcium uptake amounted to 1 nmol calcium/mg protein in both fractions. ATP stimulated the calcium uptake in both fractions up to about 30 nmol calcium/mg protein after 40 min. The uptake was markedly stimulated in the presence of calcium-precipitating anions. In the P1 fraction, phosphate was much more effective than oxalate; in the P5 fraction, both anions stimulated the calcium uptake to the same extent. A biochemical characterization of the calcium pump molecules in the P1 and the P5-P6 fractions was performed by the analysis of calcium-induced phosphoprotein formation and by 125I-labeled calmodulin overlay. The results of a typical phosphorylation experiment are shown in Fig. 5. The homogenate (lanes 1-3), the P1 fraction (lanes 4-6) and the P5 fraction (lanes 7-9) were compared. The membranes were phosphorylated in different conditions and thereafter the phosphoproteins were analysed by SDS-polyacrylamide gel electrophoresis on a gradient gel at low pH. A Laemmli-type gel electrophoresis can not be used, since the phosphointermediates of transport ATPases are acyl phosphates which are labile at alkaline pH [26]. The autoradiogram of a gel is shown. All

5 I

A

o

1

4

B

o

I

Eo

2 0 0

Proteins

5O

[ %)

Fig. 3. Distribution of the marker enzymes ( N a + + K + ) ATPase (A), glucose-6-phosphatase (B), y-glutamyltransferase (C), succinate dehydrogenase (D) and ATP-dependent, azideinsensitive calcium uptake (E) in the membrane fractions (P1P6, S). Calcium uptake was measured as described in Fig. 1. The mean enrichment factor (specific activity in fraction/ specific activity in homogenate) is plotted against the mean relative protein content (protein content of fraction/protein content of homogenate) as described by De Duve et al. [34].

150

~m

100

J" 50

j

~ o - - o

,

20

o

o

?

20

40 TJ me

( rain )

40

Fig. 4. Effect of precipitating anions on calcium uptake in the P1 fraction (A) and the P5 fraction (B). Calcium uptake was measured as described in the legend to Fig. 1. The basic medium (O) was supplemented with 5 mM Tris-ATP and 2 m g / m l albumin (O); 5 mM Tris-ATP, 5 mM oxalate and 2 m g / m l albumin (It) or 5 mM Tris-ATP, 40 mM phosphate and 2 m g / m l albumin (A).

76

H 1

23

P1 ,~5

P5 6

78

9

S

Mr

135 kOa llS kDa

Fig. 5. Autoradiogram of calcium-induced 32p-labeled phosphoproteins. The phosphoproteins were separated by SDSpolyacrylamide gradient gel electrophoresis at pH 2.4 [33]. The homogenate (lanes 1-3), the P1 fraction (lanes 4-6) and the P5 fraction (lanes 7-9) were examined. The phosphorylations were done for 10 s at 0°C, by addition of 6 btM [y-32p]ATP (13 Ci/mmol) to an incubation medium containing 30 mM imidazole-HCl (pH 7.0 at 25°C), 100 mM KCI, 5 mM flglycerophosphate and in addition 1 mM K-EGTA (lanes 1, 4, 7); 50 /~M CaCI 2 (lanes 2, 5, 8) or 50 /~M CaC12 +50 #M LaC13 (lanes 3, 6, 9). The protein concentration of the samples was 1.25 mg/ml, fl-Glycerophosphate was added in order to suppress the phosphorylation of alkaline phosphatase.

fractions contained two calcium-induced phosphoproteins with M r of 115 kDa and 135 kDa, respectively. The formation of these phosphoproteins is strongly reduced in the presence of EGTA (lanes 1,4,7). The formation of the 135 kDa phosphoprotein is maximally stimulated when 50 /~M lanthanum is included in the medium. The 135 kDa phosphoprotein is located predominantly in the P1 fraction (lanes 5,6). The formation of the 115 kDa phosphoprotein, on the other hand, is depressed in the presence of lanthanum. This phosphoprotein is predominantly located in the P5 fraction (lanes 8,9). In agreement with the data obtained for renal cortex preparations [26,30] these observations indicate that two types of calcium pump are present in the LLC-PK~ cells. The presence of calmodulin-binding proteins

was investigated. The membrane proteins were separated by SDS-polyacrylamide gel electrophoresis, and subjected to 125I-labelled calmodulin overlay, after electroblotting (Fig. 6). The major calmodulin-binding polypeptide observed by this technique migrated with an M r of 135 kDa (A). Densitometric scans of the autoradiograms are shown for homogenate, P1 and P6 fractions (B-D). The 135 kDa calmodulin-binding polypeptide is preferentially distributed in the P1 fraction. A similar 135 kDa calmodulin-binding protein was also detected in renal basal-lateral plasma membranes [33,35] and was identified with the plasma membrane calcium pump molecule.

The effect of sodium Calcium uptake was measured in a purified basal-lateral preparation from renal cortex and in a P1 fraction from LLC-PK 1 cells (Fig. 7). The effect of extravesicular sodium and of the membrane potential was investigated. The K+-iono A S_

135 kOa

8

135 kDa

1

--

/

I O

.

2

A

Fig. 6. 125I-labeled calmodulin binding in a LLC-PK 1 homogenate (A), and densitometric scannings of the relevant part of the autoradiograms for the same homogenate (B) the P1 fraction (C) and the P6 fraction (D). The samples were analysed by Laemmli-type SDS-polyacrylamide gel electrophoresis. 1251labeled calmodulin binding was detected by overlay after electroblotting to nitrocellulose [30].

77

phore valinomycin creates an intravesicular negative transmembrane potential when K ÷ is present inside the vesicle and Na ÷ is present in the extravesicular medium. Calcium uptake in the vesicles is higher in a K÷-containing medium than in an Na÷-containing medium. When valinomycin is added, calcium uptake in the basal-lateral membrane vesicles is slightly increased ( + 10%) with K ÷ in the medium, but markedly decreased ( - 4 0 % ) with Na ÷ in the medium (A). In the P1 fraction, isolated from LLC-PK 1 cells, the effects of valinomycin are much smaller. In the presence of K ÷, valinomycin did not stimulate the calcium uptake. In the Na+-containing medium a small valinomycin-induced inhibition ( - 15%) of the uptake was observed (B).

The effect of inositol 1,4,5-trisphosphate The presence of the 115 kDa calcium pump molecule and the oxalate sensitivity of the calcium uptake in the microsomal (P5-P6) fractions of LLC-PK] cells suggest that ATP-dependent calcium accumulation partly occurs in an intracellular compartment, presumably the endoplasmic reticulum. We have assessed calcium up-

take in and release from this compartment by using permeabilized LLC-PK 1 cells. As described in Materials and Methods, the plasma membrane was permeabilized with saponin. Calcium uptake in mitochondria was excluded by the absence of mitochondrial substrates and the presence of azide and oligomycin. Calcium sequestration in the cells was followed at 25°C by a calcium-sensitive macroelectrode. As shown in Fig. 8, addition of the cells resulted in a decrease of the concentration of ionized calcium in the medium to about 0.4 /~M. The uptake was completely abolished when 5 mM vanadate was included in the medium (data not shown). The amount of calcium taken up in the cells was calculated by calibration of the electrode response with known amount of CaC12. Uptake was about 5 nmol c a l c i u m / m g cell protein. A pulse addition of 6 /~M IP3 resulted in a rapid release of 60-70% of the accumulated

cells

5.~, B

/o

_g

~P3

-j

c6 }~M )

¢

5.6

o / ° 5.8

q,

6.0

6.2 I

I

25

5 TIME

/

L

i

25

5

(mtn)

i

6.~ I

Fig. 7. The effect of sodium and membrane potential on calcium uptake in pig renal basal-lateral plasma membrane vesicles (A) and the P1 fraction of LLC-PK 1 cells (B). Calcium uptake was measured as indicated in the legend to Fig. 1. Preincubation with KCI and ouabain was as described in Materials and Methods. Incubation was in 30 m M imidazoleHC1 (pH 7), 100 m M KCI, 100 m M sucrose, 5 m M Tris-azide, 1 m M K-EGTA, 1 m M ionized Mg 2+ and 2.7 # M ionized Ca 2÷ ( × ) . To this medium 5 m M Tris-ATP was added, without ( © ) or with 20 # g / m l valinomycin (e). To test the effect of Na ÷, KCI in the outside medium was replaced by NaCI, in the absence (O) or in the presence of valinomycin (I).

I

10 rain 6.6~Fig. 8. Calcium uptake in permeabilized LLC-PK 1 cells. Calcium was measured at 2 5 ° C with a calcium-sensitive electrode. The incubation medium (1 ml) consists of 130 m M KC1, 4 m M MgCI2, 1.2 m M KH2PO4, 25 m M Hepes (pH 7.4), 3 m M Tris-ATP, 10 m M creatine phosphate (Tris-salt), 8 U / m l creatine kinase, 1 / ~ g / m l oligomycin, 10 m M Tris-azide and 50 # g / m l saponin. Addition of cells (final concentration 17.10 6 cells/ml or 6.5 mg cell protein/ml) and of inositol 1,4,5-trisphosphate (6 # M ) is indicated by arrows.

78 calcium. Reuptake followed immediately. The IP~-induced release could be repeated several times with only slight attenuation of the effect, and each release was followed by reuptake. When successive IP~ additions were made at short time intervals, the concentration of medium free calcium remained at an elevated level. Reuptake resumed again following the final addition of IPs. Discussion

The cell material The LLC-PK l cell line is widely used for the study of renal transport mechanisms [1]. In order to obtain large amounts of cells (up to 10 9 cells), the LLC-PK~ cells used in this study were grown in Singletray and Multitray Units. This technique implies the use of a trypsinization procedure for cell harvesting. It was previously reported that trypsinization of the LLC-PK~ monolayer in the presence of EDTA was harmful for alkaline phosphatase and to a lesser extent for 7-glutamyltransferase [36]. In order to prevent excessive trypsinization an excess of trypsin inhibitor was used throughout the preparation. The activity of the marker enzymes measured in the LLC-PK~ homogenates (Table I) is not lower than the values obtained by other authors [37-40]. As can be observed in Table 1, the activities of all enzymes, except y-glutamyltransferase and N A D H : c y t o c h r o m e c oxidoreductase, are much lower in the LLC-PK 1 cell homogenate than in the homogenate of renal cortex. This is in agreement with the observation that the activity of several enzymes is reduced when primary renal cells are brought into culture [41]. A TP-driuen calcium uptake We report here the presence of ATP-dependent calcium transport mechanisms in LLC-PK~ cell membranes. The highest calcium uptake rate was observed when cells are disrupted by a brief sonication step (Fig. 1). Sonication was used in several studies [36,42,43] for the determination of enzyme activities, but not prior to flux measurements in vesicles. The strong influence of the homogenization procedure probably reflects a difference in vesiculation. This indicates that it is very difficult to compare flux rates in membrane preparations

obtained by different methods. The results in Fig. 2 therefore merely illustrate that ATP-driven calcium uptake is present in LLC-PK 1 cell membranes, as well as in membranes derived from renal cortex. The ATP-dependent calcium uptake is totally abolished by the calcium ionophore A23187. This indicates that calcium is taken up in the intravesicular compartment. In contrast to the renal cortex preparation, a comparatively small part of the ATP-driven calcium uptake in the LLC-PK1 cell homogenate is azide- and oligomycin-sensitive. This observation, taken together with the relatively low activity of mitochondrial enzymes in LLC-PK~ cell homogenates and the conclusion of Bonventre and Cheung [17] that under basal conditions, most of the intracellular calcium is located in nonmitochondrial stores, suggests that the role of the mitochondria in the calcium homeostasis of the LLC-PK~ cells could be less pronounced than in LLC-MK 2 cells [18] or in proximal tubule cells [44].

Characterization of the A TP-dependent calcium transport systems Azide-insensitive, ATP-dependent calcium uptake was described in both the plasma membrane and the endoplasmic reticulum of several cell types, including renal epithelial cells [30,45]. In renal proximal tubule cells the ATP-driven plasma membrane calcium pump is exclusively located at the basolateral cell pole [21,23,26,30,33,35]. In this study, calcium uptake was studied in different membrane fractions, isolated by differential centrifugation from the homogenate of LLCPK 1cells. According to marker enzyme analysis, a crude separation was obtained between fractions enriched in ( N a + + K+)-ATPase, a marker for the basal-lateral plasma membrane (P1-P3), and fractions enriched in glucose-6-phosphatase, a marker of the endoplasmic reticulum (P5-P6) (Fig. 3). Unfortunately, it was not possible to purify the plasma membrane fractions further as to obtain pure basal-lateral membranes. Attempts with sucrose or Percoll density gradients were not successful. It is however, clear from the data presented in Fig. 3 that ATP-dependent, azide-insensitive calcium transport is distributed over two different membrane fractions (P1 and P5-P6). Calcium-precipitating anions stimulate calcium

79 accumulation in membrane vesicles [50]. Calcium uptake in fractions enriched in basal-lateral plasma membranes is more stimulated by phosphate than by oxalate, in the fractions enriched in endoplasmic reticulum both anions are equally effective (Fig. 4). This observation suggests that calcium is taken up in two different types of membrane vesicle, having a different permeability for oxalate. Similar results were also obtained in renal tissue [30] and in non-epithelial tissues, such as smooth muscle [46] and were explained by the fact that plasma membranes are generally much less permeable for oxalate than the membranes from the endoplasmic reticulum. Analysis of the formation of phosphorylated intermediates is a very sensitive and specific method for identification of the calcium pump molecules of both endoplasmic reticulum and plasma membranes. The calcium pump in the endoplasmic reticulum of several cell types is characterized by a 115 kDa phosphoprotein, the formation of which is inhibited by lanthanum. The calcium pump in the plasma membrane, on the other hand, forms a 135 kDa phosphointermediate, the formation of which is stimulated by lanthanum [26,30,47]. As shown in Fig. 5, we can identify both types of phosphoprotein in the LLCPK 1 membrane preparations. The relative enrichment of both phosphoproteins in the membrane fractions is consistent with the interpretation that the 115 kDa and the 135 kDa calcium pumps are present in the endoplasmic reticulum and the plasma membranes of the LLC-PK~ cells, respectively. The plasma membrane calcium pump of most cell types is a calmodulin-regulated enzyme [48]. the calcium pump in the basal-lateral membrane preparations from renal cortex is known to have a calmodulin-stimulated ATPase-activity [22,23,35, 49] and can be detected as a calmodulin-binding protein [33,35]. As shown by 1251-labelled calmodulin overlay (Fig. 6), a 135 kDa calmodulin-binding polypeptide is found preferentially in the plasmamembrane fractions of the LLC-PK~ cells. This confirms the presence of the 135 kDa calcium pump in the LLC-PK 1 plasma membranes.

Presence of an Na ÷/ C a " + exchanger The inhibition by sodium of the ATP-dependent calcium accumulation in renal membranes [45] was interpreted as the expression of an N a + / Ca 2+ exchanger, present in the basal-lateral membranes [21,23,35]. The presence of an N a + / C a 2+ exchanger has been further documented by Van Heeswijk et al. [23], Jayakumar et al. [24] and Scoble et al. [25]. Since an ATP-dependent calcium pump is present in the same membranes [21, 23,26,30,35], the function of the N a + / C a 2+ exchanger was questioned [23]. From studies with LLC-MK 2 cells it was proposed that the N a + / Ca 2+ exchanger could serve for calcium influx [20], but the problem is not yet settled. The effect of sodium on the ATP-driven calcium uptake is explained by the fact that the N a + / C a 2+ exchanger can release the calcium accumulated in the vesicles by the calcium pump. As pointed out by SchiSnfeld et al. [51] the possibility can not be excluded that part of the calcium release is due to competition between Na ÷ and Ca 2 + for intravesicular binding sites. The observation, however, that an inside-negative transmembrane potential in the presence of an Na ÷ gradient decreased the calcium accumulation (Fig. 7) is consistent with the presence of an electrogenic N a + / C a 2+ exchanger ( 3 N a + / 1 C a 2+) carrying a net positive charge inward. The electrogenicity of the renal N a + / C a 2+ exchanger was recently reported [25], and is in agreement with the observations in heart tissue

[52]. The presence of ouabain or replacement of extracellular sodium with choline or lithium resuited in a relatively modest increase in intracellular free calcium in LLC-PK 1 cells. Na+-Ca 2+ exchange could provide an explanation for this observation [17]. It was shown that parathyroid hormone is an important modulator of the N a ÷ / Ca 2÷ exchanger activity in rat [24] and canine [25] renal cortical preparations. Further work is needed to see whether different growth conditions can induce a more pronounced expression of the N a + / C a z+ exchanger.

The effect of inositol 1,4,5-trisphosphate Inositol 1,4,5-trisphosphate, a product of the hydrolysis of phosphatidylinositol 4,5-bisphosphate, has been shown to mobilize calcium from

80

an intracellular store in a wide variety of tissues, as recently reviewed in several papers [53,54,60]. ATP-dependent calcium uptake in permeabilized LLC-PK~ cells occurs in a vanadate-sensitive, non-mitochondrial compartment. Our data suggest that this compartment represents the endoplasmic reticulum. The IP3-induced release of the accumulated calcium lends further supports to this view, in agreement with observations in other cell types, e.g., pancreatic cells [55] where the role of endoplasmic reticulum in IP3-induced calcium release was demonstrated. In permeabilized LLC-PK 1 cells more than 60% of the accumulated calcium was released by IP3, even in the presence of an operating calcium influx pump. It has been suggested that only a part of the endoplasmic reticulum is sensitive to IP3 [53,54]. This, however, depends on the cell type. For RINm5F insulinoma cells, it was demonstrated that IP3 released almost all the calcium accumulated in the endoplasmic reticulum [57]. In similar experiments on permeabilized neutrophils [58] or hepatocytes [54], only 25-50% of the calcium was released. The transient effect of IP3 is very well documented, e.g., Ref. 57. The major mechanism for terminating IP3 action is thought to be the removal of the 5-phosphate by a family of specific phosphatases [59]. Our observation that high calcium levels in the medium could be maintained by successive additions of IP3, indicates that no desensitization occurs and that the transient effect is due to IP3 breakdown. Our experiments therefore suggest that LLC-PKa cells not only contain the IP 3 receptor in the endoplasmic reticulum, but also contain the phosphatases responsible for IP3 breakdown. In a recent study [56] IP3-induced calcium release was demonstrated in permeabilized rat cortical kidney cells. Although rat kidney cortex is a heterogeneous tissue, it was proposed that the IP3 effect occurs primarily in proximal tubule cells. Our study therefore confirms that LLC-PK1 cells exhibit many properties found in the cells of the proximal tubule. In the same study [56] it was also demonstrated that addition of angiotensin IIamide to the saponin-treated kidney cells resulted in calcium release. It remains to be established whether hormone-induced changes in the phos-

phoinositide metabolism and c o n c o m i t a n t Ca2+-mobilization also occur in LLC-PK 1 cells. In conclusion, we have demonstrated that the same types of calcium transport molecule which were described for renal cortex preparations are also present in LLC-PK1 cells. The finding that IP3-induced calcium release occurs in LLC-PK~ cells suggests that this cell line provides a very useful model system, not only for the study of renal calcium handling and its regulation, but also for the study of hormone-induced intracellular calcium mobilization.

Acknowledgements We thank Mrs. M. Crabb~ and Miss L. Bauwens for excellent technical assistance. We are indebted to Prof. Dr. J.-J. Cassiman, Mr. B. Staf and Mrs. G. Deferme (Center for Human Genetics, K.U. Leuven) for valuable help and advice and to Dr. K. Amsler and Dr. D.M. Scott (Max-Planck-Institut for Systemphysiologie, Dortmund, F.R.G.) for stimulating discussions. This work was supported by grant No. 3.0042.83 from the Fonds voor Geneeskundig Wetenschappelijk Onderzoek (Belgium).

References 1 Handler, J.S. (1983) J. Exp. Biol. 106, 55-69 2 Hull, R.N., Cherry, W.R. and Weaver, G.W. (1976) In Vitro 12, 670-677 3 Mills, J.W., Macknight, A.D.C., Dayer, J.-M. and Ausiello, D.A. (1979) Am. J. Physiol. 236, C137-C162 4 Mullin, J.M., Weibel, J., Diamond, L. and Kleinzeller, A. (1980) J. Cell. Physiol. 104, 375-389 5 Rabito, C.A. and Ausiello, D.A. (1980) J. Membrane Biol. 54, 31-38 6 Sepulveda, F.V. and Pearson, J.D. (1982) J. Cell. Phys. 112, 182-188 7 Rabito, C.A. and Karish, M.V. (1982) J. Biol. Chem. 257, 6802-6808 8 Rabito, C.A. and Karish, M.V. (1983) J. Biol. Chem. 258, 2543-2547 9 Rabito, C.A. (1983) Am. J. Physiol. 245, F22-F31 10 Biber, J., Brown, C.D. and Murer, H. (1983) Biochim. Biophys. Acta 735, 325-330 11 Haggerty, J.G., Cragoe, E.J., Jr., Slayman, C.W. and Adelberg, E.A. (1985) Biochem. Biophys. Res. Commun. 127, 759-767 12 Inui, K.-I., Saito, H. and Hori, R. (1985) Biochem. J. 227, 199- 203

81 13 Brown, C.D.A. and Murer, H. (1985) J. Membrane Biol. 87, 131-139 14 Chaillet, J.R., Amsler, K. and Boron, W.F. (1986) Proc. Natl. Acad. Sci. USA 83, 522-526 15 Agus, Z.S. and Goldfarb, S. (1985) in The Kidney: Physiology and Pathophysiology (Seldin, D.W. and Giebisch, G., eds.), pp. 1323-1335, Raven Press, New York 16 Taylor, A. and Windhager, E.E. (1985) in The Kidney: Physiology and Pathophysiology (Seldin, D.W. and Giebisch, G., eds.), pp. 1297-1322, Raven Press, New York 17 Bonventre, J.V. and Cheung, J.Y. (1986) Am. J. Physiol. 250, F329-F338 18 Bode, A.B. (1972) J. Membrane Biol. 10, 45-66 19 Borle, A.B. and Snowdowne, K.W. (1982) Science 217, 252-254 20 Snowdowne, K.W. and Borle, A.B. (1985) J. Biol. Chem. 260, 14998-15007 21 Gmaj, P., Murer, H. and Kinne, R. (1979) Biochem. J. 178, 549-557 22 De Smedt, H., Parys, J.B., Borghgraef, R. and Wuytack, F. (1981) FEBS Lett. 131, 60-62 23 Van Heeswijk, M.P.E., Geertsen, J.A.M. and Van Os, C.H. (1984) J. Membrane Biol. 79, 19-31 24 Yayakumar, A., Cheng, L., Liang, C.T. and Sacktor, B. (1984) J. Biol. Chem. 259, 10827-10833 25 Scoble, J.E., Mills, S. and Hruska, K.A. (1985) J. Clin. Invest. 75, 1096-1105 26 De Smedt, H., Parys, J.B., Borghgraef, R. and Wuytack, F. (1983) Biochim. Biophys. Acta 728, 409-418 27 Streb, H. and Schulz, I. (1983) Am. J. Physiol. 245, G347-G357 28 Affolter, H. and Sigel, E. (1979) Anal. Biochem. 97, 315-319 29 Pennington, R.J. (1961) Biochem. J. 80, 649-654 30 Parys, J.B., De Smedt, H., Vandenberghe, P. and Borghgraef, R. (1985) Cell Calcium 6, 413-429 31 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 32 Martell, A.E. and Smith, R.J. (1982) Critical Stability Constants, Plenum Press, New York 33 De Smedt, H., Parys, J.B., Wuytack, F. and Borghgraef, R. (1984) Biochim. Biophys. Acta 776, 122-132 34 De Duve, C., Pressman, B.C., Gianetto, R., Wattiaux, R. and Appelmans, F. (1955) Biochem. J. 60, 604-617 35 Gmaj, P., Zurini, M., Murer, H. and Carafoli, E. (1983) Eur. J. Biochem. 136, 71-76

36 Rabito, C.A., Kreisberg, J.I. and Wight, D. (1984) J. Biol. Chem. 259, 574-582 37 Moran, A., Handler, J.S. and Turner, R.J. (1982) Am. J. Physiol. 243, C293-C298 38 Brown, C.D., Bodmer, M., Biber, J. and Murer, H. (1984) Biochim. Biophys. Acta 769, 471-478 39 Inui, K.-I., Saito, H., Takano, M., Okano, T., Kitazawa, S. and Hori, R. (1984) Biochim. Biophys. Acta 769, 514-518 40 Gstraunthaler, G., Pfaller, W. and Kotanko, P. (1985) Am. J. Physiol. 248, F536-F544 41 Burlington, H. (1959) Am. J. Physiol. 197, 68-70 42 Sepulveda, F.V., Burton, K.A. and Pearson, J.D. (1982) Biochem. J. 208, 509-512 43 Yoneyama, Y. and Lever, J.E. (1984) J. Cell. Physiol. 121, 64-73 44 Mandel, L.J. and Murphy, E. (1984) J. Biol. Chem. 259, 11188-11196 45 Moore, L., Fitzpatrick, D.F., Chen, T.S. and Landon, E.J. (1974) Biochim: Biophys. Acta 345, 405-418 46 Raeymaekers, L., Wuytack, F., Eggermont, J., De Schutter, G. and Casteels ,R. (1983) Biochem. J. 210, 315-322 47 Wuytack, F., Raeymaekers, L., Verbist, J., De Smedt, H. and Casteels, R. (1984) Biochem. J. 224, 445-451 48 Carafoli, E. and Zurini, M. (1982) Biochim. Biophys. Acta 683, 279-301 49 Gmaj, P., Murer, H. and Carafoli, E. (1982) FEBS Lett. 144, 226-230 50 Hasselbach, W. (1978) Biochim. Biophys. Acta 515, 23-53 51 Sch~nfeld, W., Menke, K.-H., SchiSnfeld, R. and Repke, K.R.H. (1984) Biochim. Biophys. Acta 770, 183-194 52 Langer, G.A. (1982) Annu. Rev. Physiol. 44, 435-449 53 Berridge, M. and Irvine, R. (1984) Nature 312, 315-321 54 Williamson, J., Cooper, R., Joseph, S. and Thomas, A. (1985) Am. J. Physiol. 248, C203-C216 55 Streb, H., Bayerd~rffer, E., Haase, W., Irvine, R.F. and Schulz, I. (1984) J. Membrane Biol. 81,241-253 56 Th6venod, F., Streb, H., Ullrich, K.J. and Schulz, I. (1986) Kidney Int. 29, 695-702 57 Prentki, M., Corkey, B.F. and Matschinsky, F.M. (1985) J. Biol. Chem. 260, 9185-9190 58 Prentki, M., Wollheim, C.B. and Lew, P.D. (1984) J. Biol. Chem. 259, 13777-13782 59 Michell, B. (1986) Nature 319, 176-177 60 SekarcM.C. and Hokin, L.E. (1986) J. Membrane Biol. 89, 193-210