Expression of the Receptor Tyrosine Kinase Ret on the Plasma Membrane Is Dependent on Calcium*

Mutations in the Ret receptor tyrosine kinase are responsible for a variety of human syndromes, including multiple endocrine neoplasia 2 and Hirschsprung’s disease. Ret is expressed as a 150-kDa precursor form in the endoplasmic reticulum and a 170-kDa mature form at the plasma membrane. Here we show that expression of p170 ret is dependent on calcium. Depletion of extracellular calcium completely blocks p170 ret expression, which is not caused by a decrease in half-life of p170 ret at the plasma membrane but by a defect in processing of p150 ret into p170 ret . This processing defect can be mimicked by treating the cells with thapsigargin, a drug that releases calcium from internal stores, indicating that reduction in luminal calcium is responsible for the processing defect. We propose that a relatively high concentration of luminal calcium is necessary for the proper folding of Ret in the endoplasmic reticulum.

Mutations in the Ret receptor tyrosine kinase are responsible for a variety of human syndromes, including multiple endocrine neoplasia 2 and Hirschsprung's disease. Ret is expressed as a 150-kDa precursor form in the endoplasmic reticulum and a 170-kDa mature form at the plasma membrane. Here we show that expression of p170 ret is dependent on calcium. Depletion of extracellular calcium completely blocks p170 ret expression, which is not caused by a decrease in half-life of p170 ret at the plasma membrane but by a defect in processing of p150 ret into p170 ret . This processing defect can be mimicked by treating the cells with thapsigargin, a drug that releases calcium from internal stores, indicating that reduction in luminal calcium is responsible for the processing defect. We propose that a relatively high concentration of luminal calcium is necessary for the proper folding of Ret in the endoplasmic reticulum.
Ret is a receptor tyrosine kinase expressed mainly in neural crest-derived cells. In most cell types Ret proteins are expressed as glycoproteins of 150 and 170 kDa. The p150 ret and p170 ret proteins differ in their subcellular localization. Cell fractionation experiments showed that both isoforms are present in membrane fractions (1). However, the 170-kDa isoform of Ret is present at the plasma membrane, because biotinylation of intact Ret-expressing cells results in biotinylation of only the 170-kDa isoform of Ret. The p150 ret isoform is endoglycosidase H-sensitive, suggesting that p150 ret is an incompletely processed form of Ret, present in the endoplasmic reticulum (2)(3)(4). Both forms of Ret are derived from a single polypeptide chain of approximately 120 kDa, as was shown by treatment of Ret-expressing cells with tunicamycin, an inhibitor of posttranslational N-linked glycosylation (4).
Mutations in the ret gene have been linked to several human syndromes, including multiple endocrine neoplasia 2A and 2B, familial medullary thyroid carcinoma, and Hirschsprung's disease (13). Hirschsprung's disease is a congenital disorder of the autonomic innervation of the gut (14). Mutational analysis has shown that specific point mutations in the intracellular domain of Ret result in impaired tyrosine kinase activity. Mutations in the extracellular domain of Ret result in dramatically reduced levels of p170 ret expression at the plasma membrane (15,16). Reduced levels of expression of functional Ret during embryonic development underlie the defects observed in Hirschsprung's disease patients, because in mice that completely lack Ret expression similar yet more severe symptoms are observed (17,18).
In the extracellular domain of Ret, a cadherin-like domain of approximately 110 amino acids is present (19). In cadherins, this domain is repeated three or four times and mediates calcium binding and calcium-dependent homophilic interactions with cadherin molecules on other cells (20). Also in the cadherin-like domain in Ret, consensus calcium-binding sequences are present. Asai et al. (3) demonstrated that a point mutation in one of these sequences results in a dramatic reduction of p170 ret expression at the plasma membrane, whereas p150 ret expression is not affected. They hypothesized that mutation of the putative calcium-binding site might interfere with processing of Ret in the Golgi complex or with Ret transport to the plasma membrane. Alternatively, it is possible that this mutant of Ret is fully processed and transported to the plasma membrane but cannot be retained at the membrane in the absence of calcium binding.
We have investigated the role of calcium in cell surface expression of Ret. We found that calcium depletion completely abolishes cell surface expression of p170 ret . Furthermore, we show by pulse-chase experiments that processing of p150 ret into p170 ret is strongly impaired under conditions of calcium depletion. Finally, we provide evidence that this impaired processing is due to improper processing of Ret in the endoplasmic reticulum (ER).

MATERIALS AND METHODS
Cell Growth Conditions-The SKP2 cell line expressing Ret proteins has been described previously (11). Cells were grown in DF12 medium supplemented with 10% fetal calf serum and antibiotics. For culture under calcium-free conditions, calcium-free Dulbecco's modified Eagle's medium supplemented with 10% Chelex-treated (calcium-free) fetal calf serum was used. Cells were washed two times with calcium-free phosphate-buffered saline (PBS) prior to addition of calcium-free medium for culture under calcium-free conditions. For restoring physio-* This work was supported by grants from the Netherlands Foundation for Chemical Research with financial support from the Netherlands Organization for Scientific Research (to D. H. J. v. W.) and from the Royal Academy of Arts and Sciences (to I. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed. E-mail: J.L.Bos@ med.ruu.nl. 1 The abbreviations used are: GDNF, glial cell line-derived neurotrophic factor; ER, endoplasmic reticulum; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; ERKZ, extracellular signal-regulated kinase z; HERRet, human epidermal growth factor receptor-ret chimaeric receptor. logical levels of extracellular calcium (1.8 mM), a stock solution of 1 M CaCl 2 was used. Thapsigargin was used at a concentration of 300 nM, and brefeldin A was used at a concentration of 5g/ml. For the analysis of Ret expression, cells were lysed, and Ret proteins were immunoprecipitated and analyzed as described previously (24). ERK2 activation was measured by the mobility shift assay as described (24). This assay measures the appearance of a slower migrating form of active phospho-ERK by Western analysis.
[ 35 S]Methionine Labeling-Subconfluent SKP2 cultures were incubated for 45 min or 3 h in methionine-free medium containing 200 Ci of [ 35 S]methionine per dish. Cells were washed with PBS and incubated for different periods of time with unlabeled methionine-containing medium with or without calcium. Cells were lysed in RIPA buffer, and Ret proteins were immunoprecipitated as described previously (24). Immunoprecipitated proteins were separated by SDS-PAGE, and radioactive proteins were visualized by autoradiography.
Biotinylation of Ret Proteins-Subconfluent cultures of SKP2 cells were rinsed with PBS, followed by incubation with 0.5 mg/ml sulfo-NHS-biotin (Pierce) in calcium-containing PBS for 30 min at room temperature. Subsequently, cells were rinsed with PBS and incubated in calcium-containing or calcium-free culture medium at 37°C. After different time periods, cells were lysed and Ret proteins were immunoprecipitated and analyzed by SDS-PAGE and Western blotting as described previously (24). For detection of biotinylated proteins streptavidin-conjugated horseradish peroxidase was used (Pierce).

RESULTS
p150 ret Is a Precursor of p170 ret -To study Ret processing and the role of calcium in Ret expression, we used SKP2 cells. The SKP2 cell line is derived from the human neuroepithelioma cell line SK-N-MC by stable transfection with a fulllength Ret expression plasmid (11). In these cells, the 150-and 170-kDa isoforms of Ret are roughly equally expressed ( Fig. 1A and Ref. 11). A precursor-product relationship between p150 ret and p170 ret was proposed previously (2-4) but has never been demonstrated. To substantiate the proposed precursor-product relation and to determine the kinetics of this process, we pulselabeled SKP2 cells for 45 min with [ 35 S]methionine, followed by a chase for different time periods. Cells were lysed, and Ret proteins were immunoprecipitated and analyzed by SDS-PAGE and autoradiography. As shown in Fig. 1B, the 45-min pulse resulted in labeling of only the 150-kDa isoform of Ret. With time, p150 ret was converted into p170 ret as judged from the decrease in labeled p150 ret and the increase of labeled p170 ret . From this result we conclude that indeed p150 ret is a precursor of p170 ret . The estimated half-life of p150 ret is about 1 h, which is similar to the estimated half-life of p170 ret , explaining the equal expression of the two products in steady state (Fig. 1A).
p170 ret Expression Is Dependent on Extracellular Calcium-Mutation of a critical amino acid in the calcium-binding domain of Ret drastically reduces p170 ret expression (3), suggesting a role for calcium in plasma membrane expression of p170 ret . To investigate the role of calcium on Ret expression, we cultured SKP2 cells in the presence and absence of extracellular calcium. In the presence of extracellular calcium, the 150and 170-kDa isoforms of Ret were expressed ( Fig. 2A). When extracellular calcium was removed by replacement of the culture medium with calcium-free medium, p170 ret expression gradually decreased. After 4 hours of incubation in the absence of extracellular calcium, hardly any p170 ret expression could be detected. Under these conditions an increase in p150 ret expression was observed ( Fig. 2A). When, after 16 h of growth in calcium-free medium, physiological levels of calcium were restored in the medium, p170 ret was re-expressed within 80 min (Fig. 2B). These results show that extracellular calcium is necessary for expression of p170 ret proteins in SKP2 cells.
Processing of p150 ret Is Dependent on Calcium-A simple explanation for the lack of p170 ret expression in the absence of extracellular calcium is that binding of calcium to p170 ret is necessary for the stability of p170 ret at the cell surface. To address this possibility we have tested the effect of extracellular calcium on the half-life of p170 ret at the plasma membrane. To that end, plasma membrane proteins of SKP2 cells were labeled with biotin, using the membrane-impermeable reagent sulfo-NHS-biotin. After biotinylation, cells were incubated for different periods of time in the presence or absence of extracellular calcium. Cells were lysed and (biotinylated) Ret proteins were immunoprecipitated with an anti-Ret antiserum and visualized by Western blotting using enzyme-conjugated streptavidin. As shown in Fig. 3A (top panel) the half-life of p170 ret at the plasma membrane in the presence of calcium is approximately 1 h, in agreement with the measured half-life of p170 ret in the pulse-chase experiment presented in Fig. 1B. However, the same half-life for Ret was found when the cells were cultured in calcium-free medium. As a control we reprobed the blot with anti-Ret antiserum. As expected, new (nonbiotinylated) p170 ret is formed in the presence of calcium, whereas in the absence of calcium the decrease in biotinylated p170 ret is not replaced by newly synthesized Ret (Fig. 3A, bottom panel). Thus, the decreased cell surface expression of p170 ret in the absence of extracellular calcium cannot be explained by a decrease in the half-life of cell surface-expressed p170 ret .
An alternative explanation for the absence of p170 ret at the cell surface may be that processing of p150 ret is very sensitive to the levels of intracellular calcium, a level that will be affected when the cells are cultured in calcium-free medium. To test this hypothesis we have performed pulse-chase experiments similarly as presented in Fig. 1B. In contrast to the result obtained in the presence of extracellular calcium, p150 ret was not processed into p170 ret when the chase was performed in the absence of extracellular calcium (Fig. 3B). Moreover, under these conditions there was no decrease in the amount of labeled p150 ret during the first 4 h and only a small decrease after 6 h. Clearly, in the absence of extracellular calcium processing of p150 ret is strongly reduced, resulting in an increased half-life of p150 ret and absence of p170 ret at the plasma membrane.
Reduced Levels of Calcium in the ER Inhibit p150 ret Processing-One possible explanation for the lack of p150 ret processing in the absence of extracellular calcium is that calcium is necessary for the proper folding of the extracellular domain of Ret. This folding process takes place in the ER, and proteins not properly folded are retained in the ER and subsequently, after a delay, degraded ("quality control", for review see Ref. 21). To test this possibility we have incubated SKP2 cells with thapsigargin. Thapsigargin is a toxin that specifically inhibits calcium ATPases in the ER membrane (22). As a result of this inhibition, calcium levels in the ER are reduced. Immunoprecipitation and Western blotting of Ret proteins from thapsigargin-treated SKP2 cells showed that in the presence of 300 nM thapsigargin p170 ret expression decreased with similar kinetics as upon removal of extracellular calcium (Fig. 4A). This result suggests that the calcium concentration in the ER determines whether p150 ret is processed.
To show that the thapsigargin-induced block in processing is  (25), this precursor form of HERRet was revealed (Fig. 4B). This faster migrating protein could not be labeled with sulfo-NHSbiotin, whereas the 155-kDa form of HERRet was efficiently biotinylated, establishing their localization in the cell and at the plasma membrane, respectively (data not shown). When brefeldin A was removed, the precursor protein was processed into the mature form of HERRet, both in the presence and absence of thapsigargin (Fig. 4B). Treatment of SKP2 cells with brefeldin A also inhibited p150 ret processing (while p170 ret protein down-regulation continued). Release of the cells from the brefeldin A block in the absence of thapsigargin restored p170 ret expression, but, as expected, release of the cells in the presence of thapsigargin did not lead to p150 ret processing (Fig.  4B). These results show that the block in p150 ret processing under reduced calcium concentrations in the ER is specific for the extracellular domain of Ret.
To further support the notion that the defect in processing is due to retention of p150 ret in the ER, we have allowed p150 ret to reach the Golgi compartment prior to calcium depletion. To that end we first pulse-labeled SKP2 cells with [ 35 S]methionine in the presence of extracellular calcium at 37°C for 45 min. This resulted in labeling of only p150 ret (Fig. 1B). Subsequently, the cells were chased for 2 h at 20°C. At this temperature no processing of proteins in the Golgi complex occurs, and thus folded p150 ret protein will accumulate in the Golgi complex (Fig. 5, lane 1). Next, we treated the cells with brefeldin A (5 g/ml) for 10 min at 37°C followed by 30 min at 37°C in the presence or absence of extracellular calcium. As a consequence proteins trapped in the Golgi complex that have already passed the quality control of the ER will be returned to the ER. After this treatment a small decrease in the size of the p150 ret proteins was observed, most likely as a result of trimming of glycan moieties, which normally occurs in the Golgi complex but which apparently did not occur at 20°C (Fig. 5, lane 2). Processing of this returned p150 ret was analyzed in the presence or absence of extracellular calcium. In the presence of extracellular calcium, p150 ret proteins that were returned to the ER were normally processed into p170 ret , showing that after removal of brefeldin A, protein processing returns to normal (Fig. 5, lanes 3-5). However, in the absence of extracellular calcium, p150 ret was also processed into p170 ret (Fig. 5,  lanes 6 -8). This result suggests that once p150 ret proteins have passed the ER quality control, reduction of calcium can no longer prevent processing, providing evidence that calcium affects an early step in p150 ret processing.
Inhibition of p150 ret Processing Abrogates Ret but Not HER-Ret Signal Transduction-GDNF stimulation of Ret-expressing cells results in p170 ret tyrosine phosphorylation and activation of different signal transduction pathways (11). From our results it can be predicted that SKP2 cells cultured in the presence of thapsigargin no longer respond to GDNF stimulation because of the absence of p170 ret expression. Indeed, although under normal culture conditions GDNF induced p170 ret tyrosine phosphorylation and ERK2 activation, GDNF stimulation of thapsigargin-treated SKP2 cells no longer resulted in p170 ret tyrosine phosphorylation or in ERK2 activation (Fig. 6). Similar results were obtained when SKP2 cells were grown in the absence of extracellular calcium prior to GDNF stimulation (data not shown).
To confirm that the extracellular (calcium-binding) domain of Ret is responsible for the effect of thapsigargin on Ret expression and signaling, we again used the SKF5 cell line. Stimulation of SKF5 cells with epidermal growth factor resulted in HERRet autophosphorylation and ERK2 activation (Fig. 6), which is in agreement with previous results (24). In contrast to p170 ret expression in SPK2 cells, culture of SKF5 cells in the presence of thapsigargin did not influence HERRet expression (Fig. 6). Moreover, epidermal growth factor-induced HERRet autophosphorylation and ERK2 activation were also not influenced by thapsigargin treatment (Fig. 6). Similar results were obtained when SKF5 cells were grown in the absence of extracellular calcium prior to epidermal growth factor stimulation (data not shown). These results show that the extracellular domain of Ret is responsible for the lack of p170 ret expression and signaling under reduced calcium conditions. DISCUSSION We have analyzed Ret expression in the stably transfected cell line SKP2, in which Ret is expressed as 150-and 170-kDa isoforms (11). Whereas p170 ret is a functional receptor located at the cell surface, p150 ret is insensitive to GDNF-stimulation and is located in the ER (3,11). Using pulse-chase experiments we now established previous proposals that p150 ret is a precursor of p170 ret (2)(3)(4). The half-life of both isoforms is approximately 1 h, which means that there is a relatively slow conversion of p150 ret into p170 ret but a relatively rapid downregulation of p170 ret from the plasma membrane as compared with for instance the platelet-derived growth factor or insulin receptor (26,27).
Based on the observation that a mutation in the putative calcium-binding domain of Ret leads to aberrant expression of p170 ret (3), we have investigated the role of calcium in p170 ret expression. We observed that depletion of extracellular calcium results in down-regulation of p170 ret but not p150 ret expression. This provided us with an excellent system for studying in further detail the reason for the aberrant cell surface expression of p170 ret . Using pulse-chase experiments we found that calcium is necessary for proper processing of p150 ret . Furthermore, by depleting intracellular calcium stores with thapsigargin, we could show that ER luminal calcium is responsible for this proper processing, compatible with the reported localization of p150 ret proteins in the ER (3). Moreover, we provide evidence that once p150 ret proteins have left the ER, calcium is no longer necessary for proper processing. The finding that calcium affects processing of Ret in the ER is supported by the notion that the extracellular domain of Ret is responsible for the calcium sensitivity.
In the ER, newly synthesized proteins are folded, and it has been shown for several proteins that high ER luminal calcium is necessary for proper folding (28,29). Calcium could be either necessary for the many calcium-binding proteins (chaperones) involved in protein folding (30 -32), or alternatively, calcium is directly involved in the three-dimensional structure of the newly synthesized protein, as shown for cadherins (33,34). A protein is exported from the ER only if a protein is properly folded ("quality control", for review see Ref. 21). The observed inhibition of p150 ret processing when ER luminal calcium is diminished is therefore most likely due to the failure to fold properly. This was further supported by the following observation: retention of p150 ret in the ER was diminished when Ret proteins were allowed to be folded in the presence of calcium and to be exported from the ER to the Golgi. When folded p150 ret proteins were subsequently returned from the Golgi to the ER using brefeldin A, they were no longer dependent on calcium for processing into p170 ret . Clearly, calcium affects p150 ret processing in the ER and does not influence processing after export from the ER.
The extracellular domain of Ret contains a cadherin-like domain of approximately 110 amino acids (1,19). In cadherins, this domain is repeated three or four times and mediates calcium binding and calcium-dependent homophilic interactions with cadherin molecules on other cells (20). For correct folding of cadherin molecules, calcium binding is also essential (33,34). It is therefore likely that calcium directly affects the folding of this domain. Indeed, a mutation of a putative calcium-binding site in Ret also inhibits p150 ret processing (3). However, despite the use of different methods, we were not able to convincingly show binding of calcium to full-length Ret proteins or to gluthatione S-transferase-Ret fusion proteins containing the putative calcium-binding domain of Ret (data not shown). Most likely, the affinity of this domain for calcium is very low, explaining not only the absence of detectable calcium binding but also the drastic effect of mild calcium depletion on p170 ret expression. An alternative explanation for the inhibition of p150 ret processing during extracellular calcium depletion is that calcium-binding chaperones are inhibited. This latter possibility is less obvious, because in general these chaperones have a very high affinity for calcium (22,(35)(36)(37) not likely to be affected by the mild methods of calcium depletion used in this study.
Reduced levels of functional Ret during embryonic development most likely underlie the defects observed in Ret-linked Hirschsprung's disease, because in mice completely lacking Ret expression, similar yet more severe symptoms are observed (17,18). Indeed, point mutations in the intracellular domain of Ret, as observed in Hirschsprung's disease, result in impaired Ret tyrosine kinase activity (16,38,39). Interestingly, Ret-linked Hirschsprung's disease is also caused by a variety of point mutations in the extracellular domain of Ret. These mutations result in dramatically reduced levels of p170 ret expression at the plasma membrane without affecting the presence of the 150-kDa form (15,16). Furthermore, the severity of the disease, as indicated by the length of the aganglionic segment, correlates with the degree of impairment of p170 ret expression (15). These results indicate that the extracellular domain of Ret is very sensitive for mutational alterations that either affect Ret processing or retention of Ret at the plasma membrane.
Our results indicate that protein processing, most likely protein folding, is a delicate step in p150 ret processing that needs high levels of calcium in the ER to occur. Even under these conditions processing of p150 ret is slow, taking at least 1 h to proceed. The similarity between the aberrant expression of p170 ret in Hirschsprung's disease associated with mutations in the extracellular domain of Ret and the aberrant expression of p170 ret in the absence of calcium may indicate that the extracellular domain mutations also cause a defect in proper folding of p150 ret .
Further research is necessary to show whether modulation of luminal calcium concentration is involved in the regulation of Ret expression in vivo. However, inositol 3-phosphate-mediated release of calcium from internal stores is a common event after receptor stimulation, and perhaps this may affect processing of Ret. In addition, gradients of extracellular calcium do exist, for instance in the skin where calcium influences keratinocyte differentiation (40, 41).