Activity of the hSPCA1 Golgi Ca²⁺ pump is essential for Ca²⁺-mediated Ca²⁺ response and cell viability in Darier disease

Changes in extracellular Ca²⁺ concentrations, as are seen in normal unperturbed epidermis or after epidermal permeability perturbation, control epidermal functions such as differentiation, barrier repair, keratinocyte cell-to-cell adhesion and keratinocyte motility (Fang et al., 1998; Mao-Qiang et al., 1997; Vasioukhin et al., 2000). Thus, transducing changes in extracellular Ca²⁺ levels into intracellular Ca²⁺ signals that control these processes is essential for keratinocyte and epidermal viability.

Increasing extracellular Ca²⁺ concentration in cultured keratinocytes mimics many of the changes in keratinocyte differentiation and cell-to-cell adhesion seen in vivo (Pillai et al., 1990; Stanley and Yuspa, 1983; Vasioukhin et al., 2000), through a well-defined pathway that uses a plasma-membrane-bound Ca²⁺ receptor (CaR) to signal intracellular Ca²⁺ release, and subsequent long-lasting Ca²⁺ influx (Fatherazi et al., 2003; Oda et al., 1998). The coordinated responses to raised extracellular Ca²⁺ levels establish a new resting cytosolic concentration that continues to increase slowly (days to weeks) as the cells are exposed to continuing high levels of extracellular Ca²⁺ (Pillai et al., 1990). Initial studies found that emptying the endoplasmic reticulum (ER) Ca²⁺ store by pharmacologically inhibiting the Ca²⁺-ATPase, SERCA2, responsible for sequestering Ca²⁺, impaired the cells' response to raised extracellular Ca²⁺ (Li et al., 1995). These studies suggested that keratinocytes, like most other non-excitable cells, used only the ER Ca²⁺ store.

However, more recent studies have demonstrated that the Golgi Ca²⁺ store, sequestered by a related Ca²⁺-ATPase, hSPCA1, also is required for Ca²⁺ signaling in response to raised extracellular Ca²⁺ concentrations. Mutations in the Golgi hSPCA1 cause another inherited skin disorder, Hailey-Hailey disease (HHD). Keratinocytes from HHD patients do not respond to raised extracellular Ca²⁺ levels and suffer from depleted Golgi Ca²⁺ stores, even though ER Ca²⁺ stores are normal (Hu et al., 2000). Further, experimental manipulation of the hSPCA1 demonstrates that keratinocytes, but not COS cells, require the hSPCA1-mediated Golgi Ca²⁺ store for normal signaling (Callewaert et al., 2003). Keratinocytes may therefore be unique among mammalian cells in requiring intact Golgi Ca²⁺ stores for Ca²⁺ signaling.

Darier disease (DD) is an acantholytic skin disease in which the ER Ca²⁺-ATPase, SERCA2 encoded by ATP2A2, is mutated (Sakuntabhai et al., 1999). Like HHD, this disease is characterized by impaired differentiation and abnormal cell-to-cell adhesion. Keratinocytes with these naturally occurring mutations allow us to determine how keratinocytes adapt to impaired ER Ca²⁺ stores, and also to examine the relative contributions of ER vs Golgi Ca²⁺ stores in the Ca²⁺-signaling cascade in response to raised extracellular Ca²⁺ levels.

Three isoforms of SERCA2 are generated by alternative splicing of exon 20 and exon 21. Isoform `a' is the cardiac sarco-endoplasmic Ca<sup>2+</sup> pump whereas isoform `b' is the major SERCA isoform in non-muscle cells, including epidermal cells (Wuytack et al., 1989; Wuytack et al., 2002). SERCA2c, a recently described splice variant, is expressed in epithelial, mesenchymal and hematopoietic cell lines, and in monocytes (Gelebart et al., 2003). SERCA2b contains an additional transmembrane domain, which places the C-terminus of the protein in the lumen of the ER, whereas this terminus is cytosolic in SERCA2a. Both the SERCA2a and SERCA2b isoforms are highly expressed in cultured keratinocytes but only SERCA2b is found in the epidermis. Loss of the SERCA2b isoform is sufficient to cause DD (Dhitavat et al., 2003). SERCA2 sequesters Ca²⁺ in the ER whereas hSPCA1 localizes to the trans-Golgi membrane and sequesters Ca²⁺ in the Golgi (Behne et al., 2003). hSPCA1 transports Ca²⁺ with an affinity comparable to that of the ER SERCA2b pump (Lytton et al., 1992; Ton et al., 2002). hSPCA1 also transports Mn²⁺ ions with high affinity (Fairclough et al., 2003; Ton et al., 2002). hSPCA1 is responsible for 67% of the Ca²⁺ uptake into the Golgi (Callewaert et al., 2003), which is consistent with the observation that ATP2C1 transcripts are highly expressed in keratinocytes (Hu et al., 2000).

In this report, we show that DD keratinocytes overexpress hSPCA1 in compensation for decreased SERCA2, and thereby retain a near-normal response to raised extracellular Ca²⁺ concentrations. These studies demonstrate that the Golgi Ca²⁺-ATPase hSPCA1 mediates a compensatory mechanism for the functional deficiency in the ER Ca²⁺-ATPase seen in DD keratinocytes and confirm that keratinocytes use the hSPCA1-controlled Golgi Ca²⁺ store to mediate Ca²⁺ signaling.

We first confirmed that SERCA2 protein was decreased in the DD keratinocytes (Fig. 1). Preliminary studies demonstrated that ATP2A2 mRNA synthesis was tightly grouped among keratinocytes taken from four unrelated adult donors (data not shown); thus keratinocytes from one normal adult were used as controls in the following experiments. The S920Y (DD2) and N767S (DD5) mutations led to a minor decrease (about 25%) in the amount of total endogenous SERCA2 protein (WT + mutated protein), whereas mutations Y203del (DD1), A998fsX33 (DD3) and S346F (DD4) caused a more profound decrease (up to 50%). These results are consistent with previous functional analysis of the S920Y and A998fsX33 mutations using overexpression of a mutated cDNA (Ahn et al., 2003; Dhitavat et al., 2003; Dode et al., 2003). We next assessed the SERCA2-controlled ER Ca²⁺ store by measuring the Ca²⁺ released by thapsigargin, which specifically and irreversibly inhibits SERCA pumps. ER Ca²⁺ stores were much lower in DD versus normal keratinocytes (Table 2). Both Ca²⁺ stored in the ER by SERCA, measured after exposure to 500 nM thapsigargin, and also total stored Ca²⁺, measured after exposure to 1 μM ionomycin, were decreased in DD keratinocytes (Table 2). Increases in cytosolic Ca²⁺ following ionomycin or thapsigargin treatment did not differ among keratinocytes taken from different DD patients (data not shown).

Increased extracellular Ca²⁺ concentrations activate the plasma membrane Ca²⁺ receptor (CaR), resulting both in Ca²⁺ release from intracellular stores and subsequent Ca²⁺ influx through plasma membrane ion channels. Keratinocytes again seem to differ from other mammalian cells in that both the hSPCA1-controlled Golgi Ca²⁺ stores and the SERCA2-controlled ER Ca²⁺ stores are central to intracellular Ca²⁺ release and subsequent Ca²⁺ influx (Behne et al., 2003; Hu et al., 2000). We therefore tested whether the Ca²⁺ response was preserved in DD keratinocytes. Although overall Ca²⁺ homeostasis is not identical in DD and normal keratinocytes, as evidenced by abnormal resting Ca²⁺ concentrations, DD keratinocytes nevertheless responded robustly to raised extracellular Ca²⁺ concentrations (Fig. 2), with an identical absolute increase in cytosolic Ca²⁺ levels and a significantly higher increase in response to raised extracellular Ca²⁺. The rate of increase in intracellular Ca²⁺ did not differ between normal and DD keratinocytes (Table 3).

Resting cytosolic Ca²⁺ concentrations varied among different patients, although Ca²⁺ values from each patient were tightly grouped, suggesting that different types of mutations might produce higher or lower cytosolic Ca²⁺ concentrations. However, simply classifying the mutations into nonsense or missense mutations did not predict resting Ca²⁺ values in our small sample (Table 3).

James Putney (Putney, 1986) first advanced the `capacitive Ca²⁺ entry' model, proposing that Ca²⁺ entered the cell in response to the emptying of intracellular Ca²⁺ stores. Thus, intracellular Ca²⁺ release is amplified by Ca²⁺ influx from the extracellular compartment, passing through ion channels located in the plasma membrane. Earlier keratinocyte studies using Ca²⁺-sensitive dyes have demonstrated that significant Ca²⁺ entry follows the emptying of the ER Ca²⁺ stores (Csernoch et al., 2000; Fatherazi et al., 2003). Since DD ER Ca²⁺ stores are depleted relative to normal keratinocytes, we compared capacitive Ca²⁺ entry in DD versus normal keratinocytes. Standard protocols for measuring capacitive Ca²⁺ entry first remove extracellular Ca²⁺ (to block non-capacitive Ca²⁺ entry), then release SERCA-dependent Ca²⁺ stores by adding thapsigargin, and finally restore extracellular Ca²⁺ and quantify capacitive Ca²⁺ influx by measuring increases in cytosolic Ca²⁺ concentrations. Because millimolar extracellular Ca²⁺ concentrations not only support capacitive Ca²⁺ influx but also might activate Ca²⁺ signaling via the CaR-mediated pathway, we restored extracellular Ca²⁺ to a final concentration of 0.06 mM (the same concentration in which the cells were grown), which does not significantly activate the keratinocyte plasma membrane CaR (Oda et al., 1998). When extracellular Ca²⁺ was restored after emptying Ca²⁺ stores, we found that capacitive Ca²⁺ entry was larger in DD relative to normal keratinocytes, consistent with diminished ER stores and normal signaling between intracellular Ca²⁺ stores and plasma membrane channels (Table 4, Fig. 3). Unlike pancreatic cells containing mutated ATP2A2 (Zhao et al., 2001), rates of Ca²⁺ increase or decrease were not different between normal and DD keratinocytes, suggesting that extrusion mechanisms such as that mediated by the plasma membrane Ca²⁺-ATPase (PMCA) or by Na⁺/Ca²⁺ exchangers were not different between DD and normal keratinocytes.

Taken together, these findings demonstrate that the Ca²⁺ signaling in response to raised extracellular Ca²⁺ is intact in DD keratinocytes, despite the loss of SERCA2-controlled Ca²⁺ stores. These findings suggest that other Ca²⁺-signaling mechanisms are likely to be upregulated in response to decreased SERCA2 levels and function. We therefore examined other mechanisms that might compensate for SERCA2 loss.

After extracellular Ca²⁺ stimulates an acute increase in intracellular Ca²⁺, cytosolic Ca²⁺ is then stabilized at a new, increased Ca²⁺ concentration by the coordinated interaction between Ca²⁺ extrusion through the Na⁺/Ca²⁺ exchangers and PMCA and re-uptake of Ca²⁺ into the ER (via the SERCA2), and into the Golgi (via the hSPCA1). Ca²⁺ signaling in DD keratinocytes therefore might be due to compensatory upregulation of Ca²⁺-permeable ion channels, plasma membrane transporters such as PMCA4, the major PMCA expressed in keratinocytes (Cho and Bikle, 1997), or intracellular Ca²⁺ ATPases such as the hSPCA1, localized to the Golgi in keratinocytes. We first examined whether DD Ca²⁺ signaling was due to upregulation of plasma membrane ion channels or transporters by assessing Ca²⁺ influx or efflux across the plasma membrane, measured using ⁴⁵Ca²⁺ influx or efflux (Table 5). Neither Ca²⁺ influx nor Ca²⁺ efflux was significantly different in DD vs. normal keratinocytes. These experiments suggest that plasma membrane components, such as the PMCA, do not compensate for decreased SERCA2 in DD keratinocytes.

hSPCA1 is central to keratinocyte Ca²⁺ signaling and regulation of resting Ca²⁺ concentrations (Behne et al., 2003; Hu et al., 2000). Because hSPCA1 constitutes the second major mechanism for sequestering intracellular Ca²⁺ in keratinocytes, we next investigated whether decreases in SERCA2 stimulated a compensatory upregulation of hSPCA1. We found that hSPCA1 was markedly upregulated in all DD keratinocytes studied (Fig. 4). These data, combined with the functional signaling characteristics of DD keratinocytes, suggested that hSPCA1 upregulation might compensate for SERCA2 dysfunction.

Plasma membrane Ca²⁺-ATPases (PMCA) also are known to regulate Ca²⁺ extrusion. However, of the five DD patients studied, only two, DD2 (S920Y) and DD3 (A998fsX33), overexpressed total PMCA or PMCA-4. The others displayed a normal (DD4, S346F; DD5, N767S) or slightly decreased (DD1, Y203del) level of PMCA, indicating that PMCA expression was variable among DD patients. These data, coupled with the functional Ca²⁺ efflux data above, suggest that changes in hSPCA1, but not PMCA or PMCA-4, underlie the compensatory changes in DD keratinocytes that allow normal Ca²⁺ signaling in response to raised extracellular Ca²⁺ concentrations.

We previously found that hSPCA1 dysfunction abrogates the Ca²⁺ response in keratinocytes containing ATP2C1 mutations (Behne et al., 2003; Hu et al., 2000). To test whether compensatory hSPCA1 upregulation enables DD keratinocytes to respond to extracellular Ca²⁺ levels, even in the face of SERCA2 dysfunction, we inactivated ATP2C1 in DD keratinocytes using small interfering RNA (siRNA). RNA interference (RNAi) is a strategy whereby genes are silenced post-translationally by treatment with sequence-specific double-stranded RNA. Cellular exposure to siRNA results in the selective degradation of the complementary native RNA (McManus and Sharp, 2002). Our siRNA oligonucleotides were designed to avoid any overlap between ATP2C1 and ATP2A2 (confirmed by BLAST analysis). To assess non-specific effects, we measured ATP2A2 cDNA expression after treatment with ATP2C1 siRNA. Only a 10% reduction in ATP2A2 cDNA was observed after treatment with ATP2C1 siRNA in normal keratinocytes, confirming that any changes in Ca²⁺ signaling were due specifically to decreases in hSPCA1. Normal and DD human keratinocytes were cultured in six-well plates to 75% confluence, then treated with siRNA (8 μg/well). Control samples were treated with transfection reagent and annealing buffer only. In preliminary experiments, we found that siRNA decreased ATP2C1 mRNA by approximately 80% within 24 hours after treatment.

Since a drop in protein levels generally lags behind a decrease in mRNA, we next assessed protein levels and compared these with changes in Ca²⁺ signaling in normal and DD keratinocytes treated with siRNA to ATP2C1. Normal keratinocytes treated with siRNA displayed normal Ca²⁺ signaling, confirmed by measuring normal hSPCA1 protein levels, 48 hours after siRNA treatment (data not shown). By 72 hours, however, hSPCA1 protein levels dropped, and Ca²⁺ signaling in normal keratinocytes correspondingly changed to resemble the abnormal Ca²⁺ signaling seen in Hailey-Hailey keratinocytes, in which the hSPCA1 is mutated (Hu et al., 2000). Specifically, baseline cytosolic Ca²⁺ concentrations increased, and the ability to respond to raised extracellular Ca²⁺ levels diminished, after exposure to siRNA (Fig. 5 and Table 6). DD keratinocytes responded more quickly to siRNA, demonstrating decreased hSPCA1 levels and abnormal Ca²⁺-signaling parameters within 48 hours of siRNA treatment. Resting Ca²⁺ concentrations, although increased, did not reach cytosolic Ca²⁺ levels seen in normal keratinocytes, demonstrating that Ca²⁺ signaling changes at 48 hours were due to selective inhibition of hSPCA1 synthesis and not non-specific toxicity. In contrast to normal keratinocytes, DD keratinocytes became pyknotic and did not cleave the Fura-2 dye to its active form at 72 hours, demonstrating that these cells were no longer viable. In both normal and DD keratinocytes, transfection-reagent-treated control cells did not differ from untreated keratinocytes (compare Table 6 to Table 3). These data confirm that hSPCA1 is essential for Ca²⁺-mediated Ca²⁺ responses. Further, these data suggest that increased hSPCA1 compensation for dysfunctional SERCA2 is essential for DD keratinocyte cell viability, because DD keratinocytes treated with siRNA to ATP2C1 died 48-72 hours after ATP2C1 was inactivated.

In this study, we investigated Ca²⁺ homeostasis in DD keratinocytes with five different causative mutations of ATP2A2. These cells all displayed decreased total and ER Ca²⁺ stores owing to decreased SERCA2 levels. In spite of decreased ER Ca²⁺ stores, we found that the DD cells preserved their Ca²⁺-signaling capability and were able to respond to raised extracellular Ca²⁺ levels. Because we know that ER Ca²⁺ is decreased in DD keratinocytes (Tables 2 and 4), it is likely that Ca²⁺ release from this compartment in response to raised extracellular Ca²⁺ is decreased in DD cells. As the hSPCA1 is upregulated in these DD keratinocytes, it appears that increased Ca²⁺ signaling via the Golgi Ca²⁺ store compensates for defective SERCA2-dependent signaling. The proposed changes in Ca²⁺ signaling pathways in DD vs normal keratinocytes are described in Fig. 6.

To date, 22 mutations of the ATP2A2 gene have been studied by site-directed mutagenesis of the ATP2A2 cDNA and overexpression in COS-1 or HEK293 cells (Ahn et al., 2003; Dode et al., 2003; Sato et al., 2004). Nonsense (Q790X, E917X) and frame-shift (1625delAG) mutations studied resulted in truncated pump proteins that were susceptible to degradation by the proteasome (Ahn et al., 2003). Most DD mutants (10 out of 12) displayed decreased expression and/or stability with the exception of two mutants (C344Y and V843F) (Ahn et al., 2003). Kinetic analyses of SERCA2 missense variants revealed that any one of the SERCA2 catalytic cycle steps (Ca²⁺ binding, autophosphorylation from ATP, conformational change, release of bound Ca²⁺ into the ER lumen and dephosphorylation) could be affected. Depending on the location of the amino acid substitution, the activity of the mutated pump may be decreased (most of the missense mutations), increased (S920Y), or abolished when the phosphorylation step is affected (T357K, G769R) (Dode et al., 2003; Sato et al., 2004). Furthermore, some missense mutants (N39D, N39T, C344Y, F487S, S920Y) inhibit the activity of native and recombinant wild-type SERCA2b by directly interacting with the wild-type pump (Ahn et al., 2003). Patient DD2 in this work bears the S920Y mutation studied by others (Ahn et al., 2003; Dode et al., 2003). Our results show that the resting cytosolic Ca²⁺ concentration in these cells was lower than that of the control but higher than in the four other DD cells studied. The kinetic analyses by Dode et al. revealed that S920Y is a unique DD mutant displaying an enhanced molecular Ca²⁺-transport activity relative to wild-type SERCA2b. The apparent affinity for Ca²⁺, however, was threefold lower relative to wild-type SERCA2b and this mutant was insensitive to the feedback inhibition of the transport cycle by accumulated lumenal Ca²⁺. Ahn et al. have shown a decreased stability of S920Y, preferential co-immunoprecipitation of S920Y with the wild-type pump, and a reduction by S920Y of the activity of wild-type SERCA2b (Ahn et al., 2003). This could explain the intermediate level of resting Ca²⁺ concentrations in DD2 cells.

All these results demonstrate that DD is caused in most cases by haploinsufficiency of SERCA2b, which could be amplified when the mutated pump has a dominant-negative effect. The analysis of L321F (Sato et al., 2004) revealed that a decreased Ca²⁺ affinity and insensitivity to the feedback inhibition without decreased expression are sufficient to cause abnormal Ca²⁺ homeostasis and DD. The range of functional alterations observed among ATP2A2 variants could account for the variable clinical features of DD.

Although Ca²⁺-mediated Ca²⁺ signaling was preserved in DD keratinocytes (Fig. 6), Ca²⁺ homeostasis was not completely normal, as evidenced by the abnormally low resting cytosolic Ca²⁺ concentrations (Table 2). Baseline cytosolic Ca²⁺ concentrations are maintained by the constant interaction of proteins that control Ca²⁺ release or influx with proteins that control Ca²⁺ extrusion or reuptake. Loss of one of the reuptake mechanisms, such as the Ca²⁺-sequestering SERCA2, resulting from DD mutations or from acute pharmacologic SERCA2 inactivation (Hu et al., 2000; Jones and Sharpe, 1994), may be expected to increase cytosolic Ca²⁺ levels, as seen in HHD keratinocytes (Behne et al., 2003; Hu et al., 2000). In fact, DD keratinocytes respond to long-term SERCA2 dysfunction by decreasing resting cytosolic Ca²⁺ levels (Table 2). These cytosolic Ca²⁺ concentrations, although lower than those of most mammalian cells, are comparable to those seen in healthy mouse keratinocytes (Li et al., 1995); and we find that these DD keratinocytes appeared to retain comparable proliferative capability even though (or perhaps because) their resting cytosolic Ca²⁺ concentrations are abnormally decreased. Decreased baseline cytosolic Ca²⁺ levels were unlikely to result from changes in plasma membrane channels or transporters, as Ca²⁺ influx and efflux were unchanged in DD compared with normal keratinocytes. However, increased hSPCA1 may contribute to changes in resting cytosolic Ca²⁺, as inactivating ATP2C1 with siRNA, in addition to abrogating the Ca²⁺ response, also increased baseline cytosolic Ca²⁺ in both normal and DD keratinocytes (compare Table 6 with Table 3), consistent with changes seen in HHD keratinocytes (Behne et al., 2003; Hu et al., 2000), in which the ATP2C1 is mutated. According to Liu and co-workers, in cells overexpressing the plasma membrane Ca²⁺-ATPase, PMCA1a, the activity of the Ca²⁺ release-activated channels (CRAC) pathway is upregulated, whereas the inositol 1,4,5-trisphosphate receptor (IP3R) and the SERCA pumps are downregulated (Liu et al., 1996). In Chinese hamster ovary (CHO) cells overexpressing PMCA4, a reduction in SERCA2b expression levels has been described (Guerini et al., 1995). It is conceivable that a compensatory mechanism relates Ca²⁺ efflux by the cell membrane ATPases to the activity of the ER Ca²⁺-ATPases, through the regulation of the respective genes. Our data, however, show that overexpression of PMCA proteins is at best small and non-uniformly observed across the range of DD mutations studied; apart from the ATP2A2 mutations, the DD keratinocyte donors differed regarding age, gender and genetic background. In addition, Ca²⁺ efflux was not increased in DD keratinocytes, suggesting that PMCA does not play a significant compensatory role.

Conversely, we have found evidence of several-fold overexpression of hSPCA1 in DD keratinocytes. Moreover, we found that compensatory overexpression of this pump in response to chronic SERCA2 deficiency appears to be essential for a normal Ca²⁺ response to raised extracellular Ca²⁺ levels and partially compensates for SERCA2 functional deficiency by enabling DD keratinocyte survival. Thus, this report adds to the growing body of information demonstrating that this previously obscure (to mammalian cell biologists) Ca²⁺-ATPase in fact plays an essential role in mammalian cell Ca²⁺ signaling. For example, Reinhardt and co-workers observed that overexpression of rat SPCA in COS-7 cells caused significant alterations in several of the cell Ca²⁺ transport molecules and dramatically increased the cell division rate (Reinhardt et al., 2004). Total PMCA protein expression was reduced along with the expression of the ER Ca²⁺-binding protein calreticulin. These changes suggest a redundant flexibility in the cell Ca²⁺ homeostatic mechanism. The pumping of excess cytoplasmic Ca²⁺ out of the cell or into the ER was necessarily reduced to compensate for increased Ca²⁺ movement into the Golgi. The Golgi of these cells compensated for this increased Ca²⁺ influx by increased expression of the Golgi Ca²⁺-binding protein CALNUC. This would seem to be a necessary adaptation to prevent Ca²⁺ cytotoxicity.

Taken together, the yeast and human data strongly suggest that SPCA is the mammalian Golgi Ca²⁺-ATPase. The opposing argument put forth by Taylor and co-workers (Taylor et al., 1997) is that the Golgi complex does not contain a unique resident Ca²⁺-ATPase. They state that all Ca²⁺ uptake into Golgi can be attributed to PMCAs in transit to the plasma membrane and to SERCAs that are not restricted to the ER. This is at odds with the data from the present study, where we show that specific inhibition of hSPCA1 abolishes Ca²⁺ response in DD keratinocytes. A major role for hSPCA1 in keratinocyte Ca²⁺ homeostasis is thus warranted.

The subcellular specialization of hSPCA1 may underlie the differences in clinical manifestations between DD and HHD. hSPCA1 deficiency causes acantholysis, whereas mutations in the gene encoding SERCA2 cause both acantholysis and apoptosis. In turn, the clinical phenotype of acantholysis plus apoptosis in DD probably reflects the varied functions of the keratinocyte hSPCA1-versus SERCA2-controlled Ca²⁺ stores.

All patients (Table 1) gave informed consent. Biopsies were taken from the unaffected abdominal skin. Control biopsies of normal skin were obtained from two age-matched controls. Patients were screened for mutations in the ATP2A2 gene. PCR amplification of the 21 exons and flanking splice sites of ATP2A2 was performed using previously published primers and protocols (Sakuntabhai et al., 1999). PCR products were sequenced on an ABI 377 automated sequencer using the BigDye chemistry (Applied Biosystems).

The S920Y mutation found in patient DD2 was identified previously and the functional analysis of this mutation has been described (Ahn et al., 2003; Dode et al., 2003). The A998fsX33 mutation found in DD3 occurs in the region specific to the SERCA2b isoform and was identified previously (Dhitavat et al., 2003). The N767S missense mutation found in DD5 has been reported previously by others (Jacobsen et al., 1999; Ruiz-Perez et al., 1999). The previously unpublished mutations, Y203del (DD1, deletion) and S346F (DD4, missense), affect highly conserved amino acid positions, respectively in the actuator domain of SERCA2 and the phosphorylation domain next to the phosphorylation site (Asp351).

Biopsies were explanted from adult normal skin (surgical skin margins) or clinically normal DD skin (punch biopsies) and primary keratinocytes were isolated and grown as previously described (Rheinwald and Green, 1975). Second to fifth passage cultured human keratinocytes were grown in 0.06 mM Ca²⁺ EpiLife medium (Cascade Biologics, Eugene, Oregon) to approximately 60-80% confluence (Boyce and Ham, 1983).

To study calcium influx (Ca_i), normal and DD keratinocytes were cultured in six-well plates to 70-80% confluence (Grando et al., 1996). Cells were washed for 10 minutes in Ca²⁺-free buffer (20 mM HEPES, 120 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1 mg/dl glucose, pH 7.4), then incubated for 1.5 hours at 37°C in EpiLife plus 0.06 mM Ca²⁺ plus 0.05 μCi ⁴⁵CaCl₂ (specific activity ∼0.8 mCi/mmol). Subsequently they were washed four times, quickly, with Ca²⁺-free buffer plus 2 mM ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA) then lysed with 0.5N NaOH. The lysate was analyzed with scintillation spectroscopy.

To study Ca²⁺ efflux, cultured cells were washed for 10 minutes with Ca²⁺-free buffer, then incubated for 1.5 hours at 37°C in EpiLife plus 0.06 mM Ca²⁺ plus 0.05 μCi ⁴⁵CaCl₂. Subsequently, they were washed four times, quickly, with Ca²⁺-free buffer plus 2 mM EGTA, then incubated for 1.5 hours at 37°C in EpiLife plus 0.06 mM Ca²⁺. Media then were collected and analyzed by scintillation spectroscopy.

Keratinocytes grown on glass coverslips were incubated in 6.26 μM cell-permeant Fura-2 acetomethoxy ester (Molecular Probes, Eugene, OR) at 37°C for 15-30 minutes, rinsed in the control solution (138 mM NaCl, 2.7 mM KCl, 0.01 M Na phosphate, 0.06 mM CaCl₂, 10 mM glucose, pH 7.4), and intracellular Ca²⁺ was monitored by ratiometric method, using the InCytIM2 Imaging System (Intracellular Imaging, Cincinnati, OH). For experiments measuring capacitive Ca²⁺ influx, cells were initially perfused with a solution containing 138 mM NaCl, 2.7 mM KCl, 0.01 M phosphate, 10 mM glucose, 0 mM CaCl₂ plus 0.5 M EGTA, pH 7.4., then perfused with the same solution with the Ca²⁺ concentration increased to 0.06 mM, without EGTA. A calibration curve was constructed using a standard calibration kit (Molecular Probes, Eugene, OR).

Table 2 relates to cells from all five DD mutants. Tables 3 and 4 contain data from DD1, DD2 and DD3. Because of the technical complexity and numbers of cells needed to perform the ⁴⁵Ca²⁺ and siRNA experiments, the ⁴⁵Ca²⁺ experiments were done with cells from DD1 and DD2, whereas the siRNA experiments were done with cells from a single mutant, DD3. DD3 was chosen for siRNA studies because its mutation occurs in the region specific to the SERCA2b isoform, which we previously have shown is sufficient to cause Darier disease (Dhitavat et al., 2003).

The rate of change (increase or decrease) was calculated for each cell as: (Peak Ca²⁺ value - baseline or recovery Ca²⁺ value)/(Peak Ca²⁺ time - baseline or recovery Ca²⁺ time), leading to a rate of increase or decrease measured as ΔnM Ca²⁺/second. Data are presented as the mean ± s.d. Statistical analysis was performed using an unpaired two-tailed Student's t-test.

For PMCA and SERCA2 detection, cell lysates in 2× Laemmli SDS sample buffer (2 μg total protein per well) were fractionated on 10% polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were blocked in PBS with 5% non-fat dry milk, 1% BSA and 0.5% Tween-20, and incubated at room temperature with the relevant primary antibody for 1 hour: mouse monoclonal anti-human SERCA2 antibody, clone IID8 (Biomol), 1:5000; mouse monoclonal anti-human PMCA, clone 5F10 (Abcam), 1:4000; and rabbit anti-human GAPDH polyclonal antibody (Abcam), 1:20,000. Chemiluminescent detection was performed using secondary reagents from the ECL Plus western blotting detection system (Amersham). The protein contents of the samples were quantified to ensure equal loading of protein into each well. For hSPCA1 detection, cells were lysed in a buffer containing 50 mM Tris-HCl pH 7.4, 150 mM KCl, 250 mM sucrose, with addition of 1 tablet of protease inhibitors per 10 ml of buffer (Complete Mini EDTA-free, Roche Diagnostics) and 30 μg total protein per well was fractionated by SDS-PAGE. Following transfer of proteins to PVDF membranes and blocking with 5% milk, 0.5% Tween-20 in PBS, samples were incubated overnight at 4°C with anti-human hSPCA1 primary antibody (Santa Cruz Biotechnology), 1:1000 in the blocking solution. Protein expression also was normalized using an antibody to β-actin (Sigma, clone AC-74), 1:5000 dilution. To quantify calcium pump expression in DD keratinocytes, densitometry was performed on the chemiluminescence photo images, using a Bio-Rad GS-710 scanner and Quantity One analysis software. Density values were normalized to loading control expression (GAPDH or β-actin) within same samples, and a percentage value for the difference between conditions was calculated.

The effectiveness of potential siRNA candidates was first predicted by choosing a G/C content of approximately 50%, and homology to ATP2C1 or ATP2A2 was checked by running the sequence of the designed siRNA against the full genome sequence in BLAST search (http://www.ncbi.nlm.nih.gov/BLAST/). After testing, one optimum sequence to silence ATP2C1 (codon position: 150, 47% G/C content, 5′-agg cuc gcc uau gac uaa cTT-3′) and one sequence to silence ATP2A2 (codon position: 2515, 57% G/C content; 5′-gcg ccg acg uaa cag cca aTT-3′) were selected. The oligoribonucleotides were synthesized by TriLink Biotechnologies (San Diego, CA). At 75% confluence, primary normal and DD keratinocytes were treated with siRNA suspended in 0.03 mM Ca²⁺, serum-free medium using the Trans-Messenger Transfection Kit (Qiagen). Control normal and DD keratinocytes were treated using only the transfection reagents. No signs of cytotoxicity were observed with this protocol in experimental or control keratinocytes. The medium was changed after 4 hours, and cells were harvested after 48 hours (DD keratinocytes) and 72 hours (normal keratinocytes). RNA was extracted using the RNAeasy Mini Kit (Qiagen) and quantified using a spectrophotometer at 260 nm. cDNA was synthesized from RNA of each sample using the TaqMan Reverse Transcription kit (Applied Biosystems) and measured by the ABI Prism 7900 HT instrument using SYBR Green as gene-amplification detection in quantitative PCR (Applied Biosystems). The 18S RNA was used as a control, housekeeping gene. Protein levels were determined on total lysates using western analysis (see above).

Statistical analysis was performed using unpaired two-tailed Student's t-tests.

L.F. and A.H. thank Nathalie Antoni, Jittima Dhitavat, Agnès Gadroy and Ariane Rochat for providing patient skin biopsies or cells, and José Enrique Mejía for his help in the preparation of this manuscript. This work was supported by the National Institutes of Health, grant 2PO1AR39488 (T.M.), the Medical Research Service of the San Francisco Veterans Administration Hospital (T.M.), and the Fondation pour la Recherche Médicale (A.H.). L.F. is the recipient of a fellowship from the French Ministry of Research.