The endoplasmic reticulum is a heterogeneous compartment with respect to the distribution of its Ca2+-handling proteins, namely the Ca2+-binding proteins, the Ca2+ pumps and the Ca2+ release channels. The nonuniform distribution of these proteins may explain the functional heterogeneity of the endoplasmic reticulum, such as the generation of spatially complex Ca2+ signals, Ca2+ homeostasis, and protein folding and quality control.
Introduction
In the past, the endoplasmic reticulum was viewed as a single, continuous, and homogeneous compartment with a uniform Ca2+ store. However, such a simplistic view does not explain the functional heterogeneity of this complex organelle. In particular, the endoplasmic reticulum must provide a pool of rapidly exchanging Ca2+ for signal generation, while concurrently maintaining areas within its lumen with stably high Ca2+ levels for proper protein folding and processing (Rooney and Meldolesi, 1996). Today it is widely accepted that although the endoplasmic reticulum is physically continuous (i.e., lumenally connected) (Subramanian and Meyer, 1997), it is spatially and functionally heterogeneous (Villa et al., 1993; Meldolesi and Pozzan, 1998; Baumann and Walz, 2001; Blaustein and Golovina, 2001). This heterogeneity may be established by the nonuniform distribution of endoplasmic reticulum Ca2+-handling proteins: (1) the Ca2+-binding proteins, such as calreticulin, calsequestrin, glucose-regulated protein 78 and 94 (Grp78/BiP and Grp94), protein disulfide isomerase (PDI),* and proteins belonging to the PDI-like family, ERp72, ERp57, and the newly identified ERp29 (Sargsyan et al., 2002); (2) the Ca2+ uptake channels, sarcoplasmic/endoplasmic reticulum Ca2+-transporting ATPases (SERCAs), and (3) the Ca2+ release channels, inositol 1,4,5-trisphosphate (InsP3) receptors and ryanodine receptors. A nonuniform distribution of these Ca2+-handling proteins necessarily divides the endoplasmic reticulum into subdomains, which extend beyond the classical divisions of rough endoplasmic reticulum, smooth endoplasmic reticulum, and nuclear envelope. Unique accumulations of Ca2+-handling proteins in the endoplasmic reticulum may determine special areas of this membrane system, involved in either Ca2+ signaling, Ca2+ homeostasis, or protein folding and processing.
Ca2+ signaling
Spatially and temporally complex Ca2+ signals generated by the endoplasmic reticulum underlie a diversity of cellular processes (Berridge et al., 2000; Johnson and Chang, 2000). Such Ca2+-dependent pathways include muscle contraction, secretion, proliferation, apoptosis, cell adhesion, differentiation, motility, cellular metabolism, fertilization, and control of gene expression (Johnson and Chang, 2000). It is an enormous task for the cell to be able to encode information in the form of Ca2+ waves and oscillations, such that signal fidelity is maintained and the correct cellular outcome is achieved. The endoplasmic reticulum is able to solve this problem by establishing a nonuniform distribution of its Ca2+-handling proteins, thus creating spatially heterogeneous Ca2+ stores within its lumen (Pezzati et al., 1997; Montero et al., 1997; Golovina and Blaustein, 2000; Blaustein and Golovina, 2001). Heterogeneity in endoplasmic reticulum lumenal Ca2+ has previously been demonstrated using electron energy loss imaging (Pezzati et al., 1997), and such store heterogeneity is sufficient to generate Ca2+ signals with enough complexity to control various Ca2+-dependent cellular processes (Johnson and Chang, 2000). Furthermore, these Ca2+ signals may play a critical role not only in the cytoplasm but also in the endoplasmic reticulum lumen (Corbett and Michalak, 2000).
The pancreatic acinar cell is a model cell type used to study the effects of the spatial heterogeneity of Ca2+-handling proteins on Ca2+-dependent events and has been extensively characterized by the group of Petersen (Belan et al., 1996; Petersen et al., 1999) and others (Kasai et al., 1993; Nathanson et al., 1994; Lee et al., 1997a,b; Leite et al., 1999). The pancreatic acinar cell has special Ca2+ release sites in its apical region, termed the trigger zone, which are enriched in InsP3 receptors (Nathanson et al., 1994; Lee et al., 1997b; Petersen et al., 1999), whereas its uptake sites, enriched in SERCAs, are located in the basal region (Lee et al., 1997a; Petersen et al., 1999). In pancreatic acinar cells, the focus has been on the distribution of Ca2+ pumps and channels and their critical role in the generation of spatially complex Ca2+ waves and oscillations. We suggest here, however, that Ca2+ signal generation and regulation is an intricate process, which likely also requires input from the endoplasmic reticulum lumenal environment and its Ca2+-binding proteins (Simpson et al., 1997; Johnson and Chang, 2000).
The major Ca2+-binding protein of the endoplasmic reticulum of smooth muscle and nonmuscle cells is calreticulin (Milner et al., 1991), and its homologue in the sarcoplasmic reticulum of striated muscle is calsequestrin. Calreticulin may influence the action of both SERCAs and InsP3 receptors (Corbett and Michalak, 2000), thereby modulating Ca2+ uptake and release, respectively, and thus regulating Ca2+ signals. In particular, Camacho's group has shown that calreticulin, in the lumen of the endoplasmic reticulum, inhibits Ca2+ uptake by the SERCA2b pump, thus inhibiting the continued generation of Ca2+ waves (Camacho and Lechleiter, 1995). This may be due to a direct interaction between calreticulin and the lumenal COOH-terminal tail of the SERCA2b isoform. In support of this interaction, calreticulin has been shown to colocalize with SERCA2b (John et al., 1998). In addition, a recent report indicates that a mathematical model, developed from single cell Ca2+ dynamics, has predicted an interaction between calreticulin and the SERCA pump (Baker et al., 2002). This model suggests that calreticulin alters the pump's affinity for Ca2+, thus regulating Ca2+ oscillations. The effects of calreticulin expression on InsP3 receptor activity may also be direct. Cell subfractionation experiments have revealed that calreticulin copurifies with InsP3 binding sites (Enyedi et al., 1993), whereas double immunolabeling experiments have shown that calreticulin colocalizes with the InsP3 receptor in the acrosome, in the equatorial segment, and in cytosolic vesicles of human spermatozoa (Naaby-Hansen et al., 2001). In summary, although the strategic placement of Ca2+ pumps and channels is imperative in generating spatially complex Ca2+ signals (Johnson and Chang, 2000), it is the responsibility of Ca2+-binding proteins to provide a releasable pool of Ca2+ near the release channels, and to modulate the activity of the pumps and channels to regulate Ca2+ waves and oscillations.
The arrangement of Ca2+-handling proteins in the endoplasmic reticulum creates specialized Ca2+-handling subdomains. For example, in astrocytes and oligodendrocytes, Ca2+ wave amplification sites exist along the endoplasmic reticulum, which are enriched in calreticulin, SERCAs, and InsP3 receptors and thus exhibit elevated Ca2+ release kinetics (Simpson et al., 1997, 1998). This organization between the InsP3 receptor and calreticulin is similar to the specialization seen in the junctional sarcoplasmic reticulum in striated muscle between calsequestrin and the ryanodine receptor (Allen and Katz, 2000; Gatti et al., 2001). Calreticulin is excluded from these junctional areas, and is found in the longitudinal sarcoplasmic reticulum, from where calsequestrin is excluded (Allen and Katz, 2000). The mechanisms underlying the heterogeneous distribution of endoplasmic reticulum Ca2+-binding proteins have only recently been elucidated for calsequestrin. The term condensation is used to refer to the head-to-tail oligomerization of calsequestrin, which is responsible for creating dense cores of the protein that are heterogeneously located throughout the sarcoplasmic reticulum (Gatti et al., 2001). Other endoplasmic/sarcoplasmic reticulum resident proteins may utilize a similar mechanism to establish a nonuniform distribution.
Multifunctionality of the endoplasmic reticulum
The heterogeneous distribution of Ca2+-handling proteins organizes the endoplasmic reticulum into various functional domains, some responsible for Ca2+ signaling, some for Ca2+ homeostasis, and others for protein folding and quality control. For example, some parts of the endoplasmic reticulum that are enriched in InsP3 receptors, SERCAs, and certain Ca2+-binding proteins may be responsible for rapid Ca2+ uptake and release, whereas other regions enriched only in Ca2+-binding proteins may be left to carry out the housekeeping functions of the endoplasmic reticulum (i.e., protein processing), which are also Ca2+-dependent but are shielded from fluctuations in Ca2+ concentration (Rooney and Meldolesi, 1996). Therefore, the distribution of resident Ca2+-binding proteins may also be potentially important in protein folding and quality control, as most of these proteins are involved in aspects of protein folding and maturation (i.e., chaperoning), and are enriched in the rough endoplasmic reticulum, the site of protein synthesis (Opas et al., 1991; Baumann and Walz, 2001), but are nonuniformly concentrated within this compartment (Baumann and Walz, 2001). Chaperones are weakly associated with one another and form a matrix in which they become embedded, resulting in their increased local concentration (Baumann and Walz, 2001). Fig. 1 compares the distribution of calreticulin to that of PDI and Grp94 within the endoplasmic reticulum of mouse embryonic fibroblasts. PDI exhibits an overlapping yet distinct distribution with calreticulin, whereas Grp94 exhibits virtually complete overlap with calreticulin (Fig. 1, A and B, respectively). The differential distribution of various lumenal Ca2+-binding proteins may be of great physiological importance for Ca2+ homeostasis. Calreticulin and Grp94 are two major Ca2+ storage proteins of the endoplasmic reticulum, by virtue of their extremely high Ca2+ capacity (Milner et al., 1991; Argon and Simen, 1999), which is not matched by any other endoplasmic reticulum protein. Fluctuations of Ca2+ concentration in the lumen of the endoplasmic reticulum, which are ultimately regulated by Ca2+-handling proteins, may have profound effects on the structure and function of integral and lumenal (peripheral) membrane proteins and likely contribute to the functional heterogeneity of the endoplasmic reticulum (Corbett and Michalak, 2000). Furthermore, endoplasmic reticulum functions involved with protein processing have been shown to underlie various diseases such as Alzheimer's disease, Parkinson's disease, and α1-antitrypsin deficiency (Kopito and Ron, 2000; Paschen and Frandsen, 2001; Rutishauser and Spiess, 2002). The unique distributions of chaperones may be significant in such protein folding pathologies.
Additionally, it was recently postulated that the endoplasmic reticulum and sarcoplasmic reticulum may play different roles in cells in which they coexist (Jaconi et al., 2000; Mesaeli et al., 2001). For example, in cardiomyocytes, the sarcoplasmic reticulum is involved in the classical role of excitation–contraction coupling, whereas the endoplasmic reticulum has been suggested to perform housekeeping functions, such as protein turnover (Mesaeli et al., 2001). The heterogeneous distribution of calsequestrin and calreticulin in the heart, along with their respective release channels, ryanodine receptors in the sarcoplasmic reticulum and InsP3 receptors in the endoplasmic reticulum, may be responsible for this duality of function. Studies on calreticulin deficient cardiomyocytes show that these cells exhibit spontaneous contraction, thus suggesting a functional sarcoplasmic reticulum (Mesaeli et al., 1999). However, calreticulin null fibroblasts show impaired Ca2+ homeostasis by the endoplasmic reticulum (Nakamura et al., 2001), and this may also be the case for calreticulin-null cardiomyocytes. At the early stages of development, the endoplasmic reticulum may play a critical role in protein and lipid synthesis as well as in the regulation of Ca2+-dependent transcriptional processes. These will play only a minor role in the mature heart. In mature muscle, the sarcoplasmic reticulum becomes responsible for regulating Ca2+uptake and release and excitation–contraction coupling.
Interestingly, heterogeneity in compartmentalization may already be evident during development. In human oocytes, for example, the endoplasmic/sarcoplasmic reticulum–associated Ca2+-binding proteins are nonuniformly distributed (Balakier et al., 2002). Calreticulin is predominant in the cell cortex, whereas calsequestrin is found throughout the entire cytoplasm (Fig. 2). Such differential distribution of calreticulin and calsequestrin indicates that oocytes have two distinct Ca2+ storage compartments: one enriched in calreticulin (and the InsP3 receptor), and the other in calsequestrin (and the ryanodine receptor). Calsequestrin may localize to certain regions by the condensation mechanism, whereas calreticulin may utilize specific protein–protein interactions to achieve its localization in the endoplasmic reticulum. Interestingly, calnexin, like calreticulin, is also predominant in the cell cortex where it is found in a peculiar trilaminar arrangement (Fig. 2). The differential distribution of these proteins may reflect their functional differences. Calnexin is a chaperone, calsequestrin is a Ca2+ storage protein, whereas calreticulin carries out both functions. Thus, certain regions of the endoplasmic reticulum may be involved in intensive protein processing required for oocyte maturation and embryo development, whereas other regions may be involved in Ca2+ homeostasis. In conclusion, the spatial heterogeneity of the endoplasmic reticulum may be established early on in development, and this warrants further investigation. Deciphering the organization of Ca2+-handling proteins in the endoplasmic reticulum may hold a clue to our understanding of the generation of the multifunctionality of this membrane system.
Acknowledgments
M. Michalak is a Canadian Institutes of Health Research (CIHR) Senior Scientist of the Alberta Heritage Foundation for Medical Research. M. Opas is a member of the Heart and Stroke/Richard Lewar Centre of Excellence. This work was supported by grants form the CIHR (to M. Michalak and M. Opas) and from the Heart and Stroke Foundations of Alberta (to M. Michalak) and Ontario (to M. Opas).
Abbreviations used in this paper: InsP3, inositol 1,4,5-trisphosphate; PDI, protein disulfide isomerase; SERCA, sarcoplasmic/endoplasmic reticulum Ca2+-transporting ATPase.