Sdmg1 is a component of secretory granules in mouse secretory exocrine tissues

Authors


Abstract

Sdmg1 is a conserved eukaryotic transmembrane protein that is mainly expressed in the gonads where it may have a role in mediating signaling between somatic cells and germ cells. In this study we demonstrate that secretory exocrine cells in the pancreas, salivary gland, and mammary gland also express Sdmg1. Furthermore, we show that Sdmg1 expression is up-regulated during pancreas development when regulated secretory granules start to appear, and that Sdmg1 colocalizes with secretory granule markers in adult pancreatic acinar cells. In addition, we show that Sdmg1 co-purifies with secretory granules during subcellular fractionation of the pancreas and that Sdmg1 and the secretory granule marker Vamp2 are localized to distinct subdomains in the secretory granule membrane. These data suggest that Sdmg1 is a component of regulated secretory granules in exocrine secretory cells and that the developmental regulation of Sdmg1 expression is related to a role for Sdmg1 in post-Golgi membrane trafficking. Developmental Dynamics 238:223–231, 2009. © 2008 Wiley-Liss, Inc.

INTRODUCTION

The murine Sdmg1 gene encodes a member of a family of transmembrane proteins that has been conserved throughout eukaryotic evolution from yeast to man (Best et al.,2008). Despite the evolutionary conservation of this gene family, little is known about the function of these proteins and what their cellular role might be. We, and others, have identified Sdmg1 in screens to identify genes that are expressed in a sexually dimorphic manner in the developing embryonic gonad at the time of gonadal sex determination (Svingen et al.,2007; Best et al.,2008). During male embryonic gonad development Sdmg1 is expressed in the supporting cell lineage from around 12.5 days post coitum (dpc), soon after the supporting cells initiate their male-specific differentiation into Sertoli cells (Svingen et al.,2007; Best et al.,2008). Sdmg1 expression is maintained in the male Sertoli cells throughout the remainder of embryonic development and into adulthood (Svingen et al.,2007; Best et al.,2008). In contrast, Sdmg1 is not expressed in the female gonad during embryonic development, but is switched on the female supporting cells, the granulosa cells, a few days after birth (Best et al.,2008). Sdmg1 expression is maintained in the ovarian granulosa cells during folliculogenesis as the oocyte grows and matures (Best et al.,2008).

Functional analysis of Sdmg1 in the SK11 Sertoli cell line suggests that the Sdmg1 protein plays a role in membrane trafficking (Best et al.,2008). Sdmg1 is predicted to contain seven transmembrane domains, and is localized to endosomes in the SK11 Sertoli cell line and in embryonic Sertoli cells (Best et al.,2008). RNAi-mediated knock-down of Sdmg1 in the SK11 Sertoli cell line perturbs the subcellular localization of Stx2 (Best et al.,2008), a member of the soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptor (SNARE) family of transmembrane proteins. The mammalian genome contains over 30 different SNAREs, and these proteins are involved in mediating fusion between membranes at different points throughout the membrane trafficking pathway (Chen and Scheller,2001; Jahn and Scheller,2006). Stx2 is implicated in mediating fusion of secretory vesicles with the plasma membrane (Hansen et al.,1999), and knock-down of Sdmg1 impairs the ability of the SK11 Sertoli cell line to masculinize germ cells isolated from female embryos, a phenotype that presumably reflecting defects in secretion (Best et al.,2008). Taken together the existing data suggests that Sdmg1 may play a role in post-Golgi membrane trafficking in the supporting cell lineage, and may be involved in mediating some of the signaling events that occur between supporting cells and germ cells during gametogenesis.

During the course of our analysis of the role of Sdmg1 in gonad development, we noticed that an electronic database of microarray-based expression data suggested that Sdmg1 might be expressed in secretory exocrine tissues, such as the pancreas, salivary glands, and mammary glands, as well as the testis (Su et al.,2002). All of these exocrine secretory tissues contain electron-dense vesicular structures known as secretory granules that store cargo for regulated secretion (Arvan and Castle,1998; Burgoyne and Morgan,2003). Regulated secretion is a specialized part of the post-Golgi membrane trafficking pathway that allows some exocrine cells to release stored enzymes and proteins into ductal systems in a controlled manner in response to a physiological stimulus. Regulated secretion is also involved in neuronal survival and learning, in hormone release from endocrine cells, in sperm capacitation, and in the cortical reaction in fertilized eggs (Arvan and Castle,1998; Burgoyne and Morgan,2003). Furthermore, developmentally important signaling molecules, such as basic fibroblast growth factor and stem cell factor are stored in secretory granules in some cell types and secreted in a regulated manner (Qu et al.,1998; de Paulis et al.,1999).

Secretory granules containing insoluble aggregates of secretory granule cargo are thought to originate from the trans-Golgi network. During secretory granule maturation, vesicles bud from these immature secretory granules to remove any mis-sorted proteins and excess solute, and the secretory granule cargo condenses further. The mature secretory granules are stored in the cytoplasm of the secretory cell and typically only fuse with the plasma membrane and secrete their cargo in response to extracellular signals (Arvan and Castle,1998; Burgoyne and Morgan,2003). During secretory granule exocytosis, there appears to be little diffusion from the secretory granule membrane into the plasma membrane and, once the secretory granule cargo is expelled, the membrane of the secretory granule remains distinct from the plasma membrane and is eventually endocytosed and recycled (Thorn et al.,2004). Around 80–90% of all protein secretion in the secretory exocrine cells of the pancreas and parotid salivary gland occurs through secretory granule exocytosis (Van Nest et al.,1980; von Zastrow and Castle,1987); therefore, proteins involved in the regulated secretory pathway are expressed at high levels in these secretory exocrine cells (Wang et al.,2004,2007).

In this manuscript, we demonstrate that the secretory exocrine cells in the pancreas, salivary gland, and mammary gland express Sdmg1. Furthermore, we show that Sdmg1 expression is up-regulated during pancreas development at the time when regulated secretory granules start to appear and that Sdmg1 is a component of regulated secretory granules in the exocrine pancreas. These data further strengthen the hypothesis that Sdmg1 plays a role in post-Golgi membrane trafficking.

RESULTS

Sdmg1 Is Expressed in Secretory Exocrine Tissues in Adult Mice

During our analysis of Sdmg1 expression in gonad development (Best et al.,2008), we noticed that electronic databases for microarray-based gene expression profiling suggest that Sdmg1 is expressed in some nongonadal adult tissues such as stomach, intestine, bladder, prostate, pancreas, salivary gland, mammary gland (Su et al.,2002). These tissues were not present in the panels of tissue used previously to assess the expression profile of Sdmg1 (Svingen et al.,2007; Best et al.,2008). We noted that many of the tissues, such as the pancreas, salivary gland, and mammary gland, contain specialized secretory exocrine cells. As Sdmg1 has a potential role in post-Golgi membrane trafficking (Best et al.,2008), we decided to investigate Sdmg1 expression in this group of tissues further.

To confirm that Sdmg1 mRNA is expressed in these secretory exocrine tissues, we performed reverse transcriptase-polymerase chain reaction (RT-PCR) on RNA isolated from adult pancreas, from adult parotid salivary gland, and from mammary gland isolated from pregnant adult females late in gestation (15.5 dpc). Heart was used as a negative control for Sdmg1 expression, and testis as a positive control (Svingen et al.,2007; Best et al.,2008). Sdmg1 expression was detected in the pancreas, salivary gland, and mammary gland in addition to the testis using this method (Fig. 1A). No expression was detected in the negative control heart tissue (Fig. 1A), or if reverse transcriptase was omitted from the RT-PCR (Fig. 1A). Thus, Sdmg1 mRNA is expressed in the pancreas, parotid salivary gland, and mammary gland in adult mice.

Figure 1.

Sdmg1 is expressed in secretory exocrine tissues. A: Reverse transcriptase-polymerase chain reaction (RT-PCR) for Sdmg1 mRNA in adult secretory exocrine tissues. Expression of Sdmg1 (221-bp RT-PCR product) was assayed in pancreas, parotid salivary gland, and mammary gland by RT-PCR. Heart was included as negative control tissue, testis as positive control tissue. Gapdh expression (452-bp RT-PCR product) was used as a positive control for RNA integrity. RT-PCR performed in the absence of reverse transcriptase (−RT) and using distilled water as a template in the PCR step were included as negative controls. B: Western blot for Sdmg1 protein in adult secretory exocrine tissues. A total of 25 μg of protein extract from pancreas, salivary gland, and mammary gland tissues were assayed for Sdmg1 protein expression; heart was included as negative control tissue, testis as positive control tissue. Sdmg1 migrates at ∼110 kDa in Western blots, the Gapdh positive control migrates at ∼37 kDa. Mammary gland was isolated from a 15.5 days post coitum (dpc) pregnant female for RT-PCR and Western blot analysis. Molecular weight markers are indicated on the left of each panel.

We next tested whether the presence of Sdmg1 mRNA in adult secretory exocrine tissues correlated with the presence of Sdmg1 protein by performing Western blots on protein extracts from the same set of adult secretory exocrine tissues using anti-Sdmg1 antibodies (Best et al.,2008). Although the predicted molecular weight of Sdmg1 is 49 kDa, N-glycosylation causes Sdmg1 to migrate at ∼110 kDa in Western blots of testis protein extracts (Best et al.,2008). Anti-Sdmg1 antibodies detected a band migrating at ∼110 kDa in protein extracts from the pancreas, salivary gland and mammary gland in addition to the testis protein extract (Fig. 1B). No bands were detected with anti-Sdmg1 antibodies in the negative control heart protein extract (Fig. 1B). Thus, Sdmg1 protein is present in the pancreas, parotid salivary gland, and mammary gland in adult mice, and correlates with the presence of Sdmg1 mRNA in these tissues.

Sdmg1 Is Localized to the Apical Cytoplasm of Secretory Exocrine Cells in the Adult Pancreas, Salivary Gland, and Mammary Gland

We next investigated which cell type in each of the pancreas, salivary gland, and mammary gland was expressing Sdmg1 by immunohistochemistry. The anti-Sdmg1 antibody used for immunohistochemistry was raised against the C-terminal cytoplasmic domain of the Sdmg1 transmembrane protein (Best et al.,2008). This antibody appears to specifically recognize endogenous Sdmg1 as assessed by the similarity between the Sdmg1 mRNA distribution obtained by in situ hybridization and Northern blotting, and the anti-Sdmg1 antibody signals obtained by immunohistochemistry and Western blotting in a variety of tissues (Best et al.,2008).

The adult pancreas is mainly composed of clusters of exocrine acinar cells that secrete digestive enzymes into collecting ducts. The digestive enzymes are stored in abundant secretory granules present in the apical cytoplasm of the pancreatic acinar cells. Immunohistochemistry for Sdmg1 in the adult pancreas suggests that Sdmg1 is expressed in the exocrine acinar cells of the adult pancreas (Fig. 2F). At high magnification, anti-Sdmg1 staining could be seen as abundant punctate structures in the apical cytoplasm of the pancreatic acinar cells (Fig. 2K). No staining was seen in the salivary gland when nonspecific rabbit IgG was used as a negative control (Fig. 2A).

Figure 2.

Sdmg1 is localized to punctate structures in the apical cytoplasm of secretory exocrine cells. Immunohistochemistry for Sdmg1 in secretory exocrine tissues. AO: Sections of pancreas (A,F,K), parotid salivary gland (B,G,L) and mammary gland (C–E,H–J,M–O) from adult mice were immunostained with rabbit IgG as a negative control (A–E), or with rabbit anti-Sdmg1 antibodies (F–O) as indicated. Mammary gland was isolated from a virgin female (C,H,M), a 15.5 days post coitum (dpc) pregnant female (D,I,N), or a lactating female (E,J,O). Bound antibodies can be visualized as a brown precipitate, nuclei are counterstained with hematoxylin (blue). Scale bars = 100 μm in A–J, 10 μm in K–O.

Like the pancreas, the parotid salivary gland is composed mainly of clusters of exocrine acinar cells that secrete their contents into collecting ducts. Digestive enzymes such as salivary amylase are stored in the abundant secretory granules present in the apical cytoplasm of the parotid salivary gland acinar cells. Immunohistochemistry for Sdmg1 suggests that Sdmg1 is present in the exocrine acinar cells of the parotid salivary gland (Fig. 2G). At high magnification, anti-Sdmg1 staining could be seen as abundant punctate structures in the apical cytoplasm of the salivary gland acinar cells (Fig. 2L). No staining was seen in the salivary gland when nonspecific rabbit IgG was used as a negative control (Fig. 2B).

We also investigated the expression of Sdmg1 protein in the adult mammary gland. The mammary gland undergoes significant changes during the reproductive life cycle of an adult female (Anderson et al.,2007). In virgin adult females, the mammary gland comprises a branched ductal epithelium with alveolar buds embedded within the mammary fat pad. During pregnancy, there is extensive proliferation of the ductal epithelium, an increase in ductal branching, and the alveolar buds differentiate into lobuloalveolar structures that are capable of milk secretion. Lactation is initiated at around the time of parturition, and after weaning, the ductal epithelium undergoes extensive cell death as the ducts regress back to their normal state (Anderson et al.,2007). The alveolar epithelial cells are responsible for milk secretion in the mammary gland, and these cells use both regulated and constitutive exocytic pathways for milk secretion (Turner et al.,1992). Secretory differentiation of the alveolar epithelial cells is initiated transiently in virgin adult mammary glands and progresses during pregnancy to full differentiation around the time of parturition (Robinson et al.,1995). Immunohistochemistry for Sdmg1 in the mammary glands isolated from virgin, pregnant, and lactating adult females suggests that Sdmg1 protein is expressed at each of these stages in mammary gland development (Fig. 2H–J). Furthermore, Sdmg1 is expressed in the alveolar epithelial cells responsible secretion in this tissue (Fig. 2H–J) and is localized to the apical cytoplasm of these cells (Fig. 2M–O). In contrast to the anti-Sdmg1 staining in the pancreas and the salivary gland, no discrete punctate structures could be distinguished in the anti-Sdmg1 staining in the mammary gland. No staining was seen in the mammary gland tissues when nonspecific rabbit IgG was used as a negative control (Fig. 2C–E).

Although the anti-Sdmg1 immunohistochemistry shown in Figure 2 is consistent with Sdmg1 expression in the pancreas, salivary gland, and mammary gland, it is possible that the anti-Sdmg1 antibody used in these experiments is cross-reacting with a different protein that is expressed in these secretory exocrine tissues. To address this possibility, we performed immunohistochemistry for Sdmg1 using a second anti-Sdmg1 antibody (Supp. Fig. S1, which is available online). This also resulted in abundant anti-Sdmg1 staining in the apical cytoplasm of pancreatic acinar cells, salivary gland acinar cells, and mammary gland alveolar epithelial cells (Supp. Fig. S1). Thus, the anti-Sdmg1 staining patterns observed using immunohistochemistry are likely to represent the localization of Sdmg1 protein in secretory exocrine tissues.

The immunohistochemistry for Sdmg1 in the pancreas, salivary gland, and mammary gland confirms the RT-PCR and Western blotting data that Sdmg1 is expressed in these tissues (Fig. 1) and shows that the cells that are expressing Sdmg1 in these tissues are the secretory exocrine cells. Furthermore, the subcellular localization of Sdmg1 in each of these secretory exocrine cells bears some resemblance to the subcellular localization of Vamp8, a SNARE associated with secretory granules in these cell types (Wang et al.,2004,2007).

Expression of Sdmg1 in the Developing Pancreas

The size and distribution of the punctate structures labeled with anti-Sdmg1 antibodies in the acinar cells of the adult pancreas is consistent with the size and distribution of secretory granules in this cell type. To investigate whether the expression of Sdmg1 might be related to the presence of secretory granules in the pancreatic acinar cells, we investigated the expression profile of Sdmg1 during embryonic development of the pancreas. The pancreas arises from evaginations of the posterior foregut endoderm, with the appearance of a dorsal, then a ventral pancreatic bud occurring at around 9.5–10.5 dpc in mice. The pancreatic endodermal epithelium, which contains distinct endocrine and exocrine cells, then proliferates and branches. The exocrine pancreatic cells start to form morphologically distinct acini and ducts around 14.5 dpc, and exocrine secretory granules start to appear at this stage (Slack,1995; Habener et al.,2005). We, therefore, examined Sdmg1 expression in the developing pancreas from 12.5–16.5 dpc, the developmental window when the exocrine pancreas morphologically differentiates into acini and acquires secretory granules. We were able to detect weak anti-Sdmg1 immunostaining in embryonic pancreas at 12.5 dpc, and strong anti-Sdmg1 immunostaining in the assembling pancreatic acini at 14.5 dpc (Fig. 3A,B,D,E). We were also able to detect robust anti-Sdmg1 immunostaining in the pancreatic acini at 16.5 dpc (Fig. 3C,F). No staining was seen at any of these developmental stages when nonspecific rabbit IgG was used as a negative control (Fig. 3G–I). Thus, up-regulation of Sdmg1 expression in the developing embryonic pancreas appears to be coincident with the formation of the exocrine acini, and the appearance of secretory granules in the developing acinar cells.

Figure 3.

Sdmg1 expression is up-regulated during embryonic pancreatic acinar cell differentiation. Immunostaining for Sdmg1 during embryonic pancreas development. AI: Cryosections of developing pancreas from 12.5 days post coitum (dpc; A,D,G), 14.5 dpc (B,E,H) and 16.5 dpc (C,F,I) embryos were immunostained with nonspecific rabbit IgG as a negative control (G–I), or with rabbit anti-Sdmg1 antibodies (A–F). Bound antibodies can be visualized as a green signal, nuclei are counterstained with DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride; red). Scale bars = 100 μm.

Sdmg1 Colocalizes With Secretory Granule Markers in the Adult Pancreas

The immunohistochemistry staining pattern for Sdmg1 in the adult pancreas, and the developmental expression profile of Sdmg1 during pancreas development is consistent with Sdmg1 being a component of the regulated secretory granules that are abundant in this cell type. We, therefore, attempted to colocalize Sdmg1 with markers of secretory granules and other membrane trafficking compartments by immunofluorescence on adult pancreas cryosections. Immunostaining adult pancreas cryosections with anti-Sdmg1 antibodies gave similar results to the anti-Sdmg1 immunohistochemistry performed on wax sections of the pancreas: Sdmg1 is predominantly localized to granular structures in the apical cytoplasm of the pancreatic acinar cells (Figs. 2F,K, 4A,D). In addition, anti-Sdmg1 immunostaining on pancreas cryosections suggests that there is also a small amount of Sdmg1 present in the basolateral cytoplasm of these cells (Fig. 4G,J).

Figure 4.

Sdmg1 colocalizes with the secretory granule marker Vamp2 in adult pancreatic acinar cells. Immunostaining for Sdmg1 and membrane trafficking markers in adult pancreatic acinar cells. A,B,D,E,G,H,J,K: Pancreas cryosections were immunostained with anti-Sdmg1 antibodies (A,D,G,J, green) and antibodies to membrane trafficking markers (B,E,H,K, red). C,F,I,L: DNA was counterstained with DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride; blue) in the merged images. A–I: Vamp2 is a marker for secretory granules (A–C), EEA1 is a marker for early endosomes (D–F), Tfrc is a marker for recycling endosomes (G–I), and Lamp1 is a marker for late endosomes/lysosomes (J–L). Scale bar = 10 μm.

We tested whether Sdmg1 colocalizes with secretory granules in the adult pancreas by double labeling pancreas cryosections with Sdmg1 and the secretory granule SNARE Vamp2 (Gaisano et al.,1996; Weng et al.,2007). Antibodies to Sdmg1 and to Vamp2 both labeled abundant punctate structures in the apical cytoplasm of pancreatic acinar cells and exhibited extensive colocalization (Fig. 4A–C). Thus, the abundant punctate structures in the apical cytoplasm of pancreatic acinar cells that are labeled with anti-Sdmg1 antibodies are likely to be secretory granules.

As we have shown that Sdmg1 is present in endosomes in Sertoli cells (Best et al.,2008), we also tested whether the abundant punctate anti-Sdmg1 staining in the apical cytoplasm of pancreatic acinar cells might represent localization to endosomes rather than to secretory granules by colocalizing Sdmg1 with the endosomal markers EEA1, Tfrc1 and Lamp1 (Clague,1998) in the adult pancreas. Whereas Sdmg1 labels abundant punctate structures in the apical cytoplasm of pancreatic acinar cells, the early endosome marker EEA1 labels sparse punctate structures in the apical cytoplasm (Fig. 4D–F). Thus, although we cannot exclude the possibility that some Sdmg1 is present in early endosomes in addition to secretory granules in the apical cytoplasm of pancreatic acinar cells, early endosome localization alone would not be sufficient to account for the abundant anti-Sdmg1 staining in the apical cytoplasm of pancreatic acinar cells. Similarly, the small amounts of anti-Sdmg1 staining present in the basolateral cytoplasm of the pancreatic acinar cells show some overlap with the recycling endosome marker Tfrc1 (Fig. 4G–I), and the late endosome/lysosome marker Lamp1 (Fig. 4K,L). Thus, although some Sdmg1 may be present in endosomes in pancreatic acinar cells, these endosomal compartments do not appear to be responsible for the abundant punctate anti-Sdmg1 staining seen in the apical cytoplasm of the pancreatic acinar cells. Rather, the abundant punctate anti-Sdmg1 staining in the apical cytoplasm of pancreatic acinar cells suggests that Sdmg1 is a component of regulated secretory granules in this cell type.

Sdmg1 Is a Component of Pancreatic Secretory Granules

To confirm whether Sdmg1 might be a component of secretory granules in pancreatic acinar cells, we followed the distribution of Sdmg1 during biochemical purification of pancreatic secretory granules. Pancreatic secretory granules can be purified by Percoll gradient separation of a particulate fraction of pancreatic postnuclear supernantant (Chen et al.,2003). The dense white band of secretory granules was collected in fraction 3 of the Percoll gradient, and microscopic inspection of the Percoll gradient fractions confirmed that large vesicles around 1 μm in diameter were enriched and abundant in this fraction (Fig. 6; and data not shown). Western blotting of the Percoll gradient fractions for the secretory granule SNARE Vamp8 (Wang et al.,2004; Weng et al.,2007) confirmed that secretory granules were enriched in fraction 3 of the Percoll gradient. We also detected some Vamp8 in lighter fractions toward the top of the Percoll gradient. It is not clear if the presence of Vamp8 in these fractions represents a different type of membrane trafficking intermediate or organelle present in the Percoll gradient, or if this might represent secretory granules that have lysed or become damaged during the purification procedure. Western blotting of the same Percoll gradient fractions for Sdmg1 showed that Sdmg1 is present in the secretory granule fraction (Fig. 5). We could not detect Sdmg1 in other fractions of the Percoll gradient, although we cannot exclude the possibility that these fractions contain levels of Sdmg1 that are below our limits of detection. However, this subcellular fractionation shows that Sdmg1 is enriched in purified populations of pancreatic secretory granules.

Figure 5.

Sdmg1 co-purifies with secretory granules during subcellular fractionation of the adult pancreas. Western blot of subcellular fractions from adult pancreas. The particulate fraction of a postnuclear supernatant from adult pancreas was separated on a Percoll gradient, and six fractions were collected (F1 from the bottom of the gradient, F6 from the top) then Western blotted with anti-Vamp8 and anti-Sdmg1 antibodies. Secretory granules (SG) are enriched in fraction F3. Protein molecular weight markers are indicated on the left of each panel.

Figure 6.

Sdmg1 is localized to discrete subdomains on the membrane of purified pancreatic secretory granules. Immunostaining of purified pancreatic secretory granules. A,C: The secretory granule fraction (F3) from the pancreas subcellular fractionation was immunostained with mouse anti-Vamp2 (red) and rabbit anti-Sdmg1 (green) antibodies. B,D: Secretory granules in the same field were visualized using differential interference contrast (DIC) light microscopy. A,B: Low magnification image of purified secretory granules. C,D: High magnification image of purified secretory granules. Anti-Vamp2 and anti-Sdmg1 signals can be seen in distinct and separate subdomains of the secretory granule membrane (arrowhead in C,D). In some granules, multiple anti-Vamp2 and anti-Sdmg1 signals appear to surround the entire secretory granule membrane (arrows in C,D). Scale bars = 10 μm.

To confirm that the co-purification of Sdmg1 and secretory granules represents Sdmg1 being a component of these structures, rather than a component of an unrelated organelle co-purifying with the secretory granules, we performed immunostaining on the secretory granule fraction from the Percoll gradient. The secretory granule fraction contains abundant vesicular structures around 1 μm in diameter. The secretory granule marker Vamp2 (Gaisano et al.,1996; Weng et al.,2007) colocalized with these vesicular structures, confirming that they are indeed secretory granules (Fig. 6A,B). As expected from data reported in the literature (Weng et al.,2007), in higher magnification images, regions of anti-Vamp2 immunostaining could be seen to be associated with the membranes rather than the lumen of the secretory granules, and anti-Vamp2 immunostaining was usually restricted to distinct regions of the secretory granule membrane rather than being homogeneously distributed around the secretory granule membrane (Fig. 6C,D). Anti-Sdmg1 immunostaining also colocalized with the 1 μm-diameter vesicles in the secretory granule fraction, suggesting that Sdmg1 is a component of secretory granules (Fig. 6C,D). Anti-Sdmg1 immunostaining is also associated with discrete regions of the secretory granule membrane in a similar manner to anti-Vamp2 immunostaining (Fig. 6C,D). Of interest, although anti-Vamp2 and anti-Sdmg1 signals were often present on the membrane of the same secretory granule, they were often present in different and separate subdomains (Fig. 6C,D, arrowheads). Some secretory granules appeared to possess multiple anti-Vamp2 and multiple anti-Sdmg1subdomains that encircled the entire granule (Fig. 6C,D, arrows). Other secretory granules possessed only anti-Sdmg1 or anti-Vamp2 staining (Fig. 6C,D), an observation that is consistent with a recent report suggesting that pancreatic acinar cells contain multiple subpopulations of secretory granules carrying distinct molecular markers (Weng et al.,2007). Taken together the co-purification of Sdmg1 with pancreatic secretory granules and the anti-Sdmg1 immunostaining of purified pancreatic secretory granules strongly suggest that Sdmg1 is a component of secretory granules in the adult pancreas.

DISCUSSION

This study describes the expression and subcellular localization of the transmembrane protein Sdmg1 in secretory exocrine tissues in the mouse. Sdmg1 is expressed in the supporting cell lineage of the testis and ovary in adult mice, but does not appear to be highly expressed in the any of the other organs previously tested (Svingen et al.,2007; Best et al.,2008). Here, we show that Sdmg1 mRNA and Sdmg1 protein are both present in three different secretory exocrine tissues: the pancreas, the salivary gland, and the mammary gland. We also show that, within these tissues, Sdmg1 protein is expressed in the cells involved in exocrine secretion. Furthermore, we show that Sdmg1 is a component of the specialized regulated secretory granules that are characteristic of these exocrine secretory cells.

The expression of Sdmg1 in the exocrine secretory cells in the pancreas, salivary gland, and mammary gland, and the localization of Sdmg1 to pancreatic secretory granules, bears some resemblance to the expression and subcellular localization of the Vamp8 SNARE (Wang et al.,2004; Weng et al.,2007). Expression of Vamp8 and Sdmg1 are also both up-regulated in a male-specific manner during embryonic gonad development (Best et al.,2008). Vamp8 is a general SNARE for regulated secretion in exocrine tissues, and Vamp8 knockout mice have defects in regulated secretion in several exocrine tissues (Wang et al.,2004; Weng et al.,2007). The similarity between the expression and subcellular localization of Sdmg1 and Vamp8 suggests that, like Vamp8, Sdmg1 may play a role in regulated secretion. However, further experiments are required to determine the function of Sdmg1 in these secretory exocrine cell types.

Immunostaining suggests that Sdmg1 is present in both secretory granules and endosomes in pancreatic acinar cells (Fig. 4), although it is not clear at present what proportion of Sdmg1 in the pancreas is present in secretory granules and what proportion is present in endosomes. In cells that lack secretory granules, Sdmg1 is localized to a subset of endosomes (Best et al.,2008). Similarly, Vamp8 is a component of secretory granules in secretory exocrine tissues (Wang et al.,2004; Weng et al.,2007), but is localized to a subset of endosomes in cell types that lack secretory granules (Wong et al.,1998; Antonin et al.,2000; Steegmaier et al.,2000). Both Sdmg1 and Vamp8 appear to have a function in cells lacking secretory granules (Antonin et al.,2000; Best et al.,2008), possibly related to the role of some endosomes in the secretory pathway in polarized epithelial cells (Ang et al.,2004; Lock and Stow,2005). However, it is not clear at present what the biochemical function of Sdmg1 might be in the endosomal membrane, or whether Sdmg1 performs a similar biochemical function in secretory granule membranes.

Our finding that anti-Sdmg1 immunostaining is localized to subdomains on the pancreatic secretory granule membrane is consistent with the staining patterns seen with anti-Vamp2 SNARE and anti-Vamp8 SNARE antibodies on secretory granule membranes (Weng et al.,2007), but distinct from the homogenous distribution of the small GTPase Rab3D around the secretory granule membrane (Chen et al.,2003). The simplest interpretation of this anti-Sdmg1 staining pattern is that there are subdomains of high and low Sdmg1 protein abundance within the secretory granule membrane. Alternatively, Sdmg1 protein may be homogenously distributed around the secretory granule membrane, but different protein complexes or protein modifications in some regions of the membrane may prevent the anti-Sdmg1 antibodies from binding. Lastly, the subdomains may be an in vitro artifact generated during the purification or fixation of the secretory granules. Subdomains of SNAREs and lipids within membranes have been reported in various organelles where they have been suggested to be involved in sorting membranes and cargo to different destinations, or in clustering similar proteins together to facilitate protein–protein interactions between membranes (Sonnichsen et al.,2000; Lang et al.,2001). Our data suggesting that secretory granule membranes may also contain discrete subdomains is, therefore, consistent with observations in other membrane trafficking systems. The data presented here also suggest that the anti-Sdmg1 and anti-Vamp2 subdomains in the secretory granule membrane are distinct from each other. Anti-Vamp2 and anti-Vamp8 subdomains also appear to be distinct from each other (Weng et al.,2007). It will be of interest to investigate whether the anti-Sdmg1 subdomains in the secretory granule membrane overlap with the anti-Vamp8 subdomains, or with any subdomains of other SNARE proteins present on the secretory granule membrane, and how these domains are arranged and oriented during secretory granule exocytosis.

EXPERIMENTAL PROCEDURES

Mice

Outbred CD1 mice (Charles River Laboratories) were naturally mated with noon on the day the vaginal plug was found termed 0.5 dpc.

RT-PCR

RNA was isolated from tissue using Trizol (Invitrogen) according to the manufacturer's instructions. RNA was treated with DNA-free DNase (Ambion), and then converted to cDNA using random primers and Superscript II reverse transcriptase (Invitrogen) according to the manufacturer's instructions. PCR primers for Sdmg1 were 5′-AGTGGCTTAAAGGAGACCATCA-3′ and 5′-TCAGCATGCGTTTCTCTATGTT-3′, PCR primers for Gapdh were 5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′.

Immunostaining

For immunofluorescence on cryosections, tissue was fixed with 4% paraformaldehyde in phosphate buffered saline (PBS) at 4°C, infiltrated with 20% sucrose in PBS, embedded in OCT compound (VWR), frozen on dry ice, then cryosectioned at 10 μm. After sectioning, the slides were briefly fixed with 3.7% formaldehyde solution in PBS, then washed with PBS, blocked (PBS, 1–10% serum, 0.1% Tween-20), and incubated with primary then secondary antibodies diluted in blocking solution. An Axioplan II fluorescence microscope (Carl Zeiss) equipped with a Coolsnap digital camera (Photometrics, Tucson, AZ) was used to acquire the images.

For immunohistochemistry on wax sections, tissue was fixed with Bouin's solution (Sigma-Aldrich) at room temperature then embedded in Paraplast wax (R.A. Lamb). Sections were cut at 7 μm, de-waxed with xylene, rehydrated through an ethanol series, and washed with PBS. Antigen retrieval was performed by boiling for 20 min in 0.1 M citrate buffer pH 6 in a microwave oven. Sections were then blocked, incubated with primary antibodies, and bound antibodies visualized using an Envision HRP-linked detection kit (Dako) followed by counterstaining with hematoxylin.

Antibodies

Antibodies raised against the C-terminal domain of Sdmg1 (Best et al.,2008) were used at 1 μg/ml for immunostaining. The SK1774 rabbit anti-Sdmg1 antibody (Best et al.,2008) was used in Supplementary Figure S1 to confirm anti-Sdmg1 immunostaining in secretory exocrine tissues; the SK1775 rabbit anti-Sdmg1 antibody (Best et al.,2008) was used for all other immunostainings and for Western blots. Goat anti-EEA1 (Santa Cruz Biotechnology), goat anti-Tfrc (Santa Cruz Biotechnology), rat anti-Lamp1 (Developmental Studies Hybridoma Bank), and nonspecific rabbit IgG (Sigma-Aldrich) were also all used at 1 μg/mL for immunostaining. Mouse anti-Vamp2 (Synaptic Systems) was used at a 1:1,000 dilution. Secondary antibodies were used according to the supplier's recommendations (Invitrogen), and DNA was stained with 2 μg/ml DAPI.

Subcellular Fractionation of the Pancreas

Subcellular fractionation of the pancreas was performed essentially as described (Chen et al.,2003). Pancreata from two adult mice were homogenized in 6 ml of ice-cold homogenization buffer (0.3 M sucrose, 25 mM Pipes pH6.8, 2 mM EGTA supplemented with protease inhibitors), then centrifuged at 300 × g for 10 min at 4°C to generate a postnuclear supernatant. The postnuclear supernatant was then centrifuged at 1,000 × g for 10 min at 4°C to generate a crude particulate fraction, which was resuspended in 50% Percoll (Sigma-Aldrich), 50% homogenization buffer, centrifuged at 100,000 × g for 20 min at 4°C, and collected in six equal fractions. A dense white band of secretory granules was visible around halfway down the gradient (fraction F3), and a tan-colored band of mitochondria visible toward the top of the gradient (fraction F5).

For immunostaining of purified secretory granule fractions, purified secretory granules were washed with ice-cold homogenization buffer, allowed to settle on Superfrost plus slides (VWR), fixed with 4% paraformaldehyde in PBS at room temperature, then blocked and incubated with antibodies as described for cryosections.

Western Blotting

Protein extracts from tissues were prepared by homogenizing tissue in PBS containing protease inhibitors at 4°C, then adding an equal volume of 2× Laemmli sample buffer (Harlow and Lane,1988). Tissue homogenates were denatured by heating to 85°C for 10 min rather than boiling to minimize the formation of high molecular weight (>200 kDa) aggregates of the Sdmg1 transmembrane protein. A total of 25 μg protein extract from each tissue was used for Western blotting. For Western blotting of Percoll gradient fractions, equal volumes of each fraction were mixed 1:1 with 2× Laemmli sample buffer and denatured by heating to 85°C for 10 min. Tris/glycine sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed as described (Harlow and Lane,1988), and proteins were transferred to Immobilon-P membrane (Millipore) using a Mini-Protean II Western blotting apparatus (Bio-Rad). The membrane was blocked with 0.25% glycine, 0.1% Tween-20 in PBS, followed by 20% nonfat skimmed milk, then incubated with antibodies diluted in 2% milk, 0.5% Tween-20 in PBS. SK1775 rabbit anti-Sdmg1 antibodies (Best et al.,2008) were used at 0.25 μg/ml for Western blotting, rabbit anti-Vamp8 antibodies (Borner et al.,2006) were used at a 1:200 dilution, and mouse anti-Gapdh antibodies (Abcam) were used at a 1:1,000 dilution. Peroxidase-linked secondary antibodies were detected using enhanced chemiluminescence.

Acknowledgements

We thank the Medical Research Council for financial support, Andrew Peden (Cambridge Institute for Medical Research, Cambridge, UK) for antibodies and discussion, and Mary O'Connell (MRC-HGU) and Andrew Peden (CIMR) for critical reading of the manuscript.

Ancillary