Journal of Cell Science partnership with Dryad

Journal of Cell Science makes data accessibility easy with Dryad

Dictyostelium Hip1r contributes to spore shape and requires epsin for phosphorylation and localization
Shannon Lea Repass, Rebecca J. Brady, Theresa J. O'Halloran

Summary

Clathrin-coated pits assemble on the plasma membrane to select and sequester proteins within coated vesicles for delivery to intracellular compartments. Although a host of clathrin-associated proteins have been identified, much less is known regarding the interactions between clathrin-associated proteins or how individual proteins influence the function of other proteins. In this study, we present evidence of a functional relationship between two clathrin-associated proteins in Dictyostelium, Hip1r and epsin. Hip1r-null cells form fruiting bodies that yield defective spores that lack the organized fibrils typical of wild-type spores. This spore coat defect leads to formation of round, rather than ovoid, spores in Hip1r-null cells that exhibit decreased viability. Like Hip1r-null cells, epsin-null cells also construct fruiting bodies with round spores, but these spores are more environmentally robust. Double-null cells that harbor deletions in both epsin and Hip1r form fruiting bodies, with spores identical in shape and viability to Hip1r single-null cells. In the growing amoeba, Hip1r is phosphorylated and localizes to puncta on the plasma membrane that also contain epsin. Both the phosphorylation state and localization of Hip1r into membrane puncta require epsin. Moreover, expression of the N-terminal ENTH domain of epsin is sufficient to restore both the phosphorylation and the restricted localization of Hip1r within plasma membrane puncta. The results from this study reveal a novel interaction between two clathrin-associated proteins during cellular events in both growing and developing Dictyostelium cells.

Introduction

Regulating transport across cellular membranes is of utmost importance to all living organisms, whether single-celled or part of a multicellular community. Clathrin-mediated endocytosis, the trafficking of clathrin-coated vesicles from the plasma membrane, is a major route of cellular internalization for eukaryotic cells (Kirchhausen, 2000; Brodsky et al., 2001; Lafer, 2002). Clathrin-dependent processes are necessary for receptor internalization, synaptic transmission, maintaining cell polarity and the propagation of immune responses and other fundamental physiological events (Ruscetti et al., 1994; Morgan et al., 2000; Granseth et al., 2006; Fölsch et al., 1999; McCormick et al., 2005; Badour et al., 2007). The importance of clathrin-mediated endocytosis is illustrated by pathologies, such as hypercholesterolemia, leukemia and a number of neurodegenerative diseases that occur when clathrin and/or clathrin-associated proteins are perturbed (Anderson et al., 1977; Floyd and De Camilli, 1998; Browne and Beal, 2006; Yao et al., 1999; Yao et al., 2000).

Coated vesicles from the plasma membrane form when clathrin triskelions assemble into clathrin-coated pits through complex interactions between a multitude of proteins, the plasma membrane and the cytoskeleton (Brodsky et al., 2001; Lafer, 2002; Perrais and Merrifeld, 2005; Smythe and Ayscough, 2006). How these factors recruit and regulate clathrin assembly has been studied extensively. By contrast, how clathrin accessory proteins interact with each other or how specific proteins influence the function of other factors within the endocytic complex remains much less clear. Understanding both how individual proteins function within clathrin-coated pits and how they influence each other is paramount to building a mechanistic model of endocytosis in eukaryotic cells.

One family of clathrin accessory proteins – the Sla2/Hip1 family – has distinct domains demonstrated to interact with both clathrin and the actin cytoskeleton (Chen and Brodsky, 2005; Legendre-Guillemin et al., 2002; Qualmann and Kessels, 2002). At the N-terminus, members of this family contain an AP180 N-terminal homology (ANTH) domain, a domain found in proteins predicted to function in trafficking events through interactions with phosphatidylinositols in the plasma membrane (Legendre-Guillemin et al., 2004; De Camilli et al., 2002; Itoh et al., 2001). Adjacent to the N-terminal region, a predicted coiled-coil region of variable length mediates protein-protein interactions, including binding to clathrin (Wesp et al., 1997; Engqvist-Goldstein et al., 1999; Mishra et al., 2001; Henry et al., 2002). Finally, the C-terminal region of the Sla2/Hip1 orthologs contains a talin-like, F-actin-binding domain, also called the THATCH domain, that has been demonstrated to interact with F-actin (Brett et al., 2006; McCann and Craig, 1997; Yang et al., 1999; Engqvist-Goldstein et al., 1999; Engqvist-Goldstein et al., 2001). The domain structure of Sla2/Hip1 family members suggests that they serve as molecular links between the plasma membrane, cortical actin and endocytic complexes (Henry et al., 2002; Engqvist-Goldstein et al., 1999; Engqvist-Goldstein et al., 2001; Legendre-Guillemin et al., 2002).

Various studies of members of the Sla2/Hip1 family have revealed a role for the proteins in clathrin-mediated endocytosis (Engqvist-Goldstein et al., 1999; Carreno et al., 2004; Yang et al., 1999; Henry et al., 2002; Chen and Brodsky, 2005; Newpher and Lemmon, 2006). The founding member of the Sla2/Hip1 protein family, Sla2p (for: `synthetic lethal with Abp1'), was one of the first proteins identified that, when mutated, led to a deficiency in endocytosis in Saccharomyces cerevisiae (Raths et al., 1993), and interactions with clathrin in yeast were identified subsequently (Henry et al., 2002; Newpher and Lemmon, 2006). Similarly, Mammalian HIP1R (for: `Huntingtin-interacting protein 1 related') colocalizes extensively with clathrin and clathrin-associated proteins at both the plasma membrane and the trans-Golgi network and binds to the clathrin light chain in vitro (Engqvist-Goldstein et al., 1999; Engqvist-Goldstein et al., 2001; Carreno et al., 2004; Chen and Brodsky, 2005). Recent studies have demonstrated the association of Sla2/Hip1 family members with other clathrin accessory proteins during endocytic events, such as cortactin in mammalian cells and Pan1p, the Eps15 homolog, in yeast (Toshima et al., 2006; Le Clainche et al., 2007).

Epsin, another clathrin accessory protein, has a membrane-binding domain at its N-terminus, an ENTH domain, which is similar in structure to the ANTH domain of HIP1R and binds to the phospholipid phosphatidylinositol (4,5)-bisphosphate (Itoh et al., 2001). Along its C-terminal domain, epsin contains multiple binding sites for clathrin and other clathrin-associated proteins (Chen et al., 1998; Kay et al., 1999). Epsin is thought to promote assembly of clathrin and to assist in the membrane invagination necessary for formation of vesicle buds (Ford et al., 2002; Hurley, 2006). In addition, through binding to the cytoplasmic tails of membrane receptors, epsin functions in the selection of cargo to be internalized (Maldonado-Baez and Wendland, 2006; Chen and De Camilli, 2005; Sigismund et al., 2005).

The social soil amoeba Dictyostelium discoideum is an excellent model system for the study of cellular trafficking events in both growing and developing cells as membrane budding and fusion is of crucial importance to the organism (Kessin, 2001; Maniak, 2003). As in other eukaryotes, Dictyostelium contains clathrin-coated vesicles that bud from the plasma membrane, as well as from intracellular organelles (O'Halloran and Anderson, 1992; Damer and O'Halloran, 2000; Wang et al., 2006). Furthermore, Dictyostelium cells have clathrin-associated proteins, such as AP180, adaptor proteins 1 to 4 and epsin (Stavrou and O'Halloran, 2006; Charette et al., 2006; Lefkir et al., 2003) (R.J.B. and T.J.O., unpublished). When nutrients are depleted, Dictyostelium cells develop into multicellular structures, offering a model system to investigate clathrin and clathrin accessory function during growth and development of eukaryotic cells.

In this study, we characterized Dictyostelium Hip1r. We uncovered essential roles for Hip1r and epsin during Dictyostelium development: both functioned in producing ovoid and environmentally robust spores. Our study revealed an indispensable contribution of epsin to both the phosphorylation state and the restricted localization of Hip1r within plasma membrane puncta. Taken together, our results identify a novel interaction between two clathrin-associated proteins important for both growth and developmental function in eukaryotic cells.

Results

Identification of Dictyostelium Hip1r, a new member of the Sla2/Hip1 family of proteins

To identify potential orthologs of the Sla2/Hip1 family of proteins, we searched the Dictyostelium gene database (Chisholm et al., 2006) (www.dictybase.org) and found a single gene (DDB0232318), located on chromosome 4, that we named hipA. The gene encodes a protein of 961 amino acids with a predicted molecular mass of 107 kDa. Analysis of the predicted reading frame revealed a gene product that shares domains with members of the Sla2/Hip1 family of proteins. At its N-terminus (amino acids 1-125), the Dictyostelium Hip1r homolog contains an ANTH domain that shares approximately 32% amino acid identity with other family members. The C-terminal portion of the protein (amino acids 762-959) has a talin-like domain, or THATCH domain, that shares 34% identity with other members of the Sla2/Hip1 family of proteins. Additionally, the central portion of the protein contains regions predicted to form coiled-coil structures. Across its entire length, Hip1r shares ∼24% identity with Sla2p of Saccharomyces cerevisiae and 20% and 19% identity with mouse HIP1R and human HIP1R, respectively. Dictyostelium Hip1r has a central proline-rich region (amino acids 247-294) and, similar to yeast Sla2p, a glutamine-rich (amino acids 312-401) region absent from the mammalian orthologs (Fig. 1).

Hip1r is a phosphorylated protein that associates with membranes

To gain insight into the function of Hip1r, we examined interactions of the protein with membranes and/or cortical structures. To test for a possible association with membranes, lysates of growing, wild-type cells were analyzed by differential centrifugation (Fig. 2A). Immunoblots stained with antibodies against Hip1r demonstrated that most of the cellular pool of Hip1r sedimented with large plasma membrane fragments and/or large membrane-bound organelles [within the low-speed pellet (LSP)], whereas the remaining population of the protein remained soluble [within the low-speed supernatant (LSS) or high-speed supernatant (HSS)].

In addition, we examined the association of Hip1r with Triton X-100 (TX-100)-extracted fractions. When cell lysates were extracted with the non-ionic detergent TX-100, a portion of Hip1r fractionated into the insoluble pellet, suggesting an association with cytoskeletal factors or with TX-100-insoluble lipids (Fig. 2B). Extraction of the TX-100-insoluble fraction with high salt shifted Hip1r into the soluble fraction, suggesting the existence of an ionic interaction with cytoskeletal elements or with TX-100-insoluble lipids (Fig. 2B).

Interestingly, immunoblots probed with antibodies against Hip1r revealed two Hip1r species. To determine whether this finding was due to phosphorylation of the Hip1r protein, cell lysates were incubated with calf intestinal phosphatase (CIP) (Fig. 2C). Okadaic acid, an inhibitor of serine/threonine phosphatases, was included in control samples. After 25 minutes of incubation of the lysate with the phosphatase, the slower migrating band of the doublet disappeared from the test sample but remained in the control sample, indicating that Hip1r is phosphorylated and is most likely modified on a serine or threonine residue. Phosphorylation of Sla2p has been reported in yeast, and HIP1 doublets have been observed in protein gels containing mammalian samples (Yang et al., 1999; Kalchman et al., 1997), but the significance of this modification remains unknown. Identification of phosphorylated Hip1r in Dictyostelium provides an opportunity to study the functional consequence of this modification in a model organism.

Fig. 1.

Domain structures of Hip1r orthologs in different species. Like other Sla2/Hip1 family members, Dictyostelium Hip1r contains an ANTH domain, a central coiled-coil region and THATCH domain as well as proline- and glutamine-rich regions. Over its entire length, Hip1r shares 24% identity with Sla2p, and 20% and 19% identity with mouse and human HIP1R, respectively. Overlap of shaded regions indicates areas of overlap between domains. Dicty, Dictyostelium discoideum; Yeast, Saccharomyces cerevisiae; Mouse, Mus musculus; Pro, proline; Gln, glutamine.

Hip1r-null cells have normal clathrin-mediated phenotypes and organization of F-actin

To examine the cellular role of Dictyostelium Hip1r, we examined the phenotype of Hip1r-null cells generated by using homologous recombination to delete the hipA gene (supplementary material Fig. S1A). Because members of the Sla2/Hip1 family of proteins participate in membrane trafficking events, we examined the Hip1r-null cells for phenotypes previously shown to be defective in cells lacking clathrin (O'Halloran and Anderson, 1992; Ruscetti et al., 1994; Niswonger and O'Halloran, 1997a; Niswonger and O'Halloran, 1997b; Wang et al., 2003). To test for a possible defect in cytokinesis, Hip1r mutant cells were assessed for growth in suspension. Unlike clathrin mutants, Hip1r-null cells were able to complete cytokinesis in suspension (supplementary material Fig. S2A). Likewise, Hip1r mutant cells were not defective in pinocytosis, as measured by the internalization of the fluid-phase marker FITC-dextran (supplementary material Fig. S2B). Once internalized, the marker recycled to the extracellular milieu with kinetics similar to those of wild-type cells (supplementary material Fig. S2C).

To determine whether loss of Hip1r expression affected clathrin localization, Hip1r-null cells expressing clathrin-GFP constructs were examined using fluorescence microscopy. Wild-type and Hip1r-null cells showed similar numbers and intensities of clathrin puncta on their plasma membranes (supplementary material Fig. S3A,B). To test for abnormalities of actin, we stained the actin cortex of wild-type and Hip1r-null cells with fluorescently labeled phalloidin and examined them by fluorescence microscopy (supplementary material Fig. S3C,D). Both cell lines demonstrated a thick band of F-actin at the cell cortex, suggesting that loss of Hip1r expression does not affect the overall organization of cortical actin. Thus, in contrast to Hip1r orthologs from other species, Dictyostelium Hip1r does not play an essential role in clathrin-mediated phenotypes in growing cells.

Hip1r mutant cells develop into fruiting bodies

The unique life cycle of Dictyostelium provided an opportunity to examine cell-specific functions for Hip1r during developmental events. When nutrients are depleted from the surrounding medium, Dictyostelium cells initiate a developmental program whereby ∼100,000 starving cells aggregate to form the fruiting body, a multicellular structure consisting of a thin stalk supporting a sorus that contains spores (Kessin, 2001). Approximately a third of the aggregating cells are destined to form the stalk, whereas the remaining cells become spores, dormant amoebae protected by a thick spore coat. When environmental conditions are suitable, these dormant amoebae emerge to produce viable cells. The Hip1r protein is expressed throughout both growth and developmental stages of the Dictyostelium life cycle, as determined by analysis of microarray data and immunoblotting (www.dictybase.org; data not shown). To test for a role in development, wild-type and Hip1r-null cells were inoculated onto agar plates containing lawns of bacteria. After they depleted the bacteria, Hip1r-null cells aggregated without delay, forming fruiting bodies that were indistinguishable from fruiting bodies formed by wild-type cells (Fig. 3A,B).

Fig. 2.

Hip1r is a phosphorylated protein that associates with membrane and Triton-X100-insoluble fractions. (A) Subcellular cell fractionation of wild-type cells. Most of the cellular pool of Hip1r associates with large membrane structures (LSPs), whereas the rest is soluble (HSS). (B) Triton-X fractionation of Hip1r. Cells were resuspended into buffer containing 0.5% Triton-X100 and fractionated, resulting in a portion of Hip1r sedimenting with the insoluble fraction (P). Extraction of the detergent-insoluble pellet with high-salt buffer shifts Hip1r to the soluble fraction (S). (C) Hip1r is phosphorylated in vivo. Cells were lysed and incubated with calf intestinal phosphatase (CIP) for 25 minutes, resulting in the disappearance of the upper band of the Hip1r doublet. The control lane includes CIP plus okadaic acid, a phosphatase inhibitor. WCL, whole-cell lysate; LSP, low-speed pellet; LSS, low-speed supernatant; HSP, high-speed pellet; HSS, high-speed supernatant; S, supernatant; P, pellet; CIP, calf intestinal phosphatase-treated sample; control, CIP plus okadaic acid.

Fig. 3.

Hip1r-null mutants develop normally into fruiting bodies. Wild-type and Hip1r-null cells were harvested and inoculated onto non-nutrient agar plates to induce the Dictyostelium developmental program. (A,B) Both wild-type (WT) and Hip1r-null cells (Hip1r-) produce fruiting bodies consisting of a round sorus, supported by a thin stalk. Bar, 0.25 mm. (C,D) The stalks from wild-type (WT) and Hip1r-null cells are similar in size and cellular organization. Bar, 10 μm.

To inspect the morphology of the cells within the developmental structure, we examined fruiting bodies on coverslips by differential interference contrast (DIC) microscopy. The stalk cells in the fruiting bodies formed by Hip1r-null mutants were similar in size and appearance to stalk cells of wild-type fruiting bodies (Fig. 3C,D). Both wild-type and mutant stalk structures comprised vacuolated cells with distinct cell borders.

Hip1r mutant cells produce spores with abnormal morphology

In contrast to the normal stalk cells within the fruiting body formed by Hip1r-null cells, the morphology of Hip1r-null spores differed dramatically from that of wild-type spores (Fig. 4A-D). Wild-type cells produced spores that were elliptical. However, Hip1r mutant spores were completely round (Fig. 4A,B).

A major component of the Dictyostelium spore coat is the polysaccharide cellulose, and Dictyostelium mutants with defects in cellulose produce round spores (Zhang et al., 2001). To determine whether a lack of cellulose was associated with the abnormal shape of the Hip1r-null spores, we stained wild-type and Hip1r mutant spores with calcofluor, a cellulose-binding reagent. The spore coat of Hip1r-null cells stained brightly with calcofluor, demonstrating the presence of cellulose within the coat of the mutant spores (Fig. 4C,D).

Hip1r-null cells are not deficient in pre-spore vesicle (PSV) fusion

Because spore coat proteins must be secreted to the extracellular face of the plasma membrane to form a protective coat, we postulated that the secretion of the proteins contained in PSVs might be defective in Hip1r-null strains (Srinivasan, 1999; West, 2003).

Preassembled spore coat proteins are stored in specialized vesicles called PSVs (Srinivasan et al., 1999; Srinivasan et al., 2000a). These vesicles accumulate within the cytoplasm of the pre-spore cell until an unidentified signal is received. Upon receipt of this signal, the PSVs fuse with the plasma membrane and deposit spore coat proteins on the extracellular surface (Srinivasan et al., 2000a). To determine whether fusion of the PSVs with the plasma membrane was defective in Hip1r mutant cells, we examined wild-type and Hip1r-null spore coats for the presence of a marker for delivery of the PSV, SP70, a spore coat protein (Gomer et al., 1986; Fosnaugh et al., 1994). Mature spores from wild-type and Hip1r-null fruiting bodies were harvested and immunostained with an antibody directed against SP70. In both wild-type and Hip1r-null spores, the antibody against SP70 labeled the patches that outlined the periphery of the spore without staining the center of the spores (Fig. 4E-T). Quantification of the intensity of the SP70-containing puncta demonstrated that wild-type and Hip1r mutant spores stained with equivalent intensities (data not shown). This staining pattern suggested that Hip1r mutant cells successfully deliver the complex of spore-coat proteins contained within PSVs to the extracellular surface of spores, eliminating failure of the PSV to form or to secrete its contents as a basis for the round-spore phenotype.

Hip1r-null spores have a disorganized spore coat

We used electron microscopy to examine the fine structure of the spores found in Hip1r-null fruiting bodies. As found by light microscopy, scanning electron microscopy showed oblong wild-type spores and round mutant Hip1r spores. When wild-type and Hip1r-null spores were washed briefly with detergent before preparation for scanning electron microscopy, most wild-type spores remained oblong, whereas the majority of Hip1r-null spores collapsed into flattened sacs, suggesting that the mutant spore coat was fragile (Fig. 5A-D).

To examine the ultra-fine structure of the spore coat, we used transmission electron microscopy. The Dictyostelium spore coat comprises three layers: an outer protein layer, a middle layer of cellulose and an inner protein layer adjacent to the plasma membrane of the dormant amoebae (West, 2003). Images of the wild-type spore coat show an outer and inner protein layer, with a middle region consisting of material that appeared to be crosslinked or arranged in fibrils (Fig. 5E). In contrast to wild-type spore coats, the spore coat formed by mutant Hip1r cells was disorganized and varied in thickness. Moreover, the coat lacked the crosslinked fibrils found in the middle layer of wild-type spore coats (Fig. 5F), indicating that Hip1r plays an essential role in forming a wild-type spore coat.

Fig. 4.

Dictyostelium Hip1r-null cells produce round, calcofluor-positive spores that have SP70 within the spore coats. (A-D) Wild-type and Hip1r-null spores were harvested, stained with the cellulose-staining dye calcofluor and viewed by phase-contrast or fluorescence microscopy. Wild-type (WT) spores are elliptical in morphology, whereas Hip1r-null spores (Hip1r-) are round. Bar, 5 μm. (E-T) Hip1r-null cells secrete outer-layer spore coat proteins. Wild-type (WT) and Hip1r-null spores were harvested, fixed and stained with an antibody against an outer spore coat protein, SP70, and viewed by DIC (E-H and M-P) or fluorescence (I-L and Q-T) microscopy. Bar, 5 μm.

Hip1r/epsin mutant cells yield spores identical to Hip1r mutant spores

The Hip1r-null phenotype was similar to that of epsin-null cells, which also develop into fruiting bodies that contain round spores. Indeed, epsin-null cells nearly phenocopy the Hip1r-null cells. As with Hip1r-null cells, epsin-null mutants are normal for actin, clathrin and other growth phenotypes but produce fruiting bodies that contain spores that are rounder than wild-type spores (R.J.B. and T.J.O., unpublished). Of all of the clathrin or clathrin accessory protein mutants in Dictyostelium, only epsin-null cells share the round-spore phenotype we observed in Hip1r-null cells (Stavrou and O'Halloran, 2006; Charette et al., 2006; Lefkir et al., 2003; Wang et al., 2003; Niswonger and O'Halloran, 1997a) (R.J.B. and T.J.O., unpublished). This raised the possibility that Hip1r and epsin function together to construct robust spores. To test for functional interactions between these two proteins, we constructed a double-null mutant that carried deletions in the genes encoding both Hip1r and epsin. As for single Hip1r and epsin-null mutants, the Hip1r epsin double-mutant cells were viable, grew normally and had normal cellular phenotypes during growth (data not shown).

Direct comparison of the spore cells by DIC microscopy confirmed that Hip1r and epsin mutants produced spores that were not as elongated as wild-type spores (Fig. 6). However, whereas Hip1r-null spores were all completely round, epsin-null spores displayed a variety of shapes, including completely round as well as more elliptical (Fig. 6A,B). All of the spores from the Hip1r epsin double-mutant cells were completely round, identical in morphology to spores made by single Hip1r-null cells (Fig. 6C,D). Thus, the double mutants were not enhanced in phenotypic defects and displayed a round spore defect identical with that of Hip1r single-mutant cells. The absence of enhanced defects in the double mutant suggested that Hip1r and epsin do not share redundant functions in either growth or in development.

To test the viability of the mutant spores, we compared germination rates between wild-type spores, single Hip1r-null spores, single epsin-null spores and Hip1r epsin double-mutant spores. Spores harvested from sori of the four cell lines were incubated with bacteria on agar plates. The plates were monitored to determine whether viable amoebae emerged from the spores to grow and clear plaques on the bacterial lawn. Spores harvested from the sori of both wild-type and mutant fruiting bodies demonstrated comparable survival rates: 70-80% of spores yielded amoebae capable of growing on bacteria (Fig. 6E).

Fig. 5.

Hip1r-null spore coats lack organized cellulose. Wild-type and Hip1r-null spores were collected, fixed and stained (see Materials and Methods). (A,B) Wild-type (WT) and Hip1r-null (Hip1r-) spores were collected and washed without detergent. Wild-type spores are elliptical, whereas null spores are round (C,D), and washing wild-type spores and Hip1r-null spores with 0.1% NP40 before fixation results in the collapse of the mutant spores. Bars, 5 μm. (E,F) TEM of spores shows a middle layer of protein arranged in fibrils or crosslinked in wild-type (arrows) spores that is absent for the spore coat of mutant spores. Bar, 500 nm.

In the natural world, the spore coat must protect the dormant amoeba from a variety of environmental onslaughts such as chemical and temperature fluctuations. Wild-type spores are encased in a tough, trilaminar spore coat that can withstand harsh conditions (Kessin, 2001; West, 2003). To test the resilience of the various null spore coats, we tested spore viability of wild-type and mutant spores after treatments with detergent and heat (Fig. 6E). After washing spores with the non-ionic detergent NP-40, we plated them on a lawn of bacteria and measured their ability to germinate and grow. Although ∼50% of wild-type and epsin-null spores germinated, the detergent treatment reduced the viability of single Hip1r-null spores and Hip1r epsin double-mutant spores to less than 20%. Heat treatment also revealed vulnerability in the Hip1r and Hip1r epsin double-mutant spores. After a brief heat treatment, ∼40% of wild-type spores and single epsin-null spores were viable, but only ∼25% of the Hip1r-null and Hip1r epsin double-mutant spores yielded viable amoebae. Thus, Hip1r epsin double-mutant spores displayed deficits identical to those of Hip1r-null spores. The similarity in defects of Hip1r-null and Hip1r epsin double-null cells suggested that epsin functions upstream of Hip1r in a specialized pathway required for spore cell development.

Hip1r localizes within plasma membrane puncta that contain epsin

To examine the relative intracellular locations of Hip1r and epsin, we used immunofluorescence microscopy. Staining permeabilized spore cells during development proved difficult as the spores were fragile and those that remained intact trapped antibody nonspecifically. Immunostaining growing cells was more informative and revealed a punctate distribution for Hip1r. Growing cells displayed some lightly stained puncta in the cytoplasm, but the brightest Hip1r fluorescence was associated with discrete puncta on the plasma membrane (Fig. 7A,B). Unlike its mammalian ortholog, Hip1r did not exhibit a juxtanuclear staining pattern. Focusing on the surface of cells expressing GFP-labeled clathrin and immunostained with antibodies against Hip1r revealed that Hip1r colocalized with clathrin structures, although not all Hip1r puncta were associated with clathrin (Fig. 7C-E).

The Dictyostelium epsin ortholog also localizes within clathrin-associated puncta on the plasma membrane (R.J.B. and T.J.O., unpublished). To examine the relative localization of Hip1r and epsin, cells expressing GFP-labeled epsin were fixed, immunostained with affinity-purified antibodies against Hip1r and examined by fluorescence microscopy. Hip1r and epsin colocalized extensively within puncta along the plasma membrane (Fig. 7F-H).

Fig. 6.

Hip1r epsin double-mutant cells display developmental phenotypes similar to those of Hip1r single-mutant cells. Wild-type (WT), epsin-null (epsin-), Hip1r-null (Hip1r-) and Hip1r epsin (Hip1r-/epsin-) double-mutant cells were harvested and inoculated onto non-nutrient agar plates to induce spore formation. Spores were examined by DIC microscopy. Wild-type spores (A) are ovoid, whereas epsin-null spores (B) are less oblong, with some round spores. Hip1r-null spores (C) and Hip1r epsin double-null spores (D) are completely round. Bar, 5 μm. (E) When treated with detergent or heat, Hip1r-null and Hip1r epsin double-mutant spores show decreased viability. Wild-type, epsin-null, Hip1r-null and Hip1r epsin double-mutant spores were harvested and left untreated, subjected to washes with the detergent NP40 or heat treated for 10 minutes. Spores were then plated on a bacterial lawn, where they germinated and grew. Cleared plaques were counted as a measure of spore germination. Error bars represent the s.e. (n=1000 spores, seven experiments).

Hip1r exhibits a functional interaction with epsin

The similar developmental phenotypes and the extensive colocalization of Hip1r and epsin within cell-surface puncta prompted us to explore the relationship between the two proteins in more detail. To test whether Hip1r influenced the localization of epsin within puncta on the plasma membrane, we examined the distribution of epsin in Hip1r-null cells. Inspection by fluorescence microscopy of Hip1r-null cells transformed with vector encoding epsin-GFP showed a distribution of epsin that was similar to that observed in wild-type cells. In both strains, epsin-labeled puncta were almost exclusively at the plasma membrane (Fig. 8A,B). Thus epsin localizes to plasma membrane puncta independently of Hip1r.

To test the reciprocal possibility, that epsin influenced the localization of Hip1r, we examined the distribution of Hip1r in epsin mutant cells. Epsin-null cells immunostained with antiserum against Hip1r revealed a mislocalization of the Hip1r protein (Fig. 8C,D). In contrast to wild-type cells, Hip1r was not concentrated at the surface of epsin-null cells. Rather, the epsin-null cells displayed an increased number of Hip1r-labeled punctate structures within the cytoplasm, with few Hip1r-labeled puncta at the cell surface. The cytoplasmic Hip1r puncta in epsin-null cells did not colocalize with clathrin (data not shown). The absence of Hip1r membrane puncta in epsin-null cells suggested that epsin could play a role in either recruiting or stabilizing Hip1r within puncta on the plasma membrane. To test directly for a physical interaction between epsin and Hip1r, we used antibodies against epsin to probe western blots of fractions immunoprecipitated by antibodies against Hip1r, but we failed to detect epsin in these fractions (data not shown). Whereas the phenotypes demonstrated that epsin interacts functionally with Hip1r, the immunoprecipitation suggested that the interaction between the two proteins is indirect or of low affinity.

The domain structure of the epsin protein includes an N-terminal membrane-binding ENTH domain and a C-terminal domain that contains multiple motifs for binding to clathrin and clathrin accessory proteins. To determine whether expression of these domains could restore the proper localization of Hip1r, we examined epsin-null cells that expressed the ENTH domain. When expressed in epsin-null cells, an N-terminal construct that included the ENTH domain was sufficient to restore the restricted localization of Hip1r into peripheral puncta in approximately half (47%) of the cells examined (Fig. 8E).

The possibility that epsin influences the phosphorylation the Hip1r protein was also examined. Lysates of wild-type, epsin-null cells and epsin-null cells expressing epsin-GFP were analyzed in immunoblots stained with antiserum against Hip1r. Although the antibodies against Hip1r recognized two bands in the wild-type and epsin-null cells expressing epsin-GFP, only a single band was stained in the epsin-null cells (Fig. 8F). These results suggested that epsin is essential for the localization of Hip1r to the plasma membrane as well as its phosphorylation state. Moreover, expression of the N-terminal ENTH domain of epsin-null cells also restored phosphorylation of Hip1r (Fig. 8F). By contrast, epsin-null cells that expressed the C-terminal domain contained only the unphosphorylated species of Hip1r (Fig. 8F). Taken together, these results show that epsin, and particularly the ENTH domain, is required for both the restricted localization of Hip1r as well as its ability to be phosphorylated in living cells.

Discussion

Understanding the complex interactions that occur between proteins during membrane trafficking events is key to discerning how eukaryotic cells grow, adapt and maintain homeostasis. Although the interactions between clathrin and various accessory proteins have been studied widely, much less is known about how individual accessory proteins collaborate to promote successful cellular events at the plasma membrane. In this study, we identified a novel interaction between two clathrin-associated proteins, Hip1r and epsin, that is crucial for the localization of Hip1r during growth, as well as for the construction of elliptical, environmentally robust spores during development.

Fig. 7.

Hip1r localizes to the plasma membrane and cytoplasm and displays an interaction with epsin. Growing cells were fixed and stained with antibody against Hip1r. (A) Cells were viewed by DIC microscopy. (B) Cells stained with antibody against Hip1r show bright Hip1r puncta at the cell cortex when viewed by fluorescence microscopy. (C-E) Focusing on the surface of cells expressing a clathrin-GFP construct (clathrin:GFP) shows partial colocalization of clathrin (green) with Hip1r (anti-Hip1r, red). Examples are indicated by arrowheads (E). (F-H) Cells expressing an epsin-GFP fusion protein (epsin:GFP, green) and stained with antibody against Hip1r (anti-Hip1r, red) exhibit extensive colocalization at the plasma membrane (arrowheads). Occasionally, puncta contain epsin but not Hip1r (arrow). Bar, 10 μm.

Roles for Hip1r in Dictyostelium spore formation

The generation of environmentally resistant spores, the end result of Dictyostelium development, is important for the organism to survive harsh conditions in nature. The coat that encloses a dormant amoeba within the spore comprises cellulose and glycoproteins (West and Erdos, 1990). Most of these glycoproteins are delivered to the exterior of pre-spore cells when PSVs, containing pre-assembled protein complexes, fuse with the plasma membrane (Srinivasan et al., 1999; Srinivasan et al., 2000a). This regulated secretion is a prerequisite for both the synthesis and the organization of cellulose in the spore coat (Srinivasan et al., 2000b; Metcalf et al., 2003). After the spore coat proteins are secreted to the extracellular side, the enzyme responsible for cellulose synthesis, cellulose synthase, is positioned in the plasma membrane and synthesizes cellulose directly into the outer layer of the spore coat (Metcalf et al., 2007; West, 2003; Metcalf et al., 2003; Srinivasan et al., 2000b).

Here, we show that Hip1r is required for the formation of a normal Dictyostelium spore coat. Although the Hip1r mutant spore coat stains with calcofluor, demonstrating the presence of cellulose, electron micrographs of the spore coat show an absence of crosslinked fibrils, suggesting that the cellulose might be incorrectly organized to form a structurally sound protective coat. Hip1r mutant spores stain with SP70, a glycoprotein housed within the pre-spore vesicle, so Hip1r is not required for the overall secretion of PSV contents. The spore coat contains a minimum of ten different glycoproteins that are secreted through at least two different pathways (Metcalf et al., 2007; West, 2003). As a scaffolding protein with binding sites for actin, the plasma membrane and a coiled-coil domain, it is conceivable that Hip1r could be required for the proper positioning of any of the many factors involved in the organization of the spore coat. Alternatively, although nothing is known about a role for endocytosis in pre-spore cells, specialized endocytosis is important for organization of the spore cell wall in yeast (Iwamoto et al., 2005). Similarly, it is possible that a specialized endocytic event involving Hip1r could prepare Dictyostelium pre-spore cells for the coordinated placement of enzymes crucial for organization of the spore coat.

Hip1r interacts with epsin

Using immunofluorescence and biochemical assays, we found that Hip1r is a phosphorylated protein that is distributed into puncta associated with the plasma membrane. A genetic interaction between the yeast ortholog of HIP1R, Sla2p, and epsin has been reported previously, but how these two proteins influence each other is unknown (Baggett et al., 2003). We identified an essential contribution of epsin to both the localization and phosphorylation state of Dictyostelium Hip1r. In wild-type cells, Hip1r displayed an extensive colocalization with epsin on the plasma membrane. In epsin-null cells, Hip1r was displaced from the plasma membrane. Furthermore, the phosphorylated form of Hip1r was absent in epsin-null cells. A functional connection between Dictyostelium Hip1r and epsin is also supported by the almost identical phenotypes exhibited by Dictyostelium epsin- and Hip1r-null mutants: during growth, the epsin- and Hip1r-null cells have a wild-type phenotype. However, during development, both mutants display a limited developmental phenotype that is unique for membrane trafficking mutants and contain round, rather than ovoid, spores within the sorus of the fruiting body. Although Hip1r epsin double-null spores were both round, epsin-null cells had a range of spore cell shapes and were more resilient to environmental stress. The similar but less severe epsin-null spore phenotype and the lack of Hip1r association with plasma membrane puncta in epsin-null cells suggests that the epsin deficits result from the absence of Hip1r at the cell cortex. If this hypothesis is true, deleting both proteins should result in phenotypes analogous to those observed for Hip1r mutant cells. Consistent with this idea, the Hip1r epsin double-null cells were similar to wild-type cells during growth and formed normal fruiting bodies that gave rise to completely round, rather than ovoid, spores. Moreover, when tested under adverse conditions, the spores harvested from the Hip1r epsin double-mutant fruiting bodies displayed a survival rate that was identical to that of Hip1r single-mutant spores. Our results suggest that Hip1r must be associated with the plasma membrane for the production of viable spores with wild-type morphology and that the developmental defect observed in epsin-null cells is due to the mislocalization of Hip1r and not solely due to the loss of epsin.

Fig. 8.

The N-terminal ENTH domain of epsin is necessary and sufficient for phosphorylation of Hip1r and membrane localization. (A,B) In wild-type (A, WT) and Hip1r-null cells (B, Hip1r-), epsin-GFP (epsin:GFP) localizes within puncta on the plasma membrane. (C,D) In contrast to wild-type cells (C), Hip1r puncta (anti-Hip1r) are not restricted to the cell periphery in epsin-null cells (D, epsin-). (E) Expression of the ENTH domain in epsin-null cells [epsin-(ENTH:GFP)] restores the peripheral localization of Hip1r. Bar, 10 μm. (F) Epsin influences the phosphorylation state of Hip1r. Immunoblots of cell lysates of WT cells, epsin-null cells expressing epsin-GFP (epsin-, epsin:GFP) and the ENTH domain of epsin (epsin-, ENTH:GFP) probed with antibody against Hip1r exhibit double bands, indicative of phosphorylation of Hip1r. In epsin-null cells (epsin-) and cells expressing only the epsin C-terminus (epsin-, C-term), Hip1r is not phosphorylated.

Based on our results, we propose that epsin acts to recruit and/or position Hip1r at the plasma membrane of growing and developing Dictyostelium cells. This functional requirement of Hip1r for epsin does not necessarily imply a physical interaction between the two proteins, and indeed we did not detect an interaction between epsin and Hip1r in immunoprecipitation experiments (data not shown). The interaction between epsin and Hip1r could be mediated through an epsin-dependent kinase required to phosphorylate and stabilize Hip1r at the membrane of pre-spore cells, where it could function to promote determinants for spore morphology. Alternatively, an epsin-dependent positioning of Hip1r at the cell cortex could allow Hip1r to be phosphorylated, perhaps by membrane-associated kinases, and thus influence its function. We favor the possibility that phosphorylation of Hip1r occurs at the plasma membrane as the N-terminal ENTH domain of epsin, which is able to rescue Hip1r phosphorylation and localization, is distributed uniformly and exclusively along the plasma membrane (R.J.B. and T.J.O., unpublished). We found that the N-terminal ENTH domain of epsin could rescue phosphorylation and Hip1r localization, suggesting that determinants within this domain are important for function. Recently, the ENTH domain alone, expressed without the C-terminal domain of epsin, was shown to supply a clathrin-independent function and interact with the actin-cdc42 pathway in yeast (Aguilar et al., 2006). Conceivably, in Dictyostelium cells, regulation of the actin cytoskeleton by epsin could influence the positioning of Hip1r at the cortex or regulate kinases that phosphorylate Hip1r.

In whole-cell lysates and in membrane fractionation, the amount of phosphorylated Hip1r was always less than the unphosphorylated species (Fig. 2A,C). However, we consistently observed a different ratio in TX-100 supernatants, where the amount of phosphorylated Hip1r was equivalent or even greater than the unphosphorylated Hip1r (Fig. 2B). This could mean that phosphorylated Hip1r preferentially associates with the TX-100-soluble pool or with TX-100-soluble lipids. Alternatively, the kinase for Hip1r could be enriched in the TX-100 supernatant, or the phosphatase for Hip1r could be diminished. Phosphorylation was not a requisite for TX-100 partitioning as the total amount of Hip1r in TX-100 supernatants and pellets was unchanged in epsin-null cells, despite the absence of the phosphorylated species (S.L.R., unpublished).

Our results reveal a novel interaction between Hip1r and epsin, a functional relationship that cannot be fulfilled by other genes in Dictyostelium Hip1r- or in epsin-null mutants. As clathrin and associated proteins have been shown to be important for cell-fate specifications in other organisms, the novel relationship between Hip1r and epsin described in this study might be representative of interactions occurring in more-complex eukaryotes during normal cellular operations and specialized processes during development. Taken together, these results advance the understanding of how cooperation between distinct clathrin accessory proteins contributes to complex functioning in eukaryotic cells.

Materials and Methods

Strain and cell culture

Dictyostelium discoideum wild-type strain Ax2 and Hip1r-null and epsin-null mutant strains were cultured in HL-5 nutrient media on Petri dishes at 20°C. Null cells were supplemented with 5 μg/ml blasticidin. Strains expressing GFP constructs were also supplemented with 10 μg/ml G418. Double-mutant strains were grown in FM media (Formedium™; Norwich, England) or HL-5 and supplemented with 5 μg/ml blasticidin.

Cloning of hipA and polyclonal antibody generation

The amino acid sequence of mouse HIP1R and yeast Sla2p (accession numbers W82687 and Z22811, respectively) were used to search the Dictyostelium genome database (www.dictybase.org) using the BLAST search program (tBALSTn). Predicted protein sequences were analyzed using the DNAStar Megalign program (DNAStar, Madison, WI).

Polyclonal antisera were raised against the ∼20-kDa C-terminal portion of Hip1r. This protein was expressed from plasmid pGEX2T-THATCH. The plasmid was constructed by amplifying this region from Dictyostelium cDNA by PCR using primers 5′-CCCGGGGGATGCTGCAAACTCATTATTGGCC-3′ and 5′-GAATTCTTGATTTTCGTCATATTGTTTCTTTC-3′. The product was cloned into the XmaI and EcoRI sites of the plasmid pGEX2T that situates the 3′ end of the hipA gene downstream of the gene encoding glutathione-S-transferase (GST), thus generating an expression plasmid for a GST C-terminal Hip1r fusion protein. pGEX2T-THATCH was transformed in Escherichia coli strain BL21 for large-scale protein purification. Purification of the fusion protein was accomplished as described previously (Vithalani et al., 1998). The purified GST-Hip1r protein was used to generate polyclonal antibodies against Hip1r in rabbits (Cocalico Biologicals, Reamstown, PA; supplementary material Fig. S1A). For immunostaining experiments, the polyclonal antibodies were affinity purified according to the manufacturer's directions using an affinity column of GST C-terminal Hip1r fusion protein made with the GST Orientation kit (Pierce, Rockford, IL; supplementary material Fig. S1B-G).

Immunofluorescence microscopy

Growing cells were harvested, adjusted to 2×106 cells/ml, plated on coverslips and allowed to adhere at 20°C for 15 minutes. After a brief wash in PDF buffer (20 mM KCl, 11 mM K2HPO4, 13.2 mM KH2PO4, 1 mM CaCl2 and 2.5 mM MgSO4, pH 6.4), cells were fixed by submerging the coverslip in 2% formaldehyde in PDF buffer at room temperature for 15 minutes, followed by permeabilization with 100% methanol at –20°C for 5 minutes. Cells were washed briefly in PDF buffer and processed for immunostaining.

For immunostaining, fixed cells on coverslips were blocked with 3% BSA in PBS solution for 15 minutes at 37°C and then incubated with affinity-purified antibodies against Hip1r (1 μg/ml) or affinity-purified antibodies against clathrin (4 μg/ml) for 60 minutes at 37°C. Coverslips were washed in PBS, followed by incubation with secondary antibody (30 μg/ml), Texas-Red-X-conjugated goat anti-rabbit IgG (Molecular Probes, Eugene, OR) for 60 minutes at 37°C. After rinsing in PBS, coverslips were mounted with mounting media onto slides and allowed to dry. As a control to confirm the specificity of the antisera, Hip1r-null cells were stained with affinity-purified antibodies against Hip1r and no fluorescence was observed (supplementary material Fig. S1B-G).

For F-actin staining, cells were fixed by submerging the coverslips in 2% formaldehyde in PDF buffer at room temperature for 15 minutes, followed by permeabilization with 0.025% TX-100 in PDF buffer for 15 minutes, followed by a brief wash in PDF buffer. Cellular F-actin was stained by incubating the fixed cells with Texas-Red-conjugated phalloidin (Molecular Probes, Eugene, OR) at a concentration suggested by the manufacturer for 25 minutes at room temperature. After rinsing in PBS, coverslips were mounted with mounting media onto slides and allowed to dry.

Images were taken on an inverted NIKON microscope TE2000 (NIKON Instruments, Dallas, TX) with a 100× 1.4 NA PlanFluor objective and Quantix camera (Roper Scientific, AZ) controlled by Metamorph software (Universal Image Corp., PA) and processed with Adobe Photoshop software (Adobe, San Jose, CA).

SDS-PAGE analysis

For gel samples, lysates of growing cells were harvested at a concentration 1×106 cells/ml. After boiling with SDS-containing sample buffer, lysates were loaded onto 7.5% or 10% SDS polyacrylamide gels, electrophoresed and transferred to nylon membranes. After blocking with 3% nonfat milk in Tris-buffered saline (TBS), membranes were incubated with polyclonal antibodies (1:1000), washed with TBS and then incubated with goat anti-rabbit secondary antibody conjugated to horseradish peroxidase (1:5000; Southern Biotech, Birmingham, AL). Membranes were developed using the Supersignal West Pico Chemiluminescent Substrate Kit (Pierce Biotechnology, Rockford, IL).

Phosphatase treatment, differential centrifugation and Triton-X-100 fractionation

To assess protein phosphorylation, cells were collected and resuspended in 1× NEB 3 buffer (New England Biolabs, Beverly, MA) containing protease inhibitors diluted to the manufacturer's instructions (Fungal Protease Inhibitor cocktail, Sigma-Aldrich, St Louis, MO). For lysis, cells were pelleted at 1500 g for 3 minutes at 4°C and resuspended at a concentration of 2×106 cells/ml in 1× NEB 3 buffer containing protease inhibitors plus 0.5% TX-100. The resuspension was divided into test and control samples. CIP, calf intestinal alkaline phosphatase (New England Biolabs, Beverly, MA) was added to both samples and okadaic acid (10 ng/ml) (Sigma-Aldrich, St Louis, MO) was added to control samples to inhibit CIP activity. Samples were incubated at 37°C for 25 minutes and analyzed by SDS-PAGE and western blotting.

For differential centrifugation, approximately 8×108 cells were collected, washed with isolation buffer [10 mM MES (pH 6.5), 50 mM potassium acetate (pH 6.5), 0.5 mM MgCl2, 1 mM EGTA, 1 mM DTT and 0.02% NaN3] and resuspended to 4×107 cells/ml in isolation buffer with protease inhibitors diluted to the manufacturer's instructions (Fungal Protease Inhibitor cocktail, Sigma-Aldrich, St Louis, MO). Cells were lysed by passing them through two pieces of Osmonics (GE Osmonics, Trevose, PA) polycarbonate membrane (pore size: 5 μm) in a Gelman Luer-Lock-style filter (Gelman Sciences, Ann Arbor, MI). The cell lysates were centrifuged at 3000 g for 10 minutes at 4°C. The resulting postnuclear supernatants were ultracentrifuged at 100,000 g for 60 minutes at 4°C to generate a high-speed supernatant and a high-speed membrane pellet. The resulting fractions were analyzed by SDS-PAGE and western blotting.

To assess TX-100 fractionation, cells were harvested by centrifugation at 3000 g for 5 minutes at 4°C. Cells were resuspended in MES buffer [20 mM MES (pH 6.5), 2 mM EGTA, 1 mM MgCl2, 1 mM DTT] containing protease inhibitors diluted to the manufacturer's instructions (Fungal Protease Inhibitor cocktail, Sigma-Aldrich, St Louis, MO), washed with isolation buffer and resuspended in MES buffer and protease inhibitors plus 0.5% TX-100 for lysis. Lysates were spun at 16,060 g at 4°C for 30 seconds to separate the TX-100-soluble and -insoluble fractions. After collecting the soluble fraction, the insoluble fraction was resuspended in MES buffer in the same volume. Fractions were analyzed by SDS-PAGE and western blotting.

Disruption of the hipA and epnA gene by homologous recombination

PCR was used to amplify genomic sequences flanking and within the coding region of the Dictyostelium hipA gene. The 5′ upstream region of the hipA gene was amplified using the primers 5′-GATGACAGAGTTTGAAGCAATTGTCC-3′ and 5′-GCAGCTTGTTGTTGTTGTAATTGTAAATTTGG-3′. The 3′ downstream region of the gene was amplified using the primers 5′-CGTGTCGAAAAGGGTAAAACAAGTGATGG-3′ and 5′-CGGTTTAAAAAAGTTACCATCAAGGC-3′. Each PCR fragment was cloned initially into the vector pCR2.1 (Invitrogen) and then subcloned into the plasmid pSP72-BSR, which carries a gene conferring blasticidin resistance (Wang et al., 2002). The resulting plasmid, pSP72BSR-HipKO, was linearized with SacI and XhoI and introduced by electroporation into Ax2 cells.

PCR was used to amplify genomic sequences flanking the coding region of the Dictyostelium epnA gene. The 5′ upstream region of the epnA gene was amplified using the primers 5′-TTAAAAAAGGTAAAGATGCAGTATTG-3′ and 5′-TTGGAAATTTGGTGTTGCTGGTG-3′, whereas the 3′ downstream region of the gene was amplified using the primers 5′-AATCAAAGTGGTGCGAATAGAAATAC-3′ and 5′-AATGATGATAGTAAAACTGATGGTAGAAG-3′. The PCR fragments were cloned into the vector pCR2.1 (Invitrogen) and then subcloned into the plasmid pSP72-BSR. The resulting plasmid, pSP72-BSR-EpsinKO, was linearized with HindIII and EcoRI and introduced by electroporation into the Ax2 wild-type strain.

For electroporation, 5×106 cells in 100 μl of buffer H-50 (20 mM HEPES, 50 mM KCl, 10 mM NaCl, 1 mM MgSO4, 5 mM NaHCO3, 1 mM NaH2PO4) was mixed with 5 μg of linearized DNA and electroporated using a Bio-Rad Gene Pulser (75 kv, 25 μF) (Bio-Rad, Hercules, CA). Each transformation reaction was diluted into HL-5 with 5 μg/ml blasticidin and plated into six 96-well plates. The resulting clonal lines were selected and screened by western blotting with antibodies against Hip1r or epsin. Replacement of the hipA or epnA gene by integration of the blasticidin cassette was confirmed by PCR analysis.

To generate a strain of Dictyostelium that harbored deletions in both the hipA gene and the epnA gene, PCR was used to amplify genomic sequences flanking and within the coding region of the hipA gene. The 5′ upstream region of the hipA gene was amplified using the primers 5′-CTCGAGTGCTGAAATTTTTACATCCAACC-3′ and 5′-AAGCTTGAAGGGTGGGTGGGTTTACG-3′. The 3′ downstream region of the gene was amplified using the primers 5′-GAGCTCGCCAAATCCGTATCCAATGC-3′ and 5′-GAATTCCCATTTACTTTCCGGAGAAATCTCG-3′. Each PCR fragment was cloned initially into the vector pCR2.1 (Invitrogen) and then subcloned into the plasmid pSP72-pyr, which carries a gene for pyrimidine biosynthesis (Wang et al., 2002). The pSP72-pyr plasmid is derived from pSP72-BSR and the pRH130 vector. The resulting plasmid, pSP72-pyr-HipKO, was linearized with XhoI and EcoRI and introduced by electroporation into DH1 cells. Each transformation reaction was diluted into FM media (Formedium, Norwich, England) and plated into six 96-well plates. The resulting clonal lines were selected and screened by western blotting with antibodies against Hip1r. The Hip1r-null cell lines derived from the DH1 wild-type strain were transformed with the plasmid pSP72-BSR-EpsinKO by electroporation. Each transformation reaction was diluted into HL-5 media with 5 μg/ml blasticidin or FM media with 5 μg/ml blasticidin and plated into six 96-well plates. The resulting clonal lines were selected and screened by western blotting with antibodies against epsin or Hip1r.

For electroporation, 5×106 cells in 100 μl of buffer H-50 (20 mM HEPES, 50 mM KCl, 10 mM NaCl, 1 mM MgSO4, 5 mM NaHCO3, 1 mM NaH2PO4) was mixed with 5 μg of linearized DNA and electroporated using a Bio-Rad Gene Pulser (75 kv, 25 μF; Bio-Rad, Hercules, CA).

Generation of green fluorescent protein (GFP) constructs

To generate a full-length epsin and GFP chimera, Dictyostelium cDNA was amplified by PCR with the primers 5′-TGGAGACTATGATTAAAAGTTATATTAAAAAAGGTAAAGATGCAGTATTGAATACACCAGAAATTGAAAGAAAGGTTAG-3′ and 5′-GCAGATCCCATGCTATTAGTATTTCTATTCGC-3′. The coding region of epsin was subcloned into pCR2.1 (Invitrogen) and subsequently ligated into the pTxGFP vector (a kind gift of Tom Egelhoff) using the KpnI and EcoRV restriction sites. To generate the epsin ENTH domain fused to GFP, the DNA encoding amino acids 1-333 was amplified by PCR using the primers 5′-TGGAGACTATGATTAAAAGTTATATTAAAAAAGGTAAAGATGCAGTATTGAATACACCAGAAATTGAAAGAAAGGTTAG-3′ and 5′-GGTCGACTTCTTCCGCCAG-3′, and pCR2.1 containing full-length epsin as a template. The epsin (1-333) PCR product was subcloned into the pCR2.1 vector. pCR2.1-Epsin (1-333) was then cut with EcoRI, blunted and cloned into the EcoRV site pTxGFP, making pTX-pCR2.1-Epsin (1-333):GFP.

Measurement of pinocytosis and secretion

Fluid-phase endocytosis was measured by uptake of FITC-dextran (2 mg/ml), as described previously (Ruscetti et al., 1994). Log-phase cells were harvested and 1×106 to 2×106 cells were grown overnight in suspension cultures until the cells reached a titer of 3×106 to 4×106 cells/ml. For pinocytosis assays, FITC-dextran was added to a final concentration of 2 mg/ml and incubated at 20°C. For secretion studies, cells were incubated with FITC-dextran for 2 hours at 20°C before taking samples from the cultures. Fluorescence of internalized FITC-dextran was quantified using a VersaFluor Fluorometer (Bio-Rad, Hercules, CA) equipped with the filter set for FITC, a 485-495 nm excitation filter and a 515-525 nm emission filter.

Development into fruiting bodies

For development of fruiting bodies, cells were harvested and washed twice with starvation buffer (20 mM MES, 0.2 mM CaCl2, 2 mM MgSO4) and plated onto either non-nutrient agar plates (1% Noble agar; Difco Laboratories) at 20°C or seeded on SM-5 plates containing a lawn of bacteria (Escherichia coli B/R) and grown at 20°C. Fruiting bodies were harvested and images were taken with either a Zeiss microscope Semi SR with 1.2× or 2.0× objectives controlled by NIH image software or with an inverted NIKON microscope TE2000 (NIKON Instruments, Dallas, TX) with 100× 1.4 NA PlanFluor objective and Quantix camera (Roper Scientific, AZ) controlled by Metamorph software (Universal Image Corp., PA). Images were adjusted using Adobe Photoshop software.

Spore viability assays

The viability of wild-type, epsin-null, Hip1r-null and Hip1r epsin double-mutant spores were assayed by harvesting approximately 2×108 cells and plating onto non-nutrient agar plates. Cells were grown at 20°C for 4 days to allow development into fruiting bodies. Spores were harvested by inverting plates and tapping gently to collect spores on the lid of the Petri dish. The collected spores were suspended into spore buffer (40 mM KH2PO4, 20 mM KCl, 2.5 mM MgCl2), washed twice by centrifugation at 12,500 g for 2 minutes at room temperature and counted with a hemocytometer. Spores were either heated to 45°C for 10 minutes or incubated with 0.5% or 0.1% NP40 detergent (Sigma-Aldrich, St Louis, MO) in spore buffer for 5 minutes or incubated with spore buffer alone for 5 minutes. Spores were washed and plated in triplicate onto SM-5 agar plates containing a lawn of bacteria (E. coli B/R) and grown for 14 days at 20°C. The viability of spores was assessed by counting the number of clear plaques formed on the bacterial lawns.

Spore fixation and staining

Approximately 2×108 growing wild-type, epsin-null, Hip1r-null and Hip1r epsin double-mutant spores were harvested, washed and plated on non-nutrient agar plates. Cells were grown at 20°C for 4 days to allow for spore formation. Spores were collected by inverting plates and tapping gently, resuspended into and then washed twice with spore buffer and placed on coverslips treated with poly-L-lysine. Spores were either viewed with phase-contrast or DIC microscopy or processed for staining. The presence of cellulose in harvested spores was detected by staining with 1 ng/ml Calcofluor (Sigma-Aldrich, St. Louis, MO).

For immunostaining experiments, spores were fixed by the addition of 2% paraformaldehyde in PDF buffer to the coverslips with harvested spores. Coverslips were incubated for 15 minutes at room temperature. Spores were washed briefly in PDF buffer and blocked with 3% BSA in PBS solution for 15 minutes at 37°C. To detect SP70, 100 μl of a 1:200 solution of antibodies against SP70 (a gift of Richard Gomer) were incubated with fixed spores for 90 minutes at 37°C. Spores were washed in PBS followed by incubation with 60 μl of a 30 μg/ml secondary antibody, Texas-Red-X-conjugated goat anti-rabbit IgG (Molecular Probes, Eugene, OR), for 60 minutes at 37°C. Spores were washed in PBS and mounted onto clean slides with mounting media and allowed to dry.

Images were taken on an inverted NIKON microscope TE2000 (NIKON Instruments, Dallas, TX) with a 100× 1.4 NA PlanFluor objective and a Quantix camera (Roper Scientific, AZ) controlled by Metamorph software (Universal Image Corp., PA) and processed with Adobe Photoshop software.

Electron microscopy (EM)

To examine spores under electron microscopy, ∼2×108 wild-type and Hip1r-null spores were harvested, washed and resuspended in 2% glutaraldehyde (EM grade) in spore buffer and incubated for 1 hour at 20°C. After this fixation, the spores were washed in spore buffer and examined either by scanning electron microscopy (SEM) or by transmission electron microscopy (TEM).

SEM spore samples were dehydrated through a graded ethanol series to 100% ethanol, critical point dried in a Tousimis Samdri-790, and coated with a gold-palladium alloy. Specimens were visualized with a Philips 515 scanning electron microscope, and images were processed with Adobe Photoshop software.

TEM spore samples were embedded in 1% agarose, post fixed in 2% osmium tetroxide for 2 hours and dehydrated through a graded ethanol series to 100% ethanol. The ethanol was replaced with epoxy resin and samples were baked. Spore samples were sectioned at less than 100 nm thickness on a RMC MT6000-XL ultramicrotome, stained with uranyl acetate and lead citrate and viewed on a Philips EM208 transmission electron microscope. Images were processed with Adobe Photoshop software.

Acknowledgments

We thank John Mendenhall of the ICMB Microscope Facility at UT Austin for the electron microscope images. We thank Richard Gomer for the kind gift of the antibody against SP70, and Tom Egelhoff for the pTxGFP vector. We also thank members of the O'Halloran and De Lozanne labs for helpful discussions and encouragement. This work is supported by NIH RO1 GM048625 to T.J.O.

Footnotes

  • Accepted September 3, 2007.

References

View Abstract