INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Yeast cells have been shown to internalize extracellular fluid, ligands bound to their receptors, and plasma membrane proteins. Internalization is time-, temperature-, and energy-dependent. Once internalized by endocytosis, most of the characterized molecules are transported to the vacuole where they are degraded. Transport to the vacuole has been proposed to pass through at least two membrane-bound compartments, which were identified by following endocytosis at low temperature (15°C), and have been termed early and late endosomes based upon the kinetics of appearance of radiolabeled, internalized pheromone in the biochemically separable organelles (Riezman, 1993; Singer-Krüger et al., 1993). The occurrence and relative densities of these kinetically defined compartments show some similarity to mammalian systems (Gorvel et al., 1991).

A genetic analysis of endocytosis has led to the identification of a number of genes that are required for at least three different steps in the pathway (Munn and Riezman, 1994; Munn et al., 1995; Riezman et al., 1996). Most of the mutants characterized thus far affect the internalization step of endocytosis. This step has been shown to be dependent upon actin (Kübler and Riezman, 1993) and a number of other gene products that affect actin function, including calmodulin (Kübler et al., 1994), a type I myosin (Geli and Riezman, 1996), two proteins with sequence homology to amphiphysin (Munn et al., 1995; Sivadon et al., 1995), which seems to play a role in clathrin-dependent endocytosis in animal cells (Shupliakov et al., 1997), and two proteins with Eps15 homology domains (Bénédetti et al., 1994; Wong et al., 1995; Wendland et al., 1996). Eps15 binds the plasma membrane clathrin adaptor complex, AP-2, and localizes to the collar of clathrin-coated pits in animal cells (Benmerah et al., 1995; Tebar et al., 1996). Mutants in clathrin also show reduced endocytic internalization (Tan et al., 1993). Remarkably, the internalization of several plasma membrane proteins seems to require their prior ubiquitination (Hein et al., 1995; Galan et al., 1996; Hicke and Riezman, 1996). This requirement may be conserved for certain endocytosed proteins in animal cells (Staub et al., 1996; Strous et al., 1996).

Transport of internalized ligand to and/or through the early endosome has been shown to be dependent upon Ypt51p and homologous proteins that are yeast homologs of the small mammalian GTPase Rab5 (Singer-Krüger et al., 1995), the SEC18 gene product that is the yeast homologue of the N-ethylmaleimide-sensitive fusion protein and requires continuous input from the secretory pathway (Hicke et al., 1997). Transport from late endosomes to the vacuole requires another small GTPase, Ypt7p, the yeast homologue of mammalian Rab7 (Schimmöller and Riezman, 1993).

While several molecular requirements for the yeast endocytic pathway have been identified (Munn and Riezman, 1994; Munn et al., 1995) that suggest similarities to animal systems, relatively little work has been published concerning the morphology of the pathway. Recent studies have attempted to follow the endocytic pathway of the -factor receptor by immunofluorescence and have identified the late endosome by immunogold localization of the receptor at the ultrastructural level (Hicke et al., 1997). Recently, cationized ferritin has been used to identify apparent internalization structures that accumulate in an endocytic mutant (Wendland et al., 1996).

To improve the morphological characterization of the yeast endocytic pathway, we have followed the internalization and intracellular targeting of positively charged Nanogold (Nanoprobes, Stony Brook, NY) by yeast spheroplasts. The positively charged Nanogold binds strongly to the cell surface and is internalized by an energy- and END gene-dependent mechanism. The Nanogold travels through small vesicles, a vesicular/tubular compartment, and a multivesicular body on its way to the vacuole. After internalization in a sec18 mutant, the internalized Nanogold is found exclusively in 30- to 50-nm diameter vesicles, most likely the primary endocytic vesicles, consistent with a vesicular mechanism of transport to the early endosome. Development of this technique should allow a more detailed analysis of the endocytic pathway and the available endocytosis mutants.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Yeast Strains, Growth Conditions, and Preparation of Spheroplasts

The wild-type strain used throughout these studies was RH144-3D (MATa his4 leu2 ura3 bar1). The following mutant strains were used: RH1995 (MATa his4 leu2 ura3 bar1 end3::URA3), and RH1737 (MATa sec18-20 his4 leu2 ura3 bar1).

All yeast strains were inoculated from saturated precultures and grown overnight in YPUAD medium (1% yeast extract, 2% peptone, 2% glucose, 40 mg/l uracil and adenine) to an OD₆₀₀ of 0.7-1.0. The cells were collected by centrifugation and converted to spheroplasts (Kübler et al., 1994) with some modifications. The cells were resuspended in one-half the original volume with 0.1 M Tris·HCl pH 9.0/10 mM 2-mercaptoethanol and incubated for 10 min at room temperature (RT). The cells were again collected by centrifugation and washed once with 10 mM Tris·HCl, pH 7.0/0.7 M sorbitol/5% glucose/0.5 × YPUAD and resuspended in the same solution to 1-2 × 10⁹ cells/ml, and then incubated with recombinant lyticase until most of the cells were converted to spheroplasts. The spheroplast preparation was centrifuged for 20 min at 1500 × g at RT and washed twice by resuspension in 10 mM Tris·HCl, pH 7.0/0.7 M sorbitol/1% glucose/0.5 × SD medium containing the appropriate additional nutrients (Dulic et al., 1991). The spheroplasts were resuspended in the last buffer to a concentration of 10⁹ spheroplasts/ml.

Incubations with Positively Charged Nanogold and Analysis by Electron Microscopy

Spheroplasts were incubated with positively charged Nanogold with two basic protocols. Spheroplasts (1 ml) were either incubated with 5 nmol of positively charged Nanogold at 0°C for 15 min and then warmed up to the indicated temperature or were preincubated at the indicated temperature before addition of 5 nmol of positively charged Nanogold. When sodium azide and sodium fluoride (10 mM each) were used, they were included in the preincubation period. The spheroplasts were then incubated for various times at the indicated temperature before fixation by direct addition of formaldehyde and glutaraldehyde to final concentrations of 3 and 0.2%, respectively.

Spheroplasts were fixed for 2 h at RT or overnight at 4°C. After fixation, samples were washed three times with 50 mM HEPES, pH 7.0/3 mM KCl. The spheroplasts were then treated with 1% metaperiodate for 30 min to avoid problems in the embedding procedure due to the remaining cells that were not converted to spheroplasts (van Tuinen and Riezman, 1987). Dehydration, infiltration, and polymerization were done according to the protocol furnished with the LR GOLD resin (London Resin, London, England). Thin sections of about 50 nm were cut and mounted on nickel grids. The positively charged Nanogold was enhanced with HQ Silver (Nanoprobes) for 4 min as described by the manufacturer. Sections were then stained with 6% uranyl acetate for 10 min. The sections were examined with a Philips 400 electron microscope (Philips Electronic Instruments, Mahwah, NJ) at 80 kV.

Labeled structures were counted for the internalization experiment at 15°C from 20 independent labeled spheroplast profiles from the indicated time points by two different people. The total numbers from the 20 sections at each time point were averaged and an SD determined. Labeled vesicles, vesicular/tubular structures (early endosomes), spherical or ovoid structures (late endosomes), and vacuoles were counted.

Quantitation of endocytosis of positively charged Nanogold by end3 spheroplasts and wild-type cells was performed by counting the number of gold particles revealed by the enhancement procedure that were on the cell surface and interior of five spheroplast sections. The total numbers were used to calculate the percentage of internalized positively charged Nanogold. The total number of gold particles counted was 4160 for wild-type spheroplasts and 4711 for end3 mutant spheroplasts.

Vesicles, early endosomes, and late endosomes that were labeled with positively charged Nanogold were counted from 20 sections at each time point and temperature for sec18 mutant spheroplasts. Nonpermissive temperature was 32°C, and permissive temperature was RT.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Positively Charged Nanogold Can Be Used as an Endocytic Marker

Wild-type yeast cells were grown overnight in rich medium, collected by centrifugation, and converted to spheroplasts in sorbitol-enriched growth medium. The spheroplasts were cooled on ice, and positively charged Nanogold was added and allowed to bind to the cell surface. After 15 min incubation on ice the sample was warmed to RT and incubated for 30 min before fixation. The fixed spheroplasts were dehydrated and embedded, and thin sections were prepared. As the size of the positively charged Nanogold is too small to be seen well by conventional electron microscopy, the gold was enhanced using HQ Silver and visualized in a Philips 400 electron microscope. After incubation of spheroplasts at RT, substantial gold labeling as a rather fine dark precipitate could been seen inside the cells in the vacuole and in a number of other intracellular membrane-bounded structures, some of which resemble the previously described late endosome (Figure 1A) (Hicke et al., 1997). The endocytic nature and specificity of the procedure were verified in several ways. Visualization of the positively charged Nanogold required enhancement because if this part of the procedure was left out, we found that no gold precipitate whatsoever was detected (our unpublished observations). Detection also required incubation with the positively charged Nanogold because enhancement performed on spheroplasts that were not incubated with gold showed only a low and characteristic background (Figure 1C). The background due to the enhancement procedure was somewhat variable and increased with the age of the enhancement solutions, but was distinguishable from the real signal resulting from enhancement of the positively charged Nanogold because the latter gave a finer precipitate (Figure 1A).

View larger version (117K): [in this window] [in a new window]   View larger version (110K): [in this window] [in a new window]   Figure 1.   Positively charged Nanogold can be used as an endocytic marker. Wild-type and end3 mutant strains were grown overnight, converted to spheroplasts, and incubated under various conditions on ice for 15 min followed by a 30-min incubation at RT. The samples were fixed and embedded and thin sections were cut. (A) Sections from wild-type cells incubated with positively charged Nanogold were enhanced with HQ Silver. (B) Sections from end3 mutant cells incubated with positively charged Nanogold and enhanced. (C) Sections from wild-type cells incubated in the absence of positively charged Nanogold, but enhanced. (D) Sections from wild-type cells incubated with positively charged Nanogold in the presence of NaN3 and NaF and enhanced. The cells were visualized in the electron microscope as described in MATERIALS AND METHODS. Bar, 200 nm.

To test for energy dependence of the internalization of the positively charged Nanogold, we incubated spheroplasts under the same conditions as above, but with addition of sodium azide and sodium fluoride to block ATP production from glycolysis and oxidative phosphorylation. Under these conditions, heavy labeling was seen on the surface of the spheroplasts, but no internalization of positively charged Nanogold occurred, showing the energy dependence of this process. Some cell surface invaginations could be seen under these conditions (Figure 1D), but we do not know whether these structures are related to bona fide endocytic internalization structures.

Due to the relatively large size of the positively charged Nanogold (1.4 nm diameter), it is most likely that the internalization occurred by endocytosis. To verify this, end3 mutant spheroplasts, which are defective for the internalization of both fluid phase and membrane-bound markers (Raths et al., 1993), were incubated as described above for the wild-type spheroplasts in the presence of sorbitol-enriched growth medium and positively charged Nanogold. The end3 mutant is defective for endocytosis at all temperatures (Bénédetti et al., 1994). Even after 30 min incubation at RT, virtually all of the positively charged Nanogold was found at the surface of the spheroplasts. This difference was quantified by counting the number of particles that were seen on the surface and interior of five complete sections of spheroplast profiles of approximately equal size. For wild-type spheroplasts, 11.5% of the gold was internal after 30 min incubation, whereas for end3 spheroplasts this amount was only 0.4% (Figure 2). We conclude from these experiments that positively charged Nanogold can be used as an endocytic marker at the ultrastructural level because it is internalized by an energy-dependent, END gene-dependent mechanism and is delivered to the vacuole, as are other known markers of endocytosis, such as -factor (Singer and Riezman, 1990), the pheromone receptors (Davis et al., 1993; Schandel and Jenness, 1994), and several transporters (Berkower et al., 1994; Volland et al., 1994; Egner et al., 1995; Lai et al., 1995).

View larger version (27K): [in this window] [in a new window]   Figure 2.   Quantitation of internalization by end3 spheroplasts. Positively charged Nanogold was incubated with wild-type and end3 spheroplasts as described in Figure 1, A and B. Gold particles were counted for five sections each on the plasma membrane and over the cell interior. The percentage of internal gold particles was calculated and plotted. The total number of cell surface particles counted was 3680 for wild-type spheroplasts and 4692 for end3 spheroplasts.

Ultrastructure of Endocytic Intermediates

To characterize intermediates in the endocytic pathway of yeast, we allowed the positively charged Nanogold to bind to yeast spheroplasts at 0°C, and then raised the temperature to 15°C, a temperature at which the transport rate through endosomes is differentially decreased compared with internalization (Singer and Riezman, 1990), and incubated for various times before fixation, embedding, and visualization of the surface-bound and internalized positively charged Nanogold. After incubation on ice the positively charged Nanogold was seen on the cell surface and in shallow plasma membrane invaginations, but not inside the cells (Figure 3A). After an 8-min incubation at 15°C, positively charged Nanogold was still found on the cell surface, sometimes in deeper cell surface invaginations (Figure 3B), but also seen inside the spheroplasts, where mainly small (~30-50 nm diameter) vesicular structures were labeled (Figure 3C). These vesicles were the most abundantly labeled structure at this time (Figure 4).

View larger version (156K): [in this window] [in a new window]   Figure 3.   Endocytosis at early time points. Wild-type cells were grown overnight, converted to spheroplasts, and incubated with positively charged Nanogold on ice for 15 min (A) and subsequently warmed up to 15°C for 8 min (B and C). The spheroplasts were prepared for electron microscopy and visualized as in Figure 1A. Bar, 200 nm.

View larger version (33K): [in this window] [in a new window]   Figure 4.   Quantitation of labeled profiles. Wild-type cells were grown overnight, converted to spheroplasts, and incubated with positively charged Nanogold on ice for 15 min and subsequently warmed up to 15°C for 0, 8, 12, or 20 min. Labeled vesicles, vesicular/tubular structures (early), oval or spherical structures (late), or vacuoles were counted. The numbers represent an average of two independent countings of the total number of labeled profiles over 20 spheroplast profiles.

After a 12-min incubation at 15°C, cell surface and vesicle labeling was still seen, but labeling of a vesicular/tubular-like structure became easily apparent (Figure 4, A and B). This structure was seen as early as after 8 min at 15°C, but labeled profiles with this structure increased greatly by 12 min (Figure 4). It was not possible to determine with our technique whether the vesicle profiles were physically connected, but the structure was often seen with a horseshoe-like shape and extended over 250-300 nm, suggesting either continuity or binding of the vesicles to an underlying framework. Indeed, labeling of both the tubular and vesicular parts of this structure was evident, with somewhat more labeling of the tubular components (Figure 6). Labeling of this structure was still apparent after 20 min of incubation, which is to be expected for an early endocytic intermediate because internalization of positively charged Nanogold probably occurred continuously during the incubation. Based on the time course of appearance of label in these structures, they may represent the morphological equivalent of biochemically defined early endosomes, through which -factor passes on its way to the vacuole (Singer-Krüger et al., 1993). These structures were often located peripherally in the cell, consistent with previous immunofluorescence studies of the -factor receptor after short times of internalization and after a block in the sec12 mutant (Hicke et al., 1997).

After 20 min, and more evidently after 90 min, all of the earlier structures were still seen, but another structure resembling the late endosome described previously (Hicke et al., 1997) became abundantly (Figure 5C and D; Figure 7) and frequently (Figure 4) labeled. This is a multivesicular structure containing internal membranes and corresponds in its morphology and location to the previously defined late endosome (Figure 8) (Hicke et al., 1997). The late endosomal compartment was relatively large, frequently spherical or oval in shape and approximately 200 nm across (Figure 8). It was most often found near the vacuole. The interior of the late endosome contained an electron-dense reticulum, most likely internal membranes. Most, if not all, of the gold labeling occurs on these internal membranes, suggesting that the internal membranes of this structure are derived from the plasma membrane.

View larger version (157K): [in this window] [in a new window]   Figure 5.   Early and late endocytic structures. Wild-type cells were grown overnight, converted to spheroplasts, and incubated with positively charged Nanogold on ice for 15 min and subsequently warmed up to 15°C for 12 min (A and B) or 20 min (C and D). The spheroplasts were prepared for electron microscopy and visualized as in Figure 1A. Bar, 200 nm.

View larger version (167K): [in this window] [in a new window]   Figure 6.   Early endosome profiles. Early endosomal profiles were taken from wild-type cells incubated and treated as in Figure 5. Bar, 200 nm.

View larger version (106K): [in this window] [in a new window]   Figure 7.   Late endocytic structures. Wild-type cells were grown overnight, converted to spheroplasts, and incubated with positively charged Nanogold on ice for 15 min and subsequently warmed up to 15°C for 90 min. The spheroplasts were prepared for electron microscopy and visualized as for Figure 1A. Bar, 200 nm.

View larger version (153K): [in this window] [in a new window]   Figure 8.   Late endosome profiles. Late endosome profiles were taken from wild-type cells treated as in Figure 5, C and D, or Figure 7. Bar, 200 nm.

In some cases, at this late time point, some vacuolar labeling was also detected. Vacuolar labeling was always less dense than that found in the late endosomes. This was expected because delivery of endocytic markers to the vacuole is greatly reduced at 15°C (Singer and Riezman, 1990) and because the volume of a late endosome is less than 1% that of the vacuole. We suggest from these studies that the positively charged Nanogold follows the following endocytic pathway; cell surface  vesicles  vesicular/tubular structure (early endosome)  multivesicular endosome (late endosome)  vacuole. This is likely to be similar to the endocytic pathway that -factor and its receptor follow (Singer-Krüger et al., 1993; Hicke et al., 1997).

Small Endocytic Vesicles Accumulate in the sec18 Mutant

One particular feature of the yeast endocytic pathway is that the internalization step is blocked partially by mutants in clathrin (Tan et al., 1993) and completely by mutants in actin and proteins that affect actin structure (Kübler and Riezman, 1993; Riezman et al., 1996). Clathrin-coated vesicles detected by electron microscopy vary in diameter from approximately 60 to 100 nm (Stoorvogel et al., 1996), including the coat. An intact actin network is also required for internalization through caveolae (Parton et al., 1994). Vesicles formed through this pathway have an apparent size of 60-70 nm in diameter. Phagocytosis, which requires actin as well, results in the internalization of much larger structures of variable size. Therefore, we wanted to determine the size and uniformity of the primary endocytic intermediates in yeast. To do this we used a mutant that could block their fusion, but still permit their formation.

Sec18p is the yeast homolog of the N-ethylmaleimide-sensitive fusion protein that is required for many vesicle fusion events in yeast in the secretory pathway (Graham and Emr, 1991) and is implicated in homotypic fusion of vacuoles (Haas and Wickner, 1996), and the sec18 mutation confers the earliest known postinternalization block in the endocytic pathway (Hicke et al., 1997). We reasoned, therefore, that the sec18 mutation is likely to block fusion of the primary endocytic vesicles with their target. Therefore, we performed an internalization experiment using positively charged Nanogold in the sec18 mutant at nonpermissive temperature. Spheroplasts from wild-type or sec18 mutant cells were shifted to 32°C for 10 min to inactivate mutant Sec18p, positively charged Nanogold was added, and the spheroplasts were incubated for a further 30 min and visualized by electron microscopy as described above. In wild-type cells, all of the above described compartments, including the vacuole (Figure 9A), were labeled, showing that the internalization protocol works at 32°C. Most of the positively charged Nanogold remained at the cell surface under these conditions in the sec18 mutant (Figure 9B). The sec18 spheroplasts accumulated a number of small vesicles of homogeneous size of which only a subset were labeled with positively charged Nanogold, consistent with the requirement for Sec18p for numerous intracellular membrane fusion events. The internalized, positively charged Nanogold was found exclusively in small vesicles of approximately 30-50 nm diameter (Figures 9C and 11) that seem to be similar or identical to those detected at early time points at 15°C (Figure 3C). No positively charged Nanogold was seen in structures resembling the early or late endosomes described above. To quantify the endocytic block and show that these vesicles accumulate due to the block imposed by the sec18 mutation, we performed positively charged Nanogold internalization experiments for various times at permissive temperature (RT) and nonpermissive temperature (32°C) in the sec18 mutant spheroplasts. At RT, vesicles accumulated with time, but this accumulation peaked at the 10-min time point. Both early and late endosomes were detected at the later time points (Figure 10). At nonpermissive temperature, only vesicles were seen, and they accumulated with time throughout the experiment. In addition, appearance of the vesicles in sec18 spheroplasts containing positively charged Nanogold was dependent upon endocytosis because it was completely blocked in an end3 sec18 double mutant (our unpublished observations). Therefore, we suggest that the sec18 mutation blocks fusion of the primary endocytic vesicles in yeast and that the latter are of a relatively small and uniform size.

View larger version (117K): [in this window] [in a new window]   Figure 9.   Small endocytic vesicles are seen in sec18 spheroplasts. Wild-type (A) or sec18 (B and C) cells were grown overnight, converted to spheroplasts, preincubated at 32°C for 10 min, and then incubated with positively charged Nanogold at 32°C for 30 min. The spheroplasts were prepared for electron microscopy and visualized as for Figure 1A. Bar, 200 nm.

View larger version (26K): [in this window] [in a new window]   Figure 10.   Small endocytic vesicles accumulate in sec18 spheroplasts. Mutant (sec18) cells were grown overnight at 24°C, converted to spheroplasts, and incubated with positively charged Nanogold on ice for 15 min and subsequently warmed up to RT or nonpermissive temperature (32°C) for 0, 5, 10, or 40 min. Labeled vesicles, vesicular/tubular structures (early), and oval or spherical structures (late) were counted from 20 spheroplast sections at each time point and temperature.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this paper we have developed the use of positively charged Nanogold to follow the endocytic pathway in yeast at the ultrastructural level. Positively charged Nanogold fulfills all of the criteria for an endocytic marker. Its internalization is time-, temperature-, and energy-dependent, and it is only found on the cell surface or within membrane-bound internal compartments. Its internalization is greater than 20-fold reduced in the end3 mutant, which has previously been shown to be required for the internalization of a number of endocytic markers, including a fluid phase marker, lucifer yellow CH (Riezman, 1985), a membrane phase marker, FM4-64 (Vida and Emr, 1995; Wendland et al., 1996), as well as several specific markers, including -factor, pheromone receptors, and permeases (Davis et al., 1993; Raths et al., 1993; Volland et al., 1994; Lai et al., 1995). Positively charged Nanogold is a nonspecific probe that has several advantages as a marker for absorptive endocytosis. It binds strongly and evenly to the surface of spheroplasts, and it can be used to generate an electron-dense aggregate that can be visualized in the electron microscope even though it is, itself, relatively small, and cannot be degraded, allowing its visualization throughout the endocytic pathway, including after delivery to the vacuole. Cationized ferritin has been used to follow endocytosis in yeast spheroplasts (Wendland et al., 1996). This marker is also electron dense; however, it is much larger than positively charged Nanogold and is subject to degradation, introducing another variable into the experimental conditions. Therefore, the use of positively charged Nanogold should be of general use in the analysis of endocytic compartments and mutants.

The use of positively charged Nanogold and mutant cells has allowed us to visualize and confirm the existence of distinct organelles along the endocytic pathway, including primary endocytic vesicles, a vesicular/tubular structure that we term an early endosome, late endosomes, and vacuoles. As in animal cells, there are two biochemically and morphologically distinct organelles in the endocytic pathway in addition to the primary endocytic vesicles. Overall, it is remarkable that the endocytic organelles from yeast appear quite similar to those described for endocytic organelles in animal cells. The late endosome, which has been visualized previously using antibodies to the -factor receptor (Hicke et al., 1997), is a multivesicular compartment similar to that found for animal cells (McDowall et al., 1989). The organelle is rather large, being one of the more conspicuous organelles in yeast after the nucleus and vacuole. After fixation, embedding, and visualization, the lumen of the organelle contains internal membranes but is not very electron dense, consistent with its distribution in low-density fractions in equilibrium density gradients (Singer-Krüger et al., 1993). The late endosome is found most frequently, but not exclusively, close to the vacuole, and profiles can sometimes be seen where the late endosome may be fusing with the vacuole, but further analysis will be necessary to address this point. It is interesting to note that most of the labeling was found on the internal structures of the late endosome, suggesting that the gold bound to plasma membrane molecules and early endosomes finds it way into the interior of this organelle. We cannot be certain, however, that these membranes are really internal. We cannot rule out that they are the result of multiple infoldings of the outer membrane of the structure, even though we did not detect any such infoldings in the multiple structures we examined.

At early time points of incubation with positively charged Nanogold at 15°C we could detect a labeled vesicular/tubular organelle. The low-temperature incubations allowed us to demonstrate that appearance of positively charged Nanogold in this structure preceded its appearance in the late endosome. This structure is not generated due to the low temperature incubations because it was also easily seen after internalization experiments at 30°C (our unpublished observations) and 32°C (Figure 9). This structure may be one of the morphological counterparts of biochemically defined early endosomes (Singer-Krüger et al., 1993) because the structure is labeled early during time course incubations and is found mainly in the cell periphery in accordance with immunofluorescence studies (Hicke et al., 1997). This structure resembles early endosomes from animal cells. In animal cells early endosomes have been postulated to be an interconnected structure with tubular and vesicular components (Hopkins et al., 1990; Stoorvogel et al., 1996). The vesicular components in animal cells are sometimes coated with clathrin (Stoorvogel et al., 1996). Our early endosomal profiles could be very similar because we found both tubular and vesicular structures that were consistently associated with each other. Even if these components are not actually interconnected, there must be some underlying structure because the profiles seen are fairly distinct and well ordered. The gold labeling of these structures was found in both the tubular and vesicular parts but was more often associated with the tubular components. Some of the vesicular profiles were clearly unlabeled. This would be expected if some of the vesicular profiles resulted from incoming vesicles from the secretory pathway, e.g., from the trans-Golgi. Endocytic traffic from early to late endosomes requires a continuous input from the secretory pathway (Hicke et al., 1997); therefore, one would expect such incoming traffic.

Putative Primary Endocytic Vesicles

Putative primary endocytic vesicles were visualized with the aid of the sec18 mutant. In sec18 cells, only small uniform vesicles of approximately 30-50 nm diameter were seen labeled with positively charged Nanogold at nonpermissive temperature. These vesicles were similar in size to other vesicles that accumulated in the sec18 mutant (Figure 11) (Kaiser and Schekman, 1990). Several arguments suggest that these are primary endocytic vesicles. First, these were the only labeled structures that we detected in sec18 spheroplasts at nonpermissive temperature, and they accumulated with time at nonpermissive temperature in the mutant spheroplasts. In wild-type cells at the same temperature all of the above described endocytic structures were labeled efficiently. If the positively charged Nanogold had reached the early endosome in the sec18 spheroplasts we would have detected it as we did in wild-type cells after similar incubation periods. Second, the sec18 block represents the first known postinternalization block in endocytosis. This was concluded from the following experiments. When -factor was internalized in the sec12 mutant, it accumulated in biochemically defined early endosomes (Hicke et al., 1997). Detection of the -factor receptor by immunofluorescence under similar conditions in sec12 cells revealed that the receptor was concentrated in peripheral, relatively large punctate structures. Similar experiments using the sec18 mutant showed accumulation of the -factor receptor in smaller, peripheral dots by immunofluorescence easily distinguishable from the structures accumulated in sec12 cells. A double mutant (sec12 sec18) showed a sec18 phenotype. These data showed that the small, peripheral dots that accumulated in sec18 cells were epistatic to and preceded the large peripheral compartment, which fractionated like the biochemically defined early endosomes. The small peripheral dots seen in sec18 cells, therefore, are likely precursors of the early endosomes seen by immunofluorescence in sec12 cells and must correspond to the vesicles seen here in sec18 spheroplasts because they contain all of the internalized label. Third, mammalian Sec18p has been shown to associate with endocytic clathrin- coated vesicles, and N-ethylmaleimide, an inhibitor of Sec18p, blocks fusion of clathrin-coated vesicles (Woodman and Warren, 1991; Steel et al., 1996). Finally, the vesicles seen in sec18 spheroplasts resemble the first endocytic intermediates observed using wild-type spheroplasts at 15°C.

View larger version (177K): [in this window] [in a new window]   Figure 11.   Putative primary endocytic vesicle profiles. Profiles of putative primary endocytic vesicles were taken from sec18 spheroplasts incubated for 30 min at 32°C with positively charged Nanogold.

The molecular requirements for the internalization step of endocytosis in yeast show similarities to several different types of endocytic internalization seen in mammalian cells, but the similarities are not close enough to any of them to be certain of a true homology in mechanism (Riezman et al., 1996). For instance, mutations in clathrin affect the internalization of ligands in both cell types, but in yeast this effect is only partial. On the other hand, actin is absolutely required for endocytosis in yeast, whereas actin depolymerization in animal cells using cytochalasin D has been proposed to affect clathrin-dependent internalization from apical, but not basolateral, plasma membrane (Gottlieb et al., 1993; Jackman et al., 1994). This does not rule out that actin is required for clathrin-mediated endocytosis at the basolateral surface because cytochalasin D may not depolymerize all cellular actin equally. In fact, a role for actin in all clathrin-mediated endocytosis has been proposed recently (Lamaze et al., 1997). Actin is also required for two other types of endocytic internalization in animal cells: induced internalization through caveolae (Parton et al., 1994) and phagocytosis (Greenberg et al., 1991). A clear distinction between the clathrin or caveolar uptake and phagocytic uptake is the size of the primary endocytic vesicles. For this reason it was important to identify the primary endocytic vesicle in yeast.

The size and regularity of the putative primary endocytic vesicles described here would be most consistent with their being derived through a coat-dependent mechanism, rather than a solely actin-based, phagocytic-like mechanism. The size of the primary endocytic structure determined by the latter mechanism depends upon the size of the particle being internalized, not upon the dimensions of an assembled coat structure, such as clathrin coats or caveolar coats. It is hard to imagine how actin could generate small, uniform vesicles independent of a coat protein. In yeast, no protein with clear sequence homology to caveolin is present; therefore, a caveolin homologue apparently plays no role in this event. On the other hand, clathrin coats could participate in endocytic internalization because clathrin mutants show a 50% block in this step. One possible role for clathrin in endocytosis that would be consistent with the partial block could be the regulation of the size of the endocytic vesicle and/or the recruitment of receptors into internalization structures. The precise role of clathrin in the process of internalization will have to await further experimentation and would benefit greatly from the detection of the internalization structures. Hopefully, some of the end mutants that affect the internalization step of endocytosis will be useful for this.