Efficient Cargo Sorting by ESCRT-I and the Subsequent Release of ESCRT-I from Multivesicular Bodies Requires the Subunit Mvb12 (original) (raw)

• Journal List
• Mol Biol Cell
• v.18(2); 2007 Feb
• PMC1783790

Mol Biol Cell. 2007 Feb; 18(2): 636–645.

Matt Curtiss

Department of Biology, University of Utah, Salt Lake City, UT 84112-9202

Charles Jones

Department of Biology, University of Utah, Salt Lake City, UT 84112-9202

Markus Babst

Department of Biology, University of Utah, Salt Lake City, UT 84112-9202

Sandra Lemmon, Monitoring Editor

Department of Biology, University of Utah, Salt Lake City, UT 84112-9202

Corresponding author.

Received 2006 Jul 10; Revised 2006 Oct 26; Accepted 2006 Nov 20.

Copyright © 2007 by The American Society for Cell Biology

Supplementary Materials

[Supplemental Material]

GUID: C1F0284B-C0B4-4483-A9BF-41E1B387E5D6

GUID: AAD8F9C7-0F3E-4A9F-AD88-1E4B4133D059

GUID: 090887C7-03BD-41EC-B69D-D3753FF4C713

Abstract

The endosomal sorting complex required for transport (ESCRT)-I protein complex functions in recognition and sorting of ubiquitinated transmembrane proteins into multivesicular body (MVB) vesicles. It has been shown that ESCRT-I contains the vacuolar protein sorting (Vps) proteins Vps23, Vps28, and Vps37. We identified an additional subunit of yeast ESCRT-I called Mvb12, which seems to associate with ESCRT-I by binding to Vps37. Transient recruitment of ESCRT-I to MVBs results in the rapid degradation of Mvb12. In contrast to mutations in other ESCRT-I subunits, which result in strong defects in MVB cargo sorting, deletion of MVB12 resulted in only a partial sorting phenotype. This trafficking defect was fully suppressed by overexpression of the ESCRT-II complex. Mutations in MVB12 did not affect recruitment of ESCRT-I to MVBs, but they did result in delivery of ESCRT-I to the vacuolar lumen via the MVB pathway. Together, these observations suggest that Mvb12 may function in regulating the interactions of ESCRT-I with cargo and other proteins of the ESCRT machinery to efficiently coordinate cargo sorting and release of ESCRT-I from the MVB.

INTRODUCTION

Eukaryotic cells continuously remove transmembrane proteins from the plasma membrane by endocytosis and deliver them to the lumen of the lysosome for degradation (for review, see Gruenberg and Stenmark, 2004; Katzmann et al., 2002). The topological problem of degrading transmembrane proteins in the lumen of the lysosome is solved by the formation of endosomal structures called multivesicular bodies (MVBs), which are formed by invagination and budding of vesicles from the limiting endosomal membrane into the lumen of the endosome. During this process, endosomal transmembrane proteins destined for degradation are sorted into forming vesicles and upon MVB fusion with the lysosome/vacuole, they are delivered to the lumen of the compartment, whereas others are maintained on the limiting membrane. A variety of studies indicate that monoubiquitination serves as a signal that directs protein cargo into the MVB pathway (Babst, 2005; Katzmann et al., 2002).

Sorting of endocytosed cell surface proteins through the MVB pathway plays an essential role in maintaining proper cell surface protein composition. In addition, this pathway acts to quickly and dramatically change the protein composition of the cell surface during processes such as differentiation and adaptation. For example, growth factor signaling is regulated in part by the controlled endocytosis and degradation of growth factor receptors, a process that is disrupted in pathological states of uncontrolled cellular proliferation in certain types of cancer (Di Fiore and Gill, 1999; Babst et al., 2000; Ceresa and Schmid, 2000; Thompson et al., 2005; Vaccari and Bilder, 2005). In addition to protein degradation, the MVB pathway also functions in the targeting and transport of lysosomal resident proteins. For example, in yeast the membrane-associated vacuolar enzyme carboxypeptidase S (CPS) is delivered from the _trans_-Golgi via MVBs to the lumen of the vacuole, the yeast organelle corresponding to the mammalian lysosome where it acts as a lumenal protease (Odorizzi et al., 1998). Thus, the MVB pathway also plays an important role in maintaining the function of the lysosomal/vacuolar compartment. Furthermore, MVBs have been shown to play an essential role in the immune response of mammals where they function in antigen presentation by dendritic cells and in the formation of exosomes (Murk et al., 2002).

Both MVB vesicle formation and sorting of ubiquitinated MVB cargo depends on the function of a group of at least 16 conserved proteins that were originally identified in the yeast Saccharomyces cerevisiae as class E vacuolar protein sorting (Vps) proteins (Babst, 2005). Deletion of each of the class E VPS genes in yeast results in the accumulation of endosomal cargo in large structures adjacent to the vacuole, called class E compartments (Raymond et al., 1992). This distinctive phenotype allows rapid identification of MVB sorting mutants in yeast. In addition to the morphological phenotype, in class E vps mutants transmembrane proteins such as CPS that are normally sorted to the vacuolar lumen are mislocalized to the limiting membrane of the vacuole (Odorizzi et al., 1998). Furthermore, the endosomal defects in class E vps mutants result in impaired transport of soluble vacuolar hydrolases from the _trans_-Golgi to endosomes, which ultimately leads to the secretion of a portion of the newly synthesized enzymes (Raymond et al., 1992).

The majority of the class E Vps proteins are constituents of three separate hetero-oligomeric protein complexes called endosomal sorting complex required for transport (ESCRT)-I, ESCRT-II, and ESCRT-III and the Vps4 ATPase complex. The ESCRT protein complexes are transiently recruited from the cytoplasm to the endosomal membrane where they function in transmembrane protein sorting and the formation of MVB vesicles. The ESCRT-I protein complex is recruited to the MVB by Vps27, a class E Vps protein that localizes to the endosome by binding to the lipid phosphatidylinositol 3-phosphate (Bache et al., 2003; Katzmann et al., 2003). On the endosomal membrane both ESCRT-I and Vps27 bind to ubiquitinated endosomal cargo (Katzmann et al., 2001; Pornillos et al., 2002; Bilodeau et al., 2003; Swanson et al., 2003; Sundquist et al., 2004; Teo et al., 2004; Hirano et al., 2006). In addition, ESCRT-I interacts with ESCRT-II, which initiates the oligomerization of at least four small coiled-coil proteins, resulting in the formation of the ESCRT-III complex (Babst et al., 2002b; Teo et al., 2006). ESCRT-III is a large endosome-associated structure that seems to function in the concentration of MVB cargo (Babst et al., 2002a). After protein sorting has been completed, the multimeric AAA-type ATPase Vps4 binds to ESCRT-III and disassembles the ESCRT-III complex in an ATP-dependent manner (Babst et al., 1998, 2002a).

ESCRT-I is a 350-kDa protein complex that has been shown to be composed of the three class E Vps proteins: Vps23, Vps28, and Vps37 (Babst et al., 2000; Katzmann et al., 2001). However, expression of these subunits in Escherichia coli does not result in the formation of a 350-kDa complex, suggesting that additional unidentified subunits might be necessary for the formation of ESCRT-I in yeast (Kostelansky et al., 2006; Teo et al., 2006). In this publication, we present evidence for a fourth subunit of the yeast ESCRT-I complex, called Mvb12. Loss of Mvb12 results in a partial defect in MVB sorting and the mistargeting of ESCRT-I to the vacuolar lumen. These data suggest that Mvb12 is not an essential factor for ESCRT-I function but that it may be required for efficient cargo sorting and the release of ESCRT-I from the MVB.

MATERIALS AND METHODS

Materials

Monoclonal antibodies specific for the hemagglutinin (HA) epitope and green fluorescent protein (GFP) were purchased from Covance (Princeton, NJ). Polyclonal antisera against Snf7 (Babst et al., 1998) and Vps23 (Babst et al., 2000) have been characterized previously.

Strains and Media

S. cerevisiae strains used in this work are listed in Table 1. Yeast strains were grown in standard yeast extract-peptone-dextrose (YPD) or synthetic medium supplemented with essential amino acids as required for maintenance of plasmids (YNB) (Sherman et al., 1979). The deletion strains were constructed by transforming a yeast strain with a DNA fragment containing the HIS3 or URA3 gene flanked by 50 base pairs specific for the 5′ and 3′ region of the corresponding gene. Yeast cells were selected for the presence of the URA3 or HIS3 gene, and the deletions were confirmed by PCR analysis of the chromosomal DNA.

Table 1.

Strains and plasmids used in this study

DNA Manipulations

Recombinant DNA work was performed using standard protocols (Sambrook et al., 1989). Transformation of S. cerevisiae was done by the lithium acetate method as described in Ito et al. (1983). The plasmids used in this study are listed in Table 1. The MVB12 gene was obtained by polymerase chain reaction (PCR) amplification of SEY6210 chromosomal DNA and inserted into the EcoRI/BamHI sites of pRS416, resulting in the plasmid pMB238. For the construction of pMB240, pCJ2, and pMC26, a DNA fragment coding for three HA epitopes was ligated with PCR products containing either MVB12 or VPS37, respectively, and inserted into BamHI/SalI-digested pRS416 or pRS426 vector. pMB239 and pMB306 were constructed by ligating a PCR product containing MVB12 with a GFP-containing fragment from the vector pEGFP-C1 (Clontech, Mountain View, CA) into the BamHI/SalI sites of the vectors pRS416 and pRS426.

Experimental Procedures

For native immunoprecipitation experiments, 10 OD600 equivalents of yeast cells were spheroplasted and then osmotic lysed in 1 ml of phosphate-buffered saline (PBS) (8 g/l NaCl, 0.2 g/l KCl, 1.44 g/l Na2HPO4, and 0.24 g/l KH2PO4, pH 7.2) containing protease inhibitors. The resulting cell extracts were centrifuged at 15,000 × g for 5 min. The resulting supernatant was incubated with antibodies (1/250 anti-HA antibody) for 1.5 h at 4°C. The antibodies were isolated by adding GammaBind G-Sepharose (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom). After incubation for 1 h at 4°C, the Sepharose was washed three times with PBS containing 0.5% Tween 20, and the antibodies together with the bound antigen were eluted by boiling the Sepharose in SDS-PAGE sample buffer. The resulting fractions were analyzed by SDS-PAGE and Western blotting. Vps23-ProtA was affinity purified from cell extract using IgG-Sepharose (GE Healthcare). Preparation of cell extract, incubation with the Sepharose, washes, and sample elution were performed as described for the immunoprecipitation experiments. Immunofluorescence microscopy was performed on fixed spheroplasted cells as described in Babst et al. (1998). Fluorescence microscopy was performed on a deconvolution microscope (DeltaVision, Applied Precision, Issaquah, WA). For gel filtration analysis, yeast cells were spheroplasted and lysed in PBS containing 0.1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride and protease inhibitor cocktail (Complete; Roche Molecular Biochemicals, Indianapolis, IN). The lysate was centrifuged at 100,000 × g, and the resulting supernatant (∼2 mg of protein) was loaded on a Sephacryl S300 column (16/60; GE Healthcare) and separated in presence of PBS. Subcellular fractionations were performed as described previously (Babst et al., 1997). Total cell extracts for Western blot analysis were obtained by glass bead lysis of yeast from 6 ml of culture (OD600 = 0.5) in SDS-PAGE sample buffer (2% SDS, 0.1 M Tris, pH 6.8, 10% glycerol, 0.01% bromphenol blue, and 5% β-mercaptoethanol).

RESULTS

Identification of Mvb12

A global analysis of protein localization in yeast identified a group of uncharacterized proteins that are localized to endosomal compartments (Huh et al., 2003). We tested the corresponding yeast deletion strains for phenotypes that are indicative of MVB mutations. Our studies identified the open-reading frame YGR206w as a factor required for the MVB sorting pathway (see phenotypic analysis in later sections of this article). YGR206w encodes a 12-kDa protein and therefore, we named this gene MVB12. Mvb12 is a basic protein (pI 8.35) that contains no obvious structural motifs and evolutionarily is poorly conserved. No clear MVB12 homologues are identified in higher eukaryotes. Sequence homology searches identify uncharacterized open-reading frames in other genomes of fungal species as potential MVB12 homologues. However, even among different yeast species the sequence conservation of Mvb12 is poor.

Mvb12 Transiently Localizes to MVBs

To study the localization of Mvb12, we constructed MVB12-GFP, a functional C-terminal GFP fusion of MVB12 (Supplemental Figure 1). We expressed this fusion protein in different strains and analyzed these strains by fluorescence microscopy (Figure 1A). Consistent with the results of the global protein localization study, we observed that in wild-type cells Mvb12-GFP localized to the cytoplasm and to small compartments consistent with endosomes.

Localization of Mvb12 to MVBs is dependent on Vps27 and ESCRT-I. (A) Fluorescent microscopy of yeast strains (wild type [WT], MBY66; _vps4_Δ, MBY64; _vps4_Δ/_vps27_Δ, MBY22; and _vps4_Δ/_vps37_Δ, MCY19) expressing MVB12-GFP (pMB239). White arrows indicate class E compartments, aberrant endosomal structures present in class E vps mutants. (B) Immunofluorescence microscopy of _vps4_Δ (MBY3) and _vps4_Δ_vps27_Δ (MBY22) expressing MVB12-GFP (pMB239) demonstrates that colocalization of Mvb12 with endosome-associated ESCRT-III subunit Snf7 is dependent on Vps27. The Snf7 protein was visualized using specific anti-Snf7 antibody and fluorescently labeled secondary antibody. (C) Fluorescence microscopy of WT and _vps4_Δ (MBY3) strains expressing VPS23-RFP (pMB320) and MVB12-GFP (pMB321) demonstrates colocalization of Mvb12 with ESCRT-I.

The AAA-type ATPase Vps4 functions in the dissociation of the ESCRTs from MVBs; therefore, mutations in VPS4 result in the accumulation of the ESCRT machinery on aberrant endosomal membranes, the class E compartments (Figure 1A, arrows; Babst et al., 1998, 2002a). The deletion of VPS4 resulted in the redistribution of Mvb12-GFP from the cytoplasm to larger compartments adjacent to vacuoles, consistent with class E compartments. Furthermore, by immunofluorescence microscopy we observed colocalization of Mvb12-GFP with the ESCRT-III subunit Snf7 in these cells, indicating that in _vps4_Δ Mvb12-GFP accumulates on endosomal compartments (Figure 1B). Interestingly, we observed that the endosomal accumulation of Mvb12-GFP in _vps4_Δ was dependent on the presence of Vps27 (Figure 1A). In _vps4_Δ_vps27_Δ cells, Mvb12-GFP localized mainly to the cytoplasm and a minor pool associated with smaller dispersed compartments. These compartments do not colocalize with Snf7, suggesting that they are not late endosomal structures (Figure 1B). Previous studies have shown that ESCRT-I localizes to the endosomal membrane by binding to Vps27 and that ESCRT-I accumulates on the membrane in a VPS4 deletion strain (Bache et al., 2003; Katzmann et al., 2003). Therefore, our localization studies suggested that Mvb12 might be associated with ESCRT-I. This idea was further corroborated by the observation that red fluorescent protein (RFP)-tagged Vps23 and Mvb12-GFP colocalized in both wild-type and _vps4_Δ strains (Figure 1C). Furthermore, in strains deleted for the ESCRT-I subunit _vps37_Δ the fluorescent signal of Mvb12-GFP was dramatically reduced compared with other strain backgrounds (Figure 1A). This observation suggested that Mvb12-GFP is unstable in ESCRT-I mutants, a result consistent with Mvb12 being an ESCRT-I subunit.

Mvb12 Is a Subunit of the ESCRT-I Complex

To further study the effect of ESCRT mutations on the stability of Mvb12, we constructed a functional C-terminal HA-tagged Mvb12 (Mvb12-HA; Supplemental Figure 1) and expressed this fusion protein in wild type and mutant strain backgrounds. Cell extracts were prepared and analyzed for the presence of Mvb12-HA by Western blot (Figure 2A). The analysis of wild-type cells showed two specific bands, a major band at ∼15 kDa that corresponds to the predicted molecular mass of Mvb12-HA (12 kDa for Mvb12 plus 3 kDa for HA-tag) and a minor band at ∼20 kDa, which may be a modified form of Mvb12-HA (Mvb12-HA*; Figure 2A). A similar band pattern was observed with cell extracts from strains deleted for VPS4, VPS27, VPS36 (encoding an ESCRT- II subunit), and SNF7 (encoding an ESCRT-III subunit). In contrast, cell extracts from the ESCRT-I mutant strains _vps23_Δ, _vps28_Δ, and _vps37_Δ exhibited a dramatically reduced amount of the major, 15-kDa band, confirming the observation by microscopy that Mvb12 is unstable in ESCRT-I mutants. However, the amount of the modified form of Mvb12-HA remained similar in all mutant strains (Figure 2A). To determine where this modification occurs, we studied the cellular distribution of Mvb12-HA and Mvb12-HA* by subcellular fractionation. For this purpose, we prepared cell extracts by spheroplasting and osmotic lysis and separated the extract by centrifugation at 15,000 × g into a soluble fraction (S) and a pellet fraction (P). Previous studies have shown that under these conditions, the MVB-associated protein pool is found in the pellet, whereas the cytoplasmic fraction remains in the supernatant (Babst et al., 1998, 2002a). The Western blot analysis of the fractionation samples showed that in wild-type cells Mvb12-HA is mainly found in the soluble, cytoplasmic fraction, whereas in cells deleted for VPS4 an increased amount of Mvb12-HA is observed in the membrane bound fraction (Figure 2B). These results are consistent with the Mvb12-GFP localization observed by microscopy, although the redistribution to the membrane in _vps4_Δ is less pronounced than expected from the microscopy data (Figure 1A). Similarly, ESCRT-I has been shown by microscopy to accumulate on endosomal membranes in _vps4_Δ (Figure 7A; Katzmann et al., 2001); however, only a small portion of Vps23 is found in the membrane fraction of this strain (Figure 2B). A likely explanation for this discrepancy between the two methods is that during the fractionation procedure, peripheral proteins such as ESCRT-I might partially dissociate from the membrane.

Mvb12 is unstable in ESCRT-I mutants and physically interacts with the ESCRT-I complex. (A) Western blot analysis using anti-HA antibodies of total cell extracts from yeast strains expressing MVB12-HA (pMB240) or as a negative control (con.) MVB12 (pMB238). The strains used are described in Table 1. (B) Subcellular fractionation of yeast strains expressing MVB12-HA either from a low copy (pMB240) or a high copy (2μ, pCJ2) plasmid. The resulting supernatant (S) and pellet (P) fractions were analyzed by Western blot for the presence of Mvb12-HA and Vps23. (C) Extracts from wild-type cells expressing MVB12 (negative control) or MVB12-HA were used for an immunoprecipitation experiment using anti-HA antibodies. Samples of the immunoprecipitated material (bound, lanes 1 and 2) and the remaining supernatants (lanes 5 and 6) were analyzed by Western blot for the presence of Mvb12-HA and Vps23. Extracts from wild-type yeast expressing MVB12-HA in addition to VPS23 (negative control) or VPS23-ProtA were subjected to affinity purification by using IgG-Sepharose. The resulting enriched material (lanes 3 and 4) and the remaining supernatants (lanes 7 and 8) were analyzed by Western blot for the presence of Mvb12-HA and Vps23.

In _mvb12_Δ, ESCRT-I is mislocalized to the vacuolar lumen. (A) Fluorescence microscopy of strains expressing plasmid encoded GFP-VPS27 (pEE27-4) and VPS23(M85T)-GFP (pMB319) or chromosomally integrated VPS23-GFP and VPS36-GFP. Arrows indicate class E compartments. (B) Western blot analysis using a GFP-specific antibody of extracts from cells expressing chromosomally encoded Vps23-GFP (WT, DKY54; _mvb12_Δ, MBY73; 2μ ESCRT-II, pMB175).

As expected from our previous results, the deletion of the VSP23 resulted in a reduced amount of Mvb12-HA in both the soluble and the pelletable pool (Figure 2B). In contrast to Mvb12-HA the amount and localization of the modified form Mvb12-HA* is not effected by the deletions of either VPS4 or VPS23. In all strains, Mvb12-HA* is found exclusively in the pelletable fraction. This result suggested that the modification of Mvb12-HA might occur on the endosomal membrane and that the modified pool of Mvb12-HA may not be associated with ESCRT-I. This is further supported by the observation that in cells overexpressing Mvb12-HA, the majority of the protein is found in the pellet in the modified form (Figure 2B). The nature of the modification is not known. However, based on the size shift Mvb12-HA* could represent a phosphorylated or monoubiquitinated form of Mvb12.

These data are consistent with the model that Mvb12 is a subunit of the ESCRT-I protein complex. To further test this model, we studied a potential physical interaction between Mvb12 and ESCRT-I by immunoprecipitation experiments (Figure 2C). We prepared extracts from cells expressing either MVB12 (negative control) or MVB12-HA and performed immunoprecipitations under native conditions by using anti-HA antibodies. The Western blot analysis of the resulting samples demonstrated the specific coimmunoprecipitation of Vps23 with Mvb12-HA (Figure 2C, lane 2). No Vps23 was immunoprecipitated from the Mvb12-containing control extract (Figure 2C, lane 1). A similar result was obtained by IgG-affinity purification. Cell extracts of strains expressing MVB12-HA together with VPS23 or a functional protein A-tagged VPS23 (VPS23-ProtA) were incubated with IgG-Sepharose, and the bound material was tested for the presence of Vps23 and Mvb12-HA. The data showed a specific copurification of Mvb12-HA with Vps23-ProtA (Figure 2C, lane 4), whereas no enrichment of Mvb12-HA was observed from the _VPS23_-expressing control extract (Figure 2C, lane 3). As a control for the presence of Mvb12-HA and Vps23 in the extracts, we performed Western blots of the supernatant after immunoprecipitation or affinity purification (Figure 2C, lanes 5–8). In summary, both immunoprecipitation and IgG-affinity purification experiments indicated a physical interaction between Mvb12 and ESCRT-I, supporting the notion that Mvb12 is a subunit of the ESCRT-I complex.

Finally, an extract from wild-type cells expressing MVB12-HA was separated on a gel filtration column, and the resulting fractions were analyzed by Western blot for the presence of Mvb12-HA and the ESCRT-I subunit Vps23. The results indicated a native molecular mass of ∼350 kDa for Mvb12-HA similar to the size of ESCRT-I (as indicated by Vps23; Figure 3A, gels 1 and 2, and D). Furthermore, in a strain deleted for MVB12 we observed a shift in the size range of the ESCRT-I subunits Vps23 and Vps37-HA to ∼170 kDa (Figure 3A, gels 3 and 4, and D). Together, these protein interaction and gel filtration data strongly suggest that Mvb12 is a subunit of the ESCRT-I complex.

Gel filtration experiments suggest that Mvb12 associates with ESCRT-I via the subunit Vps37. (A) Cytosol from different yeast strains expressing either MVB12-HA (pMB240) or HA-MVB12 (pMB301) was separated by gel filtration (Sephacryl S300), and the resulting fractions were analyzed by Western blot for the presence of Mvb12-HA, HA-Mvb12, and Vps23. Gels 8 and 9 show the results of the gel filtration analysis of wild type expressing MVB12-HA after 2 h of cycloheximide treatment. (B) Detailed analysis of gel filtration experiments from Figure 3A. Bars above the gels indicate the pooled fractions used in Figure 3A gels 2, 3, and 8. (C) Total cell extracts were prepared from yeast expressing MVB12-HA at different times after treatment with cycloheximide (0, 1, and 2 h) and analyzed by Western blot using anti-HA and anti-Vps23 antibodies. After 2 h of cycloheximide treatment the cells were prepared for gel filtration analysis and the extract loaded onto the column (L) was analyzed by Western blot. The result of this gel filtration analysis is shown in A (gels 8 and 9). (D) The table summarizes the apparent molecular masses of the analyzed proteins relative to the standard proteins thyroglobulin (670 kDa), ferritin (440 kDa), aldolase (158 kDa), albumin (67 kDa), ovalbumin (44 kDa) and myoglobin (17 kDa) (see A).

Mvb12 Seems to Associate with ESCRT-I via Vps37

Deletion of VPS37 results in the formation of a stable ESCRT-I subcomplex of ∼125 kDa that has been shown to contain Vps23 and Vps28 (Figure 3A, gel 6; Kostelansky et al., 2006; Teo et al., 2006). As shown in Figure 2A, Mvb12 is unstable in this mutant background, suggesting that Mvb12 associates with ESCRT-I via Vps37. To detect Mvb12 in _vps37_Δ, we transformed the mutant strain with a plasmid expressing from the CPS1 promoter a functional N-terminal fusion of MVB12 with the HA tag (Supplemental Figure 1). The resulting fusion protein HA-Mvb12, like Mvb12-HA, is unstable in _vps37_Δ, but because of the strong CPS1 promoter HA-Mvb12 is detectable by Western blot. Cell extract from _vps37_Δ expressing HA-MVB12 was separated by gel filtration, and Western blot analysis identified small amounts of HA-Mvb12 in the size range of 20 kDa, consistent with the model that Mvb12 requires Vps37 to associate with ESCRT-I (Figure 3A, gel 5). The deletion of VPS23 resulted in lower but detectable protein levels of Vps37-HA and Mvb12-HA. Analysis of this mutant strain by gel filtration indicated that the remaining pool Vps37-HA and Mvb12-HA form a complex of ∼195 kDa in absence of Vps23 (Figure 3A, gel 7, and D). Our result does not exclude the possibility that this 195-kDa complex contains Vps28. However, the crystal structure analysis of ESCRT-I has shown that Vps23 acts as a linker between Vps28 and Vps37, which argues against the presence of Vps28 in 195-kDa complex (Kostelansky et al., 2006; Teo et al., 2006). Together, the results suggested that Vps37 and Mvb12 form a subcomplex.

Further support for the Vps37–Mvb12 interaction was obtained by fluorescence microscopy. To visualize the low levels of Mvb12 in different ESCRT-I mutants, we transformed the mutant strains with a high-copy 2μ plasmid containing MVB12-GFP (Figure 4). In cells deleted for VPS4 the increased levels of Mvb12-GFP localized to class E compartments (Figure 4, arrows) similar to the localization found for Mvb12-GFP expressed at normal levels (Figure 1A). Previous studies have shown that the localization of ESCRT-I to MVBs is mediated by the interaction of Vps23 with Vps27 (Bache et al., 2003; Katzmann et al., 2003). As expected from these studies we found that Mvb12 required Vps23 for proper localization to the endosome (Figure 4, _vps4_Δ_vps23_Δ). The deletion of VPS28 did not change the endosomal localization of Mvb12-GFP, indicating that Vps28 is not essential for membrane association of Mvb12 (Figure 4, _vps4_Δ_vps28_Δ). In contrast, loss of Vps37 resulted in the redistribution of Mvb12-GFP to the cytoplasm, consistent with our model that Mvb12 binds via Vps37 to ESCRT-I (Figure 4, _vps4_Δ_vps37_Δ).

Mvb12 requires Vps23 and Vps37 to localize to endosomes. Fluorescent microscopy of Mvb12-GFP overexpressed from a high-copy plasmid in different mutant backgrounds (arrows indicate class E compartments).

Mvb12 Is an Unstable ESCRT-I Subunit

We determined the stability of Mvb12-HA in wild-type cells using the translational inhibitor cycloheximide. Cycloheximide was added to yeast cultures and samples were taken immediately (t = 0) and at 1, 2, and 3 h (t = 1 h, 2 h, and 3 h). The cell extracts were analyzed for the presence of Mvb12-HA, Vps23, and Vps37-HA by Western blot (Figure 5). In wild-type cells, Mvb12-HA was rapidly degraded with a half-life of <1 h. Vps37-HA was more stable, having a half-life of 2 to 3 h. In contrast, the amount of Vps23 only changed marginally over the time period of 3 h. Together, the data suggested that compared with other subunits of ESCRT-I, Mvb12 is relatively unstable.

Rapid turnover of Mvb12 by the proteasome depends on a functional MVB pathway. The translation inhibitor cycloheximide (50 mg/l) was added to yeast cultures (t = 0), and samples were taken after addition of the drug at the indicated time points. Cells were lysed using glass beads, and the resulting extracts were analyzed by Western blot for the presence of Vps23, Mvb12-HA, and Vps37-HA.

To determine which pathway is responsible for the degradation of Mvb12, we tested the stability of Mvb12-HA in the proteasomal mutant strain cim3-1 and a strain deleted for two major vacuolar peptidases Pep4 and Prb1. These experiments demonstrated that Mvb12-HA is a stable protein in the proteasomal mutant, but only a minor Mvb12-HA stabilization is observed in the strain deleted for vacuolar peptidases (Figure 5), indicating that the rapid degradation of Mvb12 is primarily dependent on the proteasomal system rather than the vacuole. This result is consistent with the observations by fluorescence microscopy that showed no detectable levels of Mvb12-GFP in the vacuolar lumen (Figure 1A).

The rapid turnover of Mvb12 suggested that ESCRT-I repeatedly exchanges the Mvb12 subunit with a newly synthesized copy. Therefore we expected that blocking translation should result in a shift in size of ESCRT-I over time from 350 to ∼170 kDa, the size of ESCRT-I in _mvb12_Δ. To test this prediction, we prepared a cell extract of a wild-type strain expressing Mvb12-HA after 2 h of cycloheximide treatment and analyzed this extract by gel filtration. The Western blot in Figure 3C illustrates that the sample loaded onto the gel filtration column contained at least fivefold less Mvb12-HA than the nontreated sample. In contrast, the amount of Vps23 remained stable during this procedure. The result of the gel filtration analysis indicated that during the cycloheximide treatment a large portion of ESCRT-I shifted to a smaller molecular mass (Figure 3A, gel 8, and B, gel 12). Vps23 showed a very broad size distribution (170–350 kDa), which is consistent with a mixture of complete ESCRT-I and ESCRT-I missing Mvb12 (Figure 3B). The remaining pool of Mvb12-HA shifted to ∼260 kDa after blocking translation for 2 h (Figure 3A, gel 9). It is expected that Mvb12-HA is only present in the larger forms of ESCRT-I that still contain this subunit. However, that Mvb12-HA shifts in size at all suggests that there is more than one subunit of Mvb12 present in each 350-kDa ESCRT-I complex.

The data so far indicated that the ESCRT-I subunit Mvb12 is degraded by the ubiquitin-proteasomal system and that the rate of degradation is faster than observed for other subunits of the ESCRT-I complex. The question remained whether the degradation of Mvb12 is connected to the function of ESCRT-I. Therefore, we analyzed the stability of Mvb12-HA in mutants that impair either the recruitment of ESCRT-I to MVBs (_vps27_Δ) or the dissociation of ESCRT-I from the endosomal membrane (_vps4_Δ). The result shown in Figure 5 indicated that both type of mutants resulted in stabilization of Mvb12-HA, suggesting that the instability of Mvb12 is a function of the transient localization of ESCRT-I to MVBs.

Mvb12 Is Required for Efficient MVB Sorting

To study potential MVB trafficking defects, we transformed the MVB12 deletion strain with plasmids containing GFP-gene fusions encoding for either CPS fused to GFP (GFP-CPS), Sna3-GFP, Ste2-GFP, or Ste3-GFP. Newly synthesized CPS and Sna3 are transmembrane proteins that are transported through the secretory pathway and via MVBs to the lumen of the vacuole. In contrast to CPS, the sorting of Sna3 into the MVB pathway seems to be independent of ubiquitination (Odorizzi et al., 1998; Katzmann et al., 2001; Reggiori and Pelham, 2001; Yeo et al., 2003). Ste2 and Ste3 are surface receptors that are endocytosed and delivered via the MVB pathway to the vacuolar lumen (Odorizzi et al., 1998; Chen and Davis, 2002; Shaw et al., 2003). Therefore, in wild-type cells the GFP fluorescence of GFP-CPS and Ste2-GFP is observed by microscopy in the lumen of the vacuole (Figure 6A and Supplemental Figure 2). In contrast, cells defective in the MVB pathway, such as the ESCRT-I mutant _vps23_Δ, accumulate GFP-CPS, Sna3-GFP, Ste2-GFP, and Ste3-GFP in aberrant endosomes (class E compartments; Figure 6A, arrows, and Supplemental Figure 2) and mislocalize cargo proteins to the vacuolar membrane (Figure 6A and Supplemental Figure 2). Surprisingly, the analysis of the _mvb12_Δ trafficking phenotype indicated that these mutant cells have a less severe defect in the MVB sorting pathway (Figure 6A and Supplemental Figure 2). Unlike in _vps23_Δ, _vps28_Δ, and _vps37_Δ, in MVB12 deletion mutant cells GFP-CPS and Sna3-GFP do not accumulate in class E compartments, and a portion of GFP-CPS and Sna3-GFP is properly localized to the vacuolar lumen. Ste2-GFP and Ste3-GFP showed almost normal trafficking in _mvb12_Δ with only a slight accumulation of the cargo in late endosomes. These data might suggest that the deletion of MVB12 affects the MVB trafficking of biosynthetic cargo more severely then the trafficking of cargo from the endocytic pathway.

The deletion of MVB12 results in a mild MVB trafficking phenotype. (A) Strains expressing either GFP-CPS (pGO45) or Ste2-GFP (pCS24) were analyzed by fluorescence microscopy. White arrows indicate class E compartments. ESCRT-II was overexpressed by transforming the cells with the high-copy plasmid pMB175 (2μ ESCRT-II). (B) Invertase plate assay of wild type (BHY10), _vps23_Δ (EEY5-2), and _mvb12_Δ (MCY9).

A hallmark of class E vps mutants is a partial secretion of vacuolar enzymes caused by the inefficient recycling of the sorting receptor Vps10 (Raymond et al., 1992; Cereghino et al., 1995; Cooper and Stevens, 1996). This secretion phenotype can be visualized using the reporter carboxypeptidase Y (CPY)-Invertase combined with a colorimetric plate assay (Paravicini et al., 1992). Wild-type cells efficiently sort CPY-Invertase to the vacuole and therefore remain white on the assay plate (Figure 6B). Class E vps mutants, such as _vps23_Δ, secrete CPY-Invertase, which results in the colorimetric reaction. In contrast, the MVB12 deletion strain remained white, indicating that CPY-Invertase was trafficked efficiently to the vacuole in this mutant strain (Figure 6B). Together, the trafficking studies indicated that Mvb12 is not an essential factor for MVB sorting; thus, unlike the other ESCRT-I subunits, Mvb12 does not belong to the group of class E Vps proteins. However, Mvb12 is necessary for the efficient transport of cargo proteins to the lumen of the vacuole.

One possible explanation for the partial MVB trafficking phenotype is that Mvb12 function might be redundant. Therefore, we tested whether overexpression of the ESCRT-I genes VPS23, VPS28, and VPS37 could suppress the GFP-CPS trafficking phenotype of an _mvb12_Δ strain. The data of these studies showed no suppression of the GFP-CPS mislocalization (data not shown), suggesting that high levels of the remaining ESCRT-I complex cannot replace the function of Mvb12. However, in similar studies we found that overexpression of the ESCRT-II genes fully suppressed all tested trafficking defects in the _mvb12_Δ mutant (Figure 6A and Supplemental Figure 2). This result is consistent with previous studies that showed suppression of ESCRT-I mutant phenotypes by overexpression of ESCRT-II (Katzmann et al., 2001) and therefore is further support of the idea that Mvb12 functions as part of ESCRT-I. However, in contrast to the complete rescue of the _mvb12_Δ phenotype, the mutant phenotype of other ESCRT-I subunits is only partially suppressed by ESCRT-II overexpression (Katzmann et al., 2001).

Deletion of MVB12 Affects the Dissociation of ESCRT-I from MVBs

Our studies have shown that deletion of MVB12 results in the formation of a stable ∼170-kDa ESCRT-I complex that is at least partially able to function in the sorting of MVB cargo. To determine whether lack of Mvb12 effects the transient association of ESCRT-I with the endosomal membrane, we studied the localization of the ESCRT-I subunit Vps23 in different strain backgrounds by using a chromosomally integrated functional GFP-tagged VPS23 fusion protein (Katzmann et al., 2001). As shown previously, Vps23-GFP localizes to the cytoplasm and endosomal compartments in wild-type cells and accumulates on class E compartments in cells deleted for VPS4 (Figure 7A). Interestingly, we found that deleting MVB12 caused a portion of Vps23-GFP to be transported into the vacuolar lumen (Figure 7A). This transport required the MVB trafficking pathway because the deletion of VPS4 in _mvb12_Δ prevented the vacuolar localization of Vps23-GFP and resulted in endosomal accumulation of the fusion protein similar as observed in the _vps4_Δ strain. Deletion of MVB12 did not effect Vps36-GFP localization (Figure 7A) but caused GFP-Vps27 to accumulate on large structures adjacent to the vacuole that most likely are late endosomal compartments (Figure 7A). No accumulation of GFP-Vps27 was observed inside the vacuole. Similar redistribution of Vps27 has been observed in other ESCRT-I mutants (Katzmann et al., 2003). Interestingly, a mutation in the UEV domain of Vps23 (methionine 85 to threonine, M85T) that previously had been shown to impair ubiquitin binding (Katzmann et al., 2001) resulted in loss of vacuolar localization of Vps23-GFP in _mvb12_Δ cells (Figure 7A). Because these cells expressed wild-type Vps23 in addition to Vps23(M85T)-GFP, the observed effect was not due to a defect in MVB trafficking. Together, these results suggest that binding to ubiquitinated cargo is necessary for the delivery of Vps23-GFP into the vacuolar lumen of _mvb12_Δ cells.

We have shown that overexpression of ESCRT-II suppresses the trafficking defect of an MVB12 deletion (Figure 6A). In contrast, we found that overexpression of ESCRT-II does not suppress the vacuolar localization of Vps23-GFP, but instead increases the severity of this mislocalization phenotype (Figure 7A). This result suggested that the increased efficiency of MVB sorting in the ESCRT-II overexpression strain also results in an increased delivery of ESCRT-I to the vacuolar lumen. The findings of these microscopy studies are further supported by anti-GFP Western blot analysis of extracts from wild-type, _mvb12_Δ-, and _mvb12_Δ-overexpressing ESCRT-II cells, each containing the chromosomally integrated VPS23-GFP fusion (Figure 7B). In comparison with wild type cells, _mvb12_Δ cells contained an increased amount of truncated Vps23-GFP, which was even further increased when ESCRT-II was overexpressed. This increase in truncated Vps23-GFP mirrors the increased GFP signal observed by microscopy in the lumen of the vacuole, suggesting that the truncated Vps23-GFP species represent degradation intermediates of vacuolar Vps23-GFP. As expected, the most prevalent degradation products are in the size range of GFP protein, which is generally a very stable protein (∼25 kDa). Furthermore, Western blot analysis showed that the amount of full-length Vps23-GFP did not change dramatically in cells deleted for MVB12, suggesting that the delivery of Vps23-GFP to the vacuole is inefficient and does not result in the depletion of cytoplasmic Vps23-GFP (Figure 7B). Because GFP is a very stable protein that is able to persist in the vacuole for an extended time period, the strong GFP signal observed by microscopy in the vacuolar lumen of _mvb12_Δ expressing Vps23-GFP (Figure 7A) is probably due to GFP accumulation over a long period rather than rapid vacuolar delivery of Vps23-GFP.

DISCUSSION

The ESCRT-I protein complex is part of the protein machinery that executes endosomal (MVB) sorting of ubiquitinated cargo and formation of MVB vesicles (for review, see Hurley and Emr, 2006). ESCRT-I localizes to the cytoplasm and is transiently recruited to the endosomal membrane. This recruitment step is mediated by the interaction of ESCRT-I with endosomal localized Vps27 (Bilodeau et al., 2003; Katzmann et al., 2003). On the endosome, ESCRT-I and Vps27 bind to monoubiquitinated cargo to initiate their sorting into MVB vesicles (Katzmann et al., 2001; Bilodeau et al., 2002). Subsequently, ESCRT-I interacts with ESCRT-II, a protein complex that functions in the formation of ESCRT-III (Babst et al., 2002a,b; Teo et al., 2006). Finally, ESCRT-I is released from the MVB membrane and recycled for further rounds of sorting.

In this study, we have found evidence for the presence of a forth subunit of yeast ESCRT-I, called Mvb12 (SGD: YGR206w). Immunoprecipitation and affinity purification experiments demonstrated a physical interaction between Mvb12 and ESCRT-I. Gel filtration analysis of yeast cell extract indicate that Mvb12 has a native molecular mass similar to that of ESCRT-I (350 kDa) and that deletion of MVB12 causes a shift in size of ESCRT-I from 350 to ∼170 kDa. Furthermore, we observed that Mvb12 is unstable in cells lacking any of the ESCRT-I subunits. Finally, Mvb12 colocalizes with ESCRT-I and the subcellular distribution of Mvb12 is consistent with the published transient MVB association of ESCRT-I (Katzmann et al., 2001). Previous studies have shown that native ESCRT-I has an apparent molecular mass of 350 kDa and is composed of the class E Vps proteins Vps23, Vps28, and Vps37 (Babst et al., 2000; Katzmann et al., 2001). Recent crystal structure analysis suggested that these three subunits form a trimeric complex with a 1:1:1 stoichiometry (Kostelansky et al., 2006; Teo et al., 2006). Because only a substructure of ESCRT-I has been analyzed by crystallography, the exact composition of the in vivo observed 350-kDa complex remains unknown.

The gel filtration data of different ESCRT-I mutants are consistent with a model in which Mvb12 interacts with ESCRT-I via the Vps37 subunit. However, the number of Mvb12 subunits present in each 350-kDa ESCRT-I complex is not clear. Deletion of MVB12 resulted in a shift in size of ESCRT-I from ∼350 to ∼170 kDa, suggesting that Mvb12 might promote dimerization of a Vps23-Vps28-Vps37 complex or that ESCRT-I contains at least 10 Mvb12 subunits. Finally, we cannot exclude the possibility that ESCRT-I contains additional not yet identified subunits that require Mvb12 to interact with ESCRT-I.

Both gel filtration analysis and immunoprecipitation experiments indicate that Mvb12 is tightly associated with ESCRT-I. Surprisingly, in wild-type cells we observed a rapid proteasome-dependent turnover of Mvb12, resulting in a half-life of Mvb12 that is much shorter then the half-life of other ESCRT-I subunits. As a consequence, when new protein synthesis was blocked, the ESCRT-I complex gradually shifted in size from ∼350 to ∼170 kDa. This result indicated that during the lifetime of an ESCRT-I complex, Mvb12 is repeatedly degraded and replaced by newly synthesized Mvb12 subunits. Interestingly, Mvb12 degradation was found to be dependent on the transient localization of ESCRT-I to the endosome. Loss of either the ESCRT-I recruiting factor Vps27 or the recycling factor Vps4 resulted in the stabilization of Mvb12. This result suggested that the rapid degradation of Mvb12 is a consequence of ESCRT-I function on the endosome.

Loss of function of Vps23, Vps28, or Vps37 resulted in a complete block of the MVB pathway (Babst et al., 2000; Katzmann et al., 2001). In contrast, deletion of MVB12 only partially affects the transport of MVB cargoes to the vacuolar lumen, indicating that Mvb12 is not an essential factor for ESCRT-I function. We found that the trafficking defects in _mvb12_Δ are more severe for cargo of the biosynthetic pathway (e.g., CPS and Sna3) than for endocytic cargo (e.g., Ste2 and Ste3). It is not clear whether this difference indicates the presence of two partially separate MVB pathways, a biosynthetic and an endocytic pathway, or whether it reflects cargo specificity of Mvb12 function.

In _mvb12_Δ cells, ESCRT-I localizes to endosomal compartments and the lumen of the vacuole, indicating that Mvb12 is not required for the recruitment of ESCRT-I to MVBs but instead seems to be play a role in the dissociation of ESCRT-I from the endosome after the function of ESCRT-I in cargo sorting is completed. However, the delivery of ESCRT-I into the MVB pathway in _MVB12_-deleted cells is not efficient enough to deplete the cells of ESCRT-I and is not the cause of the MVB-trafficking defects observed in _mvb12_Δ. Consistent with the role of ESCRT-II in cargo sorting downstream of ESCRT-I, we found that overexpression of ESCRT-II suppressed the trafficking phenotype of _mvb12_Δ. In contrast, the mislocalization of mutant ESCRT-I to the vacuolar lumen was enhanced by the overexpression of ESCRT-II, suggesting that ESCRT-II overexpression increased MVB trafficking not only of cargo but also of ESCRT-I.

Deletion of MVB12 caused ESCRT-I to be targeted to the vacuolar lumen, but it did not result in vacuolar localization of Vps27 or ESCRT-II. This observation suggests that the defect in ESCRT-I recycling in _mvb12_Δ is not a result of ESCRT-I not being able to break the interactions with other ESCRT machinery, but rather it is caused by maintaining (prolonging) the interaction with ubiquitinated cargo. This model is supported by the observation that a ubiquitin-binding mutant of Vps23 does not get delivered to the vacuolar lumen in an MVB12 deletion strain. Several proteins of the ESCRT machinery contain ubiquitin-binding domains that recognize a similar ubiquitin surface (for review, see Hurley and Emr, 2006). It has therefore been speculated that ubiquitinated cargo might be handed over from one ubiquitin-binding protein complex to the next (e.g., from ESCRT-I to ESCRT-II). However, the _mvb12_Δ phenotype suggests that cargo is sorted into the MVB pathway even when it remains bound to ESCRT-I. Furthermore, we observed that overexpression of ESCRT-II in _mvb12_Δ results in increased transport of both cargo and ESCRT-I to the vacuolar lumen, indicating that ESCRT-II does not compete with ESCRT-I for cargo. Together, our data are most consistent with a model in which sorting of ESCRT-I-bound cargo into MVB vesicles does not require the release of cargo from ESCRT-I and subsequent interaction with ESCRT-II.

Although Mvb12 plays an important role in the function of yeast ESCRT-I, the protein sequence of Mvb12 is evolutionarily poorly conserved. As a consequence, convincing Mvb12 homologues are only found among other fungal species. However, lack of sequence homology does not rule out the possibility that functional homologues of Mvb12 also exist in higher eukaryotes. Similar to Mvb12, very little sequence conservation is observed between Vps37 proteins from different species. Considering the high structural and functional conservation of other parts of the ESCRT machinery it is unlikely that this sequence diversity among different species reflects a difference in the function of ESCRT-I.

The current MVB sorting model suggests that after recruitment to the MVB, ESCRT-I binds to ubiquitinated cargo and to ESCRT-II, which initiates the formation of ESCRT-III. At some point, either during ESCRT-III formation or the dissociation of ESCRT-III by Vps4, the ESCRT-I complex dissociates from the ESCRT machinery and is released to the cytoplasm for further rounds of sorting. Our data suggest that loss of Mvb12 affects a step in ESCRT-I function after the endosomal recruitment and binding of cargo. This defect results in inefficient cargo sorting and in inefficient dissociation of ESCRT-I from the MVB. One possible function for Mvb12 could be to interact with other components of the ESCRT machinery, thereby regulating and coordinating the transient interactions of ESCRT-I with cargo and machinery. Without Mvb12, ESCRT-I might not properly activate the downstream ESCRT-II complex, which might affect the release of ESCRT-I from the ubiquitinated cargo. The rapid degradation of Mvb12 observed in wild-type cells might be a result of the ESCRT-I release from the MVB. During the release reaction, Mvb12 might dissociate from the complex and be degraded by the proteasomal system. It will be interesting to test whether the observed modification of endosomal Mvb12 plays a role in regulating the interaction between ESCRT-I, ubiquitinated cargo, and other parts of the ESCRT machinery.

Supplementary Material

ACKNOWLEDGMENTS

We thank Tamara Darsow for critical reading of the manuscript. We thank Dan Gottschling for providing the cim3-1 strain. This work has been supported by grant R01 GM-074171-01 A1 from the National Institute of Health and by a Funding Incentive Seed Grant (project 51003083) from the University of Utah.

Abbreviations used:

Footnotes

REFERENCES

• Babst M. A protein's final ESCRT. Traffic. 2005;6:2–9. [PubMed] [Google Scholar]
• Babst M., Katzmann D. J., Estepa-Sabal E. J., Meerloo T., Emr S. D. Escrt-III: an endosome-associated heterooligomeric protein complex required for mvb sorting. Dev. Cell. 2002a;3:271–282. [PubMed] [Google Scholar]
• Babst M., Katzmann D. J., Snyder W. B., Wendland B., Emr S. D. Endosome-associated complex, ESCRT-II, recruits transport machinery for protein sorting at the multivesicular body. Dev. Cell. 2002b;3:283–289. [PubMed] [Google Scholar]
• Babst M., Odorizzi G., Estepa E. J., Emr S. D. Mammalian tumor susceptibility gene 101 (TSG101) and the yeast homologue, Vps23p, both function in late endosomal trafficking. Traffic. 2000;1:248–258. [PubMed] [Google Scholar]
• Babst M., Sato T. K., Banta L. M., Emr S. D. Endosomal transport function in yeast requires a novel AAA-type ATPase, Vps4p. EMBO J. 1997;16:1820–1831. [PMC free article] [PubMed] [Google Scholar]
• Babst M., Wendland B., Estepa E. J., Emr S. D. The Vps4p AAA ATPase regulates membrane association of a Vps protein complex required for normal endosome function. EMBO J. 1998;17:2982–2993. [PMC free article] [PubMed] [Google Scholar]
• Bache K. G., Brech A., Mehlum A., Stenmark H. Hrs regulates multivesicular body formation via ESCRT recruitment to endosomes. J. Cell Biol. 2003;162:435–442. [PMC free article] [PubMed] [Google Scholar]
• Bilodeau P. S., Urbanowski J. L., Winistorfer S. C., Piper R. C. The Vps27p Hse1p complex binds ubiquitin and mediates endosomal protein sorting. Nat. Cell Biol. 2002;4:534–539. [PubMed] [Google Scholar]
• Bilodeau P. S., Winistorfer S. C., Kearney W. R., Robertson A. D., Piper R. C. Vps27-Hse1 and ESCRT-I complexes cooperate to increase efficiency of sorting ubiquitinated proteins at the endosome. J. Cell Biol. 2003;163:237–243. [PMC free article] [PubMed] [Google Scholar]
• Cereghino J. L., Marcusson E. G., Emr S. D. The cytoplasmic tail domain of the vacuolar sorting receptor Vps10p and a subset of VPS gene products regulate receptor stability, function and localization. Mol. Biol. Cell. 1995;6:1089–1102. [PMC free article] [PubMed] [Google Scholar]
• Ceresa B. P., Schmid S. L. Regulation of signal transduction by endocytosis. Curr. Opin. Cell Biol. 2000;12:204–210. [PubMed] [Google Scholar]
• Chen L., Davis N. G. Ubiquitin-independent entry into the yeast recycling pathway. Traffic. 2002;3:110–123. [PubMed] [Google Scholar]
• Christianson T. W., Sikorski R. S., Dante M., Shero J. H., Hieter P. Multifunctional yeast high-copy-number shuttle vectors. Gene. 1992;110:119–122. [PubMed] [Google Scholar]
• Cooper A. A., Stevens T. H. Vps10p cycles between the late-Golgi and prevacuolar compartments in its function as the sorting receptor for multiple yeast vacuolar hydrolases. J. Cell Biol. 1996;133:529–541. [PMC free article] [PubMed] [Google Scholar]
• Di Fiore P. P., Gill G. N. Endocytosis and mitogenic signaling. Curr. Opin. Cell Biol. 1999;11:483–488. [PubMed] [Google Scholar]
• Gruenberg J., Stenmark H. The biogenesis of multivesicular endosomes. Nat. Rev. Mol. Cell Biol. 2004;5:317–323. [PubMed] [Google Scholar]
• Hirano S., Kawasaki M., Ura H., Kato R., Raiborg C., Stenmark H., Wakatsuki S. Double-sided ubiquitin binding of Hrs-UIM in endosomal protein sorting. Nat. Struct. Mol. Biol. 2006;13:272–277. [PubMed] [Google Scholar]
• Huh W. K., Falvo J. V., Gerke L. C., Carroll A. S., Howson R. W., Weissman J. S., O'Shea E. K. Global analysis of protein localization in budding yeast. Nature. 2003;425:686–691. [PubMed] [Google Scholar]
• Hurley J. H., Emr S. D. The ESCRT complexes: structure and mechanism of a membrane-trafficking network. Annu. Rev. Biophys. Biomol. Struct. 2006;35:277–298. [PMC free article] [PubMed] [Google Scholar]
• Ito H., Fukuda Y., Murata K., Kimura A. Transformation of intact yeast cells treated with alkali cations. J. Bacteriol. 1983;153:163–168. [PMC free article] [PubMed] [Google Scholar]
• Katzmann D. J., Babst M., Emr S. D. Ubiquitin-dependent sorting into the multivesicular body pathway requires the function of a conserved endosomal protein sorting complex, ESCRT-I. Cell. 2001;106:145–155. [PubMed] [Google Scholar]
• Katzmann D. J., Odorizzi G., Emr S. D. Receptor downregulation and multivesicular-body sorting. Nat. Rev. Mol. Cell Biol. 2002;3:893–905. [PubMed] [Google Scholar]
• Katzmann D. J., Stefan C. J., Babst M., Emr S. D. Vps27 recruits ESCRT machinery to endosomes during MVB sorting. J. Cell Biol. 2003;162:413–423. [PMC free article] [PubMed] [Google Scholar]
• Kostelansky M. S., Sun J., Lee S., Kim J., Ghirlando R., Hierro A., Emr S. D., Hurley J. H. Structural and functional organization of the ESCRT-I trafficking complex. Cell. 2006;125:113–126. [PMC free article] [PubMed] [Google Scholar]
• Luhtala N., Odorizzi G. Bro1 coordinates deubiquitination in the multivesicular body pathway by recruiting Doa4 to endosomes. J. Cell Biol. 2004;166:717–729. [PMC free article] [PubMed] [Google Scholar]
• Murk J. L., Stoorvogel W., Kleijmeer M. J., Geuze H. J. The plasticity of multivesicular bodies and the regulation of antigen presentation. Semin. Cell Dev. Biol. 2002;13:303–311. [PubMed] [Google Scholar]
• Odorizzi G., Babst M., Emr S. D. Fab1p PtdIns(3)P 5-kinase function essential for protein sorting in the multivesicular body. Cell. 1998;95:847–858. [PubMed] [Google Scholar]
• Paravicini G., Horazdovsky B. F., Emr S. D. Alternative pathways for the sorting of soluble vacuolar proteins in yeast: a vps35 null mutant missorts and secretes only a subset of vacuolar hydrolases. Mol. Biol. Cell. 1992;3:415–427. [PMC free article] [PubMed] [Google Scholar]
• Pornillos O., Alam S. L., Rich R. L., Myszka D. G., Davis D. R., Sundquist W. I. Structure and functional interactions of the Tsg101 UEV domain. EMBO J. 2002;21:2397–2406. [PMC free article] [PubMed] [Google Scholar]
• Raymond C. K., Howald-Stevenson I., Vater C. A., Stevens T. H. Morphological classification of the yeast vacuolar protein sorting mutants: evidence for a prevacuolar compartment in class E vps mutants. Mol. Biol. Cell. 1992;3:1389–1402. [PMC free article] [PubMed] [Google Scholar]
• Reggiori F., Pelham H. R. Sorting of proteins into multivesicular bodies: ubiquitin-dependent and -independent targeting. EMBO J. 2001;20:5176–5186. [PMC free article] [PubMed] [Google Scholar]
• Robinson J. S., Klionsky D. J., Banta L. M., Emr S. D. Protein sorting in Saccharomyces cerevisiae: isolation of mutants defective in the delivery and processing of multiple vacuolar hydrolases. Mol. Cell Biol. 1988;8:4936–4948. [PMC free article] [PubMed] [Google Scholar]
• Rothman J. H., Stevens T. H. Protein sorting in yeast: mutants defective in vacuole biogenesis mislocalize vacuolar proteins into the late secretory pathway. Cell. 1986;47:1041–1051. [PubMed] [Google Scholar]
• Sambrook J., Fritsch E. F., Maniatis T. A Laboratory Manual. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1989. Molecular Cloning. [Google Scholar]
• Shaw J. D., Hama H., Sohrabi F., DeWald D. B., Wendland B. PtdIns(3,5)P2 is required for delivery of endocytic cargo into the multivesicular body. Traffic. 2003;4:479–490. [PubMed] [Google Scholar]
• Sherman F., Fink G. R., Lawrence L. W. Methods in Yeast Genetics: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1979. [Google Scholar]
• Shih S. C., Katzmann D. J., Schnell J. D., Sutanto M., Emr S. D., Hicke L. Epsins and Vps27p/Hrs contain ubiquitin-binding domains that function in receptor endocytosis. Nat. Cell Biol. 2002;4:389–393. [PubMed] [Google Scholar]
• Sundquist W. I., Schubert H. L., Kelly B. N., Hill G. C., Holton J. M., Hill C. P. Ubiquitin recognition by the human TSG101 protein. Mol. Cell. 2004;13:783–789. [PubMed] [Google Scholar]
• Swanson K. A., Kang R. S., Stamenova S. D., Hicke L., Radhakrishnan I. Solution structure of Vps27 UIM-ubiquitin complex important for endosomal sorting and receptor downregulation. EMBO J. 2003;22:4597–4606. [PMC free article] [PubMed] [Google Scholar]
• Teo H., Gill D. J., Sun J., Perisic O., Veprintsev D. B., Vallis Y., Emr S. D., Williams R. L. ESCRT-I core and ESCRT-II GLUE domain structures reveal role for GLUE in linking to ESCRT-I and membranes. Cell. 2006;125:99–111. [PubMed] [Google Scholar]
• Teo H., Veprintsev D. B., Williams R. L. Structural insights into endosomal sorting complex required for transport (ESCRT-I) recognition of ubiquitinated proteins. J. Biol. Chem. 2004;279:28689–28696. [PubMed] [Google Scholar]
• Thompson B. J., Mathieu J., Sung H. H., Loeser E., Rorth P., Cohen S. M. Tumor suppressor properties of the ESCRT-II complex component Vps25 in Drosophila. Dev. Cell. 2005;9:711–720. [PubMed] [Google Scholar]
• Vaccari T., Bilder D. The Drosophila tumor suppressor vps25 prevents nonautonomous overproliferation by regulating notch trafficking. Dev. Cell. 2005;9:687–698. [PubMed] [Google Scholar]
• Yeo S. C., et al. Vps20p and Vta1p interact with Vps4p and function in multivesicular body sorting and endosomal transport in Saccharomyces cerevisiae. J. Cell Sci. 2003;116:3957–3970. [PubMed] [Google Scholar]

Articles from Molecular Biology of the Cell are provided here courtesy of American Society for Cell Biology