Vanderbilt Institute of Chemical Biology



Discovery at the VICB






Targeting Proteins for Endocytosis


By: Carol A. Rouzer, VICB Communications
Published: January 24, 2019


New data reveal that Ppz phosphatases help the E3 ubiquitin ligase Rsp5 target plasma membrane proteins for internalization in yeast.


An important means by which cells modulate the activity of plasma membrane proteins (e.g., receptors, transporters, and ion channels) is through endocytosis, which simply removes the protein from the cell surface for ultimate recycling or degradation. The process of endocytosis requires at least 60 distinct proteins that carry out a multitude of functions. One of these is to identify and capture cargo proteins to be included in the endocytic vesicle. It is not entirely clear how all cargo proteins are identified and labeled, but one mechanism involves ubiquitylation, which targets a protein for internalization. In yeast, the primary E3 ubiquitin ligase responsible for endocytic targeting is Rsp5. However, a key question surrounds the mechanism by which Rsp5 identifies proteins that should be internalized. Now, Vanderbilt Basic Sciences investigator Jason MacGurn and his laboratory report a new role for Ppz phosphatases and the adaptor protein Art1 in modulating the Rsp5-mediated endocytic removal of key amino acid transporters in yeast [S. Lee et al., (2019) J. Cell Biol., published January 4, DOI: 10.1083/jcb.201712144].

Rsp5 comprises a typical C-terminal HECT E3 ubiquitin ligase domain, an N-terminal C2 domain, and three tandem WW domains that are used as a scaffold to interact with other proteins. Among the proteins that interact with Rsp5's WW domains are the arrestin-related Rsp5 adaptors (ARTs). ARTs contain a WW domain-binding PY motif and an arrestin fold domain that mediates selection of proteins to be ubiquitylated. Current evidence suggests that substrate-mediated changes in the conformation of a target protein (such as occur when a ligand binds to its receptor) are responsible for triggering activation of an ART protein, thereby enabling it to guide Rsp5 to the target under highly regulated and physiologically appropriate conditions. Among the plasma membrane proteins that are subject to Rsp5-dependent ubiquitylation (and subsequent internalization) are the methionine transporter (Mup1) and the arginine transporter (Can1). In both cases, it is the presence of the amino acid to be transported (methionine or arginine, respectively) that triggers endocytosis of the transporter. And in both cases, Rsp5 is directed to the proteins by Art1 following its activation by dephosphorylation. The identities of the phosphatases that carry out Art1 activation are as yet unclear, however, and it was this question that led the MacGurn lab on its quest.

The researchers began by establishing an assay by which they could monitor endocytosis of Mup1. They expressed Mup1 fused to green fluorescent protein (GFP) in yeast cells to enable them to visualize the protein, which was predominantly localized at the plasma membrane in the resting state (no methionine). When they added methionine to the culture medium, however, the protein quickly became internalized into small endocytic vesicles, and eventually could be seen in larger intracellular vacuoles (Figure 1, top). The investigators verified the vacuolar location of the Mup1-GFP signal by co-expressing the protein Vph1 fused to the red fluorescent protein m-Cherry. Vph1 is a vacuolar membrane protein (Figure 1, bottom). The researchers confirmed a role for Art1 in Mup1 internalization by demonstrating that it did not occur in cells carrying a genetic deletion of the gene encoding Art1, and expression of Art1 fused to GFP revealed that the protein rapidly translocates to the plasma membrane following addition of methionine to the cell culture medium.



FIGURE 1. Example of the assay used to monitor internalization of Mup1. Mup1 is expressed as a fusion with GFP, which produces green fluorescence. In cells expressing Mup1-GFP only (top) one can see the protein is mostly localized to the cell surface before adding methionine. After methionine addition, the protein is internalized into small endocytic vesicles, and then later found within larger vacuoles. The localization of the vacuoles can be better defined through the expression of Vph1M-Cherry (red fluorescence), a vacuolar membrane marker (bottom). Reproduced under the Creative Commons Attribution Noncommercial Share Alike 4.0 International License from S. Lee et al., (2019) J. Cell Biol., published January 4, DOI: 10.1083/jcb.201712144. Copyright 2019 S. Lee, et al.



Prior work in the MacGurn lab had demonstrated a role for Ppz phosphatases (Ppz1 and Ppz2) in modulating ubiquitin levels in yeast. They hypothesized that these enzymes might also be among those that dephosphorylate and activate Art1 to enable ubiquitylation and endocytosis of Mup1. They tested this hypothesis by evaluating the methionine-dependent internalization of Mup1 in yeast cells bearing genetic deletions of the genes encoding both Ppz proteins (Δppz1Δppz2). Consistent with their hypothesis, internalization of Mup1 did not occur in Ppz-deficient cells when they were exposed to methionine (Figure 2). Similarly, Ppz deficiency led to failure of Can1 internalization, which could be demonstrated by exposing the cells to canavanine, a toxic arginine mimic. Cells that cannot internalize Can1 take up excessive amounts of canavanine, thereby leading to hypersensitivity to its toxicity as seen in the case of (Δppz1Δppz2) cells (Figure 3). In contrast, the researchers found no effect of Ppz deficiency on internalization of the uracil transporter, indicating that loss of the phosphatases did not lead to a generalized failure of endocytosis.


FIGURE 2. Demonstration that Ppz phosphatases are required for internalization of Mup1. Cultures contain both wild-type (WT) cells expressing Mup1-GFP and Vph1-mCherry and cells deficient in both Ppz phosphatases (Δppz1Δppz2) expressing only Mup1-GFP. Prior to methionine addition, the Mup1-GFP is localized to the plasma membrane and the Vph1-mCherry to the vacuoles. WT cells contain both fluorescent labels, whereas Δppz1Δppz2 cells contain only the Mup1-GFP label. Following addition of methionine, vacuolar Mup1-GFP is seen only in WT cells. Reproduced under the Creative Commons Attribution Noncommercial Share Alike 4.0 International License from S. Lee et al., (2019) J. Cell Biol., published January 4, DOI: 10.1083/jcb.201712144. Copyright 2019 S. Lee, et al.




FIGURE 3. Failure of internalization of the arginine transporter Can1 leads to excessive uptake and toxicity of canavanine. WT cells can survive to some degree in the presence of canavanine, but Δppz1Δppz2 cells show decreased resistance to its toxicity, consistent with a failure of Can1 internalization. This sensitivity is reversed by expression of PPZ1, but not the G2A mutant of PPZ1, which cannot be myristoylated and does not associate with the plasma membrane. Reproduced under the Creative Commons Attribution Noncommercial Share Alike 4.0 International License from S. Lee et al., (2019) J. Cell Biol., published January 4, DOI: 10.1083/jcb.201712144. Copyright 2019 S. Lee, et al.



Further work revealed that deficiency of Ppz1 alone led to a partial suppression of Mup1 internalization and that expression of Ppz1 without Ppz2 could partially reverse the effects of the Δppz1Δppz2 double mutation on endocytosis (Figure 3). However, expression of an inactive mutant Ppz1 enzyme could not compensate for Ppz deficiency, indicating that the enzymes must be able to remove phosphates from one or more target proteins to carry out this modulatory function.

The investigators knew that Ppz1 is N-myristoylated at its N-terminus, and based on its sequence, they expected that the same would be true of Ppz2. Myristoylation mediates localization of Ppz1 to the plasma membrane. The researchers found that a G2A mutant Ppz1 protein, which cannot be myristoylated, does not go to the plasma membrane and does not enable Δppz1Δppz2 double mutant yeast cells to internalize Mup1 or Can1 (Figure 3). These findings suggest that the ability of Ppz phosphatases to activate transport internalization requires their localization to the plasma membrane.

As noted above, the MacGurn lab had previously shown that Ppz phosphatases play a role in ubiquitin homeostasis. Specifically, Δppz1Δppz2 cells fail to dephosphorylate ubiquitin at pSer57, which serves as a signal for protein degradation. Consequently, these cells contain unusually low levels of ubiquitin, a deficiency that could be remedied by expression of the G2A mutant of Ppz1. This finding suggested that membrane association of Ppz's is not required for modulation of ubiquitin homeostasis as it is for Mup1 or Can1 internalization. Furthermore, the deficiencies in transporter internalization in Δppz1Δppz2 double mutant cells could not be reversed by overexpression of ubiquitin. These findings led the investigators to conclude that membrane-associated Ppz phosphatases are involved in protein endocytosis while cytosolic Ppz's regulate ubiquitin levels.

Through the use of stable isotope labeling of amino acids in cell culture (SILAC) combined with mass spectrometry (Figure 4), the investigators found that addition of methionine to yeast cells led to dephosphorylation of Art1 at Thr93, Thr245, and Thr795. This same approach demonstrated that the levels of phosphorylation of these three amino acids were higher in Δppz1Δppz2 double mutant cells than in wild-type cells. Using site-directed mutagenesis to express Art1 proteins containing arginine (to mimic permanent phosphorylation) or alanine (in order to prevent phosphorylation) at each of these positions suggested that phosphorylation of Thr93 or Thr795 suppresses Art1's ability to modulate endocytosis of Mup1 and Can1. A previously identified modulator of Art1 activity is the kinase Npr1, which phosphorylates the protein and inhibits its ability to localize to the plasma membrane. This led the researchers to hypothesize that Ppz phosphatases reverse the effects of Npr1. However, experiments failed to support this hypothesis, consistent with prior data showing that Npr1 does not phosphorylate Art1 at Thr93 or Thr245.



FIGURE 4. SILAC protocol. Cells are prelabeled with either standard amino acids (light) or amino acids labeled with heavy isotopes (heavy). The two groups of cells are treated differently (in this example with or without methionine) and then lysed. Proteins of interest are purified, combined, and then subjected to tryptic digestion and MS analysis. Each peptide from the two groups of cells will be detected simultaneously, but will vary by molecular weight. The ratio of intensity of the heavier to the lighter peptide provides information of the relative quantities of that peptide in the cells as a result of the two different treatments. Reproduced under the Creative Commons Attribution Noncommercial Share Alike 4.0 International License from S. Lee et al., (2019) J. Cell Biol., published January 4, DOI: 10.1083/jcb.201712144. Copyright 2019 S. Lee, et al.



Further work demonstrated that loss of Ppz's had no effect on the overall levels of Art1 or its ability to interact with Rsp1. Consistently, when the investigators expressed FLAG-tagged Rsp5 so they could use the FLAG moiety as a handle to pull down the protein along with all other associated proteins, they discovered essentially no differences in Rsp5's ability to interact with adaptor proteins as a result of Ppz deficiency. Although SILAC-mass spectrometry analysis revealed changes in phosphorylation of several sites in Rsp5 and Art3 as a result of Ppz phosphatase deficiency, the researchers concluded that Ppz's did not have major direct effects on Rsp5 and/or its interacting proteins.

Finally, the investigators discovered that Ppz deficiency did not affect translocation of Art1 to the plasma membrane, a finding consistent with the fact that Ppz's do not counteract the effects of Npr1. However, through use of a ubiquitin two-hybrid system employing Art1 as prey, they detected an interaction between Art1 and Mup1 that was decreased when the wild-type Art1 protein was replaced by the T93D or T795D mutants. The Art1/Mup1 interaction was also decreased in Δppz1Δppz2 double mutant cells as compared to wild-type cells.


Together, the findings suggest that Ppz phosphatases help to activate Art1 by removing phosphate groups at T93 and T795, resulting in a structural change that promotes Art1's interaction with Mup1 (and presumably Can1 as well). The researchers proposed a model suggesting that under resting conditions, hyperphosphorylated and inactive Art1 is complexed with Rsp5 in the cytosol (Figure 5). Activation leads to translocation of the complex to the plasma membrane and partial dephosphorylation by as yet unidentified phosphatases. At the membrane, the Art1/Rsp5 complex encounters myristoylated (membrane-associated) Ppz phosphatases, which remove phosphate groups at T93, T795 (and possibly elsewhere), resulting in an increase in the ability of Art1 to associate with Mup1. This association enables Rsp5 to ubiquitylate Mup1, targeting it to the endosome. Subsequently, phosphorylation of Art1 by Npr1 results in dissociation of the Art1/Rsp5 complex from the membrane and a return to the basal state.



Figure 5. Proposed role for Art1 and Ppz phosphatases in the internalization of Mup1. In the resting state, inactive, hyperphosphorylated Art1 is associated with Rsp5, and the protein complex is soluble. Exposre of the cell to methionine leads to translocation of the complex to the plasma membrane and partial dephosphorylation. At the plasma membrane, the complex encounters membrane-associated Ppz phosphatases (i), which remove additional phosphates, leading to activation of Art1, and its association with Mup1. This allows ubiquitylation of Mup1 by Rsp5, and its ultimate internalization (ii). Npr1 then phosphorylates Art1 (iii), leading to its inactivation and dissociation from the plasma membrane. Reproduced under the Creative Commons Attribution Noncommercial Share Alike 4.0 International License from S. Lee et al., (2019) J. Cell Biol., published January 4, DOI: 10.1083/jcb.201712144. Copyright 2019 S. Lee, et al.


Clearly, there is much about this pathway that remains to be determined, such as the identities of the phosphatases that promote translocation of Art1/Rsp5 to the plasma membrane. However, these new findings add an important new piece to the puzzle of how Rsp5 selects and then targets plasma membrane proteins to the endosome.



View JBC article: Methionine triggers Ppz-mediated dephosphorylation of Art1 to promote cargo-specific endocytosis







The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2018 by Vanderbilt University