Sorting of Golgi resident proteins into different subpopulations of COPI vesicles: a role for ArfGAP1

Abstract

We present evidence for two subpopulations of coatomer protein I vesicles, both containing high amounts of Golgi resident proteins but only minor amounts of anterograde cargo. Early Golgi proteins p24alpha2, beta1, delta1, and gamma3 are shown to be sorted together into vesicles that are distinct from those containing mannosidase II, a glycosidase of the medial Golgi stack, and GS28, a SNARE protein of the Golgi stack. Sorting into each vesicle population is Arf-1 and GTP hydrolysis dependent and is inhibited by aluminum and beryllium fluoride. Using synthetic peptides, we find that the cytoplasmic domain of p24beta1 can bind Arf GTPase-activating protein (GAP)1 and cause direct inhibition of ArfGAP1-mediated GTP hydrolysis on Arf-1 bound to liposomes and Golgi membranes. We propose a two-stage reaction to explain how GTP hydrolysis constitutes a prerequisite for sorting of resident proteins, yet becomes inhibited in their presence.

Figure 1.

Accumulation of COPI vesicles in the presence of the α-SNAP ^dn mutant. (A) Vesicles formed using salt-washed Golgi membranes were compared with vesicles formed using nonwashed membranes in the absence or presence of the α-SNAP^dn mutant. Proteins from solubilized membranes or vesicles were separated by SDS-PAGE and subjected to Western blot analysis using specific primary antibodies to Mann II, p24γ₃, p24α₂, and p24β₁ followed by ECL. Note the lower amount of Mann II in the absence of the α-SNAP^dn mutant. Golgi membranes are shown in increasing amounts expressed as the percentage of starting material. (B) Vesicles were formed using salt-washed Golgi membranes in the presence or absence of exogenously added NSF/α-SNAP^wt and/or the α-SNAP^dn mutant. (C) Coatomer was depleted (Dp) from cytosol using specific monoclonal antibodies, recognizing native coatomer or mock (Mo) depleted using an irrelevant antibody and used in budding reactions with nonwashed membranes and the α-SNAP^dn mutant. Depleted cytosol was rescued (Rs) subsequently by adding back exogenously purified coatomer and compared with vesicles formed using untreated cytosol (Ctr). (D) Vesicle budding was performed using nonwashed membranes in the presence of the α-SNAP^dn mutant and tested for anterograde cargo incorporation (pIgR, RSA, and ApoE). Vesicles were also generated in the presence of an ATP depletion system (ATP⁻) for comparison. Proteins from solubilized membranes or vesicles were separated by SDS-PAGE and subjected to Western blot analysis and revealed by the ECL method using antibodies.

Figure 2.

Immunoisolation of p24-containing vesicles reveals subpopulations of COPI vesicles. (A–D) Magnetic beads precoated with either an irrelevant control antibody (Ctr) or specific (Sp) antibodies to the cytoplasmic domain of p24γ₃ were incubated with nonwashed Golgi membranes (A, left, and B) or with COPI vesicles formed from such membranes in the presence of the α-SNAP^dn mutant (A, right, and C and D). Isolated membranes and vesicles were pelleted, fixed, embedded, and processed for EM or solubilized and subjected to SDS-PAGE. After transfer to nitrocellulose, Mann II, p24γ₃, p24α₂, p24δ₁, and p24β₁ were detected using specific antibodies followed by ECL. Vesicles remaining in the supernatant (Supn) after immunoisolation were also monitored in terms of their protein content (C, left lanes). (D) Immunoisolated vesicles were probed for the anterograde cargo marker, RSA, and compared with p24γ₃, p24α₂, p24δ₁, and p24β₁. Lanes containing total Golgi (Gt) and 20% of total vesicles (V_tot) used for immunoisolation are included for comparison.

Figure 2.

Immunoisolation of p24-containing vesicles reveals subpopulations of COPI vesicles. (A–D) Magnetic beads precoated with either an irrelevant control antibody (Ctr) or specific (Sp) antibodies to the cytoplasmic domain of p24γ₃ were incubated with nonwashed Golgi membranes (A, left, and B) or with COPI vesicles formed from such membranes in the presence of the α-SNAP^dn mutant (A, right, and C and D). Isolated membranes and vesicles were pelleted, fixed, embedded, and processed for EM or solubilized and subjected to SDS-PAGE. After transfer to nitrocellulose, Mann II, p24γ₃, p24α₂, p24δ₁, and p24β₁ were detected using specific antibodies followed by ECL. Vesicles remaining in the supernatant (Supn) after immunoisolation were also monitored in terms of their protein content (C, left lanes). (D) Immunoisolated vesicles were probed for the anterograde cargo marker, RSA, and compared with p24γ₃, p24α₂, p24δ₁, and p24β₁. Lanes containing total Golgi (Gt) and 20% of total vesicles (V_tot) used for immunoisolation are included for comparison.

Figure 3.

Immunolocalization of p24 proteins in cells and vesicles. (A–C) Subcellular localization of p24β₁ (b, green) as determined by indirect immunofluorescence localization. For comparison, labeling of Mann II (red) is shown in A. The extent of overlap is shown in C. (D) Thin frozen sections of NRK cells were labeled with antibodies to p24β₁ (10 nm gold). Gold particles were found at one side of the Golgi stack, codistributing with antibodies to p24γ₃ (5 nm gold, arrows). (E and F) Vesicles bound to grids were labeled with antibodies against two different p24s and revealed by protein A gold conjugates. The majority of the labeled vesicles contained both p24 proteins of each combination. (G) Quantification of p24 proteins in pair-wise labeling experiments. Different antibodies and protein A gold sizes (pAG) were used. Vesicles labeled with only one type of gold particle (two or more of the same gold particle) are expressed as the percentage of single labeled (SL), and vesicles containing both types of gold particle (two or more different gold particles) are expressed as double labeled (DL).

Figure 3.

Immunolocalization of p24 proteins in cells and vesicles. (A–C) Subcellular localization of p24β₁ (b, green) as determined by indirect immunofluorescence localization. For comparison, labeling of Mann II (red) is shown in A. The extent of overlap is shown in C. (D) Thin frozen sections of NRK cells were labeled with antibodies to p24β₁ (10 nm gold). Gold particles were found at one side of the Golgi stack, codistributing with antibodies to p24γ₃ (5 nm gold, arrows). (E and F) Vesicles bound to grids were labeled with antibodies against two different p24s and revealed by protein A gold conjugates. The majority of the labeled vesicles contained both p24 proteins of each combination. (G) Quantification of p24 proteins in pair-wise labeling experiments. Different antibodies and protein A gold sizes (pAG) were used. Vesicles labeled with only one type of gold particle (two or more of the same gold particle) are expressed as the percentage of single labeled (SL), and vesicles containing both types of gold particle (two or more different gold particles) are expressed as double labeled (DL).

Figure 3.

Immunolocalization of p24 proteins in cells and vesicles. (A–C) Subcellular localization of p24β₁ (b, green) as determined by indirect immunofluorescence localization. For comparison, labeling of Mann II (red) is shown in A. The extent of overlap is shown in C. (D) Thin frozen sections of NRK cells were labeled with antibodies to p24β₁ (10 nm gold). Gold particles were found at one side of the Golgi stack, codistributing with antibodies to p24γ₃ (5 nm gold, arrows). (E and F) Vesicles bound to grids were labeled with antibodies against two different p24s and revealed by protein A gold conjugates. The majority of the labeled vesicles contained both p24 proteins of each combination. (G) Quantification of p24 proteins in pair-wise labeling experiments. Different antibodies and protein A gold sizes (pAG) were used. Vesicles labeled with only one type of gold particle (two or more of the same gold particle) are expressed as the percentage of single labeled (SL), and vesicles containing both types of gold particle (two or more different gold particles) are expressed as double labeled (DL).

Figure 4.

A requirement for GTP hydrolysis by Arf-1. (A) Vesicles were formed in the presence of GTP, GTPγS, or Arf-1^Q71_L. (B) Vesicles were formed in the presence of GTP, GTPγS, BeF_x, or AlF_x. Proteins from solubilized membranes or vesicles were separated by SDS-PAGE and subjected to Western blot analysis using specific primary antibodies to Mann II, β-COP, p24γ₃, p24α₂, p24δ₁, GS28, p24β₁, and the KDR followed by ECL.

Figure 4.

A requirement for GTP hydrolysis by Arf-1. (A) Vesicles were formed in the presence of GTP, GTPγS, or Arf-1^Q71_L. (B) Vesicles were formed in the presence of GTP, GTPγS, BeF_x, or AlF_x. Proteins from solubilized membranes or vesicles were separated by SDS-PAGE and subjected to Western blot analysis using specific primary antibodies to Mann II, β-COP, p24γ₃, p24α₂, p24δ₁, GS28, p24β₁, and the KDR followed by ECL.

Figure 5.

Effect of peptides derived from p24 proteins on ArfGAP1 activity. (A and B) Activity was assayed on Arf-1 preloaded with [γ-³²P]GTP in the presence of liposomes and in the absence or presence of 0.2 mM of respective peptides, and the release of [γ-³²P] was monitored. (C and D) Activity was assayed on Golgi membrane-bound Arf-1 preloaded with [γ-³²P]GTP. At time zero, 50 nM ArfGAP1 were added to each incubation except to the control, in the presence or absence of peptides at a concentration of 0.2 mM unless otherwise indicated. At different time points, the amount of [γ-³²P]GTP, which remained bound to membranes, was determined.

Figure 5.

Effect of peptides derived from p24 proteins on ArfGAP1 activity. (A and B) Activity was assayed on Arf-1 preloaded with [γ-³²P]GTP in the presence of liposomes and in the absence or presence of 0.2 mM of respective peptides, and the release of [γ-³²P] was monitored. (C and D) Activity was assayed on Golgi membrane-bound Arf-1 preloaded with [γ-³²P]GTP. At time zero, 50 nM ArfGAP1 were added to each incubation except to the control, in the presence or absence of peptides at a concentration of 0.2 mM unless otherwise indicated. At different time points, the amount of [γ-³²P]GTP, which remained bound to membranes, was determined.

Figure 6.

Effect of peptides in the vesicle budding assay and binding of ArfGAP1 to cytoplasmic domain peptides. (A and B) Vesicles were formed in the absence or presence of different p24 cytoplasmic domain antibodies. (A) Peptides were added at the following concentrations: 0.1 and 0.01 mM. The effect of each peptide was evaluated by SDS PAGE and Western blotting using specific antibodies and compared with control conditions (C) or to a mock control (HCl) containing the same amount of HCl as in experiments where peptides were added. (C and D) Synthetic peptides corresponding to the cytoplasmic domains of different p24 proteins and E19 and RER1 were attached to Sepharose beads and probed for their ability to recruit ArfGAP1 from cytosol or purified recombinant ArfGAP1. (C) Two different salt conditions were tested for 230 mM Na/40 mM KCl and 115 KOAc (KAc buffer). (D) High salt conditions were used (300 mM Na/90 mM KCl).

Figure 6.

Effect of peptides in the vesicle budding assay and binding of ArfGAP1 to cytoplasmic domain peptides. (A and B) Vesicles were formed in the absence or presence of different p24 cytoplasmic domain antibodies. (A) Peptides were added at the following concentrations: 0.1 and 0.01 mM. The effect of each peptide was evaluated by SDS PAGE and Western blotting using specific antibodies and compared with control conditions (C) or to a mock control (HCl) containing the same amount of HCl as in experiments where peptides were added. (C and D) Synthetic peptides corresponding to the cytoplasmic domains of different p24 proteins and E19 and RER1 were attached to Sepharose beads and probed for their ability to recruit ArfGAP1 from cytosol or purified recombinant ArfGAP1. (C) Two different salt conditions were tested for 230 mM Na/40 mM KCl and 115 KOAc (KAc buffer). (D) High salt conditions were used (300 mM Na/90 mM KCl).

Figure 6.

Effect of peptides in the vesicle budding assay and binding of ArfGAP1 to cytoplasmic domain peptides. (A and B) Vesicles were formed in the absence or presence of different p24 cytoplasmic domain antibodies. (A) Peptides were added at the following concentrations: 0.1 and 0.01 mM. The effect of each peptide was evaluated by SDS PAGE and Western blotting using specific antibodies and compared with control conditions (C) or to a mock control (HCl) containing the same amount of HCl as in experiments where peptides were added. (C and D) Synthetic peptides corresponding to the cytoplasmic domains of different p24 proteins and E19 and RER1 were attached to Sepharose beads and probed for their ability to recruit ArfGAP1 from cytosol or purified recombinant ArfGAP1. (C) Two different salt conditions were tested for 230 mM Na/40 mM KCl and 115 KOAc (KAc buffer). (D) High salt conditions were used (300 mM Na/90 mM KCl).

Figure 6.

Effect of peptides in the vesicle budding assay and binding of ArfGAP1 to cytoplasmic domain peptides. (A and B) Vesicles were formed in the absence or presence of different p24 cytoplasmic domain antibodies. (A) Peptides were added at the following concentrations: 0.1 and 0.01 mM. The effect of each peptide was evaluated by SDS PAGE and Western blotting using specific antibodies and compared with control conditions (C) or to a mock control (HCl) containing the same amount of HCl as in experiments where peptides were added. (C and D) Synthetic peptides corresponding to the cytoplasmic domains of different p24 proteins and E19 and RER1 were attached to Sepharose beads and probed for their ability to recruit ArfGAP1 from cytosol or purified recombinant ArfGAP1. (C) Two different salt conditions were tested for 230 mM Na/40 mM KCl and 115 KOAc (KAc buffer). (D) High salt conditions were used (300 mM Na/90 mM KCl).

Figure 7.

GTP hydrolysis–driven sorting coupled with ArfGAP1 modulation. The model is divided into four steps from left to right and then from bottom to top. (I) GDP bound to Arf-1 attached to the membrane is replaced with GTP by Arf-1 guanine exchange factor. Cytosolic coatomer binds Arf-1^GTP but is quickly released upon GTP hydrolysis by Arf-1. This step requires ArfGAP1 and is denoted in green as having a high activity (HA) on Arf-1. Resident cargo proteins bind coatomer directly through cytoplasmic domain motifs (e.g., K[X]KXX). In order for coatomer to capture resident proteins, coatomer first needs to be released from the membrane. This “sorting step” requires GTP hydrolysis by Arf-1. If resident cargo proteins are in addition induced to cluster (Weiss and Nilsson, 2000), coatomer is predicted to bind more strongly, thus favoring these over other resident proteins. (II) The activity of ArfGAP1 is downmodulated by captured resident cargo proteins. The rate of GTP hydrolysis by Arf-1 is as a consequence slowed down, enabling coatomer to remain on the membrane. The presence of resident cargo proteins promotes polymerization of the coat leading up to vesicle formation and budding (III). Though downmodulated in its activity, residual ArfGAP1 activity will permit Arf-1 to hydrolyze its GTP, and coatomer is released back into the cytosol. For simplicity, the release of coatomer, Arf-1, and ArfGAP1 from the vesicle is shown as one event.