The reversible modification regulates the membrane-binding state of Apg8/Aut7 essential for autophagy and the cytoplasm to vacuole targeting pathway

Abstract

Autophagy and the Cvt pathway are examples of nonclassical vesicular transport from the cytoplasm to the vacuole via double-membrane vesicles. Apg8/Aut7, which plays an important role in the formation of such vesicles, tends to bind to membranes in spite of its hydrophilic nature. We show here that the nature of the association of Apg8 with membranes changes depending on a series of modifications of the protein itself. First, the carboxy-terminal Arg residue of newly synthesized Apg8 is removed by Apg4/Aut2, a novel cysteine protease, and a Gly residue becomes the carboxy-terminal residue of the protein that is now designated Apg8FG. Subsequently, Apg8FG forms a conjugate with an unidentified molecule "X" and thereby binds tightly to membranes. This modification requires the carboxy-terminal Gly residue of Apg8FG and Apg7, a ubiquitin E1-like enzyme. Finally, the adduct Apg8FG-X is reversed to soluble or loosely membrane-bound Apg8FG by cleavage by Apg4. The mode of action of Apg4, which cleaves both newly synthesized Apg8 and modified Apg8FG, resembles that of deubiquitinating enzymes. A reaction similar to ubiquitination is probably involved in the second modification. The reversible modification of Apg8 appears to be coupled to the membrane dynamics of autophagy and the Cvt pathway.

Figure 1

Processing of Apg8. (A) Removal of myc epitopes connected to the carboxyl terminus of Apg8. Lysates were prepared from logarithmically growing cultures of wild-type cells (SEY6210; 1), Δapg8 cells (KVY5; 2), and Δapg8 cells that expressed myc-Apg8 (3 and 5) or Apg8-myc (4 and 6) encoded by a centromeric plasmid. Lysates were subjected to immunoblotting with Apg8-specific antibodies (α-Apg8; 1–4) or a monoclonal antibody against myc, 9E10 (α-myc; 5 and 6). (B) Cleavage of Apg8-myc in apg mutant cells. The centromeric plasmid encoding Apg8-myc was introduced into the various lines of apg mutant cells. Lysates were prepared and subjected to immunoblotting as described above. Lane numbers correspond to the designations of the apg mutants. Note that endogenous Apg8 was expressed in these mutants.

Figure 3

Identification of Apg4 as a novel cysteine protease. (A) Alignment of the amino acid sequences of Apg4 and its homologues from other organisms. An arrowhead indicates the most likely candidate for the cysteine residue at the active site of these enzymes. The homologues include two proteins from Homo sapiens (HsApg4A; deduced amino acid sequence based on an EST clone with GenBank No. W30741 and a genomic sequence from 889N15 with No. AL031177, and HsApg4B, No. AL080168), two proteins from Caenorhabditis elegans (CeApg4A and CeApg4B; Nos. Z68302 and AL110500) and one protein from Arabidopsis thaliana (AtApg4; No. AC004005). (B) Effects of the replacement of Cys159 of Apg4 on the cleavage of Apg8 and the Cvt pathway. (Top) Cleavage of Apg8-myc. Wild type of Apg4 (W.T.; 1), Apg4C159S (2), and Apg4C159A (3) were expressed from centromeric plasmids derived from pRS314 in Δapg4Δapg8 cells (KVY15) that expressed Apg8-myc encoded by a centromeric plasmid. Lysates were prepared from exponentially growing cells and subjected to immunoblotting with Apg8-specific antibodies. (Middle) The Cvt pathway. Apg4, Apg4C159S, and Apg4C159A encoded by centromeric plasmids were expressed in Δapg4 cells (KVY13). Maturation of proAPI in growing cells was examined by immunoblotting with API-specific antibodies. (Bottom) Detection of Apg4 and its variants. Apg4 and its variants were expressed from pRS426-based 2-μl plasmids in Δapg4 cells (KVY13) and detected by immunoblotting with Apg4-specific antibodies. (4) Δapg4 cells that harbored the vector only. (C) Effects of mutation of Cys159 in Apg4 on autophagy. Apg4, Apg4C159S, and Apg4C159A were expressed from centromeric plasmids in Δapg4 cells (KVY53) that expressed Pho8Δ60. Cells at the logarithmic phase of growth in SD+CA medium were transferred to nitrogen-starvation medium, SD(-N), for 4.5 h. Lysates were prepared from growing (open bars) and starved (closed bars) cells. ALP activity in each lysate was measured as described previously (Noda and Ohsumi 1998). Values shown are means ± SD of results from triplicates in each case. Numbering of columns corresponds to the numbering of lanes in B.

Figure 2

Cleavage of Apg8 by Apg4 in vitro. (A) Cleavage of Apg8-myc by Apg4 in vitro. Lysates were prepared from vegetative cultures of Δapg8 cells (KVY5), Δapg4Δapg8 cells (KVY15), and Δapg4Δapg8 cells that expressed Apg8-myc encoded by a centromeric plasmid as described in Materials and Methods. A lysate of Δapg4Δapg8 cells that expressed Apg8-myc was incubated with a lysate of Δapg8 or Δapg4Δapg8 cells at 30°C for the indicated times, and then subjected to immunoblotting with Apg8-specific antibodies. (B) Effects of protease inhibitors on cleavage of Apg8-myc. A lysate of Δapg4Δapg8 cells that expressed Apg8-myc was incubated with a lysate of Δapg8 cells at 30°C for 1 h in the presence of various protease inhibitors, as follows, with the type of protease indicated in parentheses: 1 mM NEM (cysteine; 2), 1 mM PMSF (serine; 3), 1 mM pepstatin (aspartic; 4), and 10 mM 1,10-phenanthroline (metallo-; 5). The result of the reaction in the absence of inhibitors is shown in 1. Cleavage of Apg8-myc was examined by immunoblotting with Apg8-specific antibodies. (C) Cleavage of recombinant Apg8-myc by recombinant Apg4. GST-Apg8-myc, GST-Apg4, and GST were prepared as described in Materials and Methods. GST-Apg8-myc was incubated with GST-Apg4 in the presence of 1 mM DTT (+) or in its absence (−) at 30°C for 1 h. As a control, GST-Apg8-myc was treated with GST in the presence of 1 mM DTT at 30°C for 1 h. Cleavage of GST-Apg8-myc was examined as described above.

Figure 4

Cleavage of Apg8. (A) Schematic representation of Apg8-myc fusion proteins. The amino acid sequence of the Apg8-myc junction is shown in the single-letter code. (B) Cleavage of Apg8-myc and its variants. Lysates were prepared from the logarithmically growing Δapg8 cells (KVY5) that expressed the fusion proteins depicted in A from pRS316-based plasmids. Numbering of lanes corresponds to the numbering of the proteins in A. (C) Alignment of the amino acid sequences of carboxy-terminal segments of Apg8 and representative homologues. An arrowhead indicates the Gly residue at the cleavage site of Apg8 and, possibly, of the homologues. Apg8/Aut7 (S. cerevisiae), SpApg8 (Schizosaccharomyces pombe; GenBank No. AL032684), HsGATE-16/GEF2 (H. sapiens), RnGATE-16/GEF2 (Rattus norvegicus), RnLC3 (R. norvegicus), AtApg8 (A. thaliana; No. AC006220), and Lb-Aut7 (Laccaria bicolor; Kim et al. 1999b). (D) MALDI mass spectra of Apg8 (top) and Apg8-myc (bottom). Arrowhead, Apg8 generated by cleavage by Apg4; Arrow, Apg8-myc; *contaminant.

Figure 5

Autophagy and the Cvt pathway in the presence of carboxy-terminal variants of Apg8. (A) Schematic representation of the variants of Apg8. Carboxy-terminal sequences are shown. (B and C) The Cvt pathway. (B) Apg8 and its variants were expressed in Δapg8 cells (KVY5) and in Δapg4Δapg8 cells from the pRS316 vector. (C) Apg8FGR and Apg8FA were expressed in Δapg8 cells (KVY5) from pRS316-based centromeric plasmids. Maturation of proAPI in cells at the logarithmic phase of growth was examined by immunoblotting with API-specific antibodies. Expression of each variant of Apg8 was detected with Apg8-specific antibodies. Numbering of lanes corresponds to the numbering of constructs in A. mAPI, mature API. (D) Autophagy (ALP assay). Each pRS316-based plasmid encoding Apg8 or a variant was introduced into mutants with SEY6210 background, namely, Δapg8 (KVY54) and Δapg4Δapg8 (KVY52), which expressed Pho8Δ60. Cells were cultured in SD+CA medium to logarithmic phase and transferred to nitrogen-starvation medium, SD(-N), for 4.5 h. Lysates were prepared from growing (open bars) and starved (closed bars) cells. ALP activity in each lysate was measured as described previously. Values shown are means ± SD of results from triplicates in each case. Numbering of columns corresponds to the numbering of constructs in A. v, cells harboring the vector only.

Figure 6

Distribution of Apg8. (A) Lysates were prepared from SEY6210 cells (Apg8FGR; wild type), KVY13 cells (Apg8FGR; Δapg4), KVY5 cells that expressed Apg8F (Apg8F; Δapg8), or Apg8FA (Apg8FA; Δapg8), KVY 135 cells (Apg8FGR; Δapg6), KVY118 cells (Apg8FGR; Δapg7), and KVY15 cells that expressed Apg8FG (Apg8FG; Δapg4Δapg8), as described in Materials and Methods. (Left) Solubilization. Lysates were centrifuged at 100,000 g for 1 h to generate pellets. The pellets were treated with 1 M NaCl or 1% deoxycholate (DOC) or held untreated (Non) on ice for 30 min, and then centrifuged at 100,000 g for another 1 h to generate to supernatants (S) and pellets (P). (Right) Subcellular distribution. Lysates were centrifuged at 13,000 g for 15 min to generate supernatants and pellets (LSP). The resulting supernatants were then centrifuged at 100,000 g for 1 h to generate supernatants (HSS) and pellets (HSP). The distribution of Apg8 was examined by immunoblotting with Apg8-specific antibodies. (B) The LSP and the HSP of wild-type cells (SEY6210) were treated with 1 M NaCl or held untreated (Non) on ice for 30 min, and then centrifuged 100,000 g for 1 h to generate to supernatants (S) and pellets (P). Apg8 was detected as described above.

Figure 7

Cleavage of Apg8FG* by Apg4. Solubilization of the Apg8 recovered in the LSP. (Left) The LSP was prepared from the Δapg4Δapg8 cells (KVY15) that expressed Apg8FG, as described in Fig. 6. It was treated with 1 M NaCl, 2 M Urea, 0.1 M Na₂CO₃, pH 11.5, 1% deoxycholate (DOC), or 2% Triton X-100 (TX100) or held untreated (Non) on ice for 30 min, with subsequent centrifugation at 100,000 g for 1 h to generate a supernatant (S) and a pellet (P). (Right) The HSP prepared from the Δapg4Δapg8 cells (KVY15) that expressed Apg8FG was treated with NaCl or not, and then separated to a supernatant (S) and a pellet (P). The distributions of Apg8 were examined by immunoblotting with Apg8-specific antibodies. (B and C) Release of Apg8 from the LSP of Δapg4Δapg8 cells (KVY15) expressing Apg8FG. (B) The LSP was incubated at 30°C for 1 h with the HSS fractions that were prepared from Δapg4 cells (KVY13; 1) and Δapg4 cells that expressed Apg4 (2 and 3) or Apg4C159S (4) from a multicopy vector (pRS426). The reaction with the HSS that contained Apg4 was performed in the presence (3) or absence (2) of 1 mM NEM. Then, reaction mixtures were centrifuged at 100,000 g for 1 h to separate to supernatants (S) and pellets (P). (C) The LSP was incubated with GST-Apg4 or GST at 30°C for 1 h, and then the reaction mixtures were centrifuged at 100,000 g for 1 h to generate supernatants. Apg8 was detected by immunoblotting with Apg8-specific antibodies. I, 5% of input. (D) Release of Apg8 from the HSP and LSP. The HSP and LSP were prepared from of Δapg4Δapg8 cells (KVY15) expressing Apg8FG. Both were first treated with the HSS that was prepared from Δapg4 cells that expressed Apg4, and then with 1 M NaCl or held untreated (Non). The mixtures were separated to supernatants (S) and pellets (P) by centrifugation at 100,000 g for 1 h. The distribution of Apg8 was examined by immunoblotting with Apg8-specific antibodies.

Figure 8

Mobilities of Apg8FGR, Apg8FG. and Apg8FG* (A) Standard SDS-PAGE. (1) Apg8 expressed in Δapg4 cells (Apg8FGR), (2) Apg8 generated by cleavage of Apg8-myc in vitro (Apg8FG; see Fig. 2 A), (3) Apg8 in Δapg7 cells (Apg8FG). (B) SDS-PAGE in the presence of urea. Each sample was resolved in standard sample buffer for SDS-PAGE and subjected to electrophoresis on a modified slab gel, with subsequent immunoblotting with Apg8-specific antibodies. The modified gel consisted of a standard stacking gel and a 13.5% polyacrylamide separating gel that contained the standard concentration of SDS and 6 M urea. (1–3) The same samples as in A (1–3), (4) Apg8FG* (Apg8 recovered in the LSP of Δapg4Δapg8 cells that expressed Apg8FG), (5) Apg8 (Apg8FG) released from the LSP by cleavage in vitro by Apg4 (see Fig. 7 C), (6) Apg8 (Apg8FG and Apg8FG*) in wild-type cells. *Apg8FGR, **Apg8FG, ***Apg8FG*.

Figure 9

Schematic model of the serial modification of Apg8. First, Apg8FGR (newly synthesized Apg8) is converted to Apg8FG as a result of proteolytic cleavage by Apg4. The intracellular site of this cleavage remains to be determined. Second, Apg8FG is converted to Apg8FG-X via a reaction that resembles ubiquitination and becomes tightly membrane bound. The molecule X might be buried in a membrane. Finally, Apg8FG-X is reversed to soluble or loosely membrane bound Apg8FG by proteolytic cleavage at the junction of Apg8FG-X by Apg4. Some of the reversed Apg8FG might be recycled in a subsequent conjugation reaction.