|Home | About | Journals | Submit | Contact Us | Français|
Autophagosome formation and specific substrate recruitment during Autophagy require ligation of the ubiquitin-like protein (UBL) Atg8 to the head group of the lipid phosphatidylethanolamine (PE). Atg8 lipidation is mediated by distinctive UBL cascades involving autophagy-specific E1, E2, and E3 enzymes that differ substantially in sequence from components of other UBL conjugation cascades. Structural studies are important for elucidating the roles of Atg proteins that regulate multiple steps involved in autophagy. This chapter describes methods to prepare and crystallize selected proteins and complexes involved in autophagy UBL conjugation pathways, as a guide for strategies for structural and biochemical characterization of Atg proteins.
Autophagy is characterized by the formation of double-membrane structures called autophagosomes, which sequester the cellular components to be degraded (Mizushima, 2007). This process comprises phagophore initiation and vesicle nucleation, vesicle elongation, fusion with lysosome/vacuole, and degradation (Kraft and Martens, 2012; Levine and Klionsky, 2004). Each step is tightly regulated by numerous proteins, including at least 15 distinct autophagy-related (Atg) proteins that are core components of many forms of autophagy (Hurley and Schulman, 2014; Klionsky and Schulman, 2014; Mizushima et al., 2011). Of these, Atg12 and Atg8 are ubiquitin-like proteins (UBLs) required for autophagosome formation and recruitment of specific substrates (e.g., protein aggregates and damaged organelles) to autophagomes (Geng and Klionsky, 2008). In mammals, 6 subfamilies of Atg8, including LC3, GABARAP, and GATE-16 proteins (for simplicity, referred to as Atg8 hereafter) have been identified (Geng and Klionsky, 2008; Schulman and Harper, 2009) (Ohsumi and Mizushima, 2004). Similar to ubiquitin, Atg12 and Atg8 are ligated to their targets via E1–E2–E3 enzyme cascades. Atg12 is unusual in that after activation by the E1 enzyme Atg7, it is transferred to its E2 enzyme Atg10 and is then directly ligated to its protein substrate Atg5 without the involvement of an E3 enzyme (Figure 1A–1C). Unlike Atg12, Atg8 is processed by the cysteine protease Atg4 to expose its C-terminal glycine, activated by the E1 enzyme Atg7, and then transferred to its E2 enzyme Atg3. This results in a transient Atg3~Atg8 intermediate in which the C-terminus of Atg8 is linked by a covalent but reactive thioester bond to the catalytic cysteine of Atg3 (“~” represents a covalent bond). Atg8 is then transferred from Atg3 to the head group of the lipid phosphatidylethanolamine (PE). This reaction is catalyzed by a multi-protein E3 enzyme composed of the product of the Atg12 ligation reaction bound to Atg16 (Atg12~Atg5-Atg16) (Figure 1D–1F) (Geng and Klionsky, 2008; Nakatogawa et al., 2009; Ohsumi and Mizushima, 2004). Mutations of the genes involved in these cascades are observed in human diseases such as ataxia, asthma and Crohn’s disease (Barrett et al., 2008; Hampe et al., 2007; Kim et al., 2016; Martin et al., 2012; Poon et al., 2012; Rioux et al., 2007).
Given the important role of the UBL conjugation system in autophagy and its close association with human health, the modulation of protein function in the Atg8 lipidation pathway through chemical inhibition of the E1–E2–E3 cascade has become an area of significant interest. Several promising autophagy inhibitors have been used in clinical trials for cancer (Jiang and Mizushima, 2014), and a recent study showed that an intestinal tissue-specific knockout of Atg7 prevents tumor initiation (Levy et al., 2015).
Structures of enzymes in this pathway have greatly enhanced our understanding of the molecular mechanisms underlying autophagy, and may facilitate the development of approaches to manipulate autophagy to treat diseases in the future. In this chapter, we present the strategy and detailed methodology for crystallizing several Atg proteins and their interacting partners involved in the Autophagy UBL conjugation pathway.
Seven proteins or protein conjugates are involved in the UBL conjugation cascade for autophagy: Atg7 (E1), Atg3 (E2), Atg10 (E2), Atg12~Atg5–Atg16 (E3), Atg12 (UBL), Atg8 (UBL), and Atg4 (de-conjugation enzyme). These E1, E2, and E3 enzymes are “non-canonical,” because they differ in many respects from the “canonical” E1, E2, and E3 enzymes such as those for ubiquitin, SUMO, and NEDD8. Canonical E1 enzymes contain three domains: a pseudosymmetric UBL adenylation domain for Ubl activation; a separate catalytic cysteine-containing domain for Ubl thiolester formation; and a ubiquitin-fold domain that recruits the E2 for transfer of the UBL from the catalytic cysteine on the E1 to that on the E2. In contrast, Atg7 is a symmetric homodimer, also defined by three distinct domains: a C-terminal domain (CTD) that houses both the adenylation domain and the catalytic cysteine; a flexible extreme-CTD (ECTD) extension; and, a unique N-terminal domain (NTD) responsible for recruiting the E2s. Atg7 homodimerization is mediated via its CTD (Figure 2). Atg3 and Atg10 contain unique catalytic residues around the catalytic cysteine, and they have additional secondary structures inserted within the E2 fold (Figure 2). Atg12~Atg5–Atg16 is a multi-protein complex (Figure 2) that functions as an E3 enzyme to ligate Atg8 to the head group of PE. This is in contrast to other canonical E3 enzymes that ligate their UBLs to a lysine residue of a protein substrate. The Atg12~Atg5–Atg16 complex does not share sequence similarity with any known UBL E3 ligases. Given these differences between the autophagy and canonical UBL conjugation systems, it is important to structurally characterize each component in the autophagy UBL conjugation pathway in order to understand the mechanisms that control the ligation of Atg8 to PE (Figure 3).
Table 1 lists existing crystal structures for several components and protein complexes in the UBL conjugation system.
Atg7 is composed of two domains (NTD, residues 1–289; and CTD, residues 294–630) separated by a short linker. We previously crystallized the NTD and CTD separately (Taherbhoy et al., 2011) (Figure 3). Atg7NTD and Atg7CTD coding sequences were each cloned into the pGEX-4T1 vector (GE Healthcare Life Sciences) and transformed into BL21 Gold (DE3) cells (Novagen).
To understand the molecular mechanism underlying the recognition of Atg3 by Atg7, we have crystallized Atg7NTD in complex with Atg3FRpep (Figure 3) by the following procedure:
Atg7 is the only E1 enzyme that activates the two UBLs Atg8 and Atg12 and transfers them to two distinct E2 enzymes - Atg3 and Atg10, respectively. Crystallization of the Atg7–Atg3 and Atg7–Atg10 complexes provided insights into how a single E1 enzyme recognizes the two different E2s and how active sites are juxtaposed for these noncanonical E1–E2 complexes to promote autophagy. To visualize the juxtaposition of E1 and E2 catalytic cysteines and to facilitate the crystallization process, cross-linking is performed with the homobifunctional sulfhydryl cross-linker bismaleimidoethane (BMOE) by using the following strategy. Versions of Atg3 and Atg10 containing only a single cysteine at the active site are prepared, BMOE is reacted with each E2, and excess cross-linker is removed by desalting. For Atg3, Atg7 (1–613) is added. For Atg10, a version of Atg7 (1–613) with C39S C195S C375A mutations is added to reduce background cross-linking. By this strategy, we obtain crystals of Atg7–BMOE–Atg3 and Atg7–BMOE–Atg10 (hereafter referred to as Atg7–Atg3 and Atg7–Atg10, respectively) (Figure 3). The detailed methodology is as follows:
Prepare Atg10 C133only cross-linked to Atg7 (1–613) C39S C195S C375A by the same procedure as that used for Atg7–Atg3 crosslinking, except that a 5× molar excess of BMOE is added to approximately 0.2 mM Atg10 C133only. Perform cross-linking with 1.2 μM Atg7 (1–613) C39S C195S C375A and 2.4 μM Atg10 C133only in the final buffer of 25 mM HEPES pH 7.0, 100 mM NaCl, and 5 mM DTT. Concentrate Atg7–Atg10 to 7 mg/mL.
To obtain wild-type or mutant version of the ATG5–ATG16L1 (1–69) complex, a baculovirus/insect cell expression system is used. Specifically, ATG5 wild type or the ATG5 E122D mutant is cloned into the pFastBac-HTb vector, which contains an N-terminal His tag fusion with ampicillin resistance (Life Technologies). ATG16L1 (1–69) is cloned into the pFastBac-hisMBP vector, which is modified by insertion of MBP between the His-tag and the TEV cleavage site from pFastBac-HTb. After amplification by DH5alpha cells, the ATG5 containing pFastBac-HTb plasmid and ATG16L1 (1–69) containing plasmid are transformed into DH10-EMBacY cells. DH10-EMBacY cells are DH10B E. coli containing a plasmid encoding the transposase and tetracycline resistance and the bacmid encoding the baculovirus genome, yellow fluorescent protein, kanamycin resistance, and ampicillin resistance. At this step, the ATG5 gene or ATG16L1 (1–69) gene, as well as a gentamycin marker, are transferred from the pFastBac-HTb plasmid into a bacmid encoding the baculovirus genome via transposase-mediated recombination, or transposition, in E. coli. When the ATG5 or ATG16L1 (1–69) gene is transposed into the bacmid, it gets inserted into the lacZα gene. The lacZα gene provides β-galactosidase complementation and turns colonies blue on plates containing X-gal. Therefore, the desired recombinant bacmids produce white colonies, whereas undesirable transformants that contain non-recombinant bacmids are blue when grown on plates containing X-gal. To verify blue/white selection, white colonies need to be re-streaked on selective plates with X-gal before proceeding to the next steps. The detailed methodology is as follows:
The initial virus (P1) is made by transfecting SF9 cells in plates with FuGENE HD (Promega) and harvesting the supernatant 3–5 days later, as described below:
Later-generation viruses (P2 and P3) are produced in the same manner by infecting a plate of SF9 cells and harvesting the medium 3 days later as follows:
Hi5 cells need to be maintained in HyQ-SFX serum-free medium, and the culture should be in the mid-log phase.
This study was supported by ALSAC, HHMI, NIH R01-GM077053, and American Heart Association 14POST19890021 (YQ).