ATG7 is dispensable for LC3–PE conjugation in thioglycolate-elicited mouse peritoneal macrophages

ABSTRACT Thioglycolate-elicited macrophages exhibit abundant conjugation of LC3 with PE (LC3-II). Among other autophagy-related (ATG) proteins, it is proposed that, like in yeast, both ATG5 and ATG7 are essential for LC3 conjugation. Using atg5-deficient (−/-) and atg7−/-macrophages, we provide evidence that loss of ATG5 but not of ATG7 resulted in LC3-II depletion. Accumulation of LC3-II in elicited atg7−/- macrophages in response to bafilomycin A1 validated these data. Furthermore, complete loss of ATG3 in atg7−/- macrophages demonstrated that ATG7 and ATG3 are dispensable for LC3–PE conjugation. In contrast to thioglycolate-elicited macrophages, naïve peritoneal and bone marrow-derived atg7−/- macrophages exhibited no LC3-II, even under inflammatory stimuli in vitro. Hence, the macrophage metabolic status dictates the level of LC3–PE conjugation with a supportive but nonessential role of ATG7, disclosing the eukaryotic exception from the LC3 lipidation model based on yeast data. Abbreviations: ATG: autophagy-related; BM: bone marrow; MAP1LC3/LC3: microtubule-associated protein 1 light chain 3; PE: phosphatidylethanolamine.

We investigated the consequences of ATG5 and ATG7 deficiency on LC3-II formation in murine macrophages which utilize LC3-II for canonical and/or non-canonical function(s) [2] and found diverging results between thioglycolate-elicited and non-elicited cells.

Loss of ATG5 but not ATG7 depletes LC3-II in thioglycolate-elicited macrophages
Loss of ATG5 or ATG7 protein (Figure 1Ai,Aii,Di,Dii) and their respective mRNAs ( Figure 1C) in macrophages proves the successful generation of the knockout mice. As expected, deletion of Atg5 resulted in substantial expression of LC3-I and absence of LC3-II (Figure 1Ai, Fig. S1A). However, lack of ATG7 in thioglycolate-elicited macrophages failed to inhibit LC3 lipidation with only minor reduction in LC3-II compared to WT cells ( Figure  1Aii, Fig. S1B,C). LC3-II was also found in thioglycolate-elicited atg7 −/macrophages 2 h after isolation, indicating that lipidation was not induced during culturing of the cells (Fig. S1C). After separation of cytosolic and membrane fractions, LC3-II was absent in the membrane fraction of thioglycolate-elicited atg5 −/macrophages (Figure 1Bi), whereas the amount of LC3-II in atg7 −/macrophages was comparable to WT cells. This is an unexpected finding, considering that LC3 lipidation is generally accepted as being dependent on both ATG5 and ATG7 [6].
Atg10 mRNA ( Figure 1C) and ATG10 protein expression ( Figure 1D) were comparable between all genotypes. However, ATG3 protein was depleted in atg7 −/macrophages ( Figure 1D) but only reduced by 25% at mRNA level ( Figure 1C), indicating posttranscriptional regulation of ATG3. Despite reduced Atg3 mRNA ( Figure 1C), we observed significantly more ATG3 protein in atg5 −/compared to control macrophages ( Figure 1Dii). Whether ATG5 protein hinders the translation of Atg3 or whether ATG3, as an E2-like enzyme, increases its activity to compensate for the lack of ATG5 requires further investigation. However, ATG3 depletion in atg7 −/macrophages indicated that the E2-like enzyme ATG3 is redundant without the E1-like enzyme ATG7, similar to Atg7-and Atg3-independent autophagy in Drosophila [8].
We conclude that the E2-like enzyme ATG10 does not require ATG7 for ATG12 activation and conjugation to ATG5 ( Figure 1E). Subsequent conjugation of PE to LC3 is possible in the absence of ATG7 and ATG3. Thus, the E3-like ATG12-ATG5 complex is sufficient for conjugation of PE to LC3 in murine thioglycolate-elicited macrophages.

Lipidated LC3 accumulates in thioglycolate-elicited peritoneal atg7 −/macrophages in response to bafilomycin A 1
To investigate LC3 lipidation dynamics in thioglycolateelicited peritoneal WT and atg7 −/macrophages in more detail, we starved (PBS) the cells to induce autophagy and used bafilomycin A 1 to inhibit autolysosomal LC3 degradation. We observed time-dependent accumulation of LC3-II in elicited WT and atg7 −/macrophages ( Figure 2A), reduction by starvation in both genotypes ( Figure 2B), and no induction of LC3-II in atg5 −/macrophages ( Figure 2B). To overcome individual differences in LC3 expression between independent experiments, we used LC3-II:LC3-I ratios instead of normalization to a housekeeping protein.

LC3-II is absent in non-elicited peritoneal and bone marrow-derived atg7 −/macrophages
To examine whether other inflammatory agents stimulate LC3 lipidation in atg7 −/macrophages in vivo, we compared LC3 protein expression in peritoneal macrophages elicited by thioglycolate, proteose-peptone, and concanavalin A ( Figure 3A). LC3-II was hardly detectable in proteosepeptone elicited cells but we observed LC3-II in atg7 −/macrophages elicited by concanavalin A (Figure 3Ai). However, if compared with LC3-II from thioglycolateelicited atg7 −/cells, LC3-II expression was significantly reduced in concanavalin A-elicited macrophages ( Figure   3Aii). Moreover, only thioglycolate-elicited atg7 −/macrophages exhibited significantly higher LC3-II than LC3-I expression, in contrast to WT macrophages showing increased LC3-II by all eliciting agents. We conclude that among investigated inflammatory stimuli peritoneal injection of thioglycolate induces LC3 lipidation in atg7 −/macrophages most effectively. However, low expression of LC3-II in cells isolated after concanavalin A injection may be a result of a reduced macrophage response, considering that thioglycolate is more potent than other eliciting agents to induce peritonitis and yield a greater number of cells in the peritoneum [9]. These results indicated that pronounced inflammation is necessary to boost LC3 lipidation in atg7 −/macrophages.
In line with these results, non-elicited peritoneal and bone marrow (BM)-derived atg7 −/macrophages were unable to conjugate PE to LC3 ( Figure 3B). As expected, LC3-II was absent in all types of atg5 −/macrophages. If compared with non-elicited peritoneal and BM-derived macrophages, the LC3-II:LC3-I ratios were drastically higher in thioglycolateelicited WT and atg7 −/macrophages. Consistent with data in Figure 2B, the LC3-II:LC3-I ratio was markedly reduced in elicited atg7 −/compared with WT macrophages.
We conclude that the macrophage activation status dictates the level of LC3-PE conjugation. In contrast to ATG5, the role of ATG7 and ATG3 in this process is supportive but not essential.

Animals and cells
Mice with a targeted deletion of Atg5 or Atg7 in myeloid cells were generated by crossing [17]  Peritoneal macrophages were obtained 3 days after intraperitoneal injection of 2 ml of 3% thioglycolate broth, 2 ml of 3% proteose peptone or 25 µg/ml of concanavalin A. Macrophages were isolated by peritoneal lavage with 10 ml ice-cold PBS/EDTA and cultured in high glucose (25 mM) DMEM supplemented with 10% LPDS (Sigma-Aldrich, S5394), 1% penicillin-streptomycin for 2 h to adhere to the non-treated culture dish. Thereafter, cells were washed twice with PBS and cultured in DMEM for 48 h prior to further use. For the bafilomycin A 1 treatment, cells were incubated as described before for 46-or 34 h and then coincubated with 10 nM bafilomycin A 1 for 2 or 14 h, respectively. PBS-starved cells were incubated as described before for 47 h and then incubated in PBS for 1 h. Cells were harvested simultaneously with untreated cells. BM-derived cells were differentiated to macrophages for 7 days in DMEM containing 10% of conditioned medium from L929 cells (ECACC, 85011425). Cells were lysed in RIPA buffer in the presence of 1 mM DTT and protease inhibitor cocktail. For subcellular fractionation, cells were lysed in fractionation lysis buffer by triturating cells through a 27 G needle with visual inspection of successful lysis under the microscope. After separation of the nuclear fraction (2,655 x g), the membrane fraction was obtained by centrifugation at 100,000 x g, while the cytosolic fraction was collected as supernatant. For in vitro treatments of naïve peritoneal and BM-derived macrophages, 10 nM bafilomycin A 1 , 100 ng/ml lipopolysaccharide, 5 ng/ml interferon-gamma or 0.3% thioglycolate broth were used. Conditioned media were collected from thioglycolate-elicited WT macrophages after culturing for 24 h in DMEM.