and 1 more

This report was spearheaded by Madlen Stephanie, with input from all the members of the Dagdas lab.   In this manuscript, Teh et al. investigated the role of exocyst subunit Exo70B2 in immune responses and secretion in plants. They showed that a subpopulation of Exo70B2 undergoes autophagic degradation. Recruitment of Exo70B2 to the autophagosomes are mediated by two ATG8 Interaction Motifs (AIM). Furthermore, they have shown that an immune activated MAPK phosphorylates Exo70B2 and enhances its binding affinity to the ATG8. Finally, they saw that yeast Exo70 homolog also interacts with ATG8 and undergoes autophagic recycling. Overall it is an interesting study, that (i) provides the first glimpses of a potential “exocystphagy” response and (ii) paves the way for future studies that could address the role of exocyst mediated secretion in immune responses. Although the authors generally used multiple lines of evidence to support their hypotheses, we had some concerns that if addressed could improve the manuscript:  1.     We felt that the title could have been more informative and precise. In a way, this story contains two exciting stories. It provides evidence, although not fully compelling, that Exo70B2 could function as an autophagy receptor to mediate recycling of the octameric exocyst complex. It also provides evidence that PM-associated Exo70B2 is phosphorylated by MPK3, and thereby regulates its function (the function is not defined here).  2.     It would have been useful if the authors used AIM mutants (1/2 double mutant) as controls in their autophagy studies. Especially, it would have been very useful to show that AIM1/2 double mutant doesn’t undergo autophagic degradation. Similarly, since it is known that BiFC empty vector controls have stability issues, AIM1/2 double mutant could have been a better control. Finally, infection studies would have benefited from the AIM1/2 double mutant. 3.     The authors chose an unorthodox way of doing the bacterial CoIPs. The inputs are separate proteins, rather than the protein samples mixed for CoIP experiments. In our opinion, the mixture of proteins which is used in the subsequent pull-down should be shown as input rather than the separate proteins only (see Fig. 4C,5C+F). 4.     It would be useful to stay consistent on the presentation of the IP figures. Rearranging the input and the pull-down blot in each picture confuses the reader (compare fig.4C with 5C and 5F). 5.     There is an empty input lane in Figure 5C. This figure would have benefited a lot by having the AIM mutant as a control as well as the WT protein. 6.     Labelling of cotyledons and roots were mixed in Fig.S5B.    
This report was prepared by Vienna Biocenter Summer Class 2017 PhD students as a part of their Priming Your PhD training. Please see the instructors and participating students’ names at the end of the review.   Shoemaker et al. performed a systematic genome wide CRISPR screen and identified a catalogue of factors involved in mammalian autophagy. They validated their findings by further characterizing one of the newly identified autophagy-related factors in their screen. They showed that TMEM41B, an integral ER-membrane protein is involved in maturation of the phagophore. Furthermore, using NBR1 as an autophagy substrate, they discovered a number of genes involved in a novel non-canonical autophagy pathway, which is independent of ATG7. The comprehensive results obtained in this study make the picture of mammalian autophagy more complete, and provide entry points for future studies that will dissect alternative autophagy pathways and the molecular mechanism underlying the role of TMEM41B in canonical autophagy. Genome wide CRISPR-screen provides a state of the art, unbiased approach to discover unknown genes in the autophagy process. They have developed a sensitive autophagic flux measurement reporter using the tandem-fluorescent reporters. In addition to the commonly used LC3B, they have also used known cargo receptors to identify autophagy regulators. Using this reporter system as a readout for a genome-wide screen is opening up the possibility to systematically dissect complex cellular autophagy networks. The authors were able to validate their screening approach by recovering almost all known ATG factors. Interestingly, the screen also identified a number of completely new candidates. While the authors have initially setup screens to address the pathways involved in the five major autophagy receptors, their last screening setup also allows to dissect the pathway involved in the highly debated ATG7-independent lysosomal targeting. We find this particularly interesting as is significantly contributes to our general understanding of how the network of autophagy receptors is organized. Furthermore, the description of ATG-independent alternative autophagy pathways questions the use of ATG7 KO genotypes as a control in autophagy experiments. As autophagy may still be active in ATG7 KO, this should be of concern for future experimental designs. The authors extensively stress-tested their reporter systems prior to performing the screen. They have also integrated all the reporters in the same loci to prevent expression level differences. The authors, in most cases, transparently provide numbers of cells used in the experimental setup and respective controls. Throughout the paper the authors confirm and cross-reference results obtained from other researchers in the field. They have used a broad range of experimental tools to validate their screen and results. Overall, the manuscript is of very high quality. Below are some suggestions that we hope will improve the manuscript. 1.     The authors have chosen a 60% interval for their FACS sorting range. It would be useful if they provide the original FACS plots (RFP/GFP ratio), in order for us to see how the distribution of non-infected and library-infected populations look like. This would explain why screening for 60% of the population is necessary/reasonable. Although they have used a published sgRNA library, it would have been useful if they provided more information such as how many sgRNAs/gene are in the initial library, library representation in the transduced cell pool, replicate correlation plots. In addition, previous CRISPR screen papers have used fold changes for quantifying their hits. It would be useful if the authors explain the rationale behind beta scores and give more detail on beta score calculation. 2.     There is also no information about Cas9 clonality in the K562 cells which were used for the screen. This would be useful for readers. 3.     In the validation experiments, the authors state “mock treated cells” as negative control – does this mean that no sgRNA is transduced? If so, this is not an applicable negative control, as biological response to transduction and DNA damage repair by introducing an sgRNA is not addressed. Therefore, using a sgRNA targeting a gene desert would be the correct control. 4.     Information how they normalized to 1.0 in Fig 3A is missing, as well as additional labels on Y-Axis that would enable the reader to understand whether this is a linear or log-scale. 5.     In Figure 6, Stx17 and one of the cargo receptors could have been nice controls. Additionally, it would have been nice to have EM images of autophagosomes in TMEM41B KO cells. 6.     Due to the design of the screen, one would have expected to get some hits related to lysosomal acidification. The authors did not mention this in the results or discussion. It would be nice to at least discuss this. Similarly, the authors should compare and contrast DeJesus et al., 2017 eLife paper in the discussion. This paper uses a similar CRISPR approach and looks for autophagy regulators. 7.     In the introduction, it would have been beneficial to introduce the SQSTM1 receptors in a bit more detail, since the screen is based on them.   8.     It would be useful if they clarified why they decided to use TMEM41B, since it wasn’t a very strong hit for all the receptors tested. 9.     The description of the cell line used for the screen was unclear; were single clones or a bulk population used? 10.  Figure S6 is missing; the supplementary figures go from Figure S5 to S7. 11.  In the text, RAB7A and HOPS are mentioned while referring to figure 7E. However, these factors cannot be found in the figure. 12.  The last paragraph of the results section refers to an analysis about genetic interactions between ATG7 and several factors, but no figure is ever referenced for this section. 13.  Figure 3A: in the legend, more information could be given as to what exactly is plotted and less regarding the supplementary data. It is difficult to determine the scale of the Y-axis as only two values are given.   PhD Students: Krista Briedis, Alexander Bykov, Claudia Ctortecka, Melanie         De Almeida, Philipp Dexheimer, Joachim Garbrecht, Sarah Gruenbacher, Bence Hajdusits, Felix Holstein, Bhagyshree Jamge, Friederike Leesch, Joanna Nowacka, Mina Petrovic, Anna Schmuecker, Monika Steininger, Pietro Tardivo, Szu-Hsien Wu Facilitators: Fumiyo Ikeda, Yasin Dagdas