Macroautophagy

Macroautophagy is a nearly universal process that eukaryotic cells employ to reutilize the constituents of cytoplasm and organelles. We have used the genetics of Dictyostelium to isolate mutants in genes that are essential for macroautophagy. These mutants have taught us how to design a screen and a selection for other genes that regulate macroautophagy and also provide material for studying autophagosome formation. A collection of mutants has been used to show that macroautophagy, previously thought to be essential for replication of the intracellular pathogen Legionella pneumophila, is not required for growth of this parasite. We now are asking how autophagy is linked to development and whether there are additional genes that are essential for autophagy.

Macroautophagy: Macroautophagy as a cell biological and structural phenomenon has been known for many years to manage the bulk degradation of cytoplasm and organelles. One of its interesting features is that it promotes the transfer of material from one topologically distinct compartment to another - from the cytosol to the vacuole in baker’s yeast or to lysosomes in other eukaryotic cells. An initiating structure called the Pre-autophagosomal Structure or PAS has been defined, but its origins, once thought to be from the endoplasmic reticulum, are not clear. The favorable vacuolar structure of Saccharomyces cerevisiae has been married to the genetics of that organism to understand the biochemical reactions involved in the production of autophagosomes (for reviews see (Abeliovich and Klionsky, 2001; Klionsky and Emr, 2000; Thumm, 2000). S. cerevisiae contains a single large central vacuole, which serves as lysosome, among other functions. When autophagy is induced, double-membrane autophagosomes are produced, and these fuse with the vacuole and release single membrane autophagic bodies that are degraded by resident hydrolases and proteases. When vacuolar proteases are inhibited by PMSF, one observes an accumulation of small spheres (autophagic bodies) within a larger vacuole. Using several screening methods to create mutants in which an accumulation of the small autophagic bodies is not observed, several laboratories, notably those of Ohsumi, Thumm, and Klionsky, have isolated autophagy mutants (called atg). The affected genes have been used to define a series of phosphorylation and ubiquitination-like reactions that are necessary for cytoplasm to vesicle transport (CVT) and macroautophagy. CVT encapsulates specific targets during growth, while autophagy envelopes bulk cytoplasm and organelles during starvation. Many of the molecules that mediate autophagy reside in the Pre-autophagosomal structure (PAS, also called the perivacuolar compartment (PVC)), which in budding yeast lies next to the vacuole (Suzuki et al., 2001) (Kim et al., 2002) (Noda et al., 2002). Gradient and immuno-purification studies suggest that the PAS does not contain markers from other compartments of the cell (Kim et al., 2002). The creation of an autophagosome is thought to occur by extension of an enveloping membrane, called an isolation membrane, which then closes around cytoplasmic constituents, as shown in Figure 1 (Kim et al., 2002). A similar situation may occur in mammalian cells (Mizushima et al., 2001). Autophagosomes then fuse with late endosomes or lysosomes to create autophagolysosomes and provide constituents and energy for the cell. Although there has been progress in understanding this process, especially in determining the genes involved and mapping known gene products onto physical structures, a number of questions remain. What is the origin of the PAS and of what is it composed, beyond the known molecules? Does it assemble de novo or does it derive from pre-existing membranes? How does the PAS convert into an isolation membrane and how does that membrane seal around its cargo? What is the source of the lipid that forms the membrane? How is size determined? Is there selectivity of cargo? What are the molecules that mediate fusion with the endosomal system?

The formation and fate of an autophagosome. Autophagosomes form from a pre-autophagic structure that contains many of the proteins know to be necessary for autophagy. The PAS extends to envelop cytoplasmic constituents, generating a structure call an isolation membrane. This structure then closes and is directed to the endocytic pathway. A hallmark of autophagosomes is their double membrane structure. The PAS is not known to be membrane bound (Noda et al., 2002). Autophagosomes are destined for destruction and therefore must constantly be created de novo. The diagram illustrates the situation in a mammalian or Dictyostelium cell and does not show the yeast vacuole.

Dictyostelium discoideum: The social amoeba D. discoideum develops while starving. The amoebae cannot perform the successive steps of development – chemotaxis, cell adhesion, migration, and spore formation without recycling components that were accumulated during the vegetative period of the life cycle (Kessin, 2001). This life style makes it an alternative, genetically tractable tool to study autophagy. Dictyostelium cells that are capable of macroautophagy survive at least several weeks of starvation, whereas mutants die (Otto et al., 2003, Otto et al, 2004). Dictyostelium workers think of spore formation as the paramount survival mechanism, but autophagy is likely to be even more important, particularly when single cells cannot make multicellular structures.

A great deal is known about the development of Dictyostelium – in particular the chemotaxis of the cells toward cAMP has been well studied - and many of the genes involved in the various stages have been cloned and examined (see (Kessin, 2001) for a summary). Although development has been studied in detail, one question has not been addressed – how is it fueled? Where do cells derive the energy and chemical resources to accomplish all of the movements, slug migrations, and spore formation, let alone the synthesis of new gene products, while they are starving and how is this process regulated? We can now ask how autophagy functions in Dictyostelium and how it is regulated during development. Genetic tools, including transformation with a variety of vectors containing stage-specific promoters and markers, are available. Insertional mutagenesis and other techniques are routine. Cells that carry mutations that cause a block in the formation of spores, such as autophagy mutants, do not affect vegetative growth and can be stored as frozen amoebae. The 34 MB genome is essentially completely sequenced. Dictyostelium has about 10,000 genes, considerably more than S. cerevisiae. Methods to isolate and characterize mutants that are not dependent on prior isolation from yeast have been developed.

Some characteristics of Dictyostelium autophagy mutants are shown in the figures below.

Cell survival. Cells in amino acid-free FM medium survive for at least 10 days if their autophagy pathway is intact, but die if it is not. FM is the defined minimal medium originally invented by Franke and Kessin (Franke and Kessin, 1977).

Mutating and complementing the atg5 and atg7 genes. Insertions into the genes cause severe disruption of development (A. wild type, C, atg5; E, atg7). atg5 does not develop at all in plaques on bacterial lawns (B). When the gene is restored under the control of a constitutive promoter, development is normal (D, atg5/act15::Atg5-GFP; F, atg7/act15::GFP-Atg7). Only one aggregateless phenotype is shown (atg5); atg7 is identical, as is atg1.

Fine structure of wild-type (A), atg1 mutant (B), and complemented (C) atg1 cells starved in FM medium lacking amino acids for 44 hours. (A) Wild-type cells exhibit a sparsely granular cytoplasm, few mitochondria, and vacuoles containing membranous and partially degraded granular cytoplasm indicative of autophagic activity. (B) Mutant cells contain a richly granular cytoplasm, relatively more mitochondria, and occasional vacuoles containing cytoplasmic organelles and membranous whorls. Scale bars = 1 µm.

Autophagy mutants sediment differently from wild-type on Percoll Gradients: The observation that the cytoplasm of the starving wild-type cells differs from that of starving autophagy mutants, suggested to us that we could separate mutant and wild-type cells on gradients. Using Percoll, we showed this to be the case. On preformed Percoll gradients, the autophagy mutants sediment much less quickly than the similarly starved wild type. This is true of all of the autophagy mutants tested. Vegetative (well fed) autophagy mutants behave like vegetative wild-type cells. Mixed populations produce two bands. This dramatic separation is the basis of a selection for mutants that cannot turn over their cellular constituents, either because they lack a functional component of autophagy, or because it is never induced in the starving cells.

Starved autophagy mutants are separable for starved wild-type cells on preformed Percoll gradients (Ratner and Borth, 1983). Autophagy mutants retain the sedimentation properties of vegetative cells, but make a slightly more diffuse band.

Transmission electron microscopic analysis of the incomplete autophagic structure present in atg1 mutants. Note the many small vesicles that are lined up as if on an underlying cable. Panel B is an expanded view of the box in panel A. An autophagosome containing an undigested mitochondrion and other material is on the right. Note double membranes. The bar represents 1µm.

Conclusion: We have developed a genetically tractable system to isolate mutants with defects in macroautophagy and these are being exploited to determine how the process is regulated during development.