PLoS Biology | www.plosbiology.org0161
Autophagy: A Forty-Year Search for a Missing
Gabor Juhasz*, Thomas P. Neufeld
February 2006 | Volume 4 | Issue 2 | e36
many developmental and pathological situations. Structures
targeted for autophagic destruction are sequestered
into newly emerging double-membrane vesicles called
autophagosomes, and delivered for lysosomal degradation.
Despite recent advances in understanding the molecular
mechanisms of autophagy, a long-standing question
concerning the source of the autophagic membrane remains
unresolved. Two major alternatives can be considered: the
membrane may be derived from a pre-existing cytoplasmic
organelle such as the endoplasmic reticulum (maturation
model), or assembled from constituents at its site of genesis
utophagy is the major self-degradative process in
eukaryotic cells, with fundamental roles in cellular
and organismal homeostasis, and is involved in
The stability of all biological systems—from single cells to
ecological communities—is based on the continuous turnover
of individual units. Just as new organisms are born to replace
dying ones, turnover of constituents within a cell ensures that
old or damaged macromolecules and organelles are replaced
by newly synthesized ones. This constant replacement
underlies the adaptability of biological systems, for example
allowing cells to rapidly change their metabolism in response
to a changing environment. Cellular homeostasis (the ability
to maintain a stable condition inside the cell) is therefore
based on the proper balance of synthesis and destruction.
In eukaryotic cells, various specialized cytoplasmic enzymes
are responsible for the specifi c degradation of proteins (the
ubiquitin-proteasome pathway), lipids, ribonucleic acids,
sugars etc.; these degradation events can be crucial to the
execution of cellular signaling and metabolic pathways.
In contrast, nonspecifi c degradation of these materials
occurs through autophagy (“self-eating” in Greek; in this
case at a subcellular level) . This process plays several
important roles in the life of a cell. During times of starvation,
autophagy ensures survival by randomly degrading bulk
cytoplasm including organelles to provide breakdown
products that can be used for energy and synthetic processes
. In response to a change in available nutrients, autophagy
is also used to specifi cally eliminate obsolete metabolic
organelles in several yeast species . In multicellular
organisms, autophagy is tightly controlled through several
signaling pathways [4,5], and has been integrated into
various developmental and physiological events, such as the
remodeling of cells, tissues, organs, or even the entire body.
To take an extreme example, nearly all of the larval tissues of
a metamorphosing insect are self-digested inside the pupal
case, making room and providing nutrients for the cells that
will eventually give rise to the adult insect . In mammals,
autophagy occurs in virtually all cells at a basal rate, and has
been shown to be required for the elimination of old and
nonfunctional organelles and protein complexes . Growth-
promoting hormones such as insulin inhibit autophagy ,
whereas glucagon, synthesized in response to low blood
sugar levels, induces this process . Autophagy has been
suggested to play a protective role during aging, cell death,
defense against intracellular pathogens, neurodegenerative
diseases, and tumorigenesis, emphasizing the biological and
medical importance of autophagy [2,10].
The Morphology of Autophagy
During autophagy, large membrane-bound portions of
the cytoplasm are delivered for destruction to lysosomes,
organelles loaded with various acidic hydrolases specialized
for rapid and effective degradation of cellular and
extracellular material. As early as the 1960s, the morphology
of the major pathway of autophagy, macroautophagy
(simply referred to as autophagy hereafter), was established
by electron microscopic studies [1,9]. Upon induction of
autophagy, a membrane cisterna (fold of membrane) known
as the isolation membrane (IM; sometimes referred to as
phagophore in mammals ) appears and curves around
part of the cytoplasm. Sealing of the edges of the IM results
in a unique double membrane vesicle, the autophagosome.
Soon after forming, autophagosomes fuse with a lysosome,
where degradation of the delivered material for recycling
takes place (Figure 1A and 1B). The membranes of IMs and
autophagosomes differ from other membranes in the cell
in having few intramembrane proteins, evident by electron
The dynamic membrane rearrangements of autophagy
are also unique in that topologically intracellular material
(cytoplasm) becomes converted into topologically
Unsolved Mysteries discuss a topic of biological importance that is poorly
understood and in need of research attention.
Citation: Juhasz G, Neufeld TP (2006) Autophagy: A forty-year search for a missing
membrane source. PLoS Biol 4(2): e36.
Copyright: © 2006 Juhasz and Neufeld. This is an open-access article distributed
under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in ay medium, provided the original
author and source are credited.
Abbreviations: Atg, autophagy-related gene; ER, endoplasmic reticulum; IM,
isolation membrane; PAS, preautophagosomal structure
Gabor Juhasz and Thomas P. Neufeld are in the Department of Genetics, Cell
Biology, and Development, University of Minnesota, Minneapolis, Minnesota, United
States of America
Competing Interests: The authors have declared that no competing interests exist.
* To whom correspondence should be addressed. E-mail: firstname.lastname@example.org (GJ);
PLoS Biology | www.plosbiology.org 0162
extracellular (the lumen of lysosomes) without crossing a
membrane barrier . In contrast, material delivered to the
lysosome by other routes either remains extracellular, in the
case of endocytosis (formerly referred to as heterophagy ),
or crosses a membrane, as in the case of lysosomal enzyme
Identifi cation of Genes Involved in Autophagy
Yeasts, like humans, alter their metabolism to meet the
amount and type of available nutrients. And, unlike humans,
they are much easier to study genetically. Little was known
about the molecular mechanisms of autophagy until the last
decade, when yeast genetic screens revealed ~27 autophagy-
related genes (Atg), whose products are necessary for
autophagy . Most Atg proteins localize at least transiently
to a single well-defi ned spot in the cytoplasm known as the
preautophagosomal structure (PAS), a site from which
the IM forms upon induction of autophagy. Atg proteins
form three main multiprotein complexes: an autophagy-
specifi c phosphatidyl-inositol 3-kinase (PI3K) complex, the
Atg1 kinase complex, and two closely linked ubiquitin-like
protein conjugation systems. Current research is aimed at
understanding how these complexes interact to promote
Multicellular organisms harbor homologs of most of the
yeast Atg genes, refl ecting the evolutionary conservation of
autophagy and its molecular components. There are also
some important differences. For example, IMs in yeast bud
from the single PAS, whereas in mammalian cells, IMs are
formed throughout the cytoplasm, and no defi nitive PAS is
observed. The presence of multiple homologs of some Atg
genes also suggests an as-yet unknown complexity of the
process in metazoans.
One of the ubiquitin-like proteins, Atg8 (called LC3
in mammals), appears to be associated with IMs and
autophagosomes through a lipid anchor , and therefore
represents the fi rst known protein covalently attached to
these membranes. Atg8/LC3 fused to a fl uorescent tag is now
commonly used as a light microscopic marker to follow the
formation of early autophagic structures (Figure 1C) .
Potential Membrane Sources of the IM
Despite the recent advancement of autophagy research,
one fundamental question remains unanswered: how does
the autophagic membrane form? The membranes of IMs
and autophagosomes are of the thin type (6–7 nm), like
the membranes of the endoplasmic reticulum (ER), cis-
Golgi, nuclear envelope, and inner and outer membranes
of mitochondria. In contrast, membranes of the plasma
membrane, lysosomes, and most of the Golgi are thick
(9–10 nm), due to their different lipid composition (e.g.,
more cholesterol) and higher protein content. A review
article written nearly 40 years ago by de Duve and Wattiaux
 already mentioned that the source of the sequestering
membrane “has given rise to many speculations,” including
the ER, Golgi complex, and de novo formation. Below we
consider the evidence relating to two general models of
autophagic membrane formation.
The Maturation Model
All cellular membrane is generally thought to derive from
the ER. Membrane from the ER is transported through the
secretory pathway by a continuous cycle of vesicular budding
and fusion. Based on these observations, one model posits
that the IM is derived from the ER or another pre-existing
organelle upon induction of autophagy. For example, a
portion of the ER may become cleared of ribosomes and
fold onto itself to form the IM (Figure 2A); alternatively,
vesicles may bud off from the ER and fuse together to form
the IM (Figure 2B). This theory is supported by the similar
membrane thickness of these organelles (see earlier),
by the putative identifi cation of ER proteins in IMs and
autophagosomes in mammalian cells  (but see also 
for an extended discussion of this point), and also by the
observation that IMs are often observed between parallel ER
cisternae in secretory cells . In yeast, recent molecular
genetic studies showed that genes necessary for ER traffi cking
are required for autophagy . These data are consistent
with a membrane contribution from the ER, although they
Figure 1. The Dynamic Membrane Events Involved in Autophagy
(A) Upon induction of autophagy, a membrane sac called the isolation
membrane (IM) forms and engulfs portions of the cytoplasm. Sealing of
its edges gives rise to the double-membrane bound autophagosome.
Fusion of the outer membrane with a lysosome results in formation of
an autolysosome, in which the inner autophagosomal membrane and its
contents are degraded.
(B) Starvation-induced autophagosomes (AP) and autolysosomes (AL) in
the fat body (the functional analogue of the liver) of a fruit fl y larva. Note
that APs contain intact cytoplasm, whereas the contents of ALs show
various stages of degradation.
(C) Liver cells of starved mice carrying a fl uorescently tagged LC3
transgene, labeling cup-shaped and ring-shaped structures that
correspond to IMs and autophagosomes, respectively.
Images courtesy of Ryan Scott (B) and Dr. Noboru Mizushima (C).
February 2006 | Volume 4 | Issue 2 | e36
PLoS Biology | www.plosbiology.org0163
may instead simply refl ect a genetic requirement for one or
more proteins synthesized in the ER.
The Assembly Model
Recent real-time studies have revealed that newly forming
IMs continue to elongate until the fi nal sealing of the
autophagosome . This observation is diffi cult to reconcile
with the direct maturation of an existing organelle into an
IM. An alternative model posits that the IM is assembled de
novo. In this model, nonvesicular transport (Figure 2C) or
local synthesis (Figure 2D) of lipids supplies the material for
the growing membrane. Currently, support for this model
comes mostly in the form of negative results, namely the
lack of convincing identifi cation of vesicles or membrane
cisternae fusing to the IM, despite numerous transmission
and freeze-fracture electron microscopic studies. Similarly,
conventional vesicular structures have not been observed
by electron microscopy at the PAS in yeast, although
retrograde transport of some Atg proteins has recently been
demonstrated, presumably involving a vesicular mechanism
. Interestingly, the cytoplasm in the area of the PAS
largely excludes free ribosomes that normally fi ll the cytosol
, consistent with a high lipid content or membranaceous
barrier in this region.
In many cell types, induction of autophagy can be
remarkably robust, resulting in a rapid appearance of
numerous autophagosomes throughout the cell. Whether
local synthesis or nonvesicular transport of lipids and their
assembly into membranes could be effi cient enough to
generate suffi cient amounts of membrane in a short time
is unclear. It is also possible that multiple membrane pools
contribute to the IM. The distinct steps of IM formation—
nucleation, assembly, and elongation—may also rely on
different membrane sources .
Toward a Solution
The unique molecular makeup of the autophagic
membrane—rich in lipid, poor in protein—has likely
hindered attempts to identify its source using traditional
proteocentric cellular and molecular approaches. A defi nitive
answer to this elusive question is thus likely to spring from new
technologies, such as improved in vivo lipid labeling methods
capable of specifi cally marking different cell membranes
in live cells . In addition, the recent and ongoing
identifi cation of the Atg proteins, many of which specifi cally
localize to the PAS and IM, provides a powerful new set of
tools to address this problem. This list is likely to grow as
additional factors are identifi ed, perhaps based on their co-
localization with known Atg proteins, and through genetic
screens in other organisms. Among the most promising of
the known factors is Atg9, the sole known integral membrane
protein involved in IM formation. Atg9 cycles through the
PAS and other unidentifi ed punctate structures [23,26].
As this protein is presumably membrane-bound from its
inception, live tracking of Atg9 is likely to identify at least
a subset of the contributing membrane. These reagents,
in conjunction with standard cell biology approaches, thus
represent novel means of investigating the membrane source.
For example, a specifi c pool of lipids or Atg proteins en
route to the forming and growing IM could be followed by
photobleaching recovery or pulse-chase methods [27,28]. In
addition, these tools will allow researchers to delve further
into this mysterious puzzle, addressing a number of closely
related fundamental questions: As new membrane is added
to the growing IM, is it recruited to the tips, center or entire
surface? How is this membrane targeted to the IM? Does
induction of autophagy stimulate membrane production,
or are pre-existing stores of lipid suffi cient? Are the known
Atg proteins directly involved in membrane recruitment or
synthesis, and if so what are the mechanisms involved? A new
generation of tools and a plate full of questions is attracting
new researchers to this expanding fi eld, leading to an
accelerated pace of discovery. Future work will provide plenty
of new information to ingest. ?
1. De Duve C, Wattiaux R (1966) Functions of lysosomes. Annu Rev Physiol
2. Levine B, Klionsky DJ (2004) Development by self-digestion: Molecular
mechanisms and biological functions of autophagy. Dev Cell 6: 463–477.
3. Bellu AR, Kiel JA (2003) Selective degradation of peroxisomes in yeasts.
Microsc Res Tech 61: 161–170.
4. Rusten TE, Lindmo K, Juhasz G, Sass M, Seglen PO, et al. (2004)
Programmed autophagy in the Drosophila fat body is induced by ecdysone
through regulation of the PI3K pathway. Dev Cell 7: 179–192.
5. Scott RC, Schuldiner O, Neufeld TP (2004) Role and regulation of starvation-
induced autophagy in the Drosophila fat body. Dev Cell 7: 167–178.
6. Neufeld TP (2004) Role of autophagy in developmental cell growth
and death: Insights from Drosophila. In: Klionsky DJ, editor. Autophagy.
Georgetown: Landes Bioscience. pp. 224–232.
7. Komatsu M, Waguri S, Ueno T, Iwata J, Murata S, et al. (2005) Impairment
of starvation-induced and constitutive autophagy in Atg7-defi cient mice. J
Cell Biol 169: 425–434.
8. Pfeifer U (1978) Inhibition by insulin of the formation of autophagic
vacuoles in rat liver. A morphometric approach to the kinetics of
intracellular degradation by autophagy. J Cell Biol 78: 152–167.
9. Ashford TP, Porter KR (1962) Cytoplasmic components in hepatic cell
lysosomes. J Cell Biol 12: 198–202.
10. Reggiori F, Klionsky DJ (2005) Autophagosomes: Biogenesis from scratch?
Curr Opin Cell Biol 17: 415–422.
11. Seglen PO (1987) Regulation of autophagic protein degradation in isolated
liver cells. In: Glaumann H, Ballard FJ, editors. Lysosomes: Their role in
protein breakdown. London: Academic Press. pp. 369–414.
12. Rez G, Meldolesi J (1980) Freeze-fracture of drug-induced autophagocytosis
in the mouse exocrine pancreas. Lab Invest 43: 269–277.
13. Kovacs AL, Rez G, Palfi a Z, Kovacs J (2000) Autophagy in the epithelial
cells of murine seminal vesicle in vitro. Formation of large sheets of nascent
isolation membranes, sequestration of the nucleus and inhibition by
wortmannin and 3-methyladenine. Cell Tissue Res 302: 253–261.
14. Fengsrud M, Erichsen ES, Berg TO, Raiborg C, Seglen PO (2000)
Ultrastructural characterization of the delimiting membranes of isolated
autophagosomes and amphisomes by freeze-fracture electron microscopy.
Eur J Cell Biol 79: 871–882.
Figure 2. Possible Membrane Sources of the IM
Some of the possible scenarios for IM formation are illustrated here.
According to the maturation model, upon induction of autophagy,
membrane may be directly derived from the ER by folding (A), or in
the form of vesicular transport (B). In the assembly model, membrane
may be assembled de novo at the site of IM formation, originating from
nonvesicular transport, such as micellar, as shown in (C), or local synthesis
(D). See text for further details.
February 2006 | Volume 4 | Issue 2 | e36
PLoS Biology | www.plosbiology.org0164
15. Stromhaug PE, Klionsky DJ (2001) Approaching the molecular mechanism
of autophagy. Traffi c 2: 524–531.
16. Klionsky DJ, Cregg JM, Dunn WA Jr, Emr SD, Sakai Y, et al. (2003) A unifi ed
nomenclature for yeast autophagy-related genes. Dev Cell 5: 539–545.
17. Ichimura Y, Kirisako T, Takao T, Satomi Y, Shimonishi Y, et al. (2000) A
ubiquitin-like system mediates protein lipidation. Nature 408: 488–492.
18. Mizushima N, Yamamoto A, Matsui M, Yoshimori T, Ohsumi Y (2004)
In vivo analysis of autophagy in response to nutrient starvation using
transgenic mice expressing a fl uorescent autophagosome marker. Mol Biol
Cell 15: 1101–1111.
19. Dunn WA (1994) Autophagy and related mechanisms of lysosome-mediated
protein degradation. Trends Cell Biol 4: 139–143.
20. Fengsrud M, Sneve ML, Overbye A, Seglen PO (2004) Structural aspects
of mammalian autophagy. In: Klionsky DJ, editor. Autophagy. Georgetown
(Texas): Landes Bioscience. pp. 11–25.
21. Ishihara N, Hamasaki M, Yokota S, Suzuki K, Kamada Y, et al. (2001)
Autophagosome requires specifi c early Sec proteins for its formation and
NSF/SNARE for vacuolar fusion. Mol Biol Cell 12: 3690–3702.
22. Mizushima N, Yamamoto A, Hatano M, Kobayashi Y, Kabeya Y, et al. (2001)
Dissection of autophagosome formation using Apg5-defi cient mouse
embryonic stem cells. J Cell Biol 152: 657–668.
23. Reggiori F, Tucker KA, Stromhaug PE, Klionsky DJ (2004) The Atg1-
Atg13 complex regulates Atg9 and Atg23 retrieval transport from the pre-
autophagosomal structure. Dev Cell 6: 79–90.
24. Kirisako T, Baba M, Ishihara N, Miyazawa K, Ohsumi M, et al. (1999)
Formation process of autophagosome is traced with Apg8/Aut7p in yeast. J
Cell Biol 147: 435–446.
25. Kuerschner L, Ejsing CS, Ekroos K, Shevchenko A, Anderson KI, et al.
(2005) Polyene-lipids: A new tool to image lipids. Nat Methods 2: 39–45.
26. Noda T, Kim J, Huang WP, Baba M, Tokunaga C, et al. (2000) Apg9p/
Cvt7p is an integral membrane protein required for transport vesicle
formation in the Cvt and autophagy pathways. J Cell Biol 148: 465–480.
27. Lippincott-Schwartz J, Altan-Bonnet N, Patterson GH (2003)
Photobleaching and photoactivation: Following protein dynamics in living
cells. Nat Cell Biol (Suppl): S7–S14.
28. Gaietta G, Deerinck TJ, Adams SR, Bouwer J, Tour O, et al. (2002)
Multicolor and electron microscopic imaging of connexin traffi cking.
Science 296: 503–507.
February 2006 | Volume 4 | Issue 2 | e36