Adipose triglyceride lipase and the lipolytic catabolism of cellular fat stores.
ABSTRACT Fatty acids (FAs) are essential components of all lipid classes and pivotal substrates for energy production in all vertebrates. Additionally, they act directly or indirectly as signaling molecules and, when bonded to amino acid side chains of peptides, anchor proteins in biological membranes. In vertebrates, FAs are predominantly stored in the form of triacylglycerol (TG) within lipid droplets of white adipose tissue. Lipid droplet-associated TGs are also found in most nonadipose tissues, including liver, cardiac muscle, and skeletal muscle. The mobilization of FAs from all fat depots depends on the activity of TG hydrolases. Currently, three enzymes are known to hydrolyze TG, the well-studied hormone-sensitive lipase (HSL) and monoglyceride lipase (MGL), discovered more than 40 years ago, as well as the relatively recently identified adipose triglyceride lipase (ATGL). The phenotype of HSL- and ATGL-deficient mice, as well as the disease pattern of patients with defective ATGL activity (due to mutation in ATGL or in the enzyme's activator, CGI-58), suggest that the consecutive action of ATGL, HSL, and MGL is responsible for the complete hydrolysis of a TG molecule. The complex regulation of these enzymes by numerous, partially uncharacterized effectors creates the "lipolysome," a complex metabolic network that contributes to the control of lipid and energy homeostasis. This review focuses on the structure, function, and regulation of lipolytic enzymes with a special emphasis on ATGL.
- SourceAvailable from: Choi Jeong-yi[Show abstract] [Hide abstract]
ABSTRACT: Nucleoredoxin (NRX) is a member of the thioredoxin family of proteins that controls redox homeostasis in cell. Redox homeostasis is a well-known regulator of cell differentiation into various tissue types. We found that NRX expression levels were higher in white adipose tissue of obese ob/ob mice and increased in the early adipogenic stage of 3T3-L1 preadipocyte differentiation. Knockdown of NRX decreased differentiation of 3T3-L1 cells, whereas overexpression increased differentiation. Adipose tissue-specific NRX (Adipo-NRX) transgenic mice showed increases in adipocyte size as well as number compared to wild-type mice. We further confirmed that the Wnt/β-catenin pathway was also involved in NRX-promoted adipogenesis, consistent with a previous report showing NRX regulation of this pathway. Genes involved in lipid metabolism were downregulated, whereas inflammatory genes, including those encoding macrophage markers, were significantly upregulated, likely contributing to the obesity in Adipo-NRX mice. Our results therefore suggest that NRX acts as a novel proadipogenic factor and controls obesity in vivo. Copyright © 2014, The American Society for Biochemistry and Molecular Biology.The Journal of Lipid Research 12/2014; · 4.73 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: New information concerning the effects of prolactin (PRL) on metabolic processes warrants reevaluation of its overall metabolic actions. PRL affects metabolic homeostasis by regulating key enzymes and transporters associated with glucose and lipid metabolism in several target organs. In the lactating mammary gland, PRL increases the production of milk proteins, lactose, and lipids. In adipose tissue, PRL generally suppresses lipid storage and adipokine release and affect adipogenesis. A specific case is made for PRL in the human breast and adipose tissues, where it acts as a circulating hormone and an autocrine/paracrine factor. Although its overall effects on body composition are both modest and species-specific, PRL may be involved in the manifestation of insulin resistance.Advances in Experimental Medicine and Biology 01/2015; 846:1-35. · 2.01 Impact Factor
Article: AIMing at Metabolic SyndromeCirculation Journal 01/2011; 75(11):2522-2531. · 3.69 Impact Factor
Adipose triglyceride lipase and the lipolytic catabolism of
cellular fat stores
Rudolf Zechner,1Petra C. Kienesberger, Guenter Haemmerle, Robert Zimmermann,
and Achim Lass
Institute of Molecular Biosciences, University of Graz, Austria
lipid classes and pivotal substrates for energy production in
all vertebrates. Additionally, they act directly or indirectly as
signaling molecules and, when bonded to amino acid side
chains of peptides, anchorproteins in biological membranes.
In vertebrates, FAs are predominantly stored in the form of
triacylglycerol (TG) within lipid droplets of white adipose
tissue. Lipid droplet-associated TGs are also found in most
nonadipose tissues, including liver, cardiac muscle, and
skeletal muscle. The mobilization of FAs from all fat de-
pots depends on the activity of TG hydrolases. Currently,
three enzymes are known to hydrolyze TG, the well-studied
hormone-sensitive lipase (HSL) and monoglyceride lipase
(MGL), discovered more than 40 years ago, as well as the
relatively recently identified adipose triglyceride lipase
(ATGL). The phenotype of HSL- and ATGL-deficient mice,
as well as the disease pattern of patients with defective ATGL
activity (due to mutation in ATGL or in the enzymeʼs activa-
tor, CGI-58), suggest that the consecutive action of ATGL,
HSL, and MGL is responsible for the complete hydrolysis
of a TG molecule. The complex regulation of these enzymes
by numerous, partially uncharacterized effectors creates the
“lipolysome,” a complex metabolic network that contributes
to the control of lipid and energy homeostasis.
focuses on the structure, function, and regulation of lipolytic
Kienesberger, G. Haemmerle, R. Zimmermann, and A. Lass.
Adipose triglyceride lipase and the lipolytic catabolism of cel-
lular fat stores. J. Lipid Res. 2009. 50: 3–21.
Fatty acids (FAs) are essential components of all
Supplementary key words
lipolysis • hydrolase • neutral lipid stor-
Lipid homeostasis reflects a balance of processes, de-
signed to generate fatty acids (FAs) and lipids, deliver
them from their site of origin to target tissues, and catabo-
lize them for metabolic purposes. Innumerable genes and
signal components are responsible for an integrated com-
munication network between many tissues and organs, in-
cluding adipose tissue, liver, muscles, the digestive tract,
pancreas, and the nervous system. This network ultimately
accounts for the accurate regulation of lipid and energy
homeostasis. Despite the central physiological importance
of these processes for human health, many basic mecha-
nisms regulating the synthesis, uptake, storage, and utiliza-
tion of lipids remain insufficiently characterized.
FAs are vital components of essentially all known organ-
isms. They are important substrates for oxidation and the
production of cellular energy. FAs are essential precursors
for all lipid classes, including those forming biological
membranes. Finally, they are important for protein func-
tion in acylated proteins and as ligands for nuclear recep-
tor transcription factors. In contrast to these “beneficial”
characteristics, unesterified FAs can become deleterious
for cells when present even at relatively low concentra-
tions. The chronic exposure of nonadipose cells and tis-
sues to elevated concentrations of FAs triggers adverse
effects subsumed under the term of “lipotoxicity” (1, 2).
Accordingly, when supplied with excessive nutrients, essen-
tially all eukaryotes reesterify and deposit FAs as triacylglyc-
erol (TG) droplets to provide an energy reserve for times
of nutrient deprivation and to detoxify otherwise harm-
Until recently, lipid droplets were viewed as an inert
storage pool of TG. It is now known that essentially all cells
in the body generate lipid droplets composed of neutral
lipids (TG and cholesteryl esters), phospholipids, and un-
esterified cholesterol at varying, tissue-specific concen-
trations. Additionally, numerous proteins are associated
with lipid droplets (3–5). These include structural pro-
teins, lipid-modifying enzymes, and proteins that regulate
enzyme activities. To date, the physiological role of many
of these factors remains elusive. However, from the limited
knowledge that is available, it is apparent that lipid drop-
lets represent remarkably flexible, dynamic organelles that
are used for the production of membrane components,
This research was supported by a grant from Genomics of Lipid-Associated Dis-
orders (GOLD), part of the Austrian Genome Project, Genome Research in Austria
(GEN-AU), funded by the Austrian Ministry of Science and Research and by the
Austrian Science Foundation (FWF), Grants W901-B05DK (Doktoratskolleg
Molecular Enzymolgy) and F30-B05 (SFB Lipotox).
Manuscript received 6 October 2008 and in revised form 23 October 2008.
Published, JLR Papers in Press, October 23, 2008.
1To whom correspondence should be addressed.
Copyright ©2009 by the American Society for Biochemistry and Molecular Biology, Inc.
This article is available online at http://www.jlr.org
Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
energy substrates, and signaling molecules, including lipo-
toxic compounds (6, 7). Although lipid droplets are ob-
served in many cell types, the majority of fat in mammals
is found in adipocytes of white adipose tissue (WAT). The
central contribution of WAT to the regulation of energy
homeostasis is due to both the enormous lipid storage
capacity as well as its function as an endocrine organ se-
creting numerous hormones and adipo-cytokines (8). Preva-
lent metabolic diseases such as obesity and type 2 diabetes
emerge when TG synthesis and catabolism lose synchrony.
The key process in fat catabolism and the provision of
energy substrate during times of nutrient deprivation (fasting)
or enhanced energy demand (e.g., exercise) is the hydrolytic
cleavage of stored TG, the generation of FAs and glycerol,
and their release from adipocytes. A complex, hormonally
controlled regulatory network controls the initiation of
this process, called lipolysis, and ultimately activates key
intracellular lipases to hydrolyze TG. Currently, three en-
zymes are known to have an established function in the
lipolytic breakdown of fat in adipose and nonadipose tis-
sues: adipose triglyceride lipase (ATGL), hormone-sensitive
lipase (HSL), and monoglyceride lipase (MGL).
REGULATION OF LIPOLYSIS
Numerous lipolytic and antilipolytic effectors control
the catabolism of stored fat in various tissues (9, 10). These
include hormones, cytokines, and adipokines. In adipose
tissue, the most potent stimulatory signals are catechol-
amines acting on b-adrenergic receptors (11). Mouse adi-
pocytes express three subtypes of b-adrenergic receptors
(b-ARs): b1-AR, b2-AR, and b3-AR. In human adipose tissue,
only b1and b2receptors induce lipolysis. When catechol-
amines bind to these receptors, stimulatory Gsproteins ac-
tivate adenylate cyclase, causing a rise in cAMP levels and
elevated activity of cAMP-dependent protein kinase-A
(PKA) (10, 12, 13). PKA-mediated phosphorlylation of tar-
get proteins, including lipolytic enzymes and lipid droplet-
associated proteins, induces an increased release of FAs
and glycerol from adipose tissue up to 100-fold. Other hor-
mones that stimulate PKAvia Gsprotein-coupled receptors
include glucagon, parathyroid hormone, thyrotropin,
a-melanocyte-stimulating hormone, and adrenocorticotro-
pin. Several antilipolytic factors have been shown to act
through inhibitory Giprotein-coupled receptors (10).
These factors include catecholamines acting through
a2-adrenergic receptors (11), adenosine (A1-adenosine re-
ceptor) (14), prostaglandin (E2 receptor) (15), NPY (NPY-1
receptor) (16), and nicotinic acid (GPR109A receptor)
(17). The relative distribution of a- and b-adrenergic recep-
torstherefore determines thelipolytic activityinatissue-and
cell type-specific manner.
Insulin and insulin-like growth factor represent the most
potent inhibitory hormones in lipolysis (9, 18). Their effects
are primarily communicated through the insulin receptor
(IR), polyphosphorylation of insulin receptor substrates
1–4 (IRS1–4), activation of phosphatidylinositol-3 kinase
(PI3K), and the induction of the protein kinase B/AKT
(PKB/AKT). Complexity in this essentially linear pathway
is added by the divergence at so-called critical nodes that
interact with other signaling cascades (19). Critical nodes
in the IR pathway include the IR and IRS interacting with
cytokine and extracellular signal-regulated kinase (ERK)
signaling and PI3K activating both 3-phosphoinositide-
dependent protein kinases (PDK1 and 2) as well as atypical
protein kinases C (PKCl and z). At this point, a signaling
network is established that regulates innumerablebiological
processes (possibly more than 1,000). Lipolysis is affected in
multiple steps, including the phosphorylation of phospho-
diesterase 3B, causing the degradation of cAMP and loss
of PKA activation (18).
The mechanisms through which other effectors regulate
lipolysis are less well characterized. These include tumor
necrosis factor-a (TNFa), growth hormone, the Cide
domain-containing proteins (CideN) family of proteins
(CIDEA, -B, and -C), and the CopI-ARF vesicle transport
machinery described below.
HSL, THE “CLASSIC” ENZYME IN LIPOLYSIS
The first enzyme discovered to facilitate the hormone-
induced catabolism of fat was HSL. Although the initial
observations of fasting-induced lipolytic activity in WAT of
dogs (20) and man (21) were reported as early as 1932 and
1950, respectively, it was not until the early 1960s that a
WAT-associated lipase was shown to be regulated by hor-
mones and found to be different from lipoprotein lipase
(22–25). In a landmark study, Vaughan, Berger, and
Steinberg (26) discovered two independent lipolytic activi-
ties in WAT of various mammals and designated these en-
zymes HSL and MGL. The purification of HSL, cloning of
the corresponding cDNA and gene, and high-level heterol-
ogous expression of the protein permitted an extensive
study of the biochemical properties of the enzyme, its
tissue-specific function, and its regulation by various ago-
nists and antagonists. Several comprehensive reviews have
been published recently to summarize these results (13,
HSL exhibits broad substrate specificity capable of
hydolyzing TG, diacylglycerol (DG), monoacylglycerol
(MG), cholesteryl esters (CEs), retinyl esters (REs), and
other ester substrates such as p-nitrophenyl butyrate
(31). The relative maximal hydrolysis rates are in the
range of 1: 10: 1: 4: 2 for TG: DG: MG:CE: RE. Thus,
TGs are actually the worst substrate for HSL among all
these natural lipid esters, whereas DGs are the best. HSL
slightly favors unsaturated medium-chain FAs over satu-
rated long-chain FAs in TG substrates (32). However,
the substrate specificity toward the length or saturation
grade of acyl chains within lipid esters is not very pro-
nounced. Within the TG molecule, HSL preferentially
hydrolyzes primary ester bonds in the sn-1 and sn-3 posi-
tions (33). Phosphorylation of HSL in vitro modestly
increases enzyme activity for TG and CE hydrolysis by
4Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
1.5- to 2-fold (34, 35). The activity for DG or MG hydrolysis
is not affected.
HSL gene, mRNA, and protein structure
The gene for human HSL (LIPE) spans a genomic re-
gion of 26 kb and is located on chromosome 19q13.2
(36). In addition to 10 exons that are transcribed into
HSL mRNA in all human and mouse tissues, alternative
exon usage results in a significant variation in the 5′-region
of HSL transcripts (37–40). In adipose tissue, adrenal
gland, and ovary, HSL transcription starts from multiple
exons (exons A, B, C, D, or exon 1) within a 13 kb region.
Because exons B, C, and D are noncoding, the alternative
exon usage does not change the amino acid composition
of the enzyme. In contrast, exon A contains coding infor-
mation for 43 additional amino acids, leading to an alter-
native enzyme isoform. In testis, two tissue-specific exons
(T1 and T2) are used as transcriptional start sites. Exon
T1 codes for an additional 300 amino acids, whereas T2
contains no coding sequences. The high variability in exon
usage results in various HSL mRNA and protein sizes in
adipose tissue, pancreatic b-cells, ovaries, and testis. Multiple
potential transcription factor binding elements upstream of
each transcriptional start site suggest the possibility of differ-
ential transcriptional regulation of HSL in different tissues
and under various physiological conditions.
According to the HSL domain structure model [the
three-dimensional (3D) structure of the enzyme remains
to be elucidated], the enzyme can be subdivided into three
functionalregions (41–44).The N-terminaldomain (amino
acids 1–300) is believed to mediate enzyme dimerization
(45) and interaction with FABP4, a fatty acid binding pro-
tein known to enhance HSL enzyme activity (46–48). The
C-terminal domain contains the catalytic triad composed
of serine 423, aspartate 703, and histidine 733 (numbering
relates to rat HSL, isoform 2) within an a/b hydrolase fold
typically found in many lipases and esterases. The third
domain represents the regulatory module of the enzyme.
This loop region (amino acids 521–669) contains all known
phosphorylation sites of HSL.
HSL regulation of enzyme activity
Two major mechanisms determine HSL activity: enzyme
phosphorylation by protein kinases and interaction with
auxiliary proteins. The pathway of b-adrenergic stimula-
tion involves the PKA-mediated phosphorylation of HSL.
Originally it was believed that phosphorylation at two ser-
ine residues (563 and 565) (numbering relates to rat HSL,
isoform 2) was sufficient to mediate the cAMP-dependent
activation of HSL (49, 50). Serine 565 was considered the
basal phosphorylation site and serine 563 the regulatory
site (51–53). However, PKA-mediated enzyme activation
in an HSL variant in which Ser 563 was replaced by alanine
led to the discovery of additional PKA phosphorylation
sites (54). The identification of these additional serines
that are targets for phosphorylation by PKA (Ser 659 and
Ser 660) (54), ERK (Ser 600) (55), glycogen synthase
kinase-4 (Ser 563) (56), Ca21/calmodulin-dependent
kinase II (Ser 565) (57), and AMP-activated kinase (Ser 565)
(57) has markedly increased the complexity of posttransla-
tional HSL modification and regulation. Enzymes involved
in the dephosphorylation of HSL include protein phospha-
tases 1, 2A, and 2C (58).
HSL phosphorylation by PKAin responseto b-adrenergic
stimulation induces the intrinsic HSL enzyme activity only
moderately (approximately 2-fold). This is in sharp con-
trast to findings in intact cells where b-adrenergic stimu-
lation and activation of PKA cause up to a 100-fold
induction of FA and glycerol release. Thus, in addition to
HSL modification, other mechanisms must contribute to
hormone-induced lipolysis. This finding led to the discov-
ery of perilipin (3, 59–61). Perilipin is expressed mostly
in WAT and steroidogenic tissues, where it localizes to the
surface of lipid droplets (62). b-adrenergic stimulation
of adipocytes causes the PKA-mediated polyphosphory-
lation of six defined serine residues within the protein
(Ser 81, -222, -276, -433, -492, and -517), which results in
the translocation of HSL to the lipid droplet and initiation
of hydrolysis (63–66). Although originally HSL binding to
the lipid droplet was seen in association with perilipin dis-
sociation, the characterization of perilipin-deficient mice
and functional studies with perilipin mutants redefined
and extended this “replacement” hypothesis (67–69). Dur-
ing hormone stimulation, perilipin is essential for the
recruitment of HSL to lipid droplets and full enzyme acti-
vation (65). Notably, perilipin phosphorylation is not
required for the translocation of HSL, because unphos-
phorylated perilipin mutants still recruit HSL to the surface
of lipid droplets (64). In contrast, perilipin phosphoryla-
tion is absolutely crucial for the hydrolytic activity of HSL.
Perilipin is mostly associated with small lipid droplets
within fat cells and, in fact, contributes to the fragmenta-
tion of large lipid droplets during the lipolytic process.
The latter activity of perilipin involves its phosphorylation
at serine residue 492 (70, 71).
Perilipin belongs to the PAT family of proteins (as
reviewed in Refs. 3, 72). These factors include perilipin,
adipophilin (ADRP), Tip47, S3-12, and myocyte lipid
droplet protein (MLDP, also termed OXPAT). Because
perilipin expression is mostly restricted to adipose and
steroidogenic cells, it is of great interest to determine
whether other PAT family members can accomplish a simi-
lar regulatory role for HSL in nonadipose tissues. In one
report, HSL interaction with lipotransin was shown to acti-
vate HSL-mediated lipolysis (73). However, this mecha-
nism has not been confirmed since its original observation.
HSL deficiency in mice: HSL is not alone
For more than three decades, HSL was considered to be
the only and therefore rate-limiting enzyme for the lipo-
lytic catabolism of stored fat in adipose and nonadipose
tissues. Because HSL was shown to hydrolyze both TG
and DG substrates, it was believed that the enzyme repre-
sented the only lipase activated by hormonal stimulation.
This view, however, changed when several independent
laboratories reported on the phenotype of HSL-deficient
mice (74–76). Although HSL deficiency causes infertility
Role of ATGL in lipolysis5
by guest, on August 22, 2013
in male mice, owing to a defect in sperm maturation, the
animals are normal with regard to their lipid and energy
metabolism. Unexpectedly, HSL knockout (HSL-ko) mice
were not overweight or obese. To the contrary, with in-
creased age, they had reduced WAT weight (77) and were
resistant to genetically or diet-induced obesity (78). HSL-
deficient adipocytes responded to b-adrenergic stim-
ulation and, compared with control mice, exhibited only
a moderate decrease in their capacity to release FA
(?40%) (32, 79). Importantly, HSL deficiency resulted
in DG accumulation in several tissues, indicating that
HSL is rate-limiting for DG hydrolysis (32). These findings
strongly suggested that at least one additional enzyme
acted as TG hydrolase when HSL was absent and that this
activity was either directly or indirectly “hormone sensi-
tive.” The findings also indicated that HSL was more im-
portant as DG hydrolase than as TG hydrolase.
ATGL: A NEW PLAYER IN THE LIPOLYSIS TEAM
In 2004, three groups independently published the dis-
covery of an enzyme able to hydrolyze TG and named it
ATGL (77), desnutrin (80), or calcium-independent phos-
pholipase A2z (iPLA2z) (81). Very soon after these initial
reports, ATGL orthologous genes and proteins were iden-
tified and characterized in other vertebrates, flies, fungi,
and plants (82–86). Work with the Drosophila melanogaster
enzyme “brummer,” triacylglycerol lipase-4 in Saccharomyces
cerevisiae, and sugar-dependent1 in Arabidopsis thaliana
demonstrated that each of these proteins exhibits robust
TG hydrolase activity and has a fundamental role in the
regulation of TG homeostasis in the respective organism.
ATGL exhibits 10-fold higher substrate specificity for TG
than for DG and selectively performs the first step in TG
hydrolysis, resulting in the formation of DG and FA (77).
The stereospecificity of ATGL for the chemically distinct
ester bonds within the TG molecule is currently not
known. Therefore, it is also unclear whether the DG gen-
erated by ATGL can participate in signaling processes in-
volving sn-1,2-DG, such as the activation of various PKC
isoenzymes. ATGL was also reported to have transacylase
(81, 87) and phospholipase activity (77, 81, 88) that was
shown to be lower than its TG hydrolase activity (77, 81).
In contrast to HSL, ATGL does not hydrolyze MG, CE, or
RE. Smirnova et al. (89) demonstrated that the hydrolytic
function of ATGL is not restricted to the catabolism of
lipid droplets (“adiposomes”) in adipose tissue and sug-
gested the enzyme be renamed adiposome triglyceride lipase
rather than adipose triglyceride lipase to more adequately re-
flect its function.
ATGL gene, mRNA, and protein structure
The mouse Atgl gene (Pnpla2) contains nine exons and
spans a region of approximately 6 kb on chromosome 7F5.
Transcription of the gene results in a 1.96 kb mRNA cod-
ing for a 486 amino acid protein with a molecular mass
of 54 kDa. The ten exons of the human ATGL gene
(PNPLA2) span 6.32 kb of genomic DNA, which are lo-
cated on chromosome 11p15.5. Mammalian ATGL belongs
to a gene family characterized by the presence of a patatin
domain (Pfam01734). This structural motif was designated
for patatin, the most abundant protein in the potato tuber,
with established DG, MG, and phospholipase activity, but
no TG hydrolase activity (90–92). ATGL is most closely re-
lated to a group of five genes and proteins named patatin-
like phospholipase domain-containing 1 to 5 (PNPLA1-5) (93,
94). Members of this protein family in addition to ATGL
(PNPLA2) are PNPLA1, adiponutrin (PNPLA3), GS2
(PNPLA4),andGS2-like (PNPLA5). To date,noorthologous
gene for GS2 has been identified in the mouse genome.
More distantly related members of ATGL include neu-
ropathy target esterase (NTE, PNPLA6), NTE-related ester-
ase (NRE, PNPLA7), calcium-independent phospholipase
A2g (iPLA2g, PNPLA8), and phospholipase A2 group VI
(PLA2G6, PNPLA9). Like ATGL, adiponutrin, GS2, and
GS2-like also exhibit hydrolase and transacylase activity in
in vitro assays (81, 87). Low specific phospholipase activity
was reported for ATGL, adiponutrin, and GS2-like (81, 88).
Considering the structural and functional diversity within
patatin domain-containing proteins, the name patatin-like
phospholipase domain-containing 1-9 for these proteins is
somewhat misleading and should be changed to a more
general name such as patatin domain-containing lipid hydro-
lase 1-9 (PDLH1-9).
The primary structures of the human and murine ATGL
enzyme share 84% sequence identity. Sequence identity is
particularly high within the patatin domain (.95%) har-
boring the active site of the enzyme. A schematic represen-
tation of the domain structure of ATGL is shown in Fig. 1.
Interestingly, unlike other typical TG hydrolases, the active
site of patatin domain-containing enzymes is not com-
posed of a catalytic triad. Instead, 3D structure determina-
tion of related members of the family (potato patatin)
revealed that the enzyme mechanism depends on a cata-
lytic dyad (95). In ATGL, mutational analyses identified
serine 47 as the active site nucleophile located within a
canonical GXSXG sequence (87, 89, 96). From homol-
ogy considerations, it is assumed that aspartate 166 is the
second amino acid critical for the catalytic dyad in ATGL.
Sequence and 3D-structural similarities also indicate that
the dyad is embedded within a three-layer a/b/a archi-
tecture commonly found in hydrolases/esterases (97).
Similarly, as has been shown for other patatin domain en-
zymes, the transition state in ATGL might be stabilized by
a glycine-rich oxyanion hole. The C-terminal region of
ATGL exhibits only poor homology to the other members
of the PNPLA family. A hydrophobic stretch from amino
acids 315 to 360 was proposed to mediate lipid droplet
binding (77). Additionally, two phosphorylation sites were
identified in the C-terminal region of the enzyme (serine 404
and serine 428 in human ATGL) (77, 98). The functional
roles of enzyme phosphorylation and involved protein
kinases remain unknown. Notably, the human protein is
19 amino acids longer than the mouse ortholog and contains
a proline-rich sequence on its very C terminus. Whether
6Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
this peptide stretch contributes to species-specific differences
in ATGL regulation and function requires clarification.
ATGL physiological function
The important role of ATGL in lipolysis became evident
from observations in ATGL-deficient (ATGL-ko) mice
(99). In contrast to HSL-deficient mice, ATGL-ko animals
had a severe “lipid” phenotype (for phenotype compari-
son of HSL-ko and ATGL-ko mice, see Table 1). Absence
of ATGL causes a reduction of FA release from WAT by
more than 75%. ATGL-ko mice accumulated TG in essen-
tially all organs and cell types analyzed, consistent with an
important function of ATGL in TG catabolism in multiple
tissues. Defective TG mobilization and massive TG accu-
mulation cause severe myopathy in cardiac muscle, defec-
tive thermogenesis in brown adipose tissue (BAT), and an
overall defect in energy homeostasis (see section on the
tissue-specific function of ATGL below). The excessive
TG accumulation in the heart causes cardiac dysfunction
and premature death in ATGL-deficient mice. In WAT,
TG hydrolase activity and hormone-stimulated lipolysis
were also drastically reduced, which is consistent with the
view that ATGL is “hormone sensitive” via either a direct or
an indirect mechanism.
In contrast to HSL-ko animals, ATGL-deficient male
mice are fertile, indicating that massive TG accumulation
in testis per se is not causative for male infertility. Instead,
it appears more likely that the impaired hydrolysis of DG,
CE, and RE may cause a defect in germ cell maturation in
HSL-deficient mice. The concentration of plasma FA, TG,
and ketone bodies are decreased in both fasted HSL-ko
and ATGL-ko mice, yet the absolute levels are lower in
ATGL deficiency and are also decreased in fed ATGL-ko
animals as compared with wild-type littermates. Interest-
ingly, total cholesterol and HDL cholesterol concentra-
tions are elevated in HSL-deficient mice and reduced in
ATGL-deficient mice. The reason for this unexpected dif-
ference is unclear and requires elucidation.
Taken together, the analysis of ATGL-ko mice suggested
that ATGL is rate-limiting for the first step in TG hydro-
lysis, generating DG and FA with an approximately 10-fold
higher specificity toward TG than HSL (calculated). HSL
efficiently degrades DG, generating MG and FA. The final
step, resulting in the formation of glycerol and FA, is per-
formed by MGL. Whether other TG hydrolases in addition
to ATGL and HSL also contribute to the hydrolysis of TG
in WAT was recently addressed by Schweiger et al. (100).
Complete inhibition of HSL with a specific inhibitor (pro-
vided by Novo Nordisk) resulted in an almost complete
absence of FA release in ATGL-deficient adipose tissue,
suggesting that besides ATGL and HSL, additional lipases
contribute little to the lipolytic capacity of white fat cells in
mice. The role of alternative lipases, such as Ces3 (101) or
TGH-2 (102), in WAT under specific physiological condi-
tions or their contribution to lipolysis in nonadipose tis-
sues remains to be determined.
Regulation of ATGL: hormones and cytokines
Although ATGL is expressed in most tissues of the body,
the highest levels of mRNA and enzyme activity are found
in WAT and BAT (77, 80, 81, 87, 103–108). During adipo-
cyte differentiation of 3T3-L1 cells, ATGL expression is
strongly induced, reaching maximal levels when the cells
accumulate visible lipid droplets (77, 80, 81, 88, 93, 103,
109). Compared with WAT and BAT, ATGL mRNA levels
are much lower in other tissues. Quantitative PCR analysis
revealed that adrenals, testis, cardiac muscle, and skeletal
muscle have approximately 25% of the ATGL mRNA levels
(normalized to tissue protein content) found in WAT, other
tissues around 10% (87, 109).
In contrast to a wealth of available information on the
regulation of HSL in WAT, comparatively little is known
about the molecular pathways leading to the activation of
ATGL activity. The ability of HSL-ko WAT to respond to
hormonal stimulation (32) and the finding that HSL inhi-
bition in WAT leaves a “hormone-inducible” hydrolytic ac-
tivity (100) suggest that ATGL activity is either directly or
indirectly activated by hormonal signals. Several observations
indicate that the molecular mechanism leading to ATGL
activation is different from that described for HSL. First,
TABLE 1. Phenotypes observed in ATGL-deficient and
Fat accumulation in nonadipose tissues
Tissue DG accumulation
Plasma ketone bodies
Plasma HDL cholesterol
ATGL, adipose triglyceride lipase; HSL, hormone-sensitive lipase;
ko, knockout; DG, diacylglycerol; TG, triacylglycerol. Data assembled
from (32, 74–76, 99).
Fig. 1. Representation of the structural domains of hu-
man adipose triglyceride lipase (ATGL) protein. The cru-
cial structural components for enzyme function are
indicated, including the patatin domain, a/b hydrolase re-
gion, active site serine (S47), putative aspartic acid within
the catalytic dyad (D166), potential lipid binding domain
(hydrophobic), and two established phosphorylation sites
tral lipid storage disease with myopathy are also indicated.
Role of ATGL in lipolysis7
by guest, on August 22, 2013
unlike HSL, ATGL is present on lipid droplets of adipocytes
in similar amounts in the basal state and in the activated
state (77). Second, although ATGL can be phosphorylated,
it is not a target for PKA (77, 98). Third, ATGL activity is
greatly enhanced by a protein annotated as a/b hydrolase
domain-containing protein 5 (ABHD5) or comparative
gene identification-58 (CGI-58) (96). CGI-58 does not af-
fect HSL enzyme activity (96).
To date, most studies addressing the regulation of ATGL
by hormonal or nutritional effectors have restricted their
analyses to the measurement of ATGL mRNA levels and
have not reported ATGL enzyme activities. Considering
the likely posttranscriptional regulation of ATGL and
HSL by phosphorylation and numerous modulating pro-
tein factors (see below), this is unfortunate, and conclu-
sions drawn from these data must be viewed with caution
when lipolytic activities are assumed from lipase mRNA
concentrations. As expected for a TG hydrolase active in
WAT, ATGL mRNA concentrations are markedly affected
by nutritional status, and increase during fasting and de-
crease during refeeding (80, 87, 103, 109). ATGL mRNA
levels during fasting are not paralleled by HSL mRNA levels
that are downregulated during acute fasting and increase
only after prolonged food deprivation (3–5 days) (110).
From in vitro experiments in murine 3T3-L1 adipocytes,
Villena et al. (80) concluded that glucocorticoids could
be responsible for the increase of ATGL mRNA levels in
the fasted state. The observation that ATGL mRNA is signif-
icantly downregulated in genetic models of obesity (ob/ob
and db/db mice) suggested a possible contribution of ATGL
in the pathogenesis of obesity (80, 103); however, this effect
was not observed in all studies (87).
The enormous induction of FA and glycerol release
from fat cells in response to b-adrenergic stimulation is
not associated with increased levels of either ATGL or
HSL mRNA. In fact, in some studies, isoproterenol treat-
ment of 3T3-L1 cells or isolated adipocytes causes de-
creased mRNA concentrations of both lipases (111–113).
This suggests that the b-adrenergic stimulation of lipolysis
is exclusively regulated posttranscriptionally.
Insulin treatment reduces ATGL mRNA levels in murine
3T3-L1 adipocytes (103, 109, 111). Importantly, this inhibi-
tory effect of insulin on ATGL expression was also dem-
onstrated in vivo using mouse models of systemic insulin
deficiency (streptozotocin-treated animals) and of adipose-
specific insulin receptor deficiency (109). Both mouse mod-
els exhibited increased lipolysis and increased ATGL mRNA
levels, indicating that the induction of ATGL gene expres-
sion might contribute to elevated FA mobilization under
conditions of defective insulin signaling.
Cytokines, and specifically TNFa, have multiple effects
on adipose tissue, and TNFa has been implicated in the
pathogenesis of obesity and insulin resistance (114, 115).
TNFa strongly stimulates lipolysis; as a consequence, mas-
sive lipid catabolism might contribute to the wasting seen
in cancer cachexia. The finding that TNFa is able to stimu-
late lipolysis in HSL-ko adipocytes (79) suggested that the
process was HSL-independent and that ATGL could be the
actual TNFa target lipase. However, although lipolysis is
increased, two studies have reported that both ATGL and
HSL mRNA levels decrease in 3T3-L1 adipocytes in re-
sponse to TNFa treatment (103, 111). This again suggests
a dissociation of enzyme mRNA levels and enzyme activity.
A possible explanation for low ATGL mRNA levels upon
TNFa treatment relates to the fact that TNFa suppresses
the expression of a large number of adipose-specific genes,
leading to an “adipocyte dedifferentiation” process (116).
One of these genes, peroxisome proliferator-activated
receptor-g (PPARg), is a key nuclear receptor controlling
adipocyte differentiation and metabolism (117). Kim et al.
(103) demonstrated that ATGL is a direct transcriptional tar-
getgeneforPPARg,andPPARg agonistssuch asrosiglitazone
increase ATGL mRNA levels and induce lipolysis in various
adipose models (103, 118–121). Therefore, it is conceiv-
able that the TNFa-mediated inhibition of PPARg reduces
ATGL mRNA expression. How TNFa affects ATGL enzyme
activity is currently unknown. In macrophages, ATGL has
also been shown to be a target of PPARy (122).
Contradicting views currently exist regarding the rela-
tive importance of ATGL in relation to HSL in human
WAT. Langin (9) concluded from their studies in pri-
mary human adipocytes that HSL is the major lipase for
catecholamine- and natriuretic peptide-stimulated lipoly-
sis, whereas ATGL mediates TG hydrolysis mainly during
basal lipolysis. Another study suggested that human HSL
has a higher capacity to hydrolyze TG compared with
ATGL (106). This study and a report by Ryden et al. (107)
also found that, in contrast to HSL, ATGL mRNA and pro-
tein levels in adipose tissue are unaffected by obesity and
weight reduction, arguing for a regulation of HSL, but
not ATGL gene expression in response to obesity status.
In contrast, other reports assign a crucial role to ATGL for
TG hydrolysis in human WAT and show decreased ATGL
mRNA and protein levels in obese individuals with insulin
resistance (104, 108, 123). The availability of specimens
from patients with ATGL deficiency might help to elucidate
the functional role of ATGL in human WAT.
Regulation of ATGL: lipid droplet proteins
CGI-58. Mammalian TG hydrolases that act on water/
lipid interphases frequently require cofactors for full
enzyme function. For example, pancreatic lipase forms a
complex with a colipase, and lipoprotein lipase (LPL) acts
in concert with apolipoprotein C-II (apoC-II). ApoC-II is
present on the surface of the major substrates for LPL,
the TG-rich lipoproteins VLDL and chylomicrons. By anal-
ogy, it was not totally surprising when a lipid droplet pro-
tein, CGI-58 or ABHD5, was found to activate ATGL (96).
In the presence of CGI-58, the TG hydrolase activity of
mouse ATGL is induced approximately 20-fold. Human
ATGL is also activated by CGI-58, although to a lesser de-
gree (approximately 5-fold ATGL induction). Importantly,
these findings provided a biochemical explanation for a
human disorder. In 2001, Lefevre et al. (125) discovered
that mutations in the gene for CGI-58 are causative for a
lipid storage disorder designated “neutral lipid storage
disease” or Chanarin Dorfman Syndrome (see below for
discussion of human mutations). CGI-58 was originally
8Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
identified as a homologous gene in an alignment of the hu-
man and the Caenorhabditis elegans genomes. Mouse CGI-58
is ubiquitously expressed, with the highest expression levels
found in testis and adipose tissue (96, 126).
CGI-58 is a 349 amino acid-long protein with a molecular
mass of 40 kDa. As shown in Fig. 2, the protein belongs to
the esterase/thioesterase/lipase subfamily of proteins
structurally characterized by the presence of a/b hydrolase
folds. In contrast to most other members of this family, the
putative nucleophilic serine within the canonical esterase/
lipase motif GXSXG is replaced by an asparagine in CGI-58
(125), effectively eliminating the possibility that CGI-58
functions as a lipase. The protein was shown to bind to lipid
droplets by interaction with perilipin A in a hormone-
dependent way (126–128). In nonstimulated adipocytes,
CGI-58 is tightly associated with the lipid droplet, whereas
upon b-adrenergic stimulation and concomitant phosphor-
ylation of perilipin, CGI-58 dissociates and becomes cyto-
solic (126, 128). Reducing the cAMP levels of the cell
reverses this dissociation process (126). Fluorescence reso-
nance energy transfer and bimolecular fluorescence com-
plementation experiments showed that CGI-58, once
dissociated from perilipin, colocalizes in close proximity
to ATGL (129), suggesting the involvement of CGI-58/
ATGL interaction in stimulated lipolysis. CGI-58 is not in-
volved in the vesicularization of lipid droplets during lipol-
ysis (130). In summary, these findings support the following
scenario: In the basal state, when adipocytes are not hor-
monally stimulated, CGI-58 binds to perilipin A and is un-
able to activate ATGL. Following hormonal stimulation,
perilipin is phosphorylated at several serine residues, in-
cluding serine 517, whereupon CGI-58 dissociates from
perilipin, interacts with ATGL, and activates TG hydrolysis.
Whether phosphorylation of serine 517 in perilipin or
phosphorylation of ATGL affects the respective CGI-58
binding directly is currently not known. Concomitant with
ATGL activation, HSL translocates from the cytosol to the
lipid droplet and efficiently hydrolyzes DG, the lipolytic
product of ATGL.
Activation of ATGL might not be the only physiological
activity of CGI-58. Importantly, a very recent publication
showed that in addition to its function as ATGL activator,
CGI-58 can also act as acylglycerolphosphate acyltrans-
ferase (AGPAT) (131). The role of this reaction in vivo
remains to be determined (see below for discussion of hu-
PAT proteins. The crucial role of perilipin in the ATGL/
CGI-58-mediated hydrolysis of TG became evident in an
elegant study by Miyoshi et al. (63) showing that hormone-
stimulated lipolysis depended on perilipin and ATGL. The
authors demonstrated that perilipin phosphorylation of
residue serine-517 is essential for ATGL-mediated lipolysis
and represents a prerequisite for the function of subse-
quent lipase activity of HSL.
Perilipin expression is confined to adipose tissue and
steroidogenic tissues. Lipolysis of lipid droplet-associated
TG is, however, required in many other tissues, including
those that do not express perilipin, such as skeletal and
cardiac muscle, or the liver. Accordingly, alternative mech-
anisms must exist to control TG hydrolysis by ATGL and
HSL (and possibly other lipases) in nonadipose tissues.
These mechanisms are not well understood. Recently,
two studies addressed the questions of whether and
how nonperilipin PAT proteins affect lipolysis and ATGL.
Listenberger et al. (132) demonstrated that ADRP controls
ATGL access and TG lipolysis in HEK293 cells and other
human cell lines. Bell et al. (133) studied the role of var-
ious PAT proteins in TG catabolism in hepatocyte-like
AML12 cells and found that reduced expression of ADRP
and TIP47 caused increased ATGL localization to lipid
droplets and increased lipolytic rates. These findings are
consistent with a crucial regulatory role for lipid droplet
scaffold protein regulating the substrate access of func-
Pigment epithelium-derived factor. In addition to PAT pro-
teins, other proteins found on lipid droplets are also in-
volved in the regulation of lipolysis. Surprisingly, searching
for receptors and binding proteins for pigment epithelium-
derived factor (PEDF), Notari et al. (88) identified ATGL as
a PEDF binding protein and proposed to name the enzyme
PEDF-receptor. Apparently, ATGL is highly expressed in the
pigment epithelium and can be found on the plasma mem-
brane, where it binds to PEDF and exhibits phospholipase
activity. PEDF binding might also be important in cells and
organs where ATGL is localized only within cells. For exam-
ple, hepatocytes that lack PEDF were shown to accrue neu-
tral lipid droplets, and lipid accumulation was reversed by
the reexpression of PEDF (134). These results suggest that
PEDF binds to ATGL on lipid droplets, inducing TG hydro-
lysis. The mechanism of this activation and the question of
whether PEDF mediates the activation of ATGL also in
other tissues remain to be determined.
CideN. Another group of lipid droplet binding proteins
that regulate lipolysis belongs to the CideN family. CideN
Fig. 2. Representation of the structural domains of human com-
parative gene identification-58 (CGI-58) a/b hydrolase domain-
containing protein 5 (ABHD5) (CGI-58/ABHD5). The a/b hydrolase
region and the asparagine residue replacing a serine within the con-
sensus GXSXG of lipases and esterases are indicated. Additionally,
mutations in CGI-58 associated with neutral lipid storage disease
with ichthyosis are shown. Two mutations within splice acceptor con-
sensus sequences cause protein truncations after exon 2 and exon 5.
The final amino acids of the wild-type sequences are indicated
(K43SM and P256SM).
Role of ATGL in lipolysis9
by guest, on August 22, 2013
proteins were originally discovered because of their struc-
tural similarity to DNA fragmentation factors and were be-
lieved to regulate cell death activation (135). Recently,
members of the CideN family were shown to affect lipid
droplet morphology and turnover. CideA and CideC/
Fsp27 bind to lipid droplets and colocalize with perilipin
(136, 137). Overexpression of these factors inhibits fat
catabolism and induces cellular lipid accumulation (136).
Consistent with these findings, mice that lack CideC/FSP27
have smaller, multilocular lipid droplets, decreased fat
mass, lower levels of plasma FAs, and increased insulin sen-
sitivity (138, 139). Similarly, CideA and CideB deficiency
in mice is associated with a lean phenotype (140, 141).
CideC/Fsp27 is also important for the regulation of TG
catabolism in hepatocytes, because increased protein ex-
pression in ob/ob mice or in animals infected with CideC-
expressing adenovirus causes hepatic lipid accumulation
and steatosis (142). The mechanism by which members
of the CideN family regulate the activity of lipases is cur-
Arf1-CopI. In a genome-wide RNA interference screen
in Drosophila S2 cells, Guo et al. (143) identified a large
number of genes that affect lipid droplet biogenesis and
morphology. Interestingly, the study identified a subset of
the Arf1-CopI family of vesicular transport proteins that
resulted in smaller, more disperse lipid droplets and in-
creased lipolysis, suggesting yet another currently unknown
mechanism that regulates the activity of lipolytic enzymes.
Taken together, these results suggest that lipases are
embedded in a complex “lipolysome” consisting of the ac-
tual lipolytic enzymes and numerous modulators of en-
MUTATIONS IN ATGL OR CGI-58 CAUSE NEUTRAL
LIPID STORAGE DISEASE IN HUMANS
Recently, mutations in the genes for ATGL and CGI-58
were identified and provided the molecular basis underly-
ing neutral lipid storage disease (NLSD) in humans. NLSD
is a rare, autosomal genetic disorder characterized by sys-
temic accumulation of TG in all tissues of the body. It is
diagnosed by increased TG storage in blood granulocytes
(referred to as Jordansʼ anomaly) (144). Excessive lipid
storage leads to variable forms of skeletal and cardiac
myopathy and hepatic steatosis. Additionally, some patients
suffer from ataxia, hearing loss, or mental retardation (145,
146). According to a recently proposed classification (147),
NLSD can be subdivided into two distinct groups. Depend-
ing on whether or not the patients suffer from a skin defect
(severe ichthyosis), they are diagnosed with either neutral
lipid storage disease with ichthyosis (NLSDI, also known as
Chanarin Dorfman Syndrome) or neutral lipid storage dis-
ease with myopathy (NLSDM), respectively. Importantly, this
classification finds its molecular basis in the affected genes.
Mutations in the gene for ATGL (PNPLA2) cause NLSDM,
and mutations in the gene for CGI-58 cause NLSDI.
In 2007, Fischer et al. (147) reported that mutations in
the gene for ATGL cause NLSDM. Since then, several new
mutations in the ATGL gene locus (PNPLA2) were discov-
ered (148, 149). Currently, six mutations are known to
cause aberrant ATGL proteins (indicated in Fig. 1). These
include a point mutation (Pro195Leu), four frameshift
mutations (at amino acids 160, 267, 270, or 283), and one
nonsense mutation (Asn2893). Both frameshift and non-
sense mutations result in the deletion of the C-terminal
region of ATGL. Interestingly, the patatin domain with
the active site serine 47/aspartate 166 dyad is present in
most of the truncated ATGL variants. These ATGL mutants
are enzymatically highly active and can be stimulated by
CGI-58 when artificial lipid emulsions are used as sub-
strates (147, 149, 150). However, ATGL lacking parts of
the C-terminal half of the enzyme exhibited reduced bind-
ing to cellular lipid droplets, and it is assumed that defec-
tive substrate binding in a cellular context is responsible
for the lipolytic defect. In one patient with severe myop-
athy, the mutation occurred within the patatin domain
(frameshift mutation at amino acid 160); this mutation re-
sults in a truncated protein that lacks the active site aspar-
tate 166 (148). Whether this mutation results in complete
enzyme inactivation is not known. True null mutants that
totally lack ATGL have not been found so far. More ex-
tensive biochemical characterization of ATGL deletion
mutants retaining the patatin and a/b hydrolase domain
but lacking the C-terminal region revealed a double func-
tion of the C terminus (150). First, it appears to mediate
lipid droplet binding. Second, it has an inhibitory role in
the hydrolytic reaction, because absence of the C-terminal
region generates an enzyme with higher specific activity
against artificial substrates. Whether the established phos-
phorylation sites present in this region (98) affect lipid
binding or enzyme activity is presently unknown.
The clinical observations in patients with NLSDM re-
semble in many aspects the phenotypic characteristics of
ATGL-ko mice. Both genetic deficiencies result in systemic
lipid accumulation, Jordansʼ anomaly, and myopathy. Skin
defects were not observed in either species. Two patients
with NLSDM were reported to have died from cardiac fail-
ure (147, 149), but unfortunately it is not known whether
their cardiac dysfunction resulted from excessive lipid ac-
cumulation as observed in mice. The identification and
characterization of additional patients with PNPLA2 gene
mutations will help to elucidate the role of ATGL in hu-
man cardiac physiology.
Six years before mutations in ATGL were found to cause
NLSDM, Lefèvre et al. (125) described eight mutations in
the human gene for CGI-58 in families with a confirmed
diagnosis of NLSDI. Subsequently, other groups reported
additional mutations in the gene for CGI-58 causative for
NLSDI (96, 151, 152). The locations of the known muta-
tions in the CGI-58 sequence that cause NLSDI are in-
cluded in Fig. 2. Patients share many clinical features
with those affected with ATGL deficiency, but some strik-
10Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
ing differences are apparent. All patients with CGI-58 defi-
ciency suffer from severe ichthyosis, and some of them
have developmental defects, including deformation of
the ear and mental retardation. These differences suggest
that CGI-58 has additional functions that are independent
of ATGL. Mutated CGI-58 with single amino acid substitu-
tions (Q130P, E260K) or deletions totally fail to activate
ATGL (96), suggesting that the defective ATGL stimulation
by CGI-58 is causative for the multi-tissue TG accumulation
observed in NLSDI. Interestingly, the same mutations in
CGI-58 are unable to bind to perilipin and are not associ-
ated with lipid droplets (128). Alternative explanations for
the molecular defect in NLSDI have also been proposed.
Even before CGI-58 was discovered, it was shown that the
molecular defect present in NLSDI prevents lipid remodel-
ing from neutral lipids to glycerophospholipids (153, 154).
Considering the recent finding that CGI-58 exhibits AGPAT
activity (131) and might affect phospholipid synthesis, it
seems conceivable that both defective TG hydrolysis and
phospholipid synthesis contribute to the pathogenesis of
NLSDI. However, several issues need clarification. First, the
finding that mutations in CGI-58 causing NLSDI had no
effect on its activity as acyltransferase raises the question
of whether this activity is lacking in patients with NLSDI.
Second, how does a defect in AGPAT activity of CGI-58
cause massive TG accumulation in light of the fact that
the product of the AGPAT reaction, phosphatidic acid, is
a common precursor for both TG and glycerophospho-
lipids? Normally, AGPAT deficiency results in lipodystrophy
and not excessive lipid accumulation (155).
Taken together, the clinical phenotype of patients af-
fected with both forms of NLSD and the comparison to
the phenotype of ATGL-ko mice suggest that excessive lipid
accumulation results from decreased lipolysis due to defects
in the enzyme (ATGL) or its activator (CGI-58). In the skin,
however, CGI-58 has an additional, ATGL-independent
function that is defective in NLSDI and responsible for
the development of ichthyosis. Whether this function of
CGI-58 involves the activation of another lipase, alterations
in the metabolism of phospholipids, or a completely unre-
lated activity remains to be determined.
THE TISSUE-SPECIFIC ROLE OF ATGL/CGI-58
With the availability of genetically modified mice that
lack ATGL (CGI-ko mice have not been reported to date)
and the characterization of patients with NLSDM and
NLSDI, a picture emerges of how ATGL affects lipid
metabolism and energy homeostasis. As a general conclu-
sion, it is evident that the physiological function of ATGL/
CGI-58 is not restricted to adipose tissue but is also crucially
important in many nonadipose tissues.
ATGL deficiency in mice is associated with a major de-
fect in WAT lipolysis (99). Both TG hydrolase activity in
WAT lysates and the release of FAs and glycerol from intact
tissue samples are markedly increased. As a consequence,
8-week-old ATGL-deficient mice are obese, exhibiting dou-
ble the fat mass of normal mice when kept on a normal
chow diet. In contrast to mice, human patients with ATGL
or CGI-58 deficiency are not overweight or obese. This has
been used as argument that ATGL-mediated lipolysis in
human WATis less important than in mouse WAT. However,
other explanations are also conceivable. First, defective li-
polysis in WAT can result in a concomitant downregulation
of lipogenesis. For example, the loss of WAT mass in HSL-
deficient mice is caused by a drastic reduction of lipo-
genesis due to decreased PPARg activity (156). Second,
patients might change their eating habits as a consequence
of their disease. Third, the absence of ATGL in human WAT
might induce alternative lipase activities. The analysis of
tissue samples from patients suffering from NLSDM or
NLSDI will hopefully help to assess the role of ATGL in
BAT serves as a TG storage organ with the unique ability
to generate heat by “non-shivering” thermogenesis. In
humans, brown adipocytes are abundant in neonates and
diminish with age. In rodents and hibernating animals,
BAT persists throughout life and is a major site for heat
production in response to low environmental tempera-
ture. Similarly to the situation in WAT, b-adrenergic stimu-
lation in BAT promotes the hydrolysis of stored TG by
endogenous lipases, leading to the mobilization of FAs as
fuel for thermogenesis. Defective norepinephrine and epi-
nephrine synthesis (157) or deficiency of all three known
b-adrenergic receptors (158) results in reduced lipolysis,
increased BAT mass, and severe cold sensitivity due to de-
fective thermogenesis. Remarkably, mice lacking HSL
exhibited normal thermogenesis (75) and were not cold
sensitive despite a lipolytic defect that resulted in brown
adipocyte hypertrophy due to TG and DG accumulation.
Apparently, in the absence of HSL, sufficient amounts of
FAs are mobilized for mitochondrial heat production. In
contrast, ATGL-ko mice are extremely cold sensitive and
die after cold exposure of more than 6 h, indicating that
the enzyme is essential for the provision of FA as substrate
for thermogenesis (99).
In the heart, continuous energy production is indispens-
able for the supply of ATP required for the permanent
contractile function of the beating heart. It is estimated
that 50–70% of the energy for myocardial contraction de-
rives from the oxidation of FAs (159). Because cardiomyo-
cytes do not synthesize FAs, they depend on their supply
from two exogenous sources: first, WAT-derived, circulating
unesterified FAs that are bound to plasma albumin; and
second, TG-associated FAs released by LPL from TG-rich
plasma lipoproteins. LPL is the only enzyme known to
be responsible for the hydrolysis of plasma TG-rich lipo-
proteins in peripheral cells (160, 161). Experiments in
transgenic and ko mouse models showed that the expres-
sion level of LPL in the heart largely determines the up-
take rate of FA. Increased LPL activity in the heart result
Role of ATGL in lipolysis 11
by guest, on August 22, 2013
in elevated FA utilization, increased lipid storage, de-
creased glucose utilization, and modest signs of cardiomyop-
athy (162–167). These studies and previous investigations
showed that FAs, once absorbed by the heart, are not uti-
lized directly for b-oxidation but, at least in part, are con-
verted into TG (168). Subsequently, intracellular lipase(s)
hydrolyze(s) these TG depots and provide(s) FA for oxida-
tion. Accordingly, endogenous myocardial TGs provide
a substantial amount of FAs for oxidation in perfused
rat hearts, especially under conditions of overt diabetes
(169). When hydrolyzed FAs are not utilized for oxidation,
they are again esterified to TG. This creates a futile cycle of
lipolysis and reesterification that reacts promptly to meet
increased substrate demand when energy is needed.
Cardiomyocytes express HSL, and the hormone respon-
siveness of myocardial lipolysis suggested that HSL might
be sufficient for TG hydrolysis in the heart (170, 171).
However, HSL-deficient mice do not accumulate TG in
the heart (32). In fact, overnight-fasted HSL-ko mice show
markedly reduced myocardial TG levels and increased LPL
activity (74), suggesting that both uptake of lipoprotein-
associated FAs and intramyocardial TG mobilization are
still functional when HSL is absent. ATGL-ko mice, in con-
trast, exhibit a prominent heart phenotype (99). As early
as 6 weeks after birth, ATGL-deficient hearts accumulate
lipids, as evident by increased number and size of lipid
droplets; a process resulting in massive TG accumulation
and yellowish discoloration of the heart. Lipid accumula-
tion leads to an increased heart mass, decreased contrac-
tility, severe cardiac insufficiency, and premature death
starting at about 12 weeks after birth. Remarkably, cardiac
LPL activity is upregulated, indicating that despite mas-
sive TG accumulation, the FA uptake machinery is maxi-
mally induced. However, due to low VLDL levels in plasma
during fasting, elevated LPL activities are probably not as-
sociated with increased FA uptake in cardiomyocytes. De-
creased FA absorption from plasma is counterbalanced by
increased uptake of glucose in ATGL-ko hearts. Thus, in
the absence of ATGL, the release of FA from TG is blocked,
and increased glucose utilization for energy production.
A recently discovered pathway to reduce excessive car-
diac lipids involves the synthesis and secretion of apoB-
containing lipoproteins (172). This mobilization of
myocardial TG for lipoprotein secretion is thought to
provide the heart with a “safety valve” for the disposal of
excess lipids. It has been proposed that this process re-
quires the hydrolysis of cytoplasmic TG stores and resyn-
thesis of TG in the endoplamic reticulum. Whether this
lipolytic step requires ATGL is presently unknown. Indi-
rect evidence for such an involvement is provided by the fact
that cardiac lipoprotein synthesis apparently cannot prevent
the lethal lipid accumulation in hearts of ATGL-ko mice.
Whether defective ATGL function in humans with
NLSDM or NLSDI also causes cardiac dysfunction is not
clear from the few cases known so far. Cardiomyopathy
was reported in patients with both conditions, NLSDM
(147, 148) and NLSDI (173), although it appears to be
much less severe or less frequent in the latter. Reportedly,
two patients with ATGL deficiency died from cardiac fail-
ure, but it is not known whether excessive lipid accumula-
tion caused the premature death.
The release of FA from TG within skeletal myofibrillar
lipid droplets requires lipases (174). The presence of
HSL mRNA, protein, and enzyme activity has been docu-
mented in rodent (170, 175, 176) and in human skeletal
muscle (177) by several laboratories. Compatible with its
physiological function, HSL expression in skeletal muscle
varies between fiber types, being higher in oxidative than
in glycolytic fibers (175, 176). HSL in skeletal muscle is ac-
tivated by PKA-mediated phosphorylation, a contraction-
induced mechanism involving PKC, and the ERK pathway
(175, 178, 179). An inhibitory effect of AMPK on HSL ac-
tivity in resting and contracting muscle was reported by
some (180–182) but not all studies (177). Despite the es-
tablished role of HSL for the hydrolysis of stored TG in
skeletal muscle, the absence of the enzyme in skeletal
muscle of HSL-ko mice did not result in elevated muscular
similarly to that observed in other tissues, HSL deficiency
led to increased DG levels in skeletal muscle. Expression
studies of genes involved in energy metabolism revealed
that enzymes involved in carbohydrate metabolism are up-
regulated in HSL-ko skeletal muscles, whereas enzymes
involved in FA biosynthesis are downregulated (183, 184).
In contrast to HSL-ko mice, ATGL-ko mice accumulated
TG in skeletal muscle (99), supporting the concept that in
addition to HSL, ATGL is also involved in the lipolytic cas-
cade in myocytes. The absence of ATGL in skeletal muscle
causes reduced lipolytic activity, neutral lipid droplet accu-
mulation in oxidative muscle fibers, and elevated glucose
uptake. Increased respiratory quotient (RQ) values during
fasting in ATGL-ko compared with wild-type animals indi-
cated increased glucose utilization in the absence of ATGL.
Additional evidence for a functional role of ATGL was pro-
vided by Watt et al. (185), showing that overexpression of
ATGL inskleletal muscle increases the oxidation ofFAfrom
TG stores and increases DG and ceramide production.
Apparently, in this experimental setup, the endogenous
HSL activity was not sufficient to hydrolyze excess DG. Ac-
cordingly, a dysequilibrium between the activities for ATGL
and HSL might contribute to the production of lipotoxic
intermediates and promote insulin resistance.
The crucial role of ATGL in skeletal muscle energy
metabolism in humans is strikingly supported by recent
findings in patients with NLSD. Individuals that lack ATGL
(147–149, 151) or its activator, CGI-58 (144), accumulate
TG in myocytes and develop muscle weakness and skeletal
myopathy. Myopathy appears to be consistently more severe
in patients with defective ATGL (NLSDM). The clinical
phenotype of complete ATGL deficiency is not known be-
cause patients that completely lack ATGL have not been
found so far. However, the severe myopathy in a patient
with a frameshift mutation at amino acid 160 suggests that
patients lacking the patatin domain are affected with a
more severe form of the disease than those with an intact
12 Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
patatin domain (147, 148). The observation of milder
myopathy in CGI-58 deficient individuals (NLSDI) may
be explained by the remnant ATGL activity in muscle that
is still present when the ATGL activator is absent. The avail-
able information suggests that, similarly to tissues, HSL
and ATGL work coordinately within the lipolytic pathway
to provide FAs for oxidation and ATP production.
Despite the fact that ATGL expression in hepatocytes is
relatively low compared with other tissues, both ATGL-ko
mice and patients affected with NLSDM or NLSDI develop
hepatosteatosis, suggesting that the lipase is functional in
the liver. Increased ATGL expression in HepG2 cells
causes increased lipid mobilization and a depletion of cel-
lular TG stores (186). Similarly, overexpression of CGI-58
is associated with an increased lipolytic degradation of
lipid droplets and increased secretion of TG-rich lipopro-
teins from cultivated hepatoma cells (187). Whether CGI-58
activates solely ATGL in hepatocytes or whether this pro-
cess also affects lipoprotein biogenesis and secretion in
vivo remains to be investigated. As mentioned above,
hepatic ATGL is also regulated by PEDF (134). This inter-
action affects TG hydrolysis and might establish a new
mechanism of ATGL regulation independent of CGI-58.
Other cell types and tissues with relatively high levels of
ATGL/CGI-58 and HSL expression include macrophages,
pancreatic b-cells, intestinal mucosa cells, pigment epithe-
lial cells, and testis. Future studies are needed to elucidate
the role of ATGL-mediated lipolysis and its integration
with HSL and MGL activity in these tissues.
HSL AND ATGL IN CARBOHYDATE METABOLISM
AND INSULIN ACTION
Elevated concentrations of plasma FAs and excess lipid
deposition in insulin target tissues are major predisposing
factors toward the development of insulin resistance and
overt type 2 diabetes. This raises the possibility that en-
zymes participating in lipolysis influence carbohydrate me-
tabolism and insulin sensitivity. Evidence for this concept
was observed in mice lacking either HSL or ATGL.
Studies with HSL-ko mice revealed that HSL influences
not only insulin sensitivity in peripheral tissues but also in-
sulin secretion from pancreatic islet cells. However, reports
have been controversial, probably due to the different
genetic backgrounds of mice and nonuniformity in the ex-
perimental conditions. Using hyperinsulinemic euglycemic
clamp studies, HSL-ko mice exhibited enhanced whole-
body glucose uptake and protection from diet-induced in-
sulin resistance in muscle, heart, and WAT on a high-fat
diet (188). On a chow diet, HSL-ko mice showed increased
hepatic insulin sensitivity (188, 189). These changes were
associated with decreased tissue TG concentrations and
fasting plasma FA levels (188, 189). Conversely, Mulder
et al. (190) demonstrated decreased hepatic insulin action
in HSL-ko mice. Insulin tolerance tests (ITTs) suggested
whole-body insulin resistance in HSL-ko mice (190, 191).
Consistently, ex vivo insulin-stimulated glucose transport
into soleus muscle and lipogenesis in isolated adipocytes
of HSL-ko mice were reduced (190). However, further ex-
amination indicated that the insulin sensitivity phenotype
of HSL-ko mice depends on nutritional status and mouse
gender (192). Importantly, pharmacological inhibition of
HSL reduces hyperglycemia in streptozotozin-induced
diabetic rats, suggesting improved insulin sensitivity (193).
Moderately decreased glucose tolerance and the blunted
response of plasma insulin levels upon glucose administra-
tion suggested impaired glucose-stimulated insulin secre-
tion (GSIS) from pancreatic islets (191).
Extensive investigations with isolated islets of HSL-ko
mice demonstrated that HSL plays a critical role in b-cell
lipid metabolism, particularly neutral cholesteryl ester hy-
drolysis, and lipid signaling facilitating GSIS (191, 192).
However, the effect of HSL deficiency on GSIS varied with
gender, age, nutritional state (192), and genetic back-
ground of mice (194). Recently, a mechanism for HSL-
facilitated GSIS from b-cells was suggested, wherein HSL
serves an important role by providing free cholesterol for
exocytosis of insulin granules (195). Besides its role in
whole-body lipid metabolism and GSIS from pancreatic
islets, HSL was also implicated in regulating adipokine
mRNA and protein levels (78, 189). Moreover, HSL defi-
ciency was associated with inflammation in and macro-
phage infiltration into WAT (196, 197), which is generally
accepted as a causative factor in the development of insulin
resistance in conjunction with obesity. This mechanism may
provide another route of HSL affecting insulin sensitivity.
Because of the short time period since its discovery in
2004, much less information is currently available on the
role of ATGL in carbohydrate metabolism. We previously
demonstrated that ATGL-ko mice exhibit enhanced whole-
body insulin sensitivity and glucose tolerance during ITTs
and glucose tolerance tests (GTTs), respectively. This is re-
markable, considering the massive lipid accumulation in
multiple insulin target tissues, and implies an important
role of ATGL in tissue insulin signaling and glucose metabo-
lism (99). Whether changes in the concentration of other
lipid species in ATGL-ko mice, such as DG, are involved is
currently unknown. Furthermore, 2-deoxyglucose uptake
under conditions similar to the GTT was significantly in-
creased in skeletal muscle, heart, and liver. This observa-
tion is in accordance with enhanced glucose utilization in
these tissues. The RQ was elevated in ATGL-ko mice during
fasting, and ATGL-ko mice exhibited hypoglycemia upon
short-term (4–6 h) fasting. Interestingly, plasma insulin
concentrations were consistently decreased in fed ATGL-
deficient mice. However, in vitro and in vivo studies em-
ploying overexpression and silencing of ATGL in rat
skeletal muscle failed to demonstrate a connection be-
tween ATGL and muscle insulin sensitivity (185).
Data on parameters of carbohydrate metabolism are
scarce in humans with ATGL deficiency. Interestingly, how-
ever, a Japanese patient exhibited decreased insulin levels
despite normal insulin sensitivity (149). This data supports
the concept that lipid accumulation in nonadipose tissue is
not causative for insulin resistance and is consistent with a
Role of ATGL in lipolysis13
by guest, on August 22, 2013
study in humans showing an association between a nucleo-
tide polymorphisms in the ATGL gene and increased con-
centrations of plasma FAs, TG, and glucose, as well as an
increased risk for the development of type 2 diabetes
(124). Accordingly, low ATGL activity may be associated
with improved insulin sensitivity in mice and humans.
MGL, THE FINAL STEP IN TG HYDROLYSIS
Current data suggest that MGL is the rate-limiting en-
zyme in MG degradation (198). MGL was first purified
in 1976 from rat adipose tissue (199). The enzyme specifi-
cally hydrolyzes MG and has no activity against DG or TG.
Cloning of the mouse cDNA encoding MGL revealed that
the enzyme is composed of 302 amino acids with a molecu-
lar mass of 33 kDa (200–202). Although not directly related
to other lipases in the human genome, MGL belongs to the
large family of a/b hydrolase fold proteins with a GXSXG
motif. The amino acids building the active site consist of
serine 122, aspartate 239, and histidine 269 (202). Mouse
and human MGLs share 84% identity. The enzyme is ubiq-
uitously expressed at relatively high levels and is found in
the cytoplasm, the plasma membrane, and on lipid droplets.
The high specific activity of the enzyme toward medium-
and long-chain MG and its abundance in many tissues
(highest inWAT) suggest that the final stepinTG hydrolysis
is not extensively regulated. However, definitive data on
the physiological role of MGL are missing because mouse
models that lack or overexpress MGL have not been re-
ported to date. In addition to MGL, HSL (203), a/b hydro-
and fatty acid amide hydrolase (FAAH) (205, 206) were also
identified as MG hydrolases in in vitro assays. The contribu-
tion of these enzymes to MG hydrolysis in vivo is not known.
Certain species of MG serve important signaling func-
tions. The MG 2-arachidonoyl glycerol (2-AG) belongs to
a family of compounds designated as endocannabinoids
(ECs), which are endogenous agonists of cannabinoid
(CB) receptors (207, 208). ECs are involved in the control
of many biologic processes, including behavior, appetite
regulation, pain, blood pressure, energy metabolism, in-
flammation, and cell growth. Their biological effect is
mimicked by D9-tetrahydrocannabinol, the major psycho-
active component of marijuana. ECs are amides or esters
of long-chain PUFAs and, besides 2-AG, anandamide [N-
arachidonoylethanolamine (AEA)] has been identified as
the main endogenous agonist of CB receptors (209). FAAH
is believed to be responsible for the degradation of AEA,
whereas MGL degrades 2-AG. Generally, MG, and specifi-
cally 2-AG, can be produced from three distinct lipolytic
pathways. First, in the plasma membrane, glycerophospho-
Fig. 3. Simplified summary of the lipolytic process and the involved metabolic intermediates. Several
potential utilization pathways of these intermediates are indicated. In most tissues, lipid droplet-associated
ATGL, together with its activator, CGI-58, hydrolyzes triacylglycerol (TG) to generate diacylglycerol (DG).
The contribution of hormone-sensitive lipase (HSL) to TG hydrolysis may vary among tissues and species.
Whether DGs can directly enter glycerophospholipid synthesis or operate as signaling molecules to activate
various isoforms of protein kinase C requires clarification. HSL converts DG to monoacylglycerols (MGs). It
is not known whether these MGs can contribute to signaling processes via the endocannabinoid system. The
absence of acyl-CoA monoacylglycerol acyltransferase and the high activity of monoglyceride lipase (MGL)
prevent reesterification of MG in most cell types and tissues (except intestinal mucosa cells). MGL hydro-
lyzes MG to glycerol and the last FA. Subsequent glycerol utilization in glycolytic or gluconeogenic pathways
mostly occurs in the liver.
14Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
lipid degradation via phospholipase C results in DG forma-
tion, which is subsequently hydrolyzed by the sn-1-specific
diacylglycerol lipase to yield 2-AG (210). Second, TG hy-
drolysis catalyzed by ATGL and the subsequent DG hydro-
lysis by HSL result in the production of MG within lipid
droplets. Third, hydrolysis of plasma lipoprotein-associated
TG by LPL results in the formation of FAs and MG, which
are efficiently internalized by parenchymal cells (211).
Whether all these sources contribute to the generation of
bioactive MGs is currently not known, but it is generally
assumed that MGL is the major enzyme for MG and 2-AG
degradation, thereby inactivating CB receptor ligands
(205, 206). MGL often colocalizes with CB1 receptors in
the brain (212).
Importantly, the EC system plays a major role in the con-
trol of energy homeostasis and in the pathogenesis of
metabolic diseases including obesity, metabolic syndrome,
and type 2 diabetes. ECs affect energy balance via central
orexigenic effects and by modulation of peripheral lipo-
genesis (213). CB receptor agonists are used as orexigenic
agents in patients suffering from anorexia, whereas CB an-
tagonists are used for the treatment of obesity-related disor-
ders (214). The European Commission recently approved
the CB1 receptor antagonist Rimonabant for diabetes treat-
mentand asananti-obesitydrug (Trade name:Acomplia:).
Inhibition of 2-AG-hydrolyzing enzymes, like MGL, is also
considered a promising pharmacological approach to mod-
ulating MG levels (215, 216).
CONCLUSION AND PERSPECTIVES
Recent discoveries concerning new enzymes and coacti-
vators have led to a revision of the lipolytic pathway, which
catabolizes stored TG (summarized in Fig. 3). In conclu-
sion, ATGL and HSL govern the capacity of adipose tissue
and nonadipose tissues to mobilize FAs. ATGL is the major
enzyme responsible for the first step in TG mobilization,
generating DG and FA. HSL is rate-limiting for the sub-
sequent degradation of DG. In the final hydrolytic step,
MGL releases the third FA from the glycerol backbone.
Each step provides a number of metabolic intermediates
that act as precursors for a variety of metabolic pathways
as well as bioactive molecules in cell signaling and gene
regulation. In accord with ATGLʼs specific role in initiating
the lipolytic cascade, the phenotype of ATGL-ko mice is
more severe compared with HSL-ko mice.
However, important questions remain to be answered in
order to understand the “lipolysome”: Which other lipases
besides the ones described so far contribute to the catabo-
lism of stored fat in various tissues? Although ATGL, HSL,
and MGL contribute more than 95% of the lipolytic capac-
ity in WAT, it is conceivable that other lipases such as Ces 3
or TGH2 (101, 102) contribute to TG hydrolysis under cer-
tain physiological conditions in WAT and other tissues.
How are the involved lipases, particularly ATGL, regulated
in different tissues? The mechanisms of how established
and newly identified lipid droplet proteins such as CGI-58,
PAT proteins, CideN proteins, and the Arf-COPI vesicle
transport system regulate the enzymatic activity of lipases
require elucidation. How does lipolysis affect metabolic
processes independently of FA production as energy sub-
strates? Besides their role as energy substrates, the lipolytic
products FA, DG, and MG fulfill other important physio-
logical functions as precursors of various lipid classes, such
as membrane lipids, prostaglandins and leukotrienes, and
ceramides. Many of these lipids, as well as FA and DG
per se, act as signaling molecules in signal transduction
pathways and in the regulation of gene transcription
(e.g., PPARs). An increased understanding of the role of
lipases and lipolysis controlling these processes is of utmost
importance. What is the role of ATGL and CGI-58 in the
pathogenesis of metabolic disorders including inherited
diseases like NLSD and acquired conditions like obesity,
metabolic syndrome, and type 2 diabetes? The molecular
basis of the phenotypical differences observed in ATGL de-
ficiency versus CGI-58 deficiency is currently unknown and
mice in comparison to ATGL-ko mice will help to solve this
problem. Answers to these questions will facilitate the devel-
opment of potential treatments for patients with NLSD.
Additionally, better understanding of the “lipolysome”
might lead to pharmacological treatment controlling the
release of FA and other lipolytic products involved in the
development of insulin resistance and type 2 diabetes.
The authors would like to thank Dr. Ellen Zechner and Mag.
Caroline Schober for critically reviewing the manuscript and
Mag. Renate Schreiber for her help in preparing the figures.
1. Schaffer, J. E. 2003. Lipotoxicity: when tissues overeat. Curr. Opin.
Lipidol. 14: 281–287.
2. Unger, R. H. 2002. Lipotoxic diseases. Annu. Rev. Med. 53:
3. Brasaemle, D. L. 2007. Thematic review series: adipocyte biology.
The perilipin family of structural lipid droplet proteins: stabili-
zation of lipid droplets and control of lipolysis. J. Lipid Res. 48:
4. Brasaemle, D. L., G. Dolios, L. Shapiro, and R. Wang. 2004. Pro-
teomic analysis of proteins associated with lipid droplets of basal
and lipolytically stimulated 3T3–L1 adipocytes. J. Biol. Chem. 279:
5. Londos, C., D. L. Brasaemle, C. J. Schultz, J. P. Segrest, and A. R.
Kimmel. 1999. Perilipins, ADRP, and other proteins that associate
with intracellular neutral lipid droplets in animal cells. Semin. Cell
Dev. Biol. 10: 51–58.
6. Beckman, M. 2006. Cell biology. Great balls of fat. Science. 311:
7. Martin, S., and R. G. Parton. 2006. Lipid droplets: a unified view
of a dynamic organelle. Nat. Rev. Mol. Cell Biol. 7: 373–378.
8. Ahima, R. S., and M. A. Lazar. 2008. Adipokines and the periph-
eral and neural control of energy balance. Mol. Endocrinol. 22:
9. Langin, D. 2006. Adipose tissue lipolysis as a metabolic pathway to
define pharmacological strategies against obesity and the meta-
bolic syndrome. Pharmacol. Res. 53: 482–491.
10. Holm, C., T. Osterlund, H. Laurell, and J. A. Contreras. 2000.
Molecular mechanisms regulating hormone-sensitive lipase and
lipolysis. Annu. Rev. Nutr. 20: 365–393.
11. Lafontan, M., and M. Berlan. 1993. Fat cell adrenergic receptors
and the control of white and brown fat cell function. J. Lipid Res.
Role of ATGL in lipolysis15
by guest, on August 22, 2013
12. Collins, S., W. Cao, and J. Robidoux. 2004. Learning new
tricks from old dogs: beta-adrenergic receptors teach new les-
sons on firing up adipose tissue metabolism. Mol. Endocrinol. 18:
13. Holm, C. 2003. Molecular mechanisms regulating hormone-
sensitive lipase and lipolysis. Biochem. Soc. Trans. 31: 1120–1124.
14. Larrouy, D., J. Galitzky, and M. Lafontan. 1991. A1 adenosine re-
ceptors in the human fat cell: tissue distribution and regulation of
radioligand binding. Eur. J. Pharmacol. 206: 139–147.
15. Richelsen, B. 1992. Release and effects of prostaglandins in adi-
pose tissue. Prostaglandins Leukot. Essent. Fatty Acids. 47: 171–182.
16. Bradley, R. L., J. P. Mansfield, and E. Maratos-Flier. 2005. Neuro-
peptides, including neuropeptide Y and melanocortins, mediate
lipolysis in murine adipocytes. Obes. Res. 13: 653–661.
17. Offermanns, S. 2006. The nicotinic acid receptor GPR109A
(HM74A or PUMA-G) as a new therapeutic target. Trends Pharmacol.
Sci. 27: 384–390.
18. Degerman, E., T. R. Landstrom, J. Wijkander, L. S. Holst, F.
Ahmad, P. Belfrage, and V. Manganiello. 1998. Phosphorylation
and activation of hormone-sensitive adipocyte phosphodiesterase
type 3B. Methods. 14: 43–53.
19. Taniguchi, C. M., B. Emanuelli, and C. R. Kahn. 2006. Critical
nodes in signalling pathways: insights into insulin action. Nat.
Rev. Mol. Cell Biol. 7: 85–96.
20. Quagliarello, G., and G. Scoz. 1932. The existence of a lipase in
adipose tissue. Arch. Sci. Biol. 17: 513–529.
21. Renold, A. E., and A. Marble. 1950. Lipolytic activity of adipose
tissue in man and rat. J. Biol. Chem. 185: 367–375.
22. Bjorntorp, P., and R. H. Furman. 1962. Lipolytic activity in rat
heart. Am. J. Physiol. 203: 323–326.
23. Bjorntorp, P., and R. H. Furman. 1962. Lipolytic activity in rat epi-
didymal fat pads. Am. J. Physiol. 203: 316–322.
24. Hollenberg, C. H., M. S. Raben, and E. B. Astwood. 1961. The
lipolytic response to corticotropin. Endocrinology. 68: 589–598.
25. Rizack, M. A. 1964. Activation of an epinephrine-sensitive lipolytic
activity from adipose tissue by adenosine 3′,5′-phosphate. J. Biol.
Chem. 239: 392–395.
26. Vaughan, M., J. E. Berger, and D. Steinberg. 1964. Hormone-
sensitive lipase and monoglyceride lipase activities in adipose tissue.
J. Biol. Chem. 239: 401–409.
27. Donsmark, M., J. Langfort, C. Holm, T. Ploug, and H. Galbo. 2005.
Hormone-sensitive lipase as mediator of lipolysis in contracting
skeletal muscle. Exerc. Sport Sci. Rev. 33: 127–133.
28. Haemmerle, G., R. Zimmermann, and R. Zechner. 2003. Letting
lipids go: hormone-sensitive lipase. Curr. Opin. Lipidol. 14: 289–297.
29. Kraemer, F. B., and W. J. Shen. 2002. Hormone-sensitive lipase:
control of intracellular tri-(di-)acylglycerol and cholesteryl ester
hydrolysis. J. Lipid Res. 43: 1585–1594.
30. Yeaman, S. J. 2004. Hormone-sensitive lipase—new roles for an old
enzyme. Biochem. J. 379: 11–22.
31. Yeaman, S. J. 1990. Hormone-sensitive lipase—a multipurpose
enzyme in lipid metabolism. Biochim. Biophys. Acta. 1052: 128–132.
32. Haemmerle, G., R. Zimmermann, M. Hayn, C. Theussl, G. Waeg,
E. Wagner, W. Sattler, T. M. Magin, E. F. Wagner, and R. Zechner.
2002. Hormone-sensitive lipase deficiency in mice causes diglyceride
accumulation in adipose tissue, muscle, and testis. J. Biol. Chem. 277:
33. Fredrikson, G., and P. Belfrage. 1983. Positional specificity of
hormone-sensitive lipase from rat adipose tissue. J. Biol. Chem.
34. Cook, K. G., S. J. Yeaman, P. Stralfors, G. Fredrikson, andP. Belfrage.
1982. Direct evidence that cholesterol ester hydrolase from adrenal
cortex is the same enzyme as hormone-sensitive lipase from adi-
pose tissue. Eur. J. Biochem. 125: 245–249.
35. Fredrikson, G., P. Stralfors, N. O. Nilsson, and P. Belfrage. 1981.
Hormone-sensitive lipase of rat adipose tissue. Purification and
some properties. J. Biol. Chem. 256: 6311–6320.
36. Holm, C., T. G. Kirchgessner, K. L. Svenson, G. Fredrikson, S.
Nilsson, C. G. Miller, J. E. Shively, C. Heinzmann, R. S. Sparkes,
T. Mohandas, et al. 1988. Hormone-sensitive lipase: sequence, ex-
pression, and chromosomal localization to 19 cent-q13.3. Science.
37. Blaise, R., J. Grober, P. Rouet, G. Tavernier, D. Daegelen, and D.
Langin. 1999. Testis expression of hormone-sensitive lipase is con-
ferred by a specific promoter that contains four regions binding
testicular nuclear proteins. J. Biol. Chem. 274: 9327–9334.
38. Blaise, R., T. Guillaudeux, G. Tavernier, D. Daegelen, B. Evrard, A.
Mairal, C. Holm, B. Jegou, and D. Langin. 2001. Testis hormone-
sensitive lipase expression in spermatids is governed by a short
promoter in transgenic mice. J. Biol. Chem. 276: 5109–5115.
39. Grober, J., H. Laurell, R. Blaise, B. Fabry, S. Schaak, C. Holm, and
D. Langin. 1997. Characterization of the promoter of human adi-
pocyte hormone-sensitive lipase. Biochem. J. 328: 453–461.
40. Langin, D., H. Laurell, L. S. Holst, P. Belfrage, and C. Holm. 1993.
Gene organization and primary structure of human hormone-
sensitive lipase: possible significance of a sequence homology with
a lipase of Moraxella TA144, an antarctic bacterium. Proc. Natl.
Acad. Sci. USA. 90: 4897–4901.
41. Holm,C., R.C.Davis, T.Osterlund,M.C.Schotz,andG.Fredrikson.
1994. Identification of the active site serine of hormone-sensitive
lipase by site-directed mutagenesis. FEBS Lett. 344: 234–238.
42. Osterlund, T., D. J. Beussman, K. Julenius, P. H. Poon, S. Linse,
J. Shabanowitz, D. F. Hunt, M. C. Schotz, Z. S. Derewenda, and
C. Holm. 1999. Domain identification of hormone-sensitive lipase
by circular dichroism and fluorescence spectroscopy, limited pro-
teolysis, and mass spectrometry. J. Biol. Chem. 274: 15382–15388.
43. Osterlund, T., J. A. Contreras, and C. Holm. 1997. Identification
of essential aspartic acid and histidine residues of hormone-
sensitive lipase: apparent residues of the catalytic triad. FEBS Lett.
44. Osterlund, T., B. Danielsson, E. Degerman, J. A. Contreras, G.
Edgren, R. C. Davis, M. C. Schotz, and C. Holm. 1996. Domain-
structure analysis of recombinant rat hormone-sensitive lipase.
Biochem. J. 319: 411–420.
45. Shen, W. J., S. Patel, R. Hong, and F. B. Kraemer. 2000. Hormone-
sensitive lipase functions as an oligomer. Biochemistry. 39: 2392–2398.
46. Shen, W. J., Y. Liang, R. Hong, S. Patel, V. Natu, K. Sridhar, A.
Jenkins, D. A. Bernlohr, and F. B. Kraemer. 2001. Characterization
of the functional interaction of adipocyte lipid-binding protein
with hormone-sensitive lipase. J. Biol. Chem. 276: 49443–49448.
47. Shen, W. J., K. Sridhar, D. A. Bernlohr, and F. B. Kraemer. 1999.
Interaction of rat hormone-sensitive lipase with adipocyte lipid-
binding protein. Proc. Natl. Acad. Sci. USA. 96: 5528–5532.
48. Smith, A. J., B. R. Thompson, M. A. Sanders, and D. A. Bernlohr.
2007. Interaction of the adipocyte fatty acid-binding protein with
the hormone-sensitive lipase: regulation by fatty acids and phos-
phorylation. J. Biol. Chem. 282: 32424–32432.
49. Stralfors, P., and P. Belfrage. 1983. Phosphorylation of hormone-
sensitive lipase by cyclic AMP-dependent protein kinase. J. Biol.
Chem. 258: 15146–15152.
50. Stralfors, P., P. Bjorgell, and P. Belfrage. 1984. Hormonal regula-
tion of hormone-sensitive lipase in intact adipocytes: identifica-
tion of phosphorylated sites and effects on the phosphorylation
by lipolytic hormones and insulin. Proc. Natl. Acad. Sci. USA. 81:
51. Garton, A. J., D. G. Campbell, P. Cohen, and S. J. Yeaman. 1988.
Primary structure of the site on bovine hormone-sensitive lipase
phosphorylated by cyclic AMP-dependent protein kinase. FEBS
Lett. 229: 68–72.
52. Garton, A. J., and S. J. Yeaman. 1990. Identification and role of the
basal phosphorylation site on hormone-sensitive lipase. Eur. J. Biochem.
53. Olsson, H., and P. Belfrage. 1987. The regulatory and basal phos-
phorylation sites of hormone-sensitive lipase are dephosphory-
lated by protein phosphatase-1, 2A and 2C but not by protein
phosphatase-2B. Eur. J. Biochem. 168: 399–405.
54. Anthonsen, M. W., L. Ronnstrand, C. Wernstedt, E. Degerman,
and C. Holm. 1998. Identification of novel phosphorylation sites
in hormone-sensitive lipase that are phosphorylated in response
to isoproterenol and govern activation properties in vitro. J. Biol.
Chem. 273: 215–221.
55. Greenberg, A. S., W. J. Shen, K. Muliro, S. Patel, S. C. Souza, R. A.
Roth, and F. B. Kraemer. 2001. Stimulation of lipolysis and
hormone-sensitive lipase via the extracellular signal-regulated
kinase pathway. J. Biol. Chem. 276: 45456–45461.
56. Olsson, H., P. Stralfors, and P. Belfrage. 1986. Phosphorylation of
the basal site of hormone-sensitive lipase by glycogen synthase
kinase-4. FEBS Lett. 209: 175–180.
57. Garton, A. J., D. G. Campbell, D. Carling, D. G. Hardie, R. J.
Colbran, and S. J. Yeaman. 1989. Phosphorylation of bovine
hormone-sensitive lipase by the AMP-activated protein kinase.
A possible antilipolytic mechanism. Eur. J. Biochem. 179: 249–254.
58. Olsson, H., and P. Belfrage. 1988. Phosphorylation and dephos-
phorylation of hormone-sensitive lipase. Interactions between
16Journal of Lipid Research
Volume 50, 2009
by guest, on August 22, 2013
the regulatory and basal phosphorylation sites. FEBS Lett. 232:
59. Granneman, J. G., and H. P. Moore. 2008. Location, location: pro-
tein trafficking and lipolysis in adipocytes. Trends Endocrinol. Metab.
60. Londos, C., C. Sztalryd, J. T. Tansey, and A. R. Kimmel. 2005. Role
of PAT proteins in lipid metabolism. Biochimie. 87: 45–49.
61. Tansey, J. T., C. Sztalryd, E. M. Hlavin, A. R. Kimmel, and C. Londos.
2004. The central role of perilipin a in lipid metabolism and adipo-
cyte lipolysis. IUBMB Life. 56: 379–385.
62. Greenberg, A. S., J. J. Egan, S. A. Wek, N. B. Garty, E. J.
Blanchette-Mackie, and C. Londos. 1991. Perilipin, a major
hormonally regulated adipocyte-specific phosphoprotein asso-
ciated with the periphery of lipid storage droplets. J. Biol. Chem.
63. Miyoshi, H., J. W. Perfield II, S. C. Souza, W. J. Shen, H. H. Zhang,
Z. S. Stancheva, F. B. Kraemer, M. S. Obin, and A. S. Greenberg.
2007. Control of adipose triglyceride lipase action by serine 517 of
perilipin A globally regulates protein kinase A-stimulated lipolysis
in adipocytes. J. Biol. Chem. 282: 996–1002.
64. Miyoshi, H., S. C. Souza, H. H. Zhang, K. J. Strissel, M. A.
Christoffolete, J. Kovsan, A. Rudich, F. B. Kraemer, A. C. Bianco,
M. S. Obin, et al. 2006. Perilipin promotes hormone-sensitive lipase-
mediated adipocyte lipolysis via phosphorylation-dependent
and -independent mechanisms. J. Biol. Chem. 281: 15837–15844.
65. Sztalryd, C., G. Xu, H. Dorward, J. T. Tansey, J. A. Contreras, A. R.
Kimmel, and C. Londos. 2003. Perilipin A is essential for the trans-
location of hormone-sensitive lipase during lipolytic activation.
J. Cell Biol. 161: 1093–1103.
66. Tansey, J. T., A. M. Huml, R. Vogt, K. E. Davis, J. M. Jones, K. A.
Fraser, D. L. Brasaemle, A. R. Kimmel, and C. Londos. 2003. Func-
tional studies on native and mutated forms of perilipins. A role in
protein kinase A-mediated lipolysis of triacylglycerols. J. Biol. Chem.
67. Clifford, G. M., C. Londos, F. B. Kraemer, R. G. Vernon, and S. J.
Yeaman. 2000. Translocation of hormone-sensitive lipase and
perilipin upon lipolytic stimulation of rat adipocytes. J. Biol. Chem.
68. Martinez-Botas, J., J. B. Anderson, D. Tessier, A. Lapillonne, B. H.
Chang, M. J. Quast, D. Gorenstein, K. H. Chen, and L. Chan. 2000.
Absence of perilipin results in leanness and reverses obesity in
Lepr(db/db) mice. Nat. Genet. 26: 474–479.
69. Tansey, J. T., C. Sztalryd, J. Gruia-Gray, D. L. Roush, J. V. Zee, O.
Gavrilova, M. L. Reitman, C. X. Deng, C. Li, A. R. Kimmel, et al.
2001. Perilipin ablation results in a lean mouse with aberrant
adipocyte lipolysis, enhanced leptin production, and resistance
to diet-induced obesity. Proc. Natl. Acad. Sci. USA. 98: 6494–6499.
70. Marcinkiewicz, A., D. Gauthier, A. Garcia, and D. L. Brasaemle.
2006. The phosphorylation of serine 492 of perilipin a directs lipid
71. Moore, H. P., R. B. Silver, E. P. Mottillo, D. A. Bernlohr, and J. G.
Granneman. 2005. Perilipin targets a novel pool of lipid droplets
for lipolytic attack by hormone-sensitive lipase. J. Biol. Chem. 280:
72. Ducharme, N. A., and P. E. Bickel. 2008. Lipid droplets in lipo-
genesis and lipolysis. Endocrinology. 149: 942–949.
73. Syu, L. J., and A. R. Saltiel. 1999. Lipotransin: a novel docking
protein for hormone-sensitive lipase. Mol. Cell. 4: 109–115.
74. Haemmerle, G., R. Zimmermann, J. G. Strauss, D. Kratky, M.
Riederer, G. Knipping, and R. Zechner. 2002. Hormone-sensitive
lipase deficiency in mice changes the plasma lipid profile by affect-
ing the tissue-specific expression pattern of lipoprotein lipase in
adipose tissue and muscle. J. Biol. Chem. 277: 12946–12952.
75. Osuga, J., S. Ishibashi, T. Oka, H. Yagyu, R. Tozawa, A. Fujimoto, F.
Shionoiri,N.Yahagi,F.B.Kraemer,O. Tsutsumi,et al.2000.Targeted
disruption of hormone-sensitive lipase results in male sterility and
adipocyte hypertrophy, but not in obesity. Proc. Natl. Acad. Sci. USA.
76. Wang, S. P., N. Laurin, J. Himms-Hagen, M. A. Rudnicki, E. Levy,
M. F. Robert, L. Pan, L. Oligny, and G. A. Mitchell. 2001. The adipose
tissue phenotype of hormone-sensitive lipase deficiency in mice.
Obes. Res. 9: 119–128.
77. Zimmermann, R., J. G. Strauss, G. Haemmerle, G. Schoiswohl,
R. Birner-Gruenberger, M. Riederer, A. Lass, G. Neuberger, F.
Eisenhaber, A. Hermetter, et al. 2004. Fat mobilization in adipose
tissue is promoted by adipose triglyceride lipase. Science. 306:
78. Harada, K., W. J. Shen, S. Patel, V. Natu, J. Wang, J. Osuga, S.
Ishibashi, and F. B. Kraemer. 2003. Resistance to high-fat diet-
induced obesity and altered expression of adipose-specific genes
in HSL-deficient mice. Am. J. Physiol. Endocrinol. Metab. 285:
79. Okazaki, H., J. Osuga, Y. Tamura, N. Yahagi, S. Tomita, F. Shionoiri,
Y. Iizuka, K. Ohashi, K. Harada, S. Kimura, et al. 2002. Lipolysis in
the absence of hormone-sensitive lipase: evidence for a common
mechanism regulating distinct lipases. Diabetes. 51: 3368–3375.
80. Villena, J. A., S. Roy, E. Sarkadi-Nagy, K. H. Kim, and H. S. Sul.
2004. Desnutrin, an adipocyte gene encoding a novel patatin
domain-containing protein, is induced by fasting and glucocorti-
coids: ectopic expression of desnutrin increases triglyceride hydro-
lysis. J. Biol. Chem. 279: 47066–47075.
81. Jenkins, C. M., D. J. Mancuso, W. Yan, H. F. Sims, B. Gibson, and
R. W. Gross. 2004. Identification, cloning, expression, and purifica-
tion of three novel human calcium-independent phospholipase
A2 family members possessing triacylglycerol lipase and acylglycerol
transacylase activities. J. Biol. Chem. 279: 48968–48975.
82. Eastmond, P. J. 2006. SUGAR-DEPENDENT1 encodes a patatin
domain triacylglycerol lipase that initiates storage oil breakdown
in germinating Arabidopsis seeds. Plant Cell. 18: 665–675.
83. Gronke, S., A. Mildner, S. Fellert, N. Tennagels, S. Petry, G. Muller,
H. Jackle, and R. P. Kuhnlein. 2005. Brummer lipase is an evolu-
tionary conserved fat storage regulator in Drosophila. Cell Metab. 1:
84. Kurat, C. F., K. Natter, J. Petschnigg, H. Wolinski, K. Scheuringer,
H. Scholz, R. Zimmermann,R.Leber, R. Zechner,and S.D.Kohlwein.
2006. Obese yeast: triglyceride lipolysis is functionally conserved
from mammals to yeast. J. Biol. Chem. 281: 491–500.
85. Saarela, J., G. Jung, M. Hermann, J. Nimpf, and W. J. Schneider.
2008. The patatin-like lipase family in Gallus gallus. BMC Genomics.
86. Shan, T., Y. Wang, T. Wu, J. Guo, J. Liu, J. Feng, and Z. Xu. 2008.
Porcine adipose triglyceride lipase complementary deoxyribo-
nucleic acid clone, expression pattern, and regulation by resveratrol.
J. Anim. Sci. 86: 1781–1788.
87. Lake, A. C., Y. Sun, J. L. Li, J. E. Kim, J. W. Johnson, D. Li, T. Revett,
H. H. Shih, W. Liu, J. E. Paulsen, et al. 2005. Expression, regula-
tion, and triglyceride hydrolase activity of Adiponutrin family
members. J. Lipid Res. 46: 2477–2487.
88. Notari, L., V. Baladron, J. D. Aroca-Aguilar, N. Balko, R. Heredia,
C. Meyer, P. M. Notario, S. Saravanamuthu, M. L. Nueda, F.
Sanchez-Sanchez, et al. 2006. Identification of a lipase-linked cell
membrane receptor for pigment epithelium-derived factor. J. Biol.
Chem. 281: 38022–38037.
89. Smirnova, E., E. B. Goldberg, K. S. Makarova, L. Lin, W. J. Brown,
and C. L. Jackson. 2006. ATGL has a key role in lipid droplet/
adiposome degradationin mammalian cells. EMBO Rep. 7: 106–113.
90. Andrews, D. L., B. Beames, M. D. Summers, and W. D. Park. 1988.
Characterization of the lipid acyl hydrolase activity of the major
potato (Solanum tuberosum) tuber protein, patatin, by cloning
and abundant expression in a baculovirus vector. Biochem. J. 252:
91. Jimenez-Atienzar, M., J. Cabanes, F. Gandia-Herrero, J. Escribano,
F. Garcia-Carmona, and M. Perez-Gilabert. 2003. Determination
of the phospholipase activity of patatin by a continuous spectro-
photometric assay. Lipids. 38: 677–682.
92. Senda, K., H. Yoshioka, N. Doke, and K. Kawakita. 1996. A cytosolic
phospholipase A2 from potato tissues appears to be patatin. Plant
Cell Physiol. 37: 347–353.
93. Wilson, P. A., S. D. Gardner, N. M. Lambie, S. A. Commans, and
D. J. Crowther. 2006. Characterization of the human patatin-like
phospholipase family. J. Lipid Res. 47: 1940–1949.
2005. Lipolysis: pathway under construction. Curr. Opin. Lipidol. 16:
95. Rydel, T. J., J. M. Williams, E. Krieger, F. Moshiri, W. C. Stallings,
S. M. Brown, J. C. Pershing, J. P. Purcell, and M. F. Alibhai. 2003.
The crystal structure, mutagenesis, and activity studies reveal that
patatin is a lipid acyl hydrolase with a Ser-Asp catalytic dyad.
Biochemistry. 42: 6696–6708.
96. Lass, A., R. Zimmermann, G. Haemmerle, M. Riederer, G. Schoiswohl,
M. Schweiger, P. Kienesberger, J. G. Strauss, G. Gorkiewicz, and
R. Zechner. 2006. Adipose triglyceride lipase-mediated lipolysis of
cellular fat stores is activated by CGI-58 and defective in Chanarin-
Dorfman Syndrome. Cell Metab. 3: 309–319.
Role of ATGL in lipolysis17
by guest, on August 22, 2013