A Peptide Derived from G0/G1 Switch Gene 2 Acts as Noncompetitive Inhibitor of Adipose Triglyceride Lipase*

Background: G0S2 acts as an endogenous inhibitor of ATGL. Results: We designed a short peptide that is active in the nanomolar range. The G0S2-derived peptide inhibits ATGL activity in a noncompetitive manner. Conclusion: G0S2-derived peptides act as potent and specific inhibitors of ATGL. Significance: G0S2-derived peptides provide a novel tool to modulate ATGL activity.

Most organisms store excess energy as neutral and inert triacylglycerols (TGs) 2 to ensure the availability of energy in the event of demand. These energy stores can be rapidly mobilized in a process termed lipolysis that generates glycerol and fatty acids (FAs). Glycerol can subsequently be channeled toward glycolysis or TG synthesis. FAs serve as energy substrate, represent signaling molecules, and can be used for anabolic reactions. FA metabolism is closely linked to metabolic diseases. Excess FA concentrations in the circulation lead to deleterious effects on the organism, including ectopic lipid accumulation, insulin resistance, and inflammation, summarized in the term lipotoxicity (1)(2)(3)(4). Consequently, TG hydrolysis and synthesis are tightly regulated processes that ensure an adequate supply of FAs to oxidative tissues while avoiding FA overload in ectopic tissues (5,6). Understanding the mechanisms that regulate lipolysis and concomitantly the release of FAs are of fundamental interest.
Adipose triglyceride lipase (ATGL) hydrolyzes TGs into diacylglycerols and FAs and is required for efficient intracellular lipolysis (7). Global inactivation of ATGL activity leads to massive TG accumulation in all tissues, including the heart of Atgl knock-out mice (8). Accordingly, patients with ATGL deficiency suffer from cardiac and skeletal muscle myopathy and require heart transplants at a relatively young age (9). However, genetic and small molecule-mediated inactivation of ATGL is also associated with beneficial effects on plasma lipid parameters and insulin sensitivity; Atgl-KO mice and WT mice treated with the small molecule inhibitor Atglistatin show reduced plasma FA and TG concentrations (8,10). Consequently, Atgl-KO mice display improved glucose tolerance and insulin sensitivity (8). Recently, it has been reported that Atgl-KO mice are also resistant to the development of cancer-associated cachexia, a multifactorial syndrome characterized by ongoing loss of skeletal muscle and fat mass (11).
Extensive research on the physiological role of ATGL has identified complex regulatory mechanisms for its activity on the transcriptional and post-translational level (12,13). Direct protein-protein interaction of ATGL with the protein comparative gene identification 58  increases ATGL activity up to 20-fold (14). Inhibition of ATGL due to direct interaction with the small, 103-amino acid protein G 0 /G 1 switch gene 2 (G0S2) was discovered in 2010 and confirmed using human and murine proteins (15)(16)(17)(18)(19)(20)(21). The highest G0S2 expression is found in adipose tissues, but it is also present in other tissues, including the liver and the heart (20,22,23). To date, no mutation in human G0S2 has been reported. Global and liver-specific G0s2-KO mouse models were generated very recently and display decreased adipose tissue mass (24,25), enhanced lipolysis in adipose tissue, and a decrease in hepatic TG content (26). Global and adipose tissue-specific overexpression of G0S2 leads to increased fat mass, overall reduction in lipolytic activity, and fatty liver (16,19,26).
Activation of ATGL by CGI-58 is considered as fasting response upon ␤-adrenergic stimulation and phosphorylation of PKA (12). Recent studies have strengthened the role of G0S2 as the so-called master regulator of basal and stimulated lipolysis (27). mRNA levels of G0S2 respond strongly to hormonal stimuli, yet the dose-dependent inhibition of ATGL by G0S2 itself is reported to be independent of phosphorylation events (18,20,21,28,29). Thus, regulation of the expression levels of G0S2, and consequently the protein concentration, resembles a potent mean to regulate lipid mobilization. Elegant studies of Zhang et al. (26) demonstrated an important physiological link of G0S2 in regulating adipose tissue lipolysis and liver energy homeostasis. Accordingly, G0S2 mRNA levels are regulated differently in adipose tissue and the liver. During feeding, G0S2 mRNA is detected only in low levels in the liver and at high levels in WAT where it promotes the storage of lipids. Upon fasting, G0S2 concentrations in WAT decrease, promoting lipolysis and WAT-derived FA transport to the liver and other organs. Simultaneously, the increase in liver G0S2 expression inhibits the degradation of endogenous TG stores and thus can promote fasting-induced liver steatosis (26). Consequently, tissue-specific regulation of G0S2 provides a mechanism controlling TG storage in different organs depending on the metabolic state. Notably, G0S2 is reduced in WAT of diabetic individuals (30) suggesting that changes in G0S2 expression contribute to dysregulated lipolysis observed in diabetic patients. Other physiological processes have also been associated with G0S2 and include adipogenesis (24), proliferation (31,32), apoptosis (33), immune regulation (34 -37), oxidative phosphorylation (38), tumor suppression, and cancer (33, 39 -41), but they have not been studied extensively. Our goal is to identify peptide inhibitors for ATGL based on the physiological inhibition by G0S2. To do so, we first have to understand the underlying mechanisms of ATGL inhibition by G0S2. Previous reports suggest that a truncated variant of G0S2 encompassing the residues Met-1-Gln-73 interacts with ATGL and inhibits the enzyme. Another variant lacking an extensive central region (residues Tyr-27-Leu-42) failed to immunoprecipitate with ATGL and had no inhibitory activity (20). These results already indicated that the N-terminal portion of G0S2 is essential for ATGL inhibition. However, deletion of the central region of G0S2 might have also resulted in loss-of-function due to complete disruption of the protein fold. Consequently, we first delineated the minimal sequence boundaries for a biologically active, truncated G0S2 variant. Furthermore, we describe a synthetic inhibitory peptide containing a sequence stretch derived from G0S2, and we kinetically characterize the mode of inhibition by this peptide. Our study provides first insights into the mode of inhibition and opens avenues to achieve tissue-specific inhibition of ATGL based on a synthetic peptide.
Cloning of Recombinant Proteins-Human G0S2 (hG0S2), including the complete open reading frame, was amplified from cDNA by PCR using Phusion TM polymerase (New England Biolabs, Ipswich, MA) and primers containing endonuclease cleavage sites for subsequent cloning into a modified pSUMO vector (kindly provided by Prof. Christoph D. Lima, Sloan-Kettering Institute) with a tobacco etch virus (TEV) protease cleavage site for tag removal (his 6 -smt-hG0S2). N-terminal truncations were generated by PCR using primers flanking the respective sequence of hG0S2 and containing endonuclease cleavage sites for insertion into the target vector. C-terminal truncations of hG0S2 were obtained by introducing stop codons using the QuikChange site-directed mutagenesis kit (Agilent Technol- CCTCTTCGGCGTGGTGTAAGGCCTGATGGAGACTGTGTGC GCACACAGTCTCCATCAGGCCTTACACCACGCCGAAGAGG ogies, Santa Clara, CA). Primers used are listed in Table 1. Sequences containing the coding sequences of mouse Cgi-58 and mouse Atgl were inserted in pSUMO (see above) and pASK-IBA5plus (IBA, Goettingen, Germany) vectors, respectively, as described previously (13,42). pcDNA4/HisMax vectors (Invitrogen) encoding mouse Atgl, patatin domain-containing proteins 6 and 7 (mPnpla6 and mPnpla7), hormonesensitive lipase (Hsl), and monoacylglycerol lipase (Mgl) coding sequences were generated as described previously (7,43,44). 6 -smt-G0S2 constructs were transformed into Escherichia coli BL21(DE3) CodonPlus cells (Stratagene, La Jolla, CA). Cultures were grown at 37°C on selective LB medium containing 40 g/ml kanamycin to an A 600 of 0.5. Expression was induced by the addition of 0.5 mM isopropyl-␤-D-thiogalactopyranoside at 30°C. After 15 h of induction, cells were harvested, resuspended in sucrose solution (250 mM sucrose, 1 mM EDTA, 1 mM DTT, 20 g/ml leupeptin, 2 g/ml antipain, 1 g/ml pepstatin, pH 7.0), and disrupted by sonication (SONOPLUS ultrasonic homogenizer HD 2070, Bandelin, Berlin, Germany) on ice. After centrifugation at 15,000 ϫ g for 20 min at 4°C, the supernatants were collected. Protein concentrations were determined as described below. Expression of the murine ATGL-Strep fusion (Strep-mATGL) and His 6 -Smt-mCGI-58 in E. coli is described in Refs. 13 and 42, respectively.
Purification of Recombinant Human G0S2-E. coli cells containing the His 6 -Smt-tagged hG0S2 (His 6 -Smt-hG0S2) were resuspended in buffer A (adjusted to pH 7.5) and disrupted by sonication (SONOPLUS ultrasonic homogenizer HD 2070) on ice. After centrifugation at 15,000 ϫ g for 20 min at 4°C, His 6 -Smt-hG0S2 was purified from the cellular extract by affinity chromatography using a 5-ml HisTrap TM FF column (GE Healthcare). Purified recombinant protein was eluted in 10 column volumes of a 0 -100% buffer B (adjusted to pH 7.5) gradient. Next, the His 6 -Smt tag was cleaved off by the addition of 1 mM EDTA and TEV protease in a ratio of 1:100 (protein/TEV protease). The cleavage was performed overnight at 4°C. Cleaved hG0S2 was isolated by size exclusion chromatography using a Superdex 200 column (GE Healthcare) and 15 mM Na 2 HPO 4 , 5 mM KH 2 PO 4 , 300 mM NaCl, 1 mM EDTA, 1 mM TCEP, pH 7.0, as mobile phase. Pooled fractions containing cleaved hG0S2 were dialyzed (3,500 Da cutoff) extensively against MilliQ (Millipore Corp., Billerica, MA) water at 4°C before lyophilization. The lyophilized powder was dissolved in 100% DMSO, and the concentration was determined with Pierce BCA protein assay kit as described below.
Preparation of Tissue Homogenates-Adipose tissue samples from overnight fasted wild-type and Atgl-KO mice were washed in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 ) containing 1 mM EDTA and homogenized on ice in sucrose solution (see above) using an Ultra Turrax (IKA, Staufen, Germany). After centrifugation at 10,000 ϫ g for 20 min at 4°C, the lipid-free infranatant was collected, and the protein content was determined as described below.
Preparation of Synthetic Peptide-Synthetic peptides were purchased from Peptide Specialty Laboratories GmbH, Heidelberg, Germany. The sequences of the inhibitory peptide hGW2052 and the control peptide were WKGKMVKLYVLGS-VLALFGVVLGLMETVCSPFTA and VDSADAGGGSGWLT-GWLPTWCP, respectively. The peptides were dissolved in 100% DMSO, and concentrations were determined via the absorbance at 280 nm. The final DMSO concentration in the activity assay was Յ1%. The control peptide was prepared the same way.
TG Hydrolase Activity Assay-TG hydrolase activity assay was performed as described (45). To screen for the inhibitory capacity of various hG0S2 truncations, 30 g of protein of cell lysates were incubated with the [ 3 H]triolein substrate together with 50 g of cell extract of Strep-mATGL (or COS-7-mATGL) and 2.5 g of purified mCGI-58. As a control, lysates of the His 6 -Smt tag were incubated under the same conditions. To determine the enzyme kinetics and the inhibitory mechanism of the hGW2052 peptide, the assay was performed with minor modifications. TG substrate was prepared with 1.67 mM triolein, 10 Ci/ml [9,10-3 H]triolein (PerkinElmer Life Sciences), and 190 M phosphatidylcholine/phosphatidylinositol (3:1) (Sigma) and diluted to the respective concentrations after sonication. Assays with CGI-58 co-activated ATGL were performed on a smaller scale. Briefly, cell lysates of Strep-mATGL (25 g of total protein content) were mixed with 2.5 g of purified mCGI-58 and 1% of the synthetic peptide hGW2052, con- Monoacylglycerol Hydrolase Activity Assay-Monoacylglycerol hydrolase activity of murine MGL was determined as described (46). In brief, 1% of hGW2052, the control peptide, or DMSO was added to 2 g of lysate of COS-7 cells overexpressing murine MGL in a total volume of 10 l. The mixture was incubated with 100 l of rac-1-(3)-oleoylglycerol (Sigma) as substrate for 20 min at 37°C. After incubation, the reaction was terminated by adding 100 l of chloroform. After vigorous mix-ing using a vortex and centrifugation at 10,000 ϫ g for 5 min, the glycerol release (into the upper aqueous phase) was determined using a commercial kit (free glycerol Reagent, Sigma) according to the manufacturer's protocol.
Lysophospholipase Activity Assay-Lysophospholipase activity of PNPLA6 and PNPLA7 was determined as described previously (44). In brief, 1% of the hGW2052 peptide, the control peptide, or DMSO was added to cell lysates of COS-7 cells overexpressing PNPLA6 (15 g), PNPLA7 (50 g), and lacZ in a total volume of 50 l of sucrose solution (see above). The mixtures were incubated with 50 l of 3 mM substrate 1-oleoyl-2hydroxy-sn-glycero-3-phosphocholine (Avanti Polar Lipids Inc., Alabaster, AL) in a water bath at 37°C for 20 min. Each reaction was terminated by heat inactivation at 75°C for 10 min. The released amounts of fatty acids were determined using a commercial assay kit (HR Series NEFA-HR(2), WAKO Diagnostics) according to the manufacturer's protocol.
Determination of Protein Concentration-Protein concentrations of cell extracts were determined using the protein assay Kit according to the manufacturer's instructions (Bio-Rad) using BSA as standard. The concentrations of purified proteins  NOVEMBER 21, 2014 • VOLUME 289 • NUMBER 47

JOURNAL OF BIOLOGICAL CHEMISTRY 32563
were obtained by the absorbance at 280 nm using the Nano-Drop ND-1000 spectrophotometer (PEQLAB Biotechnologie GmbH, Erlangen, Germany). Alternatively, protein concentrations were determined using the Pierce BCA protein assay kit and BSA as standard according to the manufacturer's protocol (Thermo Scientific).
Statistical Analysis-All measurements were performed in triplicates. Data are presented as means Ϯ S.D. Statistical significance was determined by the Student's unpaired t test (twotailed). Group differences were considered statistically significant for p Ͻ 0.05 (*), p Ͻ 0.01 (**), and p Ͻ 0.001 (***).

RESULTS
The N-terminal Region of G0S2 Is Essential for ATGL Inhibition-To identify the minimal sequence within G0S2 that is strictly required to exert an inhibitory effect on ATGL activity, we tested different C-and N-terminally shortened versions of human G0S2 (hG0S2) (Fig. 1A). As ATGL exhibits very low basal activity, we performed activity assays in the presence of CGI-58 to fully stimulate ATGL. Our results show that C-terminal truncations of G0S2 up to residue Val-46 completely inhibit CGI-58 activated ATGL activity (abbreviated as ATGL* in all figures) in COS-7 lysates (Fig. 1B). Inhibition was significantly reduced when hG0S2 was truncated at residue Leu-40 (Fig. 1B). These results were reproduced using proteins (ATGL, CGI-58, and G0S2) expressed in E. coli (Fig. 1C), suggesting that other mammalian proteins are not required for G0S2-mediated inhibition. It also demonstrates that mouse ATGL is also inhibited by human G0S2. Next, we investigated the amino acid residues between Leu-40 and Val-46 and their necessity for ATGL inhibition by G0S2 in more detail. We show that the G0S2 variant Met-1-Met-43 (hG0S2_1-43) harbors the longest C-terminal deletion that still exerts the full inhibitory effect on ATGL (Fig. 1C). Screening various N-terminal deletions revealed that truncations of hG0S2 until residue Tyr-27 (hG0S2_27-103) fully inhibit ATGL activity (Fig. 1D). Further truncation of the N terminus resulted in significant loss of the inhibitory effect. Truncation of G0S2 after residue Gly-30 resulted in complete loss of the capacity to inhibit ATGL. Next, we tested the effect of a combination of N-and C-terminal truncations of hG0S2. As expected, hG0S2 Leu-10 -Ala-52 and Lys-20 -Ala-52 fully retained their activity to inhibit ATGL, although hG0S2 Gly-30 -Ala-52 had lost all inhibitory activity. Interestingly, neither hG0S2 Tyr-27-Ala-52 nor hG0S2 Lys-20 -Met-43 exhibited full inhibitory capacity (Fig. 1E), as extrapolation of the findings presented above would have suggested (Fig. 1, C and D).
The Peptide Corresponding to Residues Lys-20 to Ala-52 (hGW2052) from G0S2 Inhibits ATGL in the Nanomolar Range-In our next experiments, we investigated whether synthetic peptides with sequence stretches corresponding to G0S2 are efficient in inhibiting ATGL, and we characterized the peptide-mediated inhibition kinetic of ATGL. Therefore, we designed a synthetic peptide identical to the human G0S2 sequence from Lys-20 to Ala-52 (Fig. 1A). A Trp residue was introduced at the N terminus of this peptide to facilitate concentration determination by UV spectroscopy. The resulting peptide was termed hGW2052. In agreement with the overexpressed G0S2 variant hG0S2_20-52 (Fig. 1), the synthetic peptide hGW2052 completely inhibited CGI-58-activated hATGL and CGI-58-activated mATGL expressed in COS-7 and E. coli cells, respectively (Fig. 2, A and B). This indicates that the synthesized peptide adopts a tertiary structure compatible for protein-protein interaction with ATGL. A control peptide was equally treated and did not show an inhibitory effect on ATGL (Fig. 2, A and B). Next, we tested whether hGW2052 inhibits ATGL also in the absence of CGI-58. Indeed, we found that hGW2052 inhibits ATGL independently of the presence of CGI-58 (Fig. 2C). The dose-response experiments of basal ATGLand CGI-58-stimulated ATGL activity using increasing amounts of hGW2052 revealed half-maximal inhibitory concentrations (IC 50 ) of 10 and 18 nM, respectively (Fig. 2, D and E). In agreement with these results, dose-response experiments using overexpressed, full-length hG0S2 revealed an IC 50 of 19 nM (Fig. 2F).
The Peptide hGW2052 Inhibits ATGL in a Noncompetitive Manner-To investigate the mode of ATGL inhibition by hGW2052, we performed enzyme kinetics. In agreement with previously published data, time course experiments showed a linear increase in FA release of ATGL-containing lysates incubated with an artificial TG substrate for up to 60 min (Fig. 3A) (13). Furthermore, a linear increase in apparent enzyme velocity up to 1,000 M triolein substrate was observed in saturation kinetic experiments in the absence of an inhibitor (Fig. 3B). Within that substrate concentration range, we also observed a linear increase in enzyme velocity in the presence of different inhibitor concentrations (Fig. 3, C and D). Dixon-Plot analysis (47) (plotting 1/v against inhibitor concentration at each concentration of triolein substrate) resulted in a set of intersecting lines on the x axis compatible with a noncompetitive inhibition mechanism (Fig. 3E). Using this method, we determined an apparent K i of 28 nM for hGW2052. Nonlinear regression analysis (Graph Pad Prism) using a noncompetitive inhibition model was also used to extract kinetic parameters and resulted in an apparent K i of 25 Ϯ 1 nM, which is in excellent agreement with the result from the Dixon-Plot method (Fig. 3E, insert).
The Peptide hGW2052 Is a Specific Inhibitor for ATGL-Next, we investigated the specificity of the peptide hGW2052. ATGL and HSL together account for more than 95% of intracellular TG hydrolase activity in white adipose tissue (WAT) (48). For both enzymes, small molecule inhibitors are available. Atglistatin and Hi 76 -0079 (N-methyl-phenyl carbamoyl tri- azole, Novo Nordisk, Denmark) act as specific inhibitors for ATGL and HSL, respectively (10,48,49). WAT lysates of WT and Atgl-KO mice were tested for TG hydrolase activity in the presence of Atglistatin, Hi 76 -0079, and hGW2052. Addition of the ATGL inhibitors hGW2052 or Atglistatin reduced TG hydrolase activity of WAT lysates from WT animals by 45 and 49%, respectively. Addition of the HSL inhibitor Hi 76-0079 reduced TG hydrolase activity of WAT lysates by 67%. Combined addition of hGW2052 or Atglistatin (representing ATGL inhibitors) and Hi 76-0079 (representing an HSL inhibitor) reduced TG hydrolase activity of WAT by 89 and 88%, respectively (Fig. 4A). Neither hGW2052 nor Atglistatin inhibited WAT TG hydrolase activity derived from Atgl-KO mice (Fig.  4A). These results indicate that hGW2052 and Atglistatin specifically inhibit ATGL but not HSL activity.
As ATGL is a member of the patatin-like phospholipase domain containing protein (PNPLA) family (50), we further tested whether the synthetic peptide inhibits other proteins of the PNPLA family. As shown in Fig. 4, B and C, hGW2052 did not exert any inhibitory effect on the lysophospholipase activities of PLNPLA6 or PNPLA7, up to a concentration of 1.4 M. As expected, the control peptide also did not affect the lysophospholipase activity of PNPLA6 and PNPLA7 (Fig. 4, B and C). Thus, we conclude that hGW2052 does not act as general inhibitor for members of the PNPLA protein family. We also tested for a potential effect of the inhibitory peptide hGW2052 on other extracellular (lipoprotein lipase) and intracellular (MGL and HSL) lipases. No inhibitory effect on any of these lipases was observed upon addition of up to 1.4 M hGW2052 or control peptide, respectively (Fig. 4, D-F). Together, these experiments demonstrate that the peptide hGW2052 is not a general lipase inhibitor and, among the different lipases tested, rather acts specifically on ATGL.

DISCUSSION
Abnormally high concentrations of circulating FAs have deleterious effects on the organism, including ectopic lipid accumulation, insulin resistance, inflammation, cellular dysfunction, and cell death (1,2). ATGL catalyzes the hydrolysis of TG into diacylglycerols and FA, the first step in TG catabolism, and thus strongly affects FA concentrations in the circulation. Decreased ATGL activity exerts beneficial effects such as improved glucose tolerance and insulin sensitivity and protection from cancer-induced cachexia (8,11,51,52). Therefore, it is of fundamental interest to understand the mechanisms that regulate ATGL-mediated lipolysis and concomitantly the release of potentially harmful FAs.
G0S2 has been identified to act as a physiological inhibitor of ATGL. However, very little is known about the molecular basis and the mode of ATGL inhibition by G0S2. Therefore, these questions were addressed in this study. We also identified minimal sequences of G0S2 that inhibit ATGL's TG hydrolase activity. Variants of G0S2 that were fully able to inhibit ATGL included extensive N-or C-terminal truncations of the wildtype protein (Tyr-27-Ser-103 and Met-1-Met-43, respectively). Interestingly, however, a combination of the identified N-and C-terminal boundaries (e.g. variants Lys-20 -Met-43 or Tyr-27-Ala-52) did not result in G0S2 variants that were able to inhibit ATGL efficiently. We hypothesize that a core peptide Tyr-27-Met-43 is mainly involved in ATGL inhibition, yet a few additional residues are required to mediate full function of the inhibitory peptide.
Based on our results, we furthermore developed a short 34-amino acid peptide that acts as an efficient inhibitor of mouse and human ATGL. Inhibition kinetics performed in this study showed that the inhibitory peptide hGW2052 acts in a noncompetitive manner. Long-chain acyl-coenzyme A and Atglistatin are reported to inhibit ATGL upon interaction with the lipase (10,13). Although noncompetitive inhibition is also reported for acyl-CoA, Atglistatin inhibits ATGL in a competitive manner (10,13). Dose-response experiments revealed that wild-type G0S2 and the inhibitory peptide hGW2052 act in the nanomolar range (IC 50 ϭ 10 and 18 nM, respectively). Consequently, physiologically relevant ATGL regulation by G0S2 is very plausible, and administration of hGW2052 at physiological concentrations can be envisaged. This high affinity of G0S2 is in contrast to the affinity of oleoyl-CoA for ATGL, which exhibits a much higher IC 50 value (33 M) (13). Accordingly, the potential physiological relevance of an acyl-CoA-mediated feedback mechanism still remains to be elucidated. The small molecule Atglistatin exhibits an IC 50 value in the low micromolar range, which further illustrates the potency of the inhibitory peptide.
To evaluate the specificity of the inhibitory peptide for ATGL, we analyzed the effect of hGW2052 on the activities of phospholipases, intracellular lipases, and extracellular lipases. ATGL belongs to the patatin domain containing family of proteins (PNPLA). Some family members have already been characterized and exhibit TG hydrolase activity (ATGL, also termed PNPLA2) or lysophospholipase activities (PNPLA6 and PNPLA7) (50). Although PNPLA family members show high sequence conservation within their common patatin domain, no effects on lysophospholipase activities of PNPLA6 or PNPLA7 could be observed upon incubation with the peptide hGW2052. Furthermore, our enzyme assays revealed that hGW2052 also does not inhibit HSL, MGL, or lipoprotein lipase activities and thus demonstrate high specificity for ATGL. Similar discrimination between general lipases, PNPLAs and ATGL was observed for Atglistatin, which was specifically developed and optimized to inhibit mouse ATGL (10). Long-chain acyl-CoAs, however, exhibit less specificity for lipase inhibition and have been reported to inhibit both ATGL and HSL (13,53). It should be noted that the physiological roles of long-chain acyl-CoAs are diverse and include acyl-CoA-dependent activities of some PNPLA protein family members (54). Acyl-CoAs are highly abundant metabolic intermediates involved in multiple processes and thus clearly do not represent promising lead compounds for the development of ATGL inhibitors.
Genetic disruption of Atgl results in severe heart defects due to massive TG accumulation in cardiac muscle (8,55). Thus, the application of a synthetic, competitive ATGL inhibitor, such as Atglistatin, as a therapeutic tool appears limited. Yet the single administration of Atglistatin to mice led to TG accumulation in liver but not the heart, accompanied by accumulation of the inhibitor in liver and adipose tissues (10). This demonstrates that a transient inhibition of ATGL is an effective means to inhibit TG catabolism and does not necessarily interfere with cardiac TG homeostasis. The peptide developed in this study harbors a high potential for selective inhibition of ATGL. Peptide inhibitors offer the great advantage that they can be fused with specific targeting sequences and as such allowing tissuespecific targeting (56 -58), which could prevent harmful TG accumulation in the heart or other tissues. Competitive versus noncompetitive inhibition modes are additional important differences when comparing the potential impact of the peptide inhibitor hGW2052 with Atglistatin. hGW2052 inhibits ATGL in a noncompetitive manner and therefore inhibits ATGL despite high physiological substrate concentrations. Thus, hGW2052 represents a powerful research tool for tissue-selective inhibition of ATGL and has a high potential for translation into therapeutic application.
Structural knowledge on ATGL is very limited. It is predicted to exert lipolytic activity via a catalytic dyad that resides in the patatin domain within the N-terminal half of the lipase (7,50). Previous reports demonstrated that G0S2 interacts with the N-terminal catalytic domain of ATGL (15). Here, we demonstrate that a G0S2-derived peptide does not compete with the TG substrate for ATGL binding. Similarly, inhibiting the truncated ATGL-289X variant by acyl-CoA showed that acyl-CoA binds to the N-terminal half of ATGL and is independent of substrate binding (13). In contrast, Atglistatin inhibits in a competitive mode indicating that the molecule probably binds within the substrate-binding pocket of ATGL. In this study, we further demonstrate that the inhibition of ATGL by the peptide hGW2052 works in the absence or presence of CGI-58. Similarly, ATGL inhibition by acyl-CoA and Atglistatin was reported to be independent of the presence of CGI-58 (10,13). In brief, the currently available data of the ATGL interaction partners G0S2 and the inhibitory peptide, CGI-58, acyl-CoA, and Atglistatin point toward independent interaction regions, yet all are reported to reside within the N-terminal half of ATGL (15,18,20,59). Unfortunately, no experimental threedimensional structures for any of these players are known, which would aid in the identification of interaction surfaces. Furthermore, no mutations of ATGL are known that retain catalytic activity, but they lose the ability to interact with CGI-58 or G0S2. Clearly, more biochemical and structural knowledge is required to address these open questions.