Protein lipidation in cell signaling and diseases: function, regulation and therapeutic opportunities

Summary

Protein lipidation is an important co- or post-translational modification in which lipid moieties are covalently attached to proteins. Lipidation markedly increases the hydrophobicity of proteins, resulting in changes to their conformation, stability, membrane association, localization, trafficking, and binding affinity to their co-factors. Various lipids and lipid metabolites serve as protein lipidation moieties. The intracellular concentrations of these lipids and their derivatives are tightly regulated by cellular metabolism. Therefore, protein lipidation links the output of cellular metabolism to the regulation of protein function. Importantly, deregulation of protein lipidation has been linked to various diseases, including neurological disorders, metabolic diseases and cancers. In this review, we highlight recent progress in our understanding of protein lipidation, in particular, S-palmitoylation and lysine fatty acylation, and we describe the importance of these modifications for protein regulation, cell signaling and diseases. We further highlight opportunities and new strategies for targeting protein lipidation for therapeutic applications.

Introduction

Protein lipidation is a unique co-translational or posttranslational modification that plays a critical role in cell signaling, and dynamically regulates protein functions in response to extrinsic and intrinsic cues. Lipidation modulates the function of targeted proteins by increasing their binding affinity to biological membranes, rapidly switching their subcellular localizations, affecting folding and stability, and modulating association with other proteins. Proteins can be covalently modified by at least six types of lipids, including fatty acids, isoprenoids, sterols, phospholipids, glycosylphosphatidylinositol (GPI) anchors and lipid-derived electrophiles (LDEs) (Figure 1). Fatty acids, such as 16-carbon palmitate, are building blocks for the biosynthesis of other complex lipid molecules in cells, and their intracellular concentrations are tightly regulated. Palmitate is biosynthesized by condensation reactions between acetyl-CoA and malonyl-CoA catalyzed by the fatty acid synthase (FASN) complex. Acetyl-CoA is generated from pyruvate and citrate, through pyruvate dehydrogenase and ATP citrate lyase, respectively. Malonyl-CoA is synthesized by acetyl-CoA carboxylase (ACC). Both acetyl-CoA and malonyl-CoA are byproducts of glucose metabolism and Krebs cycle (Figure 1). ACC is directly regulated by cellular metabolic regulator and energy sensor, AMP-activated protein kinase (AMPK). Therefore, lipid biosynthesis is directly involved in cellular energy homeostasis. Deregulation of lipid metabolism plays a prominent role in human disease. However, the detailed mechanism linking deregulation of lipid metabolism to cell signaling and disease remains elusive. It is possible that alterations of lipid metabolism could affect the availability of lipid donors, thus affecting the global protein lipidation levels. Protein lipidation may partially contribute to the pathological consequences of the misregulated lipid metabolism, such as in cancers and metabolic diseases. It is important to understand the detailed regulatory mechanisms, the functions and pathological relevance of protein lipidation, which will ultimately lead to new therapeutics. In the following sections, we will first summarize the diversity of protein lipidation, and then focus our discussions primarily on protein fatty acylation.

Figure 1. Protein lipidation links lipid metabolism to the regulation of protein functions.

Proteins can be modified by at least 6 types of lipid, including saturated and unsaturated fatty acids (palmitate, myristate, and palmitoleate etc.), isoprenoids (farnesyl and geranylgeranyl), GPI anchors, cholesterols, phospholipids (not shown here) and lipid-derived electrophiles (HNE etc.). Cellular lipid metabolism affects the availability of fatty acyl-CoA and other lipid derivatives, which are used as substrates for protein lipidation. The 16-carbon fatty acid, palmitate, is a critical intermediate for the biosynthesis of other lipids in the cells.

(FFA: free fatty acid; HNE: 4-hydroxynonenal; GPI: glycosylphosphatidylinositol; FPP: farnesyl pyrophosphate; GGPP: geranylgeranyl pyrophosphate; TCA cycle: tricarboxylic acid cycle or Krebs cycle).

Diversity of protein lipidation

Saturated and unsaturated fatty acids can attach to the cysteine, serine, or lysine residues of proteins, in a process known as fatty acylation. The 14-carbon myristate can be attached to the N-terminal glycine, catalyzed by N-myristoyl transferases (NMTs) as a stable and irreversible co-translational modification. Recent chemical proteomic studies suggest that more than 100 proteins are myristoylated in human cells (Thinon et al., 2014). Typically, myristoylation enhances protein-membrane association that is required for target protein proper localization and biological function. S-palmitoylation (or S-acylation) is another major form of fatty acylation, in which the 16-carbon palmitate (or other fatty acids, for example, the 18-carbon stearic acid) are attached to cysteine residues. Due to the labile nature of the thioester bond, S-palmitoylation is reversible, and S-palmitoylated proteins can undergo cycles of acylation and deacylation in response to upstream signals (Rocks et al., 2010). In addition, fatty-acyl groups can be attached to serine residues (O-palmitoylation) or the N-terminus (so called N-palmitoylation). To date, only a few proteins are known to be O- or N-palmitoylated. For example, Wnt proteins are modified on a conserved serine residue by a monounsaturated fatty acid, cis-palmitoleic acid (C16:1Δ9), mediated by a membrane-bound O-acyltransferase called Porcupine, (Clevers and Nusse, 2012). Histone H4 can be palmitoylated at Ser45 by lysophosphatidylcholine acyltransferase 1 (LPCAT1) (Zou et al., 2011). The 8-carbon octanoyl group can modify the serine residue of the signaling peptide ghrelin by ghrelin O-acyltransferase (GOAT) (Gutierrez et al., 2008). Hedgehog (Hh) proteins are examples of N-palmitoylated proteins, where the N-terminus of the signaling peptide is palmitoylated, catalyzed by Hedgehog acyltransferase (HHAT) (Buglino and Resh, 2008).

Modification of proteins by isoprenoids is known as prenylation. Most prenylated proteins contain a CAAX motif at their carboxyl terminus, in which the consensus Cys residue is modified by farnesyl (15-carbon) or geranylgeranyl (20-carbon) groups by farnesyltransferase (FTase) or geranylgeranyltransferase I (GGTase I), respectively. The Ras-family of proteins, including K-, H-, and N-Ras, are well-studied prenylated proteins (Wang and Casey, 2016). Prenylation is generally considered as an irreversible modification, which alters the membrane affinity of proteins.

Covalent modification of proteins by cholesterols is uncommon. To date, the only well characterized cholesterol-modified proteins are Hedgehog proteins with which cholesterol forms an ester bond at the C-terminal glycine of the cleaved signaling peptide. Recently, chemical probes have been developed to identify other cholesterol-modified proteins (Heal et al., 2011). Smoothened (SMO), the co-receptor of Hh pathway, is cholesterol modified on the Asp95 (Xiao et al., 2017).

Phospholipids play critical roles in membrane formation and signal transduction. However, protein phospholipid modification is rare. The only known example is the autophagy related protein Atg8/LC3, which is modified by phosphatidylethanolamine (PE). This modification is mediated by multi-step conjugation processes, and is essential for the double membrane formation of the autophagosome (Nakatogawa et al., 2007).

GPI anchor is a complex glycolipid that can be covalently attached to the C-terminus of proteins as a posttranslational modification. GPIs are synthesized in the endoplasmic reticulum (ER) by sequentially adding monosaccharides, acyl groups, and phosphoethanolamine residues to phosphatidylinositol. The attachment of GPI to the protein mediated by GPI transamidase complex (Yu et al., 2013). About 1% of the eukaryotic proteins are modified by GPI anchors (Orlean and Menon, 2007). In addition to its role in targeting protein to the membrane, GPI anchor has many other biological functions, such as mediating immune response and inflammation, depending on the GPI structural composition (Tsai et al., 2012).

Lipid-derived electrophiles (LDEs) are reactive lipid metabolites produced by lipid peroxidation or other metabolic pathways via non-enzymatic and enzymatic mechanisms. Known LDEs include 4-hydroxy-2-nonenal (HNE), 15-deoxy-Δ12, 14-prostaglandin J2 (15d-PGJ2), and 2-trans-hexadecenal (2-HD) etc. LDEs are able to form covalent adducts with nucleophilic residues of proteins, such as cysteine, lysine and histidine via Michael addition (Chen et al., 2016b). Misregulation of lipid metabolism often leads to the accumulation of LDEs, which are involved in various pathological conditions, such as inflammation, genotoxicity, and tissue degeneration.

Therefore, protein lipidation can regulate many key biological functions of proteins through diverse types of lipids and lipid metabolites. However, it has been challenging to investigate the specific function and regulation of protein lipidation under physiological and pathological conditions due to limited tools and methods. Several recent reviews have discussed the functions of protein prenylation, myristoylation, GPI modifications (Jiang et al., 2018; Wang and Casey, 2016; Zurzolo and Simons, 2016). Due to the limitation of space, rather than the importance, here we will briefly discuss different methods for studying protein lipidation, and then mainly focus on the function, regulation and potential therapeutic targets of protein fatty acylation, especially S-palmitoylation of proteins, for which exciting new progresses to understand its functions have been made in recent years..

Chemical and biochemical methods to study protein lipidation

Traditionally, radioactive isotope-labeled lipids were used to analyze protein lipidation. Although still in practice, such methods are generally not preferred, due to sensitivity and safety issues. Although a recombinant antibody that specifically recognizes the conformation of palmitoylated PSD-95 was reported (Fukata et al., 2013), development of specific antibodies recognizing other lipid modifications has not been successful. S-fatty acylation is particularly difficult for direct detection by chromatography or mass spectrometry methods, due to the high hydrophobicity of the lipid moiety and the unstable thioester linkage. For these reasons, several alternative methods have been developed to detect the modified cysteines or the tagged lipid moieties. Below we emphasize recent advances of acyl exchange methods and bioorthogonal chemical reporters.

Acyl-biotin exchange (ABE) was one of the first methods developed to detect S-acylation of cysteines, which converts the acyl modification to a stable biotin adduct. It was based on the high reactivity of the thioester bond, which can be readily removed by weak bases, such as hydroxylamine (Figure 2A). ABE has been widely used in proteomic and signaling studies. As S-acylation does not alter the protein mass significantly and electric charge remains the same, S-acylated species usually do not migrate differently from the non-acylated proteins on protein gel electrophoresis, making it difficult to directly detect the levels and the stoichiometry of acylation of a protein of interest. Acyl-PEG exchange (APE) or acyl-PEGyl exchange gel shift (APEGS) is a mass-tag labeling method to evaluate the levels of endogenous fatty acylated proteins (Percher et al., 2016). Similar to ABE, upon liberating the acylated cysteines by hydroxylamine, the resulting free thiol is capped with PEG-N-ethylmaleimide (PEG-NEM) to introduce a mass shift with respect to the non-acylated protein (Figure 2A). It allows easy detection of the mass shift by western blot, and does not require affinity enrichment. In addition, one can easily quantify the ratio of unmodified versus S-acylated proteins, or multiple sites of S-acylation. It has been shown that there are three S-acylated cysteine residues of interferon-induced transmembrane protein 3 (IFITM3) regulating its clustering in membrane compartments, which are critical for IFITM3 antiviral activity (Percher et al., 2016).

Figure 2. Chemical and biochemical methods to detect protein lipidation.

(A). Detection of S-acylated proteins by acyl exchange methods. (B). Detection of protein lipidation using bioorthogonal chemical reporters.

Although palmitoylation is often the most common type of S-acylation, other saturated and unsaturated acyl groups could be used. Since the original acyl groups are removed during the “exchange” process, it would be impossible to distinguish the different acyl groups originally attached to the proteins using ABE or APE alone. Therefore, to analyze the preference of lipid chains, lipid-based methods have to be used, such as bioorthogonal chemical reporters, which are usually terminal alkyne or azido derivatives of lipids (fatty acids, sterols or isoprenoids). Exogenous chemical reporters can be taken up, converted to acyl-CoAs or other intermediates and used as lipid donors in cells. Therefore, live cells are required for metabolic labeling of lipidation of proteins (Figure 2B). The chemical reporters can be detected by using Copper-catalyzed Azide-Alkyne Cycloaddition (CuAAC) (Click reaction) with biotin or fluorescent tags. ω-Alkynyl fatty acid probes are generally more efficient with higher sensitivity for detection compared to azido-fatty acids (Hannoush and Sun, 2010). ω-Alkynyl fatty acids with various chain lengths can be used to evaluate different type of lipidation, such as myristoylation, palmitoylation, stearoylation, and prenylation (Hannoush and Sun, 2010). In addition, chemical reporters also allow the detection of monounsaturated fatty acylation, prenylation and N- or O-acylation, which could not be detected by acyl exchange methods (Hang and Linder, 2011). Furthermore, the reported click chemistry-proximity ligation imaging method can be used to detect subcellular localization of fatty-acylated proteins (Gao and Hannoush, 2014). As adding the alkynyl group extended the fatty chain for 2 extra carbons, one has to be careful when picking the appropriate reporters. We recommend using the total carbon chain, including the 2 carbons in the alkynyl group, to represent the acyl chain length. In addition, since the metabolic labeling requires high concentrations of reporters (50 to 100μM), it would be difficult to evaluate the lipidation levels under physiological concentrations of endogenous acyl donors. Combination of the metabolic labeling of chemical reporters, as well as the acyl-exchange methods enables clearer analysis of protein lipidation.

S-palmitoylation as a reversible regulation in cell signaling

Palmitoyl-CoA is one of the most abundant fatty acyl-CoAs in cells, and a common intermediate for the biosynthesis of other complex lipid molecules. Proteomic analyses have identified over 1000 S-palmitoylated proteins. The palmitoylated proteome could be even bigger, and the bioinformatics tool SwissPalm has predicted over 5000 possible S-palmitoylated proteins (Blanc et al., 2015). Therefore, understanding the functions of S-palmitoylation is critical for cell signaling and therapeutics. It has been widely studied that S-palmitoylation plays an important role in regulating synaptic transmission. At the presynaptic side, many S-palmitoylated proteins, including GAD65, SNAP25 and CSP, are involved in neurotransmitter synthesis and release (Fukata and Fukata, 2010). Heterotrimeric G proteins, consisting of α, β and γ subunits, are signal-transducers of GPCRs. A recent study showed that S-palmitoylation of Gαi1 regulates its interaction with lipid rafts and affects its membrane microdomain localization (Alvarez et al., 2015).

S-palmitoylation is one of the best-studied lipid modifications, which impact diverse biological processes. S-palmitoylation can be dynamically regulated by zinc-finger DHHC (Asp-His-His-Cys)-containing palmitoyl acyltransferases (ZDHHC family of PATs) and acyl protein thioesterases (APT1 and APT2). In addition, some proteins can bind to palmitoyl-CoA directly and undergo PAT-independent autopalmitoylation. Below, we will focus on recent progresses of studying the novel functions and regulations of S-palmitoylation.

S-palmitoylation in regulation of immune receptor functions

S-palmitoylation of immune receptors plays an important role in the innate immune response (Chesarino et al., 2014; Mukai et al., 2016), linking lipid metabolism to host defense and immunity. The cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) pathway is critical in intracellular signaling of the innate immunity. STING is S-palmitoylated, and this modification promotes its clustering at the trans-Golgi membrane, where it recruits and activates TBK1 to phosphorylate IRF3 and induce the type I interferon response (Mukai et al., 2016). Pharmacological inhibition of STING S-palmitoylation, or mutation of its S-palmitoylation sites, blocks the type I interferon response. In addition, Golgi-localized ZDHHC proteins (ZDHHC3, 7 and 15) are potential candidate enzymes responsible for STING S-palmitoylation. Interestingly, ER-localized ZDHHC1 is also required for STING activation and recruitment of downstream effectors TBK1 and IRF3 upon DNA-virus infection (Zhou et al., 2014), implying the novel function of ZDHHC proteins in immune signaling. Further studies are needed to understand whether the ZDHHC3, 7 and 15, or ZDHHC1 play regulatory roles in STING-dependent immune responses.

Chemical proteomic studies showed that S-palmitoylation of several Toll-like receptors (TLRs) is important for their functions (Chesarino et al., 2014). For example, S-palmitoylation of TLR2 is required for its proper cell surface localization and NF-κB-dependent gene expression and cytokine production in response to PAM3CSK4. A palmitoylation-deficient mutant of TLR2 partially lost its ability to activate NF-κB-dependent gene expression. Moreover, it has been shown that S-palmitoylation of TLR2 is mediated by multiple ZDHHC-PATs, including ZDHHC2, 3, 6, 7 and 15.

S-palmitoylation-mediated regulation of cellular junction proteins, receptors and kinases

Proper establishment of cellular junctions is critical for cell identity, and often disrupted in cancers. Several junctional proteins are S-palmitoylated. For example, claudin family members contain membrane-proximal cysteines, which are S-palmitoylated (Gunzel and Yu, 2013). The palmitoylation-deficient mutant localized less to tight junctions and was localized in lysosomes instead, suggesting that palmitoylation regulates its trafficking. The junctional adhesion molecule C (JAM-C) is involved in cell migration, angiogenesis, cell-adhesion and polarity. JAM-C is palmitoylated by ZDHHC7, thus promoting junctional localization and inhibiting cancer cell migration (Aramsangtienchai et al., 2017). The cell polarity protein Scribble (SCRIB) localizes to cell-cell junctions and plays a regulatory role in determining epithelial cell polarity. We recently found that S-palmitoylation of SCRIB at two conserved membrane-proximal cysteine residues is required for its proper membrane localization (Chen et al., 2016a). ZDHHC7 is the primary acyltransferase and APT2 is a major depalmitoylating enzyme of SCRIB (Hernandez et al., 2017). Therefore, palmitoylation might be a conserved and common regulatory mechanism for regulating many junctional proteins and cell polarity.

S-palmitoylation plays important regulatory roles in GPCR protein signaling. Agonist-induced activation of the human canonical GPCR protein β2AR led to enhanced S-palmitoylation. The cycle of β2AR palmitoylation and depalmitoylation is mediated by the ZDHHC9, 14, or 18 and APT1, respectively (Adachi et al., 2016). Rhodopsin forms dimers and higher oligomers upon activation by light, and can bind to lipid rafts (raftophilicity) (Seno and Hayashi, 2017). Palmitoylation is required for dimerization-dependent raftophilicity of rhodopsin, playing an important role in the supramolecular organization and response to dim light (Seno and Hayashi, 2017). The melanocortin-1 receptor (MC1R) is melanocyte specific GPCR, and plays an important role in pigmentation. MC1R is palmitoylated at Cys315, critical for MC1R activation and protecting against melanomagenesis (Chen et al., 2017a).

Tyrosine kinases, including receptor (RTKs) and non-receptor tyrosine kinases (nRTKs), function in a wide range of key signal transduction pathways. Both RTKs and nRTKs can be modified by S-palmitoylation. The Src family of nRTKs, including Src, Yes, Fyn, Fgr, Lyn, Hck, Lck, Blk and Frk, are palmitoylated. S-palmitoylation of Lyn is required for its protective role against chromosome missegregation during the cell cycle (Honda et al., 2016). The epidermal growth factor receptor (EGFR) is an important RTK associated with tumorigenesis. S-palmitoylation of EGFR at Cys797 is required for its dimerization and activation, which is dependent on fatty acid synthase (FASN) (Bollu et al., 2015). Interestingly, EGFR is also S-palmitoylated at C-terminal Cys1025 and Cys1122, mediated by ZDHHC20 (Runkle et al., 2016). Both studies showed that S-palmitoylation plays a negative role in EGFR activation and that inhibition of EGFR S-palmitoylation may synergize with EGFR inhibitor-induced cell death.

S-palmitoylation in regulation of transcription factors

Estrogen receptor α (ERα) localizes to both the nucleus and the plasma membrane. S-palmitoylation at Cys451 is critical for ERα translocation, plasma membrane localization, and interaction with the membrane protein caveolin-1. A mouse model with palmitoylation-deficient ERα (C451A) was generated to obtain membrane-specific loss-of-function (Adlanmerini et al., 2014). Membrane-bound ERα receptor is critical in ovarian function, including fertility and vascular physiology. In contrast, nuclear ERα is required to mediate the uterine response to estrogen ligands. Therefore, S-palmitoylation of ERα has distinct tissue specific functions in vivo. Similarly, the androgen receptor (AR) is S-palmitoylated, which promotes AR membrane targeting. Several AR splicing variants are highly upregulated in castration-resistant prostate cancer (CRPC) cells. The AR8 variant was primarily localized on the plasma membrane, possibly through palmitoylation. The palmitoylation-deficient mutant of AR8 lost the membrane localization, and its ability to recruit Src and EGFR (Yang et al., 2011b). Therefore, palmitoylation-dependent membrane localization of AR variants may play a role in the development of CRPC.

The TEA domain transcription factors (TEAD1-4) are S-palmitoylated at conserved cysteine residues through a nonenzymatic autopalmitoylation mechanism (Chan et al., 2016). When incubated with physiological concentration of palmitoyl-CoA, TEAD2 undergoes efficient autopalmitoylation with apparent K_m of about 1 μM. We and others found that the palmitoyl chain is deeply buried into the conserved hydrophobic pocket (Chan et al., 2016; Noland et al., 2016). S-palmitoylation is required for TEAD binding to the transcription co-activator YAP/TAZ and mediates the transcriptional output of Hippo pathway (Chan et al., 2016). Interestingly, S-palmitoylation does not alter TEAD1 nuclear localization, and might function as a structural motif to rigidify the confirmation of TEADs. It remains unclear whether TEAD palmitoylation level is regulated through upstream signals, intracellular palmitoyl-CoA concentrations and the depalmitoylation process.

Dynamic regulation of S-palmitoylation

S-palmitoylation of proteins can be mediated by ZDHHC family of palmitoylating enzymes (Lemonidis et al., 2015). These enzymes catalyze the addition of palmitate to the substrate proteins with a two-step process: autoacylation and acyl-enzyme intermediate formation, followed by transferring the acyl chain to a specific cysteine residue in the substrate protein (Figure 3) (Jennings and Linder, 2012). In mammals, ZDHHC proteins are encoded by more than 20 distinct genes. ZDHHC-PATs are commonly localized in the ER, Golgi or the plasma membrane (Hentschel et al., 2016). Although some substrates of different ZDHHC proteins have been identified, the substrate specificity of these enzymes are not well understood (Table 1). One possibility is that the regions outside the DHHC domain determine the specificity of the substrate (Korycka et al., 2012). A recent structural study suggests that two conserved residues within the ankyrin repeat domain of ZDHHC17 play a critical role in binding to its substrate SNAP25b (Verardi et al., 2017). In addition, the intracellular localization of ZDHHCs might also specify the interactions with substrates. Genetic alterations of ZDHHC genes are observed in various diseases, including cancer and neurological disease (Lemonidis et al., 2015). As summarized in Table 2, several members of the family have been implicated in human physiological and pathological process. ZDHHC2 displays activities toward a broad range of substrates, and has been reported as a potential tumor suppressor in multiple human cancers (Jiang et al., 2015). In contrast, ZDHHC5 catalyzes EZH2 palmitoylation, and might function as an oncoprotein, contributing to the progression of p53-mutated glioma and non-small cell lung cancer (Chen et al., 2017b; Tian et al., 2015). ZDHHC9 mediates palmitoylation of N-Ras and H-Ras, and alterations of ZDHHC9 are associated with X-linked mental retardation and colorectal cancer (Mansilla et al., 2007; Raymond et al., 2007). Loss of ZDHHC13 or ZDHHC21 leads to hair defects, but these two PATs show different substrate preferences (Mill et al., 2009; Saleem et al., 2010). In addition, ZDHHC13 plays an important role in mitochondrial function and metabolism in liver, and is related to skin carcinogenesis and Huntington disease (HD) (Perez et al., 2015; Shen et al., 2017). ZDHHC21-deficiency results in endothelial inflammation and systemic inflammatory response syndrome (Beard et al., 2016). These biochemical and genetic data suggested that PATs are involved in diverse cellular functions, and substrate specificity and functional redundancy of different ZDHHC-PATs needs careful and detailed studies. Specific inhibitors may need to be developed for different PATs, which will require significant structural, biochemical and medicinal chemistry efforts.

Figure 3. The dynamic regulation of protein S-palmitoylation.

Palmitoyl acyltransferases (ZDHHC family enzymes, LPCAT) or autopalmitoylation are involved in adding palmitate to the Cys residue of proteins. The thioesterases or lipases (APTs and ABHDs) could remove the lipid chain from the proteins.

Table 1.

List of ZDHHC palmitoyl acyltransferases

Name	Localization	Known targets	Tissue distribution	Reference
ZDHHC1 (ANF377)	Early ED	Neurochondrin, EGFR	B, L, O	(Bollu, et al., 2015)
ZDHHC2 (Ream, ZNF372)	Recycling ED/PM	Lck, CKAP4/p63, CD9, CD151, SNAP25/23, PSD95, TLR2, EGFR	B, K, P, T, L, E	(Bollu, et al., 2015)
ZDHHC3 (Godz, Gramp1)	Golgi	GABA, PSD-95, eNOS, STING, TLR2, Gα, STREX	LV, S, L, B, PS, C, PL, E	(Mukai, et al., 2016; Tian, et al., 2010)
ZDHHC4 (ZNF374)	Golgi	BACE1	O, T, K, SK	(Vetrivel, et al., 2009)
ZDHHC5 (ZNF375)	PM	Flotillin-2, STREX, EZH2	O, T, K, H, SK, LN, BL	(Chen, et al., 2017; Li, et al., 2012)
ZDHHC6 (ZFN376)	ER	SSTR5, FLOT2, TLR2	K, LN	(Chesarino, et al., 2014; Yang, et al., 2010)
ZDHHC7 (ZNF370)	Golgi/PM	Glut4, Fas, CSP, GABA, PSD-95, eNOS, NCAM, SNAP25, PGR, AR, GAP43, STING, ERα, TLR2, Gα, STREX, Scribble, JAM-C	L, C, B, LV, PS, SK, K	(Aramsangtienchai, et al., 2017; Chen, et al., 2016; Chesarino, et al., 2014; Mukai, et al., 2016)
ZDHHC8	Golgi/PM	ABCA1, PICK1, GRIP1, eNOS	B, L, O, E, P, K	(Thomas, et al., 2012)
ZDHHC9 (ZNF380, CXorf11)	ER/Golgi	HRAS, NRAS, β2AR, STREX	B, PS, L, K, TL	(Adachi, et al., 2016; Tian, et al., 2010)
ZDHHC11 (ZNF399)	ER	Unkown	B, O, T, L
ZDHHC12 (ZNF400)	ER/Golgi	Unkown	ST, ascites, SK, L, PS, B
ZDHHC13	ER	HTT, SNAP25, MCAT, CTNND1, MC1R	U, B, ST, PL, T, C	(Chen, et al., 2017; Shen, et al., 2017)
ZDHHC14	ER	β2AR	B, LN	(Adachi, et al., 2016)
ZDHHC15	Golgi	CSP, PSD-95, Sortilin, CI-MPR, GAP43, STING, TLR2	B, O, K	(Chesarino, et al., 2014; Mukai, et al., 2016)
ZDHHC16	ER	c-ABL, JAB1	K, LV, SK, LN	(Zhang, et al., 2006)
ZDHHC17 (HIP14, HIP3, HYPH)	Golgi/CV/SV	SNAP25, PSD-95, GAD65, Synaptotagmin I, Huntingtin, CSP, STREX	B, U, E, L, TL	(Tian, et al., 2010)
ZDHHC18	Golgi	LCK, β2AR	T, LN, BL	(Adachi, et al., 2016; Greaves and Chamberlain, 2011)
ZDHHC19	ER	R-RAS	T	(Baumgart, et al., 2010)
ZDHHC20	PM	EGFR	B, O, T, L, C, LN, BL	(Runkle, et al., 2016)
ZDHHC21	Golgi/PM	FYN kinase, ERα, EGFR	B, T, U, E, LV	(Bollu, et al., 2015)
ZDHHC22	ER/Golgi	KCNMA1	B	(Tian, et al., 2012)
ZDHHC23		NOS1, KCNMA1	K	(Saitoh, et al., 2004)

Endosome, ED; ER, Endoplasmic reticulum; PM, plasma membrane; CV, cytosolic vesicle; SV, synaptic vesicle; B, brain; L, lung; O, ovary; K, Kidney; P, pancreas; T, testis; E, eye; LV, liver; S, Sleen; PS, prostate; C, colon; PL, placenta; SK, skeletal muscle; LN, lymph node; BL, blood; U, uterus; ST, stomach; SK, skin; H, heart; TL, thalamus.

The data of subcellular localization and tissue distribution of ZDHHCs is also collected from this reference (Ohno, et al., 2006).

Table 2.

Physiological and pathological functions of ZDHHCs

Name	Physiological and pathological functions	Substrates	References
ZDHHC2	Colorectal cancer (H) Nasopharyngeal carcinoma Neuro physiology	CKAP4/p63 (D) CD9, CD151 (H) Mir-155, PSD-95	(Fukata, et al., 2013; Jiang, et al., 2015)
ZDHHC3	Neuro physiology	D2 dopamine receptor	(Ebersole, et al., 2015)
ZDHHC4	Neuro physiology	D2 dopamine receptor	(Ebersole, et al., 2015)
ZDHHC5	NSCLC, Glioma	EZH2	(Chen, et al., 2017; Tian, et al., 2015)
ZDHHC8	Schizophrenia (H, M) Neuro physiology	D2 dopamine receptor	(Ebersole, et al., 2015)
ZDHHC9	X-linked mental retardation (H) Colorectal cancer (H)	NRAS and HRAS Unknown	(Mansilla, et al., 2007)
ZDHHC13	Skin Carcinogenesis (M)	Unknown	(Perez, et al., 2015)
ZDHHC17	Huntington disease (M) Synaptic transmission (D)	Hun tingtin (M) CSP, SNAP25 (D)	(Ohyama, et al., 2007)
ZDHHC20	Lung cancer	EGFR	(Runkle, et al., 2016)
ZDHHC21	Endothelial inflammation	PLC-β1	(Beard, et al., 2016)

H: human; M: mouse; D: Drosophila

As there are only 23 DHHC-family PATs (Fukata et al., 2004; Ohno et al., 2006), it is unlikely that they are responsible for all the palmitoylation activities in cells (more than 1,000 protein substrates are S-palmitoylated). Therefore, it is possible that many S-palmitoylated proteins are modified through ZDHHC-PAT-independent processes, such as autopalmitoylation. Previously, autopalmitoylation was considered to be a nonspecific reaction that occurred when surface cysteine residues encountered a high concentration of palmitoyl-CoA. However, our studies of TEAD proteins, together with the studies of yeast Bet3 protein and myelin P0 glycoprotein, have shown that autopalmitoylation can happen under physiological conditions, with specific cysteine residues being modified (Bharadwaj and Bizzozero, 1995; Chan et al., 2016; Turnbull et al., 2005)(Figure 4). It has been noted that ZDHHC proteins also undergo “autopalmitoylation” when bound to palmitoyl-CoA, and form acyl-enzyme adduct (Mitchell et al., 2006; Roth et al., 2006). It is possible that many proteins have such intrinsic enzyme-like activities, and could form acyl-protein adducts. As shown in both the TEAD and Bet3 structures, the modified cysteine residue is located close to the palmitate-binding pocket, similar to the lipid-bound ZDHHC20 protein structure (Rana et al., 2018). It is possible that such an acylation process can be facilitated by proximity-mediated thioester exchange.

Figure 4. Function and regulation of TEAD autopalmitoylation.

The Hippo pathway transcription factor TEAD is autopalmitoylated. Autopalmitoylation of TEAD may be regulated by intracellular concentration of palmitoyl-CoA, which is controlled by fatty acid metabolism. Palmitoylation of TEAD is required for its association with YAP and the regulation of transcriptional output of Hippo signaling.

If autopalmitoylation can happen readily in the presence of acyl-CoA, how is this process regulated? Two possible mechanisms have been suggested: thiolate formation and the availability of acyl-CoA (Dietrich and Ungermann, 2004). At physiological pH (7.2–7.4), the spontaneous formation of thiolate to the cysteine thiol group (pKa ~ 8.5) is unlikely. However, in a proper local environment of folded proteins, the nucleophilicity of a cysteine can be potentially modulated by polarization of other side chains. The cellular acyl-CoA levels could also play a role. The conserved acyl-CoA binding protein (ACBP) binds free palmitoyl-CoA and maintains the intracellular concentration (Faergeman and Knudsen, 1997), which might buffer the excessive autopalmitoylation. The “autopalmitoylated” proteins and PATs might compete with ACBP for palmitoyl-CoA binding. Thus, only specific proteins with strong acyl-CoA binding could be autopalmitoyalted. Furthermore, upstream regulators of lipid biosynthesis, such as fatty acid synthase (FASN) may indirectly regulate autopalmitoylation through regulating fatty acid biosynthesis (Figure 4).

Palmitoyl protein thioesterases 1 and 2 (PPT1 and PPT2) remove long-chain fatty acids (usually palmitate) from S-fatty acylated proteins during lysosome degradation (Segal-Salto et al., 2016). The function of PPTs is confined to the lysosomal degradation of S-fatty acylated proteins. Acyl protein thioesterases (APT1 and APT2) depalmitoylate membrane-anchored proteins and play critical roles in palmitoylation turnover. For a long time, APTs were thought to be the only cytosolic thioesterases, responsible for almost all of the depalmitoylation processes. Pharmacological inhibition of APT1/2 activities block Ras trafficking and inhibit Ras oncogenic activities (Dekker et al., 2010). Recently, specific APT1 and APT2 inhibitors were developed, providing important chemical tools for dissecting the function of APTs (Adibekian et al., 2012; Won et al., 2016). Fluorescence-based depalmitoylation probes (DPPs) have been reported to visualize the enzymatic activity of APTs in cells and explore new mechanism of depalmitoylation (Kathayat et al., 2017).

However, enzymes responsible for depalmitoylation are more diverse than previously thought. α/β hydrolase domain (ABHD) proteins have been suggested as potential depalmitoylases (Martin et al., 2011). Palmostatin B, the inhibitor of APT1/2, also inhibits FASN, PNPLA6, and ABHD proteins (Lin and Conibear, 2015). ABHD17 removes palmitate from N-Ras efficiently (Lin and Conibear, 2015). Almost at the same time, by screening 38 serine hydrolases containing all ABHD proteins, ABHD17 proteins have been identified as physiological PSD-95 depalmitoylating enzymes and regulate local PSD-95 palmitoylation cycles in neurons (Yokoi et al., 2016). The mammalian ABHD family of proteins consists of at least 19 members. The substrates and the physiological functions of most of the ABHD proteins are unknown (Lord et al., 2013). Therefore, studies of ABHD activities might reveal additional depalmitoylases involved in signaling and diseases (Table 3).

Table 3.

Localization and targets of depalmitoylating enzymes

Name	Localization	Known targets	Reference
APT1	Mainly localized to the cytosol	SNAP-23, Gsα, eNOS, H-Ras, KCMA1, GAP-43, MCAM, NMNAT2	(Milde and Coleman, 2014)
APT2	Mainly localized to the cytosol	H-RAs, GAP-43, NMNAT2, Scribble	(Hernandez, et al., 2017; Milde and Coleman, 2014)
PPT1	Lysosome	HRAS, Palmitoyl-CoA	(Linder and Deschenes, 2007)
PPT2	Lysosome	Palmitoyl-CoA	(Linder and Deschenes, 2007)
ABHD17	Endosome, plasma membrane	PSD95, NRAS, MAP6	(Lin and Conibear, 2015; Tortosa, et al., 2017; Yokoi, et al., 2016)

Dynamic palmitoylation might also play a critical role in response to the extrinsic signaling stimulations. Stimulation of some receptors can lead to increased palmitoylation turnover or changed palmitoylation levels. Palmitoylation turnover on G proteins is accelerated upon ligand binding and stimulation of their related GPCRs, thus regulating the downstream signaling activities (Barclay et al., 2005). Stimulation of the fibroblast growth factor receptor (FGFR) by FGF2 results in palmitoylation of neural cell adhesion molecule (NCAM) proteins (Ponimaskin et al., 2008). EGF stimulation can transiently inhibit depalmitoylation activity of APTs (Kathayat et al., 2017). Palmitoylation of GPCR protein MC1R is stimulated by α-melanocyte-stimulating hormone (α-MSH) and removal of α-MSH from the culture medium diminishes MC1R palmitoylation (Chen et al., 2017a).

Deacylases of fatty acylation of lysine

Besides S-fatty acylation on the cysteine residues, N-fatty acylation on the lysine residues is emerging as an important modification. Lysine residues can be modified by various acyl groups, such as acetyl, butyryl, myristoyl and palmitoyl group. Sirtuin 6 (SIRT6) displays much lower deacetylase activity in vitro compared with other sirtuin family of deacetylases. However, free fatty acids, including myristic, oleic, and linoleic acids, at physiological concentrations induce up to a 35-fold increase of SIRT6 catalytic efficiency (Feldman et al., 2013). Moreover, SIRT6’s active site can accommodate longer acyl chains, such as myristoyl, and can remove long-chain fatty acyl groups from lysine residues, raising the possibility that SIRT6 could be a deacylase for N-fatty acylation (Jiang et al., 2013). SIRT6 promotes the secretion of TNF-α by removing the fatty acylation of K19 and K20 (Jiang et al., 2013). Recently, it has been reported that SIRT6 regulates the lysine de-fatty acylation of R-Ras2 (Zhang et al., 2017). Sirt6-null mouse embryonic fibroblast cells exhibit elevated R-Ras2 lysine fatty acylation. The histone deacetylase, HDAC8 can remove a variety of acyl groups (C2-C16 acyl chains) from Lys9 of the histone H3 peptide (H3K9) (Aramsangtienchai et al., 2016), and SIRT7 can remove long-chain fatty acyl groups more efficiently than acetyl group (Tong et al., 2017). Taken together, it is possible that N-fatty acylation of lysine residues could be a novel and dynamic regulation, and SIRT6, SIRT7 and HDAC8 could be the potential de-fatty acylases. It remains unclear how many proteins are N-fatty acylated on lysine residues under physiological conditions, and the function of lysine acylation is elusive. The fatty acyltransferase catalyzing the lysine N-fatty acylation remains unknown.

Opportunities in drug discovery

As protein lipidation plays a critical role in regulating protein function, and its deregulation is involved in many human diseases, targeting protein lipidation could be an important therapeutic strategy. Proof of concept studies have shown that several lipidation-related enzymes are attractive drug targets, and many inhibitors and chemical probes have been developed over the past years (Table 4).

Table 4.

Chemical modulators of protein lipidation

Name	Structure	Mechanism and biological activity	Reference
2-Bromopalmitate (2BP)		Non-specific, irreversible alkylating palmitoyl acyltransferases	(Resh, 2006)
Cerulenin		Irreversible inhibitor of fatty acid synthase and palmitoyl acyltransferases	(Resh, 2006)
Compound V		Reversible inhibitor of palmitoylation and myristoylation.	(Jennings, et al., 2009)
Palmostatin B		Inhibitor of APT1/APT2/ABHD17 IC50 = 0.67 μM (Ras depalmitoylation)	(Dekker, et al., 2010)
ML348		Selective reversible inhibitor of APT1, IC50 = 0.21 μM	(Adibekian, et al., 2012)
ML349		Selective reversible inhibitor of APT2, IC50 = 0.144 μM	(Hernandez, et al., 2017)
IWP-L6		Porcupine inhibitor (EC50 = 0.5 nM), showing good stability in human plasma	(Wang, et al., 2013)
LGK974 (Wnt974)		Potent and specific Porcupine inhibitor (IC50 = 0.4 nM). In clinical trials	(Liu, et al., 2013)
C59		A nanomolar inhibitor of Porcupine, displaying good bioavailability	(Proffitt, et al., 2013)
ETC-159		Potent and orally available Porcupine inhibitor (IC50 = 2.9 nM), In clinical trial.	(Madan, et al., 2016)
DDD85646		A potent inhibitor of T. brucei NMT with IC50 of 2 nM	(Brand, et al., 2012)
Tipifarnib		A non-peptidomimetic quinolinone inhibitor of FTase, being tested for cancers	(End, et al., 2001)
Lonafarnib		A tricyclic derivative of carboxamide inhibitor of FTase, in clinical trial for progeria patients	(Liu, et al., 2007)
BMS-214662		An inhibitor of FTase, showing potential antineoplastic activity	(Hunt, et al., 2000)
Deltarasin		A potent inhibitor of PDEδ-KRAS interaction by occupying PDEδ farnesyl binding pocket, with a Kd of 0.04 μM	(Zimmermann, et al., 2013)
GNF-5		A selective and allosteric Bcr-Abl inhibitor (IC50 = 0.22 μM), occupying the myristate-binding site	(Zhang, et al., 2010)
GO-CoA-Tat		A peptide-based bisubstrate antagonist of GOAT, improving glucose tolerance and reducing body weight in mice	(Barnett, et al., 2010)

Targeting lipidation for cancer therapeutics

Ras proteins (K-, H, and N-Ras) are among the best known oncoproteins. H- and N-Ras are prenylated and then palmitoylated. Farnesyl transferase inhibitors have been developed to inhibit Ras prenylation, however, they failed to show significant clinical efficacy (Berndt et al., 2011). Alternatively, inhibitors of PDEδ have been developed. PDEδ binds to farnesyl chain of Ras. The inhibitors occupied the lipid binding pocket and blocked Ras binding (Zimmermann et al., 2013). H- and N-Ras undergo cycles of palmitoylation and depalmitoylation, which are important in the regulation of Ras localization and signaling activity. H-Ras and N-Ras are palmitoylated by ZDHHC9 (Swarthout et al., 2005). Small molecules have been developed to disrupt protein palmitoylation process to inhibit Ras activity (Jennings et al., 2009), but their selectivity is not optimal. Given the functional redundancy of DHHC proteins, it might be a challenge to develop a specific DHHC inhibitor to achieve therapeutic efficacy.

The dual inhibitor of APT1 and 2, palmostatin B (Pal B), has been developed to inhibit depalmitoylation of H-Ras. N-Ras is localized to internal membranes upon Pal B treatment (Dekker et al., 2010). However, the effects of Pal B may be from the inhibition of other enzymes, such as ABHD17 (Lin and Conibear, 2015). In contrast to Pal B, ML348 and ML349 are selective inhibitors of APT1 and 2 (Adibekian et al., 2012).

Although not discussed at length here, myristoylation also plays a crucial role in protein modification and control; in particular in the regulation of c-Abl kinases (Hantschel et al., 2003). Binding of myristoyl group into the conserved hydrophobic pocket locks the kinase domain in an inactive conformation. The oncoprotein Bcr-Abl lacks N-terminal myristoylation due to gene fusion, but the myristoyl-binding pocket is intact in the kinase domain, thus providing opportunities to design allosteric inhibitors. Small molecule inhibitors (GNF-2 and GNF-5) bind to myristoyl-binding pocket and alter the structural dynamics of the kinase domain, thus allosterically inhibiting the kinase. GNF-5 can overcome the resistance of kinase inhibitors (imatinib or nilotinib) (Zhang et al., 2010). Another inhibitor, DPH, binds to the myristoyl-binding site and prevents bending of the α I helix via steric hindrance, leading to c-Abl inactivation (Yang et al., 2011a). Other proteins with similar myristoyl-binding sites could be potential candidates for such strategies.

Aberrant Wnt signaling is involved in many types of cancer (Clevers and Nusse, 2012). Wnt proteins are fatty acylated by Porcupine (PORCN). Disruption of Wnt fatty acylation can suppress its secretion, thus inhibiting the growth of Wnt-dependent cancers. Several PORCN inhibitors have been developed. IWP compounds bind to PORCN and block its acyltransferase activity (Chen et al., 2009), and more potent analogues were also developed (You et al., 2016). GNF-1331 is a potent “hit” compound inhibiting Wnt secretion through binding to PORCN. Further optimization of GNF-1331 led to the identification of Wnt-C59 and GNF-6231 (Cheng et al., 2016). Furthermore, LGK974 (also known as WNT974) has been developed as a clinical candidate, with optimal pharmacokinetic and toxicity profiles. LGK974 inhibits Wnt signaling with impressive efficacy and is well tolerated in vivo (Liu et al., 2013). LGK974 is currently in clinical trials for treating patients with malignancies dependent on Wnt signaling (clinical trial number: NCT01351103).

Deregulation of the Hedgehog (Hh) pathway is associated with multiple human cancers, including medulloblastoma and basal cell carcinoma. Hh ligands are palmitoylated in the N-terminus by the Hh acyltransferase (HHAT) (Buglino and Resh, 2008), which is critical for the regulation of Hh signaling. A small molecule, RU-SKI 43, inhibits HHAT and blocks both autocrine and paracrine Hh signaling (Petrova et al., 2013). RU-SKI 43 decreases Gli-1 activation, and inhibits Akt/mTOR pathways and pancreatic cancer cell proliferation (Petrova et al., 2015). RU-SKI 43 is a useful chemical tool to explore HHAT functions in cancer. Further optimization and in vivo validation are needed before advancing to human clinical trials.

Targeting the deacylating enzymes provides an important approach for disruption of the fatty acylation cycles of S-palmitoylated proteins. Palmostatin B (Pal B) is an inhibitor of APT1/2, which can partially disrupt Ras localization and oncogenic activities (Dekker et al., 2010). However, Pal B inhibits other proteins, including ABHDs, which might be responsible for its Ras depalmitoylation activities. In addition, it has not been shown that Pal B allows in vivo studies. Recently, specific inhibitors for APT1 and APT2 have been reported. ML349 and ML348 have good in vivo pharmacokinetic properties in mice (Adibekian et al., 2012). ML349 can rescue Snail-induced Scribble mislocalization and tumor suppressor function by selectively inhibiting APT2 activity, but not APT1 (Hernandez et al., 2017).

ABHD family proteins are novel potential deacylases (Lin and Conibear, 2015; Yokoi et al., 2016). Development of potent and selective small molecule inhibitors of ABHDs will be important. In addition, SIRT6 deacetylates TNF- α, thus promoting its secretion (Jiang et al., 2013). SIRT6 might regulate the secretion of many other proteins (Zhang et al., 2016). A specific inhibitor blocking SIRT6 deacylase activity will be helpful for determining its biological function and for drug discovery.

Targeting lipidation for infectious diseases

Plasmodium falciparum, the parasite causing malaria in humans, expresses more than 400 palmitoylated proteins and over 30 myristoylated proteins (Jones et al., 2012; Wright et al., 2014). P. falciparum N-myristoyltransferase (NMT) has been validated as an attractive antimalarial drug target. NMT inhibition blocks the formation of essential parasite subcellular structures, leading to cell death (Wright et al., 2014). Trypanosoma brucei, the agent of human African trypanosomiasis, expresses a NMT that has been identified as a promising therapeutic target for sleeping sickness (Frearson et al., 2010). The inhibitor of the T. brucei NMT, DDD85646, inhibits the growth of a bloodstream form T. brucei in culture with EC₅₀ values between 0.8 and 3 nM (Brand et al., 2012).

Plasmodium PATs are also potential antimalarial drug targets. ZDHHC2 is essential for ookinete morphogenesis and malarial transmission (Santos et al., 2015). Recent studies have identified large families of PATs and palmitoylated proteins in other parasites as well, including Toxoplasma gondii, Trypanosoma brucei, Giardia lamblia and the fungal pathogen Cryptococcus neoformans (Brown et al., 2017). Therefore, inhibiting parasite PATs and thioesterases could be important therapeutic strategies for these devastating diseases.

Similarly, fatty acylation of viral and host proteins plays an important role in virus-host interactions and immune responses. Myristoylation of HIV Gag protein regulates Gag membrane binding and particle budding (Resh, 2004). A recent chemical proteomic study has identified many new acylated viral proteins during infection of the herpes simplex virus (Serwa et al., 2015). Lipidation of some host proteins may also be critical for the host resistance to viral infection. Palmitoylation of interferon-induced transmembrane protein 3 is required for its antiviral activity against influenza (Yount et al., 2010). Therefore, targeting lipidated viral or host proteins may lead to new antiviral agents.

Future outlook

There remain many unanswered questions about the regulation and function of protein lipidation, especially for fatty acylation of proteins. First, are there other classes of palmitoylating enzymes besides ZDHHC proteins? ZDHHC proteins are usually membrane-bound and residing in the ER, Golgi or plasma membrane. However, there are also many non-membrane bound S-palmitoylated proteins. The discovery of LPCAT1 as a palmitoylating enzyme for histone H4 suggested that indeed other enzymes might have “moonlight” functions as fatty acylating enzymes. Further studies may reveal new enzymatic activities involved in S-palmitoylation, and expand the current spectrum of palmitoylating enzymes. In addition, many S-palmitoylated proteins may undergo “non-enzymatic” autopalmitoylation, which will require careful characterization and validation.

Second, are there enzymes catalyzing the N-fatty acylation of Lys residues? It has been shown that many important signaling proteins can be fatty acylated on their Lys residues. Sirtuin family proteins have been shown to regulate de-fatty acylation. Eight types of short-chain fatty acylation of Lys residues in histones have been identified, whereas long-chain fatty acylation of histones has not been well studied (Sabari et al., 2017). Interestingly, histones S-palmitoylation and O-palmitoylation have been reported to regulate gene transcriptional activities (Wilson et al., 2011; Zou et al., 2011). New types of long-chain fatty acylation of histones lysine residues may exist. Therefore, it will be important to reveal the origin of N-palmitoylation on lysine resides, and to determine which enzymes catalyze such modifications.

Third, how is cellular lipid metabolism regulating protein lipidation? Lipid peroxidation products can covalently modify proteins and participate in the regulation of multiple biological processes (Chen et al., 2016b). Although many targets have been identified through chemoproteomic studies, detailed follow-up studies are required to validate these findings. In addition, fatty acid synthase (FASN), which is often overexpressed in cancers, has been identified as a potential therapeutic target. These finding suggest that cancer cells might become “addicted” to elevated lipid levels. It will be important to validate and characterize the “lipid addition” phenotype of cancers, and to reveal whether high lipid levels is linked to deregulated protein lipidation. Such work may provide insights into the role of lipidation in complex diseases, and bring novel therapeutic targets for drug development.