The LipidWeb blank

Proteolipids and Protein Lipidation

In 1951, proteins that were soluble in organic solvents such as chloroform-methanol were found in rat brain myelin by Folch and Lees (perhaps better known for devising the most common method for lipid extraction), who coined the term ‘proteolipid’. However it was another twenty years before it was shown that these contained covalently bound fatty acids and so differed from the plasma lipoproteins. Such lipid-modified proteins are now known to be widespread in nature with many important functions in animals, plants and bacteria. Proteolipids can be defined as all proteins containing covalently bound lipid moieties that include fatty acids, isoprenoids, cholesterol and glycosylphosphatidylinositols (the last have their own web page). Protein lipidation provides an interface between the lipophilic membranes of cells and otherwise hydrophilic proteins that enables signalling and many other functions at the membrane surface. The term ‘lipoprotein’ is also used to describe such compounds on occasion, but to avoid confusion this might be better reserved for the non-covalently linked lipid-protein complexes of the type found in plasma.

1.  Introducing Protein-Lipid Modifications

It is a curious but important fact that the two main types of protein with fatty acid modifications to have been described from eukaryotic organisms contain saturated fatty acid components, i.e. those with only myristoyl and those with predominantly palmitoyl moieties, each with a distinctive type of linkage, amide or thiol ester, respectively. Prenylated lipids contain an isoprenoid group linked via a sulfur atom (thiol ether bond) to the protein. Other protein-lipid links exist and are discussed below, but they tend to be less frequent.

Figure 1. Formulae of myristoylated, palmitoylated and prenylated proteins

A further important class of proteolipids contains a linkage to cholesterol in addition to N-palmitoylation, the ‘hedgehog’ signalling proteins, while bacteria contain proteolipids with N-acyl- and S-diacylglycerol groups attached to an N-terminal cysteine. Some O-acylated proteins are also known. Often, N-myristoylated and prenylated proteins have one or more additional S-acylation residue. Such proteolipids are important for the functioning of all classes of eukaryotes (animals, plants and fungi) and of bacteria.

Figure 2. Protein acylation forms

Modification with lipids occurs most often after synthesis of the proteins and the effect is to change them from a generally hydrophilic nature to one that is hydrophobic at one end at least, thus facilitating the interaction with membranes. Protein S-palmitoylation occurs at the membrane surface, while N-myristoylation occurs at the ribosomal level and S-prenylation is brought about by cytoplasmic enzymes. It is now clear that such modifications are important in determining the activities of proteins and in targeting them to specific subcellular membrane domains, including the rafts or caveolae in plasma membranes. Thus, both myristoylated and palmitoylated proteins are targeted to rafts (as are the GPI-anchored proteins), but prenylated lipids are not. It is significant that many signalling proteins (e.g. receptors, G-proteins, protein tyrosine kinases) and often their substrates are modified by lipids with implications for the relevant signalling events at the cell surface. Many of these proteolipids influence human disease states and are potential pharmacological targets. For example, deregulation of palmitoylation has been associated with heart disease, cancer, mental retardation and schizophrenia. Fatty acylation is essential for the survival, growth and infectivity of the trypanosomatids (protozoal parasites) in humans. In addition, some pathogenic organisms can hijack the protein acylation mechanism to increase the susceptibility of the host to infection.

The reversible nature of S-palmitoylation in particular as a regulator of cellular functions has drawn comparisons with other important reversible post-translational modifications such as phosphorylation and ubiquitination.

2.  N-Myristoylated Proteins

In the N-myristoylated proteins, myristic acid (14:0) specifically, which is a ubiquitous but usually minor component of cellular lipids (only 1% of the total fatty acids), is bound to the amino-terminal glycine residue of a relatively conserved sequence of the protein via an amide linkage that is comparatively stable to hydrolysis. Therefore, N-terminal acylation is believed to be an irreversible modification, although there may be exceptions. These proteolipids constitute a large family of essential eukaryotic proteins (0.5-0.8% of the total) with many different functions, and they are located either in the cytosol or in the cytosolic (inner) membrane of cells, or both. Indeed, N-glycine myristoylation mediates the targeting of the modified proteins to various membranes in the cell, including the plasma membrane, endoplasmic reticulum, Golgi, mitochondria and nuclei. The acyl group anchors the protein to membranes, although simultaneous binding to phospholipids or other membrane constituents is necessary to increase the strength of the interaction. For example, a single N-myristoylation is not sufficient to ensure membrane association and a second signal is necessary, e.g. hydrophobic residues, another membrane-bound binding partner or further N-myristoylation or S-palmitoylation (see below).

In vivo, the acyl group as the CoA ester is attached by the action of myristoyl-CoA:protein N-myristoyltransferases (NMT1 or NMT2 in humans) to the N-terminal glycine of the growing peptide as it begins to emerge from the ribosome, i.e. it is mainly a co-translational rather than a post-translational event. NMT1, but not NMT2, is essential for cell proliferation, but both NMT1 and NMT2 are probably required for cell survival. While they are both cytosolic, they can bind to the ribosomes and so may facilitate the co-translational timing of the reaction. Both have an N-terminal region with polybasic amino acid sequences, which are required for targeting to the ribosomes. There is an absolute requirement for the N-terminal glycine on the target protein, while amino acid residues two through eight are the consensus motif (Met-Gly-X-X-X-Ser/Thr/Cys) used in substrate recognition. The leader methionine residue is first removed from the nascent peptide chain by a methionine aminopeptidase (MAP) to expose the N-terminal glycine, before the N‑myristoyl transferase catalyses the formation of the stable amide bond. Myristoyl-CoA is supplied by an acyl-CoA binding protein, and this controls the specificity of the reaction by forming a complex with the transferase thereby stimulating it and preventing it from utilizing the competitor palmitoyl-CoA. In addition, it has become apparent that myristoylation can also occur post-translationally on internal glycine residues and in apoptotic cells, when N-terminal glycine residues are exposed in partially hydrolysed proteins. The latter may have implications for health and disease.

N-Myristoylation of proteins

Exceptions to these generalities are photoreceptor proteins, which are modified heterogeneously with the uncommon 12:0, 5‑14:1 and 5,8‑14:2 fatty acids as well as 14:0. N-Palmitoylated proteins have been found on occasion, usually where there is a dual lipid modification, for example, the cholesterol-linked hedgehog proteins (see below). In addition, myristoylation occurs on the ε-amine group of internal lysines in interleukin 1α and tumor necrosis factor α, and certain histone proteins can be modified similarly by various fatty acids. With these, different enzymes or iso-enzymes from those for N-terminal myristoylation are involved.

Scottish thistleMore than 150 mammalian proteins are known to be myristoylated, and these include proteins that play important roles in signaling networks, apoptosis, oncogenesis and viral replication. The functions of such proteolipids are steadily being unravelled, and it is evident that they are of critical physiological importance, for example as participants in cellular signalling. The lipid moieties are involved in regulating protein activity perhaps by modifying or stabilizing their conformations, by facilitating protein-protein interactions, and in targeting otherwise soluble proteins to the membranes and to appropriate receptors. N-Myristoylation is essential for erythrocyte formation and myelination, it is necessary for calcium release in mitochondria and is an important factor in T cell function and thence in the immune response. During apoptosis, NMTs are substrates for caspases, which are believed to regulate their location from the ribosomal and membrane regions to the cytosol and vice versa. Increased levels of N-myristoylation have been observed in certain cancers, and there have been suggestions that NMTs could be a target for therapeutic intervention. Also, N-myristoyl transferase activity is believed to play a role in infections by viruses, bacteria, fungi and parasitic protozoa, including such diseases as malaria, and again the host NMTs are seen as potential drug targets.

Myristoylation is an irreversible stable modification and the half-life of a myristoylated protein is similar to that of the nascent polypeptide chain. On the other hand, N-myristoylated proteins can be dissociated from membranes by de-S-palmitoylation, by phosphorylation, by binding of the hydrophobic moieties to competing cytosolic proteins, if the myristoyl moiety is relatively exposed, or by conformational changes, and these reactions may be forms of regulation. The low levels of myristic acid in tissues may also be a controlling factor. Macrophages contain at least one protease that targets certain N-myristoylated proteins, and similar protease activity has been reported from other tissues.

In plants, over 300 Arabidopsis proteins are potentially myristoylated, including many protein kinases, phosphatases, thioredoxins and transcription factors. As an example, the calcium-sensing molecule CBL1 requires N-myristoylation for association with the endoplasmic reticulum followed by S-acylation to promote trafficking to the plasma membrane; in contrast CBL2 is not N-myristoylated but rather is triply S-acylated to direct it to the tonoplast.

Mass spectrometry is currently a key method for characterization of N-myristoylated proteins, although the fatty acyl group can be released for analysis by conventional chromatographic means (e.g. gas chromatography) by the acidic hydrolysis conditions commonly employed to cleave peptide bonds. For example, treatment with 6M HCl or 2M HCl in 83% methanol at 100°C for several hours is required to release the N-acyl group as the free fatty acid or methyl ester, respectively. New methods involving specific chemical probes are now facilitating detection and analysis of N-myristoylated proteins and other proteolipids.

3.  S-Palmitoylated Proteins

In the S-palmitoylated proteins, palmitic acid (16:0) is linked to one or more (up to four) internal cysteine residues via labile high-energy thioester bonds. The name is something of a misnomer, as other fatty acids are sometimes present, including 16:1, 18:0 and 18:1, so the term 'S-acylation' may be preferable. For example, S-18:0 is as common as S-16:0 in plants, and viral proteins are often linked to C18 fatty acids. S-Palmitoylation is also observed in conjunction with N-terminal myristoylation or C-terminal prenylation sites. In contrast to N-myristoylation and isoprenylation, there does not appear to be any specific peptide target sequence.

The archetypal proteolipid protein (‘PLP’) or lipophilin is the main protein in the myelin of the central nervous system; it has multiple functions and was the first of its type to be identified and properly characterized. It is a highly conserved hydrophobic protein of 276 to 280 amino acids with four transmembrane segments that binds at least six palmitate groups via thioester bonds. A wide variety of different palmitoylated proteins, with many different functions are now known. For example, well over 500 different palmitoylated proteins have been identified in humans (10% of the proteome) and more than 50 in the yeast Saccharomyces cerevisiae. These can be grouped into three broad categories - poly-acylated membrane proteins (e.g. some receptors and rhodopsin), mono-acylated membrane proteins (some receptors and viral proteins), and hydrophilic proteins (such as certain protein kinases). For example in brain, 490 palmitoylation sites have been identified on 342 synaptic proteins, 44% of which are integral membrane proteins. It is now apparent that protein palmitoylation is essential for intracellular signalling and for the folding, trafficking and function of such disparate proteins as Src-family kinases, Ras family GTPases, G-proteins and G-protein coupled receptors.

A DHHC protein in a membraneThio-acylation occurs post-translation of the protein, and is catalysed by specific endomembrane-bound acyltransferases. Although there is some evidence for occasional non-enzymatic palmitoylation, enzymatic mechanisms predominate and palmitoyl S-acyltransferases were identified definitively first from yeasts and subsequently from mammalian cells. A family of such enzymes (23 in humans coded by more than 20 separate genes and a similar number in Arabidopsis) has now been characterized with a conserved cysteine-rich domain containing zinc and a distinctive aspartate-histidine-histidine-cysteine (DHHC) motif, which is required for activity. They are membrane proteins with a number of subcellular locations that span the bilayer at least four times with the DHHC domain and the N- and C-terminal domains on the cytosolic face. Additionally, there is a conserved C-terminus (PaCCT) domain in most palmitoyl transferases and this is important for their function. Mammalian DHHC proteins appear to be associated with specific subcellular locations but mainly the endoplasmic reticulum and Golgi, but in plants a high proportion are located at the plasma membrane. While many palmitoyl transferases have overlapping substrate specificities, some are highly specific for particular proteins.

Protein palmitoylation-depalmitoylation

Cysteine palmitoylation forms a thioester bond that is similar in energy to that in the palmitoyl donor, palmitoyl-CoA, so the reaction does not require an energy source such as ATP. The protein transacylases are all palmitoylated spontaneously when incubated with palmitoyl-CoA with release of CoA, suggesting that the auto-palmitoylated acyl-enzyme intermediate is involved in the transfer of the palmitoyl moiety to a substrate. Indeed, it is now established that the cysteine residue in the DHHC motif is palmitoylated for transfer to the target protein, as part of a two-step catalytic mechanism. There is an absolute requirement for long-chain acyl-CoA esters, mainly 16:0, as fatty acyl donors, though some members of the DHHC protein family have different acyl specificities and at least one can use a wide variety of fatty acyl-CoA substrates. Prior to S-acylation, modification by N-myristoylation or prenylation is often required. However, there does not appear to be a preferred substrate-sequence motif in target proteins, but rather there is a random palmitoylation process that depends simply upon how accessible cysteines on membrane-embedded proteins are to the palmitoyl transferases. Proteomic methods now enable prediction of which of these cysteine residues is likely to be palmitoylated.

In contrast to irreversible N-myristoylation, hydrolysis of S-palmitoylated proteins occurs readily and is catalysed by cytosolic thioesterases, two of which have been now been characterized, i.e. acyl protein thioesterase 1 (APT1) and LYPLA2 (APT2), each with specificities for particular classes of protein conjugates. Two lysosomal enzymes, palmitoyl protein thioesterases 1 and 2 (PPT1 & 2), have related activities. Thus, most proteolipids of this type undergo cycles of acylation-deacylation, with a half-life that is much shorter than that of the peptide per se, and this permits proteins to shuttle between membranes and other cellular compartments, for example between the plasma membrane and Golgi in both directions. This acylation-deacylation cycle is believed to have a regulatory function, and for example, some proteins are prenylated and S-acylated in an active form, while the inactive form is only prenylated. The activities of synthetic and hydrolytic enzymes are regulated dynamically by extracellular stimuli, like phosphorylation, and the level of palmitoylation is determined by a finely tuned balance between the activities of these enzymes in specific cellular locations.

Scottish thistleAs with the myristoylated proteins, palmitoylation is believed to modify protein function partly by modifying their conformations, but mainly in targeting otherwise soluble proteins to specific membranes or to appropriate receptors. The number of bound fatty acyl groups may control the strength of the interaction with membranes. For example, a hydrophobic protein with a single acylation can bind only loosely to membranes and is easily displaced. However, a second or further acylation ensures strong targeting of a protein to the cytoplasmic face of the membrane, and guarantees that it is firmly bound to a specific site on the membrane where an appropriate receptor may be located. Palmitoylation of integral membrane proteins may be a regulatory function, changing their conformation to increase their stability through protecting them from degradation by preventing ubiquitinylation and increasing their resistance to proteases. In addition, palmitoylation is believed to be an important factor in the process of trafficking proteins between organelles and in directing them to specific membrane compartments. For example, in neurons, palmitoylation targets proteins for transport to nerve terminals and may regulate trafficking at synapses; it is essential for the growth and integrity of neuronal axons and for conveying axonal signals. Protein S-palmitoylation is also a basic mechanism for control of the properties and functions of ion channels, both directly and indirectly via other signalling pathways. Thus, protein lipidation is a potent regulator of apoptotic signaling via activation of calcium channels in the plasma membrane and endoplasmic reticulum. Mitochondrial morphology and function is regulated by stearoylation of human transferrin receptor 1. Dysregulation of DHHC palmitoyl acyltransferases is known to be associated with many human diseases, including schizophrenia, X-linked mental retardation, Huntington's disease and cancer, but more knowledge of the individual enzymes is required before any can be seen as a target for drug development.

More generally, the saturated acyl moiety in palmitoylation facilitates transfer of proteins to lipid rafts, subdomains of the plasma membrane that are enriched in sphingolipids and cholesterol. In this situation, proteolipids can participate in cell signalling events, such as in T-cells of the immune system, where the unique feature of reversibility in lipid S-palmitoylation modifications is advantageous. For example in lipid rafts, Ras proteins are activated by extracellular stimuli and produce signals that lead to cell proliferation, differentiation and apoptosis. They undergo palmitoylation on the Golgi membrane, which enables them to be transported to the plasma membrane where they exert their signalling function. Eventually, the signal is attenuated by depalmitoylation and thence dissociation of the protein from the plasma membrane. Following recaptured by the Golgi, the protein can undergo a new sequence of plasma membrane targeting and signalling. S-palmitoylation of Ras proteins is also important in fungal pathogens.

In addition, palmitoylation is involved in lipoprotein metabolism. Lipoprotein particles containing apolipoprotein B (apo B), such as chylomicrons, very-low-density and low-density lipoproteins, are essential for the transport of triacylglycerols and cholesterol esters in plasma. It has been established that palmitoylation of apo B regulates the biogenesis of the nascent lipoprotein particles that contain this apolipoprotein and may regulate the amount available for lipid transport.

Scottish thistleNumerous S-acylated proteins are known to be present in higher plants, and many functions are now being reported although progress appears to have been slower than with mammalian systems. SNARE and G proteins and the cellulose synthase complex are among several proteins known to be modified in this way with effects on fertility, Ca2+ signalling, movement of potassium ions, stress signalling (e.g. the immune response), and the growth of root hairs and pollen tubes. For example, the 18 subunits of the cellulose synthase A family are multiply S-acylated (70–110 S-acyl groups), rendering it the most heavily S-acylated complex ever described. As this complex is integral to the plasma membrane and extrudes cellulose microfibrils into the extracellular environment to form the cell wall, the extrusion process is believed to enable the complex to pass through the plane of the plasma membrane to form microdomains and recruit accessory proteins for unhindered transport across the membrane. Although little appears to be known of the putative deacylation enzymes, regulatory functions of S-palmitoylation may be more important at the plasma membrane in plants than in animals.

Protein S-palmitoylation is critical for activity of many human pathogens, including fungal and bacterial infections. While parasitic protozoa and fungi possess their own palmitoyltransferases, viruses and bacteria hijack the enzymes in their hosts in order to favour their internalization, survival, and replication inside the cells. Palmitoylation of viral proteins is essential for their life cycle, and three types of membrane proteins in viruses, including many that are highly pathogenic, are known to be S-acylated, a factor that is important for the immune response. For example, palmitoylated 'spike' proteins are the main transmembrane proteins in the viral envelope, and they are involved in the entry of viruses into cells by catalysing receptor binding and/or membrane fusion. The viroporins are a second group, which are freely expressed in infected cells but not into virus particle per se to any appreciable extent. They possess one or two membrane-spanning regions, which amongst other functions serve as hydrophilic pores in membranes. A third diverse group of palmitoylated proteins produced by viruses are peripheral membrane proteins in which the fatty acid component simply anchors the modified protein to a membrane. In contrast to other proteins, glycoproteins of viral membrane in that they are palmitoylated at or near the cytoplasmic face and then remain palmitoylated.

Analysis by modern mass spectrometric methods in concert with a site-specific acyl-biotin-exchange reaction permits location of the acyl group to specific cysteine residues. In contrast to the N-acylated proteins, the fatty acids are easily released from the thiol linkage by base-catalysed transesterification for analysis by gas chromatography.

4.  Prenylated Proteins

Prenylated proteins are formed by attachment of isoprenoid lipid units, farnesyl (C15) or geranylgeranyl (C20), via cysteine thio-ether bonds, which are more stable than ester or thioester bonds, at or near the carboxyl terminus. Such proteins were first detected in fungi, but they are ubiquitous in mammalian cells where they can amount to up to 2% of the total proteins, and they are increasingly being found in other eukaryotic organisms. In addition, some proteins from pathogenic bacteria and viruses can be prenylated by their hosts, apparently as a protective measure. Isoprenylation is a stable (non-reversible) modification, which targets specific proteins to membranes and aids protein-protein interactions; it is essential for their functions. However, in contrast to the acylated proteins, the bulky branched nature of the lipid moiety of isoprenylated proteins ensures that the latter cannot be incorporated into ordered raft microdomains.

Formula of a prenylated protein

Whether a protein is prenylated is determined by specific amino acid sequence motifs at the carboxyl terminus, principally a CAAX sequence with cysteine (C) attached to two aliphatic amino acids (A) then to a variable carboxyl-terminal amino acid residue (X). The nature of the X residue determines whether a protein will be farnesylated or geranylgeranylated; with a few exception, proteins ending in serine, methionine, alanine or glutamine are farnesylated, while proteins ending in leucine or isoleucine are geranylgeranylated. A single farnesylation is usual, but some dual geranylgeranylation can occur. Proteins involved in cellular signalling and trafficking pathways are most involved, and the most important of these are probably the Ras super-family (low-molecular weight G-proteins, or guanosine 5’-triphosphate (GTP) hydrolases), which act as molecular switches for many different signal pathways including those controlling cell proliferation, adhesion, apoptosis and migration, and the integrity of the cytoskeleton.

The isoprenoid units are produced by the mevalonate pathway, as discussed on our web page dealing with cholesterol with farnesyl pyrophosphate as the branch-point in sterol/isoprene synthesis. Subsequent biosynthesis of prenylated proteins involves a concerted series of reactions in which the proteins are transported through various cellular organelles, ending mainly but not only at the plasma membrane. Prenylation occurs in the cytoplasm of the cell after synthesis of the protein per se, with farnesyl or geranylgeranyl pyrophosphate as the isoprenoid substrate, each catalysed by its own transferases, i.e. protein farnesyltransferase and protein geranylgeranyl transferase (two types GGT-1 and 2), respectively. The enzymes transfer the isoprenoid group to the cysteine residue in the CAAX box. Cleavage of the terminal tri-peptide (AAX) then occurs in the endoplasmic reticulum via a specific protease, before the new terminal cysteine is enzymically methylated at the carboxyl group with S-adenosyl methionine as the methyl donor (further increasing the hydrophobicity of the proteolipid). GGT-2 is known to transfer two geranylgeranyl units to the C-terminal double-cysteine motif (CC or CXC) of the 'Rab' family of proteins.

Prenylation of proteins

Prenylation alone tends to target proteins to endomembranes, such as the endoplasmic reticulum and Golgi. However, S-acylation (palmitoylation) of prenylated proteins can occur also, increasing the affinity for membranes and then proteins are directed from the endomembranes to the plasma membrane. As the second modification is reversible, it may function as part of a control mechanism. After they have been fully processed, these proteolipids have a high affinity for cellular membranes and possess a unique structure at their carboxyl termini, which functions as a specific recognition motif in some protein-protein interactions.

The "Ras" proteins in mammalian cells are farnesylated, while a subfamily of "Rho" proteins are usually geranylgeranylated. Prenylated proteins of the Ras family are attached to the cytoplasmic face of the cell membranes, where they transfer the signal from surface receptors to transcription factors that effect gene expression in the nucleus. An additional palmitoylation is required to bind to membranes and aids the transfer of Ras proteins from the Golgi and thence to the plasma membrane. Rho proteins can be anchored to the plasma membrane, endomembranes or endosomes by geranylgeranylation alone. Although the prenylation reaction is irreversible, some geranylgeranylated Rho proteins can be removed from membranes and in effect de-activated by binding via specific protein-protein interactions; prenylated proteins can also be trafficked between membranes by this means. However, binding between prenylated and non-prenylated proteins can serve to increase the activity of the latter, and many examples of this have now been documented.

Ultimately, degradation of prenylated proteins occurs in the lysosomal compartment of the cell and is catalysed by a prenylcysteine lyase, which is a flavin-containing monooxygenase that converts prenylcysteine to prenyl aldehyde by a novel mechanism.

As many of these proteins are involved in the development of cancer, they are the subject of much pharmaceutical interest, focusing especially on the inhibition of the prenylation reaction; a farnesyltransferase inhibitor for patients with acute myeloid leukemia is undergoing clinical trials. Defects in prenylation or its regulation have also been implicated in cardiovascular disease, neurodegenerative disorders and metabolic diseases. In addition, inhibitors of protein farnesyltransferase have been shown to be efficacious in the treatment of protozoal pathogens and other parasitic diseases in animal models, and they appear to be of value in the treatment of viral (e.g. hepatitis D) and fungal infections.

In plants, protein prenylation is required for plant growth, development and environmental responses, including the control of abscisic acid and auxin signalling and for meristem development. For example, at least 250 proteins in Arabidopsis have the CAAX sequence and have the potential to be prenylated, although this has been demonstrated experimentally for only a few. It appears that some prenylated proteins do not use the prenyl group to bind to membranes, but rather they may use it as a signal for interaction with other proteins, for example.

5.  O-Acylated Proteins/Peptides

Formula of ghrelinIn a few proteolipids, serine or threonine residues are acylated such that an O-acyl rather than an S-acyl linkage is formed. The best characterized example is ghrelin, a circulating 28-amino acid peptide hormone, which is octanoylated (C8) at a serine residue (third amino acid from the N-terminus). Ghrelin is of particular importance as a hunger-stimulating hormone produced in the human stomach and pancreas, increasing food intake and adiposity. Among many other functions that are now known, it is a potent stimulator of growth hormone from the anterior pituitary gland. Only the octanoylated protein binds to the single receptor characterized to date, i.e. growth hormone secretagogue receptor GHSR1a, and has biological activity. The ghrelin O-acyltransferase (GOAT), a member of the MBOAT family of acyltransferases that is widely involved in glycerolipid metabolism, is the enzyme that catalyses octanoylation of desacyl-ghrelin and is expressed mainly in the stomach and intestines. It is now perceived as a target for pharmaceutical intervention in the treatment of the metabolic syndrome.

The family of 'Wingless' or Wnt proteins are central mediators of embryonic development and tissue renewal in animals that influence cell proliferation, differentiation and migration and require acylation for secretion and activity. In murine Wnt3a, the most intensively studied form, there is an unusual O-acyl modification with palmitoleic acid (9-cis-hexadecenoic or 9-16:1), which has been termed a lipokine (i.e. a lipid hormone), at a conserved serine residue (but with no S-palmitoylation on a conserved cysteine residue as originally reported). All Wtn proteins, have a conserved signal sequence, which is cleaved on entry into the endoplasmic reticulum before glycosylation and fatty acylation. Following desaturation of palmitoyl-CoA to palmitoleoyl-CoA by stearoyl-CoA desaturase in the endoplasmic reticulum, the subsequent O-acylation requires an O-acyltransferase termed 'porcupine' (PORCN), another MBOAT family member, which is entirely distinct from the S-acyltransferases and is essential for intracellular trafficking and activity of Wnt proteins; one inhibitor of this enzyme is undergoing clinical trials for treatment of Wnt-driven solid tumors. The palmitoleate residue is crucial for maintaining Wnt structural integrity and assists Wnt secretion by facilitating Wnt movement from the trans Golgi network to the plasma membrane; it also binds within a hydrophobic cleft of the Wnt receptor to promote ligand-receptor interactions. Serine acylation may be important for extracellular long-range transport of Wtn proteins in lipoprotein particles. Wnt proteins are deactivated by an extracellular carboxylesterase ('Notum'), which removes the essential palmitoleic acid group.

In addition, an acyl-CoA:lysophosphatidylcholine acyltransferase I (LPCAT1), which also generates the pulmonary surfactant dipalmitoylphosphatidylcholine from lysophosphatidylcholine, catalyses O-palmitoylation of a serine residue on the protein histone H4, presumably as a means of regulating mRNA synthesis that may lead to changes in the global transcriptional activity of the cell. A rare O-acylation mechanism is present also in bacteria of the order Corynebacteriales (see below).

6.  Hedgehog Proteins linked Covalently to Cholesterol and Palmitate

Hedgehog proteins have a major role in signalling during the differentiation of cells in the development of all embryos from Drosophila to fish to humans. They were first found and studied in the insect model Drosophila melanogaster (and named for an anomalous cuticular feature reminiscent of a hedgehog's spines in a mutant), but they are now known to occur in all higher organisms. Vertebrates, for example, express three hedgehog family proteins designated 'Sonic' (Shh), 'Indian' (Ihh) and 'Desert' (Dhh) hedgehog, of which Shh is most studied. They are required for an extensive range of processes, from the control of left-right asymmetry of the body to the specification of individual cell types within the brain and to limb development. Aberrant expression and/or signalling of the Shh hedgehog proteins have been implicated in morphological abnormalities and the generation of many human cancers. A distinctive feature is that they contain cholesterol in covalent linkage, further confirmation if needed of the vital importance of this lipid in animal tissues. Proteins that are functionally analogous but structurally distinct are found in nematodes.

Formula of a 'hedgehog' signalling protein

Hedgehog proteolipids are formed post-translationally by attachment of cholesterol via an ester bond to glycine at the C-terminus, a highly conserved region of the protein, while a palmitoyl moiety is attached to a cysteine residue at the N-terminus (N-palmitoylation). The signalling proteins are synthesised initially as 46 kDa inactive propeptides, with two distinct domains, i.e. an N-terminal 'hedge' domain, which is proteolytically cleaved to a 19 kDa fragment (Hh-N) by means of the C-terminal "hog" domain in the endoplasmic reticulum. The C-terminus acts also as a cholesterol transferase to attach a cholesterol molecule covalently to the carboxyl end of the Hh amino terminal fragment, Hh-N. Finally, the nascent Hh-N is further modified by the addition of an N-palmitoyl group at Cys-24, a reaction mediated by a palmitoylacyltransferase (Hedgehog acyltransferase or Hhat, another MBOAT family member) to create a highly hydrophobic molecule that is often described as 'Hh-Np for Hh-N-processed'.

A hedgehog protein in the outer leaflet of a membraneUnlike the other lipid-modified proteins discussed above, but like the GPI-anchored proteins, the lipid moieties are located in the exoplasmic or exterior leaflet of membranes with the protein component in the extracellular region. The process is regulated both positively and negatively by various oxy-cholesterol derivatives including vitamin D3, possibly by competing for binding with sterol-sensing domains (similar to those in proteins involved in cholesterol homeostasis) in receptor and other proteins, and the interactions between sterol metabolism and hedgehog signalling are increasingly a focus for research.

Following biosynthesis in the endoplasmic reticulum and Golgi, there are mechanisms to transport the final proteolipid through the membranes and onwards to other cells often for an appreciable distance from the site of synthesis, but much remains to be learned of how this is accomplished. In Drosophila and vertebrates, a protein termed 'dispatched' (DISP), which contains a sterol-sensing domain, is required for transport across the membrane, while extracellular release and subsequent distribution of cholesterol-modified Shh is enhanced by its interaction with the secreted protein SCUBE2; the cholesterol component is essential for this purpose. Cholesterol-modified Shh is also shed from the surface of producing cells in exovesicles or “exosomes” derived from the budding of cellular membranes. In insect models, these proteolipids are transported in the form of lipoprotein complexes (lipophorins) with the lipid moieties in a phospholipid monolayer that surrounds a core of triacylglycerols and cholesterol esters.

Both lipid components are essential for the proper tissue distribution and function of the attached proteins, and the N-palmitoyl moiety in particular is required to cause the proteins to form multimeric complexes essential for biological activity. The cholesterol modification contributes to the partitioning of hedgehog proteins into plasma lipoprotein complexes for long-range transport. The mammalian Shh receptor termed 'PTCH1' is a transmembrane protein, which also contains a sterol-sensing domain and is located on the primary cilium, an antennae-like projection of the plasma membrane into the cytoplasm of the cell. When PTCH1 binds to hedgehog proteins, a second protein designated SMO is activated to enter the primary cilium and initiates down-stream signalling. While the protein moiety lacking cholesterol maintains some of the signalling capacity, loss of palmitoylation abolishes the signalling activity entirely. Inhibition of Hhat may be of therapeutic value against pancreatic cancer.

7.  Bacterial Proteolipids

Formula of a bacterial proteolipidAll bacteria contain large numbers of proteins with a unique and distinctive post-translational lipid modification (more than 2000 have been identified), and this appears to be essential for their efficient function and even for their pathogenesis via host-pathogen interactions. The lipid components consist of N-acyl- and S‑diacylglycerol groups attached to an N-terminal cysteine, i.e. it contains a thio ether bond. In Mycobacterium bovis, for example, positions sn-1 and 2 of the glycerol moiety are linked to palmitic and tuberculostearic acids, respectively, and either fatty acid can be the N-acyl moiety. Proteomic analysis of Staphylococcus aureus revealed 63 different proteolipids of this type.

As with other proteolipids, the lipid moieties act as an anchor to hold the protein tightly to a hydrophobic cellular membrane while permitting it to operate in an aqueous environment in such important activities as transport, signalling, adhesion, digestion and growth; they have an important role in nutrient and ion acquisition, enabling pathogenic species to better survive in the host. They are important constituents of the outer leaflet of the cytoplasmic membrane of Gram-positive bacteria and of the outer leaflet of the cytoplasmic and the inner leaflet of the outer membranes of Gram-negative bacteria. Like the endotoxins (lipopolysaccharides) of Gram-negative bacteria, they are potent stimulants of the human immune system, eliciting pro-inflammatory immune responses by functioning as ligands for specific receptors, especially the Toll-like receptor 2. They are thus responsible for much of the virulence of the organisms and have the potential to be used in vaccines.

The first of these to be discovered termed "Braun's lipoprotein" is one of the most abundant membrane proteins in Gram-negative cell walls. It has a molecular weight of about 7.2 kDa, and it is unique in that much of it is bound at its C-terminal end by a covalent bond to the peptidoglycan layer and is embedded in the outer membrane by its hydrophobic head so that it provides a tight link between the two layers, providing structural integrity to the outer membrane.

In the main secretory pathway, proteins destined to become lipidated have N-terminal signal peptides containing a motif known as a lipobox with an invariant cysteine residue, which directs them to the lipoprotein biogenesis machinery after transport mainly in an unfolded state. The three fatty acyl groups and the glycerol component responsible for binding to the membrane surface are derived from bacterial phospholipids, especially phosphatidylglycerol. Three enzymes are involved in the biosynthetic pathway, which occurs in the cytoplasmic (inner) membrane. The first (Lgt; phosphatidylglycerol:prolipoprotein diacylglyceryl transferase) attaches the diacylglycerol group from phosphatidylglycerol to the thiol of cysteine, the first amino acid after a signal peptide in the pro-lipoprotein. In E. coli, for example, there is an 18- to 36-amino-acid-long signal peptide, which is distinguished by a C-terminal lipobox comprising a conserved three-amino-acid sequence in front of an invariable cysteine.

Proteolipid biosynthesis in bacteria

A second enzyme (Lsp; prolipoprotein signal peptidase) then removes the signal peptide, leaving the cysteine as the new amino-terminal residue of the protein component. The third enzyme (Lnt; apolipoprotein N-acyltransferase) acylates the N-terminal amine group of the modified cysteine with a fatty acid from position sn-1 of whatever phospholipid is available (the resulting lysophospholipid is flipped back across the membrane and re-esterified). This last step always occurs in Gram-negative bacteria, but is found only rarely in Gram-positive bacteria (Lgt and Lsp are essential to all bacteria). However, other species have related Lnt-like enzymes that can acylate, or alternatively acetylate or add a peptide unit to the N-terminal cysteinyl residue. Most of the proteolipids are then transferred to the inner leaflet of the outer membrane by a complex mechanism involving five proteins (Lol pathway), which sort and translocate them via specific signal residues located C-terminally to the diacylglyceryl-cysteine. In the outer membrane, proteolipids undergo topology changes that govern the biogenesis and integrity of the membrane.

In Gram-positive bacteria, lipoprotein maturation and processing are not vital to the organism, but they are essential to their pathogenicity. It is now evident that the degree of proteolipid acylation in these species has a substantial influence on the immune response. Thus, exposure of the skin to diacylated proteolipid induces immune suppression, while exposure to triacylated proteolipid does not. In addition, the fatty acid composition of the lipid moiety influences the pro-inflammatory response.

Bacteria of the genus Mycoplasma lack a cell wall and are obligate parasites that must obtain all their lipids from the host. Recently, it has been demonstrated that otherwise cytoplasmic proteins, lacking signal peptides, are tethered to the outer membrane by a link from glutamine near the C-terminus of the protein to rhamnose and thence to a phospholipid, presumed for the moment to be phosphatidic acid. Whether other bacteria have a similar mechanism has yet to be determined.

An unusual post-translation O-acyl modification of specific proteins by mycolic acids is one means of targeting them for assembly in the outer membrane (mycomembrane) in bacteria of the order Corynebacteriales; a short linear amino acid motif for O-acylation of proteins has been revealed that seems to be preserved throughout the kingdoms of life.

Lipopeptides: A number of bacterial species produce lipopeptides of which the best known are probably the glycopeptidolipids from Mycobacteria and surfactin and related molecules from Bacillus subtilis. These are discussed in a separate web page on this site.

Note that the terms ‘lipoprotein’, ‘lipopeptide’ and ‘proteolipid’ are used interchangeably for these compounds in the literature. As discussed in the Introduction to this document to avoid confusion, I prefer to reserve the term ‘lipoprotein’ for the non-covalently linked protein-lipid complexes in plasma.

8.  Other Proteolipids

For practical reasons, the GPI-anchored proteins are discussed elsewhere in these web pages, as are the ceramides and related lipids bound to proteins in skin. In addition, N-terminal acetylation of certain membrane proteins targets them for transfer to the Golgi or lysosomes. In yeasts, a covalent conjugate of phosphatidylethanolamine with a protein designated ‘Atg8’ is involved in the process of autophagy (controlled degradation of cellular components) by promoting the formation of membrane vesicles containing the components to be degraded.

Recommended Reading

Further information on bacterial proteolipids is available at a dedicated web site - - and there is another website that deals with S-palmitoylation. An S-palmitoylation database is accessible via Blanc, M., David, F., Abrami, L., Migliozzi, D., Armand, F., Bürgi, J. and van der Goot, F.G. SwissPalm: protein palmitoylation database. F1000Res., 4, 261 (2015);  DOI.

Lipid listings Credits/disclaimer Updated: March 16th, 2018 Author: William W. Christie LipidWeb icon