Functional diversity in a lipidome

Lipids are complex molecules generated by cells through enzymatic mechanisms from simpler constituents. Each complex lipid typically consists of a head group with a unique chemical composition that is esterified to hydrophobic tails composed of fatty acyl chains or sphingoid bases. Based on their composition, lipids are classified into eight major categories (fatty acyls, glycerolipids, glycerophospholipids, sphingolipids, sterols, prenols, glycolipids or saccharolipids, and polyketides), each of which is subdivided into classes and subclasses (1). The biological functions of a lipid class are widely regarded as defined by the lipid head group. For example, phosphatidylinositols are a lipid class defined by the presence of an inositol head group, and their cellular function as signaling molecules is critically dependent on the inositol head group and modifications thereof. However, for each lipid class, there are also a number of unique molecular species, all of which have identical head groups but vary with respect to the number of carbon atoms in the acyl chains, the number of double bonds, and also the nature of the chemical linkage (ester or ether) of these chains to the head group. For example, in mammalian tissues or cells, 30 to 35 species of a specific lipid class can be detected and quantified (2), and with the advent of highly sensitive mass spectrometry, the bewildering complexity of lipid species within a single class has become apparent. Taken together across all lipid classes, mammalian cells may have thousands of individual molecular species of lipids; these have collectively been referred to as the lipidome (3).

The presence of such a lipidome is a feature of all cell types and common model organisms examined to date, including mammals, Drosophila melanogaster, Caenorhabditis elegans, and yeast, and thus appears to be conserved in evolution. However, the repertoire of lipid molecular species varies between organisms. A range of molecular species is also seen for a given lipid class both in primary cells (i.e., directly isolated from an organism) as well as cells maintained continuously in culture, although the repertoire of molecular species is distinct in these situations (4). These observations have raised a number of intriguing questions: 1) Do these individual species have distinct biochemical functions in cells? 2) What is the molecular mechanism through which individual molecular species of a given lipid influence cellular function? 3) What determines the range of molecular species for a given lipid class? In PNAS, Schuhmacher et al. (5) address some of these questions in the context of diacylglycerol (DAG), a class of glycerolipid with well-established functions in cells.

Like most other lipid classes, cells contain many molecular species of DAG that differ from each other with respect to the composition of the fatty acyl chains in the tail of the lipid. DAGs are produced by multiple enzymatic pathways; they are a key intermediate in phospholipid biosynthesis but are also a well-studied signaling intermediate in eukaryotic cells. Many cells activate phospholipase C (PLC) as a mechanism of signaling. PLC generates DAG and inositol 1,4, 5 trisphosphate (IP₃) by the hydrolysis of the signaling lipid phosphatidylinositol 4,5-bisphosphate (PIP₂). Whereas IP₃ releases Ca²⁺ from intracellular stores (6), DAG activates protein kinase C (PKC), leading to further biochemical changes and cellular effects (7). Like other lipid classes, phosphatidylinositol, the precursor of PIP₂, is also found as multiple molecular species, and therefore this is also the case with PIP₂. When mammalian cells are stimulated with agonists for G-protein–coupled receptors, multiple species of DAGs are produced with specific temporal kinetics. These include DAG species that are highly unsaturated as well as monounsaturated DAG species (reviewed in ref. 8). Previous studies have attempted to test the efficacy of these DAG species to activate PKC (9) and found that, in vitro, polyunsaturated DAGs are far more effective that monounsaturated species. However, the relevance of this finding to signaling activity within intact cells has remained unclear.

In their study, Schuhmacher et al. (5) use the tools of chemical biology to synthesize individual molecular species of DAG including stearyl arachidonyl DAG (SAG) as well as dioctanoyl DAG (DOG) (Fig. 1). SAG is the primary species of DAG produced after PLC activation in mammalian cells. The DAG species were also functionalized with additional chemical groups to allow them to be loaded onto intact cells, their spatial distribution visualized using fluorescence microscopy, and a UV-sensitive chemical cage allowed them to be released and made available to activate PKC at the plasma membrane with temporal precision. During cell signaling, the production of DAG at the plasma membrane leads to the recruitment of PKC through the interaction of its C1 domain with this lipid, a key event in PKC activation. To observe PKC recruitment in intact cells, the authors used a version of PKC tagged to green fluorescent protein and monitored its translocation to the plasma membrane. This combination of experimental features allows a precise assessment of the ability of individual species of DAG on the activation of PKC in space and time.

Fig. 1.

Chemical structure of two individual molecular species of DAG: (A) 18:0 | 20:4 DAG (1-stearoyl-2-arachidonoyl-sn-glycerol) and (B) 8:0 | 8:0 DAG (1,2-dioctanoyl-sn-glycerol). The acyl chain 18:0 is in green, 20:4 is in red, and 8:0 is in pink. The common head group (i.e., the glycerol moiety) is shown in blue.

Schuhmacher et al. (5) find that individual molecular species of DAG, when generated at the plasma membrane through photolysis of caged precursor probes, differ in their ability to recruit PKC to the plasma membrane. While SAG was most effective in PKC recruitment to the plasma membrane, DOG was the least effective. The individual isoforms of PKC (i.e., PKC α, β, γ, δ, and ε) also differed in their ability to translocate to the plasma membrane in response to uncaging of individual species of DAG. The ability of individual DAG species to recruit PKC to the plasma membrane was well correlated with their ability to modulate downstream outputs of PKC signaling (such as c-Raf and GSK3β) under equivalent conditions. Together, these findings offer compelling evidence that, in intact cells, individual molecular species of DAG differ in their ability to relay information during cell signaling.

These findings have a number of important implications. First, they imply that the biochemical activity of a lipid class is not merely a function of the head group that defines that class. Clearly, as Schuhmacher et al. (5) find, the hydrophobic tail of a lipid, acyl chains in the case of DAG, although embedded in the lipid bilayer, are able to influence the ability of its head group to interact with protein ligands recruited to the membrane. The molecular mechanism underlying this recognition of acyl chain length and unsaturation by protein ligands remains to be understood. Furthermore, this observation has implications for the common approach of using the interaction of a lipid head group with its fluorescently tagged binding domain as a reporter of lipid localization in intact cells. Indeed, Schuhmacher et al. (5) report that the ability of specific molecular species of DAG to recruit full-length PKC is not well correlated with their ability to recruit the isolated C1 domain of PKC (10). A second practical implication of these findings arises from the use of short-chain lipids to test the ability of a specific lipid class to stimulate protein activity in cell physiology experiments. Typically, short acyl chain analogs of lipids (such as DOG) are used in preference to the longer chain, unsaturated species (such as SAG) that have poor solubility and are difficult to handle in aqueous buffers used for cell physiology experiments. The interpretation of the results of such experiments will need to be reconsidered, especially in scenarios in which cell physiology is not modulated by the addition of DOG.

The findings of Schuhmacher et al. (5) raise questions on whether the selective activation of PKC by SAG is an exceptional finding. This may not be the case, because previous studies have reported additional enzymes in the PLC-triggered cycle of lipid metabolism that also show a preference for SAG. These include diacylglycerol kinase ε, which appears to prefer SAG as a substrate; phosphatidylinositol 4-phosphate 5-kinase, which appears to prefer 1-stearyl 2-arachidonyl PI4P (reviewed in ref. 11); and CDS2, which shows strong selectivity for 1-stearoyl-2-arachidonoyl-sn-phosphatidic acid as substrate (12). Although these are all enzymes and therefore expected to have substrate specificity, it is important to note that the specificity in question is at the level of the acyl chain composition of the respective substrate. Similarly, the binding of specific classes of lipids by lipid transfer proteins also appears to show acyl chain specificity (13, 14). Thus, it might be that some of the numerous molecular species of any given lipid class do indeed bind selectively to protein targets in cells.

Although the results of Schuhmacher et al. (5) demonstrate the function of SAG as an effective activator of PKC isoforms, they also highlight the need to identify functions for the other molecular species of DAG. Why are these additional molecular species present in cells, and what determines the proportion in which they are found? One possibility is that these species also selectively bind to specific, as-yet-unidentified proteins; identification of such proteins will likely aid in assigning functions to these species. A second possibility is that these additional species with varying degrees of acyl chain length and saturation do not have direct protein ligands, but by being present in fixed proportions they determine the physicochemical properties of the membrane, for example, locally in microdomains (15). Finally, because the properties of acyl chains will likely influence the lateral diffusion of lipid molecules within a leaflet of the bilayer, the acyl chain composition of a particular lipid (e.g., SAG) may also influence the spatial domain of influence of a protein (such as PKC) once it has been recruited to the membrane. SAG is principally produced by the activation of PLC at the plasma membrane, and the experiments of Schuhmacher et al. (5) elegantly describe its activity at this location. However, the subcellular distribution of the numerous other species of DAG, which have been described by total lipid extraction from cells, remains unknown. Such information coupled with the methods described by Schuhmacher et al. (5) will likely aid in understanding the functional diversity in a lipidome.

Although most lipid classes in mammalian tissues show a complex repertoire of molecular species, phosphatidylinositols are an exception to this rule. In primary cells from mammalian tissues, phosphatidylinositols are substantially enriched in 1-stearyl 2-arachidonyl acyl chains (e.g., see refs. 2 and 4), and the signaling lipids derived from PI also show this enrichment. It appears that this is not a consequence of a biosynthetic machinery that only generated 18:0/20:4 PI; rather, PI is synthesized as a heterogenous set of molecular species with diverse acyl chains, but this is subsequently remodeled to enrich for 18:0/20:4 PI (16). This remodeling is thought to occur through the Lands’ cycle (17), the sequential activity of phospholipase A2 followed by the addition of arachidonic acid at sn2 through the activity of an acyltransferase; the identification of enzymes encoding this activity is a subject of ongoing analysis (4, 18, 19). The identification of these enzymes and the ability to broaden the repertoire of molecular species of PI may also provide an alternative approach to understand the requirement of a defined repertoire of molecular species for a given lipid class in cellular function.