Twenty years ago, N-acylethanolamines (NAEs) were considered by many lipid chemists to be biological ‘artifacts’ of tissue damage, and were, at best, thought to be minor lipohilic constituents of various organisms. However, that changed dramatically in 1993, when anandamide, an NAE of arachidonic acid (N-arachidonylethanolamine), was shown to bind to the human cannabinoid receptor (CB1) and activate intracellular signal cascades in mammalian neurons. Now NAEs of various types have been identified in diverse multicellular organisms, in which they display profound biological effects. Although targets of NAEs are still being uncovered, and probably vary among eukaryotic species, there appears to be remarkable conservation of the machinery that metabolizes these bioactive fatty acid conjugates of ethanolamine. This review focuses on the metabolism and functions of NAEs in higher plants, with specific reference to the formation, hydrolysis and oxidation of these potent lipid mediators. The discussion centers mostly on early seedling growth and development, for which NAE metabolism has received the most attention, but also considers other areas of plant development in which NAE metabolism has been implicated. Where appropriate, we indicate cross-kingdom conservation in NAE metabolic pathways and metabolites, and suggest areas where opportunities for further investigation appear most pressing.
N-acylethanolamines (NAEs) are a family of functionally diverse signaling lipids. They consist of a fatty acid linked by an amide bond to ethanolamine, and are classified based on the number of carbons and degree of saturation of their acyl chain (Figure 1). Among the NAEs, N-arachidonylethanolamine or anandamide (NAE 20:4) has been the most widely studied primarily because of its role as an endogenous ligand of plasma membrane-localized cannabinoid (CB1 and CB2) receptors (Felder et al., 1993). Upon receptor binding, anandamide triggers a series of signaling events that regulate multiple physiological and behavioral processes in animals, including pain sensation, energy balance, fear, anxiety, appetite and memory (Alger, 2004). This so-called ‘endocannabinoid signaling pathway’ is now known to involve a number of canonical signaling components, such as G-proteins, calcium and potassium channels, adenylate cyclase and mitogen-activated protein kinases (Fonseca et al., 2013). In fact, Δ-9-tetrahydrocannabinol, the active ingredient of marijuana, mimics the activity of anandamide by binding to CB1 receptors in the brain to exert its strong psychotropic effects on humans (Devane et al., 1992). Other more abundant types of NAEs (e.g. ethanolamides of oleic, linoleic, linolenic and palmitic acids, Figure 1) have also received attention as lipid mediators that mostly function independently of the G-protein-coupled cannabinoid receptors (Lo Verme et al., 2005; Movahed et al., 2005).
Although most research on NAEs has focused on vertebrates, evidence for their signaling role in more primitive organisms has expanded in recent years. For instance, these small lipids have been implicated as metabolic signals that coordinate nutrient status and lifespan determination in Caenorhabditis elegans (Lucanic et al., 2011), findings that may have important implications for biomedical research on aging and obesity (De Petrocellis and Di Marzo, 2011). NAEs have also been detected in organisms as diverse as yeast (Saccharomyces cerevisiae), fresh water fish (Esox lucius and Cyprinus carpio), bivalve mollusc (Mytilus galloprovincialis), protists (Tetrahymena thermophila) and slime mold (Dictyostelium discoideum). (Natarajan et al., 1985; Sepe et al., 1998; Merkel et al., 2005; Anagnostopoulos et al., 2010; Hayes et al., 2013). In fact, some of these organisms appear to regulate their endogenous NAE levels via similar enzymatic machinery as mammalian vertebrates (Karava et al., 2005; Lucanic et al., 2011; Hayes et al., 2013). However, unlike in vertebrates, the perception, biosynthesis and molecular targets of NAEs are not yet known in other organisms. Nonetheless, the widespread occurrence of NAEs, from simple organisms to humans, suggests a highly conserved role for this group of lipids in cell signaling.
This review discusses the current state of knowledge of NAE metabolism and function in higher plants, and highlights recent experimental evidence showing that these fatty acid amides function as endogenous lipid signals that modulate plant physiological processes as diverse as seed germination, seedling establishment, plant–pathogen interactions, chloroplast development and flowering. The focus is on Arabidopsis thaliana, because the rapid progress on plant NAE research that we have witnessed in the past decade was made possible using tools developed for this model plant. For example, availability of the complete sequence of the Arabidopsis genome led to identification and functional characterization of the plant enzyme homologous to the mammalian amidohydrolase that degrades NAEs in vivo (Shrestha et al., 2003; Wang et al., 2006b). Evidence indicates that NAEs and their oxidation products influence events that define the transition from seed germination to post-germinative seedling growth. However, questions remain at the mechanistic level as to how these lipid metabolites mediate their potent actions. We hope that this review will generate enhanced interest within the plant scientific community regarding this under-studied group of small lipid mediators.
Biological Implications of NAE Occurrence in Plants
As in animals, NAEs in plants represent a small fraction of the total lipid content, and their acyl composition generally reflects the acyl groups found in membrane lipids from the same tissue (Chapman et al., 1999; Chapman, 2004; Coulon et al., 2012a; Kilaru et al., 2012). NAE content and composition vary from tissue to tissue, between plant species and with physiological conditions (Chapman et al., 1999; Tripathy et al., 1999; Venables et al., 2005). Among plant tissues surveyed, NAE content was reported to be highest in desiccated seeds of various plant species, although amounts are generally at the low parts per million (ppm) level (Chapman et al., 1999; Chapman, 2004; Venables et al., 2005; Kilaru et al., 2007). In seeds, endogenous NAE levels range from 0.5 μg g fresh weight−1 for Pisum sativum, to 35.0 μg g fresh weight−1 for Medicago truncatula (Kilaru et al., 2007; Table 1). Overall, NAEs in seeds consist mostly of NAE 18:2, NAE 16:0 and NAE 18:1, with NAE 18:2 being the most abundant of the three (Chapman et al., 1999, 2003; Chapman, 2004; Venables et al., 2005). Other types, such as NAE 12:0, NAE 14:0, NAE 18:0 and NAE 18:3, are also present but at lower amounts, and may vary from species to species (Chapman et al., 1999; Venables et al., 2005). The NAE composition in seeds reflects the acyl content of their metabolic precursor, N-acylphosphatidylethanolamines (NAPE; Coulon et al., 2012b; Kilaru et al., 2012), rather than the composition of the acyl moieties of the triacylglycerols stored in seeds. Most importantly, NAE levels in seeds decrease markedly with germination and seedling establishment (Chapman et al., 1999; Kilaru et al., 2012). As discussed below, there is accumulating evidence that this decrease in NAE levels is physiologically relevant for the embryo-to-seedling transition through interaction with abscisic acid (ABA) signaling pathways (Figure 2; Teaster et al., 2007; Keereetaweep et al., 2013).
Table 1. Endogenous levels and biological significance of selected NAE species in plants
Plant organ/tissue detected, and endogenous levels (μg g fresh weight−1)
Venables et al. (2005); Keereetaweep et al. (2013)
The NAE types and metabolites in seeds and seedlings are similar to those in animal tissues with the exception of anandamide, which has not been reported in plants. Other so-called receptor-inactive NAE types are much more prevalent than anandamide in most animal tissues, and these more abundant types match with those that are most abundant in plants (e.g. NAE 18:2, NAE 18:1 and NAE 16:0; Kilaru et al., 2010, 2011, 2012; Zoerner et al., 2011). In animals, these more abundant NAEs compete with the same enzymatic machinery that metabolizes anandamide, and have been reported to have an ‘entourage effect’ on processes regulated by anandamide (Smart et al., 2002). Seeds and seed fractions represent a natural dietary source of NAEs, which may compete with endogenous metabolites in the endocannabinoid signaling system of mammals and influence various physiological processes (Coburn et al., 1954; Berdyshev, 2000). These dietary sources of NAEs were recently shown to be perceived in the intestine and signal satiety to the brain through dopamine receptors (Tellez et al., 2013). Manipulation of NAE content in seeds or seed fractions may have nutritional or therapeutic applications, but this area has received little attention.
In vegetative tissues, the steady-state NAE content and composition are markedly different than NAE profiles of seeds, even within the same species. These differences in composition appear upon seed germination and seedling growth, such that composition of NAEs in young seedlings is similar to those in leaves of mature plants (Wang et al., 2006b; Kang et al., 2008). The marked changes that occur upon seedling establishment suggest that NAE metabolism is influenced by developmental changes in plants, and the link between NAE metabolism and composition to seedling growth has been studied extensively (Blancaflor et al., 2003; Motes et al., 2005; Teaster et al., 2007; Keereetaweep et al., 2010; Cotter et al., 2011; Kilaru et al., 2011, 2012). In seeds, the proportion of polyunsaturated NAEs decreases more substantially than saturated NAE species, and, as discussed below, there is evidence that oxidative metabolism may contribute significantly to the overall changes in NAE profiles during seedling growth.
NAE occurrence also has been assessed in cultured tobacco cells. It was through studies with cultured cells that NAEs were first implicated in plant–pathogen interactions. For instance, both NAE 12:0 and NAE 14:0 were detected at very low levels in tobacco cell cultures by HPLC and GC/MS (Chapman et al., 1998). The level of these NAEs (particularly NAE 14:0) increased up to 50-fold in the presence of fungal elicitors in both cell cultures and leaves of tobacco plants (Chapman et al., 1998; Tripathy et al., 1999). This increase in NAE 14:0 was associated with increased expression of defense genes such as phenylalanine ammonia lyase and inhibition of medium alkalinization, processes that are both involved in plant responses to pathogens (Tripathy et al., 1999). With the identification of NAEs in various plant organs and plant species including the model plant Arabidopsis, the biological roles of these lipid mediators are likely to expand beyond what is summarized in Figure 2 and Table 1. Some of these biological processes are discussed in more detail in the light of accumulating evidence that the metabolism of NAE, particularly the machinery for its degradation and oxidation, appears to be tightly linked to its in vivo function(s).
NAE Formation: Does the ‘Transacylation Phosphodiesterase’ Pathway Operate in Plants?
In mammalian systems, NAEs are synthesized and released ‘on demand’ from its membrane precursor, NAPE, via multiple pathways (summarized in Figure 3 and Table 2). NAEs in animals may be generated by the hydrolysis of NAPE in a single enzymatic step, catalyzed by a phospholipase D (PLD; Schmid et al., 1983; Petersen and Hansen, 1999; Schmid, 2000), and a calcium-activated NAPE-specific phospholipase D (NAPE-PLD), a member of the metallo-β-lactamase family with a catalytically important Zn ion, has been characterized in mouse (Mus musculus) and subsequently other mammals (Okamoto et al., 2004, 2005; Wang et al., 2006a). The enzyme NAPE-PLD is structurally and catalytically distinguishable from all other known PLD family members, and shows high specificity for NAPE among various glycerophospholipids but showed no differential selectivity for particular N-acyl species of NAPE (Wang et al., 2006a).
Table 2. List of enzymes identified in vertebrates or plants (either in tissue extracts or by direct assays of purified protein activity) that are involved in NAE metabolism
The formation of NAEs from NAPE was also predicted to proceed through sequential deacylations, yielding lyso-NAPE and glycerophospho-NAE intermediates (Natarajan et al., 1984). Indeed, a soluble form of phospholipase A2 that was capable of converting NAPE to 2-lyso-NAPE in a calcium-independent manner (Sun et al., 2004), and α/β-hydrolase 4 (ABHD4), which acts on both NAPE or lyso-NAPE to generate glycerophospho-NAE, have been discovered (Simon and Cravatt, 2006). The subsequent cleavage of glycerophosphate to yield NAE is mediated by a metal-dependent phosphodiesterase (Simon and Cravatt, 2008). An alternative pathway involves phospholipase C-mediated production of an intermediate form of NAPE, a phospho-NAE species that is subsequently dephosphorylated to NAE by the action of the tyrosine phosphatase PTPN22 and/or SH2-containing inositol-5-phosphatase (Liu et al., 2006, 2008). Further studies illustrating a bacterial endotoxin-induced increase in anandamide in NAPE-PLD knockout mice, and the different time courses of activation of various NAE biosynthetic pathways, suggested that the ‘on demand’ synthesis of NAEs, which is rapid, may utilize the phospholipase C/phosphatase pathway, while the other pathways function predominantly in maintaining basal tissue levels (Liu et al., 2008).
Despite the identification of multiple pathways in mammalian systems by which NAPE is converted to NAE, limited progress has been made in plants. In Arabidopsis, although several PLDs have been identified (Wang, 2000), no NAPE-specific PLD has been characterized so far. Evidence for PLD-mediated NAE synthesis in plants is limited to in vitro studies, in which recombinant plant PLDβ/γ isoforms, expressed in Escherichia coli, were used to demonstrate their ability to catalyze formation of NAE from NAPE (Pappan et al., 1998). When Interpro (www.ebi.ac.uk/interpro/) was used to identify homologs to the mouse NAPE-PLD (UniProt ID Q8BH82) with the beta-lactamase domain, five proteins from Viridiplantae that were yet to be characterized were identified: A8JCX4 from Chlamydomonas reinhardtii, D8SOJ8 and D8SB87 from Selaginella moellendorffii, E1ZJ92 from Chlorella variabilis and F2E5V4 from Hordeum vulgare (Hunter et al., 2012). It was surprising that no proteins with a significant match to mouse or any mammalian NAPE-PLD were identified in Arabidopsis or other dicots despite the fact that at least 51 proteins in Arabidopsis are known to contain a β-lactamase-like domain, based on the Interpro protein sequence analysis and classification (www.ebi.ac.uk/interpro). Among the other characterized mammalian enzymes that are involved in NAE synthesis, several plant homologs have been identified for mouse ABHD4 (UniProt ID Q8VD66). In Arabidopsis, a soluble acyl CoA-dependent lysophosphatidic acid acyltransferase encoded by the gene At4g24160 showed 34% identity (E-value = 3E−44) with mouse ABHD4, and contained the consensus sequence GXSXG (‘nucleophile elbow’), together with the catalytic serine nucleophile that is necessary for lyso-NAPE lipase activity (Simon and Cravatt, 2006). Another protein, encoded by the gene At4g36530, which showed significant homology (E-value = 1E−12) to mouse ABHD4, was previously shown to be chloroplast-localized (Zybailov et al., 2008; Ferro et al., 2010). Thus, this protein is likely to be involved in starch metabolism (Heyndrickx and Vandepoele, 2012) and may be a poor candidate for NAE synthesis. Given that there are several PLA2 and phospholipase C enzymes in plants (Wang et al., 2012), characterization of an NAE-specific lipase that acts on NAPE to generate lyso-NAPE or phospho-NAE, is a daunting task. Quantification of NAPE metabolites to identify the predominant NAE intermediates may be an alternative way to determine the predominant route for NAE biosynthesis in plants. Radiolabeling studies in vivo and in vitro using seeds of Gossypium hirsutum (cotton) suggest that NAPE is a metabolic precursor for NAE in plants (Chapman and Moore, 1993a; Chapman et al., 1995). However, endogenous accumulation of intermediates may be transient or occur at very low levels, and thus they may escape detection by current analytical techniques. Future research should focus on cloning and characterization of the five homologs of NAPE-PLD in the Viridiplantae clade. Furthermore, determining the ability of Arabidopsis soluble acyl CoA-dependent lysophosphatidic acid acyltransferase to accept NAPE and lyso-NAPE as substrates to generate glycerophospho-NAE, which in turn may be hydrolyzed to NAE, is an important first step in shedding light onto how NAE is synthesized in plants.
The conversion of NAPE to NAE via NAPE-PLD described above is the second reaction in a two step-pathway called the ‘transacylation phosphodiesterase’ pathway, which is the widely accepted route for NAE synthesis in animal tissues (Wang and Ueda, 2009). The first reaction in this pathway (i.e. formation of NAPE) is catalyzed by as yet to be cloned N-acyltransferases. N-acyltransferases catalyze transfer of an acyl group from a glycerophospholipid to the membrane phospholipid phosphatidylethanolamine (Figure 3). Biochemical studies in cottonseed extracts led to the proposal that the prevalent means by which NAPE is synthesized in plants occurs through direct acylation of phosphatidylethanolamine by a reverse serine hydrolase-type catalytic mechanism using free fatty acid as a substrate (Chapman and Moore, 1993b; McAndrew and Chapman, 1998). However, a putative plasma membrane-localized acyltransferase in Arabidopsis encoded by the gene At1g78690 was shown to catalyze the synthesis of NAPE from phosphatidylethanolamine and acyl CoA when expressed in E. coli (Faure et al., 2009). This Arabidopsis acyltransferase was therefore believed to represent an acyl CoA-dependent NAPE synthase distinct from the acyltransferases in cotton and animal tissues (Kim et al., 2010). However, a recent report showing that the protein product of At1g78690 also acts as a lysoglycerophospholipid O-acyltransferase has raised questions as to whether At1g78690 is a bona fide NAPE synthase in planta (Bulat and Garrett, 2011).
Hydrolysis – Termination of NAE Bioactivity by Fatty Acid Amide Hydrolase
Termination of NAE signaling is conserved between plants and animals
Although much remains to be learned about NAE formation in plants, substantial progress has been made in understanding the enzymatic machinery responsible for NAE degradation, in large part because of knowledge gained in animal systems. In animals, breakdown of NAE after its signaling role is complete is accomplished by the enzyme fatty acid amid hydrolase (FAAH) to produce free fatty acid and ethanolamine (Figure 3). FAAH belongs to a large and diverse group of enzymes called the ‘amidase signature’ (AS) family, which contain the conserved catalytic triad Ser-Ser-Lys (Patricelli and Cravatt, 1999; Patricelli et al., 1999). A major breakthrough in animal NAE research was made when the X-ray crystal structure of rat (Rattus) FAAH was solved (Bracey et al., 2002). The crystal structure of rat FAAH provided the necessary framework for interpreting the large body of biochemical studies that began in the 1980s aimed at characterizing the catalytic mechanisms not only of FAAH but also other AS family members (McKinney and Cravatt, 2005).
Transgenic studies in mammalian models with altered FAAH expression provided compelling evidence that FAAH is the primary enzyme for breakdown of these fatty acid amides in vivo. For instance, tissue extracts from mice lacking FAAH showed an up to 100-fold reduction in NAE hydrolysis. These FAAH knockout mice also showed a 10-fold increase in NAE levels in the nervous system, and this was correlated with behavioral changes indicative of altered anandamide signaling, such as reduced pain sensation (Cravatt et al., 2001). Because these FAAH knockout mice showed no other physiological defects, chemical inhibitors of the FAAH enzyme paved the way for development of drugs used for the therapeutic treatment of pain and other neurobehavioral disorders (Blankman and Cravatt, 2013).
The advances made with the mammalian FAAH described above have helped guide NAE research in plants. For example, a search of the Arabidopsis genome using the conserved AS consensus sequence of rat FAAH led to identification of At5g64440 as a putative plant FAAH gene (AtFAAH). In vitro biochemical assays using radioactive NAEs revealed that the recombinant AtFAAH enzyme did indeed hydrolyze NAEs to free fatty acids and ethanolamine (Shrestha et al., 2003). Furthermore, AtFAAH knockout and over-expressor plants had higher and lower endogenous NAEs, respectively, confirming that, as in animals, AtFAAH is a major enzyme that catalyzes the catabolism of plant NAEs in vivo (Wang et al., 2006b; Teaster et al., 2012).
The observation that exogenous NAE 12:0 mediates dose-dependent inhibition of Arabidopsis seedling development has provided a readily quantifiable biological assay for evaluating AtFAAH-altered plants (Blancaflor et al., 2003). For instance, it was shown that AtFAAH knockouts were hypersensitive to the growth inhibitory effects of NAE 12:0, which in some respects is reminiscent of the enhanced sensitivity of mice FAAH knockouts to exogenous anandamide. On the other hand, plants over-expressing AtFAAH exhibited tolerance to NAE 12:0, indicating that the high amount of AtFAAH enzyme in the transgenic plants led to rapid degradation of NAE 12:0, thereby mitigating its inhibitory effects (Wang et al., 2006b; Figure 4a). Since the cloning of AtFAAH, genes encoding FAAH proteins in other plant species, such as Medicago truncatula (MtFAAH) and Oryza sativa (OsFAAH), have been identified. Like recombinant AtFAAH, MtFAAH and OsFAAH were capable of hydrolyzing various NAEs in vitro (Shrestha et al., 2006).
FAAH over-expression suggests a role for NAE catabolism in flowering
An interesting phenotype of AtFAAH over-expressing plants was their tendency to flower earlier than wild-type (Wang et al., 2006b; Teaster et al., 2012; Figure 4b). Although the detailed pathways underlying this phenotype remain largely unknown, transcript profiling of AtFAAH over-expressors indicated a link to known flowering regulators. For example, FLOWERING LOCUS T (FT) is a key transcription factor that controls flowering time by modifying the expression of a complex network of floral genes in the vegetative meristem (Corbesier et al., 2007; Seo et al., 2011). FT and FT-like proteins are now widely believed to be major components of a mobile signal that is translocated from leaves to the shoot apical meristem via the phloem in response to endogenous signals or when environmental conditions favorable for flowering are encountered by the plant (Pin and Nilsson, 2012). Several years of research have shown that a number of signaling pathways that involve both floral activators and repressors converge at FT to orchestrate the transition to flowering (Pose et al., 2012). For instance, FT expression is triggered by binding of the B-box-containing protein CONSTANS (CO) to the FT promoter during inductive long-day conditions (Tiwari et al., 2010; Song et al., 2013). The onset of flowering is also defined by negative regulation of factors that repress FT expression, such as FLOWERING LOCUS C (FLC), a transcription factor that also participates in the transition from juvenile to adult phase. Upon transport to the shoot apical meristem, FT directly or indirectly promotes the expression of floral meristem identity genes such as LEAFY (LFY) and APETALA (AP1) to initiate flowering (Pose et al., 2012). The results of microarray and quantitative RT-PCR analysis showing increased FT transcripts and down-regulation of a suite of floral repressors in AtFAAH over-expressors are consistent with its early flowering phenotype (Teaster et al., 2012). These floral repressors included genes encoding a mannose-binding lectin superfamily protein (At3g16460), a cupredoxin superfamily protein (At3g27200), squamosa-promoter binding protein-like (At5g43270) and a TRAF-like family protein (At3g20370; Schmid et al., 2003).
There is evidence showing that FT inhibits vegetative growth by limiting leaf size (Shalit et al., 2009; Danilevskaya et al., 2010). However, despite increase FT expression, AtFAAH over-expressors showed enhanced vegetative growth (Wang et al., 2006b; Teaster et al., 2007). In 14-day-old short day-grown plants, endogenous NAE 12:0 levels were 30% less in AtFAAH over-expressors compared to wild-type, and NAE 12:0-treated wild-type plants displayed delayed flowering (Teaster et al., 2012). This indicates that the lower levels of NAE in AtFAAH over-expressors appear to (i) counteract the negative effect of FT on vegetative growth, and (ii) promote the transition from vegetative to reproductive growth. The exact mechanisms by which NAEs interact with FT to modulate early flowering remain an open question. However, annotation of FT as a phosphatidylethanolamine-binding protein opens up the intriguing possibility that FT may interact directly with ethanolamine-containing lipids such as NAE in FT transport from leaves to the vegetative meristem to exert its effect on flowering (Teaster et al., 2012).
FAAH over-expression uncouples plant responses to bacterial pathogens from NAE hydrolytic activity
In addition to early flowering, another interesting phenotype exhibited by AtFAAH over-expressors was their hypersensitivity to a range of host and non-host bacterial pathogens (Kang et al., 2008). This observation is in agreement with earlier findings showing that NAE may participate in plant defense responses. For example, it was shown previously that application of fungal elicitors triggered an increase of both NAPE and NAE in tobacco leaves and cell suspensions (Chapman et al., 1995, 1998; Tripathy et al., 1999). It is possible that NAE may have a protective role in plants, and that the lower endogenous NAEs in the AtFAAH over-expressors rendered them more susceptible to pathogen attack. However, it was shown that endogenous levels of NAEs did not change significantly after inoculation with bacterial pathogens, suggesting that NAE levels may not be as tightly linked to pathogen responses as originally thought (Kang et al., 2008). Although there is a possibility that transient changes in endogenous levels of NAE after bacterial inoculation may have been missed, the fact that hypersensitivity to bacterial pathogens may also be induced by over-expression of catalytically inactive AtFAAH indicates that NAE hydrolytic activity is independent of plant defense responses (Kim et al., 2009). A similar scenario was observed with regard to the hypersensitivity of AtFAAH over-expressors to ABA (Figure 4c), but the precise mechanisms by which AtFAAH protein confers hypersensitivity to pathogens and ABA remains unresolved.
There is accumulating evidence that pathways governing plant immunity are under tight regulation by hormones. The hormones that have been most widely studied with regard to plant immunity are salicylic acid (SA), a compound synthesized by the phenylpropanoid pathway, and jasmonic acid (JA), a lipid metabolite produced through the oxylipin biosynthetic pathway (Mosblech et al., 2009; Vogt, 2010). Upon encountering a pathogen, the levels of SA and JA have been reported to increase, and this is typically accompanied by enhanced expression of a multitude defense-related genes, such as pathogenesis-related PR genes and WRKY transcription factor genes for SA, and plant defensin 1.2 (PDF1.2) for JA (Pieterse et al., 2012). Interestingly, AtFAAH over-expressors had twofold lower SA levels, and this was accompanied by down-regulation of SA biosynthetic and SA-related defense genes including PR genes and WRKYs (Kang et al., 2008). Importantly, benzo-(1,2,3)-thiadiazole-7-carbothoic acid S-methyl ester, an SA analog, rescued the hypersensitivity of AtFAAH over-expressors to non-host pathogens and reduced the growth of pathogens (Kang et al., 2008). Taken together, these results show that increased AtFAAH negatively affects SA biosynthesis and SA-mediated defense responses. Therefore, the reduced SA levels may partly account for the hypersensitivity response of AtFAAH over-expressors to bacterial pathogens (Kang et al., 2008). However, it is not clear how enhanced NAE hydrolytic activity due to AtFAAH over-expression translates into reduced SA, or whether FAAH has additional, undiscovered activities in plant defense. The lower basal levels of ABA and JA in AtFAAH over-expressors (Kang et al., 2008) and reports showing that ABA, JA and SA have both antagonistic and synergistic effects toward one another during plant defense responses (Pieterse et al., 2012), indicate another level of complexity with regard to how AtFAAH mediates plant interactions with bacterial pathogens.
Other NAE amidases in plants
Whereas AtFAAH over-expressors displayed a number of phenotypes, AtFAAH knockouts, except for their hypersensitivity to exogenous NAE, generally resembled wild-type plants (Wang et al., 2006b). Furthermore, despite the increased seed NAE content of AtFAAH knockouts compared to wild-type, their NAE levels were still decreased during germination. This indicates the existence of alternative pathways for NAE catabolism in plants that do not involve AtFAAH. In this regard, it is worth noting that a second enzyme with FAAH activity was discovered in primates but not in other mammals such as mouse and rat in which the original FAAH protein was discovered. This primate AS protein was designated FAAH-2, and exhibited differences in enzymatic activity compared with rat FAAH discussed above (referred to as FAAH-1; Wei et al., 2006). It is also known that the hydrolysis of NAEs in mammals is not exclusively accomplished by AS-containing enzymes. A gene encoding an enzyme that resides in the lysosome with optimum NAE amidase activity at acidic pH was found to be enriched in the immune system. This enzyme, designated NAE-hydrolyzing acid amidase, is related to acid ceramidases (Tsuboi et al., 2005). So far, no plant proteins with similarity to NAE-hydrolyzing acid amidase have been identified.
In plants, there are several genes encoding proteins that have the AS consensus sequence, but only At5g64440 (AtFAAH) discussed above and At1g08980 have been characterized in detail (Kilaru et al., 2007). The latter gene encodes a protein called amidase 1 (AMI1), which has been shown to catalyze the reaction that synthesizes indole acetic acid from indole-3-acetamide, and thus appears to play a role in an alternative route for auxin biosynthesis (Neu et al., 2007; Mano et al., 2010). Although AMI1 displayed its highest activity toward indole-3-acetamide, it has been shown to have minor NAE hydrolytic activity (Pollmann et al., 2006). It remains to be determined whether the low NAE hydrolytic activity of AMI1 plays any role in NAE catabolism in plants.
The gene At5g07360, which is highly expressed in seed and shares significant sequence homology with At5g64440, may encode an FAAH protein that may function in seed development. Interestingly, At5g07360 expression is increased in ABA hypersenstive mutants (Nishimura et al., 2007), and it is tempting to speculate that the AS protein encoded by this gene exhibits similar properties to AtFAAH with regard to potential interactions with ABA (Teaster et al., 2007; Kim et al., 2009). Furthermore, the sequence of At3g25660 contains all the amino acids of the core catalytic triad, and thus represents another potential protein invovled in NAE degradation in plants. One area for future research may be to generate combinatorial mutants for these various plant AS protein-encoding genes to determine whether they participate in endogenous NAE hydrolysis.
NAE Metabolism Interacts with the ABA Signaling Pathway to Elicit Secondary Dormancy
As NAEs are most abundant in desiccated seeds, their rapid depletion during seed imbibition and germination may be an important requirement for seedling establishment to proceed normally (Chapman et al., 1999). This notion has been supported indirectly by the abnormal Arabidopsis seedling development resulting from exogenous application of various NAE types, including NAE 12:0, NAE 18:2 and NAE 18:3. Some of these seedling growth defects were accompanied by disorganization of the cytoskeleton and chloroplast degeneration (Figure 2; Blancaflor et al., 2003; Motes et al., 2005; Keereetaweep et al., 2013). Moreover, AtFAAH over-expressors, in which endogenous NAEs are depleted more rapidly than wild-type, displayed more robust seedling growth, further indicating that these small lipids are negative growth regulators (Wang et al., 2006b).
As already mentioned, NAEs may exert some of their biological effects in plants through their interaction with ABA signaling. A key piece of evidence supporting this claim is the finding that NAE 12:0 inhibited PLDα activity and blocked ABA-induced stomatal closure (Austin-Brown and Chapman, 2002). Recently, the inhibitory effect of NAE 12:0 on PLDα was linked to attenuation of the progression of senescence in detached leaves, a process that is typically promoted by ABA (Jia et al., 2013).
The interaction between NAE metabolism and ABA signaling has received further support from studies of seedling establishment and a process described as secondary dormancy. Secondary dormancy is a stage-specific growth arrest that may be induced by several abiotic stresses, and is mediated by the ABA signaling pathway that regulates the seed-to-seedling transition (Lopez-Molina et al., 2001). Several lines of evidence indicate that NAEs negatively regulate growth via ABA signaling and participate in this process of seedling growth arrest. First, microarray studies revealed that expression of a large subset of genes up-regulated by ABA (Li et al., 2006) was also increased in seedlings treated with NAE 12:0 to arrest growth (Teaster et al., 2007). Second, upon normal seedling establishment, ABA levels rapidly decreased during the seed-to-seedling transition over a time course that roughly mirrored that of NAE depletion. Third, NAE 12:0 and ABA induced a synergistic and profound negative effect on seed germination and seedling growth (see Figure 2). Fourth, NAE and ABA inhibited seedling growth within a similar window of development, and NAE effects were dependent, in part, on an intact ABA signaling pathway via ABI3 (Teaster et al., 2007). Further, AtFAAH over-expressors were hypersensitive to ABA (Figure 4c; Teaster et al., 2007). Finally, seedling growth arrest by ABA induced a rapid accumulation of NAE oxidative metabolites, which were also shown to negatively regulate seedling growth (Keereetaweep et al., 2013). Although it is not currently known how the interaction between ABA signaling and NAE metabolism is accomplished at the molecular level, a number of possibilities are emerging from both pharmacological and gene expression studies. Most notable is the potential link to ABI3, a key transcriptional regulator of several ABA-related genes (Nakashima et al., 2006). It was shown that ABI3 transcripts were strongly expressed in seedlings treated with NAE 12:0 (Teaster et al., 2007; Cotter et al., 2011). The increased expression of ABI3, whose product is typically depleted as the seed germinates, suggests that increased NAE levels maintain plants in a state of secondary dormancy (Lopez-Molina et al., 2001, 2002). The synergistic effects between NAE 12:0 and ABA further indicate that NAE acts in negative regulation of seedling growth, in part by enhancing the sensitivity of seedlings to ABA. It is worth noting that ABA has been shown to exert some of its physiological effects in plants through heterotrimeric G-proteins, calcium ions, nitric oxide, reactive oxygen species and sphingolipids, all of which are known molecular targets of NAEs in animal systems (Allen et al., 2002; Coursol et al., 2003; Fowler, 2003; Bright et al., 2006; Chai et al., 2006). However, whether these signaling elements are involved in mediating ABA–NAE interaction requires further study (Kim et al., 2010).
The ability of NAE 12:0 to inhibit PLDα activity may also explain its interaction with ABA during the seed-to-seedling transition (Austin-Brown and Chapman, 2002). Phosphatidic acid is formed as a result of PLD-mediated hydrolysis of membrane lipids, and there is a large body of evidence implicating this lipid mediator in ABA-mediated signaling events that control seed germination (Zhang et al., 2004, 2009; Katagiri et al., 2005; Mishra et al., 2006; Uraji et al., 2012). For the future, it would be interesting to determine how various PLD mutants respond to exogenous NAEs, although there are 12 isoforms of PLD in Arabidopsis with some overlapping functions, which will make this a difficult task. Furthermore, given that lower PLDα expression improves seed longevity in Arabidopsis seeds (Devaiah et al., 2007), it would be interesting to investigate how seed viability is influenced in plants with altered NAE catabolism (e.g. AtFAAH knockouts and over-expressors).
Alkamides and N-Acyl-Homoserine Lactones Mirror NAE Effects on Seedling Development
Another group of lipids that are structurally similar to NAEs, the alkamides, have been identified in higher plants, and shown to have some effects on growth that are similar to those of NAEs (Figure 1c). Unlike NAEs, which occur widely in plants, alkamides appear to be restricted in the plant kingdom to around ten families of angiosperms (Ramírez-Chávez et al., 2004; López-Bucio et al., 2007). Alkamides as a group are represented by more than 200 different structures (Ramírez-Chávez et al., 2004). Alkamides, like NAEs, comprise an acyl chain linked via an amide bond to an amine-containing head group. The nature of this alkyl amine group may vary, with butyl, isobutyl and propyl groups having been reported; notably these are alkyl groups not alcohols (Boonen et al., 2012). Perhaps the best studied alkamide is N-isobutyl-2E,4E,8Z-decatrienamine, also named affinin. Affinin is found in high quantities in plant species such as Echinacea purpurea or Echinacea angustifolia, especially in root tissues (Ramírez-Chávez et al., 2004; López-Bucio et al., 2007). NAE 12:0 and affinin share similar structural features, and also have some similar effects on plant development (Blancaflor et al., 2003; López-Bucio et al., 2006, 2007; Morquecho-Contreras et al., 2010). Therefore, it has been proposed that these lipids may share common signaling pathway partners for their various actions in plants. However, recently, it has been shown that, unlike NAE 12:0, affinin does not appear to intersect with ABA signaling, but rather with cytokinin and jasmonic signaling pathways (López-Bucio et al., 2007; Morquecho-Contreras et al., 2010; Méndez-Bravo et al., 2011). Due to the amino head group composition of alkamides, it is unlikely that NAEs and alkamides are synthesized through the same metabolic pathway. While NAEs are formed by hydrolysis of NAPE by a PLD (see above; Chapman and Moore, 1993a,b; Coulon et al., 2012a), it has been suggested that alkamides are probably synthesized from either of two amino acid precursors: valine or phenylalanine (Cortez-Espinosa et al., 2011). Thus, despite structural and functional similarities, NAEs and alkamides are now usually perceived as two distinct small lipid mediators in plants.
The N-acyl-homoserine lactones are a class of small molecules that are structurally similar to NAEs (Figure 1c). N-acyl-homoserine lactones are important signals in quorum sensing to facilitate bacterial cell-to-cell communication. Like alkamides and NAEs, N-acyl-homoserine lactones have been shown to modify seedling root architecture. Interestingly, seedlings in which AtFAAH expression is altered appear to exhibit modified sensitivity to N-acyl-homoserine lactone-induced root architectural changes, indicating that plants may metabolize these small lipids through AtFAAH (Ortíz-Castro et al., 2008). It remains to be tested whether N-acyl-homoserine lactones are substrates of FAAH.
The Lipoxygenase Pathway and Polyunsaturated NAE Metabolism
Metabolic pathways for the formation of NAE oxylipins in plants
Much of what we know about NAE metabolism and its physiological function in plants has been derived from studies of the hydrolytic pathway catalyzed by FAAH. Recently, however, focus has shifted to alternative routes for NAE metabolism, specifically its oxidation (Figure 5). In plants, polyunsaturated free fatty acids containing cis double bonds are oxidized by lipoxygenases (LOX; Feussner and Wasternack, 2002; Liavonchanka and Feussner, 2006; Schaller and Stintzi, 2009) or α-dioxygenases (Hamberg et al., 2005; Bannenberg et al., 2009; Vicente et al., 2012), generating a large group of lipid-derived metabolites, commonly referred to as oxylipins. Depending on the position (C9 or C13) of the molecular oxygen being introduced into linoleic acid (18:2) or linolenic acid (18:3) free fatty acids, LOX enzymes may be classified as 9-LOX and 13-LOX enzymes, respectively (Schneider et al., 2007). The specificities of LOX enzymes give rise to the stereospecific corresponding hydroperoxide derivatives, which are relatively unstable and further metabolized by allene oxide synthase, allene oxide cyclase, divinyl ether synthase, epoxy alcohol synthase, hydroperoxide lyase, peroxygenase and reductases, resulting in a broad range of oxylipin species (Mosblech et al., 2009). These compounds have been reported to possess diverse and significant biological activities in planta (Browse, 2009; Seltmann et al., 2010; Dave et al., 2011; Nalam et al., 2012; Wasternack and Hause, 2013).
Interestingly, in both mammals and plants, polyunsaturated NAEs were also shown to serve as substrates for enzymes in the LOX pathway despite their low endogenous levels (van der Stelt et al., 2000; Shrestha et al., 2002; Keereetaweep et al., 2010, 2013; Zheng and Brash, 2010; Kilaru et al., 2011). In plants, α/γ-NAE 18:3 and NAE 18:2 were shown to serve as suitable substrates for LOX-1 of soybean (Glycine max) to produce 13S-hydroperoxide derivatives. The hydroperoxide derivatives of α-NAE 18:3 and NAE 18:2 were also further metabolized by hydroperoxide lyase of alfalfa (Medicago sativa) to 12-oxo-N-(9Z)-dodecenoyl(ethanol)amine, (3Z)-hexenal (for NAE 18:3) and hexanal (for NAE 18:2) in vitro. Cyclized products (12-oxo-N-phytodienoylethanolamines) were also detected as a product of allene oxide synthase of flax (Linum usitatissimum) seed, using a 13S-hydroperoxide derivative of α-NAE 18:3 (13S-hydroperoxy-α-NAE 18:3) (van der Stelt et al., 2000). GC/MS analysis of lipid metabolites generated from incubation of imbibed cottonseed extracts and NAE 18:2 revealed formation of two diastereomers of α-ketols (12-oxo-13-hydroxy-N-(9Z)-octadecenoylethanolamine), indicating the presence and ability of 13-LOX and allene oxide synthase to metabolize NAE 18:2 in seed-derived tissues (Shrestha et al., 2002).
More recent studies also demonstrated oxidation of radiolabeled polyunsaturated NAEs by crude homogenates from 4- to 8-day-old Arabidopsis seedlings (Kilaru et al., 2011). This oxidation activity was significantly higher than hydrolysis activity in 4-day-old seedling extracts, while the enzyme extract from 8-day-old seedlings had more than 10-fold higher hydrolysis activity than oxidation activity (Kilaru et al., 2011). Endogenous LOX-mediated metabolites of NAE 18:2 (9NAE-hydroxyoctadecadienoic (HOD), 13NAE-HOD) were readily detected in 4-day-old AtFAAH knockout seedlings, but were not detected in desiccated seed and 4-day-old wild-type seedlings. Neither wild-type nor AtFAAH over-expressor seedlings showed detectable endogenous LOX-mediated metabolites from NAE 18:2, possibly due to AtFAAH activity decreasing NAE 18:2 availability for the oxidative pathway. However, feeding experiments with high micromolar concentrations of NAE 18:2 and NAE 18:3 resulted in quantifiable levels of LOX-mediated metabolites from both substrates (9NAE-HOD, 13NAE-HOD, (9S,12Z,10E,15Z)-9-hydroxy-10,12,15-octadecatrienoylethanolamide and (13S,9Z,11E,15Z)-13-hydroxy-9,11,15-octadecatrienoylethanolamide) in wild-type and plants with altered AtFAAH expression (Figure 5). These LOX metabolites were found to be highest in AtFAAH knockouts and lowest in AtFAAH over-expressors due to the competing hydrolysis pathway (Kilaru et al., 2011; Keereetaweep et al., 2013). The occurrence of endogenous LOX-mediated NAE oxylipins in plants suggests that oxidative metabolites of polyunsaturated NAEs may represent a group of lipid mediators involved in seedling development and post-germinative seedling growth.
Potential physiological roles of polyunsaturated NAEs and NAE oxylipins on seedling growth and development
Improvements in the ability to quantify NAEs and NAE oxidation products by sensitive analytical procedures are providing some clues on the physiological roles of polyunsaturated NAEs and NAE oxylipins in plants. For example, the NAE 12:0-induced seedling growth inhibition may in part be explained by the discovery that this short-chain saturated NAE species is a competitive inhibitor of LOX (Keereetaweep et al., 2010). As a consequence, NAE 12:0 triggers the accumulation of polyunsaturated NAEs such as NAE 18:2 and NAE 18:3, and this effect was most pronounced in AtFAAH knockouts, which catabolize NAEs less efficiently than wild-type (Kilaru et al., 2012). Interestingly, NAE 12:0 was able to mitigate the wound-induced accumulation of JA, an oxylipin pathway-derived plant hormone (Keereetaweep et al., 2010). Thus, in addition to ABA and SA, JA signaling processes in plants may be influenced by NAEs, primarily through its effect on oxylipin metabolism.
Recently, a very specific, NAE 18:3-induced bleaching of Arabidopsis cotyledons was linked to LOX-derived NAE 18:3 oxylipins, specifically 9- and 13-hydroxides and hydroperoxides of NAE 18:3 (Keereetaweep et al., 2013). Formation of these metabolites was part of the process of NAE 18:3 depletion during seedling establishment that appears to involve competition among AtFAAH, 9-LOX and 13-LOX enzymes (Figure 5). Flux through these enzymes was shown to be altered in various AtFAAH and AtLOX knockout genotypes, such that the partitioning of NAE metabolites appears to be malleable and results in a mix of free fatty acids and ethanolamide-conjugated oxylipins (Kilaru et al., 2011; Keereetaweep et al., 2013). Interestingly, plants in which seedling growth arrest was induced by ABA exhibited a more than 50-fold increase in endogenous 9- and 13-LOX-derived NAE 18:3 oxylipins compared to untreated seedlings, suggesting that the balance of oxidation of polyunsaturated NAEs may work in concert with ABA to activate secondary dormancy (Keereetaweep et al., 2013). Furthermore, given the observation that NAE 18:3 oxylipins induce chlorophyll breakdown and chloroplast degradation, it is tempting to speculate that these NAE 18:3 oxidation products may specifically influence physiological processes that participate in the seed de-greening that typically accompanies seed maturation, as it was recently shown that stay-green (SGR) genes are regulated by ABI3 (Delmas et al., 2013). Further examination of the link between NAE oxylipins and ABA signaling is warranted in light of the occurrence and potent activities of these ethanolamide-conjugated oxylipins.
The disappearance of morphologically recognizable chloroplasts in seedlings with increased NAE 18:3 oxylipins (Keereetaweep et al., 2013) suggests that this phenomenon may be related to well-characterized processes associated with chlorophyll degradation that occur during leaf senescence (Hörtensteiner, 2013). Specific chlorophyll metabolites have not been identified or quantified in seedlings with increased NAE 18:3-derived oxylipins. However, there are several characteristics of chloroplast disruption in this particular situation (Keereetaweep et al., 2013) that may distinguish this seedling growth-arrest from the general process of chlorophyll turnover. First, the effects are stage-specific, i.e. seedlings only bleach within a narrow window of development (approximately 4 days after sowing). Second, the effects are organ-specific: cotyledons, but not true leaves, bleach upon treatment and accumulation of NAE 18:3-derived oxylipins. Third, the process is not toxic; seedlings recover and re-synthesize chloroplasts in their cotyledons when the levels of NAE 18:3 and NAE 18:3 hydroxides and hydroperoxides are reduced. Finally, the partial bleaching of cotyledons (and growth arrest) may be simulated by treatment with exogenous ABA, and this results in a concomitant increase in endogenous NAE 18:3-derived oxylipin metabolites. These results were interpreted as suggesting that LOX metabolism of NAE 18:3 is part of a fine-control mechanism that exists to help regulate the development of seedlings under favorable environmental conditions (Keereetaweep et al., 2013), but more work is required to support this hypothesis. Nonetheless, it appears that the stage-specific arrest of cotyledon development is somewhat different from the processes related to chloroplast degradation and chlorophyll catabolism that accompany programmed cell death or the normal senescence syndrome (Hörtensteiner, ). Nevertheless, the relationship between pigment catabolites and NAE 18:3 oxylipins is worth investigating, as the chloroplast targets and likely molecular changes in development may share some overlapping signaling pathways.
Pressing Questions and Future Directions
While there is remarkable evolutionary conservation in the enzymatic machinery involved in metabolism of these bioactive acylethanolamides, there remain many unanswered questions regarding regulation of their formation and the precise targets of their action in plant systems. For example, is NAE formed directly from hydrolysis of NAPE by a PLD-type enzyme in plants, or indirectly through some other combination of lipolytic activities? What role does NAPE formation play in the supply of NAE metabolites? Given the occurrence and potent activities of NAE oxylipins in emerging seedlings, what factors regulate the balance between NAE hydrolysis and oxidation and resulting metabolites? What are the specific molecular targets of NAE metabolites – are there specific transcription factors, regulatory enzymes, and/or membrane-bound receptors that transduce NAE metabolite action? Or are there unique, undiscovered plant-specific targets that mediate NAE interaction with phytohormone pathways? To what extent is the NAE regulatory pathway used in modulating various plant processes, and to what degree is there overlap in the mechanism of action? The answers to these and many questions will help to uncover the nature of the role of this group of small molecule mediators in regulation of plant growth and development, and may provide insights into the evolutionary adaptation of this signaling pathway in diverse biological organisms.
We gratefully acknowledge the US Department of Energy, Office of Science, Basic Energy Sciences program (grant number DE-FG02-05ER15647) for supporting our research on plant NAEs.