The cuticular waxes from many plant species comprise roughly equal amounts of the various compound classes, with no particular class predominating. For example, alkanes, aldehydes, primary alcohols, fatty acids and alkyl esters each contribute 9–42% of the leaf wax of Zea mays (Bianchi et al., 1984). In contrast, the wax mixtures from many other plant species contain high percentages of a single compound class. Hordeum vulgare leaf wax, for example, contains 89% of primary alcohols, together with only 0.2–9% of alkanes, aldehydes, fatty acids and alkyl esters (Giese, 1975).
Compound chain length
Variation in the chain length of wax compounds is generated during synthesis of VLCFA wax precursors. This process involves several enzyme complexes in various cellular compartments. The first phase, the de novo fatty acid synthesis of C16 and C18 acyl chains, is catalysed by the soluble fatty acid synthase (FAS) complex localized in the plastid stroma (Ohlrogge and Browse, 1995; Ohlrogge et al., 1993), and proceeds through a cycle of four reactions utilizing intermediates attached to acyl carrier protein (ACP). In each cycle, comprising the condensation of a C2 moiety originating from malonyl ACP to acyl ACP, the reduction of β-ketoacyl ACP, the dehydration of β-hydroxyacyl ACP and the reduction of trans-Δ2–enoyl ACP, the acyl chain is extended by two carbons. Three different FAS complexes participate in the production of C18 fatty acids in the plastid. They differ in their β-ketoacyl-acyl carrier protein synthase (KAS) condensing enzymes, which have strict acyl chain-length specificities: KASIII (C2–C4; Clough et al., 1992), KASI (C4–C16) and KASII (C16–C18; Shimakata and Stumpf, 1982). The two reductases and the dehydratase have no particular acyl chain-length specificity and are shared by all three plastidial elongation complexes (Stumpf, 1984).
The second phase (Figure 2), the extension of the C16 and C18 fatty acids to VLCFA chains, is carried out by fatty acid elongases (FAE; von Wettstein-Knowles, 1982), multienzyme complexes bound to the endoplasmic reticulum membrane (Kunst and Samuels, 2003; Xu et al., 2002; Zheng et al., 2005). To reach the ER-associated fatty acid elongation sites, saturated C16 and C18 acyl groups must be hydrolysed from the ACP by an acyl ACP thioesterase, exported from the plastid, and esterified to CoA. Two classes of acyl ACP thioesterases, designated FATA and FATB, have been described in plants. The FATA class exhibits a strong preference for 18:1 ACP in vitro, while the FATB thioesterases predominantly use saturated fatty acids (Voelker, 1996). The involvement of the FATB thioesterase in cuticular wax biosynthesis has been confirmed by analyses of the Arabidopsis fatb mutant, which exhibits a major reduction in its wax load (Bonaventure et al., 2003). The specifics of fatty acid export from the plastid, CoA esterification and transport to the ER are not well understood. Fatty acids released from ACP by a thioesterase in the plastid undergo conversion to acyl CoAs by a long-chain acyl CoA synthetase (LACS) in the outer envelope membrane. Of the nine LACS genes annotated in the Arabidopsis genome (Shockey et al., 2002), only one, LACS9, has been demonstrated to encode a plastid envelope enzyme (Schnurr et al., 2002). However, loss of function of LACS9 does not result in reduced export of acyl groups from the chloroplast, or a wax-deficient phenotype (Schnurr et al., 2002), suggesting that the LACS isozyme primarily responsible for CoA esterification of fatty acids en route to wax biosynthesis has yet to be identified. Movement of the fatty acyl group from the thioesterase to LACS has been proposed to occur by some type of facilitated diffusion (Koo et al., 2004), but the exact mechanism of transfer is not known. An alternative model for fatty acid export from the plastid was recently suggested by Bates et al. (2007). Their radiolabelling studies revealed that 16:0 and 18:1 fatty acids synthesized de novo in the plastid can be incorporated into phosphatidylcholine (PC), perhaps by direct acylation of lyso-PC. The acyl groups removed from PC by acyl editing may then be fed into the acyl CoA pool. However, mechanistic details and the relevance of this process for epidermal wax formation have not been established.
Figure 2. Wax biosynthetic pathways. Repeated cycles of four enzymatic steps first elongate acyl CoA precursors. They are then modified by one of (up to) five different reactions into various compound classes. Preferred chain lengths are indicated by numbers. Characterized enzymes catalysing key biosynthetic steps are shown in blue (CER6, condensing enzyme=β-ketoacyl CoA synthase; KCR, β-ketoacyl CoA reductase; dehydratase, β-hydroxyacyl CoA dehydratase; CER10, enoyl CoA reductase; CER4, fatty acyl CoA reductase; WSD1, wax ester synthase; MAH1, mid-chain alkane hydroxylase).
Download figure to PowerPoint
Translocation of fatty acids to the ER, where additional acyl chain elongation and modification of VLCFAs to diverse aliphatic wax components take place, appears to involve plastid-associated membranes (PLAMs; Andersson et al., 2007). Physical manipulation of GFP-labelled ER strands, using laser scalpels and optical tweezers, experimentally verified the intimate connection between the plastid and the ER of Arabidopsis leaf protoplasts. Therefore, PLAMs have been proposed to be major routes for lipid transfer between the two organelles.
Elongation of C16 and C18 fatty acids to VLCFAs involves cycles of four consecutive enzymatic reactions analogous to those of the FAS (Figure 2), and results in a two-carbon extension of the acyl chain per cycle. The chain lengths of aliphatic wax components are typically in the range of 20–34 carbons, thus multiple elongation cycles are needed to extend the acyl chain to its final length. The differential effects of inhibitors on incorporation of radiolabelled precursors into wax components of various chain lengths, and analyses of mutants with defects in fatty acid elongation, demonstrated that sequential acyl chain extensions are carried out by several distinct FAEs with unique substrate chain-length specificities (von Wettstein-Knowles, 1993). Specificity of each elongation reaction resides in the condensing enzyme of the FAE complex (Lassner et al., 1996; Millar and Kunst, 1997). Consistent with the requirement for fatty acyl precursors of diverse chain lengths for the synthesis of cuticular waxes, a family of 21 FAE condensing enzyme-like sequences has been identified in the A. thaliana genome (Dunn et al., 2004). An unrelated ELO-like gene family of putative condensing enzymes, related to the Saccharomyces cerevisiae condensing enzymes ELO1, ELO2 and ELO3, has also been annotated (Dunn et al., 2004). It is not known how many of these putative condensing enzymes participate in wax production and how many different condensing enzymes are needed for the elongation of a C18 to a C34 fatty acyl CoA, as single condensing enzymes may catalyse multiple elongation steps. The only wax-specific condensing enzyme characterized to date is CER6 (Fiebig et al., 2000; Hooker et al., 2002; Millar et al., 1999), which is involved in the elongation of fatty acyl CoAs longer than C22.
Unlike the condensing enzymes, the other three enzyme activities of the FAE complex, the β-ketoacyl reductase, β-hydroxyacyl dehydratase and enoyl reductase, are shared by all VLCFA elongase complexes. Thus, these three enzymes have broad substrate specificities and generate a variety of acyl products used to make different classes of lipids (Millar and Kunst, 1997). Because genetic screens in Arabidopsis did not result in isolation of mutants defective in the reductases or the dehydratase, suggesting that these enzymes are essential and/or functionally redundant (Millar and Kunst, 1997), genes encoding the β-ketoacyl reductase and enoyl reductase were cloned by homology to the corresponding sequences from Saccharomyces cerevisiae (Beaudoin et al., 2002; Kohlwein et al., 2001). Two β-ketoacyl reductase (KCR) genes are present in both the A. thaliana and maize (Zea mays) genomes. The maize genes, named GL8A and GL8B (Dietrich et al., 2005; Perera et al., 2003; Xu et al., 2002), are not only expressed in the epidermis, but also in internal tissues. Attempts to generate double mutants by crossing gl8a × gl8b failed because embryos carrying both mutations were not viable. Thus, the KCR has an essential function in plants, most likely in the production of sphingolipids (Dietrich et al., 2005).
An A. thaliana single-copy gene was identified as an enoyl reductase (ECR) candidate. Heterologous expression of the putative plant ECR gene rescued the temperature-sensitive lethality of yeast tsc13-1elo2Δ cells (Gable et al., 2004), demonstrating that it encodes a functional ECR. The A. thaliana ECR gene is ubiquitously expressed, and the protein physically interacts with the Elo2p and Elo3p condensing enzymes when expressed in yeast (Gable et al., 2004). The A. thaliana ECR was shown to be identical to CER10 (Zheng et al., 2005), the protein defective in one of the original A. thaliana eceriferum mutants isolated by Koornneef et al. (1989). These eceriferum (literally ‘not bearing wax’) mutants lack epicuticular wax crystals and therefore have glossy green inflorescence stems that can easily be recognized in visual screens. Biochemical analysis of the cer10 mutant demonstrated that the ECR gene product is involved in the VLCFA elongation that is required for synthesis of all the VLCFA-containing lipids, including cuticular waxes, seed triacylglycerides and sphingolipids (Zheng et al., 2005). Although the plant dehydratase remains unknown, recent identification of the yeast β-hydroxyacyl dehydratase PHS1 (Denic and Weissman, 2007) should permit cloning and characterization of this enzyme from plants.
In addition to variations in the chain-length distributions, cuticular wax mixtures from diverse plants and plant organs also contain various constituent compound classes. These compounds vary in the nature and position of the (typically oxygen-containing) functional groups, with the extreme case of hydrocarbons that are devoid of functional groups (Jetter et al., 2006). Five or more parallel reactions (or pathways), all competing for the VLCFA CoA precursors, can be envisioned leading to these ubiquitous wax components: (i) acyl reduction, (ii) esterification with an alkyl alcohol, (iii) hydrolysis, (iv) aldehyde formation and (v) alkane formation (Figure 2). Knowledge of all the wax biosynthetic reactions will assist in their exploitation for biotechnological production of individual compounds and/or mixtures of compounds with specific combinations of functional groups.
In virtually all vascular plants, wax compound classes with predominantly even numbers of carbons are produced by the so-called acyl reduction pathway (Figure 2; Kunst et al., 2006). The most important of these compounds are primary alcohols and alkyl esters. The latter are essentially dimeric compounds, in which the primary alcohols are bonded to acyl groups, most commonly C16, C18 or VLCFAs (>C20; Figure 3). Primary alcohols are thus central metabolites of wax biosynthesis, and their formation from VLCFA CoA esters has been studied extensively.
Figure 3. Array of biosynthetic reactions leading to wax esters. First, the variety of chain lengths is generated by elongation, leading to C22 fatty acid (‘ic’) precursors in seeds (black arrows) and including all chain lengths up to C32 in epidermal cells (orange arrows). Then, individual acyl precursors are reduced to the corresponding alcohols (‘ol’) (green arrows), and alcohols and acyl CoAs of various chain lengths are combined into esters (blue arrows). Depending on the specificity of the elongase (KCS), acyl reductase (FAR) and ester synthase (WS) enzymes, various mixtures of ester isomers and chain lengths can be generated. Arabidopsis stem surface wax contains esters with predominantly C16 acyl and C22–C30 alkyl groups.
Download figure to PowerPoint
Two reduction steps are required to transform acyl precursors into primary alcohols, and aldehydes must occur as intermediates of the reaction sequence. It has been much debated whether both reduction steps are catalysed by one fatty acyl reductase (FAR), or whether two separate enzymes are necessary for alcohol formation. There is currently substantial evidence for the existence of a one-enzyme system in a number of plant species, including the green alga Euglena gracilis (Kolattukudy, 1970) as well as the angiosperms jojoba (Simmondsia chinensis; Pollard et al., 1979), pea (Pisum sativum; Vioque and Kolattukudy, 1997) and A. thaliana (Rowland et al., 2006). For example, functional expression of genes specifying alcohol-forming FARs from jojoba (Metz et al., 2000) and A. thaliana (Rowland et al., 2006) in heterologous systems demonstrated that alcohol biosynthesis from VLCFAs in these species is carried out by a single alcohol-forming FAR. In contrast, biochemical feeding experiments that allowed isolation of an aldehyde intermediate suggest that the two-step process of alcohol formation operates in Brassica oleracea (Kolattukudy, 1971). However, similar biochemical evidence from other species and molecular information supporting the two-step process in any system is currently lacking.
It is generally assumed that primary alcohols serve as precursors for ester biosynthesis. However, detailed analyses of esterified and free alcohols of various mutants of A. thaliana only recently demonstrated a clear correlation of alcohol chain lengths in both types of compounds, indicating that the free alcohols are indeed incorporated into the wax esters (Lai et al., 2007). In addition, this study revealed that the levels of free alcohols are limiting for ester formation. Thus, a pool of primary alcohols, generated in the A. thaliana epidermal cells, is available either for export towards the cuticle or for esterification with an acyl CoA. Other plant species exhibit large variations in compositions of cuticular wax esters, characterized in some cases by broad distributions of acyl and/or alkyl moieties and in other cases by relatively high preferences for certain isomers (Figure 4).
Figure 4. Diversity of acyl and alkyl compositions of wax esters from three plant species. Waxes were extracted from leaf surfaces and analysed by GC-MS (n = 3). The relative acyl composition for each ester chain length was determined from the abundances of MS fragments [RCO2H2]+, and used to calculate overall acyl and alkyl distributions across ester chain lengths.
Download figure to PowerPoint
In higher plants, mammals and bacteria, ester biosynthesis is catalysed by one of three classes of wax synthase (WS) enzymes: jojoba-type WS, mammalian WS, and WS/DGAT bifunctional enzymes. Jojoba-type WS uses a wide range of saturated and unsaturated acyl CoAs ranging from C14 to C24, with 20:1 as the preferred acyl and 18:1 as the preferred alcohol substrate (Lardizabal et al., 2000). In A. thaliana, there are 12 wax synthases with high homology to the jojoba WS, but none have yet been characterized. Mammalian WS enzymes do not have homologues in plants, and have highest activities with C12–C16 acyl CoAs and alcohols shorter than C20 (Cheng and Russell, 2004a,b). A bifunctional WS/DGAT enzyme from Acinetobacter calcoaceticus has a preference for C14 and C16 acyl CoA together with C14–C18 alcohols (Stöveken et al., 2005). Nearly a hundred WS/DGAT homologues have been identified from over 20 other micro-organisms so far (Wältermann et al., 2007), and ten sequences in the A. thaliana genome have also been annotated as WS/DGATs. One of these enzymes, WSD1, has been characterized and shown to be responsible for the formation of cuticular wax esters in A. thaliana stems (R.J., L.K., F. Li, X. Wu and A.L. Samuels, University of British Columbia, Canada, unpublished results). The enzyme utilizes mostly saturated C16 acyl CoA precursors, showing that this upstream precursor of wax production must be co-localized in the cell with the primary alcohols, which are synthesized far downstream in the wax biosynthetic pathway.
Two additional compound classes with predominantly even-numbered chain lengths, aldehydes and free fatty acids, are also found in the wax mixture of most plant species, albeit usually at relatively low concentrations. Currently, our knowledge on their biosynthesis is very limited. Formation of free fatty acids must involve hydrolysis of the elongated acyl CoA precursors (Figure 2). However, it is not clear whether this reaction occurs spontaneously or whether it is enzyme-catalysed. Aldehyde formation requires reduction of acyl CoA precursors, and may occur as an intermediate step during alcohol formation (see above), during alkane formation (see below), or independently of either of these pathways (Figure 2). Only a single wax aldehyde-forming reductase enzyme has been partially purified to date, and the gene encoding this enzyme has not been identified (Vioque and Kolattukudy, 1997).
A separate set of wax biosynthetic reactions is responsible for the formation of compounds with predominantly odd numbers of carbons (Figure 2). Examples of such compound classes include the alkanes, secondary alcohols and ketones that occur together in many wax mixtures and typically share similar chain-length distributions (Jetter et al., 2006). Early biochemical experiments led to a model describing the biosynthesis of these compounds as a two-stage process, with a first set of reactions transforming VLCFA precursors into alkanes and a second series of reactions modifying them into secondary alcohols and ketones (Kolattukudy, 1965). Subsequent experiments confirmed the central role of alkanes in this pathway (Kolattukudy, 1968; Kolattukudy and Brown, 1974; Kolattukudy et al., 1974), either as intermediates en route to mid-chain functionalized compounds or as end products if the downstream reactions are missing. Overall, the second stage of the pathway is relatively well characterized, whereas the first part remains poorly understood.
Although conversion of VLCFA precursors into alkanes could proceed directly in one reaction, the net acyl decarboxylation is apparently brought about by a sequence of transformations. This multistep pathway is supported by the fact that a number of different A. thaliana mutants with alkane-deficient cuticular wax mixtures have been described (Hannoufa et al., 1993; Jenks et al., 1995; Rashotte et al., 2001, 2004). Cloning of several of these mutated genes (CER1, CER2 and CER3/WAX2) revealed that the proteins they encode contain motifs similar to known biosynthetic enzymes (Aarts et al., 1995; Ariizumi et al., 2003; Chen et al., 2003; Kurata et al., 2003; Negruk et al., 1996; Rowland et al., 2007; Xia et al., 1996). While this suggests a potential enzymatic role for these proteins, their exact function remains unknown. Due to this lack of molecular information, it is currently not possible to predict the exact number of reaction steps involved in the conversion of acyl precursors into alkanes, the nature of these steps or the resulting intermediates.
Two alternative pathways have been proposed for the conversion of acyl compounds into alkanes, which vary in the central reaction in which a C1 unit is cleaved off (Bianchi, 1995; Bognar et al., 1984; Chibnall and Piper, 1934). The difference lies in the nature of the immediate precursor from which cleavage occurs and whether the C1 unit is CO or CO2 (decarbonylation versus decarboxylation). Only one model, which describes alkane formation as the decarbonylation of an aldehyde intermediate, has been tested experimentally to some extent (Cheesbrough and Kolattukudy, 1984). However, conclusive molecular genetic and biochemical evidence for either model is lacking, leaving alkane formation as the least understood part of wax biosynthesis.
In A. thaliana leaves, alkanes are the major odd-numbered product, while a high level of secondary alcohols and ketones accompanies alkanes in the stem wax, as well as in wax from B. oleracea leaves (Baker, 1974; Jenks et al., 1995). In these instances, a second stage of the pathway is additionally involved, transforming alkanes first into secondary alcohols and then into ketones (Figure 2). This reaction sequence is well-supported by chemical evidence correlating chain-length and isomer compositions of all three compound classes (Jenks et al., 1995), and by biochemical evidence provided by feeding experiments and detailed studies of label positions in resulting products (Kolattukudy and Liu, 1970; Kolattukudy et al., 1971, 1973). Recently, a reverse genetic approach led to the discovery of a cytochrome P450 enzyme that is involved in secondary alcohol and ketone formation in A. thaliana (Greer et al., 2007). The protein is a mid-chain alkane hydroxylase (MAH1) catalysing two consecutive reactions by first hydroxylating the central CH2 group of alkanes, and then probably re-binding the resulting secondary alcohol for a second hydroxylation of the same carbon. Overall, this confirms the original hypothesis that the pathway involves alkanes as central intermediates that may be further oxidized depending on plant species and organ.
Waxes from certain taxa and/or organs can also contain other compound classes (Figure 1), most prominently aromatic esters and compounds with two hydroxyl or carbonyl functions (diols, ketols, ketoaldehydes and diketones; Jetter et al., 2006). These wax constituents can be regarded as downstream or side products of the ubiquitous biosynthetic reactions forming the common product classes as described above. This implies that additional enzymes, expressed at high levels in certain plant species, can intercept intermediates and/or final products of the ubiquitous pathways before they are exported to the cuticle. As these enzymes can apparently handle the pre-formed wax compounds, they could be added in a modular fashion to the standard pathways in heterologous expression systems. This would allow stepwise addition and modification of secondary functional group(s), and substantially increase the chemical diversity of biotechnologically produced wax mixtures. The necessary biochemical and molecular genetic information on the biosynthesis of these compound classes is currently not available. However, cloning and characterization of the genes involved may become possible in the near future, once the standard wax biosynthetic pathways are better understood in A. thaliana, so that the rapidly growing genomic information from other species (Pennisi, 2007) can be further exploited.
With the isolation and characterization of a number of key genes involved in modification of VLCFA precursors into the diverse wax compound classes, important information on the intracellular localization of wax biosynthetic pathways has emerged. The site of primary alcohol formation appears to be the ER, as shown by localization of the alcohol-forming Arabidopsis enzyme CER4 after expression in yeast (Rowland et al., 2006). This is in contrast with mammalian FARs, which are associated with peroxisomes (Burdett et al., 1991; Cheng and Russell, 2004a,b), and therefore the localization of the CER4 FAR will have to be verified in planta. Meanwhile, the subsequent enzyme in the wax biosynthetic pathway, the wax ester synthase WSD1, has been localized to the ER, (R.J., L.K., F. Li, X. Wu and A.L. Samuels, University of British Columbia, Canada, unpublished results). Similarly, the mid-chain alkane hydroxylase MAH1 (CYP96A15), which catalyses the last two steps of the alkane-pathway in A. thaliana stems, is also confined to the ER (Greer et al., 2007). These two downstream pathway enzymes are thus co-localized with the VLCFA-generating FAEs (Kunst and Samuels, 2003; Xu et al., 2002; Zheng et al., 2005), and it is very likely that the entire wax biosynthesis process occurs in a single subcellular compartment. All the wax biosynthetic enzymes and precursors are therefore expected to be present in the ER of epidermal cells, leading to accumulation of all intermediates and products in the extensive membrane system of this organelle.