Moving beyond the ubiquitous: the diversity and biosynthesis of specialty compounds in plant cuticular waxes

Abstract

Cuticular waxes coat aerial plant surfaces to protect tissues against biotic and abiotic stress. The waxes are complex mixtures of fatty-acid-derived lipids formed on modular biosynthetic pathways, with varying chain lengths and oxygen functional groups. The waxes of most plant species contain C₂₆–C₃₂ alcohols, aldehydes, alkanes, and fatty acids together with their alkyl esters, and comparisons between diverse wax mixtures have revealed matching chain length distributions between some of these compound classes. Based on such patterns, the biosynthetic pathways leading to the ubiquitous wax constituents were hypothesized early on, and most of these pathway hypotheses have since been confirmed by biochemical and molecular genetic studies in model species. However, the most abundant wax compounds on many species, including many important crop species, contain secondary functional groups and thus their biosynthesis differs at least in part from the ubiquitous wax compounds with which they co-occur. Here, we survey the chemical structures of these species-specific specialty wax compounds based on a comprehensive CAS SciFinder search and then review relevant reports on wax compositions to help develop and refine hypotheses for their biosynthesis. Across the plant kingdom, specialty wax compounds with one, two, and three secondary functional groups have been identified, with most studies focusing on Angiosperms. Where multiple specialty wax compounds were reported, they frequently occurred as homologous series and/or mixtures of isomers. Among these, it is now possible to recognize series of homologs with predominantly odd- or even-numbered chain lengths, and mixtures of isomers with functional groups on adjacent or on alternating carbon atoms. Using these characteristic molecular geometries of the co-occurring specialty compounds, they can be categorized and, based on the common structural patterns, mechanisms of biosynthesis may be predicted. It seems highly likely that mixtures of isomers with secondary functions on adjacent carbons arise from oxidation catalyzed by P450 enzymes, while mixtures of isomers with alternating group positions are formed by malonate condensation reactions mediated by polyketide synthase or ketoacyl-CoA synthase enzymes, or else by the head-to-head condensation of long-chain acyls. Though it is possible that some enzymes leading to ubiquitous compounds also participate in specialty wax compound biosynthesis, comparisons between co-occurring ubiquitous and specialty wax compounds strongly suggest that, at least in some species, dedicated specialty wax compound machinery exists. This seems particularly true for the diverse species in which specialty wax compounds, most notably nonacosan-10-ol, hentriacontan-16-one (palmitone), and very-long-chain β-diketones, accumulate to high concentrations.

Introduction

Ubiquitous wax compounds

All primary above-ground plant surfaces are covered with cuticular waxes, complex mixtures of diverse very-long-chain (VLC) aliphatic and, in some cases, alicyclic or aromatic compounds. The wax mixtures on almost all plant species and organs have a basic set of components in common, which can be fairly easily identified as alkanes, primary alcohols, aldehydes, and fatty acids, and thus as compounds with either an oxygen functionality on one end of the hydrocarbon chain or no functional group at all. On some species, dimers of fatty acids and alcohols, or alkyl esters, are also found. All these ubiquitous wax constituents have fully saturated, straight-chain hydrocarbon backbone structures that vary in chain length and thus occur as homologous series.

Due to the structural variation of ubiquitous wax compounds, cuticular wax compositions are typically reported by providing two dimensions of information: the quantities of each compound class and the chain length distributions within each. These two dimensions define wax compound structural variation with respect to both the terminal carbon oxidation state (R) and the total carbon number (TCN) of each compound. For example, the major wax compounds from leaves of Arabidopsis thaliana (Busta et al. 2016b) and Brassica oleracea (Shepherd et al. 1995) are homologous series of fatty acids, aldehydes, and alcohols with even TCNs, and alkanes with odd TCNs (Fig. 1a). These characteristic patterns of predominant chain length parities in the different compound classes can be captured in a table that is defined by the two dimensions of ubiquitous wax compound structural variation (TCN and R, Fig. 2), to facilitate comparisons of structural signatures within and between the various homologous series, and to reveal patterns of co-occurrence and series characteristics.

Fig. 1

Structure and biosynthesis of ubiquitous wax compounds. a Structures of major ubiquitous wax compounds from Arabidopsis thaliana (A. th) and Brassica oleracea (B. ol). b Biosynthesis of ubiquitous wax compounds. De novo fatty acid biosynthesis (green) and fatty acid elongation (magenta) produce wax precursors (very-long-chain acyl-CoAs, boxed compounds), which wax functionalization pathways (blue) process into ubiquitous wax compounds (compounds in shadowed box). The grey background indicates that all these compounds have been found in plant waxes. KAS ketoacyl-ACP synthase, KAR ketoacyl-ACP reductase, HAD hydroxyacyl-ACP dehydratase, EAR enoyl-ACP reductase, KCS ketoacyl-CoA synthase, KCR ketoacyl-CoA reductase, HCD hydroxyacyl-CoA dehydratase, ECR enoyl-CoA reductase, RED reductase, AD aldehyde decarbonylase, EST esterase, FAR fatty acyl reductase, WS wax ester synthase. (Color figure online)

Fig. 2

Tabulation of major ubiquitous wax compounds from A. thaliana and B. oleracea. Columns indicate the total carbon number (TCN) and rows indicate the terminal carbon oxidation state (R) of the ubiquitous wax compounds. An entry of A. th (A. thaliana) or B. ol (B. oleracea) in a cell indicates the presence of that compound in that species’ wax

Comparisons of the wax profiles from diverse species have consistently shown that the TCNs of adjacent homologs in each homologous series virtually always differ by two, and that acids, aldehydes, and alcohols have predominantly even TCN parity, while alkanes have predominantly odd TCNs (e.g., Fig. 2). Based on these characteristic patterns, it was hypothesized early on that the biosynthetic pathways leading to the ubiquitous wax constituents may involve (1) specific chain length elongation mechanisms to generate homologous series, and (2) various head group modification reactions leading to different compound classes (Kolattukudy 1970).

First, wax elongation mechanisms were hypothesized based on the finding that the TCN parities of cuticular fatty acids are highly reminiscent of those of long-chain fatty acids (C_≤18). It was well established that the latter are formed by sequential addition of C₂ units (derived from malonyl-CoA) catalyzed by plastidial fatty acid synthase (FAS) complexes (Fig. 1b, green pathway), and so it seemed plausible that formation of VLC (C_>20) fatty acids might proceed by further addition of C₂ units catalyzed by analogous enzyme complexes. Biochemical and molecular genetic investigations have since confirmed this hypothesis. Specifically, C₁₈ fatty acyl-ACPs from plastidial de novo biosynthesis are hydrolyzed and transferred to the endoplasmic reticulum, where they are converted into acyl-CoAs and elongated by fatty acid elongase (FAE) complexes (Yeats and Rose 2013). In these, acyl-CoAs are initially condensed with malonyl-CoA by a ketoacyl-CoA synthase (KCS; Millar and Kunst 1997), and then in three further reactions the 3-keto function is removed to yield an acyl-CoA with two carbons more than the original KCS substrates (Fig. 1b, magenta pathway): a ketoacyl-CoA reductase (KCR) transforms the initial 3-ketoacyl-CoA intermediate into a 3-hydroxyacyl-CoA (Beaudoin et al. 2009), a 3-hydroxyacyl-CoA dehydratase (HCD) then converts it into an enoyl-CoA (Bach et al. 2008), which is then saturated by an enoyl-CoA reductase (ECR) to yield a saturated fatty acyl-CoA (Zheng et al. 2005). The FAE reaction cycle can be repeated such that the C₁₆ fatty acyl-CoA precursors are elongated to various even-numbered chain lengths ranging from C₂₂ to C₃₈ or more, which then serve as wax precursors.

Next, head group modification reactions were deduced from similarities between chain length patterns of wax compound classes within and across species. On one hand, it had been observed that series of aldehydes and alkanes in the same wax mixture frequently have similar homolog distributions that are offset by one carbon, such that an aldehyde series with a peak at C_n frequently accompanies an alkane series of similar shape with a peak at C_n−1 (Fig. 2). This suggested that alkanes might be biosynthesized from aldehydes via head group removal, while the even TCNs of the aldehydes suggest they may be formed by reduction of acyl-CoAs generated by FAE elongation. On the other hand, it was found that the homolog distributions of primary alcohols and fatty acids (FAs) differ from one another and from those of aldehydes and alkanes, leading to the hypothesis that they are derived from wax precursors via two additional pathways (Chibnall and Piper 1934). Finally, the co-occurrence of esters and primary alcohols and the match between chain length distributions of alkyl ester-bound primary alcohols and free primary alcohols in the few species investigated in detail so far suggested that alkyl esters likely arise from esterification of free primary alcohols (Lai et al. 2007; Shepherd et al. 1995).

The biosynthetic relationships between compound classes, derived from wax profile analyses, have been largely confirmed by molecular biological experiments in model plant species. It is now established that, in most species, at least two parallel head group modification pathways operate subsequent to elongation. On the alkane-forming pathway, a reductase (RED) converts fatty acyl-CoA precursors into aldehydes that are then available to an aldehyde decarbonylase (AD) for the production of alkanes (Bernard et al. 2012; Fig. 1b, upper blue pathway). In the parallel acyl reduction pathway, a fatty acyl-CoA reductase (FAR) generates primary alcohols (Rowland et al. 2006) that are then available to a wax ester synthase (WS) for the assembly of alkyl esters (Li et al. 2008; Fig. 1b, lower blue pathway). Although the homolog distributions of ubiquitous wax compounds strongly suggest the existence of a separate fatty acid-forming pathway, no enzymes involved in this process have been identified. However, it seems likely that a (thio)esterase (EST) may catalyze such reactions (Fig. 1b, middle blue pathway).

In summary, the biosynthesis of both ubiquitous wax compound aliphatic chains and terminal functional groups was first hypothesized based on the wax profiles of multiple species by looking for patterns of co-occurrence and comparing structural characteristics within and between homologous series. All the major hypotheses were then confirmed by biochemical and, eventually, by molecular biological investigations in model species, demonstrating that carbon chain formation by fatty acid elongation precedes head group modification. Thus, FAEs generate fatty acyl-CoAs with even TCNs that serve as precursors for the functionalization pathways to produce the ubiquitous wax compounds that are found on the surfaces of almost all plant species.

Specialty wax compounds

Many species of diverse lineages in the dicots, monocots, gymnosperms, ferns, and mosses not only bear the ubiquitous compounds described above, but also wax compounds exhibiting further structural diversity. While some specialty compounds can be isoprenoid in nature (Jetter et al. 2006) or have branches in their hydrocarbon chains (Busta and Jetter 2017), they are most commonly found as VLC linear aliphatics with one or more secondary oxygen functional groups instead of or in addition to a terminal functional group. These specialty wax compounds vary in secondary functional group position and oxidation state, thus adding dimensions of structural diversity beyond TCN and R, and resulting in substantially greater structural variation than is present among ubiquitous wax compounds. An attempt has been made to describe the diversity of specialty wax constituents (Jetter et al. 2006), but such has not been done comprehensively with systematic chemical database searching tools.

In some reports on specialty wax compound structures, especially when waxes of several species or plant lines were compared, authors put forward hypotheses on possible biosynthetic pathways leading to these compounds. It was suggested that the secondary functionalities may, depending on the exact structure of the compound, be installed before, during, or after the biosynthesis of the other portions of the molecule. For example, it was inferred that ketones in the Brassicaceae might be formed by oxygenation of ubiquitous alkanes, after elongation and head group modification (Kolattukudy et al. 1968), while other ketones occurring in various dicot species may be generated by head-to-head condensation of acyl compounds, effectively bypassing head group modification (Jetter 2000). In contrast, mid-chain β-diketones were recognized as polyketides, leading to the prediction of polyketide synthase (PKS) enzyme activity before or during elongation (Schulz et al. 2000; von Wettstein-Knowles 1995). Finally, compounds with regioselectively placed secondary functional groups in ferns and mosses were surmised to result from elongation of acyl-CoAs retaining in-chain functional groups from elongation intermediates (Busta et al. 2016a; Jetter and Riederer 2000).

However, only two of these scenarios have been corroborated by experimental evidence, and only in select species: (1) the oxygenation mechanism leading to ketones has been confirmed by molecular biological investigations, showing that a P450 enzyme, MAH1, in Arabidopsis hydroxylates alkane precursors to secondary alcohols, ketones, diols, and ketoalcohols (Greer et al. 2007), and (2) earlier biochemical and very recent genetic evidence show that cuticular β-diketones are formed by hydrolytic intercept of elongation intermediates and probably further elongation involving a PKS enzyme (Hen-Avivi et al. 2016; Schneider et al. 2016). However, it is not clear to which extent either of these mechanisms may account for the formation of the diverse specialty compound structures reported so far. Furthermore, neither of these mechanisms has been critically evaluated in the context of specialty wax compound structures from across the plant kingdom.

Specialty wax compounds have thus far been found to always co-occur with ubiquitous wax compounds and share some structural characteristics with them. Accordingly, it is also possible to use structural comparisons to evaluate potential overlap of the pathways leading to these two groups of molecules. Published models for specialty wax compound biosynthesis described above all implied that ubiquitous and specialty wax compounds are derived from the same or similar biosynthetic machinery. However, no large-scale, comprehensive comparisons of this kind had been performed, and how such comparisons may influence specialty compound biosynthesis hypotheses is unclear.

To summarize the progress that has been made in the area of specialty wax compounds and assist in identifying targets for future research, the present work aims to generate a catalog of the specialty compounds reported thus far (“Occurrence and diversity of specialty wax compounds” section), and to assess patterns of co-occurrence and the structural characteristics of these compounds to critically evaluate and further develop models for their biosynthesis (“Biosynthesis of specialty wax compounds” section).

Occurrence and diversity of specialty wax compounds

The diversity of specialty wax compounds reported thus far was assessed by querying molecular formulas in the Chemical Abstracts Service SciFinder database for reports of mono-, bi-, tri-, and tetra-functional linear aliphatic compounds with 17–52 carbons. After results from non-biological sources were filtered out, the resulting preliminary list of ca. 350 references was inspected, and reports describing compounds from animals (mostly insects), bacteria (usually < C₂₀), or geological samples were eliminated. A few reports describing VLC compounds with unsaturation(s) in the carbon chain, but without secondary oxygen functional groups, were also not considered further. After basic filtering, ca. 125 references as remained that together described the surface waxes of ca. 100 plant species bearing specialty wax compounds. A few representative reports of insect lipids and plant essential oils containing compounds with structures similar to plant specialty wax compounds were added to the list as context on the wider occurrence of specialty wax-like compounds. With this addition, the final list of reports that are reviewed in this work was obtained (Table S1).

To define and evaluate the chemical space occupied by the specialty compounds from across all selected references, tables were laid out according to specialty compound structural characteristics, which were then populated with (qualitative) data on specialty compound occurrence. Three sets of tables were necessary: one set each for compounds with one, two, or three secondary functional groups. Within each set, separate tables were dedicated to compounds with frequently observed constellations of functional groups. Within each table, each column corresponds to a TCN. Conversely, rows were grouped by oxidation states of the terminal carbon and of the secondary function, and rows within groups sorted according to the position (carbon number, C-n) of the secondary functional group(s). For example, if a species’ wax mixture contains a homologous series of C₂₄, C₂₆, C₂₈, and C₃₀ 3-ketoacids, a mixture of isomeric C₂₄ ketoacids with 1,6-, 1,7-, 1,8-, and 1,9-geometry, and a mixture of isomeric C₂₈ primary,secondary diols with 1,7-, 1,9-, 1,11-, and 1,13-geometry, the table for specialty wax compounds with one secondary functional group would be subdivided according to these structural features and would receive twelve total entries of “G.sp” (G.sp = Genus species) in cells corresponding to the structural characteristics of these compounds (Fig. 3).

Fig. 3

Example tabulation of hypothetical specialty wax compounds. For each specialty wax compound found in a species, its initials (e.g. Genus species = “G.sp”) are entered once into the cell corresponding to the structural characteristics of that specialty wax compound. Columns indicate the total carbon number (TCN). Rows indicate functional group positions (C–n) and are grouped by terminal carbon oxidation state (R). Entries are color-coded to highlight members of the same homologous series or mixture of isomers. (Color figure online)

Because the interpretation of the dataset would largely rely on identifying commonalities between homologous series and mixtures of isomers that co-occurred on the same species as well as those that recurred between taxa, all potential patterns in functional group position needed to be explored. To facilitate these comparisons, we classified each homologous series and mixture of isomers with two criteria using the positions (carbon numbers, C-n) of the secondary functional groups across that whole series or mixture: (1) according to whether the secondary functional groups were on both even- and odd-numbered carbons (EO), only on even-numbered carbons (E), or only on odd-numbered carbons (O) (Fig. 4a, function parity); and (2) according to whether functional groups appeared near the terminus of the molecules (C-n, n ≤ 3, (a)), or in the central portion of the molecules (C-n, n > 3, (b)) (Fig. 4a, function position). Homologous series or mixtures of isomers whose members had more than one functional group were described by categorizing each functionality according to (1) and (2) (Fig. 4b).

Fig. 4

System of color coding and nomenclature for specialty wax compounds. Homologous series and isomeric mixtures of specialty wax compounds reported in the literature are distinguished based on the positions of their secondary functional groups. Different structures within each box in a and b show different isomers, with –CH₂–CH₂– units in brackets representing chain building blocks occurring once or repeatedly to form homologous compounds. a Isomeric mixtures of specialty compounds were first distinguished according to whether they had functional groups on carbons with mixed or single parity (functionalities on even- and odd-numbered carbons, e.g. C-14, C-15, C-16 (EO); functions on only odd-numbered carbons; e.g. C-7, C-9, C-11 (O); or functionalities on only even-numbered carbons, e.g. C-4, C-6, C-8 (E). Next, series or mixtures were distinguished according to whether their members’ secondary functional groups were present near the terminus (C-n, n ≤ 3 (a)) or towards the middle (C-n, n > 3 (b)) of the aliphatic carbon chain. Colors were assigned to each category distinguished by these criteria, and hypothetical example isomers (on one side of each dotted line) and homologs (on opposite sides of each dotted line) are shown for each category. b For compounds that contained more than one secondary functional group, nomenclature and coloration was assigned to each function independently. R = COOR′, COOH, CHO, CH₂OH, CH₃. c Abbreviations used here and in subsequent figures. (Color figure online)

Homologous series or mixtures of isomers from the same species were highlighted with common color-coding in the tables. Thus, in the example of the ketoacids co-occurring on “Genus species” (Fig. 3), each cell containing entries corresponding to compounds from the same series or mixture was given the same background color to distinguish the three different series or mixtures of isomers as O(a), EO(b), and O(b) compounds, respectively (Fig. 3). All possible constellations of functional groups within co-occurring compounds were thus categorized and highlighted in the summary tables. In the following, the resulting structural patterns will be briefly described, proceeding from those with one secondary functional group (“Specialty compounds with a single secondary oxygen functional group” section) to those with two functionalities (“Specialty compounds with two secondary functional groups” section) and then those with three or more (“Specialty compounds with three or more secondary functional groups” section).

Specialty compounds with a single secondary oxygen functional group

From several species, homologous series of specialty compounds have been reported with one secondary oxo or hydroxyl function on C-2 of their carbon chains, thus E(a) compounds, with predominantly odd TCNs between C₁₇ and C₃₇. Such series have been encountered as surface wax components on the flowering plants Aloe arborescens (Racovita et al. 2014) and Solanum tuberosum (Szafranek and Synak 2006) (Fig. 5a, light green cells). Single compounds with similar structural characteristics (odd TCN, secondary group on C-2) have also been found on, for example, the angiosperms Laurus nobilis (Garg et al. 1992), and Prunus domestica (Mahmood et al. 2009), in esterified form on the gymnosperm Pinus radiata (Franich et al. 1985), in many plant essential oils (C₁₇ and C₁₉; Fig. 5a, Many O), as aphrodisiac pheromones on the wings of Pieris butterflies (C₂₉; Yildizhan et al. 2009), as nest identity signals from Bombus terrestris bumblebees (C₂₉; Rottler et al. 2013), and from many other insect sources (Fig. 5a, Many I). Such single compounds, without accompanying isomers/homologs, can only tentatively be classified as E(a) compounds.

Fig. 5

Catalog of specialty wax compounds with one secondary functional group. A species’ initials (e.g. Genus species = “G. sp”) are entered once for each specialty wax compound found on that species in the cell corresponding to the structural characteristics of that specialty wax compound. Columns indicate the total carbon number (TCN). Rows indicate secondary functional group position (C-n) and are grouped by terminal carbon oxidation state (R). Entries are color-coded to highlight members of the same homologous series or mixture of isomers. a Compounds with CH₃ terminal carbons. b Compounds with oxygen-bound terminal carbons. c Color coding and nomenclature. Compounds marked with an asterisk were reported in esterified form and those with two asterisks had a methoxy secondary functional group. Note that row numbering is not necessarily continuous. (Color figure online)

Further specialty wax compounds with a single secondary oxo or hydroxyl group on secondary carbons other than C-2 have been found as co-occurring isomers with functionalities on even- and odd-numbered carbons, thus EO(b) compounds, for example as mixtures of nonacosan-6-one, nonacosan-7-one, nonacosan-8-one, nonacosan-9-one, and nonacosan-10-one. These compounds typically had odd TCNs spanning C₂₃–C₃₅, occurring as homologous series and/or isomer mixtures in the surface waxes of the angiosperms Arabidopsis thaliana (Hannoufa et al. 1993; McNevin et al. 1993; Wen and Jetter 2009), Berberis aquifolium (Berg 1985), several Brassica species¹ (Eigenbrode and Pillai 1998; Laseter et al. 1968; Netting and Macey 1971; Shepherd et al. 1995), Exochorda racemosa (Holloway et al. 1976), several Fragaria spp. (Baker and Hunt 1979), Malus domestica (Verardo et al. 2003), Pisum sativum (Macey and Barber 1970), Rosa virgo (Mladenova et al. 1977), and Triticum aestivum (Racovita and Jetter 2016a; Fig. 5a, light orange cells). Compounds with the same structural characteristics have also been found on several insect surfaces, including butterfly (C₂₇–C₃₁; Yildizhan et al. 2009) and wasp wings (C₂₃; Howard and Baker 2003).

Other co-occurring monofunctional specialty wax compounds bore a secondary oxo or hydroxyl group located almost exclusively on even-numbered carbons near the middle of the chain. Such E(b) compounds were found, for example, as a mixture of hentriacontan-8-one, hentriacontan-10-one, hentriacontan-12-one, hentriacontan-14-one, and hentriacontan-16-one. Members of this class usually had odd TCNs between C₂₇ and C₃₅, as found for Rosa canina (Buschhaus et al. 2007), Solanum tuberosum (Szafranek and Synak 2006), and the ferns Osmunda regalis (Jetter and Riederer 1999a) and Azolla filiculoides (Brouwer et al. 2015) (Fig. 5a, light blue cells).

In some cases, compounds with secondary functional groups on even-numbered carbons were found, but no accompanying homologs or isomers were reported to occur with them and, thus, they cannot be assigned to any of the categories based on predominant mixture structural characteristics. Two prominent examples include: (1) nonacosan-10-ol and/or nonacosan-10-one from Clematis vitalba (Ulubelen 1970), Euonymus latifolius (Ulubelen and Baytop 1973), Foeniculum vulgare (Muckensturm et al. 1997), Fumaria parviflora (Naz et al. 2013), Nelumbo nucifera (Ensikat et al. 2011), Prunus avium (Peschel et al. 2007), Rosa canina (Buschhaus et al. 2007), and various gymnosperms including Picea omorika (Nikolić et al. 2013) and Pinus spp. (Günthardt-Goerg 1986), and (2) hentriacontan-16-one (palmitone) and other symmetrical ketones and secondary alcohols found with mainly odd TCNs (C₁₇–C₃₁) such as those on surfaces of Allium porrum (Rhee et al. 1998), Annona squamosa (Shanker et al. 2007), Ginkgo biloba, Magnoflia grandiflora, and Liriodendron tulipifera (Gülz et al. 1992), Hibiscus rosasinensis (Siddiqui et al. 2005), Hymenocallis littoralis (Abou-Donia et al. 2008), Platanus orientalis (Dhar and Munjal 1976), Thesium humile (Gharbo et al. 1962), Tridax procumbens (Verma and Gupta 1988), and Vaccinium ashei (Freeman et al. 1979) (Fig. 5a, white cells).

Numerous wax constituents with one secondary functional group and one terminal functional group have been identified. These occur with vast diversity owing to variation of oxidation states possible for the terminal group. Head groups defining fatty acids, methyl esters, primary alcohols, and aldehydes have all been reported in combination with secondary methoxy, oxo, and hydroxyl groups. Some of these compounds were found as mixtures of compounds with both 1,2- and 1,3-functional group geometry, thus EO(a) secondary functionalities, as diols from Cosmos bipinnatus flower wax (Buschhaus et al. 2013; Fig. 5b, light yellow cells).

More frequently, compounds with one terminal and one secondary functional group have been reported as mixtures of homologous structures with 1,3-functional group geometry and without accompanying isomers, thus O(a) functionalities, with almost exclusively even TCNs ranging from C₂₀ to C₃₂. These have been encountered in waxes of the angiosperms Aleo arborescens (Racovita et al. 2014), several Papaver spp. (Jetter et al. 1996), Ricinus communis (Vermeer et al. 2003), and in esterified form on the moss Funaria hygrometrica (Busta et al. 2016a), (Fig. 5b, light green cells). Single homologs of these compounds have been reported from, for example, Pegolettia senegalensis (Bohlmann et al. 1983), several Salvia spp. (Agar et al. 2008), the seed oil of Wrightia tinctoria (C₁₈; Nagalakshmi and Murthy 2015), the volatiles of aerial organs of Minuartia recurva (C₁₈; Jovanović et al. 2009) and Melicocca bijuga fruit (C₁₈; Pino et al. 2002), and neutral lipids of the microalga Chlorella sorokiniana (C₁₈; Yang et al. 2013).

Other mixtures of specialty compounds with one terminal and one secondary function had the latter on C-5, C-7, C-9 etc., thus secondary functional groups on only odd-numbered carbons some distance into the chain, and were therefore categorized as O(b) compounds. These almost always had even TCNs C₂₄–C₃₆, and were found on surfaces of the angiosperms Cerinthe minor (Jetter and Riederer 1999b), Myricaria germanica (Jetter 2000) and several Papaver spp. (Jetter et al. 1996), the gymnosperm Taxus baccata (Wen and Jetter 2007), the ferns Osmunda regalis (Jetter and Riederer 1999a) and Azolla filiculoides (Brouwer et al. 2015), and the moss Funaria hygrometrica (Busta et al. 2016a) (Fig. 5b, light blue cells). These compounds have also been found in algae, including Schizymenia dubyi (C₂₂ and C₂₄; Barnathan et al. 1998) and Nannochloropsis spp. (C₃₂ and C₃₄; Gelin et al. 1997), as well as Argemone mexicana seed oil (C₂₈ and C₃₀; Gunstone et al. 1977).

In two cases, wax constituents with one terminal function and a secondary function on either even- or odd-numbered carbons co-occurred, resulting in mixtures of compounds with 1,11-, 1,12-, 1,13-, and 1,14-geometries, thus EO(b) secondary functionalities. These compounds have been reported as C₂₈ diols from Pisum sativum (Wen et al. 2006a), and as esterified C₂₆–C₃₀ diols from Triticum aestivum (Racovita and Jetter 2016a; Tulloch 1971) (Fig. 5b, dark orange cells).

Thus, the structural diversity of monofunctional specialty wax compounds is considerable. Three types of compounds have been encountered most frequently: (1) C₂₇ and C₂₉ 14- and 15-secondary alcohols and ketones, (2) nonacosan-10-ol and nonacosan-10-one, and (3) hentriacontan-16-one. All three have frequently been found accumulating to high concentrations and predominating in their respective wax mixtures. Accordingly, many of these compounds form epicuticular wax crystals that bestow special properties upon the tissue surfaces to serve a wide range of biological functions (Koch and Ensikat 2008). Notably, these specialty wax compounds have been identified in diverse angiosperm species, but only sporadically in earlier-diverging lineages (Fig. 6). This suggests that the machinery required to make such specialty wax compounds, with functional groups in the central portion of the molecule, may be present only in angiosperms and perhaps, if it evolved fairly recently, only in certain angiosperm subgroups. In contrast, specialty wax compounds with functional groups near the chain terminus (i.e., O(a) or E(a) geometry) have been found in plants from nearly every major group of land plants, indicating that the machinery required to make these compounds was probably present in a common ancestor of land plants. Furthermore, multiple investigations of algae have found compounds with the same structural characteristics as specialty wax compounds, suggesting that the machinery required for specialty wax compound biosynthesis may have been present in plant predecessors even before the colonization of land, and that such machinery may have been recruited from other areas of metabolism to assist in the formation of the plant cuticle.

Fig. 6

Distribution of specialty wax compounds across plant orders. Phylogeny (Stevens 2001) showing plant orders in which specialty compounds have been found is listed on the left side of the matrix, and the classes of specialty compounds are listed across the top of the matrix. Each filled cell indicates that one or more specialty compounds belonging to that class have been observed in that order. In cases where compound(s) were not present as a homologous series and/or mixtures of isomers, secondary functional group isomer characteristics could not be fully determined, and such compounds could only be assigned tentatively to a class of specialty compounds. a includes occurrences of compounds with both confirmed and tentative class assignments. b shows only compounds with confirmed assignments

Specialty compounds with two secondary functional groups

Some wax mixtures contain specialty compounds with two secondary oxo or hydroxyl groups. In certain species, these mixtures comprised isomers with mid-chain functional groups in α- and β-constellations: compounds with functional groups both on adjacent carbons and separated by one CH₂ group, respectively. Thus, these are mixtures of isomers with functionalities on both odd- and even-numbered carbons (e.g., 15-hydroxynonacosan-14-one co-occurring with 16-hydroxynonacosan-14-one). Accordingly, these mixtures are best described as having EO(b)EO(b) functionalities (instead of distinguishing two sub-groups as O(b)O(b) and E(b)E(b), respectively). All the bifunctional compounds in respective mixtures usually had odd TCNs between C₂₇ and C₃₁, as found for waxes on stems of Arabidopsis thaliana (Wen and Jetter 2009), leaves of several Brassica spp. (Holloway and Brown 1977; Shepherd et al. 1995), and the wings of Pieris butterflies (Yildizhan et al. 2009) (Fig. 7a, pink cells).

Fig. 7

Catalog of specialty wax compounds with two secondary functional groups. A species’ initials (e.g. Genus species = “G.sp”) are entered once for each specialty wax compound found in that species and in the cell corresponding to the structural characteristics of that specialty wax compound. Columns indicate the total carbon number (TCN). Rows indicate functional group positions (C-n) and are grouped by terminal carbon oxidation state (R). Entries are color-coded to highlight members of the same homologous series or mixture of isomers. a Homologous series or mixtures of isomers with CH₃ terminal carbons, grouped according to whether their members’ secondary functional groups were found on alternating and adjacent carbons (α- and β-geometries) or only on alternating carbons (β-geometry). b Homologous series or mixtures of isomers with CH₃ terminal carbons and functional groups in varying positions. Filled spade: denotes the ketol grouping (i.e., one secondary hydroxyl and one secondary oxo group), while filled club: denotes the diketone grouping (two secondary oxo groups). c Homologous series or mixtures of isomers with primary functional groups and two secondary functional groups in varying positions. Filled heart: denotes the secondary/secondary diol grouping, filled diamond: denotes the ketol grouping. d Color coding and nomenclature. Asterisks within the tables indicate that the compound(s) were reported in esterified form. Note that row numbering is not necessarily continuous. (Color figure online)

Other wax mixtures contain similar specialty compounds with two secondary groups (oxo or hydroxyl), but with all isomers present having functional groups separated by one carbon (in beta-constellations) and almost exclusively on even-numbered mid-chain carbons, and thus E(b)E(b) functional groups. These compounds spanned a wide range of mainly odd TCNs from C₁₉ to C₃₇, occurring in fairly diverse taxa such as Buxus sempervirens (Dierickx 1973; Meusel et al. 2000), Carthamus tinctorius (Akihisa et al. 1994, 1997), Chrysanthemum segetum (Meusel et al. 2000), Eucalyptus globulus (Osawa and Namiki 1988; Steinbauer et al. 2009, 2004), Myricaria germanica (Jetter 2000), Nicotiana tabacum (Matsuzaki and Koiwai 1988), several Rhododendron spp. (Evans et al. 1975), and Taraxacum officinale (Guo et al. 2017) (Fig. 6a, purple cells). The wax mixtures on select organs of various grass species also contain these compounds, including Agropyron spp. (Tulloch 1976a, 1983), Andropogon hallii and Andropogon scoparius (Tulloch and Hoffman 1979), and Hordeum vulgare (Jackson 1971; Mikkelsen 1979). Select homologs or isomers from these series have been found on, for example, Eragrostis curvula (Tulloch 1982), Elymus cinereus (Tulloch and Hoffman 1977), Leymus arenarius (Meusel et al. 2000), Panicum virgatum (Tulloch and Hoffman 1980), Triticale hexaploide (Tulloch and Hoffman 1974), Triticum durum and Triticum aestivum (Cervantes et al. 2002; Racovita and Jetter 2016b; Tulloch and Hoffman 1971; Tulloch and Weenink 1969). Similar compounds were also found in pollen lipids of Helianthus annuus (Schulz et al. 2000; Ukiya et al. 2003).

There are also some reports of mixtures of bifunctional compounds with one secondary functional group always on the same, even-numbered carbon, and another group in greatly varying positions on both even- and odd-numbered carbons (Fig. 7b), and thus mixed functional groups E(b)EO(b). These were found as C₂₉ diols on Juniperus scopulorum (Tulloch and Bergter 1981), Myricaria germanica (Jetter 2000), Nelumbo nucifera and Tropaeolum majus (Koch et al. 2006), Thalictrum flavum (Jetter and Riederer 1995), Picea abies (Percy et al. 2009), Pinus radiata (Franich et al. 1979; Günthardt-Goerg 1986), and Taxus baccata (Wen et al. 2006b), and as diols, ketoalcohols, and diketones on Osmunda regalis (Jetter and Riederer 1999a, 1999b, blue-to-orange cells). In Myricaria germanica leaf wax, isomeric C₃₁ diols with similar secondary functional group geometries were identified (Jetter 2000). It should be noted that the compounds with E(b)EO(b) functional groups were recognized in the plant wax mixtures listed above only based on careful comparisons of all co-occurring homologs and isomers, to unambiguously distinguish them from mixtures restricted to E(b)E(b) functionalities.

Further examples of VLC aliphatic compounds with two secondary functional groups occur in lipid mixtures other than cuticular waxes. A C₃₆ diketoacid with secondary keto groups on odd-numbered carbons of backbones with even TCNs C₂₆ and C₂₈ (E(b)E(b)) were identified in Helianthus annuus pollen (Schulz et al. 2000; Fig. 7c). In addition, dihydroxy acids with hydroxyl groups on odd-numbered carbons of C₂₀, C₂₂, and C₂₄ backbones (O(b)O(a)) were detected in the zygospore cell walls of Chlamydomonas monoica (Blokker et al. 1999), the floral oils of Malpighia spp. (Seipold et al. 2004), and the floral oils collected by bees (Reis et al. 2007) (Fig. 7c, blue-to-green cells).

Overall, specialty wax compounds with two secondary functional groups have been found in various forms and on many species. The most frequently reported among these are the α-ketoalcohols from the Brassicas and the β-diketones from grasses and several dicot families. Both of these compound classes have also been shown to participate in the formation of epicuticular wax crystals that create superhydrophobic surfaces, which may in part explain why they are found on the surfaces of diverse plant species (Koch and Ensikat 2008). So far, both these classes of compounds have only been found in angiosperm lineages, suggesting relatively recent evolution. We note that, while β-diketones are very common on surfaces of Poaceae species, they are found in diverse other angiosperm taxa, suggesting either a common ancestor or that pathways to these compounds have evolved several times.

Specialty compounds with three or more secondary functional groups

There have been some reports of specialty wax compounds with three or more functional groups. Hydroxy- and oxo-β-diketones have been found, usually alongside β-diketones, with the β-diketo functionality almost always on even-numbered carbons and the additional hydroxyl or oxo group on various even- or odd-numbered carbons, and thus an overall E(b)E(b)EO(b) constellation. These compounds have been reported, with odd TCNs C₃₁ and C₃₃, from leaves of Agropyron spp. (Tulloch 1976a, b, 1983), Andropogon hallii and Andropogon scoparius (Tulloch and Hoffman 1979), Elymus cinereus (Tulloch and Hoffman 1977), seeds of Hordeum vulgare (Heinzen and Moyna 1988), Triticum aestivum (Racovita et al. 2016) (Fig. 8), and root wax of Corallocarpus epicaeus (Tulloch 1982). Single homologs or isomers belonging to the same series have been reported from Eucalyptus globulus (Osawa and Namiki 1988), Festuca ovina (Tulloch and Hoffman 1977), and Triticale hexaploide (Tulloch and Hoffman 1974). A structurally related tetrafunctional compound, 4-hydroxy-25-oxohentriacontane-14,16-dione, has been reported from Agropyron elongatum (Tulloch 1983).

Fig. 8

Catalog of specialty compounds with three secondary functional groups. A species’ initials (e.g. Genus species = “G.sp”) are entered once for each specialty wax compound found in that species and in the cell corresponding to the structural characteristics of that specialty wax compound. Columns indicate the total carbon number (TCN). Rows indicate third functional group position (C-n) and oxidation state and are grouped by terminal carbon oxidation state (R). Entries were color-coded to highlight members of the same homologous series or mixture of isomers. Colored cells in this figure indicate compounds with E(b)E(b)EO(b) functional groups. Note that row numbering is not necessarily continuous. (Color figure online)

Overall, not very many examples of specialty wax compounds with three or more functional groups have been encountered so far. However, hydroxy-β-diketones have been reported many times, and the structural diversity within that compound class seems to be relatively well documented.

Biosynthesis of specialty wax compounds

The second goal of this work was to review the structural characteristics of the specialty compounds tabulated above in the context of their biosynthesis, either to identify characteristics of known pathways generating the structures, or to derive pathway hypotheses for compounds of unknown biosynthetic origin. For this, the positions and oxidation states of functional groups on (co-occurring) specialty wax compounds, in comparison with co-occurring ubiquitous wax compounds, were analyzed to infer biosynthetic relationships between compound classes. Here, our discussion focuses on compounds carrying secondary functional groups on carbons of mixed parity (EO, “Specialty compound classes with secondary functional groups on even- and odd-numbered carbons” section), compounds bearing a single secondary functional group on a carbon of single parity near the chain terminus (O(a) and E(a), “Specialty compounds with functional groups on carbons of single parity near the chain terminus” section), compounds with a single secondary functional group on a carbon of single parity near the middle of the carbon skeleton (O(b) and E(b), “Compound classes with a single secondary functional group near the middle of the chain” section), and finally those that have two secondary functional groups on carbons of single parity near the middle of the carbon skeleton (O(b)O(b) and E(b)E(b), “Compound classes with two secondary functional groups of single parity near the middle of the chain” section). We also identify opportunities for future experiments and targeted analyses that will evaluate the described biosynthesis hypotheses.

Specialty compound classes with secondary functional groups on even- and odd-numbered carbons

Homologous series or isomer mixtures of compounds bearing a single EO(b) functional group (mid-chain secondary alcohols and ketone, cf. Figs. 5a and 6a) were accompanied by compounds with EO(b)EO(b) functional groups (co-occurring α- and β-isomers of diols, ketoalcohols, and diketones) for example in waxes of Arabidopsis thaliana (McNevin et al. 1993), Brassica napus and Brassica oleracea (Eigenbrode and Pillai 1998). Such co-occurrence indicates that the process(es) by which these two classes are biosynthesized are probably related and not highly regioselective. It was suggested early on that P450 enzymes with limited regiospecificity might be responsible for installation of the hydroxyl groups (Franich et al. 1979).

Since that suggestion, genetic tools have been used to investigate the biosynthesis of compounds with EO(b)EO(b) and EO(b) functional groups from Arabidopsis thaliana. A P450 enzyme, MAH1, was identified that produces secondary alcohols (compounds with EO(b) groups) from ubiquitous alkanes and can also perform further oxidation reactions to generate ketones, α- and β-diols and ketoalcohols (compounds with EO(b)EO(b) groups; Greer et al. 2007; Wen and Jetter 2009). Another P450 enzyme, CYP703A2, was identified as a component of cutin biosynthesis that also catalyzes hydroxylation of one of several adjacent mid-chain carbons on its substrate (Morant et al. 2007). Together, these genetic and biochemical studies show that swarms of positional isomers with functional groups on adjacent carbons can be formed by P450 enzymes with relatively poor regioselectivity using lipid substrates. On the other hand, they also confirm the original hypothesis first predicting enzyme activity based on chemical data (Franich et al. 1979), underlining the predictive power of our present analysis of common functional group patterns.

Given the widespread occurrence of biosynthetic P450s in diverse plant species, including many known to use lipid substrates, it seems plausible that P450 enzymes similar to MAH1 also install functional groups in other wax specialty compounds. We therefore expand the original hypothesis, first referring only to secondary alcohols, to predict that isomer mixtures with single or two hydroxyl or oxo groups on mid-chain carbons of mixed parity (EO(b) or EO(b)EO(b)) are products of P450 enzymes (Fig. 9a, orange functional groups), and that such functional groups may occur alone (secondary alcohols, ketones) or in pairs (secondary/secondary diols and sec./sec. ketoalcohols). Based on the database of specialty wax compounds presented here, functionalities likely produced by P450 enzymes are present in various combinations with other functional groups located either on the chain terminus (primary/secondary diols, prim./sec. ketoalcohols, prim./sec. ketoaldehydes; Fig. 9a, orange functional groups) or within the chain (sec./sec. diols, hydroxy-β-diketones; Fig. 9a, orange functional groups).

Fig. 9

Biosynthesis of specialty wax compounds with functional groups on carbons of mixed parity. Biosynthesis of homologous series or isomeric mixtures of very-long-chain specialty wax compounds whose members’ secondary functional groups are on carbons of mixed parity (both even- and odd-numbered carbons) has been demonstrated to be carried out by a P450 enzyme in Arabidopsis. In other species, such enzymes may also install functional groups on specialty compounds with similar structural characteristics, leading to secondary alcohols, ketones, diols, ketoalcohols, or diketones with functional groups on carbons of mixed parity in the center of the molecule (EO(b) and EO(b)EO(b)) and: a without primary functional group, b with primary functional group, and c in the presence of one (E(b)EO(b)) or two (E(b)E(b)EO(b)) additional functional groups found on carbons with single parity. d Biosynthesis of series or mixtures of compounds with secondary functional groups on carbons of mixed parity near the chain terminus may proceed via P450 oxidation (top), or possibly by hydration of an α,β-double bond (bottom). Grey background denotes compounds found in cuticular waxes. Functional group color coding corresponds to geometry groups specified in Fig. 4

Similarly, the secondary functional groups of co-occurring 1,2- and 1,3-diols also appear on both even- and odd-numbered carbons and may thus also be installed by a P450 enzyme. However, it cannot be ruled out that the mixtures of 1,2- and 1,3-functionalized compounds are generated by hydration of an α,β double bond (Fig. 9d, yellow functional groups) in the presence of a terminal functional group.

In summary, we conclude that multiple classes of specialty wax compounds are formed by P450 oxidation. Furthermore, the diversity of putative P450 products summarized here indicates that these enzymes have limited regioselectivity, discriminating only between broad regions in the aliphatic chains of their substrates to generate swarms of positional isomers centered around their preferred target carbon position.

Specialty compounds with functional groups on carbons of single parity near the chain terminus

Compounds with O(a) secondary groups

In contrast to the wax constituents described in the previous section, O(a) compounds were functionalized strictly on C-3 and not accompanied by isomers. Based on the structural characteristics of homologous series of O(a) compounds, the reactions forming their hydrocarbon backbones, primary functionalities and secondary functional groups may be deduced.

To first understand the formation of the carbon backbones in O(a) compounds, it must be noted that these specialty compounds all had a (primary) functional group on C-1 and predominantly even TCNs. With these characteristics, O(a) compounds all resemble ubiquitous wax compounds (fatty acids, aldehydes, and alcohols). This similarity suggests that the carbon backbones of O(a) compounds are formed by elongation reactions similar or identical to those of ubiquitous wax compound biosynthesis.

Next, the reactions determining the primary functional groups in the O(a) compounds can be assessed. Across diverse species, O(a) compounds have been found with terminal alcohol, aldehyde, or carboxylic acid head groups (the latter in free or esterified form), and hence with the entire set of terminal functionalities known from ubiquitous, mono-functional wax constituents. Therefore, it seems very likely that the primary functional groups on O(a) compounds are generated by machinery similar or identical to the head group modification pathways leading to the same functionalities in ubiquitous compounds. Specifically, head group modification reactions analogous to those on the alcohol- and alkane-forming pathways likely both use acyl-CoAs as substrates for the generation of O(a) compounds. Thus, either parallel or consecutive acyl-CoA reductions likely generate the primary alcohol and aldehyde head groups of the O(a) compounds, whereas acyl transfer and hydrolysis reactions likely generate O(a)-functionalized esters and acids, respectively. However, from the wax chemical data reviewed here, it is not possible to determine whether these steps are carried out by novel, dedicated enzymes in all species, or whether the ubiquitous wax compound biosynthesis machinery generates these compounds along with its normal products lacking the O(a) function.

Finally, the mechanisms introducing the secondary functional groups on O(a) compounds may be hypothesized. Since O(a) specialty compounds co-occurring in the same wax mixture all had strict 1,3-geometry and were not accompanied by 1,2- or 1,4-isomers, it seems unlikely that their secondary functionalities are introduced by P450 enzymes similar to the Arabidopsis MAH1. Instead, two alternative mechanisms seem probable. First, the O(a) structures bear strong resemblance to the 1,3-bifunctional compounds occurring as intermediates of fatty acid elongation, prompting the hypothesis that the O(a) compounds may be biosynthetically related to those intermediates (Racovita et al. 2014; Wen and Jetter 2007). Indeed, it is very plausible that the 3-keto- and/or 3-hydroxyacyl-CoAs formed in the FAE complex may serve as precursors for the formation of the 1,3-bifunctional wax compounds (Fig. 10a). Second, it is also possible that the 3-functionalized precursors result from condensation of acyl-CoA precursors with malonyl-CoA catalyzed by a polyketide synthase (PKS) instead of the elongase.

Fig. 10

Biosynthesis of specialty compounds with functional groups on carbons of single parity near the chain terminus. Biosynthesis of homologous series or isomeric mixtures of very-long-chain specialty wax compounds whose members’ secondary functional groups are on carbons with single parity (only even- or only odd-numbered carbons) has been hypothesized to occur via the functionalization of elongation intermediates (bracketed compounds). Reduction or hydrolysis of these intermediates could lead to 1,3-functionalized compounds (O(a)), while reduction and decarbonylation or hydrolysis and decarboxylation could lead to 2-functional compounds (E(a)), all of which have been found in plant waxes (indicated by grey background). Functional group color coding corresponds to geometry groups specified in Fig. 4. KCS ketoacyl-CoA synthase, KCR ketoacyl-CoA reductase, HCD hydroxyacyl-CoA dehydratase, ECR enoyl-CoA reductase, RED reductase, AD aldehyde decarbonylase, EST esterase, FAR fatty acyl-CoA reductase, DE decarboxylase

It seems very likely that 3-functionalized acyl-CoA esters generated by condensation of acyl-CoAs with malonyl-CoA, either by KCS or by PKS enzymes, serve as common precursors for O(a) compounds. For the ensuing conversion of the thioester head group into the observed primary functionalities, again two fundamentally different sets of reactions may be hypothesized. On one hand, an aldehyde-forming reductase, a fatty acid-forming thioesterase, and an alcohol-forming fatty acyl-CoA reductase may each have access to 3-functionalized acyl-CoA condensation products, and thus generate their products independently (Fig. 10, blue pathway). In case (stand-alone) PKS enzymes generate respective condensation products, then it is very plausible that a similar variety of modifying enzymes could have access to the same pool of 3-functionalized acyl-CoA precursors. Conversely, if FAE-bound KCS enzymes provide the 3-functionalized acyl-CoA precursors, then the elongation complexes would have to have rather unusual properties and either release some of their intermediates or allow easy access for the modifying enzymes to intercept the intermediates. The modifying enzyme activities implied in this scenario all resemble those of the alkane- and alcohol-forming pathways leading to ubiquitous wax compounds, and either the same or homologous enzymes may be involved in the formation of compounds with and without both O(a) secondary functionalities. This scenario thus involves special PKS or FAE enzymes, but no special modifying enzymes.

On the other hand, a thioesterase could hydrolyze 3-functionalized CoA esters into corresponding free fatty acids, which then serve as substrates for three branch reactions (Fig. 10, red arrows): (a) reduction of the carboxylic head group to 3-functionalized alcohols, (b) reduction to 3-functionalized aldehydes, or (c) decarboxylation to 2-functionalized hydrocarbons (see below). This scenario implies that only one type of enzyme, with thioesterase activity, is able to utilize the condensation products, likely because it alone can intercept respective intermediates from the FAE complex. Precedence for thioesterase access to elongation intermediates exists in the tomato enzyme MKS2, which converts long-chain 3-ketoacyl-CoAs into 3-ketoacids (Yu et al. 2010). However, there is no precedence for enzymes catalyzing reactions a) and b) with VLC free acid substrates. It must, therefore, be considered whether the free acids are first re-activated to CoA esters by long-chain acyl-CoA synthetases (LACS), before they are then converted by reductases and/or decarbonylases analogous to ubiquitous wax compound formation. In both the above scenarios, the various O(a) compounds are explained as a set of products resulting from different head group modification reactions analogous to those yielding the various classes of ubiquitous wax constituents, albeit with an additional secondary functionality near the chain terminus.

Compounds with E(a) secondary groups

As with O(a) compounds, compounds with E(a) secondary functional groups (E(a) compounds) exhibited strict secondary functional group regiospecificity, in this case on C-2. E(a) compounds had hydrocarbon chains with mainly odd TCNs, and were devoid of primary functional groups. Again, the strict functional group geometry of the E(a) compounds makes it unlikely that they are formed by P450 enzymes similar to the Arabidopsis MAH1. However, the E(a) structures resembled those of the O(a) compounds, except that the latter had an additional carbon at the chain end bearing the (primary) functional group. This structural similarity suggests a biosynthetic relationship between both compound classes, in which the O(a) compounds or their immediate precursors likely serve as substrates that can be converted into E(a) compounds by removal of their C-1 atom/functional group.

One of two possible reactions may remove the C-1 atom and its functional group from O(a) structures to convert them into E(a) compounds: decarbonylation or decarboxylation. In the first scenario, by analogy with the pathway to ubiquitous wax alkanes, 3-functionalized aldehydes may be decarbonylated to 2-functionalized hydrocarbons, by an enzyme similar or identical to CER1. Thus, this scenario could be tested by searching for 2-alkanols and 2-ketones in the waxes of the Arabidopsis cer1 mutant. Alternatively, head group cleavage could also occur from a 3-functionalized carboxylic acid, either spontaneously or enzyme-catalyzed. The secondary functionality of 3-functionalized acid precursors may facilitate cleavage of the bond between C-1 and C-2 and thus the removal of the terminal functionality. Interestingly, a similar reaction has been described for the biosynthesis of long-chain methylketones (compounds with E(a) functional groups) in tomato, and the decarboxylase catalyzing it has been characterized (Yu et al. 2010). Identification of similar enzymes that operate on very-long-chain substrates have yet to be identified.

Based on the precedence provided by both the (ubiquitous wax) alkane- and the (tomato trichome) methylketone-forming pathways, enzymatic decarbonylation and decarboxylation of 1,3-bifunctional aldehydes and acids, respectively, are the most likely biosynthetic routes to compounds with E(a) groups (Fig. 10, black arrows). Overall, we propose that E(a) structures are downstream products of O(a) compounds or their immediate precursors, and the secondary functional groups in both E(a) and O(a) compounds are therefore equivalent. From this, it follows that both E(a) and O(a) compounds should co-occur, especially if one of these compound classes is formed in substantial quantity. So far, however, it appears that the only species known to have both classes are Arabidopsis thaliana (unpublished results), Solanum tuberosum (Szafranek and Synak 2006) and Aloe arborescens (Racovita et al. 2014), and in all of them only very small amounts were found. It is interesting to note that, in these cases, the co-occurring E(a) and O(a) compounds had different chain length profiles, pointing to different specificities of the enzymes en route to both final products (i.e., involving different FAE complexes or chain-length-specific head group modifying enzymes).

So far, the potential pathways to O(a) and E(a) compounds we have described treat substrates and products with oxo versus hydroxyl secondary functional groups largely equally. However, it is possible that enzymes involved in the release or intercept of β-functionalized CoA esters exhibit preference for or against one of these secondary functional groups, or that downstream reduction or head group removal processes exhibit similar biases. It is also possible that the secondary functional groups on intermediates and/or products of these pathways may undergo redox reactions, thereby interconverting secondary oxo- and hydroxyl-bearing specialty wax constituents.

In summary, it seems likely that compounds with E(a) groups (2-functional compounds) are biosynthesized from compounds with O(a) groups (1,3-bifunctional compounds), which are in turn derived from β-functionalized acyl-CoAs intercepted or released from the elongation pathway, or else produced by PKS enzymes. We propose that, in many plant species, the same enzymes involved in ubiquitous wax compound biosynthesis may also catalyze some of the steps leading to specialty compounds with O(a) and E(a) functional groups. However, it is also conceivable that in some species, especially in species where specialty compounds with E(a) and/or O(a) groups accumulate to high levels, multiple isoforms of these enzymes exist where at least one favors substrates with C-3 function and another one substrates without secondary functional group. Since relatively highly abundant O(a) compounds do not always share chain length distributions with co-occurring ubiquitous compounds with the same head groups, such as the 1,3-diols in Papaver orientale (Jetter et al. 1996), it seems likely that such novel enzymes do exist at least in some species.

The pathway alternatives summarized here have been derived from chemical information accumulated over the past decades, and in the future will be tested by molecular biological and biochemical experiments. In the design of experiments characterizing the enzymes involved in pathways to O(a) and E(a) compounds, it will be important to select study species that bear substantial amounts of diverse compounds with such functional groups. To verify the involvement of ubiquitous wax biosynthesis enzymes, the selection of a system with characterized wax biosynthesis pathways would also have advantages. Based on these criteria and the wax profiles reported so far, Arabidopsis is an easy target for testing the involvement of ubiquitous compound biosynthesis enzymes, although Ricinus communis has higher abundance of O(a) compounds. Further analyses of diverse wax mixtures may provide more favorable candidate species. It should also be kept in mind that different plant species may synthesize these compounds via different pathways or using different enzymes, especially if these compounds are present as either major or minor wax constituents. In either case, however, the O(a) or E(a) compounds may serve as biomarkers for possible elongation disruption or PKS-catalyzed condensation events.

Compound classes with a single secondary functional group near the middle of the chain

Many plant wax mixtures contain high concentrations of even-TCN compounds with both a primary functionality and a secondary functional group on only odd-numbered carbons near the middle of the chain (O(b) functionalities), indicating that the latter are installed by a highly regioselective (and therefore not P450-mediated) process. These mixtures of O(b) compounds frequently occurred alongside compounds with O(a) or E(a) functional groups (Fig. 6, or cf. Fig. 5a, b), suggesting that all three compound classes are biosynthetically related. Since O(a) and E(a) compounds are likely derived from released 3-functionalized acyl-CoAs, it follows that the biosynthesis of O(b) compounds may also proceed via the same free intermediates.

By analogy with ubiquitous acyl-CoA esters (lacking 3-functionalities), there are two likely fates for free 3-functionalized acyl-CoAs: they can (1) enter the head group modification pathways (and become O(a) or E(a) compounds, see “Specialty compounds with functional groups on carbons of single parity near the chain terminus” section), or (2) enter the elongation pathway(s). If a 3-functionalized acyl-CoA goes through a full elongation cycle (Fig. 11a, compound in dotted box), its C-3 functional group is retained but shifted to C-5 in the product (Fig. 11a, compound in solid box). The completion of any subsequent rounds of elongation would separate the retained functional group from the head group by an additional two carbons, producing compounds with 1,7, 1,9, 1,11, etc. functional group geometries. This process can, thus, generate fatty acyl-CoA compounds with a secondary functional group exclusively on odd-numbered carbons (Fig. 11a, compounds in bold box). By passing through head group modification pathways (see “Specialty compounds with functional groups on carbons of single parity near the chain terminus” section; Fig. 11a, red and blue arrows), these acyl-CoAs can be converted into aldehydes, fatty acids, and/or alcohols with secondary functional groups present only on odd-numbered carbons (Fig. 11a, compounds in the bold, shadowed box labelled O(b)). Based on wax analyses of single species, other authors have reached similar hypotheses (for example, Busta et al. 2016a; Schneider et al. 2016; Wen and Jetter 2007). The present work demonstrates that such hypotheses hold when considering wax chemical structures from diverse species.

Fig. 11

Biosynthesis of specialty wax compounds with a single functional group near the middle of the chain. a Biosynthesis of homologous series or isomeric mixtures of very-long-chain specialty wax compounds whose members’ secondary functional group is on carbons of single parity (only even- or only odd-numbered carbons) and with a carbon number greater that three (C-n, n > 3) has been hypothesized to occur via the intercept of elongation intermediates (bracketed compounds), followed by subsequent re-entry into further elongation cycles. The products of these further elongation cycles (compounds in dashed boxes) may be available to functionalization pathways (blue), to generate compounds with secondary functional groups on only odd-numbered carbons (O(b)) or, via an alkane-forming pathway, to products with secondary functional groups on only even-numbered carbons (E(b)), all of which have been reported from plant waxes (indicated by grey background). Elongation intermediates arising from the re-entry of an intercepted intermediate (square-bracketed compounds) could also be intercepted and passed to functionalization enzymes, resulting in 1,3-compounds with an additional secondary functional group on only odd-numbered carbons (O(b)O(a)) or, via the alkane-forming pathway, 2-functional compounds with an additional secondary functional group on even-numbered carbons (E(b)E(a)). b Biosynthesis of very-long-chain ketones (E(b)) by head-to-head condensation. Functional group color coding corresponds to geometry groups specified in Fig. 4. KAS ketoacyl-ACP synthase, KAR ketoacyl-ACP reductase, HAD hydroxyacyl-ACP dehydratase, EAR enoyl-CoA reductase, PKS polyketide synthase, PKR polyketide reductase, KCS ketoacyl-CoA synthase, KCR ketoacyl-CoA reductase, HCD hydroxyacyl-CoA dehydratase, ECR enoyl-CoA reductase, RED reductase, AD aldehyde decarbonylase, EST esterase, FAR fatty acyl reductase, DE decarboxylase, LACS long-chain acyl-CoA synthetase. (Color figure online)

In some species, compounds with O(b) functional groups co-occur with E(b) structures, as exemplified by prim./sec. diols such as 1,11-C₃₀ diol accompanying nonacosan-10-ol on Osmunda regalis, Myricaria germanica and several Papaver species (Fig. 11, or cf. Fig. 5a, b). Based on this co-occurrence, these E(b) compounds can be interpreted as derivatives of the co-occurring O(b) structures lacking the C-1 atom/functional group. It is therefore very plausible that the E(b) compounds are formed from corresponding O(b) compounds or a common biosynthetic precursor by removal of the head group through decarbonylation or decarboxylation, analogous to similar processes leading to ubiquitous wax alkanes and E(a) compounds (see “Specialty compounds with functional groups on carbons of single parity near the chain terminus” section). In this way, O(b) and E(b) compounds co-occurring in the wax of some species can be recognized as biosynthetically related products of head group modification reactions, involving either hydrolysis, reduction, or head group removal. The hypothesized biosynthetic pathways to both compound classes thus share two steps: the generation of 3-functionalized acyl-CoAs and their further elongation to in-chain-functionalized acyl-CoAs. The pathways diverge only in the final head group modification steps.

Since the head group modification reactions leading to O(b) and E(b) compounds parallel those involved in ubiquitous wax compound formation, both are likely catalyzed by homologous or identical enzymes. However, there are several reports on species whose E(b) compounds had chain length profiles not matching those of the ubiquitous compound class(es) with corresponding head group. For example, the secondary alcohols of Pogonatum belangeri and several Papaver spp. (Jetter and Riederer 1996; Neinhuis and Jetter 1995) were dominated by C₂₉ (primarily nonacosan-10-ol), while the accompanying alkanes (i.e., the ubiquitous compounds with the same TCN parity also lacking C-1 functionalities) peaked at C₂₇ and C₂₅/C₂₇, respectively. This suggests that, at least in these species, the machinery involved in E(b) compound biosynthesis differs from that generating the ubiquitous alkanes.

The pathways to O(b) and E(b) functional groups delineated above are very plausible for species where several compound classes with characteristic combinations of secondary and primary functionalities co-occur, or where several homologs and isomers of E(b) compounds occur in specific patterns pointing to elongation of 3-functionalized acyl-CoAs. However, in species where single E(b) compounds rather than homolog or isomer mixtures occur, alternative pathways cannot be ruled out. One such pathway in particular had been suggested early on (Kolattukudy 1970), involving the head-to-head condensation of two fatty acyl compounds (Fig. 11b). This mechanism was then tested in a few species (B. oleracea and Pisum sativum; Kolattukudy 1966, 1968; Kolattukudy et al. 1968), and rejected there. However, comprehensive isomer analyses of the secondary alcohol and ketone products in these species showed that they do not bear E(b) structures, but instead have EO(b) compounds now thought to be formed by P450-mediated hydroxylation (see “Specialty compound classes with secondary functional groups on even- and odd-numbered carbons” section). Thus, it turns out that the head-to-head condensation mechanism has not been tested thoroughly for the formation of certain E(b) compounds so far, and biochemical investigations of select species accumulating such structures are clearly warranted.

Suitable models for investigating potential biosynthetic mechanisms involving 3-functionalized CoA elongation and head-to-head condensation should be chosen based on comprehensive analyses of the specialty compounds, to rule out species on which the target E(b) compounds are accompanied by positional isomers suggesting biosynthesis by P450 oxidation (see “Specialty compound classes with secondary functional groups on even- and odd-numbered carbons” section), or accompanied by head group modification analogs (and thus most likely products of elongated 3-functionalized acyl-CoAs, see this section above). However, many earlier publications reported only single E(b) compounds, without mentioning other wax constituents or whether any compounds remained unidentified, and without assessing detection limits for homologs and isomers. For species in such reports, it is not clear whether head-to-head condensation must still be considered an alternative for formation of the E(b) compounds. Future wax analyses reporting such structures should ensure that detailed homolog and isomer searches are performed and reported, even in the case of negative results.

Compound classes with two secondary functional groups of single parity near the middle of the chain

Since the mechanisms of elongation intermediate release and further condensation/elongation and head group modification have been invoked to explain the biosynthesis of specialty wax compounds with single secondary functional groups (O(b) and E(b) compounds, see “Compound classes with a single secondary functional group near the middle of the chain” section), it may be further hypothesized that these mechanisms may repeatedly act on the same, growing hydrocarbon precursor to generate compounds with multiple secondary functional groups. For example, after the release and further elongation of a 3-functionalized acyl-CoA, rather than head group modification taking place (and leading to O(b) and E(b) compounds), another release event could occur. This could lead to 3,7-; 3,9-; etc. bifunctional acyl-CoAs (Fig. 11a, compounds bound by straight brackets). If these CoA esters were used directly for head group reduction, the wax products would have even TCNs, a primary functionality, and secondary functional groups on C-3 as well as another odd-numbered carbon further inside the chain (Fig. 11a, compounds in the shadowed box labelled O(b)O(a)). While such compounds have been found in plant floral oils (Fig. 6c), only detailed wax analyses of more species will reveal if they also occur as cuticle components. Corresponding E(b)E(a) compounds, resulting from head group removal and thus having odd TCNs and lacking a primary functionality, have not been reported so far. The identification of these structures, especially in conjunction with other specialty wax components, will give important clues about the biosynthesis of these compounds.

It is also possible that repeat release and condensation/elongation mechanisms may occur immediately one after the other. For example, a released 3-functionalized acyl-CoA might enter a new elongation cycle, but then be released again after only the condensation step is complete. This would lead to a 3,5-bifunctional acyl-CoA, as two malonate condensations occurred without removal of the new secondary functionalities (Fig. 12a, compound in dashed box). If a 3,5-bifunctional acyl-CoA entered further cycles of elongation, then specialty compounds with two secondary functional groups near the middle of the carbon chain would be formed (Fig. 12a, CoA ester compounds in solid boxes). The wax products derived from these would have functional groups either on odd-numbered carbons (O(b)O(b)) of chains with even TCNs that bear a head group or on even-numbered carbons (E(b)E(b)) of chains with odd TCNs that lack head groups (Fig. 12a, compounds in bold, shadowed boxes). With these structural characteristics, the O(b)O(b) and E(b)E(b) compounds appear to be bifunctional analogs of the O(b) and E(b) structures discussed above (see “Compound classes with a single secondary functional group near the middle of the chain” section). Just like O(b) and E(b) compounds, O(b)O(b) and E(b)E(b) compounds also co-occur with O(a) and E(a) compounds (markers for 3-functionalized acyl-CoA intermediates (Fig. 6, cf. Figs. 5, 7). Based on these observations, it is plausible that the mechanisms leading to compounds with O(b)O(b) or E(b)E(b) functional groups are similar to those leading to single O(b) or E(b) functionalities, except that they generate not one, but two regioselectively placed secondary functional groups. By considering the structures of O(b)O(b) and/or E(b)E(b) compounds from single species, previous authors have reached similar hypotheses (Jetter and Riederer 2000; Schulz et al. 2000). The present work, through examination of wax chemical structures from species of diverse lineages, provides further support for these hypotheses.

Fig. 12

Biosynthesis of specialty wax compounds with two secondary functional groups of single parity near the middle of the chain. a Biosynthesis of homologous series or isomeric mixtures of very-long-chain wax compounds whose members’ secondary functional groups are on carbons with single parity (only even- or only odd-numbered carbons) and with carbon numbers greater than three (C-n, n > 3) has been hypothesized to occur via the intercept of fatty acid biosynthesis intermediates (bracketed compounds), followed by condensation with malonyl-CoA by a PKS enzyme, and subsequent re-entry into further elongation cycles. The products of these further elongation cycles (compounds in dashed boxes) may be available to functionalization pathways (blue) to generate compounds with two secondary functional groups on only odd-numbered carbons (O(b)O(b)), or, via an alkane-forming pathway, products with secondary functional groups on only even-numbered carbons (E(b)E(b)), some of which have been reported form plant waxes as fatty acids (indicated by grey background). Elongation intermediates arising from intercepted fatty acid biosynthesis intermediates by the method described above (square-bracketed compounds) could also be intercepted and passed to functionalization enzymes, resulting in 1,3-compounds with two additional secondary functional groups on only odd-numbered carbons (O(b)O(a)) or, via the alkane-forming pathway, 2-functional compounds with two additional secondary functional groups on even-numbered carbons (E(b)E(a)). b Biosynthesis of very-long-chain β-diketones and β-ketoalcohols (E(b) E(b)) by head-to-head condensation. Functional group color coding corresponds to geometry groups specified in Fig. 4. KAS ketoacyl-ACP synthase, KAR ketoacyl-ACP reductase, HAD hydroxyacyl-ACP dehydratase, EAR enoyl-CoA reductase, PKS polyketide synthase, PKR polyketide reductase, KCS ketoacyl-CoA synthase, KCR ketoacyl-CoA reductase, HCD hydroxyacyl-CoA dehydratase, ECR enoyl-CoA reductase, RED reductase, AD aldehyde decarbonylase, EST esterase, FAR fatty acyl reductase, DE decarboxylase, LACS long-chain acyl-CoA synthetase. (Color figure online)

Many different enzyme combinations can be envisioned that might perform repeat malonate additions without secondary functional group removal to generate the 3,5-bifunctional acyl-CoA intermediates, eventually leading to the O(b)O(b) and E(b)E(b) compounds by the route described above. In essence, the machinery generating and releasing the 3-functionalized acyl-CoAs might be a KCS enzyme (in a leaky/truncated FAE complex), a KAS enzyme (in a leaky/truncated FAS complex), or a PKS enzyme. Following release, it seems likely that either a KCS or PKS enzyme may catalyze one further malonate addition, and the resulting 3,5-bifunctional acyl-CoAs may then enter further elongation cycles to generate acyl-CoA wax precursors with O(b)O(b) functionalities. The two malonate additions may, thus, be catalyzed by the same types of enzymes or a combination of two different types (2xKCS, 2xPKS, KAS + KCS, KAS + PKS, KCS + PKS, PKS + KCS).

Acyl-CoA wax precursors with O(b)O(b) functionalities, generated by repeat release and condensation steps, could then be substrates for head group modification and give rise to O(b)O(b) wax components (Fig. 12a, compounds in shadowed boxes). Subsequently, O(b)O(b) compounds could have their C-1 function removed, either by decarbonylation or decarboxylation, to yield E(b)E(b) wax compounds. Whether these modification reactions are carried out by enzymes similar or identical to those performing ubiquitous wax biosynthesis is unclear. However, there are several examples of wax mixtures in which co-occurring E(b)E(b) compounds and wax alkanes have different chain length profiles (for example, Eragrostis curvula and Panicum virgatum, Tulloch and Hoffman 1980; Tulloch 1982), suggesting that at least some of the machinery leading to these E(b)E(b) compounds differs from that generating the ubiquitous compounds.

Finally, while the elongation and modification of 3,5-functionalized acyl-CoA esters is a reasonable mechanism for the biosynthesis for E(b)E(b) and O(b)O(b) wax compounds, it should be noted that compounds with E(b)E(b) functional groups and odd TCNs could also be formed by head-to-head condensation of an intercepted 3-functionalized fatty acyl with a simple VLC fatty acyl-CoA or fatty acid (Fig. 12b). Both biosynthesis pathways to E(b)E(b) compounds (elongation/modification of 3,5-functionalized acyl-CoAs or head-to-head condensation) require 3-functionalized fatty acid precursors; therefore, the co-occurrence of compounds with relatively short chain lengths bearing E(a) or O(a) functional groups, likely originating from 3-functionalized acyl-CoA intermediates, cannot be used as a proxy to distinguish between potential pathways to E(b)E(b) compounds. However, these two mechanisms differ in that the products of one are derived from acyl elongation processes, while in the other the products are not. Ubiquitous compounds (and probably specialty compounds) that are biosynthesized via elongation pathway(s) often occur as homologous series and alongside related compounds that differ in their head group oxidation state, for example the co-occurrence of ubiquitous aldehydes and alkanes with related chain length distributions, or the co-occurrence of 1,11-diols and nonacosan-10-ol in Papaver species. Thus, if O(b)O(b) and E(b)E(b) compounds are biosynthesized via an intercept-condensation-elongation-modification mechanism, they co-occur with related products analogous to those observed in these other classes, and thus may serve as pathway indicators. Presence or absence of O(b)O(b) compounds is thus a crucial clue for the pathways leading to E(b)E(b) structures. Specifically, the presence of head groups in O(b)O(b) compounds are biomarkers of an intercept-condensation-elongation-modification pathway to E(b)E(b) structures, rendering head-to-head condensation unlikely (but not impossible). However, there has been only one report to date describing the co-occurrence of compounds bearing E(b)E(b) and O(b)O(b) functional groups, where these compounds were reported in pollen extract of Helianthus annuus (Schulz et al. 2000). It is critical that future investigations of species bearing E(b)E(b) compounds include targeted searches for compounds with O(b)O(b) functionalities to help distinguish between biosynthesis mechanisms leading to E(b)E(b) compounds. So far, neither mechanism can be ruled out based on the chemical data reviewed here.

β-Diketones are the most commonly encountered compounds with E(b)E(b) functional groups, and their biosynthesis has been the subject of several experiments due to the important role these compounds play in the cuticles of many crop species. Early on, the highly abundant mid-chain β-diketones found on, for example, several Poaceae, Buxaceae, and Ericaceae species (Fig. 6, Table S1) were classified as polyketide products based on their functional group geometry, leading to the hypothesis that PKS-catalyzed condensation reactions install both keto groups in these E(b)E(b) structures (von Wettstein-Knowles 1995). Diverse pieces of biochemical and genetic evidence have since confirmed that medium- and/or long-chain acyls serve as substrates for condensation with malonate, confirming that at least one PKS enzyme is essential for β-diketone formation in both wheat and barley (Hen-Avivi et al. 2016; Schneider et al. 2016). Thus, the original hypothesis that at least one mid-chain functionality is introduced by PKS condensation, as predicted from chemical structures alone, has now been verified for this class of E(b)E(b) structures in at least two species. However, it remains to be seen whether similar PKS enzymes are involved in introducing functional groups in other E(b)E(b) compounds in other species as well.

Interestingly, the wheat and barley PKS genes were found in clusters together with genes also involved in β-diketone formation, including one annotated as encoding a carboxylesterase. Biochemical evidence suggested that it generates 3-functionalized fatty acids, likely by intercepting and hydrolyzing 3-functionalized acyl-ACP intermediates of plastidial fatty acid de novo biosynthesis. Accordingly, it was concluded that the two functional groups of β-diketones are introduced by different mechanisms, consecutively involving FAS intermediate intercept and subsequent PKS condensation. The resulting 3,5-diketoacyl-CoA intermediates may then be converted into the final β-diketone products, likely involving FAE-based elongation to VLC diketoacyl-CoA products (Fig. 12a). Finally, head group modification analogous to ubiquitous wax compound biosynthesis may lead either to O(b)O(b) compounds with diverse primary functional groups or, through decarbonlyative or decarboxylative removal of C-1 with its functional group, to E(b)E(b) compounds such as the β-diketones.

Conclusions

This literature survey was undertaken to first catalogue and codify specialty wax compounds reported in the literature to date, and thus to identify gaps in our knowledge around the waxes of certain plant taxa or missing wax compound structures. We further analyzed the structural features of known specialty compounds to derive hypotheses for how they might be biosynthesized, and to compare these hypotheses to those previously derived from the examination of single species. Finally, we aimed to identify and prioritize experiments for testing these hypotheses.

Specialty compounds with one, two, and three secondary functional groups have been reported from species spanning nearly the entire plant kingdom. The Angiosperms remain the best-studied group, and relatively few reports exist for other species, especially bryophytes and ferns. Based on results reported so far, it seems that these taxa contain species that do bear specialty wax compounds that have structures or occur in combinations that are biosynthetically informative. Future analyses should gravitate towards these understudied taxa to provide information on how specialty wax compounds they bear may differ from those already described, to search for expected but so far unobserved specialty compounds, to identify patterns of co-occurrence that will provide evidence that may resolve uncertainties between competing biosynthesis hypotheses, or to report the absence of specialty wax compounds.

Specialty cuticular wax compounds have in common several structural variables including TCN parity, head group oxidation state, secondary functional group oxidation state(s), and isomer characteristics of the secondary functional group(s). Analysis of such structural features exhibited by specialty compounds reported so far indicate that their carbon backbones and head groups are probably created by machinery largely similar to that generating ubiquitous compounds. The position(s) of secondary functional groups found across mixtures and homologous series of related specialty compounds suggest that they are installed by P450 enzymes (secondary groups on even- and odd-numbered carbons) or by the elongation and head group modification of acyl-CoAs with preexisting secondary functional groups (secondary groups on only even- or only odd-numbered carbons). Though these biosynthetic reactions could theoretically be carried out by the enzymes generating ubiquitous wax compounds, differences between chain length profiles of specialty compounds and co-occurring ubiquitous wax compounds suggest that at least some of the machinery leading to specialty compounds is unique. It will be important to determine the extent to which ubiquitous wax compound biosynthesis enzymes play a role in the biosynthesis of these specialty compounds.

While the hypothetical biosynthetic pathways derived from the analyses presented here largely support those that had been suggested based on single species analyses, this work provides several new perspectives. First, it seems likely that, while P450 enzymes installing secondary functional groups on wax compounds or wax precursors are not highly regiospecific, they do belong to several subgroups that have loose regiospecificity in that they restrict their activity to portions of the aliphatic chains of their substrate(s). Second, the widespread presence of compounds probably derived from the elongation and modification of acyl-CoAs with secondary functions adds to the number of possible fates for such compounds. Accordingly, there may be competition for these compounds between reduction enzymes of the elongation complex, head-group modifying enzymes, and the release and entry of these compounds into different condensation enzymes or elongation complexes. Furthermore, we speculate that the architecture of this competition is complex in at least some species. For example, the chain length distributions of 2-alcohols (mainly C₃₁), and 3-hydroxy acids (mainly C₂₈) in Aleo arborescens (Racovita et al. 2014) suggest that acyl-CoAs with in-chain functions may preferentially enter or be acted upon by different modification pathways depending on their chain length.

Finally, this work identifies a number of pressing questions surrounding specialty wax compounds, potential systems that will likely be useful in the search for answers to these questions, and in some cases, first experiments towards doing so. Chief among these are the identification and characterization of enzymes and mechanisms behind the biosynthesis of specialty compounds with secondary functional groups that are not P450 derived, for example nonacosan-10-ol, palmitone, and the various β-diketones. Also of interest are the details behind the biosynthesis of the diverse 1,3-bifunctional compounds that have been found. Elucidation the biosynthesis of these major specialty compounds will enable new investigations into their functions, into the mechanisms and factors influencing the formation of epicuticular wax crystals by specialty compounds, and into potential applications in plant biotechnology.