Composition of alkyl esters in the cuticular wax on inflorescence stems of Arabidopsis thaliana cer mutants

Authors

  • Christine Lai,

    1. Department of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, BC V6T 1Z4, Canada, and
    Search for more papers by this author
  • Ljerka Kunst,

    1. Department of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, BC V6T 1Z4, Canada, and
    Search for more papers by this author
  • Reinhard Jetter

    Corresponding author
    1. Department of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, BC V6T 1Z4, Canada, and
    2. Department of Chemistry, University of British Columbia, 6174 University Boulevard, Vancouver, BC V6T 1Z3, Canada
    Search for more papers by this author

(fax +001 604 822 6089; e-mail jetter@interchange.ubc.ca).

Abstract

Wax biosynthetic pathways proceed via the elongation of 16:0 acyl-CoA to very long-chain fatty acids (VLCFA), and by further modifications that include reduction to primary alcohols and formation of alkyl esters. We have analyzed the alkyl esters in the stem wax of ten cer mutants of Arabidopsis thaliana together with the corresponding wild types. Alkyl esters with chain lengths between C38 and C52 were identified, and the levels of esters ranged from 0.15 µg cm−2 in Wassilewskija (WS) to 1.20 µg cm−2 in cer2. Esters with even numbers of carbons prevailed, with C42, C44 and C46 favoured in the wild types, a predominance of C42 in cer2 and cer6 mutants, and a relative shift towards C46 in cer3 and cer23 mutants. The esters of all mutants and wild types were dominated by 16:0 acyl moieties, whereas the chain lengths of esterified alcohols were between C20 and C32. The alkyl chain-length distributions of the wild-type esters had a maximum for C28 alcohol, similar to the free alcohols accompanying them in the wax mixtures. The esterified alcohols of cer2, cer6 and cer9 had largely increased levels of C26 alcohol, closely matching the patterns of the corresponding free alcohols and, therefore, differing drastically from the corresponding wild type. In contrast, cer1, cer3, cer10, cer13 and cer22 showed ester alcohol patterns with increased levels of C30, only partially following the shift in chain lengths of the free alcohols in stem wax. These results provide information on the composition of substrate pools and/or the specificity of the ester synthase involved in wax ester formation. We conclude that alcohol levels at the site of biosynthesis are mainly limiting the ester formation in the Arabidopsis wild-type epidermis.

Introduction

Most above-ground plant organs are covered with a cuticle consisting of intracuticular waxes embedded in a polymer matrix of cutin, and epicuticular waxes deposited on the outer surface of the cutin layer (Walton, 1990). In many species, the epicuticular waxes form microscopic crystals that scatter light and give the surface a whitish (glaucous) appearance (Clark and Lister, 1975). The most important physiological function of the waxes is to limit non-stomatal water loss, but they are also of ecological importance as they form the interface between plants and their environment. The water-repellent cuticular waxes guard leaf, stem, flower and fruit surfaces from the accumulation of particles from the atmosphere (Saunders, 1971), and also keep them dry (Holloway, 1970), thus preventing the germination of pathogen spores (Doss et al., 1993). All these specific functions of plant cuticles can only be understood on the basis of their characteristic wax composition and biosynthetic origin.

The cuticular waxes are formed in epidermal cells on biosynthetic pathways that first generate the very long-chain carbon skeletons, and then modify them into diverse aliphatic lipid classes (Kunst et al., 2006). Wax biosynthesis utilizes 16:0 and 18:0 acyl precursors supplied by a plastid-localized elongation pathway. These are first transformed into acyl-CoA esters and then extended into very long-chain fatty acids (VLCFAs) by extraplastidial elongases. The resulting mixture of VLCFAs, with chain lengths typically ranging from C24 to C32, is modified either by the decarbonylation pathway to yield corresponding aldehydes, alkanes, secondary alcohols, alkanediols, ketols and ketones, or by the acyl reduction pathway to yield the primary alcohols found in cuticular wax. It is generally assumed that these very long-chain alcohols are involved in the formation of alkyl esters by reacting with VLCFA precursors with chain lengths longer than C20 (Kunst et al., 2006). However, direct evidence showing that, on the one hand, ester alcohols originate from the same pool as those destined for the surface wax and, on the other hand, ester acids are C20–C32 products resulting from the elongation of 16:0 acyl-CoA, is lacking.

Wax secondary alcohols, alkanediols, ketones, ketols and alkyl esters have structures with potential isomerism. All these compound classes are typically biosynthesized with characteristic product chain lengths and positions of the functional groups (Jetter et al., 2006). The exact isomer compositions have been elucidated, e.g. for alkanediols and secondary alcohols of various plant species (Franich et al., 1979; Holloway et al., 1976; Jetter, 2000; Jetter et al., 1996), and have led to hypotheses on their biosynthetic origins and the specificity of the enzymes involved in their production. The proposed reactions have largely been confirmed in those cases where the suggested biochemical processes have since been investigated (Soliday and Kolattukudy, 1978).

The composition of alkyl esters may provide especially valuable information on wax biosynthesis in different tissues of various plant species, because they occur as isomers and are inherently dimeric molecules containing acid and alcohol units. Besides, these esters are the end products of the reduction pathway (Kunst et al., 2006) and therefore are useful targets for determining the intracellular site of the final steps of wax biosynthesis. The subcellular organization of late steps on the pathway(s) and their integration with export towards the cuticle are currently unknown. The homologue composition of cuticular wax esters of diverse plant species has been studied (Holloway, 1970), but only rarely were the isomer distributions reported. Although in some cases the acyl and alkyl moieties were quantified after ester cleavage (Allebone and Hamilton, 1972; Jetter and Riederer, 1996; Sümmchen et al., 1995b), they can also be directly assessed by gas chromatography-mass spectrometry (GC-MS) using characteristic acyl fragments in the mass spectral datasets. Gülz et al. (1994) employed this technique to show that wax esters of Quercus robur leaves consist of a broad range of both acid and alcohol moieties. In a similar way, the ester compositions of Picea abies and Brassica napus var. rapifera were investigated, and found to be a complex mixture of C12–C38 acids esterified with C8–C34 alcohols, and of C16–C20 acids esterified with C26–C30 alcohols, respectively (Shepherd et al., 1995; Sümmchen et al., 1995a).

Alkyl esters also occur in the cuticular wax on inflorescence stems of Arabidopsis thaliana, and their composition is of special interest as biochemical inferences from wax analytical data can be tested using the molecular tools available in this model system. Unfortunately, only the total levels of Arabidopsis wax esters have been reported to date (Jenks et al., 1995, 1996; Rashotte et al., 2001, 2004), without further details on their chain-length distribution and isomer composition. Hence, important information is missing that could be used to assess substrate preference and availability, and also possibly the localization of substrate pools and enzymes involved in ester formation in A. thaliana.

A collection of Arabidopsis mutants with glossy inflorescence stems had been isolated in visual screens (Koorneef et al., 1989) and were later found to have altered levels or compositions of cuticular wax (Jenks et al., 1995). Among these eceriferum or cer mutants, ten have increased wax ester levels, and therefore represent ideal tools for comparative studies of alkyl ester biosynthesis. In the present work, we used GC-MS to investigate the stem wax esters of these selected cer mutants and of corresponding wild types. Our objectives were to (1) determine the chain-length distribution of the wax esters, (2) identify and quantify the acyl and alkyl units in these esters, (3) compare this information on ester isomer composition with data on free alcohols and acids in the stem wax, and (4) thereby assess substrate limitations and location of wax ester biosynthesis in epidermal cells.

Results and discussion

Qualitative and quantitative analysis of wax ester chain lengths

Cuticular waxes were extracted from inflorescence stem surfaces of ten selected cer mutants and the two corresponding wild types of A. thaliana, and alkyl esters were separated from other wax constituents by GC. Mass spectrometric detection was initially used to identify the alkyl ester homologues based on retention order, molecular ions, and comparison with spectra of selected authentic standards. In most of the samples, alkyl esters with even numbers of carbons and chain lengths ranging from C38 to C52 were identified. Additionally, odd-numbered ester homologues between C41 and C51 were detected. Mass spectral identifications were confirmed for selected samples after purification of the alkyl ester fraction on TLC (data not shown).

In a second set of GC runs, flame ionization detection was employed to quantify the absolute quantities of alkyl ester homologues in comparison with an internal standard. The Landsberg erecta (Ler) and Wassilewskija (WS) wild types showed total alkyl ester coverages of 0.17 µg cm−2 and 0.20 µg cm−2, respectively (Table 1). cer13 was the only mutant that had absolute quantities of esters similar to the wild types, whereas all the other lines showed increased levels of esters ranging from 0.28 µg cm−2 in the wax of cer24 to 1.12 µg cm−2 in cer2. Absolute quantities of total wax esters have been reported for various mutants previously (Jenks et al., 1995, 1996; Rashotte et al., 2001, 2004), and were largely confirmed by our results.

Table 1.   Coverages (µg cm−2) of ester homologues and relative composition (%) of ester isomers in the stem cuticular wax of Arabidopsis thaliana. Mean values (n = 3) and SD are given for cer mutants and corresponding wild types
 Acid chain lengthAlcohol chain length Lercer1cer2cer3cer6cer9cer10cer13WScer22cer23cer24
C42 esters  0.03 ± 0.010.07 ± 0.010.52 ± 0.170.04 ± 0.020.30 ± 0.230.13 ± 0.040.05 ± 0.010.04 ± 0.030.05 ± 0.030.16 ± 0.020.10 ± 0.010.06 ± 0.01
14286.5 ± 2.94.7 ± 0.80.7 ± 0.22.3 ± 1.12.4 ± 1.14.6 ± 1.55.4 ± 1.67.3 ± 2.54.3 ± 2.04.3 ± 1.25.9 ± 1.54.5 ± 1.0
162675.2 ± 1.590.0 ± 0.898.9 ± 0.594.0 ± 3.594.2 ± 1.989.9 ± 2.090.9 ± 1.686.8 ± 0.191.3 ± 0.094.0 ± 1.789.7 ± 0.584.0 ± 7.8
182412.4 ± 5.43.1 ± 0.70.1 ± 0.11.3 ± 1.32.5 ± 0.61.7 ± 0.73.1 ± 0.52.3 ± 0.93.0 ± 0.01.7 ± 0.73.6 ± 0.21.7 ± 1.5
20220.8 ± 1.21.0 ± 0.90.1 ± 0.11.1 ± 1.00.5 ± 0.51.3 ± 0.50.7 ± 0.62.6 ± 0.10.1 ± 0.2 0.9 ± 1.21.8 ± 1.2
22203.9 ± 0.61.2 ± 1.70.2 ± 0.41.3 ± 1.40.0 ± 0.11.4 ± 1.2 1.0 ± 1.40.6 ± 0.9  4.6 ± 5.8
24181.1 ± 1.6   0.3 ± 0.61.1 ± 1.9  0.6 ± 0.9  3.3 ± 3.3
C44 esters  0.08 ± 0.050.16 ± 0.020.44 ± 0.130.09 ± 0.040.21 ± 0.160.18 ± 0.030.12 ± 0.010.07 ± 0.020.06 ± 0.010.23 ± 0.010.17 ± 0.010.12 ± 0.02
14303.0 ± 0.23.9 ± 1.40.1 ± 0.27.0 ± 1.70.0 ± 0.00.7 ± 0.71.9 ± 0.84.9 ± 3.01.6 ± 0.86.2 ± 3.74.4 ± 1.43.4 ± 0.1
162874.5 ± 3.183.0 ± 0.788.2 ± 1.484.5 ± 7.267.8 ± 3.678.4 ± 2.187.9 ± 2.685.0 ± 2.784.3 ± 3.977.3 ± 9.363.6 ± 23.773.8 ± 17.9
182611.6 ± 0.88.3 ± 0.711.6 ± 1.55.8 ± 4.727.5 ± 2.316.7 ± 1.16.1 ± 1.06.2 ± 0.28.8 ± 0.810.8 ± 2.817.2 ± 12.411.7 ± 5.6
20246.8 ± 1.34.5 ± 0.80.1 ± 0.30.8 ± 0.83.4 ± 3.43.2 ± 0.92.8 ± 1.44.0 ± 0.23.6 ± 0.73.0 ± 0.713.4 ± 10.85.7 ± 4.4
22224.1 ± 0.80.4 ± 0.7 1.4 ± 1.11.1 ± 1.21.0 ± 1.40.9 ± 0.9 0.8 ± 1.12.7 ± 4.71.4 ± 1.95.5 ± 8.0
2420   0.5 ± 0.80.1 ± 0.2 0.4 ± 0.6 0.9 ± 0.4   
C46 esters  0.06 ± 0.040.13 ± 0.020.16 ± 0.120.14 ± 0.060.15 ± 0.070.09 ± 0.020.12 ± 0.010.05 ± 0.010.04 ± 0.000.13 ± 0.000.18 ± 0.070.06 ± 0.01
1432 1.6 ± 2.2  0.2 ± 0.30.0 ± 0.00.2 ± 0.4    1.0 ± 1.7
163045.4 ± 5.567.1 ± 5.57.8 ± 3.889.9 ± 3.11.9 ± 1.610.3 ± 1.873.7 ± 3.460.7 ± 2.654.1 ± 4.366.9 ± 2.746.9 ± 0.447.8 ± 13.4
182820.8 ± 2.69.8 ± 1.630.0 ± 3.53.9 ± 0.416.1 ± 2.328.1 ± 3.511.7 ± 2.69.5 ± 2.515.3 ± 1.812.1 ± 0.213.6 ± 3.115.2 ± 4.3
202625.3 ± 0.618.4 ± 3.455.7 ± 2.54.9 ± 2.174.3 ± 5.458.9 ± 3.412.8 ± 3.015.2 ± 0.327.0 ± 3.918.9 ± 1.728.7 ± 2.931.1 ± 13.1
22246.0 ± 1.02.9 ± 1.96.5 ± 6.20.7 ± 0.14.3 ± 2.32.8 ± 0.41.6 ± 1.67.3 ± 5.72.5 ± 0.62.2 ± 1.18.7 ± 0.23.7 ± 3.4
24222.5 ± 1.30.3 ± 0.6 0.6 ± 0.73.2 ± 2.7  7.2 ± 0.41.1 ± 1.5 2.2 ± 0.71.3 ± 1.3
Total esters  0.17 ± 0.100.35 ± 0.041.12 ± 0.360.27 ± 0.120.71 ± 0.410.29 ± 0.090.29 ± 0.020.17 ± 0.040.20 ± 0.050.56 ± 0.030.47 ± 0.080.28 ± 0.08

Individual alkyl esters with chain lengths of C42, C44 and C46 reached levels that could be quantified in all samples with the methods used, whereas all other ester homologues were present only at trace levels (<0.005 µg cm−2). In the wild-type waxes C42, C44 and C46 esters had similar absolute levels between 0.03 µg cm−2 and 0.08 µg cm−2 (Table 1). In the mutant lines, individual homologues were typically accumulated to levels between 0.01 µg cm−2 and 0.20 µg cm−2, reaching coverages of 0.52 µg cm−2 and 0.30 µg cm−2 for the C42 ester in cer2 and cer6, respectively (Table 1). The homologue patterns showed a relatively equal distribution of the three major chain lengths of C42, C44 and C46 for all investigated lines. The mutants cer1, cer9, cer10, cer13, cer22 and cer24 had similar ester chain-length distributions as those found for the wild types. In contrast, the remaining four mutant lines clearly differed from the wild types, with a predominance of the shorter ester chain length C42 in cer2 and cer6, and a relative shift towards the longer C46 homologue in cer3 and cer23 (Table 1).

Chain length distribution of acyl moieties in wax esters

The initially acquired GC-MS dataset was further used to quantify the percentages of acyl moieties within the three major ester homologues in all samples (Table 1). In the wax esters of both wild types and of most mutant lines, 16:0 acyl moieties strongly dominated, with percentages of 64–99% in the C42 and C44 homologues and 45–90% in C46 (Table 1). Only in the C46 ester of cer2, cer6, cer9 and cer23 did the 20:0 acyl moieties reach higher percentages ranging from 56% to 74%. In this ester homologue, a bimodal distribution of acyl chain lengths with maxima for 16:0 and 20:0 was found, most notably for WS, cer1, cer13, cer23 and cer24. The remaining mutant lines and the Ler wild type showed similar isomeric acyl distribution in the C46 ester, but with less pronounced maxima.

Our results show that the biosynthesis of cuticular alkyl esters in A. thaliana stems proceeds with a strong preference for the 16:0 acyl substrate. This finding may reflect either the homologue distribution of acyl precursors available at the sub-cellular site of ester biosynthesis or the acyl substrate chain-length specificity of the ester synthase. Although our chemical data do not allow a distinction between these two possibilities, it can still be concluded that the ester synthase enzyme must have access to a pool of 16:0 acyl precursors that it is likely to utilize as CoA esters.

Chain-length distribution of alkyl moieties in wax esters

In order to qualitatively assess the alcohol substrates for ester synthesis, the homologue patterns had to be compared between free and esterified alcohols in the stem waxes of respective A. thaliana wild types and cer mutants. The relative levels of the esterified alkyl units could be calculated based on the ester acyl percentages and the overall ester chain lengths (Table 1), whereas the free alcohols were quantified using flame ionization detector (GC-FID) analysis of the same wax mixtures. The wax mixtures of both wild types showed similar chain-length patterns of free primary alcohols, with a maximum for the C28 homologue and with the C30 alcohol being present at higher concentrations than the C26 alcohol (Figure 1). The mutant lines cer23 and cer24 had chain-length patterns of free alcohols identical to those of the corresponding WS wild type. A number of mutants were found to have increased percentages of C30 free alcohol and decreased levels of C26 free alcohol instead, and this shift towards longer chain lengths was more pronounced in cer3 than in cer1, cer10, cer13 and cer22. In contrast, the remaining cer2, cer6 and cer9 mutants showed an increase of C26 free alcohol with an almost complete loss of the C30 homologue. These results closely match previously reported data for all the lines investigated (Jenks et al., 1995, 1996; Rashotte et al., 2001, 2004).

Figure 1.

 Chain-length distributions of free and esterified alcohols in the cuticular wax on stems of various Arabidopis thaliana cer mutants and corresponding wild types. Asterisks directly over the bars denote significant differences (P < 0.05) of mutant alcohol percentages from the corresponding wild-type values; asterisks above curved arrows show significant (P < 0.05) differences between esterified and free alcohol levels within the mutant wax.

The homologue distributions of the esterified alcohols in the wild-type wax mixtures differed from the corresponding patterns of free alcohols by slightly increased levels of C26 alcohol and corresponding decreases of the C30 homologue (Figure 1). This shift towards shorter ester alkyl chains may either reflect a preference of the ester synthase for alcohol substrates shorter than C30, or show that a part of the C30 alcohol pool is inaccessible for the enzyme. It will therefore be of particular interest to compare the chain-length specificities of wax ester synthases from both wild types.

The waxes of cer23 and cer24 were characterized by ester isomers containing alcohol chain-length patterns similar to those of the corresponding wild type, which also mimick the bias against the C30 homologue found in WS. Consequently, these mutants do not yield new information on wax ester biosynthesis and will not be discussed further.

The mutants cer2, cer6 and cer9 had chain-length distributions of esterified alcohols that differed drastically from the corresponding wild types, but at the same time closely matched the homologue patterns of the free alcohols accompanying them in the mutant wax mixtures. All three mutants had largely increased levels of C26 alkyl chains, whereas C30 was nearly missing. The matching homologue patterns between free and esterified alcohols in these mutants show that both export of free alcohols to the cuticle and wax ester biosynthesis in A. thaliana stems rely on a common pool of alcohol precursors. Our data also indicate that ester biosynthesis proceeds without chain-length specificity for alcohol substrates between C24 and C28.

The mutants cer1, cer3, cer10, cer13 and cer22 were characterized by a shift in the chain-length distribution of free alcohols from C26 and C28 towards C30. The corresponding alkyl chains in the esters showed a similar trend in all five lines, albeit less pronouncedly. Those mutants that had the most drastic increase in free C30 alcohol also had the strongest increase of this chain length in the ester alkyl distribution. For example, a significant decrease of esterified C26 and C28 alcohols was found for cer3, whereas C30 ester alcohol accumulated to higher levels than in the Ler wild type. Overall, the ester alcohol distributions of cer1, cer3, cer10, cer13 and cer22 were intermediate between the chain-length patterns of the free alcohols of these mutants on the one hand, and the ester alcohol patterns of the corresponding wild types on the other hand. This result can be explained combining the two major inferences drawn from the other mutant and wild-type ester compositions (see above): (1) the alcohols formed by fatty acyl reduction are used as a substrate for ester biosynthesis, so that an increase in C30 alcohol causes an increase in C30 alkyl moieties in the corresponding esters; (2) either the ester synthase has a preference for alcohols with chain lengths <C30 or this substrate is not completely available to the enzyme, so that an increase in C30 alcohol can consequently cause only a partial increase in C30 alkyl moieties in the corresponding esters.

It had been suggested that the free alcohols accumulating in the Arabidopsis wild-type wax are formed by a fatty acyl reductase encoded by CER4 in A. thaliana, and this gene has recently been cloned and characterized (Rowland et al., 2006). It was shown that alcohols with chain lengths C24–C28 are formed exclusively by CER4, and can serve as substrates for further reactions including wax ester formation. As the CER4 enzyme was shown to be located in the endoplasmic reticulum (ER) when expressed in yeast, export of the (C24–C28) free alcohol products involves trafficking from the ER to the plasma membrane. We conclude that the ester synthase must either be co-localized together with CER4 in the ER, or must intercept the wax alcohols during intracellular transport (possibly in the plasma membrane).

cer4 null mutants had reduced but unexpectedly high levels of free C30 alcohol in their wax (Rowland et al., 2006), showing that this particular alcohol chain length can be biosynthesized on two parallel routes by the CER4 reductase and by an unknown enzyme or pathway. This result can now be combined with our finding that C30 alcohol is relatively inefficiently incorporated into esters in the Arabidopsis wild-type stem: we hypothesize that only the pool of C30 alcohol originating from CER4-catalyzed reduction is available at the site of ester biosynthesis. The ester synthase, rather than showing chain-length specificity biasing against this chain length, does not have access to the second pool of C30 alcohol generated on the alternative route.

Limiting substrate pools for wax ester biosynthesis in Arabidopsis wild-type stems

To further characterize the substrate pools involved in ester biosynthesis, the total quantities of primary alcohols in the wax mixtures must be considered. They can be calculated by adding the coverages of free alcohols and esterified alcohols in the respective wax mixtures. As compared with the wild-type waxes, the total alcohols were most drastically increased for cer2, cer3, cer6 and cer9 (Table 2). Unfortunately, only one of the genes mutated in these lines has been characterized to date, thereby allowing an interpretation of the biochemical effects on free alcohols and esters. CER6 is known to be a condensing enzyme involved in fatty acid elongation rounds starting with 26:0 (Millar et al., 1999). A deficiency of this activity in the cer6 mutant results in the increased formation of C26 alcohol (Jenks et al., 1995). Our data show that the increased levels of C26 alcohol substrates available for ester biosynthesis results in an increase of the corresponding esters in the cer6 mutant, suggesting that alcohol substrate levels, rather than ester synthase activity, are limiting ester biosynthesis in the wild type. It is plausible that cer2, cer3 and cer9 also have increased ester levels as a result of the increased substrate levels of either C26 or C30 alcohols.

Table 2.   Coverages (µg cm−2) of esterified acids and total alcohols in the stem wax of cer mutants and corresponding wild types of Arabidopsis thaliana. Acid coverages were calculated by summing the quantities of respective isomers over all ester homologues. Corresponding quantities of esterified alcohols were added to coverages of free alcohols to yield the total alcohol quantities. Mean values (n = 3) and SD are given
 Lercer1cer2cer3cer6cer9cer10cer13WScer22cer23cer24
  1. aTraces, i.e. <0.005 μg cm−2 detected.

Esterified acid chain length
 14tra0.01 ± 0.002tr0.01 ± 0.0040.01 ± 0.0010.01 ± 0.0040.01 ± 0.0010.01 ± 0.004tr0.02 ± 0.010.01 ± 0.0010.01 ± 0.002
 160.11 ± 0.070.28 ± 0.030.91 ± 0.270.24 ± 0.110.42 ± 0.320.26 ± 0.060.24 ± 0.020.13 ± 0.030.12 ± 0.010.42 ± 0.010.28 ± 0.020.17 ± 0.02
 180.03 ± 0.010.03 ± 0.0050.10 ± 0.040.01 ± 0.010.09 ± 0.060.06 ± 0.010.02 ± 0.0020.01 ± 0.00030.01 ± 0.0020.04 ± 0.010.06 ± 0.030.03 ± 0.01
 200.02 ± 0.010.03 ± 0.0030.09 ± 0.070.01 ± 0.010.13 ± 0.080.06 ± 0.010.02 ± 0.010.01 ± 0.00010.01 ± 0.010.03 ± 0.0030.08 ± 0.040.03 ± 0.01
 220.01 ± 0.01tr0.02 ± 0.02tr0.01 ± 0.0030.01 ± 0.004trtrtr0.01 ± 0.010.02 ± 0.010.01 ± 0.02
 Total0.17 ± 0.100.35 ± 0.041.12 ± 0.360.27 ± 0.120.66 ± 0.450.39 ± 0.090.29 ± 0.020.17 ± 0.040.15 ± 0.0020.53 ± 0.030.45 ± 0.060.25 ± 0.02
Total alcohol chain length
 20     0.01 ± 0.0010.01 ± 0.0010.01 ± 0.005 0.01 ± 0.0010.005 ± 0.0010.01 ± 0.003
 220.01 ± 0.01trtr0.01 ± 0.0020.01 ± 0.00040.01 ± 0.0040.01 ± 0.0030.01 ± 0.0010.03 ± 0.0020.01 ± 0.010.008 ± 0.0050.01 ± 0.01
 240.03 ± 0.020.05 ± 0.010.03 ± 0.030.02 ± 0.010.13 ± 0.080.05 ± 0.010.02 ± 0.0040.03 ± 0.00010.04 ± 0.0030.03 ± 0.010.075 ± 0.010.02 ± 0.01
 260.27 ± 0.170.26 ± 0.023.49 ± 1.350.20 ± 0.032.41 ± 1.101.39 ± 0.070.19 ± 0.030.30 ± 0.040.41 ± 0.090.38 ± 0.070.512 ± 0.040.24 ± 0.05
 280.52 ± 0.340.45 ± 0.052.39 ± 0.820.40 ± 0.041.08 ± 0.311.55 ± 0.040.38 ± 0.050.58 ± 0.030.66 ± 0.230.52 ± 0.070.846 ± 0.170.47 ± 0.07
 300.32 ± 0.190.57 ± 0.060.01 ± 0.0011.49 ± 0.100.05 ± 0.040.13 ± 0.010.40 ± 0.040.84 ± 0.210.45 ± 0.220.40 ± 0.080.467 ± 0.070.31 ± 0.06
 32tr0.01 ± 0.0050.01 ± 0.010.01 ± 0.01tr0.01 ± 0.010.01 ± 0.01tr0.01 ± 0.010.01 ± 0.005trtr
 Total1.16 ± 0.741.34 ± 0.135.94 ± 2.192.13 ± 0.093.68 ± 1.533.14 ± 0.111.01 ± 0.111.77 ± 0.271.60 ± 0.531.34 ± 0.231.914 ± 0.271.06 ± 0.17

Finally, the absolute quantities of esters and alcohols must be further taken into account. cer1, cer10 and cer22 all had increased levels of esters compared with Ler and WS (Table 1), whereas the total levels of alcohols were identical to those found in the wild types (Table 2). This suggests that in these mutants ester biosynthesis was enhanced as a result of the increases in the available acyl substrate without a concomitant increase of the alcohol substrate. Consequently, increased levels of ester biosynthesis caused a partial depletion of the C26 and C28 alcohol pools, and resulted in free alcohol profiles deviating from those of wild-type cuticular waxes. This interpretation is in accordance with the biochemical role of CER10, as it has been characterized as the enoyl-CoA reductase involved in VLCFA elongation starting with the 16:0 fatty acid (Zheng et al., 2005). A mutation in this gene can therefore be expected to result in increased levels of 16:0 CoA. Unfortunately, the exact biochemical functions of the proteins corresponding to the other two cer genes are unknown, so their effect on ester biosynthesis cannot be further interpreted. Overall, our results can be explained by the hypothesis that 16:0 acyl CoA levels at the site of biosynthesis, together with the alcohol substrate levels (see above), are limiting ester formation in the wild type.

Conclusions

We have shown that the pool of free alcohols, formed by the fatty acyl reductase CER4, is the substrate for wax ester biosynthesis. The ester synthase must either be co-localized with CER4 in the ER, or must intercept the alcohols during intracellular trafficking either towards or in the plasma membrane. 16:0 CoA is the predominant acyl substrate for ester synthesis in Arabidopsis stem wax. Both the pool of very long-chain alcohols and the acyl levels at the site of biosynthesis are limiting wild-type ester formation. Our chemical data show that these two types of substrates must be co-localized with the ester synthase enzyme (or enzymes). Both substrates are formed at very different stages in the wax biosynthetic pathway, 16:0 CoA being the first intermediate en route to waxes (and other compounds such as cutin and membrane lipids) and alcohols being a relatively late product of the acyl reduction branch of the pathway. Arabidopsis wild-type ester biosynthesis either proceeds with preference for alcohols with chain lengths <C30 or this substrate is not completely available to the enzyme.

Experimental procedures

Plant growth and wax extraction

Wild-type (Ler and WS ecotypes) and mutant seeds (cer1, cer2, cer3, cer6, cer9, cer10 and cer13 in the Ler background) were obtained from the Arabidopsis Biological Resource Center (http://www.arabidopsis.org). cer22, cer23 and cer24 (WS ecotype) were a gift from Dr Matthew Jenks (Purdue University; http://www.purdue.edu/). Seeds were stratified for 3–4 days at 4°C, and were then germinated on AT-agar plates (Somerville and Ogren, 1982) at 20°C under continuous light (100 μE m−2 sec−1 of photosynthetically active radiation). Seedlings (10 days old) were transplanted to soil (Sunshine Mix 5; SunGro, http://www.sungro.com/) and grown at 20°C under continuous light. All plants were grown in one batch in parallel, with multiple pots for each line containing 3–5 individuals.

For total wax extraction, inflorescence stems from 6-week-old Arabidopsis plants were harvested and flowers, siliques and cauline leaves were removed. The resulting stem sections were immediately immersed twice for 30 sec in CHCl3 at room temperature and a defined quantity of n-tetracosane was added as an internal standard. The resulting solutions were filtered, dried and stored at 4°C until they were analyzed. The extracted surface area was determined by measuring the height and diameter of the stems assuming cylindrical geometry.

Chemical analysis

Prior to GC analysis, chloroform was evaporated from the samples under a gentle stream of N2 while heating to 50°C. Then the wax mixtures were treated with bis-N,N-(trimethylsilyl)trifluoroacetamide (BSTFA; Sigma-Aldrich, http://www.sigmaaldrich.com/) in pyridine (30 min at 70°C) to transform all hydroxyl-containing compounds into the corresponding trimethylsilyl (TMSi) derivatives. The qualitative composition was studied with capillary GC (5890 N, HP-1 column 30 m, 0.32 mm i.d., film thickness = 0.1 µm; Agilent, http://www.home.agilent.com/) with He carrier gas inlet pressure programmed for constant flow of 1.4 ml min−1 and mass spectrometric detector (5973 N, Agilent). GC was carried out with a temperature-programmed injection at 50°C, oven temperature for 2 min at 50°C, then raised by 40°C min−1 to 200°C, held for 2 min at 200°C, raised by 3°C min−1 to 320°C and then held for 30 min at 320°C. Individual wax components were identified by comparison of their mass spectra with those of authentic standards (C22 and C26 alcohol esters of 16:0, 18:0, 20:0 and 22:0 acids) and literature data.

The quantitative composition of the mixtures was studied using capillary GC with flame ionization detector under the same GC conditions as above, but with H2 carrier gas inlet pressure regulated for a constant flow of 2 ml min−1. Single compounds were quantified against the internal standard by automatically integrating peak areas. The isomer composition of individual ester homologues was determined using GC-MS as described above. Averaged mixed spectra were obtained for the entire isomer mixture, and relative abundances of all fragments RCOOH2+ were quantified (Sümmchen et al., 1995b). All quantitative data are given as means and SD. Statistical analysis was performed using spss 13.0 (independent samples t-test not assuming equal variances, P < 0.05; SPSS, http://www.SPSs.com/).

Acknowledgements

This work has been supported by the National Science and Engineering Research Council (Canada), the Canada Research Chairs Program, and the Canadian Foundation for Innovation.

Ancillary