• plant cuticle;
  • epicuticular wax;
  • elongation system;
  • eceriferum;
  • desaturase


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

About 15% of the epidermal wax on Hordeum vulgare cv. Bonus barley spikes is n-alkanes. Longer homologues are greatly reduced in the eceriferum mutants, cer-a6, cer-e8, cer-n26, cer-n53, cer-n985, cer-x60, cer-yc135 and cer-yl187. Simultaneously hydrocarbons accounting for only traces in the wild-type become prominent in the mutants, although their chain-length distributions remain unchanged. Accordingly several new hydrocarbon series were identified. The two major ones were C23–C35cis monoenoic alkenes (the major 9-ene isomer was part of a homologous series including 11, 13 and 15-enes), and the novel C27–C31 cyclopropanes (the ring carbons of major isomers were 9,10 and 11,12 with lesser amounts of 13,14). Three minor series included 2- and 3-methylalkanes plus C25–C33 internally branched alkanes (methyls on carbons 9, 11, 13, 15 or 17; shorter homologues dominated by the 9 isomer, longer homologues by 11, 13 or 15 isomers). Acyl chains destined for spike waxes are synthesized via acyl and polyketide elongase systems plus associated reductive and decarbonylative/decarboxylative enzyme systems. Both elongation systems are defective in synthesizing C32 acyl chains in all nine mutants. The similarities in the position of the chemical groups (primarily on carbon 9, secondarily on carbon 11) of the alkenes, cyclopropanes and internally branched methyl alkanes imply an origin from a common, hitherto unrecognized associated pathway in barley, designated the enoic pathway. The elongation system leading to the enoic derived hydrocarbons differs from the known elongation systems by inclusion of a mechanism for introducing a double bond.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Enabling the movement of the ancestors of plants from their aquatic environment onto land required the development of a lipid coating as protection against water loss. This lipid coat or cuticle on the aerial surface of plants frequently consists of an aliphatic, outer-most layer of epicuticular wax decorating a hydrophobic layer composed of the cutin polymer in which intracuticular waxes are embedded. The lipid classes composing the surface wax generally include very-long-chain n-alkanes, aldehydes, primary alcohols, esters and free fatty acids as in Zea mays (Bianchi et al., 1984), which are often supplemented by other lipids such as the secondary alcohols and ketones characterizing Arabidopsis thaliana (Kunst and Samuels, 2003). Early experiments suggested that C18 acyl chains synthesized by fatty acid synthase (FAS) served as precursors for further elongation by a membrane-localized enzyme complex, an elongase (ELS) consisting of a condensing enzyme that added a C2 unit, giving an acyl chain that was subjected to keto reduction, dehydration and enoyl reduction reactions yielding a saturated chain substrate for the next extension. Action of one or more ELS complexes gave elongated products that (i) served as precursors for associated enzyme complexes such as the acyl reduction pathway (giving even-chain-length aldehydes, primary alcohols and esters) and the acyl decarboxylation/decarbonylation pathways (forming odd-chain-length n-alkanes, secondary alcohols and ketones), or (ii) were exported to the plant surface giving free fatty acids (von Wettstein-Knowles, 1995). These early proposals are now being substantiated by the isolation and characterization of ELS genes, namely the Arabidopsis CER6 and KCS genes encoding condensing enzymes (Fiebig et al., 2000; Todd et al., 1999), the Zea mays gl8a and gl8b genes encoding keto reductases (Dietrich et al., 2005; Xu et al., 2002) and the Arabidopsis CER10 gene encoding an enoyl reductase (Zheng et al., 2005). Genes for reductive pathway components include Arabidopsis CER4 encoding an aldehyde reductase, closely related to the Simmondsia chinensis FAR gene that participates in primary alcohol synthesis in seeds (Metz et al., 2001). From the latter species, a gene encoding a wax synthase forming esters in seeds has also been isolated (Lardizabal et al., 2000). In addition, the Arabidopsis CER5 gene encodes a plasma membrane-localized ABC transporter that is involved in moving the wax lipids toward the cuticle surface (Pighin et al., 2004).

Acyl ELS complexes differ with respect to their specificity for the elongation substrate, as do FAS complexes (von Wettstein-Knowles, 1993). Radioactive tracer, inhibitor and genetic studies established that, in barley, a minimum of three acyl ELS complexes participate in the seven C2 additions extending 18 carbon chains to give 20, 22–26 and 28–32 carbon chains Mikkelsen, 1978). The resulting acylCoAs are substrates for aldehydes, primary alcohols, esters, n-alkanes, secondary alcohols and free fatty acids. Of these, the primary alcohols are the predominating lipid class on leaves and lower leaf sheaths. On the uppermost leaf sheaths and internodes plus the spikes (floral structure), synthesis of the predominating wax lipids requires a different type of ELS. These polyketide ELS complexes omit the three tailoring reactions following specific C2 extensions, resulting in a substrate for the next elongation containing a 3-oxo group (von Wettstein-Knowles, 1993). The polyketide ELS complexes exhibit specificity for the elongation substrate, and 3-oxo groups are retained in only one or two successive elongation steps. The resulting oxoacyl-CoAs serve as substrates for (i) the decarboxylative pathway in which associated enzymes produce alkan-2-ol esters and oxoalkan-2-ol esters, and (ii) the decarbonylative pathway in which associated enzymes give rise to β-diketones, hydroxy-β-diketones and oxo-β-diketones. On wild-type Bonus barley spikes, the β-diketone lipids account for at least 50% of the epicuticular wax, and the primary alcohols and hydrocarbons account for 3–4 and 15–16%, respectively (von Wettstein-Knowles and Søgaard, 1980). The major cuticular surface of the spike that is responsible for its overall wax phenotype is that of the lemmas enfolding the seeds, while paleas, glumes, rachis and awns make minor contributions. Structural and/or wax lipid class analyses reveal that β-diketone lipids are present on all surfaces except those of the awns (Simpson and von Wettstein-Knowles, 1980; von Wettstein-Knowles, unpublished results). Hydrocarbons are more prominent constituents of awn and glume wax than they are of the total spike wax (Simpson and von Wettstein-Knowles, 1980). Moreover, awn waxes are so far the only cuticle surface in barley that has been shown to bear secondary alcohols (von Wettstein-Knowles and Netting, 1976).

A total of 1580 eceriferum (cer) mutants distributed among 79 loci were isolated in barley based on a visual decrease in the glaucousness of the leaves, uppermost leaf sheaths and internodes and/or spikes (Lundqvist and Lundqvist, 1988). The wax phenotype on each of these three surfaces is designated by ‘−’ (absent), ‘+’ (less than wild-type) or ‘++’ (as wild-type). Thus, the wild-type wax phenotypic formula is ‘++ ++ ++’ with the cuticle surfaces in the order specified above. The 55 recessive and one dominant mutation at the cer-n locus, affecting spike and uppermost leaf sheath and internode wax, exhibit a range of phenotypes that are susceptible to environmental factors. The overall spike wax composition of ten cer-n mutants was therefore investigated (Lundqvist and von Wettstein-Knowles, 1983) using plants grown under controlled conditions in the Stockholm Phytotron (Dormling et al., 1969, 1972). In the mutant waxes, the β-diketone lipids varied from wild-type amounts to 8% thereof, and the dominating C31 alkane varied from approximately 71% as in wild-type to 8% of the hydrocarbons. The effect on the two lipid classes was not correlated, however. The primary alcohol spectrums dominated by the C26 homologue could resemble that of the wild-type or have increased amounts of C20, C22, C30 or C32 homologues. In the most extreme case, the C32 alcohol accounted for 38%, versus 19% for C26. The observed changes in the β-ketoacyl ELS derived lipids were not correlated with those of the acyl ELS derived lipids, nor were those of the two acyl ELS derived lipids, the n-alkanes and primary alcohols, correlated. Interestingly, the reduction in long-chain n-alkanes was accompanied by apparent increases in long-chain minor hydrocarbons in all the mutants. This implies that a different elongation system may be involved in their synthesis, where the term ‘elongation system’ includes not only the ELS complexes but also the origin of the substrates they use. As a first step to probe this hypothesis, the identity of the minor hydrocarbons has to be established. Material for this investigation included four mutants of the cer-n locus plus a mutant at each of five additional loci, namely cer-a, − − ++; cer-e, − ++ ++; cer-x, − − ++; cer-yc, − ++ ++, and cer-yl, − − ++, in which the long-chain n-alkanes are also reduced without apparently affecting the chain-length distributions of the minor hydrocarbons. The results identify five new hydrocarbon series in wild-type and eceriferum (cer) mutant barley spike epicuticular waxes. The prominent ones are monoalkenes (C23–C35) and cyclopropanes (chain lengths C27–C31), novel wax lipids. The minor series include internal methyl branched alkanes (chain lengths C25–C33), plus 2- and 3-methyl branched alkanes. The nature of the potential elongation systems participating in synthesis of the monoenoic alkenes, the cyclopropanes and the internally branched alkanes is considered.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Hydrocarbons were isolated by preparative thin-layer chromatography (TLC) from the epicuticular waxes obtained from the spikes of the wild-type Bonus and nine cer mutants. The chain-length distributions of those from Bonus and the mutant cer-n985 are shown in Figure 1. That of the wild-type, which is analogous to previous determinations (see, for example, von Wettstein-Knowles and Netting, 1976), consists predominantly of odd-chain-length n-alkanes that are greatly reduced in cer-n985 and the other eight cer mutants. Most marked is the reduction of C31 from approximately 72% to 10%, which is compensated for by the increased relative abundance of the shorter n-alkanes, especially C23 and C25 (1–2 to 16–19%), and two other homologous series: the x series eluting just before the odd-chain-length n-alkanes and the y series eluting just before the even-chain-length n-alkanes. The objective of this study was to identify the x and y series of hydrocarbons whose syntheses were apparently not reduced in the spike waxes of cer mutants characterized by defects in the elongation system(s) leading to n-alkane and β-diketone formation. In the course of this work, several other hydrocarbon series were observed, and some of these were also identified.


Figure 1.  Hydrocarbons from the epicuticular spike waxes of the wild-type Bonus barley (a) and its mutant cer-n985 (b) as resolved by GC using a 3% SP-2100 column. Chain lengths of the n-alkanes are specified. The letters x and y denote members of two homologous hydrocarbon series that are shown herein to be alkenes and cyclopropanes, respectively.

Download figure to PowerPoint


Initially the hydrocarbons were subjected to AgNO3 preparative TLC which gave two bands, both of which were recovered and analyzed by gas chromatography (GC). The slower band contained the x series, suggesting they were unsaturated. Hydrogenation of the x series resulted in the formation of straight-chain, odd-numbered n-alkanes whose identities were confirmed by TLC, GC and mass spectrometry (MS). The chain-length distributions of the unsaturated homologues from Bonus and the nine cer mutants are given in Table 1. Examination of the data reveals that the distributions from all 10 genotypes are similar, i.e. odd homologues from C23–C37 characterize all genotypes, with C39 and C41 also being detected in some of the genotypes. The major homologue is always C29, amounting to 33% in Bonus and somewhat more in the mutants (42–54%). The mean distribution of the nine mutants that were grown at various times over a three-year period (with SEs) implies that the differences among them are not significant, i.e. the nine mutations have no effect on alkene synthesis. Combining results from the total hydrocarbon and alkene distributions (see Experimental procedures) reveals that the unsaturates represent approximately 0.3% in Bonus, and 10.9, 25.2, 18.9 and 15.7%, respectively, in cer-n26, −n53, −n625 and −n985, for example. To determine the configuration of the unsaturated bond, total spike wax of mutants was applied to the silica gel end of a hybrid TLC plate. Development with hexane separated the hydrocarbons from the other wax components. Upon encountering the AgNO3-containing gel, the mobility of the unsaturates was retarded. Co-chromatography with cis and trans 9-C23:1 standard alkenes, which had Rf values of 0.38 and 0.57, respectively, revealed that the barley unsaturates were cis alkenes. No band corresponding to alkenes was detected on the hybrid plates when wild-type awn wax having the same amount of hydrocarbons as the mutants was analyzed.

Table 1.   Monoenoic alkenes present in the epicuticular spike waxes of wild-type Bonus and nine cer mutants (weight%)
Carbon numberBonusa6e8n26n53n624n985x60yc135yl187Mean mutant value ±SE
  1. aPossibly contaminated as judged by the shape of the peak. bC29 alkenes are given as an approximate weight percentage of the total hydrocarbons.

233.0a 0.7 1.3 0.8Trace 0.8 1.4 0.6 0.2 1.10.8 ± 0.16
241.2aTrace 0.3 0.3TraceTrace 0.1TraceTrace 0.30.1 ± 0.05
2514.7 7.012.3 9.2 6.4 7.2 9.9 7.8 6.510.38.5 ± 0.76
260.7Trace 0.9TracetraceTraceTraceTraceTrace 0.50.2 ± 0.11
2714.414.919.015.714.116.914.514.512.915.815.4 ± 0.59
280.9 0.7 0.7 0.7 0.5Trace 0.5 0.8Trace 0.80.5 ± 0.11
2932.652.541.944.154.153.845.947.352.746.248.7 ± 1.53
301.6 1.6 1.4 1.3 1.4 0.4 1.5 2.7 1.4 1.31.4 ± 0.20
3119.319.818.722.720.119.120.721.123.218.720.5 ± 0.55
320.7TraceTraceTraceTraceTraceTraceTraceTrace 0.4Trace ± 0.04
334.6 1.5 2.0 2.6 1.7 1.3 3.2 2.4 1.9 2.32.1 ± 0.20
353.4 1.0 1.3 1.7 1.1 0.3 1.8 1.9 1.1 1.61.3 ± 0.17
371.8 0.4 0.2 0.7 0.4Trace 0.5 0.8Trace 0.70.4 ± 0.10
390.8   0.1Trace     0.10.2 ± 0.01
29b0.1 3.4 5.5 5.512.910.0 7.8 4.1 3.5 0.5 

To determine the positions of the double bonds in the alkenes, total hydrocarbons were subjected to OsO4 oxidation converting the alkenes to diols, which were then analyzed by GC-MS of their trimethylsilyl (TMS) derivatives. GC traces of cer-e8 hydrocarbons before and after the oxidation are shown in Figure 2. The TMS derivatives elute 3.76 equivalent chain-length (ECL) units later than the respective alkene, for example the C29-TMS elutes before the C33n-alkane with an ECL of 32.43 (Figure 2b), while the alkene elutes before the C29n-alkane with an ECL of 28.67 (Figure 2a). That only one homologous series results from OsO4 oxidation implies that only one type of alkene is present. Upon GC-MS fragmentation between α-oxytrimethylsilyl (OTMS)-substituted carbons, strong intensity ions whose m/z in combination with that of the M-15 ion reveal the number of carbons on either side of the original double bond (Capella and Zorzut, 1968). When more than one isomer is present, the mass spectrum obtained depends on the time it was taken during elution of the GC peak. This is illustrated for the cer-a6 C29-TMS GC peak in Figure 3(a,b). Two major positional isomers with double bonds at 9 or 11 (Figure 3a) are revealed by the pairs of ion fragments at 215/369 and 243/341 plus an M-15 ion at 569. In addition, the ion fragment pairs at 271/313 and 285/299 intimate the presence of lesser amounts of the 13- and 14-ene isomers. The latter numbered from the other end of the chain is a 15-ene. Such differences are summarized in Table 2A by the chemical nomenclature designation/opposite end designation, in this case 14/15-ene. In the text, however, isomers are designated as members of a homologous series differing by C2-units. The spectrum in Figure 3(b) is dominated by the 215/369 ion fragment pair indicative of a double bond at position 9, with small amounts of the ion fragment pairs 243/341 and 187/397 plus an M-15 ion at 569, implying the presence of 7- and 11-ene isomers. Table 2A summarizes all alkene positional isomers occurring in cer-yl187 epicuticular spike wax detected during MS analyses at representative time points throughout the elution of each GC peak. For all homologues, the 9-ene was the major isomer, as at peak height, the diagnostic ion fragments of any other isomer never amounted to more than 20% of those of the C9 isomer, generally to less than 10%. While the 11-, 13- and 15-enes were generally present, whether or not 17- to 23-enes also occur is unknown at present as only one of the characteristic ion fragments from each pair was present in the spectra. Analogous positional isomers characterize the monoenoic cis alkenes from cer-a6 and cer-e8.


Figure 2.  Hydrocarbons from the epicuticular spike wax of cer-e8 before (a) and after (b) formation of the diol-TMS derivatives of the alkenes by OsO4 oxidation as resolved by GC chromatography using a 3% SP-2100 column. Dotted lines are used for n-alkanes. Black and x denote the alkene homologues in (a), and their diol-TMS derivatives in (b). Chain lengths of the n-alkanes are specified.

Download figure to PowerPoint


Figure 3.  Mass spectra of the diol-TMS derivatives generated from the C29 alkene of cer-a6 epicuticular spike wax resolved by GC using a 5% OV-101 column and the VG 7070F instrument. (a) Leading side of the GC peak with two major isomers and two minor isomers. The origins of the major 11-ene 243/342 and minor 15-ene 285/299 fragments are shown. (b) Tailing side of the GC peak with one major isomer and two minor isomers. The origin of the major 9-ene ion fragment 215/369 is shown.

Download figure to PowerPoint

Table 2A.   Monoenoic alkenes from cer-yl187 barley spike epicuticular wax
Chain lengthDouble bond positions
  1. Double bond positions revealed by GC-MS of the diol-TMS derivatives of OsO4 oxidation products as deduced from mass sepctrograms taken at various times during elution of each homologue.

25911, 10/15, 12/13, (8/17)
27911, 12/15, 10/17
29911, 13, 14/15, (10/19), (8/21)
31911, 13, 15, (10/21), (8/23)

Even in the earliest studies of plant long-chain alkenes, double bond positions were identified. For example, in the wax from Papaver rhoeas anthers, cis 9-, 7- and 11-enes of C31 were identified, while C27 was the dominant long-chain n-alkane (Stránský and Streibl, 1969). In Rosa damascena flower waxes, the major alkene was C29 with cis 5 or 7 double bonds, and the major n-alkane was C31 (Wollrab, 1968). A later more detailed study on petal waxes of three Cistus species revealed that the C23–C33 alkenes were primarily cis 9-enes, but C23–C25 could also be 1-enes and C27–C33 could be 5- or 11-enes, while the major n-alkane was C25 (Gülz et al., 1979). Thus, in common with the barley monoenoic alkenes, those from other plant waxes are characterized by a series of positional isomers and homologue distributions that differ from those of the n-alkanes. Dienoic alkenes were also identified in early studies, for example in flower wax of R. damascene, but double bond positions were not characterized (Stoianova-Ivanova et al., 1971).

Branched hydrocarbons

The chain-length distributions of the saturated hydrocarbons recovered from the faster running band on the AgNO3 TLC plates are given in Table 3. These compounds from the wild-type spikes are primarily odd-chain-length n-alkanes dominated by the C31 homologue, with trace amounts of even-chain-length homologues (von Wettstein-Knowles and Søgaard, 1980). In all nine mutant waxes, the importance of C31 is markedly decreased, while the relative amounts of the C23, C25 and C27 homologues are significantly increased. In addition, the z and y homologous series of hydrocarbons eluting with ECLs of 27.34, 29.34, 31.34 and 27.86, 29.86 and 31.86, respectively, become visible components of the saturated hydrocarbons (compare Figure 1a and b). As the SP-2100 column did not adequately resolve the z and y homologues from the adjacent n-alkane, the saturated hydrocarbon fraction was treated with a molecular sieve to reduce the relative amounts of the latter. This resulted in the simultaneous removal of the y series, revealing that they were not branched hydrocarbons.

Table 3.   Saturated hydrocarbons from epicuticular waxes on spikes of wild-type Bonus and nine cer mutants after AgNO3 TLC removal of alkenes as determined by GC (weight%)
  1. aEquivalent chain length (ECL) determined from retention times versus the carbon numbers of standards during temperature-programmed analyses using a 3% SP-2100 column. Trace <0.1%. Hydrocarbons with an ECL ending in .34 belong to homologous series z, those ending in .86 belong to homologous series y, while 26.64 is 2-methylheptacosane. The abundance of the two longer y series homologues is over-estimated, especially for Bonus, as they are not resolved from other minor hydrocarbons (for example, see Figure 4).

20 1.1       0.1
21 6.5
22Trace0. 0.7
23 2.819.811.312.619.418.
24Trace1. 0.8
25 2.914.09.89.816.
26Trace0. 0.3
26.64 0.30.1  TraceTrace0.3 0.2
27 8.3
27.17 0.60.3   0.9 0.70.6 0.6
27.86 1.3
28 1.3
29 + 29.3422.117.320.221.116.515.620.919.915.417.3
29.86 8.0
31 + 31.3464.48.835.129.79.9 8.612.819.948.78.7
31.86 3.7
33 1.6
35 0.4        

The molecular sieve-treated hydrocarbons were subjected to GC-MS to determine whether some of the minor branched components shown in Figure 4 could be identified. The molecular ion is generally not visible in electron-impact MS of long-chain methylalkanes which was used in the present study. Instead an M-15 ion and others of lesser mass are formed that can be used to deduce what the molecular ion was (Blomquist et al., 1987; Hamilton, 1995). A series with ECLs of 24.65, 26.65, 28.65 and 30.65 was characterized by M-CH3 and M-C3H7 ion fragments (337 + 309, 365 + 337, 393 + 365, 421 + 393), identifying them as iso or 2-methyl alkanes, namely 2-methyltetra-, hexa-, and octacoasane plus 2-methyltriacontane. In accordance with this, 13-methyltetra- and 15-methylhexadecanote from Supelco's bacterial fatty acid methyl ester mix have ECLs of 14.65 and 16.65 on the same HP-1 column. Two members of a second series with ECLs of 27.73 and 29.73 were characterized by M-CH3 and M-C2H5 (379 + 365, 407 + 393) ion fragments, identifying them as anteiso or 3-methylalkanes, namely 3-methylhepta- and nonacosane. In addition, members of two other homologous branched series having an ECL with either an odd or even carbon number + 0.34 are seen in Figure 4, labeled ‘ib’ (internal branch). The common ECL implies an analogous position of the methyl branch in both series. The MS data revealing the positions of the methyl branches are presented below. The additional unmarked peaks shown in Figure 4 were not investigated.


Figure 4.  Segment of a GC trace of the saturated hydrocarbons from the epicuticular spike wax of cer-a6 present after treatment with a molecular sieve obtained using a methylsilicone HP column. Numbers denote the chain length of the hydrocarbon. 2-, 3- and ib- denote the presence of a methyl group on carbons 2 or 3 or internally (ib, internal branch), respectively.

Download figure to PowerPoint

Cleavage of internally branched hydrocarbons upon chemical ionization occurs on either side of the methyl group, giving rise to two ion fragments. This is illustrated for the internally methyl branched C31z homologue (methylhentriacontane) from the epicuticular wax of cer-a6 spikes in Figure 5(a). A methyl group on carbon 15 of a C31 chain gives two characteristic ion fragment doublets, namely 225 + 224 and 253 + 252, where the doublets represent the primary and secondary carbonium ion fragments (Nelson et al., 1980). Hereafter, only the member of the doublet with the higher mass will be referred to. The ion fragment pairs 197/281, 169/309 and 141/337 would result from methyl groups on carbons 13, 11 or 9, respectively. Stronger spectra were subsequently obtained using the VG Trio-2 instrument, and the relevant portion for methyl branch determination of a MS from the analogous z homologue from cer-yl187 spikes is shown in Figure 5(b). As noted above, the observed isomer distribution depends on the time during homologue elution that the spectrum is taken. A summary of the methyl positions for all the cer-yl187z homologues is given in Table 2B. For the three shorter homologues, at peak height the major isomer had a methyl on carbon 9, whereas for the C31 and C33 homologues the methyls were prominent at two positions, namely 11 + 13 and 13 + 15, respectively. The diagnostic ion fragments of all other isomers never amounted to more than 10% of the specified isomer(s). Examination of Figure 5(b) also reveals fragment ions at 323, 295 and 267, which, although unaccompanied by clear doublets for the expected smaller ion fragment, imply that smaller amounts of isomers with the methyl group on carbons 10, 12 and 14 may also be present. A complete analysis of the cer-a6z homologues gave analogous results. Identification of the second series of internally branched methyl hydrocarbons eluting after the corresponding even-chain-length n-alkane (Figure 4) is based on weak spectra from a GC-MS analysis of the C30 homologue from the epicuticular waxes of cer-a6 and cer-yl187. Both spectra revealed the ion fragment pairs 155/281, 181/253 and 222/225, indicating methyl groups on carbons 10, 12 and 14 and an M-15 fragment ion at m/z 393. The tentative identification of the C28 homologue (Figure 4) is based on its having an analogous ECL (28.34) to that of the C30 homologue (30.34).


Figure 5.  Mass spectra of internally branched methylhentriacontane, a series z homologue, from two mutants, and generated by acidic hydrogenation of the cyclopropane methylenehentriacontane, a series y homologue. (a) Complete spectrum of the homologue from cer-a6 epicuticular spike wax, resolved using a 5% OV-101 column with the VG 7070F instrument. (b) Portion of the mass spectrum of the homologue from cer-yl187 epicuticular spike wax, resolved using a methylsilicone HP column with the VG-Trio 2 instrument. (c) Portion of the mass spectrum of the compound generated by acidic hydrogenation of the cyclopropane homologue from cer-x60 epicuticular spike wax (see Figure 6c), resolved using a methylsilicone HP column with the VG-Trio 2 instrument. Only the m/z for the ion fragment of greater mass in each doublet representing the primary and secondary carbonium ions is specified in the spectra.

Download figure to PowerPoint

Table 2B.   Internally branched hydrocarbons from cer-yl187 barley spike epicuticular wax
Chain lengthCarbon numberMethyl group positions
  1. Positions of methyl group revealed by GC-MS.

2526911, 13
2728911, 13
2930911, 13, 15
313211, 139, 15
333413, 159, 11, 17

Internal methyl branched hydrocarbons in waxes from walnut leaves were first reported many years ago (Stránskýet al., 1970). The homologues ranging from C21–C33, with C27 and C29 being the most important, had methyl groups on carbons 9, 11, 13 or 15. Associated with the C28 GC peak were mass ions characteristic for methyl groups on carbons 12 and 14, indicating that the methylalkanes were not resolved from octacosane by the column used. These results are rather similar to those reported here for barley spikes, and for extracts of dried chamomile flowers where other typical wax lipids were also present (Stránskýet al., 1981).


Initial MS of the y series of homologues of the cer-e8 saturated hydrocarbons (not shown) unexpectedly implied that they might be alkenes with an even number of carbons. More interestingly, the ECLs of the y series of hydrocarbons relative to the other saturated components depended upon whether they were subjected to GC using a non-polar column (for example, SP-2100 in Figure 1) or a polar column. To illustrate, the y homologues from cer-x60 in Figure 6(a) resolved by GC on the polar HP-17 column have ECLs of 28.15, 30.15 and 32.15 versus 27.86, 29.86 and 31.86 on the SP-2100 column (Table 3) and 27.90, 29.90 and 31.90 on the HP-1 column. This effect of column polarity on retention is characteristic of cyclopropanes (see Experimental procedures). Moreover, if the y series were cyclopropanes, then the similarity of the initial MS to that of alkenes would be explained (Christie and Holman, 1966). To confirm this identification, the saturated hydrocarbons of cer-x60 epicuticular spike wax were brominated, resulting in the expected disappearance (Brian and Gardner, 1968) of the deduced cyclopropanes from the mixture as illustrated in Figure 6(b). Another aliquot of the cer-x60 saturated hydrocarbons was subjected to acidic hydrogenation, which opens a cyclopropane ring giving rise to a pair of internally branched hydrocarbons or a straight-chain n-alkane depending on which of the three carbon–carbon bonds in the ring is cleaved (McCloskey and Law, 1967). This is shown in Figure 6(c) in which the peaks representing the cyclopropane components are absent and two others are increased. For example, acidic hydrogenation of the cyclopropane y homologue eluting between the C29 and C31n-alkanes (Figure 6a) resulted in marked increases in the C30n-alkane peak and the internal methyl branched C29 peak (Figure 6c). These results are in accordance with the hypothesis that the y series of hydrocarbons comprises cyclopropanes with chain lengths of 27, 29 and 31 carbons, i.e. methyleneheptacosane, methylenenonacoasane and methylenehentricontane.


Figure 6.  Saturated hydrocarbons from the epicuticular spike wax of cer-x60 (a) as recovered from AgNO3 TLC plates, (b) after bromination, and (c) after acidic hydrogenation as resolved by GC using an HP-17 column. cp, cyclopropanes of series y; n, normal; ib, internal methyl branch in (a) and (b) = series z.

Download figure to PowerPoint

To confirm the identity of the branched compounds resulting from acidic hydrogenation as well as to determine the position of the ring, they were analyzed by GC-MS as described above for the internally branched methyl alkanes. Results for the C31 homologue from the epicuticular spike wax of cer-x60 are shown in Figure 5(c). The following ion fragment pairs resulted: 141/337, 155/323, 169/309, 183/295, 197/281 and 211/267. For the fragment ions at m/z 183, 197 and 211, only one member of the expected doublet can be seen. The relatively equal amounts of the 267 + 281, 295 + 309 and 323 + 337 ion fragments presumably reflects the randomness of the chemical acidic hydrogenation that breaks the carbon–carbon bonds in the cyclopropane ring. By comparison, the mass spectrum of the internally branched C31z homologue analyzed in Figure 5(b) from the epicuticular wax of cer-yl187 spikes reveals quite different amounts of the 281 and 309 versus the 267 and 295 ion fragments, presumably implying specificity of the enzyme participating in their synthesis. Analogous results were obtained for the cer-x60 C27 and C29z homologues as well as for the three cer-a6 homologues.

A summary of the positions of the cyclopropane rings in cer-x60 is given in Table 2C. Characteristic for all three y homologues is the approximately equal frequency of the ring on carbons 9,10 versus 11,12. Moreover, at peak height, the diagnostic ion fragments for the specified isomers amounted to at least 70% of all possible isomer fragments. The distribution of the three homologues is similar in all nine mutants with mean values ±SEs of 7.7 ± 0.4 for C27, 61.2 ± 0.8 for C29 and 31.0 ± 1.1 for C31, implying that synthesis of the cyclopropanes is not affected by these mutations. As only the major C29y homologue is detected in the wild-type saturated hydrocarbons, the increases in abundance of the cyclopropanes would appear to be significant in all the mutant hydrocarbons, with the smallest increase occurring in cer-yc135 and the largest in cer-a6, cer-n53, cer-n624 and cer-yl187. In the spike wax of cer-x60, the cyclopropanes account for 9.3% of the total hydrocarbons (weight percentage), which is similar to 8.7% for the alkenes. At best, the internally branched, odd–chain-length hydrocarbons (C25–C33) amount to 1.8%, which is more than any of the other methyl branched series identified (Figure 4).

Table 2C.   Cyclopropanes from cer-x60 barley spike epicuticular wax
Chain lengthCarbon numberCyclopropane ring carbons
  1. Carbons participating in the cyclopropane ring were identified by GC-MS of the TMS derivatives of the methyl branched products resulting from acidic hydrogenation.

  2. *In those instances where the location of the functional group differs with respect to chemical nomenclature and numbering from the opposite end of the carbon chain, the former precedes and the latter follows the solidus (/). Text in parentheses indicates that only one fragment ion of pair is present.

27289,10 + 11,1213,14
29309,10 + 11,1213,14 + (14,15/15,16)
31329,10 + 11,12(13,14)

Carbocyclic fatty acids occur in the Malvales and plants allied with this order, plus some gymnosperms. Generally they have a chain length of 18 or 17 carbons, with the ring including carbons 9 and 10 joined by either a single or double bond (Christie, 1970), and are known as cyclopropane and cyclopropene acids, respectively. While especially prevalent in seed oils, they occur also in vegetative tissues, among which roots have the highest concentration (Christie, 1970; Schmid and Patterson, 1988; Shenstone et al., 1965). In seed oils, they are primarily sequestered in triacylglycerols, but are also esterified to phospho- and glycolipids (Christie, 1970). In vegetative tissues, their concentration is always higher in neutral lipids than in the phospho- and glycolipids (Schmid and Patterson, 1988), and they may be localized in intracellular droplets (Yatsu and Kircher, 1987). Thus, the only similarity between the presently identified wax cyclopropanes and the previously known carbocyclic fatty acids in plants is the position of the ring. This generalization includes the cyclopropane fatty acids with chain lengths of 20, 21, 22 and 24, ring carbons unknown, in either sulfolipids or phosphatidyl choline of a number of early spring plants as noted by Kuiper and Stuiver (1972).

The other lipid classes in cer-a6, cer-e8, cer-x60, cer-yc135 and cer-yl187

The other lipid classes were characterized as detailed previously (Lundqvist and von Wettstein-Knowles, 1983). Results pertinent to the present study follow. The mutant waxes from cer-a6, cer-e8, cer-x60, cer-yc135 and cer-yl187 have approximately 8, 75, 16, 73 and 24%, respectively, of the β-diketone lipids compared with the wild-type per spike. That all five mutants have a ‘−’ spike wax phenotype implies that, at least in cer-e8 and cer-yc135, the total amount of wax per spike is greatly reduced. The chain-length distributions in the mutants are not affected; for example, in cer-yl187, C31 comprises 94% of the β-diketone lipids. Too little sample was available to analyze the chain-length distributions in cer-a6, cer-x60 and cer-n53. The C32 aldehyde homologue, accounting for 55% of this lipid class in wild-type, amounted to approximately 5, 37, 26 and 50% for cer-a6, cer-e8, cer-x60 and cer-yc135, respectively (there was too little material to analyze for cer-yl187). The other acyl ELS derived lipid classes are characterized as in the wild-type by prominent amounts of the shorter homologues, namely C22–C26 in the free and ester primary alcohols and C20 in the free fatty acids. The C32 homologues of the free primary alcohols amount to only 2–10%, and are almost non-existent in the other two lipid classes. The free fatty acids exhibit a rather uniform distribution from C20–C32, accounting for only 6% in the wild-type and 1–4% in the mutant waxes.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Structural similarities imply a close relationship among the alkenes, cyclopropanes and internally branched methyl alkanes

To identify the hydrocarbons other than n-alkanes present in barley spike epicuticular waxes, the wild-type Bonus and nine cer mutants induced therein were grown under precisely controlled conditions for vegetative growth in the Stockholm Phytotron. The nine mutants were chosen because preliminary investigations revealed that reduction of the longest chain homologues of the n-alkanes was accompanied by an increased abundance of several other series of hydrocarbons. Of these, the monoenoic cis alkenes with chain lengths of C23–C35 and cyclopropanes with chain lengths of C27–C31 were the most prominent. The alkenes representing 0.3% of the hydrocarbons in Bonus were increased to as little as 1% in cer-yl187 but to as much as 25% in cer-n53, with the value of 8.7% in cer-x60 being typical. In the latter mutant, the cyclopropanes constituted 9.3%, compared with less than 1.5% in Bonus. By removing the alkenes and reducing the n-alkanes and cyclopropanes, several additional series of hydrocarbons were clearly revealed, namely 2-methyl branched alkanes with chain lengths of C24–C30, 3-methyl branched alkanes with chain lengths of C24 and C26, and internally branched methyl alkanes with chain lengths of C25–C33. The chain length of the most common homologue of the alkenes and cyclopropanes was 29, while the major double bond isomer was 9, and the major cyclopropane ring carbons were either 9,10 or 11,12. Although none of the internally branched methyl alkanes is prominent, the major isomer(s) of the shorter homologues had a methyl on carbon 9, and the longer homologues had a methyl on 11 or 13 and 13 or 15. These compositional similarities among the alkenes, cyclopropanes and internally branched methyl alkanes imply a close biosynthetic relationship among them.

Relevant examples of such relationships are present in the literature. For example, cyclopropane synthase (CPS) enzymes in Escherichia coli, Mycobacterium tuberculosis and Sterculia foetida generate fatty acids with cyclopropane rings from those with cis double bonds by transfer of a methyl group from S-adenosylmethionine (Bao et al., 2002; Grogan and Cronan, 1997; Yuan and Barry, 1996; Yuan et al., 1995). A similar CPS conceivably converts the barley cis double bonds to cyclopropane rings. Many years ago, a CPS enzyme was proposed to give rise to both cyclopropane and methyl branched hydrocarbons (Christie, 1970). Recently, five CPS enzymes have been characterized in M. tuberculosis, three acting in the synthesis of the methoxymycolates and two in the synthesis of the α-mycolates (Yuan and Barry, 1996; Yuan et al., 1995). However, none of the M. tuberculosis enzymes introduces a simple methyl group, although this does occur in synthesis of the 50:50 mixture of 7- and 8-methyl C17 alkanes characterizing the algae Nostoc muscorum and Anabaena variabilis (Fehler and Light, 1972; Han et al., 1969). Radiolabelled S-adenosylmethionine and C18:1Δ11 substrates were used in the experiments to demonstrate that synthesis of the internally branched alkanes occurred via methylene addition to the double bond followed by loss of the carboxy carbon. The mechanism by which the latter occurs is unknown. No cyclopropanes were detected in these algae. In contrast to the algal system, which very likely includes a CPS, a similar barley system would have, on the basis of the observed composition, a preference for addition of the methyl group to one side of the cis double bond, i.e. onto carbon 9 rather than 10.

A quite different mechanism giving rise to internally branched methyl alkanes is the use of methylmalonyl CoA instead of malonyl CoA as the extender unit in carbon chain construction (Howard and Blomquist, 2005). Insect alkanes may contain 5, 7, 9, 11, 13 or 15 methyl branches as well as dimethyl branches that are separated by 1, 3, 7, 9, 11 or 13 methylene groups, reflecting a programmed use of the two types of extender units (Howard and Blomquist, 2005). While it cannot be eliminated, this mechanism seems unlikely to play a role in synthesis of the barley internally branched methyl alkanes. It does not explain why the major isomer has a methyl group on carbon 9 in the C25, C27 and C29 homologues, but on 11 or 13 and 13 or 15 in the C31 and C33 homologues, respectively. Likewise the presence of the deduced 10, 12 and 14 methyl branched isomers cannot be accounted for.

How are the identified alkenes, cyclopropanes and internally branched alkanes related to the other wax lipids?

The identification of three structurally related wax lipids, presumably representing three products of an associated pathway, hereafter designated the enoic pathway to distinguish it from the reductive and decarboxylative/decarbonylative pathways, raises the question as to how their synthesis is correlated with that of the other wax components. From the discussion above, the critical question appears to be the origin of the alkenes. Most likely, an elongation system including a desaturase (DES) provides substrates for the enoic pathway. Several mechanisms are conceivable, differing in the time point at which the DES inserts the double bond into the acyl chain. Given the positions of the double bonds in the barley alkenes, insertion could occur (i) before elongation takes place, (ii) during elongation, (iii) after elongation is finished, or (iv) after the acyl chain has entered an associated decarboxylation/decarbonylation pathway. The latter is not tenable, however, as synthesis of the long carbon chains characterizing the n-alkanes is defective in all nine mutants, but that of the alkenes is not. One would also expect that the same chain length would be found for the major homologues of both wax lipids if they were derived from the same associated pathway, rather than C31 for the n-alkanes and C29 for the alkenes.

With regard to the first two possibilities, that the major positional isomer is always the 9-ene suggests participation of a DES with specificity for the 9–10 position, as is the case for introduction of the first double bond into the C18 products of the fatty acid synthase in plastids (Ohlrogge and Browse, 1995). In given seeds, some of the resulting C18:1 is then elongated to C20:1 and C22:1 for storage in triacylglycerols (Kunst et al., 1992). Such a scenario was proposed many years ago for the origin of the 9–10 double bond in the C23–C33 alkenes in Cistus petals (Gülz, 1980). More recently in Drosophila, two Δ9 desaturases, functioning in the synthesis of cuticular pheromones, have been cloned and characterized. In males, DESAT1 and 2 use C16 and C14 acyl chains as substrates, respectively, giving rise to C16:1 and C14:1 chains that are elongated and decarboxylated yielding alkenes with a double bond in position 5 or 7, respectively (Dallerac et al., 2000). As alkenes with the double bond in position 9 are also synthesized, a third Δ9 DES is proposed with specificity for C18 chains. In vivo 5,9 and 7,11 dienes are synthesized by females of some species, but were not when the appropriate DESAT 1 and 2 genes were expressed in yeast, implying the presence of yet another desaturase. If similar desaturases participate in synthesis of the barley wax alkenes, several possibilities may be envisaged to account for the set of isomers identified for each homologue (Table 2A): (i) a set of 18, 20, 22 and 24 chain-length desaturases specific for the 9–10 position similar to the Drosophila example described, or (ii) a set of desaturases with different regio-specificities for the site of double bond insertion (3–4, 5–6, 7–8 or 9–10) into a C18 acyl chain. Both types of desaturases have been characterized in plants, most recently one from English ivy that exhibits both chain-length and regio-specificities (Whittle et al., 2005). In none of the latter cases are the resulting unsaturated acyl chains subsequently extended to the lengths (21–35 carbons) characterizing the barley wax alkenes. With regard to the possibility that insertion could occur after elongation has finished, it is conceivable that a DES may insert a double bond into a completed, very long acyl chain, for example C30. This would require specificity for position 9–10 or 21–22, but I do not know of an example of such.

A quite different mechanism for inserting a double bond into an acyl chain is its introduction during construction of the chain. In each round of elongation, the action of the dehydrase moieties of ELS, FAS and some polyketide synthase complexes replaces the β-hydroxyl group with a transΔ2 double bond. Two mechanisms are known that retain the double bond giving an unsaturated acyl chain that is the substrate for a specific β-ketoacyl synthase: (i) either the enoyl reductase that removes the transΔ2 double bond is absent, or (ii) a second domain of the dehydrase has isomerase activity yielding a cisΔ3 acyl chain that cannot serve as substrate for the enoyl reductase. Both occur in the formation of the Rhizobium leguminosarum NOD fatty acid C18:4; the transΔ2, Δ4 and Δ6 bonds by the first mechanism, and the cisΔ11 bond by the latter (Geiger et al., 1994). Such a scenario cannot be ruled out at present for the barley alkenes, given the prominent role of the polyketide ELS in determining barley wax composition. Such an ELS, however, must have both β-ketoreductase and dehydrase activities, and the latter would have to form cis rather than trans double bonds in the C20 and C22 extension steps. In combination with the fact that β-diketone lipid synthesis is impaired in eight of the nine analyzed mutants, the observations suggest that a second polyketide ELS system would have to function in synthesis of the new hydrocarbons.

Wax elongation systems in barley

The investigated cer loci were all selected for a decrease in the blue–grey color that is indicative of the presence of the very long, thin β-diketone tubes on the surface of the lemmas. With the exception of cer-n26 wax, the mutant waxes are characterized by a relative and/or absolute reduction in the amount of β-diketone lipids per spike. The chain-length distributions are not altered, however, with the C31 homologue accounting for 94–97% of the β-diketones. Regardless of the extent of the reduction of the β-diketone lipids, the acyl elongation system leading to the C32 homologue precursors of the n-alkanes and aldehydes is impaired in all nine mutants. Thus, mutants at six different genes perturb both the acyl and β-ketoacyl elongation systems, while the elongation system giving rise to substrates for the enoic pathway-derived alkenes, cyclopropanes and internally branched methyl alkanes is not perturbed. The most obvious feature distinguishing the elongation system leading to the enoic pathway-derived lipids from the other two elongation systems is inclusion of a mechanism for introducing the double bond, very likely due to the presence of a DES as discussed above. This is contrary to the generally accepted notion that waxes arise only from saturated substrates while the aliphatic moieties of the closely related cutin and suberin monomers are synthesized from both saturated and unsaturated acyl chains. A recent estimate implicates participation of a DES in the synthesis of 30–35% of cutin and suberin monomers in Arabidopsis (Franke et al., 2005). For suberin, these included 2-hydroxyenoic acids with chain lengths up to C26 (Kurdyukov et al., 2006). Another example of elongation following desaturation is synthesis of the C22:1 and C24:1 anacardic acids in the glandular trichomes of geraniums (Schultz et al., 1996), while formation of the C20:1 fatty acid in meadowfoam seeds is accomplished by elongation preceding desaturation (Pollard and Stumpf, 1980). The coordinated action of a DES with ELS elongation systems is thus a well characterized phenomenon.

Synthesis of odd-chain-length enoic lipids from even-chain-length acyl substrates also requires a chain-shortening reaction as occurs in synthesis of the n-alkanes, secondary alcohols and β-diketone lipids, potentially by decarboxylation or decarbonylation. Results of early inhibitor studies revealed that different enzyme systems participate in formation of the shortened acyl chains used for n-alkane and β-diketone biosynthesis (reviewed in von Wettstein-Knowles, 1993). If one of the same enzyme systems functions in enoic lipid formation, then the origin of the dominant C29 homologues compared with the C31 homologues of the other lipids must lie in differences within the elongation systems, for example in the substrate specificities of the given ELS complexes. The chain-length difference is another feature peculiar to the enzyme systems giving rise to the enoic lipids. Two other observations are relevant. (i) Synthesis of the 13, 15 and 17 carbon esterified alkan-2-ols in barley has been shown to take place via decarboxylation giving a methylketone, followed by a reduction (Mikkelsen, 1984). (ii) In plants, the C18 carbocyclic fatty acid, sterculic acid, is often accompanied by major amounts of the C17 compound, malvalic acid, that is presumed to have originated by α-oxidation of sterculic acid (Bao et al., 2002). However, it appears unlikely that such reactions may also participate in synthesis of the barley wax enoic lipids.

As the spike is a complex organ, the apparent increased abundance of the newly identified hydrocarbons could reflect a decrease in n-alkane synthesis on one or more surfaces, accompanied by an unchanged synthesis of the new hydrocarbons on the other surfaces. The latter are unlikely to include the awns as alkenes were not detectable in wild-type awn wax. In addition, examination of the data from an early analysis of awn, spike minus awn and in which total spike wax from the mutant cer-u69, the abundance and chain-length distribution of the hydrocarbons are not affected, revealed that the new hydrocarbons must be less plentiful in awn than in total spike wax of this mutant (von Wettstein-Knowles and Netting, 1976). In this regard, it is interesting that 31 of the known 79 cer loci affect wax on more than one of the three cuticle surfaces defined by the phenotypic formula (Lundqvist and Lundqvist, 1988), including four of the six loci investigated. That the alkenes are often far more abundant in the mutant waxes relative to the cyclopropanes than they are in total spike wax of the wild-type does not support the proposal of differential changes on the various cuticle surfaces. Regardless of which spike cuticle surfaces are involved, however, the present data imply the existence of a third elongation system in barley that functions in synthesis of the enoic lipids, namely the alkenes, cyclopropanes and internally branched methyl alkanes. Future experiments are required to delineate the mechanism for insertion of the double bond, the nature of the ELS, and the biosynthetic relationship among the three enoic lipids, as well as the mechanism for chain shortening and its relationship to the decarbonylation/decarboxylation pathways acting on the acyl and β-ketoacyl ELS products.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plant material

Wild-type Hordeum vulgare cv. Bonus and its cer mutants cer-a6, cer-e8, cer-n26, cer-n53, cer-n624, cer-n985, cer-x60, cer-yc135 and cer-yl187 were planted at various times over a three-year period in the Phytotron (von Wettstein, 1967) at the Swedish University of Agriculture Science, Stockholm. Plants were grown in mineral wool from Grodan (Hedehusene, Denmark) under conditions giving optimum vegetative growth (Dormling et al., 1969, 1972), i.e. a thermoperiodicity of 16 h at 15°C and 8 h at 10°C under continuous light using Osram HQI-R 250 W/NDL lamps giving 250–300 μE sec−1 m−2 at soil level. Plants were watered twice daily with Hoagland's solution containing N2 increased to 100 mg l−1. At 8 weeks, the first three or four heading spikes from each of 14 plants were harvested, and the epicuticular wax collected by dipping for 1 × 15 sec plus 1 × 10 sec in chloroform. The extracts were filtered, combined and concentrated under N2. Leaving the plants under the optimum conditions for vegetative growth gave additional spikes every 8–9 days at approximately the same developmental stage of growth, yielding 12–16 spikes per plant. At this stage, a Bonus spike bears approximately 1 mg wax (von Wettstein-Knowles and Søgaard, 1980, unpublished).

The four cer-n mutants analyzed have the following wax phenotypes in the field and the Phytotron, respectively: cer-n20 (+/− + ++,+/++ + ++), cer-n53 (− + ++, − + ++), cer-n624 (−/+ + ++, − + ++), cer-n985 (− − ++, − − ++). The solidus (/) indicates year-to-year variability, with the most frequent phenotype given first. In the Phytotron, the β-diketone lipids in the spike wax varied from wild-type amounts in cer-n26 to 17–24% of that in wild-type in the other three mutants. The C31n-alkane, accounting for approximately 71% of the hydrocarbons in wild-type, ranged from 8–14% in the four mutants. Only in cer-n985 was the C26 primary alcohol somewhat reduced (Lundqvist and von Wettstein-Knowles, 1983).


Standards used for establishing the technique that was utilized to determine the positions of the double bonds and configurations of the latter in the barley alkenes included: C18:1 methyl ester synthesized from oleic acid (Sigma-Aldrich, Broendby, Denmark) using 5% H2SO4 in methanol and incubating for 2 h at 85°C, and the alkenes 1-C22 obtained from B. Stoianova-Ivanova, (Z)-9-C23 (Sigma-Aldrich) from G.J. Blomquist, a technical-grade tricosene (9-C23:1) containing approximately 65% of the cis and 15% of the trans isomer from E. Casagrande, and a mixture of cis olefins from beeswax (Streibl et al., 1966) from K. Stránský. The latter were predominantly C31, C33 and C35 with a cis double bond at C10. The CP1 bacterial fatty acid qualitative methyl ester mix from Supelco (Sigma-Aldrich) was subjected to GC on the non-polar SP-2100 and HP-1 and polar HP-17 columns specified below. This revealed that the cyclopropane methyl esters of cis-9,10-methylenehexadecanote and cis-9,10-methyleneoctadecanote had ECLs of 16.81 and 18.81 on the SP-2100 column and 17.12 and 19.12 on the HP-17 column.

Hydrocarbons were separated from the other epicuticular wax lipids by preparative TLC utilizing 20 × 20 cm plates coated with silica gel H type 60 (Merck, KGaA, Darmstadt, Germany), developing with hexane and eluting with amylene-stabilized chloroform (Merck). Preparative 5 × 20 cm AgNO3 TLC plates prepared from a slurry of silica gel GF254 type 60 (Merck), AgNO3 and H2O (45:7.5:100 w/w/v) were developed with hexane to separate alkenes (Rf = 0.08) from the saturated hydrocarbons (Rf = 0.75). Both were recovered using chloroform. The saturated hydrocarbons were transferred into isooctane, and the n-alkane components depleted by addition of several pieces of molecular sieve 5A 1/16-inch pellets (Mikrolaboratoriet, Aarhus, Denmark). During a week at room temperature, the molecular sieve pieces were replaced with new ones. The latter were then removed, the isooctane evaporated under N2, and the remaining hydrocarbons taken up in hexane. The configuration of the double bond in the alkenes was determined using hybrid 5 × 20 cm TLC plates with the bottom 5.5 cm coated with silica gel (GF254) and the upper 14.5 cm with AgNO3–silica gel prepared as described above, and developing three times with hexane to within 1–2 cm of the top [modified from the method described by Morris (1966)].

Chemical reactions

Alkenes were dried under N2, and hydrogenated based on the procedure described by Appelqvist (1972). The tube containing the alkenes was sealed with a cap, flushed with H2, and Adam's catalyst (1 mg/500 μl methanol) injected. After shaking for 30 min at room temperature, the sample was extracted with hexane. Completeness of the reaction was ascertained by TLC. Aliquots of the saturated hydrocarbons were subjected to acidic hydrogenation by the method described for mild hydrogenation except that glacial acetic acid replaced the methanol, and the reaction was allowed to proceed for 2 h. Other aliquots of the saturated hydrocarbons were exposed to bromine in diethyl ether (1:5 v/v) for 1 h at room temperature (modified from the method described by Brian and Gardner, 1968). Then the excess bromine and diethyl ether were evaporated at 50°C under N2 and the resulting hydrocarbons taken up in hexane.

To identify the positions of the double bonds in the alkenes, the procedure described for oxidizing monounsaturated fatty acids to α-diols (Capella and Zorzut, 1968) was modified for use with the very long-chain barley epicuticular alkenes. Alkenes were oxidized by adding 2 μg OsO4 in 200 μl dioxane and allowing the reaction mixture to stand for at least 30 min before adding 3 ml of 16% Na2SO3 in water (w/v) plus 3.3 volumes of methanol. After shaking the reaction mix for 1 h, the diols were recovered by adding 2 ml of hexane and centrifuging briefly at 5000 g. The hexane layer was recovered and washed with water. Conversion of standards to α-diols was monitored by silica gel H TLC using 3% ethanol in chloroform: standards had Rf values of 0.71–0.79 and their respective α-diols had Rf values of 0.02–0.23. TMS derivatives of the dried α-diols in capped vials were prepared by adding 100 μl pyridine, 100 μl N,N-bis-trimethylsilyltrifluoroacetamide (BSTFA) and 8 μl TMS through the cap and incubating at 60°C for 20 min. The solution was concentrated using dried N2 and taken up in hexane. Table 4 shows the percentage recovery of the four standard substrates with and without the inclusion of the hexane extraction step. Excellent recoveries of the TMS derivatives of the α-diols of C18 methyl ester and the C22 and C23 alkenes were obtained, while those of the longer C31 to C35 alkenes were less efficient. The results of time-course experiments revealed that yields of the latter were not increased by extending the oxidation step beyond an hour.

Table 4.   Effect of substrate and hexane extraction on the recovery of α-diols after a 1 h OsO4 oxidation
Substrate(μg)% recovery
+ hex− hex
Methyl ester
 C22:12096 37
100107 12
200103  7
 C23:12094 11
8298 24
16389  8
 C33:17982  3

The GC analyses were performed using (i) a Hewlett-Packard 5840 instument (Agilent Technologies, Naerum, Denmark) with a 3% SP-2100 (152 × 3.2 mm) on 80/100 mesh Supelcoport (Sigma-Aldrich) column, or a Varian Vista 6000 instrument (agilent Technologies) with either (ii) an HP-17 column (10 m × 0.53 mm × 0.2 μm, cross-linked 50% phenylmethyl silicone) (Agilent Technologies) or (iii) an HP-1 column (10 m × 0.53 mm × 0.2 μm, 100% dimethyl silicone) (Agilent Technologies). Operating conditions were optimized to give maximum sensitivity. Equivalent chain lengths (ECLs) were determined using (i) the 3% SP-2100 column and programming from 140°C at 3°C min−1 to 310°C, (ii) the HP-1 column for 5 min at 60°C followed by programming at 5°C min−1 to 300°C, and (iii) the HP-17 column for 5 min at 90°C followed by programming at 5°C min−1 to 290°C. The C29 alkene as a weight percentage of the total hydrocarbons was estimated from GC traces of the latter taken after formation of the diol-TMS derivates as this homologue was reasonably well resolved from other components in this wax fraction as illustrated in Figure 2(b). Values were corrected for the weight difference between alkenes and their respective diol-TMS derivatives. GC traces of the saturated hydrocarbons recovered from AgNO3 TLC plates, as illustrated in Figure 6(a), were used to estimate the weight percentage of the cyclopropanes and branched alkanes in this wax fraction. These values were converted to weight percentage of total hydrocarbons using the already determined weight percentages of the alkenes in the total hydrocarbons.

GC-MS analyses of the alkene derivatives and branched hydrocarbons were carried out on Varian GC 7070F instrument equipped with a VG 2035 Data System and a Pye Unicam Series 204 GC (Cambridge, UK) with a 2 mm (internal diameter) × 1.5 m glass column packed with 5% OV-101 on Chromosorb W-HP, 80–100 mesh (Agilent Technologies); carrier gas (He), approximately 25 ml min−1; ion source temperature 220°C; ionization energy 70 eV. Analyses of the molecular sieve-treated hydrocarbons, cyclopropane derivatives and the internally branched saturated hydrocarbons were accomplished using a Varian GC (Mass lab) Trio-2 mass spectrometer with direct inlet, ionization energy 70 eV, using a 21 m × 0.31 mm × 0.17 μm methylsilicone HP column (Agilent Technologies).


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

I am indebted to J. Øgaard Madsen (Organic Chemistry Department, Technical University of Denmark) for obtaining and interpreting the mass spectra, to M. Petersen and M. Mortensen for outstanding technical assistance, and to G.J. Blomquist (University of Nevada, Reno, USA), B. Stoianova-Ivanova (University of Sofia, Bulgaria), K. Stránský (Academy of Sciences, Prague, Czech Republic) and E. Casagrande (AgriSense-BCS Ltd, Pontypridd, UK) for the standards. Support from the Danish Natural Sciences Research Council (grant number 272–05–0401) is gratefully acknowledged.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Appelqvist, L. (1972) A simple and convenient procedure for the hydrogenation of lipids on the micro- and nanomole scale. J. Lipid Res. 13, 146148.
  • Bao, X., Katz, S., Pollard, M. and Ohlrogge, J. (2002) Carbocyclic fatty acids in plants: Biochemical and molecular genetic characterization of cyclopropane fatty acid synthesis of Sterculia foetida. Proc. Natl Acad. Sci. USA, 99, 71727177.
  • Bianchi, G., Avato, P. and Salamini, F. (1984) Surface waxes from grain, leaves and husks of maize (Zea mays L.). Cereal Chem. 61, 4547.
  • Blomquist, G.J., Nelson, D.R. and de Renobales, M. (1987) Chemistry, biochemistry, and physiology of insect cuticular lipids. Arch. Insect Biochem. Physiol. 6, 227265.
  • Brian, B.L. and Gardner, E.W. (1968) A simple procedure for detecting the presence of cyclopropane fatty acids in bacterial lipids. Appl. Microbiol. 16, 549552.
  • Capella, P. and Zorzut, C.M. (1968) Determination of double bond position in monounsaturated fatty acid esters by mass spectrometry of their trimethylsilyloxy derivatives. Anal. Chem. 40, 14581463.
  • Christie, W.W. (1970) Cyclopropane and cyclopropene fatty acids. In Topics in Lipid Chemistry (Gunstone, F.D., ed.). London: Logos Press, pp. 149.
  • Christie, W.W. and Holman, R.T. (1966) Mass spectrometry of lipids. I. Cyclopropane fatty acid esters. Lipids, 1, 176182.
  • Dallerac, R., Labeur, C., Jallon, J.-M., Knippie, D.C., Roelofs, W.L. and Wicker-Thomas, C. (2000) A Δ9 desaturase gene with a different substrate specificity is responsible for the cuticular diene hydrocarbon polymorphism in Drosophila melanogaster. Proc. Natl Acad. Sci. USA, 97, 94499454.
  • Dietrich, C.R., Perera, M.A., Yandeau-Nelson, M.D., Meeley, R.B., Nikolau, B.J. and Schnable, P.S. (2005) Characterization of two GL8 paralogs reveals that the 3-ketoacyl reductase component of fatty acid elongase is essential for maize (Zea mays L.) development. Plant J. 42, 844861.
  • Dormling, I., Gustafsson, Å. and von Wettstein, D. (1969) Phytotron cultivation of Bonus barley. The control of maturation and grain quality. Hereditas, 63, 415428.
  • Dormling, I., Gustafsson, Å., Jung, H.R. and von Wettstein, D. (1972) Phytotron cultivation of Svalof's Bonus barley and its mutant Svalof's Mari. Hereditas, 56, 221237.
  • Fehler, S.W.G. and Light, R.J. (1972) Biosynthesis of methylheptadecanes in Anabaena variabilis. In vitro incorporation of S-[methyl-14C]adenosylmethionine. Biochemistry, 11, 24112416.
  • Fiebig, A., Mayfield, J.A., Miley, N.L., Chau, S., Fischer, R.I. and Preuss, D. (2000) Alterations in CER6, a gene identical to CUT1, differentially affect long-chain lipid content on the surface of pollen and stems. Plant Cell, 12, 20012008.
  • Franke, R., Briesen, I., Wojciechowski, T., Faust, A., Yephremov, A., Nawrath, C. and Schreiber, L. (2005) Apoplastic polyesters in Arabidopsis surface tissues – A typical suberin and a particular cutin. Phytochemistry, 66, 26432658.
  • Geiger, O., Thomas-Oates, J.E., Glushka, J., Spaink, H.P. and Lugtenberg, B.J. (1994) Phospholipids of Rhizobium contain nodE-determined highly unsaturated fatty acid moieties. J. Biol. Chem. 269, 1109011097.
  • Grogan, D.W. and Cronan, J.E. (1997) Cyclopropane ring formation in membrane lipids of bacteria. Microbiol. Mol. Biol. Rev. 61, 429441.
  • Gülz, P.-G. (1980) Alkanes and alkenes in the epicuticular waxes from Cistus plants. In Biogenesis and Function of Plant Lipids (Mazliak, P., Benveniste, P., Costes, C. and Douce, R., eds). Amsterdam: Elsevier/North Holland, pp. 275279.
  • Gülz, P.-G., Amigoni, E., Schmitz, B. and Egge, H. (1979) Untersuchungen über die Kutikularwachse in der Gattung Cistus L. (Cistaceae). VI. Die Zusammensetzung der Alkene aus Cistus-Petalen und die Lage der Doppelbindung in diesen Alkenen am Beispiel von 3 Cistus-Arten. Z. Pflanzenphysiol. 94, 3543.
  • Hamilton, R.J. (1995) Analysis of waxes. In Waxes, Chemistry, Molecular Biology and Functions (Hamilton, R.J., ed.). Dundee: The Oily Press, pp. 311342.
  • Han, J., Chan, H.W.S., and Calvin, M. (1969) Biosynthesis of alkanes in Nostoc muscorum. J. Am. Chem. Soc. 91, 51565159.
  • Howard, R.W. and Blomquist, G.J. (2005) Ecological, behavioral, and biochemical aspects of insect hydrocarbons. Annu. Rev. Entomol. 50, 371393.
  • Kuiper, P.J. and Stuiver, B. (1972) Cyclopropane fatty acids in relation to earliness in spring and drought tolerance in plants. Plant Physiol. 49, 307309.
  • Kunst, L. and Samuels, A.L. (2003) Biosynthesis and secretion of plant cuticular wax. Prog. Lipid Res. 42, 5180.
  • Kunst, l., Taylor, D.C. and Underhill, E.W. (1992) Fatty acid elongation in developing seeds of Arabidopsis thaliana. Plant Physiol. Biochem. 30, 425434.
  • Kurdyukov, S., Faust, A., Nawrath, C. et al. (2006) The epidermis-specific extracellular BODYGUARD controls cuticle development and morphogenesis in Arabidopsis. Plant Cell, 18, 321339.
  • Lardizabal, K., Metz, J.G., Sakamoto, T., Hutton, W.C., Pollard, M.R. and Lassner, M.W. (2000) Purification of a jojoba embryo wax synthase, cloning of its cDNA, and production of high levels of wax in seeds of transgenic Arabidopsis. Plant Physiol. 122, 645655.
  • Lundqvist, U. and Lundqvist, A. (1988) Mutagen specificity in barley for 1580 eceriferum mutants localized to 79 loci. Hereditas, 108, 112.
  • Lundqvist, U. and von Wettstein-Knowles, P. (1983) Phenotypic diversity of barley spike waxes resulting from mutations at locus cer-n. Carlsberg Res. Commun. 48, 321344.
  • McCloskey, J.A. and Law, J.H. (1967) Ring location in cyclopropane fatty acid esters by a mass spectrometric method. Lipids, 2, 225230.
  • Metz, J.G., Pollard, M.R., Andersen, L., Hayes, T.R. and Lassner, M.W. (2001) Purification of a jojoba embryo fatty acyl-coenzyme A reductase and expression of its cDNA in high erucic acid rapeseed. Plant Physiol. 122, 635644.
  • Mikkelsen, J.D. (1978) The effects of inhibitors on the biosynthesis of the long chain lipids with even carbon numbers in barley spike epicuticular wax. Carlsberg Res. Commun. 43, 1535.
  • Mikkelsen, J.D. (1984) Biosynthesis of esterified alkan-2-ols and β-diketones in barley spike epicuticular wax: synthesis of radioactive intermediates. Carlsberg Res. Commun. 49, 391416.
  • Morris, L.J. (1966) Separation of lipids by silver ion chromatography. J. Lipid Res. 7, 717732.
  • Nelson, D.R., Howard, R.W., McDaniel, C.A. and Blomquist, G.J. (1980) Re-analysis of the cuticular methylalkanes of Solenopsis invicta and S. richteri. Insect Biochem. 10, 409418.
  • Ohlrogge, J. and Browse, J. (1995) Lipid biosynthesis. Plant Cell, 7, 957970.
  • Pighin, J.A., Zheng, H., Balakshin, L.J., Goodwin, I.P., Western, T.L., Jetter, R., Kunst, L. and Samuels, A.L. (2004) Plant cuticular lipid export requires an ABC transporter. Science, 306, 702704.
  • Pollard, M. and Stumpf, P.K. (1980) Biosynthesis of C20 and C22 fatty-acids by developing seeds of Limnanthes alba. Chain elongation and Δ5 desaturation. Plant Physiol. 66, 649655.
  • Schmid, K.M. and Patterson, G.W. (1988) Distribution of cyclopropenoid fatty acids in malvaceous plant parts. Phytochemistry, 27, 28312834.
  • Schultz, D.J., Cahoon, E.C., Shanklin, J., Craig, R., Cox-Foster, D.L., Mumma, R.O. and Medford, J.I. (1996) Expression of a Δ9 14:0-acyl carrier protein fatty acid desaturase gene is necessary for the production of ω5 anacardic acids found in pest-resistant geranium (Pelargonium xhortorum). Proc. Natl Acad. Sci. USA, 93, 87718775.
  • Shenstone, F.S., Vickery, J.R., and Johnson, A.R. (1965) Studies on the chemistry and biological effects of cyclopropenoid compounds. J. Agric. Food Chem. 13, 410414.
  • Simpson, D. and von Wettstein-Knowles, P. (1980) Structure of epicuticular waxes on spikes and leaf sheaths of barley as revealed by a direct platinum replica technique. Carlsberg Res. Commun. 45, 465481.
  • Stoianova-Ivanova, B., Mladenova, K. and Malova, I. (1971) Long chain conjugated alkadienes; a new component of plant waxes. Phytochemistry, 10, 25252528.
  • Stránský, K. and Streibl, M. (1969) On natural waxes. XII. Composition of hydrocarbons in morphologically different plant parts. Collect. Czech. Chem. Commun. 34, 103117.
  • Stránský, K., Streibl, M. and Kubelka, V. (1970) On natural waxes. XV. Hydrocarbon constitutents of the leaf wax from walnut tree (Juglans regia L.). Collect. Czech. Chem. Commun. 35, 882891.
  • Stránský, K., Streibl, M., Ubik, K., Kohoutová, J. and Novotný, L. (1981) Die Zusammensetzung der Kohlenwasserstoffe der Kamille (Matricaria chamomilla L.). Fette Seifen Anstrichm. 83, 347354.
  • Streibl, M., Stránský, K. and Sorm, F. (1966) Über einige neue Kohlenwasserstoffe im Wachs der Honigbiene (Apis mellifica L.). Fette Seifen Anstrichm. 68, 799805.
  • Todd, J., Post-Beittenmiller, D. and Jaworski, J.G. (1999) KCS1 encodes a fatty acid elongase 3-ketoacyl-CoA synthase affecting wax biosynthesis in Arabidopsis thaliana. Plant J. 17, 119130.
  • von Wettstein, D. (1967) The phytotron in Stockholm. Studia Forest. Suecica, 44, 123.
  • von Wettstein-Knowles, P. (1993) Waxes, cutin, and suberin. In Lipid Metabolism in Plants (Moore, T.S. ed.). London: CRC Press, pp. 127166.
  • von Wettstein-Knowles, P. (1995) Biosynthesis and genetics of waxes. In Waxes, Chemistry, Molecular Biology and Functions (Hamilton, R.J., ed.). Dundee: The Oily Press, pp. 91129.
  • von Wettstein-Knowles, P. and Netting, A.G. (1976) Composition of epicuticular waxes on barley spikes. Carlsberg Res. Commun. 41, 225235.
  • von Wettstein-Knowles, P. and Søgaard, B. (1980) The cer-cqu region in barley: gene cluster or multifunctional gene. Carlsberg Res. Commun. 45, 125141.
  • Whittle, E., Cahoon, E.B., Subrahmanyam, S. and Shanklin, J. (2005) A multifunctional acyl-acyl carrier protein desaturase from Hedera helix L. (English ivy) can synthesize 16- and 18-carbon monoene and diene products. J. Biol. Chem. 280, 2816928176.
  • Wollrab, V. (1968) Über Naturwachse VIII. Olefine und Paraffine aus den Wachsen einiger Pflanzen der Familie Rosaceae. Collect. Czech. Chem. Commun. 33, 15841600.
  • Xu, X., Dietrich, C.R., Lessire, R., Nikolau, B.J., and Schnable, P.S. (2002) The endoplasmic reticulum-associated maize GL8 protein is a component of the acyl-coenzyme A elongase involved in the production of cuticular waxes. Plant Physiol. 128, 924934.
  • Yatsu, L.Y. and Kircher, H.W. (1987) Electron microscopic cytochemistry of cyclopropenoids in cottonseed (Gossypium hirsutum L.) tissues. Lipids, 22, 397400.
  • Yuan, Y. and Barry, C.E. (1996) A common mechanism for the biosynthesis of methoxy and cyclopropyl mycolic acids in Mycobacterium tuberculosis. Proc. Natl Acad. Sci. USA, 93, 1282812833.
  • Yuan, Y., Lee, R.E., Besra, G.S., Belisle, J.T. and Barry, C.E. (1995) Identification of a gene involved in the biosynthesis of cyclopropanated mycolic acids in Mycobacterium tuberculosis. Proc. Natl Acad. Sci. USA, 92, 66306634.
  • Zheng, H.Q., Rowland, O., and Kunst, L. (2005) Disruptions of the Arabidopsis enoyl-CoA reductase gene reveal an essential role for very-long-chain fatty acid synthesis in cell expansion during plant morphogenesis. Plant Cell, 17, 14671481.