The Inhibitor of wax 1 locus (Iw1) prevents formation of β- and OH-β-diketones in wheat cuticular waxes and maps to a sub-cM interval on chromosome arm 2BS

Authors


For correspondence (e-mail cristobal.uauy@jic.ac.uk).

Summary

Glaucousness is described as the scattering effect of visible light from wax deposited on the cuticle of plant aerial organs. In wheat, two dominant genes lead to non-glaucous phenotypes: Inhibitor of wax 1 (Iw1) and Iw2. The molecular mechanisms and the exact extent (beyond visual assessment) by which these genes affect the composition and quantity of cuticular wax is unclear. To describe the Iw1 locus we used a genetic approach with detailed biochemical characterization of wax compounds. Using synteny and a large number of F2 gametes, Iw1 was fine-mapped to a sub-cM genetic interval on wheat chromosome arm 2BS, which includes a single collinear gene from the corresponding Brachypodium and rice physical maps. The major components of flag leaf and peduncle cuticular waxes included primary alcohols, β-diketones and n-alkanes. Small amounts of C19–C27 alkyl and methylalkylresorcinols that have not previously been described in wheat waxes were identified. Using six pairs of BC2F3 near-isogenic lines, we show that Iw1 inhibits the formation of β- and hydroxy-β-diketones in the peduncle and flag leaf blade cuticles. This inhibitory effect is independent of genetic background or tissue, and is accompanied by minor but consistent increases in n-alkanes and C24 primary alcohols. No differences were found in cuticle thickness and carbon isotope discrimination in near-isogenic lines differing at Iw1.

Introduction

Wild emmer (Triticum turgidum ssp. dicoccoides), a tetraploid grass species, is the ancestor of both modern pasta wheat (Triticum turgidum ssp. durum) and bread wheat (Triticum aestivum) (Dubcovsky and Dvorak, 2007). Natural emmer populations occur in the Fertile Crescent across a wide range of habitats and environments. This adaptability partly results from genetic variability for morphological traits, including glaucousness, the scattering effect of visible light from wax deposited on cuticles of aerial organs. In addition to the effect on light reflectance (Reicosky and Hanover, 1978; Feldhake, 1990), cuticular waxes also determine many physical properties of the plant, such as water relations (Febrero et al., 1998), canopy temperature, and interactions with fungal pathogens and insects (Ringelmann et al., 2009). Although almost all modern wheat cultivars are glaucous, wild emmer accessions display a wide range of visual wax phenotypes across individuals, as well as between tissues within a single plant.

Wheat genetic studies have revealed two genes for glaucousness (W1 and W2) and two glaucous suppressors (Iw1 and Iw2) located on chromosome arms 2BS (W1 and Iw1) and 2DS (W2 and Iw2) (Tsunewaki and Ebana, 1999; Tsujimoto, 2001; Liu et al., 2007). The presence of either Iw1 or Iw2 is sufficient to inhibit W1 and/or W2 (Tsunewaki and Ebana, 1999). Although the original nomenclature Iw refers to ‘Inhibitor of wax’, their true effect is inhibition of glaucousness. Several quantitative trait loci have also been identified that act specifically on certain plant organs (Dubcovsky et al., 1997; Peng et al., 2000; Börner et al., 2002; Kulwal et al., 2003; Mason et al., 2010; Bennett et al., 1990). Despite these studies, none of these genes have been characterized at the molecular level, and the exact extent (beyond visual examination) by which they affect the composition and quantity of cuticular wax is also unclear. Moreover, the visual phenotype of a plant surface is often not correlated with the wax load. For example, internodes of cer-c36 barley (Hordeum vulgare) with 15–19 μg amorphous wax cm−2 are glossy, whereas its leaves with the same amount of dense crystalline wax are glaucous (von Wettstein-Knowles, 1969). In Arabidopsis thaliana, the increased wax load on WIN1 over-expression lines results in glossier leaves (Broun et al., 2004), and inconsistent correlations between glaucousness and cuticular wax load have been described in wheat (Johnson et al., 1983).

Aliphatics (hydrocarbons, esters, aldehydes and alcohols) and triterpenoids, flavonoids and phenolic lipids are common components of plant waxes. In addition, some genera such as the Triticeae contain dominating amounts of aliphatic β-diketones and related compounds. The 20-34 carbon chains characterizing these wax compounds are synthesized by the sequential action of a fatty acid synthase complex producing C16 and C18 chains that are further extended by an elongase complex of which there are two types. The first are fatty acid elongases that consist of a condensing enzyme plus three additional tailoring enzymes that prepare the growing acyl chain for the next round of elongation. The second are type III polyketide synthases (KCS), which lack the tailoring enzymes, with the result that oxygens are introduced into the acyl chain during elongation. While much knowledge has accrued in the past decade about wax biosynthetic pathways in Arabidopsis involving fatty acid elongases (Samuels et al., 2008), only early genetic/biochemical studies in barley established that an additional elongase, pkKCS, synthesizes the β-diketone aliphatics and determined the roles of the multi-functional gene cer-cqu therein (von Wettstein-Knowles, 2012).

Here a genetic approach with detailed biochemical characterization of wax compounds was used to characterize the Iw1 locus. Using synteny and a large number of F2 gametes, Iw1 was fine-mapped to a sub-cM genetic interval on wheat chromosome arm 2BS that includes a single collinear gene from the corresponding Brachypodium and rice physical intervals. The major components of flag leaf and peduncle cuticular waxes were primary alcohols (POHs), β-diketones and n-alkanes, accompanied by small amounts of aldehydes and free fatty acids plus the phenolic lipids, alkyl and methylalkylresorcinols (ARs/MARs). Iw1 inhibits formation of β- and hydroxy-β-diketones in the specified cuticles, accompanied by minor increases in n-alkanes and C24POH. The inhibitory effect is independent of genetic background and across tissues. No differences were found in cuticle thickness and carbon isotope discrimination in BC2F3 near-isogenic lines (NILs) that differ with regard to the presence of the Iw1 dominant allele.

Results

The non-glaucous phenotype in Shamrock is determined by a dominant allele of Iw1

A doubled-haploid (DH) population developed from two winter wheat varieties, Shango (glaucous, Figure 1a) and Shamrock (non-glaucous, Figure 1b), was used to map the non-glaucous phenotype to a locus on chromosome arm 2BS (Simmonds et al., 2008). The locus was originally named Viridescence (Vir) because of the bright green colour of the non-glaucous organs (Figure 1b,c). Interestingly, Jensen and Driscoll (1962) and Liu et al. (2007) also mapped a dominant inhibitor of glaucousness (Iw1) with the same phenotype to this location. As both Iw1 and Shamrock 2BS are derived from wild emmer, we hypothesized that Vir and Iw1 are the same gene.

Figure 1.

Visual differences arising from the presence of different alleles of the Iw1 gene in single tillers (a, b) and canopies (c). Shango iw1/iw1 (a), Shamrock Iw1/Iw1 (b) and Xi19 BC2F2 NIL pairs (c) in the field. In (a), peduncle (p) refers to the visible uppermost internode; the flag leaf blade and sheath are also indicated.

To test this, Shamrock was crossed to six additional glaucous hexaploid UK wheat varieties (Alchemy, Einstein, Hereward, Malacca, Robigus and Xi19) with different genetic backgrounds to that of Shango. The resulting F1 plants were non-glaucous, suggesting a dominant gene action consistent with Iw1. This was further supported by the 1:3 glaucous to non-glaucous segregation in the F2 populations (Data S1). In all six crosses, the non-glaucous phenotype mapped to chromosome arm 2BS as in the Shango cross. These results strongly support our hypothesis that Iw1 determines the non-glaucous phenotype of Shamrock.

Iw1 maps to a sub-cM interval on wheat chromosome arm 2BS

The non-glaucous phenotype had been previously mapped to an approximately 3 cM interval between SSR marker Xgwm614 and DArT marker wPt-4453 (Simmonds et al., 2008). To fine map Iw1, markers were developed from ten ESTs previously mapped to the distal deletion bin of 2BS (2BS3-0.84–1.00) (Conley et al., 2004), which has a clear syntenic relationship to the sequenced Brachypodium and rice genomes (Table S1). Using the original DH population (87 lines), Iw1 mapped between markers JIC007 and JIC012, corresponding to Brachypodium genes Bradi5g01220 and Bradi5g01130 (Figure 2 and Table S2).

Figure 2.

Shango × Shamrock genetic map of Iw1, and syntenic relationship with the Brachypodium and rice (Oryza sativa) physical maps. Syntenic markers are joined by lines. Red bars indicate the Iw1 interval in wheat and syntenic intervals in rice and Brachypodium.

An additional four markers were developed for genes between JIC007 and JIC012. Three of these markers (JIC009, JIC010 and JIC011; corresponding to Bradi5g01180, Os04g05030 and Bradi5g01160, respectively, Figure 2) were completely linked to Iw1, whereas JIC015 (Bradi5g01190) mapped to 2BS, but proximal to Iw1. This genetic position is consistent with the fact that Bradi5g01190 (and Bradi5g01200) are >95% similar to Bradi5g02390 (and Bradi5g02380), suggesting a very recent duplication of these genes into the Iw1 syntenic interval. The duplication appears to be limited to these two genes because the flanking genes, Bradi5g02400 and Bradi5g02370, are single-copy genes. The complete linkage of the JIC009–11 markers was consistent across the six F2 mapping populations generated between Shamrock and the glaucous UK varieties.

To further delimit Iw1 and increase the mapping resolution, 2111 F2 plants from the Shango × Shamrock cross were screened using EST-derived SNP markers JIC004 and JIC015 (Bradi5g01410 and Bradi5g01190). A total of 297 recombination events were identified between these markers (7.03 cM), of which 36 mapped between markers JIC007 and JIC012 (0.85 cM, Figure 2). Phenotypic evaluation of these lines and their F3 progeny determined that Iw1 remained completely linked to JIC009, JIC010 and JIC011 within this 0.85 cM interval. Twenty-nine recombination events were mapped between JIC007 and Iw1 (0.69 cM), while seven were proximal, between Iw1 and JIC012 (0.17 cM).

To break the linkage between Iw1 and the three markers (JIC009–11), 850 F2 plants from a cross between tetraploid durum wheat Langdon (iw1/iw1) and TTD140, a wild emmer accession previously shown to carry Iw1 (Rong et al., 2000), were screened using JIC007–JIC012. This identified recombinants across the Iw1 interval, with the marker order remaining consistent with the Shango × Shamrock population. The phenotypes of these lines and their F3 progeny remained completely linked to JIC009. However, three independent recombination events, between Iw1 and JIC010/JIC11 (0.18 cM) (Data S2), broke the linkage seen in the Shango × Shamrock population. An additional 25 recombinants between JIC007 and Iw1 (1.47 cM) positioned Iw1 within a 1.65 cM interval between JIC07 and JIC010/JIC011. The syntenic intervals in rice and Brachypodium include only one gene with the corresponding wheat sequence having evidence of 2BS localization, Bradi5g01180 (Os04g05010), which maps as JIC009 within the Iw1 interval.

Identification of additional candidate genes from barley genomic contigs

We screened the recently released barley genome (International Barley Genome Sequencing Consortium, 2012) using the Bradi5g01180 sequence, and identified a contig anchored at 4.85 cM, together with a series of other short contigs at the same position. This contig cluster includes three annotated genes in addition to the barley Bradi5g01180 homologue (MLOC_77461). Two of them, MLOC_20994 and MLOC_6767, are non-syntenic to rice and Brachypodium, whereas the third (MLOC_59629) is homologous to Bradi5g01130, which we mapped to wheat 2BS as JIC012 (proximal to Iw1). The two non-syntenic barley genes have significant BLASTN hits to the wheat group 2 chromosome arm assemblies, suggesting potential localization within the corresponding wheat Iw1 interval.

Iw1 affects epicuticular waxes, without altering the cuticle membrane

In numerous species, mutants affect epicuticular wax deposition and/or alter the cuticle membrane. Cryo-scanning electron microscopy was used to examine tissue surfaces of Shango and Shamrock plants to determine whether the surface structure correlated with their visible phenotypes (Figure 1a,b). The peduncles of Iw1 plants were completely devoid of any visible wax protruding from the surface, whereas plants lacking Iw1 showed a dense accumulation of tubular/rod-shaped wax structures across the entire surface (Figure 3a,b). This was also true for the abaxial side of the flag leaves, where plants differing at the Iw1 locus had the same contrasting phenotypes as seen on peduncles (Figure 3c,d). On the adaxial side of flag leaves, both plants with and without Iw1 had platelet-shaped wax deposited on the surface. However, plants lacking Iw1 also had the tubular/rod-shaped wax structures seen on the peduncle and abaxial surface (Figure 3e,f), suggesting that the adaxial cuticle surface is more complex than the abaxial leaf and peduncle surfaces. These major variations in epicuticular wax density and structure were consistent across Iw1 BC2F3 NILs derived from glaucous UK varieties (Figure S1). These results are in accordance with observations on related wheat lines (Netting and von Wettstein-Knowles, 1973), except that the presence of tubes occurred only at the base of the adaxial flag leaf blades.

Figure 3.

Iw1 affects epicuticular waxes, without altering the cuticle membrane.(a–f) Cryo-scanning electron micrographs of Shamrock (a, c, e) and Shango (b, d, f) exposed peduncles (a, b), and abaxial (c, d) and adaxial flag leaf blade surfaces (e, f).
(g–j) Close-ups of transmission electron micrographs of adaxial flag leaf blades from Iw1/Iw1 (g) and iw1/iw1 (h) BC2F3 Xi19 NILs, and Shamrock (i) and Shango (j) peduncle cuticles. Full-sized images are available as Figure S4. cw, cell wall; c, cuticle. Scale bars = 10 μm (a–f) and 500 nm (g–j).

In contrast, no differences were found in the cuticle thickness of flag leaf blades between Xi19 NIL pairs differing at Iw1 when examined by transmission electron microscopy (Figure 3g,h). Xi19 Iw1/Iw1 lines had a mean cuticle thickness of 0.105 ± 0.012 (adaxial) and 0.134 ± 0.016 μm (abaxial), which was not significantly different (= 0.52) from the thickness of corresponding surfaces in Xi19 iw1/iw1 lines (0.098 ± 0.016 and 0.144 ± 0.012 μm, respectively). The peduncle cuticle thickness and morphology were likewise indistinguishable in Shamrock (0.239 ± 0.005 μm) and Shango (0.244 ± 0.024 μm) field-grown plants (= 0.84, Figure 3i,j). These data suggest that differences in glaucousness between lines with and without Iw1 result from changes in epicuticular wax deposition.

Primary alcohols, β-diketones and n-alkanes are major components of wheat cuticular waxes

Wax composition was analysed to determine the differences that give rise to the glaucous versus non-glaucous phenotype. Thin-layer chromatography (TLC) initially disclosed a major difference between Shango and Shamrock total flag leaf waxes (Figure 4), namely that the dominating β-diketones are absent in Shamrock. This correlates with the presence versus absence of tubular structures on the examined surfaces (Figure 3c,d). A second difference between the two varieties is the Shango band labeled unknown (6) in Figure 4. Not only does its Rf not correspond to that of the hydroxy-β-diketones in the Eucalyptus sp. standard, but diagnostic ions for these aliphatics (see below) were absent when its eluate was subjected to mass spectrometry.

Figure 4.

Wax lipids in flag leaf blades of Shango (iw1/iw1), Shamrock (Iw1/Iw1) and Eucalyptus sp. visualized after TLC in hexane:ether (9:1 v/v) using primuline. Numbered arrows correspond to n-alkanes (1), β-diketones (2), aldehydes (3), primary alcohols (4), hydroxy-β-diketones (5), unknown (6), and alkylresorcinols, methylalkylresorcinols and free fatty acids (7). All samples were run on the same TLC plate.

To quantify the total and relative amounts of the β-diketones and their hydroxy derivatives if present, the total mg wax mg−1 leaf−1 was determined from OD273 measurements plus wax and leaf fresh weights (Table 1). Shango flag leaves have more than twice as much wax as Shamrock leaves. In the former, approximately 67% of the wax were β-diketone aliphatics; in the latter, these aliphatics constituted no more than 8% of the total wax. In the two varieties, the hydroxy-β-diketones account for 0.5 and 2%, respectively, of the total β-diketones (<0.03 and 0.01 μg mg−1 leaf−1). Other components common to both genotypes are n-alkanes, aldehydes, POHs and free fatty acids (Figure 4), as previously reported in other wheat waxes (Netting and von Wettstein-Knowles, 1973; Tulloch and Hoffman, 1973; Koch et al., 2006). Although they were not resolved from free fatty acids in the solvent system used, waxes of both varieties contain ARs and MARs (Figure 4).

Table 1. Spectrophotometric OD273 measurements comparing the relative abundance of β-diketones in Shango and Shamrock flag leaf epicuticular waxes from field-grown plants at 20 days post-anthesis in 2011
VarietyWax (mga)Leaf (mgb)Total Dik (mg)Percentage of Dikcμg Dik mg−1 leaf−1
βOH-ββOH-β
  1. Values are the mean of two biological replicates; Dik, β-diketones plus hydroxy-β-diketones; β, β-diketones; OH-β, hydroxy-β-diketones.

  2. a

    Dried weight of the leaf extract.

  3. b

    Estimates based on OD273 measurements.

  4. c

    Fresh wet weight.

Shango6.399594.2899.50.544.670.03
Shamrock2.819190.2197.92.120.23<0.01

The standard approach to obtain an overview of the major wax constituents is to subject silylated wax samples to gas chromatography/mass spectrometry (GC-MS). In the resulting total ion chromatogram (TIC) traces, non-silylated aliphatics, for example n-alkanes and aldehydes, are less prominent than silylated POHs, and silylated β-diketone aliphatics are underestimated versus silylated POHs (Tulloch and Hogge, 1978). Furthermore, in β-diketone-containing waxes, not all components are resolved by this technique. Between 18.5 and 20 min, three non-symmetrical peaks occur in the Shango TIC traces (Figure 5a,c), but not in the Shamrock TIC traces (Figure 5b,d). Subjecting the C31 standards hentriacontane-14,16-dione and 25-hydroxy-14,16-dione to the same conditions resulted in similarly shaped peaks eluting at the same retention times. The first eluting peak is the non-derivatized β-diketone, hentricontane-14,16-dione, the second is a mixture of silylated isomers of the same compound, and the third contains a mixture of silylated hydroxy-β-diketones (Figure 5a–c). The presence of a strong ion at m/z 100 arises from cleavages alpha to the carbonyl groups when neither has been silylated. When one of the carbonyls of the specified β-diketone is silylated, cleavage adjacent to a carbonyl gives rise to the prominent diagnostic ions m/z 325 and 353 accompanied by M+ and M-15 ions. When a 25-silylated hydroxy group is also present, cleavage adjacent thereto gives rise to strong m/z ions at 539 and 187, and, if adjacent to the carbonyl-bearing carbons, the ions m/z 325 and 441 as well as M+ and M-15 ions (Tulloch and Hogge, 1978). As (i) none of the above-mentioned diagnostic peaks is seen in the Shamrock TIC traces (Figure 5b,d), (ii) trace components in cuticular waxes absorbing at 273 nm result in over-estimation of the amount of the β-diketone aliphatics in Table 1 (von Wettstein-Knowles, 1976), and (iii) the latter are undetectable on TLC plates (Figure 3), we conclude that β-diketone aliphatics are most likely absent in the investigated Shamrock waxes.

Figure 5.

GC-MS total ion chromatograms of Shango (a, c) (iw1/iw1) and Shamrock (b, d) (Iw1/Iw1) flag leaf blades (a, b) and peduncles (c, d). The vertical axis is relative abundance; IS, internal standard (C30 ALK); asterisks indicate non-silylated. Fatty acids (FA), n-alkanes (ALK), primary alcohols (POH), aldehydes (ALD), methylalkylresorcinol (MAR), and both β- and hydroxy-β-diketones (β-Dik and OH-β-Dik) are indicated.

Figure 5(a) also shows that, among the non-β-diketone aliphatics, C28POHs are the dominant constituents of the total flag leaf waxes, amounting to 68% of Shamrock wax. This is in accordance with the Shamrock scanning electron microscopy and TLC observations (Figures 3 and 4). As plate structures were not observed on the abaxial surface of Shango leaves, the observation that C28POHs are such prominent constituents in TIC traces was unexpected, but similar situations have been described (Baker, 1982). C28POHs are essentially absent in peduncle waxes of both varieties (Figure 5c,d). Next in frequency are C24POHs plus C29 and C31 n-alkanes in both flag leaf and peduncle waxes (see also Table 2). Closer examination of the traces supplemented with the pertinent ion traces identified n-alkanes with chain lengths from C22–C33 and POHs with chain lengths from C18–C34. Using this technique, a lower abundance of C22–C32 free fatty acids and C24–C34 aldehydes was observed. To further strengthen identification and reveal peaks co-eluting with other compounds, aliquots of the waxes were subjected to reduction with NaBH4, oxidation with K2Cr2O7 or passed through NaOH columns before silylation. This resulted in the expected reduction and shifts of peak retention times (data not shown). The 53 identified wax components are listed in Table S3.

Table 2. Total wax load (μg wax mg−1 leaf−1) on field-grown flag leaf blades and peduncles, plus amounts (μg mg−1) of the six most important constituents in the parental Shango and Shamrock lines, four DH lines and four BC2F3 lines at 9 days post-anthesis in 2010
TissueWheat lineIw1 +/−Wax load (μg wax mg−1 leaf−1)n-alkanesPrimary alcoholsβ-diketoneHydroxy-β-diketone
C29C31C24C28
  1. ND, not detected.

Flag leafShango12.680.100.040.123.297.091.59
Shamrock+4.760.330.130.553.230.06ND
DH11910.810.110.050.063.565.101.46
DH93+5.320.380.130.783.450.06ND
DH2113.060.110.050.112.857.861.67
DH81+4.800.340.140.413.38<0.01ND
Alchemy7.670.120.080.133.292.680.83
Alchemy+4.430.200.260.393.09NDND
Malacca5.820.100.040.153.990.710.35
Malacca+4.800.210.090.363.68NDND
PeduncleShango6.030.040.010.07ND4.541.34
Shamrock+0.810.270.050.37ND0.040.03
DH1193.420.070.010.04ND2.320.96
DH93+1.240.460.050.61<0.010.03ND
DH214.260.040.010.05ND3.131.01
DH81+0.820.350.040.38<0.01NDND
Alchemy1.820.020.010.03ND1.230.53
Alchemy+0.270.080.060.10NDNDND
Malacca0.910.060.010.11ND0.330.40
Malacca+0.440.160.030.18<0.01NDND

Alkyl and methylalkylresorcinols are present in wheat flag leaves

While trying to identify minor components of the waxes, one with a strong m/z 282 ion plus an M+ 562 ion was noted. As an m/z 282 ion is characteristic of MARs, this component was potentially identified as a MAR with an alkyl chain of 23 (C23MAR). As the presence of these phenolic lipids had not been previously reported in wheat waxes, they were further analysed. TLC plates developed in hexane:diethyl ether (8:2 v/v) and sprayed with 0.05% Fast Blue B revealed purple spots migrating close to the origin (Rf = 0.05) in Shango and Shamrock flag leaf and peduncle wax samples. When the solvent was changed to a combination of chloroform and ethyl acetate (8.5:1.5 v/v), in which the MARs would be expected to be more soluble, the purple spots migrated with an Rf of 0.42. Treatment of total waxes with K2Cr2O7 oxidized the hydroxyl groups on the benzol ring and abolished the Fast Blue B staining of MARs on TLC plates. When silica was removed from the appropriate region of preparative TLC plates and the compounds were recovered for analysis by GC-MS and ion scanning, we identified, in addition to a homologous series of MARs (C19–C27) with the 21 and 23 homologues being most prominent, an analogous homologous series of ARs with the 23 and 27 homologues being most prominent. These are characterized by a strong m/z 286 ion and M+ ions differing by 28. ARs have been frequently reported in wheat grains (Knödler et al., 2010). Figure 6 shows ion scans at m/z 268 and 282 for the ARs and MARs, respectively, in total waxes from flag leaf blades of Shango and Shamrock.

Figure 6.

Alkylresorcinol (AR) and methylalkylresorcinol (MAR) components of waxes from flag leaf blades of Shango (a) (iw1/iw1) and Shamrock (b) (Iw1/Iw1) as disclosed by mass spectrometry ion scans at m/z 268 (blue) and 282 (red), respectively. Total ion chromatogram traces are shown in black (see Figure 5).

Effect of Iw1 on the major aliphatics in the cuticular wax of flag leaves and peduncles of DH and BC2F3 wheat lines

Flag leaf and peduncle wax samples from four DH lines arising from the Shango × Shamrock cross were collected 9 days after anthesis. The results were compared to those of the parents (Figure 5) in Table 2. The wax load on Shamrock flag leaves was 38% of that on Shango, whereas that on peduncles was only 13% of that on Shango. Analogous differences were found among the DH lines. Similar to Shamrock, lines DH93 and DH81 carrying the dominant Iw1 allele had minor amounts of β-diketones, whereas lines DH119 and DH74, both homozygous recessive, had significantly higher (< 0.001) but not identical amounts. The ratios of β-diketones to hydroxy-β-diketones in the three glaucous lines were also similar, albeit higher on the flag leaves (3.5–4.7:1) than on the peduncles (2.4–3.4:1). In comparison, the amounts of C28POHs were not significantly different among all four DH lines and the parents. Intriguingly, the amounts of n-alkanes and C24POHs were consistently greater (< 0.001) in lines with the Iw1 dominant allele.

Analogous flag leaf and peduncle waxes from five pairs of BC2F3 NILs were also analysed. Across genotypes, the effects of Iw1 alleles were consistent and similar to those seen in the Shango × Shamrock DH lines. The effect was independent of the glaucousness of the original varieties, as exemplified by the two lines with the highest (Alchemy) and lowest (Malacca) wax load (Table 2). Alchemy, with only 27 and 38% as much β-diketone aliphatics as Shango on the peduncle and flag leaf, respectively, had similar ratios of β-diketones to hydroxy-β-diketones as Shango and the DH lines lacking the dominant Iw1 allele. In Malacca, however, the ratio was dramatically reduced on both cuticle surfaces. In all five NIL pairs, the relationship among the relative amounts of the C24POHs and n-alkanes was consistent. The lines with the dominant Iw1 allele had more C24POHs and n-alkanes than those without, although the increased amount did not compensate for the reduction in the β-diketone aliphatics (Table 2). The combined observations imply that the inhibitory effect of the gene operates independently of the genetic background, and functions during formation of β-diketone aliphatics in flag leaf and peduncle wheat cuticles. Moreover, concordant small increases in the amounts of n-alkanes and C24POHs accompany the inhibition.

Effect of Iw1 throughout plant development

The presence of visible wax on glaucous varieties is clearly manifested at anthesis. To understand the effect of development on the wax profile and the gene's effect, we performed a time course experiment using flag leaves from the Alchemy, Malacca, Robigus and Xi19 NIL pairs, differing at Iw1. At the early reproductive stage (Zadoks GS31, first node detectable within leaf sheaths; Zadoks et al., 1974), plants lacked visible wax, and no obvious phenotypic differences were detected between NIL pairs (Figure 7). GC-MS analysis showed that C28POHs were almost the only component of the wax (approximately 6 μg/mg leaf) at this stage in all NILs (Figure 7d). At the GS47–49 stage (boot stage), the glaucous phenotype is visible and the flag leaf sheath is opening, but the ear has not yet emerged. C28POHs were still the dominating wax components, but interestingly amounted to only approximately 2 μg mg−1 flag leaf−1, significantly lower than at GS31. This is most likely due to the differences in leaf thickness, and hence weight, between early vegetative leaves and the flag leaf that was sampled at boot stage (no flag leaf exists at GS31). The wide array of other compounds characterizing Shango and Shamrock (Figure 5) were also detected by GC-MS. In TIC traces of NILs carrying the dominant allele of Iw1, β-diketone aliphatics were absent (Figure 7e,f), but the levels of n-alkanes (Figure 7a,b) and C24POHs (Figure 7c) were significantly higher than in glaucous lines. This led to insignificant differences in total wax load (Figure 7g) between any NIL pair as determined by GC-MS, despite the clear differences in visual wax across all varieties in the field.

Figure 7.

Time-course analysis of major cuticular wax components of field-grown BC2F3 NILs across five growth stages in 2011 [Zadoks GS31, GS47, GS51, 18 days post-anthesis (DPA) and 42 DPA].
(a, b) C29 and C31 n-alkanes, (c, d) C24 and C28 primary alcohols, (e, f) β- and hydroxy-β-diketones, and (g) total wax load. Green lines and triangles represent Iw1/Iw1 NILs; grey circles represent iw1/iw1 NILs. Asterisks indicate significance at *P < 0.05, **P < 0.01 and ***P < 0.001.

Five days later, at GS51 (first spikelet of the ear just visible above flag leaf ligule), the amounts of β-diketone aliphatics had increased, and, together with C28POHs, were the major wax components in glaucous NILs. The increase in β-diketone aliphatics continued through flag leaf development, but these compounds remained absent in NILs with the dominant Iw1 allele (Figure 7e,f). The levels of n-alkanes and C24POHs also increased during flag leaf development in both sets of NILs, as did the difference between each NIL pair. The total amount of C28POHs remained similar between NIL pairs across all time points, increasing after 18 days post-anthesis (Figure 7d).

Iw1 does not affect 13C discrimination (Δ) in flag leaves, spike or grain

To evaluate the effect of the various wax loads on water-use efficiency (WUE), we determined the carbon isotope discrimination (Farquhar et al., 1989) for tissues from field-grown plants in 2011 and 2012. Flag leaves and spikes were collected at anthesis and grains were collected at maturity. No significant effect was detected for Iw1 in either 2011 or 2012: the NIL pairs were not significantly different in any of the three tissues examined in both the Malacca or Alchemy genetic backgrounds (Figure 8). However, the Δ of flag leaves and grains was significantly greater in 2012 than in 2011 (by approximately 1.5–2.5‰, Figure 8a,b, P < 0.001).

Figure 8.

Carbon isotope discrimination of Alchemy and Malacca BC2F3 NILs in 2011 (a) and 2012 (b). Green bars represent non-glaucous Iw1/Iw1 NILs; grey bars represent glaucous iw1/iw1 NILs. No significant differences were detected between NIL pairs across tissue or years. GS, Zadoks growth stage.

Discussion

Fine mapping of Iw1

The use of synteny has been extensively exploited in wheat genetics to fine map and clone genes (Krattinger et al., 2009). This approach is proving increasingly powerful due to the generation of new genome sequences from closely related species such as Brachypodium (International Brachypodium Initiative, 2010) and the genomic contigs of barley (International Barley Genome Sequencing Consortium, 2012). The unrestricted access to the wheat chromosome arm assemblies is also a key resource. For the first time, wheat researchers have the possibility of accessing genomic contigs of all three homoeologues, a critical step for more targeted marker design. Likewise, a simple BLAST search provides evidence of putative chromosome localization (Altschul et al., 1990). These tools will allow researchers to focus marker design on those genes that have the greatest probability of lying in candidate gene intervals.

For fine mapping of Iw1, we exploited these new resources by focusing marker design to the Iw1 interval (JIC007 and JIC012), including seven genes in Brachypodium and 21 genes in rice. Use of the wheat chromosome arm assemblies allowed us to consider only those genes with evidence of 2BS localization, including the Bradi5g01180/Os04g05010 gene. This approach allowed discrimination for genes present in only one of the two sequenced genomes. For example, Bradi5g01160 and Os04g05030 both had evidence of 2BS genomic sequence, and were genetically mapped to the Iw1 interval, whereas an additional 18 genes in rice were excluded from further analysis as they showed evidence of localization to other wheat chromosomes. In an analogous manner, we used the recently released barley genomic contig to identify two additional barley genes (MLOC_20994 and MLOC_6767) with evidence of wheat group 2 chromosome localization.

The synteny-based approach greatly facilitated fine mapping of Iw1 to a sub-cM genetic interval. However, the absence of β-diketone aliphatics in Brachypodium and rice makes the existence of a homologue of the Iw1 dominant inhibitor of the β-diketone biosynthetic pathway highly unlikely in these species. Furthermore, barley, which synthesizes β-diketone aliphatics, may also lack such a gene given the very rare frequency of chromosome 2H dominant eceriferum mutants (King and von Wettstein-Knowles, 2000), or it may be located elsewhere in the genome. This implies that a gene in the sub-cM interval of 2BS has a different function than its orthologues in Brachypodium and rice. A pertinent example is the GPC gene, which encodes orthologous NAC transcription factors with divergent functions in wheat and rice (Distelfeld et al., 2012). Construction of a physical map and the final identification of Iw1 will help to answer these questions. At present, the combined genetic and in silico analysis yields three candidate genes for Iw1: Bradi5g01180/Os04g05010, MLOC_20994 and MLOC_6767.

Iw1 candidate genes

The complete linkage between JIC009–11 and Iw1 in the Shango × Shamrock population was surprising based on the large number of F2 gametes screened and the expected high levels of recombination in the distal ends of wheat chromosomes (Lukaszewski and Curtis, 1993). Moreover, the amplification of gene-based markers JIC009–11 in all the germplasms used in this study suggests that these genes have not been deleted. The complete linkage between JIC009 and Iw1 across both populations may therefore be interpreted to mean that (i) these two genes are in very close proximity, (ii) a deletion event occurred in glaucous varieties proximal to JIC009, thereby suppressing recombination between glaucous and non-glaucous lines, or (iii) the wheat homologue of Bradi5g01180/Os04g05010 is Iw1.

The three candidate genes share no homology and have different conserved domains and putative functions. The wheat homologue to Bradi5g01180/Os04g05010 encodes a protein that includes a C-terminal cystathionine β-synthase domain. These domains have no assigned biological function in plants, but gene expression-based analyses in rice and Arabidopsis suggest that some may play a role in stress response and development (Kushwaha et al., 2009). The MLOC_6767 gene encodes a laccase-like multi-copper oxidase protein that forms part of a large family of proteins with broad functions. In Arabidopsis, there is evidence that laccase contributes to lignification of stems (Berthet et al., 2011) and oxidative polymerization of flavonoids in the seed coat (Pourcel et al., 2005). The MLOC_20994 gene encodes a protein that includes a C-terminal Mu-homology domain (MHD) that defines a family of conserved endocytic adaptor proteins called muniscins (Conibear, 2010). These proteins have been characterized in yeast (Reider et al., 2009) and mammals (Uezu et al., 2007), and have been shown to interact with phospholipids and to facilitate vesicle formation. MHD proteins remain uncharacterized in plants. We are currently performing allelic diversity studies and complementation experiments to determine whether any of these genes encode Iw1.

Complete series of ARs and MARs are present in wheat flag leaves

The identification of complete homologous series of odd-numbered ARs and MARs (C19–C27) raises the interesting question as to how they are synthesized. Similarly to stilbene synthase, a type III polyketide synthase, which adds three C2 units from malonyl CoA to p-coumaric CoA to form a tetraketide intermediate that cyclizes to give resveratrol (Yu and Jez, 2008), alkylresorcinol synthases also form a tetraketide intermediate at the end of C22 acyl chains that cyclize with loss of a carbon to give an AR: 1,3-dihydroxy,5-C15-benzene (Cook et al., 2010). If methyl malonyl replaces malonyl CoA in one of the last three elongation steps, the 1,3-hydroxybenzene ring will bear a methyl group on carbon 2, 4 or 6 (MARs). Is there a specific elongation step in which methylmalonyl CoA replaces malonyl CoA, as in Pinus strobus (Schröder et al., 1998), or does this occur at any of the three steps to produce a MAR? Why does this not occur at all when ARs are synthesized, or never twice? While these are questions for the future, clearly the type III polyketide synthase(s) giving rise to the ARs and MARs of wheat cuticular wax is/are different from the pkKCS that produces the triketide intermediate with the oxygens close to the middle of the carbon chain that do not cyclize in β-diketone synthesis. Furthermore, Iw1 only inhibits synthesis of the latter. This information will be useful in trying to predict the function of Iw1.

Effect of Iw1 on water-use efficiency

At the plant physiological level, WUE refers to the instantaneous water-use efficiency of gas exchange, A/T, where A is net photosynthesis and T is transpiration. As A is dependent on the ratio between the concentration of CO2 outside and inside the leaf, relative changes in the proportion of these two parameters, and related changes in Δ, are used to estimate shifts in plant A/T. By measuring the δ13C values of plant biomass, and thus determining Δ (Farquhar et al., 1989), we assessed to what extent lw1 affected WUE during the growing season in 2011 and 2012.

Our results show that lw1 did not affect Δin flag leaves, spikes and grains in Malacca and Alchemy, indicating that WUE was unaffected in these lines. Our inter-annual data showed clear differences in WUE between 2011 and 2012 (Figure 8), which differed in the amount of rainfall and relative humidity, both of which were higher in 2012 (Figure S2). This suggests that, in the UK, a change in moisture availability has a greater effect on wheat WUE than the amount and type of organic compounds that comprise cuticular waxes. Our Δ results are not consistent with previous research that found a positive link between glaucousness and Δin wheat grain and leaves (Merah et al., 2000; Monneveux et al., 2004). Variation in the amount of cuticular wax load among varieties may be one of the reasons for these differences. Furthermore, the generic classification of non-glaucousness irrespective of the genetic source (Iw1, Iw2, w1 or w2) hinders comparisons. Future studies based on glaucous and non-glaucous germplasm with a defined genetic make-up, including quantitative and qualitative wax compositions, will help to explore the link between glaucousness and WUE in more detail.

Concluding remarks

Introgression of Iw1 into seven genetic backgrounds led to inhibition of β-diketone aliphatic synthesis and a small increase in the amount of n-alkanes and C24POHs, but cuticle thickness and water use efficiency were unaltered. How does Iw1 achieve this? The 10 min duration used for wax extraction implies that both epi- and intra-cuticular waxes were recovered, suggesting that the lack of β-diketone aliphatics does not result from a defect in their transport to the cuticle surface, but rather a defect in their biosynthesis or the regulation thereof, given the dominant nature of Iw1. If inhibition occurs at the branch point of the fatty acid elongase and pkKCS pathways, one may expect extensive re-channelling of precursors from one to the other, but this appears unlikely as the increase in n-alkanes and C24POHs is less than 10% of the decrease in β-diketone aliphatics. The Cer-cqu gene in barley maps to a similar region on chromosome arm 2HS as Iw1 on wheat 2BS; however, only one of 522 mutations is dominant (King and von Wettstein-Knowles, 2000). All 18 Cer-yy induced mutations are dominant, analogous to Iw1, but Cer-yy maps to chromosome 1H (Lundqvist and von Wettstein-Knowles, 1982; von Wettstein-Knowles, 1990). As a number of barley and wheat varieties carry dominant alleles of these inhibitors, identifying the molecular nature of Iw1 and Cer-yy is of great interest.

Experimental Procedures

Plant material

The previously published Shango × Shamrock DH population (87 lines) (Simmonds et al., 2008) was used to map EST-derived markers across chromosome arm 2BS. An F2 population generated from these parents was used to assess segregation ratios and identify recombinants across the region. Plants with recombination between markers JIC004 and JIC015 (Data S3 and Figure S3) were selfed, and homozygous F3 recombinants were selected and phenotyped. A total of 2111 Shango × Shamrock F2 plants were screened (4222 gametes).

An additional 850 F2 plants were screened from a durum wheat cross between glaucous variety Langdon (iw1/iw1) and non-glaucous wild emmer accession TTD140, which carries Iw1 (Rong et al., 2000). A similar strategy was used as for the Shango × Shamrock recombinant plants. Additional F2 populations were developed by crossing six glaucous hexaploid UK varieties (Alchemy, Einstein, Hereward, Malacca, Robigus and Xi19) to Shamrock. To generate the BC2 NILs, plants heterozygous across the Iw1 interval were selected at each generation using markers Xgwm614 and Xwmc25 (Alchemy, Einstein, Malacca and Xi19) or Xgwm614 and Xwmc154 (Hereward and Robigus), and crossed again to Shamrock. After the second back-cross, self-pollinated homozygous BC2F2 NILs were selected.

Isolation of waxes

To study the effect of Iw1 on cuticular waxes, five flag leaves and peduncles from independent replications were collected for Shango, Shamrock, the four DH lines and the five BC2 NIL pairs (Alchemy, Einstein, Hereward, Malacca and Robigus) from the 2009-2010 field plots. For the time-course analysis, three flag leaves from independent replications were collected for the four BC2 NIL pairs (Alchemy, Malacca, Robigus and Xi19) grown in 2010-2011. Samples were collected in the field, placed in pre-weighed 15 ml polypropylene tubes, and frozen on dry ice. Tubes were re-weighed to determine the wet tissue weight before freezing in liquid nitrogen and storing at −80°C.

Waxes were extracted using 5 ml chloroform (Merck, analytical grade; as are all other solvents below, https://;www.merckmillipore.com/united-kingdom/chemicals) in glass tubes with screw-cap polytetrafluoroethylene lids containing chloroform and triacontane (C30 alkane, Sigma Catalogue 263842, http://www.sigmaaldrich.com/united-kingdom.html) as an internal standard (35 μg ml−1 for leaves and 10 μg ml−1 chloroform for peduncles); samples were immersed for 10 min at room temperature and shaken three times for 10 sec. The extracts were transferred to new glass tubes and dried down in a vortex evaporator (3-2201, Buchler Instruments Inc.). Each wax sample was re-suspended in 1 ml chloroform and transferred to a pre-weighed glass vial, dried under nitrogen and then re-weighed to determine the total amount of wax extracted.

Gas chromatography-mass spectrometry

Wax samples were derivatized in a 100 μl aliquot of a pyridine and N,O-Bis (trimethylsilyl) trifluoroacetamide with 1% trimethylchlorosilane (Sigma 15238) mixture (1:1 v/v) at 75°C for 1 h; samples were vortexed every 15 min. Commercial standards [C30 alkane (Sigma 263842), a mix of C7–C40 n-alkanes (Sigma 49452), a mix of 100 μg each of 1-tetracosanol (Sigma L350), 1-hexacosanol (Sigma H2139), 1-octacosanol (Sigma O3379) and 1-triacontanol (Sigma T3777)] and samples of β-diketones (96% hentriacontane-14,16-dione) and hydroxy-β-diketones (97% 25-hydroxyhentriacontane-14,16-dione) isolated from barley (Hordeum vulgare L. cv. Bonus) spikes (von Wettstein-Knowles, 1976) were derivatized similarly.

The derivatized fraction was analysed on an Agilent GC 6890N gas chromatograph (Agilent Technologies, http://uk.rs-online.com) equipped with a 5973 inert mass selective detector. Automated split-less 3 μl injections were performed using an Agilent 7683 automatic sampler. The conditions of chromatography were inlet temperature 250°C, He carrier gas at a flow rate of 0.8 ml min−1, nominal inlet pressure of 9.27 psi, oven temperature program from 140°C (1 min) to 380°C (at 10°C min−1), then held for 5 min. The column used was a -5HT Inferno (Zebron; 7HG-G015-02, Phenomenex, http://www.phenomenex.com/) (30 m × 0.25 mm × 0.1 μm) with a 5 m guard column fitted to the front end. The retention time locking feature was used, and the method locked to the retention time of the triacontane internal standard (16.3 min). The mass spectrometer parameters using electron ionization in positive mode (70 eV), with a source temperature of 230°C and a quad temperature of 150°C, were set to the manufacturer's recommended defaults. Total ion scans used 50–500 amu; all data were processed using agilent gc chemstation software (D.03.00) in conjunction with the National Institute of Standards and Technology mass spectral library, version 8.0.

Quantification of wax compounds

Subtracting the percentages of the β-diketone aliphatics from the total wax gives the percentage attributable to the other components. The relative abundances for these compounds were calculated from GC-MS TIC peaks by automatic integration using the Custom Report function in the ChemStation software. Where compounds such as the C26 FA/C28 aldehyde and C28 FA/C30 aldehyde eluted closely together, such that individual TIC peaks could not be integrated separately, the major ion for each compound was searched and integrated separately for the relevant retention time. The derivatized β-diketone peak was integrated manually, and then the characteristic ion for the obscured C30 FA peak was integrated and subtracted from the β-diketone peak. The same approach was used to estimate the C23MAR peak that is hidden within the hydroxy-β-diketone peak. While the data presented do not take into account that not all wax aliphatics are silylated nor the differential responses of the chemical groups to flame ionization (Sternberg et al., 1962), they give a reproducible approximation of the quantities of the wax aliphatics. All together, 53 components were identified, with 26 being studied in more detail as they account for >95% of the total wax load in both Shango and Shamrock flag leaves and peduncles. Additional procedures are available as Data S4–S8.

Acknowledgements

This work was supported by the Biotechnology and Biological Sciences Research Council (grants BB/H018824/1, BB/J004553/1, BB/J004596/1 and BB/I002545/1), the John Innes Foundation, the National Institute of Agricultural Botany Trust and Earth & Life Systems Alliance. We thank Moshe Feldman (Weizmann Institute of Science, Rehovot, Israel) for providing seeds of TTD140; Hana Simkova and Jaroslav Dolezel (Institute of Experimental Biology, Olomouc, Czech Republic) for providing DNA of flow-sorted arms, the International Wheat Genome Sequencing Consortium for pre-publication access to chromosome arm assemblies, Nils Stein and Ruvini Ariyadasa (Institut für Pflanzengenetik und Kulturpflanzenforschung, Gatersleben, Germany) for barley contig information, Lionel Hill (John Innes Centre, Norwich, UK), Jack McCurley and Paul Disdle (University of East Anglia, Norwich, UK) for technical assistance, and the John Innes Centre glasshouse staff (Peter Sawdon, Lionel Perkins, Damian Alger and Barry Robertson) for plant husbandry.

Ancillary

Advertisement