Introduction

The aerial surfaces of land plants are covered with a cuticle, a continuous hydrophobic layer mainly consisting of two major types of lipids, cutin and waxes (Kunst and Samuels 2009; Buschhaus and Jetter 2012). The cutin is a polymer which consists of omega and mid-chain hydroxy and epoxy C16 and C18 fatty acids, as well as their derivatives (Samuels et al. 2008). Wax compounds are deposited both within (intracuticular waxes) and on the surface (epicuticular waxes) of the cutin matrix, where they can form as crystals (Jetter and Schaffer 2001; Kannangara et al. 2007). The waxes protect plants from non-stomatal water loss, pathogen invasion and other stresses such as dust, insects and frost (Sieber et al. 2000; Aharoni et al. 2004; Sturaro et al. 2005; Li-Beisson et al. 2013; Espana et al. 2014).

Cuticular waxes are mainly composed of VLCFAs and their derivatives including aldehydes, primary and secondary alcohols, alkanes, ketones and wax esters (Yeats and Rose 2013). The biosynthesis of wax occurs exclusively within the plastid of epidermal cells where the C16 and C18 fatty acids were produced. Then the two fatty acids are used as precursors for the generation of VLCFAs up to 38 carbons followed by the decarbonylation and acyl reduction pathways to derive all the components (Shepherd and Wynne Griffiths 2006). The extension of C16 and C18 fatty acids to VLCFAs occurs on the endoplasmic reticulum (ER) via the fatty acid elongation (FAE) complex (Lee and Suh 2013). The FAE includes four enzymes: a β-ketoacyl-CoA synthase (KCS), a β-ketoacyl-CoA reductase (KCR), a β-hydroxyacyl-CoA dehydratase (HCD) and an enoyl reductase (ECR). Each elongation cycle involves four successive enzymatic reactions: condensation, reduction, dehydration and reduction, which together extend the substrate’s carbon chain by a C2 unit in one cycle (Haslam and Kunst 2013).

Unlike the other three enzymes, the KCS condensing enzyme showed strict substrate specificity and was considered as the cycle’s rate-limiting enzyme (Paul et al. 2006). The Arabidopsis genome contains 21 FAE-like KCS members, which have been found to have different substrate specificites, such as FAE1 (James et al. 1995), KCS1 (Todd et al. 1999), CER6/CUT1/KCS6 (Millar et al. 1999; Fiebig et al. 2000; Hooker et al. 2002), KCS2/DAISY (Lee et al. 2009), KCS20 (Lee et al. 2009) and KCS9 (Kim et al. 2013). However, to date, only a few KCS genes have been cloned in rice, like WSL1 (Yu et al. 2008), ONI1 (Ito et al. 2011) and ONI2 (Tsuda et al. 2013).

Here, we used map-based cloning to identify the rice Wax crystal-sparse leaf 4 (WSL4) gene that involved in leaf cuticular wax accumulation. The wsl4 mutant exhibited reduced epicuticular wax crystals on the leaf surface compared to its wild type (WT). Amino acid sequence analysis suggested that WSL4 was a member of the KCS family. WSL4 was constitutively expressed in rice and localized in the ER, where the FAE complex was localized. Further study showed both knocks-down and knocks-out of WSL4 reduced cuticular wax loads on leaves. Together, all these data indicate that WSL4 is involved in wax biosynthesis in rice.

Materials and methods

Plant material and growth conditions

wsl4, a rice leaf wax deficient mutant, is a tissue culture-induced mutation of Japonica cv Nipponbare. For fine mapping of the wsl4 gene, the wsl4 mutant was crossed with indica cv Yuewanxian to construct an F2 mapping population. All rice plants used in this study were cultivated in the experimental field under normal growth conditions at Changping (Beijing, China) or Sanya (Hainan, China).

Scanning and transmission electron microscopy

For scanning electron microscopy (SEM) analysis, leaves and leaf sheaths were air dried. Epicuticular wax crystals were imaged on a HITACHI 8100 variable-pressure scanning electron microscope. To observe the transverse cuticle structure of wsl4, transmission electron microscopy (TEM) analysis was carried out as described previously (Mao et al. 2012).

Water loss and chlorophyll leaching assays

For water-loss assays, samples were treated as described previously (Qin et al. 2011). Each sample was weighed with a microbalance at 0, 1, 2, 3, 4, 5 and 6 h. The water loss rate was calculated based on the initial weight of the samples.

For chlorophyll leaching measurements, samples were treated according to Mao et al. (2012). The chlorophyll content was quantified with a spectrophotometer (DU-800, Beckman Coulter, USA) at 647 and 664 nm absorption peaks, respectively, according to Lolle et al. (1997).

Wax and cutin analysis

Cutin monomer analysis was performed according to Li et al. (2010). Leaves that had been used for wax extraction were re-extracted in fresh chloroform/methanol (1:1 v/v) four times for several hours each. Then the samples were lyophilized. The delipidated samples were then depolymerized using transesterification in 1 mL of 1 N methanolic HCl at 80 °C for 2 h. After the addition of 2 mL of saturated NaCl/H2O and 10 mg of dotriacontane (Fluka, USA) as an internal standard, the hydrophobic monomers were subsequently extracted three times with 1 mL of hexane. The organic phases were combined, and then the solvent was evaporated. The remaining sample was derivatized and cutin monomers were detected by gas chromatography–mass spectrometry (GC–MS). The cuticular wax was extracted from leaves and analyzed using GC–MS as described previously (Mao et al. 2012).

Molecular cloning of the WSL4 gene

All 4-leaf-stage F2 seedlings were immersed in water to identify the extreme water-sticking individuals. A total of 811 leaf wax deficient plants were collected for WSL4 gene mapping. Genomic DNA was extracted by the CTAB method and the linkage analysis was carried out using both published and newly developed markers (Supplementary Table 1), according to the sequence diversity between Nipponbare and 9311 (indica var.) available on Gramene (http://www.gramene.org/).

Complementation of the wsl4 mutant

A 7649-bp genomic fragment of WSL4 containing the full length gene region, the 3902-bp upstream fragment, and 1608-bp downstream fragment was amplified from rice genomic DNA with the primer pairs WSL4-2300-C F/R. The PCR product was subcloned into the EcoRI/SmaI sites of binary vector pCAMBIA2300 using the InFusion Advantage PCR Cloning Kit (Takara, Japan) and sequenced. The complementation vector was introduced into the wsl4 mutant by Agrobacterium-mediated transformation as described previously (Hiei and Komari 2008).

RNA interference and knock-out of WSL4

The RNA interference vector pCUbi1390-ΔFAD2 (ubiqutin promoter and FAD2 intron inserted in pCAMBIA1390) used for RNAi was described previously (Li et al. 2013). A 387-bp pair cDNA fragment was amplified by PCR primer pairs of WSL4-1390-RNAi-1 F/R and WSL4-1390-RNAi-2 F/R from the cDNA of WSL4 gene, and then subcloned into the SacI and SnaBI sites of the vector pCUbi1390-ΔFAD2. The resulting WSL4-RNAi vector construct was introduced into Nipponbare by Agrobacterium-mediated transformation.

The high specificity target site for WSL4-Crispr was chosen by using the CRISPR-P online website (http://cbi.hzau.edu.cn/cgi-bin/CRISPR) and evaluated the secondary structural of RNA. The forward and reserved primers were mixed in a 1:1 ratio and annealed from 94 to 15 °C at the rate of 0.1 °C/s to form the double strands DNA and integrate into the pCRAC vector. The constructed vector was introduced into Nipponbare by Agrobacterium-mediated transformation. Positive T1 lines were subjected to further phenotypic evaluation.

RNA isolation and quantitative PCR analysis

Total RNA was isolated from WT and transgenic lines using the ZR Plant RNA MiniPrep Kit (Zymo research, USA), following the protocol provide by the manufacturer. 1 μg of RNA was reverse-transcribed into cDNA with the QuantiTech Reverse Transcription Kit (QIAGEN, USA). Quantitative RT-PCR (qRT-PCR) was performed in the 7500 Real-Time PCR System (Applied Biosystems, USA) using the SYBR Green PCR Kit (Takara, Japan) in a reaction volume of 20 μL. The 2−ΔΔCT method was used to calculate relative changes in gene expression as described (Rao et al. 2013).

Subcellular localization and GUS analysis

The coding region of WSL4 was amplified without the stop codon and cloned into the vector 1305-35S-GFP to generate a WSL4-GFP expression construct under the control of 35S promoter. To investigate the subcellular localization of WSL4 in plant cells, Agrobacterium strain EHA105 was transformed with 35S:WSL4-GFP, ER-marker, and P19 vectors, respectively, and cultivated overnight. Then the three Agrobacteria were mixed at a ratio of 1:1:1 to a final volume of 1 mL as described (Batoko et al. 2000). After incubating for 4 h in dark, mature N. benthamiana leaves were inoculated with the mixture using a syringe. After 2 days inoculation, the protoplasts were isolated by enzyme treatment from the leaf discs. GFP fluorescence was observed at the wavelength of 488 nm under confocal microscope (LSM 700, Carl zeiss, Germany).

For the GUS assay, a 3902-bp upstream fragment of WSL4 was amplified from rice genomic DNA with primer pair of WSL4-1305-GUS F/R. The PCR product was subcloned into the EcoRI/NcoI site of binary vector pCAMBIA1305 using the InFusion Advantage PCR Cloning Kit. The GUS vector was introduced into the Kitaake by Agrobacterium-mediated transformation. Excised tissues from independent transgenic lines were used for GUS activation assay according to the method described previously (Jefferson 1987).

Results

Morphological analysis of wsl4

To identify critical genes for cuticular wax synthesis in rice, we screened the existing mutant library in our lab, which includes two types mutants generated from EMS mutagenesis and tissue culture-induced mutagenesis, and obtained a tissue culture-induced mutant derived from Oryza sativa cv. Nipponbare. The mutant, named as wax crystal-sparse leaf 4 (wsl4), exhibited extreme water adhesiveness (Fig. 1a, b) when seedlings were immersed in water. In addition, wsl4 was shorter and produced fewer tillers than WT after heading (Fig. 1c, d). F1 hybrids of a cross between wsl4 and WT exhibited the WT phenotype, and the F2 population was segregated in a 3:1 (WT: wsl4) phenotypic ratio, indicating that the wsl4 phenotype was caused by a single recessive nuclear gene.

Fig. 1
figure 1

Characterization and cuticular permeability of the wsl4 mutant. a Phenotypic comparison of Nipponbare (left) and the wsl4 mutant (right); b water adhesiveness phenotype of the wsl4 mutant; c plant height (cm) of WT and the wsl4 mutant; d tiller numbers of WT and the wsl4 mutant; e water loss rates of excised leaves of WT and the wsl4 mutant. Water loss rates were measured at 0, 1, 2, 3, 4, 5 and 6 h. Data are shown as mean ± SE for three replicates; f chlorophyll leaching assays of WT and the wsl4 mutant leaves. Chlorophyll leaching rates of WT and the wsl4 mutant were measured from 0 to 10 h after immersion in 80% ethanol solution. Data are shown as mean ± SE of three replicates

Full size image

Structural and chemical analysis of cuticular waxes

To investigate the reason for the altered leaf water stickiness phenotype of wsl4, the analysis of SEM was performed. There are substantially fewer epicuticular wax crystals on wsl4 leaf surface, while upright schistose-shaped, pyknotic and regularly wax crystals were arrayed on leaf surface of the wild type (Fig. 2a, b). Similar results were observed on the leaf sheath of wsl4 (Fig. S1). To further investigate the ultrastructure of cuticle layers on leaf surface, we performed TEM analysis. The wsl4 possessed a thicker cuticular layer than the wild type (Fig. 2c, d). Subsequently, GC–MS was employed to analyze the cutin and cuticular wax composition. There were no significant differences in total amount of cutin monomers (Fig. 2e), while the cuticular wax coverage was greatly reduced in leaves of the wsl4 mutant, compared with the wild type (Fig. 2f).

Fig. 2
figure 2

Phenotypic analysis of WT and wsl4. SEM analysis of epicuticular wax crystal patterns on the leaf surface of WT (a) and wsl4 (b). Bar 1 μm. TEM analysis of the leaf cuticle membrane of WT (c) and wsl4 (d). The leaf cuticle membrane on wsl4 appears thicker than WT. The cuticle membrane is indicated between the white arrows. Bar 200 nm. e Cutin monomer composition on leaves of WT and wsl4. FA fatty acid, ω-OH FA ω-hydroxyl fatty acid, 2HFA C16-10,16-dihydroxyl fatty acids, Error bars represent ± SE of three biological replicates. f Cuticular wax load on the leaf surfaces of WT and the wsl4 mutant analyzed by GC–MS. Asterisks denote significant differences from WT. Error bars indicate ±SE (n = 3)

Full size image

Cuticular permeability analysis

We surveyed the cuticle permeability difference between the wsl4 and the wild type, and the results showed that the excised leaf of wsl4 lost water significantly faster, and that chlorophyll leached from wsl4 leaves more quickly than the wild type (Fig. 1e, f). These results suggest that wsl4 exhibits greater permeability than the WT.

Map-based cloning of WSL4

A total of 811 F2 plants with the mutant phenotype were used for fine mapping of WSL4. The mutant site was narrowed to a 58 kb region by the InDel markers M-6283 and M-6341 on chromosome 3 (Fig. 3a). This interval contains nine annotated genes according to the Rice Genome Annotation Project database (http://rice.plantbiology.msu.edu/). A single base mutation (G to A) was found on the fifth open reading frame (ORF) of LOC_Os03g12030, causing the 312nd amino acid to change from Val to Met (Fig. 3b). To verify the identity of WSL4/LOC_Os03g12030 as the candidate gene, a 7742 bp genomic fragment amplified from Nipponbare, including the entire coding region of LOC_Os03g12030 and its promoter sequences was introduced into wsl4. The leaf water-sticking phenotype of transgenic plants was rescued (Fig. 4a) and the wax crystals were also present as WT (Fig. 4b–e). These results indicated that the abnormal phenotypes of wsl4 are caused by the mutation of LOC_Os03g12030, and we therefore designated LOC_Os03g12030 as WSL4.

Fig. 3
figure 3

Map-based cloning of the WSL4 locus. a Fine mapping of the wsl4 locus gene on chromosome 3. Markers and the number of recombinants identified are shown. cM centimorgan. b Schematic gene structure of the Loc_Os03g12030 gene. The wsl4 mutant contains a mutation in nucleotides 1281 (G → A) in the second exon of the Loc_Os03g12030 gene, causing the mutation from Val312 to Met312

Full size image
Fig. 4
figure 4

Complementation tests of WSL4. a Phenotypes of transgenic plants in complementation tests. Positive transgenic rice of wsl4 with a complementary vector containing a 7204 bp genomic fragment of WSL4 rescues the water adhesive phenotype (right) similar to WT (left), compared to wsl4 with an empty vector (middle). b–e SEM analysis. All three individual complementation tests of transgenic lines in wsl4 background (c–e) show similar wax crystal distributions on leaf surfaces to WT (b), bar 1 μm

Full size image

Sequence alignment and phylogenetic analysis of WSL4

The WSL4 protein is composed of 494 residues with a predicted molecular mass of 55.79 kDa. The WSL4 protein processes high similarity to the FAE-type KCS family: 90.02% to barley HvCUT1; 78.79% to cotton GhCER6; 78.27% to Arabidopsis AtKCS6/CUT1/CER6 and 76.16% to AtKCS5; 77.22% to potato StKCS6/CER6 and 76.81% to tomato LeCER6 (Fig. 5a). The phylogenetic tree based on FAE type KCS members in Arabidopsis and rice revealed that WSL4, AtKCS6 and AtKCS5 were located on the same branch (Fig. 5b), suggesting that WSL4 is homologous to KCS6 that catalyzes the first step of elongation reactions of the FAE complex for VLCFAs.

Fig. 5
figure 5

Amino acid sequence alignment and phylogenic analysis of the WSL4 protein with homologs from other species. a The amino acid sequence alignment of WSL4 with homologs from other species. The mutant residue of WSL4 is marked by asterisk. b The phylogenic analysis of WSL4 with homologs from other species

Full size image

Spatial and temporal expression and subcellular location of WSL4

qRT-PCR analysis showed that WSL4 was widely expressed in rice leaf, culm, panicle, sheath and young seedling (Fig. 6a). The expression pattern of WSL4 was further confirmed by GUS staining of Pro WSL4 ::GUS transgenic plants. GUS activity was detected in various tissues including root, leaf blade, sheath, stem, lamina joint, glume and flower (Fig. 6b–i). To clarify the subcellular location of WSL4, we transiently expressed the WSL4-GFP fusion protein in tobacco leaves. The green fluorescent signals from WSL4-GFP co-localized with the ER marker mCherry-HDEL, indicating WSL4 is localized in the ER (Fig. 6j).

Fig. 6
figure 6

Spatial expression pattern of WSL4 and subcellular location of WSL4. a Analysis of WSL4 expression in different tissues by qRT-PCR. The UBQ gene was used as control and error bars represent ±SE of three biological replicates. b–i Gus expression patterns of P WSL4 :GUS transgenic rice plants. GUS activity was detected in different tissues: plumule (b), root (c), leaf blade (d), sheath (e), stem (f), lamina joint (g), glume (h), and flower (i). b–h Bar 1 mm, i bar 1 μm. j Co-expression of WSL4-GFP fusion protein and the HDEL-mCherry fusion protein in N. benthamiana protoplasts was imaged by confocal microscopy with a Zeiss LSM700 fitted with green (WSL4-GFP) and red filters (HDEL-mCherry) (color figure online)

Full size image

RNA interference and knocked-out of WSL4

To further confirm the function of WSL4 involved in wax biosynthesis in rice, the WSL4-RNAi transgenic plants and WSL4-targeted knocked-out mutant lines in Nipponbare were constructed. The leaf cuticular wax density of these plants was examined with SEM. The RNAi transgenic lines showed obvious leaf water-sticking phenotype (Fig. 7a). The expression of WSL4 in RNAi transgenic lines was greatly decreased (Fig. 7b). Besides, the SEM analyses of RNAi transgenic lines revealed substantial reduction in wax crystal density (Fig. 7d–f). Furthermore, all WSL4 knocked-out mutants exhibited wax load decrease on the leaf surface comparing to the WT (Fig. 7g–i). All these data suggest that WSL4 is required for cuticular wax accumulation in rice.

Fig. 7
figure 7

RNAi and Knocked-out of WSL4 in rice. a Phenotypic comparison of Nipponbare (left) and WSL4-RNAi transgenic plants (right). b Expression analysis of the WSL4 gene in different RNAi transgenic lines and WT by qRT-PCR. The UBQ gene was used as the control; error bars represent ± SE of three biological replicates. c–f SEM analysis of WSL4-RNAi plants and WT. All three individual WSL4-RNAi transgenic lines (d–f) in Nipponbare background show significantly decreased wax crystals on the leaf surface compared to WT (c). g Sequence analysis of WSL4-knocked-out transgenic plants. h, i SEM analysis of the WT and WSL4-knocked-out transgenic plants, bar 1 μm

Full size image

Discussion

In this study, we showed that WSL4 encodes a KCS6 which is involved in the VLCFAs elongation reactions. Comparing to the WT, wsl4 displays reduced surface wax loads. Map-based cloning and nucleotide sequencing results revealed that the phenotype of wsl4 is caused by a single nucleotide substitution in the second exons of LOC_Os03g12030, causing a conservative amino acid alternation of the whole peptide. Genetic complementation confirmed that the mutation in WSL4 was responsible for the mutant phenotype. Bioinformatics analysis indicated that WSL4 belongs to the KCS family and is highly homologous to AtKCS6/CUT1/CER6 (Millar et al. 1999; Fiebig et al. 2000), HvKCS6/CUT1.3 (Weidenbach et al. 2014), StKCS6 (Serra et al. 2009), LeCER6 (Leide et al. 2007) and GhCER6 (Qin et al. 2007), most of which were previously validated as one of the FAE complex involved in the first step of VLCFA elongation reactions.

In rice, three other KCS genes, WSL1, ONI1 and ONI2, have been cloned. The T-DNA insertion mutant wsl1 exhibits a pleiotropic phenotype including sparse wax crystals. WSL1 encodes a protein of KCS family and is ubiquitously expressed in rice, the VLCFA precursors of C20–C24 of total wax load are reduced on both leaf blades and sheathes in wsl1 (Yu et al. 2008). ONI1 encodes a fatty acid elongase similar to AtFDH which functions in the synthesis of VLCFAs. oni1 produces very small shoots, lethal seedlings, and an aberrant outermost epidermal cell layer due to reducing the amount of VLCFAs and alcohols components in wax (Ito et al. 2011). ONI2 encodes another rice KCS family gene, the oni2 mutant has a reduced amount of VLCFAs and is characterized by growth cessation after germination, fused leaves, and small shoots (Tsuda et al. 2013). Both studies on ONI1 and ONI2 suggest that VLCFAs play an important role for normal development in rice. In our study, the wsl4 mutants also exhibit significant reduced wax load comparing to the WT.

In Arabidopsis, the KCS6/CUTI/CER6 gene encodes a VLCFA condensing enzyme which localizes in the epidermis cells and determines long-chain lipid content on the pollen and stems surface (Millar et al. 1999; Fiebig et al. 2000), especially for the production of C28 VLCFAs (Millar et al. 1999; Haslam et al. 2012). One WSL4 homolog in barley is HvCUT1.3 (HvKCS6). Its mutant emr1 has significantly fewer aliphatic wax constituents longer than C24 (Weidenbach, et al. 2014). In cotton, heterologous expression of the Arabidopsis protein homolog encoding gene GhCER6 in yeasts could rescue growth defect phenotypes of both elo3 deletion mutant and elo2Δelo3Δ double mutant, accompanied by detectable increase in the amount of VLCFAs products ranged from C20 to C26, suggesting GhCER6 is a functional β-ketoacyl-CoA synthase (Qin et al. 2007). lecer6 of tomato exhibits up to three- to eightfold water loss comparing to the wild type, due to the C28 VLCFAs decrease in fruit cuticular waxes constitutions (Leide et al. 2007). In potato, RNAi lines of StKCS6, all VLCFAs compounds with chain length shorter than C26 are accumulated. On the contrary, the components with chain length longer than C28 are significantly reduced. These indicated that StKCS6 is essential for the formation of lipid monomers with chain lengths longer than C28 in both suberin and wax biosynthesis (Serra et al. 2009). Because different KCS genes have different substrate specificities, loss function of one KCS gene may interrupt the fatty acid elongation and then effect wax load. In the present study, we found that WSL4 encodes a rice KCS6 homologous protein, and the wsl4 mutant displays significant reduced wax loads on leaf surface, consistent with the function of other KCSs in previous studies.

Both the WSL4-RNAi transgenic lines and WSL4 knocked-out mutants could mimic the phenotypes of wsl4 in cuticular wax decrease, confirming that WSL4 is indispensable in rice cuticular wax biosynthesis. The qRT-PCR and GUS assay indicated that WSL4 is expressed ubiquitously, similar to expression pattern of other two KCSs, viz. KCS1 and KCS9 in Arabidopsis (Todd et al. 1999; Kim et al. 2013). Transient expression of the WSL4-GFP fusion protein in tobacco revealed a reticulate network subcellular localization, corresponding to the FAE subunits localization in the ER network (Joubès et al. 2008).

In conclusion, this study demonstrated that WSL4 encoding a KCS6 protein is involved in rice leaf cuticular wax formation. The biochemical function of WSL4 in the synthesis of rice suberins, sphingolipids and phospholipids deserves further study.

Author contribution statement

LG, SSZ, ZCZ, and JMW designed the research. LG, SSZ, ZCZ, LLL, XLW and ZZ performed the research. XZ, JW, JLW and XPG managed the rice transformation. LG, SSZ, ZCZ, LLL and JMW wrote the paper.