Seeds of the rice (Oryza sativa) genotype Xiushui 11, a japonica variety, were sown in a plastic box containing rice nutrient solution (Yoshida et al., 1976) and maintained in a controlled climate room at 26 ± 2°C, 12 h light phase and 80% relative humidity. Arabidopsis (Col-0) seeds were obtained from Arabidopsis Biological Resource Center at Ohio State University. Seeds were surface sterilized with 10% bleach and sown onto half-strength MS medium. To induce germination, seeds were placed at 4°C in the dark for 3 days under high humidity. Twelve to fifteen-day-old seedlings were transferred to pots containing soil mix. Plants were cultivated in a growth chamber under a 10-h photo period/14-h dark at 22°C for vegetative stage and 16-h photoperiod/8-h dark for reproductive stage. Plants were watered twice a week with Hogland's nutrient solution but, no pesticide was applied.

Colonies of the brown planthopper (Nilaparvata lugens) were maintained on Taichung Native 1 (TN1, an indica variety without any resistant gene to herbivores and pathogens) rice seedlings in a controlled climate room at 26 ± 2°C, 12 h light phase and 80% relative humidity. The green peach aphids (GPA, M. persicae) were reared in cages on the Chinese cabbage (Brassica rapa L.). Aphid rearing was maintained in an acclimatized room with a relative humidity of 60–70%, a temperature of 20 ± 2°C and an 18:6 L: D photoperiod. For all experiments, only apterous aphids were used.

The coding sequence of the OsLCB2a1 gene was amplified from pMD18TOsLCB2a1 using OsLCB2a1-ORF-F (GCTCTAGAATGGTGAGGCTGCCCTAC) (XbaI site underlined) and OsLCB2a1-ORF-R (GGGGTACCTTGAAGCTCTTCAGTTTCTCA) (KpnI site underlined). After digestion with XbaI/KpnI, the coding sequence of the OsLCB2a1 was ligated into plant binary vector pCAMBIA1300 under the cauliflower mosaic virus 35 S promoter in sense orientation. The resulting recombinant plasmid, designated as pCAMBIA-OsLCB2a1, was introduced into Agrobacterium tumefaciens strain GV1301. A. tumefaciens mediated transformation was used to introduce the transgene OsLCB2a1 into Columbia-0 (Col-0) by floral deep method (Zhang et al., 2006). Two independent transformants in Col-0, referred to as Line OE1 and Line OE2 were used in further experiments. Transformed progenies were screened on half-strength MS medium supplemented with 0.25 mg/L hygromycin. Homozygous transgenic lines of T3 generation were used for the experiments.

Total RNA from leaf or seedling samples was extracted using extraction kit (Tiangen, Shanghai, China) and reversely transcribed to cDNA using a PrimeScript RT reagent Kit according to the manufacture's instruction (Takara). Gene-specific primers were designed with Primer-3-Plus software. qPCR was performed in a total volume of 50 μl containing 4 μl cDNA, 1.5 μl of each gene-specific primer (10 mM), 18 μl ddH₂O and 25 μl SYBR Green Super mix Reagent (BioRad). qPCR was run on an ABI 7500 Real-Time PCR System for one cycle of 95°C for 3 min followed by 40 cycles of 95°C for 10 s, and 60°C for 30 s. Normalized gene expression was calculated by the 2^−ΔΔCT method (Livak and Schmittgen, 2001). Primers used in this study are listed in Table S2.

To determine the T-DNA border sequence and insertion position of OsLCB2a1 gene in Arabidopsis over-expressed plants TAIL PCR (Liu and Whittier, 1995) was done. We used three nested T-DNA border region specific primers LBSP1, LBSP2, LBSP3 together with mixture of three arbitrary degenerate primers AD1, AD2 and AD3. Primers sequences are shown in Table S2.

LCB2 homolog sequences were obtained by BLASTp (http://blast.ncbi.nlm.nih.gov), and Phytozome 10.3 (http://phytozome.jgi.doe.gov) and representative sequences were aligned by CLUSTALW using GENETYX program (Software Development Co., Ltd., Tokyo, Japan). The aligned sequences were used for phylogenetic analysis by Maximum Likelihood (ML) method by MEGA 5.0 software. Accession numbers of different organisms used in this tree contraction in the NCBI and Phytozome Genbank were given in Table S1.

Measurement of sphingolipids was performed as previously described by Markham and Jaworski (2007) with minor modifications. Briefly, about 600 mg of shoots from 4-week-old WT and over-expressed plants were homogenized. The internal standards (C17baseD-erythro-sphingosine and C₁₂-Ceramide) were added and extracted with the isopropanol/hexane/water (55:20:25 v/v/v). After incubation at 60°C for 15 min, the supernatants were dried by nitrogen. The dried extracts were de-esterified in 33% methylamine in ethanol/water (7: 3 v/v) for 1 h of incubation at 50°C. The samples were dried under nitrogen, then dissolved in 1 ml of methanol and analyzed by the Shimadzu UFLCXR/Triple-TOF 5600 LC/MS system using an Agilent Eclipse XDB C8 column (50 92.1 mm, 1.8 l m). Peak score responding to the target analytics and internal standards were collected and processed using the Agilent Masshunter Quantitative Analysis software. The components of sphingolipids were determined as described previously (Markham and Jaworski, 2007). Three biological repeats were performed.

No-choice aphid tests were performed with 15 biological replicates per genotype. Synchronized 1-day-old nymphs were used to infest 4-week-old plants with one nymph per plant. Nymphs were transferred to the plants using a fine camel hair brush. The total number of aphids was counted 14 days after infestation.

The EPG technique (Tjallingii, 1988; Reese et al., 2000) was employed to monitor penetrating and feeding behavior of aphids on over-expressed and WT plants. A gold wire (diameter 20 mm) was attached onto the dorsum of young adult aphids using conductive water-based silver glue. The wired aphid was placed on a 4-week-old plant that was connected to a recording system via a copper electrode in the soil (Tjallingii, 2006). The EPGs were recorded in a 22°C room with constant light for 8 h. At least 15 recordings of individual aphids (one aphid per plant) were obtained for each line. The EPG data were analyzed using the PROBE 3.0 software (Wageningen University, the Netherlands) to distinguish the various waveforms. Waveform C represents the pathway phase, when the aphid stylet is penetrating through the leaf tissue; waveform E2 represents phloem sap ingestion; Waveform F is associated with derailed stylet mechanics or penetration difficulties; and waveform G indicates active uptake of water from the xylem elements.

Arabidopsis plants were grown in soil. After 4 weeks the plants were sprayed with 25 μM phytospingosine or methanol as methanol was used to prepare stock solution of phytospingosine. After 8 h rosette leaves were collected and immediately frozen in liquid nitrogen for sphingolipid measurement. Three biological repeats were performed.

Callose deposition in leaves of the transgenic Arabidopsis plant and WT was visualized as a violet color by staining with aniline blue. Leaves were infiltrated with 5 ml of a solution made of phenol, glycerol, lactic acid, water and 95% ethanol (1:1:1:1:2, v/v). Leaves in solution were incubated in a 65°C bath until they become transparent and then stained with aniline blue. The staining reaction was held in the dark for 4 h. Samples were observed by microscopy under ultraviolet field (Zhang et al., 2014).

To visualize the alteration in cuticular wax content, a Field emission scanning electron microscope was used to analyze the images of Arabidopsis leaves without any form of preparation (Zimmermann et al., 2007). The adaxial side of 4-week-old rosette leaves of OE and WT plants were collected and directly placed onto the stub.

Data were subjected to analysis of variance (ANOVA) and presented as the mean value for each treatment. Statistical analyses were performed using Data Processing System (DPS) statistical software package (Tang and Zhang, 2013; http://www.dpsw.cn; Zhu et al., 2015). A Tukey's test (P < 0.05) was performed to evaluate the treatment effects.

To investigate the possible function of sphingolipid in insect resistance response, we cloned a gene (Os01g70380) from O. sativa encoding the LCB2a subunit of SPT, designated as OsLCB2a1. The full-length cDNA of the OsLCB2a1 gene was amplified by RT-PCR. Sequence analysis showed that OsLCB2a1 contains a 1470 bp open reading frame and encodes a protein of 489 amino acids which shares high identity with LCB2a from other plants. To identify LCB2a homology, we used BLASTp to conduct sequence similarity searches using the sequence of A. thaliana (AtLCB2a, AT5G23670). CLUSTALW alignment of the amino acid sequence of OsLCB2a1 with LCB2a from various organisms showed the presence of a KBL like domain (Figure S1) and a conserved GTFTKSFG motif (amino acid positions 307–314) corresponding to the known PLP-binding site, which is commonly found in members of the α-oxoamine synthase subfamily, such as 5-aminolevulinic acid synthase, 2-amino-3-ketobutyrate CoA ligase, and 8 amino-7-oxononanoate synthase (Figure 1A; Hanada, 2003). Phylogenetic analysis showed that LCB2a can be divided into seven categories: dicot, monocot, non-flowering plant, algae, fungi, animal, and bacteria and that OsLCB2a1 belongs to monocot branch (Figure 1B). There were three copies of LCB2a in rice (Figure 1B). OsLCB2a1 has 90, 84, and 82% identity to Zea mays SPT2, N. banthamiana SPT, and A. thaliana LCB2a, respectively.

BPH is one of the most important rice pests in Asia with piercing and sucking mouthpart that are needlelike. They suck plant sap during their feeding on rice tissues. We performed the experiment on the effect of BPH infestation on OsLCB2a1 inducement. The experimental results revealed that OsLCB2a1 transcripts in rice seedling were increased with time, reaching a peak at 4 h after that declined at 8–24 h under BPH attack (Figure 2).

To examine OsLCB2a1 function, we transferred the OsLCB2a1 gene into WT Arabidopsis (Col-0) and generated the over-expression (OE) lines. RT-PCR result showed that the OsLCB2a1-OE plants but not the WT plants amplified a DNA band corresponding to the OsLCB2a1 trans-gene (Figure S2). Quantification of OsLCB2a1 expression through qRT-PCR in 4-week-old leaves of Arabidopsis plants revealed that the gene expression was increased by a 33-fold in the OE plant compared to the WT plant (Figure S3). The T-DNA border sequence and insertion position of transgenic plants were confirmed by TAIL PCR (Liu and Whittier, 1995). The insertion site located upstream 1 kb of AT5G56900 and downstream 2.7 kb of AT5G56890, without any annotated genes broken in the OE plant (Figure S4).

As SPT is the enzyme that catalyzes the first step for sphingolipid biosynthesis (Gable et al., 2000; Breslow and Weissman, 2010), we hypothesized that its overexpression could change the sphingolipid profile in Arabidopsis. Sphingolipids were extracted from leaves of 4-week-old WT and OE plants and quantitatively analyzed by LC-MS/MS. Phytosphingosine (t18:0), sphingosine (d18:1), sphinganine (d18:0), different ceramides and phytoceramides were identified by the retention time of the corresponding standards. As expected, LCBs and ceramides were increased in the OE plants compared with the WT (Figure 3). LCB composition of Arabidopsis leaves was similar to those reported previously for other Solanacea, such as tobacco and tomato (Sperling et al., 2005; Markham et al., 2006; Buré et al., 2014; Wu et al., 2015), in which trihydroxylated LCBs such as phytosphingosine was more abundant than the dihydroxylated LCBs sphinganine and sphingosine (Figure 3B). Phytoceramide (t18:0) was more abundant than ceramide (d18:1) and hydroxyceramide (d18:0), these findings were similar to Wu et al. (2015). Comparing ceramides with LCB moieties, we found a higher level of phytoceramides in OE plants than that in the WT plants (Figure 3C). Thus, the quantitative sphingolipid profiling showed that the OE plants accumulated more long-chain bases (1.3 nmolg⁻¹FW) and ceramides (6.0 nmolg⁻¹FW) compared to WT (LCBs 0.8 nmolg⁻¹FW and ceramide 4.6 nmolg⁻¹FW). These data suggested that overexpression of OsLCB2a1 in Arabidopsis increase the level of sphingolipids.

Plant callose is strongly related with plant resistance/tolerance to insect pests and pathogens (Jacobs et al., 2003; Hao et al., 2008; Zhang et al., 2014). Callose is recognizable in tissue sections through the formation of an intense yellow, UV light-induced fluorescence with the aniline blue fluorochrome (Stone et al., 2003). The experimental results revealed that the fluorescence in leaves of OE plant was stronger than that of WT (Figures 4A,B), suggesting that OsLCB2a1 over-expression may increase callose deposition to improve plant tolerance to stress. We also examined the expression patterns of callose synthase-encoding genes using qRT-PCR. Two callose synthase-encoding genes (GSL1, GSL5) were clearly up-regulated in OE Arabidopsis plants compared to WT (Figure 4C). Additionally we examined the leaf cuticular wax deposition, which play important roles in plant resistance to various biotic and abiotic stresses (Zhang et al., 2007; Wang et al., 2011; Ni et al., 2016). The experimental results revealed that the wax deposition was higher in leaf of OE than WT plants (Figure 5).

As OsLCB2a1 expression was up-regulated in rice plant under BPH attack, we inferred that OsLCB2a1 up regulation may improve plant resistance to aphids. To test this hypothesis, no choice assay was carried out in WT and OE Arabidopsis plants. The results showed that population size of the green peach aphids (M. persicae) in OE plants (Line OE1, Line OE2) was smaller than that in WT plant after 14 days (Figure 6). These results suggest that over-expression of OsLCB2a1 might improve plant resistance to aphid.

To reveal whether aphid feeding behavior was affected by the over-expression of OsLCB2a1 gene, we compared electrical penetration graph (EPG) recordings of M. persicae on WT or OE plants. The EPG parameters for our study are summarized in Table 1. All aphids started to penetrate the leaf around the same time on all tested plants, as indicated by the time to the first probe. Compared to the WT plants, M. persicae on overexpressed plant needed 1.6 times more time to the first probe to first phloem contact. Aphids on the OE plant spent significantly less time and took smaller number of time salivating into the ingesting phloem and sustain phloem sap than aphids on WT plant types (Table 1). Additionally, aphid showed a significantly longer duration of the non-probing phase on the OE than on the WT. Waveform F, associated with derailed stylet penetration, was also observed for a significantly longer time and in a larger number on the OE Arabidopsis plant (Table 1). On the other hand, the aphids spent less time taking up xylem sap from the OE plant as was indicated by a shorter time and smaller number of waveform G. Furthermore, M. persicae showed a significantly longer duration and larger number of potential drop phase on the OE plant than the WT plant.

Phytosphingosine (phyto-sph) is an intermediate in sphingolipid metabolism pathway (Figure S5) and this LCB implicated as a secondary messenger in vital signaling process in eukaryotic organism. Phyto-sph is abundant in fungi and plants, and also found in animals including humans. To find out whether exogenous insertion of sphingiod base can change the sphingolipid profile in Arabidopsis plant, we sprayed phyto-sph onto the rosette leaves of WT Arabidopsis plants and observed the sphingolipid profile. Results revealed that exogenous phyto-sph can change the sphingolipid profile in the WT plant. Figure 7 showed that the amount of three LCBs e.g., sphingosine (sph), dihydrosphingosine (Dhsph) and phyto-sph were increased significantly in the WT Arabidopsis (two-factor ANOVA, P < 0.05) (Figures 7A–C). The amount of C-18 Cer, C-24 Cer and phyto-24 Cer were also increased after treated with exogenous phyto-sph (Figures 7D–F).

To find out whether the change in sphingolipid profile due to the exogenous phyto-sph in Arabidopsis plants has any impact on aphid population growth we again carried out the no choice assay in phyto-sph treated and untreated plants and results showed that population size of the green peach aphids in phyto-sph treated WT plants were reduced after 14 days compared to the control plants (Figure 8). Taken together, these results indicated that exogenous phyto-sph could inhibit aphid infestation.

To investigate the signal transduction pathway involved in OsLCB2a1-mediated aphid resistance, we examined the expression levels of some defense-responsive genes, which are known to function in SA- and JA/ethylene-dependent pathways (Li et al., 2006; Zarate et al., 2007). Expression levels of the genes EDS1 (ENHANCED DISEASE SUSCEPTIBILITY1), PAD4 (PHYTOALEXIN DEFICIENT 4), NPR1 (NONEXPRESSOR OF PATHOGENESIS-RELATED GENES 1) related to SA-dependent pathway were remarkably higher in the OE plants than that in WT at 0, 4, 8, and 24 h under aphid attack (Figure 9). However, the transcripts of LOX2 (LIPOXYGENASE 2), VSP2 (VEGETATIVE STORAGE PROTEIN 2), EIN2 (ETHYLENE INSENSITIVE 2) and ERF1 (ETHYLENE RESPONSE FACTOR 1) on JA/ethylene-dependent pathway were significantly lower in OE lines than those in WT (Figure 9). The results suggest that OsLCB2a1 might mediate the transgenic resistance to aphids through SA-dependent signaling.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.