Identification and Classification of LEA Family Genes in Orchids and Characterization of Their Role in Callus Formation

The plant late embryogenesis abundant (LEA) proteins are abundant in seeds, play an important role in various abiotic stresses. However, there is still no information on genome-wide identification of LEA genes in orchids and their function in callus formation is almost unknown. In this study, the LEA genes from two orchids (Phalaenopsis equestris and Dendrobium officinale), were genome-wide identified, classified and characterized. A total of 57 and 59 LEA genes were identified in the genomes and these were divided into 8 and 9 groups for P. equestris and D. officinale, respectively. The LEA_1 and LEA_4 genes from P. equestris and D. officinale showed strong expression in seeds, but were significantly down-regulated in flowers and absent in vegetative organs (leaves, stems and roots). In addition, the LEA_1 and LEA_4 genes from D. officinale were abundant in the protocorm-like body (PLB) stage, while weak signals that were detected in in vitro shoots could not be detected in plantlets. The expression of these genes highlights PLBs in orchids are somatic embryos. The DoLEA36 from LEA_4 and DoLEA43 from LEA_1 were further characterized. The GFP signal of the DoLEA36-GFP fusion protein was only detected in the cytoplasm, while the GFP signal of the DoLEA43-GFP fusion protein was detected in both the cytoplasm and nucleus. This indicates that DoEA36 localizes in the cytoplasm while DoLEA43 localizes in both the cytoplasm and nucleus. Both DoLEA36 and DoLEA43 stimulated callus formation in transgenic Arabidopsis. The percentage of callus formation from 35S::DoLEA43 transgenic lines was higher than in wild type plants in two callus induction methods. Our results provide comprehensive information about the LEA gene family in orchids and genetic evidence for the involvement of LEA genes in the induction of callus, which may reveal their positive role in the maintenance of PLBs in orchids.


7
[25]. The size of exons/introns can affect splice-site recognition [26]. Hence, the composition of exons/introns of LEA genes in P. equestris and D. officinale was explored.
Interestingly, the majority of LEA genes from P . equestris and D. officinale have conserved intron/exon structures. For example, 64.9% (37 of 57) of PeLEAs and 62.7% (37 of 59) of DoLEAs contain only one exon (Figure 2). This suggests that the structure of the LEA gene in orchids is conserved and may have resulted in functional conservation during their evolution.

Expression analysis of PeLEAs and DoLEAs in different tissues
To characterize the LEA genes, the expression pattern of LEA genes among vegetative organs (roots, stems and leaves) and reproductive organs (flowers and seeds) were surveyed by comparing the FPKM values for each gene in P. equestris and semiquantitative RT-PCR in D. officinale. The NG1-3, PvLEA18, LEA_2 and LEA_3 groups in both orchids did not display specific expression patterns ( Figure 3). DoLEA12, -18, -23, -31 and DoLEA53 from NG1-3 groups were widely expressed in roots, stems, leaves, flowers, and seeds ( Figure 3). Only one PvLEA18 gene was found in both P. equestris and D. officinale, although they shared different expression patterns. PeLEA43 from the PvLEA18 group in P. equestris was strongly expressed in seeds and flowers, but D. officinale DoLEA7 from the PvLEA18 group was absent in flowers ( Figure 3). In contrast, the LEA_1 group members in both orchids displayed a seed-specific expression pattern. All seven LEA_1 group genes (PeLEA1, -23, -24, -50, -6, -7 and -8) in P. equestris, and all LEA_1 group genes from D. officinale (except for DoLEA3), showed a strong expression pattern in seeds (Figure 3). In addition, LEA_1 group members such as PeLEA6, DoLEA43 and DoLEA48 were also strongly expressed in flowers (Figure 3). Two LEA_2 group genes (PeLEA4 and PeLEA47) were found in P. equestris, PeLEA7 displayed no specific expression pattern in all the detected tissues, while PeLEA4 showed higher expression in seeds and lower expression in flowers, similar to the LEA_1 group genes (Figure 3). DoLEA13 , -17, -36, -4 and -5 from the D. officinale LEA_4 group were highly expressed in seeds, but expression was absent in all vegetative organs (Figure 3). LEA_1 and LEA_4 displayed a similar expression pattern, suggesting that they might have a similar function, whereas the other groups had diverse expression patterns.
Expression of LEA genes from D. officinale at different developmental stages PLBs are a specific developmental stage in orchids, and are regarded as somatic embryos as a result of their similar macroscopic and microscopic features [27]. We used four developmental stage samples namely PLBs, multiple shoots from PLBs, and two developmental stage plantlets of D. officinale (see details in the materials and methods section), to survey the changes of LEA genes during development. Very interestingly, most genes in LEA_1 and LEA_4 groups showed high expression at the PLB stage (T1), while their expression became lower as PLBs developed to plantlets in the two detection methods ( Figure 4). For example, DoLEA4, -13, -17, -36 and -45 from the LEA_4 group, and DoLEA9, -11, -27, -43 and -48 from LEA_1 were only strongly detected at the PLB stage but were significantly down-regulated or absent at the other three developmental stages namely T2, T3 and T4 (Figure 4). The LEA_2 group has a close phylogenetic relationship with the LEA_1 and LEA_4 groups, and most genes in this clade also displayed a similar expression pattern with genes from LEA_1 and LEA_4 groups. For instance, DoLEA2, -19 and DoLEA55 from the LEA_2 group were highly expressed at the PLB stage after semiquantitative RT-PCR analysis ( Figure 4). However, the genes from NG1-3 groups displayed diverse expression patterns in the four developmental stages.

The localization of DoLEA36 and DoLEA43
To further analysis the function of the LEA_1 and LEA_4 group genes, the DoLEA36 gene from LEA_4 and DoLEA43 from LEA_1 were selected to explore their localization and further function analysis based on their expression patterns. The empty vector of pCAMBIA1302 transformed into Arabidopsis plants was used as the positive control. The green fluorescent signals of the seedlings harboring an empty vector observed in the cytoplasm, plasma membrane and nucleus were strong ( Figure 5A). In 35S::DoLEA36-GFP seedlings, green fluorescence filled cells, which suggests that the DoLEA36-GFP fusion protein is located in the cytoplasm ( Figure 5A). In contrast, the 35S::DoLEA43-GFP lines showed green fluorescent signals in the cytoplasm and nucleus ( Figure 5A). To further verify the localization of the two proteins, the cytoplasm proteins and nucleus proteins were isolated from both 35S::DoLEA36-GFP and 35S::DoLEA43-GFP transgenic plants.
Western blot analysis showed that the GFP of 35S::DoLEA36-GFP plants was only detected in the cytoplasm ( Figure 5B). However, the GFP of 35S::DoLEA43-GFP plants was detected in the cytoplasm and nucleus ( Figure 5B). This result indicates that DoLEA36 localized in the cytoplasm while DoLEA43 localized in the cytoplasm and nucleus. DoLEA36 and DoLEA43 had different localization although they shared a similar expression pattern.
DoLEA36 and DoLEA43 have different roles in stimulating callus formation Both DoLEA36 and DoLEA43 were strongly detected in seeds and PLBs of D. officinale, but displayed different localization. Hence, the callus formation rate in the transgenic lines of DoLEA36 and DoLEA43 driven by a 35S promoter were analyzed. The RT-PCR result ( Figure   6A) and the western blot analysis ( Figure 6B) revealed that both DoLEA36 and DoLEA43 genes were transcribed and expressed successfully in transgenic lines. The hypocotyl without a cotyledon and roots was planted to plant growth regulators (PGR)-free half-strength MS medium and cultured in the dark for two days, then transferred to a 16-h photoperiod. Callus formed at the wounding site in a small number of explants. The WT plants and transgenic lines displayed a different callus induction rate ( Figure 6C-F). The callus induction rate of 35S ::DoLEA36 lines was about two-fold higher than WT plants ( Figure 6C and D) while the callus induction rate of 35S::DoLEA43 lines was three-fold higher than WT plants 14 d after wounding ( Figure 6E and F). The callus initiation of all transgenic lines exceeded 15%, while that of WT plants was only about 6% ( Figure 6).
To further explore callus induction, 6 d-old seedlings after stratification were dissected and the upper end of hypocotyls were removed, while the hypocotyl containing roots were incubated on PGR-free half-strength MS medium in the dark. The 35S::DoLEA36 lines and WT did not have statistically significantly different callus induction 12 d after wounding ( Figure 7A and B). In contrast, the callus induction rate of all 35S::DoLEA43 lines was higher than WT plants 12 d after wounding ( Figure 7C). The 35S::DoLEA43 lines showed increased callus initiation compared with WT plants ( Figure 7D). These results suggest that DoLEA36 and DoLEA43 play a role in callus formation, while DoLEA43 is more effective in callus induction than DoLEA36 .
The expression patterns of DoLEA36 and DoLEA43 under wounding stress Increasing data proved that wounding triggers callus formation [28,29]. Hence, the expression of DoLEA36 and DoLEA43, which played a role in callus formation, was explored under wounding stress. Total RNA at 0, 2, 5, 10 and 25 h after wounding were applied to qRT-PCR analysis. As expected, DoLEA36 and DoLEA43 genes were up-regulated in these time points ( Figure 8A). DoLEA36 displayed an increasing trend at 2, 5, 10 and 25 h after wounding, while DoLEA43 initially increased within 10 h then maintained its transcriptional level ( Figure 8A). Both DoLEA36 and DoLEA43 showed about 3.5-fold higher expression 2 h after wounding, while became about 15-and 52-fold higher at 10 h after wounding, respectively ( Figure 8A). This indicates that both genes were modulated by wounding.
No wounding responsive element (WUN-motif, AAATTACTA) was found in the putative promoter of DoLEEA36, while one WUN-motif was present at the -453 bp site of the DoLEA43 promoter ( Figure 8B). In addition, two WUSATAg motifs (TTAATGG) were found at -1014 and -1440 bp sites of the DoLEA43 promoter ( Figure 8B). These results thus establish that both DoLEA36 and DoLEA43 could be induced by wounding and might help explants generate callus.  Figure 8). This is similar to the expression pattern of a LEA gene from Papaver somniferum that displayed 4-fold higher expression than the control after wounding treatment for 5 h in the whole seedlings [44]. In addition, one WUN-motif was found in the promoter of DoLEA43, but none in the DoLEA36 promoter ( Figure 8B). This suggests that the regulation of expression in wounding by transcription factors might differ for the two genes. Two WUSATAg motifs were found in the DoLEA43 promoter, but no WUSATAg motif

Conclusions
In this study, a total of 57 and 59 the LEA genes from P. equestris and D. officinale were identified at the genome-wide scale, and these could be divided in 8 and 9 groups, respectively. The expression of LEA in different organs and different developmental stages were explored and showed a conserved expression pattern in some groups. Genes from the LEA_1 and LEA_4 groups were abundantly expressed in seeds and PLBs. Two LEA genes, DoLEA36 and DoLEA43, shared similar expression patterns but different localization. These two genes stimulate callus formation but displayed different roles in callus induction in transgenic Arabidopsis. Our study provides insight into the mechanism of maintenance of PLBs in orchids and plant regeneration.

Subcellular localization
The transgenic seeds of 35S::DoLEA36 and 35S::DoLEA43 were surface-sterilized in 1% NaClO for 10 min, washed in sterile distilled water six times, then sown on half-strength MS medium in Petri dishes containing 1.5% sucrose and 0.8% agar (pH 5.7). Plates were incubated at 4 °C in the dark for 2 d for stratification, then cultured in a growth chamber.
After stratification, 4 day-old seedlings were used to survey GFP fluorescence and

Callus induction
In this study, we explored callus formation via two callus induction methods. In the first method, hypocotyls from 6-day-old Arabidopsis seedlings after stratification were cut by disposable knives about 0.5 mm from the cotyledon-hypocotyl junction and at about 0.5 mm from the hypocotyl-root junction. The center of hypocotyls was transferred carefully to Petri dishes containing half-strength MS medium with 1.5% sucrose and 0.8% agar (pH 5.7) with a sterile toothpick, incubated for 2 days in the dark and grown in a growth chamber. In the second method, hypocotyls were cut once at about 2 mm from the hypocotyl-root junction, the hypocotyl-root explants were transferred to Petri dishes containing half-strength MS medium with 1.5% sucrose and 0.8% agar (pH 5.7) and incubated in the dark at 22 °C. Callus was observed at the end of the cut site. Callus on an explant was considered to be induced if the callus was visible on a Leica S8 APO stereomicroscope (Leica Microsystems Ltd., Heerbrugg, Switzerland). Callus induction was quantified as a percentage of explants forming callus. Every experiment was repeated in triplicate. At least 50 seedlings of WT and transgenic lines from each experiment were used.

Wounding treatment
PLBs growing on half-strength MS medium containing 2.0% sucrose, 0.5 mg/L NAA and 0.6% agar (pH 5.4) were used for the wounding treatment. The PLBs were sliced into 2 mm thick slices and placed on the same medium. Explants at 0, 2, 5, 10, and 25 h after wounding were harvested and total RNA was extracted using the method described above.
DoLEA36 and DoLEA43 expression was detected by qPCR analysis. Three biological replicates were performed for each sample.
Prediction of cis-responsive element in DoLEA36 and DoLEA43 promoters To further explore the gene-responsive factors, 2000 bp upstream of the initiation codon from DoLEA36 and DoLEA43 were obtained from the D. officinale genome and used to predict the cis-responsive elements by PLANTCARE [59] and PLACE [60].

Statistical analyses
All the data of induction rate was analyzed by SigmaPlot12.5 software (Systat Software Inc., San Jose, CA, USA) using one-way analysis of variance (ANOVA) followed by Dunnett's test. P < 0.05 was considered to be statistically significant.

49.
Murashige T, Skoog F: A revised medium for rapid growth and bio-assays with      Representative images of callus initiation in WT and 35S::DoLEA43-GFP lines.
Scale bar = 0.05 cm. The hypocotyls were cute once at about 2 mm from the hypocotyl-root junction, the hypocotyl-root explants were used to study the callus induction. Bars represent mean ± SD of three biological replicates. ** indicate P < 0.001 between WT and transgenic lines, respectively.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.
Supplementary materials.pdf