Brassica rapa var. Japonica, which shows multiple shoot branching during the vegetative stage, was used for BrLAS gene isolation and expression pattern analysis in different tissues and under various treatments. Seeds were germinated in plastic petri dishes containing two layers of filter paper saturated with distilled water at 23°C for 2 days in the dark then transferred to pots containing soil growth medium under artificial growth conditions at 23–25°C, a photoperiod of 16/8 h light/dark and relative humidity of 60 %. Hoagland nutrient solution was added once a week. Roots, stems, leaves, axillary buds and shoot tips were sampled from 4-week-old plants. Flowers were harvested after vernalization and shoot apical meristems (SAMs) were sampled every 7 days during shoot branching development (from 14 to 42 days after germination). Five well-grown plants were thoroughly mixed to form one sample. Three biological replicates of each tissue and organ were sampled.

For hormone treatment, 2-week-old seedlings were transferred to 4 L hydroponic containers containing half-strength Murashige and Skoog (1/2 MS) liquid medium (pH = 5.8) renewed once every 3 days. At the five-leaf-stage, seedlings were exposed to 1/2 MS solution with 100 μM ABA. For stress treatment, 4-week-old plants grown in pots were irrigated with 200 mM NaCl, 15 % (w/v) polyethylene glycol (PEG) and 100 mM hydrogen peroxide (H₂O₂), respectively. Leaves were sampled 0.5, 1, 2, 4, 8, 12, and 24 h after treatment and seedlings without treatment were used as a control. In addition, the leaves of seedlings grown under solution without any treatment were also sampled at the designated time (8:00 a.m. to 8:00 a.m. the next day). Five well-grown plants were thoroughly mixed to form one sample. Three biological replicates of each tissue and organ were sampled for RNA extraction.

A. thaliana ecotype Columbia (Col-0) and BrLAS-overexpressing Arabidopsis plants were used in this study. Seeds of transformed Arabidopsis were surface sterilized with 1% sodium hypochlorite and selected on 1/2 MS medium containing appropriate antibiotics. Homozygous lines of T₃ generations were used for further analysis.

To clone BrLAS from B. rapa, a specific pair of primers, TL-F and TL-R, were designed based on sequence information in the Brassica genome database (BRAD¹) (Wang et al., 2011) and sequence information of Arabidopsis LAS. The full-length cDNA of BrLAS was obtained via PCR using B. rapa total RNA, specific primers and Tks Gflex DNA Polymerase. The PCR products were then cloned into the pMD 18-T Vector (Takara, Japan) for sequencing.

BrLAS homologous genes were obtained by BLAST searches of the National Center for Biotechnology Information² and Brassica database (BRAD, see footnote 1) using the DNA or protein sequences of BrLAS as queries. BrLAS-related proteins were aligned using DNAMAN software and phylogenetic relationships were analyzed using MEGA 6.0 by the neighbor-joining method with 1,000 bootstrap replicates.

Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s instructions. First-strand cDNA was synthesized from total RNA using a PrimeScript^®RT reagent Kit (Takara). The real-time PCR reaction was performed using a Roche thermocycler (LightCycler 480, Roche, Germany). Each reverse transcription-PCR procedure was carried out with three biological and three technical replicates. The Chinese cabbage GAPDH gene was used as an internal control to normalize expression of the BrLAS gene and the relative quantification method was used to evaluate quantitative variation. ACTIN2 (At3g18780) was used as an internal control to normalize expression levels of abiotic stress-responsive genes in transgenic Arabidopsis plants.

To construct the overexpression vector 35S: BrLAS, the entire 1,332-bp coding sequence of BrLAS was amplified and subcloned into the pCR8/GW/TOPO entry vector (Invitrogen) according to the manufacturer’s instructions. The CDS fragment was then inserted into the pMDC32 Gateway-compatible binary vector through LR recombination reactions (Curtis and Grossniklaus, 2003). The 35S: BrLAS plasmid was then introduced into Agrobacterium tumefaciens strain GV3101 and transformed into Arabidopsis Col-0 plants using the floral dip method (Clough and Bent, 1998).

To observe subcellular localization of the BrLAS protein, a p35S: BrLAS-GFP vector containing the open reading frame (ORF) of BrLAS (without the termination codon) fused in-frame with green fluorescent protein (GFP) and driven by the 35S promoter was constructed. The p35S: BrLAS-GFP and 35S: GFP control plasmids were then introduced into GV3101 and the transformed Agrobacterium were infiltrated into leaf cells of Nicotiana benthamiana as described previously (Sparkes et al., 2006). GFP signals in leaf epidermal cells were subsequently observed under a Leica fluorescence microscope (Leica Microsystem, Heidelberg, Germany).

To determine the chlorophyll content, 16-day-old seedlings of wild-type (WT) and transgenic Arabidopsis grown on 1/2 MS medium were collected in centrifuge tubes and weighed then immersed in 1 mL pre-chilled 80% acetone and fully ground. After incubation at 4°C for 4 h and centrifugation at 5,000 × g for 10 min at 4°C, the supernatants were transferred to a 10 mL graduated tube and diluted to 3 mL with 80% acetone. A spectrophotometer was then used to measure absorbance at 645 and 663 nm, and chlorophyll a, chlorophyll b, and total chlorophyll contents were calculated as follows:

Arabidopsis seeds of WT (Col-0) and BrLAS-overexpressing lines were surface sterilized with 1% sodium hypochlorite and kept at 4°C in the dark for 72 h. To conduct a germination rate assay showing the response to ABA, sterilized seeds of WT and transgenic lines were sown on 1/2 MS plates supplemented with 0 or 2 μM ABA and grown at 22°C in a growth chamber with a photoperiod of 16/8 h light/dark. Germination rates (seedling with cotyledon) were then examined after 6-day growth. For the root growth assay, seeds of transgenic and WT lines were firstly sown on 1/2 MS plates then grown at 22°C in a growth chamber with a photoperiod of 16/8 h light/dark for 5 days. Five-day-old seedlings were then transferred to fresh 1/2 MS medium supplemented with 0 or 10 μM ABA. Seedling root lengths were then measured after 8 days growth on vertical plates.

To examine drought tolerance, transgenic and WT Arabidopsis plants were germinated on 1/2 MS plates under a 16/8 h light/dark cycle at 22°C. Seven-day-old seedlings were then transferred to pots containing an identical soil mix (1:1 vermiculite:humus) under normal growth conditions. Twenty-five-day-old seedlings were then subjected to water stress by withholding watering until distinct phenotypic differences between transgenic and WT plants were observed. Plants were then re-watered and survival rates surveyed after re-watering for 3 days. To measure ABA content, leaves of WT and transgenic plants at an identical rosette stage were collected under normal and drought stress, and the procedure for extraction, purification and quantification of ABA was performed as described by Cornish and Zeevaart (1984) with some modifications. Each experiment was performed in triplicate.

For water loss measurements, leaves of transgenic and WT plants at an identical rosette stage were removed and weighed immediately. The samples were then weighed at designated time intervals under normal light conditions. Leaves from five plants were mixed to form one sample. Three biological replicates were obtained for each transgenic line. Water loss was calculated as follows: (Leaf initial fresh weight - leaf weight at each time point)/leaf initial fresh weight.

For SEM, intact fully expanded rosette leaves of WT and transgenic plants were collected in parallel then freshly excised into thin pieces 3 mm wide and immediately fixed in 2.5% (v/v) glutaraldehyde in PBS buffer at 4°C overnight. After washing in the same phosphate buffer, the samples were post-fixed with 1% OsO₄ and dehydrated with a graded ethanol series. After trough-critical point drying, the samples were mounted and sputter-coated with gold then observed under a SEM (HITACHI S-3400N).

To examine H₂O₂ accumulation, 4-week-old rosette leaves of WT and transgenic plants under drought stress were collected and immersed in freshly prepared 1 mg mL^-1 DAB solution for 12 h in the dark. Stained leaves were then discolored with a bleaching solution (ethanol/acetic acid/glycerol = 3:1:1) and photographed.

To detect the changes in physiological indicators under drought stress, leaves of WT and T₃ transgenic plants at an identical rosette stage were harvested. For extraction of H₂O₂, O^2-, superoxide dismutase (SOD), peroxidase (POD) and chloramphenicol acetyltransferase (CAT), 0.2 g samples were placed in 1.8 mL 0.1 M PBS (pH 7.4) on ice. After centrifugation at 10,000 × g for 10 min at 4°C, the supernatants were then used to measure H₂O₂ and O^2- contents, and SOD, POD, and CAT enzyme activity using corresponding detection kits. MDA assay was performed using the thiobarbituric acid method (TBA) (Heath and Packer, 1968).

All primers used in this study are listed in Supplementary Table S1.

All experiments included three independent repeats, and significant differences analyses were performed using the t-test at significance levels of ^∗P < 0.05 and ^∗∗P < 0.01.

The full-length cDNA sequence of BrLAS was 1,866-bp long and contained a 1,332-bp ORF with a 5′-untranslated region (UTR) of 151-bp and a 3′-UTR of 383-bp. It encoded a protein of 443 amino acid residues with a predicted molecular mass of 49.278 kDa. Amino acid sequence alignment analysis revealed that BrLAS is highly homologous to GRAS genes in other species except for its two progenitor species (BnLAS in B. napus and BoLAS in B. oleracea). For example, Arabidopsis AtLAS shared 89% homology, tomato LeLs shared 81% homology and rice OsMOC1 shared 71% homology (Figure 1A). The deduced amino acid sequence of BrLAS contains a typical GRAS domain consisting of the LHRI, VHIID, LHRII, PFYRE, and SAW motifs in its C-terminus (Bolle, 2004) (Figure 1A). Here, phylogenetic analysis indicated that BrLAS groups with the AtLAS branch contained LeLS and OsMOC1 (Figure 1B). These results strongly suggest that BrLAS is a member of the GRAS gene family in B. rapa and an ortholog of LAS in Arabidopsis.

To examine the potential functions of the BrLAS gene, quantitative real-time analysis was performed to investigate expression patterns in the roots, stems, leaves, flowers, axillary buds, and shoot tips. As shown in Figure 2A, BrLAS showed relatively high levels of expression in the roots, axillary buds, and flowers. Since relatively higher expression levels were observed in the axillary buds, expression levels of BrLAS at five stages of axillary bud development (14–42 days after germination) were examined further (Figure 2B). The results revealed that transcript levels of BrLAS underwent a dramatic change during axillary bus development, suggesting a role in the formation of shoot branching.

To examine the detailed expression profiles of BrLAS under various abiotic stresses and phytohormone treatment, 4-week-old B. rapa seedlings were treated with exogenous 100 μM ABA, and NaCl, PEG, and H₂O₂ treatment. Expression levels of BrLAS were significantly induced by ABA, reaching a maximum at 8 h after treatment and up-regulated more than 20-fold compared with the control, and then decreasing sharply (Figure 2C). BrLAS expression was also induced by NaCl, reaching a peak of more than 15-fold at 8 h after treatment (Figure 2D). Similarly, expression was markedly induced by PEG, reaching a peak of more than 15-fold at 1 h after treatment then declining gradually (Figure 2E). Under H₂O₂ treatment, expression reached a peak of more than 10-fold at 4 h after treatment then decreased (Figure 2F). Moreover, the expression of BrLAS under normal conditions using solution without any treatment was also examined, and the result showed that the expression of BrLAS was affected by mechanical stress, reaching a peak of approximately fivefold at 0.5 h after tissues cut then increasing gradually, and showed the circadian rhythm, the expression levels of which at 12 and 16 h (8:00 p.m. and 00:00 a.m.) were higher than other time points during the day (Supplementary Figure S1). These data suggest that expression of BrLAS is sensitive to NaCl, PEG, H₂O₂ and exogenous ABA treatment, affected by mechanical stress and showed circadian rhythm.

To determine subcellular localization of the BrLAS protein, we constructed a p35S: BrLAS-GFP construct containing the ORF of BrLAS (without the termination codon) fused in-frame with GFP driven by the 35S promoter and transformed it into B. rapa protoplast cells by means of modified PEG treatment (Yoo et al., 2007). As shown in Figure 3, fluorescent signals of 35S: GFP were uniformly distributed throughout the cells, while BrLAS-GFP was exclusively localized in the nucleus. Consistent with the subcellular localization of other members of the GRAS gene family and possible function as a transcription factor, these findings suggest that BrLAS is a nuclear protein.

To determine the potential function of BrLAS, we constructed a 35S: BrLAS construct to transform Arabidopsis Col-0 plants. Approximately 40 transgenic plants were identified through hygromycin B resistance analysis and PCR verification with primers of BrLAS and a selection marker gene (hygromycin, HYG). Phenotypes of T₁ transgenic plants were then observed and representative individuals were propagated for further experiments. Compared with WT Arabidopsis, the morphology and developmental characteristics of the overexpression lines showed obvious changes (Figure 4 and Supplementary Table S2). During the same developmental period, BrLAS-overexpressing plants grew at a slower rate and had fewer leaves, which were darker in color than the WT plants (Figures 4A,B). Some overexpressing lines exhibited delayed bolting and flowering, to varying degrees, at the reproductive stage (Figures 4C,H), with fewer senesced rosette leaves. All changes were stably inherited by the following generations; therefore, three representative T₃ transgenic lines (OE6, OE8, and OE9) were chosen for further quantitative analysis. Rosette leaf numbers in the three overexpressing lines were significantly reduced compared to the WT plants before bolting (Figure 4F). Moreover, plant height at day 65 was markedly reduced (Figure 4G). The bolting times of all three lines were delayed by about 14 days compared with the WT (Figure 4H), and fertility was also reduced, resulting in fewer fertile siliques on the main stem (Figure 4I).

Since the majority of BrLAS-overexpressing plants grown both on plates and in soil showed smaller leaves with a darker green color than the WT, we also examined cell morphology and chlorophyll content. SEM showed that the leaf epidermis cells were significantly smaller in the transgenic lines than the WT (Figures 4D,E), suggesting that this was the cause of the smaller leaves in BrLAS-overexpressing plants. The total chlorophyll content of the three transgenic lines increased by 88.1, 74.1, and 80.1%, respectively, with increases in both chlorophyll a and chlorophyll b contents (Figure 4L), suggesting that this was the major cause of the darker green leaves in the BrLAS-overexpressing lines. Since the AtLAS gene in Arabidopsis is required for the initiation of axillary meristems, we also examined branching numbers. As shown in Figures 4J,K, numbers of rosette branches and cauline branches both decreased in the 35S: BrLAS transgenic plants compared with the WT.

Since expression of BrLAS was up-regulated with ABA, PEG, and H₂O₂ treatment (Figure 2), and considering the altered leaf morphology and delayed leaf senescence in the 35S: BrLAS transgenic plants, it is plausible that BrLAS is involved in drought stress tolerance in BrLAS-overexpressing Arabidopsis plants. To verify this hypothesis, 3-week-old transgenic and WT seedlings were subjected to drought stress. Survival rates were then surveyed after re-watering for 3 days. As shown in Figure 5A, the leaves of WT plants displayed severe wilting and yellowing, while the overexpression lines maintained good growth after drought stress. Moreover, only 30% of WT plants were able to recover from stress, while the survival rates of the OE6 and OE8 transgenic plants were 86 and 93%, respectively (Figure 5E). Consistent with these findings, water loss rates of detached leaves from the overexpression lines decreased at each time point compared with the WT plants (Figure 5C). Stomatal apertures of WT and transgenic lines under normal and drought stress conditions were also examined. Under normal conditions, no significant differences were observed; however, at 6 h after drought stress treatment, the stomatal apertures of OE6 and OE8 leaves decreased by 65.0 and 68.4%, respectively, while those of WT plants decreased by only 35.3% (Figures 5B,D). The MDA content, a sign of oxidative damage, was also measured, revealing a significant increase in the WT plants, but only a slight increase in the transgenic plants (Figure 5F). The overexpression lines also maintained a higher Fv/Fm ratio than the WT plants after drought treatment (Figure 5G). These findings confirm that overexpression of BrLAS confers drought resistance in the transgenic plants.

Abiotic stress can lead to increased accumulation of ROS, having damaging effects on plant development (Wang et al., 2012). To investigate the levels of active oxygen accumulation in the transgenic and WT plants, we performed DAB staining to determine H₂O₂ accumulation. Compared with WT plants, the overexpression lines showed a visible decrease in DAB staining intensity under drought stress treatment (Figure 6A), consistent with the lower accumulation of H₂O₂ in detached leaves of the BrLAS-overexpressing plants. Moreover, the WT plants exhibit higher contents of H₂O₂ and O^2- than the OE lines after drought treatment (Figures 6B,C). These findings indicate that the transgenic lines experienced less damage than the WT plants under stress conditions.

Antioxidant enzyme activity is another important indicator of a plant’s ability to scavenge ROS. We therefore measured relative activities of SOD, POD, and catalase CAT (Hu et al., 2012) in WT and BrLAS-overexpressing lines. Activities of SOD and POD in the transgenic plants were both higher than in the WT plants under normal conditions. Meanwhile, under drought stress, levels of SOD and POD activity increased significantly in the OE lines, but only slightly in the WT (Figures 6D,E). CAT activity was up-regulated in both the OE lines and WT following drought stress treatment, with no obvious differences between the two (Figure 6F). These findings suggest that BrLAS-overexpression confers increased drought stress resistance in transgenic plants through enhanced ROS scavenging ability.

To determine whether the enhanced drought resistance in the BrLAS-overexpressing lines was associated with ABA, seeds of transgenic and WT plants were germinated on 1/2 MS medium with or without ABA. The germination rates were then compared at designated time points. No significant differences in germination rates were observed between the transgenic and WT seeds when grown without ABA (Figures 7A,B). However, on 1/2 MS medium containing 2 μM ABA, germination rates of the transgenic seeds were lower than those of WT seeds at each time point (Figures 7A,C). It is well known that ABA inhibits plant growth (Luo et al., 2014), therefore, the differences in ABA-induced inhibition of root growth were also observed. Five-day-old seedlings were transferred to 1/2 MS plates supplemented with ABA and the primary root lengths were measured after vertical growth for 8 days. Exogenous ABA-induced inhibition of root growth was more severe in the transgenic than the WT seedlings when grown on 1/2 MS medium containing 10 μM ABA (Figures 7D,E). These findings suggest that overexpression of BrLAS in Arabidopsis enhances plant sensitivity to exogenous ABA.

To further investigate the molecular mechanism of BrLAS-enhanced drought stress tolerance, expression patterns of plant stress-related marker genes were examined by qRT-PCR (Figure 8). Abundant transcripts of several genes involved in ROS generation and scavenging; namely, SOD, POD, CAT2, ascorbate peroxidase (APX), and LOX, a lipoxygenase gene (Jin et al., 2013), were detected in BrLAS-overexpressing and WT plans under both normal and drought conditions. Moreover, while expression levels increased in both transgenic and WT plants, transcript levels were significantly higher in transgenic compared with WT plants under drought stress conditions, except for CAT. CAT was up-regulated in both OE and WT plants after stress treatment, with no obvious differences in expression levels, which is consistent with the change in catalase activity in both plants under drought stress (Figure 6F). Accumulation of ERD11 (Jensen et al., 2010) and RD29B (Yang et al., 2011), two drought-responsive genes, was also higher in the OE than the WT plants under drought stress treatment. Meanwhile, the two leaf senescence-related genes SAG13 (senescence-associated gene 13) (Garapati et al., 2015) and SAG113 (senescence-associated gene 113) (Zhang et al., 2011) were both down-regulated in BrLAS overexpressing compared to WT plants both under normal and drought stress conditions (Figure 8), suggesting that BrLAS overexpression delays leaf senescence in OE plants. Taken together, these findings suggest that BrLAS is involved in drought stress via modulation of key stress- and senescence-related genes.

The high-level expression patterns of ABA signaling related genes in 35S: BrLAS plants, under normal and drought stress conditions, suggested that the endogenous ABA content may be elevated in transgenic plants. To verify this hypothesis, we measured endogenous ABA contents of transgenic and WT plants with and without drought stress. As shown in Figure 9, the ABA contents in all lines were increased under drought stress conditions. As expected, the level of ABA in 35S: BrLAS plants were both higher than WT plants under normal and drought stress conditions. Drought stress treatment induced a dramatically increase in endogenous ABA content of 35S: BrLAS plants, which is up to approximately fourfold higher than ABA increase in WT plants. These data demonstrate that overexpression of BrLAS may affect endogenous ABA content.