The R9 thermotolerant pepper line (introduced from the World-Asia Vegetable Research and Development Center, PP0042-51) and the Arabidopsis ecotype Col-0 variety were used in this study. Pepper seedlings were grown under normal conditions (26°C/20°C day/night, 200 μmol⋅m^-2⋅s^-1 illumination intensity, thermo- and photoperiod of 16 h light /8 h dark cycle, and 70 % relative humidity) in a controlled climate chamber, and Arabidopsis seedlings were grown under 22°C/18°C day/night conditions.

The amino acid sequences of CaBiP1, 2, and 3 were downloaded from the PGD¹ (Qin et al., 2014) and PGP² (Kim et al., 2014) pepper genome databases. Arabidopsis AtBiP amino acid sequences were obtained from Genbank³ (Noh et al., 2003) and rice OsBiP sequences were downloaded from Rice Genome Annotation Project⁴ (RGAP) (Sarkar et al., 2013).

Alignment of full-length BiP amino acid sequences from pepper, rice and Arabidopsis was performed using the online program of Clustal Omega⁵, and the phylogenetic tree was constructed using MEGA 6 with the neighbor-joining method, p-distance substitution model and 1000 bootstrap replicates (Tamura et al., 2013). Functional and signaling domains in CaBiPs were identified based on published Arabidopsis and rice literatures (Noh et al., 2003; Sarkar et al., 2013). The identification of conserved motifs in CaBiP proteins was carried out using the MEME program⁶ with the following parameters: normal motif discovery mode, maximum number of motifs = 9, a motif site distribution in which each gene has none or only one motif, and a motif width between four and 200 amino acids. Structural diagrams for exon-intron analysis were generated using the online program GSDS⁷.

The CaBiP-1 ORF without a termination codon was amplified using specific primer pair GFP-CaBiP1-F and GFP-CaBiP1-R (Supplementary Table S1) from a cDNA template isolated from R9 leaf material grown under normal growth conditions, and then the PCR product was cloned into the pBI221 expression vector containing green GFP. An empty vector without CaBiP-1 was used as the control. Particle bombardment was performed to introduce recombinant plasmid into onion epidermal cells. ER-Tracker Red (Beyotime, C1041, China), a specific fluorescent probe for ER, was used to highlight this cellular component. Details of the methods can be found in our previous study (Guo et al., 2014).

A 346 bp fragment of the CaBiP1 ORF was amplified by gene-specific primer pair TRV2-CaBiP1-F and TRV2-CaBiP1-R (Supplementary Table S1) from a cDNA template isolated from R9 leaf material grown under normal growth conditions, and the PCR product was cloned into the pMD19T vector (Takara, Dalian, China). After digestion with restriction enzymes of Xba I and Kpn I, the CaBiP1 fragment was cloned into the pTRV2 virus expression vector to generate the TRV2:CaBiP1 silencing construct. The empty TRV2:00 vector without CaBiP1 was used as the control, and TRV2:CaPDS (phytoene desaturase gene) was used as a marker for gene silencing. Agrobacterium tumefaciens strain GV3101 cells containing TRV2:CaBiP1, TRV2:00 or TRV2:CaPDS were separately injected into the leaves of the R9 thermotolerant pepper line as described by Wang et al. (2013). When the photo-bleaching phenotype was evident in pepper seedlings carrying TRV2:CaPDS, the silencing efficiency of pTRV2:CaBiP1 was assessed by qRT-PCR with the primer pair of qCaBiP1-F and qCaBiP1-R (Supplementary Table S1).

The full-length CaBiP1 ORF was amplified from cDNA isolated from R9 leaf material grown under normal growth conditions with the gene-specific primer pair of CaBiP1-F and CaBiP1-R (Supplementary Table S1). The amplification product was inserted into the plant transformation binary vector pBI121 between the CaMV-35S promoter and the nos (nopaline synthase) terminator. The resultant pBI121 vector was transformed into Arabidopsis ecotype Col-0 using the floral dip method intermediated by Agrobacterium GV3101 (Clough and Bent, 1998). Transgenic plants were obtained by screening successive generations for kanamycin resistance, and T3 seeds were used for subsequent experiments.

For tissue-specific expression analysis of CaBiP genes, young leaves, flower buds, fruits (about 1 cm in length), stems and roots were collected from pepper plants grown under normal conditions. Seedlings at the six-leaf stage were used for abiotic stress treatments. For abiotic stress treatments involving abscisic acid (ABA), H₂O₂, DTT, the plants were sprayed with 0.1 mM ABA, 1 mM H₂O₂, or 15 mM DTT until leaves were thoroughly wetted, and leaves were collected at 0, 1, 3, 6, 12, and 24 h post treatment. For salt and osmotic stress experiments, the roots of the seedling were soaked in 200 mM NaCl and 200 g⋅L^-1 PEG6000, respectively, and leaves and roots were sampled at 0, 1, 3, 6, 12, and 24 h post treatment. For heat treatment, pepper seedlings were incubated at 45°C, and leaves were collected at 0, 0.5, 1, 2, 4, and 6 h post treatment. All samples were immediately frozen in liquid nitrogen and kept at -80°C for RNA extraction.

Pepper seedlings of TRV2:CaBiP1 and TRV2:00 were used for abiotic stress treatments. For ER stress, pepper seedlings were sprayed with 30 mM DTT until leaves were thoroughly wetted and incubated for 24 h. For heat stress, pepper seedlings were exposed to 45°C for 24 h then allowed to recover for 5 days under normal conditions. For osmotic stress, the seedlings were soaked in 300 g⋅L^-1 PEG for 24 h. For drought stress, pepper seedlings were deprived of water for 10 days, after which the RWC was determined. For dehydration tests, detached leaves were placed on the bench and weighted from 0 to 240 min at intervals of 30 min. For salt stress, pepper seedlings were divided into two groups, and one group was irrigated with 300 mM NaCl for 14 days, while the root of the other group was soaked in 300 mM NaCl for 24 h. After treatment, pepper leaves were sampled immediately used for determination of H₂O₂, REL, MDA, and soluble protein.

Seeds from Arabidopsis overexpressing CaBiP1 and wild-type Col-0 were germinated on MS plates with 3 mM DTT for ER stress, and 5-day-old seedlings grown on normal medium were transferred to medium with 2 mM DTT for 15 days. For heat stress, the MS plates with 7-day-old transgenic Arabidopsis seedlings were immersed in a water bath at 45°C for 50 min, then recovered at 22°C for 5 days, while 2-week-old seedlings in pots were heat-treated at 45°C for 6 h and recovered at 22°C for 7 days in a controlled temperature chamber. For salt and osmotic stress tests, Arabidopsis seeds were grown on MS medium with 100 mM NaCl or 200 mM mannitol for 4 days to determine the germination rate, and on medium with 75 mM NaCl or 150 mM mannitol for 7 days to determine root length. For salt stress experiments, 2-week-old seedlings were watered with 300 mM NaCl for 15 days to determine the survival rate. For dehydration tests, the detached Arabidopsis leaves were placed on the bench and weighted from 0 to 240 min at intervals of 30 min. Arabidopsis seedlings were deprived of water for 12 days to measure the RWC, root length, and re-watered for 2 days to determine the survival rate. In addition, 3-week-old Arabidopsis seedlings were soaked with root in 300 mM NaCl or 300 mM mannitol solutions, or sprayed with 30 mM DTT, or exposed to high temperatures of 45°C, and leaves were collected at 0, 3, and 9 h post treatments for genes expression analysis. For drought stress, 2-week-old Arabidopsis seedlings were withheld water for 10 days for gene expression analysis. All experiments were performed with three biological replicates.

Pepper leaf discs were used to measure the REL according to the method of Yin et al. (2014), and the MDA content was determined using the thiobarbituric acid reaction according to Dhindsa and Thorpe (1981). H₂O₂ levels were assessed by the DAB staining method (Dang et al., 2013), and the soluble protein content was measured by the protein dye-binding method with Coomassie light blue (Bradford, 1976). RWC was measured as described by Chakraborty et al. (2012). All experiments were performed with three biological replicates.

Total RNA was extracted from sampled plant tissues using the Trizol (Invitrogen, Carlsbad, CA, United States) method, and the residual genomic DNA was digested by RNase-free DNase I (Promega, Madison, WI, United States). The first-strand cDNA was synthesized using the PrimeScript^TM Kit according to the manufacturer’s instructions (TaKaRa, Tokyo, Japan). Primer pairs (Supplementary Table S1) were designed by NCBI Primer-BLAST⁸, and qRT-PCR was performed using SYBR^® Premix Ex Taq^TM II (TaKaRa) as described previously by Wang et al. (2013). Relative gene expression levels were analyzed according to the 2^-ΔΔCT method (Livak and Schmittgen, 2001), and CaUBI3 and AtActin2 were used as internal controls in pepper and Arabidopsis, respectively. Significance tests for differences in gene expression levels between control and stress treatments were performed using the Student’s t-test method at the α = 0.05 and 0.01 levels.

The phylogenetic tree of BiP proteins from pepper, Arabidopsis and rice showed that CaHsp70-8 and CaHsp70-7 share a close evolutionary relationship with AtBiP1 and AtBiP2, respectively, while CaHsp70-10 is more closely related to AtBiP3 and OsBiP2 (Figure 1). Therefore, CaHsp70-8, CaHsp70-7, and CaHsp70-10 were renamed as CaBiP1, CaBiP2, and CaBiP3, respectively.

Alignment of CaBiP protein sequences with AtBiPs (AtBiP1, NP_198206; AtBiP2, NP_851119; AtBiP3, NP_172382) and OsBiPs (OsBiP1, XP_015625618; OsBiP2, XP_015629631; OsBiP3, XP_015638605; OsBiP4, XP_015638801; OsBiP5, BAT04227; OsBiP6, LOC_Os01g33360) revealed high conservation in the signaling and functional domains. Seven domains and nine motifs were identified in CaBiPs based on published Arabidopsis and rice literatures (Noh et al., 2003; Sarkar et al., 2013) (Figures 2A,B). A signal peptide sequence (SP) for membrane transport was detected at the beginning of the N-terminus in CaBiPs, and the cut-off point GI (labeled by black arrow), a conserved site separating ATPase domain from the peptide-binding domain, was also found. In the N-terminal ATPase region, all the essential domains for BiP ATPase activity, including Domain 1 (motif 5) and Domain 2 (N-terminus of motif 1) for phosphate-binding, and Domain 4 (motif 2) for adenosine binding, are conserved in the predicted CaBiP proteins. Meanwhile, Domain 3 (C-terminus of motif 1) for calmodulin binding was also identified in CaBiPs. The C-terminal protein binding region has a highly conserved five residue core (shown by red arrows) that mediates hydrogen bonding with peptide substrates. However, Domain 5 (motif 6), which is essential for binding to peptide substrates, differs between CaBiP3 and CaBiP1/CaBiP2. Although CaBiP3 has the same pattern in motif distribution pattern found in CaBiP1 and CaBiP2, motif 7 containing the SP and motif 9 containing the ER retention signal are absent in CaBiP3 (Figure 2B and Supplementary Table S2). Analysis of gene structure also revealed that CaBiP3 lacks an intron that is present in CaBiP1 and CaBiP2 (Figure 2C).

Expression profiles of the three CaBiP genes in pepper root, stem, leaf, flower and fruit tissues were determined by qRT-PCR with transcript-specific primers (Supplementary Table S1). The results indicated that the three CaBiP genes were diversely expressed in roots, stems, leaves, flowers, and fruits (Figure 3A), and the overall expression level of CaBiP3 was much lower than CaBiP1 and CaBiP2 (Figure 3B). In addition, the expression levels of CaBiP1 and CaBiP2 genes were obviously higher in stem and leaf tissue than in other organs, while CaBiP3 expression was higher in stem and flower (Figure 3A).

Expression of all CaBiP genes was induced under abiotic stress conditions involving ABA, H₂O₂, DTT, heat, salt, and osmotic stresses (Figure 4). Peak expression occurred earlier in heat stress (1 h) than in other treatments (3 h), and the fold change in CaBiP3 expression was significantly higher than that of CaBiP1 and CaBiP2 in all treatments. In addition, CaBiP was expressed more highly in roots than in leaves under osmotic stress and salt stress.

Under the control of CaMV-35S promoter, the 35S:CaBiP1-GFP construct was transiently expressed in onion epidermal cells. While both the green fluorescence from fusion protein of CaBiP1 and GFP and the red fluorescence from ER-Tracker Red were detected by laser confocal microscopy, their merged yellow fluorescence was also observed (Figure 5). The results indicate CaBiP1 is located in the ER.

Pepper seedlings harboring TRV2:CaBiP1 demonstrated a silencing efficiency of over 70%, and Arabidopsis transgenic lines OE7 and OE8 were used to perform abiotic stress treatments. No visible difference was observed between either TRV2:CaBiP1 and TRV2:00 pepper lines, and the same was true for CaBiP1-OE lines and WT Arabidopsis plants (Supplementary Figure S1).

Following ER stress induced by 30 mM DTT, the H₂O₂ content in pepper leaves, as revealed by dark brown dots after DAB staining, was increased in both TRV2:CaBiP1 and TRV2:00 lines, albeit to a greater extent in the former (Figure 6A). Similar results were also observed for REL levels and MDA content (Figures 6B,C). By contrast, the soluble protein content was decreased in the leaves of both TRV2:CaBiP1 and TRV2:00 plants following ER stress, and the decrease was again more significant in the former (Figure 6D).

Under normal growth conditions, no difference was observed in the germination rate between CaBiP1-OE lines and WT Arabidopsis lines. However, under ER stress induced by 3 mM DTT, the germination rate was significantly higher in CaBiP1-OE plants (Supplementary Figure S2A). Furthermore, when the treatment was extended to 14 days, chlorosis of cotyledons was more severe in WT than in CaBiP1-OE lines (Figures 6E,F). After Arabidopsis seedlings at 5-day-old were transferred to MS medium containing 2 mM DTT, the growth was inhibited in both CaBiP1-OE and WT lines, but fresh weight and root length were obviously higher in the CaBiP1-OE plants (Figures 6G,H and Supplementary Figure S2B). Additionally, expression of some genes in the UPR pathway, such as AtbZIP60F, AtbZIP60S, AtbZIP28, and AtNF-YC2 (encoding a component of the AtbZIP28 transcription complex; Liu and Howell, 2010), were up-regulated when the Arabidopsis seedlings were exposed to DTT, and the fold change was greater in transgenic lines than in WT plants (Figure 6I).

When TRV2:CaBiP1 and TRV2:00 pepper seedlings were subjected to 45°C for 24 h, sunburn-like symptoms of heat stress was observed in TRV2:CaBiP1 leaves but not in those of TRV2:00 plants. After recovery under normal conditions for 5 days, the injured leaves in TRV2:CaBiP1 fell off, while TRV2:00 leaves resumed growth (Figure 7A). Under heat stress, H₂O₂, MDA content, and REL levels were increased in both TRV2:CaBiP1 and TRV2:00 lines, and all three indicators were markedly higher in the former (Figures 7B–D).

Arabidopsis seedlings of 1-week-old were treated at 45°C for 50 min, then recovered for 5 days under normal conditions, and only 20% of WT plants survived, compared with more than 90% of CaBiP1-OE plants (Figures 7E,F). When 2-week-old seedlings of WT and CaBiP1-OE lines in pots were subjected to 45°C for 6 h then recovered for 5 days, about 30% of WT plants survived, compared with 74.1% for OE7 and 70.4% for OE8 plants (Figures 7G,H). During heat stress treatments, both UPR-related genes (AtbZIP60S and AtbZIP28) and heat responsive genes (AtHsfA2, AtHsfA7a, and AtHsp101) were obviously induced in both WT and transgenic Arabidopsis lines. However, the induction of gene expression in transgenic lines was more intense than in WT plants. Furthermore, the expression levels of three heat response genes peaked at 3 h, while those of two UPR genes increased continuously during the treatments (Figure 7I).

After irrigation with 300 mM NaCl solution for 14 days, the leaves of CaBiP1-silenced pepper seedlings became yellowish and wilted and eventually fell off, while in TRV2:00 plants, only yellowness was observed (Figure 8A). When the root of TRV2:CaBiP1 and TRV2:00 seedlings were soaked in 300 mM salt solution for 24 h, the accumulation of H₂O₂ in TRV2:CaBiP1 leaves was higher than in TRV2:00 leaves (Figure 8B), and the MDA content was also 1.4-fold higher in TRV2:CaBiP1 leaves (Figure 8C). Similarly, REL levels were higher in TRV2:CaBiP1 (36.8%) than TRV2:00 (26.4%) (Figure 8D).

When Arabidopsis seedlings were grown on MS medium with 100 mM NaCl for 4 days, the germination rate of CaBiP1-OE lines was obviously higher than that of WT plants (Supplementary Figure S3). Additionally, in Arabidopsis seedlings grown on MS plates containing 75 mM NaCl for 7 days, the roots of CaBiP1-OE lines were significantly longer than those of WT plants (Figures 8E,F). Following watering of 2-week-old seedlings with a high concentration of NaCl (300 mM) for 15 days, CaBiP1-OE seedlings showed a higher salt tolerance than WT plants; the survival rates of CaBiP1-OE7 and OE8 plants were 2.0- and 1.9-fold higher than WT, respectively (Figures 8G,H). In terms of gene expression, salt stress-related genes (AtAPX2, AtDREB2A, and AtRD29A) and UPR-related genes (AtbZIP60S and AtbZIP28) displayed higher levels of transcription in CaBiP1-OE plants than in WT plants (Figure 8I). Interestingly, the expression of all the tested genes revealed continuous enhancement throughout the duration of salt stress experiments (Figure 8I).

After soaking in 300 g⋅L^-1 PEG6000 solution for 24 h, TRV2:CaBiP1 pepper seedlings accumulated more H₂O₂ than did TRV2:00 plants, as shown by the dark brown dots in leaves after DAB staining (Figure 9A). Similarly, the MDA content and REL levels in TRV2:CaBiP1 seedlings were obviously higher than those of TRV2:00 lines following osmotic stress (Figures 9B,C). By contrast, the soluble protein content in TRV2:CaBiP1 leaves was significantly lower than in TRV2:00 leaves (Figure 9D). Following osmotic stress, Arabidopsis seeds from CaBiP1-OE plants grown on MS medium containing 200 mM mannitol displayed a remarkably high germination rate compared with that of WT plants; seed germination in OE7 and OE8 lines reached to 84 and 78% at the 4th day, respectively, whereas that of WT was only 54% (Figure 9E). On the 7th day after germination on MS medium containing 150 mM mannitol, CaBiP1-OE Arabidopsis seedlings exhibited stronger tolerance to osmotic stress, as evidenced by a 1.8-fold increase in root length compared with WT plants (Figures 9F,G). The qRT-PCR analysis indicated that osmotic stress-related genes (AtDREB2A, AtRD29A, and AtHsfA2) were highly induced in stressed leaves (Figure 9H). Notably, AtRD29A, a marker for osmotic stress, as up-regulated by 5.0- and 3.3-fold at 9 h in OE7 and OE8 lines compared with WT plants. Similarly, CaBiP1-OE lines showed higher expression levels of ER stress-related genes (AtbZIP60S and AtbZIP28) at 3 h and 9 h following exposure to 300 mM mannitol. Additionally, expression of AtRD29A and AtHsfA2 peaked at 3 h, while expression of the other three genes increased continuously throughout the treatment (Figure 9H).

At the end of the drought stress treatment that involved withholding water for 10 days, wilting symptoms were observed in pepper leaves of TRV2:CaBiP1 lines but not TRV2:00 (Figure 10A). Furthermore, TRV2:CaBiP1 plants exhibited higher REL level and lower RWC than TRV2:00 (Figures 10B,C). Similarly, dehydration of pepper leaves placed on a bench for 4 h resulted in a rate of water loss that was higher in gene-silenced pepper seedlings compared with controls (Figure 10D).

Following the withholding of water for 12 days, Arabidopsis seedlings became severely withered, but CaBiP1-OE seedlings exhibited higher RWC and longer root than WT plants (Figures 10E,F). After re-watering for 2 days, more than 90% of CaBiP1-OE Arabidopsis seedlings survived, whereas just 59% WT plants remained alive (Figures 10G,H). In addition, under dehydration conditions, the rate of water loss for detached rosette leaves from transgenic seedlings was markedly lower than that in WT leaves (Figure 10I). Additionally, CaBiP1-OE lines showed a more obvious induction in the drought stress-related genes (AtDREB2A, AtRD29A, and AtAPX2) and ER stress-related genes (AtbZIP60S and AtbZIP28) under drought stress than did WT plants (Figure 10J).