The seeds of C. bursa-pastoris and Nicotiana tabacum were stored as previously described (Zhou et al., 2012). The seeds of A. thaliana of Columbia (Col-0) accession were obtained from the Arabidopsis Biological Resource Center (ABRI: Columbus, OH, USA). Plants were grown at stable temperature (22°C) and light conditions (16 h light/8 h dark). For cold application of C. bursa-pastoris, the 4-week-old plants were subjected to 4°C for 4, 8, or 24 h. Meanwhile the seedlings of the same age were shifted from 22 to 12°C for 4 days, 4°C for 4 days, 0°C for 2 h continuously for cold acclimation. The roots, stems and leaves were collected at each time point and frozen in liquid nitrogen immediately. For ionic liquid treatments, 2-week-old C. bursa-pastoris seedlings were soaked with 80 mM KCl, 50 mM MgCl₂, 5 mM ZnCl₂, 30 mM LiCl, 0.1 mM CuCl₂, 80 mM CaCl₂ solution for 0, 4, 8, and 24 h. For phytohormone treatments, 2-week-old A. thaliana plants were soaked with 5 μM gibberellin (GA), 300 μM methyl jasmonate (MeJA), and 500 μM salicylic acid (SA) solution for 0, 1, 6, and 24 h, respectively. The whole seedlings were harvested for RNA extraction. For the cold tolerance test of N. tabacum, 4-week-old tobacco plants were subjected to 4°C for 24 h, -4°C for 1 h and 22°C for 2 days in turn.

The genomic DNA of C. bursa-pastoris is obtained by CTAB extraction method. According to the manual of the Universal Genome Walker TM Kit (CLONTECH), the genome walker library was constructed and nested amplification was conducted using specific primers RCI35GSP1 and RCI35GSP2 paired with the adaptor primers AP1 and AP2. A 1011 bp of 5′ upstream sequence of CbRCI35 was cloned and analyzed by PLANTCARE¹ (Zhou et al., 2014).

For characterization of promoter activity, CbRCI35 promoter clones were constructed in the KpnI/NcoI site of the pCAMBIA1301 vector (CAMBIA, Australia) with primers pCbRCI35-F and pCbRCI35-R. The pCbRCI35::GUS plasmid was transformed into Arabidopsis of Col-0 plants with the Agrobacterium-mediated floral dip method. The T1 plants were selected on hygromycin (30 mg/L) media and further confirmed by PCR. The T2 lines showing 3:1 segregation were carried forward to the T3 generation. PCR positive homozygous T3 and T4 lines were used for further analyses of the CbRCI35 promoter.

For gene overexpression, CbRCI35 mRNA sequence information was previously described (Lin et al., 2007). The CbRCI35 cDNA clones were constructed in the NcoI/NheI site of the pCAMBIA1304 vector (CAMBIA, Australia) with primers CbRCI35-nco-F and CbRCI35-nhe-R, using hygromycin resistance as a selection marker. The 35S::CbRCI35 plasmid was introduced into leaf disks of tobacco using Agrobacterium EHA105. Plants transformed with the empty vector were used as a blank control. Transformants were selected on MS medium solidified with 0.8% agar containing 0.1 mg/L NAA, 1.0 mg/L 6-BA, 250 mg/L carbenicillin disodium and 30 mg/L hygromycin, and regenerated on hormone-free MS medium containing 250 mg/L carbenicillin disodium at 22°C. Transformants were identified by DNA-PCR analysis with primers Hyg-F and Hyg-R (Table 1).

The seedlings of 2.5-week-old transgenic Arabidopsis with pCbRCI35::GUS at 22°C or after 4°C treatment for 8 h were collected and soaked in GUS staining buffer containing 0.075% X-Gluc (5-bromo-4-chloro-3-indolyl-bd-glucuronic acid), 3 mM potassium ferricyanide, 7.2 mM EDTA, 57.7 mM disodium phosphate, 42.3 mM sodium phosphate, and 0.005% Triton X-100 followed by vacuum infiltration for 20 min. Afterward the seedling samples were incubated at 37°C overnight and de-stained in 75% ethanol at 60°C for several times until chlorophyll was removed. Another set of 4.5-week-old transgenic plants were stained and embedded with Technovit 7100 plastic embedding kit (Heraeus Kulzer, Germany) following the manufacturer’s instructions and semi-sectioned by Leica 2265 Rotary Microtome (Leica, Germany). The sections were observed with Zeiss Scope A1 microscope (Zeiss, Germany).

The CbRCI35 cDNA clones were constructed in the NcoI/BglII site of the pCAMBIA1302 vector (CAMBIA, Australia) with primers CbRCI35-nco-F and CbRCI35-bgl-R. The onion epidermis discs were incubated with a culture of Agrobacterium carrying the 35S::CbRCI35-GFP plasmid for 20 min. Acetosyringone (100 μM) was added to enhance the transformation efficiency. After grown in the dark at 25°C for 2 days, the GFP signal was detected by confocal laser scanning microscopy (Zeiss 710, Germany) and the images were analyzed with Zen software.

Quantitative real-time PCR (qPCR) was employed to analyze gene relative expression levels. The total RNA was extracted using Plant RNA Mini Kit (CW Biotech Inc., Ltd, China) and the cDNA was synthesized using PrimeScript^® RT Master Mix (TaKaRa, China) according to the manufacturer’s instructions. The qPCR was carried out using SYBR^® Premix Ex Taq^TM II (Perfect Real-Time; TaKaRa, China) on a StepOnePlus^TM Real-Time PCR System (Applied Biosystems) with three replicates. The PCR procedure was 95°C for 30 s, 40 cycles of 95°C for 5 s, and 60°C for 34 s, followed by 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s. The CbACTIN (HQ880662), ACTIN2 (AK230311) and NtACTIN (AJ133422) were used as the internal control for C. bursa-pastoris. A. thaliana, and N. tabacum, respectively.

Physiological indices were measured as previously described (Wu et al., 2012). In brief, leaves of tobacco were incubated in 2 ml deionized water at room temperature for 12 h and the sample conductivities (C1) were measured with a DDS-11A meter (Shanghai SUOSHEN Electrical Equipment Co. Ltd., China). Then the samples were boiled in deionized water at 100°C for 1 h and cooled to room temperature to get the sample conductivities (C2). Relative electrolyte leakage was calculated using the following formula: C1/C2 × 100%. Relative water content (%) was calculated as 100% × (FW-DW)/(TW-DW), in which the fresh weight (FW) and the turgid weight (TW) were determined before and after leaves dipped in 4 ml deionized water for 12 h at room temperature, while the dry weights (DW) were measured after samples were oven dried at 65°C for 24 h. Glucose content was assayed by measuring the NADH production using the Glucose (HK) Assay kit (Sigma–Aldrich, Inc.) depending on absorbance at 340 nm (A₃₄₀). The H₂O₂ content was measured according to A₄₁₅ of the titanium-peroxide complex as previously described (Jiang and Zhang, 2001). The superoxide dismutase (SOD) activity was determined by an inhibited photoreduction rate of nitro blue tetrazolium (NBT) (Mckersie et al., 1993). The malondialdehyde (MDA) content was measured using the thiobarbituric acid (TBA)-based colorimetric method as reported previously (Draper et al., 1993). The BioPhotometer Plus (Eppendorf, Germany) was used for electromagnetic absorbance reading in each assay. Three biological replicates were used for each experiment and the data were analyzed by Student’s t-test.

Tobacco leaf disk samples were immersed with 10 μM carboxy-2′,7′-dichloro- dihydro-fluorescein diacetate (DCFH-DA) probe at 37°C for 20 min in the dark (Ezaki et al., 2000). Samples were subsequently washed with 20 mM potassium phosphate buffer (pH 6.0) for three times to remove extra probe. Fluorescent signals were visualized using a confocal laser scanning microscope (Zeiss 710, Germany).

The RCI genes were identified according to their cold-induced expression pattern. CbRCI35 was responsive to cold temperatures in young C. bursa-pastoris plants (Lin et al., 2007). For a more detailed characterization, the expression patterns of CbRCI35 in different tissues in response to low temperature were detected in a cold acclimation assay and a time course assay. CbRCI35 showed a root-specific and temperature-dependent expression pattern in C. bursa-pastoris during cold acclimation from 12, 4 to 0°C. After 4 days of 12°C application, the CbRCI35 transcript did not change in leaves, stems or roots compared with the 22°C control. When treated at 4°C for 4 days followed by 0°C for 2 h, significant enhancements were observed only in roots and 0°C activated the highest observed transcription level (Figure 1A). In a time course assay, the basal level of CbRCI35 mRNA was highest in roots, and was lower in stems and leaves. During cold application, CbRCI35 transcription was elevated at the time point of 4 h and reached its peak after 8 h in 4°C. As before, in roots it showed the highest expression level, while in leaves and stems the gene expression pattern had a similar trend but the transcript levels were lower (Figure 1B).

The RCI genes coding peroxidases such as AtRCI3 are involved in K⁺ transportation (Kim et al., 2010) and the protein activity can be affected by ion concentration (Lai et al., 2006). We detected CbRCI35 expression in response to 80 mM KCl, 50 mM MgCl₂, 5 mM ZnCl₂, 30 mM LiCl, 0.1 mM CuCl₂, 80 mM CaCl₂ in young seedlings of C. bursa-pastoris. Consistent with the induced transcription of AtRCI3 under K⁺ deficiency, CbRCI35 expression was repressed by K⁺ application. Mg²⁺ also down-regulated CbRCI35 expression and Zn²⁺ caused a transient suppression. For Li⁺, Cu²⁺, and Ca²⁺, the transcription of CbRCI35 showed a significant transient elevation in the early stage, suggesting that CbRCI35 might play a role in Li⁺, Cu²⁺, and Ca²⁺ transportation (Figure 1C).

A fragment of 1011 bp in the promoter region of CbRCI35 was isolated from the C. bursa-pastoris genome by DNA walking (Figure 2). Sequence analysis was performed using the PLANTCARE software². Three cis-acting elements involved in gibberellin (GA), SA and MeJA responsiveness were identified, respectively. To further characterize the CbRCI35 promoter, a GUS reporter system driven by the 1011 bp pCbRCI35 promoter fragment was created (Figure 3A). Histochemical GUS staining indicated extremely low levels of GUS activity throughout the entire seedling at 22°C, and increased activity after 8 h exposure to 4°C in 2-week-old plants (Figure 3B). Correspondingly, a markedly increased level of GUS mRNA in response to cold was detected in 2-week-old seedlings (Figure 3C). For a more accurate, tissue-specific analysis, semi-thin sections of 4.5-week-old plant organs were dissected out after staining. In contrast to what is observed in younger seedlings, GUS accumulation was mainly in the roots, while comparatively lower GUS enhancement was observed in stems and leaves (Figure 3D). The GUS activity in roots was displayed in the cortex but not the epidermal or vascular tissues, implying the protective function of CbRCI35 for the cortical cells containing stored carbohydrates and other substances. These indicated that the cold-induced activity of CbRCI35 was in an age-dependent, tissue-specific manner. Considering the predicted cis-acting elements in the CbRCI35 promoter, phytohormone responsive activity was subsequently investigated using GA, SA, and MeJA treatments. Interestingly, GA and MeJA inhibited CbRCI35 transcription while SA conferred transient upregulation of gene expression in early stages and downregulation after 24 h of application (Figure 3E). The promoter characterization revealed a cold and phytohormone inducible expression pattern of CbRCI35.

Given the tissue-specific expression pattern of CbRCI35, it was relevant to study the subcellular localization of the CbRCI35 protein to further clarify its function. It was reported that AtRCI3, a highly homologous protein to CbRCI35, was a secreted protein and could be sorted to the cell wall (Kim et al., 2010). The online tool Plant-mPLoc³ suggested that CbRCI35 was localized in the cytoplasm. Using a transient expression assay, CbRCI35-GFP fusion protein driven by the 35S promoter was expressed in onion epidermal cells (Figure 4A). The CbRCI35-GFP promoter was distributed around the periphery of onion cells, while the signal of the GFP control was also detected in the nucleus (Figure 4B). After plasmolysis in 2 g/mL sucrose solution, the CbRCI35-GFP fusion signal was observed in the cytoplasm but not the cell wall (Figure 4C). These showed that unlike secretion of AtRCI3, CbRCI35 is localized in the cytoplasm, although the sequence similarity of these two proteins is as high as 95% (Lin et al., 2007).

Next, the functional analysis of CbRCI35 was carried out in cold sensitive tobacco (N. tabacum). The CbRCI35 expression level was examined in transgenic tobacco lines. Two individual CbRCI35-ox lines (L1 and L4) exhibited high transcript levels and were therefore selected for subsequent experiments (Figure 5A). Compared with wild type (WT) and empty vector transformants (EV), L1 and L4 plants showed higher capacity of cold acclimation. WT and EV seedlings both had wilting leaves while L1 and L4 showed a robust status after 1 h of -4°C exposure followed by 2 days of recovery at 22°C (Figure 5B). Corresponding physiological analyses were performed using four indices: electrolyte leakage, malondialdehyde (MDA) content, relative water content, and glucose content – all representative indicators of cellular damage under low temperatures (Zhou et al., 2014). After freezing application, relative water contents and glucose content in L1 and L4 were much higher than that of the controls, indicating the protection of cellular water and bioactive components in transgenic plants. In addition, two CbRCI35-ox lines showed significantly lower electrolyte leakage and less MDA accumulation under both chilling and freezing stress, which was a symbol of the protection of plasma membrane integrity (Figure 5C). These phenotypically demonstrated that CbRCI35 overexpression increased freezing resistance in tobacco.

Moreover, the impact of CbRCI35 on endogenous cold responsive signaling genes in tobacco was investigated. NtDREB1 and NtDREB3 belong to CBF/DREB transcription factor family, a group of key cold responsive regulators (Catala et al., 2003; Zhou et al., 2012). NtERD10a and NtERD10b are downstream CBF/DREB regulons modulated by DREBs (Kasuga et al., 2004). In normal temperatures, NtDREB3 of L1 and L4 exhibited significantly higher transcript levels than WT or EV. NtERD10a as well as NtERD10b were also upregulated (Figure 6). During treatment in 4°C, NtDREB3 expression levels of transformants were still much higher than the controls, although they remained similar compared to NtDREB3 expression at 22°C. NtERD10a and NtERD10b were remarkably induced by cold and showed much higher transcripts in L1 and L4. NtDREB1 did not exhibit obviously different expression levels from WT and EV. It can be concluded that CbRCI35 positively regulated NtDREB3. NtERD10a, and NtERD10b expression, and conferred enhanced cold induction of these two ERD10 genes in cold temperature.

Predicted as a peroxidase, CbRCI35 is supposed to regulate ROS levels in plants. Using the dichloro-dihydro-fluorescein diacetate (DCFH-DA) assay, we verified that two CbRCI35-ox tobacco lines exhibited moderately elevated ROS levels before exposure to cold temperatures compared to WT plants. After exposure to 4°C for 24 h, the ROS levels in each tobacco were similar (Figure 7A). As a representative component of ROS (Slesak et al., 2007), H₂O₂ showed content changes that were in line with DCFH-DA results (Figure 7B). We also detected the activity of SOD, a primary antioxidant enzyme in plants (Price et al., 1994). In both normal and chilling conditions, CbRCI35-ox tobacco showed higher SOD activity than WT, which can contribute to the protection of cell structure (Figure 7C). The increase of ROS at the basal level together with the stronger SOD activity after cold exposure indicated the effects of CbRCI35 in plant ROS regulation and cell protection. We further examined the expression of ROS metabolic genes in transgenic tobacco. Intriguingly, ROS scavenging and biosynthesis genes were both disrupted (Figure 8). In 22°C, transcripts of a key ROS scavenging gene NtAPX (Yan et al., 2014) were significantly lower in CbRCI35-ox plants, while that of NtSOD, another ROS detoxifying gene (Liu et al., 2016), was higher than controls. After 24 h of 4°C application, expression levels of the ROS scavenging genes NtAPX. NtCAT, and NtGST were noticeably lower in CbRCI35-ox plants. Meanwhile, ROS production modulators NtRBOHD1 and NtRBOHD2 were downregulated by CbRCI35 under normal temperature, and the cold induction of NtRBOHD1 was also significantly blocked in CbRCI35-ox tobacco (Figure 8). Taken cumulatively, these revealed that the overexpression of CbRCI35 regulates ROS homeostasis in tobacco.