Physiological, biochemical and genetic responses of Caucasian tea (Camellia sinensis (L.) Kuntze) genotypes under cold and frost stress

Plant cultivation, cold treatment and sampling

The experiments on cold and frost induction were carried out using two-year-old plants of tea cv. Kolkhida (frost sensitive) and Gruzinskii7 (frost tolerant) (Tuov & Ryndin, 2011; Gvasaliya, 2015) (Fig. 1). Plants were obtained by vegetative propagation of adult tea plants from field collections of the Federal State Budgetary Scientific Institution RRIFSC of two locations: Goitkh (GPS: N44°14′51″E 39°22′33.96″- cultivar Gruzinskii7) and Uch-Dere (GPS: N43°66′89.64″E, 39°63′14.51″- cv. Kolkhida). Both cultivars were shown to survive in −5 °C (cv. Kolkhida) and −15 °C (cv. Gruzinskii7) temperatures. Plants were grown in 2 liter polyethylene pots filled with brown forest acidic soil (pH = 5.0). According to the literature (Hao et al., 2018) cold acclimation of tea plants started when temperature decreased lower than +10. On the other hand, winter comes not immediately after optimum growing period (+18–25 °C) in natural conditions. Therefore medium temperature was selected for the control treatment of plants. Before the cold treatments plants were grown for three months in control conditions with the temperature of +12–14 °C (, with illumination regime of 14 h of light and 10 h of dark, with light intensity of 3000 lux with normal irrigation. Only healthy plants were selected for these experiments. 10 plants of each genotype were included in the study. For each assessed parameter, 2nd, 3rd and 4th mature leaves were used for sampling for each analysis.

Low temperature stress was induced using cold chambers HF-506 (Liebherr, Denmark) as follows: decreasing the temperature by 0–2 °C for 7 days (cold treatment), following decreasing the temperature by −6∼−8 °C for 5 days (frost treatment) to reveal the mechanisms of cold acclimation and frost hardening, respectively. During the treatments, the illumination regime was established as follows: 14 h of light and 10 h of dark every day, with light intensity of 3000 lux.

In order to identify morphological markers of cold tolerance, microstructural parameters of leaves were analyzed by light microscopy using Axio Imager 2 (Carl Zeiss) with the related software. Freshly prepared leaf sections were analyzed in three replicates with 10 fields of view in each.

To determine the (RWC), fresh leaves (FW) were first weighed and then dried at 105 °C for five hours (DW). RWC was calculated according to the formula: $RWC=\left(\frac{\left(FW-DW\right)}{FW}\right)\ast 100\text{\%}$ (Yamasaki & Dillenburg, 1999).

The cell membranes integrity was measured with a portable conductivity meter ST300C (Ohaus). 200 mg fresh leaf sample was immersed in 150 ml of deionized water. The measurement of electrical conductivity was done twice: before and after boiling for 60 min at 100 °C. The cell membranes integrity (CMI, %) was calculated using the formula: ${\displaystyle CMI=\frac{1-\left(\frac{L1}{L2}\right)}{1-\left(\frac{C1}{C2}\right)}\ast 100,}$ where L1 and L2 are the conductivity values before and after boiling, C1 and C2 are the relative conductivity of control before and after boiling (averaged over five replicates) (Bajji & Kinet, 2001).

The pH of the cell sap was determined potentiometrically on a pH-meter Testo 205. 1 g of fresh leaves was homogenized in 20 ml of distilled water for pH determination using a hydrogen electrode.

The water-soluble protein content was determined spectrophotometrically in five replicates according to Bradford protocol (Bradford, 1976; Bonjoch & Tamayo, 2001). The optical density of protein solution was measured at a wavelength of 595 nm on a spectrophotometer USF-01 (Russia).

Proline content in leaves (mg g⁻¹ fresh leaf mass) was evaluated spectrophotometrically by simplified ninhydrin method (Shihalyeyeva et al., 2014). Absorbance of solution was measured at 520 nm using a spectrophotometer USF-01 (Russia).

Other amino acids (arginine, tyrosine, beta-phenylalanine, leucine, methionine, valine, threonine, serine, alpha-alanine, glycine) (mg g⁻¹) as well as sugar content (mg g⁻¹) and cations (g g⁻¹) were evaluated by capillary electrophoresis on analyzer Kapel-105M (Russia) (Brykalov et al., 2019). Fold-changes of amino acids were counted as the ratio of absolute values cold/control and frost/control.

Gene expression analysis by qRT-PCR

Total RNA was extracted from the third mature leaf in three biological replicates by the guanidine method with sorption on silica columns, according to the manufacturer’s protocol (Biolabmix, Novosibirsk, Russia). The concentration and quality of RNA was determined on an IMPLEN NPOS 3.1f nano-spectrophotometer and integrity was assessed in a 1% agarose gel. RNA samples were treated with DNaseI; reverse transcription was performed using the MMLV-RT kit (Eurogen). The efficiency of DNaseI treatment and reverse transcription was tested by agarose gel electrophoresis and by qRT-PCR. The results of this verification were evaluated by the presence/absence of a PCR product in RNA samples before and after DNaseI treatment, and by observing the size of PCR fragments in RNA samples before treatment and its cDNA synthesis. Only those samples that confirmed the absence of genomic DNA contamination were included in further analysis of gene expression. This analysis included three groups of samples for each cultivar: the control group - before stress induction, and two experimental groups (cold and frost). To analyze expression differences between two cultivars we focused on the several genes which were previously reported to play important role in abiotic stress-response: ICE1, CBF1, DHN1, DHN2, DHN3 (Ban et al., 2017), NAC17, NAC26, NAC30 (Wang et al., 2016b), bHLH7, bHLH43 (Cui et al., 2018), WRKY2 (Wang et al., 2016a), LOX1, LOX6, LOX7 (Zhu et al., 2018), SnRK1.1, SnRK1.2, SnRK1.3 (Yue et al., 2015). Actin was taken as a reference gene (Table 1) and results were quantified using a Light Cycler 96 analyzer (Roche). The relative gene expression level was calculated by the Livak & Schmittgen (2001) using following algorithm: 2^−ΔΔCq, where: ${\displaystyle \Delta \Delta Cq={\left(C{q}_{geneofinterest}-C{q}_{internalcontrol}\right)}_{treatment}-{\left(C{q}_{geneofinterest}-C{q}_{internalcontrol}\right)}_{control}}$

Data analysis, visualization and relationship assessment

All analyses were repeated three times with three to five biological replicates. Statistical analyses were carried out using STATISTICA 6.0 software. One-way ANOVA and Student t-test were performed to determine significant differences between the effect of genotype and the respective treatments. For the correlation analysis, the algorithms of nonparametric statistics (Spearman coefficient) were used. The significance of the differences was evaluated by the Fisher test, LSD₀₅ and standard deviations from the mean value (Bailey, 1967).

Morphological assessment of tea cultivars

Morphological tests revealed that the thickness of the upper and lower epidermis were not significantly different between tolerant and sensitive tea plant genotypes. The total thickness of the leaf was ∼298.57 m observed in the tolerant cultivar Gruzinskii7, but only ∼235.25 m in the sensitive cv. Kolkhida (Fig. 2). However, there were significant differences between the two cultivars for thickness of spongy and palisade parenchyma. In cv. Gruzinskii7, both parenchyma layers were significantly thicker compared to the sensitive cv. Kolkhida. Gruzinskii7 was characterized by a lower stomata density as well as smaller stomata size than cv. Kolkhida.

Physiological response in tea under cold and frost

Cold treatment did not lead to changes in the CMI, RWC and pH of the cell sap in both tea cultivars. Frost treatment on the other hand resulted in a significant decrease in CMI in cv. Kolkhida, decreased RWC and increase in the pH of the cell sap in cv. Kolkhida, however; no significant changes in these parameters were observed in frost-tolerant cv. Gruzinskii7. Additionally, both cold and frost treatments resulted in increase in the water-soluble protein content by an average of three to four times with no significant differences between two cultivars (Fig. 3).

Soluble sugars content was elevated during low temperature induction and the highest concentration of sugars was reached in the tolerant cultivar during frost (Fig. 4A). Proline content was also increased significantly in both cultivars under cold induction in both cultivars (Fig. 4B). Sum of cations (NH₄⁺, Na⁺, K⁺, Mg²⁺, Ca²⁺) in the cell sap was elevated during cold and frost without significant differences between the cultivars (Fig. 4C). Among these five cations, K⁺, Mg²⁺, Ca²⁺ possessed the most pronounced changes during treatments. The highest Ca²⁺ elevation was observed in frost tolerant cultivar under Frost induction (Fig. 4E). The highest K⁺ elevation was observed under frost induction in sensitive cultivar (Fig. 4D). The increase of Mg²⁺ under treatments was not genotype-specific (Fig. 4F).

Due to cold exposure, six amino acids contents increased in comparison to the control (before stress induction) in cv. Gruzinskii7 as follows: serine: 3.24 folds, leucine and valine: 3.5 and 3.7 folds, respectively, glycine: 4.0 folds, threonine: 4.8 folds, and methionine: 5.3 folds. In the sensitive cv. Kolkhida cold exposure led to an increase only in two amino acids as follows: serine by 3.3 folds and methionine by 4.0 folds (Fig. 5).

Similarly, five amino acids increased due to frost induction compared to the control in cv. Gruzinskii7 as follows: leucine, valine, serine: 3.6–3.7 folds, threonine: 4.4 folds, and methionine: 6.55 folds. In sensitive cv. Kolkhida frost treatment resulted in increase of four amino acids: serine: 2.8 folds, valine: 3.2 folds, leucine: 3.8 folds and methionine: 4.2 folds (Fig. 5).

Relative gene expression in tea under cold and frost stress

All studied genes divided on three clusters. Cluster 1 included 11 genes with higher expression in the tolerant cultivar. Cluster 2 combined four genes with no difference between two cultivars and Cluster 3—three genes with higher expression in susceptible cultivar under stress conditions (Fig. 6).

The expression level of ICE1 gene under cold and frost increased by 1.5–1.78 folds in tolerant cultivar Gruzinskii7. Cultivar Kolkhida showed slight increase in expression of this gene with 0.88–1.01 folds under cold and frost. The accumulation of CBF1 transcripts was dramatically up-regulated by cold. The expression of this gene was genotype-specific and increased by 2400 and 907 folds in cv. Gruzinskii7 and cv. Kolkhida, respectively. After frost treatment, CBF1 expression slightly decreased in cv. Gruzinskii7.

The expression of WRKY2 in cv. Gruzinskii7 was extensively induced by cold. The expression of this gene increased by 200 folds, making it the second most important expression level after CBF1. In cv. Kolkhida, WRKY2 increased by 10 and 5 folds under cold and frost treatments, respectively.

Three genes DHN1, DHN2 and DHN3 exhibited elevated expression with significantly varied expression pattern. The highest expression of the DHN1 under cold was observed in cv. Kolkhida—by 34 folds. Frost treatment resulted in further increase of DHN1 expression in cv. Kolkhida but no elevation was observed in cv. Gruzinskii7. The significantly up-regulated expression of DHN2 and DHN3 was also observed in both cultivars in response to low-temperature treatment. The accumulation of the transcripts under cold treatment was 4 and 2 folds in DHN2 and DHN3 without significant variation between the two genotypes. However, frost treatment lead to the greater accumulation of DHN2 transcripts in cv. Gruzinskii7 and DHN3 transcripts in cv. Kolkhida.

Strong induction of NAC17, NAC26 and NAC30 transcripts was observed in response to cold and frost in both cultivars. Expression of these genes was elevated 4–8 folds (NAC17), 6–8 folds (NAC26) and 5–10 folds (NAC30) in cold and frost, respectively. The tolerant cv. Gruzinskii7 showed greater level of NAC17 and NAC26 expression.

The expression of SnRK1.1, SnRK1.2 and SnRK1.3 genes was up-regulated and varied in two cultivars under low-temperature treatment. Transcripts of these genes accumulated 2 –5 folds during cold and without further accumulation during the frost treatment. Significant differences between the two cultivars were observed in SnRK1.1 and SnRK1.3 expression. The greater accumulation of the transcripts was obtained in cv. Gruzinskii7. The expression profile of SnRK1.3 was not changed in cv. Kolkhida under cold and frost treatment.

Two genes of bHLH family were also extensively expressed in response to the low temperature induction. The accumulation of bHLH7 transcripts increased 3–4 folds under cold and frost, respectively without difference between the two genotypes. bHLH43 was strongly up-regulated in the tolerant cultivar in response to cold and frost. The accumulation of its transcripts in cv. Gruzinskii7 increased 4 folds comparing with cv. Kolkhida—1.5 folds.

Three studied LOX genes were intensively expressed in response to low-temperature treatment and their expression patterns varied significantly. LOX1 transcripts were accumulated 6 and 21 folds under cold and frost treatment, respectively with no difference between the two cultivars. However, LOX6 and LOX7 expression in response to cold was genotype-specific. LOX6 exhibited gradually increased expression pattern in the frost-tolerant cultivar with 3–7 folds accumulation under cold and frost, respectively. On the other hand, LOX7 showed higher expression level in cold-sensitive cultivar Kolkhida—26–30 folds elevation comparing with Gruzinskii7 6–9 folds.

The expression of the P5CS gene was also significantly induced by cold treatment. Greater level of the transcript accumulation was observed in cv. Gruzinskii7—16 –21 folds comparing with cv. Kolkhida—5–6 folds after cold and frost treatment, respectively.

Correlations of tea plant responses to low temperature stress

Positive correlations were observed between WRKY2 and ICE1 (r = 0.77). Strong correlations were observed between WRKY2 (r = 0.71) and threonine, as well as between ICE1 and threonine (r = 0.78).CBF1 correlated positively with methionine (r = 0.79) and serine ( r = 0.74) but negatively with RWC (r =  − 0.85). The P5CS gene correlated with methionine (r = 0.71). A positive correlation was also observed between the pH of the cell sap and leucine (r = 0.86) (Table 2).