Agronomic traits, secondary metabolites and element concentrations of Lavandula angustifolia leaves as a response to single or reiterated drought stress: How effective is the previously experienced stress?

Sixty uniform transplants of L. angustifolia were kindly provided by the Agricultural Application and Research Centre, Igdir University, Turkey. The transplants were 1 year old, and the height and life size of the transplants were homogeneous. The selected transplants were grown in a 2-L pot with a mixture of soil/peat/perlite (2:1:1, by volume). Before starting the experiments, all pots were transferred to the greenhouse in order to acclimatise the plants for 15 days. The plants were grown under the following conditions: 14-h photoperiod; mean temperature: 26−30 °C in the day, 16−20 °C in the night; and relative humidity: 60%–70%. Following the acclimatisation period, the lavender transplants were subjected to drought and full irrigations (details are provided in the section on “Application of drought treatments”). The properties of the experimental soils were as follows: saturation: 64.9%; pH: 7.57%; salt: 0.16%; CaCO_3: 8.41%; organic matter: 4.29%; K: 670 mg · kg⁻¹ and P: 19.87 mg · kg⁻¹.

The experimental design regarding the drought application treatments (single or double stress) are presented in Table 1 and Figure 2. The experimental design of Kulak (2020) and Pintó-Marijuan et al. (2017) with some minor modifications was used for the current study. Changes in the basic growth parameters, mineral content and secondary metabolites of lavender were investigated as the main targets of the present work. In the study, the measurements and sample collections were done on Day 0, Day 11, Day 17 and Day 28. Day 0 was the first day of sampling and measurement of the agronomic attributes, secondary metabolites and mineral content of the leaves of lavender, just before the drought stress was applied. According to the soil water content (SWC), the lavender plants were exposed to three cycles of drought stress for 11 days by completely ceasing irrigation. The lavender plants reached the wilting point after an 11-day drought stress, followed by subsequent drought-stress recovery for 5–6 days. In the work of Kulak (2020), the same lavender species was reported to be tolerant to a 7-day stress period. Due to the differences in experimental soils or background experiences of the lavender, the period of the stress lasted for 11 days. During the period of stress treatment of lavenders, control plants were subjected to full irrigation for 11 days.

Figure 2

This stage of drought application as a single stress was termed as the “stress stage” (Cycle 1). The “recovery stage” was namely noted as “Cycle 2”. After Cycle 2 (recovery stage), the drought-subjected and fully irrigated plants were divided into two more subgroups. In this context, four experimental groups were obtained. For each group, half of the plants were fully irrigated and the remaining half were subjected to drought for 11 days, denoted as Cycle 1 (Figure 2). At the end of each cycle, five plants from each group were randomly selected for the relevant analysis.

The gravimetric SWC (SWC_grav) was estimated according to Du and Rennenberg (2018). In this regard, SWC_grav was calculated gravimetrically after each harvest, being expressed on a dry weight basis using the following formula: SWCgrav(g H2O⋅g−1 DW)=(FW−DW)/DW {\rm{SW}}{{\rm{C}}_{{\rm{grav}}}}\left( {{\rm{g}}\;{{\rm{H}}_2}{\rm{O}} \cdot {{\rm{g}}^{ - 1}}\;{\rm{DW}}} \right) = \left( {{\rm{FW}} - {\rm{DW}}} \right)/{\rm{DW}} where FW denotes the fresh weight of the soil before drying, and DW denotes the dry weight of the soil after drying at 105 °C for 48 h.

Leaf hydration (g H₂O · g⁻¹ DW) was calculated as (FW – DW)/DW, being expressed on a dry weight basis as in the case of SWC_grav, where FW is the fresh mass and DW is the dry mass after drying the samples in an oven at 60 °C for 72 h (Du and Rennenberg, 2018).

In order to reveal the changes in basic agronomic traits, we estimated the following in a total of 15 plants, corresponding to five plants for each replicate: leaf fresh weight, leaf dry weight, leaf rehydration, SWC, leaf length, leaf width, stem length, stem fresh weight, stem dry weight, root length, root fresh weight and root dry weight.

The extraction of leaf samples was carried out according to the modified method used in the study by Celikcan et al. (2021). All chemicals used in the study were obtained from Sigma-Aldrich, St. Louis, MO, USA. In this regard, a shaker-aided sequential extraction was performed at 120 rev · min⁻¹ for 24 h at room temperature. Briefly, 3 g of finely dried leaf samples were extracted using 50 mL of methanol. The same extraction follow-up was repeated three times, and the extracts were filtered; the filtrates were collected and evaporated using a rotary evaporator (Heidolph, 94200, Bioblock Scientific, Schwabach, Germany). Until the HPLC analysis of phenolic compounds, the vacuo-dried samples were preserved at +4 °C; samples with 0.5 mg · L⁻¹ concentration were prepared.

The methanol extracts of the lavender leaves were filtered through a 0.45-μm disc prior to HPLC analysis. Of the phenolic compounds, ascorbic acid, gallic acid, catechin, vanillic acid, caffeic acid, p-coumaric acid, ferulic acid, o-coumaric acid, rosmarinic acid, salicylic acid, quercetin and kaempferol were monitored and quantified for each sample corresponding to the treatments. In this context, a (high-performance liquid chromatography (HPLC)) system (Agilent 1260; Agilent, Santa Clara, CA, USA) equipped with a diode array detector was used. The separation of the compounds was done with 10 μL of extract on a column (4.6 × 250 mm, 5 μm; ACE Generix 5C18 (GEN-7444), Scotland) thermostatted at 30 °C. The mobile phases were as follows: (A) 0.1% phosphoric acid in water and (B) HPLC-grade 100% acetonitrile. The phenolics were quantified by comparing the peaks recorded at 300 nm with the standard curves of each acid. The results were expressed as nanograms per microliter (ng · μL⁻¹).

For the essential oil extraction, approximately 0.5 g of dried leaf samples was used. Gas chromatography (GC) headspace conditions were as follows: GC cycle time: 50 min; sample volume: 3.0 mL; incubation time: 25 min; incubation temperature: 70 °C; syringe temperature: 70 °C. After optimising the running conditions, the GC apparatus equipped with an HP-5 mass spectrometry (MS) capillary column (30 m × 0.25 μm × 250 μm) and 5977 (Agilent Technologies) with mass selective detector 7890B (Agilent Technologies, Santa Clara, United States) model GC–MS was used for determining the essential oil composition of the leaf samples. An electron ionisation system with ionisation energy of 70 eV was used, and the flow rate of the carrier gas (helium) was set to be 1.0 mL · min⁻¹. Injector and MS transfer line temperatures were set at 250 °C. Column temperature was initially kept at 50 °C for 2 min, then gradually increased to 200 °C at the rate of 5 °C · min⁻¹ and ultimately increased to 250 °C at 10 °C · min⁻¹. Samples were injected automatically with split ratio 2:1. Analyses lasted for 35 min. The relevant compounds were identified with electronic libraries using reference compounds from the NIST08, Willey7n.1 and HPCH1607 libraries.

Leaf mineral content was estimated according to the method of Kaçar and Inal (2010). Briefly, fully developed leaves (from the second or third nodes) were first washed with double-distilled water and then were left for drying at 70 °C for 48 h. One gram of dried and finely powdered leaves was extracted using 3 mL 65% HNO₃ and 1 mL 30% HCl. The obtained solution was digested in a microwave, and the process was terminated by cooling for 45 min. Ultimately, the solutions were filtered, and the filtrates were made up to 50 mL with addition of double-distilled water. Until further analysis using inductively coupled plasma atomic emission spectroscopy (ICP-OES) (Optima 2100 DV; Perkin Elmer Inc., Waltham, MA, USA), the filtrates were preserved at 4 °C.

For each treatment, three replicates corresponding to 15 plants were used. The experimental data were subjected to one-way variance analysis. The means were separated using Duncan's multiple range test at the 5% probability level (p < 0.05) (SPSS22). Due to the high number of dependent and independent variables, principal component analysis (PCA) and heat map clustering were carried out (XLSTAT software and ClustVis, respectively) to visualise, correlate and discriminate the variables.

The values of SWC_grav significantly decreased from 0.39 g H₂O · g⁻¹ DW to 0.11 g H₂O · g⁻¹ DW and 0.39 g H₂O · g⁻¹ DW to 0.20 g H₂O · g⁻¹ DW in the stress and the control groups, respectively, after withholding water supply during the first cycle. In the second cycle, in the recovering stress groups, the SWC_grav increased from 0.11 g H₂O · g⁻¹ DW to 0.40 g H₂O · g⁻¹ DW. Reiterated drought stress sharply decreased the SWC_grav from 0.40 g H₂O · g⁻¹ DW to 0.048 g H₂O · g⁻¹ DW. Moreover, application of a single stress to the control group plants after the second cycle sharply decreased the values from 0.52 g H₂O · g⁻¹ DW to 0.058 g H₂O · g⁻¹ DW. Leaf hydration was positively correlated with SWC_grav (r = 0.840, p < 0.01) and decreased from 2.08 g H₂O · g⁻¹ DW to 0.88 g H₂O · g⁻¹ DW during stress (first cycle). During the second cycle, recovery significantly increased the leaf hydration status of the stress group plants. Repeated drought stress, again, caused sharp reductions. Similarly, the lowest values of leaf hydration were obtained with applying single stress to control groups, i.e. from 0. g H₂O · g⁻¹ DW to 0.058 g H₂O · g⁻¹ DW.

By exposing lavender plants to reiterated drought under greenhouse conditions (including three cycles of 11 days of drought by withholding water, followed by subsequent periods of 6 days of recovery, and then double-stressed and single-stressed periods), it was observed that stem dry weight and root length did not significantly differ (p = 0.259 and 0.169, respectively) among the single-stressed, recovered and double-stressed plants. However, other parameters such as leaf fresh weight, leaf dry weight, leaf length, leaf width, stem length, stem fresh weight, roof fresh weight and root dry weight were significantly affected by the relevant stress treatments (Table 2). Specifically, in the first cycle (single-stressed versus fully irrigated conditions, Cycle 1), drought stress reduced the values of leaf fresh weight, leaf width, leaf length, leaf width, stem length, root fresh weight and root dry weight (p < 0.05). It is interesting to note that the responses of leaf dry weight (a decrease) and root fresh weight (an increase) were only significant after recovery, i.e. in the second cycle (Cycle 2), while the other parameters did not differ between single-stressed and recovered plants. In the third cycle, there were no significant differences in leaf fresh weight, leaf dry weight, leaf length, leaf width, stem length, stem fresh weight, root length and root dry weight due to the previously experienced stress periods. Only root fresh weight significantly decreased due to the previously experienced stress period. In the third cycle, it is also interesting to note that applying drought to the fully irrigated plants caused critical reductions in leaf fresh weight, leaf dry weight, leaf length, leaf width, stem length, stem fresh weight, root fresh weight and root dry weight. For these reasons, the differences between irrigated treatments were significant. These findings might suggest that stress-subjected lavender plants had prepared themselves for the reiterated stress (double stress) in the third cycle (Cycle 3).

As a powerful tool for the discrimination of relevant agronomic traits, heat map clustering was used for reducing the dimensions of the variables, visualising and correlating the findings. Accordingly, two major clusters were noted (Figure 3A). Considering the agronomic traits corresponding to the stressed and non-stressed groups, the first one was stem length, root length, leaf length, leaf dry weight and stem dry weight. On the other hand, leaf hydration, SWC, root fresh weight, root dry weight, leaf fresh weight, leaf width and stem fresh weight were classified into the second major cluster. Regarding the clustering of the stressed and non-stressed experimental groups, here also, two major clusters were obtained. In the first major cluster, stress recovery, sampling at Day 0, stress control and control recovery groups were observed, while the other groups were classified under the second major group. However, in a sub-cluster of the second major cluster, control-stress and stress-stress were observed in the same groups. These findings might suggest that post-drought stress in the control (stress to the full irrigated plants in the third cycle) caused damage or exhibited the same effects as in the double-stressed plants, or we might note that stress priming might prepare the plants for the possible emerging stresses. Furthermore, along with PCA, a better and clear discrimination in the growth and biomass production parameters was revealed on the 2D visualisation of the plotted scores, where the two principal components accounted for 84.17% of the total variance (Figure 4A). The first and second axes explained 62.38% and 21.78% of the total variance, respectively. It is worthy to note that retaining irrigation for stressed plants after the recovery stage moved the stressed groups into similar groups with the control plants. Interestingly, withholding water supply to the control plants after the recovery stage moved the control group into a similar group with the stressed plants, as in the case of the sub-cluster of the second major cluster.

Figure 3

Figure 4

Corresponding to the relevant treatments, it was noted that the responses of the major (Ca, K, Mg and P) and trace elements (B, Mo, Fe, Zn and Mn) in the leaf tissues of lavender were statistically significant (p = 0.000), except Cu (p = 0.121). Especially, drought caused substantial increases in the concentrations of Ca, Mo, Fe, Zn and Mn, while significant reductions in the concentrations of K, P, Mg and B were noted in stress-subjected lavenders at the end of Cycle 1. In Cycle 2, recovery in the stress-subjected plants decreased the concentrations of Ca, B, Mo and Zn, while it increased the concentrations of K, P, Fe and Mn. However, the concentrations of Mg and Cu did not differ between single-stressed and recovered plants. In the third cycle, due to previously experienced stress periods, reiterated stress decreased the concentrations of Ca, Mg, B and Mo, while it increased the concentrations of K, P and Cu. However, among the trace elements, the concentrations of Fe, Zn and Mn did not differ between single-drought stress and reiterated drought stress. Corresponding to the application of stress to the fully irrigated plants after the recovery stage (Cycle 2), drought stress caused substantial increases in the concentrations of Ca, K, P, B, Mo, Zn and Mn, while it decreased the concentrations of Mg and Fe. As in the case of stress-subjected plants from the first cycle, the concentration of Cu did not differ between the full-irrigated and stress-subjected control plants. It is interesting to note that, among the elements quantified, the responses of only K and P were the same against the stress conditions corresponding to the control and stress-experienced plants exposed to the same stress after the recovery stage. The remaining elements behaved relatively differently in stress-experienced and non-stress-experienced plants (Table 3).

As in the case of agronomic traits, a heat map was constructed for clustering and correlating the mineral content and experimental groups. Regarding the elemental contents corresponding to the experimental groups, two major clusters were observed (Figure 3B). The first one was composed of Mg, Mo and B, while the later included P, K, Ca, Cu, Zn, Fe and Mn. Considering the experimental groups, two major clusters were, again, revealed. The first cluster, except the control-recovery group, was mainly composed of stress-related groups, while the second cluster, except the group subjected to stress after 11 days, was based on the control groups. In relation to mineral content discrimination, a partial and better discrimination was observed for the stress and non-stressed groups. Moreover, the relevant values were subjected to PCA, and according to the PCA, two principal components accounting for 65.97% of the total variance (F₁: 39.19%, F₂: 26.78%) were noted (Figure 4B).

The essential oil compounds identified in lavender leaves are listed in Table 4, following their elution order on the HP-5 column. The compounds delta-3-carene, camphene, 2(10)-pinene, 4(10)-thujene, o-cymene, D-limonene, eucalyptol, camphor and endo-borneol were screened in the leaves of the stressed and non-stressed groups. In the stress-subjected and fully irrigated treatments, all the identified compounds, except eucalyptol (p = 0.201), were substantially affected (p = 0.01). Of the identified compounds, eucalyptol, camphene, camphor and endoborneol were found to be the major constituents based on their percentage. With respect to the treatments, drought caused significant reductions in the percentages of camphene and endo-borneol, while the camphor percentage increased significantly in stress-subjected lavenders at the end of Cycle 1. In Cycle 2, recovery in the stress-subjected plants increased the percentage of camphene and endo-borneol and decreased the percentage of camphor. In Cycle 3, being previously stress experienced did not substantially affect the percentage of camphor between single-stress-subjected and reiterated-drought-stress-subjected plants, but prior stress experience increased the percentage of camphene and endo-borneol. With respect to the application of stress in full-irrigated lavenders after the recovery stage, drought stress reduced the percentage of camphene and camphor, in relation to the full-irrigated lavenders, while increasing the percentage of endo-borneol. Similar to the behaviour of elements, relatively, some differences were noted between the plants due to the previously imprinted stress. As reported earlier, it is interesting to note that the predominant compound eucalyptol, with an approximately estimated percentage ranging between 62.67% and 69.20%, was not substantially affected by the treatments.

Heat map clustering analysis separated the essential oil compounds into two major clusters. The first cluster was composed of camphene, camphor, D-limonene, 4(10)-thujene and o-cymene, while delta-3-carene, endo-borneol, 2(10)-pinene and eucalyptol were classified in the second cluster (Figure 3C). Regarding the stressed and non-stressed groups, here also, two major clusters were constructed. The first cluster included the following groups: stress-recovery, stress after 11 days, control-recovery and sampling at Day 0. The second cluster was composed of stress-stress, control after 11 days, control-stress, control-control and stress-control. Coupled with heat map clustering, PCA was used to discriminate the essential oil compounds corresponding to the stressed and non-stressed groups. Accordingly, two principal components accounting for 63.34% of the total variance (F₁: 34.84%, F₂: 28.49%) were obtained (Figure 4C).

Of the large and diverse number of phenolic acids, the contents of ascorbic acid, catechin, ferulic acid, o-coumaric acid, rosmarinic acid, quercetin and kaempferol were quantified using HPLC (Table 5). Except kaempferol (p = 0.201), the other quantified compounds were significantly affected by the relevant treatments (p = 0.00). It is worthy to note that the responses of the compounds against stress conditions were relatively different. Especially, in Cycle 1, drought caused significant increases in the content of ascorbic acid, catechin, caffeic acid, o-coumaric acid and rosmarinic acid, in relation to the full-irrigated plants, while the contents of ferulic acid and quercetin decreased. In Cycle 2, recovery in the stress-subjected plants increased catechin, caffeic acid, ferulic acid, o-coumaric acid and rosmarinic acid, while decreasing the content of ascorbic acid. Only quercetin content was not affected by the recovery. In Cycle 3, being previously experienced in stress decreased the content of ascorbic acid, catechin, ferulic acid and rosmarinic acid, while increasing the content of caffeic acid and o-coumaric acid. However, quercetin content was not affected substantially. Furthermore, subjecting the full-irrigated plants to stress after the recovery stage increased the content of ascorbic acid and o-coumaric acid, relative to the full-irrigated plants, but it decreased the content of catechin and rosmarinic acid. The other compounds, such as caffeic acid, ferulic acid and quercetin, were not substantially responsive to stress application.

As in the case of other attributes estimated in the study, the same scattering, clustering and discrimination tools were used for individual phenolic acids. Of these tools, the heat map clustered the relevant acids into two major groups (Figure 3D). Phenolic acids, viz. caffeic acid, kaempferol, ascorbic acid and o-coumaric acid, were included in the first major cluster, while the second major cluster included catechin, rosmarinic acid, ferulic acid and quercetin. Considering the experimental groups, the stressed and non-stressed groups were clearly discriminated from each other. In addition, PCA, in addition to the heat map clustering, also explained a high ratio of the total variation (F₁: 48.36% and F₂: 27.91%, with a total variance: 76.28%) (Figure 4D).