Introduction

Salinization generates annual millionaire losses and it is estimated that about one third of irrigated lands worldwide have been affected by this problem (^{Schwabe et al., 2006}; ^{Shabala, 2013}). Agricultural production tends to be difficult in soils with electrical conductivities above 4 dS m^-1, because several cultivated species are sensitive to salinity (^{Jenks & Hasegawa, 2005}; ^{Duarte et al., 2013}). This situation imposes new challenges for farmers who have to deal with ecosystems that vary in soil type and quality as well as availability and quality of water resources (^{Lamz & González, 2013}). Salinity is one of the environmental factors that disrupt all or some of the biochemical processes in plants, which consequently limits the productivity and quality of agricultural crops throughout the world (^{Duarte et al., 2013}).

^{Flowers & Colmer (2008)} suggested the cultivation of plants tolerant to salinity, called halophytes, as an alternative in order to diversify production and take advantage of agricultural areas affected by this problem. These plants, representing a maximum of 2 % of terrestrial species, have adapted different physiological, biochemical and molecular strategies associated with their performance in saline environments that allow them to survive and grow normally even when there are high concentrations of salts in their rhizosphere, between 5 and 20 dS m^-1 of EC (^{Parida & Das, 2005}). The interest in aromatic and medicinal plants has increased in recent years, because from their metabolism are obtained compounds that are a unique source of pharmaceutical products, food additives, flavorings, aromas, among others, with antioxidant, antiviral, antibacterial and anticancer properties (^{Ramakrishna & Ravishankar, 2011}). Rosmarinus officinalis L. is distinguished for being a species well adapted to saline environments which has been classified as a plant moderately tolerant to salinity, whose production and commercialization is increasing due to its economic importance as an aromatic species for use in fresh and dry, seasoning, essence and its content of active ingredients (^{Westervelt, 2003}; ^{Miyamoto, 2008}).

Salinity causes symptoms related to the irreversible inhibition of growth since it slows down and does not reach completion. As a result, the leaf area, the size of the plant and the accumulation of dry matter are smaller (^{Campos et al., 2011}). It is reported that saline stress directly or indirectly inhibits cell division and elongation of the cells of the root organs, stems and leaves (^{Zidan et al., 1990}).

The reduction of biomass is attributed, mainly, to the fact that salt stress affects the photosynthetic rate due to a low potential in the soil solution, ion toxicity and nutritional imbalances (^{Munns, 2002}). On the other hand, it impacts the production and accumulation of secondary metabolites of medicinal and aromatic plants (^{Beretta et al., 2011}; ^{Jordán et al., 2013} and ^{Zaouali et al., 2013}). Rosmarinic acid is one of the main phenolic compounds found in the tissues of species of the Lamiaceae family, this is the reason why they are considered as a valuable source of these compounds; in general, all its extracts have a significant antioxidant activity (^{Trivellini et al., 2016}).

Although the salinity in soils is not exclusively due to NaCl, research on the effect of different ion sources on the development of aromatic plants, as well as their effect on secondary metabolism, is scarce. Additionally, in many cases plant material in which biological properties of extracts of Rosmarinus officinalis L. had been determined has been in populations with little mention of the agronomic management they received, and given the importance of these compounds in both industry as well as for its beneficial effects on human health, it is important to understand the response of the accumulation of these compounds under controlled conditions that may favor their production.

Based on the above, the objective of the present investigation was to evaluate the vegetative growth and production of secondary metabolites in Rosmarinus officinalis L. grown in two protected environments and different salt concentrations in the nutrient solution.

Material and Methods

This research was developed during the period from April to June 2016 at Unidad Academica de Agricultura of the Universidad Autonoma de Nayarit, located at 21° 25’ 36” latitude N and 104° 53’ 28” longitude W, at 922 msnm, in a single unit Gothic multispan greenhouse covered with milky plastic with 30 % shadow and in a shade-enclosure macro-tunnel type covered with mesh of 35 % shade. Three-month-old rosemary plants of the “Arp” variety were grown in 20 x 20 black polyethylene containers with tezontle as substrate. A completely randomized experimental design with a 2 x 7 factorial arrangement was used. The factors evaluated were two production environments and seven salinity levels of the nutrient solution, generating a total of 14 treatments (Table 1). Five repetitions were made per treatment. The experimental unit consisted of a plant in a container. Each plant was manually irrigated with 250 mL every three days. The harvest of vegetative material was made in the morning, between 9 and 11 am, at 18 and 36 days after the beginning of the application of the nutritive solutions.

Two groups of variables were evaluated, with measurements at 18 and 36 days after transplant (DAT). The first group was plant height (cm) and branches·plant^-1, the following group of variables was formed by the proline content (µmol·g^-1), content of total phenols (μg CAE/100 g) and the antioxidant activity index (AAI) (% inhibition of DPPH).

Plant height was measured with a ruler based on the base of the plant and the apex of the main stem as the maximum height, while the number of branches⋅ plant^-1 was counted from the base of the stem to the apical meristem.

Proline content was determined following the procedure of ^{Bates et al. (1973)}, 0.5 g of pulverized lyophilized sample were used, macerated with 5 mL of 3 % sulfosalicylic acid; the extract was filtered and then centrifuged at 10,000 rpm for 5 min. An aliquot of 2 mL of the supernatant was measured and placed in a test tube then added 2 mL of acid ninhydrin and 2 mL of glacial acetic acid. The blank was prepared with 2 mL of sulfosalicylic acid, 2 mL of acid ninhydrin and 2 mL of glacial acetic acid. The standard curve was prepared with different concentrations of standard proline solution (20 μmoles mL^-1); from this solution an aliquot of 1 mL was measured and deposited in a flask and adjusted to 50 mL with 3 % sulfosalicylic acid. The mentioned dilution is equivalent to 400 nmoles·mL^-1. Each tube was vortexed until obtaining an emulsion. Then they were covered and placed in a double boiler for 60 minutes. When finished, they were submerged in cold water for 10 minutes. 4 mL of toluene were added to each tube and then vortexed. The upper phase was placed in a new tube and the absorbance reading was performed at 520 nm in a Thermo Scientific spectrophotometer (Model Spectronic 200, Massachusetts, USA).

The proline content is expressed in μmol·g^-1 of proline, based on the following equation:

Where: Abs extract is the absorbance obtained from the extract, blank (expressed in absorbance) and slope (expressed as absorbancemnmol^-1) determined by linear regression, Vol extract is the total volume of the extract, Vol aliquot is the volume used in the test, FW (expressed in mg) is the amount of plant material in which extraction was carried out. It is assumed that the Abs extract is within the linear range.

Total content of phenols and AAI were evaluated in fresh material. The samples were prepared according to the procedure of ^{Chizzola et al. (2008)} and ^{Nourhene et al. (2009)}; 60 % methanol (v/v) was used, 2 g of fresh material were treated with 15 mL of solvent and the extraction was carried out at 4 °C for 24 h. The extract was then filtered to separate the particles of plant material and refrigerated at 4 °C until analysis.

The total content of phenols was determined with the Folin-Ciocalteau method, according to ^{Chizzola et al. (2008)} with some modifications proposed by ^{Juárez et al. (2011)}. The reagents used were: caffeic acid, Folin-Ciocalteau reagent, sodium carbonate and ethanol. Aliquots of 0.5 mL of ethanolic extract, 1 mL of 95 % ethanol (v/v) were measured to which 5 mL of distilled water were added, and 0.5 mL of Folin-Ciocalteau reagent diluted in distilled water 1:10. After 5 min, 1 mL of sodium carbonate solution in water (5 %, v/v) was added. The samples were shaken and maintained during 30 min in the dark, then the absorbance at 725 nm was measured in a Thermo Scientific spectrophotometer (Model Spectronic 200, Massachusetts, USA). The blank was prepared following the same procedure with ethanol. Different concentrations of caffeic acid in ethanol were used for the calibration curve. The total content of phenolic compounds in the extract was expressed in μg equivalents of caffeic acid (CAE) per 100 g of fresh plant material (FPM).

The AAI of the phenolic extract was determined with the DPPH method described by ^{Chizzola et al. (2008)} and ^{Scherer & Texeira (2009)}. For each sample, an aliquot of 400 μL of the extract was measured and adjusted to 1 mL with 50 % methanol, then 1 mL of DPPH (2,2 diphenyl-1-picrylhydracil) (2.43 x 10^-4 M) was added. The samples were placed in darkness for 30 min at room temperature; the absorbance against a blank was measured at 517 nm in a Thermo Scientific spectrophotometer (Model Spectronic 200, Massachusetts, USA). The blank consisted of 500 μL of Trolox, 500 μL of methanol and 1 mL of DPPH reagent in order to obtain a total discoloration of the radical. As a reference substance for the calibration curve, 2.5 mM Trolox (6-hydroxy-2,5,8-tetramethyl-chroman-2-carboxylic acid) in methanol was measured; the concentrations for the curve were 0.1 to 2 mM of trolox in 1 mL of methanol. The standard Trolox solution was prepared under the same conditions. The results are expressed in percentage of inhibition of DPPH, according to the following equation:

Where Absc is the absorbance value of the control, Absm is the absorbance value of the sample.

Analysis of variance of each factor and variable was performed with the statistical package SAS (SAS, Inst., 2007). The comparison of means was conducted by means of the Tukey test (p<0.05).

Results and Discussion

The analysis of variance indicated significant differences at 18 DAT due to the effect of the environment on the height of the plant and highly significant differences in the number of branches⋅plant^-1. Significant differences were also obtained on the variables proline, total phenols and AAI due to the effect of the production environment, the nutrient solution and the interaction among them. The analysis of variance at 36 DAT did not indicate significant differences in the evaluated growth variables, except for plant height due to the effect of the environment interaction by nutrient solution. However, in all physiological variables evaluated, highly significant differences were found due to the effect of the environment, the nutritive solution and the environmental interaction by nutritive solution (Table 2).

Height. At 18 DAT, the highest averages were found in plants grown in the greenhouse (Table 3); In this regard, it is possible that the initial light intensity perceived in that environment has stimulated a greater accumulation of carbohydrates (^{Lambers et al., 2008}). However, the crop followed the same growing rate as long as the experiment was run in both production environments. When considering the effect of the nutrient solution on the height, the analysis of variance allows to observe a homogeneous growth at 18 and 36 DAT, the averages oscillated between 32.22 and 35.05 cm (Table 3 and 4). These values are higher than those reported by ^{Kiarostami et al. (2010)} when they used NaCl in the growing medium; these differences could be attributed to the suboptimal concentration of some nutrients that accompanied the NaCl treatments in this trial.

Number of branches. This variable was different between production environments at 18 DAT. The highest values were found in plants grown in shade-enclosure; however, at 36 DAT there were no differences observed due to environment effect (Table 3 and 4). The initial difference observed could had been influenced by the light conditions allowed by the cover of the shade-enclosure and its effect on the growing rate, since the formation of new leaves allows to intercept a greater percentage of radiation (^{Silber & Bar-Tal, 2008}). No differences were observed due to the nutrient solution effect. The results of this variable were similar to the trend observed for the variable plant height.

In this work, it is partially demonstrated that Rosmarinus officinalis L. is a plant moderately tolerant to salinity as indicated by ^{Miyamoto (2008)} and ^{Tounekti et al., (2008)}, since the saline stress promoted by nutritive solutions added with NaCl and high EC during 36 days of treatment did not reduce growth.

Proline. The proline content was different in the sampling dates in both production environments. Rosemary plants grown in the greenhouse recorded the highest proline content at 36 DAT with 2.067 μmol/g (Table 4). These values differ from those reported in Thymus vulgaris L. where the proline content increased in field conditions compared to shade-enclosure conditions, which suggests that the conditions in the greenhouse somehow intensified the induced stress, which stimulated a greater synthesis of this metabolite (^{Zrig et al., 2016}). In addition to acting as osmolyte, under various conditions of abiotic stress, proline functions as a molecular chaperone that preserves the integrity of proteins and membranes, stabilizes the pH of the cytosol and neutralizes reactive oxygen species (^{Kishor et al., 2005}; ^{Hayat et al., 2012}).

Due to the effect of the nutritive solution, statistical differences were observed at 18 and 36 DAT in the proline content (Table 3 and 4). During the period of the experiment, the lowest values in proline content were obtained in treatment 1, while the highest values were found in treatment 3 and 4 at 36 DAT (Table 4). The above indicates that the osmotic effect caused by NaCl is different from that caused by an excessive concentration of all the nutrients. Although the observed tendency of the proline to increase as the salinity increased coincided with other medicinal species when growing under high concentrations of NaCl, the proline values obtained in this research are higher than those reported in Mentha piperita, but lower than those of Satureja hortensis and Matricaria chamomilla (^{Roodbari et al., 2013}, ^{Akbari et al., 2013}; ^{Afzali et al., 2009}).

In conditions of salinity, plants need strategies to survive in hostile environments. One of the most efficient physiological mechanisms to survive stress conditions is osmotic adjustment in which tissues reduce their osmotic potential by accumulating a variety of metabolites which allow them to maintain turgor (^{Hayat et al., 2012}; ^{Wu et al., 2013}; ^{Zrig et al., 2016}). At the beginning of the experiment, a progressive increase in the proline content was observed as the concentration of the nutrient solution increased. In addition, the decrease observed over time allows to infer that the age of the plant influences its ability to synthesize and regulate the concentration of this amino acid during the time that stress lasts (^{Xu et al., 2014}).

Total content of phenols. This variable was different among production environments, where rosemary plants grown in the greenhouse obtained the highest values in the samples (163,996 - 178,244 μg CAE/100 g) (Table 3 and 4). These values differ from those of ^{Zrig et al. (2016)}, who reported 6.46 mg gallic acid g^-1 PF (average) of total phenols in Thymus vulgaris cultivated in shade house, with a contribution of 500-700 mM m^-2 s^-1 PAR. These last conditions of luminosity were similar to those recorded in greenhouse in this work, which suggests that total phenolic content could be sensitive to light. In this regard, ^{Ghasemzadeh et al. (2010)} observed that the variation of the radiation levels influences the accumulation and distribution of phenols in Zingiber officinale in which higher results were obtained with 460 - 790 μmol m^-2 s^-1 (34.16 - 39.06 mg gallic acid g^-1 PS).

On the total phenolic content in rosemary plants, significant differences were observed from 36 DAT on, where the maximum value was reported in treatment 1 (173.789 μg CAE/100 g) and the minimum in treatment 2 (157.605 μg CAE/100 g) (Table 4). The values recorded in this work are higher than those of ^{Chizzola et al. (2008)} and ^{Juárez et al. (2011)} in Thymus vulgaris (65.1 and 68.05 μg CAE/100 g, respectively), which suggests that the cultivation of rosemary is a species with a high production of phenolic compounds that does not require the manipulation of the nutrient solution with excess of salts to obtain a higher total phenolic content in rosemary cultivation.

AAI. This variable was different from 36 DAT on, where rosemary plants grown in the greenhouse environment recorded the highest values (69.02 %) (Table 4). This response was possibly a function of the photosynthetically active radiation perceived in the environment. This trend is contrary to that reported by ^{Ghasemzadeh et al. (2010)} in extracts of Zingiber officinale, where a decrease in antioxidant activity was observed when the luminous intensity increased. Given the biological importance of this type of compounds and their role as antioxidants in human nutrition, the use of different controlled environments for their production should be considered.

Due to the effect of the nutrient solution, statistical differences were observed on the AAI of rosemary plants during the time the experiment lasted. At 36 DAT there were statistically significant differences for treatments 3 and 4 in comparison to the other treatments, which obtained the lowest values (44.9 and 41.4 %) (Table 4); In this regard, it is documented that salinity generates oxidative stress, which means it accelerates the oxidation of a biological system; antioxidants reduce the adverse effects of reactive oxygen species. It is possible that these values were due to the oxidative stress induced by the high concentrations of all the nutrients present in the substrate solution, which stimulated a greater production of reactive oxygen species (^{Miller et al., 2010}).

In this investigation, Rosmarinus officinalis showed a decrease in the AAI as it increased the salinity of the growing medium, a trend similar to that observed by ^{Kiarostami et al. (2010)} and ^{Oueslati et al. (2010)}, which tends to change significantly as it increases the concentration of the nutrient solution and has a differential response in terms of the salt that causes it. Treatment 1 was maintained with the highest percentages of inhibition, which was a stress-free condition, while the lower values were recorded in treatments 3 and 4 with 44.94 and 41.47 % (Table 3 and 4). However, the percentages of DPPH inhibition observed in this experiment were superior to those reported by ^{Chizzola et al. (2008)} (22-55 %) and ^{Juárez et al. (2011)} (43.88 %) in Thymus vulgaris. Because of its properties and structure, phenolic compounds are complex extracts that present an important antioxidant activity, which is given by the sum of the antioxidant capacities of each of its components, the interaction between them and the environment in which they are met. Eventually, potentiating or inhibiting effects may occur (^{Frankel & Meyer, 2000}). The results obtained suggest that it is not advisable to induce saline stress in rosemary plants to produce antioxidant compounds of natural origin.

Conclusions

Neither the production environment, shade-enclosure or greenhouse, nor the formulation of the nutrient solution had any effects on the height of the plant and the number of branches of rosemary. The accumulation of proline in rosemary, the total phenolic content and the antioxidant activity index were highly influenced by the growing environment, being higher in the greenhouse.

The accumulation of proline was influenced by the composition of the nutrient solution, the highest values were obtained with the nutrient solution of Steiner at 640 and 870 %, while the total phenolic content and the antioxidant activity index were greater with the nutrient solution of Steiner at 75 %.