Introduction

Cistus heterophyllus subsp. carthaginensis is an Iberian-North African endemism that appears on the Spanish Red List of Threatened Plants as critically endangered (Sánchez-Gómez et al. 2002). The only known European populations of C. heterophyllus, comprising only 3 and 22 known individuals, respectively, are in the provinces of Valencia and Murcia (Spain). Due to their naturally high phenolic antioxidant content, several Cistus species have been used in folk medicine as antiinflammatory, antiulcerogenic, antimicrobial and antispasmodic agents (Barrajón-Catalán et al. 2010; and references herein). However, no previous studies have specifically addressed the phytochemistry and biological activity of C. heterophyllus.

As well as being essential for the conservation of biodiversity, in vitro culture has been shown to be effective in stimulating the accumulation of phytochemicals in plant cells by elicitation with biotic and abiotic elicitors as well as with stress-related mediators such as methyl jasmonate or salicylic acid (SA) (Matkowski 2008).

SA is a multifaceted hormone that plays an important role in the induction of stress tolerance in plants (Horváth et al. 2007; Vlot et al. 2009; Gadzovska et al. 2012). The exogenous application of SA not only provides protection against various biotic and abiotic stresses, but also enhances the growth and productivity of plants (reviewed by Hayat et al. 2010; Rivas-San Vicente and Plasencia 2011). In particular, several authors (Chen et al. 2006; Kováčik et al. 2009a) have revealed that SA provokes changes in the gene expression related with the biosynthesis and production of phenylpropanoids, such as those of phenylalanine ammonia-lyase (PAL), the first enzyme of the overall phenylpropanoid biosynthetic pathway, and class III plant peroxidases (EC 1.11.1.7; hydrogen donor: H2O2 oxidoreductase, Prxs), a key polymorphic group of heme-containing glycosylated enzymes that can oxidize a wide variety of phenolic compounds using hydrogen peroxide as the electron acceptor (Passardi et al. 2005; Almagro et al. 2009), resulting in changes in plant metabolome.

Since phenylpropanoids are required for plant growth, development and adaptation (Vogt 2010), and since, furthermore, many of these compounds, especially polyphenolics, are powerful antioxidants whose positive health-related effects in many oxidative-stress related diseases have been repeatedly reported (Soobrattee et al. 2005; Valko et al. 2007), an increase in the production of antioxidative polyphenols could be a promising approach for improving both plant vigour and their content of medicinally valuable compounds (Amoo et al. 2012, ; Pérez-Tortosa et al. 2012).

With this in mind, the purpose of this study was to analyze the influence of different SA doses (10, 100 and 1,000 μM) on C. heterophyllus plantlet growth performance, and to determine whether SA treatment improves the antioxidant properties of Cistus shoots by modifying the activity of some of the key enzymes involved in phenylpropanoid metabolism such as PAL and Prx. From a practical point of view, this work looks at the possibility of SA application in conservation programmes of rare and threatened plants in order to improve both their growth and antioxidant capacity.

Materials and methods

Plant material and medium composition

Explants were obtained from in vitro cultures of C. heterophyllus subsp. carthaginensis previously established in our laboratory and maintained by bimonthly subculture (at least seven subcultures before starting the assays). In vitro shoots were grown on half-strength MS medium (Murashige and Skoog 1962) (MS/2) without growth regulators and supplemented with casein hydrolysate (250 mg l−1), sucrose (3 % w/v) and 0.8 % Difco Bacto agar. All cultures were kept at 25 °C with a 16 h-light/8 h-dark cycle, and 100 μmol m−2 s−1 photon flux density (Sanyo, versatile environmental test chamber, MLR-351H, Japan).

SA treatments

Eighty 40-day-old shoots (five per vessel) were placed onto fresh basal medium, buffered to pH 6.0 with 3 mM MES-KOH, a concentration commonly used to supplement media for in vitro cultures (Gratão et al. 2008; Yuan et al. 2012). Hydroalcoholic SA solutions were sterilized by filtration through 0.22 μm filters (Millex GV, Millipore) and added aseptically to autoclaved media at the desired concentrations (0, 10, 100, and 1,000 μM). Equal volumes of ethanol were added to the control vessels. After autoclaving, the pH of the media were in the range of 5.7 ± 0.2. After 8 weeks, shoot tissues were immediately frozen in liquid nitrogen, pulverized in a liquid nitrogen-cooled analytical mill (IKA, Labortechnik, Germany) for 3 × 30 s intervals and stored at −80 °C until used.

Measurement of chlorophyll fluorescence and chlorophyll and carotenoid contents

Chlorophyll fluorescence measurements were made on the upper (adaxial) surface of leaflets using a PAM-210 chlorophyll fluorometer system (Heinz Waltz, Effeltrich, Germany). Plantlets were kept for 25 min in darkness to determine the minimal fluorescence level in the dark-adapted state (F O) and the maximal fluorescence in this state (F m). Calculations for maximum quantum efficiency of PSII photochemistry [(F m−F O )/F m and (F m−F O )/F O ] were carried as recommended by Maxwell and Johnson (2000) and Lichtenthaler et al. (2005).

Chlorophyll a (Chl a) and b (Chl b) and total carotenoids x + c (xanthophylls and carotenes) were extracted with 100 % methanol several times (×3) until the extract became colourless. Their levels were determined using the extinction coefficients and the equations reported by Lichtenthaler and Wellburn (1983).

Determination of stress markers

Endogenous hydrogen peroxide contents were determined by the ferric-xylenol orange (FOX1) method as described by Cheeseman (2006). Liquid nitrogen-powdered shoots (0.25 g) were homogenized using an Ultra-Turrax disperser (IKA, Germany) in 1 ml of 5 % trichloroacetic acid (TCA), and centrifuged (15,000g, 10 min, 4 °C). 25 μl of supernatant was mixed with 500 μL of FOX1 medium (0.5 mM Fe(NH4)2(SO4)2, 50 mM H2SO4, 0.2 mM xylenol orange and 200 mM sorbitol). After 30 min of incubation in the dark, the H2O2 concentration was determined based on the difference in absorption at 560 nm using a seven-point standard curve covering the range of 0.5–10 μM.

The degree of lipid peroxidation was determined by measuring thiobarbituric acid-reacting substances (TBARS) at 532 nm, with a correction for non-specific absorbance at 440 and 600 nm (Hodges et al. 1999), using the same supernatants as in the FOX1 assay. The MDA concentration was calculated using an extinction coefficient of 155 mM−1 cm−1.

The proline content was determined spectrophotometrically in a sulphosalicylic acid extract, using acid-ninhydrin reagent after Bates et al. (1973).

Extraction and quantification of free and bound salicylic acid

Extraction and analysis of free and bound SA (SAG, 2-O-β-D-glucosylsalicylic acid) were performed according to Verberne et al. (2002). Bound SA was quantified following acidic hydrolysis with 8 M hydrochloric acid. Quantification was performed using an HPLC system (Waters Alliance 2695) equipped with a Waters 2475 multi-wavelength fluorescence detector and a Phenomenex column (Torrance, CA, USA), type LUNA C18 (250 × 4.60 mm i.d., 5 μm) with a Phenomenex SecurityGuard pre-column. An internal standard, 2,3-dihydroxybenzoic acid (2,3-DHBA), was used to estimate SA recovery. The excitation and emission wavelengths used were 294 and 426 nm for SA, and 305 and 437 nm for 2,3-DHBA, respectively (Zawoznik et al. 2007).

Extraction and assay for PAL (EC 4.3.1.24) and Prx (EC 1.11.1.7)

Liquid nitrogen powdered shoots (0.6 g) were homogenized at 4 °C in cold extraction buffer [50 mM Tris-HCl, pH 8.8, 5 mM ethylenediamine tetra-acetic acid (EDTA), 5 mM ascorbic acid, 1 mM phenylmethylsulphonyl fluoride (PMSF), 0.1 % (w/v) Triton X-100, 200 μM dithiothreitol; (DTT), and 0.25 % (w/v) polyvinylpolypyrrolidone]. Homogenates were centrifuged (48,000g, 30 min, 4 °C) and the supernatants (2.5 ml) were then desalted by chromatography on PD-10 Sephadex G-25 columns (GE Healthcare, Buckinghamshire, UK), equilibrated either with 25 mM Tris-HCl (pH 8.8) containing 1 mM EDTA to determine PAL activity, or with 25 mM sodium acetate, pH 5.5 to determine Prx activity. The pellets were washed twice with 10 ml of 50 mM Tris-HCl (pH 7.5) containing 2 % (w/v) Triton X-100 and three times with phosphate buffer before incubating with 1 M KCl for 30 min with continuous shaking at 4 °C. After centrifugation (1,000g, 10 min, 4 °C), the supernatants were desalted and used for determining the ionically-bound Prx activities.

PAL activity was determined spectrophotometrically by following the conversion of l-phenylalanine into trans-cinnamic acid (ε290 = 9.5 mM−1 cm−1), according to Olsen et al. (2008). Prx activiy was estimated at 30 °C (Ferrer et al. 1990) using 0.1 mg ml−1 tetramethylbenzidine-HCl (ε652 = 39 mM−1 cm−1) as electron donor.

Determination of phenolics, flavanols, condensed tannins, and lignin

Powdered shoots (0.25 g) were mixed with 2 ml of methanol, incubated in dark condition for 1 h on a shaker at minimal speed, and centrifuged (12,000g, 5 min, 4 °C).

The supernatant was used for the determination of total phenolics (Everette et al. 2010) and flavanols (López-Arnaldos et al. 2001) using the Folin-Ciocalteu and p-dimethylamino-cinnamaldehyde (DMACA) reagents with calibration curves for caffeic acid (10–1,000 μM) and (+) catechin (100–400 μM), respectively. The pellets were washed several times with methanol and then, air dried at 60 °C. The resulting residues were used for lignin determination using the thioglycolic acid method as previously described (Hatzilazarou et al. 2006).

Condensed tannin contents were estimated by measuring the absorbance at 545 nm of the aqueous phase from the second partitioning (after incubation with HCl) carried out to obtain the bound SA according to Verberne et al. (2002). Results were expressed as cyanidin equivalents by using an ε545 = 34.7 mM−1 cm−1 (Vermerris and Nicholson 2006).

HPLC analysis for identification and quantification of individual phenols, and phenolic families, were carried out according to the elution conditions described in Sharma et al. (2005) by using a Waters Alliance 2695 separation module connected to a Waters 2996 diode array detector. Separations were performed in a LiChroCART RP-18 reversed-phase column (250 × 4 mm i.d., 5 μm particle size) supplied by Merck. Calibration curves corresponding to several standard compounds (provided by Sigma) were constructed in order to identify and quantify the peaks obtained in the chromatograms.

Antioxidant activities

Antioxidant activity of Cistus methanolic extracts was determined using the 1,1-diphenyl-2-picrylhydrazyl (DPPH) method, the ferric reducing antioxidant power (FRAP) assay, and the conjugated diene formation test as previously described (Pérez-Tortosa et al. 2012). The hypochlorous acid (HOCl) scavenging potential of the extracts was assayed using an electrophoretic assay according to Grippa et al. (2000) with minor modification. Briefly, the reaction mixture contained 90 μl of 0.05 % (w/v) bovine serum albumin in sodium phosphate buffer (100 mM, pH 7.0), 25 μl of methanolic extracts and 10 μl of 6.72 mM NaOCl solution. Control reaction media using (+)-catechin (25 μl, 2.5 mM) and ascorbic acid (25 μL, 2.5 mM) were also prepared. After a period of incubation of 15 min at room temperature, 15 μl of these samples were analyzed by non-denaturing PAGE on a 7.5 % (w/v) acrylamide gel using a Mini-Protean II Cell (Bio-Rad, Hercules, CA). Duplicate gels were stained with Coomassie brilliant blue R-250. Band intensities were quantified using ImageJ software (version 1.34s, NIH, Bethesda, MD; [http://rbs.info.nih.gov/ij].

Statistical analysis

Data are presented as mean ± standard errors (SE) from three independent experiments (with a minimum of 10 explants per treatment). Pearson linear correlation coefficients were calculated to evaluate associations among variables. The statistical significance of the differences (P < 0.05) between groups was tested using one-way analysis of variance (ANOVA) combined with Tukey’s honestly significant difference (HSD) post hoc test. All the statistical analyses were performed using the SPSS software package (version 19.0; SPSS Inc., Chicago, IL, USA).

Results

Effects of SA treatment on chl a fluorescence, photosynthetic pigment contents, and shoot growth

Chlorophyll a fluorescence is widely used for non-invasively monitoring the photosynthetic performance of plants (Maxwell and Johnson 2000; Baker 2008). Two related ratios F v/F m (where F v = F m−F 0) and F v/F 0 are often used for determining the potential quantum efficiency of PSII (Serret et al. 2001; Lichtenthaler et al. 2005). The first ratio is routinely used as an indicator of damage to the PSII reaction centre and, typical F v/F m values for non-stressed leaves ex vitro are 0.74 − 0.85 (Lichtenthaler et al. 2005). As shown in Table 1, there was a small, but not statistically significant, decrease in the F v/F m and F v/F 0 ratios in SA-treated shoots. A dose of 10 μM SA significantly increased the levels of chlorophyll a, b and carotenoids (by 30, 23 and 52 %, respectively), although their levels fell as the concentration of SA was increased, reaching levels below the control levels at the maximum SA concentration used (1,000 μM).

Table 1 Effect of salicylic acid (SA) concentrations on chl a fluorescence, photosynthetic pigment contents, and shoot growth of 40-day-old plantlets of Cistus heterophyllus grown for 2 months with or without the elicitor on MS/2 basal medium
Full size table

No major changes (Tukey’s HSD test, P < 0.05) in growth, estimated by both fresh shoot mass and shoot length, between the control, l0 μM SA, and 100 μM SA-treated explants were found. Exposure to a high SA concentration (1,000 μM) resulted in drastic growth inhibition of 30–50 % (see Supplemental Fig. 1).

Effect of SA on H2O2 level, lipid peroxidation, proline content, and accumulation of free and bound SA

To determine whether SA doses have a toxic effect, some biochemical stress markers were evaluated. No significant changes were noted in the hydrogen peroxide level in SA-treated shoots compared with the control, whereas SA at 100 μM significantly increased proline accumulation and lipid peroxidation, as measured by the TBARS assay, and the endogenous levels of both free (>2.5-fold) and bound (>eightfold) SA in the shoots (Fig. 1), although no visible phytotoxic effects on the shoots were observed. The total content of free and bound SA in 10 μM SA-treated shoots were close to the basal levels observed in the control treatments.

Fig. 1
figure 1

Effect of salicylic acid (SA) concentrations on H2O2 level, and lipid peroxidation (A), proline content (B), and accumulation of free and bound SA (C) in shoots of 40-day-old plantlets of Cistus heterophyllus grown for 2 months with or without the elicitor on MS/2 basal medium. Vertical bars represent mean ± standard error

Full size image

Effect of SA on PAL and Prx activities

SA treatments caused a significant increase (>tenfold) in PAL activity (Fig. 2). Soluble peroxidase activities also increased in a dose-dependent manner, whereas cell-wall-bound peroxidase activity increased by about 35 % compared with the control levels in 10 μM-SA-treated shoots, although at a dose of 100 μM SA the activity was in the range of the untreated controls (Fig. 2).

Fig. 2
figure 2

Effect of salicylic acid (SA) concentrations on the activities of PAL and Prx in shoots of 40-day-old plantlets of Cistus heterophyllus grown for 2 months with or without the elicitor on MS/2 basal medium. Vertical bars represent mean ± standard error. S-Prx soluble peroxidase activity, CW-Prx cell wall-bound peroxidase activity

Full size image

Effects of SA treatment on the phenolic content of Cistus extracts

Because SA treatments resulted in a substantial increase in PAL and Prx activities, its effect on the accumulation of phenolic compounds and flavanols, one of the major groups of phenolics present in Cistus species was analyzed. No major differences in the soluble total phenol content (TPC) between control and SA-treated explants were observed (Table 2). However, higher doses of SA tended to decrease the levels of soluble flavanols and increased the levels of proanthocyanidins present in Cistus shoots. HPLC analyses were carried out to verify whether SA caused a variation in the phenolic patterns of Cistus shoots. HPLC chromatograms at 280 and 310 nm showed that SA-treated shoots contained roughly the same type of phenolic compounds, even though there were differences in the concentration of individual compounds (see Supplemental Fig. 2). The HPLC phenolic profile was dominated by phenolic acids and flavonoids, whereas hydroxycinnamic acid derivatives were detected in lower amounts. As shown in Table 2, the lignin content increased twofold in 10 μM-SA-treated shoots, whereas with 100 μM-SA its content was in the range of the controls.

Table 2 Total phenol content (TPC), flavanols, proanthocyanidins, lignin, gallic acid, and hydroxycinnamic acid derivatives present in methanolic extracts of Cistus heterophyllus shoots treated with different SA concentrations
Full size table

Effects of SA treatment on the antioxidant activities of Cistus extracts

In the assessment of the effect of SA on the antioxidant activity of Cistus extracts, four different assays were carried out: DPPH radical scavenging activity, the electron donation capacity by Fe(III) reduction in the presence of TPTZ, the inhibition of lipid autoxidation by the conjugated diene method, and the hypochlorous acid scavenging potential.

As shown in Fig. 3, increasing SA concentration tends to decrease the free-radical scavenging activity, as measured by the DPPH method, and the ferric reducing capacity of the Cistus extracts, although the differences with the control were not statistically significant at lower SA concentrations. A strong correlation was obtained when these values were compared with the total flavanol content (r = 0.962 and 0.916, P < 0.01, respectively). SA also provoked a dose-dependent increase in the protection against both lipid peroxidation and HOCl-induced BSA oxidation. In all the SA treatments, the inhibition of BSA degradation was significantly greater than that observed in the 500 μM-ascorbic acid treatment (71 ± 3 % above control level) and similar to that observed with (+)-catechin at 500 μM (121 ± 16 % above control level). Good correlations were found between protection against HOCl and free SA (r = 0.810; P < 0.01) and bound SA (r = 0.783; P < 0.01), PAs (r = 0.888; P < 0.01), and total flavanol content (r = −0.606; P < 0.01).

Fig. 3
figure 3

Antioxidant activity of SA-treated Cistus shoots. Antioxidant activity of methanol extracts of Cistus heterophyllus shoots grown in vitro in the presence (10, 100 and 1,000 μM) of SA for 2 months. Each value is expressed as percentage of respective control value in the absence of SA. Values followed by the same letter on every axes are not significantly different at P < 0.05 level (one-way ANOVA followed by Tukey’s HSD test). Letter “a” denotes that the value is not significantly different (P < 0.05) from control (no SA-treated material). CD, conjugated diene assay; HClO hypochlorous acid scavenging assay, FRAP ferric reducing antioxidant power assay, DPPH DPPH scavenging assay

Full size image

Discussion

In the present study, 40-day-old C. heterophyllus shoots were cultured for 8 weeks on MS/2 solid medium in the presence of increasing doses of SA (10, 100, and 1,000 μM) in order to evaluate its possible beneficial effect on plantlet vigour, which was assessed by measuring shoot growth, photosynthetic activity and antioxidant capacity. Oxidative stress and poor development of the photosynthetic system in in vitro cultures have been cited as major factors in transplant vulnerability (Lee et al. 1985; Cassells and Curry 2001). In the present work, low doses of SA (10 and 100 μM) did not affect the maximum efficiency of PSII or shoot growth, but enhanced chlorophyll and carotenoid levels, while a dose of 1,000 μM SA caused a visible growth inhibition, and a decline in both Chl and carotenoid levels. The growth retardation effect of high SA doses is well documented both in SA-accumulating plants (Jirage et al. 2001; Mateo et al. 2006) and when the regulator is applied exogenously (Kováčik et al. 2009a; Hayat et al. 2010). Although there are few references in the literature to the effect of long-term exposure to SA on photosynthetic performance, an increase in chlorophyll concentration after the application of low SA concentrations has been reported in short-term experiments in other plant species, including Arabidopsis thaliana (Rao et al. 1997), several genotypes of cowpea (Chandra and Bhatt 1998), mustard (Fariduddin et al. 2003) and sunflower (El-Tayeb et al. 2006). Nevertheless, using low concentrations of SA, no change (Drazic and Mihailovic 2005; Abreu and Munné-Bosch 2008) and even a decrease (Moharekar et al. 2003) in photosynthetic pigments have also been mentioned. These seemingly contradictory results confirm that SA influences growth and productivity-related physiological processes in plants, such as the photosynthetic rate, depending on its endogenous levels in particular plant species under specific developmental and/or environmental circumstances (Raskin 1992; Horváth et al. 2007; Hayat et al. 2010).

In addition, it is considered that low doses of SA (5–500 μM) cause a moderate stress that affects cell redox homeostasis in a similar way to stress acclimation processes (Rao et al. 1997; Horváth et al. 2007). Indeed, it had been proposed that SA can play a pro-oxidant or antioxidant role according to its free endogenous level (Yang et al. 2004). In the present study, there were no great changes in the levels of H2O2 in the different treatments, but the extent of lipid peroxidation and the accumulation of proline, which are considered to be indicators of disturbed physiological conditions, increased in 100 μM-SA-treated shoots. Nevertheless, the absence of visible injuries in shoots treated with low doses of SA (10 and 100 μM) suggests that no severe oxidative damage was caused to plantlets under these conditions, probably due to the induction of antioxidant mechanisms. In this respect, several studies have suggested an antioxidant role for proline during stress (Hong et al. 2000; Matysik et al. 2002). In fact, proline has been found to counteract lipid peroxidation (Szabados and Savouré 2010; and references herein) by forming long-lived adducts with ROS (Alia et al. 1997). So, it is likely that SA-induced proline accumulation contributes to lowering ROS levels and, consequently, to protecting Cistus tissues against oxidative stress, as has been reported in several plant species under both normal and stress conditions, including lentil (Misra and Saxena 2009), sunflower (El-Tayeb et al. 2006) and wheat (Shakirova et al. 2003; Hussein et al. 2007). Moreover, a model that relates the proline-linked pentose phosphate pathway to the shikimate and phenylpropanoid pathways has been proposed by Shetty (2004). According to this model, proline biosynthesis coupled to the pentose phosphate pathway could stimulate the production of phenolic phytochemicals many of which are considered to be powerful antioxidants which would enhance the ROS-deactivating capacity of SA-treated plantlets.

When the endogenous SA content was evaluated, a marked increase in both free and bound SA was observed in the 100 μM-treated shoots. A recent work (Szalai et al. 2011) suggested that SA treatment induces its de novo synthesis by activating the expression of SA biosynthetic genes. This indicates that the endogenous SA content is subject to strict control and maintained at a level that would allow the induction of appropriate defence responses. Thus, it has been proposed that excessive SA accumulation can induce the programmed cell death pathway (Borsani et al. 2001), whereas, at moderately high levels, SA may be directly involved in the scavenging of free radicals acting as an antioxidant (Dinis et al. 1994; Szalai et al. 2011) and/or indirectly alter the redox balance through the activation of antioxidant responses (Horváth et al. 2007; Szalai et al. 2011).

Phenylpropanoid metabolism is one of the major pathways stimulated during stress and acclimation responses (Dixon and Pavia 1995; Winkel-Shirley 2002). Increasing information suggests that endogenous and exogenous SA induce both gene expression and the enzymatic activity of PAL (Chen et al. 2006; Kováčik et al. 2009a), which, in turn, provokes the accumulation of phenylpropanoids (Bate et al. 1994). In the present study, long-term exposure of Cistus explants to SA markedly enhanced PAL activity, although the levels of TPC remained unchanged. Nevertheless, SA induced marked changes in the content of certain phenolic compounds. As regards the family of flavanols, increasing SA concentrations decreased the levels of soluble flavanols and increased the content of soluble proanthocyanidins (PAs) present in shoots. There is general agreement that flavanols can polymerize, possibly following an oxidative process, yielding PAs that are stored in the vacuole (Pourcel et al. 2007; Hernández et al. 2009). The oxidation of flavanols could be catalyzed by both polyphenol oxidases (in the presence of oxygen) and Prx (in the presence of H2O2). In this way, the Prx-flavonoid pair constitutes a well-known important sink/buffer of excess H2O2 (Yamasaki et al. 1997; Pérez et al. 2002; Takahama 2004). However, in a recent work (Ferreres et al. 2011), this system was also seen to be involved in the fine homeostasis of H2O2 levels. Therefore, it is plausible that SA is involved in the regulation of this Prx/phenol system, which in turn, would permit the fine-tuned regulation of ROS levels, depending on its endogenous level.

Furthermore, soluble and cell wall-bound Prxs exhibited a distinct behaviour in 10 and 100 μM SA-treated shoots, suggesting that peroxidase isoforms were differentially affected by SA. These results agreed with previous investigations reporting that SA may interfere positively or negatively with different metabolic pathways mediated by Prx (Kawano 2003), and also with the induction of specific Prx genes (Almagro et al. 2009; and references herein). Thus, the differential influence of SA on Prx-mediated metabolic processes could explain the rise in the lignin content of 10 μM SA-treated shoots, which correlated with the increase in cell wall-bound Prx activity, while the accumulation of PAs in 100 μM SA-treated shoots correlated with the rise of soluble Prx activities. Accordingly, the different levels of flavanols and hydroxycinnamic acid derivatives observed in SA-treated shoots could depend, at least in part, on the SA dose-dependent stimulation and down-regulation of these Prx isoenzyme groups.

Similar relationships between TPC, polymeric phenolics and phenolic-metabolizing enzymes have been previously described and related to SA-mediated stress tolerance. For example, changes in the phenylpropanoid profile have been reported in chamomile plants as a consequence of exogenous SA application, which leads to only slight changes in TPC but stimulates the conversion of SA to gentisic acid, which, in turn, activates the expression of the mechanisms involved in the tolerance to some heavy metals (Kováčik et al. 2009b). In contrast to the above-mentioned small changes in TPC and reflecting our findings in the present study, SA treatment may also stimulate the increase in polymeric phenolics, including cell wall-immobilized, phenolics. This could affect the tolerance to stress provoked by different agents, mediating, for instance, the endogenous accumulation of some heavy metals (Kováčik et al. 2011).

Bearing in mind that many plant phenols display antioxidant properties and taking into account that the location and timing of ROS and antioxidant formation are key factors for cell signalling and defence responses (Foyer and Noctor 2009), the effect of SA on total antioxidant capacity was examined. No effect of low SA doses on the total antioxidant capacity was found when measured by the DPPH and FRAP methods, indicating that neither the free radical scavenging activity nor the reducing capacity of Cistus was affected by such SA doses. This contrasts with the results obtained for Thymus membranaceus shoots grown in similar conditions, in which SA treatments, especially at 10 μM, greatly enhanced the antioxidant capacity (Pérez-Tortosa et al. 2012). A high correlation between soluble flavanols and antioxidant activity, as assessed by the DPPH (r = 0.962; P < 0.01), and FRAP (r = 0.916; P < 0.01) methods, was found. Using the same methods, several authors have also reported significant correlation between catechin levels and antioxidant capacity in wines (Katalinić et al. 2004), green tea (Xu et al. 2004) and in different fruits, vegetables, and beverages (Floegel et al. 2011). The antioxidant activity of flavanols has been attributed to the presence of a catechol group on the B ring, which chelates redox-active metals and traps free radicals, namely superoxide radical, singlet oxygen and lipophilic alkyl peroxyl radical (Nakao et al. 1998; Pannala et al. 2001). Furthermore, PAs (the polymeric condensation products of flavanols) have also been seen to possess strong superoxide anion (O2·−), hydroxyl radical (OH·) and hypochlorous acid (HOCl) scavenging activities in both in vitro and in vivo models, as well as in human clinical studies (Cos et al. 2004; Maldonado et al. 2005; Aron and Kennedy 2008). Thus, it was not surprising to find good correlations between PA levels and the anti-lipid oxidation activity and the HOCl scavenging potential (r = 0.575, P < 0.05; r = 0.835, P < 0.01, respectively) in Cistus shoots. It is important to point out that the overall antioxidant capacity displayed by complex extracts in these assays was probably influenced by the occurrence of additive, synergistic and/or antagonistic effects of the different individual compounds. Thus, although it is not easy to attribute the antioxidant capacity displayed by plant extracts to the contribution of a specific phenolic class, it is likely that the oxidation of flavanols to PAs observed in SA-treated shoots could serve as a backup defense system against oxidative stress that could contribute to the attenuation of lipid peroxidation.

In conclusion, the changes in the phenolic contents observed in SA-treated shoots appear to result, at least partially, from the differential influence of SA on PAL and soluble and cell wall-bound Prxs. Thus, the SA dose-dependent oxidation of flavanols to PAs by soluble Prxs could contribute to attenuating oxidative stress in plants by preventing lipid peroxidation, as suggested by the high anti-lipid oxidation activity of extracts obtained from shoots. Besides growth-stimulating effects, the protection conferred against oxidative stress induced by low concentrations of exogenously applied SA could improve the vigour of Cistus plantlets, providing potentially superior material for conservation programmes. Moreover, the findings open up the possibility of using SA-treated Cistus shoot cultures as a production system for obtaining phytochemicals with pharmacological applications.