Thymus longicaulis, C. Presl

Galasso, Silvia, Pacifico, Severina, Kretschmer, Nadine, Pan, San-Po, Marciano, Sabina, Piccolella, Simona, Monaco, Pietro & Bauer, Rudolf, 2014, Influence of seasonal variation on Thymus longicaulis C. Presl chemical composition and its antioxidant and anti-inflammatory properties, Phytochemistry 107, pp. 80-90 : 81-87

2.1. ESI MS n metabolic profiling of extracts from T. longicaulis

In order to identify changes in the metabolic profile in wild plants of T. longicaulis, LC–ESI-MS/MS analyses of hydroalcoholic extracts (Jul12, Oct12, Jan13, Apr13), prepared by the maceration accelerated with ultrasound of leaf samples of herbaceous species collected at different stages of its life cycle, were carried out. The maceration in H 2 O:MeOH (1:1, v:v) facilitated the preparation of extracts in which polyphenols are abundant. Polyphenols are molecules highly biosynthesized in plants in which exercise primarily a defensive role against the overproduction of reactive oxygen species induced by abiotic stresses ( Gholizadeh and Kohnehrouz, 2010). The metabolites identified are reported in Fig. 1 and Table 1 .

Compound 1, was identified based on its fragmentation pattern, as quinic acid. In fact, the MS 2 spectrum of the [M — H] — ion at m/z 191 showed the ion at m/z 173, resulting from the dehydration of the deprotonated molecular ion and the fragment ions at m/z 127 and 111 ( Oen, 1985).

Compound 2, eluted with a retention time of 1.52 min, was putatively identified as stachyose (Galoi1–6Galoi1–6Glcoi1–2Fru). The mass spectrum showed a peak at m/z 701, due to the adduct with chloride, [M+Cl] —, whose fragmentation generated the deprotonated molecular ion at m/z 665 and a chlorinated adduct at m/z 539 formed by neutral loss of a fructose residue (162 Da) from the corresponding precursor [M+Cl] — ( Guan and Cole, 2008).

The metabolite 3, identified in the samples from July (Jul12), January (Jan13) and April (Apr13), was putatively identified as 12-hydroxyjasmonic acid hexoside (tuberonic acid), a molecule previously isolated from the leaves of Origanum vulgare and other Thymus spp. ( Koukoulitsa et al., 2006; Pereira et al., 2013). The mass spectrum showed the [M–H] — ion at m/z 387 which provided the ion at m/z 207 as base peak. This latter, due to CO 2 loss, generated the ion at m/z 163.

MS n spectra of compound 4 were in accordance with the presence of a bis -C -glycosylated flavone. The MS spectrum exhibited the deprotonated molecule at m/z 593, which produced MS 2 ions at m/z 473 [M–H-120] — (base peak; Han et al., 2008) and m/z 341 and 311. The loss of 120 Da allowed us to hypothesize the presence of a hexose residue linked by C-C bond to apigenin (341 = [aglycone+71] —; 311 [aglycone+41] —). Further collision of the ion at m/z 473 yielded m/z 353 (loss of 120 Da) indicating the presence of a second C -hexosyl moiety. The fragment ion at m/z 383 [aglycone+113] allowed us to confirm the aglycone identity ( Ferreres et al., 2003). As C -linked saccharide residues were found only in the C-6 and/or C-8 positions of flavone skeletons, compound 4 was assigned as apigenin-6,8-di- C -hexoside.

The mass spectrum of the metabolite 5 is in agreement with the presence of pinoresinol dihexoside ( Eyles et al., 2007). The fragmentation of the deprotonated molecular ion [M–H] — at m/z 681 led to the formation of fragment ions at m/z 519 and 357 for subsequent loss of two hexose residues (162 Da).

Compound 6 showed the [M–H] — ion at m/z 693. Typical neutral losses of hexose moieties bound to a phenolic oxygen atoms as 162 Da ( Ricci et al., 2010) provided MS 2 fragment ions at m/z 521 and 359. This latter provided, besides others, the fragment ion at m/z 329 according to the loss of 30 Da, reasonably CH 2 O. These findings, in line with Ricci and Piccolella (2012) which investigated the fragmentation mechanisms of furofuranic (FUR) and tetrahydrofuranic (THF) lignans, were in accordance with lariciresinol dihexoside.

Compounds 8 and 12 were putatively identified on the basis of their fragmentation pattern and relative retention times, as rosmarinic acid (12) and its hexosyl (8) derivative. The mass spectrum of the metabolite 12 showed the [M–H] — ion at m/z 359. The MS 2 spectrum of the precursor ion showed fragment ions at m/z 223 (due to the loss of a 3,4-dihydroxy vinylbenzene), m/z 197, corresponding to oi- hydroxyhydrocaffeoyl moiety, and at m/z 179 [caffeic acid-H] —. The fragmentation of the ion at m/z 197 generated, in the MS 3 experiment, a fragment ion [M–H] — at m/z 521 provided for the compound 8 the ion at m/z 359 (rosmarinic acid), leaving to assume the presence of a hexosyl derivative of the metabolite 12. The glycosylation could take place on of the aromatic unit of the oi- hydroxyhydrocaffeoyl moiety since it was recently observed that the bond of the hexose to the caffeic moiety provides a fragment at m/z 341 ( Sakalem et al., 2012), which was not detected in our analyses. Recently Vallverdú-Queralt et al., (2014), evaluating the phytochemical composition of herbaceous species with broad culinary uses such as rosemary, oregano, thyme and bay leaf, showed the presence of the hexosyl derivative of rosmarinic acid only in rosemary and oregano.

The glycosylated flavonoids constitute the majority of polyphenols in spices. Glycosylated flavonoids were commonly observed. The UV spectrum, which showed two absorption bands at 284 and 343 nm, relating to the conjugations in the B and A rings of a flavonoid nucleus, together with the MS/MS data allowed the identification of the metabolite 7 as quercetin-3- O -hexoside. The deprotonated molecular ion, [M–H] —, at m/z 463 dissociated generating the fragment ion at m/z 301 as base peak (— 162 Da) ( Hossain et al., 2010). The presence of the molecule is widely reported as constituent of species belonging to the Labiatae family ( Vallverdú-Queralt et al., 2014).

The mass spectra of the metabolites 9, 10, 11 and 15 were in agreement with the presence of glycosylated derivatives of flavone luteolin.

The MS 2 spectrum of metabolite 9 ([M–H] — at m/z 447) showed the ion at m/z 285 in agreement with the presence of a hexosyl derivative of the flavone luteolin, identified on the basis of the MS 3 fragmentation pattern. The molecule has recently been reported as a component of Thymus x citriodorus ( Pereira et al., 2013) and identified by 1D and 2D NMR spectroscopic techniques in Thymus sipyleus subsp. sipyleus var. sipyleus ( Özgen et al., 2011) .

The metabolite 10 was putatively identified as luteolin-7- O -hexuronide, previously identified as constituent of Thymus broussonettii ( Ismaili et al., 2002) and T. sipyleus ( Özgen et al., 2011) . The [M–H] — ion at m/z 461 provided the ion at m/z 285 by loss of 176 Da.

The mass spectrum of the metabolite 11 was in accordance with the presence of a disaccharide derivative of luteolin. In fact, the collision of the deprotonated molecular ion at m/z 579 gave fragment ions at m/z 417 and 285. The neutral loss of 162 Da suggested the presence of a hexose residue linked to a pentosyl derivative of luteolin, which for subsequent loss of 132 Da gave the flavone aglycone.

Compound 13 ([M–H] — at m/z 417) was putatively identified as pentosyl luteolin. The fragmentation of the [M–H] — ion gave the ion at m/z 285 for the loss of a neutral residue of 132 Da (pentose). The collision of the ion at m/z 285 gave, also in this case, several typical luteolin fragment ions ( Gouveia and Castilho, 2013) at m/ z 199 ([M — H–C 3 O 2 –H 2 O] —), 175 ([M — H–C 3 O 2 –C 2 H 2 O] —) and 241 ([M — H–CO 2] —). Since the O -glycosylation of flavones is favored at C-7 ( Cuyckens and Claeys, 2004), in the derivatives of luteolin identified the saccharide residue was tentatively positioned at C-7 carbon.

The fragmentation of the [M–H] — ion at m/z 591 provided for the compound 15 the fragment ion at m/z 531. The neutral loss of 60 Da suggested the presence of an acetyl function linked to an alcohol one. Confirming this hypothesis, the collision of the ion at m/z 531 provided, among other ions, the fragment ion at m/z 399 [531–132]. The loss of an acetyl function esterified to an alcohol group of a pentose sugar could facilitate the conversion of the pentose sugar in 2H-dihydro- γ -pyran with formation of the ion at m/z 513. The loss of the 2H-dihydro- γ -pyran moiety generated the ion at m/z 399. The presence of the ion at m/z 417 suggested, once again, the presence of a pentosyl derivative of luteolin. From these observations the metabolite 15 was identified as luteolin-7- O -acetylpentosylpentoside.

Compounds 14 and 19 were identified as two isomers of salvianolic K acid ( Lu and Foo, 2001). The MS spectrum of both compounds showed the [M–H] — ion at m/z 555, which, for subsequent decarboxylation and dehydration gave an ion at m/z 493. The loss of a 3,4-dihydroxyvinylbenzenic moiety generated the ion at m/z 359, which consisted in rosmarinic acid.

The deprotonated molecular ion, [M–H] —, at m/z 283 relative to the metabolite 16, isolated only in sample Oct12, has been identified as 4 0 -metil-apigenin (acacetin). To confirm the UV spectrum showed two bands at 231 and 330 nm and the spectrum MS/MS provided an abundant fragment at m/z 268 probably formed by homolytic cleavage and loss of CH • 3 ( Hossain et al., 2010).

Compound 17, present only in Jul12 sample, was putatively identified as a hydroxycinnamyl derivative: (oi- hydroxyhydroferuloylcaffeoyl)caffeoyl hexoside. The [M–H] — ion at m/z 697 gave fragment ions at m/z 535 and 323. The loss of a residue of 162 suggested the presence of a caffeoyl residue probably esterified to a oi- hydroxyhydroferuloyl moiety. The electron delocalisation from the fragment at m/z 535 could lead to the loss of a 2-(4-hydroxy-3-methoxyphenyl) acetaldehyde, a CO 2 unit and of one water molecule favoring the formation of the ion at m/z 323, which for cross-ring cleavage of the saccharide moiety, generated the ion at m/z 221.

The mass spectrum of the metabolite 18 presented a deprotonated molecular ion at m/z 493, whose dissociation generated fragment ions at m/z 359 (rosmarinic acid) and 295 in accordance with the presence of the isosalvianolic A acid ( Ruan et al., 2012).

2.2. Quantitative analysis of T. longicaulis extracts

The quantitative analysis allowed us to detect changes of the metabolic composition of the extracts from T. longicaulis collected in the different four months ( Fig. 2 ). The luteolin derivatives (O - pentoside, O -hexuronide and O -acetylpentosylpentoside) and rosmarinic acid, salvianolic acid K and isosalvianolic acid A were determined in all the samples analyzed. The concentration of such molecules seems to be collection time and season dependent. In fact, comparing the data obtained from the four samples, it was particularly observed that an increase in the content of rosmarinic acid and its derivatives was often related to a decrease of the content of the luteolin pentosyl derivatives. Rosmarinic acid was considerably abundant in the sample Oct12: it amounted to 3.03 mg quercetin equivalent for 1.0 g of fresh weight sample ( Table 2 ). Jul12 sample contained the lowest amount of rosmarinic acid.

2.3. Antioxidant activity of extracts T. longicaulis

The evaluation of the antioxidant capacity was carried out by using different methods given the current state of non-existence of a valid method for the determination of total antioxidant capacity of a system. One of the methods used, since the antioxidant activity of plant extracts often is related to their polyphenols amount, was the method of Folin and Ciocalteu Reagent. Although the results obtained by this method cannot be considered an absolute measure of the amount of phenolic material constituting the plant drug, they provide an indication of the reducing properties of the plant matrix. By interpolation on a calibration curve of gallic acid, it was possible to estimate the effective reducing capacity of the four T. longicaulis hydroalcoholic extracts. It was determined that the total polyphenol content (TPC) value was clearly higher in Jul12 extract (115.7 GAE (mg)/g FW) than in the other collected extracts. The TPC value undergoes profound changes that seem to strongly depend on the collection time. In fact, comparing the data of Jul12 sample with those for remaining months, it emerged that the TPC value was reduced for more than half for the Oct12 extract (69.6 GAE (mg)/g FW) and reached its lowest level for Jan13 extract (54.4 GAE (mg)/g FW). The spring season (Apr13) seemed to favor a plant metabolic activity turned to polyphenolic compounds biosynthesis (62.9 GAE (mg)/g FW).

As it is clear from Table 3 , all the investigated extracts showed an effective reducing power of the radical species target ABTS • + and DPPH •. The scavenging effectiveness seemed to be strongly dependent on the tested dose. Oct12 extract, whose polyphenolic profile analysis highlighted the massive presence of rosmarinic acid, showed an effective reducing power of the radical species target ABTS • + and DPPH •, quantifiable in the respective ID 50 values (8.91 µg/mL vs. the ABTS radical cation and 9.50 µg/mL vs. DPPH radical). The scavenging effectiveness seemed to be strongly dependent on the tested dose: both the radicals were completely converted in their reduced form at 25.0 µg/mL dose. Rosmarinic acid antioxidant effectiveness was already reported; Fadel et al., (2011) showed that the antiradical activity of rosmarinic acid was much higher than that exerted by bioflavonoids. DFT studies revealed that the high antioxidant power of the molecule could be due to the abstraction of hydrogen atoms from the ortho hydroxyl groups, characterizing both its A and B rings ( Cao et al., 2005). The sample Oct12 is the only one that contained as constituent acacetin, a well known antioxidant, anti-inflammatory and anticancer natural substance, recently proposed as a therapeutic agent for the treatment of neuropsychiatric disorders ( Lin et al., 2014). The evaluation of the Fe 3+ reducing power strongly correlated with the TPC values ( Table 3 ). In fact, it was observed that Jul12 contained the highest TEAC value (474.95 µg Trolox ® Equivalents per g of extract), followed by Oct12 sample (323.24 µg Trolox ® Equivalents per g of extract).

When ORAC assay was performed, all the extracts massively scavenged ROO • radicals at the highest doses tested and no fluorescein decay could be observed. Thus ORAC values were determined by calculating the net area under the curve (AUC) of the samples at the lowest tested dose (0.781 µg/mL). Oct12 was able to exert the strongest antiradical efficacy (776.54 µM Trolox ® Equivalents). The activity of Oct12 sample was about 1.48, 1.49, and 1.87-fold that of Jul12, Jan13, and Apr13 extracts, respectively.

Data from ABTS, DPPH, ORAC, Fe 3+ RP, and TPC assays were used to calculate RACI (Relative Antioxidant Capacity Index) value for each extract. RACI index allows the comparison of phytocomplexes antioxidant capacity derived from different chemical methods and provides a more comprehensive assessment of the whole antioxidant capacity. The ranking of antioxidant capacity of investigated extracts differed between methods. In particular, it was observed that ABTS and DPPH data had significant correlation ( Table 4 ). RACI values highlighted that the four extracts were able to exert a similar antioxidant capability. Jul12 sample was slightly more active than the other ones ( Table 5 ).

2.4. Cytotoxicity of extracts from T. longicaulis

The evaluation of cytotoxicity was carried out by application of the XTT test that allows to estimate the number of metabolically active cells present in culture and then to evaluate the effect of treatment with an exogenously added agent on the viability of the cell population ( Roehm et al., 1991).

The test is based on the ability of the compound XTT tetrazolium (yellow colored) to be metabolized by the mitochondrial dehydrogenases, active in living cells, in a water-soluble formazan derivative of (bright orange colored). The viable cells, unlike those of non-viable, reduce XTT and the amount of formazan product is proportional to the number of present cells.

In order to obtain a cytotoxic profile of the investigated extracts and to fit out non toxic and safe preparations, each extract was tested in a wide dose level range (10–250 µg/mL; Fig. 3 ). Jul12, Oct12, Jan13 and Apr13 extracts showed dissimilar behaviors towards the different cell lines (HCT-116, MDA-MB-231, U-251, CCRF-CEM and MRC-5). In HCT-116 colon carcinoma cells Jul12 extract at the highest doses (50.0, 100.0 and 250.0 µg/mL) elicited a potent cytotoxic response after 48 and 72 h exposure times. The same extract defined a decrease of cell viability of MDA-MB-23 breast cancer cells, at a dose of 250 µg/mL. All the extracts did not affect the growth of U251 glioblastoma cells, while they exercised proliferative effects on CCRF-CEM lymphoblastic cells. Jul12 and Jan13 extracts exerted antiproliferative effects at the highest dose level tested even on MRC-5 cells, suggesting the potential cytotoxicity of both these extracts. Furthermore, the evidenced cytotoxic effects were due to an extract dose far higher than those chosen for testing antioxidant and anti-inflammatory capabilities.

2.5. Anti-inflammatory activity of extracts from T. longicaulis

Extracts of T. longicaulis collected from different seasons exhibited a varying COX-2 gene expression inhibition in THP-1 macrophages whereas extract from October showed a significant COX-2 inhibition at 41.9 ± 6.6% compared to materials from other seasons ( Fig. 4 ).

The phenophase in which the leaves were collected makes significance difference concerning phenolic composition of the extracts therefrom and consequently their biological activity.

Wild plants of T. longicaulis were collected in Durazzano, a small center in the Southern Italy (altitude 286 m above sea level) characterized by a mild Mediterranean climate, identified as Csa climate on the basis of Köppen and Geiger classification. Four collection times, between 2012 and 2013 years, were considered, each of them represented the first month of season. Data from ISPRA (Istituto Superiore per la Protezione e la Ricerca Ambientale) reports show that 2012, as well as 2013, was a year warmer than the climatological mean, both globally and in Italy ( Desiato et al., 2013, 2014). The summer of 2012 was characterized by a strong lack of rainfall and strong positive thermal anomalies that have made it one of the hottest summers of the last fifty years. In the first July decade the average temperature in Durazzano was equal to 25.5 ° C with an average relative humidity of 59.4%. On 10th July the maximum temperature was 34.8 ° C. The drought condition and the absence of precipitations define a strong decrease of rosmarinic acid amount. The content of the other compounds, among those identified, was also negatively affected, with the exception of metabolites 6 and 17, present only in the summer sample, and of compound 13 which was 3.19, 1.73 and 1.41-fold that of Oct12, Jan13, and Apr13 extracts, respectively. The deleterious effect of heat stress on rosmarinic acid accumulation was previously reported by Fletcher et al. (2005), who hypothesize that the decrease of rosmarinic acid levels during continuous heat stress conditions could be due to a competing reaction. In fact the high demand for prenylquinones biosynthesis in plant tissues under stress conditions, reduces the availability of 4-hydroxyphenylpyruvate (common precursor) for rosmarinic acid production. The long drought period characterizing the summer of 2012 ended on September and in October 2012, our second collection time, typically autumnal conditions were present. The temperatures were slightly above the normal with a monthly average of 16.7 ° C, and rainfall was in line with the normal values (170 mm in 24 h). These environmental conditions, together with an average relative humidity equal to 76.2%, promoted a rosmarinic acid overproduction. Isosalvianolic acid A were also abundantly present, whereas methylapigenin was for the first time identified. It was evident that Oct12 extract lacks a good portion of flavonoid glycosides component. This latter was again present in Jan13 sample and underwent a further increase in Apr13 sample. In the first half of January 2013, the permanence of a high pressure area brought relatively mild temperatures over a large part of the Italian regions. In collection area the monthly average in terms of temperature, relative humidity and rainfall were 7.6 ° C, 77.3%, and 6.4 mm. Although the advent of the spring season was marked by low temperatures, mild climatic conditions characterized the April 2013 month with average temperature and rainfall of 15.0 ° C and 2.5 mm. On 10th April average temperature and relative humidity equal to 13.7 ° C and 70.0% were recorded. Common factors affecting the presence of flavonoid glycosides were the low precipitation rate and the high drought conditions that characterized the different considered collection times. Considering climate data, it is clear that Oct12 was a positive exception. Abiotic stresses resulting from water deficit led to reduction in photosynthesis, transpiration and other biochemical processes associated with plant growth and development ( Ramakrishna and Ravishankar, 2011). Furthermore abiotic stress leads to oxidative stress which was counteracted by the biosynthesis of phenol compounds ( Nascimento and Fett-Neto, 2010).

3. Conclusions

T. longicaulis C. Presl. is a small aromatic plant abundant in Mediterranean macchia with traditional medicinal use. The plant, known for the antimicrobial properties of its essential oils, is a rich source of polyphenol compounds. LC– MS n based metabolic profiling analyses showed marked qualitative and quantitative variations in the plant phytochemical constitution during its life cycle. Plants collected in July showed radical scavenging activity comparable to that exerted from autumnal extract, which consisted almost exclusively of rosmarinic acid. Jul12 extract was capable of exerting cytotoxic effects on colon and breast cancer cell lines comparable to or higher than those observed for vinblastine, a known anti-cancer agent. Oct12 extract, which was particularly rich in rosmarinic acid and methylapigenin, showed a marked antioxidant and anti-inflammatory potential.

Furthermore, the study shows that the effectiveness of a plant can go far beyond its balsamic period reaching optimal values in a period in which the organism in order to cope with unfavorable environmental conditions (e.g. high temperature, dryness of the soil) activates secondary metabolic pathways useful for the biosynthesis of polyphenolic antioxidants.

Figures

Fig. 1. Main metabolites identified in Thymus longicaulis extracts.

Fig. 2. TICs from Thymus longicaulis Jul12, Oct12, Jan13, Apr13 extracts and their relative MS spectrum.

Fig. 3. Cell viability (%) of Thymus longicaulis extracts (Jul12, Oct12, Jan13, Apr13) towards HCT-116, MDA-MB-231, U-251, CCRF-CEM and MRC-5 cell lines. Values are the mean ± SD of measurements were carried out on 3 samples (n = 3) analyzed twelve times.

Fig. 4. Effects of Thymus longicaulis extracts collected during 4 different seasons on COX-2 gene expression on human leukemia cell line THP-1. All extracts (Jul12, Oct12, Jan13, Apr13) were tested at 50 µg/ml. Dexamethasone (Dex) 2.5 nM served as positive control. The graph depicts compiled data (% inhibition) of 2 independent experiments (mean ± SD). The numbers below the graph show % inhibition of COX- 2 gene expression.