*Correspondence address. Geology Department, Faculty of Mathematics and Natural Sciences, Bergen University, AlleHgaten, 41, 5007-Bergen, Norway.
E-mail address:francisco.saez@geol.uib.no (F. SaHez).
Volatile oil variability in
Thymus serpylloides
ssp.
gadorensis
growing wild in Southeastern Spain
Francisco Sa
H
ez
*
Departamento de Biologn&a Vegetal (Bota&nica), Facultad de Biologn&a, Universidad de Murcia, 30,100-Espinardo (Murcia), Spain
Received 21 April 1998; accepted 27 October 1999
Abstract
Volatile oils from single plants ofThymus serpylloides ssp.gadorensiswere collected from Southeastern Spain and studied to check for chemical variability using gas chromatography (GC) and gas chromatography}mass spectrometry (GC}MS). Many of the samples showed a phenolic chemotype, while another important group had signi"cant levels of linalool. Geraniol, myrcene, caryophyllene oxide, terpinen-4-ol and 1,8-cineole were commonly present. Principal Component Analysis (PCA) and Cluster Analysis (CA) of this chemical variability separated two groups of plants characterized by either phenols or linalool, and an isolated third type with geraniol. A few samples were found to have both phenolic and non-phenolic compounds in high quantities, thus showing a mixed chemotype. Multidimensional scaling analysis (MDS) of the percentage concentration for each component of the essential oil showed that thymol, linalool, 1,8-cineole, borneol and geraniol have clear divergent vectors. ( 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Thymus serpylloides ssp. gadorensis; Lamiaceae; Essential oil; Thymol; Linalool; Geraniol; Myrcene; Caryophyllene oxide
1. Introduction
The genusThymusL. is distributed over the Eurasian continent, the northern part of Africa and southern Greenland, although it has been spread by man all over the
world. However, when SectionSerpyllum(with clearly di!erent ecological preferences, at places with little or no summer drought) is not considered, the Mediterranean basin becomes its dispersion area, with many species being adapted to both hot summer and cold continental winter conditions (Morales, 1986, 1989, 1993). The Iberian Peninsula is one of the most diversi"ed territories for this genus, where the morphological and chemical variability of thyme is as large as the number of environmental conditions displayed (Rivas-MartmHnez, 1988). Southeast Spain has traditionally been a source of
wild raw material, not only of thyme but also a large number of other aromatic plants (Alcaraz et al., 1988), to supply the#avour, cosmetics, pharmaceutical and condiment industries (Adzet et al., 1987; Stahl-Biskup, 1991; Lawrence, 1992). The economic pressure on these wild resources sometimes leads to an over-exploitation that threatens ecosystems, especially where semiarid weather conditions do not allow a quick regeneration of shrub populations.
Thymus serpylloidesBory ssp.gadorensis(Pau) Jalas has a discontinuous distribu-tion over southeastern Iberian Peninsula, mainly in Granada Province, but also in AlmermHa, Murcia and Alicante, in mountain ranges higher than 1500 m, growing on
scarcely developed soils with a varied chemical composition, from calcareous to clayish. It contains 58 chromosomes in somatic cells, and di!ers from the type subspecies in having a denser indument and being less prostrate, although it generally does not grow higher than 20 cm. Genetic relationships with other species of thyme are found frequently, leading to hybrids such as T.]aitanae ("T. serpylloidesssp.
gadorensis]T.vulgarisssp. vulgaris),T.]hieronymi("T. serpylloidesssp. gadoren-sis]T. mastichina) and T.]pastoris ("T. serpylloidesssp. gadorensis]T. zygisssp. gracilis) (Morales, 1995).
Recent studies (SaHez, 1995a, b, 1996, 1998, 1999) have shown a complex chemical picture for some Thymus species in the studied area, where the proximity of high mountains to the sea shore, the geological mosaic (metamorphic materials adjacent to siliceous or calcareous extensions, and Quaternary depressions with abundant bad-lands and karstic formations), and the bioclimatological conditions (with the high Betic rangelands performing a barrier to stop the wet wind originated at Atlantic ocean), cause considerable ecological variability that leads to chemical variability. Consequently,T. serpylloidesssp.gadorensisoften meets the possibility of producing: (a) the above mentioned botanical hybrids, not studied here, despite of their great botanical and chemical interest; (b) specimens that morphologically correspond to the type subspecies but present some chemical characteristics that deviate from the general guidelines for it, namely the presence of 1,8-cineole. This variability in the essential oil is sometimes explained if other species of thyme and the potential from hybridization are taken into account.
The volatile oil ofT. serpylloidesssp.gadorensishas previously been studied from one sample from Granada Province (Crespo et al., 1988) revealing a carvacrol (34%) chemotype that proved to be e!ective in vitro against several Gram-positive and Gram-negative bacteria (Crespo et al., 1990). One sample from AlmermHa Province was
Fig. 1. Localization of populations at the study area. Dispersal area ofThymus serpylloidesssp.gadorensisis restricted to the highest altitudes. Numbers inside parenthesis stand for the number of samples taken at each locality.
2. Materials and methods
2.1. Plant material
Aerial parts of#oweringThymus serpylloidesssp.gadorensiswere collected from the Spanish southeast (Fig. 1) from May to July of 1990}1993. The number of individuals (single plants) taken (1}5) from each locality depended on morphological variability observed. The plant material was dried at room temperature (1}2 months) and steam distilled for 2 h in a Clevenger-type apparatus. Volatile oils were kept in sealed glass tubes at 43, without anhydrous Na
2SO4, until analysis. Voucher specimens from each
locality are kept at the Herbarium of Murcia University (MUB). The determination of the taxonomical status for each sample was performed in situ. No morphological hybrids were collected for this study, although a certain variability was detected, both inter- and intra-populations.
2.2. Gas chromatography
c Fig. 2. Cluster analysis of the studied 34 individuals with remarks on the essential oil component(s) that characterize the major subgroups. Letters A}F refer to samples highlighted in Fig. 3.
following temperature program: 703C isothermal for 1 min, rising by 103C min~1up to 903C, isothermal for 1 min, followed by an increase of 103C min~1up to 2103C. Injector and detector temperature was 2503C. Carrier gas was He with a#ow rate of 1 ml min~1. Peak area and concentrations were calculated with a Hewlett Packard 3396A integrator. Essential oil component identi"cation was performed by IR spectra, Rt comparison with pure standards, and GC}MS comparison with bibliography (Jennings and Shibamoto, 1980).
For IR identi"cation, preparative GC was performed using a Perkin}Elmer 3B chromatograph (TCD detector) with a Carbowax 20M column (4 m]3 mm), a
tem-perature program of 703C}2153C at a rate of 23C min~1, and He as a carrier gas at 25 ml min~1. Injector and detector temperature was 2403C. Infrared spectro-photometer Beckman IR 4280 was used to obtain spectra. GC/MS analysis was performed on a Hewlett-Packard 5890 gas chromatograph using a 70 eV mass selective detector HP 5971, with the same column and chromatographic parameters as stated for GC. The percentage of identi"ed compounds in each sample ranged from 90 to 99%.
2.3. Statistical analysis
Ma!ei et al. (1993a, b) have previously made extensive use of statistical methods to interpret di!erent aspects of the metabolism of aromatic plants, demonstrating the usefulness of cluster analysis (CA) and principal component analysis (PCA). Another technique, multidimensional scaling (MDS), is allegedly very robust (Minchin, 1987) due to its iterative diminishment of a goodness-of-"t statistic that represents the normalized discrepancies between distances in the MDS plot and the smoothed distances predicted from dissimilarities within data being used. In the present study, all data were statistically processed with a Systat 5 software for PC. Trace quantities ((0.01%) were considered as zero value. The analysis included: (a) cluster analysis (CA; Euclidean distance, single linkage method) to establish the di!erent groups/chemotypes within the individual essential oils; (b) principal component analysis (PCA, based on a Euclidean correlation matrix) to check for partition among the identi"ed compounds; (c) multidimensional scaling (MDS, based on a Spearman correlation matrix) to study the relationships among the di!erent compounds.
3. Results and discussion
the maximum and mean percentages for each compound from 34 samples!
Compound % Site (sample) Max. % Mean$SEM
3(2) 6(1) 8(3) 8(4) 10(1) 10(2)
a-thujene * * 0.03 0.16 * * 0.38 0.1$0.0
a-pinene 0.40 0.74 1.62 0.74 3.54 3.02 4.21 1.7$0.2
Camphene 0.38 0.31 0.77 4.22 4.66 5.86 7.93 1.7$0.4
b-pinene 0.06 0.15 0.37 0.24 0.84 * 2.37 0.5$0.1
Sabinene 0.03 0.07 1.04 0.06 * * 1.56 0.3$0.1
Myrcene 0.17 1.40 0.13 0.83 17.11 30.39 30.39 5.0$1.5
a-terpinene 0.03 1.37 5.07 0.07 * * 5.07 0.8$0.2
Limonene 0.06 0.40 0.84 0.95 * * 5.61 1.2$0.3
1,8-cineole 0.12 0.35 4.05 0.25 10.52 10.19 13.52 3.2$0.7
c-terpinene 0.31 5.98 11.70 0.73 1.82 * 20.85 5.4$0.9
Linalool 0.15 0.06 29.38 24.47 1.54 * 79.74 17.4$4.5
Linalyl ac. 1.02 0.90 1.33 39.39 * * 39.39 2.7$1.2
Isobornyl ac. 1.99 1.27 0.09 1.04 0.48 * 11.10 1.4$0.4
b-caryophyllene 1.52 0.08 2.25 0.13 * * 5.22 0.7$0.2
Terpinen-4-ol 0.14 * 0.13 0.73 10.82 11.71 28.05 3.2$1.0
a-terpineol 0.27 0.06 0.34 2.48 0.54 27.02 27.02 1.5$0.8
Borneol 2.95 1.46 0.54 2.15 1.54 2.36 10.44 1.7$0.3
Geranial 1.05 0.05 0.15 1.02 7.13 4.89 7.13 1.2$0.3
Geranyl ac. 0.10 0.42 0.02 1.32 * * 17.70 0.8$0.5
Citronellol 1.95 * 0.05 0.04 * * 1.95 0.2$0.1
Geraniol 79.99 0.26 0.25 1.24 * * 79.99 7.4$3.9
Caryophyllene ox. 0.32 1.03 0.31 0.48 6.28 1.74 14.10 1.9$0.6
Viridi#orol 0.05 * 0.07 0.05 * * 0.93 0.1$0.0
Elemol 0.24 0.05 0.15 0.02 * 2.81 2.81 0.2$0.1
Spathulenol 0.10 0.46 0.07 0.07 2.04 * 5.36 0.5$0.2
Thymol 0.97 18.50 11.55 0.90 3.56 * 56.97 11.9$2.5
Carvacrol 0.13 20.13 0.77 0.87 1.37 * 27.73 3.3$1.2
of unusual concentrations of particular compounds. Thus, cluster analysis develops a total of"ve groups, marked with vertical lines at Fig. 2, and outlined as follows:
(a) The group characterized by phenols is the largest one, with the highest level of thymol in the individual named&Aulago 1'(57%), and with&Aulago 2'presenting the maximum for carvacrol (27%). Both phenols were found to co-occur not only at the same site (di!erent plants), but also in the same sample, such as&Tetica 1'(Table 1) and
'MarmHa 1'(13% thymol and 17% carvacrol). Medium to high proportions of phenolic
precursors (p-cymene andc-terpinene) are also characteristic throughout the group. (b) The linalool group is also well represented with many samples, although chemical variability here is greater than in the phenol group. The most diverse samples are &CambroHn 4' (24% linalool #39% linalyl ac.) and Bacares 1 (39% linalool#29%tr-sabinene hydrate), with&CambroHn 3'and&MorroHn 1'being peri-pheral due to the simultaneous presence of phenols and linalool.
(c) Myrcene characterizes a group of"ve samples, which, with the exception of
&Tetica' are geographically close. Levels of myrcene range from 17 to 30%, and
caryophyllene oxide, 2 to 14%) are present.
(d) Three samples were found with geraniol levels from 75 to 80%. They were collected from&Buitre'and&La Ragua'sites, which are relatively near to each other at the southwestern part of the study area.
(e) The remaining three samples in Fig. 2 present considerable chemical di!erences and have been placed apart from the rest due to an unusual combinations of essential oil components. Thus,&Aitana 2'has a mixture of phenolic and non-phenolic com-pounds (18% carvacrol#10% 1,8-cineole#10% borneol), while&Bacares 3'is char-acterized by terpinen-4-ol (28%), and &Calar Mundo bajo 2' has myrcene (30%),
a-terpineol (27%), terpinen-4-ol (12%) and 1,8-cineole (10%).
The great variability in the volatile oil of T. serpylloides ssp. gadorensisis also demonstrated by Principal Components Analysis (Fig. 3), where samples marked with
&C'and&D'represent the linalool group,&B'and the neighbouring plots are devoted to
phenols, and the lower part of the graph contains the myrcene group. Plots&A',&E'and
&F' represent the three unusual combinations of components. Complete essential oil
composition of all samples represented with a letter is shown at Table 1, also including maximum and mean values for each compound. Factor 1 re#ects 34.7% of the total variability, while Factor 2 re#ects 24.2% and Factor 3, 13.0%, jointly making 71.9% of the total. Fig. 4 shows vectors for the di!erent compounds presenting more than 8% at least in one studied sample, after MDS analysis. There is a high number of vectors represented, if this analysis is compared with similar statistical studies performed for other species of thyme growing at the same area (SaHez, 1995a, b, 1996, 1998, 1999). Vectors representing phenols, linalool, geraniol and myrcene clearly diverge, while other components such as 1,8-cineole or terpinen-4-ol take up intermediate positions.
4. Conclusion
Fig. 3. Localization of individuals for (a) factors 1 and 2, and (b) factors 1 and 3 after PCA. A"Buitre 2;
B"Tetica 1; C"CambroHn 3; D"CambroHn 4; E"Calar del Mundo bajo 1; F"Calar del Mundo bajo
2. Samples A and F overlap at a.
previously reported to be exclusively phenolic, thus, the presence of 1,8-cineole and linalool in plants morphologically within the type subspecies (not hybrids) could be regarded as "ngerprints of introgression processes with other species of thyme growing nearby, that present these chemotypes themselves (namely,T.vulgarisand T. zygisssp.gracilis).
Fig. 4. Multidimensional Scaling Analysis. Only compounds with 8% in at least one sample are represented.
geraniol, myrcene and caryophyllene oxide in Figs. 3 and 4 (with easily detectable chemical/geographical groups). The comparison of data reported here with the previously published data shows a much wider variability in essential oil composition ofT. serpylloides ssp.gadorensisthan earlier expected. The punctual appearance of medium levels oftr-sabinene hydrate, terpinen-4-ol, caryophyllene oxide and other compounds, and the remarks stated above suggest that investigations with much more plant material per site is required, to improve the knowledge about the occurrence of these components in the essential oil of this economically interesting species. Speci"c focusing of future works on morphological hybrids will presumably lead to the description of new mixed chemotypes.
Acknowledgements
The author thanks Dr. M.C. GarcmHa-Vallejo and Dr. M.C. Soriano for their help
with analytical methods. Dr. P. SaHnchez kindly modi"ed maps from Rivas-MartmHnez.
Thanks are also due to RAMON SABATER, S.A., the Laboratory of Palynology at Murcia University, and&FundacioHn SEND ECA'
References
Adzet, T., Granger, R., Passet, J., San-Martin, R., 1987. Le polimorphisme chimique dans le genreThymus: sa signi"cation taxonomique. Biochem. Syst. Ecol. 5, 269}272.
Alcaraz, F., SaHnchez-GoHmez, P., Correal, E., 1988. CataHlogo de las plantas aromaHticas, condimentarias y medicinales de la RegioHn de Murcia. MonografmHas I.N.I.A. No. 67. Madrid.
Arrebola, M.L., Navarro, M.C., JimeHnez, J., 1995. Variations in yield and composition of the essential oil of Thymus serpylloidesBory subsp.gadorensis(Pau) Jalas. J. Ess. Oil Res. 7 (4), 369}374.
Crespo, M.E., Gomis, E., JimeHnez, J., Navarro, C., 1988. The essential oil ofThymus serpylloidesssp. gadorensis. Planta Medica 54, 161}162.
Crespo, M.E., JimeHnez, J., Gomis, E., Navarro, C., 1990. Antibacterial activity of the essential oil ofThymus serpylloidessubespeciesgadorensis. Microbios 61, 181}184.
Jennings, W., Shibamoto, T., 1980. Qualitative Analysis of Flavour Fragrance Volatiles by Gas Chromatography. Academic Press, New York.
Lawrence, B.M., 1992. Chemical components ofLabiataeoils and their exploitation. In: Harley, R.M. Reynolds, T.K. (Eds.), Advances in Labiate Science. pp. 399}436.
Ma!ei, M., Mucciarelli, M., Scannerini, S., 1993a. Environmental factors a!ecting the lipid metabolism in Rosmarinus ozcinalisL. Biochem. Syst. Ecol. 21 (8), 765}784.
Ma!ei, M., Peracino, V., Sacco, T., 1993b. Multivariate methods for aromatic plants: An application to mint essential oils. Acta Horticulturae 330, 159}169.
Minchin, P.R., 1987. An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 71, 145}156.
Morales, R., 1986. TaxonommHa de los geHnerosThymus(excluida la Sect.Serpyllum) yThymbraen la PenmHnsula IbeHrica. Ruizia 3, 1}324.
Morales, R., 1989. El geHnero ThymusL. en la regioHn mediterraHnea occidental (Lamiaceae). Biocosme. Mesogeen. Nice. 6 (4), 205}211.
Morales, R., 1993. El geHneroThymusL. (Labiatae) en Africa. An. Jard. Bot. Madr. 51 (2), 205}236. Morales, R., 1995. HmHbridos deThymusL. (Labiatae) en la PenmHnsula IbeHrica. An. Jard. Bot. Madr. 53 (2),
199}211.
Rivas-MartmHnez, S., 1988. Memoria del mapa de series de vegetacioHn de Espan8a. Madrid, I.C.O.N.A., Serie TeHcnica.
SaHez, F., 1995a. Essential oil variability ofThymus zygisgrowing wild in southeastern Spain. Phytochemistry 40 (3), 819}825.
SaHez, F., 1995b. Essential oil variability ofThymus hyemalisgrowing wild in Southeastern Spain. Biochem. Syst. Ecol. 23 (4), 431}438.
SaHez, F., 1996. El geHneroThymusen el Sureste IbeHrico: estudios bioloHgicos y taxonoHmicos. Ph.D. Thesis, Department of Plant Biology (Botany), University of Murcia.
SaHez, F., 1998. Variability in essential oils from populations ofThymus zygisL. in Southeastern Spain. Journal of Herbs, Spices and Medicinal Plants 5 (4), 65}76.
SaHez, F., 1999. Essential oil variability ofThymus baeticusgrowing wild in southeastern Spain. Biochem. Syst. Ecol. 27 (3), 269}276.
