Interspecifi c variation in the allelopathic potential of the family Myrtaceae

Allelopathy is a type of biotic interference wherein a plant releases bioactive metabolites into an environment, thereby aff ecting the adjacent biota. Stressful environments stimulate the production of these metabolites. Th e present study tests the novel weapons hypothesis, which postulates that species belonging to the same genus and from the same environment have similar allelopathic eff ects. Th e aim of this study was to assess the allelopathic eff ects that the aqueous leaf extracts of 15 species belonging to fi ve genera of the Myrtaceae family have on the seed germination and initial seedling growth of lettuce (Lactuca sativa L.), tomato (Solanum lycopersicum L.) and onion (Allium cepa L.). Germination rates, average germination times, informational entropy of germination and allelopathic eff ects, as quantifi ed with a response index, were calculated. A taxonomic distance matrix based on Gower dissimilarity and a Euclidean distance matrix were constructed. Th e results revealed that all extracts from donor species signifi cantly increased average germination time or reduced the germination rate of eudicotyledonous plant species. Th e only extracts that showed no eff ect on monocotyledonous seeds were those of Campomanesia pubescens O. Berg and Psidium cinereum Mart. We conclude that eudicotyledonous and monocotyledonous plants were both signifi cantly aff ected by the presence of all extracts tested. Our results make it clear that each species behaves distinctly in relation to allelopathic activity, with no apparent grouping by genus or subtribe. Th erefore, the hypothesis was rejected, because plants from the same environment and with taxonomic proximity do not necessarily display similar production of secondary metabolites.


Introduction
In the neotropical savanna of Brazil (the cerrado), plants grow in nutrient-poor soils (Haridasan 2008).Th erefore, the replacement of predated leaves represents a high cost (Fine et al. 2006) and competition for nutrients is intense.Consequently, plant species in this ecosystem have developed defense mechanisms, such as the production of leaves only in propitious periods; the production of coriaceous leaves; the lowering of nitrogen and water content; and the elevation of quantities of phenolic compounds (Marquis et al. 2002).Predation and competition for natural resources restrict the distribution of plants, acting as an environmental fi lter by selecting species with similar traits that allow their survival under environmentally challenging conditions (Fukami et al. 2005).Th us, functional traits are generally phylogenetically conserved in plant lineages (Ackerly 2003).Th erefore, phylogenetic proximity among species and traits inherited from a common ancestry should correlate with similar responses to environmental processes (Webb et al. 2002;Núñez-Farfán et al. 2007).Th e novel weapons hypothesis postulates that plant species belonging to the same genus and subjected to the same environmental conditions tend to have similar characteristics, principally with respect to their defense mechanisms, which include the production of inhibitory secondary metabolic compounds.To test this hypothesis, we collected samples of 15 species belonging to fi ve genera of the family Myrtaceae, all from the cerrado ecosystem.Our objective was to determine whether the allelopathic eff ect is similar across the genera.
Allelopathy is a type of biotic interference wherein a plant releases bioactive metabolites, known as allelochemicals, into the surrounding environment.Th e growth of neighboring vegetation might be aff ected and a selective advantage thus aff orded the donor plant (Anaya 1999).Allelochemicals aff ect various metabolic processes in or-ganisms: altering membrane permeability (Bogatek et al. 2005) and ion absorption (Gniazdowska & Bogatek 2005); inhibiting electron transport in photosynthesis and respiration (Abrahim et al. 2000); changing enzyme activity (Singh et al. 2009); and impeding cell division (Teerarak et al. 2010).Because of these eff ects, allelopathy is recognized as an important ecophysiological process in ecosystems, infl uencing primary and secondary plant succession, as well as the structure, composition and dynamics of native or cultivated plant communities (Rizvi et al. 1992;Scrivanti et al. 2003).In addition, allelopathy plays a key role in the detection of bioactive compounds of commercial importance (Oliveros-Bastidas 2008).
Worldwide, the family Myrtaceae includes approximately 3,100 species in approximately 140 genera, divided into two subfamilies, Leptospermoideae and Myrtoideae, outlined in the second edition of the Angiosperm Phylogeny Group classifi cation system, as modifi ed by Judd et al. (1999) and later ratifi ed by Watson & Dallwitz (2007).In Brazil, the subfamily Myrtoideae comprises 23 genera and approximately 1,000 species (Cardoso & Sajo 2006).Th e fl oral inventory compiled by the Brazilian Institute of Geography and Statistics showed that the cerrado contains approximately 211 Myrtaceae species in 14 genera, making this family one of the most representative in the ecosystem.Species of Eucalyptus (subfamily Leptospermoideae), the most widely studied genus in the family Myrtaceae, have been reported to show allelopathic eff ects (Fang et al. 2009).However, there have been few studies of the allelopathic potential of Myrtaceae species found in the cerrado ecosystem, most of which belong to the subfamily Myrtoideae.In the cerrado, some Myrtaceae species have been observed to inhibit the growth of adjacent plants, indicating that the former produce allelochemicals.Th erefore, the aim of this study was to assess the allelopathic eff ects that the aqueous leaf extracts of 15 species belonging to fi ve Myrtaceae family genera found in the cerrado ecosystem have on the seed germination and initial seedling growth of Lactuca sativa (lettuce), Solanum lycopersicum (tomato) and Allium cepa (onion).

Collection area
Th e plant material used in this study was collected in the cerrado (stricto sensu) in the state of São Paulo, Brazil (21°58' to 22°00'S; 47°51' to 47°52'W).According to the Köppen climate classifi cation system, the climate in the region is type Cwa, with two well-defi ned seasons (Monteiro & Prado 2006): a wet season (from October through March) and a dry season (from April through September).Th e vegetation is characterized by a woody layer composed of trees and bushes that protrude above a clearly defi ned herbaceous layer (Ribeiro & Walter 1998).

Biological material
Th e Myrtaceae specimens found on-site were marked and observed until the period of fl owering and fruiting, enabling the species to be identifi ed.Th e leaves of each species were non-systematically collected from at least fi ve plants in the vegetative stage, during the dry season.Carlos (accession nos. 8308,8309,8319,8320,8321,8322,8310,8311,8312,8313,8314,8315,8316,8317 and 8318,respectively).Leaves were collected and dried in an incubator at 40°C for 48 h, then powdered in an electric grinder and stored in plastic containers at room temperature (≈ 25°C).

Preparation of plant extracts
Th e extracts were prepared by mixing powdered dried leaves with distilled water in the proportion 1 g powder:10 ml water, at 10% weight/volume.Th is solution was stored at ≈ 5°C for 12 h.Th e extract was then fi ltered by suction through a Buchner funnel covered with fi lter paper (Gatti et al. 2010).

Germination bioassay
Th e seeds of the target species were placed in Petri dishes (9 cm in diameter) lined with two sheets of fi lter paper moistened with 5 ml of an aqueous leaf extract or with the same volume of distilled water (control).Th e experiments were carried out with four replicates of 20 seeds per dish for each extract.Petri dishes containing lettuce and tomato seeds were maintained in a biochemical oxygen demand germination chamber at 25°C and Petri dishes containing onion seeds were maintained at 20°C.Aft er preliminary tests performed to determine the optimal germination conditions for each target species, all of the Petri dishes were maintained on a 12/12-h light/dark cycle.Germination counts were made every 12 h for 15 days.Seeds that sprouted a 2-mm primary root were classifi ed as having germinated (Borghetti & Ferreira 2004).

Seedling growth bioassay
Th e seeds used in this bioassay were fi rst germinated in water (until presenting a root length of 2-4 mm), then moved to transparent plastic boxes (500 ml, 14 × 10 cm) lined with two sheets of fi lter paper and moistened with 13 ml of extract or distilled water (control).Th e boxes were kept in biochemical oxygen demand germination chambers, under the same conditions described for the germination bioassay.Th e length of the shoot and primary roots, in mm, were measured with a digital caliper aft er seven days.

Physicochemical characteristics of the extracts
Th e osmotic potential of the extracts was measured with an osmometer (5004 MICRO-OSMETTE; Precision Systems, Natick, MA, USA).Osmolality was measured in mOsm kg −1 and converted to MPa (Larcher 2004).Th e germination and growth bioassays of lettuce, onion and tomato were carried out with polyethylene glycol (PEG) 6000 solutions in order to simulate the osmotic potential recorded in the extracts.Th e PEG solution was prepared in accordance with specifi cations given by Sun (2002), and the germination and growth bioassays followed the method described above.
Th e pH of the extracts was measured with a pHmeter (PM608; Analion, Ribeirao Preto, Brazil).Because the pH of all the samples was within the range recommended for germination and plant growth (Larcher 2004), bioassays to evaluate the infl uence of pH were not carried out.

Mathematical and statistical analysis
We calculated proportional germination, expressed as a percentage; average germination time, in days; informational entropy of germination, in bits (Ranal & Santana 2006); and the response index for the allelopathic eff ect (Zhang et al. 2010).Th e response index was calculated as follows: RI = (T.C −1 ) − 1 where RI is the response index, T is the germination rate (seeds germinated per day) for seeds exposed to the leaf extract, and C is the germination rate (seeds germinated per day) for the control seeds.
Th e design of the laboratory experiments was completely randomized, with four replications for each treatment.Th e statistical signifi cance of the diff erences between the treatments and the control were investigated by Student's t-test, for normal data, or by the Wilcoxon test, for non-normal data, both at the 5% level.
A taxonomic distance matrix of Gower dissimilarity based on genus and subtribe was constructed for the 15 donors.Subsequently, a Euclidean distance matrix of donors was created based on the mean values of each variable (proportional germination, average germination time, informational entropy of germination, shoot length and root length) for all target species.Th e correlations between donors, in terms of the taxonomic and Euclidean distances, were tested with Mantel's test (Manly 2000).A dendrogram was built from the Euclidean distances, using the group average method, to defi ne the clusters formed by the donor species with similar allelopathic eff ects.Clusters were identifi ed by a dissimilarity ≥ 50.We used ANOVA, followed by Tukey's test, in order to illustrate which variables infl uence the clustering.All statistical analyses were performed with the program R, version 2.9.1 (R Development Core Team 2009).

Results
The osmotic potential of the aqueous leaf extracts ranged from −0.1 to −0.2 MPa.Th e bioassays using PEG 6000 showed that the osmotic potential did not infl uence germinability, average germination time or seedling length (Fig. 1 and 2).
Th e results obtained in the germination bioassay are shown in Fig. 1.Regarding the fi rst eudicotyledonous species (lettuce), the aqueous leaf extracts of C. pubescens and P. cinereum did not have signifi cant eff ects on the germination rate, although they did signifi cantly increase the average germination time.In addition, lettuce seeds displayed high values for information entropy when subjected to C. pubescens extract, indicating low synchrony of the germination process.Regarding the second eudicotyledonous species (tomato), the aqueous leaf extracts of C. pubescens, E. myrcianthes and P. cinereum did not signifi cantly reduce proportional germination, although they did signifi cantly increase the average germination time.In the monocotyledon species (onion), the eff ects of the aqueous leaf extracts of E. myrcianthes, E. punicifolia, M. multifl ora, M. splendens, M. tomentosa and P. laruotteanum had signifi cant eff ects on the germination rate, and all of the extracts except C. pubescens, E. myrcianthes, P. cinereum and P. laruotteanum, had signifi cant eff ects on average germination time (Fig. 1).Hence, all of the extracts inhibited the germination of eudicotyledon and monocotyledon seeds, with the exception of the extracts of C. pubescens and P. cinereum, which did not inhibit the germination of monocotyledon seeds.All of the target species seeds had a negative response index, indicating the presence of allelopathic activity on the part of all donor species (Fig. 1).
Th e results of the growth bioassay are shown in Fig. 2.Among the eudicotyledonous species, the extracts inhibited shoot and root growth for lettuce seedlings.Regarding tomato, all of the tested extracts caused signifi cant inhibition of root growth, whereas shoot growth was reduced by 11 of the 15 extracts.In the monocotyledon species (onion), all extracts reduced root growth; however, only eight extracts altered shoot growth.
No correlation was found between the variables of Euclidean and taxonomic distances (Mantel's r = 0.15, p = 0.057).Cluster analysis of the allelopathic eff ects identifi ed four clusters of donor species-clusters A, B, C and D, comprising 1, 5, 4 and 6 species, respectively (Fig. 3).Cluster A (control) diff ered from the other clusters in terms of seed germination for lettuce and tomato; shoot length for onion, tomato and lettuce; and root length for onion and tomato.Cluster C did not diff er from cluster A in terms of lettuce seed germination or root length (Table 1).In cluster B, the extracts strongly inhibited lettuce and tomato germination, compared with only slight inhibition of such germination in cluster C. Cluster D species extracts moderately inhibited lettuce and tomato germination (Table 1).

Discussion
Th e osmotic potentials of the aqueous leaf extracts evaluated signifi cantly infl uenced neither seed germination nor seedling development (Fig. 1 and 2), indicating that the detrimental eff ects exerted by the extracts on the germi-nation and growth of the target species resulted from the presence of bioactive substances.According to Grisi et al. (2012), osmotic potentials below −0.3 MPa do not interfere in seed germination or in the initial growth of seedlings.
Th e results showed that the monocotyledon and eudicotyledon plants respond diff erently to the extracts (Fig. 1 and  2).Th e infl uence of the extracts on germination depends on the size and permeability of the seed coat (Hanley & Whiting 2005).Th e species-dependent response to the allelochemicals shows that they are fundamental not only in natural environments, where they infl uence fl oristic composition, but also in agriculture, where they can be used as selective herbicides.
Allelopathic chemicals alter plant growth and seed germination through a multiplicity of eff ects on physiological and biochemical processes, because there are hundreds of diff erent structures and many of the compounds have multiple phytotoxic eff ects.According to Singh et al. (2009), phytotoxins can aff ect enzyme activities or plant hormones, increasing amylase activity and promoting a greater release of reserves that would otherwise be provided to the embryo, extending oxidative stress and seed dormancy through the increase of abscisic acid production and inhibiting water absorption via alterations in membrane permeability.Th e same author also observed a reduction in protein content and nitrate reductase activity in corn root tissues exposed to Nicotiana plumbaginifolia leachates.Einhellig (1995) also reported that alteration in the enzymatic activity of seeds aff ects the mobility of stored compounds, thus increasing germination inhibition, or sometimes completely suppressing it.Th erefore, the observed diff erences between the control and treatment groups, in terms of the number of germinated seeds, might be attributed to the presence of allelopathic compounds.Th e seedling growth results for almost all of the target species showed that the roots were as sensitive to the leaf extracts as were the shoots.Allelochemicals can aff ect hydrogen adenosine triphosphatase in the plasma membrane, which is responsible for generating the electrochemical proton gradient and thus provides the driving force for the uptake and effl ux of ions and metabolites across the plasma membrane.Hydrogen adenosine triphosphatase inhibition results in reduction of the uptake of minerals and water by roots and, consequently, has a signifi cant eff ect on essential plant functions such as photosynthesis, respiration and protein synthesis, culminating in growth reduction (Gniazdowska & Bogatek 2005).In addition, allelochemicals have been associated with the inhibition of mitosis and disruption of organelle structure (Zhang et al. 2010), due to their eff ects on chromatin organization, altering the physical and chemical proprieties of DNA (Teerarak et al. 2012).Cell growth in plants is dependent on the mitotic process.When cell divisions are disturbed during germination, most seedlings either die before emergence or show abnormalities.
Th e eff ects of allelochemical activity can be detected at the molecular, structural, biochemical, physiological and ecological levels of plant organization (Gniazdowska & Bogatek 2005).Allelopathic compounds can induce secondary oxidative stress, manifesting as increased production of reactive oxygen species (ROS), as demonstrated by Weir et al. (2004).It is known that ROS act as signaling molecules, indicating biotic and abiotic stress (Foyer & Noctor 2005), as well as being major regulators of plant growth and development (Gapper & Dolan 2006).Certain ROSmainly ethylene and abscisic acid-have been shown to act as second messengers in plant hormone responses (Kwak et al. 2006).Ethylene and abscisic acid are both regarded as common stress hormones involved in the regulation of seed dormancy and germination (Kucera et al. 2005).Some allelochemicals of Artemisia annua have been shown to decrease protein content and increase superoxide dismutase activity (Lixiao et al. 2012).According to Oracz et al.   *Values on the same row and followed by same superscript letter do not diff erent signifi cantly from each other.
(2012), myrigalone, extracted from Myrica gale, inhibits the seed germination of Lepidium sativum by impeding the metabolism of gibberellins, the metabolism of abscisic acid, and apoplastic superoxide production, all of which are required for endosperm rupture and consequent embryo growth by elongation.Few studies have focused on the allelopathic potential of Myrtaceae species in Brazil.In a study of sesame and radish (Pina et al. 2009), the leaf extracts of E. dysenterica (Myrtaceae) were found to have no infl uence on the germination of seeds but drastically reduced the seedling elongation rate.Fresh leaf extracts of Blepharocalyx salicifolius (Myrtaceae) have been shown to reduce the survival and germination of lettuce seeds (Mairesse et al. 2007).Souza--Filho et al. (2006) observed that methanolic and hexane extracts of M. guianensis inhibited seed germination of the weeds Senna obtusifolia and Mimosa pudica.Crushed leaves and aqueous leaf extract of the species Campomanesia adamantium and E. dysenterica, tested in soil, have been found to reduce the length of sesame seedlings (Souza et al. 2007).Species of Eucalyptus-a genus originating mainly from Oceania-have been widely examined for allelopathic potential.Currently, approximately 38 Eucalyptus species of Australia have proven to display inhibitory activity against various organisms (Willis 1999).
Th e clusters of the donor species investigated in the present study were not aligned by genus or subtribe, as would be expected.Th is corroborates the results obtained by Santos & Salatino (2000), who observed the non-clustering of Annonaceae species of the cerrado with respect to the chemical composition of the leaf fl avonoids.All donor species studied here belong to the Myrteae tribe, whose systematics have been studied in an attempt to determine the best grouping within the tribe (McVaugh 1968;Salywon et al. 2002;Snow et al. 2003;Wilson et al. 2001;2005).Wilson et al. (2001) stated that the Myrteae tribe is paraphyletic.Indeed, analysis of the matK gene sequence has demonstrated that the Syzygium/Acmena group arose independently.Lucas et al. (2007) showed that the Myrteae tribe would be monophyletic if the Syzygium/Acmena group were excluded.Among the subtribes studied here, Myrtinae and Eugeniinae are paraphyletic, whereas Myrciinae is monophyletic (Lucas et al. 2005).Studies based on nuclear DNA sequencing suggest that the genus Psidium (subtribe Myrtinae) may be paraphyletic (Salywon et al. 2002), whereas Lucas et al. (2005), using nuclear and plastid DNA sequences, demonstrated that this genus is very likely monophyletic.Th e genus Blepharocalyx (subtribe Myrtinae) is extremely isolated, and further research is needed in order to establish the relevance of this group within the tribe (Lucas et al. 2005).Th e genus Eugenia (subtribe Eugeniinae) shows signs of being paraphyletic, furthermore, nuclear and plastid DNA sequence data have demonstrated that the genus Myrcia is also paraphyletic (Lucas et al. 2005).Reports in the literature show that most of the investigated genera have a tendency to be paraphyletic, which underscores the results obtained in the present study.Another factor that explains the lack of clustering of the genera studied is the variation in composition and quantity of allelochemicals in the family Myrtaceae.Keszei et al. (2008) stated that the leaves of Myrtaceae species have high concentrations of terpenes and that these compounds show substantial qualitative and quantitative variations between each taxon, population and individual.Other phytochemical studies of Myrtaceae leaves have identifi ed sesquiterpenes, triterpenes, fl avonoids and alkaloids in Calycorectes psidiifl orus (Domingues et al. 2010); fl avonoids, tannins and phenols in Plinia caulifl ora (Souza-Moreira et al. 2010); and fl avonoids in Baeckea frutescens (Kamiya & Satake 2010).
In the present study, which involved eudicotyledonous and monocotyledonous species, germination and growth were both significantly affected by the aqueous leaf extracts evaluated, demonstrating the phytotoxicity of the donor species.This information might foster the success of agricultural and agro-forestry systems, allowing the identification of interspecific associations.In addition, such extracts could be studied from the perspective of their herbicidal, insecticidal and fungicidal properties, which could make them useful in the battle against the main pests responsible for reducing productivity of those systems.
Our data could also be of use in ecological studies, because these species evaluated here can infl uence the diversity and spatial distribution of individuals in natural communities.On the basis of these results, it is clear that each donor species has a diff erent phytotoxic eff ect.Th ere was no clustering by genus or subtribe.Th erefore, plants from the same environment and with taxonomical proximity showed no similarities in the production of secondary metabolites.For this reason, the hypothesis tested here was rejected; in relation to the production of bioactive compounds, the characteristics intrinsic to each species supersede environmental conditions, as the environment was the same for all species.It was concluded that both the eudicotyledonous and monocotyledonous plants were signifi cantly aff ected by the presence of aqueous leaf extracts.