Insights into the search for new drugs from traditional knowledge: An ethnobotanical and chemical–ecological perspective

Abstract

Context: The variety of pathways for the introduction of a species into the medical collection of traditional communities has led many researchers to question the processes of selection and the use of these resources. A better comprehension of these processes will allow us to understand the cultural dynamics that are related to traditional medical practices, as well as to provide us with new ways in which to facilitate the exploration of natural products.

Objective: This study aims to test the predictive power of the plant apparency hypothesis as it relates to medicinal plant selection by the rural communities of the Caatinga and the Atlantic Forest in northeast Brazil.

Material and methods: Initially, a survey of the medicinal plants used by these communities was conducted using semistructured interviews. Subsequently, data on the life strategies and the habits of each species were collected. More detailed data on the specific plant parts indicated in interviews were also collected. A phytochemical screening for seven classes of chemical compounds was carried out to test the predictions of the plant apparency hypothesis.

Results: The medicinal plants from the Caatinga (especially those that are considered to be trees in habit) have a strong ability to accumulate quantitative compounds, and these species are most likely to be plants with significant biological activity related to these compounds; the medicinal plants from the Atlantic Forest, on the other hand, tend to have a high occurrence of qualitative compounds, especially in herbaceous life forms.

Discussion and conclusion: It was concluded that the plant apparency hypothesis does not adequately explain the selection of medicinal plants in the two environments studied. Our findings highlight some important implications for bioprospecting that need to be further tested experimentally, and through systematic studies, in different regions.

Keywords::

• Bioprospecting
• ethnopharmacology
• plant apparency hypothesis
• traditional medicine
• secondary metabolites

Introduction

The various methods of the introduction of a species into the pharmacopoeia of traditional communities (Bennett & Prance, 2000; Janni & Bastien, 2004; Palmer, 2004) has led many researchers to question the processes of selection & the use of medicinal plants (Voeks, 1996; Stepp and Moerman, 2001; Albuquerque, 2010). The complete comprehension of these processes can lead us to a better understanding of the cultural dynamics related to traditional medical practices, as well as to the development of new ways in which to facilitate the exploration of natural products.

Accordingly, some researchers have questioned whether the habits of plants and their ecological biochemistry could somehow be predictors of which plants might possess medicinal properties (Johns, 1996; Gottlieb et al., 1998; Stepp & Moerman, 2001). For example, Stepp & Moerman (2001) observed a high frequency of medicinal herbs in various traditional pharmacopoeias, suggesting that this may be related to the chemical and ecological aspects of these species. The authors suggest that the plant apparency hypothesis presents a possible scenario for understanding the patterns of selection of certain medicinal plants by different cultures.

The apparency hypothesis of plant defensive strategies was proposed by researchers studying insect–plant relationships (see Albuquerque & Lucena, 2005). Feeny (1976), for example, believes that plants have developed two types of chemical defense against herbivory and they can therefore be classified as “apparent” and “non-apparent” plants. The “apparent” species (plants that are more susceptible to attack by herbivores due to the characteristics of their habits and life cycle) developed quantitative chemical defenses that work by reducing their digestibility and not through the development of high toxicity. These plants are also most often trees and plants with long life cycles. The “non-apparent” species accumulate qualitative defenses that are highly toxic in small quantities because they are highly active compounds. These species are most commonly herbs and plants with short life cycles. For example, the defensive ecology of the family Cruciferaceae seems to be typical of herbaceous plants, where chemical resistance presents itself in the form of small quantities of toxic compounds (Feeny, 1977). Based on these generalizations, and on the selection strategies of medicinal plants used by traditional communities, the high occurrence of herbs in different pharmacopoeias is clearly present and consistent with these principles (see Stepp & Moerman, 2001).

If human populations guide the selection of medical resources, even unconsciously, then a test of the chemical–ecological strategies of these particular plants is necessary to determine if the plant apparency hypothesis can be used as a guide for future strategies in bioprospecting. Albuquerque and Lucena (2005) admit that people can also act as fodder for this theory, and thus, it is necessary to test the predictions of the plant apparency hypothesis within socioenvironmental systems.

This hypothesis is very difficult to evaluate, because there are not clear quantitative variables to test it. Only two studies have tested the relationship of this hypothesis to the ethnobotanical data for semi-arid environments (Almeida et al., 2005; Alencar et al., 2009, 2010).

Thus, this study aims to test the predictions of the plant apparency hypothesis in two different environments: an area of dry forest (the Caatinga from northeastern Brazil) and an area of humid forest (the Atlantic Forest), using the local ethnobotanical data. The following issues are raised: Does the habit and life strategy of a plant predict the occurrence of certain phytocompound classes? Can the habit, life strategy, and chemical composition of a plant explain its relative local importance, as measured by the mean of the versatility of the plant?

Material and methods

Study area

Caatinga

In the Caatinga, the study was performed in the Soledade municipality, located in the Soledade microregion and in the mesoregion of Agreste, Paraíba state, at the coordinates 07°03′26′′ S and 36º21′46′′ W. Soledade is in the Paraíba River Basin and Taperoá River sub-basin (Lacerda et al., 2005), crossed by perennial rivers, but with small flow and low potential for groundwater. The seat of the Soledade municipality is 165.5 km from the capital of Joao Pessoa, 634.9 km² in area and 521 m above sea level (Mascarenhas et al., 2005). The climate is semi-arid with warm summer rains (BSwh′), 7 to 8 months of drought and precipitation of only 400 to 600 mm per year (SEBRAE, 1998). Due to the warm and dry climate, the vegetation consists of deciduous and thorny Caatinga forests (Beltrão et al., 2005), with a dominant shrub layer and few individual trees. This study was conducted in three communities in the Caatinga where the main economic activity is subsistence agriculture: Cachoeira, Bom Sucesso, and Barrocas. Each community is located near Soledad city in the state of Paraiba; Cachoeira is 14 km away, Barrocas is about 18 km away, and Bom Sucesso is about 21 km. The distance between each city is about 7 km.

Atlantic Forest

In the Atlantic Forest, the municipality studied was Igarassu, located in the Itamaracá microregion and the mesoregion of Recife, Pernambuco state, at the coordinates 7°50′00′′ S and 34°54′30′′ W, 30 km from the state capital (CONDEPE/FIDEM, 2007). The climate is tropical, hot, and humid, with rainfall in autumn and winter (As′). The average annual temperature is 25°C, and the average annual rainfall is around 2000 mm (CONDEPE/FIDEM, 2007). The predominant vegetation includes the remnants of Atlantic Forest, scrubs, mangroves, palm trees, and areas of commercial and subsistence agriculture. There are ecological reserves in the municipality, such as the Usina São José Mata (State Law No. 9989 from January 13, 1987), with dense and widespread vegetation of 323.30 hectares located on the Transcanavieira Highway (PE-41) (Igarassu, 2007). The study was conducted in the district of Três Ladeiras, 30 km north of the municipality headquarters (Igarassu, 2007), which is inhabited by 1472 people, of which 876 are adults (436 women and 440 men) (CONDEPE/FIDEM, 2007). Residents of the Três Ladeiras village interact with the native flora in a fragments of the Atlantic Forest, which has been declared as an ecological reserve (January 13, 1987, No. 9989). The surrounding area is a monoculture cane sugar plantation from Usina de São José.

Ethnobotanical data collection

Initially, the Free Informed Consent Term (FICT) was arranged with the residents who agreed to participate in the study. The legal and ethical aspects from resolution 196/96 of the Research Ethics Committee were followed closely to obtain permission to work with and interview the local residents about the plants that are known as medicinal in their culture.

The free list technique was used in the interviews to obtain the largest number of known ethnospecies and their uses in the region (Albuquerque et al., 2008). The following question was used to acquire general information on the botanical species: “What medicinal plants do you know?” Subsequently, a semistructured interview was conducted (Albuquerque et al., 2008) to obtain information on the plant parts used, the method of preparation, and identification. The interviews were conducted in the presence of the household head (over 18 years of age) on each occasion.

In the Caatinga, ethnobotanical data were collected in the municipality of Soledade, with sampling of all households in the communities of Cachoeira, Bom Sucesso, and Barrocas, totaling 55 interviews (40 women and 15 men, between 20 and 88 years), from January to July, 2006. In the Atlantic Forest, the work was performed in the municipality of Igarassu in the community of Três Ladeiras. As a community with a high number of residents, sampling was carried out in 53% of the households, totaling 203 interviews (148 women and 55 men, between 18 and 93 years) from June 2007 to January 2008.

In the Caatinga, the study resulted in the documentation of 100 species, but only 65 were sampled. In the Atlantic Forest, 177 species were recorded and 75 were sampled (for more information, Almeida et al., 2010). The collections were carried out after each interview. In the case of species that were not located near the interviewees’ homes, subsequent collection trips were made to the forest areas with the aid of a forester. The difference in the number of species cited during the interviews and the number of collected species is due to the exclusion of traded species and species that were mentioned by only one interviewee.

Species studied were identified and deposited in the Geraldo Mariz Herbarium (UFP), Federal University of Pernambuco and Vasconcelos Sobrinho Herbarium (PEUFR), Federal Rural University of Pernambuco.

Analysis and data processing

To test the predictions of the plant apparency hypothesis, the collected plants were identified and classified according to their habits and life strategies. In the field, data on the life strategies and habits were obtained for each species, as well as for all of the parts of the plant that were indicated as important for medical use (the whole plant, stem, bark, leaves, flowers, fruits, seeds, root, or latex). For species that had more than one part mentioned, the material was collected from the specific plant organ that was mentioned in the majority of the interviews.

The life strategies of the plants were defined according to Begon et al. (1988), where the letters “r” and “K” are the parameters of the logistic equation. Individuals selected as “r” have high rates of reproduction and growths, being the more likely colonizers of an area, whereas the “K” individuals are species that have a lower potential for reproduction, but have better competitive survivability related to the density of the previous life stage and are present in constant habitats.

To categorize the type of plant habit each species possessed, the concepts proposed by Begon et al. (1988) were adopted. Trees are considered to be perennial woody plants, usually with a single axis stem. Shrubs are also perennial woody plants but are of low stature and typically have many branches. Herbs are not woody plants and have relatively small shoots. The Cactaceae, Agavaceae, and Arecaceae, which occurred only in restricted environmental contexts, were classified using the categories of universal relevance: “tree,” “shrub,” and “herb” (Brown, 1977).

Phytochemical tests were performed for the following compound classes: tannins, saponins, quinones (naphthoquinones and anthraquinones), coumarins, flavonoids, terpenoids (mono-, di-, tri-, and sesquiterpenes), and alkaloids. These compounds were selected due to their pronounced biological activity (Falkenberg, 2004; Henriques et al., 2004; Kuster & Rocha, 2004; Santos & Mello, 2004; Schenkel et al., 2004; Spitzer, 2004; Zuanazzi & Mountain, 2004) and the need to work on compounds of both low and high molecular weight, as are predicted in the plant apparency hypothesis (Feeny, 1976).

The selected compounds were divided into two categories according to Feeny (1976): quantitative compounds (those of high molecular weight and low biological activity, such as tannins, saponins, quinones) and qualitative compounds (those of low molecular weight and high biological activity, such as coumarins, flavonoids, terpenoids, and alkaloids).

Tests were performed with the plant parts described by the respondents as medicinally useful, which were dried in the dark at room temperature after collection. Afterward, 20 g of plant material was crushed and sieved through 2.5-mm mesh. The resulting powder was subjected to cold extraction in 70% ethanol for 5 days. After this period, the extract was filtered with qualitative filter paper. The crude extracts were evaporated until the solvent was completely removed and was kept in a desiccator for 1 week. Subsequently, 20 mg of the dried extract was dissolved in 1 mL ethanol.

The phytochemical testing was performed using thin-layer chromatography (TLC) using Merck silica gel with 60 chromatoplates and with an F₂₅₄ as a fluorescent indicator measuring 0.2 mm in thickness. The elution systems were specifically selected for each of the compound classes following the methods described by Wagner and Bladt (1996) and Harbone (1982). This chromatographic method was chosen because it provides reasonably rapid results and repeatability, and because it is able to qualitatively detect the compounds selected for evaluation. Only the presence and absence of each chemical class was analyzed for each plant species that was indicated as medicinal based on traditional knowledge.

Data analysis

The relative importance (RI) of each plant species was calculated according to the method proposed by Bennett and Prance (2000), with “2” being the maximum value that a species can obtain. Species that obtained higher results were considered more versatile, and present a higher number of properties and bodily systems that can subsequently be treated. This analysis was performed only in order to test if the habit of a species could be a predictor of the versatility of a plant that already possesses medicinal uses. We applied the Kruskal–Wallis test (Sokal & Rholf, 1995) to examine if the RI of local species is associated with the life strategy, habit, chemical composition, and the plant parts used (P < 0.05).

To test the plant apparency hypothesis, the G-test was used (Sokal & Rholf, 1995) to verify the percentage of species with positive results for each class of chemical compounds in relation to the life strategy and habit (P < 0.05). Statistical tests were performed using the statistical package BioEstat 5.0 (Ayres & Ayres-Júnior, 2007).

Results and discussion

From the 65 species identified in the Caatinga region, 41.5% are trees, 17% are shrubs, and 41.5% are herbs. The 75 species of the Atlantic Forest are composed of 39% trees, 30.5% shrubs, and 30.5% herbs. An analysis of the life strategies showed that both species from the Caatinga and those from the Atlantic Forest were predominantly K-strategy species [68 and 77%, respectively (Table 1)], which are considered perennial species.

Table 1.  Medicinal plants collected to study the criteria to be used for the selection and use in the Caatinga and the Atlantic Forest, northeastern Brazil.

Caatinga
Common name	Family	Scientific name	Habit	Life strategy	RI	Chemical compounds
Ameixa	Olacaceae	Ximenia americana L.	Tree	K	1.51	T, S, TE
Angico	Mimosaceae	Anadenanthera colubrina var. cebil (Griseb.) Altschul.	Tree	K	0.96	T, S, F, A
Aroeira	Anacardiaceae	Myracrodruon urundeuva Allemão	Tree	K	1.55	T, S, A
Baraúna	Anacardiaceae	Schinopsis brasiliensis Engl.	Tree	K	0.30	T, S, A
Bom nome	Celastraceae	Maytenus rigida Mart.	Tree	K	1.78	T, S, TE
Cardeiro	Cactaceae	Cereus jamacaru DC.	Tree	K	0.53	C, TE
Catingueira	Caesalpiniaceae	Caesalpinia pyramidalis Tul.	Tree	K	0.68	T, S
Cumaru	Fabaceae	Amburana cearensis (Allemão) A.C. Sm.	Tree	K	1.58	T, F, TE
Favela	Euphorbiaceae	Cnidoscolus quercifolius Pohl.	Tree	K	1.66	C, TE
Feijão brabo	Capparaceae	Capparis flexuosa (L.) L.	Tree	K	0.15	TE
Imbiratanha	Bombacaceae	Pseudobombax marginatum (A. St.-Hill., Juss. & Cambess.) A. Robyns	Tree	K	1.00	T, S, Q, C, TE
Juazeiro	Rhamnaceae	Zizyphus joazeiro Mart.	Tree	K	1.12	S, C
Jucá	Caesalpiniaceae	Caesalpinia ferrea Mart. ex Tul. var. leiostachya Benth.	Tree	K	0.53	T, S
Jurema branca	Mimosaceae	Piptadenia stipulacea (Benth.) Ducke.	Tree	K	0.49	Q, F, TE
Jurema preta	Mimosaceae	Mimosa tenuiflora (Willd.) Poir.	Tree	K	0.72	T, S, Q, C, C
Mororó	Caesalpiniaceae	Bauhinia cheilantha (Bong.) Steud.	Tree	K	1.64	S, Q, F, TE
Mulungu	Fabaceae	Erythrina velutina Willd.	Tree	K	1.22	TE
Pereiro	Apocynaceae	Aspidosperma pyrifolium Mart.	Tree	K	0.87	T, Q, C, TE, A
Quixabeira	Sapotaceae	Sideroxylon obtusifolium (Roem. & Schult.) T.D. Penn.	Tree	K	1.56	T, S, Q, TE
Umburana	Burseraceae	Commiphora leptophloeos (Mart.) J.B.	Tree	K	1.61	T, S, C, TE
Umbuzeiro	Anacardiaceae	Spondias tuberosa Arruda	Tree	K	0.15	T, S, C
Coroa de frade	Cactaceae	Melocactus sp.	Shrub	K	0.15	C, TE
Macambira de pedra	Bromeliaceae	Encholirium sp.	Shrub	K	0.15	T, S, TE
Marmeleiro	Euphorbiaceae	Croton blanchetianus Baill.	Shrub	K	0.53	T, S, C, TE
Marmeleiro branco	Euphorbiaceae	Croton sincorensis Mart. ex Müll. Arg.	Shrub	K	0.34	S, F, C, TE, A
Oliveira	Solanaceae	Nicotiana glauca Graham	Shrub	K	0.15	T, S, Q, F, C, TE, A
Pinhão brabo	Euphorbiaceae	Jatropha mollissima (Pohl) Baill.	Shrub	K	0.30	T, S, C, A
Cabeça de negro	Cucurbitaceae	Wilbrandia sp.	Herb	K	0.15	S, C, TE
Capitão	Amaranthaceae	Gomphrena basilanata Suesseng.	Herb	r	0.38	S, C, TE
Caruá	Bromeliaceae	Neoglaziovia sp.	Herb	K	0.30	T, F
Gogoia	Solanaceae	Solanum agrarium Sendtner.	Herb	r	0.46	S, C, A
Mão fechada	Selaginellaceae	Selaginella convoluta (Arn.) Spring.	Herb	r	0.15	S, Q
Salambaia	Bromeliaceae	Tillandsia recurvata (L.) L.	Herb	r	0.30	T, S, TE
Salambaia do pinhão brabo	Bromeliaceae	Tillandsia streptocarpa Baker	Herb	r	0.34	T, S, F, TE
Urtiga branca	Loasaceae	Aosa rupestris (Gardner) Weigend	Herb	r	0.95	S, C, TE
Algaroba	Mimosaceae	Prosopis juliflora (Sw.) DC.	Tree	K	0.15	T, S, F, C, TE
Cajueiro roxo	Anacardiaceae	Anacardium occidentale L.	Tree	K	1.58	T, S,Q, TE
Coqueiro	Arecaceae	Cocos nucifera L.	Tree	K	0.30	C, TE
Goiabeira	Myrtaceae	Psidium guajava L.	Tree	K	0.34	T, S, Q, F, TE, F
Graviola	Annonaceae	Annona muricata L.	Tree	K	0.15	T, S, Q, F, TE, F
Mamoeiro	Caricaceae	Carica papaya L.	Tree	K	0.49	T, S, Q, F, TE
Alfazema braba	Lamiaceae	Hyptis suaveolens Poit.	Shrub	K	0.15	T, S, Q, F, TE
Herb cidreira	Verbenaceae	Lippia alba (Mill.) N.E.Br.	Shrub	K	1.57	Q, TE
Hortelã do pará	Lamiaceae	Plectranthus barbatus Andr.	Shrub	K	0.57	T, S, Q, F, TE, A
Maracujá	Passifloraceae	Passiflora edulis Sims.	Shrub	K	0.30	S, Q, TE
Palma	Cactaceae	Opuntia ficus-indica (L.) Mill.	Shrub	K	0.15	Q, TE
Alecrim	Lamiaceae	Rosmarinus officinalis L.	Herb	K	0.88	T, S, F, TE
Arruda	Rutaceae	Ruta graveolens L.	Herb	r	1.07	T, S, Q, F, C, TE, A
Capim santo	Poaceae	Cymbopogon citratus (DC.) Stapf	Herb	K	1.30	T, S, F, C, TE, A
Espinho de cigano	Asteraceae	Acanthospermum hispidum DC.	Herb	r	0.45	S, Q, F, TE
Hortelã graúda	Lamiaceae	Plectranthus amboinicus (Lour.) Spreng.	Herb	K	0.46	S, Q, F, TE
Hortelã miúda	Lamiaceae	Mentha piperita L.	Herb	r	1.76	T, S, Q, F, C, TE
Macela	Asteraceae	Egletes viscosa (L.) Less.	Herb	r	0.38	Q, TE
Malva branca	Sterculiaceae	Waltheria rotundifolia Schrank	Herb	r	0.15	T, S, C, TE
Malva rosa	Sterculiaceae	Melochia tomentosa L.	Herb	r	0.95	T, S, Q, F, C, TE
Manjericão	Lamiaceae	Ocimum basilicum L.	Herb	r	0.60	T, Q, C, TE
Maria três babados	Lamiaceae	Hyptis sp.	Herb	r	0.15	Q, F, TE
Mastruz	Amaranthaceae	Chenopodium ambrosioides L.	Herb	r	1.19	T, S, Q, TE
Melância	Cucurbitaceae	Citrillus vulgaris Schard.	Herb	r	0.30	T, S
Mussambê	Capparaceae	Cleome spinosa Jacq.	Herb	K	0.46	S, C, TE
Novalgina	Amaranthaceae	Alternanthera dentata (Moench.) Stuchlik. ex R.E. Fr.	Herb	r	0.34	T, S, Q
Pega pinto	Nyctaginaceae	Boerhavia diffusa L.	Herb	r	0.19	S, C
Penicilina	Amaranthaceae	Alternanthera sp.	Herb	r	0.19	S, F
Saião	Crassulaceae	Kalanchoe brasilensis Camb.	Herb	r	0.84	S, F, C
Quebra pedra	Euphorbiaceae	Phylanthus sp.	Herb	r	0.57	T, S, Q, F, TE
Atlantic Forest
Common name	Family	Scientific name	Habit	Life strategy	RI	Phytocompounds
Aroeira	Anacardiaceae	Schinus terebinthifolius Raddi	Tree	K	1.60	T, S, Q, F, TE
Babatenom	Mimosaceae	Pithecellobium cochliocarpum (Gomez) Macbr.	Tree	K	1.14	S, Q, TE, A
Canela	Lauraceae	Nectandra cuspidata Ness & Mart.	Tree	K	0.95	S, Q, F, C, TE, A
Embaúba branca	Cecropiaceae	Cecropia palmata Willd.	Tree	K	0.42	S, Q, F, TE, A
Espinheiro	Mimosaceae	Acacia sp.	Tree	K	0.12	C
Imbira branca	Flacourtiaceae	Sem identificação 1	Tree	K	0.12	S, TE
Imbira vermelha	Annonaceae	Xylopia frutescens Aubl.	Tree	K	0.12	S, Q, F, TE
Lacre	Clusiaceae	Vismia guianensis (Aubl.) Choisy	Tree	K	0.42	T, S, F, TE, A
Mutamba	Sterculiaceae	Guazuma ulmifolia L.	Tree	K	0.24	S, Q, C, TE
Macaíba	Arecaceae	Acrocomia intumescens Drude	Tree	K	0.76	S, F, TE
Jenipapo	Rubiaceae	Genipa americana L.	Tree	K	0.64	C
João mole	Nyctaginaceae	Guapira sp.	Tree	K	0.24	T, S, Q, F, TE
Pau daico roxo	Bignoniaceae	Tabebuia avellanedae Lorentz ex Griseb.	Tree	K	0.54	T, S, TE
Pata de vaca	Caesalpiniaceae	Bauhinia forficata Link	Tree	K	0.30	S, Q, F, TE
Sucupira preta	Fabaceae	Bowdichia virgilioides Kunth	Tree	K	0.39	F, C, TE, A
Araça	Myrtaceae	Psidium guineense Sw.	Shrub	K	0.12	T, Q, F, TE
Cipó japecanga	Smilacaceae	Smilax rotundifolia L.	Shrub	K	0.12	S, Q, F
Chumbinho	Verbenaceae	Lantana camara L.	Shrub	K	0.19	S, Q, F
Acônito	Amaranthaceae	Pfaffia glomerata (Spreng.) Pederson	Herb	r	0.27	S, Q, F
Alho do mato	Iridaceae	Eleutherine bulbosa (Mill.) Urb.	Herb	r	0.15	S, F, TE
Cana de macaco	Zingiberaceae	Costus sp.	Herb	r	0.84	Q
Cipó cururu	Sapindaceae	Serjania sp.	Herb	K	0.12	S, Q, F, C, TE
Melão são caetano	Cucurbitaceae	Momordica charantia L.	Herb	r	0.12	Q, F
Pepaconha	Violaceae	Hybanthus sp.	Herb	r	0.51	S, Q, F
Salambaia	Bromeliaceae	Tillandsia usneoides (L.) L.	Herb	r	0.24	Q, F, C
Urinana	Fabaceae	Zornia sp.	Herb	r	0.15	F
Abacate	Lauraceae	Persea americana Mill.	Tree	K	0.62	S, Q, F
Acerola	Malpighiaceae	Malpighia glabra L.	Tree	K	0.51	A
Azeitona preta	Myrtaceae	Syrygium jambolanum (Lam.) DC.	Tree	K	0.69	T, Q, F
Carambola	Oxalidaceae	Averrhoa carambola L.	Tree	K	0.30	Q, F
Eucalipto	Myrtaceae	Eucalyptus citriodora Hook.	Tree	K	0.24	T, Q, F, A
Goiaba	Myrtaceae	Psidium guajava L.	Tree	K	0.24	T, S, Q, F, TE, A
Jaca	Moraceae	Artocarpus integrifolia L. f.	Tree	K	0.24	S, Q, F, C, TE, A
Juá	Rhamnaceae	Zizyphus joazeiro Mart.	Tree	K	0.76	A
Laranja	Rutaceae	Citrus sinensis (L.) Osbeck	Tree	K	0.93	S, Q, F, C, TE
Limão	Rutaceae	Citrus limonia Osbeck	Tree	K	0.57	S, F, C, TE, A
Manga	Anacardiaceae	Mangifera indica L.	Tree	K	0.62	T, S, Q, F, A
Mamão	Caricaceae	Carica papaya L.	Tree	K	0.46	Q, A
Pinha	Annonaceae	Annona squamosa L.	Tree	K	0.12	Q, F, A
Siriguela	Anacardiaceae	Spondias purpurea L.	Tree	K	0.27	T, Q, F
Algodão	Malvaceae	Gossypium barbadense L.	Shrub	K	0.20	F
Alcachofra	Asteraceae	Vernonia condensata Baker	Shrub	K	0.83	Q, F, TE, A
Alfavaca de caboclo	Lamiaceae	Ocimum campechianum Mill.	Shrub	r	0.69	Q, F, TE
Boa noite branca	Apocynaceae	Catharanthus roseos L. G. Don.	Shrub	k	0.27	S, Q, F, C, TE, A
Café	Vitaceae	Coffea sp.	Shrub	K	0.27	S, Q, F, TE, A
Colônia	Zingiberaceae	Alpinia zerumbet (Pers.) Burt. ex R. M. Smith	Shrub	K	0.91	S, F, TE
Herb cidreira	Verbenaceae	Lippia alba (Mill.) N.E.Br.	Shrub	K	1.86	Q, F, TE
Flor de sabugo	Caprifoliaceae	Sambucus nigra L.	Shrub	K	0.57	S, Q, F, TE, A
Hortelã grande	Lamiaceae	Plectranthus amboinicus (Lour.) Spreng.	Shrub	K	1.22	S, Q, F, TE
Insulina	Vitaceae	Cissus verticillata (L.) Nicolson & C.E. Jarvis	Shrub	K	0.12	Q, F, TE
Jurubeba	Solanaceae	Solanum paniculatum L.	Shrub	K	0.35	S, Q, F, TE, A
Liamba	Verbenaceae	Vitex agnus-castus L.	Shrub	K	0.24	S, Q, F
Mastruz	Chenopodiaceae	Chenopodium ambrosioides L.	Shrub	r	0.68	Q, F
Mussambê	Capparaceae	Cleome spinosa Jacq.	Shrub	K	0.30	S, Q, F, C
Pitanga	Myrtaceae	Eugenia uniflora L.	Shrub	K	0.59	T, Q, F
Pimenta	Solanaceae	Capsicum frutescens L.	Shrub	K	0.24	S, Q, F, A
Pimenta do reino	Piperaceae	Piper nigrum L.	Shrub	K	0.24	S, Q, F, TE, A
Pinhão roxo	Euphorbiaceae	Jatropha gossypiifolia L.	Shrub	K	0.35	S, Q, F
Rosa branca	Rosaceae	Rosa sp.2	Shrub	K	0.62	T, S, F
Salsa	Convolvulaceae	Ipomoea sp.	Shrub	K	0.15	Q, F, TE
Agrião	Brassicaceae	Nasturtium officinale R. Br.	Herb	K	0.53	Q, F, TE, A
Arruda	Rutaceae	Ruta graveolens L.	Herb	r	1.08	Q, F, TE, A
Alvomoura	Solanaceae	Solanum sp.	Herb	r	0.30	S, Q, F, TE, A
Capim santo	Poaceae	Cymbopogon citratus (DC.) Stapf	Herb	K	1.47	S, F, TE, A
Cardo santo	Papaveraceae	Argemone mexicana L.	Herb	r	0.24	F, C, A
Herb doce	Apiaceae	Pimpinella anisum L.	Herb	r	1.09	S, F, C
Hortelã pequena	Lamiaceae	Mentha piperita L.	Herb	K	2.00	Q, F, TE
Hortelã pastilha	Lamiaceae	Mentha pulegium L.	Herb	r	1.07	S, Q, F, TE
Manjericão	Lamiaceae	Ocimum basilicum L.	Herb	r	1.16	Q, F
Manjirioba	Caesalpiniaceae	Senna occidentalis (L.) Link.	Herb	r	0.76	S, Q, F, A
Pega pinto	Nyctaginaceae	Boerhavia diffusa L.	Herb	K	0.83	Q, F
Quebra pedra	Euphorbiaceae	Phyllanthus niruri L.	Herb	r	0.46	T, Q, F
Rabo de paposa	Asteraceae	Conyza sumatrensis (Retz.) E. Walken	Herb	r	0.12	Q, F, TE, A
Timpi	Phytolacaceae	Petiveria alliacea L.	Herb	r	0.54	Q, F
Veiga morta	Lamiaceae	Mentha sp.	Herb	r	0.84	Q, F

RI: Relative importance, T: Tannins, S: Saponins, Q. Quinones, F: Flavonoids, C: Coumarins, TE: Terpenoids A: Alkaloids.

RI of species

Of the five Caatinga species that obtained better results on their RI, four are native and are classified as trees: Maytenus rigida Mart. (Celastraceae) (1.78), Cnidoscolus quercifolius Pohl. (Euphorbiaceae) (1.66), Bauhinia cheilantha (Bong.) Steud. (Caesalpiniaceae) (1.64), and Commiphora leptophloeos (Mart) J.B. Gillet (Burseraceae) (1.61). The other species, Mentha piperita L. (Lamiaceae), is exotic and herbaceous and has a RI of 1.76 (Table 1). In a study conducted in the Caatinga by Alencar et al. (2010), plants with a great versatility, as measured by RI, were also native trees, including: Caesalpinia ferrea Mart. (Caesalpiniaceae) (1.81), Myracrodruon urundeuva Allemão (Anacardiaceae) (1.75), and Mimosa tenuiflora (Willd.) Poir. (Mimosaceae) (1.50). Alencar et al. (2009, 2010) also emphasize that from the 10 species with the highest RI values, nine were trees, which again shows the versatility of species with this habit.

In the Atlantic Forest, from the five species that stood out, just one was native and a tree, Schinus terebinthifolius Raddi (Anacardiaceae) (1.60). The others were exotic and herbaceous in habit, such as Mentha piperita (2.00), Lippia alba (Mill.) N.E..Br. (Verbenaceae) (1.86), Cymbopogon citratus (DC.) Stapf. (Poaceae) (1.47), and Plectranthus amboinicus (Lour.) Spreng. (Lamiaceae) (1.22). One exotic species is highlighted in both study sites, Mentha piperita (Table 1).

The trees in the Caatinga tended, on average, to receive higher scores in RI in relation to the shrubs (H = 6.7313, P < 0.05) and the herbs (H = 4.66, P < 0.05) (Table 2). These results differ from those obtained by Almeida et al. (2005), in which the highest RI scores obtained from the species studied in the other regions of the Caatinga where independent of habit, life strategy, and compounds classes, despite woody plants tending to have higher scores.

Table 2.  Summary of the Kruskal–Wallis test based on the relative importance (RI), life strategies, habit, and phytocompounds, to study the criteria for the selection and use of medicinal plants in the Caatinga and the Atlantic Forest, northeastern Brazil (mean + standard deviation).

		Caatinga	Atlantic Forest
Life strategies	K	0.74 ± 0.56	0.53 ± 0.44
	r	0.55 ± 0.42	0.69 ± 0.35
K × r		H = 0.7273, P = 0.3938	H = 3.3755, P = 0.0662
Habit	Tree	0.91 ± 0.58	0.46 ± 0.27
	Shrub	0.39 ± 0.42	0.53 ± 0.42
	Herb	0.56 ± 0.41	0.83 ± 0.50
Tree × Shrub		H = 6.7313, P = 0.0095	H = 0.0274, P = 0.8685
Tree × Herb		H = 4.66, P = 0.0309	H = 3.4273, P = 0.0641
Shrub × Herb		H = 2.9593, P = 0.0854	H = 3.6176, P = 0.0572
Phytocompounds	Quantitative	0.68 ± 0.51	0.54 ± 0.42
	Qualitative	0.72 ± 0.54	0.56 ± 0.46
Quantitative × Qualitative		H = 0.0891, P = 0.7654	H = 0.0028, P = 0.9576

Thus, in the variety of species studied in this work, the RI was independent of the variables analyzed, except in the Caatinga region where the habit of the species was related to the RI. In the Atlantic Forest, there was no significant difference for the variables analyzed (Table 2), although herbs tended to have the highest RI scores.

Apparency hypothesis: compounds classes versus habit

All of the species studied possessed at least one of the seven classes of compounds. The Caatinga species presented the following configuration: saponins (76.9%), terpenes (72.3%), and tannins (60%). In the Atlantic Forest species, the configuration was different and was constituted by: flavonoids (86.7%), quinones (76%), and saponins (54.7%), which stood out (Figure 1). We believe that the next step in future studies is to include quantitative tests for total phenols.

Figure 1.  Positive correlations among the species studied related to phytocompounds of the medicinal plants used in the Caatinga and Atlantic Forest, northeastern Brazil.

In the Caatinga, tannins were predominantly found in all of the habits studied, including trees (G = 10.13, P < 0.01), shrubs (G = 22.74, P < 0.0001) and herbs (G = 8.40, P < 0.01), whereas in the Atlantic Forest, tannins stood out only in shrubs (G = 16.70, P < 0.0001) and herbs (G = 9.52, P < 0.01) (Table 3). The plant apparency hypothesis predicts that qualitative compounds, such as alkaloids, flavonoids, and terpenoids, will have a higher occurrence in herbs. Our results were contradictory to this prediction. We found qualitative phytocompounds, flavonoids (G = 50.54, P < 0.0001) and alkaloids (G = 16.40, P < 0.01), to be more common in trees from the Caatinga, whereas in the Atlantic Forest, there was no distinction in the qualitative compounds between the different habits (see Table 3). There is a clear trend in the quantity and quality of compounds and that occur in the Caatinga and in the Atlantic Forest. We expected that the “apparent” species would tend to accumulate quantitative chemical defenses, whereas the “non-apparent” species would present qualitative chemical defenses (Feeny, 1976). The findings of our study allow us to reject the plant apparency hypothesis.

Table 3.  Summary of the G-test related to the presence of all compound classes from the medicinal plants collected to study the criteria to be used for the selection and use of medicinal plants in the Caatinga and the Atlantic Forest, northeastern Brazil.

		Tannins	Saponins	Quinones	Flavonoids	Coumarins	Terpenoids	Alkaloids
Caatinga	Tree	G = 10.13, P = 0.0015	G = 10.13, P = 0.0015	G = 32.03, P < 0.0001	G = 50.54, P < 0.0001	G = 3.78, P = 0.0518	G = 14.62, P = 0.0001	G = 16.40, P = 0.0015
	Shrub	G = 22.74, P < 0.0001	G = 25.80, P < 0.0001	G = 96.82, P < 0.0001	—	G = 181.43, P < 0.0001	—	G = 25.80, P < 0.0001
	Herb	G = 8.40, P = 0.0038	G = 0.44, P = 0.5028	G = 47.99, P < 0.0001	G = 22.79, P < 0.0001	G = 0.13, P = 0.713	G = 2.89, P = 0.08887	G = 0.19, P = 0.6574
	K	G = 0.64, P = 0.421	G = 3.20, P = 0.0735	G = 31.52, P < 0.0001	G = 27.13, P < 0.0001	G = 4.67, P = 0.0305	G = 25.42, P < 0.0001	G = 0.76, P = 0.3809
	r	G = 14.50, P = 0.0001	G = 29.95, P < 0.0001	G = 54.24, P < 0.0001	G = 30.56, P < 0.0001	G = 0.21, P = 0.6405	—	G = 4.98, P = 0.0256
Atlantic Forest	Tree	G = 1.89, P = 0.169	G = 29.99, P < 0.0001	G = 14.51, P = 0.0001	G = 8.22, P = 0.0041	G = 3.58, P = 0.0583	G = 74.44, P < 0.0001	G = 19.55, P < 0.0001
	Shrub	G = 16.70, P < 0.0001	G = 0.96, P = 0.3252	G = 22.01, P < 0.0001	—	G = 14.38, P = 0.0001	G = 14.50, P = 0.0001	G = 56.00, P < 0.0001
	Herb	G = 9.52, P = 0.002	G = 5.76, P = 0.0163	G = 0.71, P = 0.3968	G = 18.16, P < 0.0001	G = 15.95, P < 0.0001	G = 10.35, P = 0.0013	G = 79.19, P < 0.0001
	K	G = 0.90, P = 0.3402	G = 10.81, P = 0.001	G = 10.32, P = 0.0013	G = 22.28, P < 0.0001	G = 7.70, P = 0.0055	G = 13.52, P = 0.0002	G = 62.63, P < 0.0001
	r	G = 10.98, P = 0.0009	G = 3.15, P = 0.0756	G = 5.13, P = 0.0234	G = 20.92, P < 0.0001	G = 5.13, P = 0.0234	G = 15.51, P < 0.0001	G = 13.52, P = 0.0002

(—): Unable to complete the test due to zero values.

According to the results presented in our work, and based on all of the study species collected from the Caatinga and the Atlantic Forest, the habit and the life strategy of the plant are not predictors of the occurrence of certain phytocompound classes.

The herbs from Caatinga demonstrated a high occurrence of quantitative compound, such as tannins (G = 8.40, P < 0.01) and quinones (G = 47.99, P < 0.0001). In the Atlantic Forest, it was demonstrated that herbs possessed more qualitative compounds, such as flavonoids (G = 18.16, P < 0.0001), terpenoids (G = 10.35, P < 0.001), and alkaloids (G = 79.19, P < 0.0001) (Table 3).

Even with the low occurrence of alkaloids in the Caatinga species (see Figure 1), trees (G = 16.40, P < 0.01) and shrubs (G = 25.80, P < 0.0001) showed significant results for the presence of these compounds. However, Pelletier (1983) maintains that alkaloids are fairly common in plants. Contrary to our results, Levin (1976) reports that there is no difference in the percentages of alkaloids between herbaceous and woody plants, although annual plants can contain more alkaloids than perennial species. This assertion can only be checked for the medicinal plants from Atlantic Forest.

Overall, our findings for the Caatinga plants are corroborated by previous tests of the plant apparency hypothesis by using the same methodological procedures (Almeida et al., 2005; Alencar et al., 2009). Although these results lead to a rejection of the plant apparency hypothesis, a curious pattern emerges; quantitative compounds (i.e., tannins) have a broad occurrence in plants from the Caatinga, independent of habit, whereas in the Atlantic Forest there is a significant occurrence of qualitative compounds (i.e., alkaloids). In other phytochemical and herbivory studies, the findings followed the expected distributions of compounds, in contrast to the present study (see, e.g., McKey, 1984; Heil et al., 2002). This apparent discrepancy needs to be further discussed.

The production of bioactive compounds in plants is due to different environmental influences. In general, in addition to adaptations to UV radiation (Gottlieb, 1987), there are other phenomena in the environment which require special adaptations including defense mechanisms of these plants to environmental conditions, microorganisms, insects, and animals (Montanari & Bolzani, 2001), as well as adaptation to low density, humidity, and other environmental challenges.

Accordingly, bioprospecting studies for certain compounds should take into consideration the environment in which plants are developed. Gottlieb (1987), for example, argues that plants in arid and semi-arid regions, such as those in the Caatinga, tend to have favored the metabolic pathways that lead to the production of high-molecular-weight compounds (such as tannins). There is evidence, for example, that plants that are commonly used as anti-inflammatory agents in the Caatinga tend to have higher concentrations of tannic compounds, rather than other compounds that could also justify this activity, such as flavonoids (de Araújo et al., 2008). Flavonoids were very frequent in medicinal plants from both regions. This may be related to the deposition of flavonoids in the leaf surface for protection against UV radiation, heat reduction, herbivory and as an antimicrobial agent (Gottlieb, 1987).

Our findings indicate some important patterns that can be applied to bioprospecting that need to be tested with both experimental and systematic studies in different regions. The patterns indicate that medicinal plants (especially trees) from semi-arid regions (i.e. the Caatinga) are strongly orientated to accumulate quantitative compounds (such as tannins and saponins), and it is therefore more likely to find plants with significant biological activity related to these types of compounds in these habitats. The medicinal plants of wet forests (i.e. the Atlantic Forest), for example, tend to have a high occurrence of qualitative compounds (such as coumarins and alkaloids), and these compounds are especially prevalent in herbs.

Acknowledgements

The authors thank CNPq (“edital universal”) for financial support to the PhD student Cecilia de Fatima C.B.R. de Almeida and grants to U.P. Albuquerque. The Iana Moura and Reinaldo Lucena for help in ethnobotanical data collection in Soledade (PB) and the study participants from Cachoeira, Barrocas, Bom Sucesso, and Três Ladeiras for participating in this study and for valuable exchange of knowledge. The authors also thank Camila C.B.R. de Almeida, and Ana Carolina O. da Silva for technical assistance, and to the anonymous reviewers for suggestions and critical ideas.

Declaration of interest

We also thank the contribution of the project “Sustainability of remnants of the Atlantic rainforest in Pernambuco and its implications for conservation and local development”, a Brazilian-German scientific cooperation within the program “Science and Technology for the Atlantic Rainforest” funded by CNPq (590039/2006-7) and BMBF (01 LB 0203 A1).