Antimicrobial lectin from Schinus terebinthifolius leaf

Authors


Correspondence

Patrícia M.G. Paiva, Departamento de Bioquímica, CCB, Universidade Federal de Pernambuco, Avenida. Prof. Moraes Rego S/N, Cidade Universitária, 50670-420, Recife-PE, Brazil. E-mail: ppaivaufpe@yahoo.com.br

Abstract

Aims

Schinus terebinthifolius leaves are used for treating human diseases caused by micro-organisms. This work reports the isolation, characterization and antimicrobial activity of S. terebinthifolius leaf lectin (SteLL).

Methods and Results

The isolation procedure involved protein extraction with 0·15 mol l−1 NaCl, filtration through activated charcoal and chromatography of the filtrate on a chitin column. SteLL is a 14-kDa glycopeptide with haemagglutinating activity that is inhibited by N-acetyl-glucosamine, not affected by ions (Ca2+ and Mg2+) and stable upon heating (30–100°C) as well as over the pH 5·0–8·0. The antimicrobial effect of SteLL was evaluated by determining the minimal inhibitory (MIC), bactericide (MBC) and fungicide (MFC) concentrations. Lectin was active against Escherichia coli, Klebsiella pneumoniae, Proteus mirabilis, Pseudomonas aeruginosa, Salmonella enteritidis and Staphylococcus aureus. Highest bacteriostatic and bactericide effects were detected for Salm. enteritidis (MIC: 0·45 μg ml−1) and Staph. aureus (MBC: 7·18 μg ml−1), respectively. SteLL impaired the growth (MIC: 6·5 μg ml−1) and survival (MFC: 26 μg ml−1) of Candida albicans.

Conclusions

SteLL, a chitin-binding lectin, purified in milligram quantities, showed antimicrobial activity against medically important bacteria and fungi.

Significance and Impact of the Study

SteLL can be considered as a new biomaterial for potential antimicrobial applications.

Introduction

Antimicrobials have been used to treat diseases caused by bacteria and fungi and therefore have significantly contributed to a reduced mortality rate in humans and animals. The emergence of micro-organisms resistant to frequently used commercial antimicrobial drugs, such as methicillin, oxacillin and penicillin, has stimulated the evaluation of medicinal plants as sources of compounds for inhibiting the growth or survival of micro-organisms. Plant lectins with antibacterial and antifungal activities against micro-organisms that cause human diseases have been isolated from seeds, heartwood, cladodes and leaves (Santi-Gadelha et al. 2006; Oliveira et al. 2008; Sá et al. 2009; Santana et al. 2009; Costa et al. 2010; Charungchitrak et al. 2011).

The antibacterial activity of lectins occurs through the interaction with N-acetylglucosamine, N-acetylmuramic acid (MurNAc) and tetrapeptides linked to MurNAc present in the cell wall of Gram-positive bacteria or to lipopolysaccharide present in the cell walls of Gram-negative bacteria (Dziarski et al. 2000). Previous studies have revealed that isolectin I from Lathyrus ochrus seeds binds to muramic acid and muramyl dipeptide – two components commonly found in bacterial cell walls – through hydrogen bonds between ring hydroxyl oxygen atoms of the sugar and the carbohydrate-binding site of lectin as well as through hydrophobic interactions with the side chains of residues Tyr100 and Trp128 of isolectin I (Bourne et al. 1994).

The antifungal activity of lectins occurs through an interaction with the fungal cell wall, which is composed of chitin, glucans and other polymers (Adams 2004). Chitin-binding lectins can impair the synthesis and/or deposition of chitin in the cell wall as well as prevent hyphal development and spore germination (Lis and Sharon 1981; Selitrennikoff 2001; Trindade et al. 2006). It has been also suggested that small antifungal lectins such as hevein (4·7 kDa) and poutein (14 kDa) can penetrate the fungal cell wall to reach the plasma membrane, where they can block the active sites of enzymes involved in cell wall morphogenesis (Van Parijs et al. 1991; Boleti et al. 2007).

Schinus terebinthifolius Raddi (Anacardiaceae family), known as Brazilian pepper, is a tree that is distributed worldwide. The leaves are used topically in Brazil for healing and tissue repair of skin wounds. Moreover, its leaves are commonly used as an infusion for treating infections in the respiratory, digestive and urinary tracts, as well as against rheumatism and oral candidiasis (Martinez et al. 1996; Ribas et al. 2006; Lindenmaier 2008). The essential oil of the S. terebinthifolius leaf is used to treat respiratory problems, mycosis and candidal infections (topical use). Its activity has been attributed to its high concentrations of monoterpenes (Lloyd et al. 1977). S. terebinthifolius leaf essential oil has shown antibacterial activity against Bacillus subtilis, Escherichia coli, Pseudomonas aeruginosa, Shigella dysenteriae, Staphylococcus albus, Staphylococcus aureus and Staphylococcus intermedius as well as antifungal activity against Aspergillus niger, Aspergillus parasiticus and Candida albicans (El-Massry et al. 2009; Silva et al. 2010). Extracts from S. terebinthifolius leaves in ethanol and dichloromethane containing secondary metabolites such as phenols, flavones, flavonoids, xanthones, leucoanthocyanidins, flavanones and free steroids were active against E. coli, Ps. aeruginosa, Staph. aureus and C. albicans (Lima et al. 2006; El-Massry et al. 2009). Among 23 extracts from 12 Cuban plants, an aqueous extract from the leaves of S. terebinthifolius showed the highest activity against Staph. aureus, and it could inhibit the growth of B. subtilis (Martinez et al. 1996).

This work reports the isolation and characterization of a chitin-binding lectin from S. terebinthifolius leaf (SteLL) and the effect of this protein on the growth of bacteria and fungi that cause human diseases.

Materials and methods

Leaf extract

Leaves of S. terebinthifolius were collected in Recife City, State of Pernambuco, in north-eastern Brazil. A voucher specimen is archived (number 73 431) in the Instituto Agronômico de Pernambuco (IPA), Recife, Brazil.

Dried leaves (20 g) were crushed to powder (40 mesh) and suspended in 0·15 mol l−1 NaCl (200 ml). After agitation for 16 h at 4°C, the sample was filtered through gauze and centrifuged at 3000 g for 15 min. The supernatant (crude extract) was passed through activated charcoal (10%, w/v; Reagen, Paraná, Brazil), and the filtrate (leaf extract) was collected.

Isolation of Schinus terebinthifolius leaf lectin (SteLL)

Leaf extract (11·4 mg of protein) was loaded onto a chitin (Sigma-Aldrich, MO, USA) column (7·5 × 1·5 cm) equilibrated (20 ml h−1 flow rate) with 0·15  NaCl. After extensive washing with equilibrating solution (100 ml), SteLL was recovered by elution using 1·0 mol l−1 acetic acid (80 ml). The elution product was dialysed in a 10-kDa cut-off membrane (Sigma-Aldrich) against distilled water (4 h, 4°C) and in sequence against 0·15 mol l−1 NaCl (4 h, 4°C).

Total phenol content

Total phenol content of the crude extract, leaf extract and SteLL was determined using the Folin–Ciocalteu method based on the reduction of phosphomolybdic–phosphotungstic acid reagent in an alkaline medium (Morais et al. 1999). The Folin–Ciocalteu's reagent (1 : 10 solution in distilled water; 2·5 ml) and sodium carbonate (75 g l−1; 2 ml) were added to the samples (0·5 ml), and the mixtures were incubated at 50°C for 5 min. After cooling for 30 min, absorbance was measured at 760 nm. Phenol content was determined based on a standard curve of tannic acid (9·6–48 mg ml−1).

Haemagglutinating activity

In lectinology, the haemagglutinating assay is the classic tool to assess the carbohydrate-binding property of a lectin, making sure that it is active. A haemagglutinating assay was carried out in microtitre plates (TPP-Techno Plastic Products, Trasadingen, Switzerland) according to the method described by Paiva and Coelho (1992). A serial 2-fold dilution of the lectin preparation (50 μl) was prepared with 0·15 mol l−1 NaCl before incubation with a suspension (2·5% v/v) of glutaraldehyde-fixed rabbit erythrocytes (Bing et al. 1967). One haemagglutination unit (titre) was defined as the reciprocal of the highest dilution of the sample promoting full erythrocyte agglutination (Napoleão et al. 2011). Specific haemagglutinating activity was defined as the ratio between the titre and protein concentration (mg ml−1) determined according Lowry et al. (1951). Increase in specific haemagglutinating activity reveals lectin concentration and purification.

Haemagglutinating activity of SteLL (0·5 mg ml−1; 50 μl) was also evaluated in the presence of metal ions, at different pH values, or after heating to different temperatures. The effect of Ca2+ and Mg2+ was evaluated according to Pajic et al. (2002) using the haemagglutinating activity assay described above but by replacing the 0·15 mol l−1 NaCl with 0·02 mol l−1 ion solution prepared with 0·15 mol l−1 NaCl. SteLL was incubated at 25°C for 45 min in the presence of metal ions before the addition of a rabbit erythrocyte suspension (50 μl). For pH assays, aliquots of SteLL (50 μl) were serially diluted 2-fold in 0·01 mol l−1 citrate phosphate (pH 5·0–6·0), 0·01 mol l−1 sodium phosphate (pH 7·0), 0·01 mol l−1 Tris-HCl (pH 8·0–9·0) or 0·01 mol l−1 glycine–NaOH (pH 10·0–11·0); all these buffers were prepared with 0·15 mol l−1 NaCl. The effect of temperature was evaluated by incubating (30 min) SteLL at 30, 40, 50, 60, 70, 80, 90 and 100°C before conducting the haemagglutinating assay.

Polyacrylamide gel electrophoresis (PAGE)

SteLL was evaluated using PAGE in the presence of sodium dodecyl sulfate (SDS-PAGE) and β-mercaptoethanol on a 15% (w/v) gel according to Laemmli (1970). Polypeptide bands and molecular mass standards (bovine serum albumin, 66 000 Da, ovalbumin, 45 000 Da, glyceraldehyde-3-phosphate dehydrogenase, 36 000 Da, carbonic anhydrase, 29 000 Da, trypsin inhibitor, 20 000 Da, α-lactalbumin, 14 400 Da, from Sigma-Aldrich) were stained with 0·02% (w/v) Coomassie Brilliant Blue prepared with 10% (v/v) acetic acid. Glycoprotein staining was performed using Schiff's reagent (Pharmacia Fine Chemicals 1980).

Antibacterial activity

Gram-positive (Staph. aureus WDCM 00032) and Gram-negative (E. coli WDCM 00013, Klebsiella pneumoniae ATCC 29665, Ps. aeruginosa WDCM 00025 and Salmonella enteritidis MM 6247) strains were provided by the Departamento de Antibióticos, Universidade Federal de Pernambuco, Brazil. The Gram-negative Proteus mirabilis WDCM 00023 was provided by the Fundação Oswaldo Cruz, Brazil. Stationary cultures were maintained in nutrient agar (NA) and stored at 4°C. To determine the antibacterial activity, bacteria were cultured in nutrient broth (NB) and incubated while shaking at 37°C overnight. Cultures were adjusted turbidimetrically to 1·5 × 108 colony forming units (CFU) ml−1 at a wavelength of 600 nm.

Aliquots (100 μl) of leaf extract (1·9 mg ml−1 of protein) or SteLL (0·23 mg ml−1) were diluted 1 : 2 in NB (100 μl) and submitted to a series of 10 double dilutions to a final ratio of 1 : 2048. A 180-μl aliquot of each dilution was dispensed into a microtitre plate well. All wells were inoculated with 20 μl of the bacterial culture and incubated at 37°C for 24 h. Assays were performed in triplicate for each concentration. Negative control wells contained NB medium and the micro-organisms. After incubation, optical density was measured at 490 nm (OD490) using a microplate reader (Biotek Instruments Inc., VT, USA). Minimal inhibitory concentration (MIC) was determined as the lowest protein concentration at which there was ≥50% reduction in optical density relative to the control well OD490 (Amsterdam 1996).

To determine minimal bactericide concentration (MBC), inoculations (10 μl) from wells treated with leaf extract or SteLL that was found to inhibit bacterial growth were transferred to NA plates and incubated at 37°C for 24 h. The lowest protein concentration showing no bacterial growth was recorded as the MBC. The assay was performed in triplicate.

Antifungal activity

Candida albicans was obtained from the Culture Collections at University Recife Mycologia, Departamento de Micologia, Universidade Federal de Pernambuco, Brazil. Antifungal activity was evaluated using the same method used for antibacterial activity, changing the incubation temperature (28°C) and replacing the culture medium used. Sabouraud dextrose was used to determine the MIC, while Sabouraud agar was used to determine the minimal fungicide concentration (MFC) – the lowest protein concentration showing no fungal growth. Assays were performed in triplicate.

Results

From 20 g of S. terebinthifolius leaves, 3·7 g of protein was extracted using 0·15 mol l−1 NaCl. S. terebinthifolius crude extract (specific haemagglutinating activity: 80·3) was treated with activated charcoal to eliminate tannins and other phenol compounds. After a filtration step, the filtrate (leaf extract) showed no colour, but it showed higher specific haemagglutinating activity (2155) than the crude extract. To confirm that activated charcoal treatment was efficient, the total phenol content of the crude extract and leaf extract was determined. Values obtained for the crude extract and leaf extract (after activated charcoal treatment) were 0·019 mg ml−1 and 0·005 mg ml−1, respectively.

Lectin was further purified by chromatography of the leaf extract over a chitin column. Haemagglutinating activity for the leaf extract adsorbed onto the chitin column and one active protein peak (SteLL; specific haemagglutinating activity of 29467) was eluted using 1·0 mol l−1 acetic acid (Fig. 1a). From 20 g of leaf powder, 125 mg of SteLL was isolated with a yield of 1224% and 367-fold purification relative to the crude extract (Table 1). SDS-PAGE revealed that SteLL is a glycosylated polypeptide of 14·0 kDa (Fig. 1b). The same electrophoretic profile was detected after treating lectin with the reducing agent β-mercaptoethanol. Phenolic compounds were not detected in SteLL.

Table 1. Summary of SteLL isolation
SampleProtein (mg ml−1)Haemagglutinating activity (titre)Total protein (mg)Total haemagglutinating activitySpecific haemagglutinating activityPurification (fold)aYield (%)b
  1. a

    Purification fold corresponds to the ratio between specific haemagglutinating activity of the sample and specific haemagglutinating activity of crude extract. Specific haemagglutinating activity was defined as the ratio between the titre and protein concentration (mg ml−1).

  2. b

    Yield corresponds to the percentage of total haemagglutinating activity from crude extract recovered.

Crude extract25·52048370029696080·31100
Leaf extract1·94096260546406215527184
SteLL0·27881921253634790294673671224
Figure 1.

Purification of Schinus terebinthifoilus leaf lectin (SteLL). (a) Chromatography of leaf extract on chitin column. Washing step used 0·15 mol l−1 NaCl. Fractions of 2·0 ml were collected and evaluated for haemagglutinating activity and absorbance at 280 nm in spectrophotometer. (b) SDS-PAGE of SteLL under reducing conditions with β-mercaptoethanol; molecular weight markers (1) and SteLL (2) were stained with Coomassie Brilliant Blue. SDS-PAGE of SteLL stained using Schiff's reagent (3). (image_n/jam12086-gra-0001.png) 0.15 mol l−1 NaCl; (image_n/jam12086-gra-0002.png) 1 mol l−1 NaCl; (image_n/jam12086-gra-0003.png) 1 mol l−1 -acetic acid and (image_n/jam12086-gra-0004.png) log of haemagglutinating activity.

The specific haemagglutinating activity of SteLL (29467) was not altered at pH 5·0, 6·0, 7·0 and 8·0, but was reduced at pH 9·0, 10·0 and 11·0 for 58, 16 and 8, respectively. Heating to 100°C and the addition of Ca2+ and Mg2+ did not alter the haemagglutinating activity.

Antibacterial assays with leaf extract revealed an inhibitory effect on E. coli, Pr. mirabilis and Staph. aureus growth and no effect on Kl. pneumoniae, Ps. aeruginosa and Salm. enteritidis (Table 2). Bactericidal activity was only detected against Staph. aureus (MBC of 950 μg ml−1 of protein). SteLL was active against all tested bacteria, MIC values ranged from 0·45 to 28·75 μg ml−1, and MBC values ranged from 7·18 to 115 μg ml−1 (Table 2).

Table 2. Minimum inhibitory (MIC) and minimum bactericidal concentrations (MBC) of Schinus terebinthifolius leaf extract and lectin (SteLL)
BacteriaLeaf extractSteLL
MICMBCMICMBC
  1. MIC and MBC values are expressed in μg ml−1 of protein. ND: not detected. (+) Gram-positive and (−) Gram-negative bacteria.

Escherichia coli (−)12750ND28·75115
Klebsiella pneumoniae (−)NDND3·59115
Pseudomonas aeruginosa (−)NDND1·7914·37
Proteus mirabilis (−)950ND3·5914·37
Staphylococcus aureus (+)1189501·797·18
Salmonella enteritidis (−)NDND0·45115

The antifungal assay revealed that C. albicans growth was inhibited by the leaf extract (MIC of 12·75 μg ml−1 of protein) and SteLL (MIC of 6·5 μg ml−1). However, only SteLL showed fungicide activity, with an MFC value of 26 μg ml−1.

Discussion

Extraction and isolation of pharmacologically active compounds from medicinal plants have received attention in the search for new economically viable alternatives for treating human infections. Plant lectins have been reported as active components in aqueous extracts that have antimicrobial activity (Oliveira et al. 2008; Sá et al. 2009; Costa et al. 2010).

The crude extract of S. terebinthifolius leaves showed strong pigmentation, indicating the presence of a high concentration of phenol compounds such as tannins. Leaf extract obtained after treatment with activated charcoal showed no colour, a phenol content of 73% lower than that in the crude extract and specific haemagglutinating activity higher than that in the crude extract. Studies have shown that activated charcoal is efficient for adsorbing tannins and other phenol compounds (Mohan and Karthikeyan 1997; Aerts et al. 1999; Mukherjee et al. 2007).

SteLL, a glycosylated protein, was isolated with yield >100% using chromatography on a chitin matrix. The molecular mass of SteLL was smaller than that of glycosylated lectin isolated from the Myracrodruon urundeuva leaf (14·2 kDa), and the purification fold as well as the yield achieved for SteLL was higher than that obtained for this lectin (Napoleão et al. 2011).

The high yield obtained in the SteLL isolation procedure likely reflects the elimination of tannins from the crude extract after activated charcoal treatment and chitin chromatography. Phenol compounds form soluble or insoluble complexes with proteins, thereby interfering with their biological activity (Hagerman 1992). Suzuki and Mori (1989), using a combination of affinity, anion exchange and gel filtration chromatographies, isolated a lectin from the haemolymph of Pinctada fucata martensii with a yield of 200%. The authors suggested that the high recovery ratio was likely due to the elimination of haemagglutinating activity inhibitors.

SteLL haemagglutinating activity was not affected in the presence of divalent cations, and it was stable at broad pH and temperature ranges. Haemagglutinating activity of lectins from Setcreasea purpurea rhizome and from Artocarpus integrifolia fruit was also shown to be stable at pH 6·0–9·0 and until 80°C, respectively (Trindade et al. 2006; Yao et al. 2010). The pH and heat stability of SteLL may be due to its glycosylation. In a previous study, the oligosaccharide moiety of Erythrina corallodendron lectin showed dynamically stable interactions, forming long-range contacts between amino acids, which were important for maintaining the structure of this glycoprotein (Kaushik et al. 2011).

The stability of SteLL at different pH and temperatures is a physicochemical characteristic desirable for its use as an antibiotic. An antimicrobial agent should act at wide pH and temperature ranges as pathogenic bacteria are able to grow at a temperature range between 4 and 60°C and at high or low pH values (Hill et al. 1995). Additionally, fungi can also grow at a wide range of temperatures, and most fungi, including C. albicans, tolerate wide variations in pH (Gostinčar et al. 2011; Vylkova et al. 2011). SteLL was active at temperatures and pH around those found in the human body (37°C; pH 6·5–7·5), indicating its potential use for treating human infectious diseases.

Leaf extracts showed weak inhibitory and bactericidal effects on only E. coli, Pr. mirabilis and Staph. aureus. Ethanolic extracts from S. terebinthifolius leaves were also not effective antibacterial agents against Staph. aureus, showing only a bacteriostatic effect with MIC values <100 mg ml−1 (Martinez et al. 1996; Guerra et al. 2000; Lima et al. 2006). An aqueous leaf extract from this plant was shown to be active against Staph. aureus and B. subtilis through a qualitative agar diffusion assay, with growth inhibition zones of 14 and 17 mm, respectively (Martinez et al. 1996).

SteLL showed higher antibacterial activity than leaf extract against E. coli, Pr. mirabilis and Staph. aureus, and it was active against Kl. pneumoniae, Ps. aeruginosa and Salm. enteritidis, which were not affected by the leaf extract. The increment of antibacterial activity was likely due to the concentration of SteLL, indicating that it is one of the main active components present in the leaf extract. A thermo-resistant lectin isolated from Eugenia uniflora seeds inhibited the growth of Staph. aureus and Ps. Aeruginosa with an MIC (1·5 μg ml−1) similar to those determined for SteLL, but it was less effective in inhibiting (MIC of 16·5 μg ml−1) the growth of B. subtilis and E. coli (Oliveira et al. 2008).

The MBC for bactericidal drugs is generally the same or not more than fourfold higher than the MIC value. In contrast, the MBC of bacteriostatic drugs is many fold higher than their MIC (Levison 2004). Based on the MBC/MIC ratio, SteLL was a bactericide drug against E. coli, Staph. aureus and Pr. mirabilis. Regarding the effects on Kl. pneumoniae, Ps. aeruginosa and Salm. enteritidis, SteLL can be classified as a bacteriostatic agent because MBC values were 32-, 8- and 255-fold greater than MIC values, respectively. Lectins isolated from Bothrops leucurus venom and M. urundeuva heartwood can also be considered as bacteriostatic agents against Staph. aureus, because MBC values of these lectins were 15·8- and 13·9-fold greater than their MIC values, respectively (Sá et al. 2009; Nunes et al. 2011).

SteLL was more efficient in killing the Gram-positive Staph. aureus than the Gram-negative bacteria such as E. coli, Kl. pneumoniae, Ps. aeruginosa and Salm. enteritidis. Similarly, the N-acetyl-d-glucosamine-binding lectin from Araucaria angustifolia seeds was also more active against Gram-positive (Clavibacter michiganensis) than Gram-negative (Xanthomonas axonopodis) bacteria, promoting the formation of pores and severe disruption of the C. michiganensis membrane and bubbling on the X. axonopodis cell wall (Santi-Gadelha et al. 2006). The difference in the susceptibility of Gram-positive and Gram-negative bacteria may be linked to the difficulty of lectins to cross the Gram-negative bacteria outer cell wall to reach the periplasmic space (Nunes et al. 2011). Additionally, the high level of peptidoglycan (which contains GlcNac) in the cell wall of Gram-positive bacteria may provide more interaction sites for chitin-binding lectins.

SteLL was more efficient than leaf extract in inhibiting C. albicans growth than the lectins from Archidendron jiringa seeds (MIC of 56·7 μg ml−1), rhizome of Curcuma longa (MIC of 46 μg ml−1) and the lectin-rich fraction from Hypnea musciformis, which showed only weak fungistatic action (Cordeiro et al. 2006; Petnual et al. 2010; Charungchitrak et al. 2011).

Plant lectins are active against fungi by recognizing and immobilizing the micro-organisms via binding to carbohydrate components, thereby preventing their subsequent growth and multiplication (Cordeiro et al. 2006). The chitin-binding property of SteLL may be involved in its antifungal mechanism. Hevein, a chitin-binding lectin, showed inhibitory activity against C. albicans with a MIC value of 95 μg ml−1 (Kanokwiroon et al. 2008).

In conclusion, SteLL is a glycosylated and chitin-binding lectin that can be purified in milligram quantities in one chromatographic step. SteLL showed strong antimicrobial activity against species that cause human diseases. SteLL haemagglutinating activity was inhibited by N-acetylglucosamine, a component of chitin found in bacterial and fungal cell walls. The characteristics of SteLL, including its ability to interact with N-acetylglucosamine, its pH and heat stability and the insensitiveness of haemagglutinating activity to divalent cations, indicate that studies should be conducted focusing its application as a biomaterial with bactericide and fungicide properties for treating infections as well as for evaluating its toxicity in humans.

Acknowledgements

The authors express their gratitude to the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for research grants and fellowships (LCBBC and PMGP). We are also grateful to the Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco (FACEPE) and the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for financial support. F.S. Gomes would like to thank CAPES for graduate scholarship. We thank Maria Barbosa Reis da Silva for technical assistance and Dra. Rita de Cássia Pereira, from Instituto Agronômico de Pernambuco (IPA), for the identification of the botanical material.

Ancillary