Several microorganisms are responsible for great economic losses in worldagriculture. Preventive and treatment methods are applied to avoid contaminationof crops by these microorganisms, however, the use of chemical antimicrobialsdamages health and the environment. Secondary plant metabolites are safe naturalsources of antimicrobials for this application. Fabaceae family has its historydescribed in the literature as a potential source for obtaining antimicrobialbioactive. The objective of this work was to isolate bioactive compounds guidedby antimicrobial assays against bacteria and fungi in vitro.Organic extracts were prepared by eluotropic series of leaves ofAnadenanthera colubrina var. cebil andwere tested against six bacteria and six fungi phytopathogenic. Theantimicrobial assays of minimum inhibitory concentration (MIC) and minimummicrobicidal concentration (MMC) were performed at each purification step thatoccurred through HPLC-DAD, Flash Chromatography and HPLC-preparative analysis,to confirm the isolation of the bioactive. Through bioguided isolation, thecompound p-hydroxybenzoic acid was obtained, which showed activity against thephytobacteria Xanthomonas campestris pv.campestris and Acidovorax citrulli.

Vegetable diseases, caused by several pathogenic microorganisms, are one of the mainproblems faced by world agriculture[1]. Bacteria and fungi are among the main pathogens that cause agreat decrease in productivity and consequently economic losses in thissector[2]. Among themethods applied to control diseases in the field are preventive methods, which havelow efficiency, and treatment methods, such as the application of syntheticantimicrobials[3]. Thesehave caused problems related mainly to damage to human and animal health, inaddition to the accumulation of environmental contamination[4].

Secondary plant metabolites are natural sources of substances with antimicrobialproperties, which can be used as an alternative in substitution tosynthetics[5]. Thisstrategy minimizes the negative impacts associated with being considered safe forhealth and the environment, in addition to being economically advantageous due tothe low cost, they are also capable of reducing the impact of microorganisms onagriculture. In addition, plant extracts reduce the possibility of causing microbialresistance, as they are complex mixtures of metabolites[6].

Several studies have demonstrated how the species of the Fabaceae family havepotential as sources of antimicrobials for a great diversity of pathogens[7-9]. Anadenanthera colubrina var.cebil (Griseb.) Altschul, popularly known as Angico, belongingto this family, it is widely used in folk medicine because it is related to theantimicrobial properties of its leaves[10-11]. This work aimedto conduct a bioguided study on organic extracts of A. colubrinavar. cebil, to obtain compounds with antimicrobial activity againstphytopathogenic bacteria and fungi.

Leaves of A. colubrina were collected in the Catimbau NationalPark (08°34′30,96″ S e 37°14′51,76″ W), in the Northeast of Brazil. Thecollected material was oven dried at 45ºC for 72 h, then ground to obtain a thinpowder, stored in an airtight container and kept at 4ºC until use. One specimenidentified and registered by the Herbarium Dárdano de Andrade-Lima of theInstituto Agronômico de Pernambuco (IPA), under voucher IPA - 80350. The plantmaterial was registered in the Sistema Nacional de Gestão do Patrimônio Genéticoe do Conhecimento Tradicional Associado (SisGen) at number A08E18B.

A hundred grams of the powder of the leaves de A. colubrinasubjected to eluotropic series of organic solvents: cyclohexane (CHX),chloroform (CHL), ethyl acetate (EtOAc) and methanol (MeOH) in Soxhlet,respecting the boiling temperature of each solvent, and each was kept underreflux for 24 h. The extracts obtained were then filtered (Whatman n°1), and thesolvents were entirely removed on a rotary evaporator at 45°C under reducedpressure. The dry extracts were stored at 4°C hermetically sealed until use.

The organic extracts tested against 12 phytopathogens. Six bacteria:Acidovorax citrulli (Acc - strain Ac1.12),Pectobacterium carotovorum subsp.carotovorum (Pcc - strain Pcc31),Ralstonia solanacearum (Rsol - strainCM10R22), Xanthomonas campestris pv.campestris (Xcc - strain Xcc53),Xanthomonas campestris pv. malvacearum(Xcm - strain Xcv137) and Xanthomonascampestris pv. viticola (Xcv -strain Xcm11.2.1) obtained by the Collection of Cultures of thePhytobacteriology Laboratory of the Department of Agronomy of UniversidadeFederal Rural de Pernambuco (UFRPE), Brazil. For antimicrobial tests theisolates were grown in a nutrient medium of yeast dextrose - NYDA (5g.L-1 Yeast extract; 3 g.L-1 meat extract; 5g.L-1 peptone; 10 g.L-1 dextrose; 18 g.L-1de agar) for 24 h at 30°C. Six fungi: Aspergillus flavus(Af - strain 6029), Fusarium moniliforme(Fm - strain URM - 5411), Fusariumoxysporum (Fo - strain URM - 6185), F.solani (Fs - strain URM - 6264), Rhizopussotolonifer (Rs - strain URM - 6525) andVerticillium lecanii (Vl - strain URM -6171), obtained from the mycological collection of Micoteca-URM of theDepartment of Mycology of Universidade Federal de Pernambuco (UFPE), Brazil. Forantimicrobial tests, the isolates were grown in Potato Dextrose Agar - PDA (4g.L-1 potato extract; 15 g.L-1 dextrose; 18g.L-1 agar) for 5 days at 48°C.

To evaluate the antimicrobial activity of A. colubrina leaves,the organic extracts, fractions, and isolated compound were solubilized in anaqueous solution at a concentration of 100mg.ml-1 with 10% dimethylsulfoxide (DMSO) and were sterilized by filtration through a microfilter 0.22 µm(GV-Millipore).

The minimum inhibitory concentration (MIC) was determined by the microdilutionmethod (CLSI, 2011) with modifications. A serial dilution of theextract/fractions was prepared in NYD or BD and 15 μl (Absorbance 600 nm = 0.150± 0.05) of bacteria or fungi suspension was added. The concentration of thesamples ranged from 50 mg.ml-1 to 100 µg.ml-1 for organicextracts, from 6 mg.ml-1 to 22 µg.ml-1 for fractions and500 to 1 ug.ml-1 for purified compound. The samples were incubatedfor 24 h for bacteria and 48 h for fungi, at 30°C for both. As a positivecontrol, chloramphenicol was used for bacteria, and cercobin for tested fungiand sterile water with DMSO (10%) was used as a negative control. All tests wereperformed in triplicate.

To determine the minimum bactericidal or fungicidal concentration (CMB or CMF)after the microplate incubation period, 5 µL of the solution from each well wastransferred to NYDA plates and incubated again for the same period. The completeabsence of growth on the agar surface with the lowest concentration of thesample was defined as the MBC or CMF, respectively for bacteria and fungi.

The active organic extract was fractionated by flash chromatography (BiotageIsolera one, Biotage, Charlotte, NC, EUA). The separation occurred in a 50 gSNAP KP-SIL column (company Biotage, Charlotte, NC, EUA). The mobile phase was agradient of N-hexane: EtOAc, 30:70 (v / v) with 1 column volume (1 CV);N-hexane: EtOAc, 30:70 to 0: 100, (v / v) with 6 CV; MeOH: EtOAc, 0: 100 to20:80 (v / v) with 6 CV; MeOH: EtOAc, 20:80 to 80:20 (v / v) with 3 CV. The flowof the mobile phase had a flow rate of 70 ml.min-1 and scan detectionfrom 200 to 800 nm. The 38 fractions were grouped by software that evaluates thefractions according to the UV absorption spectrum, resulting in 6 fractions.

The active extract and fraction, and the isolated compound were analyzed onHPLC-DAD (1260 infinity LC System-DAD, Agilent OpenLAB CDS EZChrom Editionsoftware, version 04.05 of Agilent Technologies, Santa Clara, CA, USA) equippedwith the Zorbax, SB-C18, 5 µm and 4.6 x 250 mm column and Zorbax SB-C18pre-column of 5 µm and 4.6 x 12.5 mm. For this, samples at a concentration of 5mg.ml-1 were solubilized in methanol and filtered through 0.22 µmpolytetrafluoroethylene (PTFE) filters. The mobile phase was composed of thefollowing solutions: (A) 0.3% acetic acid with Milli-Q water (Millipore) and (B)100% acetonitrile (Merck).

The active extract was analyzed under exploratory chromatographic conditionsusing the linear-gradient method of 95 - 40% (A) between 0-30min, with a flowrate of 2.4 ml.min-1, the initial pressure of 202 Bar and ultravioletdetection (UV) from 196 to 400nm. The chromatographic conditions for analysis ofthe active fraction and the purified compound followed a linear gradient of 92 -65% (A) 0-15 min, with a flow rate of 2.4 ml.min-1, the initialpressure of 202 Bar and detection at 256 nm. In both analyzes, the identifiedpeaks had retention time and UV spectrum compared to the commercial referencestandards: gallic acid, p-coumaric acid, caffeic acid, catechin, trans-ferulicacid, quercetin 3β D-glucoside, chlorogenic acid, quercetin, rutin, and ellagicacid.

To isolate the active compound, the active fraction obtained by the flashchromatography system had its compounds separated by AutoPurification HPLCSystem ™ (model: 2767 Sample Manager, 2545 Binary Gradient Module, SystemFluidics Organizer, System Fluidics Organizer, 2489 UV / Vis Detector, MassLynxSoftware with FractionLynx Application Manager and ACQUITY QDa Detector -Waters). The fraction solubilized in methanol at a concentration of 15mg.ml-1 and filtered through 0.22 µm PTFE filters. The separationand isolation of the compounds occurred through the preparative column XBridgePrep C18 (5 µm and 10 × 100 mm), through the following mobile phase; 0.1% formicacid in Milli-Q water (A) and 100% acetonitrile (B) (Merck) with a lineargradient from 94 to 65% (A) from 0 to 9 min; 9-10 min (65%-0% A), 10-14 min (0%A), 16-20 min (0% -94% A), at room temperature, flow rate 9 ml.min-1and detection at 256 nm.

TABLE 1 shows the resultsobtained about the antimicrobial power of organic extracts from A.colubrina leaves against six phytopathogenic bacteria, according to theextraction solvent used. It is observed that all extracts showed reduced growth inthe tested phytobacteria when compared to the control. However, it is observed thatMIC ≤ 1.56 mg.ml-1 is registered in 66.6% of the phytobacteria in theethyl acetate extract, followed by 33.3% in the chloroform and methanol extracts,and the cyclohexane extract showed no activity for any of the species. Whenevaluating the MBC, it is observed that only the ethyl acetate extract hasbactericidal activity in concentrations ≤ 1.56 mg.ml-1 for A.citrulli and X. campestris pv.campestris.

The antifungal activity of the extracts against phytopathogenic fungi shown inTABLE 2. It observedthat there was a reduction in growth in the tested concentrations when compared tothe control, however, no extract was considered to have relevant inhibitory orantifungal activity due to all results being higher than 6.25 mg.ml-1. Inboth microbiological tests, it was observed that the DMSO used to solubilize theorganic extracts in an aqueous medium, did not affect the bacterial or fungal growthin the negative controls in the concentration used.

From the results obtained in the screening of antimicrobial activity, bioguidedpurification of the bioactive compound was continued. Given the observed, the EtOAcextract was selected to be analyzed on HPLC-DAD (FIGURE 1). This analysis allowed the detection of six mainpeaks (≥ 500 mAU), between the retention time (Rt) 3 and 4 min, peak 1 (λmax 228,260 and 294), between Rt 4 and 5 min, peak 2 (λmax 256), between 7 and 8 min, peak 3(λmax, 265, 354) peak 4 (λmax 256, 352) and peak 5 (λmax 256, 356), and between 10and 11 min peak 6 (λmax 256 and 370). Peak 2 is the major compound of the extract.Peak 6, when compared to retention time (Rt) and UV absorption spectrum, wasidentified as quercetin, in addition to traces of catechin and gallic acid were alsoidentified. Except for peak 6, it was not possible to identify the other peaksobtained in the HPLC-DAD analysis by the reference standards used.

The EtOAc extract was subjected to Flash Chromatography (Biotage™), this semipurification generated six fractions grouped according to the UV absorption spectrum(FIGURE 2). These weretested against the microorganisms Acc and Xccaccording to the activity recorded for the crude EtOAc extract (TABLE 3). Fraction 3 for theXcc bacteria showed activity, as it had a MIC less than 1mg.ml-1, so it was selected for the purification of the activecompound.

Thus, flash fraction 3 was analyzed and was subfractionated by HPLC-Preparative,giving rise to six subfractions that were again evaluated for their antimicrobialactivity (TABLE 4). Subfraction3 showed activity with MIC and MBC of 0.5 mg.ml-1 forXcc. In view of its antibacterial potential, subfraction 3 wasanalyzed by HPLC-DAD, and a pure compound was detected. The UV spectrum associatedwith data available in the literature indicates that this compound isp-hydroxybenzoic acid (FIGURE 3),tracked through its UV absorption spectrum using flash chromatography (Biotage™) andHPLC-Preparative (Autopurification System™).

Given the importance of developing a biopesticide, several studies have been carriedout, in search of this objective, thus, a screening carried out by Silva etal.[11] with severalplants of medicinal importance against several phytopathogens pointed to A.colubrina as an important source of antimicrobial compounds. From thisresult, a bioguided purification approach was carried out, starting with anextraction following the eluotropic order of solvents and using an invitro anti-phytopathogenic bioassay. According to Santos etal.[12] to fightbacteria, an extract with MIC > 2.0 mg.ml-1 is considered inactive,therefore, the active and promising extract for the isolation of bioactive compoundwas the one with the lowest MIC and MBC values. According to this criterion, theEtOAc extract was the most active, with MIC and MBC ≤ 1.56 mg.ml-1against the bacteria A. citrulli and X. campestrispv. campestris.

HPLC-DAD analysis of the crude EtOAc extract revealed a phenolic acid as the major(peak 2), with λ max 256 nm, it's ultraviolet (UV) absorption spectrum being verysimilar to p-hydroxybenzoic acid[13-17]. This acid waspreviously isolated and identified in Anadenanthera colubrina inthe works of Gutierrez-Lugo et al.[18] and Weber et al.[19]. It was also possible to verifythe presence of four flavonoids (FIGURE1) with characteristic UV spectra. The UV spectrum of flavonoidsshows two main peaks in the region of 240-400 nm, these two peaks are referred to asthe band I of the molecule (usually 300-380nm) and band II (usually 240-280 nm).Thus the peaks 3 (λ max, 265, 354), 4 (λ max 256, 352) and 5 (λ max 256, 356) arereferred to as quercetin derivatives and maybe quercetin 3,7-O-diglucoside (λ max256, 355), quercetin 3-O-glucoside 7-O-rhamnoside (λ max 257,358), quercetin3-methyl ether (λ max 257,358), quercetin 3-O-glucoside 7-O-rutinoside (λ max257,358) among others[20]. Peak 6λ max 256 and 370 identified through the standard, as already mentioned, isquercetin.

The purification of the bioguided EtOAc extract by the assay against phytopathogensin vitro resulted in the isolation (FIGURE 2) of the major component (peak 2 λ max 256)which when tested separately presented MIC and MBC of 0.5 mg.ml-1X. campestris pv. campestris. The differencebetween the MIC and MBC values of the extract and the isolated substance may be dueto the presence of nutritional components, common in extracts, such as proteins andsugars, which can contribute to the development of the microorganism[21]. Our data are in agreement withthe findings of Araújo et al.[22] which isolated and identified by nuclear magneticresonance (NMR) p-hydroxybenzoic acid the major compound of the ethyl acetateextract of aerial parts of A. colubrina.

Phenolic acids are part of the group of phenolic compounds, rarely occur as freeacids, are divided into benzoic, cinnamic acids and their derivatives,p-hydroxybenzoic acid is the simplest form found in nature[23]. Despite data on theantimicrobial effects of phenolic acids[24-25], studies dealingwith the anti-phytopathogenic properties of its metabolites or derivatives are stillscarce. Literature data indicate that p-hydroxybenzoic acid inhibits the growth ofplant pathogens. Cho et al.[26] found that the acid inhibits the growth of X.campestris with an IC50 of 0.136 mg.ml-1. Howeverresearch indicates that the phytopathogen Xcc may have developed afunctional degradation pathway of 4-HBA (4-hydroxybenzoate) that plays a role indetoxifying phenolic metabolites in the host during infection, however, themechanistic details and the biological significance of this phenomenon have yet tobe elucidated[27].

In fungi assays, all extracts were ≥ 6.25 mg.ml-1, however, the CHXextract was the most active with MIC and MBC of 6.25 mg.ml-1 forAspergillus flavus, Rhizopus sotolonifer, andVerticillium lecanii, the rest of the extracts presented MICand MFC ≥ 12.5 mg.ml-1. Campos et al.[28] perform a bioguided assay withA. colubrina against fungi and identify or extract hexane asthe most active, assign an antimicrobial activity to three substances, among them:β-sitosterol and β-sitosterol linoleate, both in fruits and leaves and with MICs of0.25 and 0.5 mg.ml-1 in front of Alternaria alternatarespectively. Due to the similarity with ergosterol, steroidal substances cancompete with fungal proteins involved in the synthesis of this metabolite and can belethal to the fungus[28].

The bioguided approach carried out with the extracts of the leaves of A.colubrina through in vitro antimicrobial tests led top-hydroxybenzoic acid which, in addition to being the major component of the activeextract, showed antimicrobial activity against the phytopathogensXanthomonas campestris pv. campestris with MICand MBC of 0.5 mg.ml-1. The data highlight the potential of the speciesas an alternative source of this compound to fight diseases of economic importancein agriculture.