Tecoma stans: Alkaloid Profile and Antimicrobial Activity.

AIM: Tecoma stans (L.) Kunth is a promising species in the trumpet creeper family Bignoniaceae. This study aimed at showing the antibacterial and antifungal potentials of T. stans methanolic leaf extract (TSME) correlated to its phytoconstituents.
MATERIALS AND METHODS: The antimicrobial potential of TSME was evaluated using agar diffusion method. The main alkaloids were separated on silica gel column and identified using nuclear magnetic resonance spectral analysis. Molecular docking was performed for the isolated compounds against MurD ligase, penicillin-binding protein, and dihydropteroate synthase enzyme to rationalize the observed antibacterial effect. RESULTS AND DISCUSSION: TSME showed significant antibacterial effect against all tested microorganisms with comparable minimum inhibitory concentration (MIC) to the ampicillin and gentamicin with MIC values ranging between 0.98 and 1.95 µg/mL, in addition to a promising antifungal effect when compared to amphotericin with MIC values 3.9 and 15.63 µg/mL for Aspergillus flavus and Candida albicans, respectively. Several alkaloids were separated, purified, and identified as tecostanine, 4-OH tecomanine, 5-hydroxyskytanthine, and tecomanine, which were previously isolated from T. stans. The docking study showed that the alkaloids bind in a similar fashion to the co-crystallized ligands of the crystal structures of MurD ligase. The binding poses and scores in the case of penicillin-binding protein and dihydropteroate synthase did not match the co-crystallized ligands in their crystal structures. The in silico results suggest an antibacterial mechanism that involves the inhibition of MurD ligase.
CONCLUSION: T. stans alkaloids could represent the basic skeleton for a powerful antimicrobial agent. Copyright:

Introduction

The antimicrobial properties of medicinal plants have been widely studied while searching for new antimicrobial agents combating the acquired resistance to different antibiotics.[1] The family Bignoniaceae (order Lamiales), commonly known as trumpet creeper family, Jacaranda family, Bignonia family, or the Catalpa, is a family of flowering plants that consists of 120 genera with 800 species.[2] This family is known for its iridoid, flavonoid, and steroid contents.[3] Tecoma stans is an ornamental plant showing a variety in its active constituents with reported high phenolic and flavonoid content, besides the presence of monoterpene alkaloids,[4] indoles (such as tryptophan, tryptamine, and skatole), in addition to anthranilic acid in its free form.[5] A number of pyrindane (monoterpene) alkaloids have been identified including tecomanine, 4-noractinidine, an N-nor-methyl skytanthine, boschniakine, 7-hydroxyskytanthine, and 4-hydroxytecomanine, besides S-hydroxy-skytanthine and tecomanine.[6] In addition, tecomine, 5b-hydroxy-skitanthine, and boschniakine alkaloids have been identified in T. stans collected from Egypt and Brazil.[7] Fruits and flowers showed the presence of 5-hydroxy-skytanthine hydrochloride in addition to a phenylethanoid, 2-(3,4-dihydroxyphenyl)ethyl-2-O-[6-deoxy-α-l-mannopyranosyl-4-(3,4-dihydroxyphenyl)-2-propenoate]-β-d-glucopyranoside, 4-O-E-caffeoyl-α-l-rhamnopyranosyl-(1’,3)-α/β-d-glucopyranose, E/Z-acetoside, isoacetoside, rutin, luteolin 7-O-β-d-neohespridoside, luteolin 7-O-β-d-glucopyranoside, luteolin 7-O-β-d-glucuronopyranoside, diosmetin 7-O-β-d-glucuronopyranoside, diosmetin 7-O-β-d-glucopyranoside, and diosmetin 7-O-β-d-glucuronopyranoside methyl ester.[8] Those active constituents are correlated to antioxidant, antiproliferative,[89] antidiabetic,[10] anti-inflammatory,[11] and antibacterial activities.[1213] In this study, we evaluated the antimicrobial activity of T. stans leaf extract, and then the constitutive alkaloids were isolated and characterized. Furthermore, molecular docking was performed against several bacterial proteins.

Materials and Methods

Instruments and reagents

1H-nuclear magnetic resonance (NMR) (400 Hz) and Attached proton test (APT) NMR spectra were recorded in Dimethylsulfoxide (DMSO) on Bruker AMX spectrophotometer (Bruker Optik, Ettlingen, Germany). Chemical shifts were shown in parts per million (ppm) with Tetramethylsilane as the internal standard. J values were given in Hertz (Hz). The assignment of APT spectral data depended on the comparison with reported data for similar compounds. Silica gel (Merck, 70–230 mesh) was used for column chromatography. Precoated silica gel thin layer chromatography (Merck) was run in CH2Cl2:MeOH (9.5:0.5, S1 and 9:1, S2), and spots were developed by Dragendorff’s reagent.

Plant material

Leaves of T. stans (Juss) (Bignoniaceae) were collected from Al-Zohria Garden during February 2015, and were kindly identified by Dr. Therese Labib (consultant of taxonomy at the ministry of agriculture and the former director of El-Orman Botanical Garden, Giza, Egypt).

Extraction procedure

The dried leaves (500g) were exhaustively extracted with methanol where the solvent was evaporated under vacuum leaving a residue (30g). The MeOH residue was treated with 2 M HCl then partitioned with CH2Cl2 (400mL × 5) to remove neutral impurities. The aqueous acidic layer was alkalinized with liquid ammonia (30%) then further extracted with CH2Cl2 (400mL × 5). The combined CH2Cl2 fractions were collected then dried under vacuum to yield crude total alkaloids (2g, 6.7%) as a dark brown sticky mass (TC), whereas drying of the aqueous fraction (TA) yielded 7g (23.3%), giving positive reaction with Dragendorff’s reagent. The crude extract (TC) was chromatographed over silica gel eluting with CH2Cl2 and increasing proportion of CH3OH (from 0% to 100%), where 120 fractions were collected. Fractions 10–23, eluted with CH2Cl2:MeOH (95:5) yielded pure compound 1 (15mg), whereas fractions 30–45, eluted with CH2Cl2:MeOH (90:10) yielded compound 2 (10mg). Similarly, fractions 55–65 gave compound 3 (20mg) on increasing the polarity to CH2Cl2:MeOH (80:20) followed by elution of compound 4 (15mg, fractions 76–80) with CH2Cl2:MeOH (75:25). All compounds were positive to Dragendorff’s reagent.

Antimicrobial testing

The antimicrobial effect of T. stans methanolic extract was evaluated using disc diffusion method against gram-positive bacteria (Bacillus subtilis [RCMB 010067] and Staphylococcus aureus [RCMB 010010]), gram-negative bacteria (Escherichia coli [RCMB 010052], Pseudomonas aeruginosa [RCMB 010043], and Klebsiella pneumoniae [RCMB 0100106]), in addition to fungi (Candida albicans [RCMB 05036] and Aspergillus fumigatus [RCMB 02568]) by measuring the diameters of the inhibition zones in millimeters. T. stans methanolic leaf extract (TSME) was added on a 6-mm filter paper disk, then the plates were incubated at 37°C for 24h for bacteria and at 28°C for 48h for fungi.[14] The minimum inhibitory concentration (MIC) was determined by the broth microdilution method using 96-well microplates.[15] The stock solution was dissolved in DMSO to reach a concentration of 1000 μg/mL, then a serial dilution of the tested extract (100 μL) was prepared at 500, 250, 125, 62.5, 31.25, 15.6, 7.81, 3.9, 1.95, 0.98, 0.49, 0.24, and 0.12 μg/mL. Microbial growth was indicated by the well turbidity after incubation at 37°C for 24h for the antibacterial activity and 48h at 25°C for antifungal activity. The lowest concentration showing no growth was taken as the MIC.

Molecular docking study

Four crystal structures that had been previously exploited in studying the antibacterial effects of alkaloids were applied in this study.[16] The first crystal structure with PDB ID 1UAG is for the UDP-N-acetylmuramoyl-l-alanine (UMA):d-glutamate ligase (MurD), a peptidoglycan ligase involved in cell wall synthesis.[17] The 1UAG crystal has a co-crystallized ligand, UMA, which was exploited to define the active site. Another high-resolution MurD crystal, 2 × 5O, had a co-crystallized inhibitor that occupies the same active site as UMA.[18] The amino acids in the active site in both 1UAG and 2 × 5O nearly overlap with only some subtle differences. The third structure (PDB ID: 3UDI) is for a penicillin-binding protein (PBP) with a co-crystallized open-form penicillin that was used to define the active site.[19] All these three proteins (PDB ID: 1UAG, 2 × 5O, and 3UDI), are important in bacterial cell wall synthesis. The fourth target used in this study is the dihydropteroate synthase (DHPS), which is essential in folic acid synthesis. The DHPS structure with PDB ID 3TYE[20] is co-crystallized with a sulfonamide inhibitor and another endogenous ligand, 2-amino-6-methylidene-6,7-dihydropteridin-4(3H)-one. The two ligands were used to identify two neighboring docking sites in 3TYE and termed site 1 and site 2, respectively. Hence, five docking sites were used to study the binding patterns of four potential ligands. We re-docked the original co-crystallized ligands to serve as a reference for our results. We also docked two of the ligands mentioned in the aforementioned docking study.[16] In all dockings, a grid box of dimensions 50 grid points and spacing of 0.375 was centered on the given ligand. Docking was performed via AutoDock4 (Scripps Research Institute, California, USA) implementing 100 steps of genetic algorithm while keeping all the default settings provided by AutoDock tools.[21] Visualization was carried out using Discovery Studio (Biovia, D.S., 2017, Discovery studio visualizer, San Diego, CA, USA).[22]

Results and Discussion

The results of the antimicrobial activity and the MIC determinations are summarized in Tables 1 and 2. TSME showed significant antimicrobial effect with higher activity against gram-negative bacteria. The impact of this study resides on the inhibitory effect of TSME against resistant microbial species. P. aeruginosa is among the resistant bacteria, which causes a neutrophilic response resulting in significant host damage,[23] whereas S. typhimurium is among the most commonly isolated foodborne pathogens, which result in foodborne outbreaks[24] and lead to substantial morbidity and mortality.[25] In addition, K. pneumoniae and E. coli are drug-resistant Enterobacteriaceae that are responsible for huge public health problem.[262728] A promising effect was also observed against gram-positive bacteria as S. aureus, which represents a major cause of bacterial drug-resistant infections and is linked to high mortality rate,[29] and S. pneumoniae, which is a leading cause of pneumonia, meningitis, and bacteremia worldwide,[30] where TSME showed MIC of 1.95 µg/mL. With increased incidence of invasive fungal infections, evolved the need for powerful antifungal agent. C. albicans infections ranged from superficial to life-threatening and systemic infections.[31] TSME showed MIC of 15.63 µg/mL against C. albicans compared to 0.49 µg/mL for amphotericin.

Table 1

Inhibition zone of Tecoma stans leaf methanolic extract compared with standard antibacterial and antifungal agents

Micro-organisms	Tecoma stans (mm)	Standards (µg/mL)
Fungi		Amphotericin B
Aspergillus fumigatus (RCMB 02568)	20.1 ± 0.63	23.7 ± 0.1
Candida albicans (RCMB 05036)	17.3 ± 0.58	25.4 ± 0.1
Gram-positive bacteria		Ampicillin
Streptococcus pneumoniae (RCMB 010010)	22.4 ± 1.2	23.8 ± 0.2
Bacillus subtilis (RCMB 010067)	21.3 ± 0.58	32.4 ± 0.3
Staphylococcus aureus (RCMB 010028)	20.6 ± 1.2	26.2 ± 0.3
		Vancomycin
Methicillin-resistant Staphylococcus aureus (RCMB 2658)	18.3 ± 0.25	20.3 ± 1.2
Gram-negative bacteria		Gentamicin
Pseudomonas aeruginosa (RCMB 010043)	21.3 ± 0.25	17.3 ± 0.1
Escherichia coli (RCMB 010052)	22.4 ± 1.2	19.9 ± 0.3
Salmonella typhimurium (RCMB 010072)	22.6 ± 0.25	22.3 ± 1.5
Klebsiella pneumoniae	22.1 ± 0.58	23.2 ± 0.58

Table 2

Minimum inhibitory concentration values for Tecoma stans leaf methanolic extract compared with standard antibacterial and antifungal agents

Micro-organisms	Tecoma stans (µg/ml)	Standards (µg/ml)
Fungi		Amphotericin
Aspergillus fumigatus (RCMB 02568)	3.9	0.98
Candida albicans (RCMB 05036)	15.63	0.49
Gram-positive bacteria		Ampicillin
Streptococcus pneumoniae (RCMB 010010)	1.95	0.98
Bacillus subtilis (RCMB 010067)	1.95	0.49
Staphylococcus aureus (RCMB 010028)	1.95	0.49
		Vancomycin
Methicillin-resistant Staphylococcus aureus (RCMB 2658)	7.81	3.89
Gram-negative bacteria		Gentamicin
Pseudomonas aeruginosa (RCMB 010043)	1.95	15.63
Escherichia coli (RCMB 010052)	1.95	3.9
Salmonella typhimurium (RCMB 010072)	0.98	0.98
Klebsiella pneumoniae (RCMB 0100106)	0.98	0.49

The alkaloidal content obtained by acid–base extraction from the total methanolic extract (30g, 6%) of T. stans leaves (500g) yielded two brown sticky masses; TC (2g, 6.7%) and TA (7g, 23.3%). Chromatographic investigation of TC led to the isolation of compounds 1, 2, 3, and 4, giving positive Dragendorff’s reaction and typical 1H and 13C-NMR spectra for tecomanine, 4-OH tecomanine, tecostanine, and 5-hydroxy-skytanthine, respectively, which were previously identified in T. stans.[6732]

To combat antibiotic resistance, scientists follow usually one of the following to be the target for drug discovery, either identifying new microbial proteins for which they direct their effort or designing innovative drugs that can target the existent and known proteins.[33] DHPS, which is a key enzyme that catalyzes microbial folate biosynthesis, is the target of many antibacterial and antifungal classes of drugs.[3435] Other attractive targets in drug discovery are PBPs as they activate the last steps of the biosynthesis of peptidoglycan, which are present outside the cytoplasmic membrane.[36]

Molecular docking represents one of the most effective and powerful tool in drug discovery through a virtual screening of the new bioactive material toward a particular protein.[37] In this study, docking was performed to rationalize the antibacterial activity associated with the isolated alkaloids, as the protein targets contribute to different antibacterial mechanisms. The predicted affinities for the most populated clusters per docking calculation are given in Table 3. To validate the docking approach, each co-crystallized ligand, which included natural substrates or known inhibitors, was isolated and then re-docked to compare the predicted docking pose with the experimentally defined pose. The approach was further validated through re-docking two molecules that were previously docked in a previous study[16] and comparing the scores. The validation is given in the first three rows of Table 3.

Table 3

Docking scores (in kcal/mol) of the alkaloids in the active sites of MurD ligase (PDB ID: 1UAG, 2 × 5O), penicillin-binding protein (PDB ID: 3UDI), and dihydropteroate synthase enzyme (PDB ID: 3TYE)

Source of ligand	Ligand	1UAG	2 × 5O	3TYE	3TYE	3UDI
				Site 1	Site 2
From PDBa	Co-crystallized	−7.50	−7.39	−5.03	−7.15	−9.06
Previously studied molecules[16]a	Indole-3 acetic acid	−5.50 (−5.8)	−4.30 (−5.8)	−3.95 (−5.7)	−4.97 (−5.7)	−5.51 (−6)
	Hypaphorine	−5.32 (−6.1)	−5.63 (−5.9)	−4.2 (−6)	−4.12 (−6)	−5.53 (−6.3)
	5-OH-skytanthine	−5.65	−4.85	−5.13	−4.64	−4.89
	Tecomanine	−5.60	−4.98	−5.26	−4.53	−5.24
	4-OH-tecomanine	−5.14	−5.51	−5.46	−4.70	−5.13
	Tecostanine	−5.17	−4.70	−4.96	−4.62	−4.34

aThe first three rows give the scores for the re-docking of the co-crystallized ligands and two of the previously reported molecules. The previously reported docking scores are given between brackets. For 3TYE, the previously reported scores were duplicated as the study did not distinguish the two regions of the active site

With MurD ligase (PDB ID: 1UAG), 5-OH skytanthine, tecostanine, and 4-OH tecomanine occupied one region of the binding pocket, whereas tecomanine occupied another region. Both regions are occupied by the native co-crystallized ligand, which is much larger in size than the four alkaloids. The alkaloids interact with different amino acids including hydrogen bonds with Gly73, Ser71, and Asp35 for tecomanine and with Gly137 and Asn138 for the other three alkaloids. Other hydrophobic interactions with Leu15, Pro72, Ile95, Pro142, and Ala143 could also be observed. The interaction network for tecomanine is shown in Figure 1A.

Figure 1

Two-dimensional (2D) and three-dimensional (3D) and binding interaction of best scoring compound with different protein. (A) 2D and 3D binding interaction of tecomanine in the active site of MurD ligase (PDB ID: 1UAG). (B) 2D and 3D binding interaction of 4-OH tecomanine in the active site of MurD ligase (PDB ID: 2 × 5O). (C) 2D and 3D binding interaction of tecostanine in the active site of the dihydropteroate synthase enzyme (PDB ID: 3TYE). The docking region is defined by a co-crystallized sulfonamide inhibitor (site 1). (D) 2D and 3D binding interaction of tecostanine in the active site of the dihydropteroate synthase enzyme (PDB ID: 3TYE). The docking region is defined by a co-crystallized endogenous ligand (site 2). (E) 2D and 3D binding interaction of 4-OH tecomanine in the active site of the penicillin-binding protein (PDB ID: 3UDI)

The situation with the other crystal structure of MurD ligase (PDB ID: 2 × 5O) was different from 1UAG as the co-crystallized ligand in 2 × 5O occupies a slightly different region than that of 1UAG. Interestingly, the two co-crystallized ligands meet in one region, similar to that occupied by docking tecomanine in 1UAG. Nevertheless, when the docking was repeated in 2 × 5O, tecomanine and 4-OH tecomanine scored higher in another region [Figure 1B]. On the contrary, 5-OH skytanthine managed to dock in the aforementioned overlap region and to interact with neighboring amino acids such as Asp35 (compared to the interactions of tecomanine in Figure 1A). Tecostanine had the worst pose in a region, which was not occupied by any of the co-crystallized ligands. In the case of 2 × 5O, the most prominent in terms of interactions and score is 4-OH tecomanine, scoring −5.51 kcal/mol. 4-OH tecomanine shares interactions with the co-crystallized ligand, including hydrogen bonds with His138, Lys348, Thr321, and hydrophobic interactions [Figure 1B].

The DHPS protein (PDB ID: 3TYE) has two binding pockets defined by two co-crystallized ligands. For the binding site defined by a sulfonamide inhibitor (site 1, PDB residue code: YTZ), the highest scoring molecule was 4-OH tecomanine (−5.46 kcal/mol). The scores for the four alkaloids are all comparable to that of the co-crystallized sulfonamide inhibitor (−5.03 kcal/mol). Nevertheless, they bind at a close by region that does not overlap with that occupied by the sulfonamide and thus interact with different amino acids. Interactions include hydrogen bonds with Arg254 [Figure 1C].

Docking in the second binding site of 3TYE, defined by a co-crystallized endogenous ligand (site 2, PDB residue code: XHP), did not result in high scores. Again, none of the four ligands occupied a space similar to the endogenous ligand. Hence, the alkaloids did not make any interactions similar to the co-crystallized ligand. However, all of them occupied a very similar space that nearly overlaps with the space occupied by their poses from the docking in site 1. The interactions shared by the ligands were hydrogen bonds or van der Waals interactions with Arg254, Asp101, and Thr67, in addition to hydrophobic interactions with His256 and Ile25. The most representative ligand for these interactions is tecostanine [Figure 1D].

For the PBP (PDB ID: 3UDI), the best scoring ligands were 4-OH tecomanine and tecomanine, making two interactions similar to those made by the covalently bonded co-crystallized ligand. These interactions are hydrogen bonds with Ser487 and Lys437. As an example, the interactions of 4-OH tecomanine are shown in Figure 1E. Because of the weak nature of the hydrogen bond to Lys437 (relatively longer bond), it was not detected automatically in the two-dimensional plot. In the case of 4-OH tecomanine, there is an additional hydrogen bond with Asn489, which is not formed by the co-crystallized ligand.

As the alkaloids under study are fragment-like with relatively low molecular weights, it is expected that they would have relatively lower affinities and multiple binding poses. With 1UAG or 2 × 5O, the docking scores for all alkaloids are slightly less than those of the co-crystallized ligands. In addition, there is no consistency in the binding poses of the four alkaloids across the two slightly varying crystal structures. The docking poses, however, nicely overlap with the co-crystallized ligands. This is not the case when the alkaloids were docked in 3TYE. Judging by the scores alone, one is inclined to choose DHPS (PDB ID: 3TYE) to rationalize the activity of T. stans. On the contrary, the scores for the PBP (PDB ID: 3UDI) docking are significantly lower than the co-crystallized ligand, albeit having overlapping binding poses. We thus tend to exclude DHPS and PBP because of the poor docking poses with the former and poor docking scores with the latter. Given their relatively small sizes and binding patterns, the mechanism of action of the four alkaloids likely involves the inhibition of MurD ligase, possibly rationalizing the antibacterial effect of T. stans extract and presenting a helpful base for the development of antimicrobial agents. However, alkaloids of bigger sizes need to be studied to further confirm or invalidate this hypothesis.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.