Article pubs.acs.org/jnp
Determination of Gymnemic Acid I as a Protein Biosynthesis Inhibitor Using Chemical Proteomics Angela Capolupo,†,‡ Roberta Esposito,‡ Angela Zampella,§ Carmen Festa,§ Raffaele Riccio,‡ Agostino Casapullo,‡ Alessandra Tosco,*,‡ and Maria Chiara Monti*,‡ †
Ph.D. Program in Drug Discovery and Development and ‡Department of Pharmacy, University of Salerno, 84084, Fisciano, Salerno, Italy § Department of Pharmacy, University of Napoli “Federico II”, 80131, Naples, Italy S Supporting Information *
ABSTRACT: The plant Gymnema sylvestre has been used widely in traditional medicine as a remedy for several diseases, and its leaf extract is known to contain a group of bioactive triterpene saponins belonging to the gymnemic acid class. Gymnemic acid I (1) is one of the main components among this group of secondary metabolites and is endowed with an interesting bioactivity profile. Since there is a lack of information about its specific biological targets, the full interactome of 1 was investigated through a quantitative chemical proteomic approach, based on stable-isotope dimethyl labeling. The ribosome complex was found to be the main partner of compound 1, and a full validation of the proteomics results was achieved by orthogonal approaches. Further biochemical and biological investigations revealed an inhibitory effect of 1 on the ribosome machinery.
M
isotope incorporation at the peptide level, using heavy (deuterated) or light formaldehyde and cyanoborohydride to globally label the peptide N-terminus and the Lys ε-amino groups through an extensive reductive amination (Scheme S1, Supporting Information). MS analysis of the peptide mixtures is based on a comparison of the ion intensities between the heavy and light modified peptides (ΔMW = 4 Da), allowing the relative quantification of each probed target.
any plant extractives are used as food supplements and nutraceutical preparations, attracting millions of people for their perceived health benefits. However, the systematic analysis of the biological and pharmacological properties of each component of these plant extracts is still incomplete. Among the numerous plants used for various diseases, Gymnema sylvestre R. Br. (Asclepiadaceae) is employed in Ayurvedic medicine for the treatment of obesity, diabetes, asthma, inflammation, and metabolic syndrome.1 It has been reported that G. sylvestre extracts are beneficial in reducing blood sugar levels and in weight control.2 In addition, this medicinal plant has antimicrobial, antiproliferative, antihypercholesterolemic, and hepatoprotective activities.3 The leaves and aerial parts of G. sylvestre are known to contain triterpene saponins belonging to the gymnemic acid class.1,4 Gymnemic acid I (1, Figure 1, panel A), first isolated and characterized in 1989 by Yoshikawa and co-workers,5 is one of the principal constituents of the glycosidic fraction of G. sylvestre dried leaves,6 and it is reported to possess an antisweet potential and to affect glucose uptake, increasing insulin levels in the blood plasma.7 Thus, in order to shed light on its biological potential, the interactome of 1 has been investigated through a quantitative chemical proteomic approach, using a dimethyl-isotope labeling strategy. This method gives reliable results on the full target profile of a small molecule, since its interacting proteins are revealed on the basis of their enrichment level when compared with a control sample.8−10 The dimethyl labeling strategy provides a stable© 2017 American Chemical Society and American Society of Pharmacognosy
■
RESULTS AND DISCUSSION As a first step, 1 was immobilized on an azlactone-activated polyacrylamide resin, a rigid polymeric structure with a high surface area and pore volume. In order to minimize steric hindrance, which could impair the interaction between the ligand and its potential targets during the pull-down step,11 a spacer arm (NH2−(PEG)3−NH2) was inserted between the matrix and 1 (Figure 1, panel A). The overall bead modification yield was assessed at 80%, as determined by RP-HPLC (Figure 1, panel B), corresponding to 200 nmol of 1 per mg of resin beads. During the affinity chromatographic step, the 1-bearing beads and control beads were separately incubated with 1 mg of HeLa cell lysates, and the nonspecific protein adsorption was reduced by several washing steps. Small aliquots of the two samples were eluted and analyzed by 1D-SDS-PAGE (Figure S1, Received: August 31, 2016 Published: March 3, 2017 909
DOI: 10.1021/acs.jnatprod.6b00793 J. Nat. Prod. 2017, 80, 909−915
Journal of Natural Products
Article
Figure 1. (A) Chemical structure of 1 and reaction scheme of synthesis and immobilization of 2 on resin beads. (B) HPLC profile of free 1 before and after the coupling reaction.
separated on SDS-PAGE, and the fluorescent signal from the luciferase band was quantified by Ettan DIGE software. A decrease of luciferase translation in a concentration-dependent manner of 1 was assessed (Figure 2B), suggesting a direct inhibition of ribosomal protein synthesis by this natural product. Next, the effect of 1 in a cellular system was analyzed in detail. Since the low lipid solubility of 1, due to its structural features reported, impairs the crossing of biomembranes,16 a cytopermeability experiment was carried out to measure the internalization level of 1 in a HeLa cellular system. After incubation of living HeLa cells with 1 and cell lysis, its internalization amount was assessed by RP-HPLC-MS (with the SIM of a negative ion at m/z 805.75) and by integration of the corresponding peak area (Figure S3 and Table S1, Supporting Information). As a result, a low cell internalization level, of around 4%, was measured. Pathan and co-workers17 developed a formulation mixing 1 with phospholipids, which increased its bioabsorption and improved its pharmacodynamic profile. Also a facile method to enhance its permeability is to transform the carboxylate moiety into an amide (gymnemic acid I ethyl amide 3, Figure S3, Supporting Information). Thus, the cytopermeability experiments were repeated with 3, monitoring its RP-HPLC-MS profile (with the SIM of positive ions at m/z 834.49 [M + H]+ and 856.47 [M + Na]+), resulting in around 15% internalization (Figure S4 and Table S1, Supporting Information). Finally, a cellular assay was performed to measure the capability of 1 and 3 to impair protein synthesis, using L-[35S]methionine and L-[35 S]-cysteine as protein radioactive probes.18,19 HeLa cells were treated with 1, 3, or cycloheximide, after assessing their cell viability (Figure S5, Supporting Information) and then fed with the radioactive amino acids. After 2 h, equal amounts of cell lysates were subjected to a 10% SDS-PAGE analysis (Figure 3B) and exposed to an autoradiographic film. Figure 3 shows a clear concentration-dependent drop of the radioactive protein levels in the sample treated with
Supporting Information), while the remaining fractions were digested on the beads with trypsin. The tryptic peptide mixtures were then treated with the labeling reagents, for which the excess amounts were chosen as reported by Boersema et al.10 (see also Figure S1, Supporting Information). The peptides coming from the proteins that interacted with 1 were mixed with formaldehyde and NaBH3CN (light labeling), while those coming from the control experiment were treated with deuterated formaldehyde and NaBH3CN (heavy labeling) (Scheme S1, Supporting Information). The two mixtures (same volume) were combined to give a sample that was submitted to nano-LC-ESIMS/MS.8−12 The data gathered were then analyzed using MaxQuant software to provide the results shown in Table 1. Three biological replicates were performed, and almost 70 proteins were identified (including those identified with one peptide in each biological replicate) with a heavy/light (H/L) ratio of
