Inhibitory Activities of Guaianolides from the Seeds of Byrsonima Crassifolia against Protein Glycation In Vitro

1Research Laboratory of Natural Products, School of Chemical Engineering and Extractive Industries-IPN, Unidad Profesional Adolfo Lopez Mateos, Zacatenco, D.F. CP 07758, Mexico DF 2Laboratory of Secondary Metabolites of Microorganisms, Department of Biotechnology and Bioingenieria, Cinvestav, Av. 2508 Mexico DF 3Graduates Department of Food. National School of Biological Sciences-IPN Carpio / Plan de Ayala Mexico DF


Introduction
Advanced glycation end-products (AGEs) are the final products of the nonenzymatic reaction between reducing sugars and amino groups in proteins. They are a group of complex and heterogeneous compounds that are known as brown and fluorescent cross-linking substances [1]. Reactive carbonyl compounds are intermediates in the formation of AGEs and advanced lipoxidation end products in tissue proteins in chronic disease as diabetes and its complications [2].
Byrsonima crassifolia fruit is edible and bright yellow when ripened, it has sweet taste and slightly bitter aftertaste. In México, nanche fruit is consumed as liquor, candy and jelly and is wide accepted as nutraceutical too [3]. Since prehispanic times it has been used on ethnobotanical uses include: as wound healing, anti-inflammatory, dysentery infections, and antidiabetic [3]. In the literature there are many researches on the properties and composition of B. crassifolia [4][5][6][7][8][9][10][11][12]. For these reasons, in this study the anti-AGEs activity of guaianolides was tested in vitro that may contribute to prevent diabetes or other pathogenic complications.

Experimental Section
General experimental procedures IR spectra were recorded on a Perkin-Elmer FTRI 1720X. NMR experiments were obtained on a Bruker DRX-400. The NMR data were processed using UXNMR software. HREIMS were measured on a JEOL HX 110 mass spectrometer. For Column chromatography was carried on Silica gel 60 (230-400 mesh, Merck Co. New Jersey (USA) and Sephadex LH-20 from Sigma-Aldrich (St. Louis, USA). Precoated TLC silica gel 60 F254 aluminum sheets were used, solvents used as eluents from Fermont (California, USA).       13

Bovine serum albumin (BSA)-glucose assay
The methodology was based on that of Brownlee et al. [13]. BSA (l0 mg/ml) was incubated at 37ºC with 500 mM of glucose in phosphate buffered-saline (PBS) (5 ml total volume, pH 7.4) and compound containing 0.02% sodium azide. All reagents and samples were sterilized by filtration through 0.2 μm membrane filters. The protein, the sugar and the prospective inhibitor were included in mixture simultaneously. Aminoguanidine (AG) was used as an inhibitor positive control. After 15 days of incubation the fluorescence intensity was measured at an excitation wavelength 370 nm and emission wave-length 440 nm in the test solutions.

BSA-methylglyoxal assay
This assay was based on a published method by Rahbar and Figarola [14], establish the inhibition of protein glycation at an intermediary. BSA and methylglyoxal were dissolved in phosphate buffer (100 mM, pH 7.4) to a concentration of 20 mg/ml and 60 mM, respectively. Isolated were dissolved in the same phosphate buffer. One milliliter of the BSA solution was mixed with 1ml of methylglyoxal solution and 1 of ml compounds. The mixture was incubated at 37ºC with sodium azide (0.2 g/l) as an aseptic agent. Phosphate buffer was used as a blank. AG and phloroglucinol was used as positive control inhibitor. Fluorescence of the samples was measured after seven days of incubation, using an excitation wavelength 340 nm and an emission 420 nm.

Amadorin activity
Amadorin activity was determined using a post-Amadori screening assay [15]. Lysozyme (10 mg/ml) was incubated with 0.5 M ribose in 0.1 M sodium phosphate buffer containing 3 mM sodium azide, pH 7.4 at 37ºC for 24 h. Unbound ribose was removed by dialysis against 4 1 of 0.1M sodium phosphate buffer, pH 7.4 at 4ºC for 48 h with 5-6 changes. Following dialysis, the protein concentration was determined using the Bio-Rad standard protein assay kit based on the Bradford dye-binding procedure [16]. Dialysed ribated lysozyme (10 mg/ml) was reincubated with 10 mg/ml of either 1-8 and AG in 0.1M sodium phosphate buffer containing 3 mM sodium azide, pH 7.4 at 37ºC for 15 days.

Nα-acetyl-glycyl-lysine methyl ester-o-ribose test
Assay was used to determine the ability of terpenes to inhibit the cross-linking of G.K. peptide in the presence of D-ribose [17]. G. K. peptide (80 mg/ml) was incubated with D-ribose (0.8M) and sodium phosphate buffer (0.5M, pH 7.4) under sterile conditions at 37°C for 24 h. Compounds 1-8 were added to a final concentration of 1 mM, however AG was used at 10 and 50 mM. The f1uorescence intensity was measured at an excitation wavelength of 330 nm and an emission wavelength of 415 nm.

Analysis of protein conformation changes by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE)
SDS-PAGE analysis was made according to Laemmli [18]. Protein samples was diluted with an equal volume of sodium dodecyl sulfate (SDS), terpenes and buffer (2% w/v SDS, 30% w/v glycerol, 0.25M Tris hydroxyaminomethane, pH 6.8), and then boiled for 5 min; 10 µL was loaded on a 10% gradient acrylamide slab gel (Multigel 10/20, Daiichi Pure Chemicals, Tokyo) and electrophoresed at a current of 1 mA/ lane for 1 h. After that, the gel was fixed with 60% ethanol followed by Coomassie Brilliant Blue (CBB) R-250 staining.

MALDI linear TOF MS
After BSA-glucose assay with guaianolides were initially reconstituted in trifluoroacetic acid (TFA 0.1%). An aliquot of each sample was combined (1:1) with sinapic acid matrix solution 5 mg/ml in acetonitrile (ACN) at 50%/TFA 0.1% and applied onto well MALDI plates [19]. Instrument was calibrated by adding compounds (12.0 ng; 0.5 µl) to each calibration spot. Mass spectra were obtained with a MALDI-TOF/TOF mass spectrometer (4800 Proteornics Analyzer: Applied Biosystems, Foster City, CA) in the positive ion reflector mode with a mass range between 800 and 7000 Da. Manual interpretation of tandem mass spectra was performed through the Data Explorer software TM ver. 4.4 (Applied Biosystems).

Statistical Analysis
All data are expressed as mean ± SEM. Statistical analysis was performed using commercially software sigmastat 3.5. For multiple comparisons, one way ANOVA was used followed by Tuckey and/ Dunnet's test. p-values less than 0.05 were considered to be statistically significant.
Correlations observed between H-13 and H-6, revealed that they are oriented toward β-face. The coupling constant J 1,5 , J 5,6 and J 6,7 (9.4 Hz) show that 1 has an A.B cis-fused guaianolide skeleton with trans-diaxial disposition of H-6 (β) and H-7 (α). One acetoxy and one benzoyloxy groups were clearly distinguishable in the NMR spectrum. The relative downfield shift of H-4 (δ H 5.23) was indicative that has a benzoyl moiety that was confirmed by HMBC correlations between H-4 and the carbonyl ester signal at δ Comparison of IR and 1 H NMR data of compound 1 and 2 indicated that both have typical signals of the guaianolide sesquiterpene lactone. Terpene 2 have two esters groups identified as angeloyloxy group by typical signals characteristic at δ H 6.06 (H-3', brq, J=7.5 Hz) and at δ H 1.85 (H-5', br, s). The position of the angeloyloxy group was deduced to be at C-8 (δ C 68.18) by correlation between H-8 (6.06) and C-1' (δ C 167.83) in the HMBC spectrum. The configuration of the angelate double bond were assigned as Z on the basis of a NOESY cross-peak from the H-3' to the methyl 5' pairs. The second ester was essentially identical with those of compound 1 at C-10. H-8 appeared at δ H 4.55 so, angeloxy moiety must be at C-8 since the small coupling constant value of J 7,8 3 Hz required an α-orientation of H-8, which was further confirmed by NOE experiments. The position of the angeloxy group was confirmed at C-8 (δ C 68.18) by the correlation between H-8 (δ H 4.55) and C-1'' (δ C 167.83) in the HMBC spectrum. A singlet at δ H 1.30 could be assigned to a methyl tertiary group attached to a quaternary carbon (C-10) further substituted by an oxygen function (acetyl group) these assignments were supported by HMBC H 3 -14 to C-10, C-9 and C-1. The presence of this quaternary carbon at δ C 62.27 exhibited that the methyl and acetyl groups were attached to the same carbon. Moreover, correlations between the quaternary carbons, C-4 (δ C 137.80) and C-10 (δ C 86.43) and their neighbours made clear the total skeleton of guaianolide [21]. Structure 2 was further supported by 1 H-H 1 COSY data while NOESY cross-peaks between H-6 and H-7, between H-8 and H-5 and H-9β and between H-7 and H-9α supported the relative configuration. Consequently, the structure of sesquiterpene lactone was elucidated as byrsonina D.
Compound 3 was related to the molecular formula of C 24 H 30 O 8 by HR-ESIMS. Its NMR data was closely similar to that compound 2, except for a signal corresponding to an acetyl group at C-2. The C-2 carbon experiencing lowfield shift from δ C 22.88 to δ C 76.68 on acetylation. The C-15 protons and the C-6 and C-8 protons all correlations to each other, indicating that they have syn-orientations to each other. As evident from the occurrence of one-proton δ H 5.44 coupled according to COSY to H-1 and H-3, characteristic for an acetyl proton. Acetylation of C-2 modified multiplicity of the signals of H-3 (dq, J=1.1 Hz,) and H-1 (brt, J=9.0 Hz). Therefore, compound 3 was named as byrsonina E.
The molecular formula C 23 H 24 O 6 , of the compound 4 was established by HRESIMS m/z 549.2483). The 1 H and 13 C NMR spectra were very similar to those of compound 1 (byrsonina C), signals unlike for C-8, C-9 and C-10. Angeloyloxy group was replaced by an acetyl group. Its 13 C NMR spectrum revealed the occurrence of a tetra-substituted double bond (δ C 139.92 and 134.81), an exomethylene (δ C 135. 24  Data are mean ± standard deviation of triplicate tests. The fluorescence intensity was measured at ex 370 nm and em 470 nm. The intensity of each blank was subtracted from the intensity of each sample. Concentration of an inhibitor required to inhibit 50% of the control. Calculated from linear regression equation. a p<0.05, b p<0.01, c p< 0.001; vs control aminoguanidine or * p<0.05, ** p<0.01, *** p<0.001; vs control controlphloroglucinol or orpyridoxamine n. Although the NMR spectral data of compound 5 were essentially identical with those for 4 but only two carbonyl groups were detected, suggesting that in 5 the benzoyl group is absent. The structure was assigned as byrsonina G. The 1 H NMR and 13 C NMR data gave evidence that compound 6 was analogue of 1. They only significant differences in the missing of signals at δ H 1.46 (H-8) in 6. The finding of the HMBC were finally confirmed by the result of COSY experiments, in which coupling between the proton at δ H 4.95 (ddd, J=4.1, 10.9, 1.5 Hz, H-8) and H-7 and H-9α were found. This indicated that one acetyl moiety was attached to C-8. Thus, the structure of compound 6 was identified as byrsonina H.
Compound 7 its IR spectrum displayed absorption bands at 1770 and 1710 cm -1 assignable to lactone and unsaturated ketone functions. As deduced from HREIMS and 1 H NMR and 13 C NMR data, 7 had a molecular formula of C 14 H 16 O 3 , which gave very similar chemical shifts with 5, however, analysis of 1 H NMR of 7 showed a quite different chemical shift at H-2 (δ H 2.21, m for 5) this signal was not present in 7, and, instead the compound was another guaianolide with a C-1 at δ C 134.46 and C-2 ketone function at δ C 194.35, but lacking the signals due to the acetyl or angeloyl moiety. One double bond at C-1 and C-10 was confirmed in the HMBC spectrum, with correlations between H-14 and C-1, C-2; H-3 with C-1 (δ C 134.48), C-2 (δ C 194.35; α,β-unsaturated carbonyl) established the cyclopentenone fragment. No characteristic acetyl signals were observed in the NMR spectrum of 7. From these data 7 was named byrsonina I. Compound 8 was obtained as a colourless oil. The molecular formula C 22 H 28 O 7 was determined by HR-ESIMS analysis. The IR spectrum showed absorption bands at 3440 and 1749 cm -1 assignable to hydroxyl and lactone functions respectively. 13 C NMR and 1 H NMR signals at δ C 132. 55 (C-2), δ C 140.87 (C-3), δ H 5.92 (d, J=10.0 Hz, H-2) and δ H 6.02 (d, J=10.0 Hz, H-3) indicated a double bond at C-2 to C-3. 1 H NMR resonances at δ H 1.77 (s), 1.91 (brd, J 7.5), and 6.06 (brq, J 7.5) suggest an angelate moiety. The position of the angeloxy group was confirmed at C-8 (δ C 68.73) by the correlation between H-8 (δ H 4.62) and C-1'' (δ C 167.74) in the HMBC spectrum. Two sharp three-proton singlet at δ H 1.24, (s, H-14) and δ H 1.42, (s, H-4) showed that the methyl groups could be assigned to a methyl group attached to a carbon further substituted by an oxygen function (hydroxyl group) indicating that positions 4 and 10 were fully substituted in accord with correlation of H-14 (1.24) with δ C 77.45 assigned to C-10 in the HMBC spectrum. A cross-peaks of the proton at δ H 2.68 (H-9b) with the hydroxyl substituted carbon at δ C 77.45 (C-10), and C-1 were also observed. A singlet at δ H 1.42 could be assigned to a methyl tertiary group attached to a quaternary carbon (C-4) further substituted by an acetyl group these assignments were supported by HMBC H 3 -15 to C-4, C-3 and C-5. The structure of this guaianolide, designated therein as byrsonina J (all structures are shown in Figure 1b).

Effect of guaianolides in protein glycation
In this study, several methods assay have been proposed to determine the inhibitory effect of 1-8 on AGEs formation, including assays based on the inhibition of advanced glycation end products (AGEs) cross-linking and monitoring the production of fluorescent products which is characteristic of AGEs.
Guaianolides, phloroglucinol and AG showed inhibition against AGEs formation of BSA with glucose, with an IC 50 value range of 0.512 to 1.654 µM /ml (Table 1). Figure 2 exhibit inhibitory activity of compounds 1-8 on BSA glycation at 1 µM. The BSA-glucose assay is a useful tool for determine the effects of guaianolides on the nonenzymatic glycation process. Compounds 1-4 and 8 present inhibitory activity on AGE formation in this model. Methylglyoxal (MG) is a reactive dicarbonyl, formed during glycolysis is a precursor of AGEs formation and triggers of oxidative stress. Methylglyoxal-BSA glycation inhibition was evaluated for 1-8 which show significant activity (p<0.05), compared with phloroglucinol and aminoguanidine (Table 1). However, activity of 1-8 on the formation of AGEs induced by ribose exhibited a significant inhibition (p<0.05) compared to AG. In all experiments, the compound 8 was the most active followed by 3, 2, 1, 4 and 6, 5, 7 as the less actives. Oxidative stress and the formation of AGEs contribute to cellular aging, so increased MG and AGEs cause accelerated cellular aging. MG is double-edged sword due that is a potent inducer of oxidative stress [22], and it is a major precursor of AGEs formation. AGEs also induce oxidative stress.
Rearrangement of the Schiff base to form stable complexes called Amadori product which are considered an important route to AGEs formation that has been implicated in diabetes complications. Terpenes act as inhibitors of this conversion of Amadori intermediates to AGEs. Sesquiterpene lactones 8, 3, 2, and 1 show considerable like activity at the AG, however the compounds 4, 5 and 7 have moderate Amadorin activity (Table 1). Table 1 presents the ability of the compounds 1-8 to inhibit the crosslinking of lysozyme in the presence of ribose.

Glycation of hemoglobin
HbA1c, is an Amadori product of the reaction glucosehemoglobin, is used as an indicator of metabolic control of diabetes being that glycohemoglobin levels increased in diabetics [23]. Glycated hemoglobin (GHb) is formed by non-enzymatic process of reaction of aldehyde group of hexoses with the amino-terminal group of hemoglobin [24]. Table 2 shows the amount of GHb and HbA1c. These results indicates that 1-4 and 8 have a potent glycation inhibition activity      at the early stage of hemoglobin glycation at a concentration, of 10 mM/ ml. Compounds 8, 3, and 2 had the most potent inhibitory effect with values of 37.1%, 32% and 28% respectively compared with the positive control (p<0.05). These compounds were more effective than that of AG (25.1%, 10 mM), indicating that can effectively prevent GHb and HbA 1C formation. Carboxymethyl-lysine (CML) is an advanced glycation end product formed on protein as glyco-oxidation product which is not fluorescent and not reactive [25]. CML is an indicator of the advanced stages of the Maillard reaction and are biomarkers of more extensive underlying glycative and oxidative damage to the protein. As shown in Table 3, compound 8 exhibited a potent anti-cross-linking activity in G.K. peptide-D-ribose model system. Potential sites where the most effective compounds inhibit protein glycation were indicated in Figure  3. These observations suggested that compounds 1-4, and 8 can inhibit the glycoxidative modification of proteins.

SDS-Page profile of guaianolides on protein
SD-PAGE profiles for control, BSA-guaianolides and BSA-AG. BSA-glucose produces cross-linking to give dimers into high molecular weight polymers which are shown in Figure 4. BSA cross-linking by glucose was inhibited by the addition of guaianolides. Polymers of high molecular weight were formed in a very small amount in the presence of AG and guaianolides as compared to the control inhibiting crosslinked advanced glycation endproducts causing a reduction in intensity of the dimerised peaks. This study confirms the capacity of guaianolides to avoid formation of crosslinked advanced glycation end-products in vitro. Due to the above suggest that guaianolides can be mainly attributed to the ability of compounds to react with carbonyls groups. While AG can inhibit through competitive binding to glucose as well as inhibiting of carbonyl formation [26].

Glycation of guaianolides monitored by MALDI linear TOF MS
Glycation of guaianolides was monitored by MALDI linear TOF MS. ( Figure 5) exhibit mass spectra of BSA (a) and shows the typical mass spectra of positive of ion MALDI linear TOF MS with the sinapic acid matrix (b). In Figure 6, we can observe a typical spectrum of glycated BSA where glycation of the protein was performed by incubation with glucose and Byrsonines C and J shows a different mass profile, with a relative abundance of each species. The main band at m/z 67925.163 correspond to native BSA whereas the second peak at m/z 3999.656 corresponds to a diglycated form of BSA. A third minor band at m/z 5397.565 may indicated a protonated monoglycated form. The small band at m/z 4202.943 correspond to the protonated diglycated and triglycated forms. Through MALDI linear TOF MS is established that guaianolides inhibit the glycation process likely through competitive binding to glucose.

Conclusions
The guaianolides isolated from seeds of Byrsonima crassifolia were effective for glycation inhibitory activity using different in vitro glycation assays. The antiglycation effect was due to the ability to react with dicarbonyl intermediates being the major mechanism for protein glycation inhibition. However, our results showed that compounds 1-4 and 8 also react with Amadori adducts blocking their conversion to AGEs. Based in these results, we suggest that the terpenes may prevent or improve the AGE associated chronic conditions as diabetes complications.