First report of fatty acids in Mimosadiplotricha bee pollen with in vitro lipase inhibitory activity

Chemicals and reagents

The chemicals (α-amylase, acarbose, AChE from electric eel, acetylthiocholine iodide, 5,5′-dithiobis (2-nitrobenzoic acid), physostigmine, crude porcine pancreatic lipase type II, p-nitrophenyl palmitate, and orlistate) used in this study were from Sigma-Aldrich, Darmstadt, Germany.

Sample collection

Six A. mellifera BPs (BP1 to BP6), one from each of six localities in Thailand, were collected in 2018, that were suspected, based on the available flowers near the hives, of being monofloral and derived from Camellia sinensis L. and Mimosa diplotricha in Chiangmai province (BP 1 and BP3, respectively), Helianthus annuus L. in Lopburi province (BP2), Nelumbo nucifera in Nakhon Sawan province (BP4), Xyris complanata in Udon Thani province (BP5), and Ageratum conyzoides in Lamphun province (BP6), repectively. The BP samples used were dried using a specific process and stored at room temperature (25 °C) until used.

Identification of the bee pollen

Crude extraction and partition

The extraction and partition was modified from Chantarudee et al. (2012). Each BP (140 g) was mixed with 800 mL of methanol (MeOH), shaken at 100 rpm, 15 °C for 18 h, and then centrifuged at 6,000 rpm, 4 °C, for 15 min. The supernatant was collected, while the solid residue (pellet) was re-extracted three more times in the same manner with 800 mL of MeOH each time. The supernatants were pooled and evaporated under reduced pressure at a maximum temperature of 40–45 °C to obtain the crude MeOH extracts, which were kept at −20 °C in the dark until used.

The MeOH crude extracts were sequentially partitioned by hexane (low polarity), to eliminate the lipid and non-polar compounds; DCM (medium polarity), and finally MeOH (high polarity). To this end, the six MeOH crude extracts (one for each of BP1–6) were separately dissolved in MeOH until it was not sticky and then mixed with an equal volume of hexane in a separating funnel and left to phase separate, whereupon the upper hexane phase was collected. The lower MeOH phase was then further extracted with hexane in the same manner twice more, and the hexane extracts were pooled and evaporated under reduced pressure at a maximum temperature of 40–45 °C to yield the hexane partitioned (HX) extracts of BP1–6 (bee pollen from C. sinensis L., H. annuus L., M. diplotricha, N. nucifera, X. complanata, and A. conyzoides, respectively, and designated as HXCBP, HXHBP, HXMBP, HXNBP, HXXBP, and HXABP, respectively). Meanwhile, the residual MeOH phase was then extracted with an equal volume of DCM three times in the same manner as above (except the DCM phase was the lower layer), with the pooled DCM extracts evaporated as above. The sample from this step was named the DCM partitioned extracts of C. sinensis L., H. annuus L., M. diplotricha, N. nucifera, X. complanata and A. conyzoides flower BP (BP1–6, respectively), and designated as DCMCBP, DCMHBP, DCMMBP, DCMNBP, DCMXBP, and DCMABP, respectively.

Finally, the residual MeOH phase was evaporated as above to yield the MeOH-partitioned (MT) extract of C. sinensis L., H. annuus L., M. diplotricha, N. nucifera, X. complanata and A. conyzoides flower BPs (BP1–6, respectively, and designated as MTCBP, MTHBP, MTMBP, MTNBP, MTXBP, and MTABP, respectively). All partitioned extracts were kept at −20 °C in the dark until used to test the biological activities.

In vitroα-amylase inhibitory activity

The α-amylase inhibition assay was modified from Akoro, Ogundare & Oladipupo (2017). The partitioned extract of BP was dissolved in dimethyl sulfoxide (DMSO) and subsequently diluted in MeOH at different concentrations (125, 250, 500, 1,000, and 2,000 µg/mL). Two hundred and fifty µL of α-amylase solution [0.5 units (U)/mL] dissolved in buffer [Na₂HPO₄/NaH₂PO₄ (0.02 M) and NaCl (0.006 M)] at pH 6.9 was mixed with 250 µL of the extract and incubated at 37 °C for 10 min. After that, 250 µL of the starch solution [0.5% (w/v) in d-H₂O] was added and incubated at 37 °C for 10 min. The reaction was terminated by the addition of 500 µL DNSA reagent (12 g of sodium potassium tartrate tetrahydrate in 8.0 mL of 2 M NaOH, and 20 mL of 96 mM of 3,5-dinitrosalicylic acid solution) and was heated at 85–90 °C for 5 min in a water bath. The mixture was cooled to room temperature and was diluted with 5 mL of d-H₂O, and the absorbance was measured at 540 nm (A₅₄₀) using a UV-Visible spectrophotometer (Sunrise, Tecan, Austria). Acarbose was used as the positive inhibitor. Each sample was performed and measured in triplicate. The inhibitory percentage of α-amylase was calculated using the equation given below.

Percentage of α-amylase inhibition = [{(A-B) –(C-D)} / (A-B)] ×100

where A is the A₅₄₀ after incubation without an extract, B is the A₅₄₀ after incubation without an extract and α-amylase, C is the A₅₄₀ after incubation with an extract and α-amylase, and D is the A₅₄₀ after incubation with an extract, but without α-amylase.

The % α-amylase inhibition (Y axis) was plotted against the extract concentrations (X axis) and the IC₅₀ value was obtained using regression analysis.

In vitro AChEI inhibitory activity

Evaluation of the AChEI activity was modified from Li et al. (2019) based on Ellman’s method. Firstly, 160 µL of TTB [50 mM Tris–HCl buffer pH 8 with 1% (v/v) Trition X-100], 20 µL of the extract dissolved in DMSO (500 µg/mL), and 10 µL of 0.2 U/mL AChE from electric eel dissolved in 0.1% (w/v) bovine serum albumen in TTB were mixed and incubated at 4 °C for 20 min. Then, 5 µL of 15 mM acetylthiocholine iodide in d-H₂O and 5 µL of 2 mM 5,5′-dithiobis (2-nitrobenzoic acid) (DTNB) in TTB containing 0.1 M NaCl and 0.02 M MgCl₂ were added per well and incubated at 37 °C for 20 min. The absorbance at a wavelength of 412 nm (A₄₁₂) was measured using a microplate reader (Sunrise, Tecan, Austria). Physostigmine was used as the positive control. All the reactions were performed in triplicate. The percentage inhibition was calculated as follows.

Percentage of AChE inhibition = [{(A-B) –(C-D)} / (A-B)] ×100

where A is the A₄₁₂ after incubation without an extract, B is the A₄₁₂ after incubation without an extract and AChE, C is the A₄₁₂ after incubation with an extract and AChE, and D is the A₄₁₂ after incubation with an extract, but without AChE.

The % AChEI (Y axis) was plotted against the extract concentrations and the IC₅₀ values were obtained using linear or non-linear regression analysis.

In vitro porcine pancreatic lipase inhibitory (PPLI) activity

The enzyme solution was prepared immediately before use as previously described (Jamous et al., 2018) with some modifications. Crude PPL type II was suspended in 50 mM Tris–HCl buffer pH 8 to a concentration of 2 mg/mL. The suspension was mixed and centrifuged at 16,000x g for 10 min. The clear supernatant was recovered and kept. The PPLI assay was adapted from Maqsood et al. (2017). Briefly, 100 µL of the extract at different concentrations (200, 400, 600, 800, and 1,000 µg/mL for DCMCBP, DCMMBP, DCMNBP, DCMXBP, and DCMABP; and 12.5, 25, 50, 100, and 200 µg/mL for DCMMBP2 and DCMMBP2-1) dissolved in DMSO and 600 µL of 50 mM Tris–HCl buffer (pH 8.0) were pre-incubated with 200 µL of 2 mg/mL of PPL solution at 37 °C for 30 min. Afterwards, 100 µL of 1.5 mM of p-nitrophenyl palmitate (p-NPP) in isopropanol was added and incubated at 37 °C for 2 h. Lipase activity was determined by measuring the hydrolysis of p-NPP to p-nitrophenol product via measuring the absorbance at 410 nm (A₄₁₀) using a microplate reader. Orlistat was used as the positive standard. Each sample was performed and measured in triplicate. The percentage of lipase inhibition was calculated according the following formula.

Percentage of lipase inhibition = [{(A-B) –(C-D)} / (A-B)] ×100

where A is the A₄₁₀ after incubation without an extract, B is the A₄₁₀ after incubation without an extract and lipase, C is the A₄₁₀ after incubation with an extract and lipase, and D is the A₄₁₀ after incubation with an extract, but without lipase.

The % lipase inhibition (Y axis) was plotted against the extract concentrations (X axis) and the IC₅₀ values were obtained from the graph.

Enrichment of active fractions

Among the mentioned bioactivities, the extract with the most potent in vitro PPLI activity was used for further fractionation (enrichment) by silica gel 60 column chromatography (SiG60-CC).

Chemical structure analysis by NMR

Among the fractions obtained from the SiG60-CC (250-mL size), the most active fraction for PPLI activity was evaporated and analysed. Briefly, the evaporated sample was dissolved in an appropriate deuterated solvent (Chloroform-d or MeOH-d4, Merck) at a ratio of 5–20 mg of compound to 600 µL of deuterated solvent. It was then transferred to an NMR tube and shaken until completely dissolved. The NMR spectrum was recorded on a Jeol JNM-ECZ (JNM-ECZ500R, Tokyo, Japan) 500MHz operated at 500 MHz for ¹H-NMR nuclei in order to detect the functional groups using tetramethylsilane as the internal standard. The chemical shift in δ (ppm) was assigned with reference to the signal from the residual protons in the deuterated solvents, while the chemical shift and J coupling value were determined using the MestReNova version 12.0.3 software.

Preparation of FAMEs

A portion (26.7 mg) of the fraction with the highest PPLI activity obtained from SiG60-CC (250-mL size), was added to absolute MeOH (two mL), followed by 0.5 mL of concentrated H₂SO₄. The reaction mixture was stirred and heated at 60 °C for 3.5 h. The reaction mixture was then evaporated to dryness, diluted with DCM (six mL) and extracted with saturated NaHCO₃. The DCM layer was collected and washed several times with d-H₂O until pH of the solution was 7. The combined DCM layer was dried over anhydrous Na₂SO₄ and evaporated to dryness. Prior to analysis of the prepared FAMEs by GC-FID analysis (6890N GC, California, USA), the formation of FAMEs was confirmed by ¹H-NMR analysis (JNM-ECZ500R, Tokyo, Japan), where the singlet signal of the methoxy group at δ_H 3.65 ppm was observed (Supplement 3).

GC-FID analysis

The prepared FAMEs of the fraction with the highest PPLI activity from the SiG60-CC (250-mL size) fractionation was submitted to the Food Research and Laboratory (Faculty of Science, Chulalongkorn University) for analysis of its fatty acid components by GC-FID following the AOAC method 996.06. Briefly, chromatographic analysis was performed using a GC-FID system (Aglilent 6890 N) equipped with an autosampler and split-splitless injector. A SPTM 2560 FULED SILICA capillary column with an internal diameter of 0.25 mm and 0.2 µm film thickness was used for the chromatographic separation. Helium was used as the carrier gas at 1.1 mL/min. The injector and detector temperatures were set at 260 and 250 °C, respectively. The initial GC oven temperature was 140 °C, held for 5 min, increased to 240 °C at 20 °C/min, and held at 250 °C for 0.5 min. A volume of 1.0 µL of sample was injected using the split injection mode (100:1). The peaks were identified by comparison of their relative retention times with a standard FAME mixture. The results were expressed as mg/g fatty acid.

Data analysis

Experiments were performed in triplicate. Numerical data are reported as the mean ± one standard deviation (± SD), determined in the Microsoft Excel 2019 software. One-way ANOVA and T-test were used to test for significant differences in IC₅₀ values. Tukey’s test (p < 0.05) was applied for the pairwise multiple comparisons. The statistical analyses were performed using IBM SPSS statistics version 22 for windows.

The overall procedure of BP screening and enrichment for the PPLI active component in the most active extract is summarized schematically in Fig. 1.

Palynological analysis of the six BP samples

The major floral pollen types present in BP must be identified in order to report the type of floral pollen present in the BP. The simplest way is by morphology using LM and SEM analyses. Under LM and SEM, the morphology of bee pollen was observed. The pollen grains in BP1 were convex triangular with tricolporate, and the exine ornamentation was verrucate (Figs. 2A and 3A). This is consistent with the pollen grains of C. sinensis (L.) Kuntze (Ariyarathna et al., 2011; Fan et al., 2019). The pollen grains of BP2 were close to spherical with triaperturate, and the exine ornamentation was echinate (Figs. 2B and 3B), supported it to be the pollen grains of H. annuus L. (Ali et al., 2021). The pollen grains of BP3 were spherical or prolate spheroidal in shape with four (tetrad) pollen subunits, and the exine ornamentation was tuberculate (Figs. 2C and 3C), supporting it to be the pollen grains of M. diplotricha (Lima, Silva & Santos, 2008; Peukpiboon, Benbow & Suwannapong, 2017). The pollen grains of BP4 were spherical in shape with tricolpate, and the exine ornamentation was uniformly dense reticulate (Figs. 2D and 3D), supporting that they were the pollen grains of N. nucifera (Sangsuk, Baslev & Jampeetong, 2021; Zhang et al., 2019). The pollen grains of BP5 were ellipsoidal in polar view, flattened/convex in equatorial view with monosulcate and operculate, and the exine ornamentation was reticulate (Figs. 2E and 3E), they consitent with the pollen grains of X. complanata (Da Luz et al., 2015). Finally, the pollen grains of BP6 were spherical in shape with tricolporate, and the exine ornamentation was echinate (Figs. 2F and 3F) confirming they were the pollen grains of A. conyzoides (Zafar, Ahmad & Khan, 2007; Garg, 1996).

Identification of the pollen species in each BP sample by molecular analysis

The six BP samples (BP1–6) were also identified by sequence analysis of the ITS-2 region of the rRNA genes. In each case, PCR amplification of the ITS-2 region gave an amplicon of the expected size (500 bp). After sequencing the amplicons and using them as the query sequence to BLASTn search the GenBank reference sequences, the same plant pollen species identifications were obtained as by the LM and SEM morphological analyses. Sample BP1 showed 100% nucleotide identity to the sequence of C. sinensis L. (accession # MN242039.1), BP2 at 95.33% nucleotide identity to the sequence of H. annuus L. (accession # KF767534.1), BP3–6 at 100% nucleotide identity to the sequence of M. diplotricha (accession # MH768249.1), N. nucifera (accession # FJ599761.1), X. complanata (accession # MW113223), and A. conyzoides (accession # KY700213.1), respectively.

Thus, the six BPs were confirmed to be essentially monofloral, matching the principal flowers around the hives, and with the morphological and molecular analyses congruent with each other. Thus, floral origin of each BP was clarified in this work.

The partitioned extracts of BP

For the six BPs, they were separately sequentially partitioned by MeOH, DCM, and hexane, three organic solvents with different polarities. The yield and character of all 18 obtained extracts (three solvents for each of BP1-6) are summarized in Table 1. The highest yield was obtained from the MeOH-partitioned extracts in all six samples (above 40%). Only a sticky solid form was obtained for the DCM-partitioned extracts, while an oil form was obtained in both the MeOH- and hexane-partitioned extracts.

However, only the MeOH and DCM partitioned extracts of all six types of BP were tested for the three enzyme inhibitory activities because the hexane-partitioned extracts were insoluble in each of the respective enzyme assay buffer solutions.

In vitroα-amylase inhibitory activity

The partitioned extracts were initially used at a final concentration of 2 mg/mL, and the α-amylase inhibitory activity (%) is presented as the mean ± SD in Table 2. At this concentration, DCMXBP provided the highest in vitro α-amylase inhibitory activity (54.8 ± 2.8%). The subsequent dose response assay revealed the anti α-amylase activity of DCMXBP was concentration dependent (Fig. 4A, Supplement 1) with an IC₅₀ value of 1,792.5 ± 51.0 µg/mL (Table 2), which was markedly less effective than that of acarbose, the positive control (Fig. 5A, Supplement 2) with a 63-fold lower IC₅₀ value of 28.1 ± 2.7 µg/mL. In contrast, DCMMBP and MTXBP had no real anti- α-amylase activity at this concentration of 2 mg/mL (1.19 ± 2.06% and 0.00 ± 0.00%, respectively).

In vitro AChEI activity

The partitioned extracts were initially screened for AChEI activity at a final concentration of 500 µg/mL. The in vitro AChEI activity (%) is presented as the mean ± SD in Table 2. At this concentration, DCMCBP provided the highest AChEI activity at 19.3 ± 1.5%. However, the AChEI activity of all the partitioned extracts were negligible compared to that of physostigmine, the positive control, with an over 230-fold lower IC₅₀ value of 0.082 ± 0.002 µg/mL (Fig. 5B and Table 2, Supplement 2).

In vitro PPLI activity

The partitioned extracts were initially screened for PPLI activity at a final concentration of 400 µg/mL with the results (as the % PPLI activity) presented as the mean ± SD in Table 2. This concentration was selected because it was the highest concentration that could be totally dissolved in DMSO. At this concentration, some DCM partitioned extracts showed a PPLI activity close to 50%, and so they were further evaluated at different concentrations. The PPLI activity was found to be dose-dependent (Fig. 4B, Supplement 1) and broadly similar between DCMCBP and DCMMBP, with IC₅₀ values of 458.5 ± 13.4 and 500.8 ± 24.8 µg/mL, respectively. However, they were markedly less effective than orlistat, the positive control, with an over 21,800-fold lower IC₅₀ value of 0.021 ± 0.000 µg/mL (Fig. 5C, Supplement 2).

In vitro PPLI activity of active compounds from DCMMBP

Principal chemical composition analysis (TLC and NMR) of fractions DCMMBP2-1 and DCMMBP2-2

After SiG60-CC, the chemical composition of the three obtained fractions was tested by TLC (Fig. 6). For DCMMBP2-1, no band was observed under UV light at 254 nm (Fig. 6A), but a dark blue spot was found after dipping in 3% (v/v) anisaldehyde in MeOH and heating over a hot plate (Fig. 6B). The structure of fraction DCMMBP 2-1 was first analyzed by ¹H-NMR, where the signal at δ 0.85 to 2.85 ppm and the signal at δ 5.35 ppm showed the characteristics of unsaturated free fatty acids (UFFAs; Supplement 4). Therefore, fraction DCMMBP2-1 was prepared as FAMEs for the GC-FID analysis.

In contrast, only a sharp band was observed on the TLC plate for fraction DCMMBP2-2, which indicated it was enriched to apparent homogeneity (potentially pure) compound. Therefore, the chemical structure of the compound, named compound 1, in fraction DCMMBP2-2 was analyzed by ¹H-NMR. Compared to naringenin isolated from the M. pigra L. flower BP (Khongkarat et al., 2021), the obtained NMR peaks in the chemical shift pattern were ¹H-NMR (500 MHz, MeOH-d4) δ: 7.30 (d, J = 8.5 Hz, 2H), 6.80 (d, 8.5 Hz, 2H), 5.86 (q, J = 2.2 Hz, 2H), 5.33 (dd, J = 13.0, 3.0 Hz, 1H), 3.09 (dd, J = 17.1, 13.0 Hz, 1H), and 2.67 (dd, J = 17.1, 3.0 Hz, 1H) (Supplement 5). Thus compound 1 was identified as naringenin (Fig. 7A). However, the structure of DCMMBP2-3 was not identified from the NMR results because it did not have any marked PPLI activity.