Bionmr-Based Identification of Natural Anti-Aβ Compounds In

Bioorganic Chemistry 83 (2019) 76–86

Contents lists available at ScienceDirect

Bioorganic Chemistry

journal homepage: www.elsevier.com/locate/bioorg

bioNMR-based identification of natural anti-Aβ compounds in Peucedanum T ostruthium ⁎ Alessandro Palmiolia,b, , Sara Bertuzzia, Ada De Luigic, Laura Colomboc, Barbara La Ferlaa, ⁎ Mario Salmonac, Ivano De Nonid, Cristina Airoldia,b, a Department of Biotechnology and Biosciences, University of Milano – Bicocca, P.zza della Scienza 2, 20126 Milan, Italy b NeuroMI, Milan Center for Neuroscience, University of Milano – Bicocca, 20126 Milano, Italy c Department of Molecular Biochemistry and Pharmacology – Istituto di Ricerche Farmacologiche “Mario Negri” IRCCS, Via G. La Masa 19, 20156 Milano, Italy d Department of Food, Environmental and Nutritional Sciences, University of Milano, Via Celoria 2, 20133 Milano, Italy

ARTICLE INFO ABSTRACT

Keywords: The growing interest in medicinal plants for the identification of new bioactive compounds and the formulation Alzheimer’s disease of new nutraceuticals and drugs prompted us to develop a powerful experimental approach allowing the detailed Anti-amyloidogenic compounds metabolic profiling of complex plant extracts, the identification of ligands of macromolecular targets ofbio- Chlorogenic acids medical relevance and a preliminary characterization of their biological activity. To this end, we selected Furanocoumarins Peucedanum ostruthium, a plant traditionally employed in Austria and Italy for its several potential therapeutic NMR spectroscopy applications, as case study. We combined the use of NMR and UPLC-HR-MS for the identification of the meta- Peucedanum ostruthium UPLC-HR-MS bolites present in its leaves and rhizome extracts. Due to the significant content of polyphenols, particularly chlorogenic acids, recently identified as anti-amyloidogenic compounds, polyphenols-enriched fractions were prepared and tested for their ability to prevent Aβ1-42 peptide aggregation and neurotoxicity in a neuronal human cell line. STD-NMR experiments allowed the detailed identification of Aβ oligomers’ ligands responsible for the anti-amyloidogenic activity. These data provide experimental protocols and structural information sui- table for the development of innovative molecular tools for prevention, therapy and diagnosis of Alzheimer’s disease.

1. Introduction and respiratory diseases [1]. The alcohol extract of the rhizome (Radix imperatoriae) has been used as a stimulant, a stomachic, a diuretic for The biological activity of medicinal plants is mainly due to the chronic indigestion, as well as a therapeutic for typhoid, intermittent presence of secondary metabolites, compounds with no apparent fever, paralytic conditions, and in delirium treatment [2]. Furthermore, function in the primary metabolism of the organism, but with a re- its use as anti-inflammatory agent is also documented [3]. markable importance in defense, competition and adaptation to the Coumarins are the most peculiar secondary metabolites found in environment. These molecules often have specific functions and many this species. In particular, the presence of imperatorin, ostruthol, oxy- of them present biological activities that can be useful for the treatment peucedanin hydrate, oxypeucedanin, isoimperatorin (all fur- of human diseases, therefore providing new active ingredients for anocoumarins) and ostruthin and osthole is reported [1]. Coumarins are pharmacological research. related to a wide range of biological activities such as antiinflammatory Due to the growing interest in medicinal plants for the discovery and antioxidant [4], anticancer [5], antifungal and antimicrobial [6] and rational design of new drugs, we decided to investigate the meta- activity. In addition, ostruthol acts as inhibitor of acetylcholinesterase, bolic content and the potential biological activities of the species which could imply a strong potential for the treatment of Alzheimer's Peucedanum ostruthium (vernacular name “Imperatoria”). This plant is disease (AD) [7]. traditionally employed in Austria and Italy. The parts of Peucedanum The many therapeutic applications of Imperatoria prompted us to ostruthium with documented uses are the leaves and the rhizome, further deepen its secondary metabolite content by combining NMR commonly used for the preparation of tisanes, liqueurs and bitters, or as spectroscopy and UPLC coupled with high resolution mass spectrometry powders applied on the skin to treat gastro-intestinal, cardiovascular (HR-MS).

⁎ Corresponding authors at: Department of Biotechnology and Biosciences, University of Milano – Bicocca, P.zza della Scienza 2, 20126 Milan, Italy. E-mail addresses: [email protected] (A. Palmioli), [email protected] (C. Airoldi). https://doi.org/10.1016/j.bioorg.2018.10.016 Received 5 September 2018; Received in revised form 7 October 2018; Accepted 9 October 2018 Available online 12 October 2018 0045-2068/ © 2018 Elsevier Inc. All rights reserved. A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Interestingly, a first metabolic profiling of different kinds of aqueous 2.3. Chemical characterization and hydro-alcoholic extracts revealed the significant presence of chlorogenic acids (CGAs). By dissecting the potential neuroprotective 2.3.1 NMR spectroscopy activity of both roasted and green coffee extracts, we recently demon- Freeze-dried samples were suspended in D2O at a final concentra- strated that CGAs show interesting anti-amyloidogenic properties [8]. tion of 25 mg/mL, sonicated (37 kHz, 20 min, Elmasonic P 30H, Elma Briefly, CGAs are Aβ oligomer ligands able to inhibit Aβ1-42 peptide Schmidbauer GmbH, Singen, Germany) and centrifuged (14,000 rpm, on-pathway aggregation and Aβ-induced toxicity in SH-SY5Y human 5 min, 20 °C, ScanSpeed 1730R Labogene, Lynge, Sweden). 4,4- neuroblastoma cell line, suggesting the potential application of coffee Dimethyl-4-silapentane-1-sulfonic acid (DSS, final concentration extracts for AD prevention [8]. AD is the most common form of de- 0.5 mM) was added to the supernatant as an internal reference for mentia in older population and the number of people affected is ex- concentrations and chemical shift. The pH of each sample was verified pected to increase dramatically in coming years [9]. Since symptom with a microelectrode (Mettler Toledo, Columbus, OH, USA) for 5 mm treatments occur in a stage of the disease in which neuronal damage is NMR tubes and adjusted to 7.4 with NaOD or DCl. All pH values were irreversible, there is a strong need to identify a primary prevention corrected for the isotope effect. The acquisition temperature was 25°C. treatment for AD, and Aβ oligomers are a promising target for this All spectra were acquired on an Avance III 600 MHz NMR spectrometer purpose [9]. Thus, the interest in natural compounds capable of inter- (Bruker, Billerica, MA, USA) equipped with a QCI (1H, 13C, 15N/31P and fering with toxic oligomers of Aβ peptides is growing. 2H lock) cryogenic probe. 1H NMR spectra were recorded with Based on these evidences, we decided to perform a preliminary in- cpmgpr1d, noesygppr1d, ledbpgppr2s1d pulse sequences (Bruker library) vestigation of the nutraceutical potential of Peucedanum ostruthium ex- and 256 scans, spectral width 20 ppm, relaxation delay 5 s. They were tracts for the prevention of AD. To this end, the chemical analysis done processed with 0.3 Hz line broadening, automatically phased and through NMR and UPLC-HR-MS was complemented with biochemical, baseline corrected. Chemical shifts were internally calibrated to the DSS biophysical and biological assays, the last carried out on SH-SY5Y peak at 0.0 ppm. Compound identification and assignment were done human neuroblastoma cell line. with the support of 2D NMR experiments, comparison with reported assignments and the SMA analysis tool integrated in MestreNova 2. Experimental section Software. The 1H,1H-TOCSY (Total Correlation SpectroscopY) spectra were acquired with 48 scans and 512 increments, a mixing time of All chemicals and solvents were purchased from Sigma-Aldrich or 80 ms and relaxation delay of 2 s. 1H, 13C-HSQC (Heteronuclear Single Fisher scientific and used without further purification. Quantum Coherence) spectra were acquired with 64 scans and 512 increments, relaxation delay 2 s. 2.1. Preparation of plant extracts 2.3.2 Ultra-performance liquid chromatography/electrospray ionization- Samples of Peaucedanum ostruthium (rhizome and leaves) were field- high resolution mass spectrometry (UPLC/ESI-HR-MS) collected (Barmasc, Rû Courtod, Ayas, AO, Italy, 1900 mt above sea The UPLC/ESI-HR-MS analysis was carried by coupling an Acquity level) on August 13, 2016 and identified by a local ethnobotanical UPLC separation module (Waters, Milford, MA, USA) with in-line expert (Dr. Fiorenza Cout). photodiode array (PDA) eλ detector (Waters) to a Q Exactive hybrid Samples were frozen in liquid nitrogen and grounded into a fine quadrupole-Orbitrap mass spectrometer and an HESI-II probe for elec- powder. 1 g of each sample was introduced into a cellulose thimble and trospray ionization (Thermo Scientific, San Jose, CA, USA). The ion extracted in a Soxhlet apparatus (100 mL) with a hydroalcoholic mix- source and interface conditions were: spray voltage +3.0/−2.5 kV, ture (80% EtOH) during 10 cycles. Finally, the organic phase was re- sheath gas flow 60, auxiliary gas flow 20 and temperature 300 °C,ca- moved under reduced pressure at 40 °C and the aqueous residue was pillary temperature 350 °C. Positive mass calibration was performed freeze-dried. The lyophilized extracts were stored at −20 °C. with Pierce LTQ ESI Positive Ion Calibration Solution (Thermo Scientific Pierce, Rockford, IL, USA), containing caffeine, the tetra- 2.2. Preparation of polyphenol-enriched fractions peptide MRFA and Ultramark 1621. Negative mass calibration was performed with Pierce ESI Negative Ion Calibration Solution (Thermo 2.2.1 Solid-phase extraction Scientific Pierce), containing sodium dodecyl sulfate, sodium taur- Solid-phase extraction (SPE) was performed with Amberlite® XAD®-4. ocholate and Ultramark 1621. Four μL of sample (2 mg/mL in water) A SPE cartridge loaded with 1 g of resin was washed with EtOH (3x 10 mL) were separated using a Waters Acquity BEH C18 column and equilibrated with H2Omq. A solution of extract (120 mg in 2 mL of (150 × 2.1 mm, 1.7 μm, 130 Å) (Waters, Milford, MA, USA) kept at −1 H2Omq) was loaded and the column was stirred for 12 h. Then, non-ad- 40 °C, and using 0.1 mL 100 mL of formic acid in H2O MilliQ-treated −1 sorbed fraction (fract.1 - N.A.), washing fraction (2 mL of H2O, fract. 2 - water (solvent A) and 0.1 mL 100 mL formic acid in acetonitrile W) and elution fraction (4 mL of EtOH, fract. 3 – Ext) were collected, the (solvent B). For the UPLC separation, a linear elution gradient was organic solvent was removed under reduced pressure and aqueous re- applied (isocratic 5% B for 5 min then 5–50% of solvent B in 20 min) at sidues were freeze dried obtaining fraction 1 – 3 as solid residues. a flow rate of 0.2 mL min−1. The LC eluate was analyzed by Full MS and data dependent tandem MS analysis (dd-MS2) of five the most intense 2.2.2 Preparative reverse-phase column chromatography ions (Top 5). The resolution was set at 70,000 and 17,500 and the AGC Automated flash chromatography was performed on a Biotage® targets were 1 × 106 and 1 × 105 for Full MS and dd-MS2 scan types, Isolera™ Prime system equipped with Spektra package. A solution of respectively. The maximum ion injection times were 50 ms. The MS extract sample (200 mg in 1.5 mL of MeOH) was loaded in a pre-column data were processed using Xcalibur software (Thermo Scientific) and samples SNAP-C18 (1 g) and left air-dried overnight. Column chroma- Mnova MS (Mestrelab). Metabolites were determined according to their tography was performed on SNAP KP-C18-HS (12 g) cartridge using calculated exact mass and absorption spectra. Their structures were water (solvent A) and methanol (solvent B) as eluent solvents. A linear confirmed by high resolution tandem MS (HR-MS/MS) by comparison elution gradient was applied (2% B for 2 CV, 2–100% of B in 15 CV and to reported assignments in literature or databases. 100% B for 3 CV) at flow rate of 12 mL/min. The eluate was automatically collected in fraction based on photodiode array detector 2.4. In vitro antioxidant activity signal (range 200–400 nm). Fractions were pooled in homogenous groups, organic solvent was removed under reduced pressure and re- Antioxidant activity was evaluated by mean in vitro spectro- sidues were freeze-dried obtaining fractions A-E. photometric assays reported below. Absorbance measurements were

77 A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Fig. 1. (A) UPLC-PDA chromatogram of crude Soxhlet extract obtained from Imperatoria rhizome. Absorbance at 320 nm. Metabolites were identified according to absorption spectra, calculated exact mass and HR-MS/MS. (B) 1H NMR profile of the same extract dissolved in phosphate buffer, pH 7.4. The main metabolites’ assignments are reported. The spectrum was acquired at 600 MHz, 25 °C. performed with Varian Cary 50 Scan UV–Visible Spectrophotometer working solution of ABTS+ was daily prepared by diluting the stock using disposable polymethyl methacrylate (PMMA) semimicro 10 mm- solution (1:50) reaching an absorbance of 0.70 ± 0.05 at 734 nm. Fifty cuvettes relative to a blank solution. Extract samples were diluted to μL of sample (or standards) were added in a cuvette containing 950 μL 0.2 mg/mL, and standard solutions (0–200 μg/mL) of 5-CQA were used of ABTS+ solution, and the absorbance at 734 nm was read after 30 min for calibration (linear fitting R2 = 0.9985, n = 7). Results were ex- of incubation at room temperature. pressed as μg of Chlorogenic acid Equivalent (CQA eq)/mg of freeze- dried extract. Data were reported as means ( ± SD) of triplicate mea- 2.4.3 DPPH radical scavenging sures of three independent experiments. DPPH assay is based on the scavenging of the stable free-radical 2,2- diphenyl-1-picrylhydrazyl, according to literature [25]. Briefly, 950 μL of a 2.4.1 Total reducing capacity (TRC) - Folin-Ciocalteu assay diluted solution of DPPH in buffered MeOH (100 μM in a mixture of60% Total reducing capacity was measured by Folin-Ciocalteu’s phenol MeOH and 40% acetate buffer pH 4.5, Abs 0.70 ± 0.05) and 50 μLofa assay prior described by Singleton, Orthofer, and Lamuela-Raventós diluted sample (or standard) were added into a cuvette, the absorbance at (1999) [24]. Briefly, 80 μL of diluted samples (or standards/blank) and 517 nm was read after 30 min of incubation at room temperature. 40 μL of Folin’s reagent were dispensed in a cuvette containing 400 μL of H2Omq; then 480 μL of 10.75% (w/v) Na2CO3 solution was added and 2.5. Biological activity the solution was quickly mixed. The absorbance at 760 nm was read after 30 min of incubation at room temperature. 2.5.1 Peptide synthesis and sample preparation Aβ1-42 (DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGV- 2.4.2 ABTS radical scavenging VIA) was prepared by Solid Phase Peptide Synthesis (SPPS) on a Syro I ABTS assay is based on the scavenging ability of antioxidants to the synthesizer (Biotage, Uppsala, SE) using Fmoc-protected L-amino acid long-life intense colored radical cation 2,2′-azino-bis(3-ethylben- derivatives, NOVASYN-TGA resin and a 0.1 mM scale [26]. Peptide was zothiazoline-6-sulphonic acid). A 7 mM stock solution of ABTS·+ was cleaved from the resin and was treated to obtained monomers and produced by mixing an equal amount of a 14 mM ABTS solution and a oligomers as previously described [27]. The peptide was purified by RP- 4.9 mM K2S2O8 solution in H2Omq (final concentration 7.00 mM and HPLC on a semi-preparative Jupiter C4 column (5 µM, 300 Å, 250 x 2.45 mM, respectively). The mixture was left at room temperature in 10 mm, Phenomenex) with a mobile phase of 0.1% TFA in water (eluent dark for at least 12–16 h before use and stored at 4 °C for 7 days. A A) and 0.08% TFA in acetonitrile (eluent B), using a linear gradient

78 A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Fig. 2. Structures of the main compounds identified in the crude Soxhlet extract obtained from Imperatoria rhizome. from 5 to 50% of eluent B in 40 min. Peptide identity and purity was 2.5.3 Aβ-induced cytotoxicity confirmed by MALDI-TOF analysis (model Reflex III, Bruker). The Human neuroblastoma SH-SY5Y cell line was grown in Dulbecco’s purity of peptide was always above 95%. Modified Eagle’s medium (DMEM, Lonza) supplemented with L-gluta- mine (5 mM, Gibco), antibiotics (penicillin/streptomycin 10,000 U, 2.5.2 Thioflavin T binding Lonza) and 10% heat-inactivated fetal calf serum (FCS, Gibco). The SH-

Aβ1-42 was dissolved in 10 mM NaOH, H2O and PB 50 mM (1:1:2) to SY5Y cell line was seeded in 96-well plates (105 cell/mL) and incubated 2.5 μM with or without Imperatoria extracts (25 μg/mL) and were in- overnight (37 °C, in a humidified 5% CO2 atmosphere). After com- cubated at 37 °C with 20 µM ThT (Sigma, PB 50 mM, pH 7.4) in 96-well pleting planting the medium was replaced with 1% of FCS in DMEM, to black plates (Isoplate, Perkin Elmer). ThT fluorescence was monitored for reduce cell growth.

24 h by a plate reader (Infinite F500 Tecan: excitation 448 nm, emission Aβ1–42 was dissolved in 10 mM NaOH, H2O and PBS (1:1:2) and 485 nm, 37 °C). Data were expressed as the mean of three replicates, cal- added to Imperatoria fractions (100 μg/mL in 0.5% DMSO) to obtain a culated by subtracting the relative control solutions (extracts alone) and final concentration of 10 μM for Aβ1–42 in the well. Cytotoxicity was were expressed as the percentage reduction of Aβ1–42 aggregation. evaluated after 24 h incubation, using the MTT reduction assay.

79 .Plil tal. et Palmioli A. Table 1 UPLC/HR-MS for the major extract components’ identified in the crude Soxhlet extract obtained from Imperatoria rhizome, corredated by 1H NMR data for CGAs.

2 1 # RT (min) ID Name Molecular Monoisotopic mass Exp m/z Error Abs λmax MS (+) (rel. int.) H chemical shifts (ppm) formula [M+H]+ (ppm) (nm)

1 7.50 3-CQA 3-O-caffeoylquinic acid C16 H18 O9 354.0945 355.1024 0.12 298, 328 163.0392 (1 0 0), 91.0580 (65), 285.0104 – (37) 2 9.72 5-CQA 5-O-caffeoylquinic acid C16 H18 O9 354.0945 355.1022 0.57 298, 328 163.0392 (1 0 0) 7.63 (d, H3′), 7.13 (d, H5′), 6.82 (d, H9′), 377.0839a 1.07 6.69 (d, H8′), 6.32 (d, H2′), 5.32 (ddd, H5), 4.24 (m, H3) 3 9.78 Phenylpropanoid glycoside C16 H24 O7 328.1517 329.1596 0.25 252, 325 167.1070 (1 0 0), 111.0445 (68), 125.0600 – (30) C10 H14 O2 166. 0994 167.1067 0.02 125.0600 (1 0 0), 111.0445 (78), 121.1015 aglycone (52) 4 10.08 4-CQA 4-O-caffeoylquinic acid C16 H18 O9 354.0945 355.1026 0.54 298, 328 163.0396 (1 0 0), 193.0502 (19) – 5 11.33 Scopoletin-7-O-rutinoside C22 H28 O13 500.1313 501.1607 0.82 296 193.0500 (1 0 0), 355.1026 (6) – C10 H8O4 192.0417 193.0497 0.58 – aglycone 6 11.69 5-FQA 5-O-feruloylquinic acid C17 H20 O9 368.1102 369.1175 1.38 298, 326 177.0549 (1 0 0), 145.0285 (8) – 7 12.25 Apt Apterin C20 H24 O10 424.1364 425.1442 0.15 326 187.0394 (1 0 0), 263.0917 (30), 85.0291 – (23) C14 H14 O5 262.0841 263.0917 1.28 – aglycone 8 13.33 3.4-diCQA 3.4-di-O-caffeoylquinic acid25 C H24 O12 516.1273 517.1342 0.35 298, 326 163.0392 (1 0 0) 7.49 (d, H3′b), 7.45 (d, H3′a), 6.33 (d, 499.1236b 0.17 H2′a), 6.19 (d, H2′a), 5.67 (m, H3), 4.99 (q, H4) 9 13.55 3.5-diCQA 3.5-di-O-caffeoylquinic acid25 C H24 O12 516.1273 517.1342 0.21 298, 326 163.0392 (1 0 0) 7.66 (d, H3′b), 7.64 (d, H3′a), 6.42 (d, 499.1233b 0.42 H2′a), 6.37 (d, H2′b), 5.48 (m, H3, H5),

80 4.08 (m, H4) 10 14.02 4.5-diCQA 4.5-di-O-caffeoylquinic acid25 C H24 O12 516.1273 517.1343 0.47 297, 326 163.0392 (1 0 0) 7.47 (d, H3′b), 7.40 (d, H3′a), 6.20 (d, 499.1237b 0.50 H2′b), 6.12 (d, H2′a), 5.64 (dd, H5), 5.17 (dd, H4), 4.42 (m, H3) 11 14.04 Hesp Hesperidin C28 H34 O15 610.1892 611.1957 2.25 286, 328 303.0867 (1 0 0), 85.0291 (30), 71.050 – (20), 263.0553 (18) C16 H14 O6 302.0785 303.0868 1.58 177.0549 (1 0 0), 153.0184 (68), 171.0291 aglycone (34) 12 14.40 Oxyh-Glc Oxypeucedanin hydrate C22 H26 O11 466.1470 467.1547 0.26 313 305.1024 (1 0 0), 203.0346 (23), 287.0920 – glycoside (15) C16 H16 O6 304.0941 305.1025 2.09 203.0345 (1 0 0) aglycone 13 14.80 Oxyh-Glc Oxypeucedanin hydrate C22 H26 O11 466.1470 467.1544 0.79 313 305.1024 (1 0 0), 203.0345 (30), 287.0919 – glycoside (20) C16 H16 O6 304.0941 305.1024 1.49 203.0345 (1 0 0) aglycone 14 14.97 3F.4CQA 3-O-feruloyl-4-O-caffeoylquinic C26 H26 O12 530.1419 531.1500 0.52 298, 328 177.0549 (1 0 0), 163.0392 (69), 319.0817 – acid 513.1394b 0.42 (20)

15 15.02 Peucenin gentobioside C27 H36 O14 584.2111 585.2178 0.14 297, 326 261.1124 (1 0 0), 205.0501 (58) – Bioorganic Chemistry83(2019)76–86 C15 H16 O4 260.1043 261.1043 1.24 – aglycone 16 15.20 4F.5CQA 4-O-feruloyl-5-O-caffeoylquinic C26 H26 O12 530.1419 531.1499 0.35 298, 328 177.0550 (1 0 0), 163.0392 (20), 145.0285 – acid 513.1394b 0.83 (11) 17 15.32 4C.5FQA 4-O-caffeoyl 5-O-ferruloylquinic C26 H26 O12 530.1419 531.1499 0.40 298, 327 177.0550 (1 0 0), 163.0392 (33), 145.0285 – acid 513.1394b 0.48 (10) 18 15.52 Peucenin gentobioside C27 H36 O14 584.2111 585. 2175 0.51 290, 324 205.0500 (1 0 0), 261.1123 (87) – C15 H16 O4 260.1043 261.1123 0.77 205.0500 (1 0 0) aglycone 19 16.02 3.4.5-triCQA 3.4.5-tri-O-caffeoylquinic acid34 C H30 O15 678.1579 679.1652 0.81 298, 328 163.0392 (1 0 0) – 20 16.32 Oxyh Oxypeucedanin hydrate C16 H16 O6 304.0941 305.1012 2.59 311 203.0343 (1 0 0) – (continued on next page) A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Tetrazolium solution (20 μL of 5 mg/mL, Sigma Aldrich) was added to each well and incubated for 4 h. The medium was replaced with acid- ified isopropanol (0.04 M HCl) to dissolve the purple precipitate andthe absorbance intensity was measured at 570 nm, using a plate reader (Infinite M200, Tecan). Data were expressed as percentages of controls (vehicle) for three separate replicates.

2.5.4 Statistical analysis All data were shown as mean ± standard deviation (SD). Statistical analysis was done using GraphPad Prism 4.0. One-way ANOVA, fol- lowed by Dunnett’s multiple comparison test, was used to assess the H chemical shifts (ppm) 1 – – – – – – – significance of differences between groups.

2.6. Atomic force microscopy (AFM)

Aβ1-42 was dissolved in 10 mM NaOH, H2O and PB 50 mM (1:1:2) to 2.5 μM with or without Imperatoria fractions (25 μg/mL) and were incubated at 37 °C for 24 h. At different time points, 20 μL of samples were spotted onto a freshly cleaved Muscovite mica disk and incubated

for 5 min. The disk was then washed with 5 mL of H2O and dried under a gentle nitrogen stream. Samples were mounted onto a Multimode AFM with a NanoScope V system (Veeco/Digital Instruments) operating

(+) (rel. int.) in Tapping Mode using standard antimony(n)-doped Si probes (T: 2 3.5–4.5 mm, L: 115–135 mm, W: 30–40 mm, f 0: 313–370 kHz, k: MS (30), 203.0344 (30), 347.1125 (27) (53), 287.0921 (35), 203.0347 (30), 127.0756 (26), 329.1029 (8) (12) 85.0656 (1 0 0), 145.0862(27), (48), 203.0348 103.0760 (23), 287.0921 (11) 85.0655 (1 0 0), 145.0861(27), (49), 203.0346 103.0760 (23), 287.0919 (11) 20–80 N/m) (Bruker). Samples were analyzed with the Scanning Probe Image Processor (SPIP Version 5.1.6 (released April 13, 2011) data max analysis package. SPIP software was used to analyze the distribution of Abs λ (nm) the molecular assemblies of the different populations in terms of height and diameter profiles, as previously described [28].

Error (ppm) 2.7. NMR binding studies

NMR spectra were recorded on a Bruker Avance III 600 MHz

+ equipped with a QCI cryo-probe (Bruker, Billerica, MA, USA).

m / z Preparation of a sample containing Aβ oligomers: lyophilized 509.1656 0.49595.1646 311 1.93 85.0655 (1 0 0),287.0908 145.0860 (85), 320 287.0916 387.1430 2.20271.0956 145.0861 (1 0 0), 2.18271.0961 85.0655 (72) 347.1129 299.1642 309 3.12 309 1.46 0.04 203.0342 300 (1 0 0), 85.0655 83.0498 (24) 310 (1 0 0), 185.1176 (48), 330 167.1069 203.0344 (1 0 0), 69.0707 203.0344 (21) (1 0 0) 175.0391 (1 0 0) 347.1123 0.59 347.1119 1.86 [M+H] Aβ1–42 was dissolved in 10 mM NaOD (280 µL) then diluted with 20 mM phosphate buffer (140 µL) and the pH adjusted to 7.4. After1h incubation at 37 °C, extract (final concentration 15 mg/mL) or Fraction C sample (final concentration 15 mg/mL) dissolved in 140 µL of20mM phosphate buffer (pH 7.4) was added. The pH was measured witha Microelectrode (Mettler Toledo) for 5 mm NMR tubes and adjusted to 508.1575 594.1579 286.0836 386.1360 270.0887 270.0887 298.1564 346.1047 346.1047 Monoisotopic mass Exp pH 7.4 with NaOD and/or DCl. Total sample volume was 560 µL. A basic sequence from Bruker library was employed for STD experiment. For STD, a train of Gaussian-shaped pulses of 50 ms each was 12 7 5 7 5 7 4 4 3

O O O O O O O O O employed to selectively saturate the protein envelope; the total sa- 28 19 30 19 14 22 14 14 22

H H H H H H H H H turation time of the protein envelope was adjusted by the number of 24 18 27 18 16 21 16 16 19 C C C C C C C C C aglycone aglycone Molecular formula shaped pulses and set at 3 s. The on- and off-resonance spectra were acquired in an interleaved mode with the same number of scans. The STD NMR spectrum was obtained by subtracting the on-resonance spectrum from the off-resonance spectrum. A reference experiment acquired on a sample containing only the crude extract was run under the same experimental conditions to verify true ligand binding. The effects observed in the presence of the peptide oligomers were dueto true saturation transfer because there were no signals in the STD NMR spectra obtained in the reference experiments. Oxypeaucedanin 2′-acetate- 3′glucoside Imperatorin Oxypeaucedanin 2′-acetate-3′- (6″malonyl)-glucoside Name 3. Results and discussion

3.1. Preparation and characterization of Imperatoria extracts and fractions .

+ enriched in polyphenolic compounds . +

O+H] As already mentioned, the traditional local medicine uses 2 Peucedanum ostruthium for the preparation of medicative tisanes or li- ( continued ) queurs. Thus, in order to optimize the extraction of bioactive metabo- [M−H [M+Na]

a b lites from plant leaves and rhizomes, we tested three different experi- 21 16.47 24 22.5725 Osto 22.7726 23.1027 Isoimp 23.52 Imp Ostruthol Osti Isomperatorin Ostruthin 22 17.12 23 21.35 Oxy Oxypeucedanin # RT (min) ID

Table 1 mental procedures: (a) aqueous extraction at 100 °C, (b) hydroalcoholic

81 A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Fig. 3. (A) 1H NMR spectra of fractions from SPE on XAD-4: (1) crude extract of Imperatoria rhizome obtained by Soxhlet extraction; (2) not retained fraction; (3) fraction eluted with water; (4) fraction eluted with ethanol. (B) 1H NMR spectra of fractions from RP chromatographic separation: (1) fraction A; (2) fraction B; (3) fraction C; (4) fraction D; (5) fraction E; (6) fraction F, as collected on the basis of the chromatographic UV-DAD trace depicted in Supplementary material – Fig. S6. All the spectra were recorded at 600 MHz, 25 °C, after sample dissolution in phosphate buffer, pH 7.4.

(20% ethanol) ultrasound-assisted extraction, (c) hydroalcoholic (80% 320 nm (1A) and the 1H NMR spectrum (1B) of crude Soxhlet extract ethanol) Soxhlet extraction. obtained from Imperatoria rhizome. Structures and detailed spectro- Preliminary analysis of extracts, i.e. UV–Vis, total polyphenol con- scopic data for each identified compound are reported in Fig. 2 and tent and 1H NMR metabolic profiling (Supplementary material – Figs. Table 1 respectively. S1–S3) highlighted that hydroalcoholic Soxhlet extraction of rhizome Data analyses allowed the identification of 27 polyphenolic com- afforded the best results in terms of overall yield of polar aromatic pounds in Peucedanum ostruthium crude extracts, including phenolic metabolites, allowing to obtain 43% w/w yield (430 ± 32 mg out of acids, simple coumarins, furanocoumarins and flavonoids glycosides. 1 g of dried sample, n = 3). CGAs, a group of esters formed between quinic acid and trans-hy- Metabolites contained in the crude rhizome extract were identified droxycinammic acids, mainly caffeic, p-coumaric and ferulic acids, by a combined analytical approach based on NMR spectroscopy and were found as the most abundant polyphenols. Among them we iden- UPLC-HR-MS. In particular, UPLC separation was monitored through a tified O-caffeoylquinic acids (3-CQA, 4-CQA and 5-CQA isomers), 5-O- PDA detector to reveal the characteristic polyphenol absorbances at 280 feruloylquinic acid (5-FQA) and di-O-caffeoylquinic acids (3,4-diCQA, and 320 nm. NMR spectroscopy data were complementary used for 3,5-diCQA, 4,5-diCQA isomers) as the most representative; further- metabolites identification and quantification, including non-ionizable more, di-O-feruloyl-caffeoylquinic acids (3F,4-CQA, 4F,5-CQA and compounds. Fig. 1 depicts the chromatographic UV trace extracted at 4C,5-FQA isomers) and 3,4,5-tri-O-caffeoylquinic acid (3,4,5-CQA)

82 A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Fig. 6. MTT assay determination of cell viability after 24 h treatment of SH- Fig. 4. Relative ThT fluorescence measured after 24 h incubation of Aβ1-42 SY5Y cells with a solution of Aβ 1–42 peptide (10 µM) in oligomeric form in (2.5 µM) with 20 µM ThT in absence or presence of Imperatoria fractions A, B, C absence or presence of Imperatoria fractions A, B, C or XAD (100 μg/mL). Cell and XAD (25 μg/mL). Data were expressed as the mean of three replicates, viability was expressed as percentage of the corresponding control (vehicle) and calculated by subtracting the relative control solutions (fraction alone), and data reported as mean ± SD of 3 independent experiments. ***p < 0.001 Aβ were expressed as mean ± SD of the percentage reduction of Aβ1–42 ag- vs control, °°p < 0.01, °°°p < 0.001 Aβ + tested fraction vs Aβ alone (one- gregation, °°°p < 0.001 Aβ + tested fraction vs Aβ alone (one-way way ANOVA + Dunnett post-test). CT = control (vehicle PBS); ANOVA + Dunnett post-test). DMSO = control (vehicle 0.5% DMSO in PBS).

Hesperetin-7-rutinoside (Hesperidin), whose presence in Peucedanum ostruthium rhizome had been previously described [12], was identified as the major glycosylated flavonoid in the mixture. The coumarin ostruthin (6-geranyl-7-hydroxycoumarin) and the linear furanocoumarins oxypeucedanin, oxypeucedanin hydrate, ostruthol, imperatorin and isoimperatorin were unequivocally identified as mean of monoisotopic mass and their MS(+) fragmentation pattern, in comparison with published data [1]. The presence of an angelicin- type dihydroxyfuranocoumarin glycoside (Apterin) was detected according to what observed in other Peucedanum species [13]. Furthermore, compounds eluting at 14.38 and 14.80 min retention time gave an identical mass spectrum corresponding to a molecular 2 formula of C22H26O11 and the neutral loss of 162 amu in MS spectra indicating the presence of an O-hexosyl group and a common C16H16O6 fragment, representing a psolaren-type 7-(2,3-dihydroxy-3-methylbu- toxy) coumarin aglycone. Esters or glycosides of furanocoumarins containing hydroxyl groups in the prenyl chain are seldom found in nature [14]. However, some linear furanocoumarin glycosides were found in the rhizomes of Apiacae family [15]. Thus, we tentatively identify these compounds with oxypeucedanin hydrate 3′-O-glucopyranoside [15c] and oxypeucedanin hydrate 2′-O-glucopyranoside. Compounds eluting with a retention time of 16.47 and 17.12 min

showed mass data corresponding to a molecular formula of C24H28O12 and C27H30O5 respectively and a common aglycone fragment relative to a molecular formula of C18H19O7, clearly identified as oxypeucedanin hydrate 3′-acetate, as previously found in Peucedanum ostruthium [16]. The first one showed a neutral loss of 162 amuin tandem MS experiments, leading to the identification of oxypeucedanin hydrate 3′-acetate-2′-glucopyranoside, whereas the second one gave a neutral loss of 248 amu (anhydromalonylglucose) and (less abundant) loss of 266 amu (malonylglucose), resulting in the identification of oxypeucedanin hydrate 3′-acetate-2′-(6″-malonyl) glucopyranoside. Detailed MS data and a fragmentation pathway for the identification of compounds 21 and 22 are available in Supplementary Fig. 5. AFM images recorded after 24 h incubation of Aβ 1–42 peptide (2.5 µM) material – Figs. S4 and S5. in absence or presence of Imperatoria fractions A, B, C or XAD (25 μg/mL) Moreover, a mass peak corresponding to a molecular formula of

(Scale bar: 1 nm). C16H24O7 and with a retention time of 9.78 min was observed. Upon fragmentation, it gave a neutral loss of 162 amu, generically identified were detected as minor components. Structures and isoforms of CGAs as phenylpropanoid glycoside. Compounds eluting at a retention time were determined on the basis of tandem MS(-) (data not shown), in of 15.02 and 15.52 min showed a mass peak of 585.2175 [M+H+], agreement with literature data [10]. In addition, the main CGAs were resulting in a molecular formula of C27H36O14 and gave fragment ions identified also by 1H NMR with the support of 1H,1H-TOCSY and 1H, in positive mode at 261.1123 (87%) and 205.0501 as base peak. 13C-HSQC and through comparison with reported assignments [11]. Neutral losses of 324 amu and 56 amu indicated the presence of a

83 A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Fig. 7. (A) 1H NMR and (B) STD-NMR spectra acquired on a mixture containing the total Soxhlet extract of Imperatoria rhizome (15 mg/mL) and Aβ1-42 peptide (80 μM) in oligomeric form dissolved in deuterated phosphate buffer, pH 7.4. STD spectrum was acquired with 1024 scans and 2 s saturation time at 600 MHz, 25 °C. (C) 1H NMR spectrum acquired on a sample-containing fraction C (15 mg/mL) dissolved in deuterated phosphate buffer, pH 7.4.

gentiobioside moiety and a prenyl substituent respectively, thus sug- 3.2. Effects of polyphenol-enriched fractions on Aβ peptide aggregation and gesting a prenylated aglycone with a molecular formula of C15H16O4, neurotoxicity. tentatively identified as peucenin. To the best of our knowledge, no peucenin glycoside was reported in literature, so we tentatively iden- In light of 5-CQA ability to bind Aβ1-42 and prevent its aggregation tified these compounds as chromone gentiobioside isomers, presenting and neurotoxicity [8], the identification of several isomers of CGAs, peucenin as aglycone moiety. including di-caffeoyl and caffeoyl-feruloylquinic acids, as major com- Finally, compound eluting with 11.33 min of retention time ponents of polyphenol-enriched fractions of Imperatoria rhizome ex- showed a mass peak at 501.1607 [M+H+]. Upon fragmentation, tract (A, B and C fractions from RP-chromatography and XAD fraction ions at 355.1026 (6%) and 193.0497 as base peak, with a neutral loss from SPE), prompted us to test their potential anti-amyloidogenic ac- of 148 and 308 amu, were observed, suggesting the presence of a tivity. rutinoside moiety, and a C10H8O4 aglycone, resulting in a tentative Firstly, their efficacy in inhibiting the aggregation of Aβ1-42 pep- identification of scopoletin-7-O-rutinoside, also found in Asteracae tide was verified by ThT assay [18] and AFM analysis, employed to species [17]. evaluate peptide aggregation after 24 h of co-incubation at 37 °C with In addition, both NMR and MS analysis allowed the identification of the tested fractions dissolved at a concentration of 25 μg/mL. According other metabolites and nutrients, such as choline, malic acid, γ-amino- to relative ThT fluorescence values (Fig. 4), all the fractions reduced butyric acid (GABA), L-alanine, D-glucose and sucrose. peptide aggregation. In fact, peptide co-incubation with the tested Starting from a portion of crude rhizome extract (Fig. 3A), a fractions induced a decrease of the fluorescence emitted by ThT upon fraction enriched in polyphenolic compounds were prepared by solid binding to Aβ aggregates of 71%, 82.3%, 88.3% and 75.7% for fractions phase extraction (SPE) on Amberlite XAD-4 resin. A retained XAD A, B, C and XAD respectively. fraction eluted with ethanol was achieved with a 10.3% yield. Its 1H The AFM images (Fig. 5) obtained after 24 h of incubation of 2.5 µM NMR profile is reported in Fig. 3A-4. As can be deduced from the Aβ 1–42 peptide with the same enriched fractions of Imperatoria ex- comparison with the 1H NMR spectra of the crude extract (Fig. 3A-1), tract at a concentration of 25 μg/mL showed a significant reduction of the not retained fraction (Fig. 3A-2) and the fraction eluted with peptide aggregation, in particular the formation of peptide fibrils and water (Fig. 3A-3), XAD-retained fraction (from here named as macro-aggregates, with a “potency” following the order XAD < B < “fraction XAD”) (Fig. 3A-4) showed the major content of aromatic C. In fact, comparing the AFM image of the peptide alone with that of metabolites. peptide co-incubated with fraction C, the disappearance of fibrils and Moreover, another portion of crude extract was submitted to reverse the presence of small globular structures with a diameter of 40 nm ± phase (RP) chromatographic separation on RP-C18 column, collecting 15 nm and only a few prefibrillar structures with a length of the eluate in seven fractions (A-G) (Fig. 3B), based on chromatographic 60 nm ± 20 nm were evident. The sample obtained from the peptide UV-DAD trace, as reported in Supplementary material – Fig. S6. As co-incubation with fraction B showed almost exclusively short pre-fi- inferred from 1H NMR profiles (Fig. 3B), fraction C (Fig. 3B-3) showed brillary structures of 150 nm ± 20 nm in length, while amorphous the highest enrichment in aromatic compounds. In particular, both material (big globular structures) and few pre-fibrillar structures were NMR and UPLC-HR-MS revealed that fractions XAD and C resulted present upon peptide co-incubation with XAD fraction. Fraction A in- enriched in CGAs with a complete removal of the apolar fur- duces only a very mild reduction of the peptide aggregation, at variance anocoumarins comprised from 21.33 and 23.52 min RT in the chro- with ThT assay results (Fig. 4). matogram of the total extract (Fig. 1A and S7). Fraction XAD resulted Finally, an MTT assay [19] carried out on human neuroblastoma enriched in mono-substituted and di-substituted CGAs, being 5-CQA SH-SY5Y cell line [8] allowed the evaluation of the protective effect of and 3,5-CQA the most abundant (Fig. S8), while fraction C in di-sub- Imperatoria fractions against the neurotoxicity induced by Aβ1–42 stituted CGAs, in particular di-caffeoyl and caffeoyl-feruloylquinic acid oligomers (Fig. 6). isomers (diCQA and F, CQA), flavonoids and furanocoumarin glyco- Cells were treated with 10 µM Aβ1-42 peptide in oligomeric form sides (Fig. S9). and incubated for 24 h with or without 100 µg/mL of different

84 A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Imperatoria fractions. The co-treatment with fractions B, C and XAD 4. Conclusion improved significantly cell survival (+13.5%, +21.3% and +12.5% respectively), while fraction A showed no significant effect. In conclusion, we exploited NMR and UPLC-HR-MS to characterize In addition, in view of the dramatic role of oxidative stress in AD the metabolic profile of Peucedanum ostruthium, a medicinal plant al- onset [20], we briefly characterized the in vitro antioxidant activity of ready described for its several beneficial effects on human health. our samples, measuring the total reducing capacity (Folin-Ciocalteu The detection of a conspicuous content of CGAs tickled us to test the assay) and the ability to scavenge radical species (ABTS and DPPH putative anti-amyloidogenic activity of Imperatoria, based on recent assay). Our data indicated that fraction B and C from RP chromato- findings about the preventive effect of mono-substituted CGAs against graphy showed the highest antioxidant capacity, suggesting that this Aβ peptides [8]. effect correlates with the total amount of CGAs(Supplementary mate- We employed ThT assay and AFM analysis to monitor Aβ1-42 ag- rial – Fig. S10). gregation and to study aggregates’ morphology in absence and presence of polyphenol-enriched fractions of crude Imperatoria rhizome extract. 3.3. NMR-based identification of Aβ oligomer ligands contained in extracts The sample showing the highest potency in preventing peptide ag- from Imperatoria rhizome gregation was also the most effective in preventing Aβ-induced neurotoxicity, according to MTT assays carried out on the SH-SY5Y human NMR-based ligand-receptor interaction studies were exploited to cell line. detect extract components able to bind directly Aβ1-42 oligomers as STD-NMR experiments revealed that di-substituted CQAs are the putative molecules responsible for the biological activities previously best ligands of Aβ1-42 oligomers contained in Imperatoria rhizome investigated. extract, as they showed relative STD intensities higher than mono- STD-NMR experiment [21] is a robust and reliable method to reveal substituted CQAs and thus higher affinities for the target. Our study protein ligands in complex compound mixtures [8,22]. In addition, in showed also glycosylated flavonoids and furanocoumarins as able to the last years, several works have demonstrated the suitability of this bind Aβ1-42 oligomers. The co-presence of these compounds in the experiment for the identification of ligands of Aβ oligomers andthe same extract allows obtaining a significant biological activity. characterization of their binding mode [8,22b,22c,23]. All together, these data provide important structural information for STD NMR experiments were carried out on a sample containing a the rational design of new molecules with anti-amyloidogenic activity mixture of the total Soxhlet extract of Imperatoria rhizome and Aβ and molecular tools for the specific targeting of amyloid aggregates. oligomers (nominal peptide concentration of 80 µM) dissolved in Moreover, our results demonstrate once again the power of NMR deuterated phosphate buffer (pH 7.4, 25 °C)(Fig. 7). spectroscopy in the screening of complex natural extracts for the direct Aβ1–42 peptide was dissolved in aqueous buffer according to the identification of bioactive molecules, bypassing the isolation ofthe procedure previously reported [23c] that assures its oligomeric state. single components. Briefly, lyophilized Aβ1-42 is dissolved in 10 mM NaOD to allowthe disaggregation of any preformed seeds, leading to an Aβ peptide solu- Acknowledgments tion containing monomers and some low-molecular-weight oligomers. The sample is then diluted with phosphate buffer and the pH adjusted This work was supported by Fondazione CARIPLO and Regione to 7.4, inducing immediately the peptide aggregation, to afford a Lombardia, project n° 2015-0763, and partially supported by sample enrichment in Aβ oligomers. After 1 h of incubation at 37 °C the Fondazione Sacchetti 2016-2017. sample consists mainly of soluble Aβ oligomers, as checked by AFM and The authors acknowledge Dr. Fiorenza Cout for providing us sam- NMR diffusion experiments [23c]. ples of Peucedanum ostruthium leaves and rhizome, and Alberto Gianni Crude extract was added after its dissolution in phosphate buffer, pH for his support. 7.4, 25 °C. The selective saturation of some aliphatic protons of Aβ oligomers was achieved by irradiating at −1.00 ppm (on-resonance fre- Conflicts of interest quency). If the compounds present in the test mixture are not irradiated directly (verified by blank experiments on a sample containing the mixture Authors declare no conflicts of interest. only), the detection of their NMR signals in the STD spectrum is a non- ambiguous evidence of their interaction with the receptor. Conversely, any Appendix A. Supplementary material signal from non-binding compounds is deleted in the difference spectrum (STD spectrum). Thus, the presence of resonances belonging to mono- and Supplementary data to this article can be found online at https:// di-CGAs, furanocoumarins and flavonoids in the STD spectrum (Fig. 7B) doi.org/10.1016/j.bioorg.2018.10.016. was the unequivocal demonstration of their interaction with peptide oligomers. In particular, the compounds showing the highest STD relative References intensities were di-substituted CGAs. Notably, also signals in the regions 4.5–3.2 ppm appeared in the STD spectrum. Notably, they are not signals [1] S. Vogl, M. Zehl, P. Picker, E. Urban, C. Wawrosch, G. Reznicek, J. Saukel, B. Kopp, of the free sugars glucose and sucrose, metabolites that are very abundant Identification and quantification of coumarins in peucedanum ostruthium (L.)Koch by HPLC-DAD and HPLC-DAD-MS, J. Agric. Food. Chem. 59 (9) (2011) 4371–4377. in the crude extract. In fact, the resonances in this region of the STD [2] O. Gökay, D. Kühner, M. Los, F. Götz, U. Bertsche, K. Albert, An efficient approach spectrum do not belong to these molecules, as can be deduced from the for the isolation, identification and evaluation of antimicrobial plant components form of the signals by comparing the STD spectrum (Fig. 7B) with the 1H on an analytical scale, demonstrated by the example of Radix imperatoriae, Anal. Bioanal. Chem. 398 (5) (2010) 2039–2047. NMR spectrum of the mixture (Fig. 7A). We can infer that they belong to [3] H. Joa, S. Vogl, A.G. Atanasov, M. Zehl, T. Nakel, N. Fakhrudin, E.H. Heiss, the carbohydrate moieties of glycosylated furanocoumarins and flavonoids P. Picker, E. Urban, C. Wawrosch, J. Saukel, G. Reznicek, B. Kopp, V.M. Dirsch, identified by UPLC-HR-MS. Thus, being fraction C from RP chromato- Identification of ostruthin from peucedanum ostruthium rhizomes as an inhibitor of graphy the only one containing all these molecular classes (mono- and di- vascular smooth muscle cell proliferation, J. Nat. Prod. 74 (6) (2011) 1513–1516. [4] K.C. Fylaktakidou, D.J. Hadjipavlou Litina, K.E. Litinas, D.N. Nicolaides, Natural CGAs, furanocoumarins and flavonoid glycosides), STD-NMR binding and synthetic coumarin derivatives with anti-inflammatory/antioxidant activities, studies allowed rationalizing the highest biological efficacy of this en- Curr. Pharm. Des. 10 (30) (2004) 3813–3833. riched fraction among those tested. [5] X.Y. Huang, Z.J. Shan, H.L. Zhai, L. Su, X.Y. Zhang, Study on the anticancer activity of coumarin derivatives by molecular modeling, Chem. Biol. Drug Des. 78 (4) An additional STD experiment, confirming these findings, was per- (2011) 651–658. formed on a sample containing fraction C and Aβ oligomers and is re- [6] S. Sardari, Y. Mori, K. Horita, R.G. Micetich, S. Nishibe, M. Daneshtalab, Synthesis ported in Supplementary Material – Fig. S11. and antifungal activity of coumarins and angular furanocoumarins, Bioorg. Med.

85 A. Palmioli et al. Bioorganic Chemistry 83 (2019) 76–86

Chem. 7 (9) (1999) 1933–1940. (2017) 297–306. [7] A. Urbain, A. Marston, K. Hostettmann, Coumarins from peucedanum ostruthium. [22] (a) M. Politi, M.I. Chávez, F.J. Cañada, J. Jiménez-Barbero, Screening by NMR: a as inhibitors of acetylcholinesterase, Pharm. Biol. 43 (8) (2005) 647–650. new approach for the study of bioactive natural products? The example of pleurotus [8] C. Ciaramelli, A. Palmioli, A. De Luigi, L. Colombo, G. Sala, C. Riva, C.P. Zoia, ostreatus hot water extract, Eur. J. Org. Chem. 2005 (7) (2005) 1392–1396; M. Salmona, C. Airoldi, NMR-driven identification of anti-amyloidogenic com- (b) C. Airoldi, E. Sironi, C. Dias, F. Marcelo, A. Martins, A.P. Rauter, F. Nicotra, pounds in green and roasted coffee extracts, Food Chem. 252 (2018) 171–180. J. Jimenez-Barbero, Natural compounds against Alzheimer’s disease: molecular [9] E. McDade, R.J. Bateman, Stop Alzheimer’s before it starts, Nature News 547 (7662) recognition of Aβ1–42 peptide by salvia sclareoides extract and its major compo- (2017) 153. nent, rosmarinic acid, as investigated by NMR, Chem. – Asian J. 8 (3) (2013) [10] (a) M.N. Clifford, K.L. Johnston, S. Knight, N. Kuhnert, Hierarchical scheme forLC- 596–602; MSn identification of chlorogenic acids, J. Agric. Food. Chem. 51 (10) (2003) (c) E. Sironi, L. Colombo, A. Lompo, M. Messa, M. Bonanomi, M.E. Regonesi, 2900–2911; M. Salmona, C. Airoldi, Natural compounds against neurodegenerative diseases: (b) M.N. Clifford, S. Knight, N. Kuhnert, Discriminating between the six isomersof molecular characterization of the interaction of catechins from green tea with dicaffeoylquinic acid by LC-MSn, J. Agric. Food. Chem. 53 (10) (2005) 3821–3832. Aβ1–42, PrP106–126, and Ataxin-3 Oligomers, Chem. – Eur. J. 20 (42) (2014) [11] (a) A. Tolonen, T. Joutsamo, S. Mattlla, T. Kämäräinen, J. Jalonen, Identification of 13793–13800; isomeric dicaffeoylquinic acids from Eleutherococcus senticosus using HPLC-ESI/ (d) A. Palmioli, C. Ciaramelli, R. Tisi, M. Spinelli, G. De Sanctis, E. Sacco, TOF/MS and 1H-NMR methods, Phytochem. Anal. 13 (6) (2002) 316–328; C. Airoldi, Natural compounds in cancer prevention: effects of coffee extracts and (b) F. Wei, K. Furihata, F. Hu, T. Miyakawa, M. Tanokura, Complex mixture ana- their main polyphenolic component 5-CQA on oncogenic Ras proteins, Chem. Asian lysis of organic compounds in green coffee bean extract by two-dimensional NMR J. 12 (2017) 2457–2466. spectroscopy, Magn. Reson. Chem. 48 (11) (2010) 857–865. [23] (a) R. Ahmed, B. VanSchouwen, N. Jafari, X. Ni, J. Ortega, G. Melacini, Molecular [12] L. Hörhammer, H. Wagner, D. Heydweiller, Hesperidin aus dem rhizom von Mechanism for the (−)-Epigallocatechin Gallate-Induced Toxic to Nontoxic Peucedanum ostruthium, Phytochemistry 8 (8) (1969) 1605. Remodeling of Aβ Oligomers, J. Am. Chem. Soc. (2017); [13] (a) H. Chang, Y. Okada, T. Okuyama, P. Tu, 1H and 13C NMR assignments for two (b) C. Airoldi, F. Cardona, E. Sironi, L. Colombo, M. Salmona, A. Silva, F. Nicotra, new angular furanocoumarin glycosides from Peucedanum praeruptorum, Magn. B. La Ferla, cis-Glyco-fused benzopyran compounds as new amyloid-[small beta] Reson. Chem. 45 (7) (2007) 611–614; peptide ligands, Chem. Commun. 47 (37) (2011) 10266–10268; (b) M. Eeva, J.P. Rauha, P. Vuorela, H. Vuorela, Computer-assisted, high-perfor- (c) C. Airoldi, L. Colombo, C. Manzoni, E. Sironi, A. Natalello, S.M. Doglia, mance liquid chromatography with mass spectrometric detection for the analysis of G. Forloni, F. Tagliavini, E. Del Favero, L. Cantu, F. Nicotra, M. Salmona, coumarins in Peucedanum palustre and Angelica archangelica, Phytochem. Anal. Tetracycline prevents Aβ oligomer toxicity through an atypical supramolecular 15 (3) (2004) 167–174; interaction, Org. Biomol. Chem. 9 (2) (2011) 463–472; (c) T. Yrjönen, M. Eeva, T.J. Kauppila, O. Martiskainen, J. Summanen, P. Vuorela, (d) C. Airoldi, G. D'Orazio, B. Richichi, C. Guzzi, V. Baldoneschi, L. Colombo, H. Vuorela, Profiling of Coumarins in Peucedanum palustre (L.) Moench M. Salmona, C. Nativi, F. Nicotra, B. La Ferla, Structural modifications of cis-gly- Populations Growing in Finland, Chem. Biodivers. 13 (6) (2016) 700–709. cofused benzopyran compounds and their influence on the binding to amyloid-β [14] M. Curini, G. Cravotto, F. Epifano, G. Giannone, Chemistry and biological activity of peptide, Chem. – Asian J. 11 (2) (2016) 299–309; natural and synthetic prenyloxycoumarins, Curr. Med. Chem. 13 (2) (2006) (e) C. Airoldi, S. Mourtas, F. Cardona, C. Zona, E. Sironi, G. D'Orazio, E. Markoutsa, 199–222. F. Nicotra, S.G. Antimisiaris, B. La Ferla, Nanoliposomes presenting on surface a cis- [15] (a) X. Zhao, X. Feng, X. Jia, M. Wang, Y. Shan, Y. Dong, New coumarin glucoside glycofused benzopyran compound display binding affinity and aggregation inhibi- from Angelica dahurica, Chem. Nat. Compd. 43 (4) (2007) 399–401; tion ability towards Amyloid β1-42 peptide, Eur. J. Med. Chem. 85 (2014) 43–50; (b) Y. Tada, Y. Shikishima, Y. Takaishi, H. Shibata, T. Higuti, G. Honda, M. Ito, (f) C. Guzzi, L. Colombo, A.D. Luigi, M. Salmona, F. Nicotra, C. Airoldi, Flavonoids Y. Takeda, O.K. Kodzhimatov, O. Ashurmetov, Y. Ohmoto, Coumarins and γ-pyrone and their glycosides as anti-amyloidogenic compounds: Aβ1–42 interaction studies derivatives from Prangos pabularia: antibacterial activity and inhibition of cytokine to gain new insights into their potential for Alzheimer's disease prevention and release, Phytochemistry 59 (6) (2002) 649–654; therapy, Chem. Asian J. 12 (1) (2017) 67–75; (c) S.K. Koul, K.L. Dhar, R.S. Thakur, A new coumarin glucoside from Prangos (g) A.R. Jesus, C. Dias, A.M. Matos, R.F.M. de Almeida, A.S. Viana, F. Marcelo, pabularia, Phytochemistry 18 (10) (1979) 1762–1763; R.T. Ribeiro, M.P. Macedo, C. Airoldi, F. Nicotra, A. Martins, E.J. Cabrita, (d) H.T. Chang, Y. Okada, T.J. Ma, T. Okuyama, P.F. Tu, Two new coumarin gly- J. Jiménez-Barbero, A.P. Rauter, Exploiting the Therapeutic Potential of 8-β-d- cosides from Peucedanum praeruptorum, J. Asian Nat. Prod. Res. 10 (6) (2008) Glucopyranosylgenistein: Synthesis, Antidiabetic Activity, and Molecular 577–581. Interaction with Islet Amyloid Polypeptide and Amyloid β-Peptide (1–42), J. Med. [16] A. Hiermann, D. Schantl, M. Schubert-Zsilavecz, J. Reiner, Coumarins from Chem. 57 (22) (2014) 9463–9472; Peucedanum ostruthium, Phytochemistry 43 (4) (1996) 881–883. (h) S. Merlo, E. Sironi, L. Colombo, F. Cardona, A.M. Martorana, M. Salmona, B. La [17] D.N. Olennikov, N.I. Kashchenko, C. Vennos, A new esculetin glycoside from Ferla, C. Airoldi, Cis-glyco-fused benzopyran derivatives as hit compounds for the Calendula officinalis (Asteraceae) and its bioactivity, Farmacia 65 (5) (2017) development of therapeutic and diagnostic tools against neurodegenerative dis- 698–702. eases, ChemPlusChem 79 (6) (2014) 835–843; [18] A. Hawe, M. Sutter, W. Jiskoot, Extrinsic fluorescent dyes as tools for protein (i) R. Ahmed, G. Melacini, A solution NMR toolset to probe the molecular me- characterization, Pharm. Res. 25 (7) (2008) 1487–1499. chanisms of amyloid inhibitors, Chem. Commun. 54 (37) (2018) 4644–4652. [19] T. Mosmann, Rapid colorimetric assay for cellular growth and survival: application [24] V.L. Singleton, R. Orthofer, R.M. Lamuela-Raventós, Analysis of total phenols and to proliferation and cytotoxicity assays, J. Immunol. Methods 65 (1) (1983) 55–63. other oxidation substrates and antioxidants by means of folin-ciocalteu reagent, [20] (a) G. Aliev, M.E. Obrenovich, V.P. Reddy, J.C. Shenk, P.I. Moreira, A. Nunomura, Methods in Enzymology, Academic Press, 1999, pp. 152–178. X. Zhu, M.A. Smith, G. Perry, Antioxidant therapy in Alzheimer’s disease: theory [25] O.P. Sharma, T.K. Bhat, DPPH antioxidant assay revisited, Food Chem. 113 (4) and practice, Mini Rev. Med. Chem. 8 (13) (2008) 1395–1406; (2009) 1202–1205. (b) Y. Feng, X. Wang, Antioxidant Therapies for Alzheimer's Disease, Oxid. Med. [26] C. Manzoni, L. Colombo, P. Bigini, V. Diana, A. Cagnotto, M. Messa, M. Lupi, Cell. Longevity 2012 (2012) 472932. V. Bonetto, M. Pignataro, C. Airoldi, E. Sironi, A. Williams, M. Salmona, The mo- [21] (a) M. Mayer, B. Meyer, Characterization of ligand binding by saturation transfer lecular assembly of amyloid Aβ controls its neurotoxicity and binding to cellular difference NMR spectroscopy, Angew. Chem. Int. Ed. 38 (12) (1999) 1784–1788; proteins, PLoS ONE 6 (9) (2011) e24909. (b) C. Airoldi, S. Sommaruga, S. Merlo, P. Sperandeo, L. Cipolla, A. Polissi, [27] C. Manzoni, L. Colombo, M. Messa, A. Cagnotto, L. Cantù, E. Del Favero, F. Nicotra, Targeting Bacterial Membranes: Identification of Pseudomonas aerugi- M. Salmona, Overcoming synthetic Aβ peptide aging: a new approach to an age-old nosa D-Arabinose-5P Isomerase and NMR Characterisation of its Substrate problem, Amyloid 16 (2) (2009) 71–80. Recognition and Binding Properties, ChemBioChem 12 (5) (2011) 719–727; [28] M. Messa, L. Colombo, E. del Favero, L. Cantù, T. Stoilova, A. Cagnotto, A. Rossi, (c) L. Amigoni, M. Stuknytė, C. Ciaramelli, C. Magoni, I. Bruni, I. De Noni, M. Morbin, G. Di Fede, F. Tagliavini, M. Salmona, The peculiar role of the A2V C. Airoldi, M.E. Regonesi, A. Palmioli, Green coffee extract enhances oxidative mutation in amyloid-β (Aβ) 1–42 molecular assembly, J. Biol. Chem. 289 (35) stress resistance and delays aging in Caenorhabditis elegans, J. Funct. Foods 33 (2014) 24143–24152.