SUPPLEMENTARY MATERIAL Glycosyl flavonoid profile, in vivo antidiabetic and in vitro antioxidant properties of Linaria reflexa Desf. Thamere Cherieta,b,*, Mourad Hanferc, Amel Boudjelald, Nadir Baalid, Ines Mancinie, Ramdane Seghiria, Souad Ameddahc, Ahmed Menadc, Fadila Benayachea, Samir Benayachea a Unité de Valorisation des Ressources Naturelles, Molécules Bioactives et Analyse Physicochimiques et Biologiques (VARENBIOMOL), Université des Frères Mentouri Constantine BP, 325 Route de Ain Elbey, Constantine 25017, Algérie. bFaculté des Sciences, Département de Chimie, Université de M'Sila, 28000 M'Sila, Algérie. cLaboratoire de Biologie et Environnement, Faculté des Sciences de la Nature et de la Vie, Université des Frères Mentouri Constantine BP, 325 Route de Ain Elbey, Constantine 25017, Algérie. dFaculté des Sciences, Département de Microbiologie et Biochimie, Université de M'Sila, 28000 M'Sila, Algérie. eLaboratorio di ChimicaBioorganica, Università di Trento, via Sommarive 14, I-38123 Povo-Trento, Italy. * Corresponding author e-mail: tamercheriet@umc.edu.dz Abstract: Aerial parts of Linaria reflexa, used in North-African folk medicine for treating certain skin diseases, were investigated by HPLC-DAD-ESI/MS technique able to identify the glycosyl flavonoids pectolinarin (1), isolinariin A (2), isolinariin B (3), linariin (4), isolinariin D (5) and isolinariin E (6) as the most abundant components in both hydroalcoholic (HAE) and ultrasoundassisted (UAE) extracts profiles. Metabolite 5, isolated and fully characterized by extensive NMR analysis, has been very recently reported from L. japonica together with the compound 6. Good antioxidant activities (DPPH radical scavenging, β-carotene bleaching and reducing power assays) were observed for the extracts. The remarkable antidiabetic activity displayed by UAE (300 mg/kg) has yielded the most marked decrease in blood glucose levels of the alloxan diabetic rats (-72.09 %), greater than the effects by the drug glybenclamide (-63.29 %). This study reports the first correlation of antidiabetic activity of Linaria sp. extracts with their chemical composition. Keywords: Linaria reflexa, HPLC-DAD-ESI/MS, Antidiabetic, Antioxidant, isolinariin D, isolinariin E. Contents 1. Plant material and crude extract preparation 2. Metabolite profile by HPLC-PDA/ESI/MS analysis 3. Isolation and purification of isolinariin D (5) 4. Structural characterization of isolinariin D (5) 5. Antioxidant activity of L. reflexa extracts 6. In vivo antidiabetic activity of L. reflexa extracts 7. References Table S1. Group division of rats used for in vivo anti-diabetic experiments of ultrasound- assisted extract (UAE) and aqueous extract (AQE) from L. reflexa. Table S2. In vivo Anti-diabetic evaluation: Variation in blood glucose level overnight fast, after oral administration of UAE and AQE from L. reflexa for 15 days Table S3. Effect of L. reflexa UAE and AQE dose on rat body weight after induction of diabetes. Table S4. In vivo Anti-diabetic evaluation: Variation in blood glucose and biochemical parameters after sacrifice of rats after different treatments. For abbreviations see Figure S1. DAD-HPLC chromatograms from online LC-MS analysis of UAE (A) and HAE (B) of the aerial parts of L. reflexa Desf. Figure S2. MS spectra in negative ion mode from LC-MS analysis of UAE of the aerial parts of L. reflexa Desf, associated to the chromatogram (A) in Fig.S1. Figure S3. Antioxidant activity of UAE and HAE from L. reflexa evaluated as DPPH radical scavenging property using ascorbic acid as a standard. Each value represents a mean ± SD (n=3), P<0.05. For abbreviations, see Table S1. Figure S4. Antioxidant activity by K3Fe(CN)6 test in reducing power of UAE and HAE of L. reflexa compared to quercetin and trolox as standards, evaluated as absorbance at 700 nm (a), and percentage inhibition as a function of concentration (b). Each value represents a mean ± SD (n=3), P<0.05. For abbreviations, see Table S1. Figure S5. Antioxidant activity by β-carotene bleaching effect of UAE, HAE from L. reflexa and standards (a) and percentage inhibition (b). Each value represents a mean ± SD (n=3), P<0.05 1. Plant material and crude extract preparation The aerial parts of L. reflexa Desf. were collected in March 2012 at El-Meridj near Constantine city, Algeria. A voucher specimen (n° 06/2012/SLR) has been deposited in the herbarium of the Research Unit: VARENBIOMOL. Air-dried leaves and flowers were extracted with two different methods. In the conventional HAE, 15 g of the plant material was extracted using a 250 mL (× 3) solution containing MeOH:H2O (8/2 v/v), left to macerate at room temperature for 48 h to residue 0.69 g after evaporation in vacuo of the solvents. In the UAE procedure, the plant material (15 g) was extracted with methanol (250 mL) using a sonicator Bandelin Sonorex RK510H (Germany) at 35 KHz and P180/320W for 2 h at room temperature. The solvent was evaporated in vacuo to leave 3.5 g of residue. The two extracts were used for LC-ESI/MS analysis and for antioxidant evaluation, while only UAE residue was used for the in vivo antidiabetic assays. The aqueous extract (AQE) was prepared according to the traditional method. The powder from dried L. reflexa was dissolved in boiling distilled water; the hot infusion was then left to reach room temperature (15 min) and filtered. The filtrate was used as such for biological tests on animals. The methanolic extract (UAE) was filtered, concentrated under reduced pressure to a semisolid mass free from solvent, which was administered orally after suspending in normal saline. 2. Metabolite profile by HPLC-DAD-ESI/MS analysis LC-MS was performed using a Hewlett–Packard (Palo Alto, CA, USA) Model 1100 Series liquid chromatograph coupled to both a Photo Diode Array detector (Agilent, Palo Alto, CA, USA) 1100 Series, and to an Esquire LC–ion trap mass spectrometer (Bruker Daltonics, Billerica, MA, USA) equipped with an electrospray ionization (ESI) interface. The polyphenols were separated under the following conditions: Macherey-Nagel Nucleosil C18-column (250 mm × 4.6 mm i.d.ν 5 m particle diameter, end-capped), the Photo Diode Array (PDA) detector was set at 330 nm, an 8/2 H2O (+ 1 mL of ammonium acetate 10 mM):CH3CN (+ 1 mL of ammonium acetate 10 mM) eluent was used for 30 min, then to 0:100 at 56 min; the split of the column effluent was applied to achieve a flow rate of 1 mδ/min into the mass spectrometer. High-purity nitrogen was used as the nebulizer at 35 psi, and also as the drying gas at 300°C, at a constant flow rate of 6 L/min. Full scan spectra were acquired in negative ion mode in the region m/z 100–1000. The structural information for the compositions was obtained from the in-source collisional induced dissociation fragments on the selected precursor ions. Data Analysis, version 3.0 (Bruker Daltonik GmbH) was used to analyze the mass spectra. 3. Isolation and purification of isolinariin D (5) All evaporations were carried out at room temperature at reduced pressure. Thin layer chromatography (TLC): Merck silica gel F254, Merck RP-18 (reversed phase) F254, with visualization by either UV light or by treatment with an acid solution of cerium sulfate. Flash chromatography (FC): Merck Si-60 15–25 m, εerck δichroprep RP-18; preparative thin layer chromatography (PLC) was realized on 20 × 20 cm Merck Kieselgel 60 F254 0.5 mm plates. Methanolic extract was subjected to a flash chromatography (FC) on reversed phase RP-18 with MeOH/H2O starting from 20:80 to 100:0, to obtain 25 fractions. The purification of fraction 18 by preparative liquid chromatography (PLC) eluted with chloroform/methanol (80:20) gives compound 5 (1.7 mg). Starting from the hydroalcoholic extract, chloroform, ethyl acetate and n-butanol extracts were obtained. The chloroform extract was fractioned by FC using hexane/ethyl acetate as gradient eluent to obtain 10 fractions (F1-F10). Fraction F8 (542.4 mg) was fractionated by FC (15 g) with chloroform/isopropanol as gradient eluent to obtain 17 sub-fractions; the purification of subfraction F8-11 (17.3 mg) by PLC (CHCl3/isopropanol 85:15) yielded pure compound 5 (3.2 mg), used for structural characterization. 4. Structural characterization of isolinariin D (5) 4.1. General Polarimetric data were obtained with a Bellingham & Stanley Limited ADP 440 apparatus, reporting [α]D in dm 1 deg mL g 1. UV data was recorded on Perkin Elmer instrument, version Lambda 25 UV/VIS spectrometer using methanol as solvent. IR spectra were recorded by using a FT-IR Tensor 27 Bruker spectrometer equipped with Attenuated Transmitter Reflection (ATR) device at 1 cm 1 resolution in the absorption region Δ 4,000–1,000 cm 1. A thin solid layer was obtained by evaporation of methanol solution of the sample. The instrument was purged with a constant dry nitrogen flow. Spectra processing was made using Opus software package. The intensities of signals are described as very strong (vs), strong (s), middle (m) or weak (w). NMR spectra were recorded on a Bruker-Avance 400 spectrometer by using a 5-mm BBI probe 1H at 400 MHz and 13C at 100 MHz in CD3OD ( H = 3.31 and C = 49.00 ppm), values in ppm, J values in Hz. Structural assignments are from correlation spectroscopy (COSY), heteronuclear single quantum correlation (HSQC) and heteronuclear multiple bond correlation (HMBC) experiments. Electrospray ionization (ESI) mass spectra were taken by using a Bruker Esquire-LC spectrometer with an electrospray ion source used in positive or negative ion mode by direct infusion of a methanolic solution of the sample, under the following conditions: source temperature 300°C, drying gas N2, scan range 100–1000 m/z. MALDI-TOF measurement was performed on Bruker Daltonics Ultraflex MALDI-TOF/TOF mass spectrometer equipped with a reflectron unit. Pulsed ion extraction: 30 ms; detector gain: 6.8X; electronic gain set at 100 mV/full scale; two ion sources at 25 and 20.9 kV. The acceleration voltage was set at 20 kV. For desorption of the components, a nitrogen laser beam (k = 337 nm) was focused on the template. The laser power level was adjusted to obtain high signal-to-noise ratios (22 %), while ensuring minimal fragmentation of the parent ions. Sample was diluted 1:10 in εeOH and then 0.5 δ were withdrawn and spotted onto εAδDI plate, let to air dry and covered with 1 δ of 2,5-dihydroxybenzoic acid (DHB) matrix solution (20 mg/mL in acetonitrile/water 7:3 + 0.1% TFA). After crystallization at ambient conditions, positive ion spectra were acquired in the reflectron mode, giving signal for [M+Na] ion as the most abundant. Every mass spectrum represents the average of about 120 single laser shoots. Calibration for high resolution experiments was performed on two standards: sphingomyelins (SMs, m/z of the most intense species: 731.6066674, corresponding to [M+H] of SM 18:0) and diacylglyceryltrimethylhomoserine (DGTS, m/z 712.6090962, corresponding to [M+H] of DGTS 36:0. 4.2. Data of isolinariin D (5) Yellow, amorphous powderν [α]20D -105.7 (c 0.26 g/mδ, εeOH)ν UV (εeOH) max(log )μ 204 (5100), 277 (2910), 328 (3720) nm; FT-IR (KBr) max 3407 (vs), 2928 (w), 1744 (m), 1658 (s), 1605 (vs), 1462 (s), 1363 (s), 1251 (vs), 1067 (vs) 837 (m) cm-1; 1H NMR (CD3OD, 400 MHz,): δ = 7.97 (2H, d, J = 8.7 Hz, H-2' H-6'), 7.10 (2H, d, J = 8.9 Hz, H-3' H-5'), 6.92 (1H, s, H-8), 6.70 (1H, s, H-3), 5.19 (1H, d, J = 7.6 Hz, H-1''), 5.17 (1H, dd, J = 3.5, 1.6 Hz, H-4'''), 5.00 (1H, dd, J = 10, 3.4 Hz, H-3'''), 4.71 (1H, br s, J = 1.2 Hz, H-1'''), 3.90 (6H, br s, 2 × OCH3), 1.93 (3H, s, COCH3), 1.89 (3H, s, COCH3), 1.18 (1H, d, J = 6.2 Hz, H-6'''), 13C NMR (CD3OD, 100 MHz,): δ = 183.08 (C, C-4); 170.48 (C, CH3CO), 169.84 (C, CH3CO), 165.3 (C, C-2), 163.03 (C, C-4'), 156.15 (C, C7), 152.91 (C, C-9), 152.74 (C, C-5), 132.67 (C, C-6), 128.01 (CH, C-2', C-6'), 122.90 (C, C-1'), 114.02 (CH, C-3', C5'), 106.10 (C, C-10), 102.61 (CH, C-3), 99.76 (CH, C-1''), 97.45 (CH, C-1'''), 94.03 (CH, C-8), 76.19 (CH, C-3''), 75.72 (CH, C-5''), 72.94 (CH, C-2''), 71.45 (CH, C-3'''), 70.02 (CH, C-5'''), 69.24 (CH, C-4'''), 68.38 (CH, C-4''), 68.01 (CH, C-2'''), 65.43 (CH2O, C-6''), 60.14 (6OCH3), 54.67 (4'-OCH3), 20.34 and 19.70 (2 × CH3CO), 16.37 (CH3, C-6'''); H-2', H-6' (7.97 ppm) : C-2, C-4'; H-3', H-5' (7.1 ppm) : C-4', C-1'; H-8 (6.92 ppm) : C-4, C-7, C-9, C-6, C-10; H-3 (6.70 ppm) : C-4, C-2, C-1', C-10; H-1'' (5.19 ppm) : C-7; H-4''' (5.17 ppm) : CH3CO; H-3''' (5.17 ppm) : CH3CO, C-4'''; H-1''' (4.71) : C-3''', C-2'''; 2 × OCH3 (3.9 ppm) : C-4', C-6; H-6''' (1.18 ppm) : C-4''', C-3'''. ESI(+)-MS m/z 729.1 [M+Na] , ESI(+)MS/MS (729): m/z 669, 609, 499, 415, 337; ESI(-)MS m/z 705.3 [M-H] ; ESI(-)MS/MS (705) m/z 313, 298, 283; HR-MALDI-TOFMS [M+Na]⁺ m/z 729.2040 ± 0.005 (calcd for C33H38O17Na, 729.20012). 5. Antioxidant activity of L. reflexa extracts 5.1. DPPH radical scavenging activity The ability to scavenge the stable free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH) was determined based on the method of Magalhaes et al., (2006) with minor modifications. A DPPH solution in methanol (1 mL, 0.2 mM) was mixed with L. reflexa extracts in methanol (1 mL). The reaction mixture was vortexed thoroughly and left in the dark at room temperature for 30 min. A control sample containing the same volume of solvent in place of extracts was used to measure the maximum DPPH absorbance. The absorbance of the mixture was measured at 517 nm. Ascorbic acid was used as reference. Results were expressed as percentage of inhibition of the DPPH radical according to the following equation: 5.2. Reducing power assay ⁻ ⁻ ⁻ A control A sample control Different amounts of extracts were suspended in methanol and mixed with 1 mL of 0.2 M phosphate buffer (pH 6.6) and 1 mL of 1% K3Fe(CN)6. The mixture was incubated at 50°C for 20 min, 1 mL of 10% thiazolidine-4-carboxylic acid (TCA) was added to the mixture and centrifuged at 3000 rpm for 10 min. 1.5 mL of the solution upper layer was mixed with 1.5 mL distilled water and 0.3 mL of 0.1% FeCl3, the absorbance was measured at 700 nm. The increase of reaction mixture absorbance indicates reducing power activity (Oyaizu 1986). Quercetin and 6-hydroxy2,5,7,8-tetramethylchroman-2-carboxylic acid (trolox) were used as references. 5.3. β-Carotene/linoleic acid bleaching assay A stock solution was prepared using 0.5 mg of β-carotene in 1 mδ chloroform, 25 δ of linoleic acid and 200 mg of Tween 40. The chloroform was evaporated in vacuo and then 100 mL of oxygenated distilled water were added to the residue. Each sample solution (1 g/δ, 350 δ) prepared in DMSO were added to 2.5 mL of the above mixture in test tubes which were incubated at 50°C for 24 h. BHA and trolox were used as a positive control. The absorbance values were measured at 490 nm each three hours during 24 h (Miraliakbari and Shahidi, 2008). Antioxidant activity (inhibition percentage) of the samples was calculated using the following equation: Absorbance of 6. In vivo antidiabetic activity of L. reflexa extracts carotene at 24 h ⁻ carotene For experimental design, rats were randomly divided into nine groups (Table S1), each group consisting of five animals. The drug preparations were given orally to rats of each group twice daily for 15 days. 6.1. Animals Wistar strain albino rats (180–200 g) were purchased from Algerian Pasteur Institute. They were acclimatized to animal house conditions. The animals were fed ad libitum with pellet diet (ONAB: national office of animal feed, Bejaia, Algeria) and water. They were also kept and maintained under laboratory conditions of temperature and light (24 ± 1°C and 12 h light/dark cycle), respectively. The experiments were performed according to the guidelines of the guide for care and use of laboratory animals (Committee for the Update of the Guide for the Care and Use of Laboratory Animals, 2010). 6.2. Diabetes induction Rats were injected intra-peritoneally with a single injection of alloxan monohydrate dissolved in sterile normal saline at a dose of 150 mg/kg body weight. After 3 days, hyperglycemia was confirmed by using the Accu-Chek Active Glucometer (Roche Diagnostics, Germany). Only rats with fasting blood glucose levels greater than 250 mg/dL were used in this study (Matteucci and Giampietro, 2008). 6.3. Assessment of blood glucose level and weight before sacrifice Blood was collected from the tail vein before extract administration every two days, for determination of blood glucose level using Accu-Chek Active Glucometer. The results were expressed in terms of mg/dL blood. The body weight was measured every five days. 6.4. Biochemical assay At the end of the experiment (15 days), rats were fasted overnight and sacrificed. The blood was collected and the serum was separated by centrifugation (10.000 g/15 min) and used for biochemical analysis. The concentration of glucose was measured according to the enzymatic colorimetric method, after incubation with glucose oxidase. Total lipids was determined by phosphovanilline reagent (Boudjelal et al., 2012). The triacylglycerol concentration was determined by the enzymatic hydrolysis and release of fatty acids and glycerol, with the final production of H2O2. Total cholesterol levels were determined enzymatically by cholesterol ester/oxidase (Moura 1982). The fractions of high density lipoprotein (HDL) were measured through precipitation by phosphotungstic acid. Low density lipoprotein (LDL) was calculated using the Friedewald equation. The serum urea level was determined in the presence of urease, resulting in CO2 and ammonia production. The addition of phenol-hypochloride led to an indophenol-blue complex with an absorbance at 340 nm. Creatinine was assayed by reaction with picric acid in alkaline buffer to form a yellow-orange complex, with the color intensity, determined at 492 nm (Tietz 1990). 6.5. Statistical analysis All assays of in vitro antioxidant activity were carried in triplicates and results expressed as means ± standard deviation. IC50-value ( g extract/mδ) is the effective concentration which proves 50% of activity, was calculated for each assay. Statistical comparisons were done with Student’s test. Differences were considered to be significant at P < 0.05. All data of in vivo antidiabetic activity are expressed as mean ± SD. Within group comparisons were performed by analysis of variance using ANOVA test (GraphPad Prism 5). Significant difference between control and experimental groups were assessed by Tukey’s testν P-values of less than 0.05 were considered to be significant. 7. References Boudjelal A, Henchiri C, Siracusa L, Sari M, Giuseppe R. 2012. Compositional analysis and in vivo anti-diabetic activity of wild Algerian Marrubium vulgare L. infusion. Fitoterapia 83:286–292 Magalhaes LM, Segundo MA, Reis S, Lima JLFC. 2006. Automatic method for determination of total antioxidant capacity using 2,2-diphenyl-1-picrylhyrazyl assay. Anal Chim Acta 558:310– 318 Matteucci E, Giampietro O. 2008. Proposal open for discussion: defining agreed diagnostic procedures in experimental diabetes research. J Ethnopharmacol. 115:163–172 Miraliakbari H, Shahidi F. 2008. Antioxidant activity of minor components of tree nut oils. Food Chem. 111:421–427 Moura RA. 1982. Técnicas de laboratório (2nd edn), Atheneu Editora, São Paulo. Oyaizu M. 1986. Studies on products of browning reaction prepared from glucosamine. Jpn J Nutr. 44:307–315 Tietz NW. 1990. Clinical guide to laboratory tests (2nd edn), W.B. Saunders Company: Philadelphia, 554–556. Table S1. Group division of rats used for in vivo anti-diabetic experiments of ultrasound- assisted extract (UAE) and aqueous extract (AQE) from L. reflexa Groups I II III IV-VI VII-IX Treatment Normal Control (rats treated with distilled water) Diabetic Control (rats treated with distilled water) Diabetic rats Treated with Glibenclamide (5 mg/kg b.w.) Diabetic rats Treated with L. reflexa aqueous extract (100, 200 and 300 mg/kg b.w., respectively) Diabetic rats treated with L. reflexa UAE (100, 200 and 300 mg/kg b.w., respectively) Label NC DC DTG DT-AQE DT-UAE Table S2. In vivo Anti-diabetic evaluation: Variation in blood glucose level overnight fast, after oral administration of UAE and AQE from L. reflexa for 15 days. Label groups Dose (mg/kg) NC - Blood glucose (mean±SD) (mg/dL) Number of days 1 3 89.20 ± 1.02 5 88.00 ± 1.58 a 92.20 ± 1.79 a 445.80 ± 4.02 9 85.20 ± 1.92 a 454.80 ± 3.96 11 102.20 ± 1.48 a 473.80 ± 2.68 13 89.00 ± 1.58 a 472.80 ± 4.44 86.00 ± 2.00 a 585.40 ± 3.58 <±1 a DC - 350.80 ± 3.77 DTG 5 375.40 ± 2.70b 308.20 ± 2.39b 279.40 ± 2.41 b 183.80 ± 2.59 b 169.20 ± 1.30 b 168.60 ± 1.14 b 137.80 ± 1.92 b (-) 63.29 DT- AQE 100 276.60 ± 2.70b 266.80 ± 2.50 b 249.00 ± 1.58 b 238.40 ± 1.34 b 210.40 ± 2.70 b 190.20 ± 1.70 b 183.60 ± 1.50 b (-) 33.62 200 348.40 ± 1.24b 291.40 ± 1.67 b 250.80 ± 1.64 b 236.20 ± 1.30 b 220.20 ± 1.33 b 160.40 ± 1.82 b 128.00 ± 2.17 b (-) 63.26 300 370.00 ± 1.08 b 216.00 ± 2.55 b 156.60 ± 1.52 b 130.00 ± 1.85 b 121.20 ± 2.32 b 115.80 ± 2.05 b 111.40 ± 2.51 b (-) 69.89 100 380.20 ± 2.97 b 361.60 ± 2.41 b 256.40 ±2.39 b 197.80 ± 1.48 b 167.00 ± 1.55 b 138.40 ± 2.30 b 119.00 ± 1.52 b (-) 69.70 200 370.00 ± 1.24 b 289.80 ± 1.36 b 270.40 ± 1.25 b 180.40 ± 1.05 b 158.40 ± 1.48 b 132.00 ± 1.35 b 107.80 ± 2.77 b (-) 70.86 300 381.60 ± 1.44 b 204.00 ± 2.01 b 187.20 ±1.70 b (-) 72.09 DT-UAE 401.60 ± 3.36 7 % decrease glycaemia 160.50 ± 1.80 b 124.70 ± 2.02 b 107.30 ± 1.90 b 106.50 ± 1.40 b a p < 0,001 when normal control compared to diabetic control. b p < 0,001 when treated diabetic groups compared to diabetic control. (+) 66.87 Table S3. Effect of L. reflexa UAE and AQE dose on rat body weight after induction of diabetes. For abbreviaitons, see Table S1. Groups NC DC DTG DT-AQE DT-UAE Dose (mg/kg) 5 100 200 300 100 200 300 1 130.90 ± 2.63 152.12 ± 2.96 148.83 ± 1.32 145.20 ± 1.73 141.20 ± 2.84 136.20 ± 1.90 137.70 ± 1.65 154.04 ± 2.45 158.50 ± 1.50 Body weight (g) at different days (1÷15) 5 10 135.64 ± 2.40 138.24 ± 1.70 147.28 ± 2.33 134.92 ± 2.56 140.40 ± 1.63 143.06 ± 2.05 138.58 ± 1.14 142.26 ± 2.51 136.84 ± 2.10 137.60 ± 1.90 132.73 ± 2.23 134.75 ± 2.20 135.33 ± 1.89 135.90 ± 1.42 151.80 ± 1.89 153.10 ± 1.79 155.25 ± 2.24 157.80 ± 2.84 15 146.34 ± 1.85 130.48 ± 3.20 147.08 ± 1.80 148.29 ± 1.33 140.80 ± 2.30 137.30 ± 1.99 139.20 ± 2.75 158.50 ± 1.10 162.57 ± 1.82 Table S4. In vivo Anti-diabetic evaluation: Variation in blood glucose and biochemical parameters after sacrifice of rats after different treatments.. For abbreviations, see Table S1. Label groups parameters NC DC DTG DT-AQE 100 DT-AQE 200 DT-AQE 300 DT-UAE 100 DT-UAE 200 DT-UAE 300 Glucose (mg/dL) 100.00 ± 1.53 441.00±4.22 140.20±3.29 139.75±2.22 130.60±1.20 124.00±0.94 118.40±1.00 115.30±0.66 110.00±0.20 Cholesterol (g/L) 0.56 ± 0.10 0.67±0.25 0.55±0.14 0.46±0.10 0.39±0.16 0.40±0.12 0.60±0.08 0.48±0.12 0.60±0.42 Triglycerides (g/L) 0.43±0.10 0.57±0.25 0.46±0.14 0.33±0.22 0.33±0.17 0.40±0.18 0.47±0.03 0.56±0.01 0.50±0.02 Total lipids (g/L) 2.36±0.46 8.83±1.75 3.25±0.53 1.76±1.12 1.81±1.34 1.78±0.5 1.50±0.49 1.35±0.28 1.05±0.14 LDL (g/L) 0.50±0.10 0.61±0.12 0.52±0.16 0.44±0.04 0.33±0.13 0.39±0.12 0.47±0.08 0.33±0.20 0.32±0.06 HDL (g/L) 0.30±0.12 0.45±0.07 0.43±0.03 0.35±0.10 0.38±0.16 0.33±0.14 0.45±0.12 0.42±0.20 0.32±0.05 Protéines totales (mg/L) 60.40±0.54 86.00±0.71 82.00±0.16 77.00±0.18 73.00±0.25 69.67±0.01 67.75±0.41 66.8±0.14 65.20±0.38 TGO (UI/L) 136.40±0.24 158.00±0.31 147.00±0.55 120.00±0.07 140.40±0.13 110.67±0.09 121.20±0.48 121.60±0.20 120.00±0.22 TGP (UI/L) 42.00±0.55 47.33±0.45 44.33±0.66 41.50±0.03 46.40±0.78 38.33±0.23 33.80±0.08 32.40±0.14 35.60±0.09 Urea (g/L) 0.41±0.07 0.85±0.09 0.84±0.10 0.47±0.09 0.54±0.55 0.48±0.25 0.60±0.39 0.58±0.16 0.45±0.07 Creatinine (mg/L) 6.14±0.10 8.40±0.19 7.83±0.27 7.15±0.19 7.20±0.50 7.53±0.08 7.56±0.12 7.52±0.17 6.26±0.13 Values are means ± SD (n = 5 rats) LDL: Low-density lipoproteins; HDL: High-density lipoproteins; TGO: Transaminase glutamo oxaloacetic acid; TGP: Transaminase glutamo pyruvic Figure S1. DAD-HPLC chromatograms from online LC-MS analysis of UAE (A) and HAE (B) of the aerial parts of L. reflexa Desf. Figure S2. MS spectra in negative ion mode from LC-MS analysis of UAE of the aerial parts of L. reflexa Desf, associated to the chromatogram (A) in Fig.S1. 100 IC50 : 14,39 ±0,30 UAE Inhibition (%) 80 IC50 :214,22 ±4,69 HAE Ascorbic acid 60 IC50 :856 ±20,04 40 20 0 0 200 400 600 800 1000 1200 Concentration µg/mL Figure S3. Antioxidant activity of UAE and HAE from L. reflexa evaluated as DPPH radical scavenging property using ascorbic acid as a standard. Each value represents a mean ± SD (n=3), P<0.05 Figure S4. Antioxidant activity by K3Fe(CN)6 test in reducing power of UAE and HAE of L. reflexa compared to quercetin and trolox as standards, evaluated as absorbance at 700 nm (a), and percentage inhibition as a function of concentration (b). Each value represents a mean ± SD (n=3), P<0.05 Figure S5. Antioxidant activity by β-carotene bleaching effect of UAE, HAE from L. reflexa and standards (a) and percentage inhibition (b). Each value represents a mean ± SD (n=3), P<0.05