molecules

Article Quantitative 1H-NMR Spectroscopy for Profiling Primary Metabolites in Mulberry Leaves

Qianqian Liang 1,2, Qiuying Wang 1, Yuan Wang 3, Ya-nan Wang 4, Jia Hao 1 and Miaomiao Jiang 1,*

1 Tianjin State Key Laboratory of Modern Chinese Medicine, Tianjin University of Traditional Chinese Medicine, Tianjin 300193, China; [email protected] (Q.L.); [email protected] (Q.W.); [email protected] (J.H.) 2 Guangdong Province Key Laboratory of Pharmacodynamic Constituents of TCM and New Drugs Research, Guangzhou 510632, China 3 Tianjin Zhongxin Innova Laboratories, Tianjin 300457, China; [email protected] 4 Institute of Materia Medica, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing 100730, China; [email protected] * Correspondence: [email protected]; Tel.: +86-022-5959-6163

Received: 11 January 2018; Accepted: 5 February 2018; Published: 2 March 2018

Abstract: The primary metabolites in aqueous extract of mulberry (Morus alba L.) leaves were characterized by using proton nuclear magnetic resonance (1H-NMR) spectroscopy. With the convenience of resonance assignment, GABA together with the other 10 primary metabolites was simultaneously identified and quantified in one 1H-NMR spectrum. In this study, external calibration curves for metabolites were employed to calculate the concentrations of interests. The proposed quantitative approach was demonstrated with good linearity (r2 ranged in the interval of 0.9965–0.9999), precision, repeatability, stability (RSD values in the ranges of 0.35–4.89%, 0.77–7.13% and 0.28–2.33%, respectively) and accuracy (recovery rates from 89.2% to 118.5%). The established 1H-NMR method was then successfully applied to quantify 11 primary metabolites in mulberry leaves from different geographical regions within a rapid analysis time and a simple sample preparation procedure.

Keywords: quantitative 1H-NMR; Morus alba; amino acids; saccharides; organic acids

1. Introduction Mulberry leaves are the main food of silkworm, with a long history of industrial use for more than 5000 years [1], and they are also used as functional foods for humans concerning their nutritive and medicinal values. For instance, mulberry tea has been developed to be a popular health food loved by consumers on account of its effects on modulating dyslipidemia, preventing diabetes, and maintaining health [2,3]. The extracts of mulberry leaves are reported to display a wide range of significant biopharmaceutical activities, including antidiabetic, antibacterial, anticancer, cardiovascular, hypolipidemic, antioxidant, antiatherogenic, and anti-inflammatory effects [4]. Phytochemical studies reveal that γ-aminobutyric acid (GABA) is one of the principal components isolated from the extracts of mulberry leaves [5]. GABA is well known for its essential role in the nervous system and brain development [6], and may be associated with several of the physiological activities of mulberry leaves, such as antioxidant [7], antihyperglycaemic [8], antihypertensive [9], and anti-inflammatory properties [10]. Like many other aliphatic amino acids, GABA lacks a suitable chromophore for direct ultraviolet detection, and hence, sample derivatization is required before injection to liquid chromatography [11]. However, the pre-column derivatization process is time-consuming, can add impurities into samples, and makes the analysis more complex.

Molecules 2018, 23, 554; doi:10.3390/molecules23030554 www.mdpi.com/journal/molecules Molecules 2018, 23, 554 2 of 10

As a result of the improvements in instrumentation and technology, proton nuclear magnetic resonance (1H-NMR) has been exploited as a universal and quantitative tool applied in analysis of complex natural samples [12]. An unbiased view in the composition of a given mixture can be offered by 1H-NMR spectroscopy with simultaneous access to both qualitative and quantitative information [13]. Quantitative 1H-NMR measurement has been proved robust [14] with its error reported to be less than 2.0%, which is an acceptable limit for precise, accurate quantification [15]. In addition, 1H-NMR methods provide easy sample preparation, non-destructive analysis, and relatively short analysis times. Compared to mass spectrometry, 1H-NMR has from a relatively low detection sensitivity, which can be partly ameliorated by the use of high-field magnets and CryoProbe technologies [12]. With its high capacity to characterize amino acids, organic acids, and carbohydrates without obvious UV chromophores, 1H-NMR has become a common quantification tool in many applications for profiling and determining primary metabolites in foods [16,17], plants or herbal remedies [18,19], and biofluid samples [20–22]. To the best of our knowledge, the use of a 1H-NMR method for chemical profiling of mulberry leaves has not been explored in the literature. In this paper, a rapid, reliable and facile approach based on 1H-NMR spectroscopy was developed to identify and quantify GABA together with other primary metabolites in aqueous extracts of mulberry leaves, as well as to provide an alternative method for quality control and evaluation of mulberry leaves in medicinal prescriptions, health care products and functional foods.

2. Results and Discussion

2.1. Proton Signal Assignments and Chemical Identification Representative 1H-NMR spectra of the aqueous extract of standard material (SM) are shown in Figure1. The resonance signals were assigned to 11 metabolites according to the published results and further confirmed by a series of 2D NMR spectra. The dominant primary metabolites present in the mulberry leaf extracts included four amino acids (alanine, proline, asparagine and GABA), three organic acids (acetic acid, succinic acid and fumaric acid), two carbohydrates (glucose and sucrose), and two alkaloids (choline and trigonelline). Their chemical structures are shown in Figure2 and detailed signal assignments are listed in Table1.

Figure 1. Typical 1H-NMR spectra of mulberry leaf extracts. (A) Full 1H-NMR spectrum from δ 0.0 to

δ 9.5 (TSP-d4 as an internal standard with chemical shift at δ 0.0); (B) Enlarged spectrum from δ 1.0 to δ 3.1; (C) Enlarged spectrum from δ 3.2 to δ 5.5; (D) Enlarged spectrum from δ 6.4 to δ 9.3. Peaks of 11 metabolites: 1, acetic acid; 2, alanine; 3, asparagine; 4, choline; 5, fumaric acid; 6, GABA; 7, glucose; 8, proline; 9, succinic acid; 10, sucrose; 11, trigonelline. Molecules 2018, 23, 554 3 of 10

Molecules 2018, 23, x 3 of 10

Figure 2. Chemical structures of 11 metabolites from mulberry leaf extracts. Figure 2. Chemical structures of 11 metabolites from mulberry leaf extracts.

Table 1. Proton NMR spectroscopic data for 11 metabolites in D2O.

No. MetabolitesTable δ 1. ProtonH (Multiplicity NMR spectroscopic a) data for 11 metabolitesAssignment in D 2O. 1 Acetic acid 1.91 (s) CH3 No. Metabolites2 Alanine δ H (Multiplicity3.78 (q), 1.48 (d) a)α‐CH, Assignment β‐CH3 3 Asparagine 2.86 (dd), 2.96 (dd), 4.02 (dd) β‐CH2, α‐CH 1 Acetic4 acidCholine 1.91 (s)4.11 (m), 3.52 (dd), 3.21 (s) 2‐CH CH2, 33‐CH2, 4‐N(CH3)3 5 Fumaric acid 6.50 (s) CH 2 Alanine 3.78 (q), 1.48 (d) α-CH, β-CH3 6 GABA 3.00 (t), 2.30 (t), 1.90 (m) γ‐CH2, α‐CH2, β‐CH2 5.24 (d), 4.65 (d), 3.89 (m), 3.83 (m), 3.75 (m), 3.69 (m), 2‐CH,β 11‐CH,α 6‐CH, 4‐CH, 3 Asparagine7 Glucose 2.86 (dd), 2.96 (dd), 4.02 (dd) -CH2, -CH 3.52 (m), 3.48 (m), 3.40 (m), 3.25 (dd) 3‐CH, 5‐CH 4 Choline8 Proline 4.11 (m),4.11 (m), 3.52 3.41 (dd), (m), 3.21 3.28 (m), (s) 2.35 (m), 2.04 (m) α‐CH, 2-CH δ‐CH2,2, 3-CH β‐CH22, ,γ‐ 4-N(CHCH2 3)3 5 Fumaric9 acidSuccinic acid 6.50 (s)2.41 (s) CH2 CH 7‐CH, 3‐CH, 4‐CH, 17‐CH, 10 Sucrose 5.40 (d), 4.20 (d), 4.04 (t), 3.80 (m), 3.55 (m), 3.46 (t) 6 GABA 3.00 (t), 2.30 (t), 1.90 (m) 19‐CH,γ-CH 12‐CH,2, α -10CH‐CH2, β-CH2 11 Trigonelline 9.13 (s), 8.84 (t), 8.09 (t), 4.44 (s) 1‐CH, 3‐CH, 5‐CH, 4‐CH, 9‐CH3 5.24 (d), 4.65 (d), 3.89 (m), 3.83 (m), 3.75 (m), 2-CH, 11-CH, 6-CH, 4-CH, 3-CH, 7 Glucose a Abbreviations, s = singlet, d = doublet, dd = doublet‐doublets, m = multiplet, t = triplet. 3.69 (m), 3.52 (m), 3.48 (m), 3.40 (m), 3.25 (dd) 5-CH

1 8 Proline2.2. Quantitative 4.11 H‐NMR (m), 3.41Analysis (m), and 3.28 Method (m), Validation 2.35 (m), 2.04 (m) α-CH, δ-CH2, β-CH2, γ-CH2

1 9 Succinic acidOne inherent 2.41 advantage (s) of H‐NMR is that the concentrations of CHsubstances2 are directly proportional to the signal areas without any need for response factor correction [12,13]. Routinely, 7-CH, 3-CH, 4-CH, 17-CH, 10 Sucrosethe repetition time 5.40 should (d), 4.20 be (d),five 4.04times (t), the 3.80 longest (m), longitudinal 3.55 (m), 3.46 relaxation (t) time of resonance T1 to measure 99% of the equilibrium magnetization [23]. 19-CH, 12-CH, 10-CH 11 TrigonellineA long NMR 9.13 measurement (s), 8.84 (t), time 8.09 is (t), required 4.44 (s) to completely relax all analyzed1-CH, protons. 3-CH, To 5-CH, reduce 4-CH, 9-CH3 the total time spent on multiple scans, an inadequate time of relaxation delay (D1 = 4 s) was set to a Abbreviations, s = singlet, d = doublet, dd = doublet-doublets, m = multiplet, t = triplet. acquire spectra in our work. We then employed external calibration for each metabolite to calculate their contents in extracts to avoid a potential error source of every step of sample treatment and 1 2.2. QuantitativetestingH-NMR [24]. The Analysis1H‐NMR method and Method was subsequently Validation validated with respect to linearity, precision, repeatability, stability and accuracy (Tables 2 and 3): One inherent advantage of 1H-NMR is that the concentrations of substances are directly proportional to the signalTable areas 2. The without linear regression any equations, need for LOD response and LOQ for factor 11 metabolites correction (n = 3). [12,13]. Routinely, the No. Metabolites Regression Equation Correlation Coefficients (r2) LOD (mM) LOQ (mM) repetition time should be five times the longest longitudinal relaxation time of resonance T1 to measure 1 Acetic acid y = 0.3087x + 0.1805 0.9993 0.040 0.132 99% of the equilibrium2 Alanine magnetization y = 0.5976x [ +23 0.0173]. 0.9998 0.012 0.038 A long NMR3 measurementAsparagine y = time 1.2754x is + required0.0969 to completely0.9996 relax all0.025 analyzed 0.083 protons. To reduce 4 Choline y = 0.0849x + 0.0916 0.9965 0.001 0.002 the total time spent5 Fumaric on multiple acid y = scans, 0.5585x + an 0.0085 inadequate 0.9999 time of relaxation0.003 delay (D0.0091 = 4 s) was set to acquire spectra in our work. We then employed external calibration for each metabolite to calculate their contents in extracts to avoid a potential error source of every step of sample treatment and testing [24]. The 1H-NMR method was subsequently validated with respect to linearity, precision, repeatability, stability and accuracy (Tables2 and3):

Table 2. The linear regression equations, LOD and LOQ for 11 metabolites (n = 3).

No. Metabolites Regression Equation Correlation Coefficients (r2) LOD (mM) LOQ (mM) 1 Acetic acid y = 0.3087x + 0.1805 0.9993 0.040 0.132 2 Alanine y = 0.5976x + 0.0173 0.9998 0.012 0.038 3 Asparagine y = 1.2754x + 0.0969 0.9996 0.025 0.083 4 Choline y = 0.0849x + 0.0916 0.9965 0.001 0.002 Molecules 2018, 23, 554 4 of 10

Table 2. Cont.

No. Metabolites Regression Equation Correlation Coefficients (r2) LOD (mM) LOQ (mM) 5 Fumaric acid y = 0.5585x + 0.0085 0.9999 0.003 0.009 6 GABA y = 0.7050x + 0.0319 0.9996 0.016 0.053 7 Glucose y = 2.7552x + 0.5830 0.9995 0.075 0.248 8 Proline y = 1.4021x + 0.1328 0.9982 0.036 0.121 9 Succinic acid y = 0.1945x + 0.0316 0.9993 0.004 0.013 10 Sucrose y = 3.3043x + 0.0558 0.9981 0.047 0.156 11 Trigonelline y = 0.0819x + 0.0170 0.9993 0.003 0.009

Table 3. The methodological investigation results of precision, stability, repeatability and recovery (n = 6).

Precision Stability Repeatability Recovery Metabolites RSD (%) RSD (%) RSD (%) (%) RSD (%) Acetic acid 0.35 0.28 1.32 106.1 5.46 Alanine 2.39 0.94 5.15 91.6 7.01 Asparagine 4.89 2.33 2.65 118.5 5.25 Choline 0.69 0.24 1.09 109.8 2.26 Fumaric acid 1.15 0.32 0.77 90.4 4.47 GABA 4.21 1.74 2.48 95.6 2.52 Glucose 0.68 0.75 3.82 116.2 1.53 Proline 0.90 0.79 2.00 90.7 2.83 Succinic acid 0.75 0.46 0.85 89.2 1.70 Sucrose 2.40 1.81 7.13 99.8 4.34 Trigonelline 1.27 0.30 0.80 104.0 3.07

Linearity. Six solutions of 11 metabolites in different concentrations were prepared and analyzed in triplicate. The calibration curve was constructed by plotting the ratio between the peak areas of metabolite and internal standard (x) versus the given metabolite concentration (y). The linear regression equation and correlation coefficient of each metabolite were shown in Table2. The correlation coefficients of 11 metabolites were in the range of 0.9965–0.9999, indicating good linearity of the established method. In addition, the limit of detection (LOD = 3.3σ/S) and limit of quantification (LOQ = 10σ/S) were calculated by the standard deviation of y-intercept of the regression line (σ) and the slope of the calibration curve (S) [25]. The values of LOD and LOQ for 11 metabolites were in the range of 0.001–0.075 mM and 0.002–0.248 mM, respectively. Precision. The intraday precision was determined by analyzing six replicates of the same sample in one day. The relative standard deviation (RSD) values of intraday precision for the contents of 11 metabolites ranged from 0.35% to 4.89%. Stability. The stability was evaluated by analyzing one sample for 12 h at an interval of every 2 h. The RSD values for the contents of eleven metabolites were in the range of 0.24–2.33%, which showed the analytes were stable during the tested period at ambient temperature. Repeatability. The repeatability was assessed by analyzing six samples prepared from the same batch of SM. The RSD values of the contents of 11 metabolites ranged from 0.77% to 7.13%, indicating high repeatability. Accuracy. The accuracy of the NMR method was determined by conventional recovery tests. Six replicates of mixed standard solutions were spiked into 11 known amounts of samples with the same concentration that had previously been analyzed. The spiked samples were then quantified by the established 1H-NMR methods. The recovery rate was calculated by 100% × (found amount—original amount)/spiked amount. The results revealed that the average recoveries ranged from 89.2% to 118.5% with RSD values of less than 7.01% for all the 11 metabolites. Molecules 2018, 23, 554 5 of 10

Using the developed 1H-NMR method, 11 metabolites were determined simultaneously in 54 samples of mulberry leaves from 18 origins in China (Table S1 and S2), including standard material (SM) obtained from the National Institute for Food and Drug Control (Beijing, China), Pingyao (PY), Anyang (AY), Zhoukou (ZK), Suzhou (SZ), Bozhou (BZ), Guangzhou (GZ), Jinan (JN), Qingdao (QD), Linyi (LY), Heze (HZ), Zibo (ZB), Dezhou (DZ), Anguo (AG), Xingtai (XT), Chengde (CD), Daan (DA), and Fuzhou (FZ). As shown in Figure3, all 11 metabolites were detected in the samples, except for asparagine that was absent in AY samples. Alanine in BZ, PY, XT and AY samples showed lower levels (0.13–0.39 mg/g) than those of SM (0.48 mg/g) and the other original samples. Acetic acid levels in BZ, FZ, XT, ZK and AY samples ranged from 0.87 mg/g to 1.46 mg/g, less than that of SM samples Molecules 2018, 23, x 5 of 10 (1.48 mg/g), while acetic acid in the other origins were in range of 1.74–4.15 mg/g. Proline levels ranged fromexcept 3.48 for mg/g asparagine to 11.71 that mg/g,was absent and in onlyAY samples. those ofAlanine PY and in BZ, AY PY, were XT and lower AY thansamples the standard level (3.53 mg/g).showed lower Succinic levels acid(0.13–0.39 in XT, mg/g) CD, than DZ, those HZ, of SM QD (0.48 and mg/g) ZB samplesand the other showed original significantly samples. higher Acetic acid levels in BZ, FZ, XT, ZK and AY samples ranged from 0.87 mg/g to 1.46 mg/g, less than levels (0.63–1.03that of SM mg/g) samples than (1.48 that mg/g), of while SM (0.39 acetic mg/g),acid in the while other origins BZ, FZ, were PY in andrange AY of 1.74–4.15 showed mg/g. relatively low levels. A variationProline levels of ranged asparagine from 3.48 from mg/g 0 to mg/g 11.71 mg/g, to 1.30 and mg/gonly those was of PY observed and AY were in AG,lower BZ, than AY, PY, XT, CD and LYthe samples, standard which level (3.53 was mg/g). much Succinic lower thanacid in those XT, CD, of SM DZ, and HZ, theQD otherand ZB original samples samples.showed GABA in significantly higher levels (0.63–1.03 mg/g) than that of SM (0.39 mg/g), while BZ, FZ, PY and AY BZ, PY, XT,showed AY, QD relatively and AG low samples levels. A displayedvariation of asparagine lower levels from (0.09–0.760 mg/g to 1.30 mg/g) mg/g was than observed that of in SM samples (0.80 mg/g),AG, while BZ, AY, GABA PY, XT, in CD the and other LY samples, origins which ranged was frommuch 0.91–2.30lower than those mg/g. of SM Choline and the levels other from most origins wereoriginal higher samples. than GABA that in of BZ, SM PY, (0.97 XT, AY, mg/g) QD and except AG samples the origins displayed of lower BZ, PY,levels XT, (0.09–0.76 AY, ZK and AG. Glucose andmg/g) sucrose than that varied of SM samples in wide (0.80 ranges mg/g), of while 3.41–64.23 GABA in the mg/g other and origins 0.72–65.41 ranged from mg/g, 0.91–2.30 respectively. mg/g. Choline levels from most origins were higher than that of SM (0.97 mg/g) except the origins of Fumaric acidBZ, wasPY, XT, detected AY, ZK and in lowerAG. Glucose levels and from sucrose FZ, varied AY, DAin wide and ranges SZ originsof 3.41–64.23 (0.05–0.19 mg/g and mg/g) than the standard0.72–65.41 level (5.43mg/g, respectively. mg/g). Trigonelline Fumaric acid variedwas detected in a in range lower oflevels 0.06–0.09 from FZ, mg/g, AY, DA whichand SZ paralleled to that of SMorigins (0.08 (0.05–0.19 mg/g). mg/g) These than data the indicatedstandard level that (5.43 the mg/g). contents Trigonelline of 10 primary varied in metabolitesa range of changed 0.06–0.09 mg/g, which paralleled to that of SM (0.08 mg/g). These data indicated that the contents of obviously with10 primary different metabolites geographical changed obviously origins. with different geographical origins.

Figure 3. The contents of 11 metabolites in mulberry leaves extracts from 17 origins. Figure 3. The contents of 11 metabolites in mulberry leaves extracts from 17 origins. 2.3. The Metabolic Variations of Mulberry Leaves from Different Origins

Molecules 2018, 23, 554 6 of 10

MoleculesMolecules 2018 2018, ,23 23, ,x x 66 of of 10 10 2.3. The Metabolic Variations of Mulberry Leaves from Different Origins PrincipalPrincipalPrincipal component componentcomponent analysis analysisanalysis (PCA) (PCA)(PCA) was waswas carried carriedcarried out out without withwith the quantitative thethe quantitativequantitative NMR NMRNMR data to datadata simulate toto thesimulatesimulate metabolic the the trends metabolic metabolic of all trends trends the samples of of all all the the obtained samples samples from obtained obtained different from from regions. different different Scores regions. regions. plot Scores Scores (Figure plot plot4A) (Figure (Figure showed that4A)4A) samples showed showed from that that thesamples samples same from from region the the clusteredsame same region region well, clustered clustered whereas well, well, the whereas whereas ones from the the ones LY,ones GZ, from from JN, LY, LY, HZ GZ, GZ, and JN, JN, SM HZ and SM origins were separable. On the direction of PC2, samples from HZ, LY and JN were originsHZ and were SM separable. origins were On separable. the direction On the of PC2,direction samples of PC2, from samples HZ, LY from and HZ, JN LY were and close JN were to the closeclose to to the the standard standard samples, samples, which which may may be be caused caused by by similar similar geographic geographic and and climate climate conditions conditions standard samples, which may be caused by similar geographic and climate conditions of the three ofof thethe threethree regions.regions. BiplotBiplot ofof PCAPCA (Figure(Figure 4B)4B) furtherfurther revealedrevealed thatthat glucoseglucose andand sucrosesucrose majormajor regions. Biplot of PCA (Figure4B) further revealed that glucose and sucrose major contributed to the contributedcontributed to to the the classification, classification, which which was was consistent consistent with with the the content content determination determination results. results. They They classification, which was consistent with the content determination results. They are important sources areare important important sources sources of of energy energy for for cellular cellular respiration respiration in in plants. plants. of energy for cellular respiration in plants.

Figure 4. Scores plot (A) and biplot (B) of PCA model established with the quantitative NMR data. FigureFigure 4. 4. Scores Scores plot plot ( A(A) )and and biplot biplot ( B(B) )of of PCA PCA model model established established with with the the quantitative quantitative NMR NMR data. data.

CorrelationCorrelationCorrelation analysis analysisanalysis showed showedshowed an obviously anan obviouslyobviously positive positivepositive relationship relationshiprelationship among amongamong asparagine, asparagine,asparagine, GABA, GABA,GABA, choline, alanine,choline,choline, and alanine, alanine, proline and and (Figure proline proline5 A), (Figure(Figure indicating 5A), 5A), indicatingindicating they had theythey a similar hadhad aa similar variationsimilar variationvariation tendency. tendency.tendency. It was ItIt further waswas confirmedfurtherfurther confirmed byconfirmed the results byby the ofthe pathway resultsresults ofof enrichment pathwaypathway enrichment enrichment analysis that analysis analysis asparagine, thatthat asparagine,asparagine, GABA, alanine GABA, GABA, and alanine alanine proline wereandand subjected proline proline were were to two subjected subjected pathways to to two two with pathways pathways high impact with with valueshigh high impact impact (Figure values values5B), including (Figure (Figure 5B), 5B), alanine, including including aspartate alanine, alanine, and glutamateaspartateaspartate metabolism and and glutamate glutamate (Figure metabolism metabolism5C) as well (Figure (Figure as arginine 5C) 5C) as as well andwell prolineas as arginine arginine metabolism and and proline proline (Figure metabolism metabolism5D). GABA (Figure (Figure as an intermediate5D).5D). GABAGABA in asas the anan two intermediateintermediate pathways, in andin thethe our twotwo results pathways,pathways, suggested andand thatourour inhibitingresultsresults suggestedsuggested transamination thatthat inhibitinginhibiting of GABA mighttransaminationtransamination increase its of yieldof GABA GABA in this might might plant. increase increase its its yield yield in in this this plant. plant.

FigureFigure 5. 5. Cont Cont.Cont. .

Molecules 2018, 23, 554 7 of 10 Molecules 2018, 23, x 7 of 10

FigureFigure 5. (5.A ()A Correlation) Correlation analysis analysis ofof 11 metabolites from from mulberry mulberry leaves leaves extracts; extracts; (B) Overview (B) Overview of of pathwaypathway analysisanalysis with 11 metabolites from from mulberry mulberry leaves leaves extracts; extracts; (C (C) )Alanine, Alanine, aspartate aspartate and and glutamateglutamate metabolism; metabolism; (D ()D Arginine) Arginine and and proline proline metabolism. metabolism.

3. Materials and Methods 3. Materials and Methods

3.1.3.1. Materials Materials and and Reagents Reagents Proline (≥99%), choline chloride (≥98%) and sodium acetate (≥98%) were purchased from the Proline (≥99%), choline chloride (≥98%) and sodium acetate (≥98%) were purchased from National Institute for Food and Drug Control (Beijing, China). D‐(+)‐Glucose (99.9%) was purchased the National Institute for Food and Drug Control (Beijing, China). D-(+)-Glucose (99.9%) was from Supelco (Bellefonte, PA, USA). Sucrose (≥99.5%) was purchased from Shanghai Aladdin purchased from Supelco (Bellefonte, PA, USA). Sucrose (≥99.5%) was purchased from Shanghai Biological Technology Co., Ltd. (Shanghai, China). L‐Asparagine (≥98%), succinic acid (analytical ≥ Aladdinstandard), Biological trigonelline Technology hydrochloride Co., Ltd. (analytical (Shanghai, standard), China). L‐alanine L-Asparagine (≥98%) and ( 98%),sodium succinic hydrogen acid (analyticalfumarate standard), (≥99%) were trigonelline purchased hydrochloride from Sigma (St. (analytical Louis, MO, standard), USA). γ‐ L-alanineAminobutyric (≥98%) acid and (GABA, sodium hydrogen≥99%) was fumarate purchased (≥99%) from were Shanghai purchased Yuanye from Biological Sigma Technology (St. Louis, Co., MO, Ltd. USA). (Shanghai,γ-Aminobutyric China). acidDeuterium (GABA, oxide≥99%) (D2 wasO, 99.9 purchased atom % D) from and sodium Shanghai 3‐trimethylsilyl Yuanye Biological [2,2,3,3‐d4 Technology] propionate (TSP Co.,‐d Ltd.4, (Shanghai,98 atom China).% D) were Deuterium from Cambridge oxide (D Isotope2O, 99.9 Laboratories atom % D) (Cambridge, and sodium FL, 3-trimethylsilyl USA). [2,2,3,3-d4] propionateThe (TSP- standardd4, 98 material atom % (SM) D) were of mulberry from Cambridge leaves (Morus Isotope alba LaboratoriesL) was obtained (Cambridge, from the National FL, USA). InstituteThe standard for Food material and Drug (SM) Control of mulberry (Beijing, leaves China). (Morus The other alba L)samples was obtained of mulberry from leaves the National were Institutecollected for from Food 17 and cities Drug in China Control (Table (Beijing, S3), including China). ThePingyao other (PY), samples Anyang of (AY), mulberry Zhoukou leaves (ZK), were collectedSuzhou from (SZ), 17 Bozhou cities (BZ), in China Guangzhou (Table S3),(GZ), including Jinan (JN), Pingyao Qingdao (PY), (QD), Anyang Linyi (LY), (AY), Heze Zhoukou (HZ), Zibo (ZK), Suzhou(ZB), (SZ),Dezhou Bozhou (DZ), Anguo (BZ), Guangzhou(AG), Xingtai (GZ),(XT), Chengde Jinan (JN), (CD), Qingdao Daan (DA), (QD), and Linyi Fuzhou (LY), (FZ). Heze (HZ), Zibo (ZB), Dezhou (DZ), Anguo (AG), Xingtai (XT), Chengde (CD), Daan (DA), and Fuzhou (FZ). 3.2. Sample Preparation 3.2. SampleThe Preparation mulberry leaves were dried at 80 °C for 4 h and ground to a fine powder. Six hundred microlitersThe mulberry of D2O leaves containing were 0.1177 dried mM at 80 TSP◦C‐d for4 were 4 h added and ground into 24 mg to aof fine leaf powder.powder as Six extraction hundred solvent. D2O was used for the internal lock signal and TSP‐d4 as the internal standard with a microliters of D2O containing 0.1177 mM TSP-d4 were added into 24 mg of leaf powder as chemical shift of δ 0.0. The extracts were impregnated for 1 h, and sonicated for 15 min, followed by extraction solvent. D O was used for the internal lock signal and TSP-d as the internal standard with centrifugation for 210 min (14,000 rpm) at room temperature. The supernatants4 were transferred into a chemical shift of δ 0.0. The extracts were impregnated for 1 h, and sonicated for 15 min, followed by 5 mm NMR tubes (Vineland, NJ, USA) for further analysis. centrifugation for 10 min (14,000 rpm) at room temperature. The supernatants were transferred into 5 mm3.3. NMR NMR tubesSpectroscopy (Vineland, NJ, USA) for further analysis.

3.3. NMRAll Spectroscopy the NMR spectra were acquired at 298 K on a 600 MHz Bruker AVIII HD spectrometer (Bruker BioSpin, Rheinstetten, Germany) equipped with a 5 mm BBO H&F cryogenic probe. AutomaticAll the NMR shimming spectra and were adjusting acquired of 90° at pulse 298 Klength on a were 600 performed MHz Bruker for AVIIIeach sample. HD spectrometer Standard (BrukerCarr‐Purcell BioSpin,‐Meiboom Rheinstetten,‐Gill (CPMG) Germany) pulse sequence equipped with with water a presaturation 5 mm BBO was H&F used cryogenic to record probe. the ◦ Automatic1H‐NMR shimmingspectra for filtering and adjusting the broad of signals 90 pulsefrom the length molecules were (such performed as proteins) for with each short sample. T2 Standardrelaxation Carr-Purcell-Meiboom-Gill times. A total of 64 scans (CPMG)were collected pulse into sequence 64 K data with points water over presaturation a spectral width was usedof to record12,019.2 the Hz1H-NMR with a relaxation spectra for delay filtering of 4 thes and broad an acquisition signals from time the of molecules 3.0671 s. (suchAn exponential as proteins) withline short‐broadening T2 relaxation of 0.3 times. Hz was A applied total of 64to the scans free were induction collected decay into prior 64 K to data Fourier points transformation. over a spectral width of 12,019.2 Hz with a relaxation delay of 4 s and an acquisition time of 3.0671 s. An exponential Molecules 2018, 23, 554 8 of 10 line-broadening of 0.3 Hz was applied to the free induction decay prior to Fourier transformation. Additional 1H, 1H-correlation spectroscopy (COSY), 1H-1H total correlation spectroscopy (TOCSY) and 1H, 13C-heteronuclear single quantum correlation (HSQC) spectra were recorded on the selected samples for the purpose of resonance assignment.

3.4. Quantification of the Metabolites The original NMR data were processed by MestReNova 6.1.0 (Mestrelab Research S. L., Santiago de Compostela, Spain) with automatic correction of phase and baseline. Quantitative signals of 11 metabolites were selected to integrate manually. Despite the integral areas directly proportional to the number of contributing protons, the accuracy of measurements can also be influenced by various experimental parameters and conditions, such as pulse sequence, relaxation delay, and the purity of the internal standard. We thus employed external calibration curve to calculate the content of each metabolite. Six different concentration solutions of acetic acid, alanine, asparagine, choline, fumaric acid, GABA, glucose, proline, succinic acid, sucrose and trigonelline in the ranges of 0.1848–5.1925 mM, 0.0288–0.9204 mM, 0.1650–5.2816 mM, 0.0895–2.8656 mM, 0.4298–13.7509 mM, 0.0640–2.0413 mM, 1.2417–39.7322 mM, 0.1928–6.1634 mM, 0.0350–1.1179 mM, 0.3111–9.9562 mM and 0.0265–0.8450 mM were employed to establish the linear regression equations according to the ratio between the peak areas of metabolites and internal standard (x) versus the concentration of given metabolites (y).

3.5. Statistical Analysis All data were expressed as mean values ± standard deviation and performed using GraphPad Prism 6.04 software (GraphPad Software, Inc., San Diego, CA, USA). Principal component analysis (PCA) analysis, correlation analysis and pathway analysis based on Kyoto Encyclopedia of Genes and Genomes (KEGG) database were carried out with MetaboAnalyst 3.0 (http://www.metaboanalyst.ca/).

4. Conclusions A quantitative 1H-NMR method was established to identify and determine 11 primary metabolites in the leaves of Morus alba L. External calibration curves were employed for quantitative measurements, and subsequent validation results showed the proposed approach having good linearity, precision, repeatability and accuracy. Compared to conventional HPLC-UV spectroscopy, the developed 1H-NMR approach enables a robust and non-destructive determination of primary metabolites in the extracts of mulberry leaves without any derivatization or separation procedures. Multivariate statistical analyses revealed that GABA, asparagine, choline, alanine, and proline were probably varied with geographical regions, while trigonelline had no obvious in different origins.

Supplementary Materials: Supplementary Materials are available on line. Table S1 and S2. The contents of 11 metabolites in the extracts of mulberry leaves from 18 origins. Table S3. The geographical information of 18 different regions. Acknowledgments: This work was financially supported by the Natural Science Foundation of China (No. 81573547), the project of Tianjin Applied Basic and Cutting-edge Technology Research Program (No. 15JCTPJC60000) and (No. 15JCYBJC29300), and the Open Fund of Guangdong Province Key Laboratory of Pharmacodynamic Constituents of TCM and New Drugs Research (No. 2016B 030301004). Author Contributions: M.J. designed the experiment and was responsible for the study conception; Q.L. carried out the data analysis and wrote the paper, Q.W. and Q.L. performed the experiment; Y.W. (Yuan Wang) and J.H. contributed to the laboratory experiments, and Y.W. (Ya-nan Wang) contributed to the acquirement of NMR spectra. Conflicts of Interest: There are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome. Molecules 2018, 23, 554 9 of 10

Abbreviations

1H-NMR Proton nuclear magnetic resonance GABA γ-Aminobutyric acid TSP-d4 Sodium 3-trimethylsilyl [2,2,3,3-d4] propionate D2O Deuterium oxide CPMG Carr-Purcell-Meiboom-Gill COSY 1H, 1H-correlation spectroscopy TOCSY 1H-1H total correlation spectroscopy HSQC 1H, 13C-heteronuclear single quantum correlation PCA Principal component analysis SM Standard material PY Pingyao AY Anyang ZK Zhoukou SZ Suzhou BZ Bozhou GZ Guangzhou JN Jinan QD Qingdao LY Linyi HZ Heze ZB Zibo DZ Dezhou AG Anguo XT Xingtai CD Chengde DA Daan FZ Fuzhou

References

1. Wang, D.; Dong, Z.; Zhang, Y.; Guo, K.; Guo, P.; Zhao, P.; Xiao, Q. Proteomics provides insight into the interaction between mulberry and silkworm. J. Proteome Res. 2017, 16, 2472–2480. [CrossRef][PubMed] 2. Kobayashi, Y.; Miyazawa, M.; Kamei, A.; Abe, K.; Kojima, T. Ameliorative effects of mulberry (Morus alba L.) leaves on hyperlipidemia in rats fed a high-fat diet: Induction of fatty acid oxidation, inhibition of lipogenesis, and suppression of oxidative stress. Biosci. Biotechnol. Biochem. 2010, 74, 2385–2395. [CrossRef][PubMed] 3. Pothinuch, P.; Tongchitpakdee, S. Melatonin contents in mulberry (Morus spp.) leaves: Effects of sample preparation, cultivar, leaf age and tea processing. Food Chem. 2011, 128, 415–419. [CrossRef][PubMed] 4. Gryn-Rynko, A.; Bazylak, G.; Olszewska-Slonina, D. New potential phytotherapeutics obtained from white mulberry (Morus alba L.) leaves. Biomed. Pharmacother. 2016, 84, 628–636. [CrossRef][PubMed] 5. Chen, H.W.; He, X.H.; Liu, Y.; Li, J.; He, Q.Y.; Zhang, C.Y.; Wei, B.J.; Zhang, Y.; Wang, J. Extraction, purification and anti-fatigue activity of γ-aminobutyric acid from mulberry (Morus alba L.) leaves. Sci. Rep. 2016, 6, 18933. [CrossRef][PubMed] 6. Watanabe, M.; Maemura, K.; Kanbara, K.; Tamayama, T.; Hayasaki, H. GABA and GABA Receptors in the Central Nervous System and Other Organs. Int. Rev. Cytol. 2002, 213, 1–47. [PubMed] 7. Katsube, T.; Imawaka, N.; Kawano, Y.; Yamazaki, Y.; Shiwaku, K.; Yamane, Y. Antioxidant flavonol glycosides in mulberry (Morus alba L.) leaves isolated based on LDL antioxidant activity. Food Chem. 2006, 97, 25–31. [CrossRef] 8. Thaipitakwong, T.; Numhom, S.; Aramwit, P. Mulberry leaves and their potential effects against cardiometabolic risks: A review of chemical compositions, biological properties and clinical efficacy. Pharm. Biol. 2018, 56, 109–118. [CrossRef][PubMed] Molecules 2018, 23, 554 10 of 10

9. Lee, Y.J.; Choi, D.H.; Kim, E.J.; Kim, H.Y.; Kwon, T.O.; Kang, D.G.; Lee, H.S. Hypotensive, hypolipidemic, and vascular protective effects of Morus alba L. in rats fed an atherogenic diet. Am. J. Chin. Med. 2011, 39, 39–52. [CrossRef][PubMed] 10. Aramwit, P.; Supasyndh, O.; Siritienthong, T.; Bang, N. Mulberry leaf reduces oxidation and C-reactive protein level in patients with mild dyslipidemia. Biomed. Res. Int. 2013, 2013, 787981. [CrossRef][PubMed] 11. Clark, G.; O’Mahony, S.; Malone, G.; Dinan, T.G. An isocratic highperformance liquid chromatography method for the determination of GABA and glutamate in discrete regions of the rodent brain. J. Neurosci. Methods 2007, 160, 223–230. [CrossRef][PubMed] 12. Simmler, C.; Napolitano, J.G.; McAlpine, J.B.; Chen, S.N.; Pauli, G.F. Universal quantitative NMR analysis of complex natural samples. Curr. Opin. Biotechnol. 2014, 25, 51–59. [CrossRef][PubMed] 13. Bharti, S.K.; Roy, R. Quantitative 1H-NMR spectroscopy. TrAC Trends Anal. Chem. 2012, 35, 5–26. [CrossRef] 14. Gallo, V.; Intini, N.; Mastrorilli, P.; Latronico, M.; Scapicchio, P.; Triggiani, M.; Bevilacqua, V.; Fanizzi, P.; Acquotti, D.; Airoldi, C.; et al. Performance assessment in fingerprinting and multi component quantitative NMR Analyses. Anal. Chem. 2015, 87, 6709–6717. [CrossRef][PubMed] 15. Malz, F.; Jancke, H. Validation of quantitative NMR. J. Pharm. Biomed. Anal. 2005, 38, 813–823. [CrossRef] [PubMed] 16. López-Rituerto, E.; Cabredo, S.; López, M.; Avenoza, A.; Busto, J.H.; Peregrina, J.M. A thorough study on the use of quantitative 1H-NMR in Rioja red wine fermentation processes. J. Agric. Food Chem. 2009, 57, 2112–2118. [CrossRef][PubMed] 17. Almeida, C.; Duarte, I.F.; Barros, A.; Rodrigues, J.; Spraul, M.; Gil, A.M. Composition of beer by 1H-NMR spectroscopy: Effects of brewing site and date of production. J. Agric. Food Chem. 2006, 54, 700–706. [CrossRef][PubMed] 18. Staneva, J.; Denkova, P.; Todorova, M.; Evstatieva, L. Quantitative analysis of sesquiterpene lactones in extract of Arnica montana L. by 1H-NMR spectroscopy. J. Pharm. Biomed. Anal. 2011, 54, 94–99. [CrossRef] [PubMed] 19. Pauli, G.F.; Jaki, B.U.; Lankin, D.C. A Routine Experimental Protocol for qH NMR Illustrated with Taxol. J. Nat. Prod. 2007, 70, 589–595. [CrossRef][PubMed] 20. Engan, T. Magnetic resonance spectroscopy of blood plasma lipoproteins in malignant disease: Methodological aspects and clinical relevance. Anticancer Res. 1996, 16, 1461–1467. [PubMed] 21. Ala-Korpela, M. Critical evaluation of 1H-NMR metabonomics of serum as a methodology for disease risk assessment and diagnostics. Clin. Chem. Lab. Med. 2008, 46, 27–42. [CrossRef][PubMed] 22. Bollard, M.E.; Stanley, E.G.; Lindon, J.C.; Nicholson, J.K.; Holmes, E. NMR-based metabonomic approaches for evaluating physiological influences on biofluid composition. NMR Biomed. 2005, 18, 143–162. [CrossRef] [PubMed] 23. Hürlimann, M.D. Diffusion and relaxation effects in general stray field NMR experiments. J. Magn. Reson. 2001, 148, 367–378. [CrossRef][PubMed] 24. Hohmann, M.; Koospal, V.; Bauer-Christoph, C.; Christoph, N.; Wachter, H.; Diehl, B.; Holzgrabe, U. Quantitative 1H-NMR analysis of egg yolk, alcohol, and total sugar content in egg liqueurs. J. Agric. Food Chem. 2015, 63, 4112–4119. [CrossRef][PubMed] 25. Singh, J. International conference on harmonization of technical requirements for registration of pharmaceuticals for human use. J. Pharmacol. Pharmacother. 2015, 6, 185–187. [CrossRef][PubMed]

Sample Availability: Samples of the mulberry leaves are available from the authors.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).