**1. Introduction**

*Morus* spp. (also commonly named mulberry) belong to the family Moraceae and are natively cultivated in China, as well as distributed in Japan and Korea [1]. For 5000 years, the mulberry plant (leaves) has been the sole fodder for silkworms (*Bombyx mori*) for the production of silk in China, and sericulture is one of the most important symbols of ancient China [2]. The major mulberry species, such as *Morus alba* L., *Morus multicaulis* Pitter., and *Morus atropurpurea* Roxb., are extensively used to produce mulberry leaves for silkworms after thousands of years of improvement and domestication. These species are generally divided into eight cultivar types according to their geographical distribution in China: Pearl River Basin type, Taihu Basin type, Sichuan Basin type, Middlestream of the Yangtze River type, Downstream of the Yellow River type, Loess Plateau type, Xinjiang type, and Northeast type [3].

**Citation:** Zou, X.-Y.; He, Y.-J.; Yang, Y.-H.; Yan, X.-P.; Li, Z.-B.; Yang, H. Systematic Identification of Bioactive Compositions in Leaves of *Morus* Cultivars Using UHPLC-ESI-QTOF-MS/MS and Comprehensive Screening of High-Quality Resources. *Separations* **2022**, *9*, 76. https:// doi.org/10.3390/separations9030076

Academic Editor: Ernesto Reverchon

Received: 23 February 2022 Accepted: 14 March 2022 Published: 15 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 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 (https:// creativecommons.org/licenses/by/ 4.0/).

In addition to application as plant fodder in traditional sericulture, mulberry leaves possess great developmental potential and utilization value in relation to medicine, food, and ecological protection [2]. For example, the dry leaves of *M. alba* are a traditional Chinese medicine recorded as Mori Folium in the *Chinese Pharmacopoeia* (2020 edition) to treat anemopyretic cold, lung cough, headache, and dizziness [4]. In particular, its fresh leaves serve as a general food due to enriched nutrients such as carbohydrates, protein, fats, fiber, and vitamins [5,6], which are an essential part of a healthy lifestyle. They also possess numerous essential compounds with various bioactivities that are important for the health of organisms. Previous research suggested that flavonoids, alkaloids, chlorogenic acid, free amino acids, some organic acids, etc., were the main bioactive constituents in *M.* spp. leaves (MSLs) [7–10]. The preclinical and clinical studies indicated that MSLs possessed beneficial functions, such as antidiabetic, antihyperglycemic, antioxidant, and antiobesity functions [11].

The multiple beneficial functions of MSLs explain their extensive application in food products, such as mulberry leaf tea and beverages [12–14]. MSLs are also usually pulverized as an additive to improve the flavor and nutritional value of food. Mulberry leaf kueh (桑叶粿) is a delicious pastry made of homogenized liquid of fresh MSLs and glutinous rice, which has been popular in the Chaoshan region of China for hundreds of years [15]. Nevertheless, bioactive compositions of different MSL resources have not yet been illustrated clearly or compared comprehensively by researchers. Even the same variety of MSL usually shows great variation in widely cultivated regions. Therefore, it is very important to systematically analyze the compositions in different MSL resources for screening of highly valuable *Morus* cultivars.

Liquid chromatography coupled with high-resolution mass spectrometry (LC-HRMS) methods, such as ultrahigh-performance liquid chromatography–electrospray ionization quadrupole time-of-flight mass spectrometry (UHPLC-ESI-QTOF-MS/MS), have been successfully used for efficient annotation and rapid quantification of organic ingredients in complex matrices [16–21]. These improved LC-HRMS technologies have made it easier to study bioactive components in foods and plants. This study developed an improved LC-HRMS strategy to systematically identify the main components and construct chemical profiles of MSLs using UHPLC-ESI-QTOF-MS/MS coupled with global natural product social molecular networking (GNPS). Furthermore, multiple MSL resources collected from different geographical areas were evaluated by semiquantitation of these typical components. High-quality *Morus* cultivars were finally screened based on multivariate statistical analysis methods.

#### **2. Materials and Methods**

#### *2.1. Materials and Chemicals*

A total of 90 *Morus* resources that mainly originated from eight geographical areas, including three major species, namely *Morus alba* L., *Morus multicaulis* Perr., and *Morus atropurpurea* Roxb., along with *Morus cathayana* Hemsl., *Morus bombycis* Koidz., *Morus australis* Poir., and *Morus mizuho* Hotta., were cultivated in the Mulberry Variety Resources Nursery of Sericultural Research Institute in Changsha City, the Mulberry Experimental Base in Lixian County, and the Donghuang Mulberry Seedling Base in Ningxiang City, China. The leaves of each of them were collected in June 2021. The fresh weight of mulberry leaves of each sample was first calculated and then dried at 50 ◦C to constant dry weight. The dried leaf samples were milled with a grinder and then screened through a 100-mesh sieve. The filtered powder was stored in a desiccator at room temperature in the dark before the experiment.

Commercial standards comprising isoleucine, leucine, phenylalanine, tryptophan, valine, 1-deoxynojirimycin (1-DNJ), *γ*-aminobutyric acid (GABA), chlorogenic acid, rutin, isoquercitrin, quercetin, and kaempferol were purchased from Yuanye Bio-Technology Co., Ltd. (Shanghai, China) and had over 98% purity as detected by high-performance liquid chromatography (HPLC). Acetonitrile, methanol, and formic acid were of chromatographic

grade for LC-HRMS analysis (Merck KGaA Co., Ltd., Darmstadt, Germany). Analytical ethanol and other reagents were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Wahaha purified water was used in the study.

#### *2.2. Preparation Solutions of Standards and Samples for LC-HRMS Analysis*

In total, 12 reference standards, isoleucine (1.06 mg), leucine (1.04 mg), phenylalanine (1.03 mg), tryptophan (1.01 mg), valine (1.07 mg), 1-DNJ (0.81 mg), GABA (0.98 mg), chlorogenic acid (1.17 mg), rutin (1.07 mg), isoquercitrin (0.90 mg), quercetin (0.77 mg), and kaempferol (0.83 mg), were each dissolved in 1.0 mL acetonitrile/methanol (1:1, *v*/*v*) to prepare individual stock solutions. Each solution was diluted at appropriate concentrations of 5.0−10.0 μg/mL for qualitative analysis. All the stock solutions were mixed and further diluted into a series of appropriate concentrations at a range of 0.03−100.0 μg/mL for quantitative analysis.

Next, 0.15 g powder of each sample was added to 3.0 mL 60% ethanol solution and extracted for 20 min using an ultrasonic instrument with consistent power (200 W, 40 Hz; Kunshan Ultrasonic Instrument Co., Ltd., Kunshan, China). The extract was centrifuged at 12,000 rpm for 15 min, and then 1.0 mL of supernatant was filtered by a 0.22 μm membrane and then used for LC-HRMS analysis. All of the tests on the sample were carried out in triplicate.

#### *2.3. LC-HRMS Conditions*

An Agilent 1290 UHPLC coupled with a 6545B ESI-Q-TOF/MS system (Agilent Technologies, Palo Alto, CA, USA) was employed for the qualitative and quantitative analysis of mulberry leaves. An SB C18 column (2.1 × 100 mm, 1.8 μm, Agilent Technologies, Palo Alto, CA, USA) at a consistent temperature condition of 35 ◦C was used for the chromatographic separation. A two-phase system composed of phase A (ultrapure water containing 0.1% formic acid) and phase B (acetonitrile) was adopted with a flow rate of 0.15 mL/min, and an optimized gradient elution condition was created (0–2 min, 2% B; 2–3 min, 2–10% B; 3–5 min, 10% B; 5–8 min, 10–25% B; 8–10 min, 25% B; 10–13 min, 25–55% B; 13–16 min, 55% B; 16–19 min, 55–90% B; 19–21 min, 90–98% B; 21–30 min, 98% B). The injection volume for individual samples and standard solutions was 2.0 μL.

The mass-spectrometric conditions were improved in positive and negative ESI modes. The drying gas temperature was 325 ◦C with a flow rate was 9 L/min, the nebulizer pressure was 50 psi, the sheath gas temperature was 365 ◦C with a flow rate of 11 L/min, the nozzle voltage was 0.5 kV, the VCAP voltage was 4.0 kV, the fragmentor voltage was from 80 to 200 V, the OCT1 RF Vpp was 750 V, and the skimmer voltage was 65 V. The mass data were obtained at a range of between *m*/*z* 50 and *m*/*z* 1000 Da. Fragment ions of individual mass peaks were acquired under collision energies from 5 to 40 eV. Data acquisition and processing were carried out with Agilent Acquisition software and Agilent MassHunter Qualitative Analysis (version B.08.00), respectively.

#### *2.4. GNPS Library Help to Identification of Compounds*

The ".d" data were converted into ".MGF" file format by Agilent MassHunter Quantitative Navigator (version B.08.00), and the dedicated module "Create Molecular Networking" was used to create a network on the GNPS web platform to match potential compounds by comparing MS/MS spectra in the GNPS library. This work can be accessed via the GNPS website at https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=8d05eff8ec5e4fb3a8786c501 8600d49 (25 December 2021).

#### *2.5. Quality Evaluation by Multivariate Statistical Analysis*

The SIMCA tool (14.1) was employed to analyze responsible markers in MSL samples. The criteria importance through intercriteria correlation (CRITIC) method was used to comprehensively evaluate the MSL quality [22].

### **3. Results and Discussion**

One type of bioactive analogue generally possessed a certain kind of skeleton and was found extensively in complex matrices. Thus, it is important to rapidly mark and effectively identify these substances. However, some compounds were found to have the same *m*/*z* of skeleton (isomeric structures) when MS/MS analysis was carried out. It is necessary to discriminate fragments of skeleton ions to avoid the possible misattribution of homologues. The identification of compounds was undertaken using the strategy of combining in-source collision-induced dissociation (IS-CID) with target collision-cell CID (TCC-CID) to quickly catch analogues with the same skeleton (Figure 1).

**Figure 1.** An improved strategy combining IS-CID with TCC-CID to screen analogues with the same skeleton.

The fragmentor (80–200 V) has traditionally been selected according to the structural characteristics of compounds to obtain HRMS spectra with low dissociation for parent ions. In this study, a strategy that combined IS-CID with TCC-CID based on UHPLC-ESI-QTOF-MS/MS was successfully established for fast annotation of analogues with the same skeleton. In brief, TIC containing the HRMS spectrum of a sample solution was first collected under a higher fragmentor voltage (200 V) after liquid chromatography separation. The intensity of the skeleton ion dissociated from the parent ions generally increased after improving the fragmentor, and then the RT for each potential analogue was marked by an EIC finding. Subsequently, TCC-CID at a specific collision energy (CE) was performed for the potential skeleton ion (with consistent mass to charge, *m*/*z*) to capture the MS/MS spectrum. These MS/MS spectra were compared with each other to accurately label the skeleton with unanimous fragment ions and intensity to screen analogues.

#### *3.1. Characterization of Morus Leaves*

To quickly acquire more abundant mass data of bioactive components present in these *Morus* leaves, a QC sample was prepared by mixing an equal solution of each MSL sample, and the mass information was collected by the IS-CID−TCC-CID strategy (Figure 2a,b). The typical mass characteristics of total ion chromatograms (TICs) under ESI positive and negative modes indicated that multiple bioactive compositions in MSLs potentially exist

(Table 1). Efficient identification of these bioactive ingredients is the basic foundation for understanding the beneficial effects. Therefore, GNPS was introduced to assist in the identification of potential compounds by comparison with MS/MS spectra in the online library.

**Figure 2.** TICs of *Morus* spp. leaves (QC) in −ESI (**a**) an d +ESI (**b**) modes, EIC finding potential skeletons of quercetin, kaempferol, and quinic acid at *m*/*z* 303.05 (**c**), *m*/*z* 287.05 (**d**), and *m*/*z* 191.05 (**e**), respectively.


**Table 1.** Bioactive and nutritive compositions in 90 samples of *Morus* leaves under UHPLC-ESI-QTOF-MS/MS investigation in positive and negative ESI modes.


**Table 1.** *Cont.*

<sup>a</sup> confirmed by standard substances; <sup>b</sup> annotated by the IS-CID−TCC-CID strategy; <sup>c</sup> identified by comparison through GNPS library; <sup>d</sup> tentatively identified.

#### 3.1.1. Identification of Quercetin-Type Flavonoids (QTFs)

The general nomenclature of the aglycone fragments is shown in Figure 3a. The IS-CID technique was first performed to dissociate the QTFs with a fragmentor at 200 V, and the potential QTFs were screened by EIC finding with *m*/*z* 303.05 (skeleton ion of potential quercetin) as shown in Figure 2c. Then, these skeleton ions were measured with target MS/MS by the TCC-CID technique, resulting in MS/MS spectra of individual skeleton ions (Figure 3b), which were compared with MS/MS spectra of the standard quercetin (Figure 3c). The MS/MS spectra of quercetin exhibited the consecutive loss of the neural ions of H2O and CO (Figure 3c), resulting in typical fragment ions at *m*/*z* 257.04 [M + H - H2O - CO]+, *<sup>m</sup>*/*<sup>z</sup>* 229.05 [M+H-H2O-2 × CO]+, and *<sup>m</sup>*/*<sup>z</sup>* 201.05 [M+H-2 × H2O - <sup>2</sup> × CO]+. Simultaneously, *<sup>m</sup>*/*<sup>z</sup>* 165.02 [0,2A]+, *<sup>m</sup>*/*<sup>z</sup>* 153.02 [1,3A]+, and *<sup>m</sup>*/*<sup>z</sup>* 137.02 [0,2B]+ showed the typical fractures of quercetin. Therefore, aglycones of compounds **24**, **26**, **27**, **29**, **30**, **32**, **33**, and **37** showed consistent fragment ions with quercetin that were undoubtedly classified as QTFs. The next step was to assign the linkage between quercetin aglycone and a sugar unit, which was performed under a moderate fragmentor voltage (100 V) to capture HRMS spectra of TIC with low dissociation of parent ions. QTFs possess a basic quercetin skeleton that generally joins glucoside, rutinoside, or malonyl-glucoside with C3, which were confirmed as the major phytochemicals in MSLs. The C7 position would link to some other sugar unit when C3 was occupied.

**Figure 3.** General nomenclature of the aglycone fragments (**a**); comparison of MS/MS spectrum between potential quercetin skeleton (**b**) and authorized quercetin substance (**c**); comparison of MS/MS spectrum between potential kaempferol skeleton (**d**) and authorized kaempferol substance (**e**).

Compounds **29**, **30**, and **37** were identified as rutin (que-3-rut), isoquercitrin (que-3 glu), and quercetin, respectively, by comparison with the authorized substances. Compound **24** showed a parent ion at *m*/*z* 713.16 [M + H]+, and fragment ions at *m*/*z* 551.10 [M + H - glu]+, *m*/*z* 465.10 [M + H - glu - mal]+, and *m*/*z* 303.05 [M + H - glu - mal - glu]+ were identified as quercetin-3-*O*-(6"-*O*-malonyl)-glucose-7-*O*-glucoside (que-mal-glu-glu) by referring to the GNPS library. Similarly, compound **26** had a parent ion and *m*/*z* 697.16 [M + H]+ and possessed fragment ions at *m*/*z* 611.16 [M + H - mal]+, *m*/*z* 465.10 [M +

H - mal - rha]+, and *m*/*z* 303.05 [M + H - mal - rha - glu]+. Compound **27** possessed parent ion and *m*/*z* 757.22 [M + H]+ and possessed fragment ions at *m*/*z* 611.16 [M + H - rha]+, *m*/*z* 465.10 [M + H - rha - rha]+, m/z 449.11 [M + H - rut]+, and *m*/*z* 303.05 [M + H - rha - rut]+. Compounds **26** and **27** were identified as quercetin-3-*O*-(6"-*O*-malonyl) glucose-7-*O*-rhamnoside (que-mal-glu-rha) and quercetin-3-*O*-rutinoside-7-*O*-rhamnoside (que-rut-rha), respectively, by comparing MS/MS data. Compounds **32** and **33** showed the same parent ion and fragment ion at *m*/*z* 551.10 and *m*/*z* 303.05, respectively, which indicated the existence of malonyl-glucoside in conjunction with quercetin aglycone, and both were empirically identified as quercetin-3-*O*-(6"-*O*-malonyl)-glucoside (que-mal-glu 1) and quercetin-3-*O*-(2"-*O*-malonyl)-glucoside (que-mal-glu 2), respectively.

#### 3.1.2. Identification of Kaempferol-Type Flavonoids (KTFs)

KTFs are also important bioactive compounds in MSLs that contain kaempferol aglycone and show the same sugar connection mode as QTFs. Potential KTFs were screened by an EIC finding with *m*/*z* 287.05 (skeleton ion of potential kaempferol) using the IS-CID technique with a fragmentor at 200 V (Figure 2d). Similarly, TCC-CID was performed to compare MS/MS spectra of these skeleton ions with the standard kaempferol (Figure 3d,e). The MS/MS spectra of kaempferol showed similar fragment patterns to quercetin and had diagnostic ions at *m*/*z* 121.03 [0,2B]+, which had been confirmed in our previous study [23].

As results, compounds **28**, **31**, **34**, **35**, and **36** accurately belonged to KTFs compared with MS/MS spectra of the standard kaempferol (**40**). Compound **28** had a parent ion at *m*/*z* 741.22 [M + H]<sup>+</sup> and fragment ions at *m*/*z* 595.17 [M + H - rha]+, *m*/*z* 449.18 [M + H - rha - rha]+, and *m*/*z* 287.05 [M + H - rha - rut]+; compound **31** exhibited a parent ion at *m*/*z* 595.17 [M + H]+ and fragment ions at *m*/*z* 449.18 [M + H - rha]<sup>+</sup> and *m*/*z* 287.05 [M + H - rut]+; compound **34** showed a parent ion and *m*/*z* 449.18 [M + H]+ and a fragment ion at *m*/*z* 287.05 [M + H - glu]+. These three compounds were annotated as kaempferol-3-*O*-rutinoside-7-*O*-rhamnoside (kae-rut-rha), kaempferol-3-*O*-rutinoside (kae-rut), and kaempferol-3-*O*-glucoside (kae-glu), respectively. Compounds **35** and **36** showed the same parent ion and typical fragment ion at *m*/*z* 535.11 and *m*/*z* 287.05, respectively, which indicated the existence of malonyl-glucoside in conjunction with kaempferol aglycone, and both were empirically assigned as kaempferol-3-*O*-(6"-*O*-malonyl)-glucoside (kae-mal-glu 1) and kaempferol-3-*O*-(2"-*O*-malonyl)-glucoside (kae-mal-glu 2), respectively.

#### 3.1.3. Identification of Chlorogenic Acids (CAs)

IS-CID with a fragmentor at 200 V in ESI negative mode showed a consistent skeleton ion of four potential quinic acid homologues at *m*/*z* 191.06 (Figure 2e). One of them was identified as chlorogenic acid by standard reference (compound **22**). The skeleton ions (*m*/*z* 191.06) of the other three peaks showed consistent MS/MS spectra with the skeleton ion of chlorogenic acid under TCC-CID, which indicated that they were isomers due to the variety of different substituted positions of the caffeoyl unit on quinic acid (Figure 4a).

It is very difficult to accurately annotate these positional isomers. Fortunately, Jia-Yu Zhang et al. reported these four CA isomers [24], and MS/MS spectra in this study showed consistent retention behaviors and fragment ions with them. CAs **19**, **23**, and **25** were identified as neochlorogenic acid, cryptochlorogenic acid, and 1-caffeoylquinic acid, respectively, which exhibited similar fragment patterns. For cryptochlorogenic acid, for example, the parent ion (*m*/*z* 353.09 [M - H]−) was dissociated from *m*/*z* 179.04 by loss of quinic acid and subsequently generated *m*/*z* 135.05 by loss of CO2, or was dissociated from *m*/*z* 191.06 by loss of caffeoyl unit and then generated *m*/*z* 173.05 by loss of H2O (Figure 4b). These similar fragment ions might lead to uncertain annotation; therefore, this study calculated the ratio of parent ion to skeleton ion at a specific collision energy (10 eV), and each CA isomer was efficiently discriminated in complex matrices (Figure 4c).

**Figure 4.** Fragmentation patterns (**a**), MS/MS spectrum (**b**) and ion ratios of four chlorogenic acid isomers (**c**).

3.1.4. Identification of 1-DNJ, GABA, Amino Acids, and Unsaturated Fatty Acids

1-DNJ (**3**) and GABA (**9**) were typical bioactive compounds in MSLs and were accurately identified by referring to the authorized compounds. In total, 16 amino acids (AAs), including the standard substances valine, isoleucine, leucine, phenylalanine, and tryptophan, and the initial identified AAs by accurate mass, MS/MS spectra, GNPS library, were determined. A pair of AA isomers, isoleucine and leucine, showed close RT and fragment ions. This study also provided an effective identification method that showed that isoleucine generally exhibited a higher fragment ion at *m*/*z* 69.07 than leucine in one consistent LC-HRMS setting in ESI positive mode. Additionally, unsaturated fatty acids and the other compounds in MSLs were also tentatively elucidated referring to accurate mass and fragment ions (Table 1).

#### *3.2. Qualification of Bioactive Compounds*

Semiquantitative analysis was performed using mixed standard solutions by Agilent MassHunter Quantitative Analysis (version 10.2). The calibration curve of each compound with correlation coefficients (R2) higher than 0.997 in appropriate concentration ranges was determined. Limits of detection (LOD) and quantification (LOQ) were obtained accordingly. RSD values of precision, reproducibility, and stability of the 12 compounds within 2.9% indicated that the method was suitable for the qualified requirement. Analogous or isomeric compounds that do not have standard substances were tentatively quantified using a similar authorized structure [19,25]. Specifically, the identified AAs, flavonoids, and CAs were quantified by referring to isoleucine, rutin, and chlorogenic acid, respectively.

The contents of 36 identified bioactive and nutritive compounds were quantified in 90 MSLs from different *Morus* cultivars. For the sum of 16 nutritive AAs, S55, S16, S34, S61, S40, S38, S53, S27, S05, S46, S59, S04, S20, and S37 showed higher values from 10.00 to 13.45 mg/g. Heat map analysis suggested that phenylalanine, leucine tyrosine, isoleucine, valine, proline, tryptophan, and asparagine possessed the major contents in the MSL samples (Figure 5a).

**Figure 5.** Amino acids (**a**) and typical bioactive compounds (**b**) in different *Morus* spp. leaves.

The concentrations of 20 bioactive compositions in different MSLs were calculated as shown in Figure 5b. The unique bioactive compound in MSLs was 1-DNJ, which showed over 1.00 mg/g in S64, S84, S20, S53, S56, S52, S27, and S83, from 1.00 to 1.83 mg/g. GABA had values over 1.50 mg/g in S84, S18, S10, and S48, from 1.54 to 1.82 mg/g. Chlorogenic acid possessed high content compared to all qualified compounds, and 80 of the 90 MSL samples had chlorogenic acid content of over 1.10 mg/g. Chlorogenic acid above 5.00 mg/g was detected in S34, S27, and S46 (Figure 5b). Rutin, quercetin-3-*O*-(6"-*O*-malonyl) glucoside (que-mal-glu 1), and kaemferol-3-*O*-(6"-*O*-malonyl)-glucoside (kae-mal-glu 1) were the main flavonoids in MSLs. In particular, que-mal-glu 1 showed values of more than 1.50 mg/g in 35 MSL samples, and it showed high content in S66, S70, and S36 from 3.51 to 4.38 mg/g. Most of the wild MSLs (S81–S90) showed low contents of the qualified bioactive compounds. For the sum of 20 bioactive compounds, S33, S27, S20, S50, S75, S42, S51, S24, S80, S34, S37, S52, S07, S35, S22, S59, S70, S18, S66, S49, S46, and S36 exhibited

higher values from 10.27 to 15.60 mg/g. As a result, S20, S37, S46, and S59 were screened with high values of nutritive AAs and bioactive compounds that all exceeded 10.00 mg/g in MSL samples.

#### *3.3. Important Variables of M.* spp. *Leaves*

The fact that there were several quantified datasets made the interpretation of MSL quality difficult. Therefore, partial least squares discriminant analysis (PLS-DA) was initially introduced to find important variables in these samples. First, 90 MSL samples were divided according to the species (cultivars), and the PLS-DA score scatter plot was presented (Figure 6a). The corresponding score plot combined with variable importance in the project (VIP > 1) values screened out compounds, namely quercetin-3-*O*-(6"-*O*malonyl)-glucoside (que-mal-glu 1), chlorogenic acid, kaempferol-3-*O*-rutinoside (kae-rut), rutin, kaempferol-3-*O*-(6"-*O*-malonyl)-glucoside (kae-mal-glu 1), isoleucine, valine, *N*isobutyrylglycine, and phenylalanine, which were significant constituents of the differences between MSL species (Figure 6b). Then, these samples were divided based on geographical distribution in a PLS-DA score scatter plot (Figure 6c). Similarly, these compounds mentioned above, along with proline and asparagine, were important variables (VIP > 1) for the differences between MSL original types (Figure 6d). The combined results of the PLS-DA data demonstrated that these important variables could serve as key values in evaluating the quality of different MSL samples.

**Figure 6.** PLS-DA score scatter plot for different species of *Morus* spp. leaves (**a**) and VIP value of compounds (**b**); PLS-DA score scatter plot for different original types of *Morus* spp. leaves (**c**) and VIP value of compounds (**d**).

#### *3.4. Quality Evaluation of M.* spp. *Leaves*

To comprehensively evaluate the quality of MSL samples according to the screened important variables obtained in PLS-DA analysis, the objective weight (*Wj*) of each compound was calculated according to the CRITIC method that was based on characteristic conflict (*Rj*), correlation of indicators (*rij*), amount of information (*Cj*), and standard deviation or intensity (*σj*). In short, the data matrix was established according to the standardized data and formulas (experimental values − experimental minimum)/(experimental maximum − experimental minimum), and the objective weight of each indicator was obtained (Table 2). The calculated formulas are depicted as follows [22]:

$$R\_{\vec{j}} = \sum\_{i=1}^{n} (1 - r\_{i\vec{j}}) \tag{1}$$

$$\mathbf{C}\_{\circ} = \sigma\_{\circ} \mathbf{R}\_{\circ} \tag{2}$$

$$\mathcal{W}\_{\dot{\jmath}} = \frac{\mathbb{C}\_{\dot{\jmath}}}{\sum\_{j=1}^{n} \mathbb{C}\_{j}} \tag{3}$$


**Table 2.** Comparison of intensity, conflict, information, and objective weight of each indicator.

This objective method resulted in a comprehensive evaluation of MSL resources from different species and areas of origin (Table 3). Samples S46 (M. multicaulis), S37 (*M. multicaulis*), S36 (*M. multicaulis*), S59 (*M. alba*), S20 (*M. multicaulis*), S34 (*M. multicaulis*), S27 (*M. multicaulis*), S66 (*M. alba*), S33 (*M. alba*), S52 (*M. multicaulis*), etc., had a higher score in the 90 MSL samples, which indicated the excellent potential of these *Morus* cultivars for higher value application of MSLs in the field of medicine and food.

**Table 3.** Comprehensive evaluation of *M.* spp. leaves from different areas using the CRITIC method.



#### **Table 3.** *Cont.*


**Table 3.** *Cont.*
