*Article* **Chemical Constituents of the Leaves of Butterbur (***Petasites japonicus***) and Their Anti-Inflammatory E**ff**ects**

**Jin Su Lee 1, Miran Jeong 2, Sangsu Park 3, Seung Mok Ryu 4, Jun Lee 4, Ziteng Song 5, Yuanqiang Guo 5, Jung-Hye Choi 1,2, Dongho Lee 6,\* and Dae Sik Jang 1,2,\***


Received: 30 October 2019; Accepted: 27 November 2019; Published: 29 November 2019

**Abstract:** Two new aryltetralin lactone lignans, petasitesins A and B were isolated from the hot water extract of the leaves of butterbur (*Petasites japonicus*) along with six known compounds. The chemical structures of lignans **1** and **2** were elucidated on the basis of 1D and 2D nuclear magnetic resonance (NMR) spectroscopic data, electronic circular dichroism (ECD) and vibrational circular dichroism (VCD) spectra. Petasitesin A and cimicifugic acid D showed significant inhibitory effects on the production of both prostaglandin E2 (PGE2) and NO in RAW264.7 macrophages. The expressions of inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2) were inhibited by compound **1** in RAW264.7 cells. Furthermore, compounds **1** and **3** exhibited strong affinities with both iNOS and COX-2 enzymes in molecular docking studies.

**Keywords:** *Petasites japonicus*; Asteraceae; lignan; anti-inflammation; NO; PGE2; iNOS; COX-2; molecular docking

#### **1. Introduction**

*Petasites japonicus* Maxim (Asteraceae), known as butterbur, Japanese butterbur, and giant butterbur, is used as a botanical dietary supplement in the USA. The aerial parts of *P. japonicus* have been used in traditional Japanese folk medicine as an antipyretic, antitussive, or wound healing agent [1]. The constituents of *P. japonicus* have been reported and include flavonoids [2], sesquiterpenes [3–5], triterpenes [6], and various types of phenolic compounds [7]. Moreover, the leaves or stalks of *P. japonicus* are commonly consumed as vegetables in Korea and Japan. In the course of searching for active compounds from higher plants [8,9], the leaves of *P. japonicus* were selected for a detailed study since a hot water extract of the leaves of *P. japonicus* have shown inhibitory activity against nitric oxide (NO) production in RAW 264.7 cells [half maximal inhibitory concentration (IC50) value: 19 ± 4.9 μg/mL]. Phytochemical study on the hot water extract resulted in the isolation of two new

aryltetralin lactone lignans (**1** and **2**) along with six previously known compounds (Figure 1). The chemical structures of the new lignans **1** and **2** were determined by interpretation of 1D and 2D nuclear magnetic resonance (NMR) spectroscopic data, and by electronic circular dichroism (ECD) and vibrational circular dichroism (VCD) studies.

**Figure 1.** Chemical structures of the isolates **1**–**8**.

All the isolates from the leaves of *P. japonicus* were evaluated for their inhibitory effects on lipopolysaccharide (LPS)-induced production of pro-inflammatory mediators NO and prostaglandin E2 (PGE2) in the LPS-stimulated RAW 264.7 macrophages. We describe isolation of the secondary metabolites from the leaves of *P. japonicus*, structure elucidation of the two new lignans (**1** and **2**), and anti-inflammatory effects of the isolates as well as the possible mechanism.

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

#### *2.1. General Experimental Procedures*

General experimental procedures are described in the Supplementary Materials.

#### *2.2. Plant Material*

The leaves of *Petasites japonicus* (Asteraceae) were obtained from Nature Bio Co. (Seoul, Republic of Korea), in October 2016. The plant material was identified by one of the authors (D.S.J.) and the plant specimen (PEJA-2016) has been deposited in the Laboratory of Natural Product Medicine, College of Pharmacy, Kyung Hee University.

#### *2.3. Isolation of Compounds*

The dried leaves (500 g) were extracted once with 10 L of boiled water for 4 h and the solvent was evaporated with freeze drying. The extract (100.0 g) was separated over Diaion HP-20 (Mitsubishi, Tokyo, Japan) column eluted with an H2O-acetone gradient (from 1:0 to 0:1 *v*/*v*, gradient) to give 15 fractions (K1–K15). A part of fraction K3 was fractionated with medium pressure liquid chromatography (MPLC) using Redi Sep (Teledyne Isco, Lincoln, NE, USA)-C18 cartridge (13 g, acetonitrile–H2O, 0:1 to 3:7 *v*/*v*, gradient) and purified by high performance liquid chromatography (HPLC) using YMC Pack ODS-A column (Phenomenex, Torrance, CA, USA), yielding compound **8** (7.7 mg). Fraction K6 was separated over Sephadex LH-20 (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom) column with an acetone–H2O mixture (6:4 *v*/*v*) as solvent to give three fractions (K6-1–K6-3). Fraction K6-2 was fractionated further using Sephadex LH-20 with an acetone–H2O mixture (2:8 *v*/*v*) to yield six

subfractions (K6-2-1–K6-2-6). Compound **4** (0.5 g) and **3** (27.1 mg) were obtained from fraction K6-2-4 using LiChroprep RP-18 (Merck, Kenilworth, NJ, USA) CC. Fraction K8 was loaded to Sephadex LH-20 as stationary phase eluting with EtOH–H2O mixture (1:1 *v*/*v*) to afford 12 pooled fractions (K8-1–K8-12). Compound **5** (40.3 mg), **6** (4.2 mg), and **7** (4.7 mg) were purified from fraction K8-6 by HPLC with a Luna 10 μm C18(2) 100 Å column. Subfraction K8-9 was purified using Luna 10 μm C18 (**2**) 100 Å column to obtain compound **2** (61.6 mg). Fraction K9 was fractionated further using Sephadex LH-20 column eluted with the EtOH–H2O mixture (1:1 *v*/*v*) to generate ten fractions (K9-1–K9-10). Fraction K9-6 was purified by HPLC using YMC Pack ODS-A column, yielding compound **1** (4.2 mg).

#### 2.3.1. Petasitesin A (**1**)

Dark brownish powder; [α]D23: <sup>−</sup>11.5◦ (*c* 0.1, MeOH); ultraviolet (UV) (MeOH) <sup>λ</sup>max (log <sup>ε</sup>) 204 nm (3.85), 262 nm (3.57); CD (CH3CN) λmax 214 (−10.4), 239 (2.1), 254 (−5.3); infrared (IR) (ATR) νmax 3333, 2915, 2847, 1718, 1524, 1240 cm<sup>−</sup>1; High resolution electrospray ionization mass spectrometry (HRESIMS) (HRESIMS) (negative mode) *m*/*z* 325.0714 [M−H]<sup>−</sup> (calculated for C18H13O6, 325.0712) (Figure S1); 1H and 13C NMR data (Table 1) (Figures S2 and S3); 2D NMR data (Figures S4–S7).


**Table 1.** 1H (500 MHz) and 13C NMR (125 MHz) data of compounds **1** and **2**.

#### 2.3.2. Petasitesin B (**2**)

Dark brownish powder; [α]D23: <sup>−</sup>20.6◦ (*<sup>c</sup>* 0.1, MeOH); UV (MeOH) <sup>λ</sup>max (log <sup>ε</sup>) 206 nm (4.34), 287 nm (3.48); CD (CH3CN) λmax 203 (4.0), 210 (−5.4), 222 (−4.4), and 229 (1.9); IR (ATR) νmax 3305, 1768, 1606, 1514 cm<sup>−</sup>1; HRESIMS (negative mode) *<sup>m</sup>*/*<sup>z</sup>* 343.0810 [M−H]<sup>−</sup> (calculated for C18H15O7, 343.0818) (Figure S8); 1H and 13C NMR data (Table 1) (Figures S9 and S10); 2D NMR data (Figures S11–S14).

#### *2.4. Computational Methods*

ECD and VCD calculations of compounds **1** and **2** were conducted as described previously [10,11]. In brief, their 3D models were built from Chem3D modeling. Conformational analysis was performed by the MMFF force field as implemented in Spartan'14 software (Wavefunction, Inc., Irvine, CA, USA; 2014). Geometrical optimization of the selected conformers was performed at the B3LYP/6–31 + G (d,p) level by Gaussian 09 software (Revision E.01; Gaussian, Inc., Wallingford, CT, USA; 2009). The theoretical ECD and VCD spectra were calculated at the CAM-B3LYP/SVP level with a CPCM solvent model (acetonitrile) and at the DFT [B3LYP/6–31 + G(d,p)] basis set level by the Gaussian 09 software, respectively.

#### *2.5. Measurement of NO Production*

The 3-[4¨C-dimethylthiazol-2-yl]-2,5-dipheyl tetrazolium bromide (MTT) and Griess reaction assays were used for cell viability studies and measuring nitrite levels, respectively, as reported previously [12].

#### *2.6. Measurement of PGE2*

The RAW 264.7 macrophage cell lines were pretreated with various concentrations of the extract and isolates **1**–**8** for 1 h and then stimulated with or without LPS (1 μg/mL) for 24 h. A selective COX-2 inhibitor, NS-398 (*N*-[2-(cyclohexyloxy)-4-nitrophenyl]methanesulfonamide; Sigma Aldrich, St. Louis, MO, USA) was used as a positive control for blocking PGE2 production. PGE2 levels in cell culture mediums were measured using the same methods as described in the previous paper [12].

#### *2.7. Measurement of iNOS and COX-2 Expression*

Quantitative polymerase chain reaction (qPCR) using Thermal Cycler Dice Real Time PCR System (Takara Bio Inc., Shiga, Japan) was used to determine the steady-state mRNA levels of inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2) as reported previously [12].

#### *2.8. Molecular Docking Studies*

The software AutoDock Vina with AutoDock Tools (The Scripps Research Institute, La Jolla, CA, USA: ADT 1.5.6) using the hybrid Lamarckian Genetic Algorithm (LGA) was used for performing molecular docking simulations as reported in the literature [13,14]. In short, the 3D crystal structures (resolution: 2.5 Å) of iNOS (PDB code: 3E6T) and COX-2 (PDB code: 1PXX) were obtained from the RCSB Protein Data Bank. The configurations of compounds **1** and **3** were determined by their nuclear overhauser effect spectroscopy (NOESY) spectra and time-dependent density functional theory (TDDFT) ECD calculations. Chem3D Pro 14.0 software (CambridgeSoft, Waltham, MA, USA) was used for construction of the standard 3D structures (PDB format) of compounds **1** and **3**.

#### **3. Results**

#### *3.1. Structure Elucidation of Compounds 1 and 2*

Compound **1** was obtained as a dark brownish powder, and its molecular formula was identified as C18H14O6 by HRESIMS (*m*/*z* 325.0714 [M−H]<sup>−</sup>; calculated for C18H13O6, 325.0712). It exhibited UV maxima at 262 nm and IR maxima at 3333, 1718, and 1524 cm-1, suggesting the presence of a hydroxyl, ester group, and aromatic ring. The 13C NMR spectral data of compound **1** (Table 1) exhibited 18 carbon signals including a carbonyl carbon (δ<sup>C</sup> 173.9), 12 aromatic carbons (from δ<sup>C</sup> 115.7 to 145.9), an oxygenated methylene carbon (δ<sup>C</sup> 72.4), a methine carbon (δ<sup>C</sup> 42.3), and a methylene carbon (δ<sup>C</sup> 29.0).

The remaining two quaternary carbons (δ<sup>C</sup> 160.9 and 128.2) were derived from a double bond. The 1H NMR spectrum revealed one 1,2,4,5-tetrasubstituted aromatic ring [δ<sup>H</sup> 6.72 (s, H-6) and 6.60 (s, H-3)], and the one 1,3,4-trisubstituted aromatic ring [δ<sup>H</sup> 6.63 (d, *J* = 8.0, H-5 ), 6.59 (d, *J* = 2.0, H-2 ), and 6.45 (dd, *J* = 8.0 and 2.0, H-6 )], an oxygenated methylene [δ<sup>H</sup> 4.97 (d, *J* = 17.0) and 4.89 (d, *J* = 17.5), H-9], a methine [δ<sup>H</sup> 4.53 (s, H-7 )], and a methylene [δ<sup>H</sup> 3.86 (d, *J* = 23.0) and 3.62 (overlapped), H-7]. The heteronuclear multiple bond correlation spectroscopy (HMBC) correlations of **1** (Figure 2) suggest aryltetralin lactone type lignan with a double bond at C-8 and C-8 .

**Figure 2.** Selected correlations of compounds **1** and **2**: correlation spectroscopy (COSY, ŷ) and heteronuclear multiple bond correlation spectroscopy (HMBC, →) (in acetone-*d*<sup>6</sup> and methanol-*d*4).

The absolute configuration at C-7 of compound **1** was established by comparing its experimental ECD spectrum with those calculated spectra of (7 *R*) and (7 *S*) models using the time-dependent density functional theory (TDDFT) method. The experimental ECD spectrum of compound **1** exhibited a positive Cotton effect (CE) at 239 nm (Δε +2.1) and negative CEs at 214 nm (Δε −10.4) and 254 nm (Δε −5.3). The experimental data (Figure 3) was in agreement with the calculated ECD spectrum of the (7 *R*) model, suggesting the absolute configuration of compound **1** as (7 *R*). Thus, the structure of **1** was elucidated as (*R*)-9-(3, 4-dihydroxyphenyl)-6,7-dihydroxy-4,9-dihydronaphtho[2¨C-*c*]furan-1(3*H*)-one, and was named as petasitesin A.

**Figure 3.** Comparison of experimental and calculated electronic circular dichroism (ECD) spectra of compounds **1** (**A**) and **2** (**B**).

The compound **2** was isolated as a dark brownish powder. Its molecular formula was established as C18H16O7 by HRESIMS (*m*/*<sup>z</sup>* 343.0810 [M−H]<sup>−</sup>; calculated for C18H15O7, 343.0818). The 1H and 13C NMR data of **2** were similar to those of **1** (Table 1), although the NMR solvents were different from each other due to the different solubility of the compounds. Comparison of the 13C NMR data and molecular weights from **1** and **2** suggested that the carbons C-8 and C-8 of **1** with a double bonded linkage (δ<sup>C</sup> 160.9 and 128.2) were replaced by an oxygenated quaternary (δ<sup>C</sup> 78.0) and methine (δ<sup>C</sup> 56.2) carbon atoms. The correlation spectroscopy (COSY) correlation between H-7 (δ<sup>H</sup> 4.18) and H-8 (δ<sup>H</sup> 3.24), and the HMBC experiment revealed aryltetralin lactone type lignan (Figure 4). Considering a biogenetic relationship with **1**, the absolute configuration at C-7 of **2** was suggested to be (*R*)-configuration [15]. The coupling constant of 3.0 Hz between H-7 and H-8 suggested the *cis*-geometry of C-7 and C-8 . It was further confirmed by the NOESY interaction of H-7 and H-8 [16].

**Figure 4.** Comparison of experimental and calculated vibrational circular dichroism (VCD) spectra of compound **2** (*c* 0.5 M, DMSO-*d*6).

To determine the absolute configuration C-8 of **2,** experimental ECD spectrum of **2** was compared with the calculated spectra of (8*R,*7 *R,*8 *R*) and (8*S,*7 *R,*8 *R*) models using the TDDFT method. The experimental ECD spectrum of **2** showed positive CEs at 203 nm (Δε +4.0) and 229 nm (Δε +1.9), and negative CEs at 210 nm (Δε −5.4) and 222 nm (Δε −4.4). The experimental spectrum (Figure 3) was also in agreement with the calculated ECD spectrum of (8*S,*7 *R,*8 *R*) model. Moreover, the VCD spectrum of **2** was measured additionally to establish the configuration at C-8. The conformity of the experimental IR and VCD spectra and theoretical spectra of **2** suggested the absolute configuration of **2** as (8*S,*7 *R,*8 *R*) (Figure 4). Therefore, the structure of **2** was proposed as (9*R*,3a*S*,9a*R*)-9-(3,4-dihydroxyphenyl)-6,7,3a-trihydroxy-4,9,3a,9a-tetrahydronaphtho[2,3-*c*]furan-1(3*H*)-one, and was named as petasitesin B.

Compounds **3**–**8** were identified as cimicifugic acid D (**3**) [17], fukinolic acid (**4**) [7], 3,4-dicaffeoylquinic acid (**5**) [18], 3,5-dicaffeoylquinic acid (**6**) [18], 4,5-dicaffeoylquinic acid (**7**) [18], and caffeic acid (**8**) [19] by comparison of their NMR data with those reported.

#### *3.2. Anti-inflammatory E*ff*ects of the Isolates*

As shown in Table 2, cimicifugic acid D (**3**) and the new compound **1** (petasitesin A) exhibited significant inhibitory activities against NO production with IC50 values of 12 ± 1.1 and 15 ± 1.4 μM, respectively, without affecting the cell viability (Figure S15). 4,5-Dicaffeoylquinic acid showed mild activity with an observed IC50 value of 38.9 ± 0.72 μM. On the other hand, compound **1** showed the most potent inhibitory effect on PGE2 production with an IC50 value of 17 ± 3.2 μM (Table 2) in a dose-dependent manner (Figure 5). These results suggest that compound **1** might have an anti-inflammatory effect due to inhibition of the production of NO and PGE2 which are the key inflammatory mediators of macrophages. It is worth noting that compound **1** significantly suppressed the expression of NO and PGE2 synthesis enzymes, inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2), respectively (Figure 6), in a concentration-dependent manner. The data indicate that the inhibitory effect of compound **1** on NO and PGE2 production in macrophages is related to the regulation of iNOS and COX-2 expression.


**Table 2.** Inhibitory effects of the compounds from *P. japonicus* on NO and PGE2 production in lipopolysaccharide (LPS)-induced RAW 264.7 cells.

*<sup>a</sup>* The values represent the means of the results from three independent experiments with similar patterns. l-*N*6-(1-Iminoethyl)lysine (l-NIL) and *N*-[2-(cyclohexyloxy)-4-nitrophenyl]methanesulfonamide (NS-398) were used as a positive control substance for NO [half maximal inhibitory concentration IC50) value = 1.62 ± 0.08 μM] and prostaglandin E2 (PGE2) productions (IC50 value = 3.3 ± 0.15 μM), respectively. Three known compounds, fukinolic acid, 3,4-dicaffeoylquinic acid, and 3,5-dicaffeoylquinic acid were inactive (IC50 value > 50 μM) in this assay system.

**Figure 5.** Effects of compound **1** (6.25, 12.5, 25 or 50 μM) on LPS-stimulated NO (**A**) and PGE2 (**B**) productions in RAW 264.7 macrophages. #: *p* < 0.05 as compared with the untreated group, \*: *p* < 0.05 as compared with the LPS only-treated group.

**Figure 6.** Effect of compound **1** on the expression of iNOS (**A**) and COX-2 (**B**) in RAW 264.7 macrophages. #: *p*<0.05 as compared with the untreated group, \*: *p*<0.05 as compared with the LPS only-treated group.

To better understand the molecular mechanism of inhibitory activities against NO and PGE2 production, the most active compounds **1** and **3** were subjected to molecular docking studies. The results showed that **1** and **3** had strong affinities with both NO and PGE2 synthesis enzymes, iNOS and COX-2 (Figure 7, Table 3). The binding residues and logarithms of free binding energy are given in Table 3. These results implicated that **1** and **3** may directly interact with the cavity residues of iNOS and COX-2, leading to the activity reduction of free iNOS and COX-2 enzymes. Taken together, these results indicate petasitesin A (**1**), a novel lignan isolated from butterbur leaves extract, exhibits anti-inflammatory properties by suppressing NO and PGE2 production via inhibiting the expression of iNOS and COX-2 and binding to the free iNOS and COX-2 enzymes.

**Figure 7.** Molecular docking results of compounds **1** and **3** with iNOS (**A**) and COX-2 (**B**) enzymes. Molecular docking simulations obtained at the lowest energy conformation, highlighting potential hydrogen contacts of **1** and **3**, respectively (nitrogen is blue; oxygen is red; carbon is cyan; hydrogen is gray). For clarity, only interacting residues are labeled. Hydrogen bonding interactions are shown by dashes. These figures were created by PyMOL (Schrödinger, LLC, New York, NY, USA: version 1.3).


**Table 3.** Logarithms of free binding energies (FBE, kcal/ mol) of NO inhibitors to the active cavities


#### **4. Discussion**

In the present study, we isolated two new aryltetralin lactone lignans, petasitesin A and B (**1** and **2**) from the leaves of *P. japonicus*. To the best of our knowledge, this is the first report on the isolation of the aryltetralin lactone type lignans from the leaves of *P. japonicas*. Although cimicifugic acid D (**3**) has been isolated from *Cimicifuga* spp. including black cohosh (*Cimicifuga racemosa*) and possesses vasoactive effect and hyaluronidase inhibitory activity [20,21], this is the first finding that it presents in *P. japonicus* and inhibits pro-inflammatory mediators, NO and PGE2.

A new lignan petasitesin A (**1**) showed a potent inhibitory effect on the production of both NO and PGE2 in LPS-stimulated macrophages (IC50 values < 20 μM). Our molecular docking studies reveal that petasitesin A (**1**) can interact with the cavity residues of both iNOS and COX-2. Interestingly, petasitesin A (**1**) also inhibited the mRNA expression of iNOS and COX-2 induced by LPS in macrophages. However, the molecular mechanism of action underlying the gene expression regulation by petasitesin A remains to be investigated. Considering that LPS binds to toll-like receptor 4 (TLR4), the TLR4-mediated NF-κB pathway is likely associated with the inhibition of iNOS and COX-2 expression by petasitesin A. In fact, NF-κB is a key transcriptional factor to regulate the iNOS and COX-2 gene in macrophages under the inflammatory condition. In this regard, the effect of petasitesin A on the NF-κB pathway can be further elucidated.

#### **5. Conclusions**

New lignans (compounds **1** and **2**) and six known compounds were isolated and identified from the leaves of *P. japonicus*. Petasitesin A (**1**) and cimicifugic acid D (**3**) inhibit production of inflammatory mediators NO and PGE2. PetasitesinA (**1**) inhibits iNOS and COX-2 expression, and petasitesin A (**1**) and cimicifugic acid D (**3**) have strong affinities with both iNOS and COX-2 enzymes in molecular docking studies. Thus, petasitesin A (**1**) and cimicifugic acid D (**3**) are worthy of further pharmacological evaluation for their potential as anti-inflammatory drugs.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2218-273X/9/12/806/s1, Figure S1: HR-ESI-MS spectrum of compound **1**, Figure S2: The 1H-NMR (500 MHz, CD3COCD3) spectrum of compound **1**, Figure S3: The 13C-NMR (125 MHz, CD3COCD3) spectrum of compound **1**, Figure S4: The HSQC spectrum of compound **1** in CD3COCD3, Figure S5: The COSY spectrum of compound **1** in CD3COCD3, Figure S6: The HMBC spectrum of compound **1** in CD3COCD3, Figure S7: The NOESY spectrum of compound **1** in CD3COCD3, Figure S8: HR-ESI-MS spectrum of compound **2**, Figure S9: The 1H-NMR (500 MHz, CD3OD) spectrum of compound **2**, Figure S10: The 13C-NMR (125 MHz, CD3OD) spectrum of compound **2**, Figure S11: The HSQC spectrum of compound **2** in CD3OD, Figure S12: The COSY spectrum of compound **2** in CD3OD, Figure S13: The HMBC spectrum of compound **2** in CD3OD, Figure S14: The NOESY spectrum of compound **2** in CD3OD, Figure S15: Cell viability of the isolates from *P. japonicus*.

**Author Contributions:** D.L. and D.S.J. conceived and designed the experiments; J.S.L., M.J., S.P., S.M.R., and Z.S. performed the experiments; J.S.L. and S.M.R. analyzed the data; J.L., Y.G., and J.-H.C. interpreted the data and contributed to manuscript structure and flow; and J.S.L., M.J., and Z.S. wrote the paper. All authors reviewed and confirmed the manuscript.

**Funding:** This research was supported by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture and Forestry (IPET) through Agri-Bio Industry Technology Development Program, funded by Ministry of Agriculture, Food and Rural Affairs (MAFRA), Republic of Korea (116044-1, 118046-03). This work was also supported by the National Research Foundation of Korea (NRF) grant funded by Ministry of Science and ICT (MSIT), Republic of Korea (NRF-2019R1A2C1083945).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 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/).

#### *Article*
