**Biomolecule from** *Trigonella stellata* **from Saudi Flora to Suppress Osteoporosis via Osteostromal Regulations**

**Hairul-Islam Mohamed Ibrahim 1,2,\*, Hossam M. Darrag 3,4, Mohammed Refdan Alhajhoj <sup>5</sup> and Hany Ezzat Khalil 6,7**


Received: 7 October 2020; Accepted: 2 November 2020; Published: 20 November 2020

**Abstract:** *Trigonella stellata* has used in folk medicine as palatable and nutraceutical herb. It also regulates hypocholesterolemia, hypoglycemia, and has showed anti-inflammatory activities as well as antioxidants efficacy. Osteoporosis is a one of bone metabolic disorders and is continuously increasing worldwide. In the present study, caffeic acid was isolated from *Trigonella stellata* and identified using 1 D- and 2 D-NMR spectroscopic data. Caffeic acid was investigated on osteoblast and osteoclast in vitro using mice bone marrow-derived mesenchymal cells. Caffeic acid played reciprocal proliferation between osteoblast and osteoclast cells and accelerated the bone mineralization. It was confirmed by cytotoxicity, alkaline phosphatase (ALP), alizarin red S (ARS), and Tartrate resistant acid phosphatase (TRAP) assay. Caffeic acid regulated the osteogenic marker and upregulated the osteopontin, osteocalcin, and bone morphogenic proteins (BMP). Quantitative real time PCR and Western blot were used to quantify the mRNA and protein markers. It also regulated the matrix metalloprotease-2 (MMP-2) and cathepsin-K proteolytic markers in osteoclast cells. In addition, caffeic acid inhibited bone resorption in osteoclast cells. On the other hand, it upregulate osteoblast differentiation through stimulation of extracellular calcium concentrations osteoblast differentiation, respectively. The results also were confirmed through in silico docking of caffeic acid against cathepsin-B and cathepsin-K markers. These findings revealed that caffeic acid has a potential role in bone-metabolic disorder through its multifaceted effects on osteoblast and osteoclast regulations and controls osteoporosis.

**Keywords:** *Trigonella stellata*; caffeic acid; osteoporosis; osteoblast; osteoclast; BMP

#### **1. Introduction**

Osteoporosis is a metabolic bone disease characterized by low bone mass, imbalanced bone cell types that leads to osteoarthritis (OA). Bone is a metabolically active connective tissue, ability to regenerate from the incidence and accident of fractures [1]. Bone reabsorption and osteogenic formation tends the remodeling of bone loss and balancing the bone stereotypes. Osteoclast types of cells efficiently absorb the damaged bone cells and promote osteoconductivity, as in-growth around the bone; induction of osteogenic response promote progenitor cell differentiation and mineral storage in osteoblastic

lineages. Balances between osteoblast and osteoclast would be happened by bone marrow derived cells. Bone-marrow derived mesenchymal stem cells (BM-MSCs) are multipotent, differentiated into connective tissue, and develop into mature osteoblasts [2]. Mesenchymal cells directly involved in extracellular matrix composition, mineralization, and coordinate the differentiation of osteostromal cells [3]. Osteoclasts are generated from precursor cells in presence of receptor activator of nuclear factor kappa-B ligand (RANKL) [4]. These inductions activate the inflammatory mediators which given back bone loss. Bone morphogenetic protein (BMP) and IL-10 are examples of osteogenic markers widely used in differentiation of mesenchymal stem cells into osteoblasts. Mitogen activated protein kinase (MAPK) regulates the osteoblast-specific transcription factors for differentiation process [5].

Simple analgesics are now recognized as one of the first line pharmacological treatment of uncomplicated OA. Whereas some non-steroidal anti-inflammatory drugs (NSAIDS) may show some fatal adverse effects particularly if used in long term treatment plans.

Natural products have been recently considered as a source of important therapeutic candidates, that could treat various diseases and considered effective for maintaining good health [6]. The interest in drugs derived from plants is predominantly attributed to the trust that green medicine is safe and dependable in comparison to the synthetic one. Wide ranges of natural candidates have implemented in treatment of chronic and infectious diseases [7].

Currently, a plethora of agents are available for the treatment of inflammatory disorders including OA, but some of the drugs are associated with risk of life-threatening adverse effects leading to its withdrawal from market [8]. Hence, the management of inflammatory disorders using medicines without side effects is still a challenge. In the last decades, hundreds of reports were published regarding the anti-inflammatory activities of plants that were available for alternative therapy [9]. Particularly, phenolic bioactive constituents showed significant attention due to their modulatory activities on inflammasomes [10]. Therefore, it is highly required to find out herbal products or nutraceuticals, which can be used as add-on therapy for long-term management of inflammatory disorders.

Studies have proven that leguminous plants may act as reservoirs of potential secondary metabolites of diverse therapeutic utilities and can produce an anti-inflammatory effect as well as they have significant nutritional value [11,12]. In this aspect, there is a consumption of *Trigonella* species (member leguminous herbs) because of their nutritional value. Particularly, *Trigonella foenum-graecum*, which is a small plant with several benefits, attributed to the diverse array of phytoconstituents such as phenolics and flavonoids [13]. Traditionally *Trigonella foenum-graecum* is reputed to exert anticancer, antidiabetic, antioxidant, antihyperlipidemic, and other various pharmacological effects [14].

*Trigonella stellata* (*T. stellate*) (Leguminosae) is a member of the genus *Trigonella* that is native and very common to grow wildly in Arabian Peninsula including Saudi Arabia [15]. Traditionally, *T. stellata* is a palatable herb [16] and used to treat abdominal pain, diarrhea, dysentery [17], and as nutraceutical herb [18]. Recent studies reported that *T. stellata* contains isoflavans and saponins and showed antidiabetic and anti-hyperlipidemic activities [19,20].

No previous reports have been performed to illustrate the efficacy of *T. stellata* as anti-erosion agent in vitro and in vivo platforms. Therefore, this study was aimed to investigate the *T. stellata* for the first time for such activities.

#### **2. Results**

#### *2.1. Isolation and Identification of Major Compound*

The methanol extract of shade dried aerial parts (25.0 g) was subjected to several and repeated chromatographic techniques to give the pure phenolic compound; Caffeic acid (CAF) (23.6 mg) [21] (Figure 1). The structure was elucidated by inspection of 1 D- and 2 D-NMR spectroscopic data (Table 1 and Supplementary Material) and compared with literature values [21]. This study represents the first report on the isolation of CAF from *T. stellata*.

**Figure 1.** Structure of caffeic acid (CAF) isolated from *T. stellata* (**A**). HPLC chromatogram of SubFr. 2-4-4-1 (**B**). HPLC chromatogram of collected pure CAF (**C**).


**Table 1.** NMR spectroscopic data of caffeic acid in CD3OD.

δC: chemical shift in ppm for <sup>13</sup>C-NMR, δH: chemical shift in ppm for <sup>1</sup>H-NMR; mult.: multiplicity; J in Hz: coupling constants in Hz; d: doublet, dd: doublet of doublet.

#### δ δ *2.2. E*ff*ect of* T. Stellata *Extract Against Mesenchymal Cells*

μ was not inhibited up to 50 μ The CAF at concentration >10 μM The biocompatibility of *T. stellata* crude extract (TCE) against murine bone marrow derived mesenchymal cells. Initially, the TCE was evaluated using BM-MSCs viability assay at concentrations (10, 20, 50, 100, 250, and 500 µg/mL) by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method (Figure 2A) and neutral red (NR) assay (Figure 2B). This examination revealed that the TCE metabolites play a major role in cell biocompatible. The major moiety isolated from TCE was CAF. The selection of biocompatible concentrations of CAF was carried out using MTT assay. Cell viability was not inhibited up to 50 µg/mL of TCE concentration. The CAF at concentration >10 µM decreases the cell viability with an IC50 with >10 µM concentration. The neutral red assay estimates the uptake of colour from viable cells. NR measurements of the total viable cells was correlated with MTT assay. The NR results revealed that, insignificant cell death was observed at all the tested concentrations. Even at 10 µM CAF treatment showed <25% cell death.

ith an IC50 with >10 μM concentration. The neutral red assay estimates

concentration tested from 10 μM to 500 μM. The cru was treated with osteogenic cells in concentration of 1.0 μM to 50 μM for 7 days and 14 days cultured **Figure 2.** In vitro cytotoxic evaluation of *T. stellata* crude extract (TCE) and its metabolites. (**A**) The osteostromal cells were isolated from murine bone marrow and osteogenic characters were induced. The osteogenic cells were further analyzed for cytotoxic effect against *T. stellata* crude extract. The concentration tested from 10 µM to 500 µM. The crude extract was treated with osteogenic cells for 72 h and evaluate the cell viability using MTT reagent. (**B**) The major metabolites caffeic acid (CAF) was treated with osteogenic cells in concentration of 1.0 µM to 50 µM for 7 days and 14 days cultured cells. (**C**) Alkaline phosphatase-specific activity. (**D**) Total protein content. Bars represent the mean ± SD (*n* = 4). Statistical results are shown as \* *p* < 0.05, Values compared between the PBS group with CAF at different concentrations.

Concerning, the effect of bone cell differentiation or proliferation, CAF significantly stimulated alkaline phosphatase (ALP) activity and bone cell differentiation in BM-MSCs cells. Thus, the effect of CAF on osteoblast culture from primary precursors cells was investigated. ALP was estimated at 14 days incubation (Figure 2C). The contents of ALP was increased 50% at 5-µM CAF treatments. ALP results were co-related with total protein content (Figure 2D). Total protein was raised up to 50% in CAF treated osteoblastic cells.

μM CAF treatments.

#### *2.3. CAF Regulated In Vitro Mineralization of Osteoblastic Cells*

CAF regulated the bone cell differentiation and influenced the loading of calcium on osteoblastic cells. CAF stimulated the proliferation in osteogenic cells. Thus, the effect of CAF was investigated using in vitro osteoblastic cells primary stained with alizarin red S (ARS). As shown in Figure 3A,B, differentiated osteoblasts were stained with ARS for a period of 7, 14, and 21 days, respectively. The contents of ARS staining increased 100–200% at CAF (5 and 10 µM) Figure 3C. Whereas 10 µM showed highly significant amplification of loading mineralization as well as accelerates the ARS staining up to 2-fold compared to PBS group. The inflammatory cytokines from osteoblast differentiation were quantified and found CAF insignificantly regulated the IL-10 and it was not overexpressed by the differentiation cellular modifications Figure 3D. Whereas, TNF-α was reduced significantly at 5 µM of CAF treatment in osteoblast stereotypes.

– % at CAF (5 and 10 μM) Figure 3C. Whereas 10 μM

α was reduced

–

α and IL **Figure 3.** The differentiation and in vitro mineralization of osteoblast was evaluated at CAF treated conditions. (**A**) Representative macroscopic (**B**) Microscopic observation of alizarin red S (ARS) (**C**) ARS quantification for osteoblastic Differentiation at CAF treated conditions. (**D**) Cytokines estimation at 7th day of CAF treated osteoblastic cells. TNF-α and IL-10 were estimation using ELISA kits. Bars represent the mean ± SD (*n* = 4). Statistical results are shown as \* *p* < 0.05. Values compared between the PBS group with CAF at different concentrations.

#### *2.4. Regulation of CAF on Osteoblastic Markers*

5 μM μM of CAF treatment CAF assessed on mechanistic influences related to the effect on mineralization and differentiation. It was decided to examine the quantification of mRNA and protein markers related to osteoblastic regulations (Figure 4). The protein estimation of cathepsin-B and BMP-2 in CAF treated osteoblastic cells was shown in Figure 4A. The estimation showed that, cathepsin-B protein was significantly increased at 5 µM CAF treated osteoblastic cells. Moreover, BMP-2 was also upregulated parallel to Cath-B protein (Figure 4A). Differential media increased the BMP-2 level 60% compared to control and CAF treatment increased the level up to 300% compared to control groups. These results claim to investigate more markers related to osteoblastic cells. Immunoblot based estimation of cathepsin-B, osteopontin, osteocalcin, and BMP-2 were estimated in CAF treated cells. The protein of Cath-B, osteocalcin, and BMP-2 were increased significantly in CAF treated cells (Figure 4B). On the other hand, osteopontin was not upregulated by CAF treatments. Moreover, 10 µM of CAF treatment upregulate the transcript marker of osteoblast-related morphogenic proteins, osteocalcin (250%), as well as the osteo-inductive protein BMP-2 (230%) (Figure 4C–F). The significance observation was also noted in mRNA estimation of osteoblastic markers. Osteopontin was not significantly increased against differential medium (DM) group, but it was significant against control group.

α and IL – it was quantified by quantitative real time PCR was normalized to β **Figure 4.** Effect of CAF on bone markers in osteoblast cell organization. (**A**) Cytokines estimation at 7th day of CAF treated osteoblastic cells. TNF-α and IL-10 were estimation using ELISA kits. (**B**) Effect of CAF on the protein expression of bone markers such as cathepsin-B (Cath-B), osteopontin (Opn), steocalcin (Ocn) and Bone morphogenic protein-2 (BMP-2). Protein expression were measured by Western blot analysis. (**C**–**F**) mRNA expression of bone markers on CAF treated osteoblastic cells and it was quantified by quantitative real time PCR was normalized to β-actin. (**C**) mRNA expression of cathepsin B, (**D**) mRNA expression of osteopontin, (**E**) mRNA expression of osteocalcin and (**F**) mRNA expression of BMP-2 marker, respectively. Bars represent the mean ± SD (*n* = 4). Statistical results are shown as \* *p* < 0.05.

#### *2.5. E*ff*ect of CAF on Osteoclastogenic Regulations*

to 10 μM regulate the osteoclast cell viability. At 5 and 10 μM concentration plays a 30% and 45% cell death membrane organization and DNA linearization. These results claimed that CAF at 10 μM Osteoclasts are major cell type in bone mass and play a critical role in bone rejuvenation and resorption. The negative regulation of osteoclastogenesis has recognized to be a positive treatment for bone loss and bone degenerative diseases. Therefore, it was examined for the anti-osteoclastogenic effect using CAF in RANKL-induced osteoclast cells. As shown in Figure 5A, results showed that CAF treatment concentrations tested from 0.5 to 10 µM regulate the osteoclast cell viability. At 5 and 10 µM concentration plays a 30% and 45% cell death, respectively. CAF treatment strongly inhibited the tartrate resistant acid phosphatase (TRAP) positive osteoclast cells by a dose-dependent manner. The inhibition of osteoclast differentiation was confirmed by TRAP activity (Figure 5B). It was further evaluated using a differentiation platform with RANKL (Figure 5C). As shown in Figure 5D, CAF exhibited significant DNA damage to osteoclast cells. The inhibition was profound nuclear membrane organization and DNA linearization. These results claimed that CAF at 10 µM concentration inhibits the osteoclastogenesis and accelerates RANKL induced DNA damage. Apparently, it was further examined to confirm the osteoclastogenesis mRNA and protein markers.

reduced at CAF 5 μM and 10 μM treate elevated at DM group compared to other tested groups. In addition, 10 μ Osteoclast marker investigation was done using mRNA and Western blot analysis. Western blot based quantification of TRAP, MMP-9, and cathepsin-K were estimated in CAF treated osteoclast cells. The significance observation was noted in mRNA estimation of osteoclast markers. TRAP expression was not appreciated in differential RANKL medium compared to control group, it was reduced at CAF 5 µM and 10 µM treated groups. The protein of TRAP, Cath-K, and MMP-9 were significantly decreased in CAF treated cells (Figure 5E,F). On the other hand, cathepsin-K was elevated at DM group compared to other tested groups. In addition, 10 µM of CAF treatment decreased the protein expression of MMP-9, these results revealed that CAF inhibit the migration and infiltration of bone cells. These results examined, showed that osteoclast markers were negative regulated by CAF treatment. It was significantly regulated by CAF compared to DM group. The inhibition of osteoclast genesis gives a remarkable output of *Trigonella* molecules reciprocally regulated the osteoblast and osteoclast cell types and might control the bone erosion and increases the reabsorption of senescent bone cells.

**Figure 5.** Effects of CAF on receptor activator of nuclear factor kappa-B ligand (RANKL)-induced osteoclast differentiation in mesenchymal cells. Bone-marrow derived mesenchymal stem cells (BM-MSCs) were cultured with vehicle or CAF (5 and 10 µM) in the presence RANKL (100 ng/mL) for 7 days. (**A**) Cultured cells tested against CAF treatment and analysed the cell viability. Cell viability of BM-MSCs was determined using the XTT assay (**B**) The estimation of Tartrate resistant acid phosphatase (TRAP) activity at 450 nm. (**C**) Cell differentiation effect was quantified on two intervals from 7 days to 10 days of incubation with CAF treatment. (**D**) RANKL induced osteoclast cells were treated with CAF and estimate the DAPI positive cells to confirm apoptotic induction as well as nuclear organization by fluorescence. Nuclei were stained by DAPI (blues signal). The area of DNA damage and nuclear organization was measured using ImageJ software. (**E**) Quantification of mRNA expression osteoclastic markers in CAF treated osteoclast cells using quantitative RT-PCR. The markers are TRAP, matrixmettalloprotease-9 (MMP-9), cathepsin-K (Cath-K). (**F**) Quantification of protein expression in osteoclastic cells treated with CAF using Immunoblot methods. The markers are TRAP, matrixmettalloprotease-9 (MMP-9), cathepsin-K (Cath-K). Bars represent the mean ± SD (*n* = 4). Statistical results are shown as \* *p* < 0.05.

#### *2.6. In Silico Docking of CAF against Cathepsin-B and Cathepsin-K Markers*

– − – − – Cathepsin is a cysteine residue based protease; ligand selection is based on its hot spot binding in substrate binding site of cathepsins family. These are the key factors for designing effective new ligand based inhibitors. Cathepsin-B structure showed arrangement of 18 amino acids long insertion from (Pro107-Asp124). In current study, orientations of the most potent inhibition of enzymatic activity of cathepsins-B and K were studied. It was found CAF docked similar binding regions with three-dimensional designed crystal structures of cathepsin-B and K (Figure 6A–G). The molecular modeling of CAF identified that a long hydrophobic pocket represents the potential binding site on the surface of cathepsin-K (−6.01) with two hydrogen bonds (Table 2), which is necessary for the peptide binding and excision (Figure 6B–D). The CAF docked the cathepsin-B in four hydrogen bonds, which at the site of hot spot of the receptor and made −4.43 binding energy (Figure 6E–G). Binding energy, ligand efficiency, and intermol energy were noted in Table 2.

– – **Figure 6.** In silico interaction of caffeic acid (CAF) with cathepsin family receptor molecules. Docked orientation of (**A**) CAF (CID\_637511), (**B**–**D**) 3D structure of cathepsin-K (PDB ID: 6QLM), hydrogen bonds between CAF with cathepsin-K, binding pocket of cathepsin-K. (**E**–**G**) 3D structure of Cathepsin-B (PDB ID:3AI8), hydrogen bonds between CAF with cathepsin-B, binding pocket of cathepsin B with CAF complex. Docking analysis was performed using Autodock tools (ADT) and Autodock v4.2 software.


**Table 2.** Hydrophobic interaction of caffeic acid and amino acid residues of target proteins.

Cat-K: Cathepsin-K, Cat-B: Cathespsin-B, LYS: Lysine, ASN: Asparagine, THR: Threonine, ILE: isoleucine.

#### **3. Discussion**

σ

doublet at σ agreement with the aromatic resonances at σ and 122.88, as well as two olefin carbons at σ resonance at σ TCE showed remarkable osteogenic activity and improves bone strength. There is a possibility that the TCE contains active molecules responsible for the potential activity. Further refining and identification of constituents revealed the isolation of CAF. The structure was elucidated by inspection of <sup>1</sup>H, <sup>13</sup>C, DEPT, HMQC, and HMBC spectroscopic data (Table 1 and Supplementary Material) and compared with the literature values [21]. Whereas the <sup>1</sup>H NMR spectrum of CAF (Table 1 and Supplementary Material) revealed the presence of aromatic resonances for an ABX system at σ<sup>H</sup> 7.06 (d, J = 2.04 Hz), 6.80 (d, J = 8.16 Hz), and 6.95 (dd, J = 2.04, 8.16 Hz). In addition, a pair of doublet at σ<sup>H</sup> 6.23 and 7.55 with coupling constants of 15.88 Hz was observed corresponding to two *trans*-olefin protons. By inspection of the <sup>13</sup>C NMR spectrum (Table 1 and Supplementary Material), the obtained resonances were in agreement with the aromatic resonances at σ<sup>C</sup> 127.82, 115.10,146.82, 149.47, 116.51, and 122.88, as well as two olefin carbons at σ<sup>C</sup> 115.55 and 147.06 and a carboxyl resonance at σ<sup>C</sup> 171.08, all of which were characteristic of caffeic acid. This study represents the first report on the isolation of CAF from *T. stellate*.

The obtained results establish the direct stimulatory effect of *T. stellata* metabolites on osteoblast differentiation. The potential stimulation of osteogenic properties of *T. stellata* metabolites has been proposed, but the molecular mechanism of this stimulation remains unclear. Natural products promote the osteoblast activity and suppress the osteoclast activity [22]. There are many studies from plants which showed the osteoblast activation and osteoclast inhibition processes [23–25].

–

The cytotoxic effects directly explained that the TCE was not toxic up to 100 µg/mL concentration. This biocompatibility on mesenchymal cells triggers us to investigate further on metabolites of *T. stellata*. The cytotoxic activity of CAF also explained its toxic properties and found nontoxic up to 10 µM concentration. The ALP and total protein were also reflecting this viability analysis. The results are comparable with study on medicinal plant's (*Leonurus sibiricus*) ethanol extract and other herbal molecules, which showed osteoblast differentiation and suppress osteoclast differentiation as well as bone resorption in a mouse model [26–29]. One of the studies reported with osteogenic differentiation by the compound leonurine hydrochloride from *Leonurus sibiricus*. Some other studies similar to current study observed with BMP-2 transcription factor, which played key role in bone formation and osteoblast differentiation [23,30]. The presented results demonstrate that CAF from TCE stimulated mineralization in osteoblast by enhancing transcription factors and stimulating cathepsin-B and BMP-2 signaling. Likewise, CAF suppressed RANKL-induced osteoclast differentiation and resorption capacity of cells.

The guava fruit was also effective against osteoporosis according to Chinese medicine due to the presence of polyphenolic compounds and increased bone health [31–35]. The results showed similar properties as *Rumex crispus* extract in osteoblast differentiation through runt related transcription factor 2 and suppress the RANKL induced bone loss through suppressing the RANKL signaling [36,37]. Furthermore, it was observed that the transcription factors were enhanced to activate the osteogenesis [37]. It was found that CAF enhanced the mRNA levels of stimulatory transcription factors to induce osteogenesis. BMP-2 is the osteoblast-regulating factor that induced morphogenic protein involved in differentiation; mineralization and bone strengthen mediators, such as osteopontin and osteocalcin. The mineralization explained the loading of calcium in osteoblast cells improves the bone strength and vitamin-D signaling pathways [38]. The loading of calcium was noted for 21 days and found 50% elevated storage after the 14th day of differentiation. In both tested concentrations got a similar response in differentiation and ARS uptake. These results revealed that, the osteoblast differentiation and mineralization were dose dependent manner.

It is well established that, extracellular signal regulated kinase (ERK) and p38-MAPK are osteogenic mediated signaling pathways that regulate osteoblast markers in unique differentiation. The underlying mechanism of CAF was investigated on the osteogenic activity, the study was examined whether CAF from TCE regulates the BMP-mediated activation of signaling markers and down regulated the TRAP, MMP-9, and cathepsin-K markers. These bone factors considered as major transcription factors required for bone formation, associated with the regulation of osteopontin and osteocalcin mediators [37–39]. In addition, BMP-2-mediated Smad protein stimulation regulate the transcription of osteopontin and osteocalcin markers [40,41]. In this study, CAF synergistically upregulate the BMP-2 and showed reciprocal regulation of cathepsin-B with cathepsin-K, upregulate the osteocalcin as well as bone transcription factors (Figures 5 and 6). These results suggest that the metabolic effect of CAF on BMP-2-dependent osteogenic to enhance the morphogenic protein mediated osteopontin and osteocalcin signaling axis for bone formation. It was showed that CAF treatments at low concentrations expressed an enhancement in differentiation and mineralization of osteoblast in both protein and mRNA level. Understanding the CAF effects on osteogenic markers, which delivers the outlined approach about the therapeutic potential for osteoporotic disorders. Furthermore, the regulations were confirmed by in silico docking analysis. The docking results revealed that, the structures of cathepsins-B and K are very similar, including their cleavage sites containing several "hot-spot" amino acids conserved among all types of cathepsins, along with Lysine 17 for cathepsin K and isoleucine 20 for cathepsin B receptors. The active site contains two histamine residues, which has high affinity towards substrate binding on receptors. These results revealed that CAF docked with cathepsin family receptors and plays a differential expressed ligand role in proteolytic protein osteogenic cell types.

#### **4. Materials and Methods**

#### *4.1. General Experimental Procedures and Chemicals*

<sup>1</sup>H-, <sup>13</sup>C-, and 2D-NMR spectra were measured on an AVANCE 400 NMR spectrometer (1H-NMR: 400 MHz and <sup>13</sup>C-NMR: 100 MHz, Bruker). Silica gel Column Chromatography (SCC) was performed on silica gel 60 (E. Merck, Darmstadt, Germany; 70-230 mesh). Reversed-Phase Silica Gel Column Chromatography (RPCC) was performed on a Cosmosil 75C18-OPN column (Nacalai Tesque, Kyoto, Japan; internal diameter = 50 mm, length = 25 cm, linear gradient: MeOH:H2O). Diaion HP-20 (Mitsubishi Chemical Co., Ltd. Tokyo, Japan). Pre-coated silica gel 60 F254 plates (E. Merck; 0.25 mm and 1000 µm in thickness) were used for Thin Layer Chromatography (TLC) visualization by spraying with p-anisaldehyde reagent and heated to 150 ◦C on a hotplate. High Performance Liquid Chromatography (HPLC) instrument used (Agilent, 1200 series, Germany) consisted of binary pumps, a PDA detector, and an auto sample injector, with a Chem satiation software module. The column was ZORBAX-SB-C18 (150 mm × 4.6 mm × 5 µm) (Agilent, Folsom, CA, USA). The mobile phase was in an isocratic mode using mixture of MeOH:H2O (45:55 *v*/*v*) with flow rate 1.0 mL/min. The volume of the samples injected into the HPLC system for analysis was set at 10 µL. Chemicals and reagents used were of analytical grade.

#### *4.2. Plant Material*

Aerial parts of *T. stellate* were collected from gardens of King Faisal University, Saudi Arabia (September 2016) and was identified by Dr. Mamdouh Shokry, Director of El-Zohria botanical garden, Giza, Egypt. Voucher specimen (9-16-Sept-TS) was deposited at the Herbarium museum of College of Clinical Pharmacy, King Faisal University, Saudi Arabia.

#### *4.3. Extraction and Isolation*

Shade-dried aerial parts of *T. stellata* (1.0 kg) were extracted three times with 70.0% methanol (10.0 L) by maceration at room temperature. The well-filtered extracts were combined and concentrated under reduced pressure to give an extract (103.0 g). The extract was partitioned with n-hexane (15.0 L) to give hexane fraction (74.0 g) and remaining mother liquor was concentrated to give (29.0 g) defatted extract. The defatted extract (25.0 g) was subjected to Diaion HP-20 column chromatography (1.0 kg) then it was eluted with water, 50.0% and 100.0% MeOH to obtain the fractions of water (5.0 g), 50.0% MeOH (12.0 g), and 100.0% MeOH (7.0 g). On the basis of the TLC patterns, the 50.0% MeOH-soluble fraction (12.0 g) was subjected to SCC (250.0 g, CHCl3:MeOH:H2O (15:6:1)(5.0 l)) then washing using 100.0% MeOH (2.0 l) to obtain seven sub-fractions (SubFr. 2-1 to 2-7). Sub-fraction 2-4 (1.3 g) was subjected to RPCC (100.0 g, using MeOH:H2O gradient elution) to yield six subtractions (SubFr. 2-4-1 (175.3 mg), SubFr. 2-4-2 (26.1 mg), SubFr. 2-4-3 (137.5 mg), SubFr. 2-4-4 (220.9 mg), SubFr. 2-4-5 (56.0 mg), and SubFr. 2-4-6 (330.7 mg)). SubFr. 2-4-4 (220.9 mg) was further fractionated on preparative TLC (CHCl3:MeOH:H2O (15:6:1)) the spot at R<sup>f</sup> value of 0.59 was eluted (SubFr. 2-4-4-1) and was further purified by HPLC to give pure compound; caffeic acid (CAF) (Figure 1B,C).

#### *4.4. Animals and Ethical Aspects*

Male C57BL/6j mice 4-week-old were used and weighing between 15 and 20 gm body weight, acquired from the Animal Facility of College of Science of King Faisal University. Animals were housed in cages under standard conditions and room temperature (22 ± 2 ◦C) under 12 h alternating light/dark conditions. Animal acclimatization were followed by the standard procedure of animal experimentation, and protocols were approved by the Ethics Committee on animal use of the College of Science (No. 180123).

#### *4.5. Culture of BM-MSCs*

Primary bone marrow derived mesenchymal cells (BM-MSCs) harvested from bone marrow of male C57BL/6j mice using serum free culture medium [RPMI-1640; UFG, Yanbu, Saudi Arabia). Mice were euthanized by humane control and dissect the femora bones in aseptic conditions following [42]. This isolation procedure was used in all experiments such as osteoblast, osteoclast differentiation studies, respectively.

#### *4.6. In Vitro Cytotoxicity Analyses*

#### 4.6.1. MTT Assay

The cytotoxicity of different concentrations of TCE (10, 25, 50, 100, 250, and 500 µg/mL) and CAF (0.5. 1. 2, 5, and 10 µM) were tested in vitro in osteogenic medium. After the 7th day of TCE and the 7th and 14th day of CAF exposures, cell viability was evaluated by the MTT and the NR uptake assays [43] briefly, aspirate the supernatant after CAF treatments, the MTT reagent was added with fresh serum free medium and incubate for 3 h. The formation of formazan crystals were solubilized by adding 0.1 mL dimethyl sulfoxide, and the optical density (OD) was determined at 570 nm.

#### 4.6.2. Neutral Red Assay

The effect of CAF on neutral red uptake using osteogenic cells treated with CAF and cells were aspirate with 0.5 mL of neutral red (3-amino-7-dimethylamino-2-methylphenazine hydrochloride) solution (50 µg/mL). Then, it was incubated for 2 h at 37 ◦C. Thereafter, the supernatant aspirated with PBS, neutral red dye was extracted using 0.1 % glacial acetic acid, and the OD was determined at 540 nm [44].

#### *4.7. Alkaline Phosphatase Activity*

Alkaline phosphatase activity was determined at the 10th day, using the Roy (1970) protocol [45], After CAF treatment, cell layers were scraped off using cell scrapper. Cell pellets were collected, and lysate was prepared by freeze–thaw method. After gently spinning down the cell debris, 20 µL supernatant from each sample were added to an assay mixture of p-nitrophenyl phosphate. Samples were quantified using 405 nm in Biotek elisa plate reader (Biotek Instruments Ltd., England). The specific activity of ALP was quantified by equilibrated total protein concentration.

#### *4.8. Assays of Osteoblast Di*ff*erentiation*

Osteoblast differentiation was induced by addition of differentiation medium (DM), supplemented with 50 µg/mL ascorbic acid, 10 mM β-glycerophosphate, and 10 nM dexamethasone (Sigma-Aldrich, St. Louis, MA, USA). The negative group was kept without additives [42]. Osteoblast differentiation was assessed by mineralization of calcium using the protocols of ARS dye. Osteoblasts were treated with CAF at concentrations of 0, 5, and 10 µM for 21 days. CAF treated osteoblast cells were washed with ice-cold PBS buffer and fixed in ice-cold 10% formalin for 20 min. Then, 1% alcian blue solution was used for fixation. These sections were incubated for 8 min with ARS. Mineralized cell patches observed and counted using an image analyzing system EVOS (Life Technologies, Carlsbad, CA, USA).

#### *4.9. Osteoclast Tartrate-Resistant Acid Phosphatase (TRAP) Activity Estimation*

The effect of CAF was examined on osteoclast differentiation isolated according to previously described method [40]. After the RANKL induction, cells were treated with CAF at 5 µM and 10 µM for 7 days, and results were analyzed by TRAP activity assay [26].

#### Osteoclast Apoptosis

The effect of CAF was evaluated on differentiation of pre-osteoclast cells isolated from bone marrow in mice. To generate osteoclasts, RANKL (150 ng/mL) was used for 4 days. Total TRAP activity was measured at an absorbance of 405 nm after treatment with Substrate (p-nitrophenyl phosphate) as described previously [26]. After treatment with CAF, results were obtained according to previously described protocol [46].

#### *4.10. Gene Expression Analysis by Real-Time Reverse Transcriptase Polymerase Chain Reaction (qRT-PCR)*

Total RNA extracted from CAF treated cells by the use of Trizol reagent (Invitrogen, Life Technologies, Grand Island, NY, USA). Extracted RNA were pretreated with DNAse I and quantified using Nanodrop (Thermo scientific, Waltham, MA, USA). Equiliberate the RNA concentration 300 ng/marker and used for complementary DNA preparation using SuperScript™III Reverse Transcriptase (Invitrogen, Waltham, MA, USA). Real time primer details were noted in Table 3. The relative mRNA quantification of cathepsin-B, cathepsin-K, MMP-p, BMP-2, TRAP, osteopontin, osteocalcin were evaluated by quantitative real time PCR (Applied Biosystems, Life Technologies, USA). Relative quantification of target genes were calculated using the 2−∆∆Ct method, as described by [47]. β-actin was used as an internal control.



#### *4.11. Western Blot for Protein Marker Quantification*

Protein quantification was done by Western blot analysis as described [48]. Briefly, CAF treated osteostromal cells such as osteoblast and osteoclast cells were washed with cold PBS. Cell pellets were collected by centrifugation and lysed by RIPA cell lysis buffer (Santa cruz biotechnology, USA). Cell lysates were obtained and total protein content in the supernatant was determined by the Bradford assay. It was eluted using SDS-PAGE, and the separated proteins were transferred to nitrocellulose membranes. Primary antibodies against cathepsin-B, cathepsin-K, MMP-p, BMP-2, TRAP, osteopontin, osteocalcin, and β-actin were diluted (1:1000; Cell Signaling, Danvers, MA, USA) Santa Cruz Biotechnology, Santa Cruz, CA, USA) at 4 ◦C overnight. Membranes were visualized using enhanced luminol-based chemiluminescent (ECL) substrate (Santa Cruz biotechnology, Santa Cruz, CA, USA).

#### *4.12. Cytokine Estimation by ELISA*

Production of tumor necrosis factor-α (TNF-α) and interlukin-10 (IL-10) in cell culture supernatant were evaluated. The BMP-2 and cathepsin-B were quantified in whole cell lysate of osteoblast treated with CAF using the manufacture protocol of Abcam validated kits (Abcam, Germany). Assays were performed on 96-well micro titer plates followed by manufacture instructions. The cytokine was quantified by the addition of chromogenic substrate solution (p-Nitrophenol). The reaction was measure at 450 nm. The standard curve was made in regression plot and sample concentrations were calculated in relation to the standard curve.

#### *4.13. Computational Docking Analysis*

The potential docking interaction was evaluated CAF against cathepsins-B and K were examined in silico using Autodock tools (ADT) v1.5.4 and Autodock v4.2 program (http://www.scripps.edu/mb/ olson/doc/autodock). The respective chemical structure of ligand CAF (CID\_637511), was retrieved from Pubchem database (http://www.ncbi.nlm.nih.gov/pccompound). The structures of cathepsin receptors were downloaded from the Protein Data Bank (PDB). The three dimensional structures of cathepsin-B (PDB ID: 6QLM) and cathepsin-K (PDB ID:3AI8) were retrieved from the Protein Data Bank; (http://www.pdb.org). The receptors were prepared by removing polar groups and non-amino acid residues using the protein preparation Wizard of PYMOL and python platform. The active sites of Cath-B and K were identified by Q-site Finder [49]. The competitive inhibition was studied on CAF complexed cathepsin protein family. Ligand docked to the receptor consider as rigid body, and receptor was considered as flexible factor. The ligand efficiency, intermol energy, and binding carbon numbers were recorded.

#### *4.14. Statistical Analysis*

The study was designed and evaluated in four independent experiments. The data values were expressed as means ± standard deviation (SD) (*n* = 4). Statistical analysis was conducted by student T test and one-way ANOVA regression plot compared with control group or treatment at different concentrations of TCE and CAF. All analyses were performed using Microsoft Excel office-10 and GraphPad Prism 7.0 (GraphPad Software Inc., San Diego, CA, USA), and statistical results are shown as the corrected p value (\* *p* < 0.05).

#### **5. Conclusions**

Based on the current findings, it can be concluded that CAF treatment showed promising activation on the bone metabolism via osteoblast differentiation and bone maturation in C57BL/6j bone marrow derived cells. At low concentration of CAF treatments, it enhanced the mineralization and upregulated the osteogenic marker expression in mesenchymal cells. CAF induced the morphogenic protein of bone and cathepsin-B in bone derived precursor cells. On the other hand, CAF attenuated the osteoclast formation and differentiation through the inhibition of osteoclast markers such as TRAP, Cath-K, and MMP-9 in RANKL induced osteoclast cells. Moreover, it regulates apoptotic bone loss in osteoclast cells. These findings recommend, CAF derived for the first time from *T. stellata*, might be used as potential therapeutic alternative for the treatment of bone metabolic diseases.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2223-7747/9/11/1610/s1, Figure S1–S8: 1 D- and 2 D-NMR spectroscopic data of CAF.

**Author Contributions:** H.-I.M.I.: conceptualization, funding acquisition, formal analysis, writing–review and editing and supervision, methodology; H.M.D. and M.R.A.: validation, software, and writing–original draft preparation, H.E.K.; conceptualization, methodology, writing—original draft preparation, formal analysis, writing—review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors extend their appreciation to the Deputyship for Research & Innovation, Ministry of Education in Saudi Arabia for funding this research work through the project number (IFT20094).

**Acknowledgments:** The authors are thankful to the Deputyship for Research & Innovation, Ministry of Education in Saudi Arabia for funding this research work through the project number (IFT20094).

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

#### **References**


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

© 2020 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*

### **Antibacterial Activity of** *Arbutus pavarii* **Pamp against Methicillin-Resistant** *Staphylococcus aureus* **(MRSA) and UHPLC-MS**/**MS Profile of the Bioactive Fraction**

**Nawal Buzgaia 1,2, Tahani Awin <sup>1</sup> , Fakhri Elabbar <sup>1</sup> , Khaled Abdusalam 2,3, Soo Yee Lee <sup>2</sup> , Yaya Rukayadi 2,4 , Faridah Abas 2,4 and Khozirah Shaari 2,5,\***


Received: 2 October 2020; Accepted: 3 November 2020; Published: 11 November 2020

**Abstract:** *Arbutus pavarii* Pamp is a medicinal plant commonly used by local tribes in East Libya for the treatment of many diseases, such as gastritis, renal infections, cancer and kidney diseases. In this study, the antibacterial activity of the leaf and stem bark extracts of the plant against methicillin-resistant *Staphylococcus aureus* (MRSA), as well as the metabolite profiles of the bioactive fractions, was investigated. The antibacterial activity was determined by disc diffusion method, minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC), while the microbial reduction by the bioactive fraction was evaluated using time–kill test. The bioactive fraction was further subjected to ultrahigh-performance liquid chromatography–mass spectrometry (UHPLC-ESI-MS/MS) analysis to putatively identify the chemical constituents contained therein. All the extracts and fractions showed different levels of antibacterial activity on the tested MRSA strains. The highest total antibacterial activity, i.e., 4007.6 mL/g, was exhibited by the crude leaf methanolic extract. However, the ethyl acetate fraction of the leaf showed moderate to significant antibacterial activity against MRSA at low MIC (0.08–1.25 mg/mL). Metabolite profiling of this fraction using UHPLC-ESI-MS/MS resulted in the putative identification of 28 compounds, which included phenolic acids, flavan-3-ols and flavonols. The results of this study showed that the ethyl acetate fraction of *Arbutus pavarii* leaf possessed potential antibacterial activity against MRSA and hence can be further explored for pharmaceutical applications as a natural antibacterial agent.

**Keywords:** *Arbutus pavarii*; antibacterial activity; MRSA; time–kill curves; ultrahigh-performance liquid chromatography; mass spectrometry (UHPLC- ESI-MS/MS)

#### **1. Introduction**

Millions of people are affected by contagious bacterial diseases throughout the world. These infectious diseases have persistently caused disability and death throughout mankind's history. According to the World Health Organization (WHO), approximately 50,000 people die

from bacterial infectious diseases throughout the world every year [1]. Methicillin-resistant *Staphylococcus aureus* (MRSA) is a group of Gram-positive bacteria that are distinct from other strains of *Staphylococcus aureus* [2]. MRSA is usually found in hospitals, prisons and nursing homes, where the people with open wounds and deteriorated immune systems are at greater risk of hospital-acquired infections. Although MRSA began as a hospital-acquired infection, it can be found in all communities and livestock. The terms HA-MRSA (healthcare-associated or hospital-acquired MRSA), CA-MRSA (community-associated MRSA) and LA-MRSA (livestock-associated) reflect the MRSA infections in a variety of hosts [3]. The MRSA displayed resistance against many antibiotics such as methicillin, a semisynthetic β-lactam antibiotic. Generally, the β-lactam mechanisms of resistance of MRSA strains support cross-resistance to all β-lactam antibiotics [4]. The key mechanism for resistance is the enzyme-catalyzed modification and ultimate destruction of the antibiotic, causing its dynamic efflux from cells and antibiotic target alteration [5]. Therefore, there is a high demand to develop antibiotics from natural sources based on medical plant extracts in a bid to back up the effectiveness and potency of conventional antibiotics [6]. Natural products play an important role in drug discovery, as evidenced by over 50% of all modern clinical drugs being of natural product origin [7].

Medicinal plants are rich sources of secondary metabolites with various biological properties, including antimicrobial properties [6,8]. *Arbutus pavarii* Pamp, an endemic medicinal plant species known locally as Shmar in Libya, is an evergreen shrub belonging to the Ericaceae family [9,10]. In folk medicine, it is used for the treatment of gastritis, renal infections, cancer ailments and kidney diseases [11]. Previous phytochemical studies on *A. pavarii* showed that this plant contains mainly flavonoids, tannins, glycosides, simple phenolics, triterpenes and sterols [11]. In addition, it was also reported that *A. pavarii* demonstrated strong antibacterial activity against several pathogenic bacteria [11]. However, few studies have focused on determining the effect of *A. pavarii* extracts and its fractions against resistant bacterial strains. Thus, the aim of this study was to evaluate the *A. pavarii* leaf and stem bark extracts against methicillin-resistant *Staphylococcus aureus* (MRSA). The anti-MRSA activity of the crude methanolic extract and various solvent fractions were assayed using disc diffusion assay, followed by minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) determinations, as well as time–kill curve analysis. In addition, the active fraction was subjected to ultrahigh-performance liquid chromatography–mass spectrometry (UHPLC-ESI-MS/MS) analysis for the identification of potential bioactive compounds.

#### **2. Results and Discussion**

#### *2.1. Antibacterial Activity Of A. pavarii Crude Extracts and Solvent Fractions*

Disc diffusion test was used to first screen the crude methanolic extracts and solvent fractions for presence of antibacterial activity. The appearance of zones of inhibition produced around the discs was observed, and their diameters were measured and recorded (Table 1). The standard antibiotic, 0.1% CHX, showed inhibition zones ranging from 7.00 to 10.33 mm against the bacterial strains. At the test concentration of 10 mg/mL, the crude methanolic extracts of the leaf and stem bark showed inhibition zones in the ranges of 8.00–9.67 mm and 7.00–10.00 mm, respectively. Among the different solvent fractions of the leaf, the EtOAc fraction showed the greatest activity towards all the bacterial strains, giving inhibition zones of 13.66, 12.00, 13.67 and 13.00 mm against MRSA ATCC 700699, MRSA KCCM 12255, MRSA1 and MRSA2, respectively. The same trend was observed for the stem bark fractions. However, compared to the leaf EtOAc fraction, the stem bark EtOAc fraction showed smaller inhibition zones of 8.00–9.00 mm, indicating that the stem bark either contained different bioactive constituents or lower amounts of the same bioactive constituents [12]. Other solvent fractions of the leaf and stem bark showed no to weak activities against the test bacteria. Previously, Alsabri et al. [13] investigated the antibacterial properties of solvent extracts prepared from the aerial part of *A. pavarii*. They reported that the methanol extract exhibited the highest activity against *S. aureus*, *Escherichia coli* and *Candida albicans*. The chloroform extract was active only against *S. aureus*, while the n-hexane

extract showed activity against *C. albicans*. Overall, these results indicated that the polarity of the solvent plays an important role in the extraction of the active ingredients and consequently in their potential antimicrobial activity.

The calculated relative inhibition zone diameter (RIZD) values of the test samples against the MRSA strains varied from 70.99% to 171.43%, as shown in Table 1. The RIZD value provides additional information showing the differential effects of the test extracts and fractions compared to the standard antibiotic used as a positive control. An RIZD value >100% means that the tested extract is more effective than the antibiotic. The leaf EtOAc fraction demonstrated the highest RIZD values, ranging from 125.85% to 171.43% against all MRSA strains. The higher RIZD percentages demonstrated by the leaf EtOAc fraction are a good indication that the leaf of *A. pavarii* contained most, and probably in higher amounts, of the antibacterial compounds of this plant species. It is worthy of note that the local population frequently uses the leaf material for medicinal purposes [11]. Based on the higher biological activity, the leaf and the stem bark EtOAc and n-BuOH fractions were subjected to further evaluation of their MIC, MBC and total activity values.

#### *2.2. Bacteriostatic (MIC) and Bactericidal (MBC) E*ff*ects of Bioactive Extracts and Fractions*

The antibacterial activity of the bioactive extracts and fractions was further investigated through the determination of the MIC and MBC values, as well as the total activity. The MIC and MBC values are presented in Table 2. The MIC values of the crude leaf methanolic extracts ranged between 0.08 and 1.25 mg/mL, while the MBC values ranged between 0.16 and 2.50 mg/mL. The leaf methanolic extract was more potent against the two standard MRSA strains, i.e., ATCC 700699 (MIC 0.08 mg/mL; MBC 0.16 mg/mL) and KCCM 12255 (MIC 0.63 mg/mL; MBC 1.25 mg/mL), in comparison to the clinical isolates, against which it showed MIC of 1.25 mg/mL and MBC of 2.5 mg/mL for both strains. A similar trend of potency was observed for the leaf fractions, where the standard MRSA strains were more susceptible to the fractions while the clinical isolates were less affected. The MIC and MBC for the EtOAc fractions were in the ranges of 0.08–1.25 mg/mL and 0.16–2.50 mg/mL, respectively; the MIC and MBC for n-BuOH fractions were 0.04–2.50 mg/mL and 0.08–5.00 mg/mL, respectively. In addition, among the activities exhibited by the leaf extract and fractions on the MRSA strains, the n-BuOH fraction showed the highest potency against the ATCC 700699 strain with MIC and MBC values of 0.04 and 0.08 mg/mL, respectively. The antimicrobial activity of an extract is considered very interesting and is of significant scientific value when its MIC values are lower than 100 µg/mL [14]. Hence, the present results revealed that the *A. pavarii* leaf methanolic extract and solvent fractions have moderate to significant activity against the tested MRSA strains.

On the other hand, the MIC values of the crude stem bark methanolic extract were lower than the leaf extract, ranging between 0.63 and 1.25 mg/mL, while the MBC values ranged between 1.25 and 2.50 mg/mL. The stem bark methanolic extract was more potent against the standard MRSA strain, ATCC 700699 (MIC 0.63 mg/mL; MBC 1.25 mg/mL), than against MRSA KCCM 12255 and the two clinical isolates as it showed MIC and MBC values of 1.25 and 2.5 mg/mL, respectively, against these three strains. In comparison, the stem bark EtOAc and n-BuOH fractions were less potent towards all the MRSA strains, except against the clinical isolate MRSA2 (MIC 0.63 mg/mL; MBC 1.25 mg/mL).

Overall, all the extracts and fractions showed different levels of antibacterial activity against the tested MRSA strains. This variation could be due to the different potencies of the bioactive compounds present in the extracts and fractions leading to different bacteriostatic and bactericidal effects on the bacterial strains, as reported by Qaralleh [15] and Oliveira et al. [16]. Several studies investigated the efficacy of plant extracts and their effective compounds as antibacterial agents to control infections by MRSA, suggesting that the bioactive component(s) of the plant extracts interact with enzymes and proteins of the bacterial cell membrane, causing its disruption, to disperse a flux of protons towards the cell exterior, which induces cell death or may inhibit enzymes necessary for the biosynthesis of amino acids [17].


**Table 1.** Inhibition zones of leaf and stem bark crude methanolic extracts and solvent fractions of *A. pavarii* against methicillin-resistant *Staphylococcus aureus* (MRSA) strains.

CH3OH = methanol extract, EtOAc = ethyl acetate fraction, n-BuOH = butanol fraction. Hexane and chloroform fractions showed no inhibition zones. MRSA1 and MRSA2 are clinical isolates, n.a = no activity (no inhibition zone detected). n.d = not detected. Diameter of inhibition zones in mm (including disc). Positive control: 0.1% CHX; negative control: 10% DMSO. Values are expressed as means±standard deviation (SD).


**Table 2.** Minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) values (mg/mL) of leaf and stem bark crude methanolic extracts and solvent fractions of *A. pavarii* against MRSA strains.

CH3OH = methanol extract, EtOAc = ethyl acetate fraction, n-BuOH = butanol fraction, CHX = 0.1% chlorhexidine (standard antibiotic). Hexane and chloroform fractions showed no inhibition zones. MRSA1 and MRSA2 are clinical isolates.

Besides MIC and MBC values, the antibacterial activity against the MRSA strains was also determined based on the total activity of the extracts and fractions. Total activity is defined as the volume to which the biologically active component (extracts, fractions or compounds) present in 1 g of dried plant material can be diluted and still kill the bacteria [18]. Total activity is useful for the selection of sample material for isolating bioactive compounds. Extracts or fractions with large total activity values are considered the best material for isolating potentially bioactive compounds. As shown in Table 3, the total activity values of the extracts and fractions of *A. pavarii* leaf and stem bark demonstrated high variation. The leaf methanolic extract diluted in 4007.60 mL of solvent can still inhibit the growth of MRSA ATCC 700699 (total activity: 4007.60 mL/g). The leaf n-BuOH fraction possessed higher total activity against MRSA ATCC 700699, with a value of 2235.89 mL/g. The leaf EtOAc fraction has higher total activity against MRSA2 and MRSA ATCC 700699, with 2158.97 and 1078.10 mL/g values, respectively. Both the extract and fractions of the stem bark exhibited lower total activity against all the tested MRSA strains as compared to leaf, with values ranging from 54.22 to 452.35.


**Table 3.** Total activity of leaf and stem bark crude methanolic extracts and solvent fractions of *A. pavarii* against MRSA strains.

#### *2.3. Time–Kill Curve for Ethyl Acetate Fraction of the Leaf*

A time–kill assay, using the four bacterial strains, was performed for the leaf EtOAc fraction since it exhibited a stronger antibacterial activity in comparison to the other fractions. Although MIC value gives a good indication of the efficacy of an antimicrobial agent, it provides limited information on the kinetics of the antimicrobial action [19]. A better method of assessing the bactericidal or bacteriostatic activity of an antimicrobial agent over time is by using time–kill kinetics assay, where the effect of various concentrations of the antimicrobial agent over time in relation to the growth stages of the bacteria is monitored [20]. The bacterial strains were thus exposed to the EtOAc fraction, at test concentrations of 0, 0.5, 1, 2, 4 and 8 × MIC over a period of 4 h, and the time–kill curve was plotted.

α

The assay results for MRSA ATCC 700699 (Figure 1A) revealed that the bacteria were completely killed after 4 h when a concentration of 4 × MIC (0.63 mg/mL) was used and after 2 h with the higher concentration of 8 × MIC (1.25 mg/mL). In terms of practical application, the 4 h killing time would be more preferred since the effect was obtained using a lower concentration (0.63 mg/mL) of the disinfecting agent. This condition is similar to that of a drug that exhibits a concentration-dependent bactericidal action, where the bactericidal effect is dependent on the dose of the leaf EtOAc fraction rather than on incubation time [21]. On the other hand, the time–kill curves for MRSA KCCM 12255 (Figure 1B) showed that the time–kill endpoint was achieved after 2 h incubation with a higher concentration of 8 × MIC (2.5 mg/mL). Meanwhile, in the case of the clinical isolates, as illustrated in Figure 1C,D, the time–kill endpoint could only be achieved with a concentration of 4 × MIC (5 mg/mL) after 1 h of incubation.

**Figure 1.** Time–kill curves for leaf EtOAc fraction against (**A**) MRSA ATCC 700699, (**B**) MRSA KCCM 12225, (**C**) MRSA1 and (**D**) MRSA2.

The data demonstrated that the bactericidal ability of the leaf EtOAc fraction is dependent on concentration and the bacterial strain. Generally, the time–kill kinetics results reasserted the expectation that a more concentrated sample will kill the microorganism in a shorter period of time. An increase in concentrations of plant extracts leads to an increase in the diffusion of phytochemicals into the cell membrane of bacteria, thus causing membrane destruction [22]. Furthermore, the bioactive compounds in the fraction may inhibit the synthesis of essential metabolites such as folic acid by preventing the enzymatic reaction. The protein synthesis in the microorganisms also can be inhibited if the bioactive compounds interfere and change the shape of the ribosome, which may lead to misreading of genetic code on mRNA [22]. The results of this time–kill kinetics study, together with the other results presented earlier, including disc diffusion assay, MIC, MBC and total activity determinations, reveal that the *A. pavarii* leaf possesses bacteriostatic and bactericidal effects against the tested MRSA strains, and the bioactive constituents could be largely present in the ethyl acetate fraction. Consequently, the EtOAc fraction was subjected to dereplication using UHPLC-MS/MS in order to gain an insight into the potential bioactive constituents.

#### *2.4. UHPLC-ESI–MS*/*MS Profile of the EtOAc fraction*

Several compounds from the classes of hydroxyquinone (arbutin), phenolic acid (caffeic, ferulic, gallic, rosmarinic, chlorogenic and salicylic acids), flavonoid (catechin, quercetin, dihydroquercetin, isoquercitrin, kaempferol, myricetin, rutin, naringin, neodiosmin, naringenin-7-*O*-glucoside, isovitexin-7-*O*-glucoside and delphinidin-3-*O*-rutinoside) and triterpenoid (oleanolic acid, lupeol and α-amyrin) have been previously reported to be present in *A. pavarii* [11,23,24]. In the present study, 28 compounds were putatively identified from the negative UHPLC-MS/MS spectrum of the leaf EtOAc fraction. The base peak chromatogram is shown in Figure 2, and compounds identified along with their spectral data are shown in Table 4. The results showed that the fraction was rich in phenolic compounds.

**Figure 2.** UHPL-ESI-MS/MS base peak chromatogram of the leaf EtOAc fraction of *A. pavarii* in negative ion mode.


**Table 4.** Compounds identified in the leaf EtOAc fraction of *A. pavarii*.


**Table 4.** *Cont.*

#### 2.4.1. Identification of Phenolic Acids and Derivatives

Compounds **2**, **4**, **5**, **6**, **7**, **8**, **11**, **14** and **16** were identified as gallic acid and its derivatives based on the presence of the aglycone fragment ion at *m*/*z* 169 and the characteristic fragment ions at *m*/*z* 271 and 211 in their MS/MS spectra [25]. Compound **5**, with a pseudomolecular ion at *m*/*z* 169.0131, was assigned as gallic acid, showing the characteristic base peak at *m*/*z* 125 for [M-H-CO2] −. Compounds **2**, **4** and **6**, eluting at three different retention times (0.78, 1.04 and 1.18 min, respectively), were identified as isomers of gallic acid hexoside (I-III). These compounds exhibited pseudomolecular ions at *m*/*z* 331.0668, 331.0669 and 331.0668, respectively, and all three produced a fragment ion at *m*/*z* 169 for [M-H-162]−, due to the neutral loss of a hexoxyl moiety. This agrees with previous reports by Mendes et al. [26] and Abu-Reidah et al. [27]. Meanwhile, compound **7** exhibited a pseudomolecular ion at *m*/*z* 343.0668. The compound was assigned as galloylquinic acid based on the presence of base peak at *m*/*z* 169 and fragment ion at *m*/*z* 125 for a further loss of CO2, all of which were characteristic fragment ions of gallic acid moiety [28]. Compounds **11** and **16** were identified as *di*-*O*-galloylhexose and *tri*-*O*-galloylhexose, respectively, based on similar fragmentation pattern showing losses of the corresponding number of galloyl moieties and the presence of a base peak at *m*/*z* 169 for the gallic acid aglycone.

Compound **8** has a pseudomolecular ion of *m*/*z* 315.0720, indicative of the molecular formula C13H16O9. It was identified as dihydroxybenzoic acid *O*-hexoside based on fragment ion at *m*/*z* 153 for [M-H-162]−, due to the loss of a hexoxyl moiety, and fragment ion at *m*/*z* 109 for [M-H-162-44]<sup>−</sup> indicating a further loss of CO<sup>2</sup> moiety, in agreement with Karar and Kuhnert [29]. Meanwhile, compound **14**, which exhibited a pseudomolecular ion of *m*/*z* 329.0878 and base peak at *m*/*z* 167 for [M-H-162]− for a neural loss of a hexoxyl moiety, was assigned as vanillic acid-*O*-hexoside. The assignment was supported by comparison with the fragmentation pattern previously reported by Morales-Soto et al. [30].

#### 2.4.2. Identification of Flavan-3-ol and Derivatives

Compounds **9**, **10**, **12**, **13**, **15**, **17**, **20** and **21** were identified as (epi)catechin and its derivatives based on the presence of fragment ions at *m*/*z* 289 and 125, corresponding to the (epi)catechin aglycone [31]. Compound **9**, which displayed a pseudomolecular ion at *m*/*z* 305.0663, was identified as (epi)gallocatechin based on the fragment ion at *m*/*z* 179 for [M-H-126]−, due to the characteristic loss of the trihydroxybenzene moiety [32]. Compound **10**, with pseudomolecular ion at *m*/*z* 451.1254, was identified as (epi)catechin-3-*O*-hexoside based on the fragment ion at *m*/*z* 289, for the loss of a hexoxyl moiety [33]. Compounds **13** (Rt = 5.44 min) and **15** (Rt 7.74 min) showed similar pseudomolecular ions at *m*/*z* 289.0714 and 289.0717, respectively. By comparison of their elution order with a previous study by Stöggl et al. [34], the compound eluted earlier was identified as catechin while the one eluted later was identified as epicatechin. Both compounds yielded the fragment ions at *m*/*z* 137 and 151 which were the results of retro-Diels–Alder (RDA) cleavage at ring C of the flavan-3-ol structure.

Three compounds (**12**, **17** and **21**) were identified as the dimeric forms of B-type proanthocyanidins (PAs), which could be differentiated from the A-type Pas with the extra 2 Da in their pseudomolecular ion [35]. Compound **12**, showing pseudomolecular ion at *m*/*z* 577.1334, was identified as the (epi)catechin + (epi)catechin. The compound also exhibited a fragment ion at *m*/*z* 425 ([M–H-152]−), which was due to the characteristic RDA cleavage at ring C of the dimer top unit [35]. Another fragment ion at *m*/*z* 407 ([M-H-152-18]−) due to the subsequent loss of a water molecule from the parent molecule was also observed. The presence of two other dimeric derivatives, **17** and **21**, was also indicated by the pseudomolecular ions at *m*/*z* 729.1458 (Rt = 4.03 min) and 729.1453 (Rt = 5.25 min). These compounds were identified as (epi)catechin gallate + (epi)catechin isomers based on the fragment ion at *m*/z 577 indicative of galloyl moiety losses ([M-H-152]−) from the parent ion [36]. Compound **20** at Rt = 5.20 min was identified as (epi)catechin-3-*O*-gallate. It displayed a pseudomolecular ion at *m*/*z* 441.0823. Its fragmentation pattern showed a fragment ion at *m*/*z* 289 for [M-H-169]−, which corresponded to a loss of gallic acid moiety via cleavage of the ester bond and loss of the (epi)catechin unit [37].

#### 2.4.3. Identification of Flavonols and Derivatives

The ethyl acetate fraction also contained the flavonol quercetin (**28**) and several of its derivatives (**18**, **19**, **23**, **25**, **26**, **27** and **28)**. Quercetin (**28**) was identified based on its pseudomolecular ion at *m*/*z* 301.03 and fragment ions at *m*/*z* 271, 255, 179 and 151 [36]. Compound **18**, with pseudomolecular ion at *m*/*z* 615.0997, displayed fragment ions at *m*/*z* 463 for [M-H-169]−, indicating loss of a galloyl moiety, and at *m*/*z* 301 for [M-H-331]−, indicating an additional loss of a hexoxyl moiety. Compound **18** was thus deduced to be quercetin-*O*-galloylhexoside, based on these data and data reported by Mendes et al. [26].

Compounds **19**, **23**, **25** and **26** were assigned as quercetin-3-*O*-deoxyhexosylhexoside, quercetin-3- *O*-pentoside, quercetin-3-*O*-deoxyhexoside and quercetin-3-*O*-hexoside. These compounds exhibited pseudomolecular ions at *m*/*z* 609.1463, 433.0775, 447.0931 and 463.0885, respectively. The transition of these ions to the aglycone ion (Y<sup>0</sup> −) at *m*/*z* 301 revealed the losses of the respective sugar moieties [34]. The glycosylation at the C-3 position of these compounds was determined by the higher relative abundance of their radical aglycone ion ([Y<sup>0</sup> H]<sup>−</sup> *m*/*z* 300) than the Y<sup>0</sup> − ion (*m*/*z* 301) [38]. Compound **27**, with a pseudomolecular ion at *m*/*z* 583.1099, was assigned as quercetin-*O*-(*p*-hydroxy)benzonylhexoside. The compound showed a fragment ions at *m*/*z* 463 for [M-H-120]−, indicating a loss of hydroxybenzoyl moiety, and *m*/*z* 301 ([M-H-282]−) for a further loss of hexoxyl moiety, in agreement with data reported by Jaiswal et al. [36].

Compound **22** was identified as myricetin-3-*O*-hexoside based on the presence of fragment ions at *m*/*z* 317, 316, 179 and 151, corresponding to the aglycone myricetin. The deprotonated aglycone peak observed at *m*/*z* 316.02 [M-H-162]<sup>−</sup> was due to the loss of a hexoxyl moiety [39]. Compound **24** with a pseudomolecular ion at *m*/*z* 447.093 was identified as kaempferol-3-*O*-hexoside. This compound showed characteristic fragment ion at *m*/*z* 285 due to the loss of sugar moiety and fragment ions at *m*/*z* 255 and 227 which are due to the loss of [M-162-CHO]<sup>−</sup> and [M-162-H2O-CO], respectively. Similarly, the fragment ions at *m*/*z* 179 and 151 were due to RDA cleavage of C-ring [39]. Attachment of the sugar moiety at the C-3 position of these compounds was also determined based on the relative abundance of [Y<sup>0</sup> H]<sup>−</sup> and Y<sup>0</sup> − ions [38].

#### 2.4.4. Identification of Other Compounds

Compound **1**, with a pseudomolecular ion at *m*/*z* 191.0555 (C7H12O6), was identified as quinic acid. It yielded fragment ions at *m*/*z* 171 ([M-H-H2O]−), 127.04 ([M–H-CO2-H2O]−) and 109 ([M–H-CO2-2H2O]−) [28]. Compound **3**, with pseudomolecular ion at *m*/*z* 271.0453 (C12H16O7), was identified as arbutin; it yielded a fragment ion at *m*/*z* 108 for [M-H-162]− due to loss of the hexose moiety [40].

The UHPLC-MS/MS results showed that the flavonoids and phenolic acid components are major secondary metabolites in the ethyl acetate fraction of *A. pavarii* leaf. Furthermore, among the identified compounds, several of them have been previously reported to possess antibacterial activity against MRSA. Shibata et al. [41] reported that gallic acid has antibacterial activity against MRSA with MIC value of 62.5 µg/mL. Catechins are often linked to antimicrobial effects associated with their interactions with the microbial cell membrane [42]. Cushnie et al. [43] reported that membrane disruption by catechins causes potassium leakage in MRSA strain, which is the first indication of membrane damage in microorganisms [44]. In addition, several studies have shown that the effectiveness of β-lactams can be enhanced by combining them with epigallocatechin gallate [45,46] and epicatechin gallate [47]. Meanwhile, Su et al. [48] reported that quercetin exhibited inhibitory effect against different MRSA strains, with MIC values ranging from 31.25 to 125 µg/mL, while rutin, a quercetin-3-*O*-deoxyhexosylhexoside, was reported to inhibit MRSA with MIC value of 250 µg/mL [49]. Besides, arbutin was reported to exert antibacterial activity against MRSA with MIC value of 10 mg/mL and MBC value of 20 mg/mL [50]. Therefore, the presence of these compounds, especially the flavonoids and phenolic acids, could have contributed significantly to the antibacterial activity of the leaf ethyl acetate fraction of *A. pavarii*.

#### **3. Materials and Methods**

#### *3.1. Plant Materials, Extraction and Fractionation*

The leaf and stem bark of *A. pavarii* were obtained from Al Jabal Al Akhdar region, Northeast Libya in March 2016 and identified by Dr. Abdulamid Alzerbi, a botanist at Biology Department of Benghazi University, Libya. The leaf and stem bark were dried under shade before being pulverized into a powder using a mechanical grinder (model: MX1100XT11CE, Waring, S/NoB 8643, Atlanta, GA, USA). The powdered plant material was sieved with a steel sieve (80 mesh) to obtain a uniform fine powder. For extraction, 1500 g of the ground leaf and 500 g of the stem bark were separately mixed with methanol at 1:10 solid-to-liquid ratio. The mixtures were sonicated at 35 ◦C for 60 min with a frequency of 53 kHz using an ultrasonic water bath (Branson, model 8510E-MTH, Danbury, CT, USA). The crude methanolic extracts were filtered (Whatman No. 1 filter paper, USA), and the collected filtrate was concentrated at 45 ◦C under reduced pressure using a rotary evaporator (Buchi, USA). The crude methanolic extracts were further fractionated using liquid–liquid fractionation to obtain solvent fractions of different polarities, namely hexane, chloroform, ethyl acetate (EtOAc) and n-butanol (n-BuOH) fractions (Merck, Darmstadt, Germany). The yields and physical appearance of the various extracts and fractions are tabulated in Table 5.


**Table 5.** Yields of extracts and solvent fractions of *Arbutus pavarii*.

#### *3.2. Bacterial Strains and Preparation of Inoculum*

MRSA ATCC 700699 was obtained from the American Type Culture Collection (Rockville, MD, USA) while MRSA KCCM 12255 was obtained from the Korean Culture Center of Microorganisms (Seoul, South Korea). Two clinical isolates (MRSA1 and MRSA2) were collected from the nasal swab of a 4th-year medical student from University Putra Malaysia, Malaysia. The MRSA strains are kept at the Laboratory of Natural Products (Institute of Bioscience, UPM, Malaysia). The MRSA ATCC 700699, MRSA KCCM 12255 MRSA1 and MRSA2 were grown on Mueller Hinton agar (MHA) (Difco, Franklin Lakes, NJ, USA) aerobically for 24 h at 37 ◦C, whereas inoculum cell suspension was prepared by transferring and incubating a single colony of each bacterial species in 10 mL of Mueller Hinton broth (MHB) at 37 ◦C overnight with 200 rpm agitation. Then, 1 µL of bacteria suspension was transferred to new MHB in a ratio of 1:10 to yield an inoculum size of 10<sup>6</sup> CFU/mL.

#### *3.3. Disc Di*ff*usion Assay*

Antibacterial activity was evaluated using agar diffusion assay, according to Rukayadi et al. [51]. Briefly, an inoculum of the bacterial strain was streaked on the surface of MHA plates using a sterile cotton swab. Sterile 6 mm filter paper discs (Whatman, Germany) were prewetted with 10 µL aliquot of the test extracts or fractions, prepared in DMSO at a concentration of 10 mg/mL. The discs were then placed on the inoculated plates at an appropriate distance from each other. Positive (chlorhexidine, 0.1% CHX, St Louis, MO, USA) and negative (dimethyl sulfoxide, 10% DMSO, Merck, Darmstadt, Germany) control discs were similarly prepared and placed on each test plate. Inoculated plates were subsequently incubated for 24 h at 37 ◦C and observed for inhibition zones. All experiments were conducted in triplicate, and inhibition zone diameter (IZD) was measured in mm. Antibacterial activity was expressed as the percentage of relative inhibition zone diameter (RIZD) with respect to standard antibiotic (0.1% CHX), according to Alsohaili and Al-fawwaz [52], and calculated using the following formula:

$$\text{V\%} \text{ RIZD} = [\text{IZD}\_{\text{sample}} - \text{IZD}\_{\text{negative control}}] \text{IZD}\_{\text{standard antibiotic}} \text{I} \times 100 \tag{1}$$

#### *3.4. Determination of Minimum Inhibitory Concentration (MIC) and Minimum Bactericidal Concentration (MBC) Values*

The MIC and MBC values of the test samples against the MRSA strains were established as described by the Clinical and Laboratory Standards Institute (CLSI) [19]. The determination was performed in a 96-well round-bottom microtiter plate (Greiner, Germany) using a 2-fold standard broth microdilution method with an inoculum of about 10<sup>6</sup> CFU/mL. The first well, designated as the negative control, was filled with 100 µL MHB. The second well, designated as the positive control, was filled with 100 µL of the bacterial suspension. A 100 µL aliquot of the test extract or fraction, prepared at a concentration of 10 mg/mL, was then added to the 12th well. Two-fold dilutions were then made from the 12th well down to the 3rd well. Therefore, the 12th well contained the highest concentration (5 mg/mL), while the 3rd well contained the lowest concentration (0.01 mg/mL). The plate

was then incubated aerobically for 24 h at 37 ◦C. After the incubation period, the MIC of the test extract or solvent fraction was determined. The MIC value is defined as the lowest concentration of the test sample that inhibited bacterial growth completely. For determining the MBC value, a 10 µL aliquot of the suspension in each of the 12 wells of the MIC determination was subcultured on an MHA plate. The plate was then incubated for 24 h, at 37 ◦C. After the incubation period, the plate was observed for bacterial growth and the MBC value was determined. The MBC is defined as the lowest concentration of the test sample that killed the bacterial strain completely. The MIC and MBC values were determined in duplicate. Chlorhexidine (0.1% CHX, St Louis, MO, USA) was used as a positive control. The antimicrobial activity of plant extracts may be expressed in different ways, including total activity values [18]. The total activity of the extract and the fractions was estimated as follows:

Total activity = Quantity of material extracted from 1 g of plant material/MIC (2)

#### *3.5. Time–Kill Curve*

Time–kill assay against the MRSA strains was performed according to Ramli et al. with slight modifications [22]. Briefly, the inoculum suspension ofMRSA was diluted to approximately 10<sup>6</sup> CFU/mL. The ethyl acetate fraction of the leaf was diluted with the MHB medium containing inoculum to obtain final concentrations of 0 × MIC, 0.5 × MIC, 1 × MIC, 2 × MIC, 4 × MIC and 8 × MIC for MRSA ATCC 700699 and MRSA KCCM 12255 and final concentrations of 0 × MIC, 0.5 × MIC, 1 × MIC, 2 × MIC and 4 × MIC for MRSA1 and MRSA2. Cultures (1 mL final volume) were incubated at 30 ◦C with 200 rpm agitation. At predetermined time points (0, 0.5, 1, 2 and 4 h), 10 µL aliquots were transferred to clean microcentrifuge tubes. The aliquots were serially diluted with 990 µL of 1% phosphate-buffered saline (PBS), and 20 µL was staked onto the MHA plates. The number of colonies formed on the plates after incubation at 30 ◦C for 24 h was counted and the number of CFU/mL was calculated. Assays were carried out in triplicate. The graph of log CFU/mL versus time was plotted as described by Ramli et al. [22].

#### *3.6. UHPLC-ESI-MS*/*MS Analysis*

The bioactive fraction was separated using a Hypersil Gold C18 reversed-phase column (2.1 × 100 mm, 1.9 µm, Thermo, USA) on a Thermos Scientific Ultimate 3000 (Bremen, Germany) with a mobile phase consisting of LCMS grade water (solvent A) and acetonitrile (solvent B), each containing 0.1% formic acid flowing at 0.4 mL/min. The programmed gradient system consisted of 0 min (95% A), 1 min (95% A), 20 min (5% A), 25 min (5% A), 25.1 min (95% A) and 35 min (95% A). The sample of 1 mg/mL (*w*/*v*) was prepared by dissolving 1 mg of a dried sample of the active fraction with 1 mL of methanol. The resultant mixture was then filtered using 0.22 µm Nylon membranes, and then 10 µL of the filtrate was auto-injected. The MS analysis was done on a Q-Exactive Focus Orbitrap LC-MS/MS system. The ESI-MS parameters were set as follows: negative mode, collision energy of 3.5 kV, capillary temperature 350 ◦C, auxiliary gas heater temperature 0 ◦C, sheath gas flow rate 40 arbitrary units and auxiliary nitrogen gas (99% pure) flow rate 8 arbitrary units. Then, the mass resolution was set to 70,000 full width at half maximum (FWHM) and a full scan of 150–2000 amu. The identification analysis was carried out by comparing the obtained MS/MS data with the literature.

#### *3.7. Data Analysis*

Microsoft Excel (Version 2010) was employed to perform the statistical analysis. Disc diffusion results were given as a mean ± standard deviation with three replicates.

#### **4. Conclusions**

In this study, the leaf and stem bark of *A. pavarii* were evaluated for anti-MRSA activity. The antibacterial activity was performed using disc diffusion agar test, MIC and MBC assays, in which the methanolic extracts and fractions of leaf and stem bark of *A. pavarii* demonstrated potential

antibacterial activity against the tested MRSA strains. Among the extracts and fractions, the EtOAc fraction of *A. pavarii* leaf revealed the highest antibacterial activity against all tested MRSA strains, with activity ranging from moderate to significant (MIC 0.08–1.25 mg/mL). In time–kill analysis, the MRSA strains were found to be completely killed after exposure to this fraction for 30 min to 2 h at 4× MIC and 8× MIC, revealing a remarkable capacity to inhibit or kill the MRSA strains. The UHPLC-ESI-MS/MS profiling of the bioactive fraction revealed that it contains high amounts of polyphenolic compounds. Phenolic acid and flavonoids were the main components and could be responsible for the bioactivity. The present findings add support for the traditional medicinal use of *A. pavarii* and highlight its potential as a source of natural antibacterial agents for future exploitation as natural antibiotics in the fight against MRSA prevalence.

**Author Contributions:** Conceptualization, K.S., Y.R., F.A. and N.B.; methodology, K.S., Y.R. and N.B.; investigation, N.B.; resources, K.A., F.E. and T.A.; data curation, K.S., Y.R., S.Y.L. and N.B.; Writing—Original draft preparation, N.B.; Writing—Review and editing, K.S. and S.Y.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The first author gratefully acknowledges support from the Ministry of Higher Education Libya for scholarship.

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

#### **References**


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

© 2020 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*
