Phenolic Profile and In Vitro Assessment of Cytotoxicity and Antibacterial Activity of Ziziphus spina-christi Leaf Extracts

Background: Ziziphus spina-christi is a tropical evergreen tree of Sudanese origin. The plant is commonly used in ethno-medicine for the treatment of many diseases such as malaria, digestive disorders, weakness, hepatic diseases, obesity, urinary problems, diabetes, skin infections, fever, diarrhea and insomnia. However, the pharmaceutical properties and bioactive compounds present in Z. spina-christi are still unknown. Therefore, the aim of the study was to significantly improve and deepen our knowledge about the phenolic composition of Z. spina-christi leaf extracts as well as their antibacterial and potential cytotoxic effects. Methods: The phytochemical profile of the known compounds in the methanol/water (7:3 v/v) extract of Z. spinachristi leaves were tentatively assigned on the basis of reported analytical data from three to four pieces of independent methods including retention time (tR), UV/Vis spectroscopy, high resolution mass spectrometry (HR-MS) and tandem mass spectrometry (MSn). Four bioactive compounds were purified via preparative-HPLC and identified by means of HPLC-ESI-TOF, HPLC-ESI-MSn, 1H-NMR, 13C-NMR, literature data and authentic reference standards. Antibacterial activities of aqueous and methanol leaf extracts were carried out with six bacterial strains (Bacillus subtilis, Bacillus aquimaris, Clavibacter michiganensis, Escherichia coli, Erwinia amylovora, and Pseudomonas syringae) using an agar diffusion assay. Cytotoxic effects towards the human epidermal keratinocyte cell line HaCaT and rat intestine epithelial IEC-6 cells was assess by cytoskeletal and plasma membrane integrity or analyses of mitochondrial and proliferative activities. Results: We tentatively assigned fifty-seven phenolic compounds (fifty-two of them are known) to their regioisomeric level in the methanol/water extract of Z. spina-christi leaves. To our knowledge 45 of them were not reported previously in Z. spina-christi and five for the first time in nature. Highly glycosylated flavonoids, proanthocyanidins, and chlorogenic acids were identified as the major components. Preparative-HPLC let to the isolation of four known phytochemicals including quercetin 3-O-(6-O-rhamnosyl-glucoside) (rutin) 26, quercetin 3-O-(2-O-rhamnosyl-arabinoside) 27, phloretin 3',5' di-C-glucoside 39 and quercetin 3-O-(2,6-di-O-rhamnosyl-glucoside) 50. Regarding the antimicrobial assays, the methanol extracts of the plant were found to show higher activity than the aqueous extracts indicating hydrophobic substances of Ziziphus leaves exerting antibacterial activity. The extracts revealed no cytotoxic effects towards the HaCaT or IEC-6 cells. Conclusions: Several previously known and unknown compounds were indentified in Z. spina-christi leaf extracts. The plant possesses strong antimicrobial activity and no toxicity. Our results therefore suggest that Z. spina-christi extracts might be potentially suitable for topical applications and support the use of the plant in traditional medicine to treat infectious diseases.


Background
Ziziphus spina-christi (L.) Desf. (Rhamnaceae) is a tropical evergreen tree of Sudanese origin. The plant has very interesting historical and religious aspects. It is repeatedly mentioned in Muslim as well as Christian traditions and was recorded by pilgrims visiting the Holy Land on numerous occasions. The boiled water extracts of the leaves of Z. spina-christi are used by Muslims in the cleaning of a dead body before burial suggesting antibacterial properties. In addition, the plant has been used in mummification by the ancient Egyptians [1]. It has been suggested that the plant material referred to in the Bible as the "thorns" or "bramble" (Judges 9; 14-15), "thorns" (Matthew 27: [27][28][29] and the "crown of thorn" (John 19:5) might have been derived from Z. spina-christi [2]. The Holy Quran mentions the Lote tree (Cedar) three times (XXXIV: 16; LIII: 13-18; LVI: [28][29][30][31][32], which was frequently identified as Z. spina-christi. Accordingly, this species is highly respected throughout the Middle East, has been widely used as a food and as medicinal as well as a environmental protection plant since antiquity, and is still in use at present [2,3].
Z. spina-christi is commonly used in ethno-medicine for the treatment of many illnesses such as digestive disorders, weakness, hepatic disorders, obesity, urinary problems, diabetes, skin infections, fever, diarrhea and insomnia [4,5]. In Sudanese ethno-medicine, the leaves of Z. spina-christi are used for the treatment of malaria [6]. In addition, Michel et al. reported an anti-diabetic activity of the leaves of Z. spina-christi due to their saponin and polyphenol constituents [7], which was supported in pharmacological studies by Glombitza et al. indicating that extracts of Z. spina-christi leaves or its main saponin glycoside, christinin-A, enhanced glucose consumption in diabetic rats [8]. Furthermore, Z. spina-christi leaves and fruits are reported to possess antibacterial activity [9], as well as antifungal activity on plant pathogens [10]. In addition, Adzu et al. found that root bark extracts showed significant antinociceptive activity in mice and rats [11].
The widespread and incorrect use of antibacterial agents has caused emergence of bacterial strains that are resistant to several antibiotics. Bacteria have developed different defence strategies to protect themselves from antimicrobial drugs such as alteration of the drug target, enzymatic inactivation, reduction of intracellular drug concentration by modifications in membrane permeability, or by efflux pumps [12,13]. Accordingly, new antimicrobials with greater effectiveness and better tolerability than existing drugs are urgently required for treatment of bacterial infections. Plant extracts are wellknown to combine a diversity of phytoconstituents that have a broad spectrum of antimicrobial activities, consequently plant-derived bioactive compounds counted as the main source for lead compound identification and development in pharmaceutical production [14].
Plant polyphenols comprise a large group of secondary metabolites that can extend from simple molecules, such as phenolic acids, to highly polymerized compounds such as tannins. They are constituents with a rich number of derivatives in the plant kingdom, from mosses and ferns to higher angiosperms including food plants. Many of these plantderived polyphenols are essential components in our diets [15]. They are derivatives of the pentose phosphate, mevalonic acid, shikimate, and phenylpropanoid pathways in plants [16].
Flavonoids have been reported to have numerous biological effects such as antioxidant activity [17], antimicrobial activity [18], antiinflammatory activities [19], inhibition of platelet aggregation [20], and inhibition of mast cell histamine release [21]. Moreover, antioxidant phenolics have been suggested to possess preventive functions in the progression of heart diseases and different cancers including prostate, breast, lung, colon and rectal cancers [22]. In addition, epidemiological studies have demonstrated that there is an inverse relationship between the intake of flavonoids such as myricetin, apigenin, quercetin, kaempferol and luteolin, which are naturally present in fruits, vegetables and beverages, and the lowering of the risk of suffering from cardiovascular diseases [23].
From different species of Ziziphus, peptide and cyclopeptide alkaloids, flavonoids, sterols, tannins, betulinic acid and triterpenoidal saponin glycosides have been isolated and chemically identified [24,25]. From the methanol extract of the fruits of Z. jujube and Z. spina-christi twelve compounds have been reported as representing quercetin, kaempferol, and phloretin derivatives [15]. In a qualitative as well as quantitative study, flavonoids, saponins, and triterpenic acids were isolated from the leaves of two Ziziphus species [26].
Herein, the aim of the present work was to significantly improve and deepen our knowledge about the phenolic composition of Z. spina-christi leaves using liquid chromatography coupled to electron spray ionization multi-stage mass spectrometry (HPLC-ESI-MS n ) as a powerful tool for the analysis of natural products and to evaluate the antimicrobial and cytotoxic effects of the plant extracts. The obtained results may contribute to a better understanding of the role of Z. spinachristi phenolics on biological, nutritional and medicinal properties.

Plant material
The Z. spina-christi plant samples were freshly collected in June 2010 from its natural habitats in Omdurman (15° 38 '

Sample preparation
Five grams of the shade-dried green leaves of the plants were ground in a hammer mill to fine powder and extracted with methanol/ water (7:3 v/v) using a Soxhlet apparatus (Buchi Extraction System B-811/B-811 LSV, Germany) for 5 h. These extracts were filtered through a Whatman no. 1 filter paper. The solvent was removed by evaporation in vacuo, and the extracts were stored in air-tight vial at -20 °C until required for the photochemical analysis and cytotoxicity assays. For the photochemical analysis, the dried extract was thawed at room temperature, dissolved in methanol/water (7:3 v/v; 60 mg/10 ml), filtered through a membrane filter with a pore size of 0.45 μm (Carl Roth, Germany) and used directly for LC-MS.
For determining the antimicrobial activity, two additional methanol and aqueous extracts were prepared in a similar way. The methanol extract was concentrated to dryness using rotary evaporator in vacuo to obtain the crude methanol extract. The aqueous extract was concentrated and lyophilized. The respective extracts obtained were kept in air-tight vial at -20 °C until use.

UV Irradiation
UV irradiation experiments were performed as previously reported [27].

Liquid chromatography-mass spectrometry (LC/ESI/MS n )
The Agilent 1100 LC equipment (Agilent, Karlsruhe, Germany) comprised a binary pump, an auto-sampler with a 100-µl loop, and a diode array detector with a light-pipe flow cell (recording at 254, 280, and 320 nm). This was interfaced with an ion-trap mass spectrometer fitted with an electrospray ionization source (Bruker Daltonics HCT Ultra, Bremen, Germany) operating in full-scan, auto-MS n mode to obtain fragment ion m/z. Tandem mass spectra were acquired in auto-MS n mode (smart fragmentation) using a ramping of the collision energy. Maximum fragmentation amplitude was set to 1 V, starting at 30% and ending at 200%. MS operating conditions (negative mode) had been optimized using procyanidin B1 and 5-O-caffeoylquinic acid with a capillary temperature of 365°C, a dry gas flow rate of 10 l/min, and a nebulizer pressure of 10 psi. The spectra full scan mass were performed within the range from m/z 50 up to 1500 in negative ion mode.
High-resolution LC-MS was performed using the same high performance LC equipped with a micrOTOF mass spectrometer (Bruker Daltonics, Bremen, Germany) fitted with an electrospray ionization source, and internal calibration was achieved with 10 ml of 0.1M sodium formate solution injected through a six-port valve prior to each chromatographic run. Calibration was performed using the enhanced quadratic mode.

High-performance liquid chromatography
The HPLC separation was were performed as previously reported [28].

Antibacterial activity assay
Three Gram-positive bacterial strains (Bacillus subtilis S168, Bacillus aquimaris MB-2011, and Clavibacter michiganensis GSPB 390 as well as three Gram-negative bacterial strains (Escherichia coli DH5α, Erwinia amylovora 1189, and Pseudomonas syringae pv tomato DC300) were selected as model organisms to evaluate the antibacterial activity of the methanol and aqueous Z. spina-christi crude plant extracts at a final concentrations of 100 mg/ml. The antimicrobial activity assay was performed using the agar diffusion method according to Nathan et al. with slight modification as follows: Lysogeny broth (LB) agar plates were inoculated with 200 µl of the tester organism (1 × 10 7 cfu/ml) by spreading the bacterial cell suspensions over the agar with the help of sterile glass beads. Holes (5 mm diameter) were punched into the agar with a sterile Pasteur pipette, and 50 µl of the crude extracts was pipetted into the wells [30]. The plates were incubated overnight at 28°C except for E. coli, for which incubation was done at 37°C. Inhibition of microbial growth was determined by measuring the radius of the zone of inhibition. For each bacterial strain, an equal volume of water and methanol was used as negative controls. As positive controls, the following antibiotics were tested effective against the following bacterial organism causing inhibition zones of 10.3 to 23.3 mm: 25 mg/ml kanamycin (grad ≥ 750 IU/mg; Carl Roth, Germany) for B. subtilis and B. aquimaris; 50 mg/ml ampicillin (Carl Roth, Germany) for E. coli and E. amylovora; 25 mg/ml streptomycin (grade 730 IU/mg; Serva, Germany) for P. syringae and C. michiganensis. The experiments were conducted in triplicates, and the data are given as mean values ± standard deviation (SD). Each mean value was calculated from three biological replicas and each biological replica was calculated from three technical replicas. The SD was calculated from the mean of the three biological replicas. Data were analyzed with one-way analysis of variance (ANOVA) using Origin lab software.

Cell culture
HaCaT cells at passages 40 -55 were used as a model for epidermal keratinocytes [31,32]. They were cultured in Dulbecco's Modified Eagle Medium (DMEM, DMEM, Lonza Biowhittaker, Switzerland) containing phenol red and supplemented with 10% fetal bovine serum (FBS; Lonza Biowhittaker). The IEC-6 cell line at passage 16 -40 was used as cellular model of small intestine epithelial cells (European Collection of Cell Cultures, ECACC, UK), and cultured in DMEM containing 0.1 µg/ml insulin (Sigma-Aldrich, USA) and 5% FBS. Cells were grown at 37 °C in a humified atmosphere at 8% CO₂ for keratinocytes and at 5% CO₂ for IEC-6 cells. Cell culture medium was exchanged every three to four days, confluent cell cultures were used for experimentation.

Cellular integrity analyses by acquisition of propidium iodide (PI) and counter-staining with Draq5™
PI (Carl Roth, Germany) acquisition through ruptured plasma membranes was used as a measure of necrotic cell death, while Draq5™ staining of cellular DNA served to determine total cell numbers [33]. Cells were grown on cover slips for 24 hours and then treated with 50 µl of the plant extract at final concentrations of 100 µg/ml, 10 µg/ml, and 1 µg/ml (weight/volume, final concentrations). Controls were treated with 0.5% DMSO, since all extracts were diluted with DMSO to this final concentration. After incubation of the cells with extract preparations for 24 hours at 37 °C, cells were washed with prewarmed PBS. Staining with 2 µg /µl PI (Carl Roth, Germany) diluted in the respective culture medium was for 30 minutes at 37 °C. Afterwards cells were rinsed with PBS twice before fixation with 4% paraformaldehyde (Applichem, Germany) in 200 mM Hepes, pH 7.4 for 20 minutes at RT. Then, cells were rinsed in PBS and stained with 5 µM Draq5™ (Biostatus, UK) in PBS for 45 minutes at 37 °C. Finally, cells were washed three times with PBS and then mounted in mowiol (33% glycerol and 14% Mowiol (Carl Roth, Germany) in 200 nM Tris-HCL in ultrapure H₂O at pH 8.5).
Images were taken at a Zeiss LSM 510 Meta confocal laser scanning microscope (Carl Zeiss, Germany) and analyzed using Zeiss LSM Image Browser software version 3.2.0.70.

Staining of the actin cytoskeleton with FITC-Phalloidin and counter-staining with Draq5™
Cells were fixed as described above and stained with 3 µM FITC-Phalloidin (Sigma-Aldrich, Germany) and 5 µM Draq5™ in PBS for 45 minutes at 37 °C [34]. Cells were washed three times before mounting in mowiol.

Determination of cell counts using the Image Analysis Software CellProfiler™
Images were analyzed using Cell Profiler 2.0 [35] as previously described [31,33,34].

MTT Assay
Cells were grown for 24 hours at 37 °C and the respective CO 2 concentration in 24-well plates before treatment with 50 µl plant extracts per 1 ml culture medium to achieve final concentrations of the extracts to reach 100, 10 and 1 µg/ml (wt/vol), respectively. As controls, cells were treated with 0.5% DMSO and mock-treated cells were used, to which no extracts were applied. MTT ([3-(4, 5-Dimethylthiazol-2yl)-2, 5-diphenyltetrazolium bromide; Carl Roth, Germany) was applied in PBS to the respective cell culture medium at a final concentration of 0.5 mg/ml MTT. Cells were incubated in the dark for 1.5 hours at 37 °C and the respective CO 2 concentration, before formazan crystals were dissolved in 500 µl of 100% DMSO and quantified at 570 nm using a Thermo Spectronic Genesys 10 UV-vis Spectrophotometer (Thermo Scientific™, USA).

Results and Discussion
Methanol/water (7:3 v/v) extract of the leaves of Z. spina-christi was separated by reversed-phase chromatography. Identification of compounds was obtained using high resolution-electrospray ionizationmass spectrometry (HR-ESI-MS) for determination of molecular formula and ESI with quadruple ion trap analyzer MS (ESI-QIT-MS) to obtain tandem MS data for further structure elucidation. Additionally, retention times and UV (320 nm) spectra were used for reversed phase compounds identification ( Figure 1). Identification of phenolic compounds, for which commercial standards were available was carried out by the comparison of their retention times, UV-vis spectra and mass spectral data recorded in negative ion mode, while, the identity of other compounds was elucidated using the UV-vis spectrum to assign the phenolic class [36]. For all compounds the high resolution mass data were in good agreement with the theoretical molecular formulas (Table 1S, Supporting information), all displaying a mass error of below 5 ppm, thus confirming their elemental composition. Moreover, the chromatographic elution order helped in some structural assignments as it was described previously [37]. The phytoconstituents characterized in Z. spina-christi are presented in Table 1 and their chemical structures are shown in Figure 2.
In the following section structure assignment of selected compounds is illustrated. Full assignment arguments are provided in the Supporting information.

Gallocatechin isomers (m/z 305)
At retention times 10.5 and 17.3 min, two peaks with m/z 305 were detected in the EIC and base peak chromatogram (BPC) and were assigned as gallocatechin 8 and epigallcatechin 9 (Table 1) after comparison with the retention time and fragmentation pattern of gallocatechin and epigallcatechin commercial standards.

Catechin isomers (m/z 289)
Two peaks (t R 17.8 and 24.0 min) were detected at m/z 289 in the EIC and were identified as catechin 10 and epicatechin 11 ( Table 1)  After irradiation, we found that the cis isomer in the chromatogram as peak with considerably increased intensities if compared to trans isomer from the original plant extract ( Figures 7S and 8S, Supporting information), which confirmed the presence of the cis-5-O-caffeoylquinic acids 13 [27].

Characterization of caffeic acid 4-O-glucoside (m/z 341)
One peak was readily detected at m/z 341 in the EIC and was tentatively assigned as caffeic  [44] and found that they are identical. Based on the above arguments compound 17 was identified as caffeic acid 4-O-glucoside. We have recently identified 6-O-caffeoyl-α-glucose and 6-O-caffeoyl-β-glucose in the bottle gourd (Lagenaria siceraria Stand) fruits [45]. Here we are unable to assign the anomeric structure for the glucose moiety.

Characterization of phloretin derivatives (m/z 597)
Three C-glycosides derivatives of phloretin (compounds 38-40) were detected in the EIC of Z. spina-christi leaves extract (Table 1) Figure 11S, Supporting information). Compound 40 was identified as phloretin 3', 5'-di-C-glucoside. This compound has been previously reported in Ziziphus species [15]. We speculate that compounds 38 and 39 were isomers of compound 40 alternative C-glycosylated hexoses. From the literature and our experiments we have found that the glycosides containing galactoside units are more polar than the glycosides containing glucoside units [46]. Based on the above arguments we have tentatively assigned isomer 38 as phloretin 3', 5'-di-C-galactoside (t R 32.6 min) and isomer 39 as phloretin 3'-C-glucoside 5'-C-galactoside or phloretin 3'-C-galactoside 5'-C-glucoside (t R 36.8 min) ( Table 1). To our knowledge, compound 38 was not previously reported in the plant kingdom.

Characterization of diosmetin derivatives (m/z 769)
Two peaks, 43 and 44, were detected at m/z 769 in the EIC and showed different retention times (t R 54.4 and 56.6 min). They showed in their MS 3 fragment ions at m/z 299 (base peak), 255 and 284 corresponding to diosmetin aglycone [47], which suggested that, these compounds were diosmetin glycosides (Table 1). Based on the elution order and the data previously reported [48][49][50]

Characterization isorhamnetin, kaempferol and quercetin derivatives
Di, tri, tetra and penta glycosylated flavonoids were also detected in Z. spina-christi methanol/water extract. The presence of the aglycone fragment ions at m/z 315, 285 and 301 in their MS 2 and/or MS 3 , and characteristic fragment ions (m/z 151 and 179 for quercetin, m/z 151 for kaempferol and m/z 300 for isorhamnetin) [51], in addition to their UV spectra suggested that these compounds were isorhamnetin, kaempferol and quercetin derivatives [52,53].

Characterization of the isolated constituents
In our present work, preparative HPLC was used after the purification on C 18 solid phase extraction columns to isolate four known compounds from the methanol/water (7:3 v/v) extract of Z. spina-christi leaves. The structures of the isolated compounds were elucidated by HRMS (Table S1 in the Supporting information), tandem MS (Table 1 and Figure 4 (Figure 2). Compounds 26, 27 and 39 have been previously isolated from the fruits of Ziziphus species [15]. The UV spectra, MS fragmentation and the retention time of compound 50 were compared with the compound (m/z 755) reported in the leaves of Catharanthus roseus and found that they are identical [55]. The purity of these compounds was determined as in our previous study [29] by total ion chromatograms in negative ion mode and UV chromatograms at 280 and 320 nm. The purity of the compounds were 86% for compound 26, 75% for compound 27, 56% for compound 39 and >98% for compound 50.

Antibacterial activity
The antibacterial activities of the plant extracts analysed in this study were evaluated for their efficacies against Gram-positive and Gram-negative bacteria using the agar diffusion method. As suitable model organisms B. subtilis S168, B. aquimaris MB-2011, and C. michiganensis GSPB 390 were chosen for Gram-positive and E. coli DH5α, E. amylovora 1189, and P. syringae pv tomato DC300 for Gramnegative. The results of the antibacterial activity tests of aqueous and methanol extracts of the leaves of Z. spina-christi are given in Table  2 and Figure 5. Interestingly, the plant extracts tested herein showed antibacterial activity only against the Gram-positive bacterium B. aquimaris MB2011 and the Gram-negative representative P. syringae pv tomato DC300 while the other bacterial strains were not affected at all. These results are in contrast to those of others [56]      In line with our findings, it has been shown before that many plant metabolites including flavonoids, phenols, tannins, and alkaloids exhibit moderate antimicrobial activity [57]. Our data propose that antibacterial activities might be associated with the phenolic compounds from Z. spina-christi leaves, which were identified in the aqueous and methanol extracts. Our data furthermore indicated that methanol extracts appeared to be more effective in antibacterial activity against Gram-negative bacteria as compared to the aqueous extracts of the leaves from Z. spina-christi suggesting that the bioactive metabolites of Z. spina-christi were rather hydrophobic than hydrophilic. However, B. aquimaris MB-2011 was similarly affected by the aqueous and the methanol extracts suggesting that it is either affected by at least two differently extractable compounds or that it is susceptible to a compound different from that effective against the Gram-negative strain P. syringae. Thus, if traditional use of the leaves of Z. spina-christi involves the preparation of a pre-extract in alcoholic solvents, their effectiveness may be more pronounced than simply brewing the leaves in water [3].

Cytotoxic activity
In order to use Z. spina-christi extract as a potential antibiotic and to enable safe use as topical treatment of skin or as systemic drug taken up by oral administration routes, cytotoxicity against the exposed cells and tissues must be analyzed in order to exclude potential negative, offtarget effects during treatment. Here, we have used human keratinocyte (HaCaT) and rat intestine epithelial (IEC-6) cell lines in culture to assess potential cytotoxicity exerted by leaf extracts from Z. spina-christi in vitro. For all experiments, the cells were first grown to confluence with the aim to simulate the effect of the extracts on cell monolayers which are considered representative of intact epithelia of the respective tissue, i.e., epidermis of the skin and mucosa of the gastro-intestinal tract lining.
Cytotoxicity was estimated by using the 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay, which is a routine test evaluating cell viability and proliferative activity by means of cellular enzyme activity. The MTT is converted by mitochondrial NADH-dependent dehydrogenase into an insoluble dye made of dark blue formazan crystals [58]. MTT assays are suited to determine the activity of mitochondrial enzymes reducing the yellow tetrazole MTT to purple formazan in living cells, only [59]. Therefore MTT conversion is proportional to both, cell viability and proliferation rates of treated cell cultures.
The MTT assay has been applied by us and others before. In comparison with other cytotoxicity assays such as the neutral red assay or quantification of LDH latency, the MTT assay was suitable but less sensitive than the former when CdCl 2 effects onto hepatoma cell lines were determined [60]. In another study, the above mentioned three different assays were compared and additionally ATP contents of the cultured cells were determined [61]. While both, the MTT and the LDH assay were somewhat non-suitable because of their sensitivity to inhibitory agents contained in the compound mixtures tested, it was also pointed out that the choice of assay depends on the pathway of death that is potentially initiated by the agents applied.
Here, we used the MTT assay to recapitulate the well-known cytotoxicity of DMSO by incubating cultures of HaCaT keratinocytes and IEC6 intestine epithelial cells with different concentrations thereof ( Figure 20S, Supporting information). The results demonstrate that indeed DMSO is cytotoxic if applied at concentrations above 1%, because MTT conversion declines in a concentration dependent manner. However, the data also support the notion that concentrations of up to 0.5% DMSO are non-cytotoxic, and can be used to as solvent of lipophilic compounds.
We have used this assay before to determine cell viability upon exposure of fibroblasts, keratinocytes, and intestine epithelial cells to other stressors in addition to testing for the safe use of e.g., potential drug delivery tools [33,34,62]. The conversion of MTT by HaCaT keratinocytes ( Figure 6A) is not influenced upon exposure to 100 µg or 10 µg/ml Z. spina-christi methanol/water (7:3 v/v) extract. However, upon treatment with 1 µg/ml Z. spina-christi conversion of MTT was decreased in HaCaT cell cultures. On the other hand the MTT conversion by small intestine epithelial IEC-6 cells ( Figure 6B) increased with decreasing concentration although this was not statistically significantly different from controls. For the lower concentrations of 10 µg and 1 µg/ml the values of conversion were approximately even. We conclude that cytotoxicity towards IEC-6 cell cultures increased with increasing concentration of Z. spina-christi, and it was safe to apply extracts at a final concentration of 1 µg/ml to intestine epithelial cell monolayers. In contrast, for HaCaT cells the exposure to 1 µg/ml Z. spina-christi extracts, only exerted cytotoxic effects that were detectable by a significant reduction of the MTT-conversion to approx. 75% in comparison to DMSO-treated controls.
In addition, MTT-and LDH-assays are usually backed up by additional tests that involve determination of cellular integrity as we have shown in a study in which the MTT assay was employed to test for cytotoxicity of dust samples applied to keratinocytes and fibroblasts [34]. Results achieved by the MTT assay were comparable with culture impedance measurements. In addition, phalloidin staining of the actin cytoskeleton and propidium iodide staining of the nuclei of cells with ruptured plasma membrane were performed in our previous work, again, supporting the results of the MTT-assay with regard to structural and metabolic integrity of the cells, respectively. In another study, the MTT assay was used by us to determine IC 50 values of cucurbit [7] uril which were backed up by imaging studies and supported by in vivo toxicity studies revealing highly congruent data [62]. Therefore, we believe that the MTT assay, especially when used in combination with additional measures of cytotoxicity, is a suitable method to conclude on cytotoxic effects of diverse samples such as dust simulants, drug delivery agents, or plant extracts as applied in this study.
Consequently, here we performed propidium iodide (PI) staining for both, HaCaT keratinocytes and IEC-6 intestine epithelial cells to determine the extent of potential plasma membrane rupture and the number of necrotic cells per viable cells when the cultures were treated with extracts of Z. spina-christi. This assay is based on the fact that PI intercalates with DNA, thereby visualizing nuclear DNA when PI passes ruptured membrane lipid bilayers. Therefore PI may stain for necrotic cells in which the lipid bilayer integrity of the plasma membrane is compromised [34]. With increasing concentration of Z. spina-christi extracts, the number of necrotic or ruptured cells increased for the exposed HaCaT cultures ( Figure 7A). For the IEC-6 cell cultures ( Figure 7B), the number of ruptured cells decreased upon exposure to leaf extract of Z. spina-christi. However, the numbers of ruptured cells in HaCaT or IEC6 cell cultures treated with Z. spina-christi leaf extracts remained moderate as they were determined to not even reach one percent upon 24 h of exposure.
As an additional and sensitive measure of cellular integrity, the actin cytoskeleton was analyzed using a Phalloidin Draq5™ staining.
Upon treatment with 100 µg/ml leaf extracts of Z. spina-christi, HaCaT cells ( Figure 8) showed a slightly altered actin filamentous system, and some cells detached from their neighbors indicating loosening of cellcell contacts. For the IEC-6 cell cultures (Figure 8), the cortical F-actin system remained intact upon treatment with different concentrations of the extracts prepared from Z. spina-christi.
In summary, mild cytotoxic effects can be observed on both, HaCaT and IEC-6 cell cultures upon a 24-h treatment with leaf extracts of Z. spina-christi. IEC-6 cells were less sensitive compared to HaCaT cells, suggesting Z. spina-christi extracts might be potentially suitable for topical applications.

Conclusions
Z. spina-christi is a medicinal and edible plant growing wild in Sudan and used for a long time in traditional medicine for the treatment of different diseases caused by infection. In the present work we report that Z. spina-christi leaves are a rich source of phenolic compounds including chlorogenic acids, proanthocyanidins, simple and highly glycosylated flavonoids. Fifty-seven components were tentatively detected and characterized on the basis of reported analytical data from three to four pieces of independent methods including retention time, UV/Vis spectroscopy, high resolution mass spectrometry (HR-MS n ) and tandem mass spectrometry (MS n ). To our knowledge 45 of them were not reported previously in Z. spina-christi and five for the first time in nature. Besides, four bioactive known compounds were isolated. Next, four known phenolic compounds were isolated and characterized by means of preparative-HPLC and other spectroscopic techniques. We investigated the antimicrobial activities of aqueous and methanol extracts of Z. spina-christi leaves. Methanol extracts were more effective in inhibiting bacterial growth than the aqueous extracts, which is most likely attributed to the solubility and hydrophobicity of the antibacterial plant compounds. Remarkably, the methanol extract of Z. spina-christi leaves revealed an antibacterial activity against both, a Gram-positive and a Gram-negative tester organism while the aqueous extracts also showed milder activity against the Gram-positive representative, B. aquimaris. We conclude that further studies are needed with the goal to purify the methanol extracts in order to isolate and identify the specific compound/s that exert/s antimicrobial effects, in particular, against human pathogens. Furthermore, the cytotoxicity of Z. spina-christi leaf extracts was tested with a range of different cell biological and biochemical means. The herein established cytotoxicity assessment routine revealed that short-term application of Z. spina-  christi leaf extracts barely exhibited any cytotoxic effects on human skin keratinocytes (HaCaT) or rat small intestine epithelial cells (IEC-6). This was deduced from analyses of plasma membrane integrity as well as the intracellular cytoskeletal architecture, which remained largely unaffected upon short-term treatment with the leaf extracts. Moreover, activities of the mitochondrial reductase were not affected by Z. spinachristi extract incubation of IEC6 cells or HaCaT keratinocyte cultures. In conclusion, the above outlined assessment routine has proved suitable to test for possible cytotoxic effects of crude plant extracts in a straight-forward approach, thereby allowing to conclude on the integrity of biological membranes, structural elements such as the actin cytoskeleton, and on the metabolic and proliferative activities of HaCaT keratinocyte and IEC6 intestine epithelial cells.

Additional File
High resolution MS data, full assignment arguments, proposed fragmentation pathways and MS n spectra of some compounds are provided in the Supporting information.