Fractionation and Antibacterial Evaluation of the Surface Compounds from the Leaves of Combretum zeyheri on Selected Pathogenic Bacteria

Combretum zeyheri is traditionally used for the treatment of many infections, including bacterial infections. The aim of this study was to fractionate and evaluate antibacterial activity of the crude extract of C. zeyheri, as well as the surface compounds from the leaves of C. zeyheri, in two pathogenic bacteria, Staphylococcus aureus and Pseudomonas aeruginosa. The antibacterial activities of fractions obtained from chromatographic separations were determined using broth microdilution assay on the laboratory and clinical strains of S. aureus and P. aeruginosa. The fractionation of the compounds on the leaf surface yielded 262 fractions. The fractionated compounds with similar TLC profiles were pooled together to yield 47 pools. The extract and pooled fractions CZSC151154, CZSC155160, and CZSC209213 showed significant antibacterial activity with MIC values ranging from 12.5 μg/ml to 100 μg/ml. The clinical strain of S. aureus had MIC greater than 100 μg/ml for CZSC151154 and CZSC155160. The minimum bactericidal concentration values for these fractions were also in the range of 12.5 μg/ml to 100 μg/ml. The extract and fractions CZSC151154, CZSC155160, and CZSC209213 showed a concentration-dependent inhibition of growth in S. aureus. Analyses of the CZSC209213 pool by LC-MS showed the presence of nine compounds which are (3R,7R)-1,3,7-octanetriol, (-)-tortuosamine, 11-aminoundecanoic acid, 1-piperidinecarboxaldehyde, 3-hydroxy-4-isopropylbenzyl alcohol 3-glucoside, hydroxy-isocaproic acid, oleamide, palmitic amide, phytospingosine, and sphinganine. In conclusion, C. zeyheri leaf surface compounds exhibited antibacterial activity. The crude extract and the pooled fractions showed concentration-dependent inhibition of growth on S. aureus. Results from this study indicate the potential of C. zeyheri as a source of lead compounds that may be further developed into antibacterial drugs.


Introduction
Zimbabwe is rich in fauna and fora; it has more than 5000 diferent plant species and 10% of these are believed to be of great medical importance [1]. Te use of medicinal plants in Zimbabwe has been used since time immemorial for the treatment and management of various ailments. It is mostly seen in the rural population, with more than 80 percent depending on these medicinal plants. Tis is because medicinal plants are readily available and afordable for poor communities in rural areas [2]. However, this raises concerns about the over harvesting of these plants that will lead to the extinction of some plants [3]. A group of six organisms has recently been labelled as ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacterbaumannii, Pseudomonas aeruginosa, and Enterobacter spp.) [4]. P. aeruginosa which is among the ESKAPE pathogens is resistant to many antibiotics. It is regarded as a multidrug-resistant organism as it exhibits several drug-resistant mechanisms [5]. Resistance is exhibited through the presence of drug efux pumps coupled together with low permeability of the outer membrane to drugs. S. aureus is another ESKAPE pathogen which is ubiquitous and forms part of the human microbiota and resides in the nose, inguinal areas, and also on the skin. Infection occurs when the bacteria manages to evade the protective mechanisms of the host, for example, when the host is injured [6].
Plants belonging to the Combretaceae family are the most widely used medicinal plants for traditional medicinal purposes in Africa [7]. Terminalia and Combretum species are mainly used for the management of bacterial, fungal, and viral infections; and they are also used for the treatment of bilharzia [8]. In addition to their antimicrobial properties, they possess analgesic, anticancer, anti-infammatory, and antioxidant properties, they are also used for the management of heart disease [7]. Te genus Combretum comprises 250 species; of these species, over 24 species are used for medicinal purposes [9].
Combretum zeyheri is found in southern parts of Kenya, DR Congo, and north-eastern parts of South Africa [10]. In Zimbabwe, it is distributed in the tropical grassland (savanna) regions [11]. Te tree grows in forests, rocky hillsides, and sometimes in termite hills [11]. C. zeyheri has been previously documented to possess antimicrobial properties. Te whole leaf extract was shown to have antifungal activity against C. albicans and C. krussei with MIC of 0.08 μg/ml and 0.16 μg/ml, respectively [11]. Tis shows that the C. zeyheri leaf extract contains phytochemicals that are responsible for the antimicrobial activity. Tese phytochemicals of C. zeyheri can be isolated and can serve as new chemical entities for the development of new antimicrobial drugs [12].
Plants secrete a wide range of phytochemicals on their leaf surfaces that act as allelochemicals against insects and microbes [13]. Tese secondary metabolites difuse to the surface of the leaf and accumulate in the glandular trichomes or the waxy cuticle. Tis makes plants a good source of antimicrobial compounds [14,15].

Materials and Methods
An outline summarising the procedures is shown in Figure 1.

Chemicals and Reagents.
All chemicals used in the study were obtained from Sigma-Aldrich Chemical Company (Darmstadt, Germany). All solvents used were of analytical reagent grade, and these were acetone (extractant), methanol, n-hexane, methanol, ethyl acetate, and chloroform used for column chromatography. Dimethylsulfoxide (DMSO) was used for dissolving the crude extract and fractions for assay. Silica gel mesh of size 60-120 was used for the preparation of the extract and stationary phase; fltration was performed on Whatman flter paper number 1. TLC plates were obtained from Sigma-Aldrich Chemical Co. Distilled water was collected from the Pharmacy Department. A rotary evaporator was used for fraction concentration. Te strains used were laboratory strain P. aeruginosa (ATCC 27853), P. aeruginosa clinical strain, laboratory strain S. aureus (ATCC 9144), and S. aureus clinical strain. Te clinical strains were procured from the Parirenyatwa Group of Hospitals from the College of Health Sciences and Department of Microbiological Sciences of the University of Zimbabwe, while the laboratory strains were from the Department of Biological Science at the University of Botswana. Determination of all antibacterial activities was carried out in a Class 2 biosafety cabinet.

Plant Collection and Extraction of Phytochemicals.
Te leaves were collected from Norton (Geographic Coordinates of Norton, Zimbabwe; latitude: 17°52′59′ S; longitude: 30°42′00′E; elevation above sea level: 1360 m; Mashonaland West Province of Zimbabwe). Te leaves were verifed by a taxonomist at the National Herbarium and Botanical Gardens. Te surface compounds were extracted according to Omosa et al. [16], with modifcations. Te leaves were frst washed under running water to remove any dust particles and then dipped in acetone for 15-30 seconds. Filtration was performed to remove any residues, and the acetone evaporated overnight under a cold stream of cold air to concentrate the extract until it was dry.

Tin Layer
Chromatography. TLC was used to determine the mobile phase for column chromatography. TLC plates (aluminium coated with Merck, silica gel 60 F 254 ) were developed under saturated conditions with solvents of diferent polarities, hexane:ethyl acetate 90 :10 up to 85 :15 and chloroform:methanol 90 :10 up to 80 : 20. Te surface compounds extracted at 100 μg/ml were applied 1 cm from the bottom of the plates using a micropipette and allowed to develop in the tank saturated with an eluent. Once developed, the separated compounds were observed under UV light at 254 nm and 360 nm. Sulphuric acid was also used for visualisation, and it was sprayed and heated to 110°C. Te best mobile phase was the one that showed distinct separation of compounds.

Isolation of Phytochemicals Using Column
Chromatography. Te extract was dissolved in minute amounts of methanol. It was adsorbed in an equal amount of 60-120 silica mesh and added in very small amounts until a crunchy mixture formed. Tis was left to dry overnight and ground to a fne powder using a mortar and pestle. Silica 60-120 which was four times the amount of the extract was mixed with 0.532 L of ethyl acetate to form slurry. Te slurry was carefully poured into the length of the column (4 cm × 120 cm). Te dead volume was determined and the column was left to pack overnight. Te prepared extract was poured into the column and a plug of cotton wool was placed on top to prevent disturbances as the mobile phase was being poured into the column. Solvents of diferent polarities were slowly passed through the column in order of increasing polarity to facilitate separation. Te fractions collected from each solvent polarity were concentrated using a rotary evaporator (BUCHI ™ R2, Switzerland) and poured into vials that were left to dry on the bench. Te purity of the fractions was determined using TLC.

Te Efects of Fractionated Extracts on Bacteria.
Antibacterial susceptibility tests were carried out according to the broth microdilution method according to Vangeas et al. [17]. Te bioactivity of the pooled fractions was screened using sterile polystyrene 96-well plates at the highest concentration of 100 μg/ml ( Figure 2).
Te antibiotic ciprofoxacin was used as a standard drug. Te number of viable bacteria in the sample used for the study was 1 × 10 6 cfu/ml according to the CLSI standard guidelines. Each plate accommodated 10 diferent fractions for screening. Te plate was covered with a cling wrap paper and incubated at 37°C in humid conditions for 20 hours. After 20 hours of incubation, readings were taken, and the results were analysed to give the MIC. MTT assays were carried out to serve as an indicator of growth. Yellow MTT forms insoluble purple formazan crystals in the presence of a viable cell [18]. MTT of volume 20 μl of concentration 0.2 mg/ml was added to the wells and incubated for 2 hours. Tis assay confrmed the MIC value since MTT is an indicator of growth. Te lowest concentration that showed no colour change was the MIC.  Antibiotic + cells P 1 P 2 P 3 P 4 P 5 P 6 P 7 P 8 P 9 P 10 CZSC155160 (25), and CZSC209213 (39)) that had the greatest inhibitory activities were chosen for the determination of MBC. An inoculum of bacteria from the wells with the lowest concentration to the highest concentration of the fractions showing no visible growth was streaked on the agar. Positive and negative controls were incorporated into the assay. Te MBC value was the lowest concentration that showed no visible growth in the Petri dishes.

2.7.
Time-Kill Assays. Te time-dependent and concentration-dependent growth inhibition studies were carried out on S. aureus (ATCC) which was more susceptible to all the fractions. Te efects of the diferent concentrations of the extract and the fractions on the growth rate of S. aureus were determined by using the 96-well microplate using the turbidometry method according to Banf et al. [19], with some modifcations. Te bacterial inoculum (1 × 10 6 CFU/ml) was exposed to varying concentrations of the extract and fractions which were ½ × MIC, 1 × MIC, and 2 × MIC, and MBC and OD readings were recorded using the Tecan microplate reader (Tecan, Genios Pro Microplate Reader, Grödig, Austria) after every 2 hours for the frst 10 hours and then at 20 and 24 hours.

High-Performance Liquid Chromatography/Mass Spectrometry Analyses.
Te constituents of the most potent extract that were in the CZSC209213 pool were analysed using liquid chromatography-mass spectroscopy in the Department of Pharmacy of the University of Zimbabwe. An Agilent HPLC 1260 system (Agilent, USA) was used equipped with a binary pump, autosampler, and thermostated column compartment. Chromatographic separation was performed using Agilent Poroshell 120, C18 Column, with dimensions 50 × 2.7 mm, 3 microns. Te mobile phase for the positive mode consisted of 0.1% formic acid in deionised water (A) and 0.1% formic acid in acetonitrile (B). Te total run time for each sample was 25 min. Te fow rate was 0.25 mL min −1 , and the injection volume was 5 μL. Te column temperature was maintained at 40°C. An Agilent Technologies Q-TOF 6530 Mass Spectrometer was used as the detector. Agilent Technologies Mass Hunter Software version B.07.03 (509) was used for data acquisition, instrument control, and data analysis.

Statistical Analyses.
Graphical and statistical analyses were performed using GraphPad Prism version 5. Data were expressed in the form of mean ± standard deviation of the mean. Statistically signifcant diferences between various means of controls and tests were analysed using one-way ANOVA using Dunnett's multiple comparison posttest with a p value of 0.05.

Results and Discussion
Antibacterial resistance results from misuse, irrational prescribing, and poor patient compliance. Bacterial species multiply at a faster rate and have a high ability to change genetic material, making the available antibiotics inefective [20]. Te resistance forms of the pathogenic bacteria are associated with poor clinical outcomes compared with their susceptible counterparts. Poor prognosis means longer hospital stays, increased mortality, and higher hospitalisation costs [16]. Te Infectious Diseases Society of America (ISDA) developed the acronym ESKAPE to emphasise the group of pathogens that cause hospital infections and effectively "escape" the efects of antibacterial drugs [17]. New antibacterial agents and plants that can be a useful source of lead compounds for drug development are needed. Tis study highlights the efects of surface compounds isolated from Combretum zeyheri leaves as sources of antibacterial agents. We previously showed the inhibitory efects of Combretum zeyheri leaf extracts and their S9 metabolites against Escherichia coli, Bacillussubtilis, and Candida albicans [24]. Te plant extract also showed inhibitory efects on drug efux in all three organisms, suggesting that the antimicrobial actions could be partially due to the inhibition of drug efux pumps. Te S9 metabolites of C. zeyheri also had an inhibitory efect on the growth of C. albicans, suggesting that they are inhibitors of the growth of Candida albicansin vivo.

Separation and Identifcation of Surface Compounds.
A total of 262 fractions were collected from the column chromatographic separation. Fractions showing similar TLC profles were pooled together to obtain 47 pools. Profles for the identifcation of separated fractions as well as their antibacterial efect on P. aeruginosa and S. aureus are shown are Tables 1-3.

Comparison of the Efects of Fractionated Pools and Extracts on Bacteria.
Bioactivity screening was carried out to determine the most potent pooled fractions. Te extracts were screened at the highest concentration of 100 μg/ml. Laboratory strains of P. aeruginosa and S. aureus were used in the study. Te bacterial species were incubated overnight at a concentration of 100 μg/ml of the pooled fractions. Te results were plotted as a graph of cell density versus the concentration (100 μg/ml) of the pooled fractions; the % efect of these pooled fractions on the inhibition of growth of the test pathogens was calculated from these graphs and is shown in Tables 1-3. All the pooled fractions signifcantly inhibited the growth of P. aeruginosa. Tere was a signifcant inhibition of growth of the bacteria for fraction CZSC2728 to pooled fraction CZSC133139 when compared with the control. Tere was complete inhibition of growth of the bacteria for the pooled fraction CZSC141146 to pooled fraction CZSC260261. Ciprofoxacin was used as a positive control and the MIC was 0.25 μg/ml. Te efects of the pooled fractions on S. aureus (ATCC) are also shown in Tables 1-3. CZSC2932, CZSC5051, and CZSC9596 did not signifcantly inhibit the growth of S. aureus. Fraction CZSC2728 to fraction CZSC121131 signifcantly inhibited the growth of S. aureus. Ciprofoxacin was used as the positive control drug and the MIC was 0.25 μg/ml.
Since plant extracts contain a variety of phytochemicals which include alkaloids, tannins, terpenes, and saponins, they can be separated according to their varying polarities. In the study, column chromatography was utilized for separation together with the most used method of identifcation TLC [19]. Te study showed that there was a decrease in activity after fractionation which occurred from pooled fractions CZSC2728 to CZSC133139 which had no MIC at the highest concentration of 100 μg/ml used in the study. Te antibacterial activity of the fraction CZSC209213 was increased and that of CZSC151154 and CZSC155160 was reduced compared to the original unfractionated extract. Tese results are in agreement with the trend observed in a study by Njateng et al. [21], who found that the fractionation crude extract of Polyscias fulva Hiern (Araliaceae) led to an increase in antibacterial activity of the fractions on the bacterial species S. typhi, E. aerogenes, P. aeruginosa, and E. coli. In a study by Nwodo et al. [22], the fractionation of the crude methanolic bark extract of T. indica led to an increase or reduced antibacterial activity of the fractions.  After the fractionation on the column, pooled fraction CZSC209213 was found to be more active than the extract. In a study, Milugo et al. [23] reported on the antagonistic efects of phytochemicals afecting the antioxidant activity of the crude extract of Rauvolfa cafra. In that study, the presence of saponins and alkaloids was found to reduce the antioxidant properties of the crude extract [23]. Combretum species are known to have activity against both Gram-positive and Gram-negative bacteria [7,24]. In a study by Elof et al. [25], the antibacterial activity of the whole leaf extract of C. zeyheri in both S. aureus and P. aeruginosa was reported to be 0.8 mg/ml, showing that C. zeyheri has activity against both Gram-positive and Gram-negative bacteria. In the study, both clinical and laboratory strains were used. Clinical strains were used to represent the pathogenesis of "real life" because laboratory strains may lack some pathophysiological characteristics [26]. Laboratory strains are reference strains as required by NCCLS, and clinical strains have been exposed to selective pressures in clinical settings when patients are being treated [25]. Laboratory strains were more sensitive to surface compounds compared to clinical strains, due to the selective drug pressure to which clinical strains are exposed. In a study by Mukanganyama [27], it was reported that laboratory strains were more susceptible.

MIC and MBC Determination for Fractions CZSC151154, CZSC155160
, and CZSC209213. Tree pooled fractions CZSC151154, CZSC155160, and CZSC209213 that showed the highest inhibitory activities for both bacteria in initial antibacterial screening assays were chosen for the determination of MBC. Te efects of the pooled fractions CZSC209213 on the growth of P. aeruginosa and S. aureus are shown in Figure 3. Te data for all three fractions are summarised in Table 4. Te most active pooled fraction was CZSC209213 shown in Table 4; the order of potency of the pooled fractions and the crude extract was CZSC209213 > extract > CZSC155160 > CZSC151154. Te extract had MIC of 100 μg/ml for S. aureus (clinical strain) and 25 μg/ml for S. aureus (ATCC) and both the laboratory and clinical strains of P. aeruginosa. Te second most potent pooled fraction of the three pooled fractions tested was CZSC155160 which had the same MIC value as the crude extract of S. aureus (ATCC) and both the laboratory and clinical strains of P. aeruginosa, but the MIC of S. aureus (clinical strain) was greater than 100 μg/ml used. Te least active pooled fraction of the three pooled fractions was CZSC151154 which had MIC of 50 μg/ml for both the laboratory and clinical strains of P. aeruginosa, MIC of 25 μg/ml for S. aureus (ATCC), and MIC (clinical strain) greater than 100 μg/ml for S. aureus.
Te order of potency was as follows: For MBCs, there was a two-fold decrease in the MIC for both species of bacteria. Since MICs and MBCs were not that diferent between extracts and fractions, this may signify that the phytochemicals in these pools were the most abundant and were responsible for the bioactivity or the enhanced bacteriostatic and bactericidal efects observed.

Time-Kill Assays.
Te efects of the extract and the pooled fractions on the growth curve ofS. aureus were determined using the turbidimetric method at varying concentrations. In all cases, the MIC was equivalent to the MBC. Te three pooled fractions CZSC151154, CZSC155160, and CZSC209213 and the extract showed a concentrationdependent inhibition efect on S. aureus growth (Figure 4). In all cases, the concentrations at 1 × MIC and 2 × MIC values completely inhibited growth from time t = 0 to time t = 24. Te ½ × MIC value showed that the lag phase  Te Scientifc World Journal of the bacterial growth curve was prolonged compared with the positive control. Of note was that all samples, except the control, showed a negative OD value at time zero. Tis essentially shows that the MIC and the 2 × MIC concentration had an immediate reduction efect on cell density compared to the control for the 24 hours of investigation. Te duration of the lag phase for the positive control was 2 hours. Te duration of the lag phase for ½ MIC for the extract was 11.1 hours (Figure 4(a)); for CZSC151154, it was 10 hours (Figure 4(b)), for CZSC155160, it was 11.5 hours (Figure 4(c)), and for CZSC209213, it was 12.9 hours (Figure 4(d)). Te % inhibition of the ½ MIC values are shown on the graphs with CZSC209213 having the highest % inhibition of 82% (Figure 4(d)), the extract having a % inhibition of 46% (Figure 4(a)), and a % inhibition of 44% for both CZSC151154 and CZSC155160 (Figures 4(b) and 4(c), respectively). It is expected that the Gram-negative bacteria are generally less sensitive to plant extracts and drugs due to the presence of an outer membrane which acts as a permeability barrier [7]. However, this was not observed with the clinical strains used in the study. Time-kill kinetics also provides information on the speed of bactericidal activity of a plant extract. An ideal antibacterial agent would be one that kills bacteria in a short time to prevent the bacteria from becoming resistant [18]. In the study, the high concentrations of MIC and 2 × MIC completely inhibited the growth of the bacteria and did not allow the exponential phase to be reached, and ½ × MIC value of 1 2 MIC did not completely inhibit the growth even after exposure of the bacteria for 24 hours. Te bacterial growth curve consists of three phases: the lag phase during which the bacteria initiate an infection and the exponential phase. In the exponential phase, virulence factors are synthesized and there is also active growth of the bacteria and lastly the stationary phase where the bacteria produce toxins and exotoxins and develop sensing mechanisms for spreading to other tissues [28]. Te ability of bacteria to initiate an infection lies in the presence of virulence factors, which are synthesized in the exponential phase. Terefore, it is important to note that the surface compounds of C. zeyheri prevented S. aureus from entering the exponential phase and therefore exhibited high antibacterial activity.  Te results show that CZSC209213 contained a mixture of phytoconstituents of diferent classes. It had a phospholipid (phytosphingosine), fatty acid amides (palmitic amide and oleamide), cholesterol (sphinganine), amine (11-aminoundecanoic acid), amino acid (hydroxy-isocaproic acid), alcohol ((3R,7R)-1,3,7-octanetriol), and alkaloid ((-)-tortuosamine). Te chemical formula of these compounds are provided as supplementary data at the end of the manuscript as Supplementary Table 4 and their structures are also shown in the supplementary material as Supplementary Table 5.
Medicinal plant products are likely to have many variations in phytochemical content due to several contributing factors which include the plant's identity, seasonal variations, and geographical location [29]. To ensure that health as a basic need for humans is afordable and safe, the WHO stresses the importance of qualitatively and quantitatively characterizing of medicinal plant products in traditional medicinal practise [30]. LC-MS analysis of CZSC209213 led to the identifcation of 9 compounds belonging to diferent classes of phytoconstituents. Te antibacterial activity of CZSC209213 can be attributed to the presence of phospholipids (phytosphingosine and sphinganine) in the pooled fraction. Tese phospholipids can also be found on the epidermis of the human skin and ofer protective efects against pathogens [31]. Phytosphingosine was found to be antibacterial with an MIC of 200 μg/ml on S. aureus and sphinganine with an MIC of 100 μg/ml on P. aeruginosa. In a study by Drake et al. [32], sphinganine acted by disrupting the bacterial cell wall, resulting in invaginations of the plasma membrane and loss of ribosomes in the cytoplasm [33].
Most of the compounds identifed by LC-MS from the leaves of Combretum zeyheri have been reported to have antibacterial efects. Phytosphingosine (PS) has been shown to be active against Staphylococcus aureus, the etiological cause of atopic dermatitis [34]. Phytosphingosine (PS) is a natural antimicrobial ingredient present in the stratum corneum of mammalian skin [35]. Te presence of this compound in the leaves makes it an important component of the plant, thus, justifying the use of the plant in Zimbabwean traditional medical practices. In the study by Başpınar et al. [35], it was shown to have antimicrobial activity against  Concentration-dependent inhibition of growth of bacteria is shown by all the 3 fractions and the extract. Prolongation of the lag phase occurred compared with the control (S. aureus). Te duration of the lag phase for ½ MIC for (a) was 11.1 hours, (b) was 10 hours, (c) was 11.5 hours, and (d) was 12.9 hours. Te % inhibition of the ½ MIC values is shown in the graphs. 8 Te Scientifc World Journal Gram-positive and Gram-negative bacteria and Candida strains. It has been reported that fatty acids and their amide derivatives are natural-defence agents in plants [36]. In terms of mechanism of action, the amides target protein synthesis and cause leakage of intracellular components in the organisms. It is postulated that the presence of palmitic amide and oleamide in the surface extracts of Combretum zeyheri would exert antimicrobial activities. Terefore, the fatty acids and their amides can be exploited as therapeutics that can cover a wide range of indications of bacterial infections, as shown in this study. Te derivatives of phospholipids (C16 sphinganine) and fatty acids (11-aminoundecanoic acid) and palmitic amide have also been isolated by LC-MS from mushrooms that have been reported to be responsible for the antibacterial activity against Staphylococcus aureus [37]. Te compound 2-hydroxy-isocaproic acid (HICA) was shown to have antibacterial activity against obligate anaerobic bacterial species associated with periodontal disease [38]. Te compound 1-piperidinecarboxaldehyde has been isolated from an herbal medicine that contains extracts of Zingiber ofcinale, Piper longum, and Piper nigrum [39]. Tortuosamine is an alkaloid found in Sceletium tortuosum. Traditionally, this medicinal plant is mainly masticated or smoked and is used to relieve toothache and abdominal pain, as a mood enhancer, analgesic, hypnotic, anxiolytic, thirst and hunger suppressant, and for its intoxicating/euphoric efects [40]. Te compound 3-hydroxy-4-isopropylbenzyl alcohol 3-glucoside belongs to the class of organic compounds known as terpene glycosides. Terpene derivatives have been shown to be a potential agent against antimicrobial resistance (AMR) pathogens such as A. baumannii and S. aureus [41].

. Conclusions
Te surface compounds from the leaves of C. zeyheri were found to have signifcant antibacterial activity. After fractionation, CZSC209213 had greater antibacterial activity compared to the crude extract, and CZSC151154 and CZSC155160 had lower activity than the extract. Te antibacterial activity of CZSC2728 to CZSC133139 was signifcantly reduced after fractionation. Te extracts CZSC209213, CZSC151154, and CZSC155160 showed a concentrationdependent inhibition efect on the growth of S. aureus. LC-MS analyses showed the presence of phytospingosine, palmitic amide, oleamide, spinganine, 11-amino-undecanoic acid, hydroxy-isocaproic acid, 1-piperidinecarboxaldehyde, 3hydroxy-4-isopropylbenzyl alcohol 3-glucoside, (3R,7R)-1,3,7-octanetriol, and (-)-tortuosamine. Tese compounds contribute to the antibacterial activity of the leaf surface compounds found in C. zeyheri. Further isolation of single chemical species is required to develop lead compounds with antibacterial activity.

Data Availability
Te data sets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest
Te authors have no conficts of interest to declare.