Chemical composition and antibacterial activity against Escherichia coli of extracts of a common household plant

Plants have long been a source for a wide variety of compounds with medicinal properties; of significant interest are plant antimicrobial compounds. Given the continual increase in bacterial resistance, it is imperative that new antimicrobial compounds need to be discovered. The purpose of this study was to screen some common household plants for antibacterial activity, specifically against Escherichia coli. Leaves of 16 different household plants were subjected to methanol-water extraction and the dried extract was analyzed for antibacterial activity using the standard Kirby-Bauer disk diffusion technique. Of all the plants tested, only the Zebra plant (Aphelandra squarrosa) showed a strong antibacterial activity against E. coli. This Zebra plant extract was further analyzed by gas chromatography/mass spectrometry (GC/MS). A variety of known antibacterial compounds including glycerine 46%, phytol 10%, palmitic acid 9%, hydroquinone 6%, linoleic acid 6%, catechol 4% and decanoic acid 3% were shown to be present in the Zebra plant extract along with other compounds whose structure could not be identified at this time. This provides additional avenues for future research especially in the identification of these unknown compounds present in this plant extract which may contribute to its antibacterial activity.


INTRODUCTION
Ongoing scientific research of the importance of plants in human health has led to the identification of more than 9,000 native plants and about 1,500 new species with antimicrobial properties (Swamy et al., 2016). While these numbers appear large, they constitute only a fraction of the estimated 391,000 plant species (Willis, These phytochemicals can disrupt microbial membranes, impairing cellular metabolism, control biofilm formation, inhibit bacterial capsule production, weaken bacterial virulence and even reduce microbial toxin production (Upadhyay et al., 2014;Abreu et al., 2012). The continued study of plant-based derivatives and therapies is particularly important given that a large percentage of the world-wide population depends on them for medicine and healthcare (World Health Organization, 2019). Of the 141 antibacterial drugs to receive approval between 1981 and 2014, over 58% acquired the label of -natural product‖, and in the past 33 years, 47% of approved drugs have been directly or indirectly linked to natural products (Newman and Cragg, 2016), with 11% of these medicines derived from native flowering plant species (Veeresham, 2012). With the increase in antibiotic resistance by bacteria, plants are being seen as a reservoir of potential pharmaceuticals (Gupta and Birdi, 2017). These developments explain the renewed interest in natural product-based drug discovery (Atanasov et al., 2015). Even when a product does not result in a direct medical application, it can aid the process of drug discovery by serving as a chemical template for the synthesis of other novel substances (Newman, 2008;Veersham, 2012).
In search of new plant candidates, it is preferred to find medicinal compounds in the regenerating portion of a plant, such as the leaves, thereby preserving the plants longevity and productivity over a period of time. The aim of the current study was to screen for potential antibacterial activity against E. coli in the leaves of common household plants. E. coli was selected as the bacterial target since it is a prevalent bacterium responsible for common ailments such as urinary tract infection, food-poisoning, gastroenteritis, and even some forms of meningitis and pneumonia (Croxen et al., 2013).

Plant selection
The entire PubMed database was searched using a list of 150 commonly available plants compiled from the local SummerWinds Nursery (Palo Alto, California) to find species that had not been analyzed for antibacterial phytochemicals. From this list, 16 plants (listed in Table 1) were chosen for testing, along with positive plant controls, fennel and poinsettia known to possess antibacterial activity (Cowan, 1999;Sharif et al., 2015). Plants were obtained between the months of January and April of 2018; plant identification was based on the documentation attached to each plant by the nursery along with independent verification in consultations with Prof. Arthur Grossman, Carnegie Institute, Department of Plant Biology, and databases plants.jstor.org, plants.sc.egov.usda.gov and garden.org. A filter disk devoid of plant extract was used as a negative control.

Drying and preparation of plant leaves
Freshly collected leaves were washed with tap water to remove any Kozar et al. 57 dust, and then shade dried at room temperature for 7 days. The shade dried plant materials were ground to a coarse powder and subsequently used for phytochemical extraction and GC/MS quantification.

Preparation of plant extract for antibacterial testing
One gram of ground plant leaves was added to 10 ml of a 4:1 methanol-water mixture and allowed to sit at room temperature for 24 h (Harborne, 1998;Eloff, 1998). The extract was passed through a 0.45 μm syringe filter into 150 mm × 15 mm sterile petri dishes. The filtrate was allowed to evaporate to dryness at room temperature. This dried plant extract was used to test for antibacterial activity.

Bacterial strain
The antibacterial potency of each plant extract was evaluated using the K-12 strain of Escherichia coli (Code 15-5068) obtained from Carolina Biological. Cells were inoculated and incubated at 37°C in Luria-Bertani (LB) (Thermo Fisher Scientific) broth for 12 h prior to plating. This culture was then diluted 1:100 with distilled water prior to plating.

Antibacterial activity test
The Kirby-Bauer disk diffusion method was used to test for antibacterial activity (Bauer et al., 1966;Hudzicki, 2009). Twentyfive ml of LB agar medium was poured into 100 mm × 15 mm sterile petri dishes. After the agar set, plates were marked into quadrants for the testing of four plant samples per plate. Using sterile techniques, 100 μl of the 1:100 diluted bacterial culture was then added to the agar plates and spread evenly across the entire plate surface using a sterile cell spreader. An individual 6 mm filter disk was moistened with sterile water and then swabbed across the dried plant extract. The filter disk containing the plant extract was then placed on the surface of the agar plate, ensuring complete contact with the agar. This procedure was repeated for each plant extract. The plates were subsequently incubated at 35°C for 24 h. The presence of an inhibition zone was measured and recorded (Table 1); the absence of bacterial growth around the filter disk was an indication of antibacterial activity. Each plant was tested at least twice, and each run included both the positive (fennel and poinsettia) and negative (plain filter paper) controls.

Preparation of plant sample for GC/MS
To prepare samples for GC/MS analysis, the same extraction procedure was used as for the antibacterial assay. 1.0 g of dried Aphelandra squarrosa leaves was added to 10 ml of a 4:1 methanol-water mixture and allowed to sit for 24 h. The extract was then passed through a 0.45 μm filter, and evaporated to dryness to remove any water that could damage the chromatography column. This evaporate was reconstituted in 0.1 ml of methanol, and 1.0 μl of this solution was subsequently injected into an Agilent 7890A GC/MS. A different extraction technique was used to extract more hydrophobic molecules than could be obtained with methanolwater. One gram of dried A. squarrosa leaves was added to 10 ml of methanol and allowed to sit at room temperature for 24 h. The extract was passed through a 0.45 μm syringe filter, and then evaporated to dryness. This evaporate was then reconstituted in 0.1 ml of methanol, and 1.0 μl was injected into an Agilent 7890A GC/MS. Felicia Armelbides (Agthea San Luis) 6, 6 6 NI 3 Heuchera (Purple Petticoats) 6, 6 6 NI 4 Verbena Tapien (Blue Violet) 6, 7 6.5 NI 5 Osteospermum ecklonis (African Daisy) 6, 6 6 NI 6 Matthiola incana (Harmony Purple) 6, 6 6 NI 7 Leucanthemum (Shasta Daisy ‗Becky') 6, 7 6.5 NI 8 Gaillardia (Arizona Sun) 7, 6 6.5 NI 9 Brachyscome (Swan River Daisy) 6, 7 6.5 NI 10 Lithodora diffusa (Purple gromwell) 6, 6 6 NI 11 Blechnum gibbum (Silver Lady Fern) 6, 6 6 NI 12 Pellaea rotundifolia ( Each plant species was tested in duplicate. The average zone of inhibition (in mm) is shown. (<8 mm is considered to have no antibacterial activity (no inhibition (NI)); 8-12 is weak antibacterial activity; >12 mm is strong antibacterial activity). The zone of inhibition measurement includes the disk diameter of 6 mm.

GC/MS procedure
The GC/MS analysis of the methanol and methanol-water extracts of A. squarrosa leaves were performed using an Agilent 7890A GC/MS instrument employing the following conditions: a standard DB5-MS Capillary non-polar column (30 m × 0.25 mm × 0.25 μm) with a stationary phase of (5%-phenyl)-methylpolysiloxane. Helium was used as a carrier gas at a constant flow rate of 11.3 psi (1 ml/min). The initial oven temperature was 37°C for 4 min, then increasing at a rate of 20°C/min to a final temperature of 320°C for 7 min. Total run time was 25 min. The fractions obtained from gas chromatography were then analyzed by mass spectrometry to identify the fragmentation patterns (Rukshana et al., 2017;Thangavel et al., 2015). Mass range was m/z 50-550. Identification of phytochemical compounds was achieved through retention time and mass spectrometry by comparing the mass spectra of unknown peaks with those stored in the Agilent ChemStation™ GC/MS library (version 2.00.493).

Antibacterial screening of household plants
Sixteen different household plant species were tested for antibacterial activity against the K-12 strain of E. coli by the disk diffusion method. An evaluation of the antibacterial activity of the plant extracts-the mean diameter of the zones of inhibition-is recorded in

GC/MS analysis
The phytochemical components of A. squarrosa were further analyzed by GC/MS. Since water could not be injected into the GC/MS column, two different extract preparations of the plant were analyzed. The first was a 4:1 methanol-water extract that was evaporated to dryness and then reconstituted in pure methanol prior to analysis by GC/MS. The results are shown in Figure 1 and Table 2. This extract showed several different compounds (Table 2) including glycerine (46%), cyclohexanetetrol (12%), hydroquinone (6%), catechol (4%), 3,7-dimethyl-6-octen-1-ol (3%), capric acid (3%), butyl isopropyl ether (2%), 2,3-dihydro-3,5-dihydroxy-6methyl-4H-Pyran-4-one (1%), and methyl vinyl ketone (1%), all of which are water soluble. There were three other significant peaks that could not be identified with high probability by the ChemStation™ software. The second extraction method utilized pure methanol to extract additional hydrophobic compounds. This methanolic extract produced a different set of compounds as shown in Figure 2 and Table 3. 2,3-dihydro-3,5dihydroxy-6-methyl-4H-Pyran-4-one was present in both the methanol-water and the pure methanol extracts. Otherwise, the pure methanol extraction yielded phytol (10%), palmitic acid (9%) and linoleic acid (6%); all long chain hydrocarbon compounds not particularly soluble in water. There were also seven other significant peaks that could not be identified with high probability from their retention time and fragmentation pattern using the ChemStation™ software. The percent abundance of the various peaks was calculated by the ChemStation™ software. The total abundance from Tables 2 and 3 does not add up to 100% because only peaks with more than 1% abundance were listed. There were many peaks identified by the software that appeared to be little more than background noise and were not further analyzed.

DISCUSSION
Dried leaves from a variety of common household plants were processed in a 4:1 methanol-water solution, and these extracts were tested for antibacterial activity using the disk diffusion method. This extraction solution was chosen because it has been shown to extract a wider variety of phytochemicals than other solvents (Harborne, 1998;Eloff, 1998;Banu and Catherine, 2015). Therefore, majority of the plant phytochemicals potentially responsible for any antibacterial activity were extracted using this method.
Of the 16 plants tested, only one, A. squarrosa, exhibited significant antibacterial activity to E. coli. Due to its striped leaves, this plant is commonly known as the Zebra plant. A. squarrosa is a species in the Acanthaceae family, a native to the tropical forests of Brazil. Traditionally, the leaves from this family have been used in the treatment of a variety of conditions ranging from poisonous bites (Awan et al., 2014;Muthu et al., 2006), fever, headache, vertigo, wounds (Mia et al., 2009), ulcers (Ignacimuthu et al., 2008), to bronchial disease (Krishnaraju et al., 2005). However, specifically for A. squarrosa, there has been no medicinal use reported to date.
A literature search beyond PubMed found a study by Roia and Smith (1977) who used methanolic extracts of the whole Zebra plant, finding antibacterial activity against Proteus vulgaris but not E. coli. Pierre et al. (2015), also report no activity against E. coli, although a different extraction solvent (mixture of methanol and dichloromethane) was used. In contrast, a study by Khalili et al. (2012) reports low antibacterial activity of this plant against E. coli using methanol extractions. However, Khalili et al. do not indicate which part of the plant is used; a positive plant control is not incorporated in the study; and in addition, there is no characterization of the compounds in the plant extract. When harvesting a plant for medicinal purposes, it is important to use only the leaves, as they are able to regenerate, thereby preserving the plant for further harvesting. Our study, however, used the leaves of A. squarrosa and has shown significant antibacterial activity against E. coli.
Further analysis of the A. squarrosa methanol-water leaf extracts by GC/MS identified compounds with known antibacterial activity: methyl vinyl ketone (Kai et al., 2012), glycerine (Nalawade et al., 2015), the phenolic compounds catechol and hydroquinone (Kocacaliskan et al., 2006), capric acid (Royce et al., 2013), and citronellol (Guimaraes et al., 2019). The analysis of the straight methanolic leaf extract identified phytol (Lee et al., 2001),    Knothe et al, (2019) and the long chain fatty acids; palmitic and linoleic acid (Huang et al., 2011;Desbois and Smith, 2010;Dilika et al., 2000). It is of significance to note that the more polar methanol-water solvent was able to extract more polar compounds (such as the quinones, catechol and hydroquinone), while the less polar methanol solvent was able to extract a long chain alcohol (phytol), and fatty acids (palmitic acid and linoleic acid). This demonstrates the importance of using both polar and non-polar solvents for the identification of available phytochemicals in a sample. These aforementioned compounds are found in other plants (Knothe et al., 2019, Santos et al., 2019 and are not unique to the A. squarrosa plant. The cyclohexanetetrol identified in the methanol-water preparation potentially functions as an osmotic regulator, as similar chemical structures serve this function in other plants (Garza-Sanchez et al., 2008). Comparing the two solvent leaf extracts, only one identifiable compound was present in both, namely the pyran. This pyran compound, has been shown to be an antioxidant, and most likely formed during the drying process (Yu et al., 2013).There were, however, additional peaks in the gas chromatogram for both the methanol and methanol-water extracts that were not found in the known chemical database. The characterization of these entities requires further analytical studies, something that is beyond the scope of the current study.

Conclusion
The current study screened a cohort of common household plants for antibacterial activity against E. coli, identifying A. squarrosa as having significant antibacterial properties in its leaf material. GC/MS analysis of the A. squarrosa leaf extracts identified several known compounds that could account for the observed antibacterial activity. The chromatography also showed peaks whose identity could not be determined by the current ChemStation™ chemical database. Further studies are needed for the characterization of these unknown phytochemicals.