Background: The development of antibiotic resistant bacteria stems from a number of factors, including inappropriate
use of antibiotics in human and animal health and their prolonged use as growth promoters at sub-clinical doses in
poultry and livestock production. We were interested in investigating plants that could be useful in protecting humans or
animals against diarrhoea. We decided to work on extracts of nine plant species with good activity against Escherichia coli
based on earlier work in the Phytomedicine Programme. Leaves of nine medicinal plant species with high antibacterial
activity against Escherichia coli were extracted with acetone and their minimal inhibitory concentration (MIC) values
determined using a microplate serial dilution technique against Gram-positive (Staphylococcus aureus, Enterococcus faecalis
and Bacillus cereus) and Gram-negative (Escherichia coli, Salmonella Typhimurium and Pseudomonas aeruginosa) bacteria.
Bioautography was used to determine the number of bioactive compounds in each extract. In vitro safety of the extracts
was determined using the 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide reduction assay on Vero cells.
Results: The extracts were active against all the pathogens with average MICs ranging from 0.02 to 0.52 mg/ml.
As expected E. coli was relatively sensitive, while E. faecalis and S. Typhimurium were more resistant to the
extracts (average MICs of 0.28 mg/ml and 0.22 mg/ml respectively). Cremaspora triflora and Maesa lanceolata leaf
extracts had higher activity than the other extracts against Gram-positive and Gram-negative pathogens with
mean MICs of 0.07 mg/ml and 0.09 mg/ml respectively. Extracts of Maesa lanceolata and Hypericum roeperianum
had the highest total antibacterial activity (TAA) at 1417 and 963 ml/g respectively. All extracts with the exception of that
of Maesa lanceolata, Elaeodendron croceum and Calpurnia aurea had relatively low cytotoxicity with LC50 >20 μg/ml.
Cremaspora triflora had the best selectivity index (SI) against S. aureus and E. coli of 2.87 and 1.15 respectively. Hypericum
roeperianum had a SI of 1.10 against B. cereus. Bioautography revealed 1–6 visible antimicrobial compounds that were
Conclusions: There was a weak positive, but statistically non-significant correlation between the potency of the extracts
and their cytotoxicity (R= 0.45, ρ > 0.05). The activity of the extracts on the test bacteria was in some cases not correlated
with cytotoxicity, as shown by selectivity indices >1. This means that cellular toxicity was probably not due to compounds
with antibacterial activity. Some of the extracts had a good potential for therapeutic use against the bacterial pathogens
or for application in treating diarhoea. It does not appear that activity against E. coli is a good predictor of activity against
Gram-negative rather than Gram-positive bacteria. Further investigation is in progress on C. triflora and H. roeperianum,
both of which had promising activities and potential safety based on cytotoxicity.