Infections are the world’s leading cause of premature deaths, killing almost 50 000 people every day . The pharmacological industry produces a large amount of antibiotics but the extensive and inappropriate use of antibacterial and antifungal agents led to a significant upsurge in resistance to these drugs [2, 3]. The treatment of bacterial and fungal pathogens that are drug-resistant is even more complicated in acquired immune deficiency syndrome (AIDS) patients . These developments have increased the need to search for new antibacterial products with improved activity .
Plants produce a diverse range of bioactive molecules with a wide spectrum of activities, making them a rich source of different types of compounds that could be used as medicine [6, 7]. Throughout the world, plants are used to treat many illnesses, particularly infectious diseases, and were once used as the primary medicines all over the world [8, 9]. There has been growing world-wide interest in natural and traditional medicines as an alternative form to treat infectious diseases . This is partially based on the widely held, but not necessarily correct assumption, that natural medicine is safer. The World Health Organisation (WHO) estimated that about 80% of the rural populations of the developing countries rely exclusively on plants to meet their health care needs . Nevertheless, of all the c. 250 000 species of higher plants on earth, only a fraction has been examined for all aspects of their potential therapeutic medicinal value . Furthermore, natural products and their derivatives (including those from microorganisms) represent more than 50% of all drugs in clinical use in the world . The importance of natural products in drug discovery has been discussed in many scientific papers .
Southern Africa has a rich floral diversity comprising in the order of 10% of the world’s plant diversity on less than 2.5% of Earth’s land surface . According to Goldblatt  about 80% of the plants in the Western Cape are endemic to this region. This diversity represents a very valuable resource for commercial development as well as basic scientific studies . In South Africa in particular, many rural ethnic groups rely on traditional indigenous plant knowledge to treat various diseases in both humans and livestock [17, 18]. In the order of 15% of the 24 000 taxa recorded in southern Africa are used in traditional medicines [19, 20] and an estimated 500 plant species are traded in informal medicinal plant markets . Traditional medicine remains more affordable than Western medicine and is also easily accessible by the poorer communities.
In recent years, the rate of information documented on the biological activity and chemistry of plants used in traditional medicine in southern Africa have increased  and several studies provided scientific support for the use of various African plants for treating infections and diseases . Several southern African medicinal plants have been identified in the first African Herbal Pharmacopoeia .
Contrary to expectations, thousands of publications on antimicrobial activity of plant extracts have not led to the development of any new commercial antimicrobial compounds of significant importance world-wide . Many large-scale screening programmes of the past failed to produce worthwhile plant-derived antimicrobial pharmaceutical products [5, 26]. Consequently, the major pharmaceutical companies have lost interest in screening higher plants for their biological potential .
Some reasons for the failure of these screening programmes could be:
methods such as agar diffusion that are inaccurate and misleading for plant extracts have been used by many researchers ,
many scientists following traditional healers, used water while it has been shown that the antimicrobial compounds are usually intermediate polarity compounds that are not extracted by water [28, 29],
several publications considered excessively high MIC’s (>5 mg/ml) as active [25, 30–32]
many scientists selected plant species for research studies based on traditional knowledge .
Traditional healers typically use aqueous extracts of plants which, as stated above, generally have very low activities [28, 33]. Water does not extract the antimicrobial compounds that usually have an intermediate or non-polar character . The activity of effective aqueous extracts used by traditional healers may be based on an indirect effect by stimulating the immune system of the host rather than killing the pathogens. Aqueous extracts of plants species used in ethnomedicine may therefore not have high direct antimicrobial activity and scientists may consequently have focussed on the wrong species.
Moreover, many scientists focussed on the isolation of compounds not recognising that phytomedicines contain a mixture of compounds that often acts synergistically . In our experience antimicrobial compounds isolated from extracts never had the expected activity based on the activity of crude extracts and fractions . This is probably because different plant metabolites may work in combination with other compounds to regulate microbial infections and may therefore not be effective alone .
For these reasons coupled with the large number of plant species that have not yet been examined for their antimicrobial activities and the urgent need to discover new antimicrobial agents, we decided to screen southern African plant species to identify promising antimicrobial plant extracts against six important pathogens as leads for further in-depth research. Such an approach has not been followed in southern Africa before, except partially in the work of Noristan Pharmaceuticals . The methods they used to determine antimicrobial activity are now outdated, thereby rendering comparisons impossible. Our focus was to sample leaves of representatives of southern African tree genera and families for ease of collection and sustainable use in case exciting activities are found. We used a standardised method for analysing the antimicrobial activities. A wide-screening approach offered the potential to discover plants with antimicrobial activities that are available in rural areas but which are not used traditionally. It also reduced the administrative complications of investigating plants based on its use in traditional medicine due to the new legislation to prevent biopiracy and promote benefit sharing in South Africa .
This study was part of a preliminary extensive screening programme of the Phytomedicine Programme, Department Paraclinical Sciences, University of Pretoria. The main aim of the screening was to facilitate the discovery of tree leaf extracts with high activities that may yield useful products for the herbal medicine that can be used to combat microbial infections in animals and humans. Preliminary data obtained in this study provided material for several masters and doctorate projects for students. Patents have been filed on some of the results.
The secondary aim of the screening study was to compare the antimicrobial activities of several southern African tree species at two taxonomic levels: suprafamilial (order) and suprageneric (family) to investigate if some orders and families generally contained tree species with higher antimicrobial activities compared to other taxa. Further studies could then focus on related taxa in promising orders and families to facilitate and improve the selection process. This approach may facilitate and optimise the selection of tree species for the discovery of new antimicrobial plant extracts by saving time and cost.
Given that related taxa may contain similar or related pharmacologically active compounds and therefore similar bioactivity [36, 37], the correlation between taxonomy and antimicrobial activity was investigated. The rationale was that taxa with general high activity could offer more promising leads. If good correlations were found, it could lead to a better guided approach in selecting tree species from promising families for continuing studies rather than based on ethnopharmacology. This information could also be useful for related taxa of plants growing outside South Africa. A number of closely related taxa, assumed to contain related active compounds could then be screened.
In this paper we compare the antimicrobial activities of tree extracts at order level based on the average minimum inhibitory concentration (MIC) of all the species that we have analysed within each order. Since orders are more inclusive compared to families, the mean MIC’s calculated for each of the orders should therefore be less affected by outliers. In addition, since the boundaries of orders are more accepted, a comparison at this level may reduce variation caused by changes in, or differences between classification systems. We also compared the antimicrobial activities of the species present in the four orders with the highest activities with the species present in the four orders with the lowest activities to determine the practical value of the differences between orders with high and low activity.