Methods 42 (2007) 325–329 www.elsevier.com/locate/ymeth Colorimetric broth microdilution method for the antifungal screening of plant extracts against yeasts Manjuan Liu a a,* , Veronique Seidel a, David R. Katerere b, Alexander I. Gray a Natural Products Research Group, Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, 27 Taylor Street, Glasgow, G4 0NR, UK b PROMEC Unit, MRC, P.O. Box 19070, Tygerberg 7500, Cape Town, South Africa Accepted 18 February 2007 Abstract Screening plant extracts for antifungal activity is increasing due to demand for new antifungal agents, but the testing methods present many challenges. Standard broth microdilution methods for antifungal susceptibility testing of available antifungal agents are available now, but these methods are optimised for single compounds instead of crude plant extracts. In this study we evaluated the standard NCCLS method as well as a modification which uses spectrophotometric determination of the end-points with a plate reader. We also evaluated another standard method, the EUCAST method, which is a similar microdilution assay to the NCCLS method, but uses a larger inoculum size and a higher glucose concentration in the medium as well as spectrophotometric end-point determination. The results showed that all three methods had some drawbacks for testing plant extracts and thus we modified the NCCLS broth microdilution method by including a colorimetric indicator—resazurin for end-point determination. This modified method showed good reproducibility and clear-cut end-point, plus the end-point determination needed no instruments. It enabled us to evaluate the activity of a selection of extracts from six Combretaceous plants against three Candida spp. and thus provided pharmacological evidence for some traditional uses of these plants while assisting the identification of the active ingredients.  2007 Elsevier Inc. All rights reserved. Keywords: Combretaceae; Candida; Broth microdilution; NCCLS; EUCAST; Resazurin; Natural products; Antifungal tests 1. Introduction Opportunistic fungal infections, including life-threatening invasive mycoses, have increased in incidence over the last two decades due to the prevalence of immune-suppressing disease conditions e.g., HIV-AIDS, organ transplantation and cancer. The search for new antifungal agents from natural sources has intensified in response to the limitations of currently available therapy and the emergence of drug-resistant strains [1]. The family Combretaceae consists of 20 genera of which Combretum and Terminalia (about 250 species each) are the largest and * Corresponding author. Tel.: +44 141 548 2204. E-mail address: manjuan.liu@strath.ac.uk (M. Liu). 1046-2023/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ymeth.2007.02.013 most widely used in traditional herbal medicine in Africa and Asia [2]. Both Combretum and Terminalia species are used to treat abdominal disorders, diarrhoea, venereal disease, wounds and burns [3]. The species of the Combretaceae family chosen for the present study have been shown to possess anti-Candida activity [4–8] and the main aim of this study was to evaluate the suitability of a standard antifungal susceptibility test for plant extracts. Testing for antifungal activity of natural products, especially plant extracts, presents many challenges. The diversity of testing methods and lack of clearly defined testing conditions such as inoculum size and medium type can lead to low reproducibility and difficulties in comparison with the anti-Candida activities of the same plant species. Additionally, the trailing effect (residual fungal growth) of Candida species makes the end-points less well defined in broth 326 M. Liu et al. / Methods 42 (2007) 325–329 tests than that of tests for bacteria and can affect the reliability of such tests. At the same time, the standardisation of the in vitro antifungal susceptibility testing has advanced greatly in recent years. The National Committee for Clinical Laboratory Standards (NCCLS) have set the benchmark methodology by providing laboratory tested, reproducible, consensus peer-reviewed standards. NCCLS M27A2 standard for yeasts [9] provides a broth microdilution test which could be a good screening method for plant extracts with its high through-put potential, considerable savings in media usage, and requirement of a small quantity of sample. Some recent studies have substituted the traditional agar diffusion method with the NCCLS method for screening plant extracts [10–12]. However the optimisation of test conditions is needed in order to adapt this method to test plant extracts because it is optimised for accepted antifungal agents, i. e. pure single compounds. In this study, we evaluated three end-point determination methods for the standard NCCLS method, which are the visual determination using turbidity levels as stated in the NCCLS M27-A2 protocol, the spectrophotometric determination using a plate reader [13,14], and the colorimetric determination using an oxidation-reduction indicator resazurin. We also evaluated another standard method, the EUCAST method, mainly used in European countries [15], which is a similar microdilution assay to the NCCLS method, but uses a larger inoculum size and higher glucose concentration in the medium as well as a spectrophotometric end-point determination. 2. Materials and methods 2.1. Plant collection All plants were collected in Guruve, Zambezi Valley in Northern Zimbabwe in May 1999. Voucher specimens (Table 1) have been deposited at the herbarium of the Harare Botanical Garden. 2.2. Preparation of plant extracts Dried and powdered plant materials (ca.20 g) were extracted in a Soxhlet apparatus successively with n-hexane, ethyl acetate (EtOAc) and methanol (MeOH) for ca. eight hours. Extracts were evaporated to dryness under reduced pressured at 40 C. The yields are reported in Table 2. 2.3. Microorganisms, broth medium, inoculum and sample preparation 2.3.1. Microorganisms The Candida species used in this study, Candida parapsilosis (ATCC 22019), Candida krusei (ATCC 6258) and Candida albicans (ATCC 90028), were obtained from LGC Promochem. Strains were cultured from frozen stocks and maintained at 37 C on Sabouraud dextrose agar (Sigma-Adrich, UK). 2.3.2. Preparation of broth medium Broth medium (0.5 L) used for the NCCLS test was prepared as follows: RPMI-1640 medium supplemented with glutamine and phenol red, without bicarbonate (10.4 g), 3-(N-morpholino) propanesulfonic acid (MOPS) (34.53 g) were dissolved in 400 ml distilled water, adjusting the pH to 7.0 at 25 C with 1 mol/L sodium hydroxide. Additional water was added to bring the medium to a final volume of 0.5 L, which was filter sterilized and stored at 4 C until required. The medium was further supplemented with glucose (18 g) to achieve a final concentration of 2% glucose (w/v) for the EUCAST assay. All chemicals were obtained from Sigma-Adrich, UK. 2.3.3. Preparation of inocula For the assays, organisms were subcultured once onto Sabouraud dextrose agar and incubated for 24 h at 37 C. Inocula were prepared by transferring several colonies to sterile distilled water (5 ml). The suspensions were mixed for 15 s to ensure homogeneity and subsequently diluted to match the turbidity of a 0.5 McFarland standard (i.e. OD = 0.12–0.15 at k = 530 nm, ThermoSpectronic UV1 X, corresponding to 1–5 · 106 CFU/ml). Further dilutions in sterile distilled water were made to obtain the required working suspensions (1–5 · 105 CFU/ml and 1– 5 · 103 CFU/ml for the EUCAST and NCCLS assays, respectively). Colorimetric readings of the results for the NCCLS assay employed resazurin (Sigma-Adrich, UK) as an indicator of cell growth. For this purpose, the working suspension (20 ml, 1–5 · 103 CFU/ml) was supple- Table 1 Combretum and Teminalia spp. used in this study Plants Voucher number Part used Code Combretum zeyheri Sond. Combretum fragrans F. Hoffm. Combretum elaeagnoides Klotzsch Combretum kirkii M.A. Lawson Terminalia brachystemma Welw. ex Hiern Terminalia mollis M. Laws. SRGH-DRK-5/99-cze SRGH-DRK-5/99-cfr SRGH-DRK-5/99-cel SRGH-DRK-5/99-cki SRGH-DRK-5/99-tbr SRGH-DRK-5/99-tmo Leaves Heart wood Stem Leaves Leaves Leaves CZ CF CE CK TB TM 327 M. Liu et al. / Methods 42 (2007) 325–329 Table 2 Inhibitory activity of plant extracts against Candida spp. using the modified NCCLS method with colorimetric end-point determination Code Extract Yield % MIC (mg/L) C. krusei C.albicans C. parapsilosis CZ MeOH EtOAc MeOH EtOAc MeOH EtOAc MeOH EtOAc MeOH EtOAc MeOH EtOAc 12.73 0.95 8.52 0.28 2.73 0.31 11.18 1.87 NA NA 9.25 1.37 8 250 16 >500 >500 >500 32 >500 8 63 >500 >500 0.06 32 250 32 >500 >500 >500 63 >500 8 125 >500 >500 0.12 63 500 63 >500 >500 >500 250 >500 16 125 >500 >500 0.06 CF CE CK TB TM Itraconazole mented with 0.1ml sterilized solution of resazurin (20 mg/ ml in water). 2.3.4. Preparation of samples Stock solutions of the plant extracts and the positive control drug Itraconazole (Sigma-Adrich, UK) were prepared in dimethyl sulphoxide (DMSO) at the concentrations of 100 mg/ml and 1.6 mg/ml, respectively and further diluted (1:50) in broth. 2.4. Preparation of plates Microdilution susceptibility testing was performed in flat-bottom 96-well clear plates containing broth medium (0.1 ml) in each well. Sample solutions (0.1 ml) were subsequently serially diluted two-fold in the plates with the broth, starting with the final concentration of 500 mg/L for plant extracts and 8 mg/L for Itraconazole. The working inoculum suspension (0.1 ml) was added to give a final inoculum concentration of 0.5–2.5 · 105 and 0.5–2.5 · 103 CFU/ml for the EUCAST and NCCLS assays, respectively. Itraconazole was used as the standard antifungal drug. Sterility and growth controls in the presence of organic solvents employed in sample preparation were also included. No inhibitory effects were observed in the presence of DMSO at the highest concentration used (0.5% v/v). The plates were incubated at 37 C for 24 and 48 h for the EUCAST and NCCLS assays, respectively. 2.5. Results determination Visual readings: The amount of growth in the wells containing the agent was compared visually with the growth in the growth control wells. The concentration with a prominent decrease in turbidity was determined as the MIC. Spectrophotometric readings: a plate reader (Spectra MAX190) was used to measure the amount of growth at k = 530 nm following agitation by pipetting to ensure homogeneity. The backgrounds for each sample and the growth control were also measured. The percentage of growth was calculated by the following equation: % Growth ¼ OD530 Sample  OD530 Corresponding Background OD530 Growth Control  OD530 Corresponding Background  100 MIC50, an inhibition of growth equal to or greater than 50% of that of the growth control was recorded as endpoint. Colorimetric readings: Colorimetric MIC end-points were interpreted as the lowest sample concentration that remained blue (indicating no growth) or the first dilution that changed from blue to slightly purple (equivalent to prominent growth inhibition). All assays were repeated at least three times. Because this assay is developed from the susceptibility test, which is a qualitative test instead of quantitative test, no statistical analysis is involved. The final activity of the plant extracts was presented by the highest MIC of the three tests if they are different in ± one-fold dilution otherwise the assay requires to be repeated further to ensure reproducibility. 3. Results and discussions 3.1. Evaluation of different methods The visual reading method of the NCCLS M27-A2 test was both subjective and less reproducible due to the trailing effect. When testing plant extracts, additional problems were encountered. For example, water-insoluble constituents within the extracts formed a precipitate which could be confused with the cells; the strong colour of most extracts obscured the judgment of the turbidity level. Thus, the data that resulted from this method were discounted. The use of a spectrophotometer helps to obtain an objective and rapid MIC reading. The reproducibility is also good although some negative readings (percentage growth) can be observed due to the interference of the 328 M. Liu et al. / Methods 42 (2007) 325–329 background (the colour of plant extracts) and also due to the turbidity caused by the insoluble compounds. In order to facilitate the analysis and comparison of the results from NCCLS and EUCAST methods, line charts of extract concentration against percentage growth (the mean values from three tests) were presented and the standard deviations were shown as single direction error bars. Fig. 1a shows the results obtained from the NCCLS method with the spectrphotometric end-point determination. As shown in the chart, the MIC50 can be obtained easily and the standard deviation is acceptable for most concentrations. However, this method demanded an agitation of the plates to ensure homogeneity in the wells before reading, which could affect the reproducibility and can be very tedious 160 b 160 140 140 120 120 % of growth control % of growth control a for the operator. A background plate was also necessary to obtain the absorbance contribution from the colour of the plant extracts or the turbidity of insoluble compounds. The EUCAST method gave the poorest results. For plant extracts, the higher inoculum size and glucose supplementation which was used to optimize the growth of Candida species did not show the claimed advantage of simplifying end-point determination of MICs if combined with a spectrophotometric method [16]. The readings were not reproducible; some plant extracts had the percentage growth fluctuating by around 50%, making the determination of MIC50 impossible, while some plant extracts showed a dose-dependent growth-promotion phenomenon (Fig. 1b). The reasons behind this phenomenon are not 100 80 60 40 100 80 60 40 20 20 0 0 500 250 125 62.5 500 31.25 250 125 62.5 31.25 -20 -20 Concentration mg/L Concentration mg/L CZM CFE c CZE CEM CZM CFE CFM CEE CZE CEM CFM CEE Concentration of plant extracts mg/L 500 250 125 62.5 31.25 15.63 8 4 2 1 0.50 0.25 7.81 3.91 1.95 0.98 G S A B C Plant extracts D E F G Positive control H 0.12 0.06 0.03 0.02 Concentration of positive control mg/L Fig. 1. (a) and (b) Results obtained for 6 plant extracts* against Candida krusei using (a) the NCCLS method and (b) the EUCAST method, with the spectrophotometric end-point determination. (c) An example plate showing the results obtained from our modification of the NCCLS method with the colorimetric end-point determination. G, growth control, S, sterility control. From row A to F were plant extracts: CZM, CZE, CFM, CFE, CEM, and CEE. Row G was an untreated row. Row H was Itraconazole treated row. * CZM and CZE, Methanol and EtOAc extract of Combretum zeyheri, CFM and CFE, Methanol and EtOAc extract of C. fragrans, CEM and CEE, Methanol and EtOAc extract of C. elaeagnoides. M. Liu et al. / Methods 42 (2007) 325–329 clear and we have not so far pursued the matter further. The failure of the EUCAST method also highlighted the important influence of inoculum size and medium type in the antifungal test. Good reproducibility and well-defined end-points were achieved by the colorimetric method which also requires less instrumentation and a simpler operating procedure. The deep blue indicator resazurin, which is reduced to bright pink resorufin by viable cells, dramatically reduced the interference from the plant extracts including the colour and the precipitation. Fig. 1c provides an example of a plate with results obtained from the NCCLS method after modification with the colorimetric end-point determination. Resazurin did not affect the efficacy of Itraconazole used here as the positive control and this widely used indicator has no inhibitory activity against Candida spp. [17]. 3.2. Anti-Candida activity of plant extracts Table 2 lists the MIC of 12 extracts from six Combretum and Terminalia species against three Candida species using the NCCLS method with colorimetric end-point determination. In our pre-tests, all the hexane extracts showed no activity, which could have been due to the poor solubility of non-polar ingredients in a water-based medium. However, no further testing was carried out on the hexane extracts. Among the six plants, Terminalia brachystemma, which is traditionally used to treat schistosomiasis [18], showed the highest activity against all three Candida spp. This is in agreement with a previous study [6] and worthy of further phytochemical investigation. Combretum zeyheri was the second most potent plant in our study. It is traditionally used to treat diarrhoea and its stem and root have shown substantial antimicrobial activity [3]. In agreement with the 24 well nunc plate bioassay [5], the polar extract of C. fragrans showed moderate activity in our test although it had no inhibition zone in the agar diffusion test by Masoko et al. [3]. T. mollis showed no activity in contrast to high activity reported by Masoko et al. [6] and again in agreement with the 24 well nunc plate bioassay [7]. Finally, C. eleaegnoides and C. kirkii, have no published data available as far as we are aware. C. eleaegnoides showed no activity while the methanol extract of C. kirkii showed some activity especially against C. krusei and C. albicans. 4. Concluding remarks In this study, we successfully modified the standard antifungal susceptibility test for the screening of plant extracts 329 for their anti-Candida activity using the dye resazurin. This colorimetric broth microdilution assay showed the advantage of good reproducibility and easy operation compared with other available methods. Acknowledgments We thank Allan Drummond for technical support. This work was supported in part by a scholarship from the Faculty of Science, University of Strathclyde and by personal support from Xiao Liu and Peiying Liu. References [1] D. Barret, Biochim. Biophys. Acta (BBA)-Mol. Basis Dis. 1587 (2002) 224–233. [2] D.R. Katerere, A.I. Gray, R.J. Nash, R.D. Waigh, Phytochemistry 63 (2003) 81–88. [3] M. Gelfand, S. Mavi, R.B. Drummond, B. Ndemera, The Traditional Medical Practitioner in Zimbabwe, first ed., Mambo Press, Gweru, 1985. [4] S.M. Carpano, E.D. Spegazzini, J.S. Rossi, M.T. Castro, S.L. Debenedetti, Fitoterapia 74 (2003) 294–297. [5] K. Batawila, K. Kokou, K. Koumaglo, M. Gbéassor, B. de Foucault, Ph. Bouchet, K. Akpagana, Fitoterapia 76 (2005) 264–268. [6] P. Masoko, J. Picard, J.N. Eloff, J. Ethnopharmacol. 99 (2005) 301–308. [7] F. Baba-Moussa, K. Akpagana, P. Bouchet, J. Ethnopharmacol. 66 (1999) 335–338. [8] P. Fyhrquist, L. Mwasumbi, C.A. Hæggström, H. Vuorela, R. Hiltunen, P. Vuorela, J. Ethnopharmacol. 79 (2002) 169–177. [9] National Committee for Clinical Laboratory Standards. Reference method for broth dilution antifungal susceptibility testing of yeasts; approved standard. NCCLS document M27-A2. National Committee for Clinical Laboratory Standards, Pa. Wayne, 2002. [10] M.L. Motsei, K.L. Lindsey, J. van Staden, A.K. Jäger, J. Ethnopharmacol. 86 (2003) 235–241. [11] H. Bouamama, T. Noël, J. Villard, A. Benharref, M. Jana, J. Ethnopharmacol. 104 (2006) 104–107. [12] L.V. Buwa, J. van Staden, J. Ethnopharmacol. 103 (2006) 139–142. [13] M.A. Pfaller, S.A. Messer, S. Coffmann, J. Clin. Microbiol. 33 (1995) 1094–1097. [14] B.A. Arthington-Skaggs, W. Lee-Yang, M.A. Ciblak, J.P. Frade, M.E. Brandt, R.A. Hajjeh, L.H. Harrison, A.N. Sofair, D.W. Warnock, Antimicrob. Agents Chemother. 46 (2002) 2477–2481. [15] European Committee on Antibiotic Susceptibility Method for determination of minimal inhibitory concentration (MIC) by broth dilution of fermentative yeasts. Discussion document E. Dis. 7.1. European Society of Clinical Microbiology and Infectious Diseases, Taufkirchen, Germany, 2002. [16] M. Cuenca-Estrella, W. Lee-Yang, M.A. Ciblak, B.A. ArthingtonSkaggs, E. Mellado, D.W. Warnock, J.L. Rodriguez-Tudela, Antimicrob. Agents Chemother. 46 (2002) 3644–3647. [17] R.N. Tiballi, X. He, L.T. Zarins, S.G. Revankar, C.A. Kauffman, J. Clin. Microbiol. 33 (1995) 915–917. [18] P. Mølgaard, S.B. Nielsen, D.E. Rasmussen, R.B. Drummond, N. Makaza, J. Andreassen, J. Ethnopharmacol. 74 (2001) 257–264.