170 T.S. Yam et al. / FEMS Microbiology Letters 152 (1997) 169^174 Table 1 Antimicrobial spectrum of green tea extract, 2% nominal Bacterial species S

Transcription

1 FEMS Microbiology Letters 152 (1997) 169^174 Microbiological activity of whole and fractionated crude extracts of tea (Camellia sinensis), and of tea components T.S. Yam 1, Saroj Shah, J.M.T. Hamilton-Miller * Department of Medical Microbiology, Royal Free Hospital School of Medicine, London NW3 2QG, UK Received 4 January 1997; accepted 23 April 1997 Abstract Aqueous extracts of teas (Camellia sinensis) of different types and from various sources inhibited a wide range of pathogenic bacteria, including methicillin-resistant Staphylococcus aureus. Tea extracts were bactericidal to staphylococci and Yersinia enterocolitica at well below `cup of tea' concentrations. Activity was confined to one of four fractions obtained from a green tea extract by partition chromatography. Testing of pure tea compounds and closely related chemicals suggested that the antibacterial activity of extracts of green tea can be explained by its content of epigallocatechin, epigallocatechin gallate and epicatechin gallate. In black tea extracts, theaflavin and its gallates are additional antibacterially active components. Keywords: Tea (Camellia sinensis); Crude extract; Staphylococcus aureus 1. Introduction Aqueous extracts of the dried and processed leaf of Camellia sinensis, commonly called `tea', are known to have a remarkable range of pharmacological activities [1]. However, information about the antimicrobial action of tea is scattered widely through the literature, some data are contradictory and no systematic study appears to have been carried out. We report here our ndings on some antibacterial * Corresponding author. Tel.: +44 (171) ; Fax: +44 (171) Present address: Department of Medicine, Craigavon Area Hospital, 68 Lurgan Rd, Portadown BT63 5QQ, UK. e ects of tea against a wide range of pathogenic bacteria, and likely identi cation of some active principles in the extract. 2. Materials and methods 2.1. Chemicals The following known components of tea extracts [2] or closely related compounds were obtained: (3)epicatechin (EC), (+)catechin (C), ca eine, chlorogenic acid, gallic acid, theobromine, theophylline, quercitin and kaempferol from Sigma; (+)gallocatechin (GC), (3)epigallocatechin (EGC), (3)epigallocatechin gallate (EGCG), (+)catechin gallate (CG), (3)epicatechin gallate (ECG), (3)epiafzelechin (EA) and its gallate (EAG), theasinensin A, thea a / 97 / $17.00 ß 1997 Federation of European Microbiological Societies. Published by Elsevier Science B.V. PII S (97)

2 170 T.S. Yam et al. / FEMS Microbiology Letters 152 (1997) 169^174 Table 1 Antimicrobial spectrum of green tea extract, 2% nominal Bacterial species Staphylococcus aureus (including 18 MRSA b ) 33/33 S. epidermidis 38/38 S. saprophyticus 16/16 Streptococcus pyogenes 5/5 S. pneumoniae (including 2 penicillin-resistant) 4/4 Neisseria spp. 2/2 Corynebacterium spp. 3/3 Bacillus spp. 3/3 Klebsiella pneumoniae 12/20 Helicobacter pylori 1/1 Pseudomonas aeruginosa 10/10 Stenotrophomonas maltophilia 4/4 Acinetobacter spp. 4/4 Serratia marcescens 8/8 Enterobacter spp. 22/25 Citrobacter spp. 11/14 Five species c in tribe Proteae 28/28 Salmonella spp. 3/3 Yersinia enterocolitica 2/2 Haemophilus in uenzae 2/2 Campylobacter jejuni 2/2 Streptococcus agalactiae 0/5 Enterococcus faecalis 0/4 Listeria monocytogenes 0/7 Mycobacterium spp. d 0/4 Clostridium spp. 0/4 Bacteroides spp. 0/3 Escherichia coli 8/20 Candida albicans 0/4 Filamentous fungi e 0/8 a Zone size of at least 10 mm around well containing extract. b Meticillin-resistant strains. No. of strains sensitive a /no. tested c Five Proteus mirabilis, 5Pr. vulgaris, 5Morganella morgani, 8Providencia stuartii, 5Prov. rettgeri. d Two M. fortuitum, 1M. kansasii, 1M. tuberculosis. e Two Microsporum spp., 1 each Fusarium, Trichophyton, Penicillium, Scopulariopsis, Acremonium, Aspergillus. vin, its 3-gallate and 3,3P-digallate were generous gifts from Unilever Ltd., Colworth, UK Crude extracts of tea Di erent brands of the three basic types of tea ^ green (Japanese), semi-fermented (Oolong) and black (Indian and Chinese) ^ were purchased from retail shops. Crude aqueous infusions were made by adding 2 g of leaf to 100 ml boiling water, allowing the suspension to stand for 10 min and removing solid matter by ltration. The infusions obtained (that were approximately twice the strength of a normal `cup of tea') were either used immediately or freezedried and reconstituted when required. Only fresh extracts (or freshly made-up freeze-dried material) were used in these experiments, as marked chemical changes occur when tea is allowed to stand [3] Fractionation of crude extract The method of Robertson and Bendall [4] was followed. 30 g of crushed Sencha leaf (Japanese green tea) was extracted into 500 ml boiling water. The ltrate was reduced to about 100 ml in a rotary evaporator, and extracted 5 times with equal volumes of chloroform (to remove ca eine and pigments). The aqueous phase was clari ed by centrifu-

3 T.S. Yam et al. / FEMS Microbiology Letters 152 (1997) 169^ Table 2 Inhibitory and bactericidal activities of 2% nominal extract of green tea Bacterial strain MNIC (mg ml 31 ) MIC (mg ml 31 ) MBC (mg ml 31 ) S. aureus USA a (0.05) S. aureus Oxford 0.17 b S. epidermidis a (0.11) Y. enterocolitica a (0.05) a Mean (with S.D.) for 6separate extracts. b Mean of 3 determinations. gation and then extracted 8 times with 200 ml volumes of ethyl acetate. The organic phases were combined, concentrated to about 30 ml in a rotary evaporator, 30 ml water added and the remaining ethyl acetate evaporated. The aqueous phase was then freeze-dried, giving 3.6g of solids, 1.2 g of which was dissolved in 1.5 ml methanol, and an equal volume of chloroform: petrol (1: 1) added. This mixture was loaded onto a column (2.5 cmu47 cm) of Sephadex LH-20 that had been equilibrated in chloroform:methanol:petrol (1:2:1), and eluted with the same solvent under gravity at 1 ml min 31, 10 ml fractions being collected until 1.5 l eluate had emerged. Fractions were monitored at E 280 nm, and those making up the four major peaks (A^D) were pooled Microbial strains These were from our research collection. All were clinical isolates Media These were obtained from Unipath, Basingstoke, UK Determination of antimicrobial activity The antimicrobial spectrum of tea was determined qualitatively for the non-mycobacterial species and for Candida albicans in terms of zone sizes around wells, cut in plates of IsoSensitest agar (supplemented as necessary) surface-inoculated with V10 5 cfu of various microbial species, containing 0.1 ml of extract. Strains were designated in arbitrary way as `sensitive' or `resistant' if the zones were respectively v10 or 6 10 mm in diameter after overnight incubation at 37³C under appropriate conditions. Filamentous fungi were tested in a similar way, on Sabouraud agar incubated at 30³C. Mycobacterium kansasii and M. tuberculosis H37Rv were tested in the Bactec using 7H12 medium, and M. fortuitum on Mueller Hinton agar for their ability to grow in the presence of a 1:10 dilution of tea extract (V0.6mg solids ml 31 ). Quantitative assays were carried out using a similar hole-plate assay, using various dilutions of the crude extract, and determining the `maximum non-inhibitory concentration' (MNIC) against each strain from plots of zone diameter vs log concentration (dry weight ml 31 ). The po- Table 3 Bactericidal activity of 2% nominal extract of green tea Bacterial species Time (h) required to achieve 99.9% kill a by indicated multiple of MNIC b U1 U2 U4 U8 S. aureus 12 s S. epidermidis Y. enterocolitica 1 s 24 s a Data obtained by interpolation from time-kill plots. b Numerical values of MNIC (mg ml 31 ) given in Table 2.

4 172 T.S. Yam et al. / FEMS Microbiology Letters 152 (1997) 169^174 Fig. 1. Elution pro le of crude polyphenol extract of Sencha tea on Sephadex LH-20. tency of extracts, in units ml 31 (1 unit being de- ned, arbitrarily, as that concentration giving a zone of inhibition of 15 mm diameter against S. aureus Oxford), could also be calculated from these plots. Minimum inhibitory concentrations (MICs) were determined by tube dilution in IsoSensitest broth, with an inoculum of 10 5 cfu per tube. Bactericidal activity was measured by determining the minimum bactericidal concentration (MBC), by sub-culture of 0.2 ml from tubes used in MIC experiments (see above), and by a time-kill method, as previously described [5] Serum binding MNIC was determined against S. aureus in the presence and absence of 5% human serum. % binding was calculated as: [(MNIC with serum3mnic without serum)/mnic with serum]u Results and discussion 3.1. Biological activity of di erent teas 2% (nominal) extracts were made from 12 black

5 T.S. Yam et al. / FEMS Microbiology Letters 152 (1997) 169^ teas (11 Indian, 1 China), 1 semi-fermented, and 4 green teas. For each extract, the concentration of solids was determined by dry weight measurement, and potency against S. aureus Oxford assayed. While the concentrations of the extracts were rather similar (range 6.5^4.1 mg ml 31, mean 5.5 þ 0.75 S.D.), the antibacterial potency varied widely (range 2.9^0.2 U ml 31, median 1.85). The greatest potency was shown by Red Label (an Indian tea), but Japanese Sencha was chosen for further study, as its activity was almost as great (2.7 U ml 31, provided the leaves were crushed to a powder prior to extraction), and the chemical composition of a green tea infusion is much simpler than that of a black tea [2]. This latter property is important with respect to the eventual isolation of active principle(s). Solids were extractable from crushed Sencha in a consistent manner; in 7 separate preparations, 2% nominal extracts yielded a mean of 6.5 mg solids ml 31 (S.D. 0.1) Antimicrobial spectrum Many important pathogens were sensitive to the Sencha extract (2% nominal, concentration of solids 6mgml 31 ), as shown in Table 1. Most species tested were homogeneous in their susceptibility (e.g. S. aureus, all of which were sensitive, Listeria monocytogenes, all of which were resistant), but there was some variation among di erent strains of Escherichia coli and Klebsiella pneumoniae. The pattern of activity was not like that of an antiseptic, but was di erential within Gram-negative and Gram-positive species, cocci and rods; this suggests a mode of action other than non-speci c denaturation of proteins, known to be a property (`astringency') of certain tea components [2] Quantitation of antibacterial activity Judged by the sizes of inhibition zones, staphylococci and Yersinia enterocolitica were the most sensitive Gram-positive and Gram-negative genera, respectively, of those shown on Table 1. Further tests were made on S. aureus USA12 (a meticillin-resistant strain), S. epidermidis 17 and Y. enterocolitica 1; MNIC, MIC and MBC are shown in Table 2. Reconstituted freeze-dried material had the same activity as fresh brew (data not shown). Time-kill plots showed that the extract killed both the staphylococcal strains and Y. enterocolitica in a concentration-dependent fashion, so that after 24 h, U4 and U8 respective MNICs had sterilised the cultures. The contact times with di erent concentrations of tea extract required to kill 99.9% of the original inocula are shown on Table Serum binding MNIC for S. aureus was increased almost 5-fold, from 0.2 to 1.13 mg ml 31, giving a value for serum binding of 82% Antibacterial activity of tea fractions and of pure tea components The elution pro le from the column is shown as Fig. 1. Activity (against S. aureus Oxford) was con- ned to peak D. We were not able to analyse the chemical content of the various fractions. GC, ECG, EGC, EGCG, EAG, thea avin and its 3-gallate and 3,3P-digallate all had mean MNICs for S. aureus of between and 0.05 mg ml 31. CG, theasinensin A and quercitin were less active, having MNICs of 130^180 mg ml 31, while EA, EC, C, caffeine, chlorogenic acid, gallic acid, theobromine and theophylline had no activity at 1 mg ml 31. Kaempferol was inactive at 0.2 mg ml 31, the maximum concentration that could be tested. These results are consistent with the notion that gallocatechins and their gallates are the main chemical moieties responsible for the antibacterial activity of green tea extracts [6]; the epi con guration seems in general to be the more active. As only very small amounts of the active compounds were available, we were unable to carry out more extensive investigations. It is interesting that the component parts of the active compounds EGC, EGCG, CG and ECG, namely gallic acid, C and EC, were without biological activity even at a concentration of 1 mg ml 31. In extracts of black tea, thea avin and its gallates (which are absent from green tea) will make an additional contribution to the antibacterial activity. Further studies are needed to elucidate the mode of action of tea extracts, as well as investigating the

6 174 T.S. Yam et al. / FEMS Microbiology Letters 152 (1997) 169^174 relative activities of di erent catechins and the more complex tea components such as thea avins and thearubigins. However, such work will not be possible unless these compounds become generally available. Acknowledgments We are grateful to Dr P. Collier, Dr A. Davies and Dr Y. Cai of Unilever Research Colworth Laboratory for helpful discussions and supplying pure tea components. References [1] Hamilton-Miller, J.M.T. (1995) Antimicrobial properties of tea (Camellia sinensis L.). Antimicrob. Agents Chemother. 39, 2375^2377. [2] Kirk, R.E. and Othmer, D.F. (1980) Encyclopedia of Chemical Technology, 3rd edn., Vol. 22, pp. 628^648. Wiley, New York. [3] Spiro, M. (1995) What's in a nice cup of tea? Educ. Chem. 32, 11^13. [4] Robertson, A. and Bendall, D.S. (1983) Production and HPLCanalysis of black tea thea avins and thearubigins during in vitro oxidation. Phytochemistry 22, 883^887. [5] Brum tt, W., Hamilton-Miller, J.M.T. and Shah, S. (1992) Invitro activity of RP 59500, a new semisynthetic streptogramin antibiotic, against Gram-positive bacteria. J. Antimicrob. Chemother. 30 (Suppl. A), 29^37. [6] Toda, M., Okubo, S., Ikigai, H. and Shimamura, T. (1990) Antibacterial and antihaemolysin activities of tea catechins and their structural relatives. Jap. J. Bacteriol. 45, 561^566.