International Journal of Antimicrobial Agents 22 (2003) 48 Á/53 www.ischemo.org Berberis aetnensis C. Presl. extracts: antimicrobial properties and interaction with ciproﬂoxacin Rosario Musumeci a,*, A. Speciale a, R. Costanzo a, A. Annino b, S. Ragusa c, A. Rapisarda d, M.S. Pappalardo b, L. Iauk aa Department of Microbiological and Gynaecological Sciences, Section of Microbiology, University of Catania,Via Androne 81, 95124 Catania, Italy b Department of Pharmaceutical Sciences, University of Catania, Catania, Italy c Department of Pharmacobiological Sciences, University of Catanzaro, Catanzaro, Italy d Pharmacobiological Department, University of Messina, Messina, Italy Received 27 August 2002; accepted 16 December 2002Abstract Previous research showed that berberine-containing Berberis species synthesise the substances 5?-methoxyhydnocarpin-D (5?-MHC-D) and pheophorbide a , which have no antimicrobial activity but inhibit the expression of multidrug resistant efflux pumps(MDRs) in Staphylococcus aureus and potentiate the action of berberine. The MDR pumps extrude synthetic and naturalantimicrobials from bacterial cells. We searched for these compounds in Berberis aetnensis C. Presl. (Berberidaceae), an endemicplant of the volcano Mount Etna. This work confirms the presence of pheophorbide a and permits us to hypothesise the presence of5?-MHC-D in leaf extracts. In fact, the activity of ciprofloxacin was improved when two chromatographic fractions isolated fromleaf extracts were added. These results are indicative of the presence of MDR pump inhibitors. Moreover, crude extracts were testedon several micro-organisms and showed antimicrobial activity mainly against Gram-positive bacteria and yeasts.# 2003 Elsevier Science B.V. and the International Society of Chemotherapy. All rights reserved.Keywords: Berberis aetnensis ; Berberine; Efﬂux pump inhibitors; Pheophorbide a ; 5?-MHC-D1. Introduction studies regarding a possible anti-tumour activity [1 Á/5]. The most prominent clinical uses include bacteria In traditional medicine the extracts of various Berber- related diarrhoea, parasitic intestinal infections andidaceae (Berberis aquifolium , Berberis vulgaris and ocular infections (conjunctivitis, trachoma) [2,6]. AsBerberis aristata) are used for rheumatic complaints Stermitz et al. [7,8] stated, the weak antibacterial activityand other types of chronic inflammations . Some of berberine in some bacterial strains could be due to theauthors demonstrated that these extracts have a sig- expression of multidrug resistant efflux pumps (MDRs)nificant activity against bacteria, viruses, fungi, proto- that extrude the drugs, rather than their intrinsiczoa, helminthes and chlamydia . Studies carried out activity. MDR efflux pumps are important mechanismson the properties and chemical composition of the of active and unidirectional transport. Their presenceextracts show that their principal activity is due to their makes bacterial strains resistant to various classes ofalkaloid constituents with an isoquinolinic nucleus such antibiotics and prevents the intracellular action ofas berberine, oxyacanthine, berbamine and palmatine berberinic alkaloids that are thus promptly expelled. It has been shown that berberine has febrifugal, . Several studies have shown that plants of the genushypotensive, immuno-stimulating, anti-inflammatory Berberis (Berberis repens , B. aquifolium and Berberisand antimicrobial properties and there are on-going fremontii ) producing berberine (Fig. 1a) synthesise two substances, the flavonolignan 5?-MHC-D and the por- * Corresponding author. Tel.: /39-095-31-2386; fax: /39-095-32- phyrin pheophorbide a (Fig. 1b and c), which have no5032. antibacterial activity but have an inhibiting property E-mail address: email@example.com (R. Musumeci). against MDR efflux pumps found so far in Staphylo-0924-8579/03/$30 # 2003 Elsevier Science B.V. and the International Society of Chemotherapy. All rights reserved.doi:10.1016/S0924-8579(03)00085-2
R. Musumeci et al. / International Journal of Antimicrobial Agents 22 (2003) 48 Á/53 49 by the vegetation that is characteristic of the volcanic slopes, Astragaletum siculi : xerophyte and thorny plants are prevalent. It is believed that this form of Astraga- letum established itself in the area where once forests were dominant; it is thus secondary vegetation, in substitution [14,15]. In the past some authors considered the species aetnensis a simple adaptation of B. vulgaris due to the dry conditions of the South . Currently the Kewensis Index  gives the species aetnensis its own characteristics and distinguishes it from the species vulgaris . 2. Materials and methods 2.1. Plant material Roots of B. aetnensis C. Presl. were collected on the slopes of the volcano Mount Etna (1800 Á/1900 m) in November 2001, leaves were collected in May 2001. B. aetnensis specimens were obtained thanks to the Re- gional Forest Corps Detachment of Catania-Nicolosi. A voucher specimen of the plant was deposited in the herbarium of the Pharmacobiological Department of the University of Messina (Italy). The fresh material was air-dried and powdered. 2.2. Preparation of extracts For the preparation of the ethanol and ether extractsFig. 1. Chemical structure of berberine. Chemical structure of 5?- of roots, and ethanol and chloroform extracts of leavesMHC-D. Chemical structure of pheophorbide a . of B. aetnensis exhaustive extraction of 100 g of single drugs was carried out at room temperature by macera-coccus aureus (allowing berberine to carry out its tion for 24 h on a rotating shaker with 600 ml of ethanolactivity) [7,8,10]. Because of the results obtained with 70%, with ether and with chloroform . The extractsthe Berberidaceae so far, we assayed the antimicrobial were then filtered and dried under vacuum. The residueactivity and verified the presence of eventual MDR of the ethanol and ether extracts of the roots were 4.64efflux pump inhibitors in Berberis aetnensis C. Presl., a and 0.215 g, respectively; the ethanol and chloroformberberidaceae species endemic to Mount Etna, Sicily, extracts of the leaves were 10.31 and 12.58 g, respec-Italy. Plant material was collected in the A zone of the tively. The residues of the extracts were solubilised inEtna Volcano Park. B. aetnensis C. Presl is a bushy- dimethylsulphoxide (DMSO).spiny shrub with dense twisted branches whose grey,smooth bark has fine longitudinal channels. Leaves on 2.3. Chromatographic analysis of the chloroform extractlong branches are transformed into spines 3-fid, the One hundred grams of dry powdered leaves of B.shorter branches have normal leaves. The leaves are aetnensis were first steeped in hexanes (drug/solventoblong-egg shaped, rigid, coriaceous and have spiny- ratio 1:6 w/v) for 24 h, filtered under vacuum and thenserrated margins. The flowers, yellow in colour and are steeped in CHCl3 (drug/solvent ratio 1:5 w/v) for 24 hgrouped in hanging racemes. The fruit is an oblong and subjected to a final filtration. Twenty microlitres ofberry, dark red in colour [11 Á/13]. As soon as they are chloroform extract of B. aetnensis leaves (test solution)cut, the branches and the roots have a penetrating odour were used for thin layer chromatographic (TLC) analy-and are intensely yellow in colour. B. aetnensis grows on sis. A standard solution was prepared dissolving 1 mg ofrocky slopes, on the rocky bed of water courses, at the commercial pheophorbide a (ICN) in 10 ml of CHCl3upper limit of the wooded zone and beyond. In Sicily it (reference solution). Twenty microlitres of test solutionis common on Mount Etna, in the area of vegetation and 10 ml of reference solution were laid on chromato-between 1800 and 2450 Á/2500 m. This zone is dominated graphic silica gel 60 F254 pre-coated TLC plates (Merck,
50 R. Musumeci et al. / International Journal of Antimicrobial Agents 22 (2003) 48 Á/53Germany). The mobile phase was made up of aerobic incubation for 18Á/24 h at 37 8C, the bacterialchloroform Á/methanol (9.5:0.5). The developed plate and fungal cultures were diluted to a turbidity of 0.5was then examined by UV lamp at 254 and 365 nm. McFarland (1.5 )/108 CFU/ml) and further diluted inRatio front (Rf) values were calculated from the extract saline solution to obtain an inoculum of 5 )/105 CFU/spots, corresponding to that of the reference solution. well in a final volume of 100 ml. The plates wereThe chloroform extract was separated by flash chroma- incubated aerobically at 37 8C for about 18 h. Thetography on silica gel 60 (230 Á/400 mesh, supplied by MIC is defined as the lowest concentration at whichMerck, Germany) eluting with chloroform Á/methanol there was no visible growth after incubation at 37 8C for(9.5:0.5). Two intensely coloured fractions were recov- 18 h. With the aim of verifying eventual synergisticered, one yellow and the other green. These fractions activity of ciprofloxacin with different dilutions of ourwere dried, weighed and then solubilised in DMSO, chromatographic fractions, the activity of these associa-obtaining two solutions with concentrations of 7000 (Y, tions was compared with that of ciprofloxacin plusyellow solution) and 13 500 mg/ml (G, green solution). commercial pheophorbide a (0.5 mg/ml) whose whichThe two solutions were diluted in each well to obtain the antimicrobial activity was also tested.following final concentrations: 233 mg/ml (Y 233), 23.3mg/ml (Y 23.3) and 0.76 mg/ml (Y 0.76) for solution Yand 450 mg/ml (G 450), 45 mg/ml (G 45) and 4.5 mg/ml 3. Results(G 4.5) for solution G. 3.1. Activity of the extracts2.4. Strains The root and leaf extracts generally showed a greater The activity of B. aetnensis extracts were assayed activity against Gram-positive bacteria and yeasts thanagainst seven standard Gram-positive and Gram-nega- against Gram-negative bacteria, except for P. aerugi-tive bacterial strains: S. aureus ATCC 29213, Bacillus nosa CF 03 (MIC 0/312 mg/l). The lowest MIC valuesubtilis ATCC 6603, Enterococcus faecalis ATCC obtained with the ethanol extract of roots was 312 mg/l29212, Escherichia coli ATCC 30218, E. coli ATCC against S. epidermidis . S. aureus ATCC 29213, two25922, Pseudomonas aeruginosa ATCC 27853, Steno- strains of S. epidermidis , B. subtilis ATCC 6603 and thetrophomonas maltophilia NCTC 10257 and against 14 three species of Candida were inhibited at 625 mg/l. Instrains of nosocomial origin: two strains of S. aureus (1 most cases the ether extract was more efficacious thanMet-S, 1 Met-R); four strains of Staphylococcus epider- the ethanol extract except for one strain of S. epidermi-midis (2 Met-S, 2 Met-R); three strains of E. coli ; four dis (MIC ]/10 000 mg/l), C. albicans ATCC 3183 and C.strains of P. aeruginosa (three from patients affected by parapsilosis ATCC 22019 (MIC 0/1250 mg/l in bothcystic fibrosis-CF) and Hafnia alvei . We also assayed C. cases). With the root ethanol extract the lowest MICalbicans ATCC 3183, C. parapsilosis ATCC 22019 and value was found against E. faecalis ATCC 29212 and C.C. krusei ATCC 6258. krusei ATCC 6258 (MIC 0/78 mg/l in both cases). The lowest MIC value for the leaf ethanol extract was 12502.5. Antimicrobials mg/l against S. epidermidis strains. In the remaining cases there was no difference in the activity of the We used ciprofloxacin (Cip; Bayer), amphotericin B extracts against Gram-positive, Gram-negative bacteria(AMB; Sigma) and berberine chloride (Sigma). The and yeasts. The chloroform extract of leaves was moresusceptibility breakpoints for ciprofloxacin were chosen active (MIC range0/39Á/2500 mg/l) than the ethanol oneaccording to the National Committee for Clinical (MIC range 1250Á/10 000 mg/l) both against bacterialLaboratory Standard (NCCLS 2001) . For B. strains and yeasts. The lowest MIC value obtained withsubtilis, studies are not yet adequate to develop repro- this extract against Gram-positives was 78 mg/l for bothducible results. As regards amphotericin B, the sensitiv- B. subtilis ATCC 6603 and E. faecalis ATCC 29212. S.ity breakpoints for C. albicans , C. parapsilosis and C. aureus ATCC 29213, the four strains of S. epidermidiskrusei were chosen according to the NCCLS 1997 suppl. and H. alvei were inhibited at concentrations of 156 mg/M-27A . l. Among the yeasts, the chloroform extract of leaves was particularly active against C. krusei ATCC 62582.6. Determination of the minimum inhibitory with a MIC of 39 mg/l, the lowest value found in ourconcentrations (MICs) study (data not shown in Table). The MICs for the 21 bacterial strains and the three 3.2. TLC analysis of the chloroform leaf extractCandida strains were determined following the methodof double-serial microdilution, conforming to the pro- TLC analysis of the chloroform leaf extract (Fig. 2)cedures recommended by the NCCLS 2001 . After showed a green spot (Rf 0/0.95) corresponding to the
R. Musumeci et al. / International Journal of Antimicrobial Agents 22 (2003) 48 Á/53 51Fig. 2. TLC analysis of the chloroform leaf extract of B. aetnensis C. Presl. (Lane A). The reference compound (Lane B) is a sample of pheophorbidea (eluting system CHCl3/MeOH 9.5:0.5).lipid constituents; a yellow spot (Rf 0/0.42) and a green 450. The most interesting results were seen against S.spot (Rf 0/0.24) corresponding to the reference pheo- aureus MR Br11, which, initially resistant to ciproflox-phorbide a. Based on these results, we hypothesise that acin (MIC 0/32 mg/l), was inhibited by 1 mg/l usingthe yellow spot with Rf 0/0.42 (Lane A) is the 5?-MHC- Cip/Y 233 and by 2 mg/l using Cip/G 450 associa-D and, moreover, we confirm the presence of pheo- tions. Of the Gram-negatives, P. aeruginosa P 54,phorbide a in the chloroform leaf extract of B. aetnensis initially resistant (MIC 0/64 mg/l), was inhibited by B/C. Presl. (Lane A) compared with the commercial 0.5 mg/l of both Cip/Y 233 and Cip/G 450 associa-compound (Lane B). tions. For a more accurate evaluation of the MIC values of the associations tested, we used dilutions Y 23.3 and3.3. Activity of the chromatographic fractions G 45 and the results are shown in Table 2. The synergistic activity of the two associations was con- The solutions at the lowest concentration prepared firmed for all the bacterial strains except for S.from the two coloured fractions isolated by chromato- epidermidis MR 26 and P. aeruginosa CF 09. Moreover,graphy, namely Y 0.76 and G 4.5, showed no synergistic of the the Gram-positive bacteria, the most interestingaction with ciprofloxacin and pheophorbide a (data not results were observed for S. aureus MS DG630; in fact,shown). The results of the activity of ciprofloxacin and both Cip/Y 23.3 and Cip/G 45 dilutions reduced theof its associations with dilutions Y 233 and G 450 and MIC of ciprofloxacin by 2 to 5/0.015 mg/l. As regardswith commercial pheophorbide a (PPB 0.5) against the Gram-negative microrganisms the above mentionedGram-positive and Gram-negative microrganisms are dilutions reduced the MIC of ciprofloxacin of P.shown in Table 1. For most Gram-positive and Gram- aeruginosa P 54, initially resistant, both by 64 Á/0.25negative strains there was an increase in the activity of mg/l. These results were compared with those of theciprofloxacin as demonstrated by the values of the MICs association of ciprofloxacin with commercial pheophor-when it was associated both with dilutions Y 233 and G bide a .
52 R. Musumeci et al. / International Journal of Antimicrobial Agents 22 (2003) 48 Á/53Table 1MICs (expressed in mg/l) of ciproﬂoxacin (Cip alone), Cip/Y 233, Cip/G 450, Cip/PPB 0.5 and commercial pheophorbide a (PPB alone) againstGram-positive and Gram-negative bacteriaBacterial strains MIC (mg/l) Cip alone Cip/Y 233 Cip/G 450 Cip/PPB 0.5 PPB aloneS. aureus ATCC 29213 0.25 5/0.5 5/0.5 5/0.015 500S. aureus MS DG630 2 5/0.5 5/0.5 5/0.015 250S. aureus MR Br11 32 1 2 5/0.015 250S. epidermidis MS 29 2 1 1 0.25 250S. epidermidis MR 26 0.25 5/0.5 5/0.5 0.12 4E. coli ATCC 30218 0.03 5/0.5 5/0.5 5/0.015 64E. coli ATCC 25922 0.015 5/0.5 5/0.5 5/0.015 64E. coli 116 0.015 5/0.5 5/0.5 5/0.015 125E. coli 131 2 5/0.5 5/0.5 0.06 125E. coli 147 8 1 5/0.5 0.5 125P. aeruginosa ATCC 27853 0.5 5/0.5 5/0.5 0.5 125P. aeruginosa P54 64 5/0.5 5/0.5 0.5 125P. aeruginosa CF 03 2 5/0.5 5/0.5 5/0.015 125P. aeruginosa CF 04 2 5/0.5 5/0.5 5/0.015 125P. aeruginosa CF 09 0.06 5/0.5 5/0.5 0.06 1254. Discussion antimicrobial activity of our ether extract could be explained by its better quantitative extraction of berber- Comparing the activity of the ethanol root extract ine-alkaloids and terpenoids present in B. aetnensis .with that of the ethanol leaf extract of B. aetnensis , it The chloroform gave the lowest MIC values, particu-clear that the former had greater antibacterial activity larly against Gram-positive bacteria, thus demonstrat-against Gram-positive microrganisms and the Candida ing a superior antimicrobial activity.strains. Moreover, S. epidermidis and C. krusei were the The antimicrobial activity of ciprofloxacin was im-microrganisms most susceptible to all our extracts. The proved when it was added to the dilutions of theGram-negative bacteria, perhaps due to their external chromatographic fractions of the chloroform leaf ex-lipopolysaccharide structure, were more resistant. Ether tract of B. aetnensis . These results confirm that ouris among the solvents with the greatest extraction extracts contain pheophorbide a and allow us toactivity for alkaloids and terpenoids, . The good hypothesise the presence of 5?-MHC-D.Table 2MICs (expressed in mg/l) of ciproﬂoxacin (Cip alone), Cip/Y 23.3, Cip/G 45, Cip/PPB 0.5 and commercial pheophorbide a (PPB alone) againstGram-positive and Gram-negative bacteriaBacterial strains MIC (mg/l) Cip alone Cip/Y 23.3 Cip/G 45 Cip/PPB 0.5 PPB aloneS. aureus ATCC 29213 0.25 5/0.015 5/0.015 5/0.015 500S. aureus MS DG630 2 5/0.015 5/0.015 5/0.015 250S. aureus MR Br11 32 1 2 5/0.015 250S. epidermidis MS 29 2 1 1 0.25 250S. epidermidis MR 26 0.25 0.25 0.25 0.12 4E. coli ATCC 30218 0.03 5/0.015 5/0.015 5/0.015 64E. coli ATCC 25922 0.015 5/0.015 5/0.015 5/0.015 64E. coli 116 0.015 5/0.015 5/0.015 5/0.015 125E. coli 131 2 0.25 0.06 0.06 125E. coli 147 8 1 0.25 0.5 125P. aeruginosa ATCC 27853 0.5 0.12 0.12 0.5 125P. aeruginosa P54 64 0.25 0.25 0.5 125P. aeruginosa CF 03 2 5/0.015 5/0.015 5/0.015 125P. aeruginosa CF 04 2 5/0.015 5/0.015 5/0.015 125P. aeruginosa CF 09 0.06 0.06 0.06 0.06 125
R. Musumeci et al. / International Journal of Antimicrobial Agents 22 (2003) 48 Á/53 53 The improvement in the antibacterial activity of constituent of Mahonia aquifolium . Complement Altern Med 2002;2:2.ciprofloxacin, when associated with fractions Y (Y 233  Khosla PK, Neeraj VI, Gupta SK, Satpathy G. Berberine, aand Y 23.3) and G (G 450 and G 45), is attributable to potential drug for trachoma. Rev Int Trach Pathol Ocul Tropthe inhibiting properties of pheophorbide a and 5?- Subtrop Sante Publique 1992;69:147 Á/65.MHC-D against the NorA specific MDR pumps found  Stermitz FR, Lorenz P, Tawara JN, Zenewicz LA, Lewis K.in S. aureus [7 Á/10]. Synergy in a medicinal plant: antimicrobial action of berberine potentiated by 5?-methoxyhydnocarpin, a multidrug pump in- hibitor. Proc Natl Acad Sci 2000;97:1433 Á/7.  Stermitz FR, Tawara-Matsuda J, Lorenz P, Mueller P, ZenewiczAcknowledgements L, Lewis K. 5?-Methoxyhydnocarpin-D and pheophorbide A: Berberis species components that potentiate berberine growth The authors thank the Regional Forest Corps De- inhibition of resistant Staphylococcus aureus . J Nat Prodtachment of Catania, Sicily, in particular the Comman- 2000;63:1146 Á/9.dant of the Detachment of Catania-Nicolosi, Warrant  Marshall NJ, Piddock LJV. Antibacterial efﬂux systems. Micro- biologia 1997;13(3):285 Á/300.Ofﬁcer Gianluca Ferlito, for his helpful and effective  Stermitz FR, Beeson TD, Mueller PJ, Hsiang JF, Lewis K.collaboration. Staphylococcus aureus MDR efﬂux pump inhibitors from a Berberis and a Mahonia (sensu strictu) species. Biochem Syst Ecol 2001;29:793 Á/8.References  Tutin TG, Heywood VH, Burges NA, et al. Flora Europaea, vol. I. Cambridge: University Press, 1964. Ivanovska N, Philipov S. Study of the anti-inﬂammatory action of  Pignatti S. Flora d’Italia, vol. I. Edagricole, 1982. Berberis vulgaris root extract, alkaloid fractions and pure  Index Kewensis 2.0. Oxford University Press, 1997. alkaloids. Int J Immunopharmacol 1996;18:553 Á/61.  Poli Marchese E. Piante e Fiori dell’Etna. Palermo: Sellerio, 1991. Birdsall TC, Kelly GS. Berberine: therapeutic potential of an  Poli Marchese E. La vegetazione altomontana dell’Etna. Boll Ist alkaloid found in several medicinal plants. Altern Med Rev Bot Univ Catania 1965;s. 3(5):1 Á/6. 1997;2:94 Á/103.  Marjorie Murphy C. Plant products as antimicrobial agents. Clin Sethi ML. Enzyme inhibition VI: inhibition of reverse transcrip- Microbiol Rev 1999;12:564 Á/82. tase activity by protoberberine alkaloids and structure-activity  National Committee for Clinical Laboratory Standard (NCCLS); relationship. J Pharm Sci 1983;72(5):538 Á/41. Performance Standards for Antimicrobial Susceptibility Testing. Tsai TH. Analytical approaches for traditional Chinese medicines Eleventh Informational Supplement. Villanova, PA, 2001. exhibiting antineoplastic activity. J Chromatogr B Biomed Sci  National Committee for Clinical Laboratory Standards Appl 2001;764(1 Á/2):27 Á/48. (NCCLS); Reference Method for Broth Dilution Antifungal Cernakova M, Kost’alova D, Kettmann V, Plodova M, Toth J, Susceptibility Testing of Yeasts. Approved standard. Document Drimal J. Potential antimutagenic activity of berberine, a M-27A. Wayne, PA, 1997.