zhang & rock (2004) evaluation of epigallocatechin gallate and related plant

Yong-Mei Zhang and Charles O. Rock

Type II Fatty-acid Synthasethe FabG and FabI Reductases of BacterialRelated Plant Polyphenols as Inhibitors of Evaluation of Epigallocatechin Gallate andLipids and Lipoproteins:

doi: 10.1074/jbc.M403697200 originally published online May 7, 20042004, 279:30994-31001.J. Biol. Chem.

10.1074/jbc.M403697200Access the most updated version of this article at doi:

.JBC Affinity SitesFind articles, minireviews, Reflections and Classics on similar topics on the

Alerts:

When a correction for this article is posted When this article is cited

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/279/30/30994.full.html#ref-list-1This article cites 55 references, 20 of which can be accessed free at

by guest on Novem

ber 11, 2014http://w

ww

.jbc.org/D

ownloaded from

by guest on N

ovember 11, 2014

http://ww

w.jbc.org/

Dow

nloaded from

Evaluation of Epigallocatechin Gallate and Related PlantPolyphenols as Inhibitors of the FabG and FabI Reductasesof Bacterial Type II Fatty-acid Synthase*

Received for publication, April 2, 2004, and in revised form April 27, 2004Published, JBC Papers in Press, May 7, 2004, DOI 10.1074/jbc.M403697200

Yong-Mei Zhang and Charles O. Rock

From the Protein Science Division, Department of Infectious Diseases, St. Jude Childrens Research Hospital,Memphis, Tennessee 38105

Epigallocatechin gallate (EGCG) is the major compo-nent of green tea extracts and possesses antibacterial,antiviral, and antitumor activity. Our study focused onvalidating the inhibition of the bacterial type II fattyacid synthesis system as a mechanism for the antibacte-rial effects of EGCG and related plant polyphenols.EGCG and the related tea catechins potently inhibitedboth the FabG and FabI reductase steps in the fatty acidelongation cycle with IC50 values between 5 and 15 M.The presence of the galloyl moiety was essential foractivity, and EGCG was a competitive inhibitor of FabIand a mixed type inhibitor of FabG demonstrating thatEGCG interfered with cofactor binding in both enzymes.EGCG inhibited acetate incorporation into fatty acids invivo, although it was much less potent than thiolactomy-cin, a validated fatty acid synthesis inhibitor, and over-expression of FabG, FabI, or both did not confer resist-ance. A panel of other plant polyphenols was screenedfor FabG/FabI inhibition and antibacterial activity.Most of these inhibited both reductase steps, possessedantibacterial activity, and inhibited cellular fatty acidsynthesis. The ability of the plant secondary metabolitesto interfere with the activity of multiple NAD(P)-dependent cellular processes must be taken into ac-count when assessing the specificity of their effects.

Plants are renowned for containing compounds of medicinalinterest, and there is a continuing debate over the clinical valueand safety of herbal remedies (1). Botanical extracts include alarge variety of low molecular weight secondary metabolitesderived from isoprenoid, phenylpropanoid, or fatty acid/polyketide pathways. The rich diversity of these compounds isthought to arise from an evolutionary process driven by theacquisition of resistance to microbiological attack (2). Plantsand their natural enemies co-evolve (3), so it is also expectedthat bacterial defenses have arisen to combat these agents. Themajority of plant pathogens are Gram-negative bacteria, andconsistent with the co-evolutionary hypothesis, Gram-positivebacteria are generally more susceptible to the plant secondarymetabolites than Gram-negative bacteria (4, 5). For example,most Gram-negative bacteria are refractory to plant secondarymetabolites when tested in standard susceptibility tests pro-

ducing MIC1 values in the range of 0.11 mg/ml. A primaryunderlying cause for the resistance of Gram-negative bacteriato a wide range of plant toxins is the existence of efflux pumpsthat prevent the intracellular accumulation of polyphenols (4,6). Genetic elimination of these pumps can increase the efficacyof the plant metabolites by 12 orders of magnitude (6).

Green tea and the individual compounds purified from teaextracts are among the best known plant polyphenols andpossess numerous biological activities (including antimicrobialactivity) against a variety of organisms (7). The main compo-nents of a cup of green tea (200 ml) are the well characterizedcatechins consisting of ()-epigallocatechin gallate (EGCG)(140 mg), ()-epigallocatechin (65 mg), ()-epicatechin gallate(28 mg), and ()-epicatechin (17 mg) (8). Avid tea drinkersconsume several cups a day, and although the potency of thecatechins is low, the large quantities of the polyphenols thatare ingested make it reasonable to think that they have poten-tial for antibacterial activity. An example is the correlationbetween the in vivo and in vitro susceptibility of Helicobacterpylori to green tea (9, 10).

There is a huge amount of literature on the biological andbiochemical activities of green tea extracts and isolated com-pounds, but a unifying hypothesis that accounts for their bio-logical activities has not emerged, and the underlying biochem-ical targets remain unidentified (8, 11, 12). For example, EGCGand related polyphenols inhibit the fungal and mammaliantype I fatty-acid synthase system, most likely by interactingwith their reductase and perhaps condensation subdomains(1316). Indeed, the inhibition of fatty acid synthesis is con-sistent with many biological effects of EGCG, and it is temptingto conclude that fatty acid synthesis is an important target fortea catechins and plant secondary metabolites. Bacterial fatty-acid synthase consists of multiple individual enzymes, eachencoded by a separate gene (17, 18), in contrast to the mam-malian fatty-acid synthase, which is a homodimer of singlemultifunctional polypeptide derived from a single gene (19).The difference between the bacterial and mammalian syn-thases has been exploited to establish fatty acid synthesis as atarget for antibacterial drug discovery (20, 21). There are sev-eral naturally produced antibiotics, such as cerulenin (22),thiolactomycin (TLM) (23), and CT2108A (24), and syntheticmolecules (such as isoniazid (25) and triclosan (26, 27)), all ofwhich specifically target this pathway. Some of these, such ascerulenin (28), inhibit both the type I and type II fatty-acidsynthases, whereas others, such as thiolactomycin (23), are

* This work was supported by National Institutes of Health GrantGM34496, Cancer Center (CORE) Support Grant CA 21765, and theAmerican Lebanese Syrian Associated Charities. The costs of publica-tion of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: St. Jude ChildrensResearch Hospital, 332 N. Lauderdale, Memphis, TN 38105-2794. Tel.:901-495-3491; Fax: 901-495-3099; E-mail: [email protected].

1 The abbreviations used are: MIC, minimal inhibitory concentration;EGCG, ()-epigallocatechin gallate; TLM, thiolactomycin; ACP, acylcarrier protein; FabB, -ketoacyl-ACP synthase I; FabH, -ketoacyl-ACP synthase III; FabG, -ketoacyl-ACP reductase; FabI, trans-2-enoyl-ACP reductase; 3HC, 2,2,4-trihydroxychalcone.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 279, No. 30, Issue of July 23, pp. 3099431001, 2004 2004 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org30994

by guest on Novem


ww

.jbc.org/D

ownloaded from

selective for the bacterial type II system and are potentiallyuseful therapeutics (29).

The goals of this study were to determine whether the teacatechins and related plant polyphenols were inhibitors of bac-terial type II fatty-acid synthase and to determine whether thisinhibition was related to their antibacterial properties. We findthat EGCG is a potent inhibitor of both the -ketoacyl-ACPreductase (FabG) and the trans-2-enoyl-ACP reductase (FabI)components in the bacterial type II fatty-acid synthase system,a property that is common to a broad range of plant polyphe-nols. However, the inhibition of fatty acid synthesis was not thesole reason for the antibacterial activity of the tested com-pounds in the Escherichia coli model system.

EXPERIMENTAL PROCEDURES

MaterialsAmersham Biosciences supplied [1-14C]acetyl-CoA (60Ci/mol), [1-14C]acetate (54 Ci/mol), and [2-14C]malonyl-CoA (52Ci/mol); Sigma supplied acetoacetyl-CoA, malonyl-CoA, ACP,NADPH, NADH, and chloramphenicol. The trans-2-octanoyl-N-acetyl-cysteamine was the generous gift of Rocco Gogliotti and John Domagala(Parke-Davis). His-tagged FabB, FabH, FabG, and FabI from E. coliwere purified as described previously (30, 31). All other reagents wereof the highest grade available.

Green tea extract compounds, including EGCG, ()-epigallocatechin,()-epicatechin gallate, ()-epicatechin, ()-gallocatechin gallate, ()-gallocatechin, ()-catechin gallate, ()-catechin, propyl gallate, gallicacid, butein, isoliquirtigenin, resveratrol, piceatannol, fisetin, querce-tin, rhein, and plumbagin were purchased from Sigma. Fustin, taxifo-lin, 2,2,4-trihydroxychalcone, and 7,3,4-trihydroxyisoflavone werepurchased from Indofine Chemicals Co. The compounds were dissolvedin Me2SO at 10 mM.

Bacterial StrainsStrain ANS1 (metB1 relA1 spoT1 gyrA216tolC::Tn10 R F) cultured in Tryptone broth (1% Tryptone) wasused to determine the minimal inhibitory concentrations (32). Chemi-cally defined M9 medium (33) was used in the [14C]-acetate labelingexperiments described below. ANS1 was constructed by P1-mediatedtransduction of the tolC::Tn10 element in strain EP1581 into strainUB1005 followed by selection for tetracycline resistance. The TolC-de-pendent type I secretion and multidrug efflux systems are defective inASN1 (32). Plasmid expressing fabG or fabI was constructed by movingthe XbaI-BamHI fragment of the His-tagged version of either the fabGor fabI gene (34), respectively, in pET15-b into pBluescript II KS(). Toselect for the presence of both reductases, the pBluescript plasmidcarrying the enoyl-ACP reductase gene from Clostridium acetobutyli-cum was modified to replace the AmpR gene with and KanR cassette(35). The His-tagged versions were used to confirm expression using ananti-tag antibody.-Ketoacyl-ACP Reductase (FabG) AssayThe disappearance of

NADPH, the cofactor for FabG reaction, was measured spectrophoto-metrically at 340 nm as described previously (36). The reaction mixturecontained 0.5 mM acetoacetyl-CoA, 0.2 mM NADPH, 4 g of FabGprotein, 0.1 M sodium phosphate buffer, pH 7.4, in a final volume of 300l. Test compounds were added to the reaction mixture at the concen-trations indicated in the text and figure legends. The reaction wasinitiated by the addition of acetoacetyl-CoA. A decrease in the absorb-ance at 340 nm was recorded for 2 min. The initial rate was used tocalculate the enzymatic activity. Reactions with Me2SO solvent alonewere used as controls.

Enoyl-ACP Reductase (FabI) AssayReduction of the trans-2-octa-noyl-N-acetylcysteamine substrate analog was measured spectrophoto-metrically by following the utilization of NADH at 340 nm at 25 C asdescribed previously (30). Briefly, a standard reaction contained 0.1 Msodium phosphate, pH 7.5, 100 M trans-2-octanoyl-N-acetylcysteam-ine, 200 M NADH, and 12 g of FabI in a final volume of 300 l. Adecrease in absorbance at 340 nm was measured at 25 C for the linearperiod of the assay (usually the first 12 min). Test compounds wereadded in the assay to the indicated concentrations as described in thetext and figure legends. An equal volume of the solvent was used for theuntreated control.-Ketoacyl-ACP Synthase I (FabB) and III (FabH) AssayThe cou-

pled assay of FabH as described previously (34) contained 25 M ACP,1 mM -mercaptoethanol, 65 M malonyl-CoA, 45 M [1-14C]acetyl-CoA(specific activity, 60 Ci/mol), 1 g of purified FabD, 0.1 M sodiumphosphate buffer, pH 7.0, and 20 ng of FabH protein in a final volumeof 40 l. The FabD protein was present to generate the malonyl-ACPsubstrate for the reaction. The ACP, -mercaptoethanol, and buffer

were preincubated at 37 C for 30 min to ensure the complete reductionof ACP. The reaction was initiated by the addition of FabH. Afterincubation at 37 C for 15 min, 35 l of the reaction mixture wasremoved and dispensed onto a paper filter disc (Whatman No. 3MMfilter paper). The disc was washed successively with ice-cold 10, 5, and1% trichloroacetic acid with 20 min for each wash and 20 ml of washsolution per disc. The filter discs were dried and counted for 14C isotopein 3 ml of scintillation fluid.

The condensation assay for FabB was the same as described previ-ously (37). Briefly, FabB assays contained 45 M myristoyl-ACP, 50 M[2-14C]malonyl-CoA (specific activity, 52 Ci/mol), 100 M ACP, 1 gof FabD, and 25 ng of FabB in a final volume of 20 l. ACP was reducedby 0.3 mM dithiothreitol before the other reaction components wereadded. The reaction was initiated by the addition of the enzyme. Afterincubation at 37 C for 15 min, the reaction was stopped by adding 0.4ml of the reducing reagent containing 0.1 M K2HPO4, 0.4 M KCl, 30%tetrahydrofuran, and 5 mg/ml NaBH4. The reaction contents were vig-orously mixed after the addition of the reducing reagent and incubatedat 37 C for 40 min before being extracted into 0.4 ml of toluene. The 14Cisotope in the upper phase was quantitated by scintillation counting.IC50 values were determined at a series of concentrations. A line wasdrawn between the points, and the IC50 was the interpolated concen-tration that gave 50% inhibition (see Fig. 1).

Determination of the MICThe MICs of the test compounds againstE. coli strain ANS1 were determined by a broth microdilution method.ANS1 was grown to midlog phase in 1% Tryptone broth and thendiluted 30,000-fold in the same medium. A 10-l aliquot of the dilutedcell suspension (3,0005,000 colony-forming units) was used to inocu-late each well of a 96-well plate (U-bottom with a low evaporation lid)containing 100 l of Tryptone broth with the indicated concentration ofinhibitors. The plate was incubated at 37 C for 20 h before being readwith a FusionTM universal microplate analyzer (Packard) at 600 nm.The absorbance was normalized to the solvent-treated control, whichwas considered to be 100%.

[1-14C]Acetate LabelingStrain ANS1 was grown to midlog phase inM9 minimal medium supplemented with 0.1% casamino acids, 0.4%glycerol, and 0.0005% thiamin. A 1-ml aliquot of cells was treated withthe antimicrobial compounds for 10 min at the indicated concentrationsas described in the figure legends. An equal volume of the solventMe2SO was added to the untreated control. Cells were then labeled with10 Ci of [1-14C]acetate for 1 h before being harvested by centrifugation.The cell pellets were washed with phosphate-buffered saline and resus-pended in 100 l of M9 medium. The total cellular lipids were extracted,and the incorporated 14C isotope in lipids in the chloroform phase wasquantitated by scintillation counting. Results reflect duplicateexperiments.

RESULTS

Inhibition of FabG and FabI by Tea CatechinsThe enzymesof the type II fatty-acid synthase were assayed for inhibition byEGCG, the major catechin of green tea (Fig. 1A). The elonga-tion condensing enzyme (FabB) was refractory to EGCG inhi-bition, whereas FabH, the initiating condensing enzyme, wasinhibited by EGCG with an IC50 of 40 M. The most potentlyinhibited enzyme was FabG, the NADPH-dependent ketore-ductase in the pathway, which exhibited an IC50 of 5 M. FabI,the NADH-dependent enoyl reductase, was inhibited by EGCGwith an IC50 of 15 M. FabG was used as a model enzyme todetermine the inhibitory potency of the other significant greentea catechins (Fig. 1B). EGCG was the most potent, but theother catechins, ()-epicatechin gallate, ()-gallocatechin gal-late, and ()-catechin gallate, all exhibited activity againstFabG with IC50 values ranging from 5 to 15 M. Similarly, thefour compounds exhibited activity against FabI with IC50 val-ues between 5 and 15 M (Table I). Also, we found that thegallate substitution was critical for the inhibitory activity ofthe compounds (Table I). Removal of the galloyl moiety fromany of the tea catechins resulted in complete loss of inhibitoryactivity in vitro. The galloyl group itself has no significantinhibitory activity as shown by the lack of FabG and FabIinhibition by propyl gallate and gallic acid (Table I). These dataestablish that green tea catechins have significant inhibitoryactivity against the reductase enzymes of the bacterial type II

Plant Polyphenol Antibacterial Compounds 30995

by guest on Novem


ww

.jbc.org/D

ownloaded from

fatty-acid synthase. Furthermore, the structure-activity rela-tionship shows that the portion of the EGCG molecule requiredfor inhibition is defined by the boxed area in Scheme 1.

Mechanism for Tea Catechin Inhibition of FabG and FabIThe kinetic mechanism for the inhibition of FabG and FabI wasdetermined using EGCG as the model compound (Fig. 2). BothFabG (38) and FabI (39) have compulsory ordered mechanismswith the nucleotide cofactors as the leading substrates. Thisknowledge allowed us to design a kinetic analysis that woulddistinguish between the three possible outcomes; EGCG couldbind to the free enzyme, the enzyme-substrate complex, or bothto prevent catalysis. In the first case, the inhibition patternwith respect to the cofactor would be competitive; in the second,the inhibition pattern would be non-competitive; and in thethird case, mixed-type inhibition would be observed (40). Theinhibition of FabG by EGCG was mixed with respect to NADPH(Fig. 2A). Thus, EGCG binds to both the free enzyme to preventthe binding of the nucleotide cofactor and also to the FabG-

NADPH complex to prevent the binding of the substrate. Incontrast, EGCG was a competitive inhibitor of FabI with re-spect to NADH (Fig. 2B), meaning that EGCG interferes withactivity by binding to the free enzyme and preventing thebinding of NADH. These data illustrate that EGCG inhibitsboth enzymes by association with the nucleotide cofactor bind-ing site, and with FabG, EGCG has the additional property ofbinding to the enzyme-cofactor complex.

Antimicrobial Effect of EGCGAs reported previously (5, 7),EGCG has moderate antibacterial activity. Most Gram-nega-tive bacteria are resistant to plant polyphenols; therefore weused our E. coli strain ANS1 (tolC) to eliminate the activity ofa major class of multidrug efflux pumps (6, 32). Such tolCmutants are used routinely as a platform to investigate themechanism of drug action (41). In strain ANS1, EGCG had anMIC of 500 M (Fig. 3A), although the parent wild-type strainUB1005 had the same MIC in this case. To evaluate whetherthe antibacterial properties of the catechins could be attributedto their effects on fatty acid synthesis, first we determinedwhether the compounds blocked the incorporation of acetateinto membrane fatty acids (Fig. 3B). EGCG indeed attenuatedfatty acid synthesis in vivo compared with the untreated con-trol cells and cells treated with chloramphenicol, a proteinsynthesis inhibitor. However, the effect of EGCG was notnearly as great as the effect of TLM, an established inhibitor offatty acid synthesis at the elongation condensing enzyme step(42, 43). These data demonstrate that EGCG inhibits fatty acidsynthesis in vivo, although it is not as potent when present atdouble its MIC compared with a bona fide inhibitor added tothe cells at the same relative concentration, suggesting thatfatty acid synthesis may not be the sole pathway inhibited byEGCG.

Overexpression of individual genes and the isolation of re-sistant mutants are powerful genetic tools for target validationin vivo. For example, FabB was unequivocally established asthe critical in vivo target for TLM by demonstrating that over-expression of FabB increased the resistance to TLM and that apoint mutation conferring resistance to the drug was localizedto the fabB gene (32, 42). Similarly, FabI was validated as thetriclosan target through analyzing the effects of FabI overex-pression and the isolation of resistant mutants in the fabI gene(26, 27). Unfortunately, there are no known specific inhibitorsagainst FabG to corroborate the function of our FabG expres-sion plasmid, but the construct increases FabG expression, andlike the FabB and FabI constructs with identical promoterelements is anticipated to shift the dose-response curve for anantibacterial compound that selectively targets FabG. There-fore, we examined the effect of the overexpression of FabG,FabI, or both on the MIC for EGCG in strain ANS1 (Fig. 3A).Unlike other drugs that primarily target lipid synthesis, noneof these plasmids increased the resistance of strain ANS1 toEGCG. We also attempted to raise EGCG-resistant mutantsusing the techniques described previously to isolate mutantsspecifically resistant to TLM (32) or triclosan (27). However, wewere unable to obtain colonies resistant to 1 mM EGCG. Thesedata do not support fatty acid synthesis as a primary targetfor the action of EGCG, and the inability to isolate specificmutants is consistent with the existence of multiple targetsin vivo.

FabG/FabI Inhibition by Other Plant Natural ProductsTodetermine whether the inhibition of FabG and FabI was aproperty of other natural products with the biphenyl core struc-ture diagramed in Scheme 1, we tested a panel of plant poly-phenols (Table II). The MIC values were obtained using strainANS1 to eliminate the effects of efflux pumps, and the MICs fora wild-type and tolC isogenic pair were reported previously (6).

FIG. 1. Inhibitory effects of EGCG and its analogs on fatty acidbiosynthetic enzymes. A, EGCG inhibited the enzymatic activities ofE. coli FabG (E), FabH (), and FabI () with apparent IC50 values of5, 15, and 40 M, respectively. FabB (f) was refractory to EGCG.Enzyme treated with an equal volume of the solvent (Me2SO) was takenas the 100% active control. The specific activity of FabB was 0.36 nmolof [14C]malonyl-CoA incorporated per min per g of protein, FabG was2.6 nmol of NADPH oxidized per min per g, FabH was 0.77 nmol of[14C]acetyl-CoA incorporated per min per g, and FabI was 0.33 nmol ofNADH oxidized per min per g. B, EGCG (E) and its analogs, ()-epicatechin gallate (), ()-catechin gallate (), and ()-gallocatechingallate (), potently inhibited E. coli FabG with apparent IC50 valuesbetween 5 and 15 M. The assay conditions are described under Ex-perimental Procedures. Error bars show standard deviations.

Plant Polyphenol Antibacterial Compounds30996

by guest on Novem


ww

.jbc.org/D

ownloaded from

Some of the natural products (such as coumestrol, rhein, andplumbagin) had potent antibacterial activity but were not sig-nificant inhibitors of either the FabG or FabI reductases. How-ever, in general all of the tested natural products with thebiphenyl chalcone nucleus inhibited both enzymes. Resvera-trol, piceatannol, fustin, taxifolin, and 7,3,4-trihydroxyiso-flavone were good inhibitors for FabG and FabI, but their MICvalues were high, indicating only weak activity against anytarget in vivo.

The remaining polyphenols with comparable reasonable MIC

values (Table II) were further evaluated for their effects onfatty acid synthesis at 4 times their MICs. The inhibition of[14C]acetate incorporation by butein, 3HC, fisetin, and querce-tin was between 20 and 50% in comparison to cells treated withMe2SO (Table II). The highest inhibition was observed withisoliquirtigenin, which reduced the [14C]acetate incorporationby about 75%. The FabG/FabI inhibitor with the lowest MICagainst strain ANS1, 3HC, was selected for more detailed in-vestigation to determine whether its antibacterial action couldbe linked to blocking fatty acid synthesis (Fig. 4). FabH was not

TABLE IInhibitory effects of green extract compounds on the activity of fatty acid synthetic enzymes (FabG and FabI) and bacterial growth

a IC50 and MIC values greater than 100 and 200 M, respectively, were not determined. The IC50 values were rounded to the nearest 5 M.

SCHEME 1. The portion of the EGCGmolecule that is required for its in-hibitory activity against FabG andFabI as determined from the data inFig. 1 and Table I is enclosed by thebox. This structure is rotated 90 clock-wise and stripped of the irrelevant sub-stituents to reveal a putative active nu-cleus composed of two polyhydroxyphenolrings joined by an ester linkage to gallicacid.


by guest on Novem


ww

.jbc.org/D

ownloaded from

inhibited by 3HC. FabB, FabG, and FabI were all inhibited by3HC with IC50 values of 100, 25, and 40 M, respectively (Fig.4A). Thus, 3HC was less potent than EGCG in vitro. The MICfor 3HC was 6.25 M, and the MIC was not shifted by theoverexpression of either FabG or FabI (Fig. 4B). Finally, ace-tate incorporation studies showed that fatty acid synthesis wasinhibited only at concentrations significantly higher than theMIC for the compound (Fig. 4C). Although like most of theother plant polyphenols, 3HC inhibited the reductase steps infatty acid synthesis in vitro, these data establish that the mosteffective antibacterial examined in the screen did not primarilyinhibit cell growth by blocking the type II fatty acid synthesis.Likewise, we examined the inhibition of fatty acid synthesis invivo using four other analogs with MICs below 75 M (Table II).Although all of the compounds inhibited acetate incorporationinto cellular fatty acids, none were as potent as the TLMcontrol. Thus, these polyphenols reduced fatty acid synthesis,but the extent of inhibition was not consistent with fatty acidsynthesis as the primary target for their antibacterial activity.

DISCUSSION

The emergence of multidrug resistance in pathogenic bacte-ria is a global problem that calls for the development of newantibiotics with unique cellular targets (44). Consequently, thedifferences between bacterial and mammalian fatty acid bio-synthesis are being exploited to develop the type II fatty-acidsynthase as a target for novel drug discovery (21, 45, 46). FabGis ubiquitously expressed in all bacteria, is highly conservedacross species, and is the only known isozyme that catalyzesthe essential keto reduction step in the elongation cycle (47).Although there are no known FabG inhibitors (21, 45, 46),FabG represents an ideal focus for the development of newantibiotics based on the hypothesis that FabG would be vul-nerable to inhibitors that interact with its cofactor binding site(48). Our study identifies a small molecule scaffold with potentinhibitory effects against FabG as well as FabI, an establishedtarget in bacterial fatty acid synthesis. Green tea catechins(EGCG) and related plant polyphenolic compounds are selec-

FIG. 2. The mechanism of inhibition of EGCG on E. coli FabGand FabI. A, EGCG is a mixed type inhibitor of FabG with respect toNADPH. The double-reciprocal plot of 1/v versus 1/[NADPH] at differ-ent concentrations of EGCG (6 M (f), 4 M (), and 0 M (E)) inter-cepted to the left of the 1/v axis and above the 1/[NADPH] axis, indi-cating that EGCG is a mixed type inhibitor for NADPH. B, EGCG is acompetitive inhibitor of FabI with respect to NADH. The double-recip-rocal plot of 1/v versus 1/[NADH] at different concentrations of EGCG(30 M (f), 20 M (), 10 M (), and 0 M (E)) intercepted on the 1/vaxis, indicating that EGCG is a competitive inhibitor for NADH. Theassays were performed using the conditions described under Experi-mental Procedures. Error bars show standard deviations.

FIG. 3. Inhibitory effects of EGCG on the growth and fatty acidbiosynthesis of E. coli. A, EGCG inhibited the growth of E. coli strainANS1 with an MIC at 500 M (E). The addition of constructs expressingfabG (), fabI (), or both () did not rescue the cells from EGCG. Thegrowth of cultures treated with Me2SO was considered to be 100%. B,the effects of the antimicrobial compounds on fatty acid biosynthesiswere tested by monitoring the [14C]-acetate incorporation into lipids at1 time (white bars) and 2 times (gray bars) MIC concentrations of thedrugs. When the cells were treated with EGCG at 2 times the MICconcentration (1 mM), [14C]-acetate incorporation was inhibited by 50%,suggesting that fatty acid biosynthesis was compromised in the pres-ence of EGCG. When an MIC concentration of 500 M EGCG was used,25% inhibition was observed. TLM, a known inhibitor of fatty acidbiosynthesis, exhibited a more profound inhibitory effect on [14C]-ace-tate incorporation at 5 and 10 M. No inhibition was obtained withchloramphenicol, a known inhibitor of protein synthesis, at 5 and10 M.


by guest on Novem


ww

.jbc.org/D

ownloaded from

tive inhibitors of the FabG and FabI reductase steps of bacte-rial type II fatty acid synthesis. The galloyl moiety of thecatechins is absolutely essential for inhibition. Two hydroxy-phenyl rings connected by a 23-carbon linker is the basicrequirement for inhibition as illustrated in Scheme 1. Aspectsof this core structure are also found in the hydroxydiphenylether class of potent FabI inhibitors (27), except in this case thetwo hydroxyphenyl rings are connected by a single bridgingoxygen. The kinetic analysis indicates that EGCG binds to theligand-free form of both FabG and FabI and blocks NAD(P)Hbinding. This result does not necessarily mean that EGCGoccupies the same pocket as NAD(P)H, so to advance the designof inhibitors based on this scaffold it is important to determinethe structure of the EGCG-enzyme binary complex. This struc-ture is particularly significant in light of the high degree ofconformational flexibility in FabG (38). The tight binding oftriclosan to FabI is caused by its ability to lock the reductase ina closed confirmation (31); identifying a slow binding polyphe-nol derivative with a high affinity interaction and a specificFabG conformation has the most promise as an antibacteriallead.

The green tea catechins also inhibit the reductase and con-densation reactions of the polyfunctional mammalian type Ifatty-acid synthase (1315). Thus, the same enzymatic steps

are targeted by EGCG in the mammalian enzyme as in thebacterial system, and it is tempting to speculate that the anti-cancer, proapoptotic, and hypolipidemic activities associatedwith green tea extracts and EGCG arise from the inhibition offatty acid synthesis (1315, 4951). However, these data mustbe considered correlative in the absence of a genetic validationof fatty-acid synthase as the principal cellular target for thecatechins in animal cells. These experiments are critical to theinterpretation of the biochemical data in light of the potentiallyhigh number and diversity of EGCG biochemical targets (suchas matrix metalloproteinases (52) and the laminin receptor(53)) that may also contribute to its anticancer activity.

Treatment of bacterial cells with EGCG results in the inhi-bition of fatty acid production, implicating the bacterial type IIfatty-acid synthase as a target for their antibacterial activity.However, EGCG was not as effective at blocking acetate incor-poration as established fatty-acid synthase inhibitors whencompared at 2 times their MICs. Also, the overexpression of oneor both of the reductase targets (FabG or FabI) did not conferincreased resistance to EGCG. Genetic target validation, suchas the acquired resistance to TLM or triclosan by either over-expression or mutations in the target genes fabB (32) or fabI(27), is essential to the confirmation of fatty acid synthesis as atarget for EGCG. Our experiments failed to forge a definitive

TABLE IIInhibitory effects of polyphenol compounds on the activity of fatty acid synthetic enzymes (FabG and FabI) and bacterial growth

a IC50 and MIC values greater than 100 and 200 M, respectively, were not determined. The IC50 values were rounded to the nearest 5 M.b The amount of [14C] acetate incorporated into the cells treated with Me2SO was 100%. The compounds were tested at 4 their MICs.c NT, not tested.


by guest on Novem


ww

.jbc.org/D

ownloaded from

link between the FabG/FabI reductases and the antibacterialeffect of EGCG, arguing against fatty acid synthesis as the soleantibacterial target for EGCG in vivo. We reached the same

conclusions with 3HC, a chalcone secondary metabolite with alow MIC against strain ANS1 and inhibitory activity againstFabG and FabI. In a similar study, inhibitors of fungal fattyacid synthesis were identified from plant extracts (16); how-ever, a correlation between the fatty acid inhibitory activityand the antifungal action of the compounds could not beestablished.

Understanding the potential benefits and risks of tea drink-ing is complicated not only from the great variety of secondarymetabolites present (8) but also from our data demonstratingthat a single compound from tea has multiple targets. Thisfinding opens the potential for diverse mechanisms of actiondepending on the relative importance of the targets in thedifferent experimental systems under investigation. BothFabG and FabI are members of the short chain dehydrogenase/reductase superfamily (54, 55). This family is composed of alarge cohort of proteins that bind nicotinamide nucleotide co-factors using a similar protein fold and catalyze a long list ofessential reactions in intermediary metabolism (56, 57). Ourkinetic results indicate that EGCG interferes with the bindingof the nicotinamide adenine dinucleotide cofactor in both en-zymes, suggesting that the EGCG scaffold recognizes a com-mon feature in their NAD(P) binding domains. This conceptleads to the hypothesis that EGCG and related polyphenolstructures may have inhibitory activity against other enzymesin the short chain dehydrogenase/reductase superfamily. Theeffects may even be broader if they extend to the nucleotidebinding sites in other protein families as well. The ability ofEGCG to inhibit FabH and not FabB may be understood basedon this hypothesis because FabH uses acetyl-CoA (an adeninenucleotide) as a substrate whereas FabB does not. Not only areNAD(P) cofactors essential in intermediary metabolism, butthe importance of their signaling functions in controlling cellphysiology is becoming increasingly apparent (58). For exam-ple, many of the same flavonoids that inhibit FabG/FabI (TableII) also interact with Sir2, an NAD-binding protein that influ-ences lifespan in yeast (59). These data point to the ability ofthe plant polyphenols to bind to a variety of enzymes that useNAD(P), and more work needs to be done to provide informa-tion on their multiple effects on the cell biology of differentsystems and the multiple targets presented by the diversity ofenzymes that utilize NAD(P). Also, minor changes in the hy-droxylation pattern of two rings and the distance and compo-sition of the linkers that connect them are likely to have amajor impact on the spectrum of enzymes targeted by thepolyphenols. Considering the large number of potential proteintargets for these compounds, this appears to be a dauntingtask, but nonetheless plant polyphenol secondary metabolitesmay eventually prove to be suitable chemical scaffolds for thefuture development of selective inhibitors of fatty acid synthe-sis in bacteria and other systems.

AcknowledgmentsWe thank Matt Frank and Amy Sullivan fortheir expert technical assistance.

REFERENCES

1. De Smet, P. A. (2002) N. Engl. J. Med. 347, 204620562. Dixon, R. A. (2001) Nature 411, 8438473. Rausher, M. D. (2001) Nature 411, 8578644. Lewis, K. (2001) J. Mol. Microbiol. Biotechnol. 3, 2472545. Yam, T. S., Shah, S., and Hamilton-Miller, J. M. (1997) FEMS Microbiol. Lett.

152, 1691746. Tegos, G., Stermitz, F. R., Lomovskaya, O., and Lewis, K. (2002) Antimicrob.

Agents Chemother. 46, 313331417. Hamilton-Miller, J. M. (1995) Antimicrob. Agents Chemother. 39, 237523778. Yang, C. S., and Wang, Z. Y. (1993) J. Natl. Cancer Inst. 85, 103810499. Yee, Y. K., and Koo, M. W. (2000) Aliment. Pharmacol. Ther. 14, 635638

10. Mabe, K., Yamada, M., Oguni, I., and Takahashi, T. (1999) Antimicrobiol.Agents Chemother. 43, 17881791

11. Adhami, V. M., Ahmad, N., and Mukhtar, H. (2003) J. Nutr. 133, Suppl. 7,2417S2424S

12. Higdon, J. V., and Frei, B. (2003) Crit. Rev. Food Sci. Nutr. 43, 89143

FIG. 4. Biochemical and biological effects of 2,2,4-trihy-droxychalcone. A, 2,2,4-trihydroxychalcone inhibited E. coli FabB(E), FabG (), and FabI () activity with apparent IC50 values of 100,25, and 40 M, respectively. FabH (f) was refractory to this compound.B, 2,2,4-trihydroxychalcone inhibited E. coli growth with an MIC of6.25 M. Overexpression of fabG () or fabI () did not make the cellsmore resistant to the drug compared with control strain ANS1 (E). C,[14C]-acetate incorporation was inhibited when cells were treated withincreasing concentrations of 2,2,4-trihydroxychalcone. TLM (a knownfatty acid biosynthesis inhibitor) and chloramphenicol (a known proteinsynthesis inhibitor) were used as positive and negative controls. Theassays were performed using the conditions described under Experi-mental Procedures. Error bars show standard deviations.


by guest on Novem


ww

.jbc.org/D

ownloaded from

13. Wang, X., and Tian, W. (2001) Biochem. Biophys. Res. Commun. 288,12001206

14. Wang, X., Song, K. S., Guo, Q. X., and Tian, W. X. (2003) Biochem. Pharmacol.66, 20392047

15. Li, B. H., and Tian, W. X. (2004) J. Biochem. (Tokyo) 135, 859116. Li, X. C., Joshi, A. S., el-Sohly, H. N., Khan, S. I., Jacob, M. R., Zhang, Z., Khan,

I. A., Ferreira, D., Walker, L. A., Broedel, S. E., Jr., Raulli, R. E., and Cihlar,R. L. (2002) J. Nat. Prod. 65, 19091914

17. Cronan, J. E., Jr., and Rock, C. O. (1996) in Escherichia coli and Salmonellatyphimurium: Cellular and Molecular Biology (Neidhardt, F. C., Curtis, R.,Gross, C. A., Ingraham, J. L., Lin, E. C. C., Low, K. B., Magasanik, B.,Reznikoff, W., Riley, M., Schaechter, M., and Umbarger, H. E., eds) pp.612636, American Society for Microbiology, Washington, D. C.

18. Rock, C. O., and Jackowski, S. (2002) Biochem. Biophys. Res. Commun. 292,11551166

19. Smith, S., Witkowski, A., and Joshi, A. K. (2003) Prog. Lipid Res. 42, 28931720. Heath, R. J., White, S. W., and Rock, C. O. (2001) Prog. Lipid Res. 40, 46749721. Campbell, J. W., and Cronan, J. E., Jr. (2001) Annu. Rev. Microbiol. 55,

30533222. Vance, D. E., Goldberg, I., Mitsuhashi, O., Bloch, K., Omura, S., and Nomura,

S. (1972) Biochem. Biophys. Res. Commun. 48, 64965623. Hayashi, T., Yamamoto, O., Sasaki, H., and Okazaki, H. (1984) J. Antibiot.

(Tokyo) 37, 1456146124. Laakso, J. A., Raulli, R., McElhaney-Feser, G. E., Actor, P., Underiner, T. L.,

Hotovec, B. J., Mocek, U., Cihlar, R. L., and Broedel, S. E., Jr. (2003) J. Nat.Prod. 66, 10411046

25. Banerjee, A., Dubnau, E., Quemard, A., Balasubramanian, V., Um, K. S.,Wilson, T., Collins, D., de Lisle, G., and Jacobs, W. R., Jr. (1994) Science263, 227230

26. McMurray, L. M., Oethinger, M., and Levy, S. (1998) Nature 394, 53153227. Heath, R. J., Yu, Y.-T., Shapiro, M. A., Olson, E., and Rock, C. O. (1998) J. Biol.

Chem. 273, 303163032128. Omura, S. (1976) Microbiol. Rev. 40, 68169729. Miyakawa, S., Suzuki, K., Noto, T., Harada, Y., and Okazaki, H. (1982) J.

Antibiot. (Tokyo) 35, 41141930. Heath, R. J., and Rock, C. O. (1995) J. Biol. Chem. 270, 265382654231. Heath, R. J., Rubin, J. R., Holland, D. R., Zhang, E., Snow, M. E., and Rock,

C. O. (1999) J. Biol. Chem. 274, 111101111432. Jackowski, S., Zhang, Y.-M., Price, A. C., White, S. W., and Rock, C. O. (2002)

Antimicrob. Agents Chemother. 46, 1246125233. Miller, J. H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor

Laboratory, Cold Spring Harbor, New York34. Heath, R. J., and Rock, C. O. (1996) J. Biol. Chem. 271, 109961100035. Marrakchi, H., DeWolf, W. E., Jr., Quinn, C., West, J., Polizzi, B. J., So, C. Y.,

Holmes, D. J., Reed, S. L., Heath, R. J., Payne, D. J., Rock, C. O., and Wallis,N. G. (2003) Biochem. J. 370, 10551062

36. Zhang, Y.-M., Wu, B., Zheng, J., and Rock, C. O. (2003) J. Biol. Chem. 278,5293552943

37. Garwin, J. L., Klages, A. L., and Cronan, J. E., Jr. (1980) J. Biol. Chem. 255,1194911956

38. Price, A. C., Zhang, Y.-M., Rock, C. O., and White, S. W. (2004) Structure 12,417428

39. Sivaraman, S., Zwahlen, J., Bell, A. F., Hedstrom, L., and Tonge, P. J. (2003)Biochemistry 42, 44064413

40. Cornish-Bowden, A. (1995) Fundamentals of Enzyme Kinetics, pp. 95101,Portland Press Ltd., London

41. Sulavik, M. C., Houseweart, C., Cramer, C., Jiwani, N., Murgolo, N., Greene,J., DiDomenico, B., Shaw, K. J., Miller, G. H., Hare, R., and Shimer, G.(2001) Antimicrob. Agents Chemother. 45, 11261136

42. Tsay, J.-T., Rock, C. O., and Jackowski, S. (1992) J. Bacteriol. 174, 50851343. Price, A. C., Choi, K. H., Heath, R. J., Li, Z., Rock, C. O., and White, S. W.

(2001) J. Biol. Chem. 276, 6551655944. Verhoef, J. (2003) Adv. Exp. Med. Biol. 531, 30131345. Heath, R. J., and Rock, C. O. (2004) Curr. Opin. Investig. Drugs 5, 14615346. Heath, R. J., White, S. W., and Rock, C. O. (2001) Appl. Microbiol. Biotechnol.

58, 69570347. Lai, C. Y., and Cronan, J. E. (2004) J. Bacteriol. 186, 1869187848. Wright, H. T. (2004) Structure 12, 35835949. Yeh, C. W., Chen, W. J., Chiang, C. T., Lin-Shiau, S. Y., and Lin, J. K. (2003)

Pharmacogenomics J. 3, 26727650. Brusselmans, K., De Schrijver, E., Heyns, W., Verhoeven, G., and Swinnen,

J. V. (2003) Int. J. Cancer 106, 85686251. Tian, W. X., Li, L. C., Wu, X. D., and Chen, C. C. (2004) Life Sci. 74, 2389239952. Garbisa, S., Biggin, S., Cavallarin, N., Sartor, L., Benelli, R., and Albini, A.

(1999) Nat. Med. 5, 121653. Tachibana, H., Koga, K., Fujimura, Y., and Yamada, K. (2004) Nat. Struct.

Mol. Biol. 11, 38038154. Rafferty, J. B., Simon, J. W., Baldock, C., Artymiuk, P. J., Baker, P. J., Stuitje,

A. R., Slabas, A. R., Rice, D. W., and Aarnoudse, J. G. (1995) Structure 3,927938

55. Bergler, H., Wallner, P., Ebeling, A., Leitinger, B., Fuchsbichler, S., Aschauer,H., Kollenz, G., Hogenauer, G., and Turnowsky, F. (1994) J. Biol. Chem.269, 54935496

56. Jornvall, H., Hoog, J. O., and Persson, B. (1999) FEBS Lett. 445, 26126457. Oppermann, U., Filling, C., Hult, M., Shafqat, N., Wu, X., Lindh, M., Shafqat,

J., Nordling, E., Kallberg, Y., Persson, B., and Jornvall, H. (2003) Chem.Biol. Interact. 143144, 247253

58. Berger, F., Ramirez-Hernandez, M. H., and Ziegler, M. (2004) Trends Biochem.Sci. 29, 111118

59. Howitz, K. T., Bitterman, K. J., Cohen, H. Y., Lamming, D. W., Lavu, S., Wood,J. G., Zipkin, R. E., Chung, P., Kisielewski, A., Zhang, L. L., Scherer, B., andSinclair, D. A. (2003) Nature 425, 191196


by guest on Novem


ww

.jbc.org/D

ownloaded from

zhang & rock (2004) evaluation of epigallocatechin gallate and related plant

Documents