Toxicity of Ochratoxin A, Its Opened Lactone Form and Several of Its Analogs: Structure–Activity Relationships

TOXICOLOGY AND APPLIED PHARMACOLOGY ARTICLE NO.

137, 182–192 (1996)

0071

Toxicity of Ochratoxin A, Its Opened Lactone Form and Several of Its Analogs: Structure–Activity Relationships HAO XIAO, SRINIVASA MADHYASTHA, RONALD R. MARQUARDT,1 SUZHEN LI, JAYANTHA K. VODELA,* A. A. FROHLICH, AND BARBARA W. KEMPPAINEN* Department of Animal Science, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2; and *Department of Physiology and Pharmacology, College of Veterinary Medicine, Auburn University, Auburn, Alabama 36849 Received March 13, 1995; accepted November 20, 1995

Toxicity of Ochratoxin A, Its Open Lactone Form and Several of Its Analogs: Structure–Activity Relationships. XIAO, H., MADHYASTHA, S., MARQUARDT, R. R., LI, S., VODELA, J. K., FROHLICH, A. A., AND KEMPPAINEN, B. W. (1996). Toxicol. Appl. Pharmacol. 137, 182–192. Ochratoxin A (OA); its three natural analogs, ochratoxin C (OC), B (OB), and a (Oa); and its six synthetic analogs, the epimere of OA (d-OA), the ethylamide of OA (OE-OA), decarboxylated OA (DC-OA), O-methylated OA (OM-OA), lactone-opened OA (OP-OA), and the methyl ester of Oa (M-Oa) were assayed for their toxicities in prokaryotic (Bacillus brevis) and eukaryotic (HeLa cell) systems and in animals (mouse and rat). The LC50s (mM) for HeLa cells, were 0.005 (OA), 0.009 (OC), 0.163 (d-OA), 10.1 (OE-OA), 7.6 (DC-OA), 0.83 (OM-OA), 0.054 (OB), and 0.56 (Oa). The minimum inhibitory doses (nmol/disc) for the growth of B. brevis (pH 6.5) were 8.7 (OA), 2.0 (OC), 5.5 (d-OA), 1.1 (OEOA), 54 (OB), 390 (Oa), and 90 (M-Oa) while no inhibition of the bacterial growth was observed for OM-OA, DC-OA, and OP-OA at doses as high as 350 nmol/disc. The results indicate that the toxicities of OA were associated with its isocoumarin moiety but that neither the dissociation of the phenolic hydroxyl group nor the iron-chelating properties of OA were directly related to its toxicities. The lactone carbonyl group of OA, however, appears to be involved in OA toxicity as OP-OA is found in the bile of rats injected with OA and has similar toxicity to that of OA when administered intravenously to the rat. Overall, the structure–activity studies suggest that the toxicity of OA is attributable to its isocoumarin moiety and that the lactone carbonyl group may be involved in its toxicity. q 1996 Academic Press, Inc.

Ochratoxin A (OA), 7-carboxyl-5-chloro-8-hydroxyl-3,4dihydro-3-R-methylisocoumarin-7-L-b-phenylalanine (Fig. 1 and Table 1), is a lactone containing secondary metabolite of some toxigenic species of Aspergillus and Penicillium. This mycotoxin is of concern as it is hepatotoxic, nephrotoxic, teratogenic, and carcinogenic to animals and occurs worldwide in many agricultural commodities (Kuiper-Good1

To whom correspondence should be addressed. Fax: (204) 275-0402.

182

0041-008X/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID

TOX 7747

/

6h0a$$$$81

man and Scott, 1989). The general toxicity of OA has been reviewed by Marquardt and Frohlich (1992) and Castegnaro et al. (1991). Ochratoxin A competitively inhibits the activities of succinate dehydrogenase and cytochrome C oxidase in the mitochondria of rats (Wei et al., 1985). It alters the mitochondrial membrane transportation system and competitively inhibits inner membrane ATPase activity (Meisner and Chan, 1974). OA also competitively inhibits the activity of enzymes involved in the metabolism of phenylalanine (Creppy et al., 1983a,b, 1984, 1990). The precise mode by which enzyme activity is inhibited by OA remains unknown. Several hypotheses for the mechanism of action of OA have been proposed. Chu et al. (1972) proposed that the dissociation of the phenolic hydroxyl group of OA was required for its toxicity. Unlike other halogenated phenols, however, OA does not act as an uncoupler of mitochondrial oxidative phosphorylation (Meisner and Chan, 1974; Wei et al., 1985). In a yeast assay model, OA was found to inhibit protein synthesis through competitive inhibition of phenylalanyltRNA synthase activity (Creppy et al., 1983a,b). Rahimtula et al. (1988) proposed that OA may exert its toxic effect by enhancing membrane lipid peroxidation. Their model emphasized the role of the phenolic hydroxyl group of OA in the formation of an OA–Fe3/ chelate. They postulated that this complex facilitates the shuttle of electrons from a substrate to oxygen to form the reactive oxygen species by complexing with the redox cycle of the cytochrome P-450 system (Hasinoff et al., 1990; Omar et al., 1990, 1991). Although the acute toxicity of OA can be partially prevented by the coadministration of a-tocopherol, OA-induced lipid peroxidation was not observed in OA-treated rats (unpublished data). The results of the current study also indicated that chelation of iron was not essential for the toxicity of some OA analogs and that some strong iron-chelating analogs of OA were weak cytotoxic agents. Recently, Malaveille et al. (1994) suggested that the genotoxicity of OA may involve the formation of an OA phenoxide radical and that the cytotoxicity of OA may involve the formation of thiolcontaining OA derivatives in the cells. Their results implied

02-21-96 11:44:10

toxa

AP: Tox

TOXICITY OF OCHRATOXIN A

FIG. 1. The structure of the ochratoxins and their analogs. The functional group (from R1 to R5) for each toxin is given in Table 1.

that the biotransformation of OA and the covalent modification of biomolecules by OA may be involved in the toxicity of OA. To examine the roles of the functional groups and the hydrophobic side chain of OA in its toxicity, five analogs (Fig. 1 and Table 1) were synthesized and their structures were characterized (Xiao et al., 1995). They were: the ethylamide of OA (OE-OA), the decarboxylated OA (DC-OA), the D-phenylalanine-substituted OA (d-OA, the epimere of OA), the O-methyl ether of OA (OM-OA), and the methyl ester of Oa (M-Oa) (Xiao et al., 1995). These analogs together with the natural analogs of OA (Fig. 1 and Table 1), ochratoxin B (OB, the dechlorinated form of OA), ochratoxin a (Oa, OA without phenylalanine), the ethyl ester of OA (OC), and the lactone-opened form of OA (OP-OA) were utilized in the current study to establish the structure– activity relationship of the ochratoxins. Preliminary results suggested that the anti-microbial activity, cytotoxicity, and animal toxicity of OA and its analogs varied substantially for some analogs and that the lactone of OA may be involved in its toxicity. The objective of this study was to utilize different natural and synthetic analogs of OA to determine their structure–activity relationships in order to establish the mechanism by which OA exerts its biological effects. MATERIALS AND METHODS Chemicals. Ochratoxin A and OB were isolated and crystallized from solid-phase fermentation of wheat inoculated with A. ochraceus Wilhelm (NRRL 3174) using the procedure of Xiao et al. (1996). d-OA, OE-OA, OM-OA, DC-OA, M-Oa, and Oa were synthesized using OA and Oa as starting materials (Xiao et al., 1995). The ethyl ester of OA (OC) was synthesized according to the method of van der Merwe et al. (1965a,b). The purity of each analog was greater than 99% as determined by HPLC coupled photo diode-array spectrophotometry [single peaks as detected over the full uv–vis spectrum (199–900 nm)], NMR (no H-signal other than from the compound), and crystal melting point over a narrow range, ({17C), (Xiao et al., 1995, and unpublished data). Most of the synthesized analogs were novel and no external standard were available for comparison. GC/

AID

TOX 7747

/

6h0a$$$$81

02-21-96 11:44:10

183

MS technique was not used for the underivatized OA analogs as they are essentially not volatile at the maximum operating temperature of GC. The lactone-opened OA (OP-OA) was prepared as described below. TLC plates (C18-coated reversed-phase, 5 1 20-cm and silica, 5 1 20 cm) were from Whatman (Maidstone, England). Acetic anhydride and N-methylimidazole were from Sigma (St. Louis, MO). Tryptic soy–agar was from Difco Laboratories (Detroit, MI). All other analytical grade reagents were from Canlab (Winnipeg, MB). Determination of the phenolic dissociation constants (pKa) of OA and its analogs. The pKa value for each toxin was determined spectrophotometrically with a Beckman DU-8 spectrophotometer (Irvine, CA) using the procedure described by Chu et al. (1972) with modifications. In brief, a stock solution of each toxin was prepared in dimethylsulfoxide (DMSO) at a concentration of 50–500 mg/ml depending on its solubility in the pH 4.0 buffer. Two sets of 0.1 M phosphate buffer solutions were prepared at 0.25 pH unit interval between pH 4.0 and pH 13.5; one set contained the toxin, while the other did not (reference blank). The toxin in DMSO (200 ml) was thoroughly mixed with 1.8 ml of each buffer and the pH and uv absorbancies (365 nm for OB and 380 nm for the others) were determined in quartz cuvettes after mixing. The corresponding buffer without the toxin was used as a blank for the uv absorbancy measurements. The assay was carried out as rapidly as possible in the high pH buffers (pH ú 12) so as to minimize interference from the absorbancy change produced by hydrolysis of the lactone. A standard pH meter (Radiometer, PHM-28, Copenhagen, Denmark) having a SD of { 0.01 pH units was used for the pH measurements. The experiment was replicated twice. The maximum absorbancy change at each pH value was plotted and those data that fell on the line (i.e., the linear portion of the curve) following linear regression analysis were used to calculate the pKa . The pKa was the pH at one-half of the maximum absorbancy change. Iron chelation of OA and its analogs. Ferrous (Fe2/ from FeSO4) and ferric [Fe3/ from Fe2(SO4)3] ions were both chelated by OA to form a reddish-brown color complex which was readily soluble in CHCl3 . FeSO4 (0.05 M, pH 4.5) was utilized to study the chelation capacity of OA and its analogs with the molar ratio of FeSO4 to the toxins being 5 to 1. In this study 100 ml of the toxin in DMSO was mixed with 1.9 ml of the FeSO4 solution and the iron-chelates of the toxins were quantitatively extracted with an equal volume of CHCl3 while the nonchelated and colorless Fe2/ remained in the aqueous phase. The iron-chelating ability of each toxin was graded visually according to the intensity of color in the CHCl3 extract with the grades being strong (3/), moderate (2/), weak (/), and none (0). In contrast to results obtained at pH 4.5, the iron–OA complex was not formed if the pH of the FeSO4 solution was adjusted to a pH 7.2 with a phosphate buffer, presumably due to the insolubility of the FeSO4 or to the inability of the toxins to chelate iron at a higher pH. Lactone hydrolysis. The hydrolysis of the lactone at a high pH was performed using a modification of that first observed by Xiao et al. (1995) with OM-OA and d-OA. In the current procedure, the lactone form of OA was hydrolyzed in strong base and reformed in strong acid. The ring-opened form of OA (OP-OA) was produced by adding 0.5 N NaOH directly to the toxin in DMSO for a period of 2 hr at 257C. This was sufficient time in which to completely hydrolyze the lactone. The hydrolyzed lactone of OA was neutralized to pH 6.2 with solid NaH2PO4 and dried in a vacuum drier (AS160-Savant, Farmingdale, NY) followed by methanol extraction to remove excess salts. HPLC analysis demonstrated that OP-OA was completely stable in solution at pH ú 6.0 and 257C for 24 hr. This preparation of OP-OA was utilized for the toxicity study on Bacillus brevis. The lactone was reformed by acidification to a pH õ 1 with 6 N HCl at 257C for 6 hr. Reconstruction of OA from OP-OA by acidification was prevented by the acetylation of the 3-hydroxyl group with acetic anhydrate using N-methylimidazole as the catalyst. In this procedure, OP-OA (approximately 10 mg) in 10 ml of 1 M phosphate buffer (pH 6.2) was mixed with 100 ml of Nmethylimidazole and 100 ml of acetic anhydride at 257C for 6 hr. The mixture was acidified to pH 1 with 0.5 N HCl and extracted with CHCl3 .

toxa

AP: Tox

184

XIAO ET AL.

Both OA lactone or OP-OA esters in CHCl3 fraction (10 ml) was applied to TLC for identification. The reversed-phase TLC plate (C18-coated, 20 1 5 cm, Whatman) was developed with a solvent mixture of methanol and water (7:3). The acetyl esters of OP-OA exhibited a purple fluorescence (Rf Å 0.92) while OA exhibited blue fluorescence (Rf Å 0.88). Racemization of both OA and OP-OA as indicated by the presence of duplicate bands on the TLC plate, occurred due to the formation of mixed anhydride intermediates of OA and OP-OA with acetic anhydride. Racemization of OPOA also occurred if the concentration of NaOH was too high (ú 5 N). The spectrophotometric assay for the two forms of the toxins is based on the original observation that the lactone form of OM-OA and d-OA have an uv absorption maximum (lmax) at 310 and 380 nm, while the lactone-opened toxins have an uv lmax of 290 and 345 nm, respectively (Xiao et al., 1995). Similar changes were also observed with other forms of OA. The difference in uv lmax between OA and OP-OA also makes it possible to determine the reaction kinetics for the hydrolysis of OA using an uv spectrophotometric technique. The kinetics of the reactions were followed in a quartz optical cuvette (4 ml) using a Beckman DU-8 spectrophotometer. The toxins (OA, OB, and Oa) at concentrations ranging from 80 to 200 mg were dissolved in 200 ml of DMSO, mixed in the cuvettes with 1.8 ml of 0.5 M Na2CO3 (for OA only) or 0.5 N NaOH and the absorbancy at 380 nm for OA and Oa or at 365 nm for OB was recorded at 1 min interval for the first 10 min, then every 5 min until the absorbancy reached nearly 0.1 OD unit. The initial absorbancy was adjusted to 0.9–1 optical density units. The rate constants or half-lives (T1/2) for the hydrolysis of the lactones of OA, OB, and Oa were determined by measuring the time course of the absorbancy change. The calculation was based on Ln(A) Å 0Krtrln(A7),

(1)

T(1/2) Å 0.693/K,

(2)

then

where A7 is the initial absorbancy, A is the absorbancy at any time t, and K is the first-order rate constant of hydrolysis for a given toxin in a given concentration of NaOH. High-performance liquid chromatographic (HPLC) analysis of OA and OP-OA. The two forms of OA were detected using an HPLC fitted with a Waters (Milford, MA) Nova-Pak reversed-phase column (C18, 0.46 1 25 cm, 4 mm, at 427C). The procedure was similar to that routinely used in our laboratory (Frohlich et al., 1988). The mobile phase of the HPLC system consisted of a mixture of solvents A (methanol:isopropanol, 9:1), and B (0.001 M Na2PO4:0.005 M NaH2PO4 , 1:1, pH 6.2). The compounds were eluted using a programmable HPLC gradient system (Water 600E, Milford, MA). The system (1.5 ml/min) delivered 30 to 50% A at an incremental increase of 2%/min (0–10 min), 50 to 85% A at an incremental increase of 3%/min (10–15 min), and 85% A (15–20 min). The compounds were detected using a Shimadzu (Kyoto, Japan) RF-535 fluorescence detector (excitation, 333 nm; emission, 450 nm) and recorded using a Shimadzu CR501 integrator. Cell culture and cytotoxicity assay. HeLa S3 cells (obtained originally from Dr. B. H. Sells, University of Guelph, ON, Canada) were cultured in 25-ml culture flasks using a-minimal essential media, supplemented with 10% fetal bovine serum (FBS, JRH Bioscience, Lenexa, KS) and antibiotics. The assay procedure used for these experiments was described by Terse et al. (1993). Briefly, on Day 1 cells cultured in 25-ml tissue culture flasks were trypsinized and seeded in 96-well plates. On Day 2, test chemicals (OA and its analogs in 20 ml ethanol) were diluted in 1 ml of the culture medium and six logarithmic dilutions were prepared. The culture medium in the 96-well plate by replaced with medium containing toxin and incubated at 377C for 72 hr in a CO2 incubator. The cells were stained with DiffQuik (Baxter, McGaw Park, IL), and representative areas of cells were counted with an inverted light microscope (Nikon, Garden City, NY) equipped with an eye-piece grid. Each toxin was subjected to 10-fold serial

AID

TOX 7747

/

6h0a$$$$82

02-21-96 11:44:10

dilutions and six concentrations and six doses of each concentration were used in the analysis. The concentration (mM) of toxin causing 50% lethality (LC50) was calculated by linear regression analysis (Sigma Plot, Jandel Scientific, San Rafael, CA). Antibacterial assay. B. brevis, which was used for this assay, was obtained from American Type Culture Collection (ATCC; Rockville, MD). The procedures for the assay were as described by Madhyastha et al. (1994). In brief, blank filter paper discs (6.35 mm in diameter, BBL Microbiology System, Becton Dickinson Co., Cockeysville, MD) were placed on a petri plate. The control solvent (20 ml of methanol) and the toxins in methanol (ranged from 0.5 to 200 mg/disc) were applied to the discs one drop at a time using a micropipet. Discs were allowed to dry for 15 min and then placed evenly on the agar surface of each plate previously incubated with bacteria. The pH of the medium was adjusted to 5.5, 6.5, and 7.5 depending on the experiment with 1 N HCl or 0.5 N NaOH. The plates were inverted and preincubated at 47C for 2 hr to allow the uniform diffusion of toxin into the agar. The plates were then incubated at 287C for 24 hr. The inhibition zone diameters were measured in millimeters and corrected for the diameter of the paper disc. In the first study the minimum toxic dose (MTD; nmol/disc) was determined in the pH 6.5 medium. A total of five different doses were selected for each mycotoxin so as to give a net inhibition diameter of between 0 and 15 mm. The doses were selected on the bases of an initial range finding experiment and were prepared by doubling serial dilutions of the stock solution. A plot of the data gave a linear or slightly curvilinear response with all toxins and their analogs (Sigma Plot). The minimum toxic doses for all of the compounds were obtained by extrapolation of the dose– response curve to zero. The experiment was replicated three times. In the second study three doses of each toxin were selected as shown in Table 3 and the influence of pH (pH 5.5 and 7.5) of the incubation medium on the size of the inhibition zone (mm net diameter) was determined. Assays for each dose were carried out in triplicate. The MTD was not calculated for this study due to the limited number of doses tested and the large range of toxicity of the different compounds at the two pH values. Animal studies. The toxicity of OA and its derivatives (two doses/ compound) was determined using 6-week-old mice [15–16 g of body weight (BW)] obtained from the University of Manitoba experimental mouse facility. Ten mice were used for the intraperitoneal (ip) administration of each dose of the toxin (see Table 4 for the doses of each compound). A mixture of DMSO (2 ml/g BW) and corn oil (20 ml/g BW) was the vehicle for the toxins. Each toxin was first dissolved in DMSO and then mixed with corn oil before injection. All mice were fed and watered ad libitum. The total number of deaths in each treatment group was determined at 12-hr intervals over a 72-hr period after administration of the toxin. Preliminary studies demonstrated that toxicity following ip administration of OA occurred within 6 hr and that nearly all deaths occurred within 48 hr (data not shown). A 72-hr period was therefore selected as being representative of the acute toxicity of OA as this period was of sufficient duration to include all of the acute toxicities. A second study was carried out to determine whether OP-OA could be detected in the blood, bile, or urine of rats injected intravenously (iv) with OA. A total of four female Sprague–Dawley rats weighing 300 { 20 g ({ SD) were injected with 100 mg of OA in 1 ml vehicle and two rats were injected with vehicle but no toxin. The toxins were dissolved in 200 ml ethanol and made up to 1.0 ml with 800 ml of 0.145 M saline. The toxin was administered via the carotid artery over a 2-min period and blood was collected 3 hr after infusion of the toxin into the jugular vein while bile and urine were collected from the bile duct and the ureter 2.5 to 3 hr after infusion of the toxin. The surgical procedures were as outlined by Mulder et al. (1981) except that the rats were anesthetized with pentobarbital throughout the study. The body fluids were mixed with 4 parts of methanol, centrifuged at 11,000g in a IEC Centra-M microcentrifuge (ICE, Boston, MA) for 10 min and injected onto a C-18 reversed-phase column as described above. The same experiment was repeated with four additional rats

toxa

AP: Tox

185

TOXICITY OF OCHRATOXIN A

TABLE 1 Structures, pKa Values, and Iron Chelation of Ochratoxin A (OA) and Its Analogs Structure of OA analogs OA and its analogs Ic OA I d-OA I OC I OE-OA I DC-OA I OM-OA I OB IIc Oa II M-Oa III OP-OAe III OP-OA-Ace,g

R1

R2

R3

R4

R5

pKaa

Fe2/ chelationb

COOH H COOCH2CH3 CONHCH2CH3 H COOH COOH — — COOH COOH

H COOH H H H H H — — H H

H H H H H CH3 H H H H H

Cl Cl Cl Cl Cl Cl H Cl Cl Cl Cl

— — — — — — — H CH3 — —

7.0 { 0.02 7.2 { 0.02 7.2 { 0.05 6.7 { 0.05 7.9 { 0.10 NAd 7.8 { 0.02 11.6 { 0.05 8.2 { 0.15 ND f ND f

// // /// /// 0 0 / 0 0 /// ND f

a

The pKa values { SD are for the dissociation of the phenolic hydroxyl group (R3) but not for the carboxyl group (R1, R2 or R5). Iron-chelate of OA and its analogs at pH 4.5 as determined visually by color change: strong (///), moderate (//), weak (/) and none (0). See Materials and Methods for further details. c See Fig. 1 for the basic nuclear structure of the ochratoxins. d NA, not applicable as the phenolic group is methylated. e Refer to Fig. 2, III, for structure. f ND, not determined. g Acetylated OA-OA. The 3-hydroxyl group after hydrolysis of the lactone was acetylated as outlined under Materials and Methods. b

except that the route of administration was intraperitoneal. In both groups the time of death was recorded. In a final experiment, female Sprague–Dawley rats weighing 200 { 15 g ({ SD) were administered either 100 or 500 mg of either OC, OA, or OP-OA in 1 ml saline iv. There were four treatments and four rats per treatment group. Vehicle without toxin was also administered iv to four rats. The procedures were as described above except that the jugular vein, bile duct and ureter were not cannulated and the fluids were not collected. The time of death was recorded.

RESULTS

Dissociation and Iron Chelation of Ochratoxin A and Its Analogs Ochratoxin A and all of its analogs except OM-OA contain a dissociable phenolic hydroxyl group that undergoes a red shift in the uv lmax (from 333 to 381 nm) when converted from the protonated to the dissociated form (Chu et al., 1972). The proportion of the dissociated species in solution at a given pH value can therefore be determined spectrophotometrically. The pKa values of the phenolic hydroxyl group for OE-OA, OA, d-OA, and OC as determined by this procedure were lower than those for OB, DC-OA, M-Oa, and Oa (Table 1). The higher pKa values of OB (pKa Å 7.8) and DC-OA (pKa Å 7.9) compared to OA (pKa Å 7.0) suggests that the presence of the electron-withdrawing chlorine (isocoumarin ring) and carboxyl (phenylalanine) enhanced the dissociation of the phenolic hydroxyl group. The pKa for OA, OC, and OB (7.0, 7.2, and 7.8, respectively) were similar (within 0.2 pKa units) to those in the literature (Chu et

AID

TOX 7747

/

6h0a$$$$82

02-21-96 11:44:10

al., 1972) except that a higher value for Oa was obtained in the current study (11 vs 11.6). This difference may be attributed to the unstable nature of Oa in alkaline solution (pH ú 12). Under these conditions and as discussed below, the lactone of Oa is hydrolyzed in a pH- and time-dependent manner to yield a new compound with a different uv lmax . A shift in lmax from 380 to 345 nm due to the hydrolysis of the lactone would affect the calculation of pKa values for toxins that had high values (i.e., Oa). The true pKa value because of this effect may therefore be somewhat higher than that reported in this paper and is higher than that reported by Chu et al. (1972). In the current study the linear portion of the absorbancy curve was used and the assay was carried out in a minimum time so as to reduce this effect. The iron chelation study (Table 1) indicated that among the analogs tested only OE-OA, OC, d-OA, OP-OA, OA, and OB were able to chelate Fe2/ ion at a low pH (pH 4.5). Rahimtula et al. (1988) have suggested that the chelation of iron by OA involved the coordination of the C-8-phenolic and C-12-carbonyl oxygen atoms and that this complex facilitated the reduction of Fe3/ to Fe2/. The results of current study, in contrast to those of Rahimtula et al. (1988), indicated that the C-14-carboxyl group from the phenylalanine moiety may also be involved in the formation of the toxin– iron complex. This is based on the observation that no iron was chelated by those compounds that did not contain phenylalanine (DC-OA, Oa, and M-Oa, compounds without the 14-C carboxyl group) or when the phenolic oxygen was blocked (OM-OA) (Table 1). These studies, however, do not

toxa

AP: Tox

186

XIAO ET AL.

FIG. 2. The synthetic reaction scheme for the hydrolysis of the lactones of ochratoxins with base: I, protonated lactone; II, deprotonated lactone; and III, the ring-opened hydroxycarboxylate.

The hydrolysis of the lactone ring was further confirmed as restoration of the lactone structure by reacidification was prohibited if the 3-hydroxyl group of OP-OA was acetylated by acetic anhydride using N-methylimidazole as a catalyst. Acetylated OP-OA exhibited a purple fluorescency and had a Rf value of 0.92, while OA exhibited a blue fluorescency and had a Rf value of 0.88 on reverse-phase TLC that was developed in a mixture of methanol and water (7:3). The uv lmaxs of the sodium salt of OP-OA in 0.5 N NaOH were 240 nm (e Å 7750) and 345 nm (e Å 5480). The two forms of OA were also readily separated using HPLC analysis (Fig. 3). The open form of OA (OP-OA, 5:15 min, Fig. 3B) that was produced by base hydrolysis of OA (0.5 N NaOH for 2 hr) and then neutralized with Na2HPO4 to a pH of 6.2 was almost completely stable for 12 hr. An identical amount of OA (12.4 min, Fig. 3C) was restored from OP-OA by acidification of OP-OA that was

preclude a structural requirement of all three oxygen atoms and the amide nitrogen atom for the binding of iron. Lactone Hydrolysis The hydrolysis of the lactone of OM-OA, a new analog of OA (see Table 1), was first observed during base-catalyzed hydrolysis of its methyl ester of OM-OA (unpublished observations, Xiao et al., 1995). In this study both OM-OA and its ester had an uv lmax at 290 nm instead of at 333 nm. It was rationalized that this shift should not have occurred in alkaline solution as this compound did not contain a dissociable phenolic group. A time-dependent blue shift in uv lmax from 333 to 290 nm was, nevertheless, observed when OMOA was dissolved in an alkaline solution (0.5 N NaOH). This blue shift was attributed to the hydrolysis of its intramolecular ester bond (lactone ring). In the current study a similar spectral behavior (blue shift) was also observed with OA. Ochratoxin A initially exhibited an instantaneous red shift in uv lmax from 333 to 380 nm in 0.5 N NaOH (pH ú 13) due to the dissociation of the C-8 phenolic hydroxyl group followed by a time-dependent blue shift in uv lmax from 380 to 345 nm. In a solution of 0.5 M Na2CO3 (11 ú pH ú 10), OA exhibited an identical red shift in uv lmax (from 333 to 381 nm) but not the time-dependent blue shift (from 381 to 345 nm) as seen in 0.5 N NaOH. The blue shift in the uv lmax of OA is attributed to the hydrolysis of its lactone. On the basis of this observation, it is proposed that three forms of OA can occur. The flow scheme for the protonated (I) and deprotonated (II) forms of the OA lactones and OP-OA (III) is outlined in Fig. 2. The protonated form of OA (lmax Å 333 nm, Fig. 2, I) is instantaneously dissociated (lmax Å 380 nm, Fig. 2, II) in 0.5 N NaOH followed by a timedependent hydrolysis of the lactone ring to OP-OA (lmax Å 345 nm, Fig. 2, III). The reaction was reversed by reacidification of the alkaline solution to a pH õ 1 with 6 N HCl.

AID

TOX 7747

/

6h0a$$$$83

02-21-96 11:44:10

FIG. 3. Reverse-phase HPLC elution profile of OA standard (A), lactone-opened OA, OP-OA (B), and OA reformed from OP-OA (C). The early eluting peaks are solvent peaks. See Materials and Methods for further details.

toxa

AP: Tox

187

TOXICITY OF OCHRATOXIN A

FIG. 4. The kinetic profile for the hydrolysis of the lactones of ochratoxin A (OA), ochratoxin B (OB) and ochratoxin a (Oa): (1) hydrolysis of OA lactone in 0.5 M Na2CO3 , ln A Å 00.00016rt00.1053, R2 Å 0.9998; (2) hydrolysis of OB lactone in 0.5 N NaOH, ln A Å 00.02652rt / 0.0032, R2 Å 0.9995; (3) hydrolysis of OA lactone in 0.5 N NaOH, ln A Å 00.1095rt 0 0.0676, R2 Å 0.9985; and (4) hydrolysis of Oa lactone in 0.5 N NaOH, ln A Å 00.2647rt 0 0.0115, R2 Å 0.9956. See Materials and Methods further details.

in the pH 6.2 buffer or OP-OA in the NaOH solution with 6 N HCl to a pH õ 1 for 6 hr followed by neutralization with NaH2PO4 to a pH of 6.2 for 12 hr. Ochratoxin A was readily partitioned into CHCl3 from the aqueous phase at pH 6.0, while the OP-OA remained in the aqueous phase, indicating a greater degree of ionization (extra carboxylate anion) of the latter compound.

(Table 2). OE-OA and OC were 8- and 4-fold more toxic to B. brevis, respectively, than the parent compound, OA. In contrast, no antibacterial activity was observed at a dose more than 40-fold higher than the toxic dose of OA when the carboxyl group was removed (DC-OA), the hydroxyl group was blocked (OM-OA) or the lactone was hydrolyzed (OP-OA). The absence of the chlorine atom (OB) resulted in a 6-fold reduction in the toxicity of OA to B. brevis. Interestingly, Oa and its methyl ester (M-Oa) also possessed antibacterial activity even though they were 45- and 10-fold less toxic, respectively, than OA. d-OA and OA had similar toxicities to the bacteria. Compared to the prokaryotic system, the cytotoxicity of OA and its analogs to the mammalian (HeLa) cell line yielded the same relative pattern of toxicity except for OEOA and d-OA (Table 2). Compared to OA, OE-OA and dOA were 8- or 1.5-fold more toxic to B. brevis but were 2000- and 30-fold less cytotoxic to HeLa cells, respectively. The lower toxicity of d-OA in HeLa cells relative to OA may be attributed to steric differences as discussed below. The reason for the differential effects of the two toxin in HeLa cells compared to B. brevis was not established. Possibly, OE-OA could inhibit the synthesis of some of the vital components required for the normal growth of the bacillus while such components were not required for the mammalian cells. Alternately, the different compounds could cross prokaryotic or eukaryotic cell walls by different mechanisms

TABLE 2 The Comparative Toxicity of Ochratoxin A (OA) and Its Analogs to B. brevis and HeLa Cellsa

Rate of Hydrolysis of the Lactone Group The reactivity of the lactone carbonyl group of OA and two common analogs (OB and Oa) to a hydroxyl anion (HO0), which is analogous to the bases in the binding site of the enzymes (for example, the hydroxyl, thiol and amino groups of serine, cysteine and lysine residues, respectively), was evaluated in 0.5 N NaOH (pH ú 13) and in 0.5 M Na2CO3 (pH õ 11). The rate of hydrolysis of the lactone was first order and was pH- and time-dependent (Fig. 4). The half-lives for the hydrolysis of the lactone form of Oa, OA, and OB in 0.5 N NaOH were 2.6, 6.6, and 26.1 min, respectively, while the half-life of OA was 4330 min in 0.5 M Na2CO3 , corresponding to approximately a 700-fold slower rate of hydrolysis as compared to that in 0.5 N NaOH. The results suggest that the lactone carbonyl groups of Oa and OA are much more susceptible to the nucleophilic attack than that of OB. Antimicrobial Activity and Cytotoxicity The toxicity of OA to B. brevis in a pH 6.5 incubating medium varied substantially when its structure was modified

AID

TOX 7747

/

6h0a$$$$83

02-21-96 11:44:10

Analogs

HeLa cell LC50b (mM)

B. brevis MTDc (nmol/disc)

OA OC OB d-OA Oa OM-OA DC-OA OE-OA M-Oa OP-OA

0.005 0.009 0.054 0.163 0.56 0.83 7.6 10.1 NDe ND

8.7 { 1.7 2.0 { 0.5 54 { 0.6 5.5 { 2.1 390 { 20 NI NId 1.1 { 0.2 90 { 10 NI

a

See Materials and Methods for detailed description of assays. LC50, the concentration of toxin required to reduce number of cells by 50% as determined by linear regression. A total of six doses of the toxin was used for each assay with the total number of replicates per dose being six. The mean coefficient of variation for all analysis was the LC50 { 12%. c MTD, minimum toxic dose { SD for measurable inhibition of B. brevis growth in a medium at a pH of 6.5. A total of five doses of the toxin was used for each assay with the number of assays per dose being three. The experiment was replicated three times. d NI, no inhibition attained at 350 nmol/disc. e ND, not determined. b

toxa

AP: Tox

188

XIAO ET AL.

TABLE 3 Effect of the pH of the Culture Medium on the Toxicity of OA and Its Analogsa Incubation pH (inhibition zone, mm)d Analogb

Analog concentration (nmol/disc)c

OE-OA

1.2 2.4 4.8 4.6 9.2 18.4 5 10 20 5 10 20 54 108 216 195 390 780 60 120 240 92 184 368 70 140 280

OC

OA

d-OA

OB

Oa

OM-OA

M-Oa

DC-OA

5.5 { { { { { { { { { { { { { { { { { { 0 0 6{ 0 0 0 0 0 0

6 8 11 9 12 14 6 10 14 7 10 13 6 9 11 9 12 15

7.5 2 2 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

1

0 0 3{ 9{ 10 { 14 { 0 0 0 0 0 0 0 0 0 0 0 4{ 0 0 0 4{ 5{ 7{ 0 0 0

1 1 1 1

1

1 1 1

a Values represent means of one set of analyses with each analysis being carried out in triplicate. See Materials and Methods for a description of this assay. b Analogs were ranked in ascending order according to their toxicity at pH 5.5. c Amount of analog added to the disc (nmol/disc). d Net size of inhibition zone (mm) produced by the addition of the toxin to the disc.

and at different rates. OE-OA was also considerably less toxic to mice than OA (Table 4). Ochratoxin C, which is structurally close to OE-OA, in contrast, was as cytotoxic as OA to both cell types. Ochratoxin C may, however, act as a pro-OA as it is efficiently and rapidly converted into OA by esterases in vivo (Fuchs et al., 1984). DC-OA was essentially nontoxic to both types of cells. The result with d-OA and OE-OA, although not explained, suggests that it may be possible to develop low toxicity antibiotics by the structural modification of this isocoumarin derivative. The Effect of Medium pH on the Toxicity to B. brevis The effect of pH of the culture media on the antibacterial activity of OA and its analogs (Table 3) was examined using

AID

TOX 7747

/

6h0a$$$$83

02-21-96 11:44:10

the B. brevis disc diffusion assay. The results demonstrate that pH can have a profound effect on the toxicity. In most comparisons, the toxicity of OA and its analogs (OE-OA, dOA, OB, Oa, and OM-OA) greatly increased when the pH of the incubation medium decreased from 7.5 to 5.5. Interestingly, the toxicity of the ester form of OA (OC) was not affected by the pH of the medium, while that of M-Oa and structurally similar OE-OA were substantially increased when the pH of the culture media was increased. The reason for the dramatic differences in toxicity as affected by pH was not established in the current studies. However, a similar pattern of response was obtained for OA and OC (Madhyastha et al., 1994). It was hypothesized that the much greater toxicity of OA at the lower pH was related to its ionic state with the noncharged form being more readily transported across the lipid membrane or cell walls. Ochratoxin C, which does not contain a free carboxyl group, would not be ionized at the lower pH. Therefore, its uptake, as indicated by the toxicity data, would not have been affected by the pH of the culture medium. These results are consistent with those of Kumagai (1985 and 1988) who demonstrated that the protonation of OA in the upper small intestine due to the low pH environment greatly facilitate its absorption into the systemic circulation. Toxicity of Ochratoxin A and Its Analogs to Mice The toxicity of OA and its analogs to mice was showed in Table 4. All mice in the control group survived. There was a 30 and 90% death rate during the 72-hr period following ip injection of 20 and 50 mg/kg of OA, respectively. Mice injected with 20 mg/kg of OA exhibited no abnormal signs during the first 3 hr, and no deaths during the first 24 hr. There were two deaths during the 24- to 36-hr period and another during the 48- to 72-hr period after injection. Mice

TABLE 4 Toxicity of OA and Its Analogs to Mice Analogs Control OA d-OA OE-OA DC-OA

Dosage (mg/kg BW)a

Death rateb

0 20 50 50 200 200 500 200 500

0/10 3/10 9/10 0/10 0/10 0/10 10/10 0/10 10/10

a

Analogs OP-OA OM-OA OB Oa M-Oa

Dosage (mg/kg BW)

Death rate

500 200 500 200 500 200 500 200 500

0/10 0/10 10/10 0/10 10/10 0/10 10/10 0/10 10/10

Dosage, the amount of toxin (mg) injected (ip) into mice per kg body weight (BW). b Death rate, numbers of dead mice/10 mice dosed 72 hr after injection. See Materials and Methods for the description of assay.

toxa

AP: Tox

189

TOXICITY OF OCHRATOXIN A

injected with 50 mg/kg of OA showed no apparent abnormal signs during the first 3 hr. There were six deaths between 12 and 24 hr and three between 24 and 72 hr with one survivor. All mice injected with 50 or 200 mg/kg of d-OA and 200 mg/kg of OE-OA, DC-OA, OM-OA, OB, Oa, and M-Oa survived the 72-hr test period. Symptoms of panting and hyperactivity were observed after approximately 30 to 60 min in mice injected with 200 mg/kg of OE-OA, DCOA, OB, and Oa. None of these mice died during subsequent 3-week period after which the experiment was terminated. The mice showed abnormal behavior such as panting, loss of coordination, gasping, and choking 30 min after the injection of 500 mg/kg of DC-OA, OM-OA, OB, Oa, and Moa; they then became sedated with cyanosis being observed on the lips and tails. Deaths were observed within 2 hr of injection with all mice dying within a 12-hr period. Mice injected with 500 mg/kg of OE-OA exhibited similar but less severe signs and abnormal behavior. No death occurred in this group within 3 hr of injection, while all of the mice died within a 12-hr period of injection. Autopsy of the dead mice did not show apparent organ damage. In contrast to the above analogs, all mice survived 72 hr when they were injected with OP-OA at a dose of 500 mg/kg body weight. Ochratoxin alpha has been considered to be nontoxic to the animal (Chu et al., 1972). The current study indicated that Oa was less toxic compared to OA but was toxic at a higher dose. Studies in the Rat The open form of OA (OP-OA) was detected in the bile of rats injected with OA but not in the blood or urine. The concentrations of OP-OA { SD, 6 to 7 hr after injection of OA, were 0 for blood, 0 for urine, and 106 { 9 ng/ml for bile, while the corresponding concentrations for OA { SD were 640 { 32, 305 { 84, and 100 { 16 ng/ml. Studies with the control fluids obtained from rats not injected with OA demonstrated that no peaks coeluted with OA or OPOA. Also incubation of the fluids at pH values of õ 1.2 for 4 hr resulted in the disappearance of the OP-OA peak in the bile and an increase in size of the OA peak which provides confirmation that the new peak was OP-OA. This peak did not form in any of the biological fluids when they were spiked with OA. Also the OP-OA peak was stable for more than 24 hr at physiological pH values. In this experiment all of the rats died within 4 to 6 hr following iv administration of OP-OA, while none died when the route of administration was ip or when the vehicle was administered. Clearly route of administration influences the lethality of OP-OA. In the final study two concentrations of OA, OC, and OPOA were administered to smaller rats (200 g vs 300 g) compared to those in the previous experiment. All deaths occurred within 6 hr of toxin administration. The toxin administered, its dosage and the number of deaths out of four

AID

TOX 7747

/

6h0a$$$$83

02-21-96 11:44:10

rats in each group were as follows: OA (100 mg), 04; OA (500 mg), 24; OP-OA (100 mg), 14; OP-OA (500 mg), 34; OC (100 mg), 14; OC (500 mg), 24, vehicle alone, 04. These results demonstrate that all three forms of OA were toxic when administered iv with the OP-OA form being as toxic or more toxic than OA. This effect was not caused by an interconversion of OP-OA to OA as demonstrated by clearance studies with OP-OA (data not shown). DISCUSSION

Ochratoxin A, a potent mycotoxin, has been studied since it was first discovered by van der Merwe et al. (1965a, b). The molecular mechanism of action of OA remains unclear even though several hypotheses have been proposed. The structure requirement for OA toxicity was proposed by Chu et al. (1972), using the natural analogs of OA such as OB, Oa, and OC and a synthetic analog, the O-methyl ether of OC. The report suggested that the dissociation of the phenolic hydroxyl group of OA was required for its toxicity. Creppy et al. (1983a, b, 1984, 1990) proposed that the phenylalanine moiety of OA exerted its toxic effect by competitive inhibition of enzymes involved in phenylalanine metabolism. Rahimtula et al. (1988) and Omar et al. (1990, 1991) proposed that an OA–iron complex may play an important role in OA toxicity. Results from the current study suggest a new mechanism of action of OA involving the lactone moiety. Data from the literature and the current research suggest that the phenylalanine moiety of OA functions to guide the toxin to the appropriate phenylalanine metabolizing enzymes while the isocoumarin portion of the molecule is responsible for its toxic effects. Structurally, it is logical to consider OA as an analog of this amino acid. Creppy et al. (1983a, b, 1984), for example, reported that OA competitively inhibited phenylalanine tRNA synthase with the toxicity of OA being partially prevented by the coadministration of phenylalanine. Other phenylalanine metabolizing enzymes are also specifically affected by OA (Creppy et al., 1990; Parker et al., 1982; Pitout and Nel, 1969). The results from the current study support this proposal as the toxicity of d-OA was approximately 30-fold less to HeLa cells than OA while it was essentially non-toxic to mice. OA in contrast was highly toxic to mice. Presumably d-OA, which contains the D rather than the L isomer of phenylalanine, either was not absorbed or was not metabolized by the mammalian cells. The similar toxicities of d-OA and OA in B. brevis may be attributed to the ability of bacteria to better utilize d-amino acids than mammals (Davis, 1990). Another possible explanation is that the vital components of bacteria are different from the mammalian cells (the cell wall for example). This may also explain why OE-OA is more toxic to B. brevis but much less toxic to HeLa cells as compared to OA.

toxa

AP: Tox

190

XIAO ET AL.

Studies by Roth et al. (1989, 1993) with OB, the dechlorinated form of OA, demonstrated that this compound did not inhibit phenylalanyl-tRNA synthase, further suggesting that the phenylalanine moiety of OA was not the reactive group and that the toxicity of OA is mainly attributed to its isocoumarin moiety. The bacterial and animal toxicity data for Oa and M-Oa, forms that do not contain phenylalanine, also suggest that the phenylalanine side chain is not absolutely essential for OA toxicity. Although the phenylalanine moiety may not be essential for OA toxicity it appears that the carboxyl group of OA from phenylalanine (and probably that from other amino acids) enhances the toxicity of the isocoumarin portion of the molecule. DC-OA, for example, was approximately 10-, 50-, and 1500-fold less toxic to mice, B. brevis, and HeLa cells, respectively, than OA. Overall, it would appear that the carboxyl and the Cl groups contribute to the toxicity of OA, while the steric orientation of the phenylalanine moiety apparently affects the disposition and probably the enzyme-binding of the toxin. The conclusion that there is a direct correlation between the toxicity of different analogs of ochratoxins and their dissociation constants (phenolic group) (Chu et al., 1972) or iron chelating capacity (Rahimtula et al., 1988) is only partially supported by the observations in the current study. In B. brevis there appeared to be a positive association between the toxicity of some of the analogs and their iron binding capacity. For example, OC and OE-OA had the highest iron binding capacity, OA and d-OA intermediate values, OB low values, and DC-OA and OM-OA no binding capacity. The corresponding toxicity to B. brevis also followed the same pattern. In contrast, Oa and M-Oa were toxic at high doses but were not able to bind any iron. These latter observations which were also observed for the HeLa cells and mice would suggest that iron binding is not an obligatory requirement for the toxicity of OA analogs as proposed by Rahimtula et al. (1988). According to Chu et al. (1972), the toxicity of OA is directly related to the degree that C-8 phenolic hydroxyl group is ionized. The current study, however, demonstrates that the toxicity of Oa (pKa Å 11.6) to mice is not 10,000-fold less toxic than OA (pKa Å 7) as it should be according to their hypothesis. The toxicity to mice and HeLa cells of OM-OA, an analog in which the phenolate is blocked, further suggests that a dissociable phenolic hydroxyl group in OA is not essential. OMOA was more toxic to both B. brevis and HeLa cells than DC-OA, suggesting that the carboxyl group plays a more important role in OA toxicity than its phenolate group. These observations therefore question the role of the phenolate group and iron-chelation in the toxicity of OA as hypothesized by other researchers. In this current study a new form of OA was discovered, OP-OA. This form is readily produced in vitro at high pH, is relatively stable at physiological pH and is converted to

AID

TOX 7747

/

6h0a$$$$84

02-21-96 11:44:10

OA at low pH. Similarly, OB and Oa are converted into their open forms in the presence of a strong base and into their closed form at low pH values. The open form of the toxin was detected in the bile of rats treated with OA but not in the urine or blood of the same rats. The amount of OP-OA formed was significant as approximately 50% of the total OA detected in the bile was present as the open form. The open form of OA was also highly toxic to rats when administered iv but no ip with the toxicity following iv injection being similar or even greater than that of iv injected OA. However, the open form when administered to mice ip or to B. brevis was much less toxic than OA. The reason for these differences in toxicity was not established but it may be related to the rate of uptake of the toxin by B. brevis and the rate of delivery of the toxin to the target organs (liver and kidney) in the case of mammals. Presumably the open form, being more polar than OA, would not be as readily taken up in the rat particularly when it is not delivered directly to the liver. This hypothesis is supported by the observations of Kumagai (1988) who reported that the rate of absorption of OA across the intestinal epithelium is much slower when it is in its more polar form. It is not clear, however, if the chronic effects of the toxin are the same and if route of administration of the toxin would affect them differentially. Chronic studies should therefore be carried out as the principal toxic effects of OA are mainly associated with its carcinogenic, nephrotoxic, immunosuppressive, and hepatotoxic effects (Marquardt and Frohlich, 1992). Although this is the first study that has proposed the existence of the open form of OA, studies by Trivedi et al. (1992) with an unidentified compound was probably also with OP-OA. They reported that a new compound was produced when OA was heated (1757C) in the presence of 0.1 N NaOH and that its cytotoxicity to HeLa cells was substantially reduced compared to that of the parent OA. This compound seems likely to be OP-OA even though OA was only partially converted to OP-OA according to the uv spectrum of their reaction mixture. The mode by which OA is converted to OP-OA in the cell and the means by which it exerts its toxic effect, has not been established. Studies on the rate of hydrolysis of the lactone form of Oa, OA, and OB in 0.5 N NaOH demonstrated that the rate of hydrolysis of Oa was approximately twice that of OA and 10 times that of OB. The relative in vivo toxicities of the three compounds, however, are different with OA being the most toxic, OB being 10-fold less toxic than OA, and Oa being essentially nontoxic (Chu, 1979; Marquardt and Frohlich, 1992). These results suggest that there is not a simple relationship between the toxicity of these forms and their rate of hydrolysis. Although Oa is not toxic in vivo it has been shown to be more effective in vitro than OA at inhibiting state 3 respiration in mitochondria, while OB was completely ineffective (Moore and Truelove,

toxa

AP: Tox

191

TOXICITY OF OCHRATOXIN A

1970). These latter data, in contrast to the in vivo data, are consistent with the rate of hydrolysis of the lactone as discussed above. The difference in toxicity between the in vitro and in vivo studies may be attributed to differences in several factors such as degree of uptake of the toxins by the cells, rate of clearance from the body (i.e., very rapid in vivo for Oa but not OA) and the actual toxicity once the toxins are taken up by the cell. Clearly, additional studies need to be carried out to establish the relationship between the structural characteristics of OA and its analogs and factors that affect their uptake by the cell and their toxicity once they are taken up by the cell. The research from the current study, nevertheless, has suggested that the lactone group of OA may be involved in its toxicity. In addition to OA, many biologically active lactones have been isolated from nature or chemically synthesized (Nobilec et al., 1994; Sato et al., 1992; Nozawa et al., 1981). Their activities involve the covalent modification of some vital biomacromolecules (enzymes for example) and their biological activities were lost if they were prehydrolyzed (Kam et al., 1992; Saltz et al., 1993; Luthi-Peng et al., 1992; Borgstrom, 1988). For example, the mechanism of inhibition of serine proteases by isocoumarin compounds involved the formation of a covalent acylenzyme complex between the hydroxyl group of serine in the active site of the enzyme and the lactone carbonyl group of isocoumarin (Kam et al., 1992; Hernandez et al., 1992). The adducts of OA that have been identified (Grosse et al., 1995; Xiao et al., 1995b) may be formed by a similar mechanism. Nevertheless, the open form of OA may also produce its toxic effects by another mechanism as this form of OA, in contrast to studies with other open forms of different lactones (see previous discussion), is toxic. In summary, the current studies on the structure – activity relationships of derivatives of OA and Oa have suggested that: (1) the hydroxyl, carboxyl, chlorine, and lactone groups of OA substantially affect its toxicity to B. brevis, its cytotoxicity to HeLa cells, and its toxicity to mice and rats, (2) the biological reactivity of OA may in part be associated with the lactone carbonyl group of its isocoumarin moiety, and (3) there may not be a direct relationship between the toxicity of the ochratoxins and the extent of iron chelation, the dissociation of the phenolic group, and the presence of phenylalanine side chain even though these groups may indirectly enhance the toxicity of OA. Additional research must be carried out to conclusively demonstrate the role of the lactone group in the toxic action of OA. This question could possibly be addressed by the synthesis of analogs without the lactone ring including the synthesis of compounds that maintain the bicycle ring system while removing one or both of the lactone oxygens.

AID

TOX 7747

/

6h0a$$$$84

02-21-96 11:44:10

ACKNOWLEDGMENTS This work was supported by Natural Science and Engineering Research Council (NSERC) of Canada, the University of Manitoba, and Agriculture Canada.

REFERENCE Borgstrom, B. (1988). Mode of action of tetrahydrolipstatin: A derivative of naturally occurring lipase inhibitor lipstatin. Biochim. Biophys. Acta. 962, 308–316. Castegnaro, M., Plestina, R., Dirheimer, G., Chernozemsky, I. N., and Bartsch, H. (1991). Mycotoxins, Endemic Nephropathy and Urinary Tract Tumours, Vol. 115. IARC Sci. Publ., Oxford Univ. Press, Oxford, UK. Chu, F. S., Noh, I., and Chang, C. C. (1972). Structure requirements for ochratoxin A intoxication. Life Sci. 11, 503–508. Creppy, E. E., Kern, D., Steyn, P. S., Vleggaar, R., Roschenthaler, R., and Dirheimer, G. (1983a). Comparative study of the effect of ochratoxin A analogues on yeast aminoacyl-tRNA synthases and on the growth and on the protein synthesis of hepatoma cells. Toxicol. Lett. 19, 217–224. Creppy, E. E., Stormer, F. C., Roschenthaler, R., and Dirheimer, G. (1983b). Effects of ochratoxin A metabolites on yeast phenylalanyl-tRNA synthase and on the growth and in vivo protein synthesis of hepatoma cells. Chem. Biol. Interct. 47, 239–247. Creppy, E. E., Roschenthaler, R., and Dirheimer, G. (1984). Inhibition of protein synthesis in mice by ochratoxin A and its prevention by phenylalanine. Food Chem. Toxicol. 22, 883–886. Creppy, E. E., Chakor, K., Fisher, M. J., and Dirheime, G. (1990). The mycotoxin ochratoxin A is a substrate for phenylalanine hydroxylase in isolated rat hepatocytes and in vivo. Arch. Toxicol. 64, 279–284. Davis, B. D. (1990). Biosynthesis of amino acids. In Microbiology (B. D. Davis, R. Dulbecco, H. N. Eisen, and H. S. Ginsberg, Eds.), 4th ed., pp. 93–94. Lippincott, Philadelphia. Frohlich, A. A., Marquardt, R. R., and Bernatsky, A. (1988). Quantitation of ochratoxin A: Use of reverse phase thin-layer chromatography for sample cleanup followed by liquid chromatography or direct fluorescence measurement. J. Assoc. Off. Anal. Chem. 71, 949–953. Fuchs, R., Hult, K., Peraica, M., Radic, B., and Plestina, R. (1984). Conversion of ochratoxin C into ochratoxin A in vivo. Appl. Environ. Microbiol. 48, 41–42. Grosse, Y., Baudrimont, I., Castegnaro, M., Betbeder, A-M., Creppy, E. E., Dirhemer, G., and Pfohl-Leszkowicz, A. (1995). Formation of ochratoxin A metabolites and DNA-adducts in monkey kidney cells. Chem.–Biol. Interact. 95, 175–187. Hasinoff, B. B., Rahimtula, A. D., and Omar, R. F. (1990). NADPHcytochrome-P-450 reductase promoted hydroxyl radical production by the iron (III)-ochratoxin A complex. Biochim. Biophys. Acta 1032, 78– 81. Hernandez, M. A., Powers, J. C., Golinski, J., Oleksyszyn, J., Vijayalakshmi, J., and Meyer, E. F. (1992). Effect of the 7-amino substituent on the inhibitory potency of mechanism-based isocoumarin inhibitors for porcine pancreatic and human neutrophil elastases: A 1.85-A X-ray structure of the complex between porcine pancreatic elastase and 7-[(N-tyrosylphenylalanyl)amino]-4-chloro-3-methoxyisocoumarin. J. Med. Chem. 35, 1121–1129. Kam, C. M., Kerrigan, J. E., Dolman, K. M., Goldschmeding, R., Von dem Borne A. E., and Powers, J. C. (1992). Substrate and inhibitor studies on protease 3. FEBS Lett. 297, 119–123. Kuiper-Goodman, T., and Scott, P. M. (1989). Risk Assessment of the Mycotoxin Ochratoxin A. Biomed. Environ. Sci. 2, 179–248.

toxa

AP: Tox

192

XIAO ET AL.

Kumagai, S. (1985). Ochratoxin A: Plasma concentration and excretion onto bile and urine in albumin-deficient rats. Food Chem. Toxicol. 23, 941–943. Kumagai, S. (1988). Effects of plasma ochratoxin A and luminal pH on the jejunal absorption of ochratoxin A in rats. Food Chem. Toxicol. 26, 753–758. Luthi-Peng, Q., Mark, H. P., and Hadvary, P. (1992). Identification of the active site serine in human pancreatic lipase by chemical modification with tetrahydrolipstatin. FEBS Lett. 299, 111–115. Madhyastha, M. S., Marquardt, R. R., Masi, A., Borsa, J., and Frohlich, A. A. (1994). Comparison of toxicity of different mycotoxins to several species of bacteria and yeasts: Use of Bacillus brevis in a disc diffusion assay. J. Food Protect. 57, 48–53. Malaveille, C., Brun, G., and Bartsch, H. (1994). Structure-activity studies in E. coli strains on ochratoxin A (OTA) and its analogues implicate a genotoxic free radical and cytotoxic thiol derivative as reactive metabolites. Mutat. Res. 307, 141–147. Marquardt, R. R., and Frohlich, A. A. (1992). A review of recent advances in understanding ochratoxicosis. J. Anim. Sci. 70, 3968–3988. Meisner, H., and Chan, S. (1974). Ochratoxin A, an inhibitor of mitochondrial transport system. Biochemistry 23, 2795–2800. Moore, J., and Truelove, B. (1970). Ochratoxin A, inhibition of mitochondrial respiration. Science 168, 1102–1103. Mulder, G. J., Scholtens, E., and Meijer, D. K. F. (1981). Collection of metabolites in bile and urine from the rat. In Methods in Enzymology, Detoxification and Drug Metabolism: Conjugation and Related Systems (W. B. Jakoby, Ed.), pp. 21–30. Academic Press, New York. Nobilec, E., Aniol, M., and Warzenczyk, C. (1994). Lactone 1. hydroxylation of dihydro-b-campholenolactone by Fusarium culmorum. Tetrahedron 50, 10339–10344. Nozawa, K., Yamada, M., Tsuda, Y., and Kawai, K. I. (1981). Synthesis of antifungal isocoumarins. II. Synthesis and antifungal activity of 3substituted isocoumarins. Chem. Pharm. Bull. 29, 2491–2495. Omar, R. F., Hasinoff, B. B., Mejilla, F., and Rahimtula, A. D. (1990). Mechanism of ochratoxin A stimulated lipid peroxidation. Biochem. Pharmacol. 40, 1183–1191. Omar, R. F., Rahimtula, A. D., and Bartsch, H. (1991). Role of cytochrome P-450 in ochratoxin A-stimulated lipid peroxidation. J. Biochem. Toxicol. 6, 203–209. Parker, R. W., Phillips, T. D., Kubena, L. F., Russell, L. H., and Heidelbaugh, N. D. (1982). Inhibition of pancreatic carboxypeptidase A: A possible mechanism of interaction between penicillic acid and ochratoxin A food and feed mycotoxins. J. Environ. Sci. Health 17(2), 77–91.

AID

TOX 7747

/

6h0a$$$$85

02-21-96 11:44:10

Pitout, M. J., and Nel, W. (1969). The inhibitory effect of ochratoxin A on bovine carboxypeptidase A in vitro. Biochem. Pharmacol. 18, 1837– 1841. Rahimtula, A. D., Bereziat, J. C., Bussacchini-Griot, V., and Bartsch, H. (1988). Lipid peroxidation as a possible cause of ochratoxin A toxicity. Biochem. Pharmacol. 37, 4469–4477. Roth, A., Creppy, E. E., Kane, A., Bacha, H., Steyn, P. S., Roschenthaler, R., and Dirheimer, G. (1989). Influence of ochratoxin B on the ochratoxin A inhibition of phenylalanyl-tRNA formation in vitro and protein synthesis in hepatoma tissue culture cells. Toxicol. Lett. 45, 307–314. Roth, A., Eriani, G., Dirheimer, G., and Gangloff, J. (1993). Kinetic properties of pure over produced Bacillus subtilis phenylalanyl-tRNA synthetase do not favour its in vivo inhibition by ochratoxin A. FEBS Lett. 326, 87– 91. Saltz, L., Sirott, M., Young, C., Tong, W., Niedzwiecki, D., Tzy-Jyun, Y., Tao, Y., Trochanowski, B., Wright, P., and Barbosa, K. (1993). Phase I clinical and pharmacology study of topotecan given daily for 5 consecutive days to patients with advanced solid tumors, with attempt at dose intensification using recombinant granulocyte colony-stimulating factor. J. Natl. Cancer. Inst. 85, 1499–1507. Sato, T., Nagai, K., Suzuki, K., Morioka, M., Saito, T., Nohara, C., Susaki, K., and Takebayashi, Y. (1992). A new isocoumarin antibiotic, Y05460M-A. J. Antibiot. Tokyo 45, 1949–1952. Terse, P. S., Madhyastha, M. S., Zurovac, O., Stringfellow, D., Marquardt, R. R., and Kemppainen, B. W. (1993). Comparison in vitro and in vivo biological activity of mycotoxins. Toxicon 31, 913–919. Trivedi, A. B., Doi, E., and Kitabatake, N. (1992). Detoxification of ochratoxin A on heating under acidic and alkaline conditions. J. Biotech. Biochem. 56, 741–745. van der Merwe, K. J., Steyn, P. S., Fourie, L., Scoot, D. B., and Theron, J. J. (1965a). Ochratoxin A, a toxic metabolite produced by Aspergillus ochraceus Wilh. Nature 205, 1112–1113. van der Merwe, K. J., Steyn, P. S., and Fourie, L. (1965b). Mycotoxins Part II. The constitution of ochratoxin A, B and C, metabolites of Aspergillus ochraceus Wilh. J. Chem. Soc. 7083–7088. Wei, Y. H., Lu, C. Y., Lin, T. N., and Wei, R. D. (1985). Effect of ochratoxin A on rat liver mitochondrial respiration and oxidative phosphorylation. Toxicology 36, 119–130. Xiao, H., Marquardt, R. R., Frohlich, A. A., and Ling, Y. Z. (1995). Synthesis and structural elucidation of ochratoxin A analogs. J. Agric. Food Chem. 43, 524–530. Xiao, H., Marquardt, R. R., Abramson, D., and Frohlich, A. A. (1996). Metabolites of ochratoxins in rat urine and in a culture of Aspergillus ochraceus. Appl. Environ. Microbiol. 62, in press.

toxa

AP: Tox