Identification of the diastereomers of pentobarbital N-glucosides excreted in human urine.

A study was undertaken to determine if humans excreted pentobarbital N-glucosides as urinary metabolites following oral administration of pentobarbital. (1'RS,5RS)-1-(beta-D-Glucopyranosyl)pentobarbital ((1'RS,5RS)-PTBG) was isolated from the urine of one subject. The two diastereomers, (1'RS,5R)-PTBG and (1'RS,5S)-PTBG were separated and found to be identical to synthetic standards when compared using HPLC retention times coupled with UV (with and without post-column ionization) and mass spectrometry (HPLC/MS). A HPLC method was developed for detecting and quantifying (1'RS,5R)-PTBG, (1'RS,5S)-PTBG and pentobarbital in urine. Following a single oral dose of sodium pentobarbital to male subjects (n = 6), 1.6-6.2% of the pentobarbital dose was excreted as (1'RS,5S)-PTBG over 60 hours. (1'RS,5R)-PTBG was also detected in one subject and accounted for 0.3% of the pentobarbital dose. Using a modified HPLC system, the four pentobarbital N-glucosides were resolved and analysis of a partially purified pentobarbital N-glucoside extract from one subject indicated that only (1'R,5R)-PTBG and (1'S,5S)-PTBG could be detected as urinary excretion products. These results indicate that the side chain chirality of pentobarbital may influence the observed enantioselectivity for the formation and/or urinary excretion of the pentobarbital N-glucosides.

Identification of 5-ethyl-5-(2-methylbutyl)barbituric acid as an impurity of manufacture in amobarbital.

Amobarbital [5-ethyl-5-(3-methylbutyl)barbituric acid], USP, was found to contain an impurity that was not associated with hydrolysis and decomposition of the barbiturate ring. The impurity was isolated by semipreparative HPLC and was identified as 5-ethyl-5-(2-methylbutyl)barbituric acid (1) by MS (electron impact and chemical ionization) and 1H NMR. The substitution pattern on the alkyl side chain was verified by using the achiral NMR shift reagent tris(6,6,7,7,8,8,8-heptafluoro-2,2- dimethyl-3,5-octanedionato)europium(III). Older samples of amobarbital, USP, contained greater than 6% of 1, whereas recent samples of amobarbital, USP, contained less than 1% of 1. Because the pharmacological profiles of 1 and amobarbital in rodents are comparable, the impurity probably does not constitute a clinically significant problem for humans.

Identification of phenobarbital N-glucosides as urinary metabolites of phenobarbital in mice.

Previously, the N-glucosylation of phenobarbital had been observed only in humans. The results of a species screen (mouse, rat, guinea pig, rabbit, cat, dog, pig, and monkey) found that only mice excreted the N-glucosides of phenobarbital in urine after ip administration of sodium phenobarbital. The major diastereomer excreted by the mouse had the R configuration at the C-5 position of the barbiturate ring. The N-glucoside metabolites accounted for a small percentage of the dose (approximately 0.5%). Following ip dosing of the mouse with the phenobarbital N-glucosides, free phenobarbital could be detected in the urine. Upon ip or intercerebroventricular (icv) injection of the phenobarbital N-glucosides, minimal CNS activity was observed in the mouse.

Barbital N-glucoside is not detected as a urinary excretion product of barbital in humans.

A study was undertaken to determine if humans excreted barbital N-glucoside as a urinary metabolite following oral administration of barbital. A liquid chromatography method using gradient elution was developed for detecting and quantifying barbital N-glucoside and barbital in urine. Following a single oral dose of barbital to male caucasian and oriental subjects that had previously been shown to excrete amobarbital and phenobarbital N-glucosides, no barbital N-glucoside conjugate was observed in the urine. This result indicates that N-glucosylation of barbiturates is not a general pathway for the biodisposition of barbiturates in man.

Stereochemical characterization of the diastereomers of the amobarbital N-glucosides excreted in human urine.

The stereochemistry associated with the amobarbital N-glucoside diastereomers (1a and 1b) that are excreted by humans in urine is unknown. Using X-ray crystallography, the absolute configuration of 1b was determined to be S (C-5 position of the barbiturate ring). Following oral administration of amobarbital to Caucasians and Orientals, from 5 to 25% of the dose of amobarbital was excreted in the urine as 1b. The other diastereomer, 1a, accounted for less than 0.1 to 0.2% of the dose in four individuals, with none detected in nine individuals. The rate constants, kf,1b, determined from the urinary excretion of 1b were lower than those previously reported for unresolved amobarbital N-glucosides. However, based on the urinary excretion of 1b, the rate constants, K, for elimination of amobarbital in Caucasians and Orientals were similar to those previously determined from the serum levels of amobarbital and the urinary excretion of unresolved amobarbital N-glucosides. In previous studies of the N-glucosylation of amobarbital, it is likely that a single N-glucose diastereomer, 1b, was being observed.

Stereochemical characterization of the diastereomers of the phenobarbital N-beta-D-glucose conjugate excreted in human urine.

The absolute configuration of the N-beta-D-glucoside metabolites of phenobarbital was determined by methylation of the diastereomers to make mephobarbital N-beta-D-glucosides, followed by oxidative removal of glucose to give the optical isomers of mephobarbital. Following a single oral dose of phenobarbital to two male subjects, both phenobarbital N-beta-D-glucosides were excreted in the urine. The absolute configuration (C-5 position) of the major phenobarbital N-beta-D-glucoside excreted in the urine was the S form. A pronounced stereoselective formation and/or urinary excretion occurs for the N-glucoside conjugates of phenobarbital in humans.

LC determination of the diastereomers of 1-(beta-D-glucopyranosyl)phenobarbital in human urine.

The "product enantioselectivity" associated with the urinary excretion of the phenobarbital N-glucoside conjugates has not been determined previously. A liquid chromatography method using gradient elution was developed for quantifying both phenobarbital N-glucoside conjugates, phenobarbital, and p-hydroxyphenobarbital. Following a single oral dose of phenobarbital to male Caucasian and Oriental subjects, both phenobarbital N-glucoside conjugates were observed in the urine. In seven subjects, 3.3-10.6% of the phenobarbital dose was detected as a single phenobarbital N-glucoside (S configuration at the C-5 position of the barbiturate ring). The other phenobarbital N-glucoside diastereomer accounted for less than 1.5% of the phenobarbital dose. The urinary excretion of the major phenobarbital N-glucoside diastereomer paralleled the urinary excretion of phenobarbital and was comparable in both Caucasian and Oriental subjects. These results indicate a pronounced selectivity for the formation and/or urinary excretion of the phenobarbital N-glucosides.

Synthesis of N-beta-D-glucopyranosyl derivatives of barbital, phenobarbital, metharbital, and mephobarbital.

The condensation of per(trimethyl)silylbarbital and -phenobarbital with 1,2,3,4,6-penta-O-acetyl-beta-D-glucopyranose in the presence of stannic chloride in dichloroethane gave moderate yields of the beta-coupled barbiturate N-D-glucopyranosyl derivatives. Reaction of metharbital and mephobarbital under the same conditions was unsuccessful. The homologous N-methylglucosides were prepared by reaction of the barbital and phenobarbital N-glucosyl derivatives with diazomethane. The diastereomers of the phenobarbital and mephobarbital derivatives were resolved by use of C-18 reverse-phase h.p.l.c. 1H- and 13C-n.m.r. spectroscopy, and thermospray 1.c.-m.s. proved to be the most useful methods for characterizing the barbiturate glucosides.

Stability of phenobarbital N-glucosides: identification of hydrolysis products and kinetics of decomposition.

The two diastereomers of 1-(1-beta-D-glucopyranosyl)phenobarbital, (1A) and (1B), decompose to 1-(1-beta-D-glucopyranosyl)-3-(2-ethyl-2-phenylmalonyl)urea (2A or 2B) followed by decarboxylation to 1-(1-beta-D-glucopyranosyl)-3-(2-phenylbutyryl)urea (3A and 3B) under physiological conditions of temperature and pH. The sigmoidal pH-rate profile and the Arrhenius parameters indicate that degradation takes place by hydroxide ion attack on the undissociated and monoanion forms of 1A and 1B. The rates of hydrolysis of the nonionized species of 1A and 1B are more than two orders of magnitude faster than those of common 5,5-disubstituted or 1,5,5-trisubstituted barbiturates. Molecular modeling studies suggest that rate enhancement is due to intramolecular hydrogen bonding in the transition state of the C2' hydroxyl with the tetrahedral hydrated C6 carbonyl as well as hindered rotation around the N1-C1' of phenobarbital and glucose. Based on these studies it is recommended that any data related to the quantitation of 1A and 1B be reevaluated depending on how the samples were collected, stored, and analyzed.

Isolation of chlordecone binding proteins from pig liver cytosol.

Cytosolic proteins may play an important role in the transport of water insoluble substances through the cytosol to various subcellular locations. The binding of chlordecone (CD) to pig liver cytosolic proteins was studied after the simultaneous administration of [14C]CD and [3H]cholesterol via the portal vein. The isolation of chlordecone binding proteins (CDBPs), from liver cytosol consisted of repeated ultracentrifugation, ammonium sulfate fractionation, and chromatography on Bio-Gel A 0.5m, carboxymethyl cellulose (CMC), and Sephadex G-100. Three proteins retained [14C]CD after elution from the CMC column. CDBP I and CDBP II also retained [3H]cholesterol through this stage of purification, while CDBP III did not. The molecular weights, estimated by Sephadex G-100 gel filtration, were 33,500 and 67,000 for CDBP IA and CDBP IB, 49,000 for CDBP II, and 44,000 for CDBP III. Since dissociation of bound cholesterol could not be avoided during the purification procedures, binding of cholesterol to the CDBPs eluted from Sephadex G-100 was investigated. Incubation of CDBPs with [3H]cholesterol resulted in a 2000-fold increase of 3H associated with CDBP II, an 800-fold increase with CDBP IA, a 100-fold increase for CDBP IB, and a 300-fold increase for CDBP III. The isolation characteristics, molecular weights, and cholesterol binding properties of the CDBPs are compared with cytosolic cholesterol binding proteins previously isolated. The high specific activity binding of both CD and cholesterol by CDBP I and CDBP II suggests that CD and cholesterol share a common transport pathway in the liver cytosol.

High-density lipoproteins decrease the biliary concentration of chlordecone in isolated perfused pig liver.

Chlordecone (CD) is an organochlorine pesticide associated with albumin and high-density lipoproteins (HDL) in the plasma. It is found in higher concentrations in the liver than in other tissues and is excreted in the bile. The influence of plasma HDL on the biliary excretion of CD was studied using isolated pig liver perfused with a Krebs-Ringer bicarbonate solution containing albumin, dextrose, and pig red blood cells. Within 5 min after administration into the perfusion medium of [14C]CD bound to albumin or to HDL, only 13% of the [14C]CD dose remained in the perfusate, showing that CD is rapidly taken up by the liver. After 60 min the plasma concentration was constant at 0.008% dose/ml when [14C]CD was administered bound to albumin in the absence of HDL and at 0.004% dose/ml when administered bound to HDL. The mean concentration of CD in the bile was higher when CD was administered bound to albumin in the absence of HDL (0.039% dose/ml) than when it was administered bound to HDL (0.010% dose/ml). The elimination rate constant of CD from the liver into the bile was 0.007/min whether CD was administered bound to albumin or to HDL. The addition of HDL to the perfusion system after the administration of albumin-bound CD resulted in lower biliary CD concentrations. The results suggest that HDL affects the distribution of CD between the perfusate and liver and between liver and bile. In both cases, distribution toward the liver is favored.

Chlordecone metabolism in the pig.

The biotransformation of chlordecone (CD) to chlordecone alcohol (CDOH) occurs in man and gerbils but not in rats, guinea pigs and hamsters [1, 2]. Because of the species differences in CDOH formation and the need for a suitable animal model, pigs were administered CD by intraperitoneal (i.p.) injection. Plasma, gallbladder bile, hepatic bile, liver and feces were collected and analyzed by gas chromatography for CD metabolites. CDOH was present in bile and feces with up to 85% conjugated in the bile but only 15% was conjugated in the feces. Up to 20% of the CD in plasma and bile and less than 3% in feces was in the conjugated form. Both reduction and conjugation of CD in the pig are similar to those in man.

Preferential binding of chlordecone to the protein and high density lipoprotein fractions of plasma from humans and other species.

The preferential distribution of the relatively nonpolar pesticide chlordecone (CD) to liver rather than to fat tissues in humans suggests that it may be transported in plasma differently from other organochlorine pesticides. The plasma binding of [14C] CD was investigated in vitro in human, rat, and pig plasma and in vivo in rat plasma. Protein and lipoprotein fractions were separated by serial ultracentrifugation. Heparin-manganese precipitation and agarose gel electrophoresis were also carried out to determine whether separation techniques altered CD binding to plasma components. In human plasma, the distribution of [14C] CD among proteins and high density, low density, and very low density lipoproteins (HDL, LDL, and VLDL) was 46, 30, 20, and 6%, respectively. The distribution of cholesterol in the same plasma fractions was 4, 20, 63, and 7%, respectively. In the pig and rat the order of binding was similar to that in humans, with protein greater than or equal to HDL greater than LDL greater than or equal to VLDL. Separation by heparin-Mn precipitation confirmed the results obtained by ultracentrifugation. The distribution of [14C] CD in rat lipoprotein was similar whether the CD was administered in vivo or incubated with plasma in vitro, with approximately 80% bound to HDL, 11% to LDL, and 9% to VLDL in either case. Agarose gel electrophoresis of plasma-bound [14C] CD indicated that albumin was the major component of the protein fraction responsible for CD binding. Preferential binding of CD by albumin and HDL may explain its unusual tissue distribution compared to other organochlorine pesticides such as aldrin and dieldrin, which bind preferentially to VLDL and LDL and distribute preferentially to fat tissues.