Albumin, a New Biomarker of Organophosphorus Toxicant Exposure, Identified by Mass Spectrometry
http://www.100md.com
《毒物学科学杂志》
University of Nebraska Medical Center, Eppley Institute, Omaha, Nebraska 68198–6805
University of Montana, Department of Biomedical and Pharmaceutical Sciences, Missoula, Montana 59812
ABSTRACT
The classical laboratory tests for exposure to organophosphorus toxicants (OP) are inhibition of acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) activity in blood. In a search for new biomarkers of OP exposure, we treated mice with a biotinylated organophosphorus agent, FP-biotin. The biotinylated proteins in muscle were purified by binding to avidin-Sepharose, separated by gel electrophoresis, digested with trypsin, and identified from their fragmentation patterns on a quadrupole time-of-flight mass spectrometer. Albumin and ES1 carboxylesterase (EC 3.1.1.1) were found to be major targets of FP-biotin. These FP-biotinylated proteins were also identified in mouse plasma by comparing band patterns on nondenaturing gels stained for albumin and carboxylesterase activity, with band patterns on blots hybridized with Streptavidin Alexa-680. Two additional FP-biotin targets, AChE (EC 3.1.1.7) and BChE (EC 3.1.1.8), were identified in mouse plasma by finding that enzyme activity was inhibited 50–80%. Mouse plasma contained eight additional FP-biotinylated bands whose identity has not yet been determined. In vitro experiments with human plasma showed that chlorpyrifos oxon, echothiophate, malaoxon, paraoxon, methyl paraoxon, diazoxon, diisopropylfluorophosphate, and dichlorvos competed with FP-biotin for binding to human albumin. Though experiments with purified albumin have previously shown that albumin covalently binds OP, this is the first report of OP binding to albumin in a living animal. Carboxylesterase is not a biomarker in man because humans have no carboxylesterase in blood. It is concluded that OP bound to albumin could serve as a new biomarker of OP exposure in man.
Key Words: acetylcholinesterase; butyrylcholinesterase; organophosphate; FP-biotin; albumin.
INTRODUCTION
Organophosphorus toxicants (OP) are used in agriculture as pesticides, in medical practice as antihelminthics, in the airline industry as additives to hydraulic fluid and jet engine oil, and as chemical warfare agents. These compounds are known to exert their acute effects by inhibiting acetylcholinesterase (EC 3.1.1.7, AChE). The excess acetylcholine that accumulates causes an imbalance in the nervous system that can result in death (McDonough and Shih, 1997).
Though AChE is the clinically important target of OP exposure, other proteins also form a covalent bond with OP, depending on the identity of the OP (Casida and Quistad, 2004). These secondary targets include butyrylcholinesterase, acylpeptide hydrolase, neurotoxic esterase, fatty acid amide hydrolase, arylformamidase, cannabinoid CB1 receptor, muscarinic acetylcholine receptor, and carboxylesterase. With the exception of neurotoxic esterase, whose inhibition is responsible for delayed neuropathy, the toxicological relevance of inhibition of these secondary targets is not yet understood (Casida and Quistad, 2004; Ray and Richards, 2001). Albumin has not previously been shown to bind OP in a living animal, though in vitro experiments with purified albumin have demonstrated covalent binding to diisopropylfluorophosphate, sarin, and soman (Black et al., 1999; Means and Wu, 1979; Murachi, 1963; Sanger, 1963; Schwartz, 1982). Albumin hydrolyzes chlorpyrifos oxon and paraoxon (Ortigoza-Ferado et al., 1984; Sultatos et al., 1984).
The toxic effects of a particular OP cannot be attributed entirely to inhibition of AChE. Toxic signs are different for each OP when that OP is administered at low doses (Moser, 1995). For example, a low dose of fenthion decreased motor activity in rats by 86% but did not alter the tail-pinch response, whereas a low dose of parathion did not affect motor activity but did decrease the tail-pinch response (Moser, 1995). These toxicological observations suggest that OP have other biological actions in addition to their cholinesterase-inhibitory properties.
Another confounding observation is the finding that toxic signs do not correlate with degree of AChE inhibition. Rats given doses of OP that inhibited AChE to similar levels had more severe toxicity from parathion than chlorpyrifos (Pope, 1999). There are also examples of toxic signs unaccompanied by AChE inhibition. Workers who manufacture the OP pesticide quinalphos have significantly low scores for memory, learning ability, vigilance, and motor response, though their blood AChE activity levels are the same as in control subjects (Srivastava et al., 2000). Chronic low-level exposure to OP induces neuropsychiatric disorders without inhibition of esterase activity (Ray and Richards, 2001; Salvi et al., 2003). These observations have led to the suggestion that some OP have toxicologically relevant sites of action in addition to AChE (Moser, 1995; Pope, 1999; Ray and Richards, 2001). The hypothesis arose that a given OP reacts not only with AChE, but with a set of proteins unique for each OP.
Our goal is to identify the proteins in a living animal that covalently bind the biotin-tagged OP called FP-biotin (Kidd et al., 2001; Liu et al., 1999). In this report we used tandem mass spectrometry, enzyme activity assays, gel electrophoresis, and blots to identify four FP-biotin-labeled proteins in the muscle and plasma of mice that had been injected with FP-biotin ip. We found FP-biotin-labeled albumin, carboxylesterase, BChE, and AChE. This is the first report to demonstrate that albumin is a significant target of OP binding in a living animal. Eight other proteins in mouse blood became labeled but have not yet been identified.
MATERIALS AND METHODS
Materials. FP-biotin (MW 592.3) was custom synthesized by Troy Voelker in the laboratory of Charles M. Thompson at the University of Montana (Liu et al., 1999). Purity was checked by NMR and mass spectrometry, and no evidence of contamination was detected. FP-biotin was stored as a dry powder at –70°C. Just before use, the dry powder was dissolved in 100% ethanol to a concentration of 13.3 mg/ml and diluted with saline to 15% ethanol containing 2 mg/ml FP-biotin.
Immun-Blot PVDF membrane for protein blotting, 0.2 μm (catalog #162-0177) and biotinylated molecular weight markers (catalog #161–0319) were from Bio-Rad Laboratories, Hercules, CA. Streptavidin Alexa-680 fluorophore (catalog #S-21378) was from Molecular Probes, Eugene, OR. Avidin-agarose beads (catalog # A-9207), iso-OMPA, and bovine albumin, essentially fatty acid-free (Sigma A 7511) were from Sigma-Aldrich, St. Louis, MO. Echothiophate iodide was from Wyeth-Ayerst, Rouses Point, NY. All other OP were from Chem Service Inc, West Chester, PA.
Mice. The Institutional Animal Care and Use Committee of the University of Nebraska Medical Center approved all procedures involving mice. Animal care was provided in accordance with the principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Mice completely lacking AChE protein were made by gene targeting (Xie et al., 2000) at the University of Nebraska Medical Center. Exons 2, 3, 4, and 5 of the ACHE gene were deleted to make AChE–/– mice. The AChE–/– animals are in strain 129Sv genetic background. The colony is maintained by breeding heterozygotes because AChE–/– mice do not breed (Duysen et al., 2002). Wild-type mice are littermates of AChE–/– mice. The strain 129Sv mice were used for experiments with FP-biotin.
Injection of FP-biotin into mice. Mice were injected intraperitoneally with FP-biotin dissolved in 15% ethanol to give a dose of 56 or 5 mg/kg, or with vehicle alone. The dose of FP-biotin was calculated from dry weight. No correction was made for the fact that FP-biotin is a mixture of phosphorus stereoisomers. Mice were euthanized 120 min after FP-biotin injection, by inhalation of carbon dioxide. Blood was washed out of tissues by intracardial perfusion with saline solution. Tissues from six mice were analyzed by mass spectrometry: two AChE–/– FP-biotin treated (5 and 56 mg/kg), two AChE–/– untreated, one wild-type FP-biotin treated (56 mg/kg), and one wild-type untreated. In addition wild-type mice were treated with 0, 0.5, 1.0, 5.0, or 18.8 mg/kg FP-biotin (n = 2 for each dose). Blood was analyzed by gel electrophoresis, blotting, and enzyme activity assays.
Enzyme activity. AChE activity was measured with 1 mM acetylthiocholine in the presence of 0.5 mM dithiobisnitrobenzoic acid (Ellman et al., 1961) after inhibiting BChE activity for 30 min with 0.1 mM iso-OMPA, in 0.1 M potassium phosphate pH 7.0, at 25°C. BChE activity was measured with 1 mM butyrylthiocholine. E412nm = 13,600 M–1 cm–1 at pH 7.0. Carboxylesterase activity was measured with 5 mM p-nitrophenyl acetate after inhibiting AChE and BChE with 0.01 mM eserine, and after inhibiting paraoxonase with 12.5 mM EDTA. E400nm = 9000 M–1 cm–1 at pH 7.0. AChE, BChE, and carboxylesterase units of activity are micromoles hydrolyzed per minute at pH 7.0, 25°C.
Isolation of FP-biotin-labeled protein. To prepare OP-labeled proteins for mass spectrometry, FP-biotin-labeled proteins in muscle were purified on avidin-agarose beads and separated by SDS–polyacrylamide gel electrophoresis. Proteins from mice that had
Tissues were homogenized in 10 volumes of 50 mM TrisCl pH 8.0 containing 5 mM EDTA, and centrifuged for 10 min in a microfuge at 12,000 rpm to partially clarify the suspension. A detailed example of the protocol follows. The 0.96 ml of muscle homogenate (7.4 mg protein/ml) was diluted with 3.75 ml of 50 mM TrisCl pH 8.0, 5 mM EDTA to make 1.5 mg/ml protein solution. SDS was added to make the solution 0.5% SDS. The protein solution was heated for 3 min in a boiling water bath and then diluted with buffer to make the final SDS concentration 0.2%. The protein solution was incubated with 100 μl of washed avidin-agarose beads (1.9 mg avidin/ml of beads) overnight at room temperature, with continuous inversion, to bind the FP-biotin-labeled proteins to the beads. Beads were washed three times with the TrisCl/EDTA buffer, containing 0.2% SDS, to remove nonspecifically bound protein. Twenty-five μl of 6x SDS–PAGE loading buffer (0.2 M TrisCl, pH 6.8, 10% SDS, 30% glycerol, 0.6 M dithiothreitol and 0.012% bromophenol blue) were added to the 100 μl of beads, and the mixture was heated at 85°C for 3 min. This step released the biotinylated proteins from the avidin beads. Equal amounts of the bead mixture were loaded directly into two wells of a 10–20% gradient SDS–PAGE (10-well format, 1.5-mm thick) and run for 4000 volt-hours in the cold room. The gel was stained with Coomassie blue G250 (Bio-Safe from BioRad), and destained with water. Coomassie G250 is reportedly 2–8 fold more sensitive than Coomassie R250 (BioRad specifications). To minimize contamination from keratin, the staining dish had been cleaned with sulfuric acid, and the water was Milli-Q purified. For the same reason, gloves were worn for all operations involving the gel.
Protein digestion protocol. The proteins separated on SDS–PAGE were digested with trypsin to prepare them for identification by mass spectral analysis. Gloves were worn throughout these procedures, all solutions were made with Milli-Q purified water, and all glassware, plasticware, and tools were rinsed with Milli-Q purified water to minimize keratin contamination. Each Coomassie-stained band from one lane of the SDS–PAGE was excised, placed into a separate 1.5-ml microfuge tube, and chopped into bits. The amount of gel excised was kept to a minimum. The gel bits were destained by washing with 200 μl of 25 mM ammonium bicarbonate (Aldrich) in 50% acetonitrile (synthesis grade, from Fisher). After three washes, the gel bits were colorless and had shrunken considerably. Residual liquid was removed, and the gel bits dried by evaporation in a Speedvac (Jouan). Disulfide bonds in the protein were reduced by incubating the gel bits with 10 mM dithiothreitol (molecular biology grade, from Sigma) in 200 μl of 100 mM ammonium bicarbonate for 1 h at 56°C. The gel pieces were then centrifuged, excess solution was removed, and the protein was alkylated with 55 mM iodoacetamide (Sigma) in 120 μl of 100 mM ammonium bicarbonate for 1 h at room temperature in the dark. The gel bits were again centrifuged, excess solution was removed, and the bits were washed with 200 μl of 25 mM ammonium bicarbonate in 50% acetonitrile (three times). Residual liquid was again removed and the gel bits dried by evaporation in the Speedvac. The proteins were digested in the gel with trypsin, using 12.5 ng/μl of sequencing grade trypsin (Promega) in 25 mM ammonium bicarbonate. Ninety μl of the trypsin solution were added to the dry gel bits and incubated at 4°C for 20 min, to allow the gel to re-swell. Then 60 μl of 25 mM ammonium bicarbonate were layered over each sample, and the samples were incubated at 37°C overnight (about 17 h). Peptides were extracted by incubating each reaction mixture with 200 μl of 0.1% trifluoroacetic acid (sequencing grade from Beckman) in 60% acetonitrile for one h at room temperature. Extraction was repeated three times, and the extracts for each sample were pooled. The pooled extracts were evaporated to dryness in the Speedvac, and the dry samples were stored at –20°C until analyzed.
Mass spectral analysis. Each tryptic peptide digest was resuspended in 40 μl of 5% aqueous acetonitrile/0.05% trifluoroacetic acid. A 10-μl aliquot of the digest was injected into a CapLC (capillary liquid chromatography system from Waters Corp) using 5% aqueous acetonitrile/0.05% trifluoroacetic acid (auxiliary solvent) at a flow rate of 20 μl per min. Peptides were concentrated on a C18 PepMapTM Nano-PrecolumnTM (5 mm x 0.3 mm id, 5μm particle size) for 3 min, and then eluted onto a C18 PepMapTM capillary column (15 cm x 75 μm id, 3 μm particle size both from LC Packings), using a flow rate of 200–300 nl per min. Peptides were partially resolved using gradient elution. The solvents were 2% aqueous acetonitrile/0.1% formic acid (solvent A), and 90% acetonitrile/10% isopropanol/0.2% formic acid (solvent B). The solvent gradient increased from 5% B to 50% B over 22 min, then to 80% B over 1 min, and remained at 80% B for 4 min. The column was then flushed with 95% B for 3 min and equilibrated at 5% B for 3 min before the next sample injection.
Peptides were delivered to the Z-spray source (nano-sprayer) of a Micromass Q-TOF (tandem quadrupole/time-of-flight mass spectrometer from Waters Corp.) through a 75-μm id capillary, which connected to the CapLC column. In order to ionize the peptides, 3300 volts were applied to the capillary, 30 volts to the sample cone, and zero volts to the extraction cone. Mass spectra for the ionized peptides were acquired throughout the chromatographic run, and collision-induced dissociation spectra were acquired on the most abundant peptide ions (having a charge state of 2+, 3+, or 4+). The collision-induced dissociation spectrum is unique for each peptide and is based on the amino acid sequence of that peptide. For this reason, identification of proteins using collision-induced dissociation data is superior to identification by only the peptide mass fingerprint of the protein. The collision cell was pressurized with 1.5 psi ultrapure argon (99.999%), and collision voltages were dependent on the mass-to-charge ratio and the charge state of the parent ion. The time-of-flight measurements were calibrated daily using fragment ions from collision-induced dissociation of [Glu1]-fibrinopeptide B. Each sample was post-processed using this calibration and Mass Measure (Micromass). The calibration was adjusted to the exact mass of the autolytic tryptic fragment at 421.76, found in each sample.
The mass and sequence information for each detected peptide was submitted either to ProteinLynx Global Server 1.1 (a proprietary software package, from Micromass), or to MASCOT (a public access package provided by Matrix Science at http://www.matrix-science.com). Data were compared to all mammalian entries (ProteinLynx) or just mouse entries (MASCOT) in the NCBInr database (National Center for Biotechnology Information). Search criteria for ProteinLynx were set to a mass accuracy of 0.25 Da, fixed modification of methionine (oxidation), and variable modification of cysteine (carbamidomethylation). One missed cleavage by trypsin was allowed. Search criteria for MASCOT were set to a mass accuracy of ±0.1 Da, one missed cleavage, variable modification of methionine (oxidation) and cysteine (carbamidomethylation), and peptide charge +2 and +3. Both software packages calculated a score for each identified protein based on the match between the experimental peptide mass and the theoretical peptide mass, as well as between the experimental collision-induced dissociation spectra and the theoretical fragment ions from each peptide. Results were essentially the same from both packages.
Polyacrylamide gel electrophoresis. Gradient polyacrylamide gels (4–30%) were cast in a Hoefer gel apparatus. Electrophoresis was for 5000 volt-hours (200 volts for 25 h) at 4°C for nondenaturing gels and 2500 volt-hours (100 volts for 25 h) at 4°C for gels containing 0.1% SDS.
Staining gels for BChE activity. Nondenaturing gels were stained for BChE activity by the method of Karnovsky and Roots (1964). The staining solution contained 180 ml of 0.2 M sodium maleate pH 6.0, 15 ml of 0.1 M sodium citrate, 30 ml of 0.03 M cupric sulfate, 30 ml of 5 mM potassium ferricyanide, and 0.18 g butyrylthiocholine iodide in a total volume of 300 ml. Gels were incubated, with shaking, at room temperature for 3 to 5 h. The reaction was stopped by washing the gels with water. To determine the location of albumin, activity-stained gels were stained with Coomassie blue.
Staining gels for carboxylesterase activity and albumin. Nondenaturing gels were incubated in 100 ml of 50 mM TrisCl pH 7.4 in the presence of 50 mg beta-naphthylacetate dissolved in 1 ml ethanol, and 50 mg of solid Fast Blue RR. The naphthylacetate precipitates when it is added to the buffer, but enough remains in solution that the reaction works. Though the Fast Blue RR does not dissolve, pink to purple bands develop on the gel within minutes (Nachlas and Seligman, 1949). A maximum of 30 min incubation at room temperature was needed. The gels were washed with water and photographed. This stain is primarily for carboxylesterase. Albumin gives a faint band with this method because albumin slowly hydrolyzes beta-naphthylacetate (Tove, 1962). Activity-stained gels were counterstained with Coomassie blue to verify the location of albumin.
To align bands on gels stained for enzyme activity with biotinylated bands on a PVDF membrane, the transparent activity-stained gels were placed on top of a printed image of the fluorescent bands in the PVDF membrane.
Visualizing FP-biotin-labeled proteins. For determination of the number and size of proteins labeled by FP-biotin, proteins were subjected to gel electrophoresis, transfer to a PVDF membrane, and hybridization with a fluorescent probe. The details of the procedure follow.
Proteins were transferred from the polyacrylamide gel to PVDF membrane (Immun-Blot from BioRad) electrophoretically in a tank using plate electrodes (TransBlot from BioRad), at 0.5 amps, for 1 h, in 3 l of 25 mM Tris/192 mM glycine buffer, pH 8.2, in the cold room (4°C), with stirring. The membrane was blocked with 3% gelatin (BioRad) in 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl for 1 h at room temperature. The 3% gelatin solution had been prepared by heating the gelatin in buffer in a microwave oven for several seconds. The blocked membrane was washed three times with 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl and 0.05% Tween-20, for 5 min.
Biotinylated proteins were labeled with 9.5 nM Streptavidin Alexa-680 fluorophore in 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl, 0.05% Tween-20, 0.2% SDS, and 1% gelatin, for 2 h, at room temperature, protected from light. Shorter reaction times resulted in less labeling. The SDS was found to increase the intensity of labeling. The membrane was washed twice with 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl and 0.05% Tween-20, and twice with 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl, for 20 min each, while protected from light.
Membranes were scanned with the Odyssey Infrared Imaging System (LI-COR, Lincoln, NE) at 42 microns per pixel. The Odyssey employs an infrared laser to excite a fluorescent probe, which is attached to the target protein, and then collects the emitted light. The emitted light intensity is directly proportional to the amount of probe. Both the laser and the detector are mounted on a moving carriage positioned directly below the membrane. The membrane can be scanned in step sizes as small as 21 microns, providing resolution comparable to X-ray film. Data are collected using a 16-bit dynamic range. The fluorophore is stable in the laser, making it possible to scan the membrane repeatedly, while using different intensity settings to optimize data collection for both strong and weak signals. The membrane was kept wet during scanning.
Biotinylated protein standards. The biotinylated bovine serum albumin (BSA) standard was prepared by incubating 10 μM BSA (0.5 mg/ml) with 20 μM FP-biotin in 20 mM TrisCl pH 7.4 at room temperature for 16 h. The biotinylated BChE standard was prepared by incubating 50 nM human BChE (3 units/ml; 4.2 μg/ml) with 10 μM FP-biotin, in 20 mM TrisCl pH 7.4 at room temperature for 16 h.
Amount of biotinylated albumin in mouse plasma. The percentage of FP-biotinylated albumin in mouse plasma was estimated from the relative intensities of the biotinylated albumin band and the biotinylated BChE band on a blot stained with Streptavidin Alexa-680. The concentration of BChE in mouse plasma is 0.003 mg/ml. The concentration of biotinylated BChE was calculated from the reduction in enzyme activity. The concentration of albumin in mouse plasma is 50 mg/ml. These values allowed estimation of percent biotinylated albumin in plasma. For example, when BChE activity was inhibited 35%, the biotinylated BChE band represented about 0.001 mg/ml biotinylated BChE. A biotinylated albumin band of similar intensity would contain 0.001 mg/ml biotinylated albumin. When mouse plasma had to be diluted 1000-fold to reduce the intensity of biotinylated albumin to a similar intensity as the biotinylated BChE in undiluted plasma, it was calculated that the concentration of biotinylated albumin in undiluted plasma was 1 mg/ml.
Binding various OP to human albumin. Human plasma was diluted 1:100 to reduce the albumin concentration to 10 μM. The diluted plasma was reacted in vitro with 10 mM malaoxon, paraoxon, chlorpyrifos oxon, methyl paraoxon, dichlorvos, diisopropylfluorophosphate diazoxon, echothiophate, or iso-OMPA for 1 h in 20 mM TrisCl pH 7.5 at 25°C. Then FP-biotin was added to 10 μM and allowed to react for 1 h, and 10 μl containing the equivalent of 0.1 μl plasma was loaded per lane on a nondenaturing gel. Biotinylated proteins were visualized with Streptavidin Alexa-680 after transfer to PVDF membrane.
Statistical analysis. Samples were analyzed by independent samples t-test assuming equal variances. Probability values less than 0.05 were considered significant.
RESULTS
Toxicity of FP-Biotin
The structure of FP-biotin is given in Figure 1. A dose of 56 mg/kg FP-biotin ip was lethal to AChE–/– mice and reduced plasma BChE activity from 1.9 ± 0.4 units/ml in untreated animals to 0.006 ± 0.003 units/ml post treatment, a 99.7% inhibition. A dose of 18.8 mg/kg ip was not lethal, but did cause severe cholinergic signs of toxicity. This dose reduced plasma BChE activity to 0.34 ± 0.09 units/ml, an 82% inhibition. A dose of 5 mg/kg caused only mild signs of toxicity and inhibited plasma BChE of AChE–/– mice 37%, to 1.2 ± 0.4 units/ml. In contrast, wild-type mice showed no signs of toxicity after treatment with 18.8 or 5 mg/kg FP-biotin ip, even though their plasma BChE activity was inhibited to the same extent as in the AChE–/– mice.
Plasma AChE activity in wild-type mice treated with 18.8 mg/kg FP-biotin was inhibited from a predose level of 0.30 ± 0.01 units/ml to 0.13 units/ml, a 56% inhibition. There was no inhibition at lower doses. AChE activity was not measured in AChE–/– mice, because these knockout animals have no AChE activity (Xie et al., 2000). AChE has a 10-fold lower affinity for FP-biotin compared to BChE (Schopfer, unpublished). This explains why a given dose of FP-biotin caused less inhibition of AChE than of BChE.
Plasma carboxylesterase activity was inhibited to the same extent in AChE–/– and +/+ mice. A dose of 18.8 mg/kg FP-biotin caused a reduction from the untreated values of 18.6 ± 0.6 units/ml to 3.5 units/ml, an 81% inhibition, while a dose of 5 mg/kg reduced the carboxylesterase levels to 9.2 ± 1.6 units/ml, a 50% reduction.
Fp-Biotin Does Not Cross the Blood–Brain Barrier
FP-biotin treated AChE–/– mice showed no inhibition of BChE in brain, supporting the conclusion that FP-biotin does not cross the blood–brain barrier.
Identification of FP-Biotinylated Proteins by Mass Spectrometry
Muscle proteins from mice that had been treated with FP-biotin, as well as from untreated control mice, were isolated by binding to avidin beads. The proteins were released from avidin by boiling in SDS and separated by SDS gel electrophoresis. Protein bands visible with Coomassie blue staining were excised and digested with trypsin. Fragmentation of tryptic peptides yielded amino acid sequence information characteristic of the protein. The proteins listed in Table 1 were consistently identified in three separate experiments. The probability score for correct identification (MOWSE score) was extremely high at 1157 for albumin and 655 for ES1 carboxylesterase. A MOWSE score of 69 is significant (p < 0.05), so scores of 1157 and 655 show complete confidence. Though albumin and ES1 carboxylesterase were found in samples prepared from muscle, these proteins are typically expressed at high levels in blood (Kadner et al., 1992; Peters, 1996). It is likely that they were transported from the blood into the extravascular fluid where they were not washed out by perfusion. Albumin was identified by 17 peptides. A representative mass spectrum of an albumin peptide is shown in Figure 2. These 17 peptides represented 51% of the mouse albumin sequence, leaving no doubt that albumin was labeled by FP-biotin. The untreated control tissues did not show albumin, thus demonstrating that only biotinylated albumin had bound to avidin beads. This control experiment eliminated the possibility that the albumin might have bound nonspecifically to the avidin beads or that albumin was endogenously biotinylated.
FP-biotinylated AChE and BChE were not found because these proteins are not abundant enough to give a Coomassie blue stained band on SDS gels. In this work only proteins that gave a Coomassie stained band were analyzed by mass spectrometry.
Identification of Endogenous Biotinylated Proteins
Endogenous biotinylated proteins were identified by mass spectrometry. The proteins in Table 2 were found in muscle of untreated as well as in FP-biotin treated mice. They are propionyl CoA carboxylase alpha, pyruvate carboxylase, and methylcrotonyl CoA carboxylase alpha. These endogenous biotinylated proteins bound to avidin beads and were abundant enough to be visualized as Coomassie blue bands on an SDS gel.
Blot Showing FP-Biotinylated Proteins in Mouse Plasma
A blot showing the proteins in plasma that became labeled with FP-biotin after treatment of mice with FP-biotin, is shown in Figure 3. Doses of 1 and 5 mg/kg FP-biotin were not toxic to the animals. The intense broad band in the middle of the gel resolved into three bands upon serial dilution of mouse plasma. The top band in the triplet is carboxylesterase (CE), the middle band is albumin, and the bottom band has not been identified.
A band for biotinylated BChE was visible in plasma from mice treated with 5 mg/kg but not 1 mg/kg FP-biotin. This is consistent with our finding that BChE activity was inhibited about 35% after treatment with 5 mg/kg but was not inhibited after treatment with 1 mg/kg FP-biotin.
In addition to albumin, carboxylesterase, and butyrylcholinesterase, mouse plasma contains about eight other biotinylated bands whose protein identity is unknown. The intensity of the band at the top of the gel is higher than that of mouse BChE, suggesting that this protein is more abundant than BChE and that it is highly reactive. The intensity of the other bands is equal to that of BChE or lower. Thus, mouse plasma contains at least 11 proteins that bind OP at physiological conditions, at doses of OP that produce no toxic signs.
About 1–2% of the albumin in mouse plasma is estimated to have bound FP-biotin in a mouse treated with 5 mg/kg FP-biotin ip. However, there is so much more albumin (50 mg/ml) than BChE (0.003 mg/ml) in mouse plasma that albumin consumes a significant amount of FP-biotin. Similar biotinylated band intensities were obtained in Figure 3 for BChE in undiluted plasma and for albumin diluted 1000-fold, suggesting that albumin bound 1000-fold more FP-biotin than was bound by BChE.
Activity-Stained Gels
The gels for Figures 3–5 were nondenaturing polyacrylamide gels. Nondenaturing gels were used because under these conditions the BChE tetramer of 340 kDa separated well from the 67 kDa albumin. This separation is not possible on SDS gels because albumin (67,000 MW) is 10,000 times more abundant in plasma than BChE (85,000 MW), and spreads into a broad band that overlaps with the BChE band, making it impossible to visualize BChE. A second reason for using nondenaturing gels is that nondenaturing gels allow identification of BChE and carboxylesterase based on activity. Figure 4 shows activity with beta-naphthylacetate. The intense band is carboxylesterase (EC 3.1.1.1, CE) in mouse plasma. The bubble below carboxylesterase is albumin. BChE as well as several unidentified proteins also react with beta-naphthylacetate. Figure 5 shows activity of blood proteins with butyrylthiocholine. The BChE tetramer is the intense band near the top of the gel. By aligning bands in Figures 3, 4, and 5 we confirmed the identities of biotinylated albumin, carboxylesterase, and BChE in Figure 3.
Other OP Compete with FP-Biotin for Binding to Albumin
The goal of this experiment was to determine whether other OP bind to albumin. We wanted to know whether binding to albumin was a special property of FP-biotin or whether other OP also bound to albumin. If the OP binding site is a specific tyrosine (Black et al., 1999; Sanger, 1963) then pretreatment with other OP was expected to block binding of FP-biotin. Human plasma was used in this experiment because it does not contain carboxylesterase. The absence of carboxylesterase facilitated interpretation of the results, as there was no confusion between carboxylesterase and albumin. Figure 6 shows that other OP competed with FP-biotin for binding to albumin to varying extents. Pretreatment with chlorpyrifos oxon, echothiophate, malaoxon, paraoxon, methyl paraoxon, diazoxon, dichlorvos, and diisopropylfluorophosphate (DFP) reduced the binding of FP-biotin to human albumin in vitro. The only OP tested that gave no evidence of binding to albumin was tetraisopropylpyrophosphoramide (iso-OMPA).
These results suggest that OP binding to albumin is a general property of OP. It is to be noted that in this report FP-biotin has been demonstrated to bind to albumin from three species: mouse, bovine, and human. The albumins were in mouse plasma, purified bovine serum albumin, and human plasma.
DISCUSSION
OP Labels Proteins That Have No Active Site Serine
Organophosphorus toxicants are well known as inhibitors of serine proteases and serine esterases. Enzymes such as trypsin, chymotrypsin, thrombin, AChE, BChE, acylpeptide hydrolase, and carboxylesterase have a conserved active site serine with the consensus sequence GXSXG. When the active site serine is covalently modified by OP, the enzyme loses activity. Loss of enzyme activity allows one to conveniently measure reactivity with OP.
Proteins with no catalytic activity are a novel class of OP target proteins. They have no active site serine and have been shown to bind OP in vitro, but not in living animals. The advent of quadrupole time-of-flight mass spectrometry has made it possible to positively identify albumin as a protein that binds the OP FP-biotin in living mice.
Experiments with purified proteins have documented covalent attachment of OP not only to serine, but also to tyrosine and histidine. For example, human albumin covalently binds sarin, soman, and DFP at tyrosine (Black et al., 1999; Means and Wu, 1979). Bovine serum albumin is readily phosphorylated by DFP with a stoichiometry of one DFP molecule bound per molecule of albumin (Murachi, 1963). The sequence of the albumin peptide that covalently binds DFP was reported by Sanger as ArgTyrThrLys for human and rabbit albumin, and ArgTyrThrArg for bovine albumin (Sanger, 1963). After the complete amino acid sequences of the albumins was known, these reactive tyrosines were identified as Tyr 411 in human and Tyr 410 in bovine albumin (Meloun et al., 1975; Peters, 1996). Papain binds DFP on tyrosine (Chaiken and Smith, 1969), while rabbit liver carboxylesterase binds DFP on histidine as well as on the active site serine (Korza and Ozols, 1988). Bromelain is not inhibited by DFP, but it does react with DFP, leading to the formation of a fully active, phosphorus-containing enzyme (Murachi, 1963; Murachi et al., 1965). The rat M2 muscarinic acetylcholine receptor covalently binds chlorpyrifos oxon, though the binding site has not yet been identified (Bomser and Casida, 2001; Huff et al., 1994).
Living animals have previously been demonstrated to bind OP to noncholinesterase sites. The tissue distribution of radiolabeled DFP in rabbits (Jandorf and McNamara, 1950), and radiolabeled soman in rats and mouse brain (Little et al., 1988; Traub, 1985) had no correlation with cholinesterase localization. These results have been interpreted to mean that tissue proteins other than cholinesterase are capable of binding OP poisons (Jandorf and McNamara, 1950; Kadar et al., 1985).
Existing Mass Spectrometric Methods for Diagnosis of OP Exposure
Gas chromatography combined with mass spectrometry was used to detect sarin in archived blood samples from victims of the 1995 Tokyo subway attack (Polhuijs et al., 1997). The covalently bound sarin in 0.12 to 0.5 ml of serum was released by incubation with 2 M potassium fluoride at pH 4.0. This method confirmed that the people had been exposed to OP and, in addition, identified the OP as sarin. The mass-to-charge ratios of the released OP fragments were 81, 99, and 125, values characteristic of sarin.
The GC-MS method is a significant advance over simple inhibition assays, but it has limitations. (1) The method relies on being able to release the OP from its covalent attachment site on BChE with potassium fluoride. The release step requires the catalytic machinery of BChE to be intact. Samples that have been stored in ways that denature the BChE cannot undergo fluoride-induced release of OP (Polhuijs et al., 1997). (2) Proteins other than BChE, AChE, and carboxylesterase may not be able to release OP upon treatment with potassium fluoride. Human albumin after phosphylation with sarin is not amenable to reactivation with fluoride ions (Van Der Schans et al., 2004). (3) OP-derivatized protein samples that have lost an alkyl group from the phosphonate in the process called ‘aging’ would not be capable of releasing their OP upon treatment with potassium fluoride.
GC-MS of OP metabolites is the method recommended by the Centers for Disease Control for monitoring potential exposure to nerve agents. Measurement of OP metabolites is limited by the fact that these compounds are rapidly cleared from the body. Sarin metabolites were found in urine on post-exposure days 1 and 3, and in trace amounts on day 7 in a man who had inhaled a dose that made him unconscious (Nakajima et al., 1998). Urine samples from four patients hospitalized for sarin exposure showed that most of the sarin metabolites had cleared within 24 h (Hui and Minami, 2000).
Many of these limitations were overcome in an approach where plasma BChE was purified by affinity chromatography and digested with pepsin. The peptides were separated by liquid chromatography, and the OP-derivatized peptide as well as the OP it carried were identified by electrospray tandem mass spectrometry (Fidder et al., 2002). This method is well suited to detect OP exposure and will become even more valuable if OP-derivatized proteins other than BChE are included in the analysis. A major advantage of using protein-bound OP to diagnose exposure is that the OP-derivatized proteins remain in the circulation for weeks (Cohen and Warringa, 1954; Grob et al., 1947; Munkner et al., 1961; Van Der Schans et al., 2004).
Albumin Properties
Albumin is a nonglycosylated single chain of 585 amino acids folded into three domains by 17 intrachain disulfide bonds. Albumin is the most abundant soluble protein in the body of all vertebrates and is the most prominent protein in plasma. It is synthesized in the liver and discharged into the bloodstream. Of the 360 g total albumin in a human, 40% resides in the blood and 60% in extravascular fluids of tissues (Peters, 1996). The rate of transfer to the extravascular space is about 4.5% per hour. About 40% of the extravascular albumin is in muscle. Albumin is responsible for the colloid osmotic pressure of plasma and for maintenance of blood volume and supplies most of the acid/base buffering action of plasma proteins in extravascular fluids. Albumin binds and transports metabolites and drugs. Analbuminenia, a rare deficiency of albumin, is compatible with nearly normal health in humans. The study of analbuminemia shows that albumin is helpful in coping with stress and in containing environmental and physiological toxins.
Albumin and BChE are Attractive Biomarkers for OP Poisoning
Both albumin and BChE are found in human plasma, a tissue that can be readily sampled for biomarkers of OP exposure. Both proteins bind OP, though the binding to albumin of OP other than FP-biotin has not been demonstrated in vivo. Both proteins have a long half-life in the circulation of humans, 20 days for albumin (Chaudhury et al., 2003; Peters, 1996) and 10–16 days for BChE (Cohen and Warringa, 1954; Jenkins et al., 1967; Munkner et al., 1961; Ostergaard et al., 1988). The long half-life of OP-labeled proteins contrasts with the short half-life of OP and OP metabolites in urine. This provides a significant advantage to a method that uses OP-labeled proteins for diagnosis of OP exposure.
Role of Albumin in OP Toxicity
Binding of OP to albumin could serve to scavenge OP molecules and therefore reduce the amount of OP available for reaction with AChE. The affinity of various OP for albumin is likely to vary, so the effectiveness of albumin as an OP scavenger is likely to depend on the identity of the OP.
Another consideration in evaluating the role of albumin in OP toxicity is competition with drugs for binding to the same site in albumin. Several drugs, including diazepam (Fehske et al., 1979; Peters, 1996) and ibuprofen, bind to the same region of albumin that binds OP (see Tables 3–5 in Peters, 1996). Since the free form of a drug is active, the extent of binding to albumin controls both the effect and duration of the drug. Decreased binding of a drug to albumin can cause toxicity owing to an increase in the concentration of its free form. Thus, a person being treated for intestinal worms with metrifonate (an OP that is converted nonenzymatically to dichlorvos) might be unable to tolerate a standard dose of diazepam or ibuprofen. Or conversely, a person taking ibuprofen, could be intolerant of a standard dose of metrifonate.
ACKNOWLEDGMENTS
Supported by U.S. Army Medical Research and Materiel Command grants DAMD17-01–1–0776 (to O.L.), and DAMD17-01–0795 (to C.M.T.), UNMC Eppley Cancer Center Support Grant P30CA36727, National Science Foundation grant MCB9808372 (to C.M.T.), EPS0091995 (to C.M.T.) and U.S. Army Research, Development & Engineering Command grant W911SR-04-C-0019 (to O.L.) The information does not necessarily reflect the position or the policy of the U.S. Government, and no official endorsement should be inferred. Holly Coughenour at the University of Montana provided the protocol for in-gel reduction, alkylation, and digestion of proteins. We thank the anonymous reviewer for helpful references on the interaction of OP with albumin.
NOTES
REFERENCES
Black, R. M., Harrison, J. M., and Read, R. W. (1999). The interaction of sarin and soman with plasma proteins: The identification of a novel phosphonylation site. Arch. Toxicol. 73, 123–126.
Bomser, J. A., and Casida, J. E. (2001). Diethylphosphorylation of rat cardiac M2 muscarinic receptor by chlorpyrifos oxon in vitro. Toxicol. Lett. 119, 21–26.
Casida, J. E., and Quistad, G. B. (2004). Organophosphate toxicology: Safety aspects of nonacetylcholinesterase secondary targets. Chem. Res. Toxicol. 17, 983–998.
Chaiken, I. M., and Smith, E. L. (1969). Reaction of a specific tyrosine residue of papain with diisopropylfluorophosphate. J. Biol. Chem. 244, 4247–4250.
Chaudhury, C., Mehnaz, S., Robinson, J. M., Hayton, W. L., Pearl, D. K., Roopenian, D. C., and Anderson, C. L. (2003). The major histocompatibility complex-related Fc receptor for IgG (FcRn) binds albumin and prolongs its lifespan. J. Exp. Med. 197, 315–322.
Cohen, J. A., and Warringa, M. G. (1954). The fate of P32 labelled diisopropylfluorophosphonate in the human body and its use as a labelling agent in the study of the turnover of blood plasma and red cells. J. Clin. Invest. 33, 459–467.
Duysen, E. G., Stribley, J. A., Fry, D. L., Hinrichs, S. H., and Lockridge, O. (2002). Rescue of the acetylcholinesterase knockout mouse by feeding a liquid diet; phenotype of the adult acetylcholinesterase deficient mouse. Brain Res. Dev. Brain Res. 137, 43–54.
Ellman, G. L., Courtney, K. D., Andres, V., Jr., and Feather-Stone, R. M. (1961). A new and rapid colorimetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7, 88–95.
Fehske, K. J., Muller, W. E., and Wollert, U. (1979). A highly reactive tyrosine residue as part of the indole and benzodiazepine binding site of human serum albumin. Biochim. Biophys. Acta 577, 346–359.
Fidder, A., Hulst, A. G., Noort, D., de Ruiter, R., van der Schans, M. J., Benschop, H. P., and Langenberg, J. P. (2002). Retrospective detection of exposure to organophosphorus anti-cholinesterases: Mass spectrometric analysis of phosphylated human butyrylcholinesterase. Chem. Res. Toxicol. 15, 582–590.
Grob, D., Lilienthal, J. L., Harvey, A. M., and Jones, B. F. (1947). The administration of di-isopropylfluorophosphate (DFP) to man. I. Effect on plasma and erythrocyte cholinesterase; general systemic effects; use in study of hepatic function and erythropoiesis; and some properties of plasma cholinesterase. Bull. Johns Hopkins Hosp. 81, 217–244.
Huff, R. A., Corcoran, J. J., Anderson, J. K., and Abou-Donia, M. B. (1994). Chlorpyrifos oxon binds directly to muscarinic receptors and inhibits cAMP accumulation in rat striatum. J. Pharmacol. Exp. Ther. 269, 329–335.
Hui, D. M., and Minami, M. (2000). Monitoring of fluorine in urine samples of patients involved in the Tokyo sarin disaster, in connection with the detection of other decomposition products of sarin and the by-products generated during sarin synthesis. Clin. Chim. Acta 302, 171–188.
Jandorf, B. J., and McNamara, P. D. (1950). Distribution of radiophosphorus in rabbit tissues after injection of phosphorus-labeled diisopropyl fluorophosphate. J. Pharmacol. Exp. Ther. 98, 77–84.
Jenkins, T., Balinsky, D., and Patient, D. W. (1967). Cholinesterase in plasma: First reported absence in the Bantu; half-life determination. Science 156, 1748–1750.
Kadar, T., Raveh, L., Cohen, G., Oz, N., Baranes, I., Balan, A., Ashani, Y., and Shapira, S. (1985). Distribution of 3H-soman in mice. Arch. Toxicol. 58, 45–49.
Kadner, S. S., Katz, J., and Finlay, T. H. (1992). Esterase-1: Developmental expression in the mouse and distribution of related proteins in other species. Arch. Biochem. Biophys. 296, 435–441.
Karnovsky, M. J., and Roots, L. (1964). A ‘direct-coloring’ thiocholine method for cholinesterases. J. Histochem. Cytochem. 12, 219–221.
Kidd, D., Liu, Y., and Cravatt, B. F. (2001). Profiling serine hydrolase activities in complex proteomes. Biochemistry 40, 4005–4015.
Korza, G., and Ozols, J. (1988). Complete covalent structure of 60-kDa esterase isolated from 2,3,7,8-tetrachlorodibenzo-p-dioxin-induced rabbit liver microsomes. J. Biol. Chem. 263, 3486–3495.
Little, P. J., Scimeca, J. A., and Martin, B. R. (1988). Distribution of [3H]diisopropylfluorophosphate, [3H]soman, [3H]sarin, and their metabolites in mouse brain. Drug Metab. Dispos. 16, 515–520.
Liu, Y., Patricelli, M. P., and Cravatt, B. F. (1999). Activity-based protein profiling: The serine hydrolases. Proc. Natl. Acad. Sci. U.S.A. 96, 14694–14699.
McDonough, J. H., Jr., and Shih, T. M. (1997). Neuropharmacological mechanisms of nerve agent-induced seizure and neuropathology. Neurosci. Biobehav. Rev. 21, 559–579.
Means, G. E., and Wu, H. L. (1979). The reactive tyrosine residue of human serum albumin: Characterization of its reaction with diisopropylfluorophosphate. Arch. Biochem. Biophys. 194, 526–530.
Meloun, B., Moravek, L., and Kostka, V. (1975). Complete amino acid sequence of human serum albumin. FEBS Lett. 58, 134–137.
Moser, V. C. (1995). Comparisons of the acute effects of cholinesterase inhibitors using a neurobehavioral screening battery in rats. Neurotoxicol. Teratol. 17, 617–625.
Munkner, T., Matzke, J., and Videbaek, A. (1961). Cholinesterase activity of human plasma after intramuscular diisopropyl fluorophosphonate (DFP). Acta Pharmacol. Toxicol. (Copenh.) 18, 170–174.
Murachi, T. (1963). A general reaction of diisopropylphosphorofluoridate with proteins without direct effect on enzymic activities. Biochim. Biophys. Acta 71, 239–241.
Murachi, T., Inagami, T., and Yasui, M. (1965). Evidence for alkylphosphorylation of tyrosyl residues of stem bromelain by diisopropylphosphorofluoridate. Biochemistry 4, 2815–2825.
Nachlas, M. M., and Seligman, A. M. (1949). The histochemical demonstration of esterase. J. Natl. Cancer Inst. 9, 415–425.
Nakajima, T., Sasaki, K., Ozawa, H., Sekjima, Y., Morita, H., Fukushima, Y., and Yanagisawa, N. (1998). Urinary metabolites of sarin in a patient of the Matsumoto sarin incident. Arch. Toxicol. 72, 601–603.
Ortigoza-Ferado, J., Richter, R. J., Hornung, S. K., Motulsky, A. G., and Furlong, C. E. (1984). Paraoxon hydrolysis in human serum mediated by a genetically variable arylesterase and albumin. Am. J. Hum. Genet. 36, 295–305.
Ostergaard, D., Viby-Mogensen, J., Hanel, H. K., and Skovgaard, L. T. (1988). Half-life of plasma cholinesterase. Acta Anaesthesiol. Scand. 32, 266–269.
Peters, T., Jr. (1996). All about Albumin. Biochemistry, Genetics, and Medical Applications. Academic Press Ltd., London.
Polhuijs, M., Langenberg, J. P., and Benschop, H. P. (1997). New method for retrospective detection of exposure to organophosphorus anticholinesterases: Application to alleged sarin victims of Japanese terrorists. Toxicol. Appl. Pharmacol. 146, 156–161.
Pope, C. N. (1999). Organophosphorus pesticides: Do they all have the same mechanism of toxicity J. Toxicol. Environ. Health B Crit. Rev. 2, 161–181.
Ray, D. E., and Richards, P. G. (2001). The potential for toxic effects of chronic, low-dose exposure to organophosphates. Toxicol. Lett. 120, 343–351.
Salvi, R. M., Lara, D. R., Ghisolfi, E. S., Portela, L. V., Dias, R. D., and Souza, D. O. (2003). Neuropsychiatric evaluation in subjects chronically exposed to organophosphate pesticides. Toxicol. Sci. 72, 267–271.
Sanger, F. (1963). Amino-acid sequences in the active centres of certain enzymes. Proc. Chem. Soc. 5, 76–83.
Schwartz, M. (1982). A serine protease activity of human serum albumin towards 4-methylumbelliferyl-guanidinobenzoate (MUGB) and diisopropyl fluorophosphate (DEP): Implications for the use of MUGB reactivity in amniotic fluid in prenatal diagnosis of cystic fibrosis. Clin. Chim. Acta 124, 213–223.
Srivastava, A. K., Gupta, B. N., Bihari, V., Mathur, N., Srivastava, L. P., Pangtey, B. S., Bharti, R. S., and Kumar, P. (2000). Clinical, biochemical and neurobehavioural studies of workers engaged in the manufacture of quinalphos. Food Chem. Toxicol. 38, 65–69.
Sultatos, L. G., Basker, K. M., Shao, M., and Murphy, S. D. (1984). The interaction of the phosphorothioate insecticides chlorpyrifos and parathion and their oxygen analogues with bovine serum albumin. Mol. Pharmacol. 26, 99–104.
Tove, S. B. (1962). The esterolytic activity of serum albumin. Biochim. Biophys. Acta 57, 230–235.
Traub, K. (1985). In vivo distribution of 14C radiolabeled soman [3,3-dimethyl-2-butoxy)-methylphosphorylfluoride) in the central nervous system of the rat. Neurosci. Lett. 60, 219–225.
Van Der Schans, M. J., Polhuijs, M., Van Dijk, C., Degenhardt, C. E., Pleijsier, K., Langenberg, J. P., and Benschop, H. P. (2004). Retrospective detection of exposure to nerve agents: Analysis of phosphofluoridates originating from fluoride-induced reactivation of phosphylated BuChE. Arch. Toxicol. 78, 508–524.
Xie, W., Stribley, J. A., Chatonnet, A., Wilder, P. J., Rizzino, A., McComb, R. D., Taylor, P., Hinrichs, S. H., and Lockridge, O. (2000). Postnatal developmental delay and supersensitivity to organophosphate in gene-targeted mice lacking acetylcholinesterase. J. Pharmacol. Exp. Ther. 293, 896–902.(Eric S. Peeples, Lawrence)
University of Montana, Department of Biomedical and Pharmaceutical Sciences, Missoula, Montana 59812
ABSTRACT
The classical laboratory tests for exposure to organophosphorus toxicants (OP) are inhibition of acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) activity in blood. In a search for new biomarkers of OP exposure, we treated mice with a biotinylated organophosphorus agent, FP-biotin. The biotinylated proteins in muscle were purified by binding to avidin-Sepharose, separated by gel electrophoresis, digested with trypsin, and identified from their fragmentation patterns on a quadrupole time-of-flight mass spectrometer. Albumin and ES1 carboxylesterase (EC 3.1.1.1) were found to be major targets of FP-biotin. These FP-biotinylated proteins were also identified in mouse plasma by comparing band patterns on nondenaturing gels stained for albumin and carboxylesterase activity, with band patterns on blots hybridized with Streptavidin Alexa-680. Two additional FP-biotin targets, AChE (EC 3.1.1.7) and BChE (EC 3.1.1.8), were identified in mouse plasma by finding that enzyme activity was inhibited 50–80%. Mouse plasma contained eight additional FP-biotinylated bands whose identity has not yet been determined. In vitro experiments with human plasma showed that chlorpyrifos oxon, echothiophate, malaoxon, paraoxon, methyl paraoxon, diazoxon, diisopropylfluorophosphate, and dichlorvos competed with FP-biotin for binding to human albumin. Though experiments with purified albumin have previously shown that albumin covalently binds OP, this is the first report of OP binding to albumin in a living animal. Carboxylesterase is not a biomarker in man because humans have no carboxylesterase in blood. It is concluded that OP bound to albumin could serve as a new biomarker of OP exposure in man.
Key Words: acetylcholinesterase; butyrylcholinesterase; organophosphate; FP-biotin; albumin.
INTRODUCTION
Organophosphorus toxicants (OP) are used in agriculture as pesticides, in medical practice as antihelminthics, in the airline industry as additives to hydraulic fluid and jet engine oil, and as chemical warfare agents. These compounds are known to exert their acute effects by inhibiting acetylcholinesterase (EC 3.1.1.7, AChE). The excess acetylcholine that accumulates causes an imbalance in the nervous system that can result in death (McDonough and Shih, 1997).
Though AChE is the clinically important target of OP exposure, other proteins also form a covalent bond with OP, depending on the identity of the OP (Casida and Quistad, 2004). These secondary targets include butyrylcholinesterase, acylpeptide hydrolase, neurotoxic esterase, fatty acid amide hydrolase, arylformamidase, cannabinoid CB1 receptor, muscarinic acetylcholine receptor, and carboxylesterase. With the exception of neurotoxic esterase, whose inhibition is responsible for delayed neuropathy, the toxicological relevance of inhibition of these secondary targets is not yet understood (Casida and Quistad, 2004; Ray and Richards, 2001). Albumin has not previously been shown to bind OP in a living animal, though in vitro experiments with purified albumin have demonstrated covalent binding to diisopropylfluorophosphate, sarin, and soman (Black et al., 1999; Means and Wu, 1979; Murachi, 1963; Sanger, 1963; Schwartz, 1982). Albumin hydrolyzes chlorpyrifos oxon and paraoxon (Ortigoza-Ferado et al., 1984; Sultatos et al., 1984).
The toxic effects of a particular OP cannot be attributed entirely to inhibition of AChE. Toxic signs are different for each OP when that OP is administered at low doses (Moser, 1995). For example, a low dose of fenthion decreased motor activity in rats by 86% but did not alter the tail-pinch response, whereas a low dose of parathion did not affect motor activity but did decrease the tail-pinch response (Moser, 1995). These toxicological observations suggest that OP have other biological actions in addition to their cholinesterase-inhibitory properties.
Another confounding observation is the finding that toxic signs do not correlate with degree of AChE inhibition. Rats given doses of OP that inhibited AChE to similar levels had more severe toxicity from parathion than chlorpyrifos (Pope, 1999). There are also examples of toxic signs unaccompanied by AChE inhibition. Workers who manufacture the OP pesticide quinalphos have significantly low scores for memory, learning ability, vigilance, and motor response, though their blood AChE activity levels are the same as in control subjects (Srivastava et al., 2000). Chronic low-level exposure to OP induces neuropsychiatric disorders without inhibition of esterase activity (Ray and Richards, 2001; Salvi et al., 2003). These observations have led to the suggestion that some OP have toxicologically relevant sites of action in addition to AChE (Moser, 1995; Pope, 1999; Ray and Richards, 2001). The hypothesis arose that a given OP reacts not only with AChE, but with a set of proteins unique for each OP.
Our goal is to identify the proteins in a living animal that covalently bind the biotin-tagged OP called FP-biotin (Kidd et al., 2001; Liu et al., 1999). In this report we used tandem mass spectrometry, enzyme activity assays, gel electrophoresis, and blots to identify four FP-biotin-labeled proteins in the muscle and plasma of mice that had been injected with FP-biotin ip. We found FP-biotin-labeled albumin, carboxylesterase, BChE, and AChE. This is the first report to demonstrate that albumin is a significant target of OP binding in a living animal. Eight other proteins in mouse blood became labeled but have not yet been identified.
MATERIALS AND METHODS
Materials. FP-biotin (MW 592.3) was custom synthesized by Troy Voelker in the laboratory of Charles M. Thompson at the University of Montana (Liu et al., 1999). Purity was checked by NMR and mass spectrometry, and no evidence of contamination was detected. FP-biotin was stored as a dry powder at –70°C. Just before use, the dry powder was dissolved in 100% ethanol to a concentration of 13.3 mg/ml and diluted with saline to 15% ethanol containing 2 mg/ml FP-biotin.
Immun-Blot PVDF membrane for protein blotting, 0.2 μm (catalog #162-0177) and biotinylated molecular weight markers (catalog #161–0319) were from Bio-Rad Laboratories, Hercules, CA. Streptavidin Alexa-680 fluorophore (catalog #S-21378) was from Molecular Probes, Eugene, OR. Avidin-agarose beads (catalog # A-9207), iso-OMPA, and bovine albumin, essentially fatty acid-free (Sigma A 7511) were from Sigma-Aldrich, St. Louis, MO. Echothiophate iodide was from Wyeth-Ayerst, Rouses Point, NY. All other OP were from Chem Service Inc, West Chester, PA.
Mice. The Institutional Animal Care and Use Committee of the University of Nebraska Medical Center approved all procedures involving mice. Animal care was provided in accordance with the principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Mice completely lacking AChE protein were made by gene targeting (Xie et al., 2000) at the University of Nebraska Medical Center. Exons 2, 3, 4, and 5 of the ACHE gene were deleted to make AChE–/– mice. The AChE–/– animals are in strain 129Sv genetic background. The colony is maintained by breeding heterozygotes because AChE–/– mice do not breed (Duysen et al., 2002). Wild-type mice are littermates of AChE–/– mice. The strain 129Sv mice were used for experiments with FP-biotin.
Injection of FP-biotin into mice. Mice were injected intraperitoneally with FP-biotin dissolved in 15% ethanol to give a dose of 56 or 5 mg/kg, or with vehicle alone. The dose of FP-biotin was calculated from dry weight. No correction was made for the fact that FP-biotin is a mixture of phosphorus stereoisomers. Mice were euthanized 120 min after FP-biotin injection, by inhalation of carbon dioxide. Blood was washed out of tissues by intracardial perfusion with saline solution. Tissues from six mice were analyzed by mass spectrometry: two AChE–/– FP-biotin treated (5 and 56 mg/kg), two AChE–/– untreated, one wild-type FP-biotin treated (56 mg/kg), and one wild-type untreated. In addition wild-type mice were treated with 0, 0.5, 1.0, 5.0, or 18.8 mg/kg FP-biotin (n = 2 for each dose). Blood was analyzed by gel electrophoresis, blotting, and enzyme activity assays.
Enzyme activity. AChE activity was measured with 1 mM acetylthiocholine in the presence of 0.5 mM dithiobisnitrobenzoic acid (Ellman et al., 1961) after inhibiting BChE activity for 30 min with 0.1 mM iso-OMPA, in 0.1 M potassium phosphate pH 7.0, at 25°C. BChE activity was measured with 1 mM butyrylthiocholine. E412nm = 13,600 M–1 cm–1 at pH 7.0. Carboxylesterase activity was measured with 5 mM p-nitrophenyl acetate after inhibiting AChE and BChE with 0.01 mM eserine, and after inhibiting paraoxonase with 12.5 mM EDTA. E400nm = 9000 M–1 cm–1 at pH 7.0. AChE, BChE, and carboxylesterase units of activity are micromoles hydrolyzed per minute at pH 7.0, 25°C.
Isolation of FP-biotin-labeled protein. To prepare OP-labeled proteins for mass spectrometry, FP-biotin-labeled proteins in muscle were purified on avidin-agarose beads and separated by SDS–polyacrylamide gel electrophoresis. Proteins from mice that had
Tissues were homogenized in 10 volumes of 50 mM TrisCl pH 8.0 containing 5 mM EDTA, and centrifuged for 10 min in a microfuge at 12,000 rpm to partially clarify the suspension. A detailed example of the protocol follows. The 0.96 ml of muscle homogenate (7.4 mg protein/ml) was diluted with 3.75 ml of 50 mM TrisCl pH 8.0, 5 mM EDTA to make 1.5 mg/ml protein solution. SDS was added to make the solution 0.5% SDS. The protein solution was heated for 3 min in a boiling water bath and then diluted with buffer to make the final SDS concentration 0.2%. The protein solution was incubated with 100 μl of washed avidin-agarose beads (1.9 mg avidin/ml of beads) overnight at room temperature, with continuous inversion, to bind the FP-biotin-labeled proteins to the beads. Beads were washed three times with the TrisCl/EDTA buffer, containing 0.2% SDS, to remove nonspecifically bound protein. Twenty-five μl of 6x SDS–PAGE loading buffer (0.2 M TrisCl, pH 6.8, 10% SDS, 30% glycerol, 0.6 M dithiothreitol and 0.012% bromophenol blue) were added to the 100 μl of beads, and the mixture was heated at 85°C for 3 min. This step released the biotinylated proteins from the avidin beads. Equal amounts of the bead mixture were loaded directly into two wells of a 10–20% gradient SDS–PAGE (10-well format, 1.5-mm thick) and run for 4000 volt-hours in the cold room. The gel was stained with Coomassie blue G250 (Bio-Safe from BioRad), and destained with water. Coomassie G250 is reportedly 2–8 fold more sensitive than Coomassie R250 (BioRad specifications). To minimize contamination from keratin, the staining dish had been cleaned with sulfuric acid, and the water was Milli-Q purified. For the same reason, gloves were worn for all operations involving the gel.
Protein digestion protocol. The proteins separated on SDS–PAGE were digested with trypsin to prepare them for identification by mass spectral analysis. Gloves were worn throughout these procedures, all solutions were made with Milli-Q purified water, and all glassware, plasticware, and tools were rinsed with Milli-Q purified water to minimize keratin contamination. Each Coomassie-stained band from one lane of the SDS–PAGE was excised, placed into a separate 1.5-ml microfuge tube, and chopped into bits. The amount of gel excised was kept to a minimum. The gel bits were destained by washing with 200 μl of 25 mM ammonium bicarbonate (Aldrich) in 50% acetonitrile (synthesis grade, from Fisher). After three washes, the gel bits were colorless and had shrunken considerably. Residual liquid was removed, and the gel bits dried by evaporation in a Speedvac (Jouan). Disulfide bonds in the protein were reduced by incubating the gel bits with 10 mM dithiothreitol (molecular biology grade, from Sigma) in 200 μl of 100 mM ammonium bicarbonate for 1 h at 56°C. The gel pieces were then centrifuged, excess solution was removed, and the protein was alkylated with 55 mM iodoacetamide (Sigma) in 120 μl of 100 mM ammonium bicarbonate for 1 h at room temperature in the dark. The gel bits were again centrifuged, excess solution was removed, and the bits were washed with 200 μl of 25 mM ammonium bicarbonate in 50% acetonitrile (three times). Residual liquid was again removed and the gel bits dried by evaporation in the Speedvac. The proteins were digested in the gel with trypsin, using 12.5 ng/μl of sequencing grade trypsin (Promega) in 25 mM ammonium bicarbonate. Ninety μl of the trypsin solution were added to the dry gel bits and incubated at 4°C for 20 min, to allow the gel to re-swell. Then 60 μl of 25 mM ammonium bicarbonate were layered over each sample, and the samples were incubated at 37°C overnight (about 17 h). Peptides were extracted by incubating each reaction mixture with 200 μl of 0.1% trifluoroacetic acid (sequencing grade from Beckman) in 60% acetonitrile for one h at room temperature. Extraction was repeated three times, and the extracts for each sample were pooled. The pooled extracts were evaporated to dryness in the Speedvac, and the dry samples were stored at –20°C until analyzed.
Mass spectral analysis. Each tryptic peptide digest was resuspended in 40 μl of 5% aqueous acetonitrile/0.05% trifluoroacetic acid. A 10-μl aliquot of the digest was injected into a CapLC (capillary liquid chromatography system from Waters Corp) using 5% aqueous acetonitrile/0.05% trifluoroacetic acid (auxiliary solvent) at a flow rate of 20 μl per min. Peptides were concentrated on a C18 PepMapTM Nano-PrecolumnTM (5 mm x 0.3 mm id, 5μm particle size) for 3 min, and then eluted onto a C18 PepMapTM capillary column (15 cm x 75 μm id, 3 μm particle size both from LC Packings), using a flow rate of 200–300 nl per min. Peptides were partially resolved using gradient elution. The solvents were 2% aqueous acetonitrile/0.1% formic acid (solvent A), and 90% acetonitrile/10% isopropanol/0.2% formic acid (solvent B). The solvent gradient increased from 5% B to 50% B over 22 min, then to 80% B over 1 min, and remained at 80% B for 4 min. The column was then flushed with 95% B for 3 min and equilibrated at 5% B for 3 min before the next sample injection.
Peptides were delivered to the Z-spray source (nano-sprayer) of a Micromass Q-TOF (tandem quadrupole/time-of-flight mass spectrometer from Waters Corp.) through a 75-μm id capillary, which connected to the CapLC column. In order to ionize the peptides, 3300 volts were applied to the capillary, 30 volts to the sample cone, and zero volts to the extraction cone. Mass spectra for the ionized peptides were acquired throughout the chromatographic run, and collision-induced dissociation spectra were acquired on the most abundant peptide ions (having a charge state of 2+, 3+, or 4+). The collision-induced dissociation spectrum is unique for each peptide and is based on the amino acid sequence of that peptide. For this reason, identification of proteins using collision-induced dissociation data is superior to identification by only the peptide mass fingerprint of the protein. The collision cell was pressurized with 1.5 psi ultrapure argon (99.999%), and collision voltages were dependent on the mass-to-charge ratio and the charge state of the parent ion. The time-of-flight measurements were calibrated daily using fragment ions from collision-induced dissociation of [Glu1]-fibrinopeptide B. Each sample was post-processed using this calibration and Mass Measure (Micromass). The calibration was adjusted to the exact mass of the autolytic tryptic fragment at 421.76, found in each sample.
The mass and sequence information for each detected peptide was submitted either to ProteinLynx Global Server 1.1 (a proprietary software package, from Micromass), or to MASCOT (a public access package provided by Matrix Science at http://www.matrix-science.com). Data were compared to all mammalian entries (ProteinLynx) or just mouse entries (MASCOT) in the NCBInr database (National Center for Biotechnology Information). Search criteria for ProteinLynx were set to a mass accuracy of 0.25 Da, fixed modification of methionine (oxidation), and variable modification of cysteine (carbamidomethylation). One missed cleavage by trypsin was allowed. Search criteria for MASCOT were set to a mass accuracy of ±0.1 Da, one missed cleavage, variable modification of methionine (oxidation) and cysteine (carbamidomethylation), and peptide charge +2 and +3. Both software packages calculated a score for each identified protein based on the match between the experimental peptide mass and the theoretical peptide mass, as well as between the experimental collision-induced dissociation spectra and the theoretical fragment ions from each peptide. Results were essentially the same from both packages.
Polyacrylamide gel electrophoresis. Gradient polyacrylamide gels (4–30%) were cast in a Hoefer gel apparatus. Electrophoresis was for 5000 volt-hours (200 volts for 25 h) at 4°C for nondenaturing gels and 2500 volt-hours (100 volts for 25 h) at 4°C for gels containing 0.1% SDS.
Staining gels for BChE activity. Nondenaturing gels were stained for BChE activity by the method of Karnovsky and Roots (1964). The staining solution contained 180 ml of 0.2 M sodium maleate pH 6.0, 15 ml of 0.1 M sodium citrate, 30 ml of 0.03 M cupric sulfate, 30 ml of 5 mM potassium ferricyanide, and 0.18 g butyrylthiocholine iodide in a total volume of 300 ml. Gels were incubated, with shaking, at room temperature for 3 to 5 h. The reaction was stopped by washing the gels with water. To determine the location of albumin, activity-stained gels were stained with Coomassie blue.
Staining gels for carboxylesterase activity and albumin. Nondenaturing gels were incubated in 100 ml of 50 mM TrisCl pH 7.4 in the presence of 50 mg beta-naphthylacetate dissolved in 1 ml ethanol, and 50 mg of solid Fast Blue RR. The naphthylacetate precipitates when it is added to the buffer, but enough remains in solution that the reaction works. Though the Fast Blue RR does not dissolve, pink to purple bands develop on the gel within minutes (Nachlas and Seligman, 1949). A maximum of 30 min incubation at room temperature was needed. The gels were washed with water and photographed. This stain is primarily for carboxylesterase. Albumin gives a faint band with this method because albumin slowly hydrolyzes beta-naphthylacetate (Tove, 1962). Activity-stained gels were counterstained with Coomassie blue to verify the location of albumin.
To align bands on gels stained for enzyme activity with biotinylated bands on a PVDF membrane, the transparent activity-stained gels were placed on top of a printed image of the fluorescent bands in the PVDF membrane.
Visualizing FP-biotin-labeled proteins. For determination of the number and size of proteins labeled by FP-biotin, proteins were subjected to gel electrophoresis, transfer to a PVDF membrane, and hybridization with a fluorescent probe. The details of the procedure follow.
Proteins were transferred from the polyacrylamide gel to PVDF membrane (Immun-Blot from BioRad) electrophoretically in a tank using plate electrodes (TransBlot from BioRad), at 0.5 amps, for 1 h, in 3 l of 25 mM Tris/192 mM glycine buffer, pH 8.2, in the cold room (4°C), with stirring. The membrane was blocked with 3% gelatin (BioRad) in 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl for 1 h at room temperature. The 3% gelatin solution had been prepared by heating the gelatin in buffer in a microwave oven for several seconds. The blocked membrane was washed three times with 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl and 0.05% Tween-20, for 5 min.
Biotinylated proteins were labeled with 9.5 nM Streptavidin Alexa-680 fluorophore in 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl, 0.05% Tween-20, 0.2% SDS, and 1% gelatin, for 2 h, at room temperature, protected from light. Shorter reaction times resulted in less labeling. The SDS was found to increase the intensity of labeling. The membrane was washed twice with 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl and 0.05% Tween-20, and twice with 20 mM TrisCl buffer, pH 7.5, containing 0.5 M NaCl, for 20 min each, while protected from light.
Membranes were scanned with the Odyssey Infrared Imaging System (LI-COR, Lincoln, NE) at 42 microns per pixel. The Odyssey employs an infrared laser to excite a fluorescent probe, which is attached to the target protein, and then collects the emitted light. The emitted light intensity is directly proportional to the amount of probe. Both the laser and the detector are mounted on a moving carriage positioned directly below the membrane. The membrane can be scanned in step sizes as small as 21 microns, providing resolution comparable to X-ray film. Data are collected using a 16-bit dynamic range. The fluorophore is stable in the laser, making it possible to scan the membrane repeatedly, while using different intensity settings to optimize data collection for both strong and weak signals. The membrane was kept wet during scanning.
Biotinylated protein standards. The biotinylated bovine serum albumin (BSA) standard was prepared by incubating 10 μM BSA (0.5 mg/ml) with 20 μM FP-biotin in 20 mM TrisCl pH 7.4 at room temperature for 16 h. The biotinylated BChE standard was prepared by incubating 50 nM human BChE (3 units/ml; 4.2 μg/ml) with 10 μM FP-biotin, in 20 mM TrisCl pH 7.4 at room temperature for 16 h.
Amount of biotinylated albumin in mouse plasma. The percentage of FP-biotinylated albumin in mouse plasma was estimated from the relative intensities of the biotinylated albumin band and the biotinylated BChE band on a blot stained with Streptavidin Alexa-680. The concentration of BChE in mouse plasma is 0.003 mg/ml. The concentration of biotinylated BChE was calculated from the reduction in enzyme activity. The concentration of albumin in mouse plasma is 50 mg/ml. These values allowed estimation of percent biotinylated albumin in plasma. For example, when BChE activity was inhibited 35%, the biotinylated BChE band represented about 0.001 mg/ml biotinylated BChE. A biotinylated albumin band of similar intensity would contain 0.001 mg/ml biotinylated albumin. When mouse plasma had to be diluted 1000-fold to reduce the intensity of biotinylated albumin to a similar intensity as the biotinylated BChE in undiluted plasma, it was calculated that the concentration of biotinylated albumin in undiluted plasma was 1 mg/ml.
Binding various OP to human albumin. Human plasma was diluted 1:100 to reduce the albumin concentration to 10 μM. The diluted plasma was reacted in vitro with 10 mM malaoxon, paraoxon, chlorpyrifos oxon, methyl paraoxon, dichlorvos, diisopropylfluorophosphate diazoxon, echothiophate, or iso-OMPA for 1 h in 20 mM TrisCl pH 7.5 at 25°C. Then FP-biotin was added to 10 μM and allowed to react for 1 h, and 10 μl containing the equivalent of 0.1 μl plasma was loaded per lane on a nondenaturing gel. Biotinylated proteins were visualized with Streptavidin Alexa-680 after transfer to PVDF membrane.
Statistical analysis. Samples were analyzed by independent samples t-test assuming equal variances. Probability values less than 0.05 were considered significant.
RESULTS
Toxicity of FP-Biotin
The structure of FP-biotin is given in Figure 1. A dose of 56 mg/kg FP-biotin ip was lethal to AChE–/– mice and reduced plasma BChE activity from 1.9 ± 0.4 units/ml in untreated animals to 0.006 ± 0.003 units/ml post treatment, a 99.7% inhibition. A dose of 18.8 mg/kg ip was not lethal, but did cause severe cholinergic signs of toxicity. This dose reduced plasma BChE activity to 0.34 ± 0.09 units/ml, an 82% inhibition. A dose of 5 mg/kg caused only mild signs of toxicity and inhibited plasma BChE of AChE–/– mice 37%, to 1.2 ± 0.4 units/ml. In contrast, wild-type mice showed no signs of toxicity after treatment with 18.8 or 5 mg/kg FP-biotin ip, even though their plasma BChE activity was inhibited to the same extent as in the AChE–/– mice.
Plasma AChE activity in wild-type mice treated with 18.8 mg/kg FP-biotin was inhibited from a predose level of 0.30 ± 0.01 units/ml to 0.13 units/ml, a 56% inhibition. There was no inhibition at lower doses. AChE activity was not measured in AChE–/– mice, because these knockout animals have no AChE activity (Xie et al., 2000). AChE has a 10-fold lower affinity for FP-biotin compared to BChE (Schopfer, unpublished). This explains why a given dose of FP-biotin caused less inhibition of AChE than of BChE.
Plasma carboxylesterase activity was inhibited to the same extent in AChE–/– and +/+ mice. A dose of 18.8 mg/kg FP-biotin caused a reduction from the untreated values of 18.6 ± 0.6 units/ml to 3.5 units/ml, an 81% inhibition, while a dose of 5 mg/kg reduced the carboxylesterase levels to 9.2 ± 1.6 units/ml, a 50% reduction.
Fp-Biotin Does Not Cross the Blood–Brain Barrier
FP-biotin treated AChE–/– mice showed no inhibition of BChE in brain, supporting the conclusion that FP-biotin does not cross the blood–brain barrier.
Identification of FP-Biotinylated Proteins by Mass Spectrometry
Muscle proteins from mice that had been treated with FP-biotin, as well as from untreated control mice, were isolated by binding to avidin beads. The proteins were released from avidin by boiling in SDS and separated by SDS gel electrophoresis. Protein bands visible with Coomassie blue staining were excised and digested with trypsin. Fragmentation of tryptic peptides yielded amino acid sequence information characteristic of the protein. The proteins listed in Table 1 were consistently identified in three separate experiments. The probability score for correct identification (MOWSE score) was extremely high at 1157 for albumin and 655 for ES1 carboxylesterase. A MOWSE score of 69 is significant (p < 0.05), so scores of 1157 and 655 show complete confidence. Though albumin and ES1 carboxylesterase were found in samples prepared from muscle, these proteins are typically expressed at high levels in blood (Kadner et al., 1992; Peters, 1996). It is likely that they were transported from the blood into the extravascular fluid where they were not washed out by perfusion. Albumin was identified by 17 peptides. A representative mass spectrum of an albumin peptide is shown in Figure 2. These 17 peptides represented 51% of the mouse albumin sequence, leaving no doubt that albumin was labeled by FP-biotin. The untreated control tissues did not show albumin, thus demonstrating that only biotinylated albumin had bound to avidin beads. This control experiment eliminated the possibility that the albumin might have bound nonspecifically to the avidin beads or that albumin was endogenously biotinylated.
FP-biotinylated AChE and BChE were not found because these proteins are not abundant enough to give a Coomassie blue stained band on SDS gels. In this work only proteins that gave a Coomassie stained band were analyzed by mass spectrometry.
Identification of Endogenous Biotinylated Proteins
Endogenous biotinylated proteins were identified by mass spectrometry. The proteins in Table 2 were found in muscle of untreated as well as in FP-biotin treated mice. They are propionyl CoA carboxylase alpha, pyruvate carboxylase, and methylcrotonyl CoA carboxylase alpha. These endogenous biotinylated proteins bound to avidin beads and were abundant enough to be visualized as Coomassie blue bands on an SDS gel.
Blot Showing FP-Biotinylated Proteins in Mouse Plasma
A blot showing the proteins in plasma that became labeled with FP-biotin after treatment of mice with FP-biotin, is shown in Figure 3. Doses of 1 and 5 mg/kg FP-biotin were not toxic to the animals. The intense broad band in the middle of the gel resolved into three bands upon serial dilution of mouse plasma. The top band in the triplet is carboxylesterase (CE), the middle band is albumin, and the bottom band has not been identified.
A band for biotinylated BChE was visible in plasma from mice treated with 5 mg/kg but not 1 mg/kg FP-biotin. This is consistent with our finding that BChE activity was inhibited about 35% after treatment with 5 mg/kg but was not inhibited after treatment with 1 mg/kg FP-biotin.
In addition to albumin, carboxylesterase, and butyrylcholinesterase, mouse plasma contains about eight other biotinylated bands whose protein identity is unknown. The intensity of the band at the top of the gel is higher than that of mouse BChE, suggesting that this protein is more abundant than BChE and that it is highly reactive. The intensity of the other bands is equal to that of BChE or lower. Thus, mouse plasma contains at least 11 proteins that bind OP at physiological conditions, at doses of OP that produce no toxic signs.
About 1–2% of the albumin in mouse plasma is estimated to have bound FP-biotin in a mouse treated with 5 mg/kg FP-biotin ip. However, there is so much more albumin (50 mg/ml) than BChE (0.003 mg/ml) in mouse plasma that albumin consumes a significant amount of FP-biotin. Similar biotinylated band intensities were obtained in Figure 3 for BChE in undiluted plasma and for albumin diluted 1000-fold, suggesting that albumin bound 1000-fold more FP-biotin than was bound by BChE.
Activity-Stained Gels
The gels for Figures 3–5 were nondenaturing polyacrylamide gels. Nondenaturing gels were used because under these conditions the BChE tetramer of 340 kDa separated well from the 67 kDa albumin. This separation is not possible on SDS gels because albumin (67,000 MW) is 10,000 times more abundant in plasma than BChE (85,000 MW), and spreads into a broad band that overlaps with the BChE band, making it impossible to visualize BChE. A second reason for using nondenaturing gels is that nondenaturing gels allow identification of BChE and carboxylesterase based on activity. Figure 4 shows activity with beta-naphthylacetate. The intense band is carboxylesterase (EC 3.1.1.1, CE) in mouse plasma. The bubble below carboxylesterase is albumin. BChE as well as several unidentified proteins also react with beta-naphthylacetate. Figure 5 shows activity of blood proteins with butyrylthiocholine. The BChE tetramer is the intense band near the top of the gel. By aligning bands in Figures 3, 4, and 5 we confirmed the identities of biotinylated albumin, carboxylesterase, and BChE in Figure 3.
Other OP Compete with FP-Biotin for Binding to Albumin
The goal of this experiment was to determine whether other OP bind to albumin. We wanted to know whether binding to albumin was a special property of FP-biotin or whether other OP also bound to albumin. If the OP binding site is a specific tyrosine (Black et al., 1999; Sanger, 1963) then pretreatment with other OP was expected to block binding of FP-biotin. Human plasma was used in this experiment because it does not contain carboxylesterase. The absence of carboxylesterase facilitated interpretation of the results, as there was no confusion between carboxylesterase and albumin. Figure 6 shows that other OP competed with FP-biotin for binding to albumin to varying extents. Pretreatment with chlorpyrifos oxon, echothiophate, malaoxon, paraoxon, methyl paraoxon, diazoxon, dichlorvos, and diisopropylfluorophosphate (DFP) reduced the binding of FP-biotin to human albumin in vitro. The only OP tested that gave no evidence of binding to albumin was tetraisopropylpyrophosphoramide (iso-OMPA).
These results suggest that OP binding to albumin is a general property of OP. It is to be noted that in this report FP-biotin has been demonstrated to bind to albumin from three species: mouse, bovine, and human. The albumins were in mouse plasma, purified bovine serum albumin, and human plasma.
DISCUSSION
OP Labels Proteins That Have No Active Site Serine
Organophosphorus toxicants are well known as inhibitors of serine proteases and serine esterases. Enzymes such as trypsin, chymotrypsin, thrombin, AChE, BChE, acylpeptide hydrolase, and carboxylesterase have a conserved active site serine with the consensus sequence GXSXG. When the active site serine is covalently modified by OP, the enzyme loses activity. Loss of enzyme activity allows one to conveniently measure reactivity with OP.
Proteins with no catalytic activity are a novel class of OP target proteins. They have no active site serine and have been shown to bind OP in vitro, but not in living animals. The advent of quadrupole time-of-flight mass spectrometry has made it possible to positively identify albumin as a protein that binds the OP FP-biotin in living mice.
Experiments with purified proteins have documented covalent attachment of OP not only to serine, but also to tyrosine and histidine. For example, human albumin covalently binds sarin, soman, and DFP at tyrosine (Black et al., 1999; Means and Wu, 1979). Bovine serum albumin is readily phosphorylated by DFP with a stoichiometry of one DFP molecule bound per molecule of albumin (Murachi, 1963). The sequence of the albumin peptide that covalently binds DFP was reported by Sanger as ArgTyrThrLys for human and rabbit albumin, and ArgTyrThrArg for bovine albumin (Sanger, 1963). After the complete amino acid sequences of the albumins was known, these reactive tyrosines were identified as Tyr 411 in human and Tyr 410 in bovine albumin (Meloun et al., 1975; Peters, 1996). Papain binds DFP on tyrosine (Chaiken and Smith, 1969), while rabbit liver carboxylesterase binds DFP on histidine as well as on the active site serine (Korza and Ozols, 1988). Bromelain is not inhibited by DFP, but it does react with DFP, leading to the formation of a fully active, phosphorus-containing enzyme (Murachi, 1963; Murachi et al., 1965). The rat M2 muscarinic acetylcholine receptor covalently binds chlorpyrifos oxon, though the binding site has not yet been identified (Bomser and Casida, 2001; Huff et al., 1994).
Living animals have previously been demonstrated to bind OP to noncholinesterase sites. The tissue distribution of radiolabeled DFP in rabbits (Jandorf and McNamara, 1950), and radiolabeled soman in rats and mouse brain (Little et al., 1988; Traub, 1985) had no correlation with cholinesterase localization. These results have been interpreted to mean that tissue proteins other than cholinesterase are capable of binding OP poisons (Jandorf and McNamara, 1950; Kadar et al., 1985).
Existing Mass Spectrometric Methods for Diagnosis of OP Exposure
Gas chromatography combined with mass spectrometry was used to detect sarin in archived blood samples from victims of the 1995 Tokyo subway attack (Polhuijs et al., 1997). The covalently bound sarin in 0.12 to 0.5 ml of serum was released by incubation with 2 M potassium fluoride at pH 4.0. This method confirmed that the people had been exposed to OP and, in addition, identified the OP as sarin. The mass-to-charge ratios of the released OP fragments were 81, 99, and 125, values characteristic of sarin.
The GC-MS method is a significant advance over simple inhibition assays, but it has limitations. (1) The method relies on being able to release the OP from its covalent attachment site on BChE with potassium fluoride. The release step requires the catalytic machinery of BChE to be intact. Samples that have been stored in ways that denature the BChE cannot undergo fluoride-induced release of OP (Polhuijs et al., 1997). (2) Proteins other than BChE, AChE, and carboxylesterase may not be able to release OP upon treatment with potassium fluoride. Human albumin after phosphylation with sarin is not amenable to reactivation with fluoride ions (Van Der Schans et al., 2004). (3) OP-derivatized protein samples that have lost an alkyl group from the phosphonate in the process called ‘aging’ would not be capable of releasing their OP upon treatment with potassium fluoride.
GC-MS of OP metabolites is the method recommended by the Centers for Disease Control for monitoring potential exposure to nerve agents. Measurement of OP metabolites is limited by the fact that these compounds are rapidly cleared from the body. Sarin metabolites were found in urine on post-exposure days 1 and 3, and in trace amounts on day 7 in a man who had inhaled a dose that made him unconscious (Nakajima et al., 1998). Urine samples from four patients hospitalized for sarin exposure showed that most of the sarin metabolites had cleared within 24 h (Hui and Minami, 2000).
Many of these limitations were overcome in an approach where plasma BChE was purified by affinity chromatography and digested with pepsin. The peptides were separated by liquid chromatography, and the OP-derivatized peptide as well as the OP it carried were identified by electrospray tandem mass spectrometry (Fidder et al., 2002). This method is well suited to detect OP exposure and will become even more valuable if OP-derivatized proteins other than BChE are included in the analysis. A major advantage of using protein-bound OP to diagnose exposure is that the OP-derivatized proteins remain in the circulation for weeks (Cohen and Warringa, 1954; Grob et al., 1947; Munkner et al., 1961; Van Der Schans et al., 2004).
Albumin Properties
Albumin is a nonglycosylated single chain of 585 amino acids folded into three domains by 17 intrachain disulfide bonds. Albumin is the most abundant soluble protein in the body of all vertebrates and is the most prominent protein in plasma. It is synthesized in the liver and discharged into the bloodstream. Of the 360 g total albumin in a human, 40% resides in the blood and 60% in extravascular fluids of tissues (Peters, 1996). The rate of transfer to the extravascular space is about 4.5% per hour. About 40% of the extravascular albumin is in muscle. Albumin is responsible for the colloid osmotic pressure of plasma and for maintenance of blood volume and supplies most of the acid/base buffering action of plasma proteins in extravascular fluids. Albumin binds and transports metabolites and drugs. Analbuminenia, a rare deficiency of albumin, is compatible with nearly normal health in humans. The study of analbuminemia shows that albumin is helpful in coping with stress and in containing environmental and physiological toxins.
Albumin and BChE are Attractive Biomarkers for OP Poisoning
Both albumin and BChE are found in human plasma, a tissue that can be readily sampled for biomarkers of OP exposure. Both proteins bind OP, though the binding to albumin of OP other than FP-biotin has not been demonstrated in vivo. Both proteins have a long half-life in the circulation of humans, 20 days for albumin (Chaudhury et al., 2003; Peters, 1996) and 10–16 days for BChE (Cohen and Warringa, 1954; Jenkins et al., 1967; Munkner et al., 1961; Ostergaard et al., 1988). The long half-life of OP-labeled proteins contrasts with the short half-life of OP and OP metabolites in urine. This provides a significant advantage to a method that uses OP-labeled proteins for diagnosis of OP exposure.
Role of Albumin in OP Toxicity
Binding of OP to albumin could serve to scavenge OP molecules and therefore reduce the amount of OP available for reaction with AChE. The affinity of various OP for albumin is likely to vary, so the effectiveness of albumin as an OP scavenger is likely to depend on the identity of the OP.
Another consideration in evaluating the role of albumin in OP toxicity is competition with drugs for binding to the same site in albumin. Several drugs, including diazepam (Fehske et al., 1979; Peters, 1996) and ibuprofen, bind to the same region of albumin that binds OP (see Tables 3–5 in Peters, 1996). Since the free form of a drug is active, the extent of binding to albumin controls both the effect and duration of the drug. Decreased binding of a drug to albumin can cause toxicity owing to an increase in the concentration of its free form. Thus, a person being treated for intestinal worms with metrifonate (an OP that is converted nonenzymatically to dichlorvos) might be unable to tolerate a standard dose of diazepam or ibuprofen. Or conversely, a person taking ibuprofen, could be intolerant of a standard dose of metrifonate.
ACKNOWLEDGMENTS
Supported by U.S. Army Medical Research and Materiel Command grants DAMD17-01–1–0776 (to O.L.), and DAMD17-01–0795 (to C.M.T.), UNMC Eppley Cancer Center Support Grant P30CA36727, National Science Foundation grant MCB9808372 (to C.M.T.), EPS0091995 (to C.M.T.) and U.S. Army Research, Development & Engineering Command grant W911SR-04-C-0019 (to O.L.) The information does not necessarily reflect the position or the policy of the U.S. Government, and no official endorsement should be inferred. Holly Coughenour at the University of Montana provided the protocol for in-gel reduction, alkylation, and digestion of proteins. We thank the anonymous reviewer for helpful references on the interaction of OP with albumin.
NOTES
REFERENCES
Black, R. M., Harrison, J. M., and Read, R. W. (1999). The interaction of sarin and soman with plasma proteins: The identification of a novel phosphonylation site. Arch. Toxicol. 73, 123–126.
Bomser, J. A., and Casida, J. E. (2001). Diethylphosphorylation of rat cardiac M2 muscarinic receptor by chlorpyrifos oxon in vitro. Toxicol. Lett. 119, 21–26.
Casida, J. E., and Quistad, G. B. (2004). Organophosphate toxicology: Safety aspects of nonacetylcholinesterase secondary targets. Chem. Res. Toxicol. 17, 983–998.
Chaiken, I. M., and Smith, E. L. (1969). Reaction of a specific tyrosine residue of papain with diisopropylfluorophosphate. J. Biol. Chem. 244, 4247–4250.
Chaudhury, C., Mehnaz, S., Robinson, J. M., Hayton, W. L., Pearl, D. K., Roopenian, D. C., and Anderson, C. L. (2003). The major histocompatibility complex-related Fc receptor for IgG (FcRn) binds albumin and prolongs its lifespan. J. Exp. Med. 197, 315–322.
Cohen, J. A., and Warringa, M. G. (1954). The fate of P32 labelled diisopropylfluorophosphonate in the human body and its use as a labelling agent in the study of the turnover of blood plasma and red cells. J. Clin. Invest. 33, 459–467.
Duysen, E. G., Stribley, J. A., Fry, D. L., Hinrichs, S. H., and Lockridge, O. (2002). Rescue of the acetylcholinesterase knockout mouse by feeding a liquid diet; phenotype of the adult acetylcholinesterase deficient mouse. Brain Res. Dev. Brain Res. 137, 43–54.
Ellman, G. L., Courtney, K. D., Andres, V., Jr., and Feather-Stone, R. M. (1961). A new and rapid colorimetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7, 88–95.
Fehske, K. J., Muller, W. E., and Wollert, U. (1979). A highly reactive tyrosine residue as part of the indole and benzodiazepine binding site of human serum albumin. Biochim. Biophys. Acta 577, 346–359.
Fidder, A., Hulst, A. G., Noort, D., de Ruiter, R., van der Schans, M. J., Benschop, H. P., and Langenberg, J. P. (2002). Retrospective detection of exposure to organophosphorus anti-cholinesterases: Mass spectrometric analysis of phosphylated human butyrylcholinesterase. Chem. Res. Toxicol. 15, 582–590.
Grob, D., Lilienthal, J. L., Harvey, A. M., and Jones, B. F. (1947). The administration of di-isopropylfluorophosphate (DFP) to man. I. Effect on plasma and erythrocyte cholinesterase; general systemic effects; use in study of hepatic function and erythropoiesis; and some properties of plasma cholinesterase. Bull. Johns Hopkins Hosp. 81, 217–244.
Huff, R. A., Corcoran, J. J., Anderson, J. K., and Abou-Donia, M. B. (1994). Chlorpyrifos oxon binds directly to muscarinic receptors and inhibits cAMP accumulation in rat striatum. J. Pharmacol. Exp. Ther. 269, 329–335.
Hui, D. M., and Minami, M. (2000). Monitoring of fluorine in urine samples of patients involved in the Tokyo sarin disaster, in connection with the detection of other decomposition products of sarin and the by-products generated during sarin synthesis. Clin. Chim. Acta 302, 171–188.
Jandorf, B. J., and McNamara, P. D. (1950). Distribution of radiophosphorus in rabbit tissues after injection of phosphorus-labeled diisopropyl fluorophosphate. J. Pharmacol. Exp. Ther. 98, 77–84.
Jenkins, T., Balinsky, D., and Patient, D. W. (1967). Cholinesterase in plasma: First reported absence in the Bantu; half-life determination. Science 156, 1748–1750.
Kadar, T., Raveh, L., Cohen, G., Oz, N., Baranes, I., Balan, A., Ashani, Y., and Shapira, S. (1985). Distribution of 3H-soman in mice. Arch. Toxicol. 58, 45–49.
Kadner, S. S., Katz, J., and Finlay, T. H. (1992). Esterase-1: Developmental expression in the mouse and distribution of related proteins in other species. Arch. Biochem. Biophys. 296, 435–441.
Karnovsky, M. J., and Roots, L. (1964). A ‘direct-coloring’ thiocholine method for cholinesterases. J. Histochem. Cytochem. 12, 219–221.
Kidd, D., Liu, Y., and Cravatt, B. F. (2001). Profiling serine hydrolase activities in complex proteomes. Biochemistry 40, 4005–4015.
Korza, G., and Ozols, J. (1988). Complete covalent structure of 60-kDa esterase isolated from 2,3,7,8-tetrachlorodibenzo-p-dioxin-induced rabbit liver microsomes. J. Biol. Chem. 263, 3486–3495.
Little, P. J., Scimeca, J. A., and Martin, B. R. (1988). Distribution of [3H]diisopropylfluorophosphate, [3H]soman, [3H]sarin, and their metabolites in mouse brain. Drug Metab. Dispos. 16, 515–520.
Liu, Y., Patricelli, M. P., and Cravatt, B. F. (1999). Activity-based protein profiling: The serine hydrolases. Proc. Natl. Acad. Sci. U.S.A. 96, 14694–14699.
McDonough, J. H., Jr., and Shih, T. M. (1997). Neuropharmacological mechanisms of nerve agent-induced seizure and neuropathology. Neurosci. Biobehav. Rev. 21, 559–579.
Means, G. E., and Wu, H. L. (1979). The reactive tyrosine residue of human serum albumin: Characterization of its reaction with diisopropylfluorophosphate. Arch. Biochem. Biophys. 194, 526–530.
Meloun, B., Moravek, L., and Kostka, V. (1975). Complete amino acid sequence of human serum albumin. FEBS Lett. 58, 134–137.
Moser, V. C. (1995). Comparisons of the acute effects of cholinesterase inhibitors using a neurobehavioral screening battery in rats. Neurotoxicol. Teratol. 17, 617–625.
Munkner, T., Matzke, J., and Videbaek, A. (1961). Cholinesterase activity of human plasma after intramuscular diisopropyl fluorophosphonate (DFP). Acta Pharmacol. Toxicol. (Copenh.) 18, 170–174.
Murachi, T. (1963). A general reaction of diisopropylphosphorofluoridate with proteins without direct effect on enzymic activities. Biochim. Biophys. Acta 71, 239–241.
Murachi, T., Inagami, T., and Yasui, M. (1965). Evidence for alkylphosphorylation of tyrosyl residues of stem bromelain by diisopropylphosphorofluoridate. Biochemistry 4, 2815–2825.
Nachlas, M. M., and Seligman, A. M. (1949). The histochemical demonstration of esterase. J. Natl. Cancer Inst. 9, 415–425.
Nakajima, T., Sasaki, K., Ozawa, H., Sekjima, Y., Morita, H., Fukushima, Y., and Yanagisawa, N. (1998). Urinary metabolites of sarin in a patient of the Matsumoto sarin incident. Arch. Toxicol. 72, 601–603.
Ortigoza-Ferado, J., Richter, R. J., Hornung, S. K., Motulsky, A. G., and Furlong, C. E. (1984). Paraoxon hydrolysis in human serum mediated by a genetically variable arylesterase and albumin. Am. J. Hum. Genet. 36, 295–305.
Ostergaard, D., Viby-Mogensen, J., Hanel, H. K., and Skovgaard, L. T. (1988). Half-life of plasma cholinesterase. Acta Anaesthesiol. Scand. 32, 266–269.
Peters, T., Jr. (1996). All about Albumin. Biochemistry, Genetics, and Medical Applications. Academic Press Ltd., London.
Polhuijs, M., Langenberg, J. P., and Benschop, H. P. (1997). New method for retrospective detection of exposure to organophosphorus anticholinesterases: Application to alleged sarin victims of Japanese terrorists. Toxicol. Appl. Pharmacol. 146, 156–161.
Pope, C. N. (1999). Organophosphorus pesticides: Do they all have the same mechanism of toxicity J. Toxicol. Environ. Health B Crit. Rev. 2, 161–181.
Ray, D. E., and Richards, P. G. (2001). The potential for toxic effects of chronic, low-dose exposure to organophosphates. Toxicol. Lett. 120, 343–351.
Salvi, R. M., Lara, D. R., Ghisolfi, E. S., Portela, L. V., Dias, R. D., and Souza, D. O. (2003). Neuropsychiatric evaluation in subjects chronically exposed to organophosphate pesticides. Toxicol. Sci. 72, 267–271.
Sanger, F. (1963). Amino-acid sequences in the active centres of certain enzymes. Proc. Chem. Soc. 5, 76–83.
Schwartz, M. (1982). A serine protease activity of human serum albumin towards 4-methylumbelliferyl-guanidinobenzoate (MUGB) and diisopropyl fluorophosphate (DEP): Implications for the use of MUGB reactivity in amniotic fluid in prenatal diagnosis of cystic fibrosis. Clin. Chim. Acta 124, 213–223.
Srivastava, A. K., Gupta, B. N., Bihari, V., Mathur, N., Srivastava, L. P., Pangtey, B. S., Bharti, R. S., and Kumar, P. (2000). Clinical, biochemical and neurobehavioural studies of workers engaged in the manufacture of quinalphos. Food Chem. Toxicol. 38, 65–69.
Sultatos, L. G., Basker, K. M., Shao, M., and Murphy, S. D. (1984). The interaction of the phosphorothioate insecticides chlorpyrifos and parathion and their oxygen analogues with bovine serum albumin. Mol. Pharmacol. 26, 99–104.
Tove, S. B. (1962). The esterolytic activity of serum albumin. Biochim. Biophys. Acta 57, 230–235.
Traub, K. (1985). In vivo distribution of 14C radiolabeled soman [3,3-dimethyl-2-butoxy)-methylphosphorylfluoride) in the central nervous system of the rat. Neurosci. Lett. 60, 219–225.
Van Der Schans, M. J., Polhuijs, M., Van Dijk, C., Degenhardt, C. E., Pleijsier, K., Langenberg, J. P., and Benschop, H. P. (2004). Retrospective detection of exposure to nerve agents: Analysis of phosphofluoridates originating from fluoride-induced reactivation of phosphylated BuChE. Arch. Toxicol. 78, 508–524.
Xie, W., Stribley, J. A., Chatonnet, A., Wilder, P. J., Rizzino, A., McComb, R. D., Taylor, P., Hinrichs, S. H., and Lockridge, O. (2000). Postnatal developmental delay and supersensitivity to organophosphate in gene-targeted mice lacking acetylcholinesterase. J. Pharmacol. Exp. Ther. 293, 896–902.(Eric S. Peeples, Lawrence)