|Home | About | Journals | Submit | Contact Us | Français|
Exposure to organophosphorus (OP) agents can lead to learning and memory deficits. Disruption of axonal transport has been proposed as a possible explanation. Microtubules are an essential component of axonal transport. In vitro studies have demonstrated that OP agents react with tubulin and disrupt the structure of microtubules. Our goal was to determine whether in vivo exposure affects microtubule structure. One group of mice was treated daily for 14 days with a dose of chlorpyrifos that did not significantly inhibit acetylcholinesterase. Beta-tubulin from the brains of these mice was diethoxyphosphorylated on tyrosine 281 in peptide GSQQY281RALTVPELTQQMFDSK. A second group of mice was treated with a single sublethal dose of chlorpyrifos oxon (CPO). Microtubules and cosedimenting proteins from the brains of these mice were visualized by atomic force microscopy nanoimaging and by Coomassie blue staining of polyacrylamide gel electrophoresis bands. Proteins in gel slices were identified by mass spectrometry. Nanoimaging showed that microtubules from control mice were decorated with many proteins, whereas microtubules from CPO-treated mice had fewer associated proteins, a result confirmed by mass spectrometry of proteins extracted from gel slices. The dimensions of microtubules from CPO-treated mice (height 8.7 ± 3.1 nm and width 36.5 ± 15.5 nm) were about 60% of those from control mice (height 13.6 ± 3.6 nm and width 64.8 ± 15.9 nm). A third group of mice was treated with six sublethal doses of CPO over 50.15 h. Mass spectrometry identified diethoxyphosphorylated serine 338 in peptide NS338NFVEWIPNNVK of beta-tubulin. In conclusion, microtubules from mice exposed to chlorpyrifos or to CPO have covalently modified amino acids and abnormal structure, suggesting disruption of microtubule function. Covalent binding of CPO to tubulin and to tubulin-associated proteins is a potential mechanism of neurotoxicity.
Organophosphorus (OP) agents are used as pesticides, in jet engine oil as a flame retardant, as chemical warfare agents, and for treatment of Schistosomiasis. The acute toxicity that occurs with high doses of OP is due to inhibition of acetylcholinesterase (AChE). Some people who survive acute exposure describe neurological symptoms that last long after the acute symptoms have abated. The sarin attack in the Tokyo subway occurred in 1995, but 5 years later, some victims of this attack still suffered from blurred vision, easy fatigability, difficulty in concentration, and insomnia (Kawada et al., 2005; Yanagisawa et al., 2006). Magnetic resonance imaging revealed structural changes in the brains of exposed subjects (Yamasue et al., 2007). These long-lasting symptoms could be the consequence of seizure activity and anoxia initiated by acute inhibition of AChE (McDonough and Shih, 1997).
Low doses of OP that do not inhibit AChE have also been associated with neurological dysfunction, including clinically significant extrapyramidal symptoms, anxiety, depression (Salvi et al., 2003), memory loss, and learning disability (Srivastava et al., 2000). The mechanism of low-dose toxicity is not understood. One hypothesis to explain long-lasting neurotoxicity from low doses of OP involves OP modification of proteins in the axonal transport system (Gearhart et al., 2007; Gupta et al., 1997; Prendergast et al., 2007; Terry et al., 2007). The transport system moves organelles from the cell nucleus to the axon termini and back to the nucleus. Hundreds of proteins are present in the axoplasm of neurons (Rishal et al., 2010), suggesting the involvement of a great many proteins in this transport process.
Indirect evidence supports the hypothesis that organophosphorylation of key proteins in the axonal transport system disrupts the transport mechanism in neurons (Gearhart et al., 2007; Terry et al., 2007). Such a disruption could result in loss of synaptic contacts, slow dying back of axon structures, and finally in neuron cell death as has been described by Morfini et al. (2009) to explain neurodegenerative diseases, including Parkinson’s, Alzheimer’s disease, and amyotrophic lateral sclerosis.
Rats treated chronically with low doses of chlorpyrifos have decreased rates of axonal transport measured in sciatic nerves ex vivo, implicating OP modification of tubulin, kinesin, and microtubule-associated proteins in this dysfunction (Gearhart et al., 2007; Terry et al., 2007). The present report focuses on tubulin in brain because in vitro experiments have shown that OP modification of tubulin disrupts polymerization of tubulin into microtubules (Grigoryan and Lockridge, 2009; Prendergast et al., 2007). Disruption of tubulin polymerization has been shown to result in neuron dysfunction, cellular apoptosis, and tissue damage (Conde and Caceres, 2009; Tierno et al., 2009).
Mass spectrometry analysis of a variety of proteins treated with OP in vitro identified covalent binding of OP to tyrosine and lysine residues (Grigoryan et al., 2008, 2009a,b,c). The results suggest that almost any protein can be modified by OP. Previously, OP adducts were thought to form only with the active site serine of enzymes in the serine hydrolase superfamily. In vivo studies have identified OP-tyrosine adducts in the blood of guinea pigs treated with nerve agents and have shown that the adducts are detectable 24 days after exposure at a time when OP-serine adducts of butyrylcholinesterase (BChE) are no longer detectable (Read et al., 2010).
In the present report, we treated mice with the pesticide chlorpyrifos as well as with its active metabolite, chlorpyrifos oxon (CPO). The pesticide chlorpyrifos is relatively harmless to man and rodents because it is rapidly detoxified and excreted. However, a portion of the chlorpyrifos is bioactivated through oxidative desulfuration catalyzed by cytochrome P450 enzymes to form CPO, the toxic agent that inhibits AChE (Sams et al., 2004; Tang et al., 2001) (see Fig. 1).
Our goal was to identify alterations in microtubule structure following in vivo treatment of mice with chronic low doses of chlorpyrifos and with sublethal doses of CPO. This is the first report to identify structural deficits in microtubules of animals treated in vivo with OP and furthermore to provide mass spectrometry evidence for OP labeling of tubulin in mice treated in vivo with chlorpyrifos or CPO.
Chlorpyrifos and CPO (PS-674 and MET-674B; ChemService Inc., West Chester, PA) were stored at −80°C. Guanosine 5′-triphosphate (GTP) sodium salt hydrate (G8877) ≥ 95% HPLC pure powder and 1,4-piperazinediethanesulfonic acid (PIPES, P6757) > 99% pure powder were from Sigma (St Louis, MO). Sequencing grade–modified trypsin (V5113) was from Promega (Madison, WI). Alpha-cyano-4-hydroxycinnamic acid was from Applied Biosystems (MDS Sciex; Foster City, CA). All other chemicals were of analytical grade.
All animal work was conducted in accordance with the Guide for the Care and Use of Laboratory Animals as adopted by the National Institutes of Health. Formal approval to conduct the experiments was obtained from the Institutional Animal Care and Use Committee of the University of Nebraska Medical Center. Adult wild-type mice of strain 129Sv were used in all trials. Mice were bred at the University of Nebraska Medical Center.
Adult female mice treated for 14 days with a low dose of chlorpyrifos (n = 4) (85.6 ± 11.2 days of age) weighed an average of 19.8 ± 3.2 g, while female control mice (n = 4) (90.1 ± 5.8 days of age) weighed an average of 18.2 ± 1.4 g. Adult male mice treated with a single sublethal dose of CPO (n = 3) (72 days of age) weighed an average of 24.6 ± 1.4 g, and control male mice (n = 3) (72 days of age) weighed an average of 23.8 ± 1.9 g. Two female mice were treated with six sublethal doses of CPO over a period of 50.15 h (127 days of age, 23.0 g), and two control female mice (127 days of age, 22.6 g) were treated at the same time points with an equivalent volume of ethanol.
Female mice (n = 4) were injected sc with 3 mg/kg chlorpyrifos dissolved in 3% dimethyl sulfoxide/97% corn oil. Each mouse was treated for 14 consecutive days between 10 and 11 a.m. Surface body temperature, body weight, and observations were made every 5 min through 15 min and at 30-min postdosing. The axial body temperature was measured with a digital thermometer, Thermalert model TH-5, and a surface Microprobe MT-D, Type T thermocouple (Physitemp Instruments Inc., Clifton, NJ). Blood was collected (50 μl) via the saphenous vein prior to dosing each day. On day 15, 24 h after the final dose, the mice were decapitated, plasma was collected, and the brains placed on ice. Brains were weighed and processed according to the tubulin purification protocol. The tubulin was examined for OP modifications.
In three separate experiments, a mouse was injected ip with CPO dissolved in ethanol at a dose of 3.0 mg/kg delivered in 50 μl. A control mouse was injected with ethanol only. Animals were observed for signs of toxicity following injection. About 30-min postinjection, the mice were euthanized and perfused with 50 ml of 0.1M PBS to remove blood from tissues. Tubulin from the brains of these animals was purified and used for nanoimaging and SDS gel electrophoresis. Proteins in the SDS bands were analyzed mass spectrally.
To derive samples for mass spectral analysis of CPO-labeled tubulin, mice (n = 2) were injected ip with CPO dissolved in ethanol at a dose of 2.5 mg/kg delivered in 50 μl. Control mice (n = 2) were injected with ethanol only. Each mouse received six injections at 0, 1, 22, 48, 50, and 50.15 h. Animals were observed for signs of toxicity after each injection. Mice were euthanized 5 min after the last dose by inhalation of carbon dioxide and then perfused via intracardial injection with 50 ml 0.1M PBS solution to remove blood from the tissues. Tubulin from the brains of these animals was examined for OP modifications.
AChE and BChE activities in mouse plasma and brain were measured with 1mM acetylthiocholine and 1mM butyrylthiocholine, respectively, as described (Peeples et al., 2005). Carboxylesterase activity in mouse plasma was assayed with 2.5mM p-nitrophenyl acetate after incubating 3 μl plasma for 5 min in 0.1M potassium phosphate (pH 7.0) containing 12.5mM EDTA, to inhibit paraoxonase, and 0.01mM eserine, to inhibit AChE and BChE. The reaction was started by adding 0.05 ml of 0.1M p-nitrophenyl acetate dissolved in methanol. The total reaction volume was 2.0 ml, and the temperature was 25°C. Absorbance increase at 400 nm was measured on a Gilford spectrophotometer and recorded on a computer interfaced to MacLab/200 (AD Instruments, Sydney, Australia). Activity in micromoles per minute was calculated from the extinction coefficient for p-nitrophenol of 9000M-1cm-1.
Tubulin from mouse brains was purified by two cycles of polymerization using the protocol of Shelanski et al. (1973). Brains from mice were collected and placed on ice. The brains were homogenized in four volumes of ice-cold PM buffer (80mM PIPES, 2mM magnesium chloride, and 0.5mM ethylene glycol tetraacetic acid (EGTA), pH 7.0) and centrifuged at 17,000 × g for 60 min at 4°C. The pellet was discarded, and the clear supernatant was transferred to a clean ultracentrifuge tube. The supernatant, in which no microtubules were present, was mixed with an equal volume of 8M glycerol containing 1.0mM GTP. The samples were incubated for 2 h at 37°C to assemble the microtubules. The solution was centrifuged for 60 min at 100,000 × g at 29°C to pellet the microtubules and remove the polymerization inducers glycerol and GTP. The pellet was suspended in 500 μl PM buffer, depolymerized on ice for 30 min, and centrifuged for 1 h at 4°C to remove particulates. Tubulin in the supernatant was mixed with an equal volume of 8M glycerol containing 1.0mM GTP and incubated for 45 min at 37°C to assemble the microtubules for a second time. The sample was centrifuged for 60 min at 29°C to pellet the microtubules. The pellet was resuspended in 1 ml of PM buffer and maintained at room temperature to preserve the microtubules. The protein concentration of the final microtubule pellet ranged from 3 to 7 μg/μl (depending on the preparation) as measured by absorbance at 280 nm using the NanoDrop spectrophotometer ND-1000 (NanoDrop Technologies Inc.; Thermo Scientific, Waltham, MA). Absorbance at 280 nm was compared to protein concentration estimated from the Pierce bicinchoninic acid protein assay. Both methods gave similar results. Bovine albumin was used as the standard for both assays.
Atomic force microscopy nanoimages were acquired in the Nanoimaging Core Facility at the University of Nebraska Medical Center, codirected by L.S. A portion of the microtubule pellet, prepared by two cycles of assembly, was fixed immediately by addition of glutaraldehyde to 0.25%. The fixed microtubules were stored at 4°C for 12 h before an aliquot was diluted 10-fold with double-distilled water to a protein concentration of 0.5 μg/μl. Ten microliters of the suspension was drawn from the bottom of the tube and placed in the center of a mica chip that had been treated with 1-(3-aminopropyl) silatrane. The detailed procedure for mica surface modification is given in Shlyakhtenko et al. (2003) and Lyubchenko and Shlyakhtenko (2009). The silatrane modification allows the sample to be deposited in a wide range of ionic strengths and pH. After 2-min incubation in a humid chamber, the samples were rinsed with deionized water and dried with argon gas flow. Images were acquired using an MFP-3DAtomic Force Microscope (Asylum Research, Santa Barbara, CA) operated in tapping mode. Pictures were taken from 15 different 5 × 5 μm areas on average. Image processing and measurement of microtubule height and width were performed using Femtoscan software (Advanced Technology Center, Moscow, Russia).
A portion of the microtubule pellet was denatured in 8M urea, reduced with 10mM dithiothreitol, and carbamidomethylated with 50mM iodoacetamide. The sample was diluted to 2M urea and then digested with trypsin at a ratio of 50:1 (wt/wt) at 37°C for 16 h. The salt concentration was reduced by dilution rather than by dialysis because dialysis resulted in loss of up to 95% of the tubulin protein due to sticking to the dialysis membrane. Peptides were desalted and separated by reverse phase HPLC on a 100 × 4.60 mm Phenomenex C18 column eluted with a gradient from 0 to 60% (vol/vol) acetonitrile versus 0.1% (vol/vol) trifluoroacetic acid at a flow rate of 1 ml/min for 60 min on a Waters 625 LC system. Fractions of 1 ml were collected, dried by vacuum centrifugation, and redissolved in 50 μl of 5% acetonitrile and 0.1% formic acid. Fractions that eluted between 5 and 40 min were analyzed by liquid chromatography-tandem mass spectrometry (LC-MSMS) electrospray ionization. The CPO-labeled peptide in Figure 4 was in fraction 38, and the CPO-labeled peptide in Figure 7 was in fraction 30. Details on the method for searching mass spectrometry data to identify labeled peptides are in the “Materials and Methods” section entitled Q-Trap 4000 and LTQ-Orbitrap mass spectrometry. In brief, the mass spectrometry data were compared to the National Center for Biotechnology Information (NCBI) nonredundant database using Mascot software. Diethoxyphosphate was used as a variable modification for Ser, Thr, Tyr, and Lys. Each identified peptide was manually examined to verify that the tandem mass spectral fragmentation (MSMS) spectrum supported the Mascot assignment.
Two gradient polyacrylamide gels (4–30%) were cast in a Hoefer gel apparatus to make 15 × 11–cm gels, 0.75-mm thick, with a single well that spanned the entire gel. Microtubule pellets from control and CPO-treated mouse brains in 400 μl PM buffer were denatured by addition of 80 μl of 6 × SDS gel loading buffer containing 0.2M Tris/Cl (pH 6.8), 10% SDS, 30% glycerol, 0.6M dithiothreitol, and 0.012% bromophenol blue. The samples were boiled for 3 min, and each was loaded onto a gel. About 2 mg of protein were loaded onto each gel. Electrophoresis was for 3000 V/h at 4°C. Gels were stained with Coomassie blue. Protein bands from the upper 75% of each gel were cut into slices. Gel slices were destained, dried, treated with trypsin, and the peptides extracted as previously described (Peeples et al., 2005). The dry samples were dissolved in 80 μl of 5% acetonitrile and 0.1% formic acid and placed in autosampler vials for analysis in the Q-Trap 4000 (MDS Sciex; Applied Biosystems) and the LTQ-Orbitrap (Thermo Scientific) mass spectrometers.
Peptides purified by off-line HPLC or extracted from gel slices were analyzed by LC-MSMS in the Q-Trap 4000 linear ion trap mass spectrometer using a nanospray source with a continuous flow head, at a flow rate of 0.3 μl/min (Li et al., 2009). The four most intense peptides in each cycle of analysis were fragmented by collision-induced dissociation using pure nitrogen at 40 μ Torr and a collision energy of 20–50 V. Collision voltage was determined by the Analyst software based on the size and charge state of the peptide. The data were searched against the NCBI nonredundant database using Mascot v 1.9 (Matrix Science Ltd, London, UK) for tryptic peptides with an added mass of 136 Da from CPO (Perkins et al., 1999). The CPO added mass was put into the UNIMOD public database of protein modifications (http://www.unimod.org) for use with Mascot. Mascot search parameters included peptides with one missed tryptic cleavage, carbamidomethylated cysteine as a fixed modification; diethoxyphosphorylated serine, threonine, tyrosine, and lysine; and oxidized methionine as variable modifications, a peptide mass tolerance of 1.2 Da and a fragment mass tolerance of 0.6 Da. The MSMS fragmentation spectra of peptides identified by Mascot were manually confirmed with the aid of the MS-Product algorithm from Protein Prospector (v 5.3.2 from the University of California Mass Spectrometry Facility). Peptides extracted from gel slices were also analyzed in the LTQ-Orbitrap mass spectrometer as described (Wiederin et al., 2009).
A two-independent sample t-test was performed to determine the statistical significance of differences in the dimensions of the microtubules between the treated and control groups. The data were normally distributed. SPSS software 16.0 (Microsoft Corp., Redmond, WA) was used for statistical analysis.
Mice injected daily for 14 consecutive days with a nontoxic dose of chlorpyrifos (not the oxon), 3 mg/kg sc, showed no signs of toxicity with the exception of hunched posture, which persisted for less than 15 min after treatment. Their body temperature and body weight remained normal (see Fig. 2).
Plasma AChE activity dropped from 0.3 to 0.2 U/ml after the first two doses of chlorpyrifos but rebounded to 0.45 U/ml after 9–14 days of treatment (see Fig. 3A). In contrast, plasma BChE activity steadily declined following each injection of chlorpyrifos (Fig. 3B). This result is consistent with the greater sensitivity of BChE to inhibition by CPO compared to that of AChE (Amitai et al., 1998). Plasma carboxylesterase activity was unaffected (Fig. 3C). Since inhibition of BChE has no health consequences and since AChE activity was not inhibited by the 14th day of treatment, it is concluded that the dosing regimen was nontoxic.
Tubulin purified from the brains of mice treated with a daily low dose of chlorpyrifos was digested with trypsin and analyzed by LC-MSMS. The mass spectrometry data were searched for diethoxyphosphorylated peptides with an added mass of 136 amu on tyrosine, serine, threonine, or lysine. One modified peptide was found. Figure 4 shows the MSMS spectrum for peptide GSQQY281RALTVPELTQQMFDSK from mouse beta-tubulin, covalently modified on tyrosine 281 by diethoxyphosphate. The triply charged parent ion has a mass-to-charge ratio of 854.8, which is consistent with the peptide sequence plus 136 amu from the diethoxyphosphate. The y5 to y11 ions and the internal fragments support the sequence. Internal fragments are at 157.0 (AL minus 28), 268.1 (ALT minus water), 286.2 (LTV minus 28), 367.1 (ALTV minus water), 423.0 (PELT minus water), 569.2 (PELTQ), 653.2 (LTVPEL), 697.2 (PELTQQ), 737.2 (QQMFDS), and 920.3 amu (YRALTVP diethoxyphosphorylated on Y).
Support for modification of tyrosine comes from the internal fragment at 920.3 amu, which corresponds to the sequence Y*RALTVP where Y has an added mass of 136 amu. The threonine in peptide YRALTVP can be ruled out as the site of diethoxyphosphorylation because internal fragments that include this threonine (ALT minus water, LTV minus 28, ALTV minus water, and LTVPEL) do not have an added mass of 136 amu. The second threonine in peptide GSQQY281RALTVPELTQQMFDSK can be ruled out because internal fragments that include this threonine (PELT minus water, PELTQ, and PELTQQ) do not have an added mass of 136 amu. The serine near the C-terminus can be ruled out based on the mass of the internal fragment QQMFDS. The serine near the N-terminal can be ruled out based on the mass of the doubly charged b4 ion. Precedent for the reactivity of this tyrosine in tubulin comes from in vitro experiments where the homologous tyrosine in bovine beta-tubulin, GSQQY281R, was covalently labeled by CPO (Grigoryan et al., 2009c).
Mice treated ip with one dose of 3.0 mg/kg CPO and euthanized 30-min postdosing had toxic signs, including decreased body temperature (average reduction 2.1°C ± 0.5), reduced activity, piloerection, and hunched posture. BChE and AChE activities in the plasma were inhibited 70 and 62%, respectively. Tubulin from the brains of these mice was purified and polymerized to microtubules. The proteins in an aliquot of these microtubules were separated by SDS gel electrophoreses and identified by mass spectrometry. Another aliquot of the microtubule preparation was used for nanoimaging.
Figure 5 shows the SDS gels stained with Coomassie blue from which gel slices were analyzed by mass spectrometry. The banding pattern and staining intensities are similar for the control and treated samples. The original goal of this experiment was to identify proteins modified by CPO. In anticipation that several proteins would be modified by CPO, gel slices from the upper 75% of the gels were prepared for mass spectrometry. The criterion for a positive result was finding the labeled peptide and finding an MSMS spectrum that supported the presence of diethoxyphosphate in the peptide. Labeled peptides are extremely difficult to find. The LC-MSMS analysis identified the proteins in the gel slices but did not identify any labeled peptides. However, the mass spectrometry analysis showed the surprising result that the set of proteins was not identical in the two gels. The microtubules from the CPO-treated mice were missing the six proteins listed at the bottom of Table 1. The missing proteins were well represented in the control microtubule sample but were undetectable in microtubules prepared from CPO-treated mice.
The Coomassie-stained bands in the gels in Figure 5 appeared identical, yet mass spectrometry analysis of the proteins in the gels revealed missing proteins in the microtubules from the CPO-treated mouse. This result can be explained by the fact that any gel slice contains multiple proteins. Coomassie blue staining reports the aggregate composition of the band, favoring the most abundant proteins, but mass spectrometry identifies the individual proteins in a gel band.
The mass spectrometry results in Table 1 are supported by the atomic force microscopy nanoimages in Figure 6. Microtubules prepared from control mouse brain in Figure 6A are decorated with many proteins. In contrast, microtubules prepared from the CPO-treated mouse brain in Figure 6B have few attached proteins.
The microtubules from CPO-treated animals (in Fig. 6B) appeared to be thinner than those from untreated animals (in Fig. 6A). The difference was quantified by measuring the width and height of 100 microtubules at three to four positions along each tubule for a total of 700 measurements. Figure 6C shows that the average width was 64.8 ± 15.9 nm for the control and 36.5 ± 15.5 nm for the CPO-treated microtubules. The average height was 13.6 ± 3.6 nm for the control and 8.7 ± 3.1 nm for the CPO-treated microtubules.
The atomic force microscopy nanoimaging and the mass spectrometry experiments to identify proteins that cosediment with microtubules were conducted by two researchers, each finding similar results. It is unlikely that the missing proteins were removed during purification as both control and CPO-treated samples were handled identically, and the tubulin was not exposed during the purification process to the high-salt conditions, which remove microtubule-associated proteins (Friden and Wallin, 1991).
Mice treated ip with six doses of 2.5 mg/kg CPO showed no toxic signs after injections at 0, 1, 22, 48, and 50 h. Following the injection at 50.15 h, CPO-treated mice had hunched posture, piloerection, decreased activity, ataxic gait, decrease in body temperature (average reduction 4.2°C ± 1.8), and lacrimation. Mice were euthanized 5 min after the 50.15-h dosing. BChE and AChE plasma activities were completely inhibited 5 min after the 50.15-h injection. At this same time point, BChE and AChE activities in the brain were inhibited 50 and 46%, respectively. Brains from these mice were the source of the tubulin that yielded the CPO-labeled peptide in Figure 7.
Collision-induced fragmentation identified a second tubulin peptide covalently modified by CPO. The MSMS fragmentation spectrum in Figure 7 shows peptide NS338NFVEWIPNNVK of beta-tubulin to be covalently modified on serine 338 by CPO. The masses of y and b ions exactly fit the indicated sequence. Internal fragments are at 430 (EWI), 538.3 (IPNNV), 584.3 (S338NFV with diethoxyphosphate on S), and 740 amu (VEWIPN). Support for labeling on serine comes from the internal fragment at 584.3 amu, which corresponds to the sequence S338NFV, where S has an added mass of 136 Da. In addition, the parent ion mass requires the presence of an extra 136 amu component, and there are no other nucleophilic residues in this peptide that could reasonably support reaction with CPO.
Living mice treated with a nontoxic dose of chlorypyrifos or a sublethal dose of CPO have CPO-labeled tubulin in their brains. The identification of two diethoxyphosphorylated tubulin residues (tyrosine 281 and serine 338) is direct evidence that CPO covalently reacts with tubulin in mouse brain.
In previous studies with purified bovine tubulin, 17 peptides from bovine alpha- and beta-tubulin were found to have been labeled with CPO on tyrosine and lysine. Tyrosine 281 was the most reactive residue in beta-tubulin (Grigoryan et al., 2009c). However, no serine was found to be organophosphorylated by CPO in bovine tubulin (Grigoryan et al., 2009c). When live mice were treated with chlorpyrifos or CPO, the labeled tubulin residues were tyrosine 281 and serine 338. No other residues were found to be organophosphorylated in mouse tubulin. It is possible that more sensitive methods will detect additional labeled residues in the future.
Table 2 compares the amino acid sequences of bovine and mouse tubulin peptides that include tyrosine 281 and serine 338. The sequence of peptide GSQQY281RALTVPELTQQMFDSK is conserved in both species, bovine differing from mouse by virtue of a single conservative change (alanine to serine) at the second position from the C-terminus. Tyrosine 281 is present in both bovine and mouse tubulin. Both bovine beta-tubulin peptide NSSY440FVEWIPNNVK and mouse beta-tubulin peptide NS338NFVEWIPNNVK have serine at position 338, but only bovine tubulin has tyrosine at position 340. Both peptides are labeled by CPO. The label in bovine tubulin is on tyrosine 440. The difference in target at this locus could reflect the greater nucleophilicity of tyrosine compared to serine.
Seventeen sites on bovine tubulin were labeled with CPO in vitro, while only two sites were labeled on mouse tubulin in vivo. The larger overall number of labeling sites for bovine tubulin in vitro could be partly explained by the fact that the pure bovine tubulin purchased from Cytoskeleton Inc. had been depleted of microtubule-associated proteins before it was treated with CPO. Tyrosine sites available for reaction in vitro may be unavailable in vivo. The tyrosines might be protected by interaction with other proteins or might already be phosphorylated by kinases (Westermann and Weber, 2003). Prior phosphorylation could mask other potential serine residue targets in vivo.
The homologous sequences in human beta-tubulin are nearly identical to those in bovine and mouse beta-tubulin. Sites of labeling on human tubulin are unknown, but they would be expected to be similar to those identified here.
Thirty-two proteins are listed in Table 1. Eleven of these are forms of tubulin. Two are large proteins (plectin 1 isoform 6 and Titin) with Mowse scores of about 100. These proteins are frequently identified in database searches and may be considered as nonspecific identifications. The remaining 19 proteins appear to copurify with the microtubules. These proteins are recognized as being involved in axonal transport (Maccioni and Cambiazo, 1995; Park et al., 2008; Sakamoto et al., 2008).
Six proteins in Table 1 were undetectable in microtubules prepared from CPO-treated mice. These were heat-shock protein 84 kDa, alpha-internexin, Myosin Va, dynein cytoplasmic 1 light intermediate chain, cytoskeleton-associated protein 5, and microtubule-associated protein 2 isoform 1. Heat-shock proteins function as molecular chaperones, able to direct folding of cytoskeletal proteins, such as alpha/beta/gamma-tubulin, actin, and centractin (Liang and MacRae, 1997). Internexin is among the five major types of intermediate filament proteins expressed in mature neurons. Alpha-internexin, which is highly expressed during mammalian nervous system development, can coassemble with neurofilament proteins and plays a key role in axonal outgrowth (Lariviere and Julien, 2004). Myosin is involved in fast axonal/dendritic transport and can form a complex with kinesin, a microtubule-based motor. This complex allows long-range movement of vesicles within axons and dendrites on microtubules (Langford, 2002). The formation and maintenance of neuronal synapses are dependent on the active transport of material between the cell body and the axon terminal. Cytoplasmic dynein is one of the motors for microtubule-based axonal transport, and it takes part in regulating the dynamic behavior of microtubules (Hestermann et al., 2002; Susalka and Pfister, 2000). Cytoskeleton-associated protein 5 plays a major role in organizing spindle poles. Microtubule-associated protein 2 isoform 1 is a low-molecular weight alternatively spliced variant that is more abundant in embryonic brain than in adult brain (Tucker et al., 1988). Thus, the proteins that are missing from microtubules in CPO-treated mouse brain are related to microtubule assembly, structure, stability, and function. Dissociation of these proteins from the microtubules could account for the abnormally thin microtubules in CPO-treated mice.
Our observation of a decrease in microtubule cosedimenting proteins in response to exposure to CPO is supported by other reports. Prendergast et al. (2007) found that CPO treatment of organotypic hippocampal slice cultures from rat brain results in dramatically reduced levels of microtubule-associated protein 2a and 2b. The fluorescence intensity from bound antibody was reduced up to 85% following 3 days of treatment with 1–10μM CPO. However, levels of alpha-tubulin were not altered by CPO exposure, suggesting that the general structure of the microtubules was not destroyed. A proteomic analysis of cultured N2a neuroblastoma cells found that treatment with 10μM diazinon (an OP agent) increased the levels of some proteins, decreased the levels of others, but left tubulin levels unchanged (Harris et al., 2009). It was suggested that diazinon exposure interferes with cytoskeletal networks, disrupting the structure and function of microtubules, microfilaments, and neurofilaments. Reduced levels of the cytoskeletal protein microtubule-associated protein 1B were found in differentiating rat C6 glioma cells following treatment with CPO, thus demonstrating a potent perturbation effect of CPO on the microtubule network (Sachana et al., 2008).
Microtubules are polymers of tubulin dimers that serve in a structural capacity in neurons and as tracks for transporting cell components from the nucleus to the axons and from the axons back to the nucleus. Microtubule-associated proteins bind to tubulin and serve a wide range of functions, including stabilizing, destabilizing, cross-linking, and facilitating interactions between microtubules and other proteins in the cell. Protein binding to microtubules is regulated through phosphorylation. For example, tubulin phosphorylated by calmodulin-dependent protein kinase is not capable of binding to microtubule-associated protein 2 or of polymerizing to microtubules (Wandosell et al., 1986). Detachment of microtubule-associated proteins from the microtubule is regulated by microtubule affinity–regulating kinase. Phosphorylation of the associated proteins and subsequent detachment causes destabilization and destruction of the microtubule. Hyperphosphorylation of the tau protein causes disassembly of microtubules, a phenomenon seen in Alzheimer’s disease (Alonso et al., 1994).
Phosphorylation of tubulin and tubulin-associated proteins following challenge with OP agents has been studied in vitro, in brain extracts, and in vivo, in rats and hens. Choudhary et al. (2001) found in rats that dichlorvos induced hyperphosphorylation of tubulin and microtubule-associated protein 2, resulting in destabilization of the microtubule assembly. An increase in phosphorylation of microtubule-associated protein 2 was measured in the brain of hens following a single in vivo challenge with diisopropylfluorophosphate (DFP) (Abou-Donia et al., 1993). An increase in the phosphorylation of tubulin, myelin basic protein, and tau was found following in vitro DFP treatment of brain extracts (Gupta and Abou-Donia, 1999). The DFP-treated hen had decreased concentrations of tau protein as well as decreased tubulin polymerization (Gupta and Abou-Donia, 1994).
We propose that the reduction in microtubule cosedimenting proteins for microtubules purified from CPO-treated mice resulted from direct organophosphorylation of tubulin by CPO and perhaps from organophosphorylation of the associated proteins as well. It is also possible that enhanced phosphorylation of tubulin and microtubule-associated proteins by kinases such as CaM kinase II was triggered by CPO exposure (Gupta and Abou-Donia, 1999) and that this contributed to the dissociation of some of the microtubule-associated proteins that we observed. We further propose that the dissociation of these vital microtubule-associated proteins resulted in impaired microtubule structure as demonstrated by the significant reduction in their width and height. Such structural alterations imply disruption of microtubule function. These results support a noncholinergic mechanism for OP neurotoxicity in which long-lasting neurotoxicity is due to OP modification of proteins involved in axonal transport.
The question was raised how disruption of axonal transport provides a better explanation of long-lasting neurotoxicity after OP exposure when it is likely that both the persistence of microtubule disruption and of AChE inhibition are short lived? To address this question, we offer the hypothesis developed by Scott Brady and his associates to explain neurodegenerative diseases (Morfini et al., 2009). Their hypothesis is that alterations in phosphorylation-dependent intracellular signaling mechanisms result in alterations of axonal transport. Dysregulation of axonal transport causes axons to lose connectivity to synapses and to slowly die back. The consequence is a slow neurodegeneration. The abnormality that initiates these events in diseases such as Alzheimer's and Huntington's is a mutation. We propose that phosphorylation abnormalities can also be initiated by covalent binding of OP. The present work demonstrates covalent binding by OP. Other workers have shown that treatment with OP leads to hyperphosphorylation of neurofilaments, tau, calcium/cyclic AMP response element binding protein, tubulin, and microtubule-associated protein 2 (Choudhary et al., 2001; Gupta and Abou-Donia, 1999; Gupta et al., 1997; Schuh et al., 2002) and to disruption of adenylyl cyclase signaling and serotonin-mediated signal transduction (Aldridge et al., 2003; Song et al., 1997). Thus, we propose a mechanism in which low-dose exposure to OP disrupts axonal transport for a sufficient length of time to initiate an irreversible degradation of the neuron, which in turn leads to permanent loss of neuronal function.
U.S. Army Medical Research and Materiel Command (W81XWH-07-2-0034); the National Institutes of Health (NIH) (U01 NS058056 and P30CA36727); NIH (SIG program), the UNMC Program of Excellence, the Nebraska Research Initiative; Mass spectra were obtained with the support of the Mass Spectrometry; Proteomics Core Facility at the University of Nebraska Medical Center.
Nanoimages were obtained with the support of L.S., codirector, and Dr Yuri L. Lyubchenko, director of the Nanoimaging Core Facility at the University of Nebraska Medical Centerin the Department of Pharmaceutical Sciences. The Nanoimaging Core.