Performance of a Novel High Throughput Method for theDetermination of VX in Drinking Water SamplesJennifer S. Knaack,‡,† Yingtao Zhou,‡ Matthew Magnuson,§ Erin Silvestri,§ and Rudolph C. Johnson*,‡

‡Emergency Response Branch, Division of Laboratory Sciences, National Center for Environmental Health, Centers for DiseaseControl and Prevention, 4770 Buford Highway, MS F44, Chamblee, Georgia 30341, United States§U.S. Environmental Protection Agency, National Homeland Security Research Center, 26 W. Martin Luther King Drive, MS NG16,Cincinnati, Ohio 45268, United States

ABSTRACT: VX (O-ethyl-S-(2-diisopropylaminoethyl)methylphosphonothioate) is a highly toxic organophosphorusnerve agent, and even low levels of contamination in water canbe harmful. Measurement of low concentrations of VX inaqueous matrixes is possible using an immunomagneticscavenging technique and detection using liquid chromatog-raphy/tandem-mass spectrometry. Performance of the methodwas characterized in high-performance liquid chromatography(HPLC)-grade water preserved with sodium omadine, anantimicrobial agent, and sodium thiosulfate, a dechlorinatingagent, over eight analytical batches with quality controlsamples analyzed over 10 days. The minimum reportablelevel was 25 ng/L with a linear dynamic range up to 4.0 μg/L.The mean accuracies for two quality control samples containing VX at concentrations of 0.250 and 2.00 μg/L were 102 ± 3%and 103 ± 6%, respectively. The stability of VX was determined in five tap water samples representing a range of water qualityparameters and disinfection practices over a 91 day period. In preserved tap water samples, VX recovery was between 81 and 92%of the fortified amount, 2.0 μg/L, when analyzed immediately after preparation. Recovery of VX decreased to between 31 and45% of the fortified amount after 91 days, indicating hydrolysis of VX. However, the preservatives minimized the hydrolysis rateto close to the theoretical limit. The ability to detect low concentrations of VX in preserved tap water 91 days after spikingsuggests applicability of this method for determining water contamination with VX and utility during environmental remediation.

Since the terrorist attacks of September 11, 2001, awarenessof the potential vulnerability of the water infrastructure of

the United States has increased.1 Both large and small drinkingwater and wastewater utilities could potentially be the target ofa terrorist attack.1 Intentional contamination of drinking watersystems with chemical agents could damage or disrupt drinkingwater systems and result in illness, disease, or death to thepopulation relying on those water sources.2 Chemical warfareagents (CWAs) have been used in the past for warfare andterrorism,3−5 and the potential for water to be a target matrixfor chemical terrorism exists. For instance, during the Iran−IraqWar from 1980 to 1988, Iraq used mustard gas against Iraniantroops and was later reported to have used both mustard gasand organophosphorus nerve agents (OPNAs), such as sarin,when they dropped bombs on their own Kurdish civilians in1988.5,6 Because Iraq was producing these CWAs for warfare,5

there remained a possibility that the agents could beaccidentally or intentionally released or disposed of into thewaters of the Arabian−Persian Gulf, the primary source ofwater for countries along the Gulf.7 In June 1994, sarin wasreleased in Matsumoto, Japan, causing approximately 600exposures and 7 deaths.8 Sarin was detected in a small pond inthe area,8 which suggests the potential for contamination ofwater supplies following a gaseous release of nerve agents. The

OPNA VX (O-ethyl-S-(2-diisopropylaminoethyl) methylphos-phonothioate) is more toxic and environmentally persistentthan sarin,9 and thus, there is a concern that it could beintentionally used to contaminate a drinking water system.10

The U.S. Environmental Protection Agency’s (EPA) WaterLaboratory Alliance (WLA), part of the EnvironmentalResponse Laboratory Network (ERLN), provides a nationallaboratory network with the capacity and capability to analyzesamples from both unintentional and intentional water-relatedcontamination (biological, chemical, and radiochemical con-taminants) in support of monitoring, surveillance, response,and remediation.11 In order to assess the extent ofcontamination of a drinking water system, accurate and rapidanalytical detection methods are needed to analyze samples.2

Limits of detection, data quality objectives and quality controlcriteria, selectivity for the analyte of interest, and highthroughput capability are all important factors to considerwhen selecting a method to analyze a large number of drinkingwater samples following a contamination event.12

Received: December 12, 2012Accepted: January 24, 2013Published: February 12, 2013

Technical Note

pubs.acs.org/ac

This article not subject to U.S. Copyright.Published 2013 by the American ChemicalSociety

2611 dx.doi.org/10.1021/ac3036102 | Anal. Chem. 2013, 85, 2611−2616

Traditional methods for detection of organophosphoruscompounds include high-performance liquid chromatography(HPLC)13,14 or gas chromatography15−17 coupled to massspectrometry detection, but the limits of detection obtainedwith these analytical methods are in the parts per million orparts per billion range, and lower limits are necessary forensuring water sources are remediated to the appropriatecleanup level. In addition, newer research on detection of nerveagents such as VX, using biosensors which utilize a molecularlyimprinted photonic crystal, has been limited because themethod is aimed at detecting the degradation products, not theparent compound, and cannot be used to identify the agentused in the attack.18 Recently, research has been conductedusing biomarkers, such as butyrylcholinesterase (BuChE), toverify human exposure to VX through clinical samples collectedfrom potentially exposed individuals. Examples includetechniques such as reactivation of inhibited BuChE usingfluoridation19−22 and digestion of BuChE to form phosphony-lated nonapeptides followed by analysis using electrosprayHPLC tandem mass spectrometry (HPLC-MS/MS).23

Building on the use of biomarkers for detection of VX, anovel technique which utilized immunomagnetic scavenging(IMSc) was demonstrated for quantitative analysis of VX inwater samples with quantitation down to the part per trillion(ppt) level in a high throughput format.24 In this method,antibody-coated magnetic beads were conjugated to BuChEwhich acted as a scavenger for OPNAs by forming covalentadducts between an active site serine and the OPNA. Adductswere extracted by immunomagnetic separation and enzymati-cally digested to yield small peptides that were analyzed byHPLC-MS/MS.24 This novel method provided analyticalsensitivity and identification of VX down to the ppt level andcan be used to analyze more than 500 samples in a day.24 Thismethod was demonstrated in finished tap water, but the effectsthat drinking water distribution system residual disinfectants,such as chlorine or monochloramine, have on the stability ofVX in water were not determined.The purpose of this study was to investigate the stability of

VX in drinking water samples and to characterize methodperformance for detection of VX using the IMSc method in fivetap water samples with diverse water quality characteristicscollected from across the United States. The effects of sodiumomadine, an antimicrobial preservative, and sodium thiosulfate,a dechlorinating agent, on VX stability in these water sampleswere also investigated, and the holding time for these sampleswas established. In the event that a large number of samples arerequired following a contamination event, stability of theanalyte in the sample matrix is key for accurate analysis and willimprove the chance of detecting the analyte if present.

■ MATERIALS AND METHODSChemicals and Solvents. Ten μg/mL VX (O-ethyl S-[2-

(diisopropylamino)ethyl] methylphosphonothioate, CAS50782-69-9) in isopropanol was provided to the Centers forDisease Control and Prevention by Lawrence LivermoreNational Laboratory (Livermore, CA). VX is a hazardouschemical and should only be handled by trained personnel in awell-ventilated chemical hood. After use, any solutionscontaining VX should be diluted by at least 50% with a 5%sodium hypochlorite solution. Surfaces exposed to VX solutionsshould be thoroughly decontaminated using a 2.5% sodiumhypochlorite solution prior to disposal. Unlabeled andisotopically labeled (13C4D

615N) peptides corresponding to

the active site of BuChE (FGESAGAAS) and unlabeled andisotopically labeled (13C4D

615N) peptides corresponding to VX-

BuChE adducts were synthesized at Los Alamos NationalLaboratory (Los Alamos, NM). Monoclonal antibodies againsthuman BuChE were purchased from Thermo Fisher AffinityBioReagents (Rockford, IL). The following materials werepurchased from Sigma-Aldrich (St. Louis, MO): triethanol-amine buffer solution containing 0.2 M triethanolamine withmagnesium ions, ethylenediaminetetraacetic acid (EDTA), and<0.1% sodium azide as a preservative; phosphate buffered saline(PBS), 10× concentrate diluted to 1× with HPLC-grade waterfor a final concentration of 10 mM sodium phosphate (0.9%NaCl); formic acid; dimethyl pimelimidate dihydrochloride(DMP); 0.2 M tris buffered saline (TBS), 10× solution used asa 10× solution or diluted to 1× with HPLC-grade water for afinal concentration of 20 mM Tris with 0.9% NaCl; phosphatebuffered saline with Tween 20 (PBST) containing NaCl, 2.7mM KCl, and 0.05% Tween 20; and pepsin from porcinegastric mucosa. HPLC-grade water (WS-2211) and acetonitrilewere purchased from Tedia (Fairfield, OH). Pooled humanserum prepared with tripotassium EDTA was obtained fromTennessee Blood Services (Memphis, Tennessee). Protein GDynabeads were purchased from Invitrogen (Carlsbad, CA). Allsamples were prepared in 2 mL deep 96-well plates and 250 μLshallow 96-well plates (VWR, Radnor, PA) and sealed witheither Eppendorf adhesive foils or Easy Pierce 20 μm heatsealing foils (Fisher Scientific, Fair Lawn, NJ). Sample mixingwas carried out using an Eppendorf MixMate plate mixer(VWR, Radnor, PA). Automated sample extraction wasperformed on a KingFisher automated magnetic bead processorfitted with a deepwell head attachment (VWR, Radnor, PA)and a deepwell comb (VWR, Radnor, PA). Extracted sampleswere filtered on a 10 kDa Multiscreen Ultracel-10 Filter Plate(Millipore, purchased from Fisher Scientific, Fair Lawn, NJ)fitted with a 96-well Advion collection plate (Ithaca, NY) usingan Eppendorf 5810R with a microplate bucket rotor (FisherScientific, Fair Lawn, NJ). Sample digestion was performed at37 °C in a Precision water bath (Winchester, VA).

Generation of VX Calibration Curves and QualityControl Materials. Ten μg/mL VX was diluted into TediaHPLC-grade water to create 100 μg/L and 1 μg/L spikingsolutions. These solutions were then spiked into HPLC-gradewater to create calibrators at 0.025, 0.090, 0.310, 1.13, and 4.00μg/L. Quality control (QC) materials were also made inHPLC-grade water at 2.00 μg/L (high-QC) and 0.250 μg/L(low-QC). Unspiked HPLC-grade water served as a negativecontrol.

IMSc Extraction of VX. Magnetic bead preparation andIMSc extraction were performed as previously described.24 ForIMSc extraction, a 2 mL 96-well sample plate was prepared byaliquotting 500 μL of each calibrator, quality control material,negative control, and water sample into individual wells. Eachsample was then buffered with 55 μL of 10× PBST. To extractVX from samples, 100 μL of prepared beads was added to eachsample using the automated bead processor, and the sampleplate was sealed with an adhesive foil and mixed at roomtemperature on a plate mixer for 2 h at 1400 rpm. The beadswere then washed three times in 500 μL of 1× PBST per washusing the automated bead processor. A concentrated pepsinsolution (2 mg/mL in 5% formic acid) was made 30 min priorto dilution and use. Just before use, the concentration enzymesolution was diluted to 251 μg/mL (0.63% formic acid) withHPLC-grade water. The beads were transferred into 75 μL

Analytical Chemistry Technical Note

dx.doi.org/10.1021/ac3036102 | Anal. Chem. 2013, 85, 2611−26162612

aliquots of the diluted pepsin solution in a 96-well plate usingthe bead processor. The plate was then sealed with an adhesivefoil and floated in a 37 °C water bath for 1.5 h to digest theproteins into peptides. After digestion, the plate was brieflycentrifuged to collect condensation. Ten μL of peptide internalstandard was added to each sample, and samples were thentransferred to a 10 kDa filter plate and fitted with a collectionplate. Samples were filtered by centrifugation at 3500 rpm for90 min, and the collection plate was removed and sealed with aheat sealing foil.HPLC-MS/MS and Data Analysis. Samples were analyzed

for the presence of unadducted BuChE and VX-adductedBuChE nonapeptides as previously described24 using a WatersnanoAcquity HPLC fitted with a UPLC mixer kit for 1 mmcolumns (Milford, MA). HPLC mobile phases and gradient andmass spectrometer settings were used as previously described.24

Briefly, 10 μL of sample was injected on a 50 × 1 mm AquasilC-18 column (Thermo Scientific, Waltham, MA) maintained at25 °C and a flow rate of 75 μL/min. The needle was washedwith 200 μL of a 50% mixture of water/methanol followed by600 μL of a wash containing 2% acetonitrile in water.Chromatographic separation was performed with mobilephase A (0.1% formic acid in water) and mobile phase B(0.1% formic acid in acetonitrile) as follows: 2−35% B in 1.3min, 0.8 min hold, and 35−45% B in 0.3 min. Analysis of VX-BuChE adducts was performed on an API 4000 Q Trapquadrupole-linear ion trap mass spectrometer (AppliedBiosystems, Foster City, CA) using Analyst 1.4.2 software(Applied Biosystems/MDS Sciex, Foster City, CA). Quantita-tion and confirmation ion transitions from m/z 902 → 778 andm/z 902 → 673, respectively, were monitored for VX-BuChEadducts. Only one transition, 913 → 785, was monitored forthe corresponding internal standard.Quantitation of the VX-BuChE peptide was performed using

a calibration curve. The curve was generated by plotting theratio of the analyte ion transition peak area to the internalstandard ion transition peak area against the expectedconcentration with 1/x weighting for each calibrator. Theratio of the confirmation ion transition area to the quantitationion transition area, or confirmation ratio, was calculated foreach calibrator. For a peak to be identified as the respectiveanalyte, its confirmation ratio had to be within 30% of the meanof the confirmation ratios of the calibration solutions. Qualitycontrol samples were evaluated using Westgard rules.25

Statistical analysis of data was performed using both MicrosoftExcel 2010 (Microsoft Corporation, Redmond, WA) and SASstatistical software (SAS Institute, Inc., Cary, NC).Storage Stability Studies with Finished Tap Water.

The stability of VX was determined in HPLC-grade water andin finished tap water samples that were collected from fivedifferent sources throughout the U.S. To determine optimalsample preservation additives, VX stability was also measured inthe presence of sodium omadine (64 mg/L), a microbialgrowth inhibitor, with either ascorbic acid (100 mg/L) orsodium thiosulfate (80 mg/L) to quench residual oxidants inthe samples. Residual oxidants, such as chlorine andmonochloramine, are added to drinking water to inhibitmicrobial regrowth but may interact with analytes and interferewith their determination. Samples were generated by spikingtap and HPLC-grade water to a concentration of 2.0 μg/L VX,and 500 μL aliquots of each sample were transferred into 96-well plates and stored at 4 °C for up to 91 days in the dark.Samples were analyzed for VX approximately 15 min and 5 h

after preparation and on days 7, 14, and 91. To quantitatestorage stability samples, the concentration of VX wasdetermined by relating the MS response of the quantitationion to the MS response of the internal standard. Specifically, aresponse factor (RF) for the analyte was calculated fromidentical concentrations of analyte extracted from HPLC-gradewater:

= A Q A QRF /x is,x is,x x (1)

where Ax = the integrated abundance of the VX quantitationion in HPLC grade water; Ais,x = integrated abundance of theinternal standard in HPLC grade water; Qx = quantity of VX inconcentration units in HPLC grade water; Qis,x = quantity ofinternal standard in concentration units in HPLC grade water.The concentration of VX in the sample (Cy) was then

calculated using:

=C A Q A/ (RF)y y is,y is,y

where Ay = integrated abundance of the quantitation ion of VXin the sample; Ais,y = integrated abundance of the internalstandard in the sample; Qis,y = quantity of internal standard inconcentration units in the sample. The concentration of VX ispresented as a percentage of VX fortified into the sample.

■ RESULTS AND DISCUSSIONMethod Characterization in HPLC-Grade Water.

Method characterization was performed over the course of 8analytical batches over 10 days in HPLC-grade watercontaining 80 mg/L sodium thiosulfate and 64 mg/L sodiumomadine. Each analytical batch contained five calibrators, twoquality control samples, and a blank. The retention time for theVX-BuChE adduct analyte and the corresponding internalstandard was 2.2 min. Analyte response was linear over theentire calibration range (25 ng/L−4.0 μg/L) with a coefficientof determination, R2, of at least 0.9998 for all calibration curves.The lowest standard, 25 ng/L, was chosen as the minimumreportable level (MRL). The mean measured concentration forthe lowest standard was 100 ± 6% of the actual concentration,and mean accuracies ranged between 99 and 102% of the actualconcentration for all calibrators. Other method performancecharacteristics are summarized in Table 1. The fortified

concentrations for the two quality control pools, low-QC andhigh-QC, were 0.25 and 2.0 μg/L VX, respectively. The meanmeasured accuracy for VX was 102 ± 3 and 103 ± 6% of theactual concentration for low-QC and high-QC samples,respectively. Individual measurement accuracies ranged be-

Table 1. Method Characterization Summary for VX inHPLC-Grade Water with Sodium Omadine (64 mg/L) andSodium Thiosulfate (80 mg/L) (n = 8 over 10 days)a

sample

fortifiedconc.(μg/L) measured conc. (%)

MDLb

(ng/L)MRLc

(ng/L)

upperlimit

(μg/L)

low-QC

0.25 102 ± 3

5.6 25 4.0high-QC

2.0 103 ± 6

aThe measured concentration is expressed as a percent of that fortified(x ± σn−1).

bMDL is the method detection limit. cMRL is theminimum reportable level for the method.

Analytical Chemistry Technical Note

dx.doi.org/10.1021/ac3036102 | Anal. Chem. 2013, 85, 2611−26162613

tween 98 and 106% and 94−112% for the low-QC and high-QC pools, respectively.The method detection limit (MDL) was calculated on the

basis of the procedure discussed at 40 CFR26 and is defined asthe minimum concentration of a substance that can bemeasured and reported with 99% confidence that the analyteconcentration is greater than zero.26 The MDL is calculated as

σ= × α− − =tMDL n( 1,1 0.99)

where σ = standard deviation of replicate analyses, t(n−1, 1−α=0.99) = Student’s t value for the 99% confidence levelwith n − 1 degrees of freedom, and n = number of replicates.The calculated MDL for VX in preserved HPLC-grade waterwas 5.6 ng/L (Table 1). The MRL, defined as theconcentration of the lowest standard, was 25 ng/L (Table 1).The highest reportable level, defined as the concentration of thehighest calibrator (upper limit), was 4.0 μg/L (Table 1).Samples with VX concentrations greater than 4.0 μg/L saturateactive sites on the magnetic beads and must be diluted prior toanalysis.24 The MRL and measured accuracies for high- andlow-QC samples are consistent with previously published valuesfor VX in unpreserved HPLC-grade water indicating methodrobustness.24

Storage Stability Studies. In tap water, three main typesof reactions that contribute to analyte loss during storage areoxidation by residual oxidants, microbial degradation, andhydrolysis. Chorine or chloramine is added to water to preventmicrobial regrowth in finished water. A dechlorinating agentcan be added to eliminate residual oxidant in a sample collectedfor analysis. However, once eliminated, the sample can besubject to microbial regrowth even if refrigerated. To increaseVX stability in finished tap water samples, this study utilizedsodium thiosulfate which electrochemically reduces the residualoxidant and also sodium omadine to retard microbial regrowth.Hydrolysis of VX during storage is complex because VXundergoes a pH dependent hydrolysis at an observed rate,kobsd:

27

=+

++

+

+

+

+ +−

+

k kK H

kK

Kk

KK

[H ][ ]

[ ][H ]

[OH ]

[ ][H ]

obsd H Oa

a

aOH

a

a

2

where kH2O = 2.9 × 10−4 h−1, k = 1.5 × 10−2 h−1, kOH = 30 M−1

h−1, and Ka = 2.5 × 10−9. When kobsd is plotted against pH, it isevident that kobsd remains essentially unchanged below pH 6.Thus, VX is inherently unstable in aqueous solutions but has aconsiderable half-life that allows for its measurement if thehydrolysis rate is minimized.VX concentration was measured in five tap waters as a

function of time from 30 min to 91 days. Analyte peak heightswere at least 4.0 × 104 counts per second, representing a signal-to-noise ratio of approximately 4.0 × 104, for all tap watersextracted 30 min after preparation (Figure 1). Observed firstorder rate constants were estimated by plotting theconcentration versus the logarithm of the reaction time.These rate constants are summarized in Table 2, along withwater quality characteristics. The rate constants for degradationare similar within the error in the calculation suggesting that thepreservation scheme investigated provides consistent resultsacross the water sources investigated, which were chosen torepresent a range of water quality parameters encountered. The

standard error in the rate constant reflects errors in applying afirst order rate assumption to the degradation reaction. VXundergoes many reactions, so the first order rate assumption isfor convenience but is sufficient for the present purpose, i.e., toselect a preservation approach for water samples contaminatedwith VX. Indeed, Table 1 suggests that some reactions occurimmediately after fortifying the water samples which reduce theanalytical signal after sample processing. The half-life of VX

Figure 1. Representative chromatogram of VX quantitation (A, m/z902 → 778) and confirmation (B, m/z 902 → 673) ion transitionsfrom a tap water (source water 2 from Table 2) preserved with sodiumomadine and sodium thiosulfate spiked to a concentration of 2.0 μg/LVX and extracted 30 min after preparation.

Table 2. Observed Rate Constants and Percentage of VXMeasured in Several Finished Tap Waters at RoomTemperaturea

type of source water(disinfectant) kobsd,10

−4 h−1 b

measured VXimmediately afterpreparation (%)

measured VX91 days after

preparation (%)

ground water 1c

(chlorine)3.5 ± 0.7 81 ± 1 31 ± 8

surface water 2d

(chlorine)3.8 ± 0.8 85 ± 2 30 ± 11

surface water 3e

(monochloramine)3.8 ± 0.8 92 ± 2 33 ± 6

surface water 4f

(monochloramine)3.8 ± 0.8 85 ± 3 39 ± 4

surface water 5g

(chlorine)3.8 ± 0.8 92 ± 4 45 ± 5

aThe type of source water, along with the residual disinfectant used inthe water system, is identified in the table, and water qualitycharacteristics of the finished tap waters are included as footnotes.Error in rate constant is the standard error of the regression. Thepercentage of VX measured in tap waters extracted 30 minutes afterpreparation and after 91 days of holding in the presence of sodiumomadine and sodium thiosulfate as preservatives. Percentage isreported as x ± σn‑1 for n = 3. bObserved degradation rate constantswere calculated from three replicates per time point. cTotal organiccarbon (TOC) not detected in well-field; pH 7.6; hardness, 500 mg/L;chlorine, 0.2−0.4 mg/L (monthly averages). dTOC, 1.0; pH 8.5;hardness, 130 mg/L; chlorine, 0.8 mg/L (monthly averages). eTOC,2.3; pH 7.4; hardness, 190 mg/L; monochloramine, 3.4 mg/L(monthly averages). fTOC, 7.6 mg/L; pH 9.2; hardness, 65 mg/L;monochloramine, 2.4 mg/L (monthly averages). gTOC, 0.3 mg/L; pH8.9; hardness, 17 mg/L; chlorine, 1.3 mg/L (monthly average).

Analytical Chemistry Technical Note

dx.doi.org/10.1021/ac3036102 | Anal. Chem. 2013, 85, 2611−26162614

under these conditions is about 90 days, suggesting thatlaboratories may hold the samples for extended periods of timeprior to analysis. Most drinking water methods target a holdingtime of about 28 days, regardless of analyte stability.28

■ CONCLUSIONSWhile remediation goals are inherently site specific, theperformance data suggest the applicability of this methodduring water contamination incidents involving the analyte.Risk based criteria (RBC) have been identified from existinghealth benchmarks to serve as analytical targets whendeveloping analytical methods for various chemicals, and theRBC for VX in drinking water is 0.021 μg/L.29 Even if VXsignal loss from 91 days of sample holding is factored in, theMDL is below this analytical target level, suggesting that themethod may be applicable during response and remediationactivities. Similar method performance, in terms of detectionlimit, accuracy, and precision was not reported in the literaturesummarized in the introduction. Importantly, for a wide scalecontamination incident in which many hundreds or thousandsof samples may be generated, the method is automated,enabling simultaneous processing of a large number of watersamples with minimal effort while reducing the potential fortechnical errors to occur.As a broader implication for designing a VX sampling and

analysis plan, it is worth noting that the preservation approachemployed in this method resulted in observed degradation rateconstants (Table 1) which are similar within experimental errorto slowest VX degradation reported in the literature.27 For VX,shorter holding times correspond to less loss of analytical signalfrom VX. However, longer holding times may be utilizeddepending on the goals of analysis. Indeed, since thispreservation scheme results in perhaps the slowest degradationpossible, regardless of the method employed, analytical goalsduring remediation of contaminated areas must take intoaccount the inevitability of loss of VX signal.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: RJohnson6@cdc.gov. Fax: (404) 638-5309.Present Address†Department of Pharmaceutical Sciences, Mercer University,3001 Mercer University Dr., DV- 114, Atlanta, GA 30341, U.S.Author ContributionsThe manuscript was written through contributions of allauthors. All authors have given approval to the final version ofthe manuscript.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe U.S. Environmental Protection Agency (USEPA), throughits Office of Research and Development, collaborated with theCenters for Disease Control and Prevention in the researchdescribed herein under EPA IA# DW75-92304801. Thiscontent has been peer and administratively reviewed and hasbeen approved for publication as a joint USEPA and CDCdocument. Note that approval does not signify that thecontents necessarily reflect the views of the USEPA, the CDC,the Public Health Service, or the U.S. Department of Healthand Human Services. Reference herein to any specificcommercial product, process, or service by trade name,

trademark, manufacturer, or otherwise does not necessarilyconstitute or imply its endorsement, recommendation, orfavoring by the United States government. The views andopinions expressed herein do not necessarily state or reflectthose of the United States government and shall not be used foradvertising or product endorsement purposes.

■ REFERENCES(1) Copeland, C.; Cody, B. CRS Report for Congress: Terrorism andSecurity Issues Facing the Water Infrastruture Sector (Updated),RL32189; Congressional Reseach Service: Washington, D.C.; 2006.(2) USEPA. A Water Security Handbook: Planning for and Respondingto Drinking Water Contamination Threats and Incidents, 817-B-06-001;Office of Ground Water and Drinking Water: Washington, D.C., 2006.(3) Talmage, S. S.; Munro, N. B.; Watson, A. P.; King, J. F.;Hauschild, V. In Chemical Warfare Agents: Toxicology and Treatment;Marrs, T. C., Maynard, R. L., Sidell, F. R., Eds., John Wiley & Sons,Ltd: Chichester, UK, 2007; p 91.(4) Salem, H.; Ternay, A., Jr.; Smart, J. K. In Chemical WarfareAgents: Chemistry, Pharmacology, Toxicology, and Therapeutics;Romano, J. A., Jr., Lukey, B. J., Salem, H., Eds.; CRC Press: BocaRaton, FL, 2007; pp 1−20.(5) Harris, R.; Paxman, J. A higher form of killing: The Secret History ofChemical and Biological Warfare; Random House Trade Paperbacks:New York, NY, 2002.(6) Ember, L. Chem. Eng. News 1993, 71, 8−9.(7) Khordagui, H. K. Environ. Int. 1995, 21, 363−379.(8) Morita, H.; Yanagisawa, N.; Nakajima, T.; Shimizu, M.;Hirabayashi, H.; Okudera, H.; Nohara, M.; Midorikawa, Y.; Mimura,S. Lancet 1995, 346, 290−293.(9) Munro, N. Environ. Health Perspect. 1994, 102, 18−38.(10) CDC. U.S. Department of Health and Human Services, Centersfor Disease Control and Prevention; http://emergency.cdc.gov/agent/vx/basics/facts.asp, 2006; accessed May 2012.(11) U.S. EPA. http://water.epa.gov/infrastructure/watersecurity/wla/upload/WLAResponsPlan_November2010.pdf, 2010; accessedMay 2012.(12) U.S. EPA. http://www.epa.gov/sam, 2012.(13) Ingelse, B. A.; van Dam, R. C.; Vreeken, R. J.; Mol, H. G.;Steijger, O. M. J. Chromatogr., A 2001, 918, 67−78.(14) Xu, Z. L.; Wang, Q.; Lei, H. T.; Eremin, S. A.; Shen, Y. D.;Wang, H.; Beier, R. C.; Yang, J. Y.; Maksimova, K. A.; Sun, Y. M. Anal.Chim. Acta 2011, 708, 123−129.(15) Wong, J. W.; Webster, M. G.; Halverson, C. A.; Hengel, M. J.;Ngim, K. K.; Ebeler, S. E. J. Agric. Food Chem. 2003, 51, 1148−1161.(16) Sathe, M.; Merwyn, S.; Ghorpade, R.; Agarwal, G. S.; Rao, M.K.; Rai, G. P.; Kaushik, M. P. J. Hazard. Mater. 2011, 192, 1720−1728.(17) Fang, G.; Chen, W.; Yao, Y.; Wang, J.; Qin, J.; Wang, S. J. Sep.Sci. 2012, 35, 534−540.(18) Liu, F.; Huang, S.; Xue, F.; Wang, Y.; Meng, Z.; Xue, M. Biosens.Bioelectron. 2012, 32, 273−277.(19) Degenhardt, C. E.; Pleijsier, K.; van der Schans, M. J.;Langenberg, J. P.; Preston, K. E.; Solano, M. I.; Maggio, V. L.; Barr, J.R. J. Anal. Toxicol. 2004, 28, 364−371.(20) Holland, K. E.; Solano, M. I.; Johnson, R. C.; Maggio, V. L.;Barr, J. R. J. Anal. Toxicol. 2008, 32, 116−124.(21) Solano, M. I.; Thomas, J. D.; Taylor, J. T.; McGuire, J. M.;Jakubowski, E. M.; Thomson, S. A.; Maggio, V. L.; Holland, K. E.;Smith, J. R.; Capacio, B.; Woolfitt, A. R.; Ashley, D. L.; Barr, J. R. J.Anal. Toxicol. 2008, 32, 68−72.(22) Polhuijs, M.; Langenberg, J. P.; Benschop, H. P. Toxicol. Appl.Pharmacol. 1997, 146, 156−161.(23) Fidder, A.; Hulst, A. G.; Noort, D.; deRuiter, R.; van der Schans,M. J.; Benschop, H. P.; Langenberg, J. P. Chem. Res. Toxicol. 2002, 15,582−590.(24) Knaack, J. S.; Zhou, Y.; Abney, C. W.; Prezioso, S. M.;Magnuson, M.; Evans, R.; Jakubowski, E. M.; Hardy, K.; Johnson, R.C. Anal. Chem. 2012, 84, 10052−10057.

Analytical Chemistry Technical Note

dx.doi.org/10.1021/ac3036102 | Anal. Chem. 2013, 85, 2611−26162615

(25) Westgard, J. Basic QC Practices, Third ed.; Westgard QC, Inc.:Madison, WI, 2010.(26) CFR, Part 136--Guidelines Establishing Test Procedures for theAnalysis of Pollutants; U.S. Government Printing Office: Washington,D.C., 2011.(27) Epstein, J.; Callahan, J. J.; Bauer, V. E. Phosphorus 1974, 4, 157−163.(28) NEMI. National Environmental Methods Index; http://www.nemi.gov, accessed July 2012.(29) EPA. Risk-Based Criteria to Support Validation of DetectionMethods for Drinking Water and Air, EPA/600/R-08/021, 2008;http://cfpub.epa.gov/si/si_public_file_download.cfm?p_download_id=498247, accessed August 2012.

Analytical Chemistry Technical Note

dx.doi.org/10.1021/ac3036102 | Anal. Chem. 2013, 85, 2611−26162616