Absolute quantification of free glutathione and cysteine in aquatic insects using isotope dilution and selected reaction monitoring

ORIGINAL PAPER

Absolute quantification of free glutathione and cysteinein aquatic insects using isotope dilution and selectedreaction monitoring

Christopher M. Shuford & Monica D. Poteat &David B. Buchwalter & David C. Muddiman

Received: 20 July 2011 /Revised: 8 September 2011 /Accepted: 12 September 2011 /Published online: 29 September 2011# Springer-Verlag 2011

Abstract A simple and robust isotope dilution massspectrometry-based assay was developed for the determina-tion of free cysteine and glutathione (GSH) in aquatic insects.Several experimental parameters were evaluated and opti-mized to provide specific and sensitive detection of bothcompounds by in situ derivatization with N-ethylmaleimidefollowed by acid alkylation quenching and reverse-phased liquid chromatography coupled with selectedreaction monitoring. For both targets, the assay wasevaluated over a concentration range of 0.313 to 320 μMand was demonstrated to have a quantitative dynamicrange spanning nearly three orders of magnitude, withlower limits of quantification being 0.330 μM for GSHand 0.370 μM for cysteine. Additionally, measurementswere observed to be highly reproducible over the courseof several days. When applied to the analysis of fourdifferent species of insects, large biological variationbetween and within species was observed. Differentfeeding regimens were also tested within two species ofinsects but statistical comparisons revealed no significantdifference in the levels of either compound.

Keywords Glutathione . Cysteine . Selected reactionmonitoring . Absolute quantification . Isotope dilution

Introduction

Insects are a vital and abundant part of nearly all ecosystemsand, as such, the numerous species found therein have thepotential to serve as indicators for environmental stress [1]. Amajor challenge for all organisms that rely upon oxygen formetabolic purposes is the amelioration of reactive oxygenspecies (ROS) [2]. In addition to ROS generation associatedwith normal aerobic metabolism, several environmentalagents are capable of generating ROS, including ultravioletradiation [3], trace metals [4, 5], crude oil [6], and organo-phosphates [7, 8]. In addition, recent work has shown thatdiatoms (a common food source for aquatic insects) canproduce polyunsaturated aldehydes when they senesce [9–11], which are currently receiving attention due to their rolesin oxidative stress, carcinogenesis, and aging. Thus, antiox-idant systems are important in determining performance ofspecies in changing environments.

Free glutathione (GSH) and free cysteine are thought to beparticularly important antioxidants in insects [12] given theirvital roles in maintaining appropriate redox conditions in thecell and controlling free concentrations of potentially toxicmetal ions. Consequently, these compounds may be critical indetermining pollution sensitivity differences among species[13, 14]. This is particularly true in metal contaminatedenvironments because heavy metal exposure results inoxidative stress through generation of reactive oxygen species[15, 16] and thiol scavenging [17]. Though studies of freeGSH and cysteine in aquatic insects are limited [18], thereappears to be potentially significant differences in the relativeroles of antioxidant enzymes and non-enzymatic antioxidantsin this diverse group of insects. It is clear that additional

Published in the 10th Anniversary Issue.

Electronic supplementary material The online version of this article(doi:10.1007/s00216-011-5416-2) contains supplementary material,which is available to authorized users.

C. M. Shuford :D. C. Muddiman (*)W.M. Keck FT-ICR Mass Spectrometry Laboratory,Department of Chemistry, North Carolina State University,Raleigh, NC 27695, USAe-mail: [email protected]

M. D. Poteat :D. B. BuchwalterDepartment of Environmental and Molecular Toxicology,North Carolina State University,Raleigh, NC 27695, USA

Anal Bioanal Chem (2012) 402:357–366DOI 10.1007/s00216-011-5416-2

quantitative information is needed to better understand therole of thiols across species of insects in order to identify thebest molecular markers of oxidative stress.

Determination of GSH and cysteine (free or oxidized) inbiological samples has traditionally been performed usingchromatographic separation followed by UV [19, 20],fluorometric [20, 21], or coulometric detection [22, 23].Due to the generic nature of these detection schemes, theburden of specificity is primarily placed on the chromato-graphic separation, which often results in tedious samplepreparation to eliminate co-eluting contaminants. Massspectrometry (MS) has more recently been applied as thedetection mode to measure free GSH and cysteine in avariety of physiological samples such as saliva [24], urine[25, 26], plasma [25, 27, 28], and bronchoalveolar lavagefluid [29] (for detailed review of such methods see references[30] and [31]). Tandem MS (i.e., MS/MS) detection schemeshave mainly been utilized due to their ability to provide un-paralleled molecular specificity based on unique parent andproduct ion mass combinations, which allow for muchsimpler sample preparation and shorter chromatographicrun times. Selected reaction monitoring (SRM), in particular,has been shown to provide excellent sensitivity for theseassays in the presence of complex matrices because of itscharacteristic improvement in signal-to-noise ratios.

To date, Buchwalter and coworkers have provided the onlyexample of free GSH and cysteine quantification in aquaticinsects [18]. In this report, the authors utilize alkylation withN-ethylmaleimide (NEM) to stabilize the free thiols followedby reversed-phased LC and product ion scanning on a lineartrapping quadrupole to yield MS/MS detection. One majorshortcoming of this work was the use of reserpine as theinternal standard, which is an analog of neither GSH norcysteine and, consequently, likely provides poor quantitativeaccuracy. Herein, we improve upon this method forquantifying both free GSH and cysteine by coupling isotopedilution mass spectrometry with SRM detection. In compar-ison to the previous method utilized in this unique biologicalsystem [18], the detailed method reported here allows forimproved accuracy, precision, and sensitivity through the useof stable isotope-labeled (SIL) internal standards, SRMdetection on a triple quadrupole mass spectrometer, as wellas alkylation quenching. This method was validated andutilized to make quantitative comparisons between severalspecies and between various sample treatments.

Experimental

Materials

Cysteine and glutathione were purchased in their reduced formfrom Sigma-Aldrich (St. Louis, MO) as were their stable

isotope-labeled counterparts, 13C3,15N-cysteine (SIL-cysteine)

and 13C2,15N-glutathione (SIL-GSH). Also purchased from

Sigma-Aldrichwere NEM, iodoacetamide (IAM), ammoniumbicarbonate, and acetic acid. All solvents used in these studieswere of HPLC grade, unless otherwise noted, and werepurchased from Burdick and Jackson (Muskegon, MI).

Insect collection and handling

Hydropsyche sparna (order: Trichoptera) larvae werecollected from the Eno River located in Orange County,NC. Ephemerella catawba and Ephemerella rossi (order:Ephemeroptera), Cheumatopsyche slossonae (order: Tri-choptera), and Diplectrona modesta (order: Trichoptera)larvae were collected from small streams in HaywoodCounty, NC, in Great Smoky Mountains National Park.Insects were transported back to the laboratory in a coolerwith stream water, cobble substrate, and aeration. Acclima-tion occurred for a minimum of 48 h in a walk-in cold room(12.7 °C; 12:12 light/dark photoperiod) in aerated Ameri-can Society for Testing and Materials (ASTM International,W. Conshohocken, PA) artificial very soft water (12 mg/LNaHCO3; 7.5 mg/L CaSO4·2H2O; 7.5 mg/L MgSO4;0.5 mg/L KCl). Voucher specimens were preserved in70% ethanol, and verified by independent taxonomists.Only apparently healthy specimens were used for experi-mentation.

Preparation of alkylated standards

All solutions herein were created from individual 50 mMstock solutions of the natural or SIL compounds. Thesestock solutions were created by weighing out ∼10 mg of thesolid material on an analytical balance and dissolving themin the appropriate volume of water to produce the desiredconcentration. For all solutions, the volume of solventadded was also measured by weight to ensure accuracy ofthe final concentration.

Modified standards were created for method develop-ment work by combining 200 μM cysteine or GSH with anequal volume of 250 μM alkylation reagent (NEM or IAMin 100 mM NH4HCO3, pH 8.0). The reactions wereallowed to proceed for 15 min at room temperature and,subsequently, quenched by acidifying the solution withconcentrated acetic acid (final concentration 1%, v/v). Inpreparing the calibration standards, a series of solutions wasfirst created having concentrations of 0, 0.313, 0.625, 1.25,2.5, 5 10, 20, 40, 80, 160, and 320 μM for both cysteineand GSH. Each solution also contained 10 μM SIL-GSHand 10 μM SIL-cysteine and was made in 100 mMNH4HCO3 (pH 8.0). Concentrated NEM was added toeach standard solution to a final concentration of 500 μMand was allowed to react at room temperature. The reaction

358 C.M. Shuford et al.

was quenched after 15 min by the addition of concentratedacetic acid, as stated above. All solutions of modifiedstandards were stored at −20 °C until required for analysisand were determined to be stable over the course of severalmonths.

Sample preparation

All analyses were performed with fresh/living larvae unlessspecified. Each whole specimen was homogenized briefly(∼10 s) under an atmosphere of nitrogen in 0.5 mL of 25 mMNEM. For quantitative analysis, unmodified SIL-GSH andSIL-cysteine were each added as internal standards immedi-ately prior to homogenization at a final concentration of10 μM each. The homogenate was allowed to sit for 30 min atroom temperature to ensure complete modification of all freethiols and then, unless otherwise noted, acetic acid was addedto a final concentration of 1% (v/v) in order to quench thereactions. The crude homogenate was pelleted (centrifuga-tion at 16,000×g 10 min) and, subsequently, the supernatantwas filtered using Nanosep® 3K Omega MWCO-filters(centrifugation at 16,000×g 10 min). The eluent wascollected and stored at −20 °C until required for LC-SRManalysis. Where noted, the total protein content of the crudehomogenate was measured using a Bradford assay.

Online LC-SRM

A 2D-nanoLC system equipped with an AS1 autosampler(Eksigent, Dublin, CA) was coupled to a TSQ Quantum triplequadrupole mass spectrometer (Thermo Scientific, San Jose,CA) in order to perform online LC-SRM detection. The LCsystem was configured to use only the micro-flow pumps fordirect loading and gradient elution at 8 μL/min. For allsamples, 2 μL was aspirated and loaded directly onto a 150×0.32 mm, Hypersil GOLD aQ column (Thermo Scientific,San Jose, CA) in 3% mobile phase B. Subsequently, thegradient was ramped to 50%mobile phase B over 6.5 min andthen the column was washed with 90% mobile phase B for5.5 min prior to re-equilibrating the column with the initialconditions. The column eluent was subjected to electrosprayionization (ESI) using the Ion-Max API Source in positive ionmode with a spray voltage of 3,500 V, sheath gas flow-rate often units, capillary temperature of 270 °C, and skimmer offsetof 8 V. SRM was performed using the transitions listed inTable 1 and a collision gas pressure of 1.5 mTorr (Argon).Each transition was monitored with a scan width of 1 m/zand scan time of 100 ms.

Data acquisition and processing

Unless otherwise noted, all samples were run in duplicateand in a randomized order with blanks interdigitated to

minimize carryover. For quantitative data sets, calibrationstandards were run in triplicate along with the samples andwere randomized throughout. Peak detection and integra-tion was performed using the Quan Browser in Xcalibur2.0.7 SP1, and the average peak area ratio (natural/SIL)from three independent SRM transitions were used forquantification. Correction of the observed peak areas forSIL-GSH (vide infra) was performed using the faculties ofthe software. Calibration curves were created in JMP 8.0.2(SAS Institute Inc., Cary, NC) using weighted least-squareslinear regression with a weighting factor of 1/x, where x isthe concentration of standard. Given the large dynamicrange of concentrations tested, this form of regression wasutilized to minimize the bias toward higher concentrationstandards. For both target compounds and theircorresponding SIL standards, the relative peak areas oftheir three SRM transitions were monitored throughoutthese studies in order to assess the purity of each peakmeasured. Samples were excluded from quantification if therelative peak areas (measured as ratios) of the threetransitions deviated by more than 20% from that of thestandard compounds (see Table 1).

Results and discussion

Comparison of thiol alkylation reagents

To accurately quantify the reduced forms of GSH andcysteine, it is necessary to stabilize the free thiols andprevent their oxidation during extraction and samplehandling [32, 33]. There are two common thiol alkylationreagents used to accomplish this task, IAM [34] and NEM[35]. In order to test which reagent would provide the bestperformance by LC-MS, an equimolar mixture of themodified species (GSH-NEM, GSH-CAM, cysteine-NEM,and cysteine-CAM) was created and analyzed by LC-MS.SIM was used in this case so that differences in fragmen-tation efficiencies would not bias the observations. Com-pared with the IAM-modified forms, the NEM-modified

Table 1 Optimized SRM transitions

Analyte Precursor m/z(natural/SIL)

Product m/z(natural/SIL)

CE (V) Relativeabundance

GSH-NEM 433.1/436.1 304/307 15 1

287/290 21 0.37

358/358 18 0.05

cysteine-NEM 247.1/251.1 230/233 13 1

201/204 13 0.65

158/158 21 0.58

CE collision energy

Absolute quantification of free glutathione and cysteine 359

species provided improved retention and ion abundance forboth analytes (see Fig. S1 in the Electronic supplementarymaterial (ESM)). Over the course of three replicateinjections, the average peak area for the NEM-modifiedcysteine and glutathione was measured to be 9- and 12-foldlarger than the species modified with IAM, respectively.These results are consistent with recent publications thatdemonstrate more hydrophobic reagents such as NEMallow for improved retention by RPLC and greater ESIresponse compared with hydrophilic reagents like IAM[36–44]. Based on these previous studies and the observa-tions made here, NEM was utilized for all additional methoddevelopment.

Product ion selection

Product ion spectra were acquired for the NEM-modifiedcompounds to determine the optimal fragment ions tomonitor by SRM. Following modification, each reactionmixture was diluted 2-fold with methanol to give a finalconcentration of 50 μM and then independently analyzedby direct infusion ESI. In all cases, fragmentation spectrawere acquired on the singly protonated molecular ion,[M+H+]1+. Shown in Fig. 1 are the product ion spectrumobtained for natural and SIL forms of GSH modified with

NEM. The predominant product ions in both species (304/307 m/z) results from the loss of the neutral glutamateresidue, which is consistent with the fragmentation patternpreviously observed for S-alkylated forms of GSH [26, 29].

The product ion spectrum for NEM-modified cysteine isshown in Fig. 2. Many of the observed fragment ions areconsistent with neutral losses (i.e., –NH3, –H2O, and –CHO2H) previously observed for unmodified cysteine [25,31]; however, several unique product ions not previouslyreported were observed that appear to result from the NEM-modification. For instance, the product ion at 202 m/z(206 m/z in the SIL-cysteine-NEM) most likely results fromcyclization of the modified species and subsequent frag-mentation of the NEM moiety. This cyclized product ioncan further undergo water loss to form the product ion at184 m/z (188 in the SIL-cysteine-NEM). In both cases, the4 m/z difference between the natural and SIL pairs indicatesthat the product ions contain the intact cysteine residuecontaining the heavy-labeled atoms, which support theproposed fragment ions shown. Interestingly, the 184 m/zproduct ion from the natural cysteine-NEM can also resultfrom a second pathway—neutral loss of methanoic acid(–CHO2H) and ammonia (–NH3) from the cysteine residueitself. Within the SIL species, this fragmentation pathwayleads to a product ion, 186 m/z, which only contains two

304

287 433358201

160 180 200 220 240 260 280 300 320 340 360 380 400 420 440

m/z

307

436

358290201

[M+H+]+

[M+H+]+

358

S

NH

O

NH OH

O

HO

O

H2N

H2N

O

NEM

290

201

307

**

*

A)

B)

287

201

304

358

S

NH

O

NH OH

O

HO

O O

NEM

Fig. 1 Displayed are theproduct ion spectra obtainedfor (A) natural GSH and (B)SIL-GSH, both modified withN-ethylmaleimide (NEM). Theproposed fragmentation sites areconsistent with mass shiftsobserved between the naturaland SIL product ions. Note,the product ion at 201 m/z forboth the natural and SIL speciesresults from an internal cleavage

360 C.M. Shuford et al.

heavy-labeled atoms. Consequently, there are two productions (186 and 188 m/z) that correspond to a single-production for the natural species (184 m/z).

Transition optimization and validation

After identifying the abundant fragment ions for NEM-modified GSH and cysteine, breakdown curves were pro-duced for each product ion in order to determine the optimalcollision energy to use for monitoring the SRM transitions(see Fig. S2 in the ESM). A list of all SRM transitions andtheir optimized collision energies are compiled in Table 1. Ina preliminary study to determine which of these transitionscould provide specific detection of the endogenous species, aset of the natural NEM-modified standards were analyzed byLC-SRM along with a pooled extraction sample, createdfrom five extracts of H. sparna prepared using the optimizedprotocol (vide infra), but without the use of SIL internalstandards. Shown in Fig. 3 are the extracted ion chromato-grams of the SRM transitions obtained for the standards andpooled extracts. For the NEM-modified GSH, the SRMtraces obtained with the standards are nearly identical tothose obtained for the endogenous species, which indicatesthese transitions provide specific detection of the endoge-

nous compound. However, contaminants were observed toco-elute with the endogenous NEM-modified cysteine in twoof its SRM traces (247→202 m/z and 247 m/z→184m/z) and,consequently, prevented these transitions from further use.Based on these results, three transitions were selected forboth compounds to use for quantification (see Table 1). Alsolisted for these transitions are their average relative peakareas observed across five replicate injections of the NEM-modified standards. Given the relative abundance of productions is known to remain constant for a pure analyte, therelative peak areas were continuously monitored in subse-quent quantitative measurements to identify the presence ofco-eluting contaminants (vide ante).

Alkylation quenching analysis

Given that NEM is also known to modify free amines underprolonged reaction times [45], alkylation quenching wasevaluated to determine its effect on downstream detection ofthe endogenous species. Using ten H. sparna specimens, twosets of extractions were performed: one set using acetic acidfor alkylation quenching and one set without alkylationquenching. The five biological replicates within each setwere pooled and then analyzed in triplicate by LC-SRM. In

150 160 170 180 190 200 210 220 230 240 250 260

m/z

[M+H+]+

251-15NH3

[M-15NH3]+

233

-H2O

[M-15NH3-H2O]+

215

[M-13CHO2H]+

204

-H2O

-15NH3

186

188

178

158

206

N

O

OSH+

N

O

OS

NH3+O

OH ** *

*

NH

S

OH

O

O

O+

*

** *

[M+H+]+

247-NH3

[M-NH3]+

230

-H2O

[M-NH3-H2O]+

212[M-CHO2H]+

201

-H2O

-NH3

184

174

158 202

N

O

OS

NH3+O

OH

N

O

OSH +

NH

S

OH

O

O

O+

NH

S

OH

O

O

+

NH

S

OH

O

O

+

*

** *

A)

B)

Fig. 2 Shown are the product ion spectra obtained for (A) naturalcysteine and (B) SIL-cysteine, both modified with N-ethylmaleimide.Proposed structures shown for the various product ions are consistentwith the location of isotopically labeled atoms and mass shiftsobserved between the natural and SIL product ions. Note, the ion at

184 m/z in the natural product ion spectrum form is a composite oftwo distinct molecular ions having the same nominal mass. Within theSIL product ion spectra, these two molecular ions are resolved intotwo peaks at 186 and 188 m/z

Absolute quantification of free glutathione and cysteine 361

comparing the two samples, both the NEM-modifiedcysteine and GSH gave significantly larger signal whenacetic acid was used to quench the modification reaction.More specifically, the average peak areas for modifiedcysteine and GSH were 2.0- and 94.8-fold greater, respec-tively, when alkylation quenching was employed. Theseresults demonstrate if the alkylation reaction is allowed toproceed unchecked, non-specific modification of the targetspecies—presumably at the amine groups—will have aprofound negative impact on their detectability. However, itshould be noted the use or absence of an alkylationquenching is unlikely to impact the quantitative accuracy inthese studies as the SIL internal standards will be modified atthe same rate as their endogenous counterparts. Nevertheless,alkylation quenching was adopted for quantitative studies toensure sensitive detection of both target species.

Calibration statistics

For all quantitative analysis, a series of calibration standardswere run along with the samples of interest. Representativestandard curves are shown in Fig. 4. Good linearity and

correlation was obtained across the concentration rangetested for NEM-modified cysteine; however, the slope ofthe GSH calibration curve was significantly lower thanexpected (ideal slope of 0.1 for 10 μM internal standard).Upon closer inspection of the calibration series, the peak arearatios for the upper calibration points (>10 μM) were muchlower than expected while the lower calibration points(<10 μM) gave the expected ratios. Given the small massdifference between the natural and SIL forms of GSH (Δm=3 Da), this observation suggested the natural GSH contrib-utes signal to the SIL-GSH transition due to the isotopicoverlap. To confirm this, the natural GSH standard was runby itself while monitoring both the natural and SILtransitions. For two of the three SIL transitions (436→307and 436→290), a small amount of signal was obtained thatwas approximately 0.7% of that observed for thecorresponding natural transitions (see Fig. S3 in the ESM).Although this seems small, the relative contribution to theSIL transitions becomes more significant when the naturalGSH is in large excess relative to the SIL-GSH. To remedythis, 0.7% of the peak areas measured for the naturaltransitions were subtracted from their corresponding SIL

0

20

40

60

80

100

Rel

ativ

e A

bund

ance

3 4 5 6 7 8 9Time (min)

3 4 5 6 7 8 9Time (min)

Co-eluting Contaminants247m/z 201 m/z

202 m/z

184 m/z

158 m/z

230 m/z

Standards Endogenous

0

20

40

60

80

100

Rel

ativ

e A

bund

ance

4 5 6 7 8 9 10Time (min)

4 5 6 7 8 9 10Time (min)

433 m/z 287 m/z

358 m/z

201 m/z

304 m/z

A)

B)

Fig. 3 Displayed are representa-tive SRM chromatogramsobtained for the light forms ofNEM-modified (A) cysteineand (B) GSH. The left panelsshow the traces obtained withNEM-modified standards, whilethe right panels show the sametraces obtained from a pooledsample of NEM-modifiedextracts. For cysteine, thereis a distinct peak co-elutingwith the target species that iscontaminating the 202 and184 m/z SRM transitions;consequently, these transitionswere excluded from use inquantification. In comparison,the SRM transitions forGSH-NEM are identical betweenthe standards and extracts

362 C.M. Shuford et al.

transitions. When calibration curves were produced using thecorrected peak areas, the ideal slope of 0.1 was achieved forthe individual transitions (Fig. 4C–E) as well as for theaverage of all three transitions (Fig. 4B). Using the correctedstandard curves, the limit of detection (LOD) and limit ofquantification (LOQ) were estimated according to method ofLavagnini and Magno [46]. The LOD in the y-domain (i.e.,the normalized analytical response) was defined as the upper95% confidence interval of the regression intercept while theLOQ in the y-domain was defined as ten standard deviationsabove the regression intercept. The corresponding LOD andLOQ values in the x-domain (i.e., the concentration) weresubsequently calculated using the regression equation. Bythis method, the LOD and LOQ, respectively, werecalculated to be 0.067 and 0.330 μM for GSH, and 0.075and 0.370 μM for cysteine; however, it should be notedthese values are not necessarily obtainable in a real sample asthe calibration standards were prepared in water and do not

undergo matrix-induced ion suppression. In both cases, theseresults demonstrate a quantitative, linear dynamic range ofnearly three orders of magnitude.

0

4

8

12

16

20

24

28

32

0 40 80 120 160 200 240 280 320

Ave

rag

e P

eak

Are

a R

atio

(N

AT

/SIL

)

[Cysteine] (µM)

y = 0.0999x + 0.0087R2 = 0.9991

A)

0.0

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10

y = 0.0984x + 0.0078R2 = 0.9993

y = 0.0908x + 0.0123R2 = 0.9966

B)

0.0

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10

0

4

8

12

16

20

24

28

32

0 40 80 120 160 200 240 280 320

[GSH] (µM)

Ave

rag

e P

eak

Are

a R

atio

(N

AT

/SIL

)

[GSH] (µM)

Pea

k A

rea

Rat

io (

NA

T/S

IL)

y = 0.0967x + 0.0087

y = 0.0851x + 0.0157

C)

0

8

16

24

32

0 80 160 240 320[GSH] (µM)

Pea

k A

rea

Rat

io (

NA

T/S

IL)

y = 0.0948x + 0.0105

y = 0.0836x + 0.0173

D)

0

8

16

24

32

0 80 160 240 320

[GSH] (µM)P

eak

Are

a R

atio

(N

AT

/SIL

)

y = 0.1037x + 0.0072

E)

0

8

16

24

32

0 80 160 240 320

Fig. 4 Standard curves are shown for cysteine (A) and (B) GSH. Thesewere created using the average peak area ratios measured by threeindependent SRM transitions. The insets of (A) and (B) show the low(<10 mM) calibration points for each plot. Standard curves are also shownfor the individual SRM transitions of GSH: (C) 433/436→304/307,

(D) 433/436→287/290, and (E) 433/436→358/358. In each, the blueplots are those obtained without correction for the isotopic overlap in theSIL transition and the red plots are those obtained after correction—note,no correction was necessary for cysteine or for the third transition ofGSH, 433/436→358/358, shown in (E)

Table 2 Interday and intraday statistics

Mean (nmol) SD (nmol) CV (%)

Cys GSH Cys GSH Cys GSH

Day 1 1.82 3.62 0.03 0.002 1.43 0.05

Day 2 2.02 3.59 0.23 0.11 11.31 3.04

Day 3 1.71 3.50 0.02 0.02 1.31 0.71

Day 4 1.91 3.67 0.11 0.05 5.52 1.45

Day 5 1.95 3.83 0.18 0.08 9.19 2.04

Day 6 1.84 3.70 0.00 0.19 0.12 5.00

Day 7 1.80 3.75 0.10 0.23 5.79 6.15

Total 1.87 3.66 0.10 0.11 5.51 2.99

Absolute quantification of free glutathione and cysteine 363

Intraday and intraday variability

To assess the robustness of the measurements, a pooledsample produced from five extracts of E. catawba wasquantified over the course of seven consecutive days. Theabsolute amount of GSH and cysteine measured over thecourse of 1 week are summarized in Table 2. The totalcoefficient of variation over the span of the week was 3.0%for GSH and 5.5% for cysteine. A one-way ANOVA testwas also performed in order to assess if there was anystatistical difference between days. The variance observedwithin and between days was determined to be statisticallyinsignificant for both cysteine (p value=0.3405) and GSH(p value=0.2910), indicating the assay is highly reproduc-ible over the course of a week. This is most likelyattributable to the use of SIL analogs as the internalsstandards, which mimic the temporal decrease in signal ofthe endogenous analyte (see Fig. S4 in the ESM).

Quantification across species and treatments

The method developed here was applied to the analysis of fourspecies of aquatic insects to assess the biological variabilitywithin and between species. For the trichopterans,D. modesta

and C. slossonae, comparisons were also made betweenfasted and fed larvae. All fasted larvae were deprived of foodfor 4 days post-acclimation while fed larvae were givenaccess to natural periphyton for 4 days and then fasted for1 day to purge gut contents prior to analysis. Additionally,the effects of freezing on quantification were evaluated bystoring several intact specimens of E. Catawba for 4 daysat −20 °C prior to their preparation and analysis.

A large degree of variation was observed for the absoluteamounts of free cysteine (Table 3) and GSH (Table 4)measured within each sample set. The CV for bothcompounds were greater than 15% within all species andfeeding regimens, while most were greater 25%. Thisvariability is a result of several factors: biological variation(e.g., body weight), pre-analytical variation (e.g., intakeduring acclimation), and analytical variation (e.g., extrac-tion efficiency). In general, it is possible to define theanalytical variation by performing full replicate analysis ondifferent fractions of the same sample; however, the natureof these specimens precluded such a study, as well asassessment of analyte recovery, because specimen fraction-ation would be impractical. In an attempt to minimize forbiological variation and allow for better comparisonsbetween sample sets, the absolute quantities of each analyte

Table 3 Free cysteine content across sample sets

Species Fresh/frozen Feed Number nmola pmol/mg body massa nmol/mg proteina

Diplectrona modesta Fresh Fasted 5 12.7 (30.8) 740. (31.0) 31.5 (81.7)

Fresh Fed 7 16.4 (27.6) 975 (14.2) 146 (104)

Cheumatopsyche slossonae Fresh Fasted 3 22.3 (44.9) 1,010 (29.4) 21.8 (77.0)

Fresh Fed 2 9.26 (41.0) 489 (68.2) n/ab

Ephemerella catawba Fresh Fasted 10 1.42 (115) 342 (108) 4.1 (89.9)

Frozen Fasted 9 1.71 (60.6) 410. (66.7) 5.70 (121)

Ephemerella rossi Fresh Fasted 8 3.52 (18.0) 723 (15.2) 15.0 (29.2)

a The values are reported as: mean (CV), where CV is expressed as a percentageb Bradford results were inconclusive

Table 4 Free GSH content across sample sets

Species Fresh/frozen Feed Number nmola pmol/mg body massa nmol/mg proteina

Diplectrona modesta Fresh Fasted 6 13.9 (42.8) 808 (28.1) 36.5 (66.9)

Fresh Fed 8 14.2 (22.6) 908 (28.2) 107 (109)

Cheumatopsyche slossonae Fresh Fasted 4 16.2 (38.0) 662 (15.6) 20.9 (81.2)

Fresh Fed 3 15.2 (43.1) 593 (22.1) n/ab

Ephemerella catawba Fresh Fasted 10 4.60 (27.1) 1,130 (24.6) 15.0 (67.8)

Frozen Fasted 10 3.43 (29.6) 796.1 (33.8) 8.95 (73.3)

Ephemerella rossi Fresh Fasted 9 5.31 (15.2) 1,090 (13.4) 22.1 (23.9)

a The values are reported as: mean (CV), where CV is a percentageb Bradford results were inconclusive

364 C.M. Shuford et al.

were normalized either to fresh body mass or total proteincontent (see Table S1 in the ESM for statistics on specimenbody mass and protein content). Normalization to freshbody mass decreased the observed CV in four of sevensamples sets for cysteine and five of seven for GSH (seeTables 3 and 4). On the other hand, normalization to proteincontent actually increased the observed variation for bothcysteine and GSH in nearly all samples sets. Based on thisassessment, all statistical comparisons were made usingvalues normalized to fresh body mass.

Comparing the frozen and fresh specimens of H. catawbarevealed that freezing results in significantly lower levels ofGSH (p value=0.0146). Although the same was not true forcysteine levels (p value=0.6523), these results indicatefreezing may result in oxidation or degradation that can leadto artificially low measurements for GSH. In comparing therelative amounts of these antioxidants across species, C.slossonae contained the most cysteine per milligrams ofbody mass and E. catawba contained the least. Interestingly,the opposite was observed for GSH; of the four speciestested, E. catawba contained the most GSH per mg of bodymass while C. slossonae contained the least. Despite therelatively large variation within each species, the differencesbetween them were still determined to be significant for bothcysteine (p value=0.0037) and GSH (p value=0.0020) by aone-way ANOVA test. Future studies will examine theserelationships in detail across several species.

Lastly, comparisons were performed on the fed and fastedsample sets of D. modesta and C. slossonae to determine ifthe levels of GSH and cysteine were affected by thenutritional status. For the former species, the differencesobserved between to two feeding regimens were determinedto be insignificant for both cysteine (p value=0.0504) andGSH (p value=0.4958). The same conclusion was made inthe latter species for both cysteine (p value=0.1619) andGSH (p value=0.4682). Although it is unknown whether ornot the nutritional state of these animals affects theconcentration of cysteine and GSH, the current resultsindicate any changes are small in comparison to the variationmeasured in each species.

Conclusions

An LC-SRM-based method for quantitative determination ofcysteine and GSH in aquatic insects was developed andoptimized. It was determined that modification of the freethiols by NEM allowed for superior performance of thecompounds by RPLC-MS when compared with the morehydrophilic reagent, iodoacetamide. Additionally, quenchingof the alkylation reaction was determined to be a vital step inthe sample preparation process to prevent non-specificmodification and ensure sensitive detection of the endogenous

compounds. Using SIL internal standards, the final assay forGSH and cysteine produced a dynamic range of approximatelythree orders of magnitude, good detection limits, and goodreproducibility. To validate the assay, cysteine and GSH werequantified across different species and treatments. Largebiological variation was observed across and within eachspecies, and it was determined that the respective levels ofcysteine and GSH in each specimen should be normalized tothe fresh body mass to minimize the observed variation.Despite normalization, the biological variation was such thatno significant difference could be observed between differentfeeding regimens tested in two species. However, comparisonof samples prepared from frozen tissues and fresh tissuesrevealed that freezing could result in an underestimation of thelevels of GSH. Given the simplicity of the sample preparation,this assay could easily be adapted to incorporate a differentiallabeling strategy, similar to that previously reported [26], toallow for simultaneous quantification of both free-thiolcompounds and their oxidized forms. This would allow fordirect comparison of free to oxidized ratios between sampleswithout the need for normalization to body mass or totalprotein content.

Acknowledgments Funding for this project was provided in partfrom the National Science Foundation (IOS-0919614).

References

1. Zaman K, Macgill RS, Johnson JE, Ahmad S, Pardini RS (1995)Arch Insect Biochem Physiol 29:199–209

2. Storey KB (1996) Braz J Med Biol Res 29:1715–17333. Borgeraas J, Hessen DO (2000) J Plankton Res 22:1167–11834. Lannig G, Cherkasov AS, Sokolova IM (2006) Mar Environ Res

62:S79–S825. Sokolova I, Lannig G, Cherkasov A, Flores J (2006) FASEB J 20:

A828–A8286. Reid DJ, MacFarlane GR (2003) Environ Pollut 126:147–1557. Pena-Llopis S, Ferrando MD, Pena JB (2002) Chemosphere

47:485–4978. Pena-Llopis S, Ferrando MD, Pena JB (2003) Aquat Toxicol

65:337–3609. Carotenuto Y, Wichard T, Pohnert G, Lampert W (2005) Limnol

Oceanogr 50:449–45410. Miralto A, Barone G, Romano G, Poulet SA, Ianora A, Russo GL,

Buttino I, Mazzarella G, Laabir M, Cabrini M, Giacobbe MG(1999) Nature 402:173–176

11. Pohnert G (2000) Angew Chem Int Edit 39:4352–435412. Felton GW, Summers CB (1995) Arch Insect Biochem Physiol

29:187–19713. Ahmad S (1995) Arch Insect Biochem Physiol 29:135–15714. Stadtman ER, Oliver CN (1991) J Biol Chem 266:2005–200815. Zaman K, Pardini RS (1996) Toxic Subst Mech 15:151–18116. Stohs SJ, Bagchi D (1995) Free Radical Biol Med 18:321–33617. Sies H (1999) Free Radic Biol Med 27:916–92118. Xie L, Flippin JL, Deighton N, Funk DH, Dickey DA, Buchwalter

DB (2009) Environ Sci Technol 43:934–94019. Giustarini D, Dalle-Donne I, Colombo R, Milzani A, Rossi R

(2003) Free Radic Biol Med 35:1365–1372

Absolute quantification of free glutathione and cysteine 365

20. Chen X, Zhou Y, Peng XJ, Yoon J (2010) Chem Soc Rev39:2120–2135

21. Toyo’oka T, Suzuki T, Saito Y, Uzu S, Imai K (1989) Analyst114:413–419

22. Rabenstein DL, Saetre R (1977) Anal Chem 49:1036–103923. Allison LA, Shoup RE (1983) Anal Chem 55:8–1224. Iwasaki Y, Hoshi M, Ito R, Saito K, Nakazawa H (2006) J Chrom

B 839:74–7925. Rafii M, Elango R, Courtney-Martin G, House JD, Fisher L,

Pencharz PB (2007) Anal Biochem 371:71–8126. Seiwert B, Karst U (2007) Anal Chem 79:7131–713827. Suh JH, Kim R, Yavuz B, Lee D, Lal A, Ames BN, Shigenaga

MK (2009) J Chrom B 877:3418–342728. Hellmuth C, Koletzko B, Peissner W (2011) J Chrom B 879:83–8929. Harwood DT, Kettle AJ, Brennan S, Winterbourn CC (2009) J

Chrom B 877:3393–339930. Iwasaki Y, Saito Y, Nakano Y, Mochizuki K, Sakata O, Ito R,

Saito K, Nakazawa H (2009) J Chrom B 877:3309–331731. Rafii M, Elango R, House JD, Courtney-Martin G, Darling P,

Fisher L, Pencharz PB (2009) J Chrom B 877:3282–329132. Shimada K, Mitamura K (1994) J Chrom B 659:227–24133. Camera E, Picardo M (2002) J Chrom B 781:181–20634. Dickens F (1933) Biochem J 27:1141–1151

35. Friedmann E (1952) Biochim Biophys Acta 9:65–7536. Iwasaki Y, Nakano Y, Mochizuki K, Nomoto M, Takahashi Y, Ito

R, Saito K, Nakazawa H (2011) J Chrom B 879:1159–116537. Shuford CM, Comins DL, Whitten JL, Burnett JC, Muddiman DC

(2010) Analyst 135:36–4138. Shuford CM, Muddiman DC (2011) Expert Rev Proteomics

8:317–32339. Williams DK, Meadows CW, Bori ID, Hawkridge AM, Comins

DL, Muddiman DC (2008) J Am Chem Soc 130:2122–212340. Williams DK, Comins DL, Whitten JL, Muddiman DC (2009) J

Am Soc Mass Spectrom 20:2006–201241. Frahm JL, Bori ID, Comins DL, Hawkridge AM, Muddiman DC

(2007) Anal Chem 79:3989–399542. Walker SH, Papas BN, Comins DL, Muddiman DC (2010) Anal

Chem 82:6636–664243. Bereman MS, Comins DL, Muddiman DC (2010) Chem Commun

46:237–23944. Walker SH, Lilley LM, Enamorado MF, Comins DL, Muddiman

DC (2011) J Am Soc Mass Spectrom. doi:10.1007/s13361-011-0140-x

45. Smyth DG, Konigsberg W, Blumenfeld OO (1964) Biochem J91:589–595

46. Lavagnini I, Magno F (2007) Mass Spectrom Rev 26:1–18

366 C.M. Shuford et al.