ORIGINAL PAPER

Selective determination of mono- and dihydroxylated analogsof polybrominated diphenyl ethers in marine spongesby liquid-chromatography tandem mass spectrometry

Yoshihisa Kato & Syohei Okada & Kazutaka Atobe &

Tetsuya Endo & Koichi Haraguchi

Received: 24 March 2012 /Revised: 17 May 2012 /Accepted: 20 May 2012 /Published online: 8 June 2012# Springer-Verlag 2012

Abstract A number of bioactive brominated secondarymetabolites, including hydroxylated polybrominateddiphenyl ethers, have been isolated from algae, sponges,and bacteria. In the present study, a screening method usingliquid-chromatography tandem mass spectrometry was de-veloped for the identification and selective determination ofdihydroxy (diOH), hydroxy-methoxy (OH-MeO), anddimethoxy (diMeO) analogs of tetra- to hexa-BDEs in ma-rine biota. In negative atmospheric pressure chemical ioni-zation (APCI) mode, diOH and OH-MeO analogs providedintense [M−H]− ions, whereas diMeO analogs providedcharacteristic [M−Br+O]− and [M−CH3]

− ions. This enabledthe diOH-, OH-MeO-, and diMeO-PBDEs to be distin-guished using selected reaction monitoring transitions inthe APCI source. Recoveries of 2′-OH-6-MeO-2,3′,4,5′-tet-ra-BDE in spiked marine samples were 84±5 %, with a limitof quantification at 9.1 ng mL−1 (signal-to-noise ratio010).The developed method was used to analyze two spongespecies collected from Palau, Micronesia; the concentrationratio of diOH-tetra-BDE:OH-MeO-tetra-BDE was 10:1 for

the Lamellodysidea sp., whereas it was 1:30 for the Call-yspongia sp.

Keywords LC/MS/MS . APCI . Dihydroxy-PBDE .

Hydroxy-methoxy-PBDE . Dimethoxy-PBDE .Marinesponge

Introduction

The marine environment is a rich source of halogenatedcompounds produced by marine plants, animals, and bacte-ria [1]. Brominated secondary metabolites are produced bycyanobacterial symbionts [2–4] or by bacterial Vibrio spp.[5] in sponge tissues, although the profiles of these metab-olites differ as a result of the diversity of species involved.The family Dysideidae, in particular, yields a number ofhydroxylated polybrominated diphenyl ethers (OH-PBDEs),including dihydroxy (diOH), hydroxy-methoxy (OH-MeO),and dimethoxy (diMeO) analogs of tri- to hexa-BDEs [6,7].These phenolic PBDE analogs exhibit a variety of bioactiv-ities, such as antibacterial and antifungal properties [8],cytotoxicity, and enzyme inhibition [9]. Structure–activityrelationships indicate that these activities may depend on thenumbers and positions of hydroxyl groups [7]. In addition toPBDE homologs found in marine sponges, OH-PBDEs(e.g., 6-OH-BDE47) and the corresponding MeO-PBDEs(e.g., 2′-MeO-BDE68) have been reported to be present inmarine algae [10], mussels [11], fish blood [12], as well asthe marine food web [13–18]. Although there have as yetbeen no reports of diOH-PBDEs in higher trophic organ-isms, O-methylated PBDEs (e.g., 2′,6-diMeO-BDE68) havebeen found to accumulate in shark liver [19] and whaleblubber [20].

Y. Kato : S. Okada :K. AtobeKagawa School of Pharmaceutical Sciences,Tokushima Bunri University,1314-1 Sanuki,Kagawa 769-2193, Japan

T. EndoFaculty of Pharmaceutical Sciences,Health Sciences University of Hokkaido,1757 Ishikari-Tobetsu,Hokkaido 061-0293, Japan

K. Haraguchi (*)Daiichi College of Pharmaceutical Sciences,22-1 Tamagawa-cho, Minami-ku(Fukuoka 815-8511, Japane-mail: k-haraguti@daiichi-cps.ac.jp

Anal Bioanal Chem (2012) 404:197–206DOI 10.1007/s00216-012-6132-2

The extraction, chromatographic separation, and detectionof phenolic PBDE analogs have recently been reviewed [21].In general, OH-PBDEs are routinely quantified by gas chro-matography (GC), using either electron-capture detection,electron-ionization high-resolution mass spectrometry,electron-capture negative ionization mass spectrometry, ortandem mass spectrometry (MS/MS), after derivatization ofOH-PBDEs separated from neutral PBDEs and/or MeO-PBDEs [22]. GC/MS methods for OH-PBDEs provide thegreatest selectivity in the discrimination of PBDEs and MeO-PBDEs from other halogenated organics. However, diazome-thane derivatization of OH-MeO-PBDEs for GC/MS analysiscan be duplicated by that of diOH-PBDEs. Other derivatiza-tions (e.g., acetylation with acetic anhydride) of diOH-PBDEin the presence of OH-MeO-PBDEs result in several derivat-ized products and would make the identification difficult.

For the direct determination of OH-PBDEs, several ion-ization techniques using liquid-chromatography tandemmass spectrometry (LC/MS/MS) have been developed.Hua et al. [23] have studied high-performance liquid chro-matography (HPLC) in negative electrospray ionization(ESI) mode for the quantitative analysis of OH-tri-BDE.Mas et al. [24] have further developed negative ion spray-LC/MS/MS and determined eight OH-PBDEs in an envi-ronmental setting. Chang et al. [25] have developed thequantification of phenolic PBDEs, including bisphenol Aand bromophenols, in ESI mode. Lupton et al. [26] havereported LC/atmospheric pressure chemical ionization(APCI)-MS/MS for OH-PBDEs, ranging from tri- to hexa-brominated, in negative APCI mode. Lai et al. [27] havedeveloped a robust ultra-performance LC/MS/MS for rapiddetermination of nine OH-PBDEs to study the pharmacoki-netics of 6-OH-BDE47 in ESI mode. We have also devel-oped the simultaneous determination of OH- and MeO-PBDEs in APCI mode using both ([M−H]−→Br−) and ([M−Br+O]−→Br−) transitions [28]. However, these methodswere validated for monohydroxylated PBDEs but not forthe mixed dioxygenated (diOH- and OH-MeO-PBDEs) con-geners by using different MRM transitions. An alternativeLC/MS/MS method is therefore worth considering for the

analysis of dihydroxylated analogs in the environment andfor investigating the metabolic processes of diOH-PBDEs.

The aim of this study was to develop an LC/MS/MSmethod in negative APCI mode for direct determination ofdiOH, OH-MeO, and diMeO analogs for tetra- to hexa-BDEs, as well as mono substituted OH- and MeO-tetra-BDEs. The LC/MS/MS method was validated using2′-OH-BDE68, 2′,6-diOH-BDE68 and their O-methylatedanalogs. During our survey of brominated products inmarine sponges, we found a set of dihydroxylated PBDEanalogs that were dominant in Callyspongia sp. and Lamel-lodysidea sp. collected in Palau, Micronesia. In this paper,we describe the overall LC/MS/MS profiles and concentra-tions of these oxygenated PBDEs in two sponge species andcompare the total profiles of the derivatized compoundsusing GC/MS in EI mode.

Experimental

Chemicals

Standards of 2′-hydroxy-2,3′,4,5′-tetrabromodiphenyl ether(2′-OH-BDE68) and 2′-methoxy-2,3′,4,5′-tetrabromodi-phenyl ether (2′-MeO-BDE68), were purchased from Cam-bridge Isotope Laboratories Inc. (Andover, MA, USA). Thestandard 2′,6-dimethoxy-2,3′,4,5′-tetrabromodiphenyl ether(2′,6-diMeO-BDE68) was a gift from Dr. G. Marsh (Stock-holm University, Sweden). Standards of 2′,6-dihydroxy-2,3′,4,5′-tetrabromodiphenyl ether (2′,6-diOH-BDE68) and2′-hydroxy-6-methoxy-2,3′,4,5′-tetrabromodiphenyl ether(2′-OH-6-MeO-BDE68) were synthesized by demethylationof 2′,6-diMeO-BDE68 in the presence of boron tribromide(2 M) in dichloromethane. The demethylated products (OH-MeO-tetra-BDEs) were identified by GC-MS in EI mode onthe basis of its fragmentation pattern with M+ (m/z 526) and[M−CH3Br]

+ ion (m/z 434) [29]. The [M−CH3Br]+ ion was

characteristic for 2′-MeO-6-OH-BDE68 but not for the iso-meric 2′-OH-6-MeO-BDE68. Thus, both compounds can bedistinguished by the abundance of this ion. The purities of

Table 1 Characterization of reference compounds and MRM parameters used for quantitative determination

Compound LC tR(min)

LOQ(ng mL−1)

Recovery(%)

RSD(%)

MRM transition(m/z)

DP(V)

CE(V)

Mono-substituent

2′-OH-BDE68 7.2 1.9 80–89 15 500.3→78.6 [M−H]−→Br− −15 −44

2′-MeO-BDE68 13.6 9.5 87–92 16 450.7→78.6 [M−Br+O]−→Br− −40 −58

Di-substituent

2′,6-diOH-BDE68 5.1 9.1 76–92 20 516.3→78.9 [M−H]−→Br− −35 −64

2′-OH-6-MeO-BDE68 6.6 9.1 77–92 22 530.7→78.9 [M−H]−→Br− −30 −40

2′,6-diMeO-BDE68 11.5 15 88–93 15 480.8→78.8 [M−Br+O]−→Br− −40 −66

198 Y. Kato et al.

2′-OH-6-MeO-BDE68 and 2′,6-diOH-BDE68 isolated bypreparative reversed-phase LC were >84 and >91 %, respec-tively. These compounds were used as external standards forquantification of each homolog group by LC/MS/MS.

Sponge collection

One specimen of Callyspongia sp. (phylum Porifera, orderHaplosclerida, family Callyspongia) and two specimens of

Lamellodysidea sp. (formerly known as Dysidea; phylumCeractinomorpha, order Dictyoceratida, family Dysideidae)were collected while scuba diving at Nikko Bay Korrol, Palau,in August 2005. Freshly collected sponge materials were im-mediately frozen and stored at −20 °C prior to molecular andchemical analyses. The three specimens of sponges were phy-togenetically identified by the method of Meixner et al. [30].Voucher specimens have been deposited at the Department ofAnalytical Chemistry, Daiichi College of Pharmaceutical

Fig. 1 Chemical structures andLC/MS spectra (precursor ionscan of Br−) in negative APCImode of authentic 2′,6-diOH-BDE68 (a), 2′-OH-6-MeO-BDE68 (b), and 2′,6-diMeO-BDE68 (b). The concentrationof each standard was 2 μg mL−1

Selective determination of mono- and dihydroxylated analogs 199

Sciences, as numbers SP05-319 (Callyspongia sp.), SP05-321,and SP05-322 (Lamellodysidea sp.).

Sample preparation

Sponge tissues were cut into pieces and the homogenate wasextracted with methanol/ethyl acetate (1:1, v/v) at roomtemperature. The crude extracts were filtered, and theorganic matter was determined gravimetrically. The pu-rification was performed as described previously [31].Briefly, portions of the crude extracts were purified bygel-permeation chromatography (Bio-Beads, S-X3, Bio-Rad Laboratories, Hercules, CA, USA), with elutionwith dichloromethane/n-hexane (1:1, v/v). The eluate,which contained phenolic PBDEs, was concentrated todryness and dissolved in n-hexane. The hexane extractswere partitioned in 2 M KOH/ethanol (7:3, v/v). Theorganic phase (neutral fraction) was purified by silica-gel column chromatography (0.2 g, Wako gel S-1, WakoPure Chemical Industries Ltd., Osaka, Japan), elutedwith dichloromethane/n-hexane (12:88, v/v). The aqueouslayer was acidified with 2 M HCl and back-extracted threetimes with n-hexane/diethyl ether (8:2, v/v; phenolic fraction).

Both the neutral and phenolic fractions were concentrated anddissolved in acetonitrile for LC/MS/MS analysis.

APCI-LC/MS/MS analysis

Analyses were carried out using a liquid chromatograph(Prominence 20A; Shimadzu Co., Kyoto, Japan) coupledto a tandem mass spectrometer (3200QTRAP-MS/MS sys-tem; AB SCIEX, Tokyo, Japan). The column conditions andMS/MS parameters were optimized as described previously[28]. A reversed-phase Shim-pack FC-ODS column (150×4.6 mm, i.d., 3.0 μm particle size; Shimadzu Co., Kyoto,Japan) was used. The isocratic mobile phase compositionwas acetonitrile/water (9:1, v/v) at 0.5 mL min−1. Full-scandata acquisition was performed by scanning from m/z 50 to700 (Q1 scan range) in profile mode, using a scan time of 1 swith a step size of 0.1 amu and a pause between each scan of5 ms. To select the fragmentation patterns of m/z (Q1)→m/z(Q3) ions for the multiple reaction monitoring (MRM) tran-sitions, product ion scan mass spectra were recorded bycollision-activated dissociation of selected precursor ions.The optimized parameters for each analyte are summarizedin Table 1.

GC/MS analysis for comparison

The phenolic fraction from the sponge extract was furthermethylated with diazomethane for GC/MS analysis. Themethylated phenolic fraction was analyzed using a gaschromatograph (Agilent 6980N, Agilent, Santa Clara,CA, USA) equipped with a mass-selective detector(5973i) in electron-ionization and selected-ion-monitoringmode. The GC/MS conditions were the same as describedelsewhere [31].

Quality assurance and quality control

An eight-point calibration curve for 2′,6-diOH-BDE68,2′-OH-6-MeO-BDE68 and 2′,6-diMeO-BDE68 in the con-centration range 2–2,000 ng mL−1 was used to determinelinearity. The limits of quantification (LOQs) were deter-mined using spiked samples (n05) at 5 and 20 ng mL−1.LOQs using a signal-to-noise ratio (S/N) of 10 were mea-sured for 2′-OH-6-MeO-BDE68, 2′,6-diOH-BDE68, and2′,6-diMeO-BDE68. Recoveries of the three compoundswere assessed by spiking with 10 and 50 ng of each com-pound throughout the entire extraction procedure. Therepeatability (percentage relative standard deviation(RSD%)) for all analytes was evaluated by five consecutiveinjections of a 20–50 ng mL−1 standard solution and bymeasuring the same standard solution on different days.All reported concentrations were calculated by comparingthe MRM peak areas to the peak areas produced by external

Fig. 2 Total ion chromatograms of PBDE analogs in the phenolic andneutral fractions of sponge extracts from Lamellodysidea sp. (upper)and Callyspongia sp. (lower)

200 Y. Kato et al.

standards. Each group of compounds was quantified basedon the assumption that the MRM responses of ([M+4−H]−→Br−) for tetra-, penta-, and hexa-BDE analogs are

the same as those of 2′-OH-6-MeO-BDE68, and that theMRM responses of ([M+4−Br+O]−→Br−) are the same asthose of 2′,6-diMeO-BDE68.

Fig. 3 LC/MS spectra(precursor ion Q1 and production Q3 scan) of peaks labeledaccording to Fig. 2; peaks d andg are diOH-penta-/hexa-BDEsin the phenolic fraction ofLamellodysidea sp., peaks b’and e are OH-MeO-tetra-/penta-BDEs in the phenolic fractionof Callyspongia sp., and peaksc and f are diMeO-tetra-/penta-BDEs in the neutral fraction ofLamellodysidea sp.

Selective determination of mono- and dihydroxylated analogs 201

Results

The phenolic PBDEs (i.e., 2′,6-diOH-BDE68 and 2′-OH-6-MeO-BDE68) exhibited a base peak [M+4−H]− and a weakfragment peak [M+4−Br]−, whereas the dimethoxylated PBDE(i.e., 2′,6-diMeO-BDE68) exhibited two characteristic ions,[M+4−Br+O]− and [M+4−CH3]

− (Fig. 1). The MRM param-eters, LC retention times, LOQs, recoveries, and repeatabilitiesfor the five authentic standards are shown in Table 1. Eachcongener was well separated on the reversed-phase columnwithin 15 min, and none of theMRM peaks interfered with theothers. These five standards were used for validation of themethod and for quantification of each homolog as an externalstandard. The five-point calibration curve (20–1,000 ng mL−1)of each congener was linear (R2>0.999), and the proceduralrecoveries were 77–92 % (n05) for 2′-OH-6-MeO-BDE68 inthe phenolic fraction and 88–93 % (n04) for 2′,6-diMeO-BDE68 in the neutral fraction. The LOQ of 2′-OH-6-MeO-BDE68 was 9.1 ng mL−1 (S/N010), which was about 5-foldhigher than that for 2′-OH-BDE68 (1.9 ng mL−1) (Table 1).The repeatabilities (n05, RSD%) for each standard were15–22 % for inter-day variations.

Total ion chromatograms (TICs; precursor ion scan of Br−)for brominated compounds in the phenolic and neutral fractionsof the two sponge species are shown in Fig. 2. The Q1 scans ofmajor peaks in the phenolic fraction (Fig. 3) showed intensemolecular-related ions [M+4−H]− for diOH-penta-BDEs (m/z595, peak d) and diOH-hexa-BDEs (m/z 673, peak g), OH-MeO-tetra-BDEs (m/z 531, peak b′), and OH-MeO-penta-BDEs (m/z 609, peak e), whereas the Q1 scans of the peaksin the neutral fraction gave characteristic ions, [M+4−Br+O]−

and [M+4−CH3]−, for diMeO-tetra-BDEs (m/z 483, peak c),

and diMeO-penta-BDEs (m/z 561, peak f), respectively. Nomass spectra were acquired for OH-MeO-hexa-BDEs.

Under the optimized conditions (Table 2), the MRM chro-matograms gave peaks a, d, and g for diOH-tetra-/penta-/hexa-

BDEs (tR04.5–6.5 min), peaks b, b’, e, and e’ for OH-MeO-tetra-/penta-BDEs (tR06.5–8.0 min), and peaks c and f fordiMeO-tetra-/penta-BDEs (tR011–14 min) in the extractsfrom the two sponge species (Fig. 4). No peaks were observedin the MRMs for OH-MeO-/diMeO-hexa-BDEs. Each MRMpeak for ([M+4−H]−→Br−) was synchronized with those ofthe isocratic transition ([M−H]−→Br−) at an isotope ratio ofabout 1:3 to 1:6 for [M−H]−/[M+4−H]−. In the MRMs fordiOH-tetra-BDEs, peak a was identified as 2′,6-diOH-BDE68. The signal next to peak a (Fig. 4a) represents aninterfering peak from fragment ions (e.g., [M+4−Br]−) ofdiOH-penta-BDEs (peak d).

In the MRMs for OH-MeO-tetra-BDEs, peak b’(tR07.35 min) was tentatively identified as 2′-MeO-6-OH-BDE68 in extracts of Callyspongia sp. (Fig. 4a), whereasthe other peak b (tR06.60 min) was identified as the isomeric2′-OH-6-MeO-BDE68 in the extracts of Lamellodysidea sp.Similarly, in the MRMs for OH-MeO-penta-BDE, peak e wasobserved in the Lamellodysidea sp. and an additional peak e’was observed in the Callyspongia sp. (Fig. 4b); these wereidentified as the isomeric 2′-MeO-6-OH and/or 2′-OH-6-MeOanalogs. In the MRMs for diMeO-PBDEs, two major compo-nents (peaks c and f) were observed in both species, of whichpeak c had the same tR as 2′,6-diMeO-BDE68.

The identification of each homolog group in the phenolicfraction was confirmed by GC/MS in EI mode, followingderivatization with diazomethane (Fig. 5); there was onetetra-BDE homolog, two penta-BDE homologs, and onehexa-BDE homolog in the methylated phenolic fraction.

Using three external standards, we quantified threehomolog groups of diOH-, OH-MeO-, and diMeO-PBDEs in the extracts from three marine sponge species(two genus) using APCI-LC/MS/MS. Table 3 shows theconcentrations (in micrograms per gram wet weight,(ww)) of the mono-oxygenated (OH- and MeO-) tetra-BDE, and di-oxygenated (diOH-, OH-MeO-, and

Table 2 MRM parameters usedfor quantitative determination ofPBDE homologs

Compound Chemicalformula

Precursorion

Production

MRM transition(m/z)

DP(V)

CE(V)

Tetra-BDE homolog

diOH-tetra-BDE C12H6O3Br4 [M+4−H]− Br− 516.7→78.9 −30 −40

OH-MeO-tetra-BDE C13H8O3Br4 [M+4−H]− Br− 530.7→78.9 −30 −40

diMeO-tetra-BDE C14H10O3Br4 [M+4−Br+O]− Br− 482.8→78.9 −30 −40

Penta-BDE homolog

diOH-penta-BDE C12H5O3Br5 [M+4−H]− Br− 594.6→78.9 −30 −40

OH-MeO-penta-BDE C13H7O3Br5 [M+4−H]− Br− 608.6→78.9 −30 −40

diMeO-penta-BDE C14H9O3Br5 [M+4−Br+O]− Br− 560.7→78.9 −30 −40

Hexa-BDE homolog

diOH-hexa-BDE C12H4O3Br6 [M+4−H]− Br− 672.5→78.9 −30 −40

OH-MeO-hexa-BDE C13H6O3Br6 [M+4−H]− Br− 686.5→78.9 −30 −40

diMeO-hexa-BDE C14H8O3Br6 [M+4−Br+O]− Br− 638.7→78.9 −30 −40

202 Y. Kato et al.

diMeO-) tetra-, penta-, and hexa-BDEs in the two ge-nus. For mono-oxy-substituted PBDEs, OH-tetra-BDEs(62 μg g−1 ww in the Lamellodysidea sp. (SP05-322)and 3.0 μg g−1 ww in the Callyspongia sp.) weredetected at higher levels than the corresponding MeO-tetra-BDEs (10 and 1.8 μg g−1 ww, respectively). In theLamellodysidea sp.(SP05-322), the levels of diOH-tetra-BDE (31 μg g−1 ww), diOH-penta-BDE (264 μg g−1

ww), diOH-hexa-BDE (196 μg g−1 ww) were higherthan those of the corresponding OH-MeO analogs (lessthan 4.9 μg g−1 ww). In contrast, the phenolic fractionof the Callyspongia sp. contained higher amounts ofOH-MeO-tetra-BDEs (6.0 μg g−1 ww) than of diOH-tetra-BDEs (0.2 μg g−1 ww) or diMeO-tetra-BDEs(0.4 μg g−1 ww). The percentage contributions (%) of

diOH-, OH-MeO-, and diMeO-tetra-BDEs were 5:1:94in the Lamellodysidea sp. and 3:92:5 in the Callyspon-gia sp. (Table 3). The total percentage contributions ofmonohydroxy- and dihydroxy-substituted homologs were5.7 % (72 μg g−1 ww) and 93 % (1,174 μg g−1 ww),respectively, in the Lamellodysidea sp., and 40 %(4.8 μg g−1 ww) and 60 % (7.3 μg g−1 ww), respec-tively, in the Callyspongia sp.

Discussion

Until now, dihydroxylated or hydroxy-methoxy derivativesof PBDEs in the phenolic fractions from marine sampleshave been identified as diMeO-PBDEs by GC/MS,

Fig. 4 MRM chromatograms of diOH-, OH-MeO-, and diMeO-substituted tetra-, penta-, and hexa-BDEs in marine sponges: Lamello-dysidea sp. (a) and Callyspongia sp. (b). Peak labeling is according to

Fig. 2. Stationary phase, 150×4.6 mm, i.d. Shim-pack FC-ODS col-umn; mobile phase, acetonitrile:water (9:1, v/v) at a flow rate of0.5 mL min−1

Selective determination of mono- and dihydroxylated analogs 203

following derivatization, and thus quantified as a mixture[31]. The present study demonstrates that dihydroxy, hy-droxy-methoxy, and dimethoxy analogs of each PBDEcould be separately determined by MRM of LC/MS/MS innegative APCI mode, so that the concentration ratios of thethree analogs can be calculated in underivatized fractions.

The method was validated using 2′,6-diOH-BDE68,2′-OH-6-MeO-BDE68, and 2′,6-diMeO-BDE68 (Fig. 2).The LOQs (9.1 ng mL−1) of two phenolic standards byLC/MS/MS in the negative APCI mode are comparable tothe LOQ of 2′-OH-6-MeO-BDE68 by GC/MS in EI modeafter O-methylation. This allows the determination of diOH-PBDEs in human fluid matrices (e.g., blood). The presentLOQ for diOH-BDE68 is slightly higher than the LOQsreported by Mas et al. [24] which were measured for eightOH-PBDEs (0.7–4.6 g mL−1) in negative ion-spray ioniza-tion mode, and also those reported by Lai et al. [27] whoreported LOQs of 0.2–2.8 ng mL−1 for nine OH-PBDEs,

using the same mode. Higher ionization intensities mayoccur in higher brominated congeners such as penta-BDEor hexa-BDE [32]. The accuracies of spiked samples andrepeatabilities for diOH-PBDEs (n03, RSD<22 %) in thisstudy are similar to those for OH-PBDEs obtained usingother methods [24,25]. The present LC/MS/MS methodwith an isocratic mobile phase may give incomplete separa-tion of OH-PBDE isomers, compared with the ultra-performance liquid chromatography method used by Laiet al. [27] who demonstrated efficient separation of nineOH-PBDEs using a gradient system. However, these couldnot be directly compared with our target analytes. In ourstudy, the two isomeric 2′-OH-6-MeO- and 2′-MeO-6-OH-BDE68s were separated from each other under the HPLCconditions used. Furthermore, OH-MeO-tetra-BDE (peak b’)can be distinguished by MRM from OH-tetra-BDE (2′-OH-BDE68), which it was close to in terms of LC retention times(tRs of 2′-OH-BDE68 and 2′-MeO-6-OH-BDE68 were 7.2and 7.35min, respectively).Mono- and dihydroxylated PBDEanalogs from the marine food web can therefore be screenedusing the APCI-LC/MS/MS, without derivatization. We frac-tionated the phenolic and neutral compounds by partitioningwith KOH and n-hexane. However, the fractionation follow-ing the GPC procedure may not be necessary, because the bothproducts can be separated by HPLC (Fig. 2).

The developed method was applied to two marinesponges with different profiles of dihydroxylated analogs,from tetra- to hexabrominated. The Q1 scan (TIC) indicatedthat the diOH analogs of penta- and hexa-BDEs are pre-dominant in the Lamellodysidea sp, whereas OH-MeO-tetra-BDEs are selectively present in the Callyspongia sp.The interfering peak which appeared in the MRMs of OH-MeO-BDEs may be caused by the fragment ion [M−Br]− ofdiOH-PBDEs. In fact, the Lamellodysidea sp. producedlarge amounts of diOH analogs (at levels two orders ofmagnitude higher than that for the OH-MeO analogs). Asa result, the MRM ([M+4−H]−→Br−) for OH-MeO-tetra-BDEs is interfered with by that for the diOH-penta-BDEs.

In the MRMs of OH-MeO-tetra-/penta-BDEs, the pres-ence of two isomeric congeners is confirmed by GC/MSfollowing derivatization of the phenolic fraction (Fig. 5).The methylated phenolic fraction gave one peak for thetetra-BDE homolog, two peaks for the penta-BDE homo-logs, and one peak for the hexa-BDE homolog, originatingfrom the diOH and/or OH-MeO analogs. LC/MS/MSshowed that a higher amount of diOH-tetra-BDE than ofOH-MeO was observed in the Lamellodysidea sp., wherediOH/OH-MeO was 10:1 for the tetra-BDE homologs, and100:1for penta-BDE. In contrast, a higher amount of OH-MeO-tetra-BDE was observed in the Callyspongia sp. (theratio, 1:30 for the tetra-BDE homologs). These findingssuggest that the formation ratios of these phenolicPBDEs are species-dependent and support the

Fig. 5 GC/MS (EI) TIC of brominated products in the methylatedphenolic fraction (upper) and neutral fraction (lower) from Lamellody-sidea sp. extracts. The phenolic fraction (Fig. 2) was methylated withdiazomethane. The peaks in this fraction are estimated as the sum ofdiOH and OH-MeO analogs (one peak for tetra-, two peaks for penta-,and one peak for hexabrominated homologs). IS internal standard, 4′-MeO-BDE121

204 Y. Kato et al.

hypotheses that the bacterial O-methylation of diOH-PBDEs in the sponges may vary with the degree ofbromination [33].

Mixed analogs of mono- and di-OH-tetra-BDEs havealso been isolated from Dysidea spp. [5–8] and marine algae[10]. The present study also supported the suggestion that inmost species the major component has the diphenyl etherstructure oxygenated at the 2,6′-position. One of the meth-ylated products, 2′,6-diMeO-BDE68, has been found toaccumulate in fish [31], shark liver [19], and whale blubber[16,19,20]. The potential source may be the sponge bromi-nated products in the phenolic fraction, but these phenolicproducts have not been detected, or were only present intrace amounts, in higher trophic organisms [28], probablybecause of the lower persistence and higher water solubilityof phenolic PBDEs, and they accumulate as diMeO-tetra-BDEs following bacterial methylation of the hydroxylgroup.

The simultaneous determination of OH-, OH-MeO-, anddiOH-PBDEs in phenolic fractions are of great importancein monitoring studies of animals and humans, because theseanalogs exhibit a variety of bioactivities, depending on thestructure (the presence of one or two hydroxyl groups inPBDE molecules) [7–9]. For example, diOH-PBDEs and

OH-MeO-PBDEs are active against the Gram-positive bac-terium Bacillus subtilis, whereas in their methylated analogs(diMeO-PBDEs), such activity is reduced [7]. In anotherreport, OH-MeO-PBDEs were shown to have a greaterbiological effect than those of the corresponding diOH-PBDEs [34]. The proposed LC/MS/MS method should bean efficient tool for elucidating the O-methylation of diOH-PBDEs, or microsomal demethylation of diMeO-PBDEs.

Conclusions

This study demonstrated that negative APCI-LC/MS/MS isan efficient tool for the selective determination of diOH-PBDEs and the corresponding O-methylated analogs inmarine biota. The method provides better selectivity fordiOH-PBDEs and OH-MeO-PBDEs in the MRM transition([M−H]−→Br−), and HPLC could be used to resolve thehomologs. The APCI-LC/MS/MS method was applied totwo sponge extracts with different profiles of a set of diOH,OH-MeO, and diMeO analogs, within homologs and be-tween species. The proposed approach will broaden moni-toring studies dealing with the sources, metabolism, andfates of diOH-PBDEs in wildlife and humans.

Table 3 Determination ofdihydroxy-, hydroxy-methoxy-,and dimethoxy-PBDEs in ma-rine sponges using APCI-LC/MS/MS

ND not detected (less than0.05 μg g−1 wet weight)

Homologs Concentrations (μg g−1 wet weight)

Lamellodysidea sp. Callyspongia sp.

SP05-321 SP05-322 SP05-319

Mono-substituted

OH-tetra-BDE 55.4 61.8 3.01

MeO-tetra-BDE 2.19 10.4 1.82

Total 57.6 (6.1 %) 72 (5.7 %) 4.8 (40 %)

Di-substituted

Tetra-BDE homolog

diOH-tetra-BDE 37.9 30.7 0.20

OH-MeO-tetra-BDE 3.79 4.93 5.98

diMeO-tetra-BDE 381 639 0.36

Total 423 (44.7 %) 675 (53.7 %) 6.5 (54.2 %)

Penta-BDE homolog

diOH-penta-BDE 276 264 0.58

OH-MeO-penta-BDE 3.33 18.5 0.06

diMeO-penta-BDE 7.98 20 ND

Total 287 (30.3 %) 303 (24.1 %) 0.64 (5.3 %)

Hexa-BDE homolog

diOH-hexa-BDE 170 196 0.06

OH-MeO-hexa-BDE ND ND ND

diMeO-hexa-BDE ND ND ND

Total 170 (18.0 %) 196 (15.6 %) 0.06 (0.5 %)

Total 947 (100 %) 1,256 (100 %) 12 (100 %)

Selective determination of mono- and dihydroxylated analogs 205

Acknowledgment This research was financed by Grants-in Aid fromthe Japan Society for the Promotion of Science (B20404006 andC23510083).

References

1. Gribble WG (2003) Chemosphere 52:289–2972. Unson MD, Faulkner DJ (1993) Experientia 49:349–3533. Unson MD, Holland ND, Faulkner DJ (1994) Mar Biol 119:1–124. Malmvarn A, Zebuhr Y, Kautsky L, Bergman K, Asplund L (2008)

Chemosphere 72:910–9165. Elyakov GB, Kuznetsova T, Mikhailov VV, Maltsev II, Voinov

VG, Fedoreyev SA (1991) Experientia 47:632–6336. Handayani D, Edrada RA, Proksch P, Wray V, Witte L, Van Soest

RW, Kunzmann A, Soedarsono (1997) J Nat Prod 60:1313–13167. Hanif N, Tanaka J, Setiawan A, Trianto A, de Voogd NJ, Murni A,

Tanaka C, Higa T (2007) J Nat Prod 70:432–4358. Sionov E, Roth D, Sandovsky-Losica H, Kashman Y, Rudi A,

Chill L, Berdicevsky I, Segal E (2005) J Infect 50:453–4609. Liu H, Namikoshi M, Meguro S, Nagai H, Kobayashi H, Yao X

(2004) J Nat Prod 67:472–47410. Haraguchi K, Kotaki Y, Relox JR, Romero ML, Terada R (2010) J

Agric Food Chem 58:12385–1239111. Malmvärn A, Marsh G, Kautsky L, Athanasiadou M, Bergman Å,

Asplund L (2005) Environ Sci Technol 39:2990–299712. Marsh G, Athanasiadou M, Bergman Å, Asplund L (2004) Envi-

ron Sci Technol 38:10–1813. Kelly BC, Ikonomou MG, Blair JD, Gobas FA (2008) Environ Sci

Technol 42:7069–707714. Sinkkonen S, Rantalainen AL, Paasivirta J, Lahtipera M (2004)

Chemosphere 56:767–77515. Covaci A, Losada S, Roosens L, Vetter W, Santos FJ, Neels H,

Storelli A, Storelli MM (2008) Environ Sci Technol 42:8654–866016. Vetter W, Stoll E, Garson MJ, Fahey SJ, Gaus C, Muller JF (2002)

Environ Toxicol Chem 21:2014–2019

17. Pettersson A, van Bavel B, Engwall M, Jimenez B (2004) ArchEnviron Contam Toxicol 47:542–550

18. Stapleton HM, Dodder NG, Kucklick JR, Reddy CM, SchantzMM, Becker PR, Gulland F, Porter BJ, Wise SA (2006) Mar PollutBull 52:522–531

19. Haraguchi K, Hisamichi Y, Kotaki Y, Kato Y, Endo T (2009)Environ Sci Technol 43:2288–2294

20. Marsh G, Athanasiadou M, Athanassiadis I, Bergman Å, Endo T,Haraguchi K (2005) Environ Sci Technol 39:8684–8690

21. Covaci A, Voorspoels S, Ramos L, Neels H, Blust R (2007) JChromatogr A 1153:145–171

22. Lacorte S, Ikonomou MG, Fischer M (2010) J Chromatogr A1217:337–347

23. Hua W, Bennett ER, Letcher RJ (2005) Environ Int 31:621–63024. Mas S, Jauregui O, Rubio F, de Juan A, Tauler R, Lacorte S (2007)

J Mass Spectrom 42:890–89925. Chang H, Wan Y, Naile J, Zhang X, Wiseman S, Hecker M, Lam

MHW, Giesy JP, Jones PD (2010) J Chromatogr A 1217:506–513

26. Lupton SJ, McGarrigle BP, Olson JR, Wood TD, Aga DS (2010)Rapid Commun Mass Spectrom 24:2227–2235

27. Lai Y, Chen X, Lam MHW, Cai Z (2011) J Chromatogr B879:1086–1090

28. Kato Y, Okada S, Atobe K, Endo T, Matsubara F, Oguma T,Haraguchi K (2009) Anal Chem 81:5942–5948

29. Athanasiadou M, Marsh G, Athanassiadis I, Asplund L, BergmanÅ (2006) J Mass Spectrom 41:790–801

30. Meixner MJ, Luter C, Eckert C, Itskovich V, Janussen D, vonRintelen T, Bohne AV, Meixner JM, Hess WR (2007) Mol Phylo-genet Evol 45:875–886

31. Haraguchi K, Kato Y, Ohta C, Koga N, Endo T (2011) J AgricFood Chem 59:13102–13109

32. Lagalante AF, Oswald TD (2008) Anal Bioanal Chem 391:2249–2256

33. Allard AS, Remberger M, Neilson AH (1987) Appl EnvironMicrobiol 53:839–845

34. Oda T, Liu H, Namikoshi M (2005) Mar Drugs 3:119–125

206 Y. Kato et al.