Arsenic Demethylation by a C·As Lyase in Cyanobacterium Nostoc sp.PCC 7120Yu Yan,†,‡,∥ Jun Ye,†,∥ Xi-Mei Xue,† and Yong-Guan Zhu*,†,§

†Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021,People’s Republic of China‡University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China§State Key Lab of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences,Beijing 100085, People’s Republic of China

*S Supporting Information

ABSTRACT: Arsenic, a ubiquitous toxic substance, exists mainly asinorganic forms in the environment. It is perceived that organo-arsenicals can be demethylated and degraded into inorganic arsenicby microorganisms. Few studies have focused on the mechanism ofarsenic demethylation in bacteria. Here, we investigated arsenicdemethylation in a typical freshwater cyanobacterium Nostoc sp.PCC 7120. This bacterium was able to demethylate monomethy-larsenite [MAs(III)] rapidly to arsenite [As(III)] and also had theability to demethylate monomethylarsenate [MAs(V)] to As(III).The NsarsI encoding a C·As lyase responsible for MAs(III)demethylation was cloned from Nostoc sp. PCC 7120 andheterologously expressed in an As-hypersensitive strain Escherichiacoli AW3110 (ΔarsRBC). Expression of NsarsI was shown to confer MAs(III) resistance through arsenic demethylation. Thepurified NsArsI was further identified and functionally characterized in vitro. NsArsI existed mainly as the trimeric state, and thekinetic data were well-fit to the Hill equation with K0.5 = 7.55 ± 0.33 μM for MAs(III), Vmax = 0.79 ± 0.02 μM min−1, and h =2.7. Both of the NsArsI truncated derivatives lacking the C-terminal 10 residues (ArsI10) or 23 residues (ArsI23) had a reducedability of MAs(III) demethylation. These results provide new insights for understanding the important role of cyanobacteria inarsenic biogeochemical cycling in the environment.

■ INTRODUCTION

Arsenic (As) is a ubiquitous toxic and carcinogenic substance inthe environment.1 Arsenic exposure through food and water arecausing human health problems, including lung and skincancers, in many parts of the world.2,3 Arsenic was ranked thefirst on the National Priorities List (NPL) of HazardousSubstances (http://www.atsdr.cdc.gov/SPL/index.html) by theAgency for Toxic Substances and Disease Registry (ATSDR).The biotransformations of inorganic arsenic into organo-

arsenicals are important parts of global arsenic biogeochemicalcycle. Inorganic As, including arsenate [As(V)] and arsenite[As(III)], are dominant chemical species in the environment.4,5

However, microorganisms can convert inorganic As intoorganic species.6 The microbial methylation detoxifies As(III)by sequentially producing less toxic MAs(V), dimethylarsenate[DMAs(V)], trimethylarsine oxide [TMAsO(V)], and nontoxicvolatile trimethylarsine [TMAs(III)].7,8 However, MAs(III)and dimethylarsenite [DMAs(III)] produced in the methyl-ation process are more toxic than As(III) for bothmammalian9,10 and microbial cells.11 As(III) S-adenosylme-thionine methyltransferases (ArsMs) from bacteria, archaea,and eukaryotes (except higher plants) catalyzing the bio-methylation have been characterized.12−16 The widespread and

early-evolved ArsMs continually methylate the environmentalarsenic. In addition, substantial amounts of organoarsenicals,such as monosodium methylarsonic acid (MSMA), were largelyintroduced into the environment by their use as herbicides andpesticides during the past half century.17 Nonetheless, themajority of environmental arsenic exists as inorganic forms.This is because the demethylation of organoarsenicals may beextensive in both terrestrial and aquatic environments, andmicrobes have been found to be involved in As demethyla-tion.18,19 Based on As species analysis, studies have been doneon the transformation of methylarsenicals in the soils oforganoarsenic-contaminated sites, such as cotton fields20 andgolf course greens,21 where high levels of MSMA were used asherbicides. Moreover, the biodegradation of organoarsenicalshas also been evaluated in anaerobic sludge22 and lakes.23,24

Furthermore, methylarsenic-decomposing microorganisms havebeen classified and isolated from soil and water,25 and it hasbeen shown that only a few types of microbes, such as

Received: July 11, 2015Revised: September 23, 2015Accepted: November 6, 2015Published: November 6, 2015

Article

pubs.acs.org/est

© 2015 American Chemical Society 14350 DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

Montrachet wine yeast,26 Candida humicola,27 and Mycobacte-rium neoaurum28 are capable of As demethylation. A novel two-step pathway of MSMA degradation by reducing MAs(V) toMAs(III) and demethylating MAs(III) to As(III) in sequencewas demonstrated recently.29 The two steps are performed bydifferent bacterial species isolated from Florida golf course soil.The gene, namely arsI (As-inducible gene),30 responsible forthe second-step MAs(III) demethylation was identified fromthe environmental MAs(III)-demethylating isolate Bacillus sp.MD1, and the ArsI protein confers MAs(III) resistance and hasFe2+-dependent C·As lyase activity.11 Its ortholog fromThermomonospora curvata has been crystallized to determinethe three-dimensional structure.31

Ancient cyanobacteria as photosynthetic prokaryotes areglobally widespread, found from seawater and freshwatersystems to soils and even deserts.32 Cyanobacteria often growin heavy metal(loid)-contaminated habitats due to their abilityto resist metal toxicity.33 Previous studies have indicated thatcyanobacteria have multiple As biotransformation pathwaysand, in particular, make great contributions to As methylation.14

However, it is still unclear whether cyanobacteria are associatedwith the reverse reaction, As demethylation, not to mention themolecular mechanism. Nostoc sp. PCC 7120 (Nostoc) is atypical filamentous nitrogen-fixing cyanobacterium. Nostoc hasbeen extensively studied as a model organism for the geneticanalysis of photosynthetic and nitrogen fixation processes. Thesequencing of the entire genome of Nostoc has been completedin 2002.34 For As metabolism, Nostoc can methylate As(III)into methylated arsenicals35 and produce arsenosugars.36

Because of its available genomic and As metabolic information,we focused on the molecular mechanism underlining Asdemethylation by Nostoc in this study. Herein, we investigatedmethylarsenical degradation pathway of Nostoc and focused onthe molecular mechanism of As demethylation. By rapidlydemethylating As, cyanobacteria including Nostoc play asignificant role in As biogeochemical cycling. Our results alsoshed light on the molecular mechanism of As metabolism incyanobacteria.

■ MATERIALS AND METHODSOrganisms and Cultivations. Nostoc was kindly provided

by Professor Wen-Li Chen, Huazhong Agricultural University,and cultivated in BG11 medium without nitrate.37 Experimentswere carried out in a light-equipped shaker incubator under a12 h dark−light cycle with a light intensity of approximately 50μmol photons (m−2 s−1) at day and night temperatures of 27and 25 °C, respectively.Unless otherwise noted, E. coli strains were grown at 37 °C

aerobically in Luria−Bertani (LB) medium with shaking at 180rpm. For plasmid construction and replication, E. coli strainDH5α (Promega) was used. The As(III) hypersensitive E. colistrain AW3110 (DE3)38 [ΔarsRBC; ArsR-repressor, ArsB-As(III) efflux pump, and ArsC-As(V) reductase] was used forresistance assay and functional verification of NsarsI. E. colistrain Rosetta (DE3) (Novagen) was used for proteinexpression.Chemical Reagents. The MAs(III) was prepared by

reduction of MAs(V) as previously described.11 All otherreagents were obtained from commercial sources and ofanalytical grade or better.Arsenic Demethylation by Nostoc. When the cultures

were in the exponential phase, Nostoc was treated with 1 μMMAs(V) or MAs(III) in triplicate. The samples were harvested

by centrifugation (6500g, 5 min) at each different timeintervals. The supernatants were filtrated through 0.22 μmdisposable filters and kept at −80 °C until analysis. The cellswere washed with precooled MES buffer [1 mM K2HPO4, 5mM MES, and 0.5 mM Ca(NO3)2 at pH 6.2]39 to removeapoplastic As and then lyophilized. The cellular As wasextracted by deionized water (18.2 ΩM cm−1) as describedbelow. After the appropriate 0.01 g of freeze-dried cells hadbeen transferred to lysing tubes with 0.5 g glass beads, 300 μLof deionized water was added into the tubes. Sequentially thecell lysis was carried out with Fastprep-24 (MP Biomedicals)for 60 s at a speed of 6.5 m s−1. The cell lysis liquid wascentrifuged (13400g, 1 min) at 4 °C, and then the supernatantswere carefully pipetted into a new tube. The above steps wererepeated three times, and the supernatants were combined andfiltrated through 0.22 μm filters to determine the As species.

Cloning of arsI from Nostoc. To identify the putative arsIin Nostoc, we used the protein sequence of confirmed ArsI inBacillus sp. MD1 (accession no. KF899847.1) as a query toperform a similar search (BALSTP) against the proteins ofNostoc. The result showed that alr1104 (accession no.BAB73061) annotated as glyoxalase was an ortholog of ArsI.Thus, we amplified alr1104 (named as NsarsI) from thegenomic DNA, which was isolated from Nostoc by using a plantgenomic DNA extraction kit (Bioteke) according to themanufacturer’s instructions. The NsarsI consisting of the startcodon and excluding the stop codon was amplified by PCRwith primers 5′-CATATGTCCGTTATGAAAACACACG-3′(NcoI restriction site underlined) and 5′-CTCGAGAGCA-CAACATGACTTC-3′ (XhoI restriction site underlined). ThePCR products were purified and cloned into plasmid pMD19-Tsimple vector (Takara) to result in the plasmid pMD19T-NsarsI. After digestion with NdeI and XhoI, NsarsI was clonedinto the NdeI−XhoI-digested expression vector pET22b(Novagen) to generate plasmid pET22b−NsarsI. ThepET22b−NsarsI was subsequently transformed into the E. coliRosetta (DE3) or AW3110 (ΔarsRBC).

Overexpression and Purification of NsArsI in E. coli.Rosetta cells bearing pET22b−NsarsI were incubated at 37 °Cin LB medium containing ampicillin (100 μg mL−1) andchloramphenicol (30 μg mL−1). Overnight cultures of E. coliwere diluted 100-fold into 2 L of fresh LB medium withampicillin and chloramphenicol. Exponential phase cells wereinduced by 1 mM isopropyl β-D-1-thiogalactopyranoside(IPTG) to express His-tagged NsArsI. After being continuallycultured at 16 °C overnight, the cells were harvested bycentrifugation (4 °C, 6000g, 15 min) and resuspended withHEPES buffer [20 mM HEPES, 500 mM NaCl, 3 mM tris(2-carboxyethyl)phosphine (TCEP), pH 7.2]. The cells weredisrupted by a single pass through a French press at 10 MPaand centrifuged at 4 °C and 13000g for 1 h. The supernatantwas gravity-loaded onto a Ni-NTA agarose column (Qiagen)pre-equilibrated with HEPES buffer and washed with 20 mMimidazole in HEPES buffer. The protein was then eluted with15 mL of HEPES buffer containing 500 mM imidazole, andfurther purified by size-exclusion chromatography (HiLoad 16/600 Superdex 75 pg, GE Healthcare Life Sciences) with HEPESbuffer. The elution volume of NsArsI was compared with theelution profile of molecular-mass standards (BioRad) toinvestigate its oligomer status. The purified NsArsI wasidentified by sodium dodecyl sulfate-polyacrylamide gelelectrophoresis (SDS-PAGE) and concentrated by ultrafiltra-tion (10 kDa cutoff Amicon Ultrafilter, Millipore).

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14351

MAs(III) Resistance Assays in E. coli. MAs(III) resistanceassays were performed as previously described.11 Single colonyof AW3110 (ΔarsRBC) bearing either pET22b or pET22b-NsarsI was incubated into 5 mL of LB medium containingantibiotics with shaking (200 rpm) at 37 °C overnight. Thecells were diluted 100-fold into 30 mL of fresh LB medium withantibiotics and induced by adding 1 mM IPTG when theoptical density at 600 nm (OD600 nm) reached 0.6−0.8. Afterincubation at 16 °C for an additional 18 h, the induced cellswere washed with an equal volume of ST medium (10-foldconcentrated ST 10−1 medium23) and diluted 50-fold into 30mL of fresh ST medium supplemented with antibiotics and0.2% D-glucose. The cells were cultured with the indicatedconcentrations of MAs(III) at 30 °C with shaking at 200 rpm.The OD600 nm was monitored by using an ultraviolet−visiblespectrophotometer (UV-6300 double beam spectrophotome-ter, Mapada).MAs(III) Demethylation Assays in E. coli. AW3110

(ΔarsRBC) bearing either pET22b or pET22b−NsarsI wasused to measure MAs(III) demethylation. The IPTG-inducedcells were incubated for an additional 6 h and then transferredto ST 10−1 medium containing the appropriate antibiotics and0.5 μMMAs(III). After incubation at 25 °C for another 1 h, thesamples were centrifuged at 13400g for 2 min; the supernatantswere filtrated through 0.22 μm filters, the pellets were washedand lysed with deionized water using the method forcyanobacteria as described above. The MAs(III)-demethylatingactivity was analyzed by determining As species in both themedium and cells.MAs(III) Demethylation Assays Using Purified NsArsI.

In vitro reaction of MAs(III) demethylation with 10 μMpurified NsArsI was performed in triplicate in a 100 mM MOPSbuffer (pH 7.0) at 30 °C for 1 h. The reaction mixture alsocontained 3 mM TCEP, 1 mM cysteine, 0.1 mM Fe2+, and 10μM MAs(III). Each assay mixture was immediately terminatedby adding 40 mM EDTA at the indicated time point anddivided into two parts. One portion was directly diluted 10-foldand filtered through a 0.22 μm filter to monitor the As speciesduring the reaction process, and the other portion was used todetermine the nature of NsArsI-bound arsenicals as describedbelow. The samples were passed through BioGel P6 columns(BioRad) pre-equilibrated with MOPS buffer, followed bydilution with 6 M guanidine HCl at once to denature proteinand release the bound As, the species of which weredetermined. For kinetic assays, the reaction was performedwith the indicated concentrations of MAs(III). The amounts ofAs(III) generated at 1 min by 10 μM NsArsI were used asapproximate initial rates. Nonlinear regression analysis wasperformed with SigmaPlot 12.5.MAs(III) Demethylation of NsArsI Truncated Deriva-

tives in Vivo and in Vitro. The results from multiplesequence alignment of ArsI homologues showed that there islittle conserved sequence in C-termini. In addition, arsI124,which is an arsI derivative from Bacillus sp. MD1 encoding onlythe N-terminal 124 residues still has the C·As lyase activity.11

Herein, two truncated derivatives of NsArsI including ArsI10and Ars23 were constructed to verified whether the C-terminalresidues of NsArsI influence the MAs(III) demethylationactivity. The forward primer used for truncated derivativeswas the one for full-length NsArsI. The reverse primers were 5′-CTCGAGTTCCCCAGAAGAACTTAC-3′ (XhoI restrictionsite underlined) and 5′-CTCGAGCGCTGTATCTGCTAC-3′(XhoI restriction site underlined) for ArsI10 and ArsI23,

respectively. The other conditions for gene cloning, proteinpurification, and functional verification in vivo and in vitro ofthese derivatives were the same to the full-length NsArsI unlessotherwise indicated.

Arsenic Speciation Analysis. Arsenic speciation wasassayed by high-performance liquid chromatography−induc-tively coupled plasma mass spectrometry (HPLC−ICP-MS,7500a, Agilent Technologies) using the previous instrumentparameters.12 A Jupiter 5 μ C18 300A reverse-phase column(250 mm × 4.6 mm, Phenomenex) was used for arseniccompounds separation. The mobile phase (pH 5.95) consistedof 3 mM malonic acid, 5 mM tetrabutylammonium hydroxide,and 5% methanol at a flow rate of 1.0 mL min−1. The ICP-MSwas tuned for the monitoring of m/z 75 (arsenic). At the sametime, m/z 77 and m/z 82 (selenium) were used for monitoringthe interference by ArCl, and m/z 72 (germanium) and m/z103 (rhodium) were used as internal standards to compensatefor any instrument drift. Arsenic species in the samples wereidentified by retention times, compared with those of thestandards composed of As(III), MAs(III), DMA(V), MAs(V),and As(V). Arsenic amounts were quantified by externalcalibration curves with peak areas.

■ RESULTSNostoc Demethylation of Both MAs(III) and MAs(V) to

As(III). Biodemethylation of MAs(III) by Nostoc was measuredduring a 24 h period. During the first 12 h, the concentration ofAs(III) increased rapidly in the medium, concomitant with thedecrease of MAs(III) (Figure 1A). As(III) concentration in themedium did not increase beyond 12 h. During the wholeincubation period, MAs(V) was shown to be the major Asspecies in the cells (Figure S1) and may be the product ofMAs(III) oxidation by enzymatic or nonenzymatic reactions. Itis also possible that the detected MAs(V) was formed byMAs(III) oxidation during sample preparation. However,As(III), accounting for only 6 to 25% of the total As in cells,was relatively less, and its peak appeared in the third hour.As(V) was also detected in the cells and showed a similar trendto As(III). There was no inorganic arsenic detected in thecontrol without Nostoc inoculation. MAs(V) demethylation byNostoc was measured during 3 weeks of treatment. The rate ofdemethylation for MAs(V) was much slower than that forMAs(III) (Figure 1B). The increase in the quantity of As(III)coincided with the decrease of the MAs(V) in the mediumduring the initial 15 days, and there was no further increase inAs(III) from 15 to 21 days. Quite a small amount of MAs(III)was also detected for 1−9 days, accounting for 0.4−1.1% of thetotal arsenic in the medium.

Cloning of NsarsI. A putative 450 bp length NsarsI wasyielded by PCR using the genomic DNA of Nostoc. NsarsIencodes NsArsI of 150 residues, with a predicted molecularmass of 17.35 kDa. The analysis of sequence alignment showedthat NsArsI had a high degree of sequence homology (45%identity and 83% similarity) to ArsI from Bacillus sp. MD1.NsArsI contained the histidine(7)−histidine(65)−glutamicacid(117) motif as a putative divalent metal (Fe2+) bindingsite and a cysteine pair at residues 98−99 as a putativeMAs(III) binding site (Figure S2). Besides, the first residue(His7) of the putative Fe2+ binding site was not conserved andwas replaced by glutamine residue in the putative ArsI orthologfrom Phycicoccus jejuensis.

E. coli AW3110 Bearing NsarsI Conferment of MAs(III)Resistance. When NsarsI was expressed in E. coli AW3110,

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14352

which had neither chromosomal arsRBC nor orthologous arsI,it conferred resistance to MAs(III) (Figure 2A). When grownon MAs(III)-free ST medium, AW3110 cells bearing eitherpET22b or pET22b−NsarsI showed almost no difference ingrowth rate. At 1 and 3 μM of MAs(III), AW3110 expressingNsarsI grew significantly faster than that bearing pET22b at the7 to 24 h time points (t test, P < 0.05). AW3110 bearingpET22b−NsarsI still grew slightly faster than that bearingempty plasmid at 9 μM of MAs(III), although both of AW3110were strongly inhibited at higher concentration MAs(III).MAs(III) Demethylation to As(III) by E. coli AW3110

Expressing NsarsI. To elucidate whether NsarsI has MAs(III)demethylation activity in vivo, we incubated AW3110expressing NsarsI in ST medium containing 1 μM MAs(III),and the products of demethylation reactions were quantified inthe culture medium (Figure 2B). After AW3110 bearing NsarsIhad been cultured for 1 h, MAs(III) became undetected, andAs(III) was the dominant arsenic species in the medium.Meanwhile, only inorganic arsenic, including As(III) andAs(V), was detected in the cells lysates of AW3110 expressingNsarsI (Figure 2C). There were trace amounts of As(III)detected in both the medium and the cells of AW3110 bearingpET22b. It was probably coming from the slightly contami-

nated MAs(III) used in the experiments because the initialsubstrate also contained a trace amount of As(III).

MAs(III) Demethylation to As(III) by Purified NsArsI inVitro. To make the mechanism of the enzyme morecomprehensible, we analyzed changes of arsenic species invitro assayed with purified NsArsI over 1 h. Purified His-taggedNsArsI as soluble protein had a molecular mass of

Figure 1. MAs(III) and MAs(V) demethylation by Nostoc. (A) Asspecies in BG11 medium with Nostoc exposed to MAs(III) for theindicated times (0, 1, 3, 6, 9, 12, and 24 h). (B) As species in BG11medium with Nostoc exposed to MAs(V) for the indicated times (0, 1,3, 6, 9, 12, 15, 18, and 21 days). The error bars indicate the standarderror of triplicate assays.

Figure 2. MAs(III) resistance and demethylation by AW3110(ΔarsRBC) expressing pET22b-NsarsI. (A) The growth curves ofAW3110 bearing pET22b-NsarsI or pET22b. Absorbance wasmonitored at 600 nm. (B) As species in ST medium cultured withAW3110 exposed to 0.5 μM MAs(III) for 1 h at 25 °C. (C) HPLC−ICP-MS analysis of the As species in the cell lysate of AW3110 asdescribed in the Materials and Methods section. AW3110 bearingpET22b was the control.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14353

approximately 17 kDa, which was basically coincided with thepredicted molecular mass (Figure S3). Once purified NsArsIwas added into the reaction, the quantity of MAs(III) declinedsharply, although there was only a small increase in As(III).With the reaction time prolonging in 1 h, the concentration ofAs(III) continually increased, and total soluble arsenic graduallyrecovered to the level without adding NsArsI (Figure 3A).

Hence, we predicted that MAs(III) was quickly and stronglybound to the enzyme,40 so arsenic species bound to NsArsI wasanalyzed. The results showed that MAs(V) was present in theentire reaction and might be the product of MAs(III) oxidationduring sample preparation, while As(III) was mainly observedin the last 40 min (Figure 3B).NsArsI Mainly Existed as a Trimer. Size-exclusion

chromatography of NsArsI resulted in a main peak at 69.4mL and a secondary peak at 60.9 mL (Figure S4). The elutionprofile of NsArsI corresponded to a mass of 50 kDa and 106kDa by comparison with molecular-mass standards (Figure4A). Although the predicted NsArsI monomer mass was about17 kDa. Thus, NsArsI probably mainly existed as a trimer with alittle mixture of hexamer in solution. In agreement with thechromatography, the kinetic of NsArsI for MAs(III)demethylation was in accord with the Hill equation with an h

value of 2.7 rather than Michaelis−Menten eq (Figure 4B). Therest kinetic parameters of K0.5 and Vmax for MAs(III) were 7.55± 0.33 μM and 0.79 ± 0.02 μM min−1, respectively.

Truncated NsArsI Derivatives Were Active. There werenonconserved residues located in the C-termini of NsArsI, andthe previous studies indicated that removal of these residues inother As resistance enzymes such as ArsM improve proteinpurification and crystallization instead of losing function.41

Therefore, we constructed two truncated NsArsI derivatives(ArsI10 and ArsI23) to verify whether the C-terminal residueshave an impact on function. When AW3110 expressing ArsI10or ArsI23 was incubated in the medium with 3 μM MAs(III),MAs(III) was completely converted to As(III), and the rate wasalmost as the same as AW3110 expressing the full-lengthNsArsI (ArsIall) (Figure 5A). However, in the in vitro reactionwith the addition of 30 μM MAs(III), the purified ArsI10 orArs23 only demethylated 2.1% or 15.6% MAs(III) while ArsIallwas able to demethylate up to 29.6% MAs(III) (Figure 5B).This suggested that the nonconserved residues, despite notbeing critical for MAs(III) demethylation, still had negativeeffects on the efficiency of MAs(III) demethylation.

Figure 3. MAs(III) demethylating functional verification of NsArsI.(A) Changes in As species in MOPS buffer (pH 7.0) for assays ofMAs(III) demethylation by NsArsI at 37 °C over 1 h. (B) HPLC−ICP-MS analysis of the enzyme-bound arsenicals of NsArsI in vitroassays under the same conditions in (A) for the indicated times.

Figure 4. Possibility of the oligomeric state of NsArsI and the kineticsof the MAs(III) demethylation reaction catalyzed by NsArsI. (A) Thestandard curve for protein molecular mass using analytical size-exclusion chromatography on a HiLoad 16/600 Superdex 75 column.The values for Ve of NsArsI shown as red dots are 60.9 and 69.4 mL.(B) The kinetic curve of NsArsI for MAs(III) demethylation with theindicated MAs(III) concentrations (0.5, 2, 3, 4, 5, 10, 20, and 40 μM).

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14354

■ DISCUSSIONTo our knowledge, NsArsI is the first identified andcharacterized MAs(III) demethylase from photosyntheticorganisms (cyanobacteria). Our results showed that Nostoccould biotransform both MAs(V) and MAs(III) into As(III).For MAs(V) demethylation by Nostoc, the detection ofMAs(III) at the early stage indicated that MAs(V)demethylation was a two-step process: MAs(V) reductionand MAs(III) demethylation. The previous research showedthat two microbes were required for MAs(V) reduction andMAs(III) demethylation, respectively.29 Herein, our resultsshowed that Nostoc could carry out both steps. The tinyamounts of intermediate MAs(III) suggested that NsArsIdemethylated MAs(III) to As(III) as soon as MAs(III) wasproduced through MAs(V) reduction. Consequently, MAs(V)reduction may be the rate-limiting step of MAs(V)demethylation by Nostoc. However, the enzyme in Nostocresponsible for MAs(V) reduction to MAs(III) is still notknown. For MAs(III) demethylation by Nostoc, As(III) as themain product of MAs(III) demethylation was not extensivelyaccumulated within the Nostoc cells becauseAs(III) can beextruded by efflux via Acr3 or ArsB.42 However, in Nostoc cells,MAs(V) was detected as the main arsenic species after exposureto MAs(III). Enzymatic MAs(III) oxidation may be one of the

important sources of the intracellular MAs(V). Recently, ArsHwas found to be a trivalent organoarsenical oxidase,43 but theArsH ortholog was not found in Nostoc, and MAs(III) might benonspecifically oxidized by other oxidases. Therefore, ArsI isconsidered to be the dominant enzyme involved in MAs(III)detoxification in Nostoc. Indeed, our results clearly showed thatMAs(III) was almost entirely demethylated to As(III) and theneffluxed to the medium.MAs(III) demethylation is a pathway to detoxify more

acutely toxic MAs(III) to As(III). MAs(III) in the environmentis mainly generated by ArsM.44 The widespread distribution ofarsMs implies that arsenic methylation occurred on early Eartheven before the atmosphere became oxidizing.1,11 MAs(III) asthe initial product of As methylation was stable in the absenceof oxygen, similar to how the organisms on early Earth musthave evolved MAs(III) resistance mechanisms to protectthemselves. Cyanobacteria with oxygenic photosynthesis werethe earliest important producers of atmospheric oxygen.45,46

When cyanobacteria were under MAs(III) selective pressure,the evolution of ArsI using self-generated molecular oxygen tobreak the C−As bond of MAs(III) seems quite economical andeffective. Although MAs(III) can be rapidly oxidized in thepresent oxidizing atmosphere, MAs(III) still exists in thenatural environment. For example, a high concentration ofMAs(III) was found in groundwater contaminated by herbicideproduction,47 and MAs(III) could also be detected infreshwaters.48 In addition, the widely used MASA wasconstantly reduced by various organisms to MAs(III) as itsactive form.49 Thus, MAs(III) demethylation catalyzed by ArsImake cyanobacteria stay away from the poison of MAs(III)throughout the ages.Because methylation and demethylation are two opposite

arsenic metabolic pathways, it should be noted that two of thepathways coexist in Nostoc because they have both arsM35 andarsI genes. After exposure to higher As(III) concentrations(e.g., 100 μM), Nostoc methylated As(III) to DMAs(V) andvolatile TMAs(III) with ArsM35 or further producedarsenosugars with unknown enzymes.36 MAs(III), a predictedintermediate in the pathway of methylation, was not detectedduring As methylation in Nostoc.35 It is likely that the producedMAs(III) could be quickly demethylated to As(III) by ArsI oroxidized to MMA(V) by enzymatic or nonenzymatic reactions.When Nostoc was treated with lower concentrations of As(III)(e.g., 10 μM), As(III) cannot be methylated but was detoxifiedby direct efflux out of cells or oxidation to As(V) as describedby Yin et al.35 The present study suggested that As(III)produced by demethylation of MAs(III) at 1 μM is not enoughto remethylate to organoarsenicals. In short, cyanobacteria haveevolved versatile detoxification mechanisms to deal with variousarsenic species. Because MAs(III) is more toxic than As(III),MAs(III) demethylation with higher reaction rate and lowerconcentration of substrate than As(III) methylation seems tobe critical for the survival of cyanobacteria in arsenic-contaminated environments. Furthermore, MAs(III) is consid-ered as a primitive antibiotic, and cyanobacteria are “smart”enough to rid themselves of MAs(III) by demethylation.ArsI is considered a nonheme Fe(II)-dependent extradiol

dioxygenase with C·As lyase activity.11 Our results indicatedthat NsArsI exists as a trimer in the solution. It is not surprisingthat the extradiol dioxygenases always exist in a range ofoligomeric states;50 for example, gallate dioxygenase fromPseudomonas putida KT2440 was also a trimer composed ofthree identical subunits.51 In accordance with the trimeric state

Figure 5. MAs(III) demethylation of MAs(III) by full-length NsArsIand truncated derivatives. (A) In vivo MAs(III) demethylation assayswere conducted in the ST medium exposed to 3 μM MAs(III) andinoculated AW3110 expressing arsIall, arsI10, or arsI23. AW3110bearing pET22b was used as the control. (B) In vitro assays ofMAs(III) demethylation using purified proteins (ArsIall, ArsI10, andArsI23) were tested in MOPS buffer (pH 7.0) containing 30 μMMAs(III). The reaction solution without adding NsArsI was thecontrol.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14355

of NsArsI, kinetic characterization of NsArsI suggested thatthere are multiple ligand-binding sites in one protein molecule.The turnover number for NsArsI was only 0.079 min−1, and thepossible reason for the low value is that both the enzyme andthe substrate for MAs(III) demethylation are readily oxidized.11

The C-terminal truncated derivatives of NsArsI with differentlength of nonconserved residues deletion showed restrictedactivity for C·As cleavage. Meanwhile, the C-terminal tail ofAkbC as a type I extradiol dioxygenase is responsible forsubstrate specificity by blocking the entrance and plays a role inshielding the binding pocket from oxidation and the correctpositioning of the substrate.52 Hence, we suggest that the C-terminal residues of NsArsI may contribute to the catalyticactivity by helping the substrate positioning for enzymaticattack, although these residues are not directly involved in thecatalytic process. More studies are needed for the furtherunderstanding of reaction mechanism of NsArsI by thecombined use of structural, kinetic, and computationalapproaches. ArsI10 with the highest activity of the truncatedderivatives may be the first choice for use in crystallization todetermine the crystal structure.

■ ENVIRONMENTAL IMPLICATIONS

Methylated arsenicals such as MSMA are broad-spectrumherbicides that are widely used in agriculture, and approx-imately 1 360 000 kg of MSMA have been applied annually inthe United States (www.epa.gov/oppsrrd1/REDs/organic_arsenicals_red.pdf). Although the usage of MSMA is nowrestricted to sod farms, golf courses, highway rights-of-way, andcotton fields by the United States Environment ProtectionAgency (USEPA),53 the repeated application of MSMA has ledto the contamination of soils.54 In addition, other organo-arsenicals containing aromatic compounds such as roxarsone(4-hydroxy-3-nitrophenylarsonic acid) have been extensivelyused as growth promoters in intensive animal farming.Although the USEPA has withdrawn approval of roxarsone inanimal feed formulations55 in light of environmental concernsof roxarsone used in broiler poultry feed,56 it may still be widelyused in the animal industry in other parts of the world. Themicrobial degradation of organoarsenicals will be an importantprocess in controlling the behavior and fate of organoarsenics inthe environment. Except for catalyzing MAs(III) demethyla-tion, ArsI also has the ability to cleave the C·As bond intrivalent roxarsone and other aromatic arsenicals.11 Cyanobac-teria are abundant in the aquatic environments and soils. In ourstudy, Nostoc with NsarsI was shown to degrade methylatedarsenic into As(III), and it is very likely that Nostoc can degradearomatic arsenicals, too. Therefore, it is believed thatcyanobacteria are likely involved in organoarsenic degradationand may play a critical role in arsenic biogeochemistry.In conclusion, we have demonstrated that photosynthetic

organism Noctoc is capable of demethylation of MAs(III), andwe functionally characterized a novel C·As lyase, NsArsI, fromNostoc. NsArsI exhibited the ability of C·As cleavage todemethylate MAs(III). MAs(III) demethylation is an importantmechanism of MAs(III) detoxification in Nostoc. Cyanobacteria,including Nostoc, are well-known for their contribution toAs(III) methylation14,35 and arsenosugars36 and arsenolipids57

biosynthesis. Our results of cyanobacteria involved inmethylarsenic demethylation open up new horizons in the Asbiotransformation by cyanobacteria. Given the high abundanceof cyanobacteria in aquatic systems, wetlands, and soils, these

photosynthetic prokaryotic organisms may play a significantrole in As biogeochemical cycling.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.est.5b03357.

Figure S1: As species in the cell lysate of Nostoc exposedto MAs(III) for the indicated times. Figure S2: Multiplealignment of ArsI homologues. Figure S3: Graph of SDS-PAGE for purified full-length NsArsI (ArsIall) and twotruncated derivatives (ArsI10 and ArsI23). Figure S4:Elution profiles of NsArsI and molecular-mass standardsfrom a HiLoad 16/600 Superdex 75 column. (PDF)

■ AUTHOR INFORMATIONCorresponding Author*Phone: 86-592-6190977; fax: 86-592-6190977; e-mail:ygzhu@iue.ac.cn.

Author Contributions∥These authors contributed equally to this work.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe thank Professor Wen-Li Chen of Huazhong AgriculturalUniversity for providing Nostoc sp. PCC 7120. This work issupported by the National Natural Science Foundation ofChina (31270161), the Chinese Academy of Sciences(XDB15020402), and the Natural Science Foundation ofFujian Province (2014J01141).

■ REFERENCES(1) Zhu, Y. G.; Yoshinaga, M.; Zhao, F. J.; Rosen, B. P. Earth abidesarsenic biotransformations. Annu. Rev. Earth Planet. Sci. 2014, 42,443−467.(2) Abernathy, C. O.; Thomas, D. J.; Calderon, R. L. Health effectsand risk assessment of arsenic. J. Nutr. 2003, 133 (5), 1536S−1538S.(3) Tchounwou, P. B.; Patlolla, A. K.; Centeno, J. A. Carcinogenicand systemic health effects associated with arsenic exposure-a criticalreview. Toxicol. Pathol. 2003, 31 (6), 575−588.(4) Bissen, M.; Frimmel, F. H. Arsenic-a review. Part I: occurrence,toxicity, speciation, mobility. Acta Hydrochim. Hydrobiol. 2003, 31 (1),9−18.(5) Oremland, R. S.; Stolz, J. F. The ecology of arsenic. Science 2003,300 (5621), 939−944.(6) Thomas, D. J.; Waters, S. B.; Styblo, M. Elucidating the pathwayfor arsenic methylation. Toxicol. Appl. Pharmacol. 2004, 198 (3), 319−326.(7) Challenger, F. Biological methylation. Adv. Enzymol. Relat. AreasMol. Biol. 2006, 12, 429−491.(8) Bentley, R.; Chasteen, T. G. Microbial methylation of metalloids:arsenic, antimony, and bismuth. Microbiol. Mol. Biol. R. 2002, 66 (2),250−271.(9) Petrick, J. S.; Ayala Fierro, F.; Cullen, W. R.; Carter, D. E.;Vasken Aposhian, H. Monomethylarsonous acid (MMA III) is moretoxic than arsenite in Chang human hepatocytes. Toxicol. Appl.Pharmacol. 2000, 163 (2), 203−207.(10) Mass, M. J.; Tennant, A.; Roop, B. C.; Cullen, W. R.; Styblo, M.;Thomas, D. J.; Kligerman, A. D. Methylated trivalent arsenic speciesare genotoxic. Chem. Res. Toxicol. 2001, 14 (4), 355−361.(11) Yoshinaga, M.; Rosen, B. P. A C·As lyase for degradation ofenvironmental organoarsenical herbicides and animal husbandry

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14356

growth promoters. Proc. Natl. Acad. Sci. U. S. A. 2014, 111 (21),7701−7706.(12) Qin, J.; Rosen, B. P.; Zhang, Y.; Wang, G.; Franke, S.; Rensing,C. Arsenic detoxification and evolution of trimethylarsine gas by amicrobial arsenite S-adenosylmethionine methyltransferase. Proc. Natl.Acad. Sci. U. S. A. 2006, 103 (7), 2075−2080.(13) Qin, J.; Lehr, C. R.; Yuan, C.; Le, X. C.; McDermott, T. R.;Rosen, B. P. Biotransformation of arsenic by a Yellowstonethermoacidophilic eukaryotic alga. Proc. Natl. Acad. Sci. U. S. A.2009, 106 (13), 5213−5217.(14) Ye, J.; Rensing, C.; Rosen, B. P.; Zhu, Y. G. Arsenicbiomethylation by photosynthetic organisms. Trends Plant Sci. 2012,17 (3), 155−162.(15) Wang, P. P.; Sun, G. X.; Zhu, Y. G. Identification andCharacterization of Arsenite Methyltransferase from an Archaeon,Methanosarcina acetivorans C2A. Environ. Sci. Technol. 2014, 48 (21),12706−12713.(16) Kuramata, M.; Sakakibara, F.; Kataoka, R.; Abe, T.; Asano, M.;Baba, K.; Takagi, K.; Ishikawa, S. Arsenic biotransformation byStreptomyces sp. isolated from rice rhizosphere. Environ. Microbiol.2015, 17 (6), 1897−1909.(17) Matteson, A. R.; Gannon, T. W.; Jeffries, M. D.; Haines, S.;Lewis, D. F.; Polizzotto, M. L. Arsenic retention in foliage and soilafter monosodium methyl arsenate (MSMA) application to turfgrass. J.Environ. Qual. 2014, 43 (1), 379−388.(18) Von Endt, D.; Kearney, P. C.; Kaufman, D. D. Degradation ofMSMA by soil microorganisms. J. Agric. Food Chem. 1968, 16 (1), 17−20.(19) Sanders, J. G. Microbial role in the demethylation and oxidationof methylated arsenicals in seawater. Chemosphere 1979, 8 (3), 135−137.(20) Bednar, A.; Garbarino, J.; Ranville, J.; Wildeman, T. Presence oforganoarsenicals used in cotton production in agricultural water andsoil of the southern United States. J. Agric. Food Chem. 2002, 50 (25),7340−7344.(21) Feng, M.; Schrlau, J. E.; Snyder, R.; Snyder, G. H.; Chen, M.;Cisar, J. L.; Cai, Y. Arsenic transport and transformation associatedwith MSMA application on a golf course green. J. Agric. Food Chem.2005, 53 (9), 3556−3562.(22) Sierra Alvarez, R.; Yenal, U.; Field, J. A.; Kopplin, M.; Gandolfi,A. J.; Garbarino, J. R. Anaerobic biotransformation of organoarsenicalpesticides monomethylarsonic acid and dimethylarsinic acid. J. Agric.Food Chem. 2006, 54 (11), 3959−3966.(23) Maki, T.; Watarai, H.; Kakimoto, T.; Takahashi, M.; Hasegawa,H.; Ueda, K. Seasonal dynamics of dimethylarsenic acid degradingbacteria dominated in Lake Kibagata. Geomicrobiol. J. 2006, 23 (5),311−318.(24) Maki, T.; Hirota, W.; Ueda, K.; Hasegawa, H.; Azizur Rahman,M. Seasonal dynamics of biodegradation activities for dimethylarsinicacid (DMA) in Lake Kahokugata. Chemosphere 2009, 77 (1), 36−42.(25) Maki, T.; Takeda, N.; Hasegawa, H.; Ueda, K. Isolation ofmonomethylarsonic acid-mineralizing bacteria from arsenic contami-nated soils of Ohkunoshima Island. Appl. Organomet. Chem. 2006, 20(9), 538−544.(26) Crecelius, E. A. Arsenite and arsenate levels in wine. Bull.Environ. Contam. Toxicol. 1977, 18 (2), 227−230.(27) Cullen, W. R.; McBride, B. C.; Pickett, A. W. Thetransformation of arsenicals by Candida humicola. Can. J. Microbiol.1979, 25 (10), 1201−1205.(28) Lehr, C. R.; Polishchuk, E.; Radoja, U.; Cullen, W. R.Demethylation of methylarsenic species by Mycobacterium neoaurum.Appl. Organomet. Chem. 2003, 17 (11), 831−834.(29) Yoshinaga, M.; Cai, Y.; Rosen, B. P. Demethylation ofmethylarsonic acid by a microbial community. Environ. Microbiol.2011, 13 (5), 1205−1215.(30) Zhang, Y. B.; Monchy, S.; Greenberg, B.; Mergeay, M.; Gang,O.; Taghavi, S.; van der Lelie, D. ArsR arsenic-resistance regulatoryprotein from Cupriavidus metallidurans CH34. Antonie van Leeuwen-hoek 2009, 96 (2), 161−170.

(31) Nadar, S. V.; Yoshinaga, M.; Kandavelu, P.; Sankaran, B.; Rosen,B. P. Crystallization and preliminary X-ray crystallographic studies ofthe ArsI C-As lyase from Thermomonospora curvata. Acta Crystallogr.,Sect. F: Struct. Biol. Commun. 2014, 70 (6), 761−764.(32) Bahl, J.; Lau, M. C.; Smith, G. J.; Vijaykrishna, D.; Cary, S. C.;Lacap, D. C.; Lee, C. K.; Papke, R. T.; Warren-Rhodes, K. A.; Wong,F. K.; McKay, C. P.; Pointing, S. B. Ancient origins determine globalbiogeography of hot and cold desert cyanobacteria. Nat. Commun.2011, 2, 163.(33) Fiore, M. F.; Moon, D. H.; Trevors, J. T. Metal resistance andaccumulation in Cyanobacteria. In Wastewater Treatment with Algae;Wong, Y. S., Tam, N. F. Y., Eds.; Springer-Verlag: Berlin, 1998; pp111−124.(34) Kaneko, T.; Nakamura, Y.; Wolk, C. P.; Kuritz, T.; Sasamoto, S.;Watanabe, A.; Iriguchi, M.; Ishikawa, A.; Kawashima, K.; Kimura, T.Complete genomic sequence of the filamentous nitrogen-fixingcyanobacterium Anabaena sp. strain PCC 7120. DNA Res. 2001, 8(5), 205−213.(35) Yin, X. X.; Chen, J.; Qin, J.; Sun, G. X.; Rosen, B. P.; Zhu, Y. G.Biotransformation and volatilization of arsenic by three photosyntheticcyanobacteria. Plant Physiol. 2011, 156 (3), 1631−1638.(36) Miyashita, S. i.; Fujiwara, S.; Tsuzuki, M.; Kaise, T.Cyanobacteria produce arsenosugars. Environ. Chem. 2012, 9 (5),474−484.(37) Rippka, R.; Deruelles, J.; Waterbury, J. B.; Herdman, M.; Stanier,R. Y. Generic assignments, strain histories and properties of purecultures of cyanobacteria. J. Gen. Microbiol. 1979, 111 (1), 1−61.(38) Carlin, A.; Shi, W.; Dey, S.; Rosen, B. P. The ars operon ofEscherichia coli confers arsenical and antimonial resistance. J. Bacteriol.1995, 177 (4), 981−986.(39) Zhang, S. Y.; Sun, G. X.; Yin, X. X.; Rensing, C.; Zhu, Y. G.Biomethylation and volatilization of arsenic by the marine microalgaeOstreococcus tauri. Chemosphere 2013, 93 (1), 47−53.(40) Dheeman, D. S.; Packianathan, C.; Pillai, J. K.; Rosen, B. P.Pathway of human AS3MT arsenic methylation. Chem. Res. Toxicol.2014, 27 (11), 1979−1989.(41) Marapakala, K.; Ajees, A.; Qin, J.; Sankaran, B.; Rosen, B. P.Crystallization and preliminary X-ray crystallographic analysis of theArsM arsenic(III) S-adenosylmethionine methyltransferase. ActaCrystallogr., Sect. F: Struct. Biol. Cryst. Commun. 2010, 66 (9), 1050−1052.(42) Rosen, B. P.; Tamas, M. J. Arsenic transport in prokaryotes andeukaryotic microbes. Adv. Exp. Med. Biol. 2010, 679, 47−55.(43) Chen, J.; Bhattacharjee, H.; Rosen, B. P. ArsH is anorganoarsenical oxidase that confers resistance to trivalent forms ofthe herbicide monosodium methylarsenate and the poultry growthpromoter roxarsone. Mol. Microbiol. 2015, 96 (5), 1042−1052.(44) Thomas, D. J.; Rosen, B. P. Arsenic methyltransferases. InEncyclopedia of Metalloproteins; Kretsinger, R. H., Uversky, V. N.,Permyakov, E. A., Eds.; Springer New York: New York, 2013; pp 138−143.(45) Rasmussen, B.; Fletcher, I. R.; Brocks, J. J.; Kilburn, M. R.Reassessing the first appearance of eukaryotes and cyanobacteria.Nature 2008, 455 (7216), 1101−1104.(46) Lyons, T. W.; Reinhard, C. T.; Planavsky, N. J. The rise ofoxygen in Earth/’s early ocean and atmosphere. Nature 2014, 506(7488), 307−315.(47) McKnight Whitford, A.; Chen, B.; Naranmandura, H.; Zhu, C.;Le, X. C. New method and detection of high concentrations ofmonomethylarsonous acid detected in contaminated groundwater.Environ. Sci. Technol. 2010, 44 (15), 5875−5880.(48) Azizur Rahman, M.; Hasegawa, H. Arsenic in freshwatersystems: Influence of eutrophication on occurrence, distribution,speciation, and bioaccumulation. Appl. Geochem. 2012, 27 (1), 304−314.(49) Chen, J.; Sun, S.; Li, C. Z.; Zhu, Y. G.; Rosen, B. P. Biosensorfor organoarsenical herbicides and growth promoters. Environ. Sci.Technol. 2014, 48 (2), 1141−1147.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14357

(50) Vaillancourt, F. H.; Bolin, J. T.; Eltis, L. D. The ins and outs ofring-cleaving dioxygenases. Crit. Rev. Biochem. Mol. Biol. 2006, 41 (4),241−267.(51) Nogales, J.; Canales, A.; Jimenez Barbero, J.; García, J. L.; Díaz,E. Molecular Characterization of the Gallate Dioxygenase fromPseudomonas putida KT2440 THE PROTOTYPE OF A NEWSUBGROUP OF EXTRADIOL DIOXYGENASES. J. Biol. Chem.2005, 280 (42), 35382−35390.(52) Cho, H. J.; Kim, K.; Sohn, S. Y.; Cho, H. Y.; Kim, K. J.; Kim, M.H.; Kim, D.; Kim, E.; Kang, B. S. Substrate binding mechanism of atype I extradiol dioxygenase. J. Biol. Chem. 2010, 285 (45), 34643−34652.(53) Agency, E. Organic arsenicals: product cancellation order andamendments to terminate uses. Fed. Regist. 2009, 74, 50187−50194.(54) Mahoney, D. J.; Gannon, T. W.; Jeffries, M. D.; Matteson, A. R.;Polizzotto, M. L. Management considerations to minimize environ-mental impacts of arsenic following monosodium methylarsenate(MSMA) applications to turfgrass. J. Environ. Manage. 2015, 150,444−450.(55) U.S. Food and Drug Administration. On September 30, USFDAwithdraws approval of three out of the four arsenicals used in animalfeed formulations including roxarsone, carbarsone and arsanilic, 2013;http://www.centerforfoodsafety.org/files/20130930_docketfda-2009-p-0594_signed-arsenic-cp-response_94793.pdf.(56) Fisher, D. J.; Yonkos, L. T.; Staver, K. W. EnvironmentalConcerns of Roxarsone in Broiler Poultry Feed and Litter in Maryland,USA. Environ. Sci. Technol. 2015, 49 (4), 1999−2012.(57) Xue, X. M.; Raber, G.; Foster, S.; Chen, S. C.; Francesconi, K.A.; Zhu, Y. G. Biosynthesis of arsenolipids by the cyanobacteriumSynechocystis sp. PCC 6803. Environ. Chem. 2014, 11 (5), 506−513.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b03357Environ. Sci. Technol. 2015, 49, 14350−14358

14358