Bioresource Technology 169 (2014) 644–651

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Long-term effect of the antibiotic cefalexin on methane productionduring waste activated sludge anaerobic digestion

http://dx.doi.org/10.1016/j.biortech.2014.07.0560960-8524/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Department of Civil and Environmental Engineering,Graduate School of Engineering, Tohoku University, Sendai, Miyagi 980-8579, Japan.

E-mail addresses: luxq@epl1.civil.tohoku.ac.jp (X. Lu), zhenguangyin@163.com(G. Zhen), yyli@epl1.civil.tohoku.ac.jp (Y.-Y. Li).

Xueqin Lu a, Guangyin Zhen a,b,⇑, Yuan Liu c, Toshimasa Hojo a, Adriana Ledezma Estrada c, Yu-You Li a,c

a Department of Civil and Environmental Engineering, Graduate School of Engineering, Tohoku University, Sendai, Miyagi 980-8579, Japanb National Institute for Environmental Studies, Onogawa 16-2, Tsukuba, Ibaraki 305-0053, Japanc Graduate School of Environmental Studies, Tohoku University, Sendai, Miyagi 980-8579, Japan

h i g h l i g h t s

� Long-term effect of cefalexin onmethane production during sludgefermentation was assessed.� Cefalexin exhibited a temporary

inhibition in methane production,followed by a marked recover.� The highest methane yield was

450 mL at 1000 mg-CLX/L after157 days of digestion, rising by 63.8%.� Extracellular polymeric substances

protected microbes for self-growthand fermentation.

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 June 2014Received in revised form 13 July 2014Accepted 14 July 2014Available online 21 July 2014

Keywords:Methane productionCefalexinExtracellular polymeric substances (EPS)Ultraviolet visible (UV–Vis) spectra

a b s t r a c t

Long-term experiments herein were conducted to investigate the effect of cefalexin (CLX) on methaneproduction during waste activated sludge (WAS) anaerobic digestion. CLX exhibited a considerable inhi-bition in methane production during the initial 25 days while the negative effect attenuated subse-quently and methane production recovered depending on CLX doses used (600 and 1000 mg/L). Thehighest methane yield reached 450 mL at 1000 mg-CLX/L after 157 days of digestion, 63.8% higher thanCLX-free one. Stimulated excretion of extracellular polymeric substances (EPS) by CLX served as micro-bial protecting layers, creating a suitable environment for microbes’ growth and fermentation. Furtherexamination via ultraviolet visible (UV–Vis) spectra also verified the elevated slime EPS, LB-EPS andTB-EPS indicated by UV-254 in the presence of CLX. Unlike the commonly accepted adverse effect, thisstudy demonstrated the beneficial role of CLX in methane production, providing new insights into its trueenvironmental impacts.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

A high proportion of antibiotics can enter aquatic ecosystemsdirectly, through the discharge of wastewater from the pharma-ceutical industries, households and hospitals, or indirectly, by

leaching and runoff of agricultural soils amended with manurefrom livestock (Sara et al., 2013). The release of antibiotics isbecoming of considerable concern since it can cause negativeeffects not only on the environment but also on human health.Cefalexin (CLX), as one of the most prescribed antibiotics, belongsto the first generation cephalosporin type used in human medicineto treat aspiratory path infection, urine pathway infection and shintissue infection and also in veterinary medicine due to itsenhanced oral activity (Zhai, 2012). The widespread and overuseof CLX has lead to its detection in the aquatic environment. For

Table 1Characteristics of excess sludge and seed sludge used in this study.

Items Seed sludge WAS

pH 6.90–6.91 6.44–6.47Moisture content (%) 84.81 90.43TSS (g/L) 72.14 21.00VSS (g/L) 67.16 4.76TCOD (mg/L) 94890.26 30374.18SCOD (mg/L) 14591.76 27.32NH4

+-N (mg/L) 1369.18 22.34TPS (mg/L) 24064.87 4385.68SPS (mg/L) 1439.81 �TPN (mg/L) 23670.67 8111.93SPN (mg/L) 2273.32 1194.21

TSS: total suspended solids; VSS: volatile suspended solids; TCOD: total chemicaldemand oxygen; SCOD: soluble chemical demand oxygen; TPN: total protein; SPN:soluble protein; TPS: total polysaccharide; SPS: soluble polysaccharide.

X. Lu et al. / Bioresource Technology 169 (2014) 644–651 645

example, the concentration of CLX present in municipal wastewa-ter was found to be 339.4–375 ng/L (Guo et al., 2010); other stud-ies reported by Estrada et al. (2012), as well as Saravanane andSundararaman (2009) also showed that the effluent from a phar-maceutical drugs factory in India nearly reached 29 mg/L aftertreatment, suggesting that the potential effect of antibiotics inthe environment requires further attention.

Anaerobic digestion (AD), a well-established technology, is themost frequently used biological process for sludge treatment(Zhen et al., 2014). Owing to heavy use of pharmaceutical drugs,high concentrations in the slurry influent to wastewater treatmentplants (WWTPs) will inevitably result in the effects on the mixedpopulation of anaerobic bacteria during the anaerobic processdue to these molecules’ biological residual activity (Sara et al.,2013). A recent study by Ince et al. (2013) confirmed that numbersof active bacteria and methanomicrobiales were negatively corre-lated with the concentrations of antibiotic oxytetracycline, inwhich around 50–60% reduction in biogas production wereobtained with respect to the control depending on the applieddoses. Over the last few years, a considerable amount of workhas been done on assessing the effect of antibiotics on methaneproduction during anaerobic digestion, however, the results abouttheir definite functions are inconsistent, and in some cases evencontradictory. Masse et al. (2000), for instance, studied the effectof six antibiotics (tylosin, lincomycin, tetracycline, sulphameth-azine, penicillin and carbadox) on the psychrophilic anaerobicdigestion of swine manure slurries in sequencing batch reactors(SBRs). The results indicated that presence of penicillin and tetra-cycline caused 35% and 25% decrease in methane output, respec-tively; in contrast, the slurries from pigs receiving otherantibiotics did not significantly influence methane production.Similarly, Mitchell et al. (2013) obtained the significant inhibitionof anaerobic digestion of cattle manure containing florfenicol atconcentrations of 6.4, 36 and 210 mg/L, where biogas yield reducedby 5%, 40% and 75%, respectively; whereas sulfamethazine over theevaluated concentration range of 0.28–280 mg/L imparted noobserved effect on biogas production during the 40 days’ incuba-tion. Another study by Sara et al. (2013) investigating the effectof three classes of veterinary antibiotics (danofloxacin, micospec-tone and ceftiofur) demonstrated that the former two showedobvious inhibition on pig manure anaerobic digestion, reducingbiogas production by 10–15% and 18–23%, respectively; compara-tively, ceftiofur induced a much less substantial impact mainlybecause of its large biodegradability. For sulfonamide, Mohringet al. (2009) did not observe any statistically significant inhibitoryeffect on anaerobic digestion. The above inconsistent and evencontradictory findings might be attributed to differences in thetypes and concentrations of antibiotics used, sensitivity of analysisutilized to measure effects (Amin et al., 2006), and also durations ofexposure tests.

Obviously, although numerous investigations have been carriedout to explore the effect and fate of different types of antibiotics inanaerobic treatment, up to date, the questions on whether and/orhow antibiotics, particularly CLX will affect sludge anaerobic diges-tion still remain unknown. Moreover, most of current studies weremainly focused on the short-term response of anaerobic process toantibiotics (Beneragama et al., 2013; Shi et al., 2011; Mitchell et al.,2013), and the useful information on the long-term toxic impactsof CLX is definitely limited (Estrada et al., 2012; Zhai, 2012). It istherefore of considerable importance to make additional effortfor better understanding the long-term effect and exact role ofCLX on/in digestion efficiency and stability of anaerobic digestionprocess. Hence, the main objective of this study was to explorethe potential influence of the presence of CLX on methane produc-tion from WAS during anaerobic digestion. The inhibitory effect ofCLX was assessed by monitoring methane production, variations of

extracellular polymeric substances (EPS), and accumulation ofindividual volatile fatty acids (VFAs). Removal of CLX was evalu-ated based on measured bulk phase concentrations of CLX.

2. Methods

2.1. Test materials

Antibiotic cefalexin (CLX) used in this study was purchasedfrom Wako (99% purify), Japan. Chemical molecular formula ofCLX is C16H17N3O4S�H2O; the molecular weight is 365.40 g/mol;and the solubility is 1790 mg/L. The waste activated sludge(WAS) was withdrawn from the secondary sedimentation tank ofa municipal WWTP in Sendai, Japan. The sludge samples weretransferred immediately to the laboratory and stored at 4 �C inorder to maintain sample freshness. The seed sludge was collecteddirectly from discharged sludge from a 5-L lab-scale completedstirred treatment reactors (CSTR) operated at 35 ± 1 �C in our lab.The reactor was made of plexiglass and controlled to 35 ± 1 �C bywater jackets and heaters. The feed sludge was comprised of thick-ened WAS. Feeding and drawing pumps were operated 6–12 timesper day by using a timer-controller. The inoculum was degassedand incubated in a water bath (35 ± 1 �C) (BT 100, Yamato, Japan)under anaerobic condition by 5–7 days until methane contentreached about 55%. The principle characteristics of the WAS andseed sludge are given in Table 1.

2.2. Biochemical methane potential (BMP) assay

Biochemical methane potential (BMP) assay (Shi et al., 2011;Zhen et al., 2014) was conducted in the series of 120 mL sealedglass serum bottles to evaluate the potential effect of CLX onWAS anaerobic digestion. The effective working volume was75 mL. The seed sludge to WAS ratio was 2:1 (volume:volume).To determine dose–response of methane output to CLX and pro-vide the possible guides for addressing the CLX-related issues incase of emergency, the investigations of a broad range of antibioticconcentrations were performed herein. CLX was initially negligiblein the WAS samples and was added into assays with the final con-centrations of 0, 50, 200, 400, 600, 1000, and 2000 mg/L. The pH ofall samples was then carefully neutralized to around 7.0 before thestart-up of reactors. A control with 75 mL of inoculums was con-ducted to determine the biogas production from endogenous inspi-ration. After being sealed with butyl rubber stoppers secured withaluminum crimp, the bottles were flushed with high-purity nitro-gen gas at the flow rate of 0.5 L/min for about 2 min to ensure ananaerobic atmosphere before the beginning of fermentation. Afterthat, all bottles were placed in a water bath (BT 100, Yamato,

Duration time (days)

0 25 50 75 100 125 150

Cum

ulat

ive

met

hane

(mL)

0

100

200

300

400

500

Duration time (days)0 10 20 30 40

Cum

ulat

ive

met

hane

(mL)

0

5

10

15

20

Fig. 1. Net cumulative methane production at different CLX concentrations in thecourse of duration time. Values are the means from duplicate digesters.

646 X. Lu et al. / Bioresource Technology 169 (2014) 644–651

Japan) at 35 ± 1 �C under shaking of 100 ± 1 rpm. The biogas yield,during the digestion period about 157 days, was measured throughreleasing the pressure in the bottles using 10–50 mL glass syringes.Biogas composition (CH4 and CO2) was determined using gas chro-matograph (GC) equipped with a thermal conductivity detector(GC-8A, Shimadzu, Japan). All digestions have been done induplicate to ensure the reproducibility of experimental data, andaverage values were then given.

The modified Gompertz model, the most suitable model fordescribing the relationship between bacterial growth and meta-bolic production, was employed for analysis of the BMP data. Themaximum cumulative biogas production (Hm, mL), the maximummethane production rate (Rm, mL/d), and the lag-phase (k, day)were estimated by fitting the data into the Gompertz model asshown below.

Y t ¼ Hm � exp � expRmeHmðk� tÞ þ 1

� �� �ð1Þ

where Y is the accumulative biogas yield [mL CH4]; Hm is themaximum cumulative gas production (mL); Rm is the maximummethane production rate (mL/d); k is the lag-phase time (day);and e is exponential 1, which is 2.71828.

2.3. Methane producing activity

Batch tests to determine the methane producing activity (MPA)of digested sludge from BMP tests with synthetic wastewaters of2000 mg-COD/L of sodium acetate (model VFA compound) wereperformed. The synthetic wastewater consisted of (g/L of distilledwater) 1000 MgCl2�6H2O, 375 CaCl2�2H2O, 1250 NH4Cl, 2180K2HPO4, 1700 KH2PO4, 2500 NaHCO3, 500 Cystein�HCl�H2O,5 FeCl2�4H2O, 0.425 CoCl2�6H2O, 0.175 ZnCl2, 6 H3BO3, 0.15H3BO3, 1.25 MnCl2�2H2O, 0.1 NiCl2�6H2O, 0.0675 CuCl2�2H2O,0.0625 NaMoO4�2H2O, 12.5 EDTA, and 1.25 resazurin; resazurinwas used as a redox indicator. Prior to use, the synthetic wastewa-ter was boiled for around 2 h to remove the dissolved oxygen. Syn-thetic wastewater of 60 mL and 20 mL inocula incubated duringBMP tests were added into each serum bottle. The bottles weresealed with rubber stoppers secured by aluminum crimp, andflushed with nitrogen gas to remove traces of oxygen in headspace.Meanwhile, 1 mL Na2S�9H2O (250 mg/L), used as the reducingagent, was injected into each bottle to obtain an absolutely anaer-obic condition. After that, all bottles were incubated in a waterbath (100 ± 1 rpm) at 35 ± 1 �C. By the analysis of methane produc-tion, the long-term impact of CLX on the methanation wasobtained.

2.4. Extracellular polymeric substances

Extracellular polymeric substances (EPS) of sludge before andafter fermentation were extracted according to our previousresearch (Zhen et al., 2013a). Three types of EPS, i.e., slime EPS,loosely bound EPS (LB-EPS) and tightly bound EPS (TB-EPS) werecollected, respectively. After filtered through a 0.45 lm syringe-driven filter (Millipore Co., USA), proteins (PN) and polysaccharides(PS) in different EPS fractions were then analyzed. Furthermore,the UV–Vis spectra of EPS were recorded utilizing a UV–Vis spec-trophotometer (DR 5000, HACH Co., USA) in a 1 cm quartz cuvette.The UV–Vis spectra were collected with absorbance scanning from190 to 500 nm at 2 nm increments and were presented as absor-bance. The UV–Vis spectra of deionized water were collected toeliminate the background noise. Besides, the UV-254 absorbanceof different EPS fractions were also collected based on UV–Visspectra to roughly estimate the contents of conjugated aromaticmolecules.

2.5. Other analytical methods

The pH, moisture content, NH4+-N, TSS and VSS were estimated

following the Standard Methods (APHA, 1998). Both TCOD andSCOD (after filtration through 0.45 lm) were determined by thesemi-automated colorimetric method at 600 nm with a DR 5000UV–Vis spectrophotometer. PN were measured using the modifiedLowry method (Frolund et al., 1996), and its absorbance was mea-sured at 750 nm. PS were stained with the phenol/H2SO4 method(Dubois et al., 1956), and its absorbance was measured at490 nm, using glucose as the standard. The concentrations of VFAsincluding acetic, propionic, butyric, isobutyric, valeric and isova-leric acids were determined by a Agilent 6890 gas chromatographyequipped with a flame ionization detector (FID) using helium ascarrier gas (38.6 kPa, 5.5 mL/min) in a DB-WAXetr capillary col-umn (30 m � 0.53 mm I.D., 1 lm). The biogas composition (CH4

and CO2) was analyzed using a gas chromatograph (SHIMADZUGC-8A) equipped with a thermal conductivity detector (TCD,80 mA) and an 2 m stainless steel column packed with PorapakQ. Helium was the carrier gas at a flow rate of 30 mL/min. CLX con-centrations were determined with a high-pressure liquid chroma-tography (HPLC) equipped with a UV detector (262 nm) (WaterAssociates, USA). Spherisorb ODS-2.25 cm � 4.6 cm column wasused as the stationary phase. The mixture of 0.1 mM NaH2PO4

and methanol (75: 25) was used as the mobile phase. The flow ratewas 0.4 mL/min with the column temperature of 30 �C andretention time of 1.6 min.

3. Results and discussion

3.1. Biochemical methane potential (BMP) assay

3.1.1. Methane productionThe methane production during anaerobic digestion is com-

monly used as an indicator of anaerobic digester performance.Cumulative methane production (CMP) during anaerobic digestiontrials for each dose of CLX tested (50, 200, 400, 600, 1000, and2000 mg/L) was recorded and compared with no-CLX control.Fig. 1 shows the variation of the methane yield (mL) obtainedagainst digestion time (day), and corresponding methane contentis presented in Fig. 1S. The modified Gompertz model (Eq. (1))was used to fit the data of each CMP curve. Model lines presentedin Fig. 1 are based on the best fit of the maximum cumulative gasproduction (Hm, mL), maximum methane production rate (Rm, mL/d)and the lag-phase time (k, day) (Fig. 2S in Supplementaryinformation).

0 500 1000 1500 20000.004

0.006

0.008

0.010

MPA

(gC

OD

/gV

SS. d

ay)

CLX concentration(mg/L)

(a)

2100 50(b)

X. Lu et al. / Bioresource Technology 169 (2014) 644–651 647

As illustrated in Fig. 1, a dose–effect was found; with the risingconcentration of CLX the net methane yield decreased substan-tially at the beginning (0–25 days), suggesting that addition ofCLX has exerted some pressure on methanogens, which reducedthe activity of microbes (Shi et al., 2011). As a result of this, a tem-porary adverse impact on methane generation occurred (see theinset of Fig. 1), and a lag-phase thus could be observed at this stage.As the anaerobic digestion process advanced the acclimation toCLX increased the tolerance of anaerobic inocula to CLX toxicity.Dreher et al. (2012), investigating the effect of antimicrobial chlor-tetracycline (CTC) on the anaerobic digestion of swine manureslurry, similarly observed the rising acclimation of microbial con-sortia to CTC after a lag-phase of 56 days. In this respect, the activ-ities of methanogens were effectively revived after a short periodof acclimation depending on the CLX concentrations applied, theadverse impact induced by CLX ceased gradually and the genera-tion of methane began to increase for the duration of experiment(i.e., 26–157 days). For instance, the methane yield with 200 and600 mg/L of CLX at the end of fermentation (157 days) was 5.7%and 30.3% higher than the CLX-free control, respectively. Moreover,methane yield could be further increased and reached the peak of450.84 mL as 1000 mg/L of CLX was added, increasing by around63.8% with respect to the control. But a sharp drop of methaneyield was detected when increasing CLX into 2000 mg/L. Evidentlythe effect of CLX was no strict dose-dependent, in conformity withthe results obtained from Pearson’s correlation (p > 0.05) (Fig. 2S inSupplementary information). From the abovementioned results itcan be noted that the increased CLX concentration does not alwayscause the elevated inhibition to methane production; the use of anappropriate dose of CLX might be, sometimes, helpful to enhancemethane production. Similar findings for CLX or other kinds ofantibiotics were also reported in the literature, such as theincreased methane yield in the presence of antibiotic tylosin dur-ing batch anaerobic swine manure digestion throughout 216-daystudy period (Stone et al., 2009). Guo et al. (2012) reported thatbiogas output from pig manure in the presence of sulfadiazine sur-passed the control after 42 days of digestion in semi-continuousexperiments at 38 �C. Further, Arikan et al. (2006) determinedthe fate of oxytetracycline during mesophilic anaerobic digestionof manure from medicated calves and did not observe significantlynegative effects on process stability, i.e., biogas methane content orreductions of volatile solids and soluble organic carbon. The resultssupported the standpoints of Amin et al. (2006) and others thatshort-term assays are not sufficient to explore the potential influ-ences of antibiotics on complex microbial systems because of lackof considerations of adaption of the biomass. Unlike the commonlybelieved adverse effect related to antibiotics observed before, theresults in this study demonstrated the beneficial role of CLX inmethane production, providing new insights into its true environ-mental impacts.

0 2 4 6 8 10 12 14 160

300

600

900

1200

1500

1800

CLX

con

cent

ratio

n (m

g/L)

Duration time (days)

200 600 1000 2000

Fig. 2. (a) MPA values at different CLX concentrations, and (b) removal of CLXduring 16 days of MPA tests.

3.1.2. Specific methanogenic activitySludge anaerobic digestion usually undergoes solubilization of

particulate organic-carbon, hydrolysis, acidification and methana-tion. Methanogenesis, as the central step of digestion, is consideredthe rate-limiting for fermentation of soluble substrates and meth-ane production (Ponsa et al., 2008). It can be anticipated that dif-ferent the activity of methane-producing microbes, relying ontheir sensitivity toward CLX, would differ in methane yield duringanaerobic process. To further evidence this, the effect of CLX on themethanation was assessed through methane production activity(MPA) tests using synthetic wastewaters with sodium acetate assubstrate (2000 mg-COD/L). CLX dose applied during MPA testsranged from 0 to 2000 mg/L and digestion period was 16days. The MPA was then calculated according to the slope of

methanogenesis in the test, and was expressed in terms of g-CODper gram VSS per day (Fig. 2a).

As given in Fig. 2a, the addition of CLX yielded an evident, butlimited influence on the activity of methanogens. An abrupt dropin the MPA from 0.010 to 0.007 g-COD/g VSS-day could be foundas CLX concentration increased from 0 to 200 mg/L. When furtherelevating CLX dose to 2000 mg/L, however, the inhibitory effectwas not provoked too much and the MPA exhibited a slight butgradual reduction. The results agreed well with the trend of meth-ane yield during the beginning of BMP tests in the inset of Fig. 1,but apparently conflicted with the ultimate methane yield at day157 (Fig. 1). Considering these observations in the context of activ-ity of methanogens, as well as ultimate methane yield, we canspeculate that the impact of CLX is time dependent, showing thatshort-term tests, e.g., MPA procedure used herein and previously,might be not suitable for disclosing the true role of antibiotics insludge anaerobic digestion. Similar observation in testing effectsof sulfonamide and tetracycline antibiotics on soil microbial activ-ity has been obtained (Thiele-Bruhn and Beck, 2005). Incipientinhibition observed in both MPA tests (16 days) and the first25 days of BMP might be due to the inhibition of CLX to proteinsynthesis (Zhai, 2012) involved in methanogens metabolismsbecause of a sudden change in external environment, which tem-porarily deteriorated their MPA and reduced methane production.As the incubation progressed (i.e., 26–157 days), the methanogensand precursor fermentative populations became more establishedand the production of methane therefore increased (Fig. 1)(Beneragama et al., 2013; Stone et al., 2009). Mechanisms ofmicrobes resistance to antibiotics might be ascribed to the

648 X. Lu et al. / Bioresource Technology 169 (2014) 644–651

development of resistance in existing bacteria and/or an increasein abundance of resistance genes, etc. as hypothesized in the liter-ature (Amin et al., 2006; Rôças and Siqueira-Jr, 2012). For example,Angenent et al. (2008) found that the levels of macrolide–lincosa-mide–streptogramin B (MLSB)-resistant bacteria in the biomass ofanaerobic sequencing batch reactor seeded with a inoculum thatpreviously had not been exposed to the antimicrobial tylosin,increased substantially during the first 69 days of operation.Rôças and Siqueira-Jr (2012) detected the presence of 6 antibioticresistance genes in infected dental root canal strains. Morerecently, the group of Resende et al. (2014) reported a persistenceof antibiotic resistant bacteria during anaerobic digestion treat-ment of cattle manure operated at ambient temperatures.Although several antibiotic resistance genes in anaerobic bacteriahave been identified, knowledge regarding CLX resistance genesin the literature is very rare. Nevertheless, the enhanced methaneyield after 157 days of digestion in our study (Fig. 1) could stillbe best explained by the acclimation of inoculum to CLX or thepresence of CLX resistance genes. This is because that daily useof CLX to treat aspiratory path infection, urine pathway infectionand skin tissue infection (Zhai, 2012) will inevitably bring aboutits liberation into environmental receptors (Estrada et al., 2012),e.g., WWTPs, hence benefiting the growth of CLX resistantmicroorganisms in WAS.

3.1.3. CLX removalThe levels of CLX in MPA tests throughout the entire digestion

were detected in order to ascertain its fate and possible removalin the anaerobic digester, and the resulting data is depicted inFig. 2b. It can be found that the CLX was effectively removed after16 days of digestion even the considerable inhibition of MPA wasachieved as noted in Fig. 2a, and the corresponding removal effi-ciencies of CLX in the different assays ranged from approximately65 to 92% by the end of experiments depending on the concentra-tions used. The numerous observations in effective removal of anti-biotics during anaerobic process are available in previouspublications. Stone et al. (2009) achieved a 57% and nearly 100%reduction for chlortetracycline and tylosin in swine manurethrough 216 days. Álvarez et al. (2010) reported 57.8–67.7%removal in pig manure at day 21. Other groups, such as Shi et al.(2011) and Sara et al. (2013) also gave the rough 50% and 70%reduction in tetracycline/sulfamethoxydiazine, and ceftiofur overthe duration of tests, respectively. The transformation and substan-tial removal of CLX obtained in this study might result from tworeasons as follows: (i) CLX, containing one NH2 and one COOHactive groups that provide grafting sites and have a broad antimi-crobial spectrum (Qi et al., 2012), is biodegradable and could bemetabolized by the one or more of active bacterial groups duringbiological process; and (ii) on the other hand, part of CLX remainedentrapped/adsorbed by sludge flocs or solid matters associatedwith their strong affinity, similar to other types of antibiotics(Álvarez et al., 2010; Sara et al., 2013), to cations, particles andorganics (Loke et al., 2003); for instance, Álvarez et al. (2010) notedthat the fraction adsorbed onto the solids for oxytetracycline andchlortetracycline, was around 18–20 and 25–27 mg/L, respectively.Owing to the rapid CLX mineralization or bio-adsorption (e.g., viaEPS), the performance of CLX-bearing anaerobic systems canrecover after a short-term inhibition, which, therefore, wouldallow methane production to resume during the later phase ofexposure experiments (Fig. 1).

3.2. TSS and VSS reduction

Fig. 3 presents the changes of TSS and VSS along with risingcefalexin dose after 157 days of fermentation. The initial TSS andVSS were determined to be 51.9 ± 0.4 mg/L and 48.0 ± 0.3 mg/L

on average, respectively. Regardless of the concentrations of CLX,the marked reduction in both TSS and VSS took place over theduration of digestion. By the termination of tests, TSS and VSSdecreased to 38.9–42.6 mg/L and 34.6–38.4 mg/L, with corre-sponding removal efficiencies ranging from 17.8 to 24.8% and20.4 to 25.9%, respectively. The aggressive and deleterious effectsthat CLX exerted on digestion of sludge, as expected, were not con-siderable. Comparatively, higher TSS and VSS removal occurredwhen 1000 mg/L CLX was dosed with a lower VSS/TSS ratio of0.89, in accordance to the data regarding methane yields observedbefore (Fig. 1). Our results contradict those of a previous report (Shiet al., 2011), where the presence of antibiotics adversely restrictedthe fermentation of total solids (TS) (p < 0.05). This might bemainly attributable to the striking differences in biodegradabilityof antibiotics investigated and doses, as well as the types of inocu-lums used and their resistance capabilities to toxins. Further evi-dences to support the resistance mechanisms of anaerobicmicrobes to CLX are thus required.

3.3. Role of extracellular polymeric substances (EPS)

3.3.1. Variations of different EPS fractions during anaerobic digestionEPS, secreted by the multispecies community of microorgan-

isms, are located in the out layer of sludge flocs, and consist mainlyof PN and PS, as well as in some cases lipids, nucleic acids and otherbiopolymers in minor contents (Frolund et al., 1996). It is com-monly believed that EPS have a crucial influence on sludge dewa-tering and sedimentation (Zhen et al., 2013a). Over the pastdecade, the essential role played by EPS in the toxicity responseof microbial aggregates to chemical toxins has drawn much moreattention in scientific communities. Information about EPS changesduring anaerobic digestion thus is very important for better under-standing of the toxicity response of microbes to antibiotics. Fig. 4compares the overall variations of different EPS fractions includingslime EPS, LB-EPS and TB-EPS before and after 157 days of diges-tion. The addition of CLX contributed to a stepwise rise in EPS aspresented in Fig. 4a, in particular in slime EPS at the beginning ofdigestion. It might be associated with the acute toxicity of CLX,which resulted in the lysis of some planktonic cells in bulk liquoror absorbed on the surface of sludge flocs and, subsequent releaseof biopolymers. Of course, possibly because of the toxic impact, ashort but clear lag-phase in methane production depending onthe doses of CLX added occurred (inset of Fig. 1).

Furthermore, it is generally believed that the sludge with highEPS, which can embed microcolonies through both electrostaticand hydrophobic forces that contribute to the strength and cohe-siveness of a floc (Henriques and Love, 2007), would have a stifftexture and thereby be more difficult to be hydrolyzed duringanaerobic process. The effective rupture of EPS is thereby regardedas the prerequisite for sludge solubilization and succeeding anaer-obic digestion. As expected, an appreciable decrease in EPS, espe-cially in both LB-EPS and TB-EPS was identified irrespective ofthe concentrations of CLX (Fig. 4b), implying that sludge hadundergone normal decomposition and fermentation processes.However, it does not mean that the higher rupture of EPS, the bet-ter the performance of digestion is. EPS, which accumulate on thebacterial cell surface, beyond influencing sludge dewatering canalso serve as a protective layer against external stress, and as car-bon and energy source during starvation. Hence, if EPS matrix aresubjected to severe destruction or are removed markedly duringanaerobic process, more microbial cells trapped within sludge flocsas well as cell membranes would be liberated and become moreexposed to toxins. In this situation, CLX would become more per-meable and may diffuse into the cells faster, thus causing the awfullysis and death of microbes. The hypothesis is strongly supportedby the findings herein (Fig. 4b) that the sludge flocs, at moderate

32

36

40

44

48

52 Before After

TSS

(mg/

L)

Cefalexin (mg/L)

32

36

40

44

48

Cefalexin (mg/L)

VSS

(mg/

L)

Before

After

0.88

0.89

0.90

0.91

0.92

0.93

VSS/TSS

0 50 200 400 600 10002000 0 50200 400 600 10002000

Fig. 3. TSS and VSS removals after 157 days of digestion.

0

1000

2000

3000

4000

5000 (a)

Solu

tion

conc

entra

tion

(mg/

L)

Cefalexin (mg/L)

PS PN

Slime EPS

LB- EPS

TB- EPS

0 50 200 400 600 1000 20000

1000

2000

3000

4000

5000

0 50 200 400 600 1000 2000

(b)

Solu

tion

conc

entra

tion

(mg/

L)

Cefalexin (mg/L)

Fig. 4. Variations of different EPS fractions before (a) and after (b) 157 days of digestion.

X. Lu et al. / Bioresource Technology 169 (2014) 644–651 649

doses of CLX ranging from 600 to 1000 mg/L, had an abundant butnot superfluous level of EPS (i.e., LB-EPS and TB-EPS), which corre-sponded to the increased methane outputs (Fig. 1). This could bedue to the fact that the presence of EPS can adsorb/bind CLX andretard its diffusion into the cells to some extent; on the other hand,anaerobic microbes, after experiencing an adaption process to CLX,can retain viability at the high concentration of CLX and utilize CLXas carbon sources for self-growth and metabolisms while in turnstimulating the excretion of EPS and subsequent fermentation ofWAS. Because of the complementary roles played by EPS andCLX, the performance of WAS anaerobic digestion thus maintainedstable, and methane production was promoted greatly.

Obviously, the results obtained here confirm the idea that thefloc EPS matrix play a vital role in protecting microbial cells fromthe chemical toxins through impeding the access of toxins intocells, as has been hypothesized by Henriques and Love (2007)and Zhen et al. (2013b). Similar protective mechanisms have alsobeen observed in a recent study by Guibauda et al. (2012), whocomparing the resistance of anaerobic granular biofilms to Cd(II)and Pb(II) concluded that EPS bound metal irons predominantlyvia proton exchange, coordination, chelation and precipitation,which provided protection for the bacteria embedded in the bio-films. Hessler et al. (2012) also reported that acting as a barrierto the cell membrane, capsular EPS permitted attachment of tita-nium dioxide (TiO2) nanoparticles to the cell, while delaying cellu-lar damage caused by the production of reactive oxygen species.Thanks to the key functions of EPS demonstrated herein and inthe literature, the microbes involved in sludge flocs survive welleven under environmental shocks, which is the core reason forenhanced methane yield observed here as well.

3.3.2. UV–Vis spectroscopic properties of different EPS fractionsTo further approach to the protective mechanisms of EPS, UV–

Vis spectroscopy was used to characterize EPS derived from the

sludge before and after 157 days of digestion. Typical UV–Vis spec-tra of different EPS fractions is given in Fig. 3S in Supplementaryinformation. UV–Vis absorbance values decreased steadily accord-ing the wavelength, agreeing well with those of Zhen et al. (2013a,2014) who successfully applied the UV–Vis spectroscopic tech-nique for charactering the sludge dewatering and humificationprocess. The spectra usually include two main absorption bonds:the first one located at k of 200–223 nm and the second presentat k between 250 and 270 nm. The first bond belongs to the pres-ence of carboxylic bonds while the second bond might be attribut-able to n ? p⁄ and p ? p⁄ electron transition of C@O and C@Cstructure of substituted benzenes or polyphenols (Chen et al.,2002).

To better understand the potential differences in UV–Vis spec-troscopic properties of sludge EPS before and after anaerobic pro-cess, the absorption of dissolved organic matters at 254 nm, i.e., arough indicator of the density of conjugated aromatic molecules(UV-254), is thus given in detail (Fig. 5). The change of UV-254reflects the possible degradation, synthesis or transformation ofkey organics (e.g., humic and aromatic compounds) involved inslime EPS, LB-EPS and T-EPS. As exhibited in Fig. 5a, UV-254 valuesof different EPS fractions before anaerobic digestion exhibited asimilar trend to EPS (i.e., PN and PS) with increasing CLX doses(Fig. 4a), demonstrating the deleterious influences of CLX on dam-age of partial planktonic cells. Only discrepancy between UV-254values and measured EPS seems that the former increased compar-atively fast with increased CLX. One possible explanation might bethat the absorbance in 254 nm achieved here not only resultedfrom PN and PS, but also from CLX. Furthermore, Fig. 5b showsthe variation of absorbance of sludge flocs’ EPS after anaerobicdigestion. As expected, the absorbance of different EPS fractionsdropped first, followed by a considerable increase as the appliedCLX elevated from 600 to 1000 mg/L. High absorbance observedherein, agreeing well with increased EPS content (Fig. 4b), strongly

0.00.51.01.52.02.53.03.5

0 50 200 400 600 1000 20000 50 200 400 600 1000 2000

UV

-254

(m-1

)

Cefalexin (mg/L)

Slime EPS LB-EPS TB-EPS

(a) (b)

0.00.51.01.52.02.53.03.5

UV

-254

(m-1

)

Cefalexin (mg/L)

Fig. 5. UV–Vis absorbance of the dissolved organic matters in the EPS of sludge before (a) and after (b) 157 days of anaerobic digestion.

650 X. Lu et al. / Bioresource Technology 169 (2014) 644–651

proved the central role of EPS in keeping stable performance ofsludge fermentation and stimulating methane production (seeFig. 4S in Supplementary information). Based on the aforemen-tioned observations, it can be noted that unlike the commonlyaccepted adverse effect, this study demonstrated the beneficialrole of CLX in restoration of EPS matrix, stability of microbial con-sortia, as well as methane production, providing innovativeinsights into its true environmental impacts.

3.4. Volatile fatty acids (VFAs) assessment

Volatile fatty acids (VFAs) are considered as the key parameterinfluencing the performance of anaerobic digestion (Stone et al.,2009). VFAs, including acetic, propionic, butyric, isobutyric, valeric,and isovaleric acids, before and after anaerobic digestion weredetermined. Fig. 6 gives the initial and final VFAs concentrationsin this test. At the beginning of test, the total VFAs level (sum ofacetic, propionic, butyric, isobutyric, valeric, and isovaleric acids)were usually below 750 mg/L independent of the CLX concentra-tion with acetic and propionic acids being the dominant compo-nents, and even in some cases were less than 250 mg/L. After157 days of digestion, the VFAs accumulation occurred, particu-larly when low concentrations of CLX were used. Total VFAsreached up to 3569, 2337, 3457 and 3746 mg/L with CLX concen-trations of 0, 50, 200 and 400 mg/L, respectively. Similar findinghas been documented by Amin et al. (2006) for erythromycin-trea-ted sludge. In addition, a recent study conducted by Beneragamaet al. (2013) also verified that the presence of antibiotics in dairymanure caused the significant accumulation of VFAs compared

0

500

1000

1500

2000

2500

3000

3500

4000

VFA

s (m

g/L)

Cefalexin (mg/L)

Valeric acid Isovaleric acid Butyric acid Isobutyric acid Propionic acid Acetic acid

0 50 200 400 600 1000 2000

Before

After

Fig. 6. Variations of VFAs under different dosages of CLX before and after 157 daysof digestion.

with a no-antibiotic control. This elevated VFAs level suggestedthat antibiotic CLX reduced methane output at low concentrationsmainly via hindering the utilization of VFAs by homoacetogenicbacteria or aceticlastic methanogens instead of the conversion ofamino acids, sugars and fatty acids into VFAs through fermentativeor acid-producing bacteria, as already observed by Stone et al.(2009). Presumably because of the significant inhibition of metha-nogenesis process induced by antibiotics, methane production thusdropped. Comparatively, the VFAs accumulation seemed to beeffectively alleviated, as CLX was further increased to 600 mg/Land above. Efficient conversion of VFAs, in particular acetic acids,by aceticlastic methanogens such as Methanosaeta spp. into meth-ane and carbon dioxide was observed, which corresponded to thepH variation (Fig. 5S in Supplementary information) and increasedmethane yield. Whereas even though rapid use of acetic acids tookplace throughout the duration of experiment, the degradation ofother VFAs, especially propionic acids, was not actively progress-ing. Propionic acids consistently kept a high level by the end of testregardless of applied CLX dose, possibly owing to high Gibbs freeenergy value (DG00) of +76.1 kJ/mol (Thauer et al., 1977) as wellas the relatively low activity of propionate degraders such asGram-negative bacteria (e.g., Syntrophobacter species) as suggestedby Amin et al. (2006) and Stone et al. (2009). Based on the currentresults obtained here, we still cannot predict how CLX shiftedmicrobial populations or developed resistance of microbes to thepresence of CLX during long-term experimental conditions. Hence,further investigation in this regard is suggested. Nevertheless, theprocess stability of the digesters in this situation has been main-tained without obvious upset or failure except the lag phaseobserved before, providing new information about the true rolesof CLX on methane generation from sludge.

4. Conclusion

Cefalexin (CLX) inhibited methane production during the initial25 days through inhibiting methanogenesis while the negativeeffect attenuated subsequently and methane production recovered.The highest methane yield reached 450 mL at 1000 mg-CLX/L after157 days of digestion, 63.8% higher than CLX-free control. Furtherexamination via UV–Vis spectra verified the elevated slime EPS,LB-EPS and TB-EPS in the presence of CLX. Stimulated excretionof EPS by CLX served as microbial protecting layers, creating asuitable environment for microbes’ growth and methanogenesisprocess. These observations provide better understanding of theoverall importance of EPS in CLX microbial toxicity.

Acknowledgements

The authors wish to thank the China scholarship council (CSC)for the partial support of this study.

X. Lu et al. / Bioresource Technology 169 (2014) 644–651 651

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.biortech.2014.07.056.

References

Álvarez, J.A., Otero, L., Lema, J.M., Omil, F., 2010. The effect and fate of antibioticsduring the anaerobic digestion of pig manure. Bioresour. Technol. 101, 8581–8586.

Amin, M.M., Zilles, J.L., Greiner, J., Charbonneau, S., Raskin, L., Morgenroth, E., 2006.Influence of the antibiotic erythromycin on anaerobic treatment of apharmaceutical wastewater. Environ. Sci. Technol. 40, 3971–3977.

Angenent, L.T., Mau, M., George, U., Zahn, J.A., Raskin, L., 2008. Effect of the presenceof the antimicrobial tylosin in swine waste on anaerobic treatment. Water Res.42, 2377–2384.

APHA, 1998. Standard Methods for the Examination of Water and Wastewater,20th ed. A.P.H. Association, Washington, DC, USA.

Arikan, O.A., Sikora, L.J., Mulbry, W., Khan, S.U., Rice, C., Foster, G.D., 2006. The fateand effect of oxytetracycline during the anaerobic digestion of manure fromtherapeutically treated calves. Process Biochem. 41, 1637–1643.

Beneragama, N., Lateef, S.A., Iwasaki, M., Yamashiro, T., Umetsu, K., 2013. Thecombined effect of cefazolin and oxytetracycline on biogas production fromthermophilic anaerobic digestion of dairy manure. Bioresour. Technol. 133, 23–30.

Chen, J., Gu, B., LeBoeuf, E.J., Pan, H., Dai, S., 2002. Spectroscopic characterization ofthe structural and functional properties of natural organic matter fractions.Chemosphere 48, 59–68.

Dreher, T.M., Mott, H.V., Lupo, C.D., Oswald, A.S., Clay, S.A., Stone, J.J., 2012. Effects ofchlortetracycline amended feed on anaerobic sequencing batch reactorperformance of swine manure digestion. Bioresour. Technol. 125, 65–74.

Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1956. Colorimetricmethod for determination of sugars and related substances. Anal. Chem. 28 (3),350–356.

Estrada, A.L., Li, Y.Y., Wang, A., 2012. Biodegradability enhancement of wastewatercontaining cefalexin by means of the electro-Fenton oxidation process. J.Hazard. Mater. 227–228, 41–48.

Frolund, B., Palmgren, R., Keiding, K., Nielsen, P.H., 1996. Extraction of extracellularpolymers from activated sludge using a cation exchange resin. Water Res. 30,1749–1758.

Guibauda, G., Bhatia, D., dAbzac, P., Bourven, I., Bordas, F., Hullebusch, E.D.v., Lensc,P.N.L., 2012. Cd(II) and Pb(II) sorption by extracellular polymeric substances(EPS) extracted from anaerobic granular biofilms: evidence of a pH sorption-edge. J. Taiwan Inst. Chem. E 43 (3), 444–449.

Guo, W., Wang, H., Shi, Y., Zhang, G., 2010. Sonochemical degradation of theantibiotic cephalexin in aqueous solution. Water SA 36, 651–654.

Guo, J., Ostermann, A., Siemens, J., Dong, R., Clemens, J., 2012. Short term effects ofcopper, sulfadiazine and difloxacin on the anaerobic digestion of pig manure atlow organic loading rates. Waste Manage. 32, 131–136.

Henriques, I.D.S., Love, N.G., 2007. The role of extracellular polymeric substances inthe toxicity response of activated sludge bacteria to chemical toxins. Water Res.41, 4177–4185.

Hessler, C.M., Wu, M.Y., Xue, Z., Choi, H., Seo, Y., 2012. The influence of capsularextracellular polymeric substances on the interaction between TiO2

nanoparticles and planktonic bacteria. Water Res. 46, 4687–4696.

Ince, B., Coban, H., Turker, G., Ertekin, E., Ince, O., 2013. Effect of oxytetracycline onbiogas production and active microbial populations during batch anaerobicdigestion of cow manure. Bioprocess Biosyst. Eng. 36, 541–546.

Loke, M.L., Jespersen, S., Vreeken, R., Halling-Sørensen, B., Tjørnelund, J., 2003.Determination of oxytetracycline and its degradation products by high-performance liquid chromatography–tandem mass spectrometry in manure-containing anaerobic test systems. J. Chromatogr. B 783, 11–23.

Masse, D.I., Lu, D., Masse, L., Droste, R.L., 2000. Effect of antibiotics on psychrophilicanaerobic digestion of swine manure slurry in sequencing batch reactors.Bioresour. Technol. 75, 205–211.

Mitchell, S.M., Ullman, J.L., Teel, A.L., Watts, R.J., Frear, C., 2013. The effects of theantibiotics ampicillin, florfenicol, sulfamethazine, and tylosin on biogasproduction and their degradation efficiency during anaerobic digestion.Bioresour. Technol. 149, 244–252.

Mohring, S.A.I., Strzysch, I., Fernandes, M.R., Kiffmeyer, T.K., Tuerk, J., Hamscher, G.,Kiffmeyer, T.K., Tuerk, J., Hamscher, G., 2009. Degradation and elimination ofarious sulfonamides during anaerobic fermentation: a promising step on theway to sustainable pharmacy? Environ. Sci. Technol. 43, 2569–2574.

Ponsa, S., Ferrer, I., Vazquez, F., Font, X., 2008. Optimization of the hydrolytic-acidogenic anaerobic digestion stage (55 �C) of sewage sludge: influence of pHand solid content. Water Res. 42, 3972–3980.

Qi, X., Gunawan, P., Xu, R., Chang, M.W., 2012. Cefalexin-immobilized multi-walledcarbon nanotubes show strong antimicrobial and anti-adhesion properties.Chem. Eng. Sci. 84, 552–556.

Resende, J.A., Silva, V.L., Oliveira, T.L.R.d., Fortunato, S.d.O., Carneiro, J.d.C., Otenio,M.H., Diniz, C.G., 2014. Prevalence and persistence of potentially pathogenicand antibiotic resistant bacteria during anaerobic digestion treatment of cattlemanure. Bioresour. Technol. 153, 284–291.

Rôças, I.N., Siqueira-Jr, J.F., 2012. Antibiotic resistance genes in anaerobic bacteriaisolated from primary dental root canal infections. Anaerobe 18, 576–580.

Sara, P., Giuliana, D.I., Michele, P., Maurizio, C., Luca, C., Fabrizio, A., 2013. Effect ofveterinary antibiotics on biogas and bio-methane production. Int. Biodeterior.Biodegrad. 85, 205–209.

Saravanane, R., Sundararaman, S., 2009. Effect of loading rate and HRT on theremoval of cephalosporin and their intermediates during the operation ofmembrane bioreactor treating pharmaceutical wastewater. Environ. Technol.30, 1017–1022.

Shi, J.C., Liao, X.D., Wu, Y.B., Liang, J.B., 2011. Effect of antibiotics on methane arisingfrom anaerobic digestion of pig manure. Anim. Feed Sci. Technol. 166–167,457–463.

Stone, J.J., Clay, S.A., Zhu, Z., Wong, K.L., Porath, L.R., Spellman, G.M., 2009. Effect ofantimicrobial compounds tylosin and chlortetracycline during batch anaerobicswine manure digestion. Water Res. 43, 4740–4750.

Thauer, R.K., Jungermann, K., Decker, K., 1977. Energy conservation in chemotrophicanaerobic bacteria. Bacteriol. Rev. 41 (1), 100–180.

Thiele-Bruhn, S., Beck, I.C., 2005. Effects of sulfonamide and tetracycline antibioticson soil microbial activity and microbial biomass. Chemosphere 59, 457–465.

Zhai, Q.Z., 2012. Inclusion of cefalexin in SBA-15 mesoporous material and releaseproperty. Mater. Sci. Eng., C 32, 2411–2417.

Zhen, G., Lu, X., Li, Y.-Y., Zhao, Y., 2013a. Innovative combination of electrolysis andFe(II)-activated persulfate oxidation for improving the dewaterability of wasteactivated sludge. Bioresour. Technol. 136, 654–663.

Zhen, G., Lu, X., Niu, J., Su, L., Chai, X., Zhao, Y., Li, Y.Y., Song, Y., Niu, D., 2013b.Inhibitory effects of a shock load of Fe(II)-mediated persulfate oxidation onwaste activated sludge anaerobic digestion. Chem. Eng. J. 233, 274–281.

Zhen, G., Lu, X., Li, Y.Y., Zhao, Y., 2014. Combined electrical-alkali pretreatment toincrease the anaerobic hydrolysis rate of waste activated sludge duringanaerobic digestion. Appl. Energy 128, 93–102.