lable at ScienceDirect

Journal of Cleaner Production 258 (2020) 120993

Contents lists avai

Journal of Cleaner Production

journal homepage: www.elsevier .com/locate/ jc lepro

Insight into the microbial and genetic responses of anammox granulesto spiramycin: Comparison between two different dosing strategies

Jing-Wu , Nian-Si Fan *, Ye-Ying Yu , Yi-Jun He , Yi-Heng Zhao , Quan Zhang ,Bao-Cheng Huang , Ren-Cun Jin **

Laboratory of Water Pollution Remediation, School of Life and Environmental Sciences, Hangzhou Normal University, Hangzhou, 311121, China

a r t i c l e i n f o

Article history:Received 16 December 2019Received in revised form2 March 2020Accepted 6 March 2020Available online 9 March 2020

Handling Editor: Prof. Bing-Jie Ni

Keywords:AnammoxSpiramycinExposure strategyMicrobial communityAntibiotic resistance genes

* Corresponding author.** Corresponding author.

E-mail addresses: fns1907@163.com (N.-S. Fan), jr

https://doi.org/10.1016/j.jclepro.2020.1209930959-6526/© 2020 Elsevier Ltd. All rights reserved.

a b s t r a c t

The potential application of anaerobic ammonium oxidation (anammox) to treat antibiotic-containingwastewaters has attracted public attentions. The responses of anammox to a typical antibiotic (spi-ramycin, SPM) with different exposure strategies (elevating-concentration strategy in R1 and repeated-exposure strategy in R2) were evaluated on the aspects of general performance, microbial communityand genes. Results showed that anammox bacteria could adapt to low concentrations (�1 mg L�1) ofSPM, while high-concentration (�3 mg L�1) SPM showed inhibitory influences. The nitrogen removalefficiency (NRE) quickly decreased from 90.21% to 76.29% when R1 was exposed to 5 mg L�1 SPM, andthen returned to a higher level within a short period. On the contrary, R2 could not recover with theinhibition of 5 mg L�1 SPM, and the abundance of Planctomycetes declined from 18.96% to 15.41% on Day105. Combined with the changes in key genes and functional bacteria, R1 had a relatively stable per-formance and higher resistance to SPM, indicating the elevating-concentration strategy might be moreeffective to gain the tolerance of anammox process to high-concentration antibiotics. The present studyprovides a guide for the stable operation of anammox process treating antibiotic-containingwastewaters.

© 2020 Elsevier Ltd. All rights reserved.

1. Introduction

Antibiotics are irreplaceable for human health and the survivalof animals and plants. However, most antibiotics are not completelydegraded or absorbed by organisms, and large amounts of antibi-otics could be detected in natural environments (Sarmah et al.,2006). Extensive antibiotics induce the presence and trans-formation of antibiotic resistance bacteria (ARB) and antibioticresistance genes (ARGs), which have become emerging pollutants(Rizzo et al., 2013). Where antibiotics and ARGs content are high iswastewater treatment plant (WWTP). Li et al. (2013) detectedsulfonamides (22.7 mg kg�1 dw) and macrolides (85.1 mg kg�1 dw)in sludge collected from sewage treatment plants. Mao et al. (2015)detected 30 pairs of sulfonamide, tetracycline, macrolide or qui-nolone ARG in two activated sludge WWTPs in China. Su et al.(2017) collected 161 sewage samples from 32 WWTPs in China. In

czju@aliyun.com (R.-C. Jin).

total, 381 different ARGs were detected with the average concen-tration of 1.79 � 1012 copies L�1. Under the stress of various anti-biotics, the biological activity of functional bacteria is suppressed,thus influencing the stable operation of wastewater treatmentsystems. SPM is one of the common antibiotics, mainly used to treatoropharynx, respiratory and genitourinary infections (Zhu et al.,2014). It can also be used to increase the growth rate of livestockand poultry (Mayor et al., 2017). The concentrations of SPM in soiland wastewater were as high as 289.4 mg kg�1 and 41 mg L�1 (Liuet al., 2014), respectively. Although SPM residues arise public con-cerns, the effective and economic measures to eliminate SPM arestill scarce.

As an efficient and sustainable nitrogen removal pathway,anaerobic ammonium oxidation (anammox) has been extensivelyinvestigated. However, field application and operation are limited byvarious factors, such as the low growth rate and cell yield ofanammox bacteria (AnAOB) and its susceptibility to the surroundingenvironment. Previous studies have verified the inhibitory effects ofheavy metals, Mn, Cu, Pb, Zn, Cr, Cd (Zhang et al., 2015b; Xu et al.,2017; Manish et al., 2019), organic/inorganic matters chemical ox-ygen demand (COD), phosphate, nitrite (Molinuevo et al., 2009;

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 1209932

Zhang et al., 2016; Qi et al., 2018), nanoparticles, nanoscale zero-valent iron, Cr NPs, Cu NPs, NiO NPs (Zhang et al., 2017a; Xu et al.,2019; Marcos et al., 2019), etc. As one of widespread contami-nants, the influences of antibiotics on anammox were also investi-gated. Zhang et al. (2019c) reported that AnAOB could tolerateerythromycin (ERY) below 1mg L�1, but 10 mg L�1 ERY significantlyinhibited the bioactivity of AnAOB. Shi et al. (2017) found that2.0 mg L�1 OTC significantly inhibited the AnAOB. Fan et al. (2019)reported that the inhibitory thresholds of tetracycline on anam-mox were 1 mg L�1. Norfloxacin in trace concentration (1 mg L�1)significantly suppressed anammox system (Zhang et al., 2019d).Inhibition of anammox bacteria was directly proportional to chlor-amphenicol (CAP) concentrations (Phanwilai et al., 2020). Thecombined effects of various antibiotics were also investigated. Forexample, Zhang et al. (2019a) reported that anammox performancewas inhibited when oxytetracycline and sulfamethoxazole concen-tration reached 1 mg L�1. Nevertheless, the effect of SPM on generalperformance and microbial community on anammox granularsludge have never been reported. Moreover, different exposurescenarios could also cause the difference.

Based on the above, different concentrations (0e5 mg L�1) ofSPM were added with two dosing strategies (elevating concentra-tion strategy and repeated exposure strategy) to evaluate its long-term influence on anaerobic ammonium bioreactors in this study.The focuses of this study are: (i) determining the effects of SPM inthe different dosing strategies on the performance of anammoxprocess and activity of AnAOB; (ii) revealing the effects of SPMstress on the microbial community structures of anammox sys-tems; and (iii) illuminating the changes in ARGs and key functionalgenes, and the correlations among these genes and dominantbacteria. The outcomes of this work not only comprehensivelyinterpret the effects of SPM on anammox, also provide a theoreticalguide for the practical application of anaerobic ammonium oxida-tion bioreactors.

2. Materials and methods

2.1. Seed sludge

Collected from a high-loading anammox reactor, the inoculatedsludge had been stably cultivated for more than two years in thelaboratory. The nitrogen loading rate (NLR) and nitrogen removalrate (NRR) were 5.68 ± 0.4 kg N m�3 d�1 and 5.38 ± 0.4 kg N m�3

d�1, respectively. Candidatus Kuenenia was the dominant AnAOB inthis system. Based on the previous studies (Jin et al., 2013), reactorswere fed with the synthetic wastewater, consisting of substrates,inorganic solution, and trace elements (Table S1). Equimolaramounts of ammonium and nitrite were supplied as substrates inthe forms of (NH4)2SO4 and NaNO2, respectively.

2.2. Operation strategy

Seed sludge of 700 mL was inoculated into three upflowanaerobic sludge blanket (UASB) reactors (Fig. S1) with 1.5 Leffective volume (Fan et al., 2019), including a control reactor (R0)and two experimental reactors (R1 and R2). In natural environ-ments, antibiotics commonly accumulated in two different ways.Antibiotics concentrations inwastewaters gradually increasewith acontinuous exposure to antibiotics, such as hospital wastewater.Another situation is periodically receiving the wastewater con-taining high-concentration antibiotics, such as pharmaceuticalfactory producing various types of antibiotics. Therefore, accom-panied with the synthetic wastewater, R1 was fed with gradientconcentrations of SPM (elevating concentration strategy), and R2was repeatedly exposed to the same concentration of SPM

(repeated exposure strategy) (Table 1). Considering the antibioticdegradation, SPM was daily added to maintain a relatively stableconcentration in systems. For the initial sludge, the mixed liquorsuspended solids (MLSS) concentration was 4.54 g L�1 and themixed liquor volatile suspended solids (MLVSS) concentration was2.79 g L�1. The hydraulic retention time (HRT) is 2.25 h and theinfluent pHwas controlled at 7.2e8.4 with phosphate buffer (Zhanget al., 2019b). To avoid the influences of light and temperaturevariation, these continuous up flow systems were operated in adark room maintained at 35 �C.

2.3. Chemical analysis and anammox activity measurements

The concentrations of NH4þ, NO2

� and NO3� were determined

using the method of APHA. A digital pH meter was used formeasuring the pH (INESA, Shanghai, China). Extraction of extra-cellular polymeric substances (EPS) in samples were using pyrolysis(Zhang et al., 2015a). The fluorenone-H2SO4 method and the Folie-phenol reagent method was used to measure the polysaccharidecontent and protein content of EPS, they are based on glucose andbovine serum albumin (BSA), respectively (Zhang et al., 2017b).

According to Xu et al. (2017), the specific anammox activity(SAA) was measured with serum bottles, three 120 mL of serumbottles containing 100 mg L�1 NH4

þ-N and NO2�-N were prepared in

advance. 5 mL sludge samples were taken from reactors andwashed three times to remove residual nitrogen with a phosphatebuffer solution at the end of each phase. The serum bottles wereaerated by argon gas for 10 min and then maintained anaerobicconditions by sealing. All serum bottles were incubated on a shakerat 35 �C, 180 rpm. Samples were collected every 30 min to monitorthe rate of substrates consumption.

2.4. Gene quantification and isolation of ARB

At the end of each stage, the mixed sludge samples werecollected and stored at �20 �C for short term. Extraction of totalDNA through Power Soil DNA Kit (MoBio Laboratories, Carlsbad,CA). The Nano-drop 2000 spectrophotometer (Thermo Scientific,USA) was used to measure DNA concentration and quality. Thesubsequent detection and quantification of 5 ARGs (ermB, ermF,ermQ, ereA and mphA) and 6 functional genes (hdh, hzsA, nirS, nosZand norB) were learn from the method of Fan et al., (2019). Theprimers used in this study and corresponding amplification con-ditions were shown in Table S2.

The ARB was isolated based on the previous study (Ullah et al.,2019). Briefly, mixed with 30% glycerol, 2 mL of sludge samples wasand stored at �80 �C. The anammox granular sludge was gentlygrinded by sterile forceps. After centrifuged at 5000 rpm for 1 min,100 mL of supernatant was collected as bacterial suspension. Thebacterial solutionwasdilutedwith10-folddilution.OnaLBagarplatecontaining 50mg L�1 SPM, bacterial suspension in each dilutionwasspread, and then placed at 35 �C for 12 h. The 16S rRNA gene wereamplifiedwithuniversal primers according to standardPCRprotocol.Purified PCR products were sequenced and comparedwith GenBankDatabase through Basic Local Alignment Search Tool (BLAST).

2.5. Microbial community analysis

To amplify the V3eV4 region of 16S rRNA gene, the bacterialprimers 338F (50-ACTCCTACGGGAGGCAGCAG-30) and 806R (50-GGACTACHVGGGTWTCTAAT-30) were used, and 2% agarose gelelectrophoresis were used to check the PCR products. The IlluminaMi-Seq platform (Majorbio, Shanghai, China) was used for high-throughput sequencing. Deposit the sequence data in the NCBISequence Read Archive (SRA) database under BioProject

Table 1Operation strategy of SPM used in this study.

Phase R0 (Control) R1 (Elevating-concentration strategy) R2 (Repeated-exposure strategy)

Concentration (mg L�1) Duration (d) Concentration (mg L�1) Duration (d) Concentration (mg L�1)

Stabilization 0 1e40 0 1e40 0Phase 1 0 41e105 0.5 41e105 5Phase 2 0 106e125 0 106e126 0Phase 3 0 126e150 0.5 126e150 0Phase 4 0 151e170 1 151e170 5Phase 5 0 171e195 0 171e195 5Phase 6 0 196e215 3 196e215 5Phase 7 0 215e225 0 215e225 5Phase 8 0 225e245 5 225e245 5

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 120993 3

PRJNA576972, and the registration number is SRR10259579-600.

2.6. Statistical analysis

To test the significance of the results, an analysis of variance(ANOVA) was performed during the whole experiment. The SAA,EPS, and qPCR tested were performed in triplicate. The relation-ships between dominant bacteria, functional genes and ARGs wereassessed using SPSS 21.0 (SPSS Inc., USA). The statistical significancewas marked with * (p < 0.05) and the highly statistical significancewas marked with ** (p < 0.01). Correlation analysis results arepresented through heat maps.

3. Results and discussion

3.1. Nitrogen removal performance

General performance of three reactors were illustrated in Fig. 1.

Fig. 1. General performances of experimental reactors during different phases. (A) control renitrogen load rate; NRR: nitrogen removal rate; NRE: nitrogen removal efficiency. RS: NO2

�-

Before adding antibiotics, three UASB bioreactors were stabilizedfor 40 days. NRE maintained at 92.3 ± 0.8% in all reactors, and NLRand NRR were 6.1 ± 0.5 kg N m�3 d�1 and 5.7 ± 0.9 kg N m�3 d�1,respectively. During the whole experimental period, NRE of R0 wascomparatively stable with the average value of 91.6 ± 4.8%. Theoccasional slight fluctuations might be due to the temperaturevariations caused by the malfunction of equipment (Fig. 1A). Lottiet al. (2015) proposed that temperature was an key factoraffecting anammox activity. For example, a dramatical increase inthe effluent concentration of 131.9 mg L�1 of NH4

þ-N on Day 48 wasmainly attributed to the low temperature (28 ± 2 �C) in the pre-vious days. The theoretical values of stoichiometric ratio of RS(NO2

�-N conversion to NH4þ-N depletion) is 1.32 and RP (NO3

�-Nconversion to NH4

þ-N depletion) is 0.26. RS and RP of R0 were alsorelatively stable throughout the experimental period. The RS/RP was(1.18 ± 0.21):(0.21 ± 0.05), close to the values of 1.18:0.24 (Yang andJin, 2013).

During phase 1, the NRE of R1 maintained at 93.7 ± 2.6%,

actor; (B) dosing with elevating strategy; (C) dosing with repeated strategy. NLR: totalN conversion/NH4

þ-N depletion; RP: NO3�-N production/NH4

þ-N depletion.

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 1209934

indicating that 0.5 mg L�1 of SPM does not significantly affect theanammox performance (Fig. 1B). When the SPM raised to 1 mg L�1

from Day 151, NRE decreased to 58.6%, and the NH4þ-N and NO2

�-Nconcentrations in effluent increased to 105.5 mg L�1 and104.5 mg L�1 after 6 days, respectively. The stoichiometric ratios ofR1 were 0.63 (RS) and 0.23 (RP), which also deviated from thetheoretical values. Afterwards, NRE maintained at the low level,followed by a rebound to 93.5% after a recovery period of 15 dayswithout SPM. Under the stress of 3 mg L�1 SPM, NRE dropped againto 59.6 ± 3.5%. Correspondingly, the concentrations of NH4

þ-N ineffluent was 129.0 mg L�1 and the concentrations of NO2

�-N ineffluent was 130.4 mg L�1. It took 10 days to recover, which wasshorter than that of the previous phase. In the last phase (5 mg L�1

SPM), the NRE temporarily decreased to 76.3% on Day 232, and thenquickly recovered to more than 90%. In general, the low concen-trations (0.5 mg L�1) of SPM did not significantly affect nitrogenremoval performance of systems. However, high SPM concentra-tions (5 mg L�1) exhibited inhibitory effects, which requireddifferent periods of time to recover in both reactors. As the SPMconcentration increased, the bacterial resistance also increased;thus, its inhibition intensity on nitrogen removal performancereduced, and corresponding recovery period became shorter.

In R2, the first addition of SPM caused a decrease of NRE to83.1 ± 2.9% after 25 days, and it reached to the lowest value of 14.8%on Day 94 (Fig. 1C). RS and RP reached to 1.68 and 1.41 respectively.Simultaneously, the effluent concentrations of NH4

þ-N was276.3 mg L�1 and NO2

�-N was 244.3 mg L�1. Such deterioration ofnitrogen removal performance lasted for 72 days, implying that itcould not restore automatically this time. Considering the wholeexperimental schedule, reducing influent substrate concentrationstrategy was adopted to revive the performance. After 30 days,effluent substrate concentrations dropped back to 29.5 mg L�1

(NH4þ-N) and 13.5 mg L�1 (NO2

�-N), and NRE recovered to 92.7% onDay 150, slightly higher than R1 (phase 7). The difference in theresponses of anammox on 5 mg L�1 SPM was mainly attributed tothe pervious long-standing adaptation (addition of 5 mg L�1 SPMfrom Day 40 to Day 105). In phase 4, NRE showed a slight decreasefrom 91.4% to 58.7% within four days, and then quickly recovered toa higher level and remain stable (92.9 ± 3.4%) to the end. Thesecond exposure to high-concentration SPM did not trigger anysignificant fluctuations in performance, reconfirming that thesludge in R2 already had the resistance to 5 mg L�1 SPM.

Based on the nitrogen removal performance,�1mg L�1 could beregarded as a long-term inhibitory concentration of SPM foranammox. Zhang et al. (2019b) verified that AnAOB can completelyresist erythromycin below 1 mg L�1, indicating that anammoxprocess might tolerate the concentration of macrolide antibioticsbelow 1 mg L�1. Although SPM in higher concentrations had sig-nificant inhibitory effects on the nitrogen removal performance ofanammox, it could recover to the initial level through an adaptationperiod. The restorability of SPM inhibition on anammox impliedthat ARGs probably played an important role during the recoveryprocess (Rodríguez-Rojas et al., 2013). Similar phenomenon wasobserved in anammox systems under oxytetracycline (Shi et al.,2017) and erythromycin (Zhang et al., 2019c) inhibition, indi-cating that anammox bacteria can keep stable activity at relativelylow concentrations (1 mg L�1) of oxytetracycline and erythromycin.R2 in this experiment can endure 5 mg L�1 SPM, probably becauseanammox process might be more tolerant to the SPM inhibitionthan oxytetracycline and erythromycin.

3.2. Changes in anammox activity and EPS

In this study, SAA and EPS were important indicators for eval-uating the effects of SPM on AnAOB. As the first barrier against

microbial stress, EPS has a certain protective effect on microor-ganisms. The internal environment of anammox particles is rela-tively stable because it has a compact spherical structure and a highcontent of EPS. When external toxic compounds invade, such asantibiotics, EPS provides a high resistance. (Zhang et al., 2015a).Polysaccharides (PS) and proteins (PN) are the hydrophiliccomposition and a hydrophobic composition within EPS, respec-tively. A reduction in PN/PS led to the poor hydrophobicity, andmight reduce the flocculation capacity of the sludge. According toFig. 2A, the flocculation ability of both experimental reactorsdecreased and were lower than the control reactor due to the stressof SPM. Meng et al. (2019) also proposed that the low PN/PS ratiocould lead to poor sludge settling capacity. From Day 40 to Day 65,the PS in R1 increased by 42.8%, and the PN in R2 increased by 13.5%.The initial exposure of SPM stimulated the secretion of EPS in thesludge (Jia et al., 2017). During Days 65e105, the total EPS of R1 andR2 dropped by 16.3% and 14.2%, respectively, probably because thecontinuous antibiotic pressures destroyed the bacterial cell struc-ture and caused cell lysis (Meng et al., 2019). SPM was added againafter 25 days of recovery. For self-protection, the EPS in R1increased by 11.7% and that of R2 increased by 34.3%. Simulta-neously, the bioactivity and NRE were also stable at a high level(Figs. 2A and 1).

The initial SAA of the seed sludge was 92.7 ± 7.7 mg N g�1 VSSd�1. When 0.5 mg L�1 of SPMwas firstly added in R1, SAA remaineda comparable value to that in R0 on Day 65. Then, it decreased to85.1 mg N g�1 VSS d�1 after 40 days, which was 63.9% of the controlreactor. Although the NRE was not affected during this period,0.5 mg L�1 of SPM had a certain negative effect on anammox ac-tivity. After a period of recovery, adding 1 mg L�1 SPM caused theactivity of R1 (138.8 ± 22.2 mg N g�1 VSS d�1) dropped to380.6 ± 36.4 mg N g�1 VSS d�1 (36.5% of the control), which was inline with the anammox performance. Similarly, the inhibition onSAA could also recover to a considerable level to the control inphase 5. Increasing SPM concentration to 3 mg L�1, SAA of R1declined to 84.7 mg N g�1 VSS d�1, which was about 49.6% of thecontrol. Further adding 5 mg L�1 SPM, SAA of R1 was even slightlyhigher than R0 on Day 245, indicating that SAA had recoveredquickly in this phase. After the first addition of 5 mg L�1 SPM to R2,SAA of R2 decreased to 47.19 ± 14.1 mg N g�1 VSS d�1 on Day 105,which was 24.7% of the control (Fig. 2B). After a long period ofrecovery, SAA in R2 remained stable at a high level of297.3 ± 28.2 mg N g�1 VSS d�1, which was 68.6% higher than thecontrol (176.3 ± 9.5 mg N g�1 VSS d�1). The changes in SAA wasgenerally consistent with that of nitrogen removal performance.Comparatively, 5 mg L�1 of SPM exhibited a more significant in-hibition on anammox in R2, indicating that using elevating con-centration strategy was more tolerant to pressure of the high-concentration SPM due to a previous long-term acclimation.

3.3. Changes in ARGs and functional genes

There were 5 ARGs detected and quantified by qPCR in thisstudy, including methylase genes (ermB, ermF, ermQ), esterase gene(ereA) and phosphatase gene (mphA). Combined with the intI1 and6 functional genes (hdh, hzsA, nirS, nosZ and norB), their quantifi-cation results were illustrated in Fig. 3 and Fig. 4. Most ARGs in R0generally maintained comparatively stable and relatively lowerlevels.

After R1 was addedwith 0.5 mg L�1 SPM, the abundance of ARGs(ermB, ermF and ermQ) increased significantly compared to those inthe control reactor. The abundance of ermF significantly increasedfrom 3.07� 103 copies ng�1 DNA (Day 40) to 2.44� 104 copies ng�1

DNA (Day 105), and finally reached to 5.37 � 104 copies ng�1 DNA.As a representative of horizontal mobile elements, intI1 usually

Fig. 2. Changes in the EPS (A) and SAA (B) of anammox granules in reactors during the experimental period. (a): R0; (b): R1; (c): R2.

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 120993 5

located on plasmids and transposons to promote horizontaltransfer of resistance genes, and is associated with multidrugresistance genes (Gillings et al., 2014). When R1 was added0.5 mg L�1 SPM, the abundance of intI1 increased from 2.80 � 103

copies ng�1 DNA to 2.94 � 104 copies ng�1 DNA (Day 105), which isconsistent with the increasing trend of ereA, mphA and ermF. Adramatical reduction in the abundances of all ARGs occurred fromDay 125, which mainly due to the interruption about two weeks ofSpring Festival Holiday. From Day 170, their abundances continuedto increase or maintained at higher levels, with increasing SPMconcentrations. For example, the abundance of ermF raised from9.34� 102 copies ng�1 DNA (Day 170) to 6.5� 104 copies ng�1 DNA(Day 215), and finally it was much higher than the control reactor.

Changes in the abundances of ermB, ermF, and ermQ in R2exhibited similar trends. As a representative, the abundance ofermF significantly increased from initial 7.24� 103 copies ng�1 DNAto 9.90� 104 copies ng�1 DNA (Day 105). Afterwards, its abundancedramatically decreased, which might be related to the reactorsstopped for a while due to holiday. This phenomenon occurred inthe fluctuations of all the detected ARGs. With the second additionof SPM, the abundance of ermF rebounded from Day 195 andmaintained at a relatively high level to the end. It suggested thatbacteria in the reactor became tolerant to the continuous dosage of5 mg L�1 SPM. Variations in the abundance of ereA were similar. Inaddition, the abundance of intI1 also increased to 1.97 � 104 copiesng�1 DNA on Day 195. The similarity between the trends of intI1and ARGs implied that intI1 involved in the transfer of resistancegenes.

For the anammox-related functional genes, the abundances ofhzsA, hdh, and nirS had similar trends (Fig. 4). The abundances of allfunctional genes in R0 generally increased, and remained higher

values than those in the experimental reactors. During the initialaddition of 0.5 mg L�1 SPM in R1 (Days 40e105), the abundances ofhzsA and hdh firstly decreased and then increased 4.20� 105 copiesng�1 DNA and 6.49 � 108 copies ng�1 DNA, respectively. Therebound of their abundances might be related to the increase in theresistance of AnAOB, which was also verified by the high ARGsabundances on Day 105. Meanwhile, the abundance of nirS in R1rapidly decreased from 2.75 � 108 copies ng�1 DNA to 1.27 � 108

copies ng�1 DNA, implying that the second step of denitrification(NO2

�/NO) was partially inhibited. With higher concentrations ofSPM, the abundances of hzsA and hdh generally decreased, whilethat of nirS slightly increased, but still much lower than its initialvalue. In R2, first addition of 5 mg L�1 SPM reduced the abundancesof hzsA to 4.86 � 104 copies ng�1 DNA and reduced the abundancesof hdh to 4.04 � 107 copies ng�1 DNA at Day 65. Similarly, theabundance of nirS also decreased. These values in R2 were lowerthan those in R1, confirming that the inhibitory effects of SPMincreased with the increasing concentrations. Afterwards, abun-dances of three genes recovered and maintained at a high level,which was consistent with NRE (Fig. 1C). Abundances of AMX, norBand nosZ also exhibited a similar trend. In the end, the abundancesof these functional genes in both R1 and R2were lower than those inR0, indicating that both strategies had inhibitory effect on thefunctional genes and potentially corresponding host bacteria.

3.4. Response of microbial community to the SPM stress

The analysis on the Chao1 and ACE indices (related to richnessrichness) and Shannon index (related to diversity) of microbialcommunity was performed. According to Table S3, the Chao1 andACE indices of R1 were continuously increasing during the first two

Fig. 3. Quantification results of ARGs and intI1.

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 1209936

phases, indicating that the population richness and diversity in R1increased in this period. Afterwards, both decreased when SPMconcentration was higher than 3 mg L�1, suggesting high concen-trations of SPM resulted in some unsuitable microbial death anddisappear. On the other hand, the degradation products of antibi-otics might also influence the microbial community (Halling-Sørensen et al., 2002), which needs further investigation accord-ing to related experiences (Shojaie et al., 2018; Ashouri et al., 2019).Suffering frommore antibiotic pressure, R2 had lower diversity andrichness than R1, after two doses of 5 mg L�1 SPM, both the ACE andChao1 indices decreased significantly (t-test, p < 0.01). Afterexposed to 0.5 and 5 mg L�1 for 65 days, the ACE and Chao1 in tworeactors were 448 and 394 (R1), and 356 and 346 (R2), respectively.It showed that the higher concentration of SPM inhibited the di-versity of anammox system more significantly, which shows nodifference with the results of previous work of Du et al. (2018).Compared with other groups, the diversity trend of R0 was muchweaker throughout the experiment. During the whole period, theShannon index had a similar trend to the changes in the ACE andChao1 indices, which had also been observed in soil microbialcommunity (Zhao et al., 2019). The predominant phyla in threereactors were Chloroflexi, Planctomycetes and Proteobacteria,occupied more than 70% of the total abundance (Fig. 5A). On Day40, The abundance of Planctomycetes in R0was 20.7% and increasedto 54.2% after 25 days.

The abundance of Planctomycetes related to nitrogen removal in

R1 decreased from 20.65% to 14.99% after adding 0.5mg L�1 SPM for25 days, it rebounded to 35.89% in the next 40 days. It reflected thatPlanctomycetes was initially inhibited by the pressure of 0.5 mg L�1

SPM, followed by a recovery after a short period of adaptation. Onthe contrary, Chloroflexi and Proteobacteria has the opposite trend.After adding 0.5 mg L�1 SPM for 20 days, both abundancesincreased by 11.5%, 38.5%, respectively, implying their higherresistance to SPM. At genus level (Fig. 5B), Candidatus Kuenenia, as arepresentative in Planctomycetes, was the dominant functionalbacteria related to nitrogen removal of anammox process (Zhanget al., 2018). Variations in its abundance was consistent with thatof Planctomycetes, and also similar to the changes in the reactorperformance. For example, after adding 3 mg L�1 SPM for 20 days,the abundance of Candidatus Kuenenia in R1 reduced from 30.97%to 23.55%. Correspondingly, the NRE of R1 reduced from 93.81% to59.6%. Generally, Planctomycetes was proportional to the NRE withthe correlation coefficient of 0.64 (p < 0.05).

In R2, the abundance of Planctomycetes continued to decreasewith the dosing of 5 mg L�1 SPM, and decreased to 15.41% at Day105, which is consistent with the NRE of R1 falling to 27.05%(Fig. 1B). As the resistance of R2 to SPMwas gradually increased, theabundance of Planctomycetes in R2 remained high during thesubsequent addition of 5 mg L�1, and the NRE remained more than90%. The abundance of Candidatus Kuenenia in R0 rose from 18.96%to 53.3% in the first 25 days, while that in R1 and R2 dropped to13.38% and 17.37%, respectively. On Day 105, the abundance of

Fig. 4. Quantification results of functional genes.

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 120993 7

Candidatus Kuenenia in R2 dropped to 15.37%, while that in R1increased to 34.62%. This difference was in line with their corre-sponding performances (Fig. 1), that NRE of R1 remained stable andthat of R2 decreased to 25.65%. The abundance of Planctomycetes inR2 finally reached to 38.16%, which was much higher than that of R1(26.91%). Compared with R2, R1 under a constant pressure ofincreasing-concentration SPM maintained a relatively stable oper-ation for a long time and higher antibiotics resistance. Combinedwith the nitrogen removal performance, it can be concluded thatelevating-concentration strategy is more efficient to increase theantibiotic tolerance of rectors.

It noted that after adding SPM for 20 days, the abundance ofLimnobacter in R1 and R2 increased by 38.8% and 85.7%, respectively.Limnobacter is a thiosulfate-oxidizing bacterium. Under SPM stress,anammox cells secreted more EPS, which might provide the pro-liferation of heterotrophic microorganisms like Limnobacter (Springet al., 2001). Other carbon and energy sources for this microor-ganism might be the degradation products of antibiotics and me-tabolites of bacteria in system, which needs further investigationand verification. In addition, other low-abundance (<0.1%) bacteriawere not affected by SPM with a generally increasing total abun-dance. In order to further determine the potential resistance bac-teria, isolation of ARBs was carried out. Only one strain(Pseudomonas oceani) was isolated under aerobic condition

(Fig. S1), and no colony was observed under anaerobic condition.Unfortunately, this strain was not detected by the high-throughputsequencing; thus, its role in the systems remained unknown thatneeds subsequent research.

3.5. Correlation analysis among ARGs, functional genes anddominant bacteria

The correlation among ARGs, functional gene and dominantbacteria was analyzed and illustrated in Fig. 6 and Table S4. In total,33 pairs showed significant correlations. Among them, intI1showed a significantly positive correlation with the six genes,including two ARGs (ereA (r ¼ 0.784, p < 0.01), mphA (r ¼ 0.726,p < 0.01)), and five functional genes (hdh (r ¼ 0.570, p < 0.01), hzsA(r ¼ 0.689, p < 0.01), AMX (r ¼ 0.614, p < 0.01), nosZ (r ¼ 0.732,p < 0.01) and norB (r ¼ 0.541, p < 0.01)). IntI1 encodes a class 1integrase that promotes the acquisition, exchange and expressionof genes in the gene cassette, which can promote the horizontaltransfer of resistance genes (Gao et al., 2017). Liu et al. (2012) re-ported that played an important role in the transfer of tetracyclineresistance genes. Candidatus Kuenenia was positively correlatedwith ereA (r ¼ 0.460, p < 0.05),mphA (r ¼ 0.448, p < 0.05) and intI1(r ¼ 0.445, p < 0.05), indicating that it might carry these ARGsduring the long-term experimental period. The ermB, ermQ and

Fig. 5. Microbial community analysis at both phylum level (A) and genus level (B).

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 1209938

ermF genes encode methylase, so they were significantly positivelycorrelated with each other with the ecoefficiency of r ¼ 0.818(p < 0.01), r¼ 0.627 (p < 0.01) and r¼ 0.584 (p < 0.01), respectively.Coordination of the resistance genes with same mechanism wasalso reported in other studies, like tetracycline and oxytetracycline(Bai et al., 2019; Zhang et al., 2019a). The proteins encoded by hzsA,hdh and nirS catalyze the anammox reaction; thus, positive corre-lations were observed among these functional genes. The variationsin their abundances were related to the performance of theanammox system. For example, the abundances of hzsA and hdh inR0 and R1 on Day 105 were significantly higher than those in R2.Simultaneously, the NRE of R0 and R1 were also much higher thanthat of R2, due to the inhibition of 5 mg L�1 SPM on R2. Moreover,some resistance genes (such as ereA and mphA) were positively

correlated with functional genes. Fan et al. (2019) also observed asimilar phenomenon, that the tetracycline resistance gene tetMwaspositively related to all functional genes (r> 0.75, p< 0.05), and tetBalso has correlation with hdh (r ¼ 0.808, p < 0.05) and nirS(r ¼ 0.795, p < 0.05).

Based on the previous microbial and genetic analysis, theinfluencingmechanism of SPM on anammox could be speculated asfollow: SPM entered the AnAOB by combining the surface acceptor,and then inhibited protein synthesis by stimulation the dissociationof peptidyl-tRNA from ribosomes (Menninger and Otto,1982); thus,AnAOB function and anammox activity were influenced, leading tothe deterioration of process performance. The decrease in theabundance of AnAOB provided more space and substrates for otherbacteria, which promoted the change in microbial community

Fig. 6. Correlation analysis among ARGs and dominant genera.

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 120993 9

structure. On the other hand, SPM resistance genes could transferthrough mobile genetic elements (integron). The microorganismsreceived resistance genes would tolerate the pressure of high-concentration SPM, and enriched in the system.

4. Conclusions

Based on the investigation on the effects of SPM, anammoxbacteria could adapt and tolerate to the low concentrations (lessthan1 mg L�1) of SPM, which was mainly attributed to the defenseof EPS and ARGs. SPM with higher concentrations (3e5 mg L�1)significantly inhibited anammox activity, leading to decline of ni-trogen removal capacity. Simultaneously, the abundance of Planc-tomycetes (Candidatus Kuenenia is its representative at genus level)also decreased, and most ARGs and intI1 increased. In comparison,the reactor fed by increasing concentrations of SPM has a betterresistance to SPM.When first exposed to 5 mg L�1 SPM, R1 could beself-recovered within a short recovery period, while R2 could not.Correspondingly, the performance fluctuations in R2 were moresignificant. This study revealed the performance, microbial andgenetic effects on anammox process under different exposurescenarios. The findings of this word provide a comprehensive un-derstanding on the response of anammox to the SPM pressure, andalso suggest that the anammox has the potential that remove ni-trogen from sewage containing high-concentration SPM after aneffective acclimation.

Declaration of competing interest

The authors declare that they have no known competingfinancial interests or personal relationships that could haveappeared to influence the work reported in this paper.

CRediT authorship contribution statement

Jing-Wu: Investigation. Nian-Si Fan: Investigation, Writing -review & editing, Funding acquisition. Ye-Ying Yu: Data curation.Yi-Jun He: Methodology. Yi-Heng Zhao: Validation. Quan Zhang:Software. Bao-Cheng Huang: Writing - review & editing. Ren-CunJin: Writing - review & editing, Funding acquisition.

Acknowledgement

The authors are grateful for financial support from the National

Natural Science Foundation of China (51878231), the Natural Sci-ence Foundation of Zhejiang Province (LQ20E080014) and theScience and Technology Development Program of Hangzhou(20191203B11).

Appendix A. Supplementary data

Supplementary data to this article can be found online athttps://doi.org/10.1016/j.jclepro.2020.120993.

References

Ashouri, R., Ghasemipoor, P., Rasekh, B., Yazdian, F., Mofradnia, S., Fattahi, M., 2019.The effect of ZnO-based carbonaceous materials for degradation of benzoicpollutants: a review. Int. J. Environ. Sci. Technol. 16 (3), 1729e1740.

Bai, Y., Xu, R., Wang, Q.-P., Zhang, Y.-R., Yang, Z.-H., 2019. Sludge anaerobic digestionwith high concentrations of tetracyclines and sulfonamides: dynamics of mi-crobial communities and change of antibiotic resistance genes. Biores. Technol.276, 51e59.

Du, L., Cheng, S., Hou, Y., Sun, X., Zhou, D., Bo, L., 2018. Influence of sulfadimethoxine(SDM) and sulfamethazine (SM) on anammox bioreactors: performance eval-uation and bacterial community characterization. Biores. Technol. 267.S0960852418307272-.

Fan, N.S., Zhu, X.L., Wu, J., Tian, Z., Bai, Y.H., Huang, B.C., Jin, R.C., 2019. Decipheringthe microbial and genetic responses of anammox biogranules to the single andjoint stress of zinc and tetracycline. Environ. Int. 132, 105097.

Gao, P., Gu, C., Wei, X., Li, X., Chen, H., Jia, H., Liu, Z., Xue, G., Ma, C., 2017. The role ofzero valent iron on the fate of tetracycline resistance genes and class 1 inte-grons during thermophilic anaerobic co-digestion of waste sludge and kitchenwaste. Water Res. 111, 92e99.

Gillings, M.R., Gaze, W.H., Pruden, A., Smalla, K., Tiedje, J.M., Zhu, Y.-G., 2014. Usingthe class 1 integron-integrase gene as a proxy for anthropogenic pollution.ISME J. 9, 1269.

Halling-Sørensen, B., Sengeløv, G., Tjørnelund, J., 2002. Toxicity of tetracyclines andtetracycline degradation products to environmentally relevant bacteria,including selected tetracycline-resistant bacteria. Arch. Environ. Contam. Tox-icol. 42, 263e271.

Jia, F., Yang, Q., Liu, X., Li, X., Li, B., Zhang, L., Peng, Y., 2017. Stratification of extra-cellular polymeric substances (EPS) for aggregated anammox microorganisms.Environ. Sci. Technol. 51, 3260e3268.

Jin, R.C., Yang, G.F., Zhang, Q.Q., Ma, C., Yu, J.J., Xing, B.S., 2013. The effect of sulfideinhibition on the ANAMMOX process. Water Res. 47, 1459e1469.

Li, W., Shi, Y., Gao, L., Liu, J., Cai, Y., 2013. Occurrence, distribution and potentialaffecting factors of antibiotics in sewage sludge of wastewater treatment plantsin China. Sci. Total Environ. 445e446, 306e313.

Liu, M., Zhang, Y., Yang, M., Tian, Z., Ren, L.R., Zhang, S.J., 2012. Abundance anddistribution of tetracycline resistance genes and mobile elements in anoxytetracycline production wastewater treatment system. Environ. Sci. Technol.46, 7551e7557.

Liu, M., Ding, R., Zhang, Y., Gao, Y., Tian, Z., Zhang, T., Yang, M., 2014. Abundance anddistribution of Macrolide-Lincosamide-Streptogramin resistance genes in ananaerobic-aerobic system treating spiramycin production wastewater. WaterRes. 63, 33e41.

Lotti, T., Kleerebezem, R., Loosdrecht, M.C.M., 2015. Effect of temperature change on

Jing-Wu et al. / Journal of Cleaner Production 258 (2020) 12099310

anammox activity. Biotechnol. Bioeng. 112, 98e103.Manish, K., Anindita, G., Santanu, M., 2019. Metal removal, partitioning and phase

distributions in the wastewater and sludge: performance evaluation of con-ventional, upflow anaerobic sludge blanket and downflow hanging spongetreatment systems. J. Clean. Prod. 25, 119426.

Mao, D., Yu, S., Rysz, M., Luo, Y., Yang, F., Li, F., Hou, J., Mu, Q., Alvarez, P.J., 2015.Prevalence and proliferation of antibiotic resistance genes in two municipalwastewater treatment plants. Water Res. 85, 458e466.

Marcos, L.S.O., Maria, I., Xavier, Q., Roy, N.L., Binoy, K.S., Luis, F.O.S., 2019. Nano-particles from construction wastes: a problem to health and the environment.J. Clean. Prod. 219, 236e243.

Mayor, M.A.G., Gonz�alez, G.P., Martínez, R.M.G., Hernando, P.F., Alegría, J.S.D., 2017.Synthesis and characterization of a molecularly imprinted polymer for thedetermination of spiramycin in sheep milk. Food Chem. 221, 721e728.

Meng, Y., Sheng, B., Meng, F., 2019. Changes in nitrogen removal and microbiota ofanammox biofilm reactors under tetracycline stress at environmentally andindustrially relevant concentrations. Sci. Total Environ. 668, 379e388.

Menninger, J.R., Otto, D., 1982. Erythromycin, carbomycin, and spiramycin inhibitprotein synthesis by stimulating the dissociation of peptidyl-tRNA from ribo-somes. Antimicrob. Agents Chemother. 21, 811e818.

Molinuevo, B., García, M.C., Karakashev, D., Angelidaki, I., 2009. Anammox forammonia removal from pig manure effluents: effect of organic matter contenton process performance. Biores. Technol. 100, 2171e2175.

Phanwilai, S., Piyavorasakul, S., Noophan, P., Daniels, K.D., Snyder, S.A., 2020. Inhi-bition of anaerobic ammonium oxidation (anammox) bacteria by addition ofhigh and low concentrations of chloramphenicol and comparison of attached-and suspended-growth. Chemosphere 238, 124570.

Qi, P.-Q., Li, J., Dong, H.-Y., Wang, D., Bo, Y.-T., 2018. Performance of anammoxprocess treating nitrogen-rich saline wastewater: kinetics and nitrite inhibition.J. Clean. Prod. 199, 493e502.

Rizzo, L., Manaia, C., Merlin, C., Schwartz, T., Dagot, C., Ploy, M.C., Michael, I., Fatta-Kassinos, D., 2013. Urban wastewater treatment plants as hotspots for antibioticresistant bacteria and genes spread into the environment: a review. Sci. TotalEnviron. 447, 345e360.

Rodríguez-Rojas, A., Rodríguez-Beltr�an, J., Couce, A., Bl�azquez, J., 2013. Antibioticsand antibiotic resistance: a bitter fight against evolution. Int. J. Med. Microbiol.303, 293e297.

Sarmah, A.K., Meyer, M.T., Boxall, A.B.A., 2006. A global perspective on the use, sales,exposure pathways, occurrence, fate and effects of veterinary antibiotics (VAs)in the environment. Chemosphere 65, 725e759.

Shi, Z.-J., Hu, H.-Y., Shen, Y.-Y., Xu, J.-J., Shi, M.-L., Jin, R.-C., 2017. Long-term effects ofoxytetracycline (OTC) on the granule-based anammox: process performanceand occurrence of antibiotic resistance genes. Biochem. Eng. J. 127, 110e118.

Shojaie, A., Fattahi, M., Jorfi, S., Ghasemi, B., 2018. Synthesis and evaluations ofFe3O4eTiO2eAg nanocomposites for photocatalytic degradation of 4-chlorophenol (4-CP): effect of Ag and Fe compositions. Int. J. Integrated Care9 (2), 141e151.

Spring, S., K€ampfer, Peter, Karl, H.S., 2001. Limnobacter thiooxidans gen. nov., sp.nov., a novel thiosulfate-oxidizing bacterium isolated from freshwater lakesediment. Int. J. Syst. Evol. Microbiol. 51, 1463e1470.

Su, J.-Q., An, X.-L., Li, B., Chen, Q.-L., Gillings, M.R., Chen, H., Zhang, T., Zhu, Y.-G.,2017. Metagenomics of urban sewage identifies an extensively shared antibioticresistome in China. Microbiome 5, 84.

Ullah, R., Yasir, M., Bibi, F., Abujamel, T.S., Hashem, A.M., Sohrab, S.S., Al-Ansari, A.,Al-Sofyani, A.A., Al-Ghamdi, A.K., Al-Sieni, A., Azhar, E.I., 2019. Taxonomic di-versity of antimicrobial-resistant bacteria and genes in the Red Sea coast. Sci.Total Environ. 677, 474e483.

Xu, J.J., Zhang, Z.Z., Chen, Q.Q., Ji, Z.Q., Zhu, Y.H., Jin, R.C., 2017. The short- and long-term effects of Mn(2þ) on biogranule-based anaerobic ammonium oxidation(anammox). Biores. Technol. 241, 750e759.

Xu, J.-J., Cheng, Y.-F., Xu, L.-Z.-J., Liu, Y.-Y., Zhu, B.-Q., Fan, N.-S., Huang, B.-C., Jin, R.-C.,2019. The revolution of performance, sludge characteristics and microbialcommunity of anammox biogranules under long-term NiO NPs exposure. Sci.Total Environ. 649, 440e447.

Yang, G.F., Jin, R.C., 2013. Reactivation of effluent granular sludge from a high-rateAnammox reactor after storage. Biodegradation 24, 13e32.

Zhang, Z.Z., Zhang, Q.Q., Guo, Q., Chen, Q.Q., Jiang, X.Y., Jin, R.C., 2015a. Anaerobicammonium-oxidizing bacteria gain antibiotic resistance during long-termacclimatization. Biores. Technol. 192, 756e764.

Zhang, Z.Z., Zhang, Q.Q., Xu, J.J., Deng, R., Ji, Z.Q., Wu, Y.H., Jin, R.C., 2015b. Evaluationof the inhibitory effects of heavy metals on anammox activity: a batch teststudy. Biores. Technol. 200, 208e216.

Zhang, Z.Z., Xu, J.J., Hu, H.Y., Shi, Z.J., Ji, Z.Q., Deng, R., Shi, M.L., Jin, R.C., 2016. Insightinto the short- and long-term effects of inorganic phosphate on anammoxgranule property. Biores. Technol. 208, 161e169.

Zhang, Z.-Z., Xu, J.-J., Shi, Z.-J., Bai, Y.-H., Cheng, Y.-F., Hu, H.-Y., Jin, R.-C., 2017a.Unraveling the impact of nanoscale zero-valent iron on the nitrogen removalperformance and microbial community of anammox sludge. Biores. Technol.243, 883e892.

Zhang, Z.-Z., Xu, J.-J., Shi, Z.-J., Cheng, Y.-F., Ji, Z.-Q., Deng, R., Jin, R.-C., 2017b.Combined impacts of nanoparticles on anammox granules and the roles ofEDTA and S2- in attenuation. J. Hazard Mater. 334, 49e58.

Zhang, Z.Z., Cheng, Y.F., Xu, L.Z.J., Bai, Y.H., Jin, R.C., 2018. Anammox granules showstrong resistance to engineered silver nanoparticles during long-term exposure.Biores. Technol. 259, 10e17.

Zhang, Q.-Q., Bai, Y.-H., Wu, J., Xu, L.-Z.-J., Zhu, W.-Q., Tian, G.-M., Zheng, P., Xu, X.-Y.,Jin, R.-C., 2019a. Microbial community evolution and fate of antibiotic resistancegenes in anammox process under oxytetracycline and sulfamethoxazolestresses. Biores. Technol. 293, 122096.

Zhang, Q.Q., Zhao, Y.H., Wang, C.J., Bai, Y.H., Wu, D., Wu, J., Tian, G.M., Shi, M.L.,Mahmood, Q., Jin, R.C., 2019b. Expression of the nirS, hzsA, and hdh genes andantibiotic resistance genes in response to recovery of anammox processinhibited by oxytetracycline. Sci. Total Environ. 681, 56e65.

Zhang, X., Chen, Z., Ma, Y., Chen, T., Zhang, J., Zhang, H., Zheng, S., Jia, J., 2019c.Impacts of erythromycin antibiotic on Anammox process: performance andmicrobial community structure. Biochem. Eng. J. 143, 1e8.

Zhang, X., Chen, Z., Ma, Y., Zhang, N., Pang, Q., Xie, X., Li, Y., Jia, J., 2019d. Response ofAnammox biofilm to antibiotics in trace concentration: microbial activity, di-versity and antibiotic resistance genes. J. Hazard Mater. 367, 182e187.

Zhao, S., Qiu, S., Xu, X., Ciampitti, I.A., Zhang, S., He, P., 2019. Change in strawdecomposition rate and soil microbial community composition after strawaddition in different long-term fertilization soils. Appl. Soil Ecol. 138, 123e133.

Zhu, P., Chen, D., Liu, W., Zhang, J., Shao, L., Li, J.A., Chu, J., 2014. Hydroxylation andhydrolysis: two main metabolic ways of spiramycin I in anaerobic digestion.Biores. Technol. 153, 95e100.