ENVIRONMENTAL BIOTECHNOLOGY

Effect of hydraulic retention time (HRT) on the biodegradationof trichloroethylene wastewater and anaerobic bacterialcommunity in the UASB reactor

Ying Zhang & Xin Wang & Miao Hu & Pengfei Li

Received: 19 July 2014 /Revised: 11 September 2014 /Accepted: 12 September 2014# Springer-Verlag Berlin Heidelberg 2014

Abstract This study utilizes the unique merits of an 8-L labo-ratory upflow anaerobic sludge blanket (UASB) reactor fortreating synthetic wastewater containing trichloroethylene(TCE). The reactor was operated at different hydraulic retentiontimes (HRT) of 25, 20, 15, 10, and 5 h. TCE removal efficiencydecreased from 99 to 85 % when the HRT was lowered downfrom 25 to 5 h, as well as chemical oxygen demand (COD)removal efficiency (from 95 to 84.15 %). Using Illumina 16SrRNA gene MiSeq sequencing, we investigated the evolution ofbacterial communities in the anaerobic sludge under five differ-ent conditions of HRT. In total, 106,387 effective sequences ofthe 16S rRNA gene were generated from 5 samples that widelyrepresented the diversity of microbial community. Sequenceanalysis consisting of several novel taxonomic levels rangingfrom phyla to genera revealed the percentages of these bacterialgroups in each sample under different HRTs. The differencesfound among the five samples indicated that HRT had effects onthe structures of bacterial communities and the changes of bac-terial communities associated with the effect of HRT on theperformance of the reactor. Sequence analyses showed thatBacteroidetes and Firmicutes were the dominant phyla. It isnotable that the classDehalococcoidiawas found in the samplesat HRT of 5, 10, 20, and 25 h, respectively, in which there weresome dechlorination strains.Moreover, a tremendous rise of TCEremoval efficiency from HRTof 5 h to HRT of 10 h was found.

Keywords Anaerobic granular sludge . Hydraulic retentiontime . Trichloroethylene . Upflow anaerobic sludge blanketreactor . High-throughput sequencing

Introduction

With the widespread use of chlorinated hydrocarbons, trichlo-roethylene (TCE) is one of the problematic contaminantsobserved in the groundwater. Because TCE is a dense non-aqueous phase liquid (DNAPL), when discharged into agroundwater system, it sinks beneath the water table andresides on the bottom of the aquifer, often within a lowpermeability layer (Liang et al. 2007). The toxic and persistentTCE causes a serious health threat to humans and the envi-ronment. Under anaerobic conditions, TCE can bemicrobiallydechlorinated to cis-dichloroethene (cis-DCE), vinyl chloride(VC), and finally to ethene in pure cultures, mixed cultures,and groundwater environments (Friis et al. 2007). Of theproducts, VC is the only one considered as the largest threatto human health and the environment (Hwu and Lu 2008).Apparently, the treatment of completely dechlorinating TCEto ethene is required for groundwater bioremediation.

Anaerobic treatment has been widely used due to its par-ticular advantages (Cheng et al. 2010). Because of its simpledesign, easy construction and maintenance (Diamantis andAivasidis 2007), short hydraulic retention time (HRT), greatefficiency, and low operating cost (Bao et al. 2009), theupflow anaerobic sludge blanket (UASB) reactor has attainedsuccessful applications. Numerous studies have demonstratedthat UASB reactor has been applicable to treating aromaticand aliphatic chlorinated chemicals, such as tetrachloroethene(PCE) (Hwu and Lu 2008), tetrachloroethane (Basu andGupta 2010), and trichlorotoluene (Sponza and Atalay2005). Furthermore, many studies have shown the biodegra-dation of TCE in the anaerobic reactors (Siggins et al. 2011a,b; Mitra and Gupta 2013; Ozdemir et al. 2007).

An understanding of the microbial composition ofsludge is important for understanding the mechanism ofbiological treatment processes (Hesham et al. 2011). Mo-lecular fingerprinting techniques such as denaturing

Y. Zhang (*) :X. Wang :M. Hu : P. LiSchool of Resource and Environment, Northeast AgriculturalUniversity, Harbin 150030, Chinae-mail: zhangyinghr@hotmail.com

Appl Microbiol BiotechnolDOI 10.1007/s00253-014-6096-6

gradient gel electrophoresis (DGGE) and polymerasechain reaction (PCR)-cloning approach have been widelyused to evaluate microbial community structure in watersupply systems (Sekar et al. 2012). However, these com-munity analysis techniques assess a mere snapshot of thedominant members, with little information on taxa withmedium to low abundances (Zhang et al. 2012). High-throughput sequencing is a second-generation DNA se-quencing platform in investigating the microbial diversityof the environmental samples, as it provides enough se-quencing depth to cover the complex microbial commu-nities (Shendure and Ji 2008). Recent studies have beenapplied to analyze microbial communities in soil (Roeschet al. 2007), wastewater (McLellan et al. 2010), andmarine water (Qian et al. 2011). High-throughput se-quencing can provide significant insights into the evolu-tion of microbial community resulting from the variationof the different conditions.

The influential factors in the treatment of wastewater con-taminated with TCE by a UASB reactor have less beenresearched. Furthermore, there was limited information onthe effect of HRT on the performance of the UASB reactortreating synthetic wastewater containing TCE. To our knowl-edge, no previous study has utilized high-throughput sequenc-ing to investigate the microbial community of TCE wastewa-ter in the UASB reactor. Therefore, the specific objective ofthis study is to investigate the influences of HRT on theperformance of the UASB reactor treating TCE and the chang-es of bacterial communities in the reactor at varying HRTs.

Materials and methods

Experimental appliance and synthetic wastewater

The UASB reactor was made of a transparent acrylic glasssheet as described by Zhang et al. (2014). One-liter TCE-acclimated anaerobic sludge was seeded in the reactor. Theinfluent chemical oxygen demand (COD) was kept constant at(3000±20)mg L−1 by adding glucose and lactate as primarysubstrates. The composition and characteristics of syntheticTCE wastewater employed in the present study were de-scribed by Zhang et al. (2014).

HRT study

The UASB reactor was operated at five different HRTs (25,20, 15, 10, and 5 h). For each HRT, the reactor was operatedfor 20–25 days at steady-state condition. The concentrationsof influent COD and TCE were (3000±20)mg L−1 and36.5 mg L−1, respectively. Influent alkalinity was kept at(1850±50)mg CaCO3L

−1 and pH was 7.0±0.2, respectively.The composition of other chemicals used in the synthetic

wastewater was the same as that described by Zhang et al.(2014).

DNA extraction, PCR amplication, and high-throughputsequencing

The bacterial communities in anaerobic sludge at each HRTcondition were investigated by Illumina high-throughput se-quencing. Sludge samples were collected at HRT of 25, 20,15, 10, and 5 h, respectively. DNA in the anaerobic sludgesamples was extracted by 3S DNA Isolation Kit for Environ-mental Samples (Bocai Biology, Shanghai, China) accordingto the manufacturer’s protocol. The DNA extracts were storedat −20 °C for the following PCR amplification. The V3 andV4 regions of the 16S rDNA gene were selected for PCR. Theprimers were 338 F (5′ACTCCTACG GGAGGCAGCA-3′)and 806R (5′GGACTACHVGGGTWTCTAAT-3′) (Masoudet al. 2011). The 20-μL PCR reaction mixture was composedof 4 μL 1× FastPfu Buffer, 2.5 mM dNTPs, 5 μM each offorward and reverse primers, 0.4 μL FastPfu Polymerase(TransGen Biotech, Beijing, China), and 10 ng DNA tem-plate. The PCR protocol consisted of an initial 2-min denatur-ation at 95 °C, followed by 25 cycles of denaturing at 94 °Cfor 30 s, annealing at 55 °C for 30 s, extension at 72 °C for30 s, and completed with a final extension at 72 °C for 5 min.PCR products were examined on a 2 % (w/v) agarose gel, andthe band was extracted and purified with the AxyPrepDNAGel (Axygen, CA, USA) and PCR Clean-up System.Amplicons from different sludge samples were sent out forpyrosequencing on the Illumina MiSeq platform at the Shang-hai Majorbio Bio-Pharm Technology Co., Ltd (Shanghai,China).

Sequence analysis and phylogenetic classification

All the raw reads were treated with the following processes.For quality control, the reads which contained one or moreambiguous bases (“N”) were removed firstly. Illumina se-quencing generated a pair of reads from the two ends(paired-end reads) for one DNA fragment. A self-writtenscript was developed to complementarily reverse one of thepaired-end reads and then compared with another one. Thenthe tag sequences were sorted into different individual filesaccording to the barcodes of all samples (Mao et al. 2013).

Sequence date was processed by read trimming and iden-tification of V3–V4 sequences, followed by filtering andassigning the operational taxonomic units (OTUs). OTUswere identified with a cutoff of 97 % identity. The reads fromfiltered OTUs are processed using Quantitative Insights intoMicrobial Ecology (QIME) program, to construct a represen-tative sequence for each OTU (Yadav et al. 2014). The repre-sentative sequences were assigned at different taxonomiclevels (from phylum to genus) to the SILVA dataset of bacteria

Appl Microbiol Biotechnol

following the Bayesian approach and cutoff of 97 %. Theclusters were constructed at a 3 % dissimilarity cutoff andserved as OTUs for generating predictive rarefaction modelsand for determining the abundance-based coverage estimators(ACE) and Chao 1 richness and Shannon diversity indices.The Venn diagram with shared and unique OTUs was used todepict the similarity and difference among the five communi-ties. The Venn diagram was using the R package (http://www.R-project.org/). The interrelationships between the bacterialcommunities of different samples at the five different HRTlevels were visualized using principal component analysis(PCA). It was conducted by R to group the bacterialcommunities of different samples. Redundancy analysis(RDA) was performed to determine the links between HRTand removal efficiencies and pyrosequencing data on the typeof genus. The RDA was carried out in R using the veganpackage (Sheik et al. 2012).

Nucleotide sequence accession numbers

The sequences obtained from pyrosequencing were submittedto the NCBI Sequence Read Archive database under theaccession numbers SRA148475.

Analytical methods

The influent and effluent COD, alkalinity, pH, and rate ofbiogas production were measured daily. Cumulative gas andmethane gas were measured using liquid displacement meth-od. TCE and its degradation products were analyzed by gaschromatograph (GC) (SHIMADZU GC-2014C) with a flameionization detector (FID) as described by Zhang et al. (2014).Measurements of COD and alkalinity were performed accord-ing to the Standard Methods (American Public Health Asso-ciation 1998).

One-way analysis of variance (ANOVA) for TCE andCOD removal was carried out to compare the mean valuesfor different operation conditions and to examine whetherthere were any significant differences among the means atthe 95 % confidence level. The ANOVA analysis was per-formed using the SPSS 17.0 software. P<0.05 was consideredstatistically significant.

Results

Effect of HRT on the performance of reactor

The UASB reactor was operated at the five different HRTs(25, 20, 15, 10, and 5 h) to understand the influence of changein HRT on reactor performance and the biodegradation ofTCE. The loading rates of COD in UASB reactor were 2.88,

3.60, 4.80, 7.20, and 14.40 kg m−3 day−1 at the HRTs of 25,20, 15, 10, and 5 h, respectively.

Figure 1a showed the changes of the COD for thefive different HRT pilot phases in the UASB reactor. Ingeneral, the results from the one-way ANOVA testshowed that there was a significant difference forCOD removal when the HRT was decreased from 25to 5 h in the reactor. The effluent COD concentrationincreased from (140±20) to (500±50)mg L−1 with thedecrease in HRT from 25 to 5 h. Meanwhile, the aver-age COD removal efficiency was also declined (from 95to 84.15 %) when the HRT was decreased from 25 to5 h. The changes of effluent alkalinity concentrationsduring the five different pilot phases of HRTs wereshown in Fig. 1b. In general, a decreasing trend wasobserved in the generation of alkalinity (from (1880±20) to (1790±14)mg CaCO3L

−1) in the effluent withthe decrease of HRT. The effect of various HRTs (25,20, 15, 10, and 5 h) on biogas production was alsoinvestigated (Fig. 1c). As HRT decreased every time,there would be a sharp increase of biogas production,and biogas methane content decreased significantly. Thepercentage of methane gas from (65±2)% (HRT=25 h)gradually decreased to (50±1)% (HRT=5 h).

TCE degradation products

The TCE removal efficiency decreased from 99 % (corre-sponding to 0.5 mg L−1) to 85 % (corresponding to5 mg L−1) as the HRT was lowered from 25 to 5 h (Fig. 1d).cis-DCE and VC sequentially accumulated as prevalent inter-mediates. As the TCE loading rate was increased from35.04 mg L−1 day−1 (HRT=25 h) to 175.2 mg L−1 day−1

(HRT=5 h), the concentrations of cis-DCE and VC were alsoincreased (Fig. 1e).

Structure of bacterial community under different HRTconditions

Overall bacterial phylogeny and diversity of bacterialphylotypes

Since the functions and behaviors of bacteria influenced theprocess performance of UASB, further investigation byIllumina pyrosequencing was performed to analyze the bacte-rial community diversities and phylogenetic structures of theanaerobic sludge under different HRT conditions. In total,106,387 effective sequences of the 16S rRNA gene weregenerated from 5 samples that widely represented the diversityof microbial community. Over 16,000 sequences were obtain-ed for each sample. The sequence information of samples andcalculated microbial diversity index were listed in Table 1.The numbers of OTUs, highest species richness (Chao 1),

Appl Microbiol Biotechnol

ACE, and Shannon-Weaver index at cutoff of 3 % were verysimilar and had no significant changes among the five HRTtreatments.

Tags that differed by not more than 3 % were clustered intoOTUs to calculate rarefaction. Rarefaction analysis wasemployed to standardize and compare observed taxon

Fig. 1 Variation of a influent and effluent COD and COD removalefficiency, b influent and effluent alkalinity, c biogas production andCH4 concentration in UASB reactor, d the influent and effluent

concentrations of TCE and TCE removal efficiency, and e the concentra-tions of TCE, cis-DCE, and VC under different HRTs

Table 1 Different indices ofsamples for five different HRTs Sample Effective reads OTUs ACE Chao 1 Shannon-Weaver

HRT1 (5 h) 18,795 256 300 285 3.45

HRT2 (10 h) 25,474 258 289 281 3.35

HRT3 (15 h) 20,445 278 306 301 3.73

HRT4 (20 h) 16,358 237 279 286 3.47

HRT5 (25 h) 25,315 278 307 305 3.41

Appl Microbiol Biotechnol

richness between samples and to identify whether the samplewas unequally sampled. In this study, rarefaction curvesshowed similar patterns for all samples (Fig. 2). To evaluatethe distribution of OTUs among the different HRT samples, aVenn diagram was constructed (Fig. 3). This showed that 173OTUs, embracing 53.6 % of the sequences, were common toall five samples. The majority of the shared OTUs belonged toBacteroidetes and Firmicutes. Further, the number of specialOTUs of the five samples was one, six, four, four, and four atthe HRTs of 5, 10, 15, 20, and 25 h, respectively. They wererepresented by a low percentage of sequences. As shown inFig. 4, PCA analysis showed significant separation of thebacterial communities under different HRT conditions. Thecumulative percentage variance of species explained by thefirst axis was 55.37 %. The first axis and the second axisexplained 90.31 % variance of species. PCA analysis indicat-ed that the microbial communities were largely determined byHRT.

Taxonomic complexity of the bacterial community

Bacteria from the five samples demonstrated similar diversi-ties but different abundances. Figure 5a showed the relativebacterial community abundance on the phylum level. In total,14 identified phyla were observed, and Bacteroidetes,Firmicutes, Chloroflexi, Synergistetes, and Proteobacteriawere the dominant phyla. Bacteroidetes and Proteobacteriawere the dominant phyla in previous studies on TCE degra-dation, and some of known dechlorination strains which wereemployed to treat TCE belonged to the phylum Bacteroidetesand Proteobacteria (Humphries et al. 2005; Gu et al. 2004).The relative abundance of the phyla (especial for the predom-inant phyla) was altered, resulting from the variation in HRT.Firmicutes was highly enriched under the HRT of 25 h andaccounted for 35.68 % of the total bacteria, while it showed adecreasing trend as HRT decreases. In the sample at HRT of5 h, the percentage of Firmicutes was only 18.5 %. Thepercentage of Bacteroidetes in the sample at HRT of 25 hwas 27.43 %, and it showed an increasing trend as HRTdecreases. In the sample at HRT of 5 h, Bacteroidetes wasthe most predominant phylum (41.22%) instead ofFirmicutes(18.5 %). The percentage of Chloroflexi was kept high all thetime.

Further analysis was made to show the relative bacterialcommunity abundance on the class level (Fig. 5b). The per-centages of bacterial groups in each sample varied dependingon the HRTs at class level. While the percentage of the classesvaried among samples, the dominant class was vadinHA17 allthe time, which belonged to Bacteroidetes. The percentage ofvadinHA17 in all bacterial communities increased sharply asHRT increases from 5 to 25 h followed by Clostridia,Bacteroidia, Bacilli, Synergistia, Anaerolineae, and

Negativicutes. It is noteworthy that Dehalococcoidia can befound little in HRT of 5, 10, 20, and 25 h, respectively.

There were 20 orders detected in the total bacterial popu-lation (Fig. 5c). The percentages of each of these bacterialgroups varied depending on the HRT, and there was highvariability between each sample. The most dominant bacterialorder was vadinHA17_norank, which accounted for morethan 17 % for all the samples, followed by Clostridiales,Bacteroidales, Lactobacillales, Synergistales, Anaerolineales,and Selenomonadales. Their percentages varied clearly during

Fig. 2 Rarefaction curves based on the 16S rRNA gene sequencing ofthe HRTs

Fig. 3 Venn of the bacterial communities of five different HRT sampleson OTUs at 3 % distance

Appl Microbiol Biotechnol

different HRTs. A total 25 families were detected in the totalbacterial population, and most of the sequences belonged toseven families. Obviously, vadinHA17_norank was the mostdominant bacterial family, followed by Streptococcaceae,Synergistaceae, Anaerolineaceae, Porphyromonadaceae,Eubacteriaceae, and Acidaminococcaceae. The percentageof these bacterial families changed among HRT conditionsbut did not show a clear variation trend. Streptococcaceae andEubacteriaceae species were obviously enriched in HRTof 25and 20 h compared the other samples but were both sharplydecreased after the decrease of HRT. On the contrary,Synergistaceae and Anaerolineaceae were enriched in HRTfor 5 and 10 h and increased as HRT decreases.

A total 112 genera were classified among the test samples.vadinHA17_norank, Lactococcus, and Petrimonas obviouslydominated in the samples of all HRTs. The largest percentageof Lactococcuswas the sample at HRTof 10 h (up to 14.9 %),and it showed an increasing trend as the HRT decreases. Themost OTUs of Petrimonas were in the sample at HRT of 25 h(up to 2135), and the least OTUs were in the sample at HRTof15 h (up to 952). The Dehalobacter comprises a genus-levelgroup of bacteria within the phylum Chloroflexiwhich can befound in the five samples. The most OTUs of Dehalobacterwere in the sample at HRT of 10 h (up to 19). Spirochaeteswas found in all five samples, especially in HRTof 10 h (up to46). The Acetobacterium was found little in this study exceptat HRT of 20 h and was abundant in HRT of 15 h (up to 11).Clostridiawas found in large amount in all samples, especial-ly in HRTof 10 and 15 h, and Anaerolineaewas also detected.Syntrophobacter was present in all samples and the numbersof OTUs were large, especially in HRT of 20 and 25 h. The

percentage of Desulfovibrio changed between hydraulic con-ditions but did not show a clear variation trend.

As can be seen from Fig. 6, the RDA triplot, describing therelationship between major bacterial genera and HRT andremoval efficiencies, showed that several bacteria of the gen-era vadinBC27 wastewater-sludge group, Propionicicella,Longilinea, Alistipes, and Caldisericum had almost no signif-icant relationships with TCE and COD removal efficiencies.In addition, Lactococcus sp. was negatively correlated withTCE and COD removal efficiencies, while the generaClostridium, Desulfovibrio, Eubacterium, Petrimonas,Syntrophobacter, and Thermanaerovibrio had a positive cor-relation with them. Especially, the genera Clostridium andDesulfovibrio had nearly the same coordinates that cannotdistinguish them in the graph, and the genera Alistipes andCaldisericum had similar coordinates. When combining theinfluence of HRT and removal efficiencies and bacterial gen-era, the first and second axes explained 42.44 and 5.39 % ofthe variation in the bacterial communities, respectively.

Discussion

In the anaerobic biological treatment of the organic wastewa-ter system, the majority of the COD removal is mainly throughthe production and release of methane and carbon dioxide(Wang et al. 2004). The reason for the decreases in CODand TCE removal efficiency, while reducing the HRT, wasthat there was insufficient contact time available for sludgegranules to mineralize organic matters and intermetabolites.Ozdemir et al. (2007) documented that the level of the remov-al of COD and TCE was above 80 and 85%, respectively, in aUASB reactor. In general, a decreasing trend was observed inthe generation of alkalinity in the effluent with the decrease ofHRT. Higher alkalinity generation may be caused by theconversion of CO2 to bicarbonate at high pH condition prev-alent within the reactor (Chui et al. 1994). As HRT decreasedevery time, there would be a sharp increase of biogas produc-tion and biogas methane content decreased significantly. Thiswas probably attributed to the fact that the microorganism’sadaption to the new environment resulted from HRT shifting.Conventionally, biogas methane content has been employedas a secondary parameter of monitoring bioreactor stability(Enright et al. 2007). However, there was evidence to suggestthat in the presence of an alternative electron acceptor such asa chlorinated alkene, decreased methanogenesis was not aneffective indicator of metabolic stress, and continued stablebioreactor performance may continue, although methane pro-duction would be greatly reduced (Zhuang and Pavlostathis1994).

The TCE removal efficiency decreased as HRT waslowered from 25 to 5 h (Fig. 1d). A similar trend was observed

Fig. 4 PCA analysis of all five samples based on the composition ofbacterial communities

Appl Microbiol Biotechnol

during the degradation of PCE by Sponza and Atalay (2005).Similarly, there was significant difference for TCE removalfrom the one-way ANOVA test. As the HRT decreased from25 to 5 h, the TCE loading rate was increased from 35.04 to175.2 mg L−1 day−1. An excess amount of TCE could not becompletely degraded and also increased the risk of microbialcontact with TCE, resulting in the increase of the effluent TCE

concentration. A similar report indicated that the effluent TCEconcentration increased with decreasing HRT since TCE re-quired a longer detention time for complete degradation(Fliermans et al. 1988). This was likely because TCE as arecalcitrant compound took a longer detention time for com-plete biodegradation, because it was not a direct process thatTCE was biodegraded into ethene (Misra and Gupta 2001).

Fig. 5 Bacterial community structures under different HRTs on the a phylum level, b class level, c order level, d family level, and e genus level. Theabundance is presented in terms of a percentage of the total effective bacterial sequences in the sample

Appl Microbiol Biotechnol

Table 2 showed the TCE removal efficiency in several anaer-obic reactors. By contrast, our reactor had a good perfor-mance, especially at HRT of 25 h.

The concentration of cis-DCE was more than that of VC,which demonstrated that the degradation process of TCE tocis-DCE was fast, but the degradation process of cis-DCE toVC was slow. The detection of cis-DCE and VC showed thatthe degradation pathway of TCE in the UASB reactor was theanaerobic reductive dechlorination, which replaced a hydro-gen atomwith a chlorine atom. This result was consistent withthe findings of many other experts and scholars (Prakash andGupta 2000; Chui et al. 1994; Ndon et al. 2000).

In this study, Illumina pyrosequencing has revealed theimpacts of HRTon the structure and composition of microbialcommunity. The numbers of OTUs, Chao 1, ACE, andShannon-Weaver index showed that HRT had no significantinfluence on the diversity of the sludge bacterial community.This finding was similar to the bacterial diversity in a movingbed biofilm reactor (MBBR) under different HRTs (5, 10, and15 h) reported by Calderón et al. (2012). Rarefaction analysiswas employed to standardize and compare observed taxonrichness between samples and to identify whether a samplewas unequally sampled. In this study, rarefaction curvesshowed similar patterns for all samples (Fig. 2) and suggestedthat the bacterial community was showed well as they becamerelatively flat while the number of sequences analyzed in-creased. The Venn diagram showed that good homologyamong five bacterial communities on the species diversitywas due to the same feed water. PCA analysis showed a

significant separation of the bacterial communities under dif-ferent HRTconditions. PCA analysis indicated that the micro-bial communities were largely determined by HRT.

Bacteria from the five samples demonstrated similar diver-sities but different abundances. Previous reports showed thatmembers of Firmicutes were weak and could not resist strongshear imposed on them (Hu et al. 2012). The decrease of HRTcould bring the strong shear to the UASB system, whichwould explain why the abundance of Firmicutes was de-creased with decreasing HRTs. Members of Chloroflexi areimportant in the UASB reactor as they play an ecological rolein the degradation of carbohydrates and cellular materials (Huet al. 2012). This can explain why the percentage ofChloroflexi was kept high all the time. Moreover, its abun-dance in samples was high as expected. Besides the abilitysaid above, some of known dechlorination strains which wereemployed to treat various chlorinated hydrocarbons like TCEand cis-DCE belong to the phylum Chloroflexi (Gu et al.2004). It is noteworthy that Dehalococcoidia can be foundlittle in HRT of 5, 10, 20, and 25 h, respectively. There aresome dechlorination strains belonging to the classDehalococcoidia (Löff ler e t a l . 2013) , such asDehalococcoides ethenogenes 195 which was the first knownorganism capable of complete dechlorination of TCE to eth-ene (Maymo-Gatell et al. 1997).

The potential function of dominant genera also showed andrevealed the whole bacterial evolution of the biodegradationof TCE under different HRTs. Lactococcus and Petrimonasobviously dominated in the samples of all HRTs. Species ofthe genus Lactococcus could produce lactate by fermentationof glucose (Castelló et al. 2009). Petrimonaswas described asa mesophilic, anaerobic, fermentative bacterium and observedin UASB reactors (Castelló et al. 2009). Besides, it also couldproduce lactate by fermentation of glucose (Grabowski et al.2005). Therefore, it makes sense that Lactococcus andPetrimonas could degrade COD and produce electron donor(lactate) for TCE degradation, which would explain why therewere large amounts of Lactococcus and Petrimonas in allsamples and the variation trend of COD (Fig. 1a) and alkalin-ity (Fig. 1b) was in agreement with our results. TheDehalobacter comprise a genus-level group of bacteria withinthe phylum Chloroflexi which can be found in the five sam-ples. The most OTUs of Dehalobacter were in the sample atHRT of 10 h (up to 19). A previous study indicated thatDehalobacter utilized only H2 as an electron donor and PCEand TCE as electron acceptors in an anaerobic respirationprocess (Holliger et al. 1998). It was interesting to note thetremendous rise of TCE removal efficiency from HRT of 5 hto HRT of 10 h (Fig. 1d). It is clear that the variation of TCEremoval efficiency is associated with the numbers ofDehalobacter. Spirochaetes was found in all five samples,especially in HRTof 10 h (up to 46). Spirochaetes populationseither produce acetate from H2 and CO2 or ferment

Fig. 6 RDA of ordination diagram (triplot) showing the relationshipbetween environmental variables, samples, and bacterial genus in thepyrosequencing data; arrows, purple triangles, and circles denote envi-ronmental variables (TCE and COD removal efficiencies), types ofgenus, and samples. First axis is horizontal, second axis is vertical.HRT1, HRT2, HRT3, HRT4, and HRT5: HRT of 5, 10, 15, 20, and25 h, respectively

Appl Microbiol Biotechnol

carbohydrates and other complex substrates to acetate andother substances (Leadbetter et al. 1999). The Acetobacteriumwas detected little in this study except at HRTof 20 h and wasabundant in HRT of 15 h. The Acetobacterium was reportedpreviously that it was particularly enriched in the cis-DCEculture and somewhat in the VC enrichment, but not in theTCE and PCE enrichments. In this study, it was not surprisingthat they were present in the samples and could explain thedegradation products of TCE (Duhamel and Edwards 2006). Itwas observed that the most concentration of cis-DCE was inHRT of 15 h (Fig. 1e). A previous study indicated thatClostridia and Anaerolineae were both anaerobic bacteria(Yamada et al. 2006). Clostridia and Anaerotruncus couldbe responsible for hydrogen production (Chang et al. 2006;Castelló et al. 2009). Proteiniphilum has been observed in amethane fermentation reactor and described as proteolyticanaerobic bacteria (Chen and Dong 2005). Syntrophobacterwas present in all samples, and the numbers of OTUs werelarge, especially in HRT of 20 and 25 h. It could facilitate thestable production of CH4 (Siggins et al. 2012) and associatewith the variation trend of methane content (Fig. 1d). Themaintenance of an anaerobic environment is critical for con-tinued growth of Desulfovibrio, of which all known strainswere obligate anaerobes (Hernandez-Eugenio et al. 2000).The percentage of Desulfovibrio changed between hydraulicconditions but did not show a clear variation trend.Veillonellaceae has been shown to possess some capabilityfor hydrogen fermentation by using glucose or lactate assubstrate. Shida et al. (2012) observed a similarity with rep-resentatives of Veillonellaceae in a reactor fed with glucose ata controlled pH with an HRT of 2 h.

The potential function of these dominant genera mentionedabove was revealed as follows. Firstly, most species in thegenera Lactococcus, Spirochaetes, Proteiniphilum, andVeillonellaceae could produce lactate by fermentation of glu-cose or metabolize lactate to H2 which served as electrondonors to the dechlorinating culture. Subsequently, membersin the Dehalobacter genus and bacteria in class ofDehalococcoidia, which served as electron acceptors in ananaerobic respiration process, were capable to reduce TCE tocis-DCE and VC. Besides, members of the genera (such asPetrimonas, Clostridia, Anaerolineae, Proteiniphilum, and

Desulfovibrio) could maintain anaerobic conditions and pro-vide other nutrients for the growth of dechlorination strains inthe reactor. This can clearly indicate the whole bacterial evo-lution of the biodegradation of TCE.

In conclusion, the results of this study showed that TCE-containing wastewater could be treated effectively by a UASBreactor at different HRTs varying between 25 and 5 h. As HRTshortened from 25 to 5 h, COD removal efficiencies decreasedfrom 95 to 84 %, and TCE removal efficiency decreased from99 to 85 %. In this study, Illumina pyrosequencing has re-vealed the impacts of HRT on the microbial communitystructure and composition. The percentages of bacterialgroups in each sample varied depending on the HRTs atdifferent taxonomic levels. The potential function of dominantgenera also showed and revealed the whole bacterial evolutionof the biodegradation of TCE and varied depending on theHRTs. Illumina pyrosequencing provided more comprehen-sive information about bacterial communities at differentHRTs due to its capacity to identify a greater number ofsequences than traditional DNA approaches. The variationof bacterial communities associated with the effect of HRTon the performance of reactor. The information will assist inimproving the operation and control of UASB system to helpdegrade TCE.

Acknowledgments This work was supported by the National NaturalScience Foundation of China (No. 41073071) and University Science andTechnology Innovation Team Construction Projects of HeilongjiangProvince (2013TD003).

References

AWWA (1998) Standard methods for the examination of water andwastewater. AWWA, Washington, 20

Bao J, Liu H, Wang Y, Zhang L (2009) A granulation model usingdiosgenin wastewater in an upflow anaerobic sludge blanket reactor.Front Earth Sci China 3(3):361–366

Basu D, Gupta SK (2010) Biodegradation of 1, 1, 2, 2-tetrachloroethanein upflow anaerobic sludge blanket (UASB) reactor. BioresourTechnol 101(1):21–25

Calderón K, Martín-Pascual J, Poyatos JM, Rodelas B, González-Martínez A, González-López J (2012) Comparative analysis of thebacterial diversity in a lab-scale moving bed biofilm reactor

Table 2 TCE removal in different anaerobic reactors

Reactor Concentration of TCE (mg/L) TCE removal rates (%) HRT (h); T (°C) References

UASB reactor 36.5 99 to 85 25 to 5; 30 This study

UASB reactor 39 80 4.5; 35 Ozdemir et al. 2007

Expanded granular sludge bed (EGSB) 30 79 24; 15 Siggins et al. 2011b

EGSB 60 99 24; 35 Siggins et al. 2011c

Anaerobic hybrid reactor (AHR) 100 81.51 28; at ambient temperature Misra and Gupta 2001

AHR 50 99.93 18; 27±0.5 Mitra and Gupta 2013

Appl Microbiol Biotechnol

(MBBR) applied to treat urban wastewater under different opera-tional conditions. Bioresour Technol 121:119–126

Castelló E, García y Santos C, Iglesias T, Paolino G, Wenzel J,Borzacconi L, Etchebehere C (2009) Feasibility of biohydrogenproduction from cheese whey using a UASB reactor: links betweenmicrobial community and reactor performance. Int J HydrogenEnergy 34(14):5674–5682

Chang JJ, Chen WE, Shih SY, Yu SJ, Lay JJ, Wen FS, Huang CC (2006)Molecular detection of the clostridia in an anaerobic biohydrogenfermentation system by hydrogenase mRNA-targeted reverse tran-scription-PCR. Appl Microbiol Biotechnol 70(5):598–604

Chen S, Dong X (2005) Proteiniphilum acetatigenes gen. nov., sp. nov.,from a UASB reactor treating brewery wastewater. ISME J 55(6):2257–2261

Cheng J, Zhu X, Ni J, Borthwick A (2010) Palm oil mill effluenttreatment using a two-stage microbial fuel cells system integratedwith immobilized biological aerated filters. Bioresour Technol101(8):2729–2734

Chui HK, Fang HH, Li YY (1994) Removal of formate from wastewaterby anaerobic process. J Environ Eng 120(5):1308–1320

Diamantis VI, Aivasidis A (2007) Comparison of single-and two-stageUASB reactors used for anaerobic treatment of synthetic fruit waste-water. Enzym Microb Technol 42(1):6–10

Duhamel M, Edwards EA (2006) Microbial composition of chlorinatedethane-degrading cultures dominated by Dehalococcoides. FEMSMicrobiol Ecol 58(3):538–549

Enright AM, Collins G, O’Flaherty V (2007) Low-temperature anaerobicbiological treatment of toluene-containing wastewater. Water Res41(7):1465–1472

Fliermans CB, Phelps TJ, Ringelberg D, Mikell AT, White DC (1988)Mineralization of trichloroethylene by heterotrophic enrichmentcultures. AEM 54(7):1709–1714

Friis AK, Heimann AC, Jakobsen R, Albrechtsen HJ, Cox E, Bjerg PL(2007) Temperature dependence of anaerobic TCE-dechlorinationin a highly enrichedDehalococcoides-containing culture. Water Res41(2):355–364

Grabowski A, Tindall BJ, Bardin V, Blanchet D, Jeanthon C (2005)Petrimonas sulfuriphila gen. nov., sp. nov., a mesophilic fermenta-tive bacterium isolated from a biodegraded oil reservoir. IJSEM55(3):1113–1121

Gu AZ, Hedlund BP, Staley JT, Strand SE, Stensel HD (2004) Analysisand comparison of the microbial community structures of twoenrichment cultures capable of reductively dechlorinating TCEand cis-DCE. Environ Microbiol 6(1):45–54

Hernandez-Eugenio G, Fardeau ML, Patel BK, Macarie H, Garcia JL,Ollivier B (2000) Desulfovibrio mexicanus sp. nov., a Sulfate-reducing bacterium isolated from an upflow anaerobic sludge blan-ket (UASB) reactor treating cheese wastewaters. Anaerobe 6(5):305–312

Hesham AEL, Qi R, Yang M (2011) Comparison of bacterial communitystructures in two systems of a sewage treatment plant using PCR-DGGE analysis. J Environ Sci 23(12):2049–2054

Holliger C, Hahn D, Harmsen H, Ludwig W, Schumacher W, Tindall B,Zehnder AJ (1998)Dehalobacter restrictus gen. nov. and sp. nov., astrictly anaerobic bacterium that reductively dechlorinates tetra-andtrichloroethene in an anaerobic respiration. Arch Microbiol 169(4):313–321

HuM,Wang X,Wen X, Xia Y (2012)Microbial community structures indifferent wastewater treatment plants as revealed by 454-pyrosequencing analysis. Bioresour Technol 117:72–79

Humphries JA, Ashe AM, Smiley JA, Johnston CG (2005) Microbialcommunity structure and trichloroethylene degradation in ground-water. Can J Microbiol 51(6):433–439

Hwu CS, Lu CJ (2008) Continuous dechlorination of tetrachloroethene inan upflow anaerobic sludge blanket reactor. Biotechnol Lett 30(9):1589–1593

Leadbetter JR, Schmidt TM, Graber JR, Breznak JA (1999) Acetogenesisfrom H2 plus CO2 by spirochetes from termite guts. Science283(5402):686–689

Liang C, Wang ZS, Bruell CJ (2007) Influence of pH on persulfateoxidation of TCE at ambient temperatures. Chemosphere 66(1):106–113

Löffler FE, Yan J, Ritalahti KM, Adrian L, Edwards EA, KonstantinidisKT, Spormann AM (2013)Dehalococcoides mccartyi gen. nov., sp.nov., obligately organohalide-respiring anaerobic bacteria relevantto halogen cycling and bioremediation, belong to a novel bacterialclass, Dehalococcoidia classis nov., order Dehalococcoidales ord.nov. and family Dehalococcoidaceae fam. nov., within the phylumChloroflexi. IJSEM 63(Pt 2):625–635

Mao Y, Xia Y, Zhang T (2013) Characterization of Thauera-dominatedhydrogen-oxidizing autotrophic denitrifying microbial communitiesby using high-throughput sequencing. Bioresour Technol 128:703–710

Masoud W, Takamiya M, Vogensen FK, Lillevang S, Al-Soud WA,Sørensen SJ, Jakobsen M (2011) Characterization of bacterial pop-ulations in Danish raw milk cheeses made with different startercultures by denaturating gradient gel electrophoresis and pyrose-quencing. Int Dairy J 21(3):142–148

Maymo-Gatell X, Chien YT, Gossett JM, Zinder SH (1997) Isolation of abacterium that reductively dechlorinates tetrachloroethene to ethene.Science 276(5318):1568–1571

McLellan SL, Huse SM, Mueller-Spitz SR, Andreishcheva EN, SoginML (2010) Diversity and population structure of sewage-derivedmicroorganisms in wastewater treatment plant influent. EnvironMicrobiol 12(2):378–392

Misra C, Gupta SK (2001) Hybrid reactor for priority pollutant-trichloroethylene removal. Water Res 35(1):160–166

Mitra S, Gupta SK (2013) Biodegradation of trichloroethylene in anaer-obic hybrid reactor. Environ Prog Sustain 32(4):1055–1060

Ndon UJ, Randall AA, Khouri TZ (2000) Reductive dechlorination oftetrachloroethylene by soil sulfate-reducing microbes under variouselectron donor conditions. Environ Monit Assess 60(3):329–336

Ozdemir C, Dursun S, KaratasM, Sen N, Sahinkaya S (2007) Removal oftrichloroethylene (TCE) in up flow anaerobic sludge blanket reac-tors (UASB). Biotechnol Biotechnol Equip 21(1):107

Prakash SM, Gupta SK (2000) Biodegradation of tetrachloroethylene inupflow anaerobic sludge blanket reactor. Bioresour Technol 72(1):47–54

Qian PY, Wang Y, Lee OO, Lau SC, Yang J, Lafi FF, Wong TY (2011)Vertical stratification of microbial communities in the Red Searevealed by 16S rDNA pyrosequencing. ISME J 5(3):507–518

Roesch LF, Fulthorpe RR, Riva A, Casella G, Hadwin AK, Kent AD,Triplett EW (2007) Pyrosequencing enumerates and contrasts soilmicrobial diversity. ISME J 1(4):283–290

Sekar R, Deines P, Machell J, Osborn AM, Biggs CA, Boxall JB (2012)Bacterial water quality and network hydraulic characteristics: a fieldstudy of a small, looped water distribution system using culture-independent molecular methods. J Appl Microbiol 112(6):1220–1234

Sheik CS, Mitchell TW, Rizvi FZ, Rehman Y, Faisal M, Hasnain S,Mclnerney MJ, Krumholz LR (2012) Exposure of soil microbialcommunities to chromium and arsenic alters their diversity andstructure. PLoS One 7(6):e40059

Shendure J, Ji H (2008) Next-generation DNA sequencing. NatBiotechnol 26(10):1135–1145

Shida GM, Sader LT, Cavalcante de Amorim EL, Sakamoto IK,Maintinguer SI, Saavedra NK, Silva EL (2012) Performance andcomposition of bacterial communities in anaerobic fluidized bedreactors for hydrogen production: effects of organic loading rateand alkalinity. Int J Hydrogen Energy 37(22):16925–16934

Siggins A, Enright AM, O’Flaherty V (2011a) Methanogenic communitydevelopment in anaerobic granular bioreactors treating

Appl Microbiol Biotechnol

trichloroethylene (TCE)-contaminated wastewater at 37°C and15°C. Water Res 45(8):2452–2462

Siggins A, Enright AM, O’Flaherty V (2011b) Low-temperature (7°C)anaerobic treatment of a trichloroethylene-contaminated wastewa-ter: microbial community development. Water Res 45(13):4035–4046

Siggins A, Enright AM, O’Flaherty V (2011c) Temperature depen-dent (37–15°C) anaerobic digestion of a trichloroethylene-contaminated wastewater. Bioresour Technol 102(17):7645–7656

Siggins A, Enright AM, Abram F, Botting C, O’Flaherty V (2012) Impactof trichloroethylene exposure on the microbial diversity and proteinexpression in anaerobic granular biomass at 37° C and 15° C.Archaea 2012:2012

Sponza DT, Atalay H (2005) Treatment of trichlorotoluene in ananaerobic/aerobic sequential reactor system. Process Biochem40(1):69–82

Wang J, Huang Y, Zhao X (2004) Performance and characteristics of ananaerobic baffled reactor. Bioresour Technol 93(2):205–208

Yadav TC, Khardenavis AA, Kapley A (2014) Shifts in microbial com-munity in response to dissolved oxygen levels in activated sludge.Bioresour Technol 165:257–264

Yamada T, Sekiguchi Y, Hanada S, Imachi H, Ohashi A, Harada H,Kamagata Y (2006) Anaerolinea thermolimosa sp. nov., Levilineasaccharolytica gen. nov., sp. nov. and Leptolinea tardivitalis gen.nov., sp. nov., novel filamentous anaerobes, and description of thenew classesAnaerolineae classis nov. andCaldilineae classis nov. inthe bacterial phylum Chloroflexi. IJSEM 56(6):1331–1340

Zhang T, Shao MF, Ye L (2012) 454 Pyrosequencing reveals bacterialdiversity of activated sludge from 14 sewage treatment plants. ISMEJ 6(6):1137–1147

Zhang Y, Liu Y, Hu M, Jiang Z (2014) Acclimation of the trichloroeth-ylene anaerobic degradation granular sludge and the degradationcharacteristics in an upflow anaerobic sludge blanket reactor. WaterSci Technol 69(1):120–127

Zhuang P, Pavlostathis S (1994) Effect of chlorinated alkenes on thereductive dechlorination and methane production processes. WaterSci Technol 30(7):85–94

Appl Microbiol Biotechnol