Chemical Engineering Journal 223 (2013) 932–943

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /ce j

Properties of biofilm in a vermifiltration system for domestic wastewatersludge stabilization

1385-8947/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.cej.2013.01.092

⇑ Corresponding author. Tel.: +86 021 65984275.E-mail address: xingmeiyan@tongji.edu.cn (M. Xing).

Xiaowei Li, Meiyan Xing ⇑, Jian Yang, Yongsen LuKey Laboratory of Yangtze Water Environment for Ministry of Education, State Key Laboratory of Pollution Control and Resources Reuse, National Engineering Research Centerfor Urban Pollution Control, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China

h i g h l i g h t s

" The VF biofilms had less organic matter and microbial biomass than the BF biofilms." The VF biofilms had higher microbial enzyme activities than the BF biofilms." Proteobacteria seemed to be the important contributors in the VF system." The relationships between some indexes in VF biofilm was different from BF biofilm." The presence of earthworm was a main reason to change the VF biofilm property.

a r t i c l e i n f o

Article history:Received 27 June 2012Received in revised form 26 January 2013Accepted 30 January 2013Available online 9 February 2013

Keywords:VermifilterBiofilmMicrobial enzyme activityPolymerase chain reaction–denaturinggradient gel electrophoresisScanning electron microscopy

a b s t r a c t

Vermifiltration is an alternative and low-cost technology for stabilizing excess sludge from domesticwastewater treatment plants. The biofilm properties of a vermifilter (VF) with earthworms, Eisenia fetida,for domestic wastewater sludge (DWS) treatment were studied. A biofilter (BF) without earthwormsserved as the control. VF biofilms had lower levels of suspended solids (SSs), volatile SS, C, H, N and S con-tents, protein-like groups, and total viable cell numbers and larger humic acid-like fractions and protease,dehydrogenase, lipase, and amylase activities compared with BF biofilms. Furthermore, VF biofilms fea-tured richer diversity in their microbial community and more populations of Proteobacteria than BF bio-films. The relationships between organic matter and microbial eco-physiological indices in VF biofilmswere significantly different from those in BF biofilms. Overall findings indicated that earthworm presenceremarkably decreases organic matter contents and microbial biomass and improves microbial enzymeactivities and the community structure of VF biofilms.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

The introduction of earthworms to filtration systems, termedvermifiltration systems, was first advocated by José Toha [1,2]. Asan extension of vermicompost for solid wastes, vermifilters (VFs)were developed to treat a mixture of solids and liquids from house-hold or animal wastes with high organic pollution [3,4]. Severalstudies have been conducted to investigate the use of vermifiltersin wastewater treatment, including synchronous treatment of sew-age and sludge [2,5–11]. Recent studies have shown that VF is afeasible and efficient technology for stabilizing excess sludge fromdomestic wastewater treatment plants (WWTPs) [12–14]. Thereduction of volatile suspended solids (VSSs) using VF reaches56.2–66.6%, meeting up with the criteria for aerobic and anaerobicsludge stabilization (>40%) [14]. The capacity of VF to treat

domestic wastewater sludge (DWS) can be attributed to thevermicomposting process that occurs within the system andearthworm consumption of solid organic waste on the bedsurface [3].

DWS treatment and disposal pose challenges for domesticWWTPs worldwide due to environmental, economic, social, and le-gal concerns [12,15]. Several mechanical, physical, and chemicaltreatment processes, including ultrasonic, thermal, and ozonepre-treatment, require large amounts of capital and high opera-tional costs [14]. Compared with other technologies used in DWStreatment, such as anaerobic and aerobic digestion [12,15], VF isa low-cost ‘‘bio-safe’’ technique, and thus is more suitable forwastewater and DWS treatment of WWTPs in developing countries[4,13,14]. Previous studies have focused on treatment perfor-mance, earthworm–microorganism interactions, and organic-mat-ter distribution and transformation in vermifiltration systems [10–14]. However, biofilm properties in the vermifiltration system havenot been fully investigated.

X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943 933

The functions of conventional biofilters (BFs) rely heavily on themetabolism of microorganisms present in the biofilm, which is themost fundamental characteristic of BFs [14]. In a vermifiltrationsystem, the reduction and stabilization of DWS is involved by thejoint action of earthworms and microorganisms in the biofilm[14]. The relationship between microorganisms and earthwormsduring the vermicomposting process has been widely investigated[17]. In a vermicomposting system, microbes are responsible forthe biochemical degradation of organic matter of sewage sludge,whereas earthworms drive this process by conditioning the sub-strate and altering its biological activity [18,19]. However, littleinformation about the organic matter composition and microbialeco-physiological properties of biofilm in vermifiltration systems,as well as the influence of earthworm inoculation on VF biofilmproperties, is available.

A range of parameters, such as suspended solids (SSs), VSS, via-ble cell numbers, and activities of protease, dehydrogenase, lipase,glucosidase, and amylase, is indicative of biofilm development andactivity. VSS corresponds to the quantity of biofilm and representsthe non-ash part of the total SS [20]. Viable cells are quantified byextracting phopholipids and analyzing phosphate contents(cleaved from phospholipids); this parameter correlates to viablebiomass [20]. Phospholipids have the advantages that their con-centration remains fairly constant in relation to cell biomass andonce a cell dies the phospholipids have a short half-life. The prote-ase activity arising from the depolymerization of dissolved organicnitrogen from N-containing compounds is assumed to be a criticalpoint in the N cycle because polymers are not accessible to micro-organisms [21]. Dehydrogenase activity is essential in both miner-alization and transformation of organic C [22,23] and thus plays anessential role during the initial stages of oxidation of organic mat-ter by transferring hydrogen electrons from substrates to acceptors[20]. Lipase, glucosidase, and amylase are associated with carbonturnover in a wide range of ecosystems [24]. Thus, enzymes playan essential role in the biochemical transformation that involvesthe decomposition of organic matter in soil [25] and biologicalwastewater treatment processes [26]. Enzyme activities serve asindicators of microbiological activities in the biofilm.

Some advanced analytical techniques, such as fluorescence anal-ysis of humic acid-like (HAL) fraction, polymerase chain reaction–denaturing gradient gel electrophoresis (PCR–DGGE), and scanningelectron microscopy (SEM), are often used to further understandthe microbial characteristics contained within the biofilm. Unlikethe traditional plate-counting method, PCR–DGGE can provide amore comprehensive analysis of bacterial compositions, includingculturable and non-culturable microbes [27]. SEM can directly re-veal the microorganism profile and biofilm microstructure frommagnifications of 10 times to more than 500,000 times [12,28].

The objectives of the study are (1) to determine the organicmatter composition and microbial eco-physiological characteris-tics of VF biofilms using various techniques and (2) to investigatethe effect of earthworm presence on biofilms by comparing biofilmproperties in VFs with and without earthworms.

2. Materials and methods

2.1. Vermifilter system

A cylindrical VF consisting of a Perspex tubing (20 cm in diam-eter and 120 cm in depth) was set up (Supporting informationFig. S1) and assembled as previously described in Zhao et al. [14].The tubing contained a 100 cm filter bed of ceramic pellets (10–20 mm in diameter). A layer of plastic fiber was placed on top ofthe filter bed to avoid direct hydraulic influence on the earth-worms and ensure even influent distribution.

The VF was inoculated with Eisenia fetida at an initial earth-worm density of approximately 40 g L�1, while a conventional BFwas set up without earthworms as the control. The hydraulic load-ings of the filters were kept at 4 m d�1 during the experimentalperiod. The influent sludge was obtained from the aeration tankof a domestic WWTP (Quyang WWTP, Shanghai) and diluted to aconstant organic load of approximately 1.12 kg VSS m�3 d�1 usingtap water. After passing through the filter bed, the treated sludgeentered into a sedimentation tank below the VF, and the superna-tant in the sedimentation tank was recycled. The ratio of VSS to SSand the pH in the initial sludge were 70.4 ± 3.3% and 7.5 ± 0.4%,respectively. The initial earthworms were randomly picked froma Donghai farm, Shanghai, China, and cultured with cow dung assubstrate. After acclimation for approximately 30 d, the filterswere continuously operated for 330 d.

2.2. Biofilm analyses

2.2.1. SamplingCeramic pellet samples were collected from the filter bed in the

BF and VF at depths of 12.5, 37.5, 62.5, and 87.5 cm at the end ofthe experiment to evaluate biofilm properties. Samples from theBF were designated as B1, B2, B3, and B4, while those from theVF were designated as V1, V2, V3, and V4; all samples were rinsedwith sterile water. The biofilm in the wash water was collected andcentrifuged for 10 min at 8000 rpm and 4 �C. Settled biofilm sam-ples were then used for further analysis.

2.2.2. Organic matter compositionThe SS and VSS contents of the biofilm samples in the BF and VF

were assessed according to Chinese standard methods [14].Settled biofilm samples were freeze-dried and filtered using

0.15 mm sieves. C, H, N, and S contents in the samples were mea-sured in triplicate using the elemental analyzer Vario EL I(Germany).

2.2.3. Fluorescence analysisHAL fractions were extracted from freeze-dried and sieved bio-

film samples as previously described [16], and fluorescence excita-tion–emission matrix (EEM) spectra from aqueous solutions of theHAL fractions were obtained. The HAL fractions had a concentra-tion of 100 mg L�1 after overnight equilibration at 25 �C. The pHwas adjusted to 8 using 0.05 mol L�1 NaOH and measured by anF-4600 fluorescence spectrophotometer (Hitachi, Japan). Emissionand excitation slits were set to 5 nm bandwidths, and a scan speedof 12,000 nm min�1 was selected for both monochromators. EEMspectra were recorded by scanning the emission wavelength be-tween 250 nm and 500 nm, while the excitation wavelength wassequentially increased from 200 nm to 400 nm. Surfer 8.0 softwarewas used to analyze the fluorescence spectral data.

2.2.4. Total viable cell numberThe total viable cell number is represented as the quantity of

phospholipids. Extraction of phospholipids and release of phos-phate from phospholipids, as well as measurement of nanomoleconcentrations of phosphate released from phospholipids, wereperformed as previously described by Findlay et al. [29].

2.2.5. Enzymatic activitiesBiofilm samples from the two filters were analyzed for protease,

dehydrogenase, glucosidase, lipase, and amylase activities. Prote-ase activity was measured from the tyrosine derivatives generatedfrom 1 g of sample after incubation with sodium caseinate for10 min at 40 �C and subsequent reactions with Folin Ciocalteau re-agent [21]. Dehydrogenase activity was quantified according to themethod previously described by Zhang et al. [20], which involves

934 X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943

reduction of triphenyltetrazolium chloride (TTC) to triphenyl for-mazan (TPF) at 37 �C for 2 h in the dark. Estimation of glucosidaseactivity was carried out following the method descried by Airaet al. [21], which involves the colorimetric estimation of p-nitro-phenol (PNP) formed by the hydrolysis of p-nitrophenyl-b-d-gluco-piranoside (PNG) at 37 �C for 60 min. Lipase and amylase activitieswere measured using p-nitrophenyl palmitate and starch as sub-strates, respectively, according to the methods described by Yuet al. [26] and Winkler and Stuckmann [30].

2.2.6. Bacterial community structureBacterial genomic DNA was extracted from the biofilm samples

using a Fast DNA Spin Kit for soil (QBIOgene, Carlsbad, CA, USA).For denaturing gradient gel electrophoresis (DGGE) analysis,

nested polymerase chain reaction (PCR) was performed on the totalDNA using the bacterial universal primers 8f and 1492r [14] alongwith 341f with a 40 bp GC clamp and 534r [31]. DGGE was performedusing the D-Code System (Bio-Rad, USA). Polyacrylamide gels of 8%were prepared with denaturing gradients ranging from 35% to 55%,and then run at 60 �C and 120 V for 6 h. DGGE fingerprints were ana-lyzed using Quantity One software (version 4.4, Bio-Rad).

Clear and high-intensity representative bands were excisedfrom DGGE gels using a previously described method [14] and thensequenced by Shanghai Sangon Biotech Co., Ltd.

2.2.7. SEMForty healthy earthworms were randomly obtained from the VF

system and kept in culture dishes in the dark for 24 h to obtain acast. Casts in the culture dishes and those that adhered to theearthworms were collected.

SEM images of the cast and biofilms were obtained using aVEGA TS 5136MM scanning electron microscope (Tescan S.R.O.,Czech Republic).

2.3. Data analysis

Statistical analyses were carried out using SPSS 13.0 (SPSS Inc.,USA). Significant difference between samples were analyzed usingANOVA procedures, and main effects were separated by Duncan’smultiple range tests. The Pearson correlation coefficient (rp) wasused to estimate linear correlations. The value of rp was always be-tween +1 and �1, where +1 denotes a perfect positive correlation,�1 indicates a perfect negative correlation, and 0 indicates an ab-sence of relationship.

The Shannon index was calculated to show the structural diver-sity (richness and evenness) of the microbial community based onthe number of bands present and the relative intensities of thebands in each lane [7,32] using the following equation:

H ¼ �X

pi log pi ¼ �Xs

i¼1

ðni=NÞ logðni=NÞ

where ni is the height of peak and N is the sum of all peak heights inthe curve.

Nucleotide sequences were subsequently compared with thosedeposited in the GenBank database. Sequences determined in thisstudy and those obtained from the DNA database were alignedusing CLUSTAL W. Distance matrix analyses were carried out usingp-distance, and neighbor-joining trees were constructed usingMEGA version 4.0. The sequences generated in this study weredeposited in the National Center for Biotechnology Information(NCBI, America) under accession numbers JQ235757 to JQ235776.

Eco-physiological profiles were designed as sum-ray plots bycombining microbial biomass, enzyme activities, numbers ofbands, and Shannon diversity index values (H) of the DGGE pro-files. Every combined index was divided by the maximum values

of the index in the samples, and normalized to eliminate differ-ences in absolute values among different indices. The integratedarea of the plot for each biofilm sample was measured using AdobeAcrobat 9� (Adobe Systems Incorporated, CA, USA).

3. Results and discussion

3.1. Organic matter composition

VF biofilms had lower SS and VSS contents compared with BFbiofilms (Supporting information Fig. S2), implying that VF biofilmshave lower organic matter contents than BF biofilms. Moreover, VFbiofilms had lower C, H, N, and S contents than BF biofilms (Support-ing information Fig. S3), indicating that VF biofilms had low protein-like (N-containing) and carbohydrate-like (C- and H-containing)groups. The results showed that VF biofilms degraded more organicmatter and had lower possibility of bio-clogging than BF biofilms.

These findings may be attributed to the vermicomposting pro-cess that occurred in the VF due to the presence of earthworms[3]. During vermicomposting, the earthworms act as mechanicalblenders and by comminuting the organic matter, and modify itsphysical and chemical status, gradually reducing its C:N ratio,increasing the surface area exposed to microorganisms and makingit much more favorable for microbial activity and further decom-position [18]. Therefore, earthworms promote organic matter deg-radation in the biofilms and deferred VF bio-clogging through thejoint action of earthworms and microorganisms. The SS, VSS, C,H, N, and S contents in the BF and VF biofilm samples decreasedwith depth, indicating an increase in the degree of degradation oforganic matter in the biofilms.

3.2. Fluorescence analysis

Fluorescence EEM spectra of HAL fractions isolated from VF andBF biofilms are shown in Fig. 1. The spectra of biofilms from thetwo filters were characterized by several fluorophores, but fea-tured their own excitation/emission wavelength pairs (EEWPs)and specific fluorescence intensity (SFI) (Table 1). HAL fractionsfrom BF biofilm samples showed two main peaks (peaks 1 and2), whereas three main peaks (peaks 1–3) were found in VF biofilmsamples, except in the V1 sample. According to Chen et al. [33],peak 1 belonged to an aromatic protein region, which could in-clude tryptophan and BOD5, peak 2 pertained to soluble microbialby-product-like materials, and peak 3 fell in the region related tothe amount of HAL compounds.

VF biofilm samples yielded lower SFIs in peaks 1 and 2 andshowed the presence of an additional peak (peak 3), except inthe V1 sample, implying that HAL fractions from VF biofilms havethe formation of humic acid-like materials, and less amount of pro-tein-like group than those from BF biofilms. These findings arelikely due to the fact that earthworm activity accelerates the humi-fication of organic matter [18].

The amount and quality of HAL components in composts andvermicomposts are considered to be important indicators of theirbiological maturity and chemical stability [34]. Thus, the resultsobtained show that the maturity and stability of VF biofilms seemsto be better than those of BF biofilms, confirming and complement-ing the findings on organic matter composition in this study.

3.3. Microbial biomass

The total viable cell numbers of the biofilms in the two filtersare shown in Fig. 2. The total viable cell numbers of the VF biofilmswere significantly lower than those of BF biofilms, except for theV4 sample. These results suggest that VF biofilm samples have

-

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

B1 - -1- -2- -4- -3- -1

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

V1

-- -1

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

B2-- -1- -2- -4- -3

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

V2

- -1- -2- -4- -3- -1- -2

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

B3 -- -1- -2- -4

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

V3

-- -1- -2

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

B4

250 300 350 400 450 500

Em

200

220

240

260

280

300

320

340

360

380

400

Ex

04080120160200240280320360400440480

V4

Fig. 1. Fluorescence excitation–emission matrix spectra (as contour maps) of humic acid-like fractions extracted from the BF (B) and VF (V) biofilms.

X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943 935

lower microbial biomasses than BF biofilm samples because earth-worm activity causes a decrease in microbial biomass contained inVF biofilms. This result is in accordance with previous studiesbased on total phospholipid fatty acids analysis in vermicompo-sting [35]. Digestion of organic material by earthworms has nega-tive effects on microbial biomass production [35]. Earthworms

may also reduce the microbial biomass by enhancing the depletionof resources for microbes [17].

The total viable cell numbers of the VF and BF biofilm samplesgradually decreased with depth ranging from 41.09 g L�1 to32.80 g L�1 and 58.50–35.45 g L�1, corresponding to the depth dis-tributions of SS, VSS, and organic elemental contents. These results

Table 1Ex/Em maxima of fluorescence excitation–emission matrix spectra (as contour maps)of humic acid-like fractions extracted from BF and VF biofilms.

Sample Peak 1 Peak 2 Peak 3

Ex/Ema SFIb Ex/Em SFI Ex/Em SFI

B1 280/350 145 235/355 44.1B2 280/345 187.2 240/350 59.2B3 280/345 145.3 235/350 44.9B4 280/355 128.1 235/355 43.1V1 280/350 158.5 240/355 54.27V2 280/345 114.9 240/355 41.9 325/420 32.1V3 280/350 83.4 235/355 22.69 365/445 34.6V4 280/355 105.1 240/355 38.5 335/435 30.85

a Ex/Em represented the Excitation/Emission wavelength pairs.b SFI referred to the specific fluorescence intensity.

100

75

50

25

00.0 0.3 0.6 0.9 1.2

dc

cb

ab

ab

a

a

Dep

th (

cm)

Dep

th (

cm)

Dehydrogenase

(mg h-1 g-1 VSS)

100

75

50

25

00 1 2 3 4 5

dc

cb

b

b

a

a

Dep

th (

cm)

Lipase

(mg min-1 g-1 VSS)

100

75

50

25

00 10 20 30 40 50 60 70

d

c

c

bc

aba

ab

a

ab

Total viable cell

mmol P L-1

Fig. 2. Depth distribution in microbial biomass and enzymatic activities in BF a

936 X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943

may be due to the fact that the available organic matter for micro-organisms decreased with depth.

3.4. Microbial enzyme activities

Protease, dehydrogenase, glucosidase, lipase, and amylaseactivities of the BF and VF biofilms are given in Fig. 2. The protease,dehydrogenase, and lipase activities of VF biofilms were signifi-cantly higher than those in BF biofilms, implying that organic mat-ter degradation and transformation in VF biofilms occur faster thanin BF biofilms. These findings further verify the findings on organicmatter composition above.

The results were attributed to earthworm inoculation in the VFsystem. Earthworms exert considerable effects on the microbial

0.0 0.3 0.6 0.9 1.2

e

dd

c

c

b

b

Protease

(mg min-1 g-1 VSS)

a

0.0 0.5 1.0 1.5 2.0

c

b

ab

ab

aba

abab

Glucosidase

(mg min-1 g-1 VSS)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

cbc

bbc

bcb

ba

Amylase

(mg min-1 g-1 VSS)

VF BF

nd VF biofilms. Different letters indicate significant differences at P < 0.05.

X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943 937

enzymatic activities of VF biofilms via two possible mechanisms ofaction: (1) by increasing substrate availability and (2) by activatingmicrobial metabolism [21]. Several microbial enzymes whoseactivities depend on substrate availability, such as glucosidase,could be induced by the substrate [21]. Thus, the increase in sub-strate availability caused by earthworms could result in an in-crease in microbial enzyme activity. Moreover, enzyme activityhas been widely reported to be greater in casts of earthworms thanin undigested material [36,37].

The activities of protease, dehydrogenase, lipase, and amylase inVF biofilms decreased with depth, ranging from 1.05 mg min�1 g�1

to 0.44 mg min�1 g�1 VSS, 1.01–0.63 mg h�1 g�1 VSS, 3.80–1.98 mg min�1 g�1 VSS, and 2.78–2.00 mg min�1 g�1 VSS. Thedepth distribution of enzymatic activities in BF biofilms was con-sistent with those in VF films, ranging from 0.96 mg min�1 g�1 to0.14 mg min�1 g�1 VSS, 0.83–0.57 mg h�1 g�1 VSS, 3.23–1.50 mg min�1 g�1 VSS, and 2.48–1.48 mg min�1 g�1 VSS. These re-sults were likely caused by the decline in available organic mattercontents for microbial action with depth (Supporting informationFigs. S2 and S3).

Fig. 3. Analysis of microbial community structures of VF and BF biofilms. (A) DGGE profi

3.5. Microbial community structure

PCR–DGGE has been widely used to reveal bacterial popula-tions in various environments and has been proposed as asemi-quantitative measure of bacterial diversity [13,14,32]. Themicrobial community compositions of BF and VF biofilms wereanalyzed by PCR–DGGE coupled with sequence analysis of 16SrDNA gene fragments of dominant bands. The results are givenin Figs. 3 and 4.

Approximately 45 and 70 bands were detected in the BF and VFbiofilms, respectively, indicating that VF biofilms have higher bac-terial richness than BF biofilms (Table 2). Similarly, Shannon indi-ces revealed that bacterial richness was higher in VF biofilms thanin BF biofilms, which is in agreement with previous results[14,32,38]. Moreover, cluster analysis revealed that the DGGEbanding pattern in VF biofilms had low similarity with BF biofilms(Fig. 3b and c).

Different banding patterns and intensities obtained from VF andBF biofilm samples were likely due to the presence or absence ofearthworms. Earthworms influence the bacterial community struc-

le; (B) Diagram of the DGGE profile; and (C) Phylogenetic tree of the DGGE profile.

band10 Terrimonas sp. (HM124372.1)

band16 Uncultured Sphingobacteriales bacterium (HM592607.1)

band9 Uncultured bacterium (HM460642.1)

Uncultured bacterium (GU388397.1) Uncultured bacterium (GQ396872.1)

band14 band13

band8 Uncultured Bacteroidetes bacterium (EU283360.1) band1 Flavobacterium sp. (GU295968.1)

band7 Uncultured Bacteriodetes bacterium (GQ469315.1)

band12 Uncultured Bacteroidetes bacterium (FJ916293.1)

band6 Uncultured Xanthomonadaceae bacterium (EU642089.1)

band18 Uncultured Legionella sp. (GU979447.1)

Uncultured Pseudomonas sp. (GQ183209.1) band15 band17 Uncultured actinobacterium (EU298500.1)

band4 Uncultured Acidobacteria bacterium (AM491119.1)

band20 Uncultured bacterium (GU732118.1)

band2 Uncultured Geobacter sp. (DQ395028.1)

band19 Uncultured eubacterium (AF495397.1)

band3 Uncultured delta proteobacterium (EU980148.1)

band5 Uncultured Bdellovibrio sp. (FJ542862.1)

band11 Uncultured bacterium (DQ673352.1)

0.1

Gammaproteobacteria

Betaproteobacteria

Deltaproteobacterium

Firmicutes

Acidobacteria

Actinobacteria

Bacteroidetes

Fig. 4. Phylogenetic tree based on neighbor-joining analysis of gene sequences from the V3 region of bacterial with a length of 197 bp. Band 1–20 (accession numbersJQ235757–JQ235776) represent the sample sequences.

Table 2Functional diversity of bacterial community in BF and VF biofilms based on theShannon index (H).

Filter beddepth (cm)

VF BF

Number ofbands

Shannonindex (H)

Number ofbands

Shannonindex (H)

12.5 16 6.10 11 5.7837.5 17 5.93 11 5.6762.5 19 6.09 12 5.3087.5 18 6.07 11 5.67

938 X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943

ture of VF biofilms in five possible ways: (1) burrowing improvesthe aerobic condition for aerobic microorganism [18]; (2) selectivefeeding causes a decrease in the abundance of some bacteria [39];(3) production of mucus and casts that are enriched with availableforms of C, N and P [21] changes the abundances of some microbialpopulations; (4) generation of indigenous gut-associated micro-flora increases some microorganisms species [32]; and (5) compe-tition for organic food inhibits the growth of some microbe [17].

Dominant bands recovered from BF and VF biofilms were indi-vidually identified as different members using bands in the Gen-Bank database as comparison. The phylogenetic positions ofthese bands are illustrated in neighbor-joining tree shown inFig. 4. All identified bands showed similarities ranging from 96%to 100% with previously identified gene sequences, six of which(bands 1, 7, 8, 10, 12, and 16) belonged to Bacteroidetes. Band 1 be-

longed to Flavobacterium sp. of class Flavobacteria, and bands 10and 16 belonged to order Sphingobacteriales. Among the bands thatwere identified as Proteobacteria, bands 6, 15, and 18 belonged toXanthomonadaceae, Legionella, and Pseudomonas of class Gamma-proteobacteria, respectively, while bands 2 and 3 were clusteredwith Geobacter sp. of Betaproteobacteria, and Bdellovibrio sp. of classDeltaproteobacterium, respectively. Bands 4 and 19 ascribed to classAcidobacteria of Acidobacteria, and genus Eubacterium of Firmicutes,while Actinobacteria was observed in the sequences of band 17.Other bands (e.g., bands 9, 11, 13, 14, and 20) were not identifiedin any bacterium group and likely belonged to uncultured bacte-rium groups.

In the present study, microbial communities of VF and BF bio-films were composed of mostly Bacteroidetes and Proteobacteriafollowed by uncultured bacteria. Small amounts of bacteria wereobserved and classified according to phylogenetic clusters relatedto Acidobacteria, Firmicutes, and Actinobacteria. Similar findingshave been reported in previous studies [7,14]. Additionally, Vivaset al. [32] found that the most abundant phylum in vermicompo-sting was Proteobacteria. Fracchia et al. [40] reported that Bacterio-detes was the predominant bacteria in vermicomposting. Danonet al. [41] found that Proteobacteria and Bacteriodetes were thedominant populations in compost.

Several additional bands found in VF biofilms were not pres-ent in BF biofilms, including bands 2 (Geobacter sp), 3 (Deltapro-teobacterium), 6 (Xanthomonadaceae), 11 (uncultured bacterium),13 (Alphaproteobacterium), 16 (Sphingobacteriales), and 18

Fig. 5. SEM micrographs of the BF biofilm (B), the VF biofilm (V) and the earthworm cast (C). The photographs were recorded at the indicated magnification (6 or 10 k�).

X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943 939

(Actinobacterium) (Figs. 3 and 4). This result implies that VF bio-films appear to have more populations of Proteobacteria than inBF biofilms because indigenous gut-associated microflorae fromearthworms contribute to the microbial community. The similar

findings have been demonstrated in cured vermicompost [32].These results support previous findings that phylum Proteobacte-ria is an important contributor in vermifiltration or vermicompo-sting processes [14].

940 X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943

3.6. SEM

Typical SEM micrographs of VF and BF biofilm samples, as wellas the earthworm-cast sample, are shown in Fig. 5. The BF biofilmsample appeared to have a loose, flaky structure, while the earth-worm-cast sample exhibited a distinct physical appearance charac-terized by a predominantly spherical cell-like structure, which wasmore compact than in the BF biofilm. The microstructure of VF bio-film sample was more compact than that of the BF biofilm sample,but more loosely than that noted in the earthworm cast, was looserthan that of the earthworm cast, implying that the earthworm castaffects the microstructure of VF biofilms and may be an importantconstituent of VF biofilms. These results confirm that earthworm

Fig. 6. Microbial eco-physiological indexe

casts are an important factor that results in higher activities ofmicrobial enzymes in VF biofilms compared with BF biofilms.

3.7. Microbial eco-physiological profiles

Microbial eco-physiological indices are shown in Fig. 6. Theintegrated areas of the plots in VF biofilms were 425.40, 397.65,294.24, and 264.62 mm2. These values are higher than those ob-tained from the BF biofilms, which were 431.19, 272.60, 202.84,and 170.14 mm2, except for the V1 sample. The results obtainedshow that the microbial eco-physiological characteristics of VF bio-films are superior to those of BF biofilms. The integrated areas ofthe plots in BF and VF biofilm samples gradually decreased with

s of the biofilms in BF (B) and VF (V).

X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943 941

depth, confirming the findings on microbial biomass, enzymeactivities, and bacterial community structure in this study.

3.8. Correlation analysis of several parameters in determining biofilmproperties

Correlation analysis was used to investigate the relationship ofindexes in BF and VF biofilm properties. The results of the variousparameters are listed in Tables 3 and 4.

In BF biofilms, a strong correlation was found between theparameters on organic matter composition, such as VSS, SS, N, C,H, and total contents, but no relationship was established betweenthe S content and other organic matter indices. No significant cor-

Table 3Pearson correlation coefficients for linear regression between the various parameters abou

VSS SS Organic elemental composition

N C S H

SS 0.991**

N 0.978* 0.949*

C 0.973* 0.971* 0.976*

S 0.135 0.262 -0.050 0.123H 0.979* 0.969* 0.988* 0.998** 0.072Total 0.977* 0.976* 0.976* 1.000** 0.137 0.998**

MB 0.927 0.894 0.981* 0.966* �0.132 0.976*

DH 0.839 0.785 0.933 0.893 �0.324 0.910PR 0.958* 0.911 0.981* 0.919 �0.148 0.941GL 0.956* 0.984* 0.906 0.966* 0.370 0.953*

AM 0.825 0.744 0.910 0.807 �0.449 0.841LP 0.995** 0.988* 0.957* 0.948 0.172 0.954*

ARh 0.991** 0.976* 0.994 0.990* 0.057 0.996**

a The total of N, C, S and H contents.b Microbial biomass.c Dehydrogenase.d Protease.e Glucosidase.f Amylase.g Lipase.h The integrated areas of the plots.

* Correlation is significant at the 0.05 level (2-tailed).** Correlation is significant at the 0.01 level (2-tailed).

Table 4Pearson correlation coefficients for linear regression between the various parameters abou

VSS SS Organic elemental composition

N C S H

SS 1.000**

N 0.886 0.887C 0.998** 0.998** 0.911S �0.973* �0.973* �0.761 �0.959H 0.981* 0.979* 0.822 0.971* �0.988*

Total 0.996** 0.996** 0.923 0.999** �0.951* 0.969*

MB 0.897 0.900 0.900 0.914 �0.799 0.793DH 0.923 0.925 0.961* 0.942 �0.816 0.836PR 0.495 0.499 0.802 0.543 �0.285 0.343GL �0.837 �0.836 �0.496 �0.804 0.940 �0.897AM 0.956* 0.954* 0.922 0.960* �0.912 0.963*

LP 0.912 0.914 0.979* 0.934 �0.795 0.827ARh 0.781 0.784 0.955* 0.816 �0.619 0.666

a The total of N, C, S and H contents.b Microbial biomass.c Dehydrogenase.d Protease.e Glucosidase.f Amylase.g Lipase.h The integrated areas of the plots.

* Correlation is significant at the 0.05 level (2-tailed).** Correlation is significant at the 0.01 level (2-tailed).

relation was found between the N content and other organic mat-ter indexes in VF biofilms, but the S content had a strongly negativecorrelation with VSS, SS, H, and total contents. Moreover, microbialbiomass had a significant correlation with N, C, H, and total con-tents in BF biofilms. In contrast, no significant correlation wasfound between the microbial biomass and any organic matter in-dex in VF biofilms. These observations show that the organic mat-ter compositions in BF and VF biofilms were very different,implying that the presence of earthworms significantly changesthe organic matter composition of VF biofilms. The organic matterin BF biofilms was affected by protein-like (N-containing groups)and carbohydrate-like (C- and H-containing materials) compoundsand was present in microbial viable cells. In contrast, the organic

t BF biofilm properties.

MBb Microbial enzyme activities

Totala DHc PRd GLe AMf LPg

0.962*

0.885 0.979*

0.920 0.942 0.9050.970* 0.867 0.747 0.8360.803 0.911 0.941 0.948 0.6510.953* 0.888 0.789 0.949 0.946 0.7990.991** 0.967* 0.899 0.964* 0.946 0.861 0.974*

t VF biofilm properties.

MBb Microbial enzyme activities

Totala DHc PRd GLe AMf LPg

0.9120.945* 0.985*

0.559 0.764 0.785�0.789 �0.580 �0.579 0.052

0.967* 0.801 0.876 0.507 �0.7480.939 0.970* 0.997** 0.808 �0.544 0.8860.826 0.928 0.956* 0.929 �0.318 0.774 0.969*

942 X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943

matter of VF biofilms was mostly influenced by carbohydrate-likematerials possibly because the presence of earthworms promotesthe decomposition of carbohydrate-like groups and microbial cellsin VF biofilms but has little influence on the degradation of N- andS-containing groups. The results confirm the findings on organicmatter composition and microbial biomass in this study.

According to Tables 3 and 4, the correlations between themicrobial enzyme activities and the organic matter indexes in BFand VF biofilms were different. In BF biofilms, a significant correla-tion was found between protease activity and VSS and N contents,between glycosidase activity and VSS, SS, C, H, and total contents,and between lipase activity and VSS, SS, N, H, and total contents. Astrong correlation was found between dehydrogenase activity andN and total contents, between amylase activity and VSS, SS, C, H,and total contents, and between lipase activity and N content inVF biofilms. Moreover, the integrated area was significantly corre-lated with the indices, such as VSS, SS, N, C, H, and total contents,microbial biomass, and activities of protease and lipase in BF bio-films. However, no correlation was found between N content andactivities of dehydrogenase and lipase in VF biofilms. These resultsindicate that the presence of earthworms significantly affects themicrobial eco-physiological indices of VF biofilms, resulting inchanges in the relationship between the indices of microbial en-zyme activities and organic matter. Microbial enzyme activitieswere simultaneously influenced by the available amount of organicmatter of the substrates and the action of earthworms in VFbiofilms.

Previous studies have shown that a highly significant correla-tion may be found between microbial biomass and enzyme activi-ties, such as glucosidase, cellulose, and protease [21,24]. However,no significant correlation was found in the present study, exceptbetween microbial biomass and dehydrogenase activity in BF,and between microbial biomass and activities of protein and lipasein VF. These results may be due to the fact that DWS treated by BFand VF was mainly composed of microbial cells. Thus, the biofilmscame from a mixture of influent sludge, earthworm cast, or mi-crobes growing in the filters, causing the relationships betweenthem to become more complex.

In summary, the correlations of several indices on biofilm prop-erties in VF biofilms were very different from those in BF biofilms,such as between organic matter parameters and between indicesof microbial enzyme activities and organic matter. The results con-firm that the presence of earthworms significantly changes therelationship between organic matter and microbial eco-physiolog-ical indices on biofilm properties.

4. Conclusion

VF biofilms had lower organic matter contents and microbialbiomasses found on the filter bed but had higher enzymatic activ-ities compared with BF biofilms, implying that organic matter deg-radation was much faster in VF biofilms. Moreover, VF biofilmsfeatured richer diversity in their microbial community than BF bio-films, and Proteobacteria was an important contributor to the ver-mifiltration system. Additionally, the presence of earthwormssignificantly changed the relationships between organic matterand microbial eco-physiological indexes. The results obtainedshow that earthworms can change the biofilm properties of VF sys-tems through their burrowing action, selective feeding, mucus andcast production, and organic food competition with microbes.

Acknowledgements

The research was funded by the National Science Foundation ofChina (51109161), the PhD Programs Foundation of Ministry of

Education of China (20110072120029), the National Water Pollu-tion Control and Management Technology Major Projects(2011ZX07316004), and the Fundamental Research Funds for theCentral Universities (0400219187). The Open Analysis Fund forLarge Apparatus and Equipments of Tongji University (2012055),and the National Spark Program of China (2010GA680004) is alsoacknowledged.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.cej.2013.01.092.

References

[1] M.L. Aguilera, Purification of wastewater by vermifiltration, Doctoral Thesis,University of Montpellier 2, France, 2003, pp. 188.

[2] Y. Li, P. Robin, D. Cluzeau, M. Bouché, J. Qiu, A. Laplanche, M. Hassouna, P.Morand, C. Dappelo, J. Callarec, Vermifiltration as a stage in reuse of swinewastewater: monitoring methodology on an experimental farm, Ecol. Eng. 32(2008) 301–309.

[3] M. Taylor, W.P. Clarke, P.F. Greenfield, The treatment of domestic wastewaterusing small-scale vermicompost filter beds, Ecol. Eng. 21 (2003) 197–203.

[4] S.D. Ghatnekar, M.F. Kavian, S.M. Sharma, S.S. Ghatnekar, G.S. Ghatnekar, A.V.Ghatnekar, Application of vermin-filter-based effluent treatment plant (pilotscale) for biomanagement of liquid effluents from the Gelatin industry, Dyn.Soil, Dyn. Plant 4 (2010) 83–88.

[5] R.J. Hughes, J. Nair, G. Ho, The risk of sodium toxicity from bed accumulation tokey species in the vermifiltration wasterwater treatment process, Bioresour.Technol. 100 (2009) 3815–3819.

[6] P. Tomar, S. Suthar, Urban wastewater treatment using vermin-biofiltrationsystem, Desalination 282 (2011) 95–103.

[7] L. Wang, F. Guo, Z. Zheng, X. Luo, J. Zhang, Enhancement of rural domesticsewage treatment performance, and assessment of microbial communitydiversity and structure using tower vermifiltration, Bioresour. Technol. 102(2011) 9462–9470.

[8] L. Wang, Z. Zheng, X. Luo, J. Zhang, Performance and mechanisms of amicrobial–earthworm ecofilter for removing organic matter and nitrogen fromsynthetic domestic wastewater, J. Hazard. Mater. 195 (2011) 245–253.

[9] R.K. Sinha, G. Bharambe, U. Chaudhari, Sewage treatment by vermifiltrationwith synchronous treatment of sludge by earthworms: a low-cost sustainabletechnology over conventional systems with potential for decentralization,Environmentalist 28 (2008) 409–420.

[10] M. Xing, X. Li, J. Yang, Treatment performance of small-scale vermifilter fordomestic wastewater and its relationship to earthworm growth, reproductionand enzymatic activity, Afr. J. Biotech. 9 (44) (2010) 7513–7520.

[11] M. Xing, J. Yang, Y. Wang, J. Liu, F. Yu, A comparative study of synchronoustreatment of sewage and sludge by two vermifiltration using epigeicearthworm Eisenia fetida, J. Hazard. Mater. 185 (2011) 881–888.

[12] M. Xing, L. Zhao, J. Yang, Z. Huang, Z. Xu, Distribution and transformation oforganic matter during liquid-state vermiconversion of activated sludge usingelemental analysis and spectroscopic evaluation, Environ. Eng. Sci. 28 (2011)619–626.

[13] M. Xing, X. Li, J. Yang, B. Lv, Y. Lu, Performance and mechanism ofvermifiltration system for liquid-state sewage sludge treatment usingmolecular and stable isotopic techniques, Chem. Eng. J. 197 (2012) 143–150.

[14] L. Zhao, Y. Wang, J. Yang, M. Xing, X. Li, D. Yi, D. Deng, Earthworm–microorganism interactions: a strategy to stabilize domestic wastewatersludge, Water Res. 44 (2010) 2572–2582.

[15] D. Fytili, A. Zabaniotou, Utilization of sewage sludge in EU application of oldand new methods – a review, Renew. Sust. Energy Rev. 12 (2008) 116–140.

[16] X. Li, M. Xing, J. Yang, Z. Huang, Compositional and functional features ofhumic acid-like fractions from vermicomposting of sewage sludge and cowdung, J. Hazard. Mater. 185 (2011) 740–748.

[17] J. Domínguez, State of the art and new perspectives on vermicompostingresearch, in: C.A. Edwards (Ed.), Earthworm Ecology, second ed., CRC Press,Boca Raton FL USA, 2004, pp. 401–424.

[18] J. Domínguez, C.A. Edwards, Vermicomposting organic wastes: a review, in:S.H.S. Hanna, W.Z.A. Milkhail, (Eds.), Soil Zoology for Sustainable Developmentin the 21st Century, Cairo, 2004.

[19] M. Xing, X. Li, J. Yang, Z. Huang, Y. Lu, Changes in the chemical characteristicsof water-extracted organic matter from vermicomposting of sewage sludgeand cow dung, J. Hazard. Mater. 205–206 (2012) 24–31.

[20] L. Zhang, M. Xing, Y. Wu, Z. Huang, J. Yang, Spatial distributions of biofilmproperties and flow pattern in NiiMi process, Bioresour. Technol. 102 (2011)1406–1414.

[21] M. Aira, F. Monroy, J. Domínguez, Earthworms strongly modify microbialbiomass and activity triggering enzymatic activities during vermicompostingindependently of the application rates of pig slurry, Sci. Total Environ. 385(2007) 252–261.

X. Li et al. / Chemical Engineering Journal 223 (2013) 932–943 943

[22] R. Kizikaya, Dehydrogenase activity in Lumbricus terrestris casts andsurrounding soil affected by addition of different organic wastes and Zn,Bioresour. Technol. 99 (2008) 946–953.

[23] M. Molina-Muñz, J.M. Poyatos, B. Rodelas, C. Pozo, M. Manzanera, E. Hontoria,J. Gonzalez-Lopez, Microbial enzymatic activities in a pilot-scale MBRexperimental plant under different working conditions, Bioresour. Technol.101 (2010) 696–704.

[24] M. Aira, F. Monroy, J. Domínguez, Microbial biomass governs enzyme activitydecay during aging of worm-worked substrates through vermicomposting, J.Environ. Qual. 36 (2007) 448–452.

[25] J. Tao, B. Griffiths, S. Zhang, X. Chen, M. Liu, F. Hu, H. Li, Effects of earthwormson soil enzyme activity in an organic residue amended rice–wheat rotationagro-ecosystem, Appl. Soil Ecol. 42 (2009) 221–226.

[26] G.H. Yu, P.J. He, L.M. Shao, D.J. Lee, Enzyme activities in activated sludge flocs,Appl. Microbiol. Biotechnol. 77 (2007) 605–612.

[27] K.L. Ho, Y.C. Chung, C.P. Tseng, Continuous deodorization and bacterialcommunity analysis of a biofilter treating nitrogen-containing gases fromswine waste storage pits, Bioresour. Technol. 99 (2008) 2757–2765.

[28] J.H. Priester, A.M. Horst, L.C. Van De Werfhorst, J.L. Saleta, L.K. Mertes, P.A.Holden, Enhanced visualization of microbial biofilms by staining andenvironmental scanning electron microscopy, J. Microbiol. Meth. 68 (2007)577–587.

[29] R.H. Findlay, G.M. King, L. Watling, Efficacy of phospholipids analysis indetermining microbial biomass in sediments, Appl. Environ. Microbiol. 55(1989) 2888–2893.

[30] U.K. Winkler, M. Stuckmann, Glycogen, hyaluronate, and some otherpolysaccharides greatly enhance the formation of exolipase by Serratiamarcescens, J. Bacteriol. 138 (1979) 663–670.

[31] G. Muyzer, E.C. De Waal, A.G. Uitterlinden, Profiling of complex microbialpopulations by denaturing gradient gel electrophoresis analysis of polymerasechain reaction-amplified genes coding for 16S rRNA, Appl. Environ. Microb. 59(1993) 695–700.

[32] A. Vivas, B. Moreno, S. Garcia-Rodriguez, E. Benitez, Assessing the impact ofcomposting and vermicomposting on bacterial community size and structure,and microbial functional diversity of an olive-mill waste, Bioresour. Technol.100 (2009) 1319–1326.

[33] W. Chen, P. Westerhoff, J.A. Leenheer, K. Booksh, Fluorescence excitation–emission matrix regional integration to quantify spectra for dissolved organicmatter, Environ. Sci. Technol. 37 (2003) 5701–5710.

[34] E. Romero, C. Plaza, N. Senesi, R. Nogales, A. Polo, Humic acid –like fractions inraw and vermicomposted winery and distillery wastes, Geoderma 139 (2007)397–406.

[35] M. Gómez-Brandón, M. Aira, M. Lores, J. Domínguez, Changes in microbialcommunity structure and function during vermicomposting of pig slurry,Bioresour. Technol. 102 (2011) 4171–4178.

[36] K. Parthasarathi, L.S. Ranganathan, Aging effect on enzyme activities inpressmud vermicasts of Lampito mauritii (Kinberg) and Eudrilus eugeniae(Kinberg), Biol. Fertil. Soils 30 (2000) 347–350.

[37] P. Mora, E. Miambi, J.J. Jiménez, T. Decaëns, C. Rouland, Functional complementof biogenic structures produced by earthworms, termites and ants in theneotropical savannas, Soil Biol. Biochem. 37 (2005) 1043–1048.

[38] B. Sen, T.S. Chandra, Do earthworms affect dynamics of functional responseand genetic structure of microbial community in a lab-scale compostingsystem?, Bioresour Technol. 100 (2009) 804–811.

[39] J.P. Curry, O. Schmidt, The feeding ecology of earthworms – a review,Pedobiologia 50 (2007) 463–477.

[40] L. Fracchia, A.B. Dohrmann, M.G. Martinotti, C.C. Tebbe, Bacterial diversity in afinished compost and vermicompost: difference revealed by cultivation-independent analyses of PCR-amplified 16S rRNA genes, Appl. Microbiol.Biotechnol. 71 (2006) 942–952.

[41] M. Danon, I.H. Franke-Whittle, H. Insam, Y. Chen, Y. Hadar, Molecular analysisof bacterial community succession during prolonged compost curing, FEMSMicrobiol. Ecol. 65 (2008) 133–144.