Accepted Manuscript

Investigation on the difference between biofilm morphologies of the vermifilter

and conventional biofilter with the flow cytometer

Wanyin Di, Meiyan Xing, Jian Yang

PII: S0960-8524(16)30673-3
DOI: http://dx.doi.org/10.1016/j.biortech.2016.05.033
Reference: BITE 16529

To appear in: Bioresource Technology

Received Date: 20 March 2016

Revised Date: 9 May 2016
Accepted Date: 11 May 2016

Please cite this article as: Di, W., Xing, M., Yang, J., Investigation on the difference between biofilm morphologies
of the vermifilter and conventional biofilter with the flow cytometer, Bioresource Technology (2016), doi: http://
dx.doi.org/10.1016/j.biortech.2016.05.033

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Investigation on the difference between biofilm morphologies of the vermifilter and

conventional biofilter with the flow cytometer

Wanyin Di, Meiyan Xing*, Jian Yang

Institute of Biofilm Technology, Key Laboratory of Yangtze River Water Environment,

Ministry of Education, College of Environmental Science and Engineering, Tongji

University, Shanghai 200092, China

Abstract

With the demand of new sludge reduction processes, the vermifilter (VF) was studied

based on the conventional biofilter (BF). Biofilm morphology was investigated using a

new technique, the flow cytometer (FCM), to find a way to optimize VF structure. VF

was inoculated with Eisenia fetida, packed with ceramsites, and operated stably at the

organic load of 1.2 kg-VSS m-3 d-1 with BF as the control. Compared with BF, VF had

about 13% more removal efficiency of excess sludge and 45% shorter biofilm update

period. FCM profile showed the morphology of microbial cells in VF biofilms was

significantly different from that in BF in upper layers, with decreases of average

refractive index (about 72%) and size (about 22%), and suggested it was better to keep

earthworm there to remove rod-shaped microorganisms with other filter media in lower

layers to remove spherical ones combining findings in SEM images and extracellular

*
Corresponding author. Tel./fax: +86 21 65984275.

E-mail addresses: xmy7705@163.com

1
polymeric substances.

Keywords: earthworms; biofilm; update period; refractive index; EPS

1. Introduction

In China, the increasing amount of sewage sludge poses serious challenges from

environment protection, economic development and government regulations for

wastewater treatment plants (WWTPs) and municipalities; and increasing attentions

have been paid on improved or novel sludge reduction processes (Jin et al., 2014; Yang

et al., 2015). Conventional technologies for sewage sludge treatment and disposal are

capital intensive, and ecological technologies are considered to be cost-effective and

environment-friendly. As for small towns, the vermifilter (VF) is considered to be a

promising technology for excess sludge reduction (Zhao et al., 2010).

It is well accepted that the inoculation of earthworms Eisenia foetida into the

conventional biofilter (BF) could enhance the removal efficiency of organic matter in

excess sludge (Li et al., 2014; Zhao et al., 2014; Arora and Kazmi, 2015), and the

enhanced removal efficiency of volatile suspended solids (VSS) in liquid-state sludge

was about 13.86% ~ 16.6% at the organic load of 1.10~1.28 kg-VSS m-3 for ceramicite

VF (Xing et al., 2012; Xing et al., 2014; Xu et al., 2014). When the food supply is not

successive, the biomass of E. foetida would get a peak value at first, and then fall. If the

organic matter could be supplied successively, the biomass of E. foetida could become

2
larger and larger. Actually, though the sludge was supplied continuously at a constant

concentration for VF, the biomass of E. foetida stayed almost the same during spring and

summer after an adaptive phase of about two months (Xing et al., 2012). Since that there

were relatively stable sludge reduction levels in VF and BF (Xu et al., 2014), then as

compared with BF, VF had additional reduced sludge, most of which could be utilized as

energy to maintain respiration of earthworms. Accordingly, the respiration rate of

earthworms could be evaluated. When microorganisms in VF live in balance with nature,

biofilm growth could also come into a stable stage, too. When earthworm biomass,

biofilm and VSS reduction level were all stable, the biofilm update period could be

evaluated. However, there have been few reports about the update period of biofilms.

E. foetida are able to break large sludge particles into small fragments, such as 2000

m into 630~2000 m during vermicomposting (Ziegler and Zech, 1992), and 10~200

m into 0~2 m in VF (Zhao et al., 2010); and they could transform insoluble organic

materials to soluble ones (Edwards, 2004).The fragmentation process will put pressure

on some microbial cells harbored in sludge, and leave some rolling and digging marks in

VF biofilms, which could be observed by morphological differences between VF and BF

biofilms (Wang et al., 2015). Actually, viable cells in VF are less than that in BF at the

same depth, opposite to the compared results of dehydrogenase, protease, lipase,

amylase, and glucosidase (Li et al., 2013). If microbial cells in VF grow much better and

could produce higher enzymatic activities than those in BF, they would have better

3
refractive indexes (Liu et al., 2016). A microbial cell injured by some inhibitors will

become wrinkly and shrunken, stick to microbial cell walls badly, and have a lower

refractive index (Zheng et al., 2009; Jiang et al., 2010); and the inhibitors could be

antibacterial peptides in the earthworm mucus (Bilej et al., 2010), Therefore,

morphological characteristics of microbial cells in VF could be used to reflect the

earthworm growth state and feeding taste as well as the biofilm morphology, which

could be evaluated using a new parameter of the refractive index.

The microbial cell refractive index is a key biophysical parameter of microbial cell

morphology, which has been extensively studied in cell biology and disease diagnosis

(Liu et al., 2016). It is correlated to cell biophysical properties like mechanical, electrical

and optical characteristics, and could represent intracellular substances and their

concentrations in a cell. Bacteria are the majority in nature, and they are typically

0.5~5.0 m in length. Considering the too small sizes of bacteria, there would be too

much deviation to analyze them by conventional optical microscope techniques, and the

process is time-consuming and labor-intensive. Thus, a more convenient and accurate

micro technique is required, and the flow cytometer (FCM) technique is recommended

herein, which hasnt been used in previous studies about vermifiltration.

FCM has been used to descript cell morphology since 1960s (Wang et al., 2010).

Because FCM has high-throughput capacity and could maintain the ability for

interrogations at the single-cell level, microbiologists have used it in the environmental

4
microbiology field since 1970s (Bailey et al., 1977), and find it could enumerate more

bacteria than the conventional culture method (Lange et al., 1997). However, because it

is difficult to interpret signals from very small objects, its application in environmental

microbiology has been hitherto rather limited, e.g. bacteria monitoring in drinking water

(Hammes et al., 2010). With the development of various pretreatment techniques

(Falcioni et al., 2006; Brown et al., 2015), it could be possible to enumerate bacteria and

even viruses by FCM for activated sludge samples. Similarly, VF biofilms could also be

analyzed by FCM after proper pretreatment.

Moreover, the feeding behavior of E. fetida could also result in a small number of

microbial cells with a special morphology. E. fetida could ingest some rob-shaped

microorganisms such as Pseudomonas aeruginosa, Bacterium antitratum, Enterobacter

aerogenes, E. cloacae, Proteus vulgaris, P. mirabilis, P. rettgeri, Escherichia coli,

Bacillus subtilis, and B. cereus, and some spherical microorganisms such as Candida

tropicalis, C. krusei C. albicans, and Cryptococcus neoformans (Parthasarathi et al.,

2007); but some spherical microorganisms such as fungal spores and Staphylococcus

aureus could not be ingested during passage through earthworm guts (Edwards, 2004;

Dedeke et al., 2010). Therefore, it is necessary to find the possible shape differences

between microbial cells in VF and BF, and the scanning electron microscope (SEM)

technique is recommended herein. SEM has been applied in previous studies (Zhao et al.,

2010; Wang et al., 2015), but they cannot give any information about shape differences

5
between microbial cells of VF and BF.

It is generally known that zoogloeae are main components of biofilm sludge and

activated sludge, and extracellular polymeric substances (EPS) are important for

zoogloeae growth. Accordingly, the content and structure of EPS could be changed by

the presence of earthworms. EPS have four fractions: (1) slime, (2) loosely bound EPS

(LB-EPS), (3) tightly bound EPS (TB-EPS) and (4) pellet (Yu et al., 2008). And the

ratio of LB-EPS to TB-EPS (LB/TB) is usually used to judge the sludge aggregation

level (Liu et al., 2010). Therefore, morphological change of biofilms could also be

characterized by the change of EPS, which has not been studied in previous reports.

The main aim of present study was to analyze update periods and morphological

characteristics of VF and BF biofilms when both ecosystems were at a stationary phase.

The morphological characteristics of biofilms were analyzed using a new parameter of

microbial cell reflective index as well as the size from FCM, and differences were

further analyzed via SEM images and EPS.

2. Material and methods

2.1. Vermifilter design

Two biofilters were set up with total height of 1.2 m. Each filter consisted of four

columns with 4 ceramsite beds (25 cm in height) and 4 bed intervals (5 cm in height)

alternately. A cone water distributor was on the top of each reactor to keep an even water

6
distribution. The excess sludge was (1) got from a secondary settling tank, (2) diluted in

a conditioning tank, (3) sieved through a 10-mesh sieve to exclude big leaves and

particles which easily blocked pipes, (4) pumped by peristaltic pumps, (5) and sent to

VF evenly at last. After passing through the filter beds, the treated sludge entered a

sedimentation tank, where the supernatant was recycled to the conditioning tank.

Earthworms belonged to E. fetida, and were selected from nearby Wangjun

Earthworm Farm, where they were fed on cow dung. Before earthworms were put into

BF to create VF, it took about a week to change their feeding habitat by increasing the

proportion of excess sludge in feedstuff daily. Then vital adult earthworms were selected

and inoculated at an initial density of 32 g L-1. Their bodies were 4~5 cm in length, and

had fresh red color.

Ceramsites were the filter media. They were brick-red pellets, in shapes of sphere and

ellipse with equivalent spherical diameters in 10~15 cm. Their chemical compositions

were the same to ceramsites of Wang et al. (2015).

Fresh excess sludge was used herein, and it came from Quyang WWTP, Shanghai,

China, where excess sludge was produced from synthetic domestic wastewater. The

sludge was neutral, in liquid state, and about 5000 mg-VSS L-1.

The experiment was conducted in late spring, and last for 2 months. Reactors were

placed in a shed with the room temperature at 15C ~ 25C. Details about related

parameters and treatment performance were listed in Table 1.

7
2.2. Sampling

Each reactor comprised four same beds, designed as Bed 1, 2, 3 and 4 from up to down.

Vi (or Bi; i is 1, 2, 3, or 4) was short for Bed i of VF (or BF). Each bed had sampling

ports with the same design. Biofilm sludge samples were withdrawn randomly for each

bed fortnightly, and rinsed by distilled water to form solutions for each bed and each

filter. At the end of the experiment, biofilm samples were collected randomly from 1

liter fillers at each bed, and stored at 4C after proper pretreatment.

2.3. Physico-chemical analysis

The VSS and SS were determinated according to Standard Methods (APHA, 2005). At

the end of the experiment, Zeta potentials were detected using a Zeta potential meter

(Zetasizer Nano Z, Malvern Instruments Ltd., UK), and SEM images were taken by the

environmental SEM (XL30 ESEM-TMP, Philips EDAX, America).

2.4. FCM analysis

According to the methods of Brown et al. (2015), the sample pretreatment process was

improved. Fresh liquid-state sludge samples were fixed at a final concentration of 0.5%

glutaraldehyde for 25 minutes at 4C in the dark. Then samples were washed by

centrifugation (9000 rpm 4C) with sterile TE-buffer (10 mM Tris-HCl ; 1 mM EDTA;

pH 8.0) for 3 times. The supernatant was substituted with fresh TE. Then surfactants -

8
polyoxyethylene - sorbitan monooleate (Tween 80, Sigma) and onic dispersants -

sodium pyrophosphate (SP, Sigma) were added to destroy sludge flocs at final

concentrations of 5% Tween 80 and 10mM SP, and the mixed solution was kept warm at

37C for 15 min. Next, all samples were washed by centrifugation according to primary

steps. Microbial cells with size less than 50 m were kept by sieving. At last, three 1 mL

dilutions (1/500, 1/5000, 1/10000) were prepared in special tubes. All samples were sent

to the flow cytometry (FACSVerse, BD Biosciences, USA), which was equipped with an

Ar laser (488 nm). And each sample was pumped at 6 m s-1 for about 1.5 min.

2.5. Analysis on EPS

EPS extraction was improved according to several methods (Sheng et al., 2010). There

were 4 continuous steps to extract each fraction of EPS. (1) The 30 mL sludge samples

were centrifuged at 6000 rpm for 15 min; the supernatant was filtered through the

0.45-m mixed cellulose ester membrane (same method to filtration below) and the

slime layer was in the filtrate. (2) The residue was restored to the initial volume with 2

mM buffer (Na3PO4: NaCl: KCl = 2: 4: 9: 1, pH 7.0) of 70C; and the solution was kept

warm at 70C for 2 min, centrifuged (9000 rpm, 4C) for 20 min again; the supernatant

was filtered, and LB layer of EPS was in the filtrate. (3) The residue was restored with 2

mM buffer; and the solution was soften by resin cation (60 g g-1-VSS), centrifuged (6000

rpm, 4C) for 30 min, smashed by sonication at 180 W for 10 min, and centrifuged

9
(12000 rpm, 4C) for 30 min again; the supernatant was filtered, and TB layer of EPS

was in the filtrate. (4) The pellet layer of EPS was in the ultimate residue. Each fraction

of EPS was quantified by the total contents of the protein and polysaccharide, and EPS

were quantified by the total contents of all fractions. The polysaccharide was quantified

by anthrone colorimetry (Doutre et al., 1978), and the protein was quantified by

modified Folin-Lowry method (Frolund et al., 1995).

2.6. Determination on biofilm update periods

In a balanced ecosystem of filters, sludge input could output the products of respiration

in kinds of forms but different to SS, like kinds of gases, soluble matters and energy

substitutes in organism bodies. The difference between sludge reduction levels of VF

and BF has been ascribed to earthworm ingestion and earthworm-microorganism

interaction (Zhao et al., 2010; Li et al., 2014). Therefore, earthworm respiration rate

could be evaluated by Eq (1). Similarly, biofilm respiration rates in VF and BF could

also be inferred, which were shown in Eq (2) and Eq (3) respectively. The update periods

of biofilms in VF and BF could be inferred according to dimensional analysis.

Rw = (EV - EB) * I * Q * p / (10000 Bw) (1)

Rbb = EB * I * Q / (100 Bbb) (2)

Rbv = (EV * I * Q / 100 - Rw * Bw) / Bbv (3)

Ubb = 1 / Rbb (4)

10
Ubv = 1 / Rbv (5)

Where Rw is the earthworm respiration rate, kg-VSS kg-1-earthworm d-1 ; EV and EB are

the sludge removal efficiencies of VF and BF respectively, %; I is the concentration of

the influent excess sludge (IES), kg-VSS m-3; Q is the influent flow,m3 d-1; p is the

proportion responsible for earthworms feeding in the additional reduced sludge between

VF and BF, %, and 21 herein as suggested by Li et al. (2014); Bw is the fresh weight of

total earthworms, kg-earthworm m-3 ; Rbb and Rbv are respiration rates of BF and VF

biofilms respectively, d-1; Bbb and Bbv are biofilm contents in BF and VF respectively,

kg-VSS m-3; Ubb and Ubv are update periods of BF and VF biofilms respectively, d.

2.7. Data analysis

Data were expressed as mean standard deviation on time series analysis at the end of

experiment. And scientific notation was applied in the data with certain significant digits.

All the statistical analysis was done using Excel 2013. Student's t-test (double tails, pair

type) was used to test differences between performance parameters of VF and BF.

Pearson correlation analysis was used to test correlations between data of different

variables. If the probability of t-test is no more than 0.05, the two samples are thought to

be different significantly. Figures were major drawn with Origin Pro 8.5 (OriginLab Co.,

USA) and FlowJo 7.3.1 (FlowJo LLC, Oregon).

11
3. Results and discussion

3.1. Analysis on biofilm contents

According to Table 1, the average sludge removal efficiency of VF was about 13%

higher than that of BF, which was similar to the result of Xing et al. (2014). And the

influence of earthworms has aroused great concern on the difference comparison of

microbial structure (Xu et al., 2014) and biochemical indexes (Li et al., 2013). However,

variation of biofilm contents could express more points, not only about the difference

presentation but also about the process improvement suggestions.

The excess sludge was intercepted when it touched the media, and then degraded

quickly in theory. However, many biofilms grew based on microorganisms in excess

sludge. Therefore, IES would experience a long journey when biofilms still have the

chance to grow and spread in filter media.

Fig. 1(a) shows the decreasing variation trends of biofilm sludge contents (expressed

by SS) with depth at VF and BF respectively, and the average content of biofilm sludge

was less in VF than in BF at the same level. The variation trends in Fig.1(a) indicates

that there is much more interspace for biofilms to grow in the lower layer of VF and BF

respectively, and VF has larger interspace to hold more IES than BF at the same level. In

Fig. 1(a), probabilities of t-test were 0.01, 0.00, 0.01 and 0.01 for 4 pairs of beds at the

same level from up to down respectively, so interspace size occupied by biofilm sludge

12
is influenced significantly by the present of earthworms, and so is the remaining one.

Fig. 1(b) shows the decreasing variation trends of biofilm biomass content (expressed

by VSS) with depth at VF and BF respectively, and the average content of biofilm

biomass was less in VF than in BF at the same level. The variation trends in Fig. 1(b)

indicates that there are more opportunities for existent microorganisms in biofilms to

grow in the lower layer of VF and BF respectively, and VF could provide more

opportunities to attract new biofilm settlers than BF at the same level. Similar to Fig.

1(a), the present of earthworms also influences the biofilm biomass significantly except

the pair of V1 and B1 (the probability of t-test was 0.07.) which could be ascribed to the

unstable thriving of microorganisms in earthworm casts (Edwards, 2004).

When Fig. 1 (a) and (b) were compared, it could be found that VSS was degraded

faster than SS, which supports the feeding habit of earthworms (Edwards, 2004). When

the depth distribution of earthworm biomass were considered (Li et al., 2014), it could

be found that earthworm biomass is greatly influenced by the availability of organic

matter in the supply, like the organic percentage (VSS/SS ratio) in biofilms.

Fig. 1(c) shows organic percentage and depth distributions of VF and BF biofilms.

Organic percentages were 44%, 45%, 48% and 52% for biofilms in Bed 4, 3, 2 and 1 of

VF respectively, and 52%, 55%, 59% and 63% for those of BF respectively. Thus,

organic percentages are 40% ~ 55% and 50% ~ 65% for biofilms of VF and BF

respectively. However, organic percentages are all more than 65% in IES and the

13
effluent of BF and VF according to Table 1. Therefore, the process is inferred to be

instant for sludge interception and degradation, and intercepted sludge is accumulated

little in filters, especially in lower layers. And the average absolute difference between

organic percentages in VF and IES or effluent was larger in VF than that between BF

and IES or effluent, indicating degradation is stronger in VF than in BF, and thus

morphological change is larger in VF than in BF based on IES or effluent.

Fig. 1(d) shows biofilm content differences between BF and VF according to Fig. 1 (a)

and (b), and they were about 2.0, 2.4, 1.4 and 1.2 g-VSS L-1 or 0.7, 2.4, 1.5 and 1.5 g-SS

L-1 at Bed 1, 2, 3 and 4 respectively. There was a decreasing biofilm content difference

from Bed 2 to 4, indicating that difference is lessen with depth, and earthworm

inoculation could exert better effect at an upper layer. Nevertheless, the difference in

Bed 1 was less than in Bed 2, and reasons could be as follows: (1) the influent of Bed 2

was evener than that of Bed 1 by the redistribution of Bed 1, which could bring less

disturbances for earthworms and more reasonable food distribution for biofilms; (2)

there is the largest earthworm biomass in Bed 1 of VF (Li et al., 2014), and they could

produce lots of fresh earthworm casts full of available nutrients and higher microbial

biomass for Bed 2 after the water transportation (Edwards, 2004). As compared with

Bed 3 and 4, VSS differences were larger in Bed 1 and 2, where earthworms also have

the larger biomass (Li et al., 2014). Earthworm inoculation helped prevent biofilm

overgrowth and interspace clogging in each bed of VF, especially for Bed 1 and 2, so it

14
is reasonable that biofilms were updated faster in VF than in BF as shown in Table 2.

Table 2 shows earthworm respiration rate, biofilm update rates and periods of VF and

BF on the whole according to Eqs. (1~5). Biofilm update periods were about 6.4 and

11.6 days for VF and BF respectively. VF had 45% less biofilm update period than BF,

which could be supported by enzymatic activity differences between VF and BF (Xu et

al., 2011). Earthworm respiration rate was larger than Rbv, which is consistent with the

view that earthworm respiration rate is larger than the surrounding (Edwards, 2004).

3.2. Morphological analysis by FCM

3.2.1. Analysis on VF biofilms by FCM

Two parameters are usually used for morphological analysis of microbial cells by FCM,

the FSC-A and SSC-A. FSC-A and SSC-A are used to reflect the relatively sizes and

refractive indexes of microbial cells respectively (Ng, 2004). And special number and

type of cells could be determined by setting gates in FSC-A and SSC-A scatter diagrams.

For same cells, their scatters would gather in a certain range.

Fig. 2 shows the FSC-A and SSC-A distributions of biofilms at different beds of VF.

V1 took on a ray distribution similarly and others took on 2(4) distribution similarly.

When FSC-A was less than 80K, cells had the lower SSC-A values in V1 than in other

any bed of VF, indicating small cells were inhibited more significantly in V1 than in any

other bed of VF. When FSC-A was more than 150K, cells had decreasing SSC-A values

15
in the order of V1, V2, V3 and V4, and the variation trend went up in V1 but down in

other beds, indicating large cells with higher refractive indexes are thriving in V1. And it

may be the predation that results in small cells with lower refractive indexes and large

cells with higher refractive indexes in V1 according to the size relationship between

predators and preys (Yurista et al., 2014).

FSC-A medians, SSC-A means and cell abundance of VF were shown in Table 3. In

Table 3, FSC-A median increased from V1 to V2, and then decreased from V2 to V4; V1

had the lowest FSC-A median in VF. Microbial cells are usually in a state with lower

refractive indexes and smaller sizes at the harmful habitat (Ma et al., 2009). V1 had

lowest FSC-A median and almost lowest SSC-A mean in VF, so it could be speculated

that state of microbial cells get relatively worst in the upper layer of VF. The variation

trend of FSC-A median with depth is consistent with the SS difference distribution in Fig.

2(d), and the Pearson coefficient was 0.91. In Table 3, SSC -A mean decreased slightly

from V1 to V2, and then increased from V2 to V4; V4 had highest SSC -A mean in VF.

The variation trend of SSC-A mean with depth is overall consistent with that of

interspace size (Fig. 2(a)) and inorganic percentage in biofilms (Fig. 2(c)), but opposite

to that of earthworm biomass (Li et al., 2014) and enzyme activities (Xu et al., 2011).

Additionally, Table 3 shows a decreasing cell abundance with depth in VF, which was

consistent with variation trend of viable cell number (Li et al., 2013) and biofilm

biomass (Fig. 2(b)) with depth. The Pearson coefficient was 0.86 between biofilm

16
biomass content and cell abundance of VF, indicating that biofilm biomass could also be

used to reflect the abundance in VF to some extent.

3.2.2. Analysis on BF biofilms by FCM

Fig. 3 shows the depth distribution of microorganisms in BF biofilms. Unlike B2, B3

and B4, B1 had more cells with higher refractive indexes. The similar 2(6) distributions

could be found in B2, B3 and B4. When FSC-A was less than 80K or more than 150K,

average SSC-A decreased with depth on the whole, indicating small and large cells could

survive better in an upper layer. When SSC-A was more than 50K, cell number and

diversity of cell morphologies decreased with depth from B1 to B3, and slightly

increased from B3 to B4, indicating filter media could provide a relatively healthy

habitat for biofilms if air exchange is strong enough therein.

FSC-A medians, SSC-A means and cell abundance of BF were shown in Table 3.

FSC-A median increased from B1 to B2 slightly, and then decreased from B3 to B4. B4

had the lowest FSC-A median in BF. SSC-A mean decreased with depth in BF. B4 had

lowest FSC-A median and SSC-A mean in BF, so it could be speculated that the state of

microbial cells are relatively worse in the lower layer of BF. B1 had almost highest

FSC-A median and highest SSC-A mean in BF, so it could be speculated that the state of

microbial cells are relatively better in the upper layer of BF. Moreover, similar to VF, BF

had the diminishing cell abundance with depth as shown in Table 3. The Pearson

coefficient was 0.95 between biofilm biomass content and cell abundance for BF, a

17
stronger correlation between biomass and abundance than that of VF.

3.2.3. Analysis on VF and BF biofilms by FCM

According to Fig. 2 and Fig. 3, there is the most difference between cell morphologies of

V1 and B1, which was shown in Fig. 4 (a~c). The SSC-A and FSC-A distributions of V1

and B1 could be distinguished in Fig. 4(a), where V1 had a narrow distribution band that

almost be covered by B1 distribution, indicating less types of cell states in V1. B1 has

more small microbial cells (< 50K) than V1 (Fig. 4(b)) which could confirm the

fragmentation function of earthworms (Zhao et al., 2010). Most microbial cells had low

refractive indexes both in V1 and B1, and B1 had the higher cell count at any SSC-A

(Fig. 4(c)), indicating that the presence of earthworms could reduce microbial cells with

various refractive indexes. Taken together, it could be found that at the presence of

earthworms, small cells are reduced greatly and significantly, whereas large microbial

cells are increased significantly but in fewer cell states.

According to Table 3, there were differences between SSC-A means and FSC-A

medians of VF and BF. The variation trends of SSC-A mean and FSC-A median were

both similar in VF and BF from Bed 2 to 4. There was the largest absolute difference

between SSC-A means or FSC-A medians of B1 and V1. Compared with BF, VF had

smaller SSC-A means in Bed 1 and 2, and larger ones in Bed 3 and 4, indicating there is

a turning point at depth of 50 cm for the qualitative change of external stimulation on

biofilms by the presence of earthworms. Compared with VF, BF had larger FSC-A

18
medians in Bed 1, and smaller ones from Bed 2 to 4, indicating that when earthworms

are inoculated in BF, there are phenomena about particle fragmentation and cell

ingestion at Bed 1, and about slight particle aggregation and cell growth in following

beds. Taken together, there is a relatively great change for microbial habitat in upper

layers by the presence of earthworms.

3.3. Analysis on shapes of microbial cells

In order to investigate the microbial habitat morphology, mixed biofilm samples of Bed

1 and 2 were analyze by SEM. According to SEM images, there were more rod-shaped

microbial cells than spherical ones in BF than in VF. Many rod-shaped microorganisms

were 15~45 m in length and 3~7 m in width and many spherical ones were 5~15 m

in diameter.

There are many rod-shaped microbial cells in excess sludge, and E. fetida could also

ingest many rod-shaped microorganisms (Parthasarathi et al., 2007; Edwards, 2004). So

it is possible for rod-shaped microbial cells to be reduced greatly in VF. According to

SEM images, small rod-shaped microbial cells are inferred to be reduced more than the

large ones by the presence of earthworms.

Furthermore, more compact structure could be found in VF than in BF according to

SEM images, which could verity the activity of earthworms and was not conflicted with

the conclusion that VF had smoother biofilms than BF (Wang et al., 2015).

19
3.4. Analysis on EPS of biofilms

As an important part of zoogloeae in EPS, EPS could protect microorganisms from

direct exposure to earthworms and change the biofilm surface property. Actually, there

was the compact structure in VF and loosen one in BF through SEM analysis. Thus EPS

was influenced directly under external stimuli with biofilm surface change. Table 4

shows the contents of EPS at different fractions for BF and VF biofilms. The contents of

EPS were about 102.23 1.36 mg g-1-VSS and 141.47 1.27 mg g-1-VSS for VF and

BF biofilm samples respectively during the experimental phase, which indicates that

microorganisms in VF may be stimulated by the presence of earthworms, and result in

the higher update rate of biofilms in VF than in BF (Table 2).

According to Table 4, parts of the slime, LB, TB and pellet accounted for about 0.09%,

10.16%, 30.00% and 59.57% respectively in BF, and about 0.05%, 9.63%, 39.00% and

51.31% respectively in VF. Thus there are differences about structure and content of EPS

between VF and BF. As for the EPS extracted from sludge flocs, the LB/TB ratio plays

an important role on the flocs absorption (Sheng et al., 2010).The LB/TB ratios were

about 0.33 and 0.25 in VF and BF respectively, which indicates that zoogloeae in VF

may be destroyed by earthworms. And a lower LB/TB ratio indicates a stronger

absorption of sludge flocs and a larger change about the Zeta potential of influent sludge

flocs (Andreadakis, 1993). The Zeta potentials were about -15.32 mV, -12.77 mV and

20
-7.00 mV for IES, BF and VF effluent respectively. Both changes about the Zeta

potential and the LB/TB ratio could testify the peeling probability of many zoogloeae

under the influence of earthworm inoculation.

4. Conclusions

Morphological differences between BF and VF biofilms were analyzed through

parameters of the VSS removal efficiency, biofilm content, earthworm respiration rate,

biofilm update period, microbial size (FSC-A), microbial refractive index (SSC-A),

microbial cell shape, EPS content, LB/TB ratio, and Zeta potential. And the FCM profile

was first applied in present vermifiltration studies. Results suggested that earthworms

could live better at an upper layer of about 50 cm in thickness, and it was better to keep

earthworm there to remove small microorganisms in rod shape with other filter media in

the lower layer to remove other microorganisms, especially the spherical ones.

Acknowledgement

The research was supported by the National Natural Science Foundation of China (No.

51109161). Authors acknowledge the valuable suggestions made by Professor Dr.

Yongsen Lu during the paper writing.

References

1. Andreadakis, A. D., 1993. Physical and chemical properties of activated sludge floc.

21
Water Res., 27: 1707-1714.

2. APHA, AWWA, WEF, 2005. Standard methods for the examination of water and

wastewater, twenty-first ed. American Public Health Association, Washington DC.

3. Arora, S., Kazmi, A. A., 2015. The effect of seasonal temperature on pathogen

removal efficacy of vermifilter for wastewater treatment. Water Res., 74: 88-99.

4. Bailey, J. E., Fazel-Makjlessi J., Mcquitty D. N., Lee Y. N., Allred J. C., Oro J. A.,

1977. Characterization of bacterial growth by means of flow microfluorometry. Science,

198: 1175-1176.

5. Bilej, M., Prochzkov, P., ilerov, M., Joskov, R., 2010. Earthworm immunity.

Springer US.

6. Brown, M. R., Camzuli S., Davenport R. J., Petelenz-Kurdziel E., vres L., Curtis

T.P., 2015. Flow cytometric quantification of viruses in activated sludge. Water Res., 68:

414-422.

7. Dedeke, G. A., Omemu O., Aladesida A. A., Museliu F., 2010. Comparative

Microbial Analysis of Earthworm Casts Collected From Ikenne, Ogun State, Nigeria.

Eth. J. Environ. Stud. Manage., 3(3): 57-63.

8. Doutre, D. A., Hay G. W., Hood A., Vanloon G. W., 1978. Spectrophotometric

methods to determine carbohydrates in soil. Soil Biol. Biochem., 10(6): 457-462.

9. Edwards, C. A., 2004. Earthworm ecology, 2nd edition. CRC Press, Boca Raton,

American.

22
10. Falcioni, T., Manti A., Boi P., Canonico B., Balsamo M., Papa S., 2006. Comparison

of disruption procedures for enumeration of activated sludge floc bacteria by flow

cytometry. Cytom. Part B-Clin. Cy., 70(3): 149-153.

11. Frolund, B., Griebe T., Nielsen P. H., 1995. Enzymatic activity in the activated

sludge floc matrix. Appl. Microbiol. Biotechnol., 43(4): 755-761.

12. Hammes, F., Goldschmidt F., Vital M., Wang Y., Egli T., 2010. Measurement and

interpretation of microbial adenosine tri-phosphate (ATP) in aquatic environments.

Water Res., 44(13): 3915-3923.

13. Jiang, Y., Wang Y., Su Z., Yang L., Guo W., Liu W., Zuo J., 2010. Synergistic

induction of apoptosis in HeLa cells by the proteasome inhibitor bortezomib and histone

deacetylase inhibitor SAHA. Mol. Med. Rep., 2010, 3(4): 613-619.

14. Jin, L., Zhang G., Tian H., 2014. Current state of sewage treatment in China. Water

Res, 66: 85-98.

15. Lange, J. L., Thorne P. S., Lynch N., 1997. Application of flow cytometry and

fluorescent in situ hybridization for assessment of exposures to airborne bacteria. Appl.

Environ. Microbiol., 63(4): 1557-1563.

16. Lautenschlager, K., Boon N., Wang Y., Egli T., Hammes F., 2010. Overnight

stagnation of drinking water in household taps induces microbial growth and changes in

community composition. Water Res., 44(17): 4868-4877.

17. Li, X. W., Xing M. Y., Yang J., Dai X. H., 2014. Earthworm eco-physiological

23
characteristics and quantification of earthworm feeding in vermifiltration system for

sewage sludge stabilization using stable isotopic natural abundance. J. Hazard. Mater.,

276: 353-361.

18. Li, X. W., Xing M. Y., Yang J., Lu Y. S., 2013. Properties of biofilm in a

vermifiltration system for domestic wastewater sludge stabilization. Chem. Eng., 223:

932-943.

19. Liu, P. Y., Chin L. K., Ser W., Chen H. F., Hsieh C. M., Lee C. H., Sung K. B., Ayi

T. C., Yap P. H., Liedberg B., Wang K., Bourouina T., Leprinc-Wang Y.. 2016. Cell

refractive index for cell biology and disease diagnosis: past, present and future. Lab

Chip, 16: 634-644.

20. Liu, X. M., Sheng G. P., Luo H. W., Zhang F., Yuan S. J., Xu J., Zeng R. J., Wu J.

G., Yu H. Q., 2010. Contribution of extracellular polymeric substances (EPS) to the

sludge aggregation. Environ. Sci. Technol., 44(11): 4355-4360.

21. Ma, J., Wang K., Xing H., 2009. Influences of cyclin dependent kinase on

proliferation and apoptosis in hepatic carcinoma cells. New Medicine, 40(5): 288-291.

22. Ng, V. L., 2004. Practical flow cytometry, 4th Edition. Shock, 21(1).

23. Parthasarathi, K., Ranganathan L.S., Anandi V., Zeyer J., 2007. Diversity of

microflora in the gut and casts of tropical composting earthworms reared on different

substrates. J Environ. Biol., 28(1):87-97.

24. Sheng, G., Yu, H., Li, X., 2010. Extracellular polymeric substances (EPS) of

24
microbial aggregates in biological wastewater treatment systems: a review. Biotechnol.

Adv., 28: 882-894.

25. Wang, Y., Hammes F., Roy K. D., Verstraete W., Boon N., 2010. Past, present and

future applications of flow cytometry in aquatic microbiology. Trends Biotechnol.,

228(8): 416-424.

26. Wang, Y., Xing M., Yang, J., Lu, B., 2015. Addressing the role of earthworms in

treating domestic wastewater by analyzing biofilm modification through chemical and

spectroscopic methods. Environ. Sci. Pollut. Res., 18(6): 14760-14772.

27. Xing, M., Zhao, C., Yang J., Lv, B., 2014. Feeding behavior and trophic relationship

of earthworms and other predators in vermifiltration system for liquid-state sludge

stabilization using fatty acid profiles. Bioresour.Technol. 169: 149-154.

28. Xing, M.Y., Li, X. W., Yang, J., Lv, B.Y., Lu, Y. S., 2012. Performance and

mechanism of vermifiltration system for liquid-state sewage sludge treatment using

molecular and stable isotopic techniques. Chem. Eng., 197: 143-150.

29. Xu, T., Xing, M. Y., Yang, J., Lv, B. Y., Duan, T., Nie, J., 2014. Tracking the

composition and dominant components of the microbial community via polymerase

chain reaction-denaturing gradient gel electrophoresis and fluorescence in situ

hybridization during vermiconversion for liquid-state excess sludge stabilization.

Bioresour. Technol., 167: 100-107.

30. Xu, Z., Xing, M. Y., Yu, F., Yan, Q., Yang, J., 2011. Investigation on Biofilm

25
Activity in Vermifilter with Filter Media of Different Diameter. Energy Procedia,

4196-4204.

31. Yang, G., Zhang, G., Wang, H., 2015. Current state of sludge production,

management, treatment and disposal in China. Water Res., 78: 60-73.

32. Yu, G. H., He, P. J., Shao, L. M., He P. P., 2008. Toward understanding the

mechanism of improving the production of volatile fatty acids from activated sludge at

pH 10.0. Water Res., 42(18): 4637-4644.

33. Yurista, P. M., Yule D. L., Balge M., Vanalstine J. D., Thompson J. A., Gamble A.

E., Hrabik T. R., Kelly J. R., Stockwell J. D., Vinson M. R., 2014. A new look at the

Lake Superior biomass size spectrum. Can. J. Fish. Aquat. Sci., 71: 1324-1333.

33. Zhao, C. H., Xing, M.Y., Yang, J., Lu, Y. S., Lv, B.Y., 2014. Microbial community

structure and metabolic property of biofilms in vermifiltration for liquid-state sludge

stabilization using PLFA. Bioresour. Technol., 151: 340-346.

34. Zhao, L., Wang, Y., Yang, J., Xing, M., Li, X., Yi, D., Deng, D., 2010.

Earthworm-microorganism interactions: A strategy to stabilize domestic wastewater

sludge. Water Res., 44: 2572-2582.

35. Ziegler, F., Zech, W., 1992. Chemical changes of beech litter during passage through

the gut of the lumbricid earthworm Eisenia fetida (SAV.). Soil Sci. Plant Nutr., 155:

69-70.

26
Tables

Table 1. Parameters about design, operation and treatment effect of VF and BF

Item Parameter Value

VF BF
Filter media Height (m) 1.00 1.00
Diameter (cm) 20.0 20.0
Effective size d10 (mm) 15.4 15.4
Porosity (%) 49 49
Working Volume (L) 31.4 31.4
E. fetida Initial density1 (g L-1) 32 /
Stable density1 (g L-1) 26.0 0.5 /
Excess sludge pH 7.1 0.3 7.1 0.3
VSS (mg L-1) 5000 80 5000 80
Water content (%) 99.4 0.3 99.4 0.3
Control factors Hydraulic load (m d-1) 4.0 0.1 4.0 0.1
Influent organic matter2 (mg L-1) 298 11 298 11
Organic load (kg-VSS m-3 d-1) 1.2 0.1 1.2 0.1
Treatment effect VSS removal efficiency3 (%) 54.1 1.0 40.9 2.4
VSS/SS in IES4 0.75 0.05 0.75 0.05
VSS/SS in the effluent4 0.66 0.03 0.71 0.04

Note: 1 based on fresh weight; 2 based on the volatile suspended solid (VSS); 3 the

probability of t-test is 0.00 ( 0.05); 4 IES is short for influent excess sludge, and SS

is short for suspended solid.

27
Table 2. Parameters of biofilm update period calculation

Parameter Value Parameter Value

Bw (kg) 816.4 Rb (d-1) 0.09

Bbb (kg) 179.0 Rbv (d-1) 0.16

Vbb (kg) 124.0 Ubb (d) 11.6

Rw (kg kg-1- earthworm d-1) 1.28 Ubv (d) 6.4

Note: weight is balanced by VSS except the fresh weight of earthworms.

28
Table 3. SSC-A means, FSC-A medians and abundances of microbial cells at different

beds of VF and BF.

FSC-A median Abundance

Bed No. SSC-A mean
(100) (cell L-1)

V1 2.13E+03 2.68E+02 1.14E+10

V2 2.11E+03 3.91E+02 1.00E+10

V3 2.68E+03 3.49E+02 8.73E+09

V4 3.52E+03 2.97E+02 5.31E+09

B1 7.66E+03 3.42E+02 1.28E+10

B2 2.97E+03 3.58E+02 1.14E+10

B3 1.57E+03 3.34E+02 4.62E+09

B4 1.72E+03 2.97E+02 5.86E+09

29
Table 4. Fraction contents of EPS in BF and VF biofilms

Absolute content (mg g-1-VSS) Relative content (%)

Fraction
BF VF BF VF

Slime 0.09 0.01 0.08 0.00 00 00

LB 10.16 0.17 13.63 0.28 10 0 10 0

TB 30.67 0.41 55.17 2.43 30 0 39 2

Pellet 61.31 0.84 72.59 2.52 60 0 51 2

Total content 102.23 1.36 141.47 1.27 100 0 100 0

Note: EPS is short for extracellular polymeric substances; LB denotes the loosely bound

layer of EPS; TB denotes the tightly bound layer of EPS.

30
Figure Captions

Fig. 1. (a) Biofilm sludge contents (expressed by SS) at VF and BF beds. (b) Biofilm

biomass content (expressed by VSS) at VF and BF beds. (c) Organic percentages

(VSS/SS ratio) in biofilms of VF and BF beds. (d) The content differences between VF

and BF biofilms at different beds.

Fig. 2. Morphological distributions of VF biofilms at different beds by FCM analysis

Fig. 3. Morphological distributions of BF biofilms at different beds by FCM analysis

Fig. 4. (a) SSC-A and FCS-A distribution of V1 and B1. (b) Count and FSC-A

distribution of V1 and V2. (c) Count and SSC-A distribution of V1 and V2.

31
Fig. 1. (a) Biofilm sludge contents (expressed by SS) at VF and BF beds. (b) Biofilm

biomass contents (expressed by VSS) at VF and BF beds. (c) Organic percentages

(VSS/SS ratio) in biofilms of VF and BF beds. (d) The content differences between VF

and BF biofilms at different beds.

32
Fig. 2. Morphological distributions of VF biofilms at different beds by FCM analysis

33
Fig. 3. Morphological distributions of BF biofilms at different beds by FCM analysis

34
Fig. 4. (a) SSC-A and FCS-A distribution of V1 and B1. (b) Count and FSC-A

distribution of V1 and V2. (c) Count and SSC-A distribution of V1 and V2.

35
Highlights:

The update period of biofilms was shortened by the presence of earthworms.

Biofilms were negatively simulated in upper beds of VF.

Biofilms were positively simulated in lower beds of VF.

Rod-shaped and spherical microbial cells stood out in BF and VF respectively.

The LB/TB ratio in EPS decreased when earthworms were present.

36