ORIGINAL RESEARCH PAPER

Bioaugmentation of a sequencing batch biofilm reactorwith Comamonas testosteroni and Bacillus cereus and theirimpact on reactor bacterial communities

Zhongqin Cheng • Mei Chen • Liqun Xie •

Lin Peng • Maohua Yang • Mengying Li

Received: 8 June 2014 / Accepted: 5 September 2014

� Springer Science+Business Media Dordrecht 2014

Abstract The immobilization of microorganisms is

essential for efficient bioaugmentation systems. The

performance of Bacillus cereus G5 as biofilm-forming

bacteria and Comamonas testosteroni A3 a 3,5 dini-

trobenzoic acid (DNB)-degrading strain] in labora-

tory-scale sequencing batch biofilm reactors (SBBRs)

treating DNB synthetic wastewater has been exam-

ined. The microbial diversity in the reactors was also

explored. The reactor R3 inoculated with B. cereus G5

and C. testosteroni A3 together not only improved the

removal of contaminants, but also exhibited obvious

resistance to shock loading with DNB during later

operations. Pyrosequencing was used to evaluate

bacterial communities in three reactors. Comamonas

was predominant in the reactor R3, indicating the

effect of G5 in promoting immobilization of A3 cells

in biofilms. Those microbial resources, e.g.G5, which

can stimulate the self-immobilization of the degrading

bacteria offer a novel strategy for immobilization of

degraders in bioaugmentation systems and show

broader application prospects.

Keywords Bacillus cereus �Bacterial communities �Bioaugmentation � Biofilm-forming bacteria �Comamonas testosteroni � 3,5-Dinitrobenzoic acid �Pyrosequencing � Sequencing batch biofilm reactors

Introduction

Bioaugmentation, which introduces new metabolic

functions by the addition of bacteria or genetic

information into wastewater treatment systems, is a

possible way to accelerate the removal of undesired

compounds (Limbergen et al. 1998; Boon et al. 2003).

A major obstacle to efficient bioaugmentation is the

often insufficient establishment of the desired func-

tions within the microbial community because of the

wash-out of inoculated microbes, which limits the

applications and developments of bioaugmentation

(Park et al. 2009). The immobilization of microor-

ganisms has therefore been proposed as a novel

strategy for preventing wash-out of the degraders in

wastewater treatment systems. Bacterial biofilms can

act as a natural medium for this kind of immobilization

process (Lawniczak et al. 2011; Abe et al. 2013).

Bacterial biofilms are formed when unicellular organ-

isms come together to form a community that is

attached to a solid surface and are encased in an

exopolysaccharide matrix (Mah and O’Toole 2001).

In multiple bacterial species biofilms, microorganisms

can form intimate or special relationships with other

Zhongqin Cheng and Mei Chen have contributed equally to

this paper.

Z. Cheng � M. Chen � L. Xie � L. Peng �M. Yang � M. Li (&)

School of Biology and Basic Medical Sciences, Soochow

University, Suzhou 215123, People’s Republic of China

e-mail: lmysuda@163.com

123

Biotechnol Lett

DOI 10.1007/s10529-014-1684-1

microbial species by the gene expression exchanges to

a certain extent, and also show stronger adaptability

and stability in wastewater treatment systems.

One of main factors for multi-species biofilms

development is biofilm-forming bacteria that can co-

aggregate with bacteria from different genera. Mix-

tures of appropriate biofilm-forming strains with

degrading species will help the degraders to immobi-

lize on biofilms and, consequently, promote degrada-

tion (Gioia et al. 2004; Duque et al. 2011; Li et al.

2013). The earliest and extensive studies about

biofilm-forming bacteria mainly focused on oral

biofilms or dental plaques. More than 1,000 bacterial

strains from the oral cavity were demonstrated to have

co-aggregation ability (Kolenbrander 2000). Biofilm

microflora present in both natural and wastewater

environments were explored, and several new strains

with a bridging function have been found; for exam-

ple, Micrococcus luteus 2.13 (Rickard et al. 2004),

Acinetobacter johnsonii S35 (Malik et al. 2003a),

Acinetobacter cacoaceticus (Simoes et al. 2008) and

Bacillus cereus G5 (Cheng et al. 2014).

Bacillus cereus G5, which mediates the integration

of an exogenous 3,5-DNBA (3,5-dinitrobenzoic acid)-

degrading bacterium, Comamonas testosteroni A3 (Li

et al. 2008), into biofilms and contributes to increasing

the biofilm biomass, has considerable potential in

bioaugmentation applications (Cheng et al. 2014). The

aim of this study was to exploit the performance of

B. cereus G5 in a laboratory-scale sequencing batch

biofilm reactor (SBBR) treating synthetic wastewater

containing 3,5-DNBA as a model xenobiotic com-

pound. At the same time, the microbial diversity in all

of the reactors was explored.

Materials and methods

Isolation of indigenous bacteria from activated

sludge

Activated sludge samples were collected from a

municipal sewage treatment plant in Suzhou, China.

The samples were vortexed in sterile water, and 200 ll

was spread on petri plates containing LB agar

medium. The plates were incubated at 30 �C for

72 h and the bacterial colonies were picked and

purified based on colony morphology. All isolates

were selected and stored at -80 �C for future use.

Preparation of bacterial suspensions

Bacteria were first cultured on solid LB medium at

30 �C for 18-24 h then inoculated into flasks (250 ml)

containing 100 ml liquid LB medium and shaken at

180 rpm at 30 �C for 24 h. Cells were harvested by

centrifugation at 5,0009g for 5 min, washed twice in

PBS buffer (0.1 M, pH 7), and resuspended in the

same buffer. Finally, the OD660 of the cell suspension

was adjusted to 1 for future use.

Co-aggregation assay

Co-aggregation reactions were estimated by a modi-

fied method described by Malik et al. (2003b). Cell

suspensions (OD660 = 1) of bacterial partners were

prepared as described above. Equal volumes (5 ml) of

co-aggregating partner suspensions were mixed in

15 ml centrifuge tubes and incubated at 25 �C. Pure

suspensions of single bacterial strains (10 ml) were

incubated under similar conditions as controls. Sam-

ples were cultured for 2 and 20 h, subsequently

centrifuged at 5009g for 1 min, and then the OD660

of the supernatant was measured. The co-aggregation

index representing the co-aggregation rate (%), was

calculated as follows:

ðOD660x þ OD660yÞ=2� OD660ðxþyÞðOD660x þ OD660yÞ=2

� 100 %

where OD660x and OD660y represent the OD660 value

for each of the individual pure-cultures and OD660

(x ? y) represents theOD660 value for the mixtures

(Buswell et al. 1997).

Quantification of biofilm biomass

Tubes (10 9 75 mm) containing 2 ml 5 % (v/v) LB

medium were inoculated with 2 ll previously pre-

pared bacterial cell suspensions (half volume of each

in the case of two bacterial strains), and the liquid

cultures were incubated at 30 �C with shaking at

150 rpm for 20 h. Quantification of biofilms was

performed using a Crystal Violet stain described by

Zhu and Mekalanos (2003). Briefly, after 20 h incu-

bation, the medium containing suspended cells was

removed and the tubes rinsed twice with distilled

water, and the remaining attached biomasses were

stained for 30 min with 50 ml 0.1 % (w/v) Crystal

Biotechnol Lett

123

Violet in water. The tubes were washed thoroughly

with water and dried overnight. The retained Crystal

Violet was dissolved in 10 ml ethanol/acetone (4:1

v/v), and absorbance was measured at 570 nm.

Sequencing batch biofilm reactors (SBBRs)

Three Plexiglas SBBRs (22 cm 9 14 cm 9 16 cm)

with 5 l working volume were set up. The carrier

material used in this experiment was sphere-like,

porous and 10 cm internal diam., in which strips of

PVC (polyvinyl chloride) films (0.3 mm thick) were

filled, accounting for about 30 % of the effective

volume of the reactors. The activated sludge used for

the inoculation was collected from the same municipal

sewage treatment plant mentioned in materials and

methods. These three reactors were fed with equiva-

lent 3,5-DNBA synthetic wastewater (100, 200, 300,

400 and 500 mg l-1 successively), and the carriers

were always kept in the reactors. Then, the reactor

NO.1 (R1) was seeded with 200 mg activated sludg e

l-1; the reactor NO.2 (R2) was seeded with 150 mg

activated sludge l-1 and 50 mg 3,5-DNBA-degrading

bacterium Comamonas testosteroni A3 l-1; and the

reactor NO.3 (R3) was seeded with 100 mg activated

sludge l-1, 50 mg C. testosteroni A3 l-1, 50 mg B.

cereus G5 l-1. In order to develop sufficient biofilms

on the carrier surfaces, these reactors were aerated for

24 h after inoculation. Subsequently, half of the

synthetic wastewater in the reactors was changed

every 12 h, which was done six times.

After this start-up stage, the medium in the reactors

was emptied and fresh synthetic wastewater was added

every 24 h. Subsequently, the reactors operated on a

24 h cycle consisting of the following stages: a 30 min

filling phase, a 22 h aeration period, a 1 h settling

phase, and a 30 min drawing phase. The reactors ran at

26 �C (±2 �C) and the hydraulic retention time (HRT)

was maintained for 24 h. To avoid dissolved O2 (DO)

limitation, the DO concentration was maintained

higher than 3 mg l-1 during aeration. The SBBRs

were run continuously for approx. 33 days in total.

Effluent samples analysing

Effluent samples were collected during the entire

operation at pre-defined time intervals from the top of

the reactors and were centrifuged for 4 min at

50009g. 3,5-DNBA concentrations were determined

by the absorbance at 241 nm (Li et al. 2008). All

analytical values were reported as an average of three

independent measurements.

Pyrosequencing analysis of microbial diversity

Genomic DNA was extracted from approx. 1 ml

biofilm sample using a soil DNA kit (Omega Bio-Tek

Inc. USA) following the manufacturer’s instructions.

To analyze the taxonomic composition of bacterial

community, the universal 16S rRNA gene primers (F

50-AACGCGAAGAACCTTAC-30 and R 50-CGGTGT

GTACAAGACCC-30) were chosen for the amplifica-

tion and subsequent pyrosequencing of the PCR

products. The PCR was carried out in triplicate 25 ll

reactions with 2.5 ll 10-fold reaction buffer, 2 ll

template DNA, 0.4 lM forward and reverse primers,

0.625 U Takara Pyrobest polymerase and 0.2 mM

dNTPs. Certified DNA-free PCR water was added to

make 25 ll. The amplification program was: initial

denaturation at 94 �C for 4 min, followed by 25 cycles,

where 1 cycle consisted of 94 �C for 30 s (denatur-

ation), 53.5 �C for 30 s (annealing) and 72 �C for 30 s

(extension), and a final extension of 72 �C for 7 min.

PCR products were pooled and visualized on agarose

gels (1 % in TBE buffer) containing ethidium bromide,

and purified with a DNA gel extraction kit (Axygen,

China). Amplicon pyrosequencing was performed

using a 454 Life Sciences Genome Sequencer FLX

Titanium instrument (Roche, NJ, USA) by Personalbio

Technology Co., Ltd., Shanghai, China.

Results and discussion

Bacillus cereus G5 co-aggregates with new

isolates from activated sludge and promotes

biofilm formation

To ensure biofilm development with Bacillus cereus

G5 in the reactors, the co-aggregation reaction and

biofilm-forming reaction between G5 and thirteen new

isolates (L2, L7, L7R, L8, L15, L25, L31, L47, L48,

L51, L61, LF3 and LH1) from activated sludge were

exploited. First, the abilities of these new isolates to

degrade 3,5-DNBA were analyzed and all showed

negative results (results not shown). Of all 13 pairwise

combinations, eight pairs showed a[50 % co-aggre-

gation rate after a 2 h co-cultivation, and nine pairs

Biotechnol Lett

123

exhibited a[60 % co-aggregation rate after a 20 h co-

cultivation period, indicating that B. cereus G5 co-

aggregated with these isolates and that the co-aggre-

gating ability increased with time (Fig. 1).

Biofilm biomass from co-aggregating mixtures of

B. cereus G5 with each of these thirteen strains, along

with their monocultures, was quantified in 5 % LB

(, simulating the general wastewater nutritional condi-

tion). The results showed that co-aggregating pairs of B.

cereus G5 with 12 of all 13 strains produced greater

biofilm biomass than any other monoculture (Fig. 2).

The t test results demonstrated that the calculated

P value was less than 0.01 when comparing the values

from B. cereus G5 co-cultured mixtures with those

from monocultures, indicating that the co-cultured

biofilm biomass increased significantly at this low

nutritional condition.

B. cereus G5 could not only co-aggregate with

these isolates from the wastewater treatment system

but also promote biofilm formation, thereby indicating

its usefulness for bioaugmentation.

3,5-DNBA degradation performance of SBBRs

inoculated with B. cereus G5

To test the bridging ability of B. cereus G5 in SBBRs

and evaluate its function in bioaugmentation, three

parallel 3,5-DNBA-degrading reactors were set up

according to procedure given in the previous section.

The performance of the three reactors was monitored

for 33 days and data were presented in Fig. 3.

During the initial biofilm-forming period

(0–4 days), activated sludge and inoculated bacteria

quickly attached to strips and formed a visible biofilm

at the surfaces in all reactors. 3,5-DNBA in R2 and R3

decreased to 13.1 and 12.4 mg l-1, respectively, on

the first day, while 3,5-DNBA in R1 was still

87.5 mg l-1. On the third day the concentration of

3,5-DNBA in R1 also decreased (21.7 mg l-1). Thus

R2 and R3 could remove 3,5-DNBA quickly because

of the inoculated degrader B. cereus A3, while the

degradation reaction in R1 was probably attributable

to bacteria from the activated sludge.

After this biofilm-forming period, the reactors were

emptied on the 4th day and 3,5-DNBA degradation

went ahead by the biofilms attached on the carrier

Fig. 1 Co-aggregation indexes of co-aggregating pairs of

B. cereus G5 with thirteen new isolates from raw activated

sludge at 2 and 20 h. The bacteria were incubated at 20 �C for or

20 h. The error bars represent ± SD of the assay performed in

triplicate. Thirteen new isolates were L2, L7, L7R, L8, L15,

L25, L31, L47, L48, L51, L61, LF3 and LH1

Fig. 2 Biofilm biomass

comparison of co-

aggregating pairs of

B. cereus G5 with different

strains. The bacteria were

incubated at 30 �C at

1509g for 20 h in 5 % LB.

The error bars

represent ± SD of the assay

performed in triplicate.

Thirteen new isolates were

L2, L7, L7R, L8, L15, L25,

L31, L47, L48, L51, L61,

LF3 and LH1

Biotechnol Lett

123

materials. On the 5th day, 3,5-DNBA in R1, R2 and R3

was 15.3, 8 and 6.1 mg l-1 respectively, which

showed that biofilms in three reactors were well-

developed and the degradation rates in R2 and R3 were

a little more than that in the control reactor R1.

Keeping 3,5-DNBA at a low influent concentration

(100 or 200 mg l-1) for 12 days, the effluent 3,5-

DNBA concentrations from three reactors were rela-

tively stable and the removal rates would reach more

than 90 %, showing no obvious differences among

three reactors.

On the 17th day, shock-loading with 300 mg 3,5-

DNBA l-1 was applied to these reactors and conse-

quently the effluent concentrations from R1 and R2

increased to 149.8 and 147.5 mg l-1, respectively.

With this shock loading, the effluent 3,5-DNBA

concentration reduced gradually in the following days,

and tended towards stability with the value of

*10 mg l-1. Average 3,5-DNBA removal rates were

*70 % in R1 and *72 % in R2, respectively, which

were lower than those in the first two weeks, showing

an acute toxic effect of 3,5-DNBA at higher concen-

trations. In contrast, the corresponding concentration

of 3,5-DNBA in R3 was 84.8 mg l-1, and the average

3,5-DNBA removal rate was 84 %, which suggested

that R3 exhibited resistance to 3,5-DNBA loading

shock. Overall, with the influent 3,5-DNBA concen-

tration increasing (400 and 500 mg l-1) during later

operations, similar change patterns on 3,5-DNBA

concentration and the removal rate appeared. The

average removal rates of R1 and R2 were similar

(29–33 days; 75.9 % in R1 and 76.7 % in R2),

suggesting enough microbes from the raw activated

sludge capable of degrading 3,5-DNBA might have

emerged in R1. In terms of R3, the average removal

rate was still higher than 80 % during later operations

(23–28 days, 81.5 %; 29–33 days, 83 %). The t-test

results demonstrated that the P value were both less

than 0.01 when comparing the removal rates in R3 and

those in R1 and R2, indicating that R3 had a better and

more stable degrading efficiency.

Difference in composition of bacterial

communities in the three reactors

The different performances of these reactors predicted

the differences in bacterial communities, which were

worthy of further exploration. The presence of the

degrading strain (C. testosteroni A3) and the biofilm-

forming strain (B. cereus G5) in biofilms on the strip

carriers was determined by pyrosequencing. Biofilm

samples were taken from the carriers in each of three

reactors on the 30th day. Pyrosequencing analysis

yielded about 10,000 sequences from every reactor

after removal of short and low-quality reads with an

average read length of 414 bp. Rarefaction curves

were generated at 3 % cut-off to make a comparison of

species richness among the three samples. The shape

Fig. 3 3,5-DNBA degradation performance of sequencing

batch biofilm reactors (SBBRs). The SBBRs were operated as

the following mode, with a cycle (24 h) consisting of a 30 min

filling phase, a 22 h aeration period, a 1 h settling phase, and a

30 min drawing phase. a 3.5-DNBA (mg l-1), b Removal rates

of 3,5-DNBA (%) R1: with activated sludge only, R2:with

C. testosteroni A3 and activated sludge,R3:with C. testosteroni

A3, B. cereus G5 and activated sludge

Biotechnol Lett

123

of the rarefaction curves indicated that bacterial

richness of every reactor was complete (Fig. 4).

Done up to here.

Genera with relative abundances higher than 1 %

within total bacteria were sorted and estimated for the

bacterial community diversity and the related result

about the diversity of three reactors was illuminated as

shown in Fig. 5.

In general, Comamonas was predominant in R2 and

R3, accounting for 49 % (R2) and 79 % (R3),

respectively, which are much higher than that in R1

(1.9 %). At the same time, the percent (11 %) of

Bacillus in R3 was higher than those in R2 and R1.

Conversely, the richness of bacterial communities in

R1was obviously greater than those in R2 and R3,

which was attributable to the action of selection from

3,5-DNBA in the reactors during the whole running

period. C. testosteroni A3, as 3,5-DNBA-degrading

bacterium, propagated more quickly in R2 and R3 and

slowed down growth of other bacteria and, conse-

quently, indicated less richness in the reactor. These

results identified that both B. cereus G5 and C.

testosteroni A3 were immobilized successfully in

biofilms in R3, which was in agreement with the

outperformance of R3.

In this way, immobilizing specific-pollutant

degrading strains into biofilms mediated by biofilm-

forming bacteria might be an efficient approach for

colonization of the degraders in bioaugmentation

treatment systems and show broader application

prospects.

Fig. 4 Rarefaction curves of biofilm samples from three

reactors. The number of operational taxonomic units (OTUs)

with 3 % cut-off values was plotted as a function of the number

of sequences sampled. R1: with activated sludge only, R2: with

C. testosteroni A3 and activated sludge, R3: with C. testosteroni

A3, B. cereus G5 and activated sludge

Fig. 5 Relative abundance of the main genera identified in

different biofilm samples from three reactors. Only genera with

a relative abundance more than 1 % are shown. R1: with

activated sludge only, R2: with C. testosteroni A3 and activated

sludge, R3: with C. testosteroni A3, B. cereus G5 and activated

sludge

Biotechnol Lett

123

Acknowledgments This work was supported by the National

Natural Science Foundation of China (no. 51079094) and

Suzhou Environment Protection Agency (no.C201302).

References

Abe A, Furukawa S, Watanabe S, Morinaga Y (2013) Yeasts

and lactic acid bacteria mixed-specie biofilm formation is a

promising cell immobilization technology for ethanol fer-

mentation. Appl Biochem Biotechnol 171:72–79

Boon N, Top EM, Verstraete W, Siciliano SD (2003) Bioaug-

mentation as a tool to protect the structure and function of

an activated-sludge microbial community against a

3-chloroaniline shock load. Appl Environ Microbiol 69:

1511–1520

Buswell CM, Herlihy YM, Marsh PD, Keevil CW, Leach SA

(1997) Coaggregation amongst aquatic biofilm bacteria.

J Appl Microbiol 83:477–484

Cheng ZQ, Meng XX, Wang HC, Chen M, Li MY (2014) Iso-

lation and characterization of broad spectrum coaggregat-

ing bacteria from different water systems for potential use

in bioaugmentation. PLoS ONE 9:e94220

Diana DG, Fambrini L, Coppini E, Fava F, Barberio C (2004)

Aggregation-based cooperation during bacterial aerobic

degradation of polyethoxylated nonylphenols. Res Micro-

biol 155:761–769

Duque AF, Bessa VS, Carvalho MF, Castro PM (2011) Bioaug-

mentation of a rotating biological contactor for degradation

of 2-fluorophenol. Bioresour Technol 102:9300–9303

Kolenbrander PE (2000) Oral microbial communities: biofilms,

interactions, and genetic systems. Annu Rev Microbiol

54:413–437

Lawniczak L, Kaczorek E, Olszanowski A (2011) The influence

of cell immobilization by biofilm forming on the biodeg-

radation capabilities of bacterial consortia. World J

Microbiol Biotechnol 27:1183–1188

Li MY, Peng L, Ji Z, Xu JL, Li SP (2008) Establishment and

characterization of dual-species biofilms formed from a

3,5-dinitrobenzoic-degrading strain and bacteria with high

biofilm-forming capabilities. FEMS Microbiol Lett

278:15–21

Li CY, Li Y, Cheng XS, Feng LP, Xi CW, Zhang Y (2013)

Immobilization of Rhodococcus rhodochrous BX2 (an

acetonitrile-degrading bacterium) with biofilm-forming

bacteria for wastewater treatment. Bioresour Technol

131:390–396

Limbergen VH, Top EM, Verstraete W (1998) Bioaugmentation

in activated sludge: current features and future perspec-

tives. Appl Microbiol Biotechnol 50:16–23

Mah TC, O’Toole GA (2001) Mechanisms of biofilm resistance

to antimicrobial agents. Trends Microbiol 9:34–39

Malik A, Sakamoto M, Hanazaki S, Osawa M, Suzuki T,

Tochigi M, Kakii K (2003a) Coaggregation among non-

flocculating bacteria isolated from activated sludge. Appl

Environ Microbiol 69:6056–6063

Malik A, Sakamoto M, Ono T, Kakii K (2003b) Coaggregation

between Acinetobacter johnsonii S35 and Microbacterium

esteraromaticum strains isolated from sewage activated

sludge. J Biosci Bioeng 96:10–15

Park JY, Choi YJ, Moon S, Shin DY, Nam K (2009) Micro-

bubble suspension as a carrier of oxygen and acclimated

bacteria for phenanthrene biodegradation. J Hazard Mater

163:761–767

Rickard AH, Gilbert P, Handley PS (2004) Influence of growth

environment on coaggregation between freshwater biofilm

bacteria. J Appl Microbiol 96:1367–1373

Simoes LC, Simoes M, Vieira MJ (2008) Intergeneric coag-

gregation among drinking water bacteria: evidence of a

role for Acinetobacter calcoaceticus as a bridging bacte-

rium. Appl Environ Microbiol 74:1259–1263

Zhu J, Mekalanos JJ (2003) Quorum sensing-dependent biofilms

enhance colonization in Vibrid cholerae. Dev Cell 5:

647–656

Biotechnol Lett

123