Applied Microbiology and Biotechnology

, Volume 95, Issue 1, pp 255–262

Isolation and characterization of a Klebsiella oxytoca strain for simultaneous azo-dye anaerobic reduction and bio-hydrogen production


  • Lei Yu
    • Advanced Laboratory for Environmental Research and Technology
    • Department of ChemistryUniversity of Science and Technology of China
    • Department of Biology and ChemistryCity University of Hong Kong
    • Advanced Laboratory for Environmental Research and Technology
    • Department of ChemistryUniversity of Science and Technology of China
  • Michael Hon-Wah Lam
    • Advanced Laboratory for Environmental Research and Technology
    • Department of Biology and ChemistryCity University of Hong Kong
  • Han-Qing Yu
    • Department of ChemistryUniversity of Science and Technology of China
  • Chao Wu
    • Department of ChemistryUniversity of Science and Technology of China
Environmental biotechnology

DOI: 10.1007/s00253-011-3688-2

Cite this article as:
Yu, L., Li, W., Lam, M.H. et al. Appl Microbiol Biotechnol (2012) 95: 255. doi:10.1007/s00253-011-3688-2


A facultative anaerobic bacteria strain GS-4-08, isolated from an anaerobic sequence batch reactor for synthetic dye wastewater treatment, was investigated for azo-dye decolorization. This bacterium was identified as a member of Klebsiella oxytoca based on Gram staining, morphology characterization and 16S rRNA gene analysis. It exhibited a good capacity of simultaneous decolorization and hydrogen production in the presence of electron donor. The hydrogen production was less affected even at a high Methyl Orange (MO) concentration of 0.5 mM, indicating a superior tolerability of this strain to MO. This efficient bio-hydrogen production from electron donor can not only avoid bacterial inhibition due to accumulation of volatile fatty acids during MO decolorization, but also can recover considerable energy from dye wastewater.


AnaerobicBio-hydrogenDecolorizationElectron donorKlebisiella oxytocaMethyl Orange (MO)


Intensive research has been recently conducted in treatment of azo-dye wastewaters, which are known to be toxic, carcinogenic and causing serious environmental pollution (O’Nell et al. 2000; Méndez-Paz et al. 2005). Biological method has been recognized as an inexpensive, environmental friendly and sustainable way for dye wastewater treatment (Rajaguru et al. 2000; Yu et al. 2011). However, azo dyes are mostly xenobiotic compounds that contain one or more azo groups. Decolorization, i.e., breakage of the azo bond, is the first step of azo dye degradation (van der Zee and Villaverde 2005). Anaerobic–aerobic treatment has long been a common practice for azo dye biodegradation. In the anaerobic process, however, methanogenesis can be seriously inhibited by azo dye or its reduced products, resulting in VFA accumulation and low biogas production. And the increased VFA concentration would, in turn, further inhibit the biogas production (Isık and Sponza 2004; Méndez-Paz et al. 2005). As a result, almost no valuable energy products can be recovered from dye-containing wastewaters, leading to a vast waste of potentially precious resources.

This problem might be addressed if some powerful bacterium is available that can simultaneously reduce the dyes and convert VFA into valuable products (ethanol or hydrogen gas). However, no such single strains have been reported so far. Fortunately, a bacterium capable of simultaneous dye reduction and hydrogen production, strain GS-4-08, was discovered and successfully isolated by us. In this study, we characterize this strain and evaluate its performances in azo dye decolorization and bio-hydrogen production. To the best of our knowledge, this is the first report on simultaneous anaerobic reduction of azo dyes and bio-hydrogen production by a single strain.

Materials and methods

Chemicals and cultivation media

The azo dyes, Acid Red 88 (AR88), Reactive Black 5 (RB5), Disperse Orange 3 (DO3), Direct Red 81 (DR81), were purchased from Sigma-Aldrich without further purification. N’,N-dimethyl-p-phenylenediamine (DPD) was procured from Aladdin. Other chemicals were of analytical grade and were procured from Sinopharm Chemical Reagent (China).

The Luria–Bertani (LB) medium was used for aerobic cultivation. The basal medium was used as the screening medium, which contained (g l−1 of tap water): glucose (1.0), K2HPO4 (1.11), KH2PO4 (2.09), NaHCO3 (0.5), NH4Cl (0.28), MgSO4·7H2O (0.1), yeast extract (0.1). Azo-dye decolorization and sucrose fermentation tests were performed using a simulated dye wastewater containing (g l−1 of tap water): sucrose (6.84), Methyl Orange (MO, 0.033), K2HPO4 (1.0), KH2PO4 (0.5), (NH4)2SO4 (2.0), MgSO4·7H2O (0.1), 1 ml of a trace element solution containing (mg l−1) FeSO4·7H2O (5), ZnSO4·7H2O (0.011), MnCl2·4H2O (0.1), CuSO4·5H2O (0.392), Co(NO3)2·6H2O (0.248), NaB4O7·10H2O (0.177), NiCl2·6H2O (0.025) (O’Nell et al. 2000).

Bacteria isolation

An azo-dye tolerant strain, GS-4-08, was isolated from an anaerobic sequencing batch reactor for treatment of dye wastewater (Yu et al. 2011). Firstly, a 2-ml mixture of anaerobic sludge was inoculated into 100 ml screening medium with 0.1 g l−1 MO, and incubated at 30 °C and pH 7.0 under static conditions. After the color of medium disappeared, a new round of screening operation was performed with an inoculation ratio of 3% (v/v). The screening process was repeated five times, and finally the concentration of target bacteria was 1,000-fold diluted. A sloop of 50 μl aliquots of the dilution was plated on LB agar containing 0.1 mM MO. After purification by single colony isolation, stain GS-4-08 was identified and preserved at −70 °C in LB with 30% glycerol.

16S rRNA sequencing and phylogenetic analysis

To identify the strain, the bacteria culture was subjected to 16S rRNA sequencing and phylogenetic analysis. DNA from 50 ml of samples was isolated and purified using a spin column bacteria genomic DNA reagent kit (Sangon Biotech Co., Ltd, China). Then 16S rRNA gene was amplified by using the polymerase chain reaction (PCR) method with two universal primers, 1492R and 27F. The PCR amplification products were used for the determination of 16S rRNA gene sequence (Sangon Biotech CO., Ltd., China). The sequences amplified from this strain were compared with those in the GenBank nucleotide database by using BLAST program packages. A phylogenetic tree was constructed from distance matrix via neighbor-joining method by using the program of MEGA 4.1.

Decolorization and hydrogen production tests

The strain was first cultivated under aerobic conditions at 30 °C overnight in LB medium and then harvested by centrifugation (8,000 × g, 25 min). The pellet was washed twice in 10 ml Na–K phosphate buffer (pH = 7.0) and then re-suspended in 20 ml fresh Na–K phosphate buffer. Cell suspension was transferred into 250-ml serum bottles containing simulated dye wastewater (100 ml) with an initial cell mass of approximately 0.3 ± 0.01 g l−1 (cell dry weight [CDW]). After being purged with N2 gas to ensure the anaerobic condition, the serum bottles were sealed with butyl rubber stoppers and incubated statically under anaerobic conditions at 30 °C and pH 7.0. The sucrose (20 mM) and MO (0.1 mM) were used as electron donor and acceptor, respectively, unless otherwise stated. Then, a series of tests on dye decolorization and hydrogen production were performed.

Analytical methods

Liquid samples were extracted from serum bottles at regular time interval for analysis of the residual dyes and reduced products. The dye concentrations in supernatant were determined by a UV/visible spectrophotometer (UV-2550; Shimadzu, Japan). The concentration of aromatic amines was analyzed by an HPLC (Waters 2695; Waters Inc., USA) according to Yu et al. (2011). Reduced sugars were determined using the anthrone method (Koehler et al. 1952), while the CDW was measured according to the Standard methods (APHA 1998). VFA concentration was detected by a GC-FID (GC-6890N; Agilent Inc., USA) equipped with a fused-silica capillary column (DB-FFAP) as reported previously (Mu et al. 2007). Gas production was periodically measured by releasing the pressure in bottles using a glass syringe according to the method described by Owen et al. (1979). The hydrogen was measured using a gas-tight syringe (1.0 ml injection volume), analyzed by a GC-FID (GC-9790; FuLi Inc., China) coupling with a 5-Å molecular sieve column with argon as the carrier gas.

The bacteria samples were collect at the end of the exponential growth phase, washed by 50 mM phosphate buffer solution (PBS) for three times, and then fixed using 2.5% glutaraldehyde overnight. After decanting the fix solution, the samples were rinsed with PBS for three times again. The washed samples were dehydrated by ethanol solution, then vacuum-dried and coated by Au prior to SEM (S-4700; Hitachi, Japan) observation.


Morphological and phylogenetic analysis of strain GS-4-08

The SEM images show that the isolated strain had rod shape and was atrichous (Fig. 1). A further biochemical assay reveals that it is a Gram-negative and facultative anaerobic bacterium (data not shown). According to the 16S rRNA gene (approximately 1,400 bp) sequence analysis, this strain belongs to the Klebsiella oxytoca species. Analysis against the NCBI database using BLAST program packages identified Klebsiella species as the closest known relatives to strain GS-4-08 (accession no. FJ816026), exhibiting a highest homology of 99%. A detailed phylogenetic placement of the isolate is shown in Fig. 2. This strain has been deposited at China General Microbiological Culture Collection Center under the provisions of the Budapest Treaty, accession Number 5237.
Fig. 1

SEM images of strain GS-4-08 at resolution of: a 5,000, b 20,000
Fig. 2

Phylogenetic tree of genus Klebsiella based on 16S rRNA gene sequence. The tree was constructed using the neighbor-joining method. GenBank sequence accession numbers were given in parentheses. Numbers at the nodes show the % bootstrap value

Decolorization characteristics of K. oxytoca GS-4-08

In this study, the electron donors showed a positive effect on the decolorization rate (DR) for this strain. The addition of sugars significantly promoted the decolorization process by strain GS-4-08, while organic acids exhibited a relative lower acceleration effect (Fig. 3a and b). Sucrose was determined as the most favorite electron donor for this strain to decolorize the model dye. Moreover, the appropriate sucrose concentration for the azo dye decolorization was also investigated (Fig. 4a). The addition of sucrose from 0 to 20 mM enhanced the decolorization activity, while further increase the sucrose concentration showed no significant promotion. About 86.93 ± 1.83% of the MO (0.1 mM) was decolorized by genus GS-4-08 within 24 h when 20 mM sucrose was added as the electron donor. The optimum pH and temperature for MO decolorization were determined as 7 °C and 35 °C (Fig. 3c and d), the subsequent experiments were all performed under this condition unless otherwise mentioned.
Fig. 3

MO decolorization rates with different substrates of a organic acids, b sugars, and at various c temperatures, and d pH values (the concentrations of electron donor and MO were 20 and 0.1 mM, respectively, in the batches)
Fig. 4

MO decolorization efficiencies at different initial concentrations of a sucrose, b cell and c MO

The decolorization performances under varied biomass concentration were also investigated in this study. As shown in Fig. 4b, the specific decolorization rates (SDRs) in all the batches increased significantly in the initial 14 h, and then declined gradually till the end of cultivation at 48 h. The overall decolorization efficiencies increased with the initial biomass concentration. Low decolorization efficiencies of 35.53 ± 1.61 and 54.63 ± 1.03% were obtained, respectively, at low CDW batches (0.5 and 1.5 g l−1). The appropriate initial CDW for decolorization of 0.1 mM MO was 3.0 g l−1, exceeding which no obvious enhancement of SDR was observed. A similar result was obtained by Cosmarium sp. in biodegradation of Malachite Green in a previous study (Daneshvar et al. 2007).

To determine the maximum tolerable concentration of MO for K. oxytoca GS-4-08, the decolorization performances under different initial MO concentrations (0.1–0.5 mM) were investigated. At an initial MO concentration of 0.1 mM, the decolorization efficiency was about 86.93 ± 1.83% after 24 h cultivation and reached 96.86 ± 1.48% after 30 h. Further increase the dye concentration lowered the decolorization efficiency, but the dyes could be almost entirely decolorized when extending the incubation time. As shown in Fig. 4c, at initial dye concentration of 0.2 mM the decolorization efficiencies reached 96.53 ± 1.25% after 48 h, while 96 h was needed to achieve 92.96 ± 0.88% decolorization at an initial dye concentration of 0.3 mM. When the initial dye concentration was further increased to 0.5 mM, a rapid decolorization was achieved in the initial 48 h. But, after that it slowed down, and finally reached a decolorization efficiency of about 75%. This result implies that K. oxytoca G-04-08 can tolerate a MO concentration as high as 0.5 mM.

In addition to MO, GS-4-08 also showed good decolorization ability toward a broad range of azo dyes (Table 1). The highest SDR at 131.16 mg g cell−1 h−1 was obtained when Methyl Red (MR) was used as the model azo dye, while the Orange II was the hardest for decolorization by strain GS-4-08. It was noticed that the maximum SDRs of AR88 and CR were 8.36 and 8.47 mg g cell−1 h−1, respectively, despite that they both own two sulfonate groups. The DO3 formed a new adsorption peak at 382 nm (close to λmax of 443 nm) and there was no residual absorbance at λmax. Accordingly, a shift of color from red to yellow was observed. It infers that the azo bonds of DO3 are likely to be completely broken and formed a new coloration radical. The similar phenomenon was also observed for decolorization of Reactive Red 2 and Reactive Red 4 by anaerobic granular sludge in other studies (van der Zee et al. 2001).
Table 1

Decolorization rates of various azo dyes by K. oxytoca GS-4-08

aC0 = initial dye concentration

bThe solution color turns into yellow, resulting in a λmax = 382 solution

Bio-hydrogen production from electron donors by K. oxytoca GS-4-08 in MO decolorization

MO decolorization and sucrose degradation in a typical run are illustrated in Fig. 5, where the initial MO, sucrose and CDW concentrations were 0.1 mM, 20 mM and 3.0 g l−1, respectively. In this study, MO was degraded to 4-aminobenzenesulfonic acid (4-ABS) and DPD (Fig. 5a), while sucrose degradation was accompanied by the formation of acetate, ethanol (Fig. 5b) and hydrogen (Fig. 5c), simultaneously. As shown in Table 2, the calculated hydrogen yields by K. oxytoca GS-4-08 varied only slightly and non-systematically, so did the acetate and ethanol.
Fig. 5

Profiles of a MO decolorization and intermediate formation, b sucrose degradation and metabolic products formation, and c hydrogen volume and production rate. The initial concentrations of MO and sucrose were 0.1 and 20 mM, respectively, and freely suspended cells of K. oxytoca GS-4-08 were used in the test

Table 2

Effect of initial dye concentration on the sucrose degradation by K. oxytoca GS-4-08



YH2,(t=96 h)




























CMO initial MO concentration (mM), DE degradation efficiency (%), YH2,(t=96 h) actual hydrogen yield (mmol g sucrose−1), YHAc acetic acid yield (mmol g sucrose−1), YEtOH ethanol yield (mmol g sucrose−1)


The isolated strain K. oxytoca GS-4-08 exhibited considerable anaerobic decolorization ability to a broad range of azo dyes. The highest SDR for MR was 131.15 mg g cell−1 h−1 under a high initial concentration of 1 mM. This high performance might be due to fact that MR could easily penetrate though the cell membranes and be degraded efficiently (Chen et al. 2010), while those dyes with highly polar sulfonate groups cannot penetrate through the cell membrane and thus have lower SDRs (Table 1). Notably, no decolorization occurred under shaking aerobic condition, indicating an anaerobic culturing nature of this bacterium toward azo dyes. Electron donor is necessary in dye anaerobic decolorization, and the type of applied electron donor can directly affect the decolorization performance. Sucrose was identified as the favorite electron donor for MO decolorization, with a highest SDR of 4.04 ± 0.09 mg g cell−1 h−1 achieved at a sucrose concentration of 20 mM. Higher sucrose concentration caused no further increase in SDR, possibly because that although the bacteria preferentially utilize sucrose as carbon source for anaerobic growth, a too high sucrose concentration may result in the “latent period” (the initial 6 h cultivation as shown in Fig. 4a) and lower the decolorization efficiency (Wang et al. 2009).

The strain GS-4-08 has comparable decolorization ability to other reported strains. A number of bacteria like Klebsiella pneumoniae RS-13, Kocuria rosea MTCC 1532 and Enterobacter sp. hEC3, fungi Ceriporia lacerate P2 have been previously reported to have a considerable ability of dye decolorization (Parshetti et al. 2010; Wang et al. 2009; Wong and Yuen 1996; Xu et al. 2008). Compared with these bacteria, the genus GS-4-08 in this study showed a much superior decolorization capability (Table 3).
Table 3

Comparison between the decolorization ability of K. oxytoca GS-4-08 and those reported in the literature







Klebsiella pneumoniae RS-13



Wong and Yuen (1996)

Klebsiella oxytoca GS-4-08



This study


Kocuria rosea MTCC 1532



Parshetti et al. (2010)

Ceriporia lacerate P2



Xu et al. (2008)

Klebsiella oxytoca GS-4-08



This study


Enterobacter sp. hEC3



Wang et al. (2009)

Klebsiella oxytoca GS-4-08



This study

Values presented in this study are obtained in addition of electron donor (sucrose) concentration of 20 mM, cultivation time of 24 h and CDW concentration of 3.0 g l−1, except when otherwise noted

aDye concentration in reference/in this study (mg l−1)

bAverage decolorization efficiencies after 24 h except for MR decolorization (%)

cDegradation time were 12 and 7 h in the reference and in this study, respectively

It has been reported that high concentration of azo dye during anaerobic degradation would inhibit the methanogenesis, which may induce the accumulation of VFAs and further inhibit biogas production (Isık and Sponza 2004). However, no obvious VFA accumulation was observed in this study even at high MO concentration of up to 0.5 mM (Table 2). Only low concentration of acetate (0.17–0.19 g l−1) was detected in all batch runs, suggesting that the produced VFAs have been effectively utilized for hydrogen and ethanol production. The hydrogen yields remained almost unchanged (ranging slightly from 8.04 to 9.33 mmol g−1 sucrose) independent of the initial MO concentration (Table 2). These results suggest that this strain possesses a remarkable high hydrogen production capability as well as azo dye adaptive ability.

Table 4 compares the hydrogen yields and volumetric hydrogen production rates of K. oxytoca with bacteria of same species or engineered bacteria (Long et al. 2005; Mathews et al. 2010; Liu and Fang 2007; Wu et al. 2008). K. oxytoca showed a maximum hydrogen yield of 1.60 mol H2 mol−1 glucose, which was considerably higher than the others except for the engineered Escherichia coli stains (Mathews et al. 2010). It can efficiently produce hydrogen even at a high MO concentration of 0.5 mM. The maximum hydrogen yield (9.33 mmol g−1 sucrose) was comparable to other Klebsiella species. This high ability of hydrogen production confers several advantages: first, it avoids the accumulation of VFA, and enables considerable energy recovery during dye wastewater treatment; moreover, the efficient conversion of electron donor would release more electrons for MO degradation and in turn further promote MO decolorization. This study demonstrates that K. oxytoca GS-4-08 is a powerful strain that possesses excellent dye reduction and hydrogen production capabilities, and has a great potential for dye wastewater treatment application.
Table 4

Comparison between the hydrogen production yields and rates obtained in this study to those reported in the literature








Klebsiella oxytoca GS-4-08




This work


Klebsiella oxytoca HP1



Long et al. (2005)


Engineered Escherichia coli strain


HFo, HAc

Mathews et al. (2010)


Klebsiella pneumoniae DSM 2026




Liu and Fang (2007)


Klebsiella sp. HE1



2,3-BDL, EtOH

Wu et al. (2008)

EtOH ethanol; HFo formatic acid; HAc acetic acid; 1,3-PD1,3-propanediol; 2,3-BDL 2,3-butanediol

aMaximum hydrogen yield (mol H2 mol−1 glucose)

bMaximum volumetric hydrogen production rate (mmol H2 l−1 h−1)


The authors wish to thank the NSFC-JST joint project (21021140001), the National 863 Program of China (2008BADC4B18), the JSNSF (BK2010256), the Chinese Universities Scientific Fund and the Suzhou Environmental Protection Technology project for the partial support of this study.

Copyright information

© Springer-Verlag 2011