Abstract

Endotoxins are potential toxics impacting human health through respiration derived in wastewater treatment plants (WWTPs), yet the formation of endotoxins during wastewater treatment processes is still lacking research. In our study, the distribution of endotoxins and bacterial community structure in the wastewater of three full scale pharmaceutical WWTPs were explored using the limulus amebocyte lysate (LAL) test and MiSeq technique. Results showed that higher endotoxin activities in the influent of Plant A and Plant C (560 and 1140 EU/mL), stemming from the fermentation process, were found compared to that of Plant B (135 EU/mL), coming from the process of chemical synthesis. During the anaerobic treatment and the cyclic activated sludge system (CASS) in the three WWTPs, the endotoxin activity increased, while it declined in the aerobic treatment system. In all bioreactors, the relative abundance of Gram-negative bacteria accounted for 50.0–94.6%. Bacteria with high lipopolysaccharide (LPS) in LAL assay were found at the genus level of Bacteroides, Enterococcus, Desulfovibrio, and Megasphaera.

INTRODUCTION

Exposure to airborne endotoxins, due to their adverse effect on human health, has attracted much attention in several occupational environments such as livestock farming, vegetable greenhouses, and wastewater treatment plants (WWTPs). Epidemiological investigations suggest a higher prevalence of gastrointestinal tract symptoms, respiratory symptoms, fatigue, and headaches among employees in the above occupations (Duquenne et al. 2014). Endotoxin, also called Lipopolysaccharide (LPS), constitutes the outer leaflet of the outer membrane of most Gram-negative bacteria and some cyanobacteria (Anderson et al. 2002). LPS consists mainly of three parts: Lipid A, core sugars and O-antigen repeats (Caroff & Karibian 2003). Lipid A is responsible for the major bioactivity of endotoxin. The health-based recommended occupational exposure limit (HBROEL) for endotoxin is 50 EU/m3 (Visser et al. 2006). Endotoxin activity is determined by the composition and structure of LPS, such as limulus amebocyte lysate (LAL) test, HPLC, rFC and ELISA methods. The LAL test is the most widely used analytical method, and is based on enzymatic coagulation of the blood of a primitive marine arthropod, i.e., the horseshoe crab (Chiominto et al. 2014).

In WWTPs, airborne endotoxins are generated during several unit operations such as surface aeration, mechanical sludge removal, and some cleaning practices (Visser et al. 2006). As a component of Gram-negative bacteria, which are an abundant and important class of microorganisms present in WWTP effluent, endotoxins are transmitted to the ambient air in wastewater droplets (Gangamma et al. 2011). There are lots of studies on airborne endotoxins about the sampling analysis methods, exposure levels, as well as corresponding influence factors and relevant prevention measures (Forestier et al. 2012; Paba et al. 2013). However, few studies have focused on the fate of endotoxins in the wastewater, which is the origin of airborne endotoxins during wastewater treatment processes, although the endotoxin activity in wastewater has a big influence on the exposure level of airborne endotoxin.

Among WWTPs, biological treatment is the most popular treatment process. Activated sludge or biofilm in the biological treatment units includes eukaryotes (protozoa and fungi), bacteria, archaea and viruses (Zhang et al. 2012). Endotoxin can be released from the Gram-negative bacteria into wastewater during the multiplication, death and lysis process (Huang et al. 2011), which varies with different kinds of bacteria (Zhang-Sun et al. 2015). However, a constituent of LPS can be utilized as a carbon source by some filamentous bacteria, which indicates that endotoxin in wastewater can be degraded by some kinds of bacteria (Kragelund et al. 2008). The protozoa existing in the biological treatment process could reduce the release of endotoxin by feeding on the bacteria and their corpses (Kamika & Momba 2015). In addition, Guizani (2010) surveyed three plants that mainly receive domestic wastewater and the results showed that organic matter of larger size (100 KDa-0.1 μm) exhibited higher endotoxin concentration. Presently, few studies on pharmaceutical wastewater plants have been reported. Therefore, to highlight the relationship between LPS and the microbial community structure of activated sludge or biofilm during biological treatment is valuable for further understanding and controlling the release and degradation process of LPS.

This study aims to discover the distribution of endotoxin in wastewater at three pharmaceutical WWTPs using the LAL test. At the same time, the microbial communities of the activated sludge or biofilm in biological treatment units have also been analyzed through 16 s rRNA gene sequencing technology.

MATERIAL AND METHODS

Investigated wastewater treatment plants

The three WWTPs mainly receive wastewater generated during the production of pharmaceuticals. Plant A has a daily processing capacity of 500 m3 and mainly treats the wastewater generated during the production of Vitamin C. The treatment processes of Plant A are a combination of anaerobic and aerobic treatment (Figure 1(a)). Plant B, with a processing capacity of 900 m3/day, mainly receives wastewater derived from the production of Vitamin A and Vitamin E. In Plant B, biological treatment units (a cyclic activated sludge system (CASS) and a moving bed biofilm reactor (MBBR)) follow an air flotation tank (Figure 1(b)). Plant C mainly receives a kind of mixed pharmaceutical wastewater derived from a biological fermentation process, including high-end biotech drugs, injections and oral preparations and some other drugs. In Plant C, there are two independent influent sources with high chemical oxygen demand (COD) concentration and low COD concentration wastewater. The total daily influent quantity of the two sections is about 1700 m3 with a ratio of 0.4–0.6 (high/low), in which the flux of the high COD concentration section and low COD concentration section are about 680 m3/d and 1020 m3/d, respectively. The influent of the high COD concentration section was treated using a hydrolytic acidification tank and CASS. The influent of the low COD concentration section was treated by a hydrolytic acidification unit. After the independent treatment process, both influents were combined into the same process, which contained an anaerobic and an aerobic units and an air flotation unit. The treatment procedures and sampling locations of Plant C are showed in Figure 1(c).

Figure 1

Wastewater treatment process and water quality index results of studied samples in Plant A (a), Plant B (b), Plant C (c). Sampling locations are shown in the figure: sampling point of wastewater, ●; sampling point of activated sludge, ◆.

Figure 1

Wastewater treatment process and water quality index results of studied samples in Plant A (a), Plant B (b), Plant C (c). Sampling locations are shown in the figure: sampling point of wastewater, ●; sampling point of activated sludge, ◆.

Analytical methods

All the wastewater samples were collected in plastic bottles, placed in a container filled with ice, and transported to the laboratory. Samples of 50 mL in several centrifuge tubes were centrifuged at 4000 g and the supernatant was stored at −20 °C. After collecting all samples, various chemical parameters such as COD, ammonia nitrogen, nitrate nitrogen and chloride ions were analyzed according to the State Environmental Protection Administration (EPA) Standard Methods for the Analysis of Water and Wastewater (EPA 2002).

As shown in Table 1, activated sludge and biofilm samples were taken from the three WWTPs. All sludge was settled on site to be concentrated and then was fixed in 50% (v/v) ethanol aqueous solution. The fixed samples were immediately transported to the laboratory for further treatment.

Table 1

The sampling locations and the corresponding numbers of samples

Plant ID Sampling location Type of samples Sampling time 
A-S1 Anaerobic tank Activated sludge 04/2014 
A-S2 Aerobic tank Activated sludge 04/2014 
B-S1 CASS Activated sludge 04/2014 
B-S2 MBBR Biological film 04/2014 
C-S1 CASS Activated sludge 05/2014 
C-S2 Anaerobic tank Activated sludge 05/2014 
C-S3 Aerobic tank Activated sludge 05/2014 
Plant ID Sampling location Type of samples Sampling time 
A-S1 Anaerobic tank Activated sludge 04/2014 
A-S2 Aerobic tank Activated sludge 04/2014 
B-S1 CASS Activated sludge 04/2014 
B-S2 MBBR Biological film 04/2014 
C-S1 CASS Activated sludge 05/2014 
C-S2 Anaerobic tank Activated sludge 05/2014 
C-S3 Aerobic tank Activated sludge 05/2014 

Endotoxin activity assay

The endotoxin activities of samples were analyzed using chromogenic Limulus (Tachypleus tridentatus) Amebocyte Lysate (LAL) endpoint assay kits (Xiamen Houshiji Ltd, China). Approximately 2–4 serial dilutions (1:100–1:100,000) of each sample in pyrogen-free water (Milli-Q UF water) were analyzed based on the expected endotoxin concentration. The diluted samples were mixed with LAL, incubated for 10 min at 37 °C, mixed with chromogenic substrate, and incubated for another 6 min; the reaction was then stopped with acid. The absorbance of yellow color was determined at 545 nm. The endotoxin activity was obtained from a standard curve using endotoxin standards from the Escherichia coli O111:B4 strain.

To test the validation of the endotoxin detection, positive controls, spiked with 0.10 EU/mL of endotoxins, were used to determine the recovery ratio. Spike ratios of each sample had to be between 50% and 200% to demonstrate a range of insignificant interference, and to determine appropriate sample dilutions.

Depyrogenated dilution glass tubes, depyrogenated reaction glass tubes and depyrogenated tips for pipettes were purchased from Xiamen Houshiji Ltd, China. Other glassware used in this study was made endotoxin-free by heating at 250 °C for 1 h.

Procedure of 16S rRNA gene amplicon sequencing and data analysis

Samples of activated sludge and biofilm for 16S rRNA gene amplicon sequencing were processed by DNA extraction, 16S rRNA gene polymerase chain reaction (PCR) amplification and PCR products purification. To amplify and sequence the V1-V2 hypervariable region of the 16S rRNA gene, forward primer (50-AGAGTTTGATYMTGGCTCAG-30) and reverse primer (50-TGCTGCCTCCCGTAGGAGT-30) were selected and different 8-base barcodes and a guanine were linked to the 50 end of each primer. Then the purified products were sent for sequencing using the Illumina sequencing platform (Miseq, Illumina Inc., USA). The acquired data was processed by the Sickle and Mothur program to remove the low quality of sequence and reduce noises. Finally, the filtered sequences were assigned to a taxon by the RDP classifier. Microbial abundance at the genus level was mapped using heatmap modules in ‘R’ statistical packages.

RESULTS AND DISCUSSION

Pollutant and physicochemical gradients along treatment

Throughout the whole process of the three WWTPs, both the concentrations of COD and NH4+-N were generally decreased along the treatment process. Each treatment unit in the three WWTPs showed a capacity for COD removal. Above all, the total removal rate of COD was 96.5%, 97.7% and 98.4% at Plant A, Plant B and Plant C, respectively. The removal of NH4+-N mainly occurred in the process of aerobic biological treatment. The total removal rate of NH4+-N was 96.5%, 97.7%, and 92.8% in Plant A, Plant B and Plant C, respectively, which showed the three WWTPs had a good NH4+-N removal efficiency. However, the denitrification process did not show a good performance in Plant A in which the NO3-N concentration of effluent was 180.43 mg/L. Compared to Plant A and B, Plant C showed a good denitrification efficiency with a lower NO3-N concentration.

Endotoxin distribution in the three WWTPs

Increases of endotoxin activities during WWTPs treatment

The endotoxins in pharmaceutical wastewater following each unit at the three WWTPs were measured using an LAL test. Results of the endotoxin activities of wastewater at the three WWTPs are shown in Figure 2(a)2(c). In Plant A, the endotoxin activity of samples was increased from 560 EU/mL in influent to 3657 EU/mL in effluent, respectively, showing a total increase of 5.5 times (Figure 2(a)). In Plant B, the endotoxin activity of samples was increased from 135 EU/mL to 4737 EU/mL in the influent and effluent, which showed an increase of 34 times (Figure 2(b)). The treatment process of Plant C is more complicated, and the endotoxin activity in the influent could be calculated as follows, in which the parameters Q means the influent quantity (m3/day) of the two sections treating the high and low COD concentration influent, respectively.  
formula
(1)
Figure 2

The endotoxin activities of the wastewater during each process and the corresponding recoveries of LAL assay in (a) Plant A, (b) Plant B, and (c) Plant C.

Figure 2

The endotoxin activities of the wastewater during each process and the corresponding recoveries of LAL assay in (a) Plant A, (b) Plant B, and (c) Plant C.

The result showed that the average endotoxin activity in the influent of Plant C was 1133 EU/mL. The endotoxin activity in the final effluent was 33,867 EU/mL, showing a total increase of about 29 times (Figure 2(c)). The recoveries of all spiked samples after dilutions during the LAL assay are also shown in Figure 2. The recovery of all samples was in the range of 50–200%, which validated the LAL assay for treated water following each unit of the three WWTPs.

The endotoxin activities in the influent of the three WWTPs ranged from 135 to 1133 EU/mL, which was due to quite different types of pharmaceutical production processes. As mentioned above, the influent of Plant A came from the process of biological Vitamin C fermentation using acetic acid bacteria (Gram-negative bacteria). Similarly, in Plant C, the influent of a high-concentration equalization basin also had a high endotoxin activity of 2510 EU/mL. The influent high COD concentration in Plant C also came from the process of biological fermentation. The influent low COD concentration in Plant C showed a relatively low concentration at 392 EU/mL, which mainly came from the washing of floors, fermentation tanks and the domestic wastewater of the factory. In Plant B, the endotoxin activity of the influent was the lowest (135 EU/mL), in which plant a chemical synthesis method was used to produce Vitamin A and Vitamin C.

Fate of endotoxins during biological treatment process

As shown in the three WWTPs, an increase in the endotoxin activities in the wastewater of the anaerobic unit and CASS were observed (Figure 2). For example, the endotoxin activities increased from 560 to 28,633 EU/mL during the anaerobic treatment in Plant A, and from 143 to 34,400 EU/mL and from 10,323 to 87,467 EU/mL during CASS in Plant B and C, respectively. Generally, the endotoxin in wastewater originates from Gram-negative bacteria in the treatment units. Lipopolysaccharide comprises about 3.4% of the average content of a Gram-negative bacteria cell (Han et al. 2003), which can be released into the surrounding environment during the growth and lysis period (Gorbet & Sefton 2005). The presence of toxics (such as antibiotics), phage infection, increased temperature, amino acids and other adverse conditions could prompt the production of LPS (Han et al. 2003). A relationship between the viability of Escherichia coli O111 and total LPS under several typical sanitizers also confirmed that E. coli O111 responds to factors that hinder viability by producing more LPS in its outer membrane.

In contrast, during the aerobic treatment, endotoxins were substantially removed. Decrease of endotoxin activities from 34,400 to 4737 EU/mL and from 28,633 to 3657 EU/mL was found during the moving bed bioreactor (MBBR) treatment in Plant B, and the aerobic treatment in plant A, respectively, which is similar to the previous study of Guizani et al. (2009). In their experiments, LAL test results showed that the endotoxin activity decreased after 12 hours aeration in a batch test of rejected wastewater from sludge treatment facilities. Some specialized bacteria (such as H. hydrossis, H. hydrossis-like filaments and Bacteroidetes) in WWTP are involved in degradation of sugars that participate in the conversion of LPS (Kragelund et al. 2008). Saddler et al. (1979) found a strain of slime mold, Physarum polycephalum, with enzymes that could reduce LPS from a variety of bacteria. Furthermore, an extract made from a Neisseria gonorrhoeae strain could degrade the polysaccharide part of LPS produced by itself (Apicella et al. 1978).The protozoa in wastewater treatment have an ability to degrade LPS because the protozoa feed on bacteria (Kamika & Momba 2014). Other than bacteria, mammalian viruses could degrade LPS (Apicella et al. 1978; Saddler et al. 1979).

Increases of endotoxin during DAF treatment process

Dissolved air flotation (DAF) is an important industrial wastewater treatment unit, which separates sludge or particles from mixed liquor using gas bubbles (Zhang et al. 2014). As shown in Figure 2, the endotoxin activities of wastewater increased in the DAF treatment process of Plant B and Plant C.

In Plant B, the endotoxin activity of wastewater increased from 143 to 152 EU/mL during separation of suspended solids in the DAF. However, in Plant C, the endotoxin activity of wastewater in the DAF was increased by one order of magnitude (from 5,980 to 33,864 EU/mL). Higher endotoxin production in the latter DAF was attributed to its role of separating sludge after clarification, which included high levels of bacteria cells. Furthermore, the total activity of endotoxins in water is composed of free-endotoxins and bound-endotoxins (Zhang et al. 2013). Endotoxin, an amphiphilic molecule that has both hydrophobic and hydrophilic groups, may adhere to other suspended particles forming bound-endotoxin. To some extent, DAF unit may remove some bound-endotoxins along with particles or sludges. In addition, some other treatments including soil treatment, hydrophobic membranes and coagulation can help remove endotoxins from pharmaceutical wastewater in accordance with Guizani (2010). In the future, highlights for how to control endotoxin emission from wastewater treatment units need more attention.

Microbial community structure

Microbial community at the phylum level

After denoising, filtering out chimeras, and removing the chloroplast sequences, the library size of each sample was normalized to 21,696 sequences. The 21,696 selected effective bacterial sequences in each sample were assigned to different taxa levels from genus to phylum using the RDP classifier.

Figure 3 shows the analysis of bacterial composition at the phylum level. At the phylum level, a total of 29 bacteria phyla were identified in all samples. Proteobacteria was the most abundant phylum, accounting for 9.8–85.1% of the total effective bacterial sequences, followed by the abundance of Bacteroidetes (7.4–19.1%) and Firmicutes (1.3–4.9%). The total of the three groups was 74.6%, which was similar to a few previous studies on community structure of biological samples in WWTPs (Zhang et al. 2012; Ju et al. 2014). The phyla with an average abundance above 1% included Thermotogae (2.1%), Acidobacteria (1.3%), Spirochaetes (1.3%), and Synergistetes (1.1%). The other phyla with relatively low average abundance include Actinobacteria (0.9%), Chloroflexi (0.7%), Tenericutes (0.2%), Chlorobi (0.4%), and Fusobacteria (0.1%).

Figure 3

Abundance of different phyla in the seven samples. The abundance is presented in terms of percentages in total effective bacterial sequences in a sample. Taxa represented occurred at >0.1% abundance in at least one sample. Other phyla refer to the taxa with their maximum abundance of <0.1% in any sample.

Figure 3

Abundance of different phyla in the seven samples. The abundance is presented in terms of percentages in total effective bacterial sequences in a sample. Taxa represented occurred at >0.1% abundance in at least one sample. Other phyla refer to the taxa with their maximum abundance of <0.1% in any sample.

According to Bergry's Manual of Systematic Bacteriology (Brenner et al., 2007), relative abundance of the Gram-negative bacteria was calculated and is listed in Figure 4. Gram-negative bacteria accounted for 50.0–94.6% of all bacteria. Especially in the CASS, the abundance of Gram-negative bacteria in activated sludge was larger than 90%. As mentioned above, Gram-negative bacteria differ from Gram-positive bacteria in that they have an outer cell wall layer that contains LPS. It has been estimated that a single Gram-negative bacterium contains 3.5 million LPS molecules occupying an area of about 4.9 μm2 (Anderson et al. 2008). LPS molecules are released from bacteria during bacteria growth and multiplication, especially when bacteria die and lysis occurs (Visser et al. 2006). Many Gram-negative bacteria produce LPS-laden bilayered spheres known as membrane vesicles or blebs. The high amount of Gram-negative bacteria in the activated sludge or biofilm could be a source of LPS, which may explain why the endotoxin activities of wastewater had a huge increase while going through the biological treatment units.

Figure 4

Abundance of Gram-negative bacteria and Gram-positive bacteria in the seven samples.

Figure 4

Abundance of Gram-negative bacteria and Gram-positive bacteria in the seven samples.

Analysis of bacterial community at the genus level

At the genus level, a total of 354 operational taxonomic units (OTUs) in seven samples were constructed and a total of 26 major genera (with relative abundance >1%) based on ribosomal database project (RDP) analysis are shown in a heat map (Figure 5). Dendrograms prepared using the Bray-Curtis similarity index are also shown integrated with the heat map. Among the 26 major genera, there were eight genera with an average abundance >1% in all seven samples. Five genera (>1%) such as Tepidiphilus (11.6%), Thauera (5.6%), Hyphomicrobium (4.9%), Desulfovibrio (3.7%), Methyloversatilis (1.0%) belong to the Proteobacteria phyla. Other genera such as Spirochaetes (3.7%), Proteiniphilum (2.2%), and Kosmotoma (2.0%) belong to Spirochaetes, Bacteroidetes and Thermotogae phyla, respectively.

Figure 5

Heat map of genera (abundance >1%) in seven samples. The color bar indicates the range of the percentage of a genus in a sample, based on the color key (log10 scale) at the bottom right corner. Genera with abundance >1% were selected in each sample. Clustering of the cases was based on the genera abundance (>1%) data shown in the top of heat map. The full color version of this figure is available in the online version of this paper, at http://dx.doi.org/10.2166/wst.2018.162.

Figure 5

Heat map of genera (abundance >1%) in seven samples. The color bar indicates the range of the percentage of a genus in a sample, based on the color key (log10 scale) at the bottom right corner. Genera with abundance >1% were selected in each sample. Clustering of the cases was based on the genera abundance (>1%) data shown in the top of heat map. The full color version of this figure is available in the online version of this paper, at http://dx.doi.org/10.2166/wst.2018.162.

LPS share a common architecture but vary considerably in structural motifs from one genus, species and strain to another (Caroff et al. 2002). The biological activity of LPS is determined by its structure, such as the number and chain lengths of the fatty acids in lipid A, the phosphate substituents and the length and chemical composition of its O-antigen components (Wolny et al. 2011; Zhang-Sun et al. 2015). The LAL assay used in our study also shows different responses to LPS from different bacteria.

In our study, several kind of bacteria such as Bacterides and Enterococcus with an LPS that had higher biological activities were found. The Bacteroides showed an average abundance of 0.3% in seven samples. Delahooke et al. (1995) found that in the LAL assay, LPS activities of Bacterides fragilis strains were higher than those of E. coli O18 LPS. The Desulfovibrio, which belong to sulfate-reducing bacteria, have an average abundance of 3.7%. The Desulfovibrio, which were curved and mobile microorganisms, also have higher endotoxin activities because the LPS from Desulfovibrio can induce higher levels of cytokines than E. coli J5 (Wolny et al. 2011; Zhang-Sun et al. 2015). Therefore, endotoxin activities of wastewater in biological treatment units had a relationship with the bacterial community of activated sludge or biofilms.

Correlation of the removal efficiency of a pollutant with bacterial community

To find the impact of bacterial community on the endotoxin distribution discipline at WWTPs, the relationship of the main bacteria with abundance >1% at phylum level and the removal efficiency of pollutant was illustrated by Redundancy analysis (RDA).

As shown in Figure 6, the phyla of Actinobacteria, Chloroflexi and OD1 had positive correlation with the removal rate of endotoxin in wastewater, whereas the phyla of Proteobacteria, Acidobacteria, Bacteroidetes, Thermotogae, Spirochaetes and Synergistetes had the negative correlation with the removal rate of endotoxin in wastewater. According to Bergry's Manual of Systematic Bacteriology (Brenner et al., 2007), the Actinobacteria belong to Gram-positive bacteria, which has only a single plasma membrane without the LPS. Some Gram-negative bacteria such as Chloroflexi and Deinococcales-Thermus have two cellular membranes but also lack LPS in accordance with the studies on bacterial cell envelope architecture (Botero et al., 2004; Sutcliffe, 2010). This indicated that the presence of LPS-lacking bacteria could contribute to high endotoxin removal rate. In general, the formation and release mechanisms of endotoxins are complex in biological process, and further research on the correlation of the endotoxin removal efficiency with bacterial community is needed.

Figure 6

Redundancy analysis of the abundance of typical dominant bacteria at phylum level and the removal efficiency of pollutants in biological treatment units.

Figure 6

Redundancy analysis of the abundance of typical dominant bacteria at phylum level and the removal efficiency of pollutants in biological treatment units.

CONCLUSIONS

The influent of WWTPs in Plant A and C with the process of biological fermentation had relatively higher endotoxin activities of 560 and 1140 EU/mL, while Plant B, using the process of chemical synthesis, had relatively lower endotoxin activity of 135 EU/mL. The effluent endotoxin activities of the three pharmaceutical WWTPs ranged from 3,657 to 33,867 EU/mL. The high amount of Gram-negative bacteria (50.0–94.6%) and the special bacteria with the high-toxicity LPS in the activated sludge or biofilm may be the reason why the effluent showed high endotoxin activity. Furthermore, the endotoxin could be degraded during the aerobic treatment in the three WWTPs, which inspired us that aerobic treatment had advantages to reduce the risk of exposure to endotoxins in water treatment. RDA analysis indicated that at the phylum level the abundance of Actinobacteria, Chloroflexi and OD1 had positive correlations with the removal rate of endotoxin. Considering the toxicity of endotoxin emissions from some specific work sites, such as biological fermentation, DAF and sludge treatment units in pharmaceutical WWTPs, efficient control strategies are suggested, such as designing WWTPs as fully enclosed, collecting and treating the endotoxin and odorous gases together by chemical process in the future.

ACKNOWLEDGEMENTS

This work was supported by the National Science & Technology Support Program of China (No. 2014BAC08B04) and the National Science Foundation of China (No. 2014BAC08B00). We thank Nanjing University, Yixing Environmental Research Institute for technical assistance in this study.

REFERENCES

REFERENCES
Anderson
,
W. B.
,
Slawson
,
R. M.
&
Mayfield
,
C. I.
2002
A review of drinking-water-associated endotoxin, including potential routes of human exposure
.
Canadian Journal of Microbiology
48
(
7
),
567
587
.
Anderson
,
W. B.
,
Mayfield
,
C. I.
&
Huck
,
P. M.
2008
Endotoxin release from biologically active bench-scale drinking water anthracite/sand filters
.
Journal of Water Supply Research and Technology-AQUA
57
(
8
),
585
597
.
Apicella
,
M. A.
,
Breen
,
J. F.
&
Gagliardi
,
N. C.
1978
Degradation of polysaccharide component of Gonococcal lipopolysaccharide by gonococcal and meningococcal sonic extracts
.
Infection and Immunity
20
(
1
),
228
234
.
Botero
,
L. M.
,
Brown
,
K. B.
,
Brumefield
,
S.
,
Burr
,
M.
,
Castenholz
,
R. W.
,
Young
,
M.
&
McDermott
,
T. R.
2004
Thermobaculumterrenum gen. nov., spnov.: a non-phototrophic gram-positive thermophile representing an environmental clone group related to the Chloroflexi (green non-sulfur bacteria) and Thermomicrobia
.
Archives of Microbiology
181
(
4
),
269
277
.
Brenner
,
D. J.
,
Krieg
,
N. R.
&
Staley
,
J. T.
2007
2007 Bergey's Manual of Systematic bacteriology
.
Springer
,
New York, NY, USA
.
Caroff
,
M.
&
Karibian
,
D.
2003
Structure of bacterial lipopolysaccharides
.
Carbohydrate Research
338
(
23
),
2431
2447
.
Caroff
,
M.
,
Karibian
,
D.
,
Cavaillon
,
J. M.
&
Haeffner-Cavaillon
,
N.
2002
Structural and functional analyses of bacterial lipopolysaccharides
.
Microbes and Infection
4
(
9
),
915
926
.
Delahooke
,
D. M.
,
Barclay
,
G. R.
&
Poxton
,
I. R.
1995
A reppraisal of the biological-activity of bacteroides LPS
.
Journal of Medical Microbiology
42
(
2
),
102
112
.
Duquenne
,
P.
,
Ambroise
,
D.
,
Goerner
,
P.
,
Clerc
,
F.
&
Greff-Mirguet
,
G.
2014
Exposure to airborne endotoxins among sewer workers: an exploratory study
.
Annals of Occupational Hygiene
58
(
3
),
283
293
.
EPA of China
2002
Water and Wastewater Monitoring Analysis Method
,
4th edn
.
Chinese Environment Science Publisher
,
Beijing
.
Forestier
,
D.
,
Lecornet
,
E.
,
Mosqueron
,
L.
&
Lambolez
,
L.
2012
Abstract Exposure to bioaerosols for wastewater treatment plant workers: Prioritization of the areas and tasks involving the greatest exposure, and prevention
.
Environnement Risques & Sante
11
(
2
),
137
148
.
Gangamma
,
S.
,
Patil
,
R. S.
&
Mukherji
,
S.
2011
Characterization and proinflammatory response of airborne biological particles from wastewater treatment plants
.
Environmental Science & Technology
45
(
8
),
3282
3287
.
Gorbet
,
M. B.
&
Sefton
,
M. V.
2005
Endotoxin: the uninvited guest
.
Biomaterials
26
(
34
),
6811
6817
.
Guizani
,
M.
2010
Control of endotoxins and their fate during wastewater reclamation
.
PhD thesis report
.
Hokkaido University
,
Japan
.
Guizani
,
M.
,
Dhahbi
,
M.
&
Funamizu
,
N.
2009
Survey on LPS endotoxin in rejected water from sludge treatment facility
.
Journal of Environmental Monitoring
11
(
11
),
1935
1941
.
Han
,
L.
,
Enfors
,
S. O.
&
Haggstrom
,
L.
2003
Escherichia coli high-cell-density culture: carbon mass balances and release of outer membrane components
.
Bioprocess and Biosystems Engineering
25
(
4
),
205
212
.
Kamika
,
I.
&
Momba
,
M. N. B.
2015
Effect of nickel on nutrient removal by selected indigenous protozoan species in wastewater systems
.
Saudi Journal of Biological Sciences
22
(
2
),
147
156
.
Kragelund
,
C.
,
Levantesi
,
C.
,
Borger
,
A.
,
Thelen
,
K.
,
Eikelboom
,
D.
,
Tandoi
,
V.
,
Kong
,
Y.
,
Krooneman
,
J.
,
Larsen
,
P.
,
Thomsen
,
T. R.
&
Nielsen
,
P. H.
2008
Identity, abundance and ecophysiology of filamentous bacteria belonging to the Bacteroidetes present in activated sludge plants
.
Microbiology
154
(
Pt3
),
886
894
.
Paba
,
E.
,
Tranfo
,
G.
,
Corsetti
,
F.
,
Marcelloni
,
A. M.
&
Iavicoli
,
S.
2013
Indoor exposure to airborne endotoxin: a review of the literature on sampling and analysis methods
.
Industrial Health
51
(
3
),
237
255
.
Saddler
,
J. N.
,
Coote
,
J. G.
&
Wardlaw
,
A. C.
1979
Degradation of bacterial lipopolysaccharide by the slime-mold physarum-polycephalum
.
Canadian Journal of Microbiology
25
(
2
),
124
129
.
Visser
,
M. J.
,
Spaan
,
S.
,
Arts
,
H. J.
,
Smit
,
L. A.
&
Heederik
,
D. J.
2006
Influence of different cleaning practices on endotoxin exposure at sewage treatment plants
.
Annals of Occupational Hygiene
50
(
7
),
731
736
.
Wolny
,
D.
,
Lodowska
,
J.
,
Jaworska-Kik
,
M.
,
Kurkiewicz
,
S.
,
Weglarz
,
L.
&
Dzierzewicz
,
Z.
2011
Chemical composition of Desulfovibrio desulfuricans lipid A
.
Archives of Microbiology
193
(
1
),
15
21
.
Zhang
,
C.
,
Liu
,
W.
,
Sun
,
W.
,
Zhang
,
M.
,
Qian
,
L.
,
Li
,
C.
&
Tian
,
F.
2013
Endotoxin contamination and control in surface water sources and a drinking water treatment plant in Beijing, China
.
Water Research
47
(
11
),
3591
3599
.
Zhang
,
Q.
,
Liu
,
S.
,
Yang
,
C.
,
Chen
,
F.
&
Lu
,
S.
2014
Bioreactor consisting of pressurized aeration and dissolved air flotation for domestic wastewater treatment
.
Separation and Purification Technology
138
(
138
),
186
190
.

Author notes

Xuemeng He and Lili Ding contributed equally to all aspects of conceptualizing planning, sample and data collection, data analysis and preparation of the manuscript.