- Email: [email protected]
Use of food waste-recycling wastewater as an alternative carbon source for denitrification process: A full-scale study
Bioresource Technology 245 (2017) 1016–1021
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Use of food waste-recycling wastewater as an alternative carbon source for denitrification process: A full-scale study
MARK
Eunji Kima,1, Seung Gu Shinb,1, Md Abu Hanifa Jannata, Jovale Vincent Tongcoa, Seokhwan Hwanga,⁎ a School of Environmental Science and Engineering, Pohang University of Science and Technology (POSTECH), 77 Cheongam-Ro, Nam-Gu, Pohang, Gyeongbuk, South Korea b Department of Energy Engineering, Gyeongnam National University of Science and Technology (GNTECH), Jinju, South Korea
A R T I C L E I N F O
A B S T R A C T
Keywords: Full-scale Biological nitrogen removal Denitrification External carbon source Next generation sequencing
Using organic wastes as an alternative to commercial carbon sources could be beneficial by reducing costs and environmental impacts. In this study, food waste-recycling wastewater (FRW) was evaluated as an alternative carbon source for biological denitrification over a period of seven months in a full-scale sewage wastewater treatment plant. The denitrification performance was stable with a mean nitrate removal efficiency of 97.2%. Propionate was initially the most persistent volatile fatty acid, but was completely utilized after 19 days. Eubacteriacea, Saprospiraceae, Rhodocyclaceae and Comamonadaceae were the major bacterial families during FRW treatment and were regarded as responsible for hydrolysis (former two) and nitrate removal (latter two) of FRW. These results demonstrate that FRW can be an effective external carbon source; process stabilization was linked to the acclimation and function of bacterial populations to the change of carbon source.
1. Introduction Removal of nitrogen is one of the important objectives in modern wastewater treatment systems. If nitrogen compounds are discharged at high loading without appropriate treatment, they can lead to critical environmental problems such as water pollution and eutrophication. Biological denitrification process is a proven technology to treat nitrite and nitrate-contaminated wastewater. During denitrification in anoxic conditions, nitrates are reduced to harmless nitrogen gas. This process is mediated by heterotrophic bacteria that use organic carbon as an electron donor (Fernández-Nava et al., 2010). Previous studies suggested that to achieve complete denitrification, the influent should have a ratio of carbon to nitrogen (C/N) ≥ 13, and a ratio of influent chemical oxygen demand to nitrogen (COD/N) ≥ 15 in full-scale pre-denitrification (Beccari et al., 1983; KomorowskaKaufman et al., 2006). However, the C/N ratio of wastewater is often lower than these values, so nitrogen removal is limited by the lack of organic carbon. (Sun et al., 2010). Therefore, to achieve complete nitrogen removal, the supply source of organic carbon should be sufficient. In many commercial wastewater treatment plants (WWTPs), external carbon sources such as methanol, ethanol and acetate are used to increase denitrification efficiency (Zhang et al., 2016a). Methanol has
⁎
1
Corresponding author. E-mail address: [email protected] (S. Hwang). These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.biortech.2017.08.168 Received 30 June 2017; Received in revised form 28 August 2017; Accepted 29 August 2017 Available online 01 September 2017 0960-8524/ © 2017 Elsevier Ltd. All rights reserved.
been the most commonly-used external carbon source because of its low cost ($0.33/kg) (Sun et al., 2010) but causes a long lag phase during the start-up of methanol-fed denitrifying system (Fernández-Nava et al., 2010; Nyberg et al., 1992). Ethanol and acetate improve denitrification reaction almost instantaneously when used, but their costs are too high (ethanol: $0.84/kg; acetate: $1.03/kg) (Sun et al., 2010; Zhang et al., 2016a). Thus, many studies have been conducted to test alternative carbon sources for denitrification, such as crude syrup, industrial wastewater from ice cream production or beet-sugar processing, and solid carbon sources like wheat straw and plant prunings (Karanasios et al., 2016; Zhang et al., 2016b). However, the pre-treatment required was complicated and lengthy; furthermore, the number of studies that have been conducted for application in practical wastewater treatment plants is limited (Karanasios et al., 2016). Therefore, an inexpensive and readily-available carbon source is still needed. Food waste-recycling wastewater (FRW) is generated during the recycling of food waste to produce animal feedstock or fertilizer. FRW contains high levels of organic materials, and therefore may be a good carbon source (Shin et al., 2015). The biodegradability of FRW in anaerobic digestion has been estimated to be 90.0% (Shin et al., 2015). Thus, the use of FRW by anaerobic/anoxic heterotrophs in biological denitrification process could achieve organic waste reduction and
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
free water and stored at −20 °C before being used in further molecular analysis.
waste-to-resources conversion. One study has reported successful use of a fermentative product from FRW as a denitrification carbon source in a laboratory-scale batch process (Kim et al., 2016). However, use of FRW in a commercial WWTP has not been validated as an alternative external carbon source in the long term, which should inevitably involve variation in FRW characteristics (coefficient of variation (CV) > 20%) (Shin et al., 2015). In addition, little is known about how the microbial community of heterotrophic denitrifying bacteria responds to use of FRW. Therefore, this study aimed to investigate denitrification performance in a full-scale WWTP that treats sewage wastewater by using FRW as external carbon source. Bacterial populations associated with the biological denitrification process were also investigated in relation to the change of carbon source, because utilization of each organic component in FRW may cause changes in bacterial community structures and growth of specific microorganisms. The relationship between bacterial populations and the use of different carbon sources may improve understanding of the denitrification process.
2.3. Physicochemical analyses The pH of each sample was measured using a benchtop pH meter (Cole Parmer, Vernon Hills, IL). COD and solid (total solids, TS and volatile solids, VS) contents were quantified using Standard Methods (APHA-AWWA-WEF, 2005). Total nitrogen and phosphorus concentrations were determined respectively using the Total Nitrogen HR kit and Total Phosphorus kit (C-MAC, Daejeon, South Korea) based on the Standard Methods. The samples were filtered through Ministar-RC membrane filters (Sartorius, porosity 0.45 μm) for solubility measurements, such as ionic compound and volatile fatty acids (VFAs). NH4+and NO3- were quantified on two identical ion chromatographs using METROSEP C4 250/4.0 Metrohm and a METROSEP A Supp 5 100/4.0 columns (Personal 790 IC, Metrohm, Switzerland). Ethanol and VFAs (C2–C6) were measured using a gas chromatograph (6890 Plus, Agilent, Palo Alto, CA) equipped with an Innowax capillary column (Agilent) and a flame ionization detector. The carrier gas was He at a flow rate of 2.5 mL/min with a split ratio of 10:1. All analytical methods were performed in duplicate, and results are presented as means.
2. Materials and methods 2.1. Full-scale WWTP operation A full-scale WWTP that treats sewage generated by the steel-making industry in Pohang, South Korea was chosen for this study. The biological nitrogen removal system in this facility consists of an AnaerobicAnoxic-Aerobic (A2O) process (Fig. 1), and temperature was maintained at 25 °C by a heating system. The WWTP was designed to treat up to 4000 m3 sewage/d, and treated an average of 2300 m3 sewage/d sewage during this study. It utilized 2 ton/d of a butanol- and methanol-rich commercial product called recovered carbon source 45 (RCS45) as the sole external carbon source until the end of October 2015 (day 0). This WWTP diversified the carbon source by using 8 ton/ d of FRW generated from a local facility that recycles food waste to animal feed, as the main carbon source from November 2015 to May 2016 (days 1–220). Therefore, over a period of seven months, wastewater was collected from the storage tank and the FRW was sampled regularly from the food waste-recycling plant upon transport.
2.4. 16S rRNA sequencing and metagenomics analysis To reveal the bacterial populations in the anoxic tank, the 16S rRNA gene was sequenced using an Ion Torrent PGM™ System (Life Technologies, Carlsbad, USA) according to the manufacturer’s instructions. For each DNA sample, the V4 hypervariable region of the 16S rRNA gene was PCR-amplified (initial denaturation at 95 °C for 5 min, followed by 30 cycles at 95 °C for 30 s, annealing at 57 °C for 30 s, and a final extension 72 °C for 7 min) using the bacterial universal primers 518f (CCAGCAGCCGCGGTAATACG) and 805r GACTACCAGGGT ATCTAAT). Results were barcode-labeled using the Ion Plus fragment library kit (Life Technologies). For clonal amplification of the library, emulsion PCR was performed using an Ion PGM™ Hi-Q™ template kit and a OneTouch™ 2 instrument (Life Technologies). The template-positive ion sphere particles enriched with Ion OneTouch™ ES were loaded on a 316™ chip using the Ion PGM™ Hi-Q™ sequencing kit, and sequenced in the Ion PGM™ sequencer operated by Torrent Suite™ software (version 5.0.4). A total of 278,283 sequence reads was obtained and analyzed by the Ion Reporter™ software with default parameter settings. Ion Reporter software was used to match the results to reference databases (MicroSEQ(R) 16S Reference Library version 2013.1; Greengenes version 13.5). Sequence reads <200 bp were discarded.
2.2. Sampling and DNA extraction Samples were collected three times a week from the A2O process at each of the anaerobic tank, the anoxic tank and the internal recycling port. All samples were distributed in 200-mL sterile plastic bottles, transported to the laboratory within 1 h and stored in a refrigerator at 4 °C. The total genomic DNA was extracted from the anoxic tank samples in duplicate using an automated nucleic acid extractor (Magtration System 6GC, Precision System Science, Chiba, Japan). Before extraction of genomic DNA, the raw samples were centrifuged at 16,000g for 5 min, then washed three times to remove cell debris or residual wastewater materials. The purified DNA was eluted with 100 μl nuclease-
2.5. Statistical analysis Non-metric multidimensional scaling (NMDS) was performed based on the 16S rRNA gene quantification results sequenced by Ion Torrent
Fig. 1. Plug-flow of the biological nitrogen removal process.
1017
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
content in the influent. In the FRW, the average COD was 51.7 ± 17.3 g/L and the average VS was 14.7 ± 7.0 g/L during this study period (Table 1). These values are lower than those in other studies (COD: 136.3–148.7 g/L, VS: 82.0–104.9 g/L) (Kim et al., 2013; Shin et al., 2015), because the FRW used in this study was diluted during the process of recycling food waste. Nevertheless, the FRW contained considerable organic contents, and a high VS/TS ratio of 60.2%, which is suitable for bioconversion in bioprocesses. The average COD concentration was 597.7 g/L in RCS45 and 51.7 g/L in and FRW. Therefore, about 1 ton of RCS45 can be replaced by 11 tons of FRW. Specifically, the FRW contained 2.1 ± 5.2 g carbohydrate/L (14.2% of VS), 5.4 ± 6.0 g protein/L (35.1%), and 3.5 ± 5.6 g lipid/L (23.7%), where VS includes both suspended and liquid phases. This result indicates that more than half of the organics in the FRW are carbohydrate, protein, or lipid, which are likely biodegradable in anoxic/anaerobic conditions (Shin et al., 2015). The RCS45 consisted of 65.9% water; and included organics such as 1-butyl alcohol (18.1%), methyl alcohol (9.9%), 1,4-butanediol (5.8%), n-propyl alcohol (0.2%) and acetic acid (0.1%) by weight. In contrast, FRW contained ethanol and VFAs, which can be easily utilized by heterotrophic denitrifying bacteria because these chemicals are used directly in bacterial metabolism (Zhang et al., 2016a).
Table 1 Physical and chemical characteristics of the wastewater and FRW. Parameter
pH COD TS VS VS/TS SS VSS Carbohydrate Protein Lipid TN TP TVFA + Ethanol NH4-N NO3-N NO2-N a
Units
Value
– g-CODcr/L g/L g/L g/L g/L g/L g/L g/L g/L g/L g/L g/L mg/L mg/L mg/L
a
Wastewater
FRW
7.2 0.16 ± 0.03 0.3 ± 0.1 0.13 ± 0.03 – 0.08 ± 0.01 0.08 ± 0.01 0.01 ± 0.01 0.04 ± 0.02 0.11 ± 0.01 0.02 ± 0.01 0.01 ± 0.01 0.0 ± 0.0 12.8 ± 1.7 0.32 ± 0.1 0.31 ± 0.1
5.9 51.7 ± 17.3 24.6 ± 8.2 14.8 ± 7.0 60.2 7.9 ± 8.9 7.2 ± 8.0 2.1 ± 5.2 5.4 ± 6.0 3.5 ± 5.6 1.9 ± 1.0 0.5 ± 0.4 11.7 ± 5.3 596.6 ± 219.6 0.1 ± 0.2 1.3 ± 2.0
Average of 120 tests ± standard deviation.
PGM. NMDS ordination using Sorensen (Bray-Curtis) distance was conducted to visualize how the similarity of bacteria community structures in the anoxic tank changed over time. Each main matrix was processed for ordination such that the stress (<10) and instability (<10−4) criteria were met (McCune and Grace, 2002). NMDS analysis was conducted in PC-ORD 5.0 (MjM software, OR, USA).
3.2. Denitrification performance of the full-scale plant The denitrification efficiency in the anoxic tank was 100% when RCS45 was used as a carbon source (day 0), and was stable when FRW was used as the external carbon source for seven months (days 1–220) (Fig. 2a). Although the nitrate loading increased during the operation period, the mean denitrification efficiency first increased to 97.2%, then increased to 99.3% after 98 days. Ethanol and VFAs such as acetate and propionate are organic materials found in FRW. Although they exist as a small fraction of the COD, they could be used for biological denitrification as energy and carbon sources. Easily-biodegradable COD is a requirement for treating nitrate and ethanol, and VFAs can be used as a more efficient carbon source compared to other organic compounds. Propionate was a more persistent VFA than ethanol and acetate in the initial period (Fig. 2b). This trend indicates that the microorganisms in the anoxic tank preferred ethanol and acetate over propionate. This preference can be linked to the ease in which certain carbon substrates can be assimilated into metabolic pathways. Ethanol is converted by bacterial cells to acetylCoA, which enters the tricarboxylic acid cycle. Acetate has a simple pathway because it is degraded directly by the β-oxidation process to
3. Results and discussion 3.1. Physicochemical characteristics of the wastewater and alternative carbon source The physicochemical characteristics of the wastewater and FRW used in this study are varied over time (Table 1). The wastewater had COD/N ≈ 6.8, which is much less than the recommended value of 15. To achieve complete nitrogen removal, the full-scale WWTP used in this study needed an external carbon source (RCS45). The average COD was 0.16 ± 0.03 g/L and the average VS was 0.13 ± 0.03 g/L; these results indicate that the FRW has lower organic content than domestic sewage or municipal wastewater (Chan et al., 2009). Also, the CV of COD and VS in wastewater were 18.8% and 23.1%, respectively; these small values indicate that the wastewater has minor variation of organic
Fig. 2. Nitrate-N removal efficiencies (a) and TVFA + Ethanol residual mass (b) in the anoxic tank.
1018
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
study the relative abundance of Flavobacteriaceae significantly decreased from 7.2% on day 0 to 0.2% on day 47 after the external carbon source was changed, but then the relative abundance of Flavobacteriaceae gradually increased to 5.4% on day 185. Also, the abundance of Neisseriacea increased between day 96 and day 203, and accounted for 3.5–7.3% (average 5.7%) of reads, However, Neisseriaceae only appeared at very low level (0.02%) on day 47. Members of Neisseriaceae, such as Aquaspirillum and Chromobacterium, were reported as potential denitrifying bacteria in biological wastewater treatment systems (Zumft, 1997). The most dominant group changed from Saprospiraceae to Rhodocyclaceae between days 185 and 293. Also, the abundance of Comamonadaceae increased to 8.4% and 11.2%, respectively. These trends are similar to previous reports regarding WWTP, which revealed members of Rhodocyclaceae, such as Zoogloea, Azoarcus and Thauera, and also members of Comamonadaceae, such as Acidovorax, Comamonas, Diaphorobacter and Simplicispira as dominant denitrifying bacteria (Nielsen et al., 2009; Willems, 2014). Bacteria in these groups reduce nitrate, and some members of Rhodocyclaceae and Comamonadaceae can use a variety of carbon sources including alcohols, organic acids, and amino acids. Addition of a dose of ethanol or acetate to denitrifying reactors favors growth of specific bacterial genera such as Acidovorax, Comamonas, Azoarcus, and Thauera species within the families Comamonadaceae and Rhodocyclaceae (Lu et al., 2014); this result means that bacteria in these genera are responsible for the stable and high denitrification efficiency throughout the reactor operation by using carbon sources including ethanol and VFAs. Therefore, the changes in population may be associated with their ability to use ethanol and VFAs (Fig. 2b). Understanding of the bacterial community structure obtained in this study can provide useful information for start-up of a denitrification process. Results also suggest that inoculum sources can be compared by understanding how microbial communities respond to them, and that this understanding can be used to minimize disturbance of the communities. However, further studies on how specific microorganisms can dominate, and direct evidence of how they affect nitrate removal are recommended.
acetyl-CoA (Elefsiniotis and Li, 2006). Furthermore, nitrogen removal efficiency increases immediately after acetate addition (Isaacs and Henze, 1995). Presumably, metabolization of propionate entails pathways more complex than those used for ethanol and acetate (Constantin and Fick, 1997; Pereira et al., 2004; Yatong, 1996) or entails several sequential β-oxidation reactions. 3.3. Bacterial population identified by ion torrent PGM system To monitor the dynamics of the bacterial population, seven samples were collected from the anoxic tank and analyzed using Ion Torrent PGM System. The seven samples were collected on days 0, 47, 96, 131, 164, 185 and 203; the first sample was taken before the use of FRW; the other six were taken to track temporal bacterial shifts after the carbon source was changed. Analysis using Ion Reporter™ software yielded 278,283 sequence reads. The reads were classified into bacterial 285 families at a 92% sequence identity threshold. Among them, only 29 families had relative abundance >0.5% in all bacterial reads. At the family level, these 29 dominant groups accounted for 0–37.5% of the classified sequences (Fig. 3). Families Chitinophagaceae (10.6 ± 3.7%), Xanthomonadaceae (4.6 ± 3.2%), Hyphomicrobiaceae (2.2 ± 1.8%) and Planctomycetaceae (1.8 ± 1.7%) were detected in all seven samples. However, at day 0, in which RCS45 was utilized as sole carbon source, Methylococcaceae was dominant (12.0%; Fig. 3). Methylococcaceae are type-I methanotrophs that belong to class Gammaproteobacteria, which can use methane and methanol as sole carbon and energy sources but cannot use substrates that contain carbon-carbon bonds (Bowman, 2014). Therefore, Methylococcaceae contributed to nitrate removal by using the methanol in the RCS45. When the FRW was used, the most abundant groups changed during the operating period. First, Eubacteriacea (37.5%) and Saprospiracea (36.9%) were the most dominant groups in day 47. The relative abundance of Prevotellaceae increased from 0.06% on day 0 to 1.1% on day 47. Bacteria in these families been identified as hydrolysers. These changes indicate that bacterial populations were mainly affected by the composition of FRW, such as protein and carbohydrate. Under anaerobic conditions, members of Eubacteriaceae degrade cellulose, which is a major component of organic waste (Keating et al., 2012). The Saprospiraceae form one of the core microbial communities in WWTPs, in which high abundance of Saprospiraceae improves the hydrolysis of complex organic compounds such as carbon polymers, particularly protein, during wastewater treatment (Chen et al., 2016). Members of Provotellaceae are saccharolytic anaerobic bacteria, presumably involved in the hydrolysis and acidogenesis of carbohydrates (Maspolim et al., 2015). Provotellaceae species have been linked to the breakdown of proteins and carbohydrates, and produce cellulolytic enzymes such as carboxymethyl cellulase and xylanase (Nyonyo et al., 2014). Microbial hydrolysis releases simple carbon substrates that may supply carbon for organisms involved in nutrient removal. This hydrolysis may be the rate-limiting step for these important processes (Morgenroth et al., 2002). Although hydrolyzing bacteria such as Eubacteriacea, Saprospiracea and Provotellaceae cannot use complex carbon sources like FRW directly to remove nitrogen, their presence can significantly affect the growth and activity of aerobic and denitrifying bacteria. Although Saprospiracea remained the most dominant group until the day 164, the relative abundance of other groups increased as they adapted to the conditions. For example, a previous study has shown that Flavobacteriaceae was dominant population in the beginning, gradually decline as environment changed, but became a dominant group in the system over time (Zhang et al., 2013). Similarly, in this
3.4. Temporal variation in bacterial community structure Changes in bacterial community structure were visualized by NMDS analysis based on the relative abundances of each families, because this method can detect relationships among variables, and it is the most generally-effective ordination method for ecological community data (McCune and Grace, 2002). For the plot, the NMDS result had cumulative R2 = 0.89, which explained 82.0% of the cumulative variance in metagenomics community structures (Fig. 4a). Also, it showed 3.48 final stress and 0.00002 final instability, which indicate that its accuracy was acceptable (McCune and Grace, 2002). The NMDS result indicated that bacterial community structure shifted continuously over the period, and that significant shifts of bacterial community structure occurred three times: from day 0 to day 47, from day 47 to day 164, and from day 164 to day 185 (Fig.4a). The first shift from day 0 to day 47 was from Methylococcacea to Eubacteriacea and Saprospiraceae. This shift is mainly explained by the change of carbon source from RCS45 to FRW. This result concurs with a previous study in which a change of carbon source was one of the factors that affected microbial community structure during biological wastewater treatment (Lu et al., 2014). The second shift from day 47 to day 164 (cluster 1) was a suppression of Eubacteriacea and Saprospiraceae (Fig. 4b). This observation suggests that if specific microorganisms are suppressed without
1019
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
Fig. 3. Bacteria populations at the family level in the anoxic tank, as identified using Ion Torrent PGM next-generation sequencing. Bacteria populations with relative abundance <0.5% categorized as Minor groups.
Fig. 4. The bacterial profiles based on relative abundance at family level: (a) NMDS ordination plot and (b) Cluster dendrogram. Black arrows: shifts in community structure. Yellow dotted line: 80% of information remaining criterion. d: day. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Acknowledgements
any change of environment, their functions could be replaced by other microorganisms without major reduction in process efficiency in this system. The third shift from day 164 to day 185 was a dramatic increase in Rhodocyclaceae. This shift is mainly explained by adaption of the bacteria community like denitrifying bacteria.
This work was financially supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (No. 20163030091540). This work was also supported by ‘Human Resources Program in Energy Technology’ of the KETEP Grant, funded by the MOTIE, Republic of Korea (No. 20144030200460).
4. Conclusions References Stable denitrification performance over a period of seven months was achieved in a full-scale WWTP using FRW as alternative carbon source. Propionate was the most persistent VFA at the initial period, but was completely utilized after 19 days. Increase in the relative abundance of Eubacteriacea, Saprospiraceae, Rhodocyclaceae and Comamonadaceae after change to FRW suggests that bacterial acclimation to new carbon source may be responsible for the stability. Further study to elucidate how specific bacterial communities process different organic components would be valuable to improve process control.
APHA-AWWA-WEF, 2005. Standard Methods for the Examination of Water and Wastewater, 21th ed. American Public Health Association, Washington, D.C. Beccari, M., Passino, R., Ramadori, R., Tandoi, V., 1983. Kinetics of dissimilatory nitrate and nitrite reduction in suspended growth culture. J. Water Pollut. Control Federation 58–64. Bowman, J.P. 2014. The family Methylococcaceae. In: The Prokaryotes, Springer, pp. 411–440. Chan, Y.J., Chong, M.F., Law, C.L., Hassell, D.G., 2009. A review on anaerobic–aerobic treatment of industrial and municipal wastewater. Chem. Eng. J. 155 (1–2), 1–18. Chen, Y., Zhao, Z., Peng, Y., Li, J., Xiao, L., Yang, L., 2016. Performance of a full-scale modified anaerobic/anoxic/oxic process: High-throughput sequence analysis of its microbial structures and their community functions. Bioresour. Technol. 220, 225–232. Constantin, H., Fick, M., 1997. Influence of C-sources on the denitrification rate of a high-
1020
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
Sci. Technol. 45 (6), 25–40. Nielsen, P.H., Daims, H., Lemmer, H., 2009. FISH handbook for biological wastewater treatment. Water Intelligence Online 8 9781780401775. Nyberg, U., Aspegren, H., Andersson, B., Jansen, J.l.C., Villadsen, I., 1992. Full-scale application of nitrogen removal with methanol as carbon source. Water Sci. Technol. 26 (5–6), 1077–1086. Nyonyo, T., Shinkai, T., Mitsumori, M., 2014. Improved culturability of cellulolytic rumen bacteria and phylogenetic diversity of culturable cellulolytic and xylanolytic bacteria newly isolated from the bovine rumen. FEMS Microbiol. Ecol. 88 (3), 528–537. Pereira, M., Sousa, D., Mota, M., Alves, M., 2004. Mineralization of LCFA associated with anaerobic sludge: kinetics, enhancement of methanogenic activity, and effect of VFA. Biotechnol. Bioeng. 88 (4), 502–511. Shin, S.G., Han, G., Lee, J., Cho, K., Jeon, E.-J., Lee, C., Hwang, S., 2015. Characterization of food waste-recycling wastewater as biogas feedstock. Bioresour. Technol. 196, 200–208. Sun, S.-P., Nàcher, C.P.i., Merkey, B., Zhou, Q., Xia, S.-Q., Yang, D.-H., Sun, J.-H., Smets, B.F., 2010. Effective biological nitrogen removal treatment processes for domestic wastewaters with low C/N ratios: a review. Environ. Eng. Sci. 27 (2), 111–126. Willems, A., 2014. The family Comamonadaceae. In: The Prokaryotes, Springer, pp. 777–851. Yatong, X., 1996. Volatile fatty acids carbon source for biological denitrification. J. Environ. Sci. 8 (3), 257–269. Zhang, C.S., Tang, J.H., Zhang, K.F., Rong, H.W., He, J.M., 2013. Start-up of denitrifying dephosphatation with nitrite as electron acceptor and microbial community analysis. Adv. Mater. Res.. Trans Tech Publ. 280–285. Zhang, H., Jiang, J., Li, M., Yan, F., Gong, C., Wang, Q., 2016a. Biological nitrate removal using a food waste-derived carbon source in synthetic wastewater and real sewage. J. Environ. Manage. 166, 407–413. Zhang, Y., Wang, X.C., Cheng, Z., Li, Y., Tang, J., 2016b. Effect of fermentation liquid from food waste as a carbon source for enhancing denitrification in wastewater treatment. Chemosphere 144, 689–696. Zumft, W.G., 1997. Cell biology and molecular basis of denitrification. Microbiol. Mol. Biol. Rev. 61 (4), 533–616.
nitrate concentrated industrial wastewater. Water Res. 31 (3), 583–589. Elefsiniotis, P., Li, D., 2006. The effect of temperature and carbon source on denitrification using volatile fatty acids. Biochem. Eng. J. 28 (2), 148–155. Fernández-Nava, Y., Marañón, E., Soons, J., Castrillón, L., 2010. Denitrification of high nitrate concentration wastewater using alternative carbon sources. J. Hazard. Mater. 173 (1–3), 682–688. Isaacs, S.H., Henze, M., 1995. Controlled carbon source addition to an alternating nitrification-denitrification wastewater treatment process including biological P removal. Water Res. 29 (1), 77–89. Karanasios, K.A., Vasiliadou, I.A., Tekerlekopoulou, A.G., Akratos, C.S., Pavlou, S., Vayenas, D.V., 2016. Effect of C/N ratio and support material on heterotrophic denitrification of potable water in bio-filters using sugar as carbon source. Int. Biodeterior. Biodegradation 111, 62–73. Keating, C., Cysneiros, D., Mahony, T., O'Flaherty, V., 2012. The hydrolysis and biogas production of complex cellulosic substrates using three anaerobic biomass sources. Water Sci. Technol. 67 (2), 293–298. Kim, D.-H., Lee, M.-K., Jung, K.-W., Kim, M.-S., 2013. Alkali-treated sewage sludge as a seeding source for hydrogen fermentation of food waste leachate. Int. J. Hydrogen Energy 38 (35), 15751–15756. Kim, H., Kim, J., Shin, S.G., Hwang, S., Lee, C., 2016. Continuous fermentation of food waste leachate for the production of volatile fatty acids and potential as a denitrification carbon source. Bioresour. Technol. 207, 440–445. Komorowska-Kaufman, M., Majcherek, H., Klaczyński, E., 2006. Factors affecting the biological nitrogen removal from wastewater. Process Biochem. 41 (5), 1015–1021. Lu, H., Chandran, K., Stensel, D., 2014. Microbial ecology of denitrification in biological wastewater treatment. Water Res. 64, 237–254. Maspolim, Y., Zhou, Y., Guo, C., Xiao, K., Ng, W.J., 2015. Determination of the archaeal and bacterial communities in two-phase and single-stage anaerobic systems by 454 pyrosequencing. J. Environ. Sci. 36, 121–129. McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach, OR. Morgenroth, E., Kommedal, R., Harremoës, P., 2002. Processes and modeling of hydrolysis of particulate organic matter in aerobic wastewater treatment–a review. Water
1021
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Use of food waste-recycling wastewater as an alternative carbon source for denitrification process: A full-scale study
MARK
Eunji Kima,1, Seung Gu Shinb,1, Md Abu Hanifa Jannata, Jovale Vincent Tongcoa, Seokhwan Hwanga,⁎ a School of Environmental Science and Engineering, Pohang University of Science and Technology (POSTECH), 77 Cheongam-Ro, Nam-Gu, Pohang, Gyeongbuk, South Korea b Department of Energy Engineering, Gyeongnam National University of Science and Technology (GNTECH), Jinju, South Korea
A R T I C L E I N F O
A B S T R A C T
Keywords: Full-scale Biological nitrogen removal Denitrification External carbon source Next generation sequencing
Using organic wastes as an alternative to commercial carbon sources could be beneficial by reducing costs and environmental impacts. In this study, food waste-recycling wastewater (FRW) was evaluated as an alternative carbon source for biological denitrification over a period of seven months in a full-scale sewage wastewater treatment plant. The denitrification performance was stable with a mean nitrate removal efficiency of 97.2%. Propionate was initially the most persistent volatile fatty acid, but was completely utilized after 19 days. Eubacteriacea, Saprospiraceae, Rhodocyclaceae and Comamonadaceae were the major bacterial families during FRW treatment and were regarded as responsible for hydrolysis (former two) and nitrate removal (latter two) of FRW. These results demonstrate that FRW can be an effective external carbon source; process stabilization was linked to the acclimation and function of bacterial populations to the change of carbon source.
1. Introduction Removal of nitrogen is one of the important objectives in modern wastewater treatment systems. If nitrogen compounds are discharged at high loading without appropriate treatment, they can lead to critical environmental problems such as water pollution and eutrophication. Biological denitrification process is a proven technology to treat nitrite and nitrate-contaminated wastewater. During denitrification in anoxic conditions, nitrates are reduced to harmless nitrogen gas. This process is mediated by heterotrophic bacteria that use organic carbon as an electron donor (Fernández-Nava et al., 2010). Previous studies suggested that to achieve complete denitrification, the influent should have a ratio of carbon to nitrogen (C/N) ≥ 13, and a ratio of influent chemical oxygen demand to nitrogen (COD/N) ≥ 15 in full-scale pre-denitrification (Beccari et al., 1983; KomorowskaKaufman et al., 2006). However, the C/N ratio of wastewater is often lower than these values, so nitrogen removal is limited by the lack of organic carbon. (Sun et al., 2010). Therefore, to achieve complete nitrogen removal, the supply source of organic carbon should be sufficient. In many commercial wastewater treatment plants (WWTPs), external carbon sources such as methanol, ethanol and acetate are used to increase denitrification efficiency (Zhang et al., 2016a). Methanol has
⁎
1
Corresponding author. E-mail address: [email protected] (S. Hwang). These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.biortech.2017.08.168 Received 30 June 2017; Received in revised form 28 August 2017; Accepted 29 August 2017 Available online 01 September 2017 0960-8524/ © 2017 Elsevier Ltd. All rights reserved.
been the most commonly-used external carbon source because of its low cost ($0.33/kg) (Sun et al., 2010) but causes a long lag phase during the start-up of methanol-fed denitrifying system (Fernández-Nava et al., 2010; Nyberg et al., 1992). Ethanol and acetate improve denitrification reaction almost instantaneously when used, but their costs are too high (ethanol: $0.84/kg; acetate: $1.03/kg) (Sun et al., 2010; Zhang et al., 2016a). Thus, many studies have been conducted to test alternative carbon sources for denitrification, such as crude syrup, industrial wastewater from ice cream production or beet-sugar processing, and solid carbon sources like wheat straw and plant prunings (Karanasios et al., 2016; Zhang et al., 2016b). However, the pre-treatment required was complicated and lengthy; furthermore, the number of studies that have been conducted for application in practical wastewater treatment plants is limited (Karanasios et al., 2016). Therefore, an inexpensive and readily-available carbon source is still needed. Food waste-recycling wastewater (FRW) is generated during the recycling of food waste to produce animal feedstock or fertilizer. FRW contains high levels of organic materials, and therefore may be a good carbon source (Shin et al., 2015). The biodegradability of FRW in anaerobic digestion has been estimated to be 90.0% (Shin et al., 2015). Thus, the use of FRW by anaerobic/anoxic heterotrophs in biological denitrification process could achieve organic waste reduction and
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
free water and stored at −20 °C before being used in further molecular analysis.
waste-to-resources conversion. One study has reported successful use of a fermentative product from FRW as a denitrification carbon source in a laboratory-scale batch process (Kim et al., 2016). However, use of FRW in a commercial WWTP has not been validated as an alternative external carbon source in the long term, which should inevitably involve variation in FRW characteristics (coefficient of variation (CV) > 20%) (Shin et al., 2015). In addition, little is known about how the microbial community of heterotrophic denitrifying bacteria responds to use of FRW. Therefore, this study aimed to investigate denitrification performance in a full-scale WWTP that treats sewage wastewater by using FRW as external carbon source. Bacterial populations associated with the biological denitrification process were also investigated in relation to the change of carbon source, because utilization of each organic component in FRW may cause changes in bacterial community structures and growth of specific microorganisms. The relationship between bacterial populations and the use of different carbon sources may improve understanding of the denitrification process.
2.3. Physicochemical analyses The pH of each sample was measured using a benchtop pH meter (Cole Parmer, Vernon Hills, IL). COD and solid (total solids, TS and volatile solids, VS) contents were quantified using Standard Methods (APHA-AWWA-WEF, 2005). Total nitrogen and phosphorus concentrations were determined respectively using the Total Nitrogen HR kit and Total Phosphorus kit (C-MAC, Daejeon, South Korea) based on the Standard Methods. The samples were filtered through Ministar-RC membrane filters (Sartorius, porosity 0.45 μm) for solubility measurements, such as ionic compound and volatile fatty acids (VFAs). NH4+and NO3- were quantified on two identical ion chromatographs using METROSEP C4 250/4.0 Metrohm and a METROSEP A Supp 5 100/4.0 columns (Personal 790 IC, Metrohm, Switzerland). Ethanol and VFAs (C2–C6) were measured using a gas chromatograph (6890 Plus, Agilent, Palo Alto, CA) equipped with an Innowax capillary column (Agilent) and a flame ionization detector. The carrier gas was He at a flow rate of 2.5 mL/min with a split ratio of 10:1. All analytical methods were performed in duplicate, and results are presented as means.
2. Materials and methods 2.1. Full-scale WWTP operation A full-scale WWTP that treats sewage generated by the steel-making industry in Pohang, South Korea was chosen for this study. The biological nitrogen removal system in this facility consists of an AnaerobicAnoxic-Aerobic (A2O) process (Fig. 1), and temperature was maintained at 25 °C by a heating system. The WWTP was designed to treat up to 4000 m3 sewage/d, and treated an average of 2300 m3 sewage/d sewage during this study. It utilized 2 ton/d of a butanol- and methanol-rich commercial product called recovered carbon source 45 (RCS45) as the sole external carbon source until the end of October 2015 (day 0). This WWTP diversified the carbon source by using 8 ton/ d of FRW generated from a local facility that recycles food waste to animal feed, as the main carbon source from November 2015 to May 2016 (days 1–220). Therefore, over a period of seven months, wastewater was collected from the storage tank and the FRW was sampled regularly from the food waste-recycling plant upon transport.
2.4. 16S rRNA sequencing and metagenomics analysis To reveal the bacterial populations in the anoxic tank, the 16S rRNA gene was sequenced using an Ion Torrent PGM™ System (Life Technologies, Carlsbad, USA) according to the manufacturer’s instructions. For each DNA sample, the V4 hypervariable region of the 16S rRNA gene was PCR-amplified (initial denaturation at 95 °C for 5 min, followed by 30 cycles at 95 °C for 30 s, annealing at 57 °C for 30 s, and a final extension 72 °C for 7 min) using the bacterial universal primers 518f (CCAGCAGCCGCGGTAATACG) and 805r GACTACCAGGGT ATCTAAT). Results were barcode-labeled using the Ion Plus fragment library kit (Life Technologies). For clonal amplification of the library, emulsion PCR was performed using an Ion PGM™ Hi-Q™ template kit and a OneTouch™ 2 instrument (Life Technologies). The template-positive ion sphere particles enriched with Ion OneTouch™ ES were loaded on a 316™ chip using the Ion PGM™ Hi-Q™ sequencing kit, and sequenced in the Ion PGM™ sequencer operated by Torrent Suite™ software (version 5.0.4). A total of 278,283 sequence reads was obtained and analyzed by the Ion Reporter™ software with default parameter settings. Ion Reporter software was used to match the results to reference databases (MicroSEQ(R) 16S Reference Library version 2013.1; Greengenes version 13.5). Sequence reads <200 bp were discarded.
2.2. Sampling and DNA extraction Samples were collected three times a week from the A2O process at each of the anaerobic tank, the anoxic tank and the internal recycling port. All samples were distributed in 200-mL sterile plastic bottles, transported to the laboratory within 1 h and stored in a refrigerator at 4 °C. The total genomic DNA was extracted from the anoxic tank samples in duplicate using an automated nucleic acid extractor (Magtration System 6GC, Precision System Science, Chiba, Japan). Before extraction of genomic DNA, the raw samples were centrifuged at 16,000g for 5 min, then washed three times to remove cell debris or residual wastewater materials. The purified DNA was eluted with 100 μl nuclease-
2.5. Statistical analysis Non-metric multidimensional scaling (NMDS) was performed based on the 16S rRNA gene quantification results sequenced by Ion Torrent
Fig. 1. Plug-flow of the biological nitrogen removal process.
1017
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
content in the influent. In the FRW, the average COD was 51.7 ± 17.3 g/L and the average VS was 14.7 ± 7.0 g/L during this study period (Table 1). These values are lower than those in other studies (COD: 136.3–148.7 g/L, VS: 82.0–104.9 g/L) (Kim et al., 2013; Shin et al., 2015), because the FRW used in this study was diluted during the process of recycling food waste. Nevertheless, the FRW contained considerable organic contents, and a high VS/TS ratio of 60.2%, which is suitable for bioconversion in bioprocesses. The average COD concentration was 597.7 g/L in RCS45 and 51.7 g/L in and FRW. Therefore, about 1 ton of RCS45 can be replaced by 11 tons of FRW. Specifically, the FRW contained 2.1 ± 5.2 g carbohydrate/L (14.2% of VS), 5.4 ± 6.0 g protein/L (35.1%), and 3.5 ± 5.6 g lipid/L (23.7%), where VS includes both suspended and liquid phases. This result indicates that more than half of the organics in the FRW are carbohydrate, protein, or lipid, which are likely biodegradable in anoxic/anaerobic conditions (Shin et al., 2015). The RCS45 consisted of 65.9% water; and included organics such as 1-butyl alcohol (18.1%), methyl alcohol (9.9%), 1,4-butanediol (5.8%), n-propyl alcohol (0.2%) and acetic acid (0.1%) by weight. In contrast, FRW contained ethanol and VFAs, which can be easily utilized by heterotrophic denitrifying bacteria because these chemicals are used directly in bacterial metabolism (Zhang et al., 2016a).
Table 1 Physical and chemical characteristics of the wastewater and FRW. Parameter
pH COD TS VS VS/TS SS VSS Carbohydrate Protein Lipid TN TP TVFA + Ethanol NH4-N NO3-N NO2-N a
Units
Value
– g-CODcr/L g/L g/L g/L g/L g/L g/L g/L g/L g/L g/L g/L mg/L mg/L mg/L
a
Wastewater
FRW
7.2 0.16 ± 0.03 0.3 ± 0.1 0.13 ± 0.03 – 0.08 ± 0.01 0.08 ± 0.01 0.01 ± 0.01 0.04 ± 0.02 0.11 ± 0.01 0.02 ± 0.01 0.01 ± 0.01 0.0 ± 0.0 12.8 ± 1.7 0.32 ± 0.1 0.31 ± 0.1
5.9 51.7 ± 17.3 24.6 ± 8.2 14.8 ± 7.0 60.2 7.9 ± 8.9 7.2 ± 8.0 2.1 ± 5.2 5.4 ± 6.0 3.5 ± 5.6 1.9 ± 1.0 0.5 ± 0.4 11.7 ± 5.3 596.6 ± 219.6 0.1 ± 0.2 1.3 ± 2.0
Average of 120 tests ± standard deviation.
PGM. NMDS ordination using Sorensen (Bray-Curtis) distance was conducted to visualize how the similarity of bacteria community structures in the anoxic tank changed over time. Each main matrix was processed for ordination such that the stress (<10) and instability (<10−4) criteria were met (McCune and Grace, 2002). NMDS analysis was conducted in PC-ORD 5.0 (MjM software, OR, USA).
3.2. Denitrification performance of the full-scale plant The denitrification efficiency in the anoxic tank was 100% when RCS45 was used as a carbon source (day 0), and was stable when FRW was used as the external carbon source for seven months (days 1–220) (Fig. 2a). Although the nitrate loading increased during the operation period, the mean denitrification efficiency first increased to 97.2%, then increased to 99.3% after 98 days. Ethanol and VFAs such as acetate and propionate are organic materials found in FRW. Although they exist as a small fraction of the COD, they could be used for biological denitrification as energy and carbon sources. Easily-biodegradable COD is a requirement for treating nitrate and ethanol, and VFAs can be used as a more efficient carbon source compared to other organic compounds. Propionate was a more persistent VFA than ethanol and acetate in the initial period (Fig. 2b). This trend indicates that the microorganisms in the anoxic tank preferred ethanol and acetate over propionate. This preference can be linked to the ease in which certain carbon substrates can be assimilated into metabolic pathways. Ethanol is converted by bacterial cells to acetylCoA, which enters the tricarboxylic acid cycle. Acetate has a simple pathway because it is degraded directly by the β-oxidation process to
3. Results and discussion 3.1. Physicochemical characteristics of the wastewater and alternative carbon source The physicochemical characteristics of the wastewater and FRW used in this study are varied over time (Table 1). The wastewater had COD/N ≈ 6.8, which is much less than the recommended value of 15. To achieve complete nitrogen removal, the full-scale WWTP used in this study needed an external carbon source (RCS45). The average COD was 0.16 ± 0.03 g/L and the average VS was 0.13 ± 0.03 g/L; these results indicate that the FRW has lower organic content than domestic sewage or municipal wastewater (Chan et al., 2009). Also, the CV of COD and VS in wastewater were 18.8% and 23.1%, respectively; these small values indicate that the wastewater has minor variation of organic
Fig. 2. Nitrate-N removal efficiencies (a) and TVFA + Ethanol residual mass (b) in the anoxic tank.
1018
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
study the relative abundance of Flavobacteriaceae significantly decreased from 7.2% on day 0 to 0.2% on day 47 after the external carbon source was changed, but then the relative abundance of Flavobacteriaceae gradually increased to 5.4% on day 185. Also, the abundance of Neisseriacea increased between day 96 and day 203, and accounted for 3.5–7.3% (average 5.7%) of reads, However, Neisseriaceae only appeared at very low level (0.02%) on day 47. Members of Neisseriaceae, such as Aquaspirillum and Chromobacterium, were reported as potential denitrifying bacteria in biological wastewater treatment systems (Zumft, 1997). The most dominant group changed from Saprospiraceae to Rhodocyclaceae between days 185 and 293. Also, the abundance of Comamonadaceae increased to 8.4% and 11.2%, respectively. These trends are similar to previous reports regarding WWTP, which revealed members of Rhodocyclaceae, such as Zoogloea, Azoarcus and Thauera, and also members of Comamonadaceae, such as Acidovorax, Comamonas, Diaphorobacter and Simplicispira as dominant denitrifying bacteria (Nielsen et al., 2009; Willems, 2014). Bacteria in these groups reduce nitrate, and some members of Rhodocyclaceae and Comamonadaceae can use a variety of carbon sources including alcohols, organic acids, and amino acids. Addition of a dose of ethanol or acetate to denitrifying reactors favors growth of specific bacterial genera such as Acidovorax, Comamonas, Azoarcus, and Thauera species within the families Comamonadaceae and Rhodocyclaceae (Lu et al., 2014); this result means that bacteria in these genera are responsible for the stable and high denitrification efficiency throughout the reactor operation by using carbon sources including ethanol and VFAs. Therefore, the changes in population may be associated with their ability to use ethanol and VFAs (Fig. 2b). Understanding of the bacterial community structure obtained in this study can provide useful information for start-up of a denitrification process. Results also suggest that inoculum sources can be compared by understanding how microbial communities respond to them, and that this understanding can be used to minimize disturbance of the communities. However, further studies on how specific microorganisms can dominate, and direct evidence of how they affect nitrate removal are recommended.
acetyl-CoA (Elefsiniotis and Li, 2006). Furthermore, nitrogen removal efficiency increases immediately after acetate addition (Isaacs and Henze, 1995). Presumably, metabolization of propionate entails pathways more complex than those used for ethanol and acetate (Constantin and Fick, 1997; Pereira et al., 2004; Yatong, 1996) or entails several sequential β-oxidation reactions. 3.3. Bacterial population identified by ion torrent PGM system To monitor the dynamics of the bacterial population, seven samples were collected from the anoxic tank and analyzed using Ion Torrent PGM System. The seven samples were collected on days 0, 47, 96, 131, 164, 185 and 203; the first sample was taken before the use of FRW; the other six were taken to track temporal bacterial shifts after the carbon source was changed. Analysis using Ion Reporter™ software yielded 278,283 sequence reads. The reads were classified into bacterial 285 families at a 92% sequence identity threshold. Among them, only 29 families had relative abundance >0.5% in all bacterial reads. At the family level, these 29 dominant groups accounted for 0–37.5% of the classified sequences (Fig. 3). Families Chitinophagaceae (10.6 ± 3.7%), Xanthomonadaceae (4.6 ± 3.2%), Hyphomicrobiaceae (2.2 ± 1.8%) and Planctomycetaceae (1.8 ± 1.7%) were detected in all seven samples. However, at day 0, in which RCS45 was utilized as sole carbon source, Methylococcaceae was dominant (12.0%; Fig. 3). Methylococcaceae are type-I methanotrophs that belong to class Gammaproteobacteria, which can use methane and methanol as sole carbon and energy sources but cannot use substrates that contain carbon-carbon bonds (Bowman, 2014). Therefore, Methylococcaceae contributed to nitrate removal by using the methanol in the RCS45. When the FRW was used, the most abundant groups changed during the operating period. First, Eubacteriacea (37.5%) and Saprospiracea (36.9%) were the most dominant groups in day 47. The relative abundance of Prevotellaceae increased from 0.06% on day 0 to 1.1% on day 47. Bacteria in these families been identified as hydrolysers. These changes indicate that bacterial populations were mainly affected by the composition of FRW, such as protein and carbohydrate. Under anaerobic conditions, members of Eubacteriaceae degrade cellulose, which is a major component of organic waste (Keating et al., 2012). The Saprospiraceae form one of the core microbial communities in WWTPs, in which high abundance of Saprospiraceae improves the hydrolysis of complex organic compounds such as carbon polymers, particularly protein, during wastewater treatment (Chen et al., 2016). Members of Provotellaceae are saccharolytic anaerobic bacteria, presumably involved in the hydrolysis and acidogenesis of carbohydrates (Maspolim et al., 2015). Provotellaceae species have been linked to the breakdown of proteins and carbohydrates, and produce cellulolytic enzymes such as carboxymethyl cellulase and xylanase (Nyonyo et al., 2014). Microbial hydrolysis releases simple carbon substrates that may supply carbon for organisms involved in nutrient removal. This hydrolysis may be the rate-limiting step for these important processes (Morgenroth et al., 2002). Although hydrolyzing bacteria such as Eubacteriacea, Saprospiracea and Provotellaceae cannot use complex carbon sources like FRW directly to remove nitrogen, their presence can significantly affect the growth and activity of aerobic and denitrifying bacteria. Although Saprospiracea remained the most dominant group until the day 164, the relative abundance of other groups increased as they adapted to the conditions. For example, a previous study has shown that Flavobacteriaceae was dominant population in the beginning, gradually decline as environment changed, but became a dominant group in the system over time (Zhang et al., 2013). Similarly, in this
3.4. Temporal variation in bacterial community structure Changes in bacterial community structure were visualized by NMDS analysis based on the relative abundances of each families, because this method can detect relationships among variables, and it is the most generally-effective ordination method for ecological community data (McCune and Grace, 2002). For the plot, the NMDS result had cumulative R2 = 0.89, which explained 82.0% of the cumulative variance in metagenomics community structures (Fig. 4a). Also, it showed 3.48 final stress and 0.00002 final instability, which indicate that its accuracy was acceptable (McCune and Grace, 2002). The NMDS result indicated that bacterial community structure shifted continuously over the period, and that significant shifts of bacterial community structure occurred three times: from day 0 to day 47, from day 47 to day 164, and from day 164 to day 185 (Fig.4a). The first shift from day 0 to day 47 was from Methylococcacea to Eubacteriacea and Saprospiraceae. This shift is mainly explained by the change of carbon source from RCS45 to FRW. This result concurs with a previous study in which a change of carbon source was one of the factors that affected microbial community structure during biological wastewater treatment (Lu et al., 2014). The second shift from day 47 to day 164 (cluster 1) was a suppression of Eubacteriacea and Saprospiraceae (Fig. 4b). This observation suggests that if specific microorganisms are suppressed without
1019
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
Fig. 3. Bacteria populations at the family level in the anoxic tank, as identified using Ion Torrent PGM next-generation sequencing. Bacteria populations with relative abundance <0.5% categorized as Minor groups.
Fig. 4. The bacterial profiles based on relative abundance at family level: (a) NMDS ordination plot and (b) Cluster dendrogram. Black arrows: shifts in community structure. Yellow dotted line: 80% of information remaining criterion. d: day. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Acknowledgements
any change of environment, their functions could be replaced by other microorganisms without major reduction in process efficiency in this system. The third shift from day 164 to day 185 was a dramatic increase in Rhodocyclaceae. This shift is mainly explained by adaption of the bacteria community like denitrifying bacteria.
This work was financially supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (No. 20163030091540). This work was also supported by ‘Human Resources Program in Energy Technology’ of the KETEP Grant, funded by the MOTIE, Republic of Korea (No. 20144030200460).
4. Conclusions References Stable denitrification performance over a period of seven months was achieved in a full-scale WWTP using FRW as alternative carbon source. Propionate was the most persistent VFA at the initial period, but was completely utilized after 19 days. Increase in the relative abundance of Eubacteriacea, Saprospiraceae, Rhodocyclaceae and Comamonadaceae after change to FRW suggests that bacterial acclimation to new carbon source may be responsible for the stability. Further study to elucidate how specific bacterial communities process different organic components would be valuable to improve process control.
APHA-AWWA-WEF, 2005. Standard Methods for the Examination of Water and Wastewater, 21th ed. American Public Health Association, Washington, D.C. Beccari, M., Passino, R., Ramadori, R., Tandoi, V., 1983. Kinetics of dissimilatory nitrate and nitrite reduction in suspended growth culture. J. Water Pollut. Control Federation 58–64. Bowman, J.P. 2014. The family Methylococcaceae. In: The Prokaryotes, Springer, pp. 411–440. Chan, Y.J., Chong, M.F., Law, C.L., Hassell, D.G., 2009. A review on anaerobic–aerobic treatment of industrial and municipal wastewater. Chem. Eng. J. 155 (1–2), 1–18. Chen, Y., Zhao, Z., Peng, Y., Li, J., Xiao, L., Yang, L., 2016. Performance of a full-scale modified anaerobic/anoxic/oxic process: High-throughput sequence analysis of its microbial structures and their community functions. Bioresour. Technol. 220, 225–232. Constantin, H., Fick, M., 1997. Influence of C-sources on the denitrification rate of a high-
1020
Bioresource Technology 245 (2017) 1016–1021
E. Kim et al.
Sci. Technol. 45 (6), 25–40. Nielsen, P.H., Daims, H., Lemmer, H., 2009. FISH handbook for biological wastewater treatment. Water Intelligence Online 8 9781780401775. Nyberg, U., Aspegren, H., Andersson, B., Jansen, J.l.C., Villadsen, I., 1992. Full-scale application of nitrogen removal with methanol as carbon source. Water Sci. Technol. 26 (5–6), 1077–1086. Nyonyo, T., Shinkai, T., Mitsumori, M., 2014. Improved culturability of cellulolytic rumen bacteria and phylogenetic diversity of culturable cellulolytic and xylanolytic bacteria newly isolated from the bovine rumen. FEMS Microbiol. Ecol. 88 (3), 528–537. Pereira, M., Sousa, D., Mota, M., Alves, M., 2004. Mineralization of LCFA associated with anaerobic sludge: kinetics, enhancement of methanogenic activity, and effect of VFA. Biotechnol. Bioeng. 88 (4), 502–511. Shin, S.G., Han, G., Lee, J., Cho, K., Jeon, E.-J., Lee, C., Hwang, S., 2015. Characterization of food waste-recycling wastewater as biogas feedstock. Bioresour. Technol. 196, 200–208. Sun, S.-P., Nàcher, C.P.i., Merkey, B., Zhou, Q., Xia, S.-Q., Yang, D.-H., Sun, J.-H., Smets, B.F., 2010. Effective biological nitrogen removal treatment processes for domestic wastewaters with low C/N ratios: a review. Environ. Eng. Sci. 27 (2), 111–126. Willems, A., 2014. The family Comamonadaceae. In: The Prokaryotes, Springer, pp. 777–851. Yatong, X., 1996. Volatile fatty acids carbon source for biological denitrification. J. Environ. Sci. 8 (3), 257–269. Zhang, C.S., Tang, J.H., Zhang, K.F., Rong, H.W., He, J.M., 2013. Start-up of denitrifying dephosphatation with nitrite as electron acceptor and microbial community analysis. Adv. Mater. Res.. Trans Tech Publ. 280–285. Zhang, H., Jiang, J., Li, M., Yan, F., Gong, C., Wang, Q., 2016a. Biological nitrate removal using a food waste-derived carbon source in synthetic wastewater and real sewage. J. Environ. Manage. 166, 407–413. Zhang, Y., Wang, X.C., Cheng, Z., Li, Y., Tang, J., 2016b. Effect of fermentation liquid from food waste as a carbon source for enhancing denitrification in wastewater treatment. Chemosphere 144, 689–696. Zumft, W.G., 1997. Cell biology and molecular basis of denitrification. Microbiol. Mol. Biol. Rev. 61 (4), 533–616.
nitrate concentrated industrial wastewater. Water Res. 31 (3), 583–589. Elefsiniotis, P., Li, D., 2006. The effect of temperature and carbon source on denitrification using volatile fatty acids. Biochem. Eng. J. 28 (2), 148–155. Fernández-Nava, Y., Marañón, E., Soons, J., Castrillón, L., 2010. Denitrification of high nitrate concentration wastewater using alternative carbon sources. J. Hazard. Mater. 173 (1–3), 682–688. Isaacs, S.H., Henze, M., 1995. Controlled carbon source addition to an alternating nitrification-denitrification wastewater treatment process including biological P removal. Water Res. 29 (1), 77–89. Karanasios, K.A., Vasiliadou, I.A., Tekerlekopoulou, A.G., Akratos, C.S., Pavlou, S., Vayenas, D.V., 2016. Effect of C/N ratio and support material on heterotrophic denitrification of potable water in bio-filters using sugar as carbon source. Int. Biodeterior. Biodegradation 111, 62–73. Keating, C., Cysneiros, D., Mahony, T., O'Flaherty, V., 2012. The hydrolysis and biogas production of complex cellulosic substrates using three anaerobic biomass sources. Water Sci. Technol. 67 (2), 293–298. Kim, D.-H., Lee, M.-K., Jung, K.-W., Kim, M.-S., 2013. Alkali-treated sewage sludge as a seeding source for hydrogen fermentation of food waste leachate. Int. J. Hydrogen Energy 38 (35), 15751–15756. Kim, H., Kim, J., Shin, S.G., Hwang, S., Lee, C., 2016. Continuous fermentation of food waste leachate for the production of volatile fatty acids and potential as a denitrification carbon source. Bioresour. Technol. 207, 440–445. Komorowska-Kaufman, M., Majcherek, H., Klaczyński, E., 2006. Factors affecting the biological nitrogen removal from wastewater. Process Biochem. 41 (5), 1015–1021. Lu, H., Chandran, K., Stensel, D., 2014. Microbial ecology of denitrification in biological wastewater treatment. Water Res. 64, 237–254. Maspolim, Y., Zhou, Y., Guo, C., Xiao, K., Ng, W.J., 2015. Determination of the archaeal and bacterial communities in two-phase and single-stage anaerobic systems by 454 pyrosequencing. J. Environ. Sci. 36, 121–129. McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach, OR. Morgenroth, E., Kommedal, R., Harremoës, P., 2002. Processes and modeling of hydrolysis of particulate organic matter in aerobic wastewater treatment–a review. Water
1021