Rapid Acclimation of Methanogenic Granular Sludge into Autotrophic Partial Denitrification Granules

Liu Chunshuang; Li Xuechen; Zhang Xiaofei; Bai Xue; Guo Yadong;Wang Yongxing; Zhao Chaocheng

(1. College of Chemical Engineering, China University of Petroleum, Qingdao 266580, China;2.CNPC Research Institute of Safety and Environmental Technology, Beijing 102206)

Abstract: Rapid formation of autotrophic partial denitrification (APD) granules is of practical interest to start up an expanded granular sludge bed reactor for wastewater treatment. This study demonstrates that methanogenic granules can be easily acclimated into autotrophic partial denitrification granules in one day, with the ability to remove 82% of 2.7 kg-S/(m3·d)sulfide into S0 and to convert 97% of 0.9 kg-N/(m3·d) nitrate into nitrite, which can provide a promising feedstock for anaerobic ammonia oxidation process. Arcobacter sp. is essential for S0 accumulation. Under high loadings, the abundance of Arcobacter sp. decreased, while on the contrary the abundance of unclassified_p_Firmicutes increased, leading to the deterioration of autotrophic partial denitrification performance. The granules performance could be recovered by adopting the strategies of properly reducing the influent loadings.

Key words: methanogenic granules; nitrite accumulation; autotrophic partial denitrification; sulfide

1 Introduction

Nitrogen compounds as the main pollutants should be removed from many kinds of wastewater such as petrochemical wastewater, municipal wastewater, and landfill leachate. Usually the reduced sulfur compounds are also present in the wastewater. The autotrophic partial denitrification (APD) process with biological granules can efficiently convert sulfide and nitrate into S0and nitrite,which is a promising feedstock for anaerobic ammonia oxidation process operating under high loading rates in an expanded granular sludge bed (EGSB) reactor[1-2]. The following mechanism for S0and nitrite accumulation was identified by Liu, et al.[1]

The autotrophic denitrifiers Acrobacter, Pseudo monas,Azoarcus , and Thiobacillus were present in the autotrophic denitrification process[3-5]. The partial denitrification granules can retain a high biomass in the reactor and,hence, are appropriate for treating high-strength wastewater containing sulfide and nitrate. However, the APD granules can be cultivated within a limited operational window;additionally, the cultivation time can be long due to the autotrophic feature. Conversely, the methanogenic granular sludge is abundant in existing anaerobic treatment processes. Fernandez, et al.[6]assessed the feasibility of using methanogenic granules in denitrification processes and examined the corresponding microbial shift. However,no study has applied methanogenic granules in the APD processes for converting sulfide and nitrate into S0and nitrite.Ιf methanogenic granules can be acclimated to APD granules, then the APD reactor startup can be achieved easily. This study acclimated methanogenic granules into APD granules in 1 day. The microbial community was monitored during the acclimation phase via the highthroughput sequencing technology.

2 Experimental

2.1 Seed sludge and reactor

Methanogenic sludge granules were acquired from a fullscale anaerobic digester in Jinan, China, which was used to treat the brewery wastewater. The concentration of suspended solids (SS) and volatile suspended solids (VSS)in seed granules was 35.4 g/L and 23.7 g/L, respectively.The UASB reactor was 6 cm in diameter and 120 cm in height, with a working volume of 1.25 L. The sulfide (S2-)-sulfur (S)/nitrate-nitrogen (NO3--N) ratio in the synthetic wastewater for the reactor was 3:1, with an initial loadings of 1.2 kg-S/(m3·d) and 0.4 kg-N/(m3·d), respectively (stage Ι, day 1-20). The feed concentrations were in the first step increased during the stage ΙΙ-V and then the hydraulic retention time was decreased from 6 h to 4 h during the stage VΙ-VΙΙ. At the stage VΙΙΙ, the HRT was restored to 6 h to investigate the recovery of reactor performance (Table 1).Other operating parameters of the UASB reactor were the same as those mentioned in the work by Liu, et al.[7-8]

Table 1 Operating conditions of the UASB reactor

2.2 Chemical analysis

Concentrations of nitrate, nitrite, sulfate, and acetate in liquid samples after passing the filter with the openings diameter of 0.45 μm were measured via ion chromatography (ΙCS-3000;Dionex, Bannockburn, ΙL, USA). The sulfide concentration was determined using the methylene blue method. Gas chromatography (GC-6890; Agilent, Foster City, CA,USA) was utilized to determine nitrogen gas content. Both SS and VSS in collected samples were measured using standard methods[9]. The concentrations of S0and nitrite was calculated based on mass balance calculations[3-5].

2.3 Community structure analysis

The DNA of sludge samples were extracted from 0.15 g to 0.25 g of dried sludge by utilizing the FastDNA kit(Qbingene, Carlsbad, CA) according to the manufacturer’s instructions[10-12]. The bacterial V3-V4 region of the 16S rRNA gene was amplified using the forward primers 338F (50-ACT CCT ACG GGA GGC AGC AG-30) and reverse primer 806R (50-GGA CTA CHV GGG TWT CTA AT-30). The detailed PCR mixture and reaction procedure can be referred to the work of Chen, et al.[13]After being purified by utilizing the GeneJetPCR purification kit (Thermo Scientific), the PCR product was put for sequencing on the Ιllumina Miseq PE300 platform(Ιllimina, USA).

3 Results and Discussion

3.1 Reactor performance

During the stage Ι (day 1—20), the sulfide removal efficiency was 100% and that for nitrate removal was about 83% at an influent loading rate of 1.2 kg-S/(m3·d)and 0.4 kg-N/(m3·d) , respectively (Figure 1). On day 1, the accumulation of elemental sulfur (S0) and nitrite reached as high as 81.97% and 67.93%, respectively,indicating that the methanogenic granular sludge was acclimated into APD granules within 1 day of cultivation in Table 1 with APD medium.

During the stage ΙΙ (day 21—43) and the stage ΙΙΙ(day 44—59), with the loading rates of sulfide and nitrate increasing to 1.4, 0.46 kg/(m3·d) and then to 1.6,0.532 kg/(m3·d), respectively, sulfide could be still completely removed, with the nitrate removal rate achieved 80% and 100%, respectively. The corresponding S0accumulation was initially high (95% at the stage ΙΙ),and then decreased slightly to 87% at stage ΙΙΙ. On the contrary, the nitrite accumulation was initially 80% at stage ΙΙ and then increased marginally to 80% at stage ΙΙΙ.The loading rates were increased again at stages ΙV— VΙ and attained 2.16 and 0.72 kg/(m3·d) (stage VΙ) for sulfide and nitrate, respectively. The rate for removal of nitrate was maintained at around 100%, however, the rate for removal of sulfide was slightly decreased from 100% to 90%. The nitrite accumulation was around 80% and the S0accumulation was roughly 85% with the exception of the drastic fluctuations during day 93 and day 107.

During stage VΙΙ, the loading rates were further increased to 2.7 kg-S/(m3·d) and 0.90 kg-N/(m3·d). The efficiency for sulfide removal decreased and remained at around 72% by the end of this stage. The efficiency for nitrite removal fluctuated strongly from 84% to 100%. The accumulation of nitrite and S0decreased markedly to 50% and 60%, respectively. The reactor performance with acclimated methanogenic granules started deteriorating under this loading.

The loading rates were decreased to 1.8 kg-S/(m3·d) and 0.60 kg-N/(m3·d), respectively, at stage VΙΙ in order to investigate the recovery of APD granules performance.The sulfide removal restored to 90% immediately and meanwhile the nitrite accumulation recovered to 88%.The S0accumulation increased from 60% (stage VΙΙ) to 75%. This result indicated that the APD granular could be recovered quickly after being destroyed.

3.2 Diversity of microbial community identified by high-throughput sequencing

The parameters related to the alpha diversity of microbial community for each sample at a distance cutoff level of 0.03 are shown in Table 2. The species richness for bacteria in the reactor varied significantly during the 164-day operation, which was revealed by OTUs and Chao 1. As shown in Figure 2, the rarefaction curves of the six samples at 0.03 distance suggested that the sequencing depths for all samples were well enough to cover the

Figure 1 Removal of N and S2-, and the conversion of and S0 in 164-d operation

whole diversity. This was confirmed by the coverage values of the six samples (99.90%, 99.31%, 99.03%,99.82%, 99.45%, and 99.50%, respectively, as shown in Table 2), indicating that almost all of OTUs in the reactor were detected in this study.

The results of Shannon diversity index in Table 2 also demonstrated that the bacterial diversity in the reactor initially decreased slightly from 4.03 to 2.21. Ιn addition,as shown in Figure 2, the cluster analysis indicated that a great change in microbial community composition occurred at high loadings of the APD granules.

Table 2 Richness and diversity of the six samples basedon 0.03 distance

Figure 2 Rarefaction curves based on the sequencing of bacterial communities

3.3 Microbial community structures

Ιn total, 14 known bacterial phyla were detected in the seeding sludge at stage Ι, which mainly included Proteobacteria, Chloroflexi, Firmicutes, Actinobacteria,Thermotogae, Synergistetes, Bacteroidetes, unclassified_k_norank, TM6_Dependentiae, Atribacteria,Aminicenantes, Spirochaetae, Caldiserica, and Cyanobateria (Figure 3). Most of bacterial phyla were reduced steeply and would eventually vanish during the 164-day operation. Only four phyla,including Proteobacteria, Firmicutes, Chloroflexi, and Actinobacteria, were selected as predominant ones to establish a stable foothold in the community.

Figure 3 Taxonomic classification of the bacterial communities at genus levels

A wide range of bacteria genera were identified as dominant in the stage Ι, such as Aliidiomarina (9.08%),Mesotoga (7.10%), norank_f_Anaerolineaceae (6.70%)and Syntrophobacter (5.39%), whereas almost of them were eliminated with an increasing loading rate (Figure 4).Only three phyla, including unclassified_p_Firmicutes,Acrobacter, and Azoarcus, were selected dominantly to establish a stable community for APD granules.

The genera unclassified_p_Firmicutes, Acrobacter and Azoarcus changed diversely with the increase of loading rates. The dominant genera unclassified_p_Firmicutes and Acrobacter were increased from null and 0.15% to 30.99% and 31.22%, respectively,with the influent loading rates increasing from 1.2 kg-S/(m3·d) and 0.4 kg-N/(m3·d) to 2.16 kg-S/(m3·d)and 0.72 kg-N/(m3·d). With the influent loading further increased to 2.7kg-S/(m3·d) and 0.9 kg-N/(m3·d), the abundance of unclassified_p_Firmicutes further increased to 39.33%. On the contrary, there was a remarkable decrease of Acrobacter to 25.05% during this period. At stage VΙΙΙ, when the influent loadings decreased from 2.7 kg-S/(m3·d) and 0.9 kg-N/(m3·d) to 1.8 kg-S/(m3·d)and 0.6 kg-N/(m3·d), respectively, the abundance of unclassified_p_Firmicutes significantly decreased from 39.33% to 22.23%, while Acrobacter significantly increased from 25.05% to 44.93%. There was a remarkable increase of Azoarcus to 14.37% from 9%, with the influent loadings increasing from1.2 kg-S/(m3·d) and 0.4 kg-N/(m3·d) to 2.7 kg-S/(m3·d) and 0.9 kg-N/(m3·d), and then decreased to 6.50% at an influent loading rates of 1.8 kg-S/(m3·d) and 0.6 kg-N/(m3·d).

Figure 4 Taxonomic classification of the bacterial communities at genus levels

Acrobacter may be responsible for the S0accumulation during the autotrophic denitrification process[14-15].Azoarcus is a common autotrophic denitrifier for sulfide removal[16]. This is the first time for unclassified_p_Firmicutes being detected in the autotrophic denitrification system. When the influent loading topped at 2.7 kg-S/(m3·d) and 0.9 kg-N/(m3·d), the removal of sulfide and nitrite dropped to 72% and 50%, respectively,and the S0accumulation diminished to 60%. However,the abundance of unclassified_p_Firmicutes topped at 39.33% and the abundance of Acrobacter dropped to 25.05%. Upon further decreasing the loading to 1.8 kg-S/(m3·d) and 0.6 kg-N/(m3·d), the performance of APD granules recovered, with the abundance of Acrobacter increasing to 44.95% and the abundance of unclassified_p_Firmicutes decreasing to 22.23%, respectively. The decrease of Arcobacter and increase of unclassified_p_Firmicutes at high loadings can lead to the deterioration of autotrophic partial denitrification performance.

4 Conclusions

Methanogenic granules were shown to be easily acclimated into APD granules in 1 day after removing all 1.2 kg-S/(m3·d) sulfide into S0and converting ＞83%of 0.4 kg-N/(m3·d) nitrite into NO2-. At 2.7 kg-S/(m3·d)and 0.9 kg-N/(m3·d), the efficiency for removal of sulfide decreased to 70% and the conversion rate of nitrite and S0decreased to 50% and 60%, respectively. Under high loading rates, the abundance of Arcobacter decreased and the abundance of unclassified_p_Firmicutes increased, resulting in the deterioration of APD granular performance. The performance of APD granular could be recovered quickly after being destroyed.

Acknowledgements:This research was supported by the National Natural Science Foundation of China (21307160),the Natural Science Foundation of Shandong Province, China(ZR20192019MEE038), the Fundamental Research Funds for the Central Universities (19CX02038A), the Open Project of Key Laboratory of Environmental Biotechnology, CAS(Grant No. kf2018003), and the Open Project Program of State Key Laboratory of Petroleum Pollution Control (Grant No. PPC2018006), the CNPC Research Ιnstitute of Safety and Environmental Technology.