Low-Temperature Adapted Nitrifying Microbial Communities of Finnish Wastewater Treatment

: In this study, the microbial community of nitrifying activated sludge adapted to Finnish climate conditions was studied to clarify the microbial populations involved in low-temperature nitriﬁcation. Microbial community analysis of ﬁve full-scale wastewater treatment plants (WWTPs) showed several di (cid:11) erences compared to WWTPs from other countries with a similar climate. In particular, very low abundance of ammonium oxidizing bacteria (AOBs) (altogether < 0.25% of total community) as well as typical NOBs ( < 0.35%) and a high abundance of orders Cytophagales and Micrococcales was observed in all Finnish WWTPs. To shed light on the importance of autotrophic and heterotrophic nitrifying processes, laboratory studies of activated sludge were carried out with a presence of and a lack of organic carbon in wastewater at 10 (cid:6) 1 (cid:14) C. Two di (cid:11) erent sludge retention times (SRTs) were compared to determine the e (cid:11) ect of this operational parameter on low-temperature nitrogen removal. The important role of previously reported Candidatus Nitrotogaarctica for nitrite oxidizing in cold climate conditions was conﬁrmed in both full-scale and laboratory scale results. Additionally, potential participation of Dokdonella sp. and Flexibacter sp. in nitrogen removal at low-temperatures is proposed. Operation at SRT of 100 days demonstrated more stable and e (cid:14) cient nitrogen removal after a sharp temperature decrease compared to 14 days SRT.


Introduction
Today, the main responsibility of conventional municipal wastewater treatment systems is to remove organic nutrients such as nitrogen and phosphorus from wastewaters in order to prevent environmental pollution due to human activities.Conventional biological nitrogen removal, based on nitrification and denitrification, is an underlying process in the majority of wastewater treatment plants (WWTPs).The stable and high efficiency of these WWTPs depends on the microbial composition and diversity of functional microbial groups in activated sludge [1].Consequently, design and operational control of existing WWTPs are based on the gained knowledge of the optimal parameters for the functional microbial communities.A successful nitrification (the oxidation of ammonia to nitrite, and further to nitrate) is traditionally related to sequential activity of ammonia-oxidizing and nitrite-oxidizing bacteria (AOB and NOB, respectively) [2].Temperature is considered to be one of the key factors affecting growth of the most-described AOBs and NOBs with an optimum value being around 28 • C, however cold-adapted AOB and NOB species have been reported in several studies [3,4].
In Finland, average the temperature of wastewater is in the range of 5-12 • C during most of the year and 14-21 • C in the shorter and warmest period, whilst stable and efficient nitrification can be reached at down to 8 • C [5,6].There are only a few studies on activated sludge microbial communities operated at this temperature range.According to Gonzalez-Martinez et al. [7], the diversity of bacteria and archaea in WWTPs decreases with the decrease of operational temperature.Decrease of AOB and ammonium oxidizing archaea (AOA) diversities due to a low temperature effect was also reported by Urakawa et al. [8].Accordingly, low concentrations of conventional nitrifiers have been detected in activated sludge of WWTPs in northern Finland as well as in pilot WWTPs operated at 10 ± 2 • C [7,9].Unknown temperature-tolerant AOA and heterotrophic nitrifiers were proposed as possible contributors [9,10].However, at the moment there is no available data on the key nitrifying bacteria and archaea of low-temperature adapted municipal-activated sludge and thus operational parameters of WWTPs are not optimized for cold regions.At the same time, Kruglova et al. [5,6] reported high sensitivity of nitrifying processes to temperature fluctuation at typical operation conditions while operation at longer sludge retention time (SRT) increased the resistance of the nitrifying community to low-temperature stress accompanied by microbial population shifts.
The objectives of this research were to study the microbial community structure of activated sludge adapted to Finnish climate conditions.Clarifying the groups of microorganisms, which are important for low-temperature nitrification, can help to improve day-to-day operation and troubleshooting of the activated sludge processes at WWTPs.To shed light on the importance of autotrophic and heterotrophic nitrifying processes, laboratory studies were carried out with the presence and lack of organic carbon in wastewater.In parallel, experiments with long SRT were performed in order to evaluate whether optimizing this operational parameter could stabilize efficient nitrogen removal in WWTPs in cold regions.

Sampling
The full-scale activated sludge samples were obtained from five large highly nitrifying WWTPs of Finland.The laboratory low-temperature-adapted activated sludge was obtained from four pilot reactors operated in water laboratory of Aalto University, Espoo.The sampling locations are presented in Figure 1.
Water 2020, 12, x FOR PEER REVIEW 2 of 17 value being around 28 °C, however cold-adapted AOB and NOB species have been reported in several studies [3,4].In Finland, average the temperature of wastewater is in the range of 5-12 °C during most of the year and 14-21 °C in the shorter and warmest period, whilst stable and efficient nitrification can be reached at down to 8 °C [5,6].There are only a few studies on activated sludge microbial communities operated at this temperature range.According to Gonzalez-Martinez et al. [7], the diversity of bacteria and archaea in WWTPs decreases with the decrease of operational temperature.Decrease of AOB and ammonium oxidizing archaea (AOA) diversities due to a low temperature effect was also reported by Urakawa et al. [8].Accordingly, low concentrations of conventional nitrifiers have been detected in activated sludge of WWTPs in northern Finland as well as in pilot WWTPs operated at 10 ± 2 °C [7,9].Unknown temperature-tolerant AOA and heterotrophic nitrifiers were proposed as possible contributors [9,10].However, at the moment there is no available data on the key nitrifying bacteria and archaea of low-temperature adapted municipal-activated sludge and thus operational parameters of WWTPs are not optimized for cold regions.At the same time, Kruglova et al. [5,6] reported high sensitivity of nitrifying processes to temperature fluctuation at typical operation conditions while operation at longer sludge retention time (SRT) increased the resistance of the nitrifying community to low-temperature stress accompanied by microbial population shifts.
The objectives of this research were to study the microbial community structure of activated sludge adapted to Finnish climate conditions.Clarifying the groups of microorganisms, which are important for low-temperature nitrification, can help to improve day-to-day operation and troubleshooting of the activated sludge processes at WWTPs.To shed light on the importance of autotrophic and heterotrophic nitrifying processes, laboratory studies were carried out with the presence and lack of organic carbon in wastewater.In parallel, experiments with long SRT were performed in order to evaluate whether optimizing this operational parameter could stabilize efficient nitrogen removal in WWTPs in cold regions.

Sampling
The full-scale activated sludge samples were obtained from five large highly nitrifying WWTPs of Finland.The laboratory low-temperature-adapted activated sludge was obtained from four pilot reactors operated in water laboratory of Aalto University, Espoo.The sampling locations are presented in Figure 1.
The main characteristics of WWTPs are presented in Table 1.The main characteristics of WWTPs are presented in Table 1.Full-scale sampling campaigns were organized over the spring and summer periods.Laboratory studies were carried out over autumn-winter periods.One litre of full-scale activated sludge for each sample was delivered to the Water Laboratory of Aalto University 24 h after sampling.From pilot reactors, 0.5 litre of activated sludge were collected for each sample.After collection and delivery to the laboratory, the samples were well mixed, and each sample was divided into two replicates which were immediately centrifuged for 10 min at 5000 rpm.Supernatant was removed and pellets were frozen at −20 • C for further processing.

Full-Scale Study
Full-scale samples were collected twice: in March-May (spring samples) when wastewater temperature was rising from −7 to −12 • C (15 • C in one location) and in the July-August period (summer samples), when the wastewater temperature was between −17 and −22 • C.

Pilot Study
Four pilot reactors (two equal Sequencing batch reactors (SBRs) and two equal Membrane Bioreactors (MBRs)) were operated at 10 ± 1 • C with synthetic municipal-like wastewater.The details of pilot reactor design are presented elsewhere [5,9].The main operational characteristics during the experimental period are presented in  [5].To remove the source of organic carbon from wastewater, yeast extract and peptone were replaced with acetate in two steps.First, yeast extract and peptone were removed while the concentration of CH 3 COOH* 3 H 2 O was increased to 392 mg L −1 .Next, all three components were totally removed, and the amount of molecular carbon was compensated by NaHCO 3 .
Samples from pilot reactors were taken first after the activated sludge from WWTP 1 was adapted to the laboratory conditions.The second sampling was performed after the organic carbon source was removed from the wastewater and the process performance was showing stable results.The main characteristics of the pilot-scale reactors and synthetic wastewater quality on the sampling days are presented in Table 2.
The quality of the influent, effluent and activated sludge was monitored regularly in accordance with Finnish Standards (SFS).The analyses and Standard Methods affiliations are listed in Table S1 of Supplementary materials.

Samples Processing
Frozen samples were sent to the Granada University in Spain for DNA extraction.DNA isolation was done using the FastDNA SPIN Kit for Soil.Extraction was made according to the manufacturer's instructions (MP Biomedicals, Solon, Ohio, United States).The primer pair 28F (5 -GAGTTTGATCNT GGCTCAG-3 )-519R (5 -GTNTTACNGCGGCKGCTG-3 ) was employed to amplify the V1-V2-V3 hypervariable regions of the 16S rRNA gene of the domain bacteria.For the samples from pilot studies, in addition to the primer pair, Arch519wF (5 -CAGCM GCCGCGGTAA-3 )-Arch1017R (5 -GGCCATGCACCWCCTCTC-3 ) was used to amplify archaeal hypervariable regions V4-V6 of 16S rRNA [11].The PCR steps were as follows: for bacteria 94 • C for 5 min, followed by 32 cycles at 94 DNA samples were sent for analysis to the United States at the Lubbock, Texas, at the Research and Testing Laboratory (hereinafter RTL).Next generation sequencing (Illumina Miseq apparatus, Illumina, San Diego, CA, USA) was applied to describe bacterial communities and diversity of nitrifying bacteria in activated sludge samples.The sequencing target was variant regions V1, V2, and V3 of the bacterial 16S rRNA gene.Illumina MiSeq Reagent Kit v3 reagent kit (Illumina, San Diego, USA) was used for sequencing.The primers used were 28F-519R pairs of primers.
The sequence data obtained after Illumina sequencing were processed and quality filtered in RTL.Subsequently, those sequences length was less than half of the expected sequence length with the primers used deleted.In addition, sequences that did not have the intact barcode or the separable bar code were removed [13].
The remaining sequences were clustered into operational taxonomic units (OTUs) in RTL.Clustering was done using the UPARSE algorithm and Edgar's methodology [14].Central OTU has been used for taxonomic classification.The taxonomic classification for OTUs was obtained using the USEARCH global search algorithm.Taxonomic information has been retrieved from RTL's own database.The database used is based on the National Center for the Biotechnology Information (NCBI) database.In the taxonomic assay, the percentages of sequences were 97% for bacterial species and 95% for archaea.
Results from RTL were analysed by both Excel 2016 software and QIIME (Quantitative Insights Into Microbial Ecology, [15] with 1.9.1 according to RTL's methodology [13].

Microbial Communities of Finnish WWTPs
Altogether, 228 species from 285 genera of 125 families, 66 orders and 32 classes belonging to 17 bacterial phyla were identified in full-scale samples.On the phylum level, the studied full-scale bacterial communities of activated sludge had a typical structure.Proteobacteria were the most dominant phylum found in all the samples.The other abundant phyla were actinobacteria (in most of the samples) or acidobacteria (in WWTP2), followed by Bacteroidetes and Firmicutes (Figure 2).Proteobacteria are commonly reported as the most abundant bacterial phyla of WWTP-activated sludge as well as Acidobacteria, Actinobacteria, Bacteroidetes and Fermicutes, which are reported as the next most dominant phyla [16][17][18].
Clear clustering of the samples from each WWTP, despite the small differences between the seasons, was observed using principle coordinates analyses (PCoA) based on Bray-Curtis distances (Figure S1 of Supplementary material).On the class level, most of the Proteobacteria in the samples were represented by alpha-and betaproteobacteria (from 30% to 88% of total community together).The dominance of these two classes in the active sludge has also been reported in other studies of WWTP microbial communities [17,19,20].
Spring and summer microbial communities on order level are presented in Tables 2S and 3S of Supplementary Materials, respectively.The most abundant bacteria (>1% of microbial community) in both spring and summer samples of all WWTPs belonged to the orders Cytophagales, Sphingobacteriales, Lactobacillales, Clostridiales, Rhizobiales, Rhodobacteriales, Burkholderiales and Rhodocyclales.In addition, in summer samples, high abundance of Micrococcales in all WWTPs was observed.
Orders Sphingobacteriales, Clostridiales, Rhizobiales, Burkholderiales and Rhodocyclales were also reported previously in literature as core orders of WWTP-activated sludge [7,17,21].Lactobacillales and Rhodobacteriales were found in the study of seven WWTPs from northern Finland, where a high abundance of these bacteria was also reported in all studied influents [7].In addition, both orders were reported in high abundance in Danish WWTPs with similar yearly temperatures to Finland, ranging from +7 to +20 °C [16,22].
Therefore, the main differences between the structures of the studied microbial communities Proteobacteria are commonly reported as the most abundant bacterial phyla of WWTP-activated sludge as well as Acidobacteria, Actinobacteria, Bacteroidetes and Fermicutes, which are reported as the next most dominant phyla [16][17][18].
Clear clustering of the samples from each WWTP, despite the small differences between the seasons, was observed using principle coordinates analyses (PCoA) based on Bray-Curtis distances (Figure S1 of Supplementary material).On the class level, most of the Proteobacteria in the samples were represented by alpha-and betaproteobacteria (from 30% to 88% of total community together).The dominance of these two classes in the active sludge has also been reported in other studies of WWTP microbial communities [17,19,20].
Spring and summer microbial communities on order level are presented in Tables S2 and S3 of Supplementary Materials, respectively.The most abundant bacteria (>1% of microbial community) in both spring and summer samples of all WWTPs belonged to the orders Cytophagales, Sphingobacteriales, Lactobacillales, Clostridiales, Rhizobiales, Rhodobacteriales, Burkholderiales and Rhodocyclales.In addition, in summer samples, high abundance of Micrococcales in all WWTPs was observed.
Orders Sphingobacteriales, Clostridiales, Rhizobiales, Burkholderiales and Rhodocyclales were also reported previously in literature as core orders of WWTP-activated sludge [7,17,21].Lactobacillales and Rhodobacteriales were found in the study of seven WWTPs from northern Finland, where a high abundance of these bacteria was also reported in all studied influents [7].In addition, both orders were reported in high abundance in Danish WWTPs with similar yearly temperatures to Finland, ranging from +7 to +20 • C [16,22].
Therefore, the main differences between the structures of the studied microbial communities compared to other reported WWTP data include high abundance of orders Cytophagales and Micrococcales.
Five bacterial genera presented in the highest abundance in most of the samples are shown in Figure 3. Four of them, including Candidatus Microthrix, Trichococcus, Rhodobacter and Hyphomicrobium, were previously reported as the most abundant activated sludge bacteria found in the study of 20 WWTPs in Denmark [22].In addition, high abundance of denitrifying acidobacterium Geothrix was observed in three WWTPs with a sharp increase of up to 58% of the community in summer samples of WWTP2.Hyphomicrobium, were previously reported as the most abundant activated sludge bacteria found in the study of 20 WWTPs in Denmark [22].In addition, high abundance of denitrifying acidobacterium Geothrix was observed in three WWTPs with a sharp increase of up to 58% of the community in summer samples of WWTP2.Despite the fact that Geothrix sp. is strictly anaerobic Fe (III)-reducing bacteria, it was also reported in dominant abundance, representing most of the acidobacteria in Danish WWTP-activated sludge as well as in northern Finnish WWTPs and seems to be the typical genus for activated sludge of WWTPs in Nordic countries [7,19].The contribution of Geotrix in biological removal processes is unclear but all the studied WWTPs are using ferrous or ferric sulphate coagulants for phosphorus precipitation.The dosing points for coagulants are in sand removal and in the beginning of secondary clarifiers.Additionally, in WWTP2, polyaluminum chloride was dosed to primary clarification during the summer months, which could affect bacterial community formation.Finally, Geothrix sp. was recently reported among dominating species in activated sludge, facing the selective pressure of several antibiotics due to its wide antibiotic resistance [23].Further studies are needed to clarify the engineering value of this bacterial genus.
Nitrifying bacteria were represented by ammonium oxidizers Nitrosomonas, Nitrosovibrio and Nitrosospira and nitrite oxidizers Candidatus Nitrotoga arctica and Nitrospira.In general, typical nitrifiers presented in unexpectedly low abundance in most of the WWTP samples.The overall abundance of AOBs was below 0.25% in all the bacterial communities and less than 0.35% of NOB Despite the fact that Geothrix sp. is strictly anaerobic Fe (III)-reducing bacteria, it was also reported in dominant abundance, representing most of the acidobacteria in Danish WWTP-activated sludge as well as in northern Finnish WWTPs and seems to be the typical genus for activated sludge of WWTPs in Nordic countries [7,19].The contribution of Geotrix in biological removal processes is unclear but all the studied WWTPs are using ferrous or ferric sulphate coagulants for phosphorus precipitation.The dosing points for coagulants are in sand removal and in the beginning of secondary clarifiers.Additionally, in WWTP2, polyaluminum chloride was dosed to primary clarification during the summer months, which could affect bacterial community formation.Finally, Geothrix sp. was recently reported among dominating species in activated sludge, facing the selective pressure of Water 2020, 12, 2450 9 of 16 several antibiotics due to its wide antibiotic resistance [23].Further studies are needed to clarify the engineering value of this bacterial genus.
Nitrifying bacteria were represented by ammonium oxidizers Nitrosomonas, Nitrosovibrio and Nitrosospira and nitrite oxidizers Candidatus Nitrotoga arctica and Nitrospira.In general, typical nitrifiers presented in unexpectedly low abundance in most of the WWTP samples.The overall abundance of AOBs was below 0.25% in all the bacterial communities and less than 0.35% of NOB bacteria altogether were identified in all the samples except WWTP5 spring samples, where Ca.Nitrotoga arctica presented in above 4% of total community (Figure 4).No increase in nitrifying bacteria was observed with increasing temperature.Furthermore, in three WWTPs, the amount of AOB and NOB noticeably decreased during summer compared to the spring season.At the same time, nitrification efficiency remained stable for all the experimental period with over 97% in four WWTPs and 93% in WWTP2 of ammonia transformation to nitrate, showing that the identified autotrophic nitrifying bacteria had little or no effect on low-temperature nitrification (Table 2).In the study of 50 Danish WWTPs [16], despite the seasonal variations, the abundance of total AOBs were reported above 2%, as well as the abundance of Nitrospira being above 2%, however no other typical NOBs were found in activated sludge [16].
Previous studies of AOBs have shown that the most often reported factors decreasing the diversity of the AOB community are the temperature, ammonium concentration and the presence of industrial wastewaters [1].For instance, ammonia-low wastewater may explain the absence of N. europaea and Nitrosococcus mobilis.Nitrosomonas and Nitrospirae are typically the most abundant nitrifying bacteria of activated sludge from Nordic countries [22].However, lower abundance of these genera and presence of Ca.Nitrotoga arctica in all WWTPs may indicate the effect of low temperatures on the overall nitrifying communities, since the growth optimum for Ca.Nitrotoga arctica is 10 °C with the possible growth range 4-22 °C [3].This also may explain the lower abundances of this species in summer samples.Further studies with more samples from different  No increase in nitrifying bacteria was observed with increasing temperature.Furthermore, in three WWTPs, the amount of AOB and NOB noticeably decreased during summer compared to the spring season.At the same time, nitrification efficiency remained stable for all the experimental period with over 97% in four WWTPs and 93% in WWTP2 of ammonia transformation to nitrate, showing that the identified autotrophic nitrifying bacteria had little or no effect on low-temperature nitrification (Table 2).In the study of 50 Danish WWTPs [16], despite the seasonal variations, the abundance of total AOBs were reported above 2%, as well as the abundance of Nitrospira being above 2%, however no other typical NOBs were found in activated sludge [16].
Previous studies of AOBs have shown that the most often reported factors decreasing the diversity of the AOB community are the temperature, ammonium concentration and the presence of industrial wastewaters [1].For instance, ammonia-low wastewater may explain the absence of N. europaea and Nitrosococcus mobilis.Nitrosomonas and Nitrospirae are typically the most abundant nitrifying bacteria of activated sludge from Nordic countries [22].However, lower abundance of these genera and presence of Ca.Nitrotoga arctica in all WWTPs may indicate the effect of low temperatures on the overall nitrifying communities, since the growth optimum for Ca.Nitrotoga arctica is 10 • C with the possible growth range 4-22 • C [3].This also may explain the lower abundances of this species in summer samples.Further studies with more samples from different seasons of the year must be considered to support these findings.

Acclimation of Activated Sludge to Laboratory Conditions
Activated sludge from WWTP1 was transferred to laboratory reactors to study the role of autotrophic and heterotrophic nitrifiers in a cold-temperature-adapted microbial community.The experimental period was started with an organic carbon load typical for Finnish full-scale processes (municipal-like wastewater) with a following reduction in the organic carbon source.The main operational parameters of the influent and their effect on the ammonia oxidation and nitrogen oxides formation in SBR and MBR reactors, are presented in Figure 5. experimental period was started with an organic carbon load typical for Finnish full-scale processes (municipal-like wastewater) with a following reduction in the organic carbon source.The main operational parameters of the influent and their effect on the ammonia oxidation and nitrogen oxides formation in SBR and MBR reactors, are presented in Figure 5. Efficient removal of ammonia was reached in all of the reactors, however adaptation properties between MBR and SBR reactors were noticeably different.Predictably, activated sludge in MBRs showed faster adaptation to 10°C conditions and total ammonium removal was reached after four weeks, while in SBRs, the removal of ammonium and the appearance of nitrogen oxides were limited for 11 weeks and then gradually increased over the five following weeks.These findings are in accordance with previously reported data showing that longer SRT may significantly improve the temperature adaptation process in activated sludge [5,6].Faster development of the nitrifying community can be explained by differences in microbial composition of activated sludge of SBR and MBR as well as by possible wash-out of slow growing nitrifiers in the SBRs.

Effect of Organic Carbon Source on the Dynamics of Activated Sludge Microbial Community
Consequently, to nitrification efficiency, the schedule of microbial communities sampling was planned.First microbial community samples were collected on the fifth week of the experiment, representing the activated sludge community under the effect of the temperature decrease, while MBRs were nitrifying and no nitrification was observed in SBRs.The additional sample of SBR microbial communities was taken after the appearance of the nitrification during Week 9.The next sampling for microbial analyses was performed in Week 13 of the experiment, representing the community during the decrease of the organic carbon load in the wastewater.The final samples were taken during Week 20, when all of reactors were nitrifying with the inorganic wastewater.Efficient removal of ammonia was reached in all of the reactors, however adaptation properties between MBR and SBR reactors were noticeably different.Predictably, activated sludge in MBRs showed faster adaptation to 10 • C conditions and total ammonium removal was reached after four weeks, while in SBRs, the removal of ammonium and the appearance of nitrogen oxides were limited for 11 weeks and then gradually increased over the five following weeks.These findings are in accordance with previously reported data showing that longer SRT may significantly improve the temperature adaptation process in activated sludge [5,6].Faster development of the nitrifying community can be explained by differences in microbial composition of activated sludge of SBR and MBR as well as by possible wash-out of slow growing nitrifiers in the SBRs.

Effect of Organic Carbon Source on the Dynamics of Activated Sludge Microbial Community
Consequently, to nitrification efficiency, the schedule of microbial communities sampling was planned.First microbial community samples were collected on the fifth week of the experiment, representing the activated sludge community under the effect of the temperature decrease, while MBRs were nitrifying and no nitrification was observed in SBRs.The additional sample of SBR microbial communities was taken after the appearance of the nitrification during Week 9.The next sampling for microbial analyses was performed in Week 13 of the experiment, representing the community during The dynamics of bacterial populations on order level are presented in Figure 7. Similar to the full-scale data, Cytophagales and Micrococcales seem to be the most important functional groups in conventional activated sludge due to their abundance in SBR samples.The increase of nitrifying efficiency also positively correlated with the abundance of Cytophagales, but negatively affected the abundance of Micrococcales.Additionally, the growth of Xanthomonadales and unclassified bacteria most closely related to Proteobacteria was observed with the increase in nitrifying efficiency and the decrease in organic carbon.The dynamics of bacterial populations on order level are presented in Figure 7. Similar to the full-scale data, Cytophagales and Micrococcales seem to be the most important functional groups in conventional activated sludge due to their abundance in SBR samples.The increase of nitrifying efficiency also positively correlated with the abundance of Cytophagales, but negatively affected the abundance of Micrococcales.Additionally, the growth of Xanthomonadales and unclassified bacteria most closely related to Proteobacteria was observed with the increase in nitrifying efficiency and the decrease in organic carbon.
conventional activated sludge due to their abundance in SBR samples.The increase of nitrifying efficiency also positively correlated with the abundance of Cytophagales, but negatively affected the abundance of Micrococcales.Additionally, the growth of Xanthomonadales and unclassified bacteria most closely related to Proteobacteria was observed with the increase in nitrifying efficiency and the decrease in organic carbon.In MBR samples, the abundance of Micrococcales noticeably decreased during the experiment.Conversely, Cytophagales dramatically increased during the experimental period, showing the positive correlation of the abundance with the decrease of the organic carbon source.
Sphingobacteriales was negatively correlated with the decrease in the organic carbon source and the increase of the nitrifying efficiency in both SBR and MBR reactors.It could be concluded that both Micrococcales and Sphingobacteriales are not the important contributors in low-temperature nitrification.However, slow-growing Sphingobacteriales might be important heterotrophic participants in the nitrogen-removing cycle.
The dynamics of bacterial communities on the genus level (>0.5% relative abundance at least one sample) is presented in Figure 8.The increase of Rhodobacter and Sphingobacterium during the experiment was observed in all the reactors, showing the selective advantage of these heterotrophic/mixotrophic genera from studied conditions.
Noticeable increase of Dokdonella sp. from order Xanthomonodales was observed with the increase of nitrifying activity and a decrease of the organic carbon source in SBR reactors.In MBR reactors, there was a sharp increase of Flexibacter sp. from order Cytophagales during the experimental period in MBRs and the vast majority of this species in final samples with no organic carbon in wastewater was observed.The role of these two genera in activated sludge was not discussed in previous studies of WWTP microbial compositions.
None of the cold-adapted AOB species previously reported in the literature were found in the activated sludge samples of this study.The appearance of NOB was observed with only one representing species Ca.Nitrotoga arctica in all the samples from Week 20 sampling (up to about 0.3% of the SBR community and 0.6% of MBR community).Since the seed sludge was taken during the autumn period, in consistency with the full-scale data, the temperature of the initial community operation was too high for this species' selective growth.After operation for 20 weeks at 10 • C Ca. Nitrotoga arctica could appear in the community as an active NOB.

Archaea
AOAs were expected to contribute to low temperature nitrification, since their range of adaptation temperatures is much wider compared to AOBs [24].Recently, Gonzalez-Martinez et al. [7] showed the positive correlation between ammonium oxidation in WWTPs of northern Finland with unclassified Euryarchaeaota genera.However, in this study, no clear correlation was observed.A study of the full-scale activated sludge sample from WWTP1 showed that unclassified Archaea represented 1.4% of the community and 1% was represented Methanobrevibacter sp.From pilot studies, archaea were identified only in one SBR sample and two MBR from the first and the last sampling dates.Since the results were not repeatable between the pilot reactors, no conclusions could be drawn from these data.
Zhang et al. [25] speculated that the lack of the data on the diversity of AOA in activated sludge could be due to the problems during the amplification.Consequently, in this study, five samples out of sixteen had low amplification rates and four had no amplification.Therefore, further studies of AOA contributions in low-temperature nitrification are needed with primary PCR protocol optimization.

Conclusions
A study of five full-scale Finnish WWTPs shows the typical highly nitrifying activated sludge bacterial community structure for low-temperature operated WWTPs on order level.It was observed that the most abundant bacterial genera identified in full-scale activated sludge elsewhere in Northern Europe were also typical for Finnish municipal WWTPs.However, several differences were observed, most probably caused by lower temperature conditions and specific wastewater composition and operational conditions typical for Finland.In particular, very low abundance of typical AOBs and NOBs and high abundance of orders Cytophagales and Micrococcales distinguished studied WWTPs.
The important role of previously reported Candidatus Nitrotoga arctica for nitrite oxidizing in cold climate conditions was confirmed in this study.Further studies on genera Dokdonella sp. and Flexibacter sp. are proposed for deeper understanding of the nitrogen removal cycle in low-temperature adapted WWTPs.It can be concluded that prolongation of sludge retention time from 14 days to 100 days is beneficial for nitrogen-removal efficiency at sharp temperature fluctuations.Further studies of AOA contribution with developed analytical methods are suggested.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2073-4441/12/9/2450/s1, Figure S1: Bray-Curtis PCoA plot showing the clustering of the samples from different WWTPs (clusters demonstrated by dashed lines), Table S1: Operational parameters, analysed during the study and Standard Methods, Table S2: Microbial communities of Finnish wastewater treatment plants (WWTPs) during the spring season, Table S3: Microbial communities of Finnish wastewater treatment plants (WWTPs) during the summer season.
Funding: This research received no external funding.
• C for 30 s, 60 • C for 40 s and 72 • C for 60 s and a final elongation step at 72 • C for 5 min; for archaea preheating during 5 m at 95 • C; 40 cycles of 30 s at 95 • C, 40 s at 57 • C, 90 s at 72 • C, and a final elongation step for 25 s at 80 • C [12].

17 Figure 2 .
Figure 2. Relative abundance of bacterial phyla in Finnish wastewater treatment plants (WWTPs 1-5) during the spring and summer season.

Figure 2 .
Figure 2. Relative abundance of bacterial phyla in Finnish wastewater treatment plants (WWTPs 1-5) during the spring and summer season.

Water 2020 ,
12, x FOR PEER REVIEW 9 of 17

Figure 3 .
Figure 3. Relative abundance of the dominant genera in activated sludge of Finnish wastewater treatment plants (WWTPs) during the spring and summer season.

Figure 3 .
Figure 3. Relative abundance of the dominant genera in activated sludge of Finnish wastewater treatment plants (WWTPs) during the spring and summer season.

Water 2020 ,
12, x FOR PEER REVIEW 10 of 17

Figure 4 .
Figure 4. Relative abundance of nitrifying bacteria in studied WWTPs.

Figure 4 .
Figure 4. Relative abundance of nitrifying bacteria in studied WWTPs.

Water 2020 ,
12, x FOR PEER REVIEW 11 of 17

Figure 5 .
Figure 5. Ammonium oxidizing efficiency during the experimental period.The results presented as average value between the two reactors.The error bars represent standard errors.Wastewater composition is changed during Weeks 12-14.

Figure 5 .
Figure 5. Ammonium oxidizing efficiency during the experimental period.The results presented as average value between the two reactors.The error bars represent standard errors.Wastewater composition is changed during Weeks 12-14.

Figure 6 .
Figure 6.Bray-Curtis principle coordinates analyses (PCoA) plot showing the dynamics of microbial community of seed sludge under operation in pilot MBR and SBR reactors with different wastewater composition.

Figure 6 .
Figure 6.Bray-Curtis principle coordinates analyses (PCoA) plot showing the dynamics of microbial community of seed sludge under operation in pilot MBR and SBR reactors with different wastewater composition.

Figure 7 .
Figure 7. Relative abundance of dominant bacterial orders in pilot reactors during the experimental period.

Figure 7 .
Figure 7. Relative abundance of dominant bacterial orders in pilot reactors during the experimental period.

Figure 8 .
Figure 8. Dynamics of the most abundant bacterial genera in SBR and MBR pilot reactors before and after the change of carbon source in wastewater.

Figure 8 .
Figure 8. Dynamics of the most abundant bacterial genera in SBR and MBR pilot reactors before and after the change of carbon source in wastewater.

Table 2 .
Initial synthetic wastewater at the beginning of the experiment contained 130.8 mg L −1 of CH 3 COOH* 3 H 2 O, 209.7 mg L −1 of yeast extract and 184.68 mg L −1 of peptone as a carbon source.Other components included NH 4 Cl, KH 2 PO 4 , CaCl 2 •2H 2 O, MgSO 4 •7H 2 O, trace nutrients and NaHCO 3

Table 2 .
The main operational parameters of the pilot-scale reactors.