The identification of the microorganisms present in the nitrifying sludge was carried out at different taxonomic levels. At the phylum level, Proteobacteria were identified as the dominant phylum in all cycles (cycle 39 = 62%, cycle 40 = 67%, cycle 47 = 50%, cycle 66 = 67%, cycle 99 = 78%, cycle 119 = 68%) (Fig. 2a). Acidobacteriota, Chloroflexi, Bacteroidota, and Actinobacteriota also remained among the most representative phyla despite having changes in their abundance along operation cycles. At the family level, Rhodanobacteraceae was the dominant in five of the six analyzed cycles (Fig. 2b). In cycle 99, Nitrosomonadaceae was the dominant family and always remained within the families with the highest abundance in the other cycles. Its relative abundance increased from 2.2% in cycle 39 without BT to values between 6.8 and 62.9% during BT addition in cycles 40–119, showing a marked increment after BT was introduced. Regarding genera, Pseudofulvimonas, Nitrosomonas, two uncultured genera of the Blastocatellaceae and Anaerolineaceae families, and an unidentified genus of the Xanthobacteraceae family were the most representative genera throughout the experimentation (Fig. 2c).

Relative abundance of nitrifying sludge during the operation of the SBR reactor at the (a) phylum, (b) family, and (c) genus levels

At the species level, 315 OTU’s (Operational Taxonomic Unit) were identified as different species throughout all the cycles of operation. In the control cycle (cycle 39), 134 species were identified and in the last analyzed cycle (cycle 119), 140 species were detected (Fig. 3). The cycle with the highest number of species detected was cycle 47 with 221 and the least number of species was detected in cycle 99 with only 106. These results are reflected in the Shannon diversity and Chao 1 richness analyses, indicating that in cycle 47 there was a greater diversity and richness of species (Table 1). In the same way, the less diversity of species and the lowest value of Chao 1 richness were registered in cycle 99. The Pielou’s evenness index ranged between 0.822 and 0.873 in all cycles except in cycle 99, showing stability in the evenness of the consortium. This coincides with the slight changes observed in the abundance of microbial populations in the different cycles of operation (Fig. 2c). However, in cycle 99, the Pielou’s evenness index dropped to 0.690, indicating a higher predominance of species over others. This is related to the higher abundance of Nitrosomonas in this cycle.

Bacterial population dynamics of the nitrifying sludge during the operation of the SBR reactor. Upset diagram of the frequency of OTUs through the operation cycles. The black circles of the matrix represent the OTUs present in a single SBR operation cycle, the connected black circles indicate the presence of the same OTUs in different operation cycles. The top of bar plot shows the number of detected OTUs per intersection and the lateral bar diagram shows the total number of OTUs identified in each operation cycle. The blue bar indicates the bacteria core of the nitrifying sludge. The green bars indicate the bacteria not detected in cycle 39 but detected from the first cycles of exposure to benzotriazole (cycles 40 or 47) up to cycle 119. The red bars indicate the bacteria present in cycle 39 and up to cycles 40 or 47

To identify the core of bacteria that made up the nitrifying sludge, a search was carried out for the species that were detected in all operation cycles, showing that 39 different OTU’s constituted the core (Fig. 3). Despite BT exposure, the sludge core remained above 58% of abundance in all cycles (Online Resource 1).

It was possible to identify at species level 38 OTU’s of the core. They belong to 26 different families, wherein six families were dominant: Rhodanobacteraceae 26.6%, Nitrosomonadaceae 9.0%, Blastocatellaceae 8.8%, Anaerolineaceae 5.6%, Xanthobacteraceae 3.7%, and Trueperaceae 2.3% (Fig. 4). Before BT exposition, the dominant families of the core with an abundance percentage greater than 3% were: Rhodanobacteraceae 28.9%, Blastocatellaceae 7.0%, Xanthobacteraceae 4.8%, Fimbriimonadaceae 3.9%, Sphingomonadaceae 3.6%, and Chitinophagaceae 3.1%. However, in the first cycle of BT exposition, the families Fimbriimonadaceae, Sphingomonadaceae, and Chitinophagaceae decreased in abundance to 2.0%, 1.6%, and 0.4%, respectively. In subsequent cycles, none of these families exceeded 1.6% of abundance. On the other hand, the Nitrosomonadaceae and Anaerolineaceae families increased their abundance after exposure to BT from 1.7% (cycle 39) to 20.0% (cycle 40) and 1.7–6.2%, respectively. The Nitrosomonadaceae family was the second most dominant of the nitrifying sludge core during exposure to BT (Fig. 4). The Trueperaceae family managed to increase its abundance until cycle 47 from 1.5 to 4.0%, maintaining its position as the sixth most dominant family in the cycles with exposure to BT and preserving an average of 2.3% after cycle 47 (Fig. 4).

Relative abundance of the families of bacteria that make up the core of the nitrifying sludge during the cycles of exposure to benzotriazole. Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software (BoxPlotR; http://shiny.chemgrid.org/boxplotr/); whiskers extend to 5th and 95th percentiles, outliers are represented by dots. n = 5, corresponds to the sample points (cycle 40 to 119)

According to Fig. 3, some bacteria were favored (green bar) or sensitive (red bar) to BT exposure. In this sense, eight bacteria were detected from the first exposure cycle (cycle 40) and remained present in subsequent cycles. These eight species were identified as heterotrophs belonging to the genera Limnobacter, Thauera, Pajaroellobacter, Iamia, OLB14, Lautropia, an uncultured genus from order Gaiellales, and an OTU from class Alphaproteobacteria (Fig. 5a). Other four bacteria were detected since cycle 47 and were able to remain present in the consortium up to cycle 119 (Fig. 3). These four OTU’s were identified as heterotrophic bacteria belonging to the genera 67 − 14 (order Solirubrobacterales), JG30-KF-CM45, Bauldia, and a member of the Propionibacteriaceae family (Fig. 5b). These twelve species would represent the heterotrophic bacteria favored by exposure to BT.

Relative abundance of heterotrophic bacteria prior to the addition of benzotriazole (cycle 39) and during different cycles of SBR operation with BT feeding (cycle 40 to cycle 119). (a) Bacteria detected since cycle 40 after BT feeding and that remained up to cycle 119. (b) Bacteria detected since cycle 47 and that remained up to cycle 119. (c) Bacteria previously reported in the literature as detected in media with hydrocarbon compounds

Meanwhile, other twelve identified bacteria belonging to the genera SH-PL14, Legionella, Gemmatimonas, Steroidobacter, Candidatus_Nucleicultrix, Terrimonas, WPS-2, an uncultured genus from Pirellulaceae family, an uncultured genus from order Microtrichales, an uncultured genus from phylum Armatimonadota and two bacteria from the Microscillaceae and Chitinophagaceae families resulted to be BT sensitive species as they were only detected in cycle 39 prior to BT exposure (Fig. 3). Likewise, three bacteria of the genera Flavobacterium, Fluviicola, and one species of the Reyranellaceae family, as well as six identified bacteria belonging to the genera Planctomicrobium, Legionella, Mesorhizobium, SM1A02, AKYG1722, and a member of the Holosporaceae family, were sensitive to exposure to BT, as they were only detected up to cycles 40 or 47, respectively.

In the group of AOB present in the nitrifying sludge, the Nitrosomonas and Nitrosospira genera were identified (Fig. 6a). Both genera were present in all operation cycles but not all species remained present in all cycles. As previously observed (Fig. 2c), Nitrosomonas was the second most abundant genus detected in BT exposure cycles. In the Nitrosomonas genus, Nitrosomonas mobilis was identified in five of the six cycles with an abundance greater than 1.7% during the cycles with exposure to BT and less than 0.5% in the control cycle. Nitrosomonas ureae was detected in cycles 47 and 119 with an abundance lower than 1% while two other unidentified species of the genus Nitrosomonas were detected in one cycle (Nitrosomonas uncultured sludge and Nitrosomonas unidentified). To taxonomically assign these species within the genus Nitrosomonas, the assembled sequences of both groups N_Un (Nitrosomonas unidentified) and N_us (Nitrosomonas uncultured sludge) were used to make a BLAST using the 16 S ribosomal RNA sequences database (Bacteria and Archaea) from NCBI (Online Resource 2, Table S1). The OTU identified as N_Un is most closely related to Nitrosomonas aestuarii while the OTUs of the N_us group are related to Nitrosomonas europaea. Furthermore, during the taxonomic assignment, it was not possible to assign thirteen OTUs (N1-N13) among the different characterized species of the genus Nitrosomonas from the SILVA database (Online Resource 2, Table S2). However, these OTUs maintained the highest abundance in all cycles within the genus Nitrosomonas (Fig. 6b). Prior to exposure to BT in the control cycle, these OTUs only reached 1% of abundance in the nitrifying sludge but after exposure to BT (cycles 40 to 119), their abundance was higher and reached its greater value in cycle 99 with more than 60% of the total population. As observed in Fig. 6b, the relative abundance of these thirteen OTUs decreased from cycle 40 to cycle 66. This coincides with the transitory phase (cycles 40–79) observed in the reactor where nitrification activity was altered. The higher abundance of these OTUs obtained in cycle 99 might be related to the recovery of metabolic function enabling effective nitrification (cycles 80–130). These results indicate a key contribution of these thirteen unassigned AOB species of the genus Nitrosomonas (N1-N13) in the ammonium oxidizing process. Additionally, the sequences of the best hits obtained using BLAST (Online Resource 2, Table S2) were used for the alignment analysis in M-COFFEE (https://tcoffee.crg.eu/) and PhyML 3.0 softwares to establish the possible phylogenetic relationships (Online Resource 3). Although the values ​​supporting the correlation were low, the results suggested that eleven of the thirteen OTUs are more associated with the species Nitrosomonas europaea, while N10 is associated with Nitrosomonas mobilis. Interestingly, N3 seems to be more related to the genus Nitrosospira, however the sequence size used greatly limited the association due to the high sequence similarity in both genera. This suggests that this group of thirteen OTUs belonging to the family Nitrosomonadaceae might be species that have not been previously reported, although the analysis of the V4 region is limited and further analysis is required.

Relative abundance of AOB prior to the addition of benzotriazole (cycle 39) and during different cycles of SBR operation with BT feeding (cycle 40 to cycle 119). (a) Ammonium oxidizing autotrophic bacteria abundance. (b) Most abundant unassigned AOB species of the genus Nitrosomonas (N1-N13)

In the genus Nitrosospira, Nitrosospira briensis was only detected in cycle 66 with an abundance of 0.46% (Fig. 6a). Another two OTUs were detected, however, it was not possible to assign them to any reported species of the Nitrosospira genus from the SILVA database. These OTUs remained present in all cycles with an abundance no more than 1.2%.

Regarding the heterotrophic bacteria present in the sludge, as previously described, twelve identified heterotrophs were favored by BT exposure (Fig. 5a and b). The group of eight bacteria that were detected since cycle 40 maintained as a whole, a relative abundance of 2.2% in total in cycle 40, with a higher total abundance of 4.3% in cycle 47 when the complete consumption of BT was carried out (Fig. 5a). In the same way, it is possible that the four heterotrophic species detected since cycle 47 and during all subsequent cycles would also be involved in the BT biodegradation process (Fig. 5b). The highest overall abundance of these four heterotrophic bacteria was observed in cycle 47 with a value of 1.2% in total, followed by cycle 99 with 0.8% of total abundance.

A search was made for identifying species or genera previously reported in the literature as associated with the BT biotransformation process or capable of growing in media with hydrocarbon compounds (Fig. 5c). Thirteen different species belonging to eleven different genera were identified in the sludge. The most frequent and abundant detected species were related to the Mycobacterium genus, followed by species of the genus Burkholderia-Caballeronia-Paraburkholderia. It is important to note that these species did not exceed 0.6% of individual abundance in each cycle. On the other hand, in the first exposure cycle to BT (cycle 40), the abundance of species of the genera Nocardioides, Pseudomonas, Rhodanobacter, and Stenotrophomonas was increased. While in cycle 47 the increase in abundance of the genera Rhodococcus, Bacillus, and Novosphingobium was favored. However, despite their abundance increased, these genera were not detected in all subsequent cycles. The genera Novosphingobium, Sphingomonas, Pseudomonas, and Rhodanobacter were detected in the last operation cycle (cycle 119) with an abundance that did not exceed 0.2%.