Physico-chemical changes after real wastewater introduction
From the point of wastewater treatment efficacy, nitrogen removal is a crucial process and the researches on the nitrogen removal communities treating various sewage types are commonly performed (Kowalchuk et al. 1997; Rotthauwe et al. 1997; Nicolaisen and Ramsing 2002; Wertz et al. 2008; Vanparys et al. 2007; Attard et al. 2010).
In the case of coke wastewater, the situation is more complex. This sewage is regarded as difficult for biological treatment; thus, the bacterial community able to treat this wastewater needs to be specialized, easily adapting and relatively diverse in the case of the necessity of recalcitrant removal. The RBC biofilm bacterial community in this experiment was performing stable nitrogen removal from the beginning of the experiment on synthetic coke wastewater for 719 days. After this period, the real coke wastewater was introduced to the system and the experiment was performed further as it has been described in detail by Cema et al. (2016). Figure 2 presents the nitrogen removal in RBC system during the shift of the feeding medium change.
From 520th day of research and during period I of research, the average nitrogen concentration in the influent was equal to 930.5 ± 263.0 mg/L. The removal efficiency in RBC was 71.1 ± 20.3% with very efficient and stable ammonium nitrogen removal equal to 97.8 ± 6.5% (Table 1). Also, the phenol removal was very efficient at 99.7 ± 0.2% with an influent concentration equal from 175.6 to 32.3 mg/L. After the shift to the real coke wastewater, the sudden process breakdown was observed (Fig. 2). The nitrogen removal drops from over 80% at the end of the period I to almost 0 at the beginning of period II. Also, phenol concentration was fluctuating. The rebuild of the nitrogen removal took over 150 days (the results of nitrogen fluctuation in the reactor were described in detail in Cema et al. 2016).
The microbial community shift after real wastewater introduction
The aim of this study was to present the shift of the stable performing bacterial community after the medium change from synthetic to real coke wastewater. In order to analyze the qualitative changes of the microbial community structure, we compared the 16S rRNA gene amplicons data derived from two samples taken during process performance strictly before (632 days) and after the medium change on day 823 (the last sample collected in the system fed with real wastewater). Alpha diversity values for samples taken at days 632 and 823 were 29.48 and 21.23, respectively. Rarefraction curves for each of the samples were almost asymptotic suggesting that the majority of taxonomic diversity was uncovered. Both alpha diversity and rarefraction curves were examined using MG-RAST (Supp. 1, Fig. 1). Bacteria constitute 98.6% of the total community in sample 632 and 98.8% in sample 823. Such a result is obtained in most research performed on the bacterial communities in wastewater systems (Ma et al. 2015).
At class level, 36 bacterial classes were found in 632 sample, and 33 bacterial classes in sample 823. The proportions of the 8 major classes are presented in Fig. 3. Reads affiliated to Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria constituted the majority of the communities. STAMP comparison at the class level shows significant differences between analyzed samples (P = 95%, Fig. 4). Gammaproteobacteria, Actinobacteria, Planctomycetia, and Gemmatimonadetes were overrepresented significantly in sample 632, whereas Alphaproteobacteria, Betaproteobacteria, Flavobacteria, and Deinococci were overrepresented in sample 823 (Fig. 4). The predominance of Betaproteobacteria in sample 823, 29.74 % vs. 14.08 % in sample 632, may be explained with the removal of betaproteobacterial representatives from the system, due to a change in feeding medium, which in turn caused a decrease in nitrogen removal efficacy. In most cases, Betaproteobacteria are found to be predominant (Ma et al. 2015; Felföldi et al. 2010), but in WWTPs dealing with phenol-rich wastewater (such as coke wastewater), Alphaproteobacteria and Gammaproteobacteria were predominant (Figuerola and Erijman 2007; Wang et al. 2012) and Alphaproteobacteria can exceed Gammaproteobacteria (Xia et al. 2010). These results may also be attributed to the treatment method used. As it has been stated (Felföldi et al. 2010), Proteobacteria are predominant in most cases in WWTP systems, both in biofilm and activated sludge, but it is possible that the ratio of the particular Proteobacteria classes differs significantly when the analysis is performed with DNA sequencing or the other method (such as T-RFLP).
Research revealed the decrease of the Actinobacteria and Planctomycetia members during the process. Their abundance was reduced from 6.36 to 1.40% in the case of Actinobacteria and from 4.54 to 1.74% in the case of Planctomycetia. These results also underline the difference between wastewater systems based on activated sludge and biofilm. Moreover, these results could be linked with the decrease of nitrogen removal efficacy which in biofilm systems can be performed at large scale by Anammox process performed by Planctomycetia. In case the of activated sludge classes, Firmicutes, Chlorobi, and Chloroflexi are often dominant in the system (Felföldi et al. 2010; Ma et al. 2015; Wang et al. 2015), while for biofilm, these classes are not recognized as the main groups.
As it was previously stated by Cydzik-Kwiatkowska and Zielińska (2016) and Nascimento et al. (2018) also in the case of this research, the change of feeding wastewater can be the reason of the class proportion shift. It could be caused by the concentration changes in organic matter or recalcitrant substances.
Bacterial diversity and abundance were analyzed more specifically at the genus level (Tables 2 and 3). Table 2 showed genera which number increased during the process of coke wastewater treatment. The most significant change can be observed for Acinetobacter, Parabacteroides, Riemerella, and Pseudomonas. The number of these genera increased respectively: 567, 77, 36.17, and 18.17 times, during the experiment time when the medium was changed from synthetic to real coke wastewater. These results support the previous research of Liu et al. (2015). They have isolated Acinetobacter from activated sludge treating coke wastewater which was able to remove ammonium, nitrite, and nitrate with a very high efficacy. Our results can support the thesis that real medium causes bacterial community change, seen especially in the group of nitrifiers, to prevent nitrogen removal breakdown. The activity of autotrophic nitrogen-removing bacteria in RBC (such as Nitrosomonas, Table 3) was suppressed by the harmful real coke wastewater directed to the system and such change created environment suitable for heterotrophic nitrogen-removing bacteria, such as Acinetobacter. Also, the increase of Pseudomonas bacteria number supports this thesis due to the fact that both Pseudomonas and Acinetobacter are common coke wastewater treatment plant inhabitants and they are also able to remove thiocyanates as a sole carbon source (Boucabeille et al. 1994; Huang et al. 2013). It is worth mentioning that heterotrophic nitrifiers have attracted increasing attention as this group of bacteria grow faster and is less sensitive towards high loads of ammonium and organic matter than autotrophic nitrifiers (Liu et al. 2015). It is known that heterotrophic nitrification could be performed under a wide range of C/N ratio (Yang et al. 2017), but the research in this field are still scarce; thus, it would be reasonable to undertake such study especially in the field of coke wastewater treatment, where this process seems to be common (Liu et al. 2015; Lü et al. 2012; Yang et al. 2017).
The increase of Parabacteroides seems to be a natural process, because this anaerobic bacterium and its close relatives have been already identified in the biofilm system performing nitrogen and phosphorus removal (Feng et al. 2013); thus, it could be suspected that they are the natural biofilm inhabitant and the conditions in the system caused by the medium change was suitable for their multiplication. Their increase in the biofilm could be explained with their anaerobic metabolism which locates these bacteria in the bottom of the biofilm. Parabacteroides was protected against the harmful real coke wastewater influence by the upper layers of the biofilm in which aerobic nitrifiers were located.
Wastewater treatment plants are a potential source of pathogenic bacteria that can be spread not only via water and soil but they also can be aerosolized, becoming a source of airborne infections (Yang et al. 2019). NGS-based studies of pathogens present in WWTP communities as well as their presence and quantification in the effluent have already been recognized as a good solution for bacterial diseases prevention method (Lu et al. 2015). As WWTP bacterial communities are highly diverse, they can be a source not only of the common pathogens such as Bacillus (B. cereus) or Escherichia coli (Uhrbrand et al. 2017) but they can function as a reservoir of less known but comparably dangerous microorganisms. This study revealed the presence of several bacteria recognized as pathogenic but relatively less known and common.
The genus Riemerella is mainly known as a bird disease agent (mainly pigeons and ducks) (Segers et al. 1993; Rubbenstroth et al. 2013). But it was also found in wastewater treatment systems with activated sludge (Meli et al. 2016). This result supports the statement that all types of WWTPs could be the reservoirs of pathogenic bacteria.
The number of bacterial genera, Microbacterium, Planctomycetes, Gramella, Prosthecobacter, Subtercola, Sphingobacterium, and Pirellula and Leifsonia, decreased in sample 823 more than tenfold in comparison with sample 632 (Table 3).
Although Microbacterium is known to be a pathogenic bacteria, according to Chung et al. (2016), Microbacterium was isolated from the textile wastewater and able to degrade polyvinyl alcohol together with the other isolated strain of Paenibacillus. Moreover, it is also described as sulfamethoxazole degrader (Bouju et al. 2012). Thus, it could be suspected that real coke wastewater caused the decrease of these bacteria due to their higher sensitivity (as potentially pathogenic bacteria), but in the first part of the experiment on the synthetic wastewater, this bacterium could be responsible for recalcitrant and other wastewater compounds removal.
The decrease of Planctomycetes as the representatives of ammonium removal bacteria seems to be justified also at this taxonomical level. High toxic load in the feeding medium was the probable reason why these bacteria number was reduced. This change probably caused the change of autotrophic nitrogen removal into heterotrophic. It is worth mentioning that Pirellula, in which the number also decrease during the experiment, belongs to marine Planctomycetes (Glöckner et al. 2003) but probably is not linked with anaerobic ammonia oxidation; however, it could be suspected that the mechanism of harmful influence of real coke wastewater in both microorganisms is similar.
Gramella, a representative of Bacteroidetes, possesses NosZ genes linked with denitrification (Jung et al. 2013), so it could be suspected that in this research Gramella was responsible for denitrification but its number decreased due to the harmful real coke wastewater influence on the nitrogen removal bacteria. The same factor caused the decrease of Subtercola which according to the patent of Vanotti et al. (2009) is one of 16 critical genera in the activated sludge performing nitrification of high ammonium concentration at low temperature; thus, it is possible that the fluctuations of ammonia nitrogen concentration and that higher than in the first part of the experiment temperature of the real wastewater were the factors causing the decrease of these bacteria number.
According to the previous research, there are some representatives of Prosthecobacter (P. algae sp. nov.) isolated from activated sludge able to nitrate reduction to nitrite (Lee et al. 2014); thus, it could be suspected that bacteria belonging to this genus could be responsible for nitrogen removal processes in wastewater treatment and real coke wastewater was also the factor causing their number decrease.
According to previous statements, Sphingobacterium is relatively ubiquitous. These bacteria have been isolated from Antarctic soil, compost, and clinical samples also with opportunistic infections (Holmes et al. 1982; Yabuuchi et al. 1983; Shivaji et al. 1992; Kim et al. 2006). Also, Leifsonia and its close relatives were isolated from the environmental samples such as sewage sludge compost (Vaz-Moreira et al. 2008), soil (Madhaiyan et al. 2010), and water (Han et al. 2013). In the case of Han et al.’s (2013) research, these bacteria caused septicemia. Decrease of these genera number could be caused by high toxic load in real coke wastewater, but still, its presence in the wastewater treatment system underlines the fact that wastewater and systems for its treatment can be a pathogen reservoir.