The effects of a newly isolated quorum quenching (QQ) bacteria (Bacillus sp. T5) on the microbial community has been evaluated via the Illumina sequencing method. Membrane bioreactors (MBRs) operated with this novel QQ bacterium to evaluate the improvement in the performance of MBR. Anti-biofouling effect of T5 was enhanced as 71% compared to the control reactor. Also, QQ bacteria did not have any negative effect on the removal of organics during the process. Gram-negative bacteria were found to be dominant over Gram-positive bacteria. Proteobacteria, Actinobacteria, Bacteroidetes, Acidobacteria, Firmicutes, and Chloroflexi were dominant phyla in the control and QQ reactors. The proportion of Alphaproteobacteria was most significant among Proteobacteria. The relative abundances of Actinobacteria, Acidobacteria, and Firmicutes were significantly affected by Quorum quenching mechanism. On the other hand, QQ activity of Bacillus sp. T5 significantly influenced the relative abundance of Proteobacteria, Bacteroidetes, and Chloroflexi. The QQ process appeared to generate variations in the structure of the microbial community. According to the results of the molecular analyses, the syntrophic interaction of Bacillus sp. T5 and indigenous Gram-negative and Gram-positive bacterial community is critical to the performance of MBRs.
MBRs are commonly employed in advanced wastewater treatment processes that are designed to separate the solid and liquid stages, and they have been proven to deliver a better performance than conventional activated sludge processes (Drews 2010; Meng et al. 2010). However, biofilm forms on the membrane surface in MBRs, resulting in biofouling. Despite the efforts of many researchers, it is not yet possible to apply physicochemical treatment methods to completely remove the biofilm that is generated by MBRs, and this continues to represent a significant challenge that undermines the sustainability of this technology (Malaeb et al. 2013).
Researchers have demonstrated that cell-to-cell signals, which are referred to as quorum sensing (QS) signals, finely regulate the biofilm that develops on the membrane surface during the operation of an MBR (Davies et al. 1998). As such, interfering with these signals through quorum quenching (QQ) can serve to reduce the amount of biofouling that occurs in MBRs, and many researchers have turned their attentions to blocking these signals as a means of controlling biofouling (Yeon et al. 2009; Kim et al. 2013b; Köse-Mutlu et al. 2015; Ergön-Can et al. 2017).
The main biological agents that are employed in biological treatments are microorganisms. Some researchers have examined how QQ bacteria generate QQ enzymes and have found that they can be employed to foster an approach to managing biofouling that is stable, environmentally friendly, and economically viable (Huang et al. 2008). To this end, there is a distinct need for ongoing research to focus specifically on QQ bacteria and their potential to reduce biofouling in MBRs.
While a study has been conducted that analyzed microbial 16S rRNA sequences in various MBR operating conditions, the effect that the QQ had on the MBR communities was not specifically examined in detail. Only very few studies exist in the literature on this field with high-throughput sequencing technologies (Lim et al. 2012; Kim et al. 2013a). In light of the fact that the effect of QQ mechanisms have on the composition of a microbial community and the subsequent implications for membrane fouling remains relatively unknown, there is a need to examine the dynamics of the microbial community, assess the associated metabolic products, and consider the influence of the operation parameters during the QQ process. In addition to improving existing knowledge and understanding of biofouling, in-depth research could also potentially identify the role QQ bacteria plays during the process of membrane fouling. Unfortunately, the molecular methods that are commonly used to analyze microbial communities, such as terminal restriction fragment length polymorphism (T-RFLP) and denaturing gradient gel electrophoresis (DGGE), are incapable of generating comprehensive data about the microorganisms that are present. However, one method that does hold significant potential in this area is that of Illumina sequencing. Illumina is a high-throughput sequencing technology that generates comprehensive 16S RNA data for a given site and covers diverse microbial populations, which has been found to enhance sensitivity and reduce the cost of sequencing (Logares et al. 2014). In addition, it has the ability to deal with all DNA materials. As such, it represents a very powerful tool that can be employed to develop detailed insights into the occurrence of microbial and functional genes in MBRs.
The purpose of this study was to examine the anti-biofouling effect of a new QQ bacterium, Bacillus sp. T5, in an MBR.
The aims of the research were as follows:
characterize the highly effective QQ bacterium Bacillus sp. T5;
assess the anti-biofouling effect of T5 in immobilization media in MBR; and
use the Illumina sequencing method to determine how QQ impacts the relative abundance of the microbial community during the QQ-MBR processes as a means of better understanding membrane biofouling. This approach may be helpful for understanding the microbial relationships that exist between QQ bacterium Bacillus sp. T5 and bacterial community structures in MBRs and how these hinder the improvement of both the stability and efficiency of the membrane bioreactors.
METHODS AND MATERIALS
QQ bacteria: Bacillus sp. T5
Quorum quenching bacterium Bacillus sp. T5 (Accession no. KR705939) was obtained from the Çamaltı Saltern (İzmir, Turkey) via an enrichment culture method (Yavuztürk Gül & Koyuncu 2017). Gram staining was carried out according to the procedure described by Beveridge (2001). The growth of strain T5 was examined both in autoclaved Luria-Bertani (Miller, USA) (LB) medium and synthetic wastewater. The growth curve of the T5 was plotted by measuring the optical density at a wavelength of 660 nm (OD660). The specific growth rate of Bacillus sp. T5 was determined based on the growth curve (Nair 2007).
Preparation of sodium alginate beads
As a cell immobilization media sodium alginate (Sigma-Aldrich) beads were used. The isolated Bacillus sp. T5 was cultured in LB broth at 31 °C at 180 rpm for overnight. Then, T5 culture was centrifuged 9,000 rpm for 15 min, washed and resuspended in 8 mL sterile distilled water (Kim et al. 2013b). To make 4% sodium alginate-T5 suspension 8,2 mg/mL T5 was mixed with 192 mL sterile sodium alginate suspension. T5-sodium alginate suspension was dropped in 3% CaCl2 solution through a pumping system (Supplementary Figure S1, available with the online version of this paper). Formed beads were left in CaCl2 solution for 8 h to increase the stiffness of the beads and beads were washed with sterile distilled water. The average diameter of the beads was 3 mm. Empty beads were prepared with the same procedure without the addition of T5 for the control reactor.
Measurement of AHL degrading activity of the free T5 and T5 beads
The N-acyl-homoserine lactone (C8HSL) degrading activity of T5 was measured according to bioassay method described in elsewhere (McLean et al. 2004; Lee et al. 2013). Agrobacterium tumefaciens was used as reporter strain. By the degradation rate of the C8-HSL, the potential quorum quenching activity of Bacillus sp. T5 and T5 immobilized sodium alginate beads were determined. Briefly, C8HSL was added to 30 mL 50 mM Tris-HCL (pH = 7) buffer to obtain a C8HSL final concentration of 200 nM. 50 pieces of T5 immobilized beads were added to C8HSL-Tris-HCL mixed solution to evaluate the beads activity. Free T5 activity were measured after centrifuging of T5 and centrifuged T5 was added to C8HSL-Tris-HCl solution. Samples were taken at certain periods of time in an hour. The bioassay was used to measure the residual concentrations of C8HSL. Bioassay plates were prepared according to described in previous studies. The residual amounts of C8HSL were calculated based on the calibration curve obtained by color zone sizes, corresponding to each standard concentration of the C8HSL.
Operation of MBR system
Activated sludge was taken from a wastewater treatment plant (İstanbul, Turkey) and acclimated to the synthetic wastewater before MBR operation. Two laboratory-scale MBRs were operated in parallel under a constant flux of 48 L/(m2 h). Control reactor had empty beads whereas QQ reactor had T5 immobilized beads. Reactor volumes and mixed liquor suspended solid (MLSS) concentrations were 4.5 L and 13,300–13,600 mg/L, respectively. The submerged polyvinylidene fluoride (PVDF) hollow fiber membrane modules with an effective area of 88 cm2 were used in both reactors. Hydraulic retention time (HRT) and sludge retention time (SRT) were set to 10 h and 30 d, respectively. The composition of the synthetic wastewater was as follows: glucose, 500 mg/L; urea, 100 mg/L; (NH4)2SO4, 50 mg/L; KH2PO4, 50 mg/L; MgSO4 7H2O, 50 mg/L; NaCl, 50 mg/L; CaCl2 2H2O, 10 mg/L; and NaCO3, 100 mg/L.
Deoxyribonucleic acid (DNA) was extracted from 500 mg of collected samples using a PowerSoil DNA isolation kit (Mo Bio Laboratories, USA) following the manufacturer's procedure. Concentrations of extracted DNA were determined via a NanoDrop UV–vis spectrophotometer (Thermo Scientifics, USA). 515F (50-GTGCCAGCMGCCGCGGTAA-30) and 806R (50-GGACTACVSGGGTATCTAAT-30) primers specific for V4 region (length, ca. 250 bp) of the rRNA gene were selected, required Illumina adapters and barcode sequences were added to the primers. Extracted DNA was amplified using polymerase chain reaction (PCR) following the amplification protocol, which is as follows: initial denaturation for 3 min at 94 °C, followed by 20 cycles of 45 s at 94 °C, 30 s at 53 °C, 90 s at 65 °C, and a final elongation step of 10 min at 65 °C (Shahi et al. 2016). All DNA samples were further purified using the Wizard DNA Clean-Up System (Promega) in accordance with the manufacturer's protocol. The samples were then quantified using Qubit 2.0 Fluorometer (Invitrogen, NY, USA). 16S rRNA genes were sequenced following the Illumina method (Illumina, Inc., CA, USA) with paired-end read cycles. Sequence analysis and the identification of operational taxonomic units (OTUs) were obtained using the methods suggested by Giongo et al. (2010) and Fagen (2012). At least 80% of sequence similarity was considered as the domain and phylum. OTUs abundance matrices for each taxonomic rank were created using the total number of reads, which showed 16S rRNA sequences matching with the database, and matrices of each sample were divided by the total number of pairs for normalizing varying sequencing depths.
MLSS, chemical oxygen demand (COD) and total Kjeldahl nitrogen (TKN) were determined according to APHA/AWWA/WEF (1998). The surface and cross section of the empty and QQ beads were visualized by scanning electron microscopy (SEM). Biofilm layers of the membrane surfaces were observed using confocal scanning laser microscopy (CLSM, C1 plus, Nikon, Japan).
R 3.1.1 analysis was used to conduct statistical analyses (www.r-project.org). Histogram, q-q plots and the Shapiro-Wilk's test were performed to examine data normality. Variance homogeneity was also investigated by using the Levene's test. One-way analysis of variance (ANOVA) or independent-samples t-test was used to check against the variations in QQ bacterium Bacillus sp. T5 and bacterial community structures. To provide multiple comparisons, the Tukey's test was used. Values of tests were pointed out as mean and standard deviation. Important difference was detected at the p < 0.05 level of importance.
RESULTS AND DISCUSSION
Molecular characteristics of Bacillus sp. T5
Gram staining, QQ activity of free and immobilized T5 and growth rate in LB medium and in synthetic wastewater (ww) of T5 were evaluated to observe microbial characteristics. As shown in Figure 1, Bacillus sp. T5 growth rate was higher in the LB medium than the synthetic ww. The specific growth rate of T5 during the exponential growth phase were 0.45 h−1 and 0.23 h−1 in LB medium and synthetic ww, respectively. As shown in Figure S2 (available with the online version of this paper), T5 was a Gram-positive and rod-shaped bacterium.
The AHL degradation results obtained from the free cell and T5 immobilized beads are presented in Figure 2. Bacillus sp. T5 degraded all the C8-HSL tested and exhibited great activity. The amount of the remaining C8-HSL concentration was negligible for the free T5.
Immobilization of Bacillus sp. T5 and its anti-biofouling effect on MBR performance
The beads were measured as 3 mm in diameter and had a spherical shape with a smooth surface and uniform size. In Figure 3, the surface and the cross-sectional SEM images show the morphologies of the vacant beads and QQ-beads. Rod-shaped T5 were spread on the bead surface. Before MBR operation, AHL degradation potential of the QQ-beads containing 8.2 mg T5/cm3 cubbyhole and the vacant beads were also measured. As shown in Figure 2, the degradation efficiency of C8HSL with the QQ-beads was measured to be 75% in the reaction time of 70 min. The adsorption of C8HSL on the vacant beads was negligible.
The mitigate of biofouling during the MBR operation was analyzed via monitoring transmembrane pressure (TMP). TMP rise was delayed substantially after applying T5 immobilized sodium alginate beads in the QQ reactor (Figure 4). Since TMP rises due to biofouling formation during MBR operation, TMP values were controlled to evaluate mitigation of biofouling in control and QQ reactors. Prevention efficiency of T5 was calculated using the areas under the TMP curves. As shown in Figure 4, during 12 days of MBR operation, TMP was reached to 480 mbar in control reactor, whereas TMP of QQ reactor was reached to 110 mbar. QQ reactor decreased TMP rise at least four times regarding the control reactor. This was the evidence of T5 immobilized QQ beads, inhibited biofilm formation and decreased TMP. According to the calculated area under the TMP curves; QQ effect of Bacillus sp. T5 resulted in 71% decrease in TMP values compared with the control. Inhibition of biofilm formation on membrane surface was confirmed visually via CLSM after 12 days of operation. In Figure 5, biofilm thickness formed is compared. According to the CLSM image from the QQ reactor was thinner than that from the control reactor. These data strongly supported the evidence of QQ effect on the inhibition of biofouling.
Our previous study (Yavuztürk Gül & Koyuncu 2017) was conducted with different immobilization media, namely QQ-fiber through the encapsulation of the very same QQ bacteria, Bacillus sp. T5. QQ-fiber was applied to a submerged MBR and resulted in 25% decrease in biofilm formation. According to the results, sodium alginate beads as an immobilization media increased the QQ effect of T5 from 25% to 71% on MBR. These results indicated that QQ bacterium, Bacillus sp. T5, could effectively mitigate the biofouling in MBR and QQ beads when T5 has higher QQ activity potential than the study based on QQ-fiber with same bacteria. These results are consistent with previous studies reporting that sodium alginate beads were more effective immobilization media with high QQ activity compared to the vessels (Oh et al. 2013; Köse-Mutlu et al. 2015).
To determine whether QQ mechanism has any side effect on COD and TKN removal efficiencies, biodegradation of organics of control and QQ reactors during the microfiltration process was also monitored. As shown in Figure S3 (available with the online version of this paper), the differences in COD and TKN removal efficiencies of QQ and control reactors were negligible. As a result, these findings suggested that QQ mechanism of T5 did not have any negative effect on the removal of organics.
Quorum quenching effects on microbial community structure
Microbial composition of the mixed liquor in control and QQ reactor was analyzed at the different taxonomic level during the MBR operation. Figure 6 shows the relative abundances in phyla in the studied samples. Proteobacteria (35 ± 2.8%), Actinobacteria (12 ± 2.3%), Bacteroidetes (11.7 ± 2.6%), Acidobacteria (4.2 ± 0.9%), Firmicutes (6.3 ± 1.5%), and Chloroflexi (5.2 ± 0.9%) were determined to be the dominant phyla in the control and QQ reactors, respectively. Proteobacteria and Actinobacteria were the most dominant phylum in the activated sludge, which was in accordance with former research that revealed Proteobacteria to be a dominant phylum in activated sludge (Miura et al. 2006; Teplitski et al. 2004; Lim et al. 2012; Kim et al. 2013a). Moreover, among Proteobacteria, the proportion of Alphaproteobacteria (38.2 ± 0.6%) and Betaproteobacteria (21.5 ± 0.4%) was much higher than Deltaproteobacteria (12.3 ± 0.3%) and Gammaproteobacteria (8.2 ± 0.3%). According to molecular analysis in MBR, most of the dominant phyla belonged to Gram-negative bacteria, which again confirmed the importance of these bacterial groups in the biodegradation of signal molecules.
Relative abundances of the microbial community were different in the control and QQ reactor, as expected. It is clear from Figure 6 that the relative abundances of Actinobacteria, Acidobacteria, and Firmicutes were significantly affected by QQ mechanism in QQ reactor, whereas the same phyla in control reactor were not affected (p < 0.05). On the other hand, QQ activity of Bacillus sp. significantly influenced the relative abundance of Proteobacteria, Bacteroidetes, and Chloroflexi. However, there are no significant differences in the relative abundances of the same phyla in the control reactor. This suggested that QQ bacterium Bacillus sp. T5 has the ability to change the proportion of certain microbial groups in the mixed liquor. Regarding the dominant bacterial strains in the MBR, Bacteroides sp., Acinetobacter sp., Defluvibacter sp., Desulfitobacterium sp., Streptococcus sp., Flavobacterium sp., Staphylococcus sp., Pseudomonas sp., and Bacillus sp. species were in abundance (Figure 7). These species constitued more than 50% of the dominant bacterial strains. Acinetobacter and Flavobacterium (Zhang et al. 2006) have been reported as one of the pioneer genera in the membrane fouling in a laboratory-scale MBR. Other genera such as Brevundimonas, Acinetobacter, Sphingomonas, and Aquaspirillum have also been reported to be present in low abundance in the mixed liquor but dominant on biofilm (Zhang et al. 2004). According to our results, relative abundance of Acinetobacter and Flavobacterium much higher in the mixed liquor of the QQ reactor than the control reactor and increased during the QQ process. These findings imply that QQ bacterium T5 may prevent the attachment of specific bacterial, especially Proteobacteria, groups to the membrane surface.
During MBR operation, MLSS and COD removal rate were maintained at constant levels. However, the relative abundance of microbial communities in mixed liquor demonstrated high diversities for each reactor. Furthermore, we demonstrated that QQ mechanism had a great effect on microbial taxa in the QQ reactor compared to the control one. This may be an indicator for a strong evidence of close association between microbial community structure with QQ and biofouling characteristics. Microbial diversities and their relative abundances in the community provide significant information about the main phyla involved in the inhibition of biofilm process in the QQ reactor. Consequently, the assessment of indigenous bacteria and the probable effects of quorum quenching on the population dynamics of MBR are necessary for understanding the QQ process.
The findings of this study indicated that Bacillus sp. T5 had a strong impact on the performance of MBR. Anti-biofouling effect of T5 immobilized beads was evaluated as 71%. The results of the molecular investigation revealed that Actinobacteria, Acidobacteria, Firmicutes, Proteobacteria, Bacteroidetes, and Chloroflexi were very sensitive to the QQ process of Bacillus sp. T5 and this was ultimately reflected in the performance of the membrane bioreactors. Therefore, molecular analysis of microbial community in MBR is strongly suggested for improving MBR systems and membrane biofouling mitigation.
This work supported by The Scientific and Technological Research Council of Turkey (TUBITAK) (Project No. 114Y706).