SEARCH

SEARCH BY CITATION

Keywords:

  • pyrosequencing;
  • electron-donor availability;
  • electron acceptor;
  • autotrophs;
  • heterotrophs;
  • community structure

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgement
  7. References
  8. Supporting Information

Using two membrane biofilm reactors in which hydrogen (H2) was the only exogenous electron donor, we studied the microbial community structure of biofilms composed primarily of denitrifying bacteria (DB) and sulfate-reducing bacteria (SRB). In steady-state EDvSS, H2 availability was restricted and varied. In steady-state EAvSS, the input nitrate (inline image) concentration was varied relative to a fixed sulfate (inline image) concentration. SRB co-existed with DB, even when inline image reduction was absent due to restricted H2 availability. UniFrac and principal coordinate analysis indicated that H2 availability and electron-acceptor loadings framed the microbial community structure, with H2 availability having a greater impact. In EDvSS, restricted H2 availability favored heterotrophic DB (i.e. Burkholderiales) compared with autotrophic DB (e.g. Hydrogenophilales and Rhodocyclales). In EAvSS, inline image reduction lowered the relative abundance of some DB (e.g. Hydrogenophilales), and the biofilm was colonized by Desulfovibrionales and Bacteroidales. Reinforcing the impact of H2 availability, EAvSS showed a higher microbial diversity and more even distribution among microbial groups than did EDvSS. Thus, the biofilm community in a H2-fed biofilm with DB and SRB became more heterotrophic when the H2 availability was constrained, while low inline image loading allowed more inline image reduction, causing a shift to more SRB.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgement
  7. References
  8. Supporting Information

Nitrate (inline image) and sulfate (inline image) are chemical oxyanions that coexist in a variety of waters. inline image is considered a water contaminant, usually generated by agricultural run-off or sewage discharge, because it causes methemoglobinemia in infants and spurs eutrophication of surface waters (US EPA, 2012). inline image is a natural constituent of water and is not normally viewed as a contaminant (US EPA, 2011). inline image and inline image often coexist in water due to anthropogenic activities (e.g. agricultural leaching of fertilizers, wastewater discharges), natural mineralogy (e.g. inline image minerals such as sodium sulfate, magnesium sulfate, and calcium sulfate), and atmospheric deposition of SO2 or NOx (van Breemen & van Dijk, 1988; Lovett, 1994).

Given the common co-occurrence of inline image and inline image in water, studies focusing on interactions of these two oxyanions are of high relevance for water-quality improvement by microbiological means, because many bacteria utilize inline image and inline image as electron acceptors to generate energy for their growth. Denitrification, the respiratory reduction of inline image to N2 gas, is a stepwise process catalyzed by a set of well-known reductase enzymes (Payne, 1973; Knowles, 1982; Rittmann & McCarty, 2001). Respiratory sulfate reduction relies on a different set of reductases to stepwise inline image, ultimately generating hydrogen sulfide (H2S) (Peck, 1959), which is a corrosive and toxic substance.

As summarized by Payne (1981) and Mateju et al. (1992), the denitrifying bacteria (DB) are spread in many phylogenetic genera that include autotrophs and heterotrophs. Some common autotrophic denitrifiers are in the genera Thiobacillus, Paracoccus, Ferrobacillus, and Leptothrix. Pseudomonas and Azonexus are examples of heterotrophic denitrifiers, while facultative DB are represented by Hydrogenophaga. Muyzer & Stams (2008) summarized the major sulfate-reducing bacteria (SRB) and their phylogenetic relationships. Typical SRB belong to the orders Desulfovibrionales, Desulfobacterales, Syntrophobacterales, Desulfotomaculum, Desulfosporomusa, and Desulfosporosinus. Also, SR microorganisms are present in the Archaea domain: Archaeoglobus, Caldivirga, and Thermocladium are some representative examples.

The ability of DB and SRB to coexist is determined by differences of their growth rates (Tang et al., 2012a) and thermodynamics (Rittmann & McCarty, 2001). Because inline image respiration is energetically more favorable than inline image respiration, DB growth rates are faster than SRB growth rates (Tang et al., 2012a), and this provides DB an advantage over SRB when they compete for common resources, such as an electron donor and space (Ontiveros-Valencia et al., 2012). The selection of DB over SRB in mixed communities has been a practical strategy to control SRB, and the addition of inline image has been used to minimize inline image reduction and H2S production in sewers (Bentzen et al., 1995; Garcia de Lomas et al., 2005). However, studies have demonstrated that some SRB strains, such as Desulfovibrio and Desulfomicrobium, were able to remain in biofilms exposed to inline image, even though others (e.g. Desulfobacter and Desulfobulbus) disappeared immediately after inline image addition, leading to rapid DB enrichment in sulfidogenic biofilms (Mohanakrishnan et al., 2011). Thus, the response of SRB to inline image addition appears to be genus specific, with some SRB strains able to coexist despite selective pressure from inline image.

The hydrogen (H2)-based membrane biofilm reactor (MBfR) has been successfully applied for microbial reduction of diverse sets of oxidized contaminants (e.g. Lee & Rittmann, 2002; Nerenberg & Rittmann, 2002; Chung et al., 2006a, b, 2007; Ziv-El & Rittmann, 2009; Zhang et al., 2010). In the MBfR, H2 is delivered to autotrophic bacteria by diffusion through the wall of bubbleless gas-transfer membranes. The outside of the membrane wall provides an ideal habitat for H2-oxidizing bacteria, which form a strong and stable biofilm (Lee & Rittmann, 2002; Nerenberg et al., 2008; Ziv-El & Rittmann, 2009). The microbial ecology of biofilms in H2-fed biofilms has been studied for many different sets of electron acceptors (Chung et al., 2008; Nerenberg et al., 2008; van Ginkel et al., 2010; Zhang et al., 2010; Zhao et al., 2011), but most of the previous studies have not addressed the presence and diversity of SRB.

Recently, Ontiveros-Valencia et al. (2012) studied the coexistence of DB and SRB in H2-fed MBfR biofilms. Higher electron availability (controlled by the H2 pressure supplied to the membrane) led to complete denitrification and an increase in DB (quantified by quantitative polymerase chain reaction, qPCR, targeting nitrite reductase genes). inline image reduction occurred only when the inline image effluent concentration was driven below 0.1 mg N L−1 and SRB increased (as assayed by qPCR targeting the dissimilatory sulfite reductase alpha subunit gene or dsrA) at higher H2 pressures when H2 availability was limiting. However, SRB were present in the H2-fed biofilms whether or not inline image was being reduced because of their metabolic diversity (Ontiveros-Valencia et al., 2012).

Here, we expand our understanding of the microbial ecology beyond the presence and abundance of SRB and DB in the biofilms of the H2-based MBfR. We evaluate the microbial community structure and the factors producing changes in the important genera/orders of autotrophic-founded biofilms containing DB and SRB. In particular, we identify SRB that are especially able to coexist in DB-dominated biofilms, including in situations in which inline image reduction does not occur, and we show how the onset of inline image reduction affects some DB taxonomic groups more than others.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgement
  7. References
  8. Supporting Information

Reactor configuration and continuous operation

Following Ontiveros-Valencia et al. (2012), we set up two MBfRs each composed of two glass tubes interconnected with Norprene tubing (model 06404-15,16,26; Masterflex) and plastic fittings. The total membrane surface area of each MBfR was 94 cm2, which was distributed in a main bundle of 49- 25-cm-long polypropylene fibers (Teijin, Ltd., Japan) and 10–25 cm long for ‘coupon’ fibers set up for biofilm samples. The total liquid volume was 60 mL; liquid was circulated through both MBfRs at a rate of 150 mL min−1, and they were operated at room temperature (25 ± 1 °C). We analyzed biofilm samples from the two MBfRs described in Ontiveros-Valencia et al. (2012). Both MBfRs were inoculated with activated sludge from the Mesa Northwest Wastewater Treatment Plant, for which the microbial composition has been described previously (Li et al., 2011). Table 1 summarizes the operating conditions for both MBfRs. The inline image influent concentration was held constant for both MBfRs (~ 46 mg L−1). One MBfR was operated with a set of increasing H2 pressures, which allowed us to control the electron-donor (i.e. H2) availability for a fixed ratio of the two acceptors. This set of experiments is identified as the electron-donor-varied steady states, EDvSS. For the second MBfR, the input concentration of inline image was varied, while the inline image concentration and H2 pressure were held constant. This allowed us to evaluate the effect of electron-acceptor availability, and this set of experiments is identified as the electron-acceptor-varied steady states, EAvSS. EDvSS and EAvSS were operated with continuous influent flow rates of 0.67 and 0.17 mL day−1, respectively. The corresponding hydraulic retention times were 89 and 352 min. Due to the higher flow rate in EDvSS, electron-acceptor loading rates for this reactor were higher for EDvSS than for EAvSS (Table 1); this led to H2 limitation in EDvSS, but not in EAvSS.

Table 1. Operating conditions and function metrics for EDvSS and EAvSS. The tested variables are indicated by the shaded squares. Experimental H2 fluxes and electron-acceptor (inline image and inline image) removal fluxes are from Ontiveros-Valencia et al. (2012). The maximum H2 delivery capacities of the polypropylene fibers at a given pressure were calculated from Tang et al. (2012b)
ReactorSample IDH2 pressure (atm)Maximum H2 delivery capacity (g H2 m−2 day−1)Experimental H2 flux (g H2 m−2 day−1)inline image influent concentration (mg N L−1)inline image loading (g N m−2 day−1)inline image loading (g inline image m−2 day−1)Nitrate removal flux (g N m−2 day−1)Sulfate removal flux (g inline image m−2 day−1)
EDvSS1a2.00.420.34101.04 ± 0.044.9 ± 0.210.510
1b2.70.560.47   0.810
1c3.00.630.56   1.040
1d3.70.780.80   1.082.56
EAvSS2a2.70.560.15100.261.2 ± 0.070.260
2b  0.21200.55 0.420
2c  0.1710.13 0.130.97
2d  0.1450.02 0.021.12
2e  0.20100.26 0.260.61
2f  0.33250.68 0.680

We monitored the concentrations and reduction kinetics for inline image and inline image, as described in Ontiveros-Valencia et al. (2012). Once the reactors reached a steady-state condition (5–10% variation in inline image and inline image effluent concentrations over at a minimum of 10 days), we took samples of the biofilm for DNA extraction (Ontiveros-Valencia et al., 2012). The biofilm samples represented an area of 0.8–1 cm2, which is large enough that localized heterogeneities did not bias the phylogenetic distributions (Ziv-El et al., 2012).

Pyrosequencing and sequence analysis

To investigate the major DB and SRB phylotypes found in the biofilm and their relationship with the bioreactor performance, we sent all DNA samples for pyrosequencing at the Research and Testing Laboratories LLC (Texas, USA), which performed amplicon pyrosequencing using a standard 454/GS-FLX Titanium (Sun et al., 2011). The Bacteria domain was targeted by selecting the V6 and V7 regions of the 16S rRNA gene with primers 939F (5′-TTGACGGGGGCCCGCAC-3′) and 1492R (5′-TACCTTGTTACGACTT-3′) (Zhao et al., 2011). The potential presence of Archaea was not determined. We processed the raw data using QIIME 1.4.0 suite (Caporaso et al., 2010a) and removed sequences having fewer than 200 bps, homopolymers of more than 6 bps, primer mismatches, or an average quality score lower than 25. We picked the operational taxonomic unit (OTU) using the Greengenes 16S rRNA gene database with uclust (Edgar, 2010) based on ≥ 97% identity, removed OTUs that contain less than two sequences (singletons) from our analysis, and aligned the representative sequence of each OTU to the Greengenes database using PyNast (DeSantis et al., 2006; Caporaso et al., 2010b). The potentially chimeric sequences were identified using ChimeraSlayer (Haas et al., 2011), and a python script in QIIME was employed to remove the chimeric sequences. To assign taxonomy to OTUs, we used the ribosomal database project (RDP) classifier with a 50% confidence threshold (Wang et al., 2007). We constructed Newick-formatted phylogenetic trees using FastTree (Price et al., 2009).

For the purpose of eliminating heterogeneity related to having different numbers of sequences among the samples, we subsampled the OTU table by randomly selecting 10 different times 740 sequences per sample, which was the lowest number of sequences found in one sample. We created 10 iterations for every 10 sequences and repeated this process until we reached 740 selected sequences in each sample. The diversity and evenness within each subsample of 740 sequences was calculated from rarified OTU tables with the mean of the last ten iterations of each sample. We averaged the estimates for the 10 iterations we created for every 10 sequences, compiled the averages, and produced rarefaction plots.

We used a set of metrics to characterize the microbial communities of the two MBfRs in terms of diversity and evenness. While a higher value for the Shannon diversity index indicates greater microbial diversity, a value for the Simpson metrics near one shows an even distribution of bacterial groups within the sample. The OTU richness was estimated by calculating Chao1, which determines the asymptote on an accumulative curve, predicting how many OTUs would be present if a high number of sequences had been collected, and the phylogenetic relationships using PD (Faith, 1992), which estimates the cumulative branch lengths from random OTUs.

To evaluate the overall community composition, we quantified the fraction of unique branch lengths from the total branch length of the phylogenetic tree using the unweighted UniFrac distance matrix (Lozupone et al., 2006). The unweighted option accounts only for the presence or absence of microbial phylotypes. We generated principal coordinate analysis (PCoA) plots and unweighted pair group method arithmetic mean (UPGMA) plots (Lozupone et al., 2006) using jack-knifed beta diversity that subsampled each sample at a depth of 740 sequences. Sequence data sets are available at NCBI/Sequence Read Archive (SRA) under study with accession number SRP018321. Individual sample files have the following accession numbers: SAMN01902537 - SAMN01902546.

Results and discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgement
  7. References
  8. Supporting Information

Community function

Table 1 summarizes the results of the reduction of inline image and inline image for EDvSS and EAvSS for the steady states when DNA samples were taken. The < 10% differences between the experimental H2 fluxes and the maximum H2 delivery fluxes point out that H2 was limiting in EDvSS (Ontiveros-Valencia et al., 2012). Thus, the reductions of inline image and inline image depended on the H2 pressure applied to the membranes in EDvSS. Starting with the lowest H2 pressure, the removal flux for inline image increased with greater H2 pressure until inline image was completely removed. Then, inline image was reduced as H2 became available for the SRB (EDvSS 1d).

In EAvSS, the experimental H2 flux always was at least 20% less than the maximum H2 delivery flux (Tang et al., 2012b), which indicates that H2 delivery was not limiting in the biofilm. While the H2 concentration changes within the biofilm (e.g. being at higher concentrations near the fiber surface than near the liquid side), the H2 that could be delivered at the gas pressures utilized in EAvSS was more than enough to supply all the H2 needed by the DB and SRB in the biofilm. In all the cases except EAvSS 2b, the inline image removal flux equaled the inline image loading (Table 1), which means that denitrification was complete. Significant rates of inline image reduction occurred only for the three lowest inline image loadings (EAvSS 2c, 2d, and 2e).

Forces driving the biofilm microbial community structure elucidated by UniFrac and PCoA

Pyrosequencing generated a total of 48 524 high-quality sequences with a median length of 355 bp for 16S rRNA gene for all the biomass samples of EDvSS and EAvSS. Figure 1 shows the results of the unweighted UniFrac analysis for an overall community comparison. All biofilm samples from EAvSS formed a cluster (highlighted in red), while three of four biofilm samples from EDvSS (1a to c) formed another cluster (highlighted in blue). Sample 1d, which clustered closer to the samples from EAvSS, was the only steady state, in which inline image reduction was observed for EDvSS; hence, the overall community was dramatically affected when inline image reduction took place. The blue group corresponds solely to biofilm samples with denitrification as the predominant microbial respiratory process (Table 1, samples 1a–c).

image

Figure 1. Clustering based on the unweighted UniFrac analyses. The branch length represents the distance between biofilm samples in UniFrac units, as indicated by the scale bar. 1a–d correspond to EDvSS, with 1a = 0.42 g H2 m−2 day−1, 1b = 0.56 g H2 m−2 day−1, 1c = 0.63 g H2 m−2 day−1, and 1d = 0.78 g H2 m−2 day−1. 2a–f correspond to EAvSS, with 2a = 10 mg N L−1, 2b = 20 mg N L−1, 2c = 5 mg N L−1, 2d = 1 mg N L−1, 2e = 10 mg N L−1, and 2f = 25 mg N L−1.

Download figure to PowerPoint

We performed PCoA for the sequences obtained for all biofilm samples. Figure 2 shows the unweighted PCoA, which is based only on the presence or absence of phylotypes. Again, all the samples from EAvSS grouped together, having relatively low values of PC1. The biofilm samples with the highest removal flux for inline image (EAvSS 2f) were slightly distant from the rest of the samples on the PC2 vector. For EDvSS, the effect of H2 availability on the biofilm structure showed a clear gradient (1a[RIGHTWARDS ARROW]1b[RIGHTWARDS ARROW]1c[RIGHTWARDS ARROW]1d), in which the samples with the least H2 availability (samples 1a and 1b) showed the highest magnitudes for PC1, while the samples with the greatest H2 availability became more like EAvSS on the PC1 axis.

image

Figure 2. Principal coordinate analysis (PCoA) based on the unweighted UniFrac analyses. PC1 and PC2 axes represent 25.68% and 14.40% of the variance within the microbial community. 1a–d correspond to EDvSS, with 1a = 0.42 g H2 m−2 day−1, 1b = 0.56 g H2 m−2 day, 1c = 0.63 g H2 m−2 day−1, and 1d = 0.78 g H2 m−2 day−1. 2a-2f correspond to EAvSS, with 2a = 10 mg N L−1, 2b = 20 mg N L−1, 2c = 5 mg N L−1, 2d = 1 mg N L−1, 2e = 10 mg N L−1, and 2f = 25 mg N L−1.

Download figure to PowerPoint

Ontiveros-Valencia et al. (2012) concluded that H2 availability for EDvSS and electron-acceptor loading (or inline image influent concentration in these experiments with a constant influent flow rate) for EAvSS, respectively, were the critical factors affecting the removal fluxes for inline image and inline image. The UniFrac and PCoA analyses support these conclusions, but also reflect how the community structure behaved. PCoA analysis demonstrates that H2 availability caused greater variance among the samples than electron-acceptor loading, which is well illustrated by the trends along the PC1 axis. UniFrac showed evidence for microbial community clustering in the two MBfR reactors when inline image reduction was significant within the biofilm.

Along with electron-donor availability and electron-acceptor loading rates, other factors affect the structure of the microbial community in the biofilm. For instance, the profiles of dissolved components such as H2, inline image, and inline image also have significance. As modeled by Tang et al. (2012a), the H2 concentrations are higher near the fiber surface, allowing a higher concentration of DB and SRB than at the liquid side, which is mostly populated by inert compounds and heterotrophs. The profiles of the electron acceptors inline image and inline image vary accordingly the respective biomass fractions of DB and SRB: the inline image concentration nonlinearly declines from the liquid side to the fiber side of the biofilm due to the high density of DB near the fiber surface, but the inline image concentrations do not decline much in the biofilm because of a smaller fraction of SRB than DB.

As discussed above, H2 availability and electron-acceptor loading rates allow a higher or lower abundance of DB and SRB within the biofilm. For example, higher H2 availability leads to more accumulation of DB. However, once complete denitrification is achieved, SRB are able to compete with DB for H2 and space near the fiber surface.

The sequential order of the experiments influenced the community structure. Following the steady states favoring inline image reduction (EAvSS2c and 2d), the biofilm community retained SRB despite the introduction of inline image and was still capable of reducing inline image (EAvSS2e). The SRB also remained in the biofilm in a subsequent steady state without inline image reduction (EAvSS2f).

Supporting Information Fig. S1 and Table S1 show that the microbial diversity was higher for EAvSS over EDvSS based on number of OTUs, Chao1, and Shannon indices. Thus, H2 limitation restricted diversity and led to fewer dominant phylotypes. Lastly, the evenness and PD werehigher for EAvSS than for EDvSS (Table S1 Simpson metrics and Fig. S2, respectively).

Heterotrophic and autotrophic DB dominance

The different degrees of H2 availability for EDvSS and EAvSS led to different microbial communities (Fig. 3). Figure S3 shows the microbial community for both MBfRs at the phylum and class level. The community of EDvSS was dominated by mostly heterotrophic DB (Burkholderiales) when H2 was severely restricted (EDvSS 1a and 1b); however, once the limitation for H2 was relieved, DB capable of autotrophic metabolism, such as Hydrogenophilales (chemoautotrophic bacteria that respire inline image and oxidize H2) and Rhodocyclales (a highly versatile microbial group with representative chemolithoautotrophic bacteria such as Paracoccus denitrificans and Methyloversatilis), outcompeted the heterotrophic ones (EDvSS 1c and 1d). The dominance of heterotrophic Burkholderiales when H2 was severely limited suggests that the community relied more on organic donors available from soluble microbial products (SMP) released by the autotrophs (Merkey et al., 2009; Ni et al., 2011). The growth of heterotrophic bacteria has been associated with the production of SMP by autotrophic bacteria (e.g. Kindaichi et al., 2004; Ni et al., 2011; Tang et al., 2012a). The abundance of heterotrophic bacteria has even reached 50% in an autotrophic nitrifying biofilm (Kindaichi et al., 2004).

image

Figure 3. Relative abundances of the most abundant microbial phylotypes at the order level for EDvSS and EAvSS. The EDvSS and EAvSS letter and number codes show the chronological order of samples. Samples for EAvSS are shown according to increasing inline image concentration. The sum does not add up to 100% in all cases because minor phylotypes are not shown.

Download figure to PowerPoint

Without restrictions on H2 for EAvSS, the largest DB representation was by phylotypes related to Rhodocyclales and Hydrogenophilales, with Burkholderiales was in third place, but at significantly lower abundance. This indicates that the biofilm community of DB in EAvSS was predominantly autotrophic.

The heterotrophic and autotrophic DB phylotypes in EDvSS and EAvSS are represented at the genus level in Fig. 4. For EDvSS, heterotrophic microorganisms, including Aquabacterium-like phylotypes (Fig. 4 sample 1b) and Dechloromonas-like phylotypes (1a–b), were prevalent with severe H2 limitation, while Methyloversatilis-like phylotypes (methylotrophic microorganisms capable of utilizing CO2 as carbon source) increased with increasing H2 availability (1b–d). Zhao et al. (2011) similarly found that the microbial community moved toward mixotrophic in a H2-fed biofilm when H2 delivery was limited in a denitrifying and perchlorate-reducing community. In EAvSS, Methyloversatilis was the most abundant DB genus, reinforcing the autotrophic conditions under H2 nonrestriction, and it showed a positive correlation with the increase in inline image concentration.

image

Figure 4. Relative abundances of the most abundant microbial phylotypes at the genus level for EDvSS and EAvSS. The EDvSS and EAvSS letter and number codes show the chronological order of samples. Samples for EAvSS are shown according to increasing inline image concentration.

Download figure to PowerPoint

Competition between DB and SRB: a deeper insight by pyrosequencing analysis

In EAvSS, Rhodocyclales, Hydrogenophilales, and Burkholderiales generally increased with higher inline image concentration, but Hydrogenophilales and Burkholderiales declined as inline image reduction became more important (Fig. 3). The DB community of EAvSS was clearly distinct from the DB community of EDvSS, and Rhodocyclales was the largest DB phylotype in EAvSS. In EDvSS, DB phylotypes were better competitors for H2 than SRB (e.g. Desulfovibrionales), which only showed higher relative abundances once H2 became available to them after complete denitrification (H2 pressure > 3 atm).

Using qPCR, Ontiveros-Valencia et al. (2012) reported a rise of nirS-containing denitrifiers with higher H2 availability in EDvSS. However, pyrosequencing was able to reveal which phylotypes correlated with the increase in nirS-containing denitrifiers. The nirS-containing denitrifiers in our system were Rhodocyclales, Hydrogenophilales, and Burkholderiales (Saunders et al., 2000; Matsuzaka et al., 2003; Beller et al., 2006; Yoshida et al., 2010). Burkholderiales decreased while Hydrogenophilales increased with greater H2 availability. Hence, the increase in nirS-containing denitrifiers with higher H2 availability observed by Ontiveros-Valencia et al. (2012) was correlated with the increase in Hydrogenophilales.

Despite the lack of active inline image reduction, the biofilm samples of Ontiveros-Valencia et al. (2012) showed similar abundances of SRB in EAvSS. One possibility is that SRB were actively reducing inline image in a process known as ammonification (Dalsgaard & Bak, 1994; Moura et al., 2007). However, ammonium was not detected in the MBfR effluents, which suggests that the SRB potentially were respiring oxygen (Dilling & Cypionka, 1990; Marschall et al., 1993) or fermenting organics (Widdel & Hansen, 1991). The apparent lack of inline image reduction also might be attributed to sulfide oxidation by DB. However, sulfur-driven autotrophic denitrification (Shao et al., 2010), for which the final product of respiration is N2, oxidizes sulfide to S° (Reyes-Avila et al., 2004; Chen et al., 2009a, b, 2010) or to inline image (Shao et al., 2010). Both cases were unlikely for our biofilm samples because (1) inline image reduction should have been suppressed by competition from denitrification (Tang et al., 2012a) and (2) pyrosequencing did not reveal DB known to do sulfide oxidation (e.g. Thiobacillus denitrificans, Thiothrix, Thiomicrospira denitrificans, Sulfurimonas denitrificans, Paracoccus denitrificans (Shao et al., 2010)). Furthermore, we did not observe the loss of inline image, which would have occurred if the oxidation product was S°. Although not carrying out denitrification or ammonification, SRB coexisted with DB even when inline image suppressed inline image reduction.

Dominant SRB phylotypes and effect of inline image reduction on the microbial community

Sulfate-reducing bacteria were represented by phylotypes most closely related to Desulfovibrionales (Fig. 3). In EDvSS, Desulfovibrionales became more prominent at the highest H2 availability (EDvSS 1d), but Desulfovibrionales were significantly reduced as the inline image concentration increased in EAvSS (from EAvSS 2d to 2f). Desulfovibrionales, which have high metabolic versatility (Dilling & Cypionka, 1990; Widdel & Hansen, 1991), could remain in the biofilm community even though they are dominated by DB and denitrification was happening (Fig. 3 samples 2a, b, e, and f), a trend also seen in other systems (e.g. Gu et al., 2005; Fields et al., 2006; Mohanakrishnan et al., 2011). SRB-containing orders Desulfobacterales and Desulfuromonadales also were present (at < 2% and < 1% relative abundances) in EAvSS, but not in EDvSS (Fig. 3); this reflects the greater diversity of SRB in EAvSS. It also illustrates how pyrosequencing allowed us to detect subtle impacts of inline image concentration on SRB; these abundance trends correlated well with results with the qPCR assay of the dsrA gene (Ontiveros-Valencia et al., 2012).

Consistent with the UniFrac analysis (Fig. 1), inline image reduction had a clear impact on framing the microbial community beyond DB and SRB. At the highest inline image reduction rates (EDvSS 1d and EAvSS 2c and 2d), the relative abundance of phylotypes similar to Holophagales decreased (Fig. 3). Holophagales are homoacetogens also capable of utilizing inline image as its electron acceptor (Coates et al., 1999; Drake et al., 2002). The loss of Holophagales with high inline image reduction likely reflects a competition with SRB for H2 in EDvSS and space within the biofilm in EAvSS. On the other hand, inline image reduction appeared to favor phylotypes closely related to Bacteroidales (in the phylum Bacteroidetes; EDvSS 1d and EAvSS 2c and 2d). Bacteroidales participate in the mineralization of organic matter (Nagata, 2008), particularly proteins and carbohydrates (Church, 2008). The correlation of the abundances of Bacteroidales and Desulfovibrionales during inline image reduction suggests that these microorganisms established a cooperative relationship. Most likely, Bacteroidales utilized SMP (Ni et al., 2011) released by SRB-like Desulfovibrionales during inline image reduction (Tang et al., 2012a). Ziv-El et al. (2012) also observed significant abundance of Bacteroidales and attributed their presence to the production of acetate by fermentation of complex organic molecules (e.g. decaying biomass and SMP).

In conclusion, H2 availability and inline image loading significantly shaped the microbial community structure in the MBfR. H2 availability (in EDvSS) had a greater impact than inline image loading (in EAvSS) on community structure; this included a decline in microbial diversity as H2 delivery was restricted. Furthermore, the onset of inline image reduction strongly modified the microbial community, with communities experiencing inline image reduction being distinct from those without inline image reduction. When denitrification was the major microbial respiratory process due to H2 restriction in EDvSS, DB (Burkholderiales, Rhodocyclales, and Hydrogenophilales) outcompeted SRB, although SRB were present (mostly Desulfovibrionales). However, the DB phylotypes responded differently to H2 availabilities, with the autotrophic phylotype Methyloversatilis becoming more important with greater H2 availability. Under nonlimiting H2 conditions (in EAvSS), SRB declined with increasing inline image loadings, but survived within the biofilm. Lastly, inline image reduction showed a negative impact on the homoacetogen Holophagales, which demonstrates competition between SRB for electron donor in EDvSS and space in EAvSS, and a positive impact on the heterotroph Bacteroidales, which might grow by utilizing SMP released during inline image reduction.

Our findings demonstrate relationships between DB and SRB, along with their interactions with other members of the microbial community. The biofilm community was affected by the availability of H2 as an inorganic electron donor; the biofilm became more heterotrophic when the H2 availability was below 0.56 g H2 m−2 day−1. Likewise, a relatively low inline image loading allowed more inline image reduction and caused the microbial community to shift to more SRB.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgement
  7. References
  8. Supporting Information

This research was funded by the Environmental Security Technology Certification Program (ESTCP) by grant ER-200541 and by the Consejo Nacional de Ciencia y Tecnologia (CONACYT) as scholarship granted to Aura Ontiveros-Valencia.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgement
  7. References
  8. Supporting Information
  • Beller HR, Chain PSG, Letain TE, Chakicherla A, Larimer FW, Richardson PM, Coleman MA, Wood AP & Kelly DP (2006) The genome sequence of the obligately chemolithoautotrophic, facultative anaerobic bacterium Thiobacillus denitrificans. J Bacteriol 188: 14731488.
  • Bentzen G, Smith AT, Benett D, Webster NJ, Reinholt F, Sletholt E & Hobson J (1995) Controlling dosing of nitrate for prevention of H2S in a sewer network and the effects of the subsequent treatment processes. Water Sci Technol 31: 293302.
  • van Breemen N & van Dijk HFG (1988) Ecosystems effects of atmospheric deposition of nitrogen in the Netherlands. Environ Pollut 54: 249274.
  • Caporaso JG, Kuczynski J, Stombaugh J et al. (2010a) QIIME allows analysis of high-throughput community sequencing data. Nat Methods 7: 335336.
  • Caporaso JG, Bittinger K, Bushman FD, DeSantis TZ, Andersen GL & Knight R (2010b) PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics 26: 266267.
  • Chen C, Wang AJ, Ren NQ, Lee DJ & Lai JY (2009a) High-rate denitrifying sulfide removal process in expanded granular sludge bed reactor. Bioresour Technol 100: 23162319.
  • Chen C, Wang AJ, Ren NQ, Den XL & Lee DJ (2009b) Optimal process pattern for simultaneous sulfur, nitrogen and carbon removal. Water Sci Technol 59: 833837.
  • Chen C, Ren NQ, Wang AJ, Liu LH & Lee DJ (2010) Functional consortium for denitrifying sulfide removal process. Appl Microbiol Biotechnol 86: 353358.
  • Chung J, Nerenberg R & Rittmann BE (2006a) Bio-reduction of selenate using a hydrogen-based membrane biofilm reactor. Environ Sci Technol 40: 16641671.
  • Chung J, Nerenberg R & Rittmann BE (2006b) Bio-reduction of soluble chromate using a hydrogen-based membrane biofilm reactor. Water Res 40: 16341642.
  • Chung J, Rittmann BE, Wright WF & Bowman RH (2007) Simultaneous bio-reduction of nitrate, perchlorate, selenate, chromate, arsenate, and dibromochloropropane using a hydrogen-based membrane biofilm reactor. Biodegradation 18: 199209.
  • Chung J, Krajmalnik-Brown R & Rittmann BE (2008) Bioreduction of trichloroethene using a hydrogen-based membrane biofilm reactor. Environ Sci Technol 42: 477483.
  • Church MJ (2008) Resource control of bacterial dynamics in the sea. Microbial Ecology in the Oceans, 2nd edn (Kirchman DL, ed.), pp. 335382. Wiley & Sons, Hoboken, NJ.
  • Coates JD, Ellis DJ, Gaw CV & Lovley DR (1999) Geothrix fermentans gen. nov., sp. nov., a novel Fe(III)-reducing bacterium from a hydrocarbon-contaminated aquifer. Int J Sys Bacteriol 49: 16151622.
  • Dalsgaard T & Bak F (1994) Nitrate reduction in a sulfate-reducing bacterium, Desulfovibrio desulfuricans, isolated from rice paddy soil: sulfide inhibition, kinetics, and regulation. Appl Environ Microbiol 60: 291297.
  • DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, Keller K, Huber T, Dalevi D, Hu P & Andersen GL (2006) Greengenes, a chimera-checked 16s rRNA gene database and workbench compatible with ARB. Appl Environ Microbiol 72: 50695072.
  • Dilling W & Cypionka H (1990) Aerobic respiration in sulfate-reducing bacteria. FEMS Microbiol Lett 71: 123128.
  • Drake HL, Kusel K & Matthies C (2002) Ecological consequences of the phylogenetic and physiological diversities of acetogens. Antonie Van Leeuwenhoek 81: 203213.
  • Edgar RC (2010) Search and clustering orders of magnitude faster than blast. Bioinformatics 26: 24602461.
  • Faith DP (1992) Conservation evaluation and phylogenetic diversity. Biol Conserv 61: 110.
  • Fields MW, Bagwell CE, Carroll SL, Yan T, Liu X, Watson DB, Jardine PM, Criddle CS, Hazen TC, Zhou J (2006) Phylogenetic and functional biomakers as indicators of bacterial community responses to mixed-waste contamination. Environ Sci Technol 40: 26012607.
  • Garcia de Lomas J, Corzo A, Gonzalez JM, Andrades JA, Iglesias E & Montero MJ (2005) Nitrate promotes biological oxidation of sulfide in wastewaters: experiment at plant scale. Biotechnol Bioeng 4: 801811.
  • van Ginkel SW, Lamendella R, Kovacik WP Jr, Santo Domingo JW & Rittmann BE (2010) Microbial community structure during nitrate and perchlorate reduction in ion-exchange brine using the hydrogen-based membrane biofilm reactor (MBfR). Bioresource Technol 101: 37473750.
  • Gu B, Wu W-M, Ginder-Vogel MA, Yan H, Fields MW, Zhou J, Fendorf S, Criddle CS & Jardine P (2005) Bioreduction of uranium in a contaminated soil column. Environ Sci Technol 39: 48414847.
  • Haas BJ, Gevers D, Earl AM et al. (2011) Chimeric 16s rRNA sequence formation and detection in Sanger and 454-pyrosequenced PCR amplicons. Genome Res 21: 494504.
  • Kindaichi T, Ito T & Okabe S (2004) Ecophysiological interaction between nitrifying bacteria and heterotrophic bacteria in autotrophic nitrifying biofilms as determined by microautoradiography fluorescence in situ hybridization. Appl Environ Microbiol 70: 16411650.
  • Knowles R (1982) Denitrification. Microbiol Rev 46: 4370.
  • Lee KC & Rittmann BE (2002) Applying a novel autohydrogenotrophic hollow-fiber membrane biofilm reactor for denitrification of drinking water. Water Res 36: 20402052.
  • Li G, Park S, Kang D, Krajmalnik-Brown R & Rittmann BE (2011) 2-4-5 Trichlorophenol degradation using a novel TiO2-coated biofilm carrier: roles of adsorption, photocatalysis, and biodegradation. Environ Sci Technol 45: 83598367.
  • Lovett GM (1994) Atmospheric deposition of nutrients and pollutants in North America: an ecological perspective. Ecol Appl 4: 629650.
  • Lozupone C, Hamady M & Knight R (2006) UniFrac - an online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinformatics 7: 371385.
  • Marschall C, Frenzel P & Cypionka H (1993) Influence of oxygen on sulfate-reduction and growth of sulfate-reducing bacteria. Arch Microbiol 159: 168173.
  • Mateju V, Cizinska S, Krejci J & Janoch T (1992) Biological water denitrification – a review. Enzyme Microb Technol 14: 170183.
  • Matsuzaka E, Nomura N, Maseda H, Otagaki H, Nakajima-Kambe T, Nakahara T & Uchiyama H (2003) Participation of nitrite reductase in conversion of NO2- to NO3- in a heterotrophic nitrifier, Burkholderia cepacia NH-17, with denitrification activity. Microbes Environ 18: 203209.
  • Merkey BV, Rittmann BE & Chopp DL (2009) Modeling how soluble microbial products (SMP) support heterotrophs in autotroph-based biofilms. J Theor Biol 259: 670683.
  • Mohanakrishnan J, Wegener Kofoed MV, Barr J, Yuan Z, Schramm A & Louise Meyer R (2011) Dynamic microbial response of sulfidogenic wastewater biofilm to nitrate. Appl Microbiol Biotechnol 91: 16471657.
  • Moura JG, Gonzalez P, Moura I & Fauque G (2007) Dissimilatory nitrate and nitrite ammonification by sulphate-reducing eubacteria. Sulphate-reducing Bacteria. Environmental and Engineered Systems (Barton LL & Hamilton WA, eds), pp. 241264. Cambridge University Press, New York.
  • Muyzer G & Stams AJM (2008) The ecology and biotechnology of sulphate-reducing bacteria. Nat Rev Microbiol 6: 441454.
  • Nagata T (2008) Organic matter bacteria interactions in seawater. Microbial Ecology in the Oceans, 2nd edn (Kirchman DL, ed.), pp. 207242. Wiley & Sons, Hoboken, NJ.
  • Nerenberg R & Rittmann BE (2002) Perchlorate as a secondary substrate in a denitrifying hollow-fiber membrane biofilm reactor. Water Sci Technol 2: 259265.
  • Nerenberg R, Kawagoshi Y & Rittmann BE (2008) Microbial ecology of a hydrogen-based membrane biofilm reactor reducing perchlorate in the presence of nitrate or oxygen. Water Res 42: 11511159.
  • Ni B-J, Rittmann BE & Yu H-Q (2011) Soluble microbial products and their implications in mixed culture biotechnology. Trends Biotechnol 29: 254263.
  • Ontiveros-Valencia A, Ziv-El M, Zhao H, Feng L, Rittmann BE & Krajmalnik-Brown R (2012) Interactions between nitrate-reducing and sulfate-reducing bacteria coexisting in a hydrogen-fed biofilm. Environ Sci Technol 46: 1128911298.
  • Payne WJ (1973) Reduction of nitrogenous oxides by microorganisms. Bacteriol Rev 37: 409452.
  • Payne W (1981) Denitrification. Wiley, New York, NY, pp. 214.
  • Peck HD Jr (1959) The ATP-dependent reduction of sulfate with hydrogen in extracts of Desulfovibrio desulfuricans. P Natl Acad Sci USA 45: 701708.
  • Price MN, Dehal PS & Arkin AP (2009) FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Mol Biol Evol 26: 16411650.
  • Reyes-Avila J, Razo-Flores E & Gomez J (2004) Simultaneous biological removal of nitrogen, carbon and sulfur by denitrification. Water Res 38: 33133321.
  • Rittmann BE & McCarty PL (2001) Environmental Biotechnology: Principles and Applications. McGraw-Hill Companies, Inc, New York.
  • Saunders NFW, Ferguson SJ & Baker SC (2000) Transcriptional analysis of the nirS gene, encoding cytochrome cd1 nitrite reductase, of Paracoccus pantotrophus LMD 92.63. Microbiology 146: 509516.
  • Shao M-F, Zhang T & Han-Ping Fang H (2010) Sulfur-driven autotrophic denitrification: diversity, biochemistry, and engineering applications. Appl Microbiol Biotechnol 88: 10271042.
  • Sun Y, Wolcott RD & Dowd SE (2011) Tag-encoded FLX amplicon pyrosequencing for the elucidation of microbial and functional gene diversity in any environment. High-Throughput Next Generation Sequencing. Methods Mol Biol 733: 129141.
  • Tang Y, Ontiveros-Valencia A, Liang F, Zhou C, Krajmalnik-Brown R & Rittmann BE (2012a) A model to understand the onset of sulfate reduction in denitrifying membrane biofilm reactors. Biotechnol Bioeng 100: 763772. DOI: 10.1002/bit.24755.
  • Tang Y, Zhou C, van Ginkel S, Ontiveros-Valencia A, Shin J & Rittmann BE (2012b) Hydrogen-Permeation Coefficients of the Fibers Used in H2-Based Membrane Biofilm Reactors. J Membrane Sci 407-408: 176183.
  • US Environmental Protection Agency (2011) Sulfate in drinking water. Retrieved on February 19, 2012 from http://water.epa.gov/drink/contaminants/unregulated/sulfate.cfm
  • US Environmental Protection Agency (2012) Basic information about nitrate in drinking water. Retrieved on February 19, 2012 from: http://water.epa.gov/drink/contaminants/basicinformation/nitrate.cfm
  • Wang Q, Garrity GM, Tiedje JM & Cole JR (2007) Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol 73: 52615267.
  • Widdel F & Hansen TA (1991) The dissimilatory sulfate-and sulfur-reducing bacteria. The Prokaryotes, 2nd edn (Balows A, Trüper HG, Dworking M, Harder W & Schleifer K-H, eds), pp. 583616. Springer-Verlag, New York, NY.
  • Yoshida M, Ishii S, Otsuka S & Senoo K (2010) nirK-Harboring denitrifiers are more responsive to denitrification inducing conditions in rice paddy soil than nirS-Harboring bacteria. Microbes Environ 25: 4548.
  • Zhang H, Ziv-El M, Rittmann BE & Krajmalnik-Brown R (2010) Effect of dechlorination and sulfate reduction on the microbial community structure in denitrifying membrane-biofilm reactors. Environ Sci Technol 44: 51595164.
  • Zhao H, van Ginkel S, Tang Y, Kang D-W, Rittmann BE & Krajmalnik-Brown R (2011) Interactions between perchlorate and nitrate reductions in the biofilm of a hydrogen-based membrane biofilm reactor. Environ Sci Technol 45: 1015510162.
  • Ziv-El M & Rittmann BE (2009) Systematic evaluation of nitrate and perchlorate bioreduction kinetics in groundwater using a hydrogen-based membrane biofilm reactor. Water Res 43: 173181.
  • Ziv-El M, Popat SC, Kai C, Halden RU, Krajmalnik-Brown R & Rittmann BE (2012) Managing methanogens and homoacetogens to promote reductive dechlorination of trichloroethene with direct delivery of H2 in a membrane biofilm reactor. Biotechnol Bioeng 109: 22002210.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgement
  7. References
  8. Supporting Information
FilenameFormatSizeDescription
fem12107-sup-0001-FigS1-S3-TableS1.docxWord document226K

Fig. S1. Number of unique, shared, and total OTUs per reactor.

Fig. S2. Rarefraction curves at 95% confidence.

Fig. S3. Relative abundances of phylotypes at the phyla and class level for EDvSS and EAvSS.

Table S1. Diversity and evenness metrics for EDvSS and EAvSS at a similarity level of 95%.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.