Introduction

Brazil is a tropical country with an average temperature of 25°C and has environmental conditions that favor the use of anaerobic technology. In the last 15 years, several wastewater treatment plants that use anaerobic processes as their core technology have been implemented in Brazilian municipalities. Some examples include the domestic wastewater treatment plants (WWTP) such as the following with the correspondent population: Mangueira (Recife PE), 18,000 inhabitants; the Onça (Belo Horizonte MG), one million inhabitants; the Piçarrão (Campinas SP), 200,000 inhabitants; and the Atuba Sul (Curitiba PR), 370,000 inhabitants (Jordão et al. 2007). The Mangueira upflow anaerobic sludge blanket (UASB) 810 m3 reactor with eight parallel cells has been in operation since 1997. The reactor has been monitored since the beginning of the operation, and it has proven to be a robust and stable reactor (Florencio et al. 2001; Morais et al. 2004; Barros et al. 2008).

An evaluation of the microorganisms diversity is important to understand their behavior in a UASB reactor that operates under full-scale conditions. Most studies have been conducted under bench-scale testing conditions (Pholchan et al. 2010; Fernández et al. 2009); when conducted at full-scale conditions, the diversity that is discussed only includes specific groups or conditions (Angenent et al. 2002; Werner et al. 2011).

In general, microorganisms belonging to the domains Archaea and Bacteria, which are present in anaerobic reactors used for wastewater treatment, are difficult to grow in the laboratory. Therefore, the identification of bacterial populations has been accomplished through the use of molecular tools, such as the cloning and sequencing of conserved genes like rRNA, gel electrophoresis denaturing gradient (DGGE), fluorescent in situ hybridization (FISH), and recently, through metagenomic analysis (Krause et al. 2008; Schlüter et al. 2008). Because of this, the presence of certain species in the sludge can be determined by 16S rRNA coding sequences without the need to cultivate the microorganisms.

The aim of this work was to learn more about the microorganisms present in a full-scale WWTP. Our objective was to identify microorganisms belonging to the domains Bacteria and Archaea that were present in a UASB reactor and to evaluate the population differences occurring along several levels of a sludge bed in combination with specific methanogenic activity.

Materials and methods

Sludge sampling

The anaerobic sludge used in this study was obtained from a full-scale UASB reactor, which is located at the Mangueira WWTP in Recife, Pernambuco state, Brazil. The reactor operated with an 8-h hydraulic retention time (HRT). The sludge was collected at five sampling points in the reactor (from bottom to top): A (0.8 m), B (1.3 m), C (1.8 m), D (2.3 m) and E (2.8 m). At the time of the sampling, the reactor had not been subjected to excess sludge discharge during the previous 8 months.

Specific methanogenic activity (SMA)

The SMA was determined in accordance with the works of Florencio et al. (1993) and Field et al. (1987). A mixture of acetic, propionic and butyric acids at a ratio of 1:1:1 on COD, a solution of macro and micronutrients and sludge were added into 500-mL reaction flasks with 450 mL as the working volume, resulting in a final concentration of 4 g COD L−1, 200 mL L−1 and 5 g VSS L−1, respectively.

Physico-chemical parameters

The parameters were determined according to APHA (1992) twice a week to monitor the reactor temperature, pH, suspended solids (total, fixed and volatile), COD (total and soluble), alkalinity (total and partial), volatile acids and redox potential. The parameters were determined in reactor influent and effluent for 3 months prior to the collection of sludge samples.

Total DNA extraction and PCR

The extraction protocol proposed by Moreira et al. (2005) was used for DNA extraction. The PCR reaction for the 16S rRNA gene was performed according to Nielsen et al. (1999) and Kudo et al. (1997) for the domains Bacteria and Archaea, respectively. For the domain Bacteria, the following primers were used:

  1. (i)

    968f (5′-AACGCGAAGAAC CTTAC-3′);

  2. (ii)

    1392r (5′-ACGGGCGGTGTGTAC-3′).

For the domain Archaea, the following primers were used:

  1. (i)

    1100F (5′-AACCGTCGACAGTCAGGYAACGAGCGAG-3′);

  2. (ii)

    1400r (5′-CGGCGAATTCGTGCAAGGAGCAGGGAC-3′).

The amplicons were purified with Wizard SV Gel and a PCR Clean-Up System kit (Promega Co.) in accordance with the manufacturer’s instructions.

Cloning and sequencing

The purified amplicons that corresponded to portions of the 16S rRNA gene were cloned in a pGEM-T vector (Promega Co.) in accordance with the manufacturer’s instructions; recombinant plasmids were introduced into the E. coli TOP10 strain by using the calcium chloride method. The transformed cells were selected on an LB + ampicillin medium that was supplemented with IPTG and X-gal (Sambrook and Russell 2001). White recombinant colonies were grown in an LB + ampicillin medium, and plasmids were extracted by the miniprep method. The presence of the insertion of the 16S rRNA gene was verified by the NotI restriction enzyme (New England Biolabs) in accordance with manufacturer’s instructions. The plasmid DNA (200–500 ng) was used as a PCR labeling template by using the BigDye® Terminator Cycle Sequencing kit (Applied Biosystems), with 3.2 pmol of the M13 primer in 10 μL of the final volume. The steps included initial denaturation at 94°C for 2 min, followed by 40 cycles of denaturation at 94°C for 15 s, annealing at 55°C for 10 s and extension at 60°C for 4 min. The amplification products were precipitated with ethanol/EDTA/sodium acetate and resuspended in formamide according to the manufacturer’s instructions. The sequencing of both strands was done on an ABI Prism 3100 Genetic Analyzer device (Applied Biosystems).

16S rRNA gene sequence analysis

The BioEdit 7.0.5 program and the Staden Package set of programs (http://staden.sourceforge.net/) were used to assess the quality of the sequences and the consensus sequences assembly. The consensus sequences were compared with the database of the Ribosomal Database Project–RDP (http://rdp.cme.msu.edu/) by using the Classifier program (confidence threshold = 95%) and the BLASTn program in their default settings for the nucleotide nr/nt collection GeneBank database from the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/). For species identification, a 97% minimum similarity between sequences was followed.

Results and discussion

UASB reactor performance and sludge methanogenic activity

The results presented in Table 1 indicate a strong relationship between the solids profile and the lengthy previous period without excess sludge discharge. The amount of fixed solids was significantly high and represented approximately 50% of the total solids, whereas the VSS concentrations (16.9–40.2 g L−1) were relatively low compared to values of around 50–60 g VSS L−1 found in other full-scale UASB reactors that treat domestic sewage (van Haandel et al. 2006). The concentrations of VSS tended to decrease from A to C and to increase again upwards. The high accumulation of solids possibly caused a short circuit inside the reactor, and it is possible that the withdrawal of the sludge samples was not uniform due to some sampling difficulties that were caused by pipe clogging. The accumulation of inert material inside the reactor can be attributed to the WWTP’s inefficient removal of a large amount of grit entering the sewerage. The population that is served by the WWTP lives in a low-income area where the streets are not paved and the region’s precipitation is very high (above 2,000 mm annually).

Table 1 Temperature, pH, solids and UASB reactor sludge SMA values
Full size table

Nevertheless, the average COD removal efficiency (based on total influent and filtered effluent COD) was 83%, which confirms the robustness of the reactor and the previous results (Florencio et al. 2001; Morais et al. 2004; Barros et al. 2008). However, the average removal efficiency was only 23% for the suspended solids, which is much lower than the previously obtained value of 75% (Morais et al. 2004) due to solids washout.

Regardless of the VSS concentrations, the SMA values apparently decreased from the bottom to the top, although the differences were not significant. The exception is perhaps at point E, which showed a small loss of activity probably due to the presence of less substrates and the older age of the sludge.

Reactor population belonging to the domain Bacteria

16S rRNA libraries were produced by cloning the amplification fragments generated with primers 968f and 1392r at different levels and bacterial clones were subjected to DNA sequencing. Over 54 different nucleotide sequences were produced ranging from 93 to 100% similarity to sequences deposited in the RDP database and NCBI database. Bacterial diversity was quite similar among the reactors levels (Table 2). However, for identification the minimum similarity accepted was 97%. The five identified bacterial phyla were distributed as follows: 42% belonged to the Actinobacteria phylum, 13% to the Proteobacteria phylum, 9% to the Chloroflexi phylum, 6% to the Firmicutes phylum and 4% to the Bacteroidetes phylum. The 26% of the remaining sequences did not fit in any of the phyla (Fig. 1).

Table 2 Clones representative of the different nucleotides sequences produced by the amplification of sludge total DNA with primers for bacteria
Full size table
Fig. 1
figure 1

Bacteria domain phyla distribution scheme for levels A, B, C, D and E of the UASB reactor. Representatives found: ( ) Actinobacteria—Nocardioides dokdonensis, Propionicimonas paludicola, Brooklawnia cerclae and Tessaracoccus sp., Cellulomonas sp. Pseudonocardia petroleophila. ( ) Proteobacteria—Brachymonas denitrificans, Propionivibrio limicola, Thiobacillus sp., Desulfovibrio sp. ( ) Chloroflexi—class Anaerolineae. ( ) Firmicutes—Succiniclasticum ruminis, Megasphaera elsdenii. ( ) Bacteroidetes—order Bacteroidales. ( ) Unclassified

Full size image

Actinobacteria are a group of Gram-positive bacteria that play an important role in the decomposition of organic matter and in the carbon cycle (Stackebrandt et al. 1997). Representatives of this phylum were found in all levels of the reactor. Members of the family Nocardioidaceae, such as Nocardioides dokdonensis and Propionicimonas paludicola, were found at levels A and B, respectively. Nocardioides dokdonensis is a species that is still not well known and belongs to a genus that was initially identified in soil, where many species are known to degrade hydrocarbons present in raw oil (Schippers et al. 2005). Propionicimonas paludicola is a facultative bacterium, but it grows best under anaerobic conditions by fermenting organic matter to propionate (Akasaka et al. 2003).

Members of the family Propionibacteriaceae, such as Brooklawn cerclae and Tessaracoccus sp. KSS-17Se, were found in all levels of the reactor. Brooklawn cerclae was isolated from groundwater that was contaminated with mixtures of chlorinated solvents and polycyclic aromatic hydrocarbons. They are mesophilic and produce mainly propionate and acetate from glucose (Bae et al. 2006). Tessaracoccus sp. KSS-17Se belongs to a genus of facultative anaerobes that form tetrads of cocci. An example is Tessaracoccus bendigoensis, which uses various substrates in its metabolism, such as glucose, sucrose, fructose, mannose, trehalose, valeric acid, butyric acid and propionic acid. These microorganisms are mesophilic and were previously isolated from aerobic-activated sludge systems (Maszenan et al. 1999).

Three sequences that show similarities to Cellulomonas sp. d20 were found in levels A and B of the reactor. This genus includes hydrolytic fermentative bacteria that produce acetate, ethanol and formate from complex polysaccharides, such as cellulose (Guyot 1986; An et al. 2005).

A similarity to the D3bac sequence at Genebank, which was only detected at level D, was found for the bacterium Pseudonocardia petroleophila. This genus was described by Henssen in 1957 as belonging to a group of biochemically versatile bacteria; some species may oxidize hydrocarbons, whereas others turn sulfide into sulfate or degrade cellulose (Huang et al. 2002). However, the sequence D3bac showed only 89% similarity to that bacterium. A search for similarity in RDP by using the program Classifier was also conducted. In this database, isolated D3bac was classified at the level of the phylum Actinobacteria. These results indicate that D3bac probably corresponds to a bacterium that has not yet been described but is comparable to Pseudonocardia petroleophila.

It is also worth mentioning that bacteria belonging to the Actinobacteria group were prevalent at levels A, B and C. As described, these microorganisms are mostly fermentative hydrolytic and fermentative; they are probably the most responsible for the degradation of complex organic matter that is fed into the reactor, which leads to the generation of precursor compounds for methane formation. The emergence of microorganisms that are involved in the degradation of hydrocarbons and raw oil may be the result of different substrates entering the reactor. The presence of some oily effluents or other leakages in the ground should be noted. The streets and sidewalks were not paved, and there were informal commercial shops and small-scale industrial activities in the study area. The liquid waste was able to reach the sewerage due to an inefficient drainage system that could not separately collect the rain water.

There was great diversity in the phylum Proteobacteria, which was represented by aerobic, anaerobic, photoautotrophic, photoheterotrophic and chemolithotrophic bacteria (Oren 2004). This large group of cultured bacteria is divided into five major classes: α-Proteobacteria, β-Proteobacteria, γ-Proteobacteria, δ-Proteobacteria and ε-proteobacteria. In this study, representatives of all of the classes were found, except α-Proteobacteria. In the class β-Proteobacteria, Brachymonas denitrificans was found at level A of the reactor. Initially isolated from the aerobic system (Hiraishi et al. 1995), this strain has the ability to remove phosphorus from the system under aerobic conditions and can remove nitrogen under anoxic conditions (Shi and Lee 2007). It is possible that these microorganisms were using the small amounts of oxygen that entered the reactor because they were found only at the bottom portion of the reactor.

Propionivibrio limicola strain GolChi1 T was found only in level E. It is a mesophilic anaerobic bacterium and was isolated in culture-metabolizing hydroaromatic compounds, such as hydro-quinic acid and shikimic acid. These acids are important precursors to the biosynthesis of lignin and tannins in plants, which produce propionate and acetate (Brune et al. 2002). In this work, the easily degradable organic matter that could have reached the upper part of the reactor had probably already been removed earlier. Therefore, the only remaining compounds were those that were more difficult to degrade, which explains the occurrence of this bacterium.

A clone from the γ-proteobacteria class that is 100% similar to Thiobacillus sp. was found at level C of the reactor. This organism is related to the oxidation of hydrogen sulfide in the treatment of wastewaters in reactors and of sulfur springs (Visser et al. 1997; Ravichandra et al. 2007). The presence of this bacterium in the UASB is interesting because it has been described as chemoautotrophic aerobic, which indicates that oxygen should occasionally be available in the influent that is being consumed inside the reactor. The sulfate-reducing bacteria (SRB) Desulfovibrio sp. A45 from the δ-Proteobacteria class was found at level C of the UASB. These bacteria are a group of microorganisms that live in various anaerobic environments, which include soil, domestic, industrial and mining wastewaters (Chang et al. 2001). The SRB can use many substrates, which include hydrogen, lactate, formate, malate, fumarate, pyruvate, alcohols, volatile fatty acids and various types of hydrocarbons and phenolic compounds (Rueter et al. 2002). Both sulfate and the previously described substrates are commonly present in anaerobic reactors. During complex organic matter degradation at level C of the UASB reactor, those substrates became available and could be more easily used as electron donors for the existing SRB.

A search for the D8bac, D9bac, D11bac, E6bac and E10bac sequences in the Genebank database showed maximum similarity with non-cultivable bacteria. A search in the RDP Classifier indicated that the sequences belonged to the phylum Chloroflexi and class Anaerolineae. Studies of the phylum Chloroflexi have shown that its members play an important role in the formation of granules in UASB reactors (Yamada et al. 2005).

Members of the phylum Firmicutes were found at levels A, D and E of the reactor. Succiniclasticum ruminis, which was found at level A, was first isolated from cow rumen. It is a gram-negative mesophilic bacterium that ferments succinate to propionate (van Gylswyk 1995). Megasphaera elsdenii was found at level D. It comprises a group of bacteria that live in the rumen and intestine, which are also found in anaerobic reactors. They can ferment glucose and lactate that produce valerate, propionate and butyrate (Bouallagui et al. 2004).

It is worth mentioning that the top and bottom levels showed the greatest diversity of microorganisms belonging to domain Bacteria and the highest VSS concentrations (Table 1). Layer B showed the least diversity of bacteria, despite having a VSS content comparable to that of layer E. This indicates that the reactor’s hydrodynamic regime probably affected the selection of these populations because the lowest diversity was expected to occur at layer E. At this level, the substrates should be less diversified and more difficult to degrade. It is possible that preferential paths associated with biogas and liquid movement along the height of the reactor resulted in substrate diversity at the top of the reactor. However, it is also important to consider the limitations of this work when looking for these statements since only one sample from each reactor level was taken.

Archaea population in the reactor

16S rRNA libraries were produced by cloning the amplification fragments generated with primers 1100F and 1400R at different levels and bacterial clones were subjected to DNA sequencing. Over 56 different nucleotide sequences were produced ranging from 94 to 100% similarity to sequences deposited in the RDP database and NCBI database. Archaea diversity was quite similar among the reactors levels (Table 3). According to the Classifier program, 42% of the obtained sequences belonged to the Methanomicrobiales order, 17% to the Methanobacteriales order and 12% to the Methanosarcinales order (Fig. 2). The other 17% were grouped with confidence within the phylum Euryarchaeota. However, when comparing the sequences of this last group with the Genebank database, we found organisms belonging to the genera Methanobacterium, Methanosaeta, Methanospirillum, Methanobrevibacter, and Methanolinea and the species Methanobacterium beijingense.

Table 3 Clones representative of the different nucleotides sequences produced by the amplification of sludge total DNA with primers for Archaea
Full size table
Fig. 2
figure 2

Archaea domain order distribution scheme for levels A, B, C, D and E of the UASB: ( ) Methanomicrobiales—Methanospirillum sp.; ( ) Methanobacteriales—Methanobacterium sp., Methanobrevibacter sp., Methanobacterium beijingense; ( ) Methanosarcinales—Methanosaeta sp.; ( ) Unclassified

Full size image

Methanobacterium sp. was found at levels A, B, C and D of the reactor but not in the highest level. Methanospirillum sp. and Methanobrevibacter sp. were only found at levels D and E, respectively. These hydrogenotrophic genera were also found in anaerobic reactors that were used for treating various effluents, such as those from breweries, paper mills, food manufacturing plants, phenolic compounds, whiskey production and domestic sewage (Leclerc et al. 2004). They can use H2, CO2 and formate in their metabolism for energy and CH4 production, which may have occurred in this study because these substances were detected in the upper part of the reactor.

Methanosaeta sp. was found at levels A, C, D and E of the reactor. This Archaea plays an important role in removing carbon from the system, and it can use acetate, methanol and methylamines as energy source (Laloui-Carpentier et al. 2006). This genus has also been identified in various reactors that were used for the treatment of sewage, effluent from slaughterhouses and the production of wine and paper (Leclerc et al. 2004). The fact that this genus was not detected in level B may have been caused by the hydrodynamic behavior of the reactor because the preferred substrates for this group of organisms should be available at this level of the reactor.

A new species identified as Methanolinea tarda was recently isolated from an anaerobic culture that was enriched with propionate in domestic sewage sludge in Japan. The strain uses H2 and formate as energy source and produces CH4 (Imachi et al. 2008). One of the sequences obtained in this study was observed only at level A of the reactor and showed 97% similarity to the genus Methanolinea.

Methanobacterium beijingense was found in an anaerobic reactor that was used to treat effluents from the production of beer in Beijing, China. This organism also uses H2 and formate as an energy source for the production of CH4 (Ma et al. 2005). The sequence D15arc that was found at level D showed 94% similarity to this Archaea. The relatively low percentage of similarity was mitigated by the high quality level (PHRED values between 35 and 49), which suggests that this may be a sequence of a bacterium that has not yet been described but is closely related to Methanobacterium beijingense.

We observed the marked presence of sequences that were similar to a clone of the order Methanomicrobiales, which is related to a non-cultivable organism. This clone was found at all levels of the reactor and represents 42% of all the sequences of Archaea; eight of the nine sequences were identified at level B. By using the RDP Classifier, 20 sequences with 87% similarity to the family Methanomicrobiaceae were found, and some of them showed little similarity to the genus Methanoculleus. It is notable that the sample from level B showed the lowest specific methanogenic activity and the lowest diversity level of microorganisms belonging to the domain Archaea. We also detected a reduced level of bacteria diversity at this level.

Conclusion

The results for the domain Archaea confirm the occurrence of major groups of methane-producing microorganisms in the reactor used in this study. The traditional process via acetate and H2/CO2 must have been used by microorganisms to produce methane.

The wide diversity of Bacteria and Archaea was higher in full-scale UASB upper and bottom layers showing the highest VSS levels, with middle’s less diversity and lowest SMA values. Reactor hydrodynamics influenced microorganisms selection and substrate transport. Actinobacteria hydrolysing complex organics dominated the lower layers. Archaea was well detected but with the highest SMA (0.37 g CODCH4/g SSV.d, based on VFA mixture) in the bottom. The majority of Archaea indicated predominance of methane production pathway from H2–CO2. Despite similar bottom’s diversity, the top SMA was lower due the predominance of Proteobacteria like the sulfate reducers that competed with methanogens for hydrogen and acetate.