of Evolution the bacterial community during granules formation in denitrifying reactors followed by molecular, culture-independent techniques
C. Etchebehere*, A. Cabezas*, P. Dabert** and L. Mux*
* Ctedra de Microbiologa, Facultad de Ciencias, Facultad de Qumica, UDELAR, Gral Flores 2124 CC1157, Montevideo,Uruguay (E-mail: [e-mail]; [e-mail]; [e-mail]) ** Laboratoire de Biotecnologie de LEnvironnement, Institut National de la Recherche Agronomique, Av. Des Etangs 11100, Narbonne, France (E-mail: [e-mail]) Abstract The microbial community in two acetate-fed denitrifying reactors, inoculated with methanogenic sludge, was monitored by 16S rDNA-based methods (SSCP and FISH). Both reactors converged to similar, stable communities. The predominant organisms belonged to the genera Thauera, Paracoccus and Denitrobacter, detected both by molecular and culture-based methods. Keywords Denitrification; FISH; granulation; inocula; sludge bed anoxic reactor; SSCP
Water Science and Technology Vol 48 No 6 pp 7579 IWA Publishing 2003
Introduction
Anoxic UASB reactors have been successfully used for nitrate removal in integrated systems (anaerobic, aerobic, anoxic) treating effluents with high organic and nitrogen content (Borzacconi et al., 1999; Morgan-Sagastume et al., 1994). However, there are few reports on the parameters that affect granulation of the denitrifying biomass (Cuervo-Lpez et al., 1999; Etchebehere et al., 2002). The effect of inocula on sludge granulation in denitrifying reactors was studied in two laboratory scale UASB denitrifying reactors, fed with acetate and nitrate and seeded with methanogenic granules (Etchebehere et al., 2002). The evolution of the microflora was studied by culture-dependent methods, and organisms belonging to the same species within the genus Thauera were isolated from both reactors, suggesting a role for such organisms in acetate-fed denitrifying reactors. However, in order to evaluate the denitrifying community, it is necessary to apply methods that detect both culturable and non-culturable organisms. Recently, several 16S rDNA-based methods have been developed. Among them, fluorescent in situ hybridization (FISH) reveals whole cells and has been used to analyse the microbial community structure of many ecosystems (Amann et al., 1990). Single Strand Conformation polymorphism (SSCP) is based on the different electrophoretic mobility of 16S rDNA molecules with different sequences (Orita et al., 1989) and was successfully applied to study population dynamics in anaerobic digesters (Zumstein et al., 2000) and in phosphorus removal SBR (Dabert et al., 2001). The aim of the present work was to study the evolution of the microflora developed in both reactors by culture-independent methods.
Materials and methods
Two upflow reactors (4.6 L) were seeded with intact and ground granules from an anaerobic UASB reactor treating brewery wastewater, and fed with acetate and nitrate (COD/N-NO3 = 4) (Etchebehere et al., 2002).
Sludge SSCP analysis
Sludge sample DNA was extracted using a commercial kit (UltraClean soil DNA kit, Mo Bio). The V3 region of the bacterial 16S rDNA was amplified according to Zumstein et al.
75
(2000). The PCR products were heat denatured and separated according to their size and structure by capillary electrophoresis in an automatic sequencer (Abi prism 310 Perkin Elmer, Applied Biosystems). Data processing was performed with Genescan software (Applied Biosystems). Microbial communities appeared as profiles of peaks where each peak represents a dominant organism in the microbial population from the ecosystem. The characterisation of the predominant profile peaks was performed by cloning and SSCP analysis of the cloned DNA as described previously (Dabert et al., 2001). Clones with single peak profiles, matching total DNA SSCP profile peaks, were selected for sequence identification. Sequencing was performed as described by Zumstein et al. (2000). The sequences were compared with all sequences available in databases using BLAST Search at NCIB.
SSCP analysis of isolated strains
76
C. Etchebehere et al.
Sixteen denitrifying strains were isolated from both reactors (Etchebehere et al., 2002) and grouped according to their ARDRA profile. 16S rDNA of the most frequently isolated strains was analysed by SSCP. Strains showing different SSCP profiles were selected and their 16S rDNA was partially sequenced as previously described (Etchebehere et al., 2001). The sequences were compared with all sequences available in databases using BLAST Search at NCIB.
Fluorescent in situ hybridisation (FISH)
FISH was performed on sludge samples as described by Amann et al. (1990), using probes: Alf968 to detect alpha Proteobacteria (Neef, 1997); BET42a and GAM42a (Manz, 1992) to detect beta and gamma Proteobacteria respectively; Eub338 directed to the Bacteria Domain and Arch915 directed to the Archaea Domain (Amann et al., 1990). For quantitative cells analysis, results were expressed as a percentage of cells targeted by a probe versus the total number of cells counterstained with 4,6-diamidino-2-phenylindole (DAPI).
Results and discussion
Evolution of the microbial community in denitrifying reactors
The results of the sludge SSCP analysis at different times of the denitrifying reactors (R1 and R2) are shown in Figure 1. As expected, important shifts in the microbial community structure were detected during the first period of reactor operation, suggesting adaptation of the methanogenic population to denitrifying conditions. This was in accordance with reactor performances (Etchebehere et al., 2001). The high number of peaks detected in the inoculum reflects the high diversity of bacterial organisms generally found in this kind of community, capable of degrading complex substrates in methanogenic conditions (Zumstein et al., 2000). An apparently less diverse community was selected in both reactors during operation, as shown by the smaller number of dominant peaks. Furthermore, the community in both reactors evolved in the same way. Over all profiles, seven dominant peaks were observed: Peaks A, D, F and G were detected at different times, indicating that organisms represented by these peaks persisted during reactor operation. Moreover, peaks E and G increased during time, suggesting that organisms represented by these peaks became dominant.
Identification of the SSCP peaks, comparison with clones and strains SSCP profiles
In order to identify the predominant peaks two strategies were followed: cloning and SSCP analysis of the clones, and SSCP analysis of the strains isolated from the same reactors. For the cloning approach two sludge samples were selected and a library was performed for each one, library B from sample R1 (t = 4 months) and library C from R1 (t = 11
t=0 inoculum inoculum
C. Etchebehere et al.
D
F E G
R1 R2
t=3 months
A
B
C
D
R1 R2
A
B
C
F E G
t=4 months
D
F E
G
t=10 months
R1 R2
A
B
C
D
F E
G
R1
A B
t=11 months
C
R2
Figure 1 SSCP profiles of reactor sludge samples, at different operation times. R1: reactor inoculated with intact granules, R2: reactor inoculated with ground granules
months). The SSCP analysis of the clones showed that seventeen clones out of forty matched peaks present in the sludge SSCP profile. Eleven out of sixteen strains, isolated from both reactors at different times, presented SSCP profiles matching peaks of the sludge SSCP profile (Table 1). 16S The rDNA partial sequence analysis of the clones and strains allowed us to assign peaks A, D and F, to organisms belonging to the genera Paracoccus (peak A) and Thauera
Table 1 Comparison of clones and strains SSCP profiles with sludge SSCP peaks. Organisms identification according to 16S rDNA partial sequence comparison (BLAST Search)
SSCP profile from clones and strains Number of clones from library C Number of clones from library B Strains Sludge SSCP peak Sequence identification of representative clones and strains
1 2 3 4
1 10 1 1
2 2
6 1 4
F G A D
Thauera Denitromonas Paracoccus Thauera
77
Percentage of DAPI stained cells (%)
80 70 60 50 40 30 20 10 0
Ba ct e ria alp ha
ae a
Figure 2 FISH analysis of reactor samples R1(
Ar ch
) and R2 (
) taken at 11 months
(peaks D and F) in the alpha and beta subgroup of Proteobacteria. These peaks were detected in the SSCP analysis of all the sludge samples, indicating that these organisms were present in both reactors and persisted during operation. Peak G, which increased in both reactors during operation, matched peaks from SSCP profiles of ten clones from Library C and one strain, isolated towards the end of reactor operation. These clones and strain showed sequences closely related to the genus Denitromonas in the beta Proteobacteria, suggesting that this organism was enriched in the conditions of the reactors.
FISH
In both reactors a high proportion of the Bacteria cells hybridised with the alpha and beta Proteobacteria probes (Figure 2). No organism hybridising with the gamma subgroup was detected (data not shown). A low proportion of Archaea cells were detected in both samples. The sum of the alpha and beta cell percentages was higher than the Bacteria hybridising cells in the sample from R1, this inconsistency may be due to some non-specific hybridisation of the probes used and/or to the high quantification error caused by the presence of aggregated cells.
Conclusions
78
In both denitrifying reactors the microbial community evolved in the same way, independently of the physical structure of the inoculum. A specialised denitrifying community was selected in both reactors with persistence of organisms from the genus Thauera and Paracoccus, as detected by culture and non-culture based methods. In spite of the high diversity of known denitrifiers, these genera prevailed in both UASB denitrifying reactors, suggesting that they present a selective advantage under the operation conditions. Several strains isolated during reactor operation and characterised as belonging to these genera, presented adhesion properties, as reflected by aggregates formation in liquid media and colony adhesion to solid media. It can be speculated that these strains were selected during granule formation because of their ability to adhere. Sludge flotation problems were reported in these reactors (Etchebehere et al., 2002), probably due to bubble retention inside the aggregates. The same observation was made in pure cultures of the isolates. Organisms identified as Denitromonas increased and became dominant during reactor operation. Pure cultures of these strains did not form aggregates during growth. Further work is necessary to investigate the role of this organism in UASB denitrifying reactors.
be
C. Etchebehere et al.
ta
No organisms affiliated to the genus Pseudomonas were detected by culture and non culture based methods, this indicates the difference between the denitrifying microbiota present in the denitrifying reactors and the most frequent denitrifiers isolated from soil (Cheneby et al., 2000). Results from FISH analysis showed that a community enriched in alpha and beta Proteobacteria was selected in both reactors; in accordance with the SSCP analysis and with previous culture-based analysis of these communities (Etchebehere et al., 2002). A low amount of Archaea cells were detected at the end of reactor operation. This was expected since Archaea were present in high proportions in the methanogenic inoculum (Etchebehere et al., 2002). However, their amount decreased with time during reactor operation in denitrifying conditions, probably due to the high redox potential present in denitrifying conditions. SSCP allows the monitoring of reactor microbial community dynamics without prior knowledge of their composition. However, it cannot be used to quantify microbial populations because of potential PCR biases. By using the FISH technique, it is possible to count the orga...