An outbreak of nonflocculating catabolic populations caused the breakdown of a phenol-digesting activated-sludge process

Abstract

Activated sludge was fed phenol as the sole carbon source, and the phenol-loading rate was increased stepwise from 0.5 to 1.0 g liter-1 day-1 and then to 1.5 g liter-1 day-1. After the loading rate was increased to 1.5 g liter-1 day-1, nonflocculating bacteria outgrew the sludge, and the activated-sludge process broke down within 1 week. The bacterial population structure of the activated sludge was analyzed by temperature gradient gel electrophoresis (TGGE) of PCR-amplified 16S ribosomal DNA (rDNA) fragments. We found that the population diversity decreased as the phenol-loading rate increased and that two populations (designated populations R6 and R10) predominated in the sludge during the last several days before breakdown. The R6 population was present under the low-phenol-loading-rate conditions, while the R10 population was present only after the loading rate was increased to 1.5 g liter-1 day-1. A total of 41 bacterial strains with different repetitive extragenic palindromic sequence PCR patterns were isolated from the activated sludge under different phenol-loading conditions, and the 16S rDNA and gyrB fragments of these strains were PCR amplified and sequenced. Some bacterial isolates could be associated with major TGGE bands by comparing the 16S rDNA sequences. All of the bacterial strains affiliated with the R6 population had almost identical 16S rDNA sequences, while the gyrB phylogenetic analysis divided these strains into two physiologically divergent groups; both of these groups of strains could grow on phenol, while one group (designated the R6F group) flocculated in laboratory media and the other group (the R6T group) did not. A competitive PCR analysis in which specific gyrB sequences were used as the primers showed that a population shift from R6F to R6T occurred following the increase in the phenol-loading rate to 1.5 g liter-1 day-1. The R10 population corresponded to nonflocculating phenol-degrading bacteria. Our results suggest that an outbreak of nonflocculating catabolic populations caused the breakdown of the activated-sludge process. This study also demonstrated the usefulness of gyrB-targeted fine population analyses in microbial ecology.

FIG. 1

Phylogenetic trees based on the nucleotide sequences of 16S rDNA and gyrB genes, showing the relationships among the bacteria isolated from the phenol-digesting activated sludge. Pseudomonas putida JCM 6156 was used as the outgroup. The 16S rDNA sequences of previously described strains were retrieved from GenBank, and the accession numbers were as follows: Comamonas testosteroni, D87101; Comamonas terigena, AF078772; Comamonas acidovorans, AB020186; Comamonas sp. strain E6, AB008429; Variovorax paradoxus, D30793; Rubrivivax gelatinosus, D16213; Burkholderia cepacia, L28675; Neisseria gonorrhoeae, X07714; and P. putida, D37924. The gyrB sequences were retrieved from the ICB database (16); the accession numbers of these sequences were gy00024, gy10051, gy00022, gy10391, gy10066, gy10003, gy10339, gy00036, and gy10157, respectively. Bars = 0.025 substitution per amino acid site. f, floc formed; t, turbid; —, not grown.

FIG. 2

(a) Changes in the phenol concentration (□) and TOC concentration (⧫) in response to stepwise increases in the phenol-loading rate. (b) Changes in the specific phenol-oxygenating activity of activated sludge in response to stepwise increases in the phenol-loading rate.

FIG. 3

(a) TGGE profiles of the partial 16S rDNA fragments, showing shifts in major bacterial populations in phenol-digesting activated sludge in response to stepwise increases in the phenol-loading rate. (b) Drawing of the TGGE gel from panel a, showing the bands excised for sequence analysis.

FIG. 4

Dynamics of the R6F (□), R6T (⧫), and R10 (○) populations in the phenol-digesting activated sludge, as determined by the gyrB-targeted cPCR. The total direct count (▴) is also shown. Each datum point is the mean based on two or three determinations, and each error bar indicates the standard error.