ABSTRACT: The population dynamics of Archaea after flooding of an Italian rice field soil were studied over 17 days. Anoxically incubated rice field soil slurries exhibited a typical sequence of reduction processes characterized by reduction of nitrate, Fe(3+), and sulfate prior to the initiation of methane production. Archaeal population dynamics were followed using a dual approach involving molecular sequence retrieval and fingerprinting of small-subunit (SSU) rRNA genes. We retrieved archaeal sequences from four clone libraries (30 each) constructed for different time points (days 0, 1, 8, and 17) after flooding of the soil. The clones could be assigned to known methanogens (i.e., Methanosarcinaceae, Methanosaetaceae, Methanomicrobiaceae, and Methanobacteriaceae) and to novel euryarchaeotal (rice clusters I, II, and III) and crenarchaeotal (rice clusters IV and VI) lineages previously detected in anoxic rice field soil and on rice roots (R. Grosskopf, S. Stubner, and W. Liesack, Appl. Environ. Microbiol. 64:4983-4989, 1998). During the initiation of methanogenesis (days 0 to 17), we detected significant changes in the frequency of individual clones, especially of those affiliated with the Methanosaetaceae and Methanobacteriaceae. However, these findings could not be confirmed by terminal restriction fragment length polymorphism (T-RFLP) analysis of SSU rDNA amplicons. Most likely, the fluctuations in sequence composition of clone libraries resulted from cloning bias. Clonal SSU rRNA gene sequences were used to define operational taxonomic units (OTUs) for T-RFLP analysis, which were distinguished by group-specific TaqI restriction sites. Sequence analysis showed a high degree of conservation of TaqI restriction sites within the different archaeal lineages present in Italian rice field soil. Direct T-RFLP analysis of archaeal populations in rice field soil slurries revealed the presence of all archaeal lineages detected by cloning with a predominance of terminal restriction fragments characteristic of rice cluster I (389 bp), Methanosaetaceae (280 bp), and Methanosarcinaceae/rice cluster VI (182 bp). In general, the relative gene frequency of most detected OTUs remained rather constant over time during the first 17 days after flooding of the soil. Most minor OTUs (e.g., Methanomicrobiaceae and rice cluster III) and Methanosaetaceae did not change in relative frequency. Rice cluster I (37 to 30%) and to a lesser extent rice cluster IV as well as Methanobacteriaceae decreased over time. Only the relative abundance of Methanosarcinaceae (182 bp) increased, roughly doubling from 15 to 29% of total archaeal gene frequency within the first 11 days, which was positively correlated to the dynamics of acetate and formate concentrations. Our results indicate that a functionally dynamic ecosystem, a rice field soil after flooding, was linked to a relatively stable archaeal community structure.
