Development of a novel cultivation technique for uncultured soil bacteria

Abstract

In this study, a new diffusion bioreactor was developed to cultivate hidden bacterial communities in their natural environment. The newly developed method was investigated to cultivate microbial communities from the forest soil, and the results were evaluated against traditional culture methods and compared to the results of a pyrosequencing-based molecular survey. The molecular analysis revealed that a diverse bacterial population was present in the soil sample. However, both the newly developed method and the traditional method recovered more than 400 isolates, which belonged to only four phyla: Proteobacteria, Firmicutes, Actinobacteria, and Bacteroidetes. Although these isolates were distributed over only four major phyla, the use of the newly developed technique resulted in the successful cultivation of 35 previously uncultured strains, whereas no such strains were successfully cultivated by the traditional method. Furthermore, the study also found that the recovery of uncultured bacteria and novel isolates was related to sampling season, incubation period, and cultivation media. The use of soil collected in summer, a prolonged incubation period, and low-substrate modified media increased the recovery of uncultured and novel isolates. Overall, the results indicate that the newly designed diffusion bioreactor can mimic the natural environment, which permits the cultivation of previously uncultured bacteria.

Figure 1

Scheme of the optimized protocol for the cultivation of previously uncultured microorganisms from forest soil.

Figure 2

Design of the diffusion bioreactor for the cultivation of previously uncultured bacteria. (a) Plastic container perforated throughout with holes; (b) polycarbonate membrane glued inner chamber; (c) schematic diagram of the diffusion bioreactor; (d) overall experimental setup using the newly developed diffusion bioreactor.

Figure 3

Relative abundances and taxonomic identification of bacteria revealed by pyrosequencing analysis. A total of 31 bacterial phyla were detected, 15 of which were previously known phyla and 16 were unclassified phyla.

Figure 4

Taxonomic distribution of total pure cultured bacteria isolated from forest soil samples. The bacterial diversity of pure isolates was associated with one of eight classes: Actinobacteria, Bacilli, Sphingobacteria, Flavobacteria, Cytophagia, α-Proteobacteria, β-Proteobacteria, and γ-Proteobacteria.

Figure 5

Comparison of the recovery rate of bacterial isolates using the diffusion bioreactor and the traditional method. (a) Comparison of previously uncultured and novel isolates obtained using the diffusion bioreactor and traditional method; (b) effect of soil sampling period in the recovery of novel/uncultured isolates using both the diffusion bioreactor and the traditional method. The bars represent the standard error of the replicate experiments with duplicate samples.

Figure 6

Evaluation of incubation time and medium composition on the recovery of novel/uncultured isolates. (a) Effect of incubation period; (b) effect of various media used during the cultivation period. D: Represents diffusion bioreactor; and T: represents the traditional method. In this study, isolates sharing a 16S rRNA gene sequence similarity below the threshold value of 98.7–99.0% were considered to be novel strains, and isolates affiliated with the previously uncultured strains during the 16S rRNA gene sequence analysis were considered to be previously uncultured bacteria.

Figure 7

Phylogenetic tree (maximum likelihood tree) based on nearly complete 16S rRNA gene sequences showing the taxonomic affiliations of previously uncultured bacteria isolated using the newly developed diffusion bioreactor with representatives of different established phylogenetic groups. The numbers at the nodes indicate the percentage of 1,000 bootstrap replicates. The bacterial taxa depicted in italics represent reference strains. GenBank accession numbers of the 16S rRNA gene sequences are shown in parentheses. The scale bar represents 0.1 substitutions per nucleotide position.