In this study, we examined the phylogenetic diversity in a 13,001-clone library derived from an undisturbed tall grass prairie site in central Oklahoma. We used the data set to access low-abundance, rarely sampled microbial species in soil and examined their phylogenetic affiliations, similarity to current global 16S rRNA gene inventories, and relationship to more abundant members of the soil microbial community.
Based on our evaluation of the novelty of rare clones (Fig. ), their phylogenetic affiliations (Fig. ; see also Files S4 to S6 in the supplemental material), as well as their relationships to more abundant members of the community (Fig. ), we broadly identify two main groups within the KFS rare biosphere: those with close relatives among the more abundant members of the KFS bacterial community, and those that belong to unique, phylogenetically distinct lineages with no close sequence similarity to more abundant members of KFS. Using 15% sequence divergence from the closest abundant relative within the KFS data set as an empirical “uniqueness” cutoff, members of group I represent 83.6 to 92.1% and members of the second group represent 7.9 to 16.4% of the total number of rare KFS clones (at n = 1 and n ≤ 5, respectively) (Fig. ). Similar to more abundant members of the community, members of group I belong to common, well-described, and well-sampled soil lineages. On the other hand, members of the unique group II usually belong to novel bacterial phyla, novel lineages within previously described phyla and candidate phyla, or are members of lineages that are ubiquitous in specific environments but rarely encountered in soils. We reason that these novelty and uniqueness patterns provide clues regarding the origins and potential ecological roles of members of the soil's rare biosphere.
The close sequence similarity between nonunique members of the rare biosphere (group I) and dominant OTUs within the community argues against an old, evolutionary distinct origin for this fraction of the rare biosphere, as previously suggested (32
). Various lines of ecological evidence suggest that this group of nonunique, nonnovel members of the rare biosphere acts as a backup system and readily responds to seasonal variations encountered in soil temperature, pH, light exposure, and nutrient levels. The constant seasonal promotion of some members of group I rare species to be members of the dominant (and hence readily identifiable) taxa in soil, together with the seasonal demotion of some of the more abundant taxa in soil, is probably responsible for the observation that seasonal variations often result in significant changes in the phylogenetic affiliations of the most numerous members of the community, leading to statistically detectable differences between seasonal clone libraries from the same soil (12
). However, these seasonal cyclic changes rarely affect the fundamental soil community structure, and in all seasons, the sampled soils will still have their distinctive community composition (16
). We also reason that this fine-tuning function of group I of the rare biosphere is responsible for the fact that within the thousands of soil studies conducted so far (see reference 16
for a review), no two clone libraries have had exactly the same community composition, and exact (100%) sequence matches between the most abundant species and database-deposited sequences (that broadly represent a global repository of more abundant members of soil and other communities) are very rarely encountered. The variations in physical and geochemical characteristics between different soils always select for different species as the most dominant members of the community. Therefore, in all soil surveys reported so far, dominant species identified almost always belong to a previously unencountered strain, species, or genus within well-recognized soil lineages (and hence the tail end of the curve in Fig. never reaches zero).
Within group II of the rare biosphere in the KFS data set (rare bacterial taxa with no close relatives within the dominant species), a fraction belongs to well-described phylogenetic lineages that are prevalent in other ecosystems but are rarely encountered in soil (phyla Chlorobia
candidate phylum BRC-1, Salinibacter
, and Clostridiales
-affiliated clones, and clones belonging to the Sup-05 lineage within the Gammaproteobacteria
) (see Files S4 to S6 in the supplemental material). In addition, we speculate that since members of this group are present in a far less than ideal habitat, the majority will be present in an extremely low number and escaped detection in this study. We suggest that taxa belonging to this fraction of group II of the rare biosphere (together with other species recruited via immigration) respond to more drastic disturbances that could occur in the ecosystem. For example, desertification has been shown to consistently result in an increase in the numbers of organisms from the Deinococcus
), which is otherwise rarely detected in other soil ecosystems. A change in redox potential could regenerate (among other changes) Clostridiales
-affiliated cells (or spores) present in KFS in low abundance. More drastic and sustained disturbances (e.g., the occurrence of a major hydrocarbon spill and the development of anaerobic conditions in soil) initiate more radical promotion, demotion, and recruitment processes, resulting in a completely different community composition.
Finally, a fraction of group II of the rare biosphere belongs to novel bacterial lineages (phyla and subphyla) with no close relatives in the entire global 16S rRNA gene data set currently available (members of candidate phyla KFS1 to KFS5 and novel lineages within different bacterial phyla and candidate phyla) (see Files S3, S5, and S6 in the supplemental material). The ecological role of members of these novel, unique lineages is not yet known. It has been suggested that members of this group fulfill specific crucial, yet unknown functions within soil ecosystems (14
). Alternatively, this fraction of the rare biosphere might represent remnants of microbial evolution that, although currently out-competed in all global ecosystems, possess an exceptional ability to survive and escape extinction.
The comprehensive data set obtained in this study should prove extremely valuable in future research aimed at understanding community dynamics in response to environmental fluctuations, as well as a starting point for elucidating the physiological capabilities and metabolic potential of the numerous novel, as-yet-uncultured lineages in the rare biosphere. We are currently evaluating the effect of simulated global warming on the dynamics of the KFS bacterial community at different levels of phylogenetic resolution, ranging from the phylum level (using quantitative PCR) to the species level (using Phylochip, a comprehensive 16S rRNA gene microarray [4
]). Further, targeted metagenomic approaches such as fluorescent in situ hybridization coupled to fluorescence-activated cell sorting and multiple displacement amplification, or microfluidic separation of individual cells coupled to multiple displacement amplification, are two promising approaches that could help in elucidating the metabolic potential of novel yet-uncultured groups with low abundance in soil and other complex environments (23
This work provides an overall assessment of the phylogenetic diversity and evolutionary relationships between rare and more abundant members of the soil biosphere. The data demonstrate that even in extensively studied habitats, the rare biosphere harbors novel lineages (with no representatives in the database) and unique lineages (that are evolutionary distinct, dissimilar to more abundant members of the community). We anticipate that similar efforts in different soils will greatly expand our understanding of the nature of the soil rare biosphere. Similarly, future efforts examining the rare biosphere in ecosystems currently estimated to have a higher level of yet-unexplored bacterial diversity (see reference 20
for a list of these environments) will greatly expand our understanding of the phylum-level global bacterial diversity. The identification of multiple novel bacterial phyla and subphyla within one of the most intensively studied and sampled habitats on earth clearly indicates that while the probability of identifying novel bacterial groups within numerically abundant members of various microbial communities appears to be nearing saturation, the potential of identifying novel lineages, genes, and genomes, as well as potentially novel metabolic abilities and microbial secondary metabolites within the rare biosphere, is just starting to be realized.