The variability of the 16S rRNA gene in bacterial genomes and its consequences for bacterial community analyses

Abstract

16S ribosomal RNA currently represents the most important target of study in bacterial ecology. Its use for the description of bacterial diversity is, however, limited by the presence of variable copy numbers in bacterial genomes and sequence variation within closely related taxa or within a genome. Here we use the information from sequenced bacterial genomes to explore the variability of 16S rRNA sequences and copy numbers at various taxonomic levels and apply it to estimate bacterial genome and DNA abundances. In total, 7,081 16S rRNA sequences were in silico extracted from 1,690 available bacterial genomes (1-15 per genome). While there are several phyla containing low 16S rRNA copy numbers, in certain taxa, e.g., the Firmicutes and Gammaproteobacteria, the variation is large. Genome sizes are more conserved at all tested taxonomic levels than 16S rRNA copy numbers. Only a minority of bacterial genomes harbors identical 16S rRNA gene copies, and sequence diversity increases with increasing copy numbers. While certain taxa harbor dissimilar 16S rRNA genes, others contain sequences common to multiple species. Sequence identity clusters (often termed operational taxonomic units) thus provide an imperfect representation of bacterial taxa of a certain phylogenetic rank. We have demonstrated that the information on 16S rRNA copy numbers and genome sizes of genome-sequenced bacteria may be used as an estimate for the closest related taxon in an environmental dataset to calculate alternative estimates of the relative abundance of individual bacterial taxa in environmental samples. Using an example from forest soil, this procedure would increase the abundance estimates of Acidobacteria and decrease these of Firmicutes. Using the currently available information, alternative estimates of bacterial community composition may be obtained in this way if the variation of 16S rRNA copy numbers among bacteria is considered.

Figure 1. 16S rRNA within-genome similarity and copy numbers in bacterial genomes.

Upper panel: the similarity of genomes with various copy numbers: the values indicated represent the first, the second and the third quartile. Lower panel: distribution of 16S rRNA copy numbers per genome in 1,690 sequenced bacterial genomes.

Figure 2. 16S rRNA copy numbers in bacterial genomes.

16S rRNA copy numbers in bacterial phyla (classes), selected families (n>6), genera (n>3), and species (n>6). Open triangles indicate minima and maxima, whiskers the 10th and 90th quantile and boxes the 25th and 75th quantile. Median is indicated as a horizontal line and mean as a dot.

Figure 3. Sizes of bacterial genomes.

Sizes of genomes in bacterial phyla (classes), selected families (n>6), genera (n>3), and species (n>6). Open triangles indicate minima and maxima, whiskers the 10th and 90th quantile and boxes the 25th and 75th quantile. Median is indicated as a horizontal line and mean as a dot.

Figure 4. Mean pairwise sequence similarity of 16S rRNA sequences.

Similarity within bacterial genomes, among genomes belonging to the same species and among species belonging to the same genus.

Figure 5. OTU counts and the percentage of OTUs harboring multiple bacterial taxa at various levels of 16S rRNA similarity.

Upper panel: OTU counts at various levels of 16S rRNA similarity; the dashed line and the dotted line indicate the number of bacterial species and genera in the Dataset S2. Bottom panel: the percentage of OTUs harboring 16S rRNA sequences belonging to multiple bacterial species or genera.

Figure 6. Abundance of bacterial 16S rRNA sequences, genomes and DNA in forest litter and soil.

Relative abundance of bacterial 16S rRNA sequences in the amplicon pool from Picea abies litter and soil (Baldrian et al., 2012), and estimates of the relative abundance of bacterial genomes and DNA. The estimates were calculated using the values of 16S rRNA copy numbers and genome sizes of the closest hits to each bacterial OTU.