Evaluation of 16S rRNA Gene Primer Pairs for Monitoring Microbial Community Structures Showed High Reproducibility within and Low Comparability between Datasets Generated with Multiple Archaeal and Bacterial Primer Pairs

Abstract

The application of next-generation sequencing technology in microbial community analysis increased our knowledge and understanding of the complexity and diversity of a variety of ecosystems. In contrast to Bacteria, the archaeal domain was often not particularly addressed in the analysis of microbial communities. Consequently, established primers specifically amplifying the archaeal 16S ribosomal gene region are scarce compared to the variety of primers targeting bacterial sequences. In this study, we aimed to validate archaeal primers suitable for high throughput next generation sequencing. Three archaeal 16S primer pairs as well as two bacterial and one general microbial 16S primer pairs were comprehensively tested by in-silico evaluation and performing an experimental analysis of a complex microbial community of a biogas reactor. The results obtained clearly demonstrate that comparability of community profiles established using different primer pairs is difficult. 16S rRNA gene data derived from a shotgun metagenome of the same reactor sample added an additional perspective on the community structure. Furthermore, in-silico evaluation of primers, especially those for amplification of archaeal 16S rRNA gene regions, does not necessarily reflect the results obtained in experimental approaches. In the latter, archaeal primer pair ArchV34 showed the highest similarity to the archaeal community structure compared to observed by the metagenomic approach and thus appears to be the appropriate for analyzing archaeal communities in biogas reactors. However, a disadvantage of this primer pair was its low specificity for the archaeal domain in the experimental application leading to high amounts of bacterial sequences within the dataset. Overall our results indicate a rather limited comparability between community structures investigated and determined using different primer pairs as well as between metagenome and 16S rRNA gene amplicon based community structure analysis. This finding, previously shown for Bacteria, was as well observed for the archaeal domain.

Keywords: 16S microbial community; amplicon sequencing; archaea; metagenome; microbial communities; next-generation sequencing.

Figure 1

Percentage of sequences classified as archaeal in the datasets generated applying the primers targeting the archaeal 16S ribosomal RNA gene region. 100% refers to all sequences in the separate datasets while the percentage of reads classified as Archaea is indicated by the boxes, standard deviation between the n = 5 replicates is indicated by the bars.

Figure 2

Sequence reduction for the compared 16S rRNA gene primer pairs during the different steps of the mothur analysis pipeline. 100% refers to the number of all sequences in each dataset at the beginning of the Mothur sequence annotation pipeline.

Figure 3

Sequence compositions of the communities observed by the different archaeal and prokaryotic primer pairs on the order level. Shown are the five replicates of each primer pair as well as the 16S rRNA gene sequences extracted from the metagenome (16S-MG). Annotation was done using the SILVA v123 database. Taxa with abundance below 1% were grouped as “other”.

Figure 4

Sequence compositions of the communities observed by the different bacterial and prokaryotic primer pairs on the class level. Shown are the five replicates of each primer pair as well as the 16S rRNA gene sequences extracted from the metagenome data pair (16S-MG). Annotation was done using the SILVA v123 database. Taxa with abundance below 1% were grouped as “other”.

Figure 5

Alpha diversity (Shannon numbers equivalent) observed on the 99% OTU level in the tested archaeal (A) and bacterial (B) datasets amplified with the primer pairs (n = 5) after removal singletons and low abundant OTUs (below 0.2%).

Figure 6

Heat map derived from dissimilarity matrix of Bray-Curtis distances between archaeal (A) and bacterial (B) community compositions as observed by the primer pairs and the extracted metagenomic 16S rRNA gene sequences. Similarity between samples is indicated as a color gradient from higher similarity (red) to lower similarity in (white).

Figure 7

RDA based on Hellinger-transformed taxon count data after Bray-Curtis dissimilarity calculation. Clear separation of the different primer pairs can be observed between the different archaeal (A) and bacterial (B) communities observed in the sample. The eight (Archaea) and 10 (Bacteria) taxa contributing most of the variance of the dataset are shown as vectors.