Combining whole-genome shotgun sequencing and rRNA gene amplicon analyses to improve detection of microbe–microbe interaction networks in plant leaves

Abstract

Microorganisms from all domains of life establish associations with plants. Although some harm the plant, others antagonize pathogens or prime the plant immune system, support the acquisition of nutrients, tune plant hormone levels, or perform additional services. Most culture-independent plant microbiome research has focused on amplicon sequencing of the 16S rRNA gene and/or the internal transcribed spacer (ITS) of rRNA genomic loci, which show the relative abundance of the microbes to each other. Here, we describe shotgun sequencing of 275 wild Arabidopsis thaliana leaf microbiomes from southwest Germany, with additional bacterial 16S and eukaryotic ITS1 rRNA amplicon data from 176 of these samples. Shotgun data, which unlike the amplicon data capture the ratio of microbe to plant DNA, enable scaling of microbial read abundances to reflect the microbial load on the host. In a more cost-effective hybrid strategy, we show they also allow a similar scaling of amplicon data to overcome compositionality problems. Our wild plants were dominated by bacterial sequences, with eukaryotes contributing only a minority of reads. Microbial membership showed weak associations with both site of origin and plant genotype, both of which were highly confounded in this dataset. There was large variation among microbiomes, with one extreme comprising samples of low complexity and a high load of microorganisms typical of infected plants, and the other extreme being samples of high complexity and a low microbial load. Critically, considering absolute microbial load led to fundamentally different conclusions about microbiome assembly and the interaction networks among major taxa.

Obtaining deeper insights into microbiome diversity using a simple method to block host and non-targets in amplicon sequencing

“…AmpStop can also be used by researchers who do choose PNAs, as blocking oligos design essentially follows the same procedure 16 . Studies of leaves of the wild plant A. thaliana found that bacterial fraction of extracted DNA are typically very low but range up to about 25% 20 . We found that more blocking cycles (15 vs. 10) were necessary to efficiently block non-target amplification in real leaf samples, but not in mock communities with 5% bacterial DNA.…”

“…Some host amplification can be advantageous because it can be used to quantitatively estimate bacterial loads 18 and having quantitative information has can change inferred ecological relationships between plant microbiota 20 . When we only blocked chloroplast in the 16S V3/V4 region and allowed mitochondrial amplification, the fraction of bacterial reads was proportional to the bacterial DNA load in mock communities.…”

“…Additionally, they can be dropped into almost any library preparation pipeline. Indeed, this method, which we first applied in 2016 11 , has since been used successfully in multiple studies 19,20 but its applicability and accuracy has not yet been broadly tested.…”

Accurate and sensitive detection of microbial eukaryotes from whole metagenome shotgun sequencing

“…wild Arabidopsis thaliana leaves (Regalado et al 2020). From these samples, we detect 37 different eukaryotic species and 25 different eukaryotic families (Table S7) We apply the EukDetect pipeline to public datasets from the human gut microbiome and the plant leaf microbiome, detecting eukaryotes uniquely present in each environment.…”

“…DNA and RNA sequencing from the IHMP IBDMDB project was downloaded from the NCBI SRA under Bioproject PRJNA398089 (Lloyd-Price et al 2019). Sequencing data from the Arabidopsis thaliana leaf microbiome was downloaded from the European Nucleotide Archive under accession PRJEB31530 (Regalado et al 2020).…”