Leaf-inhabiting fungal and bacterial endophytes are at their most diverse in tropical rainforest plant hosts, with some influencing host plant fitness as either symbionts or pathogens. Endophyte activity and community composition is thought to depend on competition amongst co-occurring species for resources. Here, we reveal the strength of competitive interactions between endophytes by combining e-DNA metabarcoding to characterize the community with culturing assays to ascertain their potential activity. The endophyte community associated with the understory palm Astrocaryum sciophilum was determined by extracting eDNA from 120 leaflets of eight plants located in a lowland rainforest in French Guiana. This revealed 516 fungal and 606 bacterial Operational Taxonomic Units (OTUs). Co-occurrence analysis of the most abundant OTUs revealed that direct fungal-fungal interactions were overall more negative than bacterial-bacterial interactions. This was confirmed by C-score calculations for the whole endophyte community, revealing significantly greater levels of species segregation for fungi but not bacteria when compared with simulated random communities. Following the culturing of 131 fungal, and 66 bacterial Astrocaryum endophytes collected from the same leaves, Minimal Inhibitory Concentrations of ethyl acetate culture extracts were evaluated against the fungus Trichophyton rubrum and the bacterium Staphylococcus aureus.  Overall, a minority of extracts demonstrated antipathogen activity, with greater antifungal activity expression for both fungi and bacteria when compared with antibacterial activity. In order to explore whether this activity translated into competitive structuring of the community, a BLAST was conducted to match metabarcoding sequences with Sanger derived culture sequences. This demonstrated that cultures displaying antipathogen activity were significantly more likely to co-occur with fungi if they were bacteria, and bacteria if they were fungi. Nevertheless, overall correlation values averaged around zero. These results indicate that for Astrocaryum palms, fungal endophytes are likely to play a greater role in determining colonization success of subsequent endophytes whether these be fungal or bacterial, but that overall endophyte communities do not display strong patterns of community structuring through competition. Endophyte communities are thus likely to reach relative stasis in mature leaves, where the diverse community of fungi and bacteria amongst other factors prevent the subsequent establishment of pathogens.

120 leaflets were sampled from eight Astrocaryum sciophilum palms. They were surface sterilized following a modification of the protocol detailed by Weber et al., (2004). eDNA was extracted from the leaves using a NucleoSpin Plant II (96) kit (Macherey-Nagel, Düren, Germany).  PCRs were performed using two primers; ITS1 nuclear rDNA primers to target fungi (Fwd: ITS5 GGAAGTAAAAGTCGTAACAAGG (Epp et al., 2012) and a modified version of Rev: 5.8S_Fungi CAAGAGATCCGTTGTTGAAAGTK, Taberlet et al. 2018), and 16S rDNA (V5-V6) primers to target bacteria (Fwd: GGATTAGATACCCTGGTAGT and Rev: CACGACACGAGCTGACG (Fliegerova et al., 2014)).

Amplicons were pooled and libraries were prepared with the TruSeqNano PCR free Illumina kit and were sequenced on Paired-end (2X250 bp) in the Illumina Miseq platform (Illumina, San Diego, CA, USA) at the INRA Genotoul-GetPlaGe core facility (Toulouse, France) using the Paired-end MiSeq Reagent Kit V3 (Illumina, San Diego, CA, USA), following the manufacturer’s instructions.

This results in the production of matched .fasta files (R1&R2). These files also contain data from other experiments put on the same run. These are removed using an NGS filter file. These are also provided.

The generated reads were subject to a data-curation pipeline using scripts in R (R Development Core Team, 2013), and the OBITools package (Boyer et al., 2016). Firstly, paired-end read assembly was conducted, before reads were assigned to samples, and were dereplicated. Low-quality sequences (with a pair alignment score of 50 or below) and singletons were removed prior to further analysis. Pairwise distances between the remaining sequences were then calculated using the sumaclust function, which determines pairwise sequence dissimilarity using the raw number of mismatches. Sequence clusters with similarity of 97% or above were defined as operational taxonomic units (OTUs) which are commonly used as proxies for species during molecular analysis, with this 97% threshold used in the majority of fungal studies (Coissac 2012).

Taxonomic assignment of OTUs was conducted using reference databases with the ecotag function, which uses global alignment of sequences against full-length references. For fungi the reference database was obtained by running an in-silico PCR with the ecoPCR program (Ficetola et al., 2010) on Genbank (release 197; ftp://ftp.ncbi.nlm.nih.gov/genbank) using the primer pairs employed here. Taxonomic assignment yielding the highest similarity score was kept, with similarity scores for Fungi ranging from 0.5 to 1 with sequences in the National Center for Biotechnology Information (NCBI) database. For Bacteria, taxonomic assignment was conducted using the SILVA database (Quast et al., 2012), with similarity parameters kept at 97%.

OTUs which could represent contaminants arising during extraction (present within the extraction negative controls) and during PCR (present within PCR negative controls) were removed from the dataset by comparing occurrence in samples. The influence of chimeras formed during PCR and sequencing was reduced by comparing occurrence in samples with that in the blank controls, detecting for non-used tag combinations, as follows. Potential contaminant OTUs were detected by calculating the mean sum of reads in blanks, divided by the mean sum of reads across both blanks and samples. For each OTU, the acceptable level of OTU abundance present in blanks was set to be 5% or less than the total number of reads. Above this threshold, an OTU was more likely to be a contaminant, and as such it was excluded. OTUs occurring only in negative controls were also removed, with their random distribution across plates representative of any cross-contamination of genetic information. This resulted in the removal of 137 bacterial (833 reads) and 140 fungal OTUs (2457 reads). 

All OTUs which were identified as non-fungal following taxonomic assignment were removed from the fungal database. Bacterial sequences which were identified as originating from ribosomes or chloroplasts were removed from the bacterial dataset, since these likely represented plant host genetic material rather than that of bacterial endophytes (representing 2,140,196 reads). We excluded fungal samples with less than 100 reads from subsequent analysis. These low-quality samples likely result from PCR amplification failure. This filter was chosen as a result of the majority of samples averaging a much higher read count. This resulted in the removal of 27 samples, with subsequent analysis for the fungal communities performed on a dataset of 73 leaflets. In comparison, bacterial read counts per sample were much lower, due to the dominance of ribosome and chloroplast data. Once these were removed, all samples containing a minimum of 1 read were kept in the dataset, resulting in the removal of one leaflet, with subsequent analysis for the bacterial communities performed on a dataset of 119 leaves.

Following clustering of sequence reads with a 97% similarity cut-off, and removal of contaminants, 554 OTUs (400,548 reads) were obtained for fungi, and 606 OTUs for bacteria (14,034 reads). Subsequent analysis was performed on these reduced datasets. These are contained within the .csv files provided.

To extract the sequence data from .fasta files in order to determine sequences contained in each sample, the NGS filter files are required. These have also been provided.

The post filtering files (OTU x sample) tables are included for both bacteria (astro_bact_dryad.csv) and fungi (astro_fungi_dryad.csv)