Community proteogenomics reveals insights into the physiology of phyllosphere bacteria

Abstract

Aerial plant surfaces represent the largest biological interface on Earth and provide essential services as sites of carbon dioxide fixation, molecular oxygen release, and primary biomass production. Rather than existing as axenic organisms, plants are colonized by microorganisms that affect both their health and growth. To gain insight into the physiology of phyllosphere bacteria under in situ conditions, we performed a culture-independent analysis of the microbiota associated with leaves of soybean, clover, and Arabidopsis thaliana plants using a metaproteogenomic approach. We found a high consistency of the communities on the 3 different plant species, both with respect to the predominant community members (including the alphaproteobacterial genera Sphingomonas and Methylo bacterium) and with respect to their proteomes. Observed known proteins of Methylobacterium were to a large extent related to the ability of these bacteria to use methanol as a source of carbon and energy. A remarkably high expression of various TonB-dependent receptors was observed for Sphingomonas. Because these outer membrane proteins are involved in transport processes of various carbohydrates, a particularly large substrate utilization pattern for Sphingomonads can be assumed to occur in the phyllosphere. These adaptations at the genus level can be expected to contribute to the success and coexistence of these 2 taxa on plant leaves. We anticipate that our results will form the basis for the identification of unique traits of phyllosphere bacteria, and for uncovering previously unrecorded mechanisms of bacteria-plant and bacteria-bacteria relationships.

Fig. 1.

Experimental strategy applied to characterize the phyllosphere microbiota. All analyses described were conducted from identical pools of cells as starting material. The photograph shows leaves of soybean plants; the electron micrograph shows the surface of an Arabidopsis leaf.

Fig. 2.

Taxonomic composition of the bacterial community in the Soybean 2 sample. A phylogenetic tree calculated from informative marker genes of completely sequenced organisms serves as a reference onto which the estimated coverage of the most abundant clades present in the Soybean 2 sample is projected. Coverage is estimated based on the quantity of marker genes found in the metagenome data and is indicated by red dots (13). A selection of typical representatives of the clades is listed to the right, annotated according to the 16S rRNA gene-sequencing results (Table S2). Archaea contributed only 0.35% to the microbial community of the sample and were identified as members of the mesophilic Crenarchaeota (group 1.1b) by 16S rRNA gene sequencing. The low contribution of eukaryotes (0.58%) to the analyzed phyllosphere community in the soybean sample is in accordance with the design of the microbial harvesting procedure, which included a physical depletion step for eukaryotic cells.

Fig. 3.

Rarefaction analysis of 16S rRNA gene-sequence data to estimate microbial diversity based on a cutoff <97% sequence identity for delineation of operational taxonomic units (OTUs). (A) Comparison of the composite phyllosphere dataset of this study with published samples covering at least 500 sequences each: farm soil (20), termite gut (19), coastal seawater (17), human gut (18). (B) Rarefaction curves of the individual phyllosphere samples and the joint (composite) phyllosphere dataset.

Fig. 4.

Conserved and specifically enriched proteome functions (spectral counting of Pfam domains) per host-plant type. Pfam domains drawn close to a vertex are preferentially and specifically found on that respective plant. Selected examples are highlighted and discussed in the text. Examples of common phyllosphere proteome (i.e., not enriched): 1, PF00120, glutamine synthetase catalytic domain; 2, PF02469, fasciclin domain; 3, PF00593, TonB-dependent receptor; 4, PF07715, TonB-dependent receptor plug domain. Specific proteome enrichments: 5, PF00027, cyclic nucleotide-binding domain; 6, PF03328, HpcH/HpaI aldolase/citrate lyase family; 7, PF00210, ferritin-like domain (e.g., bacterioferritins); 8, PF05067, manganese containing catalase; 9, PF06823, protein of unknown function (DUF1236); 10, PF00669, bacterial flagellin N terminus; 11, PF00128, α-amylase, catalytic domain; 12, PF03413, peptidase propeptide and YPEB domain; 13, PF05443, ROS/MUCR transcriptional regulator protein; 14, PF05532, CsbD-like (general stress response); 15, PF00011, Hsp20/alpha crystallin family; 16, PF02566, OsmC-like (e.g., organic hydroperoxide detoxification); 17, PF00700, bacterial flagellin N terminus; 18, PF01584, CheW-like (chemotaxis signaling); 19, PF00532, periplasmic binding and sugar binding domain; 20, PF00502, phycobilisome protein (light harvesting).