Gallionellaceae in rice root plaque: metabolic roles in iron oxidation, nutrient cycling, and plant interactions

Abstract

In waterlogged soils, iron plaque forms a reactive barrier between the root and soil, collecting phosphate and metals such as arsenic and cadmium. It is well established that iron-reducing bacteria solubilize iron, releasing these associated elements. In contrast, microbial roles in plaque formation have not been clear. Here, we show that there is a substantial population of iron oxidizers in plaque, and furthermore, that these organisms (Sideroxydans and Gallionella) are distinguished by genes for plant colonization and nutrient fixation. Our results suggest that iron-oxidizing and iron-reducing bacteria form and remodel iron plaque, making it a dynamic system that represents both a temporary sink for elements (P, As, Cd, C, etc.) as well as a source. In contrast to abiotic iron oxidation, microbial iron oxidation results in coupled Fe-C-N cycling, as well as microbe-microbe and microbe-plant ecological interactions that need to be considered in soil biogeochemistry, ecosystem dynamics, and crop management.

Keywords: iron oxyhydroxides; iron-oxidizing bacteria; iron-reducing bacteria; rice rhizosphere.

Fig 1

Plaque iron content compared to porewater Fe(II), and mineral composition. (A) Plaque iron content per dry root mass over the growing season plotted for 12 paddies sampled. (B) Porewater Fe(II) over the growing season plotted by and colored by paddy [see legend in (A)]. (C) Plaque Fe mineral composition at harvest, by paddy, in order of decreasing ferrihydrite.

Fig 2

(A) Log 10 median rank relative abundance curve of top 50 plaque organisms. (B) Box and whisker plot showing square-root transformed relative abundance of OTUs with median relative abundance above 0.5% in plaque. Boxes are colored according to known metabolism with orange/yellow representing FeOB, pink/purple FeRB, blue methylotrophs, and green cellulose degraders. See Tables S1 and S2 for abundance data, ranking calculations, and taxonomic classification of OTUs.

Fig 3

Enrichment factors in plaque (A) and rhizosphere (B) of most abundant FeOB and FeRB OTUs over growing season (DPT = days past transplant). Enrichment factor was calculated as relative abundance in plaque or rhizosphere divided by relative abundance in bulk soil, and the median is given for each time point. Boxes are shaded according to enrichment factor, with those below 1 (not enriched) in white with gray text. (No plaque sample available for 88 DPT.) OTUs are ranked by median relative abundance in plaque.

Fig 4

Relative read abundance of top FeOB and FeRB in plaque (with median relative abundance above 0.5%) over time for each biosphere. Abundance in bulk soil plotted with black solid line, rhizosphere plotted in tan short-dashed line, and plaque plotted in orange long-dashed line. Error bars show standard error. Asterisks indicate differential abundance in the plaque and bulk soil where adj. P < 0.1 based on DESeq2 analysis. Bulk 20 DPT n = 12, rhizosphere 20 DPT n = 12, plaque 20 DPT n = 3, bulk 42 DPT n = 11, rhizosphere 42 DPT n = 10, plaque 42 DPT n = 6, bulk 71 DPT n = 11, rhizosphere 71 DPT n = 11, plaque 71 DPT n = 4, bulk 88 DPT n = 12, rhizosphere 88 DPT n = 12, bulk 98 DPT n = 11, rhizosphere 98 DPT n = 11, and plaque 98 DPT n = 4.

Fig 5

Comparison of selected FeRB and FeOB abundance with plaque iron content. Scatter plots show plaque relative abundance of Anaeromyxobacter OTU0 versus plaque Fe (A and B), and of Gallionella/Ferrigenium OTU7 and plaque Fe (C and D). Points are colored by paddy and shapes represent sampling time (DPT). In (A and C), all plaque data are shown and the blue line shows linear regression and the gray shaded region shows 95% CI. In (B and D), temporal trends ar shown for paddies with data from multiple time points.

Fig 6

Maximum likelihood tree of Gallionellaceae FeOB genomes based on 13 concatenated ribosomal protein sequences. Genomes from this study are shown with a star, and isolate genomes are shown by arrows only. While Gallionella 9BH_112 is closely related to the rice paddy isolate F. kumadai AN22, the other genomes in this study are more distant from isolates (in Sideroxydans). Support values based on 1,000 bootstraps. Further information on classification and tree construction, including all genomes in the tree, are available in Hoover et al. (35).

Fig 7

Summary of selected functional genes in Gallionellaceae FeOB genomes. Boxes are colored by number of genes (no. of genes) or gene sets (if >1 gene mentioned) present in each genome. See Table S4 for polysaccharide hydrolysis genes. “N fix” = nitrogen fixation.