The soil microbial food web revisited: Predatory myxobacteria as keystone taxa?

Abstract

Trophic interactions are crucial for carbon cycling in food webs. Traditionally, eukaryotic micropredators are considered the major micropredators of bacteria in soils, although bacteria like myxobacteria and Bdellovibrio are also known bacterivores. Until recently, it was impossible to assess the abundance of prokaryotes and eukaryotes in soil food webs simultaneously. Using metatranscriptomic three-domain community profiling we identified pro- and eukaryotic micropredators in 11 European mineral and organic soils from different climes. Myxobacteria comprised 1.5-9.7% of all obtained SSU rRNA transcripts and more than 60% of all identified potential bacterivores in most soils. The name-giving and well-characterized predatory bacteria affiliated with the Myxococcaceae were barely present, while Haliangiaceae and Polyangiaceae dominated. In predation assays, representatives of the latter showed prey spectra as broad as the Myxococcaceae. 18S rRNA transcripts from eukaryotic micropredators, like amoeba and nematodes, were generally less abundant than myxobacterial 16S rRNA transcripts, especially in mineral soils. Although SSU rRNA does not directly reflect organismic abundance, our findings indicate that myxobacteria could be keystone taxa in the soil microbial food web, with potential impact on prokaryotic community composition. Further, they suggest an overlooked, yet ecologically relevant food web module, independent of eukaryotic micropredators and subject to separate environmental and evolutionary pressures.

Fig. 1. Screening of pro- and eukaryotic micropredators in soils.

a Proportion of major identified micropredator SSU rRNA normalized to total SSU rRNA. b Proportion of major identified micropredator SSU rRNA normalized to total micropredator SSU rRNA. Error bars show standard deviation of replicates. For sites see Table 1.

Fig. 2. Screening of Myxococcales and predatory protist taxa.

a Proportion of identified myxobacteria families SSU rRNA normalized to overall Myxococcales SSU rRNA. b Proportion of predatory protist SSU rRNA normalized to total predatory protist SSU rRNA. For sites see Table 1.

Fig. 3. Prey spectrum of myxobacteria.

a Overview of lysis in predation assays. + Visible lysis ≥ 1 mm in diameter. - No clear lysis after 14 days. b Lysis of C. basilensis colony by M. fulvus (7 days) (c) Lysis of M. luteus colony by M. fulvus (7 days) (d) Lysis of E. coli colony by H. ochraceum (5 days) (e) Lysis of M. luteus colony by H. ochraceum (7 days) (f) Lysis of P. putida colony by C. robustus (4 days) (g) Lysis of D. acidovorans colony by K. flava (4 days).

Fig. 4. Comparison of organic and mineral soils.

Predator:prey ratio of major identified micropredator SSU rRNA normalized to SSU rRNA of prey bacteria. Average in organic soils (excluding MO samples) on the left; average in mineral soils on the right. Area of boxes is proportional to abundance of SSU rRNA. Numbers show predator:prey ratios of micropredator SSU rRNA. Lysobacter data are not shown due to low abundances. Groups were tested for differentially expressed sequences (*p < 0.05; **p < 0.01).

Fig. 5. Simplified soil microbial food web.

Left: Traditional microbial food web with separate roles of prokaryotic and eukaryotic organisms. Right: Microbial food web containing a separate module, independent of eukaryotic organisms. Straight arrows: links between trophic levels. Bent arrows: release of nutrients.