Application of stable-isotope labelling techniques for the detection of active diazotrophs

Abstract

Investigating active participants in the fixation of dinitrogen gas is vital as N is often a limiting factor for primary production. Biological nitrogen fixation is performed by a diverse guild of bacteria and archaea (diazotrophs), which can be free-living or symbionts. Free-living diazotrophs are widely distributed in the environment, yet our knowledge about their identity and ecophysiology is still limited. A major challenge in investigating this guild is inferring activity from genetic data as this process is highly regulated. To address this challenge, we evaluated and improved several ¹⁵ N-based methods for detecting N₂ fixation activity (with a focus on soil samples) and studying active diazotrophs. We compared the acetylene reduction assay and the ¹⁵ N₂ tracer method and demonstrated that the latter is more sensitive in samples with low activity. Additionally, tracing ¹⁵ N into microbial RNA provides much higher sensitivity compared to bulk soil analysis. Active soil diazotrophs were identified with a ¹⁵ N-RNA-SIP approach optimized for environmental samples and benchmarked to ¹⁵ N-DNA-SIP. Lastly, we investigated the feasibility of using SIP-Raman microspectroscopy for detecting ¹⁵ N-labelled cells. Taken together, these tools allow identifying and investigating active free-living diazotrophs in a highly sensitive manner in diverse environments, from bulk to the single-cell level.

Figure 1

Comparison of ARA and ¹⁵N₂ tracer assay. ARA performed on a forest soil sample incubated with either fructose or artificial root exudates (RE) (panels A) and a cyanobacterial culture (A. torulosa) (panels B). Legend for panels A and B is located in panel A. ‘Soil, C₂H₄ contr.’ indicates an ethylene consumption control performed on the forest soil sample as described in Experimental procedure. ‘Soil, C₂H₂’ refers to soil samples incubated with acetylene in the ARA. Graphs depict average ethylene concentration (ppm) ± standard error. Panels C depict the incorporation of ¹⁵N from ¹⁵N₂ gas (average δ¹⁵N ± standard error) in the same soil samples used for ARA. ‘Contr’. indicates incubation with lab air.

Figure 2

Average ± standard error of ¹⁵N enrichment in bulk soil and different microbial biomolecules measured by EA‐IRMS. Forest soil samples were incubated for up to 14 days in an atmosphere containing ¹⁵N₂, or in lab air as a control.

Figure 3

Proportion of bacterial 16S rRNA copies recovered from each of the SIP gradient fractions in ¹⁵N‐RNA or ‐DNA SIP experiments of soil samples: RNA‐SIP (CsTFA density gradients; panel A); primary DNA‐SIP (CsCl density gradients; panel B); and secondary DNA‐SIP (CsCl density gradients with bis‐benzimide; panel C). Values on the Y‐axis represent the proportion of the 16S rRNA copies out of the total number of copies of the entire gradient. Blue‐ and red‐shaded areas indicate the fractions where unlabelled and labelled template is expected to concentrate, respectively. The legend for all panels is located in panel A. “¹⁵N” in the legend refers to samples incubated in artificial atmosphere containing ¹⁵N₂ gas, “¹⁴N” refers to control incubations in lab air.

Figure 4

Principal coordinates analysis (PCoA) depicting the Morisita–Horn dissimilarities among microbial communities in the different SIP fractions from RNA (panel A) and the secondary DNA gradient (panel B). Samples from each SIP gradient are connected by a line to aid interpretation. Large circles depict fractions where the majority of ¹⁵N‐enriched RNA (ranging from 1.785 to 1.820 g mL⁻¹) and DNA (ranging from 1.665 to 1.695 g mL⁻¹) were expected to be found. Legend for both panels can be found in panel A.

Figure 5

Fold change (log₂) of OTUs between labelled and unlabelled fractions in each SIP‐gradient: RNA‐SIP gradients from ¹⁵N₂ incubated samples (panel A); RNA‐SIP gradients from ¹⁴N₂ incubated samples (panel B); DNA‐SIP gradients from ¹⁵N₂ incubated samples (panel C); DNA‐SIP gradients from ¹⁴N₂ incubated samples (panel D). Each dot represents an OTU, which passed sparsity filtering (see Experimental procedure). The X‐axis shows the mean normalised counts of the OTUs across all samples for each displayed phylum, while the Y‐axis is the mean log₂ fold change across all gradients. Red dots denote OTUs enriched in the labelled fractions compared to the control, with log₂ fold change of >0.25 and an adjusted P‐value of <0.1.

Figure 6

Normalised OTU abundances of enriched OTUs along the density gradients from RNA‐SIP (panel A) and DNA‐SIP (panel B). Shaded areas indicate where labelled template is expected to concentrate.

Figure 7

Differences in Raman spectra of unlabelled (blue lines) and ¹⁵N‐labelled (red lines) bacterial cells. Means (bold lines) and standard error (light bands) are depicted (n = ∼ 30). Numbers in boxes indicate peaks with significant shifts in response to ¹⁵N incorporation in E. coli. AU, arbitrary unit.