Rhizopine biosensors for plant-dependent control of bacterial gene expression

Abstract

Engineering signalling between plants and microbes could be exploited to establish host-specificity between plant-growth-promoting bacteria and target crops in the environment. We previously engineered rhizopine-signalling circuitry facilitating exclusive signalling between rhizopine-producing (RhiP) plants and model bacterial strains. Here, we conduct an in-depth analysis of rhizopine-inducible expression in bacteria. We characterize two rhizopine-inducible promoters and explore the bacterial host-range of rhizopine biosensor plasmids. By tuning the expression of rhizopine uptake genes, we also construct a new biosensor plasmid pSIR05 that has minimal impact on host cell growth in vitro and exhibits markedly improved stability of expression in situ on RhiP barley roots compared to the previously described biosensor plasmid pSIR02. We demonstrate that a sub-population of Azorhizobium caulinodans cells carrying pSIR05 can sense rhizopine and activate gene expression when colonizing RhiP barley roots. However, these bacteria were mildly defective for colonization of RhiP barley roots compared to the wild-type parent strain. This work provides advancement towards establishing more robust plant-dependent control of bacterial gene expression and highlights the key challenges remaining to achieve this goal.

FIGURE 1

Analysis of rhizopine‐inducible promoters. (A) The rhizopine receiver plasmid pSIR04 and compatible promoter GFP fusion plasmids were each mobilized into Azorhizobium caulinodans AcLP for assessment of rhizopine‐inducible expression by flow‐cytometry after 24‐h induction. Error bars represent one SEM (n = 3). Independent two‐tailed student's t‐tests were used to compare means. Not significant (ns p > 0.05), **p < 0.01. (B) Alignment of three unique mocB promoter sequences from Sinorhizobium meliloti (Sm) L5‐30, Rhizobium pisi (Rp) CECT 4113, and phyllobacterium (P) sp. Tri‐38. The transcriptional start site (+1) was identified by 5′‐RACE (see Figure S2) and a promoter reduction experiment was performed to determine the minimal functional promoter region (see Figure S3). (C) Alignment of three unique mocD promoter sequences from Sm L5‐30, Rp CECT 4113, and P. sp. Tri‐38. (D) MEME output for the identification of a loosely conserved 11‐bp direct repeat motif separated by a 13‐bp spacer in the three unique mocB and mocD promoters

FIGURE 2

Functionality of rhizopine biosensor plasmids is restricted to rhizobial alpha‐proteobacteria. (A) SI‐inducible GFP expression was tested in a range of bacteria (Sm, Sinorhizobium meliloti; Rlv, Rhizobium leguminosarum bv. viciae; Ac, Azorhizobium caulinodans; Ml, Mesorhizobium loti; Ab, Azospirillum brasilense; Bv, Burkholderia vietnamienses; Ao, Azoarcus olearius; Hs, Herbaspirillum seropedicae; Ec, E. coli; Pf, Pseudomonas stuzeri) carrying the rhizopine biosensor plasmid pSIR02. Fold‐induction was calculated as relative fluorescence units (GFP fluorescence intensity/OD λ600nm) for bacteria grown with 10 μM SI supplemented into the growth media divided by RFU for non‐induced cells. (B) SI transport assay for E. coli DH5a carrying pSIR02. (C) Test of rhizopine‐inducible GFP expression in Sm CL150 and extra‐cytoplasmic sigma factor mutants carrying pSIR02. (D) Assessment mocR promoter functionality in E. coli DH5a using transcriptional reporter plasmids. Pempty is a randomly generated 30‐bp nucleotide sequence that serves as a negative control. (E) A constitutively expressed copy of mocR was introduced into E. coli DH5a carrying pSIR02 to test whether this would permit functionality of the rhizopine‐inducible GFP expression. Error bars represent one SEM (n = 3). Independent two‐tailed students t‐tests were used to compare means, except for in panel (a) where fold‐induction for each strain was subject to Bonferroni adjusted one sample t‐tests under the null hypothesis that μ = 1. Not significant (ns p > 0.05), **p < 0.01, ***p < 0.001. Bacteria in reporter assays were induced for 24‐h prior to measurement.

FIGURE 3

Tuning intBC expression for improved biosensor functionality. (A) Growth curves for Azorhizobium caulinodans AcLP and strains carrying rhizopine biosensor plasmids in UMS media supplemented with 10 mM NH₃Cl as a sole source of nitrogen and 20 mM Na succinate as a sole carbon source. Growth statistics including mean generation time (MGT) was calculated using the R package Growthcurver. (B) a series of promoters described in Sinorhizobium meliloti (MacLellan et al., 2006) were transcriptionally fused to GFP and tested for strength of GFP expression using fluorescence assays. RFU is defined as relative fluorescence units. Pempty is a randomly generated 30‐bp nucleotide sequence that serves as a negative control. (C) Dose response curve for GFP induction in AcLP carrying pSIR02 or pSIR05. Bacteria were induced for 24‐h prior to measurement. Error bars represent one SEM (n = 3). Independent two‐tailed students t‐tests were used to compare means with Pempty serving as a reference group for comparisons in panel (B). Not significant (ns p > 0.05), **p < 0.01, ***p < 0.001

FIGURE 4

In situ rhizopine‐dependent expression on RhiP barley roots. (A) Diagram of AcCherry, a derivative of Azorhizobium caulinodans AcLP harbouring a constitutively expressed mCherry gene integrated in single‐copy into the chromosome. (B) Confocal microscopy images of single AcCherry cells carrying the rhizopine biosensor plasmids pSIR02 or pSIR05 colonizing the root surface of RhiP barley. Approximately 2 × 10⁸ bacteria were inoculated onto RhiP barley, which was grown 7‐dpi prior to harvest. (C) Flow‐cytometry quantification of rhizopine‐inducible GFP expression in populations of AcCherry cells carrying pSIR02 or pSIR05 colonizing the root associated (RA, defined as root surface and endosphere) and rhizosphere (RS, defined as sand surrounding the roots) fractions of wild‐type (WT) and RhiP barley roots (n = 3, shades of orange represent each replicate). Only bacteria exhibiting mCherry fluorescence >5000 a.u. (mCherry+) were considered for analysis. The upper 99th percentile of GFP fluorescence intensity in mCherry+ bacteria isolated from wild‐type barley was used to define a threshold for GFP+ cells. (D) Median GFP fluorescence intensity from the GFP+ populations of AcCherry carrying pSIR02 or pSIR05. Error bars represent one SEM. Independent two‐tailed student's t‐tests were used to compare means. Not significant (ns p > 0.05), *p < 0.05

FIGURE 5

Colonization effectiveness and stability of expression on RhiP barley roots. (A) Approximately 10⁵ bacteria for each strain were independently inoculated onto RhiP barley, which was grown 7‐dpi prior to harvest. Bacteria isolated from the root associated (RA, defined as root surface and endosphere) and rhizosphere (RS, defined as sand surrounding the roots) fractions exhibiting mCherry fluorescence >5000 a.u. (mCherry+) were counted by flow‐cytometry. Colonization effectiveness was measured as the percentage of AcCherry cells carrying a biosensor plasmid compared to the wild‐type control, AcCherry devoid of plasmids. (B) Analysis of silencing analysis for AcCherry carrying pSIR02 and pSIR05. Cell suspensions recovered from the colonization assays were seeded 1:10 in UMS media and induced for 24‐h with 10 μm SI prior to analysis by flow‐cytometry. The upper 99th percentile of GFP fluorescence intensity in a non‐induced free‐living culture (Figure 3C) was used to define a threshold for GFP+ cells. Error bars represent one SEM. Independent two‐tailed student's t‐tests were used to compare means (n = 7). Not significant (ns p > 0.05), *p < 0.05