Role of auxin during intercellular infection of Discaria trinervis by Frankia

Abstract

Nitrogen-fixing nodules induced by Frankia in the actinorhizal plant Discaria trinervis result from a primitive intercellular root invasion pathway that does not involve root hair deformation and infection threads. Here, we analyzed the role of auxin in this intercellular infection pathway at the molecular level and compared it with our previous work in the intracellular infected actinorhizal plant Casuarina glauca. Immunolocalisation experiments showed that auxin accumulated in Frankia-infected cells in both systems. We then characterized the expression of auxin transporters in D. trinervis nodules. No activation of the heterologous CgAUX1 promoter was detected in infected cells in D. trinervis. These results were confirmed with the endogenous D. trinervis gene, DtAUX1. However, DtAUX1 was expressed in the nodule meristem. Consistently, transgenic D. trinervis plants containing the auxin response marker DR5:VENUS showed expression of the reporter gene in the meristem. Immunolocalisation experiments using an antibody against the auxin efflux carrier PIN1, revealed the presence of this transporter in the plasma membrane of infected cells. Finally, we used in silico cellular models to analyse auxin fluxes in D. trinervis nodules. Our results point to the existence of divergent roles of auxin in intercellularly- and intracellularly-infected actinorhizal plants, an ancestral infection pathways leading to root nodule symbioses.

Keywords: AUX1; PIN; actinorrhiza; auxin; endosymbiosis; model; nitrogen fixation; systems biology.

Figure 1

Auxin influx transport inhibitor 1-NOA negatively affects nodulation in D. trinervis. Percentage of plants showing nodules after inoculation with Frankia in the presence of 1-NOA from before inoculation (black line/solid box), from the moment of inoculation (dark gray line/solid squares) or without inhibitor (gray line/empty squares). Nodulation was not observed in non-inoculated control plants (empty diamonds). Letters indicate statistically different groups based on the Tukey-Kramer multiple comparison procedure (P < 0.01).

Figure 2

Immunolocalization of PAA in D. trinervis nodules. (A–C). A strong signal is detected in cells infected by Frankia. No signal is present in the vascular bundle or in non-infected cells. (B–D) No signal is detected in control sections incubated with the secondary antibody alone. Scale bars: A, B: = 50 μm; C, D: 25 μm.

Figure 3

DR5:VENUS:NLS expression pattern during lateral root and nodule development. (A–C) Lateral root: DR5 activation is gradually concentrated at the tip of developing lateral roots. (D) a mature lateral root showing VENUS fluorescence concentrated in the columella. (E–H) Nodule primordia; no concentration gradient of VENUS expression is observed in the tip of the emerging nodule primordium. (G) Longitudinal section of a mature nodule showing DR5 activation in the meristematic region. Scale bars:100 μm.

Figure 4

DtAux1 encodes a putative orthologue of CgAUX1/AtAUX1. (A) Exon-intron structure of DtAUX1 compared to AtAUX1 and CgAUX1. Exons are displayed as black boxes. (B) Maximum likelihood phylogeny obtained with the coding sequences of AUX-LAX genes. DtAUX1 clusters together with CgAUX1 and DgAUX1 and belongs to the AUX/LAX subclass of auxin influx carriers. Numbers indicate the percentage of bootstrap support.

Figure 5

Histochemical localization of β-glucuronidase (GUS) activity in D. trinervis roots expressing a ProDtAUX1:GUS construct. (A) Non-inoculated lateral root; blue staining is detected in the root tip. (B) Cross section of an inoculated root 5 days after inoculation (dai) showing two nodule primordia growing from the pericycle at opposite xylem poles. DtAUX1 expression is observed in the meristematic cells (asterisk). (C) Longitudinal section of a fully developed nodule 21 dai. Cells containing Frankia hyphae are stained in purple. GUS activity is intense in the meristematic region, still visible in the infection zone (Inf), in non-infected cells, and non-detectable in the fixation zone (Fix). (D–E) Magnified images of (C). (D) Meristematic zone. (E) Infection zone: DtAUX1 expression is limited to non-infected cortical cells (arrowheads) surrounding the highly hypertrophied infected cells (IC). (F) Fixation zone: no DtAUX1 activation is detected in the hypertrophic cells filled with Frankia that have already differentiated nitrogen-fixing vesicles (VC). Sections (B–F) were stained with toluidine blue. Scale bars: 100 μm (A–C), 50 μm (D–F).

Figure 6

Immunolocalization of PIN1-like proteins. (A) Longitudinal section of a mature nodule incubated with Anti-PIN1 antibodies and FITC labeled secondary antibodies: a strong signal is detected in the plasma membrane of hypertrophied cortical cells infected by Frankia. (B) Control section incubated with the secondary antibody alone where no signal is detectable. Scale bars: 50 μm.

Figure 7

Modeling auxin fluxes in D. trinervis nodules. (A) Example of digitization of nodule tissue. Images of nodule tissue autofluorescence (top panel) were used as a base for manual digitization of the geometry in a virtual representation of the nodule (Perrine-Walker et al., 2010). Information regarding the infection of the cells was added in the virtual tissue (infected cells are denoted in red, uninfected cells in yellow and meristematic cells in blue). (B–E) Distribution of auxin in a virtual nodule with meristematic cells expressing DtAUX1 and infected cells expressing a PIN1-like auxin efflux carrier. Auxin flux simulations were conducted with either no initial auxin and basal auxin production throughout the tissue (B), basal initial auxin level throughout the tissue (C), or an auxin source located within the Frankia compartment (to simulate auxin production by Frankia, D,E). In all cases, auxin was predicted to accumulate in non-infected cells of D. trinervis actinorhizal nodules rather than in infected and meristematic cells. Exploration of the model parameter space did not reveal a parameter configuration for which the biological auxin accumulation could be reproduced by the model based on DtAUX1 and DtPIN1 localization (data not shown). This suggests that the mechanisms leading to auxin accumulation in those cells do not rely exclusively on D. trinervis PIN1-like and DtAUX1.