A Medicago truncatula Autoregulation of Nodulation Mutant Transcriptome Analysis Reveals Disruption of the SUNN Pathway Causes Constitutive Expression Changes in Some Genes, but Overall Response to Rhizobia Resembles Wild-Type, Including Induction of TML1 and TML2

Abstract

Nodule number regulation in legumes is controlled by a feedback loop that integrates nutrient and rhizobia symbiont status signals to regulate nodule development. Signals from the roots are perceived by shoot receptors, including a CLV1-like receptor-like kinase known as SUNN in Medicago truncatula. In the absence of functional SUNN, the autoregulation feedback loop is disrupted, resulting in hypernodulation. To elucidate early autoregulation mechanisms disrupted in SUNN mutants, we searched for genes with altered expression in the loss-of-function sunn-4 mutant and included the rdn1-2 autoregulation mutant for comparison. We identified constitutively altered expression of small groups of genes in sunn-4 roots and in sunn-4 shoots. All genes with verified roles in nodulation that were induced in wild-type roots during the establishment of nodules were also induced in sunn-4, including autoregulation genes TML2 and TML1. Only an isoflavone-7-O-methyltransferase gene was induced in response to rhizobia in wild-type roots but not induced in sunn-4. In shoot tissues of wild-type, eight rhizobia-responsive genes were identified, including a MYB family transcription factor gene that remained at a baseline level in sunn-4; three genes were induced by rhizobia in shoots of sunn-4 but not wild-type. We cataloged the temporal induction profiles of many small secreted peptide (MtSSP) genes in nodulating root tissues, encompassing members of twenty-four peptide families, including the CLE and IRON MAN families. The discovery that expression of TML2 in roots, a key factor in inhibiting nodulation in response to autoregulation signals, is also triggered in sunn-4 in the section of roots analyzed, suggests that the mechanism of TML regulation of nodulation in M. truncatula may be more complex than published models.

Keywords: Medicago truncatula; RDN1; SUNN; TML; autoregulation of nodulation; small secreted peptides.

Figure 1

Diagram of experimental procedure designed to increase signal to noise ratio: (A) Twenty plants per genotype (t = 0 h) were collected and rhizobia added to the growth apparatus immediately after collection of the 0 h samples. Additional samples of 20 plants per genotype were collected at 12, 24, 48, and 72 h after the 0 h samples. Ten plants from each collection were used to determine average root length and 2 cm segments representing the zone of development of the first nodules (red zone) were collected from the remaining 10 plants. At 0 h, this region started 1 cm from the root tip, where the first full-length root hairs were present. At later time points, this region was determined by calculating the average root growth since t = 0 and adding this distance to 1 cm. (B) Cross-sections of roots harvested during the first three time points. Cell division for nodule formation was occasionally observed at 24 h (red arrow). By 48 h, nodule cell division was observed in all plants.

Figure 2

Genes with constitutively altered expression in roots of AON mutants sunn-4 and rdn1-2. Heat map of average fragments per kilobase of transcript per million mapped reads (FPKMs) of genes identified by DeSeq2 with altered expression levels in AON mutants compared to wild-type (A17) that were consistent across all times and conditions (control = no rhizobia; +R = with rhizobia). Each row is independently scaled from minimum to maximum values; underlying data are in Supplemental Dataset 2, sheet F1. Expression of some genes was altered in both mutants, while for others the difference was only found in one mutant. Some genes had higher expression in the mutants and some had lower. The geneID (v4) and annotation are given. For five geneIDs, the annotation in v5 better matched the transcript structure; the v5 geneID is also given for these, with two pairs of geneIDs merged into two larger genes in v5. ** Also found to be similarly different in shoots of AON mutants (see Figure 2).

Figure 3

Genes with constitutively altered expression in shoots of sunn-4 and rdn1-2. Heat map of average fragments per kilobase of transcript per million mapped reads (FPKMs) of genes identified by DESeq2 with altered expression levels in AON mutants compared to wild-type (A17) that were consistent across all times and conditions (control = no rhizobia; +R = with rhizobia). Each row is independently scaled from minimum to maximum values; underlying data are in in Supplemental Dataset 2, sheet F2. Expression of some genes was altered in both mutants, while for others the difference was only found in one mutant. Some genes had higher or lower expression in the mutants. The geneID (v4) and annotation are given. For seven geneIDs, the annotation in v5 better matched the transcript structure; the v5 geneID is also given for these, with two pairs of geneIDs merged into two larger genes in v5. ** Also found to be similarly different in roots of AON mutants (see Figure 2).

Figure 4

Rhizobia-induced expression of nodulation pathway genes in roots of wild-type (A17) and/or AON mutants sunn-4 and rdn1-2. Heat map of average fragments per kilobase of transcript per million mapped reads (FPKMs) of 58 known nodulation genes with patterns of expression that changed with rhizobial inoculation (+R) or with genotype. Each row is independently scaled from minimum to maximum values; underlying data are in Supplemental Dataset 2, Sheet F3. Induction was detected in all three lines for some genes (n = 40) and for only the AON mutants for others (n = 16). Two genes were altered in rdn1-2 only.

Figure 5

AON genes TML2 and TML1 are induced by rhizobia in both wild type and AON mutant root segments. RNA-seq data shows early induction in response to rhizobia for TML2 (A) and later induction for TML1 (B). The fragments per kilobase of transcript per million mapped reads (FPKMs) (black dots) and means (red lines) of three biological replicates are shown for time points 0 through 72 h post-inoculation (hpi) for uninoculated (−R; wild-type A17 and sunn-4) and inoculated (+R; wild-type, sunn-4, and rdn1-2) root segments. qPCR verified the induction of TML2 (C) and TML1 (D) in both A17 and sunn-4. Expression levels were significantly higher in both wild-type and sunn-4 at 16 hpi for TML2 and at 72 hpi for TML1 (*, p < 0.05; Kruskal–Wallis test with Bonferroni correction). The relative expression of three biological replicates (black dots = data points; blue line = means) of these genes is shown. Shading shows the standard error of the mean.

Figure 6

Strong induction of an Isoflavone 7-O-Methyltransferase (Medtr7g014510) in nodulating root segments of wild type plants is not seen in sunn-4plants. (A) fragments per kilobase of transcript per million mapped reads (FPKMs) (black dots) and means (red lines) from three biological replicates for Medtr7g014510 from RNA-seq of wild-type and the AON mutants sunn-4 and rdn1-2 over the first 72 h post-inoculation (hpi) with rhizobia (+R) compared to uninoculated controls (−R). (B) qPCR analysis of Medtr7g014510 in wild-type and sunn-4 showing expression levels relative to 0 h. Blue line is mean. Post-inoculation times points were significantly higher than the 0 h samples, although the extent of induction was 3- to 9-fold less in sunn-4 (*, p < 0.05; Kruskal–Wallis test). Shading is the standard error of the mean.

Figure 7

Gene expression induced by rhizobia in shoots. (A) Heat map of average fragments per kilobase of transcript per million mapped reads (FPKMs) of genes showing increased expression during the first 72 h of nodulation. Each row is independently scaled from minimum to maximum values; underlying data in in Supplemental Dataset 2, Sheet F6. Eight genes were induced in shoots of wild type plants. Seven of these were also induced in shoots of AON mutant plants sunn-4 and rdn1-2. Three additional genes were induced only in sunn-4. Two genes that were also induced in nodulating roots are indicated by “**”. Graphical representation of selected genes is shown in (B,C) with FPKMs (black dots) and their means (red lines) from three biological replicates. Shading is the standard error of the mean. (B) Transcription factor gene Medtr3g111880 was induced by 24 h in shoots of wild-type but not sunn-4, while in roots expression was induced in all three lines. (C) Three genes induced in shoots of sunn-4.