Dissecting hormonal pathways in nitrogen-fixing rhizobium symbioses

Nitrogen is a key element for plant growth. To meet nitrogen demands, some plants establish an endosymbiotic relationship with nitrogen-fixing rhizobium or <em>Frankia </em>bacteria. This involves formation of specialized root lateral organs, named nodules. These nodules are colonized intracellularly, which creates optimal physiological conditions for the fixation of atmospheric nitrogen by the microbial symbiont. Nitrogen-fixing endosymbioses are found among four related taxonomic orders that together form the nitrogen-fixation clade. Within this clade, nodulation is restricted to ten separate lineages that are scattered among mostly non-nodulating plant species. This limited distribution suggests that genetic adaptations that allowed nodulation to evolve occurred in a common ancestor. A major aim of the scientific community is to unravel the evolutionary trajectory towards a nitrogen-fixing nodule symbiosis. The formation of nitrogen-fixing root nodules is best studied in legumes (Fabaceae, order Fabales); especially in <em>Lotus japonicus</em> and <em>Medicago truncatula</em>, two species that serve as model. Legumes and <em>Parasponia</em> (Cannabaceae, order Rosales) represent the only two lineages that can form nodules with rhizobium bacteria. Studies on <em>M. truncatula, L. japonicus </em>and <em>Parasponia</em> showed, amongst others, that nodule formation is initiated upon perception of rhizobial secreted lipo-chitooligosaccharide (LCO) signals. These signals are structurally related to the symbiotic signals produced by arbuscular mycorrhizal fungi. These obligate biotropic fungi colonize roots of most land plants and form dense hyphal structures inside existing root cortical cells. Rhizobial and mycorrhizal LCOs are perceived by LysM-domain-containing receptor-like kinases. These activate a signaling pathway that is largely shared between both symbioses. Symbiotic LCO receptors are closely related to chitin innate immune receptors, and some receptors even function in symbiotic as well as innate immune signaling. In <strong>Chapter 2</strong>, I review the intertwining of symbiotic LCO perception and chitin-triggered immunity. Furthermore, I discuss how rhizobia and mycorrhiza might employ LCO signaling to modulate plant immunity. In a perspective, I speculate on a role for plant hormones in immune modulation, besides an important function in nodule organogenesis. In legumes, nodule organogenesis requires activation of cytokinin signaling. Mutants in the orthologous cytokinin receptor genes <em>MtCRE1 </em>and <em>LjLHK1 </em>in <em>M. truncatula </em>and <em>L. japonicus</em>, respectively, are severely affected in nodule formation. However, how cytokinin signaling is activated in response to rhizobium LCO perception and to what extent this contributes to rhizobium LCO-induced signaling remained elusive. In <strong>Chapter 3</strong>, I show that the majority of transcriptional changes induced in wild-type <em>M. truncatula</em>, upon application of rhizobium LCOs, are dependent on activation of MtCRE1-mediated cytokinin signaling. Among the genes induced in wild type are several involved in cytokinin biosynthesis. Consistently, cytokinin measurements indicate that cytokinins rapidly accumulate in <em>M. truncatula </em>roots upon treatment with rhizobium LCOs. This includes the bioactive cytokinins isopentenyl adenine and <em>trans</em>-zeatin. Therefore, I argue that cytokinin accumulation represents a key step in the pathway leading to legume root nodule organogenesis. Strigolactones are plant hormones of which biosynthesis is increased in response to nutrient limitation. In rice (<em>Oryza sativa</em>) and <em>M. truncatula</em>, this response requires the GRAS-type transcriptional regulators NSP1 and NSP2. Both proteins regulate expression of <em>DWARF27</em> (<em>D27</em>), which encodes an enzyme that performs the first committed step in strigolactone biosynthesis. NSP1 and NSP2 are also essential components of the signaling cascade that controls legume root nodule formation. In line with this, I questioned whether the NSP1-NSP2-D27 regulatory module functions in rhizobium symbiosis. In <strong>Chapter 4, </strong>I show that in <em>M. truncatula</em> <em>MtD27 </em>expression is induced within hours after treatment with rhizobium LCOs. Spatiotemporal expression studies revealed that <em>MtD27 </em>is expressed in the dividing cells of the nodule primordium. At later stages, its expression becomes confined to the meristem and distal infection zone of the mature nodule. Analysis of the expression pattern of <em>MtCCD7 </em>and <em>MtCCD8</em>, two additional strigolactone biosynthesis genes, showed that these genes are co-expressed with <em>MtD27 </em>in nodule primordia and mature nodules. Additionally, I show that symbiotic expression of <em>MtD27 </em>requires MtNSP1 and MtNSP2. This suggests that the NSP1-NSP2-D27 regulatory module is co-opted in rhizobium symbiosis. Comparative studies between legumes and nodulating non-legumes could identify shared genetic networks required for nodule formation. We recently adopted <em>Parasponia</em>, the only non-legume lineage able to engage in rhizobium symbiosis. However, to perform functional studies, powerful reverse genetic tools for <em>Parasponia </em>are essential. In <strong>Chapter 5</strong>, I describe the development of a fast and efficient protocol for CRISPR/Cas9-mediated mutagenesis in <em>Agrobacterium tumefaciens</em>-transformed <em>Parasponia andersonii</em> plants. Using this protocol, stable mutants can be obtained in a period of three months. These mutants can be effectively propagated <em>in vitro</em>, which allows phenotypic evaluation already in the T0 generation. As such, phenotypes can be obtained within six months after transformation. As proof-of-principle, we mutated <em>PanHK4</em>, <em>PanEIN2</em>, <em>PanNSP1 </em>and <em>PanNSP2</em>. These genes are putatively involved in cytokinin and ethylene signaling and regulation of strigolactone biosynthesis, respectively. Additionally, orthologues of these genes perform essential symbiotic functions in legumes. <em>Panhk4 </em>and <em>Panein2 </em>knockout mutants display developmental phenotypes associated with reduced cytokinin and ethylene signaling. Analysis of <em>Pannsp1 </em>and <em>Pannsp2 </em>mutants revealed a conserved role for NSP1 and NSP2 in regulation of the strigolactone biosynthesis genes <em>D27 </em>and <em>MAX1 </em>and root nodule organogenesis. In contrast, symbiotic mutant phenotypes of <em>Panhk4 </em>and <em>Panein2 </em>mutants are different from their legume counterparts. This illustrates the value of <em>Parasponia</em> as comparative model - besides legumes - to study the genetics underlying rhizobium symbiosis. Phylogenetic reconstruction showed that the <em>Parasponia </em>lineage is embedded in the non-nodulating <em>Trema </em>genus. This close relationship suggests that <em>Parasponia</em> and <em>Trema </em>only recently diverged in nodulation ability. In <strong>Chapter 6</strong>, I exploited this close relationship to question whether the nodulation trait is associated with gene expression differentiation. To this end, I sequenced root transcriptomes of two <em>Parasponia </em>and three <em>Trema </em>species. Principal component analysis separated all <em>Parasponia </em>samples from those of <em>Trema</em> along the first principal component. This component explains more than half of the observed variance, indicating that the root transcriptomes of two <em>Parasponia</em> species are distinct from that of the <em>Trema</em> sister species <em>T. levigata</em>, as well as the outgroup species <em>T. orientalis</em> and <em>T. tomentosa</em>. To determine, whether the transcriptional differences between <em>Parasponia</em> and <em>Trema</em> are relevant in a symbiotic context, I compared the list of differentially expressed genes to a list of genes that show nodule-enhanced expression in <em>P. andersonii</em>. This revealed significant enrichment of nodule-enhanced genes among genes that lower expressed in roots of <em>Parasponia</em> compared to <em>Trema</em>. Among the genes differentially expressed between <em>Parasponia </em>and <em>Trema </em>roots are several involved in mycorrhizal symbiosis as well as jasmonic acid biosynthesis. Measurements of hormone concentrations, showed that <em>Parasponia </em>and <em>Trema </em>roots harbor a difference in jasmonic acid/salicylic acid balance. However, mutants in jasmonic acid biosynthesis are unaffected in nodule development. Therefore, it remains a challenge to determine whether the difference in root transcriptomes between <em>Parasponia </em>and <em>Trema </em>are relevant in a symbiotic context. In <strong>Chapter 7</strong>, I review hormone function in nitrogen-fixing nodule symbioses in legumes, <em>Parasponia</em> and actinorhizal species. In this chapter, I question whether different nodulating lineages recruited the same hormonal networks to function in nodule formation. Additionally, I discuss whether nodulating species harbor genetic adaptations in hormonal pathways that correlate with nodulation capacity.