**5. Autoregulation of nodulation: Too much of a good thing can be bad**

The presence of the rhizobia together with their Nod factor signal molecule initiates the nod‐ ulation infection process. Root hair penetration is the most common form of rhizobia inva‐ sion. The bacteria attach to emerging root hairs, which begin to deform and eventually encapsulate some of the bacteria, which are continuously dividing (Callaham and Torrey 1981; Turgeon and Bauer 1985). This process happens in as little as 6 – 8 h post-inoculation (Yao and Vincent 1969; Bhuvaneswari et al., 1981; Bhuvaneswari and Solheim 1985; Turgeon and Bauer 1982, 1985). Specialized structures, called infection threads, begin to form and provide a passage way for the bacteria to enter the root (reviewed by Gage 2004). These in‐ fection threads are predominately comprised of plant cell wall components and they permit

A Comprehensive Survey of International Soybean Research - Genetics, Physiology, Agronomy and Nitrogen

As the process of rhizobia infection occurs, cortical cells in the root begin to divide and eventually give rise to the nodule primordium (Calvert et al., 1984; Mathews et al., 1989). The position of the nodule primordium is typically adjacent to the radial cells of the xylem, and away from the phloem. This positioning is thought to be largely dependent on plant hormone levels, namely gradients of the gaseous hormone, ethylene (Heidstra et al., 1997; Gresshoff et al., 2009; Lohar *et al.,* 2009). Additional tissues, including vascular tissues and central nodule tissues that are composed of both invaded and non-invaded cells, also devel‐ op to form the nodule structure (Newcomb et al., 1979; Calvert et al., 1984; Ferguson and

Infection threads initiating in the root hair eventually grow and extend towards the dividing nodule primordium located in the root cortex. Once there, rhizobia located at the tip of the infection threads are released into an infection droplet that separates and is released into the cytoplasm of the host cell. Within the cytoplasm, the rhizobia are encapsulated by a special‐ ized plant-derived membrane, known as the peribacteroid membrane, making what is com‐

Ultimately, the dividing bacteria differentiate into what are known as bacteroids, which are highly specialized and whose main purpose is to fix atmospheric di-nitrogen gas. Inside the mature nodule, the bacteroids use a nitrogenase enzyme complex to fix the di-nitrogen into forms of nitrogen that the plant can use, such as ammonia. The ammonia, which is toxic to the plant, is then quickly converted into compounds such as glutamate or ureides that are non-toxic and are safely transported throughout the plant. Legume nodules provide the ide‐ al setting for this process as they establish a peripheral oxygen barrier, via physical and met‐ abolic barriers, to create a low-oxygen environment that is essential for nitrogenase activity

The nodules formed on the roots of soybean plants are referred to as 'determinate' nodules. They are spherical and lack a persistent meristem, unlike indeterminate nodule structures that can form on other legume species, particularly those from temperate growing regions (Ferguson et al., 2010). The life-span of a soybean nodule is typically a few weeks, after which they senesce and are replaced by new nodule structures developing on the growing root system. Following nodule senescence, the bacteroids can re-differentiate and become

the bacteria to continue proliferating within the host plant.

monly referred to as the symbiosome (Udvardi and Day 1997).

new inoculum for the soil (Gresshoff and Rolfe 1978).

Reid 2005).

Relationships

34

to occur.

A number of genes that are required for proper nodule formation have been elucidated (re‐ viewed in Caetano-Anollés and Gresshoff 1991; Ferguson et al., 2010; Ferguson, 2012). The loss of any of these genes typically results in a reduced, or a complete lack of, nodule devel‐ opment. In addition to these positive regulators of nodule formation, there are also a num‐ ber of external and internal factors that act as negative regulators of nodulation. Mutants unable to synthesise or perceive these factors exhibit increased nodule numbers, often refer‐ red to as a hyper- or super-nodulation phenotype (Figure 1). Many of these factors function in the Autoregulation of Nodulation (AON) pathway, which is a mechanism used by the host plant to regulate its nodule numbers (reviewed in Reid et al., 2011a). Indeed, less than 10% of all rhizobia infection events result in the establishment of a fully functional nodule, largely due to AON. By controlling nodule development in this way, the host plant can bal‐ ance its need to acquire nitrogen against its ability to expend energy establishing and main‐ taining nodules. Supernodulating plants lacking AON are typically developmentallystunted (when inoculated with a compatible rhizobia strain) and yield poorly as a result of this balance being disrupted (Figure 2).

**Figure 1.** Roots of wild-type (WT) and supernodulating mutant (nod++) soybean plants exhibiting mature nodule structures as a result of a symbiotic interaction with *Bradyrhizobium japonicum*.

**Figure 2.** Soybean plants growing in a field in Toowoomba, Queensland, Australia. Mutants unable to form nodules (nod- ) are stunted and pale compared with wild-type (WT) plants due to their inability to establish a symbiotic interac‐ tion with nitrogen-fixing *Bradyrhizobium japonicum*. Supernodulating mutants (nod++) are significantly stunted in stature as a result of investing too much energy into forming nodule structures.

The AON pathway involves long-distance root-shoot signaling initiated during nodule de‐ velopment by the synthesis of a root-derived signal (Gresshoff and Delves, 1986; Delves et al, 1986; Reid et al., 2011a). Recent work has indicated that this signal is likely a CLV3/ESRrelated (CLE) peptide(s) hormone (Okamoto et al., 2009; Mortier et al., 2010; Reid et al., 2011b; Lim et al., 2011). In soybean, these CLE peptides are called Rhizobia Induced CLE1 (RIC1) and RIC2 (Reid et al 2011b; Lim et al., 2011). Grafting and over-expression experi‐ ments have shown that these signals travel to the shoot (Delves et al., 1986; Reid et al., 2011b), likely via the xylem, where they, or a product of their action, are perceived by a LRR receptor kinase, called the Nodulation Autoregulation Receptor Kinase (NARK) in soybean (*e. g.* , Searle et al., 2003). NARK may act in a complex with other receptors, such as CLAVA‐ TA2 and KLAVIER (Miyazawa et al., 2010; Krusell et al., 2011). This perception results in the production of a novel Shoot-Derived Inhibitor (SDI). The SDI signal subsequently travels from the shoot back down to the roots, likely via the phloem, where it acts to inhibit further nodulation events (Reid et al 2011a). It has recently been established in soybean that SDI is small (*<*1 kDa), Nod factor- and NARK-dependent, heat stable, and is unlikely to be an RNA or a protein (Lin et al., 2010, 2011a). Recent work using soybean has also identified a number of novel components that may interact with NARK directly, or that may function down‐ stream of NARK to regulate the AON process. These include genes identified using site-di‐ rected mutagenesis that encode Kinase-Associated Protein Phosphatases, called *GmKAPP1* and *GmKAPP2* (Miyahara et al., 2008), and genes identified using complete transcriptome analyses (RNAseq), such as the putative Ubiquitin Fusion Degradation protein, *GmUFD1a* (Reid et al., 2012).

Additional genes and factors also regulate nodule numbers. Root-specific genes in pea (*PsNOD3*; Postma et al., 1988) and *L. Japonicas* (*LjRDH1*, Ishikawa et al., 2008; *LjTML*, Magori et al., 2009) may have a role in the biosynthesis or translocation of RIC1 and RIC2, or in the perception of SDI. However, the identity of these genes is not yet known. Loss of function of *LjASTRAY*, which encodes a bZIP transcription factor (Nishimura et al., 2002), or *MtEFD*, which encodes an ERF transcription factor (Vernié et al., 2008), also results in increased nod‐ ule numbers. Whether these genes function in the AON pathway remains to be determined.

Ethylene and nitrate are also known to inhibit nodule development (Carroll et al., 1985; Lor‐ teau et al., 2001; Ferguson et al., 2005 a,b; Ferguson et al., 2011). Mutations to ethylene sensi‐ tivity or response genes, such as *LjETR1* and *LjEIN2/MtEIN2*, result in increased nodule formation (Penmetsa et al., 2008; Gresshoff et al., 2009; Lohar et al., 2009). Interestingly, an additional CLE peptide identified in soybean that negatively regulates nodule development, called Nitrogen-Induced CLE1 (NIC1) is highly similar to RIC1 and RIC2, but is induced by nitrate, not rhizobia (Reid et al., 2011b). Both the RICs and NIC1 appear to be perceived by GmNARK, only the nitrate-induced CLE exhibits little-to-no mobility and is perceived in the root, whereas the rhizobia-induced CLE undergoes long distance transport and is perceived in the shoot.
