**3. Common signaling pathway in mycorrhizal and rhizobial symbiosis**

In legumes, the establishment and development of AMS and RNS require a set of common symbiosis genes [59, 60], including a conserved SymRK protein from different species and several essential TFs. When LCOs and COs from bacteria and/or fungi are recognized, SymRK is activated to associate with a set of essential proteins like HMGR1 to regulate both AMS and RNS or interact with SymRK-interacting protein 2 (SIP2) which is specifically involved in RNS. Currently, some interacting proteins of SymRK have been confirmed to participate in RNS, but whether they also take part in AMS remains unknown. As critical components of CSSP, SymRK and other receptor complexes could promote the signaling pathway downstream by triggering nuclear calcium spiking and activating CCaMK. CCaMK could interact with and phosphorylate the downstream transcription factor CYCLOPS [24, 61]. Meanwhile, DELLAs bind the CCaMK-CYCLOPS complex to promote the expression of *RAM1* and regulate AMF colonization [59]. On the other hand, DELLAs associate CCaMK-CYCLOPS with NSP1-NSP2 to enhance the expression of *NIN* which regulates RNS positively [62, 63]. Recently, several other important transcription factors, for example, PHRs and SPXs have been identified to involve in AMS under different conditions of Pi [31, 32, 64].

### **3.1 SymRK and its interacting proteins involved in AMS and RNS**

As a typical LRR-RLK, SymRK was identified as an important membranelocalized receptor kinase required for activating a series of physiological responses in the symbiosis between AMF, rhizobial, Frankia bacteria, and their corresponding host plants [65, 66]. Studies have found only SymRK but not Nod factor receptors (NFRs) overexpression triggers the expression of AM-related genes, and the *symrk* mutant fails to form arbuscule, indicating that SymRK plays a crucial role in the exchange of signals and in the decision between the development of AMS or RNS [67]. According to polymorphisms and amino acid length of SymRK in both legumes and non-legumes, SymRK exists in at least three different structural versions. Rice

and tomato, two plant species widely used for AMS study, have a shorter form of SymRK that is sufficient for AMS but cannot fully complement *symrk* mutant in legumes for rhizobial endosymbiosis [68]. Based on the sequence comparison, the extracellular domain of SymRK homologs might play important roles in determining


**Figure 2.**

*Multiple sequence alignment of extracellular domain of SymRK from legumes and non-legumes. The orange box represents the signal peptide, the blue box represents malectin-like domain, the green box represents the conserved Gly-Asp-Pro-Cys (GDPC) sequence, and the purple box represents leucine-rich repeat (LRR) domain.*

their specific functions (**Figure 2**). In legumes, such as *L. japonicus*, *M. truncatula,* and *Glycine max*, great conservation with about 75% identities is found in the extracellular domain of SymRK homologs. Such conservation might pinpoint its specific role in RNS. However, the sequence identities of SymRK homologs between legumes and non-legumes are reduced to about 40−50%. Althoug, SymRK is a central player in CSSP required for both RNS and AMS, the sequence differences at its extracellular region might give direct evidence that responds to either mycorrhizal signals or rhizobial signals. However, what determines the functional difference between legumes and non-legumes is of great interest to be studied in future, and the functional difference might be related to the evolution of AMS and RNS.

Although the *symrk* knock-out mutant completely loses the ability to allow rhizobial infection, the root hairs of *symrk* mutant plants were observed to be exaggerated after rhizobial attachments [21, 69], indicating that SymRK plays an essential role in determining rhizobial infection but not rhizobial attachment. Due to the key role of SymRK in both AMS and RNS, several SymRK-interacting proteins and protein modifications of SymRK have been studied to elucidate the precise function of SymRK in RNS and/or AMS (summarized in **Table 1**). HMGR1 (3-Hydroxy-3- Methylglutaryl Coenzyme A Reductase1) was identified as an interacting protein of MtDMI2 (SymRK homolog in *M. truncatula*), suggesting that mevalonate biosynthesis was involved in mediating the function of SymRK to initiate calcium spiking and symbiotic gene expression in response to both rhizobia and AMF [70]. Transcriptome expression analysis characterized that SYMREM1 (Symbiotic Remorin 1) from *M. truncatula* could interact with MtDMI2 and may act as a scaffold protein for assembly of receptor complexes involved in rhizobial infection [71, 81]. In addition to interacting with DMI2/SymRK, SYMREM1 was also identified to interact with other receptor kinases in legumes, such as MtNFP/LjNFR5 and MtLYK3/LjNFR1 [81]. Thus, the function of SYMREM1 might provide corresponding structural support for the molecular network of NFR1/5-SymRK receptor complexes.

SymRK-interacting protein 1 (SIP1) is a major AT-rich sequence binding (ARID) transcription factor [72]. Two major splicing forms, SIP1 and SIP1L (a longer variant of the SIP1 transcripts) in *L. japonicus* were characterized. Interestingly, SIP1 was found to interact with SymRK, while SIP1L could not. Both SIP1L and SIP1 could specifically bind to the promoter of *NIN* to positively regulate symbiosis. Knockdown of SIP1using RNAi technology in transgenic hairy roots resulted in impairment in the nodule and arbuscular development, suggesting an important role of SIP1 in the CSSP [72, 73]. As a typical mitogen-activated protein kinase (MAPKK), SymRK-interacting protein 2 (SIP2) could specifically interact with SymRK homologs from different legumes [74]. Although SymRK and SIP2 have a strong interaction, the inter-phosphorylation between them was not detected [74, 82]. Additional studies identified that LjMPK6 is a phosphorylation target of SIP2 [83], and SymRK could inhibit the phosphorylation activity of SIP2, therefore SymRK might negatively regulate the SIP2-MPK6 signaling cascade. Like SIP1, SIP2 also functions as a positive regulator in RNS conformed by RNA interference methods. However, SIP2 is not required for AMF colonization, suggesting that SymRK-SIP2 interaction might be specific for rhizobium infection [74]. Interestingly, SymRK was also identified to directly associate with and suppress the kinase activity of LjBAK1, a homolog protein of Arabidopsis AtBAK1 that functions as a coreceptor for multiple MAMP receptors. The function of SymRK in suppressing LjBAK1-mediated immunity is required for rhizobial infection, but whether this suppression favors AMF infection is to be determined [80].


#### **Table 1.**

*Summary of interacting proteins of SymRK. The green dot represents the confirmed function of SymRK in AMS or RNS, while the orange dot represents the function of SymRK in AMS/RNS that was not confirmed or studied.*

SymRK-Interacting E3 ligase (SIE3) is a protein containing CTLH/CRA/RING domains, which mediates the ubiquitination of SymRK, but does not mediate the protein degradation of SymRK in an *in vitro* ubiquitination assay [76]. It is possible that ubiquitinated SymRK may allow sustained signal transduction to downstream host responses. In addition, SIE3 plays a positive role in the SymRK-mediated signaling pathway in RNS. Further study showed that SIE3 can interact with SIP1 and form a homodimer via Cys266 residue [77, 78]. SEVEN IN ABSENTIA4 (SINA4) was identified as another E3 ligase that could interact with SymRK, and coexpression of

#### *Receptor Kinases and Signal Pathway in the Arbuscular Mycorrhizal Symbiosis DOI: http://dx.doi.org/10.5772/intechopen.107261*

SymRK and SINA4 caused SymRK relocalization [75]. On the contrary, SINA4 is able to mediate the degradation of SymRK and plays a negative role in RNS by inhibiting the development of infection threads [75]. It seems like SINA4 might be a key protein working in a negative feedback loop by suppressing excessive symbiotic signal responses. However, whether SIE3 and SINA4 also regulate AMS needs to be further explored.

At present, two E3 ubiquitin ligases of Plant U-Box (PUB) family, PUB1 and PUB2, have also been proved that could interact with DMI2 in *M. truncatula* [79, 84]. PUB1 could interact with two significant symbiotic receptors (LYK3 and DMI2) and modulate the establishment of both AMS and RNS [38]. In addition, PUB1 has been shown to be directly phosphorylated by LYK3 and acts as a negative regulator to inhibit rhizobial infection and nodulation [85]. Although the E3 ligase activity of PUB1 is necessary for negative regulation in RNS and AMS, DMI2 is not a ubiquitination substrate of PUB1 [79], only LYK3 could be degraded by PUB1. It is possible that the involvement of PUB1 in the early symbiosis signal pathway might be a strategy used by plants to actively suppress the excessive infection by rhizobia and AMF. The other PUB-type E3 ubiquitin ligase, MtPUB2, was identified to be a direct regulator of DMI2 in *M. truncatula* and could enhance the ligase activity of MtPUB2 via phosphorylation at Ser421, then the activated MtPUB2 directly ubiquitinates MtDMI2 for degradation in vitro. These studies demonstrated that MtDMI2-MtPUB2 forms a negative feedback loop that displays an important role in nodulation homeostasis [79]. As a key receptor kinase involved in both RNS and AMS, the direct phosphorylation target and protein modification of SymRK need to be further elucidated.

#### **3.2 Transcription factor complexes regulate arbuscule branching**

Nuclear calcium oscillations are essential components of signals leading to AMS and RNS in host plant root cells. A couple of proteins and GRAS domain TFs cooperatively mediate calcium signals and induce symbiotic process. CCaMK (as known as DMI3 in *M. truncatula*), the initiation of calcium spiking perception in nucleus, is required in both AMS and RNS [86]. A gain-of-function of CCaMK leads to exaggerated symbiosis response by forming spontaneous nodules in the absence of rhizobial in *L. japonicus*. Interestingly, the gain-of-function of CCaMK could induce rhizobial and AMF infection, as well as calcium spiking even in the mutant plants, that is, *symrk*, *castor* or *pollux*, suggesting that calcium oscillations mediated by CCaMK are a downstream response of symbiosis pathway [61, 87, 88]. The activation of CCaMK by calcium always needs two steps, direct calcium binding to three EF-hand motifs, and calmodulin (CaM) binding to the kinase domain [89]. Thus, different levels of CCaMK activity are needed in AMS and RNS. As it has been confirmed that Thr265 residue is an essential autophosphorylation site for CCaMK activation [90]. CCaMK autophosphorylated at Thr265 and disrupts the hydrogen bonds network with residues of side chains, leading to nodule organogenesis and AMF infection after two EF-hand motifs bind to calcium. CaM binds the kinase domain and stimulates the activity of CCaMK, which is only required for rhizobial infection but not for AMF colonization. Consistent with that CCaMK with kinase domain only could restore the symbiotic entry of AMF in *ccamk* mutant in *L. japonicus* [88, 91]. In addition, rice OsDMI3 is able to functionally complement the AMS in *M. truncatula*, but partially restore the nodulation phenotype in legumes [92].

Downstream and phosphorylated by CCaMK [24], CYCLOPS/IPD3 is required for rhizobial infection, nodule development [93, 94], AMF infection, and arbuscule formation [24, 95]. Acting as a member in CSSP, OsCYCLOPS could complement the AMS and RNS phenotype of *cyclops-3* mutant in *L. japonicus*, indicating functional conservation of CYCLOPS in legumes and non-legumes [24]. Recently, IPD3 and IPD3-LIKE are identified functional redundancy in AMS in *M. truncatula*, the double mutant of *IPD3* and *IPD3L* can form arbuscular but hyphal entry into epidermis cells is impaired [96, 97]. What's more, the development of AMS is remarkably reduced in *ipd3*/*ipd3l* under a high concentration of Pi treatment [97]. Recent research has demonstrated that other important regulators in Pi uptake, SPXs, and PHRs also take part in regulating AMF colonization under different degrees of phosphate [31, 32, 64]. The formation of arbuscule is reduced in the absence of PHRs. In low Pi conditions, PHRs could bind to P1BS cis-element in promotor of AM-associated genes and induce AM-mediated Pi uptake of host plants; however, SPXs could bind to PHRs, leading to the suppression of transcription of genes downstream in Pi-sufficient conditions. Besides, PHRs could not only regulate genes including *RAM1*, *PT11*, *and WRI5A*, which are required for arbuscule formation or nutrient exchange [31], but they also target genes like *CERK1*, *SymRK*, *NSP2*, etc., that are essential for signal perception of AMF entry in the early stage of symbiosis [64].

The CSSP plays a conserved role in regulating AMS and RNS, and plants discriminate between such processes by CCaMK-CYCLOPS complex promoting different GRAS domain TFs through DELLA proteins. Exogenous GA treatment could inhibit infection threads formation and nodule development, as well as hyphal entry and arbuscule formation in *L. japonicus* [98]. As integrators of GA signaling, DELLAs positively regulate rhizobial infection and arbuscule formation, acting as a bridge to complex CCaMK/DMI3-CYCLOPS/IPD3 with NSP1-NSP2 or RAM1 [25, 98]. In RNS, DELLAs work downstream of DMI3-IPD3, enhancing the intensity of phosphorylation of IPD3 by DMI3, then interact with NSP2 to promote the DMI3-IPD3-NSP1- NSP2 complex formation and induce the expression of symbiotic associated genes like *NIN* [99]. While in AMS, DELLAs interact with CCaMK-CYCLOPS complex, activating RAM1 via DIP1 (DELLA Interacting Protein 1) in rice, and RAD1 (Required for Arbuscule Development) in *M. truncatula* and *L. japonicus* [28, 59, 100, 101]. NSP1 specifically functions in Nod factor signaling, and NSP2 may have a minor role in AMS [35]. By contrast, RAM1 is only required in AMS to support arbuscule branching and has no role in Nod factor signaling [27]. Interestingly, the experiment in *Nicotiana benthamiana* showed that RAM1 may compete NSP2 with NSP1, indicating that RAM1 and NSP1 may be the first step downstream of CSSP to distinct Myc factors and NFs signaling [102]. Meanwhile, other GRAS-type transcription factors, maybe RAD1, are involved in Myc factors signaling, for NSP2 only has a weak function in AMS [27, 101, 102].
