**3. Plant genetic control over development and functioning of mutualistic symbioses of legumes**

#### **3.1. Legume genes involved in development of RN and AM symbioses**

The complex developmental processes which lead to the formation of intercellular and subcellular symbiotic compartments in RN and AM symbioses are controlled by both macro- and microsymbiont. Genetic systems of the symbionts are highly integrated, because some genes and gene products of one partner can switch certain genetic programs in another partner, still the development and function of the symbioses is reliant to the greatest extent on the plant. Developmental genetics of RNS is now well described because both plants and nodule bacteria can be subjected to genetic analysis during nitrogen-fixing nodule formation and functioning. There has been less investigation of AM systems. Mainly this is due to the difficulties encoun‐ tered in culturing AM-fungi, caused by their obligate symbiotic lifestyle and impossibility of using selective media. Additionally, genetic analysis of AM fungi is more complex because of their heterokaryotic nature and lack of sexual process [34, 145].

The plant genes involved in development of RN and AM symbioses may be divided into two groups, according to approach which was used for the gene identification. The first group, *Sym* genes [146], had been identified with the use of formal genetic analysis (started from selection of plant mutants defective in nodule development). The other group of genes comprises nodulins (from *nodule*) [147], mycorrhizins (from *mycorrhiza*) [148, 149], and symbiosins (from *symbiosis*) [149, 150]. These genes were identified by mo‐ lecular genetic methods, through identification of proteins and/or RNAs synthesized *de novo* in root nodules (nodulins) or roots colonized by AM fungi (mycorrhizins). The genes whose expression is induced during the development of both endosymbioses, RNS and AM are called symbiosins [150].

Genes of these groups are suspected to play different roles in the processes which may be referred collectively as "management of microsymbionts" inside plant roots. Specifically, the products of some nodulin genes represent the structural elements of newly constructed temporary compartments developed during symbiosis [151]. The other nodulin genes may play essential roles in modulating the hormonal status of the developing nodules [152, 153]. Resently, *in silico* and microarray-based transcriptome profiling approaches have allowed identification of nodulins, mycorrhizins and symbiosins, which are being activated in response to an AM fungal signal, or by either rhizobial or AM fungal stimulus, respectively [149]. Several hundred genes were found to be activated at different stages of either symbiosis, with almost 100 genes being co-induced during nodulation and in AM formation. These co-induced genes representing the common evolutionary bases of AM and nodular symbioses can be associated with those cellular functions which are required for symbiotic efficiency, such as the facilitation of nutrient transport across the perisymbiotic membranes that surround the endosymbiotic bacteroids in root nodules and the arbuscules in AM roots [150]. However, it should be remembered that although most of the nodulins/mycorrhizins/symbiosins were already cloned and sequenced, functions for many of them have been identified only preliminary using the sequence data of the encoding genes and location of the gene products in the symbiotic compartments.

Still, most of nodulins, mycorrhizins and symbiosins seem to play a subordinate role in the regulatory scheme of symbiosis, nevertheless being indispensable for its functionality and stability. In turn, the major, regulatory role in realization of symbiotic programmes is to be assigned to *Sym* genes. These genes, in contrast to nodulin genes, are usually not expressed outside symbiotic structures and there are many examples of the high functional and sequence homologies between them in different legumes. First genes of this group had been identified using the spontaneous mutants from natural legume populations [154] and afterwards using the experimentally induced mutants defective in nodulation or nitrogen fixation (Nod– and Fix– phenotypes) [155]. Afterwards, it was demonstrated that mutations in some of these genes also affect the ability of plant to form AM [148]. The presence of such common genes necessary for both AM and RN symbioses development suggested that both endosymbioses were more closely related than it was suspected before. The cloning and sequencing of the common symbiotic genes helped to understand that AM and RN symbioses share the overlapping signaling pathways, which probably were established during evolution the AM symbiosis and was recruited afterwards into the RN symbiosis development [35].

The large sizes of genomes of crop legumes (e.g. soybean *Glycine max* (L.) Merr. and pea) in which the formal genetics of symbioses was initially developed, as well as low capability for genetic transformation, complicate greatly the cloning of symbiotic genes, analysis of their primary structures and the gene manipulations. Therefore, in early 1990s the new legume species, *Lotus japonicus*[156] and *Medicago truncatula* [157, 158] have been introduced in studies as model plants. These species are characterized by small genomes (470 – 500 Mb; [159]) and can easily be genetically transformed [158, 160-162]. In addition, the short lifecycle and high seed productivity made them attractive and convenient model objects for studying molecular bases of RN and AM symbioses.

Genetic analysis in model legumes as well as in crop legumes was started with experimental mutagenesis. Large-scale programs of insertion, chemical and X-rays mutagenesis, performed by different research groups, resulted in generation of numerous symbiotic mutants in *L. japonicus* and *M. truncatula* [163, 164] which allowed researchers to identify and characterize a series of *Sym* genes. The genes involved at the initial stages of nitrogen-fixing symbiosis (early *Sym* genes) were of primary interest, allowing dissection of the mechanisms by which the NF signal is perceived and transduced by host plants [165]. It turned to be that after the perception of NF, the nodulation process follows the same signalling pathway as AM does, with slight differences, though.

#### **3.2. Common Symbiosis Pathway (CSP)**

tered in culturing AM-fungi, caused by their obligate symbiotic lifestyle and impossibility of using selective media. Additionally, genetic analysis of AM fungi is more complex because of

The plant genes involved in development of RN and AM symbioses may be divided into two groups, according to approach which was used for the gene identification. The first group, *Sym* genes [146], had been identified with the use of formal genetic analysis (started from selection of plant mutants defective in nodule development). The other group of genes comprises nodulins (from *nodule*) [147], mycorrhizins (from *mycorrhiza*) [148, 149], and symbiosins (from *symbiosis*) [149, 150]. These genes were identified by mo‐ lecular genetic methods, through identification of proteins and/or RNAs synthesized *de novo* in root nodules (nodulins) or roots colonized by AM fungi (mycorrhizins). The genes whose expression is induced during the development of both endosymbioses, RNS

Genes of these groups are suspected to play different roles in the processes which may be referred collectively as "management of microsymbionts" inside plant roots. Specifically, the products of some nodulin genes represent the structural elements of newly constructed temporary compartments developed during symbiosis [151]. The other nodulin genes may play essential roles in modulating the hormonal status of the developing nodules [152, 153]. Resently, *in silico* and microarray-based transcriptome profiling approaches have allowed identification of nodulins, mycorrhizins and symbiosins, which are being activated in response to an AM fungal signal, or by either rhizobial or AM fungal stimulus, respectively [149]. Several hundred genes were found to be activated at different stages of either symbiosis, with almost 100 genes being co-induced during nodulation and in AM formation. These co-induced genes representing the common evolutionary bases of AM and nodular symbioses can be associated with those cellular functions which are required for symbiotic efficiency, such as the facilitation of nutrient transport across the perisymbiotic membranes that surround the endosymbiotic bacteroids in root nodules and the arbuscules in AM roots [150]. However, it should be remembered that although most of the nodulins/mycorrhizins/symbiosins were already cloned and sequenced, functions for many of them have been identified only preliminary using the sequence data of the encoding genes and location of the gene products in the symbiotic

Still, most of nodulins, mycorrhizins and symbiosins seem to play a subordinate role in the regulatory scheme of symbiosis, nevertheless being indispensable for its functionality and stability. In turn, the major, regulatory role in realization of symbiotic programmes is to be assigned to *Sym* genes. These genes, in contrast to nodulin genes, are usually not expressed outside symbiotic structures and there are many examples of the high functional and sequence homologies between them in different legumes. First genes of this group had been identified using the spontaneous mutants from natural legume populations [154] and afterwards using the experimentally induced mutants defective in nodulation or nitrogen fixation (Nod– and

 phenotypes) [155]. Afterwards, it was demonstrated that mutations in some of these genes also affect the ability of plant to form AM [148]. The presence of such common genes necessary for both AM and RN symbioses development suggested that both endosymbioses were more

their heterokaryotic nature and lack of sexual process [34, 145].

and AM are called symbiosins [150].

178 Plant Breeding from Laboratories to Fields

compartments.

Fix–

The data obtained during the last fifteen years allowed reconstruction the symbiotic signaling pathway which starts in RN and AM symbioses with recognition the Nod and Myc factors, respectively, and goes on as the signal transduction inside the root. In legumes Nod and Myc factors are most likely perceived by specific receptor complexes [35]. The receptor for NF is considered to be a heterodimer composed of at least two LysM containing receptor kinases [166-168]. Alike, receptor for Myc factor (which is not known yet) also supposed to be consisted of similar receptor kinases, or even include one or more kinases participating in Nod factor signaling [169]. The system of receptor kinases perceiving signal molecules of microsymbionts seems to be complicated, with some receptor complexes being necessary not only for starting the interactions, but also on later stages, during penetration bacteria into the root cortex through root hair. Moreover, some receptor kinases could non-specifically bind Nod and Myc factors, which results in intensified growth of lateral roots [43, 169]. Probably, the diversity of receptor kinases should complement the variability of soil microorganisms and increase the specificity of interactions with mutually beneficial ones.

After the first step of reception of Nod or Myc factor, the symbiotic signal is being transmitted to the common pathway, named Common Symbiosis Pathway [170]. The first player in this pathway is LRR-receptor kinase, or SymRK (symbiotic receptor kinase), which is required for both RN and AM symbioses development [171, 172]. The ligand for this receptor kinase is not known yet. Interestingly, the activity of this kinase is also required for proper progression of late symbiotic stages, at least for rhizobial infection [173]. SymRK kinase domain has been shown to interact with 3-hydroxy-3-methylglutaryl CoA reductase 1 (HMGR1) from *M. truncatula* [174], and an ARID-type DNA-binding protein [175]. These results suggest that SymRK may form protein complex with key regulatory proteins of downstream cellular responses. Symbiotic Remorin 1 (SYMREM1) from *M. truncatula* and SymRK-interacting E3 ligase (SIE3) from *L. japonicus* have also been shown to interact with SymRK [176, 177].

The symbiosis receptor kinase SymRK acts upstream of the Nod and Myc factor-induced Ca2+ spiking in the perinuclear region of root hairs within a few minutes after NF application [178]. Perinuclear calcium spiking involves the release of calcium from a storage compartment (probably the nuclear envelope) through as-yet-unidentified calcium channels. To date it is known that the potassium-permeable channels might compensate for the resulting charge imbalance and could regulate the calcium channels in plants [179-183]. Also, nucleoporins NUP85 and NUP133 (to date described only in *Lotus*) are required for calcium spiking, although their mode of involvement is currently unknown. Probably, NUP85 and NUP133 might be a part of specific nuclear pore subcomplex that plays a crucial role in the signal process requiring interaction at the cell plasma membrane and at nuclear and plastid organellemembranes to induce a Ca2+ spiking [184, 185]. Recently, the third constituent of a conserved subcomplex of the nuclear pore scaffold, NENA, was identified as indispensable component of AM and RN endosymbiotic development [186].

The calcium spiking is characteristic for both RN and AM symbioses formation [187]. These Ca2+ spikes are supposed to activate a calcium- and calmodulin-dependent protein kinase (CCaMK) that is also required for NF signaling and AM development [188]. This kinase contains an autoinhibition domain, removing of which leads to a spontaneous ac‐ tivation of downstream transcription events and induction of nodule formation in the ab‐ sence of rhizobia [189]. Thus, CCaMK appears to be a general manager for both symbioses activating different cascades of signaling for N2-fixing symbiosis and AM in response to different Ca2+ spiking, because the next steps of nodulation signaling are in‐ dependent from those of AM: the mutations in downstream *Sym* genes do not change the mycorrhizal phenotype of the legume. Interestingly, mutations in any *Sym* genes do not influence the defense reactions, suggesting that signaling pathways of mutualistic symbioses and pathogenesis are sufficiently different.

The calcium-calmodulin-dependent protein kinase (CCaMK) is supposed to be also involved in legume interactions with PGPR and/or endophytic bacteria as it was shown using inocula‐ tion of *M. truncatula* by *P. fluorescens* that *MtDMI3* gene (encoding for CCaMK) regulates intercellular root colonization by bacteria as well as expression of some plant housekeeping genes known earlier as mycorrhizins [190].

The CCaMK is known to form a complex with CYCLOPS, a phosphorylation substrate, within the nucleus [35]. *cyclops* mutants of *Lotus* severely impair the infection process induced by the bacterial or fungal symbionts, and are also defective in arbuscule development [149]. During RNS, *cyclops* mutants exhibit the specific defects in infection-thread initiation, but not in the nodule organogenesis [191], indicating that CYCLOPS acts in an infection-specific branch of the symbiotic signaling network [35]. *Cyclops* encodes a protein with no overall sequence similarity to proteins with known function, but containing a functional nuclear localization signal and a carboxy-terminal coiled-coil domain.

It is supposed that CCaMK with help of CYCLOPS probably phosphorilates the specific transcription factors already present in cell: NSP1 and NSP2, which influence the changes of expression in several genes related to the symbiosis development [192, 193]. The activity of these proteins leads to the transcriptional changes in root tissues, for instance, increasing the level of early nodulins ENOD40, ENOD11, ENOD12, ENOD5, which are known to be the potential regulators of infection thread growth and nodule primordium formation [165, 194, 195). Also, the changes in cytokinin status of plant are detected, followed by up-regulation of genes encoding for RN symbiosis-specific cytokinin receptors [196-198]. Moreover, transcrip‐ tion regulators NIN and ERN are to be induced specifically downstream of the early NF signaling pathway in order to coordinate and regulate the correct temporal and spatial formation of root nodules [199-202].

The presented genes are responsible for the signal cascade which is aimed to induce the nodulin, mycorrhizin and symbiosin genes responsible for building the symbiotic struc‐ tures and implementing their biochemical functions. It is supposed that the signaling pathway did not appear *de novo* in legumes when they become able to form nodules, but was developed from already existed system of AM formation into which the novel, nod‐ ule-specific genes were recruited. Still, new genes had been involved in RN symbiosis development, especially those encoding the receptors recognizing hormones (e.g. cytoki‐ nins) and hormone-like molecules (Nod factors).

#### **3.3. Autoregulation of symbioses formation**

known yet. Interestingly, the activity of this kinase is also required for proper progression of late symbiotic stages, at least for rhizobial infection [173]. SymRK kinase domain has been shown to interact with 3-hydroxy-3-methylglutaryl CoA reductase 1 (HMGR1) from *M. truncatula* [174], and an ARID-type DNA-binding protein [175]. These results suggest that SymRK may form protein complex with key regulatory proteins of downstream cellular responses. Symbiotic Remorin 1 (SYMREM1) from *M. truncatula* and SymRK-interacting E3 ligase (SIE3) from *L. japonicus* have also been shown to interact with SymRK [176, 177].

The symbiosis receptor kinase SymRK acts upstream of the Nod and Myc factor-induced Ca2+ spiking in the perinuclear region of root hairs within a few minutes after NF application [178]. Perinuclear calcium spiking involves the release of calcium from a storage compartment (probably the nuclear envelope) through as-yet-unidentified calcium channels. To date it is known that the potassium-permeable channels might compensate for the resulting charge imbalance and could regulate the calcium channels in plants [179-183]. Also, nucleoporins NUP85 and NUP133 (to date described only in *Lotus*) are required for calcium spiking, although their mode of involvement is currently unknown. Probably, NUP85 and NUP133 might be a part of specific nuclear pore subcomplex that plays a crucial role in the signal process requiring interaction at the cell plasma membrane and at nuclear and plastid organellemembranes to induce a Ca2+ spiking [184, 185]. Recently, the third constituent of a conserved subcomplex of the nuclear pore scaffold, NENA, was identified as indispensable component

The calcium spiking is characteristic for both RN and AM symbioses formation [187]. These Ca2+ spikes are supposed to activate a calcium- and calmodulin-dependent protein kinase (CCaMK) that is also required for NF signaling and AM development [188]. This kinase contains an autoinhibition domain, removing of which leads to a spontaneous ac‐ tivation of downstream transcription events and induction of nodule formation in the ab‐ sence of rhizobia [189]. Thus, CCaMK appears to be a general manager for both symbioses activating different cascades of signaling for N2-fixing symbiosis and AM in response to different Ca2+ spiking, because the next steps of nodulation signaling are in‐ dependent from those of AM: the mutations in downstream *Sym* genes do not change the mycorrhizal phenotype of the legume. Interestingly, mutations in any *Sym* genes do not influence the defense reactions, suggesting that signaling pathways of mutualistic

The calcium-calmodulin-dependent protein kinase (CCaMK) is supposed to be also involved in legume interactions with PGPR and/or endophytic bacteria as it was shown using inocula‐ tion of *M. truncatula* by *P. fluorescens* that *MtDMI3* gene (encoding for CCaMK) regulates intercellular root colonization by bacteria as well as expression of some plant housekeeping

The CCaMK is known to form a complex with CYCLOPS, a phosphorylation substrate, within the nucleus [35]. *cyclops* mutants of *Lotus* severely impair the infection process induced by the bacterial or fungal symbionts, and are also defective in arbuscule development [149]. During RNS, *cyclops* mutants exhibit the specific defects in infection-thread initiation, but not in the nodule organogenesis [191], indicating that CYCLOPS acts in an infection-specific branch of

of AM and RN endosymbiotic development [186].

180 Plant Breeding from Laboratories to Fields

symbioses and pathogenesis are sufficiently different.

genes known earlier as mycorrhizins [190].

Autoregulation of symbiosis development is an important process that takes place after successful mutual partners' recognition and signal exchange. For RN symbiosis, it is con‐ sidered that legume host controls the root nodule numbers by sensing the external and internal cues. A major external cue is the soil nitrate, whereas a feedback regulatory sys‐ tem in which earlier formed nodules suppress further nodulation through shoot-root communication is an important internal cue. The latter is known as the autoregulation of nodulation (AON), and is believed to consist of two long-distance signals: a root-derived signal that is generated in infected roots and transmitted to the shoot; and a shoot-de‐ rived signal that inhibits nodulation systemically [203, 204]. Therefore, AON represents a strategy through which the host plant can balance the symbiotrophic N nutrition with the energetically more "cheap" combined N nutrition.

Recent findings on autoregulation of nodulation suggest that the root-derived ascending signals to the shoot are short peptides belonging to the CLE peptide family [205, 206]. The leucine-rich repeat receptor-like kinase HAR1 of *Lotus* and its homologues in *M. truncatula* and *P. sativum* (*SUNN* and *Sym29*, respectively) mediate AON and also the nitrate inhibition of nodulation, presumably by recognizing the root-derived signal [207-210]. Other genes, like *ASTRAY*, *KLAVIER* and *TML* in *Lotus*, and *RDN1* in *M. truncatula*, are also supposed to play a sufficient role in AON [211-214].

Very little is still known about the plant regulation of mycorrhization process. In split-root systems on alfalfa (*Medicago sativa*), inoculation of one half of a split-root system with the fungus *Glomus mosseae* significantly reduced later AM colonization on the other half. A similar suppressive effect on mycorrhization was observed after inoculation with *Sinorhizobium meliloti* [215]. Furthermore, prior addition of purifed rhizobial Nod factors on one half signifcantly reduced mycorrhization on the other half of the split-root system, and reciprocally, prior mycorrhization on one side suppressed nodule formation on the other side of the splitroot system. Together these data point to a common autoregulation circuit for both symbioses [210]. It was suggested that Nod factor signaling, as well as mycorrhizal Myc factor signaling, induces expression or post-translation processing of CLE peptides, which likely function as ascending long-distance signals to the shoot [210]. Also, it was demonstrated that mutations in *HAR1* of *Lotus* and corresponding orthologues in other legumes increase both nodulation and mycorrhization suggesting the shared role of these orthologous genes in controlling the rate of root colonization by microsymbionts. Thereby, not only the local signal transduction (CSP) but the systemic autoregulation is common for the RN and AM symbioses.

#### **3.4. Next stage of development the genetics of symbioses**

The next-coming step of development the genetics of symbioses is studying gene networks on intergenomic level, i.e. the coordinated expression of plant and microbe genes. For AM, with use of the new molecular approaches, in particular transcriptomics, a series of AM fungal genes has been identified, having altered expression levels during the AM formation [216-219]. Still it is not well studied at which stages of fungal-plant interaction the complementary partners' genes are induced or repressed, and so the use of plant mutants impaired at different steps of AM development might be a challenging approach to reveal the pattern of plant and fungal genetic cooperation [220]. The same research aimed at identification of plant-rhizobial gene interactions with the use of plant and microbe mutants is also in progress.

It has been recently observed that *Medicago truncatula* showed significantly lower efficiency of nitrogen fixation than its close relative *Medicago sativa* L. [221]. The number of nodules formed on the roots of *M. truncatula* was less than that of *M. sativa*, and the nitrogen fixation measured on plants at the beginning of flowering (as well as specific N2 fixation (μgN h-1 mg nodule-1)) was significantly lower. The reasons for the low efficiency in nitrogen fixation were partially a result of low relative efficiency (electron allocated to N2 versus H+ ), and slow nitrogen export from nodules in *M. truncatula* when compared to *M. sativa*. This might be connected with a low malate concentration in the nodule tissue of *M. truncatula*, and thus insufficient carbon provision for asparagine formation (fixed nitrogen is to be added to malate to form asparagine) [221]. Therefore, Sulieman and Schulze [221] suggest that improvement the malate formation in *M. truncatula* nodules could help improving the effectiveness of nitrogen fixation.

According to these data, genes encoding for enzymes of malate synthesis should be good candidates for markers to be used as selection and breeding aimed at improvement of symbiotic properties in *M. truncatula*. But, in different species potential markers of symbiotic effectiveness could be found among genes of different functional groups. Our original data on sequencing alleles of symbiotic genes in pea (*Pisum sativum*) varieties with different symbiotic effectiveness suggest that polymorphism of genes belonging to CSP does not correspond to symbiotic properties of pea varieties analysed (Zhukov V.A., unpublished results). Perhaps, good candidates for markers of symbiotic effectiveness could be found during large-scale screening by transcriptome sequencing in different pea genotypes, which is now underway in our laboratory.
