Identiﬁcation and Characterization of Common Bean ( Phaseolus vulgaris ) Non-Nodulating Mutants Altered in Rhizobial Infection

: The symbiotic N 2 -ﬁxation process in the legume–rhizobia interaction is relevant for sustainable agriculture. The characterization of symbiotic mutants, mainly in model legumes, has been instrumental for the discovery of symbiotic genes, but similar studies in crop legumes are scant. To isolate and characterize common bean ( Phaseolus vulgaris ) symbiotic mutants, an ethyl methanesulphonate-induced mutant population from the BAT 93 genotype was analyzed. Our initial screening of Rhizobium etli CE3-inoculated mutant plants revealed different alterations in nodulation. We proceeded with the characterization of three non-nodulating ( nnod ), apparently monogenic/recessive mutants: nnod (1895), nnod (2353) and nnod (2114). Their reduced growth in a symbiotic condition was restored when the nitrate was added. A similar nnod phenotype was observed upon inoculation with other efﬁcient rhizobia species. A microscopic analysis revealed a different impairment for each mutant in an early symbiotic step. nnod (1895) formed decreased root hair curling but had increased non-effective root hair deformation and no rhizobia infection. nnod (2353) produced normal root hair curling and rhizobia entrapment to form infection chambers, but the development of the latter was blocked. nnod (2114) formed infection threads that did not elongate and thus did not reach the root cortex level; it occasionally formed non-infected pseudo-nodules. The current research is aimed at mapping the responsible mutated gene for a better understanding of SNF in this critical food crop.


Introduction
A key to the success of the legume family, comprising around 700 genera and accounting for one-third of the primary crop production in the world, was the evolution of symbioses with mycorrhizal fungi and nitrogen-fixing bacteria, collectively known as rhizobia, that facilitate plant nutrient acquisition [1][2][3]. The symbiotic N 2 -fixation (SNF) process in the legume-rhizobia interactions is ecologically and economically relevant for sustainable agriculture; it is a valuable source of nitrogen, which is an essential element for crop production [4]. This symbiotic process, occurring in N-limited soils, takes place in root-developed specialized organs called nodules. Nodulation and SNF are complex processes that are tightly regulated, both in rhizobia and in its specific legume host. The first step in this interaction is the communication between the two symbionts through molecular signals, namely flavonoids secreted by the plant root to the rhizosphere that trigger the compatible rhizobia biosynthesis of lipochitooligosaccharides, known as nodulation factors selected symbiotic mutants, with stable altered phenotypes, by analyzing the presence or absence of known structures developed in different steps of the symbiosis, such as rhizobial infection organogenesis.
Our screening of the mutant lines inoculated with Rhizobium etli CE3 revealed lines with different alterations in nodulation. We characterized three selected mutant lines, appearing to be monogenic and recessive, showing a stable non-nodulating phenotype. A phenotypic analysis of these mutants evidenced distinct alterations in each one of them that affected the different early steps of the rhizobial infection process: the rhizobia-induced root hair deformation, the infection chamber formation and infection thread development.

Screening for Altered Nodulation Phenotype in the Mutagenized Common Bean Population
In this work, 1692 M 4 lines from an EMS mutant population of the wild type (WT) BAT 93 P. vulgaris genotype [21] were screened for rhizobia-symbiosis mutants. Seeds from each line were inoculated with R. etli CE3 [38]. After 4 weeks, when the nodulation is expected to be at the maximum, mutant plants were carefully removed from their pots and their nodulation capacity was recorded after a visual comparison with that of BAT 93 wild type (WT) plants. In the initial screening, 93 lines with an altered nodulation phenotype, rated as an increased nodulation, decreased nodulation or no nodulation, were selected. In a second screening performed to confirm their nodulation-impaired phenotype, most of the mutant lines showed a wild-type nodulation phenotype. Thus, we decided to further analyze 36 mutant lines that showed a non-nodulation phenotype, by exploring the segregation of nodulation (wt) vs. non-nodulation (nnod) in the M 5 progeny. We screened 9 to 12 M 5 plants derived from each of two M 4 plants of the selected mutant lines (Supplementary Table S1). The majority-23 out of 36 lines analyzed-showed normal nodulation (wt). The remaining 13 lines showed variable wt:nnod segregation, ranging from 2:1 to 7:1. The mutant lines, 08IS-1895, 08IS-2353 and 08IS-2114, showed the closest 3:1 segregation expected for a monogenic and recessive mutation. Therefore, we selected these three mutant lines, hereafter denominated nnod(1855), nnod(2353) and nnod(2114), for further phenotypic analysis.

Growth, Root Phenotype and Nodulation Capacity of Nnod Mutants
Contrasting with the WT plants, the R. etli CE3-inoculated nnod(1895), nnod(2353) and nnod(2114) mutant plants showed symptoms of nitrogen deficiency, such as reduced growth and chlorotic leaves ( Figure 1A,B). However, the growth of the three mutant lines was fully restored when a sufficient amount of nitrate was supplied ( Figure 1B). The latter indicates that the growth impairment of inoculated nnod mutants was due to the lack of symbiotic nitrogen fixation.
Seedlings from the WT and from each of the nnod mutants were inoculated with three different rhizobia strains, which are effective and common bean symbionts: R. etli CE3 [38], Sinorhizobium fredii NGR234 [39] and Rhizobium phaseoli CIAT652 [40]. Each bear a construct for the constitutive GUS gene expression. After 25 days post-inoculation (dpi), the root growth and nodulation capacity were evaluated.
As expected, the WT plants demonstrated effective nodulation with each of the rhizobia inoculum ( Figure 3A). The nnod(1895) and nnod(2353) mutant plants were unable to develop nodules when inoculated with any of the rhizobia strains used ( Figure 3A), thus indicating that their nnod phenotype resulted from the mutation and was not dependent on the rhizobia inoculum. The nnod(2114) plants did not form nodules when S. fredii NGR234 or R. phaseoli CIAT652 were used as inoculum. However, in R. etli CE3-inoculated nnod(2114) plants, a few white nodules, or pseudo-nodules, could occasionally be observed (Figures 2A and 3A). These pseudo-nodules were 50% smaller than nodules formed in the BAT 93 plants ( Figure 3B). The GUS reporter gene signal revealed a full R. etli CE3 infection in WT nodules that was evident even in small and immature nodules ( Figure 3C,D). Contrastingly, no R. etli CE3/pGUS infection was detected in nnod(2114) pseudo-nodules ( Figure 3C,D). Our results demonstrated that from the three common-bean nnod mutants analyzed, only nnod(2114) mutant plants inoculated with R. etli CE3 were capable of forming empty, non-functional pseudo-nodules.

Comparative Microscopical Analysis of Inoculated BAT 93 and Nnod Mutants' Inoculated Roots at the Early Stages of the Symbiotic Process
In order to gain information about the symbiotic process alterations of the analyzed nnod mutants, a detailed microscopic evaluation of the morphological features that distinguish the different steps of rhizobia infection, during the early symbiotic stage, was carried out. The morphological features of the primary host infection were examined by light microscopy after the GUS-staining of R. etli CE3/pGUS 25 dpi roots. The rhizobial infection process starts with the rhizobia attachment to the root, mainly RH tips, that triggers pre-infection responses such as RH swelling and curling, which are required for the rhizobia colonization [2]. The evaluation of the RH deformation revealed a similar amount of characteristic and effective deformed or curled RH in WT, and nnod(2353) and nnod(2114)-inoculated roots ( Figure 4A). However, the nnod(1895)inoculated roots demonstrated decreased RH curling and a higher number of non-effective or "spatula-like" [41] RH deformations ( Figure 4A,B). These results ( Figure 4) indicate that the nnod(1895) mutant was impaired in the initial step (RH curling) of the symbiotic pathway, which blocks signal transduction pathways triggered by rhizobia.
.c: In the next step of rhizobia infection, bacteria are enclosed in the curled RH pocket known as the infection chamber or microcolony, which allows for the formation and development of IT [42]. A difference in the infection chamber formation was observed in the nnod(2353)-inoculated roots. Though the nnod(2353)-inoculated roots formed a similar number of effective RH curling as WT and nnod(2114) ( Figure 4A), its GUS-stained roots demonstrated a high amount of small, blue dots that higher microscope magnification identified as infection chambers (average of 7.3 infection chambers per root cm) that were arrested at this step and thus unable to develop into normal IT ( Figure 5). The absence of visible blue dots in WT or nnod(2114)-inoculated roots indicates that in these plants, the normally formed infection chambers were not arrested and developed to the next steps of the symbiotic pathway: the formation of the IT. As mentioned above, the nnod(2114) was the only common bean mutant whose plants form empty pseudo-nodules after R. etli CE3 inoculation (Figure 3), thus indicating that these mutant plants did recognize rhizobial signaling, were infected by rhizobia and partially activate nodule organogenesis. Our root hair deformation analysis demonstrated a slight increase in the effective root hair formation in nnod(2114) compared to WT-inoculated plants ( Figure 4A). The analysis of formed IT in the WT demonstrated its normal development, with the IT reaching the root cortex cell layer, while the nnod(2114)-inoculated roots only showed the truncated IT unable to reach the root cortex ( Figure 6A). In addition, the analysis of 10 dpi roots revealed an increased number of truncated IT formed in the nnod(2114) compared to normal IT formed in WT roots ( Figure 6B). These results indicate that nnod(2114) plants were able to recognize rhizobia, as evidenced by the effective RH deformation and infection chamber formation, but the development of the fully formed IT was impaired, and the IT were thus incapable of reaching the root cortex; a few of these abnormal infections may progress into empty pseudo-nodule formation.

Discussion
The symbiosis between plants and nitrogen-fixing bacteria is a unique model of an intimate interaction that has major ecological impacts. The advent of plant genomics and refined genetic resources has contributed to the current understanding of various essential protagonists of this complex symbiosis. Particularly in the model legumes, M. truncatula and L. japonicus, more than 200 symbiotic genes have been identified [6]. However, for agriculturally important legume crops such as P. vulgaris, such relevant genetic information remains scant. In this work, we analyzed an EMS mutant population of the BAT 93 genotype, developed by Porch et al. [21], to identify and characterize mutants affected in the rhizobia N2-fixing symbiosis. Using this approach, non-nodulating common bean mutants were identified several years ago [17,20,31,33]; however, the precise physiological characterization identifying a specific impaired symbiotic or genetic analysis identifying the responsible mutated genes have not been reported. In addition, Park and Buttery [19] identified the nitrate-tolerant common bean super-nodulating mutants; these have been recently molecularly and genetically characterized by Ferguson et al. [34]. Based on previous research on the soybean, using comparative-and functional-genomics approaches-two alleles of the common bean gene PvNARK (NODULE AUTOREGULATION RECEPTOR KINASE) were identified as responsible for the supernodulation phenotype of two common bean mutants in each of the R32 and SV45 cultivars [34]. The soybean and common bean NARK genes, as well as orthologous genes from different legumes, code for specialized LRR-receptor kinases that are key components of the AON, a systemic signaling pathway that controls the nodule number. The AON begins with the production of root-specific CLE peptides that are transported via the xylem to the shoot and are perceived by NARK which, in turn, leads to the production of a shoot-derived inhibitor signal that is transported to the root, where it acts to inhibit further nodulation events [5,7,43]. To our knowledge, this is the only fully characterized common bean rhizobial-symbiotic mutant.
Our initial screening of the R. etli CE3-inoculated M 4 common bean mutant plants demonstrated several lines with altered nodulation, such as increased, decreased or no nodulation, as compared to the WT BAT 93-inoculated plants. However, most of these mutants nodulated normally in the subsequent generations, thus indicating that a large number of plants were erroneously selected on the basis of nonheritable characteristics. Thus, we proceeded to characterize three mutant lines with a non-nodulating phenotype (nnod), indicative of a monogenic, recessive mutation in each line and that has been stable up to the M 9 generation. The diminished growth rate observed in the inoculated nnod(1855), nnod(2353) and nnod(2114) mutant plants, as compared to the WT BAT 93 plants, was restored when plants were grown in a non-symbiotic condition, with optimal nitrate content added. This result, which is in agreement with results from the common bean nonnodulating mutants previously analyzed [17,20,44], led us to conclude that the nnod(1855), nnod(2353) and nnod(2114) mutants can assimilate and utilize inorganic N for normal plant growth, but their inability to nodulate prevents the N fixation and growth under rhizobia symbiosis conditions. Plants from nnod(1855) and nnod(2353) mutant lines displayed the non-nodulating phenotype across the three different rhizobia species (R. etli CE3, S. fredii NGR234 and R. phaseoli CIAT652) used as inoculum; these are effective symbionts for WT common bean plants [38][39][40]. Occasionally, a few small/white pseudo-nodules were observed in nnod(2114) mutant plants inoculated with R. etli CE3 but not with the other rhizobia species. The observed nnod(2114) pseudo-nodules were empty without rhizobia invasion. These results, also reported for the previously analyzed common bean mutants [31], indicate that the non-nodulating phenotype results from the mutation in each mutant line and not from the rhizobia inoculum used.
In this work, we went beyond the observation of the presence/absence of nodules in the non-nodulating mutants to discover at which stage of the rhizobia infection or nodule organogenesis processes were impaired. Our phenotypic analysis revealed variations among the mutants, indicating that each mutant bears a different genetic defect, blocking different steps of the rhizobial infection process, as depicted in Figure 7.
The current knowledge about legume genes relevant for the rhizobia symbiosis [6] provides a comprehensive set of genes that were classified according to their participation in different symbiotic stages, such as early signaling, host range restriction and rhizobial infection. This information, as well as the phenotypes of legume mutants impaired in early symbiotic genes and their similarity with the phenotype of the common bean symbiotic mutants here reported, allowed us to propose possible candidates of mutated genes causal for the defects observed in each mutant.
Contrasting with RH deformations observed in the WT and the other nnod mutants, the nnod(1895) mutant demonstrated less effective RH curling, which results from the altered rhizobia communication and initial interaction with the legume root [2], and a higher amount of "spatula-like" non-effective RH deformations. The RH showing aberrant-ineffective-deformations are unable to trap rhizobia and to form the infection chamber required for allowing the rhizobia entry via the IT formation and development [41,45]. Our results indicate that the common bean nnod(1895) mutant is capable of the rhizobia recognition and is impaired in the very early steps of rhizobia infection (Figure 7). The analysis of model legume mutants revealed that defects in specific NF-receptors do not allow compatible rhizobia recognition by the legume root; thus, no RH deformation is observed [46]. Instead, if legume genes participating in the rhizobia infection signaling pathway are mutated, a phenotype of the decreased effective RH curling deformation and increased ineffective RH deformation has been observed. The possible candidate genes that we propose for the nnod(1895) mutant are: SYMRK (SYMBIOSIS RECEPTOR-LIKE KINASE) [47], NUP133, NUP85 (NUCLEOPORINS) [48,49], ERN (ETHYLENE RESPONSE FACTOR REQUIRED FOR NODULATION) [45] and AGO5 (ARGONAUTE 5) [41]. The mutants Ljsymrk, Ljnup85, Ljnup133 and Mtern1/ern2 double mutant are non-nodulating mutants that show increased and non-effective-"spatula like", swelling and branching-RH deformation, and no infection chambers or IT development [41,45,[47][48][49]. In addition, P. vulgaris roots silenced for AGO5 showed an increase in the "spatula-like" RH deformation [41]. The nnod(2353) and nnod(2114) mutants are impaired in the rhizobial infection process (Figure 7). These mutants are able to form RH effective deformations, which are important for trapping rhizobia, allowing bacteria to divide and form a microcolony known as the infection chamber [6,42]. In addition, the nnod(2114) forms IT, but only truncated ITs are observed in the epidermis layer, being unable to develop and reach the root cortex cell layer. Few empty pseudo-nodules were rarely formed in the R. etli CE3-inoculated nnod(2114) roots, but not when S. fredii NGR234 or R. phaseoli CIAT652, the compati-ble symbionts, were used as the inoculum. A similar phenotype was reported for the Ljnup133 mutant that formed few ineffective white nodules when inoculated with two compatible rhizobia, but not with a third strain tested [48]. We propose the following as possible candidate genes' causal for the rhizobial infection defects of the nnod(2353) and nnod(2114) mutants: EPR (EXOPOLYSACHARIDE RECEPTOR) [50], CYCLOPS [51], CBS (CYSTATHIOINE-β-SYNTHASE-like domain-containing) [52], FLOT4 (FLOTILIN) and REM1 (REMORIN) [53], NPL (NODULATION PECTATE LYASE GENE) [54], the ARPC1 (ACTIN-RELATED PROTEIN COMPONENT 1) [55], SCARN (SUPPRESOR OF cAMP RECEPTOR DEFECT-NODULATION) [56], NAP (NCK-ASSOCIATED PROTEIN1) and PIR (121F-SPECIFIC P53 INDUCIBLE RNA) [57] genes that participate in the actin remodeling required during rhizobial infection, VPY (VAPYRIN) [58], LIN(LUMPY INFEC-TIONS)/CERBERUS [59], RPG (RHIZOBIUM-DIRECTED POLAR GROWTH) [60] and IEF (INFECTION-RELATED EPIDERMAL FACTOR) [61] genes that are necessary for the IT polar growth into the cortex. The model legume mutants impaired in one of these rhizobial infection genes show a similar phenotype-effective RH deformation, infection chamber formation and development of ITs that are truncated, branched and not progressed beyond the epidermis-as those are observed for common bean nnod(2353) and nnod(2114) mutants. Some of the model legume mutants are also capable of starting nodule organogenesis, forming small, white, empty pseudonodules [50,54,[57][58][59][60][61], similar to the nnod(2114) mutant. A hypothesis to explain this is that the signal (NFs) sent by rhizobia in the IT is necessary and sufficient to trigger this process [51].
Contrasting with WT and the other two nnod mutants, a short root phenotype was observed in the nnod(2114) mutant plants. This phenotype could be associated with the nnod phenotype, as in the LjNap1 and LjPir1 mutants with a shorter root phenotype that was observed but not further analyzed [57]. Another possibility is that the short root phenotype is caused by another unrelated mutated gene present in the nnod(2114) mutant line, something that is common in mutant populations generated by EMS treatment. The analysis of short root and nnod phenotypes in individuals from the progeny of a back cross of nnod(2114) × BAT 93 will distinguish between these two possibilities, revealing whether the two phenotypes co-segregate or not.
The current research in our group focuses on identifying the mutated gene of nnod(1985), nnod(2353) and nnod(2114), which is responsible for the rhizobial symbiosis impairment in each mutant. For this, we consider the strategies of the whole genome sequence or targeted exon capture/sequencing. The whole genome sequence approach, using appropriate algorithms for a bioinformatic analysis, has been successfully used to identify a mutated gene of interest in different plants. For, example Addo-Quaye et al. [62] reported the identification of a mutated gene participating in GA biosynthesis using this approach in EMS-generated Sorghum bicolor mutants. We propose to do a comparative genome sequence analysis between genome sequences from gDNA pools from wt or nnod individuals obtained from the backcross BAT 93 × nnod mutant. The targeted exon capture/sequencing [63] has proven to be successful for different plant species. The whole exome sequence implies a considerable initial investment to build the required platform. However, we plan to use this method, focusing in exomes of the candidate mutated genes that we have proposed for our nnod mutants.
The achievement of the identification of the causal mutated genes will contribute to a better understanding of the relevant genes for the rhizobia infection process in the common bean. In addition, symbiotic mutants have been instrumental for identifying genes that participate in different related regulatory pathways of the rhizobia symbiosis [6]. An example of the latter has recently been published [64] by using the only common bean symbiotic mutant: Pvnark [34]. The analysis of Pvnark and Gmnark mutants, using physiological, reciprocal grafting and split-root approaches, determined that the AON pathway is required for the inhibition of the nodulation by phosphorus deficiency [64]. Thus, the identification of the mutated genes in the common bean non-nodulating mutants analyzed in this work would allow for the use of these mutants as a tool to answer research questions about rhizobial symbiosis regulation in the most important legume for human consumption [4].

Plant Material and Growth Conditions
The common bean (P. vulgaris) BAT 93 wild type genotype, representative of the Mesoamerican gene pool [35], was used in this work. This genotype was developed at the International Center for Tropical Agriculture (CIAT, Cali Colombia) and derived from a double cross involving four Mesoamerican genotypes.
The screening of symbiotic mutants was performed in 1692 M 4 lines of the EMS mutant population developed by Porch et al. [21] in the BAT 93 background. These seeds were bulk-harvested from plants grown in the field in Puerto Rico.
Seeds were surface sterilized in 10% V/V commercial sodium hypochlorite for 5 min and rinsed with distilled water. Then seeds were germinated on sterile moistened filter paper at 30 • C in darkness for two days. Germinated seedlings were planted in pots with wet sterile vermiculite/agrolite (1:1) and grown in glasshouse with semi-controlled conditions (temperature 25-30 • C, 70% humidity and natural illumination).
For SNF evaluation immediately after planting, plantlets were inoculated with 1 mL of saturated liquid culture of the desired rhizobial strain (see below), applied directly to the root. Plants were watered, once a week, with 100 mL per plant N-free (50%) Summerfield nutrient solution [65] and subsequently with tap water 2 times per a week. For the nonsymbiotic condition, a full nutrient Summerfield solution (7 mM N-content) was used to water the plants.

Screening for Symbiotic Mutants
Two seeds from each M 4 mutant line were used for the initial screening. Two seeds were surface sterilized, germinated and planted in 20 cm diameter pots with sterile vermiculite, with two seeds per line in each pot. In addition, pots with two WT BAT 93 seeds each, were included for a comparison of nodulation capacity. Plants were inoculated with Rhizobium etli CE3 strain, as described. After 28 days of growth, the plants were uprooted and visually inspected for nodulation, comparing mutated lines with WT BAT 93 plants. Plants of mutated lines with visible changes in their symbiotic phenotype were re-planted in the same pots and grown to maturity. To boost seed production, NPK-granulated fertilizer was added and the full nutrient Summerfield solution was used for watering. The M 5 seeds were collected from each line and two seeds per line were re-screened to confirm the stability of their altered nodulation phenotype.

Observation (Analyses) of Infection and Nodulation Events
WT and nnod mutants were analyzed after 25 dpi to examine the symbiosis phenotype. Five plants per genotype were inoculated with different strains of rhizobia. In order to probe whether the rhizobia colonized and established a symbiosis with the plants, roots and nodules were immersed in a GUS staining solution (0.05% 5-bromo-4-chloro-3-indolyl-β-dglucuronic acid, 100 mm sodium phosphate buffer (pH 7), 0.5 mm potassium ferrocyanide, 0.5 mm potassium ferricyanide, 10 mm Na2EDTA and 0.1% Triton X-100) and incubated for 3 h at 37 • C. Roots were then cleared in diluted 10% NaClO for 10 min, and rinsed in phosphate buffer. Nodule images were taken using a ZEISS SterREO Discovery microscope and measured with a digital Vernier caliper.

Root Analysis
The main root at 25 dpi in WT and nnod mutant plants was measured from digital pictures. These images were processed with SmartRoot software [70]. Fifteen plants of each phenotype were analyzed.
The root hair phenotype was analyzed with a bright field microscope (ZEISS, Axioskop 2). Two segments of the root from WT and nnod mutants were taken after GUS staining for counting the number of effective or non-effective root hairs per cm. Two plants per genotype were analyzed.

Statistical Analysis
All results are reported were analyzed by one-way ANOVA followed using Dunnett's multiple comparisons test (p < 0.05) or one-tailed t tests (p < 0.05) with the software GraphPad Prism version 9.4.1 for Windows, GraphPad Software, San Diego, CA, USA (www.graphpad.com) (accessed on 1 September 2022).

Conclusions
In the present study, the EMS-generated mutant population from the BAT 93 P. vulgaris genotype [21] was screened for mutants altered in the rhizobia SNF process. Several mutant lines, inoculated with Rhizobium etli CE3, revealed different symbiotic alterations. From these, three mutant lines with a stable non-nodulating (nnod) phenotype, indicative of a single recessive mutation, were characterized. Their phenotypic analysis revealed an impairment in a different step of the rhizobial infection process in each of the mutants. The nnod(1895) mutant is impaired in root hair deformations, which is an initial step after the plant sensing of the Nod-Factors rhizobial signal. The nnod(2353) mutant showed normal root hair curling and rhizobia entrapment to form infection chambers but was blocked at this step. The nnod(2114) formed infection threads that did not elongate and thus did not reach the root cortex level; it occasionally formed non-infected pseudo-nodules. Several possible candidate genes for each of the mutants were proposed.