Nitrogen-Fixing Symbiotic Paraburkholderia Species: Current Knowledge and Future Perspectives

: A century after the discovery of rhizobia, the ﬁrst Beta-proteobacteria species (beta-rhizobia) were isolated from legume nodules in South Africa and South America. Since then, numerous species belonging to the Burkholderiaceae family have been isolated. The presence of a highly branching lineage of nodulation genes in beta-rhizobia suggests a long symbiotic history. In this review, we focus on the beta-rhizobial genus Paraburkholderia , which includes two main groups: the South American mimosoid-nodulating Paraburkholderia and the South African predominantly papilionoid-nodulating Paraburkholderia . Here, we discuss the latest knowledge on Paraburkholderia nitrogen-ﬁxing symbionts in each step of the symbiosis, from their survival in the soil, through the ﬁrst contact with the legumes until the formation of an efﬁcient nitrogen-ﬁxing symbiosis in root nodules. Special attention is given to the strain P. phymatum STM815 T that exhibits extraordinary features, such as the ability to: (i) enter into symbiosis with more than 50 legume species, including the agriculturally important common bean, (ii) outcompete other rhizobial species for nodulation of several legumes, and (iii) endure stressful soil conditions (e.g., high salt concentration and low pH) and high temperatures.


Introduction
Nitrogen is one of the five chemical elements essential to any living organism. Organic compounds, such as nucleic and amino acids, all contain nitrogen. The largest nitrogen reservoir on Earth is atmospheric dinitrogen gas (N 2 ) [1,2]. Although it might seem contradictory, N 2 cannot be directly used by most organisms, including plants, which makes it a limiting factor in crop productivity [3]. The triple bond between the two nitrogen atoms can only be broken by certain groups of microorganisms (prokaryotes and archaea), which transform the N 2 molecule into ammonia that is then ionized into ammonium, a nitrogenous form that plants can readily metabolize [4]. The term "biological nitrogen fixation" (BNF) refers to this process, which can be performed with the nitrogenase enzyme by both symbiotic and free-living bacteria [5]. Indeed, leguminous plants can host microsymbionts that fix N 2 , called "rhizobia", in root (and sometimes stem) nodules [6][7][8][9]. To date, only the Alphaand Beta-classes of Proteobacteria have been recognized as containing genuine rhizobia [10].
In this review, we summarize the most recent findings on N 2 -fixing symbionts (Betarhizobia) belonging to the Beta-proteobacterial genus Paraburkholderia, detailing the information, to date, on their geographical distribution and phylogenetic relatedness, since their discovery 20 years ago up to the present day. We draw attention to key processes occurring in different environments where rhizobia are located prior to and during the symbiosis (i.e., soil, rhizosphere, plant root and nodules) and discuss the mechanisms employed to survive abiotic and biotic stresses, their interaction with host plants until the allocation and differentiation of the free-living rhizobia into N 2 -fixing bacteroids. In particular, we focus on the strain Paraburkholderia phymatum STM815 T , which possesses exceptional features that allow it to outperform other rhizobia during the different stages of the symbiosis. beta-rhizobia strains fixed less N 2 than the non-symbionts, suggesting that they had become progressively adapted to a non-free-living, symbiotic lifestyle more typical of rhizobia in general [18].
At this time, symbiotic beta-rhizobial strains were still classified as Burkholderia, alongside human, animal, and plant pathogens, such as members of the Burkholderia cepacia complex (BCC) or the B. pseudomallei group [18]. There have been several efforts to highlight the differences between the pathogenic Burkholderia and environmental strains. For instance, phylogenetic studies based on 16S ribosomal RNA (rRNA) and multilocus sequence analysis (MLSA) of housekeeping genes revealed the separation of these two lineages [18,[39][40][41]. The difference in GC content between the two groups was viewed as further confirmation of this separation, as plant-associated Burkholderia have a lower GC content than the pathogenic BCC, while they also lack major virulence factors as well as several secretion systems [40]. However, it was only in 2014 that Sawana and colleagues performed a phylogenetic analysis based on 21 conserved protein sequences of 45 sequenced Burkholderia species and comparisons of conserved sequence indels, which they then used as the basis for their proposal to reclassify the environmental species (including the symbionts) into a new genus: Paraburkholderia [26]. Subsequently, additional environmental Burkholderia sensu lato species/strains have been accommodated into new genera delineated within the Burkholderiaceae family, such as Caballeronia, Trinickia, Mycetohabitans, Robbsia, and Pararobbsia [42][43][44][45]. Of these genera, only strains from Paraburkholderia and Trinickia contain proven N 2 -fixing symbiotic strains [44]. Since the discovery of beta-rhizobia, there has been a global effort to isolate, sequence, and characterize new rhizobial strains in order to not only trace the evolutionary history of this polyphyletic group but also to comprehend the molecular mechanisms driving this symbiosis [18,46]. A clear example is P. phymatum, for which more than a hundred strains have been isolated from all over the world and from various plant hosts since the discovery of the first strain (STM815 T ). At present, however, of these, only the strain STM815 T has a genome sequence available (Table 1) [47].  STM3622, STM3623, STM3631,  STM3665, STM3666, STM3667, STM3668,  STM3669, STM3674, STM3675, STM3676,  STM4205, STM4207, STM4208, STM4211,  STM4212, STM4214, STM4216, STM4217,  STM4219, STM4221, STM4223, STM4225,  STM4302, STM4303, STM4305, STM4308,  STM4312, STM4313, STM4316, STM4317,  STM4318, STM4319, STM4320, STM4322,  STM4323, STM4324, STM4325, STM4326,  STM4327, STM4328, STM4332, STM4333,  STM4334, STM4335, STM4337, STM4338,  STM4339, STM4342, STM4343, STM4344,  STM4345, STM4346, STM4337, STM6016,  STM6019, STM6023, STM6028, STM6025,  STM6027, STM6031, STM6017,  * Contradictory information: it was reported to be isolated from Machaerium lunatum, although it was never proven to nodulate it; at present, M. pudica is considered the plant host. 1 Identified as being of the same genotype as P. phymatum STM815 T on the basis of 16S rRNA. 2 Additionally identified by repetitive extragenic palindromic (REP)-PCR fingerprinting. 3 Additionally identified by whole protein profile. 4 Additionally identified by PCR-RFLP. 5 Additionally identified by recA PCR amplification.

Geographical Distribution and Phylogeny of Paraburkholderia Species
Tracking the distribution of Paraburkholderia symbionts can be useful to determine the extent to which they can act as nodulating symbionts of legumes as well as to gain a better understanding of the evolution of alpha-and beta-rhizobia. Paraburkholderia species have been isolated globally. However, more specifically, with regard to symbiotic species, they have been isolated from nodules of native Mimosa species in South (Brazil, Uruguay, French Guiana, and Venezuela) and Central (Costa Rica and Panama) America [31,32,35,55,56], while in South Africa, they have been found mostly in association with papilionoid legumes [36,[57][58][59][60][61][62][63]. Although a recent study has found the southern African indigenous mimosoid Vachellia karroo to be capable of forming nodules in association with Paraburkholderia strains [64], several of these strains are recognized, at present, as members of P. tuberum sensu stricto [65]. Additionally, Paraburkholderia have also been isolated from invasive legume nodules in China, Taiwan, India, Australia, and New Zealand [32,50,52,60,66,67]. However, the only confirmed centers of symbiotic diversification of native legume-nodulating Paraburkholderia species are in South and Central America (e.g., the Cerrado and Caatinga biomes in Brazil) and in the southern tip of South Africa (the CCR) [13,18,20,29,59,[61][62][63][68][69][70][71][72]. Paraburkholderia-Mimosa symbioses exhibit some level of specificity; yet, competition studies between Paraburkholderia, C. taiwanensis, and Rhizobium species using different plant hosts and nitrogen-free conditions revealed that Paraburkholderia are the most dominant species, not due to the plant preference, but rather due to physical environmental factors [29,68]. One of the most important factors appears to be soil pH, as acidic soils are often associated with low nutrient availability, whereas neutral-and higher-pH soils are often linked to higher fertility. In low-pH Brazilian soils, Paraburkholderia are the predominant symbionts of the Mimosa species [69,73,74], as well as species in some other mimosoid genera, such as the Piptadenia group and Calliandra [48,75,76]. This contrasts with other centers of diversity of these legumes, such as Mexico, in which the native Mimosa and Calliandra species preferentially nodulate with Rhizobium/Sinorhizobium in the largely alkaline soils [76,77], and in India where the native Mimosa species (M. hamata, M. himalyana) nodulate with Sinorhizobium (Ensifer) and do not interact with the beta-rhizobial symbionts of the invasive M. pudica [52]. Paraburkholderia are similarly absent as symbionts of endemic Mimosa spp. in the high-pH soils of the metal-rich highland region of Minas in Uruguay; however, in this instance, they are nodulated with diverse species of Cupriavidus rather than with alpha-rhizobia [78]. Geology and altitude, particularly in relation to humidity and temperature, have also been shown to be ecological drivers determining the distribution of nodulating Paraburkholderia species in both of their major centers of diversity [29,58,61]. In fact, in several studies, it has been shown that the diversity of Paraburkholderia species associated with Mimosa species in central Brazil and CCR papilionoid legumes in South Africa changes with elevation and is, therefore, associated with environmental conditions rather than the ability of Paraburkholderia symbionts to adapt to infertile acidic soils [29,57,59,61].
Several Mimosa-associated strains were originally tentatively placed within P. tuberum on the basis of 16S rRNA and recA phylogenies; these strains still separated according to host/geography on the basis of the symbiotic loci, which led to the definition of two symbiovars, sv. papilionideae and sv. mimosae [48]. A symbiovar is characterized by a specific subset of symbiotic loci that enables a set of strains (that could be members of multiple rhizobial species) to effectively interact with a specific host [96]. Recent investigations of these Mimosa-nodulating P. tuberum strains have, however, delineated these into the new species P. youngii and added several strains to the species P. atlantica [89], with the strains recognized, at present, as part of P. tuberum sensu stricto, all originating from South Africa only [65,89].
The presence of a deeply branched lineage of nodulating genes in Beta-proteobacteria suggests that beta-rhizobial nodulation genes have a long and independent evolutionary history [29,71]. In the case of the nif and fix genes, better congruence is found between their phylogenies and that of the 16S rRNA and housekeeping genes [18,29,68,81]. It is believed that the set of nif and fix genes in both mimosoid and papilionoid-nodulating Paraburkholderia have been acquired from diazotrophic Burkholderia, which they, in turn, might have received from alpha-rhizobia [18,58,68,81]. However, the gene encoding the transcriptional regulator NifA, appears to have a different origin, implying that multiple HGT events may have occurred [82]. Paraburkholderia rhizobia have evolved primarily due to the frequent lateral transfer of symbiosis genes within the Cape Fynbos biome, which gives the South African species a polyphyletic origin [71]. In contrast, the monophyletic character of South American Paraburkholderia indicates that their symbiotic genes might have been acquired simultaneously and from a different origin than those acquired by South African species [97], and probably less recently. Different biogeographic patterns and legume subfamilies on the two once-connected continents may have aided the divergence of mimosoid and papilionoid-nodulating Paraburkholderia. The origin of Fynbos Paraburkholderia in South Africa and the origin of South American mimosoid-nodulating Paraburkholderia in Cerrado and Caatinga overlapped during the Eocene, about 30 million years ago [62,98,99]. Paraburkholderia populations in Asia and Oceania are closely related to South American strains and were most probably brought over with their invasive Mimosa species (M. pudica, M. pigra, and M. diplotricha) [32,51]. The exceptions are the isolates from the South African native legume, Dipogon lignosus, which is invasive in New Zealand and Australia; strains isolated from invasive D. lignosus have been placed within the species P. dipogonis [67] and are related to the South African papilionoid-nodulating Paraburkholderia species, P. rhynchosiae and P. dilworthii [67]. According to Bontemps et al., the nodulation genes of the South American Paraburkholderia symbionts were acquired around the time when legumes began to diversify 65-50 million years ago, implying that the evolution of symbiotic Paraburkholderia in their South American center has an old and stable genetic history [18,29,100].

High Competitiveness of Paraburkholderia in the Rhizosphere
Rhizobia persist for a long time in three separate soil environments: bulk soil, the rhizosphere, and the plant root. Rhizobial tolerance to the harsh surroundings and rivalry with other species for survival are examples of the numerous pressures that soil bacteria must endure in these environments [101,102]. In addition, as plants restrict the number of nodules developed in their roots by a system called autoregulation of nodulation (AON), compatible rhizobia need to compete to form these nodules. Hence, competitiveness is commonly observed as a feature directly influencing the strain's relative fitness [7]. As previously mentioned, beta-rhizobia have consistently outcompeted alpha-rhizobia in nodulating Mimosa species. This dominance, which is also influenced by the host species and geography, could possibly be explained by the inherent higher competitiveness of betarhizobia [7,68]. In this section, both abiotic and biotic challenges that rhizobia encounter in the rhizosphere prior to root colonization are described, as well as the strategies and traits that Paraburkholderia employ to overcome them.

Abiotic Factors and Stress Responses
Environmental stresses, such as temperature, salinity, drought, pH, or the presence of toxic compounds, have a significant impact on rhizobial survival. In situations where these stresses are encountered, bacteria rely on stress tolerance mechanisms and the production of specific enzymes to remove damaging components [103,104]. Stress tolerance strategies include the influence of physical-chemical soil properties, the regulation of cell envelope composition, solute accumulation, and biofilm and exopolysaccharide (EPS) production [103].
Temperature is one of the factors that can affect the survival of rhizobia in soil, the molecular communication between host and symbiont, as well as the development of the nodule [103]. The ideal growth temperature for symbiotic Paraburkholderia species is 28 • C, while some species, namely, P. atlantica, P. franconis, and P. phymatum, tolerate temperatures as high as 37 • C [87,105]. This trait already gives them a distinct advantage over most alpha-rhizobia, whose optimal temperature ranges from 25 • C to 30 • C [103]. It is well-known that alpha-rhizobia exposed to a high-temperature stress respond by activating their survival program, which involves the production of EPS, chaperones, and heat-shock proteins [103,106].
Drought and desiccation hinder the uptake of water, removing most of the water from the cells and causing osmotic stress [103]. Compared to non-symbiotic bacteria, rhizobia are often more susceptible to salt and osmotic stress [107]. Paraburkholderia have, however, been shown in several studies to better tolerate this type of stress compared to alpha-rhizobia. In comparison to the highly stress-resistant alpha-rhizobium Rhizobium tropici CIAT899 [108], P. phymatum GR01, which was isolated from common bean in semi-arid soils in Morocco [49] ( Table 1), not only survives better in free-living conditions at higher salt concentrations and in a hyperosmotic medium, but also exhibits higher nodule competitiveness under these conditions. The higher survival rate of the Paraburkholderia strain could be attributed to the accumulation of the disaccharide trehalose, the sugar alcohol mannitol, and the amino acid alanine, which are known as osmoprotectants [109]. The most well-known response to desiccation by species of Paraburkholderia is the production of EPS, which includes the production of cepacian (CEP), the major EPS produced by the Burkholderiaceae family. It is synthesized by both symbiotic species, such as P. phymatum and P. caribensis, as well as endophytic species, such as P. phytofirmans or P. xenovorans [110,111]. In P. xenovorans LB400 T , CEP has been related to desiccation tolerance [110,112]. Other examples include Paraburkholderia isolates from the South African legume D. lignosus, which presented greater tolerance to water stress induced by different concentrations of polyethylene glycol (PEG) than Bradyrhizobium or Rhizobium species [60].
Another abiotic factor that negatively impacts rhizobial persistence and survival in the rhizosphere is acidity [48,69,75,107,113]. Interestingly, Paraburkholderia and Burkholderia species are found enriched in acidic soils [48,69,75,114]. Organic acids secreted by plant roots, such as oxalate or citrate, contribute to the acidification of the environment and, at the same time, can be utilized by bacteria as carbon sources. Therefore, a link between their pH tolerance and their ability to grow using oxalate as a carbon source has been suggested [114][115][116].
Heavy metals are released into the environment with the use of pesticides, fertilizers, herbicides, insecticides, or fungicides, along with other human practices that can harm beneficial rhizobial communities [107]. Copper, nickel, cadmium, zinc, chromium, and lead are the most common contaminants [117]. Rhizobia tolerate these in a variety of ways, e.g., by producing EPS and biopolymers, but also by changing the properties of these compounds, thereby converting them into less toxic forms [103]. Carbonyl, carboxyl, and hydroxyl groups in the EPS matrix of P. xenovorans have been shown to scavenge and complex metals, such as iron and zinc [110]. Recently, Soares Neto et al., 2022, showed that P. atlantica is the dominant symbiont in nodules of several Mimosa species grown in heavy metal-contaminated soils, suggesting that symbiosis with Paraburkholderia might play a significant role in healing soils containing excesses of toxic metals [74].

Competition between Soil Bacteria for Survival and Root Nodulation
Bacterial diversity is lower in the rhizosphere than in bulk soil; however, the rhizospheric population is nonetheless more active due to competition between its constituents and the consumption of plant root exudates (REs) [118]. On a local scale, competition reduces biodiversity while increasing ecological stability. Soil bacteria have developed a variety of competitive mechanisms in order to prevent rival strains from colonizing a common niche. These strategies can be categorized into two types: the first is exploitative or passive competition, which involves rapid growth to consume the available resources resulting in metabolic alterations that profit the original strain and that can be exploited against competitors, while the second is interference or direct competition that involves direct antagonism to exclude competitors [119][120][121][122].

Exploitative or Passive Competition
Basic cellular functions, such as the assimilation of nutrients and energy sources, surface structure formation, signaling, and proliferation, can influence rhizobial nodulation competitiveness [123]. Bacteria with high metabolic adaptability are more likely to survive as they can better metabolize a wider range of nutrients and plant-secreted compounds [122]. REs provide a source of carbon and energy in the rhizosphere, and bacteria able to assimilate these compounds can change the environment through a variety of metabolic processes [119,124]. Members of the Burkholderiaceae have been shown to be metabolically versatile, which allows them to persist in a wide range of environments and confers a competitive advantage [7,68,125]. Oxalotrophy, a trait found in several symbiotic Paraburkholderia species, such as P. phymatum STM815 T , P. phenoliruptrix LMG 22037 T , P. caribensis LMG 18531 T , and P. tuberum STM678 T , has been recognized as an example of metabolic diversity in the Paraburkholderia group. It contributes to the effective colonization of lupin and maize by P. phytofirmans LMG 22487 T , for instance, by allowing the bacterium to utilize oxalate, one of the main RE components in soil-grown plants that can be toxic to strains that cannot metabolize it [116]. The production of secondary metabolites that improve direct nutrient consumption also contributes to exploitative competition. An example is the production of siderophores, which has been shown in Sinorhizobium meliloti Rm5000 to inhibit competitor strains through the siderophore sequestering of environmental iron [121,126,127]. Previous transcriptome studies looking at the responses of two beta-rhizobia (P. phymatum STM815 T and C. taiwanensis LMG 19424 T ) and one alpha-rhizobium (Rhizobium mesoamericanum STM3625) to M. pudica REs showed that only P. phymatum up-regulates the expression of a putative siderophore (Bphy_4034-4047), hinting at an important role of this siderophore in the first steps of the symbiotic interaction [128].

Interference or Active Competition
Interference competition, also referred to as active competition, employs antagonistic mechanisms to inhibit competitors [119]. It includes increasing access to space through enhancing motility, disrupting the competitiveness of other bacteria by inhibiting their quorum sensing (QS) systems, and eliminating competitors through the production of antibiotics or secreted toxins [120].
As demonstrated by the pilA1 integral pilin subunit mutant strains in S. meliloti Sm1021 or the chemotaxis receptors mcpB and mcpC mutant in Rhizobium leguminosarum bv. viciae VF39SM, motility and chemotaxis can affect the nodulation efficiency in alpharhizobia and lead to a significant decrease in nodulation competitiveness [129,130]. Within the Paraburkholderia genus, it has been demonstrated that enhanced motility and EPS formation may be features that give P. phymatum STM815 T a competitive advantage in nodulation [7,68,125]. In this species, both flagella biosynthesis and chemotaxis have been shown to be positively controlled by the alternative RNA polymerase sigma factor RpoN or σ 54 , which regulates nitrogen utilization systems under nitrogen-deplete conditions [131]. During inter-bacterial competition experiments between beta-rhizobial strains, being one of the fastest-growing strains may have contributed towards the higher competitiveness of P. phymatum STM815 T in nodulating common bean, siratro, and cowpea when compared to four other Paraburkholderia strains (P. tuberum STM678 T , P. mimosarum LMG 23256 T , P. diazotrophica LMG 26031 T , and P. sabiae LMG 24235 T ) and one Trinickia strain (T. symbiotica LMG 26032 T ) [125]. Bacterial cell motility is also regulated by QS, which is usually based on N-acylhomoserine lactone (AHL) molecules and contributes to balancing the density of rhizobial relocation into the nodule in relation to proliferation in the root zone [123,132]. The P. phymatum strains, STM815 T and GR01, possess the BraI/R QS system, which, although not essential for the nodulation of common bean or M. pudica, may provide a competitive advantage by utilizing AHLs produced by neighboring bacteria [133]. The BraI/R system regulates EPS production, which may provide a competitive advantage to P. phymatum. Indeed, the P. phymatum EPS CEP has been proposed to contribute to nodule competitiveness in planta, as a CEP mutant was impaired in attaching to common bean roots and formed less nodules compared to the wild-type and complemented strains [111].
The eradication of competitors can also occur through the production of antimicrobial compounds that can either be released into the environment (contact-independent systems) or directly injected into competitors by contact between cells (contact-dependent systems). Indeed, rhizobia can produce and secrete secondary metabolites, such as antimicrobial peptides or antibiotics, which can inhibit growth or actively kill other bacteria [119]. While bacteriocins are peptidic toxins, antibiotics are secondary metabolites that are byproducts of metabolic pathways with a non-ribosomal origin [134]. Bacteriocins have a narrow spectrum of activity; therefore, they usually target closely related species without harming the producing cell [135][136][137]. In the Paraburkholderia genus, genes potentially coding for bacteriocins have been found in the newly identified nitrogen-fixing P. lycopersici Tne-862 T and in several endophytic species, such as P. phytofirmans and P. xenovorans [138,139]. Conversely, the Burkholderiaceae family is known to produce secondary metabolites with antimicrobial activity. The most highly investigated species and strains in this regard belong to Burkholderia sensu stricto (that includes the BCC and B. pseudomallei groups), which suggest that newly derived genera, such as Paraburkholderia, may have an undiscovered potential for novel antimicrobial production [140]. To date, bioinformatic analyses have found putative genes coding for antibiotics; however, only the production of sulfazecin (a monobactam antibiotic) in Paraburkholderia acidophila has been demonstrated [141,142]. Contact-dependent mechanisms usually employ secretion systems to deliver toxins by cellto-cell contact into the target cell. One of the best-known contact-dependent mechanisms employs the type 5b or two-partner secretion system (T5SSb or TPSS), also known as the contact-dependent growth inhibition (CDI) system [143][144][145][146]. In Burkholderia dolosa, a new class of CDI was described that involves a gene cluster called bcpAIOB, where BcpA is an adhesin with the toxin encoded in the C-terminal domain, BcpI is the immunity protein that can form a tight complex with BcpA, BcpB is a beta-barrel outer-membrane protein that secretes the toxin, while the protein BcpO still has an unknown function [143,144,[146][147][148]. This gene cluster is present in several Paraburkholderia species, including P. xenovorans and P. rhizoxinica, and in the symbiotic P. phymatum [144]; however, nothing is known about its role in symbiotic N 2 -fixing rhizobia. The other secretion system that plays a role in inter-and intra-bacterial competition is the type 6 secretion system (T6SS). The T6SS possesses some elements that are similar to a phage tail spike. The genes encoding the core structure are annotated as "tss" (from type-six secretion) and code for the baseplate (TssAEFGK), the membrane complex (TssJLM), a tail tube (TssD or Hcp), a tail tip (TssI or VgrG), and a contractile sheath (TssBC). Similar to the phage spikes, T6SS penetrates the target cells delivering multiple effector proteins [149]. The T6SS is frequently present (in nearly 25%) in Gram-negative bacteria and was originally discovered in the alpha-rhizobial symbiont R. leguminosarum bv. trifolii, where it prevented the nodulation of the non-host legume pea [150][151][152]. Two complete T6SS clusters have been identified in the genome of the symbiotic P. phymatum, and both have been shown to be partially responsible for the high competitiveness of P. phymatum when nodulating cowpea roots in the presence of other beta-rhizobia [153]. The so-called T6SS-3 is more akin to that present in pathogenic Burkholderia species and can be activated at 37 • C [105]. The second T6SS cluster in P. phymatum, T6SS-b, is mainly present in soil bacteria and is positively controlled by the sigma factor RpoN under N-limiting conditions [105,131]. The expression of T6SS-b is induced by citrate (a metabolite found in common bean RE), temperatures of 20/28 • C, and is in contact with germinated common bean seeds, suggesting a possible function during early symbiosis [105].

Interaction with the Plant Host
A successful rhizobium-legume symbiosis requires the initiation of the interaction between free-living bacteria and a plant host before the rhizobia can nodulate the legume and perform BNF inside the root nodules. Through specific signaling molecules, compatible symbionts and host legumes detect and recognize each other. This interaction then triggers nodule organogenesis, the entry of the rhizobia into the plant, and the eventual accommodation of the rhizobia within the nodule, which require changes in gene expression and metabolism in both partners [154][155][156][157][158]. As beta-rhizobia have only been discovered relatively recently, there is still little information about the molecular mechanisms underlying beta-rhizobial symbiosis. The following section covers recent discoveries in this area, unraveling the genetic network used by nitrogen-fixing Paraburkholderia to interact with a plant host.

"Opening the Gate": Rhizobium-Legume Recognition
Rhizobia have two different routes by which they infect plants. The less common route is crack entry, where bacteria occupy the intercellular spaces between epidermal and cortical cells [13,159]. The second route that is more frequently encountered involves the infection of root hairs and is mediated by rhizobial nodulation factors (NFs) [160]. The transcriptional regulator, NodD, a member of the LysR family of transcriptional activators, is activated by specific plant flavonoids and controls the expression of nod genes, the structural genes that code for the NFs [156,161]. On legume hosts, they function as specialized morphogenic signal molecules that induce early responses, such as root hair deformation, calcium-spiking, and the division of cortical cells, eventually resulting in effective nodules [79,161]. NFs are composed of a chitin oligomer backbone of two to six units of β-(1,4)-linked N-acetyl-D-glucosaminyl residues that are N-acetylated on their non-reducing ends [162]. The nod genes are divided into two sets: "common" genes (nodA, nodB, nodC, nodI, and nodJ) present in all rhizobial species that use the NF for host communication, and "species-specific" genes. The "common" genes synthesize the core structure of the NFs: nodA encodes an acyl transferase, nodB codes for a deacetylase, and nodC encodes a N-acetylglucosaminyl transferase ( Figure 1, Table 2) [163]. Additionally, present in all rhizobia, the nodIJ genes encode an ABC transporter involved in the secretion of NFs. The set of "species-specific" genes are noe, nol, and additional nod genes. Their products are responsible for producing a wide range of modifications to the NF core by introducing different residues, for example, N-methyl by NodS, sulfate groups by NodH, or a O-carbamoyl by NodU (Figure 1, Table 2) [161,164,165]. As a result of these changes, rhizobia can increase the diversity of NFs, a trait that has been considered a sign of rhizobia adjusting to different host plants [166]. Alpha-and beta-rhizobia share the nod core genes; however, they differ in the presence of the species-specific genes. The beta-rhizobial nod genes are induced by RE, as has been shown for P. phymatum and C. taiwanensis with M. pudica REs [128]. In this study, P. phymatum, which is able to nodulate over 50 legume species, showed a higher induction of nod genes expression than C. taiwanensis and R. mesoamericanum [47,49,125,128]. The Paraburkholderia species nodulating mimosoids often have a single copy of each nod gene arranged in the operon nodBCIJHASU [82,167] ( Figure 1). The nodUSDABC genes in papilionoid legume-nodulating Paraburkholderia species are separated from nodIJ by the insertion of nif genes and transposases, and some nod genes are duplicated, such as nodB and nodC [82,167] (Figure 1). They also seem to lack nodH and possess nolO, suggesting that their NFs are not sulfated, but have carbamoyl groups. Common to both Paraburkholderia groups is the presence of nodSU, which indicates that methyl and carbamoyl groups are potentially added to the NF core [82,167]. In addition, the gene noeM (coding for a fatty acid hydrolase) has been proposed as a nodulation gene specific for rhizobia nodulating M. pudica, as it is present in all strains nodulating M. pudica (Ensifer/Sinorhizobium, Mesorhizobium, Neorhizobium, and Rhizobium in alpharhizobia and Paraburkholderia, Trinickia, and Cupriavidus in beta-rhizobia) [168].
To colonize the root, rhizobia form microcolonies or biofilms on the roots. Within the curled root hairs, rhizobial cells migrate through infection threads (ITs) toward the nodule primordia, where the rhizobia are released and differentiate into N 2 -fixing bacteroids [118,154,155,[169][170][171]. The main components of the root biofilm are EPS [172,173]. These molecules bind to the host EPS protein receptor 3 (EPR3), thereby inactivating the defense signaling pathway [174,175]. In beta-rhizobia, the role of EPS in the induction of root hair curling and the formation of ITs has not yet been studied. Recently, in P. phymatum STM815 T , the bceN gene that encodes a protein with GDP-D-mannose 4,6-dehydratase enzyme activity was shown to be involved in the synthesis of the polysaccharide CEP, which has been shown to play a role in the first steps of establishing a symbiosis with common bean (see above) [111]. It differs from the typical EPS found in alpha-rhizobia that is encoded by the exo/exs and wge/wga gene clusters, both of which are absent in the genomes of Paraburkholderia strains [111]. Table 2. Function of nod gene products in the biosynthesis of the Nod factors (NFs) produced by rhizobia, found in Paraburkholderia genomes from Figure 1.   [176]. P. phymatum STM815 and P. tuberum WSM4176 genomes are used as query sequences.  [176]. P. phymatum STM815 and P. tuberum WSM4176 genomes are used as query sequences.

The Steps towards Nodule Development
The recognition of NF-NF receptors (NFRs) (a transmembrane LysM-type serine/ threonine receptor kinase) activates calcium oscillation in the plant cell nucleus, which activates gene expression that allows the initiation of the bacterial infection at the epidermis and promotes plant cell division in the cortex [155,177]. Nodule organogenesis requires hormonal signaling, particularly auxin and cytokinin, which are important for plant growth, development, embryogenesis, and tropism [155,[178][179][180].
Rhizobia can modulate the activity of these plant hormones either via production or degradation during the colonization process. Several phenomena related to the production of plant hormones by microorganisms have been reported within the alpha-rhizobia. Indole-3-acetic acid (IAA) is the form of auxin most studied [181]. Bacteria can synthesize IAA in a tryptophan-dependent or independent manner [182]. The beta-rhizobial symbiont P. phymatum STM815 T has recently been shown to produce IAA via the indole-acetamide (IAM) pathway [183]. The genes involved in this pathway (iaaMH) are up-regulated in the presence of M. pudica RE and common bean nodules compared to free-living conditions, suggesting a role of iaaMH in the early steps of symbiosis [128,184]. Rhizobia can also promote nodulation by decreasing ethylene levels in the plant root nodule with the enzyme 1-aminocyclopropane-1-carboxylate (ACC) deaminase. The ACC deaminase, encoded by the gene acdS in bacteria, catalyzes the degradation of ACC, an immediate precursor of ethylene, into ammonium and α-ketobutyrate [180]. In addition, rhizobitoxin (RTX) is a phytotoxin that acts as an inhibitor of the plant ACC synthase, involved in the biosynthesis of ethylene in host roots, leading to a reduction in ACC production and decreased plant ethylene levels. The dihydrorhizobiotoxin desaturase enzyme, encoded by the rtxC gene, is involved in the final step of RTX synthesis [185]. The acdS gene has been found in genomes of both alpha-and beta-rhizobia harboring nodC, being present in 100% of the Paraburkholderia and Bradyrhizobium genomes studied [186]. In contrast, rtxC was found less frequently in Paraburkholderia (33.3%) and Bradyrhizobium (45.1%) [186]. In P. phymatum STM815 T , acdS and the whole RTX coding operon were up-regulated in the presence of M. pudica RE [128]. Hence, possessing both mechanisms to reduce ethylene levels could potentiate the nodulation ability of a strain.
It would appear as if certain rhizobial enzymes involved in carbon metabolism play an important role during symbiosis. Mutants of P. phymatum for phosphoglycerate mutase (Bphy_0266), an enzyme important for glycolysis and gluconeogenesis, and fructose 1,6biphosphatase (Bphy_0685), a regulatory enzyme in gluconeogenesis, are not able to induce ITs or nodules, although they still possess the ability to cause root hair deformation [187]. Some of the possible explanations for this observation are that either the accumulation of intermediates would interfere during nodulation, or that growth by gluconeogenesis is indispensable for P. phymatum during the early steps of nodulation [187].

"Achieving the Goal": Differentiation to BNF-Performing Bacteroids
Once rhizobia are released from the ITs inside the cortical cells, they grow and divide infecting the plant cells. Within the plant cell, bacterial cells are enclosed in a newly formed compartment surrounded by the plant-derived symbiosome membrane (also called the peribacteroid membrane, originating from the plasma membrane of the infected plant cell). The symbiosome membrane together with the enclosed bacteroids is referred to as the symbiosome [13,157,188,189]. In the nodule, bacteria differentiate into N 2 -fixing bacteroids. At this stage of differentiation, the genes involved in N 2 -fixation, the so-called nif genes, are found to be up-regulated. These genes are often clustered on symbiotic plasmids or in the chromosome as genomic islands [171]. The nif genes are involved in the synthesis of the nitrogenase enzyme that catalyzes, with a high energetic cost, the conversion of N 2 to biologically useful ammonia, following the reaction [157]: In most of the cases, BNF is conducted by molybdenum (Mo) nitrogenases; although alternative nitrogenases (vanadium or iron-only) exist in nature [190], they have never been found in rhizobia. The transcriptional regulator NifA activates the expression of the nif genes. NifA belongs to the enhancer binding protein (EBP) family and works together with the sigma factor RpoN. As proteins belonging to the EBP family, NifA has a central conserved AAA+ ATPase domain that interacts with RpoN, flanked by an amino-terminal regulatory and a carboxy-terminal DNA-binding domains. The DNA-binding domain recognizes and binds to an upstream activator sequence (UAS) (TGT-N 10 -ACA) roughly 100 bp upstream of the transcription start site (TSS). RpoN associates reversibly with the core RNA polymerase that recognizes promoter consensus sequences at −12 and −24 bp relative to the TSS and requires a DNA loop to interact with NifA [5,191,192]. P. phymatum mutants in nifA and rpoN are impaired in fixing N 2 [184]. During symbiosis with P. vulgaris, RpoN not only activates the expression of nif genes, but it also regulates the hydrogenase cluster, respiration-related genes (see below), and genes involved in ammonium transport and nitrogen metabolism [193]. In Bellés-Sancho et al., 2021, a combined metabolomics and dual RNA-sequencing (RNA-seq) approach on nodules was used to unravel the role of P. phymatum NifA during symbiosis with P. vulgaris [194]. In addition to the common response of nitrogen deprivation from non-fixing bacteroids, NifA was found to regulate other important symbiotic characteristics, such as stress response, transport of C 4 -dicarboxylates, and auxin biosynthesis, all of which were validated by growing P. phymatum in vitro. By using dual RNA-seq and therefore looking at the transcript profile of both partners simultaneously, the nifA mutant was shown to also induce significant changes in common bean, such as an increase in brassinosteroids and the down-regulation of genes involved in AON. These results are in line with the previously reported higher number of nodules that clustered together in common beans inoculated by a nifA mutant [184,194]. In the follow-up study, the role of NifA as a repressor of bacterial auxin biosynthesis was demonstrated by mutant analysis and the construction of reporter constructs. The P. phymatum NifA represses the expression of the iaaMH genes in all stages of symbiosis. Moreover, in the absence of NifA, the overproduction of IAA and IAM induced the clustering phenotype and contributed to hypernodulation, suggesting a role of bacterial auxin in plant colonization. In P. phymatum, NifA also controls the fixABCX genes [194], which are present in both alphaand beta-rhizobia and are involved in the transfer of electrons to nitrogenase. The gene products FixA and FixB are homologous to beta and alpha subunits of mammalian electrontransfer flavoproteins (ETFs), whereas FixC and FixX are homologous to ETF cognate acceptor ETF-quinone oxidase (ETF-QO), which usually transfers electrons to terminal electron acceptors [82,195].
Rhizobia are aerobic bacteria that require oxygen to respire; however, oxygen can permanently damage the nitrogenase enzyme if its active site becomes oxidized. N 2 -fixation in nodules must, therefore, occur in a microaerophilic environment, and bacteria adopt several strategies to create this low-oxygen environment [196]. Under these circumstances, rhizobia have been shown to use a high-affinity cytochrome cbb 3 oxidase encoded by the fixNOQP and fixGHIS genes, which functions as a terminal oxidase at very low oxygen concentrations [9,167]. Interestingly, no genes coding for this cytochrome oxidase have been detected in beta-rhizobial genomes [82,184]. Instead, an RNA-seq analysis of P. phymatum-common bean nodules and from P. phymatum grown in microaerobic conditions showed a strong induction in the expression of genes coding for a potential cytochrome o ubiquinol oxidase (cyo). Indeed, their disruption led to impaired nitrogenase activity, consistent with the role of this oxidase in functional nodules [184]. The nif genes of P. phymatum are strongly up-regulated in microoxic conditions compared to aerobic conditions, but are not activated by nitrogen limitation [184]. Nonetheless, in free-living conditions, even between Paraburkholderia diazotrophic species there is a difference in N 2 -fixing activity, as upon the presence of some fixed nitrogen in the growth media, symbiotic strains fix less than non-symbiotic species, such as Burkholderia vietnamiensis TVV75, P. tropica, and P. unamae [35]. These results suggest that there are differences in the nitrogen-fixation regulation between Paraburkholderia strains that are free-living and those that are symbiotic diazotrophs [18,35]. Yet, the ability of Paraburkholderia symbionts to fix nitrogen ex planta already marks a fundamental distinction from alpha-rhizobia. One of the possible reasons is the presence of the nifV gene in diazotrophic Paraburkholderia. Indeed, this gene is not present in the majority of alpha-rhizobial genomes, except for a few Bradyrhizobium species that possess the ability to fix N 2 ex planta [82,197,198]. The nifV gene encodes a homocitrate synthase that catalyzes the condensation of acetyl-coenzyme A (acetyl-CoA) and α-ketoglutarate to form R-homocitrate that is needed for the FeMo-cofactor of the nitrogenase enzyme. When an alpha-rhizobium establishes a symbiosis with a host legume, the plant produces homocitrate using the FEN1 homocitrate synthase and supplies it to the bacteroid [9,[197][198][199]. This dependency on the plant is a strategy used by plants to control the N 2 -fixing activity of the bacteroids [171,200]. In P. phymatum, nifV is expressed in symbiosis with papilionoid and mimosoid legumes and also in free-living conditions and when homocitrate is abundant in all plant hosts inoculated with P. phymatum. However, NifV was not essential during symbiosis with its original host M. pudica, suggesting that the plant may assist in providing homocitrate. General transcriptome studies comparing P. phymatum symbiotic with free-living gene expression profiles have shown an over-representation of genes related to "inorganic ion transport" and "energy production and conversion" up-regulation during symbiosis with common bean [184]. Nonetheless, the opposite trend is observed in alpha-rhizobial bacteroids that down-regulate genes involved in these two functional categories [200]. Plants provide bacteroids with carbon and energy sources in the form of C 4 -dicarboxylates, primarily malate and succinate. Malate is carried across the symbiosome membrane and is imported into the bacteroid by the dicarboxylic acid transporter, DctA, which is controlled by the DctB/C sensor-regulator system. The tricarboxylic (TCA) cycle is then used to metabolize malate, producing reduced electron carriers, such as NADH and FADH 2 , required for ATP production during N 2 -fixation [9,200,201]. Acetyl-CoA from the TCA cycle is metabolized to polyhydroxybutyrate (PHB) in lowoxygen conditions, where it is employed as carbon storage alongside glycogen and lipids, and to maintain redox balance [171,195]. In analogy to alpha-rhizobia, the PHB-associated protein and PHB synthesis genes are activated in P. phymatum during nitrogen shortage, whereas PHB depolymerase and genes that may be implicated in PHB stability are activated during symbiosis with P. vulgaris [184]. In alpha-rhizobia, the developmental program in bacteroids prevents rhizobia from assimilating ammonia fixed from the BNF, thereby giving all their reduced nitrogen to the plant [157]; however, it has not been proven in betarhizobial symbionts. Ammonia is transported through the symbiosome membrane from the bacteroids by simple diffusion, unspecific protein channels, channels for NH 3 or cations, or aquaporin-like channels [157]. Then, in plant cells, ammonia is assimilated as amino acids, mainly glutamine and asparagine, or ureides [9,157]. In alpha-rhizobia, additional pathways have been characterized and their importance in symbiosis has been proven. For instance, in the photosynthetic Bradyrhizobium sp. ORS278, the genes ccbL1/cbbS1 encoding a ribulose 1,5 bis-phosphate carboxylase oxygenase (or RuBisCO) have been identified and their role in the carbon fixation pathway confirmed. Gourion and coworkers demonstrated that RuBisCO is necessary for an efficient symbiosis between Bradyrhizobium sp. ORS278 and Aeschynomene indica [202]. Interestingly, P. phymatum STM815 T also carries ccbL1 (Bphy_6497, actually annotated as rbcL) on plasmid pBPHY01, and this gene is upregulated during symbiosis compared to free-living conditions [184]. However, whether Bphy_6497 is important for a functional symbiosis is not known. The NiFe hydrogenase that recycles part of the energy used in the N 2 -fixation process by oxidizing the H 2 released during BNF is up-regulated in common bean nodules induced by P. phymatum and has been shown to improve BNF [184,203]. An isocitrate lyase, involved in the glyoxylate shunt pathway, was also up-regulated in P. phymatum bacteroids [184]. Similar to alpha-rhizobial host legumes [204,205], specific metabolic requirements are needed depending on the plant hosts colonized by P. phymatum. Several host-specific intermediates belonging to distinct metabolic pathways accumulated in the P. phymatum symbioses with common bean, cowpea, siratro, and M. pudica; for example, intermediates from the biosynthesis of valine, leucine, and isoleucine were found in common bean nodules, metabolites involved in pyrimidine biosynthesis in cowpea nodules, three C 4 -dicarboxylic acids (fumarate, malate, and tartaric acid) in siratro nodules, and lastly metabolites involved in flavonoid and isoflavonoid biosynthesis in M. pudica nodules [206].

Concluding Remarks
Since the identification of Paraburkholderia strains as N 2 -fixing symbionts twenty-two years ago, the discovery of new rhizobial strains has not receded, as there is an ongoing international effort to isolate and characterize rhizobia from tropical soils and indigenous hosts, which might not have obvious commercial value. This practice stems not only from research groups attempting to understand the evolutionary path of rhizobia, but also from the industrial sector seeking to identify elite microorganisms capable of competing in the soil, surviving harsh environmental conditions, and fixing N 2 , with the final aim of employing them as bioinoculants. Several studies have focused on achieving a better understanding of the molecular mechanisms involved in the interaction between host legumes and rhizobia, with a particular emphasis on the common and distinct mechanisms used by alpha-and beta-rhizobia. It is interesting to note that the Paraburkholderia beta-rhizobial strain with the most information available is P. phymatum, as it has been continuously used as the main model of beta-rhizobial interactions with plants. This strain is outstanding for its ability to outcompete rival strains for nodulation of several crops [7,68,125,153], to survive in harsh soil conditions [109], and for its exceptionally broad legume host range (over 50 legume species) [35,47,125], all qualities needed to compete with native soil rhizobia. Moreover, P. phymatum possesses mechanisms employed by plant growth-promoting rhizobia (PGPR) to promote plant growth, such as the production of auxin [183,194] and free-living N 2 -fixation [183,184,194,206]; it also possesses the potential for ACC deaminase activity and iron uptake since putative genes related to both activities are present in its genome [128]. These additional features of the plant growth-promotion potential of P. phymatum, such as phosphate solubilization or siderophore production, could also support its application as a bioinoculant in agriculture. In this discovery process, functional genomic approaches have greatly contributed towards a better understanding of the molecular mechanisms underlying each step leading to the establishment of a successful beta-rhizobial symbiosis. However, even with this significant progress made at the genomic level, the importance of the identified genes remains to be characterized phenotypically. Creating mutant strains lacking or overexpressing these genes would be the preferred approach to demonstrate their role during symbiosis, or for the symbiotic condition being investigated. Since the ultimate goal is to apply these strains in the soil as agricultural inoculants, the following step would be to translate this rhizobial research to the field, where different environmental settings/stresses could be applied to discover to what extent each Paraburkholderia strain can contribute towards plant growth promotion. Indeed, in the case of P. phymatum, even if it is not used as a plant inoculant in the future, with a greater understanding of its competitiveness and N 2 -fixing mechanisms, the genetic traits involved in these phenotypes could still be used to genetically modify and create synthetic inoculant strains aimed at improving crop yield.