RESEARCHOpen Access icon OPENOpen Access license

A Promiscuity Locus Confers Lotus burttii Nodulation with Rhizobia from Five Different Genera

    Authors and Affiliations
    • Mohammad Zarrabian1
    • Jesús Montiel1 2
    • Niels Sandal1
    • Shaun Ferguson1
    • Haojie Jin1
    • Yen-Yu Lin3
    • Verena Klingl3
    • Macarena Marín3
    • Euan K. James4
    • Martin Parniske3
    • Jens Stougaard1
    • Stig U. Andersen1
    1. 1Department of Molecular Biology and Genetics, Aarhus University, Denmark
    2. 2Center for Genomic Sciences, National Autonomous University of Mexico. Cuernavaca, Mexico
    3. 3Faculty of Biology, University of Munich, Großhaderner Straße 2-4, 82152, Planegg-Martinsried, Germany
    4. 4The James Hutton Institute, Invergowrie, Dundee DD2 5DA, U.K.

    Published Online:


    Legumes acquire access to atmospheric nitrogen through nitrogen fixation by rhizobia in root nodules. Rhizobia are soil-dwelling bacteria and there is a tremendous diversity of rhizobial species in different habitats. From the legume perspective, host range is a compromise between the ability to colonize new habitats, in which the preferred symbiotic partner may be absent, and guarding against infection by suboptimal nitrogen fixers. Here, we investigate natural variation in rhizobial host range across Lotus species. We find that Lotus burttii is considerably more promiscuous than Lotus japonicus, represented by the Gifu accession, in its interactions with rhizobia. This promiscuity allows Lotus burttii to form nodules with Mesorhizobium, Rhizobium, Sinorhizobium, Bradyrhizobium, and Allorhizobium species that represent five distinct genera. Using recombinant inbred lines, we have mapped the Gifu/burttii promiscuity quantitative trait loci (QTL) to the same genetic locus regardless of rhizobial genus, suggesting a general genetic mechanism for symbiont-range expansion. The Gifu/burttii QTL now provides an opportunity for genetic and mechanistic understanding of promiscuous legume-rhizobia interactions.

    Copyright © 2022 The Author(s). This is an open access article distributed under the CC BY 4.0 International license.

    Symbiosis with nitrogen‐fixing rhizobia in root nodules enables legumes to access atmospheric nitrogen. In most cases, rhizobial entry into root cells requires recognition of rhizobial Nod factors (NFs) that are signaling molecules secreted in response to perception of plant flavonoids (Oldroyd 2013). In turn, host membrane-bound NF receptors initiate downstream signal transduction pathways initiating rhizobial infection and nodule organogenesis (Madsen et al. 2003; Murakami et al. 2018; Radutoiu et al. 2003). Plants produce complex mixtures of flavonoids (Liu and Murray 2016). Likewise, rhizobia secrete many different NF species (D'Haeze and Holsters 2002) and both flavonoid and NF pools may change dynamically over time within the same plant accession or rhizobia strain (Ferguson et al. 2020; Kelly et al. 2018; Liu and Murray 2016). This provides an intricate system that, along with bacterial effectors, exopolysaccharides (EPS), and the corresponding plant detection systems, allows fine-tuning of legume-rhizobium compatibility (Kawaharada et al. 2015; Kusakabe et al. 2020; Yang et al. 2010). Successful and efficient nitrogen fixation requires full compatibility of symbiotic partners. Symbiotic compatibility can affect early stages of infection, determining whether or not nodules are formed, or later stages of nodule development, affecting nitrogen-fixing efficiency or nodule senescence (Perret et al. 2000; Wang et al. 2012; Yang et al. 2017).

    Intercellular infection is thought to represent a more ancient and less-advanced infection mode than intracellular infection through root hairs (Sprent 2007). Specific legumes are typically infected either intra- or intercellularly. However, at least some legumes maintain genetic programs for both types of infection. These include Sesbania rostrata, in which the infection mode can change in response to flooding (Herder et al. 2006), and Lotus species, in which intercellular infection appears to serve as a backup function to the preferred intracellular infection route through root-hair infection threads. This phenomenon was observed in Lotus japonicus Gifu as rare infection events of spontaneous nodules in a NF receptor–deficient genetic background (Madsen et al. 2010). It has since been found in the L. japonicus Gifu interaction with Agrobacterium pusense (formerly Rhizobium sp. strain) IRBG74 (Montiel et al. 2020), and in Lotus burttii interactions with Sinorhizobium fredii HH103 and Rhizobium leguminosarum Norway (Acosta-Jurado et al. 2016b; Liang et al. 2019). Generally, intracellular root-hair infection appears to offer more stringent scrutiny of the rhizobial partner, whereas the compatibility requirements for intercellular infection (known as ‘crack entry’) appear to be more relaxed (Madsen et al. 2010; Sprent 2007).

    In Lotus spp., both inter- and intracellular infection depend on NF signaling (Acosta-Jurado et al. 2016b; Montiel et al. 2020; Quilbé et al. 2022), except in rare cases, if organogenesis is activated in the absence of NF signaling (Madsen et al. 2010). Candidate gene approaches relying on interspecific variation in NF receptors have been used to demonstrate their roles in determining compatibility (Bozsoki et al. 2020; Gysel et al. 2021; Radutoiu et al. 2007). Effectors and secretion system components that deliver effectors into plant cells also affect compatibility. For instance, the bradyrhizobial NopP effector is recognized by soybeans carrying the Rj2 nucleotide-binding site-leucine-rich repeat receptor resistance gene, leading to termination of infection (Sugawara et al. 2018), and the Bradyrhizobium elkanii NopF effector prevents infection in L. japonicus but not in L. burttii, (Kusakabe et al. 2020). Rhizobial effectors can also promote symbiotic interactions, as exemplified by the B. elkanii effector Bel2-5, which confers the ability to nodulate soybeans deficient in NF perception (Ratu et al. 2021). Rhizobial genes that affect NF and EPS production can also influence host range, as demonstrated by the Sinorhizobium fredii HH103 mucR1, syrM, nolR, and nodD2 mutants (Acosta-Jurado et al. 2016a). Remarkably, syrM, nolR, and nodD2 mutations induce a shift from inter- to intracellular infection in L. burttii and extend the host range to include L. japonicus Gifu (Acosta‐Jurado et al. 2019, 2020). In addition, Lotus intraspecific variation was exploited to identify Pxy as a regulator of symbiotic compatibility downstream of EPS signaling (Kawaharada et al. 2021). Also in Lotus spp., S. fredii HH103 is able to form functional nodules on L. burttii, whereas it induced ineffective nodules on L. japonicus Gifu and a quantitative trait loci (QTL) was mapped to chromosome 1 near the Nfr1 gene (Sandal et al. 2012). Likewise, L. burttii was also more permissive than L. japonicus Gifu in its interaction with R. leguminosarum Norway (Grossmann et al. 2012).

    Here, we investigate natural variation in rhizobial host range between the two Lotus species L. japonicus Gifu and L. burttii, focusing on host control of symbiotic compatibility.


    L. burttii nodulates with rhizobia from five different genera.

    It was previously reported that S. fredii HH103 forms functional nodules on L. burttii but not on L. japonicus Gifu (Sandal et al. 2012). In order to determine if L. burttii is generally more permissive in its symbiotic interactions than Gifu, we examined the nodulation phenotypes of Gifu and L. burttii with a wide range of rhizobia from different genera, including Sinorhizobium, Azorhizobium, Bradyrhizobium, Rhizobium, Allorhizobium, and Mesorhizobium (Fig. 1). These distantly related rhizobial species produce NF with different chemical modifications at the nonreducing and reducing ends (Bek et al. 2010; D'Haeze and Holsters 2002; Renier et al. 2011) (Table 1). We observed large variation in nodule numbers and structures, which included nodule primordia “bumps”, white nodules, and, in some cases, small or more developed pink nodules (Fig. 2A and D). Among the 42 strains tested, only the cognate Lotus symbiont Mesorhizobium loti R7A induced development of pink nodules in both Gifu and L. burttii (Fig. 2A and D). At the other extreme, S. meliloti nodulated neither. Sinorhizobium fredii NGR234, which is compatible with a very broad range of legumes (Pueppke and Broughton 1999), was unique in promoting a larger number of pink nodules on Gifu than on L. burttii (Fig. 2D; Supplementary Fig. S1). None of the remaining 39 strains formed pink nodules with Gifu. In contrast, 30 of the 39 strains formed at least some pink nodules with L. burttii (Fig. 2A and D). The 30 strains that nodulate L. burttii comprise five of the genera tested, indicating that the symbiont range of L. burttii is broad and that L. burttii is considerably more promiscuous in its symbiotic interactions than Gifu, regardless of the diverse composition of the NFs produced by the rhizobial strains (Table 1).

    Fig. 1.

    Fig. 1. Phylogenetic distribution of rhizobia used in this study. Phylogenetic tree of alpha and beta-rhizobia adapted from Sprent et al. (2017), to highlight the number of species used in this work from each rhizobial genus to evaluate the nodulation capacity of Lotus japonicus Gifu and Lotus burttii.

    Download as PowerPoint

    Table 1. Composition and chemical decorations in the nonreducing and reducing ends of NF produced by rhizobial species used in this study and closely related strains

    Fig. 2.

    Fig. 2. Nodulation and shoot phenotypes. A, Presence (filled squares) or absence (unfilled squares) of bumps (B), white nodules (W), and pink nodules (P) on Gifu and Lotus burttii plants at 5 weeks postinoculation (wpi) with 42 different rhizobial strains. In column S, the black and red arrows indicate a significant increase or decrease of the shoot length with respect to mock-treated plants, respectively. Student's t test, P < 0.01. B to E, Violin dot plots showing the number of bumps, white nodules, pink nodules, and shoot length in Gifu and L. burttii harvested at 5 wpi with the rhizobial species indicated in A. Dashed lines in green and orange highlight the average shoot length in mock-treated plants of Gifu and L. burttii, respectively.

    Download as PowerPoint

    L. burttii forms non-beneficial interactions with diverse rhizobial species.

    Efficient nitrogen fixation by rhizobia in fully developed nodules has a positive impact on the plant host, reflected in a more vigorous growth of the aerial part (Lindstrom and Mousavi 2019). Accordingly, both Gifu and L. burttii plants nodulated by M. loti R7A showed a significantly greater shoot length compared with mock-treated plants (Fig. 2A and E). None of the 41 remaining Gifu-rhizobial associations had a positive impact on shoot length. In contrast, L. burttii growth was significantly enhanced by 10 different rhizobial species, nine of which developed pink nodules. However, inoculation with R. leguminosarum USDA2356 (strain 33), which did not induce any pink nodules, resulted in significantly longer shoots than the mock control (Fig. 2A and E), suggesting a growth-promoting effect independent of nitrogen fixation.

    Despite formation of pink nodules, 17 L. burttii–rhizobia interactions resulted in plants displaying similar shoot length to mock-treated plants, whereas a significant reduction in plant growth was observed in plants nodulated by nine rhizobial species (Fig. 2A and E). These results show that, although a wide range of rhizobial species developed pink nodules with L. burttii, most of them did not have a positive impact on growth and instead resulted in neutral or negative outcomes. This finding prompted us to further analyze the correlation between rhizobial colonization and plant growth. We used a subset of 11 L. burttii–rhizobia combinations, which included strains from the five different genera that formed pink nodules with L. burttii. For each of the 11 different L. burttii–rhizobia associations, only the shoot length of plants with fully developed pink nodules was compared with mock-treated plants and to plants nodulated by M. loti R7A (Fig. 3A). Additionally, the nodule structure and bacteroid occupancy in the nodule cells were visualized by light microscopy (Fig. 3B to M). Only L. burttii plants harboring pink nodules colonized by S. americanus and Allorhizobium undicola (strains 7 and 35) showed comparable shoot growth to plants nodulated by M. loti R7A (Fig. 3A). In contrast, plants nodulated by Bradyrhizobium sp. strain ORS285 (strain 38) exhibited a significant reduction in the shoot length, while the remaining L. burttii–rhizobia interactions did not affect the length of the aerial part (Fig. 3A). Nodule sections revealed successful rhizobial colonization by all 11 strains tested, though to different extents. L. burttii nodules were heavily colonized by R. leguminosarum SM140B, R. leguminosarum SM144A, and Bradyrhizobium sp. strain ORS285 (Fig. 3G, H and L), but none of these strains had a positive effect on plant shoot length (Fig. 3A).

    Fig. 3.

    Fig. 3. Shoot length and nodule histology of Lotus burttii plants. A, Violin dot plots showing the shoot length of L. burttii plants with pink nodules at 5 weeks postinoculation with rhizobial species from different genera. Letters a and c below the violin plots indicate nonsignificant differences between the shoot length compared with mock-treated or R7A-inoculated plants, respectively; letter b indicates significant difference compared with mock-treated and R7A-inoculated plants. Analysis of variance, Tukey P < 0.01. The gray dashed line shows the average shoot length in mock-treated plants. B to M, Nodule histology with representative images of pink nodules developed on L. burttii by different rhizobial strains. Scale bar = 1 mm for nodules and 100-μm nodule sections.

    Download as PowerPoint

    L. burttii allows nodulation with inefficient nitrogen fixers.

    These results confirm that the permissive nature of L. burttii that allows successful nodulation by a broad range of rhizobial species does not always result in growth promotion. Importantly, none of the 5-weeks-postinoculation (wpi) nodules showed green coloration indicative of nodule senescence. Therefore, the lack of plant growth promotion in certain L. burttii–rhizobia associations could be explained by either a delay in the development of nitrogen-fixing nodules, an inefficient nitrogen fixation, or both. To address these possibilities, we first recorded the number of pink nodules in the 11 L. burttii–rhizobia associations under analysis at earlier timepoints, 2 and 4 wpi. Pink nodules were commonly detected at 2 wpi with M. loti R7A and to a lesser extent with R. leguminosarum SM155B (Fig. 4A). However, at this timepoint some pink nodules were also observed with S. fredii HH103, R. leguminosarum SM144A, R. leguminosarum viciae, and Allorhizobium undicola LMGT. Except for M. loti R7A and R. leguminosarum SM155B, most of the plants with pink nodules were found at 4 wpi for any L. burttii–rhizobia interaction (Fig. 4A). The clear delay in the formation of pink nodules could explain the lack of a contribution to plant growth in several of the L. burttii–rhizobia associations. In addition, we evaluated the nitrogen-fixation efficiency by acetylene reduction assays in the pink nodules recorded at 5 wpi for the 11 L. burttii–rhizobia associations. Interestingly, in nodules colonized by Sinorhizobium, Bradyrhizobium, Allorhizobium, and Mesorhizobium species, the acetylene reduction levels were comparable or even higher than in nodules developed by the cognate rhizobial partner M. loti R7A. In contrast, the nitrogen fixation was remarkably low in nodules induced by any R. leguminosarum strain (Fig. 4B).

    Fig. 4.

    Fig. 4. Nodulation kinetics and nitrogen fixation in Lotus burttii–rhizobia interactions. A, The number of pink nodules recorded at 2 and 4 weeks postinoculation (wpi) in L. burttii plants inoculated with Sinorhizobium fredii HH103, S. fredii USDA257, S. americanus CFEI156, Rhizobium leguminosarum SM155B, R. leguminosarum SM140B, R. leguminosarum SM144A, R. leguminosarum viciae, Allorhizobium undicola LMGT, Mesorhizobium plurifarium PMS0804, Bradyrhizobium sp. strain ORS285, and B. pachyrhizi PMS0802. The total and nodulated (parenthesis) number of plants are indicated below the violin plots. B, Nitrogen fixation was evaluated per milligram of fresh weight of nodules (FWN) by acetylene reduction assays in pink nodules formed at 5 wpi in the L. burttii–rhizobia associations mentioned above. Student's t test of nitrogen fixation between nodules colonized by M. loti R7A and other rhizobial strains. One asterisk (*) indicates P < 0.05, three (***) P < 0.001.

    Download as PowerPoint

    A single genetic locus near nfr1 confers L. burttii promiscuity.

    With the exception of S. fredii NGR234, Gifu only formed large, pink nodules with its cognate efficient nitrogen fixer M. loti that contributed positively to plant growth. The situation was much more complex for L. burttii, which engaged in many different interactions with variable outcomes for the plant. To understand the genetics underlying L. burttii promiscuity, we inoculated at least 17 Gifu × L. burttii recombinant inbred lines (RILs) (Sandal et al. 2012) with the 11 diverse strains mentioned above. In addition, 70 RILs were inoculated with R. leguminosarum Norway, which nodulates L. burttii roots intercellularly but does not form nodules with Gifu (Grossmann et al. 2012; Liang et al. 2019) (Fig. 3; Supplementary Dataset S1).

    We scored the average number of pink nodules for each RIL-rhizobium combination (Supplementary Dataset S1) and analyzed the resulting data using R/qtl. TM0002, the marker previously reported to be associated with the S. fredii HH103 Gifu/L. burttii QTL (Sandal et al. 2012), was identified as the highest peak in the QTL analyses for all 12 strains (Fig. 5). Interestingly, the single locus that appears to be responsible for allowing L. burttii nodulation with all 12 strains tested was located on chromosome 1, near the NF receptor gene Nfr1.

    Fig. 5.

    Fig. 5. Quantitative trait loci (QTL) analysis of Gifu × L. burttii RIL nodulation. The trait analyzed is the average number of pink nodules. A, Sinorhizobium fredii HH103. B, S. fredii USDA257. C, S. americanus CFEI156. D, Rhizobium leguminosarum SM155B. E, R. leguminosarum SM140B. F, R. leguminosarum SM144A. G, R. leguminosarum viciae. H, Allorhizobium undicola LMGT. I, Mesorhizobium plurifarium PMS0804. J, Bradyrhizobium sp. strain ORS285. K, B. pachyrhizi PMS0802. L, R. leguminosarum Norway. The upper and lower dashed lines indicate 5 and 10% false discovery rate levels, respectively. Blue areas indicate the results of 1,000 QTL analyses on permuted data, which were used to determine the significance thresholds. Dashed vertical gray lines indicate chromosome ends.

    Download as PowerPoint

    A Gifu/L. burttii nfr1 domain swap construct complements the nfr1 mutant in a Gifu background but does not extend its host range.

    Our analysis revealed that an Nfr1-linked locus or Nfr1 itself is involved in determining permissiveness towards diverse rhizobial species in L. burttii. We sequenced the extracellular region of the L. burttii Nfr1 gene and found two mis-sense substitutions compared with the Gifu gene (T124A and Y213D, Gifu/L. burttii [Supplementary Fig. S2B]). This made Nfr1 a strong candidate gene for explaining the capacity of L. burttii to establish symbiotic associations with multiple rhizobial genera. To examine this possibility, we created a BinG1 domain swap construct in which the Nfr1 extracellular domain from Gifu was replaced by the corresponding fragment from L. burttii Nfr1. This construct was introduced into Gifu wild type and Ljnfr1-1 mutant backgrounds using Agrobacterium rhizogenes hairy-root transformation. The BinG1 construct was functional and rescued Gifu nfr1-1 nodulation with M. loti R7A (Fig. 6). In contrast, it did not enable either the Gifu wild type nor the Gifu nfr1-1 mutant to nodulate with S. fredii HH103 (Fig. 6). A functional L. burttii Nfr1 extracellular domain in the Gifu background was thus insufficient for extending the rhizobial host range of Gifu. Besides the differences in the ectodomain, the amino acid sequence of NFR1 in L. burttii and L. japonicus (Gifu) shows one substitution in the transmembrane domain (E260K, Gifu/L. burttii [Supplementary Fig. S2A]). The relevance of this modification was tested in transgenic roots, following the approach described above but using the full L. burttii Nfr1 sequence. The construct restored the nodulation in Ljnfr1-1 plants inoculated with M. loti R7A but was unable to extend the nodulation ability with S. fredii HH103 (Supplementary Fig. S2B), indicating that LbNfr1 is not sufficient to expand the symbiotic association of L. japonicus with S. fredii HH103.

    Fig. 6.

    Fig. 6. Influence of the Nfr1 genotype on nodulation with Sinorhizobium fredii HH103. Violin dot plots show the number of pink nodules found on Lotus burttii, L. japonicus Gifu, and the Ljnfr1-1 mutant transformed with the BinG1and LjNfr1 constructs or the empty vector at 5 weeks postinoculation with Mesorhizobium loti R7A and S. fredii HH103. BinG1 = Nfr1 extracellular domain of Gifu replaced by the corresponding fragment from L. burttii Nfr1 expressed under LjNfr1 promoter; LjNfr1 = Gifu Nfr1 sequence expressed under its native promoter. The number of plants tested is indicated below the violin plots.

    Download as PowerPoint

    Nodulation of L. burttii with R. leguminosarum Norway is a dominant trait.

    Recently, it was shown that R. leguminosarum Norway colonizes L. burttii roots intercellularly (Liang et al. 2019) but is unable to infect L. japonicus Gifu (Grossman et al. 2012). That report, along with the data presented in this study confirms the wide host range of L. burttii with diverse rhizobial species. To assess if this nodulation promiscuity is a recessive or dominant trait, the F1 progeny of crosses between L. burttii and L. japonicus Gifu were inoculated with R. leguminosarum Norway green fluorescent protein. Gifu remained without nodules 6 wpi. The F1 progeny developed nodules regardless of the parental combination (Fig. 7A) and all phenotyped F1 plants were heterozygous (Fig. 7B). These results show that the nodulation phenotype of L. burttii is dominant.

    Fig. 7.

    Fig. 7. Lotus burttii × L. japonicus Gifu F1 genotype and nodulation phenotype with Rhizobium leguminosarum Norway. A, Violin dot plots show the nodule numbers at 6 weeks postinoculation with R. leguminosarum Norway on L. burttii, L. japonicus Gifu, and the F1 progeny from crosses between these two genotypes. The number of plants tested are described below the violin dot plots. B, Agarose gel with PCR products amplified with a set of primers for the Lotus power marker TM1203, using as template DNA isolated from L. burttii (Lb), L. japonicus Gifu (Lj Gifu), and the F1 progeny from their crosses (Lb × Lj Gifu).

    Download as PowerPoint


    The legume-rhizobia symbiosis is an illustrative example of a highly specific and stringent molecular dialog between symbiotic partners. Despite this fact, it is known that certain rhizobial species, such as S. fredii NGR234, are able to establish symbiotic associations with diverse legumes from distant phylogenetic genera (Pueppke and Broughton 1999). In the last decades, both rhizobial and plant components defining legume-rhizobia compatibility have been identified (Walker et al. 2020). This study shows that a single locus in L. burttii is responsible for promiscuous interactions across rhizobial genera.

    L. burttii is a promiscuous symbiotic partner engaged in both beneficial and detrimental associations.

    Lotus burttii is related to L. japonicus, a model legume with extensive genetic and transcriptomic resources (Fukai et al. 2012; Kamal et al. 2020; Małolepszy et al. 2016; Mun et al. 2016; Urbański et al. 2012). Unlike L. japonicus, which seems to be nodulated by a narrow range of rhizobial partners, L. burttii established 30 associations with diverse rhizobia that culminated in the formation of pink nodules. Excluding the L. burttiiM. loti interaction, only nine rhizobial species had marginal but significantly positive contributions to host plant growth, while the rest of the rhizobial interactions either did not affect or had a negative impact on L. burttii growth. A recent study in which soil suspensions were used as inoculum revealed that the nodule microbiome of L. burttii plants with starvation symptoms was composed by bacteria belonging to what was taxonomically defined as Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium. In contrast, the nodule microbiota of starved plants in L. japonicus and L. corniculatus was dominated by Mesorhizobium species (Crosbie et al. 2022). These findings confirm the broad permissiveness of L. burttii to establish associations with diverse rhizobial genera. Except for R. leguminosarum Norway (Liang et al. 2018), the rhizobial strains used in our study are efficient nitrogen fixers in the symbiotic associations with their cognate plant hosts (Buendia-Claveria et al. 1989; Cavassim et al. 2020; Dreyfus and Dommergues 1981; Gao et al. 2004, 2015; Lajudie et al. 1994, 1998a and b; Moeskjær et al. 2021; Mora et al. 2014; Pueppke and Broughton 1999; Ramírez-Bahena et al. 2009). The suboptimal outcomes of many of these L. burttii–rhizobia l associations may be caused either by the delay observed in their nodulation kinetics, low levels of nitrogen fixation, or both. In some cases, we even observed poor nitrogen fixation in pink and well-colonized nodules, e.g., strain R. leguminosarum SM144A (Figs. 2, 3, and 4). Other strains produced relatively few pink nodules that fixed at very high levels and contributed positively to plant growth, e.g., strain Allorhizobium undicola LMGT (Figs. 2, 3, and 4). These findings indicate that additional compatibility elements acting at later stages of nodulation could be required for efficient nodule functioning and nitrogen fixation and transfer (Walker et al. 2020). Similar examples of inefficient legume-rhizobia symbioses have been documented in Medicago spp. Sinorhizobium meliloti 1021 is an efficient nitrogen-fixer in Medicago sativa nodules and is also able to form pink nodules with the Medicago truncatula accessions A17 and R108, but with poor nitrogen fixation performance (Kazmierczak et al. 2017; Terpolilli et al. 2008). Similarly, ineffective mutants of S. meliloti were comparable to their effective counterparts in colonizing M. sativa nodules (Amarger 1981). However, legumes possess mechanisms to reward or penalize the effectiveness of their rhizobial partners hosted within nodule cells. In soybean nodules, where inefficient nitrogen fixation was mimicked by substituting nitrogen for argon, the population and growth of rhizobia was drastically lower with respect to nodules in which nitrogen fixation was performed efficiently (Kiers et al. 2003). Likewise, four-generation experiments conducted with 12 M. truncatula genotypes inoculated with a mixture of three rhizobial strains from their native range revealed an increase in the relative frequency of more beneficial rhizobial strains, estimated by the nodule number and size (Heath and Tiffin 2009). A recent study shows that pea nodules containing a partly effective strain are sanctioned when the plant is coinoculated with an effective strain, but they will retain nodules hosting the intermediate-fixing strain when the only alternative was a nonfixer (Westhoek et al. 2021).

    For some legumes, a relatively low capacity to colonize new habitats seems to be related to a scarce presence of compatible symbionts (Parker 2001). This idea is supported by the high invasiveness of certain woody legumes that possess a broad compatibility with diverse rhizobial strains (Richardson et al. 2000). A comparative study among congeneric acacias revealed that those considered invasive associate with a significantly greater number of rhizobial strains than the natural and noninvasive acacias in Australia (Klock et al. 2015). However, studies conducted in distinct geographical regions show that differential invasiveness of Acacia species is not always determined by a broad promiscuity with rhizobial strains (Keet et al. 2017; Klock et al. 2016). Similar approaches could be taken with L. burttii to further understand the contribution of host range to legume adaptiveness.

    Molecular players restricting the compatibility of legume-rhizobium associations.

    Root nodule symbiosis encompasses different checkpoints at which suitable symbionts are scrutinized, from the early infection to development of functional, nitrogen-fixing nodules (Walker et al. 2020). The rhizobial host range is determined by the perception of specific root flavonoids, along with certain rhizobial effectors and genes that contribute to NF and EPS production (Acosta-Jurado et al. 2016a, 2019, 2020; Kusakabe et al. 2020; Ratu et al. 2021; Sugawara et al. 2018). From the plant side, the first step in symbiotic partner discrimination is the recognition of specific NFs by NF receptors (Amor et al. 2003; Radutoiu et al. 2003; Smit et al. 2007). The next level of selectivity is imposed by scrutiny of EPS produced by rhizobia. In Lotus spp., this relies on the EPS receptor LjEpr3 (Kawaharada et al. 2015), and EPS signaling appears to be of general importance across legumes. The incompatibility of S. meliloti Rm41 with M. truncatula A17 is abolished by incorporating the succinoglycan-coding exo gene of the compatible S. meliloti 1021 (Simsek et al. 2007). Similarly, EPS composition confers different levels of rhizobial resistance towards the antimicrobial M. truncatula nodule-specific cysteine-rich peptides (NCRs) produced in the nodule cells of certain legumes to impose terminal differentiation of bacteroids (Arnold et al. 2018; Montiel et al. 2017). In this regard, the presence of a functional NCR allele in the M. truncatula A17 accession restricts its symbiotic association with S. meliloti Rm41. This incompatibility is not present in the M. truncatula DZA315 accession that possesses a nonfunctional NCR allele (Wang et al. 2017; Yang et al. 2017).

    The broad host range in L. burttii is not explained by any of the plant regulators mentioned above. Unlike legumes of the inverted repeat–lacking clade, in which terminal differentiation of bacteroids is orchestrated by NCRs, this peptide family is absent in Lotus spp. (Kereszt et al. 2018). Our QTL analyses with data generated from 18 Gifu × L. burttii RILs inoculated with 12 diverse rhizobial strains showed that a locus near microsatellite marker TM0002 confers the symbiotic promiscuity of L. burttii. It is unlikely that LjEpr3 is responsible for the extended nodulation capacity of L. burttii, since this gene is not located near TM0002. By contrast, the NF receptor gene Nfr1 was an obvious candidate, since it is located near the TM0002 locus. However, Gifu plants expressing a functional extracellular domain or the full L. burttii Nfr1 did not result in host range expansion to include nodulation with S. fredii HH103. The L. burttii symbiotic associations are established with rhizobial strains producing NFs with very diverse decorations (Table 1) (Bek et al. 2010; D'Haeze and Holsters 2002; Renier et al. 2011), suggesting that minor changes to NF receptors may not be the most likely cause of L. burttii promiscuity. An alternative explanation is that other components linked to the TM0002 marker are responsible, e.g., the presence of several resistance proteins in Glycine max restrict strain-specific interactions with rhizobia (Walker et al. 2020). Likewise in Lotus accessions it was recently found that Bradyrhizobium elkanii USDA61 mutants disrupted in different effector proteins of the type III secretion system are affected at different checkpoints in their symbiotic association with L. burttii, L. japonicus Gifu, and L. japonicus MG-20 (Kusakabe et al. 2020). However, the broad promiscuity of L. burttii is unlikely to be linked to a missing resistance gene, since the F1 progeny of L. burttii and Gifu crosses retained the nodulation capacity with R. leguminosarum Norway, while Gifu wild-type plants were unable to develop nodules. Therefore, the genetic components responsible for the pronounced symbiotic promiscuity of L. burttii remain elusive.


    Nodulation phenotyping.

    For germination, seeds were scarified and surface-sterilized with 0.5% sodium hypochlorite for 15 min and were rinsed several times with distilled water. Seeds were kept in sterile water for 1 h, before sowing on wet filter paper. The germinated plants were transferred and grown on 1/4 B&D medium (Broughton and Dilworth 1971), where the surface of the agar slope was covered with filter paper. Rhizobia were grown in yeast mannitol agar, except for R. leguminosarum strains, which were grown in tryptone yeast medium. The strains were diluted to an optical density at 600 nm (OD600) of 0.02 before inoculation with 50 μl per plant, by pipetting the suspension directly on the root. The nodulation phenotype was recorded at 35 days postinoculation.

    QTL analysis.

    We used L. japonicus Gifu × L. burttii RILs (Sandal et al. 2012; Shah et al. 2016). For rough mapping, at least 17 RILs with balanced genotypes were chosen (Supplementary Dataset S1). For R. leguminosarum Norway, a larger set of 70 lines were used. Genotype and phenotype data was imported into R/qtl version 3.4.2 (R Project for Statistical Computing) using the read.cross command. After converting to RIL format, the genetic map and missing genotype values were estimated using and mqmaugment, respectively. Multiple QTL mapping was then conducted, using 1,000 permutations to determine significance thresholds.

    Hairy-root transformation with nfr1 constructs.

    The full L. burttii Nfr1 and BinG1 constructs were based on the L. japonicus Nfr1 complementation construct (carrying the entire LjNfr1 gene driven by its own promoter), which was modified using standard cloning techniques and was transferred into the pIV10 integration vector (AM235368). The constructs were transformed into Agrobacterium rhizogenes AR12 (Hansen et al. 1989). BinG1 was constructed as follows. A DNA fragment from position 4,090 to position 4,993 of the Nfr1 gene (AJ575246/AJ575247) was substituted with the corresponding fragment from L. burttii produced by PCR. The two L. japonicus Gifu/L. burttii polymorphisms identified in the Nfr1 extracellular region are both contained within the L. burttii fragment included in the BinG1 construct.

    Genotyping and phenotyping of F1 progeny of L. burttii and L. japonicus crosses.

    Crosses using L. burttii as mother and L. japonicus Gifu as father or the converse were generated. Seeds of L. burttii B-303, L. japonicus Gifu B-129, and the F1 progeny were scarified and surface-sterilized as described earlier. Six-day-old seedlings were transferred into tulip shaped Weck jars (Weck 745) containing 300 ml of sterilized sand-vermiculite mixture supplemented with 40 ml of FAB medium (500 μM MgSO4·7H2O, 250 μM KH2PO4, 250 μM KCl, 250 μM CaCl2·2H2O, 100 μM KNO3, 25 μm Fe-EDDHA [Duchefa Biochemie, Haarlem, The Netherlands]). After 2 days, each plant was inoculated with 1 ml of a R. leguminosarum Norway suspension (OD600 = 0.005) or 1 ml of FAB medium as a mock control. Plants were grown under a long-day photoperiod for 6 weeks and were phenotyped using a MZ16 FA stereomicroscope (Leica). For genotyping, genomic DNA was extracted from leaves lysed in liquid nitrogen. Lysates were suspended in 500 μl of extraction buffer (2% wt/vol CTAB, 1.42 M NaCl, 20 mM EDTA, 100 mM Tris-HCl, pH 8) supplemented with 3.1 μl beta-mercaptoethanol and were incubated at 65°C for 20 min. Suspensions were mixed with 300 μl of chloroform and were centrifuged at 18,407 × g for 5 min. Supernatants were mixed with 1/10 volume of 3 M NaOAc and were centrifuged at 18,407 × g for 15 min. Pellets were washed twice with 70% ethanol, were air-dried and re-suspended in 50 μl of distilled water. The DNA was used as template in PCR reactions with the Lotus marker TM1203 (forward: TTGAATAAGGCTCATAGATCC, reverse: CTTCAGTTTGGGTTTCAAGC) (Sato et al. 2001) and was verified by agarose gel electrophoresis.

    Acetylene reduction assay.

    Acetylene reduction assays were carried out as described by Reid et al. (2016). Pink nodules from 10 plants on a plate were collected and placed in a 5-ml glass GC vial. A syringe was used to replace 500 μl of air in the vial with 2% acetylene and nodules were incubated for 30 min. Two milliliters of gas were removed from the vials and the ethylene was quantified with a SensorSense (Nijmegen) ETD-300 ethylene detector operating in sample mode with 2.5-liter/h flow rate and 6-min detection time. The ethylene levels were divided by the fresh weight of the nodules.


    In this study, we have shown that L. burttii exhibits a remarkably broad host range that is controlled by a single, dominant genetic locus near the TM0002 marker.


    R Project for Statistical Computing:

    The author(s) declare no conflict of interest.


    Funding: This work was funded by Danish National Research Foundation grant DNRF79.

    The author(s) declare no conflict of interest.