Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Inactivation of Duplicated Nod Factor Receptor 5 (NFR5) Genes in Recessive Loss-of-Function Non-Nodulation Mutants of Allotetraploid Soybean (Glycine max L. Merr.)

Inactivation of Duplicated Nod Factor Receptor 5 (NFR5) Genes in Recessive Loss-of-Function... Abstract Chemically induced non-nodulating nod139 and nn5 mutants of soybean (Glycine max) show no visible symptoms in response to rhizobial inoculation. Both exhibit recessive Mendelian inheritance suggesting loss of function. By allele determination and genetic complementation in nod139 and nn5, two highly related lipo-oligochitin LysM-type receptor kinase genes in Glycine max were cloned; they are presumed to be the critical nodulation-inducing (Nod) factor receptor similar to those of Lotus japonicus, pea and Medicago truncatula. These duplicated receptor genes were called GmNFR5α and GmNFR5β. Nonsense mutations in GmNFR5α and GmNFR5β were genetically complemented by both wild-type GmNFR5α and GmNFR5β in transgenic roots, indicating that both genes are functional. Both genes lack introns. In cultivar Williams82 GmNFR5α is located in chromosome 11 and in tandem with GmLYK7 (a related LysM receptor kinase gene), while GmNFR5β is in tandem with GmLYK4 in homologous chromosome 1, suggesting ancient synteny and regional segmental duplication. Both genes are wild type in G. soja CPI100070 and Harosoy63; however, a non-functional NFR5β allele (NFR5β*) was discovered in parental lines Bragg and Williams, which harbored an identical 1,407 bp retroelement-type insertion. This retroelement (GmRE-1) and related sequences are located in several soybean genome positions. Paradoxically, putatively unrelated soybean cultivars shared the same insertion, suggesting a smaller than anticipated genetic base in this crop. GmNFR5α but not GmNFR5β* was expressed in inoculated and uninoculated tap and lateral root portions at about 10–25% of GmATS1 (ATP synthase subunit 1), but not in trifoliate leaves and shoot tips. Introduction Many legume species fulfill their nitrogen need through a symbiotic relationship with their microsymbiont, commonly called the ‘rhizobium’, leading to atmospheric nitrogen fixation and nitrogen gain. This symbiotic ability contributes to the global application of legumes (i.e. soybean, pea, bean, peanuts, chickpea and forages) for human food and animal feed supply. The legume symbiosis displays host strain specificity, i.e. only certain combinations of plant and microsymbiont are compatible for establishment of the nitrogen-fixing symbiosis (Markmann and Parniske 2009). Host specificity is determined by a complex set of signal exchanges, as plants secrete specific (iso)flavonoid ‘cocktails’ that induce the biosynthesis of bacterial-controlled lipochitin–oligosaccharides (Nod factors) that act as a plant morphogen and mitogenic factor after perception by specific plant receptors (Oldroyd and Downie 2008). Spontaneous mutants showing an altered nodulation pattern have been isolated in several legumes, e.g. rj1 in soybean (Williams and Lynch 1954) and sym2 in pea (Geurts et al. 1997). However, it was induced mutations and subsequent positional cloning of nodulation genes in Lotus japonicus and Medicago truncatula that helped develop a working paradigm of nodule initiation and downstream signaling pathways. Specifically, putative Nod factor receptor genes [namely those encoding Nod factor receptor 1 (NFR1) and 5 (NFR5)] were identified in L. japonicus (Madsen et al. 2003, Radutoiu et al. 2003), M. truncatula [for MtLYK3 and MtNFP (Ben Amor et al. 2003, Limpens et al. 2003)] and Pisum sativum [PsSym10 (Geurts et al. 1997, Madsen et al. 2003)]. All encoded LysM-type serine/threonine receptor kinases characterized by 2–3 more or less conserved LysM domains in the N-terminal region, a transmembrane domain and a highly conserved intracellular kinase domain. They belong to a larger receptor family that includes chitin receptor, related to plant pathogen perception (Zhang et al. 2007). Lotus japonicus NFR1 and NFR5 were suggested to play a role in host determination, transforming the non-hosts M. truncatula and Lotus filicaulis into hosts for Mesorhizobium loti (Radutoiu et al. 2007). Carroll et al. (1986) described the isolation of three non-nodulation soybean mutants (nod49, nod139 and nod772) in cultivar Bragg following ethylmethane sulfonate mutagenesis. nod49 and nod772 formed a complementation group with the naturally occurring mutant rj1, whereas nod139 was mutated in a separate gene (Mathews et al. 1989a). Another soybean non-nodulation mutant (nn5 in cultivar Williams) is allelic to nod139; independently the locus has been assigned the rj6 designation (Pracht et al. 1993). nod139 and nn5 fail to show the earliest morphological changes such as deformation and curling of root hairs in response to rhizobial inoculation and consequently fail to develop nodules. However, nod139 established a symbiotic interaction with mycorrhizal fungi (Mathews et al. 1989b), indicating that the mutation affected an early stage of nodule initiation. Back-crossing of nod139 or nn5 to their parent varieties Bragg or Williams, respectively, demonstrated only one functional wild-type allele at the nod139 locus, but crossing and progeny analysis between nod139 and Glycine soja (CPI100070 and PI468.397), or between nn5 and G. max Harosoy63 identified two unlinked functional wild-type loci which complemented the nod139 and nn5 mutations (Gresshoff and Landau-Ellis 1994, Pracht et al. 1993, Searle, 2001). Thus varieties Bragg and Williams appear to harbor one functional wild-type copy of the nod139 locus, whereas G. soja accessions and variety Harosoy63 have two functional copies, reflecting the ancient allotetraploid nature of soybean (Schlueter et al. 2007a). The loci were named nod139-1 and nod139-2 (Searle 2001), and correspond to the dominant wild-type Rj5 and Rj6 alleles reported by Pracht et al. (1993). The soybean genome is considered to be paleopolyploid, resulting from sequential genome duplications about 50 as well as 13 Mya (Schlueter et al. 2007a, Schmutz et al. 2009). Duplication occurred in segmental regions, accompanied by local genomic as well as sequence divergence. For example, the soybean autoregulation of nodulation (AON) receptor kinase gene GmNARK (Searle et al. 2003) is located on soybean chromosome 12 in a homeologous region of GmCLV1A (on chromosome 11). These genes share 91.5% of their nucleotide sequence, are both transcriptionally active (with a seemingly functional open reading frame), but differ in apparent function, as GmNARK mutations are recessive, and not complemented by GmCLV1A. Similar observations came from the comparison of duplicated regions of the FAD2 gene in soybean (Schlueter et al. 2007b). One concludes that duplicated sequences provide the opportunity for sequence divergence and possible metabolic, temporal and spatial specificity. As mutations in genes coding for LysM-type receptor kinases resulted in a similar Nod− Myc+ phenotype (Ben Amor et al. 2003, Limpens et al. 2003, Madsen et al. 2003, Radutoiu et al. 2003), and the non-allelic non-nodulating nod49 mutant has been shown to be impaired in NFR1 function (Indrasumunar 2007), a combined candidate gene approach involving mapping, allele sequencing, complementation and overexpression was used to clone the soybean NFR5 genes conditioning a non-nodulation phenotype in nod139 and nn5. Results Genetic mapping of nod139 loci Non-nodulation mutant nod139 (in the cv. Bragg background) was crossed to two wild soybean accessions, G. soja CPI100070 and PI468.397 (both capable of nodulation), and F1 plants were allowed to self-fertilize to produce two F2 mapping populations. F2 segregation of the nodulation phenotype was 15 : 1 (nod+ vs. nod−) in both populations, indicating two unlinked functional copies of nod139 in both G. soja parents. These results demonstrated that, as in soybean variety Harasoy63 (Pracht et al. 1993) and G. soja PI468.397 (Gresshoff and Landau-Ellis 1994), there are two functional copies of the nod139 gene in G. soja, referred to as nod139-1 and nod139-2. Using bulk segregant analysis and simple sequence repeats (SSRs), nod139-2 was mapped to chromosome 1 near markers Satt408, Sat036 and Satt071. Analysis of 75 F2 progeny, 15 from the nod139 × PI468.397 mapping population and 60 from the nod139 × CPI100070 mapping population, demonstrated that Satt408 is 8.6 cM from nod139-2 and Sat_036 is on the other side of the locus, 10.6 cM away. Satt071 was only polymorphic in the nod139 × PI468.397 mapping population and could only be loosely mapped on 15 non-nodulating F2 plants. No recombination events were identified between nod139 and Satt071, thereby indicating that Satt071 was within 3 cM of nod139-2. To identify the location of nod139-1, markers Satt509 and Satt426 were selected from the homologous chromosomal region to nod139-2 on chromosome 11. To construct a genetic map around nod139-1, Satt426 and Satt509 were mapped on 60 F2 non-nodulating plants from the nod139 × CPI100070 mapping population. Satt509 and Satt426 were on the same side of nod139-1, with Satt426 2.5 cM and Satt509 3.3 cM away from the locus. Markers Satt071, Satt426 and Satt509 were too far from the nod139-1 and nod139-2 loci to be considered as anchors for ‘chromosome walks’. However, functionally (i.e. affecting very early nodulation but not mycorrhizal symbiotic steps), a likely candidate for the nod139 loci was NFR5, as its presumed molecular partner, namely the LysM-type receptor kinase GmNFR1, was already defined by the nod49/rj1 mutations (unlinked to nod139/nn5) (Indrasumunar 2007). Accordingly we initiated a candidate gene isolation approach using bacterial artificial chromosome (BAC) cloning, allele sequencing and genetic complementation. Isolation of BAC clones carrying the NFR5 genes of soybean Primers designed in accordance with the conserved regions in orthologous NFR5 sequences of Lotus, Medicago and pea successfully amplified two DNA fragments with sizes of around 1.0 and 2.5 kb from genomic DNA of cultivar Bragg. The expected size of the PCR product based on the LjNFR5 sequence was 1,049 bp (NB: the larger PCR fragment stemmed from GmNFR5β* that was discovered later to contain a 1,407 bp retroelement insertion GmRE-1). At first only the 1.0 kb PCR product was cloned and sequenced for comparison with LjNFR5, showing 81% identity at the nucleotide level; it was chosen as the putative GmNFR5 probe to screen the BAC library of wild-type soybean variety PI437.654 (Tomkins et al. 1999). BAC library screening resulted in 25 positively hybridizing BAC clones, but, after checking these by PCR, only 15 were positive. Six positive BAC clones (88H1, 138J20, 153C20, 167H3, 69M16 and 179K5) were chosen for further analysis and BAC sequencing. Isolation of the putative GmNFR5α and GmNFR5β genes Sequencing of six positive BAC clones by primer walking revealed two highly related (95% similarity at the nucleotide level) DNA stretches similar to LjNFR5, named GmNFR5α and GmNFR5β (Supplementary Fig. S1). The presence of two genomic copies is in agreement with the duplicated nature of the soybean genome (Shoemaker et al. 1996, and arguments above). As there are two functional copies of nod139 (Gresshoff and Landau-Ellis 1994, Searle 2001), the two highly related sequences were assumed to represent these loci, and were considered to be good candidates for the location of the nod139/nn5 mutations. Gene and protein characteristics of GmNFR5α and GmNFR5β DNA sequences 871 and 1,580 nucleotides upstream of the GmNFR5α and GmNFR5β putative translation start sites, respectively, were obtained from BAC clones (69M16, 138J20 and 179K5 for GmNFR5α, and 88H1, 153C20 and 167H3 for GmNFR5β). The 3′ untranslated regions (UTRs) of 807 nucleotides (GmNFR5α) and 367 nucleotides (GmNFR5β) were also sequenced from these BAC clones. Analysis of the genomic sequences using the FGENESH program defined gene structures without introns, as outlined in Supplementary Fig. S1. Both predicted genes consisted of a single exon with 1,794 and 1,797 nucleotides of coding sequence for GmNFR5α and GmNFR5β, respectively. Analysis of the GmNFR5α and GmNFR5β gene sequences through the SMART website and comparison with LjNFR5 (Madsen et al. 2003) predicted a typical receptor kinase protein structure consisting of an extracellular receptor domain (containing three LysM domains) with an N-terminal signal peptide, a transmembrane domain and a cytoplasmic kinase domain (Fig. 1). Fig. 1 View largeDownload slide Domain structure and sequence alignment of NFR5 proteins. The predicted amino acid sequences of the soybean NFR5α and NFR5β proteins were compared with each other and with that of the corresponding sequence from L. japonicus, P. sativum and M. truncatula. Conserved amino acids are highlighted. Key protein domains (SP, signal peptide, LysM 1, 2, 3, LysM domains; TM, transmembrane domain; KD, kinase domain) are underlined. The nonsense mutation in GmNFR5α and the transposon insertion in GmNFR5β are indicated with a star and an arrowhead, respectively. Kinase catalytic subdomains are indicated with Roman characters. Fig. 1 View largeDownload slide Domain structure and sequence alignment of NFR5 proteins. The predicted amino acid sequences of the soybean NFR5α and NFR5β proteins were compared with each other and with that of the corresponding sequence from L. japonicus, P. sativum and M. truncatula. Conserved amino acids are highlighted. Key protein domains (SP, signal peptide, LysM 1, 2, 3, LysM domains; TM, transmembrane domain; KD, kinase domain) are underlined. The nonsense mutation in GmNFR5α and the transposon insertion in GmNFR5β are indicated with a star and an arrowhead, respectively. Kinase catalytic subdomains are indicated with Roman characters. The GmNFR5α and GmNFR5β genes encode predicted proteins of 598 and 599 amino acids, respectively, compared with 596 amino acids for LjNFR5. At the N-terminus, a hydrophobic stretch of 28 amino acids is predicted to act as a transit peptide, compared with 26 amino acids for LjNFR5 (Fig. 1). The transit domain is followed by three motifs of a LysM domain (Fig. 1; Supplementary Fig. S2A) with similarity to the LysM domains found in bacterial peptidoglycan-binding proteins and chitinases from yeast (Kluyveromyces lactis) and algae (Volvoc carteri; Madsen et al. 2003). A predicted transmembrane region occurs between the extracellular domain and the C-terminal cytoplasmic kinase domain (Fig. 1). The GmNFR5α and GmNFR5β kinase domains have motifs associated with functional serine/threonine kinases, but motifs VII and VIII may be strongly modified or absent (Supplementary Fig. S2B). Specifically, an aspartic acid residue commonly conserved in domain VII is missing, and domain VIII, comprising the activation loop, is absent. The NFR5 proteins of soybean shared 71–74% overall amino acid sequence identity with the Medicago, Lotus and pea sequences (Fig. 1). The sequence identity was higher (79–82%) in the transmembrane/kinase domains and lower (64–67%) in the extracellular domain, which is believed to be responsible for ligand binding and thus determination of host range (Radutoiu et al. 2007). nod139 and nn5 carry nonsense mutations in GmNFR5α Sequence characterization of GmNFR5α from nod139 and nn5 showed the presence of point mutations (Supplementary Fig. S1). Single nucleotide substitutions (nonsense) were identified within the GmNFR5α coding region of both mutants. The nod139 mutation was identified as a nonsense mutation occurring at nucleotide +1,013 from the translational start site (ATG), resulting in the conversion of a leucine codon to a nonsense codon at position 338 of the GmNFR5α protein (Fig. 1). This L338* mutation eliminates most of the kinase domain, presumably rendering the resultant protein non-functional. Allele nn5 also truncated the GmNFR5α protein by a nonsense mutation at +1,506 bp from the starting ATG, converting a tryptophan codon to a stop codon at position 502 (Fig. 1). A natural mutation of GmNFR5β in the parental lines Bragg and Williams Genetic analysis of nod139 and nn5 indicated that recessive alleles of the GmNFR5α/β genes were responsible for the non-nodulation phenotype. Both genes were functional in G. soja and Harosoy63, but one was non-functional in the parental lines Bragg and Williams (Gresshoff and Landau-Ellis 1994, Pracht et al. 1993). DNA sequencing of both G. soja CPI100070 GmNFR5 genes showed that both were wild type (open reading frame and expressed), confirming genetic segregation data. In contrast, nodulation-capable cultivars Bragg and Williams carried a 1,407 bp long insertion sequence (GmRE-1) at the same position of the GmNFR5β gene (Supplementary Fig. S3). The insertion has the characteristics of a non-autonomous retroelement (Kunze et al. 1997); it has long terminal repeats of 214 bp and a non-perfect duplication of the 11 bp target site. Retroelement insertion occurred at nucleotide +295 from the translational start site (ATG), resulting in the conversion of a leucine to a nonsense at position 99 (LysM1, extracellular domain) of the GmNFR5β protein. The resultant protein, if stable, would lack its kinase domain, transmembrane domain and most of the extracellular domain. Complementation of nod139 and nn5 Complementation of the two GmNFR5 mutant alleles was done by Agrobacterium rhizogenes ‘hairy root’ transformation (Kereszt et al. 2007). Chimeric root systems formed by K599 transformation are composed of both transformed and non-transformed roots, as about 50% of the roots formed in nod139 and nn5 plants failed to nodulate. This frequency is consistent with expression of 35SGUS (β-glucuronidase) in soybean chimeric plants. Such roots were not considered in further quantitative characterization. Future hairy root studies will benefit from the use of visually selective markers for co-transformation (such as the DsRED vector) giving red coloration to transformed roots. Roots without red color thus can be excised and eliminated from final data sets. The non-nodulation phenotype of mutants nod139 and nn5 (Fig. 2) was complemented by wild-type GmNFR5α and GmNFR5β genes by A. rhizogenes K599 transformation. Wild-type GmNFR5β was derived from G. soja CPI100070. These results also confirmed the root control of the non- nodulation phenotypes (Mathews et al. 1992). Nodulation parameters of complemented roots were similar (50–100 nodules per plant) and independent of the type of promoter used [i.e. the native promoters (548 and 1,203 bp, respectively) and the constitutive 35S promoter (Table 1)]. Notably, the 35S overexpressing roots [with 10–80 times higher levels of GmNFR5 mRNA than control levels as measured by quantititative real-time PCR (qRT-PCR)] nodulated normally, suggesting that nodule number was not limited by GmNFR5 RNA expression level. Additionally, reciprocal complementation with the GmNFR1α gene into mutants nod139 and nn5, and the GmNFR5α and GmNFR5β genes into mutants nod49 and rj1 did not result in nodulation, showing distinct functions for the two putative receptor components. Fig. 2 View largeDownload slide Complementation of the nod139 and nn5 mutation. Transformed root systems were scored 35 d after inoculation with B. japonicum CB1809. (A) Transgenic roots of nod139 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (B) Root system of nod139 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (C) Root system of nod139 transformed with K599 carrying full-length GmNFR5α driven by the 35S promoter of CaMV. (D) Individual hairy roots of nod139 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. (E) Nodulation of nod139 transformed with K599 carrying full-length GmNFR5β driven by the 35S promoter of CaMV. (F) Composite plants (on the right side) transformed with the empty vector show the growth- and nitrogen deficiency-related phenotype associated with the absence of nodulation, while individuals carrying the wild-type transgene have normal development (left side). (G) Transgenic roots of nn5 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (H) Root system of nn5 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (I) Root system of nn5 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. Bars represent 2 cm in A, B, C, G, H and I; 1.75 cm in D; 1 cm in E; 6 cm in F. Fig. 2 View largeDownload slide Complementation of the nod139 and nn5 mutation. Transformed root systems were scored 35 d after inoculation with B. japonicum CB1809. (A) Transgenic roots of nod139 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (B) Root system of nod139 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (C) Root system of nod139 transformed with K599 carrying full-length GmNFR5α driven by the 35S promoter of CaMV. (D) Individual hairy roots of nod139 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. (E) Nodulation of nod139 transformed with K599 carrying full-length GmNFR5β driven by the 35S promoter of CaMV. (F) Composite plants (on the right side) transformed with the empty vector show the growth- and nitrogen deficiency-related phenotype associated with the absence of nodulation, while individuals carrying the wild-type transgene have normal development (left side). (G) Transgenic roots of nn5 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (H) Root system of nn5 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (I) Root system of nn5 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. Bars represent 2 cm in A, B, C, G, H and I; 1.75 cm in D; 1 cm in E; 6 cm in F. Table 1 Average number of nodules on transgenic roots induced on wild-type and mutant plants transformed with the GmNFR5α and GmNFR5β genes   Empty vector (no GmNFR5)a  GmNFR5α + native 548 bp promoter  GmNFR5α + 35S promoter  GmNFR5β + native 1,203 bp promoter  GmNFR5β + 35S promoter  Nodule number per plant  nod139  0 ab  86 cd  100 d  82 cd  91 d  Bragg  58 b  59 b  72 bc  61 bc  75 bc  Nodule number per root  nod139  0 a  22 c  28 c  20 c  25 c  Bragg  6 b  8 b  9 b  8 b  9 b  Nodule number per mg root dry weight  nod139  0.0 a  1.8 c  2.1 c  1.6 c  2.1 c  Bragg  0.3 b  0.5 b  0.5 b  0.5 b  0.5 b    Empty vector (no GmNFR5)a  GmNFR5α + native 548 bp promoter  GmNFR5α + 35S promoter  GmNFR5β + native 1,203 bp promoter  GmNFR5β + 35S promoter  Nodule number per plant  nod139  0 ab  86 cd  100 d  82 cd  91 d  Bragg  58 b  59 b  72 bc  61 bc  75 bc  Nodule number per root  nod139  0 a  22 c  28 c  20 c  25 c  Bragg  6 b  8 b  9 b  8 b  9 b  Nodule number per mg root dry weight  nod139  0.0 a  1.8 c  2.1 c  1.6 c  2.1 c  Bragg  0.3 b  0.5 b  0.5 b  0.5 b  0.5 b  aAt least 20 replicates for each treatment were scored 35 d after inoculation with B. japonicum strain CB1809. bNumbers followed by the same letter for the same measured parameter are not significantly different at the probability 0.05. View Large Expression analysis of the soybean NFR5 genes Using common qRT-PCR primers for both GmNFR5α and GmNFR5β, mRNA levels were 1.7–1.8% of GmActin5 levels in uninoculated roots of 2-week-old wild-type Bragg and supernodulating nts1007 plants, and slightly lower (1.2–1.4%) in mutant nod139 plants. In inoculated roots, expression slightly increased to 2% of that of GmActin5 in Bragg, but decreased to 1.3–1.4% in nts1007. These differences were not significant. However, expression in mature nodules and leaves was significantly lower (0.3–0.4% of GmActin5). More detailed analysis using specific primers for GmNFR5α and GmNFR5β RNA allowed insights into differential expression. The mRNA level of GmNFR5α was 10–25% of that of GmATS1 in uninoculated and inoculated portions of lateral and tap roots in 2-week-old wild-type Bragg (Fig. 3A). In contrast, the expression of GmNFR5β* (note: affected by the GmRE-1 insertion) was much lower than that of GmNFR5α, being <1% of that of GmATS1 (Fig. 3B). One presumes the transcript to be non-coding because of the insertion into the exon. In general, there is a down-regulation of GmNFR5α/β* transcript levels in inoculated tap root portions, though that effect may not be significant as lateral roots, responding strongly to inoculation for ENOD40-1 (Fig. 3C), had no response. Trifoliate leaves and the shoot tip region had extremely low or undetectable transcript levels for both genes. The presence of the retroelement in GmNFR5β* severely lowered the transcript level in all tissues. We conclude that expression of GmNFR5α and GmNFR5β* is predominantly root localized and marginally inoculation sensitive, as measured on this whole root section level. Cell type-specific regulation may have been swamped by non-responding basal root tissues. Fig. 3 View largeDownload slide Expression levels of GmNFR5α (A), GmNFR5β (B) and GmEnod40-1 (C) genes relative to that of GmATS1 (ATP synthase subunit 1) in tap root (TR), lateral root (LR), trifoliate leaves (TF) and shoot tips (STIP). The transcript level was determined by qRT-PCR. Section 1, 0–2 cm from the root tip; section 2, 2–4 cm from the root tip; section 3, 4–6 cm from the root tip. Data are the average of two technical repeats and three biological repeats. Error bars represent the standard error. Fig. 3 View largeDownload slide Expression levels of GmNFR5α (A), GmNFR5β (B) and GmEnod40-1 (C) genes relative to that of GmATS1 (ATP synthase subunit 1) in tap root (TR), lateral root (LR), trifoliate leaves (TF) and shoot tips (STIP). The transcript level was determined by qRT-PCR. Section 1, 0–2 cm from the root tip; section 2, 2–4 cm from the root tip; section 3, 4–6 cm from the root tip. Data are the average of two technical repeats and three biological repeats. Error bars represent the standard error. To visualize the expression of GmNFR5 genes in soybean roots and nodules, we cloned their native promoters in front of the GUS reporter gene in vector pCAMBIA1305.1, and transformed the reporter constructs into wild-type Bragg and its supernodulating mutant nts1007 via A. rhizogenes-mediated transformation (Kereszt et al. 2007). The expression patterns determined by the two promoters were similar and were not affected by the genotype of the transformed plant. In uninoculated roots, GUS activity was observed in most root tissues (Fig. 4A); however, the intensity of the staining decreased with the age of the root. In inoculated root, the expression of the GUS reporter gene increased during the early steps of nodule development (see arrows in Fig. 4B, C), but decreased in mature nodules (Fig. 4D). Fig. 4 View largeDownload slide Localization of NFR5 gene expression in soybean roots. (A) Cross-section of an uninoculated transgenic root. (B) Cross-section of an inoculated root with a developing side root. The arrow indicates a bacterial infection site with elevated GUS expression. (C) A hairy root developed on the supernodulation mutant nts1007 with nodules at different developmental stages. The primary root and the mature nodules show weak staining which is more intense in the younger tissues. Arrows show enhanced expression associated with nodule initiation which is also observed in developing nodules (arrowhead). (D) Cross-section of a nodule and the root shown in C; GUS expression can be observed in the cortex. Bars represent 300 μm in A and B; 3 mm in C; 1 mm in D. Fig. 4 View largeDownload slide Localization of NFR5 gene expression in soybean roots. (A) Cross-section of an uninoculated transgenic root. (B) Cross-section of an inoculated root with a developing side root. The arrow indicates a bacterial infection site with elevated GUS expression. (C) A hairy root developed on the supernodulation mutant nts1007 with nodules at different developmental stages. The primary root and the mature nodules show weak staining which is more intense in the younger tissues. Arrows show enhanced expression associated with nodule initiation which is also observed in developing nodules (arrowhead). (D) Cross-section of a nodule and the root shown in C; GUS expression can be observed in the cortex. Bars represent 300 μm in A and B; 3 mm in C; 1 mm in D. Sequence comparison of GmNFR5α and GmNFR5β in several soybean genotypes Single-nucleotide polymorphisms (SNPs) were found in the GmNFR5α (Supplementary Fig. S4) and GmNFR5β (Supplementary Fig. S5) sequences among various soybean genotypes. Several polymorphisms were identified within the coding sequence, the promoter region and 3′ UTR of soybean cv. Bragg (GQ340451), Williams (GQ340452), PI437.654 (GQ340449) and G. soja CPI100070 (GQ340450). Specifically, four SNPs, two being synonymous, were detected in the coding sequence of GmNFR5α in those soybean cultivars. The GmNFR5α sequences of Williams and cultivar PI437.654 are identical. Glycine soja CPI100070 and cv. Bragg have one and four SNPs, respectively, compared with Williams and PI437.654. GmNFR5β is less conserved than GmNFR5α. GmNFR5β of cv. Williams (GQ340458) and Bragg (GQ340457) is naturally mutated by a retroelement insertion (Supplementary Fig. S3), while G. soja CPI100070 (GQ340456) and PI437.654 (GQ340455) were wild type for this gene. Many known soybean cultivars also lacked the retroelement, allowing analysis of known breeding pedigrees. Transposable elements that comprise much of the repetitive DNA sequences in the genome have been reported in soybean, such as transposon G. max (Tgm; Vodkin et al. 1983, Rhodes and Vodkin 1988), a mariner-like element (Soymar1; Jarvik and Lark 1998) and a copia/Ty1-like retroelement (SIRE-1; Laten and Morris 1993, Laten et al. 1998). The size of these retroelements ranged from 1.6–12 kb (Tgm), 3.5 kb (Soymar1) and 11 kb (SIRE-1). Beside the retroelement insertion, comparison of the GmNFR5β coding sequence in cv. Bragg and Williams with G. soja CPI100070 and PI437.654 showed that there are seven SNPs, two of them translated into similar amino acids, while the other five translated into different amino acids (Supplementary Fig. S5). These did not affect the apparent symbiotic abilities and host range of the wild type. Unlike GmNFR5α, where cv. Bragg has four SNPs compared with cv. Williams, these two cultivars have the same GmNFR5β sequence not only in the coding sequence, but also in the promoter region and the 3′ UTR. Glycine soja CPI100070 and PI437.654, which have wild-type versions of GmNFR5β, have only one SNP in the coding sequence of GmNFR5β. However, in the 3′ UTR (the first 180 bp after the stop codon), G. soja CPI100070 has the same sequence as cv. Bragg and Williams, but has three SNPs compared with soybean cultivar PI437.654. This further verifies that modern cultivars of soybean are closely related and show some polymorphism to more ancestral types commonly called ‘Glycine soja’. However, at times wild accessions are used in soybean breeding, introducing ancestral gene blocks into modern cultivars. Widespread occurrence of the GmRE-1 retroelement insertion in the GmNFR5β gene The distribution of the GmRE-1 insertion was analysed using genomic DNA of available ancestral, first- and second-generation soybean lines. As expected, the PCR fragment was amplified from the Bragg and Williams parental lines and their mutants, but was absent in G. soja and cultivar Harosoy63 (Fig. 5A), which had already been shown by genetics to carry the wild-type alleles of both genes (Gresshoff and Landau Ellis 1994, Pracht et al. 1993). Fig. 5 View largeDownload slide Widespread distribution and the origin of a putative retroelement insertion in the NFR5β gene of US soybean cultivars. (A) Bragg, Williams, nod139 and nn5 have the mutant allele, but the mutant allele was absent in Harosoy63 and G. soja CPI100070. (B) Explanation of the origin of the putative retroelement insertion on cultivars Bragg and Williams. Red highlight, soybean with mutant allele; green highlight, soybean with wild-type of GmNFR5β. The mutant allele of Bragg was derived from D49-2491 (sibling of Lee). Williams inherited the mutant allele from L57-0034 (hashed red means ‘presumptive carrier’). Fig. 5 View largeDownload slide Widespread distribution and the origin of a putative retroelement insertion in the NFR5β gene of US soybean cultivars. (A) Bragg, Williams, nod139 and nn5 have the mutant allele, but the mutant allele was absent in Harosoy63 and G. soja CPI100070. (B) Explanation of the origin of the putative retroelement insertion on cultivars Bragg and Williams. Red highlight, soybean with mutant allele; green highlight, soybean with wild-type of GmNFR5β. The mutant allele of Bragg was derived from D49-2491 (sibling of Lee). Williams inherited the mutant allele from L57-0034 (hashed red means ‘presumptive carrier’). The parents of cultivar Bragg, i.e. Jackson and D49-2491 (a sibling of Lee), and the parents of D49-2491 (S100 and CNS) (Fig. 5B) were tested for the mutant retroelement allele. An amplification product of the same size and sequence as for Bragg, Williams and their mutants was detected in Lee and CNS, indicating that the origin of the insertion allele was in the other parent of D49-2491, namely line CNS. Cultivar Wayne, which was derived from CNS and is a parent of Williams, did not carry the insertion allele. However, other ancestors such as Clark, Richland and Volstate (Fig. 5A), which have no known relationship to CNS, possess the insertion sequence in the GmNFR5β gene. A search of the soybean genome (http://www.phytozome.net/soybean; see Schmutz et al. 2009) gave multiple locations for GmRE-1. Three of these loci (other than GmNFR5β*) showed high sequence similarity (97%); moreover, part of the retroelement (ranging between 50 and 350 bp) was present at several positions on all chromosomes with 80–99% sequence similarity (Supplementary Table S2). Thus GmRE-1 belongs to a moderately repeated family, having diverged since the ancestral segmental genome duplication leading to the two GmNFR5 loci. Discussion This study demonstrates that soybean can be analyzed effectively, combining mutant studies, electronic databases, genomic approaches and functional complementation. Specifically we were able to resolve a long-standing paradox involving recessive inheritance in a duplicated genome. Gene discovery in important legume crops such as soybean has been hampered by their larger genome size, endoduplication and tetraploidy, and difficulties with genetic complementation by transformation. Using loss-of-function mutants with a non-nodulation phenotype, and genetic mapping coupled with a gene candidate approach, the GmNFR5α and GmNFR5β genes of soybean encoding putative Nod factor receptors were cloned and characterized. GmNFR5 was duplicated in the genome, raising the question of how apparent loss-of-function mutants, such as nod139 and nn5, could be isolated. GmNFR5α and GmNFR5β are located on soybean chromosomes 11 and 1, in tandem with GmLYK7 and GmLYK4, respectively (Zhang et al. 2007). LYK4 and LYK7 genes encode related LysM receptor kinases (Limpens et al. 2003), suggesting ancestral segmental duplication prior to allopolyploidization (NB: in pea and M. truncatula LYK3 and LYK4 genes were proposed to be the entry receptor for nodulation) (Limpens et al. 2003). Many examples of two independent genes controlling the same trait can be found in soybean (Palmer and Kilen 1987). Shoemaker et al. (1996) also reported that hybridization to soybean genomic DNA by 280 randomly chosen PstI genomic clones determined that >90% of the probes detected more than two fragments and nearly 60% detected ≥3 fragments. From these findings, Stacey et al. (2004) suggested that <10% of the genome may be single copy sequence and that large amounts of the genome may have undergone genome duplication in addition to the tetraploidization event. Analysis and complementation of the mutants Hairy roots overexpressing GmNFR5α and GmNFR5β from the 35S promoter (with 10–80 times higher GmNFR5 mRNA levelss than control levels as measured by qRT-PCR) did not significantly increase nodule number in nod139 or Bragg in comparison with the hairy roots transformed with GmNFR5α and GmNFR5β driven by their native promoter, whether expressed per plant, per unit root or per unit root mass (Table 1). As overexpression of GmNFR1α (Indrasumunar 2007) increased nodulation and GmNFR5α/GmNFR5β did not, we conclude that transcription of GmNFR5α and GmNFR5β is not limiting nodule initiation. Zhang et al. (2007) found, based on sequence identification provided from this laboratory, that transcript levels of GmNFR5α and GmNFR5β in lateral or primary roots were three and five times higher, respectively, than the transcript level of GmNFR1α. As proposed by Madsen et al. (2003) and Radutoiu et al. (2003, 2007), NFR1 and NFR5 may act in concert as Nod factor receptors; therefore, increasing the transcript level of GmNFR1α that was limited will increase nodulation significantly, while increasing the transcript level of GmNFR5 that already is abundant will not result in a significant increase of nodulation. Related to this, an imbalance between NFR1 and NFR5 expression was suggested to cause the reduced nodulation efficiency observed in M. loti-inoculated L. filicaulis plants transformed with LjNFR1 and LjNFR5, separately (Radutoiu et al. 2007). The wild type and mutants respond differently in nodulation. Nodulation parameters of complemented root systems of mutant nod139 were significantly higher than wild-type Bragg (Table 1). Differential AON may be the cause of this difference, because all hairy roots of Bragg nodulated, whereas only about 50% did in nod139 roots. Thus complemented nod139 root systems may have triggered less AON, leading to increased nodule numbers. Unlike GmNFR1 which has only one functional copy in soybean (Indrasumunar 2007), both GmNFR5 genes are functional and complement each other. Comparison of protein identity shows that GmNFR5α and GmNFR5β share marginally higher identity (94%) than GmNFR1α and GmNFR1β (90%). In the extracellular domain, GmNFR5α and GmNFR5β also have higher similarity (91%) than GmNFR1α and GmNFR1β (82%). However, both pairs of genes have equally high identity (96%) in the kinase domain. Cross-complementation of GmNFR1 into GmNFR5 mutants (nod139, nn5), and GmNFR5 into GmNFR1 mutants (nod49, rj1) showed that these genes did not complement each other. Protein characteristics of GmNFR5α and GmNFR5β Protein domain structure and topology predict that the extracellular LysM domains would be involved directly or indirectly in binding of Nod factors, thus initiating signal transduction through intracellular kinases (Limpens et al. 2003, Madsen et al. 2003, Radutoiu et al. 2003, Radutoiu et al. 2007). Comparison of extracellular regions between GmNFR5α and GmNFR5β shows amino acid identities of 92, 91 and 90% in LysM1, LysM2 and LysM3, respectively. Detailed comparison of LysM motifs between GmNFR5 and orthologous NFR5 proteins of L. japonicus, P. sativum and M. truncatula reveals that GmNFR5 has higher similarity to LjNFR5 (57–79%) than the NFR5 homolog of Medicago (44–71%) and pea (53–73%); at the same time the NFR5 homologs of Medicago and pea share high similarity (72–78%) in the LysM motifs (Supplementary Fig. S2A). Soybean, lotus, pea and Medicago represent different nodulation types (determinate vs. indeterminate) and bacterial cross-inoculation groups. Madsen et al. (2003) speculated that NFR5 is not activated by an activation loop, but rather that ligand or interactor binding immediately activates the kinase. Alternatively, NFR5 might lack kinase activity and require an associated kinase, perhaps NFR1, to phosphorylate downstream substrates (Madsen et al. 2003). Radutoiu et al. (2007) showed that the Nod factor specificity of legume–rhizobial symbioses is determined by both NFR1 and NFR5. Transferring L. japonicus NFR1 and NFR5 to M. truncatula enables nodulation of the transformants by the L. japonicus symbiont M. loti. In addition, the specificity for different rhizobial symbionts of two different Lotus species is a function of a single amino acid residue (Leu118) within LysM2 domains of NFR5. Furthermore, LysM2 of NFR5 is the most diverged LysM domain among NFR1 and NFR5 homologs found in other plant species (Madsen et al. 2003; Supplementary Fig. S2A). Zhukov et al. (2008) also showed that a pea mutant (RisNod4) carrying an L77 to F substitution in the LysM1 domain of PsSym37 (PsNFR1) displays a restrictive symbiotic phenotype. RisNod4 mutants develop nodules only in the presence of a ‘Middle East’ Rhizobium strain producing O-acetylated Nod factors, indicating the PsNFR1 involvement in Nod factor recognition. These results provide strong evidence that NFR1 and NFR5 are indeed the Nod factor receptors. To confirm this conclusion, further studies on LysM domain swap between GmNFR1/5 and LjNFR1/5 are needed. In addition, further biochemical analyses of the binding affinities are still required to support the conclusion. Retroelement (GmRE-1) insertion in the NFR5β gene in many soybean cultivars Analysis of the amplification results and the pedigree of the tested soybean cultivars revealed that at least six ancestral lines (CNS, Richland, Volstate, Peking, Perry and Dorman), thought to be unrelated, carry the same insertion. We note that detailed pedigree and breeding histories are difficult to obtain. CNS, Richland, Peking and Dunfield are claimed to be of Chinese origin and thus might have common ancestors! Since these plants represent at least 20% of the genetic base of North American soybean lines (Gizlice 1994), the presence of a common retroelement insertion event implies that the genetic diversity of these cultivars is even lower than predicted from breeding and historical data. Interestingly, although most of the non-US cultivars tested were devoid of the insertion allele, the Japanese cultivar Enrei (Fig. 5A), of unknown pedigree, also carried the insertion, indicating a common ancestor(s) with the North American lines. Soybean cv. Enrei was the wild-type parent of non-nodulation mutants (En115, En1282 and En1314) and supernodulation mutant En6500 (Francisco and Akao 1993). En1314 and En1282 were mutated in the nfr1 gene (Ikeda et al. 2008). En115 did not produce root cortical cell division in response to inoculation of rhizobia (Francisco and Akao 1993) and, in this respect, is similar to nod139, which fails to initiate cortical cell division (Mathews et al. 1989b). Retroelement (GmRE-1) insertion in GmNFR5β facilitated the isolation of non-nodulation mutant lines Most of soybean nodulation mutants were derived from wild-type soybean, which already have the natural mutation in GmNFR5β (Bragg, Williams and Enrei). This is understandable due to its polyploid history; many examples of duplicated genes can be found in soybean (Palmer and Kilen 1987, Schmutz et al. 2009). In the case of GmNFR5, if ethylmethane sulfonate mutagenesis had been applied to a soybean cultivar that still had both wild-type copies of GmNFR5 (i.e. Harosoy63 and G. soja CPI100070, Illini and Jackson), non-nodulation phenotypes are unlikely to be found. Expression study of GmNFR5 Expression of GmNFR5α and GmNFR5β* was root specific, as determined by qRT-PCR (Fig. 3) and GUS fusion (Fig. 4). Trifoliate leaves and the shoot tip region had extremely low transcript levels for both genes. Expression determined by the presumed GmNFR5α and GmNFR5β promoters was observed in most root tissues of uninoculated roots (Fig. 4A). Inoculated roots showed increased expression of GUS during the early steps of nodule development (see arrows in Fig. 4B, C) except mature nodules (Fig. 4D). Similar results were obtained for reporters of the receptor-like kinase genes, M. truncatula Nod-Factor Perception (MtNFP; Arrighi et al. 2006) and M. truncatula Nodulation Receptor Kinase (MtNORK; Bersoult et al. 2005). However, the expression of MtNORK and MtNFP persisted in mature M. truncatula nodules, whereas the expression of GmNFR5 in soybean nodules did not (Fig. 4D; consistent with qRT-PCR data; see above). This observation most probably relates to the difference between indeterminant nodules containing a persistent meristem (as in M. truncatula) compared with mature determinant nodules devoid of a meristem in soybean. By the cloning of GmNFR1 (Indrasumunar 2007) and GmNFR5, all available non-nodulation mutants of soybean were successfully characterized. Despite extensive mutational screens, there is only a limited number of non-nodulation mutants of soybean compared with the diploid model legumes L. japonicus and M. truncatula. This most certainly is due to the allotetrapoloid nature of soybean. For example, we identified two copies of GmPoltergeist (A. Miyahara unpublished data), G. max nodulation receptor kinase (GmNORK) (GQ336811, GQ336812; Supplementary Fig. S6) and G. max kinase- associated protein phosphatase (GmKAPP) (Miyahara et al. 2008). Such functional duplication would prevent the detection of a visible phenotype of loss-of-function mutants. The availability of the soybean genome, candidate gene sequences from model legumes and soybean loss-of-function mutants strengthens their respective utility for the analysis of nodulation biology. Furthermore duplicated genomes, though seemingly recalcitrant to genetic analysis, evolve independently by insertions or point mutations, often leading to genetic inactivation of duplicated alleles and diploidized inheritance. Materials and Methods Soybean genotypes, F2 mapping populations and Bradyrhizobium strain Soybean plants were grown as described by Carroll et al. (1985a, 1985b). Mutant nod139 (Carroll et al. 1986, Mathews et al. 1989a) was derived from G. max variety Bragg. Non-nodulation G. max mutant nod139 was also crossed to G. soja PI468.347 (Kolchinsky et al. 1997) to produce the F2 mapping population of 481 plants. A second F2 mapping population was produced by crossing nod139 to G. soja accession CPI100070 to produce a F2 mapping population of 960 plants. The G. soja accession was selected from the Plant Introduction/Quarantine Unit, Commonwealth Scientific and Industrial Research Organization Division of Plant Industry (Canberra, Australia). All plants were grown as described by Carroll et al. (1985a, 1985b) and were inoculated 7 and 14 d after germination with Bradyrhizobium japonicum CB1809 (Biocare Pty. Ltd., Melbourne, Australia). The nodulation phenotype of plants was scored 28 d after germination. SSR marker detection Bulk DNA or individual plant DNA was used as template in PCRs with selected SSR oligonucleotide primers. The oligonucleotide sequence of selected primers was obtained from Cregan et al. (1999). Primers were used either in multiplex PCRs to amplify five SSR markers simultaneously or as separate pairs to PCR amplify a single SSR marker. Single SSR marker amplification was performed in 20 μl PCRs containing 0.3 μM of each primer, 0.075 μCi of [α-33P]dATP (3,000 Ci mmol−1), 20 μM of dNTPs, 10 mM Tris–HCl (pH 8.3), 2.0 mM MgCl2, 0.5 U of Taq polymerase (Perkin-Elmer, Foster City, CA, USA) and 50 ng of soybean genomic DNA. Isolation of the GmNFR5 probe Primers to amplify the ortholog of L. japonicus NFR5 (namely GmNFR5) were designed in accordance with the conserved sequence of LjNFR5 and orthologous NFR5 sequences of M. truncatula and P. sativum. PCR amplification was performed in triplicate in a 25 μl volume containing 50 ng of genomic DNA of Bragg, 10× PCR buffer, 2.5 mM MgCl2, 200 mM dNTPs, 0.5 U of Taq DNA polymerase recombinant (Invitrogen, Mount Waverley, VIC, Australia) and 0.2 μM of both forward and reverse oligonucleotide primers. Amplification was performed using the following cycling condition: one cycle of 94°C for 2 min, 35 cycles of 94°C for 10 s, 55°C for 30 s and 68°C for 2 min, followed by one cycle of 68°C for 7 min. Specific PCR products were cloned into pCR®4TOPO® vector (Invitrogen) according to the manufacturer’s instructions. After plasmid purification, the plasmid-containing PCR products were sequenced on an ABI 377 sequencer at the Australian Genome Research Facility, Brisbane, Australia. Sequences were viewed using the Chromas 1.45 software package (Griffith University, Australia). The sequences of PCR products were then compared with the sequence of LjNFR5. When the sequence of PCR products had high similarity (>60%) to LjNFR5, the PCR products were used as probes to screen a BAC library of soybean PI437.654 (Tomkins et al. 1999). Isolation of the GmNFR5α and GmNFR5β genes A good candidate probe with high homology to the LjNFR5 sequence was used to screen the BAC library PI437.654 (Tomkins et al. 1999). Positively hybridizing BAC clones were then ordered from Clemson University Genome Institute; subsequent BAC analysis and DNA sequencing were performed to clone the complete putative gene sequences of GmNFR5α and GmNFR5β. Primers used for BAC sequencing are listed in Supplementary Table S1. Sequence analysis of nod139 and nn5 mutants Gene-specific primers (Supplementary Table S1) were designed in accordance with the DNA sequence of GmNFR5α and GmNFR5β to amplify both genes from soybean varieties Bragg, Williams, nod139, nod49, nn5 and G. soja CPI100070. PCR products were cloned into pCR®4TOPO® vector (Invitrogen). To find the mutations, the sequences of both GmNFR5α and GmNFR5β mutants were compared with the sequences of their wild-type parent (Bragg and Williams) and G. soja CPI100070. A PCR test for detection of the GmNFR5β insertion allele Soybean seed lines for detection of retroelement insertion in GmNFR5β were provided by Dr. Sally Dillon and Dr. P. K. Lawrence from the Australian Tropical Grains Germplasm Centre and Professor Randall Nelson from USDA-ARS and University of Illinois, USA. These soybean seeds were: Adams, Altona, Amsoy, Blackhawk, Clark, CNS, Corsoy, Dorman, Dragon, DT84(1), Dunfield, Enrei, Forrest, Fukuyutaka, Harosoy, Harosoy63, Hawkeye, Illini, Jackson, Kent, Lee, Lincoln, Ogden, PI437.654, Palmeto, Peking, Perry, Richland, S100, Valiant, Volstate and Wayne. The insertion amplicons were amplified using specific primers listed in Supplementary Table S1. The forward primer was in the retroelement region (348–369 nucleotides downstream from the initiation ATG of GmNFR5β) and the reverse primer was in the transmembrane domain (2,184–2,204 nucleotides downstream of the ATG). PCR amplification was performed in a 25 μl volume containing 50 ng of genomic DNA, 10× PCR buffer, 2.5 mM MgCl2, 200 mM dNTPs, 0.5 U of Taq DNA polymerase recombinant (Invitrogen) and 0.2 μM of both forward and reverse oligonucleotide primers. Samples were heated to 94°C for 2 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 59°C for 30 s, elongation at 68°C for 2 min and a final extension at 68°C for 10 min. Amplified products were separated by electrophoresis in 1% agarose gels (Scientifix, Cheltenham, VIC, Australia) in 1× TAE buffer and were detected by fluorescence under UV light (302 nm) using the BioDoc-It™ Imaging System. Bioinformatics analysis Sequence analysis of GmNFR5α and GmNFR5β was performed using: BCM Search Launcher (http://searchlauncher.bcm.tmc.edu/multi-align/multi-align.html), ENTREZ/BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi), BOXSHADE (http://www.ch.embnet.org/software/BOX_form.html), SPIDEY (http://www.ncbi.nlm.nih.gov/spidey/), ClustalW (http://www.ebi.ac.uk/clustalw/), Gene structure and promoter predictions (http://linux1.softberry.com/berry.phtml?topic=fgenesh&group=programs&subgroup=gfind), FGENESH program (http://linux1.softberry.com/berry.phtml?topic=fgenesh&group=programs&subgroup=gfind), SMART (http://smart.embl-heidelberg.de/), Hydrophobic analysis (http://www.cbs.dtu.dk/services/SignalP/), Transmembrane prediction (http://ccb.imb.uq.edu.au/svmtm/svmtm_predictor .shtml), Map Viewer (http://www.ncbi.nlm.nih.gov/mapview/maps.cgi?TAXID=3847&MAPS=default&CHR), Primer3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3.cgi) and Phytozome (http://www.phytozome.net/soybean). Hairy root transformation To confirm that the mutations in the NFR5 genes were responsible for the non-nodulation phenotype in nod139 and nn5, complementation by hairy root transformation (Kereszt et al. 2007) was carried out. The wild-type clones of NFR5α and NFR5β genes derived from the genome of G. soja CPI100070, driven either by their own respective promoters (putatively selected at 548 bp for GmNFR5α and at 1,203 bp for GmNFR5β), or by the cauliflower mosaic virus (CaMV) 35S promoter, were cloned into binary vector pCAMBIA1305.1 and introduced into mutant plants via A. rhizogenes cucumopine strain K599. Nodule numbers of transgenic roots were determined 35 d after inoculation. RNA extraction and qRT-PCR For the qRT-PCR experiment, RNA samples were isolated from the leaves and roots of 2-week-old inoculated (1 week post-inoculation) and uninoculated wild-type Bragg, nod139 and nts1007 plants grown in a glasshouse, as well as from nodules collected 2 weeks after inoculation. More detailed expression studies were conducted on root segments, trifoliate leaves and shoot tips of uninoculated and inoculated Bragg plants. The inoculated plants were treated with a 4-day-old YMB culture of B. japonicum strain CB1809 immediately after sowing. Germination and seedling growth were conducted in an air-conditioned glasshouse maintained at day and night temperatures of 28 and 23°C, with a 16 h photoperiod. Taproot segments (TR), lateral root segments (LR), trifoliate leaves (TF) and shoot tips (STIP) were harvested from inoculated and uninoculated Bragg 14 d post-sowing. For collection of the taproot segments, the first 2 cm of the root tips was harvested as taproot 1 (TR1), and the adjoining 2 cm segments were taken as taproot 2 (TR2) and taproot 3 (TR3), respectively. Similarly three fragments collected from one lateral root were correspondingly labeled as lateral root 1 (LR1), lateral root 2 (LR2) and lateral root 3 (LR3). RNA extraction was performed using the NucleoSpin® RNA Plant Kit (Macherey Nagel, Düren, Germany) according to the manufacturer’s instructions. RNA concentrations were determined using a NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). First-strand cDNA was synthesized from 1 μg of total RNA using SuperscriptIII Reverse Transcriptase (Invitrogen) according to the manufacturer’s instructions. The cDNA templates were diluted and added to SYBR green PCR master mix (Applied Biosystems , Foster City, CA, U.S.A). containing 0.2 μmol of each primer (Table 1). qRT-PCR was carried out using an ABI PRISM 7900HT thermocycler (Applied Biosystems) under the following conditions: 95°C for 10 min followed by 35 cycles of 95°C for 15 s and 60°C for 1 min. A heat dissociation profile from 60 to 95°C was conducted at the end of each PCR to verify the specificity of the reaction. Individual values for each sample were generated by averaging data from two biological and two technical replicates. Primers for qRT-PCR were designed using Primer Express (Applied Biosystems , Foster City, CA, U.S.A.). Histochemical localization of GUS activity The promoter of both genes (880 bp for GmNFR5α and 970 bp for GmNFR5β) was integrated in front of the GUS reporter gene in vector pCAMBIA1305.1 and transformed into wild-type line Bragg and its supernodulating mutant nts1007 with the help of A. rhizogenes-mediated transformation. GUS activity was determined histochemically as described earlier (Nontachaiyapoom et al. 2007). Hairy roots were first vacuum infiltrated then stained overnight at 37°C in staining solution containing 0.1 mM X-Gluc, 100 mM sodium phosphate buffer (pH 7), 5 mM EDTA, 1 mM K4Fe(CN)6, 1 mM K3Fe(CN)6 and 0.1% (v/v) Triton X-100. For sectioning, tissues were fixed in 0.5% paraformaldehyde in 100 mM sodium phosphate buffer (pH 7) on ice under vacuum for 45 min, rinsed three times in 100 mM sodium phosphate buffer (pH 7) at room temperature, stained in X-Gluc solution, embedded in 3% agarose, and sectioned to 100 μm thickness with a vibratome (Lancer Series 1000; Ted Pella, Irvine, CA, USA). Statistical analysis Analysis of variance (ANOVA) was used to identify if there were significant effects of the treatments (empty vector, native promoter and 35S promoter) on the variables: nodule number per plant, nodule number per root and nodule number per root dry weight. Where significant effects were found, the least significant difference separation procedure was used to separate the differences. Funding This work was supported by the Australian Research Council [Centre of Excellence grant CEO34821]; the Queensland State Government [Smart State Innovation Scheme]; the University of Queensland Strategic Fund; AusAID [PhD scholarship to A.I.]; Sugar Research and Development Corporation (SRDC) [scholarship to I.S.]. Acknowledgments We thank Professor Randall Nelson, USDA-ARS and University of Illinois, USA for providing seeds of CNS, Richland, Volstate, Dunfield and non-nodulation mutant nn5. Soybean seeds lines for retroelement detection were generously provided by Dr Sally Dillon and Dr P. K. Lawrence from Australian Tropical Grains Germplasm Centre. We also thank CILR staff for help and suggestions. Abbreviations Abbreviations AON autoregulation of nodulation BAC bacterial artificial chromosome CaMV cauliflower mosaic virus GUS β-glucuronidase LYK LysM receptor kinase NFR Nod factor receptor qRT-PCR quantitative real-time PCR SNP single nucleotide polymorphism SSR simple sequence repeat UTR untranslated region. References Arrighi JF,  Barre A,  Ben Amor B,  Bersoult A,  Soriano LC,  Mirabella R, et al.  The Medicago truncatula lysine motif-receptor-like kinase gene family includes NFP and new nodule-expressed genes,  Plant Physiol. ,  2006, vol.  142 (pg.  265- 279) Google Scholar CrossRef Search ADS PubMed  Ben-Amor B,  Shaw SL,  Oldroyd GE,  Maillet F,  Penmetsa RV,  Cook D, et al.  The NFP locus of Medicago truncatula controls an early step of Nod-factor signal transduction upstream of a rapid calcium flux and root hair deformation,  Plant J. ,  2003, vol.  34 (pg.  495- 506) Google Scholar CrossRef Search ADS PubMed  Bersoult A,  Camut S,  Perhald A,  Kereszt A,  Kiss GB,  Cullimore JV.  Expression of the Medicago truncatula DMI2 gene suggests roles of the symbiotic nodulation receptor kinase in nodules and during early nodule development,  Mol. Plant-Microbe Interact. ,  2005, vol.  18 (pg.  869- 876) Google Scholar CrossRef Search ADS PubMed  Carroll BJ,  Mcneil DL,  Gresshoff PM.  Isolation and properties of soybean [Glycine max (L.) Merr] mutants that nodulate in the presence of high nitrate concentrations,  Proc. Natl Acad. Sci. USA ,  1985, vol.  82 (pg.  4162- 4166) Google Scholar CrossRef Search ADS   Carroll BJ,  Mcneil DL,  Gresshoff PM.  A supernodulation and nitrate-tolerant symbiotic (Nts) soybean mutant,  Plant Physiol. ,  1985, vol.  78 (pg.  34- 40) Google Scholar CrossRef Search ADS PubMed  Carroll BJ,  McNeil DL,  Gresshoff PM.  Mutagenesis of soybean [Glycine max (L.) Merr.] and the isolation of nonnodulating mutants,  Plant Sci. ,  1986, vol.  47 (pg.  109- 114) Google Scholar CrossRef Search ADS   Cregan PB,  Mudge J,  Fickus EW,  Marek LF,  Danesh D,  Denny R, et al.  Targeted isolation of simple sequence repeat markers through the use of bacterial artificial chromosomes,  Theor. Appl. Genet. ,  1999, vol.  98 (pg.  919- 928) Google Scholar CrossRef Search ADS   Francisco PB,  Akao S.  Autoregulation and nitrate inhibition of nodule formation in soybean cv. Enrei and its nodulation mutants,  J. Exp. Bot. ,  1993, vol.  44 (pg.  547- 553) Google Scholar CrossRef Search ADS   Geurts R,  Heidstra R,  Hadri AE,  Downie JA,  Franssen H,  van Kammen AB, et al.  Sym2 of pea is involved in a nodulation factor-perception mechanism that controls the infection process in the epidermis,  Plant Physiol. ,  1997, vol.  115 (pg.  351- 359) Google Scholar CrossRef Search ADS PubMed  Gizlice Z,  Carter TEJr,  Burton JW.  Genetic base for North American public soybean cultivars released between 1947 and 1988,  Crop Sci. ,  1994, vol.  34 (pg.  1143- 1151) Google Scholar CrossRef Search ADS   Gresshoff PM,  Landau-Ellis D.  Gresshoff P.  Molecular mapping of soybean nodulation genes,  Plant Genome Analysis: Current Topics in Plant Molecular Biology ,  1994 Boca Raton, FL CRC Press(pg.  97- 112) Ikeda S,  Rallos LEE,  Okubo T,  Eda S,  Inaba S,  Mitsui H,  Minamisawa K.  Microbial community analysis of field-grown soybeans with different nodulation phenotypes,  Appl. Environ. Microbiol. ,  2008, vol.  74 (pg.  5704- 5709) Google Scholar CrossRef Search ADS PubMed  Indrasumunar A.  Molecular cloning and functional characterisation of soybean (Glycine max L.) nod factor receptor genes,  2007 Australia The University of Queenslandpg.  162   PhD Thesis Jarvik T,  Lark KG.  Characterization of soymar1, a mariner element in soybean,  Genetics ,  1998, vol.  149 (pg.  1569- 1574) Google Scholar PubMed  Kereszt A,  Dongxue L,  Indrasumunar A,  Nguyen CDT,  Nontachaiyapoom S,  Kinkema M, et al.  Agrobacterium rhizogenes-mediated transformation of soybean to study root biology,  Nature Protocols ,  2007, vol.  2 (pg.  948- 952) Google Scholar CrossRef Search ADS PubMed  Kolchinsky A,  LandauEllis D,  Gresshoff PM.  Map order and linkage distances of molecular markers close to the supernodulation (nts-1) locus of soybean,  Mol. Gen. Genet. ,  1997, vol.  254 (pg.  29- 36) Google Scholar CrossRef Search ADS PubMed  Kunze R,  Saedler H,  Lonnig WE.  Plant transposable elements,  Adv. Bot. Res. ,  1997, vol.  27 (pg.  331- 470) Laten HM,  Morris RO.  SIRE-1, a long interspersed repetitive DNA element from soybean with sequence similarity to retrotransposons: initial characterization and partial sequence,  Genetics ,  1993, vol.  134 (pg.  153- 159) Laten HM,  Majumdar A,  Gaucher EA.  SIRE-1, a copia/Ty1-like retroelement from soybean, encodes a retroviral envelope-like protein,  Proc. Natl Acad. Sci. USA ,  1998, vol.  95 (pg.  6897- 6902) Google Scholar CrossRef Search ADS   Limpens E,  Franken C,  Smit P,  Willemse J,  Bisseling T,  Geurts R.  LysM domain receptor kinases regulating rhizobial Nod-factor-induced infection,  Science ,  2003, vol.  302 (pg.  630- 633) Google Scholar CrossRef Search ADS PubMed  Madsen EB,  Madsen LH,  Radutoiu S,  Olbryt M,  Rakwalska M,  Szcyglowski K, et al.  A receptor-like kinase gene involved in legume perception of rhizobial signal molecules,  Nature ,  2003, vol.  425 (pg.  637- 640) Google Scholar CrossRef Search ADS PubMed  Markmann K,  Parniske M.  Evolution of root endosymbiosis with bacteria: how novel are nodules?,  Trends Plant Sci. ,  2009, vol.  14 (pg.  77- 86) Google Scholar CrossRef Search ADS PubMed  Mathews A,  Carroll BJ,  Gresshoff PM.  A new non-nodulation gene in soybean,  J. Hered. ,  1989, vol.  80 (pg.  357- 360) Google Scholar CrossRef Search ADS   Mathews A,  Carroll BJ,  Gresshoff PM.  Development of Bradyrhizobium infections in supernodulating and non-nodulating mutants of soybean (Glycine max (L.) Merrill),  Protoplasma ,  1989, vol.  150 (pg.  40- 47) Google Scholar CrossRef Search ADS   Mathews A,  Carroll BJ,  Gresshoff PM.  Studies on the root control of non-nodulation and plant growth of non-nodulating mutants and supernodulating mutant of soybean (Glycine max (L.) Merr.),  Plant Sci. Limerick ,  1992, vol.  83 (pg.  35- 43) Google Scholar CrossRef Search ADS   Miyahara A,  Hirani TA,  Oakes M,  Kereszt A,  Kobe B,  Djordjevic MA, et al.  Soybean nodule autoregulation receptor kinase phosphorylates two kinase-associated protein phosphatases in vitro,  J. Biol. Chem. ,  2008, vol.  283 (pg.  25381- 25391) Google Scholar CrossRef Search ADS PubMed  Nontachaiyapoom S,  Scott PT,  Men AE,  Kinkema M,  Schenk PM,  Gresshoff PM.  Promoters of orthologous Glycine max and Lotus japonicus nodulation autoregulation genes interchangeably drive phloem-specific expression in transgenic plants,  Mol. Plant-Microbe Interact. ,  2007, vol.  20 (pg.  769- 80) Google Scholar CrossRef Search ADS PubMed  Oldroyd GE,  Downie JA.  Coordinating nodule morphogenesis with rhizobial infection in legumes,  Annu. Rev. Plant Biol. ,  2008, vol.  59 (pg.  519- 46) Google Scholar CrossRef Search ADS PubMed  Palmer RG,  Kilen TC.  Wilcox JR.  Qualitative genetics and cytogenetics,  Soybean: Improvement, Production and Uses, 2nd edn. Agronomy Monograph 16 ,  1987 Madison, WI American Society of Agronomy, Crop Science Society of America, Soil Science Society of America(pg.  23- 48) Pracht JE,  Nickell CD,  Harper JE.  Genes controlling nodulation in soybean: Rj5 and Rj6,  Crop Sci. ,  1993, vol.  33 (pg.  711- 713) Google Scholar CrossRef Search ADS   Radutoiu S,  Madsen LH,  Madsen EB,  Felle HH,  Umehara Y,  Gronlund M,  Sato S, et al.  Plant recognition of symbiotic bacteria requires two LysM receptor-like kinases,  Nature ,  2003, vol.  425 (pg.  585- 592) Google Scholar CrossRef Search ADS PubMed  Radutoiu S,  Madsen LH,  Madsen EB,  Jurkiewicz A,  Fukai E,  Quistgaard EMH, et al.  LysM domains mediate lipochitin–oligosaccharide recognition and NFR genes extend the symbiotic host range,  EMBO J. ,  2007, vol.  26 (pg.  3923- 3935) Google Scholar CrossRef Search ADS PubMed  Rhodes PJ,  Vodkin LO.  Organization of the Tgm family of transposable elements in soybean,  Genetics ,  1988, vol.  120 (pg.  597- 604) Google Scholar PubMed  Schlueter JA,  Lin J-Y,  Schlueter SD,  Vasylenko-Sanders IF,  Deshpande S,  Yi J, et al.  Gene duplication and paleopolyploidy in soybean and the implications for whole genome sequencing,  BMC Genomics ,  2007, vol.  8 pg.  330  Google Scholar CrossRef Search ADS PubMed  Schlueter JA,  Vasylenko-Sanders IF,  Deshpande S,  Yi J,  Siegfried M,  Roe BA, et al.  The FAD2 gene family of soybean: insights into the structural and functional divergence of a paleopolyploid genome,  Plant Genome ,  2007, vol.  1 (pg.  14- 26) Schmutz J,  Cannon SB,  Schlueter J,  Ma J,  Mitros T,  Nelson W, et al.  Genome sequence of the paleopolyploid soybean,  Nature ,  2009  in press Searle IR. ,  Towards the map-based cloning of nodulation genes from soybean ,  2001 Australia School of Molecular and Microbial Sciences: The University of Queensland Searle IR,  Men AE,  Laniya TS,  Buzas DM,  Iturbe-Ormaetxe I,  Carroll B, et al.  Long-distance signaling in nodulation directed by a CLAVATA1-like receptor kinase,  Science ,  2003, vol.  299 (pg.  109- 112) Google Scholar CrossRef Search ADS PubMed  Shoemaker RC,  Polzin K,  Labate J,  Specht J,  Brummer EC,  Olson T, et al.  Genome duplication in soybean (Glycine subgenus soja),  Genetics ,  1996, vol.  144 (pg.  329- 338) Google Scholar PubMed  Stacey G,  Vodkin L,  Parrott WA,  Shoemaker RC.  National Science Foundation-sponsored workshop report. Draft plan for soybean genomics 1,  Plant Physiol. ,  2004, vol.  135 (pg.  59- 70) Google Scholar CrossRef Search ADS PubMed  Tomkins JP,  Mahalingam R,  Smith H,  Goicoechea JL,  Knap HT,  Wing RA.  A bacterial artificial chromosome library for soybean PI 437654 and identification of clones associated with cyst nematode resistance,  Plant Mol. Biol. ,  1999, vol.  41 (pg.  25- 32) Google Scholar CrossRef Search ADS PubMed  Vodkin LO,  Rhodes P,  Goldberg RB.  A lectin gene insertion has the structural features of a transposable element,  Cell ,  1983, vol.  34 (pg.  1023- 1031) Google Scholar CrossRef Search ADS PubMed  Williams LF,  Lynch DL.  Inheritance of a nonnodulating character in the soybean,  Agron. J. ,  1954, vol.  46 (pg.  28- 29) Google Scholar CrossRef Search ADS   Zhang X-C,  Wu X,  Findley S,  Wan J,  Libault M,  Nguyen,  H T, et al.  Molecular evolution of lysine motif-type receptor-like kinases in plants,  Plant Physiol. ,  2007, vol.  144 (pg.  623- 636) Google Scholar CrossRef Search ADS PubMed  Zhukov V,  Radutoiu S,  Madsen LH,  Rychagova T,  Ovchinnikova E,  Borisov A, et al.  The pea Sym37 receptor kinase gene controls infection-thread initiation and nodule development,  Mol. Plant-Microbe Interact. ,  2008, vol.  21 (pg.  1600- 1608) Google Scholar CrossRef Search ADS PubMed  © The Author 2009. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: [email protected] http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Plant and Cell Physiology Oxford University Press

Inactivation of Duplicated Nod Factor Receptor 5 (NFR5) Genes in Recessive Loss-of-Function Non-Nodulation Mutants of Allotetraploid Soybean (Glycine max L. Merr.)

Download PDF
14 pages

Loading next page...
 
/lp/oxford-university-press/inactivation-of-duplicated-nod-factor-receptor-5-nfr5-genes-in-T8CKiEAh8Q

References (52)

Publisher
Oxford University Press
Copyright
© The Author 2009. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: [email protected]
ISSN
0032-0781
eISSN
1471-9053
DOI
10.1093/pcp/pcp178
pmid
20007291
Publisher site
See Article on Publisher Site

Abstract

Abstract Chemically induced non-nodulating nod139 and nn5 mutants of soybean (Glycine max) show no visible symptoms in response to rhizobial inoculation. Both exhibit recessive Mendelian inheritance suggesting loss of function. By allele determination and genetic complementation in nod139 and nn5, two highly related lipo-oligochitin LysM-type receptor kinase genes in Glycine max were cloned; they are presumed to be the critical nodulation-inducing (Nod) factor receptor similar to those of Lotus japonicus, pea and Medicago truncatula. These duplicated receptor genes were called GmNFR5α and GmNFR5β. Nonsense mutations in GmNFR5α and GmNFR5β were genetically complemented by both wild-type GmNFR5α and GmNFR5β in transgenic roots, indicating that both genes are functional. Both genes lack introns. In cultivar Williams82 GmNFR5α is located in chromosome 11 and in tandem with GmLYK7 (a related LysM receptor kinase gene), while GmNFR5β is in tandem with GmLYK4 in homologous chromosome 1, suggesting ancient synteny and regional segmental duplication. Both genes are wild type in G. soja CPI100070 and Harosoy63; however, a non-functional NFR5β allele (NFR5β*) was discovered in parental lines Bragg and Williams, which harbored an identical 1,407 bp retroelement-type insertion. This retroelement (GmRE-1) and related sequences are located in several soybean genome positions. Paradoxically, putatively unrelated soybean cultivars shared the same insertion, suggesting a smaller than anticipated genetic base in this crop. GmNFR5α but not GmNFR5β* was expressed in inoculated and uninoculated tap and lateral root portions at about 10–25% of GmATS1 (ATP synthase subunit 1), but not in trifoliate leaves and shoot tips. Introduction Many legume species fulfill their nitrogen need through a symbiotic relationship with their microsymbiont, commonly called the ‘rhizobium’, leading to atmospheric nitrogen fixation and nitrogen gain. This symbiotic ability contributes to the global application of legumes (i.e. soybean, pea, bean, peanuts, chickpea and forages) for human food and animal feed supply. The legume symbiosis displays host strain specificity, i.e. only certain combinations of plant and microsymbiont are compatible for establishment of the nitrogen-fixing symbiosis (Markmann and Parniske 2009). Host specificity is determined by a complex set of signal exchanges, as plants secrete specific (iso)flavonoid ‘cocktails’ that induce the biosynthesis of bacterial-controlled lipochitin–oligosaccharides (Nod factors) that act as a plant morphogen and mitogenic factor after perception by specific plant receptors (Oldroyd and Downie 2008). Spontaneous mutants showing an altered nodulation pattern have been isolated in several legumes, e.g. rj1 in soybean (Williams and Lynch 1954) and sym2 in pea (Geurts et al. 1997). However, it was induced mutations and subsequent positional cloning of nodulation genes in Lotus japonicus and Medicago truncatula that helped develop a working paradigm of nodule initiation and downstream signaling pathways. Specifically, putative Nod factor receptor genes [namely those encoding Nod factor receptor 1 (NFR1) and 5 (NFR5)] were identified in L. japonicus (Madsen et al. 2003, Radutoiu et al. 2003), M. truncatula [for MtLYK3 and MtNFP (Ben Amor et al. 2003, Limpens et al. 2003)] and Pisum sativum [PsSym10 (Geurts et al. 1997, Madsen et al. 2003)]. All encoded LysM-type serine/threonine receptor kinases characterized by 2–3 more or less conserved LysM domains in the N-terminal region, a transmembrane domain and a highly conserved intracellular kinase domain. They belong to a larger receptor family that includes chitin receptor, related to plant pathogen perception (Zhang et al. 2007). Lotus japonicus NFR1 and NFR5 were suggested to play a role in host determination, transforming the non-hosts M. truncatula and Lotus filicaulis into hosts for Mesorhizobium loti (Radutoiu et al. 2007). Carroll et al. (1986) described the isolation of three non-nodulation soybean mutants (nod49, nod139 and nod772) in cultivar Bragg following ethylmethane sulfonate mutagenesis. nod49 and nod772 formed a complementation group with the naturally occurring mutant rj1, whereas nod139 was mutated in a separate gene (Mathews et al. 1989a). Another soybean non-nodulation mutant (nn5 in cultivar Williams) is allelic to nod139; independently the locus has been assigned the rj6 designation (Pracht et al. 1993). nod139 and nn5 fail to show the earliest morphological changes such as deformation and curling of root hairs in response to rhizobial inoculation and consequently fail to develop nodules. However, nod139 established a symbiotic interaction with mycorrhizal fungi (Mathews et al. 1989b), indicating that the mutation affected an early stage of nodule initiation. Back-crossing of nod139 or nn5 to their parent varieties Bragg or Williams, respectively, demonstrated only one functional wild-type allele at the nod139 locus, but crossing and progeny analysis between nod139 and Glycine soja (CPI100070 and PI468.397), or between nn5 and G. max Harosoy63 identified two unlinked functional wild-type loci which complemented the nod139 and nn5 mutations (Gresshoff and Landau-Ellis 1994, Pracht et al. 1993, Searle, 2001). Thus varieties Bragg and Williams appear to harbor one functional wild-type copy of the nod139 locus, whereas G. soja accessions and variety Harosoy63 have two functional copies, reflecting the ancient allotetraploid nature of soybean (Schlueter et al. 2007a). The loci were named nod139-1 and nod139-2 (Searle 2001), and correspond to the dominant wild-type Rj5 and Rj6 alleles reported by Pracht et al. (1993). The soybean genome is considered to be paleopolyploid, resulting from sequential genome duplications about 50 as well as 13 Mya (Schlueter et al. 2007a, Schmutz et al. 2009). Duplication occurred in segmental regions, accompanied by local genomic as well as sequence divergence. For example, the soybean autoregulation of nodulation (AON) receptor kinase gene GmNARK (Searle et al. 2003) is located on soybean chromosome 12 in a homeologous region of GmCLV1A (on chromosome 11). These genes share 91.5% of their nucleotide sequence, are both transcriptionally active (with a seemingly functional open reading frame), but differ in apparent function, as GmNARK mutations are recessive, and not complemented by GmCLV1A. Similar observations came from the comparison of duplicated regions of the FAD2 gene in soybean (Schlueter et al. 2007b). One concludes that duplicated sequences provide the opportunity for sequence divergence and possible metabolic, temporal and spatial specificity. As mutations in genes coding for LysM-type receptor kinases resulted in a similar Nod− Myc+ phenotype (Ben Amor et al. 2003, Limpens et al. 2003, Madsen et al. 2003, Radutoiu et al. 2003), and the non-allelic non-nodulating nod49 mutant has been shown to be impaired in NFR1 function (Indrasumunar 2007), a combined candidate gene approach involving mapping, allele sequencing, complementation and overexpression was used to clone the soybean NFR5 genes conditioning a non-nodulation phenotype in nod139 and nn5. Results Genetic mapping of nod139 loci Non-nodulation mutant nod139 (in the cv. Bragg background) was crossed to two wild soybean accessions, G. soja CPI100070 and PI468.397 (both capable of nodulation), and F1 plants were allowed to self-fertilize to produce two F2 mapping populations. F2 segregation of the nodulation phenotype was 15 : 1 (nod+ vs. nod−) in both populations, indicating two unlinked functional copies of nod139 in both G. soja parents. These results demonstrated that, as in soybean variety Harasoy63 (Pracht et al. 1993) and G. soja PI468.397 (Gresshoff and Landau-Ellis 1994), there are two functional copies of the nod139 gene in G. soja, referred to as nod139-1 and nod139-2. Using bulk segregant analysis and simple sequence repeats (SSRs), nod139-2 was mapped to chromosome 1 near markers Satt408, Sat036 and Satt071. Analysis of 75 F2 progeny, 15 from the nod139 × PI468.397 mapping population and 60 from the nod139 × CPI100070 mapping population, demonstrated that Satt408 is 8.6 cM from nod139-2 and Sat_036 is on the other side of the locus, 10.6 cM away. Satt071 was only polymorphic in the nod139 × PI468.397 mapping population and could only be loosely mapped on 15 non-nodulating F2 plants. No recombination events were identified between nod139 and Satt071, thereby indicating that Satt071 was within 3 cM of nod139-2. To identify the location of nod139-1, markers Satt509 and Satt426 were selected from the homologous chromosomal region to nod139-2 on chromosome 11. To construct a genetic map around nod139-1, Satt426 and Satt509 were mapped on 60 F2 non-nodulating plants from the nod139 × CPI100070 mapping population. Satt509 and Satt426 were on the same side of nod139-1, with Satt426 2.5 cM and Satt509 3.3 cM away from the locus. Markers Satt071, Satt426 and Satt509 were too far from the nod139-1 and nod139-2 loci to be considered as anchors for ‘chromosome walks’. However, functionally (i.e. affecting very early nodulation but not mycorrhizal symbiotic steps), a likely candidate for the nod139 loci was NFR5, as its presumed molecular partner, namely the LysM-type receptor kinase GmNFR1, was already defined by the nod49/rj1 mutations (unlinked to nod139/nn5) (Indrasumunar 2007). Accordingly we initiated a candidate gene isolation approach using bacterial artificial chromosome (BAC) cloning, allele sequencing and genetic complementation. Isolation of BAC clones carrying the NFR5 genes of soybean Primers designed in accordance with the conserved regions in orthologous NFR5 sequences of Lotus, Medicago and pea successfully amplified two DNA fragments with sizes of around 1.0 and 2.5 kb from genomic DNA of cultivar Bragg. The expected size of the PCR product based on the LjNFR5 sequence was 1,049 bp (NB: the larger PCR fragment stemmed from GmNFR5β* that was discovered later to contain a 1,407 bp retroelement insertion GmRE-1). At first only the 1.0 kb PCR product was cloned and sequenced for comparison with LjNFR5, showing 81% identity at the nucleotide level; it was chosen as the putative GmNFR5 probe to screen the BAC library of wild-type soybean variety PI437.654 (Tomkins et al. 1999). BAC library screening resulted in 25 positively hybridizing BAC clones, but, after checking these by PCR, only 15 were positive. Six positive BAC clones (88H1, 138J20, 153C20, 167H3, 69M16 and 179K5) were chosen for further analysis and BAC sequencing. Isolation of the putative GmNFR5α and GmNFR5β genes Sequencing of six positive BAC clones by primer walking revealed two highly related (95% similarity at the nucleotide level) DNA stretches similar to LjNFR5, named GmNFR5α and GmNFR5β (Supplementary Fig. S1). The presence of two genomic copies is in agreement with the duplicated nature of the soybean genome (Shoemaker et al. 1996, and arguments above). As there are two functional copies of nod139 (Gresshoff and Landau-Ellis 1994, Searle 2001), the two highly related sequences were assumed to represent these loci, and were considered to be good candidates for the location of the nod139/nn5 mutations. Gene and protein characteristics of GmNFR5α and GmNFR5β DNA sequences 871 and 1,580 nucleotides upstream of the GmNFR5α and GmNFR5β putative translation start sites, respectively, were obtained from BAC clones (69M16, 138J20 and 179K5 for GmNFR5α, and 88H1, 153C20 and 167H3 for GmNFR5β). The 3′ untranslated regions (UTRs) of 807 nucleotides (GmNFR5α) and 367 nucleotides (GmNFR5β) were also sequenced from these BAC clones. Analysis of the genomic sequences using the FGENESH program defined gene structures without introns, as outlined in Supplementary Fig. S1. Both predicted genes consisted of a single exon with 1,794 and 1,797 nucleotides of coding sequence for GmNFR5α and GmNFR5β, respectively. Analysis of the GmNFR5α and GmNFR5β gene sequences through the SMART website and comparison with LjNFR5 (Madsen et al. 2003) predicted a typical receptor kinase protein structure consisting of an extracellular receptor domain (containing three LysM domains) with an N-terminal signal peptide, a transmembrane domain and a cytoplasmic kinase domain (Fig. 1). Fig. 1 View largeDownload slide Domain structure and sequence alignment of NFR5 proteins. The predicted amino acid sequences of the soybean NFR5α and NFR5β proteins were compared with each other and with that of the corresponding sequence from L. japonicus, P. sativum and M. truncatula. Conserved amino acids are highlighted. Key protein domains (SP, signal peptide, LysM 1, 2, 3, LysM domains; TM, transmembrane domain; KD, kinase domain) are underlined. The nonsense mutation in GmNFR5α and the transposon insertion in GmNFR5β are indicated with a star and an arrowhead, respectively. Kinase catalytic subdomains are indicated with Roman characters. Fig. 1 View largeDownload slide Domain structure and sequence alignment of NFR5 proteins. The predicted amino acid sequences of the soybean NFR5α and NFR5β proteins were compared with each other and with that of the corresponding sequence from L. japonicus, P. sativum and M. truncatula. Conserved amino acids are highlighted. Key protein domains (SP, signal peptide, LysM 1, 2, 3, LysM domains; TM, transmembrane domain; KD, kinase domain) are underlined. The nonsense mutation in GmNFR5α and the transposon insertion in GmNFR5β are indicated with a star and an arrowhead, respectively. Kinase catalytic subdomains are indicated with Roman characters. The GmNFR5α and GmNFR5β genes encode predicted proteins of 598 and 599 amino acids, respectively, compared with 596 amino acids for LjNFR5. At the N-terminus, a hydrophobic stretch of 28 amino acids is predicted to act as a transit peptide, compared with 26 amino acids for LjNFR5 (Fig. 1). The transit domain is followed by three motifs of a LysM domain (Fig. 1; Supplementary Fig. S2A) with similarity to the LysM domains found in bacterial peptidoglycan-binding proteins and chitinases from yeast (Kluyveromyces lactis) and algae (Volvoc carteri; Madsen et al. 2003). A predicted transmembrane region occurs between the extracellular domain and the C-terminal cytoplasmic kinase domain (Fig. 1). The GmNFR5α and GmNFR5β kinase domains have motifs associated with functional serine/threonine kinases, but motifs VII and VIII may be strongly modified or absent (Supplementary Fig. S2B). Specifically, an aspartic acid residue commonly conserved in domain VII is missing, and domain VIII, comprising the activation loop, is absent. The NFR5 proteins of soybean shared 71–74% overall amino acid sequence identity with the Medicago, Lotus and pea sequences (Fig. 1). The sequence identity was higher (79–82%) in the transmembrane/kinase domains and lower (64–67%) in the extracellular domain, which is believed to be responsible for ligand binding and thus determination of host range (Radutoiu et al. 2007). nod139 and nn5 carry nonsense mutations in GmNFR5α Sequence characterization of GmNFR5α from nod139 and nn5 showed the presence of point mutations (Supplementary Fig. S1). Single nucleotide substitutions (nonsense) were identified within the GmNFR5α coding region of both mutants. The nod139 mutation was identified as a nonsense mutation occurring at nucleotide +1,013 from the translational start site (ATG), resulting in the conversion of a leucine codon to a nonsense codon at position 338 of the GmNFR5α protein (Fig. 1). This L338* mutation eliminates most of the kinase domain, presumably rendering the resultant protein non-functional. Allele nn5 also truncated the GmNFR5α protein by a nonsense mutation at +1,506 bp from the starting ATG, converting a tryptophan codon to a stop codon at position 502 (Fig. 1). A natural mutation of GmNFR5β in the parental lines Bragg and Williams Genetic analysis of nod139 and nn5 indicated that recessive alleles of the GmNFR5α/β genes were responsible for the non-nodulation phenotype. Both genes were functional in G. soja and Harosoy63, but one was non-functional in the parental lines Bragg and Williams (Gresshoff and Landau-Ellis 1994, Pracht et al. 1993). DNA sequencing of both G. soja CPI100070 GmNFR5 genes showed that both were wild type (open reading frame and expressed), confirming genetic segregation data. In contrast, nodulation-capable cultivars Bragg and Williams carried a 1,407 bp long insertion sequence (GmRE-1) at the same position of the GmNFR5β gene (Supplementary Fig. S3). The insertion has the characteristics of a non-autonomous retroelement (Kunze et al. 1997); it has long terminal repeats of 214 bp and a non-perfect duplication of the 11 bp target site. Retroelement insertion occurred at nucleotide +295 from the translational start site (ATG), resulting in the conversion of a leucine to a nonsense at position 99 (LysM1, extracellular domain) of the GmNFR5β protein. The resultant protein, if stable, would lack its kinase domain, transmembrane domain and most of the extracellular domain. Complementation of nod139 and nn5 Complementation of the two GmNFR5 mutant alleles was done by Agrobacterium rhizogenes ‘hairy root’ transformation (Kereszt et al. 2007). Chimeric root systems formed by K599 transformation are composed of both transformed and non-transformed roots, as about 50% of the roots formed in nod139 and nn5 plants failed to nodulate. This frequency is consistent with expression of 35SGUS (β-glucuronidase) in soybean chimeric plants. Such roots were not considered in further quantitative characterization. Future hairy root studies will benefit from the use of visually selective markers for co-transformation (such as the DsRED vector) giving red coloration to transformed roots. Roots without red color thus can be excised and eliminated from final data sets. The non-nodulation phenotype of mutants nod139 and nn5 (Fig. 2) was complemented by wild-type GmNFR5α and GmNFR5β genes by A. rhizogenes K599 transformation. Wild-type GmNFR5β was derived from G. soja CPI100070. These results also confirmed the root control of the non- nodulation phenotypes (Mathews et al. 1992). Nodulation parameters of complemented roots were similar (50–100 nodules per plant) and independent of the type of promoter used [i.e. the native promoters (548 and 1,203 bp, respectively) and the constitutive 35S promoter (Table 1)]. Notably, the 35S overexpressing roots [with 10–80 times higher levels of GmNFR5 mRNA than control levels as measured by quantititative real-time PCR (qRT-PCR)] nodulated normally, suggesting that nodule number was not limited by GmNFR5 RNA expression level. Additionally, reciprocal complementation with the GmNFR1α gene into mutants nod139 and nn5, and the GmNFR5α and GmNFR5β genes into mutants nod49 and rj1 did not result in nodulation, showing distinct functions for the two putative receptor components. Fig. 2 View largeDownload slide Complementation of the nod139 and nn5 mutation. Transformed root systems were scored 35 d after inoculation with B. japonicum CB1809. (A) Transgenic roots of nod139 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (B) Root system of nod139 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (C) Root system of nod139 transformed with K599 carrying full-length GmNFR5α driven by the 35S promoter of CaMV. (D) Individual hairy roots of nod139 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. (E) Nodulation of nod139 transformed with K599 carrying full-length GmNFR5β driven by the 35S promoter of CaMV. (F) Composite plants (on the right side) transformed with the empty vector show the growth- and nitrogen deficiency-related phenotype associated with the absence of nodulation, while individuals carrying the wild-type transgene have normal development (left side). (G) Transgenic roots of nn5 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (H) Root system of nn5 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (I) Root system of nn5 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. Bars represent 2 cm in A, B, C, G, H and I; 1.75 cm in D; 1 cm in E; 6 cm in F. Fig. 2 View largeDownload slide Complementation of the nod139 and nn5 mutation. Transformed root systems were scored 35 d after inoculation with B. japonicum CB1809. (A) Transgenic roots of nod139 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (B) Root system of nod139 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (C) Root system of nod139 transformed with K599 carrying full-length GmNFR5α driven by the 35S promoter of CaMV. (D) Individual hairy roots of nod139 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. (E) Nodulation of nod139 transformed with K599 carrying full-length GmNFR5β driven by the 35S promoter of CaMV. (F) Composite plants (on the right side) transformed with the empty vector show the growth- and nitrogen deficiency-related phenotype associated with the absence of nodulation, while individuals carrying the wild-type transgene have normal development (left side). (G) Transgenic roots of nn5 transformed with A. rhizogenes strain K599 carrying the empty vector pCAMBIA1305.1. (H) Root system of nn5 transformed with K599 carrying full-length GmNFR5α behind its own 548 bp native promoter. (I) Root system of nn5 transformed with K599 carrying full-length GmNFR5β behind its own 1,203 bp native promoter. Bars represent 2 cm in A, B, C, G, H and I; 1.75 cm in D; 1 cm in E; 6 cm in F. Table 1 Average number of nodules on transgenic roots induced on wild-type and mutant plants transformed with the GmNFR5α and GmNFR5β genes   Empty vector (no GmNFR5)a  GmNFR5α + native 548 bp promoter  GmNFR5α + 35S promoter  GmNFR5β + native 1,203 bp promoter  GmNFR5β + 35S promoter  Nodule number per plant  nod139  0 ab  86 cd  100 d  82 cd  91 d  Bragg  58 b  59 b  72 bc  61 bc  75 bc  Nodule number per root  nod139  0 a  22 c  28 c  20 c  25 c  Bragg  6 b  8 b  9 b  8 b  9 b  Nodule number per mg root dry weight  nod139  0.0 a  1.8 c  2.1 c  1.6 c  2.1 c  Bragg  0.3 b  0.5 b  0.5 b  0.5 b  0.5 b    Empty vector (no GmNFR5)a  GmNFR5α + native 548 bp promoter  GmNFR5α + 35S promoter  GmNFR5β + native 1,203 bp promoter  GmNFR5β + 35S promoter  Nodule number per plant  nod139  0 ab  86 cd  100 d  82 cd  91 d  Bragg  58 b  59 b  72 bc  61 bc  75 bc  Nodule number per root  nod139  0 a  22 c  28 c  20 c  25 c  Bragg  6 b  8 b  9 b  8 b  9 b  Nodule number per mg root dry weight  nod139  0.0 a  1.8 c  2.1 c  1.6 c  2.1 c  Bragg  0.3 b  0.5 b  0.5 b  0.5 b  0.5 b  aAt least 20 replicates for each treatment were scored 35 d after inoculation with B. japonicum strain CB1809. bNumbers followed by the same letter for the same measured parameter are not significantly different at the probability 0.05. View Large Expression analysis of the soybean NFR5 genes Using common qRT-PCR primers for both GmNFR5α and GmNFR5β, mRNA levels were 1.7–1.8% of GmActin5 levels in uninoculated roots of 2-week-old wild-type Bragg and supernodulating nts1007 plants, and slightly lower (1.2–1.4%) in mutant nod139 plants. In inoculated roots, expression slightly increased to 2% of that of GmActin5 in Bragg, but decreased to 1.3–1.4% in nts1007. These differences were not significant. However, expression in mature nodules and leaves was significantly lower (0.3–0.4% of GmActin5). More detailed analysis using specific primers for GmNFR5α and GmNFR5β RNA allowed insights into differential expression. The mRNA level of GmNFR5α was 10–25% of that of GmATS1 in uninoculated and inoculated portions of lateral and tap roots in 2-week-old wild-type Bragg (Fig. 3A). In contrast, the expression of GmNFR5β* (note: affected by the GmRE-1 insertion) was much lower than that of GmNFR5α, being <1% of that of GmATS1 (Fig. 3B). One presumes the transcript to be non-coding because of the insertion into the exon. In general, there is a down-regulation of GmNFR5α/β* transcript levels in inoculated tap root portions, though that effect may not be significant as lateral roots, responding strongly to inoculation for ENOD40-1 (Fig. 3C), had no response. Trifoliate leaves and the shoot tip region had extremely low or undetectable transcript levels for both genes. The presence of the retroelement in GmNFR5β* severely lowered the transcript level in all tissues. We conclude that expression of GmNFR5α and GmNFR5β* is predominantly root localized and marginally inoculation sensitive, as measured on this whole root section level. Cell type-specific regulation may have been swamped by non-responding basal root tissues. Fig. 3 View largeDownload slide Expression levels of GmNFR5α (A), GmNFR5β (B) and GmEnod40-1 (C) genes relative to that of GmATS1 (ATP synthase subunit 1) in tap root (TR), lateral root (LR), trifoliate leaves (TF) and shoot tips (STIP). The transcript level was determined by qRT-PCR. Section 1, 0–2 cm from the root tip; section 2, 2–4 cm from the root tip; section 3, 4–6 cm from the root tip. Data are the average of two technical repeats and three biological repeats. Error bars represent the standard error. Fig. 3 View largeDownload slide Expression levels of GmNFR5α (A), GmNFR5β (B) and GmEnod40-1 (C) genes relative to that of GmATS1 (ATP synthase subunit 1) in tap root (TR), lateral root (LR), trifoliate leaves (TF) and shoot tips (STIP). The transcript level was determined by qRT-PCR. Section 1, 0–2 cm from the root tip; section 2, 2–4 cm from the root tip; section 3, 4–6 cm from the root tip. Data are the average of two technical repeats and three biological repeats. Error bars represent the standard error. To visualize the expression of GmNFR5 genes in soybean roots and nodules, we cloned their native promoters in front of the GUS reporter gene in vector pCAMBIA1305.1, and transformed the reporter constructs into wild-type Bragg and its supernodulating mutant nts1007 via A. rhizogenes-mediated transformation (Kereszt et al. 2007). The expression patterns determined by the two promoters were similar and were not affected by the genotype of the transformed plant. In uninoculated roots, GUS activity was observed in most root tissues (Fig. 4A); however, the intensity of the staining decreased with the age of the root. In inoculated root, the expression of the GUS reporter gene increased during the early steps of nodule development (see arrows in Fig. 4B, C), but decreased in mature nodules (Fig. 4D). Fig. 4 View largeDownload slide Localization of NFR5 gene expression in soybean roots. (A) Cross-section of an uninoculated transgenic root. (B) Cross-section of an inoculated root with a developing side root. The arrow indicates a bacterial infection site with elevated GUS expression. (C) A hairy root developed on the supernodulation mutant nts1007 with nodules at different developmental stages. The primary root and the mature nodules show weak staining which is more intense in the younger tissues. Arrows show enhanced expression associated with nodule initiation which is also observed in developing nodules (arrowhead). (D) Cross-section of a nodule and the root shown in C; GUS expression can be observed in the cortex. Bars represent 300 μm in A and B; 3 mm in C; 1 mm in D. Fig. 4 View largeDownload slide Localization of NFR5 gene expression in soybean roots. (A) Cross-section of an uninoculated transgenic root. (B) Cross-section of an inoculated root with a developing side root. The arrow indicates a bacterial infection site with elevated GUS expression. (C) A hairy root developed on the supernodulation mutant nts1007 with nodules at different developmental stages. The primary root and the mature nodules show weak staining which is more intense in the younger tissues. Arrows show enhanced expression associated with nodule initiation which is also observed in developing nodules (arrowhead). (D) Cross-section of a nodule and the root shown in C; GUS expression can be observed in the cortex. Bars represent 300 μm in A and B; 3 mm in C; 1 mm in D. Sequence comparison of GmNFR5α and GmNFR5β in several soybean genotypes Single-nucleotide polymorphisms (SNPs) were found in the GmNFR5α (Supplementary Fig. S4) and GmNFR5β (Supplementary Fig. S5) sequences among various soybean genotypes. Several polymorphisms were identified within the coding sequence, the promoter region and 3′ UTR of soybean cv. Bragg (GQ340451), Williams (GQ340452), PI437.654 (GQ340449) and G. soja CPI100070 (GQ340450). Specifically, four SNPs, two being synonymous, were detected in the coding sequence of GmNFR5α in those soybean cultivars. The GmNFR5α sequences of Williams and cultivar PI437.654 are identical. Glycine soja CPI100070 and cv. Bragg have one and four SNPs, respectively, compared with Williams and PI437.654. GmNFR5β is less conserved than GmNFR5α. GmNFR5β of cv. Williams (GQ340458) and Bragg (GQ340457) is naturally mutated by a retroelement insertion (Supplementary Fig. S3), while G. soja CPI100070 (GQ340456) and PI437.654 (GQ340455) were wild type for this gene. Many known soybean cultivars also lacked the retroelement, allowing analysis of known breeding pedigrees. Transposable elements that comprise much of the repetitive DNA sequences in the genome have been reported in soybean, such as transposon G. max (Tgm; Vodkin et al. 1983, Rhodes and Vodkin 1988), a mariner-like element (Soymar1; Jarvik and Lark 1998) and a copia/Ty1-like retroelement (SIRE-1; Laten and Morris 1993, Laten et al. 1998). The size of these retroelements ranged from 1.6–12 kb (Tgm), 3.5 kb (Soymar1) and 11 kb (SIRE-1). Beside the retroelement insertion, comparison of the GmNFR5β coding sequence in cv. Bragg and Williams with G. soja CPI100070 and PI437.654 showed that there are seven SNPs, two of them translated into similar amino acids, while the other five translated into different amino acids (Supplementary Fig. S5). These did not affect the apparent symbiotic abilities and host range of the wild type. Unlike GmNFR5α, where cv. Bragg has four SNPs compared with cv. Williams, these two cultivars have the same GmNFR5β sequence not only in the coding sequence, but also in the promoter region and the 3′ UTR. Glycine soja CPI100070 and PI437.654, which have wild-type versions of GmNFR5β, have only one SNP in the coding sequence of GmNFR5β. However, in the 3′ UTR (the first 180 bp after the stop codon), G. soja CPI100070 has the same sequence as cv. Bragg and Williams, but has three SNPs compared with soybean cultivar PI437.654. This further verifies that modern cultivars of soybean are closely related and show some polymorphism to more ancestral types commonly called ‘Glycine soja’. However, at times wild accessions are used in soybean breeding, introducing ancestral gene blocks into modern cultivars. Widespread occurrence of the GmRE-1 retroelement insertion in the GmNFR5β gene The distribution of the GmRE-1 insertion was analysed using genomic DNA of available ancestral, first- and second-generation soybean lines. As expected, the PCR fragment was amplified from the Bragg and Williams parental lines and their mutants, but was absent in G. soja and cultivar Harosoy63 (Fig. 5A), which had already been shown by genetics to carry the wild-type alleles of both genes (Gresshoff and Landau Ellis 1994, Pracht et al. 1993). Fig. 5 View largeDownload slide Widespread distribution and the origin of a putative retroelement insertion in the NFR5β gene of US soybean cultivars. (A) Bragg, Williams, nod139 and nn5 have the mutant allele, but the mutant allele was absent in Harosoy63 and G. soja CPI100070. (B) Explanation of the origin of the putative retroelement insertion on cultivars Bragg and Williams. Red highlight, soybean with mutant allele; green highlight, soybean with wild-type of GmNFR5β. The mutant allele of Bragg was derived from D49-2491 (sibling of Lee). Williams inherited the mutant allele from L57-0034 (hashed red means ‘presumptive carrier’). Fig. 5 View largeDownload slide Widespread distribution and the origin of a putative retroelement insertion in the NFR5β gene of US soybean cultivars. (A) Bragg, Williams, nod139 and nn5 have the mutant allele, but the mutant allele was absent in Harosoy63 and G. soja CPI100070. (B) Explanation of the origin of the putative retroelement insertion on cultivars Bragg and Williams. Red highlight, soybean with mutant allele; green highlight, soybean with wild-type of GmNFR5β. The mutant allele of Bragg was derived from D49-2491 (sibling of Lee). Williams inherited the mutant allele from L57-0034 (hashed red means ‘presumptive carrier’). The parents of cultivar Bragg, i.e. Jackson and D49-2491 (a sibling of Lee), and the parents of D49-2491 (S100 and CNS) (Fig. 5B) were tested for the mutant retroelement allele. An amplification product of the same size and sequence as for Bragg, Williams and their mutants was detected in Lee and CNS, indicating that the origin of the insertion allele was in the other parent of D49-2491, namely line CNS. Cultivar Wayne, which was derived from CNS and is a parent of Williams, did not carry the insertion allele. However, other ancestors such as Clark, Richland and Volstate (Fig. 5A), which have no known relationship to CNS, possess the insertion sequence in the GmNFR5β gene. A search of the soybean genome (http://www.phytozome.net/soybean; see Schmutz et al. 2009) gave multiple locations for GmRE-1. Three of these loci (other than GmNFR5β*) showed high sequence similarity (97%); moreover, part of the retroelement (ranging between 50 and 350 bp) was present at several positions on all chromosomes with 80–99% sequence similarity (Supplementary Table S2). Thus GmRE-1 belongs to a moderately repeated family, having diverged since the ancestral segmental genome duplication leading to the two GmNFR5 loci. Discussion This study demonstrates that soybean can be analyzed effectively, combining mutant studies, electronic databases, genomic approaches and functional complementation. Specifically we were able to resolve a long-standing paradox involving recessive inheritance in a duplicated genome. Gene discovery in important legume crops such as soybean has been hampered by their larger genome size, endoduplication and tetraploidy, and difficulties with genetic complementation by transformation. Using loss-of-function mutants with a non-nodulation phenotype, and genetic mapping coupled with a gene candidate approach, the GmNFR5α and GmNFR5β genes of soybean encoding putative Nod factor receptors were cloned and characterized. GmNFR5 was duplicated in the genome, raising the question of how apparent loss-of-function mutants, such as nod139 and nn5, could be isolated. GmNFR5α and GmNFR5β are located on soybean chromosomes 11 and 1, in tandem with GmLYK7 and GmLYK4, respectively (Zhang et al. 2007). LYK4 and LYK7 genes encode related LysM receptor kinases (Limpens et al. 2003), suggesting ancestral segmental duplication prior to allopolyploidization (NB: in pea and M. truncatula LYK3 and LYK4 genes were proposed to be the entry receptor for nodulation) (Limpens et al. 2003). Many examples of two independent genes controlling the same trait can be found in soybean (Palmer and Kilen 1987). Shoemaker et al. (1996) also reported that hybridization to soybean genomic DNA by 280 randomly chosen PstI genomic clones determined that >90% of the probes detected more than two fragments and nearly 60% detected ≥3 fragments. From these findings, Stacey et al. (2004) suggested that <10% of the genome may be single copy sequence and that large amounts of the genome may have undergone genome duplication in addition to the tetraploidization event. Analysis and complementation of the mutants Hairy roots overexpressing GmNFR5α and GmNFR5β from the 35S promoter (with 10–80 times higher GmNFR5 mRNA levelss than control levels as measured by qRT-PCR) did not significantly increase nodule number in nod139 or Bragg in comparison with the hairy roots transformed with GmNFR5α and GmNFR5β driven by their native promoter, whether expressed per plant, per unit root or per unit root mass (Table 1). As overexpression of GmNFR1α (Indrasumunar 2007) increased nodulation and GmNFR5α/GmNFR5β did not, we conclude that transcription of GmNFR5α and GmNFR5β is not limiting nodule initiation. Zhang et al. (2007) found, based on sequence identification provided from this laboratory, that transcript levels of GmNFR5α and GmNFR5β in lateral or primary roots were three and five times higher, respectively, than the transcript level of GmNFR1α. As proposed by Madsen et al. (2003) and Radutoiu et al. (2003, 2007), NFR1 and NFR5 may act in concert as Nod factor receptors; therefore, increasing the transcript level of GmNFR1α that was limited will increase nodulation significantly, while increasing the transcript level of GmNFR5 that already is abundant will not result in a significant increase of nodulation. Related to this, an imbalance between NFR1 and NFR5 expression was suggested to cause the reduced nodulation efficiency observed in M. loti-inoculated L. filicaulis plants transformed with LjNFR1 and LjNFR5, separately (Radutoiu et al. 2007). The wild type and mutants respond differently in nodulation. Nodulation parameters of complemented root systems of mutant nod139 were significantly higher than wild-type Bragg (Table 1). Differential AON may be the cause of this difference, because all hairy roots of Bragg nodulated, whereas only about 50% did in nod139 roots. Thus complemented nod139 root systems may have triggered less AON, leading to increased nodule numbers. Unlike GmNFR1 which has only one functional copy in soybean (Indrasumunar 2007), both GmNFR5 genes are functional and complement each other. Comparison of protein identity shows that GmNFR5α and GmNFR5β share marginally higher identity (94%) than GmNFR1α and GmNFR1β (90%). In the extracellular domain, GmNFR5α and GmNFR5β also have higher similarity (91%) than GmNFR1α and GmNFR1β (82%). However, both pairs of genes have equally high identity (96%) in the kinase domain. Cross-complementation of GmNFR1 into GmNFR5 mutants (nod139, nn5), and GmNFR5 into GmNFR1 mutants (nod49, rj1) showed that these genes did not complement each other. Protein characteristics of GmNFR5α and GmNFR5β Protein domain structure and topology predict that the extracellular LysM domains would be involved directly or indirectly in binding of Nod factors, thus initiating signal transduction through intracellular kinases (Limpens et al. 2003, Madsen et al. 2003, Radutoiu et al. 2003, Radutoiu et al. 2007). Comparison of extracellular regions between GmNFR5α and GmNFR5β shows amino acid identities of 92, 91 and 90% in LysM1, LysM2 and LysM3, respectively. Detailed comparison of LysM motifs between GmNFR5 and orthologous NFR5 proteins of L. japonicus, P. sativum and M. truncatula reveals that GmNFR5 has higher similarity to LjNFR5 (57–79%) than the NFR5 homolog of Medicago (44–71%) and pea (53–73%); at the same time the NFR5 homologs of Medicago and pea share high similarity (72–78%) in the LysM motifs (Supplementary Fig. S2A). Soybean, lotus, pea and Medicago represent different nodulation types (determinate vs. indeterminate) and bacterial cross-inoculation groups. Madsen et al. (2003) speculated that NFR5 is not activated by an activation loop, but rather that ligand or interactor binding immediately activates the kinase. Alternatively, NFR5 might lack kinase activity and require an associated kinase, perhaps NFR1, to phosphorylate downstream substrates (Madsen et al. 2003). Radutoiu et al. (2007) showed that the Nod factor specificity of legume–rhizobial symbioses is determined by both NFR1 and NFR5. Transferring L. japonicus NFR1 and NFR5 to M. truncatula enables nodulation of the transformants by the L. japonicus symbiont M. loti. In addition, the specificity for different rhizobial symbionts of two different Lotus species is a function of a single amino acid residue (Leu118) within LysM2 domains of NFR5. Furthermore, LysM2 of NFR5 is the most diverged LysM domain among NFR1 and NFR5 homologs found in other plant species (Madsen et al. 2003; Supplementary Fig. S2A). Zhukov et al. (2008) also showed that a pea mutant (RisNod4) carrying an L77 to F substitution in the LysM1 domain of PsSym37 (PsNFR1) displays a restrictive symbiotic phenotype. RisNod4 mutants develop nodules only in the presence of a ‘Middle East’ Rhizobium strain producing O-acetylated Nod factors, indicating the PsNFR1 involvement in Nod factor recognition. These results provide strong evidence that NFR1 and NFR5 are indeed the Nod factor receptors. To confirm this conclusion, further studies on LysM domain swap between GmNFR1/5 and LjNFR1/5 are needed. In addition, further biochemical analyses of the binding affinities are still required to support the conclusion. Retroelement (GmRE-1) insertion in the NFR5β gene in many soybean cultivars Analysis of the amplification results and the pedigree of the tested soybean cultivars revealed that at least six ancestral lines (CNS, Richland, Volstate, Peking, Perry and Dorman), thought to be unrelated, carry the same insertion. We note that detailed pedigree and breeding histories are difficult to obtain. CNS, Richland, Peking and Dunfield are claimed to be of Chinese origin and thus might have common ancestors! Since these plants represent at least 20% of the genetic base of North American soybean lines (Gizlice 1994), the presence of a common retroelement insertion event implies that the genetic diversity of these cultivars is even lower than predicted from breeding and historical data. Interestingly, although most of the non-US cultivars tested were devoid of the insertion allele, the Japanese cultivar Enrei (Fig. 5A), of unknown pedigree, also carried the insertion, indicating a common ancestor(s) with the North American lines. Soybean cv. Enrei was the wild-type parent of non-nodulation mutants (En115, En1282 and En1314) and supernodulation mutant En6500 (Francisco and Akao 1993). En1314 and En1282 were mutated in the nfr1 gene (Ikeda et al. 2008). En115 did not produce root cortical cell division in response to inoculation of rhizobia (Francisco and Akao 1993) and, in this respect, is similar to nod139, which fails to initiate cortical cell division (Mathews et al. 1989b). Retroelement (GmRE-1) insertion in GmNFR5β facilitated the isolation of non-nodulation mutant lines Most of soybean nodulation mutants were derived from wild-type soybean, which already have the natural mutation in GmNFR5β (Bragg, Williams and Enrei). This is understandable due to its polyploid history; many examples of duplicated genes can be found in soybean (Palmer and Kilen 1987, Schmutz et al. 2009). In the case of GmNFR5, if ethylmethane sulfonate mutagenesis had been applied to a soybean cultivar that still had both wild-type copies of GmNFR5 (i.e. Harosoy63 and G. soja CPI100070, Illini and Jackson), non-nodulation phenotypes are unlikely to be found. Expression study of GmNFR5 Expression of GmNFR5α and GmNFR5β* was root specific, as determined by qRT-PCR (Fig. 3) and GUS fusion (Fig. 4). Trifoliate leaves and the shoot tip region had extremely low transcript levels for both genes. Expression determined by the presumed GmNFR5α and GmNFR5β promoters was observed in most root tissues of uninoculated roots (Fig. 4A). Inoculated roots showed increased expression of GUS during the early steps of nodule development (see arrows in Fig. 4B, C) except mature nodules (Fig. 4D). Similar results were obtained for reporters of the receptor-like kinase genes, M. truncatula Nod-Factor Perception (MtNFP; Arrighi et al. 2006) and M. truncatula Nodulation Receptor Kinase (MtNORK; Bersoult et al. 2005). However, the expression of MtNORK and MtNFP persisted in mature M. truncatula nodules, whereas the expression of GmNFR5 in soybean nodules did not (Fig. 4D; consistent with qRT-PCR data; see above). This observation most probably relates to the difference between indeterminant nodules containing a persistent meristem (as in M. truncatula) compared with mature determinant nodules devoid of a meristem in soybean. By the cloning of GmNFR1 (Indrasumunar 2007) and GmNFR5, all available non-nodulation mutants of soybean were successfully characterized. Despite extensive mutational screens, there is only a limited number of non-nodulation mutants of soybean compared with the diploid model legumes L. japonicus and M. truncatula. This most certainly is due to the allotetrapoloid nature of soybean. For example, we identified two copies of GmPoltergeist (A. Miyahara unpublished data), G. max nodulation receptor kinase (GmNORK) (GQ336811, GQ336812; Supplementary Fig. S6) and G. max kinase- associated protein phosphatase (GmKAPP) (Miyahara et al. 2008). Such functional duplication would prevent the detection of a visible phenotype of loss-of-function mutants. The availability of the soybean genome, candidate gene sequences from model legumes and soybean loss-of-function mutants strengthens their respective utility for the analysis of nodulation biology. Furthermore duplicated genomes, though seemingly recalcitrant to genetic analysis, evolve independently by insertions or point mutations, often leading to genetic inactivation of duplicated alleles and diploidized inheritance. Materials and Methods Soybean genotypes, F2 mapping populations and Bradyrhizobium strain Soybean plants were grown as described by Carroll et al. (1985a, 1985b). Mutant nod139 (Carroll et al. 1986, Mathews et al. 1989a) was derived from G. max variety Bragg. Non-nodulation G. max mutant nod139 was also crossed to G. soja PI468.347 (Kolchinsky et al. 1997) to produce the F2 mapping population of 481 plants. A second F2 mapping population was produced by crossing nod139 to G. soja accession CPI100070 to produce a F2 mapping population of 960 plants. The G. soja accession was selected from the Plant Introduction/Quarantine Unit, Commonwealth Scientific and Industrial Research Organization Division of Plant Industry (Canberra, Australia). All plants were grown as described by Carroll et al. (1985a, 1985b) and were inoculated 7 and 14 d after germination with Bradyrhizobium japonicum CB1809 (Biocare Pty. Ltd., Melbourne, Australia). The nodulation phenotype of plants was scored 28 d after germination. SSR marker detection Bulk DNA or individual plant DNA was used as template in PCRs with selected SSR oligonucleotide primers. The oligonucleotide sequence of selected primers was obtained from Cregan et al. (1999). Primers were used either in multiplex PCRs to amplify five SSR markers simultaneously or as separate pairs to PCR amplify a single SSR marker. Single SSR marker amplification was performed in 20 μl PCRs containing 0.3 μM of each primer, 0.075 μCi of [α-33P]dATP (3,000 Ci mmol−1), 20 μM of dNTPs, 10 mM Tris–HCl (pH 8.3), 2.0 mM MgCl2, 0.5 U of Taq polymerase (Perkin-Elmer, Foster City, CA, USA) and 50 ng of soybean genomic DNA. Isolation of the GmNFR5 probe Primers to amplify the ortholog of L. japonicus NFR5 (namely GmNFR5) were designed in accordance with the conserved sequence of LjNFR5 and orthologous NFR5 sequences of M. truncatula and P. sativum. PCR amplification was performed in triplicate in a 25 μl volume containing 50 ng of genomic DNA of Bragg, 10× PCR buffer, 2.5 mM MgCl2, 200 mM dNTPs, 0.5 U of Taq DNA polymerase recombinant (Invitrogen, Mount Waverley, VIC, Australia) and 0.2 μM of both forward and reverse oligonucleotide primers. Amplification was performed using the following cycling condition: one cycle of 94°C for 2 min, 35 cycles of 94°C for 10 s, 55°C for 30 s and 68°C for 2 min, followed by one cycle of 68°C for 7 min. Specific PCR products were cloned into pCR®4TOPO® vector (Invitrogen) according to the manufacturer’s instructions. After plasmid purification, the plasmid-containing PCR products were sequenced on an ABI 377 sequencer at the Australian Genome Research Facility, Brisbane, Australia. Sequences were viewed using the Chromas 1.45 software package (Griffith University, Australia). The sequences of PCR products were then compared with the sequence of LjNFR5. When the sequence of PCR products had high similarity (>60%) to LjNFR5, the PCR products were used as probes to screen a BAC library of soybean PI437.654 (Tomkins et al. 1999). Isolation of the GmNFR5α and GmNFR5β genes A good candidate probe with high homology to the LjNFR5 sequence was used to screen the BAC library PI437.654 (Tomkins et al. 1999). Positively hybridizing BAC clones were then ordered from Clemson University Genome Institute; subsequent BAC analysis and DNA sequencing were performed to clone the complete putative gene sequences of GmNFR5α and GmNFR5β. Primers used for BAC sequencing are listed in Supplementary Table S1. Sequence analysis of nod139 and nn5 mutants Gene-specific primers (Supplementary Table S1) were designed in accordance with the DNA sequence of GmNFR5α and GmNFR5β to amplify both genes from soybean varieties Bragg, Williams, nod139, nod49, nn5 and G. soja CPI100070. PCR products were cloned into pCR®4TOPO® vector (Invitrogen). To find the mutations, the sequences of both GmNFR5α and GmNFR5β mutants were compared with the sequences of their wild-type parent (Bragg and Williams) and G. soja CPI100070. A PCR test for detection of the GmNFR5β insertion allele Soybean seed lines for detection of retroelement insertion in GmNFR5β were provided by Dr. Sally Dillon and Dr. P. K. Lawrence from the Australian Tropical Grains Germplasm Centre and Professor Randall Nelson from USDA-ARS and University of Illinois, USA. These soybean seeds were: Adams, Altona, Amsoy, Blackhawk, Clark, CNS, Corsoy, Dorman, Dragon, DT84(1), Dunfield, Enrei, Forrest, Fukuyutaka, Harosoy, Harosoy63, Hawkeye, Illini, Jackson, Kent, Lee, Lincoln, Ogden, PI437.654, Palmeto, Peking, Perry, Richland, S100, Valiant, Volstate and Wayne. The insertion amplicons were amplified using specific primers listed in Supplementary Table S1. The forward primer was in the retroelement region (348–369 nucleotides downstream from the initiation ATG of GmNFR5β) and the reverse primer was in the transmembrane domain (2,184–2,204 nucleotides downstream of the ATG). PCR amplification was performed in a 25 μl volume containing 50 ng of genomic DNA, 10× PCR buffer, 2.5 mM MgCl2, 200 mM dNTPs, 0.5 U of Taq DNA polymerase recombinant (Invitrogen) and 0.2 μM of both forward and reverse oligonucleotide primers. Samples were heated to 94°C for 2 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 59°C for 30 s, elongation at 68°C for 2 min and a final extension at 68°C for 10 min. Amplified products were separated by electrophoresis in 1% agarose gels (Scientifix, Cheltenham, VIC, Australia) in 1× TAE buffer and were detected by fluorescence under UV light (302 nm) using the BioDoc-It™ Imaging System. Bioinformatics analysis Sequence analysis of GmNFR5α and GmNFR5β was performed using: BCM Search Launcher (http://searchlauncher.bcm.tmc.edu/multi-align/multi-align.html), ENTREZ/BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi), BOXSHADE (http://www.ch.embnet.org/software/BOX_form.html), SPIDEY (http://www.ncbi.nlm.nih.gov/spidey/), ClustalW (http://www.ebi.ac.uk/clustalw/), Gene structure and promoter predictions (http://linux1.softberry.com/berry.phtml?topic=fgenesh&group=programs&subgroup=gfind), FGENESH program (http://linux1.softberry.com/berry.phtml?topic=fgenesh&group=programs&subgroup=gfind), SMART (http://smart.embl-heidelberg.de/), Hydrophobic analysis (http://www.cbs.dtu.dk/services/SignalP/), Transmembrane prediction (http://ccb.imb.uq.edu.au/svmtm/svmtm_predictor .shtml), Map Viewer (http://www.ncbi.nlm.nih.gov/mapview/maps.cgi?TAXID=3847&MAPS=default&CHR), Primer3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3.cgi) and Phytozome (http://www.phytozome.net/soybean). Hairy root transformation To confirm that the mutations in the NFR5 genes were responsible for the non-nodulation phenotype in nod139 and nn5, complementation by hairy root transformation (Kereszt et al. 2007) was carried out. The wild-type clones of NFR5α and NFR5β genes derived from the genome of G. soja CPI100070, driven either by their own respective promoters (putatively selected at 548 bp for GmNFR5α and at 1,203 bp for GmNFR5β), or by the cauliflower mosaic virus (CaMV) 35S promoter, were cloned into binary vector pCAMBIA1305.1 and introduced into mutant plants via A. rhizogenes cucumopine strain K599. Nodule numbers of transgenic roots were determined 35 d after inoculation. RNA extraction and qRT-PCR For the qRT-PCR experiment, RNA samples were isolated from the leaves and roots of 2-week-old inoculated (1 week post-inoculation) and uninoculated wild-type Bragg, nod139 and nts1007 plants grown in a glasshouse, as well as from nodules collected 2 weeks after inoculation. More detailed expression studies were conducted on root segments, trifoliate leaves and shoot tips of uninoculated and inoculated Bragg plants. The inoculated plants were treated with a 4-day-old YMB culture of B. japonicum strain CB1809 immediately after sowing. Germination and seedling growth were conducted in an air-conditioned glasshouse maintained at day and night temperatures of 28 and 23°C, with a 16 h photoperiod. Taproot segments (TR), lateral root segments (LR), trifoliate leaves (TF) and shoot tips (STIP) were harvested from inoculated and uninoculated Bragg 14 d post-sowing. For collection of the taproot segments, the first 2 cm of the root tips was harvested as taproot 1 (TR1), and the adjoining 2 cm segments were taken as taproot 2 (TR2) and taproot 3 (TR3), respectively. Similarly three fragments collected from one lateral root were correspondingly labeled as lateral root 1 (LR1), lateral root 2 (LR2) and lateral root 3 (LR3). RNA extraction was performed using the NucleoSpin® RNA Plant Kit (Macherey Nagel, Düren, Germany) according to the manufacturer’s instructions. RNA concentrations were determined using a NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). First-strand cDNA was synthesized from 1 μg of total RNA using SuperscriptIII Reverse Transcriptase (Invitrogen) according to the manufacturer’s instructions. The cDNA templates were diluted and added to SYBR green PCR master mix (Applied Biosystems , Foster City, CA, U.S.A). containing 0.2 μmol of each primer (Table 1). qRT-PCR was carried out using an ABI PRISM 7900HT thermocycler (Applied Biosystems) under the following conditions: 95°C for 10 min followed by 35 cycles of 95°C for 15 s and 60°C for 1 min. A heat dissociation profile from 60 to 95°C was conducted at the end of each PCR to verify the specificity of the reaction. Individual values for each sample were generated by averaging data from two biological and two technical replicates. Primers for qRT-PCR were designed using Primer Express (Applied Biosystems , Foster City, CA, U.S.A.). Histochemical localization of GUS activity The promoter of both genes (880 bp for GmNFR5α and 970 bp for GmNFR5β) was integrated in front of the GUS reporter gene in vector pCAMBIA1305.1 and transformed into wild-type line Bragg and its supernodulating mutant nts1007 with the help of A. rhizogenes-mediated transformation. GUS activity was determined histochemically as described earlier (Nontachaiyapoom et al. 2007). Hairy roots were first vacuum infiltrated then stained overnight at 37°C in staining solution containing 0.1 mM X-Gluc, 100 mM sodium phosphate buffer (pH 7), 5 mM EDTA, 1 mM K4Fe(CN)6, 1 mM K3Fe(CN)6 and 0.1% (v/v) Triton X-100. For sectioning, tissues were fixed in 0.5% paraformaldehyde in 100 mM sodium phosphate buffer (pH 7) on ice under vacuum for 45 min, rinsed three times in 100 mM sodium phosphate buffer (pH 7) at room temperature, stained in X-Gluc solution, embedded in 3% agarose, and sectioned to 100 μm thickness with a vibratome (Lancer Series 1000; Ted Pella, Irvine, CA, USA). Statistical analysis Analysis of variance (ANOVA) was used to identify if there were significant effects of the treatments (empty vector, native promoter and 35S promoter) on the variables: nodule number per plant, nodule number per root and nodule number per root dry weight. Where significant effects were found, the least significant difference separation procedure was used to separate the differences. Funding This work was supported by the Australian Research Council [Centre of Excellence grant CEO34821]; the Queensland State Government [Smart State Innovation Scheme]; the University of Queensland Strategic Fund; AusAID [PhD scholarship to A.I.]; Sugar Research and Development Corporation (SRDC) [scholarship to I.S.]. Acknowledgments We thank Professor Randall Nelson, USDA-ARS and University of Illinois, USA for providing seeds of CNS, Richland, Volstate, Dunfield and non-nodulation mutant nn5. Soybean seeds lines for retroelement detection were generously provided by Dr Sally Dillon and Dr P. K. Lawrence from Australian Tropical Grains Germplasm Centre. We also thank CILR staff for help and suggestions. Abbreviations Abbreviations AON autoregulation of nodulation BAC bacterial artificial chromosome CaMV cauliflower mosaic virus GUS β-glucuronidase LYK LysM receptor kinase NFR Nod factor receptor qRT-PCR quantitative real-time PCR SNP single nucleotide polymorphism SSR simple sequence repeat UTR untranslated region. References Arrighi JF,  Barre A,  Ben Amor B,  Bersoult A,  Soriano LC,  Mirabella R, et al.  The Medicago truncatula lysine motif-receptor-like kinase gene family includes NFP and new nodule-expressed genes,  Plant Physiol. ,  2006, vol.  142 (pg.  265- 279) Google Scholar CrossRef Search ADS PubMed  Ben-Amor B,  Shaw SL,  Oldroyd GE,  Maillet F,  Penmetsa RV,  Cook D, et al.  The NFP locus of Medicago truncatula controls an early step of Nod-factor signal transduction upstream of a rapid calcium flux and root hair deformation,  Plant J. ,  2003, vol.  34 (pg.  495- 506) Google Scholar CrossRef Search ADS PubMed  Bersoult A,  Camut S,  Perhald A,  Kereszt A,  Kiss GB,  Cullimore JV.  Expression of the Medicago truncatula DMI2 gene suggests roles of the symbiotic nodulation receptor kinase in nodules and during early nodule development,  Mol. Plant-Microbe Interact. ,  2005, vol.  18 (pg.  869- 876) Google Scholar CrossRef Search ADS PubMed  Carroll BJ,  Mcneil DL,  Gresshoff PM.  Isolation and properties of soybean [Glycine max (L.) Merr] mutants that nodulate in the presence of high nitrate concentrations,  Proc. Natl Acad. Sci. USA ,  1985, vol.  82 (pg.  4162- 4166) Google Scholar CrossRef Search ADS   Carroll BJ,  Mcneil DL,  Gresshoff PM.  A supernodulation and nitrate-tolerant symbiotic (Nts) soybean mutant,  Plant Physiol. ,  1985, vol.  78 (pg.  34- 40) Google Scholar CrossRef Search ADS PubMed  Carroll BJ,  McNeil DL,  Gresshoff PM.  Mutagenesis of soybean [Glycine max (L.) Merr.] and the isolation of nonnodulating mutants,  Plant Sci. ,  1986, vol.  47 (pg.  109- 114) Google Scholar CrossRef Search ADS   Cregan PB,  Mudge J,  Fickus EW,  Marek LF,  Danesh D,  Denny R, et al.  Targeted isolation of simple sequence repeat markers through the use of bacterial artificial chromosomes,  Theor. Appl. Genet. ,  1999, vol.  98 (pg.  919- 928) Google Scholar CrossRef Search ADS   Francisco PB,  Akao S.  Autoregulation and nitrate inhibition of nodule formation in soybean cv. Enrei and its nodulation mutants,  J. Exp. Bot. ,  1993, vol.  44 (pg.  547- 553) Google Scholar CrossRef Search ADS   Geurts R,  Heidstra R,  Hadri AE,  Downie JA,  Franssen H,  van Kammen AB, et al.  Sym2 of pea is involved in a nodulation factor-perception mechanism that controls the infection process in the epidermis,  Plant Physiol. ,  1997, vol.  115 (pg.  351- 359) Google Scholar CrossRef Search ADS PubMed  Gizlice Z,  Carter TEJr,  Burton JW.  Genetic base for North American public soybean cultivars released between 1947 and 1988,  Crop Sci. ,  1994, vol.  34 (pg.  1143- 1151) Google Scholar CrossRef Search ADS   Gresshoff PM,  Landau-Ellis D.  Gresshoff P.  Molecular mapping of soybean nodulation genes,  Plant Genome Analysis: Current Topics in Plant Molecular Biology ,  1994 Boca Raton, FL CRC Press(pg.  97- 112) Ikeda S,  Rallos LEE,  Okubo T,  Eda S,  Inaba S,  Mitsui H,  Minamisawa K.  Microbial community analysis of field-grown soybeans with different nodulation phenotypes,  Appl. Environ. Microbiol. ,  2008, vol.  74 (pg.  5704- 5709) Google Scholar CrossRef Search ADS PubMed  Indrasumunar A.  Molecular cloning and functional characterisation of soybean (Glycine max L.) nod factor receptor genes,  2007 Australia The University of Queenslandpg.  162   PhD Thesis Jarvik T,  Lark KG.  Characterization of soymar1, a mariner element in soybean,  Genetics ,  1998, vol.  149 (pg.  1569- 1574) Google Scholar PubMed  Kereszt A,  Dongxue L,  Indrasumunar A,  Nguyen CDT,  Nontachaiyapoom S,  Kinkema M, et al.  Agrobacterium rhizogenes-mediated transformation of soybean to study root biology,  Nature Protocols ,  2007, vol.  2 (pg.  948- 952) Google Scholar CrossRef Search ADS PubMed  Kolchinsky A,  LandauEllis D,  Gresshoff PM.  Map order and linkage distances of molecular markers close to the supernodulation (nts-1) locus of soybean,  Mol. Gen. Genet. ,  1997, vol.  254 (pg.  29- 36) Google Scholar CrossRef Search ADS PubMed  Kunze R,  Saedler H,  Lonnig WE.  Plant transposable elements,  Adv. Bot. Res. ,  1997, vol.  27 (pg.  331- 470) Laten HM,  Morris RO.  SIRE-1, a long interspersed repetitive DNA element from soybean with sequence similarity to retrotransposons: initial characterization and partial sequence,  Genetics ,  1993, vol.  134 (pg.  153- 159) Laten HM,  Majumdar A,  Gaucher EA.  SIRE-1, a copia/Ty1-like retroelement from soybean, encodes a retroviral envelope-like protein,  Proc. Natl Acad. Sci. USA ,  1998, vol.  95 (pg.  6897- 6902) Google Scholar CrossRef Search ADS   Limpens E,  Franken C,  Smit P,  Willemse J,  Bisseling T,  Geurts R.  LysM domain receptor kinases regulating rhizobial Nod-factor-induced infection,  Science ,  2003, vol.  302 (pg.  630- 633) Google Scholar CrossRef Search ADS PubMed  Madsen EB,  Madsen LH,  Radutoiu S,  Olbryt M,  Rakwalska M,  Szcyglowski K, et al.  A receptor-like kinase gene involved in legume perception of rhizobial signal molecules,  Nature ,  2003, vol.  425 (pg.  637- 640) Google Scholar CrossRef Search ADS PubMed  Markmann K,  Parniske M.  Evolution of root endosymbiosis with bacteria: how novel are nodules?,  Trends Plant Sci. ,  2009, vol.  14 (pg.  77- 86) Google Scholar CrossRef Search ADS PubMed  Mathews A,  Carroll BJ,  Gresshoff PM.  A new non-nodulation gene in soybean,  J. Hered. ,  1989, vol.  80 (pg.  357- 360) Google Scholar CrossRef Search ADS   Mathews A,  Carroll BJ,  Gresshoff PM.  Development of Bradyrhizobium infections in supernodulating and non-nodulating mutants of soybean (Glycine max (L.) Merrill),  Protoplasma ,  1989, vol.  150 (pg.  40- 47) Google Scholar CrossRef Search ADS   Mathews A,  Carroll BJ,  Gresshoff PM.  Studies on the root control of non-nodulation and plant growth of non-nodulating mutants and supernodulating mutant of soybean (Glycine max (L.) Merr.),  Plant Sci. Limerick ,  1992, vol.  83 (pg.  35- 43) Google Scholar CrossRef Search ADS   Miyahara A,  Hirani TA,  Oakes M,  Kereszt A,  Kobe B,  Djordjevic MA, et al.  Soybean nodule autoregulation receptor kinase phosphorylates two kinase-associated protein phosphatases in vitro,  J. Biol. Chem. ,  2008, vol.  283 (pg.  25381- 25391) Google Scholar CrossRef Search ADS PubMed  Nontachaiyapoom S,  Scott PT,  Men AE,  Kinkema M,  Schenk PM,  Gresshoff PM.  Promoters of orthologous Glycine max and Lotus japonicus nodulation autoregulation genes interchangeably drive phloem-specific expression in transgenic plants,  Mol. Plant-Microbe Interact. ,  2007, vol.  20 (pg.  769- 80) Google Scholar CrossRef Search ADS PubMed  Oldroyd GE,  Downie JA.  Coordinating nodule morphogenesis with rhizobial infection in legumes,  Annu. Rev. Plant Biol. ,  2008, vol.  59 (pg.  519- 46) Google Scholar CrossRef Search ADS PubMed  Palmer RG,  Kilen TC.  Wilcox JR.  Qualitative genetics and cytogenetics,  Soybean: Improvement, Production and Uses, 2nd edn. Agronomy Monograph 16 ,  1987 Madison, WI American Society of Agronomy, Crop Science Society of America, Soil Science Society of America(pg.  23- 48) Pracht JE,  Nickell CD,  Harper JE.  Genes controlling nodulation in soybean: Rj5 and Rj6,  Crop Sci. ,  1993, vol.  33 (pg.  711- 713) Google Scholar CrossRef Search ADS   Radutoiu S,  Madsen LH,  Madsen EB,  Felle HH,  Umehara Y,  Gronlund M,  Sato S, et al.  Plant recognition of symbiotic bacteria requires two LysM receptor-like kinases,  Nature ,  2003, vol.  425 (pg.  585- 592) Google Scholar CrossRef Search ADS PubMed  Radutoiu S,  Madsen LH,  Madsen EB,  Jurkiewicz A,  Fukai E,  Quistgaard EMH, et al.  LysM domains mediate lipochitin–oligosaccharide recognition and NFR genes extend the symbiotic host range,  EMBO J. ,  2007, vol.  26 (pg.  3923- 3935) Google Scholar CrossRef Search ADS PubMed  Rhodes PJ,  Vodkin LO.  Organization of the Tgm family of transposable elements in soybean,  Genetics ,  1988, vol.  120 (pg.  597- 604) Google Scholar PubMed  Schlueter JA,  Lin J-Y,  Schlueter SD,  Vasylenko-Sanders IF,  Deshpande S,  Yi J, et al.  Gene duplication and paleopolyploidy in soybean and the implications for whole genome sequencing,  BMC Genomics ,  2007, vol.  8 pg.  330  Google Scholar CrossRef Search ADS PubMed  Schlueter JA,  Vasylenko-Sanders IF,  Deshpande S,  Yi J,  Siegfried M,  Roe BA, et al.  The FAD2 gene family of soybean: insights into the structural and functional divergence of a paleopolyploid genome,  Plant Genome ,  2007, vol.  1 (pg.  14- 26) Schmutz J,  Cannon SB,  Schlueter J,  Ma J,  Mitros T,  Nelson W, et al.  Genome sequence of the paleopolyploid soybean,  Nature ,  2009  in press Searle IR. ,  Towards the map-based cloning of nodulation genes from soybean ,  2001 Australia School of Molecular and Microbial Sciences: The University of Queensland Searle IR,  Men AE,  Laniya TS,  Buzas DM,  Iturbe-Ormaetxe I,  Carroll B, et al.  Long-distance signaling in nodulation directed by a CLAVATA1-like receptor kinase,  Science ,  2003, vol.  299 (pg.  109- 112) Google Scholar CrossRef Search ADS PubMed  Shoemaker RC,  Polzin K,  Labate J,  Specht J,  Brummer EC,  Olson T, et al.  Genome duplication in soybean (Glycine subgenus soja),  Genetics ,  1996, vol.  144 (pg.  329- 338) Google Scholar PubMed  Stacey G,  Vodkin L,  Parrott WA,  Shoemaker RC.  National Science Foundation-sponsored workshop report. Draft plan for soybean genomics 1,  Plant Physiol. ,  2004, vol.  135 (pg.  59- 70) Google Scholar CrossRef Search ADS PubMed  Tomkins JP,  Mahalingam R,  Smith H,  Goicoechea JL,  Knap HT,  Wing RA.  A bacterial artificial chromosome library for soybean PI 437654 and identification of clones associated with cyst nematode resistance,  Plant Mol. Biol. ,  1999, vol.  41 (pg.  25- 32) Google Scholar CrossRef Search ADS PubMed  Vodkin LO,  Rhodes P,  Goldberg RB.  A lectin gene insertion has the structural features of a transposable element,  Cell ,  1983, vol.  34 (pg.  1023- 1031) Google Scholar CrossRef Search ADS PubMed  Williams LF,  Lynch DL.  Inheritance of a nonnodulating character in the soybean,  Agron. J. ,  1954, vol.  46 (pg.  28- 29) Google Scholar CrossRef Search ADS   Zhang X-C,  Wu X,  Findley S,  Wan J,  Libault M,  Nguyen,  H T, et al.  Molecular evolution of lysine motif-type receptor-like kinases in plants,  Plant Physiol. ,  2007, vol.  144 (pg.  623- 636) Google Scholar CrossRef Search ADS PubMed  Zhukov V,  Radutoiu S,  Madsen LH,  Rychagova T,  Ovchinnikova E,  Borisov A, et al.  The pea Sym37 receptor kinase gene controls infection-thread initiation and nodule development,  Mol. Plant-Microbe Interact. ,  2008, vol.  21 (pg.  1600- 1608) Google Scholar CrossRef Search ADS PubMed  © The Author 2009. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: [email protected]

Journal

Plant and Cell Physiology – Oxford University Press

Published: Dec 9, 2009

Keywords: Gene duplication Glycine max GmNFR5 Nodulation Retroelement Symbiosis

There are no references for this article.