Intracellular Catalytic Domain of Symbiosis Receptor Kinase Hyperactivates Spontaneous Nodulation in Absence of Rhizobia^,

Constitutive activity of the intracellular kinase domain of symbiosis receptor kinase hyperactivates nodule organogenesis in a legume.

Abstract

Symbiosis Receptor Kinase (SYMRK), a member of the Nod factor signaling pathway, is indispensible for both nodule organogenesis and intracellular colonization of symbionts in rhizobia-legume symbiosis. Here, we show that the intracellular kinase domain of a SYMRK (SYMRK-kd) but not its inactive or full-length version leads to hyperactivation of the nodule organogenic program in Medicago truncatula TR25 (symrk knockout mutant) in the absence of rhizobia. Spontaneous nodulation in TR25/SYMRK-kd was 6-fold higher than rhizobia-induced nodulation in TR25/SYMRK roots. The merged clusters of spontaneous nodules indicated that TR25 roots in the presence of SYMRK-kd have overcome the control over both nodule numbers and their spatial position. In the presence of rhizobia, SYMRK-kd could rescue the epidermal infection processes in TR25, but colonization of symbionts in the nodule interior was significantly compromised. In summary, ligand-independent deregulated activation of SYMRK hyperactivates nodule organogenesis in the absence of rhizobia, but its ectodomain is required for proper symbiont colonization.

In rhizobia-legume symbiosis, the host plant responds to rhizobial elicitation by developing root nodules and accommodating the symbionts intracellularly in the nodule interior. At the molecular level, the response is initiated with the recognition of rhizobial Nod factors by Nod factor receptors of the host plants (Limpens et al., 2003; Madsen et al., 2003; Broghammer et al., 2012). A compatible interaction triggers the Sym pathway and generates the calcium spikes as a signature of symbiotic signal transduction (Ehrhardt et al., 1996; Sieberer et al., 2012; Oldroyd, 2013). Proteins acting upstream of Ca²⁺ spiking are Doesn’t Make Infections2 (DMI2), Symbiosis Receptor Kinase (SYMRK; Endre et al., 2002; Stracke et al., 2002), the cation channels DMI1, CASTOR, and POLLUX (Ané et al., 2004; Charpentier et al., 2008), and three nucleoporins (NUP85 and NUP133 [Kanamori et al., 2006; Saito et al., 2007] and NENA [Groth et al., 2010]). A nuclear calcium- and calmodulin-dependent protein kinase (CCaMK) or DMI3 (Lévy et al., 2004; Mitra et al., 2004) decodes the Ca²⁺ spiking and phosphorylates a transcription factor Interacting Protein of DMI3 (IPD3) or CYCLOPS (Messinese et al., 2007; Yano et al., 2008; Singh et al., 2014), which along with several other transcription factors, like Nodulation Signaling Pathway1 (NSP1; Smit et al., 2005), NSP2 (Kaló et al., 2005), ERF Required for Nodulation1 (Middleton et al., 2007), and Nodule Inception (NIN; Schauser et al., 1999; Marsh et al., 2007), orchestrates the gene expression required for rhizobial infection and nodule organogenesis.

Constitutively active forms of the cytokinin receptor Lotus Histidine Kinase1 (LHK1; Tirichine et al., 2007), CCaMK (Gleason et al., 2006; Tirichine et al., 2006a), CYCLOPS (Singh et al., 2014), or NIN (Soyano et al., 2013) can result in spontaneous nodulation, showing that nodule organogenesis does not require rhizobial infection. In this model, CCaMK-induced nodule organogenesis is mediated by LHK1, because gain-of-function LHK1 induces spontaneous nodules in loss-of-function ccamk mutants (Tirichine et al., 2007; Hayashi et al., 2010; Madsen et al., 2010). CCaMK is placed upstream of all pathways required for this organogenic program, because autoactive CCaMK versions can compensate for the loss of upstream genes, like SYMRK, CASTOR, POLLUX, or NUP85/NUP133, that are involved in calcium spike generation, indicating that the primary function of these genes is the activation of CCaMK (Madsen et al., 2010).

In all tested eurosids, SYMRK contains three Leu-rich repeat (LRR) motifs, a malectin-like domain in the extracellular region, and a protein kinase catalytic domain in the intracellular region (Markmann et al., 2008). Cross species complementation tests showed that SYMRKs from eurosids, including nodulating and nonnodulating lineages, can restore symbiosis in symrk null mutants, indicating that the extra cytoplasmic region of SYMRK does not mediate specificity in rhizobia-legume recognition (Gherbi et al., 2008; Markmann et al., 2008). In this report, we show that the intracellular kinase domain of SYMRK from Arachis hypogaea (AhSYMRK-kd) hyperactivates spontaneous nodulation in Medicago truncatula in the absence of Sinorhizobium meliloti. Both CCaMK and ipd3/cyclops were required for triggering nodule organogenesis in the absence of rhizobia. The effect was not specific for SYMRK from A. hypogaea, because intracellular kinase domain of SYMRK from M. truncatula (MtSYMRK-kd) could similarly autoactivate nodule organogenesis in M. truncatula. In the presence of S. meliloti, colonization of nodules was rarely noted, leaving most of the nodules empty. Thus, ligand-independent deregulated activation of the intracellular kinase domain of SYMRK is capable of inducing nodule organogenesis in the absence of rhizobia, but the ectodomain of the receptor is important for proper colonization by S. meliloti.

RESULTS AND DISCUSSION

Cross Species Complementation of TR25 by AhSYMRK

Earlier, we reported isolation of AhSYMRK from A. hypogaea, a basal legume that is supported by crack invasion (Samaddar et al., 2013). It has 84% sequence similarity with SYMRK from an infection thread (IT) -supported legume M. truncatula (MtSYMRK) and falls at the point of divergence of legumes and nonlegumes in a distance tree (Supplemental Fig. S1). Here, we show AhSYMRK-mediated complementation of TR25, an symrk knockout mutant of M. truncatula (Endre et al., 2002). The empty vector-transformed TR25 roots did not show rhizobial colonization or nodulation upon inoculation with S. meliloti-expressing monomeric red fluorescent protein (mRFP; Fig. 1A). Under identical conditions, 35S::AhSYMRK-GFP-transformed TR25 roots responded by IT formation (Fig. 1B; Supplemental Fig. S2) as well as development of nodules (Fig. 1, C–J), indicating qualitative restoration of both organogenesis and rhizobial infection in TR25 (Supplemental Table S1). Nodules harvested 6 weeks after inoculation (WAI) were spherical in shape (Fig. 1C), whereas those harvested 8 WAI had the characteristic cylindrical shape of indeterminate nodules (Fig. 1D). In both cases, nodules remained white and ineffective, because functional restoration of symrk knockout mutants demands its expression under native promoter (Limpens et al., 2005). The TR25/AhSYMRK nodules (11 of 20) showed extensive growth of ITs in the central tissue of the nodule (Fig. 1E) with infrequent release of symbionts in some cells (Fig. 1, F and G). In a significant number of nodules (9 of 20), intracellularization of the symbionts was complete (Fig. 1, H and I), although the bacteroids remained small and resembled freshly released S. meliloti (Fig. 1J; Supplemental Fig. S3), indicating that ectopic expression of AhSYMRK in the entire nodule affects proper development of symbiosomes. Wild-type M. truncatula (A17) roots expressing 35S::AhSYMRK-GFP, however, developed nodules (Fig. 1K, inset) with properly differentiated elongated symbiosomes (Fig. 1, L and M; Supplemental Fig. S3), suggesting that native expression of endogenous SYMRK can dominate over the ectopically expressed SYMRKs for proper development of nodules and symbionts. Our observations add to the evidence showing that specificity of recognition of bacterial partners in rhizobia-legume or frankia-actinorhizal interactions is independent of the source of SYMRK as long as the SYMRK proteins have a malectin-like domain and three LRRs in their ectodomain (Gherbi et al., 2008; Markmann et al., 2008).

Figure 1.

Complementation of TR25 by AhSYMRK. TR25 roots were transformed with 35S::AhSYMRK-GFP and infected with S. meliloti expressing mRFP. A, Transgenic control roots lacking AhSYMRK cassette showing no nodules. GFP fluorescence image is shown. Bar = 2 mm. B, Root hairs of TR25/AhSYMRK 2 WAI with S. meliloti in merged images of bright-field and mRFP fluorescence. ITs (arrow) can be seen inside the curled root hairs. Bar = 200 µm. C and D, Nodulated roots of TR25/AhSYMRK 6 (C) and 8 (D) WAI shown as merged images of bright-field and mRFP fluorescence. Bottom shows enlarged views of spherical nodules (C) and cylindrical nodules (D) with bright-field (bottom left) and merged images of GFP and mRFP fluorescence (bottom right). mRFP fluorescence in inner nodule tissue indicates the presence of S. meliloti. Bars in C, inset, and D = 2 mm. Bar in D, inset = 200 µm. E to M, Longitudinal sections (30 µm) of 6-week-old nodules formed on TR25/AhSYMRK (E–J) or A17/AhSYMRK (K–M) shown as merged images of red (S. meliloti expressing mRFP) and blue (Calcofluor; cell wall) fluorescence. E to G, TR25/AhSYMRK nodules where ITs (arrowheads) occupy the central tissue (E) with infrequent release of symbionts (F and G). H to J, TR25/AhSYMRK nodules where symbionts were released from ITs (H) without being differentiated into elongated symbiosomes (I and J). K to M, A17/AhSYMRK transgenic nodules. K, Inset is shown as the merged image of GFP and mRFP fluorescence. Nodule interior filled with symbionts (K) showing fully differentiated symbiosomes (L and M). Symbiosomes in J and M are indicated by double arrowheads. Bars in E, H, and K = 100 µm. Bars in F, I, and L = 20 µm. Bars in G, J, and M = 5 µm. Bar in K, inset = 1 mm.

AhSYMRK-kd Hyperactivates the Nodule Organogenic Program in M. truncatula in the Absence of S. meliloti

Deregulated activation of CCaMK dispenses the requirement of SYMRK for triggering nodule organogenesis in the absence of rhizobia (Hayashi et al., 2010; Madsen et al., 2010). We reasoned that deregulated activation of SYMRK could trigger the organogenic program by activating CCaMK in the absence of S. meliloti. To check this possibility, we complemented TR25 with AhSYMRK-kd(573–883) that has been shown to be active in vitro (Samaddar et al., 2013). AhSYMRK-kd is expected to have deregulated activity because of (1) loss of context of its expression in the absence of the transmembrane domain or (2) loss of its ectodomain- or juxtamembrane domain-imposed regulation (Oh et al., 2009; Antolín-Llovera et al., 2014).

As indicated in Supplemental Table S1, TR25 complemented with 35S::AhSYMRK-kd by Agrobacterium rhizogenes-mediated transformation showed profuse nodulation in the absence of S. meliloti, with an average of 28.2 ± 5.22 (mean ± sem) nodules per root system. Under identical conditions, there were 4.5 ± 1.14 nodules per root system in the S. meliloti-infected TR25 plants complemented with 35S::AhSYMRK. On average, AhSYMRK-kd-induced spontaneous nodulation was 6-fold more than S. meliloti-induced nodulation in TR25 roots transformed with AhSYMRK (Supplemental Fig. S4). This is in contrast to other instances of spontaneous nodulation events, where the efficiency of spontaneous nodulation was similar or lower than the S. meliloti-induced nodulation (Gleason et al., 2006; Tirichine et al., 2006a, 2006b; Singh et al., 2014).

In most cases, the spontaneous nodules in TR25/AhSYMRK-kd roots grew as merged clusters (Fig. 2, A–D), causing distortion in the nodulated roots; however, in several instances, they were also found to grow as solitary spherical nodules (Fig. 2E). The distorted roots with excessive nodulation indicated that the spontaneous nodulation in AhSYMRK-kd-transformed roots may lack control over both nodule numbers and spatial position. In fact, spontaneously nodulated TR25 roots resembled S. meliloti-infected supernodulated M. truncatula roots in super numeric nodule (sunn) or sickle mutants that define distinct genetic pathways (Penmetsa and Cook, 1997; Penmetsa et al., 2003; Schnabel et al., 2005). SUNN encodes an LRR receptor-like kinase and is a key component of autoregulation of nodulation, which regulates nodule number by a long-distance negative feedback system (Reid et al., 2011). The sickle mutant is defective in ethylene perception, and insensitivity to ethylene is thought to be causal to its supernodulation phenotype (Penmetsa and Cook, 1997). These autoregulatory pathways could be compromised under deregulated activation of SYMRK, leading to hyperactivation of spontaneous nodulation (Fig. 2, A–D). This is unlike spontaneous nodulation under deregulated activation of CCaMK, where nodule organogenesis seemed to be autoregulated (Tirichine et al., 2006a). Interestingly, SYMRK mutants were identified as suppressors of hypernodulation in Lotus japonicus, which already suggested SYMRK to have a possible role in regulating nodule numbers (Murray et al., 2006).

Figure 2.

Spontaneous nodule formation by overexpression of AhSYMRK-kd(573–883). A to F, TR25 roots were transformed with 35S::AhSYMRK-kd-GFP. Both bright-field (top) and GFP fluorescence (bottom) images are shown for spontaneous nodules as either clusters (A–D) or solitary (E) in TR25/AhSYMRK-kd roots at 8 weeks after transplantation in vermiculite. F, No nodules in the presence of inactive kinase in TR25/AhSYMRK-kd (K625E)-GFP roots. Bars in A to D and F = 2 mm. Bar in E = 500 µm. G to O, Longitudinal section of spontaneous nodules on TR25/AhSYMRK-kd roots (H–L) compared with infected nodules on TR25/AhSYMRK (G) or A17 (M–O) roots. G and H, Toluidine blue-stained image. I, Bright-field image. J to O, Merged images of red (propidium iodide; DNA) and blue (Calcofluor; cell wall). Bars in G to I, J, and M = 100 µm. Bars in K and N = 20 µm. Bars in L and O = 5 µm. Arrowheads indicate peripheral vascular bundles in nodules, and the arrow indicates ITs. ic, Infected cell containing bacteroids; iz, infected zone; n, nucleus. P to U, TR25 and A17 roots were transformed with 35S::MtSYMRK-kd-GFP or 35S::AhSYMRK-kd-GFP. Both bright-field (left) and GFP fluorescence (right) images are shown for spontaneous nodules that are either solitary (P, R, and T) or in clusters (Q, S, and U) scored 3 weeks after transplantation in sterile agar plates on TR25/MtSYMRK-kd roots (P and Q), TR25/AhSYMRK-kd roots (R and S), and A17/AhSYMRK-kd roots (T and U). V and W, TRV25 or ipd3-1 roots were transformed with 35S::AhSYMRK-kd-GFP. Both bright-field (left) and GFP fluorescence (right) images are shown. No nodules were observed on transformed roots of TRV25/AhSYMRK-kd (V) or ipd3-1/AhSYMRK-kd (W) under identical growth conditions in agar plates. Bars in P to W = 1 mm.

Nodule organogenesis was not observed in the absence of S. meliloti in full-length AhSYMRK-transformed TR25 roots, indicating that deregulated activity of AhSYMRK-kd initiated the nodule organogenic program in TR25/AhSymRK-kd roots in the absence of any rhizobial trigger (Supplemental Table S1). To clarify whether the effect was caused by the enzymatic activity of AhSYMRK-kd, we repeated these experiments with 35S::AhSYMRK-kd K625E, where the kinase was inactivated by mutation of the invariant ATP binding Lys (Samaddar et al., 2013). Spontaneous nodules did not develop on TR25/AhSYMRK-kd K625E roots (Fig. 2F), indicating that nodulation under rhizobia-free conditions was solely caused by the deregulated catalytic activity of AhSYMRK-kd and not kinase-independent signals originating from the scaffold of the overexpressed kinase polypeptide.

The spontaneous nodules on TR25/AhSYMRK-kd, both solitary (n = 22) and merged (n = 30; Fig. 2, H and I), showed peripheral vascular bundles connected at their proximal end to the root vasculature, which was similar to S. meliloti-infected nodules (n = 10) generated in TR25/AhSYMRK roots (Fig. 2G). Therefore, spontaneous nodules follow the genuine nodule organogenic program that is elicited by rhizobia, but unlike rhizobia-induced nodules (Fig. 2G), they had an empty interior devoid of any bacteria (Fig. 2, H and I). Confocal images also revealed the bacteroid-free interior of the spontaneous nodules generated in TR25/AhSYMRK-kd roots (Fig. 2, J–L) compared with S. meliloti-infected nodules in wild-type A17 roots, where the nodule interior was full of bacteroids (Fig. 2, M–O).

To clarify whether the observed spontaneous nodulation was specific for AhSYMRK-kd, we transformed TR25 with 35S::MtSYMRK-kd(572–882)-GFP and cultivated them in sterile agar plates in the absence of S. meliloti. As shown in Figure 2P, MtSYMRK-kd, which has 91% sequence identity with AhSYMRK-kd, could successfully trigger 6.9 ± 0.97 spontaneous nodules per root system in TR25 (Supplemental Table S2) with the characteristic tendency of generating merged nodules (Fig. 2Q). This indicates that the property of triggering spontaneous nodulation by deregulated SYMRK activity is not a specific property of AhSYMRK-kd. Under this growth condition, TR25 plants transformed with AhSYMRK-kd showed 7.0 ± 1.22 nodules per root system (Fig. 2, R and S; Supplemental Table S2), showing the functional equivalence of AhSYMRK-kd and MtSYMRK-kd. It may be relevant to mention here that deregulated activity of neither AhSYMRK-kd (Samaddar et al., 2013) nor AhCCaMK-kd (Sinharoy and DasGupta., 2009) could induce spontaneous nodulation in A. hypogaea, where development of aeschynomenoid nodules are thought to be coupled with endocytosis of symbionts (S. Saha, A. Dutta, and M. DasGupta, unpublished data).

To clarify whether AhSYMRK-kd could generate spontaneous nodules in the presence of endogenous SYMRK, it was overexpressed in the wild-type M. truncatula (A17) plant. An average of 3.4 ± 0.51 spontaneous nodules developed per root system (Fig. 2T; Supplemental Table S2), and the nodulation was also noted to occur in clusters (Fig. 2U), indicating that endogenous SYMRK does not interfere with spontaneous nodule organogenesis. The number of nodules per root system was consistently higher in TR25 compared with that noted in A17 (Supplemental Fig. S5), suggesting a possible dominant negative effect of endogenous SYMRK.

Finally, it was important to understand whether spontaneous nodulation in the presence of autoactive SYMRK was dependent on functional CCaMK. We, therefore, transformed 35S::AhSYMRK-kd-GFP in TRV25 (ccamk null) mutants of M. truncatula (Lévy et al., 2004) and cultivated the plants in the absence of S. meliloti. As indicated in Figure 2V, TRV25/AhSYMRK-kd roots failed to develop any spontaneous nodule, indicating that nodule organogenesis triggered by AhSYMRK-kd involves a functional CCaMK. This observation was consistent with the role of SYMRK in the activation of calcium spiking as an upstream member of the Sym pathway (Miwa et al., 2006) and the role of CCaMK as the central regulator of nodule organogenesis (Hayashi et al., 2010; Madsen et al., 2010). Recently, it has been reported that nodule organogenesis in the presence of autoactive CYCLOPS bypasses the requirement for CCaMK, suggesting that CCaMK mediated phosphorylation of CYCLOPS to activate nodule organogenesis (Singh et al., 2014). However, autoactive CCaMK can rescue nodule organogenesis in null cyclops background, suggesting that alternative targets of CCaMK are able to substitute for CYCLOPS in nodule organogenesis (Yano et al., 2008). We, therefore, transformed 35S::AhSYMRK-kd-GFP in the ipd3-1 (cyclops null) mutant of M. truncatula (Horváth et al., 2011) to clarify whether CYCLOPS could be bypassed for generating spontaneous nodules by deregulated activation of AhSYMRK-kd. As shown in Figure 2W, in the absence of S. meliloti, we could not detect spontaneous nodules in ipd3-1/AhSYMRK-kd plants. This was unlike the case where the autoactive form of CCaMK could generate spontaneous nodules in cyclops mutant plants. Thus, activation of nodule organogenesis in the absence of rhizobia by deregulated activity of AhSYMRK-kd was dependent on CCaMK and CYCLOPS, which function downstream to calcium spiking in the Sym pathway.

Early Nodulin11 Expression Was Triggered by AhSYMRK-kd

M. truncatula Early Nodulin11 (MtENOD11) is a molecular marker for both early preinfection responses and later infection-related processes occurring within root as well as nodule tissues (Journet et al., 2001). To further assess the gain-of-function activity of AhSYMRK-kd, we introduced both 35S::AhSYMRK and 35S::AhSYMRK-kd into A17 containing pMtENOD11-GUS. In the absence of rhizobia, AhSYMRK-transformed roots occasionally showed nonsymbiotic GUS expression in the root cap with no activity in the root hairs (Fig. 3, A–C). In contrast, under the same conditions, almost 65% (13 of 21) plants transformed with AhSYMRK-kd showed scattered GUS expression in lateral roots (Fig. 3, E and F), suggesting that ENOD11 is activated by the deregulated activity of AhSYMRK-kd. The GUS activity was limited to a zone of epidermal cells associated with root hair outgrowth (Fig. 3G), which is similar to what is known as the normal expression zone of ENOD11 in S. meliloti-infected or Nod factor-treated roots (Journet et al., 2001). Additionally, in AhSYMRK-kd-overexpressed roots, ENOD11 expression was localized in the apex of the spontaneous nodules (Fig. 3H), which is identical to its expression zone in S. meliloti-induced nodules in plants overexpressing intact AhSYMRK (Fig. 3D). Thus, ectopic expression of AhSYMRK-kd did not affect the restriction of expression of ENOD11 within a strict pattern that was triggered by S. meliloti. The similarity in expression pattern of ENOD11 between S. meliloti-induced nodules and AhSYMRK-kd-induced nodules confirms that the spontaneous nodules follow the same organogenic signaling pathway and that they are genuine nodules. It should be noted that spontaneous nodules in M. truncatula, triggered by deregulated activation of 35S::MtCCaMK, also have properly localized expression of ENOD11, indicating that ectopic expression of neither autoactive SYMRK (this work) nor autoactive CCaMK (Gleason et al., 2006) was a hindrance toward proper development of spontaneous nodules.

Figure 3.

Induction of ENOD11 by AhSYMRK-kd. 35S::AhSYMRK-GFP or 35S::AhSYMRK-kd-GFP was introduced in M. truncatula (A17) pMtENOD11-GUS lines. Nonsymbiotic pMtENOD11-GUS expression in the absence of S. meliloti (A–C) and symbiotic GUS expression in the nodule apex in S. meliloti-infected nodules in pMtENOD11-GUS line/AhSYMRK roots (D). E to H, Scattered pMtENOD11-GUS expression in pMtENOD11-GUS line/AhSYMRK-kd roots in the absence of S. meliloti (E and F) with restricted expression in epidermal root hairs (G) or apices of spontaneous nodules (H). Bars in A and E = 2 mm. Bars in B and F = 500 µm. Bars in C and G = 50 µm. Bars in D and H = 1 mm.

AhSYMRK-kd Supports IT-Mediated S. meliloti Invasion But Fell Short of Releasing the Symbionts in the Nodule Cells

TR25/AhSYMRK-kd roots responded by formation of root hair curls and ITs within 2 WAI with S. meliloti (Fig. 4, A and B). The number of ITs observed in TR25/AhSYMRK-kd roots was 1.8 times higher than that observed in TR25/AhSYMRK roots (Supplemental Fig. S2). The deregulated activity of AhSYMRK-kd was solely responsible for generating these features, because root hair curling or ITs are not detected in TR25 (Endre et al., 2002). Like nodule organogenesis, the rhizobial response features are also expected to be mediated through CCaMK, because constitutive activity of CCaMK dispenses the requirement of SYMRK for successful infection of rhizobia (Hayashi et al., 2010; Madsen et al., 2010). Whatever the case, it is clear that signals required for triggering root hair curl or IT formation, which are primarily guided by membrane-bound receptors, do not require SYMRK to be anchored in the membrane in its native location.

Figure 4.

Features of S. meliloti invasion and colonization in AhSYMRK-kd-transformed TR25 roots. TR25 roots were transformed with 35S::AhSYMRK-kd-GFP and infected with S. meliloti expressing mRFP. A and B, Root hairs at 2 WAI shown as merged images of GFP and mRFP fluorescence; ITs (arrows) can be seen inside the curled root hairs (A). Magnified image showing the shepherds crook (double arrowhead) formation (B). Bars in A and B = 100 µm. C to G, White nodules developed in TR25/AhSYMRK-kd roots at 6 WAI shown as bright-field (top) and merged (bottom) images of GFP and mRFP fluorescence. Numerous nodules in a root system with rare colonization of rhizobia (indicated by arrowhead; C), uninfected spherical nodules (D), rhizobial colonization in nodule apex (indicated by arrowheads; E and F), and nodule interior (G). Bar in C = 5 mm. Bars in D to G = 1 mm. H and I, Longitudinal section of empty nodules developed in TR25/AhSYMRK-kd roots showing proper (H) and improper (I) vasculature. Bright-field (H) and toluidine blue-stained (I) images. v, Vasculature. Bars in H and I = 100 µm. J to N, Longitudinal section of nodules developed in TR25/AhSYMRK-kd roots with rhizobial colonization in the nodule apex (J–M) or nodule interior (N) shown as merged images of red (S. meliloti expressing mRFP) and blue (Calcofluor; cell wall). Magnified views of regions marked by solid, dotted, and dashed boxes in J are shown in K to M, respectively. ITs in L to N are indicated by staggered arrowheads. Bars in J and N = 100 µm. Bar in K = 50 µm. Bars in L and M = 10 µm.

Efficiency of nodule organogenesis in TR25/AhSYMRK-kd roots remains similar in the presence of S. meliloti (Supplemental Table S1). However, the phenomenon of merged nodule formation was rarely seen in the presence of the symbiont, indicating that an S. meliloti-induced signal affects the proximity of nodule primordia generation (Fig. 4C). S. meliloti-derived fluorescence was undetectable in the majority of the nodules formed in TR25/AhSYMRK-kd roots, indicating that most nodules are uncolonized by the symbionts (Fig. 4, C and D). Sectioning these nodules revealed their bacteroid-free interiors, but otherwise, they resembled genuine nodules with peripheral vascular bundles, like those seen in spontaneous nodules in the absence of S. meliloti (Fig. 4H). Intriguingly, in the presence of S. meliloti, a significant number of these empty nodules (23 of 30) had improper development of vasculature, indicating that the presence of rhizobia adversely affected the progress of nodule organogenesis (Fig. 4I). These observations are similar to observations in cyclops/CCaMKT265D and cerberus/CCaMKT265D roots, where spontaneous nodules with genuine nodule structure were generated in the absence of rhizobia but epidermal arrest of rhizobial invasion adversely affected nodule organogenesis in its presence (Yano et al., 2008, 2009). Thus, similar to CYCLOPS and CERBERUS, the ectodomain of SYMRK seems to have a role in the concerted progression of S. meliloti infection processes and nodule organogenesis.

We could detect S. meliloti in the nodule apex in only 11% of nodules (Fig. 4, E and F), and we could rarely (in 2% of nodules) detect colonization in nodule interior (Fig. 4G), indicating an overall uncoupling of the nodule organogenic program with symbiont colonization in the absence of the SYMRK ectodomain. Sectioning of nodules (n = 20) with symbionts colonized in the apex (Fig. 4J; Supplemental Fig. S6A) revealed entangled root hairs ending in an infection patch in the epidermal layer (Fig. 4K; Supplemental Fig. S6B). The magnified images of nodule-associated ITs are shown in Figure 4L and Supplemental Figure S6C. From these huge infection patches, ITs were rarely found to expand into the cortex without proper intracellular release of the symbionts (Fig. 4M; Supplemental Fig. S6D). It is possible that these infection patches are the same as the infection pockets noted in the symrk-14 mutant of L. japonicus, where these pockets are suggested to constitute an intermediate step toward attempting successful intracellular infection of nodule cortical cells (Kosuta et al., 2011). Such intermediate infection pockets preceding colonization of the nodule interior are also noted in nena-1 mutants (Groth et al., 2010). In 2% of nodules where could we detect colonization of S. meliloti in the nodule interior, sectioning (n = 10) revealed extensive IT formation throughout the nodule (Fig. 4N; Supplemental Fig. S6E), which is similar to what was noted in TR25/AhSYMRK roots (Fig. 1, E and F). However, unlike TR25/AhSYMRK nodules, symbionts were rarely released from ITs into the nodule cells in AhSYMRK-kd-transformed roots, indicating an important role of SYMRK ectodomain in intracellularization of rhizobia. Such extensive growth of ITs with rare or no release of symbionts is similar to the aberrant cortical infection phenotypes of SYMRK RNA interference lines in Sesbania rostrata or M. truncatula (Capoen et al., 2005; Limpens et al., 2005). Overall, we show that the unrestrained kinase activity of SYMRK in TR25 can successfully restore the root hair curls and epidermal ITs but that restoration of rhizobial colonization in the nodule interior was highly compromised in the absence of its ectodomain.

CONCLUSION

In conclusion, we have shown that deregulated activity of the intracellular kinase domain of SYMRK can activate spontaneous nodulation in the absence of rhizobia but that the ectodomain of SYMRK synchronizes the organogenic process with the internalization of the symbiont in its presence. Unlike autoregulated spontaneous nodulation under constitutive activity of CCaMK, deregulated activation of SYMRK hyperactivates nodule organogenesis and seems to have overcome the controls of both nodule number and the spatial positioning (Fig. 2, A–D). Whether the hypernodulation is caused by loss of autoregulation of nodulation through long-distance signaling (Nishimura et al., 2002) or loss of the negative regulatory role of NIN (Marsh et al., 2007) or hormones, like ethylene (Penmetsa and Cook, 1997), remains to be clarified. Whatever the underlying mechanism, it is clear that deregulated SYMRK activity adversely affects the signaling networks that have a physiological control over nodule number and spatial position.

MATERIALS AND METHODS

Plant and Rhizobial Strains

Medicago truncatula A17 and TRV25 seeds were from Jeanne Harris, TR25 seeds were from Giles Oldroyd and Christian Roger, pMtENOD11-GUS seeds were from David G. Barker, TRV25 seeds and Agrobacterium rhizogenes strain MSU440 were from Douglas R. Cook, ipd3-1 seeds were from Peter Kalo, and pBHR-mRFP-Sinorhizobium meliloti 2011 was from Ton Bisseling and Erik Limpens.

Constructs

The full-length Arachis hypogaea SYMRK (Samaddar et al., 2013), AhSYMRK-kd(573–883), and M. truncatula SYMRK-kd(572–882) were cloned into pENTR-D-TOPO (Life Technologies) and recombined into pK7FWG2 using LR-Clonase (Life Technologies; Karimi et al., 2002), thereby generating 35S::AhSYMRK-GFP, 35S::AhSYMRK-kd-GFP, and 35S::MtSYMRK-kd-GFP. AhSYMRK-kd(573–883) in pENTR-D-TOPO was used to generate AhSYMRK-kd K625E by site-directed mutagenesis and subsequently recombined into pK7FWG2, generating 35S::AhSYMRK-kd K625E-GFP (for details, see Supplemental Methods S1).

Generation of Composite M. truncatula Plants and Scoring Nodulation

M. truncatula seedlings were transformed with indicated constructs in A. rhizogenes MSU440 following a standard procedure (Boisson-Dernier et al., 2001). Transgenic hairy roots were screened for GFP fluorescence, and composite transgenic plants were transplanted in either agar plates containing buffered nodulation medium or vermiculite pots. Spontaneous nodulation was scored after 8 weeks in vermiculite pots or after 3 weeks in agar plates. For nodulation, S. meliloti expressing mRFP (Smit et al., 2005) was applied to plants 7 d after transferring to vermiculite. For observing ITs, plants were harvested 2 WAI, and for scoring nodules, they were harvested 6 WAI (details in Supplemental Methods S1).

Phenotypic Analysis, Histochemical Staining, and Confocal Microscopy

Images of whole-mount nodulated roots were captured using a Leica stereo fluorescence microscope M205FA equipped with a Leica DFC310FX digital camera (Leica Microsystems). Histological assay for checking GUS expression was performed according to Sinharoy et al., 2009. For microscopy of the nodule interior, 30-µm sections of fresh nodules were generated with a rotary microtome (RM2235; Leica Microsystems). Sections were stained with toluidine blue (0.05% [w/v]; Lobachemie) and imaged under an Olympus IX71 microscope. For confocal microscopy, sections were stained with indicated stains and imaged using a Leica TCS SP5 II AOBS microscope (Leica Microsystems). All digital micrographs were processed using Adobe Photoshop CS5 (details in Supplemental Methods S1).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. Phylogenetic position of AhSYMRK.

• Supplemental Figure S2. ITs in TR25/AhSYMRK and TR25/AhSYMRK-kd roots 2 WAI with S. meliloti.

• Supplemental Figure S3. Ultrastructure of S. meliloti-infected nodules formed on AhSYMRK-transformed A17 or TR25 roots.

• Supplemental Figure S4. Spontaneous nodule development in TR25 (nonnodulating symrk mutant of M. truncatula) by overexpression of AhSYMRK-kd in plants grown in vermiculite.

• Supplemental Figure S5. Spontaneous nodule development in A17 (the wild type) and TR25 (symrk mutant) of M. truncatula by overexpression of AhSYMRK-kd or MtSYMRK-kd in plants grown in sterile agar plates.

• Supplemental Figure S6. Ultrastructure of nodules developed on TR25/AhSYMRK-kd roots infected with S. meliloti.

• Supplemental Table S1. Spontaneous nodule development in TR25 (nonnodulating symrk mutant of M. truncatula) by overexpression of AhSYMRK-kd in plants grown in vermiculite.

• Supplemental Table S2. Spontaneous nodule development in TR25 (symrk mutant), TRV25 (ccamk mutant), ipd3-1 (cyclops mutant), and A17 (the wild type) of M. truncatula by overexpression of AhSYMRK-kd or MtSYMRK-kd in plants grown in sterile agar plates.

• Supplemental Methods S1. Methods in detail.

Glossary

IT

infection thread

LRR

Leu-rich repeat

WAI

weeks after inoculation