A nonsymbiotic root hair tip growth phenotype in NORK-mutated legumes: implications for nodulation factor-induced signaling and formation of a multifaceted root hair pocket for bacteria

Abstract

The Medicago truncatula Does not Make Infections (DMI2) mutant is mutated in the nodulation receptor-like kinase, NORK. Here, we report that NORK-mutated legumes of three species show an enhanced touch response to experimental handling, which results in a nonsymbiotic root hair phenotype. When care is taken not to induce this response, DMI2 root hairs respond morphologically like the wild type to nodulation factor (NF). Global NF application results in root hair deformation, and NF spot application induces root hair reorientation or branching, depending on the position of application. In the presence of Sinorhizobium meliloti, DMI2 root hairs make two-dimensional 180 degrees curls but do not entrap bacteria in a three-dimensional pocket because curling stops when the root hair tip touches its own shank. Because DMI2 does not express the promoter of M. truncatula Early Nodulin11 (ENOD11) coupled to beta-glucuronidase upon NF application, we propose a split in NF-induced signaling, with one branch to root hair curling and the other to ENOD11 expression.

Figure 1.

Cartoon of Root Hair Cytoarchitecture. Simple representations of the cytoarchitecture of developmental stages of root hairs. (A) Growing root hair. The subapex of the root hair is filled with cytoplasm (c), and the nucleus (n) is at the base of this area. The shank of the root hair is filled with the central vacuole (v) and cortical cytoplasm. (B) Early growth-terminating root hair. The first sign of growth termination is that the central vacuole overtakes the nucleus and, therefore, that the subapical region with dense cytoplasm is getting shorter. (C) Late growth-terminating root hair. The central vacuole expands more and more into the subapex and smaller vacuoles, or extensions of the central vacuole appear into the remaining cytoplasm. (D) Full-grown root hair. The nucleus has lost its fixed position in the root hair. The vacuole completely fills the root hair and is surrounded by a thin layer of cytoplasm.

Figure 2.

Medium Change–Induced Touch Response in M. truncatula. (A) Young growing M. truncatula dmi2-1 root hair just before medium change. (B) M. truncatula dmi2-1 root hair in the same region on the root as the one in (A), just after medium change. In just 1 min, the cytoplasm has reorganized such that it evenly surrounds the central vacuole, creating a cytoarchitecture of a full-grown root hair. (C) Percentage of M. truncatula root hairs with a cytoarchitecture of a growing root hair before and after medium change. Before medium change, this percentage in the wild type and dmi2-1 is ∼100%. After medium change, this percentage decreases in both the wild type and dmi2-1. In time, the wild type shows recovery, whereas in dmi2-1, the typical cytoarchitecture does not recover. Monitored were six roots per wild type and dmi2-1, respectively, >50 root hairs per root. Bars in (A) and (B) = 10 μm.

Figure 3.

Time Series of a Growing M. truncatula dmi2-1 Root Hair before and after Medium Refreshment via Gentle Perfusion. The gentle way of refreshing the growth medium does not disturb root hair cytoarchitecture and growth and, therefore, root hair morphology. Images were taken every 10 min, except of the one 1 min after medium refreshment. Bar = 15 μm.

Figure 4.

Time Series of a M. truncatula Wild-Type and dmi2-1 Root Hair before and after Touch with a Needle. (A) Wild-type root hair. Time 0 is the root hair just after needle touch. Needle touch does not affect the cytoarchitecture and/or root hair growth. (B) dmi2-1 root hair. Immediately after touch with the needle, vacuoles appear, which increase in time, and the root hair stops growing. Images were taken every 30 s. Bars = 15 μm.

Figure 5.

Percentage of Wild-Type and dmi2-1 Root Hairs Showing Has and RHD in Hairs That Were Growing at the Time of Rough NF Application. Forty-five minutes after application, there is already a clear difference between the wild type and dmi2-1. A significant higher percentage dmi2-1 root hairs shows swelling of the root hair tip. At 180 min after application, 59% of the dmi2-1 root hairs, which were growing at the time of NF application (but became growth terminating because of the medium change), show NF-induced RHD. In the wild type, only 12% of the root hairs, which were growing at the time of NF application, show NF-induced RHD. Monitored were six roots for the wild type and dmi2-1, respectively, >50 root hairs per root.

Figure 6.

Medium Change–Induced Touch Response in M. sativa and L. japonicus NORK mutants. Medium change experiment as described in Figure 2, but with other species and their respective NORK mutant. Both for M. sativa (A) and L. japonicus (B), the same trend is visible. Medium change induces a decrease in the percentage of root hairs with a cytoarchitecture of a growing root hair, from which the wild types are able to recover and the NORK mutants not. Monitored were six roots for M. sativa wild type and MNNN1008 and four roots for L. japonicus wild type and Cac41.5, respectively, >50 root hairs per root.

Figure 7.

Difference of L. japonicus Wild Type and Cac41.5 When Grown in Liquid Medium between Glass Slides. (A) Wild-type root. The root hairs have a normal diameter and show a normal elongation pattern. (B) L. japonicus Cac41.5 root before medium change. Most of the root hairs are short, swollen, and/or branched. Only root hairs with a proper shape and cytoarchitecture similar to the M. truncatula root hair in Figure 2A were followed during the medium change experiments. Bars = 30 μm.

Figure 8.

Responses of M. truncatula dmi2-1 Pro_MtENOD11:GUS Root Hairs to Spot Application of 10⁻⁹ M NF. (A) Time series of reorientation of root hair growth after spot application of 10⁻⁹ M NF on the side of the tip of a growing M. truncatula dmi2-1 root hair. At 20 min, the reorientation is visible, and at 40 min, it is pronounced. (B) Reoriented M. truncatula dmi2-1 root hair carrying the Pro_MtENOD11:GUS reporter gene, 60 min after spot application and 24 h after GUS reaction. In all of 13 root hairs, no GUS was detected, showing that the root hair does not express MtENOD11 after NF spot application. (C) Time series of the formation of a root hair branch after spot application of 10⁻⁹ M NF 30 μm below the tip of a growing M. truncatula dmi2-1 root hair. At 17 min after application, the first sign of the formation of a branch is visible, which becomes more pronounced later in time. (D) The same root hair as in (C), 60 min after spot application and 24 h after GUS reaction. In all five root hairs, no GUS was detected, showing that the root hairs do not express Pro_MtENOD11:GUS after NF spot application. (E) Pro_MtENOD11:GUS expression in a wild-type root hair, 60 min after NF spot application and 24 h after GUS reaction. Clearly, the reorientation toward the site of application and the GUS expression are visible. (F) Growing C71 (dmi1-1) root hair, 45 min after NF spot application. (G) Growing TRV25 (dmi3-1) root hair, 55 min after NF spot application. Both the C71 and TRV25 root hairs show root hair reorientation toward the site of application in a timeframe comparable to wild-type root hairs. Arrowheads point to the site of application, and bars = 10 μm.

Figure 9.

Number of Root Hair Curls per Centimeter Inoculated Root. (A) For the wild type and dmi2-1, a similar number of curled root hairs per centimeter of inoculated root were counted. Error bars are standard deviations. (B) Of the wild-type curled root hairs, 14.5% had a complex three-dimensional bacteria-entrapping structure. The other 85.5% were single-faceted two-dimensional 180° curls without entrapped bacteria. On dmi2-1, all root hair curls were single-faceted two-dimensional 180° curls without entrapped bacteria.

Figure 10.

M. truncatula Wild-Type and dmi2-1 Root Hairs Curl around S. meliloti 2011-GFP. (A) Reoriented wild-type root hair in the presence of S. meliloti. (B) Bright-field image of a wild-type root hair curl with an infection thread. Note the complex multifaceted three-dimensional structure of the curl. (C) Bright-field image of a curled dmi2-1 root hair. Arrowhead points to the furrow at 180° curling. (D) Projection of 30 images from a confocal laser scanning microscope Z-stack of a wild-type curled root hair, entrapping a GFP-expressing bacterial colony (arrow). The cell wall (red) was counterstained with 0.1% propidium iodine. Note the multifaceted three-dimensional structure of the root hair curl. (E) Projection of 35 images of a Z-stack from a wild-type curled root hair with a bacterial colony in a closed pocket. (F) Projection of 20 images of a Z-stack from a dmi2-1 root hair, curling in the presence of bacteria but unable to entrap them. Note the bacteria on the outside of the root hair. Monitored were four roots per wild type and dmi2-1. Bars = 15 μm.

Figure 11.

Putative NF-Induced Signal Transduction Pathways. (A) Pathway proposed by Catoira et al. (2000). (B) Pathway proposed by Wais et al. (2000). The difference between these two pathways is the positioning of Ca²⁺ spiking between DMI1-DMI2 and DMI3, and the combination of Has and root hair branching into RHD. (C) New pathway, in which our results are integrated into the already published results. From our results, we can conclude that DMI2/NORK, DMI1, DMI3, and its downstream Pro_MtENOD11:GUS expression are not required for the early NF-induced morphological responses. Therefore, we propose that very soon after NF perception via LysM proteins and before DMI2/NORK, DMI1, and DMI3, the pathway forms a split. One part, including the rapid NF-induced Ca²⁺ influx, branches off to root hair reorientation/Has and RHD, and the other part via DMI2/NORK to Pro_MtENOD11:GUS expression. Because we only used Pro_MtENOD11:GUS as a NF reporter gene, we can only make a statement about this gene, and we do not want to exclude the possibility that in DMI2/NORK mutants other NF-induced genes are expressed.