Signaling mechanisms are vital for all living organisms. Plants and animals use long-distance signals to coordinate and adjust their growth in response to endogenous and environmental cues. These signals transmit messages throughout the whole organism to achieve biological homeostasis. Plants use hormones, RNA, proteins, short peptides and lipids for long distance signaling in defense against pathogens, in response to numerous abiotic and biotic stresses and in developmental processes such as flowering branching and nodulation[5, 6]. Our lab uses Medicago truncatula to study long distance regulation; specifically the root-to-shoot and shoot-to-root signals that control nodule number and to understand the regulatory network involved in this process.
M. truncatula is an excellent model to study legume biology due to its small diploid genome (500 Mb), self-fertility, ease of transformation, short life cycle, high level of natural diversity and a wealth of genomic resources. In addition, the bacterial and fungal symbionts of M. truncatula that lead to the fixation of nitrogen and the increased uptake of phosphorus are well-characterized[8, 9]. In both rhizobial and arbuscular mycorrhizal symbioses, the establishment and maintenance of symbiosis requires expensive plant resources, specifically energy. To balance this expense with other plant needs, legumes have a negative feedback inhibition system called autoregulation of nodulation (AON)[11, 12]. Through AON, the early symbiotic events occurring in a root and leading to nodule organogenesis or to arbuscule formation systematically affect later symbiotic interactions through transportable signals. Efforts have been focused on locating time and space-specific root and shoot events including sensor, integrator and effector molecules. Grafting and split root inoculation systems can be very informative when combined with current molecular genetic tools to decipher the signaling. However, very little grafting and split root work has been done in model legumes, with historical work in less genetically tractable plants such as bean, pea, soybean, clover and vetch, some of which have stems much larger than those in model systems.
Establishment of symbiosis in one part of a legume root affects further symbiotic events in other parts of the root inoculated later, and this phenomenon was initially elucidated using split root experiments. In these split root systems, two roots in one plant are partitioned in time and space allowing prior inoculation of one root system (Root A) to systematically regulate nodulation from the separate inoculation of the second root (Root B). Around thirty years ago, it was reported that the suppression of nodule development in the Root B side of the split root system in soybean is associated with prior inoculation of the Root A side. Five years later, Olsson et al. reported the lack of systemic suppression of nodulation in supernodulating soybean mutants. Using a split root experiment, Tang, Robson and Dilworth showed that iron is required for nodule initiation in lupine, emphasizing the direct and indirect impact of mineral nutrient deficiency on symbiosis. Application of either rhizobia or Nod factors to the Root A side of a split root system inhibits nodulation in the B root system, suggesting that Nod factors are enough to elicit the autoregulatory responses in vetch. In clover a non-nodulating strain of Rhizobium trifolii inoculated on Root A was unable to inhibit infection by the wild type strain on Root B, suggesting a minimum requirement of Nod factor to initiate the plant inhibitory response. In work by Laguerre, et al. one root system was nodulated with a nitrogen-fixing bacterial partner while the other root system was nodulated with non-fixing partner, resulting in a plant that compensated for the local nitrogen limitation in the root with non-fixing bacteria. The same group had shown that in split root plants when one root is in a nitrogen-limited condition and the other receives nitrogen, both nitrogen fixation activity and net nitrogen uptake by the root system in the nitrogen-limited condition was higher in the M. truncatula sunn-2 mutant versus wild type plants. The authors suggested a secondary response of growth stimulation of pre-existing nodules in the wild type and sunn-2 mutant. Autoregulation signals initiated by either nodulation or mycorrhization on Root A in alfalfa systemically influence both rhizobial and arbuscular mycorrhizal colonization of Root B in a split root system without preferential selection. Also in alfalfa inoculated with mycorrhizae, isoflavonoid levels are systematically regulated in the uninoculated Root B upon prior inoculation of Root A, suggesting the involvement of isoflavonoids in the long distance autoregulation of arbuscular mycorrhizal symbiosis.
Developing a model of signal transduction by comparison across these experiments is difficult due not only to the use of many less tractable and less well developed molecular genetic systems with both determinate and indeterminate nodule development, but also to the use of a broad range of compartmentalization techniques to separate the split roots in various growth systems. For instance, PVC piping elbows have been used in soybean and split root tubes have been used in soybean and vetch[5, 13, 15]. Split root plate assays were done using Trifolium subterraneum and Lotus japonicus by separating the roots with plastic dividers supported with 0.6% water agar or by removing the center of the agar to create separate root environments. The limitations of these techniques include inability to consistently control various factors known to affect nodule regulation, such as ethylene concentration in plate systems[22, 23] and rhizobial cross contamination. Moreover, the effect on nodulation of root exposure to light, balancing the size of the root systems before treatment, the types of containers and the composition of the growth media including the amount of water and the concentration of various root exudates that affect nodulation were not consistently controlled. The above approaches were also targeted for very small laboratory scale applications, often with only 3–5 replicates (plants) per experiment.
In addition to the unintentionally introduced variables in the above experiments, many key factors in autoregulation remain unexplored in a single system/species. These factors include the time the AON signal takes to suppress nodulation and the stages of nodule initiation targeted by the AON signal. Except for the nitrogen experiments described above, the split root technique has not been used in M. truncatula. We were unable to find efficient examples of the use of the technique in model plants with growth parameters similar to M. truncatula (small stems approximately 0.1 cm in diameter). Hence improving the existing split root protocol to consistently generate large numbers of grafted plants was imperative for our AON investigations.
Another technique, reciprocal grafting, is also a valuable tool to study the remote and local interactions of various genotypes and systemic signals. For shoot-to-root reciprocal grafting in Arabidopsis, wedge grafting has been commonly employed, and adapted for other model plants with slight modifications[25–30]. Shoot-to-root reciprocal grafting allows researchers to examine the systemic signals and separate gene functions in above and below ground parts of the plant. The major limitations of the technique, especially in small plants, is the need for agar, parafilm, medical tubing or other physical support materials which interfere with inspection of the graft union and slow the healing of the union, sometimes influencing later plant development. The success rate for M. truncatula reciprocal grafts is reported to be as low as 8% and we have observed a rate of 50% depending on genotype in our previous work[27, 29, 30]; Lucinda Smith personal communication].
Despite the low success rate, reciprocal grafting is quite informative. Reciprocal grafting between a Zn hyperaccumulator, Thlaspi caerulescens, and a Zn nonaccumulator, Thlaspi perfoliatum, showed the relative importance of roots and shoots in Zn hyperaccumulation and hypertolerance. The discovery of Flowering Locus T (FT) protein as a long distance signal moving from the leaf to the apex through phloem to induce flowering in Arabidopsis was done with grafting. Grafting analysis provided evidence that the shoot genotype controls the supernodulating phenotype in the autoregulation defective mutants har1 and klv in Lotus japonicus[34, 35], sym29 in Pisium sativum, nark in Glycine max, sunn and lss in M. truncatula[27, 29]. For example, grafting sunn and lss scions on wild type (A17) rootstock produced a hypernodulation phenotype whereas the reciprocal grafting of A17 scion on either sunn or lss rootstock gave wild type nodulation phenotype[27, 29]. Grafting also revealed the action of the r oot d etermined n odulation 1 mutant rdn1 in which, unlike the examples above, the root genotype controls the hypernodulation phenotype. In cases of root-determined hypernodulation, the cause could be a defect in either the synthesis or transmission of the root derived factor or the transport and/or perception of the shoot derived descending inhibitory signal. Distinguishing between these possibilities requires a plant with roots of two different genotypes. Working in pea, researchers used approach grafting between wild type pea and lines with mutations affecting AON to reveal that early nodulation events prior to root hair curling cannot induce the AON signal, demonstrating that AON starts after root hair curling but before visible cortical and pericycle cell division. However approach grafting, in which two complete plants are joined in the stem region, adds the complication of having two shoots of different genotypes that may vary in their vascular connections to the roots and their production of the unknown shoot derived inhibitory signal compared to a single shoot, making the findings from these experiments difficult to interpret definitively. Therefore we developed an inverted-Y grafting technique to provide an extra dimension to the split root experiment by partitioning the two roots of the same plant not only in time and space but also in genotype. This experimental approach allows for the dissection of function of the gene products involved in the regulatory circuit without the complications created by approach grafting.
This methodology report describes these highly efficient split root and inverted-Y grafting protocols in M. truncatula. Our techniques provide simple ways of generating many root systems to dissect long distance signaling. These can be either split root experiments where the effect of a treatment on one root is detected on the second root or inverted-Y graft experiments where plants with two different root genotypes are used to study the root-to-shoot and shoot-to-root signals and individual gene actions. We also report a modification of the inverted-Y technique that allows rapid generation of a large number of reciprocally grafted plants with a single shoot and single root.