Shedding light on an underground problem.

I this issue of PNAS, Nishimura et al. demonstrate that ASTRAY, the legume homolog of HY5, a transcription factor that mediates light signaling and photomorphogenesis above ground, also affects the interaction between legume roots and their rhizobial endosymbionts underground (1). At first glance, this is a confusing result. What is a transcription factor that mediates light signaling doing regulating an underground (dark!) plant–microbe interaction? In fact, in Arabidopsis, HY5 regulates root architecture in addition to shoot photomorphogenetic events such as greening and hypocotyl elongation (2). Still, fitting ASTRAY into the nodulation pathway is a puzzle. Nodulation-deficient mutants have allowed the identification of genes required for nodulation, but few of these genes have been cloned. Until recently, progress was hampered by the fact that these mutations were scattered throughout the crop legumes. With the advent of two model legumes, Lotus japonicus and Medicago truncatula, previously scattered efforts to investigate multiple crop legumes gradually turned into an intense focus on the genetics and genomics of these two species (3, 4). Extensive genetic screens of L. japonicus and M. truncatula have generated a large number of nodulation mutants, which, added to the wealth of nodulation mutants in pea (the original genetic organism!), have provided the foundation for a genetic analysis of nodulation (5–8). The recent cloning of LjNin1, which encodes a putative transcription factor required for L. japonicus nodulation (9), and an LRR-type receptor required for nodulation of L. japonicus, M. truncatula, alfalfa, and pea has provided the first molecular insight into this genetic pathway (10, 11). The cloning of these nodulation genes has its roots in research begun more than 140 years ago. Legume root nodules first attracted the attention of researchers at about the same time that Darwin was investigating light signaling in oat seedlings and Koch was laying down his postulates. We now know that nodule formation is a complex morphogenetic process initiated as the result of a molecular conversation between plant and bacterial partners. A nodulation signal from the bacteria induces cells in the cortex of the plant root to start dividing to form a nodule primordium, which the bacteria subsequently infect. Cells of the nodule primordia differentiate into the many cell types that make up a mature nodule and organize into a radially symmetric organ containing a vascular system that connects with the root vasculature. Inside the nodule, the internalized bacteria differentiate into a specialized form that can fix atmospheric nitrogen. The decision of whether to nodulate is influenced by the availability of nitrogen in the soil, ethylene levels in the root, how many nodules it already has, and probably much more (12–14). Interestingly, although the bacterial partner signals when and where to form a nodule, nodule formation is actually a plant developmental program. We know this because of two curious observations: (i) plant mutants exist that can nodulate in the absence of rhizobia (15, 16); and (ii) application of auxin transport inhibitors or localized cytokinin application can induce the formation of pseudonodules (17, 18). How did legumes acquire a nodule developmental program? The fossil record is no help here, but looking at phylogenies can be. Thousands of legumes nodulate, but only one nonlegume. Are special genes that exist only in nodulating legumes required for nodulation? Or were genes required for other aspects of plant development recruited to function in a nodulation pathway? That nodulation may have evolved as many as three independent times within the legume family suggests existing genes or pathways may have predisposed legumes to form root nodules (19). Symbiotic genes like astray that also have effects outside the nodule provide a glimpse of preexisting genes or pathways that were recruited to help form a nodule. It is the pleiotropic genes that, by maintaining an original function yet displaying their new role in nodulation, reveal their prenodule origins. Other pleiotropic symbiotic genes have been described, most notably the ethylene-insensitive mutant sickle (20) and the root architecture mutant har1 (21). The phenotypes of these mutants have demonstrated the link between other aspects of plant development and nodulation. astray is the first of these pleiotropic symbiotic genes to have been cloned (1). Like its Arabidopsis homolog, HY5, ASTRAY mediates photomorphogenetic development. When the HY5 or ASTRAY genes are mutated, light-grown plants behave as if they were in the dark. In addition, astray mutants form excessive nodules. Several years ago, Kijne and colleagues observed that vetch roots grown in light were blocked from nodulating by an ethylene-mediated process (22, 23). This effect is less severe in other legumes but can occur there as well. Might astray mediate the light inhibition of vetch nodulation? This is a possibility because ethylene can inhibit the supernodulation of astray roots (24). ASTRAY’s dual role in light signaling and nodulation indicates participation in two pathways. Does it interact with the same pathway components when it functions in light signaling and nodulation? The role of HY5, ASTRAY’s Arabidopsis homolog, is well understood in light signaling. In the dark, COP1 binds HY5, targeting it for degradation. When light is perceived by cryptochromes (blue UV-A light) or phytochromes (red far red light), COP1 levels in the nucleus fall and HY5 is free to bind promoters and activate transcription (25, 26). Does HY5 interact with COP1 in roots? When grown on Petri plates, roots are exposed to the light, but this is an artificial situation. In the soil, roots encounter little light. Might HY5 mediate another environmental signal in roots? COP1 has been shown to interact with other transcription factors as well as HY5. Perhaps in the root, HY5 can interact with other WD40 domain proteins besides COP1 (Fig. 1). If HY5 does mediate a nonlight signal in roots, do HY5 and ASTRAY mediate