Does a Common Pathway Transduce Symbiotic Signals in Plant-Microbe Interactions?

Abstract

Recent years have witnessed major advances in our knowledge of plant mutualistic symbioses such as the rhizobium-legume symbiosis (RLS) and arbuscular mycorrhizas (AM). Some of these findings caused the revision of longstanding hypotheses, but one of the most solid theories is that a conserved set of plant proteins rules the transduction of symbiotic signals from beneficial glomeromycetes and rhizobia in a so-called common symbiotic pathway (CSP). Nevertheless, the picture still misses several elements, and a few crucial points remain unclear. How does one common pathway discriminate between - at least - two symbionts? Can we exclude that microbes other than AM fungi and rhizobia also use this pathway to communicate with their host plants? We here discuss the possibility that our current view is biased by a long-lasting focus on legumes, whose ability to develop both AM and RLS is an exception among plants and a recent innovation in their evolution; investigations in non-legumes are starting to place legume symbiotic signaling in a broader perspective. Furthermore, recent studies suggest that CSP proteins act in a wider scenario of symbiotic and non-symbiotic signaling. Overall, evidence is accumulating in favor of distinct activities for CSP proteins in AM and RLS, depending on the molecular and cellular context where they act.

Keywords: arbuscular mycorrhiza; legume nodulation; plant–microbe interactions; signaling pathways; symbiosis.

FIGURE 1

The common symbiotic pathway. Ten proteins (in green) have been identified as essential for both RLS and AM. In contrast with the conventional representation of signal transduction along the common symbiotic pathway (A), CSP protein location in different plant cell districts (B) highlights the existence of significant discontinuities between three subsets of CSP members. The receptor-like kinase SYMRK is believed to interact with other co-receptors (such as NFR1 and NFR5) at the cell membrane; its cytoplasmic domain binds the mevalonate (MVA) biosynthetic enzyme HMGR1(3-Hydroxy-3-Methylglutaryl CoA Reductase 1), linking SYMRK activation to MVA synthesis in the vicinity of the plasmalemma. Three nucleoporins (NUP85, NUP133, and NENA) localize at the nuclear pore complex, but their role in signal transduction as well as their functional connection with the previous set of CSP proteins remains unclear. The final stretch of the CSP localizes to the nucleoplasm: here, repeated oscillations in Ca²⁺ concentration (spiking) are generated with the contribution of MCA8 (a nuclear envelope-bound ATP-powered Ca²⁺ pump) and the potassium channels encoded by CASTOR and POLLUX genes. Such Ca²⁺-mediated signals are believed to activate the nuclear kinase CCaMK. In turn, CYCLOPS phosphorylation by CCaMK preludes to the regulation of gene expression. (C) Presents a proposal model of receptor complex assortment and CSP protein role in two separate pathways for the transduction of Nod factors or chito-tetraose in legume root hairs and atrichoblasts, respectively. Localized on the plasma membrane of root hairs (left), the extracellular LysM domain of NFR1 and NFR5 directly bind Nod factors (Broghammer et al., 2012). NFR1 and NFR5 form a complex with SYMRK (Ried et al., 2014) and HMGR1 (Kevei et al., 2007). MVA, produced upon HMGR1 activation (Venkateshwaran et al., 2015) is small enough to diffuse through the nuclear pore complex without the involvement of nucleoporins (NUPs); nevertheless, the inclusion of three NUPs in the CSP opens the possibility that at least one additional unknown protein (A) is involved, which could be activated as a consequence of MVA production and translocated to the nucleus through the action of NUPs. Ca²⁺ spiking activation in the nucleoplasm is generated by the recursive release of Ca²⁺ through unidentified channel proteins (C) in the nuclear envelope, combined with the continuous action of ATP-powered Ca²⁺ pumps like MCA8 (Engstrom et al., 2002; Capoen et al., 2011). To consider a minimal number of unidentified proteins, we can assume Ca²⁺ channels are directly activated by nuclear-imported A. CASTOR/POLLUX has been proposed to act in concert with Ca²⁺ channels (Oldroyd, 2013). The resulting Ca²⁺ spiking activates CCaMK through a RLS-specific conformational change requiring calmodulin (Shimoda et al., 2012; Poovaiah et al., 2013), which then modulates the activity of gene expression regulators, allowing the establishment of RLS (Oldroyd, 2013). A parallel pathway acts in atrichoblasts (right), where chito-tetraose (CO4) released by glomeromycetes is recognised by a complex possibly including CERK1 (Miyata et al., 2014), SYMRK, and HMGR1. Also in this case an additional protein (B) is proposed to be activated by MVA. B is then translocated to the nucleoplasm, where it activates Ca²⁺ spiking signals with a distinct, AM-specific signature (Kosuta et al., 2008; Russo et al., 2013). Consequently, CCaMK is activated in an AM-specific mode (Shimoda et al., 2012; Poovaiah et al., 2013), and its activity regulates AM-specific gene expression.