Bay et al., in this issue of New Phytologist (pp. 54–60), have reviewed the literature on cyanobacterial associations with mosses and then carried out experiments with selected species of feather mosses that have a vital role to play in primary production in boreal forests. They have confirmed data from other systems that the production of the motile form of Nostoc filaments, hormogonia, (Fig. 1) can be induced by secretions of host mosses. In addition, they show that some, as yet, unidentified chemical signal is needed for the hormogonia to move towards and colonize the host moss. Such signalling systems are well known from other nitrogen (N) fixing symbioses, especially some legumes (reviewed by Oldroyd, 2013), as well as other bryophytes such as liverworts and hornworts in which the cyanobacteria are endophytic, in specialized slime cavities of the gametophyte tissue (Meeks, 2003). In the endophytic bryophyte associations, cyanobacterial growth is slowed, as is photosynthesis and ammonium assimilation, while the heterocyst frequency and rate of N fixation are elevated from four- to five-fold, all compared to the free-living growth state (Meeks, 1998). However, there are no reports of a dramatic increase in heterocyst frequencies of cyanobacteria in epiphytic associations or of changes in photoautotrophic and N metabolism. Thus, current knowledge indicates that establishment of both epiphytic and endophytic bryophyte associations is dictated by signals from the plant partner that modulate development and behaviour of the hormogonial stage of the cyanobacterial partner, while altered cyanobacterial metabolism and enhanced N fixation occur only in the endophtytic associations. The latter assumption needs to be addressed.
Two further results of significance from Bay et al. are first, using 15N, N is rapidly transferred from the Nostoc to the host, and second, that the signalling process is geared to the N requirements of the host. In the endophytic hornwort association, N is made available from the cyanobacterium to the plant partner as ammonium (Meeks, 1998), most likely via plant transfer cells that are elaborated into the slime cavity after infection by the cyanobacterium (Rodgers & Stewart, 1977). The form of N transferred to the moss was not identified by Bay et al. and the efficiency of transfer is not known in either of the endophytic or epiphytic bryophyte associations. When there is an external supply of N, fixation is not required and when the N demand is very high, in particular when sporophytes are being produced, colonization by cyanobacteria is stimulated.
‘Has cyanobacterial N fixation been undervalued?’
Combined N has been shown to inhibit production of a hormogonium inducing factor (HIF) in the endophytic hornwort association, but it does not hinder infection if added after production of HIF (Campbell & Meeks, 1989). Combined N also represses N fixation after infection in the hornwort association (Meeks, 1998). Thus, again, there are clear similarities in the epiphytic and endophytic bryophyte associations in control of cyanobacterial development and behaviour by the plant partner in response to its metabolic demands. On a much larger scale, legume trees (species of Inga) in a lowland tropical forest, were found to adjust their nodulation and N fixation from zero to high levels depending on demand (Barron et al., 2011). These observations are important when trying to assess the extent of N fixation at the ecosystem level, rather than in agricultural systems. The problems of this notoriously difficult task have recently been considered by Vitousek et al. (2013). These workers have discussed the diversity of regulatory strategies and suggested that new models are needed in order to understand the effects of anthropogenic global change on the N cycle.