In the absence of any source of N other than atmospheric N2, the rate of N2 fixation could be adjusted in response to environmental changes either through a variation in the amount of nodule tissue (number of nodules and/or their size) or through a variation in nitrogenase activity (Hartwig, 1998). In D. trinervis, the number and distribution of root nodules is tightly regulated by an autoregulatory feedback mechanism that is induced before onset of N2 fixation (Valverde & Wall, 1999b; Wall, 2000). The rate of nodule growth (and the proportion of nodule biomass) is more likely to be regulated by another feedback mechanism, related to the N content of the plant (Valverde et al., 2000; Wall, 2000). In this study, we show that the general plant growth (shoot height, plant d. wt and shoot biomass allocation) of symbiotic D. trinervis was limited to a greater extent than nodule formation by P (Figs 1, 2, 4). Whereas accumulation of plant biomass was impaired at concentrations of P < 50 µM (Fig. 1), nodule initiation was not affected by such low P supply (Fig. 3). Moreover the ability of seedlings to regulate the number and location of nodules was not altered (Figs 3, 4), suggesting that autoregulation of nodulation (Valverde & Wall, 1999b) was not affected by the concentration of P in mineral solution. However the size of nodules and the proportion of nodule tissue was stimulated by P (Fig. 3). These data suggest that the feedback regulation of nodulation in D. trinervis, which might be mediated by a systemic mechanism linked to the N status of the plant (Parsons et al., 1993; Wall, 2000), is modulated by the availability of P. In order to support this hypothesis we found that when the proportion of nodule biomass in the plant was plotted as a function of the N : P ratio in leaves, a strong correlation was found (Fig. 9), suggesting that this ratio affects the feedback regulation of nodulation and nodule growth. Similar results were observed in soybean (Israel, 1987; Drevon & Hartwig, 1997), clover (Almeida et al., 2000; Hellsten & Huss-Danell, 2001) and alfalfa (Drevon & Hartwig, 1997). A negative correlation between leaf or shoot N : P ratio and the proportion of nodule tissue was also evident in studies where other N sources where present, or where mycorrhizal fungi colonized nodulated roots (Jha et al., 1993; Russo et al., 1993). Thus, although other fluxes of N and P are present in a symbiotic plant, the leaf N : P ratio would seem to be a better indicator of nutrient status for the purpose of feedback regulation of nodulation (Wall et al., 2000) through a systemic response as proposed by Parsons et al. (1993). A possible mechanism for this response would be that a phloem-delivered signal of N-status would interact with a metabolic pathway whose rate is limited by P (Wall, 2000). In this way, the positive effect of P on the growth of symbiotic plants could be explained as a specific effect of P on the regulation of nodule growth, rather than as a general positive effect on host plant growth (Robson, 1983). The fact that the proportion of nodule tissue was dependent on P content (Figs 3c, 9) gives support to such an explanation. If nodule growth were not affected by P, there is no reason why the proportion of nodule biomass in the plant should have changed in response to P supply, that is, nodule growth would have been accommodated to plant growth rate. A time course analysis of changes in plant biomass and nodule biomass following a change in P supply (Israel, 1993) would give additional support to such hypothesis. Variation of P supply would modify the rate of nodule growth and, subsequently the rate of plant growth would reach a new stationary growth rate value. The underlying biochemical mechanism for the participation of P in this proposed regulation of nodulation remains unknown and deserves further studies.
It is assumed that P is required for nodule initiation (Marschner, 1995; Huss-Danell, 1997). However, our results suggest that this would not be the case for D. trinervis. The formation of nodules was not altered when the supply of P was limiting plant growth (Figs 1, 3). Seedlings that formed a normal number of nodules with low P supply could also regulate their distribution (Fig. 4), whilst, seedlings that did not receive any external supply of P and that showed a clear P deficiency for growth, could develop some nodules (Fig. 6). Differentiation of Frankia into vesicles was observed within these nodules, although carbohydrate metabolism was apparently impaired in noninfected cells of the symbiotic tissue (Fig. 7). These results collectively indicate that D. trinervis has a higher P requirement for vegetative and nodule growth than for nodule initiation and early development. This was not the case for other actinorhizal plants like Alnus and Casuarina, which are more dependent on an adequate P supply to establish the symbiosis with Frankia (Sanginga et al., 1989; Yang, 1995; Wall et al., 2000). In these actinorhizal hosts, a good P supply stimulates the formation of nodules, as was noted for Alnus glutinosa by Quispel (1958). The response of early nodulation to P supply might be related to the content of P in the seeds. A high P content in the seeds could facilitate the establishment of the symbiosis though further growth would be more dependent on external P supply. In account with this possibility, we observed that nodule development was re-established in P deficient seedlings once P nutrition was resumed (Fig. 8), suggesting that some infection events had occurred at the moment of inoculation but that their development was arrested by lack of P (Valverde & Wall, 1999b).
Seeds of D. trinervis had 6–16 times more P than those of Alnus incana, Alnus acuminata and Casuarina cunninghamiana (Table 1). In these genera, and contrary to what we present here for D. trinervis, positive effects of P supply on early nodulation have been observed (Yang, 1995; Wall et al., 2000).
The average content of extractable P in the site where seeds were collected was 4.4 µg per gram of soil (Chaia, 1997). Thus, seeds of D. trinervis (with a P content of approx. 4 mg of P per g) concentrate P approx. 1000-fold. This accumulation of P in seeds could be interpreted as an adaptive strategy to ensure initial growth of the seedling after germination and establishment of the symbiosis with Frankia. The requirement for external P would arise later on and may be acquired directly by root exploration of soil or be complemented by the establishment of an additional symbiosis with mycorrhizal fungi. Nodulated roots of Discaria trinervis were always found to be mycorrhizal in nature (S. Fontenla pers. comm.).