Silencing ribulose-1,5-bisphosphate carboxylase/oxygenase expression does not disrupt nitrogen allocation to defense after simulated herbivory in <em>Nicotiana attenuata</em>

Abstract

Plant defense strategies against herbivore attack are generally categorized into resistance and tolerance strategies¹ and several theories have been proposed to explain these strategies.² One of the best experimentally supported theories is the optimal defense theory which states that defenses are adaptive but have allocation costs, since resources can either be invested into growth and fitness or into defense.³5 This growth-defense trade-off can be minimized by using inducible defenses which are only produced upon attack.²67 Consequently, plants optimize their allocation of growth-limiting resources to resistance and tolerance strategies or to growth and reproduction according to herbivore pressure.⁸ These growth-defense trade-offs are thought to be more stringent in the case of limited resources such as nitrogen (N) which affects the growth and fitness of plants and of the herbivores that feed on them. At the molecular level, components needed for growth such as sugars, amino acids and proteins, may also serve as a direct source of carbon and N for the biosynthesis of defensive chemicals.⁹12 It has been hypothesized that one major source of N for the production of defensive metabolites may be proteins,⁹10 and among these one potentially important N source is the photosynthetic protein ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO), which is the most abundant protein in foliar tissues¹³15 and is thought to serve as a N storage protein.¹⁶ This hypothesis has rarely been tested at the level of individual compounds, due to technical difficulties in comparing the allocation into different classes of molecules such as large proteins and small molecular weight defensive metabolites (e.g., alkaloids, polyamine conjugates) in a common currency.

One major resistance mechanism of the wild tobacco Nicotiana attenuata, a model plant for the study of plant-herbivore interactions, is the production of the N-containing defense metabolites nicotine and phenolamides, caffeoyl-putrescine (CP) and dicaffeoyl-spermidine (DCS). In a recent study, we used ^{N pulse-labeling and a novel LC-MS quantification method for proteins to quantify growth-defense trade-offs using N as a common currency after simulated herbivory using oral secretions (OS) of the specialist tobacco hornworm} Manduca sexta.¹⁷ Our results showed that although OS-elicited N. attenuata plants show large decreases in total soluble protein (TSP) and RuBisCO pools, the N required for synthesis of the metabolically dynamic defensive phenolamides, caffeoyl-putrescine (CP) and dicaffeoyl-spermidine (DCS), does not come from the putative storage protein RuBisCO, but likely from recently assimilated N.

Here we use a previously described transgenic line of N. attenuata stably silenced in the expression of RuBisCO using an antisense construct,¹⁸ hereafter called asRUB plants. At the whole plant level, these asRUB plants were shown to have a reduced accumulation of the RuBisCO protein, a 25% reduction in photosynthesis rates, and stalk lengths were transiently shorter, but they eventually attained the heights of empty vector-transformed plants.¹⁸ In the present study, we quantify N based growth-defense trade-offs in asRUB plants and compare them to WT plants in the same experimental design used in the previous study.¹⁷ Briefly, soil-grown plants were pulse-labeled with 5.1 mg of ^{N, supplied as KNO}₃, 3 d before repeated wounding and elicitation with 1:5 diluted M. sexta OS in glasshouse conditions. The oldest sink, youngest source and transition rosette leaves were wounded 1x per day for three consecutive days and on each day 10 µL of diluted OS was added immediately to the wound. Four days after the first wounding and OS elicitation (W+OS), the locally elicited sink leaf (hereafter referred to as young rosette leaf, yRL) was harvested and analyzed. Leaf total N was measured by isotope ratio mass spectrometry (IRMS), TSP was measured by the Bradford assay, nitrate was measured following ref.¹⁹ and absolute pools and ^{N incorporation into RuBisCO and nicotine and phenolamides were measured respectively by LC-MS and UPLC/UV/ToF-MS as described in ref.17}

Our data demonstrates that asRUB plants have similar whole leaf N accumulation and TSP pools as WT plants, with a slight decrease of whole leaf N and a sharp decrease of TSP after OS elicitation (Fig. 1A). This suggests that even though asRUB lines have decreased accumulation of the large subunit of RuBisCO (LSU, see Table 1), their TSP pools follow the same dynamics as WT plants after OS elicitation. Additionally, it suggests that N not invested in RuBisCO in asRUB is at least partially diverted to other soluble proteins. Interestingly, although asRUB are silenced for accumulation of LSU, they invest similar amounts of N into CP and significantly greater amounts into DCS after OS elicitation compared with WT (Fig. 1B, Table 1). Investment of ^{N follows the same general pattern as N investment, with significantly higher amounts of N allocated to DCS in OS-elicited as}RUB lines compared with WT (Fig. 1C). These findings are consistent with our previous conclusions that recently assimilated N is invested into phenolamide production after M. sexta OS elicitation and that, at least in the case of CP and DCS, the N for the synthesis of defense metabolites is not derived from RuBisCO degradation (see Table 1 for results of the statistical analysis and absolute pools of individually measured metabolites). It is also striking that though the total amounts of N and TSP are similar in both lines, asRUB leaves accumulate significantly more nitrate than WT leaves; after simulated herbivory the amounts of nitrate decrease to the same extent in both lines (Fig. 1). In view of the conclusions outlined above, it’s tempting to speculate that the higher investments into DCS in the transgenic line are derived from this recently acquired N-pool. By using recently assimilated N, the plants might be able to react more quickly to herbivore attack, by rapidly increasing defense biosynthesis and decreasing and remobilizing N invested in proteins, and together these responses could function to anticipate the readjustments required after the loss of resources.²⁰21

Figure 1. Silencing of ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) does not alter major foliar N pools but decreased N and ^{N investment into RuBisCO is accompanied by an increase in defense metabolite investment after OS-elicitation of sink leaves (yRL).} a) RuBisCO-silenced lines (asRUB) have similar total N pools and total soluble protein as wild-type (WT), but silencing causes increases in total nitrate pools. Asterisks indicate differences among treatments in a two-way ANOVA (**: P ≤ 0.01; ***: P ≤ 0.001). Letters represent significant differences found using the minimum adequate model (n = 5). asRUB plants show increased N-(b) and ^N-(c) partitioning to dicaffeoyl-spermidine (DCS). N/^{N investment in residual TSP (TSP - (SSU + LSU)), RuBisCO large (LSU) and small (SSU) subunit, nicotine, caffeoyl-putrescine (CP) and DCS in the yRL was calculated by multiplying the proportion of N/N in each compound with the concentration of the compound for each leaf. Concentrations of each analyte were measured as described in the text and total pools were calculated based on leaf mass. Plants were elicited as described in text and the yRL was harvested 4 d after the first wound and OS elicitation. Note that N was not measured in the residual TSP. c - control, OS - treatment with wound and oral secretions of} Manduca sexta, TSP - total soluble protein

The total amounts of nitrate measured in asRUB plants are much smaller than the amounts of DCS that accumulate after simulated herbivory (Fig. 1A, Table 1), but this may be due to the high turnover rate of the nitrate pool, and a single time-point measurement would not reflect a constant higher N-supply from this pool, which could be used for DCS biosynthesis. Our results are in contrast to earlier publications which suggest that RuBisCO is primarily a storage protein, providing N for defense metabolite production after herbivory¹⁶; and also to studies in N. tabacum silenced for RuBisCO expression, where RuBisCO-silenced plants had higher nitrate levels than WT but invested less into N-containing metabolites (nicotine), due to carbon limitation.²² In the present study, constitutive and OS-induced nicotine levels are slightly reduced in asRUB leaves, but the difference between the genotypes was not significant (Fig. 1B, Table 1). Similarly, though the nicotine levels were significantly increased after OS elicitation in both asRUB and WT leaves, the ^{N incorporation into nicotine was overall low and did not differ significantly between the transgenic line and WT (}Fig. 1C). Thus in N. attenuata we cannot find a direct growth-defense trade-off based on nicotine after simulated herbivory with OS from a specialist herbivore which is adapted to high levels of nicotine. It is noteworthy that these previous studies in cultivated tobacco silenced for RuBisCO have focused on the effects of carbon or N limitations on growth-defense trade-offs, but did not analyze the effects of herbivory on the allocation to growth and N-based defenses.²²23 Therefore, it is possible that in the absence of herbivore attack, resources are redirected to metabolic functions other than defense, to allow plants to better tolerate growth under carbon- or N-limited conditions. Another possible explanation is that herbivory-induced increases in plant N allocation to defense are stronger in N. attenuata compared with cultivated tobacco, since the former is a pioneer species, adapted to growth in a desert environment with high herbivore pressure by both nicotine-tolerant specialists and nicotine-sensitive generalist herbivores. In this context, we suggest that after elicitation by OS of the nicotine-tolerant M. sexta, N. attenuata invests N into the rapidly inducible phenolamides which have been shown to be effective against this herbivore,²⁴25 while maintaining high nicotine levels which are effective against generalist herbivores.

Based on these results we propose that the OS-elicited dramatic decline in soluble proteins in N. attenuata is mainly a tolerance mechanism, possibly a strategy to decrease the digestibility of leaf tissue to attacking herbivores or of facilitating the reallocation of N to other plant parts.⁸12 We speculate that protein-derived N is mainly reallocated to roots, thus enabling regrowth after herbivore attack (Fig. 2) as previously shown for carbon in wild tobacco and tomato: where newly assimilated carbon was transported to the roots after simulated leaf herbivory, probably to be used for post-herbivory re-growth, instead of being transported to the young leaves.²⁶27

Figure 2. Hypothetical model of nitrogen (N)-flow after herbivory. Total pools of ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) and total soluble proteins (TSP) decrease after herbivory. N originating from these proteins is not used for the biosynthesis of rapidly induced N-containing defensive phenolamides (caffeoyl-putrescine,CP and dicaffeoyl-spermidine, DCS), but probably redirected to the root; while phenolamides are synthesized from recently assimilated N, possibly nitrate.

Furthermore, our results suggest that the N-demand for defense metabolite production itself does not seem to be the direct signal leading to a major switch in metabolism after herbivory which results in a decline in growth functions and increased investment in defenses, since the transgenic line used here (asRUB) had similar or even higher investments in N-containing defenses as WT plants, despite decreased accumulation of the most abundant foliar protein in plants.

These findings are also consistent with the idea that the production of defense metabolites is adaptive: if plants have more free nutrients (in this case, nitrate) available they have a higher capacity for the rapid production of N-containing defense metabolites to protect their valuable photosynthetically active tissues. In the case of N. attenuata asRUB plants, these additional investments in defenses after OS elicitation are not more costly for the plant, since OS-elicited asRUB plants were not significantly different from WT in growth parameters measured (final stalk length, maturation of the first stem leaf, S1; Table 1). This is consistent with a previous study on this asRUB line which showed that despite decreased photosynthetic rate and lower levels of RuBisCO which affected initial growth rates, these plants eventually achieved the same stalk height as WT plants, possibly due to increased activation of RuBisCO which compensates for lower foliar pools of this protein.¹⁸28 It is unclear however, what effects the silencing of RuBisCO would have on plant growth after longer periods of herbivore damage and also in the case of intra-specific competition. It would be interesting to test whether the proposed tolerance mechanism above, based on N reallocation from proteins, is affected by the silencing of this important photosynthetic protein in the long-term in more natural settings.

Overall, our data suggest that there is no direct resource-based trade-off between N-demand for defense metabolite production and a decrease in protein content within the leaf, but that the regulation of trade-offs between defense metabolite production and plant development and tolerance is rather based on whole-plant signaling mechanisms which remain to be elucidated. The availability of free N (such as nitrate) might also play a role which needs to be further investigated.

Disclosure of Potential Conflicts of Interest

Acknowledgments

Glossary