•Here, we examined the impact of jasmonate (JA) treatment, branching and phloem girdling on 13C and 15N import, invertase activity and polyphenol accumulation in juvenile tissues of unbranched and branched hybrid poplar saplings (Populus nigra × P. deltoides).
•The import of 13C to juvenile tissues was positively correlated with invertase activity at the treatment site and enhanced by JA. Both invertase activity and 13C import were greater in shorter, younger branches and smaller, younger leaves. By contrast, JA treatments, branching and girdling had little or no impact on 15N import.
•In poplar saplings with multiple lateral branches, we observed almost no 13C movement from subtending source leaves into lateral branches above them, with or without JA treatment. The presence of potentially competing branches, treated with JA or not, girdled or not, had no impact on carbohydrate (CHO) import or polyphenol accumulation in target branches.
•We conclude that poplar branches comprise modules that are relatively independent from each other and from the stem below in terms of CHO movement, carbon-based defence production and response to elicitors. By contrast, branches are closely linked modules in terms of nitrogen movement. This should produce trees that are highly heterogeneous in quality for herbivores.
Nitrogen can also be involved in responses to wounding because it forms a part of defence metabolites or because of its requirement as part of enzymes necessary for metabolism or tissue repair. But the translocation of N per se from sources to sinks to support defence responses has not been well studied. Expression of wound-induced genes in young poplar leaves has been linked to N supply. Rennenberg et al. (2010) and Gómez et al. (2010) reported an increase in short-term nitrogen export from MeJA-elicited poplar leaves, from 12.5% to 22.7%. In other species, such as native tobacco, the nitrogenous defences themselves are transported via the vascular system in response to attack (Schwachtje et al., 2006).
In the study reported here, we used 13C and 15N labels simultaneously to determine whether nitrogen and CHO translocation from sources to wound sites is increased by treatment with the wound hormone JA, and the degree to which their movement is constrained by cell wall invertase activity, stem girdling and the presence of potentially competing sinks in branches. Our aim was to determine whether or not increased import of C and/or N resources is a component of the wound response in poplar foliage.
Materials and Methods
Hybrid poplar saplings (Populus nigra L.× P. deltoides W. Bartram ex Marshall; also known as Populus × canadensis Moench (pro sp.); clone OP-367, Segal Ranch, WA, USA) were grown from cuttings as described (Arnold & Schultz, 2002). Saplings were grown from cuttings in a glasshouse at Pennsylvania State University in 9-l containers of Metromix 250 soil-less potting media with supplemental lighting (400 W sodium vapour lamps, average daily irradiances of 800–1200 μmol m−2 s−1for 15 h d−1). Glasshouse experiments employed saplings > 30 cm tall with at least 29 leaves. Each leaf was assigned a leaf plastochron index (LPI) number, beginning with the most recently emerged ‘index’ leaf (≥ 2 cm), which is designated LPI0 (Larson & Isebrands, 1971). This method allows leaves to be classified as rapidly expanding sink leaves (LPI0–LPI4), leaves making the sink-to-source transition (LPI5–LPI7) or fully expanded source leaves (≥ LPI8). In P. deltoides, the primary vascular connections (orthostichy) are five LPI units apart such that growth of LPI3 is largely supplied by LPI8 and of LPI2 by LPI7, etc. (Larson, 1977) and with dye labelling studies we confirmed this to be also true for the hybrid. Under our growth conditions, LPI3 leaves were rapidly expanding sink leaves throughout the experiment.
JA was used as a surrogate for wounding and herbivory. It is very difficult to generate identical amounts of insect damage in feeding studies because of inherent variation in consumption rates, confounding duration with degree of damage and requiring unrealistically large sample sizes (n = 50–100). This can be resolved by using the wound hormone JA. While we realize that JA may not fully mimic responses to individual herbivores, work by our lab and others has shown broad overlap in poplar responses to the effects of JA and herbivory (Arnold & Schultz, 2002; Arnold et al., 2004; Major & Constabel, 2007). Its use in this study allows us to identify general patterns that can be explored later with specific herbivores. JA (Sigma J-2500, St. Louis, MO, USA) was solubilized in 10% EtOH(aq) with 0.125% (w/v) Triton X-100 (Arnold & Schultz, 2002; Arnold et al., 2004). This JA concentration ensured that an induced response could be detected clearly before targeted sink leaves matured into source leaves. There were no visible effects of JA application. Individual leaves or whole branches received a single application consisting of spraying an individual leaf or whole branch until runoff was observed on the leaf surface; runoff was collected to make sure that other leaves were untreated.
In order to quantify the long-distance flow of carbon resources in poplar saplings, leaves were exposed to 13CO2-rich air for 90 min in 500 ml chambers containing c. 370 ppm 13CO2 within each chamber, which was generated by acidifying NaH13CO2. The import of 13C was determined by isotope ratio mass spectrometry after 24 h as previously described (Arnold & Schultz, 2002; Arnold et al., 2004). This timeframe was chosen because it had been shown to be a peak period of 13C incorporation in aspen leaves (Kleiner et al.,1999) and we found it responsive to herbivory in the hybrid we used (Arnold & Schultz, 2002; Arnold et al., 2004). We previously confirmed that this incubation period generated a 13C pulse of sufficient magnitude to be traced throughout the samplings while minimizing potential ‘chamber’ effects on incubated LPI8 leaves. The absence of 13CO2 leaks was confirmed by examining the δ13C signature of unlabelled plants that were interspersed among the experimental saplings. To quantify the flow of nitrogen, each poplar sapling was fertilized with 15 mg Na15NO3 in a 50 ml solution by adding to the soil. The Na15NO3 solution was ‘watered-in’, saturating the 5 l of soil-less potting media and evenly distributing the stable-isotope label. This provided a sufficient 15N pulse while elevating soil Na+ concentration < 5 mM, significantly less than that required to trigger osmotic stress responses observed to affect N metabolism in poplar (see Ehlting et al., 2007). All isotope ratio mass spectrometry was carried out at University of California-Davis Stable Isotope Laboratory. The isotopic enrichment of harvested tissues was reported in the conventional δ notation in parts per thousand (‰), where δ = (Rsample/Rstandard− 1) × 1000, for example, 13C:12C and 15N:14N, using the international standards Pee Dee Belemnite for C and atmospheric nitrogen for N.
In selected saplings, long-distance flow by phloem of 13C-carbohydrates was disrupted by steam girdling. Girdling was done with a clothes steamer (Proctor-Silex, Washington, NC, USA) with a modified 8 mm nozzle. Three-second steam bursts aimed at stems from four directions caused them to blacken within 24 h. The effectiveness of steam girdling was confirmed by 13C-tracer experiments as described previously (Arnold et al., 2004). Little or no damage to other tissues appeared to occur because above the girdle site the leaves did not wilt, the stems and petioles maintained physical strength, and the tissues remained healthy well beyond the experiment’s duration.
The activity of cell wall invertase (CWI) in select leaves was quantified colorimetrically and expressed as μmol sucrose cleaved per gram of tissue (wet mass) per minute, using the methods of Arnold & Schultz (2002).
Polyphenols were isolated from poplar leaves using the method described in Appel et al. (2001). We focused on condensed tannins (proanthocyanins) as wound-induced defences because their biosynthesis has been well characterized and is known to be elicited by wounding and JA (Constabel & Major, 2005; Philippe & Bohlmann, 2007), and because we had established previously that induced tannin production co-occurs with elevated sink strength in hybrid poplar foliage (Arnold & Schultz, 2002). Condensed tannin concentrations were measured as extracted proanthocyanins by the acid butanol method (Hagerman & Butler, 1989). The concentration of total reactive polyphenols present in leaf extract was determined using the Folin-Denis assay (Appel et al., 2001), which predominately measures phenolic glycoside concentrations in this poplar clone (T. Schaeffer, pers. comm.). In both cases, standard curves were developed by using standards purified from hybrid poplar (clone OP-367) foliage (Arnold & Schultz, 2002; Arnold et al., 2004).
Carbon and nitrogen transport to sink leaves in unbranched poplar saplings
We quantified the flow of 15N and 13C isotopes and invertase activity in poplar saplings in intact plants and compared that to plants in which phloem transport was disrupted by stem girdling. Petioles of LPI3 and LPI8 leaves of each plant were banded for identification. Banded sink leaves developed rapidly during the course of the experiment such that by day 5 they were LPI3, 4, or 5 depending on the treatment they received. Experiments were concluded before banded sink leaves became mature. Individual plants were randomly assigned to control or treatment groups.
There were 24–28 replicate trees per treatment. LPI3 leaf length was measured at the start and end of the experiment. On the morning of day 1, we girdled stems below the LPI3 sink leaf and applied a soil drench of 15N to roots. In the late afternoon of the same day, LPI8 source leaves were treated with 13C. Twenty-four hours after 13C application, LPI3 and LPI8 leaves were harvested from half the replicates and used for isotope and invertase analysis. Forty-eight hours after 13C application, LPI3 and LPI8 leaves were harvested from the remaining half of the replicates and used for condensed tannin analysis. Because the isotope data were not normally distributed, the effects of girdling, invertase activity and LPI3 starting length on the import of 15N and 13C and concentrations of condensed tannins were evaluated using the nonparametric Wilcoxon Two-Sample Test with P-values reported for one-sided Z tests (PROC NPAR1WAY; SAS Institute, Cary, NC, USA). The linear relationships among traits were characterized by Pearson correlation coefficients using (PROC CORREL; SAS Institute).
Carbon and nitrogen transport to sink leaves in branched poplar saplings
In order to mimic the effect of apical meristem damage that occurs naturally in the field from herbivory or other sources of mechanical damage, sylleptic branches were created on saplings by removing the apical meristem just below the LPI8 leaf two weeks before the start of experiments, producing plants with three to four lateral branches and no terminal axis. Two similar-sized lateral branches arising from the middle section of the saplings on opposing sides of the stem axis were designated as target and nontarget branches. The target branch was sampled for isotope, invertase and defence chemistry and received 5 mM JA or a control solution over the entire length of the branch. The nontarget branch was manipulated to potentially alter its C demand by application of JA or girdling to cause interruption of phloem flow to/from the nontarget branch. The length of each branch was recorded at the start and end of the experiment. There were four treatment groups: unbranched saplings; branched saplings with unaltered nontarget branches; branched saplings with nontarget branches girdled at the base; and branched saplings with an entire nontarget branch treated with JA. Each treatment group comprised 15–16 plants. On the morning of day 1, branches were girdled, roots received the 15N drench, and branches received JA or control solution sprays. In the late afternoon of the same day, 13CO2 was fed to a group of 5–10 source leaves located immediately below the lowest branching point on the main stem. On day 2, target branches were harvested from half of the plants in each treatment group and the petioles and stems were used for isotope analysis. On day 3, target branches on the remaining plants were sampled for condensed tannin and total polyphenol analysis. Because the isotope data were not normally distributed, the effects of secondary branch treatment and the start and end lengths of the primary branch on import of 15N and 13C was evaluated using the nonparametric Wilcoxon Two-Sample Test with P-values reported for one-sided Z tests (PROC NPAR1WAY; SAS Institute). Linear relationships among traits were characterized by Pearson correlation coefficients using PROC CORREL (SAS Institute).
Girdling alters 13C import but not 15N import to treatment sites
We quantified the flow of 15N and 13C toward LPI3, a known sink leaf on unbranched glasshouse-grown poplar saplings, with and without phloem girdling just below LPI3 (Fig. 1a). Import of 13C from the LPI8 source leaf was reduced in girdled plants to just above the background isotope levels of –30 (Fig. 1b; P < 0.0001 for comparison with controls; n = 24). By contrast, 15N import by the LPI3 sink leaf from roots was affected only slightly by girdling (Fig. 1b; P < 0.0421; n = 25). In ungirdled plants, 13C import of LPI3 sink leaves was positively related to invertase activity (Fig. 1c; R = 0.46, P = 0.0470; n = 17) but stem girdling eliminated the relationship. Girdling caused an increase in condensed tannins in the LPI3 sink leaf (Fig. 1d; P < 0.0019; n = 14).
Younger, smaller leaves were significantly stronger 13C importers than were larger, more mature leaves (Fig. 2). Import of 13C to the LPI3 leaf was strongly and inversely related to its starting size, with smaller leaves importing more (Fig. 2; R = − 0.67, P = 0.0003; n = 44).
Branching alters 13C import but not 15N import to treatment sites
We compared 13C and 15N movement from source leaves to shoot apices in branched and unbranched plants (Fig. 3). At the time of the experiments, the lateral branches had several fully expanded leaves and a cluster of three to five expanding terminal leaves. Treating terminals with JA increased the flow of 13C from source leaves to shoot apices of unbranched plants, in this case the stem and petioles (Fig. 3a; P < 0.0465; n = 14). In the branched saplings, however, movement of 13C from labelled source leaves below the branch point into the target branch was minimal, irrespective of JA treatment (Fig. 3a). While statistically different from unlabelled baseline, the δ13C values of target lateral branches increased only an average of 4.5% (as compared with 350% for the targets in unbranched trees); only a tiny fraction of the CHOs present in source leaves was shared with lateral branches. It was not surprising then to find that girdling phloem at the base of the nontarget branch point had no significant effect on 13C movement into branches (Fig. 3a). In short, we found that although branches imported statistically significantly more 13C from source leaves below branches compared to background unlabelled plants, the actual amounts were miniscule. By contrast, 15N was transported from roots to lateral branches just as it was in unbranched plants, in amounts unaffected by JA exposure (Fig. 3b). Girdling also had no effect on 15N import, an unsurprising result given that most N transport occurs in xylem.
Despite the minor import of 13C from lower main stem sources, leaves on lateral branches had higher concentrations of condensed tannins (35.8 mg vs 11.1 mg g−1; P < 0.003; n = 26) and Folin-Denis reactive total polyphenols (0.104 vs 0.072 mg g−1; P < 0.003; n = 26) than did the leaves of the same age from unbranched plants. This accumulation resulted in dramatic reddening of young leaves of lateral branches, which was also observed in leaves located above stem girdling on unbranched saplings in this study and in a previous one (Arnold et al., 2004). These unusually high concentrations of polyphenols were unaffected by direct JA treatment, the presence of potentially competing branches, or girdling treatments designed to disrupt CHO flow between branches (data not shown).
We were interested in whether branch length, like leaf length, influenced what little import occurred. Because there were no statistically significant differences in 13C and 15N import among the branched-tree treatments, we combined data from all treatments and found that shorter branches imported more 13C than longer branches (Fig. 4a; R = − 0.471, P = 0.0006; n = 49). By contrast, 15N import was not correlated with branch length (Fig. 4b).
There was also a strong inverse relationship between two measures of leaf polyphenols and the initial length of the primary branch (Fig. 4c,d). Leaves on shorter branches had higher concentrations than did longer branches of both condensed tannins (Fig. 4c; R = − 0.618, P < 0.0001; n = 59) and total phenolics (Fig. 4c; R = − 0.584, P < 0.0001; n = 59). Polyphenol accumulation in leaves of the primary branch was also inversely related to branch length at the end of the experiment (data not shown).
In this study, we examined the import of carbon and nitrogen by sink tissues on hybrid poplar saplings. By labelling roots with 15N and source leaves on unbranched plants that were girdled or not, we evaluated the role of phloem in nitrogen transport from roots. By labelling roots with 15N and source leaves below sylleptic branches with 13C, we examined the response of branches when the sink strength of potential competitor branches was enhanced by JA treatment or reduced by girdling. We asked whether target tissue traits such as size influenced JA-elicited CHO movement, and determined whether JA treatment enhances nitrogen movement (as 15N) from roots to sink leaves, as we have previously shown for the movement of CHOs from source leaves to sink leaves. Finally, we assessed the impacts of differential CHO transport on accumulation of polyphenols in youthful tissues.
Plant architecture influences long-distance CHO movement
Sink tissues such as juvenile leaves, buds and reproductive structures, draw from a common pool of CHOs. Resource availability ultimately depends upon each tissue’s relative sink strength, proximity to sources and other sinks, and vascular connectivity. Therefore, we expected that we might find competition among branches. We discovered that source–sink relationships and factors influencing production of leaf phenolics are not that simple.
Only miniscule amounts of 13C labelled CHOs moved from trunk source leaves into branches, irrespective of the treatments that the branches received. This means that branches were largely independent of subtending leaves and acquired C from other sources (e.g. stems) or from their own mature source leaves, at least during the growth period we studied. As a result, phenolic accumulation in branch leaves and CHO import were not influenced when we isolated the elicited or neighbouring branch from the trunk by girdling. We also found that blocking transport to/from neighbouring stems had no impact on import, indicating little or no connectivity among branches. These results are consistent with a view of plants as collections of relatively independent modules (Haukioja, 1991). Scarascia-Mugnozza et al. (1999) also found little export of assimilates between branches and main stem leaves in a survey of four poplar species and hybrids, and Fisher et al. (1983) saw little transfer of label to axillary branches from subtending leaves in P. deltoides.
Branches and their leaves may have drawn on CHOs stored in the stem closest to leaves as they began to grow, and then from their own source leaves. We did not investigate the role of stem storage in transport or elicited responses, and did not begin labelling until the sylleptic branches were at least 7 cm long. However, supplying CHOs to phenolic synthesis from local stem storage (as opposed to more distant sources) would explain the evident disconnect between the main stem and leaves on branches in our study. Nearby stem resources have been shown to support spring leaf flush in full-grown trees of another Populus species (Landhausser & Lieffers, 2003). Poplar branches, at least as generated by removing apical dominance, are relatively independent modules with respect to CHO supply and do not appear to compete for CHOs.
Carbohydrate import and polyphenol accumulation depend on branch and leaf size and age
The import of 13C by leaves on unbranched trees was inversely related to leaf size (Fig. 2) and to branch length on branched trees, even though the latter import was small (Fig. 4). This pattern is consistent with the view that smaller, younger leaves and branches are stronger sinks and that CHO import from distant sources is influenced by sink tissue age and sink strength. By the time we sampled branches, they may have developed independence from distant sources that were important when they began to grow.
We observed a striking and visible build-up of constitutive phenolics in branch leaves, independent of elicitation treatment (Fig. 4). Although unrelated to growth rate during the experiment, these polyphenol concentrations were inversely related to starting branch length. Elevated polyphenol concentrations, particularly anthocyanins, are frequently associated with regrowth following coppicing or browsing. This phenomenon, in which regenerating tissues exhibit juvenile traits including high polyphenol concentrations, has been called ‘juvenilization’ (Bryant et al., 1983; Lindroth et al., 2007). While we did not record the ages of the branches (i.e. the date on which they began to grow), shorter branches behaved as though they were younger (more ‘juvenile’) than long branches and accumulated more polyphenols, as is typical of juvenile leaves and stems. The fact that branch growth rates did not differ among branches (so relatively short branches at the first measurement were also short at the last) lends credence to the view that shorter branches differed from longer branches in age, not growth. Functions of these constitutive accumulations remain speculative, but protection from UV light and enhancing defence have been proposed. Polyphenol production in young leaves, which is transcriptionally regulated by sugars (Rolland et al., 2002), may reflect high sink strength and CHO import.
Nitrogen transport is not increased in response to JA
All plant defences against herbivores as well as many responses to stresses require N either as part of the defence (e.g. enzymes with direct defensive roles, including proteinase inhibitors and chitinases, or metabolites such as alkaloids) or as biosynthetic enzymes. Although poplar does not produce N-intensive secondary metabolites, it does produce proteinase inhibitors, chitinases and vegetable storage proteins in response to wounding (Constabel & Major, 2005). Therefore, we hypothesized that responses to JA would increase transport of N to elicitation sites to support production of defensive proteins or construction of enzymatic machinery for production of C-based defences. Developing leaves are strong sinks for N, especially in the spring. In trees, N reserves stored overwinter in the bark and wood support up to 90% of new leaf growth (Wetzel et al., 1989; Sagisaka, 1993; Stepien et al., 1994; Millard, 1996; Frak et al., 2002). There is circumstantial evidence that N transport and partitioning can be altered by JA treatments lasting 1 wk or more (e.g. Beardmore et al., 2000; Rossato et al., 2002; Meuriot et al., 2003). JA clearly influences transport of defensive nitrogenous defences in plants producing them, such as members of the Solanaceae (Schwachtje et al., 2006).
However, we found no evidence for increased N flow from roots toward JA-treated tissues in poplar saplings over a 24 h period (Fig. 3). The amount of 15N transported from roots to leaves or branches was only slightly disrupted by girdling, indicating that 15N transport is largely restricted to xylem. The slight decrease in 15N import we observed with girdling may reflect reduced xylem to phloem transport above the girdle. This is consistent with results from another study of a different hybrid poplar, in which xylem N above the girdle was not higher after girdling, although N concentration was eventually higher below the girdle in xylem, a tissue that we did not examine (Cooke et al., 2003). We cannot rule out the possibility that N was mobilized from other sources (e.g. stems; Wetzel et al., 1989; Fan et al., 2009) because that N would not be labelled and hence would not be detected using our methods. It is also possible that the amount of 15N transported from roots to leaves or branches was increased or decreased by our treatments at earlier or later time points, or differed seasonally. However, our results indicate that it is unlikely that treatment-enhanced sink strength supplies more N for the production of defences or other JA-elicited responses in young leaves of this hybrid poplar in the short term. Davis et al. (1991) showed that assimilate movement dictates remote sites of wound-induced gene expression in poplar leaves. Our results indicate that this effect does not include translocating N itself and is probably a signalling phenomenon, not movement of resources for defence.
Several studies suggest that attacked or elicited leaves export CHOs to roots as a way of sequestering valuable resources from consumption by herbivores (Babst et al., 2005, 2008; Schwachtje et al., 2006; Newingham et al., 2007; Kaplan et al., 2008; Gómez et al., 2010). These studies used short-lived radioisotopes (11C, 13N) to study changes in resource flow occurring within minutes following elicitation. Our study, by contrast, found CHO movement toward elicited sites over 24–48 h. This apparent contradiction may have several explanations.
The 11C studies typically have not stated the sink/source status of the leaves being studied. Because leaves may undergo a rapid transition from sink to source as they age, we have been careful to work with leaves whose source/sink status is clearly established. In poplar, the sink–source transition occurs within 24 and 48 h; they change very rapidly from being importers to exporters (Larson et al., 1980; Dickson & Larson, 1981). On any plant, exporters normally export a good deal of their resources to the roots. Finding export from a leaf of unknown sink/source status to the roots in response to elicitation is not unexpected.
All leaves may export to the roots as soon as they are able, depending on demands of growth points or other aboveground sinks. Some of that material is eventually recycled aboveground. In fact, this is precisely what occurs in poplars (Ghirardo et al., 2011). The materials we see translocated to attack sites after 24 h may be recycled from belowground sources. Future studies should follow C acquired in leaves to both places – below- and aboveground – over a time course, to resolve this issue.
We conclude that sink strength in poplar saplings based on leaf age facilitates the reallocation of CHOs (but not N) among leaves within individual stems (branches) but not (or minimally) among them. This movement can be enhanced by treating sinks with JA, but does not spur CHO movement across the stem–branch boundary or between branches. A limit to CHO reallocation among branches in response to damage makes them relatively independent modules, which has important implications for the spatial patterning of plant defence and herbivore foraging. For example, this would make induced defences in trees largely a local phenomenon, slowing systemic defence development, creating a spatially heterogeneous food resource, and allowing herbivores to minimize exposure to defensive chemistry by moving to new branches (Schultz, 1983). And we have provided further evidence that heterogeneity in distribution of C-based responses to elicitation by JA or wounding (Tuomi et al., 1984; Tuomi et al., 1988; Haukioja, 1991; Honkanen et al., 1994) is dependent on factors influencing sink/source interactions.
We thank Ellen Bingham, Karen Bingham, Elizabeth Bozak, Clayton Coffman, Jeff Heath, Kathyrn Pickering, Matt Pye, Lucy Rubino and Colleen Yunker for help with experiments. We also thank Melody Kroll, Abbie Ferrieri, Dean Bergstrom and Clayton Coffman who made helpful comments on the manuscript and Karla Carter who assisted with references and formatting. This research was supported by National Science Foundation grants IOS-0614893 (T.M.A.) and IOS-0614890 (J.C.S. and H.M.A.).