Abstract
Phenolic compounds often accumulate in foliar tissues of deciduous woody plants in response to previous insect defoliation, but similar responses have been observed infrequently in evergreen conifers. We studied the effects of defoliation on the foliar chemistry of Scots pine (Pinus sylvestris L.) and cocoon mass, and survival of the pine sawfly (Diprion pini L.). In two successive years, needles were excised early in the season leaving only the current-year shoot intact (defoliated trees); untreated entire shoots served as controls (control trees). A year after the second defoliation, pine sawfly larvae were transferred to the trees. Delayed induced resistance in Scots pine in response to defoliation was indicated by (1) reduced cocoon mass in defoliated trees and (2) increased concentrations of phenolics and soluble condensed tannins in the foliage of defoliated trees compared with controls. Myricetin-3-galactoside, which showed the strongest induced response (104% and 71% increase in current-year (C) and previous-year (C+1) needles) of the compounds analyzed, also entered the regression model explaining variation in sawfly performance. Other compounds that entered the model, e.g., (+)-catechin, showed weaker responses to defoliation than myricetin-3-galactoside. Hyperin, condensed tannins and quercitrin showed strong induced responses in C or C+1 needles, or both, but these compounds did not explain the variation in sawfly performance. Accumulation of phenolics is sometimes associated with the reduced foliage nitrogen (N) concentrations in deciduous trees, and our results suggest that this may also be the case in evergreen conifers. Based on the earlier findings that defoliation reduces needle N concentration and N deficiency results in the accumulation of the same phenolic compounds, i.e., myricetin and quercetin glycosides, and soluble condensed tannins, we suggest that the accumulation of phenolics in defoliated trees occurred in response to the reduced foliar N concentration.
Introduction
Herbivory and other foliar damage to woody plants can induce the production and the accumulation of defensive chemicals in the same growing season (rapidly induced resistance) or during the following growing seasons (delayed induced resistance). Induced defenses against herbivores and pathogens may have several advantages compared with constitutive defenses (Agrawal and Karban 1999), and may have evolved in response to variable risk of herbivore and pathogen attack (Haukioja 1980, Haukioja and Neuvonen 1985, Karban and Baldwin 1997). The general response of plants to both wounding and herbivore damage involves the induction of phenylpropanoid metabolism and the accumulation of phenolic compounds (Bernards and Båstrup-Spohr 2008). However, plant responses induced by artificial mechanical defoliation may not always correspond to responses elicited by insect damage (Haukioja and Neuvonen 1985, Karban and Baldwin 1997, Reymond et al. 2000).
Delayed inducible resistance (DIR) occurs in many long-lived perennial species that occasionally experience heavy defoliation, e.g., during herbivore outbreaks (Haukioja 1980). However, some evergreen conifers, such as Scots pine (Pinus sylvestris L.), appear to have no effective DIR, as indicated by the damage caused by Diprionid sawflies that occasionally attack mature pine forests (Niemelä et al. 1984, 1991). Although, under certain conditions, Pinus species may exhibit damage-induced responses that weakly counteract attacks by sawflies (Lyytikäinen 1992, Raffa et al. 1998), in most cases no effective rapid or delayed induced resistances against sawflies have been found (Niemelä and Tuomi 1993, Lyytikäinen 1994, Krause and Raffa 1995, Björkman et al. 1997, McMillin and Wagner 1997, Lyytikäinen-Saarenmaa 1999). According to Niemelä et al. (1984, 1991), DIR may be more pronounced in deciduous trees than in evergreen trees. Results of a recent meta-analysis support this suggestion (Nykänen and Koricheva 2004).
We tested whether the defoliation of Scots pine trees in the two previous growing seasons could result in the delayed induced resistance against the pine sawfly (Diprion pini L.). In deciduous trees, delayed induced accumulation of phenolics has sometimes been associated with the reduced foliar nitrogen (N) concentration, e.g., in defoliated birch trees (Tuomi et al. 1984) as well as in individual branches (Tuomi et al. 1988b). Previously, we found a decline in N and polyamine putrescine concentrations in the undamaged foliage of defoliated Scots pine trees compared with control trees, whereas foliar concentrations of sucrose, starch and some phenolic compounds increased (Roitto et al. 2003). In some studies, the larval performance of D. pini has negatively correlated with the concentrations of certain phenolic compounds in needles (see review by Mumm and Hilker 2006). Specifically, we compared the performance of D. pini on the previously defoliated and undefoliated trees studied by Roitto et al. (2003) to determine if the induced changes in needle phenolic composition explain sawfly performance.
Materials and methods
Defoliation
Scots pine (P. sylvestris) trees growing in dry, nutrient-poor sandy soil in northern Finland on the island of Hailuoto (65°03′ N and 24°36′ E) were artificially defoliated in two successive years, June 22–26, 1998 and June 21–26, 1999. Thirty trees aged 8–25 years were randomly selected from a pool of trees ~ 2 m in height and at least 3 m apart. Selected trees were randomly divided into two groups (control and defoliation) that showed no differences in age distribution or height. Fifteen trees were defoliated by excising all the 1-year-old and older needles at the time when the current shoot (apical meristem) was fully elongated, but when current-year needle growth was just starting. Another group of fifteen trees was left intact (controls).
Ground vegetation and physical and chemical characteristics of the soil indicated that the site was nutrient-poor. The ground vegetation consisted of patches of dwarf shrubs, lichens and mosses. The pH of the mineral soil (sand) was 4.8 and the total N concentration was below 0.01% of soil dry matter. The soil organic layer was patchy and thin (0.5–1.0 cm) (Kuikka et al. 2003).
Feeding experiment
In 2000, a feeding experiment was performed with the larvae of pine sawfly (D. pini). The sawflies originated from natural populations from Lestijärvi (63°32′ N and 24° 37′ E), Ilomantsi (62°40′ N and 30°56′ E) and Kiihtelysvaara (62°28′ N and 30°13′ E) in central Finland. In May 2000, cocoons were collected from each location. A number of females were caught from each population and placed in mesh bags with branches from local undamaged Scots pine trees. The mesh bags prevented copulation, and hence the females produced only male progeny. The progeny of four females from each population was randomly selected for this trial. Of the 15 previously defoliated trees, 12 were randomly selected for the rearing experiment together with 15 control trees. The progeny of each female (early instar larvae) was evenly divided between a control and a defoliated tree, resulting in a total of four trees for each sawfly origin in both treatments.
On July 20, when the current-year needles were about the same length as the 1-year-old needles, one lateral shoot bearing current-year (C) and previous-year (C+1) needles was chosen from the mid-crown of each defoliated tree and a control tree and enclosed in a mesh bag with 20 larvae. The larvae were transferred to fresh shoots when necessary. Most of the larvae pupated by the end of August. The shoots that still had few larvae were brought to the laboratory on August 30. The shoots were supplied with water to encourage the development of unpupated larvae. The cocoons were collected and stored in paper bags at 19 °C until all the larvae had pupated (September 12). The cocoons were maintained on Sphagnum moss at 19 °C and at a relative humidity of 70–80% until mid-November when they were placed in plastic vials at 5 °C and at a relative humidity of 70–80% to simulate conditions under snow cover. In mid-March, the cocoons were transferred to paper bags and placed outdoors under snow cover until the snow thawed naturally on April 19. They were then weighed and placed in plastic vials containing fresh Sphagnum until eclosion began on May 22. The mass and number of eclosed sawflies were recorded.
Needle sampling
Needles for phenolic analyses were collected on August 10, 2000. Four needle samples (each about 1 g) were picked from each tree. The needle samples included the current-year and the previous-year needles from lateral shoots in the mid-crown of each tree, and the corresponding partially chewed needles were sampled from the mesh bags. Although our experiment was not designed to test for the rapidly induced resistance (i.e., we did not control for the shading effect of the mesh bags), we analyzed the partly chewed needles to determine if damage by natural herbivory altered the needle chemistry of the previously defoliated trees compared with that of the intact trees. To obtain freshly chewed needles, the mesh bags were opened so that the needles being eaten at the time could be collected. The mesh bags with larvae were then moved to a new shoot to avoid the effect of needle sampling on the feeding of the larvae. Intact needles were collected for N analyses on August 25.
Chemical analyses
Phenolics were extracted and analyzed as described by Roitto et al. (2003). Briefly, 5 mg of freeze-dried and finely ground needles was extracted with 400 μl of methanol in an Ultra-Turrax homogenizer. Samples were centrifuged for 3 min at 10,000g, and the extraction was repeated three times, after which the extracts were combined, and the methanol was evaporated under N2. The dry extracts were dissolved in 400 μl of water:methanol (1:1 v/v) and filtered with a syringe (0.2 μm) in preparation for the high performance liquid chromatography (HPLC) analysis. A Hewlett–Packard HPLC system (HP 1050 Series, Avodale, PA) equipped with a photodiode array detector (HP 1040A Series) controlled with a computerized HP data system was used. Individual phenolic compounds were separated with a 3 μm HP Hypersil ODS II column (60 × 4.6 mm internal diameter). The samples were then eluted with solvent A (1.5% tetrahydrofuran + 0.25% orthophosphoric acid in H2O) and solvent B (100% methanol) with the following gradient: 0–5 min 0% of B in A, 5–10 min 0–20% of B in A, 10–20 min 20–40% of B in A, 20–40 min 30–50% of B in A. The flow rate was 2 ml min−1, the injection volume was 20 μl, and the temperature of the oven was 30 °C. Identification and quantification of compounds were based on their retention times and spectral characteristics compared with those of commercial standards. The glycosides and diacylated derivatives of myricetin, quercetin and kaempferol were quantified against myricetin-3-rhamnoside (myricitrin), quercetin-3-galactoside (hyperin) (Roth, Karlsruhe, Germany) and kaempferol-3-rhamnoside (Apin Chemicals, Abingdon, Oxon, UK), respectively. A standard of (+)-catechin (Aldrich, Steinheim, Germany) was used to quantify catechins. Confirmation of the neolignan compounds was based on their retention times and spectral data following HPLC/MS analysis. Hereafter, total HPLC phenolics refer to the sum of all phenolic compounds analyzed by HPLC. Condensed tannins from the methanol extracts were determined by an acid butanol assay (Porter et al. 1986) and were quantified against purified condensed tannins of Betula nana L. Needle N concentrations were analyzed by a dynamic flash combustion technique (EA 1110 Elemental Analyzers, CE Instruments, Milan, Italy).
Statistical analyses
As sawfly origin did not affect larval growth (P > 0.05), the data from the three populations were pooled in subsequent analyses. The effect of treatment on cocoon mass was tested by the analysis of variance (ANOVA). As progeny of each female was reared on one control tree and one defoliated tree, we considered mother as a blocking factor. As measurements on needle age classes within a tree were not independent from each other, statistical differences in the needle N concentrations were analyzed by repeated measures ANOVA, with needle age as a within-subjects factor, treatment as a between-subjects factor and a tree as a subject. As we measured multiple secondary compounds from the same trees, we performed a protected ANOVA on the data. This approach combines multivariate analysis (MANOVA) and ANOVA (univariate analyses are performed if multivariate analysis yields a significant result) and is less conservative than the Bonferroni correction used to correct P values from the multiple tests of a single hypothesis (Scheiner 1993). As MANOVA (Pillai’s trace) indicated a significant defoliation × needle age interaction (P < 0.05), we proceeded to ANOVA. Statistical differences in individual secondary compounds were tested by repeated measures ANOVA with needle age (two levels) and sawfly damage (damaged versus intact needles from the same trees) as within-subjects factors and treatment (defoliation versus control) as a between-subjects factor.
The relationships between cocoon viability (mass after winter), N concentration and concentrations of phenolics were assessed by the Pearson correlation coefficients. Correlations were computed separately for current-year (C) and previous-year (C+1) needles of control and defoliated trees. A pooled correlation coefficient (r) was also computed and is presented as a common r for C plus C+1 needles in control and defoliated trees. As concentrations of N and phenolics had such diverse relationships with cocoon mass, we also studied the effects of phenolics and N on cocoon viability by multiple regression analysis. As we had no a priori expectations about the importance of the individual phenolics and N on cocoon viability in combination with each other, we performed stepwise regression analysis where the compounds (phenolics or N) were selected solely on the basis of statistical criteria: a compound is included in the model if the F probability is small (P < 0.05) and compounds already in the regression model are removed if their probability of F becomes large (P > 0.1) as new compounds are entered in the model. In the analysis, current-year and previous-year needles were included in the same model, i.e., the software selected the best possible equation among the compounds measured in current-year and previous-year needles. We performed this analysis separately for the intact needles of control and defoliated trees. Statistical analyses were performed with SPSS Version 15 (SPSS, Chicago, IL), except for the pooled correlation coefficients which were computed with MetaWin Version 2.1 software.
Results
Sawfly performance
Cocoon masses of larvae (mean ± SE) grown on defoliated trees were 9% lower than those grown on control trees (53.5 ± 1.4 versus 58.8 ± 1.4 mg; F1,11 = 16.62, P = 0.002). There was no significant difference in developmental time between treatments. On September 1, the proportion of pupated larvae was 95 ± 2.8% in the control trees and 87.5 ± 6.7% in the defoliated trees (F1,11 = 2.05, P = 0.180). The mass of eclosed D. pini on the repeatedly defoliated trees was significantly reduced compared with the control values (8.1 ± 0.3 versus 9.3 ± 0.3 mg; F1,11 = 15.9, P = 0.002). The proportion of live-eclosing sawflies was 84 ± 5% on control trees and 80 ± 4% on defoliated trees (F1,11 = 0.44, P = 0.52).
Needle N and phenolics
Defoliation in two preceding summers tended to decrease foliar N concentration in current-year needles (control: 9.6 ± 0.3, defoliated: 8.5 ± 0.4 mg N g−1) and in previous-year needles (control: 8.3 ± 0.2 mg N g−1, defoliated: 7.7 ± 0.4 mg N g−1) (ANOVA, defoliation: F1,22 = 3.61, P = 0.071, needle age: F1,22 = 19.6, P < 0.001, defoliation × needle age interaction: F1,22 = 0.8, P = 0.381). In general, the concentrations of N and phenolics were positively correlated in control trees and negatively correlated in defoliated trees (Table 2, last row).
Needle concentrations of total HPLC phenolics and soluble condensed tannins were higher in defoliated trees than in control trees (Figure 1, Table 1). Concentrations of several phenolics increased in response to defoliation (Figure 2C–F and I, Table 4). Many of these compounds showed a significant needle age × defoliation interaction (Table 1, Figure 2). Unlike most compounds analyzed, the concentration of DC-astragalin was higher in current-year needles than in previous-year needles (Figure 2H).
(A) Soluble condensed tannins and (B) HPLC phenolics in control and defoliated P. sylvestris trees (mean concentration ± SE). The concentrations are given for current-year (C) and previous-year needles (C+1), and for intact needles and those being damaged by larvae.
Concentrations (mean ± SE) of phenolic compounds in the previously defoliated and control P. sylvestris trees. The concentrations are given for current-year (C) and previous-year needles (C+1), and for intact needles and those being damaged by larvae.
Results of ANOVA for the effects of (1) defoliation (intact needles in control versus defoliated trees), (2) needle age (current-year versus previous-year needles), and (3) damage by sawfly larvae (intact versus partly eaten needles) on the concentrations of phenolic compounds. Significant differences are indicated by asterisks as *, P < 0.05; **, P < 0.01 and ***, P < 0.001. Abbreviations: MC, monocoumaryl; DC, dicoumaryl; neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran; neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran and Def., defoliation.
| Source | DF | Flavanols and tannins | Neolignans | ||||||
| Gallocatechin I | (+)-Catechin | Catechin derivative I | Catechin derivative II | Condensed tannins | Neolignan I | Neolignan II | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 3.61 | 4.07 | 7.75* | 0.13 | 8.74* | 22.70*** | 3.83 | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 133.39*** | 242.7*** | 175.8*** | 2.74 | 201.10*** | 88.92*** | 4.36 | |
| Age × Def. | 1 | 5.63* | 4.05 | 9.23* | 0.81 | 1.63 | 12.27** | 0.78 | |
| Error | 15 | ||||||||
| Eaten | 1 | 0.04 | 30.6*** | 5.44* | 0.05 | 1.67 | 6.11* | 0.47 | |
| Eaten × Def. | 1 | 0.04 | 1.20 | 0.07 | 0.03 | 0.50 | 0.17 | 0.67 | |
| Age × Eaten | 1 | 0.48 | 2.23 | 1.73 | 0.09 | 12.66** | 0.84 | 0.09 | |
| Age × Eaten × Def. | 1 | 0.00 | 0.28 | 0.99 | 0.32 | 0.81 | 0.037 | 0.34 | |
| Error | 14 | ||||||||
| Flavonols | |||||||||
| Quercitrin | Hyperin | Myricetin-3-galactoside | MC-isoquercitrin I | MC-isoquercitrin II | DC-astragalin | Total HPLC phenolics | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 37.76*** | 38.68*** | 29.15*** | 2.566 | 22.94*** | 0.068 | 59.04*** | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 228.12*** | 209.34*** | 13.15** | 11.07** | 10.38** | 219.61*** | 20.34*** | |
| Age × Def. | 1 | 31.67*** | 23.12*** | 0.53 | 0.13 | 2.35 | 0.73 | 6.51* | |
| Error | 15 | ||||||||
| Eaten | 1 | 11.28** | 1.13 | 2.65 | 0.23 | 4.06 | 14.91** | 3.33 | |
| Eaten × Def. | 1 | 0.00 | 0.03 | 0.76 | 0.28 | 0.15 | 0.06 | 0.46 | |
| Age × Eaten | 1 | 10.86** | 0.21 | 0.00 | 0.80 | 0.01 | 14.38** | 4.23 | |
| Age × Eaten × Def. | 1 | 0.04 | 0.41 | 0.65 | 1.29 | 0.01 | 1.92 | 3.2 | |
| Error | 14 | ||||||||
| Source | DF | Flavanols and tannins | Neolignans | ||||||
| Gallocatechin I | (+)-Catechin | Catechin derivative I | Catechin derivative II | Condensed tannins | Neolignan I | Neolignan II | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 3.61 | 4.07 | 7.75* | 0.13 | 8.74* | 22.70*** | 3.83 | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 133.39*** | 242.7*** | 175.8*** | 2.74 | 201.10*** | 88.92*** | 4.36 | |
| Age × Def. | 1 | 5.63* | 4.05 | 9.23* | 0.81 | 1.63 | 12.27** | 0.78 | |
| Error | 15 | ||||||||
| Eaten | 1 | 0.04 | 30.6*** | 5.44* | 0.05 | 1.67 | 6.11* | 0.47 | |
| Eaten × Def. | 1 | 0.04 | 1.20 | 0.07 | 0.03 | 0.50 | 0.17 | 0.67 | |
| Age × Eaten | 1 | 0.48 | 2.23 | 1.73 | 0.09 | 12.66** | 0.84 | 0.09 | |
| Age × Eaten × Def. | 1 | 0.00 | 0.28 | 0.99 | 0.32 | 0.81 | 0.037 | 0.34 | |
| Error | 14 | ||||||||
| Flavonols | |||||||||
| Quercitrin | Hyperin | Myricetin-3-galactoside | MC-isoquercitrin I | MC-isoquercitrin II | DC-astragalin | Total HPLC phenolics | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 37.76*** | 38.68*** | 29.15*** | 2.566 | 22.94*** | 0.068 | 59.04*** | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 228.12*** | 209.34*** | 13.15** | 11.07** | 10.38** | 219.61*** | 20.34*** | |
| Age × Def. | 1 | 31.67*** | 23.12*** | 0.53 | 0.13 | 2.35 | 0.73 | 6.51* | |
| Error | 15 | ||||||||
| Eaten | 1 | 11.28** | 1.13 | 2.65 | 0.23 | 4.06 | 14.91** | 3.33 | |
| Eaten × Def. | 1 | 0.00 | 0.03 | 0.76 | 0.28 | 0.15 | 0.06 | 0.46 | |
| Age × Eaten | 1 | 10.86** | 0.21 | 0.00 | 0.80 | 0.01 | 14.38** | 4.23 | |
| Age × Eaten × Def. | 1 | 0.04 | 0.41 | 0.65 | 1.29 | 0.01 | 1.92 | 3.2 | |
| Error | 14 | ||||||||
Results of ANOVA for the effects of (1) defoliation (intact needles in control versus defoliated trees), (2) needle age (current-year versus previous-year needles), and (3) damage by sawfly larvae (intact versus partly eaten needles) on the concentrations of phenolic compounds. Significant differences are indicated by asterisks as *, P < 0.05; **, P < 0.01 and ***, P < 0.001. Abbreviations: MC, monocoumaryl; DC, dicoumaryl; neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran; neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran and Def., defoliation.
| Source | DF | Flavanols and tannins | Neolignans | ||||||
| Gallocatechin I | (+)-Catechin | Catechin derivative I | Catechin derivative II | Condensed tannins | Neolignan I | Neolignan II | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 3.61 | 4.07 | 7.75* | 0.13 | 8.74* | 22.70*** | 3.83 | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 133.39*** | 242.7*** | 175.8*** | 2.74 | 201.10*** | 88.92*** | 4.36 | |
| Age × Def. | 1 | 5.63* | 4.05 | 9.23* | 0.81 | 1.63 | 12.27** | 0.78 | |
| Error | 15 | ||||||||
| Eaten | 1 | 0.04 | 30.6*** | 5.44* | 0.05 | 1.67 | 6.11* | 0.47 | |
| Eaten × Def. | 1 | 0.04 | 1.20 | 0.07 | 0.03 | 0.50 | 0.17 | 0.67 | |
| Age × Eaten | 1 | 0.48 | 2.23 | 1.73 | 0.09 | 12.66** | 0.84 | 0.09 | |
| Age × Eaten × Def. | 1 | 0.00 | 0.28 | 0.99 | 0.32 | 0.81 | 0.037 | 0.34 | |
| Error | 14 | ||||||||
| Flavonols | |||||||||
| Quercitrin | Hyperin | Myricetin-3-galactoside | MC-isoquercitrin I | MC-isoquercitrin II | DC-astragalin | Total HPLC phenolics | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 37.76*** | 38.68*** | 29.15*** | 2.566 | 22.94*** | 0.068 | 59.04*** | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 228.12*** | 209.34*** | 13.15** | 11.07** | 10.38** | 219.61*** | 20.34*** | |
| Age × Def. | 1 | 31.67*** | 23.12*** | 0.53 | 0.13 | 2.35 | 0.73 | 6.51* | |
| Error | 15 | ||||||||
| Eaten | 1 | 11.28** | 1.13 | 2.65 | 0.23 | 4.06 | 14.91** | 3.33 | |
| Eaten × Def. | 1 | 0.00 | 0.03 | 0.76 | 0.28 | 0.15 | 0.06 | 0.46 | |
| Age × Eaten | 1 | 10.86** | 0.21 | 0.00 | 0.80 | 0.01 | 14.38** | 4.23 | |
| Age × Eaten × Def. | 1 | 0.04 | 0.41 | 0.65 | 1.29 | 0.01 | 1.92 | 3.2 | |
| Error | 14 | ||||||||
| Source | DF | Flavanols and tannins | Neolignans | ||||||
| Gallocatechin I | (+)-Catechin | Catechin derivative I | Catechin derivative II | Condensed tannins | Neolignan I | Neolignan II | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 3.61 | 4.07 | 7.75* | 0.13 | 8.74* | 22.70*** | 3.83 | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 133.39*** | 242.7*** | 175.8*** | 2.74 | 201.10*** | 88.92*** | 4.36 | |
| Age × Def. | 1 | 5.63* | 4.05 | 9.23* | 0.81 | 1.63 | 12.27** | 0.78 | |
| Error | 15 | ||||||||
| Eaten | 1 | 0.04 | 30.6*** | 5.44* | 0.05 | 1.67 | 6.11* | 0.47 | |
| Eaten × Def. | 1 | 0.04 | 1.20 | 0.07 | 0.03 | 0.50 | 0.17 | 0.67 | |
| Age × Eaten | 1 | 0.48 | 2.23 | 1.73 | 0.09 | 12.66** | 0.84 | 0.09 | |
| Age × Eaten × Def. | 1 | 0.00 | 0.28 | 0.99 | 0.32 | 0.81 | 0.037 | 0.34 | |
| Error | 14 | ||||||||
| Flavonols | |||||||||
| Quercitrin | Hyperin | Myricetin-3-galactoside | MC-isoquercitrin I | MC-isoquercitrin II | DC-astragalin | Total HPLC phenolics | |||
| F | F | F | F | F | F | F | |||
| Between-subjects | |||||||||
| Defoliation | 1 | 37.76*** | 38.68*** | 29.15*** | 2.566 | 22.94*** | 0.068 | 59.04*** | |
| Error | 15 | ||||||||
| Within-subjects | |||||||||
| Needle age | 1 | 228.12*** | 209.34*** | 13.15** | 11.07** | 10.38** | 219.61*** | 20.34*** | |
| Age × Def. | 1 | 31.67*** | 23.12*** | 0.53 | 0.13 | 2.35 | 0.73 | 6.51* | |
| Error | 15 | ||||||||
| Eaten | 1 | 11.28** | 1.13 | 2.65 | 0.23 | 4.06 | 14.91** | 3.33 | |
| Eaten × Def. | 1 | 0.00 | 0.03 | 0.76 | 0.28 | 0.15 | 0.06 | 0.46 | |
| Age × Eaten | 1 | 10.86** | 0.21 | 0.00 | 0.80 | 0.01 | 14.38** | 4.23 | |
| Age × Eaten × Def. | 1 | 0.04 | 0.41 | 0.65 | 1.29 | 0.01 | 1.92 | 3.2 | |
| Error | 14 | ||||||||
Needles that were partially damaged by larvae had higher (+)-catechin (Figure 2B) and DC-astragalin (Figure 2H) concentrations than intact needles. The effect of feeding damage on some phenolics varied with needle age (significant interaction between needle age and damage, Table 1). The needle damage × defoliation treatment effect was not significant, indicating that the defoliation history of the trees did not influence the localized responses within the trees.
N and phenolics versus larval growth
The estimated effect of the identified phenolics on cocoon mass was generally negative, but some positive effects were found (Table 2). All the compounds in the intact needles that were entered in the stepwise regression model had a negative effect on cocoon mass (Table 3): two compounds in control trees and three compounds in defoliated trees. Correlation analyses did not reveal any unambiguous relationship between N and larval growth (cocoon mass) (Table 2). Furthermore, a regression analysis did not identify N among the group of variables that explained the changes in cocoon mass (Table 3). In needles partially eaten by sawfly larvae, we found only two phenols that affected cocoon mass: in previous-year needles of control trees both neolignan II and gallocatechin showed negative effects on cocoon mass. None of the phenolic compounds analyzed in previous-year needles of the previously defoliated trees or in current-year needles of either group of trees affected cocoon mass (data not shown).
Results of correlations (correlation coefficients = r) between concentrations of phenolics, nitrogen (N) concentration and cocoon viability (cocoon mass after winter). Results are given separately for current-year (C) and previous-year (C+1) needles in control and defoliated trees. Pooled r values are given for each of these classes. Abbreviations: MC, monocoumaryl; DC, dicoumaryl; neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran and neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran.
| Compound | C needles | C+1 needles | ||||||
| Control | Defoliated | Control | Defoliated | |||||
| Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | |
| Nitrogen | 0.001 | 0.45 | −0.65 | 0.63 | ||||
| Neolignan I | 0.09 | 0.02 | −0.16 | −0.14 | −0.45 | 0.32 | −0.63 | −0.53 |
| Neolignan II | −0.59 | 0.39 | −0.06 | 0.25 | −0.87 | 0.74 | 0.12 | 0.44 |
| Quercitrin | 0.19 | 0.73 | −0.35 | −0.88 | −0.40 | 0.31 | −0.43 | −0.01 |
| Hyperin | −0.09 | 0.30 | 0.20 | −0.36 | −0.50 | 0.18 | −0.25 | 0.03 |
| DC-astragalin | 0.39 | 0.10 | −0.35 | −0.24 | 0.41 | −0.38 | −0.61 | −0.17 |
| (+)-Catechin | 0.63 | 0.33 | −0.39 | 0.12 | 0.34 | −0.01 | −0.76 | −0.36 |
| Catechin derivative I | −0.18 | 0.06 | −0.28 | −0.47 | −0.42 | 0.79 | −0.21 | 0.20 |
| Catechin derivative II | 0.50 | 0.53 | 0.32 | 0.10 | 0.11 | −0.29 | −0.48 | −0.05 |
| Myricetin-3-galactoside | −0.32 | −0.31 | −0.60 | −0.16 | 0.26 | −0.31 | 0.12 | 0.44 |
| Gallocatechin | 0.48 | −0.22 | 0.22 | −0.32 | −0.42 | 0.26 | −0.08 | −0.03 |
| Condensed tannins | −0.20 | 0.24 | −0.25 | −0.45 | 0.13 | 0.08 | −0.02 | −0.29 |
| MC-isoquercitrin I | −0.04 | 0.30 | −0.60 | −0.06 | −0.29 | 0.40 | −0.40 | −0.30 |
| MC-isoquercitrin II | −0.66 | 0.28 | 0.06 | 0.00 | −0.73 | 0.51 | 0.01 | −0.12 |
| Pooled r | 0.01 | 0.21 | −0.01 | −0.24 | −0.30 | 0.24 | −0.24 | −0.06 |
| Compound | C needles | C+1 needles | ||||||
| Control | Defoliated | Control | Defoliated | |||||
| Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | |
| Nitrogen | 0.001 | 0.45 | −0.65 | 0.63 | ||||
| Neolignan I | 0.09 | 0.02 | −0.16 | −0.14 | −0.45 | 0.32 | −0.63 | −0.53 |
| Neolignan II | −0.59 | 0.39 | −0.06 | 0.25 | −0.87 | 0.74 | 0.12 | 0.44 |
| Quercitrin | 0.19 | 0.73 | −0.35 | −0.88 | −0.40 | 0.31 | −0.43 | −0.01 |
| Hyperin | −0.09 | 0.30 | 0.20 | −0.36 | −0.50 | 0.18 | −0.25 | 0.03 |
| DC-astragalin | 0.39 | 0.10 | −0.35 | −0.24 | 0.41 | −0.38 | −0.61 | −0.17 |
| (+)-Catechin | 0.63 | 0.33 | −0.39 | 0.12 | 0.34 | −0.01 | −0.76 | −0.36 |
| Catechin derivative I | −0.18 | 0.06 | −0.28 | −0.47 | −0.42 | 0.79 | −0.21 | 0.20 |
| Catechin derivative II | 0.50 | 0.53 | 0.32 | 0.10 | 0.11 | −0.29 | −0.48 | −0.05 |
| Myricetin-3-galactoside | −0.32 | −0.31 | −0.60 | −0.16 | 0.26 | −0.31 | 0.12 | 0.44 |
| Gallocatechin | 0.48 | −0.22 | 0.22 | −0.32 | −0.42 | 0.26 | −0.08 | −0.03 |
| Condensed tannins | −0.20 | 0.24 | −0.25 | −0.45 | 0.13 | 0.08 | −0.02 | −0.29 |
| MC-isoquercitrin I | −0.04 | 0.30 | −0.60 | −0.06 | −0.29 | 0.40 | −0.40 | −0.30 |
| MC-isoquercitrin II | −0.66 | 0.28 | 0.06 | 0.00 | −0.73 | 0.51 | 0.01 | −0.12 |
| Pooled r | 0.01 | 0.21 | −0.01 | −0.24 | −0.30 | 0.24 | −0.24 | −0.06 |
Results of correlations (correlation coefficients = r) between concentrations of phenolics, nitrogen (N) concentration and cocoon viability (cocoon mass after winter). Results are given separately for current-year (C) and previous-year (C+1) needles in control and defoliated trees. Pooled r values are given for each of these classes. Abbreviations: MC, monocoumaryl; DC, dicoumaryl; neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran and neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran.
| Compound | C needles | C+1 needles | ||||||
| Control | Defoliated | Control | Defoliated | |||||
| Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | |
| Nitrogen | 0.001 | 0.45 | −0.65 | 0.63 | ||||
| Neolignan I | 0.09 | 0.02 | −0.16 | −0.14 | −0.45 | 0.32 | −0.63 | −0.53 |
| Neolignan II | −0.59 | 0.39 | −0.06 | 0.25 | −0.87 | 0.74 | 0.12 | 0.44 |
| Quercitrin | 0.19 | 0.73 | −0.35 | −0.88 | −0.40 | 0.31 | −0.43 | −0.01 |
| Hyperin | −0.09 | 0.30 | 0.20 | −0.36 | −0.50 | 0.18 | −0.25 | 0.03 |
| DC-astragalin | 0.39 | 0.10 | −0.35 | −0.24 | 0.41 | −0.38 | −0.61 | −0.17 |
| (+)-Catechin | 0.63 | 0.33 | −0.39 | 0.12 | 0.34 | −0.01 | −0.76 | −0.36 |
| Catechin derivative I | −0.18 | 0.06 | −0.28 | −0.47 | −0.42 | 0.79 | −0.21 | 0.20 |
| Catechin derivative II | 0.50 | 0.53 | 0.32 | 0.10 | 0.11 | −0.29 | −0.48 | −0.05 |
| Myricetin-3-galactoside | −0.32 | −0.31 | −0.60 | −0.16 | 0.26 | −0.31 | 0.12 | 0.44 |
| Gallocatechin | 0.48 | −0.22 | 0.22 | −0.32 | −0.42 | 0.26 | −0.08 | −0.03 |
| Condensed tannins | −0.20 | 0.24 | −0.25 | −0.45 | 0.13 | 0.08 | −0.02 | −0.29 |
| MC-isoquercitrin I | −0.04 | 0.30 | −0.60 | −0.06 | −0.29 | 0.40 | −0.40 | −0.30 |
| MC-isoquercitrin II | −0.66 | 0.28 | 0.06 | 0.00 | −0.73 | 0.51 | 0.01 | −0.12 |
| Pooled r | 0.01 | 0.21 | −0.01 | −0.24 | −0.30 | 0.24 | −0.24 | −0.06 |
| Compound | C needles | C+1 needles | ||||||
| Control | Defoliated | Control | Defoliated | |||||
| Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | Cocoon mass | N conc. | |
| Nitrogen | 0.001 | 0.45 | −0.65 | 0.63 | ||||
| Neolignan I | 0.09 | 0.02 | −0.16 | −0.14 | −0.45 | 0.32 | −0.63 | −0.53 |
| Neolignan II | −0.59 | 0.39 | −0.06 | 0.25 | −0.87 | 0.74 | 0.12 | 0.44 |
| Quercitrin | 0.19 | 0.73 | −0.35 | −0.88 | −0.40 | 0.31 | −0.43 | −0.01 |
| Hyperin | −0.09 | 0.30 | 0.20 | −0.36 | −0.50 | 0.18 | −0.25 | 0.03 |
| DC-astragalin | 0.39 | 0.10 | −0.35 | −0.24 | 0.41 | −0.38 | −0.61 | −0.17 |
| (+)-Catechin | 0.63 | 0.33 | −0.39 | 0.12 | 0.34 | −0.01 | −0.76 | −0.36 |
| Catechin derivative I | −0.18 | 0.06 | −0.28 | −0.47 | −0.42 | 0.79 | −0.21 | 0.20 |
| Catechin derivative II | 0.50 | 0.53 | 0.32 | 0.10 | 0.11 | −0.29 | −0.48 | −0.05 |
| Myricetin-3-galactoside | −0.32 | −0.31 | −0.60 | −0.16 | 0.26 | −0.31 | 0.12 | 0.44 |
| Gallocatechin | 0.48 | −0.22 | 0.22 | −0.32 | −0.42 | 0.26 | −0.08 | −0.03 |
| Condensed tannins | −0.20 | 0.24 | −0.25 | −0.45 | 0.13 | 0.08 | −0.02 | −0.29 |
| MC-isoquercitrin I | −0.04 | 0.30 | −0.60 | −0.06 | −0.29 | 0.40 | −0.40 | −0.30 |
| MC-isoquercitrin II | −0.66 | 0.28 | 0.06 | 0.00 | −0.73 | 0.51 | 0.01 | −0.12 |
| Pooled r | 0.01 | 0.21 | −0.01 | −0.24 | −0.30 | 0.24 | −0.24 | −0.06 |
Results of stepwise regression analysis where cocoon mass is explained by compounds on the basis of statistical criteria. The standardized regression coefficient (β) and its significance (t test) are reported for each compound selected in the model. The coefficient of explanation (R2) and its significance are reported for each model. Modeling was performed separately for the intact needles of control and defoliated trees. Current-year (C) and previous-year needles (C+1) were included in the same model. Abbreviations: neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran and neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran.
| Treatment | Compound | β | t | P | R2 | P |
| Control | Neolignan II (C+1) | −0.92 | −8.20 | < 0.001 | 0.94 | 0.001 |
| Myricetin-3-galactoside (C) | −0.43 | −3.81 | 0.012 | |||
| Defoliated | (+)-Catechin (C+1) | −0.72 | −12.65 | < 0.001 | 0.99 | < 0.001 |
| Neolignan I (C+1) | −0.52 | −9.25 | < 0.001 | |||
| Catechin derivative I (C) | −0.38 | −6.88 | < 0.001 |
| Treatment | Compound | β | t | P | R2 | P |
| Control | Neolignan II (C+1) | −0.92 | −8.20 | < 0.001 | 0.94 | 0.001 |
| Myricetin-3-galactoside (C) | −0.43 | −3.81 | 0.012 | |||
| Defoliated | (+)-Catechin (C+1) | −0.72 | −12.65 | < 0.001 | 0.99 | < 0.001 |
| Neolignan I (C+1) | −0.52 | −9.25 | < 0.001 | |||
| Catechin derivative I (C) | −0.38 | −6.88 | < 0.001 |
Results of stepwise regression analysis where cocoon mass is explained by compounds on the basis of statistical criteria. The standardized regression coefficient (β) and its significance (t test) are reported for each compound selected in the model. The coefficient of explanation (R2) and its significance are reported for each model. Modeling was performed separately for the intact needles of control and defoliated trees. Current-year (C) and previous-year needles (C+1) were included in the same model. Abbreviations: neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran and neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran.
| Treatment | Compound | β | t | P | R2 | P |
| Control | Neolignan II (C+1) | −0.92 | −8.20 | < 0.001 | 0.94 | 0.001 |
| Myricetin-3-galactoside (C) | −0.43 | −3.81 | 0.012 | |||
| Defoliated | (+)-Catechin (C+1) | −0.72 | −12.65 | < 0.001 | 0.99 | < 0.001 |
| Neolignan I (C+1) | −0.52 | −9.25 | < 0.001 | |||
| Catechin derivative I (C) | −0.38 | −6.88 | < 0.001 |
| Treatment | Compound | β | t | P | R2 | P |
| Control | Neolignan II (C+1) | −0.92 | −8.20 | < 0.001 | 0.94 | 0.001 |
| Myricetin-3-galactoside (C) | −0.43 | −3.81 | 0.012 | |||
| Defoliated | (+)-Catechin (C+1) | −0.72 | −12.65 | < 0.001 | 0.99 | < 0.001 |
| Neolignan I (C+1) | −0.52 | −9.25 | < 0.001 | |||
| Catechin derivative I (C) | −0.38 | −6.88 | < 0.001 |
Discussion
The reduced cocoon mass and mass of hatched D. pini fed on the repeatedly defoliated Scots pine trees suggest delayed induced resistance. In contrast, Niemelä et al. (1984, 1991) had observed no DIR in the mature foliage of Scots pine against Diprionid sawflies. Niemelä et al. (1991) had concluded that Scots pine lacks DIR or Diprionid sawflies, including D. pini, as specialized folivores of Pinus are highly tolerant to inducible chemical changes in foliage.
Several studies have shown that the presence of DIR is related to changes in the nutrient status of the defoliated trees (Tuomi et al. 1984, 1988b, Niemelä et al. 1991, Watt et al. 1991, McMillin and Wagner 1997, Hódar et al. 2004). Accordingly, in the study by Niemelä et al. (1991), the absence of DIR in Scots pine foliage was associated with no change or slightly increased N concentrations in needles of the previously defoliated trees. In contrast, in our experiment carried out a year after the second defoliation treatment, N concentrations were still lower in the current-year (−11.5%) and the previous-year (−7.2%) needles of the previously defoliated trees compared with control values. In a previous study carried out in 1999 by Roitto et al. (2003) on the same control and defoliated trees, the decline in N concentrations in current-year needles of the previously defoliated trees was greater (−19%) than that observed in our study. In both years of our study, we found that the previous defoliation treatments enhanced the production of several phenolic compounds.
We suggest that severe needle loss at the nutrient-poor study site altered Scots pine needle chemistry leading to adverse performance of D. pini. At the study site, low concentrations of N (< 0.01%) prevail in the dry, sandy soil (Kuikka et al. 2003). Soil nutrient supply has been shown to correlate inversely with phenol production in Pinus species (Gebauer et al. 1998, Lavola et al. 2003, Kraus et al. 2004, Saxon et al. 2004). In Betula spp., N deficiency resulted in the induction of largely the same phenolic compounds as in our study, i.e., myricetin and quercetin glycosides and soluble condensed tannins (Keski-Saari and Julkunen-Tiitto 2003, Keski-Saari et al. 2005). It is known that resource availability and external stresses can differentially affect different groups of secondary compounds as well as individual compounds (e.g., Gershenzon 1984, Koricheva et al. 1998, Keski-Saari and Julkunen-Tiitto 2003, Lavola et al. 2003, Keski-Saari et al. 2005). The mechanisms that regulate these plastic changes in phenolic composition are now being studied (e.g., Arnold et al. 2004, Harding et al. 2005, Fritz et al. 2006). In deciduous trees, the mechanism underlying delayed induced resistance has been linked to the oxidative properties of certain phenolic compounds (Ruuhola et al. 2007). Similar to our results, the activities of oxidative peroxidase enzymes increased in the undamaged needles of the defoliated trees in 1999 (Roitto et al. 2003).
The phenolic compounds that were entered in the regression models of sawfly performance had a negative effect on larval growth (standardized regression coefficient, β varying from −0.38 to −0.92). Although the most important compounds were different in the models for control and defoliated trees, the models of intact needles explained over 93% of the variation in larval growth. However, the effects of specific phenolic compounds in pine needles on defoliating insects and insect adaptations to phenolics are not well known (Mumm and Hilker 2006). In addition, any single measure of phenolics is a snapshot in time (see Gayler et al. 2008). Although protein content of the food is important for larval growth (Mattson 1980), N did not have a consistent effect on larval growth among control or defoliated trees.
It is possible that mechanical damage induces less specific rapid responses than feeding or oviposition by herbivores (Reymond et al. 2000, Schröder et al. 2007). This postulated difference does not, however, resolve the discrepancy between our results and those of Niemelä et al. (1984, 1991) who also applied mechanical defoliation. We found that mechanical damage induced some phenolic compounds more than others. Few studies have examined if damage, artificial or natural, preferentially induces defensive effective compounds over nondefensive compounds (Tuomi et al. 1988a).
Among the phenolic compounds that we had analyzed, myricetin-3-galactoside showed the strongest induced increase (104%) in the current-year (C) needles and a 71% increase in the previous-year (C+1) needles (Table 4), and significantly explained the variation in sawfly performance among control trees (Table 3). The other compounds that explained variation in sawfly performance ((+)-catechin, catechin derivative II, and neolignans I and II) showed weaker inducible responses compared with myricetin-3-galactoside. In contrast, hyperin, condensed tannins and quercitrin, which showed a strong response in C or C+1 needles, or both explained none of the variations in sawfly performance within the treatment groups (Table 3). Consequently, the compounds that increased in the defoliated trees compared with the control trees (Table 4) were not the same compounds that explained the variation in cocoon mass (Table 3). It thus seems that, in our study, phenolic compounds were induced irrespective of their effects on larval growth. Raffa (1991) found a disproportionate increase in monoterpenes that were most deleterious to bark beetles during conifer subcortical induced responses. We searched for such associations as supporting evidence that the observed inducible responses are related to specific defensive responses against the folivores. However, only a weak or no such pattern was detected, so we could not reject the null hypothesis that phenolic compounds are induced irrespective of their effects on larval growth. Consequently, our results appear to follow the carbon–nutrient balance models that predict both defensive as well as nondefensive compounds to accumulate in response to defoliation.
Percent change in the concentrations of phenolics in current-year (C) and previous-year (C+1) needles in defoliated trees compared with control trees. Abbreviations: MC, monocoumaryl; DC, dicoumaryl; neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran and neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran.
| Compound | C | C+1 |
| Quercitrin | 36 | 94 |
| Hyperin | 99 | 87 |
| Myricetin-3-galactoside | 104 | 71 |
| MC-isoquercitrin I | 26 | 14 |
| MC-isoquercitrin II | 45 | 70 |
| DC-astragalin | 11 | −41 |
| (+)-Catechin | 37 | 28 |
| Catechin derivative I | 19 | 36 |
| Catechin derivative II | 15 | −3 |
| Gallocatechin | 10 | 26 |
| Condensed tannins | 90 | 34 |
| Neolignan I | 18 | 48 |
| Neolignan II | 14 | 16 |
| Compound | C | C+1 |
| Quercitrin | 36 | 94 |
| Hyperin | 99 | 87 |
| Myricetin-3-galactoside | 104 | 71 |
| MC-isoquercitrin I | 26 | 14 |
| MC-isoquercitrin II | 45 | 70 |
| DC-astragalin | 11 | −41 |
| (+)-Catechin | 37 | 28 |
| Catechin derivative I | 19 | 36 |
| Catechin derivative II | 15 | −3 |
| Gallocatechin | 10 | 26 |
| Condensed tannins | 90 | 34 |
| Neolignan I | 18 | 48 |
| Neolignan II | 14 | 16 |
Percent change in the concentrations of phenolics in current-year (C) and previous-year (C+1) needles in defoliated trees compared with control trees. Abbreviations: MC, monocoumaryl; DC, dicoumaryl; neolignan I, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-β-glucopyranosylphenyl)-2,3-dihydrobenzofuran and neolignan II, 3-hydroxymethyl-5-(γ-hydroxy-n-propyl)-2(3′-methoxy-4′-O-α-rhamnopyranosylphenyl)-2,3-dihydrobenzofuran.
| Compound | C | C+1 |
| Quercitrin | 36 | 94 |
| Hyperin | 99 | 87 |
| Myricetin-3-galactoside | 104 | 71 |
| MC-isoquercitrin I | 26 | 14 |
| MC-isoquercitrin II | 45 | 70 |
| DC-astragalin | 11 | −41 |
| (+)-Catechin | 37 | 28 |
| Catechin derivative I | 19 | 36 |
| Catechin derivative II | 15 | −3 |
| Gallocatechin | 10 | 26 |
| Condensed tannins | 90 | 34 |
| Neolignan I | 18 | 48 |
| Neolignan II | 14 | 16 |
| Compound | C | C+1 |
| Quercitrin | 36 | 94 |
| Hyperin | 99 | 87 |
| Myricetin-3-galactoside | 104 | 71 |
| MC-isoquercitrin I | 26 | 14 |
| MC-isoquercitrin II | 45 | 70 |
| DC-astragalin | 11 | −41 |
| (+)-Catechin | 37 | 28 |
| Catechin derivative I | 19 | 36 |
| Catechin derivative II | 15 | −3 |
| Gallocatechin | 10 | 26 |
| Condensed tannins | 90 | 34 |
| Neolignan I | 18 | 48 |
| Neolignan II | 14 | 16 |
In conclusion, we found evidence for the DIR in Scots pine in terms of reduced sawfly performance on defoliated trees that were associated with reduced N concentration and increased concentrations of some phenolic compounds in the foliage. Needle N concentration did not explain sawfly performance, whereas changes in the concentrations of some phenolic compounds correlated with sawfly performance. Increased concentrations of the monitored compounds, however, seemed to have no clear relationship to their adverse effects on sawflies. Instead, the phenolics that accumulated in defoliated trees were largely the same compounds that have been reported to accumulate in response to N deficiency. Consequently, the underlying mechanisms of DIR in response to the previous defoliation may differ from the mechanisms underlying the rapidly induced response to leaf damage. Our results highlight the need for a deeper physiologic understanding of the role of phenolic compounds in plant defenses.
This study was financially supported by the Foundation for Research of Natural Resources in Finland and by the Academy of Finland (Projects #40951, #50451, and #80486).
