Effect of phenology on susceptibility of Norway spruce (Picea abies) to fungal pathogens

Authors


E-mail: paal.krokene@skogoglandskap.no

Abstract

Ceratocystis polonica and Heterobasidion parviporum are important fungal pathogens in Norway spruce (Picea abies). Tree susceptibility to these pathogens with respect to phenology was studied using artificial fungal inoculations at six stages of bud development, and assessed by measuring phloem necroses in the stems of 2- and 8-year-old trees. Tree capacity for resistance was assessed by measuring phloem nonstructural carbohydrates at each stage. Phloem necroses were significantly larger in trees with fungal versus control inoculations and increased significantly over time. Changes in nonstructural carbohydrates occurred in the trees; a significant decline in starch and a slight but significant increase in total sugars occurred over time. These results suggest that susceptibility to fungal pathogens and carbohydrate levels in the stems of the trees were related to fine-scale changes in bud development. A trade-off may occur between allocation of starch (the major fraction of the stem carbohydrate pool) to bud development/shoot growth versus defence of the stem. Previous tests of plant defence hypotheses have focused on herbivory on plants growing under different environmental conditions, but the role of phenology and the effect of pathogens are also important to consider in understanding plant resource allocation patterns.

Introduction

To grow or defend is a fundamental dilemma faced by all plants (Herms & Mattson, 1992). If resources are limited, the allocation of resources to growth versus defence, the timing of allocation, and the type of resources allocated all play important roles in plant life history strategy and fitness. Thus, plant defence has been the subject of in-depth research for more than 60 years. Several hypotheses have served as frameworks for investigation, but a unifying theory of plant defence has not emerged (reviewed by Stamp, 2003). Even the growth–differentiation balance hypothesis, or GDBH (Loomis, 1932; Lorio, 1986; Herms & Mattson, 1992), which is considered to be the most mature of plant defence hypotheses (Stamp, 2003, 2004), has only recently been methodically tested (e.g. Glynn et al., 2007).

Certain areas of plant defence research have also received less attention over the years, such as the effect of pathogens or phenology on whole-plant defensive responses (but see Barton & Koricheva, 2010). The majority of plant defence research has focused on insect herbivory (Mattson, 1980; Coley et al., 1985; Coley & Barone, 1996; Stamp, 2003), while the cost of defence against pathogens has received less attention (Coley & Barone, 1996). Plant defence of stems and other woody tissues is also often overlooked, as is the role of phenology in resource allocation to tissues other than leaves. However, defence of leaves versus woody tissue and the timing of defence relative to phenology are also important components of trade-offs in plant resource allocation. A better understanding of stem defences and effects of phenology on susceptibility to pathogens will improve the ability to test theories of plant defence.

Several theories of plant defence have been formulated over the years to explain plant resource allocation and resistance under varying environmental conditions such as nutrient, light or water availability (reviewed by Stamp, 2003). The GDBH proposes a constant trade-off between growth and differentiation at the cellular level. Growth refers to any process that requires cell division and elongation, such as producing roots, leaves or woody tissue. Differentiation refers to all processes that change the structure or function of pre-existing cells, such as producing trichomes or secondary metabolites. The GDBH therefore describes a trade-off in where and when resources are allocated; for example, the physical process of cell expansion must occur before subsequent lignification of cell walls, independent of the actual level of resources available for allocation to either function (Lorio, 1986; Lorio & Sommers, 1986; Herms & Mattson, 1992). Although the hypothesis is generally tested by varying resource availability, the framework of the GDBH could also be used under constant environmental conditions to understand plant defence in relation to phenology, and to observe plant allocation to stem defences against pathogens.

This study addresses the question of whether tree susceptibility to fungal pathogens in the main stem changes in relation to bud development. The effects of two of the most economically important fungal pathogens of Norway spruce (Picea abies): Heterobasidion parviporum, a major root and butt rot pathogen of primarily Norway spruce (Korhonen & Stenlid, 1998), and Ceratocystis polonica, a virulent blue-stain fungus vectored by the European spruce bark beetle, Ips typographus (Solheim, 1993), were studied. Fungal inoculations were performed under greenhouse and field conditions and tree susceptibility was compared in relation to phenology for two age classes. The following hypotheses were tested: (i) tree susceptibility to a fungal pathogen, measured by the amount of stem tissue necrosis, will increase over the course of bud development; and (ii) phloem nonstructural carbohydrate levels will decline over the course of bud development as resources are invested in shoot growth, lessening the potential for allocation to defence in the stem.

Materials and methods

Greenhouse experiment

Two-year-old containerized Norway spruce seedlings were obtained from a nursery, potted in 8-cm-diameter pots with a 3:1 mixture of peat and perlite in April 1998, and placed in a greenhouse at Hogsmark Experimental Farm, Ås, Norway. Plants were watered regularly and fertilized with a complete nutrient solution delivered with the water (based on 50% commercial fertilizer Red Superba and 50% ammonium sulphate, providing micronutrients and NPK in the proportions 100:14:72). There was no artificial lighting or heating in the greenhouse. On each of six occasions between 30 April and 26 June 1998, 20 plants were inoculated in different bud development phases with actively growing mycelium of the blue-stain fungus C. polonica (isolate NFLI 93-208/115) cultured on malt agar medium (2% malt, 1·5% agar) or with sterile malt agar as a control. On the first three occasions bud development was determined on a scale from 0 to 6 according to Krutzsch (1973) (e.g. 0 = dormant buds, 3 = burst of bud scales, 6 = elongation of shoot, basal needles not yet spread), but at later stages bud development was expressed as the percentage of final shoot length that was completed at the time of inoculation (Table 1). This was done because it can be difficult to accurately assess the later stages of bud development using Krutzsch’s scale. Plants were inoculated through a small wound (3 × 7 mm) midway up on the first internode. Inoculum was placed beneath the bark, and Parafilm was wrapped around the inoculation wound to seal the stem and prevent contamination and excessive drying of phloem and sapwood. Eight weeks after inoculation the plants were harvested and phloem necrosis lengths measured. Plant height, current year shoot length, and diameter at the inoculation site were measured at the time of inoculation and at harvesting. Temperature in the greenhouse was not measured during the experiment, but other studies at the same site have shown that mean daily greenhouse temperatures are 6·7, 4·6 and 4·8°C higher than outside temperatures in May–June, July and August, respectively (P. Krokene, Norwegian Forest and Landscape Institute, unpublished data). Greenhouse temperatures were thus extrapolated by adding these values to the outside temperatures.

Table 1.   Date, bud development stage and accumulated temperature sum at the time of inoculation for 8-year-old field-grown Norway spruce trees and 2-year-old trees grown in a greenhouse
Inoculation time8-year-old2-year-old
Inoculation dateBud developmentaDDbInoculation dateBud developmentaDDb
  1. aFor the first three inoculation times bud development was assessed according to Krutzsch (1973). At later stages bud development was expressed as the percentage of final shoot length that was completed at the time of inoculation. Data are presented as mean (±SD), except for the first three inoculation times in 2-year-old plants, where the range is presented.

  2. bAccumulated degree-days above 5°C at the time of inoculation.

118 May1·9 (±0·57)14430 April0150
23 June3·3 (±0·93)24013 May2–3299
315 June4·5 (±0·74)33118 May5–6379
430 June67% (±13·8)4632 June77% (±11·5)570
510 July86% (±9·7)56910 June70% (±14·1)682
623 July100% (±0·0)69526 June87% (±13·7)923

Field experiment

Four full-sib families of 8-year-old Norway spruce saplings growing in a progeny test (Skrøppa & Lindgren, 1994), planted out as 2-year-old seedlings in 1992 at the Hogsmark Experimental Farm, were selected. On each of six occasions between 18 May and 23 July 1998, 10 trees from each family were inoculated with fungal mycelium on malt agar. Each tree was inoculated once with H. parviporum (isolate NFLI 87-257/1, using a 8-mm cork borer) approximately 10 cm above the soil line and once with C. polonica (NFLI 93-208/115, using a 5-mm cork borer) midway up on the third internode. Inoculation dates were chosen so that bud development stages matched between the 2- and 8-year-old trees for each round of fungal inoculation (Table 1). Ten weeks after inoculation the trees were harvested and phloem necrosis lengths measured. Temperature data were obtained from the meteorological station at the Norwegian University of Life Sciences 3·9 km away. Mean daily temperatures at these two sites are very similar during the growth season (R2 = 0·97, = 0·98x−0·096; P. Krokene, Norwegian Forest and Landscape Institute, unpublished data).

Nonstructural carbohydrate analysis

At each bud development stage, five uninoculated 2-year-old seedlings were sampled for analysis of carbohydrates by removing all bark/phloem on the lower 10 cm of the stem. Because of the destructive sampling method, samples were not collected from inoculated trees. In the 8-year-old trees, bark/phloem samples were collected from four trees per family (two uninoculated trees and two trees that were to be inoculated) at each bud development stage (= 16 per stage). For both age classes, samples were taken immediately before inoculation. All samples were frozen on liquid nitrogen and stored at −40°C until analysis. Each plant was analysed separately as described in Dalen et al. (2001). Briefly, frozen samples were homogenized in ethanol, rinsed, dried, redissolved in water and filtered before HPLC analysis of total sugars (glucose, fructose and sucrose). Starch was extracted from the residue with 60% perchloric acid and the concentration determined colorimetrically at 650 nm using iodide.

Statistical analysis

Data were analysed using the jmp 8·0 software package (SAS Institute) and pasw Statistics 18 (SPSS Inc.). Data were log(+ 1) transformed before analysis to correct for unequal variance and departures from normality. For clarity and ease of comparison, data are graphically presented using inoculation time as a surrogate for the date of inoculation; however, data for each age class were analysed using the actual date of inoculation as the independent variable (see Table 1). A general linear model with treatment (fungal inoculation versus control) as a factor and inoculation date as a covariate was used to determine significant treatment effects. Phloem necrosis length was further analysed using linear, quadratic and cubic regressions, and a partial F-test was performed to select the best-fitting model. When no significant difference occurred between models, the simplest model was chosen. Carbohydrate data were analysed using linear regression.

Results

Neither mortality nor visible foliar discoloration occurred for the 2-year-old greenhouse seedlings during the 8-week fungal incubation period. Phloem necrosis length differed significantly between control trees and trees inoculated with fungus (F3, 235 = 49·840, P < 0·0001, Fig. 1a). For trees inoculated with C. polonica, necrosis length changed significantly over bud development phases (= 0·858 + 0·0197x−0·000312x2, R2 = 0·210, P < 0·0001). Using this function, necrosis lengths were longer when inoculations were performed during the more active phases of bud and shoot elongation, peaking at day 32 of the experiment (between inoculation times 3 and 4). A marked drop in necrosis length occurred at the last inoculation time. Although a partial F-test revealed no significant difference between quadratic and cubic regressions, in this case a cubic regression provided a better visual fit than the quadratic expression shown above. Using a cubic regression (= 0·909 + 0·00552+ 0·000331x2−0·00000725x3, R2 = 0·233, < 0·0001), peak necrosis length occurred at day 37 (between inoculation times 4 and 5). For trees with control inoculations, necrosis length decreased slightly but significantly throughout the experiment (= 0·916−0·004x, R2 = 0·335, < 0·0001).

Figure 1.

 Phloem necrosis length following fungal inoculation for 2-year-old (a) and 8-year-old (b) Norway spruce (Picea abies) trees. Bud development and date at each inoculation time are given in Table 1. Black bars represent inoculation with Ceratocystis polonica in both panels. White bars represent inoculation with sterile control in panel (a) and grey bars represent inoculation with Heterobasision parviporum in panel (b). Data are mean ± 1 SEM.

The 8-year-old field trees showed a similar pattern of phloem symptoms as the 2-year-old trees. Trees inoculated with C. polonica had longer lesions than trees inoculated with H. parviporum (F3, 469 = 40·658, < 0·0001, Fig. 1b). Again, an initial increase in phloem necrosis length for trees inoculated with C. polonica was followed by a decline (= 1·278 + 0·0164x−0·000223x2, R2 = 0·138, < 0·0001) (Fig. 1b). For trees inoculated with H. parviporum, phloem lesion lengths showed no significant change over time (Fig. 1b).

Fungal inoculation did not seem to affect growth in the 2-year-old trees, as there was no effect of treatment on diameter or height growth following inoculation (F1, 228 = 0·95, = 0·33 and F1, 228 = 2·23, = 0·14, respectively). There was no significant difference in current-year shoot length between trees inoculated early or late in the season [16·9 ± 2·62 and 16·7 ± 3·54 cm shoot growth, respectively (mean ± SD)], indicating that all trees had completed height growth at the time of harvesting.

Significant changes in phloem nonstructural carbohydrates over the course of bud development occurred in both tree age classes. Starch showed a particularly consistent and significant decline over time in 8-year-old trees (= 2·327−0·008x, R2 = 0·615, < 0·0001; Fig. 2a) and also a significant decline over time in 2-year-old trees (= 2·311−0·004x, R2 = 0·097, = 0·009; Fig. 2a). Total sugars (the sum of glucose, fructose and sucrose) initially represented a much smaller component of nonstructural carbohydrates than starch, but showed a slight but significant increase over time in 8-year-old trees (= 1·712 + 0·002x, R2 = 0·190, < 0·0001; Fig. 2b) and in 2-year-old trees (= 2·002 + 0·002x, R2 = 0·177, < 0·0001; Fig. 2b).

Figure 2.

 Levels of (a) starch and (b) total sugars (glucose, fructose and sucrose) in Norway spruce (Picea abies) phloem prior to fungal inoculation. Solid lines represent 2-year-old trees and dashed lines represent 8-year-old trees. Bud development and date at each inoculation time are given in Table 1. Data are mean ±1 SEM for each inoculation time.

Temperature increased gradually up until the last inoculation date in both experiments (Fig. 3). Although temperature sums increased linearly over time, this did not cause a similar linear increase in tree susceptibility to fungal inoculation (Fig. 3). On the contrary, phloem lesion length had begun to decline in both age classes when temperatures reached their maximum. In 2-year-old trees lesion length declined sharply at the last inoculation time, even though the temperature sum continued to increase (Fig. 3). There was no correlation between mean daily temperatures 7, 14 or 21 days after each inoculation date and lesion length (R2 = 0·0003–0·029 for the different intervals, = 0·75–0·97).

Figure 3.

 Mean daily outside temperatures during the experiment (thick solid line) and accumulated temperature sums above 5°C for the greenhouse (solid line) and field trials (dashed line). Inoculation dates for the greenhouse (filled triangles) and field trial (open triangles) are indicated above the x-axis.

Discussion

The susceptibility of Norway spruce to a fungal pathogen in the main stem varied significantly with bud development in both 2- and 8-year-old trees. Susceptibility to C. polonica increased during early bud development, and then decreased as growth slowed in both age classes. This suggests that resistance in the stem is affected both by the resource demands of bud development and by fine-scale phenological changes that lead to trade-offs in resource allocation to growth versus defence. The larger phloem necroses in 8-year-old trees perhaps resulted from these trees having more tissue to support during bud development and shoot growth than smaller 2-year-old individuals. Heterobasidion parviporum caused a small but insignificant increase in necrosis length over the course of bud development in 8-year-old trees, and this study and others demonstrate that H. parviporum has rather low virulence when inoculated in the phloem of Norway spruce (Lindberg & Johansson, 1991; Kvaalen & Solheim, 2000). Control inoculations were not performed for the 8-year-old trees in this study, but only minor necroses were observed following control inoculations of sterile agar in 2-year-old trees and previous studies demonstrated that control inoculations cause a negligible wounding response at different times of the year in Norway spruce trees of varying ages (Krokene & Solheim, 1997, 1998, 2001). It is therefore concluded that C. polonica inoculations in 8-year-old trees also demonstrate that the level of susceptibility to fungal pathogens is related to tree phenology.

In order to assess Norway spruce capacity for resistance, phloem nonstructural carbohydrates were measured at each bud development stage. Phloem nonstructural carbohydrates, particularly starch, are frequently remobilized for resin production and other stem defences in Norway spruce (Christiansen & Ericsson, 1986; Horntvedt, 1988). In both age classes, measurements were done on uninoculated trees and represent a snapshot in time of carbon compounds available for growth or defence. It was hypothesized that phloem carbohydrates would decline over the course of bud development as resources were used for growth. A significant decline in starch occurred for both 2- and 8-year-old trees, but total sugars (the sum of glucose, fructose and sucrose) showed a slight increase over time. Starch represents a much greater fraction of the stem total carbohydrate pool than sugars (Li et al., 2002; Hoch et al., 2003) and is frequently the main compound reallocated to growth (Schädel et al., 2009). The present data suggest a trade-off between allocation of starch to bud development and shoot growth versus defence of the stem. Alternatively, the observed increase in total sugars could improve host nutritional quality and thereby increase stem susceptibility to pathogens (e.g. Wargo, 1972). However, the increase in total sugars was slight relative to the decrease in starch, and this alternate explanation cannot account for the sharp increase in tree resistance that occurred at the end of bud development in the 2-year-old trees, when starch levels continued to decrease. Support for a trade-off in starch allocation to growth versus defence in Norway spruce comes from Christiansen & Ericsson (1986), who observed a decline in phloem starch following inoculation with C. polonica and concluded that starch reserves were remobilized for resin production. Neither Christiansen & Ericsson (1986) nor Horntvedt (1988) found a correlation between Norway spruce resistance and starch levels at the time of inoculation, but neither of these studies explicitly accounted for changes in susceptibility or resistance as a result of phenology.

Previous tests of plant defence hypotheses, such as the GDBH, have focused mainly on herbivory, but the role of phenology and the effect of pathogens are clearly also important to consider in understanding plant resource allocation patterns. The present results indicate that bud phenology influences whole-plant defensive responses and that stem susceptibility to C. polonica is correlated with fine-scale changes in bud development in different age classes. In accordance with the GDBH, when bud development slowed, tree resistance in the stem increased immediately. This increase occurred over a very short time period, particularly in 2-year-old trees where the time span between minimum and maximum resistance to this fungal pathogen was only 2 weeks. Although overall growth did not differ between fungus- and control-inoculated 2-year-old trees at each time point, rapid changes in stem susceptibility over time highlight the influence of fine-scale changes in phenology on resource allocation. It is therefore suggested that future tests of plant defence theories explicitly account for changes in phenology or ontogeny (Barton & Koricheva, 2010).

A better understanding of the role of phenology and the effect of pathogens on plant defence has both theoretical and management implications. In Norway spruce, phenology and tree defence are important in the context of climate change. Climate change is likely to shift bud development earlier in the spring, thereby affecting tree susceptibility to herbivores and pathogens at different times. In addition, increasing temperatures may affect herbivores directly by changing their development time and voltinism (e.g. Ayres & Lombardero, 2000; Berg et al., 2006). Norway spruce is the main host for the spruce bark beetle (I. typographus), an important forest pest in northern Europe that vectors the fungus C. polonica. Under climate change scenarios, beetles are projected to switch from 1 to 2 generations per year in Scandinavia (Lange et al., 2006; Jönsson et al., 2009). The likely occurrence of a second beetle generation later in the summer, when trees may be much more susceptible to stem pathogens like C. polonica (Horntvedt, 1988), emphasizes the need for detailed knowledge of how phenology affects tree resistance. This study is the first comprehensive attempt to assess the role of phenology in Norway spruce susceptibility to a pathogen, and provides useful information for predicting the future impact of both I. typographus and fungal pathogens on Norway spruce.

Acknowledgements

We thank Gry Alfredsen, Olaug Olsen, Geir Østreng and Torild Wickstrøm for excellent technical assistance. Financial support for this work was provided by the Research Council of Norway (grant 199346/I10). EL was supported by a NSF-GRFP Nordic Research Opportunity grant.

Ancillary