Reserves Accumulated in Non-Photosynthetic Organsduring the Previous Growing Season Drive Plant Defensesand Growth in Aspen in the Subsequent Growing Season

Ahmed Najar & Simon M. Landhäusser &

Justin G. A. Whitehill & Pierluigi Bonello & Nadir Erbilgin

Received: 10 June 2013 /Revised: 26 November 2013 /Accepted: 3 December 2013 /Published online: 24 December 2013# Springer Science+Business Media New York 2013

Abstract Plants store non-structural carbohydrates (NSC),nitrogen (N), as well as other macro and micronutrients, intheir stems and roots; the role of these stored reserves in plantgrowth and defense under herbivory pressure is poorly under-stood, particularly in trees. Trembling aspen (Populustremuloides) seedlings with different NSC and N reservesaccumulated during the previous growing season were gener-ated in the greenhouse. Based on NSC and N contents, seed-lings were assigned to one of three reserve statuses: Low N–Low NSC, High N–Medium NSC, or High N–High NSC. Inthe subsequent growing season, half of the seedlings in eachreserve status was subjected to defoliation by forest tentcaterpillar (Malacosoma disstria) while the other half was leftuntreated. Following defoliation, the effect of reserves wasmeasured on foliar chemistry (N, NSC) and caterpillar perfor-mance (larval development). Due to their importance in her-bivore feeding, we also quantified concentrations of phenolicglycoside compounds in foliage. Seedlings in Low N-LowNSC reserve status contained higher amounts of inducedphenolic glycosides, grew little, and supported fewer caterpil-lars. In contrast, aspen seedlings in High N-Medium or HighNSC reserve statuses contained lower amounts of inducedphenolic glycosides, grew faster, and some of the caterpillarswhich fed on these seedlings developed up to their fourthinstar. Furthermore, multiple regression analysis indicated that

foliar phenolic glycoside concentration was related to reservechemistry (NSC, N). Overall, these results demonstrate thatreserves accumulated during the previous growing season caninfluence tree defense and growth in the subsequent growingseason. Additionally, our study concluded that the NSC/Nratio of reserves in the previous growing season represents abetter measure of resources available for use in defense andgrowth than the foliar NSC/N ratios.

Keywords Aspen . Constitutive and induced defenses .

Non-structural carbohydrates . Nitrogen . Phenolic glycosides

Introduction

Trees deploy a combination of anatomical (e.g., polyphenolicparenchyma cells, or traumatic resin ducts) and biochemical(e.g., phenolic glycosides) defense mechanisms againstherbivores and pathogens. These defenses can be constitutive orinduced (Eyles et al. 2010; Franceschi et al. 2005). Constitutivedefenses are always present in a tree to discourage attackers,while induced responses are triggered by tissue damage and canlimit further injury from attacking organisms (Bonello et al.2006; Eyles et al. 2010). However, the development ofthese defense responses is metabolically costly for plants andrequires resources in the form of carbohydrates and nutrients(Frost et al. 2008).

In woody plants, non-structural carbohydrates, such asstarch and soluble sugars, as well as amino acids, other macroor micronutrients are stored in non-photosynthetic organs,such as branches, stems, and roots. Evidence emergingfrom studies that focus on functions of reserves in plantphysiological processes, such as growth and defense, hassuggested that reserves can be critical for plant survival whenplants are under severe carbon stress resulting from limitationsin photosynthesis (e.g., due to drought) or destruction of

A. Najar : S. M. Landhäusser :N. Erbilgin (*)Department of Renewable Resources, University of Alberta,442 Earth Sciences Building, Edmonton, AB, Canadae-mail: erbilgin@ualberta.ca

J. G. A. Whitehill : P. BonelloDepartment of Plant Pathology, The Ohio State University,Columbus, OH, USA

J. G. A. WhitehillMichael Smith Laboratories, University of British Columbia,Vancouver, BC, Canada

J Chem Ecol (2014) 40:21–30DOI 10.1007/s10886-013-0374-0

photosynthetic organs (e.g., by defoliation) (Dunn et al. 1990;Landhäusser and Lieffers 2012; McDowell et al. 2008; Salaet al. 2012; Sampedro et al. 2011). For instance, after repeatedsevere defoliations by herbivores, the basic functions of aplant and the renewal of the foliage are supported by thereserves accumulated in the non-photosynthetic organs(Babst et al. 2005, 2008; Donaldson et al. 2006). If thesereserves are depleted or low before new photosynthetic organsare fully functional, plant growth and survivorship canbe compromised even under resource-rich environments(Landhäusser and Lieffers 2002; Landhäusser et al. 2012a;Sprugel 2002; Stevens et al. 2008; Zhao et al. 2008). In lightof these studies, it is critical to investigate the role of plantreserves in explaining plant growth and defense under herbivorypressure. Such investigations are particularly lacking inlong-lived trees that have large storage organs (stems androots), and are exposed to multiple and chronic biotic stressesthroughout their lifetime (Goodsman et al. 2013; Landhäusserand Lieffers 2002, 2012; Sala et al. 2012).

In this study, we tested whether the interaction betweentrembling aspen (Populus tremuloides Michx.) and its majorinsect defoliator, the forest tent caterpillar (Malacosomadisstria Hübner) (Lepidoptera: Lasiocampidae) is mediatedby reserves accumulated in aspen’s non-photosynthetic organsduring the previous growing season. Aspen is a prominent treespecies in the boreal forests of North America, similarly to itssister species, Populus tremulaL., the European aspen, whichhas similar ecological functions in Europe. Aspen has been afocal species in studies exploring patterns of variation insecondary metabolites and growth with respect to growingconditions and resource availability (Babst et al. 2005, 2008;Bryant et al. 1987; Donaldson et al. 2006; Donaldson andLindroth 2007; Osier and Lindroth 2001). Likewise, studies inP. tremula have established relationships between plantsecondary compounds and insect herbivores (Bernhardssonet al. 2013; Jansen et al. 2009). However, these studies gen-erally focused on foliar chemistry and its effects on insectperformance or investigated interactions between currentsources and sinks, and largely did not account for reservesaccumulated in non-photosynthetic organs during theprevious growing seasons. More recently, Stevens et al.(2008) provided strong evidence that biomass accumulationin previous growing seasons can influence both herbivory andplant responses to herbivore damage, suggesting that biomassaccumulation – and its content – are important features ofsubsequent plant-herbivore interactions.

Therefore, the objective of this study was to investigatewhether reserves such as nitrogen (N) or non-structuralcarbohydrates (NSC) accumulated in non-photosyntheticorgans during the previous growing season can mediatechemistry of aspen foliage, particularly NSC, N, phenolicglycosides, and herbivore performance in the subsequentgrowing season.

Methods and Materials

Seedling Generation and Treatment Applications We gener-ated aspen seedlings with different N and NSC reserve accu-mulations in their stems and roots (hereafter referred to asreserves) using the methods described in Landhäusser et al.(2012a) and Schott et al. (2013) (Table 1). Briefly, we usedseeds collected near Edmonton (Alberta, CA) (53°32′N113°30′W) and sowed them on 29-May-2009 into eightstyroblocks (Beaverplastic, Alberta) with 66 cavities (cavitysize: 5 cm dia×15 cm depth; 220 ml vol.). The plantingsubstrate used was ten parts peat, two parts perlite, and onepart clay particles. Greenhouse conditions were 18:6 h L:Dphotoperiod and 60 % relative humidity at 24 °C during thecourse of seedling growth.

Germination of seeds occurred within 2 d and germinantswere misted with water during the first 2 wk. On 14-June-2009, a single fertilization took place using N-P-K (10-52-10)with chelated micronutrients (Plant Prod Co. ON, Canada) at1 g L−1 concentration. We used fertilizers with high P con-centration to facilitate early establishment of seedlings. From28-June-2009 to 12-July-2009 we fertilized all aspen seed-lings twice with a more balanced fertilizer, N-P-K (15-30-15)with chelated micronutrients at 1 g L−1concentration (PlantProd Co.). In each of two fertilization treatments, a fertilizer(N-P-K: 10-52-10 or 15-30-15) was mixed with 6 L of water,and equally distributed among seedlings (66) in eachstyroblock.

On 15-July-2009, we moved four of eight styroblocks ofseedlings outside the greenhouse, while the remaining fourstayed inside. Seedlings inside the greenhouse experiencedlower light levels (40–50 % less) in addition to different

Table 1 Various treatment combinations produced aspen seedlings withdifferent nitrogen (N) and non-structural carbohydrate (NSC) reserves intheir stems and roots

Fertilizationregimesa

Shoot growthinhibitor

Light levels N and NSC status

High Applied Low High N-Medium NSC

Low Applied Low Low N-Low NSC

High Absent Low High N-Medium NSC

Low Absent Low Low N-Low NSC

High Applied High High N-High NSC

Low Applied High Low N-Low NSC

High Absent High High N-High NSC

Low Absent High Low N-Low NSC

a Seedlings were grown either inside (low light) or outside (high light) thegreenhouse and were subjected to either low or high fertilizationtreatments and either treated with shoot growth inhibitor or not. N=8for of each of eight treatment combinations (location (2)×fertilization(2)×shoot growth inhibitor (2))

22 J Chem Ecol (2014) 40:21–30

temperature, wind, and ambient moisture conditions than didseedlings outdoors. These additional factors were not evalu-ated in the current study. We assigned two of the inside andtwo of the outside styroblocks to either low (0.2 g L−1) or high(2 g L−1) fertilization (N-P-K: 15-30-15) treatments, applied attheir respective concentrations once a week for 4 wk untilAugust 10, as described above. One week after the fertilizertreatments started, we treated seedlings in one styroblock ofeach treatment (location×fertilization) combination once witha shoot growth inhibitor (active ingredient: Paclobutrazol;Bonzi®, Plant Growth Regulator, Syngenta Crop ProtectionCanada, Inc. ON, Canada). This treatment procedure isdescribed in more detail by Landhäusser et al. (2012b).

On 16-Aug-2009, we moved the seedlings inside thegreenhouse to the outside to go through the natural hardeningand dormancy process.When seedlings were dormant bymid-November, we took ten seedlings from each of eight treatmentcombinations (location (2)×fertilization (2)×shoot growthinhibitor (2)) (total 80 seedlings, N=10/treatment combina-tion) to evaluate seedling characteristics as described below.The seedlings remaining in styroblocks were packed in plasticbags, put in wax-coated cardboard boxes, and stored at −4 °Cuntil 12-Apr-2010.

Chemical andGrowthCharacteristics of Dormant Seedling Wemeasured dormant seedlings for shoot height, root collardiameter, root volume, total dry weight of root and shoot,and root-to-shoot ratio. Shoot height was determined by mea-suring the length of a shoot from the root collar to the tip of theterminal bud. Root volume was measured using the waterdisplacement method (Harrington et al. 1994) after removalof planting medium from the roots. Following these measure-ments, the roots and shoots were oven dried at 70 °C for 72 h.Dry weight was recorded for each shoot and root system, androot-to-shoot ratios were calculated.

We ground dried roots and shoots in a Wiley Mill (ThomasScientific Wiley Laboratory Mill, NJ, U.S.A.). Plant materialwas sieved through 40 mesh (0.4 mm) and pooled to obtain a

composite sample of non-photosynthetic tissues per plant.Using this pool, NSC was determined following the methoddescribed by Chow and Landhäusser (2004). Briefly, sugarswere extracted three times from ground tissues using hot 85%ethanol, and then analyzed colorimetrically using phenol-sulphuric acid at 490 nm. Following sugar extraction, thestarch in the remaining residue was digested with α-amylase(ICN 190151, from Bacillus lichenformis) and amyloglucosidase(Sigma A3514, from Aspergillus niger), and glucose equivalentswere determined colorimetrically with peroxidase-glucoseoxidase-o-dianisidine (Sigma Glucose Diagnostic Kit 510A).Non-structural carbohydrates were the sum of water solublesugars and starch. We summed individual sugar and starchcompounds but we used them jointly in our statistical anal-yses as sugars and starches can be transferred from one stateto the other. Furthermore, from an herbivore perspective,sum of both would provide a better reflection of whatherbivores are feeding on during defoliation. Nitrogen wasdetermined using the Kjeldahl method (Bremner andMulvaney 1982). Micro (Mn, Fe, Zn, Al, Cu, Pb) and macro(P, K, Ca, Mg, S) nutrients were measured using the micro-wave digestion / ICP OES method [EPA 3051, Instrumenta-tion: Microwave=CEM MARS Express; ICP OES=SpectroCiros, (Kalra and Maynard 1991)].

We grouped seedlings in terms of their similarity in N andNSC contents across the eight treatment combinations usinglinear discriminant analysis (LDA) (Table 2). LDA incorpo-rated vertical and radial growth, including total dry weight ofstem and roots, and chemical (NSC, N, micro and macronu-trients) characteristics and converged the seedlings into threereserve statuses: Low N-Low NSC, High N-Medium NSC, orHigh N-High NSC (Fig. 1; Table 2). Seedlings in Low-N werecharacterized by low N, low NSC, high NSC/N ratio, lowmicro (Al) and macro (Ca, K, P) nutrients, slow growth, andhad the smallest root volume, root collar diameter, and totaldry weight. Seedlings in High-N were characterized by highN, low NSC/N ratio, high micro (Mn, Fe, Pb, Cu, Zn) andmacro (Mg, S) nutrients, and faster growth. Seedlings in High-

Table 2 Chemical (content) andgrowth characteristics of aspenseedlings used in the greenhousestudy with different nitrogen (N)and non-structure carbohydrate(NSC) reserves

aMeans with the same letterwithin a row are not significantlydifferent at α=0.05. N=32 forLow N-Low NSC, and 16 foreach of the other two reservestatuses

Characteristics of aspenseedlings duringdormancy

Seedling reserve status (Mean±Standard Error)a F P

Low N–LowNSC

High N–MediumNSC

High N–HighNSC

N (mg) 12.2±0.4c 38.1±2.4 b 46.8±2.2 a 253.77 <0.0001

NSC (mg) 640.7±21.2 c 1527.8±120.0 b 2018.4±150.0 a 91.26 <0.0001

NSC/N ratio 53.0±1.3 a 39.5±1.6 b 43.5±2.6 b 17.08 <0.0001

Shoot height (cm) 25.0±1.2 c 37.7±2.4 b 57.3±2.1 a 65.35 <0.0001

Root volume (cm3) 3.9±0.4 b 11.7±1.3 a 9.5±1.0 a 22.38 <0.0001

Root collar diameter (mm) 3.5±0.1 b 5.6±0.2 a 5.3±0.2 a 44.63 <0.0001

Total dry weight (g) 1.9±0.1 b 5.5±0.5 a 5.6±0.4 a 33.49 <0.0001

Root:shoot ratio 2.6±0.2 a 2.7±0.3 a 1.1±0.2 b 41.89 <0.0001

J Chem Ecol (2014) 40:21–30 23

N were further distinguished from one another by their NSCcontent; seedlings in the High NSC had relatively higher NSCthan the seedlings in the Medium NSC. Further, seedlings inHigh N (Medium or High NSC) had the largest root volume,root collar diameter, and total dry weight, and seedlingsin the High N–High NSC reserve status had the lowestroot-to-shoot ratio compared to the seedlings in the other tworeserve statuses.

Defoliation Experiments with Forest Tent Caterpillar afterLeaf Flush On 12-Apr-2010, we selected 64 seedlings inLow N and 64 seedlings in High N (32 in Medium NSC, 32in High NSC) from remaining seedlings. These seedlingswere thawed, potted in 500 ml pots, containing a mix ofpeat:clay:perlite (5:1:1), and placed in a growth chamber.Conditions in the growth chamber were set at 16:8 h L:Dphotoperiod and 60 % RH at 22 °C. Seedlings in each of thethree reserve statuses were divided in two groups: one groupwas randomly assigned to the herbivory treatment while theother served as an untreated control.

Egg masses of forest tent caterpillar were collected from anoutbreak population in Prince George, British Columbia(Canada) by Dr. Staffan Lindgren (University of NorthernBritish Colombia, Prince George). Egg masses weredisinfected in 6 % sodium hypochlorite (Grisdale, 1985) andstored at 4 °C. When buds started to flush, egg masses weretied to the seedlings to mimic the natural synchrony betweenhatching and budbreak. Once egg masses were placed, meshbags were put over individual seedlings to prevent the cater-pillars from escaping. Untreated control seedlings also werebagged. Larval emergence occurred from May 3 to May 13,

2010. At the end of the first instar development, we removed80 % of larvae present on the foliage and kept about 100 2ndinstars per seedling. About 90% of the larvae died before theyreached the 3rd instar in some seedlings. Starting with the 3rdinstar, we kept the same number of larvae (20) on eachseedling. The maximum instar reached was recorded for eachseedling and used as an indicator of treatment effect on larvaldevelopment. One day after the feeding experiment (over 2mo of feeding on the aspen foliage) was terminated (mid July2010), the remaining foliage of all aspen seedlings, includingthose seedlings that were not exposed to herbivore feeding,was collected, placed into paper bags, and stored at -40 °C forthe analysis of phenolic glycosides, N, and NSC. Nitrogen andNSC analyses of leaf tissues were conducted as describedabove.

Extraction and Quantification of Phenolic Glycosides of AspenFoliage We freeze dried foliage for 72 h, ground it inliquid nitrogen, and stored the resulting powder at −40 °Cuntil extraction. For the extraction, ~25 mg of leaf tissuepowder were added to 1.5 ml methanol and placed in anultrasonic bath at 4 °C for 30 min, and then centrifuged at13,000 rpm (12,000g in an Eppendorf microfuge) for15 min. The supernatant was transferred to 1.5 ml glassvials and stored at −20 °C.

Phenolic glycosides were separated on a Thermo ODSHypersil column (250 mm length, 4.6 mm inner diam. and5 μM particle size) mounted on an Alliance 2690 HPLCseparation module (Waters, MA, USA) equipped with a Wa-ters 996 Photodiode Array Detector and an autosampler. Theautosampler was set at 4 °C, and the column temperature at28–30 °C. The mobile phase was water/acetic acid (98:2, v/v)(phase A) and methanol/acetic acid (98:2, v/v) (phase B). At aflow rate of 1 ml min−1, the elution program was as follows(percentages refer to the proportion of phase B): 0–35 % (0–20 min), 35–65 % (20–35 min), 65–80 % (35–36 min), 80–100 % (36–37 min), 100 % (37–38 min), 100–0 % (38–39 min), 0 % (39–40 min). The injection volume was 15 μl,and the scanning range was 200–400 nm, but the 274 nmchannel was used for data monitoring and processing.

Quantification was made possible through the use ofauthentic standards: tremulacin, salicin, and salicortin wereprovided by APIN chemicals Ltd. (Oxfordshire, UK), whiletremuloidin was provided by Dr. Richard Lindroth (Universityof Wisconsin-Madison). Standards were pooled in a concentrat-ed stock solution that was diluted to generate a standard curvewith five points. The standard pools were run three times, amean was recorded and a linear regression line fitted with aresulting R2=0.989 (N=12).

Variables Used in Statistical Analysis We did not conduct anychemical analyses of the non-photosynthetic organs fromaspen seedlings after the herbivory experiment; rather, we

Fig. 1 Linear Discriminant Analysis (LDA) based on the growth andchemical characteristics of aspen seedlings prior to dormancy. indi-cates centroids of different reserve statuses of seedlings. SDW shoot dryweight, RDW root dry weight, SH shoot height, TNK total nitrogencontent, NSC non-structural carbohydrates contents, RCD root collardiameter. Remaining abbreviations represent macro or micro nutrients.Length of vectors indicates relationship between a particular variable anda particular reserve status. Each point represents a seedling. N=32 forLow N-Low NSC, and 16 for the other two reserve statuses

24 J Chem Ecol (2014) 40:21–30

used various seedling characteristics, including chemical (N,NSC, micro and macro nutrient contents) and plant growthbefore dormancy to explain foliar chemistry (N, NSC, andphenolic glycoside concentrations) and forest tent caterpillarperformance after dormancy. Chemicals measured in seed-lings without herbivore feeding were considered constitutive,including N, NSC, and other micro and macro nutrients,whereas those measured in seedlings after herbivory wereconsidered induced chemicals, including N, NSC, and pheno-lic glycosides. From these measurements, we derived thefollowing variables: (a) N reserve, (b) NSC reserve, (c)NSC/N reserve ratio, (d) nutrient reserve (except N), (e) foliarN, (f) foliar NSC, (g) foliar NSC/N ratio, and (h) foliarphenolic glycosides.

Statistical Analyses The effect of reserve (NSC, N, nutrient)and foliar (NSC, N) chemicals was examined on the amountof constitutive and induced phenolic glycosides and caterpillarperformance as well as interactions between reserve and foliarchemicals using Regularized Discriminant Analysis (RDA)[(rdaTest package in R (Legendre and Durand 2010)]. RDA isa non-symmetric method that operates by extracting the com-ponents of the explanatory matrix in a way that makes themclosely correlated with the response matrix and then it repeatsthe same protocol with the components extracted from theresponse matrix, in order to correlate them with the explana-tory matrix. In general, RDA requires two matrices: first, theexplanatory variables matrix, which included, reserve N,reserve NSC, reserve nutrients, and reserve NSC/N ratio;second, the response variable matrix, which included four phe-nolic glycosides (tremulacin, tremuloidin, salicin, salicortin) andfoliar NSC/N ratio. Arrows at a sharp angle (<90°) indicate thattwo or more variables are correlated, arrows at a right angle(=90°) or at an obtuse angle (>90°) indicate that variables areinversely correlated. ANOVAs were applied at α=0.05 usingthe linear models (lm and anova functions in the basic “stats”package fromR) to test for differences across the three classes ofseedlings, and means were compared using LSD.test function

(“agricolae” package in R). The Pearson correlation coefficientwas used to assess the associations between the amounts ofphenolic glycosides and larval performance. To explore theroles of reserve N, reserve NSC, foliar N, foliar NSC, andinduced phenolic glycoside amounts on the larval response,we used multiple linear regression using different explanatoryand response variables.

Results

Influence of Reserves on Following Year Constitutive FoliarChemistry We observed several notable interactions amongreserve chemicals (NSC, N, NSC/N ratio) accumulated duringthe previous growing season and foliar chemicals (NSC, N,phenolic glycosides) of control plants – those unexposed toherbivore feeding – in the subsequent growing season. First,seedlings in High N (Medium or High NSC) reserve statusesyielded foliage with higher amounts of NSC (Tables 2 and 3).Second, the trend was opposite for N where seedlings in LowN-Low NSC produced foliage with higher amounts of N(Tables 2 and 3). Third, seedlings with lower NSC/N reserveratio (High N-Medium or High NSC) produced foliage withhigher NSC/N ratio (Tables 2 and 3). Fourth, seedlings withhigher NSC/N reserve ratio (Low N-Low NSC) had lowerfoliar NSC/N ratio in their foliage (Tables 2 and 3). Finally,RDA showed that amounts of salicin and tremuloidin wereassociatedwith foliar NSC/N ratio (angle between explanatory –foliar NSC/N ratio – and response – salicin and tremuloidin –variables was lower than 90°), while amounts of tremulacin andsalicortin were associated with NSC/N ratio of reserves (anglebetween explanatory –NSC/N ratio of reserves – and response –tremulacin and salicortin – variables was lower than 90°)(Fig. 2a).

Influence of NSC/N Reserve Ratio on Induced PhenolicGlycosides in Foliage When the role of NSC/N reserveratio on induced amounts of phenolic glycosides in foliage

Table 3 Mean (± SE) concentra-tions of constitutive nitrogen (N),non-structural carbohydrate(NSC) reserves and defensechemistry following leaf flush inleaves of aspen seedlings in thecontrol (unexposed to herbivorefeeding) treatment

a Values with the same letter withina row are not significantly differentat α=0.05. N=32 for Low N-LowNSC, and 16 for each of the othertwo reserve statuses

Chemicals Mean concentration (mg/g tissue)±standard errora F P

Low N–LowNSC

High N–MediumNSC

High N–HighNSC

N 1.9±0.1 a 1.3±0.1 b 1.3±0.1 b 16.28 <0.0001

NSC 21.8±0.6 b 30.3±1.1 a 30.9±0.9 a 46.06 <0.0001

NSC/N ratio 12.8±0.8 b 24.4±1.8 a 27.1±2.1 a 36.98 <0.0001

Total phenolic glycosides 218.1±5.4 199.3±10.3 206.2±8.4 1.76 0.18

Salicin 9.6±0.6 a 5.5±0.7 b 4.3±0.4 b 24.04 <0.0001

Salicortin 105.1±4.3 95.9±5.8 106.0±4.9 1.07 0.35

Tremulacin 68.6 ±2.2 68.8±4.0 69.9 ±3.0 0.05 0.95

Tremuloidin 34.8±2.6 29.1 ±3.7 26.0±2.3 2.45 0.09

J Chem Ecol (2014) 40:21–30 25

was examined using RDA, amounts of phenolic glycosideswere related to the NSC/N reserve ratio, not to the foliar NSC/N ratio (angle between response – salicin, tremuloidin,tremulacin, and salicortin – and explanatory – NSC/N ratioof reserves – variables was lower than 90° vs. angle betweenexplanatory – the foliar NSC/N ratio – and the same responsevariables was wider than 90°) (Fig. 2b). Furthermore, seed-lings with higher NSC/N reserve ratio (LowN-LowNSC) hadlower foliar NSC/N ratio, but had higher induced amounts ofphenolic glycosides than those seedlings with lower NSC/Nreserve ratio (High N-Medium or High NSC) (Table 4).

Influence of Reserves and Plant Growth on Foliar PhenolicGlycosides and Herbivore Performance We conducted mul-tiple regression analysis to evaluate the relationship of reservechemistry (NSC, N) and plant growth with foliar phenolicglycosides and found that 88% of the variation in the amountsof foliar phenolic glycosides was explained by the amounts ofreserves (NSC, N) accumulated during the previous growingseason and plant growth traits (F2,61=10.07, P=0.025). How-ever, the relationship between the individual explanatory

variables and amounts of phenolic glycosides was not signif-icant (data not shown). Further investigations on the relation-ship between reserve chemistry and larval performance in amultiple regression analysis indicated that 94% of variation inlarval performance was explained by amounts of NSC and Nstored in the reserves (F2,61=20.52, P=0.007) (Table 5).Larval performance was positively correlated to the amount ofreserve N and negatively correlated to the amount of reserveNSC.

Influence of Foliar Chemistry on Herbivore PerformanceForest tent caterpillar feeding on Low N-Low NSC seedlingsconsistently showed reduced instar development, indicating anegative role of constitutive foliar phenolic glycosides onlarval performance in control seedlings (Fig. 3a). However,we did not observe such consistent negative relationship be-tween larval development and the other two seedling groups(Medium or High NSC) as instar development was morescattered (Fig. 3a). Amounts of induced foliar phenolicglycosides in plants exposed to herbivory did not influencefolivore caterpillar performance (Fig. 3b).

-1 0 1 2

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

Foliar NSC/N

salicortin

tremulacin

tremuloidin

PMn

MgK

Ca

NSC/NReserves

-0.5 -0.25 0 0.25 0.5 0.75 1

-0.5

-0.25

0

0.25

0.5

0.75

salicin

Canonical axis 1 : 43.6%, p= 0.004

Can

on

ical

axi

s 2:

44

%,

p=

0.04

Low N – LowNSCHigh N – Medium NSCHigh N – High NSC

a

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

-1.5

-1.0

-0.5

0.0

0.5

1.0

Foliar NSC/N

salicin

salicortin

tremulacin tremuloidin

PMnMg

K

Ca

NSC/N Reserves

-1 -0.5 0 0.5 1

-1

-0.5

0

0.5

1

Low N – LowNSCHigh N – Medium NSCHigh N – High NSC

Canonical axis 1: 56.01 %, p=0.001

Can

on

ical

axi

s 2:

25

.1

%, p

=0.5

b

Fig. 2 Regularized Discriminant Analysis (RDA) of a the constitutive(unexposed to herbivore feeding) and b the induced (after herbivorefeeding) levels (concentrations) of chemicals in aspen foliage in relation

to the nutrient and carbohydrate contents of seedlings prior to dormancy.Each point represents the mean of eight seedlings. Continuous vectors:response matrix, dashed lines: explanatory matrix

Table 4 Mean (± SE) concentra-tions of induced nitrogen (n), non-structural carbohydrate (NSC)reserves and defense chemistryfollowing leaf flush in leavesof aspen seedlings defoliatedby herbivores

a Values with the same letter withina row are not significantly differentat α=0.05. N=32 for Low N-LowNSC, and 16 for each of theother two reserve statuses

Chemicals Mean concentration (mg/g tissue)±standard errora F P

Low N–LowNSC

High N–MediumNSC

High N–HighNSC

N 1.7±0.1 1.5±0.06 1.5±0.1 2.61 0.08

NSC 20.1±0.5 b 20.7±1.3 b 23.7 ±1.0 a 5.98 0.004

NSC/N ratio 12.7 ±0.6 b 15.5±1.8 a 18.1±1.8 a 5.25 0.008

Total phenolic glycosides 299.8±8.8 a 216.0 ±15.1 b 220.1±15.9 b 17.01 <0.001

Salicin 11.2±0.8 a 6.2±0.5 b 6.4±0.5 b 17.83 <0.0001

Salicortin 133.0±4.7 a 100.1±7.4 b 106.5±7.4 b 9.13 0.0001

Tremulacin 114.8±4.7 a 81.2 ±8.1 b 77.8±8.4 b 11.03 <0.001

Tremuloidin 40.7±3.2 a 28.5±1.9 b 29.5±2.0 b 5.29 0.007

26 J Chem Ecol (2014) 40:21–30

Further investigations on whether caterpillar performance canbe explained as a function of induced amounts of foliarchemistry in a multiple regression analysis found that 87 %of the variation in the average final instar reached was due tothe variation in the amounts of NSC, N and phenolic glyco-sides in foliage (F2,61=8.58, P=0.03) (Table 6). The regres-sion model also indicated that amount of foliar N had apositive, whereas amounts of phenolic glycosides had a neg-ative impact on larval performance. The impact of amount offoliar NSC was not significant. In a different multiple regres-sion model, we also found that the relationship betweeninduced amounts of phenolic glycosides and foliar chemicals(N and NSC) was not significant (df: 2, 61, R2=0.35, P=0.34)(data not shown).

Discussion

The current study provided evidence that the reserves (NSCand N) stored in the stems and roots of aspen seedlings duringthe previous growing season altered the NSC, N, and phenolicglycoside concentrations in the new aspen foliage, and signif-icantly influenced the performance of caterpillars feeding onthe foliage in the subsequent growing season. First, depending

on the levels of reserves accumulated, NSC/N reserve ratioaffected both constitutive and induced chemical responses innew aspen foliage. At the constitutive level, aspen seedlingshad similar amounts of constitutive foliar phenolic glycosides,regardless of their NSC/N reserve ratios. However, whenchallenged with defoliation, the ratio of NSC/N reservesaccumulated during the previous growing season mediatedaspen defense and caterpillar performance. Seedlings with ahigh NSC/N reserve ratio (Low N-Low NSC reserve status)contained higher amounts of induced phenolic glycosides andsupported fewer caterpillars, which in turn did not developpast the second instar. In contrast, seedlings with a lowNSC/Nreserve ratio (High N-Medium or High NSC reserve statuses)produced foliage with a lower amount of induced phenolicglycosides, but the relationship between these seedlings andcaterpillar performance was not significant. Overall, theseresults show that NSC/N ratio prior to dormancy can influencethe amounts of secondary compounds produced by aspen afterdormancy – e.g., by mediating the activity of N-based en-zymes used for the synthesis of carbon-based defenses(Gershenzon 1994) – and thus should be taken into consider-ation when predicting future plant-herbivore interactions.

Table 5 Results of multiple regression analyses of the reserve of non-structural carbohydrates (NSC) and reserve nitrogen (N) of aspenseedlings as predictor variables and larval performance of forest tentcaterpillar as response variable

Coefficients Standarderror

Lower95 %

Upper95 %

t-test Pa

Intercept 0.95 0.34 −0.0002 1.90 2.78 0.05

Reserve N 0.17 0.03 0.081 0.25 5.43 0.005

Reserve NSC −3.42 1.15 −6.610 −0.22 −2.97 0.041

aN=64

1 2 3 4 5

200

250

300

350

Average Caterpillar Instar

To

tal P

hen

olic

Gly

cosi

des

(m

g/ g

DW

)

r= - 0.63 , p= 0.09

Low N–Low NSC High N–Medium NSC High N–High NSC

1 2 3 4

180

190

200

210

220

230

240

Average Caterpillar Instar

r= - 0.81, p= 0.01

Low N–Low NSC High N–Medium NSC High N–High NSC

a b

Fig. 3 Relationship between (a) the constitutive and (b) induced levels (concentrations) of phenolic glycosides (mg per gram dry weight of tissue) inaspen foliage and the average instar achieved by forest tent caterpillars. Each point represents the mean of eight seedlings

Table 6 Results of multiple regression analyses of foliar non-structuralcarbohydrates (NSC), foliar nitrogen (N), and induced foliar phenolicglycosides of aspen seedlings as predictor variables and larvalperformance of forest tent caterpillar as response variable

Coefficients Standarderror

Lower95 %

Upper95 %

t-test Pa

Intercept 3.06 4.37 −9.05 15.18 0.70 0.52

Foliar NSC −0.01 0.15 −0.43 0.42 −0.05 0.96

Foliar N 2.91 0.79 0.73 5.09 3.71 0.02

Phenolicglycosides

−0.02 0.006 −0.04 −0.005 −3.54 0.02

aN=64

J Chem Ecol (2014) 40:21–30 27

However, the role of these secondary compounds in herbivorefeeding needs further investigation as NSC/N ratio prior todormancy can explain only part of herbivore feeding on aspenseedlings (Bryant et al. 1983; Lindroth et al. 2001).

Second, the current study provides evidence for plasticityin the allocation of foliar NSC between defense and growth inthe subsequent growing season, depending on the amounts ofNSC reserves accumulated in the previous growing season.Without herbivores, foliar NSC amount was higher in aspenseedlings with higher amount of NSC reserves than in seed-lings with lower amount of NSC reserves. However, whenseedlings were defoliated, foliar NSC concentrations weresimilar among seedlings, regardless of their prior reservestatuses. Even thoughwe do not know the exact mechanism(s)of NSC reserve allocation between growth and defense afterdormancy in aspen seedling, we suggest two possiblemechanisms that may explain our findings. First, some por-tions of foliar NSC in seedlings with high NSC reserves wereprobably allocated to support growth rather than defense, asthese seedlings had the lowest amount of induced phenolicglycosides, but the highest growth rate. This conclusion issupported by the results of earlier studies in the same systemwhich demonstrated allocation of NSC between non-photosynthetic organs and foliage following above-grounddisturbance (Landhäusser and Lieffers 2002, 2012).

Second, the reduction of foliar NSC after defoliation inseedlings with higher amount of constitutive foliar NSCcould also represent a ‘strategic retreat’ of resources (Babstet al. 2005; Orians et al. 2011). According to this view,plants reallocate their resources from the damaged photo-synthetic organs to non-photosynthetic organs, thusavoiding total loss of these resources to herbivores. Uponapplication of jasmonic acid – a plant hormone known tomimic herbivory – in aspen seedlings, Babst et al. (2005)tracked starch stored in leaves by using 11C and found that thestarch was translocated from the foliage to the stems and roots.We currently do not know whether the same resource retrievalmechanism occurred in our system, or the fate of theseretracted resources, as we did not measure reserves after thedefoliation experiment.

Finally, in the current study, aspen seedlings contained twomajor (tremulacin and salicortin), and two minor (salicin andtremuloidin) phenolic glycosides. Although aspen also con-tains condensed tannins, to date, studies have provided littleevidence to support the role of condensed tannins againstmajor defoliators of aspen, including forest tent caterpillar(Barbehenn et al. 2007; Barbehenn and Constabel 2011;Lindroth and St. Clair 2013). However phenolic glycosidesin aspen – in particular tremulacin and salicortin – have beenshown to reduce feeding, growth and survival of insect herbi-vores (Hemming and Lindroth 2000; Lindroth and St. Clair2013). Supporting the findings of earlier studies, we found aninverse relationship between caterpillar performance and

constitutive phenolic glycoside concentrations in the foliage.However, since induced amounts of phenolic glycosidesalone did not affect caterpillar performance, both N andphenolic glycosides appear to be important in caterpillarfeeding, agreeing with earlier studies, which showed thatforest tent caterpillar can thrive on foliage with highamounts of carbohydrates and nutrients in the presenceof high amounts of secondary metabolites in the samefoliage (Couture et al. 2011; Hwang and Lindroth 1997;Lindroth et al. 1993).

In conclusion, we showed that plant reserves stored in theprevious growing season can mediate plant defense andgrowth, and that reserves contained in non-photosynthetictissues prior to leaf-out should not be viewed simply aspassive transport or storage tissues, but rather as functionalparts of plants when assessing the response of plants to defo-liation stress. Although we do not know how long reserves,stored during the previous growing season, impact seedlingresponses, the growing conditions and the ability of aspen tostore reserves and nitrogen will likely affect the defense capa-bility of aspen in the subsequent growing seasons. As a resultof stress such as drought, which will limit carbon assimilationand root water and with that likely nutrient uptake in aspencarbon and N reserves in tissues will be low during thedormant season (Galvez et al. 2013). This in turn couldinfluence the defense ability of aspen in the following growingseason. Furthermore, generating seedlings with high reservescould be a useful strategy in agroforestry and forest conserva-tion and reclamation, as early seedling mortality due to her-bivory is one of the most difficult challenges in reforestationand afforestation (Macdonald et al. 2012). Aspen seedlingswith high carbohydrate and nutrient reservesmight have bettersurvival ability than seedlings without such reserves right afterplanting (Landhäusser et al. 2012a). Specifically NSC/N ratioin seedling reserves can be used to target specific seedlingcharacteristics. For example, seedlings with high NSC/Nreserve ratio (Low N-Low NSC) can be well defended,but grow slowly while seedlings with low NSC/N ratio(High N-Medium or High NSC) can grow faster and escapefrom competition from other plant species, yet may not bewell defended. Further studies are needed to establish howlong the effect of the stored reserves would last beyond thefirst year.

Acknowledgments We acknowledge the contribution of Dr. StaffanLindgren (University of Northern British Columbia) and Amy Nixon(University of Alberta) in providing egg masses of forest tent caterpillarfor the herbivory experiment. We also thank Dr. Richard Lindroth(University of Wisconsin, Madison) for supplying tremuloidin forphenolic glycoside analysis in HPLC. Dr. J Karst (University ofAlberta) provided comments on the earlier version of this paper.This study was funded by Natural Sciences and Engineering ResearchCouncil of Canada – Discovery to SML and NE. Dr. GuillaumeBlanchet helped with the statistical analysis and Pak Chow conductedthe carbohydrate analysis.

28 J Chem Ecol (2014) 40:21–30

References

Babst BA, Ferrieri RA, Gray DW, Lerdau M, Schlyer DJ, Schueller M,Thorpe MR, Orians CM (2005) Jasmonic acid induces rapid chang-es in carbon transport and partitioning in Populus. New Phytol 167:63–72

Babst BA, Ferrieri RA, Thorpe MR, Orians CM (2008) Lymantria disparherbivory induces rapid changes in carbon transport and partitioningin Populus nigra. Entomol Exp Appl 128:117–125

Barbehenn RV, Constabel CP (2011) Tannins in plant–herbivore interac-tion. Phytochemistry 72:1551–1565

Barbehenn RV, Jones CP, Yip L, Tran L, Constabel CP (2007) Limitedimpact of elevated levels of polyphenol oxidase on tree-feedingcaterpillars: assessing individual plant defenses with transgenic pop-lar. Oecologia 154:129–140

Bernhardsson C, Robinson KM, Abreu IN, Benedicte SJ, Albrectsen BR,Ingvarsson PK (2013) Geographic structure in metabolome andherbivore community co-occurs with genetic structure in plant de-fence genes. Ecol Let 16:791–79

Bonello P, Gordon TR, Herms D, Wood DL, Erbilgin N (2006)Nature and ecological implications of pathogen-induced system-ic resistance in conifers: a novel hypothesis. Phys Mol PlantPath 68:95–104

Bremner, J.M., and Mulvaney, C.S. 1982. Nitrogen-total. In: Page, A.L.(ed) Methods of soil analysis. Part 2. Chemical and microbiologicalproperties. American Society of Agronomy, Madison, Wisconsin,pp 595–624

Bryant JP, Chapin FS, Klein DR (1983) Carbon/nutrient balanceof boreal plants in relation to vertebrate herbivory. Oikos 40:357–368

Bryant JP, Clausen TP, Reichardt PB, McCarthy MC, Werner RA (1987)Effect of nitrogen fertilization upon the secondary chemistry andnutritional value of quaking aspen (Populus tremuloides Michx.)leaves for the large aspen tortrix (Choristoneura conflictana(Walker)). Oecologia 73:513–517

Chow PS, Landhäusser SM (2004) A method for routine measurementsof total sugar and starch content in woody plant tissues. Tree Phys24:1129–1136

Couture JJ, Meehan TD, Lindroth RL (2011) Atmospheric change altersfoliar quality of host trees and performance of two outbreak insectspecies. Oecologia 168:863–876

Donaldson JR, Kruger EL, Lindroth RL (2006) Competition- andresource-mediated trade-offs between growth and defensive chem-istry in trembling aspen (Populus tremuloides). New Phytol 169:561–570

Donaldson JR, Lindroth RL (2007) Genetics, environment, and theirinteraction determine efficacy of chemical defense in tremblingaspen. Ecology 88:729–739

Dunn JP, Potter DA, Kimmerer TW (1990) Carbohydrate reserves, radialgrowth, and mechanisms of resistance of oak trees to phloem-boringinsects. Oecologia 83:458–468

Eyles A, Bonello P, Ganley R, Mohammed C (2010) Inducedresistance to pests and pathogens in trees. New Phytol 185:893–908

Franceschi VR, Krokene P, Christiansen E, Krekling T (2005)Anatomical and chemical defences of conifer bark beetles and otherpests. New Phytol 167:353–375

Frost CJ, Mescher MC, Carlson JE, De Moraes CM (2008) Plant defensepriming against herbivores: getting ready for a different battle. PlantPhysiol 146:818–824

Galvez DA, Landhäusser SM, Tyrée MT (2013) Low root reserve accu-mulation during drought may lead to winter mortality in poplarseedlings. New Phytol 198:139–148

Gershenzon J (1994)Metabolic costs of terpenoid accumulation in higherplants. J Chem Ecol 20:1281–1328

Goodsman DW, Lusebrink I, Landhäusser SM, Erbilgin N, Lieffers VJ(2013) Hierarchies in carbon availability, defense chemistry andsusceptibility to fungal invasion along tree boles. New Phytol 197:586–594

Grisdale D (1985) Malacosoma disstria. In: Singh P, Moore RF (eds)Handbook of insect rearing. Elsevier. Amsterdam, The Netherlands,pp 369–379

Harrington JT, Mexal JD, Fisher JT (1994) Volume displacement pro-vides a quick and accurate way to quantify new root production.Tree Planter’s Notes 45:121–124

Hemming JDC, Lindroth RL (2000) Effects of phenolic glycosides andprotein on gypsy moth (Lepidoptera: Lymantriidae) and forest tentcaterpillar (Lepidoptera: Lasiocampidae) performance and detoxi-cation activities. Environ Entomol 29:1108–1115

Hwang SY, Lindroth RL (1997) Clonal variation in foliar chemistry ofaspen: effects on gypsy moths and forest tent caterpillars. Oecologia111:99–108

Jansen JJ, Allwood JW, Marsden-Edwards E, van der Putten WH,Goodacre R, van Dam NM (2009) Metabolomic analysis of theinteraction between plants and herbivores RID C-3707–2011 RIDB-3281–2008. Metabolomics 5:150–161

Kalra, Y,P., and Maynard, D.G. 1991. Methods manual for forest soil andplant analysis. Information Report NOR-X319

Landhäusser SM, Lieffers VJ (2002) Leaf area renewal, root retention andcarbohydrate reserves in a clonal tree species following above-ground disturbance. J Ecol 90:658–665

Landhäusser SM, Lieffers VJ (2012) Defoliation increases risk of carbonstarvation in root systems of mature aspen. Trees- Struc Func 26:653–661

Landhäusser SM, Pinno BD, Lieffers VJ, Chow PS (2012a) Partitioningof carbon allocation to reserves or growth determines future perfor-mance of aspen seedlings. For Ecol Manag 275:43–51

Landhäusser SM, Rodriguez-Alvarez J, Marenholtz EH, Lieffers VJ(2012b) Effect of stock type characteristics and time of planting onfield performance of aspen (Populus tremuloidesMichx.) seedlingson boreal reclamation sites. New For 43:679–693

Legendre, P., and Durand, S. 2010. rdaTest, Canonical redundancy anal-ysis. R package version 1.7

Lindroth RL, Kinney KK, Platz CL (1993) Responses of deciduous treesto elevated atmospheric CO2: productivity, phytochemistry, andinsect performance. Ecology 74:763–777

Lindroth RL, Roth SK, Nordheim EV (2001) Genotypic variation inresponse of quaking aspen (Populus tremuloides) to atmosphericCO2 enrichment. Oecologia 126:371–379

Lindroth RL, St Clair SB (2013) Adaptations of quaking aspen (PopulustremuloidesMichx.) for defense against herbivores. For Ecol Manag299:14–21

Macdonald, S.E., Quideau, S.A., and Landhäusser, S.M. 2012.Rebuilding boreal forest ecosystems after industrial disturbance.In: Reclamation and restoration of boreal ecosystems: attainingsustainable development. Edited by Dale Vitt and Jag Bhattia.Cambridge University Press, pp 123-160

McDowell N, Pockman WT, Allen CD, Breshears DD, Cobb N, Kolb T,Plaut J, Sperry J,West A,WilliamsDG,Yepez EA (2008)Mechanismsof plant survival and mortality during drought: why do some plantssurvive while others succumb to drought? New Phytol 178:719–739

Orians CM, Thorn A, Gómez S (2011) Herbivore-induced resourcesequestration in plants: why bother? Oecologia 167:1–9

Osier TL, Lindroth RL (2001) Effects of genotype, nutrient availability,and defoliation on aspen phytochemistry and insect performance. JChem Ecol 27:1289–1313

Sala A, Woodruff DR, Meinzer FC (2012) Carbon dynamics in trees:feast or famine? Tree Phys 32:764–775

Sampedro L, Moreira X, Zas R (2011) Costs of constitutive andherbivore-induced chemical defences in pine trees emerge onlyunder low nutrient availability. J Ecol 99:818–827

J Chem Ecol (2014) 40:21–30 29

Schott KM, Pinno BD, Landhäusser SM (2013) Premature shoot growthtermination allows nutrient loading of seedlings with an indetermi-nate growth strategy. New For 44:635–647

Sprugel DG (2002) When branch autonomy fails: Milton’s Lawof resource availability and allocation. Tree Phys 22:1119–1124

Stevens MT, Kruger EL, Lindroth RL (2008) Variation in tolerance toherbivory is mediated by differences in biomass allocation in aspen.Funct Ecol 22:40–47

Zhao W, Chen S, Lin G (2008) Compensatory growth responses toclipping defoliation in Leymus chinensis (Poaceae) under nutrientaddition and water deficient conditions. Plant Ecol 196:85–99

30 J Chem Ecol (2014) 40:21–30