Red Alder Defense Mechanisms Against Western Tent Caterpillar Defoliation

Canadian Journal of Forest Research

Red alder defense mechanisms against western tent caterpillar defoliation

Journal: Canadian Journal of Forest Research

Manuscript ID cjfr-2020-0343.R1

Manuscript Type: Article

Date Submitted by the 19-Sep-2020 Author:

Complete List of Authors: Boateng, Kennedy; University of Victoria, Centre for Forest Biology Hawkins, Barbara; University of Victoria Constabel, Peter; Centre for Forest Biology and Biology Department Yanchuk, Alvin; British Columbia Ministry of Forests and Range Fellenberg,Draft Christin; University of Victoria, Centre for Forest Biology Keyword: red alder, insect resistance, western tent caterpillar, phenolics, oregonin

Is the invited manuscript for consideration in a Special Not applicable (regular submission) Issue? :

Red alder defense mechanisms against western tent caterpillar defoliation

Kennedy Boateng a, B. J. Hawkins a,*, C. Peter Constabel a, Alvin D. Yanchuk b and Christin Fellenberg a

a Centre for Forest Biology and Department of Biology, University of Victoria, 3800 Finnerty Rd., Victoria, BC, Canada V8P 5C2. b BC Ministry of Forests, Lands, Natural Resource Operations and Rural Development, P.O Box 9518 STN PROV GOVT, Victoria, BC, Canada V8W 9C2.

e-mail addresses: [email protected] (Kennedy Boateng), [email protected] (Barbara Hawkins), [email protected] (C. Peter Constabel), [email protected] (Alvin Yanchuk), [email protected] (Christin Fellenberg).

*Corresponding author: Barbara J. Hawkins. e-mail address: [email protected]

Draft

ABSTRACT

Red alder (Alnus rubra Bong.) is a tree of high economic and ecological importance but subject to severe defoliation during episodic outbreaks of tent caterpillars (Malacosoma spp.). We evaluated variation in western tent caterpillar (M. californicum Packard, 1864)

(WTC) resistance among and within red alder populations and clones, and investigated potential defense mechanisms. Bioassay feeding trials were conducted with WTC on 20 red alder clones from 10 provenances (two clones per provenance). Phenology and quality of red alder leaves were analyzed to determine if budburst, leaf chemical content, water content or physical traits are determinants of WTC preference. Alder clones differed in percent leaf area eatenDraft by WTC and in leaf defense traits. The concentrations of total phenolics, condensed tannin and the diarylheptanoid oregonin negatively correlated with the percent leaf area eaten by the caterpillars, and a potential threshold was observed, above which the concentration of each of the chemicals appeared to reduce

WTC feeding. Particularly, foliar concentrations of oregonin greater than 20 % of leaf dry weight were consistently associated with reduced feeding. The effects of oregonin concentration in red alder leaves on tent caterpillar feeding is a novel finding.

Keywords: red alder; Alnus rubra; insect resistance; defense traits; western tent caterpillar; Malacosoma californicum; phenolics; oregonin

1. Introduction

Plants and insects have evolved together for hundreds of millions of years. To

resist insect attack, plants have developed a variety of morphological, biochemical and

molecular defense traits that may be constitutive or induced. This variation in defense

traits among individual plants can lead to differences in the resistance of the plants to

insect attack.

Physical leaf characteristics of deciduous plants, such as thickness, toughness, and

trichome density, have been related to plant defense against herbivory as they can affect

the palatability and digestibility of leaf tissue (Kudo 2003, War et al. 2012,). Thin, tender leaves with high nutritional valueDraft are generally preferred by insect larvae (Gómez et al. 2008). Leaf nutritional value is influenced by traits such as nitrogen (N), sugar and water

content (Hwang and Lindroth, 1997, Haukioja, 2005, Despland and Noseworthy 2006).

Leaf nutrients and water content have a significant influence on the feeding, survival,

growth rate and reproduction of plant herbivores (Yang et al. 2007, Haviola et al. 2012),

and lepidopteran caterpillars have sugar-sensitive chemoreceptors for perceiving and

selecting food with high sugar, especially sucrose, levels (Panzuto et al. 2001, Despland

and Noseworthy 2006, Noseworthy and Despland 2006). The ratio of sugar: protein in

leaves can also correlate with consumption by lepidopteran larvae (Haukioja et al. 2002).

In addition to nutritive compounds, plant defense chemicals influence insect

feeding on leaves. Phenolic compounds, including condensed tannin, hydrolyzable

tannin and phenolic glycosides, are an important group of defense chemicals that can be

poisonous or deter insect herbivores (Mithöfer and Boland 2012, War et al. 2012, 2018).

Concentrations of these chemicals can negatively affect the consumption, development

and fecundity of insect herbivores (Haviola et al. 2012, Lindroth and St. Clair 2013), but their efficacy is very dependent on herbivore gut chemistry, as oxidation of phenolics is often involved in their bioactivity (Appel, 1993). Phenolics are widely distributed in the plant kingdom, and tannins are particularly abundant in forest trees (Mole, 1993).

Plant genotype, developmental stage, and environment, as well as the interactions among these factors, determine the expression of the traits listed above, and thus the quality of plants as food for herbivores (Osier and Lindroth 2006, Lindroth and St. Clair,

2013). For example, trees that have early or late budbreak relative to herbivore emergence are subject to less defoliation than trees in which budbreak is synchronized with larval emergence (DonaldsonDraft and Lindroth 2008). This indicates that leaf developmental stage affects the nutritional quality of leaves via multiple effects on leaf protein, water content, defense compound concentration, and toughness (Ossipov et al.

2001, Riipi et al. 2002, Sarfraz et al. 2013).

Red alder (Alnus rubra) is the only true tree in this genus in North America, and the species has received much attention in recent years because of the ecological benefits provided by its ability to fix atmospheric N2 through a symbiosis with the actinomycete,

Frankia, and an increase in commercial use. Red alder is a pioneer species, forming pure stands on disturbed sites, and is the most common hardwood in the Pacific Northwest

(Harrington 2006). Red alder establishment is favoured by high light levels and exposed mineral soil, and it has rapid early growth but a relatively short lifespan (Harrington

2006). The species is tolerant of wet soils and grows commonly in riparian areas

(Harrington 2006), and it makes a unique contribution to the conifer-dominated forests of the Pacific Northwest in terms of soil improvement, wildlife habitat and productivity of

riparian or disturbed sites (DeBell 2006). Insect pests are not usually a major concern for

red alder, but outbreaks of defoliating insects can cause growth reductions (Harrington

2006). Red alder is attacked and defoliated during tent caterpillar outbreaks, and

successive years of defoliation can cause reduced growth, branch dieback, top-kill, and

mortality in severe cases (US Forest Service 2011). The main defoliator of red alder is the

western tent caterpillar (WTC), an oligophagous, univoltine insect native to the Pacific

Northwest (Sarfraz et al. 2013). Defensive mechanisms of plants against tent caterpillars

have been studied in species such as birch, aspen and poplar (Lindroth and Bloomer

1991; Constabel et al. 2000; Major and Constabel 2006, 2007; Donaldson and Lindroth 2008) but little is known about theseDraft mechanisms in alder. Red alder’s leaf and bark chemical defenses against vertebrate browsing have

been studied (Robbins et al. 1987, McArthur et al. 1993, González-Hernández et al. 2000,

Ballhorn et al. 2017) as have the effects of herbivory on foliage quality for insects

(Williams and Myers 1984, Safraz et al. 2013). Myers and Williams (1984) suggested

that a reduction in red alder foliage quality due to three years of high WTC defoliation

reduced the growth of larvae. Jackrel and Wootten (2015) reported that red alder leaf

consumption by leaf roller caterpillars decreased with increased leaf C: N ratio. Red alder

leaves contain substantial amounts of condensed tannins and other phenolic compounds

that significantly reduce digestibility of foliar protein for deer (Robbins et al. 1987,

McArthur et al. 1993), but we have little understanding of how red alder defends itself

against tent caterpillars. The objective of this study was to take an integrated approach

and investigate red alder genetics, foliar chemistry and physical defenses, concurrently, to

determine the variation in resistance to WTC among and within red alder populations, and to explore defense mechanisms of red alder against this lepidopteran pest.

2. Materials and methods

2.1. Red alder clone selection

This study was motivated by observations in the summer of 2013 of apparent genetic variation in the resistance of red alder to defoliation by WTC in a 19-year-old provenance-progeny test planted near Bowser, British Columbia (49° 29′ N, 124° 40′ W,

50 m asl). Because the trees at the Bowser site were too tall to be suitable for a study of defense mechanisms in the field, clones from a red alder clone bank established in 2006 at the Cowichan Lake Research DraftStation (Mesachie Lake, BC, Canada, 48.83o N, 124.14o, 174 m) were used for this study. The clone bank contained several individuals each, of clones from 100 of the high breeding value red alder families from the Bowser provenance-progeny test site. Two clones from each of ten provenances (twenty clones in total) were selected (Table 1) from trees that exhibited a range of damage by WTC at the

Bowser site. Three individual trees (ramets) per clone were sampled for this study at the

Cowichan Lake Research Station (60 trees in total).

2.2. Bioassay feeding trials

To determine if significant differences existed among the 20 clones of red alder in resistance to defoliation by WTC, choice and no-choice bioassay WTC larvae feeding trials were first conducted with leaves from the 60 selected trees. Overwintered WTC egg masses were collected from red alder trees on Quadra Island, B.C. (50o 11' N, 125o 15' W,

10 m) in February of 2014 and transported to the University of Victoria where they were

kept in a cold room at 4-5 oC. Egg masses were moved to the greenhouse on March 24

and on April 11, 2014 and after hatching, the larvae were reared in separate batches on

red alder leaves in the greenhouse under a 16 hr photoperiod and an average temperature

of 25 oC until the third instar.

Feeding trials began on April 15, 2014 with second instar larvae from the March

24 hatch. Sixty trees from the 20 clones (3 ramets/clone) were sampled from the clone

bank. Three branches were collected from a mid-canopy, south-facing position on each of

the selected ramets (N=180). The collected branches were placed in water in plastic bags

to keep leaves fresh. The branches were stored in a cold room at 4-5 oC overnight before being used the next morning for Draftthe feeding experiments. The feeding trials were conducted in a greenhouse and a laboratory (University of Victoria, Victoria, British

Columbia, Canada) (48.46o N, 123.31o W, 60 m).

In the greenhouse experiment (choice bioassay), a twig (branchlet) with three

leaves of similar size was cut from each of the three collected branches per tree, and

placed in a 500 mL vial filled with water. The samples were randomly arranged in nine, 1

m2 mesh cages (20 vials per cage) with the restriction that no more than two twigs from

the same clone were placed in the same cage to avoid the confounding effect of cage.

Three third or second instar WTC larvae were placed on the leaves of each twig to feed

for 48 h. The larvae were free to move within the cage. At the end of the feeding period,

the percent leaf area eaten for each twig was visually assessed in five percent increments.

The experiment was repeated on April 28 with second instar larvae from the April 11

hatch, and again with on May 15 and May 28, 2014 with third instar larvae from the

March 24 and April 11 hatches, respectively.

On the same dates, a laboratory (no-choice bioassay) feeding experiment was conducted. Four, 1.8 cm diameter leaf disks per branch were placed on wet Whatman filter paper in a labeled 10 cm diameter Petri dish and the Petri dishes (N=180) were randomly arranged on a laboratory bench, with a limitation that dishes from the same tree could not be clustered together. Three second or one third instar WTC larvae were placed in each dish to feed for 48 h. Subsequently, the leaf disks were scanned with an Epson perfection v750 scanner (Epson Canada Ltd, Markham ON, Canada) and the percent leaf area eaten was calculated using WinRHIZO Pro version software (Regent Instruments

Inc., QC, Canada). This experiment was repeated on the same four dates with the same branches as the greenhouse experiment.Draft Five clones with the highest and five clones with the lowest mean percent leaf area eaten in 2014 were selected for further bioassay feeding trials in 2015. The 2014 protocol for the choice feeding experiment in the greenhouse was used on May 6 and

May 20, 2015 to test for consistency in the ranking of clones by percent leaf area eaten.

Leaves were collected from four branches per tree, thus a total of 120 samples (10 clones x 3 ramets x 4 branches) were used for each feeding trial in 2015, with second instar and third instar larvae used on May 6 and May 20, respectively.

2.3. Leaf traits and phenology measurement of red alder clones

A suite of phenological, physical, and chemical measurements shown to affect feeding or performance of leaf herbivores were made on leaves collected from the trees used for the bioassay feeding trials on the same dates as the feeding trials.

2.3.1. Budburst measurement

To test the relationship of bud phenology to feeding by WTC, average

phenological stage of the entire trees sampled for the feeding trials was monitored on

March 25 and April 9, 2014, and April 2, 2015. Bud burst was classified into seven stages

(0 = buds fully closed, 1 = buds swollen, 2 = buds swollen and leaf tips exposed, 2.5 =

buds partially opened and half of leaves exposed, 3 = buds fully opened and leaves fully

exposed but not unfolded, 4 = leaves fully exposed and unfolded, and 5 = leaves fully

expanded).

2.3.2. Leaf thickness, toughness, and trichome measurements

Leaf thickness of three leaves from each branch sampled for the bioassay feeding trials was measured directly fromDraft hand-sections of leaves using a compound microscope on the four test dates in 2014 and two dates in 2015 (Smith et al. 2006). Leaf thickness

was measured at the mid-point between the leaf base and tip, and between the central

vein and the leaf margin (Corlissen et al. 2003). Leaf toughness was assessed by

measuring the force (g mm-2) needed to punch a hole in each leaf blade (Feeny 1970). A

1.8 cm diameter leaf disk of each leaf used for thickness measurements (three leaves per

branch) was cut from half-way between the leaf base and tip, midway between the central

vein and the leaf margin for the toughness measurement. The leaf disks were also

observed under a dissecting microscope to estimate the number of trichomes on both the

adaxial and abaxial surfaces. The numbers of trichomes were categorized into few (<50

per disk), medium (50-100 per disk), and many (>100 per disk).

2.3.3. Leaf defense chemical analyses

Concentrations of red alder foliar condensed tannin (CT) were measured in leaves

collected on April 29 and May 28, 2014. The foliar CT analysis was repeated on leaves

collected on May 6 and 20, 2015 in conjunction with measurements of total phenolic

(TP) and oregonin (Ore) concentrations. The concentration of Ore, a diarylheptanoid, was assessed as this compound is a major phenolic of red alder with demonstrated antimicrobial activity (Saxena et al. 1995, Dahija et al. 2014). Four or five leaves were sampled from nodes two and three behind the tip of each collected branch. Each tree had three and four branches sampled in 2014 and 2015, respectively. The midrib and large veins of all the leaves were excised and discarded. The remaining tissues were pooled by branch and immediately flash-frozen in liquid nitrogen. Tissue samples were freeze-dried for 72 h and stored at -20oC until analysis. To extract phenolic phytochemicals,Draft a 20-25 mg sample of leaf tissue was steeped in 1.5 mL of 80% HPLC-grade methanol (MeOH) with 4 steel ball bearings (2 mm diameter) and homogenized using a Precellys 24 homogenizer (Bertin Technologies,

Toulouse, France) at 2 x 45 s, 5500 rpm. Samples were then sonicated using a VWR model 75 T ultrasonic bath (VWR, Mississauga, ON, Canada) for 10 min at a frequency of 40 kHz and centrifuged at 1500 rpm for 5 min. The supernatant was removed and the tissue sample was re-extracted two more times in 1 mL 80% MeOH. The supernatants

(3.5 mL) were combined and stored at -20oC.

Total phenolic (TP) concentrations in leaf samples were analyzed and quantified with the Folin-Ciocalteu assay (Peters and Constabel 2002). Leaf sample extracts were centrifuged and 20 µL of each extract was removed and placed into a microcentrifuge tube. Then, 100 µL of Folin-Ciocalteau's reagent, 500 µL of 20% sodium carbonate solution (20 g anhydrous Na2CO3 in 100 mL ddH2O) and 380 µL ddH2O were added to each sample and mixed. After 20 min, or when solutions were no longer turbid, 200 µL

of each solution was removed and placed into wells of a 96 well plate. Eighty percent

(80%) MeOH was used as a blank and absorbance at 735 nm was read using a VictorTM

X5 2030 multilabel reader (PerkinElmer, Waltham, MA, USA). Absorbance values were

compared to a standard curve made with authentic Ore (Sigma, Aldrich) to estimate the

TP concentration of each sample in µg mg-1 dry weight.

Foliar condensed tannin (CT) was analyzed in all samples with the butanol-HCl

assay (Porter et al. 1986). First, 400 µL of extract from each leaf sample was mixed with

2 mL of 95% butanol, 5% HCl and 67 µL Fe reagent. A 200 µL aliquot of the reagent-

sample mixture was removed and placed into wells of a 96-well plate as the unheated control. The remaining sample wasDraft heated in a water bath at 95oC for 40 min and allowed to cool at room temperature for 15 min. Next, 200 µL from each heated sample was

removed and placed into the 96-well plate. Absorbance at 550 nm of each heated and

unheated sample was recorded using a VictorTM X5 2030 multilabel reader

(PerkinElmer, Waltham, MA, USA) and unheated control values were subtracted from

heated samples to correct for background. CT concentration was calculated using a

standard curve made with purified trembling aspen CT (Mellway et al. 2009).

For Ore analysis, 20 µL of the methanolic extracts from above were loaded onto a

Phenomenex Luna C18 column (250 x 4.6 mm, particle size 5 µm) and separated using a

Beckman System Gold® HPLC equipped with a 126 solvent module, a model 168 diode

array detector fitted with a deuterium UV lamp with a wavelength range of 190-600 nm

and a Beckman 508 auto-sampler (Beckman Coulter, Mississauga, ON, Canada). Eluent

A and B (water and methanol) were each acidified with 4% formic acid. Elution was

performed using a linear gradient 30% (v v-1) B in A to 50% (v v-1) B in A within 35 min

with detection at 280 nm and a flow rate of 1 mL min -1. Ore was identified and its concentration in each sample was calculated in µg mg-1 dry weight by retention times and

UV spectra based on an authentic standard (Sigma, Aldrich).

2.2.3.4. Nutritional elements and water measurement

Carbon, N, macro- and micronutrient concentrations, and water content were measured in leaves collected on April 29 and May 28, 2014 and on May 6 and 20, 2015.

Four to five leaves were harvested from the branches collected for the feeding trials, bulked by branch and oven dried at 65oC for 48 h. The dried leaves were ground to a powder and stored at room temperature until analysis. For leaf N and C concentrations, 8 mg of leaf powder was packed inDraft a 10 x 10 mm tin capsule (Elemental Microanalysis Ltd. Cambridge, UK), and N and C concentration measured using a FlashEA 1112 Series

NC analyzer (Thermo Finnigan, San Jose, CA, USA). Leaf percent moisture content was measured using the difference in leaf fresh mass and dry mass, divided by dry mass. Leaf macro- and micronutrients, except N and C, were analysed at the B.C. Ministry of

Environment Analytical Laboratory in Victoria.

Leaf sugar content was measured in the same leaves used for CT analysis in 2014 using an enzymatic assay by Campbell et al. (1999) with slight modifications. For sugar extraction, 10 mg of lyophilized leaf tissue was steeped in 1 mL of 80% ethanol with 4 steel ball bearings (2 mm in diameter) and homogenized using a Precellys 24 homogenizer (Bertin technologies, Toulouse, France) at 2 x 45 s, 5500 rpm. The samples were well mixed by vortexing twice for 5 s, followed by incubation in water at 75oC for 4 h. After incubation, the samples were mixed and centrifuged for 2 min at 10,000 rpm.

The supernatant was removed and the extraction was repeated in 1 mL of 80% ethanol.

First and second extracts were combined and speed-vacuumed to less than 20% volume

to remove ethanol. Then dH2O was added to bring extract volume up to 200 µL for each

sample.

Glucose, fructose, and sucrose assays were conducted following the protocol of

Campbell et al. (1999) with slight modifications. A buffer, ATP (adenosine triphosphate)

solution, and NAD (nicotinamide adenine dinucleotide) reagent mixture were prepared

for one 96 well plate. The buffer contained 50 mM triethanolamine, 5 mM MgSO4, 0.02%

BSA (bovine serum albumin), and 0.5 mM DTT (dithiothreitol) with pH adjusted to 7.6

with HCl. ATP solution was made with 300 mg ATP disodium salt and 300 mg Na2CO3

dissolved in 6 mL of ddH2O. NADDraft reagent mixture was made by dissolving 15.3 mg of

NAD in 7 mL of buffer, followed by addition of 14 mL ddH2O, 700 µL of ATP solution,

33 µL (40 U) of HK (hexokinase) enzymes (Sigma-H5625), and 3.5 µL (21 U) of

G6PDH (Glucose-6-phospate dehydrogenase) enzymes (Sigma-G8404).

For glucose analysis, 5 µL of sample extract was removed and placed into a well

of a 96 well plate. 200 µL NAD reagent mixture was added to each well and incubated at

room temperature for 15 min. Absorbance of each sample was measured at 355 nm using

a VictorTM X5 2030 multilabel reader (PerkinElmer, Waltham, MA, USA) to determine

production of NADH (nicotinamide adenine dinucleotide hydrogen).

After completion of the glucose assay, the fructose assay was initiated by adding

5 µL of PGluI (phosphoglucoisomerase) reagent mixture to each sample in the 96 well

plate. The PGluI mixture contained 20 µL (110 U) of PGluI (Sigma-P5381) dissolved in

500 µL ddH2O. Samples were then incubated at room temperature for 15 min, followed

by absorbance reading at 355 nm to determine total NADH. Absorbance values were then

compared to standard curves made with D-glucose and D-fructose for glucose and

fructose, respectively, to estimate the concentrations for each sample in mg mL-1 dry weight. The actual fructose concentration was calculated as the difference between glucose concentration and the estimated fructose concentration for each sample.

For sucrose analysis, enough enzyme solution for one 96 well plate was prepared by dissolving 11.3 mg of invertase (Sigma-I4504) in 1 mL of ddH2O (4000 U enzyme

-1 units mL ), followed by addition of 3 mL of glucose buffer and 6 mL of dH2O (=10 mL at 400 U mL-1). Then, 10 µL of each sample was placed into a microtitre well and 100 µL of enzyme reagent mixture was added to each sample and mixed. Samples were incubated at room temperature forDraft 15 min. After incubation, 100 µL of a reagent mixture containing 15.3 mg of NAD, 3.5 mL of glucose buffer, 7 mL of dH2O, 700 µL of ATP solution, 33 µL (40 U) of HK and 3.5 µL (21 U) of G6PDH was added to each sample and mixed well. The samples were then incubated for 15 min at room temperature and absorbance at 355 nm was read using a multilabel plate reader to determine production of

NADH. Absorbance values were compared to standard curves made with pure sucrose to estimate sucrose concentration of each sample in mg mL-1 dry weight. Since sucrose is hydrolysed to D-glucose and D-fructose with the enzyme invertase, actual sucrose concentration was obtained by subtracting the value of glucose concentration from the estimated sucrose concentration for each sample.

2.4. Data and statistical analysis

Data from the 2014 and 2015 bioassay feeding trials and red alder leaf trait measurements were analysed by ANOVA for all clones, provenances, sampling dates, and environments, by year (Tables 2 and 3). Provenance, clone, date, and environment

were treated as fixed effects, while individual tree (ramet) was treated as a random effect.

All data met assumptions of normality and homogeneity of variance, which were

evaluated using residual plots and F-tests, respectively. Respective error terms (Tables 2

and 3) were used to examine the significance of the effects of date, provenance, clone

nested within provenance, and their interactions on percent leaf area eaten and the

measured leaf traits. Least significant difference (LSD) tests were performed to determine

differences among the means of leaf damage and leaf traits for the individual red alder

provenances and clones. Bioassay feeding trial results in 2014 were analyzed for all 20

clones (Table 2A). From this analysis, five clones with the highest and five clones with the lowest percent leaf area eatenDraft were selected for analysis of leaf defense traits in 2014 (Table 2B), and for analysis of leaf damage and defense traits in 2015 (Table 3).

Regression and Pearson correlation was also performed to determine correlations

among the individual leaf traits by using the measures for individual trees. The general

linear model (PROC GLM), regression (PROC REG), and correlation (PROC CORR)

procedures of SAS 9.3 version (SAS Institute Inc., Cary NC, USA) were used to perform

ANOVA, regression and correlation analyses, respectively. Statistical significance was

determined using α = 0.05. Threshhold concentrations of CT, TP and Ore, above which

little feeding damage was observed, were determined using classification and regression

tree analysis (CART). The tree function in R version 3.0.3 was used for the CART

analysis.

3. Results

3.1. Leaf consumption by WTC in 2014 and 2015

Analysis of data from all 2014 greenhouse and laboratory experiments showed leaf area eaten by WTC varied significantly among the selected red alder provenances (Tables 4 and 5). On average, provenances P9, P11 and P50 had significantly less damage than provenances P51 and P55 in 2014 (F = 9, 1321 = 2.57, p = 0.0062) (Table 4). LSD tests showed leaf area eaten by WTC larvae also differed among the selected clones in 2014

(Table 4), with C9, C38, C12, and C79 being highly eaten; and C18, C34, C90, and C133 being least eaten. The difference in damage between clones nested within provenance was not significant (Table 5). Also, there was no significant variation in leaf area eaten among individual trees (ramets) Draftfrom the same clones.

Percent leaf damage differed significantly among the four test dates in 2014 and between the two feeding environments (Table 5). On average, the caterpillars consumed more leaf area early in the season on April 15 (26.3 + 1.6%) when leaves were still small, than on April 28 (19.2 + 0.8%), May 15 (20.5 + 0.7%) or May 28 (23.1 + 1.5%).

Caterpillars damaged a substantially greater area of whole leaves during assays carried out in the greenhouse, compared to leaf discs in the laboratory. On average, about 6% more leaf area was eaten in the greenhouse (25.2 ± 0.9%) than in the laboratory (19.2 ±

0.8%).

.A repeat of the feeding experiment in 2015 with the five highest and five lowest damage clones assessed in 2014 again showed that leaf area eaten by WTC varied significantly among clones (Tables 4 and 6). The results were similar to those in 2014, with clones C9,

C38, and C162 eaten more; and clones C34, C90, and C124 eaten less (Table 4). Mean

damage of the 10 clones analyzed in both years was significantly correlated between the

two years (p < 0.01, R = 0.6). In 2015, mean leaf damage on May 20 (47.9 ± 2.3%) was

on average greater than damage on May 6 (9.4 ± 0.6%), but the clone x date interaction

term was significant (Table 6) because ranking of clones C38 and C90 fluctuated

between the two sampling dates.

3.2. Leaf traits and phenology

Leaf thickness and time of leaf bud burst (BB) varied significantly among the selected

red alder clones within each provenance in 2014 but the effect of provenance was not significant (Table 5). Leaves of C162,Draft C125, and C90 were thicker than leaves of C115, C133, C34, C12, and C9 (Table 7). Clones C12, C133, and C115 had the most advanced

bud burst, while C9 and C34 had the latest bud burst (Table 7). There was no significant

effect of date or date x clone(provenance) interactions on any of the physical leaf traits

measured in 2014 (Table 5).

In 2015, mean leaf thickness, leaf toughness, and stage of bud burst varied

significantly among clones (Table 6). In 2015, bud burst occurred earliest in clones C79

and C12; and latest in C18, C34, and C162 (Table 8). On average, clone C133 had thick

leaves and clone C38 had tough leaves, while clones C9 and C12 had thin, soft leaves

(Table 8). Leaf thickness increased with time, but there were no significant differences in

mean leaf toughness or stage of bud burst between the two sampling dates (Table 6). Bud

burst values were negatively correlated with both leaf thickness (R = -0.24, p = 0.003)

and toughness (R = -0.2, p = 0.003) indicating that leaves that developed early in the

spring were thin and soft. Leaf thickness and toughness were positively correlated (R =

0.44, p < 0.0001).

3.3. Leaf chemistry

CT concentration in leaves varied significantly between the clones within provenance in

2014, but did not differ among provenances or among dates (Table 5). In 2015, leaf CT

concentrations again varied significantly among clones (Table 6). C90 had the highest CT

concentration, while C38 had among the lowest CT concentrations in both years (Tables

7 and 8). Ore was the dominant phenolicDraft compound detected by HPLC in the red alder leaves, comprising 15.7 – 38.7% of foliar TPs, on average. Mean Ore concentration

varied significantly among clones in 2015 (Table 6), and was highest in C79, C90, and

C124; and lowest in C12, C34, and C38 in 2015 (Table 8). TP concentration also varied

significantly among clones in 2015 (Table 6). Clone C90 again had the highest TP

concentration, while clones C9, C34, and C162 had the lowest TP concentrations (Table

8).

The concentrations of CT, Ore, and TP were significantly different between the

two sampling dates in 2015. Mean concentrations on May 6 for CT (17.8 ± 0.8 µg mg-1),

Ore (143.4 ± 6.4 µg mg-1), and TP (514 ± 5.4 µg mg-1) were higher than the concentrations on May 20, (11± 0.4 µg mg-1, 118.7 ± 4.7 µg mg-1, and 475 ± 3.8 µg mg-

1, respectively).

Concentrations of the measured leaf phenolic compounds were strongly and positively correlated with each other (r = 0.49 to 0.52, p < 0.0001), and were all

negatively correlated with foliar N concentration (r = -0.31 to -0.42, p < 0.0001). CT and

Ore concentrations also had significant positive correlations with foliar C concentration (r

= 0.47 and 0.48, respectively, p < 0.0001).

In 2014, leaf N, Ca, and Cu concentrations varied significantly among the alder

clones within provenances but there was no significant variation among provenances

(Table 5). N

concentration was lowest in clones C90, C125, and C133 (Table 7). Clone C12 had a

higher Ca concentration than C38 and C115, while clones C90 and C125 had the highest

Cu concentrations (Table 7). Fructose and glucose concentrations also varied among clones within provenances but thereDraft was no significant provenance effect (Table 5). Sucrose concentration did not differ significantly between clones within provenance or at

the provenance level (Table 5). Fructose and glucose concentrations were highest in

clone C38 (Table 7). Glucose concentrations were higher on April 29 (20.92 mg ml-1)

than on May 28, 2014 (16.85 mg ml-1) while sucrose concentrations were highest in May

(7.1 mg ml-1 on April 29 and 10.4 mg ml-1 on May 28) (Table 7). There was no

significant effect of provenance or clone within provenance on leaf moisture content, or

on the concentrations of C, P, K, S, Mn, Na, Fe or Zn in 2014.

The 2015 leaf chemical analyses showed significant variation in N, C, Ca, and Cu

concentrations among the ten selected clones (Table 6). N concentration was highest in

clones C34 and C162, and lowest in C9 and C12 (Table 8). C concentration was highest

in C90 and C162 (Table 8). Foliar Ca concentration was highest in C90, while clones

C124, C38, and C90 had the highest foliar Cu concentrations (Table 8).

Sampling date had a significant effect on leaf moisture content, and N and C concentrations in 2015 (Table 6). Mean leaf moisture content and N concentration were higher on May 6 (65.36 ± 1.67% and 2.8 ± 0.05%, respectively) than May 20 (62.21 ±

1.53% and 2.4 ± 0.02 %, respectively). However, C concentration was higher on May 20

(50.23 ± 0.33 %) than May 6 (49.9 ± 0.28 %).

3.4. Correlation between measured leaf traits, phenology, and WTC feeding

Although the percent leaf area eaten and eight of the measured leaf traits varied significantly among clones within provenances in 2014, no individual leaf trait was significantly correlated with the percent leaf area eaten by WTC. In 2015, most measured leaf traits varied significantly amongDraft clones (Table 6) but only the concentrations of foliar phenolic compounds were significantly correlated with leaf area eaten by WTC.

The concentrations of phenolic compounds were negatively correlated with percent leaf area eaten, with 17% of the total variation accounted for by CT concentration (r = -0.41, p < 0.0001), 19% of variation accounted for by TP concentration (r = -0.44, p < 0.0001) and 8% of the variation in damage accounted for by Ore concentration (r = -0.38, p =

0.0003). Highly damaged clones such as C38, C162, and C12 consistently had low concentrations of CT, Ore, and TP, while less damaged clones such as C79, C90, and

C124 had high concentrations of these chemicals (Table 8). Plotting leaf area eaten against the concentrations of foliar phenolic compounds shows a threshold above which the concentration of each of the three chemicals was associated with less WTC damage

(Fig. 1). Leaf consumption by WTC larvae was less when mean concentrations of CT,

TP, and Ore were above 18 + 0.48, 515 + 3.5, and 200 + 4.0 µg mg-1, respectively.

4. Discussion

4.1. Genetic variation in red alder leaf traits, and consumption by WTC

Leaf area eaten by WTC varied significantly among the selected red alder clones and

provenances but the clone x date x environment interactions were not significant,

indicating that WTC preference for individual clones was generally consistent across

dates and experimental environments. Plant genotype can affect plant defenses through

its influence on chemical and physical traits (Mansfield et al. 1999, Osier and Lindroth

2006, Lindroth and St. Clair 2013, Sarfraz et al. 2013). Genetic variation in red alder in

defense against tent caterpillars has not been studied previously, but significant differences in forest tent caterpillarDraft (M. disstria) larval preference and performance have been shown among trembling aspen (Populus tremuloides) clones in bioassay feeding

trials (Hwang and Lindroth 1997, Robison and Raffa 1994) and in the field (Donaldson

and Lindroth, 2008). Osier et al. (2000) found variation among aspen clones in leaf

consumption by gypsy moth larvae (Lymantria dispar) and mountain birch families were

also shown to vary in damage by autumnal moth larvae (Epirrita autumnata) (Haviola et

al., 2012). The red alder clones in our study were of the same age, grown in a common

environment, and the sampled leaves were from the same season, strongly indicating that

variation in genotype contributed to the consistent, significant among-clone variation

observed in WTC feeding preference for red alder.

WTC feeding on leaves in mesh cages in the greenhouse was significantly greater

than in the petri dishes of the lab assay. Caterpillars were observed to be much more

active in the cages and to exhibit gregarious behaviour. With 60 WTC in each cage and

three WTC per petri dish, clustering of WTC on twigs in the cages resulted in higher rates of feeding damage.

Most measured leaf traits in red alder varied significantly among the selected clones but only CT, Ore, and TP concentrations were significantly correlated with leaf area eaten by WTC in the 2015 experiment. CT is a minor component of the TP profile, which may explain the absence of a significant correlation with leaf damage in 2014, although the same trends were evident. The 2015 results show a threshold above which the concentration of each of the three chemicals, particularly CT and Ore, appeared to reduce WTC feeding. Significant genetic variation in concentrations of foliar phenolic compounds has been reported inDraft other deciduous plant species (Hwang and Lindroth 1997; Keinanen et al. 1999; Haviola et al. 2012, 2006; Laitinen et al. 2002), and high concentrations of phenolic compounds are thought to increase plant resistance to insect herbivores by serving as a deterrent, reducing digestibility or through toxic effects

(Mithöfer and Boland 2012). In red alder, leaf roller caterpillars and black slugs consumed less of leaves with high CT, lignin, and C: N content, and preferred leaves with high crude protein content (Jackrel et al. 2015, Ballhorn et al. 2017). Other studies of the effects of phenolic compounds on caterpillar feeding and performance have had variable results, depending on the host species. For example, birch tannins decreased leaf consumption rates by tussock moth larvae by 20% and negatively influenced autumnal moth leaf consumption (Haukioja et al. 2002, Kopper et al. 2002), while gypsy moth consumption was positively related to CT concentration but negatively related to phenolic glycosides (Osier et al. 2000). Biological effects are specific to individual phenolic compounds and likely insects as well.

The effects of foliar Ore concentration on insect herbivory have never been

reported. Our finding of a negative relationship between leaf Ore concentration and WTC

feeding is an intriguing indication that foliar Ore may be a strong defensive compound

against WTC in red alder. Ore, a diarylheptanoid glycoside, was first detected in the bark

and wood of alder by Karchesy and Laver (1974) and is a dominant phenolic constituent

in alder leaves (González-Hernández et al. 2000, Dahija et al. 2014, Jackrel et al. 2016).

Mean foliar Ore content in this study was similar to that reported for red alder leaves

previously (9.64 ± 1.0 %) (González-Hernández et al. 2000). This compound constituted

up to 19% of dry weight and up to 39% of leaf TPs in our study. Ore may have herbivore defense properties that reduce leafDraft digestibility because of its structural similarity to platyphylloside (González-Hernández et al. 2000). Diarylheptanoids in alder leaves,

especially Ore, have also been reported to have antimicrobial activities (Saxena et al.

1995, Dahija et al. 2014, Jackrel et al. 2016) but there has been no report of their effect

on insect feeding. Interestingly, diarylheptanoids found in Alnus species are reported to

act as antioxidants and Alnus leaf extracts are used in many traditional medicines for their

anti-inflammatory, antitumour, antiobesity, and antioxidant effects (Park et al. 2010, Sati

et al. 2011). Reduced consumption of leaves with high concentrations of CT, Ore, and TP

by WTC larvae is likely due to the deterrent characteristics or toxicity of these chemicals.

In contrast to phenolic constituents, other measured leaf traits varied consistently

among clones but were not correlated with WTC damage. We believe we are the first to

report on clonal variation in glucose, sucrose, and fructose concentrations in red alder;

however, we observed no significant relationship between WTC feeding and any of the

three sugars analyzed or leaf sugar: N ratio. In mountain birch leaves, total concentrations

of sugars peaked in July at 155 mg g-1 of leaf dry mass (Riipi et al., 2002). By contrast, the total concentration of sugars in alder leaves sampled in spring for this study was only

34 mg g-1. The sugar concentration in alder clones may be too low to affect WTC feeding; however, it was interesting to note that two of the clones with the highest feeding damage(C9 and C38) also had the highest sucrose concentrations. Our results suggest that while leaf nutritional status may play a significant role in insect herbivore performance measures such as growth, survival and reproduction (Hemming and

Lindroth 1995, Haviola et al. 2012), it may not be the best indicator for predicting leaf preference and consumption rates.

4.2. Date variation in red alderDraft leaf traits and consumption by WTC The percent leaf area eaten on the first assessment in 2014 was highest among the four feeding dates, which may be explained by the small leaf area and leaf tenderness early in the growing season. In 2015, more leaf area was eaten on May 28 than May 6, which may be attributed to the high caterpillar mortality that occurred during the May 6 feeding experiment.

Young and medium-age leaves have higher nutritional values than older leaves

(Radwan et al. 1978, DeBell and Radwan 1994, Ballhorn et al. 2017), and concentrations of most mineral elements and moisture content declined as the alder leaves matured, although C and sucrose concentrations increased over the same period. Leaf phenolic compound concentrations in our study also declined from early to late May in 2015, consistent with the pattern of seasonal variation in red alder foliar phenolic concentrations observed by others (Radwan et al. 1978, González-Hernández et al. 2000), and in other deciduous trees such as trembling aspen and birch (Manfield et al., 1999;

Riipi et al., 2002, 2004; Millard and Grelet 2010; Haviola et al., 2012). This decline may

be caused by high carbon allocation to production of defense compounds early in the

season to protect new leaves, and subsequent dilution as leaves expand, and is often

observed in early succession plants subject to insect attacks, where carbon allocation to

production of most defense compounds declines during active differentiation and growth

(Mansfield et al. 1999, Riipi et al. 2004).

5. Conclusions

We have shown that genetic variation exists in WTC feeding preference for red alder

leaves and in red alder leaf defense traits. Resistance to insects involves a myriad of complex traits, and although defenseDraft mechanisms not evaluated in this study may be involved in defense against WTC, we have shown that genotype is an important feature

of resistance determined by variation in concentrations of phenolic compounds (CT, Ore,

and TP). The concentrations of foliar phenolic compounds were negatively correlated

with the percent leaf area eaten by WTC, and a threshold was observed, above which the

concentration of each of the three chemicals appeared to reduce WTC feeding. This

indicates that sufficient levels of these secondary metabolites in red alder leaves are

required to affect WTC feeding on red alder. Ultimately, such phenolics will need to be

purified and tested for their anti-herbivore activity directly.

The effect of Ore concentration in red alder leaves on tent caterpillar feeding is a

novel finding. A future study with WTC larvae feeding on leaves with a known range of

purified Ore would confirm the role of Ore in influencing red alder leaf quality in relation

to WTC larval consumption and performance.

6. Acknowledgements

The authors gratefully acknowledge funding from the NSERC CREATE Program in

Forests and Climate Change, and from NSERC Discovery Grants to BJH and CPC. We thank staff at the Cowichan Lake Research Station and the UVic Centre for Forest

Biology for their technical assistance and support. In particular, thanks go to Dr. Lynn

Yip, S. Robbins, and B. Binges for all their help with lab analyses and greenhouse support. The authors also acknowledge the helpful advice of Dr. Cosmin Filipescu,

Canadian Forest Service, throughout this project.

Draft

References

Appel, H. 1993. Phenolics in ecological interactions: the importance of oxidation. Journal of Chemical Ecology 7: 1521-1552.

Ballhorn, D. J., Elias, J. D., Balkan, M. A., Fordyce, R. F. and Kennedy, P. G. 2017. Colonization by nitrogen-fixing Frankia bacteria causes short-term increases in herbivore susceptibility in red alder (Alnus rubra) seedlings. Oecologia 184: 497-506.

Campbell, J.A., Hansen, R.W. and Wilson, J.R. 1999. Cost-effective colorimetric microtitre plate enzymatic assays for sucrose, glucose and fructose in sugarcane tissue extracts. Journal of the Science of Food and Agriculture 79: 232-236.

Constabel, C. P., Yip, L., Patton, J. J. and Christopher, M. E. 2000. Polyphenol oxidase from hybrid poplar. Cloning and expression in response to wounding and herbivory. Plant Physiology 124: 285-295.

Dahija, S., Čakar, J., Vidic, D., Maksimović, M. and Parić, A. 2014. Total phenolic and flavonoid contents, antioxidant and antimicrobial activities of Alnus glutinosa (L.) Gaertn., Alnus incana (L.) MoenchDraft and Alnus viridis (Chaix) DC. extracts. Natural Product Research 28: 2317-2320.

DeBell, D.S. 2006. History of research and attitudes about the biology and management of red alder. In Red alder—a state of knowledge. Edited by Robert L. Deal, and Constance A. Harrington. General Technical Report. PNW-GTR-669. Portland, OR. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. pp. 3-18.

DeBell, D.S. and Radwan, M.A. 1984. Foliar chemical concentrations in red alder stands of various ages. Plant Soil 77: 391-394.

Donaldson, J.R. and Lindroth, R.L. 2008. Effects of variable phytochemistry and budbreak phenology on defoliation of aspen during a forest tent caterpillar outbreak. Agricultural and Forest Entomology 10: 399-410.

González-Hernández, M.P., Starkey, E.E. and Karchesy, J. 2000. Seasonal variation in concentration of fiber, crude protein, and phenolic compounds in leaves of red alder (Alnus rubra): Nutritional implication for cervids. Journal of Chemical Ecology 26(1): 293-301.

Gómez, S., Onoda, Y., Ossipov, V. and Stuefer, J.F. 2008. Systemic induced resistance: a risk-spreading strategy in clonal plant networks? New Phytologist 179(4): 1142-1153.

Harrington, C.A. 2006. Biology and ecology of red alder. In Red alder—a state of knowledge. Edited by Robert L. Deal, and Constance A. Harrington. General

Technical Report. PNW-GTR-669. Portland, OR. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. pp. 21-43

Haukioja, E. 2005. Plant defenses and population fluctuations of forest defoliators: Mechanism-based scenarios. Annales Zoologici Fennici. 42: 313-325.

Haukioja, E., Ossipov, V. and Lempa, K. 2002. Interactive effects of leaf maturation and phenolics on consumption and growth of a geometrid moth. Entomologia Experimentalis of Applicata 104: 125-136.

Haviola, S., Neuvonen, S., Rantala, M.J., Saikkonen, K., Salminen, J., Saloniemi, I., Yang, S. and Teija Ruuhola, T. 2012. Genetic and environmental factors behind chemistry of the mature mountain birch. Journal of Chemical Ecology 38: 902-913.

Hemming, J.D.C. and Lindroth, R.L. 1995. Intraspecific variation in aspen phytochemistry: effects on performance of gypsy moths and forest tent caterpillars. Oecologia 103(1): 79-88.

Hwang, S.H. and Lindroth R.L. 1997. Clonal variation in foliar chemistry of aspen: effects on gypsy moths and forestDraft tent caterpillars. Oecologia 111(1): 99-108 Jackrel, S. L., Morton, T. C. and Wootton, J. T. 2016. Intraspecific leaf chemistry drives locally accelerated ecosystem function in aquatic and terrestrial communities. Ecology 97: 2125-2135.

Jackrel, S.L. and Wootton, J.T. 2015. Cascading effects of induced terrestrial plant defence on aquatic and terrestrial ecosystem function. Proceedings of the Royal Society B: Biological Sciences 282(1805): 20142522.

Karchesy, J. J. and Laver, M. L. 1974. Structure of oregonin, a natural diarylheptanoid xyloside in red alder (Alnus rubra). J. Chem. Soc., Chem. Commun. 16: 649-650

Keinanen, M., Julkunen-Tiitto, R., Mutikainen, P., Walls, M., Ovaska, J. and Vapaavuori, E. 1999. Trade-offs in phenolic metabolism of silver birch: effects of fertilization, defoliation, and genotype. Ecology 80: 1970-1986.

Kopper, B.J., Jakobi V.N., Osier, T.L. and Lindroth, R.L. 2002. Effects of paper birch condensed tannin on white tussock moth (Lepidoptera: Lymantriidae) performance. Environmental Entomology 31(1): 10-14.

Kudo, G. 2003. Variation in leaf traits and susceptibility to insect herbivory within a Salix miyabeana population under field conditions. Plant Ecology 169: 61-69.

Laitinen, M., Julkunen-Tiitto, R. and Rousi, M. 2000. Variation in phenolic compounds within a birch (Betula pendula) population. Journal of Chemical Ecology 26(7): 1609- 1621.

Lindroth, R.L. and St. Clair, S.B. 2013. Adaptations of quaking aspen (Populus tremuloides Michx.) for defense against herbivores. Forest Ecology and Management 299: 14-21.

Lindroth, R.L. and Bloomer, M.S. 1991. Biochemical ecology of the forest tent caterpillars responses to dietary protein and phenolics glycosides. Oecologia 86(3): 408-413.

Major, I.T. and Constabel, C.P. 2007. Shoot–root defense signaling and activation of root defense by leaf damage in poplar. Canadian Journal of Botany 85: 1171–1181.

Major, I.T and Constabel, C.P. 2006. Molecular analysis of poplar defense against herbivory. Comparison of wound‐ and insect elicitor induced gene expression. New Phytologist 172: 617–635. Mansfield, J.I., Curtis, P.S., Zak,Draft D.R. and Pregitzer, K.S. 1999. Genotypic variation for condensed tannin production in trembling aspen (Populus tremuloides, salicaceae) under elevated CO2 and in high- and low-fertility soil. American Journal of Botany 86(8): 1154 -1159.

Mellway, R.D., Tran, L.T., Prouse, M.B., Campbell, M.M. and Constabel, C.P. 2009. The wound-, pathogen-, and UV-B -responsive MYB134 gene encodes an R2R3 MYB transcription factor that regulates proanthocyanidin synthesis in poplar. Plant Physiology 150: 924-941.

Mithöfer, A. and Boland, W. 2012. Plant defense against herbivores: chemical aspects. Annual Review of Plant Biology 63: 431-450.

McArthur, C., Robbins, C.T., Hagerman, A.E., and Hanley, T.A. 1993. Diet selection by a ruminant generalist browser in relation to plant chemistry. Canadian Journal of Zoology 71: 2236-2243.

Mole, S. 1993. The systematic distribution of tannins in the leaves of angiosperms: A tool for ecological studies. Biochemical Systematic and Ecology 8: 833-846.

Myers, J.H. and Williams, K.S. 1984. Does tent caterpillar attack reduce the foliar quality of red alder foliage? Oecologia 62(1): 74-79

Noseworthy, M.N. and Despland, E. 2006. How do primary nutrients affect the performance and preference of forest tent caterpillars on trembling aspen? The Canadian Entomologist 138: 367-375.

Osier, T.L. and Lindroth, R.L. 2006. Genotype and environment determine allocation to and cost of resistance in quaking aspen. Oecologia 148: 293-303.

Osier, T. L. and Lindroth, R. L. 2001. Effects of genotype, nutrient availability, and defoliation on aspen phytochemistry and insect performance. Journal of Chemical Ecology 27: 1289-1313.

Osier, T.L., Hwang, S.Y. and Lindroth, R.L. 2000. Effect of phytochemical variation in quaking aspen Populus tremuloides clones on gypsy moth Lymantria dispar performance in the field and laboratory. Ecological Entomology 25: 197-207.

Ossipov, V., Haukioja, E., Ossipova, S., Hanhimäki, S. and Pihlaja, K. 2001. Phenolic and phenolic-related factors as determinants of suitability of mountain birch leaves to an herbivorous insect. Biochemical Systematics and Ecology 29(3): 223-240.

Panzuto, M., Lorenzetti, F., Mauffette, Y. and Albert, P.J. 2001. Perception of aspen and sun/shade sugar maple leaf soluble extracts by larvae of Malacosoma disstria. Journal of Chemical Ecology 27: 1963-1978. Park, D., Kim, H.J., Jung, S.Y., DraftYook, C.S., Jin, C. and Lee, Y.S. 2010. A new diarylheptanoid glycoside from the stem bark of Alnus hirsuta and protective effects of diarylheptanoid derivatives in human HepG2 cells. Chemical and Pharmaceutical Bulletin 58: 238-241.

Peters, D.J. and Constabel, C.P. 2002. Molecular analysis of herbivore-induced condensed tannin synthesis: cloning and expression of dihydroflavonol reductase from trembling aspen (Populus tremuloides). The Plant Journal 32: 701-712.

Porter, L.J., Hrstich L.N. and Chan B.G. 1986. The conversion of procyanidins and prodelphinidins to cyanidin and delphidin. Phytochemistry 25: 223–230.

Radwan, M.A., Ellis, W.D., Walter, D. and Crouch, G.L. 1978. Chemical composition and deer browsing of red alder foliage. Portland, Or. : Pacific Northwest Forest and Range Experiment Station, US Dept. of Agriculture, Forest Service. pp 1-6.

Riipi, M., Haukioja, E., Lempa, K., Ossipov, V., Ossipova, S. and Pihlaja, K. 2004. Ranking of individual mountain birch trees in terms of leaf chemistry: seasonal and annual variation. Chemoecology 14: 31-43.

Riipi, M., Ossipov, V., Lempa, K., Haukioja, E., Koricheva, J., Ossipova, S. and Pihlaja, K. 2002. Seasonal changes in birch leaf chemistry: are there trade-offs between leaf growth and accumulation of phenolics? Oecologia 130: 380-390.

Robbins, C.T., Hanley, T.A., Hagerman, A.E., Hjeljord, O., Baker, D.L., Schwartz, C.C., and Mautz, W.W. 1987. Role of tannins in defending plants against ruminants: Reduction in protein availability. Ecology 68: 98 - 107.

Robison, D.J. and Raffa, K.F. 1994. Characterization of Hybrid poplar clones for resistance to the forest tent caterpillar. Forest Science 40(4): 686-714.

Sarfraz, R.M., Kharouba, H.M. and Myers, J.H. 2013. Tent caterpillars are robust to variation in leaf phenology and quality in two thermal environments. Bulletin of Entomological Research 103: 522-529.

Sati, S.C., Sati, N. and Sati, O.P. 2011. Bioactive constituents and medicinal importance of genus Alnus. Pharmacognosy Reviews 5: 174-183.

Saxena, G., Farmer, S., Hancock, R.E.W. and Towers, G.H.N. 1995. Antimicrobial compounds from Alnus rubra. International Journal of Pharmacognosy 33(10): 33-36.

Smith, A.H., Pinkard, E.A., Hunter, G.C., Wingfield, M.J. and Mohammed, C.L. 2006. Anatomical variation and defence responses of juvenile Eucalyptus nitens leaves to Mycosphaerella leaf disease. Australasian Plant Pathology 35(6): 725-731.

US Forest Service, 2011. Western tent caterpillar silken tents in early spring. Forest health protection, Rocky MountainDraft region. War, A.R., Taggar, G.K., Hussain, B., Taggar, M.S., Nair, R.M. and Sharma, H.C. 2018. Plant defence against herbivory and insect adaptations. AoB PLANTS 10: ply037; doi: 10.1093/aobpla/ply037.

War, A.R., Paulraj, M.G. and Sharma, H.C. 2012. Mechanisms of plant defense against insect herbivores. Plant signal behav. 7(10): 1306-1320.

Williams, K.S. and Myers, J.H. 1984. Previous herbivore attack on red alder may improve food quality of webworm larvae. Oecologia 63:166-170.

Yang, S., Haviola, S., and Ruuhola, T. 2007. Temporal and spatial variation in mountain birch foliar enzyme activities during the larval period of Epirrita autumnata. Chemoecology 17: 71-79.

1 Table 1. Red alder clone and tree numbers sampled from a clone bank at the Cowichan Lake 2 Research Station, and the latitude, longitude, elevation, number, and name of their provenances 3 Provenance # - name Clone no. Tree no.1 Lat. Long. Elev. (m) (o N) (o W)

P8-Nitinat Flats C9 083-3-3 48 50 124 40 30 C125 084-2-3 P9-Cowichan C34 093-3-1 48 46 123 39 150 C18 095-3-1 P11-Sarita Lake C133 112-3-5 48 55 124 52 40 C17 113-3-1 P15-Cassidy C124 152-2-3 49 03 123 56 107 C94 158-2-4 P35-Culliton Creek C13 351-2-5 49 53 123 11 250 C79 352-2-4 P41-Quadra Island C7 412-3-2 50 17 125 22 50 C100 413-2-5 P50-San Josef Main C90 502-3-3 50 40 128 04 20 C28 504-3-3 P51-NE 62 C38 Draft512-2-5 50 43 127 59 170 C12 514-2-4 P52-Kingcome Inlet C4 521-2-2 51 30 126 08 30 C42 523-3-5 P55-Bachelor Bay C162 552-2-2 52 22 126 55 30 C115 553-2-4 4 1. Tree number - BC Ministry of Forests code: provenance/family-replication-seedling code of 5 the parent at the Bowser test site. 6

8 Table 2. Mixed model analysis of variance (ANOVA) for testing the effects of A. red alder clone 9 nested within provenance (Prov), sampling date, experimental environment (Envt), and their 10 interactions on the percent leaf area eaten by WTC for 20 clones sampled in 2014, with the 11 respective error terms; and B. red alder clone nested within provenance (Prov), sampling date, 12 and their interactions on the measured leaf traits of the subset of 10 red alder clones sampled in 13 2014, with the respective error term. df = degrees of freedom. 14 Source of variation df Error term

A. Prov 9 Clone(Prov) Clone(Prov) 10 Tree(Clone(Prov)) Tree(Clone(Prov)) 40 Error Envt 1 Envt *Tree(Clone(Prov)) Envt*Prov 9 Envt*Tree(Clone(Prov)) Envt*Clone(Prov) 10 Envt*Tree(Clone(Prov)) Date 3 Date*Clone(Prov) Date*Prov 27 Date*Clone(Prov) Date*Clone(Prov) 30 Date*Envt*Tree(Clone(Prov)) Envt*Date 3 Date*Envt*Tree(Clone(Prov)) Envt*Date*Clone(Prov) 30 Date*Envt*Tree(Clone(Prov)) Date*Envt*Tree(Clone(Prov)) Draft 120 Error

B. Prov 9 Clone(Prov) Clone(Prov) 10 Tree(Clone(Prov)) Date 1 Clone(Prov) *Date Date*Prov 9 Clone(Prov)*Date Clone(Prov)*Date 10 Error Tree(Clone(Prov)) 20 Error

15 16

17 Table 3. Mixed model analysis of variance (ANOVA) for testing the effects of red alder clone, 18 tree nested within clone, sampling date, and their interactions on the percent leaf area eaten by 19 WTC and the measured leaf traits of the 10 red alder clones sampled in 2015, with the respective 20 error term. df = degrees of freedom. 21 Source of variation df Error term

Clone 9 Tree(Clone) Tree(Clone) 18 Tree(Clone)*Date Date 1 Tree(Clone)*Date Date*Clone 9 Tree(Clone)*Date Date*Tree(Clone) 18 Error 22 23

Draft

24 Table 4. Mean percent leaf area eaten (damage) (+ S.E) of red alder clones averaged over 25 greenhouse and laboratory bioassay feeding experiments with WTC, and all sampling dates in 26 2014 and 2015 (N=1331 and 240 in 2014 and 2015, respectively). Overall damage class (high = 27 H, medium = M, and low = L) was assessed based on mean percent area eaten in the two years. 28 Leaves were collected from eight or nine-year-old red alder trees at the Cowichan Lake Research 29 Station in April and May, 2014 and 2015. Within year, clonal means of leaf damage followed by 30 the same letter are not significantly different (p < 0.05). 31 Provenance # - Name Clone # 2014 damage 2015 damage Overall % % damage class P8-Nitinat Flats C9 28.2 + 3.5 a 47.9 + 3.5 a H C125 21.7 + 3.4 abc M P9-Cowichan MF C34 18.4 + 2.3 c 24.2 + 4.5 b L C18 17.8 + 1.6 c 29.7 + 5.9 b L P11-Sarita Lake C133 16.7 + 2.0 c 29.8 + 5.3 b L C17 21.4 + 2.3 abc L P15-Cassidy C124 19.9 + 2.3 bc 21.9 + 3.6 b L C94 20.6 + 2.7 bc M P35-Culliton Cr C13 23.5 + 3.1 abc M C79 25.8 + 2.9 ab 24.4 + 5.5 b H P41-Quadra Is C7 20.3 + 2.3 bc M C100 Draft22.1 + 3.0 abc M P50-San Josef Main C90 17.8 + 2.4 c 20.7 + 3.7 b L C28 20.9 + 2.5 bc L P51-NE 62 C38 29.2 + 3.3 a 33.5 + 7.3 ab H C12 25.0 + 2.7 ab 30.7 + 6.2 b H P52-Kingcome Inlet C4 22.6 + 3.5 abc M C42 22.5 + 2.9 abc M P55-Bachelor Bay C162 27.0 + 2.8 a 33.4 + 5.7 ab H C115 24.1 + 2.6 ab H

Table 5. P-values from the mixed model analysis of variance testing the effects of red alder clones nested within provenances (Prov), sampling dates, sampling environments (Envt), and their interactions on the percent leaf area eaten by WTC (Damage) (20 clones); and the measured leaf traits and elements showing significant differences (10 clones) of the red alder ramets in 2014. Thick = thickness, Tough = toughness, CT = condensed tannin, BB = budburst, and Water = moisture content. P-values in bold indicate statistical significance (p < 0.05).

Source Variable Damage Thick Tough BB Water CT N C

Prov 0.04 0.88 0.44 0.25 0.93 0.44 0.76 0.07 Clone(Prov) 0.95 0.0001 0.10 <0.001 0.14 0.008 0.02 0.26 Tree(Clone (Prov)) 0.33 <0.0001 0.53 <0.001 0.88 0.25 0.20 0.003 Envt <0.0001 Prov*Envt 0.27 Envt*Clone(Prov) 0.11 Draft Date 0.03 0.49 0.15 0.78 <0.001 0.31 0.27 0.08 Date*Prov 0.04 0.23 0.18 0.17 0.99 0.29 0.06 0.15 Date*Clone(Prov) 0.20 0.08 0.53 0.15 0.15 0.02 0.16 0.25 Envt*Date <0.0001 Envt*Date*Clone(Prov) 0.19

Ca Cu Mg Al Fru Suc Glu

Prov 0.73 0.18 0.03 0.04 0.73 0.72 0.71 Clone(Prov) 0.02 0.005 0.81 0.71 0.02 0.16 0.01 Tree(Clone (Prov)) 0.18 0.0003 0.16 0.21 0.18 0.52 0.32 Date 0.08 0.009 0.41 0.28 0.08 0.03 0.01 Date*Prov 0.10 0.009 0.79 0.007 0.10 0.31 0.78 Date*Clone(Prov) 0.10 0.005 0.07 0.02 0.18 0.52 0.48

Table 6. P-values from the mixed model analyses of variance testing the effects of red alder clone and sample tree nested within clone, sampling date, and their interactions on the percent leaf area eaten (damage) by WTC; and the measured leaf traits of the selected red alders on two dates in May, 2015. Thick = thickness, Tough = toughness, CT = condensed tannin, TP = total phenolic, Ore = oregonin, BB = budburst, and Water = moisture content. P-values in bold indicate statistical significance (p < 0.05).

Source Variable Damage Thick Tough BB Water CT TP Ore

Clone 0.01 <0.0001 0.002 <0.0001 0.21 <0.0001 0.003 0.0004 Tree(Clone) 0.16 0.91 0.58 <0.0001 0.40 0.27 0.57 0.77 Date <0.0001 0.005 0.95 0.06 <0.001 <0.0001 <0.0001 0.04 Date*Clone 0.01 0.77 0.40 0.29 0.12 0.003 0.22 0.51 Date*Tree(Clone) 0.74 0.50 0.50 0.50 0.50 0.50 0.50 0.50 Draft N C Ca Cu

Clone 0.005 0.0006 0.03 0.005 Tree(Clone) 0.13 0.22 0.005 <0.0001 Date <0.001 0.0002 Date*Clone 0.09 0.04 Date*Tree(Clone) 0.50 0.50

Table 7. Mean (± SE) of select leaf traits measured in 2014 in a subset of 10 red alder clones. Leaves were collected from eight-year- old red alder trees at the Cowichan Lake Research Station in April and May 2014. LSD test results of mean differences in each trait among clones are shown by letters (p < 0.05). Highest and lowest values for each trait are indicated by bold and underline, respectively. Thickness = leaf thickness, Toughness = leaf toughness, BB = leaf budburst stage, Water = percent moisture content, CT = condensed tannin, mineral elements, Fruc = fructose concentration, Suc = sucrose concentration, Gluc = glucose concentration.

Mean (±SE) of measured leaf trait

Clone Thickness Toughness BB Water (%) CT (µg/mg) N (%) C (%) (mm) (g/mm2) C9 0.13±0.005f 132.1±2.5 3.0±0.001f 64.4±1.5 15.2±1.4b 2.6±0.04bcd 50.3±0.2 C12 0.15±0.005e 111.7±2.4 4.3±0.071a 65.8±1.0 10.8±1.0bcd 2.5±0.06d 49.8±0.1 C18 0.20±0.004bc 149.2±1.2 3.5±0.001e 64.1±1.9 9.9±1.7cd 2.7±0.10abc 49.7±0.2 C34 0.18±0.004d 144.9±1.4 3.0±0.001f 65.8±2.0 16.8±4.2b 2.6±0.09bcd 49.7±0.1 C38 0.20±0.006bc 136.4±0.1 3.8±0.001d Draft66.9±1.4 7.7±0.6cd 2.5±0.08cd 49.8±0.1 C90 0.20±0.001ab 117.6±4.6 3.4±0.036e 66.4±1.7 27.8±1.8a 2.2±0.08e 49.9±0.1 C115 0.19±0.004cd 118.2±3.0 3.9±0.036c 66.5±1.2 7.6±1.1d 3.0±0.10a 50.2±0.1 C125 0.21±0.001a 136.6±7.3 3.5±0.001e 64.5±2.3 25.4±1.3a 2.2±0.08e 50.6±0.1 C133 0.19±0.003cd 134.9±2.8 4.2±0.071b 64.9±1.7 15.5±3.2b 2.1±0.12e 49.8±0.2 C162 0.21±0.006a 138.7±2.3 3.5±0.001e 68.3±2.0 13.9±3.2bc 2.8±0.09ab 50.0±0.2

Ca (%) Cu (mg/kg) Mg (%) Al (mg/kg) Fruc (mg/ml) Suc (mg/ml) Glu (mg/ml) C9 0.49±0.01ab 17.4±1.5c 0.26±0.01ab 18.4±0.73cde 4.9±0.7b 10.6±1.6ab 16.8±0.8c C12 0.54±0.01a 18.3±1.8c 0.24±0.01bc 19.2±1.29cd 5.5±0.5b 8.7±0.9abc 17.1±1.3c C18 0.50±0.02ab 18.8±1.5c 0.21±0.008c 29.8±1.57b 5.4±0.2b 8.9±0.8abc 18.2±0.3bc C34 0.51±0.02ab 18.7±2.1c 0.22±0.009c 35.1±1.86a 5.1±0.2b 7.9±0.5c 16.0±1.0c C38 0.39±0.03c 20.8±1.3ab 0.20±0.009c 32.0±3.19ab 13.3±3.6a 11.4±1.6a 27.1±3.8a C90 0.50±0.03ab 24.9±2.1a 0.27±0.01ab 16.0±1.18def 5.9±0.3b 8.4±0.7bc 18.1±0.6bc C115 0.47±0.02bc 23.6±2.0ab 0.21±0.02c 12.8±0.67f 7.3±0.3b 7.5±0.5c 21.5±0.9b C125 0.48±0.03bc 24.2±2.5a 0.24±0.013b 23.3±2.75c 5.7±0.4b 7.8±0.6c 16.7±1.1c C133 0.52±0.05ab 17.8±1.7c 0.27±0.013a 20.0±2.74cd 6.3±0.4b 8.9±0.5abc 18.4±0.9bc C162 0.54±0.03ab 18.8±1.2bc 0.22±0.003c 14.4±0.56ef 6.3±0.5b 7.8±0.5c 19.6±1.4bc

Table 8. Mean (± SE) of select leaf traits measured in 2015 in 10 red alder clones. Leaves were collected from nine-year-old red alder trees at the Cowichan Lake Research Station in April and May 2015. LSD test results of mean differences in each trait among clones are shown by letters (p < 0.05). Highest and lowest values for each trait are indicated by bold and underline, respectively. Thickness = leaf thickness, Toughness = leaf toughness, BB = leaf budburst stage, Water = percent moisture content, CT = condensed tannin, TP = total phenolic, Ore = oregonin. TP expressed as Ore equivalent.

Mean ±SE of measured leaf trait Clone Thickness Toughness BB Water (%) CT (µg/mg) TP (µg/mg) Ore (µg/mg) (mm) (g/mm2) C9 0.13±0.005e 85.9±2.3f 3.3±0.130c 65.1±0.54 18.0±1.8b 454.9±11.1e 134.9±13.6b C12 0.14±0.008e 84.5±1.0f 3.5±0.005a 64.6±0.35 12.9±1.0c 514.4±8.4bc 85.25±9.5de C18 0.19±0.001bc 109.1±2.3bcd 2.0±0.001f 62.9±0.55 12.6±0.8c 489.3±7.0 d 123.1±7.8bc C34 0.18±0.006d 103.9±3.2cde 2.0±0.001f 64.9±0.67 5.6±0.3d 451.7±13.5e 73.1±3.3e C38 0.19±0.002bc 131.5±1.7a 3.5±0.067b Draft64.1±1.39 6.6±0.8d 501.3±11.1cd 78.9±8.7de C79 0.19±0.002ab 110.9±2.2bcd 4.8±0.005a 61.4±0.39 17.7±1.3b 494.8±7.6cd 191.6±12.2a C90 0.19±0.003bc 102.5±1.7e 2.5±0.001d 63.9±0.26 24.4±1.7a 540.2±4.2a 172.1±5.3a C124 0.19±0.002bc 103.6±3.5de 3.5±0.001b 63.3±0.37 14.1±0.4c 529.6±5.1ab 184.7±12.1a C133 0.20±0.002a 114.4±5.3b 2.2±0.051e 62.2±0.53 13.3±0.8c 516.0±6.1bc 129.4±6.9b C162 0.18±0.002cd 111.5±1.8bc 2.0±0.001f 64.9±0.45 12.6±0.7c 439.9±10.6e 100.4±5.2cd

N (%) C (%) Ca (%) Cu (mg/kg)

C9 2.4±0.04d 50.1±0.11bc 0.50±0.02bc 11.8±0.41f C12 2.3±0.08d 49.8±0.09de 0.46±0.02cd 13.3±0.32de C18 2.6±0.05bc 49.5±0.07f 0.43±0.02de 12.3±0.14ef C34 3.2±0.13a 50.1±0.06cd 0.39±0.01ef 17.0±0.02cd C38 2.6±0.06b 50.0±0.07ef 0.35±0.01f 17.8±0.01b C79 2.4±0.04cd 50.2±0.03ab 0.52±0.02b 15.4±0.32c C90 2.4±0.05cd 50.3±0.05a 0.63±0.02a 19.0±0.24b C124 2.8±0.08b 50.3±0.05ab 0.35±0.02f 22.3±1.34a C133 2.4±0.09cd 50.3±0.07ab 0.44±0.01d 12.5±0.09def C162 3.0±0.11a 50.4±0.08a 0.44±0.02d 13.9±0.16d

Draft

120 R² = 0.1656 100

80 60 40

20 Leaf area eaten area eaten Leaf(%) 0 0 10 20 30 40 -20 a) Condensed tannin concentration (ug mg-1 DW)

120 R² = 0.1912 100 80 60 40 20 Draft Leaf area eaten area eaten Leaf(%) 0 350 400 450 500 550 600

b) -1 Total phenolics concentration (ug mg DW)

120 R² = 0.0761 100 80 60 40 20 Leaf area eaten area eaten Leaf(%) 0 0.00 50.00 100.00 150.00 200.00 250.00 300.00 350.00 Oregonin concentration (ug mg-1 DW) c)

Figure 1. Regression plot showing the relationship between red alder leaf CT (a), TP (b), and Ore (c) concentrations and percent leaf area eaten by WTC for individual red alder trees from 10 clones on two dates in May, 2015. The dotted vertical lines suggest the threshold, above which the respective foliar phenolic concentration was associated with less WTC feeding. N=240. Note: Total phenolics concentration (b) expressed as Ore equivalents. Condensed tannins are quantified as aspen tannin equivalents. DW = dry weight.