Interspecific variation and elevated CO2 influence the relationship between plant chemical resistance and regrowth tolerance

Abstract To understand how comprehensive plant defense phenotypes will respond to global change, we investigated the legacy effects of elevated CO2 on the relationships between chemical resistance (constitutive and induced via mechanical damage) and regrowth tolerance in four milkweed species (Asclepias). We quantified potential resistance and tolerance trade‐offs at the physiological level following simulated mowing, which are relevant to milkweed ecology and conservation. We examined the legacy effects of elevated CO2 on four hypothesized trade‐offs between the following: (a) plant growth rate and constitutive chemical resistance (foliar cardenolide concentrations), (b) plant growth rate and mechanically induced chemical resistance, (c) constitutive resistance and regrowth tolerance, and (d) regrowth tolerance and mechanically induced resistance. We observed support for one trade‐off between plant regrowth tolerance and mechanically induced resistance traits that was, surprisingly, independent of CO2 exposure. Across milkweed species, mechanically induced resistance increased by 28% in those plants previously exposed to elevated CO2. In contrast, constitutive resistance and the diversity of mechanically induced chemical resistance traits declined in response to elevated CO2 in two out of four milkweed species. Finally, previous exposure to elevated CO2 uncoupled the positive relationship between plant growth rate and regrowth tolerance following damage. Our data highlight the complex and dynamic nature of plant defense phenotypes under environmental change and question the generality of physiologically based defense trade‐offs.

the need to test at multiple resource levels, and measure not only growth rate, but net assimilation and secondary metabolism, the GDB has proven difficult to test but still provides a useful framework of plant defense at the physiological level (Stamp, 2004). In general, trade-offs between tolerance and chemical resistance arise as a result of plant allocation strategies meant to optimize fitness in a variable environment (Züst & Agrawal, 2017). Therefore, understanding the environmental conditions under which trade-offs manifest is of critical importance.
Further, elevated CO 2 suppresses the synthesis of jasmonic acid and stimulates the production of salicylic acid, compromising the plant's ability to mount an induced resistance response (Ode, Johnson, & Moore, 2014).
Changes in phytohormonal signaling pathways also mediate plant growth and regrowth tolerance following damage under elevated CO 2 (Guo et al., 2012). In general, by increasing photosynthesis and water use efficiency, elevated CO 2 positively affects plant growth rates (Drake, Gonzalez-Meler, & Long, 1997;Ainsworth & Long, 2005;Robinson et al., 2012;Bazzaz, Ackerly, Woodward, & Rochefort, 2013). However, the direct effects of elevated CO 2 on plant regrowth tolerance following damage can be negative (Guo et al., 2012;Lau & Tiffin, 2009;Marshall, Avila-Sakar, & Reekie, 2008;Wilsey, 2001) partially because of increased nutrient limitation under elevated CO 2 paired with phytohormonal suppression. Studies that explore the integrated influence of elevated CO 2 on the relationships between resistance and tolerance are sorely lacking.
Even less is known about the lingering effects of past CO 2 enrichment on plants. Though not ecologically plausible, the modulation of exposure to environmental change drivers such as elevated CO 2 partially reveals energetic allocation decisions made by plants under future conditions, and the persistence of those responses.
Extrapolations based on the substantial below-ground carbon sink and increased soil microbial turnover that develops in response to elevated CO 2 predict mixed but lingering effects of elevated CO 2 on plant regrowth tolerance (Hungate, Johnson, & Dijkstra, 2006;Stiling, Moon, & Rossi, 2013). To our knowledge, only two studies have examined plant responses to elevated CO 2 beyond the cessation of enrichment and found lasting effects on aspects of root morphology such as fine root hairs (Stiling et al., 2013) and increases in regrowth tolerance following fire (Bain & Day, 2019). These studies follow plant and arthropod communities in the years following enrichment cessation, yet how plant physiological properties will respond to abrupt changes in CO 2 enrichment over the course of a growing season remains to be tested.
Here, we investigate the legacy effects of elevated CO 2 on the chemical resistance traits and regrowth tolerance of four milkweed species (Asclepias). Specifically, we examined the effects of elevated CO 2 on four hypothesized trade-offs between the following: (a) initial growth rate and constitutive chemical resistance, (b) initial growth rate and mechanically induced chemical resistance, (c) constitutive chemical resistance and regrowth tolerance following damage, and (d) regrowth tolerance and mechanically induced chemical resistance. To our knowledge, no theory exists to predict the interaction between resistance and regrowth tolerance strategies under changing carbon supplementation. Nevertheless, we predicted that elevated CO 2 would induce higher growth rates and regrowth rates and depress constitutive secondary metabolites following the GDB hypothesis and mitigate, in part, any trade-off between chemical resistance traits and regrowth tolerance in milkweed. By analyzing changes in plant tolerance and resistance chemistry, we aimed to improve our understanding of how future environmental conditions may influence the defensive phenotype of plants, with implications for the herbivore communities that damage them.

| Study system
The four milkweed, Asclepias, species used in our study (A. syriaca, A. speciosa, A. incarnata, and A. curassavica) originate from North and Central America (Woodson, 1954) and support herbivores that range from phloem-feeding insects such as oleander aphids (Aphis nerii) to chewing insects capable of removing large amounts of tissue, like monarch caterpillars (Danaus plexippus), and long horn beetles (Tetraopes spp.). Most milkweed herbivores specialize within the genus because Asclepias produce a well-characterized suite of defenses against herbivory.
To physically deter feeding by arthropod herbivores, milkweed plants exude latex, produce trichomes, and increase leaf toughness (Agrawal & Fishbein, 2006;Agrawal & Konno, 2009;Hochwender, Marquis, & Stowe, 2000;Zalucki, Brower, & Alonso-M, 2001). However, milkweeds are best known for synthesizing a class of toxic steroids known as cardenolides that disrupt Na + /K + -ATPase in the Na + /K + -channels of animal cells (Agrawal, Petschenka, Bingham, Weber, & Rasmann, 2012). The composition and concentration of cardenolides produced constitutively by milkweed plants vary substantially within and among milkweed species (Agrawal et al., 2012;Rasmann & Agrawal, 2011). Damage induces quick increases in cardenolide concentrations and changes in cardenolide composition (Malcolm & Zalucki, 1996). Regrowth following damage also plays a prominent role in the defensive phenotype of milkweeds (Agrawal & Fishbein, 2008;Tao, Ahmad, Roode, & Hunter, 2016). Despite a growing body of work illustrating the effects of environmental change on milkweed chemistry and milkweed growth (Matiella, 2012;Tao, Berns, & Hunter, 2014;Vannette & Hunter, 2011), no study to date has explored the interplay between milkweed chemical resistance traits (both constitutive and induced) and regrowth tolerance under future environmental conditions. We grew four species of milkweed under ambient (400 ppm) and elevated (760 ppm) concentrations of atmospheric CO 2 at the University of Michigan Biological Station (UMBS). To manipulate atmospheric CO 2 concentrations, we used an outdoor array consisting of 40 open-top chambers, with 20 chambers maintained at ambient CO 2 , and 20 chambers maintained at elevated CO 2 from May through August of 2015. Chambers were 1 m high cubes with an octagonal top of diameter of 0.8 m composed of a PVC frame and clear plastic walls following a modified design of Drake, Leadley, Arp, Nassiry, and Curtis (1989).
We chose Asclepias species that vary in foliar cardenolide concentrations. Specifically, we included A. incarnata (low cardenolide), A. speciosa, A. syriaca (both medium cardenolide), and A. curassavica (high cardenolide). Seeds of A. speciosa and A. curassavica were obtained from commercial sources (Prairie Moon Nurseries, Winona, USA), and seeds of A. incarnata and A. syriaca were collected locally (Cheboygan county, MI). We surface-sterilized all seeds following a six-week cold stratification period (for all but tropical A. curassavica) and germinated seeds on moist filter paper for 1 week. We planted seedlings in 983 cm 3 Deepots TM (6.9 cm diameter by 35.6 cm height) containing Metromix 360 (Sun Gro Horticulture, Vancouver, BC) and Osmocote Controlled Release Fertilizer [N:P:K:16:16:16 ppm N (g/g)] (ICL Specialty Fertilizers, Dublin, USA) on May 5, 15. Germinated seedlings were watered daily and grown in the UMBS greenhouse for two weeks before they were moved to randomly assigned chambers in the CO 2 array. Once in the array, potted plants were maintained under their CO 2 treatments for three months. To minimize the entrance of herbivores into the chambers, we placed fine mesh coverings over the openings of each chamber and physically removed any herbivores that we observed during daily visual inspections.
Within each chamber, we grew as many as seven plants of each milkweed species. Low germination success limited the number of A. speciosa and A. syriaca used in this study, and not all milkweed species were represented in every chamber. Overall, our eight treatments (2 CO 2 treatments × 4 milkweed species) combined for a total of 442 plants, with exact replicate numbers reported in Table 1.
Using a LI-COR 320 IRGA (LI-COR, Lincoln, USA), we monitored atmospheric CO 2 concentrations daily in the 20 elevated CO 2 chambers and in one randomly selected ambient CO 2 chamber.
Concentrations of CO 2 were adjusted throughout the day to maintain the target of 760 ppm in each elevated chamber. The ambient temperature inside each chamber was recorded every hour using a thermochron datalogger (Thermochron, Baulkham Hills, Australia).

| Simulated damage and growth measures
Three months following the initial transfer of plants into the array, we simulated clipping/mowing by cutting all plants at the soil line.
Many milkweed habitats important to the specialist herbivores associated with milkweed are located near roadways and agricultural fields that are regularly mowed. Properly timed mowing can improve reproduction and decrease predator abundance of certain milkweed specialists, including the monarch butterfly (Haan & Landis, 2019 days since the seedling had been transferred to soil) following Agrawal and Fishbein (2008). Similarly, to calculate plant regrowth rate following mechanical damage, we divided the mass of the regrowth material by 21 days (the length of time plants were allowed to regrow following damage). Differences in regrowth rate following damage are important for the competitive success and ultimate fitness of plants (Züst & Agrawal, 2017).

| Chemical analyses and resistance classifications
We collected samples of the original aboveground foliage and the regrowth foliage of each plant for cardenolide analysis using established methods (Tao & Hunter, 2012;Vannette & Hunter, 2011;Zehnder & Hunter, 2009

| Statistical analyses
In all analyses that follow, we used either linear mixed models (LMMs; Lme4 package) or generalized linear mixed models (GLMMs; Lme4 package). To account for variation among chambers and the nonindependence of plants grown within the same individual chamber, we included chamber identity as a random effect in all of our models described below. This design allows us to test our hypotheses at the level of plant individuals to capture relevant variation in our analyses, while accounting for multiple plants within chambers. We performed all statistical tests in R version 3.6.0 (R Core Team, 2018) and selected models using likelihood ratio tests (Burnham & Anderson, 2002). Variables were transformed to best achieve normality of error as tested by the We used an LMM with log-transformed initial foliar cardenolide concentrations as the dependent variable and square root-transformed growth rate prior to clipping, CO 2 treatment, and milkweed species as fixed effects. An interaction between growth rate prior to clipping and CO 2 indicates a difference between the CO 2 treatments in the extent to which growth rate correlates with the production of cardenolides.

| Plant growth rate before damage and mechanically induced resistance of regrowth tissues
We used an LMM with log-transformed foliar cardenolide concentrations of the regrowth foliage as the dependent variable and square root-transformed growth rate prior to clipping, CO 2 treatment, and milkweed species as fixed effects. An interaction between initial growth rate and CO 2 indicates a difference between CO 2 treatments in the potential trade-off between plant growth rate before damage and chemical resistance after damage.

| Chemical resistance before damage and regrowth tolerance
Likewise, we ran an LMM with square root-transformed regrowth rate as the response variable and log-transformed initial foliar cardenolide concentrations, CO 2 treatment, and milkweed species as fixed effects. An interaction between initial foliar cardenolide concentration and CO 2 indicates a difference between atmospheres in the relationship between initial plant chemical resistance and regrowth.

| Regrowth tolerance and the mechanically induced resistance of regrowth tissues
Lastly, we ran an LMM with log-transformed regrowth foliar cardenolide concentrations as the response variable and square roottransformed regrowth rate, CO 2 treatment, and milkweed species as fixed effects. A significant interaction between CO 2 treatment and regrowth rate would signify a difference between the two atmospheres in any correlation between the two defense traits.

| Elevated CO 2 , milkweed species, and plant growth and resistance profiles
While the trade-off model framework described above provided some information on how growth rates and chemical resistance responded to our treatments, we also performed the following additional analyses to ask further questions about defense phenotypes.
To determine the effects of our treatments on plant growth rate prior to damage and regrowth rate after damage, we used CO 2 treatment, the probability of regrowth, and milkweed species as fixed effects and square root-transformed growth rates (mg/day) as response variables. Not all milkweed individuals regrew following damage. We therefore used generalized linear mixed models with binomial error distributions and logit link functions to assess the effects of plant species and CO 2 treatment on the proportion of milkweed plants that regrew following damage.
We then examined how CO 2 treatment and species influenced the relationship between growth rate prior to damage and regrowth rate following damage, using an LMM with square root-transformed regrowth rate as the response variable and square root-transformed initial growth rate, CO 2 treatment, and species as fixed effects.
Plant chemical defense encompasses not only the total concentration of defense compounds but also the diversity of chemical species produced. We therefore examined the relationships between cardenolide community diversity and growth rates.

| Plant growth rate and constitutive resistance
Milkweed growth rate prior to damage was unrelated to foliar constitutive cardenolide concentrations prior to damage (initial growth rate: F 1,195 = 2.72, p = .100, Figure 1a; Table 2). Elevated CO 2 had no effect on this nonsignificant relationship (CO 2 *initial growth rate: F 1,195 = 0.46, p = .499).

| Plant growth rate before damage and mechanically induced resistance
Instead of a trade-off between growth rate prior to damage and the mechanically induced chemical resistance of regrown tissues following damage, we found a positive relationship that weakened (became less steep) under elevated CO 2 (CO 2 *initial growth rate: F 1,215 = 5.33, p = .022, Figure 1b; Table 2).

| Constitutive resistance before damage and regrowth tolerance after damage
Similarly, we observed a weak positive relationship between constitutive chemical resistance and regrowth tolerance (constitutive resistance: F 1,208 = 3.66, p = .057, Figure 1c; Table 2). Model selection eliminated models containing the influence of CO 2 on this relationship.

| Regrowth tolerance and mechanically induced resistance of regrown tissues
In contrast to the first three potential trade-offs, we observed a significant trade-off between regrowth tolerance and the mechanically induced chemical resistance of regrown foliage (Regrowth rate*milkweed species: F 1,215 = 7.18, p = .0001, Figure 1d; Table 2).
The trade-off was determined by two of the four milkweed species (A. incarnata and A. speciosa). As above, our selection process eliminated models containing the influence of CO 2 on this relationship.
In other words, future atmospheric concentrations of CO 2 uncoupled the relationship between regrowth tolerance following damage and initial growth rate before damage. Following mechanical damage, only 278 of the 442 plants (63%) regrew aboveground tissue.
In those plants that did regrow following damage, mechanically induced resistance varied substantially by milkweed species (species: F 3,215 = 8.59, p < .0001, Figure 5b; Note: Model selection was performed using maximum likelihood. Tables were produced with the R package LmerTest, using type III sums of squares with Satterthwaite approximation for degrees of freedom, random effects estimates ± 1 standard deviation, and fixed effects parameter estimates ± 1 standard deviation. F I G U R E 2 Elevated CO 2 increased initial milkweed growth rate but had no lasting effects on regrowth rate following damage. The effects of CO 2 treatment and milkweed species on (a) initial growth rate prior to damage (mg dry mass of above-ground tissue/64 days) and (b) nonsignificant effects of elevated CO 2 and milkweed species on regrowth rate following damage (mg dry mass of above-ground tissue/21 days). In boxplots, dark lines represent the median, box boundaries represent first and third quartiles, and whiskers extend to the most extreme data point less than 1.5 times the interquartile range from the box. Milkweed species codes are the same as above. Data are grouped by species and CO 2 treatment for ease of interpretation; however, the interaction term was not retained in our models of regrown plants N.S.
The diversity of cardenolides produced constitutively among milkweed species increased by 24% under elevated CO 2 (CO 2 : F 1,68 = 4.08, p = .047, Figure 5c; Table 4). Despite a species-specific effect of elevated CO 2 on the total concentration of constitutive resistance, there was no such effect on the diversity of cardenolides produced constitutively (species*CO 2 : F 3,206 = 2.04, p = .109, Figure 5c; Table 4). Conversely, the diversity of cardenolides produced in the mechanically induced resistance profiles of both A.
When comparing the composition of cardenolide communities among individuals before and after damage, the difference between constitutive and mechanically induced foliar tissue was the strongest driver of community dissimilarity as determined by PerMANOVA (resistance type: F 1, 410 = 55.38, p = .001, R 2 = 0.15, Figure 6; Table 5). There were slight differences between these two resistance profiles among milkweed species driven by elevated CO 2 (resistance type*species*CO 2 : F 2, 410 = 2.39, p = .001, R 2 = 0.013, Figure 6; Table 5), and these slight differences likely represent the changes in cardenolide diversity detected above.

| D ISCUSS I ON
Our study reveals the limitations of a trade-off framework at the physiological level when considering how complex defense phenotypes respond to environmental change. Of the four hypothesized trade-offs among aspects of plant growth and resistance framing the study, we found support for only one between regrowth tolerance and mechanically induced chemical resistance (foliar cardenolide concentration following mechanical damage). varied substantially among milkweed species, presumably reflecting species-specific allocation patterns to defense following damage.
In contrast to expected trade-offs, we found positive relationships among some growth and resistance traits. However, the positive relationship between growth rate prior to damage and mechanically induced chemical resistance was weaker under previous exposure to elevated CO 2 . Our data add to a growing body of work that demonstrates the complex nature of plant growth and resistance relation- This finding supports previous studies that have reported negative relationships between milkweed growth and cardenolide production (Hochwender et al., 2000;Tao et al., 2016;Züst, Rasmann, & Agrawal, 2015). However, ours is the first study within the milkweed system to show interspecific differences in regrowth tolerance and mechanically induced resistance relationships following damage.
Interestingly, previous exposure to elevated CO 2 had no effect on   likely as a result of suppressed phytohormonal signaling pathways (Guo et al., 2012;Ode et al., 2014).
Despite finding no influence of elevated CO 2 on three of the four relationships between growth and resistance in our study, elevated CO 2 altered aspects of both milkweed growth and resistance independently. Notably, elevated CO 2 uncoupled the positive relationship between initial plant growth rate and regrowth tolerance following damage (Figure 3). Often plants with high innate growth rates can regrow faster following damage (Rosenthal & Kotanen, 1994).
However, in our study, those plants that were fast growing under elevated CO 2 did not maintain a proportionately high level of regrowth TA B L E 4 ANOVA tables of linear mixed effects models describing the relationships between the diversity of constitutive and induced cardenolides and growth rates dependent on milkweed species and elevated CO 2  (Woodson, 1954). Critically, the composition of cardenolide communities produced by milkweed can alter monarch interactions with natural enemies, such as a prevalent protozoan pathogen (Decker, Roode, & Hunter, 2018;Decker, Soule, Roode, & Hunter, 2019;Sternberg et al., 2012). Given the conservation importance of roadside milkweed patches that are regularly mowed throughout N. America, changes in regrowth tissue chemical quality could have implications for monarch populations. Yet, attempts to predict how migratory monarchs that depend on roadside milkweed corridors will perform under global environmental change remain challenging (Zipkin, Ries, Reeves, Regetz, & Oberhauser, 2012).
Our study, though comprehensive in its investigation of growth and chemical resistance before and after damage, does not incorporate the entire suite of defenses expressed by milkweeds.
Additional direct and indirect defenses include trichomes, latex, leaf toughness, and volatile emissions that attract natural enemies (Agrawal & Fishbein, 2006;Agrawal & Konno, 2009;Hochwender et al., 2000;Meier & Hunter, 2019;Zalucki et al., 2001). This suite of defense strategies may also generate resource-based trade-offs and alter plant-herbivore interactions (Züst & Agrawal, 2017;Züst et al., 2015). Thus, further studies exploring the fitness costs of regrowth tolerance and multiple defenses under future environmental conditions, and the responses of herbivore populations to these changes, are greatly needed.
On an evolutionary timescale, the influence of resource clines has illustrated the existence of trade-offs between growth and resistance, lending broad support to the RAH (Coley & Chapin, 1985;Endara & Coley, 2011;Strauss & Agrawal, 1999).
Currently, no well-established theory makes predictions about how trade-offs among defense traits will respond to rapid environmental change within one generation. In our study, the identity of the milkweed species determined our ability to detect a trade-off between regrowth tolerance and resistance following mechanical damage, and previous exposure to elevated CO 2 weakened a positive relationship between innate growth rate and constitutive defense. Given the rapid rate of environmental change predicted globally (Stocker et al., 2013), studies measuring the rate of plant resistance and growth evolution as well as which environmental change drivers are crucial determinants of plant fitness will be vital to predicting plant-insect interactions. This knowledge can be used to inform policy decisions which reduce the use of pesticides (Strauss & Murch, 2004) and improve weed control programs (Williams, Walsh, & Boydston, 2004).

CO N FLI C T O F I NTE R E S T
Authors have no sources of conflict of interest.

AUTH O R CO NTR I B UTI O N S
LED and MDH designed the experiment, collected and analyzed the data. LED wrote the manuscript, the MDH contributed significantly to drafts and approved the final version.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available in the Dryad Digital Repository: https://doi. org/10.5061/dryad.v6wwp zgs3