1 | Introduction
Suburban landscapes consist of a mix of human-created infrastructure and fragmented natural communities. In the temperate deciduous forest biome of northeastern North America, small forest stands are common components of the many parks, preserves, and private holdings in suburban areas. In conjunction with the lawns and fields that also commonly occur in suburbia, this landscape provides ideal habitat for white-tailed deer (Odocoileus virginianus ) . Hunting is very limited in suburbia; consequently, deer densities can be extremely high . The influence of high deer pressure on suburban forest species is of particular interest because of the huge extent of urbanizing landscapes; suburban forests now contain a large share of many regions’ biodiversity .
Many consequences of very high deer densities for forest plants and their natural communities are well documented. For example, herbivory by deer causes tissue loss that can lead to increased mortality rates or decreased reproduction; they may eat entire plants in the form of seedlings or seeds like oak acorns; seedlings can be trampled by deer, and deer can otherwise disturb the forest floor as they move across it and bed down at night (reviewed in . These effects vary with the level of deer browse pressure and among plant species . In very severely browsed forests, there may be extremely low abundance of any plants below the browse line, which increases sunlight penetration and recruitment opportunities for less shade tolerant plants, and has a cascade of other indirect effects in the forest community . Defense and stress responses of long-lived, woody plant species also are likely to be strongly affected by high deer browse pressure, but this possibility has not been well studied. Here, we compared the metabolomic profiles of indigenous and nonindigenous woody species that had been exposed to or protected from deer in a forest community in suburban New Jersey, USA.
Woody plants have characteristics particularly attractive and vulnerable to deer. They are chronically exposed to repeated browsing, they have green foliage throughout the growing season or even year-round for evergreens, and their buds provide highly nutritious forage throughout the winter. Thus, high deer browse pressure is associated with depletion of the woody component of forest understories .
Not surprisingly, woody species invest in various mechanical and chemical defenses that deter browsing . Deer browse has been demonstrated to be greater when defense levels are lower , and induced defense responses can occur within individuals in response to deer browse . Additionally, the physiological costs of producing defenses against browse can affect plant fitness , resulting in evolutionary responses at the population level over time . In some instances, on the other hand, plant chemical defense may be unaffected by deer browse or may even decrease . In suburban forests, where deer browse pressure is particularly high, effects on chemical defenses of long-lived woody plants are likely strong and, as demonstrated for the more commonly studied woody plant defenses against insect herbivory , defenses against deer browse could also influence allocation of resources to plant growth and reproduction .
Advances in metabolomic research have increasingly enabled investigation of metabolome-wide responses of plants to stressors , with growing application of metabolomics to plants in natural ecological communities . There has been limited ecometabolomic research on woody plants so far, but the studies have been wide-ranging, including investigations of abiotic drivers of population or species variation in plant chemistry , the role of plant defenses in phylogenetic diversification , metabolomic niche differentiation among sympatric tropical species , ontological variation in foliar chemistry , and disease resistance . However, there has been very little attention paid to plant metabolomics associated with deer herbivory, with just one study, showing that white-tailed deer browse less frequently on nonindigenous invasive plants that are chemically dissimilar to indigenous plants in the community .
Significant quantitative and qualitative variations are commonly observed in defense responses of different plant species; the final defense phenotypes result from genetic factors and/or environmental (light, nutrient availability, geography, etc.) gradients . Additionally, the type of defense strategy that a plant species employs is optimized by the severity of herbivore pressure and the feeding mode of the attacking herbivore (chewing herbivores vs piercing-sucking herbivores) , as well as the plant’s mode of growth; i.e. annuals versus perennials or fast-growing versus slow-growing species . Therefore, variation among species’ chemical defense responses to abundant deer is expected, and could have important consequences for forest community structure. Species that are less preferred or even avoided by deer altogether may have an ecological advantage beyond avoidance of tissue loss. Many secondary metabolites, especially those rich in nitrogen, are costly for plants to produce , so if upregulation of defense-related chemistry is a consequence of heavy, chronic deer browsing, browsed plants could experience added stress. If severe enough, these metabolic costs could contribute to shifts in species dominance within the community. Since some nonindigenous invasive plant species are less preferred by deer compared to indigenous species , it is even possible that costly, metabolome-wide responses to deer in browsed plants could facilitate invasion by nonindigenous plant species that deer avoid. Indeed, deer are acknowledged as invasion facilitators in some cases , but the possible role of metabolomic responses to deer in this facilitation has not been considered until recently . Despite the multiple hypotheses put forward to explain the variability in defense phenotypes in plant communities and the growth-defense tradeoff , studies that compare the global metabolomic responses and/or constitutive/induced defense responses of indigenous and nonindigenous species are lacking.
A first step in addressing the consequences for plant communities of metabolome-wide responses to high deer pressure is to document the metabolomic profiles of a variety of species in a community. In our study, we tested four long-lived, woody or semi-woody species that occur in one forest preserve, both in plots that had been fenced for over five years to exclude deer and in plots that were open to deer. Two species are indigenous (Nyssa sylvatica and Lindera benzoin ), and two are nonindigenous, invasive species (Rosa multiflora andEuonymus alatus ). The forest is situated in a suburban landscape (more information given in the Materials and Methods section) with very high deer density, and all four species experience deer browse, but at different rates.
We predicted that fenced and unfenced plants would differ in their global metabolite profiles and signaling pathways involved in defense. We also expected that the metabolomic divergence between fenced and unfenced plants would be more pronounced for species that are more highly preferred and affected by deer.