Principal Investigators Sharon Collinge, Biology and Environmental Studies Program, University of Colorado Paul Gepts, Dept. of Agronomy and Range Science, University of California at Davis Rick Grosberg, Section of Evolution and Ecology, University of California at Davis Susan Harrison, Dept. of Environmental Science and Policy, University of California at Davis Kevin Rice, Dept. of Agronomy and Range Science, University of California at Davis Maureen Stanton, Section of Evolution and Ecology, University of California at Davis Denise Thiede, Center for Population Biology, University of California at Davis

• Kevin J. Rice, Dept. of Agronomy and Range Science, UC-Davis
• Phone # (530) 752-8529
• FAX# (530) 752-4361
• e-mail: kjrice@ucdavis.edu
• Maureen Stanton, Section of Evolution and Ecology, UC-Davis
• Phone # (530) 752-2405
• FAX# (530) 752-1449
• e-mail: mlstanton@ucdavis.edu

Project overview

Recent technological advances, ranging from remote sensing to DNA chips, have fostered unprecedented advances in biology. However, these advances have yet to be integrated and applied to a question of fundamental interest to plant population and conservation biology--- How is biodiversity shaped and maintained by ecological and evolutionary forces interacting over a range of spatial and temporal scales? A complete understanding of the genetics of adaptation requires not only information on genome structure but also how differences in gene expression affect population fitness. Similarly, the long-term effects of habitat fragmentation on plant community structure may depend critically on the evolutionary capacity of resident species to respond to changes in gene flow and selection. Even so, conservation biology has tended to dichotomize evolutionary and ecological approaches, and to deal separately with population-scale and landscape-scale. Given the powerful tools that have recently become available, the field is now poised to make major progress in understanding and conserving plant biodiversity. This will require multidisciplinary teams of scientists, working on scales ranging from molecules to landscapes, and dedicated to focusing on the linkages between their disciplines.

In nature, plant diversity is shaped by ecological and evolutionary processes that are played out in spatially complex environments. On an ecological time-scale, species diversity is maintained by the spatial interplay between immigration, competition, and local extinction. Over evolutionary time, species diversity arises through the divergence of lineages; this spatially dependent process is determined by the interplay between natural selection, gene flow, and genetic drift. However, studies of plant populations and communities rarely examine how the spatial complexity of the environment affects these processes. The sizes, shapes and juxtaposition of different types of habitat determine the spatial distribution of individuals and selective forces, and thus these factors should greatly influence the balance between natural selection and gene flow, as well as between extinction and recolonization. Experiments conducted in small garden plots or other artificially homogeneous environments cannot disentangle these dynamics. Instead, understanding how landscape structure affects these opposing processes will require conducting experiments at spatial scales over which natural habitat variation actually occurs. An important goal of our collaborative project is to understand the role of landscape structure in plant adaptation.

Our collaboration integrates across three traditionally separate fields of study. Landscape ecologists are documenting the distribution of appropriate habitats, the nature of the boundaries between habitats, and how these patterns determine the isolation or connectedness of plant populations. Their description of physical and biotic components of the landscape will allow the population biologists on the team to conduct experiments characterizing the structure of the selective environment, by determining how local heterogeneity favors different traits. Together, these landscape and population studies document the likelihood of immigration and the strength of selection on immigrant genotypes. The spatial genetic structure of populations reflects the historical balance between these processes. To complete the linkages, molecular and ecological geneticists are planning to measure genetic similarity among populations and across substrate boundaries to infer the longer-term consequences of landscape structure for genetic drift, selection, and gene flow. Working collaboratively, we hope to dissect the linkages between landscape structure, the structure of plant populations and communities, and the forces that engender, maintain, or diminish plant biodiversity over ecological and evolutionary time scales. This information can guide efforts to conserve and restore plant biodiversity in the face ofincreasing habitat fragmentation and environmental change.

To foster a common research vision and focus, we propose a model system approach. Just as molecular and population geneticists have used model organisms (e.g., Arabidopsis) to address the most challenging questions in genetics, we will use a model ecosystem to understand spatial patterns and dynamics of plant biodiversity across multiple spatial and temporal scales.

The Model Ecosystem -- Our project is centered in the Donald and Sylvia McLaughlin Reserve, a new UC Natural Reserve System (UCNRS-provide link) site with striking physical gradients in microclimate, natural and human-caused substrate mosaics, and diverse plant communities that contain both rare species and introduced weeds. The McLaughlin Reserve is located at the junction of Napa, Lake, and Yolo Counties, in north-central California. The 7,000-acre reserve encompasses two major geologic formations (Franciscan and Great Valley), two watersheds (Putah and Cache Creeks) and four major vegetation types (oak woodland, nonserpentine chaparral, serpentine chaparral, and grasslands). Overlain on this natural substrate variation is a mosaic of human land use, yielding grazed and ungrazed grasslands, relatively pristine habitats and reclaimed mining areas.

Although a very new part of the UCNRS, McLaughlin provides a unique field laboratory for our project because of unusually comprehensive, long-term monitoring that is ongoing at the site. The environmental database for the reserve includes extensive baseline surveys, GIS maps, and 15 years of monitoring records on the site's geology, hydrology, climate, air and water quality, vegetation, rare plants, aquatic ecology, and wildlife.

A key component of McLaughlin's heterogenous ecosystem, and a focus for our collaborative research effort, is soil derived from serpentine (ultramafic) rock. Serpentine soils are high in magnesium and iron, yet often poor in macro-nutrients. Although serpentine soils are hostile for the growth and survival of many plant species, these habitats are of special interest in conservation because they harbor many of the endemic and rare plants that are native to California. Indeed, the McLaughlin Reserve lies within one of the world's richest regions for serpentine-endemic plants. Serpentine soils on the reserve support extensive chaparral communities, with several shrub endemics and 15-20 endemic herbs. Summer-moist habitats on serpentine support rare plants such as swamp larkspur, serpentine sunflower, bare monkeyflower, and Cleveland's butterweed. Non-serpentine areas of the Reserve include extensive blue oak woodland, with an understory of introduced annual grasses but also containing many native herbaceous species.

At McLaughlin, the boundaries between serpentine and non-serpentine-associated plant communities are often dramatic and abrupt, creating a striking spatial template that we can use to understand how landscape structure affects ecological and evolutionary processes. Serpentine landscape is structured on two spatial scales in this system. Serpentine-associated plant populations may occupy either a gradient in which ecological conditions change over a short distance, or may be subdivided among widely spaced, discrete patches. We expect the balance between gene flow, selection, extinction, and recolonization to differ between these extremes. Accordingly, we expect that strategies for conserving plant biodiversity will differ between types of landscape structure. Along the margins of serpentine outcrops, there is often an abrupt transition between markedly different soils and plant communities. This edge environment is ideal for addressing how gradients in biotic and abiotic conditions influence the balance between diversifying selection and homogenizing gene flow. Other serpentine communities, such as the isolated spring-seeps that occur within serpentine outcrops, consist of small discrete patches. These seeps and their specialized flora are an excellent system for examining how small population size, spatial isolation, and turnover (local extinction and recolonization) shape the dynamics and genetic diversity of plant metapopulations.

I) Edges and gradients-- Plant diversity and adaptation in a heterogeneous landscape

Landscape ecology focuses on spatial components of physical and biotic heterogeneity. As such, it is remarkably compatible with the study of selection and gene flow, the premier factors influencing the genetic architecture of plant species. However, we know little about the effects of complex landscape patterns on both selection and gene flow. In particular, we have a very poor understanding of how physical selective agents (such as soil characteristics) and biotic selective factors (such as competition) interact to shape adaptive responses in plant populations within spatially variable environments. Theory predicting that evolutionary responses should depend strongly on whether populations are increasing (source populations) or decreasing (sink populations) remains largely untested. The McLaughlin reserve provides a perfect outdoor laboratory for such analysis with its fine-scale mosaic of dramatically different microhabitats, including moist seeps, serpentine meadows and reclaimed mining areas. Usually lacking in previous field studies on plant microevolution is the accurate delineation of the "selective topography". Most studies on adaptation assume very simplistic selective gradients and ignore the typically complex mosaic nature of selection forces in heterogeneous landscapes. By combining the expertise and research efforts of landscape ecologists, plant evolutionary biologists and molecular geneticists, we are integrating a realistic characterization of spatial variation in physical and biotic selection with a rigorous experimental determination of adaptive response. This analysis of selective topography and adaptive responses, when coupled with a molecular marker analysis of gene flow, will produce a spatially explicit site map of expected fitness differentials that reflect the balance between selection and gene flow. By using experimental reciprocal transplants and demographic monitoring described below, we will then test the validity of these "fitness maps". Finally, for a truly comprehensive hierarchical analysis of plant microevolution, it is critical to understand the genetic basis of any adaptive response. Using molecular linkage mapping, we will examine the possibility that if we detect any rapid adaptive change, it may reflect the action of relatively few genes of major phenotypic effect (i.e. quantitative trait loci or QTLs).

A) Adaptation to physical stress across serpentine soil boundaries

To characterize the complex patchwork of serpentine soil edges and boundaries, we plan to 1) map this soil mosaic using a global positioning system (GPS), 2) incorporate field data on soil characteristics into a geographic information system (GIS), and 3) use geostatistical techniques to quantify the spatial structure of substrate variation. After we have gained an understanding of the spatial structure in substrate variation associated with the transition from serpentine to non-serpentine soils, we then can ask focused questions about how plants adapt to such variation. A primary question is whether plants experience diversifying selection along serpentine/non-serpentine edges. That is, are different traits favored on the two soil types? Several observations suggest that serpentine soils may exert unique selection pressures on plants, compared with other soil types. First, serpentine soils are characterized by unusually high levels of magnesium and often contain heavy metals; both features make these soils toxic to non-adapted plants. Second, plant community composition often changes dramatically across boundaries between serpentine and non-serpentine soils. Third, a number of plant species native to California (e.g. Clarkia gracilis have differentiated into distinct subspecies that are found either on or off of serpentine substrates. Despite this compelling evidence, patterns of natural selection have never been contrasted between serpentine and non-serpentine habitats. The need to do so is urgent, given the recent arrival of Bromus hordeaceus, a species that is currently colonizing serpentine substrates and may threaten native endemics.

To learn about consequences of gene flow and local adaptation for diversification in response to soil type, we plan to do comparative analysis of plant species with different mating systems and natural distributions. Using outcrossing native herbs (e.g. Clarkia concinna, C. purpurea, C. gracilis), partially selfing native herbs (e.g. Collinsia sparsiflora, Gilia tricolor) and self-pollinating, invasive grasses (e.g. Bromus hordeaceus, ...) as focal species, we will characterize how patterns of selection actually change across serpentine/non-serpentine boundaries. To this end, we will use phenotypic selection gradient analysis, in concert with controlled matings and reciprocal transplants between serpentine and non-serpentine locations. Seeds from genetically distinct lineages will be transferred between locations along the gradient will demonstrate how selection acts on migrant genotypes. Reciprocal crosses between individuals on opposite sides of the boundary will allow us to assess selection on ecotypic "hybrids" that could be generated by pollen movement. Two generations of crosses incorporating migrants and hybrids will reveal how maternal responses to the local environment influence the pattern of selection on immigrant and native genotypes. Previous studies of adaptation to stress by Stanton and Thiede suggest that serpentine substrates will favor earlier flowering, relative to non-serpentine substrates. Disruptive selection on flowering time could restrict gene flow between soil types and promote genetic differentiation in an out-crossing species. In selfing species, gene flow by pollen should be restricted, even between nearby habitats. Should we identify distinct ecotypes adapted to serpentine and non-serpentine soils, we will work to develop QTL markers for genes important to edaphic differentiation.

In companion studies, we will focus on recently invading species to examine the potential for rapid evolution of serpentine tolerance. Self-fertilization minimizes gene flow via pollen, so migrantts dispersing in by seed will contribute most to gene flow across the boundary. Previous work by Rice and Harrison suggests that landscape structure may influence the likelihood of adaptation to serpentine. While sharp edges between fertile soils and "hard" serpentine patches may facilitate dispersal, they also create an adaptational barrier that is difficult to overcome. In contrast, more gradual transitions between soil types may promote the invasion of serpentine habitat by acting as intermediate "breeding grounds" for genotypes tolerant of serpentine. Our transplant experiment will determine how the sharpness of the soil gradient affects the speed and direction of plant adaptation.

The low organic content of many serpentine soils can reduce water retention capacity, potentially exposing plants on those soils to late-season drought stress in California's Mediterranean climate. We will examine physiological mechanisms of adaptation to water stress by using stable isotope technology and pressure chamber techniques to estimate water-use efficiency, examine plant water relations, and determine soil water availability. Based on our previous studies, we predict that increased root allocation and increased seed size will occur in populations adapted to serpentine. The strength of selection necessary to produce such adaptive changes depends on the rate of gene flow across the boundaries between serpentine and non-serpentine soils. Molecular markers that are neutral to selection will provide this critical information on the patterns and rates of gene flow within the serpentine mosaic.

Conservation and restoration application - A better understanding of how environmental heterogeneity affects genetic structure of plant populations is needed for effective conservation and restoration efforts. Our interdisciplinary study will determine how landscape mosaics affect plant biodiversity on both ecological and evolutionary time scales. Our results will shed light on why the diversity of native plants is so high on altered soil types like serpentine, and how invasion of such sites by non-native plants can be prevented.

B) Adaptation to interspecific competition across productivity gradients

Gradients of soil productivity, interacting with interspecific competition, form one of the primary forces shaping plant biodiversity. However, ecologists disagree about the relative importance of competition and abiotic stress in controlling plant distributions and in producing adaptation. This current debate creates a context within which to revisit classical work on plant adaptation to serpentine using modern conceptual and analytical approaches. Serpentine soils tend to be far less fertile than nearby non-serpentine soils. In tandem field surveys of plant community diversity, plant density, herbivory, and disturbance, we will characterize the way in which biotic factors map onto the physical features of the transition between serpentine and non-serpentine soils. Data on plant diversity and density at these edges will allow us to characterize each boundary as abrupt or gradual, an attribute that will influence adaptation within populations. In addition, quantitative data on community structure will characterize potential changes in competition across the gradient. Understanding the nature of this competitive environment is central to characterizing the nature of divergent selection across the boundary between serpentine and non-serpentine soils.

Since nutrient enrichment is a widespread aspect of environmental change, there are strong practical implications to understanding how species richness and stability respond to productivity gradients. However, new theory in this area has yet to be tested in natural systems. In particular, there has been no coherent attempt to understand how changes in the spatial distribution of species diversity at the landscape level influences the evolution of component species. For example, one might expect genetic differentiation in competitive ability if species clusters are stable in space and time, while phenotypic plasticity is more likely to evolve if the competitive environment is more unpredictable.

We propose to use existing serpentine gradients and vegetation patch structure to independently manipulate substrate and community diversity in order to address these basic gaps in our knowledge. As a first step, we plan to document the spatial architecture of species diversity across substrate gradients of soil productivity. Special attention will be paid to how patterns in plant biological diversity are affected by landscape variables, such as patch size, shape, isolation, and landscape context. Complementing this community-based approach, careful monitoring will show how changes in the spatial arrangement of species result in concomitant changes in the patterns of selection acting on focal plant species. Simultaneously, we will measure how these changes in community and population structure affect patterns of gene flow and thus the likelihood of local genetic differentiation. Although even the most basic texts in evolution emphasize the importance of interspecific competition as a primary agent of natural selection, we have less than a handful of studies in natural plant populations that have really examined adaptation to competition. Of these few studies, none have combined experimental studies of adaptive response and gene flow with the detailed spatial analysis of the competitive regime that we propose.

Conservation and restoration application - We anticipate that the most important management lesson from this part of the study will be that biodiversity at the landscape level and biodiversity at the genetic level are inextricably linked. Patterns of genetic diversity in the component species determine the long-term viability of the plant community, and can be used to predict community-level responses to environmental change.

C) Identifying genetic mechanisms of plant adaptation

A long-standing debate in evolutionary biology has revolved around the genetic basis for adaptive phenotypic variation. Specifically, how many genes control expressed variation in adaptive traits? Recent molecular studies of crops have shown that relatively few genes of large effect (Quantitative Trait Loci, or QTLs) control phenotypic variation. By comparison there is very little information on QTLs in wild species, and there have been no integrated analyses of the fitness ramifications of phenotypic variation under QTL control. The speed with which plants can adapt to novel selection regimes depends, in part, on whether adaptation is conferred by many genes of small effect, or by a few genes of large effect.

Although very little is known about the genetic basis for adaptation to serpentine soils, some lessons can be taken from studies of mine waste areas. In some plants, tolerance to toxic heavy metals such as copper, lead, or zinc is controlled by 1-3 genes of major effect. Species that already contain rare tolerance alleles at these loci can rapidly adapt to soils contaminated with heavy metals, whereas species lacking these alleles rarely occur on such soils. Thus, the genome architecture of adaptation plays an important role in determining a species' evolutionary potential.

Using the field studies to identify the spatial adaptive landscape, we will use molecular linkage mapping to examine the genetic basis for adaptation to spatial variation in soil type and interspecific competition. Linkage maps will be established in segregating generations created by crossing serpentine and non-serpentine ecotypes. To understand the role of QTLs for adaptation in heterogeneous environments, we will include traits that might be highly plastic in their phenotypic expression (e.g., seed number) as well as traits that might be expected to be more genetically "hardwired" (e.g., seed weight, flowering time, or water-use efficiency). If adaptation to serpentine is based on QTLs with major effect, adaptive shifts could take place rapidly. In addition, the boundary between serpentine and non-serpentine soils should be associated with steeper changes in allele frequencies for markers linked to selectively important loci. In contrast, finding many loci of small fitness effects would indicate that adaptation to serpentine soils is likely only after many generations of genetic recombination, mutation and selection. Our interdisciplinary collaboration will represent the most comprehensive study to date on the fitness consequences of QTL variation in natural plant populations.

Conservation and restoration application - If QTLs exert a major influence on adaptation to serpentine and competition, molecular markers tightly linked to QTLs could potentially be used to screen populations for potentially useful genetic variants. Molecular technology may provide the key for genetically informed conservation and restoration of native plant communities.

II) Evolutionary Dynamics of Metapopulations

Many species exist as small isolated populations in patchy habitats. Metapopulation theory predicts that the regional survival of such species depends on a balance between local colonization and extinction. The spatial isolation and temporal turnover of populations should also strongly influence genetic structure and evolutionary potential in patchily distributed species. By determining the size of suitable habitat patches and their spatial arrangement, the structure of the landscape influences the balance between extinction and re-colonization. We predict that small, isolated patches are likely to be genetically depauperate either because seed dispersal and pollen flow are restricted over long distances or because small size and limited immigration enhance demographic stochasticity.

To determine how landscape structure affects metapopulation dynamics, we will examine genetic structure and demographic stochasticity in one or more of the approximately 25 species that are endemic to serpentine seeps at McLaughlin Reserve. Previous work by Harrison, Maron and Huxel on 5 of these species (Helianthus exilis, Senecio clevelandi, Astragalus clevelandi, Delphinium uliginosum, and Mimulus nudatus) demonstrated that isolated populations were more prone to extinction and less likely to be re-colonized. Motivated by this baseline data, we will ask how demographic stochasticity and limited immigration interact to affect genetic structure in the metapopulation. To what extent does immigration alleviate the demographic and genetic problems associated with being an isolated or small population? We will use molecular markers to describe how historical patterns of immigration and extinction have shaped existing spatial genetic structure in the metapopulation, demographic monitoring to characterize annual variability in population growth, and experimental crosses and transplants between populations to examine the demographic and genetic consequences of immigration.

A) Landscape features and their effect on spatial genetic structure of metapopulations

Using nuclear and cytoplasmic DNA markers, we will characterize the spatial genetic structure for populations of serpentine seep specialists. We predict small or isolated populations will be genetically depauperate due to genetic drift or limited gene flow, respectively. Maternally inherited cytoplasmic markers, reflecting seed dispersal, will likely show stronger spatial structure than nuclear markers that are affected by both pollen and seed movement. Distance along stream drainages may be a better correlate of genetic similarity among populations than straight-line distances, especially for cytoplasmic markers. Such results would be consistent with ecological data suggesting that dispersal and colonization among seeps are infrequent, and represent important limitations on the distribution of the seep-specialist plant species. Using geostatistical analysis we will test whether landscape features alone can explain the pattern of genetic variability or whether additional abiotic features of seeps affect the pattern. Harrison's unique long-term data documenting newly founded populations allows us to test the prediction that recently colonized populations will contain only a fraction of the variability found in longer-established populations because founder effects deplete local genetic variability. An alternative hypothesis is that genetic variation in recently founded populations represents a non-random subset of the variation present in the metapopulation as a whole. This pattern may result either from selection for colonist genotypes or from colonists arriving from a single source population.

B) Demographic variability in a metapopulation landscape

Do smaller, isolated populations exhibit more demographic variability? Does the nature of the edge or the surrounding habitat influence immigration or extinction? Annual variability in rainfall is common in California and is likely to contribute substantially to demographic variability in serpentine seep communities. One short-lived seep specialist, Mimulus nudatus, shows dramatic year to year variability in aboveground population size and density in response to rainfall patterns. To characterize the nature of demographic variability associated with landscape attributes, we will follow the fates of individuals in multiple populations to estimate transition probabilities associated with particular life cycle stages and project a population's growth rate in multiple seasons. Geostatistical analysis will reveal what attributes of the landscape structure influence local demography. In addition, we will test the hypothesis that populations exhibiting more variable demography are genetically depauperate due to population bottlenecks. If this prediction is not met, it would suggest an important role for the seed bank in storing genetic variation, an issue we will examine in future work.

C) Demographic consequences of immigration: selection on seed and pollen immigrants

The two mobile stages in the life cycle, pollen movement and seed dispersal, have different ecological and genetic consequences. For example in small populations, pollen immigrants are likely to have enhanced success because they decrease pollen limitation or reduce the probability of biparental inbreeding, a phenomenon that can limit reproductive output in small populations. Alternatively, if seep populations exhibit local adaptation, then pollen immigrants may suffer decreased vigor, a phenomenon termed outbreeding depression. Experimental pollinations between populations will reflect how immigration can affect the genetic variability of nuclear markers by changing the transition probabilities associated with seed production and seedling vigor and altering population growth rates. In contrast to pollen movement, long-distance seed dispersal can serve to buffer small populations both genetically and numerically by directly affecting the number of potential recruits in the population. Reciprocal transplants of seeds will reveal whether immigrants increase population growth rate or whether local adaptation opposes the numerical and genetic consequences of long-distance seed dispersal. Experimental pollinations and reciprocal transplants will allow us to assess the demographic consequences of these two types of immigration and to what extent selection against immigrants affects the genetic structure among patches. These experiments, in concert with genetic marker studies, will allow us to place our demographic data in a metapopulation context. These analyses will add to the growing evidence for how genetic variation in patchily distributed species is shaped by the processes of extinction and re-colonization.

Conservation and restoration application - Many rare plants are locally abundant, but have small geographic ranges and inhabit naturally sparse, patchy habitats. Our study will shed light on the importance of conserving multiple, interconnected local populations in such species.

Summary --With support from the Packard and Mellon Foundations, we are using an interdisciplinary, model system approach to investigate the ecological and evolutionary mechanisms that promote plant biological diversity at multiple spatial and temporal scales, and to ask how these mechanisms are dynamically linked between scales. Our study will integrate fundamental advances with application of our results to conservation and restoration. The project will be used to train a new generation of students to integrate novel conceptual and technical approaches to studying diversity at multiple scales of biological organization.

Back to the McLaughlin Reserve Research Proposal Home Page.