Rapid evolution during climate change: demographic and genetic constraints on adaptation to severe drought.

Populations often vary in their evolutionary responses to a shared environmental perturbation. A key hurdle in building more predictive models of rapid evolution is understanding this variation-why do some populations and traits evolve while others do not? We combined long-term demographic and environmental data, estimates of quantitative genetic variance components, a resurrection experiment and individual-based evolutionary simulations to gain mechanistic insights into contrasting evolutionary responses to a severe multi-year drought. We examined five traits in two populations of a native California plant, Clarkia xantiana, at three time points over 7 years. Earlier flowering phenology evolved in only one of the two populations, though both populations experienced similar drought severity and demographic declines and were estimated to have considerable additive genetic variance for flowering phenology. Pairing demographic and experimental data with evolutionary simulations suggested that while seed banks in both populations probably constrained evolutionary responses, a stronger seed bank in the non-evolving population resulted in evolutionary stasis. Gene flow through time via germ banks may be an important, underappreciated control on rapid evolution in response to extreme environmental perturbations.

Pollinator-mediated interactions in experimental arrays vary with neighbor identity.

PREMISE OF THE STUDY: Local ecological conditions influence the impact of species interactions on evolution and community structure. We investigated whether pollinator-mediated interactions between coflowering plants vary with plant density, coflowering neighbor identity, and flowering season. METHODS: We conducted a field experiment in which flowering time and floral neighborhood were manipulated in a factorial design. Early- and late-flowering Clarkia unguiculata plants were placed into arrays with C. biloba neighbors, noncongeneric neighbors, additional conspecific plants, or no additional plants as a density control. We compared whole-plant pollen limitation of seed set, pollinator behavior, and pollen deposition among treatments. KEY RESULTS: Interactions mediated by shared pollinators depended on the identity of the neighbor and possibly changed through time, although flowering-season comparisons were compromised by low early-season plant survival. Interactions with conspecific neighbors were likely competitive late in the season. Interactions with C. biloba appeared to involve facilitation or neutral interactions. Interactions with noncongeners were more consistently competitive. The community composition of pollinators varied among treatment combinations. CONCLUSIONS: Pollinator-mediated interactions involved competition and likely facilitation, depending on coflowering neighbor. Experimental manipulation helped to reveal context-dependent variation in indirect biotic interactions.

Effects of range-wide variation in climate and isolation on floral traits and reproductive output of Clarkia pulchella.

PREMISE OF THE STUDY: Plant mating systems and geographic range limits are conceptually linked by shared underlying drivers, including landscape-level heterogeneity in climate and in species' abundance. Studies of how geography and climate interact to affect plant traits that influence mating system and population dynamics can lend insight to ecological and evolutionary processes shaping ranges. Here, we examined how spatiotemporal variation in climate affects reproductive output of a mixed-mating annual, Clarkia pulchella. We also tested the effects of population isolation and climate on mating-system-related floral traits across the range. METHODS: We measured reproductive output and floral traits on herbarium specimens collected across the range of C. pulchella. We extracted climate data associated with specimens and derived a population isolation metric from a species distribution model. We then examined how predictors of reproductive output and floral traits vary among populations of increasing distance from the range center. Finally, we tested whether reproductive output and floral traits vary with increasing distance from the center of the range. KEY RESULTS: Reproductive output decreased as summer precipitation decreased, and low precipitation may contribute to limiting the southern and western range edges of C. pulchella. High spring and summer temperatures are correlated with low herkogamy, but these climatic factors show contrasting spatial patterns in different quadrants of the range. CONCLUSIONS: Limiting factors differ among different parts of the range. Due to the partial decoupling of geography and environment, examining relationships between climate, reproductive output, and mating-system-related floral traits reveals spatial patterns that might be missed when focusing solely on geographic position.

Seasonal changes in physiological performance in wild Clarkia xantiana populations: Implications for the evolution of a compressed life cycle and self-fertilization.

PREMISE OF THE STUDY: One explanation for the evolution of selfing, the drought escape hypothesis, proposes that self-fertilization may evolve under conditions of intensifying seasonal drought as part of a suite of traits that enable plants to accelerate the completion of their life cycle, thereby escaping late-season drought. Here, we test two fundamental assumptions of this hypothesis in Clarkia xantiana: (1) that a seasonal decline in precipitation causes an increase in drought stress and (2) that this results in changes in physiological performance, reflecting these deteriorating conditions. METHODS: We examined seasonal and interannual variation in abiotic environmental conditions (estimated by ambient temperature, relative humidity, predawn leaf water potentials, and carbon isotope ratios) and physiological traits (photosynthesis, conductance, transpiration, instantaneous water-use efficiency, ascorbate peroxidase and glutathione reductase activities, quantum yield of photosystem II, PSII potential efficiency) in field populations of C. xantiana in 2009 and 2010. KEY RESULTS: In both years, plants experienced intensifying drought across the growing season. Gas exchange rates decreased over the growing season and were lower in 2009 (a relatively dry year) than in 2010, suggesting that the temporal changes from early to late spring were directly linked to the deteriorating environmental conditions. CONCLUSIONS: Seasonal declines in transpiration rate may have increased survival by protecting plants from desiccation. Concomitant declines in photosynthetic rate likely reduced the availability of resources for seed production late in the season. Thus, the physiological patterns observed are consistent with the conditions required for the drought escape hypothesis.

The geography of demography: long-term demographic studies and species distribution models reveal a species border limited by adaptation.

Potential causes of species' geographic distribution limits fall into two broad classes: (1) limited adaptation across spatially variable environments and (2) limited opportunities to colonize unoccupied areas. Combining demographic studies, analyses of demographic responses to environmental variation, and species distribution models, we investigated the causes of range limits in a model system, the eastern border of the California annual plant Clarkia xantiana ssp. xantiana. Vital rates of 20 populations varied with growing season temperature and precipitation: fruit number and overwinter survival of 1-year-old seeds declined steeply, while current-year seed germination increased modestly along west-to-east gradients in decreasing temperature, decreasing mean precipitation, and increasing variation in precipitation. Long-term stochastic finite rate of increase, λ(s), exhibited a fourfold range and varied among geologic surface materials as well as with temperature and precipitation. Growth rate declined significantly toward the eastern border, falling below 1 in three of the five easternmost populations. Distribution models employing demographically important environmental variables predicted low habitat favorability beyond the eastern border. Models that filtered or weighted population presences by λ(s) predicted steeper eastward declines in favorability and assigned greater roles in setting the distribution to among-year variation in precipitation and to geologic surface material. These analyses reveal a species border likely set by limited adaptation to declining environmental quality.

Experimental studies of adaptation in Clarkia xantiana. II. Fitness variation across a subspecies border.

Because the range boundary is the locale beyond which a taxon fails to persist, it provides a unique opportunity for studying the limits on adaptive evolution. Adaptive constraints on range expansion are perplexing in view of widespread ecotypic differentiation by habitat and region within a species' range (regional adaptation) and rapid evolutionary response to novel environments. In this study of two parapatric subspecies, Clarkia xantiana ssp. xantiana and C. x. ssp. parviflora, we compared the fitness of population transplants within their native region, in a non-native region within the native range, and in the non-native range to assess whether range expansion might be limited by a greater intensity of selection on colonists of a new range versus a new region within the range. The combined range of the two subspecies spans a west-to-east gradient of declining precipitation in the Sierra Nevada of California, with ssp. xantiana in the west being replaced by ssp. parviflora in the east. Both subspecies had significantly higher fitness in the native range (range adaptation), whereas regional adaptation was weak and was found only in the predominantly outcrossing ssp. xantiana but was absent in the inbreeding ssp. parvifilora. Because selection intensity on transplants was much stronger in the non-native range relative to non-native regions, there is a larger adaptive barrier to range versus regional expansion. Three of five sequential fitness components accounted for regional and range adaptation, but only one of them, survivorship from germination to flowering, contributed to both. Flower number contributed to regional adaptation in ssp. xantiana and fruit set (number of fruits per flower) to range adaptation. Differential survivorship of the two taxa or regional populations of ssp. xantiana in non-native environments was attributable, in part, to biotic interactions, including competition, herbivory, and pollination. For example, low fruit set in ssp. xantiana in the east was likely due to the absence of its principal specialist bee pollinators in ssp. parviflora's range. Thus, convergence on self-fertilization may be necessary for ssp. xantiana to invade ssp. parviflora's range, but the evolution of outcrossing would not be required for ssp. parviflora to invade ssp. xantiana's range.

The joint evolution of mating system, floral traits and life history in Clarkia (Onagraceae): genetic constraints vs. independent evolution.

Genetic correlations caused by pleiotropy or linkage disequilibrium may influence the joint evolution of multiple traits as populations or taxa diverge. The evolutionary transition from outcrossing to selfing has occurred numerous times and is often accompanied by phenotypic and genetic changes in multiple traits such as flower size, pollen-ovule ratio, stigma and anther maturity and the age of reproductive maturity. Determining whether the recurring patterns of multitrait change are because of selection on each trait independently and/or the result of genetic correlations among traits can shed light on the mechanism that accounts for such convergence. Here, we evaluate whether floral traits are genetically correlated with each other and/or with whole-plant traits within- and between-populations and taxa. We report results from a greenhouse study conducted on two pairs of sister taxa with contrasting mating systems: the autogamously selfing Clarkia exilis and its predominantly outcrossing progenitor C. unguiculata and the autogamous Clarkia xantiana ssp. parviflora and its outcrossing progenitor C. xantiana ssp. xantiana. We examined variation within and covariation among maternal families in three populations of each taxon with respect to the age at first flower, the rate of successive flower production and the number of days between bud break and anther dehiscence and stigma receptivity within individual flowers. Based on phenotypic divergence between sister taxa, bivariate regressions, correlations among maternal family means and analysis of covariance (ancova), we did not find unilateral support indicating that genetic constraints govern the joint distribution of floral and whole-plant traits.

Formation of monoterpenes in Antirrhinum majus and Clarkia breweri flowers involves heterodimeric geranyl diphosphate synthases.

The precursor of all monoterpenes is the C10 acyclic intermediate geranyl diphosphate (GPP), which is formed from the C5 compounds isopentenyl diphosphate and dimethylallyl diphosphate by GPP synthase (GPPS). We have discovered that Antirrhinum majus (snapdragon) and Clarkia breweri, two species whose floral scent is rich in monoterpenes, both possess a heterodimeric GPPS like that previously reported from Mentha piperita (peppermint). The A. majus and C. breweri cDNAs encode proteins with 53% and 45% amino acid sequence identity, respectively, to the M. piperita GPPS small subunit (GPPS.SSU). Expression of these cDNAs in Escherichia coli yielded no detectable prenyltransferase activity. However, when each of these cDNAs was coexpressed with the M. piperita GPPS large subunit (GPPS.LSU), which shares functional motifs and a high level of amino acid sequence identity with geranylgeranyl diphosphate synthases (GGPPS), active GPPS was obtained. Using a homology-based cloning strategy, a GPPS.LSU cDNA also was isolated from A. majus. Its coexpression in E. coli with A. majus GPPS.SSU yielded a functional heterodimer that catalyzed the synthesis of GPP as a main product. The expression in E. coli of A. majus GPPS.LSU by itself yielded active GGPPS, indicating that in contrast with M. piperita GPPS.LSU, A. majus GPPS.LSU is a functional GGPPS on its own. Analyses of tissue-specific, developmental, and rhythmic changes in the mRNA and protein levels of GPPS.SSU in A. majus flowers revealed that these levels correlate closely with monoterpene emission, whereas GPPS.LSU mRNA levels did not, indicating that the levels of GPPS.SSU, but not GPPS.LSU, might play a key role in regulating the formation of GPPS and, thus, monoterpene biosynthesis.