TREEPEACE

FROM HOLOCENE TO ANTHROPOCENE: THE PACE OF MICROEVOLUTION IN TREES

TREEPEACE

Summary

TREEPEACE aimed at assessing evolutionary changes in European oaks as a result of past environmental changes at different time frames (holocene and anthropocene). The project explored different experimental settings in ancient and extant oak landscapes that were suitable for detecting evolutionary changes in time and space. Whenever technically feasible, changes were tracked at both the genomic and phenotypic levels. TREEPEACE was based on the rationale that the very high level of genetic diversity of oak species may generate substantial evolutionary changes under strong selection pressures, even in one generation.

We experimentally demonstrated that authentic ancient oak DNA can be extracted from subfossil and archaeological wood remains, and that chloroplast DNA is rather abundant and can be sequenced, while the recovery rate of nuclear DNA is lower and sequencing requires more efforts. We found that the ancient chloroplast haplotypes matched the most frequent haplotypes found in nearby extant populations, suggesting that there was no population replacement. Ongoing investigations regarding admixture in ancient and modern DNA will allow to decipher whether persistence occurred also at the nuclear genome, or whether gene flow between lineages contributed to significant population changes.

In parallel, by comparing extant oak populations stemming from different present climates, we found genomic footprints of population divergence, some of which can be assigned to natural selection induced by temperature driven gradients. Our results provide evidence that introgression of Q. robur genes into the Q. petraea genome facilitated adaptation of the latter species to cooler and/or wetter climates. Phenotypic monitoring conducted in a common garden on the same populations showed clinal differentiation for growth and leaf phenological traits, while limited divergence was observed for xylem anatomy, and physiology and hydraulic related traits.

On a much shorter time span, we investigated whether the “natural” warming following the Little Ice Age in Europe has triggered evolutionary changes. To do so, we took advantage of the existence of oak stands that were born under the Little Ice Age (more than 350 years ago). Our first results show that progenies of old oak populations flush later than progenies of younger oak populations suggesting recent evolution, and age structured cohorts show differences in genes related to biotic interactions.

We performed a similar analysis by comparing populations artificially transplanted two centuries ago in Europe with their source populations. Here the spatial transfer mimics the temporal climatic changes. These investigations were conducted in Quercus rubra, a species native of North America, that has been widely introduced in Europe. Our results show significant divergence between both gene pools for growth, phenology and reproduction, again suggesting adaptive evolution over few generations.

Finally we used analytical models derived from quantitative evolutionary assumptions to make predictions about evolutionary change in Q. petraea and Q. robur at a very short time span, over two successive generations. Growth, leaf morphology physiology, and defence related traits exhibited significant predicted changes whereas phenology, water metabolism, structure and resilience-related traits did not. However, the direction of the selection response and the potential for adaptive evolution differed between the two species, Quercus petraea being prone to expansion while Q. robur is entering decline.

Major results and achievements

Retrieving ancient organelle and nuclear genomes of woody species

In the Dow of the TREEPEACE project, we mentioned that « the retrieval of genomic footprints from ancient wood remains would be challenging and risky », given the background knowledge and experience in paleogenetics of woody species at the time TREEPEACE was launched. There had been a few earlier reports about single marker genotypic footprints retrieved from wood samples, and exclusively from choroplast genomes.

We established an extraction protocol to recover ancient DNA from subfossil and archaeological wood by optimizing techniques previously used for other plant tissues and animal remains. For aDNA authentication we analyzed DNA degradation patterns. DNA was highly fragmented across all samples and characterized by abundant depurination and deamination. Degradation kinetics of organelle DNA was similar to nuclear DNA. This was a surprising finding contrasting with differing degradation kinetics of mitochondrial and nuclear DNA previously described for animals. To our knowledge this is the first time that such an approach has been successfully set up and applied on wood. This approach represents a milestone for ancient DNA research and molecular archaeology, opening for further investigation of similar material in other biological systems.

Evidence of evolutionary change and adaptive shifts at contemporary time scales (9, 12, 20, 23)

At present, sessile oak and pedunculate oak, witness increased growth, reproduction and earlier bud burst and later leaf senescence. We showed that these changes –of highly integrated traits- correspond also to adaptive shifts contributing to higher fitness. Similar shifts were also detected in Quercus rubra, a species that was introduced in Europe about two centuries ago. However we did not detect changes in other traits, despite our deep phenotypic dissection of complex traits, especially traits related to water metabolism. Lack of detection does not mean lack of occurrence change, but may also be due to reduced statistical power to detect it. We also found footprints of genetic changes triggered by biotic interactions, in addition to abiotic induced selection, suggesting that climate change may indirectly lead to variation of pathogens, insects or microbiome in general. Taken together these results suggest that adaptive shifts are currently going on even in long generation species.

Evidence of adaptive introgression in oaks (21, 24)

Oaks are known to exhibit extensive interspecific gene flow. Earlier investigations have shown how hybridization facilitated dispersion during postglacial colonization. For the first time we showed in TREEPEACE, how hybridization followed by recurrent backcrossing has resulted in adaptive introgression in Quercus petraea, enhancing colonization and adaptation of this species at higher elevation and at more northern latitudes. These results have important implications in the context of ongoing climate change. Temperate white oaks in Europe (Q. robur and Q. petraea) have also mediterranean congeneric species (Q. pubescens, Q. faginea, Q. frainetto) better adapted to drier climates, and that are also undergoing polewards contemporary migration. It is highly likely that these movements will generate more frequent contacts and hybridization with temperate oaks, ultimately facilitating adaptation of the latter to warmer and drier climates.

Persistence of oak populations during the holocene (14)

The comparison of ancient chloroplast DNA footprints with modern fingerprints in different locations across Europe, revealed that oak maternal lineages were maintained since the early establishment of populations during postglacial colonization. Ongoing investigations regarding admixture in ancient and modern DNA will allow to decipher whether persistence occurred also at the nuclear genome, or whether gene flow between lineages contributed to significant population changes.

Detecting epigenetic signatures in wood archeological remains (22)

In a second study, we addressed the question if epigenetic signatures can be tracked in ancient plant materials using post-mortem damage patterns. We extended a statistical methylation score originally proposed to trace cytosine methylation in mammal sequence data to accommodate the three methylation contexts common in plants. We applied this score to a range of tissues (wood, cobs, and grains) and species (oak, maize, and barley), spanning both desiccated and waterlogged archaeological samples for which ancient DNA sequences had previously published by ourselves and other colleagues. This demonstrated that genuine DNA methylation signatures can be characterized in ancient plant remains, which opens new avenues for investigating the plant evolutionary response to farming, pollution, epidemics, and changing environmental conditions.

Assessing vulnerability to embolism in oak shoots (2, 12)

Assessing vulnerability to embolism in oaks - the major drought tolerance trait - is a current challenge in tree ecology due to the presence of long vessels in oak wood. We therefore developed a prototype, called CAVI1000 (http://sylvain-delzon.com/caviplace/), which is a cavitron allowing with its 1m diameter rotor to measure vulnerability curves to embolism on woody species with long vessels. This cavitron is the only of this size worldwide. We validated this new technology by comparing the vulnerability curves obtained with the CAVI1000 with direct observations of embolized vessels in stems using X-ray based microCT in a synchrotron.

Predicting evolutionary change of forest trees in natura over two successive generations (7, 13, 26)

For the first time in forest trees we implemented a method based on evolutionary quantitative genetics principles to estimate basic genetic parameters in stands undergoing natural and human mediated selection pressures. These parameters allowed to make prediction of genetic shifts of traits contributing to the fitness of trees. The approach requires the reconstruction of genetic relationhsips among trees over one or two generations, the assessments of micro environmental variables, and the use of general mixed models. We also provided a way for empirical validation of the predicted shifts.

Improving demographic models within species complexes (6, 24)

We improved available demographic modeling approaches based on Approximate Bayesian Computation to infer the evolutionary history of oaks, including the probability of past gene flow in European white oaks, and population changes associated to the introduction of the North American population of the red oak Q. rubra in Europe (8). The most recent developments of this work allowed us to explicitly take into account this past demography for detecting regions under selection, in an optimal, albeit unconventional, way.

Opening new research avenues

Paleogenomics of forest trees

  • The experience gained in the ancient DNA analysis of oak wood samples opens the door to paleogenomics of forest trees as a new approach to retrace the evolutionary trajectories of trees during the Holocene (14). We first highlighted the optimal conditions for DNA preservation in wood samples, and then showed how choroplast and nuclear genomic footprints may witness historical events like population persistence or admixture during tree colonisation. During the project we contributed to a review paper on plant paleogenomics that emphasizes new development derived from ancient DNA analysis (19)
  • An individual based forward simulation program was developped to monitor evolutionary changes of allelic frequencies, trait values and fitness under user defined evolutionary scenarios (1, 17). The simulation engine called METAPOP offers an invaluable tool to compare the sensitivity of observed divergence and/or genetic change to various evolutionary drivers and mechanisms (gene flow, selection pressures, drift, assortative mating). The model has been extended at the multitrait scale and includes now also plasticity. We thus anticipate that it will be used by larger community of evolutionary biologists.

Evolutionary quantitative genetics

  • Revitalizing and promoting provenance research to tackle adaptive response of trees to climate change. Provenance tests are large scale common garden experiments that were established decades ago with the aim to identify appropriate seed sources for afforestation. We demonstrated how they could be used today to anticipate their likely adaptive or plastic response to ongoing environmental changes, despite the very heterogeneous data extracted from the multiple plantations installed in different countries under different silvicultural regimes (5, 25). Our case study in Quercus petraea has now inspired other attemps in broadleaves and conifers (27,28).
  • Implementing quantitative genetics to predict short term genetic changes in natura. Quantitative has been traditionally used in tree breeding programs, within experimental plantations controlling environmental variation and using controlled crosses. Within TREEPEACE we showed how genetic relatedness between trees in natural stands could be reconstructed, thus allowing to estimate key genetic parameters as heritability, genetic correlations, predicted genetic responses (10, 13, 26). These findings offer various fundamental and more applied uses in evolutionary genetics and operational forestry, but need further refinements of the whole methodology.


Microevolution of oaks along different time frames

Microevolution during the Holocene. Allochronic approach

We recovered ancient DNA from subfossil and archaeological oak wood remains collected from across Europe and from throughout the Holocene (550 to 9800 year old material). We reconstructed 21 ancient chloroplast haplotypes from three time periods (Neolithic, Bronze and Middle Age) and nine different localities (14). These genomes represent the first ancient chloroplast genomes of a European tree. Ancient chloroplast haplotypes matched the most frequent haplotypes found in nearby extant populations. Although the sequence coverage of nuclear chromosomes was overall low (‹‹1%), we managed to identify 14 samples with oak endogenous DNA content compatible with further enrichment and / or deep sequencing. Based on the first screening results, we selected 3 ancient oak individuals from 3 different localities to generate three low coverage nuclear genomes (2 Medieval, 1 Bronze Age). These genomes are currently analyzed and compared to nearby modern individuals of pedunculate and sessile oak.

Microevolution during the Holocene. Synchronic approach

A systematic scan for genomic and phenotypic differentiation was conducted within 18 oak populations sampled along latitudinal and latitudinal gradients mimicking ongoing climatic changes (21, 12). Whole genome scans among Q. petraea populations revealed outlier SNPs that also exhibited strong interspecific divergence between Q. petraea and Q. robur, suggesting introgression between the two species. Genetic clines for associated SNPs corroborated the contribution of introgression from Q. robur to the adaptive divergence of Q. petraea populations (21).
Phenotypic monitoring conducted in a common garden on the same populations showed clinal differentiation for growth and leaf phenological traits, while no divergence was observed for xylem anatomy, physiology and hydraulic related traits. While differences between mediterranean and temperate oak species were observed for xylem vulnerability, no population variation between mediterranean and temperate Q. petraea populations was detected for resistance to embolism (2, 12, 20).

Microevolution during the Anthropocene. Contemporary evolution

Evolutionary change due to population transfer mimicking climate change We compared 38 introduced and 73 range-wide native populations of Quercus rubra, a native species of Eastern North America that was introduced in Europe at the end of the 17th century. After tracking the original seed source by Bayesian clustering analysis with SNP markers, we showed that introduced populations diverged significantly from their source for growth, leaf phenology and fruiting (8, 23).
Evolutionary change due to warming since the little ice age We sampled in three Q.petraea forests, four age structured cohorts (360 years, 180 years, 50 years and 10 years) with 50 trees per cohort. Whole genome sequencing of the cohorts indicated that genes involved in defence pathways exhibited allelic frequencies changes. At the phenotypic level, there were noticeable spring phenology differences among the cohorts. Data of this experiment are currently being analysed (no publication available yet).

Microevolution during the Anthropocene. Instantaneous evolution

We estimated in situ genetic parameters allowing to predict genetic changes over two generations in a mixed Q. petraea - Q.robur stand. In both species, growth, leaf morphology, physiology, and defence-related traits displayed significant predicted genetic shifts, whereas phenology, water metabolism, structure and resilience-related traits did not (7, 13, 26).

Microevolution in action. Simulations

The modeling platform METAPOP was upgraded by redesigning the core of the simulation engine, thus facilitating the introduction of new functionalities and additional evolutionary processes that are needed to serve the objectives of TREEPEACE (17, 1). The engine considers now multivariate phenotypes and includes explicitly a module about assortative mating and plasticity.

Back to top of page