Resurrection of Dormant Daphnia magna: Protocol and Applications

Long-term studies enable the identification of eco-evolutionary processes that occur over extended time periods. In addition, they provide key empirical data that may be used in predictive modelling to forecast evolutionary responses of natural ecosystems to future environmental changes. However, excluding a few exceptional cases, long-term studies are scarce because of logistic difficulties associated with accessing temporal samples. Temporal dynamics are frequently studied in the laboratory or in controlled mesocosm experiments with exceptional studies that reconstruct the evolution of natural populations in the wild. Here, a standard operating procedure (SOP) is provided to revive or resurrect dormant Daphnia magna, a widespread zooplankton keystone species in aquatic ecosystems, to dramatically advance the state-of-the-art longitudinal data collection in natural systems. The field of Resurrection Ecology was defined in 1999 by Kerfoot and co-workers, even though the first attempts at hatching diapausing zooplankton eggs date back to the late 1980s. Since Kerfoot's seminal paper, the methodology of resurrecting zooplankton species has been increasingly frequently applied, though propagated among laboratories only via direct knowledge transfer. Here, an SOP is described that provides a step-by-step protocol on the practice of resurrecting dormant Daphnia magna eggs. Two key studies are provided in which the fitness response of resurrected Daphnia magna populations to warming is measured, capitalizing on the ability to study historical and modern populations in the same settings. Finally, the application of next generation sequencing technologies to revived or still dormant stages is discussed. These technologies provide unprecedented power in dissecting the processes and mechanisms of evolution if applied to populations that have experienced changes in selection pressure over time.


Introduction
Long-term studies are critical to understanding ecological and evolutionary processes in nature and in assessing how species respond to and persist during environmental change 1 . This is because eco-evolutionary processes happen across generations and changes in the environment occur over long time spans. Furthermore, long-term studies provide key empirical data that improve the accuracy of predictive modeling to forecast evolutionary responses of natural ecosystems to environmental changes 2 . The accuracy of these models is critical to implement management and conservation strategies to preserve biodiversity and ecosystem services.
Excluding a few exceptional cases (e.g., Galapagos Darwin finches 3 and algae 4 ), long-term studies are largely limited to species with short generation time that can be propagated in the laboratory 5,6,7,8 . Hence, the processes underpinning evolutionary dynamics remain elusive. Because of logistic difficulties associated with accessing temporal samples, empirical data are studied more frequently in a spatial than in a temporal context, and temporal eco-evolutionary processes are inferred or modeled from spatial data. This approach is known as 'space-fortime' substitution 9 , whereby space is adopted as a surrogate to study temporal evolutionary dynamics. The main limitation of the 'space-for-time' substitution is that rates of adaptation at different spatial scales differ from temporal variation in the same population; hence, inferences based on replacement of time with space are biased 10 .
A powerful alternative that allows studying evolutionary dynamics in natural ecosystems over time is the analysis of ecological and genetic changes in species producing dormant stages 11 . These dormant stages accumulate to form stratified biological archives that can be accurately dated and paleolimnologically characterized 12,13 . Importantly, these dormant stages can be resuscitated and used in laboratory experiments, where their evolutionary response to environmental change can be directly measured. Historical populations can be competed against their modern evolved descendants to study fitness changes and the function of genes evolving in step with environmental change 14,15,16 . Dormant stages include seeds, cysts, spores, and egg banks. Although the first studies on resuscitated dormant eggs dates back to the late 1980s 17 , and a handful of studies have applied this technique in the early 1990s 18,19 , the field of Resurrection Ecology has been formally established by the seminal paper of Kerfoot and co-workers in 1999 Daphnia are freshwater zooplankters present in the majority of lotic habitats 23 . Daphnia species are either obligate asexual or cyclical parthenogens. D. magna is a cyclical parthenogen that reproduces clonally under favorable environmental conditions 24 . When environmental conditions deteriorate, male production occurs and sexual recombination leads to the formation of fertilized eggs that enter a state of dormancy protected from the environment by a chitin case called ephippium. A proportion of these dormant eggs hatch when favorable environmental conditions return. However, a large proportion of the dormant egg bank never has a chance to hatch and thus build up biological archives over time. Dormant stages remain buried in the sediment of lakes and ponds and can be resurrected for the study of evolutionary dynamics over extended time periods. Because dormant eggs of D. magna are the result of sexual recombination, they are a good representation of the natural genetic diversity of the species 25 . Moreover, they can be maintained via clonal reproduction in the laboratory. These characteristics provide the unique advantage of isogenic model organisms, while retaining the natural genetic diversity.
Two key studies are presented to demonstrate the advantages of directly comparing historical and modern descendants of the same population of D. magna experiencing environmental selection pressure over time. D. magna specimens were resurrected from Lake Ring (Denmark), a shallow (5 m depth; surface 22 ha) mixed pond that has experienced an increase in average temperature and heat waves occurrence over time. D. magna (sub)populations were resurrected along this temporal gradient spanning 60 years  and studied to investigate evolutionary response to temperature warming. In the first study in a common garden experiment, changes in fitness-linked life history traits were measured in response to an increase in temperature of +6 °C, in line with the predictions of the Intergovernmental Panel for Climate Change for the upcoming 100 years 26 . In the second study, a mesocosm experiment was used to measure the competitive abilities of the three (sub)populations under warming. These experiments combined show that in presence of warming as the only stress, all life history traits and populations show a high level of plasticity and have equal competitive abilities. These findings suggest that warming as a single stress does not impose significant fitness costs, at least in the population studied here.

Sampling of Sediment Cores
1. Sample sediment from lakes or ponds using a piston corer. This protocol used Big Ben 27 , a core tube of approximately 1.5 m in length with an internal tube diameter of 14 cm. Big Ben consists of a piston on a rope and a corer head, to which rods are attached to drive the tube into the sediment. A core catcher aids the support of the core tube when full of sediment. To extrude the sediment, a framework keeps the core tube upright and stationary, and a modified bottle jack is used to push the piston upwards during the extrusion process (Supplementary Video 1).
1. For shallow ponds of less than 1 m in depth, use a plexiglass gravity corer of no more than 6 cm of diameter manually pushed into the sediment. 2. For deep lakes (>6 m of depth), use Livingston piston corers 28 or single-drive Griffith sediment corers with the aids of an anchored pontoon boat. The Livingstone-type drive rod piston corer can be used in water up to about 30 m deep to collect successive one-meter drives of soft to consolidated lake sediment. The single-drive Griffith corer consists of a simple but robust core head that connects standard polycarbonate tubes to Livingstone drive rods. The corers are pushed into the sediment with the rods, and a piston provides the suction needed for the recovery of sediment (Figure 2). 3. For retrieving continuous, undisturbed cores, use vibracoring. These corers work on a variety of water depths and can retrieve core samples of different lengths, depending on sediment lithology. Low amplitude vibration that is transferred from the vibracore head down through the attached barrel or core tube liquefies sediments, enabling the core barrel attached to the vibracore unit to penetrate into the liquefied sediments. Some vibracorers are small, lightweight, and portable, others are large heavy units that can only be deployed from large vessels. The choice of corers depends on the lithology of sediment.
2. Slice the core horizontally in incremental layers of 1 cm or less using a flat metal surface (Supplementary Video 1). Sediment corers like the one used here are designed to reduce hydrostatic pressures at extrusion, reducing disturbance of the sediment layers. When other corers are used, the outer rind of each sediment layer may be removed with a cookie-cutter sort of blade to limit contamination among layers. 3. Collect each sediment layer in a separate sampling bag (Supplementary Video 1), and store in dark and cold (4 °C) conditions. 4. Collect a minimum of 5 g of sediment from all layers for radiometric dating. As radiometric dating is an established protocol for a detailed description of the dating assay, refer to existing publications Figure 2: Cartoon of the piston coring procedure. Piston corer, a hollow tube with an internal sliding seal (the piston) that produces a weak vacuum. When the piston touches the sediment-water interface, the weight pushes the core barrel into the sediment and the vacuum causes the sediment being cored to enter and move up the tube without disturbing the sediment layers. Please click here to view a larger version of this figure.

Sieving of Sediment Layers
1. Using a precision scale, weigh each sediment layer for future reference. Use the surface area and weight to calculate the species density in the lake. 2. Sieve each sediment layer using two geological sieves piled on top of each other. The first sieve has a mesh size of 1 mm and separates clay, large invertebrates and particulate matter, e.g. seeds, plants, and insects, from the remainder of the sediment. The second sieve has a mesh of 125 µm and separates D. manga ephippia and small particulate from the remainder of the sediment (Supplementary Video 2). 3. Transfer small aliquots of the sediment fraction collected onto the 125 µm mesh sieve to a white background tray. Depending on the type of sediment, smaller or larger aliquots of sediment may be transferred at each time.
1. Add small volumes (up to 200 mL) of medium to the sampling white tray to re-suspend the transferred sediment fraction and facilitate eye spotting of ephippia ( Figure 3). The medium used to resuspend the sediment can be dechlorinated tap water, borehole water, COMBO 29 , or ADaM medium (Aachener Daphnien) 30 1. With a disposable Pasteur pipette or using microdissection forceps, transfer individual ephippia to Petri dishes filled with 10 mL of medium.
Where l x is the proportion of survival at age x, b x is the number of neonates produced per surviving individual at age x, and r is the intrinsic rate of natural increase. 9. Perform statistical analysis using commercially available software. Here, use R 31 to plot reaction norms (phenotypic expression of a single genotype across environments) for each life history trait, using the ggplot2 package. Calculate mortality rates per population via survival analysis (R package rms; https://cran.r-project.org/web/packages/rms/rms.pdf). Finally, perform an analysis of variance (ANOVA, Table 2) in R 31 to assess whether the effect of temperature on traits can be explained by evolution (differences among populations), plasticity (response to treatment), or evolution of plasticity (population x treatment).
2. Competition experiment 1. Transfer ten hatchlings per population from stock cultures to common garden conditions: 20 °C, long photoperiod (16:8 light:dark regime), feed daily with 0.8 mg C/L of Chlorella vulgaris, and renew medium every second day. Maintain common garden conditions for at least two generations (ca. 45 days) to reduce interference from maternal effects. 2. Randomly assign five juveniles of 24 to 48 h from the second brood of the second generations of each hatchling to experimental mesocosms (20 L plastic aquaria filled with medium), in triplicates, at a density of 10 animals/L. 3. Expose the mesocosms to 24 °C, 16:8 L:D regime, and feed daily with 0.8 mg C/L of Chlorellavulgaris in a controlled temperature chamber or incubator for a minimum of four weeks (>3 clonal generations). 4. Cull each mesocosm weekly, refreshing 10% of the medium to simulate a population dynamic that Daphnia may encounter in the natural environment. Impose culling by removing a known volume of medium and animals from each aquarium (1.2 L in this case), and by replacing the culled volume with fresh medium. Start the culling regime after the experimental animals reach maturity (day 10). 5. At the end of the fourth week, sample 32 animals from each mesocosm to assess shifts in genotypic composition as compared to the initial inoculum.  Table 1) were used following established protocols 32,33 . 5. Genotype amplified fragments on a fragment analyzer. 6. Conduct fragment analysis with a commercial or freely available software using a suitable size standard.   (1960-1970; 1970-1985; >1999) are shown under warming (24 °C) as compared to modern temperature regimes (18 °C). Please click here to view a larger version of this figure.

Mesocosm experiment
After four weeks of selection, represented by warming at 24 °C, the frequency of the three (sub)populations did not change significantly (χ 2 = 0.55, P = 0.76) as compared to the initial inoculum (Figure 8). Among the 30 genotypes inoculated in the mesocosm experiment, the majority was identified after four weeks of selection (Figure 9). Specifically, 70% of the inoculated genotypes were recovered, compatible with Poissonian expectation of recovering at least one representative of each genotype in a sample of 32 individuals.  , are shown after four weeks of exposure to warming (24 °C) as compared to an initial equal frequency of genotypes (dotted line). Names on the x-axis are the inoculated genotypes ID, grouped per (sub)population (blue, 1960-1970; green, 1970-1985 Table 2: Analysis of variance(ANOVA). Analysis of variance testing whether changes in life history traits and population growth rate of the resurrected (sub)populations exposed to warming are explained by evolutionary adaptation (populations), plasticity (temperature treatment), and their interaction term (evolution of plasticity). Significant p-values (p<0.05) are shown in bold.

Supplementary Video 3: Decapsulation.
Under a stereomicroscope, D. magna ephippia are opened with microdissection forceps by applying pressure on the spine of the chitin case. The inner egg membrane is delicately removed and resting eggs gently transferred with a Pasteur pipette to a Petri dish containing 10 mL of medium. Please click here to download this file.

Supplementary Video 4: Hatching.
After exposure to a long photoperiod and 20 °C, embryo development resumes between 48 h and few weeks. When development is complete, the embryos break free from the egg shell and freely swim in the medium. Please click here to download this file.

Discussion
Due to the high thermal conductivity of water, freshwater ecosystems are at higher risk of biodiversity loss than terrestrial ecosystems in the face of global warming 34 . It is, therefore, critical to understand the response of keystone species in these ecosystems and identify coping mechanisms to survive thermal stress. The understanding of these mechanisms at the species and community levels can help predict how species are impacted by global warming and how the effect on individual species cascades to other trophic levels. Ultimately, understanding mechanisms of responses to global warming enables the identification of remediation strategies to mitigate extinctions.
The case studies presented here show that the response of D. magna to temperature increase is pervasively mediated by plasticity in life history traits and that response to temperature increase alone does not impose clear fitness costs, at least in the population studied here. High plasticity in life history traits is supported by non-significant differences in competing abilities of the (sub)populations in the presence of warming. However, longer-term competition experiments on multiple populations may be necessary to generalize these findings.
Resurrection of dormant stages provides an unprecedented resource to study mechanisms of adaptation and trajectories of a species' evolution through time 10 . Zooplankton species benefit from a rapid generation time (about 2 weeks), and the viability of dormant stages, which allows an ancestor to compete headtohead against its own descendants, or to 'replay' evolution that starts from various past states. Resurrection ecology essentially enables the investigation of whether a particular evolutionary outcome is contingent on some prior event. The identification of the genetic elements of evolution is currently possible in laboratory experiments using microorganisms for which 'ancestral lines' are frozen and resuscitated for comparative analysis with their evolved descendant 6 . However, one of the main limitations of working with laboratory organisms is that the 'ancestral state' is an already shifted baseline. The study of dormant stages allows the sampling of specimens from time predating any stress event (e.g., pristine environmental conditions) and to measure evolutionary trajectories from undisturbed environmental conditions to various past states until modern times. In recent years, the study of DNA polymorphism in resurrected or still dormant zooplankton stages has provided important insights into past demographic and adaptive processes that have contributed to the genetic makeup of present-day populations 14,16,25,33,35,36 . With the higher accessibility of high throughput sequencing technologies, the genome and transcriptome of resurrected or still dormant stages can be sequenced and the type and number of genetic changes accumulated in evolving populations over time measured.
The resurrection SOP presented here has important applications in the field of multi-omics on two levels. Multi-omics technologies can be applied to resurrected specimens, allowing an exhaustive analysis of the molecular elements involved in adaptive responses to environmental selection pressure. In addition, omics technologies can be applied to decapsulated but still dormant stages. So far, the application of high throughput sequencing technologies to resting stages has been limited by the requirement of a large amount of input material. These limitations are being lifted 37 . With the lowering requirements for input material and progress in nanofluidics, whole genome sequencing (WGS) is now possible from as little as 1 ng or a few pg of starting material 38 . The use of whole genome amplification (WGA) and whole transcriptome amplification (WTA) techniques, enabling the enrichment of DNA and RNA from very small amounts of tissue, has revolutionized both metagenomics 39,40 and medical research 41 . These technologies applied to decapsulated dormant eggs enable the exceeding of limitations associated with viability of dormant stages and the investigating of extended time periods (e.g., centuries).
The resurrection of invertebrate communities producing resting stages enables the alignment of community histories with known changes in the natural landscapes, or with environmental changes inferred from analyses of the sediments orsoils 2 . The analysis of community changes in response to environmental change provides us with the ability to quantify eco-evolutionary feedbacks 42 that have substantial consequences on population persistence 43 , trophic interactions 44 , community assembly 45 , and changes in ecosystem functions and services 46 . Finally, accurate predictions about biological responses to environmental change are paramount to guide the protection of biodiversity 47 . Current predictive models are inaccurate in this respect because they do not take into account important biological mechanisms such as demography, dispersal, evolution, and species interactions. Understanding how these processes change over time and using this information as a prior in forecast modelling will improve our ability to predict species and community persistence in the face of environmental change Some of the SOP steps presented here are readily transferable to other Daphnia species. These are: sampling, establishment of clonal lines, and experimental design. However, other steps of the SOP may require further optimization tailored to the species under study. Decapsulation is often applied to D. magna specimens to improve hatching success. However, this approach may not be suitable for smaller specimens. Hatching stimuli can also vary among species 48 and among conspecific specimen 49 . Hence, an ad hoc optimization of the hatching steps of the SOP may be required prior to applications to other crustaceans. Whilst the hatching success of the D. magna population resurrected from Lake Ring (30.5% across the sedimentary archive) is in line with previous results 49 , hatching success varies with the preservation state of the sediment, the species 50,51 , and the geographic origin of the sediment 48 . Future studies on the mechanisms that regulate entry and progression through the phases of diapause is required to identify optimal hatching stimuli tailored to different species.
Finally, background knowledge of the study system, in particular the presence of the species of interest over time, is advisable. This can be achieved via historical records. If historical records are not available, sampling and screening of the surface layers of the lake sediment prior to core sampling is advisable, although it may provide information only on the most recent history.

Disclosures
The authors have nothing to disclose