The genomics revolution is rapidly materializing as reduction in the cost per base pair is realized through improvements to high-throughput sequencing platforms, bioinformatics, and general analytical methodology (Glenn, 2011). In preparation for the coming wave of genome sequencing needs, the Smithsonian Institution has established the Global Genome Initiative (GGI;  http://www.ggi.si.edu/), which seeks to collect, voucher, and preserve genome-quality tissues from at least one species in each family and 50% of all the genera of life on Earth by 2020. For the green plant branch of the tree of life, GGI has started a program called GGI–Gardens that has as its mission the collection of genome-quality leaf samples, as well as their associated vouchers, from plants housed in gardens, greenhouses, and arboreta (hereafter called “Gardens”). In line with this goal, GGI–Gardens is forming partnerships with Gardens from around the world to form an international consortium of living plant collections that will work together to preserve samples to facilitate scientific research around the world. Here we outline a workflow for the collection and storage of leaf tissue and vouchers to be used by any organization interested in preserving genome-quality tissue for use by current and future generations of researchers. The intended audience for this article includes both botanists who have never collected tissue for genomics work, but wish to do so, and garden staff who plan to voucher their living collections.

Natural history collections serve a critical function for a variety of studies including, but not limited to, biodiversity, conservation, and evolution. For plants, herbaria and their associated collections (i.e., libraries, illustrations, photographs) represent the cornerstone of natural history collections. Currently, there are more than 350,000,000 plant and fungi specimens stored within nearly 3000 herbaria around the world (Thiers, 2016), and this number continues to grow. Herbarium collections enable essential research of plant and fungal diversity, but these collections are often ill-equipped for genomic applications (e.g., whole genome sequencing). Until recently, natural history collections have had few available resources to preserve genome-quality tissues. Zimkus and Ford (2014) surveyed genetic resource curation from 45 natural history collections across 39 institutions and found varying practices in place from sampling acquisition to final storage. Only recently have best practices been developed for repositories (see ISBER, 2012), and in many cases these have been optimized for animal or human biological or environmental samples. For instance, a recent methodology (Wong et al., 2012) identified a four-star standard system for vertebrate genomic tissue sampling; however, although basic collection practices have been identified (see Spooner and Ruess, 2014), no such method or standard exists yet for plants.

There are already in operation several well-funded and globally important programs for the collection and preservation of plant genetic resources, such as the U.S. Department of Agriculture (USDA) Germplasm Resources Information Network (GRIN;  http://www.ars-grin.gov/); the Millennium Seed Bank hosted by Kew Gardens ( http://www.kew.org/science-conservation/collections/millennium-seed-bank), which has samples of more than 36,000 species (van Slageren, 2003); and the Svalbard Global Seed Vault ( http://www.regjeringen.no/en/topics/food-fisheries-and-agriculture/landbruk/svalbard-global-seed-vault), which houses viable, duplicate samples of important crops and their wild relatives (Fowler, 2008). These organizations reflect a global commitment to the preservation of genetic resources with similar, but in some cases fundamentally different purposes (i.e., preserve and share genetic resources of crops and their wild relatives through GRIN). However, there has been no concerted effort to sample plant biodiversity across the globe and store tissue for future use in basic research that harnesses genome sequences. To accommodate this need, the Global Genome Biodiversity Network (GGBN; Seberg et al., 2016) was founded as a consortium of biorepositories to facilitate the storage of genome-quality tissues for natural history collections. GGI, including GGI–Gardens, is part of the GGBN, and the workflow described here is meant for tissues from living plants that are destined for a GGBN facility; however, it can be modified to suit the needs of other preservation methods. The workflow serves as a template for ongoing comparative studies of optimal plant DNA/ RNA sampling and preservation techniques for long-term storage and genomic research.

METHODS AND RESULTS

The workflow described here was tested over the course of six months. Collection from sites in 2015 resulted in ca. 900 voucher specimens by teams of two to four individuals. Each team was led by an experienced field biologist to make sure the plants were properly collected and that information taken down in the field was accurate. The remaining members were interns. Participation by a garden staff member is essential for successful and efficient collecting expeditions. During this time, collecting sites included the Smithsonian Institution Gardens and their greenhouses, the Department of Botany (National Museum of Natural History [NMNH]) greenhouse, the U.S. National Arboretum and South Farm (both USDA), and the United States Botanic Garden (Architect of the Capitol) and production facility. For each voucher specimen, tissue suitable for the isolation of DNA/RNA was collected and preserved in two ways: silica gel desiccant and liquid nitrogen. At least two high-quality photographic images were taken (habit and close up) for each specimen, and all collection information and images were uploaded into the NMNH electronic record management database ( http://collections.nmnh.si.edu/search/botany/). Collection procedures were ideally situated for teams of three individuals and have been outlined in Appendix 1, which includes a step-by-step protocol for collection needs prior to, during, and immediately following sampling. We describe an “any or all” approach to sampling procedures from living collections in Gardens for genomic research according to five steps: (a) preparation of living collections for genome tissue sampling, (b) voucher specimen handling and collection, (c) genetic sample handling and collection, (d) biorepository handling and storage, and (e) common sampling problems and some solutions. These GGI-Gardens workflow steps are summarized diagrammatically in Fig. 1, and a sample collection sheet is shown in Fig. 2.

Preparing living collections for genome tissue sampling—Internationally, Gardens hold an incredible amount of living plant diversity, including more than 450,000 taxa in more than 3300 Gardens (Botanic Gardens Conservation International, 2016). The collections are often identified to species level and may include rare taxa that are otherwise difficult to sample for genomic applications. A list of species in a living collection is extremely helpful to assist with indexing collected vouchers, identifying sampled tissues, and determining the original source of the material. If possible, sampling from a living collection can be optimized by identifying gaps in desired specimens and demarcating them in the garden before collecting. When possible, an experienced field botanist or member of the garden staff should accompany any sampling team to ensure collections are properly identified and appropriate material is collected in a timely fashion. Collections can be tracked through an organized note-taking system (see Fig. 2). GGI–Gardens has also replicated this collection sheet as a custom project in the iNaturalist web application environment ( http://www.inaturalist.org).

Voucher specimen handling and collection —The term “voucher specimen” in this manuscript refers to a plant (or part of a plant) that has been collected, pressed, dried, labeled, mounted on archival paper, and deposited into a recognized herbarium (e.g., a member of Index Herbariorum [ http://sweetgum.nybg.org/science/ih/]). The most important resources for sampling voucher specimens from a living collection are experience, time, and the proper tools. As with most botanical field collection methods, collectors should bring tools including, but not limited to, those listed and defined in Table 1. High-quality photographs are essential for GGI–Gardens collections to confirm determinations. Three photographs are typically taken, including (1) the specimen tag within the living collection that includes reference data for tracking provenance, (2) a photo of the specimen habit in the context of the living collections, and (3) a close-up photo showing detailed reproductive and vegetative material.

Genetic sample handling and collection —A genetic sample is collected at the same time a voucher specimen is prepared, but to preserve the intrinsic properties of DNA/RNA essential for genome sequencing, special handling and preparation methods are required. For a variety of reasons (summarized in Nagy, 2010), genomic DNA/RNA may degrade rapidly once removed from a living plant. Because of this, we recommend the preparation of two distinct genetic samples (a) silica-preserved and (b) liquid nitrogen–preserved. Chase and Hills (1991) and Corthals and Desalle (2005) discuss the biochemical and macromolecular merits of each preservation method, respectively, and outline infrastructural limitations for collectors. Many of the challenges for collections stored in liquid nitrogen are obvious, but the proximity of many Gardens to facilities that provide access to liquid nitrogen can mitigate these concerns. Together, GGI and GGBN can assist Gardens that wish to collect tissues from their living collections and preserve them for genomic research. A diagram showing recommended field collection equipment and resources is provided in Fig. 1, and a list is given in Table 2.

Biorepository handling and storage —Prepared genetic samples must be transferred to a biorepository promptly to prevent sample degradation and to ensure access for research applications. Biorepositories can be identified through the GGBN Data Portal (GGBN, 2011). Preparation and storage of genetic samples depend upon the ultimate destination and use of the samples. Each GGBN biorepository may have a different preference for storage based upon their capacity and infrastructure. For GGI–Gardens, collections are deposited into GGBN-affiliated biorepositories. Relevant legal and political considerations regarding data and material sharing are described in Seberg et al. (2016).

Common sampling problems and solutions —Insufficient material occurs when plants are small and/or reproductive material is limited. Some simple solutions include high-quality photographs and the use of older flowers or fruits as well as repeated visits to obtain the necessary material. Incorrect names are common because nomenclature (the naming of plants) is a living entity and therefore changes constantly; this may also occur as a result of staff error or quality control issues. In addition, Gardens have at their disposal a variety of nomenclatural verification materials with varying degrees of authority and accuracy. Errors in nomenclature can be easily corrected later when the data are being entered into the institutional names database. For taxonomic determinations, GGI–Gardens uses publications by experts and, as a back up, The Plant List (2013), which currently includes 17,020 genera in 642 families. The challenge of taxonomy is reflected in The Plant List by the “unassessed” status given to ca. 22.7% of species names for angiosperms alone; however, it is a widely accessible and up-to-date list of plant taxonomic names. GRIN ( http://www.ars-grin.gov/) remains an important alternative resource for Gardens and other personnel for problematic unassessed taxon names. Misidentified plants are more common than we would like and more difficult to deal with. The use of an incorrect name for a collection means that later it may be misused in a research project and may lead to confusion during research applications. This problem can be minimized in two ways. First, determinations in Garden living collections are reliable, and mistakes usually result from collectors who misinterpret labels; it is therefore easier to confirm determinations with an expert. Second, we are currently adding barcode sequencing (Kress et al., 2005) to our workflow to validate taxonomic identifications and to add to the global barcode effort. Sample cross-contamination is a concern for any approach to sampling tissue for genomic research. Possible contaminants in a garden sampling workflow could be introduced by poor handling of material, insufficient protective equipment, and lack of sterilization methods. Our best practices outline sterilization methods using flame and the handling of equipment with disposable gloves to reduce the impact of contamination between samples. These handling techniques are common for many greenhouse procedures, but less common for botanical field collection; some practical methods have been outlined for orchids (Fitch, 2004).

Fig. 1.

Diagram of the workflow associated with GGI–Gardens genomic tissue sampling. (A) Genomic tissue sampling begins in a garden or greenhouse. Targeted specimens are collected with pruning shears, and a voucher specimen is prepared using a plant press. Detailed notes and photographs are taken in the field describing the living specimens and their locality. (B) All sampled specimens also contribute tissue for preservation on silica gel and in liquid nitrogen. Silica-preserved tissues are placed into labeled envelopes and kept in a resealable bag filled with silica gel. Liquid nitrogen is brought into the garden/greenhouse in a transportation Dewar flask. Tissues are placed into GGBN-approved cryogenic storage tubes, to which a barcode label is affixed. After wrapping cryo-tubes in aluminum foil, they are deposited into the Dewar and then sent to a biorepository for storage. (C) Preserved and stored genome-quality tissue samples can be accessed for research including whole genome sequencing. Data for taxa sampled by GGI–Gardens can be accessed by searching the GGBN web portal (GGBN, 2011). (D) All voucher specimens are mounted and deposited into a recognized herbarium. Future collections are informed by GGI Gap analyses, which allow GGI-Gardens to target priority families and genera that contribute to the GGI mission.

Voucher preparation methods do not exist for living collections in Gardens, and guidance for genome-quality tissue collection has not been standardized for plant samples. Our protocol is based upon an adaptation of existing protocols that have been independently developed for genetic research. This protocol differs from alternatives because it ensures the long-term viability of genome-quality tissues by immediately flash-freezing in liquid nitrogen. It is important to note that liquid nitrogen is difficult to transport into the field, particularly for remote field sites; consequently, the availability of botanical diversity within Garden living collections facilitates a large-scale, organized effort to collect and preserve genome-quality tissues as we have outlined here. This represents a distinct advantage over traditional collection strategies that can present barriers to flash-freezing in liquid nitrogen in other contexts. Furthermore, GGI–Gardens collections contribute to a well-curated network of biorepositories that provide guidance documentation and develop best practices for physical genomic collections (Seberg et al., 2016). In addition to contributing fresh, genome-quality tissue to global biorepositories, this protocol also synergizes with Gardens that wish to voucher their collections and improve their accessibility and visibility for research but do not know how to begin. We considered alternative collecting materials for genomic applications (e.g., RNAlater; QIAGEN, Carlsbad, California, USA); however, the cost was preventative at our scale of collection when compared to readily accessible liquid nitrogen facilities (i.e., the GGBN–partnered NMNH Biorepository, Smithsonian Institution). This and many other GGBN-partnered biorepositories may provide assistance with access to liquid nitrogen for partnerships that contribute to GGBN (e.g., GGI-Gardens). Our data collection includes all available provenance data for living collections. However, one important disadvantage of our genome-quality tissue collection protocol is that material is not directly sourced from wild populations; however, we think that this protocol allows for unparalleled accessibility of material from across the plant tree of life.

CONCLUSIONS

Future problems and solutions are certain to arise, particularly with regard to rare or specialized living collections. In some cases, difficult-to-collect plants (e.g., palms) may require specialized collection procedures. Significant documentation exists for making vouchers from such plants, and a quick literature search will be of great help to Gardens. In many cases, the GGI–Gardens website ( http://ggi.si.edu/ggi-gardens) will serve as an information repository for such best practices and to curate resources and documentation. As biobanks, DNA banks, and biorepositories emerge globally to address the research and infrastructural needs of comparative genomics research, a standard for sampling must be implemented to ensure consistency across samples used for high-impact genome-sequencing efforts. Several studies have attempted to establish a standard for plant DNA/RNA collection and preservation (e.g., Chase and Hills, 1991; Särkinen et al., 2012; Gaudeul and Rouhan, 2013; Neubig et al., 2014); however, none establish a workflow for tissue sampling that will contribute to a widespread genome biorepository network as we have here. We recommend this sampling workflow for GGI–Gardens partners, which will help to refine the optimal sampling protocol and standards that will define plant genome research.

Fig. 2.

GGI–Gardens collection sheet for voucher and genetic samples.

Table 1.

List of materials and their function for voucher specimen handling and collection associated with the GGI–Gardens project.

Programs such as GGI–Gardens help Gardens show that their collections are critical to the future of scientific research. Furthermore, the availability of properly preserved leaf tissue for genomic work will greatly benefit downstream science. The workflow outlined here will facilitate a global sampling effort through voucher preparation at Gardens. Whether Gardens simply want to generate a set of voucher specimens from their collections or contribute genetic specimens for genetic or genomic research, we have provided a set of necessary steps to enable discovery from living collections. This list of resources will grow as the GGI–Gardens program continues to add international Garden partners. Membership in the GGI–Gardens consortium is available to interested Gardens that meet the criteria of the GGI–Gardens Memorandum of Cooperation ( http://ggi.si.edu/ggi-gardens). Any questions can be addressed to the first or last author.

Table 2.

List of materials and their function associated with genetic sampling handling and collection with the GGI–Gardens project.