Field collections and sampling strategies
All collections occurred at the Teller Wildlife Refuge in western Montana, USA (46.3219 N 114.1292 W). The Teller Wildlife Refuge is a 525-ha property that comprises riparian, floodplain, wetland, and upland habitat. Showy milkweed grows in numerous patches across the property, concentrated near irrigation ditches, on the floodplain, and in old fields where sub-irrigation provides enough water. We collected 6 individual leaves each from 20 randomly selected showy milkweed plants once on May 18th, 2016 before flowering, and again on September 6th, 2016 when seed pods were almost ripe. Plants were selected from a 500 m2 area of a field with a patchy distribution. All plants sampled were at least 3 m. apart. During the September collection we also collected as much of the root system as possible within the upper 20 cm of soil by destructively sampling each plant. Tissue was kept on ice upon collection, and then stored at -20 °C until further processing. Roots were first washed with tap water to remove all visible soil, and fine roots (<1 mm in diameter) were retained for further processing. Bulk root collections weighed between 0.3 and 0.8 g dry weight, per plant. Bulk leaf collections weighed between 1.2 and 2.5 g dry weight per plant. To remove microbial DNA from root and leaf exterior and characterize only colonizing microbes, all root and leaf tissue was surface sterilized in 70% ethanol for 1 min., 0.5% NaOCl for 1 min. and then rinsed 2X in sterile water. Imprints of sterilized tissue were made on growth medium to confirm the sterilization procedure. The absence of growth observed indicated successful surface sterilization. After sterilization, leaves and roots from the September collection were separated into two groups based on the two sampling strategies we chose to compare.
For Strategy 1, homogenizing after subsampling , we removed one leaf disc from 6 leaves per plant (at random locations) using an 11.5 mm in diameter corkborer and placed the six discs per plant into single 1.2 ml Eppendorf tubes. This resulted in a total surface area of approximately 62 mm2 per plant, with a dry weight of approximately 30.0 mg of leaf tissue. For roots, approximately 30.0 mg of dry root segments per plant was placed into individual 1.2 ml Eppendorf tubes. As many DNA extraction kits recommend a maximum volume of approximately 20-30.0 mg dry weight or 50.0 mg fresh weight, we have found this to be a commonly used standard (e.g., Busby et al., 2016; Gdanetz and Trail 2017; Haas et al ., 2018; Barge et al., 2019; Bunn et al., 2019), however, many studies lack sufficient detail on final extraction volumes. Subsampled root and leaf tissues were then freeze-dried using a Labconco Freezone benchtop freeze-dry system (Labconco, Kansas City, MO, USA). Subsampled tissue was homogenized using a 1600 MiniG tissue homogenizer and cell lyser (Spex SamplePrep, Metuchen, NJ, USA).
For Strategy 2, homogenizing before subsampling, all remaining root and leaf tissue from the September collection was subsequently freeze-dried and homogenized for each plant using the freeze-dry system and MiniG as described above. Approximately 30.0 mg dry weight per plant was then subsampled from the homogenized pool of root or leaf tissue and placed in 1.2 ml Eppendorf tubes for DNA extraction. All remaining foliar tissue from five plants collected in September (> 40X more tissue per plant, equaling 1241-2467 mg per plant) was then divided into approximately 250 mg replicates for additional DNA extractions to assess the extent of undersampling.