Optimal filter pore size
We determined the size distribution of intracellular eDNA by separating the states of H. halys eDNA from the slurry using a peristaltic pump and filter combination. Intracellular eDNA is typically targeted using filters with pore sizes in the 1–10 µm range (Turner, Barnes, Xu, Jones, et al., 2014). We thus used polycarbonate track etched (PCTE) filters with 1, 3, 5, 8, and 10 µm pore sizes to identify the optimal capture pore size of intracellular eDNA. We chose PCTE filters because the manufacturing process of these membranes results in a highly consistent pore density and uniform pore size, making inadvertent loss or capture of eDNA due to presence of other pore sizes less likely.
We carried out the experiment in parallel using twenty Nalgene 150 ml plastic vacuum filter units (Nalgene Nunc International, Rochester, NY, USA) with each filter size represented by five replicates, four for experimentation and one as a negative control. We paired each filter unit with a 100 ml glass bottle containing 80 ml of deionized water and 15 µl of slurry (no slurry was added to the negative control). After filtering, we immediately extracted intracellular DNA using the HotSHOT method, which lyses cells via incubation at 95ºC in sodium hydroxide for one hour (Truett et al., 2000), then stored extractions at -20 C prior to qPCR analysis. We carried out qPCR in triplicate using an Applied Biosystems StepOne Plus real-time PCR machine (Applied Biosystems, Foster City, California, United States) amplifying 96bp of the internal transcribed spacer subunit 1 (ITS1) nuclear DNA via the TaqMan assay used in Valentin et al . (2016; 2018). We used 20 μl reactions with 500 nanomolar (nM) concentration of each primer, 250 nM of the probe, 1X TaqMan® Environmental Master Mix 2.0, and 2 μl of DNA, following a reaction protocol with an initial denaturing step of 96°C for 10 min, followed by 40 cycles of denaturing for 15 s and annealing and extension at 60°C for 1 min. We converted the DNA quantity data to copy number and averaged values for each qPCR technical replicate by sample. We then evaluated the response of within and between filter copy number to filter pore size using two iterations of generalized linear models (glm; R v.3.6.1). First, we treated filter pore size as a categorical variable, assessing the variability within and among filter size treatment groups. We then assessed whether mean eDNA abundance was positively or negatively correlated with filter pore sizes (in microns) by repeating the analysis and treating filter size as continuous variable.