Escape from native range enemies can give invasive species a competitive edge according to the enemy-release hypothesis. While more commonly associated with predators and herbivores, release from belowground microbial antagonists has been recently demonstrated to benefit invasive plants. Biogeographic variation in dominance and comparisons of soil communities suggest that invasive European Phragmites australis may have also benefitted from belowground enemy release in North America. Here we examine the effects of native range (Europe) versus introduced range (North America) soil communities on European native and North American introduced P. australis using a reciprocal inoculation seedling growth experiment. Contrary to the enemy-release hypothesis, we found that North American-introduced P. australis was sensitive to soil community origin in that the seedlings grown in European soil communities (native) had higher total biomass than seedlings grown in North American soil communities (introduced). This pattern was not observed in the European native P. australis seedlings which had similar biomass when grown with North American or European soil communities. Notably, introduced P. australis had higher biomass than native P. australis regardless of which soil community it was grown in, suggesting a growth-defense tradeoff. Though the relative abundance of mutualists and pathogens composition did not differ between the two ranges, an indicator analysis revealed that mutualistic fungi and bacteria were key components of European soil communities but not in North American communities. Interestingly, North American soil communities had lower β-diversity than European communities suggesting higher levels of community conservation amongst North American populations. This research represents the first evidence of growth-defense trade-offs in introduced P. australis and offers a novel mechanism for understanding the invasion of P. australis in North America.

Field sampling

North American and European Phragmites seed and soil were collected from three different habitats, tidal brackish wetland, freshwater marsh, and inundated urban areas. All plant material was confirmed to be haplotype “M” using visual identification methods (Swearigan & Saltonstall 2010). This stratification allowed us to detect the effects of range-level differences in soil community across our different response variables and account for environmental filtering through sampling. North American sites were located along the mid-Atlantic coast of the United States, the urban (NA-U) and freshwater (NA-F) sites were located around the greater Philadelphia, PA (USA) area, while the tidal site (NA-T) was located 150 km south on the eastern shore of Maryland along King’s Creek, MD (USA) (Table 1). This site is a well-studied and extensively sampled site previously used in a multitude of Phragmites studies and ranges in salinity between 0.1-8 ppt (Meadows & Saltonstall 2007; Mozdzer & Zieman 2010; Yarwood et al. 2016). European seeds and soil were collected along a 23-km transect running along the Bay of Aarhus in Denmark (Table 1). The freshwater (EU-F) and urban (EU-U) sites were located inland from the bay while the tidal site (EU-T) was on the water’s edge and experienced salinities ranging from 14-23 ppt. A total of 6-8 inflorescences were collected at each site in the autumn of 2017 and were subsequently stripped to bare seed, pooled, and stored at room temperature until germination. Soil inocula were obtained through bulking and homogenizing 6 soil cores (20.3 cm diameter x 30.5 cm depth) collected at the base of randomly selected Phragmites shoots to ensure a site-level representative sampling of the rhizosphere-associated soil community. During homogenization, the inocula was processed to remove live plant material such as rhizomes and roots. The soil was stored at 4°C until the time of use. A soil import permit was acquired from the Animal and Plant Health Inspection Service (APHIS) a subdivision of the United States Department of Agriculture (USDA). All soil was handled and disposed of in accordance with the issued permit.

Greenhouse Experimental Set-up & Design

All seeds were surface sterilized in two batches in the first week of January 2018 in 70% ethanol and then 3% bleach solution and plated on 2% water-agarose (no added nutrients) for germination in a Conviron growth chamber set to 30°C (16 hr)-15°C (8 hr) day/night cycle (Conviron, Pembina, North Dakota). Two weeks post-plating, seedlings (~1 cm tall) were potted in 6.35 cm diameter x 35.6 cm 1L pots filled with 700 g (dry) of autoclave-sterilized Sphagnum peat moss as well as 100 g (wet) of the corresponding soil inocula (sterilized treatments received 100g of sterilized inocula). Earth Juice Microblast-liquid micronutrients were added to each pot during planting to help seedlings overcome transplant shock at a concentration of 660uL/L of water (Earth Juice, Chico, CA, USA). Seedlings were planted in the center of the pot with enough media to cover all visible root tissue. Each pot was placed in its own 13.9-cm tall water reservoir to ensure minimal cross-contamination between different soil communities.

The seeds and soil from each of the six populations (NA-U, NA-F, NA-T, EU-U, EU-F, and EU-T + sterilized soil) were planted in a fully factorial manner, giving a total of 42 different soil x seed combinations, and replicated four times (6 seed populations x (6 soil populations) x 4 replicates = 144 pots) (Fig. 1). Pots were arranged in racks corresponding to each inoculum source to minimize chances of accidental cross contamination. Pots were randomly rotated every two weeks between the two benches and pots were randomly moved within each rack as to reduce environmental effects. All pots were top watered every other day until the water reservoir was filled, and average daily light intensity (0.504 ± 0.019 mmol m^-2 s^-1) and temperature (25.31°C ± 0.10) were recorded using the Onset HOBO Pendant^® Temperature/Light 8K Data Logger (Onset, Bourne, MA). Dead seedlings were replaced with the corresponding seedlings from maintained populations up until day 120; thereafter pots with dead plants were omitted from the study. Plants were harvested on day 230 after planting. If seedlings accumulated less than 0.5g of biomass (dry weight) at the time of harvest, they were omitted from the study.

Harvest was performed 230 days post-planting. Aboveground biomass was clipped at the soil line and oven-dried at 65°C to constant mass. Belowground biomass (roots and rhizomes) was washed from the associated soil using a hose and a series of 3mm and 1mm sieves and dried at 65°C to constant mass (Roots Lab¹). The dried aboveground biomass and the dried roots and rhizomes were sorted and weighed separately and summed to obtain total biomass.

Soil community sequencing

Ten grams of rhizosphere soil was collected from each pot planted with NA-U plants during harvest. After collection, samples were immediately flash-frozen in liquid nitrogen and stored at -80ºC. In total, 21 soil samples were collected, 3 samples from each of the soil sources (NA-U, NA-T, NA-F, EU-U, EU-T, EU-F, Sterilized). Total DNA was extracted using the Qiagen DNeasy Powersoil Kit, extractions were performed following the provided kit protocol, with a small modification of using 200 ml of phenol:chloroform:isoamyl alcohol instead of the provided bead solution to maximize DNA yields (Sigma-Aldrich, St. Louis, MO, USA; Qiagen, Hilden, Germany).

DNA was quantified using QuantiFlour dsDNA System (Promega, Madison, WI). 12.5 ng of DNA was used as input into Illumina’s 16S / ITS metagenomics Library preparation workflow. The V4 region of the bacterial 16S rRNA gene sequences and ITS2 region of fungal rRNA were amplified using the primer pair containing the gene‐specific sequences and Illumina adapter overhang nucleotide sequences. 16S rRNA was amplified using forward primer 515f (5’TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTGCCAGCMGCCGCGGTAA) and reverse primer 806r (5' GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGGAC-TACHVGGGTWTCTAAT). ITS rRNA amplification was amplified using forward primer ITS3 (TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGCATCGATGAAGAA-CGCAGC) and reverse primer ITS4 (GTCTCGTGGGCTCGGAGATGTGTATAAGAGACA-GTCCTCCGCTTATTGATATGC).

Amplicon PCR was performed to amplify the template out of input DNA samples. Briefly, each 25 μL of polymerase chain reaction (PCR) reaction contains 12.5 ng of sample DNA as input, 12.5 μL 2x KAPA HiFi HotStart ReadyMix (Kapa Biosystems, Wilmington, MA), and 5 μL of 1 μM of each primer. PCR reactions were carried out using the following protocol: an initial denaturation step performed at 95°C for 3 min followed by 25 cycles of denaturation (95°C, 30 s), annealing (55°C, 30 s), and extension (72°C, 30 sec), and a final elongation of 5 min at 72°C. The PCR product was cleaned up from the reaction mix with Mag-Bind RxnPure Plus magnetic beads (Omega Bio-tek, Norcross, GA).  A second index PCR amplification, used to incorporate barcodes and sequencing adapters into the final PCR product, was performed in 25 μL reactions, using the same master mix conditions as described above. Cycling conditions were as follows: 95°C for 3 minutes, followed by 8 cycles of 95°C for 30”, 55°C for 30” and 72°C for 30”. A final, 5 minutes elongation step was performed at 72°C.  The libraries were normalized with Mag-Bind® EquiPure Library Normalization Kit (Omega Bio-tek, Norcross, GA) and then pooled. The pooled library ~600 bases in size was checked using an Agilent 2200 TapeStation and and sequenced (2 x 300 bp paired-end read setting) on the MiSeq (Illumina, San Diego, CA). Results were analyzed via Illumina BaseSpace 16S metagenomics application using module version Isis v2.5.35.6 and Illumina-curated version of the May 2013 Greengenes taxonomic database.

Data analyses

All data were analyzed in RStudio version 4.0.3. We assessed the effect of soil range, seed range, and their interaction on total biomass, shoot growth, shoot emergence, and percent guild composition (“endophyte”, “plant pathogen/plant parasite”, “mycorrhizae”, and “plant saprotroph”) using either linear mixed-effects models (LME) or generalized least squares models (GLS) using the “nlme” R package (Pinheiro et al. 2023). The model selection process starts with a base GLS model for backward selection using soil range * seed range as fixed effects. Then, we tested whether the random effect of soil and seed habitat creates a better model than a model without the incorporation of random effects. We then selected the model with the lowest AIC value. We determined whether there are heterogeneous variances across seed and soil habitats and fit an alternative variance structure to the model. Again, AIC values are assessed between models, and the best model is selected. We then run a type III ANOVA on the best model to assess the significance of soil and seed range in explaining the variance of the response variable and perform model validation by plotting fitted values against residuals, each of the fixed effects against residuals, and a histogram of the residuals alone. Tukey’s HSD values were also obtained for the soil range*seed range comparisons using the function “glht” in the R package “multcomp” (Hothorn et al. 2008).

Indicator analysis was used to assess the level of conservation within North American and European 16s/ITS soil communities. We performed indicator analysis on16s/ITS genus-level OTU reads of the sampled soil communities at a probability threshold of p<0.01 using the “indval” function from the R package “labdsv” (version 2.1-0) (Roberts 2023).

Permutational analysis of variance (PERMANOVA) and principal coordinate analysis (PCoA) was performed to assess whether 16s and ITS soil communities on the genus level from North America and Europe differed significantly from one another. PERMANOVA was performed using Bray-Curtis dissimilarity values for the 16s/ITS soil communities using the “adonis” function in the R package “vegan”, Tukey’s HSD values were also attained for the group comparisons using the function “TukeyHSD” in R package “vegan” (Oksanen et al. 2022). Principal coordinate analysis (PCoA) was performed for the 16s/ITS soil communities on the genus level in R using the function “metaMDS” in R package “vegan”, on Bray-Curtis distance matrices (Oksanen et al. 2022). To assess multivariate homogeneity of dispersion between North American and European 16s/ITS soil communities was calculated using the mean distance of each sample to the centroid for both in R using the “betadisper” function in R package “vegan” (Oksanen et al. 2022).

FUNGuild was used to assign ecological roles to fungal OTUs at the species level (Nguyen et al. 2016). FUNGuild was run using a Python 3 environment (Van Rossum & Drake 2009). The FUNGuild output assigned each fungal OTU a guild classification that lists known ecological roles/functions. This output was parsed to extract OTUs that fall into the guild classifications of “endophyte”, “plant pathogen/plant parasite”, “mycorrhizae”, and “plant saprotroph”. The counts of each of the extracted OTUs were acquired for each sample. All OTUs belonging to the same classification were then summed and divided by total OTUs in each respective sample to derive the percent composition of each guild.