Soil disturbance affects plant growth via soil microbial community shifts

Recent advances in climate research have discovered that permafrost is particularly vulnerable to the changes occurring in the atmosphere and climate, especially in Alaska where 85% of the land is underlain by mostly discontinuous permafrost. As permafrost thaws, research has shown that natural and anthropogenic soil disturbance causes microbial communities to undergo shifts in membership composition and biomass, as well as in functional diversity. Boreal forests are home to many plants that are integral to the subsistence diets of many Alaska Native communities. Yet, it is unclear how the observed shifts in soil microbes can affect above ground plant communities that are relied on as a major source of food. In this study, we tested the hypothesis that microbial communities associated with permafrost thaw affect plant growth by growing five plant species found in Boreal forests and Tundra ecosystems, including low-bush cranberry and bog blueberry, with microbial communities from the active layer soils of a permafrost thaw gradient. We found that plant growth was significantly affected by the microbial soil inoculants. Plants inoculated with communities from above thawing permafrost showed decreased growth compared to plants inoculated with microbes from undisturbed soils. We used metagenomic sequencing to determine that microbial communities from disturbed soils above thawing permafrost have differences in taxonomy from microbial communities in undisturbed soils above intact permafrost. The combination of these results indicates that a decrease in plant growth can be linked to soil disturbance driven changes in microbial community membership and abundance. These data contribute to an understanding of how microbial communities can be affected by soil disturbance and climate change, and how those community shifts can further influence plant growth in Boreal forests and more broadly, ecosystem health.

above thawing permafrost showed decreased growth compared to plants inoculated with microbes 23 from undisturbed soils. We used metagenomic sequencing to determine that microbial communities 24 from disturbed soils above thawing permafrost have differences in taxonomy from microbial 25 communities in undisturbed soils above intact permafrost. The combination of these results indicates 26 that a decrease in plant growth can be linked to soil disturbance driven changes in microbial 27 community membership and abundance. These data contribute to an understanding of how microbial 28 Soil disturbance affects plant growth via microbes 5 We conducted the growth experiment in the Institute of Arctic Biology Greenhouse at the 128 University of Alaska Fairbanks (UAF), consisting of five plant species: Vaccinium vitis-idaea (low-129 bush cranberry), Vaccinium uliginosum (bog blueberry); Picea mariana (black spruce); Ledum 130 groenlandicum (Labrador tea); and Chamerion angustifolium (fireweed), hereafter all plants will be 131 referred to by their colloquial names. We obtained all seeds from Alaskan sources, and we surface-132 sterilized them prior to planting. We placed five to ten seeds of one plant type into a pot (SC10 133 Cone-tainers, Ray Leach; USA) filled with a mixture of densely packed sterile vermiculite and 134 Canadian peat (ProMix), with 1.5 g of soil containing the treatment microbial soil inoculant ( Figure  135 1). Sixty-four pots were planted for each plant type, with 16 unique soil inoculants per each of four 136 treatments, for a total of 320 pots. We randomized all pots across seven RL98 trays (Ray Leach, 137 USA) to reduce possible bias from temperature or light gradients within the greenhouse. Plants were 138 maintained under 12hr light cycles and watered once daily. 139 Following the first month of growth past germination, we measured the plants biweekly for 140 the first three months, recording their leaf count and height to the nearest millimeter. We stopped 141 black spruce needle counts in November 2018, when needle counts exceeded 200 and we were 142 unable to accurately measure. By October 2018, all the fireweed (n = 61) had started to decline in 143 height measurements, so on 10 October 2018, we clipped above ground living plant material at the 144 soil surface and then oven dried plants 55 ˚C to a constant mass before determining biomass 145 produced. The bog blueberry, black spruce, Labrador tea, and low-bush cranberry continued to grow 146 and be measured monthly. We then harvested above ground biomass per plant type in March 2020, 147 and below ground biomass in April 2020. We oven dried plants at 55˚ C to a constant mass before 148 weighing. 149 150

Statistical Analysis 151
All statistical analyses were performed in R version 3.5.2 (R Core Team, 2014

DNA Extractions and Metagenomic Sequencing 162
In order to analyze the microbial communities used to inoculate plants in the growth 163 experiment, we performed shotgun metagenomic sequencing on each individual core. To do this, we 164 extracted and purified total genomic DNA from approximately 250 mg of soil per homogenized soil 165 core using the DNeasy PowerSoil kit (Qiagen; Germany) following manufacturer instructions. We 166 quantified the yield and the quality of the DNA extracted using a NanoDrop One spectrophotometer 167 (Thermo Scientific; USA) and a Qubit (Thermo Scientific; USA). Following DNA extractions, we 168 randomly divided the 48 cores into four sequencing runs. We prepared the DNA sequencing library 169 using the Oxford Nanopore Technologies Ligation Kit with Native barcodes to multiplex 12 samples 170 (SQK-LSK109, ONT; UK). We diluted sample DNA to 400 ng for input and followed the kit 171 according to the manufacturer instructions. We then sequenced DNA using a MinION and R9.4.1 172 and R9.5 flow-cells (FLO-MIN106) ( Table S1). The sequencing runs each lasted 48 hours. 173 174

Plant Growth Experiment 194
Our results indicate that soil microbial communities from across the thaw gradient 195 differentially affected plant productivity (Figures 2-6; Tables S3-S8). Most plant types responded 196 negatively when grown in soils inoculated with microbial communities from the most disturbed 197 FPES soil compared to either inoculants from semi disturbed or undisturbed soils, or the sterile 198 treatment ( Figure S2; Table S8).  (Table S3), there was no significant difference 205 between height or leaf number for bog blueberry plants grown in UD, SD, or ST treatments. While 206 the source of inoculant was a significant factor in above ground and below ground biomass, the 207 responses did not follow the same trends as height and leaf count. For above ground biomass the post 208 hoc Tukey HSD test showed that SD plants had more mass on average and differed significantly 209 from MD and ST, SD plants did not differ significantly in biomass production (Table S8). 210 Low-bush Cranberry. Low-bush cranberry showed the same trends to bog blueberry when 211 grown in soils from the most disturbed treatment site ( Figure 3). Low-bush cranberry height and leaf 212 count significantly decreased when grown with inoculant from the MD site compared to when grown 213 with inoculant from the SD or UD sites, or with ST inoculant (Table S4, Table S8). There was no 214 significant difference in either height or leaf count between low-bush cranberry grown in UD, SD, or 215 ST treatments. Above ground biomass increased when plants were grown with soil from the SD 216 compared to plants inoculated with UD, MD, and ST treatments. Low-bush cranberry below ground 217 biomass did not differ significantly depending on the soil inoculant treatment. 218 Labrador Tea. Labrador tea plants displayed nearly identical trends to both bog blueberry 219 and low-bush cranberry when grown in soils inoculated with microbial communities from the most 220 disturbed treatment site (Figure 4). At the time of final measurements, the mean height and leaf count 221 of Labrador tea grown in the MD treatment was significantly smaller compared to low-bush 222 cranberry grown SD, UD, or ST treatments (Table S5). There was no observed difference between 223 Labrador tea grown in UD, SD, or ST treatments. Labrador tea above ground biomass showed that 224 MD plants weighed in at a significantly smaller amount than plants grown in soils inoculated with 225 Soil disturbance affects plant growth via microbes 8 UD or SD treatments. No differences were observed between below ground biomass measures (Table  226 S8). 227 Fireweed. In contrast to cranberry, bog blueberry, and Labrador tea, fireweed did not show a 228 significant difference in mean leaf count or height ( Figure 5); however, fireweed plants grown in 229 inoculant from MD site showed a significant decrease in above ground biomass compared to those 230 grown in UD, SD, or ST treatment soils ( Figure 5; Table S6, Table S8).

Microbial Community Analysis 241
We sequenced 48 metagenomes through four multiplexed MinION runs, and after 242 demultiplexing, total sample reads showed a mean read length of 2594bp, a N50 of 5,531bp, and a 243 mean quality score of 13 (Table S2). 244 The analysis using Kraken 2 and then Bracken resulted in a mean of 41% (range, 30-60%) of 245 reads being classified as bacterial (60,034 reads). The percent classified did not show a correlation to 246 the read depth of the sample ( Figure S3). Of those classified reads, we identified the 10 most 247 common bacterial families across all samples, based on normalized relative abundance, to be: Acidobacteriaceae was found to be significantly enriched in the UD soil microbial communities, 256 present at a mean relative abundance of 18% compared to the SD (5%) and MD (>5%) soil 257 communities ( Figure 7C). The family Comamonadaceae was found to be present at significantly 258 Soil disturbance affects plant growth via microbes 9 higher relative abundances within the MD soil microbial communities with a mean of 16%, 259 compared to in UD (5%) and SD (8%) soils ( Figure 7B). We can attribute the changes in plant growth to the variation in microbial communities across 280 the active layer of soil above the permafrost thaw gradient. We hypothesized that if the soil microbial 281 communities were to change depending on the level of associated permafrost thaw, so would the 282 growth of plants in that soil. Consistent with our hypothesis, plants (specifically bog blueberry, low-283 bush cranberry, Labrador tea, and fireweed) grown with microbial inoculum from MD soils exhibited 284 lower growth than plants grown in soils with no added microbes or with inoculants from the UD or 285 SD soils. These results were similar to a study performed on boreal plant species using microbial 286 inoculums from thermokarst bogs (disturbed soils) and permafrost plateaus (undisturbed soil). 287 Schütte et al. (2019) showed that bog blueberry and marsh-cinquefoil (Potentilla palustris) grew 288 significantly worse when inoculated with soil microbes from thermokarst bogs than when inoculated 289 by permafrost plateau soil microbes. 290 Soil disturbance affects plant growth via microbes 10 We found no apparent relationship between black spruce growth and the initial soil microbial 291 inoculant. These results correspond to a study (Sniderhan and Baltzer, 2016) analyzing the growth of 292 black spruce across a lateral permafrost thaw gradient in Scotty Creek, Canada, which found that the 293 lateral thaw rate of permafrost did not appear to be a driver of black spruce growth dynamics. 294 Controlled warming experiments meant to simulate the quickly warming northern latitudes have 295 shown that black spruce shoot length tends to increase with warming air and soil temperatures 296 Enterobacterales, are both ubiquitous in soil microbiomes. 326 In contrast, the MD soil cores displayed a much higher relative abundance of bacteria from 327 the order Comamonadaceae, a subclass of Proteobacteria, than either the UD or SD cores. Members 328 of the order Comamonadaceae are known to exhibit pathogenic effects on a variety of plants, 329 including in agricultural settings (Willems et al., 1991). It is possible that the higher presence of 330 groups including known plant-pathogenic bacteria found in the MD soil microbiomes is leading to 331 the decrease in associated plant growth, caused by a disruption in nutrient cycling and direct 332 alterations to the plant rhizospheres. It is important to note that our results suggest that a decrease in 333 plant growth can be linked to changes in the taxonomic diversity of microbial communities, but 334 further experimentation investigating the functional genetic potential of the microbial communities 335 will be required to elucidate the mechanisms that underlie the patterns described here. 336 Having an increased understanding of how microbial communities that reside above 337 permafrost affect plant growth is important for predicting effects of permafrost thaw on plant 338 communities and ecosystem health. Specifically, for potentially predicting effects on plants, such as 339 bog blueberry, low-bush cranberry, and Labrador tea, that are relied on as common food sources for    Comamonadaceae, per soil core. Solid line represents mean relative abundance per treatment group. 560 (C) Relative abundance of bacterial family, Acidobacteriaceae, per soil core. Solid line represents 561 mean relative abundance per treatment group. 562