Native soil amendments combined with commercial arbuscular mycorrhizal fungi increase biomass of Panicum amarum

Coastal dune restorations often fail because of poorly performing plants. The addition of beneficial microbes can improve plant performance, though it is unclear if the source of microbes matters. Here, we tested how native soil amendments and commercially available arbuscular mycorrhizal (AM) fungi influenced performance of Panicum amarum, a dominant grass on Texas coastal dunes. In a greenhouse experiment, we manipulated the identity of native soil amendments (from P. amarum, Uniola paniculata, or unvegetated areas), the presence of soil microbes in the native soil amendments (live or sterile), and the presence of the commercial AM fungi (present or absent). Native soils from vegetated areas contained 149% more AM fungal spores than unvegetated areas. The commercial AM fungi, when combined with previously vegetated native soils, increased aboveground biomass of P. amarum by 26%. Effects on belowground biomass were weaker, although the addition of any microbes decreased the root:shoot ratio. The origin of native soil amendments can influence restoration outcomes. In this case soil from areas with vegetation outperformed soil from areas without vegetation. Combining native soils with commercial AM fungi may provide a strategy for increasing plant performance while also maintaining other ecosystem functions provided by native microbes.

www.nature.com/scientificreports/ to consider plant identity when collecting soil amendments for restoration, because there may be a build-up of species-specific pathogens that makes it unadvisable to collect soil near the species that is being restored.
As an alternative to using native soil amendments to introduce soil microbes, commercial AM fungi can be added to restorations. Using commercial AM fungi can increase plant performance 16 while avoiding the negative effects of native pathogens. However, commercial AM fungi may not provide the same benefits as locally-adapted mutualists 26,27 . Many commercial mixes likely contain early successional fungi that may act more like parasites than mutualists 26 . To combat this, suppliers are producing more diverse mixes of AM fungi derived from native ecosystems that may provide greater and more tailored benefits to plants. When remnant prairie AM fungi communities containing higher diversity were added to prairie restorations, plant biomass and ecosystem functions increased compared to lower diversity treatments 28,29 .When added to prairie restorations, more diverse AM fungi from remnant prairies yielded greater plant performance as well as increasing other ecosystem functions when compared to treatments 28,29 . However, it is unclear whether the benefits from these AM fungi extend to ecosystems other than native tallgrass prairies.
If added together, commercial AM fungi will likely interact with native soil microbial communities to influence plant performance. The addition of commercial AM fungi can significantly alter native soil community composition 30 . Such alterations may lead to greater plant productivity if they increase the diversity of plant mutualists. For instance, increased AM fungal species richness can increase plant productivity 31 . However, whether microbial diversity positively influences plant productivity can depend on the identity of the microbes 32 , so it is important to document how increased diversity resulting from the addition of both commercial and native microbes influences plants. AM fungi can also protect plants from soil pathogens 33 . Therefore, adding commercial AM fungi with native soil amendments may help increase plant performance by decreasing the effect of pathogens that may be present in native soils while increasing the possibility of including necessary locally-adapted microbiota.
Texas Gulf Coast dunes border the Gulf of Mexico and extend into coastal tallgrass prairies. In addition to providing habitat for insects, birds and mammals, these dune systems protect inland ecosystems from wind and waves during coastal storms 34 . Dunes also protect human developments from storm disturbance resulting in a high economic value. With forecasted sea level rise and the increased intensity and frequency of storms, dune stabilization has become a primary focus for restoration practitioners 3 . Revegetation is often essential for the stabilization of coastal foredunes [35][36][37][38][39] . Panicum amarum is a dominant grass species that helps stabilize dunes with its above-and below-ground biomass, and it is often used in restoration. However, restorations with P. amarum sometimes suffer from high plant mortality and erosion 37 . Panicum amarum associates with AM fungi 15 .
In this study, we tested how native soil amendments and a diverse commercial AM fungi mix derived from native prairies influenced the performance of a dominant grass species, Panicum amarum (Elliott), that is commonly used in Gulf Coast dune restorations. Dune soils are often nutrient-poor, and both soil organic matter and nutrients can be limited to areas around plants 40,41 . Dune restorations along the Gulf Coast of Texas, USA, are often conducted in areas with no plants, and AM fungi as well as other plant associated microbes are rarer in unvegetated dunes compared to vegetated dunes 10,16 . We hypothesized that (i) the effect of native soil amendments on P. amarum performance will depend on soil identity (i.e., whether the soil was collected underneath a conspecific, heterospecific, or no plant), and (ii) the addition of commercial AM fungi will interact with native soil amendments to increase growth of Panicum amarum.

Methods and materials
To test how native and commercial soil microbes influenced the performance of Panicum amarum we established a fully factorial greenhouse experiment where we manipulated the identity of native soil amendments (from P. amarum, Uniola paniculate (Linnaeus), or unvegetated areas), the presence of soil microbes in the native soil amendments (live or sterile; to tease apart effects of differences in soil communities and abiotic soil properties), and the presence of commercial AM fungi (present or absent).
Collection of native soil amendments. We collected soil for the native soil amendments from vegetated dunes on Galveston Island, TX. All research complied with institutional, local, and national guidelines for the collection and research use of plant and soil microbes. To obtain a wide range of soil communities, we selected three sites that were approximately 3 km apart. At each site, we collected eight replicates of each soil with three different "soil identities": the rooting zone of Panicum amarum, the rooting zone of Uniola paniculata (another common dune grass), and bare ground that was free of vegetation for at least one meter. Replicate soil samples were collected at least three meters apart from one another and were kept separate to avoid pseudoreplication. Prior to planting, the soil was stored in a 4 °C cold room to slow microbial processes.

Plant propagation.
To maintain similarity to restoration practices and maintain control of the plants life history we chose to propagate plants from seed. Seeds of P. amarum were acquired from Roundstone Native Seed (Uptown, KY). Prior to planting, we surface sterilized the seeds in a 0.08% hypochlorite solution for 10 min and rinsed them with deionized water. Seeds were planted in sterilized play sand (Quikcrete Atlanta, GA) and watered every 2-3 days. Play sand was sterilized by autoclaving twice at 121 °C for one hour, with 24 h in between cycles.
Experiment. Upon emergence of the first true leaf, P. amarum seedlings were transferred into 262 mL conical pots (5 cm diameter × 17.8 cm depth; Stuewe & Sons, Tangent, OR) filled with 225 mL of sterilized play sand. Pots were lined with paper towels to stop sand from draining out of the pots but allow water to pass through. When transplanting the seedlings, we added 25 ml of native soil amendments to the root zone from one of the www.nature.com/scientificreports/ three soil identities (from P. amarum, U. paniculata, or unvegetated areas) that was either live or sterile (presence or absence of native soil microbes). We sterilized the native soil amendments by autoclaving at 121 °C for one hour, twice, with 24 h in between cycles. We added 5 g of commercial AM fungi (MycoBloom, Lawrence, KS) to half of the pots in each native soil amendment identity × native microbe presence combination. MycoBloom was developed to preserve the natural diversity of AM fungal communities found in remnant Midwest prairies; therefore, we expect that this commercial AM fungi is more likely to have positive effects on native (nonagricultural) plant performance than other commercial AM fungi which are targeted to non-C4 plant species. Statistical analyses. We tested the effects of our treatments using ANOVAs with the response variables of aboveground plant biomass, belowground plant biomass, total plant biomass, and root:shoot ratio. We used ANOVAs with the factor site, followed by native soil amendment identity, native microbe presence, commercial AM fungal presence, and all possible interactions. Site was excluded from interactions as we were interested only in its main effect. Analysis of variance was conducted using type I (sequential) sum of squares. The order in which we analyzed the variables did not qualitatively change the results. Root:shoot ratio was log-transformed to improve normality. Samples that died during the experiment were removed from the analysis of root:shoot ratio. We tested how native soil amendment identity influenced the abundance of native AM fungal spores using ANOVA (type I sum of squares) with the main effects of native soil amendment identity and site. Changing the order in which the main effects were analyzed did not qualitatively effect the results. Spore abundance was logtransformed to improve normality. All statistics were conducted in the base R packages 43 .

Results
Across all treatments, the presence of commercial AM fungi increased the aboveground biomass of P. amarum by 13.6% (F 1,274 = 5.21, P = 0.02, Table 1), but the degree of increase depended on the identity of the native soil amendment (F 2,274 = 4.78, P = 0.01, Table 1, Fig. 1). When commercial AM fungi were added to soils from U. paniculata aboveground biomass of P. amarum increased by 35.5%. However, commercial AM fungi did not cause a significant increase in aboveground biomass when native soil amendments came from bare ground. The interaction between commercial AM fungi and the identity of the native soil amendment produced qualitatively similar results for total biomass, although results from the post hoc test revealed no significant differences (Tukey's HSD P = 0.30) (Fig. S1). The effect of the native soil amendments on aboveground and total biomass was not dependent on the presence of native microbes in the soils (i.e., no significant main or interactive effect of native soil microbe presence) (aboveground F 2,274 = 0.40, P = 0.67 and total F 2,274 = 0.53, P = 0.59). The presence of native microbes did interact with commercial AM fungi to influence P. amarum belowground biomass (F 2,274 = 3.75, P = 0.05) and root:shoot ratio (F 2,255 = 4.08, P = 0.04, Table 1). There was a trend for commercial AM fungi to decrease belowground biomass for plants with no native microbes by 15.2% (Fig. S2); however, the posthoc test revealed no significant difference (Tukey's HSD: P = 0.29). In the absence of native soil microbes, the addition of commercial AM fungi significantly decreased root:shoot by 19.4% (F 2,255 = 4.08, P = 0.04, Table 1, Fig. 2). In fact, the presence of any microbial addition (AM fungi or native soil microbes) significantly decreased the root:shoot ratio by an average of 17.5% relative to the sterile native soil, no commercial AM fungi control (F 2,255 = 4.08, P = 0.04, Table 1, Fig. 2). The identity of native soil amendments also influenced root:shoot ratio (F 2,274 = 4.62, P = 0.01, Table 1). Native soil amendments from P. amarum produced a lower root:shoot ratio than amendments from bare ground or U. paniculata (Fig. 3).

Discussion
The use of native soil amendments in combination with commercially available AM fungi derived from a native prairie significantly influenced the performance of the Gulf Coast dune grass, Panicum amarum. The commercial AM fungi increased aboveground biomass for P. amarum, but only when it was grown with native soil collected from the native dune grasses P. amarum or U. paniculata. Changes in root:shoot ratio also occurred in our treatments, with plants containing a higher proportion of roots in the absence of native or commercial microbes. Belowground biomass, however, was relatively unresponsive to our treatments -although there was a trend for the commercial AM fungi to decrease belowground biomass in the absence of live native microbes. In sum, the use of live native soil from vegetated areas in combination with a diverse commercial AM fungi mix derived from native prairie increased aboveground biomass without significantly changing belowground biomass, while also increasing the abundance of native AM fungal spores.
Our results indicate that soil origin is important to consider when using native soil amendments in restorations. We found that in order for native soil amendments to enhance the effects of the commercial AM fungi on aboveground biomass, the soil amendments had to come from vegetated areas. Contrary to our predictions, this effect was not driven by the presence of native soil microbes. This is particularly surprising as P. amarum has many mycorrhizal-dependent congeners. Therefore, some other property of the native soils was driving the observed response. We suspect that nutrient availability was greater in soils from vegetated areas than unvegetated areas, as has been found in other studies 44 . Even though we added in a relatively small amount of the native soil Figure 3. Effect of the identity of native soil amendments on the root:shoot ratio for P. amarum. Bars indicate treatment means ± SE. Letters denote significant differences between native soil amendments estimated using linear models (P < 0.05).

Figure 4.
Effect of native soil amendment identity on the abundance of arbuscular mycorrhizal fungal spores. Spores were extracted from 25 ml of soil. Bars indicate treatment means ± SE. Letters denote significant differences between native soil amendments estimated using linear mixed-effects models (P < 0.05). www.nature.com/scientificreports/ amendments, a small increase in soil nutrients may have had a big impact on plant performance in the nutrientpoor sandy background soil. If the native vegetated soils did increase nutrient availability, they may have provided a greater pool of nutrients for commercial AM fungi to access. The identity of the native soil amendments also influenced the root:shoot ratio of P. amarum, with the lowest ratio in soils from P. amarum. While it is unclear what is driving this relationship, it is possible that soils from conspecifics have a greater availability of nutrients needed by P. amarum, reducing the need for the plants to produce more roots to seek out nutrients. The role of soil nutrients in driving P. amarum performance could be resolved by future work that measures nutrient availability in conspecific, heterospecific, and unvegetated soils followed by tests of how the addition of certain nutrients influence plant performance. This could lead to the development of fertilizers designed to increase plant performance in dune restorations. In contrast to native soil amendments, commercial microbes are more commonly used in restoration projects. Commercial microbes are easily obtained and, unlike native soils, unlikely to contain soil enemies that may decrease plant performance. However, there are concerns that commercial microbes may not provide the same benefits as native microbes that are locally adapted to site-specific conditions 27,29 . Our results demonstrate a case where commercial microbes increased plant performance more than native microbial inocula. However, this may not be true for all commercial inocula. Many commercial suppliers of AM fungal innoculum focus on fungi that are selected to improve agricultural crop performance. In contrast, the commercial AM fungi we used was developed from native prairies. This too may have generated a mismatch between fungi and plant-host as P. amarum is a coastal dune species. Our findings raise the question of how local native microbes need to be in order to provide the greatest benefits to restorations. This may be especially true as species overlap may occur between environments but local adaptation and life history may alter plant-microbe interactions. While our experiment was not designed to answer this question, we found that microbial communities from native prairies outperformed locally derived dune microbial communities. Future work that explicitly tests for local adaptation in microbial communities and documents a wider range of responses would greatly inform the management of microbial communities in restorations 45 .
There has been growing interest in using native soils, along with their native microbes, to improve restoration outcomes 17,46 . When the efficacy of native soil amendments is in question, our results suggest that bet-hedging by adding both a native commercial AM fungi mix and native soil amendments may be an effective strategy for maximizing plant performance. Even though our live native soil amendments contained AM fungi, the native AM fungi did not provide the same benefits as the commercial AM fungi. Different AM fungi can have different effects on plant performance 47 , so there may have been a difference in composition of AM fungal communities in the native soils and the commercial mix. Another likely explanation for the weak effect of dune-derived AM fungi is that their abundance in native soils may have been too low to elicit immediate strong responses from plants. Alternatively, interactions between the dune-derived AM fungi and native soil pathogens may have produced a neutral effect on plant growth. Finally, not all AM fungi are beneficial to their plant hosts in all situations 48 . These ideas could be tested with experiments that manipulate the composition, quantity, and presence of AM fungi and soil pathogens in trials with P. amarum.
Our experiment may underestimate the effects of plant-microbe interactions as they are often contextdependent 8 , and mutualisms may have stronger effects in more stressful environments 48 . Therefore, the positive effects of microbes may be even stronger in the relatively harsh dune environment than they were in the greenhouse. The lack of effects on belowground biomass may also have been limited by pot size, potentially masking any differences between treatments. Furthermore, we were unable to measure many other responses of interest that may be influenced by microbes, including long-term plant performance, plant resistance to and resilience following disturbance, and ecosystem functions such as carbon sequestration and soil development 49 . Future experiments that test the effect of commercial and native microbes on restoration performance in the field are needed. Importantly, while the microbes in native soil amendments only had weak effects on plant performance in our experiment, they may provide critical services that we were unable to measure in our experiment that would be evident in a longer-term field experiment 20 .
Given the importance of plant success for restorations, finding an effective way to increase plant performance is critical. By demonstrating how native soil amendments and a diverse commercial AM fungi mix independently and interactively influenced P. amarum, we are able to suggest a strategy for optimizing P. amarum growth. Using commercial AM fungal derived from native ecosystems in combination with soils collected from vegetated dunes can help grow more robust plants which may provide greater ecosystem services, including habitat for other species and buffering of inland areas from the sea during large storm events. Importantly, only small amounts of native soil amendments are needed to gain this benefit (less than 25 ml per plant), so the disturbance caused by collecting soils from intact dunes need not be widespread. We found no evidence of native dune microbes negatively affecting P. amarum, indicating no need for sterilization of native soil amendments. Furthermore, native microbes can contribute to other important to ecosystem functions, such as improving carbon and nitrogen cycling, stabilizing soil, and promoting plant diversity 20,50 . Therefore, restoration strategies that combine commercial mutualists with native soils have the potential to improve multiple restoration outcomes 51 .