Root exudates and rhizosphere microbiomes jointly determine temporal shifts in plant-soil feedbacks

Plants influence numerous soil biotic factors that can alter the performance of later growing plants - defined as plant-soil feedback (PSF). Here, we investigate whether PSF effects are linked with the temporal changes in root exudate diversity and the rhizosphere microbiome of two common grassland species ( Holcus lanatus and Jacobaea vulgaris) . Both plant species were grown separately establishing conspecific and heterospecific soils. In the feedback phase, we determined plant biomass, measured root exudate composition, and characterized rhizosphere microbial communities weekly (eight time points). Over time, we found a strong negative conspecific PSF on J. vulgaris in its early growth phase which changed into a neutral PSF, whereas H. lanatus exhibited a more persistent negative PSF. Root exudate diversity increased considerably over time for both plant species. Rhizosphere microbial communities were distinct in conspecific and heterospecific soils and showed strong temporal patterns. Bacterial communities converged over time. Using path-models, PSF effects could be linked to the temporal dynamics of root exudate diversity, whereby shifts in rhizosphere microbial diversity contributed to temporal variation in PSF to a lesser extent. Our results highlight the importance of


Introduction
Plants have the capacity to influence their local abiotic and biotic soil environment (Bennett & Klironomos, 2019).These plant-driven changes in soil properties can influence the growth of conspecific or heterospecific plants that subsequently grow in the same soila process known as plant-soil feedback (PSF) (van der Putten et al., 2013).Beside abiotic factors, such as temperature, moisture and nutrient availability, the composition of soil microbial communities plays an important role in determining whether the PSF will be positive or negative (Kaisermann et al., 2017;Bennett & Klironomos, 2019).The accumulation of beneficial microorganisms, such as mycorrhizal fungi, typically leads to positive PSFs whereas the accumulation of pathogens commonly suppresses plant growth resulting in negative PSFs.Typically, plant-soil interactions exhibit short-and long-term temporal dynamics, and thus the direction and magnitude of PSFs can strongly depend on time (Bezemer et al., 2018;Dudenhöffer et al., 2018;Thakur et al., 2021;Zhao et al., 2021).To better understand temporal PSF dynamics, it is vital to simultaneously monitor the changes in plant performance and soil characteristics over time.Here, we focus on two key inter-dependent factors which are likely to drive temporal shifts in PSF: alterations of root exudation profiles and shifts in rhizosphere microbial communities.To this end, we examine the importance of these two factors in driving temporal shifts in PSF effects during the growth of two plant species grown in conspecific and heterospecific soils.

Accepted Article
The rhizosphere (the area surrounding plant roots in soil (Philippot et al., 2013)) is a hotspot of dynamic transformation of nutrients and contains complex populations of microorganisms (Philippot et al., 2013;Ling et al., 2022).Plant roots and soil microorganisms directly interact through root exudates, which typically comprise of primary metabolites such as sugars, amino acids, and organic acids, as well as a diverse set of secondary metabolites (Rovira, 1969;van Dam & Bouwmeester, 2016;Oburger & Jones, 2018).Besides being a source of carbon and nitrogen, root exudates have tremendous effects on soil microorganisms and the rhizosphere processes, such as nutrient mobilization.They are involved in the establishment of beneficial symbioses, and the production of signalling and defence compounds (Baetz & Martinoia, 2014;Oburger et al., 2014;van Dam & Bouwmeester, 2016).The composition and diversity of root exudates substantially differ among plant species and among individuals of the same species.Even within a single plant individual, exudate composition may vary among various plant developmental stages (Haichar et al., 2014;Zhalnina et al., 2018).However, current knowledge of variation in root exudates is mainly based on a handful of crop species like Avena barbata (Zhalnina et al., 2018), Zea mays (Hu et al., 2018), and Triticum aestivum (Oburger et al., 2014) as well as from the model species Arabidopsis thaliana (Chaparro et al., 2013) and lately from tree species (Weinhold et al., 2022).Only recently, studies have begun focusing on root exudates of common European grassland species confirming strong species identity effects (Steinauer et al., 2016;Herz et al., 2018;Dietz et al., 2019;Delory et al., 2021).Moreover, we still know little about how species-specific differences in root exudate composition and their diversity change over time when plants are exposed to different microbial Changes in root exudation patterns are known to affect the composition and relative abundance of microbial communities in the rhizosphere (Philippot et al., 2013;Zhao et al., 2021).Due to the continuous developmental changes of the plant individual, soil microbial communities underlie strong temporal dynamics (Hannula et al., 2019).The few studies that have examined the temporal variability in soil microbial communities indicate that their composition can vary at the scale of days (Zhang et al., 2011), months (Lauber et al., 2013;Hannula et al., 2019), and seasons (Mellado-Vázquez et al., 2019).Thus, it is likely that a seedling, a juvenile, an adult, or a senescing plant shape their soil microbial community differentially.Consequently, a single plant individual may get exposed to different soil microbial communities with different functional roles over its life history.A succeeding plantindependent of being of the same or of a different speciesis often exposed to the soil microbial community that was left behind by the plant that grew previously in the soil.Measuring how fast the succeeding plant-individual is capable to either adapt or re-shape the soil microbial community to its own benefit, is key to predict the strength and temporal dynamics of PSFs.For instance, seedlings and juvenile plants are considered to be more sensitive to the soil microbial legacy of the previous plant than adult plants (Elger et al., 2009;Hannula et al., 2021).The exact involvement of root exudate composition in shaping the soil microbial community over time is also poorly understood.Thus, it is of great importance to simultaneously study the effects of temporal changes during plant growth on root exudate dynamics along with soil microbial dynamics.

Accepted Article
Here, we experimentally examine whether temporal shifts in PSF are linked with temporal changes in root exudate composition and the rhizosphere microbial community of two common grassland species representing two different plant functional groups, the grass Holcus lanatus and the forb Jacobaea vulgaris.These species were chosen because previous studies have shown that both species grow worse in soil of the other species than in own soil (Bezemer et al., 2006;van de Voorde et al., 2012;Bezemer et al., 2018).We investigate root exudate and microbial dynamics in soils by studying their composition and diversity during the growth of these two plants.
Previous studies have shown that both species exhibit negative conspecific PSFs (Bezemer et al., 2006;van de Voorde et al., 2012;Bezemer et al., 2018).During the conditioning phase, we grew both plant species separately in pots establishing conspecific ("home-soil") and heterospecific ("away-soil") soil for both plant species (Figure 1).Hereafter, in the feedback phase, we grew individual plants from the seedling stage in soil conditioned by conspecifics or heterospecifics in a full-factorial design for 10 weeks.From week 3 onwards, we destructively harvested a subset of plants weekly to determine plant biomass.We expected (1) negative conspecific PSF effects on plant growth in both plant species in early plant growth stages due to a higher susceptibility of young plants to soil pathogens (Hersh et al., 2012).
Further, we expected (2) a shift in the strength and direction of PSF effects over time.For each harvest, we also analysed the root exudation profiles, using untargeted Liquid Chromatography-Time of Flight -Mass Spectrometry (LC-qToF-MS) and the microbial community composition (bacteria and fungi) of the rhizosphere soil.For these measures, we hypothesized (3) that

Article
temporal shifts in PSF effects in early growth stages depend on how plants alter their rhizomicrobiome through changes in their root exudation profile.
Seeds of both species were surface-sterilized (1 min in 2.5% sodium hypochlorite solution and rinsed with water afterwards) and germinated for 2 weeks on sterile glass beads in a temperature-controlled climate chamber set at 24°C light (16 hr), 20°C dark (8 hr) and 60% relative humidity.
In the conditioning phase, 80 1-L pots (10 x 10 x 11 cm) were filled with 1 kg sieved (1 cm mesh) and homogenized soil.40 seedlings of H. lanatus and 40 seedlings of J. vulgaris were transplanted to individual pots and randomly placed in a climate-controlled greenhouse set at 21°C light (16 hr), 16°C dark (8 hr) and 60% relative humidity.In week one to three 20 ml, in week four to

Accepted
Article six 50 ml and in week seven to ten 70 ml of demineralized water was added to each pot every second day.After 10 weeks of soil conditioning, we collected the conditioned soil of each pot and homogenized the soils by sieving (mesh size 1cm) per plant species (Figure 1).
In the feedback phase, we used J. vulgaris grown in tissue cultures to preclude variation in root exudate profiles due to genetic differences.In a climate room, J. vulgaris cuttings from a single genotype were asexually propagated in tissue culture using MS medium (Murashige and Skoog medium) with 100 mg/L benzylaminopurine (BAP) (16:8 h light:dark photoperiod, 20 °C).After 4 weeks, the cuttings were grown in MS medium without BAP for 10 days to produce roots.The genotype that was propagated was formerly collected from Meijendel (Wassenaar), The Netherlands.Seeds of H. lanatus were again purchased from Cruydt-Hoeck (Nijeberkoop, The Netherlands).Seeds for germination were treated as described above.We filled 2 L-pots (diameter: 12.3 cm, height: 13 cm) with 1.62 kg of sterile soil from the same field (γ-irradiated >25 KGray, Synergy Health) and 0.18 kg of conditioned soil, resulting in a 9:1 ratio.We then planted one individual per pot of J. vulgaris on J. vulgaris -conditioned soil ("Jacobaea home soil") and on H. lanatus -conditioned soil ("Jacobaea away soil").Further, we planted one individual of H. lanatus on H. lanatus -conditioned soil ("Holcus home soil") and on J. vulgaris -conditioned soil ("Holcus away soil").Plants that died within the first 10 days were replaced by new seedlings or cuttings.
Afterwards, plants were left to establish for 2 weeks before the first harvest.
Pots were watered every second day (week 3 -7; 50 ml and in week 8 -10; 70 ml) with demineralized water.The pots were kept in the greenhouse under the same conditions as above.We grew three replicate pots for each plant/soil  4) and time point (8) resulting in 96 pots.We destructively harvested plants for eight consecutive weeks always on Tuesdays between 8-10 AM (two to four hours after sunrise; February -April 2018; Wageningen, Netherlands: 51°58'12.00"N 5°40'0.01"E) (Figure 1).During each harvest we collected root exudates and rhizosphere samples for molecular identification of soil microbial community and weighed root and shoot biomass (details below).

Root exudate collection
To capture root exudate compounds, alive roots were carefully separated from the soil by continuous and gentle rinsing with deionized water until roots were separated from mineral particles (protocol adapted after (Oburger et al., 2014)).This method might cause potential root damage and thus leaking of cell contents, although the effects on exudate composition is not yet clear (Williams et al., 2021).The plants´ roots were submerged for 10 min into 100 ml of deionized water in glass flask wrapped with aluminium foil to avoid any light effects.Again, roots were gently rinsed with deionized water.After the washing procedure, roots were then placed in the final sampling solution (100 ml deionized water, containing 0.01 g l −1 Micropur classic (Katadyn®, Switzerland) and kept under greenhouse conditions for the entire sampling period (4 h).Thereafter, the sampling solution was filtered through 7 µm (Whatman™ folded filters, Ø 150 mm, 595 ½, Sigma-Aldrich) to remove remaining soil particles and further filtered through a sterile 0.2µm syringe filter (Whatman™ Puradisc™ 30 syringe filters, Sigma-Aldrich) with a cellulose acetate membrane.The samples were stored at -20°C and lyophilised at -

Solid Phase extraction of root exudates and sample processing
The freeze-dried root exudates were dissolved in 2ml of 5% methanol (LC-MS grade) in ultrapure water and sonicated for 10 minutes at ambient temperature in an ultrasonic bath, followed by a centrifugation step at 6000g for 10 minutes (after (Strehmel et al., 2014).The supernatant was transferred in a fresh 2ml tube.For every sample, a SPE cartridge packed with C18 column material (Chromabond 200mg/3ml, Machery-Nagel) was conditioned with 1ml of pure methanol followed by 1ml of 2% formic acid in water.The dissolved root exudates were transferred from the 2ml tube to the conditioned column.The column was washed with 1ml ultra-pure water followed by one elution step with 2% formic acid in pure methanol.The eluates were evaporated to dryness using a Speed Vac at 40°C and resolved in 150µl 70% methanol followed by sonification and centrifugation (10min, 6000g).Finally, the supernatant was transferred in a LC glass vial.

Accepted Article
Metabolites were analysed on a liquid chromatography quadrupole time-offlight mass spectrometer (LC-qToF-MS; Bruker impact HD; Bruker Daltonik, Bremen, Germany) with an electrospray ionization source operated in negative mode.Instrument settings were as follows: capillary voltage, 2500 V; nebulizer, 2.5 bar; dry gas temperature, 220°C; dry gas flow, 11 L min-1; scan range, 50 -1500 m/z; acquisition rate, 3 Hz.We used sodium formate clusters (10 mM solution of NaOH in 50 / 50% [v/v] isopropanol / water containing 0.2% formic acid) to perform mass calibration.For further annotation mass spectra (MS²) of selected pooled samples were collected in positive and negative MSMS mode.

LC-MS data processing and metabolite prediction
We followed the LC-MS data processing protocol described in (Ristok et al., 2019) with minor changes.We converted the LC-qToF-MS raw data to the mzXML format by using the CompassXport utility of the DataAnalysis vendor software.Subsequently, we trimmed each data file by excluding the same non-informative regions at the beginning and end of each run using the msconvert function of ProteoWizard v3.0.10095 (Chambers et al., 2012).We performed peak picking, feature alignment, and feature group collapse in R v3.3.3 (RStudio Team, 2020) using the Bioconductor packages 'xcms' (Smith et al., 2006;Tautenhahn et al., 2008;Benton et al., 2010) and 'CAMERA' (Kuhl et al., 2012).We used the following 'xcms' parameters: peak picking method "centWave" (snthr = 10; ppm = 5; peakwidth = 4, 10); peak grouping method "density" (minfrac = 0.5; bw = 6, 3; mzwid = 0.01); retention time correction method "symmetric".We used 'CAMERA' to annotate adducts, fragments, and isotope peaks with the following parameters: extended rule set (https://gitlab.com/R_packages/chemhelper/-/tree/master/inst/extdata); perfwhm = 0.6; calcIso = TRUE; calcCaS = TRUE, graphMethod = lpc.Lastly, we collapsed each annotated feature group, hereafter referred to as 'metabolite' which is described by mass-to-charge ratio (m/z) and retention time (rt), using a maximum heuristic approach.In detail, this means that the intensity values of the feature that most often displayed the highest intensity across all samples represent the feature group.We performed pre-processing with 'xcms' and 'CAMERA' for each species and sampling season.We merged all created feature lists by retention time and mass-to-charge values.
For each feature, we allowed for a retention time window of 10 seconds and a mass deviation of 5 ppm.

Soil sampling, DNA extraction and sequencing
Samples for molecular analysis were collected from the rhizosphere soil prior to root exudate collection.Therefore, plant roots were gently shaken and the soil adhering to roots was carefully brushed from the roots, homogenized by mixing, collected to an Eppendorf tube and immediately frozen in liquid nitrogen and stored at −80 °C prior to DNA-extraction.DNA was extracted from 0.75 g of soil using the PowerSoil DNA Isolation Kit (Mo Bio Laboratories, Carlsbad, CA, USA) following the manufacturer's protocol and the quantity of DNA was measured using a Nanodrop spectrophotometer (Thermo Scientific, Hudson, NH, USA).PCRs were performed using approximately 100 ng of DNA primers ITS4ngs and ITS3mix targeting the ITS2 region of fungal genes (Tedersoo et al., 2015) and the primers 515F and 806R (Caporaso et al., 2012;Apprill et al., 2015;Parada et al., 2016)  Bacterial sequences were analysed using the Hydra pipeline (Hollander, 2017) and fungi using the PIPITS pipeline (Gweon et al., 2015).In both pipelines, sequences were paired using VSEARCH and quality was filtered using standard parameters of the pipelines (i.e.min overlap 20bp, primers need to be exact matches, quality score over 28 used).For fungi, the ITS2 region was extracted using ITSx (Nilsson et al., 2015).Short reads (<100bp) were removed, and sequences were clustered based on a 97% similarity threshold using VSEARCH.For fungi, chimeric sequences were removed by comparing with UNITE uchime database.The representative fungal sequences were identified using the RDP classifier against the UNITE database (Nilsson et al., 2019) and representative bacterial sequences using SINA classification tool with SILVA database.
Only OTUs belonging to bacteria or fungi were kept in the analysis (protists, plants, archaea, mitochondria and chloroplast were removed).For both bacterial and fungal OTU tables, samples with less than 1000 reads or more than 80 000 reads were removed and OTUs present in less than three

Statistical analysis
We used linear models to test the effects of plant growth (time), soil conditioning (soil) and its interaction on shoot and root biomass.Effect of PSF on plant biomass were calculated as: ln(plant dry mass (g) in "home-soil") at time x − ln(plant dry mass (g) in "away-soil") at time x, where ln is natural logarithm.Thus, negative values indicate that plants grow better in "awaysoil".The feedback effect was calculated separately for each replicate (i.e.ln("home-soil" replicate 1) − ln("away-soil" replicate 1) etc).We further used comparisons of means for treatment-specific effects (Tukey's HSD test; p < 0.05).Tukey's tests were performed using the multcomp package (Hothorn et al., 2008).
To test the effects of duration of plant growth and soil condition on bacterial and fungal community compositions and on root exudate composition, we ran permutational multivariate analysis of variance (PERMANOVA, based on Bray-Curtis dissimilarities, 999 permutations) using the adonis2 function in the vegan package (Oksanen et al., 2020).For visualization, we applied a nonmetric multidimensional (NMDS) analysis (using the metaMDS function in the vegan package (Oksanen et al., 2020)) of the dissimilarities (based on Bray-Curtis dissimilarities) in root exudate and microbial community composition using the ggplot2 package (Wickham, 2021)

Plant biomass and plant-soil-feedback effects
Shoot and root biomass increased significantly over the experimental period for both plant species (Table 1, Figure 2a -d).In week 3 and 4, shoot biomass of J. vulgaris was higher (week 3: +43 %, week 4: + 24%) in "away-soil" than in "home-soil" (Figure 2a), whereas from week 5 onwards, J. vulgaris shoot and root biomass values were slightly higher in "home-soil" (Figure 2a, c).Thus, strong negative "home-soil" feedback in the early growth phase (< -1) could be observed, but the strength changed over time (F = 3.99; p = 0.010).
The negative feedback effect rapidly diminished, and became slightly, but not significantly, positive (Figure 2e).Shoot biomass of H. lanatus was higher in "away-soil" over the entire 10 weeks of the experiment (Figure 2b).Root biomass was mostly higher in "away-soil" except in week 6 (ranging between + 2 and + 45% increase) (Figure 2d), resulting in a negative "home-soil" feedback effect that did not significantly change over time (F = 0.52; p = 0.807; Figure 2f).However, for H. lanatus the strength of the negative feedback diminished over time.

Root exudate composition and diversity
NMDS analysis revealed strong compositional changes of root exudates composition over time for both J. vulgaris and H. lanatus (Table 2 and 3; Figure 3).Differences in soil conditioning of both plant species did not affect the root exudate composition (Table 2 and 3; Figure 3).However, the interaction of time and soil conditioning was significant for J. vulgaris (Table 2 and 3; Figure 3a).Further, the diversity of root exudates increased significantly over time for both plant species (Table S1

Soil microbial composition and diversity
Rhizosphere bacterial community composition changed significantly with plant growth for both plant species and the two differently conditioned soils (Table 2; Figure 4).Further, soil bacterial communities in "home"and "away-soil" of both J. vulgaris and H. lanatus converged over time (Figure 4).This was particularly evident for composition of bacteria in H. lanatus where bacterial composition in the "home-soil" (Holcus growing in Holcus soil) remained relatively constant, while the composition in the "away-soil" (Holcus growing in Jacobaea soil) started differently but moved into direction of the home soil over time (Figure 4).The diversity (Simpson´s diversity index) of bacterial communities of J. vulgaris was not affected by plant growth or soil conditioning.However, the bacterial diversity within the rhizosphere of H. lanatus was significantly higher at week 9 when grown in its "home-soil", and significantly lower in week 6 only when grown in its "away-soil" (Table S1, S2; Figure S1 b, e).
The fungal community in the rhizosphere of J. vulgaris was strongly affected by plant growth and to a lesser extent by soil conditioning (Table 2; Figure 5 a).However, plant growth and soil conditioning led to strong compositional changes in fungal communities in the rhizosphere of H. lanatus (Table 3; Figure 5 b).The diversity (Simpson´s diversity index) of fungal communities in the rhizosphere of J. vulgaris was not significantly affected by plant growth or soil conditioning, however, it was significantly lower in "away-soil" for H. lanatus compared to "home-soil" (Table S1, S2; Figure S1 c, f).

Accepted Article
Linkages between plant-soil feedback, root exudate and soil microbial diversity Our path models suggest that temporal variation in plant-soil feedback in both plant species over the growth period depends on temporal shifts in root exudate diversity (Figure 6), although the pathways differed between J. vulgaris (through home soil) and H. lanatus (through "away-soil").More specifically, we found stronger positive effects of root exudate diversity on temporal increase in the slightly positive plant-soil feedback of older J. vulgaris plants in "home-soil" (Figure 6 a).By contrast, root exudate diversity in "awaysoil" sustained negative PSF on H. lanatus during its growth (Figure 6 b), however, the effect size of negative PSF tended to decrease over time (Figure 2 f).Moreover, path models revealed that rhizosphere fungal and bacterial diversity were not influenced by root exudate diversity in either "home-" or "away-soils" for both plant species (Figure 6).However, we found that a moderate increase in fungal diversity over time in "home-soils" for J. vulgaris constrained its temporal shift towards neutral (slightly positive) plant-soil feedback (Figure 6 a).Direct effects of rhizosphere microbiomes on plant-soil feedback in H. lanatus were further non-significant irrespective of "home-" or "away-soils" pathway (Figure 6 b), whereas bacterial diversity in "home-soils" of J. vulgaris directly influenced (negatively) its plant-soil feedback independent of its growth period (Figure 6 a).

Discussion
In this study, we examined the temporal variation in plant-soil feedback effects (PSF), root exudates and soil microbial communities within the rhizosphere of two common grassland species.Confirming our hypothesis, we found strong

Article
negative conspecific feedback effects for both plant species in their early plant growth stages.Notably, the negative conspecific PSF effect of J. vulgaris shifted in week 5 to a neutral (slightly) positive PSF effect (Figure 2e).The strength of the conspecific PSF effects for H. lanatus also declined over time but the PSF of this species remained negative.These findings are consistent with previous studies and provide strong evidence that the strength of PSFs strongly depend on the plant growth stage of a plant (Kardol et al., 2013;Bezemer et al., 2018;Hannula et al., 2021).Emphasizing the importance of considering such temporal shifts in PSF strength when designing and performing PSF experiments.
Beside temporal changes in PSF effects, the root exudation profile of both plant species strongly depended on the plant growth stage.Especially, J. vulgaris showed very distinct composition of root exudates in the first few weeks of growth, whereas they became temporally more similar when the plants grew larger (from week 5 onwards).The root exudation profile of H. lanatus changed more gradually.Previous studies with common grass species, reported that plant biomass is positively correlated with carbon rhizodeposition and thus that root exudation increases when plants grow larger (Baptist et al., 2015).The exudation profile of Arabidopsis thaliana also has been shown to strongly vary among different plant stages (Chaparro et al., 2013).The root exudate diversity of both plant species in our study increased considerably with the developmental stage of the plant individual, thereby emphasizing that the process of root exudation is highly dynamic (Sasse et al., 2018).Interestingly, both the composition and diversity of root exudates did not differ between "home-" and "away-soil".This indicates that the soil legacy of the previous plant (i.e., the changed microbiome) appears to Plant root exudates have been shown to shape soil microbial communities in the rhizosphere (Steinauer et al., 2016;Oburger & Jones, 2018;Sasse et al., 2018).Soil microbes respond to these plant-derived metabolites which in turn determines their success to establish within the rhizosphere.Due to the changes in root exudation profiles over plant growth, we expected that soil microbial communities would exhibit strong temporal dynamics.In our study, both soil bacterial and fungal community composition varied strongly on weekly basis.This is in line with previous studies, reporting temporal changes of soil microbial communities at the scale of days (Zhang et al., 2011) to months (Lauber et al., 2013;Hannula et al., 2019).Our results can thus confirm that a seedling, a juvenile or an older plant shape their soil microbial community differently at least for the two grassland species used in our study.
Furthermore, the soil bacterial community composition of both plant species appeared to be different between "home-" and "away-soil".For J. vulgaris, it seemed that the soil bacterial community was more dissimilar between "home-" and "awaysoil" in the early stages of plant growth (week 3 -7) whereas they were more similar in a later plant developmental stage (week 8 -10).
The soil bacterial community of H. lanatus was highly distinct between "home-" and "awaysoil" at the start of the experiment, and converged only at week 10.Moreover, for H. lanatus, the temporal pattern shows that in home soil, the bacterial community remains relatively constant over time (i.e. the bacterial community associated to H. lanatus), while in the away soil, the host plant H lanatus, steers the bacterial community associated to J. vulgaris (the previous plant) towards the H. lanatus community.Why bacterial communities in both

Accepted
Article soils in which J. vulgaris was growing changed much more over time, remains further examination.We speculate that this is driven by changes in chemical composition in root exudates, but we cannot conclude this from our study.
However, what our study shows is that host plants can affect the structure and development of soil bacterial communities over a period of several weeks.
Similarly, the composition of the soil fungal community of both plant species differed among plant developmental stages, whereas "home -" and "awaysoil" of H. lanatus, but not J. vulgaris, led to distinct soil fungal communities over the course of the entire experiment.These results are in line with previous studies, reporting that temporal changes in plant growth are leading mainly to changes in soil bacterial communities of J. vulgaris whereas soil fungal communities in grasses like H. lanatus showed to be affected stronger (Hannula et al., 2021).
Temporal changes in root exudate diversity seemed to strongly affect the variation in PSF in both plants.More specifically, the temporal variation in PSF in both plant species was mainly driven by the exudate diversity in J. vulgarisconditioned-soil.For J. vulgaris, increasing root exudate diversity in "homesoil" coincided positively with changes in PSF effects, whereas for H. lanatus this was true for its "away-soil".Previous studies have indeed shown that J. vulgaris can shape a distinct rhizosphere environment potentially through specific exudation dynamics (Kowalchuk et al., 2006).Our results suggest that it could relate to the diversity of root exudates and their effects on the rhizosphere microbiome as revealed by our path models.Future experiments are required to establish causal relationships between root exudate diversity and the role of specific compounds in the exudates on rhizosphere microbiomes.This is particularly important for plant species such as J.

Accepted
Article vulgaris, where we know now that both root exudate diversity and plant-soil feedback changes over its growth period.
In conclusion, this study demonstrates that the direction and magnitude of plant-soil feedbacks depends on plant growth stages.Especially, J. vulgaris showed strong directional changes in its PSFsfrom negative to neutral (slightly positive) -in the early life stage, thereby highlighting the importance to consider temporal variability in PSF studies.We examined the importance of two potential key factors driving temporal shifts in PSF and found the root exudation profiles of both plant species to greatly depend on the plant growth stage.Furthermore, both soil bacterial and fungal community composition varied strongly on a weekly basis and were different between "home-" and "away-soils" (conspecific and heterospecific), which may also have contributed to temporal variation in PSF.Moreover, through linking these results in pathmodels, we could link shifts in PSF effects to temporal dynamics of root exudate diversity whereas changes of soil bacterial and fungal diversity effects on root exudate diversity and PSFs need to be investigated in more detail in future experiments.Further, both shifts in root exudates and soil microbial communities could cause nonlinear shifts in PSF effects which would be recommended to test for in future studies.The importance of PSFs is increasingly recognized among ecologists and more detailed studies on root exudate metabolites (e.g., identity, concentrations, and individual functions) and their specific effects on temporal PSF effects together with how they associate with soil microbiome are urgently needed to understand and predict the magnitude and direction of PSFs.(Holcus lanatusconditioned soil), turquoise= "home-soil" (Jacobaea vulgaris conditioned soil), (B) plant growth on root exudate composition of Holcus lanatus within both soils.Big panels of (A) and (B) display temporal shifts of root exudate composition, whereas in small panel of (A) circles = "away-soil" (Holcus lanatus -conditioned soil), triangles = "home-soil" (Jacobaea vulgaris conditioned soil), in small panel of (B) displays the same temporal shifts  (Jacobaea vulgaris -conditioned soil), brown = "home-soil" (Holcus lanatusconditioned soil).Big panels of (A) and (B) display temporal shifts of soil bacterial community composition, whereas small panels display the same temporal shifts including individual samples (small dots), and large dots represent averaged centroids.In small panel of (A) circles = "away-soil" (Holcus lanatus -conditioned soil), triangles = "home-soil" (Jacobaea vulgaris conditioned soil), and in small panel of (B) circles = "away-soil" (Jacobaea vulgaris -conditioned soil), triangles = "home-soil" (Holcus lanatusconditioned soil).Stress values are given for each NMDS.Asterisks represent significance levels (n.s.= not significant; *P < 0.05; **P ≤ 0.01; ***P ≤ 0.001).
Each bacterial community composition had a total sample size of 48.In small panel of (A) circles = "away-soil" (Holcus lanatusconditioned soil), triangles = "home-soil" (Jacobaea vulgarisconditioned soil), and in small panel of (B) circles = "away-soil" (Jacobaea vulgarisconditioned soil), 13653040, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/pce.14570by Universität Bern, Wiley Online Library on [20/02/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 13653040, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/pce.14570by Universität Bern, Wiley Online Library on [20/02/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Accepted Article communities in the soil, and how these changes affect plant-microbial interactions in the soil and ultimately PSF effects.
targeting the V4 region of the 16Sr RNA gene in bacteria were used.The PCR products 13653040, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/pce.14570by Universität Bern, Wiley Online Library on [20/02/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License This article is protected by copyright.All rights reserved.Accepted Article were purified using Agencourt AMPure XP magnetic beads (Beckman Coulter) and adapters and barcodes were added to enable multiplexing with Nextera XT DNA library preparation kit set A (Illumina, San Diego, CA, USA).The final PCR product was purified again with AMPure beads and quantified using a Nanodrop spectrophotometer before equimolar pooling.Pooled libraries were sequenced using Miseq PE250 technology at McGill University and Genome Quebec Innovation Center.Extraction negatives were used and further sequenced.A mock community, containing 10 fungal species, was included to investigate the accuracy of the bioinformatics analysis.
13653040, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/pce.14570by Universität Bern, Wiley Online Library on [20/02/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseAcceptedArticlesamples with relative abundance of less than 0.05% were removed.These cut-off values were derived from inspection of mock communities consisting of 10 fungal species Cumulative sum scaling (CSS) was used to normalize the data.
13653040, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/pce.14570by Universität Bern, Wiley Online Library on [20/02/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Accepted Article have far less effect on the root exudation profile of the subsequent plant individual than the plant growth stage of the individual.

Table 1 :
Linear model: table of F-and p-values on the effects of time and soil type on shoot and root biomass of Jacobaea vulgaris and Holcus lanatus

Table 2 :
PERMANOVA model: table of R², F-values and p-values for the effect of time and soil type on root exudate, baterial and fungal community compositon of 13653040, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/pce.14570byUniversität Bern, Wiley Online Library on [20/02/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)onWiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseThis article is protected by copyright.All rights reserved.Time" was used as categorical factor.Significant results (P < 0.05) are highlighted in bold.

Table 3 :
PERMANOVA model: table of R², F-values and p-values for the effect of time and soil type on root exudate, baterial and fungal community compositon of