Resuscitation of the microbial seed bank alters plant-soil interactions

While microorganisms are recognized for driving belowground processes that influence the productivity and fitness of plant populations, the vast majority of bacteria and fungi in soil belong to a seed bank consisting of dormant individuals. Still, plant performance may be affected by microbial dormancy through its effects on the activity, abundance, and diversity of soil microorganisms. To test how microbial seed banks influence plant-soil interactions, we purified recombinant resuscitation promoting factor (Rpf), a bacterial protein that terminates dormancy. Then, in a factorially designed experiment, we applied the Rpf to soil containing field mustard (Brassica rapa), an agronomically important plant species. Plant biomass was ~33 % lower in the Rpf treatment compared to plants grown with an unmanipulated microbial seed bank. In addition, Rpf reduced soil respiration, decreased bacterial abundance, and increased fungal abundance. These effects of Rpf on plant performance were accompanied by shifts in bacterial community composition, which may have diluted mutualists or resuscitated pathogens. Our findings suggest that changes in microbial seed banks may influence the magnitude and direction of plant-soil feedbacks in ways that affect above- and below-ground biodiversity and function.

findings suggest that changes in microbial seed banks may influence the magnitude and direction 22 of plant-soil feedbacks in ways that affect above-and below-ground biodiversity and function. Yet, the influence of microbial seed banks on plant-soil interactions remains to be determined. 48 Microbial seed banks may prevent local extinctions when exposed to fluctuating and stressful conditions that are typical in soil habitats (Shoemaker & Lennon, 2018). If microbial  Resuscitation is an essential process that regulates microbial seed-bank dynamics (Lennon,  that cross-link peptidoglycan, which may serve as signaling molecules (Dworkin, 2014). Under laboratory conditions, Rpf can stimulate the activity of some microbial taxa at picomolar 74 concentrations, but it has also been shown to inhibit the growth of other microorganisms (Mukamolova et al., 1998;Mukamolova et al., 2002), which is not surprising given that Rpf is In this study, we explored how seed-bank dynamics affect plant-microbe interactions by 80 manipulating the process of resuscitation. Throughout the lifespan of a host plant (Brassica rapa), we applied purified recombinant Rpf to soils containing a complex microbial community.

82
The experiment was guided by a conceptual model that considers how active and inactive microorganisms interact and potentially create feedbacks with plants ( Fig. 1  Quercus rubra) and sugar maple (Acer saccharum) within the Norman Upland bedrock 108 physiographic unit (Schneider, 1966). Chemical and physical characteristics of the soil are listed in Table S1. intensity (1100 µmol m -2 s -1 ) with controlled humidity (60 %) and temperature (28 °C). Pots were arranged in a fully randomized design where each plant was assigned to a resuscitation 118 treatment (i.e., -Rpf vs. +Rpf). All pots were watered with equal amounts of filtered and deionized water every other day.  To overexpress Rpf, we grew the E. coli expression host in Lysogeny Broth (LB) with appropriate antibiotics (ampicillin 100 µg mL -1 , kanamycin 15 µg mL -1 , and tetracycline 12.5 µg 146 mL -1 ). During logarithmic growth, recombinant protein production was induced with Isopropyl β-D-1-thiogalactopyranoside (IPTG) (100 μM final concentration). We confirmed 148 overexpression of Rpf with Western blots (Fig. S1). Cells were then collected by centrifugation, lysed by sonication, and filtered through a Ni-NTA Purification System (Invitrogen) using a 10 150 mL gravity fed column with a 2 mL resin bed to purify recombinant Rpf with the N-terminus polyhistidine-tag. Recombinant Rpf protein was washed with 5 mM imidazole buffer (300 mM 152 NaCl, 50 mM Tris-HCl, 5 mM imidazole) and then eluted with 125 mM imidazole buffer. Rpf protein was purified by buffer exchange using a 10 mL Zeba Spin Desalting Columns (Thermo 154 Fisher) with protein buffer (20 mM Tris-HCl, 100 mM NaCl) following manufacturer's instructions and then passed through a 0.2 µm syringe filter before adding to soil substrate as  In addition to our main study, we conducted a smaller scale experiment with Arabidopsis 172 thaliana, a relative of B. rapa that also belongs to the Brassicaceae. Because this species is amenable to being grown axenically in the absence of soil, we were able to test for the direct  (Gardes & Bruns, 1993) and the 5.8S reverse primer (CGCTGCGTTCTTCATCG) (Vilgalys & 208 Hester, 1990) to amplify the ITS gene region. qPCR assays were performed using Eppendorf Mastercycler Realplex system using the previously reported thermal cycle conditions: 15 min at 210 95 °C, followed by 40 cycles of 95 °C for 1 min, 30 s at 53 °C for annealing, followed by 72 °C for 1 min (Fierer et al., 2005). The coefficients of determination (r 2 ) of our assay ranged from 212 0.95 and 1, while amplification efficiencies fell between 0.93 and 0.99. Based on melting curve analyses, we found no evidence for primer dimers. We estimated fungal and bacterial abundance 214 based on the estimated gene copy number from their respective standard curves generated from bacterial and fungal isolates as described elsewhere (Lau & Lennon, 2012). We performed a two-216 tailed Student's t-test to determine the effect of the Rpf treatment on soil bacterial (16S rRNA) and fungal (ITS) gene copy abundances, as well as the fungal to bacterial ratio (F : B).

218
Microbial diversity -To account for the variation in metabolic activity, we characterized 220 bacterial communities from the soil substrate using pools of RNA and DNA. DNA is a relatively stable molecule contained in intact cells irrespective of their metabolic status. Accordingly, we 222 interpret 16S rRNA sequences recovered from the DNA pool as the "total" community. In contrast, RNA is a more ephemeral molecule that is required for protein synthesis by actively 224 growing cells. Therefore, we interpret 16S rRNA sequences recovered from the RNA pool after complementary DNA (cDNA) synthesis as the "active" community (Jones & Lennon, 2010). We 226 assume that dormant individuals can create discrepancies between the active and total composition of a given sample, but we do not attempt to use RNA and DNA to directly Paired-end raw 16S rRNA sequence reads were assembled into contigs using the Needleman 246 algorithm (Needleman & Wunsch, 1970). We then trimmed the resulting sequences using a moving average quality score (window = 50 bp, score = 35), in addition to removing long 248 sequences, ambiguous base calls, and sequences that matched Archaea, chloroplasts, and other non-bacteria. We also removed chimeric sequences that were detected using the UCHIME   (Fig. 3). While respiration increased throughout the experiment in both treatments, based on marginal means, soil respiration was 24 % lower in the +Rpf treatment than in the -Rpf 302 treatment.

330
Since the production of Rpf is thought to be restricted to the Actinobacteria, we also evaluated how treatments affected the diversity of taxa within this phylum. Rpf had no effect on  (Table S3). There was a small and marginally significant reduction in the relative abundance of actinobacterial sequences in the +Rpf treatment compared to the -Rpf 336 treatment (t9 = 1.46, P = 0.083) (Fig. S4). Based on PERMANOVA, actinobacterial composition was significantly affected by both metabolic status (i.e., active vs total) (F1, 19 = 16.08, r 2 = 0.33, 338 P = 0.001) and the Rpf treatment (F1, 19 = 3.25, r 2 = 0.07, p = 0.010) (Fig. S5). Indicator species analysis revealed that OTUs belonging to the Acidothermus, Catenulispora, and Mycobacterium 340 were associated with the -Rpf treatment, while only a single taxon belonging to the Solirubrobacterales had a significant association with the +Rpf treatment (Fig. S6)

344
Dormancy is an important life-history trait that influences the diversity, composition, and function of soil microbial communities. After generating recombinant protein with a gene from 346 an environmental isolate, we applied resuscitation promoting factor (Rpf) to a soil community to test the hypothesis that microbial seed-bank dynamics alter plant growth and fitness traits by 348 driving changes in the belowground microbial community (Fig. 1). The Rpf treatment decreased plant biomass (Fig. 2) most likely by altering the activity (Fig. 3), abundance (Fig. 4), and 350 composition (Fig. 5) of soil microbial communities. These findings are consistent with the view that soil microbial seed banks can influence plant performance, perhaps by disrupting 352 interactions with beneficial microorganisms or through the recruitment of pathogens. In the following sections, we discuss these findings while exploring other potential ways in which Rpf 354 may affect plant-microbe interactions in soil environments.

356
Rpf indirectly affects plant performance -Resuscitation of microbial seed banks led to a 33 % reduction in plant biomass (Fig. 2). For the following reasons, we infer that Rpf effects on plant

Rpf altered fungal-bacterial interactions -Rpf may have altered plant performance by
370 modifying fungal-bacterial interactions. We observed a nearly three-fold increase in fungal abundance in response to Rpf additions (Fig. 4). As a result, soil F : B ratio increased under Rpf 372 treatment (Fig. 4), which may also explain the reduction in soil microbial activity (Fig. 3) given that there are often differences in carbon-use efficiency between bacteria and fungi (Sakamoto &    While our study provides empirical support for the notion that microbial seed banks can affect plant performance, it also raises a number of questions for future exploration. We observed  (Fig. 1). Nevertheless, our study highlights a potentially important but overlooked component of plant-microbe interactions. It is estimated that soil microbial 448 communities can be dominated by dormant taxa, yet changes in activity can be fast, suggesting that seed banks may be an important factor contributing to plant-soil feedbacks, which is thought 450 to be an important mechanism maintaining landscape patterns of biodiversity (Bever, 2003).  individuals that were exposed to weekly additions of recombinant Rpf (+Rpf) to those exposed to 652 a protein buffer control (-Rpf). Black symbols represent the mean ± 95 % confidence intervals.
Grey symbols represent the individual observations. 654 Fig. 3. Effect of resuscitation promoting factor (Rpf) on soil microbial activity. Soil respiration 656 was measured after applying Rpf (+Rpf) or protein buffer control (-Rpf) to soils on a weekly basis. Symbols represent the mean ± 1 SE x ̅ .  and total (i.e., DNA) Actinobacteria from soil that were exposed to +Rpf and -Rpf treatments at 668 the end of a six-week experiment. The ellipses were generated by 'ordiellipse' function using the standard deviation of PCoA point scores to visualize the spread of each treatment.