Soil microbial community coalescence and fertilization interact to drive the functioning of the legume–rhizobium symbiosis

Merging entire microbial communities has important consequences for diversity and community function (Castledine et al., 2020; Sierocinski 2017), but been largely overlooked in studies on plant-associated microbes (Rillig 2015; Rillig, Lehmann, 2016). This phenomenon of coalescence can provide plants that depend mutualistic with novel partners complementary properties functions (Mueller & Sachs, 2015). Moreover, agricultural contexts, merging usually occurs under elevated soil nutrients due to fertilization mixing associated ploughing. While the touted as promising plant production (Kapagianni 2019; Rillig 2016), we still lack basic knowledge how their respond fertilizer application. The symbiosis between most legumes N2-fixing rhizobia represents a particularly interesting study system explore assembly functioning are affected by coalescence. In this symbiosis, supply additional reduced nitrogen return photosynthates (Denison Kiers, 2011), after mass multiplication root nodules, released into where they survive saprobes 2011). means rhizobial ensembles distinct populations derived from local legume species which may have evolved different degrees specificity (Sprent 2017). Community thus differs other systems such methanogenic bacterial (Sierocinski coalescing not cohesive units, subsets meta-community sustained surrounding hosts. deliberate soils therefore lead interactions Mansour, many might promote enhanced symbiotic (Simonsen For legume–rhizobium reassembly following will be highly dependent environment. Abiotic conditions like N, P water availability play pivotal roles driving composition symbiont assemblages via: (a) environmental filtering (Vuong (b) modulation resource allocation nodules (Wendlandt 2019) (c) influence size through improved nutrition (Robinson, 1994). instance, larger roots often correlate higher (e.g. Dinnage 2019). Thus, increased stemming interactive effects nodules. high nutrient decrease selectivity (Weese 2015), potentially allowing colonization wider range rhizobia. Early efficient also positive feedbacks enhance growth Boyle 2021). Increasing chances rhizobia, via diversity, suggests applied interest agriculture ecological restoration. There is need understand levels interact determine functioning, only because naturally occurring constantly reorganizes (Rillig, its potential application tool functionality plant–microbial mutualisms (Gravuer Scow, 2021; Wubs 2018). To our knowledge, outcomes living context variable nodule performance studied. cross-factorial pot experiment, aimed wide parameters structure communities. We tested conditioned cultivated wild crop rooibos Aspalathus linearis. Rooibos an ideal test it grows fluctuating levels, associates symbionts (primarily Mesorhizobium; Hassen 2012; Le Roux 2017; Ramoneda 2020) grown nurseries manipulated mixed) prior transplanting plantations (see Figure S1). Given proximity (Le Ramoneda, Roux, 2020), events likely occur field conditions. Understanding interplay fertility management species. Here, specifically wanted whether: rhizosphere collected synergistic seedling mineral growth, addition impacts dynamics process, (rhizobial coalescence) predominantly promotes compositional or relative abundance changes whether (d) benefits more terms biological fixation (BNF) upon fertilization. sensitivity status (Lötter 2014), (Hassen hypothesized would owing establishment new symbionts. Soils were five farms adjacent rooibos, located Sandstone Fynbos vegetation Suid Bokkeveld (Northern Cape, South Africa). 5.5 33 km apart 2020 details) managed small-scale farmers producing organic tea. around each >3-year-old shrubs beneath thought similar age older. Cultivated than 100 m within farm Methods S1 full details). Acid-scarified surface-sterilized seeds Nortier variety over 8 months 2 L pots ventilated polytunnel. experiment followed randomized nested block design main factor ‘farm origin soil’, factors ‘habitat soil’ ‘fertilization’. had corresponded (and by) ecologically habitats, namely ‘cultivated’ (hereafter ‘uncultivated’) 1:1 (v:v) mixture these (‘mixed’; see ‘fertilization’ comprised 7.5 g sieved sheep dung potted soil. calculated amount increase total N content 30% 50% (Table Each 30 treatments [farm (5) × habitat (3) (2)] replicated 10 times, totalling 300 experimental units. Treatment replicates allocated spatial blocks arrangement respect treatment was randomized. See S2 complete setup maintenance. stem, branches, leaves seedlings separated air-dried dry matter fractions. stem branch lengths measured, number branches counted. Prior drying, thoroughly rinsed sieves pressurized water. first paper before transferred cotton pads placed top silica gel tubes desiccation storage weighing grinding. Ca, Cu, K, Mg concentrations, concentrations δ13C δ15N signatures grinded subsamples leaves, determined hot HNO3-extraction Microwave Liquid digestion System (MLS Turbowave, MWS GmbH, Heerbrugg, Switzerland) measured inductively coupled plasma optical emission spectrometry (ICP-OES) ICPE-9800 (Shimadzu), ion ratio spectrometer (ICP-IRMS) Delta V Advantage (ThermoFisher Scientific) respectively. Total genomic DNA extracted 25 mg powder all seedling. dried pulverized 7 mm diameter glass bead ml Eppendorf TissueLyser II (Qiagen). NucleoSpin® Plant Mini Kit (Macherey-Nagel) according manufacturer's protocol. Root nodule-associated characterized using sequencing data two regions: 716 bp- 743 bp-long fragment (depending genus) phylotaxonomic marker gyrB, amplified primers gyrB343F gyrB1043 (Martens 2008), 455 nodA gene, NodAunivF145u NodARbrad (Moulin 2001). gene region triplicate reactions template. gyrB useful analyses including information taxonomic breadth gyrB343F/1043 primer pair characterization diversity). amplicons (i.e. unit) tagged barcoded universal (BUP) libraries 96 multiplexed samples analysed single-molecule real-time (SMRT) circular consensus (ccs) Sequel I sequencer, standard protocol (PacBio®, slight modifications second short PCR Nextera XT indices (Illumina), bp paired-end sequenced Illumina MiSeq sequencer (Methods S2). operational units (OTUs) at 99% nucleotide sequence identity both datasets S2), generated OTU tables rarefaction median read across (gyrB: 1,410 reads, nodA: 8,720 reads). Four <200 reads dataset 12 <300 excluded further analyses. OTUs assigned genera BLASTn searches (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Details quality curation described al. (2020), summarized S3. verified absence 16S rRNA (results shown). Statistical carried out generalized mixed model: Response ~ Fertilization + (Habitat soil)/Farm Fertilization*((Habitat soil)/Farm), random = 1|Block R package nlme (v3.1-149; Pinheiro 2020). Significance results models running ANOVAs model fits. summarize effect magnitude change nutritional α sizes Cohen's d coefficient (Hojat Xu, 2004), implemented cohen.d effsize (v0.8.0; Torchiano, farms, allowed calculating errors check deviation zero. Since did reference acquired none corrected leaf Gubsch (2011) comparisons farms. fertilized seedlings, accounted signature added way, could use lower surrogate N2 fixation, biologically fixed dung. iNEXT (v2.0.20) used correct unequal effort among intra- extrapolation sequences per sample (Hsieh calculate measures: richness Simpson's diversity. abundances correction effort, phyloseq (v1.16.2, McMurdie Holmes, 2013). evenness Pielou's (J'), dominance 1 − J'. order visualize structure, ran principal coordinate (PCoA) Bray–Curtis, Sorensen, weighed unweighed UniFrac distances permutational multivariate analysis variance (PERMANOVA) 9,999 permutations Bray–Curtis dissimilarities markers, vegan (v2.5-5, Oksanen Differences assembled ‘cultivated’, ‘uncultivated’ ‘mixed’ level pairwise PERMANOVA (function pairwise.adonis pairwiseAdonis; Martinez Arbizu, quantify interaction (gyrB) (nodA) taxa, differential pairs ‘mixed versus uncultivated level, ancom (v2.1, Mandal ANCOM approach chosen alternative approaches accommodate shown low false discovery rate (Weiss Those showed statistically significant increases origins counted, mean numbers responsive compared Welch two-sample t test. topmost abundant recorded, frequency occurrence particular OTU) mixtures those individual component soils, chi-square (χ2) tests. An indicator taxon taxa preferentially (high BNF), multipatt indicspecies (v1.7.6; De Caceres Legendre, 2009). this, unfertilized divided quartiles, quartile lowest identified. relationship responses (mineral growth) assembly, distance-based redundancy (dbRDA) run ordistep (v2.5-5) dissimilarities, forward/backward selection optimize model. list modelled reported Table S3, besides soils’ ‘fertilization’, whose coded dummy variables. Overall, positive, evident average (Figure 1). Fertilizer strongest performance, influencing nearly growth- nutrition-related response variables 1; S4). Fertilized accumulated (mean ± SD; 7.44 4.36 g) (4.23 2.54 g), contents. exerted multiple ecophysiological (P concentration, δ13C), BNF (δ15N isotopic signatures; S4), concentration content, acted synergistically mixing, enhancing leaf, matter, BNF, (except P, underwent dilution; S3). Cu comparison 1a), whereas stimulated 1b). 161,155 931,254 quality-filtered clustered 355 Mesorhizobium, Bradyrhizobium Rhizobium 93 Mesorhizobium OTUs. bacteria (gyrB marker) significantly SD: 31 OTUs) mixtures, (28 (25 OTUs). Mixing always highest richness, nor altered Soil S5). led intermediate regardless 2). Despite variation, sizeable (note R2) determinant (PERMANOVA, soil’: R2 0.188, p < 0.001; nodA, 0.187, 0.001), locations ‘geographic soil’; S5 summary PERMANOVA). dissimilarity measures giving weight (Sorensen distances) revealed large (9.8%–21.2% explained PCoA axes 2; S6). By contrast, based (Bray–Curtis UniFrac) twice much (28.7% 55.9% Upon skewed distribution disappear predicted. fact, became even while equally 3a). less 3b; S7). When rank-abundance same (χ2 6.97, 0.656), 22.81, 0.004). broke single (rhizobial) (bacterial) structural predicted present alone. Some 155) benefitted despite being less-abundant 3b). 99) soils. caused assembling becomes when comparing differentially separately 4a; S8), representing case minor biotic (mixed communities) abiotic (nutrient availability) identified 15 found signatures, fixers. These mostly belonged subdominant any community) Bradyrhizobium, 4b; addition, communities, 4c), indicating synergy changes. Together, findings support view enable non-dominant colonize symbiosis. size, nutrition, nodulation signatures) correlated 5; positively negatively signatures. strong 5a), 5b). Rhizobial bigger fraction BNF) observed number, very weak. appeared structured biomass, generally pronounced patterns. Our consistent indication shifts coalesced show newly established BNF. Following major alterations propagule pools, together host plants, breakdown (or few) rhizobium strain(s) rooibos. diverse richer evenness), received atmospheric N2. Bacterial improve reconfiguration Previous non-additive (Brandt 2013; Burns Hendriks Ma 2018), reflecting complex (a)biotic factors. A consequence dilution (Hendriks Hol reduction inoculum below known reduce activity beneficial well antagonistic fungi affecting (Mendes 2011; St-Denis limitation unlikely role study, long history illustrated dominant shared Instead, plant. aware previously unavailable (Booth 2006; require testing. rates mixing). More interestingly, added, resulting evenness, reducing few Why fertilizer? Growth demands 1994; Wang S3), promoted reliance turn increasing contrasts common observation (Keller Lau, However, general observations involve typical nutrient-poor (Stock Lewis, 1986). overall suggest taxonomically source pools promoting sampling associations (Loreau Hector, 2001; Thrall 2008). likelihood functionally opportunities encounter (Dinnage providing niche space assemble. propose establish promotion head start), pool select once become deficient. cause build-up mutualist leading effectiveness. patterns components. dominate rather configuration ones productive community; agreement observations, Rocca (2020) recently freshwater marine low-abundance becoming prominent members. skewed, emphasizing significance infection pressure (Singleton Tavares, that, genotype legumes. consistently 155; 3b), others took advantage created 12; S8). limiting demand exudation heterotrophic microbes, expected (Guyonnet (often rare) explain why configurations soils) widespread events, importance (Rhizobium Bradyrhizobium), growing again non-dominant, gained shows core should targeted prospecting bioinoculants. phosphorus P- K-rich manure, subsequent deficiency disadvantaged nutrient-efficient, slow-growing favoured fast growing, explaining fractions (Zahran, 1999). one As provider translate amplify (in nutrients) feedbacks. Such mechanistic target restoration Similar efforts inocula healthy readily obtained nurseries. caveat small physiological parameters. Field trials nurseries, transplantation arable fields, address relevant implications production. combined agro-ecological engineering (Bender raising coalescence, thereby amplifying biotrophic growth. thank Dr Peter Manning anonymous reviewers valuable feedback provided previous drafts manuscript. J.R. acknowledges assistance generating analysing gryB Genetic Diversity Center (GDC) Functional Genomics (FGC), processing material Sandra Wenger. J.L.R. funding Macquarie University's Faculty Science Engineering Department Biological Sciences. H.A.G. grateful project former employment ETH current FUB. All authors acknowledge Noel Oettlé, Cecilia Bester, Bokkeveld, organized Heiveld Ltd. farmer cooperative. Ruan Kemp help translating abstract Afrikaans. Open Access Funding Eidgenossische Technische Hochschule Zurich. declare no conflict interest. conceived adapted molecular genetic methodology; J.R., S.S. data; B.F. E.F. gave critical inputs manuscript; drafted manuscript incorporated co-authors. contributed final approval publication. Statement inclusion: Traditional stakeholders design. research discussed workshop seek considered deliver farming. Raw available National Biotechnology Information (NCBI) http://www.ncbi.nlm.nih.gov/bioproject/PRJNA630687 (ETH Zurich, Please note: publisher responsible supporting supplied authors. Any queries (other missing content) directed corresponding author article.