The role of helper in facilitating mycorrhization of Biserrula pelecinus L., a pasture legume new to Australia

David James Orchard

Graduate Diploma in Science (), University of New England, Armidale, NSW, Australia

Bachelor of Arts (Hons), Australian National University, Canberra, ACT, Australia

Submitted to Charles Sturt University for the degree of Doctor of Philosophy

School of Agricultural and Wine Sciences

Faculty of Science

March, 2020

Contents

List of Tables ...... iv List of Figures ...... vi Certification of Authorship ...... ix Acknowledgements ...... x Abstract ...... xi Chapter 1. helper bacteria in the of Biserrula pelecinus L...... 1 1.1. Introduction ...... 1 1.2. -microbe interactions in the rhizosphere...... 2 1.3. Microbial diversity in the ...... 7 1.4. Mycorrhiza helper bacteria ...... 10 1.5. The receptivity hypothesis ...... 21 1.6. Biserrula as a target species ...... 28 1.7. Conclusion ...... 31 1.8. Aims ...... 32 1.9. Thesis structure ...... 32 Chapter 2. Observation and quantification of arbuscular in the of the pasture legume Biserrula pelecinus L...... 34 2.1. Introduction ...... 34 2.2. Format of experiments ...... 36 2.3. Experiment 1: Quantification of mycorrhizas in the roots of field- and trap culture-grown biserrula ...... 36 2.4. Experiment 2: Identification of mycorrhizas in the roots of field- and trap culture-grown biserrula ...... 41 2.5. Experiment 3: Recruitment of AM fungi to the root environment of biserrula in a pasture environment ...... 44 2.6. Results ...... 47 2.7. Discussion ...... 54 2.8. Conclusion ...... 57 Chapter 3. Identification of mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L...... 59 3.1. Introduction ...... 59 3.2. Materials and methods ...... 61

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. ii 3.3. Results ...... 70 3.4. Discussion ...... 88 3.5. Conclusion ...... 92 Chapter 4. Ecological interactions in the rhizosphere of Biserrula pelecinus L. and their influence on MHB activity ...... 93 4.1. Introduction ...... 93 4.2. Materials and methods ...... 96 4.3. Results ...... 103 4.4. Discussion ...... 108 4.5. Conclusion ...... 113 Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. across a two-season legume-cereal rotation ...... 114 5.1. Introduction ...... 114 5.2. Materials and methods ...... 116 5.3. Results ...... 123 5.4. Discussion ...... 140 5.5. Conclusion ...... 144 Chapter 6. Summary and conclusions ...... 146 6.1. Introduction ...... 146 6.2. Aims of the research ...... 146 6.3. Key findings ...... 147 6.4. Synthesis ...... 149 6.5. Future research ...... 154 6.6. Conclusion ...... 155 References ...... 158 Appendix 1. Mycorrhiza helper bacteria in ...... 193 Appendix 2. Structure of experiments addressing ecological interactions in the rhizosphere (Chapter 4)……………………………………………………………..202

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. iii List of Tables

Chapter 1.

Table 1-1. Evidence of a helper effect on mycorrhization, growth of fungal , or plant growth parameters in arbuscular mycorrhizas ...... 14

Chapter 2.

Table 2-1. Primers and sequencing conditions for nested PCR ...... 44 Table 2-2. Primers and sequencing conditions for ITS diversity profiling ...... 46 Table 2-3. Frequency and intensity of mycorrhization in field- and pot-grown biserrula ...... 48 Table 2-4. BLAST matches for DNA ...... 50 Table 2-5. Relative abundance of common AM fungi in rhizosphere, bulk, and field ...... 52 Table 2-6. Diversity index values for ITS diversity profiles by compartment ...... 55

Chapter 3.

Table 3-1. Diversity profiling primers and sequencing conditions ...... 62 Table 3-2. 16S primers and sequencing conditions ...... 65 Table 3-3. Genus-specific primers used in this study...... 65 Table 3-4. Bacterial cultures used in each trial ...... 67 Table 3-5. Diversity index values for 16s diversity profiles by soil compartment ...... 72 Table 3-6. Differentially abundant bacterial taxa in the biserrula rhizosphere and the surrounding field soil ...... 76 Table 3-7. BLAST matches for putative MHB strains ...... 83 Table 3-8. Functional traits of putative MHB strains ...... 86

Chapter 4.

Table 4-1. Combinations of MHB strains used in the in vitro assay ...... 98 Table 4-2. Treatments used in the in vivo populated rhizosphere screening ...... 100 Table 4-3. Treatments used in the in vivo rhizobial inoculation trial ...... 101 Table 4-4. Nodulation factor primers used in this study ...... 102 Table 4-5. Effect of bacterial and mesorhizobial inoculation on nodule count...... 106

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. iv

Chapter 5.

Table 5-1. Samples collected during the rotation trial ...... 121 Table 5-2. Bacterial sequence and species richness by treatment ...... 129 Table 5-3. Diversity index values by treatment...... 130

Appendix 1.

Table A1-1. Evidence of a helper effect on mycorrhization, growth of fungal mycelium, or plant growth parameters in ectomycorrhizas ...... 195

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. v List of Figures

Chapter 1.

Figure 1-1. A simplified diagrammatic representation of the rhizosphere environment ………………………………………………………………………………………….4

Chapter 2.

Figure 2-1. Aerial view of the study site at Beckom, NSW, Australia. The approximate course of Kildary Creek, which marks the western boundary of the paddock, is denoted by the dotted blue line. The yellow rectangle denotes the sampling area. Areas outside of this rectangle were severely affected by waterlogging during the first collection period...... 38 Figure 2-2. Study site at the time of sampling. Mature, flowering of Biserrula pelecinus (a) were present in the pasture; however, large areas of the paddock had been affected by waterlogging (b). Samples were not taken from severely affected areas of the paddock...... 39 Figure 2-3. Photomicrographs of isolated from a biserrula trap culture, including intact (a,b,c) and manually broken (d,e) spores. Two distinct morphotypes were observed (a). Morphotype A (b, d) was irregular in shape and was in most cases the larger of the two (c. 160-240 µm diameter). Morphotype B (c, e) was smaller (c. 100- 170 µm diameter) and more globose. Spores of morphotype A were isolated from sporocarps and typically did not possess subtending hyphae. Scale bars represent 200 µm (a) or 100 µm (b-e)...... 43 Figure 2-4. Photomicrographs of crushed biserrula root sections displaying mycorrhizal structures. A = arbuscule (Arum type). V = vesicle. H = longitudinal ...... 49 Figure 2-5. Proportional composition of the ITS diversity profiling dataset by , as assessed by (a) total ITS sequence count and (b) total count of operational taxonomic units...... 53

Chapter 3.

Figure 3-1. Proportional composition of the 16s diversity profiling dataset by phylum, as assessed by (a) total 16s sequence count and (b) total count of operational taxonomic units (OTUs)...... 73 Figure 3-2. Operational taxonomic units (OTUs) identified as differentially abundant in the biserrula rhizosphere and the surrounding field soil...... 74

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. vi Figure 3-3. Operational taxonomic units (OTUs) identified as differentially abundant in the rhizosphere and the adjacent bulk soil (a,b) and in the bulk soil and the surrounding field soil (c,d)...... 75 Figure 3-4. The effect of bacterial treatment on in vitro germination of mosseae spores...... 77 Figure 3-5. The effect of bacterial treatment on in vitro hyphal elongation of Funneliformis mosseae spores...... 78 Figure 3-6. Photomicrographs of Funneliformis mosseae spores before and after germination...... 79 Figure 3-7. Photomicrographs of Funneliformis mosseae secondary spores (a, b) and incipient spore development in vitro (c)...... 80 Figure 3-8. Photomicrographs of Funneliformis mosseae colonies incubated for 20 weeks...... 81 Figure 3-9. The effect of bacterial treatment on in vivo colonisation of Biserrula pelecinus by Funneliformis mosseae...... 82 Figure 3-10. Osmotic stress tolerance of putative mycorrhiza helper bacteria exposed to increasing levels of NaCl...... 87 Figure 3-11. Colonies of Bacillus aquimaris/vietnamensis grown on unamended agar (a), nutrient agar amended with 8% NaCl (b), and nutrient agar amended with 10% NaCl (c) ...... 87

Chapter 4.

Figure 4-1. The effect of mycorrhiza helper bacteria on the in vitro hyphal growth of Funneliformis mosseae when applied singly or in combination, showing (a) log- transformed means ± SE, and (b) log-transformed means and ranges with significance groups...... 104 Figure 4-2. The effect of helper bacteria on mycorrhizal colonisation in biserrula when inoculated singly or with a rhizosphere solution, showing (a) mean percent colonisation value ± SE, and (b) means, ranges, and significance groups...... 107 Figure 4-3. The effect of helper bacteria and Mesorhizobium ciceri bv. biserrulae on percent mycorrhizal colonisation in biserrula when inoculated singly or in co- inoculation treatments, showing (a) mean percent colonisation value ± SE, and (b) means, ranges, and significance groups...... 108

Chapter 5.

Figure 5-1. Plants of biserrula (left) and (right) prior to harvesting...... 119 Figure 5-2. The arrangement of the glasshouse rotation trial. Coloured markers indicate first-season treatments: biserrula (orange), wheat (blue), and no (white). Replicates

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. vii are in 3 × 3 blocks. A fifth replicate was harvested but was not sequenced. The watering system shown here was not used during the study...... 120 Figure 5-3. Proportional abundance of bacterial phyla across all 64 samples, measured by operational taxonomic unit (OTU) count (a) and total sequence count (b) per phylum...... 125 Figure 5-4. Individual plot (a) and variable plot (b) of correlation matrix PCA of log- transformed data from all 63 samples. The two principal components making the greatest contribution to observed variance have been retained. The variation explained by each PC is indicated on the axes...... 126 Figure 5-5. Contribution of each variable to PC1 (a) and PC2 (b). The dashed red line indicates the average expected contribution of each variable assuming all phyla contributed evenly (2.94%). For each PC, all but one variable below the cut-off has been omitted...... 127 Figure 5-6. Diversity profiles of initial (Y0), first-season cultured (Y1NILCU, Y1BISCU, Y1WHCU), and first-season non-cultured (Y1NILNC, Y1BISNC, Y1WHNC) samples for values of q=0-5, where q=0 is equal to S, q=1 is equal to H, and q=2 is equal to Simpson diversity...... 133 Figure 5-7. Operational taxonomic units (OTUs) identified as differentially abundant between cultured and non-cultured sample pairs taken from biserrula (a, b), wheat (c, d) and the no host treatment (e, f)...... 134 Figure 5-8. Operational taxonomic units (OTUs) identified as differentially abundant between Y0 samples and first-season wheat...... 135 Figure 5-9. Operational taxonomic units (OTUs) identified as differentially abundant between Y1 nil and Y2NN (a), Y2NB (b), and Y2NW (c); between Y1 biserrula and Y2BN (d), Y2BB (e), and Y2BW (f); and between Y1 wheat and Y2WN (g), Y2WB (h), and Y2WW (i)...... 136 Figure 5-10. Operational taxonomic units (OTUs) identified as differentially abundant between Y2NN and Y2WW (a, b), and between Y2NW and Y2BB (c, d)...... 139

Appendix 2.

Figure A2-1. A diagrammatic outline of experiments conducted in Chapter 4………201

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. viii Certification of Authorship

I hereby declare that this submission is my own work and to the best of my knowledge and belief, understand that it contains no material previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at Charles Sturt University or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by colleagues with whom I have worked at Charles Sturt University or elsewhere during my candidature is fully acknowledged. I agree that this thesis be accessible for the purpose of study and research in accordance with normal conditions established by the Executive Director, Library Services, Charles Sturt University or nominee, for the care, loan and reproduction of thesis, subject to confidentiality provisions as approved by the University.

Name: David James Orchard Date: 27/03/2020

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. ix Acknowledgements

I would first like to acknowledge the contributions of my supervisors, Dr Ben Stodart (CSU), Dr Sandra Savocchia (CSU), Dr Gavin Ash (USQ), and Dr Bree Wilson (USQ), who helped to steer this project through all its various detours, and offered guidance, suggestions, valuable feedback, and practical support throughout.

I would also like to thank Mike and Velia O’Hare of “Greendale,” Beckom, for the (repeated) use of their property for the collection of soil and plant material. I hope this wasn’t too much of a burden.

Thanks also to all in the NaLSH and the QAP over the last four years for discussion and occasional practical assistance.

To my parents, the greatest possible gratitude for your constant support through frequently trying times. To my mum, thanks for the statistical discussion and guidance, without which this project would certainly have been overwhelming. To both, thanks for accompanying me on various collection trips and for the assistance in setting up the glasshouse trial.

To Chris, thanks for the valuable distraction of time spent wandering in the wilderness, or what passes for wilderness in this part of the world, and for the constant stream of plants (fungi, insects, etc.) to identify, which gave me something other than statistics to think about.

This project was supported by an Australian Postgraduate Award and by a Graham Centre Research Scholarship Top-Up.

Finally, I would like to dedicate this thesis to the memory of my brother, Stephen, who died as it was nearing completion. I’m sorry we never had a chance to properly discuss your work or mine.

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. x Abstract

Arbuscular mycorrhizas (AM) promote the growth of their host plants by enhancing nutrient uptake and conferring resistance to biotic and abiotic stresses. Consequently, stimulation of AM colonisation represents a promising means by which plant productivity in may be increased. A substantial body of evidence now supports the contention that certain rhizosphere bacteria can act to promote the formation and persistence of mycorrhizas. These mycorrhiza helper bacteria (MHBs) have been observed both to promote the pre-symbiotic development of mycorrhizal fungi in vitro and to increase the rate of root colonisation in vivo. To date, MHBs have been isolated from a relatively restricted range of AM associations and scope exists to investigate their role in other plant systems. The principal aim of this research project was to determine whether associated with Biserrula pelecinus L. (biserrula), a pasture legume selected for use in low-rainfall and high-stress agriculture, are capable of promoting mycorrhizal colonisation in that host. Subsidiary aims included (1) establishing whether this helper effect persists in non-sterile conditions; (2) identifying potential mechanisms involved in the helper effect; and (3) identifying constraints on the implementation of these helpers in terms of their survival in the rhizosphere. Microscopic analysis of biserrula plants grown in pastures and in synthetic trap cultures revealed low levels of mycorrhization (~20%) relative to functionally equivalent pasture species. Genetic diversity profiling targeting the fungal internal transcribed spacer (ITS) region suggested that Funneliformis mosseae may be the principal, though not exclusive, fungal partner in field conditions. Subsequent isolation and screening of plant-associated bacteria identified six taxa capable of stimulating in vivo mycorrhization of biserrula by F. mosseae, belonging to the genera , Bacillus, Microbacterium, and . Two strains of Pseudomonas were observed to facilitate low levels of sporulation by F. mosseae in the absence of a plant host, challenging the conventional description of AM fungi as obligate symbionts.

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. xi No single functional trait could be identified to account for the observed helper effects; however, the involvement of the common pathway (CSP) was suggested by the observation that a Pseudomonas koreensis subgroup taxon fostered nodule formation in biserrula in axenic conditions. Experimental co-inoculation of MHBs with a commercial rhizobium did not identify any synergistic interactions but did identify Mesorhizobium ciceri bv. biserrulae WSM1497 as an MHB strain. Inoculation of MHBs into non-sterile soil demonstrated that the helper effect could be diminished, in some cases, by microbial competition. A potential synergy was observed between the P. koreensis subgroup taxon and a Bacillus cereus group taxon in promoting pre- symbiotic growth of F. mosseae. While this project was successful in identifying bacterial taxa capable of acting as MHB strains in both sterile and non-sterile soil, obstacles to the practical implementation of these microorganisms remain. Diversity profiling revealed that the rhizosphere of biserrula is highly variable; this, combined with the observed reduction in the magnitude of certain helper effects in non-sterile conditions, and the suppression of taxa closely related to identified MHB strains, suggests that successful implementation of MHBs in agriculture may necessitate management of the whole rhizosphere. This, in turn, is likely to entail careful implementation of agricultural rotations.

The role of helper bacteria in facilitating mycorrhization of the pasture legume Biserrula pelecinus L. xii Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L.

1.1. Introduction A mycorrhiza is a persistent association between a and the root of a plant, typically involving a bidirectional exchange of : the fungus obtains photosynthetic carbon necessary for its growth and reproduction and in return enhances the uptake of mineral nutrients, most notably , by the plant (Chen et al. 2018). In addition, the fungus may confer resistance to biotic and abiotic stresses, including , drought, and certain soil pollutants (Lenoir et al. 2016). These beneficial effects, particularly in conditions hostile to plant growth, make mycorrhizal associations promising avenues by which the productivity of agricultural plants could potentially be increased (Smith and Smith 2011a). The tissues of the plant and fungus, as well as the surrounding soil of the rhizosphere, are often intensively colonised by bacteria (Turrini et al. 2018, Bakker et al. 2013), and this microbial background is known to interact with the mycorrhizal symbiosis in ways that affect its establishment and persistence (Hoeksema 2010). Among the microorganisms known to influence the success of mycorrhizal colonisation are the mycorrhizal helper bacteria (MHBs), bacteria found to promote the formation of mycorrhizas in their plant hosts (Frey-Klett et al. 2007, Garbaye 1994). Implementation of these helpers in agricultural settings may boost the productivity of crop and pasture plants and the benefits of this practice are likely to be felt more acutely in plants that are exposed to hostile soil conditions. The pasture legume Biserrula pelecinus L. (biserrula) meets this description: selected for use in acid soils and low-rainfall conditions (Loi et al. 2014), biserrula is subject to the many stresses associated with Australian dryland agriculture, including nutrient limitation, salinity, and sodicity (Carberry et al. 2011, Rovira 1992). By cultivating the mycorrhizal symbiosis in this way, it may be possible to alleviate these stresses and maximise the productivity of biserrula pastures.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 1 1.2. Plant-microbe interactions in the rhizosphere

1.2.1. The rhizosphere effect The roots of terrestrial plants, including crop and pasture species, traverse a soil environment intensively colonised by microorganisms and in so doing provide a unique habitat for soil microbes (Figure 1-1.). A plant’s rhizosphere, defined as the volume of soil immediately surrounding and directly affected by the activity of the plant root (Raaijmakers et al. 2009, Barea et al. 2005), is home to an assemblage of microbial life that is often both quantitatively and qualitatively distinct from the surrounding bulk soil. Bacteria, for example, are characteristically more abundant in this rhizosphere zone but are thought to be less diverse in the bulk soil (Dennis et al. 2010). Bacteria, fungi, and other microorganisms are recruited to the root environment by a combination of abiotic factors (including soil pH, nutrient concentration, and moisture availability) and biotic processes such as rhizodeposition (Philippot et al. 2013). Rhizodeposition, including the release of breakdown products from dead root cells and from live ones, enriches the rhizosphere with sugars, amino acids, organic acids, and a range of other compounds, including plant hormones and other molecules involved in signalling (Dennis et al. 2010). The constitution of rhizodeposits varies between plant taxa and consequent variations in root microbiota have been observed at the level of plant family, species and genotype (Bakker et al. 2013, Kristin and Miranda 2013). Motile bacteria are drawn from the bulk soil into the -enriched rhizosphere; consequently, the root may select for hypermotility (Barahona et al. 2010). In a sense, roots can be said to curate a subset of the total bulk soil microbiota adapted to the conditions in the rhizosphere, a phenomenon known as the rhizosphere effect (Bakker et al. 2013). The organisms that respond positively to this curation make up a plant’s “rhizosphere zoo” (Buée et al. 2009) and participate in a complex web of plant-microbe and microbe- microbe interactions. Our understanding of the microbial community composition of the rhizosphere has historically been limited by the efficacy of in vitro cultivation techniques. As the pool of species recoverable in cultures may represent less than one percent of total community diversity (Burke et al. 2008, Leveau 2007), these are likely to give a highly selective and artificial view of the rhizosphere. Genetic screening of rhizosphere communities has emerged as a means of counteracting this limitation and the application of molecular methods, including metabarcoding, metagenomics, and transcriptomics, is now

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 2 beginning to offer a picture of the total diversity of rhizosphere microbial communities and their webs of interactions (Orgiazzi 2015, Bakker et al. 2013). While many of the interspecific interactions in this fraction of the soil microbiota are at present only poorly understood, if they are known at all, and many may ultimately exert negligible effects on their plant hosts, a small number have been found to have measurable impacts, both positive and negative, on plant growth and survival (Raaijmakers et al. 2009). These microorganisms, which may influence plant growth, longevity, and reproductive success, can thus be said to form part of the root’s extended phenotype (Garcia and Kao-Kniffin 2018). In agricultural settings, these impacts may rise to economic significance: in an Australian context, plant diseases are responsible for a total estimated economic loss of more than AUD 1.5 billion per year in wheat, oilseeds, and barley using current control methods (Murray and Brennan 2012, Murray and Brennan 2009a, Murray and Brennan 2009b). While these plant pathogens almost certainly share their rhizosphere habitat with beneficial soil microorganisms, the commercial impacts of beneficial species are poorly known. Beneficial rhizosphere microorganisms include -fixing , plant growth-promoting rhizobacteria (PGPR) and mycorrhizal fungi (Raaijmakers et al. 2009). All three of these classes of beneficial microbe contribute to (Chen et al. 2018, Berruti et al. 2016, Lugtenberg and Kamilova 2009, Long 1989). In addition, mycorrhizal fungi and certain PGPR species are known to enhance plant disease resistance (Rufo et al. 2016, Zhou et al. 2014, Medeiros et al. 2011) and confer tolerance to abiotic stresses including drought (Lim and Kim 2013, Bresson et al. 2013, Wu et al. 2006a) and the accumulation of toxic minerals (Khan et al. 2016, Guo and Chi 2014, Fomina et al. 2006). Rhizobia have been employed to enhance fixation of nitrogen by agricultural legumes for more than a century, and there has been a move to commercialise many PGPR species in the creation of new bio-fertilisers (Trivedi et al. 2017, Parnell et al. 2016, Bashan et al. 2014). The economic value of beneficial species – particularly those beneficial species already present in agricultural soils – has not been addressed, but the contributions they make to plant nutrition and overall survivorship suggest that they play a significant role in crop and pasture health. It has long been assumed that the composition of the rhizosphere and the balance of beneficial and deleterious microbes therein has an important part to play in determining plant growth outcomes (Schippers et al. 1987). Mycorrhizas are perhaps the most ubiquitous of these

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 3 beneficial microbes and make the greatest contribution to plant health, making their thorough study an economic priority.

Figure 1-1. A simplified diagrammatic representation of the rhizosphere environment. As the root of a legume traverses the soil, it forms a rhizosphere (brown, dashed line), an environment enriched with chemical exudates that stimulate microbial activity. This narrow zone is the site of numerous interspecific interactions involving the host plant, bacteria such as rhizobia and mycorrhiza helper bacteria, and fungi, including mycorrhizal species.

1.2.2. The mycorrhizal symbiosis A mycorrhiza is a symbiotic association between a fungus and the root of a plant. Seven classes of mycorrhiza are recognised, differing according to the of the fungal partner, their effect on plant morphology, and the identity of the host plant (Smith and Read 2008). Of these, arbuscular mycorrhizas (AM) are the most widespread and are of the greatest economic consequence, being found in an estimated 72% of all terrestrial plant species, including most pasture and crop plants (Brundrett and Tedersoo 2018). This review will primarily consider AM associations; where additional evidence from ectomycorrhizal (ECM) systems is available, this is given in Appendix 1. The fungi that form AM associations are asexual obligate symbionts of the phylum

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 4 (Schüßler et al. 2001), represented worldwide by approximately 313 known species in 11 families and 26 genera (Schüßler and Walker 2019, Redecker et al. 2013), though this figure likely underestimates their actual diversity and an upper estimate of 1600 species worldwide has been cited (Van der Heijden et al. 2015). Several species of the Glomeromycota are known to be cosmopolitan and to form associations with a wide range of plant taxa (Parniske 2008). Mycorrhizal development begins with a bidirectional exchange of diffusible signalling molecules (Gutjahr and Parniske 2013). exuded by plant roots have been found to induce branching and growth of hyphae in several AM fungi (Besserer et al. 2006) and are considered to be the principal plant contribution to this signal exchange, though similar results have been obtained for the AM fungus Gigaspora gigantea using two hydroxy fatty acids from carrot roots (Nagahashi and Douds 2011). Among the active compounds produced by AM fungi in this pre- symbiotic phase are short-chain chitin oligomers (COs), which have been found to induce Ca2+ spiking in roots and activation of symbiosis- associated genes (Genre et al. 2013), and lipochitooligosaccharides (LCOs), which are reported to enhance mycorrhization in several plant families and have been shown to stimulate root growth and branching in M. truncatula (Maillet et al. 2011) via the same symbiosis-associated genetic pathways. These LCOs (often referred to as Myc-LCOs) strongly resemble nodulation (Nod) factors produced by rhizobia prior to symbiosis (Bonfante and Requena 2011). Nod factors have similarly been shown to promote mycorrhization in M. truncatula (Xie et al. 1995). Myc-LCOs were long considered the principal fungal contribution to signal exchange; however, the involvement of both LCOs and short-chain COs in forming symbioses suggests that this exchange of molecules may depend on more than a single Myc factor. Root contact with these molecules activates the signal transduction pathway controlled by the so-called DMI (“does not make infections”) genes, components of the common symbiotic (SYM) pathway, a set of genes associated with both mycorrhizal and rhizobial symbiosis.. DMI genes appear to be involved in regulating calcium spiking in early symbiotic contact (Parniske 2008). The economic significance of these symbioses derives from their positive effects on host plant growth and survivorship. While plant responses to mycorrhizal colonisation may occasionally be negative (Hoeksema et al. 2010), AM associations are typically framed as true mutualisms (Bonfante and Anca 2009). The fungus benefits by obtaining

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 5 photosynthetic carbon and the host plant benefits chiefly by improved nutrition (Chen et al. 2018, Corrêa et al. 2015, Smith and Smith 2011b). Secondary benefits to the plant include increased resistance to pathogens (Rufo et al. 2016, Liu et al. 2007, Filion et al. 2003, Borowicz 2001) and tolerance to environmental stresses including drought (Lenoir et al. 2016, Wu et al. 2006a, Sylvia et al. 1993, Davies et al. 1992) and toxic soil minerals (Lenoir et al. 2016, Fomina et al. 2006, Leyval et al. 1997, Dueck et al. 1986). A mycorrhizal plant may assign as much as 20% of its photosynthate to its fungal symbiont and receive as much as 90% of its phosphorus in return (Van der Heijden et al. 2015). This exchange may be especially critical in Australian pasture systems, where phosphorus is often limiting (Schefe et al. 2015). Additionally, AM fungi are reported to supply nitrogen, , copper, potassium, and other nutrients (Berruti et al. 2016, Hodge and Storer 2014, Smith and Smith 2011b, Leigh et al. 2008, Cavagnaro 2008), though these aspects of the plant-fungus association have received less attention. Transfer of nutrients between partners takes place chiefly across a symbiotic interface formed in root cortical cells by fungal structures called arbuscules, which serve as the key diagnostic feature of active AM associations. AM fungi have been reported to produce more than 100 metres of hyphae per cm3 of soil (Miller et al. 1995), and it is in large part this intensive colonisation that gives them their greater scavenging capacity relative to plant roots. The beneficial effects of mycorrhizal colonisation have typically been studied in isolation from the background microbiota. This is in spite of a developing awareness that the success of AM colonisation is contingent to some degree on the surrounding . A meta-analysis of studies examining the effects of mycorrhizal inoculation on plant responses found that, in experimental settings, AM colonisation was more positive when the background microbiota was more complex (Hoeksema et al. 2010). In natural populations, bacteria occupy every compartment of the plant- fungus symbiosis, including the interior and exterior of fungal spores (Agnolucci et al. 2015, Bianciotto et al. 2003) and the mycelium (Toljander et al. 2006, Andrade et al. 1997), as well as the surrounding soil and the plant roots. Certain bacterial taxa and assemblages have been found to live in strict association with mycorrhizal structures (Turrini et al. 2018). These organisms are embedded in the mycorrhiza and interact with it constantly, and within this pool of mycorrhizosphere bacteria are a small number known to promote or maintain the mycorrhizal symbiosis. These are the mycorrhiza or mycorrhization helper bacteria (MHB) (Frey-Klett et al. 2007, Garbaye 1994). The

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 6 ubiquity of these bacteria-mycorrhiza interactions and the presumed importance of MHB species in natural systems has led to suggestions that mycorrhizas no longer be considered bipartite mutualisms between plants and fungi but tripartite mutualisms involving plants, fungi and the surrounding microbiota (Bonfante and Anca 2009).

1.3. Microbial diversity in the mycorrhizosphere

1.3.1. The mycorrhizosphere effect The transformation of a rhizosphere into a mycorrhizosphere adds another dimension to an already complex system (Rambelli 1973). Rhizosphere microbial composition is determined in part by the exudation patterns of the root and any alteration in the of the root (by, for example, the addition of a mycorrhizal fungus) may result in qualitative or quantitative changes in these patterns. These changes are expected to alter the existing microbial equilibrium in the rhizosphere (Meyer and Linderman 1986a), culminating in the mycorrhizosphere effect. AM fungi may contribute directly to the mycorrhizosphere effect by secretion of the glycoprotein (Cameron et al. 2013). Further impacts may come through direct fungus- bacteria interactions, which occupy all points of the spectrum between and antagonism: many rhizosphere fungi produce antibiotic compounds (Brakhage and Schroekh 2011), for example, while others provide habitat or nutrition for rhizosphere bacteria (Scheublin et al. 2010). Attempts to quantify AM-associated changes to host-plant microbial communities have a long but uneven history. Early experiments based on selective cultures (Meyer and Linderman 1986a) and phospholipid fatty acid analysis (PLFA) (Rillig et al. 2006) offered a partial picture of the mycorrhizosphere microbiota but were limited by the available technologies. Meyer and Linderman (1986a) reported that the presence of AM fungi increased the total bacterial community size in roots of both Zea mays and Trifolium subterraneum and the abundance of fluorescent pseudomonads (FLPs) in mycorrhizal roots of Z. mays but were limited to those classes of bacteria recoverable in cultures. A similar restriction was encountered in the use of PLFA techniques, which identified an association between phylogenetic groups of AM fungi and distinct soil microbial communities but could resolve only broad taxonomic groups (e.g. Gram- positive bacteria or actinomycetes), due to the limitations of PLFA analysis in separating bacteria at the level of genus or species (Rillig et al. 2006). The use of

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 7 genetic approaches, involving the culture-independent screening of bacterial communities based on their DNA, developed as a result of these limitations (Leveau 2007). While their usefulness in determining whole-community composition is limited, culture-dependent techniques remain effective at screening selected bacterial strains for functional traits, including the production of degradative enzymes and phytohormones (Fernandez Bidondo et al. 2016). Culture-based methods have been used to assess the selectivity of bacterial attachment to hyphae in a Daucus carota (carrot) – spp. association (Toljander et al. 2006), finding that two root-colonising bacteria, Pseudomonas fluorescens SBW25 and Paenibacillus brasilensis PB177, preferentially adhered to the vital hyphae of Glomus sp. MUCL 43205 and the putative hypha- degrading saprotroph Bacillus cereus VA1 adhered preferentially to non-vital hyphae of irregularis (formerly Glomus intraradices) MUCL 43194. These studies offer insights into the selectivity of the mycorrhizosphere and to bacterial contributions to plant health; nevertheless, it should be remembered that these effects, and the species that produce them, are embedded in a complex and diverse microbial community that remains poorly understood (Shelef et al. 2019). While molecular methods have been used extensively to detect and characterise AM fungi themselves (Bianciotto et al. 2011), equivalent work using current genetic methodologies has not been carried out for their bacterial associations. Most attempts at characterising AM-associated bacterial communities have employed culture-dependent methods (Taktek et al. 2015, Scheublin et al. 2010, Welc et al. 2010, Lugo et al. 2008, Wang et al. 2007, Ravnskov et al. 1999, Andrade et al. 1997) or have limited their focus to spore-associated communities (Agnolucci et al. 2015, Long et al. 2008, Xavier and Germida 2003). Other studies have provided only very limited taxonomic resolution of the bacterial communities under study (Zhang et al. 2010, Lioussane et al. 2010, Singh et al. 2008, Marschner and Baumann 2003, Marschner et al. 2001). Bacterial genera occurring in association with AM fungi in multiple studies include Pseudomonas and Bacillus (Taktek et al. 2015, Agnolucci et al. 2015, Lioussane et al. 2010, Lugo et al. 2008, Andrade et al. 1997), Burkholderia/Ralstonia (Taktek et al. 2015, Xavier and Germida 2003), and Arthrobacter (Lugo et al. 2008, Andrade et al. 1997), along with members of the order Rhizobiales (Agnolucci et al. 2015; Taktek et al. 2015). These associations may not be universal, however: Ravnskov et al. (1999) and Long et al. (2008) did not find a positive association between AM fungi and Pseudomonas.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 8 Genetic screening of AM-associated bacterial communities has been conducted on a smaller scale for a Medicago truncatula (barrel medic) association using multiple AM strains and three genotypes of M. truncatula (J5: Myc+/Nod+; TRV48: Myc+/Nod-; TRV25: Myc-/Nod-) to compare mycorrhizal and non-mycorrhizal roots (Pivato et al. 2009, Offre et al. 2008, Offre et al. 2007). The low diversity of bacteria reported (all members of the order Burkholderiales in the Betaproteobacteria) suggests that the true community composition was not recovered using the methods employed in these studies. A later study using the same three M. truncatula genotypes reported a greater abundance of FLPs possessing Type III secretion systems in mycorrhizal root environments (Viollet et al. 2011). These secretion systems are implicated in pathogenicity and also in plant-growth promotion and MHB effects.

1.3.2. Genetic approaches to soil microbial community analysis The screening of bacterial communities using molecular methods is intended to circumvent the limitations and biases of culture-based approaches (Leveau et al. 2007). Metagenomics, defined as the study of genomes derived directly from environmental samples (Thomas et al. 2012), has emerged as a powerful tool for the analysis of microbial communities, allowing both taxonomic characterisation and functional analysis. This functional analysis often centres on the identification of novel compounds, including antimicrobials and other gene products of commercial interest (Daniel 2005). Metagenomes have been assembled from many soil environments, including forest, woodland, river valley and cultivated habitats (Daniel 2005). Studies specifically targeting rhizosphere microbiomes have been conducted for the model plant Arabidopsis thaliana (Hirsch and Mauchline 2012), Glycine max (soybean) (Mendes et al. 2014), the pasture legume Lotus japonicus (Unno and Shinano 2013) and the grey mangrove Avicennia marina (Alzubaidy et al. 2016). As yet, no true metagenomic comparison of mycorrhizal and non-mycorrhizal microbiomes has been made. As metagenomic libraries are derived directly from environmental samples, they are assumed to contain the collective genomes of all organisms contained in that environment (Handelsman et al. 1998). It has been estimated that a gram of soil contains around 1000 Gbp of microbial DNA; consequently, metagenomic methods generate vast libraries of molecular data (Mocali and Benedetti 2010). While the analytical possibilities they offer are equally vast, thorough analysis of metagenomic data is often impractical. For this reason, most studies of rhizosphere and

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 9 mycorrhizosphere microbiomes employ methods aimed primarily at identification of bacterial taxa. As most mycorrhizospheres remain uncharacterised, these metabarcoding methods still supply data of considerable value (Orgiazzi 2015). In recent years, pyrosequencing of amplified 16S rRNA obtained directly from soil samples, targeting key taxonomic marker regions, has emerged as the principle method of characterising mycorrhizospheres and their surrounding bulk soils (Uroz et al. 2012; Vik et al. 2013; Nguyen and Bruns 2015). These studies identified between 1244 and 47,292 operational taxonomic units (OTUs) in their samples, likely varying according to sampling methodology, soil type, and host plant. While many of these OTUs only poorly matched known bacterial taxa and consequently could not be reliably assigned to species, sufficient taxonomic resolution was achieved to identify many to the level of genus or family. This can form the basis of subsequent functional analysis. While they do not offer the same scope for functional analysis that metagenomic methods provide, 16S rRNA sequencing methods do have the potential to offer valuable insights into the composition of mycorrhizospheres.

1.4. Mycorrhiza helper bacteria

1.4.1. Characteristics of mycorrhiza helper bacteria Mycorrhiza helper bacteria (MHB) are bacteria found to exert a positive effect on the formation or functioning of mycorrhizas. Following Frey-Klett et al. (2007), the abbreviation MHB is used here to indicate bacterial strains that exert either or both of these effects, but a distinction can be made between mycorrhization helper bacteria, which increase the colonisation of plant roots by mycorrhizal fungi, and mycorrhiza helper bacteria, which work to sustain existing mycorrhizas. This distinction has been only imperfectly observed in the literature to date. Consequently, though most experimental attention has been paid to bacteria involved in fostering mycorrhizal development (i.e. mycorrhization helper bacteria), the term “mycorrhiza helper bacteria” has become standard and is used here. The scientific study of the helper effect has a long history, but many host plants remain unexplored and the mechanism or mechanisms underpinning this effect remain poorly understood. Additionally, traits of agronomic interest, such as the specificity of MHB action and the dosage required, have not been studied extensively and evidence is generally drawn from one or a few mycorrhizal systems.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 10 The first report of a bacterium exerting a helper effect in experimental conditions comes from Mosse (1959, 1962), who reported that the presence of Pseudomonas species in the root environment of Trifolium (clover) facilitated mycorrhizal colonisation and development of arbuscules by “Endogone sp.” (=Funneliformis mosseae: Rodrigues and Rodrigues [2015], Schüßler and Walker [2010]). In the absence of the bacterium, the fungal partner was unable to form penetrative structures (appressoria) and enter the host root, even when the root had been mechanically damaged to encourage access. This observation is unique: while later experiments have reported an enhancement in growth or an increase in colonisation induced by MHB strains, this remains the only report of a helper strain facilitating mycorrhizal colonisation in an otherwise non-infective system. Weaker stimulatory effects were noted for the exogenous application of a commercial preparation of pectinase and cellulose and, separately, for the application of the chelating and macerating agent ethylenediaminetetra-acetic acid (EDTA). To date, more than 35 studies have been conducted to screen bacterial strains for helper effects in AM symbioses (Table 1-1.). Two standard protocols have emerged from this process: the in vitro confrontation assay (generally modified from Duponnois and Garbaye [1990] and Dunstan et al. [1998]) and the glasshouse- or microcosm-based in vivo co-inoculation experiment. Confrontation assays are employed primarily to determine the effect of candidate MHBs on the pre-symbiotic phase of fungal growth. This phase begins with spore germination and includes the subsequent elongation of the hyphae and growth of the mycelium. Most evidence derived from this process concerns increases in hyphal length or colony size, though several studies have reported alterations in fungal architecture rather than increased growth (Aspray et al. 2013, Deveau et al. 2007, Aspray et al. 2006a). Changes in fungal architecture such as increased hyphal branching may lead to greater contact between the fungus and the plant host and consequently greater mycorrhization (Frey-Klett et al. 2007). Confrontation assays have been used for bulk screening and assessment of bacterial strains in culture (e.g. Dunstan et al. 1998). By contrast, co-inoculation experiments are designed to assess the impacts of selected bacterial cultures on the formation of mycorrhizas in live plants. This entails measuring the number of mycorrhizal roots or the total area of root colonised under single inoculation (fungus or bacteria) and dual inoculation (fungus and bacteria) treatments. The evidence for mycorrhization given in Table 1-1. derives principally from these co-inoculation experiments. Plant growth

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 11 parameters – chiefly root and shoot dry weights – are occasionally employed as proxies for mycorrhization or as measures of the efficacy of mycorrhizas, though improvements in mycorrhization do not always correspond with enhancements in plant growth (Founoune et al. 2002b, Dunstan et al. 1998). Identified MHB strains do not form a monophyletic grouping: species belonging to genera in the (including Pseudomonas and Rhizobium) and Firmicutes (chiefly Bacillus) make up the most commonly studied MHBs (Table 1-1.). This diversity includes both Gram-negative and Gram-positive species. The Gram-negative genus Pseudomonas remains the best studied: the model MHB P. fluorescens BBc6R8 was the first helper strain to have its genome sequenced (Deveau et al. 2014). Nevertheless, members of Gram-positive genera Bacillus have been repeatedly observed to enhance either pre-symbiotic fungal growth or mycorrhization (Mamatha et al. 2002, Toro et al. 1997). The capacity of these species to serve as helper bacteria is unlikely to be a consequence of shared evolutionary history but may represent convergence in a shared environment. It must be noted that the dependence on culture-based methods of screening means that the species found so far to exert helper effects likely do not represent the sum of all MHB strains; rather, they represent the MHB strains found in the cultivable fraction of the soil microbiota. The degree of specificity involved in AM helper effects is unclear. Specificity may affect any of the interactions involved in the tripartite mutualism: fungus-plant, fungus- MHB and MHB-plant. A single study reported that Paenibacillus spp. isolated from wheat enhanced mycorrhization in wheat plants, whereas the candidate MHB strain Paenibacillus brasilensis PB177, isolated from maize, did not (Arthurson et al. 2011). In another case, MHB strains selectively enhanced mycorrhizal fungi while suppressing a pathogenic species: UPMP3 and Burkholderia cepacia UPMB3 enhanced the in vitro spore germination and hyphal growth of R. irregularis UT126 and G. clarum BR152B, while inducing growth defects in the Ganoderma boninense PER 71 (Sundram et al. 2011). By contrast, at least one MHB appears to operate in both ECM and AM associations (Duponnois and Plenchette 2003): P. monteilii HR13 improved colonisation of the dual mycorrhizal shrub Acacia holosericea by species of Pisolithus and by R. irregularis. The behaviour and localisation of MHB strains within the symbiosis following dual inoculation has not been studied extensively in AM systems. Likewise, of the few available studies documenting dosage responses following MHB inoculation, only one

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 12 concerns an AM symbiosis. In an association between F. mosseae and wheat, MHB strains Paenibacillus polymixa B1-B3 were found to operate at dosages of 106 and 108 cells g-1 , while P. polymixa B4 only exerted a helper effect at the higher dosage (Arthurson et al. 2011). While MHB strains are often found to decline in abundance rapidly after inoculation, this decline was found to be less swift when P. fluorescens 92rk was co-inoculated with F. mosseae (Gamalero et al. 2004) and when P. monteilii was co-inoculated with Gl. fasciculatum (Singh et al. 2013). This effect may be due to the ability of rhizosphere bacteria to feed on the storage , which has been identified in both AM and ECM fungi (Nehls et al. 2010, Ocón et al. 2007), and has been found to serve as a chemoattractant for the MHB strain P. fluorescens BBc6R8 (Deveau et al. 2010, Frey-Klett et al. 2007).

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 13 Table 1-1. Evidence of a helper effect on mycorrhization, growth of fungal mycelium, or plant growth parameters in arbuscular mycorrhizas

Bacterial strains Host plant(s) AMF Reference AC AG AL AS AZ BA BE BR BU CO CU EN KL ME PA PE PN PS RH SI SR ST ? Trifolium parviflorum + = Trifolium pratense + Trifolium glomeratum + Trifolium subterraneum = Fm Mosse 1962 Dactylis glomerata + Triticum aestivum + Cucumis sativa + Allium cepa + Bagyaraj and Solanum lycopersicum Gfa = Menge 1978 Asp = Pennisetum Gfa + Rao et al. 1985 americanum Gi.m + Meyer and Trifolium subterraneum Mix + Linderman 1986b None (in vitro) Fm + Azcón 1987 Will and G. + = F+ Sylvia 1990 Fm ~ Azcón et al. Medicago sativa Gfa ~ 1991 Gca = Paula et al. Ipomoea batatas Gcl + = 1992 Multiple spp. Mix + Von Alten et

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 14 Table 1-1. Evidence of a helper effect on mycorrhization, growth of fungal mycelium, or plant growth parameters in arbuscular mycorrhizas

Bacterial strains Host plant(s) AMF Reference AC AG AL AS AZ BA BE BR BU CO CU EN KL ME PA PE PN PS RH SI SR ST ? al. 1993 Glycine max Fm + + Xie et al. 1995 Toro et al. Allium cepa Ri + = 1997 Gco = ~ Requena et al. Anthyllis cytisoides Ri ~ = 1997 Barea et al. Solanum lycopersicum Fm + 1998 Budi et al. Sorghum bicolor Fm + 1999 Zea mays = Vosátka and Gfi Solanum tuberosum + Gryndler 1999 Fester et al. Hordeum vulgare Ri + + + 1999 Abdel-Fattah and Sorghum bicolor Ri + Mohamedin 2000 Gfa = Carica papaya Mamatha et al. Fm + Morus alba 2002 Gca + Duponnois and Acacia spp. Ri + Plenchette 2003 Fm Vivas et al. Lactuca sativa = Ri 2003

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 15 Table 1-1. Evidence of a helper effect on mycorrhization, growth of fungal mycelium, or plant growth parameters in arbuscular mycorrhizas

Bacterial strains Host plant(s) AMF Reference AC AG AL AS AZ BA BE BR BU CO CU EN KL ME PA PE PN PS RH SI SR ST ? Gamalero et al. Solanum lycopersicum Fm ~ 2004 Babana and Triticum aestivum Ri + ~ Antoun 2005 Vivas et al. None (in vitro) Fm F+ 2005 Medicago truncatula Fm # Pivato et al.

Solanum lycopersicum Gi.r # 2009 Al Kim et al. Capsicum annuum Gcl P+ 2010 Ri Gag =

Pelargonium Gfa = Alam et al.

graveolens Fm = 2011 Ri = Artursson et al. Triticum aestivum Fm ~ 2011 Ri F+ F+ Sundram et al. None (in vitro) Gcl F+ F+ 2011 Fernández Glycine max Ri + Bidondo et al. 2011 Gfa = Kumar et al. Sorghum sp. Gag = 2012 Coleus forskohlii Gfa P+ Singh et al.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 16 Table 1-1. Evidence of a helper effect on mycorrhization, growth of fungal mycelium, or plant growth parameters in arbuscular mycorrhizas

Bacterial strains Host plant(s) AMF Reference AC AG AL AS AZ BA BE BR BU CO CU EN KL ME PA PE PN PS RH SI SR ST ? 2013 Vafadar et al. Stevia rebaudiana Ri P+ P+ P+ 2014 Fernández Daucus carota Ri = = = = = ~ = + = Bidondo et al. 2016 Omirou et al. Vigna unguiculata Mix ± 2016 Imperiali et al. Triticum aestivum Ri = 2017 de Novais et None (in vitro) Fm F= F= F= F= F= al. 2019 Leyva-Rojas et None (in vitro) Pina F+ F~ al. 2020 Bourles et al. Tetraria comosa Cet + 2020 Experimental evidence of a promotional effect exerted by bacterial strains on the formation of arbuscular mycorrhizas. Effects on mycelial growth are reported where mycorrhization was not studied and are indicated by the prefix (F). Effects on plant growth are reported where neither mycelial growth nor mycorrhization were studied and are indicated by the prefix (P). (+) Positive effect on mycorrhization. (=) Neutral or non-significant effect on mycorrhization. (~) Mixed effect of multiple bacterial strains on mycorrhization (strain-specific effect). (±) Mixed effect of single strain on mycorrhization in multiple plant-fungus systems (mycorrhiza-specific effect) or in varying environments (condition-specific effect). (#) Both strain-specific and host-specific effects. Bacteria are Acetobacterium (AC), Agrobacterium (AG), Alcaligines (AL), Azospirillum (AS), Azotobacter (AZ), Bacillus (BA), Brevibacillus (BE), Bradyrhizobium (BR), Burkholderia (BU), Cohnella (CO), Curtobacterium (CU), Enterobacter (EN), Klebsiella (KL), Methylobacterium (ME), Pandoraea (PA), Paenibacillus (PE), Pantoea (PN), Pseudomonas (PS), Rhizobium (RH), Sinorhizobium (SI), Streptomyces (SR), Stenotrophomonas (ST), and unknown (?). AMF species are sp. (Asp), Acaulospora longula (Al), Claroideoglomus etunicatum (Cet), Funneliformis mosseae (Fm), Gigaspora margarita (Gi.m), Gigaspora rosea (Gi.r), Glomus spp. (G.), (Gag), Glomus caledonium (Gca), Glomus clarum (Gcl), Glomus coronatum (Gco), Glomus fasciculatum (Gfa), Glomus fistulosum (Gfi), indica (Pina) and (Ri). aP. indica is not a true AMF but is functionally equivalent (Leyva-Rojas et al. 2020).

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 17 The establishment of the confrontation assay and co-inoculation experiment as the principal means of studying MHBs has meant that evidence for the existence of a helper effect takes one of two forms: alterations in the pre-symbiotic growth of the fungal partner or increases in the extent of root colonisation. These effects do not always correspond, however, and the behaviour of bacteria in culture does not necessarily mirror its behaviour in situ (Bowen and Theodorou 1979). This suggests that the results of in vitro confrontation assays may be unreliable proxies for in vivo colonisation rates. Similarly, in vivo colonisation studies conducted in axenic environments may themselves be poor proxies for the behaviour of MHB strains in real-world situations. Only three studies investigate the efficacy of MHBs in non-sterile AM symbioses. Omirou et al. (2016) reported that co-inoculation of an AM fungus and a Bradyrhizobium isolate in Vigna unguiculata (cowpea) resulted in an increase in AM colonisation and plant growth relative to a single inoculation treatment. By contrast, Paula et al. (1992) and Imperiali et al. (2017) both found that the helper effect was diminished in non-sterile soils. In the latter study, which involved the inoculation of a Pseudomonas strain into a wheat mycorrhiza, the effect was eliminated entirely. The complexity and unfamiliarity of the rhizosphere is implicated in the failure of many promising microbial biofertilisers to maintain their efficacy in the transition from laboratory to field conditions (Parnell et al. 2016), and direct study of the action of biological agents in non-sterile conditions is required to overcome this limitation. This principle has not been applied in the case of helper bacteria. Instead, following the codification of the helper effect by Garbaye (1994), experimental attention began increasingly to focus on possible mechanisms of action. The evidence for these is reviewed in the following sections.

1.4.2. The helper effect Evidence for the existence of a bacterial helper effect was first reviewed by Garbaye (1994), who identified five possible modes of action:

1. Increased root receptivity, possibly by the hormonal stimulation of root development or the enzymatic degradation of walls. 2. Manipulation of plant-fungus signalling and recognition mechanisms. 3. Improvement in fungal growth and survival, particularly in the pre- symbiotic stage.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 18 4. Modification of soil conditions in the rhizosphere, for example by solubilisation of nutrients, protection from toxins, or alterations to soil pH. 5. Increased germination of fungal propagules.

Crucially, these effects are not mutually exclusive, and the ‘helper effect’ may in fact be a composite of multiple independent effects. The volume of evidence so far collected for each of these hypotheses varies considerably. The standardisation of the in vitro confrontation assay has tended to focus attention on the interactions between bacteria and fungi in their pre-symbiotic state. For this reason, Garbaye’s third hypothesis has received the most experimental attention. The capacity of MHB strains to improve mycelial growth under experimental conditions has been extensively demonstrated (see Table 1-1.). The volume of evidence collected for this effect has led to the suggestion that enhanced pre-symbiotic growth is the principal mode of action of MHB species (Frey-Klett and Garbaye 2005). Crucially, however, observed increases in fungal growth in vitro do not reliably translate to enhancements in mycorrhization in vivo (Labbé et al. 2014, Dunstan et al. 1998, Bowen and Theodorou 1979). Additionally, recent research has suggested that mutants of P. fluorescens BBc6R8 deficient in Type III secretion promote mycelial growth in vitro but do not promote mycorrhization (Cusano et al. 2011). This evidence suggests that enhancements in mycorrhization and improvements in mycelial growth are at most only partially correlated, and that Garbaye’s third hypothesis is not strongly supported. The modification of rhizosphere soil conditions (Garbaye’s fourth hypothesis) has been only sparsely studied. In its broadest sense, this hypothesis covers enhancements in soil nutrition, inhibition of pathogens, alleviation of toxic effects, and alterations in abiotic factors such as pH. Many MHB strains are reported to solubilise and to produce iron-chelating siderophores (Artursson et al. 2006), and a nutritive effect of MHB inoculation has been suggested; however, phosphate solubilisation and siderophore production are also associated with PGPR species (Lugtenberg and Kamilova 2009), many of which lack a demonstrated MHB effect. While enhancements in root growth are assumed to increase opportunities for mycorrhizal colonisation (Probanza et al. 2001), little evidence directly links improvements in fungal nutrition to enhancements in mycorrhization, particularly in AM systems. A role for siderophore production has not been demonstrated and evidence for phosphate solubilisation is poor:

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 19 in a Triticum aestivum–R. irregularis mycorrhiza, phosphate-solubilising bacteria were observed to exert positive, neutral, and negative effects on colonisation, depending on the strain involved (Babana and Antoun 2005). Stronger evidence exists for a role for toxin resistance and inhibition of pathogens in certain helper effects: Brevibacillus brevis has been shown to enhance mycelial growth and spore germination in conditions of high zinc and cadmium stress (Vivas et al. 2005), and several MHB strains have been shown to antagonise pathogens in cultures (Singh et al. 2013, Sundram et al. 2011). Alterations in pH have not been studied in AM systems. The possibility of an effect of MHB species on spore germination (Garbaye’s fifth hypothesis) was investigated by Mosse (1959), who reported that germination of F. mosseae was enhanced in the presence of soil, perhaps due to microbial activity or, in autoclaved soils, microbial exudates. Azcón (1987) reported that co-inoculation with whole cultures or filtrates of an unidentified rhizosphere bacterium enhanced hyphal development but not germination of F. mosseae spores. Similarly, Barea et al. (1998) reported no enhancement of germination by Pseudomonas strains positive or negative for the antifungal compound 2,4-diacetylphloroglucinol (DAPG). Treatment with isolated DAPG reduced mycelial growth but did not affect germination. In a more recent study, germination of R. irregularis spores was not affected by co-inoculation of two Paenibacillus strains (Fernández Bidondo et al. 2011). In cultures contaminated with toxic quantities of zinc and cadmium, F. mosseae spores showed greater germination and mycelial growth when co-inoculated with an MHB from Zn- and Cd- contaminated soils, Brevibacillus brevis, relative to a bacteria-free control (Vivas et al. 2005). In contrast with previous studies, the MHB employed here appeared to enhance germination even in treatments without Cd or Zn. Improvements in spore germination were also reported by Xavier and Germida (2003) and Sundram et al. (2011). While a general enhancement of germination by MHB strains is not supported by the evidence, a species-specific effect cannot be ruled out. The contradictory evidence described here suggests that Garbaye’s third, fourth, and fifth hypotheses cannot be accepted as universally true, though they may each account for a subset of the observed helper effects. Garbaye’s first and second hypotheses, improved root receptivity and altered plant-fungus signalling, are here treated as components of a broader root receptivity hypothesis. Early work in this area focused primarily on exudation of biomolecules by MHB strains, including enzymes and phytohormones. Additionally, more recent work has suggested a role for direct

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 20 secretion of bacterial effectors and targeted manipulation of host immune behaviours and gene transcription programmes in this broader concept of root receptivity. These early results and new directions are discussed in detail in the following sections.

1.5. The root receptivity hypothesis

1.5.1. Bacterial biomolecules and root receptivity The model of root receptivity proposed by Garbaye (1994) extended from earlier work suggesting a role for degradative enzymes in the penetration of plant tissues by fungi. Mosse (1962), the first to report a helper effect exerted by a Pseudomonas strain, found that application of Pectinol, a commercial product containing both cellulolytic and pectinolytic components, induced some root penetration of a host Trifolium species by F. mosseae, though less than was observed when live Pseudomonas cultures were co- inoculated. Improved penetration was also found in the application of EDTA, a chelating agent which Mosse (1962) suggested may macerate the root tissues and allow colonisation. This effect was weaker than the effect obtained with Pectinol, suggesting that there may be an enzymatic effect over and above the maceration of the root wall, but the poor performance of the Pectinol solution relative to live cultures indicates that the helper effect is not purely due to the action of these enzymes. Little attention has been paid to the enzymatic hypothesis since Mosse (1962), though a recent study suggests it is not well supported: Fernández Bidondo et al. (2016) reported that fewer than half of 25 bacterial isolates obtained from AM propagules exhibited either cellulolytic or pectinolytic activity and that the presence of these activities did not correspond with MHB-like effects. The most effective MHB strains identified in that study did not exhibit pectinolytic, cellulolytic, or xylanolytic activity. A role for the indole-3-acetic acid (IAA) in MHB activity was suggested by its ability to increase the initiation of short roots in Douglas fir, an ECM host (Duponnois 1992). This auxin-mediated enhancement in mycorrhization may also apply to AM associations, though AM fungi are not known to produce IAA in the quantities reported for ECM species (Gopinathan and Raman 1992). IAA was not found in the hyphae of R. irregularis, though trace amounts were found in its spores (Ludwig-Müller and Güther 2007). Nevertheless, IAA appears to be implicated in AM formation: pea and tomato mutants deficient in IAA biosynthesis have been found to produce fewer mycorrhizas (Gutjahr 2014). In the IAA-deficient bushy mutant of Pisum sativum (pea),

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 21 poor mycorrhization was linked to low production of signalling molecules, which may be a consequence of reduced auxin concentrations (Foo 2013). This evidence suggests a role for bacterial IAA in at least some helper effects, though IAA production and helper effects are not precisely correlated: Fernández Bidondo et al. (2016), for example, reported that IAA production was observed among both helper strains and strains with neutral effects on mycorrhizal colonisation, as well as several strains that significantly inhibited fungal spore germination. For this reason, IAA alone is insufficient to account for all helper effects. Exogenous application of purified nodulation (Nod) factors and soybean flavonoids have both been found to stimulate mycorrhizal formation in Glycine max (soybean) (Xie et al. 1995). Nod factors resemble the Myc factors (i.e. LCOs) produced by mycorrhizal fungi and are assumed to stimulate mycorrhization in the same manner. Nod factors are produced by rhizobia and may account for the MHB effects reported for species of the Rhizobiales. The role of flavonoids in mycorrhization remains underexplored, though they are exuded by plant roots prior to rhizobial, mycorrhizal, and actinorrhizal symbioses and may serve as signalling molecules (Abdel-Lateif et al. 2012). Nod factors were found to stimulate production of flavonoids in soybean (Xie et al. 1995); production of Nod or Nod-like factors may thus account for some helper effects. Several additional compounds of interest have been identified from ECM systems, but most are not known from AM fungi or their helpers. An exception is the common carbohydrate trehalose, first implicated in an ECM helper effect (Duponnois and Kisa 2006) and later isolated from the AM helper Paenibacillus validus. Co-inoculation of R. irregularis with P. validus has been shown to prolong the life of the fungus in vitro and enable it to produce viable spores, an observation that is unique among AM fungi (Hildebrandt et al. 2002). Saccharides produced by P. validus were screened for their effects in stimulating fungal growth and spore production but, while raffinose and trehalose both increased hyphal growth, these compounds alone did not produce the persistence and sporulation observed for live cultures. Evidence so far assembled suggests that it is unlikely that any single molecule is solely responsible for the helper effect. Garbaye’s hormone- and enzyme-focused root receptivity hypothesis is not sufficient to account for the action of MHBs. Nevertheless, root receptivity to fungi remains an important consideration in discussions of the helper effect. Alternative models of root receptivity are considered here.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 22 1.5.2. Root receptivity and host-plant transcriptional activity The formation of an entails significant restructuring of the plant cell wall, and this entails a collaboration in which the plant is an active partner (Balestrini and Bonfante 2014). This process begins with the pre-symbiotic exchange of signal molecules: plant-produced compounds (e.g. strigolactones) induce hyphal branching in the fungus and fungal signal molecules (short-chain COs and LCOs) activate symbiosis-associated transcriptional pathways in the host plant (Genre et al. 2013, Murray et al. 2013). The collaboration continues after hyphal penetration of the root and is involved in the formation of the symbiotic interface: it has been shown, for example, that the silencing of two vesicle-associated membrane protein (VAMP) genes in Medicago truncatula prevents the formation of arbuscules (Ivanov et al. 2012). These genes are implicated in the formation of the periarbuscular membrane; without this participation from the host plant, mycorrhization cannot proceed. The root receptivity hypothesis, as framed by Garbaye (1994), concerns itself mainly with root penetration and does not account for the enduring nature of the plant-fungus collaboration. Bacteria may act at several stages of the mycorrhization process, for example by contributing to signalling or hyphal branching, stimulating symbiosis-associated genes, or suppressing defence responses in the plant host. To date, these possibilities remain underexplored, but upregulation of symbiosis-associated genes has been observed in Medicago truncatula in response to colonisation by P. fluorescens (Sanchez et al. 2004). In that study, 12 genes upregulated during colonisation by F. mosseae BEG12 were screened following inoculation with the MHB P. fluorescens C7R12 and the nodulating bacterium Sinorhizobium meliloti 2011. The MHB was found to cause a >2.5-fold increase in expression of six of those 12 genes, compared to only two for S. meliloti. The greatest increase in expression was observed for two enzymes implicated in plant responses to biotic and abiotic stresses, a glutathione-S-transferase (Gullner et al. 2018) and a germin-like protein (Dunwell et al. 2008), as well as a component of the nuclear protein import pathway, importin subunit beta 1, which has itself been implicated in drought tolerance in Arabidopsis thaliana (Luo et al. 2013). Smaller increases were observed for a nodulin 26-like aquaporin and two genes of unknown function. While the ecological significance of this upregulation is unknown, and was typically smaller than that induced by AM colonisation, this study suggests a role for MHBs in altering host-plant transcriptional activity. It is unclear, however, how the

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 23 upregulation of genes associated with plant responses to biotic stress could lead to increased receptivity of the root to mycorrhizal colonisation. In its broadest sense, root receptivity describes the hospitability of the root environment to colonising microorganisms. To avoid disease, plant tissues need to be inhospitable to invading pathogens; consequently, microbes that attempt to colonise plant tissues are confronted by a suite of immune mechanisms. Nevertheless, many bacteria are able to form persistent microcolonies known as on root surfaces (Danhorn and Fuqua 2007). These biofilms allow molecular communication between bacteria in the colony, facilitating changes in gene regulation that allow colonies to adapt to adverse conditions (‘quorum sensing’) (Soto et al. 2006). A consists of multiple layers of bacteria surrounded by a mucus envelope (Danhorn and Fuqua 2007), which, in tandem with adhesin-like proteins, adheres the colony to the root (Boyd et al. 2014), and subsequently protects the bacterial colony. The mucus may be acidified, as in the case of Rhizobium leguminosarum bv. viciae (Russo et al. 2006), or contain antimicrobial agents like surfactin, which has been located in biofilms of a Bacillus subtilis strain (Bais et al. 2004). Biofilm formation is assumed to be common among rhizosphere bacteria (Walker et al. 2004) and has been relatively well studied in and PGPR species (see, for example, Ansari and Ahmad 2019, Planchamp et al. 2015, Sang and Kim 2014, Cho et al. 2013, Poupin et al. 2013, Sathyapriya et al. 2012, Rosas et al. 2009). To date, Paenibacillus polymixa B1 and B2 are the only bacterial strains so far studied in the context of biofilm formation with known MHB effects (Arthurson et al. 2011, Timmusk et al. 2005); however, isolates of Bacillus subtilis (Bais et al. 2004), B. cereus (Majed et al. 2016) and Pseudomonas fluorescens (Lemos et al. 2015, Boyd et al. 2014) have been observed to form biofilms, and these parent species are known to contain MHB strains. Current evidence indicates that the formation of biofilms is more likely to raise defences than lower them. Colonisation of plant roots by bacteria, whether pathogenic or growth-promoting, is consistently reported to activate host immune behaviours (Vogel et al. 2016, Planchamp et al. 2015, Cho et al. 2013, Poupin et al. 2013,Sathyapriya et al. 2012), though the number of genes differentially regulated ranges from <10 to >400 and immune activity can be local (i.e. confined to root tissues) or systemic (Vogel et al. 2016, Planchamp et al. 2015). This difference may correspond to the PGPR status of the bacterium. Vogel et al. (2016) report, for example, that Methylobacterium extorquens PA1, a member of a genus containing few PGPR strains,

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 24 induced only a very weak immune response in Arabidopsis thaliana, whereas melonis Fr1, a close relative of several PGPR species known to confer disease resistance, triggered pronounced transcriptional changes. It should be noted, however, that inoculation of maize roots with the PGPR Pseudomonas putida KT2440 has been found to confer resistance against the maize anthracnose fungus (Colletotrichum graminicola) despite provoking only a weak immune response (Planchamp et al. 2015). Thus, the correspondence between transcriptional response and PGPR status may not be exact. Additionally, if PGPR status and host immune response are in fact correlated, the brevity of observed responses requires some explanation. Pseudomonas aeruginosa, an occasional PGPR conferring disease tolerance, has been observed to provoke a defence response in Elaeis guineensis (oil palm), including the upregulation of chitinase and β- 1,3-glucanase, but this response was found to peak five to seven days after infection and to decline thereafter (Sathyapriya et al. 2012). Following this, P. aeruginosa expanded from its surface biofilm into the interior tissues and vascular system of the host. Paenibacillus polymyxa, a PGPR and MHB, follows a similar pattern, initially colonising the root tips before proceeding into the intercellular spaces of the root cortex (Timmusk et al. 2005). This suggests that host plants undergo a process of accommodation to the bacterial coloniser. It has been suggested that this may result from the manipulation of host immune responses by bacterial secretion of effector molecules. Many effectors are targeted at aspects of the host immune response (Le Fevre et al. 2015). It is possible that MHBs use these secretory mechanisms to suppress responses to their own colonisation, and in so doing facilitate colonisation by mycorrhizal fungi.

1.5.3. Root receptivity and bacterial secretion Bacteria possess a number of mechanisms that function to transport small molecules out of the cell . These substrates may remain embedded in the outer membrane or they may be secreted into the extracellular environment. In some cases, molecules may be transported directly into the cytoplasm of neighbouring cells (Costa et al. 2015). The substrates transported in this way vary greatly in their structure and function but have often been found to inhibit aspects of the target cell’s defensive response, creating conditions suitable for further invasion by the bacterium (Green and Mecsas 2016). Nine secretion systems have so far been identified; of these, eight are found in Gram-

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 25 negative bacteria (Types I-VI and Types VIII-IX) and two (Type IV and Type VII) are found in Gram-positive bacteria (Burdette et al. 2018, Veith et al. 2017, Green and Mecsas 2016, Abby et al. 2016). Secretion systems are of interest in the study of human disease (Coburn et al. 2007) and plant pathogenesis (Chang et al. 2014), but have also been proposed as possible modes of action for MHB species, facilitating chemical communication between the symbiotic partners (Viollet et al. 2017, Cusano et al. 2011). In the context of plant pathogenesis and plant growth promotion, the Type III (T3SS), Type IV (T4SS) and Type VI (T6SS) secretion systems have attracted the most interest to date (Mitter et al. 2016). These mechanisms have been identified in both PGPR and pathogenic species; in the latter, virulence has been shown to depend on functional Type III secretion (Liu et al. 2015). The T3SS, T4SS, and T6SS have also been located in rhizobia through genomic sequencing (Nelson and Sadowsky 2015). Type III secretion is especially common in plant-associated microbial communities: an analysis of metagenomic databases reported a strong association between T3SS genes (hrp1 and hrp2) and bacterial metagenomes derived from plant communities (Barret et al. 2013). In the same analysis, T3SS and T6SS genes were commonly found among members of the phylum Proteobacteria, which contains many soil- and rhizosphere- colonising genera, including Pseudomonas and Burkholderia. Though they differ considerably in structure (Chen et al. 2015b, Diepold and Armitage 2015), the T3SS and T6SS both function by transporting proteins, known as effectors, into the cytoplasm of target cells via a syringe-like injectisome (Costa et al. 2015; Green and Mecsas 2016). By contrast, the T4SS translocates both DNA and protein substrates, including virulence factors and antibiotic-resistance genes (Wallden et al. 2010). In plants, most effectors studied to date appear to act against components of the pathogen-associated molecular pattern-triggered immunity (PTI) system (Sohn et al. 2012). PTI is an immune mechanism that is activated in response to certain recognised indicators of infection, including flagellin, the primary component of the bacterial flagellum (Zipfel and Robatzek 2010). Suppression of this response facilitates further infection of the host. Effectors may act against hormone synthesis, reactive oxygen species (ROS) production, gene transcription, or other essential biological activities; they may function as enzymes, toxins, or transcription factors, activating targeted genes (Le Fevre et al. 2015). Several species encoding more than 50 effectors have been identified (Koebnik and Lindeberg 2012). The pathogen Xanthomonas campestris pv versicatoria secretes more than 25 effectors; the degree of redundancy built into this

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 26 system is unknown, but at least three of these (AvrBsT, XopB and XopS) are implicated in suppression of host immunity (Schulze et al. 2012, Szczesny et al. 2010). Pathogens are not alone in deploying effectors: the PGPR strain P. fluorescens MFE01 secretes toxins that act against co-inhabiting bacteria, which may account for the biocontrol effect of this strain, though the same toxins may also act against eukaryotic cells (Ryu 2015). Further afield, the ECM fungus bicolor secretes the effector MiSSP7 which functions to repress the activity of jasmonic acid-associated signalling pathways in Populus hosts and may be necessary for mycorrhiza formation (Plett et al. 2014). Jasmonic acid (JA) is implicated in defensive responses to pathogens and the suppression of this response appears to facilitate mycorrhizal colonisation. JA is implicated principally in defence responses to necrotrophic pathogens (van der Does et al. 2013) and mycorrhizas may also require the suppression of salicylic acid-dependent signalling pathways, which primarily respond to biotrophic pathogens (Pozo and Azcón-Aguilar 2007). In support of this, tobacco plants with reduced levels of salicylic acid have been found to exhibit greater levels of AM colonisation (Medina et al. 2003). While the mechanisms governing fungal secretion depart in significant details from bacterial secretion, MHB effectors may operate in a similar fashion to L. bicolor. If MHB strains are able to suppress PTI or the hypersensitive response in order to facilitate their own colonisation, they may confer a similar benefit on mycorrhizal fungi. The involvement of bacterial secretion, and specifically the T3SS, in the helper effect was suggested initially by experiments undertaken in ECM systems. These are not considered here; however, a single study in a Medicago truncatula–F. mosseae symbiosis suggests that this effect may also apply to AM systems. Using T3SS+ and T3SS- strains of P. fluorescens C7R12, Viollet et al. (2017) demonstrated that the helper effect was contingent on a functional T3SS. This followed an earlier finding that pseudomonads positive for T3SS-associated genes were enriched in the rhizosphere of mycorrhizal varieties of M. truncatula relative to non-mycorrhizal genotypes (Viollet et al. 2011). The effectors involved in this T3SS-mediated effect are not known and it is unclear whether secretion is involved in the interaction between the bacterium and the plant or between the bacterium and the fungus (or, conceivably, both). Colonisation of both plants (Sanchez et al. 2004) and fungi (Deveau et al. 2007) by bacteria has been found to alter host gene regulation and there may be a role for Type III effectors in this process.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 27 While the available evidence is limited, this points to a possible role for Type III secretion in a subset of observed helper effects. By contrast, the T4SS and T6SS are yet to receive any attention in the context of MHB interactions. The T3SS and T6SS alone or together cannot account for all reported helper effects, as they are confined to Gram- negative bacteria (Costa et al. 2015) and both Gram-positive and Gram-negative helpers are known (Table 1-1.). If direct bacterial secretion is involved in the helper effects exerted by Gram-positive taxa, it is most likely to be via the T4SS. The T4SS has been isolated from plant-associated bacteria and remains a candidate for study (Goessweiner- Mohr et al. 2013, Wallden et al. 2010).

1.6. Biserrula as a target species Biserrula pelecinus L. (biserrula) is a pasture legume of Mediterranean origin, first collected for agronomic assessment in 1993 and first commercialised in Australia in 1997 (Loi et al. 2014, Nichols et al. 2012). Two cultivars now exist: B. pelecinus cv. Casbah and B. pelecinus cv. Mauro, released in 1997 and 2002 respectively (Nichols et al. 2007, Loi et al. 2006). Biserrula is one of 47 commercialised legume species currently employed in Australian agriculture and one of 30 to have first been domesticated in Australia. This abundance reflects, in part, the diversity of climate and soil types under cultivation in Australia (Nichols et al. 2012). Biserrula was initially selected for its tolerance of acid conditions and soils with low moisture-holding capabilities (Loi et al. 2014) and it remains of interest primarily in low-rainfall agriculture and soils with high acid stress. Its tolerance to drought extends in part from its rooting depth: roots of biserrula may reach 2 metres in depth and regularly extend below the roots of subterranean clover (Loi et al. 2005, Carr et al. 1999). In addition to the environmental stresses for which it was selected, the use of biserrula in Australian dryland agriculture exposes it to conditions of nutrient limitation, salinity, and sodicity (Carberry et al. 2011, Rovira 1992). Thus, there may be scope for microbial enhancement via the promotion of mycorrhization by helper bacteria. An unknown factor in this equation is the extent to which MHB strains operate in the acid soils to which biserrula is adapted, as these environments may favour fungi at the expense of bacteria (Rousk et al. 2009). To date, research into the species has focused on identifying the agronomic advantages of implementing biserrula in pasture systems. Biserrula is both hard-seeded and a prolific producer, and these seeds require considerable force to break down

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 28 (Hackney et al. 2013). Consequently, a large proportion of these seeds survive ingestion by sheep and very few are lost in “false breaks,” minor rainfall events that are sufficient to germinate the seeds of many species but not to sustain the subsequent growth of the plant. The premature germination of seed in response to false breaks is a significant limitation of the more widely planted subterranean clover (Loi et al. 2001). Additionally, biserrula exhibits low methanogenic potential relative to many other commercial legumes, making it a potentially valuable contributor to reductions in agricultural greenhouse gas emissions (Banik et al. 2019, Banik et al. 2013). Biserrula is not preferentially grazed by sheep, particularly in spring (Nichols et al. 2007). The tendency of sheep to preferentially graze co-occurring weeds rather than biserrula itself is widely construed as an advantage, as it suppresses weed growth and reproduction. The infrequent occurrence of photosensitisation remains the most significant limitation to the widespread adoption of the species (Kessell et al. 2015). No current estimates of total area sowed to biserrula area are available but roughly 1.5 million ha in Western Australia had been sowed to some combination of biserrula, French serradella, yellow serradella, and several newly developed clovers within five years of their commercialisation, suggesting that there is an appetite for new pasture solutions (Loi et al. 2005). In NSW, biserrula remains at present only infrequently utilised. A survey of 300 farmers and 33 farm advisors in central and southern NSW reported only a limited familiarity with biserrula: only 3-4% of landholders in each of the surveyed regions and 13% of advisors reported having a “good knowledge” of any member of a pool of 10 recently domesticated legumes, including biserrula. Nevertheless, more than 25% of respondents in each of the surveyed regions indicated an intention to use some combination of arrowleaf clover, balansa clover, and biserrula in the future (Hackney et al. 2008). The lack of technical knowledge concerning the species and its management remains a barrier to more extensive implementation. The widespread use of pasture legumes in Australian agricultural systems extends largely from their ability to increase the productivity of agricultural land by fixation of atmospheric nitrogen (Howieson et al. 2000). The abundance of pasture legumes in Australia reflects the naturally low levels of nitrogen in Australian soils, and biserrula is typically employed in ley-farming rotations for the purpose of enhancing the growth of the subsequent crop (Nichols et al. 2012). In common with other legumes, biserrula enters into symbiotic associations with N2-fixing bacteria termed rhizobia, which form a polyphyletic grouping across two bacterial classes, the and

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 29 Betaproteobacteria (Franche et al. 2009). Biserrula displays an unusually high specificity in its rhizobial associations: unlike legumes with longer histories of cultivation in Australia, biserrula will not enter into symbiotic associations with commonly occurring background rhizobial populations. Biserrula is reportedly nodulated by three strains of Mesorhizobium – M. australasicum, M. opportunisticum and M. cicero bv. biserrulae (Peix et al. 2015) – though a single report of nodulation by Rhizobium leguminosarum exists (Vicente et al. 2009). Screening of nodules for bacterial diversity returned a large number of bacterial isolates, potentially representing previously unknown nodulating species or biovars of known species, but these have not been identified (Vicente and Pérez-Fernández 2016). There is evidence that lateral transfer of a symbiosis island containing the genes nodA and nifH from nodulating Mesorhizobium strains is sufficient to enable nodulation by formerly non-nodulating strains (Nandasena et al. 2006). Both genes are implicated in symbiotic activity: nodA encodes an N-acyltransferase involved in synthesis (Guasch-Vidal et al. 2016) while nifH encodes a component of nitrogenase, the key enzyme in (Gaby and Buckley 2014). Nandasena et al. (2007) report that multiple Mesorhizobium strains capable of nodulation evolved in situ within six years of the introduction of biserrula into Western Australia, most likely as a consequence of horizontal transfer of this symbiosis island. This transfer may also account for the isolated record of nodulation by R. leguminosarum. While efforts have been made to identify constraints on the rhizobial symbiosis in biserrula (see Howieson and Ballard 2004), no information is presently available concerning its mycorrhizal associations. It is not clear whether biserrula exhibits a similar specificity in its relationships with soil fungi as it does with rhizobia. No published reports of mycorrhizal associations between biserrula and naturally occurring or artificially inoculated fungi exist. A single study does exist describing an attempt at inoculating biserrula, along with a suite of other pasture legumes, with a commercial mycorrhizal product containing four species of AM fungus (Kidd et al. 2016). That attempt yielded no significant mycorrhization for any of the legumes under study, most likely as a consequence of a deficiency, such as a lack of infective propagules, in the inoculum. Given the crucial role mycorrhizas play in phosphorus uptake, particularly in the nutrient-poor soils for which the implementation of biserrula is encouraged, this is a significant knowledge gap. The presence of mycorrhizas in biserrula must be established before any study of mycorrhiza-associated bacteria can be conducted.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 30 Several crop and pasture legumes, including lucerne, soybean, barrel medic and a number of Trifolium species, have been studied in the context of MHB activity, with mixed or positive effects reported for strains of Pseudomonas, Bradyrhizobium, Rhizobium, and Paenibacillus (Omirou et al. 2016, Fernández Bidondo et al. 2011, Pivato et al. 2009, Vivas et al. 2003, Xie et al. 1995, Azcón et al 1991, Meyer and Linderman 1986a, Mosse 1962), but given the high specificity of association between biserrula and its nodulating bacteria, and given the close resemblance of nodulation pathways to mycorrhization pathways, it is not clear that results from these legumes can be generalised to biserrula.

1.7. Conclusion The ability of certain bacteria to promote the formation and function of arbuscular mycorrhizas in axenic conditions is well established. Less clear is whether this effect can be maintained in a populated rhizosphere, where putative MHBs are likely to be subjected to intense interspecific competition. Of the dozens of bacterial taxa identified as MHBs in in vitro confrontation assays and axenic in vivo inoculation trials, very few have subsequently been screened in non-sterile conditions. In two out of three cases where non-sterile screening has been carried out, the helper effect was diminished or eradicated. This issue is an urgent one if MHB strains are to be successfully implemented into agricultural settings. In a similar vein, while many potential modes of action have been proposed to explain MHB activity, ranging from the exudation of tissue-degrading enzymes and the production of plant hormones to the direct secretion of protein effectors into fungal or plant cells, the mechanism or mechanisms underpinning the helper effect remain obscure. These are significant gaps in the existing knowledge of the rhizosphere processes that affect plant growth. Frey-Klett et al. (2007) suggested that the first priority of MHB research should be to identify helper strains in previously unexplored plant-fungus systems. To some extent, this advice has been followed (see Table 1-1.), but the majority of and host species remain unexplored. This includes many species of economic consequence, both established and emerging. Biserrula pelecinus belongs to the latter category and represents an ideal candidate for study: utilised in low-rainfall and high- stress agriculture, there is likely to be considerable scope for enhancing its productivity by the use of beneficial microorganisms. The study of helper effects in biserrula promises to expand our understanding of both a pasture species of economic importance

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 31 and the helper effect itself. Owing to the peculiar microbial dependencies of the biserrula root and the paucity of data regarding its non-rhizobial associations, it will first be necessary to demonstrate that biserrula is capable of forming mycorrhizas. Only then can studies be carried out to determine whether bacteria associated with biserrula are capable of acting as helpers in both sterile and non-sterile conditions. Subsequently, there is scope for experimental work aimed at deepening our understanding of the dynamics of the biserrula rhizosphere.

1.8. Aims The principal aim of this research project was to determine whether cultivable rhizobacteria were capable of promoting the colonisation of the pasture legume Biserrula pelecinus L. by mycorrhizal fungi. To do this, it was first necessary to demonstrate that biserrula was capable of forming mycorrhizal associations. The aims of subsequent experiments included (1) determining whether this helper effect was maintained in non-sterile environments mimicking real-world conditions, which entailed developing an understanding of the dynamics of the biserrula rhizosphere; (2) characterising the identified helper strains in terms of certain functional traits thought to be linked to the helper effect; and (3) identifying any potentially synergistic or detrimental microbe-microbe interactions to which the helpers might be exposed.

1.9. Thesis structure This thesis consists of a review of the available literature followed by four experimental chapters and a final summary and conclusion. Chapter 2 describes experiments conducted with the aim of determining whether biserrula forms mycorrhizas, the extent of this colonisation in natural and artificial conditions, and the identity of the fungal partner or partners involved in the symbiosis. Chapter 3 investigates the potential for biserrula-associated rhizobacteria to promote in vitro growth of arbuscular mycorrhizal fungi and in vivo mycorrhizal colonisation. Additionally, it describes experiments conducted to determine whether those bacteria facilitate sporulation by the fungus in vitro and to characterise those bacteria in terms of their functional traits and tolerances. Chapter 4 concerns the activity of the MHBs identified in the previous chapter in a biserrula rhizosphere populated by other microorganisms. The experiments described by

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 32 this chapter were conducted with the aim of determining whether helpers maintain their efficacy in non-sterile environments. Chapter 5 contains a detailed study of the dynamics of the rhizosphere across a two- season biserrula-wheat rotation, conducted with the aim of determining whether helper strains are likely to effectively colonise biserrula roots and persist across seasons and changes in host plant. Lastly, chapter 6 reviews the key findings of the research project and suggests directions for future research.

Chapter 1. Mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L. 33 Chapter 2. Observation and quantification of arbuscular mycorrhizas in the roots of the pasture legume Biserrula pelecinus L.

2.1. Introduction Arbuscular mycorrhizas (AM) are estimated to occur in at least 72% of all terrestrial plant species, including most crop and pasture plants (Brundrett and Tedersoo 2018, Van der Heijden et al. 2015). This value carries a significant caveat, however: direct observations of mycorrhizas in roots are available for only a small fraction of the world’s total floristic diversity. The estimate of 72% was derived by assigning mycorrhizal status to whole plant families based on observations from approximately 10,000 species (Brundrett 2009), representing 2.8% of known plant taxa (World Flora Online 2019). Inferences as to the mycorrhizal status of a particular plant are complicated by the existence of conflicting reports for individual species and of inconsistencies within plant families: approximately 8% of all terrestrial plants belong to families possessing both AM and non-mycorrhizal species (Brundrett and Tedersoo 2018). Complicating matters further, species widely regarded as non-mycorrhizal have been known to form rudimentary AM associations under certain conditions (Cosme et al. 2018). Consequently, while these estimates provide valuable insights into the ubiquity of AM associations, they cannot be used to infer that any given species is capable of forming mycorrhizas. Biserrula is a pasture legume belonging to the family (subfamily ). While its unusually stringent rhizobial associations are well documented (Peix et al. 2015), nothing is known of its mycorrhizal status. A single study exists describing an unsuccessful attempt to inoculate biserrula, along with a number of other legume species, with a commercial mycorrhizal product containing four species of AM fungus (Kidd et al. 2016), but the authors suggest that this may be a failing of the inoculum and not the host. As the family Fabaceae is known to harbour AM, ectomycorrhizal (ECM), non-mycorrhizal, and mixed AM-ECM species (Brundrett 2009), direct observation is required to determine firstly whether biserrula forms

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 34 mycorrhizal associations and, if it does, to which category those associations belong. No published accounts of mycorrhizal observations from Biserrula pelecinus are available. Numerous screening methods have been devised to allow both observation and quantification of mycorrhizal colonisation, based on visual estimation or subsampling of roots for microscopic observation, but the gridline intersect method facilitates the rapid screening of large root samples with an acceptably low margin of error, assuming a large enough sample is surveyed (Giovannetti and Mosse 1980). Beyond observing and quantifying mycorrhizal colonisation, it is valuable to identify the fungal partners participating in the symbiosis. While AM fungi are considerably less diverse than terrestrial plants and are consequently assumed to be generalists capable of infecting most, or perhaps all, AM hosts (Vályi et al. 2016, Lee et al. 2013), there is evidence to suggest that plants may selectively curate their AM communities (Torrecillas et al. 2012, Yang et al. 2012). It is unclear whether these selective communities confer greater benefit to the host plant than communities assembled from non-preferred species. It is possible, however, that optimising the AM symbiosis may mean ensuring that the preferred species are available to the plant. The use of DNA-based diversity profiling methods in a pasture environment allows for the assessment of host selectivity in situ and also permits the recovery of cryptic species that may not be distinguishable on the basis of morphology, or may not produce observable structures under the screening conditions (Chen et al. 2018). The persistence of mycorrhizal fungi in pastures is dependent on the conditions to which that pasture is subjected. Common management practices may profoundly alter the structure of mycorrhizal communities and the abundance and colonisation capacity of their constituent fungi (Turrini et al. 2016, Verbruggen et al. 2013). Soil fertilisation can reduce the abundance of AM fungi (Santos et al. 2006), while tilling has been observed to alter AM community structure (Lu et al. 2018, Alguacil et al. 2008) and to result in reduced spore density and root colonisation (Galvez et al. 2001). Furthermore, AM colonisation in both target and non-target species has been shown to be suppressed by the use of glyphosate-containing herbicides (Helander et al. 2018, Zaller et al. 2015). In addition to the effects of agricultural land management, environmental factors must also be considered. AM species diversity has been reported to increase following short- term waterlogging (Yang et al. 2016), although waterlogging has also been observed to suppress mycorrhizal root colonisation (Mendoza et al. 2005, Escudero and Mendoza 2005), and persistent waterlogging, as in a wetland environment, can prevent AM

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 35 colonisation altogether (Wirsel 2004). Pronounced seasonal fluctuations in AM diversity and abundance are also commonly observed and these seasonal effects may be primary determinants of AM community composition and colonisation success (Berruti et al. 2018, Santos et al. 2006, Escudero and Mendoza 2005). As a result of these factors, any sampling and screening exercise produces, at most, a snapshot of mycorrhizal colonisation at a particular time and under a particular set of land-use conditions. Nevertheless, this is a necessary precursor to any attempts to optimise the mycorrhizal symbiosis and confer the greatest benefit to the host plant.

2.2. Format of experiments This chapter presents three experiments conducted to determine whether Biserrula pelecinus is capable of forming mycorrhizal symbioses in natural (pasture) and artificial (trap culture) environments, and, subsequently, to characterise, quantify, and identify the fungal partner or partners involved in those symbioses. These experiments follow on from informal observations of mycorrhizal components in biserrula roots. The first experiment was designed with the aim of determining whether natural mycorrhizal colonisation could be observed in the roots of pasture-grown plants obtained by soil coring, determining whether colonisation could be induced in pot-grown plants exposed to an established trap culture, and determining the extent of mycorrhizal infection in each case. In the second experiment, microscopic analysis of selected mycorrhizal roots was utilised with the aim of characterising mycorrhizal structures in plant tissues, in conjunction with PCR-based methods of identifying the fungus or fungi responsible for producing these mycorrhizas. A third experiment made use of DNA-based diversity profiling methods in field samples with the aim of identifying putative mycorrhizal partners overlooked by the previous experiment.

2.3. Experiment 1: Quantification of mycorrhizas in the roots of field- and trap culture-grown biserrula The aim of this experiment was to determine whether biserrula plants form mycorrhizal associations and subsequently to assess the extent of mycorrhizal colonisation in biserrula roots obtained from field-collected root samples and from pot-based trap cultures. It was expected that, in common with other pasture legumes (Treseder 2013), biserrula would be found to harbour arbuscular mycorrhizas; however, owing to the

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 36 observed variability in colonisation rates among these species, the likely extent of colonisation could not be known in advance. 2.3.1. Materials and methods

2.3.1.1. Field site The study site selected for use in this research project was located near Beckom, a village in the central Riverina of New South Wales, Australia (34°13'42.1"S, 146°58'50.5"E) (Figure 2-1.). At the time of collection, the paddock was managed as part of a ley-farming rotation, alternating between biserrula and wheat. According to state-wide soil mapping accessed via the eSPADE platform (Office of Environment and Heritage [OEH] 2020), the site is a moderately fertile chromosol (Australian Soil Classification) or red brown earth (Great Soil Groups). Elevation is approximately 200 m above sea level. Modelling of soil characteristics for the upper 30 cm of the soil profile suggests a component of 30-40%, a silt component of 15-20%, and a sand component of 30-45%. The soil organic carbon concentration was modelled at 0.5- -1 1.5%, exchangeable Na at 2-4%., and total P at 200-300 mg kg . Soil pH (CaCl2) was estimated at 5.0-6.0 and the cation exchange capacity at 15-20 cmolc kg-1. The accuracy of this modelling is considered to be moderate (OEH 2018). Sampling took place in November 2016, at which time a mature biserrula pasture was present in the site (Figure 2-2.). The identification of the pasture species as Biserrula pelecinus cv. Casbah was confirmed by reference to available industry literature (Hackney et al. 2013). Climate data was not available for the site itself but was obtained from the town of West Wyalong, located approximately 40 km NNE. The mean annual rainfall for West Wyalong is 478.7 mm (Bureau of Meteorology 2019), however, in the June and September prior to collection, the town received rainfall of 162.4 and 166.6 mm respectively. The site is low-lying and bounded on its western edge by Kildary Creek. As a result, it was subject to waterlogging prior to collection. This killed or stunted the growth of many biserrula plants (Figure 2-2.). Sampling on this occasion and all subsequent occasions was confined to a small area of the paddock showing the least pronounced impacts of waterlogging (Figure 2-1.)

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 37

Kildary Creek

Figure 2-1. Aerial view of the study site at Beckom, NSW, Australia. The approximate course of Kildary Creek, which marks the western boundary of the paddock, is denoted by the dotted blue line. The yellow rectangle denotes the sampling area. Areas outside of this rectangle were severely affected by waterlogging during the first collection period.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 38 a

b

Figure 2-2. Study site at the time of sampling. Mature, flowering plants of Biserrula pelecinus (a) were present in the pasture; however, large areas of the paddock had been affected by waterlogging (b). Samples were not taken from severely affected areas of the paddock.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 39 2.3.1.2. Plant materials and trap cultures Roots for analysis were obtained from both field samples and trap culture-grown plants. To obtain field samples, 10 soil cores to a depth of 10 cm were taken in a linear transect across the study site. Each core was taken from within the biserrula pasture and at least 1 m away from the nearest non-biserrula plant. These cores were subsequently washed separately through a 72 µm sieve to separate the root fragments from the soil. Roots retained in the sieve were kept for analysis. To obtain trap culture-grown roots, approximately 20 seeds of Biserrula pelecinus var. Casbah were sown into each of ten 18 cm diameter pots containing soil from a pre-existing trap culture and then thinned to four plants per pot after germination. The culture used was known to contain propagules of AM species with high 18s ribosomal DNA sequence similarity to Funneliformis caledonium, F. geosporum, F. fragilistratum, F. mosseae and Rhizophagus irregularis, as well as several unidentified Glomus/Funneliformis strains (Wilson 2010). Plants were grown to maturity for 12 weeks in a glasshouse environment. Subsequently, the soil was washed out and the roots retained for study.

2.3.1.3. Quantification of mycorrhizas Roots obtained from soil cores and trap cultures were cleared and stained for microscopic inspection according to a method adapted from Brundrett (2008). Roots from each of the 10 cores and 10 trap culture-grown plants were divided into sections of approximately 1 cm in length and subsamples weighing 1.5-2 g taken. The roots were then cleared in 10% KOH for 10 min (field-collected roots) or 5 min (trap culture roots) at 121°C before being rinsed for 5 min in deionised water and transferred to the staining solution. Root fragments were stained using a 0.03% solution of Chlorazol black E in a 1:1:1 mixture of lactic acid, glycerol, and deionised water. They were allowed to stain in the solution for 72 h at room temperature and then destained in a solution of 50% glycerol for 1 week at room temperature prior to examination. Percent colonisation (PC) of each root system by AM fungi was assessed by means of the gridline intersection method (Giovannetti and Mosse 1980). Briefly, a subsample of stained roots from each of the 20 samples was transferred to a 90 mm Petri dish marked with a 1.27 cm grid and viewed under a stereomicroscope (Nikon SMZ745T) at 40-80× magnification. Gridlines were scanned vertically and horizontally and roots marked as either mycorrhizal or non-mycorrhizal where they intersected with the gridlines. For each sample, 100 observations were recorded. In cases where the initial

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 40 subsample did not produce 100 intersections, an additional subsample was drawn from the sample. Arbuscules, vesicles, spores, and mycelium were counted individually. In addition to the gridline intersection method, five samples from the field soil and five from the trap cultures were subjected to a further process of quantification. Briefly, 30 root fragments were drawn at random from each of the 10 samples and transferred to two slides at a density of 15 to a slide. These slides were then viewed under a compound microscope (Nikon Eclipse Ni-U) at 400-1000× magnification. For each 30- fragment sample, the percentage of assessed roots possessing mycorrhizas (F%), the intensity of mycorrhizal colonisation in the root system as a whole (M%), and the abundance of arbuscules in the root system as a whole (A%), were determined according to the method of Trouvelot et al. (1986, Institut national de la recherche agronomique [INRA] 2001).

2.3.1.4. Microscopic characterisation of mycorrhizal structures A subsample of 20 mycorrhizal roots was taken from each of the 20 samples employed in the study (10 from field soils and 10 from trap cultures) for further microscopic examination to confirm the presence of mycorrhizal features and to detect the presence of poorly stained or otherwise cryptic AM fungi. Root sections were transferred to slides in groups of 10 and examined at 400-1000× magnification using a compound microscope (Nikon Eclipse Ni-U). Photomicrographs of key mycorrhizal features were captured using a Nikon DS-Fi2 camera.

2.4. Experiment 2: Identification of mycorrhizas in the roots of field- and trap culture-grown biserrula The aim of this experiment was to identify the fungal taxa responsible for the mycorrhizal colonisation observed in biserrula roots obtained from field samples and trap cultures using DNA-based methods.

2.4.1. Materials and methods

2.4.1.1. AM fungal materials Root fragments taken from each of the 20 samples described in section 2.3.1.4., but not included in the subsample for staining, were examined under a dissecting microscope for the presence of AM spores. Where spores were located, they were separated from

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 41 their attached hyphae using fine forceps and transferred to a 90 mm diameter Petri dish for further examination. Where necessary, spores were extracted from sporocarps using forceps and a fine Taklon brush (Princeton Brush Company, Princeton, New Jersey, USA). Spores isolated from field-grown plants displayed little variation in coarse morphology and were assigned to a single morphotype. Spores isolated from plants grown in trap cultures were assigned to one of two morphotypes based principally on variations in size and shape (Figure 2-3.). The field-isolated morphotype could not be distinguished microscopically from morphotype B in the trap culture samples. Many spores were observed to have partly or wholly sloughed their outer wall layers. For this reason, the number of apparent layers in the spore wall was not used to define morphotypes. Multiple representatives of each morphotype were subsequently selected for surface disinfection, DNA extraction, and sequencing.

2.4.1.2. Surface disinfection of spores Spores to be surface disinfected were transferred to microcentrifuge tubes containing 0.5 ml of 80% ethanol and shaken for 30 sec to remove coarse debris, after which the ethanol solution was decanted and the spores allowed to air-dry. Spores were then treated with 0.5 ml of a disinfectant solution of 2% chloramine-T for 20 min, followed by 0.5 ml of an antibiotic solution containing 0.001% gentamicin, 0.002% streptomycin with added of Tween 20 for an additional 20 min. This solution was then decanted and the spores rinsed in five changes of sterile deionised water.

2.4.1.3. DNA extraction DNA extraction was attempted from 18 surface-disinfected AM spores (six from field roots and 12 from trap culture roots) representing each of the observed morphotypes. Spores to be extracted were transferred individually to 0.5 ml microcentrifuge tubes containing 4.0 µl of 5× MyTaq reaction buffer (Bioline, London, UK) and 12 µl of sterile deionised water and crushed using sterilised metal needles to discharge their contents. Tubes were incubated at 94°C for four min to denature DNAse (International Culture Collection of (Vesicular) Arbuscular Mycorrhizal Fungi [INVAM] 2017a). All spores and reagents were kept on ice when not in use.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 42 a b

c

d e

Figure 2-3. Photomicrographs of spores isolated from a biserrula trap culture, including intact (a,b,c) and manually broken (d,e) spores. Two distinct morphotypes were observed (a). Morphotype A (b, d) was irregular in shape and was in most cases the larger of the two (c. 160-240 µm diameter). Morphotype B (c, e) was smaller (c. 100-170 µm diameter) and more globose. Spores of morphotype A were isolated from sporocarps and typically did not possess subtending hyphae. Scale bars represent 200 µm (a) or 100 µm (b-e).

2.4.1.4. Nested PCR and sequencing The extraction solutions for each of the 18 spores were subjected to nested PCR targeting small subunit rRNA using the primers AML1-AML2 and NS31-AM1 (Xiang et al. 2016). For the initial reaction, 2 µl of a 5 µM stock of each of the primers AML1 and AML2 was added to the extraction solutions prior to cycling. Amplification was carried out on a Bio-Rad S1000 Thermal Cycler (Gladesville, NSW, Australia). For the

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 43 subsequent reaction, 1 µl of the first-reaction product was added to a solution containing 4.0 µl of 5× MyTaq reaction buffer, 11 µl of sterile deionised water, and 2 µl of each of the primers NS31 and AM1 prior to cycling. Primers and cycling conditions are described in Table 2-1. Sanger sequencing of PCR products was performed by the Australian Genome Research Facility (Melbourne, Victoria, Australia). Sequences were trimmed to remove poor quality reads using the Chromas trace viewing software (Technelysium, South Brisbane, Queensland, Australia) and aligned using the BioEdit sequence alignment editor (Hall 1999). Identifications based on sequence similarity were obtained by searching the nucleotide collection contained within the GenBank database (Clark et al. 2016) using the Basic Local Alignment Search Tool (BLAST; Altschul et al. 1990) optimised for highly similar sequences.

Table 2-1. Primers and sequencing conditions for nested PCR Primer pair No. cycles Initial Denaturing Annealing Extension Final 94°C 94°C 50°C 72°C 72°C AML1-AML2 30 3 min 60 sec 60 sec 60 sec 10 min 94°C 94°C 58°C 72°C 72°C NS31-AM1 30 3 min 60 sec 60 sec 60 sec 10 min

Forward primer AML1 ATCAACTTTCGATGGTAGGATAGA Reverse primer AML2 GAACCCAAACACTTTGGTTTCC Forward primer NS31 TTGGAGGGCAAGTCTGGTGCC Reverse primer AM1 GTTTCCCGTAAGGCGCCGAA

2.5. Experiment 3: Recruitment of AM fungi to the root environment of biserrula in a pasture environment The aim of this experiment was to characterise the diversity of fungi belonging to the phylum Glomeromycota in the rhizosphere of biserrula grown in pasture and to determine whether these species were selectively recruited to that environment. As AM fungi inhabit plant roots, it was expected that they would be enriched in the rhizosphere relative to the field and bulk soils.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 44

2.5.1. Materials and methods

2.5.1.1. Field site Field sampling was conducted at the study site described in section 2.3.1.1. Sampling time and site conditions were the same as those described in that section.

2.5.1.2. Diversity profiling Six soil cores to a depth of 10 cm were taken from a 200 m linear transect across the field site. Cores were taken within the biserrula pasture at the base of biserrula plants and at least 1 m away from the nearest non-biserrula plant. Three additional soil samples (designated field soil) were taken from at least 1 m away from the nearest biserrula plant, also to a depth of 10 cm. The soil cores were divided to produce bulk soil and rhizosphere soil samples according to the following procedure. First, the six cores were passed through a 500 µm sieve to separate the loose soil (bulk soil) from the root systems. Roots with closely adhering soil were collected from the mesh, transferred to glass bottles containing sterile deionised water, and shaken for 10 min to detach the soil from the root surface. This solution was passed again through the 500 µm sieve to separate the roots from the soil and the soil solution retained. Aliquots of this solution were then transferred to centrifuge tubes and centrifuged at 5000 RPM for 3 minutes. The excess water was decanted and the rhizosphere soil retained. To ensure consistency between treatments, as it was uncertain whether this process resulted in the loss of fungal material from the sample, an equivalent rinsing and centrifugation procedure was applied to bulk and field soil samples. Subsamples weighing 0.25 g were taken from each of the 15 samples (6 rhizosphere, 6 bulk soil, 3 field soil) and used for extraction of total fungal DNA. Extractions were carried out using the DNeasy PowerLyzer PowerSoil Kit (QIAGEN, Chadstone, Victoria, Australia) according to the manufacturer’s instructions. DNA was quantified using a NanoDrop (Thermo Fisher Scientific, Scoresby, Victoria, Australia) and each sample diluted to approximately 10 ng DNA µl-1 prior to sequencing.

2.5.1.3. Sequencing conditions PCR amplification and sequencing was performed at the Australian Genome Research Facility (Melbourne, Victoria, Australia), targeting the internal transcribed spacer region

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 45 (ITS1-ITS2) using the primers and conditions outlined in Table 2-2. Thermocycling was conducted with an Applied Biosystem 384 Veriti and using AmpliTaq Gold 360 (Life Technologies, Australia). Illumina indexing of the amplicons was achieved in a second PCR utilising TaKaRa Taq DNA Polymerase (Takara Bio, Mountain View, CA, USA). Indexed amplicon libraries were quantified by fluorometry (QuantiFluor, Promega, Fitchburg, WI, USA) and normalised. An equimolar pool was created and adjusted to 5nM for sequencing on an Illumina MiSeq (San Diego, CA, USA) with a V3, 600 cycle kit (2 x 300 base pairs paired-end). Paired-ends reads were assembled by aligning the forward and reverse reads using PEAR (version 0.9.5) (Zhang et al. 2014). Primer sequences were removed, and trimmed sequences were processed using Quantitative Insights into Microbial (QIIME 1.8) (Caporaso et al. 2010), USEARCH (version 7.1.1090) (Edgar et al. 2011, Edgar 2010), and UPARSE (Edgar 2013) software. Within USEARCH, sequences were quality filtered, full length duplicate sequences removed, and reads sorted by abundance. Singletons or unique reads in the data set were discarded. ITS sequences were clustered prior to chimera filtering using the Unite database (Kõljalg et al. 2005) as reference. Reads were mapped back to OTUs with a minimum identity of 97% to obtain the number of reads per OTUS. was assigned by Qiime using the Unite database. Unless mentioned, default settings were used for all procedures.

Table 2-2. Primers and sequencing conditions for ITS diversity profiling Target No. cycles Initial Denaturing Annealing Extension Final 95°C 94°C 55°C 72°C 72°C ITS1F-ITS2 35 7 min 30 sec 45 sec 60 sec 7 min

Forward primer ITS1F CTTGGTCATTTAGAGGAAGTAA Reverse primer ITS2 GCTGCGTTCTTCATCGATGC

2.5.1.4. Statistical analysis Statistical analysis of abundance data was carried out using version 3.6.0 of the R statistical software program (R Core Team 2019). Measurements of alpha diversity were carried out on unrarefied sequence data, owing to the known tendency of rarefaction to

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 46 introduce additional bias (Willis 2019, McMurdie and Holmes 2014). Species richness (S) values were obtained by counting the total number of OTUs represented in each sample. Shannon-Weiner index (H) values were obtained using the diversity function in the package Vegan (Oksanen et al. 2019). Effective species numbers were derived for each sample by taking the exponential of the Shannon-Weiner index value (ExpH). Diversity indices were calculated for both total soil fungi and the Glomeromycota alone. To identify taxa selectively recruited to each soil compartment, sequence data were centred log-ratio (CLR) transformed (Gloor et al. 2017) and subjected to ANOVA-Like Differential Expression analysis (ALDEx) using the ALDEx2 package (Gloor et al. 2016, Fernandes et al. 2014, Fernandes et al. 2013). Two conditions were compared at a time and the results of Welch’s t-test used to identify differences in taxon abundance. To ensure a robust estimation of effect size, 1000 Monte-Carlo iterations were used to generate all aldex.clr objects. Benjamini-Hochberg corrected p values (q) were calculated to take into account the effect of multiple comparisons. In line with recommendations (Gloor et al. 2017), effects were considered significant at q < 1.

2.6. Results

2.6.1. Experiment 1: Quantification of mycorrhizas in the roots of field- and trap culture-grown biserrula Evidence of mycorrhization was observed for 8 of 10 field samples and 9 of 10 trap culture samples subjected to the gridline intersection method. Two field samples and one trap culture-grown sample lacked any evidence of mycorrhization; one additional field sample and one trap culture sample lacked arbuscules but displayed other AM structures. Differences in percent colonisation between field and trap culture samples were contingent on the standard of evidence accepted for mycorrhization, suggesting that this decision may bias assessments of colonisation. When assessed by the presence of any AM component (spores, vesicles, arbuscules, or mycelium; Figure 2-4.), field and trap culture roots returned mean percent colonisation of 19.2 ± 4.6 and 23.5 ± 3.5 respectively. When assessed by arbuscules alone, values were more divergent: 4.8 ± 1.3 PC for field roots and 12.2 ± 2.4 PC for trap culture-grown samples. Among mycorrhizal field samples, percent colonisation ranged from 4.0 to 42.0. Among mycorrhizal trap culture samples, percent colonisation ranged from 10.0 to 37.0. No

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 47 cryptic or non-staining mycorrhizal features were observed under a compound microscope. When analysed according to the method of Trouvelot et al. (1986), field-grown and pot-grown plants returned similar values for the frequency of mycorrhizal colonisation (F%) but strongly dissimilar values for the abundance of arbuscules in the root system (A%) (Table 2-3.). The lower values for arbuscule abundance obtained by this method relative to the gridline intersection reflect the weighting scale established by Trouvelot et al. (1986). Arbuscules were observed in 3.67% of field-grown root fragments and 10.0% of pot-grown root fragments, roughly in line with the values obtained from the gridline intersection method. In spite of a similar colonisation frequency between the sample types, colonisation intensity (M%) was greater in pot-grown samples. It should be noted that root fragments were drawn from the entire root system of pot-grown samples but only from the top 10 cm of soil in field samples. Consequently, comparisons between sample types should be considered indicative and not definitive.

Table 2-3. Frequency and intensity of mycorrhization in field- and pot-grown biserrula Sample type Parameter Range Mean ± s.e. Field F% 6.67 – 36.67 21.33 ± 5.01 M% 0.73 – 4.87 3.04 ± 0.71 A% 0.01 – 0.37 0.17 ± 0.07 Pot F% 6.67 – 40.0 20.0 ± 5.58 M% 2.0 – 11.9 6.69 ± 1.76 A% 1.50 – 7.98 4.46 ± 1.40 Frequency and intensity of mycorrhization according to the method of Trouvelot et al. (1986), where F% is the percentage of assessed roots possessing mycorrhizas, M% is the intensity of mycorrhizal colonisation in the root system as a whole, and A% is the abundance of arbuscules in the root system as a whole. Values are ranges and means (± s.e.) of five samples for each sample type.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 48

H A

A A H H V A A

H

H

V A V A A

Figure 2-4. Photomicrographs of crushed biserrula root sections displaying mycorrhizal structures. A = arbuscule (Arum type). V = vesicle. H = longitudinal hypha.

2.6.2. Experiment 2: Identification of mycorrhizas in the roots of field- and trap culture-grown biserrula Sequences for small subunit RNA obtained from AM spores attached to root-associated mycelium derived from field and trap cultures were between 424 and 484 bp in length after trimming. Spores represented both of the morphotypes depicted in Figure 2-3. All

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 49 sequences, irrespective of origin, were found to be identical where they overlapped. A BLAST search performed on a consensus sequence returned 100% matches to Funneliformis mosseae (Table 2-4.). Direct comparison of this sequence to sequences belonging to species reported to be contained in the initial trap culture returned values between 94.18% (for Rhizophagus irregularis; 98% query coverage) and 99.21% (for F. geosporum; 99% query coverage). Consequently, it was assumed that, while the trap culture was known to contain multiple AM taxa, all spores obtained in this study belonged to F. mosseae.

Table 2-4. BLAST matches for spore DNA Total Query Percent Accession Description E value score coverage identity MN726657.1 Funneliformis mosseae 708 100% 0.0 100.00% isolate GB13043 small subunit ribosomal RNA gene, partial sequence MN726663.1 Funneliformis mosseae 708 100% 0.0 100.00% isolate GN17023 small subunit ribosomal RNA gene, partial sequence MN726662.1 Funneliformis mosseae 708 100% 0.0 100.00% isolate CB15132 small subunit ribosomal RNA gene, partial sequence KU136426.1a Funneliformis caledonium 682 99% 0.0 98.95% isolate E8_P2 18S ribosomal RNA gene, partial sequence KU136432a Funneliformis geosporum 688 99% 0.0 99.21% isolate H3_P2 18S ribosomal RNA gene, partial sequence AJ276085.2a Glomus fragilistratum 676 99% 0.0 98.69%

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 50 Table 2-4. BLAST matches for spore DNA Total Query Percent Accession Description E value score coverage identity partial 18S rRNA gene, clone pWD114-3-3 MH571752.1a Rhizophagus irregularis 573 98% 2e-163 94.18% strain IR27 small subunit ribosomal RNA gene, partial sequence BLAST matches for a consensus DNA sequence obtained from spores isolated from the roots of biserrula plants. The superscript (a) denotes a species reported to be contained in the initial trap culture but assumed to be absent from the sequenced DNA.

2.6.3. Experiment 3: Recruitment of AM fungi to the root environment of biserrula in a pasture environment Diversity profiling returned a total of 800,140 ITS sequences, corresponding to 1,707 operational taxonomic units (OTUs). Sequence counts per sample ranged from 48,316 to 65,660 for rhizosphere samples, 45,358 to 66,291 for bulk soil samples, and 44,771 to 49,145 for field soil samples. OTU counts per sample ranged from 441 to 665 for rhizosphere samples, 576 to 699 for bulk soil samples, and 505 to 786 for field soil samples. OTUs were assigned to six phyla, 28 classes, 83 orders, 150 families, and 240 genera. 328 OTUs were identified to species level. The most abundant phylum by both sequence and OTU count was the (Figure 2-5.), accounting for 80.11% of all sequences and 47.22% of all recorded OTUs. It was followed by the (8.98% of sequences, 26.01% of OTUs) and (4.82% of sequences, 4.22% of OTUs). These were the most abundant phyla across all samples regardless of origin, and only minor variation in their relative abundance was observed between soil compartments. Fungi that could not be assigned to a phylum accounted for 4.66% of all sequences and 16.75% of all OTUs. The three most abundant taxa, all members of the Ascomycota, together accounted for more than 30% of all sequences. These were Plectosphaerella cucumerina (13.98%), Talaromyces sp. (9.30%), and oxysporum (8.03%). Diversity profiling also returned 167 sequences belonging to 11 members of the non-fungal phylum Cercozoa.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 51 Restricting the dataset to members of the phylum Glomeromycota yielded a total of 2,135 sequences belonging to 22 OTUs, representing 0.27% of all ITS sequences and 1.29% of all OTUs returned. All four recognised orders of the Glomeromycota (Redecker et al. 2013) were present in the data. The greatest diversity was observed in the Archaeosporales (12 OTUs), followed by the (7 OTUs), (2 OTUs), and Paraglomerales (1 OTU). Only three OTUs were identified to species level: two strains of F. mosseae and one of Rhizophagus invermaius. Total sequence counts for each OTU ranged from 7 to 621 but most were present only in low concentrations: 17 of the 22 taxa were recorded fewer than 100 times and 13 were recorded fewer than 50 times. The most abundant taxon was a strain of F. mosseae, accounting for 29.07% of AM sequences, followed by four species identified to the genus Archaeospora, accounting for 16.25%, 11.43%, 7.87%, and 6.37% respectively (Table 2-5.). All 22 OTUs were present in both the rhizosphere and bulk soils whereas only eight were present in the field soil. Despite this, ALDEx analysis did not identify any differentially enriched taxa between soil compartments, likely due to low total sequence numbers for all taxa and high between-replicate heterogeneity of the samples.

Table 2-5. Relative abundance of common AM fungi in rhizosphere, bulk, and field soils Sequence counts Consensus ID Rhizosphere Bulk soil Field soil Total (mean ± s.e.) (mean ± s.e.) (mean ± s.e.) Funneliformis mosseae 621 54.2 ± 26.89 48.3 ± 24.01 2.0 ± 1.62 Archaeospora sp. 341 50.8 ± 42.27 7.0 ± 5.91 - Archaeospora sp. 244 4.3 ± 2.4 36.3 ± 19.0 - Archaeospora sp. 168 9.8 ± 4.87 18.0 ± 8.77 0.3 ± 0.40 Archaeospora sp. 136 8.50 ± 4.97 13.50 ± 7.81 1.33 ± 1.26 Fungal ITS sequence counts and mean of total counts per sample (± s.e.) for each soil compartment for the most abundant taxa of the Glomeromycota identified in the diversity profiling exercise.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 52 Figure 2-5. Proportional composition of the ITS diversity profiling dataset by phylum, as assessed by (a) total ITS sequence count and (b) total count of operational taxonomic units.

Diversity index values were calculated for all soil compartments (Table 2-6.). Indices for the rhizosphere treatment were affected by low sequence counts in a single sample. This sample returned only a single sequence belonging to an AM fungus, whereas the remaining five samples returned between 85 and 381 sequences belonging to between 10 and 17 species. A sample in the field soil treatment failed to return any AM sequences, but the small sample size of three, as opposed to the six samples in each of the rhizosphere and bulk soil treatments, and the consequent high variability of the sequence and species counts impede interpretation. The remaining two field soil samples returned abundances of 3 and 42 and species counts of 2 and 10 respectively. In both cases, the lower abundances of AM fungi corresponded to lower sequence counts for total soil fungi. These samples have not been omitted from the data as the cause of this variation is unknown and it is unclear whether such an omission is justified. Consequently, the variation around the mean introduced by these samples complicates certain comparisons of the rhizosphere and field soils. By contrast, however, clear

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 53 distinctions can be made between the bulk soil and field soil. Though similar to the field soil in total fungal diversity, bulk soil samples were markedly more diverse than field soil samples in representatives of the Glomeromycota (ExpH 7.95 ± 1.16 > 3.72 ± 2.29). Rhizosphere samples appear less diverse in total fungi than bulk and field soil samples when measured by effective species number (ExpH 17.80 ± 3.19 < 25.44 ± 1.13) but are similar to the bulk soil in AM diversity (ExpH 6.46 ± 1.41 for rhizosphere samples, compared to 7.95 ± 1.16 for bulk soil samples), pointing to a role for selective recruitment in shaping the rhizosphere mycota. Despite this evidence for the effect of selective recruitment, AM fungi were not enriched in the rhizosphere relative to the bulk soil, contrary to expectations. The similar AM values for rhizosphere and bulk soils suggest that the sampling method did not clearly distinguish between these two compartments.

2.7. Discussion

2.7.1. Quantification of mycorrhizas As anticipated, biserrula was found to harbour AM associations. Formation of mycorrhizas was observed in 80% of field-grown samples and 90% of pot-grown samples exposed to trap cultures, using the gridline intersection method. Approximately one-fifth of all intersections showed some evidence of mycorrhization, regardless of the source of the sample, with slightly greater variation between root systems observed for field samples. Giovannetti and Mosse (1980) identify colonisation rates below 20% as “low”, however, comparisons with percent colonisation values reported for other host plants are complicated by methodological inconsistencies, including the length of the incubation period. Among N2-fixing herbaceous plants, reported percent colonisation values vary from 15.0 to 96.0, while restricting this selection to plants colonised by a species of Funneliformis returns values between 18.6 and 64.7 (Treseder 2013). On the basis of the present study, biserrula appears to demonstrate a diminished tendency to form mycorrhizas. It is comparable to the value reported (24%) for sun-exposed Trifolium pratense (McGee 1990), however, in that study, plants were grown for only six weeks and not 12, as in the present case. The PC and F% values reported here are likewise comparable to those reported for plants of Vigna luteola grown for only 21 days (18.6%; Hernández et al. 2000). It is noteworthy that shaded plants of T. pratense yielded a higher value of 40% root colonisation (McGee 1990). All biserrula plants

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 54 assessed in the present study, whether collected from the study site or grown in an unshaded glasshouse, can be considered sun-exposed. Consequently, greater colonisation may be observed in different growing conditions. Observed colonisation was similar in both field and trap culture samples when assessed by the presence of any fungal component but differed considerably when arbuscules alone were accepted as evidence of mycorrhization. In the gridline intersection data, arbuscules were recorded from only 46 of the 1000 assessed intersections in the field samples, compared to 122 from 1000 in the trap culture samples. This phenomenon was repeated in the Trouvelot data. Field samples were instead dominated by intraradical hyphae, while vesicles and spores, like arbuscules, were encountered only sporadically. The practice of assessing colonisation by the presence or absence of arbuscules has clear advantages, as it points to the ongoing exchange of nutrients between the symbiotic partners (Brundrett 2009), but is complicated by the fact that arbuscules are transient structures that may persist for only a few days (Hartmann et al. 2019). Consequently, they may be absent from older and field-collected roots, or, as here, present but at lower densities. The use of any AM component to evaluate mycorrhization may yield a truer assessment of percent colonisation but introduces an additional complication: hyphae present in the roots may be non-infective saprophytes rather than mutualists (Brundrett 2009). The use of trap culture samples alongside field-collected samples in the present case offers reassurance that biserrula does form active mycorrhizas, though it remains unclear whether the exchange of nutrients persists in pasture. As a consequence of this difficulty, along with a tendency for many methods of quantification to overestimate colonisation (Giovannetti and Mosse 1980), reported percent colonisation values should be considered approximations.

Table 2-6. Diversity index values for ITS diversity profiles by soil compartment All fungi AM fungi Soil compartment Index Range Mean (± s.e.) Range Mean (± s.e.) Rhizosphere S 441 – 665 540.33 ± 35.07 1 – 17 11.33 ± 2.29 H 1.98 – 3.44 2.79 ± 0.19 0 – 2.38 1.64 ± 0.35 ExpH 7.35 – 31.25 17.80 ± 3.19 1.0 – 10.79 6.46 ± 1.41 Bulk soil S 576 – 699 647.17 ± 17.03 10 – 23 15.17 ± 1.76

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 55 Table 2-6. Diversity index values for ITS diversity profiles by soil compartment All fungi AM fungi Soil compartment Index Range Mean (± s.e.) Range Mean (± s.e.) H 3.01 – 3.33 3.23 ± 0.05 1.73 – 2.54 2.03 ± 0.13 ExpH 20.31 – 28.02 25.44 ± 1.13 5.62 – 12.68 7.95 ± 1.16 Field soil S 505 – 786 626.67 ± 83.28 0 – 10 4.0 ± 3.06 H 3.06 – 3.53 3.28 ± 0.14 0 – 2.11 0.9 ± 0.63 ExpH 21.34 – 31.42 26.98 ± 3.77 1.0 – 8.26 3.72 ± 2.29 Values for species richness (S), Shannon-Weiner index (H), and effective species number (ExpH) calculated for all soil compartments, showing ranges and means ± standard error. Values were calculated separately for all fungi and for AM fungi of the Glomeromycota alone.

2.7.2. Identification of mycorrhizas All spores isolated from field and trap culture samples were identified as belonging to F. mosseae. Additional cryptic species may have been present but not accounted for if they failed to produce spores or did not stain clearly. Microscopic examination revealed densely branched arbuscules of the Arum type (Figure 2-4.), extending from longitudinal hyphae with few observable coils (Karandashov et al. 2004). This appears to be a function of host plant: F. mosseae has previously been reported to produce Arum-type arbuscules in a legume host, Trifolium subterraneum (Dickson 2004), but Paris-type arbuscules in the non-legume species Linum usitatissimum (Rodrigues and Rodrigues 2015). Arbuscules observed in the present study showed little variation in coarse morphology, adding support to the suggestion that only a single species was present. Again, however, this may be a function of the host plant and not the fungal partner or partners. The picture is complicated by the diversity profiling data. While F. mosseae returned the highest relative abundance in the dataset, three other common taxa, all members of the genus Archaeospora, returned similar patterns of distribution. Analysis of the diversity profiling data did not identify differentially abundant taxa in this case and could not distinguish between active mycorrhizas and residual material in the soil; however, the scarcity (or, in two cases, absence) of Archaeospora sequences in the field soil suggests that an active association with the roots of biserrula is likely. Species of Archaeospora are known from legume hosts (Shi et al. 2012) and have previously been

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 56 observed as a co-dominant or secondary mycorrhiza with F. mosseae in the roots of the Australian legumes Lotus australis and L. pedunculatus (Tibbett et al. 2008), suggesting that host incompatibility and intraspecific competition are unlikely to account for the absence of Archaeospora species from microscopic observations. That these taxa were not observed in the processes of identification and quantification may point to a limitation of the methodologies in use and not to a genuine absence. Arbuscules of Archaeospora spp. are reported to stain weakly in conventional stains and vesicles have not been observed in pot cultures (INVAM 2017b). Additionally, their spores are typically smaller than those of F. mosseae and are translucent (INVAM 2017b), potentially impeding observation. Disagreements between molecular and microscopic methods in the quantification of Glomus and Archaeospora mycorrhizas have been previously reported (Shi et al. 2012). In all, 22 fungi of the phylum Glomeromycota were present in both the bulk and rhizosphere soils while only eight were recorded from the field soil. While sequence counts for many of these taxa were too low to confidently rule out the effect of random noise, this pattern of distribution raises the possibility that many more AM fungi were active in the biserrula root than could be isolated as spores and subsequently sequenced.

2.8. Conclusion This study presents the first direct observation of mycorrhizal colonisation in biserrula. The experiments outlined here demonstrate that biserrula is capable of forming mycorrhizas in both field and laboratory conditions, while highlighting some limitations inherent in the methods of quantification that are commonly used to assess mycorrhization. These results, derived as they were from a single site with a single land-use history, and at a single sampling time, require additional validation, which is beyond the scope of the present study. Additionally, it must be noted that the dominance of F. mosseae in the examined roots does not preclude the possibility that other AM taxa may, under other circumstances, form colonies in a biserrula host. A more systematic screening method, inoculating biserrula plants with axenic cultures of known AM species, would be needed to determine the range of fungi capable of colonising this host species. While the trap culture employed in this study was known to contain multiple AM species, the success of F. mosseae may reflect factors other than host preference, including the aggressiveness with which it colonises the root or its proportional representation among the propagules in the culture. Alternatively, it may

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 57 be an artefact of the screening methodology. Nevertheless, by confirming that biserrula is mycorrhizal, this study forms a basis for future efforts to fine-tune its microbial interactions with the aim of facilitating improvements in nutrient uptake and plant growth. The low values for percent colonisation reported here, together with the low sequence counts of AM species in field conditions, suggest that biserrula represents a prime candidate for intervention of this kind. With this in mind, experiments were designed to determine whether bacteria associated with the roots of biserrula were capable of acting as mycorrhiza helper bacteria, enhancing the germination and growth of AM fungi in vitro and the colonisation of biserrula roots in vivo. Following on from this work, additional experiments were designed to determine whether this helper effect persists in a populated rhizosphere.

Chapter 2. Observation and quantification of mycorrhizas in the roots of biserrula 58 Chapter 3. Identification of mycorrhiza helper bacteria in the rhizosphere of Biserrula pelecinus L.

3.1. Introduction Mycorrhiza or mycorrhization helper bacteria (MHB) are bacterial strains known to promote the formation or persistence of mycorrhizas in the roots of a host plant (Garbaye 1994). Consequently, they may act to amplify the beneficial effects of the mycorrhizal mutualism, which range from enhancements in host nutrition to protection against biotic and abiotic stressors (Jung et al. 2012). Beginning with Mosse (1959, 1962), who first observed the helper effect in a Pseudomonas strain, a large body of evidence has accumulated in support of the proposition that certain soil bacteria can positively influence both arbuscular mycorrhizal (AM) and ectomycorrhizal symbioses. It is now understood that AM fungi are embedded in complex webs of bacterial-fungal and bacterial-bacterial interactions, which may confer beneficial effects on the symbiotic partners and, subsequently, on their host plant (Turrini et al. 2018). Evidence for this helper effect, reviewed most recently by Frey-Klett et al. (2007), is derived from in vitro and in vivo experimentation, which report increases in the pre-symbiotic growth of the fungus or in the rate of mycorrhizal colonisation in response to bacterial co- inoculation. A single account also exists of a helper strain (Paenibacillus validus) prolonging the life of an AM fungus (Rhizophagus irregularis) and permitting it to sporulate in culture, despite AM fungi typically being regarded as obligate symbionts (Hildebrandt et al. 2002). Furthermore, helper bacteria may promote spore germination, although evidence for this is inconsistent, with both positive (Sundram et al. 2011, Xavier and Germida 2003, Mosse 1959) and neutral (Fernández Bidondo et al. 2011, Barea et al. 1998, Azcón 1987) results recorded. The mechanism or mechanisms by which MHB strains exert their effects are similarly unclear, though many possibilities have been proposed, including the stimulation of plant signalling pathways (Xie et al. 1995), the secretion of plant hormones such as indole-3-acetic acid (Rudawska and Gay 1995), enzymatic activity (Mosse 1962), and direct secretion of protein effectors

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 59 (Viollet et al. 2017). Given that helpers are non-monophyletic, and have been identified among both Gram-negative (e.g. Pseudomonas) and Gram-positive (e.g. Paenibacillus) taxa, it seems unlikely that any singular mode of activity will account for all observed helper effects. Recent efforts to screen large numbers of cultured soil bacteria have led to the identification of helpers of potential agronomic significance (see, for example, Fernández Bidondo et al. 2016). It is not clear, however, to what extent the action of these helper strains is host specific. Arthurson et al. (2011) reported, for example, that the candidate MHB strain Paenibacillus brasilensis PB177, isolated from maize plants, failed to enhance mycorrhization in wheat, unlike other Paenibacillus species isolated from wheat plants. For this reason it may be necessary to screen each potential host plant for compatible isolates. Furthermore, to be of use in an agricultural setting, a candidate MHB must be able to colonise and persist in the rhizosphere of the target plant. Given that different host plants are known to support radically different rhizosphere microbiomes (Bakker et al. 2013, Dennis et al. 2010), the selection of suitable strains depends upon an understanding of which bacterial taxa are selectively recruited by the host root. This is partly accomplished by screening bacteria isolated directly from the rhizosphere environment, or even within plant tissues, but may be complemented by the use of genetic diversity profiling methods to determine the extent of the recruitment effect. The beneficial effects of the mycorrhizal symbiosis are likely to be greatest in soil deficient in nutrients and in conditions in which the host plant is subjected to the most pronounced biotic and abiotic stresses. The greatest benefit of the implementation of MHB strains is likely to be seen in plant taxa that are not naturally prolific in forming mycorrhizal associations. The pasture legume Biserrula pelecinus L. (biserrula) satisfies both of these criteria. Biserrula was selected for its adaptation to acid soils and low- rainfall environments (Loi et al. 2014); its use in Australian dryland agriculture additionally subjects it to conditions of nutrient limitation, salinity, sodicity, and other stressors, which are ubiquitous in Australia (Carberry et al. 2011, Rovira 1992). Mycorrhizas may consequently provide the resilience the species needs to reach its maximum potential; however, observed rates of mycorrhizal colonisation in biserrula plants in pasture (~20%; section 2.6.1.) suggest that there is considerable room for enhancement through the use of co-inoculated helper strains. In this study, we aim to identify candidate MHBs by subjecting cultivable rhizosphere and root-endophytic bacteria to in vitro and in vivo screening using the AM

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 60 fungus Funneliformis mosseae as a model species. Additionally, we aim to determine whether these bacteria are selectively recruited to the biserrula rhizosphere in a pasture setting and whether they are capable of facilitating the in vitro formation of secondary spores in F. mosseae. We also provide some functional characterisation of selected strains in terms of traits that may be relevant to the tripartite (plant-fungus-bacteria) mutualism.

3.2. Materials and methods

3.2.1. Field site and soil sampling Sampling took place in November 2016 and under the same conditions as previously described (section 2.3.1.1). Sampling took place at the same time and under the same conditions as previously described. Six soil cores were taken, each to a depth of 10 cm, in a linear transect across the field site. Cores were taken within the biserrula pasture and at least 1 m away from the nearest non-biserrula plant. These cores were used both for diversity profiling and to obtain cultivable bacteria for screening. Field materials were used for this purpose with a view to obtaining in culture bacterial taxa known from metabarcoding data to form close associations with biserrula. Three additional field soil samples were taken from at least 1 m away from the nearest biserrula plant, also to a depth of 10 cm.

3.2.2. Diversity profiling The soil cores were divided to produce bulk soil and rhizosphere soil samples according to the procedure described in section 2.5.1.2. The same samples used in that diversity profiling exercise were used again here. Subsamples weighing 0.25 g were taken from each of the 15 samples (6 rhizosphere, 6 bulk soil, 3 field soil) and used for extraction of total bacterial DNA. Extractions were carried out using the DNeasy PowerLyzer PowerSoil Kit (QIAGEN, Chadstone, Victoria, Australia) according to the manufacturer’s instructions. DNA was quantified using a NanoDrop (Thermo Fisher Scientific, Scoresby, Victoria, Australia) and each sample diluted to approximately 10 ng DNA µl-1 prior to sequencing.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 61 3.2.3. Sequencing conditions PCR amplification and sequencing was performed at the Australian Genome Research Facility (Melbourne, Victoria, Australia), targeting the 16S rRNA gene (V1 to V3) using the primers and conditions outlined in Table 3-1. The V1-V3 region was preferred over the V3-V4 region owing to experimental evidence indicating that a greater diversity of microorganisms is obtained by this approach (Zheng et al. 2015). Thermocycling was conducted with an Applied Biosystem 384 Veriti and using AmpliTaq Gold 360 (Life Technologies, Australia). Illumina indexing of the amplicons was achieved in a second PCR utilising TaKaRa Taq DNA Polymerase (Takara Bio, Mountain View, CA, USA). Indexed amplicon libraries were quantified by fluorometry (QuantiFluor, Promega, Fitchburg, WI, USA) and normalised. An equimolar pool was created and adjusted to 5nM for sequencing on an Illumina MiSeq (San Diego, CA, USA) with a V3, 600 cycle kit (2 x 300 base pairs paired-end). Paired-ends reads were assembled by aligning the forward and reverse reads using PEAR (version 0.9.5) (Zhang et al. 2014). Primer sequences were removed, and trimmed sequences were processed using Quantitative Insights into Microbial Ecology (QIIME 1.8) (Caporaso et al. 2010), USEARCH (version 7.1.1090) (Edgar et al. 2011, Edgar 2010), and UPARSE (Edgar 2013) software. Within USEARCH, sequences were quality filtered, full length duplicate sequences removed, and reads sorted by abundance. Singletons or unique reads in the data set were discarded. Sequences were clustered followed by chimera filtering using the “rdp_gold” database as the reference. To obtain the number of reads in each OTU, reads were mapped back to OTUs with a minimum identity of 97%. Taxonomy was assigned by QIIME using the Greengenes database (DeSantis et al. 2006). Unless mentioned, default settings were used for all procedures

Table 3-1. Diversity profiling primers and sequencing conditions Target No. cycles Initial Denaturing Annealing Extension Final 95°C 94°C 50°C 72°C 72°C 16S V1-V3 29 7 min 45 sec 60 sec 60 sec 7 min

Forward primer 27F AGAGTTTGATCMTGGCTCAG Reverse primer 519R GWATTACCGCGGCKGCTG

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 62 3.2.4. AM fungi and bacterial cultures Bacteria were isolated from the rhizosphere, rhizoplane, root endosphere, and nodules of biserrula. To obtain rhizosphere cultures, a subsample weighing 5 g of the rhizosphere soil obtained by sieving was transferred to a sterile glass container and immersed in sterile deionised water to a final volume of 0.5 L. To obtain rhizoplane cultures, a similar solution was produced using 5 g of soil-free roots retained in the sieve during the production of the rhizosphere solution. To obtain root endosphere cultures, 2 g of soil-free roots were surface disinfected in a 30% solution of commercial bleach (1.3% NaOCl) for 10 min, rinsed three times with sterile H2O, immersed in 200 ml of sterile H2O, and crushed with a sterilised pestle. Nodules were removed from these roots after disinfection and rinsing and transferred to a separate solution, where they were crushed in the same manner. All solutions obtained by these methods were plated onto nutrient agar (Thermo Fisher Scientific, Scoresby, Victoria, Australia) and tryptic soy agar (Merck, Bayswater, Victoria, Australia) using a 10-fold dilution series of 101-1010. Non-specific media were utilised to obtain the widest possible range of taxa for future investigation. Individual cultures were transferred to new 90 mm Petri dishes and re-streaked until pure cultures were obtained. Spores of AM fungi were isolated from a mycorrhizal trap culture utilising biserrula as a host plant. The culture was maintained for 4 generations, with each generation consisting of a 3-month growth period followed by approximately 1 month of drying. The initial soil was inherited from a pre-existing trap culture reported to contain propagules of the AM species Funneliformis caledonium, F. geosporum, F. fragilistratum, F. mosseae and Rhizophagus irregularis, as well as several unidentified Glomus/Funneliformis taxa (Wilson 2010). The rate of sporulation in each of the first three generations was inadequate for experimental purposes. Spores were isolated by a method of wet sieving and decanting adapted from Pacioni (1992). Briefly, 100 g of soil was added to 1 L of water, manually shaken, allowed to rest for several seconds in order for the heavier sediment to settle, and finally poured through nested 1 mm, 250 µm, and 72 µm sieves, aided by a jet of water. Material collected from each sieve was transferred to 90 mm Petri dishes for inspection. Spores were separated from any attached hyphae using fine forceps. Where necessary, spores were extracted from sporocarps using forceps and a fine Taklon brush (Princeton Brush Company, Princeton, New Jersey, USA). Subtending hyphae were trimmed so that no more than 30 µm remained (see Figure 3-6.). Isolated spores were assigned to morphotypes, which corresponded to the

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 63 two morphotypes previously identified in biserrula trap cultures (see Figure 2-3.), and transferred by pipette to 1.5 ml collection tubes.

3.2.5. Surface disinfection of spores Once isolated from the soil, groups of approximately 50 AM spores were transferred to tubes containing 0.5 ml of 80% ethanol and surface disinfected according to the method described in section 2.4.1.2. When not in use, spores were stored in tubes containing 1 ml of sterile water at 4°C.

3.2.6. Identification of AM fungi and bacteria DNA was extracted from bacterial cultures grown for 48 h in nutrient broth (Thermo Fisher Scientific, Scoresby, Victoria, Australia) using the Sigma GenElute Bacterial Genomic DNA Kit (Sigma-Aldrich Pty Ltd, Castle Hill, NSW, Australia) according to the manufacturer’s instructions. Isolated DNA was subjected to 16S PCR using the primers 8F-1492R (Table 3-2.). For each reaction, 2 µl of DNA template was combined with 4 µl of 5× MyTaq reaction buffer (Bioline, London, UK), 10 µl of sterile deionised water, and 2 µl of a 5 µM stock of each primer, for a final volume of 20 µl. Following initial denaturation at 95°C for 2 min, samples were subjected to 30 cycles of the following: 95°C for 30 sec, 56°C for 30 sec, and 72°C for 80 sec. This was followed by a final extension step at 72°C for 5 min. Amplification was carried out on a Bio-Rad S1000 Thermal Cycler (Gladesville, NSW, Australia). Subsequently, bacterial isolates found to have a positive effect on in vivo AM colonisation were subjected to a second round of PCR, using primers specific to each genus (Table 3-3.). Following identification, accessions were lodged in the Biosecurity Collections at the New South Wales Department of Primary Industries (Orange, NSW, Australia). Representatives of each AM spore morphotype were selected for DNA extraction. In total, DNA was obtained from 6 surface-disinfected AM spores according to the following method. Spores were transferred individually to 0.5 ml microcentrifuge tubes containing 4.0 µl of 5× MyTaq reaction buffer (Bioline, London, UK) and 12 µl of sterile H2O and crushed using sterilised metal needles to discharge their contents. Tubes were then heated to 94°C for four minutes to denature DNAse (International Culture Collection of (Vesicular) Arbuscular Mycorrhizal Fungi [INVAM] 2017a). All spores and reagents were kept on ice when not in use. The resulting solution was subjected to nested PCR targeting the small subunit rRNA gene using the primers AML1- AML2 and NS31-AM1 under the reaction conditions described by Xiang et al. (2016).

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 64 For each reaction, 2 µl of a 5µM stock of each primer was added to the extraction solution. For the second reaction, 1 µl of the initial PCR product was used. Sanger sequencing of PCR products was performed by the Australian Genome Research Facility (Melbourne, Victoria, Australia). Sequence processing was conducted according to the method described in section 2.4.1.4. A list of 16S identifications of bacterial strains obtained by these methods is given in Table 3-4. All fungal spores yielded identical sequences and were identified as Funneliformis mosseae (BLAST matches as previously listed in Table 2-3).

Table 3-2. 16S primers and sequencing conditions Target No. cycles Initial Denaturing Annealing Extension Final 95°C 95°C 56°C 72°C 72°C 16S V1-V3 30 2 min 30 sec 30 sec 80 sec 5 min

Forward primer 8F AGAGTTTGATCCTGGCTCAG Reverse primer 1492R GGTTACCTTGTTACGACTT

Table 3-3. Genus-specific primers used in this study Genus Primers Sequences PCR Conditions PsEG30F 5’-ATYGAAATCGCCAARC-3’ (Mulet et al. Pseudomonas PsEG790R 5’-CGGTTGATKTCCTTGA-3’ 2009) Ps-for 5’-GGTCTGAGAGGATGATCAGT-3’ (Mulet et al. Pseudomonas Ps-rev 5’-TTAGCTCCACCTCGCGGC-3’ 2009) GPRA-UF2 5’-GGSAAGGGSKCNGTNATGCG-3’ (Waasbergen et Arthrobacter GPRA-UR2 5’-CCTTSCCCTGSCCNARYT-3’ al. 2000) A19-F2 5’-GTCATGCGCCTGGGCGACGA-3’ (Waasbergen et Arthrobacter A1-R1 5’-CTTGCCCTGGCCGAGTTGGT-3’ al. 2000) B-K1/F 5’-TCACCAAGGCRACGATGCG-3’ (Wu et al. Bacillus B-K1/R1 5’-CGTATTCACCGCGGCATG-3’ 2006b) CITSF 5’-TCGGAAGGTGCGGCTGGATC-3’ (Kaewklom et Bacillus CITSR 5’-AAGGCATCCACCGTGCGCCC-3’ al. 2014) Genus-specific primers used to identify putative mycorrhiza helper bacteria. Suitable primers for Microbacterium could not be identified.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 65 3.2.7. In vitro effect of bacterial cultures on pre-symbiotic growth of AM fungi Five surface-disinfected spores of F. mosseae were arranged in a pentagon in a 90 mm Petri dish containing 1% water agar (BD BactoAgar: Bacto Laboratories, Mt Pritchard, NSW, Australia) either by itself as a control or amended with 1 ml of a bacterial solution containing 1 × 107 CFU ml-1 of an individual bacterial strain (Table 3-4.). To obtain this concentration, 0.5 ml of an overnight culture of each strain was plated onto tryptic soy agar using a ten-fold dilution series of 101–106. From this series, the subset of dilutions yielding 25–250 colonies per Petri dish was used to estimate the total number of CFU ml-1 in the undiluted sample by multiplying the colony count by the appropriate dilution factor, in line with observations that values within this range return the lowest rates of error (Tomasiewicz et al. 1980). A mean value for CFU ml-1 was obtained from five replicates for each isolate. This was used to determine the volume of bacterial culture necessary to obtain the final concentration. Subsequently, five replicate Petri dishes were arranged in a randomised design generated with the DiGGeR extension for R (Coombes, 2002) and incubated at 25°C in the dark. After seven days, spores were photographed under a stereomicroscope (Nikon SMZ745T) using a Nikon DS-2Mv camera at 40-80× magnification and measurements of hyphal length taken in ImageJ (U. S. National Institutes of Health, Bethesda, Maryland, USA). Measurements were obtained by first photographing a metal scale at the same magnification and averaging 10 measurements of 1 mm to obtain a final scale. A germination count was also recorded from each sample. A spore was considered to have germinated when fresh hyphal growth could be seen emerging from the cut end of the subtending hyphae or through the spore wall.

3.2.8. In vitro formation of secondary spores Five surface-disinfected spores of F. mosseae were arranged in a pentagon in a 90 mm Petri dish containing 1% water agar (BD BactoAgar: Bacto Laboratories, Mt Pritchard, NSW, Australia). To the centre of each pentagon was added 0.5 ml of a bacterial suspension containing 1 × 107 CFU ml-1 of an individual bacterial strain (Table 3-4.). One Petri dish per replicate did not receive the bacterial suspension as a control. As before, Petri dishes were subsequently incubated at 25°C in the dark and five replicates were arranged in a randomised design generated with the DiGGeR extension for R (Coombes, 2002). After 20 weeks, Petri dishes were examined for incipient or developed spores. Where secondary spores appeared large enough to germinate (c. 90 µm or larger), they were removed from the Petri dishes, sterilised according to the

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 66 above protocol, and transferred to fresh Petri dishes containing 1% water agar. These secondary spores were incubated at 25°C in the dark and germination assessed after 7 days, as described above.

Table 3-4. Bacterial cultures used in each trial Culture 16s ID Location In vitroa In vivo BAC1 Bacillus Re + + BAC2 Bacillus Ni + + VAR1 Variovorax paradoxus Rs + + MIC1 Microbacterium Rp + + MIC2 Microbacterium Ni + + MIC3 Microbacterium Rp + - MIC4 Microbacterium Re + - PAE1 Paenibacillus Ni + + CHR1 Chryseobacterium Re + + PAN1 Pantoea/Erwinia Re + - PSE1 Pseudomonas Ni + + PSE2 Pseudomonas Re + + ART1 Arthrobacter Re + + MICC1 Micrococcus Re + - LYS1 Lysinibacillus Rs + - RHO1 Rhodococcus Rs + - ACI1 Acidovorax Ni + + SER1 Serratia marcescens Rs + + LEI1 Leifsonia Rp + - Bacterial taxa utilised in each experiment in the present study, along with 16s identifications obtained by BLAST searches and the soil or root compartment from which the strain was isolated, where Rs = rhizosphere, Rp = rhizoplane, Re = root endosphere, and Ni = nodule interior. aIncludes screening for secondary spores.

3.2.9. In vivo effect of bacterial cultures on symbiotic development of biserrula mycorrhizas Plants of B. pelecinus var. Casbah were grown in square forestry tubes (6 cm × 18 cm) containing 800 g of a sandy loam sterilised by autoclaving twice for a total of 2 h. Seeds

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 67 were surface-disinfected by soaking in a 30% solution of commercial bleach (1.3% NaOCl) with added Tween 20 for 10 minutes and rinsed five times in sterile water. These seeds were transferred to Petri dishes containing sterile, moistened filter paper. Seeds were germinated by incubating at 25°C and subsequently sown at a rate of one seed per pot after 4 days. Plants were grown in growth cabinets (Conviron Adaptis, Conviron Asia Pacific, Melbourne, Australia) under a 12/12 h daily light/dark cycle at temperatures of 24°C /14°C and a relative humidity of 75%. Plants were watered as necessary with sterile water. Bacterial cultures (Table 3-4.) were inoculated into pots immediately after the pots received their soil. Each pot received 30 ml of one of 12 bacterial suspensions containing approximately 1 × 107 CFU ml-1, or, in two control treatments per replicate, an equal volume of sterile water. Following seedling emergence, mycorrhizal propagules were pipetted directly onto the surface of the growing root at a rate of 200 surface-disinfected spores per plant. One control treatment per replicate did not receive spores to confirm that sterilisation had killed any AM spores pre-existing in the soil. For each of 5 replicates, the following 14 treatments were used: 12 bacterial cultures, each with 200 added spores; 1 sterile H2O (no culture) with 200 added spores (Control+); and

1 sterile H2O (no culture) without added spores (Control-). After 16 weeks, plants were removed from their pots and the roots isolated and rinsed to remove all soil. These roots were subsequently cleared and stained according to the following method adapted from Brundrett (2008). Roots were divided into sections of approximately 1 cm in length and subsamples weighing approximately 2 g taken. These were then cleared in 10% KOH for 5 min at 121°C, rinsed for 5 min in deionised water, and transferred to the staining solution. Root fragments were stained using a 0.03% solution of Chlorazol black E in a 1:1:1 mixture of lactic acid, glycerol, and deionised water for 72 h at room temperature and then destained in a solution of 50% glycerol for 1 week at room temperature prior to examination. Percent mycorrhizal colonisation (PC) was assessed by the gridline intersection method (Giovannetti and Mosse 1980). Stained roots were transferred to 90 mm diameter Petri dishes marked with a 1.27-cm grid and viewed under a stereomicroscope (Nikon SMZ745T) at 40-80× magnification. Gridlines were scanned vertically and horizontally and roots marked as either mycorrhizal or non-mycorrhizal where they intersected with the gridlines. Arbuscules, vesicles, spores, and mycelium were counted individually. In total, 100 observations were taken from each root system. Where

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 68 necessary, roots were discarded and additional roots drawn from the sample to ensure the total of 100 observations was met. 3.2.10. Functional characterisation of helper bacteria Bacterial strains found to exert a helper effect in vivo were further characterised using solid-media assays aimed at detecting chitinolytic (Murthy and Bleakley 2012), pectolytic, amylolytic, proteolytic (Hankin and Anagnostakis 1975), and cellulolytic (Liang et al. 2014) activity. Each putative helper strain was screened on five Petri dishes. To assess the suitability of the putative helpers for use in the saline environments to which biserrula may be exposed, osmotic stress tolerance was assessed for each strain by plating 300 µl of an overnight culture onto nutrient agar containing 0%, 2%, 4%, 6%, 8%, or 10% added NaCl. Cultures were incubated at 25°C for 72 h prior to measurement of colony diameter. Five colonies per Petri dish were measured. The colonies selected were those nearest to the points of an identical pentagon marked on each Petri dish. Each colony was photographed under a stereomicroscope (Nikon SMZ745T) using a Nikon DS-2Mv camera at 20-80× magnification and measured in ImageJ (U. S. National Institutes of Health, Bethesda, Maryland, USA). A single value for mean colony diameter was obtained for each Petri dish. This process was carried out for each of four replicates. The presence or absence of secretion systems in identified helper strains was assessed by querying the public databases MacSyDB/TXSSdb (Abby et al. 2016), EffectiveDB (Eichinger et al. 2016), and SecReT6 (Li et al. 2015). Additionally, DNA extracted from two helper strains identified as belonging to the Pseudomonas fluorescens complex was subjected to PCR targeting components of the Type III and Type VI secretion systems, using the primers HRCR8092-HRCT8986 (T3SS; Mazurier et al. 2015), HRCRVERST2-HRCTRHIZ2 (T3SS; Mazurier et al. 2006), and Pfl01_2045F-Pfl01_2045R (T6SS; Decoin et al. 2014).

3.2.11. Statistical analyses All statistical analyses were carried out using version 3.6.0 of the R statistical software program (R Core Team 2019). Measurements of alpha diversity were carried out on unrarefied sequence data, owing to the known tendency of rarefaction to introduce additional bias (Willis 2019, McMurdie and Holmes 2014). Species richness (S) values were obtained by counting the total number of OTUs represented in each sample. Shannon-Weiner index (H) values were obtained using the diversity function in the

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 69 package Vegan (Oksanen et al. 2019). Effective species numbers were derived for each sample by taking the exponential of the Shannon-Weiner index value (ExpH). To identify taxa selectively recruited to each soil compartment, sequence data were centred log-ratio (CLR) transformed (Gloor et al. 2017) and subjected to ANOVA-Like Differential Expression analysis using the ALDEx2 package (Gloor et al. 2016, Fernandes et al. 2014, Fernandes et al. 2013). Two conditions were compared at a time and the results of Welch’s t-test used to identify differences in taxon abundance. To ensure a robust estimation of effect size, 1000 Monte-Carlo iterations were used to generate all aldex.clr objects. Benjamini-Hochberg corrected p values were calculated to take into account the effect of multiple comparisons. Effects were considered significant at q < 1. Certain non-significant effects are also reported, following the suggestion of Gloor (2018) that effects found to be non-significant according to their q values be reported when their effect size is > 1. In the present study, a higher threshold (> 2) has been employed to account for high between-replicate variation in effect size. In vitro germination data were analysed using a binomial generalised linear model. To determine the effects of bacterial strain on in vitro hyphal growth, a single mean value for each Petri dish was obtained and subjected to ANOVA using the aov function in R. Owing to low germination, one taxon (a strain of Paenibacillus) was removed from the hyphal length data. Similarly, uneven germination across all taxa precluded any analysis of the effect on hyphal elongation of the position of each spore in the dish. Data from the osmotic stress assay was used to produce line graphs showing the decline in colony diameter with increasing NaCl concentration. Percent colonisation values derived from the in vivo screening were analysed using Poisson regression. Estimated marginal means for each treatment were subjected to pairwise comparison using Tukey’s method in the package emmeans (Lenth 2019). Tukey-adjusted p values < 0.05 were considered significant. The negative control (Control-) was removed from the dataset as no mycorrhizal colonisation was observed.

3.3. Results

3.3.1. Diversity profile Diversity profiling returned a total of 1,067,045 16s V1-V3 sequences, corresponding to 11,127 operational taxonomic units (OTUs). Sequence counts per sample ranged from 70,228 to 121,494 for rhizosphere samples, 21,366 to 65,008 for bulk soil samples, and

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 70 63,153 to 75,056 for field soil samples. OTU counts per sample ranged from 4,331 to 6,312 for rhizosphere samples, 3,655 to 5,534 for bulk soil samples, and 3,413 to 5,713 for field soil samples. OTUs were assigned to 32 phyla, 88 classes, 135 orders, 174 families, and 230 genera. 118 OTUs were identified to species level. The most abundant phyla were the , accounting for 37.80% of all sequences and 17.15% of all recorded OTUs, and the Proteobacteria, accounting for 26.62% of all sequences and 19.63% of all OTUs (Figure 3-1.). The taxon with the greatest relative abundance was a strain of Bradyrhizobium (3.28% of all sequences). No other taxon accounted for more than 2% of the total sequence count. Additionally, 943 sequences belonging to 25 OTUs were assigned to the kingdom Bacteria but could not be assigned to a phylum and 77,133 sequences belonging to 2,619 OTUs could not be assigned to a kingdom. Shannon diversity values and effective species were similar across all treatments (Table 3-5). Analysis of the 167 OTUs contributing more than 1000 sequences to the dataset identified 26 that were differentially abundant in the rhizosphere relative to the field soil (Figure 3-2.). By contrast, no OTUs were found to be differentially abundant in the rhizosphere relative to the bulk soil at q < 1, and only one (Variovorax paradoxus, q=0.08) was found to differ between the bulk soil and the field soil (Figure 3-3.). Of the 26 taxa found to be differentially abundant between the rhizosphere and field soil samples, only five were enriched in the rhizosphere; the remaining 21 OTUs were significantly more abundant in the field soil (Table 3-6.). Median positive effects were observed Mesorhizobium sp. (3.86, q=0.03) and Variovorax paradoxus (3.49, q=0.02), as well as strains of the Comamonadaceae (1.95, q=0.10), (1.89, q=0.07), and Actinomycetales (1.70, q=0.09), indicating that these strains were recruited to the biserrula rhizosphere. The largest median negative effects were observed for two strains of Bacillus (both -3.07, q=0.03, 0.05), one strain of the Bacillales (-2.78, q=0.04), one strain of the Actinomycetales (-2.76, q=0.05), and one of the Bacillaceae (-2.67, q=0.06). The OTU identified as belonging to the Micrococcaceae most closely matched species of Arthrobacter or Paenarthrobacter. As the effect plots make clear, large effect sizes were associated with high dispersion; this between-sample heterogeneity likely impeded the detection of selectively recruited taxa. Following the suggestion of Gloor (2018), effects found to be non-significant according to their Benjamini-Hochberg corrected p values but with large median effect sizes (in this case, > 2 or < -2) are reported here. Two such OTUs, one belonging to the family Microbacteriaceae (2.46, q=0.17) and one to the genus Pseudonocardia (2.45,

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 71 q=0.16), were found to be non-significantly enriched in the biserrula rhizosphere relative to the field soil. Three were identified in the comparison of the rhizosphere and bulk soils: species of Mesorhizobium (q=0.15), Arthrobacter or Paenarthrobacter (q=0.17), and the Microbacteriaceae (q=0.22) were observed to be enriched in the rhizosphere relative to the bulk soil. In the comparison of the bulk and field soils, four OTUs were observed to exceed the cut off for effect size, namely two members of the Bacillaceae (q=0.11, 0.13) enriched in the field soil relative to the bulk soil, and species of Mesorhizobium (q=0.21) and Pseudonocardia (q=0.25) enriched in the bulk soil relative to the field soil. In all cases, however, it cannot be ruled out that these non- significant observations were false positives.

Table 3-5. Diversity index values for 16s diversity profiles by soil compartment

Soil compartment Index Range Mean (± s.e.)

Rhizosphere S 4,331 – 6,312 5551 ± 289.68 H 6.66 – 7.15 6.94 ± 0.08 Exp(H) 778.94 – 1269.65 1049.61 ± 77.82 Bulk soil S 3,655 – 5,534 4645.67 ± 305.67 H 6.82 – 7.20 7.06 ± 0.07 Exp(H) 919.96 – 1346.06 1172.50 ± 77.89 Field soil S 3,413 – 5,713 4410.67 ± 681.20 H 6.39 – 7.14 6.71 ± 0.22 Exp(H) 595.85 – 1262.77 864.84 ± 203.03 Values for taxon richness (S), the Shannon-Weiner index (H) and effective species number (Exp(H)) calculated for all soil compartments, showing ranges and means ± standard error.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 72

Figure 3-1. Proportional composition of the 16s diversity profiling dataset by phylum, as assessed by (a) total 16s sequence count and (b) total count of operational taxonomic units (OTUs). Phyla accounting for less than 0.1% of the total sequence or OTU count have been omitted for clarity.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 73 3.3.2. In vitro screening assay Germination of F. mosseae spores was significantly enhanced in 11 of the 19 bacterial treatments relative to the control (Figure 3-4.). The most pronounced increase in germination was observed in response to a strain identified as Variovorax paradoxus (VAR1), a plant growth-promoting species reported from the mycosphere of both AM (Lecomte et al. 2011) and ectomycorrhizal fungi (Boersma et al. 2010). The remaining taxa included two strains each of Bacillus (BAC1 and BAC2), Microbacterium (MIC1 and MIC2), and Pseudomonas (PSE1 and PSE 2), and single strains of Acidovorax (ACI1), Chryseobacterium (CHR1), Arthrobacter (ART1), and Serratia marcescens (SER1). One bacterial treatment, a strain of Paenibacillus (PAE1), appeared to suppress spore germination in vitro. Only three of the 25 spores co-inoculated with that strain germinated within seven days.

a b

Figure 3-2. Operational taxonomic units (OTUs) identified as differentially abundant in the biserrula rhizosphere and the surrounding field soil. Points marked in red are differentially abundant between treatments (q < 0.1), while points marked in grey or black are non- significant. The Bland-Altman plot (a) displays the relationship between difference and abundance. The effect plot (b) maps the relationship between difference and dispersion. Points in red above the line in (b) are OTUs significantly enriched in the rhizosphere; points in red below the line are OTUs significantly enriched in the field soil. Data are centred log-ratio transformed partial 16S rRNA gene sequence counts. Analysis was restricted to OTUs occurring > 1000 times in the dataset.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 74

a b

c d

Figure 3-3. Operational taxonomic units (OTUs) identified as differentially abundant in the rhizosphere and the adjacent bulk soil (a,b) and in the bulk soil and the surrounding field soil (c,d). Single point marked in red and indicated by an arrow (c,d) is differentially abundant between treatments (q < 0.1), while points marked in grey or black are non- significant. The Bland-Altman plots (a,c) displays the relationship between difference and abundance. The effect plot (b,d) maps the relationship between difference and dispersion. Single point marked in red below the line in (d) is significantly enriched in the bulk soil. Data are centred log-ratio transformed 16S rRNA gene sequence counts.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 75 Table 3-6. Differentially abundant bacterial taxa in the biserrula rhizosphere and the surrounding field soil OTU Consensus ID Effect size q OTU2 Mesorhizobium species 3.860773 0.03 OTU5 Micrococcaceae 1.886778 0.07 OTU11 Bacillales -1.65149 0.08 OTU12 Solirubrobacterales -1.39979 0.10 OTU16 Variovorax paradoxus 3.491501 0.02 OTU20 Actinomycetales -2.75867 0.05 OTU40 Bacillus species -3.06935 0.05 OTU45 Bacillaceae -2.66977 0.06 OTU46 Planococcaceae -1.37392 0.09 OTU50 Rhodospirillaceae -1.86035 0.07 OTU55 Gaiellaceae -2.30672 0.08 OTU59 Bacillales -2.78098 0.04 OTU75 Candidatus Koribacter species -1.61381 0.09 OTU103 Nocardioidaceae -1.73194 0.08 OTU106 Comamonadaceae 1.953747 0.10 OTU136 Sinobacteraceae -2.36849 0.05 OTU174 Sporosarcina species -1.70927 0.09 OTU208 TM7-1 -2.23686 0.05 OTU337 Rhodospirillaceae -2.15235 0.08 OTU515 Unassigned -1.92242 0.10 OTU633 Bacillus species -3.0716 0.03 OTU688 Sporosarcina -1.67799 0.07 OTU3196 Gaiellaceae -1.61012 0.09 OTU3655 Planococcaceae -2.15414 0.06 OTU3901 Planococcaceae -1.69025 0.09 OTU10612 Actinomycetales 1.703586 0.09 Bacterial operational taxonomic units (OTUs) derived from 16S diversity profiling and found by ALDEx analysis to be enriched in either the biserrula rhizosphere or in the surrounding field soils. OTUs with positive values for effect size were enriched in the rhizosphere, while OTUs with negative values for effect size were enriched in the field soil.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 76 Figure 3-4. The effect of bacterial treatment on in vitro germination of Funneliformis mosseae spores. Values on the vertical axis are mean germination rates ± standard error. Treatment names on the horizontal axis correspond to taxa identified in Table 3-4. Where treatments share a lower-case letter, values are not significantly different according to Tukey’s method of pairwise comparison (p < 0.05).

Hyphal elongation was significantly enhanced in four bacterial treatments relative to the control (Figure 3-5.). The greatest increase in hyphal growth was observed for two strains of Pseudomonas (PSE1 and PSE 2), followed by one strain of Bacillus (BAC1) and one of Microbacterium (MIC1). All four strains additionally promoted germination of the spores. Despite returning the highest in vitro germination rate, V. paradoxus did not enhance hyphal growth. The Paenibacillus strain observed to suppress spore germination was not included in the hyphal elongation data, however, the three spores that did germinate in that treatment had produced only short hyphae (<0.2 mm in length) by the conclusion of the experiment.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 77 In all but one case, germination occurred by extension of the new hypha through the remnants of the subtending hypha. One spore was observed to germinate by extension of the hypha directly through the spore wall (Figure 3-6.).

Figure 3-5. The effect of bacterial treatment on in vitro hyphal elongation of Funneliformis mosseae spores. Values on the vertical axis are mean hyphal lengths ± standard error. Treatment names on the horizontal axis correspond to taxa identified in Table 3-4. Where treatments share a lower-case letter, values are not significantly different according to Tukey’s method of pairwise comparison (p < 0.05).

3.3.3. In vitro development of secondary spores Secondary spores of F. mosseae were obtained from two bacterial treatments, both strains of Pseudomonas (PSE1 and PSE2) (Figure 3-7.). In total, nine spores were obtained from the two treatments, measuring between 92 µm and 113 µm in diameter. Germination values of 33.3% were returned for secondary spores of both treatments, corresponding to one from three in PSE1 and two from six in PSE2. Germinant spores were between 105 µm and 113 µm in diameter. Structures resembling incipient spores were observed in most treatments, with the exception of Acidovorax sp. (ACI1), but did not develop to completion. Two treatments, Chryseobacterium sp. (CHR1) and

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 78 Acidovorax sp. (ACI1), appeared to induce structural alterations in the fungal colony (Figure 3-8.). Whereas most observed colonies were linear and sparsely branched, these treatments yielded intricately branched (CHR1) or radial colonies (ACI1). Interpretation of these differences was complicated by low rates of germination, however. In particular, only three spores germinated in the ACI1 treatment.

a b c

Sh

Sh Sh

d e f

Eh

Sh Eh

g Eh h i Eh

Sh Eh

Figure 3-6. Photomicrographs of Funneliformis mosseae spores before and after germination. Spores (a-d) are ungerminated and show subtending hyphae (Sh) trimmed to ≤ 30 µm in length. Spores (e-h) are newly germinated and show emerging hyphae (Eh) extending through the trimmed subtending hyphae. Spore (i) is a germinated spore exhibiting direct extension of the emerging hypha through the spore wall. The location of the subtending hyphae is also indicated. Black bars represent c. 100 µm.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 79 a b c Is

Ps

Ss

Ss Ps Ps

Figure 3-7. Photomicrographs of Funneliformis mosseae secondary spores (a, b) and incipient spore development in vitro (c). Labels indicate the primary (or parent) spores (Ps), secondary spore (Ss), and incipient terminal spore (Is). Primary and secondary spores in (a) did not grow in close proximity and were grouped manually. All images are from treatments exposed to cultures of Pseudomonas granadensis. Black bars represent c. 100 µm.

3.3.4. In vivo screening for mycorrhizal colonisation Six bacterial strains were observed to enhance mycorrhizal colonisation in vivo (Figure 3-9.). These were subsequently identified as a Pseudomonas strain belonging to the P. koreensis subgroup of the P. fluorescens lineage (Pascual et al. 2015) (PSE1); a Bacillus strain (BAC1) belonging to the B. cereus group (Liu et al. 2017); Pseudomonas granadensis (PSE2), also a member of the P. koreensis subgroup of the P. fluorescens lineage (Pascual et al. 2015); an Arthrobacter strain that could not be identified to species (ART1); a member of the B. aquimaris/B. vietnamensis cluster (Noguchi et al. 2004) (BAC2); and Microbacterium xylanilyticum (MIC1). BLAST results and voucher numbers for these putative MHBs are given in Table 3-7. Additional confirmation of these identifications was obtained by searching the curated 16s database EzBioCloud (Yoon et al. 2017). These results were not found to differ from those obtained by searching GenBank. All examined root systems exhibited some mycorrhizal colonisation. Raw percent colonisation values for untreated biserrula ranged from 11% to 26% (mean = 17.6 ± 2.77%), consistent with the previously reported values (see section 2.6.1.). The lowest value for a single root system was 8, recorded from V. paradoxus, which also returned the lowest mean colonisation rate, although this did not significantly differ from the control. The highest value for a single root system was 71%, recorded from both the P. koreensis subgroup taxon (PSE1) and the Bacillus cereus group taxon (BAC1). These two treatments returned the highest mean colonisation rates, significantly higher than Arthrobacter sp. (ART1), the Bacillus aquimaris/B. vietnamensis cluster taxon (BAC2),

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 80 and Microbacterium xylanilyticum (MIC1). Pseudomonas granadensis (PSE2) bridged the two clusters. In the PSE1 treatment, some nodulation of the root was observed in four of five replicates, though these nodules were inactive. This strain was isolated from the interior of a biserrula nodule.

a b

c d

Figure 3-8. Photomicrographs of Funneliformis mosseae colonies incubated for 20 weeks. Spores exposed to most treatments formed linear, sparely branched colonies, as in (a) Serratia marcescens and (b) the control treatment. Spores exposed to Chryseobacterium sp. (c) yielded densely branched colonies, while Acidovorax sp. (d) yielded radial colonies. All scale bars represent 100 µm.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 81 Figure 3-9. The effect of bacterial treatment on in vivo colonisation of Biserrula pelecinus by Funneliformis mosseae. Values on the vertical axis are log-transformed least-squares mean percent colonisation values. Bars represent 95% confidence intervals. Where arrows overlap, groups are not significantly different according to Tukey’s method of pairwise comparison. Treatments indicated on the horizontal axis correspond to taxa identified in Table 3-4.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 82 Table 3-7. BLAST matches for putative MHB strains BLAST matches Strain Voucher No. Consensus ID Total Query Percent Accession Description E value score coverage identity ART1 DAR85049 Arthrobacter sp. MN710453.1 Arthrobacter bambusae strain TS31 16S 933 100% 0.0 100.00% ribosomal RNA gene, partial sequence MN710431.1 Arthrobacter gyeryongensis strain TS07 16S 933 100% 0.0 100.00% ribosomal RNA gene, partial sequence KT369918.1 Arthrobacter nicotinovorans strain YF01- 933 100% 0.0 100.00% 5(4) 16S ribosomal RNA gene, partial sequence MH828349.1 Arthrobacter boritolerans strain Sneb1989 933 100% 0.0 100.00% 16S ribosomal RNA gene, partial sequence BAC1 DAR85051 Bacillus cereus groupa MN733060.1 Bacillus cereus strain CPO 4.232 16S 1186 100% 0.0 100.00% ribosomal RNA gene, partial sequence MN181263.1 Bacillus thuringiensis strain T1-9 16S 1186 100% 0.0 100.00% ribosomal RNA gene, partial sequence MN428214.1 Bacillus toyonensis strain PK4-3 16S 1186 100% 0.0 100.00% ribosomal RNA gene, partial sequence MN428213.1 Bacillus mobilis strain PK4-2 16S ribosomal 1186 100% 0.0 100.00% RNA gene, partial sequence MN421526.1 Bacillus proteolyticus strain SR6-9 16S 1186 100% 0.0 100.00% ribosomal RNA gene, partial sequence

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 83 Table 3-7. BLAST matches for putative MHB strains BLAST matches Strain Voucher No. Consensus ID Total Query Percent Accession Description E value score coverage identity MN421024.1 Bacillus wiedmannii strain SK1-22.1 16S 1186 100% 0.0 100.00% ribosomal RNA gene, partial sequence MK734308.1 Bacillus pacificus strain WTB116 16S 1186 100% 0.0 100.00% ribosomal RNA gene, partial sequence MK734052.1 Bacillus albus strain Q1 16S ribosomal RNA 1186 100% 0.0 100.00% gene, partial sequence BAC2 DAR85052 Bacillus aquimaris/B. MN661275.1 Bacillus aquimaris strain ABC6 16S 1567 100% 0.0 100.00% vietnamensis clusterb ribosomal RNA gene, partial sequence MF981077.1 Bacillus vietnamensis strain Ahv-Q1 16S 1567 100% 0.0 100.00% ribosomal RNA gene, partial sequence MIC1 DAR85050 Microbacterium JF700460.1 Microbacterium xylanilyticum strain HR87 1373 100% 0.0 100.00% xylanilyticum 16S ribosomal RNA gene, partial sequence PSE1 DAR85053 Pseudomonas koreensis MN181240.1 Pseudomonas koreensis strain P9-76 16S 1424 100% 0.0 100.00% subgroup (P. fluorescens ribosomal RNA gene, partial sequence lineage)c CP038438.1 Pseudomonas fluorescens strain LBUM677 8549 100% 0.0 100.00% chromosome, complete genome MK240436.1 Pseudomonas moraviensis strain WTB8 16S 1424 100% 0.0 100.00% ribosomal RNA gene, partial sequence

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 84 Table 3-7. BLAST matches for putative MHB strains BLAST matches Strain Voucher No. Consensus ID Total Query Percent Accession Description E value score coverage identity MH211298.1 Pseudomonas reinekei strain PgBe208 16S 1424 100% 0.0 100.00% ribosomal RNA gene, partial sequence MH071318.1 Pseudomonas granadensis strain WTT2 16S 1424 100% 0.0 100.00% ribosomal RNA gene, partial sequence PSE2 DAR85054 Pseudomonas MG022037.1 Pseudomonas granadensis strain PF81 RNA 505 100% 5e-139 99.28% granadensis polymerase sigma factor (rpoD) gene, partial cds BLAST matches and voucher numbers for consensus DNA sequences obtained from mycorrhiza helper bacteria. Redundant matches and matches not identified to species level have been omitted. a(Liu et al. 2017). b(Noguchi et al. 2004). c(Pascual et al. 2015).

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 85 3.3.5. Functional characterisation of putative helper strains Results of the solid media assays and database searches are given in Table 3-8. No single trait was shared across all helper strains, suggesting that multiple pathways may lead to enhancement of the mycorrhizal symbiosis, or that the most significant trait was not assessed in the present study. In the osmotic stress test, only one strain, Bacillus aquimaris/B. vietnamensis (BAC2), was capable of growth at NaCl concentrations above 4% (Figure 3-10.). This species remained capable of colony formation at 10% NaCl, although its mean colony diameter in that treatment was only 1.5% of its mean value under 0% NaCl (Figure 3-11.). This is consistent with the description of this species cluster as moderately halophilic or halotolerant (Hernández-González and Olmedo-Álvarez 2016, Noguchi et al. 2004). Microbacterium xylanilyticum (MIC1) was incapable of growth above 2% NaCl. One strain, Arthrobacter sp. (ART1), showed little decline in diameter between 0% and 2% NaCl but a steep decline between 2% and 4% NaCl. The remaining species, with the exception of BAC2, declined in approximation of a Type III survivorship curve (Demetrius 1978).

Table 3-8. Functional traits of putative MHB strains

Strain Identity Am Ce Ch Pe Pr T3SS T4SS T6SS

ART1 Arthrobacter sp. + + + - - - - - BAC1 Bacillus cereus group - + + - + - - - BAC2 Bacillus aquimaris/vietnamensis + + - - + - - - MIC1 Microbacterium xylanilyticum ------PSE1 Pseudomonas koreensis subgroup - - + + - - - - PSE2 Pseudomonas granadensis - - + + - - - - Results of solid media assays and database searches aimed at characterising functional traits of identified mycorrhiza helper bacteria. Traits assayed are amylolytic activity (Am), cellulolytic activity (Ce), chitinolytic activity (Ch), pectinolytic activity (Pe), proteolytic activity (Pr), and possession of Type III (T3SS), Type IV (T4SS), and Type VI (T6SS) secretion systems. The notation system indicates evidence (+) or lack of evidence (-) of a particular trait in that strain.

Database searches did not identify any secretion system components in the helper taxa. This was due, in part, to a lack of data for the relevant taxa and was complicated by poor resolution of the identities of several of the helper taxa. Likewise, PCR targeting secretion system components in the two pseudomonads resulted in no amplification of

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 86 DNA. Genes linked to Type VI secretion have been identified in P. koreensis sensu stricto (Lin et al. 2016) but may not be present in the members of the P. koreensis subgroup under study here.

Figure 3-10. Osmotic stress tolerance of putative mycorrhiza helper bacteria exposed to increasing levels of NaCl. Points on the vertical axis are mean values (± standard error) for colony diameter across four replicates. Treatments indicated on the horizontal axis are percent concentrations of NaCl in the growth medium.

a b c

Figure 3-11. Colonies of Bacillus aquimaris/vietnamensis grown on unamended nutrient agar (a), nutrient agar amended with 8% NaCl (b), and nutrient agar amended with 10% NaCl (c). All scale bars represent 100 µm.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 87 3.4. Discussion This study identified six plant-associated bacterial taxa capable of promoting the colonisation of biserrula roots by the AM fungus Funneliformis mosseae. This is the first observation of a helper effect in a biserrula host; consequently, this study extends the range of plant-microbe systems found to exhibit an MHB effect. The most pronounced effects were observed in two members of the Pseudomonas koreensis subgroup of the P. fluorescens group (PSE1 and PSE2), and in one member of the Bacillus cereus group (BAC1). Under these treatments, mean colonisation rates increased to 63.6 ± 3.83% (PSE1), 60.0 ± 3.11% (BAC1), and 49.8 ± 3.48% (PSE2), compared to the control value of 17.6 ± 2.77%. A role for Pseudomonas spp., and particularly the fluorescent pseudomonads, in promoting mycorrhizal growth and colonisation is long established: they have been identified as helpers in both AM (Babana and Antoun 2005, Fester et al. 1999, Vosátka and Gryndler 1999, Barea et al. 1998) and ectomycorrhizal (Mediavilla et al. 2016, Founoune et al. 2002a, Frey-Klett et al. 1997) systems. The magnitude of the increase in mycorrhizal colonisation produced by Pseudomonas strains varies considerably between studies, from a less than 1.5-fold increase in Solanum tuberosum (potato, 14.2% to 18.9%: Vosátka and Gryndler 1999) and S. lycopersicum (tomato, 35% to 52%: Barea et al. 1998) to a more than 2.5-fold increase in Acacia holosericea (29.7% to 76.7%: Founoune et al. 2002a) and transformed ladanifer vitroplants (6.98 ± 2.73% to 18.55 ± 3.48%: Mediavilla et al. 2016). Intermediate values were observed for the crop grasses Triticum aestivum (wheat, 4.5% to 9.8%: Babana and Antoun 2005) and Hordeum vulgare (barley, approximately 20% to 40%: Fester et al. 1999). The increases reported in the present study (3.6-fold for the mean value of PSE1 and 2.8- fold for the mean value of PSE2) exceed these prior observations. Likewise, members of the B. cereus group have previously been observed to promote mycorrhization (Vivas et al. 2003, Bending et al. 2002), and other Bacillus species are known to function as helpers (Zhao et al. 2014, Vivas et al. 2005, Mamatha et al. 2002). The 3.4-fold increase observed in biserrula exceeds the 1.3- to 1.6-fold increases reported for an ECM association (Zhao et al. 2014), as does the 2.2-fold increase returned for a second Bacillus strain, a member of the Bacillus aquimaris/vietnamensis species cluster. These values are exceeded, however, by the 4- to 5-fold increases observed for several strains in a luteus-Pinus sylvestris (Bending et al. 2002). Comparison with the remaining studies is impeded by methodological differences. This study also

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 88 provided the second observation of a helper effect in Microbacterium: the 2.1-fold increase recorded for M. xylanilyticum is comparable to the 1.6-fold to 2.0-fold increases observed for Microbacterium strains in Zea mays (Soni et al. 2014). Additionally, the first observation of a helper effect in Arthrobacter in an AM system was recorded. This effect was stronger in biserrula (2.3-fold increase) than in the Pinus sylvestris ectomycorrhiza previously investigated (1.7-fold increase: Rózycki et al. 1994). In the present study, promotion of spore germination and pre-symbiotic growth in vitro were used to justify narrowing the field of bacteria employed in in vivo colonisation tests. While some means of restricting the taxa under study is practically necessary, results indicate that this approach may be ineffective and may exclude certain taxa that are subsequently found to have a helper effect. In particular, in vitro germination rate proved to be a poor predictor of in vivo colonisation success. Of the 11 taxa found to promote germination in vitro, five subsequently failed to promote root colonisation. Curiously, Variovorax paradoxus returned both the highest in vitro germination rate and the lowest in vivo percent colonisation value, though this was not significantly lower than the control. Similarly, the Paenibacillus strain observed to suppress germination in vitro did not appear to suppress colonisation in vivo. This in vitro suppression effect was extreme but not unprecedented for Paenibacillus: P. favisporus TG1R2 has previously been observed to roughly halve the in vitro germination rate of Rhizophagus irregularis relative to an uninoculated control (Fernández Bidondo et al. 2011). In that case, as in the present study, the effect observed in cultures was not apparent in the subsequent in vivo trial; indeed, the ‘suppressive’ Paenibacillus strain was found to act as an MHB in Glycine max (soybean). Disagreements between in vitro germination studies and in vivo colonisation screenings have also been reported for Pseudomonas sp. F113, which acts as an MHB strain in vivo but does not promote in vitro germination of F. mosseae (Barea et al. 1998). Hyphal elongation fared slightly better, though neither BAC2 (B. aquimaris/B. vietnamensis) nor ART1 (Arthrobacter sp.), which were found to promote colonisation of biserrula roots in vivo, significantly increased hyphal growth in vitro. Hyphal elongation is presumed to increase fungus-root contacts and therefore opportunities for infection (Frey-Klett et al. 2007) but this does not appear to translate reliably to increases in root colonisation. A suppressive effect on in vitro hyphal growth of the

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 89 ectomycorrhizal fungus Lactarius rufus has also been reported for a Paenibacillus strain, Paenibacillus sp. EJP73 (Aspray et al. 2013, Deveau et al. 2007), but this was observed to correlate with an increase in hyphal branching density, which may increase fungus-root contacts and create more opportunities for mycorrhizal colonisation (Aspray et al. 2013). This could account for the helper effect previously reported for that strain (Aspray et al. 2006b, Poole et al. 2001), suggesting that hyphal elongation alone is inadequate as a predictor of MHB status, although a similar increase in colony branching density observed in the present study for Chryseobacterium sp. (CHR1) and Acidovorax sp. (ACI1), did not correspond to an increase in mycorrhizal colonisation. Had hyphal elongation alone been used to justify excluding candidates from screening, the helpers BAC2 and ART1 would not have been identified. The observation of secondary spore formation in vitro, in the absence of a plant host, is novel for F. mosseae and has previously been observed only once in an AM system in response to bacterial inoculation. The earlier study employed the helper strain Paenibacillus validus and successfully promoted spore development in the AM fungus Rhizophagus irregularis (Hildebrandt et al. 2002, Hildebrandt et al. 2006). In the present study, two members of the P. koreensis complex were observed to prolong the life of the fungus and facilitate sporulation. In common with that study, in vitro secondary spores observed in this case were scarce (only nine in total and six of those infertile) and were considerably smaller than those observed for plant-based trap cultures. These low spore numbers, together with inconsistency between replicates, precluded detailed analysis, and the mechanism of action is unknown. While AM fungi have traditionally been considered obligate biotrophs (Requena et al. 2007), spore formation has been observed in response to plant exudates in the absence of living material (Liu et al. 2019b). It is possible, though unconfirmed, that bacteria may mimic this function, perhaps through exudations of their own. Growth promotion, but not sporulation, has also been observed in response to bacterial degradation of certain culture media (Abdellatif et al. 2019), which may act as an alternative source of carbon. The results of the osmotic stress assay suggest that, with the exception of a single taxon (B. aquimaris/vietnamensis), the helpers identified in the present study may be unsuited to implementation in severely saline conditions. This was most obvious in the case of M. xylanilyticum, which did not tolerate NaCl concentrations above 2%. The ionic concentrations used in this assay do not closely resemble real-world conditions, however: assuming a standard conversion factor of 640 ppm = 1 dS/m, a solution of 2%

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 90 NaCl is equivalent to 31.25 dS/m, far in excess of the conditions biserrula, and indeed most crop and pasture plants, would be expected to tolerate (NSW Environment, Energy and Science 2008). Therefore, while the taxa under study, with the exception of B. aquimaris/vietnamensis, are unlikely to obtain a competitive advantage in saline soils, they are also unlikely to disappear altogether. Screening for additional environmental tolerances, such as acid conditions or heavy metals, should help to define the range of conditions to which each helper is suited. These assays should be paired with additional screenings for functional traits of relevance to MHB activity. The traits identified in the present study offer no clear, singular explanation for the helper effect. No single examined trait was shared among all helpers, raising the possibility that multiple mechanisms might account for promotion of AM growth and colonisation. Alternatively, it is possible that the correct trait was not examined. The absence of secretion systems from the identified helpers suggests that the recent focus on this pathway may be artificially limiting. While Type III secretion has been identified as a mechanism by which fluorescent pseudomonads may exert their helper effects (Viollet et al. 2017, Cusano et al. 2011), the Pseudomonas spp. under consideration here could not be shown to possess secretion system components. This finding should not be considered definitive, however, as genomic data for the relevant taxa was absent from databases and it is not certain that the secretion-specific primers utilised in this study are suitable for all species of the large P. fluorescens complex. Nevertheless, it remains possible, even likely, that the helper effect is multifactorial. The appearance of nodules in one treatment (PSE1: P. koreensis subgroup taxon) suggests that one underexplored factor may be a dual promotional effect on both the mycorrhization and nodulation pathways. Nodulation factors isolated from rhizobia have previously been observed to stimulate activity in shared elements of both pathways (Oláh et al. 2005, Xie et al. 1995) and nod factor analogues have been identified in non-rhizobial symbionts belonging to the genera Methylobacterium and Burkholderia (D’Haeze and Holsters 2002). It is possible that the production of these factors or their analogues may account for some reported helper effects. The taxa identified as MHB strains in the present study were isolated from the rhizoplane (one taxon), root endosphere (three taxa), and the interior of the nodules (two taxa) of biserrula plants, justifying the description of these species as “plant-associated”. It is not clear in all cases, however, that this is an exclusive association: taxa may be both plant-associated and ubiquitous in the general soil environment. To counter this,

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 91 the diversity profiling data offers some measured evidence to suggest that certain helper strains may be selectively recruited to the biserrula rhizosphere. A significant increase in the relative abundance of a taxon belonging to Arthrobacter or Pseudarthrobacter was observed in rhizosphere samples. Non-significant effects were observed for strains of the Microbacteriaceae (2.46, q=0.17) and for a Pseudomonas strain closely matching members of the P. koreensis subgroup of the P. fluorescens complex (1.57, q=0.20). In the latter case, the effect fell below the cut off size used in this study; however, it still exceeded the effect size cut off proposed by Gloor (2018). By contrast, Bacillus gave no indication of being recruited to the rhizosphere: five strains of the Bacillales, including two identified to the genus Bacillus and one to the family Bacillaceae, were found to be enriched in the field soil relative to the rhizosphere. Interpretation of this data is limited by the necessary use of different gene targets in the diversity profiling exercise and the identification of individual strains. For this reason, it cannot be stated confidently that these recruited (or potentially recruited) taxa represents the identified helper strains. It may be worth noting that the P. koreensis subgroup taxon, the Microbacterium species, and one of the Bacillus species were isolated from the interior of nodules removed from biserrula roots. If they are highly specialised for that environment, they may not appear in the diversity profiling data, as sampling focused on the rhizosphere soil.

3.5. Conclusion In summary, this study identifies candidate rhizobacteria that, should they be implemented in an agricultural setting, could act to promote mycorrhization and subsequent growth of biserrula plants. In the case of the B. aquimaris/B. vietnamensis taxon, this beneficial effect could extend to otherwise inhospitably saline environments. Significant hurdles still need to be overcome before this can become a reality, however. The axenic experiments used here to classify bacteria as MHB strains represent a highly artificial interpretation of the rhizosphere environment. The diversity profiling data demonstrates the extreme complexity of a plant rhizosphere in a pasture setting. It is this environment that a bacterial strain must contend with if it is to be considered a true MHB. To this end, experiments were conducted both to more thoroughly explore the dynamics of the biserrula rhizosphere in pasture and to determine whether the candidate helper bacteria identified in this study were capable of exerting their effects in a populated rhizosphere.

Chapter 3. Identification of mycorrhiza helper bacteria in the biserrula rhizosphere 92 Chapter 4. Ecological interactions in the rhizosphere of Biserrula pelecinus L. and their influence on MHB activity

4.1. Introduction The colonisation of a root by a mycorrhizal fungus results in the development of a distinctive microhabitat in the soil known as the mycorrhizosphere (Rambelli 1973). Microorganisms may be recruited to this environment, or to exposed fungal tissues (the mycosphere), in a manner similar to the curation of the root by the rhizosphere effect (Bakker et al. 2013). The terms “mycorrhizosphere effect” and “mycosphere effect” have been coined for this reason (Linderman 1988). While it is not possible to precisely define the limits of the mycorrhizosphere, hyphae associated with AM mutualisms have been reported to extend up to 12 cm from the host plant (Schnepf et al. 2008) and, nearer the plant, to produce as much as 100 m of hyphae per cm3 of soil (Miller et al. 1995). Taken together, these values suggest that the mycorrhizosphere may have both diffuse and locally intensive effects. Bacteria occupying these compartments are known to interact with the mycorrhiza in ways that may be beneficial or detrimental to the symbiosis (Johansson et al. 2004, Garbaye 1991). The abundance of bacteria within and around components of the plant-fungus symbiosis in soil has led to the description of mycorrhizal associations as “tripartite mutualisms” (Bonfante and Anca 2009). The axenic screening methods commonly used to identify mycorrhiza helper bacteria produce a simplified example of a tripartite mutualism, in which the interactions between the fungus and the plant are presumed to be mediated by the action of the candidate helper bacterium. This, however, is likely to be a highly artificial view of the mycorrhizal symbiosis: in a natural or agricultural setting, helper bacteria do not act in isolation but instead are embedded in the rhizosphere, an environment rich in microbes and hence in interspecific interactions (Raaijmakers et al. 2009). For this reason, the total microbiota surrounding the mycorrhizal partners could be regarded collectively as the third member of the tripartite mutualism. In addition to species with beneficial effects on the symbiosis (e.g. MHB strains), there are likely to be others that have

Chapter 4. Ecological interactions in the biserrula rhizosphere 93 negative or neutral effects (Frey-Klett et al. 2011). In this context, the action of a single bacterial taxon may be of less significance in the establishment and persistence of mycorrhizas than the cumulative effect of the soil microbiota. Ecological interactions in the rhizosphere remain, on the whole, poorly understood (Shelef et al. 2019), and direct experimentation is needed to elucidate the effects of these interactions on the organisms responsible for forming mycorrhizas. To date, screening of candidate helper bacteria in AM systems has been based almost exclusively on axenic confrontation assays and the few published exceptions using non-sterile soil have produced inconsistent results: in these conditions, co-inoculation of AM fungi and MHB strains has been reported both to increase (Omirou et al. 2016) and to decrease (Paula et al. 1992) the magnitude of the helper effect, or to eliminate it entirely (Imperiali et al. 2017). Evidence from ectomycorrhizas (ECM) is similarly scarce. In the association between the ECM fungus Piloderma croceum and the MHB Streptomyces sp. AcH 505, the addition of a microbial soil filtrate has been reported to reduce the abundance of the helper bacterium under some soil conditions (Kurth et al. 2013). Similarly, populations of the helper strain Pseudomonas fluorescens BBc6R8 inoculated into nursery soil were found to decline rapidly in density (Frey-Klett et al. 1997). This cannot be assumed to translate into a diminished helper effect, however, as MHBs have been found to act at concentrations as low as 30 CFU g of soil dry matter-1 in nursery or field conditions (Frey-Klett et al. 1997) and an inverse dose-response relationship has been observed for BBc6RB (Frey-Klett et al. 1999). The applicability of these findings to other systems, including AM systems, is unclear. Consequently, little is known regarding the activity of helper strains in real-world arbuscular mycorrhizas. The emphasis on single-strain axenic confrontation assays means also that the potential for synergistic effects in the application of multiple MHB strains has gone largely unexplored. Exceptions to this are the report of synergistic enhancements in the root fresh weight of tomato plants when co-inoculated with two strains of Pseudomonas fluorescens (92rk and P190r) alongside the AM fungus Funneliformis mosseae BEG12 (Gamalero et al. 2004), and the failure of co-inoculated Bacillus and Rhizobium strains to produce similar synergies (Requena et al. 1997). In the latter study, co-inoculation was found to significantly impair the helper effect of a Bacillus strain on a Glomus coronatum-Anthyllis cytisoides mycorrhiza, reducing colonisation from 51% to 13%. The same effect was not observed for Rhizophagus irregularis.

Chapter 4. Ecological interactions in the biserrula rhizosphere 94 Among the many ecological interactions taking place in the rhizosphere are some that are partially understood. Many of the signalling mechanisms underlying the rhizobial symbiosis in legumes have now been identified (Wang et al. 2018) and an interaction of these signalling pathways with the mycorrhizal symbiosis has been identified. A helper effect has been observed for Rhizobium (Fernández Bidondo et al. 2016) as well as Sinorhizobium/Ensifer and Bradyrhizobium (Xie et al. 1995). In the latter study, the authors observed a similar effect following the exogenous application of isolated nodulation (Nod) factors, lipochitooligosaccharides exuded by rhizobia in response to flavonoid signals from the plant root. In the legume Medicago truncatula, this signalling exchange has been observed to stimulate the genes NFP, DMI1, DMI2, DMI3, and NSP1, components of a common symbiosis pathway shared by rhizobia and arbuscular mycorrhizas (Oláh et al. 2005). DMI1, DMI2, and DMI3 are necessary for mycorrhiza formation; DMI1 and DMI2 are also involved in growth, which may increase contacts with mycorrhizal propagules (Oláh et al. 2005). Thus rhizobia generally may be expected to act as MHBs, assuming these results hold true for host plants other than M. truncatula. The effect has also been seen in reverse: the helper strain Pseudomonas monteilli has been reported to increase nodulation in Acacia spp., perhaps by stimulation of the common symbiotic pathway (Duponnois and Plenchette 2003). In a prior experiment (section 3.3.4.), a P. koreensis subgroup taxon was found both to act as an MHB in the roots of Biserrula pelecinus L. and to stimulate production of nodule-like structures in the absence of rhizobia. It is possible that the secretion of Nod factor analogues by the P. koreensis subgroup taxon is key to its helper effect. Co- inoculation of this species with Mesorhizobium may be a viable path to enhancing the growth of biserrula in pasture. The experiments that follow were designed to address the following questions, which should bring us closer to a full understanding of the activity of mycorrhizal helper bacteria in the biserrula rhizosphere: (1) Are helper bacteria capable of working in tandem to promote mycorrhizal growth? If so, does this result in effects that are additive or synergistic? (2) Do helpers maintain their efficacy in a fully populated rhizosphere, or does the combined effect of the diverse ecological interactions in that environment (including various forms of microbial antagonism and competition) act to suppress the helper effect? (3) Do helper bacteria stimulate the common symbiosis pathways shared by mycorrhizas and rhizobia, leading to enhanced nodulation? (4) Does Mesorhizobium act as a helper bacterium, and can it work in tandem with helper

Chapter 4. Ecological interactions in the biserrula rhizosphere 95 species to facilitate enhanced mycorrhization? Unravelling the complex web of ecological interactions surrounding the mycorrhizal symbiosis in situ is a necessary precursor to the effective implementation of helper strains in real-world environments.

4.2. Materials and methods A diagrammatic outline of the experiments described here is given in Appendix 2.

4.2.1. Field site and rhizosphere soil Soil was collected from the field site described in section 2.3.1.1. Sampling took place in May 2018. At the time of collection, the paddock had been recently tilled, and only small areas of natural regrowth of biserrula were present. Approximately 40 L of soil was taken in total from four locations within the pasture. Sampling was confined to the upper 10 cm of the soil profile. Tools and containers used in sampling and collection were first disinfected with 80% ethanol. The soil was transferred to 12 pots, each measuring 16 cm in diameter, at a rate of 2.7 kg per pot. Each pot was then sown with approximately 20 seeds of Biserrula pelecinus cv. Casbah, which were first surface disinfected by soaking in a 30% solution of commercial bleach (1.3% NaOCl) with Tween 20 for 10 min and rinsed five times in sterile water. Biserrula plants were grown in a glasshouse environment and manually watered every 2-3 days with deionised water for 22 weeks, between June 2018 and November 2018. Recorded temperatures increased over the growing period, ranging from 9.6°C to 27.7°C with a mean of 21.2°C. Plants were then harvested and the roots and their adhering soil retained. A rhizosphere solution was obtained by rinsing 500 g of roots and soil in 5 L of sterile water and passing the solution through a 32 µm sieve to remove coarse material and any AM spores that might be present in the soil. At the time of harvest, the biserrula plants possessed both flowers and immature seedpods.

4.2.2. AM fungi and bacterial cultures Four bacterial strains were selected for use in these experiments, representing each of the genera previously found to harbour taxa producing a helper effect in biserrula: Arthrobacter sp. (ART1), Bacillus cereus group taxon (BAC1), Microbacterium xylanilyticum (MIC1), and P. koreensis subgroup taxon (PSE1). These strains were isolated from the rhizoplane (MIC1) and root endosphere (ART1, BAC1) of biserrula plants grown in pasture, and from the interior of surface-disinfected nodules (PSE1).

Chapter 4. Ecological interactions in the biserrula rhizosphere 96 Cultures were maintained as 30% glycerol stocks and stored at -80°C. To obtain liquid cultures for inoculation, stocks were plated onto nutrient agar (Thermo Fisher Scientific, Scoresby, Victoria, Australia), incubated at 25°C for 48 h or until sufficient growth was observed, and then transferred to nutrient broth (Thermo Fisher Scientific, Scoresby, Victoria, Australia). Commercial inoculum of Mesorhizobium ciceri bv. biserrulae WSM1497 was obtained as both a freeze-dried powder (EasyRhiz WSM1497 biserrula inoculant) and a moist formulation employing peat as a carrier (New Edge Microbials, Albury, NSW, Australia). The presence of actively growing bacteria in these products was confirmed by plating each formulation onto a medium consisting of 5 g tryptone

(BD Micro), 3 g yeast extract (Oxoid), 1.3 g CaCl2, and 15 g agar ((BD BactoAgar: Bacto Laboratories, Mt Pritchard, NSW, Australia) per L (Beringer 1974) using a ten- fold dilution series of 101-106. A solution of WSM1497 was obtained by dissolving 30 ml of the freeze-dried formulation and 100 g of the accompanying protectant in 2 L of sterile water. The final inoculation volume of 30 ml consisted of 0.1 ml of this solution suspended in sterile water. Spores of the AM fungus F. mosseae were obtained from biserrula trap cultures by a process of wet sieving and decanting adapted from Pacioni (1992; see section 3.3.3.). Spores used in the in vitro screening were trimmed so that no more than 30 µm of subtending hypha was retained and then surface disinfected prior to use according to th method described in section 2.4.1.2. When not in use, groups of 50 spores were stored in tubes containing 1 ml of sterile water at 4°C. The process of identifying these spores is described in section 3.3.3.

4.2.3. In vitro effect of MHB combinations on germination and hyphal elongation Five surface-disinfected spores of F. mosseae were arranged in a pentagon in a Petri dish containing 1% water agar (BD BactoAgar) either by itself as a control or amended with 1 ml of a bacterial solution. All possible combinations of the four selected strains were screened (Table 4-1.). The final concentration of each solution was adjusted to 1 × 107 CFU ml-1 according to the method described in section 3.2.7., with each constituent strain making an equal contribution to the colony count. Petri dishes were incubated at 25°C in the dark. Four replicate dishes were arranged in a randomised design generated with the DiGGeR extension for R (Coombes, 2002). After seven days, a germination count was taken and spores were photographed under a stereomicroscope (Nikon SMZ745T) using a Nikon DS-2Mv camera at 40-80× magnification. Measurements of

Chapter 4. Ecological interactions in the biserrula rhizosphere 97 hyphal length were obtained using ImageJ (U. S. National Institutes of Health, Bethesda, Maryland, USA) by first photographing a metal scale at the same magnification and averaging 10 measurements of 1 mm to obtain a final reference scale. A line was then fitted to the emergent hypha to provide a measurement of its length. A spore was considered to have germinated when fresh hyphal growth could be seen emerging from the cut end of the subtending hyphae or through the spore wall.

Table 4-1. Combinations of MHB strains used in the in vitro assay Bacterial strains Treatment ART1 BAC1 MIC1 PSE1 A + - - - AM + - + - B - + - - BA + + - - BM - + + - BMA + + + - M - - + - P - - - + PA + - - + PABM + + + + PB - + - + PBA + + - + PBM - + + + PM - - + + PMA + - + + Control (Ctrl) - - - - All combinations of helper bacteria screened for their in vitro effects on germination and hyphal elongation in spores of the AM fungus Funneliformis mosseae.

Chapter 4. Ecological interactions in the biserrula rhizosphere 98 4.2.4. In vivo MHB screening

4.2.4.1. Plant materials and growth conditions The efficacy of mycorrhiza helper bacteria on percent mycorrhizal colonisation of biserrula in a populated rhizosphere and the effect of co-inoculation of helper strains and M. ciceri bv. biserrulae WSM1497 on mycorrhizal colonisation and nodulation were studied in two separate in vivo trials. For each trial, biserrula plants were grown in square forestry tubes (6 cm × 18 cm) containing 800 g of a sandy loam sterilised by autoclaving twice for a total of 2 h over two days. Pots were arranged according to randomised trial designs generated using the DiGGeR extension for R (Coombes, 2002). Four pre-germinated seeds of biserrula were sowed per pot and were thinned to single plants after seedling emergence. Prior to sowing, seeds were surface-disinfected by soaking in a 30% solution of commercial bleach (1.3% NaOCl) with a drop of added Tween 20 for 10 min, then rinsed five times with sterile water and transferred to Petri dishes containing sterile, moistened filter paper. Seeds were germinated by incubating at 25°C for three days. After germination, 200 surface-disinfected spores of F. mosseae were pipetted directly onto the root surface during sowing. Pots were then inoculated as described for each trial below. Plants were grown in growth cabinets (Conviron Adaptis, Conviron Asia Pacific, Melbourne, Australia) under a 12/12 h daily light/dark cycle at temperatures of 24/14°C and a relative humidity of 75%. Plants were watered as necessary with sterile water.

4.2.4.2. In vivo effect of MHB strains on mycorrhizal colonisation in a populated biserrula rhizosphere Following mycorrhizal inoculation, each pot received one of the treatments described in Table 4-2, namely a bacterial inoculation (at a rate of approximately 1 × 107 CFU ml-1) with or without an additional rhizosphere solution. For all microbial inoculations, a volume of 30 ml was used. In pots that did not receive the rhizosphere inoculation, 30 ml of sterile water was added instead. One pot in each of four replicates received only the rhizosphere solution and one received only sterile water as a control. After 15 weeks, plants were removed from their pots and the roots isolated and rinsed to remove all soil. The roots were subsequently cleared and stained according to a method adapted from Brundrett (2008) and described in section 2.3.1.3. This process yielded stained root fragments of 1 cm in length. Percent mycorrhizal colonisation (PC)

Chapter 4. Ecological interactions in the biserrula rhizosphere 99 was assessed by the gridline intersection method (Giovannetti and Mosse 1980). As previously described (section 2.3.1.3.), stained roots were transferred to 90 mm Petri dishes marked with a 1.27 cm grid and viewed under a stereomicroscope (Nikon SMZ745T) at 40-80× magnification. Gridlines were scanned vertically and horizontally and roots marked as either mycorrhizal or non-mycorrhizal where they intersected with the gridlines. Arbuscules, vesicles, spores, and mycelium were counted individually. In total, 100 observations were taken from each root system. Where necessary, roots were discarded and additional roots drawn from the sample to ensure the total of 100 observations was met.

Table 4-2. Treatments used in the in vivo populated rhizosphere screening Treatment Bacterial strain Rhizosphere solution ART1 Arthrobacter sp. - BAC1 Bacillus cereus group taxon - MIC1 Microbacterium xylanilyticum - PSE1 Pseudomonas koreensis subgroup taxon - RHI None + ART1×RHI Arthrobacter sp. + BAC1×RHI Bacillus cereus group taxon + MIC1×RHI Microbacterium xylanilyticum + PSE1×RHI Pseudomonas koreensis subgroup taxon + Control None - All treatments screened for their in vivo effects on mycorrhizal colonisation in biserrula in the populated rhizosphere screening

4.2.4.3. In vivo effect of MHB strains and mesorhizobial inoculation on nodule- formation in biserrula Each pot received one of the treatments described in Table 4-3, namely a bacterial inoculation (at a rate of approximately 1 × 107 CFU ml-1) with or without a solution of WSM1497. For all microbial inoculations, a volume of 30 ml was used. In pots that did not receive the mesorhizobial inoculation, 30 ml of sterile water was added instead. One

Chapter 4. Ecological interactions in the biserrula rhizosphere 100 pot in each of four replicates received only the WSM1497 solution and one received only sterile water as a control. After 15 weeks, plants were removed from their pots and the roots isolated and rinsed to remove all soil. Nodules and nodule-like structures (NLSs) were counted, then removed from the roots and dissected. The interior colour was used to identify each sample as either active (pink or red), inactive (white), or potentially senescent (green), according to the established methodology (Pate 1958). Percent mycorrhizal colonisation was assessed according to the methodology described above.

Table 4-3. Treatments used in the in vivo rhizobial inoculation trial Treatment Bacterial strain WSM1497 inoculation ART1 Arthrobacter sp. - BAC1 Bacillus cereus group taxon - MIC1 Microbacterium xylanilyticum - PSE1 Pseudomonas koreensis subgroup taxon - MES None + ART1×MES Arthrobacter sp. + BAC1×MES Bacillus cereus group taxon + MIC1×MES Microbacterium xylanilyticum + PSE1×MES Pseudomonas koreensis subgroup taxon + Control None - All treatments screened for their in vivo effects on formation of nodules and nodule- like structures in biserrula in the rhizobial inoculation screening

4.2.5. Nodulation (nod) factors DNA extracted from each of the bacterial strains, as well as from the commercial Mesorhizobium inoculum, was screened for the presence of nod factor components using primers targeting the nodC and nodD genes (Table 4-4.). For the helper strains, DNA was extracted from cultures grown for 48 h in nutrient broth (Oxoid) using the Sigma GenElute Bacterial Genomic DNA Kit (Sigma-Aldrich Pty Ltd, Castle Hill, NSW, Australia) according to the manufacturer’s instructions. Mesorhizobium DNA was obtained from commercial freeze-dried and peat-based inocula using the DNeasy

Chapter 4. Ecological interactions in the biserrula rhizosphere 101 PowerLyzer PowerSoil Kit (QIAGEN, Chadstone, Victoria, Australia) according to the manufacturer’s instructions. In all cases, 2 µl of DNA template was combined with 4 µl of 5× MyTaq reaction buffer (Bioline, London, UK), 10 µl of sterile deionised water, and 2 µl of a 5 µM stock of each primer, to make a final reaction volume of 20 µl. PCR products were visualised by agarose gel electrophoresis using GelRed (Biotium) as a stain. Additionally, a search of the Pseudomonas Genome Database (Winsor et al. 2016) was conducted using variations on the terms nod and nodule and targeting members of the P. koreensis subgroup to determine whether any evidence of nod factors could be identified in annotated genomes.

Table 4-4. Nodulation factor primers used in this study Target Primers Sequences PCR Conditions NodCfor540 5’-TGATYGAYATGGARTAYTGGCT-3’ (Sarita et al. 2005) nodC NodCrev1160 5’-CGYGACARCCARTCGCTRTTG-3’ NodCF 5’-AYGTHGTYGAYGACGGTTC-3’ (Sarita et al. 2005) nodC NodCI 5’-CGYGACAGCCANTCKCTATTG-3’ NBA12 5’-GGATSGCAATCATCTAYRGMRTGG-3’ (Laguerre et al. 1996) nodD NBF12 5’-GGATCRAAAGCATCCRCASTATGG-3’ Primers targeting genes involved in the synthesis of nodulation factors

4.2.6. Statistical analyses All statistical analyses were carried out using version 3.6.0 of the R statistical software program (R Core Team 2019). In vitro germination counts were analysed by Poisson regression. Pairwise comparison of estimated marginal means derived from this model was carried out using Tukey’s method in the package emmeans (Lenth 2019). Tukey- adjusted p values < 0.05 were considered significant. To determine the effect of combining helper strains on in vitro hyphal growth, a single mean value for hyphal length for each Petri dish was obtained and subjected to one-way ANOVA using the aov function in R. Values were log-transformed prior to analysis. Untransformed percent colonisation and nodule count data from both in vivo trials were similarly subjected to one-way ANOVA using the aov function in R. Outliers and zero values were removed from the data where necessary. Multiple pairwise comparisons were

Chapter 4. Ecological interactions in the biserrula rhizosphere 102 performed using Tukey’s honest significant difference test in the package agricolae (de Mendiburu 2019), which was also used to visualise results. As before, Tukey-adjusted p values < 0.05 were considered significant. The suitability for ANOVA of all data obtained in this experiment was assessed by means of Bartlett’s test of homoscedasticity and the Shapiro-Wilk test of normality, using the bartlett.test and shapiro.test functions in R.

4.3. Results

4.3.1. In vitro effect of MHB combinations No single strain or combination of strains was found to significantly enhance germination in vitro, contrary to prior experimental results (section 3.3.2.), perhaps owing to the reduced sample size, which impeded estimation of statistical significance. By contrast, hyphal elongation was enhanced in 14 of the 15 bacterial single-strain or combination treatments relative to the control (Figure 4-1.). Only the single-species Arthrobacter (A) treatment failed to enhance hyphal elongation relative to the control, despite previously exhibiting an in vitro promotional effect (section 3.3.2.). The highest value for hyphal elongation was obtained in the P. koreensis subgroup taxon × B. cereus group taxon (PB) treatment. The log-transformed mean value for that treatment (0.65) was greater than the sum of the respective means for the single-strain Pseudomonas (0.33) and Bacillus (0.12) treatments, pointing to the possibility of a synergistic effect between those two taxa. No significant differences were observed between the remaining P. koreensis subgroup taxon combinations. Similarly, the analysis did not identify any significant differences between the single-strain B. cereus group taxon treatment and any of its combinations (aside from PB), however, two treatments containing Arthrobacter (BA and BMA) showed a diminished helper effect relative to treatments containing the Pseudomonas strain (PB, PABM, PBM, PBA). The promotional effect of M. xylanilyticum (M) was enhanced when paired with both P and B (in PABM and PBM) but not when paired with P or B alone. Arthrobacter, which did not produce a helper effect on its own, was found to enhance hyphal growth relative to the control in all of its combinations. Significant differences were observed between the single-strain Arthrobacter treatment and the combined treatments PA, PBM, and PABM, but not between A and the treatments AM, BA, BMA, or PMA.

Chapter 4. Ecological interactions in the biserrula rhizosphere 103 Figure 4-1. The effect of mycorrhiza helper bacteria on the in vitro hyphal growth of Funneliformis mosseae when applied singly or in combination, showing (a) log-transformed means ± SE, and (b) log-transformed means and ranges with significance groups. Treatments on the x axis are identified in Table 4-1. Groups are significantly different according to Tukey’s honest significant difference test (p < 0.05) where they do not share a letter.

Chapter 4. Ecological interactions in the biserrula rhizosphere 104 4.3.2. The helper effect in a populated rhizosphere All four MHB strains were found to increase percent mycorrhizal colonisation in biserrula, regardless of whether they were inoculated singly or co-inoculated with a solution of rhizosphere soil (Figure 4-2.). The rhizosphere solution also produced a significant mean increase in colonisation rate relative to the control when inoculated in the absence of a helper strain. When co-inoculated with a helper strain, however, the rhizosphere soil solution produced inconsistent results. While the P. koreensis subgroup taxon and M. xylanilyticum were unaffected by the solution, both Arthrobacter sp. and the B. cereus group taxon demonstrated reduced efficacy in promoting mycorrhizal colonisation in co-inoculation treatments. Mean percent colonisation values for Bacillus, Arthrobacter, and Microbacterium strains co-inoculated with the rhizosphere solution were not significantly different to the values recorded for the rhizosphere solution alone. No significant differences were observed in the relative frequency of individual mycorrhizal components (arbuscules, vesicles, spores, or mycelia). Counts of these separate components were combined to produce a single percent colonisation score for each root system examined.

4.3.3. The effect of MHB strains on nodule formation Nodule-like structures (NLSs) were not observed in the ART1, BAC1, or MIC1 treatments, which did not receive an inoculation of M. ciceri bv. biserrulae, nor in the control treatment, which received neither rhizobia nor bacteria (Table 4-5.). Formation of NLSs was observed in response to inoculation with PSE1 in the absence of rhizobia, but all specimens examined in this treatment were inactive (white). The number of inactive NLSs in this treatment was not significantly different to the number observed in the MES and BAC1×MES treatments, both of which received rhizobial inoculum. Co- inoculation of PSE1 with Mesorhizobium was observed to significantly increase the number of inactive nodules/NLSs relative to Mesorhizobium alone and to the remaining co-inoculation treatments, as well as to significantly increase the total nodule/NLS count relative to the other bacterial treatments, but did not increase the number of active (pink) nodules. The proportion of nodules showing evidence of activity remained consistent across treatments. No senescent (green) nodules were observed.

Chapter 4. Ecological interactions in the biserrula rhizosphere 105 4.3.4. The effect of Mesorhizobium on mycorrhizal colonisation Inoculation of biserrula with Mesorhizobium alone resulted in an increase in percent mycorrhizal colonisation on par with Arthrobacter sp., M. xylanilyticum, and the B. cereus group taxon (Figure 4-3.). Co-inoculation of helper strains with Mesorhizobium did not significantly affect the magnitude of the observed helper effect in any treatment. All single strains and co-inoculation treatments under study enhanced mycorrhizal colonisation in vivo.

4.3.5. Nodulation (Nod) factors PCR products were observed in Mesorhizobium strains derived from both the freeze- dried and peat-based inoculants using the nodC primers NodCfor540-NodCrev1160. These products measured approximately 600 bp in length, consistent with expectations (Sarita et al. 2005). No products were obtained in Mesorhizobium using the remaining primer sets and no products were obtained from the non-rhizobial helper strains using any of the three primer pairs. Database searches returned no evidence of nod genes in annotated P. koreensis subgroup genomes.

Table 4-5. Effect of bacterial and mesorhizobial inoculation on nodule count Treatment nodW nodP nodT %Active ART1 - - - - BAC1 - - - - MIC1 - - - - PSE1 3.50c - 3.50c - MES 12.75bc 28.50a 41.25ab 68.09a ART1×MES 16.50b 13.00a 29.50b 45.53a BAC1×MES 11.25bc 20.75a 32.00b 64.96a MIC1×MES 16.75b 19.00a 35.75b 49.22a PSE1×MES 30.50a 26.25a 56.75a 46.41a Control - - - - The effect of inoculating mycorrhiza helper bacteria and Mesorhizobium ciceri bv. biserrulae on formation of nodules or nodule-like structure in biserrula. Columns are inactive nodules (nodW), active nodules (nodP), total nodules (nodT), and percentage of total nodules which were active (%Active). Treatments are identified in Table 4-3. Values in each column sharing a superscript letter are not significantly different according to Tukey’s honest dignificant difference test (p < 0.05). Treatments returning zero values were omitted from the analysis.

Chapter 4. Ecological interactions in the biserrula rhizosphere 106

Figure 4-2. The effect of helper bacteria on mycorrhizal colonisation in biserrula when inoculated singly or with a rhizosphere solution, showing (a) mean percent colonisation value ± SE, and (b) means, ranges, and significance groups. Treatments on the x axis are identified in Table 4-2. Groups are significantly different according to Tukey’s honest significant difference test (p < 0.05) where they do not share a letter.

Chapter 4. Ecological interactions in the biserrula rhizosphere 107

Figure 4-3. The effect of helper bacteria and Mesorhizobium ciceri bv. biserrulae on percent mycorrhizal colonisation in biserrula when inoculated singly or in co-inoculation treatments, showing (a) mean percent colonisation value ± SE, and (b) means, ranges, and significance groups. Treatments on the x axis are identified in Table 4-3. Groups are significantly different according to Tukey’s honest significant difference test (p < 0.05) where they do not share a letter.

4.4. Discussion The results described here reaffirm previous observations that the taxa under study are capable of promoting the colonisation of biserrula by the AM fungus F. mosseae. Despite the failure of these species to promote spore germination in vitro in the present study, and the failure of Arthrobacter sp. to promote in vitro hyphal elongation, all four strains acted as mycorrhiza helper bacteria in vivo, regardless of co-inoculation

Chapter 4. Ecological interactions in the biserrula rhizosphere 108 treatment. Additionally, a single combination, P. koreensis subgroup taxon × B. cereus group taxon, was found to enhance hyphal elongation in vitro in a manner suggestive of a synergistic interaction. Aside from a single prior observation involving two strains of P. fluorescens (Gamalero et al. 2004), the potential for synergies between helper strains has not previously been investigated. While the mechanisms underlying the helper effect have not been elucidated, the capacity for two unrelated bacterial taxa (one Gram positive, one Gram negative) to produce a synergistic effect of this kind may reflect differences in the means by which they exert their effects, suggesting that multiple compatible methods of enhancing AM fungal growth can act simultaneously. Complicating this interpretation is the fact that the sole prior observation of a synergistic helper effect occurred in two strains of the same species, and that comparable effects were not observed between other similarly disparate species, such as the Gram-negative Pseudomonas with the Gram-positive Microbacterium or Arthrobacter. Nevertheless, while caution is clearly warranted in pronouncing any given taxon an MHB on the basis of in vitro experimentation alone, given the failure of the taxa under study to demonstrate those effects in the present case, the observed synergy offers a promising avenue towards future efforts to optimise the mycorrhizal symbiosis in biserrula. The capacity of these MHB species to retain their helper effects in a populated rhizosphere was similarly promising. The ability of the rhizosphere solution to enhance mycorrhizal colonisation on its own was an unexpected result, but is consistent with earlier observations that the beneficial effects of AM colonisation may depend for their greatest expression on the existence of a complex microbial background (Hoeksema et al. 2010). In spite of this, it was noted that co-inoculation of the soil with both a rhizosphere filtrate and an MHB strain occasionally produced a helper effect that was diminished relative to the MHB strain alone. This was true of both the B. cereus group taxon and the Arthrobacter species. This suggests that competitive effects may be significant in determining whether a species is capable of exerting a helper effect in real-world conditions. Recent evidence points to a role for diffusible volatile organic compounds (VOCs) in interspecific bacterial interactions (Schmidt et al. 2015). These interactions may be beneficial or antagonistic, or they may exhibit aspects of both tendencies. For example, VOCs exuded by Collimonas pratensis and Serratia plymuthica have been shown to promote the growth and motility of P. fluorescens Pf0-1 and simultaneously to stimulate the production of secondary metabolites antagonistic to

Chapter 4. Ecological interactions in the biserrula rhizosphere 109 Bacillus spp. (Garbeva et al. 2014). Chemical signalling using VOCs has also been observed between fungi and bacteria. Fungal VOCs exuded by the pathogenic species Fusarium culmorum have been shown to induce changes in gene expression in the rhizosphere bacterium S. plymuthica PRI-2C (Schmidt et al. 2017). Similarly, bacterial VOCs produced by Pseudomonas aeruginosa enhance the growth of Aspergillus fumigatus (Briard et al. 2016). This phenomenon may account for the promotional effect of MHB strains in axenic confrontation assays. In a populated rhizosphere, however, the activity of a bacterium is likely to depend on the total composition of all diffusible volatiles in the rhizosphere (Hol et al. 2015). These may modify or, potentially, counteract the helper effect. It should also be noted that the taxa under study were isolated from the of biserrula plants that showed no evidence of a helper effect: while the plants harbouring the helper strains were not themselves assessed for percent colonisation, the mean percent colonisation value obtained for the pasture from which they were collected was 19.2 ± 4.6% (section 2.6.1.). The reasons for this disparity are unclear, but may reflect the dosage employed in the present study, approximately 30 × 107 CFU, greatly in excess of the concentrations expected in natural conditions. This saturation of the rhizosphere may have magnified the helper effect and helped to absorb the impacts of interspecific competition. The effects of dosage on the helper effect remain unclear. While an earlier study in an ECM system reported a significant helper effect in response to bacterial concentrations as low as 30 CFU g of soil dry matter-1, and not at the initial peak of 4 × 104 CFU g of soil dry matter-1, the precise timing of this effect was unclear (Frey-Klett et al. 1997). The authors note that a non-significant divergence in rates of mycorrhizal colonisation between inoculated and control plants began earlier, when a higher concentration of the helper strain was still present in the soil. Thus it is possible that the promotion of mycorrhization by the MHB was induced by the higher dosage of the helper strain but that this effect only met the threshold for statistical significance after a further period of decline. The role of dosage in the timing of helper effects remains to be fully investigated. Another factor in the inconsistent effect of MHB strains may be the localisation of the bacterium within the plant root. The Pseudomonas strain, for example, was isolated largely from the interior of nodules. With nodules scarce in the in vivo treatments, the strain may have associated more generally with the root and may therefore have exerted a greater influence over the behaviour of the root or the AM fungus. Complicating this

Chapter 4. Ecological interactions in the biserrula rhizosphere 110 is the fact that co-inoculation with rhizobia, and the consequent increase in nodule count, did not diminish the helper effect. A further consideration concerns the stability of the rhizosphere over the growing season. The inoculum used in the present study represented a snapshot of the biserrula at a particular point in the life cycle of biserrula (flowering/seeding). In this respect, it may not adequately represent the dynamics of the rhizosphere. While rhizosphere communities have been described as quasi-stationary, as they tend to reach a relatively static equilibrium (Kuzyakov and Razavi 2019), this can be disrupted by changes in environmental conditions (Huang et al. 2018, Moreno- Espindola et al. 2018) or by changes in the developmental stage of the host plant (Chaparro et al. 2014). In a pasture setting, this may entail an exchange of bacteria between the rhizosphere and the bulk soil; in the present study, with no bulk soil diversity to draw from, this exchange may not have been possible. Consequently, helper bacteria deployed in a pasture setting may be exposed to a more dynamic environment than this experiment could simulate. The close association of the Pseudomonas strain with nodules was confirmed by the MHB × M. ciceri bv. biserrulae WSM1497 axenic co-inoculation experiment. Not only did Pseudomonas induce the formation of nodule-like structures in the absence of the mesorhizobial inoculant, as previously observed (section 3.3.4.), it also significantly increased the number of inactive nodules relative to WSM1497 alone, and increased the total number of nodules relative to the remaining bacterial co-inoculation treatments. Perhaps surprisingly, it did not significantly increase the number of active nodules in the root system. It is possible that the dosage of WSM1497 was insufficient or that the inoculum was inadequate in some way (growth on tryptone-yeast-CaCl2 medium was observed to be poor), but this has not been confirmed. Alternatively, it is possible that the level of nitrogen already present in the soil, which was not measured, was sufficient to suppress nodule activity. High soil N has been observed to suppress nodule formation (Friel and Friesen 2019, Xia et al. 2017); however, it is unclear how this phenomenon affects the activity of NLSs induced by a non-rhizobial bacterium. Nevertheless, these results support the suggestion that some helper bacteria exert their effects via stimulation of the common symbiotic pathway, with the caveat that this phenomenon has not yet been demonstrated in real-world conditions. In support of this, an attempt was made to identify genes responsible for the production of Nod factor lipochitooligosaccharides in the helper strains. While the negative results for Arthrobacter, Bacillus, and Microbacterium were expected, the failure to identify nodC

Chapter 4. Ecological interactions in the biserrula rhizosphere 111 or nodD genes in Pseudomonas was perhaps surprising. Nod factor analogues have previously been identified in non-rhizobial taxa of the genera Methylobacterium and Burkholderia using primers targeting components of the nodABC complex (Moulin et al. 2001, Sy et al. 2001); the failure to locate them in Pseudomonas may indicate that they are absent – a possibility given added weight by the absence of annotated nod factor genes in public genomes – or that they are sufficiently distinct that established primers cannot amplify them. It is also noteworthy that only nodC genes, and not nodD genes, were identified in Mesorhizobium, and that one nodC primer pair failed to amplify. This suggests that the design of the established primers and the choice of targets may account for the negative findings. Correcting this deficiency to allow for the identification of Nod factor components in a wider range of host species should be a target for future Nod factor research.. Finally, it was observed that Mesorhizobium by itself was capable of acting as an MHB strain. This is the first observation of a helper effect in Mesorhizobium. It follows an earlier study describing a host-specific but generally negative effect of members of Mesorhizobium on the in vitro growth of strains of the ECM fungus Laccaria parva (Obase 2019). While the effect was less pronounced than that observed for Pseudomonas with or without rhizobia and Bacillus with added rhizobia, it was comparable in magnitude to the effect produced by the remaining single MHB strains and MHB × Mesorhizobium treatments. Mesorhizobium thus joins Rhizobium, Sinorhizobium/Ensifer, and Bradyrhizobium as nitrogen-fixing and nodule-forming genera also found to stimulate mycorrhizal colonisation. Methylobacterium and Burkholderia also belong to this complex (Kim et al. 2010, Deveau et al. 2007, Moulin et al. 2001, Sy et al. 2001). The effect observed for Mesorhizobium – an approximate doubling of mean colonisation rate – was similar to that observed for Rhizobium etli (Fernández Bidondo et al. 2016) and Sinorhizobium (Ensifer) fredii strain NGR234 (identified as Rhizobium sp. NGR234; Xie et al. 1995), and slightly more pronounced than the effect exerted by Bradyrhizobium (Xie et al. 1995). Co-inoculation of an MHB with Mesorhizobium was not observed to enhance the helper effect. This is perhaps surprising in the cases of Arthrobacter, Bacillus, and Microbacterium, as those species show no evidence of interacting with the common symbiotic pathway and likely enhance mycorrhizal colonisation via different means. It is consistent, however, with the prior observation of a decrease in the helper effect for a Bacillus strain co-inoculated with Rhizobium (Requena et al. 1997). The finding of a helper effect in Mesorhizobium

Chapter 4. Ecological interactions in the biserrula rhizosphere 112 carries the same caveat identified above: Mesorhizobium was present in a pasture soil that showed no evidence of a helper effect. While the studies described here identify previously unknown interactions occurring in the biserrula root system, the unexpected or inconsistent nature of some of these findings points to an ongoing need to elucidate the complex workings of the rhizosphere.

4.5. Conclusion In summary, the experiments described here provide additional support to the idea that certain plant-associated bacteria are capable of promoting mycorrhization in biserrula, and that this effect can be maintained, albeit with some reduction in efficacy in certain cases, in a populated rhizosphere. This study further identifies M. ciceri bv. biserrulae as a candidate MHB, as it was shown to enhance mycorrhization of biserrula in vivo, although not in the presence of a populated rhizosphere, and it did not enhance the helper effect exerted by co-inoculated MHB strains. It was also demonstrated that a rhizosphere soil filtrate may itself exert a beneficial effect on the establishment of a mycorrhizal symbiosis, thus acting, collectively, as an MHB consortium. Additionally, the proposition that certain combinations of MHB strains may act synergistically was investigated with positive in vitro results for one treatment, P. koreensis subgroup taxon × B. cereus group taxon. While it was not in scope of the present study to explore this avenue fully, it represents a promising means by which the mycorrhizal symbiosis may be optimised in a host plant such as biserrula. These findings also raised more questions that must be answered before the identified helper bacteria can be implemented practically. The most pressing of these concerns is the apparent efficacy of the MHB strains in a populated rhizosphere, when observational evidence from biserrula pastures could not identify an active helper effect. Several explanations for these results have been suggested, namely the increased dosages employed in the study, possible changes in the localisation of bacteria within the plant root relative to real-world conditions, and differences in the dynamics of the rhizosphere between real and simulated environments. While these studies supply evidence to suggest that microbial competition does not necessarily impede the helper effect, it is necessary to develop a greater understanding of how the biserrula rhizosphere operates in a pasture setting. To this end, an experiment was designed to explore the dynamics of root-associated microbial communities in a two-season biserrula-wheat rotation.

Chapter 4. Ecological interactions in the biserrula rhizosphere 113 Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. across a two-season legume-cereal rotation

5.1. Introduction The global drive to increase agricultural production has fuelled interest in the potential for soilborne microorganisms to improve plant access to soil nutrients (Trivedi et al. 2017, Parnell et al. 2016). These microorganisms, including plant growth-promoting rhizobacteria (PGPR) and mycorrhizal fungi, exert their influence by improving host plant nutrition or by conferring resistance to biotic and abiotic stresses (Raaijmakers et al. 2009). In specialised cases, such as the mycorrhiza helper bacteria (Frey-Klett et al. 2007), they may exert this influence indirectly, by promoting the growth of other beneficial microorganisms. The study of these organisms is not new: cultures of Rhizobium, for example, have been employed to enhance fixation of nitrogen by agricultural legumes for more than a century (Bashan 1998); however, there has been a recent move to commercialise many PGPR species in the creation of new biofertilisers (Bashan et al. 2014). These new technologies must overcome a number of challenges if they are to be effectively implemented. In addition to questions of practicality, any new biofertiliser must be both efficacious and versatile: that is, the product must be able to maintain its effectiveness between laboratory testing and field implementation and it must respond appropriately to changes in the plant-soil-microbe interaction over time (Parnell et al. 2016). This formulation reiterates a position that has been argued for some time: that an understanding of the complex ecology of the rhizosphere is necessary for the success of any novel biofertiliser or biocontrol agent (Whipps 2001). The rhizosphere is a chemically, physically, and microbiologically distinct environment surrounding the root of a plant (Barea et al. 2005). Microorganisms are recruited to the root environment by a combination of abiotic factors (including soil pH and moisture availability) and biotic processes such as rhizodeposition (Philippot et al. 2013). Rhizodeposition enriches the rhizosphere with sugars, amino acids, organic acids and a range of other compounds, including plant hormones and other molecules

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 114 involved in signalling (Dennis et al. 2010). The constitution of rhizodeposits varies between plant taxa and consequent variations in root microbiota have been observed at the level of plant family, species, and genotype (Bakker et al. 2013, Kristin and Miranda 2013). Additionally, the structure of the root-associated microbial community may change with the developmental stage of the host (Chaparro et al. 2014). In a sense, roots can be said to curate a subset of the total bulk soil microbiota adapted to the conditions in the root environment. The organisms that respond positively to this curation make up a plant’s “rhizosphere zoo” (Buée et al. 2009) and participate in a complex web of plant-microbe and microbe-microbe interactions. Applying a biofertiliser may thus amount to introducing a new species or group of species to an environment that is already extensively colonised and curated. To meet the requirements of efficacy and versatility, a beneficial microbe or microbial community must be able to find a niche in this environment and maintain its positive effect on plant growth despite changes in the surrounding microbiota and, in many farming systems, changes in the host plant. This is further compounded by the fact that methods of soil cultivation and seasonal factors can also influence the structure of the rhizosphere community (Huang et al. 2018, Moreno- Espíndola et al. 2018, Gdanetz and Trail 2017, Yin et al. 2017, Yurgel et al. 2017). The selection of suitable organisms is complicated by our poor knowledge of the ecology of the rhizosphere. Our understanding of the microbial community composition of the rhizosphere has historically been limited by the efficacy of culture-dependent methods, which give a highly selective and artificial view of the rhizosphere (Rappé and Giovannoni 2003). The development of technologies for soil metabarcoding (Orgiazzi 2015) offers a valuable corrective, allowing the identification of recalcitrant taxa in environmental samples. The proportion of species that cannot be recovered in conventional cultures is often estimated at 99% or greater (Burke et al. 2008, Leveau 2007), however direct comparisons between communities using culture-dependent and culture-independent methods are scarce and are often based on older technologies that produce correspondingly low yields of DNA (Shokralla et al. 2012). A molecular study of bacterial diversity in a Scottish pasture system yielded 275 DNA sequences (McCaig et al. 1999), for example, whereas the application of next-generation sequencing technologies permits the recovery of millions of sequences from equivalent samples (Orgiazzi 2015). While the direct study of potentially beneficial traits often demands in vitro methods, it is important that these should be understood in the context of the total rhizosphere community. The comparison of communities derived by culture-dependent

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 115 methods with those derived by direct DNA sequencing should help to illuminate what is lost when studies are restricted to controlled environments. In the present study, metabarcoding techniques were applied to bacteria from cultured and non-cultured samples of rhizosphere soil taken from a simulated legume- cereal rotation in order to determine (1) whether the host plants selectively curated a stable rhizosphere microbiota; (2) whether this microbiota persisted across seasons of the rotation, despite variations in host plant; and (3) how apparent community structure was affected by culturing. This last comparison, which has been carried out to determine the extent to which culturing misrepresents the true composition of a microbial community, may help to explain why many promising biofertiliser strains fail in the transition from laboratory to field settings (Trivedi et al. 2017). The host plants selected as models for this study were Biserrula pelecinus L. cv. Casbah (biserrula) and Triticum aestivum L. var. EGA Gregory (wheat). Biserrula is a pasture legume of Mediterranean origin, first collected for agronomic assessment in 1993 and first commercialised in Australia in 1997 (Loi et al. 2014, Nichols et al. 2012). It is one of 47 commercialised legume species currently grown in Australia and one of 30 to have first been domesticated in Australia (Nichols et al. 2012). In common with many of these legumes, biserrula is utilised for its association with N2-fixing rhizobia and is often grown in ley-farming systems in rotation with wheat; consequently, any organism applied as biofertiliser will have to contend with regular changes in host plant. As biserrula is of interest primarily in low-rainfall agriculture and soils prone to abiotic stresses (Loi et al. 2014), it represents a prime target for microbial enhancement. It is hoped that this study will provide baseline knowledge concerning the effects of host- plant succession on the soil microbial community in a legume-crop rotation.

5.2. Materials and methods

5.2.1. Field site and soil sampling Soil was collected from the field site described in section 2.3.1.1. Sampling took place in May 2018. At the time of collection, the paddock was tilled but unsown, with small areas of natural regrowth of biserrula. Approximately 50 L of soil was collected from each of four widely spaced locations within the pasture, ensuring that sampling locations were away from paddock boundaries and heavily trafficked areas. Each collection was subsequently assigned to a replicate in a pot-based rotation experiment.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 116 Soil was extracted from the upper 10 cm of the profile and homogenised by passage through a 12.7 mm square steel mesh and manually mixing. A subsample of approximately 25 ml was taken from each of the four reps and stored at -20°C prior to DNA extraction. The soil was then transferred to pots measuring 18 cm in height and diameter (2.7 kg per pot), which were sealed at the base to retain soil. All tools used in soil processing were disinfected in advance and between replicates either by soaking in a solution of 30% commercial bleach (1.3% available NaOCl) for 1 hour or by washing in water and spraying with 80% ethanol. Containers used to store soil were first disinfected with 80% ethanol.

5.2.2. Plant material and growth conditions Seeds of B. pelecinus var. Casbah and T. aestivum var. EGA Gregory were obtained from an existing depository at Charles Sturt University (Wagga Wagga, NSW, Australia). Seeds were surface disinfected before germination by transferring them to sterile 50 ml specimen containers containing 30 ml of a 30% solution of commercial bleach (1.3% available NaOCl). Containers were gently shaken on a magnetic stirrer for 10 (biserrula) or 15 (wheat) minutes, then rinsed in five changes of sterile water before being transferred to 90 mm Petri dishes containing filter papers pre-moistened with sterile water. Seeds were sown after one week of germination. Final plant densities were based on common sowing rates of the two plant species, namely 60 kg ha-1 for wheat (Department of Environment and Primary Industries, 2012) and 7 kg ha-1 for biserrula (Loi 20017), corresponding to 3 plants per pot for wheat and 12 plants per pot for biserrula. Pots were sown at higher densities and thinned back to these field numbers after seedling emergence. Plants were grown in a glasshouse environment and manually watered every 2-3 days with deionised water. Plants were harvested shortly before the onset of flowering (Figure 5-1.). The trial took place between June and November of 2018.

5.2.3. Trial design Pots were arranged in four replicates according to a randomised two-stage factorial design generated by the DiGGeR extension for R (Coombs 2002) (Figure 5-2.). The trial was divided into two ‘seasons’ or growing periods, the first lasting 11 weeks and the second 10 weeks. All plants were harvested at or near the onset of flowering to minimise the known effect of a host plant’s developmental stage on its associated

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 117 microbiome (Chaparro et al. 2014). Soil was sampled for total microbial DNA extraction after each of the two seasons.

5.2.3.3. First season and sampling Four replicates, each consisting of three pots of each of the three treatments (no host, biserrula, and wheat) were grown for a period of 11 weeks, after which the pots were sampled. Harvesting was carried out by taking three cores from each pot to a depth of 10 cm using cork borers pre-sterilised with 80% ethanol. In the plant treatments, cores were taken from within the root zone of the plant and consisted principally of root material and adhering soil. The three cores were combined to produce a single sample from each pot. Samples from each of the three pots per treatment were then bulked to produce a single 50 ml sample for each replicate. The first growing period extended from June to August of 2018.

5.2.3.4. Second season and sampling The sampled pots from the first growing period were retained and re-sown to simulate the second season of a pasture rotation. To simulate each possibility in this rotation, the three pots belonging to each of the three initial treatments were assigned to one of three additional treatments – nil, biserrula, or wheat – to produce a total of nine treatments per rep: nil × nil, nil × biserrula, nil × wheat, biserrula × nil, biserrula × biserrula, biserrula × wheat, wheat × nil, wheat × biserrula, and wheat × wheat. Plants were grown for an additional 10 weeks prior to sampling. For this second season, 25-ml samples were taken from each of the nine treatments. The second growing period extended from August to November of 2018.

5.2.4. Cultured and non-cultured samples To determine the effects of culturing on bacterial community structure, the soil samples taken at the end of the first growing period were divided to produce two 25-ml subsamples, one to be extracted directly and one to be enriched by incubation in lysogeny broth (LB-Lennox) (Lennox 1955) for 24 hours at 25°C prior to DNA extraction. Thus, for each of the three first-season treatments (nil, biserrula, and wheat), cultured and non-cultured equivalents were produced. In total, 64 soil samples were taken for DNA extraction (Table 5-1.). All samples were stored at -20°C prior to DNA

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 118 extraction, with the exception of the samples to be cultured, which were transferred to LB-Lennox medium immediately after collection and extracted after 24 hours.

Figure 5-1. Plants of biserrula (left) and wheat (right) prior to harvesting. Harvesting was carried out before the onset of flowering. First-season plants were grown for 11 weeks and second-season plants for 10 weeks to account for differences in flowering time.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 119 Figure 5-2. The arrangement of the glasshouse rotation trial. Coloured markers indicate first-season treatments: biserrula (orange), wheat (blue), and no host (white). Replicates are in 3 × 3 blocks. A fifth replicate was harvested but was not sequenced. The watering system shown here was not used during the study.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 120 Table 5-1. Samples collected during the rotation trial Sample Description Initial samples taken from each of four soil collections (one for each Y0 replicate of the pot trial). Samples taken from each of three treatments (no host, biserrula, wheat) at the end of the first growing period. Four replicates of each treatment Y1 were taken. Samples were further subdivided to produce cultured (Y1CU) and non-cultured (Y1NC) samples. Y1NILNC Y1 non-cultured no host. (Also identified as Y1 no host). Y1BISNC Y1 non-cultured biserrula. (Also identified as Y1 biserrula). Y1WHNC Y1 non-cultured wheat. (Also identified as Y1 wheat). Y1NILCU Y1 cultured no host. Y1BISCU Y1 cultured biserrula. Y1WHCU Y1 cultured wheat. Samples taken from each of nine treatments (described below) at the end Y2 of the second growing period. Four replicates of each treatment were taken. Y2NN Y2 no host × no host. Y2NB Y2 no host × biserrula. Y2NW Y2 no host × wheat. Y2BN Y2 biserrula × no host. Y2BB Y2 biserrula × biserrula. Y2BW Y2 biserrula × wheat. Y2WN Y2 wheat × no host. Y2WB Y2 wheat × biserrula. Y2WW Y2 wheat × wheat. Descriptions and abbreviations for all metabarcoding samples referred to in this chapter.

5.2.5. DNA extraction and sequencing Subsamples weighing 0.25 g were taken from each of the 64 experimental soil samples taken during the trial. Total genomic DNA was extracted using the Mo Bio PowerLyser Soil DNA Kit (Qiagen, Venlo, Netherlands) according to the manufacturer’s

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 121 instructions. DNA was quantified using a NanoDrop (Thermo Fisher Scientific, Waltham MA, USA) and each sample diluted to approximately 10 ng DNA µl-1. Sequencing was carried out on the Illumina MiSeq platform at the Australian Genome Research Facility (Melbourne, Victoria, Australia), targeting the 27F-19R region of the bacterial 16s rRNA gene. Sequencing conditions and primer details are described in section 3.2.3. and Table 3-1. Extractions were performed within 72 hours of collection.

5.2.6. Statistical analysis All statistical analyses were carried out using version 3.6 of the R statistical software program (R Core Team 2019). Measurements of alpha diversity were carried out on unrarefied sequence data to minimise bias (Willis 2019, McMurdie and Holmes 2014). Species (taxon) richness (S) was calculated as the total count of all operational taxonomic units (OTUs) present in a sample. Shannon-Weiner index (H) values and Pielou’s evenness (J) values were obtained for all samples using the Vegan package (Oksanen et al. 2019). In order to provide more intuitive comparisons between samples, Hill numbers were obtained for all samples by calculating the exponential of H (Jost 2006). The resulting values for effective species number (ExpH) reflect the number of taxa necessary to obtain the observed values for H, assuming complete evenness of distribution across those taxa. Additional comparisons between cultured and non- cultured samples were facilitated by deriving diversity profiles for each sample according to the method established by Leinster and Cobbold (2012). Values of 0-5 for the sensitivity parameter q were used to depict the shape of each community. These profiles were plotted using the function div.profile in the package hilldiv (Alberdi and Gilbert 2019). To aid in curve differentiation, the proportional change in diversity value between q=0 (species richness, S) and q=1 (effective species number, ExpH) was calculated for each treatment and subjected to ANOVA, using the aov function in R. Multiple pairwise comparisons were performed utilising Tukey’s honest significant difference test. Tukey-adjusted p values were considered significant at < 0.05. Following the analysis pathway outlined by Gloor et al. (2017), relative sequence abundance data were subjected to a centred log-ratio (CLR) transformation using the package compositions (van den Boogaart et al. 2019) prior to performing principal component analysis (PCA) and differential abundance analysis. Correlation matrix PCA was carried out on CLR-transformed data using the core R stats library and visualised using the package factoextra (Kassambara and Mundt 2017). PCA was carried out at the

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 122 level of phylum (i.e. the total number of sequences per phylum in each sample) to ensure that the number of observations (n=63) exceeded the number of variables (p=33). Differential abundance analysis was carried out between Y1NC and Y1CU, between Y0 and Y1 samples, between Y1 and Y2, and between all Y2 samples. Sequence data from these treatments were subjected to ANOVA-Like Differential Expression analysis (ALDEx) using the ALDEx2 package for R (Gloor et al. 2016, Fernandes et al. 2014, Fernandes et al. 2013). This method generates a reproducible measure of the difference in abundance of a taxon between two samples (effect size). Analysis was confined to the most abundant taxa in each pair undergoing comparison. The results of Welch’s t-test were used to identify differences in taxon abundance. To ensure a robust estimation of effect size, 1000 Monte-Carlo iterations were used to generate all aldex.clr objects. Taxa identified as differentially abundant between samples were visualised using Bland- Altman plots. Benjamini-Hochberg corrected p values were calculated to take into account the effect of multiple comparisons. Effects were considered significant at q < 0.1.

5.3. Results

5.3.1. Sequencing A total of 2,876,393 sequence reads (280-510 bp) was obtained, corresponding to 13,021 operational taxonomic units (OTUs). Sequence counts per sample ranged from 13,788 to 27,269 for initial (Y0) samples, 15,449 to 59,306 for first-harvest cultured (Y1CU) samples, 17,867 to 34,436 for first-harvest non-cultured (Y1NC) samples, and 9,013 to 216,328 for second-harvest (Y2) samples. OTU counts per sample ranged from 1,183 to 1,634 for initial (Y0) samples, 258 to 706 for Y1CU, 1,488 to 2,115 for first- harvest non-cultured (Y1NC) samples, and 2,115 to 6,230 for second-harvest (Y2) samples. Sequence and taxon counts varied considerably between samples (Table 5-2.). OTUs were assigned to 33 named phyla, 98 classes, 141 orders, 187 families, and 324 genera. The most abundant phyla by OTU count were Proteobacteria (21.54%), Actinobacteria (20.76%), Chloroflexi (11.57%), Firmicutes (9.11%), Planctomycetes (8.04%), Bacteroidetes (6.13%), and Gemmatimonadetes (5.01%) (Figure 5-3a.). By sequence count, the dominant phyla were Actinobacteria (39.81%), Proteobacteria (30.69%), and Firmicutes (11.92%) (Figure 5-3b.). The relative abundance of the dominant phyla was consistent across all non-cultured treatments but differed in Y1CU

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 123 samples, where Proteobacteria (57.61%) and Firmicutes (38.96%) increased their share of the total sequence count at the expense of Actinobacteria (2.81%). One sample, which exhibited stunted growth in its first season, was removed from subsequent analyses. Consequently, values for Y2WN are derived from three samples and not four. The sample was not removed from Y1 as those samples were bulked from three pots and were less obviously affected.

5.3.2. Principal component analysis (PCA) Principal component analysis of CLR-transformed data revealed a strong separation of cultured samples from non-cultured samples and a weaker tendency for non-cultured samples to cluster by harvest (Figure 5-4.). In the plot of individuals (Figure 5-4a.), cultured treatments form a discrete grouping on the right while non-cultured (Y0, Y1, and Y2) samples form a continuum across the left of the plot. Within this continuum, Y0 separates cleanly from Y2 but not from Y1. While Y1 forms a close grouping, it overlaps Y2, which shows a wider variance (particularly within Y2 biserrula × biserrula) and accounts for most of the continuum. PCA did not distinguish between treatments within each harvest group. The selected PCs together account for 50.1% of variation, with most of that (40.0%) accounted for by PC1. The variables making the greatest contributions to PC1 include Chloroflexi, Gemmatimonadetes, and Planctomycetes (Figure 5-5a.). The variable making the greatest contributions to PC2 was Cyanobacteria, followed by WS2, Chlorobi, and OP3 (Figure 5-5b.).

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 124 Figure 5-3. Proportional abundance of bacterial phyla across all 64 samples, measured by operational taxonomic unit (OTU) count (a) and total sequence count (b) per phylum.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 125 Figure 5-4. Individual plot (a) and variable plot (b) of correlation matrix PCA of log- transformed data from all 63 samples. The two principal components making the greatest contribution to observed variance have been retained. The variation explained by each PC is indicated on the axes.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 126 Figure 5-5. Contribution of each variable to PC1 (a) and PC2 (b). The dashed red line indicates the average expected contribution of each variable assuming all phyla contributed evenly (2.94%). For each PC, all but one variable below the cut-off has been omitted.

5.3.3. Effect of culturing and host plant on bacterial community diversity Bacterial community diversity was assessed by means of species richness (S) (Table 5- 1.), the Shannon-Weiner index (H), Pielou’s evenness (J), and effective species number (ExpH) (Table 5-2.). Additionally, community diversity and structure in cultured and non-cultured samples were compared by means of diversity profiles, which incorporate species richness (q=0), exponential Shannon (q=1), and Simpson diversity (q=2) (Figure 5-6.). Cultured samples produced lower values across all diversity indices

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 127 relative to their non-cultured equivalents. This difference is most clearly observed in the comparative values for ExpH and is consistent across all treatments (the two host plants and nil). Despite returning S values in the range of 258-706, mean ExpH for cultured treatments exceeded single digits only in the biserrula treatment (12.21 ± 2.19). The gulf between species richness and effective species number is indicative of a substantial skew in these samples, in which a very small number of species account for a large proportion of each sample. This is represented numerically by the values for J, which were consistently lower in cultured treatments than in their non-cultured equivalents. When the proportional decrease in value between q=0 (S) and q=1 (ExpH) was compared for all treatments, significant differences (p < 0.01) were observed between cultured and non-cultured treatments for each host plant and between cultured treatments and initial (Y0) samples. This decrease was greatest for the cultured treatments and corresponds with the lower J values reported for these samples. Non- cultured treatments did not differ significantly from Y0 samples, nor did host plants produce significant differences within the cultured and non-cultured treatments. Diversity values were lower in cultured treatments than non-cultured treatments for all plotted values of q. Alpha diversity values, including values of ExpH (Table 5-3.) increased between Y0 and Y1 and again, more noticeably, between Y1 and Y2. The increase in values of S between Y1 and Y2 suggests that sampling in the first growing period did not approach completion: many of the taxa present in Y2 samples were not detected in Y1. The increase between Y0 and Y1 occurred across all treatments but was less pronounced in the wheat treatment. The increase between Y1 and Y2 was much greater and resulted in Y2 ExpH values more than twice as large as those reported from Y1 samples. Among Y2 samples, the lowest mean ExpH values were found among second-season biserrula treatments (Y2WB, Y2BB, and Y2NB). The highest mean values were reported from Y2BN, Y2NW, and Y2NN. Statistical comparisons between Hill numbers were not performed owing to established limitations in the proposed tests, which are known to return biased p values (Butturi-Gomes 2017), but the values obtained here, while not clearly divergent, may point to a slight tendency for biserrula to reduce microbiome diversity.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 128

Table 5-2. Summary of bacterial sequence counts and species richness by treatment Sequence counts S Treatment Range Mean s.e. Range Mean s.e. Y0 12490-25640 17951.25 2760.99 1181-1632 1382.5 96.11 Y1NILNC 16384-32328 22236.25 3477.82 1488-2145 1714 138.74 Y1BISNC 19007-30756 25861.25 2728.80 1616-2105 1852.75 117.19 Y1WHNC 18945-30872 26469.5 2596.89 1509-1889 1751.5 90.54 Y1NILCU 15059-58757 38836 9725.83 339-549 431.5 44.95 Y1BISCU 15544-45377 32746.25 6489.40 337-706 538 80.63 Y1WHCU 16528-39264 27626 4730.27 258-464 384 45.55 Y2NN 32771-58276 49091.25 5882.09 3269-4969 4209.25 352.76 27277- Y2NB 59179.5 15611.71 3195-5587 4152 508.26 102107 42917- Y2NW 78796 34164.33 3761-5835 4530.25 452.11 181264 Y2BN 26427-93409 57693 14332.54 3200-5567 4385.75 536.99 Y2BB 9013-78701 52420 16239.90 2115-4908 3713.5 582.18 Y2BW 30070-81603 55136.5 11894.97 3565-4662 4116 224.57 Y2WN 20192-71462 46742.67 14828.67 2550-4816 3879 682.87 30504- Y2WB 84295.25 44478.13 3351-6230 4338 670.97 216328 Y2WW 14733-53819 38357.75 9353.46 2720-4282 3614.25 406.25 Bacterial sequence counts and species richness (S) for all treatments, showing ranges, means and standard error. Treatments are described in Table 5-1.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 129

Table 5-3. Diversity index values by treatment H J Exp(H) Treatment Range Mean s.e. Range Mean s.e. Range Mean s.e. 5.28 0.75 196.66 Y0 5.56 0.096 0.77 0.011 264.01 23.18 5.70 0.79 299.37 5.82 0.78 337.08 Y1NILNC 5.92 0.068 0.80 0.004 374.66 26.81 6.11 0.80 452.10 5.88 0.79 357.42 Y1BISNC 5.99 0.059 0.80 0.002 397.84 24.01 6.13 0.80 458.42 5.61 0.77 273.01 Y1WHNC 5.84 0.079 0.78 0.005 346.74 25.42 5.96 0.79 387.36 1.77 0.30 5.86 Y1NILCU 2.00 0.12 0.33 0.019 7.50 0.91 2.31 0.39 10.11 2.13 0.36 8.40 Y1BISCU 2.46 0.18 0.39 0.019 12.21 2.19 2.89 0.44 17.90 1.22 0.22 3.39 Y1WHCU 1.99 0.31 0.33 0.049 8.34 2.24 2.59 0.44 13.33 6.83 0.83 923.47 Y2NN 6.97 0.071 0.84 0.004 1074.05 78.03 7.16 0.84 1280.77 6.05 0.73 423.56 Y2NB 6.68 0.23 0.80 0.026 852.19 171.25 7.13 0.85 1246.94 6.87 0.81 963.58 Y2NW 6.98 0.045 0.83 0.0072 1081.99 47.86 7.08 0.85 1183.41 6.84 0.83 931.10 Y2BN 7.02 0.08 0.84 0.007 1129.71 94.00 7.18 0.86 1314.75 6.31 0.76 549.36 Y2BB 6.68 0.15 0.82 0.024 821.72 117.15 7.02 0.88 1118.54 6.66 0.80 779.58 Y2BW 6.85 0.075 0.82 0.012 954.85 70.47 7.02 0.85 1119.96 6.72 0.83 828.54 Y2WN 6.92 0.11 0.84 0.008 1026.31 106.10 7.08 0.86 1191.82 6.37 0.76 583.65 Y2WB 6.64 0.12 0.80 0.016 777.55 88.84 6.87 0.84 960.45 6.68 0.83 796.75 Y2WW 6.84 0.072 0.84 0.014 942.58 67.70 7.00 0.88 1095.92 Shannon entropy (H), Pielou’s evenness (J), and effective species numbers (calculated as the exponential of H) for all treatments, showing ranges, means and standard error.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 130

5.3.4. Effect of culturing and host plant on OTU abundance The effect of culturing was most pronounced in the biserrula treatment (Figure 5-7.), where 69 of the 143 OTUs occurring more than 500 times in the first-season dataset displayed differential abundances between cultured (20 OTUs enriched) and non- cultured samples (49 OTUs enriched). This was greater than the effect observed for the no host treatment (53 taxa, 20 enriched in cultures) and the wheat treatment (31 taxa, 16 enriched in cultures). In total, 26 taxa returned differential abundances between cultured and non-cultured samples across all sample types (no host, biserrula, wheat). Of these, 15 exhibited negative effects (increased in cultures) and 11 exhibited positive effects (decreased in cultures). The four largest negative effects were consistent between all treatments, although the order was found to vary. These were Bacillus cereus (effect sizes -6.77 to -9.71), Bacillus sp. (-5.79 to -8.32), Lysinibacillus boronitolerans (-5.66 to -6.17), and an unidentified species of the Enterobacteriaceae (-7.01 to -8.03). The largest positive value for all treatments was recorded from a species of Kaistobacter (3.93 to 5.12); the ranking of positive effects otherwise differed considerably between sample types. The 15 taxa consistently enriched in cultures belonged to four families: Bacillaceae (four taxa in Bacillus and two in Lysinibacillus), Enterobacteriaceae (four taxa, including Erwinia sp.), (three species of Pseudomonas), and Clostridiaceae (one species each of Clostridium and Tepidibacter). By contrast, the 11 taxa observed to consistently decline in cultured treatments were distributed among eight families: two OTUs in each of the Geodermatophilaceae (including one in the genus Geodermatophilus), Nocardioidaceae, and Streptomycetaceae, and one from each of the Hyphomicrobiaceae (Rhodoplanes sp.), Koribacteraceae (Candidatus Koribacter sp.), Nocardioidaceae, Pseudonocardiaceae, and (Kaistobacter sp.). Comparison of Y0 and Y1 returned no evidence of recruitment by biserrula but some evidence of a recruitment effect in wheat (Figure 5-8.), One OTU, a species of the Oxalobacteraceae most similar to Massilia sp., was observed to be differentially abundant between the Y0 samples and the Y1 biserrula treatment (4.00, q=0.03); however, this taxon showed a similar decline in the Y1 wheat (3.35, q=0.03) and Y1 nil (3.88, q=0.05) treatments. No other taxa were observed to differ in the Y1 biserrula and Y1 nil treatments relative to Y0. By contrast, six additional taxa were identified as differentially abundant in Y1 wheat. Two of these, Geodermatophilus sp. (2.47, q=0.07)

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 131 and another species of the Geodermatophilaceeae (2.54, q=0.08), returned higher abundances in the Y0 treatment. The remaining four, all members of the Actinomycetales, were enriched in wheat. Ordered by effect size, these were Kribbella sp. (-4.78, q=0.05), Promicromonospora sp. (-4.69, q=0.02), Amycolatopsis sp. (-3.45, q=0.04), and Saccharothrix sp. (-2.31, q=0.10). The Promicromonospora species was also found to be enriched in Y1 wheat when compared to the Y1 no host treatment (4.14, q=0.04) and the Kribbella species was found to be enriched in Y1 wheat relative to Y1 biserrula (4.84, q=0.04). No other differences were recorded among Y1 samples. Second season microbiomes differed markedly from their Y1 counterparts (Figure 5-9.). Of the 483 taxa occurring > 1000 times in the Y1 and Y2 data, 320 were found to be differentially abundant between Y1 biserrula and the Y2BN treatment, 306 between Y1 biserrula and Y2BB, and 332 between Y1 biserrula and YBW. Likewise, 308 taxa were found to be differentially expressed between Y1 wheat and the Y2WN treatment, 325 between Y1 wheat and Y2WB, and 343 between Y1 wheat and Y2WW. In the no host treatment, 294 taxa were found to differ between Y1 no host and Y2NN, 305 between Y1 no host and Y2NB, and 324 between Y1 no host and Y2NW. After two consecutive seasons of growth, wheat exhibited a slight recruitment effect when compared to soil maintained for two season without a host plant (Figure 5- 10.). Two taxa, both members of the Actinomycetales, were found to be enriched in Y2WW relative to Y2NN: Saccharothrix sp. (4.25, q=0.05) and Streptomyces sp. (2.78, q=0.07). No further evidence of a recruitment effect following two seasons of wheat was observed. Likewise, biserrula grown for two consecutive seasons was found to differ only from a single treatment, Y2NW (Figure 5-10.). Nine taxa were found to differ between these two treatments: three enriched in the rhizosphere of biserrula (negative effect) and six diminished (positive effect). Arranged by effect size, the species demonstrating a potential antipathy to biserrula were Bacillus sp. (3.16, q=0.04), members of the Conexibacteraceae (1.94, q=0.07) and Sphingomonadaceae (1.89, q=0.07), Rhodoplanes sp. (1.74, q=0.08), a second Bacillus sp. (1.66, q=0.08), and Geodermatophilus sp. (1.30, q=0.10). Those demonstrating a potential affinity to biserrula were Pseudomonas sp. (-1.56, q=0.09), Stenotrophomonas sp. (-1.51, q=0.09), and a member of the Micrococcaceae most closely matching species of Arthrobacter and Paenarthrobacter (-1.41, q=0.10). Significant effects (q < 1) were confined to the most abundant taxa (> 8000 reads in total) in the dataset.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 132 Differential abundance analysis did identify a number of additional non-significant (q > 0.1) effects with large median effect sizes. It has been argued (Gloor 2018) that any OTU with an effect size > 1 should be reported, irrespective of q value. In the present case, however, high between-replicate variation in effect size discouraged the inclusion of these taxa.

Figure 5-6. Diversity profiles of initial (Y0), first-season cultured (Y1NILCU, Y1BISCU, Y1WHCU), and first-season non-cultured (Y1NILNC, Y1BISNC, Y1WHNC) samples for values of q=0-5, where q=0 is equal to S, q=1 is equal to H, and q=2 is equal to Simpson diversity. Replicates are displayed individually: panels correspond to the first (a), second (b), third (c), and fourth (d) replicates.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 133 a b

c d

Figure 5-7. Operational taxonomic units (OTUs) identified as differentially abundant between cultured and non-cultured sample pairs taken from biserrula (a, b), wheat (c, d) and the no host treatment (e, f). Points marked in red are differentially abundant between treatments (q < 0.1), while points marked in grey or black are non-significant. The Bland-Altman plots (a, c, e) display the relationship between difference and abundance. The effect plots (b, d, f) map the relationship between difference and dispersion. Data are centred log-ratio transformed 16S rRNA gene sequence counts. [Figure continues over page].

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 134 e f

Figure 5-7. [Figure continues from previous page].

a b

Figure 5-8. Operational taxonomic units (OTUs) identified as differentially abundant between Y0 samples and first-season wheat. Points marked in red are differentially abundant between treatments (q < 0.1), while points marked in grey or black are non- significant. In the effect plot (b), red points above the upper line indicate taxa relatively enriched in Y0, whereas red points below the lower line indicate taxa relatively enriched in wheat. Data are centred log-ratio transformed 16S rRNA gene sequence counts.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 135 a b c

Figure 5-9. Operational taxonomic units (OTUs) identified as differentially abundant between Y1 nil and Y2NN (a), Y2NB (b), and Y2NW (c); between Y1 biserrula and Y2BN (d), Y2BB (e), and Y2BW (f); and between Y1 wheat and Y2WN (g), Y2WB (h), and Y2WW (i). Points marked in red are OTUs differentially abundant between treatments (q < 0.1), while points marked in grey or black are non-significant. Red points above the upper line in each plot are significantly enriched in the Y2 treatment. Red points below the lower line are significantly enriched in the Y1 treatment. Data are centred log-ratio transformed 16S rRNA gene sequence counts. [Figure continues over page].

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 136 d e f

Figure 5-9. [Figure continues from previous page].

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 137 g h i

Figure 5-9. [Figure continues from previous page].

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 138 a b

c d

Figure 5-10. Operational taxonomic units (OTUs) identified as differentially abundant between Y2NN and Y2WW (a, b), and between Y2NW and Y2BB (c, d). Points marked in red are differentially abundant between treatments (q < 0.1), while points marked in grey or black are non-significant. The Bland-Altman plots (a, c) displays the relationship between difference and abundance. The effect plots (b, d) map the relationship between difference and dispersion. Data are centred log-ratio transformed 16S rRNA gene sequence counts.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 139 5.4. Discussion For a microbe to be of use in improving plant nutrition, it should, ideally, be selectively recruited to the rhizosphere by its host plant. This study did not demonstrate clear host- specific recruitment of bacterial taxa by either biserrula or wheat, and the few examples that were observed tended to be found in the most extreme comparisons. The observed recruitment effects between Y2 wheat × wheat (Y2WW) and Y2 no host × no host (Y2NN), and between Y2 biserrula × biserrula (Y2BB) and Y2 no host × wheat (Y2NW), suggest that the curation of the “rhizosphere zoo” (Buée et al. 2009) may, in some circumstances, only become apparent when the host plant has been absent from the soil for some time. The possibly affinity between biserrula and members of the family Micrococcaceae suggested in the Y2BB and Y2NW comparison is consistent with previous observations, as is the antipathy between biserrula and Bacillus spp. (see section 3.3.1.). Pseudomonas, identified in the same comparison as a possible biserrula associate, was found in the earlier study to potentially contain species both enriched and diminished in the rhizosphere of biserrula. Several additional taxa displayed evidence of possible recruitment or suppression in certain comparisons but these effects demonstrated high dispersion and were found to be non-significant (q > 0.1). Additional replication may be needed to establish whether these effects are genuine. The decline of a Rhodoplanes species in the biserrula × biserrula treatment relative to the nil × wheat treatment cannot easily be explained, nor can the decline observed across all Y1 treatments for a member of the Oxalobacteraceae. Both Rhodoplanes (Gkarmiri et al. 2017, Srinivas et al. 2014) and members of the Oxalobacteraceae (Li et al. 2014, Ofek et al. 2012, Green et al. 2007) are commonly enriched in rhizospheres. The strong effect of host plant demonstrated in previous studies (Bakker et al. 2013, Kristin and Miranda 2013), including by the legume species Trifolium pratense (Hartman et al. 2017) and Glycine max (Liu et al. 2019a), was not apparent here; however, the present study did provide additional support for the previously reported association between wheat and members of the Actinomycetales (Conn and Franco 2004, Coombs and Franco 2003). Species of Kribbella, Promicromonospora, Amycolatopsis, Saccharothrix, and Streptomyces were all observed to be enriched in wheat treatments relative to treatments lacking a host plant. On the other hand, two members of the Geodermatophilaceae – also members of the Actinobacteria – were found to decline in wheat treatments, and this effect has not previously been reported. Wheat belonging to the Actinobacteria are of agronomic interest owing to

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 140 their potential to promote plant growth (Franco et al. 2007): members of the genus Streptomyces, in particular, are known to contribute to host plant nutrition and disease resistance (Viaene 2016). While one taxon found in the present study, a species of Kribbella, was enriched in wheat relative to biserrula, the remaining taxa did not significantly differ between wheat and biserrula treatments. This suggests that, if they are found to have PGPR potential, they may be effectively shared between the rhizospheres of these two host plants. The same is true of the enrichment in a single biserrula comparison of a species of Stenotrophomonas, a genus known to harbour PGPR strains (Ryan et al. 2009). It is unclear whether the apparent reduction in effective species (ExpH) number recorded for second-season biserrula treatments is a genuine effect or a statistical artefact; however, it is consistent with prior observations. As previously reported (see section 3.3.1.), biserrula was observed to selectively recruit only five taxa to its rhizosphere. By contrast, 21 taxa were found to decrease in number in the rhizosphere treatment. In the present study, a similar effect was observed in the Y2BB and Y2NW comparison: three OTUs increased after two successive seasons of biserrula, whereas six decreased. A reduction in alpha diversity may be a function of the nutritional enrichment of the rhizosphere: this phenomenon is known to recruit copiotrophic bacteria, which are less diverse than the found to make up much of the soil microbiota, from the bulk soil (Ho et al. 2017). While mean values for S and J in Y2 biserrula were within the range reported for other treatments, individual replicates demonstrated both comparatively low species richness and high unevenness, suggesting that these treatments were inconsistently affected by recruitment. This, however, requires further investigation. Likewise, it is unclear whether the association between no host treatments and higher ExpH represents a genuine increase in diversity in fallow periods. Fallow periods have previously been found to increase microbial diversity relative to certain crop species(Peralta et al. 2018). It is possible that these rest periods allow for the recovery of taxa diminished by recruitment. Cultured samples produced lower values than their non-cultured equivalents in species richness, Shannon entropy, and effective species number. They were also less even than non-cultured samples: that is, they tended to be dominated by small cohorts of highly responsive taxa. This was reflected in the low mean effective species numbers recorded for these treatments, which varied from 7.5 to 12.21, despite mean values for S ranging from 384 to 538. The seven most abundant taxa in the cultured treatments

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 141 accounted for more than 75% of all sequence reads in those samples; the most abundant OTU, an unidentified member of the Enterobacteriaceae, accounted for 38.2% of all cultured sequences. By contrast, it took 281 taxa to reach 75% in the non-cultured samples and the most abundant non-cultured taxon, a species of Kaistobacter, a genus reportedly enriched in the rhizospheres of several plants of agricultural importance and identified as a potential contributor to disease-suppressive soils (Gkarmiri et al. 2017, Wu et al. 2017, Liu et al. 2016), accounted for a comparatively small 7.2% of all non- cultured sequences. Cultured samples also separated strongly from non-cultured samples in phylum- level principal component analysis. Contributions to PC1, which accounted for 40.0% of observed variation, were distributed across many taxonomic groups: in all, 15 of 34 phyla provided contributions above the 2.94% expected average. Of these, the first five more than doubled the expected contribution and varied in their individual contributions only slightly. These were, in declining order, Chloroflexi, Planctomycetes, Gemmatimonadetes, Acidobacteria, and Actinobacteria. Two of these phyla – Gemmatimonadetes and Acidobacteria – were first recorded from 16s rRNA datasets and had no cultured representatives prior to the advent of molecular surveys (Youssef et al. 2015). Members of the Planctomycetes are similarly resistant to culturing using non- selective media (Lage and Bondoso 2012). Major contributing variables to PC2 included Cyanobacteria and Chlorobi, both of which were observed to decline in cultured samples. In the differential abundance analysis, culturing consistently affected taxon abundance in 18.18% of all taxa. In biserrula, this increased to 48.25%, greater than the values for wheat (21.68%) and the no-host treatment (37.06%), suggesting that the biserrula rhizosphere may be particularly susceptible to misrepresentation by culture- dependent methods. A reason for this has not been identified. The taxa consistently found to increase under culturing included four species of Bacillus, several species of which were previously found to decline in the biserrula rhizosphere (section 3.3.1.). The enriched Bacillus spp. included B. cereus; a member of the B. cereus group has previously been identified as an MHB strain in biserrula (see Chapter 3.). The artificial enrichment of Bacillus spp. in cultures, an effect which, in the case of B. cereus, was most pronounced in biserrula (-9.71, q < 0.01), has implications for the identification and subsequent implementation of helper strains: the use of in vitro screening methods leads inevitably to the selection of highly responsive taxa, such as B. cereus, which may

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 142 not gravitate to the rhizosphere in real-world conditions. The diminished helper effect observed in a populated rhizosphere for the B, cereus group taxon (section 4.3.3.) may partly be explained by an aversion to the rhizosphere environment. This aversion is clearly not absolute, however: the B, cereus group taxon was isolated from the root endosphere of biserrula. Furthermore, biserrula has been found, in the present case and in a previous study (section 3.3.1.), to demonstrate a slight apparent affinity for species of Arthrobacter or Paenarthrobacter; however, the Arthrobacter species previously identified as an MHB strain was also found to produce a diminished helper effect in a populated rhizosphere. The separation of samples by harvest and not by treatment in CLR-transformed PCA, alpha diversity measures, and differential abundance analysis points to an instability in the microbiome over time that superseded any effect of host plant. This finding has implications for the use of MHB strains in rotations, suggesting that they may need regular reintroduction to persist in these unstable environments. The release of nutrients and the restructuring of soil communities following the removal of the host plant cannot account for the observed increase in diversity as it was also found in treatments without a first-season host plant. The cause of this instability in the rhizosphere cannot be conclusively identified; however, two factors relevant to the application of biofertilisers in pastures suggest themselves. Studies based on both molecular analysis and profiles have reported reduced total bacterial biomass in response to tilling (Zhang et al. 2014, Helgason et al. 2010, Feng et al. 2003), as well as selection for certain taxonomic groups (Yin et al. 2017, Yang et al. 2012) and functional traits (Schmidt 2018, Smith et al. 2016). The soil utilised in this experiment had been recently tilled, which may have artificially reduced microbial diversity in the early treatments. Alternatively, it is possible that bacterial abundance was affected by increasing temperature and day length over the trial period. The trial took place from winter to late spring: while it was conducted in a partially controlled environment, mean daily maximum temperatures increased from 15.95 ± 0.40°C during the first growing period to 23.70 ± 0.26°C during the second. Experimental warming (up to 5°C) has not consistently been found to increase microbial biomass (Deangelis et al. 2015, Schindlbacher et al. 2011); however, an influence of season on microbiome composition is known (Huang et al. 2018, Moreno-Espíndola et al. 2018). Additionally, it is possible that the use of separate soil collections for each replicate may have had a confounding effect, as variations in and physical properties can influence the structure

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 143 of the soil microbiome (Hermans et al. 2020). A final caveat concerns the method of soil sampling, which, while reliably returning a large volume of root material, may have confounded results by incorporating some amount of bulk soil. The results of the present study strongly suggest that culturing produces an interpretation of bacterial community composition that underestimates diversity and is biased towards a small number of highly responsive taxa. While primer-based amplification of soil microbial DNA is itself biased towards cultured species (Hug 2018, Lloyd et al. 2018), the use of a metabarcoding approach in the present study revealed a diversity of bacterial life that the culture-dependent method could not access. It should be noted, however, that this study employed only one set of culturing conditions and modifying the choice of medium or the incubation parameters may yield yet another interpretation of community structure. It is also possible to bring many bacterial taxa into culture by the use of selective media. There has been recent success in cultivating members of the Planctomycetes, here a major contributor to the distinction between cultured and non-cultured samples, by using a culture medium formulated with, among other ingredients, vitamin B12 and β-lactam antibiotics (Lage and Bondoso 2012). While the use of selective cultures may not facilitate as comprehensive a view of the rhizosphere microbiota as molecular methods, it is a desirable goal in the pursuit and subsequent screening of beneficial bacteria for agricultural applications. The apparent adaptation of Gemmatimonadetes to dry conditions (DeBruyn et al. 2011), for example, raises the possibility that it may harbour species of interest in promoting the growth of biserrula. Difficulties in bringing representatives of the phylum into culture form an impediment to investigations of this kind.

5.5. Conclusion The identification of taxa selectively recruited to the rhizosphere of a host plant can serve as a starting point in the isolation of plant growth-promoting microorganisms. Alternatively, the screening of strains for PGPR effects in axenic experimental conditions can serve as a starting point in the identification of microbial agents capable of enhancing plant growth in real-world conditions. Ultimately, it is the function of a microorganism in the context of the recruited community and the surrounding soil that is relevant to plant health and behaviour in a pasture setting. A significant limitation of nucleic acid-based techniques is the difficulty encountered in assigning an ecological

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 144 function to the sequences recovered (Morales and Holben 2011); at the same time, laboratory-based techniques may identify specific functional traits in taxa of interest but struggle to predict the behaviour of those taxa in situ (Parnell et al. 2016). The identification of helper effects in strains of Bacillus and the subsequent discovery that Bacillus spp. are diminished in the biserrula rhizosphere serves to demonstrate this point. While a Bacillus strain was observed to maintain their helper effects in a populated rhizosphere (section 4.3.3.), albeit with diminished effect, the present study raises concerns as to how it would fare in a real-world pasture. The general decline in abundance exhibited by Bacillus strains in the rhizosphere of biserrula suggests it should not be expected to endure across phases of a rotation. Historical efforts towards the development of biofertilisers have typically focused on single species or taxonomically restricted cohorts (e.g. rhizobia); more recently, however, an increased recognition of the importance of rhizosphere interactions in bacterial behaviour has shifted focus towards the development of synthetic microbial communities (Xavier 2011). Synthetic communities can be constructed as model systems for the study of ecological interactions or as function-first communities intended to exert a particular effect in situ, including the promotion of plant growth (Großkopf and Soyer 2014). The small cohort of taxa displaying affinities to wheat or biserrula identified in the present study and in the earlier diversity profiling exercise (section 3.3.1.) represent candidates for inclusion in functional synthetic communities. However, more needs to be known about their role in the rhizosphere; further, additional experimental evidence is needed to unravel the temporal dynamics of these apparent associations and whether they persist across the entire developmental cycle of the host plant. Metabarcoding could not determine whether the observed associations were to the benefit (or, indeed, the detriment) of the host plant. This is a question that can only be addressed by further screening, and while recent innovations in the cultivation of ‘uncultivable’ bacteria, most notably ichips (Nichols et al. 2010, Berdy et al. 2017), allow for at least the partial domestication of many recalcitrant taxa, lab- and field-based investigations of synthetic communities are often bound by the same culture dependency that has limited past explorations of rhizosphere diversity (Cairns et al. 2018). Thus, while metabarcoding techniques highlight the limitations of culture- dependent methods, they may not entirely obviate the need for those methods. Robust outcomes may demand an effective integration of the two.

Chapter 5. Characterisation of the rhizosphere microbiota of Biserrula pelecinus L. 145 Chapter 6. Summary and conclusions

6.1. Introduction This thesis presents the first investigation into a potential role for plant-associated bacteria in the promotion of mycorrhizal colonisation in the pasture legume Biserrula pelecinus L. (biserrula), an acid-tolerant species selected for implementation into crop- pasture rotations in low-rainfall environments, and valued for its weed-suppressive capabilities and low methanogenic potential. The research presented in these pages extends the range of bacterial taxa known to act as mycorrhiza helper bacteria and the range of host plants known to harbour helper species. It offers insights into the behaviour of these beneficial microorganisms in the rhizosphere of a pasture plant in a ley farming system and into potential mechanisms by which MHBs may exert their effects. This chapter reviews and contextualises the principal findings of this research project and suggests future directions for research into mycorrhiza helper bacteria both in the context of biserrula and in general.

6.2. Aims of the research The principal aim of this research project was to identify whether bacteria recruited to the rhizosphere of biserrula were capable of acting as mycorrhiza helper bacteria by promoting the pre-symbiotic growth of AM fungi or by increasing the rate of mycorrhizal colonisation. Subsidiary objectives included determining whether the helper effect persisted in a populated rhizosphere, identifying potential mechanisms involved in the helper effect, and deepening our understanding of the rhizosphere effect as it pertains to the curation, persistence, and stability of the biserrula rhizobiome in the context of a pasture rotation. The ultimate concern of this last objective was to determine whether helper strains inoculated into an agricultural setting would be capable of maintaining their effects on the mycorrhizal symbiosis in that environment. In the following pages, the experimental findings detailed in chapters 3–6 of this thesis are discussed with respect to these aims, and this discussion is additionally situated

Chapter 6. Summary and conclusions 146 within a consideration of two broader concerns: (1) the elucidation of the workings of the helper effect, and (2) the potential for practical implementation of these helper bacteria in Australian agriculture.

6.3. Key findings

6.3.1. Chapter 2. Observation and quantification of mycorrhizas in biserrula The experiments described in this chapter were conducted to determine whether biserrula is capable of forming mycorrhizal associations, to quantify mycorrhizal frequency in a field environment and in a synthetic trap culture, and to identify the fungal partners involved in this symbiosis. The key findings were as follows: 1. Biserrula formed arbuscular mycorrhizas in pasture and in laboratory-based trap cultures. 2. Colonisation rates were similar for both environments (19.2 ± 4.6% in field samples and 23.5 ± 3.5% in trap cultures); however, arbuscule frequency varied according to sample origin (field: 4.8 ± 1.3%; cultures: 12.2 ± 2.4%).

3. Measured colonisation rates were low in biserrula relative to equivalent N2- fixing herbaceous plants. 4. Mycorrhizal field root segments viewed microscopically were dominated by a single fungal morphotype. Spores associated with these mycorrhizas were consistently identified by DNA barcoding as Funneliformis mosseae. The two morphotypes observed in trap culture roots were also identified as F. mosseae. 5. Diversity profiling recovered greater AM diversity associated with biserrula. Taxa identified in the rhizosphere included F. mosseae and several species of Archaeospora. Most identified AM taxa occurred only in rhizosphere and bulk soil samples, and were not detected in the field soil. These results collectively form the first direct evidence of mycorrhization in biserrula. The low colonisation values point to the potential advantages to be gained by implementation of helper bacteria.

6.3.2. Chapter 3. In vitro and in vivo MHB screening The screening of bacteria isolated from plant tissues and the adjacent soil for helper effects and relevant functional traits returned these significant findings:

Chapter 6. Summary and conclusions 147 1. More than half of the bacterial taxa studied here (11 of 19) promoted in vitro germination of F. mosseae spores. A single strain of Paenibacillus was found to suppress germination. Four of 19 strains increased hyphal elongation in vitro. 2. Two strains of Pseudomonas facilitated secondary spore formation in vitro. 3. Six taxa were found to increase percent colonisation in vivo: Arthrobacter sp., Bacillus cereus group taxon, B. aquimaris/vietnamensis, Microbacterium xylanilyticum, Pseudomonas koreensis subgroup taxon, and P. granadensis. Effects varied from a 2.1-fold to a 3.6-fold increase in colonisation rate. 4. Putative helpers were not found to share a single functional trait that could account for all observed helper effects. No secretion system components were identified in any of the putative MHBs. 5. Diversity profiling data demonstrated that bacteria closely related to or identical with certain helper taxa may be enriched in the biserrula rhizosphere whereas others may be suppressed; however, the reduced taxonomic resolution offered by the NGS method impeded further analysis. These studies provide the first evidence of a helper effect in biserrula and only the second observation of in vitro spore formation by an AM fungus.

6.3.3. Chapter 4. Ecological interactions in the biserrula rhizosphere Putative MHB strains were screened in combination with soil filtrates and rhizobial additives to determine whether the helper effect could overcome microbial antagonism in the rhizosphere. The following observations were made: 1. A synergistic effect on hyphal elongation was observed in vitro in the interaction between the Pseudomonas koreensis subgroup taxon and the Bacillus cereus group taxon. 2. All identified helpers maintained their efficacy in a populated rhizosphere, although the effect was diminished for Arthrobacter sp. and Bacillus cereus group taxon. 3. Inoculation with the Pseudomonas koreensis subgroup taxon in the presence of Mesorhizobium ciceri bv. biserrulae increased the total number of nodule-like structures in the biserrula root, but not the number of active nodules. 4. The soil filtrate and the inoculum of Mesorhizobium ciceri bv. biserrula both acted as MHBs.

Chapter 6. Summary and conclusions 148 5. Nod factors were identified in Mesorhizobium using primers targeting nodC but not nodD genes. They were not identified in the Pseudomonas strain found to promote nodulation using any primer pair.. The results obtained here extend the range of identified helpers further by demonstrating a helper effect for Mesorhizobium and for a rhizosphere soil filtrate that appeared to act as an MHB consortium. The observation of an in vitro synergistic interaction of unrelated helper strains is novel.

6.3.4. Chapter 5. Characterisation of the biserrula rhizobiome in a pasture rotation This study explored the dynamics of the biserrula rhizosphere across a simulated two- season pasture rotation. It also compared rhizosphere diversity of enriched and non- enriched soil samples. The following significant results were recorded: 1. Enriched samples were less diverse than non-enriched samples and tended to be dominated by small cohorts of responsive taxa. This effect was most extreme in biserrula treatments. 2. No evidence of host-specific recruitment by biserrula was observed in the first growing season. A stronger effect was observed for wheat. 3. Recruitment by biserrula was only apparent in comparison to a treatment that had not contained biserrula for two seasons – that is, when comparing the biserrula × biserrula and no host × wheat treatments.. 4. Enrichment was observed to increase the abundance of Bacillus spp., including B. cereus, while biserrula appeared to suppress these taxa. 5. Bacterial community composition changed markedly between growing periods, possibly as a consequence of agricultural management or the effects of temperature and day length. 6. High between-replicate variance was observed, suggesting that the historical failure to replicate sufficiently has yielded misleading interpretations of the soil microbiome.

6.4. Synthesis This project recorded the first direct evidence of the formation of arbuscular mycorrhizas in biserrula, offering a novel pathway by which the growth of the plant might be promoted. AM associations are commonly regarded as true mutualisms: they confer benefits on their host plants ranging from improved uptake of mineral nutrients

Chapter 6. Summary and conclusions 149 to defence against pathogens and resistance to abiotic stresses (Jung et al. 2012). Consequently, maximising plant potential may mean successfully managing the mycorrhizal symbiosis. In the case of biserrula, which exhibited poor mycorrhizal colonisation (~20%) relative to similar herbaceous legumes of agricultural significance (Treseder 2013), promotion of the AM symbiosis represents a promising means by which plant health and nutrition might be improved. While microscopic examination of the colonised biserrula roots and genetic analysis of associated spores yielded evidence of only a single AM fungus, F. mosseae, metabarcoding revealed a comparatively diverse consortium of AM fungi in the roots of field-grown biserrula plants. Thus, while biserrula may be highly specific in its rhizobial requirements (Peix et al. 2015), it may function as a mycorrhizal generalist. More than five decades of research now point to a beneficial effect of select bacterial strains on the colonisation rate of mycorrhizal fungi (reviewed in Frey-Klett et al. 2007 and Garbaye 1994). While MHBs have been identified in similar herbaceous legumes, including Glycine max (Fernández Bidondo et al. 2011), Medicago truncatula (Pivato et al. 2009), and M. sativa (Azcón et al. 1991), biserrula had not previously been investigated as a target plant. In the present study, six bacterial taxa isolated from the rhizosphere soil and from plant tissues were found to promote in vivo colonisation of biserrula by F. mosseae. These taxa were identified as Arthrobacter sp., Bacillus cereus group taxon, B. aquimaris/vietnamensis, Microbacterium xylanilyticum, Pseudomonas koreensis subgroup taxon, and P. granadensis. The most pronounced effect, a 3.6-fold increase in percent colonisation, was reported for the P. koreensis subgroup taxon, exceeding effects previously reported for Pseudomonas spp., with the unusual exception of the first MHB identified, which reliably facilitated mycorrhization in a host that otherwise resisted all attempts at colonisation (Mosse 1962). Candidate helper bacteria were also subjected to in vitro screening to determine their effects on the germination and pre-symbiotic growth of the AM fungus. In these assays, 11 taxa were found to promote germination of F. mosseae spores and four were found to foster hyphal elongation. These results, most notably the germination results, proved unreliable as predictors of subsequent in vivo colonisation success. Variovorax paradoxus, which returned the highest in vitro germination rate, and was strongly recruited to the biserrula rhizosphere relative to the field soil, subsequently demonstrated poor in vivo performance. The reverse was true of a strain of Paenibacillus, which suppressed in vitro germination while having no effect on

Chapter 6. Summary and conclusions 150 colonisation of the root. While it is clearly necessary to reduce the diversity of cultivable rhizosphere bacteria to a manageable level, valuable MHB strains may be excluded if these results are overemphasised. A second in vitro study yielded a more unusual result. The two strains of Pseudomonas identified in the in vivo experiment as helpers were observed to facilitate the sporulation of F. mosseae in culture. This is noteworthy, as AM fungi are conventionally regarded as obligate symbionts, incapable of completing their life cycle in the absence of a host plant (Requena et al. 2007). This observation is unique for both Pseudomonas and F. mosseae, having been recorded only once before, in an association of Paenibacillus validus and Rhizophagus irregularis (Hildebrandt et al. 2002). More recently, sporulation of an AM fungus was observed in response to the application of plant exudates (Liu et al. 2019b), but the formation of spores in the absence of any plant material at all so far appears to be confined to these two systems. While yields and subsequent germination were both poor (only three viable spores were produced), the capacity of a bacterial taxon to overcome the dependency of an AM fungus on a host plant may be a significant factor in certain examples of the helper effect. In the osmotic stress screening, only one strain, B. aquimaris/B. vietnamensis, was found to tolerate elevated NaCl levels. Additional environmental screenings are necessary to fully define the set of conditions each helper can be expected to tolerate. Attempts to identify the mechanisms responsible for the helper effect did not identify a single trait shared by all MHBs; however, all studied modes of enzymatic activity (amylolytic, cellulolytic, chitinolytic, pectinolytic, and proteolytic) were observed in at least one taxon. This agrees with prior experimental evidence, which has not reliably correlated any particular mode of enzymatic activity with the helper effect (Fernández Bidondo et al. 2016). Similarly, the results obtained here did not validate the recent focus on a role for Type III secretion in helper effects (Viollet et al. 2017, Cusano et al. 2011): no Type III secretion components were identified in any of the taxa under study. These results point to the likelihood that the helper effect is, in reality, a cluster of individual effects all promoting mycorrhizal symbiosis. A potential mode of action was identified for one of the helper strains, the P. koreensis subgroup taxon, which was observed to promote formation of nodule-like structures in the biserrula root in the absence of rhizobia and to significantly increase the number of nodules produced when co-inoculated with rhizobia, providing support for the suggestion that stimulation of the common symbiotic pathway linking

Chapter 6. Summary and conclusions 151 mycorrhization and nodulation could act as a helper mechanism (Xie et al. 1995). While the structres produced by this taxon were inactive, this may be a function of the relative dosages of the two inoculants or evidence of a deficiency in the rhizobial inoculum. Alternatively, it may be a consequence of the level of nitrogen in the soil, though high soil N is typically observed to suppress nodule formation, not simply nodule activity (Friel and Friesen 2019, Xia et al. 2017). While it was beyond the scope of the present study to explore all aspects pertaining to the induction of nodules by MHB strains, or to attempt the purification of Nod factors from the Pseudomonas strain, this suggests a productive avenue for future experimentation. This might productively be expanded to include the identification of functional nif genes and active nitrogen fixation, which has been identified in certain Pseudomonas strains (Li et al. 2017). While many helpers are identified on the basis of axenic screening methods alone (Frey-Klett et al. 2007), MHB strains that are to be employed in real-world conditions face extreme competition in the “battlefield” of the rhizosphere (Raaijmakers et al. 2009). Consequently, screening of MHBs in populated rhizospheres may yield results that are truer to life. In AM systems, with rare exceptions (Paula et al. 1992, Omirou et al. 2016, Imperiali et al. 2017), this has not been attempted. In the present study, co- inoculation with a rhizosphere soil filtrate reduced the efficacy of two helper strains (Arthrobacter sp. and the B. cereus group taxon), without eliminating the helper effect, but did not affect two others (M. xylanilyticum and the P. koreensis subgroup taxon). The filtrate itself was also found to enhance mycorrhizal colonisation, suggesting that the label MHB can be extended to certain soil consortia and not simply single species. Similar results were obtained when helper strains were co-inoculated with the nodule- forming bacterium M. ciceri bv. biserrulae. Application of the rhizobial inoculum in concert with a helper strain neither enhanced nor diminished the observed helper effect, while application of M. ciceri bv. biserrulae on its own produced a helper effect. Thus Mesorhizobium was found to join Rhizobium, Bradyrhizobium, and Sinorhizobium/ Ensifer as rhizobial taxa observed to act as MHBs (Fernández Bidondo et al. 2016, Xie et al. 1995). A discrepancy was noted between the colonisation values obtained in the in vivo experiments and the comparatively low values observed in the pasture from which the helper strains were isolated. While this disparity may simply be a matter of the high dosage of each MHB strain used in these screenings, the dynamics of the rhizosphere remain poorly known (Shelef et al. 2019), and the efficacy of a microbial inoculum may

Chapter 6. Summary and conclusions 152 depend on factors that are not currently known. In some cases, these factors may mimic those long known to affect the use of rhizobial inoculants (Weber 1977), including the impacts of soil fertility and competition with indigenous microbes on the activity of the inoculum. In a simulated two-season ley-farming rotation, biserrula was not observed to exhibit a strong recruitment effect on soil bacteria. Further, discontinuities between the composition of the rhizosphere in season one and season two suggest that external effects may be more significant than host-plant recruitment in structuring the rhizosphere microbiome. In an agricultural setting, these effects include tilling and environmental change (Huang et al. 2018, Moreno-Espíndola et al. 2018, Yin et al. 2017, Yurgel et al. 2017). This phenomenon may help to explain the apparent failure of the helpers to promote mycorrhizal colonisation in the pasture from which they were isolated, which had recently experienced profound environmental change in the form of extensive waterlogging. While the recruitment effect was less pronounced than has been reported for other host plants (Bakker et al. 2013, Kristin and Miranda 2013), the present evidence points to a possible selection effect, including the enrichment of certain members of the families Micrococcaceae and Microbacteriaceae, and the genus Pseudomonas, in the biserrula rhizosphere. High between-replicate variation was also noted, suggesting that failure to replicate, a practice that has been common (Prosser 2010), may have exaggerated certain recruitment effects. In comparing cultured and uncultured samples, it was observed that culturing resulted in microbial communities dominated by small cohorts of highly responsive taxa, accompanied by markedly reduced overall diversity. This was consistent with expectations: it is estimated that <1% of all soil bacterial taxa can be recovered in conventional cultures (Burke et al. 2008, Leveau et al. 2007). While additional methods of cultivation are available, including targeted cultures and novel technologies such as the ichip (Nichols et al. 2010, Berdy et al. 2017), the approach to identifying MHB taxa has often depended on bulk screening of cultured bacteria recovered using a single culture medium, such as TSA (Frey-Klett et al. 2007). The results obtained in the present study suggest that this process may yield taxa that are, in real-world conditions, ecologically irrelevant to the AM-plant interaction by virtue of occurring only in low numbers. In the present case, it was observed that Bacillus species, including B. cereus, were enriched in cultures, whereas evidence from diversity profiling suggests that these taxa are more likely to be suppressed in real-world biserrula pastures. The response of a bacterium to the addition of a culture medium is governed by the rate at which it can

Chapter 6. Summary and conclusions 153 adjust its proteome to suit its new environment (Pavlov and Ehrenberg 2013); taxon- specific variations in this rate likely account for why certain bacteria thrive and others decline following enrichment. While these results suggest that the taxa obtained by culture-dependent methods may not be selectively recruited to the plant rhizosphere, which may have implications for their use in agriculture, it should be noted that Bacillus spp. were still observed to promote mycorrhization in sterile and non-sterile conditions.

6.5. Future research A number of avenues for future research suggest themselves, both in pursuit of effective MHB inoculants in biserrula and in deepening our understanding of the helper effect more generally. Questions still to be answered concern the practical efficacy of helper strains, the ecology and physiology of the symbiotic partners in the MHB interaction and the optimal means of implementing them in practical situations.

6.5.1. Diversity and efficacy of MHB strains  Quantification by assessment of plant growth parameters of the benefits provided by AM associations to biserrula hosts, in terms of nutritional improvements and resistance to biotic and abiotic stress. Identifying an optimal level of colonisation with respect to plant biomass would be valuable.  Inoculation of biserrula with propagules of additional AM species to determine the range of fungi capable of forming mycorrhizas in this host.  Identification of additional helpers, including strains sourced from the native range of biserrula (Mediterranean Europe and North and East Africa).  Investigation of the persistence of helper bacteria in the biserrula rhizosphere over time. This would complement the findings concerning rhizosphere stability addressed in the present research project. More detailed study of the localisation of helper strains in the root would likewise be valuable.

6.5.2. Ecology and physiology of MHB strains  Screening of the helper bacteria for additional functional traits. Numerous alternative traits have been identified, including IAA and siderophore production, that, although poorly correlated with MHB activity as a whole, may account for certain individual helper effects. This could be extended to

Chapter 6. Summary and conclusions 154 include community-based effects to account for microbial interactions in the rhizosphere.  Studies into in vitro carbon acquisition by F. mosseae in contact with bacteria may help to explain the formation of secondary spores observed here. Owing to the scarcity of viable spores, this study may warrant repetition on a larger scale.  Continued exploration of the potential role of Nod factor analogues in promoting mycorrhization, including attempted isolation of these factors from helper strains suspected of harbouring them and development of suitable primers targeting Nod factors in a wider range of host species.

6.5.3. Implementation of MHB strains in agriculture  Validation of field colonisation results in additional pasture environments to reflect a diversity of land-use histories and edaphic factors.  Inoculation of MHB strains in situ to determine their practical efficacy.  Screening of the identified helpers in alternative plant hosts, particularly species grown in rotation with biserrula (e.g. wheat) or in mixed pastures with biserrula (e.g. Medicago and Trifolium species). This would allow for the selection of helpers that conferred benefits across the entire agricultural enterprise, not simply a single phase of a rotation.  Studies exploring alternative means of practical implementation of MHB strains in agriculture, for example as seed coats or in dual inoculation with rhizobia.  Variation in the applied dosage of the MHB strains to extrapolate a dose- response curve for each helper.

6.6. Conclusion The investigation into the role of mycorrhiza helper bacteria in facilitating colonisation of biserrula by arbuscular mycorrhizas presented in this thesis identified microorganisms that, if implemented into agricultural practice, may help to maximise the growth potential of a pasture legume typically grown in environments of high abiotic stress. Screening of plant-associated bacteria for helper effects necessarily entailed the use of culture-dependent methods. The complementary use of culture-

Chapter 6. Summary and conclusions 155 independent diversity profiling allowed those results to be situated in their proper context, in the full complexity of the rhizosphere. Initial experimental work was concerned with establishing whether biserrula was capable of forming mycorrhizas; subsequently, attention was focused on identifying the fungal partners involved in these symbioses. The use of microscopic methods, principally the gridline intersection method, indicated that rates of colonisation in biserrula were consistently low in both field conditions and in synthetic trap cultures. Pairing these methods with a genetic diversity profiling approach facilitated the recovery of a suite of cryptic species that were not observed during microscopic examination or spore isolation and could not otherwise be quantified, and challenged previous observations suggesting that the roots were dominated by a single AM fungus, F. mosseae. The results of these experiments suggested that biserrula may represent an ideal candidate for microbial enhancement via MHB activity, assuming the low values for mycorrhizal colonisation observed here can be validated in additional field sites. Subsequent screening of root-associated bacteria from biserrula plants for helper activity was necessarily dependent on culture-based methods. These revealed that a number of taxa isolated from the rhizoplane, root endosphere, and nodules of biserrula could act to promote pre-symbiotic growth of F. mosseae in vitro and colonisation of the biserrula root in vivo. The magnitude of these effects often exceeded previous reports, suggesting that a productive synergy between these helpers and AM fungi could be implemented to enhance biserrula growth. Genetic diversity profiling offered indications that some, but not all, of the identified helper bacteria may have been selectively recruited to the rhizosphere of biserrula. This may aid practical implementation of these microorganisms in an agricultural setting. On the other hand, certain candidate MHB strains appeared to be suppressed in the rhizosphere. Diversity profiling also gave an indication of the complexity of the rhizosphere and the number of interspecific interactions any helper strain inoculated into that environment is likely to be subjected to. To determine whether these interactions were likely to impede or negate the activity of the identified helpers, a screening exercise was conducted using a bacterial filtrate from biserrula roots to simulate a rhizosphere. This experiment demonstrated that, while the rhizosphere could, in some cases, diminish the magnitude of the helper effect, it did not eliminate it entirely. Indeed, the rhizosphere filtrate was itself observed to promote mycorrhization. Similarly, application of a biserrula-specific rhizobial inoculant was found to increase mycorrhizal colonisation.

Chapter 6. Summary and conclusions 156 Co-inoculation of this solution with MHB strains did not influence the helper effect but did, in the case of the P. koreensis subgroup taxon, increase the rate of nodulation. This observation suggests that certain helpers may stimulate genes responsible for both the mycorrhizal and rhizobial symbioses. Simulating a two-season pasture rotation allowed for the identification of a number of issues that are likely to complicate efforts to implement these MHB strains in pastures. The capacity of environmental change to disrupt the rhizosphere microbiota suggests that the practical application of helper species in agriculture may be contingent on a thorough understanding of the behaviour of these bacteria in the rhizosphere, namely their response to change in season, host plant, or any additional factor that may upset the equilibrium of the rhizosphere microbiota. Proper coordination of rotations with the application of MHB inoculum appears to be necessary and this requires additional investigation. Similarly, while the results presented here offer insights into how certain taxa may exert their helper effects – for example, by stimulation of the common symbiotic pathway – many observed helper effects cannot readily be explained with reference to the commonly proposed mechanisms of helper action, such as Type III secretion or enzymatic activity. This, too, requires further investigation. This project has demonstrated that certain root-associated bacteria are capable of promoting mycorrhization in biserrula and that this effect, though it may in some cases be diminished, can overcome the potential antagonism associated with microbial competition in the rhizosphere. Given the low rates of mycorrhizal colonisation recorded for untreated biserrula, including plants grown in a field setting, this represents a significant step towards maximising the productivity of biserrula pastures.

Chapter 6. Summary and conclusions 157 References

 Abby, S. S. et al. Identification of protein secretion systems in bacterial genomes. Sci. Rep. 16, 23080; 10.1038/srep23080 (2016)  Abdel-Fattah, G. M. & Mohamedin, A. H. Interactions between a vesicular- arbuscular mycorrhizal fungus (Glomus intraradices) and Streptomyces coelicolor and their effects on sorghum plants grown in soil amended with chitin of brawn scales. Biol. Fertil. Soils 32, 401–409; 10.1007/s003740000269 (2000)  Abdel-Lateif, K., Bogusz, D. & Hocher V. The role of flavonoids in the establishment of plant roots endosymbioses with arbuscular mycorrhiza fungi, rhizobia and Frankia bacteria. Plant Signal. Behav. 7, 636–41; 10.4161/psb.20039 (2012)  Abdellatif, L., Lokuruge, P. & Hamel, C. Axenic growth of the arbuscular mycorrhizal fungus Rhizophagus irregularis and growth stimulation by coculture with plant growth-promoting rhizobacteria. Mycorrhiza 29, 591–598; 10.1007/s00572-019-00924-z (2019)  Agnolucci, M., Battini, F., Cristani, C. & Giovannetti, M. Diverse bacterial communities are recruited on spores of different arbuscular mycorrhizal fungal isolates. Biol. Fertil. Soils 51, 379–389; 10.1007/s00374-014-0989-5 (2015)  Alam, M. et al. Synergistic effect of arbuscular mycorrhizal fungi and Bacillus subtilis on the biomass and essential oil yield of rose-scented geranium (Pelargonium graveolens). Arch. Agron. Soil Sci. 57, 889–898, 10.1080/03650340.2010.498013 (2011)  Alberdi, A. & Gilbert, M. T. P. hilldiv: an R package for the integral analysis of diversity based on Hill numbers. bioRxiv 10.1101/545665 (2019)  Alguacil, M. M. et al. The impact of tillage practices on arbuscular mycorrhizal fungal diversity in subtropical crops. Ecol. Appl. 18, 527–36; 10.1890/07- 0521.1 (2008)  Altankhuu, K. & Kim, J. Massilia solisilvae sp. nov., Massilia terrae sp. nov. and Massilia agilis sp. nov., isolated from forest soil in South Korea by using a newly developed culture method. Int. J. Syst. Evol. Microbiol. 67, 3026-3032; 10.1099/ijsem.0.002076 (2017)  Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410; 10.1016/S0022- 2836(05)80360-2 (1990)

References 158  Alzubaidy, H. et al. Rhizosphere microbiome metagenomics of gray mangroves (Avicennia marina) in the Red Sea. Gene 576, 626–36; 10.1016/j.gene.2015.10.032 (2016)  von Alten, H., Lindemann, A. & Schönbeck, F. Stimulation of vesicular- arbuscular mycorrhiza by or rhizosphere bacteria. Mycorrhiza 2, 167–173; 10.1007/BF00210586 (1993)  Andrade, G., Mihara, K. L., Linderman, R. G. & Bethlenfalvay G. J. Bacteria from rhizosphere and hyphosphere soils of different arbuscular-mycorrhizal fungi. Plant Soil 192, 71–79; 10.1023/A:1004249629643 (1997)  Ansari, F. A. & Ahmad, I. Fluorescent Pseudomonas -FAP2 and Bacillus licheniformis interact positively in biofilm mode enhancing plant growth and photosynthetic attributes. Sci. Rep. 9, 4547; 10.1038/s41598-019-40864-4 (2019)  Arthurson, V., Hjort, K., Muleta, D., Jäderlund, L. & Granhall, U. Effects on Glomus mosseae root colonization by Paenibacillus polymyxa and Paenibacillus brasilensis strains as related to soil P-availability in winter wheat. Appl. Environ. Soil Sci. 2011, Article 298097; 10.1155/2011/298097 (2011)  Artursson, V., Finlay, R. D. & Jansson, J. K. (2006) Interactions between arbuscular mycorrhizal fungi and bacteria and their potential for stimulating plant growth. Environ. Microbiol. 8, 1–10; 10.1111/j.1462-2920.2005.00942.x  Aspray, T. J. et al. Mycorrhization helper bacteria: a case of specificity for altering ectomycorrhiza architecture but not ectomycorrhiza formation. Mycorrhiza 16, 533–541; 10.1007/s00572-006-0068-3 (2006)  Aspray, T. J., Jones, E. E., Davies, M. W., Shipman, M. & Bending, G. D. Increased hyphal branching and growth of ectomycorrhizal fungus Lactarius rufus by the helper bacterium Paenibacillus sp. Mycorrhiza 23, 403–410; 10.1007/s00572-013-0483-1 (2013)  Aspray, T. J., Jones, E. E., Whipps, J. M. & Bending, G. D. Importance of mycorrhization helper bacteria cell density and metabolite localization for the Pinus sylvestris-Lactarius rufus symbiosis. FEMS Microbiol. Ecol. 56, 25–33; 10.1111/j.1574-6941.2005.00051.x (2006)  Azcón, R. Germination and hyphal growth of Glomus mosseae in vitro: effects of rhizosphere bacteria and cell-free culture media. Soil Biol Biochem 19, 417– 419; 10.1016/0038-0717(87)90032-0 (1987)  Azcón, R., Rubio, R. & Barea, J. M. Selective interactions between different species of mycorrhizal fungi and Rhizobium meliloti strains, and their effects

on growth, N2-fixation (15N) and nutrition of Medicago sativa L. New Phytol 117, 399–404; 10.1111/j.1469-8137.1991.tb00003.x (1991)  Babana, A. H. & Antoun, H. Biological system for improving the availability of Tilemsi phosphate rock for wheat (Triticum aestivum L.) cultivated in Mali. Nutr. Cycl. Agroecosyst. 72, 147–157; 10.1007/s10705-005-0241-7 (2005)

References 159  Bagyaraj, D. J., Menge, J. A. Interaction between a VA mycorrhiza and Azotobacter and their effects on the rhizosphere microflora and plant growth. New Phytol. 80, 567–573; 10.1111/j.1469-8137.1978.tb01588.x (1978)  Bais, H. P., Fall, R. & Vivanco, J. M. Biocontrol of Bacillus subtilis against infection of arabidopsis roots by Pseudomonas syringae is facilitated by biofilm formation and surfactin production. Plant Physiol. 134, 307–319; 10.1104/pp.103.028712 (2004)  Bakker, P. A. H. M., Berendsen, R. L., Doornbos, R. F., Wintermans, P. C. A. & Pieterse, C. M. J. The rhizosphere revisited: root microbiomics. Front Plant Sci. 4, Article 165; 10.3389/fpls.2013.00165 (2013)  Balestrini, R. & Bonfante, P. Cell wall remodeling in mycorrhizal symbiosis: a way towards biotrophism. Front Plant. Sci. 5, Article 237; 10.3389/fpls.2014.00237 (2014)  Banik, B. K., Durmic, Z., Erskine, W. & Revell, C. Anti-methanogenic advantage of biserrula (Biserrula pelecinus) over subterranean clover (Trifolium subterraneum) from in vitro fermentation is maintained across growth stages and cutting treatments. Crop Pasture Sci. 70, 263–272; 10.1071/CP18069 (2019)  Banik, B. K. et al. Variability of in vitro ruminal fermentation and methanogenic potential in the pasture legume biserrula (Biserrula pelecinus L.). Crop Pasture Sci. 64, 409–416; 10.1071/CP13073 (2013)  Barahona, E. et al. Efficient rhizosphere colonization by Pseudomonas fluorescens f113 mutants unable to form biofilms on abiotic surfaces. Environ. Microbiol. 12, 3185–3195; 10.1111/j.1462-2920.2010.02291.x (2010)  Barea, J. M. et al. Impact on arbuscular mycorrhiza formation of Pseudomonas strains used as inoculants for biocontrol of soil-borne fungal plant pathogens. Appl. Environ. Microbiol. 64, 2304–2307 (1998)  Barea, J-M., Pozo, M. J., Azcón, R. & Azcón-Aguilar, C. Microbial co- operation in the rhizosphere. J. Exp. Bot. 56, 1761–1778; 10.1093/jxb/eri197 (2005)  Barret, M., Egan, F. & O’Gara, F. Distribution and diversity of bacterial secretion systems across metagenomic datasets. Env. Microbiol. Rep. 5, 117– 126; 10.1111/j.1758-2229.2012.00394.x (2013)  Bashan, Y. Inoculants of plant growth-promoting bacteria for use in agriculture. Biotechnol. Adv. 16, 729-770; 10.1016/S0734-9750(98)00003-2 (1998)  Bashan, Y., de-Bashan, L. E., Prabhu, S. R. & Hernandez, J. P. Advances in plant growth-promotinh bacterial inoculant technology: formulations and practical perspectives (1998-2013). Plant Soil 378, 1–33; 10.1007/s11104-013- 1956-x (2014)  Bending, G. D., Poole, E. J., Whipps, J. M. & Read, D. J. Characterisation of bacteria from Pinus sylvestris– mycorrhizas and their effects on root–fungus interactions and plant growth. FEMS Microbiol. Ecol. 39, 219– 227; 10.1111/j.1574-6941.2002.tb00924.x (2002)

References 160  Berdy, B., Spoering, A. L., Ling, L. L., Epstein, S. S. In situ cultivation of previously uncultivable microorganisms using the ichip. Nat. Protoc. 12, 2232- 2242; 10.1038/nprot.2017.074 (2017)  Beringer, J. E. R factor transfer in Rhizobium leguminosarum. J. Gen. Microbiol. 84, 188–198; 10.1099/00221287-84-1-188 (1974)  Berruti, A., Bianciotto, V. & Lumini, E. Seasonal variation in winter wheat field soil arbuscular mycorrhizal fungus communities after non-mycorrhizal crop cultivation. Mycorrhiza 28, 535–548; 10.1007/s00572-018-0845-9 (2018)  Berruti A, Lumini E, Balestrini R, Bianciotto V (2016) Arbuscular mycorrhizal fungi as natural biofertilizers: let’s benefit from past successes. Front. Microbiol. 6, Article 1559; 10.3389/fmicb.2015.01559  Besserer, A., et al. Strigolactones stimulate arbuscular mycorrhizal fungi by activating mitochondria. PLoS Biol. 4, e226; 10.1371/journal.pbio.0040226 (2006)  Bianciotto, V., Lumini, E., Bonfanta, P. & Vandamme, P. ‘Candidatus Glomeribacter gigasporarum’ gen. nov., sp. nov., an of arbuscular mycorrhizal fungi. Int. J. Syst. Evol. Microbiol. 53, 121–124; 10.1099/ijs.0.02382-0 (2003)  Bianciotto, V., Lumini, E., Orgiazzi, A., Borriello, R. & Bonfante, P. Metagenomics applied to arbuscular mycorrhizal fungal communities in Metagenomics: current innovations and future trends (ed. Marco, D.) 161–178 (Caister Academic Press, 2011)  Bonfante, P. & Anca, I. A. Plants, mycorrhizal fungi, and bacteria: a network of interactions. Annu. Rev. Microbiol. 63, 363–83; 10.1146/annurev.micro. 091208.073504 (2009)  Bonfante, P. & Requena, N. Dating in the dark: how roots respond to fungal signals to establish arbuscular mycorrhizal symbiosis. Curr. Opin. Plant Biol. 14, 451–7; 10.1016/j.pbi.2011.03.014 (2011)  van den Boogaart, K. G., Tolosana-Delgado, R. & Bren, M. compositions: Compositional Data Analysis. R package version 1.40-3. https://CRAN.R- project.org/package=compositions (2019)  Borowicz, V. A. Do arbuscular mycorrhizal fungi alter plant–pathogen relations? Ecology 82, 3057–3068; 10.1890/0012-9658(2001)082[3057: DAMFAP]2.0.CO;2 (2001)  Bourles, A. et al. Co-inoculation with a bacterium and arbuscular mycorrhizal fungi improves root colonization, plant mineral nutrition, and plant growth of a Cyperaceae plant in an ultramafic soil. Mycorrhiza. Advance online publication; 10.1007/s00572-019-00929-8 (2020).  Bowen, G. D. & Theodorou, C. Interactions between bacteria and ectomycorrhizal fungi. Soil Biol. Biochem. 11, 119–126. (1979)  Boyd, C. D. et al. Structural features of the Pseudomonas fluorescens biofilm adhesin LapA required for LapG-dependent cleavage, biofilm formation, and cell surface localization. J. Bacteriol. 196, 2775–2788; 10.1128/JB.01629-14 (2014)

References 161  Brakhage, A. A. & Schroeckh, V. Fungal secondary metabolites – Strategies to activate silent gene clusters. Fungal Genet. Biol. 48, 15–22; 10.1016/j.fgb.2010.04.004 (2011)  Bresson, J., Varoquaux, F., Bontpart, T., Touraine, B. & Vile, D. The PGPR strain Phyllobacterium brassicacearum STM196 induces a reproductive delay and physiological changes that result in improved drought tolerance in Arabidopsis. New Phytol. 200, 558–569; 10.1111/nph.12383 (2013)  Briard, B., Heddergott, C. & Latgé, J. Volatile compounds emitted by Pseudomonas aeruginosa stimulate growth of the fungal pathogen Aspergillus fumigatus. mBio 7, e00219-16; 10.1128/mBio.00219-16 (2016)  Brulé, C. et al. Survival in the soil of the ectomycorrhizal fungus Laccaria bicolor and effect of a mycorrhiza helper Pseudomonas fluorescens. Soil Biol. Biochem. 33, 1683–1694; 10.1016/S0038-0717(01)00090-6 (2001)  Brundrett, MC. Methods for identifying mycorrhizas. Mycorrhizal associations: the web resource https://mycorrhizas.info/method.html (2008)  Brundrett, M. C. Mycorrhizal associations and other means of nutrition of vascular plants: understanding the global diversity of host plants by resolving conflicting information and developing reliable means of diagnosis. Plant Soil 320, 37–77; 10.1007/s11104-008-9877-9 (2009)  Brundrett, M. C. & Tedersoo, L. Evolutionary history of mycorrhizal symbioses and global host plant diversity. New Phytol. 220, 1108–1115; 10.1111/nph.14976 (2018)  Budi, S. W., van Tuinen, D., Martinotti, G. & Gianinazzi, S. Isolation from the Sorghum bicolor mycorrhizosphere of a bacterium compatible with arbuscular mycorrhiza development and antagonistic towards soilborne fungal pathogens. Appl. Environ. Microbiol. 65, 5148–5150 (1999)  Buée, M., De Boer, W., Martin, F., van Overbeek, L. & Jurkevitch, E. The rhizosphere zoo: An overview of plant-associated communities of microorganisms, including phages, bacteria, archaea, and fungi, and of some of their structuring factors. Plant Soil 321, 189–212; 10.1007/s11104-009-9991-3 (2009)  Bünemann, E. K., Marschner, P., Smernik, R. J., Conyers, M. & McNeill, A. M. Soil organic phosphorus and microbial community composition as affected by 26 years of different management strategies; 10.1007/s00374-007-0254-2 (2008)  Burdette, L.A. et al. Developing Gram-negative bacteria for the secretion of heterologous proteins. Microb. Cell Fact. 17, 196; 10.1186/s12934-018-1041-5 (2018)  Burke, D. J., Dunham, S. M. & Kretzer, A. M. Molecular analysis of bacterial communities associated with the roots of Douglas fir (Pseudotsuga menziesii) colonized by different ectomycorrhizal fungi. FEMS Microbiol. Ecol. 65, 299– 309; 10.1111/j.1574-6941.2008.00491.x (2008)  Butturi-Gomes, D., Petrere, M., Giacomini, H. C. & Zocchi, S. S. Statistical performance of a multicomparison method for generalized species diversity

References 162 indices under realistic empirical scenarios. Ecol. Indic. 72, 542-552; 10.1016/j.ecolind.2016.08.054 (2017)  Cairns, J. et al. Construction and characterization of synthetic bacterial community for experimental ecology and . Front. Genet. 9, 312; 10.3389/fgene.2018.00312 (2018)  Cameron, D. D., Neal, A. L., van Wees, S. C. M. & Ton, J. Mycorrhiza- induced resistance: more than the sum of its parts? Trends Plant Sci. 18, 539– 545; 10.1016/j.tplants.2013.06.004 (2013)  Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 7, 335–336; 10.1038/nmeth.f.303 (2010)  Carberry, P., Bruce, S., Walcott, J. & Keating, B. Innovation and productivity in dryland agriculture: a return-risk analysis for Australia. J. Agric. Sci. 149, 77–89; 10.1017/S0021859610000973 (2011)  Carr, S. J., Loi, A., Howieson, J. H. & Porqueddu, C. Attributes of Biserrula pelecinus L. (biserrula): A new pasture legume for sustainable farming on acidic sandy soils in Mediterranean environments in Dynamics and sustainability of Mediterranean pastoral systems (ed. Etienne, M) 87–90 (CIHEAM, 1999)  Cavagnaro, T. R. The role of arbuscular mycorrhizas in improving plant zinc nutrition under low soil zinc concentrations: a review. Plant Soil 304, 315–325; 10.1007/s11104-008-9559-7 (2008)  Chang, J. H., Desveaux, D. & Creason, A. L. The ABCs and 123s of bacterial secretion systems in plant pathogenesis. Annu. Rev. Phytopathol. 52, 317–345; 10.1146/annurev-phyto-011014-015624 (2014)  Chaparro, J. M., Badri, D. V. & Vivanco, J. M. Rhizosphere microbiome assemblage is affected by plant development. ISME J. 84, 790-803; 10.1038/ismej.2013.196 (2014)  Chelius, M. K. & Triplett, E. W. Dyadobacter fermentans gen. nov., sp. nov., a novel gram-negative bacterium isolated from surface-sterilized Zea mays stems. Int. J. Syst. Evol. Microbiol. 50, 751-758; 10.1099/00207713-50-2-751 (2000)  Chen, L., Zou, Y., She, P. & Wu, Y. Composition, function, and regulation of T6SS in Pseudomonas aeruginosa. Microbiol. Res. 172, 19–25; 10.1016/j.micres.2015.01.004 (2015)  Chen, M., Arato, M., Borghi, L., Nouri, E. & Reinhardt, D. Beneficial services of arbuscular mycorrhizal fungi – from ecology to application. Front. Plant Sci. 9, 1270; 10.3389/fpls.2018.01270 (2018)  Cho, S. M., Kang, B. R. & Kim, Y. C. Transcriptome analysis of induced systemic drought tolerance elicited by Pseudomonas chlororaphis O6 in Arabidopsis thaliana. Plant Pathol. J. 29, 209–220; 10.5423/PPJ.SI.07.2012.0103 (2013)  Clark, K., Karsch-Mizrachi, I., Lipman, D. J., Ostell, J. & Sayers, E. W. GenBank. Nucleic Acids Res. 4, D67–72; 10.1093/nar/gkv1276 (2016)

References 163  Coburn, B., Sekirov, I. & Finlay, B. B. Type III secretion systems and disease. Clin. Microbiol. Rev. 20, 535–549; 10.1128/CMR.00013-07 (2007)  Conn, V. M. & Franco, C. M. Analysis of the endophytic actinobacterial population in the roots of wheat (Triticum aestivum L.) by terminal restriction fragment length polymorphism and sequencing of 16S rRNA clones. Appl. Environ. Microbiol. 70, 1787–1794; 10.1128/aem.70.3.1787-1794.2004 (2004)  Coombs J. T. & Franco, C. M. Isolation and identification of actinobacteria from surface-sterilized wheat roots. Appl. Environ. Microbiol. 69, 5603–5608; 10.1128/aem.69.9.5603-5608.2003 (2003)  Coombes, N. DiGGer 1.0.2. New South Wales Department of Primary Industries Biometrics http://nswdpibiom.org/austatgen/software/ [Accessed 13 May 2018] (2002)  Corrêa, A., Cruz, C. & Ferrol, N. Nitrogen and carbon/nitrogen dynamics in arbuscular mycorrhiza: the great unknown. Mycorrhiza 25, 499–515; 10.1007/ s00572-015-0627-6 (2015)  Cosme, M., Fernández, I., van der Heijden, M. & Pieterse, C. M. J. Non- mycorrhizal plants: the exceptions that prove the rule. Trends Plant Sci. 23, 557–587; 10.1016/j.tplants.2018.04.004 (2018)  Costa, T. R. D., Felisberto-Rodrigues, C., Meir, A. & Prevost, M. S. Secretion systems in Gram-negative bacteria: structural and mechanistic insights. Nature Rev. Microbiol. 13, 343–359; 10.1038/nrmicro3456 (2015)  Cusano, A. M. et al. Pseudomonas fluorescens BBc6R8 type III secretion mutants no longer promote ectomycorrhizal symbiosis. Environ. Microbiol. Rep. 3, 203–210; 10.1111/j.1758-2229.2010.00209.x (2011)  D’Haeze, W. & Holsters, M. Nod factor structures, responses, and perception during initiation of nodule development. Glycobiology 12, 79R-105R. 10.1093/glycob/12.6.79r (2002)  Danhorn, T. & Fuqua, C. Biofilm formation by plant-associated bacteria. Annu. Rev. Microbiol. 61, 401–22; 10.1146/annurev.micro.61.080706.093316 (2007)  Daniel, R. The metagenomics of soil. Nature Rev. Microbiol. 3, 470–478; 10.1038/nrmicro1160 (2005)  Davies, F. T. Jr, Potter, J. R. & Linderman, R. G. Mycorrhiza and repeated drought exposure affect drought resistance and extraradical hyphae development of pepper plants independent of plant size and nutrient content. J. Plant Physiol. 139, 289–294; 10.1016/S0176-1617(11)80339-1 (1992)  Deangelis, K. M. et al. Long-term forest soil warming alters microbial communities in temperate forest soils. Front. Microbiol. 6, 104; 10.3389/fmicb.2015.00104 (2015)  DeBruyn, J. M., Nixon, L. T., Fawaz, M. N., Johnson, A. M. & Radosevich, M. Global biogeography and quantitative seasonal dynamics of Gemmatimonadetes in soil. Appl. Environ. Microbiol. 77, 6295-6300; 10.1128/AEM.05005-11 (2011)

References 164  Decoin, V. et al. A type VI secretion system is involved in Pseudomonas fluorescens bacterial competition. PloS one 9, e89411; 10.1371/journal.pone.0089411 (2014)  Demetrius, L. Adaptive value, entropy and survivorship curves. Nature 275, 213–214; /10.1038/275213a0 (1978)  Dennis, P. G., Miller, A. J. & Hirsch, P. R. Are root exudates more important than other sources of rhizodeposits in structuring rhizosphere bacterial communities? FEMS Microbiol. Ecol. 72, 313–327; 10.1111/j.1574- 6941.2010.00860.x (2010)  Department of Environment and Primary Industries. Growing wheat. Agriculture Victoria http://agriculture.vic.gov.au/agriculture/grains-and-other- crops/crop-production/growing-wheat [Accessed 15 January 2018] (2012)  DeSantis, T. Z. et al. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Appl. Environ. Microbiol. 72, 5069– 5072; 10.1128/AEM.03006-05 (2006)  Deveau, A. et al. The mycorrhiza helper Pseudomonas fluorescens BBc6R8 has a specific priming effect on the growth, morphology and gene expression of the ectomycorrhizal fungus Laccaria bicolor S238N. New Phytol. 175, 743– 755; 10.1111/j.1469-8137.2007.02148.x (2007)  Deveau, A. et al. Role of fungal trehalose and bacterial thiamine in the improved survival and growth of the ectomycorrhizal fungus Laccaria bicolor S238N and the helper bacterium Pseudomonas fluorescens BBc6R8. Env. Microbiol. Rep. 2, 560–568; 10.1111/j.1758-2229.2010.00145.x (2010)  Deveau, A. et al. Genome sequence of the mycorrhizal helper bacterium Pseudomonas fluorescens BBc6R8. Genome Announc. 2, e01152-13; 10.1128/ genomeA.01152-13 (2014)  Dickson, S. The Arum–Paris continuum of mycorrhizal symbioses. New Phytol. 163, 187–200; 10.1111/j.1469-8137.2004.01095.x (2004)  Diepold, A. & Armitage, J. P. Type III secretion systems: the bacterial flagellum and the injectisome. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 370, 20150020; 10.1098/rstb.2015.0020 (2015)  Ditengou, F. A., Raudaskoski, M. & Lapeyrie, F. Hypaphorine, an indole-3- acetic acid antagonist delivered by the ectomycorrhizal fungus Pisolithus tinctorius, induces reorganisation of actin and the microtubule in Eucalyptus globulus ssp. bicostata root hairs. Planta 218, 217–25; 10.1007/s00425-003-1095-3 (2003)  van der Does, D. et al. Salicylic acid suppresses jasmonic acid signaling downstream of SCFCOI1-JAZ by targeting GCC promoter motifs via transcription factor ORA59. Plant Cell 25, 744–761; 10.1105/tpc.112.108548 (2013)  Dueck, T. A., Visser, P., Ernst, W. H. O. & Schat, H. Vesicular-arbuscular mycorrhizae decrease zinc-toxicity to grasses growing in zinc-polluted soil. Soil Biol. Biochem. 18, 331–333; 10.1016/0038-0717(86)90070-2 (1986)

References 165  Dunstan, W. A., Malajczuk, N. & Dell, B. Effects of bacteria on mycorrhizal development and growth of container grown Eucalyptus diversicolor F. Muell. seedlings. Plant Soil 201, 241–249. (1998)  Dunwell, J., Gibbings, J. G., Mahmood, T. & Saqlan Naqvi, S. M. Germin and germin-like proteins: evolution, structure, and function. Crit. Rev. Plant Sci. 27, 342–375; 10.1080/07352680802333938 (2008)  Duponnois, R. Les bacteries auxiliaires de la mycorhization du Douglas (Pseudotsuga menziesii (Mirb.) Franco) par souche S238. Thesis, University of Nancy I, France (1992)  Duponnois, R. & Garbaye, J. Some mechanisms involved in growth stimulation of ectomycorrhizal fungi by bacteria. Can. J. Bot. 68, 2148–2152; 10.1139/b90-280 (1990)  Duponnois, R. & Garbaye, J. Mycorrhization helper bacteria associated with the Douglas fir–Laccaria laccata symbiosis: effects in aseptic and in glasshouse conditions. Annales des Sciences Forestières 48, 239–251; 10.1051/forest:19910301 (1991)  Duponnois, R. & Kisa, M. The possible role of trehalose in the mycorrhiza helper effect. Can. J. Bot. 84, 1005–1008; 10.1139/B06-053 (2006)  Duponnois, R. & Plenchette, C. A mycorrhiza helper bacterium enhances ectomycorrhizal and endomycorrhizal symbiosis of Australian Acacia species. Mycorrhiza 13, 85–91; 10.1007/s00572-002-0204-7 (2003)  Echbab, H. et al. Helper bacteria associated with Casuarina cunninghamiana- Frankia symbiosis: selection of isolates for their effects on plant growth in axenic conditions. Acta Bot. Gallica 151, 429–440; 10.1080/12538078.2004.10515445 (2004)  Edgar, R. C. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461; 10.1093/bioinformatics/btq461 (2010)  Edgar, R. C. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10, 996–998; 10.1038/nmeth.2604 (2013)  Edgar, R. C., Haas, B. J., Clemente, J. C., Quince, C. & Knight, R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194– 2200; 10.1093/bioinformatics/btr381 (2011)  Eichinger, V. et al. EffectiveDB—updates and novel features for a better annotation of bacterial secreted proteins and Type III, IV, VI secretion systems. Nucleic Acids Res 44, D669–D674; 10.1093/nar/gkv1269 (2016)  Escudero, V. & Mendoza, R. Seasonal variation of arbuscular mycorrhizal fungi in temperate grasslands along a wide hydrologic gradient. Mycorrhiza 15, 291–299; 10.1007/s00572-004-0332-3 (2005)  Feng, Y. et al. Soil microbial communities under conventional-till and no-till continuous cotton systems. Soil Biol. Biochem. 35, 1693-1703; 10.1016/ j.soilbio.2003.08.016 (2003)  Fernandes, A. D. et al. Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective

References 166 growth experiments by compositional data analysis. Microbiome 2, 15; 10.1186/2049-2618-2-15 (2014)  Fernandes, A. D., Macklaim, J. M., Linn, T. G., Reid, G. & Gloor, GB. ANOVA-like differential gene expression analysis of single-organism and meta-RNA-Seq. PLoS ONE 8, e67019; 10.1371/journal.pone.0067019 (2013)  Fernández Bidondo, L., et al. Pre-symbiotic and symbiotic interactions between Glomus intraradices and two Paenibacillus species isolated from AM propagules. In vitro and in vivo assays with soybean (AG043RG) as plant host. Soil Biol. Biochem. 43, 1866–1872; 10.1016/j.soilbio.2011.05.004 (2011)  Fernández Bidondo, L., et al. Cultivable bacteria associated with infective propagules of arbuscular mycorrhizal fungi. Implications for mycorrhizal activity. Appl. Soil Ecol. 105, 86–90; 10.1016/j.apsoil.2016.04.013 (2016)  Fester, T., Maier, W. & Strack, D. Accumulation of secondary compounds in barley and wheat roots in response to inoculation with an arbuscular mycorrhizal fungus and co-inoculation with rhizosphere bacteria. Mycorrhiza 8, 241–246; 10.1007/s005720050240 (1999)  Filion, M., St-Arnaud, M. & Jabaji-Hare, S. H. Quantification of Fusarium solani f. sp. phaseoli in mycorrhizal bean plants and surrounding mycorrhizosphere soil using real-time polymerase chain reaction and direct isolations on selective media. Biol. Control 93, 229–235; 10.1094/PHYTO.2003.93.2.229 (2003)  Fomina, M. A., Alexander, I. J., Colpaert, J. V. & Gadd, G. M. Solubilization of toxic metal minerals and metal tolerance of mycorrhizal fungi. Soil Biol. Biochem. 37, 851–866; 10.1016/j.soilbio.2004.10.013 (2005)  Foo, E. Auxin influences strigolactones in pea mycorrhizal symbiosis. J. Plant Physiol. 170, 523–528; 10.1016/j.jplph.2012.11.002 (2013)  Founoune, H. et al. Mycorrhiza helper bacteria stimulate ectomycorrhizal symbiosis of Acacia holosericea with Pisolithus alba. New Phytol. 153, 81–89; 10.1046/j.0028-646X.2001.00284.x (2002a)  Founoune, H. et al. Interactions between ectomycorrhizal symbiosis and fluorescent pseudomonads on Acacia holosericea: isolation of mycorrhiza helper bacteria (MHB) from a Soudano-Sahelian soil. FEMS Microbiol. Ecol. 41, 37–46; 10.1111/j.1574-6941.2002.tb00964.x (2002)  Franche, C., Lindström, K. & Elmerich, C. Nitrogen-fixing bacteria associated with leguminous and non-leguminous plants. Plant Soil 321, 35–59; 10.1007/s11104-008-9833-8 (2009)  Franco, C. et al. Actinobacterial endophytes for improved crop performance. Australas. Plant. Pathol. 36, 524-531; 10.1071/AP07067 (2007)  Frey-Klett, P. et al. Bacterial-fungal interactions: hyphens between agricultural, clinical, environmental, and food microbiologists. Microbiol. Mol. Biol. Rev. 75, 583–609; 10.1128/MMBR.00020-11 (2011).  Frey-Klett, P., Churin, J. L., Pierrat, J. C. & Garbaye, J. Dose effect in the dual inoculation of an ectomycorrhizal fungus and a mycorrhiza helper bacterium in

References 167 two forest nurseries. Soil Biol. Biochem. 31, 1555–1562; 10.1016/S0038- 0717(99)00079-6 (1999)  Frey-Klett, P. et al. Ectomycorrhizal symbiosis affects functional diversity of rhizosphere fluorescent pseudomonads. New Phytol. 165, 317–328; 10.1111/j.1469-8137.2004.01212.x (2005)  Frey-Klett, P. & Garbaye, J. Mycorrhiza helper bacteria: a promising model for the genomic analysis of fungal–bacterial interactions. New Phytol. 168, 1–4; 10.1111/j.1469-8137.2005.01553.x (2005)  Frey-Klett, P., Garbaye, J. & Tarkka, M. The mycorrhiza helper bacteria revisited. New Phytol. 176, 22–36; 10.1111/j.1469-8137.2007.02191.x (2007)  Frey-Klett, P., Pierrat, J. C. & Garbaye, J. Location and survival of mycorrhiza helper Pseudomonas fluorescens during establishment of ectomycorrhizal symbiosis between Laccaria bicolor and Douglas fir. Appl. Environ. Microbiol. 63, 139–144; 10.1016/S0038-0717(99)00079-6 (1997)  Friel, C. A. & Friesen, M. L. Legumes modulate allocation to rhizobial nitrogen fixation in response to factorial light and nitrogen manipulation. Front. Plant Sci. 10, 1316; 10.3389/fpls.2019.01316 (2019)  Gaby, J. C. & Buckley, D. H. A comprehensive aligned nifH gene database: a multipurpose tool for studies of nitrogen-fixing bacteria. Database 2014, bau001; 10.1093/database/bau001 (2014)  Galvez, L., Douds, D., Drinkwater, L. & Wagoner, P. Effect of tillage and farming system upon VAM fungus populations and mycorrhizas and nutrient uptake of maize. Plant Soil 228, 299–398; 10.1023/A:1004810116854 (2001)  Gamalero, E. et al. Impact of two fluorescent pseudomonads and an arbuscular mycorrhizal fungus on tomato plant growth, root architecture and P acquisition. Mycorrhiza 14, 185–192; 10.1007/s00572-003-0256-3 (2004)  Garbaye, J. Biological interactions in the mycorrhizosphere. Experientia 47, 370–375 (1991)  Garbaye, J. Helper bacteria: a new dimension to the mycorrhizal symbiosis (Tansley Review No. 76). New Phytol. 128, 197–210; 10.1111/j.1469- 8137.1994.tb04003.x (1994)  Garbaye, J. & Bowen, G. D. Stimulation of mycorrhizal infection of Pinus radiata by some microorganisms associated with the mantle of ectomycorrhizas. New Phytol. 112, 383–388; 10.1111/j.1469- 8137.1989.tb00327.x (1989)  Garbeva, P., Hordijk, C., Gerards, S. & de Boer, W. Volatile-mediated interactions between phylogenetically different soil bacteria. Front. Microbiol. 5, 289; 10.3389/fmicb.2014.00289 (2014)  Garcia, J. & Kao-Kniffin, J. Microbial group dynamics in plant rhizospheres and their implications on nutrient cycling. Front. Microbiol. 9, 1516; 10.3389/fmicb.2018.01516 (2018)  Gdanetz, K. & Trail, F. The wheat microbiome under four management strategies, and potential for endophytes in disease protection. Phytobiomes J. 1, 158-168; 10.1094/PBIOMES-05-17-0023-R (2017)

References 168  Genre, A. et al. Short-chain chitin oligomers from arbuscular mycorrhizal fungi trigger nuclear Ca2+ spiking in Medicago truncatula roots and their production is enhanced by strigolactone. New Phytol. 198, 179–189; 10.1111/nph.12146 (2013)  Germida, J. J., Siciliano, S. D., Renato de Freitas, J. & Seib, A. M. Diversity of root‐associated bacteria associated with field‐grown canola (Brassica napus L.) and wheat (Triticum aestivum L.). FEMS Microbiol. Ecol. 26, 43–50; 10.1111/j.1574-6941.1998.tb01560.x (1998)  Giovannetti, M. & Mosse, B. An evaluation of techniques for measuring vesicular arbuscular mycorrhizal infection in roots. New Phytol. 84, 489–500; 10.1111/j.1469-8137.1980.tb04556.x (1980)  Gkarmiri, K. et al. Identifying the active microbiome associated with roots and rhizosphere soil of oilseed rape. Appl. Environ. Microbiol. 83, e01938-17; 10.1128/AEM.01938-17 (2017)  Gloor, G. ANOVA-Like Differential Expression tool forhigh throughput sequencing data. Bioconductor https://bioconductor.org/packages/devel/ bioc/vignettes/ALDEx2/inst/doc/ALDEx2_vignette.pdf [Accessed 7 February 2020] (2018)  Gloor, G. B., Macklaim, J. M. & Fernandes, A. D. Displaying variation in large datasets: a visual summary of effect sizes. J. Comput. Graph. Stat. 10.1080/10618600.2015.1131161 (2016)  Gloor, G. B., Macklaim, J. M., Pawlowsky-Glahn, V. & Egozcue, J. J. Microbiome datasets are compositional: and this is not optional. Front. Microbiol. 8, 2224; 10.3389/fmicb.2017.02224 (2017)  Goessweiner-Mohr, N., Arends, K., Keller, W. & Grohmann, E. Conjugative type IV secretion systems in Gram-positive bacteria. Plasmid 70, 289–302; 10.1016/j.plasmid.2013.09.005 (2013)  Gopinathan, S. & Raman, N. Indole 3-acetic acid production by ectomycorrhizal fungi. Indian J. Exp. Biol. 30, 142–143. (1992)  Grayston, S. J., Wang, S., Campbell, C. D. & Edwards, A. C. Selective influence of plant species on microbial diversity in the rhizosphere. Soil Biol. Biochem. 30, 369-378; 10.1016/S0038-0717(97)00124-7 (1998)  Green, E. R. & Mecsas, J. Bacterial secretion systems – An overview. Microbiol. Spectr. 4, VMBF-0012-2015; 10.1128/microbiolspec.VMBF-0012- 2015 (2016)  Green, S. J., Michel, F. C., Hadar, Y. & Minz, D. Contrasting patterns of seed and root colonization by bacteria from the genus Chryseobacterium and from the family Oxalobacteraceae. ISME J. 1, 291–299; 10.1038/ismej.2007.33 (2007)  Großkopf, T. & Soyer, O. S. Synthetic microbial communities. Curr. Opin. Microbiol. 18, 72-77; 10.1016/j.mib.2014.02.002 (2014)

References 169  Guasch-Vidal, B. et al. Nod factor production and abiotic stress in Rhizobium in Beneficial plant-microbial interactions: ecology and applications (eds. González, M. B. R. & Gonzalez-López, J.) 70–98 (CRC press, 2016)  Gullner, G., Komives, T., Király, L. & Schröder, P. Glutathione S-transferase enzymes in plant-pathogen interactions. Front. Plant Sci. 9, 1836; 10.3389/ fpls.2018.01836 (2018)  Guo, J. & Chi, J. Effect of Cd-tolerant plant growth-promoting rhizobium on plant growth and Cd uptake by Lolium multiflorum Lam. and Glycine max (L.) Merr. in Cd-contaminated soil. Plant Soil 375, 205–214; 10.1007/s11104-013- 1952-1 (2014)  Gutjahr, C. Phytohormone signaling in arbuscular mycorhiza development. Curr. Opin. Plant Biol. 20, 26–34; 10.1016/j.pbi.2014.04.003 (2014)  Gutjahr, C. & Parniske, M. Cell and developmental biology of arbuscular mycorrhiza symbiosis. Annu. Rev. Cell Dev. Biol. 29, 593–617. 10.1146/annurev-cellbio-101512-122413 (2013)  Hackney, B., Dear, B., Li, G., Rodham, C. & Tidd, J. Current and future use of pasture legumes in central and southern NSW – results of a farmer and advisor survey in Global issues, paddock action: proceedings of the 14th Australian agronomy conference (ed. Unkovich, M. J.) http://www.agronomyaustralia proceedings.org/index.php/2008 [Accessed 9 December 2017] (Australian Society of Agronomy, 2008)  Hackney, B., Rodham, C. & Piltz, J. Using biserrula to increase crop and livestock production. (Meat & Livestock Australia, 2013)  Haichar, F. Z. et al. Plant host habitat and root exudates shape soil bacterial community structure. ISME J. 2, 1221-1230; 10.1038/ismej.2008.80 (2008)  Hall, T. A. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acid S. 41, 95–98 (1999)  Handelsman, J., Rondon, M. R., Brady, S. F., Clardy, J. & Goodman, R. M. Molecular biological access to the chemistry of unknown 0095 SP soil microbes: a new frontier for natural products. Chem. Biol. 5, R245–9; 10.1016/s1074-5521(98)90108-9 (1998)  Hankin, L. & Anagnostakis, S.L. The use of solid media for detection of enzyme production by fungi. Mycologia 67, 597–607; 10.2307/3758395 (1975).  Hartman, K., van der Heijden, M.G., Roussely-Provent, V., Walser, J.-C. & Schlaeppi, K. Deciphering composition and function of the of a legume plant. Microbiome 5, 2; 10.1186/s40168-016-0220-z (2017)  Hartmann, R. M. et al. Insights into the complex role of GRAS transcription factors in the arbuscular mycorrhiza symbiosis. Sci. Rep. 9, 3360; 10.1038/s41598-019-40214-4 (2019)  Hegazi, N. A. et al. Plant-fed versus chemicals-fed rhizobacteria of lucerne: plant-only teabags culture media not only increase culturability of rhizobacteria but also recover a previously uncultured Lysobacter sp., Novosphingobium sp.

References 170 and Pedobacter sp.. PLoS ONE 12, e0180424; 10.1371/journal.pone.0180424 (2017)  van der Heijden, M. G. A., Martin, F. M., Selosse, M. A. & Sanders, I. R. Mycorrhizal ecology and evolution: the past, the present, and the future. New Phytol. 205, 1406–1423; 10.1111/nph.13288 (2015)  Helander, N. et al. Glyphosate decreases mycorrhizal colonization and affects plant-soil feedback. Sci. Total Environ. 642, 285–291; 10.1016/j. scitotenv.2018.05.377 (2018)  Helgason, B. L., Walley, F. L. & Germida, J. J. Long-term no-till management affects microbial biomass but not community composition in Canadian prairie agroecosystems. Soil. Biol. Biochem. 42, 2192-2202; 10.1016/j.soilbio. 2010.08.015 (2010)  Hermans, S.M. et al. Using soil bacterial communities to predict physico- chemical variables and soil quality. Microbiome 8, 79; 10.1186/s40168-020- 00858-1 (2020)  Hernández, G., Cuenca, G. & Garcia, A. Behaviour of arbuscular-mycorrhizal fungi on Vigna luteola growth and its effect on the exchangeable (P-32) phosphorus of soil. Biol. Fertil. Soils 31, 232–236; 10.1007/s003740050650 (2000)  Hernández-González, I. L. & Olmedo-Álvarez, G. Draft whole-genome sequence of the type strain Bacillus aquimaris TF12T. Genome Announc. 4, e00640-16; 10.1128/genomeA.00640-16. (2016)  Hildebrandt, U., Janetta, K. & Bothe, H. Towards growth of arbuscular mycorrhizal fungi independent of a plant host. Appl. Environ. Microbiol. 68, 1919–1924; 10.1128/AEM.68.4.1919-1924.2002 (2002)  Hildebrandt, U., Ouziad, F., Marner, F. J. & Bothe, H. The bacterium Paenibacillus validus stimulates growth of the arbuscular mycorrhizal fungus Glomus intraradices up to the formation of fertile spores. FEMS Microbiol. Lett. 254, 258–267; 10.1111/j.1574-6968.2005.00027.x (2006)  Hirsch, P. R. & Mauchline, T. H. Who's who in the plant root microbiome? Nature Biotechnol 30, 961–962; 10.1038/nbt.2387 (2012)  Ho, A., Di Lonardo, P. & Bodelier, P. L. E. Revisiting life strategy concepts in environmental microbial ecology. FEMS Microbiol. Ecol. 93, fix006; 10.1093/ femsec/fix006 (2017)  Hodge, A. & Storer, K. Arbuscular mycorrhiza and nitrogen: implications for individual plants through to ecosystems. Plant Soil 386, 1–19; 10.1007/s11104- 014-2162-1 (2015)  Hoeksema, J. D. et al. A meta-analysis of context-dependency in plant response to inoculation with mycorrhizal fungi. Ecol. Lett. 13, 394–407; 10.1111/j.1461-0248.2009.01430.x (2010)  Hol, W. H. G. et al. Non‐random species loss in bacterial communities reduces antifungal volatile production. Ecology 96, 2042–2048; 10.1890/14-2359.1 (2015)

References 171  Honda, H., Sugiyama, H., Saito, I. & Kobayashi, T. High cell density culture of Rhodococcus rhodochrous by pH-stat feeding and dibenzothiophene degradation. J. Ferment. Bioeng. 85, 334-338; 10.1016/S0922-338X(97)85685- 1 (1998)  Howieson, J. & Ballard, R. Optimising the legume symbiosis in stressful and competitive environments within southern Australia—some contemporary thoughts. Soil Biol. Biochem. 36, 1261–1273; 10.1016/j.soilbio.2004.04.008 (2004)  Howieson J., O’Hara G. W. & Carr S. J. Changing roles for legumes in Mediterranean agriculture: developments from an Australian perspective. Field Crop Res. 65, 107–122; 10.1016/S0378-4290(99)00081-7 (2000)  Huang, G., Li, L., Su, Y. & Li, Y. Differential seasonal effects of water addition and nitrogen fertilization on microbial biomass and diversity in a temperate desert. CATENA 161, 27-36; 10.1016/j.catena.2017.09.030 (2018)  Hug, L. A. Sizing up the uncultured microbial majority. mSystems 3, e00185- 18; 10.1128/mSystems.00185-18 (2018)  Imperiali, N. et al. Combined field inoculations of Pseudomonas bacteria, arbuscular mycorrhizal fungi, and entomopathogenic nematodes and their effects on wheat performance. Front. Plant Sci. 8, 1809; 10.3389/fpls.2017.01809 (2017)  Institut national de la recherche agronomique [INRA]. Mycorrhiza manual prepared for the workshop “Arbuscular mycorrhizal fungi in plant production systems: detection, taxonomy, conservation and ecophysiology.” https://www2. dijon.inrae.fr/mychintec/Protocole/protoframe.html [Accessed 1 December 2020] (2001)  International Culture Collection of (Vesicular) Arbuscular Mycorrhizal Fungi [INVAM]. DNA extraction from single spores. INVAM https://invam.wvu.edu/ methods/dna-analysis/dna-extraction [Accessed 25 February 2020] (2017)  INVAM. 2017. Archaeospora J.B. Morton & Redecker. INVAM http://fungi.invam.wvu.edu/the-fungi/classification/archaeosporaceae/archaeos pora.1.html [Accessed 25 February 2020] (2017)  Ivanov, S. et al. Rhizobium–legume symbiosis shares an exocytotic pathway required for arbuscule formation. Proc. Natl. Acad. Sci. USA 109, 8316–8321; 10.1073/pnas.1200407109 (2012)  Izumi, H., Anderson, I. C., Alexander, I. J., Killham, K. & Moore, E. R. B. Endobacteria in some ectomycorrhiza of Scots (Pinus sylvestris). FEMS Microbiol. Ecol. 56, 34–43; 10.1111/j.1574-6941.2005.00048.x (2006)  Izumi, H. et al. Bacteria associated with ectomycorrhizas of slash pine (Pinus elliottii) in south-eastern Queensland, Australia. FEMS Microbiol. Lett. 282, 196–204; 10.1111/j.1574-6968.2008.01122.x (2008)  Izumi, H. & Finlay, R. D. Ectomycorrhizal roots select distinctive bacterial and ascomycete communities in Swedish subarctic forests. Environ. Microbiol. 13, 819–830; 10.1111/j.1462-2920.2010.02393.x (2011)

References 172  Jain, N. K. & Roy, I. Effect of trehalose on protein structure. Protein Sci. 18, 24–36; 10.1002/pro.3 (2009)  Johansson, J. F., Paul, L. R. & Finlay, R. D. Microbial interactions in the mycorrhizosphere and their significance for . FEMS Microbiol. Ecol. 48, 1–13; 10.1016/j.femsec.2003.11.012 (2004)  Jost, L. Entropy and diversity. Oikos 113, 363-375; 10.1111/j.2006.0030- 1299.14714.x (2006)  Jung, S. C., Martinez-Medina, A., Lopez-Raez, J. A. & Pozo, M. J. Mycorrhiza-induced resistance and priming of plant defenses. J. Chem. Ecol. 38, 651–664; 10.1007/s10886-012-0134-6 (2012)  Kaewklom, S., Chueakhalm, W., Suthirawut, S. & Aunpad, R. Development of a novel PCR primer to differentiate and identify Bacillus subtilis and closely related species isolated from Thai fermented foods. Food Biotechnol. 28, 354- 368; 10.1080/08905436.2014.964252 (2014)  Karandashov, V., Nagy, R., Wegmüller, S., Amrhein, N. & Bucher, M. Evolutionary conservation of a phosphate transporter in the arbuscular mycorrhizal symbiosis. PNAS 101, 6285–6290; 10.1073/pnas.0306074101 (2004)  Kassambara, A. & Mundt, F. factoextra: extract and visualize the results of multivariate data analyses. R package version 1.0.5. https://CRAN.R- project.org/package=factoextra (2017)  Kataoka, R., Taniguchi, T. & Futai, K. Fungal selectivity of two mycorrhiza helper bacteria on five mycorrhizal fungi associated with Pinus thunbergii. World J. Microbiol. Biotechnol. 25, 1815–1819; 10.1007/s11274-009-0082-7 (2009)  Kataoka, R., Taniguchi, T., Ooshima, H. & Futai, K. Comparison of the bacterial communities established on the mycorrhizae formed on Pinus thunbergii root tips by eight species of fungi. Plant Soil 304, 267–275; 10.1111/j.1574-6968.2008.01122.x (2008)  Kessell, A. E., Ladmore, G. E. & Quinn, J. C. An outbreak of primary photosensitisation in lambs secondary to consumption of Biserrula pelecinus (biserrula). Aust. Vet. J. 93, 174–178; 10.1111/avj.12318 (2015)  Khan, A. et al. Bacillus pumilus enhances tolerance in rice ( L.) to combined stresses of NaCl and high boron due to limited uptake of Na+. Environ. Exper. Bot. 124, 120–129; 10.1016/j.envexpbot.2015.12.011 (2016)  Kidd, D. et al. Rhizosphere carboxylates and morphological root traits in pasture legumes and grasses. Plant Soil 402, 77–89; 10.1007/s11104-015- 2770-4 (2016)  Kim, K. et al. Synergistic effects of inoculating arbuscular mycorrhizal fungi and Methylobacterium oryzae strains on growth and nutrient uptake of red pepper (Capsicum annuum L.). Plant Soil 327, 429–440; 10.1007/s11104-009- 0072-4 (2010)

References 173  Koebnik, R. & Lindeberg, M. Comparative genomics and evolution of bacterial type III effectors in Effectors in plant-microbe interactions (eds. Martin, F. & Kamoun, S.) 55–76 (John Wiley & Sons, 2012)  Kõljalg, U. et al. UNITE: a database providing web-based methods for the molecular identification of ectomycorrhizal fungi.New Phytol. 166, 1063– 1068; 10.1111/j.1469-8137.2005.01376.x (2005)  Kristin, A. & Miranda, H. The root microbiota—a fingerprint in the soil? Plant Soil 370, 671–686; 10.1007/s11104-013-1647-7 (2013)  Kumar, G. P. et al. Evaluation of fluorescent Pseudomonas spp. with single and multiple PGPR traits for plant growth promotion of sorghum in combination with AM fungi. Plant Growth Regul. 67, 133–140; 10.1007/s10725-012-9670-x (2012)  Kurth, F. et al. Detection and quantification of a mycorrhization helper bacterium and a mycorrhizal fungus in plant-soil microcosms at different levels of complexity. BMC Microbiol. 13, 205; 10.1186/1471-2180-13-205 (2013)  Kuzyakov, Y. & Razavi, B. Rhizosphere size and shape: temporal dynamics and spatial stationarity. Soil Biol. Biochem. 135, 343–360; 10.1016/j.soilbio.2019.05.011 (2019).  Labbé, J. L., Weston, D. J., Dunkirk, N., Pelletier, D. A. & Tuskan, G. A. Newly identified helper bacteria stimulate ectomycorrhizal formation in Populus. Front. Plant Sci. 5, Article 579; 10.3389/fpls.2014.00579 (2014)  Lage, O. M. & Bondoso, J. Bringing Planctomycetes into pure culture. Front. Microbiol. 3, 405; 10.3389/fmicb.2012.00405 (2012)  Laguerre, G. et al. Typing of rhizobia by PCR DNA fingerprinting and PCR- restriction fragment length polymorphism analysis of chromosomal and symbiotic gene regions: application to Rhizobium leguminosarum and its different biovars. Appl. Environ. Microbiol. 62, 2029-36; 10.1128/AEM.62.6. 2029-2036.1996 (1996)  Le Fevre, R., Evangelisti, E., Rey, T. & Schornack, S. Modulation of host cell biology by plant pathogenic microbes. Annu. Rev. Cell Dev. Biol. 31, 201–29; 10.1146/annurev-cellbio-102314-112502 (2015)  Lee, E. H., Eo, J. K., Ka, K. H. & Eom, A. H. Diversity of arbuscular mycorrhizal fungi and their roles in ecosystems. Mycobiology 41, 121–125; 10.5941/MYCO.2013.41.3.121 (2013)  Lehr, N. A., Schrey, S. D., Bauer, R., Hampp, R. & Tarkka, M. T. Suppression of plant defence response by a mycorrhiza helper bacterium. New Phytol. 174, 892–903; 10.1111/j.1469-8137.2007.02021.x (2007)  Leinster, T. & Cobbold, C. A. Measuring diversity: the importance of species similarity. Ecology 93, 477-489; 10.1890/10-2402.1 (2012)  Lemos, M., Gomes, I., Mergulhão, F., Melo, L. & Simões, M. The effects of surface type on the removal of Bacillus cereus and Pseudomonas fluorescens single and dual species biofilms. Food Bioprod. Process. 93, 234–241; 10.1016/j.fbp.2014.08.009 (2015)

References 174  Lennox, E. S. Transduction of linked genetic characters of the host by bacteriophage P1. Virology 1, 190-206; 10.1016/0042-6822(55)90016-7 (1955)  Lenoir, I., Fontaine, J., & Sahraoui, A. L.-H. Arbuscular mycorrhizal fungal responses to abiotic stresses: a review. Phytochemistry 123, 4–15; 10.1016/j.phytochem.2016.01.002 (2016)  Lenth, R. emmeans: estimated marginal means, aka least-squares means. R package version 1.3.4. https://CRAN.R-project.org/package=emmeans (2019)  Leveau, J. H. J. The magic and menace of metagenomics: prospects for the study of plant growth-promoting rhizobacteria. Eur. J. Plant Pathol. 119, 279– 300; 10.1007/s10658-007-9186-9 (2007)  Leyval, C., Turnau, K. & Haselwandter, K. Effect of heavy metal pollution on mycorrhizal colonization and function: physiological, ecological and applied aspects. Mycorrhiza 7, 139–153; 10.1007/s005720050174 (1997)  Li, H.-B. et al. Genetic diversity of nitrogen-fixing and plant growth promoting Pseudomonas species isolated from sugarcane rhizosphere. Front. Microbiol. 8, 1268; 10.3389/fmicb.2017.01268 (2017)  Li, J. et al. SecReT6: a web-based resource for type VI secretion systems found in bacteria. Environ. Microbiol. 17, 2196–2202; 10.1111/1462-2920.12794 (2015)  Li, X., Rui, J., Mao, Y., Yannarell, A. & Mackie, R. Dynamics of the bacterial community structure in the rhizosphere of a maize cultivar. Soil Biol. Biochem. 68, 392-401; 10.1016/j.soilbio.2013.10.017 (2014)  Liang, Y.-L., Zheng Z., Wu, M., Wu, Y. & Feng, J.-X. Isolation, screening, and identification of cellulolytic bacteria from natural reserves in the subtropical region of china and optimization of cellulase production by Paenibacillus terrae ME27-1. BioMed. Res. Int. 2014, 512497; 10.1155/2014/512497 (2014)  Lim, J. H. & Kim, S. D. Induction of drought stress resistance by multi- functional PGPR Bacillus licheniformis K11 in pepper. Plant Pathol. J. 29, 201–208; 10.5423/PPJ.SI.02.2013.0021 (2013)  Lin, H. et al. Genome sequence of Pseudomonas koreensis crs05-r5, an antagonistic bacterium isolated from rice paddy field. Front. Microbiol. 7, 1756; 10.3389/fmicb.2016.01756 (2016)  Linderman, R. G. Mycorrhizal interactions with the rhizosphere microflora: the mycorrhizosphere effect. Phytopathol. 78, 366–371 (1988)  Lioussanne, L., Perreault, F., Jolicoeur, M. & St-Arnaud, M. The bacterial community of tomato rhizosphere is modified by inoculation with arbuscular mycorrhizal fungi but unaffected by soil enrichment with mycorrhizal root exudates or inoculation with Phytophthora nicotianae. Soil Biol. Biochem. 42, 473–483; 10.1016/j.soilbio.2009.11.034 (2010)  Liu, F. et al. Soil indigenous microbiome and plant genotypes cooperatively modify soybean rhizosphere microbiome assembly. BMC Microbiol. 19, 201; 10.1186/s12866-019-1572-x (2019a)

References 175  Liu, J. et al. Arbuscular mycorrhizal symbiosis is accompanied by local and systemic alterations in gene expression and an increase in disease resistance in the shoots. Plant J. 50, 529–544; 10.1111/j.1365-313X.2007.03069.x (2007)  Liu, P., Zhang, W., Zhang, L. Q., Liu, X., Wei, H. L. Supramolecular structure and functional analysis of the Type III secretion system in Pseudomonas fluorescens 2P24. Front. Plant. Sci. 6, Article 1190; 10.3389/fpls.2015.01190 (2015)  Liu, X. et al. Using community analysis to explore bacterial indicators for disease suppression of tobacco bacterial wilt. Sci. Rep. 6, 36773; 10.1038/srep36773 (2016)  Liu, X., Feng, Z., Zhu, H & Yao, Q. Exogenous abscisic acid and root volatiles increase sporulation of Rhizophagus irregularis DAOM 197198 in asymbiotic and pre-symbiotic status. Mycorrhiza 29, 581–589; 10.1007/s00572-019- 00916-z (2019b)  Liu, Y. et al. Proposal of nine novel species of the Bacillus cereus group. Int. J. Syst. Evol. Microbiol. 67, 2499–2508; 10.1099/ijsem.0.001821 (2017)  Lloyd, K. G., Steen, A. D., Ladau, J., Yin, J. & Crosby, L. Phylogenetically novel uncultured microbial cells dominate earth microbiomes. mSystems 3, e00055-18; 10.1128/mSystems.00055-18 (2018)  Loi, A. Biserrula. Pastures Australia https://keys.lucidcentral.org/keys/v3/ pastures/Html/Biserrula.htm [Accessed 24 February 2020] (2007)  Loi, A. et al. Important ecological traits for selecting Biserrula pelecinus L. (biserrula) genotypes for their potential introduction into agricultural systems. Grass Forage Sci. 70, 519–529; 10.1111/gfs.12139 (2014)  Loi, A., Howieson, J. H. & Carr, S. J. Register of Australian herbage plant cultivars. Biserrula pelecinus L. (biserrula) cv. Casbah. Aust. J. Exp. Agric. 41, 841–842; 10.1071/EA01075_CU (2001)  Loi, A., Howieson, J. G., Nutt, B. J. & Carr, S. J. A second generation of annual pasture legumes and their potential for inclusion in Mediterranean-type farming systems Aust. J. Exp. Agric. 45, 289–299; 10.1071/EA03134 (2005)  Loi, A., Nutt, B. J., Revell, C. K., Sandral, G. A. & Dear, B. S. Mauro: a mid to late maturing cultivar of biserrula (Biserrula pelecinus). Aust. J. Exp. Agric. 46, 595–597; 10.1071/EA05033 (2006)  Long, L., Zhu, H., Yao, Q. & Ai, Y. Analysis of bacterial communities associated with spores of Gigaspora margarita and Gigaspora rosea. Plant Soil 310, 1–9; 10.1007/s11104-008-9611-7 (2008)  Long, S.R. Rhizobium-legume nodulation: life together in the underground. Cell 56, 203–214; 10.1016/0092-8674(89)90893-3 (1989)  Ludwig-Müller, J. & Güther, M. as signals in arbuscular mycorrhiza formation. Plant Signal. Behav. 2, 194–196; 10.1111/j.1399- 3054.2006.00812.x (2007)

References 176  Lu, X., Lu, X. & Liao, Y. Effect of tillage treatment on the diversity of soil arbuscular mycorrhizal fungal and soil aggregate-associated carbon content. Front. Microbiol. 9, 2986; 10.3389/fmicb.2018.02986 (2018)  Lugo, M.A. et al. Arbuscular mycorrhizal fungi and rhizospheric bacteria diversity along an altitudinal gradient in South American puna grassland. Microb. Ecol. 55, 705–713; 10.1007/s00248-007-9313-3 (2008)  Lugtenberg, B. & Kamilova, F. Plant-growth-promoting rhizobacteria. Annu. Rev. Microbiol. 63, 541–556; 10.1146/annurev.micro.62.081307.162918 (2009)  Luo, Y. et al. An Arabidopsis homolog of importin β1 is required for ABA response and drought tolerance. Plant J. 75, 377–389; 10.1111/tpj.12207 (2013)  Mahdi, O., Eklund, B. & Fisher, N. Stenotrophomonas maltophilia: laboratory culture and maintenance. Curr. Protoc. Microbiol. 32, Unit–6F.1; 10.1002/9780471729259.mc06f01s32 (2014)  Maillet, F. et al. Fungal lipochitooligosaccharide symbiotic signals in arbuscular mycorrhiza. Nature 469, 58–63; 10.1038/nature09622 (2011)  Majed, R., Faille, C., Kallassy, M. & Gohar, M. Bacillus cereus biofilms— same, only different. Front. Microbiol. 7, 1054; 10.3389/fmicb.2016.01054 (2016)  Mamatha, G., Bagyaraj, D. J. & Jaganath, S. Inoculation of field-established mulberry and papaya with arbuscular mycorrhizal fungi and a mycorrhiza helper bacterium. Mycorrhiza 12, 313–316; 10.1007/s00572-002-0200-y (2002)  Mamoun, M. & Olivier, J. M.. Dynamique des populations fongiques et bactériennes de la rhizosphere de noisetiers truffiers. II. due fer et répartition taxonomique chez les Pseudomonas fluorescens. Agronomie 9, 345–351 (1989)  Marschner, P. & Baumann, K. Changes in bacterial community structure induced by mycorrhizal colonisation in split-root maize. Plant Soil 251, 279– 289; 10.1023/A:1023034825871 (2003)  Marschner, P., Crowley, D. E. & Lieberei, R. Arbuscular mycorrhizal infection changes the bacterial 16 S rDNA community composition in the rhizosphere of maize. Mycorrhiza 11, 297–302; 10.1007/s00572-001-0136-7 (2001)  Mazurier, S., Lemunier, M., Hartmann, A., Siblot, S. & Lemanceau, P. Conservation of type III secretion system genes in Bradyrhizobium isolated from soybean, FEMS Microbiol. Lett. 259, 317–325; 10.1111/j.1574- 6968.2006.00290.x (2006)  Mazurier, S. et al. Type III secretion system and virulence markers highlight similarities and differences between human- and plant-associated pseudomonads related to Pseudomonas fluorescens and P. putida. Appl. Environ. Microbiol. 81, 2579–2590; 10.1128/AEM.04160-14 (2015)

References 177  McCaig, A. E., Glover, L. A. & Prosser, J. I. Molecular analysis of bacterial community structure and diversity in unimproved and improved upland grass pastures. Appl. Environ. Microbiol. 65, 1721-1730 (1999)  McGee, P. A. Survival and growth of seedlings of coachwood (Ceratopetalum apetalum): Effects of shade, mycorrhizas and a companion plant. Aust. J. Bot. 38, 583–592; 10.1071/BT9900583 (1990)  McMurdie, P. J. & Holmes, S. Waste not, want not: why rarefying microbiome data is inadmissible. PLOS Comput. Biol. 10, e1003531; 10.1371/journal.pcbi.1003531 (2014)  Medeiros, F. H. V. et al. Transcriptional profiling in cotton associated with Bacillus subtilis (UFLA285) induced biotic-stress tolerance. Plant Soil 347, 327–337; 10.1007/s11104-011-0852-5 (2011)  Mediavilla, O., Olaizola, J., Santos-del-Blanco, L., Oria-de-Rueda, J. A. & Martín-Pinto, P. Mycorrhization between L. and Bull is enhanced by the mycorrhiza helper bacteria Pseudomonas fluorescens Migula. Mycorrhiza 26, 161–168; 10.1007/s00572-015-0657-0 (2016)  Medina, M. J. H. et al. Root colonization by arbuscular mycorrhizal fungi is affected by the salicylic acid content of the plant. Plant Sci. 164, 993–998; 10.1016/S0168-9452(03)00083-9 (2003)  Mendes, L. W., Kuramae, E. E., Navarrete, A. A., van Veen, J. A. & Tsai, S. M. Taxonomical and functional microbial community selection in soybean rhizosphere. ISME J. 8, 1577–1587; 10.1038/ismej.2014.17 (2014)  de Mendiburu, D. Agricolae: statistical procedures for agricultural research. R package version 1.3-1. https://CRAN.R-project.org/package=agricolae (2019)  Mendoza, R., Escudero, V. & García, I. Plant growth, nutrient acquisition and mycorrhizal symbioses of a waterlogging tolerant legume (Lotus glaber Mill.) in a saline-sodic soil. Plant Soil 275, 305–315; 10.1007/s11104-005-2501-3 (2005)  Meyer, J. R. & Linderman, R. G. Selective influence on populations of rhizosphere or rhizoplane bacteria and actinomycetes by mycorrhizas formed by Glomus fasciculatum. Soil Biol. Biochem. 18, 191–196; 10.1016/0038- 0717(86)90026-X (1986)  Meyer, J. R. & Linderman, R. G. Response of subterranean clover to dual- inoculation with vesicular-arbuscular mycorrhizal fungi and a plant growth- promoting bacterium, Pseudomonas putida. Soil Biol. Biochem. 18, 185–190; 10.1016/0038-0717(86)90025-8 (1986)  Miller, R. M., Jastrow, J. D. & Reinhardt, D. R. External hyphal production of vesicular-arbuscular mycorrhizal fungi in pasture and tallgrass prairie communities. Oecologia 103, 17–23; 10.1007/BF00328420 (1995)  Mitter, B., Pfaffenbichler, N. & Sessitsch, A. Plant-microbe partnerships in 2020. Microb. Biotechnol. 9, 635–640; 10.1111/1751-7915.12382 (2016)  Mocali, S. & Benedetti, A. Exploring research frontiers in microbiology: the challenge of metagenomics in soil microbiology. Res. Microbiol. 161, 497– 505; 10.1016/j.resmic.2010.04.010 (2010)

References 178  Mogge, B., Loferer, C., Agerer, R., Hutzler, P. & Hartmann, A. Bacterial community structure and colonization patterns of Fagus sylvatica L. ectomycorrhizospheres as determined by fluorescence in situ hybridization and confocal laser scanning microscopy. Mycorrhiza 9, 271–278; 10.1007/PL00009991 (2000)  Morales, S. E. & Holben, W. E. Linking bacterial identities and processes: can 'omic' analyses be more than the sum of their parts? FEMS Microbiol. Ecol. 75, 2–16; 10.1111/j.1574-6941.2010.00938.x (2011)  Moreno-Espíndola, I. P. et al. The bacterial community structure and microbial activity in a traditional organic milpa farming system under different soil moisture conditions. Front. Microbiol. 9, 2737; 10.3389/fmicb.2018.02737 (2018)  Mosse, B. The regular germination of resting spores and some observations on the growth requirements of an Endogone sp. causing vesicular-arbuscular mycorrhiza. Trans. Brit. Mycol. Soc. 42, 273–286; 10.1016/S0007- 1536(56)80033-8 (1959)  Mosse, B. The establishment of vesicular-arbuscular mycorrhiza under aseptic conditions. J. Gen. Microbiol. 27, 509–520; 10.1099/00221287-27-3-509 (1962)  Moulin, L., Munive, A., Dreyfus, B. & Boivin-Masson C. Nodulation of legumes by members of the beta-subclass of Proteobacteria. Nature 411, 948– 950; 10.1038/35082070 (2001)  Mulet, M., Bennasar, A., Lalucat, J. & García‐Valdés, E. An rpoD‐based PCR procedure for the identification of Pseudomonas species and for their detection in environmental samples. Mol. Cell. Probes 23, 140–147; 10.1016/j.mcp.2009.02.001 (2009)  Murray, G. M. & Brennan, J. P. The current and potential costs from diseases of wheat in Australia. Grains Research and Development Corporation (GRDC) https://grdc.com.au/Resources/Publications/2009/10/The-Current-and- Potential-Costs-from-Diseases-of-Wheat-in-Australia [Accessed 16 November 2016] (2009)  Murray, G. M. & Brennan, J. P. The current and potential costs from diseases of barley in Australia. Grains Research and Development Corporation (GRDC) https://grdc.com.au/Resources/Publications/2009/10/The-Current-and- Potential-Costs-from-Diseases-of-Barley-in-Australia [Accessed 16 November 2016] (2009)  Murray, G. M. & Brennan, J. P. The current and potential costs from diseases of oilseed crops in Australia. Grains Research and Development Corporation (GRDC) https://grdc.com.au/Resources/Publications/2012/06/The-Current-and- Potential-Costs-from-Diseases-of-Oilseed-Crops-in-Australia [Accessed 16 November 2016] (2012)  Murray, J. D., Cousins, D. R., Jackson, K.J. & Liu, C. Signaling at the root surface: the role of cutin monomers in mycorrhization. Mol. Plant 6, 1381– 1383; 10.1093/mp/sst090 (2013)

References 179  Murthy, N. & Bleakley, B. Simplified method of preparing colloidal chitin used for screening of chitinase- producing microorganisms. Internet J. Microbiol. 10, 14186 (2012)  Muyzer, G., de Waal, E. C. & Uitierlinde, A. G. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16s rRNA. Appl. Environ. Microbiol. 59, 695–700 (1993)  Nagahashi, G. & Douds, D. D. The effects of hydroxy fatty acids on the hyphal branching of germinated spores of AM fungi. Fungal Biol. 115, 351–358; 10.1016/j.funbio.2011.01.006 (2011)  Nandasena, K. G., O’Hara, G. W., Tiwari, R. P. & Howieson, J. G. Rapid in situ evolution of nodulating strains for Biserrula pelecinus L. through lateral transfer of a symbiosis island from the original mesorhizobial inoculant. Appl. Environ. Microbiol. 72, 7365–7367; 10.1128/AEM.00889-06 (2006)  Nandasena, K. G., O’Hara, G. W., Tiwari, R. P., Sezmiş, E. & Howieson, J. G. In situ lateral transfer of symbiosis islands results in rapid evolution of diverse competitive strains of mesorhizobia suboptimal in symbiotic nitrogen fixation on the pasture legume Biserrula pelecinus L. Environ. Microbiol. 9, 2496– 2511; 10.1111/j.1462-2920.2007.01368.x (2007)  Nehls, U., Göhringer, F., Wittulsky, S. & Dietz, S. Fungal carbohydrate support in the ectomycorrhizal symbiosis: a review. Plant Biol. 12, 292–301; 10.1111/j.1438-8677.2009.00312.x (2010)  Nelson, M. S. & Sadowsky, M. J. Secretion systems and signal exchange between nitrogen-fixing rhizobia and legumes. Front. Plant Sci. 6, Article 491; 10.3389/fpls.2015.00491 (2015)  Nguyen, N. H. & Bruns, T. D. The microbiome of Pinus muricata ectomycorrhizae: community assemblages, fungal species effects, and Burkholderia as important bacteria in multipartnered symbioses. Microb. Ecol. 69, 914–921; 10.1007/s00248-015-0574-y (2015)  Nichols, D. et al. Use of ichip for high-throughput in situ cultivation of “uncultivable” microbial species. Appl. Environ. Microbiol. 76, 2445-2450; 10.1128/AEM.01754-09 (2010)  Nichols, P. G. H. et al. New annual and short-lived perennial pasture legumes for Australian agriculture—15 years of revolution. Field Crop Res. 104, 10–23; 10.1016/j.fcr.2007.03.016 (2007)  Nichols, P. G. H. et al. Temperate pasture legumes in Australia—their history, current use, and future prospects. Crop Pasture Sci. 63, 691–725; 10.1071/CP12194 (2012)  Noguchi, H. et al. Bacillus vietnamensis sp. nov., a moderately halotolerant, aerobic, endospore-forming bacterium isolated from Vietnamese fish sauce. Int. J. Syst. Evol. Microbiol. 54, 2117–2120; 10.1099/ijs.0.02895-0 (2004)  de Novais, C. B., Giovannetti, M., de Faria, S. M. & Sbrana, C. Two herbicides, two fungicides and spore-associated bacteria affect Funneliformis

References 180 mosseae extraradical mycelium structural traits and viability. Mycorrhiza 29, 341–349. 10.1007/s00572-019-00901-6 (2019)  NSW Environment, Energy and Science. Dryland salinity: identifying saline sites. https://www.environment.nsw.gov.au/research-and-publications/publicati ons-search/book-2-dryland-salinity-identifying-saline-sites [Accessed 23 March 2020] (2008)  Obase, K. Extending the hyphal area of the ectomycorrhizal fungus Laccaria parva co-cultured with ectomycorrhizosphere bacteria on nutrient agar plate. Mycoscience 60, 95–101; 10.1016/j.myc.2018.12.002 (2019)  Ocón, A., Hampp, R. & Requena, N. Trehalose turnover during abiotic stress in arbuscular mycorrhizal fungi. New. Phytol. 174, 879–891; 10.1111/j.1469- 8137.2007.02048.x (2007)  Ofek, M., Hadar, Y. & Minz, D. Ecology of root colonizing Massilia (Oxalobacteraceae). PLOS ONE 7, e40117; 10.1371/journal.pone.0040117 (2012)  Offre, P. et al. Identification of bacterial groups preferentially associated with mycorrhizal roots of Medicago truncatula. Appl. Environ. Microbiol. 73, 913– 921; 10.1128/AEM.02042-06 (2007)  Offre, P. et al. Microdiversity of Burkholderiales associated with mycorrhizal and nonmycorrhizal roots of Medicago truncatula. FEMS Microbiol. Ecol. 65, 180–192; 10.1111/j.1574-6941.2008.00504.x (2008)  Oh, S. & Lim, Y. W. Effect of bacteria on the growth of Tricholoma in vitro culture. Mycorrhiza 28, 411–419; 10.1007/s00572-018- 0828-x (2018)  Oksanen, J. et al. vegan: Community Ecology Package. R package version 2.5.5. https://CRAN.R-project.org/package=vegan (2019)  Oláh, B., Brière, C., Bécard, G., Dénarié, J. & Gough, C. Nod factors and a diffusible factor from arbuscular mycorrhizal fungi stimulate lateral root formation in Medicago truncatula via the DMI1/DMI2 signalling pathway. Plant J. 44, 195–207; 10.1111/j.1365-313X.2005.02522.x (2005)  Olivier, J. M. & Mamoun, M. Dynamique des populations fongiques et bactériennes de la rhizosphere de noisetiers truffiers. I. Relations avec le statut hydrique du sol. Agronomie 8, 711–717. (1988)  Omirou, M., Fasoula, D. A. & Ioannides, I. M. Bradyrhizobium inoculation alters indigenous AMF community assemblages and interacts positively with AMF inoculum to improve cowpea performance. Appl. Soil Ecol. 108, 381– 389; 10.1016/j.apsoil.2016.09.018 (2016)  Orgiazzi, A. 2015. Soil biodiversity and DNA barcodes: opportunities and challenges. Soil Biol. Biochem. 80, 244–250; 10.1016/j.soilbio.2014.10.014 (2015)

References 181  Pacioni, G. 16 wet-sieving and decanting techniques for the extraction of spores of vesicular-arbuscular fungi. Method. Microbiol. 24, 317–322; 10.1016/S0580-9517(08)70099-0 (1992)  Pantanella, F. et al. Violacein and biofilm production in Janthinobacterium lividum. J. Appl. Microbiol. 102, 992-999; 10.1111/j.1365-2672.2006.03155.x (2007)  Parnell, J. J. et al. From the lab to the farm: an industrial perspective of plant beneficial microorganisms. Front Plant. Sci. 7, 1110; 10.3389/fpls.2016.01110 (2016)  Parniske, M. Arbuscular mycorrhiza: the mother of plant root endosymbiosis. Nature Rev. Microbiol. 6, 763–775; 10.1038/nrmicro1987 (2008)  Pascual, J., García-López, M., Bills, G. F. & Genilloud, O. Pseudomonas granadensis sp. nov., a new bacterial species isolated from the Tejeda, Almijara and Alhama Natural Park, Granada, Spain. Int. J. Syst. Evol. Microbiol. 65, 625–632 10.1099/ijs.0.069260-0 (2015)  Paula, M. A., Urquiaga, S. & Sijqueira, J. O. Synergistic effects of vesicular- arbuscular mycorrhizal fungi and diazotrophicus bacteria on nutrition and growth of sweet potato (Ipomoea batatas). Biol. Fertil. Soils 14, 61–66; 10.1007/BF00336304 (1992)  Pavlov, M. Y. & Ehrenberg, M. Optimal control of gene expression for fast proteome adaptation to environmental change. PNAS 110, 20527–20532; 10.1073/pnas.1309356110 (2013)  Peix, A., Ramírez-Bahena, M. H., Velázquez, E. & Bedmar, E. J. Bacterial associations with legumes. CRC Crit. Rev. Plant Sci. 34, 17–42; 10.1080/07352689.2014.897899 (2015)  Peralta, A. L., Sun, Y., McDaniel, M. D. & Lennon, J. T. Crop rotational diversity increases disease suppressive capacity of soil microbiomes. Ecosphere 9, e02235; 10.1002/ecs2.2235 (2018)  Philippot, L., Raaijmakers, J. M., Lemanceau, P. & van der Putten, W. H. Going back to the roots: the microbial ecology of the rhizosphere. Nature Rev. Microbiol. 11, 789–799; 10.1038/nrmicro3109 (2013)  Pivato, B. et al. Bacterial effects on arbuscular mycorrhizal fungi and mycorrhiza development as influenced by the bacteria, fungi, and host plant. Mycorrhiza 19, 81–90; 10.1007/s00572-008-0205-2 (2009)  Planchamp, C., Glauser, G. & Mauch-Mani, B. Root inoculation with Pseudomonas putida KT2440 induces transcriptional and metabolic changes and systemic resistance in maize plants. Front. Plant Sci. 5, Article 719; 10.3389/fpls.2014.00719 (2015)  Plett, J. M. et al. Effector MiSSP7 of the mutualistic fungus Laccaria bicolor stabilizes the Populus JAZ6 protein and represses jasmonic acid (JA) responsive genes. Proc. Natl. Acad. Sci. USA 111, 8299-8304; 10.1073/pnas.1322671111 (2014)  Poole, E. J., Bending, G. D., Whipps, J. M. & Read, D. J. Bacteria associated with Pinus sylvestris–Lactarius rufus ectomycorrhizas and their effects on

References 182 mycorrhiza formation in vitro. New Phytol. 151, 743–751; 10.1046/j.0028- 646x.2001.00219.x (2001)  Poupin, M. J., Timmermann, T., Vega, A., Zuñiga, A. & González, B. Effects of the plant growth-promoting bacterium Burkholderia phytofirmans PsJN throughout the life cycle of Arabidopsis thaliana. PLoS ONE 8, e69435; 10.1371/journal.pone.0069435 (2013)  Pozo, M. J. & Azcón-Aguilar, C. Unraveling mycorrhiza-induced resistance. Curr. Opin. Plant Biol. 10, 393–398; 10.1016/j.pbi.2007.05.004 (2007)  Probanza, A. et al. Effects of inoculation with PGPR Bacillus and Pisolithus tinctorius on Pinus pinea L. growth, bacterial rhizosphere colonization, and mycorrhizal infection. Microb. Ecol. 41, 140–148; 10.1007/s002480000081 (2001)  Prosser, J. I. Replicate or lie. Environ. Microbiol. 12, 1806-1810; 10.1111/j.1462-2920.2010.02201.x (2010)  R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing https://www.R-project.org/ (2019)  Raaijmakers, J. M., Paulitz, T. C., Steinberg, C., Alabouvette, C. & Moënne- Loccoz, Y. The rhizosphere: a playground and battlefield for soilborne pathogens and beneficial microorganisms. Plant Soil 321, 341–361; 10.1007/s11104-008-9568-6 (2009)  Rambelli, A. The rhizosphere of mycorrhizae in Ectomycorrhizae: their ecology and physiology (ed. Marks, G. C.) 299–350 (Elsevier, 1973)  Rao, N. S. S., Tilak, K. V. B. R. & Singh, C. S. Effect of combined inoculation of Azospirillum brasilense and vesicular-arbuscular mycorrhiza on pearl millet (Pennisetum americanum). Plant Soil 84, 283–286; 10.1007/BF02143192 (1985)  Rappé M. S. & Giovannoni, S. J. The uncultured microbial majority. Annu. Rev. Microbiol. 57, 369-394; 10.1146/annurev.micro.57.030502.090759 (2003)  Ravnskov, S., Nybroe, O. & Jakobsen, I. Influence of an arbuscular mycorrhizal fungus on Pseudomonas fluorescens DF57 in rhizosphere and hyphosphere soil. New Phytol. 142, 113–122 (1999)  Requena, N., Jimenez, I., Toro, M. & Barea, J. M. Interactions between plant- growth-promoting rhizobacteria (PGPR), arbuscular mycorrhizal fungi and Rhizobium spp. in the rhizosphere of Anthyllis cytisoides, a model legume for in mediterranean semi-arid ecosystems. New Phytol. 136, 667– 677; 10.1046/j.1469-8137.1997.00786.x (1997)  Requena, N., Serrano, E, Ocón, A. & Breuninger, M. Plant signals and fungal perception during arbuscular mycorrhiza establishment. Phytochemistry 68, 33–40; 10.1016/j.phytochem.2006.09.036 (2007)  Rillig, M. C., Mummey, D. L., Ramsey, P. W., Klironomos, J. N. & Gannon, J. E. Phylogeny of arbuscular mycorrhizal fungi predicts community composition of symbiosis-associated bacteria. FEMS Microbiol. Ecol. 57, 389–395; 10.1111/j.1574-6941.2006.00129.x (2006)

References 183  Rodrigues, K. M. & Rodrigues, B. F. Endomycorrhizal association of Funneliformis mosseae with transformed roots of Linum usitatissimum: germination, colonization, and sporulation studies. 6, 42–49; 10.1080/21501203.2015.1024777 (2015)  Rosas, S. B. et al. Root colonization and growth promotion of wheat and maize by Pseudomonas aurantiaca SR1. Soil Biol. Biochem. 41, 1802–1806; 10.1016/j.soilbio.2008.10.009 (2009)  Rousk, J., Brookes, P. C. & Bååth, E. Contrasting soil pH effects on fungal and bacterial growth suggests functional redundancy in carbon mineralisation. Appl. Environ. Microbiol. 75, 1589–1596; 10.1128/AEM.02775-08 (2009)  Rovira, A. D. Dryland mediterranean farming systems in Australia. Aust. J. Exp. Agric. 32, 801–809; 10.1071/EA9920801 (1992)  Rózycki, H., Kampert, M., Strzelczyk, E., Li, C. Y. & Perry, D. A. Effect of different soil bacteria on mycorrhizae formation in pine (Pinus sylvestris L.). Folia For. Pol. Ser. A For. 36, 91–102. (1994)  Rudawska, M. & Gay, G. Effect of an IAA overproducer mutant of the fungus Hebeloma cylindrosporum Romagnesi on the early stages of ectomycorrhizal infection and carbohydrate content in seedlings of Pinus pinaster (Ait.) Sol. Acta Soc. Bot. Pol. 64, 255–264; 10.5586/asbp.1995.034 (1995)  Rufo, R., Batlle, A., Camprubi, A., Montesinos, E. & Calvet, C. Control of rubus stunt and stolbur diseases in Madagascar periwinkle with mycorrhizae and a synthetic antibacterial peptide. Plant Pathol. 66, 551–558; 10.1111/ppa.12612 (2016)  Russo, D. M. et al. Proteins exported via the PrsD-PrsE Type I secretion system and the acidic exopolysaccharide are involved in biofilm formation by Rhizobium leguminosarum. J. Bacteriol. 188, 4474–4486; 10.1128/JB.00246- 06 (2006)  Ryan, R. P. et al. The versatility and adaptation of bacteria from the genus Stenotrophomonas. Nat. Rev. Microbiol. 7, 514-525. 10.1038/nrmicro2163 (2009)  Ryu, C. M. Against friend and foe: Type 6 effectors in plant-associated bacteria. J. Microbiol. 53, 201–208; 10.1007/s12275-015-5055-y (2015)  Sanchez, L. et al. Common gene expression in Medicago truncatula roots in response to Pseudomonas fluorescens colonization, mycorrhiza development and nodulation. New Phytol. 161, 855–863; 10.1111/j.1469- 8137.2003.00997.x (2004)  Sang, M. K. & Kim, K. D. Biocontrol activity and root colonization by Pseudomonas corrugata strains CCR04 and CCR80 against Phytophthora blight of pepper. BioControl 59, 437–448; 10.1007/s10526-014-9584-9 (2014)  Santos, J. C., Finlay, R. D. & Tehler, A. Molecular analysis of arbuscular mycorrhizal fungi colonising a semi-natural grassland along a fertilisation gradient. New Phytol. 172, 159–168; 10.1111/j.1469-8137.2006.01799.x (2006)

References 184  Sathyapriya, H., Sariah, M., Siti Nor Akmar, A. & Wong, M. Root colonisation of Pseudomonas aeruginosa strain UPMP3 and induction of defence-related genes in oil palm (Elaeis guineensis). Ann. Appl. Biol. 160, 137–144; 10.1111/j.1744-7348.2011.00525.x (2012)  Schefe, C. R. et al. 100 Years of superphosphate addition to pasture in an acid soil—current nutrient status and future management. Soil Res. 53, 662–676; 10.1071/SR14241 (2015)  Scheublin, T. R., Sanders, I. R., Keel, C. & van der Meer, J. R. Characterisation of microbial communities colonising the hyphal surfaces of arbuscular mycorrhizal fungi. ISME J. 4, 752–763; 10.1038/ismej.2010.5 (2010)  Schindlbacher, A. et al. Experimental warming effects on the microbial community of a temperate mountain forest soil. Soil. Biol. Biochem. 43, 1417- 1425; 10.1016/j.soilbio.2011.03.005 (2011)  Schippers, B., Bakker, A. W. & Bakker, P. A. H. M. Interactions of deleterious and beneficial rhizosphere microorganisms and the effect of cropping practices. Ann. Rev. Phytopathol. 25, 339–358; 10.1146/annurev.py.25.090187.002011 (1987)  Schmidt, R., Cordovez, V., de Boer, W. et al. Volatile affairs in microbial interactions. ISME J. 9, 2329–2335; 10.1038/ismej.2015.42 (2015)  Schmidt, R. et al. Fungal volatile compounds induce production of the secondary metabolite Sodorifen in Serratia plymuthica PRI-2C. Sci. Rep. 7, 862; 10.1038/s41598-017-00893-3 (2017)  Schmidt, R., Gravuer, K., Bossange, A. V., Mitchell, J. & Scow, K. Long-term use of cover crops and no-till shift soil microbial community life strategies in agricultural soil. PLoS ONE 13, e0192953; 10.1371/journal.pone.0192953 (2018)  Schnepf, A., Roose, T. & Schweiger, P. Growth model for arbuscular mycorrhizal fungi. J. R. Soc. Interface 5, 773–784; 10.1098/rsif.2007.1250 (2008)  Schrey, S. D., Schellhammer, M., Ecke, M., Hampp, R. & Tarkka, M. T. Mycorrhiza helper bacterium Streptomyces AcH 505 induces differential gene expression in the ectomycorrhizal fungus . New Phytol. 168, 205–216; 10.1111/j.1469-8137.2005.01518.x (2005)  Schulze, S. et al. Analysis of new type III effectors from Xanthomonas uncovers XopB and XopS as suppressors of plant immunity. New Phytol. 195, 894–911; 10.1111/j.1469-8137.2012.04210.x (2012)  Schüβler, A., Schwarzott, D. & Walker, C. A new fungal phylum, the Glomeromycota: phylogeny and evolution. Mycol. Res. 105, 1413–1421; 10.1017/S0953756201005196 (2001)  Schüßler, A. & Walker, C. The Glomeromycota. A species list with new families and new genera. AMF Phylogeny http://www.amf-phylogeny.com [Accessed 24 February 2020] (2010)

References 185  Schüßler, A. & Walker, C. Archaeospora ecuadoriana sp. nov. from a mountainous biodiversity hotspot area in Ecuador, and transfer of Palaeospora spainiae to Archaeospora, as A. spainiae comb. nov.. Mycorrhiza 29, 435–443; 10.1007/s00572-019-00913-2 (2019)  Shelef, O., Hahn, P. G., Getman-Pickering, Z. & Martinez Medina, A. Coming to common ground: the challenges of applying ecological theory developed aboveground to rhizosphere interactions. Front. Ecol. Evol. 7, 58; 10.3389/fevo.2019.00058 (2019)  Shi, P., Abbott, L. K., Banning, N. C. & Zhao, B. Comparison of morphological and molecular genetic quantification of relative abundance of arbuscular mycorrhizal fungi within roots. Mycorrhiza 22, 501–513; 10.1007/s00572-011-0425-8 (2012)  Shinde, S. et al. Pseudomonas fluorescens increases mycorrhization and modulates expression of antifungal defense response genes in roots of aspen seedlings. BMC Plant Biol. 19, 4; 10.1186/s12870-018-1610-0 (2019)  Shokralla, S., Spall, J. L., Gibson, J. F. & Hajibabaei, M. Next‐generation sequencing technologies for environmental DNA research. Mol. Ecol. 21, 1794-1805; 10.1111/j.1365-294X.2012.05538.x (2012)  Singh, B. K. et al. Relationship between assemblages of mycorrhizal fungi and bacteria on grass roots. Environ. Microbiol. 10, 534–541; 10.1111/j.1462- 2920.2007.01474.x (2008)  Singh, R., Soni, S. K. & Kalra, A. Synergy between Glomus fasciculatum and a beneficial Pseudomonas in reducing root diseases and improving yield and forskolin content in Coleus forskohlii Briq. under organic field conditions. Mycorrhiza 23, 35–44; 10.1007/s00572-012-0447-x (2013)  Smith, C. R. et al. Microbial community responses to soil tillage and crop rotation in a corn/soybean agroecosystem. Ecol. Evol. 6, 8075–8084; 10.1002/ece3.2553 (2016)  Smith, F. A. & Smith, S. E. What is the significance of the arbuscular mycorrhizal colonisation of many economically important crop plants? Plant Soil 348, 63–79; 10.1007/s11104-011-0865-0 (2011)  Smith, S. E. & Read, D. J. Mycorrhizal symbiosis, 3rd edition (Academic Press, 2008)  Smith, S. E. & Smith, F. A. Roles of arbuscular mycorrhizas in plant nutrition and growth: new paradigms from cellular to ecosystem scales. Annu. Rev. Plant Biol. 62, 227–250; 10.1146/annurev-arplant-042110-103846 (2011)  Sohn, K. H., Hughes, R. K., Piquereza, S. J., Jones, J. D. G. & Banfield, M. J. Distinct regions of the Pseudomonas syringae coiled-coil effector AvrRps4 are required for activation of immunity. PNAS 109, 16371–16376; 10.1073/pnas.1212332109 (2012)  Soto, M. J., Sanjuán, J. & Olivares, J. Rhizobia and plant-pathogenic bacteria: common infection weapons. Microbiology 152, 3167–3174; 10.1099/mic.0.29112-0 (2006)

References 186  Srinivas, A., Sasikala, C. & Ramana, C. V. Rhodoplanes oryzae sp. nov., a phototrophic alphaproteobacterium isolated from the rhizosphere soil of paddy. Int. J. Syst. Evol. Microbiol. 64, 2198 – 2203; 10.1099/ijs.0.063347-0 (2014)  Sundram, S., Meon, S., Seman, I. A. & Othman, R. Symbiotic interaction of endophytic bacteria with arbuscular mycorrhizal fungi and its antagonistic effect on Ganoderma boninense. J. Microbiol. 489, 551–557; 10.1007/s12275- 011-0489-3 (2011)  Sy, A. et al. Methylotrophic Methylobacterium bacteria nodulate and fix nitrogen in symbiosis with legumes. J. Bacteriol. 183, 214–220; 10.1128/JB.183.1.214-220.2001  Sylvia, D. M., Hammond, L. C., Bennett, J. M., Haas, J. H. & Linda, S. B. Field response of maize to a VAM fungus and water management. Agron. J. 85, 193–198; 10.2134/ agronj1993.00021962008500020006x (1993)  Szczesny, R. et al. (2010) Suppression of the AvrBs1-specific hypersensitive response by the YopJ effector homolog AvrBsT from Xanthomonas depends on a SNF1-related kinase. New Phytol. 187, 1058–1074; 10.1111/j.1469- 8137.2010.03346.x  Taktek, S. et al. Trapping of phosphate solubilizing bacteria on hyphae of the arbuscular mycorrhizal fungus Rhizophagus irregularis DAOM 197198. Soil Biol. Biochem. 90, 1–9; 10.1016/j.soilbio.2015.07.016 (2015)  Tampakaki, A. P. Commonalities and differences of T3SSs in rhizobia and plant pathogenic bacteria. Front. Plant Sci. 5, Article 114; 10.3389/fpls.2014.00114 (2014)  Thomas, T., Gilbert, J. & Meyer, F. Metagenomics - a guide from sampling to data analysis. Microb. Inform. Exp. 2, 3; 10.1186/2042-5783-2-3 (2012)  Tibbett, M. et al. The diversity of arbuscular mycorrhizas of selected Australian Fabaceae. Plant Biosyst. 142, 420–427; 10.1080/11263500802151124 (2008)  Timmusk, S., Grantcharova, N. & Wagner, E. G. H. Paenibacillus polymyxa invades plant roots and forms biofilms. Appl. Eviron. Microbiol. 71, 7292– 7300; 10.1128/AEM.71.11.7292–7300.2005 (2005)  Toljander, J. F., Artursson, V., Paul, L. R., Jansson, J. K. & Finlay, R. D. Attachment of different soil bacteria to arbuscular mycorrhizal fungal extraradical hyphae is determined by hyphal vitality and fungal species. FEMS Microbiol. Lett. 254, 34–40; 10.1111/j.1574-6968.2005.00003.x (2006)  Tomasiewicz, D. M., Hotchkiss, D. K., Reinbold, G. W., Read, R. B. & Hartman, P. A. The most suitable number of colonies on plates for counting. J. Food Prot. 43, 282–286; 10.4315/0362-028X-43.4.282 (1980)  Toro, M., Azcón, R. & Barea, J. M. Improvement of arbuscular mycorrhiza development by inoculation of soil with phosphate-solubilizing rhizobacteria to improve rock phosphate bioavailability (P32) and nutrient cycling. Appl. Environ. Microbiol. 63, 4408–4412 (1997)  Torrecillas, E., Alguacil, M. M., Roldán, A. Host preferences of arbuscular mycorrhizal fungi colonizing annual herbaceous plant species in semiarid

References 187 Mediterranean prairies. Appl. Environ. Microbiol. 78, 6180–6186; 10.1128/AEM.01287-12 (2012)  Treseder, K. K. The extent of mycorrhizal colonization of roots and its influence on plant growth and phosphorus content. Plant Soil 371, 1–13; 10.1007/s11104-013-1681-5 (2013)  Trivedi, P., Schenk P. M., Wallenstein M. D. & Singh B. K. Tiny microbes, big yields: enhancing food crop production with biological solutions. Microb. Biotechnol. 10, 999-1003; 10.1111/1751-7915.12804 (2017)  Trouvelot, A., Kough, J. L. & Gianinazzi-Pearson, V. Mesure du taux de mycorhization VA d’un système radiculaire: Recherche de méthodes d’estimation ayant une signification fonctionnelle in Physiological and genetical aspects of mycorrhizae (eds. Gianinazzi-Pearson, V. & Gianinazzi, S.) 217–221 (INRA Press, 1986)  Turrini, A. et al. Changes in the composition of native root arbuscular mycorrhizal fungal communities during a short-term cover crop-maize succession. Biol. Fertil. Soils 52, 643–653; 10.1007/s00374-016-1106-8 (2016)  Turrini, A., Avio, L., Giovannetti, M. & Agnolucci, M. Functional complementarity of arbuscular mycorrhizal fungi and associated microbiota: the challenge of translational research. Front. Plant Sci., 9, 1407; 10.3389/fpls.2018.01407 (2018).  Unno, Y. & Shinano, T. Metagenomic analysis of the rhizosphere soil microbiome with respect to phytic acid utilization. Microbes Environ. 28, 120– 127; 10.1264/jsme2.me12181 (2013)  Uroz, S. et al. Effect of the mycorrhizosphere on the genotypic and metabolic diversity of the bacterial communities Involved in mineral in a forest soil. Appl. Environ. Microbiol. 73, 3019–3027; 10.1128/AEM.00121-07 (2007)  Uroz, S., Oger, P., Morin, E. & Frey-Klett, P. Distinct ectomycorrhizospheres share similar bacterial communities as revealed by pyrosequencing-based analysis of 16S rRNA genes. Appl. Environ. Microbiol. 78, 3020–3024; 10.1128/AEM.06742-11 (2012)  Vafadar, F., Amooaghaie, R. & Otroshy, M. Effects of plant-growth-promoting rhizobacteria and arbuscular mycorrhizal fungus on plant growth, stevioside, NPK, and chlorophyll content of Stevia rebaudiana, J. Plant Interact. 9, 128– 136; 10.1080/17429145.2013.779035 (2014)  Vályi, K., Mardhiah, U., Rillig, M. C. & Hempel, S. Community assembly and coexistence in communities of arbuscular mycorrhizal fungi. ISME J. 10, 2341–2351; 10.1038/ismej.2016.46 (2016)  Veith, P.D., Glew, M.D., Gorasia, D.G. & Reynolds, E.C. Type IX secretion: the generation of bacterial cell surface coatings involved in virulence, gliding motility and the degradation of complex biopolymers. Mol. Microbiol. 106, 35–53; 10.1111/mmi.13752 (2017)  Verbruggen, E., van der Heijden, M. G. A., Rillig, M. C. & Kiers, E. T. Mycorrhizal fungal establishment in agricultural soils: factors determining

References 188 inoculation success. New Phytol. 197, 1104–1109; 10.1111/j.1469- 8137.2012.04348.x (2013)  Vicente, C., Pérez-Fernández, M. A., Pereira, G. & Tavares-de-Sousa, M. M. Biodiversity of root-nodule bacteria associated with the leguminous plant Biserrula pelecinus. Soil Sci. 174, 424–429; 10.1097/SS.0b013e3181b655a2 (2009)  Vicente, C. S. L. & Pérez-Fernández, M. A. Broad environmental tolerance of native root-nodule bacteria of Biserrula pelecinus indicate potential for soil fertility restoration. Plant Ecol. Divers. 9, 299–307; 10.1080/17550874.2016.1182226 (2016)  Vik, U. et al. Different bacterial communities in ectomycorrhizae and surrounding soil. Sci. Rep. 3, 3471; 10.1038/srep03471 (2013)  Viollet, A. et al. Fluorescent pseudomonads harboring type III secretion genes are enriched in the mycorrhizosphere of Medicago truncatula. FEMS Microbiol. Ecol. 75, 457–467; 10.1111/j.1574-6941.2010.01021.x (2011)  Viollet, A. et al. Pseudomonas fluorescens C7R12 type III secretion system impacts mycorrhization of Medicago truncatula and associated microbial communities. Mycorrhiza 27, 23–33; 10.1007/s00572-016-0730-3 (2017).  Vivas, A., Barea, J. M. & Azcón, R. Brevibacillus brevis isolated from cadmium- or zinc-contaminated soils improves in vitro spore germination and growth of Glomus mosseae under high Cd or Zn concentrations. Microb. Ecol. 49, 416–424; 10.1007/s00248-004-0044-4 (2005)  Vivas, A., Marulanda, A., Ruiz-Lozano, J. M., Barea, J. M. & Azcón, R. Influence of a Bacillus sp. on physiological activities of two arbuscular mycorrhizal fungi and on plant responses to PEG-induced drought stress. Mycorrhiza 13, 249–256; 10.1007/s00572-003-0223-z (2003)  Viaene, T., Langendries, S., Beirinckx, S., Maes, M. & Goormachtig, S. Streptomyces as a plant's best friend? FEMS Microbiol. Ecol. 92, fiw119. 10.1093/femsec/fiw119 (2016)  Vogel, C., Bodenhausen, N., Gruissem, W. & Vorholt, J. A. The Arabidopsis transcriptome reveals distinct but also overlapping responses to colonization by phyllosphere commensals and pathogen infection with impact on plant health. New Phytol. 212, 192–207; 10.1111/nph.14036 (2016)  Vosátka, M. & Gryndler, M. Treatment with culture fractions from Pseudomonas putida modifies the development of Glomus fistulosum mycorrhiza and the response of potato and maize plants to inoculation. Appl. Soil. Ecol. 11, 245–251; 10.1016/S0929-1393(98)00151-6 (1999)  van Waasbergen, L. G., Balkwill, D. L., Crocker, F. H., Bjornstad, B. N., & Miller, R. V.. Genetic diversity among Arthrobacter species collected across a heterogeneous series of terrestrial deep-subsurface sediments as determined on the basis of 16S rRNA and recA gene sequences. Appl. Environ. Microbiol. 66, 3454–3463; 10.1128/aem.66.8.3454-3463.2000 (2000)

References 189  Walker, T. S. et al. Pseudomonas aeruginosa-plant root interactions. Pathogenicity, biofilm formation, and root exudation. Plant Physiol. 134, 320– 331; 10.1104/pp.103.027888 (2004)  Wallden, K., Rivera-Calzada, A. & Waksman, G. Type IV secretion systems: versatility and diversity in function. Cell Microbiol. 12, 1203–1212; 10.1111/j.1462-5822.2010.01499.x (2010)  Wang, M.-Y., Xia, R.-X., Wu, Q.-S., Liu, J.-H. & Hu, L.-M. Influence of arbuscular mycorrhizal fungi on microbes and enzymes of soils from different cultivated densities of red clover. Ann. Microbiol., 57, 1–7; 10.1007/BF03175042 (2007)  Wang, Q., Liu, J. & Zhu H. Genetic and molecular mechanisms underlying symbiotic specificity in legume-rhizobium interactions. Front. Plant Sci. 9, 313; 10.3389/fpls.2018.00313 (2018)  Warmink, J. A. & van Elsas, J. D. Selection of bacterial populations in the mycosphere of Laccaria proxima: is type III secretion involved? ISME J. 2, 887–900; 10.1038/ismej.2008.41 (2008)  Weber, D. Limitations to field application of Rhizobium inoculants in Genetic engineering for nitrogen fixation (eds. Hollaender, A. et al.) 433 (Springer, 1977)  Welc, M., Ravnskov, S., Kieliszewska-Rokicka, B. & Larsen, J. Suppression of other soil microorganisms by mycelium of arbuscular mycorrhizal fungi in root-free soil. Soil Biol. Biochem. 42, 1534–1540; 10.1016/j.soilbio.2010.05.024 (2010)  Whipps, J. M. Microbial interactions and biocontrol in the rhizosphere. J. Exp. Bot. 52, 487–511; 10.1093/jexbot/52.suppl_1.487 (2001)  Will, M. E. & Sylvia, D. M. Interaction of rhizosphere bacteria, , and vesicular-arbuscular mycorrhizal fungi with sea oats. Appl. Environ. Microbiol. 56, 2073–2079 (1990)  Willis, A. D. Rarefaction, alpha diversity, and statistics. Front. Microbiol. 10, 2407; 10.3389/fmicb.2019.02407 (2019)  Wilson, B. A. L. The effect of dryland salinity on arbuscular mycorrhizal fungi in southern New South Wales. Thesis, Charles Sturt University, Wagga Wagga, Australia (2010)  Winsor, G. L. et al.. Enhanced annotations and features for comparing thousands of Pseudomonas genomes in the Pseudomonas genome database. Nucleic Acids Res. 44, D646–D653; 10.1093/nar/gkv1227 (2016)  Wirsel, S. G. Homogenous stands of a wetland grass harbour diverse consortia of arbuscular mycorrhizal fungi. FEMS Microbiol. Ecol. 48, 129–138; 10.1016/j.femsec.2004.01.006 (2004)  World Flora Online. An online flora of all known plants. World Flora Online http://www.worldfloraonline.org/ [Accessed 13 October 2019] (2019)

References 190  Wu, J. et al. Analysis of bacterial communities in rhizosphere soil of continuously cropped healthy and diseased konjac. World J. Microbiol. Biotechnol. 33, 134; 10.1007/s11274-017-2287-5 (2017)  Wu, Q., Xia, R. & Hu, Z. Effect of arbuscular mycorrhiza on the drought tolerance of Poncirus trifoliata seedlings. Front. Forest. China 1, 100; 10.1007/s11461-005-0007-z (2006)  Wu, X. Q. et al. Effects of ectomycorrhizal fungus Boletus edulis and mycorrhiza helper Bacillus cereus on the growth and nutrient uptake by Pinus thunbergii. Biol. Fertil. Soils 48, 385–391; 10.1007/s00374-011-0638-1 (2012)  Wu, X.-Y., Walker, M., Hornitzky, M. & Chin, J. Development of a group- specific PCR combined with ARDRA for the identification of Bacillus species of environmental significance. J. Microbiol. Methods 64, 107–119; 10.1016/j.mimet.2005.04.021 (2006)  Xavier, J. B. Social interaction in synthetic and natural microbial communities. Mol. Syst. Biol. 7, 483. 10.1038/msb.2011.16 (2011)  Xavier, L. J. C. & Germida, J. J. Bacteria associated with Glomus clarum spores influence mycorrhizal activity. Soil Biol. Biochem. 35, 471–478; 10.1016/S0038-0717(03)00003-8 (2003)  Xia, X., Ma, C., Dong, S., Xu, Y. & Gong, Z. Effects of nitrogen concentrations on nodulation and nitrogenase activity in dual root systems of soybean plants. J. Soil Sci. Plant Nutr. 63, 470-482; 10.1080/00380768.2017.1370960 (2017)  Xiang, D., Chen, B. & Li, H. Specificity and selectivity of arbuscular mycorrhizal fungal polymerase chain reaction primers in soil samples by clone library analyses. Acta Agr. Scand. B.-S. P. 66, 333–339; 10.1080/09064710.2015.1117649 (2016)  Xie, Z. P. et al. Rhizobial nodulation factors stimulate mycorrhizal colonization of nodulating and nonnodulating soybeans. Plant Physiol. 108, 1519–1525 (1995)  Yang, H., Koide, R. T. & Zhang, Q. Short-term waterlogging increases arbuscular mycorrhizal fungal species richness and shifts community composition. Plant Soil 404, 373–384; 10.1007/s11104-016-2850-0 (2016)  Yang, H., Zang, Y., Yuan, Y., Tang, J. & Chen, X. Selectivity by host plants affects the distribution of arbuscular mycorrhizal fungi: Evidence from ITS rDNA sequence metadata. BMC Evol. Biol. 12, 50; 10.1186/1471-2148-12-50 (2012)  Yang, P. et al. The type three secretion system facilitates migration of Burkholderia terrae BS001 in the mycosphere of two soil-borne fungi. Biol. Fertil. Soils 52, 1037–1046; 10.1007/s00374-016-1140-6 (2016)  Yang, S. K. et al. Soil microbial community analysis of between no-till and tillage in a controlled horticultural field. World J. Microbiol. Biotechnol. 28, 1797; 10.1007/s11274-011-0933-x (2012)

References 191  Yin, C. et al. Bacterial communities on wheat grown under long-term conventional tillage and no-till in the Pacific Northwest of the United States. Phytobiomes J. 1, 83-90; 10.1094/PBIOMES-09-16-0008-R (2017)  Yoon, S. H. et al. Introducing EzBioCloud: A taxonomically united database of 16S rRNA and whole genome assemblies. Int. J. Syst. Evol. Microbiol. 67, 1613–1617; 10.1099/ijsem.0.001755 (2017)Youssef, N. H., Couger, M. B., McCully, A. L., Guerrero Criado, A. E. & Elshahed, M. S. Assessing the global phylum level diversity within the bacterial domain: a review. J. Adv. Res. 6, 269-282; 10.1016/j.jare.2014.10.005 (2015)  Yurgel, S. N. et al. Variation in bacterial and eukaryotic communities associated with natural and managed wild blueberry habitats. Phytobiomes J. 1, 102-113; 10.1094/PBIOMES-03-17-0012-R (2017)  Zaller, J., Heigl, F., Ruess, L. & Grabmaier, A. Glyphosate herbicide affects belowground interactions between earthworms and symbiotic mycorrhizal fungi in a model ecosystem. Sci. Rep. 4, 5634; 10.1038/srep05634 (2015)  Zhang, B. et al. Short-term effect of tillage and crop rotation on microbial community structure and enzyme activities of a clay loam soil. Biol. Fertil. Soils 50, 1077; 10.1007/s00374-014-0929-4 (2014)  Zhang, H. et al. Communities of arbuscular mycorrhizal fungi and bacteria in the rhizosphere of Caragana korshinkii and Hippophae rhamnoides in Zhifanggou watershed. Plant Soil 326, 415–424; 10.1007/s11104-009-0022-1 (2010)  Zhang, J., Kobert, K., Flouri, T. & Stamatakis, A. PEAR: a fast and accurate Illumina Paired-End reAd mergeR. Bioinformatics 30, 614–620; 10.1093/bioinformatics/btt593 (2014)  Zhao, L., Wu, X. Q., Ye, J. R., Li, H. & Li, G. E. Isolation and characterization of a mycorrhiza helper bacterium from rhizosphere soils of poplar stands. Biol. Fertil. Soils 50, 593–601; 10.1007/s00374-013-0880-9 (2014)  Zheng, W. et al. An accurate and efficient experimental approach for characterization of the complex oral microbiota. Microbiome 3, 48; 10.1186/s40168-015-0110-9 (2015)  Zhou, D. M., Wang, K. P., Liu, H. X., Gu, C. & Guo, J. H. Field evaluation of different application methods of the mixture of Bacillus cereus strain AR156 and Bacillus subtilis strain SM21 on pepper growth and disease resistance. Biocontrol Sci. Technol. 24, 1451–1468; 10.1080/09583157.2014.945899 (2014)  Zipfel, C. & Robatzek, S. Pathogen-associated molecular pattern-triggered immunity: veni, vidi...? Plant Physiol. 154, 551–554; 10.1104/pp.110.161547. (2010)

References 192 Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas

The review of the literature presented in Chapter 1 is concerned principally with the role of MHB strains in AM associations. Where additional evidence is available from ECM systems, this is presented here. Headings in this appendix refer to subsections in Chapter 1. The evidence discussed here should be considered supplementary to the evidence for AM associations described in the specified subsection. Evidence for a helper effect in ECM systems is summarised in Table A1-1.

1.3.1. The mycorrhizosphere effect  Paragraph 2: An attempt to quantify ECM-associated change in microbial communities was conducted using fluorescent in situ hybridisation and confocal laser scanning microscopy (FISH-CLSM; Mogge et al. 2000). This study reported dense colonisation of the ECM mantle by alpha-, beta-, and in Fagus sylvatica colonised chiefly by Lactarius spp. and , but was unable to provide finer taxonomic resolution.  Paragraph 3: Culture-dependent screening of ECM-associated bacteria has also been conducted. Frey-Klett et al. (2005) screened two FLPs obtained from the mycosphere (mycorrhizosphere + mycorrhizal fungus) and bulk soil of a Pseudotsuga menziesii (Douglas fir) – Laccaria bicolor S238N ECM association for a range of traits, reporting that phosphate-solubilising and siderophore- producing FLPs were enriched in the mycosphere environment relative to the bulk soil. Likewise, a later study (Uroz et al. 2007) screened bacterial strains isolated from the bulk soil, mycorrhizosphere and symbiotic mantle of a Quercus petraea (oak) – ECM association for their mineral-weathering potential, reporting that strains of Burkholderia and

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 193 Collimonas were most efficient and that the mycorrhizosphere and fungal mantle showed the highest mineral-weathering potential.  Paragraph 4: Genetic characterisation of mycorrhiza-associated communities has been conducted for ECM systems. Some early genetic approaches to characterising bacterial communities utilised DNA extracted from cultured bacterial communities (Izumi et al. 2008, Izumi et al. 2006). More recent studies have extracted DNA directly from mycorrhizal root tips, bypassing the limitations of the culture-dependent approach. To date, these studies have focused on a narrow range of ECM communities: Pseudotsuga menziesii (Douglas fir) in association with members of the (Burke et al. 2008); Pinus thunbergii (Japanese black pine) in association with eight ECM fungi (Kataoka et al. 2008); Betula pubescens (downy birch) in association with five ECM fungi (Izumi and Finlay, 2011); Quercus petraea (oak) in association with Scleroderma citrinum and Xerocomellus (Xerocomus) pruinatus (Uroz et al. 2012); Persicaria (Bistorta) vivipara (alpine bistort) in association with unspecified ECM fungi (Vik et al. 2013); and Pinus muricata (Bishop pine) in association with five ECM fungi (Nguyen and Bruns 2015). With the exception of Kataoka et al. (2008), which did not compare ECM and non-ECM roots, and Burke et al. (2008), these studies reported clear community differences between ECM and non-ECM roots. Nguyen and Bruns (2015) attribute the Burke et al. (2008) result to the older sequencing methods employed by the earlier study (t- RFLP and Sanger sequencing) and to the use of non-surface sterilised roots. These studies also report differences as to which groups of bacteria predominate in the ectomycorrhizosphere. For example, Uroz et al. (2012) report high concentrations of Acidobacteria and Actinobacteria in the ECM association, while Vik et al. (2013) report high concentrations of Acidobacteria in ECM but lower concentrations of Actinobacteria than in the surrounding soil, and Burke et al. (2008) report no association of Acidobacteria or Actinobacteria with ECM tissues. Kataoka et al. (2008) found significantly greater diversity in ECM root tips occupied by the early-stage coloniser Cenococcum geophilum than late stage colonisers Russula spp. and Suillus sp., suggesting a species-dependent effect or an effect of succession. The same study reported greater bacterial density in extraradical mycelium than in ECM root tips.

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 194 Table A1-1. Evidence of a helper effect on mycorrhization, growth of fungal mycelium, or plant growth parameters in ectomycorrhizas

Bacterial strains Host species Fungal partner(s) Reference AR BA BE BU CL FR PE PS RA RO SE SR ? Pseudotsuga menziesii Laccara laccata + ~ Duponnois and Garbaye 1991 = = = = Pinus sylvestris Rózycki et al. 1994 Laccaria laccata + – + – Pinus radiata Rhizopogon luteolus ~ Garbaye and Bowen 1989 Pseudotsuga menziesii Laccaria bicolor + Frey-Klett et al. 1997 Eucalyptus diversicolor Laccaria spp. ± ± ± Dunstan et al. 1998 Pseudotsuga menziesii Laccaria bicolor = Frey-Klett et al. 1999 Pseudotsuga menziesii Laccaria bicolor ± Brulé et al. 2001 Pinus sylvestris Lactarius rufus =a ~ +a = Poole et al. 2001 Pinus sylvestris Suillus luteus ~ – = – Bending et al. 2002 Acacia holosericea Pisolithus alba + Founoune et al. 2002a A. holosericea Pisolithus alba ~ Founoune et al. 2002b Pisolithus spp. + Acacia spp. Duponnois and Plenchette 2003 Scleroderma spp. + Picea abies Amanita muscaria ~ Schrey et al. 2005 Pinus sylvestris + Laccaria bicolor = = + + Pinus sylvestris Lactarius rufus = – = = Aspray et al. 2006a Contaminating spp. ~ = = = Pinus sylvestris Lactarius rufus + + Aspray et al. 2006b None (in vitro) Pisolithus alba F+ Duponnois and Kisa 2006 None (in vitro) Laccaria bicolor F= F+ F– F~ F+ Deveau et al. 2007 Wilcoxina mikolae F= F+ None (in vitro) Kataoka et al. 2009 Rhizopogon sp. F= F=

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 195 Table A1-1. Evidence of a helper effect on mycorrhization, growth of fungal mycelium, or plant growth parameters in ectomycorrhizas

Bacterial strains Host species Fungal partner(s) Reference AR BA BE BU CL FR PE PS RA RO SE SR ? Cenococcum geophilum F– F= Pisolithus tinctorius F= F– F+ F+ Pinus thunbergii Boletus edulis P+ Wu et al. 2012 None (in vitro) Lactarius rufus * Aspray et al. 2013 Populus spp. Laccaria bicolor ~ Labbé et al. 2014 Pisolithus tinctorius + = Populus deltoides Zhao et al. 2014 Lactarius insulsus + – Cistus ladanifer Boletus edulis + Mediavilla et al. 2016 None (in vitro) Tricholoma matsutake – – – – + – – Oh and Lim 2018b None (in vitro) Laccaria parva # – ± – Obase 2019b Populus tremuloides Laccaria bicolor + Shinde et al. 2019 Experimental evidence of a promotional effect exerted by bacterial strains on the formation of ectomycorrhizas. Effects on mycelial growth are reported where mycorrhization was not studied and are indicated by the prefix (F). Effects on plant growth are reported where neither mycelial growth nor mycorrhization were studied and are indicated by the prefix (P). (+) Positive effect on mycorrhization. (=) Neutral or non-significant effect on mycorrhization. (–) Negative effect on mycorrhization. (~) Mixed effect of multiple bacterial strains on mycorrhization (strain-specific effect). (±) Mixed effect of single strain on mycorrhization in multiple plant-fungus systems (mycorrhiza-specific effect) or in varying environments (condition-specific effect). (#) Both strain-specific and host-specific effects. (*) Reduced colony size, increased branching density. Bacteria are Arthrobacter (AR), Bacillus (BA), Brevibacillus (BE), Burkholderia (BU), Collimonas (CL), Frankia (FR), Paenibacillus (PE), Pseudomonas (PS), Ralstonia (RA), Rhodococcus (RO), Serratia (SE), Streptomyces (SR), and unknown (?). aStudy did not clearly distinguish between Bacillus and Paenibacillus. bOh and Lim (2018) and Obase (2019) screened representatives of more than 20 genera, finding that most suppressed fungal growth in vitro. Taxa not previously investigated were excluded from the table for reasons of space.

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 196 1.4.1. Characteristics of mycorrhiza helper bacteria  Paragraph 5: Additional evidence concerning the specificity of helper activity is drawn from ECM and ectendomycorrhizal systems. Bacillus subtilis, for example, has been reported to enhance the in vitro pre-symbiotic growth and in vivo mycorrhization of the ecetendomycorrhizal fungus Wilcoxina mikolae while suppressing the pre-symbiotic growth and mycorrhization of another ECM fungus, a species of Rhizopogon (Kataoka et al. 2009). Similarly, P. fluorescens BBc6R8 has been found to enhance mycorrhization between Eucalyptus diversicolor and Laccaria fraterna while inhibiting mycorrhization between E. diversicolor and L. laccata (Dunstan et al. 1998) and P. monteilii HR13 has been found to enhance the pre-symbiotic growth of Pisolithus strains but not Scleroderma strains (Duponnois and Plenchette 2003). Further, Streptomyces sp. AcH505 was reported to increase in vitro mycelial growth in Amanita muscaria while retarding it in the pathogenic fungi obscura and Heterobasidion annosum (Schrey et al. 2005). By contrast, a later study (Lehr et al. 2007) found that Streptomyces sp. AcH505 suppressed 11 of the 12 screened pathogenic Heterobasidion strains but not H. abietinum 331. This strain suffered no ill effects from antibiotic WS-5995 B, produced by the MHB strain, and received a benefit to infectivity. Strains of Ochrobactrum and Bacillus were found to serve as helper bacteria in Casuarina cunninghamiana – Frankia actinorrhizas (Echbab et al. 2004) but have not been assessed for their ability to enhance AM formation in that system.  Paragraph 6: The behaviour and localisation of helper strains following inoculation has received more attention in ECM systems than in AM systems. For example, Frey-Klett et al. (1997) studied a Douglas fir – L. bicolor ECM system and found that the MHB, P. fluorescens BBc6R8 was randomly distributed along the root and was not strongly associated with L. bicolor sporocarps or mycorrhizal tissues. By contrast, Zhao et al. (2014) studied an ECM association between poplar (Populus deltoides) and the fungi Pisolithus tinctorius and Lactarius insulsus and reported extensive colonisation of short roots and surface hyphae by their study strain, Bacillus sp. DZ18. Bacteria were not observed inside the root. More generally, in a Pinus sylvestris – Lactarius rufus ECM, the MHB Paenibacillus sp. EJP73 was only able to exert its helper effect in direct contact with the symbiotic partners (Aspray et al. 2006b).

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 197 Populations of the MHB P. fluorescens BBc6R8 have been found to decline rapidly after inoculation into nursery soil (Frey-Klett et al. 1997) but this decline is less swift when co-inoculated with the ECM fungus Laccaria bicolor S238N (Deveau et al. 2010).  Paragraph 6: Dosage responses have received additional study in ECM systems. Inoculation experiments in glasshouse and nursery-based Douglas fir – L. bicolor systems found that doses of the MHB strain P. fluorescens BBc6R8 as low as 30 colony-forming units (CFU) g-1 soil produced a significant helper effect (Frey-Klett et al. 1997). A later experiment conducted in the same system used three dosages of the same MHB (8x105, 8x107 and 8x109 CFU m2) and generated an inverse dose-response effect, wherein the lowest dose produced the most pronounced benefits to the host plant in terms of shoot length and root dry weight (Frey-Klett et al. 1999; note that no significant increase in mycorrhization was reported by this study). Later work has complicated this picture: in a Petri dish-based microcosm, Aspray et al. (2006b) found that Paenibacillus sp. EJP73 enhanced mycorrhization between Pinus sylvestris and Lactarius rufus at all concentrations tested (105,107, 109 and 1010 CFU ml-1) whereas Burkholderia sp. EJP67 only demonstrated a helper effect at 107 and 109 CFU ml-1.  Paragraph 6: Trehalose has also been found to enhance radial growth of the ECM fungus Pisolithus alba IR100 (Duponnois and Kisa 2006).  Paragraph 7: Some additional evidence concerning non-sterile conditions: while P. fluorescens BBc6R8 has been found to maintain its helper effect in field and nursery soils (Frey-Klett et al. 1997), co-inoculation of the MHB Streptomyces sp. AcH 505 with a soil filtrate has been found to reduce the abundance of the helper strain under certain circumstances (Kurth et al. 2013).

1.4.2. The helper effect  Paragraph 2: Inconsistencies between in vitro and in vivo results have been documented in ECM assays. Bowen and Theodorou (1979) observed that the behaviour of co-inoculated bacteria and ECM fungi in cultures bore no resemblance to their behaviour in the rhizoplane. More recently, Dunstan et al. (1998) reported an in vivo enhancement in mycorrhization between Eucalyptus diversicolor and both L. laccata and L. fraterna co-inoculated with Bacillus

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 198 subtilis MB3, despite a negative effect of this strain on L. fraterna growth and a neutral effect on L. laccata growth in cultures. P. fluorescens, which enhanced the in vitro growth of both L. laccata and L. fraterna significantly enhanced mycorrhization in L. laccata but inhibited it in L. fraterna. In a later study (Kataoka et al. 2009), Ralstonia basilensis had no significant effect on the growth of the ECM fungus Cenococcum geophilum in cultures but inhibited the formation of mycorrhizas in pots. Similarly, Labbé et al. (2014) identified 19 fluorescent pseudomonads with beneficial effects on L. bicolor S238N colony diameter but only three of these demonstrated an ability to enhance mycorrhization and one was found to inhibit the formation of mycorrhizas.  Paragraph 3: Evidence of a nutritive MHB effect can be seen in ECM systems. Co-inoculation of the MHB strain P. fluorescens BBc6R8 with the ECM fungus L. bicolor S238N on a nutrient-poor medium (water agar) showed a stimulatory effect on fungal growth and survival whereas co-inoculation on a rich medium (10% tryptic soy agar) showed an inhibitory effect on fungal growth, suggesting that the MHB effect of that strain may relate in part to improvements in fungal nutrient status (Brulé et al. 2001). Siderophore production, however, is poorly correlated: Founoune et al. (2002b) screened 32 fluorescent pseudomonads obtained from soil, fungus, rhizobium, nematode and plant compartments of an Acacia holosericea – Pisolithus alba ECM for their effects on mycorrhization and shoot and root biomass, and found that 15 isolates enhanced mycorrhization and at least one growth parameter, while 6 enhanced mycorrhization without enhancing growth and 11 enhanced growth without affecting mycorrhization. The authors reported that plant growth promotion was strongly associated with particular pyoverdine siderovars, suggesting that iron accumulation was critical for growth in the experimental conditions. By contrast, the imprecise association between improvements in plant growth and mycorrhization suggests that siderophore production was not principally responsible for the observed MHB effect.  Paragraph 3: An effect on pH has been studied in ECM systems. Olivier and Mamoun (1988) and Mamoun and Olivier (1989) report alterations in pH associated with iron chelation by siderophores.  Paragraph 3: Additional evidence of a role for selective antagonism comes from the MHB strain Streptomyces sp. AcH 505, which produces the antifungal

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 199 compound WS-5995B and has been shown to suppress the activity of many pathogenic Heterobasidion strains (Lehr et al. 2007). Interestingly, one additional strain, H. abietinum 331, escaped this suppressive effect in cultures and showed an enhanced colonisation of Picea albes (Norway spruce) roots when co-inoculated with the MHB. It is possible that certain helper effects occur due to selective resistance to bacterial antifungals; the reduced competition these fungi face in the rhizosphere may facilitate infection.

1.5.1. Bacterial biomolecules and root receptivity  Paragraph 2: Evidence for the involvement of IAA in some helper effects: greater plant colonisation was reported for an IAA-overproducing mutant of the ECM fungus Hebeloma cylindrosporum relative to its wild-type strain (Rudawska and Gay 1995).  Paragraph 4: Several biomolecules of interest have been identified in ECM systems. Hypaphorine, which is exuded by Pisolithus tinctorius and functions to suppress development (Ditengou et al. 2003), has been suggested as a candidate (Founoune et al. 2002a), but this avenue has not subsequently been pursued. A role in the MHB effect for trehalose, a common carbohydrate implicated in many biological processes including stress resistance and found in many fungi (Jain and Roy 2009), was suggested by a study reporting greater in vitro growth of Pisolithus alba when co-inoculated with the helper strain P. monteilii and in cultures containing trehalose but lacking the helper strain (Duponnois and Kisa 2006). No equivalent effect was reported for cultures containing D-glucosamine, oxalic acid, malonic acid, citric acid, carboxymethyl cellulose, starch or chitin. The ECM fungi Hebeloma crustuliniforme and were previously shown to respond positively to malic acid and citric acid (Duponnois and Garbaye 1990), suggesting a variable or perhaps fungus-specific effect of these exudates. The model MHB P. fluorescens BBc6R8 shows strong towards trehalose and showed improved growth when treated with trehalose in a dose-response manner (Deveau et al. 2010). Finally, exposure of L. bicolor S238N to unspecified volatiles derived from Paenibacillus sp. EJP73 was found to reduce colony size but increase branching, which may subsequently increase root-fungus contacts (Aspray et al. 2013). The volatiles in this case have not been identified, and this contrasts with

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 200 earlier work (Aspray et al. 2006b) which found no effect of bacterial cell suspensions despite a positive effect of live cultures.

Appendix 1. Mycorrhiza helper bacteria in ectomycorrhizas 201 Appendix 2. Structure of experiments addressing ecological interactions in the rhizosphere (Chapter 4)

Figure A2-1. A diagrammatic outline of experiments conducted in Chapter 4, which concern interactions between MHB strains and other microorganisms in the biserrula rhizosphere. Pink hexagons represent starting materials, rectangles represent preparatory and experimental stages, and green rhomboids represent outputs.

Appendix 2. Structure of experiments addressing ecological interactions 202