Potential for suppression of Rhizoctonia root rot is influenced by nutrient (N and P) and carbon inputs in a highly calcareous coarse-textured topsoil

Keywords: calcareous, microbial diversity, nitrogen, phosphorus, soil-borne diseases, Rhizoctonia, stubble management, suppression.

A large proportion of cereal production in Australia occurs in semiarid regions on coarse-textured soils that have many edaphic limitations to agricultural productivity (McKenzie et al. 2004). These limitations include not only inherent low water holding capacity (WHC) but also poor fertility exacerbated by low amounts of organic matter (OM) and hence limited cycling of essential nutrients such as nitrogen (N). Additionally, in alkaline coarse-textured topsoils that comprise a high proportion of carbonate, the ‘tie-up’ of nutrients such as phosphorus (P) and zinc, osmotic pressure of soil solution and toxicity from high concentrations of bicarbonate, are all factors likely to constrain plant growth and production (Rengasamy 2010). This complex of edaphic limitations to crop production is characteristic of alkaline calcareous soils (>2 million ha) that support many low-input rain-fed farming systems in southern Australia, which are important contributors to regional and national grain production. Furthermore, such edaphic constraints not only directly affect crop growth and productivity but will affect soil microorganisms (Lauber et al. 2008; Kamble et al. 2014) and indirectly influence plant–soil microorganism interactions (Hartman and Tringe 2019). Low carbon (C) content and poor fertility in particular are likely to limit microbial activity and growth in sandy soils and may support less diverse microbial communities. Indeed, these factors may contribute to the dominance of fungal soil-borne root pathogens in continuous cereal crops on coarse-textured soils (Gupta et al. 2012), especially Rhizoctonia solani AG8 which appears to thrive at low soil moistures typical of coarse-textured soils in semiarid environments (MacNish and Neate 1996; Gill et al. 2001a, 2001b).

Rhizoctonia bare patch resulting from root disease caused by the soil-borne fungus R. solani AG8 is a major biological constraint to cereal crop production in the semiarid cropping regions of southern and Western Australia, and in Pacific North-West regions of the USA (MacNish and Neate 1996; Paulitz 2006; Murray and Brennan 2009, 2010). Roots of plants infected with R. solani show characteristic brown and rotten root tips (spear-tips) that restrict the growth of the root system and result in reduced ability of plants to access water and nutrients. Additionally, growth response to applied fertiliser has been shown to be greater for plants with healthy roots than those with root disease (Wall et al. 1994). Despite several significant advances in the control of Rhizoctonia bare patch, viz. reduced tillage and crop rotation based management practices and some chemical and biocontrol options, complete and reliable control of the disease has not been achieved. Biological disease suppression mediated by a diverse community of soil bacteria and fungi has been suggested to reduce or remove root disease naturally, even in the presence of pathogen, and suppression of cereal root diseases has been identified in agricultural fields in southern and Western Australia (Roget 1995; Penton et al. 2014; Gupta et al. 2019), as well as in soils of the Pacific North-West cropping region in the United States (Lucas et al. 1993; Yin et al. 2013; Schillinger and Paulitz 2014).

It has been widely reported that addition of OM to soil can increase suppression of soil-borne root diseases (van Overbeek et al. 2012; Bonanomi et al. 2018), and this is principally attributed to the influence of C in the OM, especially readily available or labile C. Inputs of C can improve root disease suppression by (i) a general increase in total microbial activity, and thus competitive exclusion of pathogen; (ii) increased abundances of specific microbial communities that reduce pathogens and (iii) induced disease tolerance capacity in plants (Gupta et al. 2011; Schlatter et al. 2017; Vida et al. 2020). Long-term studies of the effects of conservation agriculture practices, including crop residue (cereal stubble) retention for increasing C inputs and microbial activity, have been shown to alter microbial community composition resulting in suppression of soil-borne diseases such as Rhizoctonia bare patch and take-all (Cook 2007; Gupta et al. 2019). There are a range of soil organisms identified as being involved in the development of root-disease suppression (Weller et al. 2002). Soil organisms isolated from agricultural soils in southern Australia that have been shown to act individually or in concert to decrease root infection by Rhizoctonia include bacteria (Exiguobacterium acetylicum, Pantoea agglomerans, Microbacterium spp., Pseudomonas spp., Streptomyces sp. and Bacillus spp.), fungi (Penicillium griseofulvum and Trichoderma spp.) and mycophagous amoebae (Gupta et al. 1999; Barnett 2005; Yang et al. 2005a, 2005b; Barnett et al. 2006).

The development of disease suppression can take 7–10 years, often with a disease peak in the intervening 3–5 years (Roget 1995; Cook 2007). However, in highly alkaline soils of the upper Eyre Peninsula region of South Australia, practices including crop stubble retention, reduced or no-till, crop rotation and enhanced nutrient supply changed microbial population activity and diversity after 8 years but disease suppression of Rhizoctonia and take-all did not develop (Cook et al. 2012). Furthermore, although a controlled environment study of three soils from the same region indicated that addition of OM increased the potential for soil-borne root disease suppression by reducing root infection and increasing DNA of potentially suppressive organisms (E. acetylicum, P. agglomerans and Microbacterium and Trichoderma spp.), the effect was least in the soil that was the most calcareous (Davey et al. 2019). This limited disease suppression potential was postulated to be the result of edaphic limitations, in particular low P availability, that reduced both plant and soil organism growth and function despite C inputs. Although varied effects of N nutrition in agriculture on crop root disease expression have been described (Srihuttagum and Sivasithamparam 1991; Lucas et al. 1993; Smiley et al. 1996), and the application of P fertiliser has been observed to reduce incidence of cereal root diseases (Brennan 1989, 1992), the causes of these effects are poorly understood in terms of the interactions among nutrients, plants, pathogens and potential biocontrol biota (Dordas 2008).

Current understanding of the driving forces behind interactions among soil properties, OM and fertiliser additions to soil and communities of soil microorganisms is limited (Johnson et al. 2003), and hinders the development of management options for suppression of soil-borne diseases. Thus, the aim of the work reported here was to assess how the potential for disease suppression to Rhizoctonia root rot, in a highly calcareous soil from a semiarid agricultural region in southern Australia, was influenced by N or P addition with or without C amendment.

Three experiments were undertaken using a highly calcareous coarse-textured soil collected by randomly sampling the top 100 mm depth across an agricultural field located near Streaky Bay on the Eyre Peninsula (32′48.956°S, 134′10.175°E) in South Australia. The climate of this cereal production region is classified as semiarid, characterised by hot dry summers and cool wet winters (Nix 1975; Cawood and McDonald 1996), with long-term annual median rainfall (370 mm) being less than half of the annual potential evapotranspiration. The majority of the soil profiles in the upper Eyre Peninsula around Streaky Bay are classified as Calcarasols according to the Australian classification system and tend to have a highly calcareous alkaline topsoil (Isbell 2002; Hall et al. 2009). The farming systems are mainly low-input based cereal–annual legume pasture rotations (often utilising no N fertiliser and low rates of P) and are highly susceptible to cereal root-disease caused by R. solani AG8. The soil was air dry when collected over the summer fallow period, following a wheat (Triticum aestivum) crop in the previous growing season, and was coarsely sieved through a 2-mm mesh to remove stones and other large material. Commercial laboratory analysis of a subsample reported a pH_(H2O) of 8.4, electrical conductivity (EC_1:5) of 0.208 dS m^–1, extractable nitrate-N and ammonium-N (Searle 1984) of 49 and 4 mg kg^–1 respectively, extractable potassium and P (Colwell 1963) of 403 and 48 mg kg^–1 respectively, a calcium carbonate content (Martin and Reeve 1955) of 73%, an organic C content (Walkley and Black 1934) of 2.25%, and extractable (DTPA) micronutrients (Rayment and Lyons 2011) of 0.20 mg kg^–1 copper, 1.32 mg kg^–1 zinc, 5.6 mg kg^–1 manganese and 4.28 mg kg^–1 iron. Further subsamples of soil were analysed for P using resin extraction and diffusive gradients in thin films (DGT) methodologies that are considered highly appropriate for assessing plant-available P, particularly in calcareous soils (Mason et al. 2010). Resin extractable P was 2 mg kg^–1 (<6 is considered P deficient) and DGT C_E was 298 ug L^–1 (<800 is considered P deficient). Field capacity water content of the soil was determined as 24% w/w using a 1-m column technique (Marshall and Holmes 1979).

For each experiment, bulk amounts of dry soil were moistened with nutrient solutions (made using sterilised reverse osmosis (RO) water) and thoroughly mixed. Soil for all experiments received trace element solution (H₃BO₃ and MnSO₄·4H₂O at 0.31 mg kg^–1 soil; ZnCl₂, CuCl₂·H₂O and MoO₃ at 0.01 mg kg^–1 soil) plus another nutrient solution containing MgSO₄·7H₂O (0.06 g kg^–1 soil) and ferric citrate (0.01 g kg^–1 soil). After the specific treatments for each experiment were imposed (as described in the following sections), the soil was again mixed, divided and allocated into small pots (0.3 L; 0.1 m height) to be used for growing plants in a bioassay.

A bioassay was used in each experiment to assess potential for expression of soil-borne disease suppression (Wiseman et al. 1996; Roget et al. 1999). Following set up of the treatments for an experiment, the soil in each bioassay pot was inoculated with two 10-mm agar plugs of R. solani AG8 strain W19 (isolated from diseased wheat roots) placed at approximately half pot-height equidistant from each other and incubated for 2 weeks in a controlled environment room at 15°C with a 12-h day/night regime. Along with the inoculated bioassay pots, a second set of uninoculated pots were set up and designated healthy controls. The soil utilised in this study had previously been assessed as causing minimal or no R. solani root infection in wheat seedlings (Davey 2013). Wheat (T. aestivum cv. Yitpi) plants were then grown in all the pots of soil from surface-sterilised seeds planted at seven per pot and thinned to five plants after emergence. Plants were grown for 4 weeks and soil maintained at 75% WHC (18% w/w) by watering to weight every 2–3 days. After 4 weeks, shoots were cut off at the base of the stem, dried in an oven for 4 days at 60°C and weighed to enable an average dry weight per plant to be calculated. Roots were extracted from soil, washed and stored frozen before being rated for disease. Following disease rating roots were dried in an oven for 4 days and dry weight recorded.

Roots were visually scored for severity of Rhizoctonia root rot (McDonald and Rovira 1983) on a scale of 0 (no infection) to 5 (severe infection), and also assessed for percent root infection (PRI) using the method described by Barnett et al. (2006). Briefly this involved counting the total number (No.) of seminal roots, the number of seminal roots infected by Rhizoctonia root rot and truncated to less than the depth of the pot, and the number infected with Rhizoctonia but not truncated. Greater weighting was given to truncated roots than those infected but not truncated. These counts were then used to calculate PRI as follows:

Four treatments were set up for this experiment using a fully randomised complete block design (RCBD) with replicates as the blocks: (i) nil addition of N and P, (ii) 100 mg ammonium-N kg^–1 soil added as a solution of (NH₄)₂SO₄ (0.12 mg kg^–1 soil) plus CaCl₂ (0.07 mg kg^–1 soil) and K₂SO₄ (0.06 mg kg^–1 soil) in order to balance cation additions with other treatments, (iii) 100 mg nitrate-N kg^–1 soil added as a solution of Ca (NO₃)₂ (0.15 mg kg^–1 soil) and KNO₃ (0.06 mg kg^–1 soil) and (iv) 50 mg P per kg soil added as a solution of KH₂PO₄ (0.02 mg kg^–1 soil). All solutions were made up using sterilised RO water. Four replicates of each treatment were set up for use in the bioassay with or without the addition of sucrose at 1 g per 100 g dry soil (1% soil dry weight).

Four P treatments were set up in this fully RCBD experiment: nil applied P and three rates of applied P (5, 25 and 50 mg P kg^–1 soil) as Na₂HPO₄ in solution. Soil from each treatment was divided into three sets of six replicate pots (300 mL) and adjusted to 75% WHC using RO water before use in the previously described bioassay.

In addition to shoot and root dry weight determination and root disease assessment as described for Experiment 1, the negative effect of the disease on plant growth in the bioassay in Experiment 2 was assessed as:

Sets of replicates of soil from the treatments were destructively sampled at three stages during Experiment 2. The inoculated zero P treatment only was sampled immediately before inoculation and incubation (i.e. at the start of the bioassay) to assess the baseline amount of DNA present for the soil microorganisms of interest. Inoculated and uninoculated P treatments were sampled during the bioassay (i) immediately before sowing but after the 2-week incubation period and (ii) at the termination of the bioassay after 4 weeks of plant growth.

Soil samples were sent to the SARDI Molecular Diagnostic laboratory (https://pir.sa.gov.au/research/services/molecular_diagnostics) for quantification of DNA of selected organisms using real time PCR (Ophel-Keller et al. 2008). The pathogen assayed was R. solani AG8 and the putative beneficial organisms P. agglomerans, Microbacterium spp., Trichoderma Group A (sect. Pachybasium) and Trichoderma Group B (sect. Trichoderma) (Kullnig-Gradinger et al. 2002; Barnett et al. 2006; Ophel-Keller and Barnett 2006; Vinale et al. 2008b). These specific organisms were investigated because they had been detected in other agricultural soils from the same region where the soil in this study was collected (Davey 2013) and have been shown to be associated with decreased root infection in wheat in a bioassay (Davey et al. 2019).

Further subsamples of soils from after the 2-week incubation of the bioassay (before sowing) were analysed to assess microbial activity and diversity using a modified Microresp^® technique (Campbell et al. 2003; Knox et al. 2009). Microbial catabolic activity was measured as CO₂ production in response to the addition of specific C-containing substrates. Average well colour development (AWCD; Spectromax M4, Moleculardevices.com) from Microresp^®, and the respiration values for different substrates were normalised against average CO₂ response for all 15 C substrates for each sample (Garland 1997). This average metabolic response represented the overall functional capability of soil heterotrophic microbial communities. Microbial diversity was assessed using the C substrate utilisation data from which ‘community-level physiological profiles’ (CLPP) were constructed (Campbell et al. 2003) and heatmaps to illustrate the relative responses to various C substrates as influenced by the treatments were generated (Jones et al. 2019).

Pathogen effectiveness in this study was specified as a measure of ability to cause disease on plant roots per unit of pathogen, and was estimated as:

where t₁ is the soil sampling immediately before sowing following the initial 2-week incubation period for the bioassay and PRI is measured at the conclusion of the bioassay after 4 weeks of growth.

Bulk amounts of the highly calcareous soil amended with finely chopped wheat stubble at 2.8 or 5.6 mg dry matter (DM) kg^–1 dry soil and unamended (control) soils were incubated moist (75% WHC) at 15°C for 6 weeks before P was applied as Na₂HPO₄ in solution at either 25 or 50 mg P kg^–1 soil, or no P was applied. Soil from each treatment was then put into pots for a bioassay as previously described with six replicates per treatment and additional sets of replicates for soil sampling and set out in a fully RCBD.

Sets of replicates from each of the nine treatments were destructively sampled at three stages during Experiment 3: (i) when the bioassay experiment was initially set up at the end of the 6-week incubation period following stubble addition, (ii) immediately before sowing in the bioassay which was after a further 2 weeks of incubation with or without pathogen inoculation and (iii) at the end of the bioassay when plants that had been grown for 4 weeks were sampled and processed for measurement of dry weights and assessment of root infection as described for the previous experiments. Soil subsamples from the first two sampling times were assessed for quantitative analysis of soil microorganism DNA and for microbial activity and catabolic diversity as previously described.

Data were analysed as an RCBD using ANOVA in Genstat Tenth Edition (VSN International Ltd). Normality and distribution of data were assessed before analysis. Least significant difference (l.s.d.) at P = 0.05 was used for comparison of treatment means and interaction effects. Multivariate analysis on the CO₂ evolution and AWCD Microresp^® data was undertaken using PRIMER-E software (https://www.primer-e.com/; Auckland, New Zealand) (Clarke and Gorley 2006). The patterns of CLPP derived from the Microresp^® data were analysed using multivariate statistical techniques such as nonmetric multidimensional scaling analysis (NMDA) to evaluate the relative degree of similarity among environmental samples (Knox et al. 2009). Significant differences in patterns of CLPP were tested for treatments with permutational multivariate analysis of variance (PERMANOVA) (Anderson 2001) and analysis of similarity (ANOSIM) (Clarke and Ainsworth 1993).

There was no significant plant growth response to addition of N as nitrate in either the uninoculated (Fig. 1) or inoculated bioassays (Fig. 2a), and there was a slight but significant increase in plant dry weight for plants given ammonium-N in the uninoculated bioassay (Fig. 1). Addition of nitrate-N in the absence of sucrose (labile C) increased root infection caused by R. solani AG8 (Fig. 2b) whereas ammonium-N did not, and this possibly contributed to the dry weight of plants in this treatment with addition of ammonium-N being greater than those with nitrate-N (Fig. 2a). Sucrose addition markedly reduced root infection for the nitrate-N treatment but slightly increased it in the ammonium-N treatment.

Fig. 1.  Root (RW) and shoot (SW) dry weight (mg plant^–1) for 28-day-old wheat plants from Experiment 1 grown in a highly calcareous soil with no added N or P (Nil) or with N or P added (as NH₄, NO₃ or PO₄) and without pathogen (R. solani) inoculation (healthy controls).

Fig. 2.  (a) Root (RW) and shoot (SW) dry weights (mg plant ^–1) and (b) percent root infection (PRI) and root score for 28 day old wheat plants from Experiment 1 grown in a highly calcareous soil with no added N or P (Nil) or with N or P added (as NH₄, NO₃ or PO₄), and with or without addition of sucrose at 1% soil dry weight. All treatments were inoculated with the pathogen R. solani AG8. Different letters indicate significant differences according to l.s.d. of treatment means at P ≤ 0.05.

Addition of P in the absence of sucrose caused a marked increase in shoot and root growth in the uninoculated (Fig. 1) and inoculated bioassays (Fig. 2a), despite also increasing PRI in the inoculated bioassay (Fig. 2b). Addition of sucrose with or without P addition markedly suppressed root infection to negligible levels (Fig. 2b).

The qualitative visual scoring of root disease symptoms generally resulted in comparable rankings of the treatment effects on root disease expression as those obtained using the quantitative technique of measuring PRI (Fig. 2b), suggesting that this could prove a useful method for disease assessment where rapid screening of large numbers of plants is needed.

After the initial 2-week incubation period of the bioassay (before sowing), the amount of R. solani AG8 DNA for the P addition treatments was unchanged from that measured initially, although it had significantly decreased in the treatment where no P had been added (Table 1a). During this same period, DNA for Microbacterium spp. was maintained at the quantity measured initially, irrespective of P application, with the exception of a significant decrease where 50 units of P had been applied (Table 1a). The DNA for P. agglomerans at this sampling time was not affected by P addition (Table 1a). The DNA for Trichoderma Group B was below detection in all of the treatments although it was initially present in small amounts (Table 1a), and DNA for Trichoderma Group A was below detection in all soils sampled at this time (Table 1a). Microbial activity (Microresp^® CO₂ data) and catabolic diversity (CLPP) of soils at this sampling time demonstrated some significant increases with P application, and this was particularly evident for the highest P rate with several amino acid C sources and some carboxylic acid C sources, but not for carbohydrate C sources except raffinose (Fig. 3).

Table 1.  Amounts of DNA for specific putative beneficial soil organisms P. agglomerans, Microbacterium spp., Trichoderma groups A and B, and the root rot disease causing organism R. solani AG8 in soil sampled from pots in the bioassay for Experiment 2 (a) at the start of the bioassay (Initial) and 2 weeks after the soils had been inoculated with R. solani AG8 and had different rates of P applied (0, 5, 25 or 50 mg P kg soil^–1), and (b) after a further 4 weeks during which wheat plants were grown
Different letters show significant differences between treatment means at P ≤ 0.05

Fig. 3.  Heat map of microbial community functional diversity (or community-level physiological profiles) expressed as AWCD (average well colour development from Microresp^®) for soils from Experiment 2 after 2 weeks of incubation following inoculation with the pathogen R. solani AG8 and addition of P (as orthophosphate) at different rates (0, 5, 25 and 50 mg P kg soil^–1). High substrate utilisation responses are indicated by darker and low responses by lighter colours. The l.s.d. (P < 0.05) values are given where F-test showed significance. Respiration values for different substrates were normalised against average CO₂ response for all 15 carbon substrates for each sample.

Inoculation of the soil with R. solani AG8 changed the microbial profile of the soil. Soil microbial catabolic diversity (CLPP) significantly differed between soils incubated for 2 weeks with and without R. solani AG8 addition. The NMDA analysis showed distinct grouping between soils with and without inoculation (Fig. 4). Inoculation with R. solani AG8 resulted in significant increases in CO₂ evolution (Microresp^®) from a range of amino acid C sources compared with uninoculated soils (Table 2). Within the grouping of communities from the inoculated soils there was a tendency for those at zero P and the lowest P rate to cluster closer to the inoculated grouping and away from those at the higher P rates (Fig. 4). NDMA analysis across all soils (inoculated and uninoculated) sampled at this time indicated no significant effect of P on microbial community diversity (Fig. 4).

Fig. 4.  Multidimensional scaling plot based on Bray–Curtis similarities for log(x + 1)-transformed carbon source utilisation data (community-level physiological profile) from soils sampled in Experiment 2 after 2 weeks of incubation with and without inoculation with R. solani AG8 and addition of P (as orthophosphate) at different rates (0, 5, 25 and 50 mg P kg soil^–1). Significant difference between R. solani uninoculated and inoculated treatments (Permanova P ≤ 0.001) is shown by the solid oval line grouping. Trend within inoculated grouping for lower P rates (0 and 5) to cluster separately to higher P rates (25 and 50) rates is indicated by dotted line groupings.

Table 2.  Average CO₂ evolution (μg CO₂-C g^–1 soil over a 5 h test; Microresp^®) in response to key groups of carbon sources for soils from Experiment 2 after 2 weeks of incubation either with or without R. solani inoculation (Rs) and with different rates of applied P (0, 5, 25 or 50 mg P kg soil^–1)
aa, amino acids; ns, not significant. l.s.d. values are given for main effects of phosphorus application (P) and R. solani inoculation (Rs) for C sources with significant differences at P < 0.05

By the end of the bioassay, after 4 weeks of growth of plants, the amount of DNA for R. solani AG8 and the putative suppressive organisms Microbacterium spp. had increased in all treatments compared with the previous sampling, although it was still significantly lower in the soil without added P than where P had been applied (Table 1b). The amount of R. solani AG8 DNA increased as the rate of P applied increased, as did the DNA of Microbacterium spp. (Table 1b). The DNA for P. agglomerans remained similar across P treatments (Table 1b) and amounts did not differ from those measured at the first sampling (Table 1a). The DNA for Trichoderma Group A was detected after the bioassay, but the amount was not affected by P treatment, and Trichoderma Group B was below detection (Table 1b).

The positive effect of P fertiliser on plant growth in this highly calcareous soil seen in Experiment 1 was confirmed in Experiment 2 (Fig. 5a), and this effect appeared to partially compensate for the fact that P addition increased root infection in this experiment (Fig. 5b) as in the previous one. Consequently, the negative effect of Rhizoctonia root disease on relative reduction in plant dry weight was less at the higher rate of applied P (Fig. 6). Furthermore, the effectiveness of the pathogen did not significantly increase when P was added, with a mean value across all treatments of 28 and a range of 25–30 units of root infection per unit of pathogen DNA.

Fig. 5.  (a) Root (RW) and shoot (SW) dry weight and (b) percent root infection for plants grown in a highly calcareous soil (Experiment 2) after inoculation with R. solani and addition of P (as orthophosphate) at different rates (0, 5, 25 and 50 mg P kg soil^–1). Different letters show significant differences for the treatment means at l.s.d._(0.05) (shoot = 18.91, root = 15.01, PRI = 0.05).

Fig. 6.  Reduction in plant dry weight per percent root infection (relative to a healthy control) for the applied P treatments (0, 5, 25 and 50 mg P kg soil^–1) inoculated with R. solani in Experiment 2. Different letters indicate significant differences at P ≤ 0.05.

As in Experiment 2, the addition of P to the highly calcareous soil in Experiment 3 significantly increased DNA of Microbacterium spp. but not P. agglomerans or Trichoderma Group A, as measured during the bioassay after 2 weeks incubation with R. solani AG8 (Table 3). Furthermore, at this sampling time, where the soil had been amended with stubble there was a significant increase in the DNA of both Microbacterium spp. and P. agglomerans, with greatest amounts observed at the higher stubble addition (Table 3). Addition of P did not significantly affect the amount of R. solani AG8 DNA detected after 2 weeks of incubation (data not shown), whereas it decreased in soils that had been amended with stubble (Table 3).

Table 3.  Amounts of DNA for specific putative beneficial organisms Pantoea agglomerans, Microbacterium spp., Trichoderma Group A and pathogen R. solani AG8 in soil from bioassay pots for Experiment 3 after 2 weeks of incubation with R. solani inoculum plus addition of phosphorus (0, 25 or 50 mg P kg soil^–1) and previous 6 weeks of incubation with 0, 2.8 or 5.6 g stubble dry matter (DM) kg soil^–1
Trichoderma B was not detected in any treatment. The two columns for Microbacterium spp. show data for significant (P < 0.001) main effect of P (left column) and significant (P = 0.008) interaction with stubble (right column). Means for the P treatments are shown for all other organisms where only the stubble main effect was significant (P < 0.001). Different letters show significant differences between treatment means at P ≤ 0.05

Stubble amendment per se followed by incubation for 6 weeks had no effect on the amount of DNA for the native population of R. solani AG8 (Table 4) but increased the DNA for P. agglomerans, Trichoderma Group A and Microbacterium spp., with the latter increasing concurrently with increasing rate of stubble amendment (Table 4). Trichoderma Group B was largely absent from these soils. Microbial activity (Microresp^® CO₂ data) and diversity (CLPP) for stubble-amended soils at this sampling time (6 weeks after incubation with stubble) showed no systematic variation between the treatments (PERMANOVA data not shown).

Table 4.  Amounts of DNA for specific putative beneficial soil organisms P. agglomerans, Microbacterium spp., Trichoderma groups A and B, and the root rot disease causing organism R. solani AG8 in soil from Experiment 3 after 6 weeks incubation with 0, 2.8 or 5.6 g wheat stubble dry matter (DM) kg soil^–1
Different letters indicate significant differences for each organism between stubble rates at P = 0.05. ns, not significant

However, CLPP data did indicate significant changes in the microbial community structure during the bioassay after inoculation with R. solani AG8 and a further 2 weeks of incubation with added P fertiliser Fig. 7), in particular where stubble had been previously added (Fig. 7) and there were some significant interaction in combination with P fertiliser (Fig. 7). The main effect of stubble addition varied depending on the different types of C substrates with some significant positive and negative shifts in the use of specific amino acids and carbohydrates (Fig. 7). There was a strong similarity in the community profiles (CLPP) for soils that received P but were not amended with stubble (see within the black ring in Fig. 8). In soil amended with stubble at 2.8 g kg^–1, community profiles were different to those for soil without added stubble; in addition the profiles in the 25P and 50P treatment differed (indicated by the pink and grey rings in Fig. 8; PERMANOVA – for P treatments coefficient of variation (CV) = 1.27, P = 0.002, for C treatments CV = 1.28, P = 0.001; for P × C interaction CV = 1.91, P = 0.003). Similar to the previous experiment, a significant effect of addition of P on CLPP was observed with some C sources (Fig. 7).

Fig. 7.  Heat map of microbial community functional diversity (or community-level physiological profiles) expressed as AWCD (average well colour development from Microresp^®) for soils from Experiment 3 after 6 weeks of incubation with stubble (0, 2.8 and 5.6 g kg^–1) followed by 2 weeks of incubation after addition of P (as orthophosphate) at different rates (0, 5, 25 and 50 mg P kg soil^–1) and inoculation with the pathogen R. solani AG8. High substrate utilisation responses indicated by darker and low responses indicated by lighter colours. The l.s.d. (P < 0.05) values for Stubble and P fertiliser main effects and Stubble × P interactions are given where F-test showed significance. Respiration values for different substrates were normalised against average CO₂ response for all 15 carbon substrates for each sample.

Fig. 8.  Multidimensional scaling plot based on Bray–Curtis similarity for log(x + 1)-transformed carbon source utilisation profiles for soils from Experiment 3 after 6 weeks of incubation with stubble (0, 2.8 and 5.6 g kg^–1) followed by 2 weeks of incubation with 25 and 50 mg P kg soil^–1. All treatments were inoculated with R. solani AG8 inoculum.

The overall effect of P addition to stubble-amended soil on plants in the bioassay was less Rhizoctonia infection on wheat seedling roots compared with that in soil without P or stubble addition (Table 5). Effectiveness of R. solani AG8 (PRI per unit of pathogen DNA) was not influenced by P addition but was decreased by 50% with stubble amendment: it was 26 units of root infection per unit of pathogen DNA in the nil stubble treatment and reduced to 14 and 13 where soils had been amended with 2.8 or 5.6 g stubble DM respectively.

Table 5.  Treatment means for the main effects in Experiment 3 on percent root infection (PRI), root score and root (RW) and shoot (SW) dry weights (mg plant^–1) of plants that had been grown for 4 weeks in bioassay pots after 2 weeks of incubation with R. solani inoculum plus addition of phosphorus (0, 25 or 50 mg P kg soil^–1) and previous 6 weeks of incubation with 0, 2.8 and 5.6 g stubble DM kg soil^–1
Root scored on scale of 1 (least disease) to 5 (greatest disease) as explained in text. Different letters indicate significant differences for each organism between stubble rates or P at P = 0.05. ns, not significant

Abiotic stress conditions, including edaphic constraints such as availability of C, nutrient deficiencies and other chemical toxicities, can influence the composition and functioning of the soil microbiome and plant–microbe interactions (van Elsas et al. 2002). It is clear from the results of this study that the investigated highly calcareous topsoil presents both abiotic and biotic constraints to the potential for development of suppression for root rot disease, which are linked to limited bioavailability of C and P. Whereas N application alone did not appear to have a major effect on plant growth or potential for disease suppression, indicating the system was not N limited. The study highlighted that labile C input (sucrose addition), P fertiliser addition and the synergy between C added in wheat stubble and P fertiliser were important factors in determining the responses of plants, pathogen and putative disease-suppressing microorganisms that ultimately lead to root infection by R. solani AG8. These factors, as well as the agronomic implications of the results, are discussed further in the following sections.

The positive effects on plants (seen in Experiment 1) of slightly increased plant growth and significantly reduced root infection following addition of a labile C source (sucrose), even in the absence of added nutrients (N or P), highlight that availability of C is a key biotic constraint to the potential for root disease suppression in this soil. This is despite it being a coarse-textured soil that should afford a low level of protection to OM decomposition (Baldock and Skjemstad 2000). Since the soil is from a low-rainfall environment, where annual inputs from crop residues are relatively small in amount and of low quality in terms of C to nutrient ratio, it may be that the predominant forms of organic C in this soil are largely recalcitrant (Baldock et al. 1992). Furthermore, it has been suggested that some calcareous soils occlude soluble organic C (Schmidt et al. 2012), hence, although the organic C content of the soil in this study was reasonably high for a semiarid environment (~2.5%) it is likely that much of that C is not readily available (Demoling et al. 2007). Available C is a key driver for biological activity in soil, and more specifically the effect of OM addition for increased root disease suppression and enhanced beneficial biota is well documented (Alabouvette 1999; Hoitink and Boehm 1999; Bonanomi et al. 2018). Indeed, an effect of C inputs on potential for root-disease suppression in the soil used for this study has previously been reported, although other edaphic constraints were still considered to be limiting (Davey et al. 2019). Rhizoctonia solani AG8 is known to be highly and competitively saprophytic (Sneh et al. 2013), and possibly could access more recalcitrant C sources in the soil OM which might give it an advantage over other microorganisms in a soil where available C is limited.

Although the addition of labile C markedly suppressed root infection in Experiment 1, results from Experiment 2 for putative beneficial organisms suggest that the soil may be inherently limited in terms of a biotic component. Trichoderma species are considered important antagonists for R. solani AG8 (Vinale and Sivasithamparam 2020) but DNA for the two Trichoderma groups (Trichoderma A and B) was either below detection or detected intermittently in small amounts, even in the presence of plants. Similarly, in the presence of plants, the DNA for Microbacterium spp. also increased, but not P. agglomerans for which DNA amounts were constant. The rhizosphere is well recognised as an environment to support growth and function of microorganisms via labile C and other nutrients in root exudates (Bais et al. 2006; de Boer et al. 2006). As mentioned before there are many other putative beneficial organisms that were not assessed in this study but have been isolated under bioassay conditions from other suppressive soils in southern Australia, and if these were assessed it would more fully characterise the suppressive potential of this soil. Indeed, research using metagenomic and transcriptomics tools suggests the involvement of a wide community of bacteria (Yin et al. 2013; Hayden et al. 2018) and fungi (Penton et al. 2014) in the suppression of Rhizoctonia root rot; including several fungi belonging to the Xylariaceae, Bionectriaceae and Hypocreaceae families that have known antifungal capability and were found to be the dominant communities in suppressive compared with nonsuppressive field soils.

The different dynamics of soil organisms measured in this study during plant growth highlights the complex of plant–pathogen–beneficial interactions involved in the expression of disease suppression. In fact, inoculation with pathogen per se for the bioassay also caused a change in catabolic diversity from the indigenous microbial population, as evident by the altered CLPP at 2 weeks after inoculation in Experiment 2, which emphasises the pathogen–soil microbiome interaction. Recent research using ‘omics’ tools have improved our understanding of the general response of microbiomes to multiple influences in natural and agricultural ecosystems (Delgado-Baquerizo et al. 2018); however, the specific drivers for microbial community change in relation to pathogen dynamics, plant–microbe interactions and consequences to disease suppression are yet to be fully resolved (Penton et al. 2014; Delgado-Baquerizo et al. 2020). It is highly likely in this soil also that edaphic factors such as the high pH and salt and carbonate toxicity will influence diversity and function of organisms (Wakelin et al. 2008; Rath et al. 2019), but currently there appears to be limited evidence as to the effects of these specific edaphic factors on potential for suppression of soil-borne root disease (Höper et al. 1995; Senechkin et al. 2014).

Overall, addition of labile C per se did appear to improve potential for disease suppression in this soil, resulting in substantial reduction in root infection and associated small increases in plant growth, although the addition of P had an even greater influence on plant growth and root infection, as discussed in the next section.

The plants were responsive to P application as predicted from the DGT and resin-extraction based assessments of soil P status which indicated likely P deficiency for crop production (Mason et al. 2010) and confirmed that the Colwell P test (which indicated P sufficiency for this soil) may not be suitable for measuring P status in certain soils due to extraction of P from nonlabile sources (Mason et al. 2008). It is well known that soluble P rapidly forms insoluble complexes in these highly calcareous soils of southern Australia (Bertrand et al. 2003; Lombi et al. 2005). Furthermore, the finding that availability of P is a major constraint for plant growth in this highly calcareous soil was not unexpected given reports from other glasshouse and field studies demonstrating that calcareous soils from the region where the soil in this study was collected are highly responsive to fertiliser P, particularly in fluid form (Holloway et al. 2001).

It is evident that P availability in this highly calcareous soil directly constrains the Rhizoctonia pathogen since pathogen DNA decreased 2 weeks after inoculation in the absence of P application, suggesting that the R. solani AG8 population was constrained by P deficiency. However, although P addition maintained pathogen DNA at a constant amount for 2 weeks after inoculation, it was only after the introduction of plants in the bioassay that DNA of R. solani AG8 increased and was responsive to P, as was the DNA for Microbacterium spp. which are commonly associated with plant rhizospheres (Barnett et al. 2006). Also, DNA for Trichoderma Group A was below detection until after plants were grown. Trichoderma species are well documented as antagonistic fungi to a variety of pathogens including R. solani (Grosch et al. 2006; Verma et al. 2007; Vinale et al. 2008a), as well as being plant growth promoting microorganisms (Avis et al. 2008; Vinale et al. 2008b; Vinale and Sivasithamparam 2020). The results from Experiment 2 therefore suggest a role for the plant rhizosphere in both stimulating putative suppressive organisms, potentially via provision of C-rich substrates and other signal molecules in root exudates (Jones et al. 2009; Carvalhais et al. 2015) and increasing P availability in this highly calcareous soil via solubilisation in the rhizosphere (Richardson et al. 2009). Along with the measured increases in DNA of specific soil organisms after the bioassay in Experiment 2 there were some observed changes to the community physiological profiles, indicative of root induced effects on catabolic diversity (Grayston et al. 1998), specifically where P was in sufficient supply.

Despite no increase in DNA for measured organisms before plants in the bioassay, added P was observed to rapidly increase overall microbial activity (Microresp^®) 2 weeks after addition. In P-limited soils from moist tropical forests in Costa Rica, addition of P increased microbial activity over a 30-day period (Cleveland et al. 2002), and in tropical pasture soils in Mexico addition of P increased both microbial activity (CO₂ evolution) and C-immobilisation in the microbial biomass after 20 days (Galicia and Garcia-Olivia 2004). The increased microbial activity observed with added P in this current study coincided with a shift in the CLPP from chiefly carbohydrates with no P addition to include a range of amino acids and some carboxylic acids with P addition, especially at the higher rates of P; and this supported an observed trend for the communities of organisms at higher rates of P (25 and 50) to group together and shift slightly away from those receiving low (5) or no P. Given the limited suite of putative organisms measured it is difficult to speculate as to whether the observed trend for community to shift in response to high P is likely to increase or decrease potential for suppression, but it is clear that the application of P, in the absence of sufficient labile C (as seems to be the inherent state for this soil as discussed earlier), increases the abundance of Rhizoctonia pathogen as well as root infection and general microbial activity, which should potentially lead to increased pathogen effectiveness. However, the effectiveness of the pathogen (PRI per unit of pathogen DNA) did not increase, which was possibly attributable to greater root growth in response to applied P, but also the community changes induced by P could have favoured suppressive organisms that were not assessed in this study. Indeed, addition of nutrients to soil can cause changes in microbial biomass C and N which are linked to microbial community structure (Wardle 1992) and addition of P has been reported to shift the microbial community structure in some soils (Liu et al. 2012). Furthermore, the results for this study indicate that if the applied P rate is high enough (greater than ~25 kg P ha^–1) any increased root infection is unlikely to compromise shoot growth. The increase in plant growth despite the presence of the pathogen and root infection is partly a synergistic effect in which a nutrient-sufficient plant tends to be healthier and more robust. Plants that experience greater stress including nutrient deficiencies (Dubuis et al. 2005) or lack of moisture or suboptimal soil temperature (Duveiller et al. 2007) are reported to be more susceptible to disease; supporting the observation in this study that the plants receiving no or low P had a greater reduction in plant dry weight per unit of root infection than plants receiving higher rates of P.

In summary, the observations of (i) apparent stable pathogen effectiveness and no response in P. agglomerans populations regardless of P application, (ii) an increase in overall soil microbial activity and plant root infection with increasing addition of P alone, (iii) a lack of detectable Trichoderma populations and (iv) increased populations of soil organisms in the presence of plant roots, indicate that the indigenous biota in this highly calcareous soil may be constrained in the development of suppression to Rhizoctonia root rot, not only by the availability of P but also C. Other studies (Roget et al. 1999; Gupta et al. 2009b) have reported that addition of labile C as sucrose without any additional nutrition in calcareous soils had limited beneficial effect in suppressing Rhizoctonia disease, which supports the suggestion that any effect of added C could potentially be improved by the addition of limiting nutrients such as P.

Given observations of the effect of labile C addition to the soil on potential for disease suppression in Experiment 1 of this study, it was not surprising that the addition of stubble per se decreased disease severity on plant roots in the bioassay and the amount of pathogen inoculum (DNA for R. solani AG8). The link between OM and labile C input to soil and an increased potential for root disease suppression is well documented as mentioned earlier, and has specifically been demonstrated for cereal stubble inputs in some Australian soils and farming systems (Roget 1995; Pankhurst et al. 2002a, 2002b; Davey et al. 2019). Further evidence for increased suppression potential from addition of stubble per se is seen in this study since abundance of the putatively suppressive organisms, P. agglomerans and Microbacterium spp., increased after 6 weeks of incubation with stubble, and Trichoderma A was detected after it was initially below detection. The increase in some organisms indicates likely increased population size, although Microresp^® data at this time were too variable to confirm any increased microbial activity or diversity. It is possible that any labile C input from stubble addition to a C-limited soil may have been depleted after 6 weeks, and therefore sampling earlier after stubble addition in Experiment 3 could have improved the likelihood of measuring any significant differences associated with stubble-derived labile C effects on microbial community composition (Gupta and Reddy 2010). However, the significant responses in the CLPP data reflecting the changes in microbial community structure after a further 2 weeks of incubation following addition of P fertiliser suggests a combined influence of C and P fertiliser addition in modifying the microbial functional capacity.

The effect of P addition alone on soil organisms in the bioassay in Experiment 3 was similar to Experiment 2; that is, population size (DNA) for beneficial organisms increased, with less of an effect than stubble alone, whereas the pathogen R. solani AG8 did not increase. Furthermore, the CLPP data indicated that microbial communities significantly changed in response to P, confirming the trend observed in Experiment 2. Whereas with stubble amendment plus P, the increase in the beneficial species combined with the reduction in R. solani AG8 confirms the earlier suggestion that both P and C are limiting the biota, although C appears a key input. Stubble plus P addition also reduced root infection reflecting a potential increase in disease suppression. These results indicate that general suppression (Baker 1968), via antagonistic actions including antibiosis (Fravel 1988) and competition for resources and space (Paulitz 2000), and specific suppression (Mazzola 2002) are likely to both be acting to reduce the effect of P on increasing infection that was seen under low C availability in Experiment 2.

In summary, stubble amendment plus P fertiliser acted in synergy to increase the disease-suppressive potential of this highly calcareous soil by addressing abiotic and biotic edaphic constraints. Addition of stubble without fertiliser P in this highly calcareous soil had no effect on the native population of R. solani AG8 but it did result in an increase in the populations of specific potentially suppressive organisms. Effectiveness of R. solani AG8 (PRI per unit of pathogen DNA) was not influenced by P addition alone but was decreased by 50% when P was added to a stubble-amended soil, as was severity of Rhizoctonia root rot (i.e. PRI). Shifts in microbial community structure that are likely to contribute to general disease suppression appeared to occur when both C and P limitations were addressed. These findings suggest that future research, investigating changes in microbial community composition following C and nutrient addition with functional characterisation using metagenomic techniques, would help identify specific members of the community and/or functional characteristics that contribute to the suppression of soil-borne root diseases.

Application of N alone in this soil did not apparently improve the potential for suppression of disease as evident by the absence of any reduction in root infection relative to unfertilised plants in the inoculated bioassay. However, there were some differences between the two fertiliser N-sources in terms of effects on plant growth and root infection in this bioassay in relation to labile C, whereby plants given nitrate-N were smaller and had higher root infection than those given ammonium-N in the absence of labile C addition (sucrose), but the reverse was the case in the presence of sucrose. There is conflicting evidence regarding the effects of N fertilisers on Rhizoctonia disease–plant growth interactions (MacNish and Neate 1996). A greater increase in mycelial growth of R. solani given nitrate-N relative to ammonium-N has been reported (Ghini et al. 2007), so that higher infection could potentially occur where nitrate is applied, as was the case in the present study in the absence of added C. There are field studies in Australia that also suggest accumulation of mineral N in the surface soil during the autumn period increases disease severity (Gupta and Roget 2011) although others have reported the opposite effect (MacNish 1985). Studies on the severity of take-all under different N forms found that ammonium increased the populations of Pseudomonads, which were antagonistic to take-all, while nitrate increased populations of those that were deleterious to plant growth (Sarniguet et al. 1992).

Overall, N, at the rate applied in Experiment 1 of this study and in either nitrate or ammonium form, did not appear to greatly limit plant growth per se in this highly alkaline soil. However, this study only investigated the individual effects of added P or N and not the effect in combination which has been shown to be important for wheat production on grey calcareous soils (Hancock et al. 2003). Further work to determine the effects of N fertiliser form on potential for root disease suppression in these alkaline soils is merited once P limitations have been addressed using fertiliser application.

The results from this study suggest that increasing stubble and fertiliser P inputs in semiarid farming systems on highly calcareous topsoils should improve the potential for development of root disease suppression where the pathogen R. solani AG8 is present. Indeed, it has been proposed that the long-term adoption of crop management practices that supply greater biologically-available C inputs, either through crop residues or addition of composts and organic manures, can support higher levels of suppression in low OM soils in semiarid southern and Western Australia (Gupta et al. 2011). However, at present there is limited evidence from agronomic field trials in semiarid southern Australia regarding the effectiveness of such management strategies for developing strong disease-suppressive capacity in all soil types (Gupta et al. 2009a, 2012), and some research suggesting that management practices that included stubble retention over the medium term (8 years) did not develop suppression in calcareous soils in the Eyre Peninsula region of South Australia (Cook et al. 2012). It has been said that agronomic trials present the outcome of complex synergistic interactions among many interventions working together (Watt et al. 2006) and this may partly explain why disease suppression is not necessarily improved in some trials, in particular in soils with abiotic or chemical constraints. For example, the pot studies reported here were undertaken under optimal well-watered conditions and did not mimic the seasonal fluctuations in surface soil moisture and temperature in low-rainfall environments typical of southern Australia that are known to affect the incidence and severity of cereal root-rot diseases (Gill et al. 2001a, 2001b, 2002). The rates of P used in this study were somewhat higher than those widely applied in the field by most farmers in the region. Addition of C as stubble was based on average input of 5.6 mg C kg^–1 soil given an average yield of 1.6 t ha^–1 on the Eyre Peninsula and a harvest index of 0.38. Recent yields in this region on the highly calcareous soils are often much lower than the average and when several poor seasons occur in succession, inputs from stubbles will be very low. The soil–plant–microorganism relationships that underpin development of disease suppression and how these are affected by seasonal and annual supply of water in semiarid farming systems remains to be unravelled. Research presented here highlights the need for future targeted and temporally-based approaches preferably in a field environment to further investigate the role of specific differentially expressed microbial taxa (and functional genes) that contribute to the suppression of Rhizoctonia root rot and to ultimately develop suitable management options.

Addition of P fertiliser with labile C input in a highly calcareous topsoil can alleviate edaphic constraints to the growth of plants, as well as increase the activity and diversity of soil microorganisms including the abundance of soil-borne pathogens and some putative beneficial organisms; all of which can enhance the potential for suppression of Rhizoctonia root rot. However, the application of P in the absence of sufficient labile C is likely to increase root infection and the amount of DNA for R. solani AG8, although this may not necessarily increase the effectiveness of the pathogen. Furthermore, if the applied P rate is high enough (greater than ~25 kg P ha^–1) any increased root infection is unlikely to compromise shoot growth, suggesting that improved plant nutrition could mitigate effects of root disease. Overall, results from these studies suggest that management practices which increase the quantity of C in topsoil and biological availability of that C, such as stubble retention but also possibly crop rotation choice or green manure strategies, when combined with strategic P fertiliser decisions can improve the potential for development of suppression to Rhizoctonia root rot in highly calcareous topsoils by addressing biotic and abiotic edaphic constraints. It is recognised that outcomes from agronomic interventions are the result of multiple plant–soil synergistic and antagonistic interactions, and it is therefore difficult to definitively predict the extent of, and time frame for, development of disease-suppressive soil in any farming system. In semiarid systems, the influence of dynamic seasonal and annual water deficits on plant growth, soil community function and amount of residues, may be particularly important in relation to the potential development of disease suppression in the field, but the interactions are not well understood.

The authors declare no conflicts of interest.