The diversity of microfungi associated with grasses in the Sporobolus indicus complex in Queensland, Australia

There are five closely related Sporobolus species, collectively known as weedy Sporobolus grasses (WSG) or the rat’s tail grasses. They are fast growing, highly competitive, unpalatable weeds of pastures, roadsides and woodlands. An effective biological control agent would be a welcomed alternative to successive herbicide application and manual removal methods. This study describes the initial exploratory phase of isolating and identifying native Australian microfungi associated with WSG, prior to evaluating their efficacy as inundative biological control agents. Accurate species-level identification of plant-pathogenic microfungi associated with WSG is an essential first step in the evaluation and prioritisation of pathogenicity bioassays. Starting with more than 79 unique fungal morphotypes isolated from diseased Sporobolus grasses in Queensland, Australia, we employed multi-locus phylogenetic analyses to classify these isolates into 54 fungal taxa. These taxa belong to 22 Ascomycete families (12 orders), of which the majority fall within the Pleosporales (>24 taxa in 7 families). In the next phase of the study, the putative species identities of these taxa will allow us to prioritise those which are likely to be pathogenic based on existing literature and their known ecological roles. This study represents the first step in a systematic, high-throughput approach to finding potential plant pathogenic biological control agents.


Introduction
Five species of Sporobolus grasses (S. africanus, S. fertilis, S. jacquemontii, S. natalensis, and S. pyramidalis) are collectively recognised as the weedy Sporobolus grasses (WSG) or rat's tail grasses across Australia (Palmer, 2012;Biosecurity Queensland, 2018).These five species, together with the native Australian species S. blakei, S. creber, S. elongatus, S. laxus, and S. sessilis (Simon and Jacobs, 1999;Peterson et al., 2014), belong to the S. indicus complex (Hetherington and Irwin, 1999;Peterson et al., 2017).The WSG spread prolifically, are unpalatable and lack nutrition for grazing livestock, which reduces carrying capacity for graziers and outcompetes native plants for resources (Palmer, 2012;Ansong et al., 2015).The WSG are morphologically similar and difficult to distinguish from closely related native Sporobolus species (Hetherington and Irwin, 1999;Biosecurity Queensland, 2018).This complicates management strategies, particularly the targeted application of herbicides.In addition, areas requiring WSG management are often vast and in remote locations.WSG are prime targets for biological control (biocontrol) solutions, either separately or in combination with herbicide application and manual control (Yobo et al., 2009;Lawrie, 2011;Lock, 2018;Sutton, 2019).Of the five WSG, S. natalensis is the primary target for control in Queensland as it impacts agriculture, pasture and biodiversity, from the New South Wales border to the Cape York Peninsula (Business Queensland, 2020).
Observations of WSG populations across several years of surveys at multiple locations in Queensland, revealed evidence of S. natalensis dieback or die-off, reduced fecundity of plants, lesions, and other symptoms of fungal disease (Vitelli et al., 2017).As a result of these and other surveys, 79 fungal isolates were collected from symptomatic tissues of S. indicus complex grasses, and transferred onto artificial medium, with the goal of testing them as potential biocontrol agents for the WSG.Many of the isolates in the collection are certainly novel and/or cryptic species, based on based on multi-locus sequence analysis.This is the first step in prioritising these fungi for testing as potential biocontrol agents for WSG.

Sample collection and fungal isolation
Between early 2017 and mid-2021, plant tissue samples were collected from areas with infestations of WSG in Queensland (Table S1).Specifically, S. indicus complex grasses with disease symptoms (leaf chlorosis, leaf and stem lesions, and root death) were collected (Figure 1).Additional samples were collected from Sporobolus spp.cultivated in glasshouses at the Ecosciences Precinct, Dutton Park, QLD, Australia.A 5 cm sample of symptomatic plant material was surface sterilised by submersion in a solution of 70% v/v ethanol and 1% v/v sodium hypochlorite for 30 s, followed by 70% v/v ethanol, then rinsed twice in sterile distilled water, and dried in a laminar flow cabinet on sterile filter paper (as per Bills, 1996).Three to six leaf segments (~2 mm 2 ) from each sample were placed on potato dextrose agar (PDA) amended with either streptomycin (sPDA; 50 mg L −1 ) or chloramphenicol (cPDA; 200 mg L -1 ) and incubated in the dark at room temperature (23−25 °C).As mycelia developed, isolates were sub-cultured onto fresh cPDA, followed by PDA plates for growth of pure cultures.Reference isolates generated in this study were deposited in the culture collection at the Queensland Plant Pathology Herbarium (BRIP), Dutton Park, QLD, Australia.

DNA extraction, amplification, and sequencing
Genomic DNA was extracted with either the Isolate II Plant DNA Kit (Bioline) or the ZymoBIOMICS DNA Miniprep Kit (Zymo Research) as per the manufacturers' instructions, from ~100 mg of mycelium scraped from agar plates.Oligonucleotide primers and Polymerase Chain Reaction (PCR) conditions used to amplify and sequence the targeted loci are listed in the Supplementary Material Table S2.PCRs were performed either with the MyTaq DNA Polymerase (Bioline) or with Phusion HF Master Mix (New England Biolab) according to manufacturer's instructions, using 10 mmol of each primer and 2−3 mL neat DNA extract, to a total reaction volume of 25 mL per reaction.PCR products were purified with the ISOLATE II PCR and Gel Kit (Bioline) according to manufacturer's instructions, eluted with sterile distilled H 2 O, and submitted to Macrogen Inc. (Seoul, South Korea) for bidirectional sanger sequencing.

Phylogenetic analysis
DNA sequence chromatograms of the sequenced loci were viewed, edited and assembled in Geneious Prime v. 11.1.2(Biomatters Ltd., Auckland, New Zealand) and deposited in GenBank (Supplementary Material Table S1).DNA sequences were aligned with selected reference sequences downloaded from NCBI (Supplementary Material S3) using the MAFFT algorithm (Katoh et al., 2002) as implemented in Geneious.An initial two-loci phylogeny was constructed with a combined alignment of the internal transcribed spacer (ITS) region and 28S large subunit ribosomal RNA (LSU) sequences with Ustilago abaconensis ex-type CBS 8380 as the outgroup.Based on this initial identification, a more in-depth phylogenetic analysis was undertaken for each family based on DNA sequences from additional nuclear loci including the 18S small subunit ribosomal RNA (SSU), partial region of the glyceraldehyde-3phosphate dehydrogenase (gapdh), RNA-directed polymerase II subunit 2 (rpb2), beta-tubulin (tub2), and translation elongation factor 1-a (tef1-a).All phylogenies were constructed using maximum likelihood (ML) with the RAxML v. 7.2.8 (Stamatakis, 2014) plug-in in Geneious starting from a random tree topology.The nucleotide substitution model used was General Time-Reversible (GTR) with a gamma-distributed rate variation.In addition, the Bayesian analysis was performed using the MrBayes v.3.2.1 (Huelsenbeck and Ronquist, 2001) plug-in in Geneious.To remove the need for a priori model testing, the Markov chain Monte Carlo (MCMC) analysis was set to sample across the entire GTR model space with a gammadistributed rate variation across the nucleotide sites.Ten million random trees were generated using the MCMC procedure with four chains.The sample frequency was set at 2000 and the temperature of the heated chain was 0.1."Burn-in" was set at 25%, after which the log-likelihood values were stationary.

Phylogenetic analysis
In total, 79 fungal isolates were analysed in this study (Supplementary Material Table S1).Sequences from ITS and LSU resolved these isolates into 22 families representing 12 orders, all within the Ascomycota (Figure 2).Subsequent phylogenies used loci that were informative for each family.

Dothideales
The phylogenetic analysis of four gene regions (ITS, LSU, rpb2, and tub2) identified two isolates as Aureobasidium (Saccotheciaceae; Figure 5).One isolate (BRIP 70138) was isolated from the seed of S. natalensis and is closely related to the ubiquitous A. melanogeum.The other isolate (BRIP 68300) was isolated from leaf tissue of S. natalensis, and is a novel species sister to A. mangrovei, a species described from plant debris in a freshwater habitat in Oman (Alaraimi et al., 2019).

Glomerellales
The phylogenetic analysis of three gene regions (ITS, gapdh, and tub2) identified ten isolates as Colletotrichum (Glomerellaceae; Figure 6).Six of these isolates (BRIP 68238,BRIP 68239,BRIP 69018,BRIP 69684,and BRIP 70190) clustered in the same clade as the ex-type strain of Co. karsti, whilst two isolates (BRIP 70194, and BRIP 71165) clustered in the same clade as the ex-type strain of Co. gigasporum.Two isolates are likely to represent taxonomic novelties, one (BRIP 68820) in the Co. gloeosporioides species complex, and the other (BRIP 68299) in the Co. graminicola species complex.

Helotiales
The phylogenetic analysis of four gene regions (ITS, LSU, SSU, and rpb2) identified one isolate (BRIP 69689) isolated from the roots of S. natalensis as sister to the yellow rot fungus Scytalidium sphaerosporum (Helotiales; Figure 7).

Hypocreales
Three isolates isolated from the leaves of S. natalensis were identified to belong to three families within Hypocreales.The phylogenetic analysis of three gene regions (LSU, rpb2, and tef1a) identified one isolate (BRIP 70643) as a novel genus within Clavicipitaceae (Figure 8A).The phylogenetic analysis of four genes regions (ITS, LSU, rpb2, and tef1-a) identified one isolate (BRIP 66083) in the same clade as the ex-type strain of Fusarium proliferatum (Nectriaceae; Figure 8B).The phylogenetic analysis of three gene regions (ITS, LSU, and tub2) identified isolate BRIP 68235 as a novel species in Parasarcocladium (Sarocladiaceae; Figure 8C).

Pleosporales
Most of the isolates examined in this study belonged to six families within Pleosporales.The phylogenetic analysis of two gene regions (ITS and rpb2) identified three isolates to belong to Didymellaceae (Figure 10A).One species (BRIP 70507, BRIP 70661, BRIP 70883, and BRIPs 70564-70566) is a novel Epicoccum, isolated from stem, roots and leaves of S. elongatus and S. natalensis.This species was identified as closely related to E. multiceps.One isolate (BRIP 70187) was phylogenetically close to Leptospherulina argentinensis and L. gaeumannii.Another isolate (BRIP 65632) was identified as L. queenslandica (Crous et al., 2022).
The phylogenetic analysis of four gene regions (ITS, LSU, SSU, tef1-a) resolved 11 isolates into five species within Phaeosphaeriaceae (Figure 10D).Three isolates (BRIP 70506, BRIP 70650, and BRIP 70656) represent two novel species in Phaeosphaeria.One isolate (BRIP 70642) is a novel species in Parastagonospora, and another isolate (BRIP 70189) is a novel Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS and LSU sequences of the isolates in this study.RAxML bootstrap (bs) values greater than 70% are given at the nodes.In bold font are the isolates from this study and ex-type strains are indicated with T .species in Phaeosphaeriopsis.Interestingly, during this study, 30 isolates representing Stagonospora tauntonensis (Crous et al., 2022) were isolated from multiple Sporobolus spp.collected across the state of Queensland.Six of these isolates were included in the phylogeny.
The phylogenetic analysis of the rDNA (ITS and LSU) found 12 isolates belonged to Pleosporaceae (Figure 11A).Three isolates (BRIP 68520, BRIP 68540, and BRIP 70508) clustered within Alternaria sect.alternata.Additional phylogenetic analysis based on three gene regions (gapdh, rpb2, and tef1-a) Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, LSU, tef1-a and tub2 sequences of related genera from Pestalotopsidaceae (Amphisphaeriales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated with T .

FIGURE 5
Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, LSU, rpb2 and tub2 sequences of related genera in Saccotheciaceae (Dothideales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated by T .

B A FIGURE 4
Phylogenetic trees inferred from RAxML analyses of strains in Chaetosphaeriales based on concatenated alignments of (A) ITS, LSU and tef1-a sequences of related species in Chaetosphaeriaceae; and (B) ITS and LSU sequences of related species in Neoleptoporella (incertae sedis).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated with T .

FIGURE 6
Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, tub2 and gapdh sequences of related genera from the Colletotrichumgleosporioides, Co. graminicola, and Co. boninense species complexes in Glomerellaceae (Glomerellales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated by T .

Magnaporthales
The phylogenetic analysis of four gene regions (ITS, LSU, SSU, and tef1-a) identified three isolates (BRIP 69687,BRIP 69698,and BRIP 70178) as Magnaporthiopsis meyeri-festucae (Figure 12).Magnaporthiopsis meyeri-festucae was associated with summer Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, LSU, rpb2 and SSU sequences of related genera from Scytalidium (Helotiales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font is the isolate from this study, and ex-type strains are indicated with T .patch-like disease symptoms in Festuca spp.(Poaceae) in the USA (Luo et al., 2017).This study represents a new host record for M. meyeri-festucae, and the first report in Australia (Wong et al., 2022).

Discussion
Grass-associated endophytes have been used for the biocontrol of insects and weeds in various agricultural and environmental systems.Neotyphodium spp.(Clavicipitaceae) has been applied to tall fescue (Festuca arundinacea) and perennial ryegrass (Lolium perenne) for the biocontrol of herbivorous insect pests, e.g., the root aphid Aploneura lentisci and wheat sheath miner Cerodontha australis (Saikkonen et al., 2006;Young et al., 2013).Other entomopathogenic fungi, including Beauveria (Cordycipitaceae; Mascarin and Jaronski, 2016), Parametarhizium (Clavicipitaceae; Gao et al., 2021) and Clonostachys (Bionectriaceae; Vega et al., 2008) have been used successfully to control insect pests.The fungal pathogen, Stagonospora convolvuli, produces metabolites that are toxic to crop pathogens (Boss et al., 2007a) and weeds (Boss et al., 2007b).There is evidence that endophyte-containing pasture Phylogenetic trees inferred from RAxML analyses of strains in Hypocreales based on concatenated alignments of (A) rpb2, tef1-a and LSU sequences of related genera from Clavicipitaceae; (B) ITS, LSU, rpb2, tef1-a and tub2 sequences of related species from the Fusarium fujikuroi species complex (Nectriaceae); and (C) ITS, LSU, tub2 sequences of related genera from Sarocladiaceae (Hypocreales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font is the isolate from this study, and ex-type strains are indicated with T .

FIGURE 9
Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, LSU, rpb2 and tef1-a sequences of related genera from Elsinoaceae (Myriangiales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font is the isolate from this study, and ex-type strains are indicated with T .grasses are more competitive against invading weeds (Saikkonen et al., 2013).Several fungal pathogens, including Colletotrichum (Glomerellaceae), Alternaria (Pleosporaceae), and Phoma (Didymellaceae), have been tested as bioherbicides for the control of target weeds (Morin, 2020).Currently in Australia, there is only one commercially registered product, Di-Bak ® Parkinsonia, which contains three species, Lasiodiplodia pseudotheobromae, Macrophomina phaseolina and Neoscytalidium novaehollandiae in the Botryosphaeriaceae, that is available for the biocontrol of the invasive legume Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, LSU, SSU and tef1-a sequences of related genera from Magnaporthaceae (Magnaporthiales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated with T .
Parkinsonia aculeata (Galea, 2021).As individual plants may host hundreds of fungal species, finding a potential biocontrol agent among them is like searching for a needle in a haystack (Peay et al., 2016;Sutton, 2019).
In this study, we resolved the identities of 79 isolates of microfungi from S. indicus species across the state of Queensland.These microfungi were identified based on multi-locus sequence analysis, which is faster and arguably more accurate than using morphology alone, especially when taxonomic characters are lacking, unavailable or uninformative, and where taxonomic novelty is encountered.
Several fungi, including Microdochium dawsonorium, Pestalotiopsis etonensis, and Neopestalotiopsis nebuloides, have been recently described from Sporobolus spp. in Australia (Crous et al., 2020;Crous et al., 2021;Crous et al., 2022).These three fungal species were shown not to be pathogenic or host-specific against WGS (Lock, 2018;Kukuntod, 2020).Consequently, other isolates of Microdochiaceae and Pestalotiopsidaceae found in this study Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, LSU, SSU and rpb2 sequences of related species in Myrmecridiaceae (Myrmecridiales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated with T .
Metagenomic approaches reveal a greater diversity of fungal taxa than isolation methods (Peay, 2014;Persǒh, 2015;FIGURE 14 Phylogenetic tree inferred from a RAxML analysis based on a concatenated alignment of ITS, LSU and tef1-a sequences of related species in Trichosphaeriaceae (Trichosphaeriales).RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated with T .

FIGURE 15
Phylogenetic trees inferred from RAxML analyses of strains in Xylariales based on concatenated alignments of (A) ITS, LSU, rpb2 and tef1-a sequences of related species in Hypoxylaceae; and (B) ITS, LSU and rpb2 sequences of related species in Microdochiaceae.RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated with T .
FIGURE 1 (A-D) Examples of disease symptoms on Sporobolus spp.Plants targeted for sampling, and (E) sampling sites for this study in Queensland, Australia generated using Google Earth Pro©.

FIGURE 7
FIGURE 7 FIGURE 11Phylogenetic trees inferred from RAxML analyses of strains in Pleosporales based on concatenated alignments of (A) ITS and LSU sequences of related species in Alternaria, Bipolaris, Curvularia, and Exserohilum (Pleosporaceae); (B) gapdh, rpb2 and tef1-a sequences of related species in the Bipolaris, Curvularia, and Exserohilum (Pleosporaceae); (C) LSU, ITS, rpb2 and tub2 sequences of related species in Pyrenochaetopsidaceae; and (D) ITS, LSU, rpb2 and tef1-a sequences of related species from Roussoellaceae.RAxML bootstrap (bs) values greater than 70% and Bayesian posterior probabilities (pp) greater than 0.8 are given at the nodes (bs/pp).In bold font are the isolates from this study, and ex-type strains are indicated with T .