New Fusarium species from the Kruger National Park, South Africa

Abstract Three new Fusarium species, F. convolutans, F. fredkrugeri, and F. transvaalense (Ascomycota, Hypocreales, Nectriaceae) are described from soils collected in a catena landscape on a research supersite in the Kruger National Park, South Africa. The new taxa, isolated from the rhizosphere of three African herbaceous plants, Kyphocarpa angustifolia, Melhania acuminata, and Sida cordifolia, are described and illustrated by means of morphological and multilocus molecular analyses based on sequences from five DNA loci (CAL, EF-1 α, RPB1, RPB2 and TUB). According to phylogenetic inference based on Maximum-likelihood and Bayesian approaches, the newly discovered species are distributed in the Fusarium buharicum, F. fujikuroi, and F. sambucinum species complexes.


Introduction
Fungi are common colonisers of the plant rhizobiome and endosphere, where they play a key role in modulating the interactions between plant roots and soil (Zachow et al. 2009;Visioli et al. 2014). The direct and indirect interaction between fungal growth in the rhizosphere and its effect on plant growth and health is well docu-mented (Havlicek and Mitchell 2014;Hargreaves et al. 2015;Lareen et al. 2016). Such effects include either a positive feedback by producing plant growth promoting factors, solubilising and stimulating nutrient uptake by plant roots or by inhibiting the growth of concomitant pathogenic organisms (Schippers et al. 1987;Mommer et al. 2016). Conversely, deleterious effects have also been observed, either related to the presence of pathogenic fungal species or caused by fungal-induced modifications of plant root functions, impeding root growth or negatively altering nutrient availability (Schippers et al. 1987;Mommer et al. 2016). Likewise, plants can select and harbour a particular fungal community on its roots via root exudates (Lareen et al. 2016;Sasse et al. 2018), while abiotic influences including water availability, climate and season, soil type, grazers and other animals, orchestrate the development of a unique fungal diversity (Philippot et al. 2013;Havlicek and Mitchell 2014;Hargreaves et al. 2015;Lareen et al. 2016).
The genus Fusarium Link (Hypocreales, Nectriaceae) includes a vast number of species, commonly recovered from a variety of substrates including soil, air, water and decaying plant materials; being also able to colonise living tissues of plants and animals, including humans; acting as endophytes, secondary invaders or becoming devastating plant pathogens (Nelson et al. 1994). In addition to their ability to colonise a multiplicity of habitats, Fusarium is a cosmopolitan genus, present in almost any ecosystem in the world, including human-made settings such as air and dust in the indoor environment or even in hospitals (Perlroth et al. 2007; Aydogdu and Asan 2008;Pinheiro et al. 2011).
Being common inhabitants of plant root ecosystems, fusaria and, particularly  Aoki), have been regularly studied for their interactions with the rhizobiome, motivated mainly by the importance of these organisms as soil-borne plant pathogens and the need to develop effective control mechanisms (Larkin et al. 1993;Hassan Dar et al. 1997;Pal et al. 2001;Fravel et al. 2003;Idris et al. 2006;Díaz Arias et al. 2013). Similarly, abundant data is available regarding the ecology and distribution of plant-associated fusaria, particularly related to pathogenic species or commonly isolated endophytes (Leslie and Summerell 2006). Little attention has however been given to the occurrence of non-pathogenic fungal species, including Fusarium spp. in root microbial communities (Zakaria and Ning 2013;Jumpponen et al. 2017;LeBlanc et al. 2017), while comprehensive DNA sequence-based surveys have been directed mostly to the study of highly relevant and abundant rhizosphere fungal genera such as Trichoderma Pers., Verticillium Nees or mycorrhizal fungi (Zachow et al. 2009;Bent et al. 2011;Ruano-Rosa et al. 2016;Saravanakumar et al. 2016).
The Kruger National Park (KNP) in South Africa is one of the largest natural reserves in Africa, encompassing a number of non-manipulated landscapes, with almost no human alteration (Carruthers 2017). Recently, four research "supersites" have been identified and established in KNP, each of these supersites representing unique geological, ecological and climatic features of the park (Smit et al. 2013). A multidisciplinary study was conducted in KNP aimed to determine functioning and interaction between abiotic and biotic components, as well as soil properties, hydrology and other processes that determine the structure, biodiversity and heterogeneity of a catena or hill slope ecosystem on one of these "supersites", located deep inside the KNP (data not published). In order to assess the microbial soil population and community dynamics, mainly focused on bacteria, several rhizosphere samples were obtained from diverse African plants on one of these exceptional protected savannah landscapes. From these collections, interesting fusaria were isolated from the root ecosystem of three native African herbaceous plants i.e. Kyphocarpa angustifolia (Moq.) Lopr. (Amaranthaceae), Melhania acuminata Mast. (Malvaceae) and Sida cordifolia Linn. (Malvaceae). According to their unique morphological traits and clear phylogenetic delimitations, these isolates are described here as three new Fusarium species.

Study site and sampling
During March 2015, rhizosphere soil from three herbaceous plants was collected in the Southern Granites "supersite" catena (Stevenson-Hamilton supersite) in the KNP, between 25°06'28.6S, 31°34'41.9E and 25°06'25.7S, 31°34'33.7E (Fig. 1). A catena consists of different soil types observed from a crest to a valley bottom with a wetland or drainage exhibiting different water retention capabilities due to the slope or aspect (topography) and the depth of underlying geological rocks (Brown et al. 2004, Van Zijl andLe Roux 2014). The main characteristics of the Stevenson-Hamilton supersite are described in detail by Smit et al. (2013). Briefly, in this site, a single catena landscape covers approximately 1 km from top to bottom and consists of a hill slope, a sodic site (or grazing lawn), a riparian and floodplain area and a dry drainage line. Three species of plants were selected for sampling occurring at the two extremes of the catena. Two of these species (Kyphocarpa angustifolia and Sida cordifolia) occurred at both top and bottom sites while Melhania acuminata only occurred at the top site. The soil (100 mm depth) at the top of the slope is Clovelly with a high percentage of sand (90%) and a low cation exchange capacity (CEC) (mean sodium concentration of 1062 mg/kg) and pH (mean 5.85). The soil at the bottom of the slope is of the Sterkspruit type, with higher clay content thus higher CEC (mean sodium concentration of 3802 mg/kg) and higher pH (mean The arrows indicate the location of the four research "supersites" (adapted from Smit et al. 2013). Sampling site is indicated with a black star. The inset shows the location of the KNP within South Africa, indicated by a grey box. 6.4). Rhizosphere soil of 10 plants of the same species occurring at each top or bottom site was sampled using a core soil sampler. A total of 50 samples consisting of ca. 200 g of soil from the roots of each plant were taken, deposited in zip-lock plastic bags and kept on ice in a cool bag at approximately 5 °C until analysed in the laboratory.

Isolation of Fusarium strains
Soil samples were mixed thoroughly and sieved to remove large elements. Fine soil particles were uniformly spread and distributed over the surface of pentachloronitrobenzene agar (PCNB; also known as the Nash-Snyder medium, recipe in Leslie and Summerell 2006) supplemented with streptomycin (0.3 g/l) and neomycin sulphate (0.12 g/l) and malt-extract agar (MEA; recipes on Crous et al. 2009) on 9 mm Petri dishes and incubated at 24 °C for 10 d under a natural day/night photoperiod. Each soil sample was processed in duplicate. Fungal growth was evaluated daily and growing colonies were transferred to fresh Potato Dextrose Agar (PDA; recipe in Crous et al. 2009). Colonies were evaluated for their macro-and microscopic characteristics and a total of 19 fungal cultures showing features typical of Fusarium were subjected to single spore isolation as described previously (Sandoval-Denis et al. 2018). Single spore isolates were finally transferred and maintained in Oatmeal Agar plates and slants (OA; recipe in Crous et al. 2009). Fungal strains isolated in this study were deposited in the collection of the Westerdijk Fungal Biodiversity Institute (CBS; Utrecht, the Netherlands), the working collection of Pedro W. Crous (CPC), held at CBS (Table 1); and voucher specimens were deposited in The South African National Collection of Fungi (NCF) (Mycology Unit, Biosystematics Division, Plant Protection Institute, Agricultural Research Council, Pretoria, South Africa).

Morphological characterisation
Fusarium isolates were characterised morphologically according to procedures described elsewhere Summerell 2006, Sandoval-Denis et al. 2018). Colonial growth rates and production of diffusible pigments were evaluated on PDA, colony features were also recorded on corn-meal agar (CMA; recipe in Crous et al. 2009) and OA. Colour notations followed those of Rayner (1970). For the study of micro-morphological features, cultures were grown for 7-10 d at 24 °C, using a 12 h light/dark cycle with near UV and white fluorescent light. Aerial and sporodochial conidiophores and conidia and formation of chlamydospores were evaluated on Synthetic Nutrient-poor Agar (SNA; Nirenberg 1976) and on Carnation Leaf Agar (CLA; Fisher et al. 1982). Measurements and photomicrographs were recorded from a minimum of 30 elements for each structure, using sterile water as mounting medium and a Nikon Eclipse 80i microscope with Differential Interference Contrast (DIC) optics and a Nikon AZ100 dissecting microscope, both equipped with a Nikon DS-Ri2 high definition colour digital camera and the Nikon software NIS-elements D software v. 4.30.

DNA isolation, amplification and sequencing
Isolates were grown for 7 d on MEA at 24 °C using the photoperiod described above. Fresh mycelium was scraped from the colony surface and subjected to total DNA extraction using the Wizard® Genomic DNA purification Kit (Promega Corporation, Madison, WI, USA), according to the manufacturer's instructions. Fragments of five DNA loci were amplified using primers and PCR conditions described   (Li et al. 2015). Phylogenetic analyses were based on Maximum-likelihood (ML) and Bayesian (B) analyses, both algorithms run on the CIPRES Science Gateway portal (Miller et al. 2012). Evolutionary models were calculated using MrModelTest v. 2.3 using the Akaike information criterion (Nylander 2004;Posada and Crandall 1998). For ML, RAxML-HPC2 v. 8.2.10 on XSEDE was used (Stamatakis 2014), clade stability was tested with a bootstrap analysis (BS) using the rapid bootstrapping algorithm with default parameters. The B analyses were run using MrBayes v. 3.2.6 on XSEDE (Ronquist and Huelsenbeck 2003) using four incrementally heated MCMC chains for 5M generations, with the stop-rule option on and sampling every 1000 trees. After convergence of the runs (average standard deviation of split frequencies below 0.01) the first 25% of samples were discarded as the burn-in fraction and 50% consensus trees and posterior probabilities (PP) were calculated from the remaining trees.
Phylogenies were first made individually for each locus dataset and visually compared for topological incongruence amongst statistically supported nodes (ML-BS ≥ 70% and B-PP ≥ 0.95) (Mason-Gamer and Kellogg 1996, Wiens 1998), before being concatenated for multi-locus analyses using different locus combinations according to strains and DNA sequences currently available in public databases, in addition to previously published phylogenies (O'Donnell et al. 2000Herron et al. 2015;Lupien et al. 2017;Moussa et al. 2017, Sandoval-Denis et al. 2018. A further 232 sequences representing 72 taxa were retrieved from GenBank and included in the phylogenetic analyses, while an additional 58 DNA sequences were obtained from 24 fungal strains requested from the CBS and NRRL (Agricultural Research Service, Peoria, IL, USA) culture collections (Table 1). All alignments and trees generated in this study were uploaded to TreeBASE (https://treebase.org).

Phylogenetic analyses
Pairwise DNA alignments and BLAST searches using EF-1α and RPB2 sequences showed that the 19 isolates from KNP belonged to three different species complexes of the genus Fusarium i.e. the F. buharicum Jacz. ex Babajan & Teterevn.-Babajan species complex (FBSC; two isolates), the F. fujikuroi Nirenberg species complex (FFSC; two isolates) and the F. sambucinum Fuckel species complex (FSAMSC; 15 isolates). According to these results, sequences of related taxa and lineages were retrieved from GenBank and incorporated into individual phylogenetic analyses for each species complex.
Multi-locus analyses were carried out in order to further delimit the KNP Fusarium isolates amongst the known diversity in their respective species complexes. With the exception of the FFSC, the topologies observed from ML and B analyses of single and multi-locus datasets were highly congruent, with only minor differences affecting unsupported nodes on the trees (all trees available in TreeBASE). The characteristics of the different alignments and tree statistics for all the species complexes are shown in Table 2.
The analysis of the FBSC included sequences of EF-1α, RPB1 and RPB2 loci from 18 isolates representing 10 taxa, including members of the Fusarium torreyae T. Aoki, J.A. Sm., L.L. Mount, Geiser & O'Donnell species complex (FTYSC) and Fusarium lateritium Nees species complex (FLSC) as outgroup (Fig. 2). The four ingroup taxa resolved with high statistical support. Two KNP isolates from K. angustifolia obtained from the bottom site of the catena (CBS 144207 and 144208) clustered in a sister relationship with the clade representing Fusarium sublunatum Reinking, but were genetically clearly delimited.
The phylogeny of the FFSC included sequences of CAL, EF-1α, RPB1, RPB2 and TUB loci from 48 strains and 44 taxa, including two outgroups (F. oxysporum CBS 716.74 and 744.97) (Fig. 3). The phylogeny showed a clear delimitation between the biogeographic clades recognised in this species complex (African, American and Asian clades sensu O' Donnell et al. 1998). Both American and Asian clades where shown as monophyletic with high ML-BS and B-PP support; in contrast, the African clade was resolved as polyphyletic, comprising two distinct and highly supported lineages. A terminal, speciose clade (African A) encompassing 17 taxa and a basal clade (African B), close to the American clade which included the ex-type of Fusarium dlaminii Marasas, P.E. Nelson & Toussoun (CBS 119860) and a sister terminal clade (ML-BS=100, B-PP=1) comprising two KNP isolates from M. acuminata (CBS 144209 and 144495) and two unidentified African Fusarium isolates (CBS 144210 and NRRL 26152). From the loci used here, only TUB resolved both African clades as sister groups; however, its monophyly was not supported by clade stability measurements (data not shown). Conversely, individual CAL, EF-1α and RPB2 phylog-  The FSAMSC was studied using combined RPB1 and RPB2 sequences. The phylogeny included 35 isolates from 20 taxa, including the two outgroups Fusarium circinatum Nirenberg & O'Donnell (CBS 405.97) and Fusarium fujikuroi Nirenberg (NRRL 13566) (Fig. 4)  The clades including KNP isolates and corresponding to previously undisclosed lineages of Fusarium are described in the taxonomy section as the three novel species, F. convolutans, F. fredkrugeri and F. transvaalense. Description. Colonies on PDA growing in the dark with an average radial growth rate of 2.1-4.8 mm/d, 4.4-5.8 mm/d and 4.6-6.3 mm/d at 24, 27 and 30 °C, respectively; reaching 11-28 mm diam. in 7 d at 24 °C and a maximum of 23-37 mm diam. in 7 d at 30 °C. Minimum temperature for growth 12 °C, maximum 36 °C, optimal 27-33 °C. Colony surface white to cream coloured, flat and highly irregular in shape, velvety to felty, with scant and short aerial mycelium; colony margins highly irregular to rhizoid, with abundant white to grey submerged mycelium. Reverse white, straw to yellow diffusible pigment produced between 21-33 °C, scarcely produced and turning luteous to orange at 36 °C. Colonies on CMA and OA incubated in the dark reaching 40-48 mm diam. in 7 d at 24 °C. Colony surface white to cream coloured, flat or slightly elevated at the centre, velvety to dusty; aerial mycelium abundant, short and dense, concentrated on the colony centre; margins membranous and regular, buff to honey coloured, without aerial mycelium. Reverse ochreous without diffusible pigments. Sporulation scant from conidiophores formed on the aerial mycelium, sporodochia not formed. Conidiophores on the aerial mycelium straight or flexuous, smooth-and thin-walled, simple, mostly reduced to conidiogenous cells borne laterally on hyphae or up to 50 μm tall, bearing terminal single or paired monophialides; phialides subulate to subcylindrical, smooth-and thin-walled, 15.5-22 μm long, (3.5-)4-5 μm at the widest point, with inconspicuous periclinal thickening and a shortflared collarette; conidia clustering in discrete false heads at the tip of monophialides, lunate to falcate, curved or somewhat straight, tapering gently toward the basal part, robust; apical cell often equal in length or slightly shorter than the adjacent cell, blunt to conical; basal cell papillate to distinctly notched, (1-2-)3-septate, hyaline, thin-and smooth-walled. One-septate conidia: 24 × 4.5 μm; two-septate conidia: 24.5 × 6 μm; three-septate conidia: (25.5-)29-36.5(-38.5) × (4-)5-6.5(-7.5) μm. Chlamydospores abundantly formed, globose to subglobose, smooth-and thick-walled, (9.5-)11-13.5(-14) μm diam.; terminal or intercalary in the hyphae or conidia, often borne laterally at the tip of elongated, cylindrical, stalk-like projections, solitary or in small clusters. Sterile, coiled, sometimes branched hyphal projections abundantly formed laterally from the substrate and aerial mycelium.
Distribution. Madagascar, Niger and South Africa. Etymology. In honour and memory of Dr. Frederick J. Kruger, pioneer of forest hydrology, fynbos ecology and invasive species and fundamental for the collections included in this study. Notes. This species is genetically closely related to F. dlaminii, both species having similar colonial morphology, optimal growth conditions and biogeography. Moreo-ver, both species exhibit relatively short aerial phialides producing conidia in heads, somewhat resembling those produced by F. oxysporum rather than most members of the FFSC (Leslie and Summerell 2006;Marasas et al. 1985). However, besides exhibiting much faster growth rates, F. fredkrugeri presents clearly distinctive morphological features such as the production of only one type of aerial conidia (vs. two types in F. dlaminii: allantoid to fusiform and 0-septate; and napiform 0-1-septate); orange to pink sporodochia, produced on carnation leaves but also abundantly on the agar surface (vs. orange sporodochia, produced only on the surface of carnation leaves in F. dlaminii) (Leslie and Summerell 2006). Additionally, F. fredkrugeri produces shorter and less septate sporodochial conidia ((1-)3-4-septate and up to 45 μm long in the latter species vs. mostly 5-septate and up to 54 μm long in F. dlaminii) while chlamydospores are not produced. The latter feature, coupled with the somewhat more complex conidiophores also clearly differentiates F. fredkrugeri from F. oxysporum.

Fusarium transvaalense Sandoval-Denis, Crous & W.J. Swart, sp. nov.
MycoBank: MB825104 Fig. 7 Diagnosis. Different from most species in FSAMSC by its slender sporodochial conidia with tapered and somewhat rounded apex; its smooth-to tuberculate, often pigmented chlamydospores and the formation of large mycelial tufts on OA.
Distribution. Australia and South Africa Etymology. After Transvaal, the name of a former colony and Republic located between the Limpopo and Vaal rivers, currently a province of South Africa and where this species was found. From Latin trans meaning "on the other side of " and Vaal a South African river. Notes. Fusarium transvaalense exhibits a sporodochial conidial morphology typical of members of FSAMSC with marked dorsiventral curvature and tapered ends. Several species in FSAMSC form comparable conidia in culture i.e. F. crookwellense L.W. Burgess, P.E. Nelson & Toussoun, F. sambucinum, F. sporotrichioides Sherb., F. venenatum Nirenberg and F. culmorum (Wm.G. Sm.) Sacc. However, with the exception of F. sporotrichioides, the conidia of most species above-mentioned, differ by being more robust and often more pointed apically. Fusarium transvaalense differs from F. sporotrichioides by the absence of pyriform aerial conidia.
Two strains NRRL 13829 and NRRL 31008, previously identified as F. brachygibbosum Padwick showed different degrees of genetic similitude with the new species. While NRRL 31008 clustered within F. transvaalense, NRRL 13829 formed a clearly delimited sister linage. Morphologically, F. transvaalense exhibits significant differences allowing its separation from F. brachygibbosum. Both species produce sporodochial conidia with similar septation and sizes; however, F. brachygibbosum commonly exhibits a bulge in the middle portion of the conidia (Padwick 1945), a feature not present in F. transvaalense. In addition, the latter species produces comparatively larger sporodochial conidia, when elements with the same degree of septation are compared; its chlamydospores are smaller, smooth-walled to markedly tuberculate and pigmented (7-8 μm vs. 10.7-15.3 μm, smooth-walled and hyaline in F. brachygibbosum) and has a distinctive colonial growth on OA, forming large, pustule-like hyphal tufts, a feature not reported for F. brachygibbosum (Padwick 1945).

Discussion
In this study, three new Fusarium spp. were introduced, isolated from rhizosphere soils of three native African shrubs in a protected savannah ecosystem deep inside the Kruger National Park, South Africa. Some remarkable differences were noted regarding the distribution of the novel fungal species and their respective hosts on this particular site. For instance, F. transvaalense, which exhibited the greatest relative abundance, was found in high quantities from the rhizospheres of the three hosts sampled, showing a considerable genetic diversity. Interestingly, this species was only on the top of the catena, even when two of its hosts, K. angustifolia and S. cordifolia, were found and sampled either at the top and bottom sites. Similarly, F. fredkrugeri was recovered only from soils under M. acuminata, a host species which occurred only at the top location. In contrast, F. convolutans was found in the rhizosphere of K. angustifolia, occurring only at the bottom of the catena, while none of the three fungal species was found associated with S. cordifolia at the bottom of the site. Nevertheless, not being an objective of this work, it was not possible to categorically assign these new species to specific hosts or locations. Likely, these fungi could be in low abundance and thus not detectable using the current methods. However, plant species composition varies considerably through a catena ecosystem, in relation to the different soil characteristics, pH gradient and water availability, which also greatly influence microbial and animal biodiversity (Lareen et al. 2016;Mohammadi et al. 2017). However, the full patterns of variation between locations on this particular catena still need to be systematically assessed and compared. As evidenced here, certain differences do exist between the soils at the upper and bottom locations of the Stevenson-Hamilton supersite, which might explain the fungal diversity variation observed here. The cation exchange capacity (CEC; capacity of a soil to hold exchangeable cations) varies considerably between sampling sites, basically depending on the proportion of sand versus clay content of each soil type (Ketterings et al. 2007;Van Zijl and Le Roux 2014). It is known that CEC greatly impacts the soil's ability to retain essential nutrients and prevents soil acidification (Ketterings et al. 2007). Nutrient content also increased from the top to the bottom of the slope which is consistent with the increase in CEC. Nutrient poor soils are also a driver of biological diversity and most likely influenced fungal diversity in these particular locations (Havlicek andMitchell 2014, Mapelli et al. 2017).
The three Fusarium species, described here, were not associated with any visible symptomatology on their hosts. However, they cannot be ruled out as pathogens since they were not assessed for pathogenicity against the sampled plants nor any other putative host species at the same locations. Likewise, it is unknown if these fungi exert any beneficial or deleterious effect on their ecosystems. These are important unsolved questions that need further evaluation. However, as shown by phylogenetic analyses, each of the three new species was in close genetic proximity with well-known plant pathogenic Fusarium spp. on their respective species complexes, which could suggest a potential pathogenic role. Fusarium convolutans clustered within the FBSC, together with three known plant pathogenic Fusarium spp. i.e. F. buharicum, a pathogen of Hibiscus cannabinus L. and Gossypium L.; F. sublunatum, known to affect banana and Theobroma cacao L. in Central America (Gerlach andNirenberg 1982, Leslie andSummerell 2006) and a newly discovered although unnamed phylogenetic species causing wilt, crown and root rot of Hibiscus moscheutos L. (Lupien et al. 2017). Fusarium transvaalense belonged to the FSAMSC, a genetically diverse group common in temperate and subtropical zones (Leslie and Summerell 2006). Fusarium sambucinum, the conserved type species of the genus (Gams et al. 1997) being an aggressive plant pathogen and one of the most important agents of potato dry rot (Peters et al. 2008); while the latter species and several others in the complex have been reported causing disease on diverse crops, including many cereals and fruits (Leslie and Summerell 2006).
Fusarium fredkrugeri is here recognised and formally proposed as a new species. Although the clade representing this taxon had already been identified as a distinct unnamed phylogenetic species by O'Donnell et al. (2000), it had not been given a formal description pending the collection of additional isolates. Two other African isolates previously determined to belong to this clade i.e. CBS 144210 from Striga hermonthica (Del.) Benth. in Madagascar and NRRL 26152 from an unknown substrate in Niger, were incorporated into the analyses, although the latter strain is not viable anymore (NRRL, pers. comm.), thus not available for morphological assessment. Strain CBS 144210, however, is known as a pathogen of the 'purple witchweed', a parasite plant common to sub-Saharan Africa and known to devastate Sorghum bicolor (L.) Moench and Oryza sativa L. plantations (O'Donnell et al. 2000;Yoshida et al. 2010). As previously demonstrated by O'Donnell et al. (2000), our phylogenetic results showed that the clade comprising F. fredkrugeri and its sister species F. dlaminii does not cluster within the main African core of species in the FFSC. Thus, despite the African origin of our isolates, the predicted biogeographic patterns did not match the observed phylogeny. It has been hypothesised that this should not be the result of genetic markers tracing different phylogenies, but the consequence of losing the phylogenetic signal due to saturated sites and introns (O'Donnell et al. 2000). However, the inclusion in our analysis of additional, highly informative and slowly evolving loci such as RPB1 and RPB2 yielded similar results, which points out the need to re-evaluate the phylogeographic arrangement of this important species complex including the vast new data generated during the last 20 years that challenges the established assumptions (Kvas et al. 2009;Walsh et al. 2010;O'Donnell et al. 2013;Laurence et al. 2015). Nevertheless, although rather unlikely, alternative factors such as anthropogenic dispersion of F. fredkrugeri, its host or additional invasive alternative hosts, cannot be rejected as an explanation for the discordance between biogeography and phylogenetic results. However, these scenarios are difficult to imagine given the characteristics of the sampled site, not being an agroecosystem but a protected, isolated zone, with minimal human intervention (Smit et al. 2013).
This study is a new example of how easily new Fusarium spp. can be found when mycological studies are directed to neglected natural ecosystems of minimal anthropogenic disturbance (Phan et al. 2004;Leslie and Summerell 2011;Summerell et al. 2011;Burgess 2014, Laurence et al. 2015. Although irrelevant for some researchers, finding and properly describing new species, regardless of whether they have little or no pathogenic or mycotoxigenic potential, is of utmost importance to improve our understanding on the diversity, biogeographic and phylogeographic patterns of such a complex and heterogeneous genus as Fusarium. In addition, this study remarks on the significance and need to further stimulate the exploration of conserved, non-manipulated natural environments (supersites) and their potential impact on biodiversity research on the fungal kingdom.

Aknowledgments
Todd J. Ward and James Swezey (Agricultural Research Service, Peoria, IL, USA) are thanked for providing strains. We kindly thank Kerry O'Donnell (Mycotoxin Prevention and Applied Microbiology Research Unit, Agricultural Research Service, US Department of Agriculture, Peoria, IL, USA) for providing DNA sequence datasets. Mercia Coetzee (Central University of Technology, Bloemfontein, South Africa) is thanked for her technical support in the field. Alejandra Giraldo (Westerdijk Fungal Biodiversity Institute, Utrecht, The Netherlands) is thanked for her assistance with fungal isolation. Eddie Riddell and Navashni Govender (SANParks) are acknowledged for their research support in the Kruger National Park. We also thank Konstanze Bensch (Mycobank curator) and Uwe Braun (Geobotanik und Botanischer Garten, Martin-Luther-Universität Halle-Wittenberg, Halle, Germany) for their help regarding Latin names.