First report of Fusarium boothii from pecan ( Carya illinoinensis ) and camel thorn ( Vachellia erioloba ) trees in South Africa

Fusarium boothii forms part of the Fusarium graminearum species complex (FGSC), the important grain pathogen group that causes Gibberella ear rot of maize and Fusarium head blight of wheat. It is known to infect many grain crops such as maize, wheat and barley. Moreover, this pathogen is a 15-ADON mycotoxin producer and thus of During isolations the Hoopstad of the two trees, pecan camel thorn, isolates of the FGSC constituted of a Fusarium Pecan ( Carya in the semi-arid to arid regions the DNA sequence comparisons of α isolates


Introduction
Fusarium head blight (FHB) is a disease of wheat (Triticum aestivum L.) where florets or entire spikes of the wheat plant are infected, giving them a bleached appearance or discolouration at the base of the head (Goswami and Kistler, 2004).
On maize, F. graminearum s.l. causes a disease known as Gibberella ear rot (GER) (Boutigny et al., 2012;Munkfold, 2003;Sampietro et al., 2010Sampietro et al., , 2012. This is due to the occurrence of the sexual state of F. graminearum, namely G. zeae, on the symptoms (Boutigny et al., 2012;Munkfold, 2003). The fungus colonizes the cob from the tip and rot progresses downward (Munkfold, 2003). Similarly to FHB, various mycotoxins are produced in the cob that are dangerous to those consuming the cobs or derived food.
The F. graminearum species complex (FGSC) consists of 15 described species contributing to FHB and GER in various parts of the world (Aoki et al., 2012). Of these, F. graminearum s. str. is the species most commonly associated with disease (Goswami and Kistler, 2004). Although morphologically indistinguishable, the species can be separated by twelve phylogenetic markers (Aoki et al., 2012). Some species such as F. graminearum s. str. occurs globally, while others have only been reported from single countries such as F. aethiopicum O'Donnell, Aberra, Kistler & T.
Aoki (Aoki et al., 2012). The various species also have certain unique chemotypes whereby they have the genetic potential to produce different combinations of mycotoxins (Aoki et al., 2012).
A number of studies has started to elucidate the distribution of members of the FGSC throughout South Africa on grain crops such as wheat, barley and maize (Boutigny et al., 2011a(Boutigny et al., , 2011b(Boutigny et al., , 2012Lamprecht et al., 2011 Rudimentary geographical patterns of these species may exist, and a level of host preference to certain crops has been observed. Of these, F. acaciae-mearnsii is the only species that occurs on a different type of host other than cereals, namely wattle (Acacia mearnsii De Wild.) and eucalypt (Eucalyptus grandis W. Hill ex Maiden) trees (Aoki et al., 2012).
In the past surrounding native vegetation, or naturalized or invasive non-native plants, have been shown to play important roles in the epidemiology of diseases, where these are often alternative hosts where the pathogens can survive, increase in numbers in the absence of the known host plants or generate additional genetic diversity. Absence of such data impacts on our understanding of the current geographical distribution and possible movements of pathogens, and their potential to become genetically more diverse on these hosts. This could negatively affect disease management programmes and mycotoxin risk assessments.
North-western areas of South Africa range from arid to semi-arid. These areas are planted with agricultural crops such as maize, wheat, sunflowers, potatoes and groundnut. Few tree crops are grown in these areas due to lack of water, with pecan nuts [Carya illinoinensis (Wangenh.) K. Koch] one of the few. Pecan is rapidly becoming a large industry in South Africa (Erasmus, 2011). Not many serious diseases are as yet published from this crop. Vegetation surrounding crops in the North-western areas of South Africa comprise of grassland and predominantly Vachellia and Senegalia tree species, which was known as Acacia in the past (Kyalangalilwa et al., 2013). Of these, Vachellia erioloba (E. Mey.) P. J. H. Hurter (camel thorn), previously named Acacia erioloba E. Mey. (Kyalangalilwa et al., 2013), is the dominant form.
A pilot survey was done in the Hoopstad area of South Africa (Free State province) to investigate the co-infection of fungal endophytes and latent pathogens between non-native pecan trees and native V. erioloba. Endophytes are organisms living asymptomatically inside tree tissues, while latent pathogens are plant pathogens that have an endophytic phase (Slippers and Wingfield, 2007). The presence of pathogens known to infect the crops in the area was also investigated in these unrelated trees. A number of Fusarium isolates were obtained from these trees during the survey. Fusarium are well known fungi that include devastating plant pathogens of various crops, mycotoxin producers and also pathogens of humans (Leslie and Summerell, 2006). What was of special interest was a number of isolates morphologically resembling species in the Fusarium graminearum species complex (FGSC). The aim of this study was to determine which species in the complex these isolates represent and if they include pathogens known from South Africa. Possible disease reactions on pecan leaves were also assessed in bioassays.

Collection of samples
Ten trees of V. erioloba and pecan, respectively, were sampled at two sites about 500 m apart in the Hoopstad area, Free State. Ten branches were randomly selected from each tree, and ten leaves were cut from each branch. Ten pieces per leaf and per branch (c. 4 mm diam.) were plated onto 2% Potato Dextrose Agar (Biolab, Merck Millipore, South Africa) after surface sterilization (1 min wash with sterile water, 5 min submersion in 3% sodium hypochloride followed by 5 min in 70% ethanol, final rinse in sterile water). Resultant colonies were purified onto PDA plates and identified morphologically. Isolates resembling Fusarium spp. were single spored and maintained on PDA. Representative isolates were deposited in the National Collection of Fungi, Agricultural Research Council, Pretoria, South Africa.

Identification with DNA sequence comparisons
DNA was extracted from scraped mycelium of six representative and morphologically distinct seven-day-old cultures ( Fig. 1) using the method developed by Möller et al. (1992). PCR amplicons and sequences of the Translation Elongation Factor 1-α and β-tubulin genes were obtained following the protocols of O'Donnell et al. (2000,2004) using the Robust PCR kit (KAPA Biosystems). Amplification products were visualized on 1% agarose gels (Cleaver Scientific, AEC-Amersham, South Africa) containing Gelred DNA stain (Biotium, Anatech, South Africa) under UV illumination using a Geldoc XR+ imaging system (Bio-Rad, South Africa).
Up to 20 ng/µl of PCR amplicons, purified using the EXO/SAP Amplicon Purification system (Werle et al., 1994), were used for sequencing reactions with the BigDye Terminator v3.1 cycle sequencing ready reaction kit (Applied Biosystems, Foster City, CA, USA). Sequencing reactions were purified with EDTA/Ethanol precipitation and run on an ABI 3130XL genetic analyzer (Applied Biosystems).

obtained from Dr Kerry O'Donnell (United States
Department of Agriculture). Maximum likelihood analyses with 1000 bootstrap replicates were done on the datasets with MEGA 6.06 using appropriate models also obtained with MEGA. Initial analyses were done separately on the TEF1-α and βtubulin datasets, but the datasets were combined since both gave congruent phylogenetic trees and were proven previously to be combinable (O'Donnell et al., 2000(O'Donnell et al., , 2004.

Bioassays
Two repeats of fresh leaf bioassays (Keith and Zee, 2010) with five isolates (Fig. 2) in the FGSC were done on fresh, mature pecan leaves to test if the isolates had an ability to cause lesions on this host. Similar assays on V. erioloba leaves proved too difficult due to the small size of the leaf pinnules of the compound leaves. Agar blocks (4 mm diam) with mycelium were placed on 10 surface sterilized leaves for each isolate at wounds created by a sterile needle. Clean agar blocks were used for negative controls. Leaves were kept in moist chambers at room temperature and leaf lesions were scored after three weeks (1=no lesion, 2=small lesion, 3=intermediate lesion, 4=lesion covered leaf). The average scale for each isolate and the negative control were calculated and represented as a graph.

Collection of samples
The number of Fusarium isolates (33)

Identification with DNA sequence comparisons
The combined dataset including TEF1-α and β-tubulin DNA sequences consisted of 71 taxa and 991 characters. In the FGSC, the isolates grouped with F. boothii (Bootstrap support 84%) based on TEF1-α and β-tubulin DNA sequences (Fig. 1). This is different from results of the Brazilian isolates that grouped with F. graminearum s. str. and F. austroamericanum, similar to results from Lazarotto et al. (2014a).

Bioassays
All isolates of the FGSC tested formed lesions in the bioassays (Fig. 2-3). No lesions formed in the control inoculations. The degree of lesions formed varied between isolates with some completely consuming leaves (Fig. 3). The inoculated fungi could be re-isolated from the lesions and their identity were confirmed using DNA sequencing.

Discussion
This study reports the presence of the cereal pathogen F. boothii from tree hosts for the first time. These trees include the common native tree V. erioloba, and the nonnative and unrelated tree crop pecan. This study also indicated that F. boothii coinfected these two unrelated hosts where they co-occur. F. boothii is only known from barley, maize and wheat in several countries in Europe, South America, the U.S.A. and South Africa (Boutigny et al., 2011a(Boutigny et al., , 2011b(Boutigny et al., , 2014Desjardins and Proctor, 2009;Malihipour et al., 2012;Sampietro et al., 2010Sampietro et al., , 2012Tóth et al., 2005) and has never before been reported from a native plant. Two other different types of hosts include soybean [Glycine max (L.) Merr.] (Chiotta et al., 2015) and tomato (Solanum lycopersicum L.) (Gomes et al., 2015).
The fact that the important pathogen F. boothii was found in trees has important implications for the epidemiology of the disease F. boothii causes on barley, maize and wheat. Although the maize fields adjacent to the pecan and V. erioloba trees were not sampled during the survey, it is likely that F. boothii occurred in these fields as it is known from maize, wheat and barley throughout the Free State, North West and Northern Cape (Boutigny et al., 2011a(Boutigny et al., , 2011b(Boutigny et al., , 2012. It is not yet known what other hosts it could infect, and on what it may be present in the Hoopstad area, including other agricultural crops such as sunflowers and groundnut. The host range of F. boothii most likely is much wider than anticipated if the fungus can infect such unrelated and physiologically different hosts. These alternative hosts represent unrecognized sources of pathogen inoculum for the next season of growth, could potentially be sources where the fungus could generate genetic variation developing more virulent and toxigenic strains, or be asymptomatic vessels with which to move the pathogen (Swett and Gordon, 2015). Widely occurring V. erioloba trees could also represent a natural corridor connecting different populations of the pathogen between cultural lands and areas.
Fusarium acacia-mearnsii was described from the non-native trees A.
mearnsii and E. grandis in the KwaZulu-Natal province, South Africa (O'Donnell et al., 2004). This fungus was able to cause lesions on these trees during artificial inoculations, and was associated with lesions in the field (Roux et al., 2001). During national surveys on wheat, maize and barley, this species was detected from wheat in Kwazulu-Natal, and it is able to infect a cereal naturally (Boutigny et al., 2011a). It is thus also able to infect such different types of hosts as has been found for F. boothii in this study. It would have been expected that F. acacia-mearnsii was a likely candidate to infect V. erioloba and pecan in this study, but instead another member of the FGSC was detected. It could be possible that the geographic range of F. acaciamearnsii does not yet include the Hoopstad area in South Africa, as previous studies only linked it to Kwazulu/Natal (Roux et al., 2001;Lamprecht et al., 2011;Boutigny et al., 2011).
There are few significant diseases caused by Fusarium spp. from pecan. F. equiseti has been isolated from seedling necrosis, wilt and root rot symptoms in Brazil and were shown to be pathogenic (Lazarotto et al., 2014b), while F. solani was shown to cause root necrosis (Hsu and Hendrix, 1973). Fusarium species of the F. chlamydosporum species complex, F. oxysporum species complex, F. proliferatum, and F. austroamericanum and F. graminearum s. str. of the F. graminearum species complex were isolated from diseased pecan in Brazil (Lazaretto et al., 2014a).
Bioassays from this study indicated that the endophytically isolated F. boothii can cause lesions on pecan leaves, but more extensive surveys will be necessary to determine if F. boothii is really associated with any naturally occurring disease symptoms on this host.
Another concern that must be investigated, is the association of F. boothii with pecan nuts. F. boothii produces 15-ADON trichothecenes that poses a serious health hazard (Aoki et al., 2012). Our study did not include isolations from pecan nuts, but it must be determined if this fungus can infect and survive in nuts either in the field or during storage, and if it will produce significant levels of mycotoxins. Very few and especially recent studies report on the infection and mycotoxins levels of pecan nuts (Huan and Hanlin, 1975;Terabe et al., 2008).
Only one member of the FGSC is primarily known from trees, i.e. F. acaciamearnsii (Aoki et al., 2012;Roux et al., 2001), while the rest are predominantly known from cereals. Other types of hosts for members of the FGSC include vine, pampas grass, carnations, giant cane, fern, banana, grape ivy, and soil (Aoki et al., 2012). However, F. graminearum s. str. and F. austroamericanum have been reported from pecan in Brazil (Lazaretto et al., 2014a), together with F. boothii from this study. F. meridionale has been found from an orange twig (Aoki et al., 2012), but it is unknown if the fungus occurs naturally on this tree. The presence of four species in the FGSC on woody hosts besides cereals, raises the question if other species in the FGSC could also have such wide and diverse host ranges. Such unexpected host associations has been shown for the important pine pathogen F. circinatum that has also been shown to naturally occur on maize and grasses, a fact that sheds new light on its evolutionary development and epidemiology (Swett and Gordon, 2009;Swett et al., 2014).
The aims of disease management programmes against FHB and GER include the prediction of disease threats and the potential mycotoxin contamination risk (Sampietro et al., 2010). Knowledge of the current species composition and substrate or host associations in a particular area facilitates this because the various species in the FGSC produces known mycotoxins and have fairly well-studied plant health risks.
If there are unprecedented hosts and niches of these pathogens in an area that could harbor problem species, this would compromise these management efforts. This is especially so when there are no typical symptoms when the fungi occur asymptomatically, as was the case in this study. Inadvertent movement of species, genetic diversity of species already known to occur in the area, or new chemotypes of species that can produce other mycotoxins, could also be produced on these unexpected hosts, or they could be natural corridors of movement overlooked when only human-mediated movement are taken into account.
It must be determined if the occurrence of F. boothii on pecan and a native Vachellia species is significant. Unexpected hosts can be refugia for this pathogen that is thought to occur predominantly on grass-like crops. It is known that species in the FGSC can survive on plant debris in the winter (Goswami and Kistler, 2004) and plant matter from these new hosts could provide additional inoculum. The extent of such co-infection could be localized around particular, already infected fields, or perhaps be more extensive and naturally occurring. Hypotheses on evolution, origin, movement, and possible sources of infection are largely based on research from agricultural, food, veterinarian and human substrates. Therefore, the occurrence of known species from unexpected hosts, substrates and niches influences our understanding of these problem fungi. Such questions would have to be pursued with more extensive and targeted sampling.

Acknowledgments
Dr Piet van Wyk was invaluable in sampling efforts in the Hoopstad area. DNA datasets for the FGSC were kindly provided by Dr K. O'Donnell (USDA, USA).     Control PPRI 19334 PPRI 19346 PPRI 19347 PPRI 19336 PPRI 19335 Lesion scores Isolates Lesion score