Black root rot of cotton in Australia: the host, the pathogen and disease management

Cotton belongs to the genus Gossypium in the Malvaceae family (USDA 2013b), and of the 17 native species of Gossypium in Australia (Brubaker et al. 1999), none produces significant fibre for commercial production. The main commercial cotton grown in Australia today is derived from Gossypium hirsutum (upland cotton, the majority of commercial cotton in Australia) and Gossypium barbadense (pima cotton), which were introduced to Australia as a source of textile fibre from the Americas (Brubaker et al. 1999). Historically, in 1788, Governor Phillip brought seed for commercial cotton production to Australia and cotton was exported for the first time in 1831. The shortage caused by the American Civil War of 1862–65 increased the demand for cotton produced in the state of Queensland (Healy 1923); however, only with the development of irrigation in the 1960s was a stable Australian cotton industry established in northern New South Wales (NSW) and southern Queensland (CCC-CRC 2012a). Currently, 95% of Australia’s cotton growers plant transgenic varieties (CCC-CRC 2012a) and Australian production is the sixth largest in the world, standing at 4.5 million 480-lb bales, with the highest yield worldwide of 4645 lb (2107 kg) per hectare in 2012 (USDA 2013a).

Black root rot is a seedling disease caused by the soil-borne fungal pathogen Thielaviopsis basicola, a species with a worldwide distribution (Nag Raj and Kendrick 1975; USDA 2013b), with the first reported case on cotton in Sacaton, Arizona, in 1922 (King and Presley 1942). Diseased cotton plants show stunted or slow growth early in the season compared with surrounding healthy plants, and delayed flowering or maturity. Belowground symptoms include blackening of the roots and reduced number of lateral roots (Allen 2001).

Although there were no records of root rot in the first 200 years of cotton growing in Australia, it was known as a serious cotton disease in North America, and good crop management was recommended (Healy 1923). In 1930, T. basicola was first observed in Australia, when it was isolated in Queensland from sweet pea, which was grown in the cooler season (Simmonds 1966). In 1981–82, it was reported from other plants in Australia, including tobacco, bean and pine (Warcup and Talbot 1981 and Sampson and Walker 1982, cited in Allen 1990). It was first detected in cotton in 1989 in north-western NSW (Allen 1990), and since then the pathogen has quickly spread to all cotton-growing areas of NSW, most likely by movement of pathogen spores attached to footwear, equipment or machinery. By 2004, T. basicola reached all cotton-growing regions in NSW and Queensland, and the disease was declared an Australian pandemic (Nehl et al. 2004a). Black root rot has had a significant impact on the Australian cotton industry, with delayed crop maturity and yield losses as high as 1.5 bales per acre (i.e. 705 bales per ha) (Jhorar 2004). The 2010–11 cotton pathology surveys (Allen et al. 2012a) showed that black root rot was found in 93% of the farms and 83% of the fields surveyed in NSW.

The origin of T. basicola in Australia is unclear. The disease has been recorded on a wide range of agriculturally and horticulturally important species, both native and exotic (Honess 1994) and has been reported in all states except the Northern Territory. Nevertheless, it is probably not endemic, as despite extensive screening there are no records of natural occurrence of T. basicola in undisturbed Australian soils (Pattemore and Aitken 2000; Harvey et al. 2003). This is in contrast to its presence in uncultivated soil and plants (not always pathogenic to plants) across Europe and the USA (Yarwood 1981). It is possible that the pathogenic strains found on cotton crops were introduced to Australia via the importation of cotton-processing machinery from California (Honess 1994) or in peat (Graham and Timmer 1991). It has been suggested that the infection of native species probably occurred through horticultural practice, using peat or potting mix infected with the pathogen, since peat moss or peat-based media has been identified as a source of infection in greenhouse nurseries (Graham and Timmer 1991; O’Brien and Davis 1994). Thielaviopsis basicola has been recovered from greenhouse air samples (Graham and Timmer 1991) and can be spread by insect vectors (Stanghellini et al. 1999; El-Hamalawi 2008a, 2008b).

In addition to T. basicola, other cotton seedling pathogens, including Rhizoctonia, Pythium, and Fusarium are among the major causes of cotton seedling mortality, which reached 36% in 2011 in NSW (Allen et al. 2012a). Lesions caused by T. basicola may open up the roots for infection by other seedling pathogens that can cause mortality (Allen et al. 2012b); however, it is unclear whether T. basicola alone can cause mortality in cotton. There are some indications in the literature that, under favourable conditions, T. basicola can kill cotton seedlings at the later stages of their development (Hillocks 1992). It has been shown to cause vascular necrosis in the presence of the root-knot nematode, Meloidogyne incognita, in cotton in the USA, leading to increased mortality of cotton seedlings compared with either pathogen alone (Walker et al. 1998). No documentation has been found of M. incognita or other root nematode associated with black root rot in Australia (L. Pereg, unpubl. data). Plant parasitic nematodes of the species Helicotylenchus dihystera were found within roots of Australian cotton (G. hirsutum) without any correlation to fungal disease (Knox et al. 2006). However, a relationship of T. basicola with nematodes or other pests affecting roots cannot be ruled out. While disease surveys in Australia showed that there is no general association between black root rot incidence and seedling mortality, black root rot has been shown to cause a decrease of up to 46% in yield in experimental fields in Australia (Nehl et al. 2004a).

Regular disease surveys of cotton fields in NSW have shown a dramatic increase in the incidence of black root rot caused by T. basicola since it was first observed on cotton in Australia in 1989 (Allen 1990). Cotton disease occurrence and severity on farms, in fields within farms, and on plants within fields has been assessed in Australia in all cotton-growing regions for over a decade (Allen et al. 2012a; CCC-CRC 2012b). Black root rot has been surveyed yearly in Murrumbidgee, Lachlan, Macquarie, Namoi, Gwydir, Macintyre, Bourke/Walgett (NSW), Darling Downs, St George, Theodore/Moura, Emerald and from 2009–10 in Burdekin (Queensland). Surveys since the 2004–05 cotton-growing season show that incidence and disease severity are higher in NSW than in Queensland. In NSW, while all of the growing areas tested have the disease, the Namoi and Macquarie have shown the highest prevalence of black root rot consistently since 2004. Disease severity in the major cotton-growing valleys in NSW (Macintyre, Gwydir, Namoi and Macquarie) has been relatively steady between the 2004–05 season, with 66% of fields and 24% of plants surveyed showing disease, and the 2008–09 season, with 65% of fields and 32% of plants showing disease. In the 2009–10 season, 93% of farms visited and 58% of the fields surveyed in NSW were affected by black root rot, and in 2010–11 the proportions increased to 93% of farms visited and 83% of the fields surveyed in NSW. In Queensland, although the disease was reported to be present in all cotton-growing areas except Burdekin, disease prevalence was relatively low, and at least since 2004–05, black root rot has not been detected in Theodore and Emerald (CCC-CRC 2012b).

The lower prevalence of the disease in Queensland than in NSW is most likely due to the higher temperatures in Queensland at the start of the cotton-growing season. Similarly, in Arizona, black root rot has been more prevalent at higher elevations, where soil temperatures are cooler at planting, than at lower elevations (Mauk and Hine 1988). The disease appears to be most severe early in the growing season when soil temperature is <24°C (Rothrock 1992); high soil water content and poorly drained soils were also reported to enhance disease severity (King and Presley 1942; Rothrock 1992). As soil temperatures increase later in the season and plants resumes growth, the diseased cortical tissue sloughs off the dead cortical cells and roots elongate (King and Presley 1942; Mathre et al. 1966; Mauk and Hine 1988). It is noteworthy that this sloughing-off of the infected tissue leaves the root white, potentially causing marked underestimation of disease prevalence if field surveys are delayed. This was suggested to be the case with the recording of disease in Queensland in the 2010–11 season, as surveys were delayed due to severe flooding (Allen et al. 2012a). In addition to disease severity, the survival of spores of T. basicola also depends on soil parameters, including texture, temperature and moisture (Rothrock 1992), with lower survival of spores at higher temperature of 24−28°C than at 10−18°C. A similar trend was described in both naturally infested and inoculated soils (Papavizas and Lewis 1971; Rothrock 1992).

Movement of fungal spores with irrigation water was suggested as a possible explanation for the steady increase in the prevalence of the disease within farms and fields in Australia (Nehl et al. 2004a). An increase in disease prevalence has been observed constantly in black root rot surveys before 2004 (Nehl et al. 2004a) and from the 2004–05 to 2010–11 seasons (CCC-CRC 2012b). Inoculum of T. basicola was observed in irrigation water and in floating crop residues (Nehl et al. 2004a). Spores of T. basicola may be dispersed attached to soil adhering to the floating residues or possibly in the vascular tissue of the cotton residues (Nehl et al. 2004a). The latter may be possible if internal colonisation of mature plant stems occurs, as may happen occasionally (King and Presley 1942; Mauk and Hine 1988), and might release large amounts of reproduction bodies of T. basicola to the soil. The incidence of black root rot in the largest cotton production areas in NSW could have a high impact on the Australian economy, as Australia has been one of the world’s largest cotton exporters (Dowling 2003; USDA 2013a).

The fast spread of the disease in Australian cotton is rooted in the biology of both the pathogen T. basicola and its cotton host. Understanding the way the pathogen is interacting with the plant, and factors that enhance or suppress the pathogen growth and disease severity in cotton, could lead to improved integrated management of black root rot.

Thielaviopsis basicola is a soil-borne, filamentous, hemibiotrophic fungus (Mims et al. 2000). Fungal hemibiotrophs start their infection cycle with a biotrophic phase and then move to a necrotrophic phase. Although T. basicola is generally considered an obligate parasite (Hood and Shew 1997), it can also associate with hosts in a non-pathogenic manner (Yarwood 1974) and is capable of limited saprophytic utilisation of soil organic matter (Gayed 1972; Chittaranjan and Punja 1994).

The life cycle of T. basicola has been described for tobacco (Hood and Shew 1997), pansy (Mims et al. 2000) and cotton (Mauk and Hine 1988). Based on these descriptions, the cycle can be divided into six major steps: (i) germination of spores; (ii) growth of the germ tube towards roots; (iii) attachment to root surface—the first contact by the pathogen and initial host–pathogen recognition; (iv) differentiation of the pathogen into infection structures and penetration into the host cells; (v) establishment of a biotrophic phase; and (vi) conversion to necrotrophy (root rotting) and the production of new spores. Changes in both organisms during each step of the cycle indicate constant communication between the host and the pathogen.

Communication between cotton host (G. hirsutum) and T. basicola early in the infection cycle was detected using two-dimensional gel electrophoresis separation of cotton root proteomes (Coumans et al. 2009, 2010). It was shown that (1) new proteins are expressed in the cotton root as early as 1 day after infection with T. basicola, and (2) T. basicola is capable of adapting its proteome to germinate and grow according to nutrients available in its environment (Coumans et al. 2009, 2010). Moreover, another study by the same group on the interaction of T. basicola with its hosts showed that germ tubes and hyphae of T. basicola grow specifically in the direction of the host when challenged with germinating cottonseeds (Pereg 2011).

Thielaviopsis basicola produces two types of spores, endoconidia (also known as phialospores) and chlamydospores (also known as aleuriospores). Figure 1a presents the two types of spores produced by T. basicola. In culture, endoconidia are produced within 24 h and chlamydospores within 3 days (Shew and Meyer 1992), with some variations depending on the isolate and culture conditions. The presence of the two types of spores and the shape of the chlamydospores are good morphological tools for the identification of the species. Since the endoconidia have all the essential features found in the Chalara complex, Nag Raj and Kendrick (1975) have placed T. basicola in the genus Chalara, with the species name Chalara elegans. However, DNA sequencing techniques have shown that it belongs to the species T. basicola, and C. elegans is now considered a synonym of T. basicola (Paulin-Mahady et al. 2002).

Fig. 1.  Spores produced by Thielaviopsis basicola. (a) Chlamydospores (large, thick-walled, brown chains) and endoconidia (small, cylindrical thin-walled and stained blue with Cotton Blue) produced by T. basicola, and (b) endoconidial germination in culture. Each single chlamydospore in the chain measures ~10–16 μm in length and 5–8 μm in width. Endoconidia appear cylindrical, truncated at each end, and measure 8–20 μm in length and 4–6 μm in width.

The endoconidia are produced from phialides, and liberated single endoconidia are hyaline, cylindrical with rounded ends and variable in size (Delvecchio et al. 1969) and can only survive in the soil for a few months, with <1% survival in the soil after 15 months (Schippers 1970). The chlamydospores are produced at the tip of hyphae in chains composed of thick-walled melanised (dark) compartments surrounded by a distinct outer wall (Delvecchio et al. 1969). Their thick walls and the presence of melanin protect them from adverse conditions such as extreme temperatures, low moisture, UV radiation and microbial lysis, allowing them to survive in the soil for years as resting spores (Tsao and Bricker 1966). The persistence of the stress-resilient chlamydospores in the soil makes black root rot so hard to eradicate once it reaches a field. Control measures to enhance sanitation, such as ‘come clean go clean’ (Maas 2011; Allen et al. 2012b), have been recommended; however, these methods have little effect once the pathogen reaches a farm, in particular where irrigation is employed as discussed above. There are limited management options currently available for reducing black root rot (Table 1).

Table 1.  Management strategies for controlling black root rot: potentials and limitations
The information in this table is summarised from the review text. All references are given throughout the text, including: Allen et al. 2012b; Maas 2011; Matthiessen and Kirkegaard 2006; Nehl et al. 2004a, Nehl et al. 2004b; Jhorar 2004; Harrison and Shew 2001; Wheeler et al. 1999; Brubaker et al. 1999; Candole and Rothrock 1998; Zaki et al. 1998; Kaufman et al. 1988; Wheeler et al. 1997; Arthur 1996; Butler et al. 1996; Rothrock 1992; King and Presley 1942

Germination of fungal spores is thought to occur in response to signalling, either physical (thigmotropic) or chemical (Prell and Day 2001). Figure 1b presents germinated spores of T. basicola in culture. Endoconidia and chlamydospores of T. basicola were shown not to germinate in fallow soil (Papavizas and Adams 1969) but germinated in the presence of root extracts, some sugars or specific stimulatory substances such as natural lecithins or their constituents, unsaturated fatty acids or unsaturated triglycerides (Mathre and Ravenscroft 1966; Papavizas and Adams 1969). Lindeman and Tousson (1968) showed that germination of T. basicola only occurred in the immediate vicinity of the host (G. hirsutum). In addition, spores of a cotton isolate of T. basicola germinated and reproduced in response to exudates from several hosts (cotton and some legume hosts), including wheat, which had been considered a non-host, as it did not show disease symptoms (Rothrock and Nehl 2000). A decade later, laboratory experiments have shown that wheat can host T. basicola (Pereg 2011) and this will be discussed below.

Root hairs are the primary penetration sites for T. basicola, but penetration has also been reported through other root epidermal cells (Jones 1991; Nan et al. 1992) and through wounds (Baard and Laubscher 1985; Punja et al. 1992). Histological studies revealed that, in cotton seedlings, T. basicola invades the root cortex (Mathre et al. 1966). Root invasion occurs during the first 2–8 weeks of the cotton growth (Hillocks 1992), causing disease symptoms in crop, whereas older roots appear to be more resistant to infection.

Histological studies of the infection process of susceptible and resistant cultivars of tobacco by T. basicola indicated a dynamic interaction in the establishment of the parasitic relationship (Hood and Shew 1997). Spore germination, germ tube growth and penetration of root tissue were similar in the two cultivars, with penetration of root hairs and epidermal cells observed, with epidermal cells the most commonly observed site of infection (Hood and Shew 1996). However, hyphae advancement and lesion development were limited in the resistant cultivar, whose root system outgrew the effects of the initial inoculation (Hood and Shew 1996). In cotton, in vitro studies at 24°C have shown that endoconidia germinated within 6 h of inoculation onto G. barbadense seedling roots and penetration of host tissue occurred within 12 h after inoculation; chlamydospores germinated after 24 h of incubation and host tissue was penetrated within 36 h (Mauk and Hine 1988). Tissue colonisation occurred immediately after penetration, and 10 days after inoculation and incubation, seedlings were stunted, roots were decayed, and the height of the plants was significantly reduced compared with controls (Mauk and Hine 1988). Production of chlamydospores is associated with the necrotrophic phase, and they are produced throughout the root cortex and on the surface of the roots and adjacent soil. The vascular tissue of roots is usually not invaded (Walker et al. 1998), allowing for the survival of the plant host (Hood and Shew 1997; Mims et al. 2000).

Thielaviopsis basicola demonstrates diverse phenotypes when grown in culture (Fig. 2). Colony pigmentation of T. basicola has been described as grey, olive, dark blackish brown or black (Stover 1950; Ellis 1971). Other than brown and grey types, a white (albino) phenotype and ‘sectoring’ of older colonies were observed, where pigmented types developed albino sectors with each variant capable of giving rise to the other during subculturing (Punja and Sun 1999). Moreover, differences in colony appearance, colour, growth rate, production of spores, length of chlamydospore chains and virulence have been observed within axenic cultures grown from single chlamydospore (Huang and Patrick 1971). The number and size of individual spores in the chlamydospore chains vary among isolates of T. basicola (Punja and Sun 1999). Moreover, differences have also been observed among adjacent spores in the same chlamydospore chain. These morphological variants appear to arise frequently during growth in culture of some wild-type isolates and may show variations in pathogenicity (Huang and Patrick 1971). Isolates within a morphological group were not unique to any given geographical region or host of origin (Punja and Sun 1999). Variation in colony morphology, colour and size has also been observed when growing T. basicola in the presence of different plant root extracts (Coumans et al. 2010). Therefore, substantial care is required when attempting to classify T. basicola according to its appearance in culture.

Fig. 2.  Four isolates of Thielaviopsis basicola displaying various morphologies in culture. Top left, cotton isolate; top right, carrot isolate; bottom left, lettuce isolate showing an albino variant; bottom right, lupin isolate.

Sexual reproduction has not been demonstrated in T. basicola (Paulin-Mahady et al. 2002) and it is not known whether isolates may include more than one mating type (Geldenhuis et al. 2006). Since a teleomorph of the barley pathogen Septoria passerinii was found ~125 years after the description of the anamorph (Ware et al. 2007), it is possible that a cryptic sexual cycle of T. basicola may be found, and if it exists, it may explain the high variation observed within cultures of the same isolate and the variations among isolates. The genetic basis for the morphological differences is unknown; however, they may have evolved as a strategy for improving the long-term survival of T. basicola in adverse conditions or new environments.

Australian isolates from two different cotton-growing regions clustered into distinct regional groups when analysed for genetic variability using random amplification of polymorphic DNA (RAPD) (Pattemore and Aitken 2000). Lettuce and peat isolates were also found to cluster together and were distinct from other host groups, indicating that peat was probably the source of T. basicola found in lettuce soils. The isolate from tobacco soil clustered separately from isolates from other hosts (Pattemore and Aitken 2000). Isolates from similar geographic regions or hosts formed distinct groups when analysed for genetic variability using hierarchical clustering analysis of RAPD-PCR results (Punja and Sun 1999). The fact that different isolates of T. basicola from the same host share greater genetic similarity and morphology than those from different hosts indicates that host selection was likely to be determined by genotype (Punja and Sun 1999).

It has been suggested that the high intra-specific variation of strains of T. basicola may be due to a high mutation rate and/or the presence of transposable genetic elements (Punja and Sun 1999) and that T. basicola may contain double-stranded RNA (dsRNA) of viral origin, although some strains may carry two or more distinct, serologically unrelated viruses (Subramanian 1983). The presence of the virus seems to have no profound effect on the growth or morphology of the host, or on pathogenicity of the fungus (Subramanian 1983), although altered fungal physiology and virulence have been reported (Bottacin et al. 1994; Punja 1995). There is no information on the presence of dsRNA viruses in T. basicola in Australia. There is a gap in the information on the genotypes of T. basicola existing in Australia; analysis of isolates from fields across cotton-growing areas in NSW and southern Queensland (L. Pereg, S. Cooper and K. Kirkby, unpubl. data) will shed light on the diversity of this pathogen in Australia and the virulence of distinct isolates towards cultivated cotton.

In a recent study using genomics (internal transcribed spacer sequence analysis, ITS) and proteomics (total protein mapping) tools, various isolates of T. basicola were grouped according to host of origin, namely cotton, carrot or lettuce, irrespective of geographical origin (Coumans et al. 2011). Evidence is accumulating to suggest that isolates of T. basicola show some degree of host specificity or preference (Meyer and Shew 1991; O’Brien and Davis 1994; Coumans et al. 2011; Pereg 2011), which may lead to differentiation of isolates of T. basicola into intra-specific groups.

Other than persistence of the resilient spores of T. basicola in soil, one of the reasons that black root rot is hard to control is the wide host range of the pathogen; T. basicola has a host range of >230 species (Hood and Shew 1997; Farr and Rossman 2013) and is commonly found on plants of the Fabaceae, Malvaceae, Solanaceae and Cucurbitaceae families (Otani 1962; Yarwood 1974, 1981; Shew and Meyer 1992; Subramanian 1968). Of particular importance is the fact that it causes disease on several crop plants, e.g. cotton (Mathre et al. 1966), tobacco, several grasses (Gayed 1972), groundnut, bean (Tabachnik et al. 1979), chicory (Prinsloo et al. 1991), carrots (Punja et al. 1992), citrus (Graham and Timmer 1991), tomato (Koike and Henderson 1998) and pineapple (Wilson Wijeratnam et al. 2005). Ornamental plant species affected include pansies (Viola carnula), sweet pea (Lathyrus odoratus) and Nemesia (O’Brien and Davis 1994).

Cotton and carrot isolates of T. basicola were significantly more virulent towards cotton than were isolates from lettuce (Coumans et al. 2011), and these differences in virulence were associated with proteomic differences between cotton, carrot and lettuce isolates. Two lettuce isolates tested on 13 hosts of T. basicola were found to have clear differences in the susceptibility of 11 lettuce cultivars tested (O’Brien and Davis 1994). The same study showed that disease on bean plants was severe, and on watermelon, cucumber and rock-melon moderate or low, whereas other plants such as eggplant and capsicum showed complete resistance to black root rot. In this trial, the cotton cultivar Gossypium hirsutum cv. Siokra-1-4 showed no disease (O’Brien and Davis 1994). Seven different isolates of T. basicola caused highly variable disease severity on a susceptible tobacco cultivar (Meyer and Shew 1991). In other studies, isolates of T. basicola highly pathogenic on poinsettia have been found to be moderately pathogenic on beans and non-pathogenic on tobacco, whereas tobacco isolates were found to be non-pathogenic on beans and poinsettia (Keller and Shanks 1955; Lloyd and Lockwood 1963). Although isolates of T. basicola from Australian cotton are not limited to causing black root rot on cotton and have also been found to be highly pathogenic towards lupin, pansy and soybean, these isolates are non-pathogenic towards lettuce (Mondal et al. 2004; Pereg 2011), thus displaying host specificity that could possibly be exploited in the search for fungal pathogenicity factors and mechanisms of host resistance.

Thielaviopsis basicola exhibits three modes of interactions with plants (Pereg 2011); isolates may (1) infect the roots and cause disease; (2) infect the roots but not cause disease; or (3) not infect the roots. Pursuant to this organisation, it was suggested that plants be placed into three categories: susceptible hosts, non-susceptible hosts, and non-hosts to T. basicola. Non-susceptible hosts are those in which chlamydospores of T. basicola were detected on healthy-looking roots of the plants. It is unclear whether such infected hosts would remain unaffected by the disease under any given conditions. Nevertheless, it was found that wheat is a non-susceptible host to several isolates of T. basicola, including cotton isolates when tested in soil under controlled condition (Pereg 2011).

Rotation with non-host crops, especially monocots, has been shown to reduce the density of T. basicola in the soil and the incidence of black root rot (Holtz and Weinhold 1994). That study found larger populations of T. basicola in soil in California planted for ≥3 years to cotton than in soil rotated to other crops including barley, wheat or safflower (Holtz and Weinhold 1994). Hairy vetch (Vicia villosa) can be planted during winter as a cover crop to possibly improve soil properties (Rogers and Giddens 1957); seed cotton yield increased by up to 162 kg ha^–1 year^–1 in a cotton production system with legume as a cover crop (Scott et al. 1990). Rotation with other host plants, such as planting soybean in cotton farming systems, may contribute to the cumulative increase of inoculum with time (Mondal et al. 2004). Rotation with a non-host for up to 3 years is one of the current recommendations for black root rot management in cotton (Allen et al. 2012b) (Table 1). It should be noted that, previously, wheat was not considered a host, and therefore was often planted in rotation with cotton in T. basicola-infested fields (Nehl et al. 2004a; Coumans et al. 2010). The finding that wheat is a non-susceptible host (Pereg 2011) can explain the observations that black root rot severity on cotton was similar in rotation and without rotation with wheat in long-term experiments (Nehl et al. 2004a). It can also explain why the proteome response of a cotton isolate of T. basicola to wheat was closer to its response to the cotton host than to vetch (non-host) (Coumans et al. 2010). This highlights the importance of reliable techniques to analyse the interactions between the plant and the fungal isolates (Pereg 2011).

Certain plants have been identified that were more attractive to various isolates of T. basicola than to others, irrespective of the virulence of these isolates on these plants (Pereg 2011). This may represent a component of resistance that could be manipulated. Cotton seedlings were found to be exceptionally attractive to isolates of T. basicola from different origins (carrot, lettuce, lupin, cotton), including isolates that did not cause disease symptoms on cotton (Pereg 2011). A study using protein mapping and ITS analysis demonstrated that, in Australia, strains of T. basicola originated from descendants of a single strain or groups of closely related strains associated with specific hosts (Coumans et al. 2011). Accumulated evidence from these and other studies of T. basicola–host adaptation (Punja and Sun 1999) could explain the difficulties in controlling the disease. The question arises whether different cultivars of commercial cotton differed in their ability to stimulate hyphal growth and in their resistance to infection by various isolates of T. basicola, with potential use in disease management.

Current disease management options recommended for black root rot in Australia are limited to delaying the sowing time to avoid cool conditions that favour the disease early in the season, crop rotation with non-hosts, biofumigation with a green manure crop of vetch, summer flooding, seed treatment with chemicals that induce systemic resistance (acibenzolar-S-methyl) and other soil preparation techniques (Allen et al. 2012b). The development of an integrated disease management program that takes into account pathogen biology and its interactions with cotton under diverse conditions is essential for controlling infection by T. basicola and its persistence in the soil.

Once introduced into the soil, the spores of T. basicola are very persistent under diverse environmental conditions. The thick-walled spores (chlamydospores) produced by the pathogen can survive in the soil for years and start the disease cycle once a host is available. The spores have been easily spread in flood and irrigation water and on cotton residues attached to machinery, equipment and footwear. Farm hygiene practices such as ‘come clean go clean’ (clean footwear and machinery with anti-fungal substances to reduce spore spreading) is always recommended to avoid further spreading. Some of the measures include: cleaning soil and crop debris from vehicles, machinery and footwear and applying an appropriate disinfectant before coming onto, or when leaving, a farm; correctly disposing of crop by-products, residues and trash; retaining tail-water and runoff water on-farm and keeping it out of river systems; and for the farmer to maintain communication with the relevant industry and neighbours about contamination issues (Nehl et al. 2004a; Allen et al. 2012b). However, once the spores have reached a farm, it is very difficult to eradicate them.

Soil fumigation with fungicides could possibly be used to control black root rot. The fungicides do not eradicate the fungal pathogen, but can reduce and suppress T. basicola present in the soil (Matthiessen and Kirkegaard 2006). In the USA, fungicide seed treatments were found to be effective in years with cool and wet early seasons. The systemic, sterol-inhibiting fungicides myclobutanil and triadimenol showed some efficacy in controlling T. basicola when used as cottonseed treatments (Kaufman et al. 1988; Arthur 1996; Butler et al. 1996). Other mixtures of fungicides were found to be efficient in controlling black root rot and seedling disease complexes caused by several pathogens, mainly T. basicola (Arthur 1996; Wheeler et al. 1999). Cotton yield increased when seed was treated with a mix of triadimenol, captan and metalaxyl (Wheeler et al. 1997), and cotton stands increased significantly using a commercial mixture of the fungicides metalaxyl, triadimenol and thiram (Zaki et al. 1998). To date, fumigation has not been a practical control measure for Australian cotton farms, since fumigants do not penetrate and disperse well in the clay soils common to Australian cotton farms, and in-furrow application of several fungicides has been shown to be ineffective (Jhorar 2004; Matthiessen and Kirkegaard 2006). Furthermore, some fungicides have been shown to have a phytotoxic effect on cotton, delaying emergence and slowing plant growth (Jhorar 2004). Pre-treating seeds with fungicides such as triazole or host resistance-inducing chemicals such as acibenzolar-S-methyl (benzothiadiazole) can reduce the incidence of black root rot and improve seedling survival while avoiding the complications associated with soil treatment (Minton et al. 1982; Toksoz et al. 2009). It has shown some promise in pot trials as a method of reducing the severity of black root rot caused by T. basicola when used as a seed treatment immediately before sowing (Mondal et al. 2000). Large-scale field evaluations in cotton fields in Australia demonstrated that, while acibenzolar-S methyl as a seed treatment does not provide complete control by itself, it can induce an increased level of resistance to T. basicola in cotton under moderate disease pressure but it is less effective under high disease pressure (Allen 2007). It can thus be considered a valuable component of an integrated disease management strategy, especially when planting in warmer soil temperatures.

Although efforts have been made toward developing cotton cultivars resistant to fungal diseases, no such cultivar has progressed to the stage of commercial use. An Australian study showed that seven commercial varieties of cotton (G. barbadense and six cultivars of G. hirsutum) were all susceptible to infection by an isolate of T. basicola from cotton-growing soils, with no significant difference in the extent of disease apparent between the cultivars (Honess 1994). By contrast, some native Australian diploid Gossypium species show various levels of resistance to black root rot (Nehl et al. 1998), and the diploid G. arboreum (PI 1415) also shows high resistance to T. basicola (Wheeler et al. 1999). Partial and high resistances to T. basicola have been identified in G. arboreum and G. herbaceum variants, respectively (Wheeler and Gannaway 2007), and crossbreeding, followed by genetic analysis, was done in an attempt to identify quantitative trait loci of these two varieties (Niu et al. 2008). However, being distant diploid relatives of the commercial tetraploid cottons, crossbreeding for resistance poses some difficulties (Brubaker et al. 1999).

The development of transgenic plants with general antifungal genes has also been a priority for cotton industries. Magainin, an antimicrobial peptide isolated from the African clawed frog (Xenopus laevis), has broad-spectrum, antimicrobial activity (Kristyanne et al. 1997). The isoform magainin 2 has shown effects on the ultrastructure of some plant pathogens, completely inhibiting growth of T. basicola in culture (Kristyanne et al. 1997). Tobacco and banana plants transformed with a synthetic analogue of magainin, MSI-99, have been shown to have enhanced resistance to fungal pathogens (Chakrabarti et al. 2003), making it a strong candidate for engineering into cotton and other susceptible plant species to enhance fungal resistance. A cotton cultivar was transformed with a gene from Trichoderma virens (Gv29-8), encoding an endochitinase, and selected transgenic lines demonstrated significant levels of resistance against fungal pathogens, including T. basicola (Howell 2003).

Until recently, sources of resistance for the purposes of crossbreeding were lacking in Australia (Wang and Davis 1997; Allen 2001). More recently, there have been plans to test a genetically modified cotton line that contained a plant defensin gene, nad1, derived from the ornamental tobacco, Nicotiana alata (OGTR 2006). This gene encodes a defensin protein, NaD1, which inhibits the growth of fungi, including T. basicola (OGTR 2006). However, it has been reported that continuous expression of some defensins may have toxic effects, such as abnormal morphology, reduced fertility and reduced cell growth. The presence of mature NaD1 causes abnormal growth of transgenic cotton, resulting in distorted leaves and short internodes. Using a modified defensin, where NaD1 is combined with a C-terminal propeptide domain (CTPP), eliminated the toxic effects in the host plant (Anderson et al. 2009, cited in Stotz et al. 2009). Another defensin, Rs-AFP2, originating from Raphanus sativus (radish) has been shown to suppress several fungal pathogens in vitro, including strains of Fusarium oxysporum and Verticillium dahliae (Vilas Alves et al. 1994). A crude extract from R. sativus seeds has also been shown to suppress T. basicola. Rs-AFP2 reduces growth of some fungi by interacting with glucosylceramide in their cell membrane, increasing membrane permeability (Thevissen et al. 1999, 2003). Although a variety of transgenic plants containing the gene encoding Rs-AFP2 have been produced, including cotton, this did not lead to the release of a commercial product, possibly due to the toxic nature of the protein. To overcome the adverse effects that constitutive expression of defensins inflicts on transgenic plants, a line of cotton was developed expressing the Arabidopsis NPR1 protein (AtNPR1), which was shown to confer resistance against several pathogens, including T. basicola (Kumar et al. 2013). The roots of AtNPR1-overexpressing transgenic plants exhibited stronger and faster induction of several defence genes, particularly PR1, thaumatin, glucanase, LOX1, and chitinase, thus inducing the plant’s own defence system. These transgenic plants also performed better than the wild type plant when they were challenged with T. basicola, showing reduced disease symptoms, significantly higher shoot and root mass, longer shoot length, and greater boll-set (Kumar et al. 2013). These results are promising; however, for practical applications, further studies need to be done to test NPR1-based defence technology in a more controlled manner in order to estimate the metabolic costs to the transgenic plants.

There are several points to consider in the development of transgenic plants resistant to black root rot. Once a disease-suppressive gene is successfully expressed in transgenic plants, the gene product can potentially provide protection throughout the life of the plant. Use of the cry genes, originally isolated from Bacillus thuringiensis, in cotton to protect the plant from bollworm (Helicoverpa sp.) is an example of this type of approach (Perlak et al. 2001; Pray et al. 2002; Pyke 2007). However, black root rot is a seedling disease, affecting the plant mainly in the first few weeks post-planting, and disease levels decline later in the growing season when temperatures increase (Rothrock 1992). Therefore, ongoing production of the anti-microbial or defence proteins is not required for the entire life span of the plant and may put an unnecessary metabolic burden on the growing plant, slowing growth and reducing yield. If a transgenic solution is sought for seedling diseases, a system in which the introduced gene expression is switched off after several weeks should be considered. Also, taking into account the wide range of isolates of T. basicola that affect commercial cotton, breeding and genetic manipulation should be directed towards multiple-resistance to a mixture of strains of T. basicola.

Thielaviopsis basicola has a worldwide distribution and is probably a natural soil inhabitant, being found in virgin soil far removed from crops (Stover 1950; Yarwood 1981).

The severity of black root rot is primarily dependent on the susceptibility of the host plant, the strain of T. basicola and the inoculum concentration at the time of infection. There appears to be a positive correlation between disease severity and inoculum density (Tabachnik et al. 1979; Holtz and Weinhold 1994); however, disease incidence approaches maximum as the population of T. basicola increases above 100 cfu g^–1 soil (Holtz and Weinhold 1994; Nehl et al. 2004a). Unfortunately, it appears that estimation of the populations of T. basicola alone is not an accurate tool for predicting black root rot incidence in Australian cotton-growing systems tested (Nehl et al. 2004a).

There are several other abiotic and biotic factors that have the potential to either suppress or promote T. basicola or black root rot and that could be manipulated in attempting to control the disease. Abiotic factors can influence the severity of black root rot caused by T. basicola. The optimal pH range for growth of T. basicola in culture is 4.0–6.5 (Punja 1993). Soil pH affects the solubility of ions in the soil, increasing it if acid, thus indirectly affecting the distribution and activity of soil microorganisms (Kaufmann and Williams 1964). Soils with pH ≥5.6 and alkaline clay soils are considered conducive, increasing black root rot severity. Soils with pH <5.2 suppress black root rot, decreasing its severity (Bateman 1962; Hillocks 1992; Meyer et al. 1994; Harrison and Shew 2001). Soil type may influence disease prevalence, with weathered ground moraine being suppressive and weathered molasse conducive to black root rot of burley tobacco in Switzerland (Stutz et al. 1989). A large proportion of cotton in Australia is produced in heavy texture, alkaline clay soil (McKenzie 1994 as reported in Nehl et al. 2004a), and black root rot in cotton is more severe in medium clay soils than in very heavy clays or lighter clays, although its severity has not necessarily correlated with water-holding capacity (Nehl et al. 2000).

Black root rot of cotton has been reported to be less severe in well-drained soils (King and Presley 1942). However, survival of T. basicola in soils with water-holding capacities of 45% is lower than in soils with ≤15% (Papavizas and Lewis 1971). Flooding as a control measure in cotton fields has been demonstrated to decrease the severity of black root rot by up to 98%, especially when applied before planting (Jhorar 2004). In Australia, summer flooding is recommended for disease management (Allen et al. 2012b) but is constrained not only by high costs but also by terrain and the availability of water, and due to the risk of disease spread through runoff.

The levels of exchangeable calcium, aluminium, nitrogen and a variety of other ions in soils can affect the severity of black root rot. Both alkaline and acidic soils containing high levels of calcium were reported to promote black root rot (Meyer and Shew 1991; Oyarzun et al. 1998). Acidic soils that suppress the disease generally contain high levels of aluminium, phosphates or nitrogen, reported to inhibit spore germination and hyphal growth (Meyer et al. 1994; Delgado et al. 2006). High levels of aluminium at pH ≤5.0 also inhibit chlamydospore production (Fichtner et al. 2006). In burley tobacco production systems, acidifying fertilisers containing nitrogen are generally recommended as a control measure for T. basicola (Harrison and Shew 2001). However, manipulation of soil pH and ionic concentrations to levels that suppress black root rot is not often a solution available for growers, as such levels are often not optimal for plant growth.

Temperature and inoculum density influence disease development in cotton. The optimum temperature for growth of the pathogen in culture is 20−30°C (Lucas 1955; Mauk and Hine 1988) and survival rates of T. basicola are greater at temperatures of 10−18°C than at 24−34°C (Papavizas and Lewis 1971). However, black root rot is worse at cooler temperatures of 16−20°C and is prevalent in temperatures ≤26°C (Mauk and Hine 1988; Rothrock 1992), presumably since the lower temperature is stressing the growing plant and favouring the pathogen (Lloyd and Lockwood 1963; Mauk and Hine 1988). There may be some adaptation of strains of T. basicola to regional conditions, with an isolate from peat in New Zealand having a temperature optimum of 22°C, compared with 27°C for a cotton isolate from Narrabri, Australia (Honess 1994). Nonetheless, the temperature at the time of planting can influence the severity of black root rot on any crop and it can be used as a method of controlling the disease. The damage to cotton has been shown to be particularly severe when there is an extended period of cool weather in the spring or if crops are planted too early, whereas planting when temperatures are higher can reduce severity even though inoculum levels of T. basicola may be higher (Rothrock 1992; Jhorar 2004).

Biotic factors can also influence the severity of black root rot caused by T. basicola. The use of biofumigation crops has been examined for control of black root rot in cotton. Incorporation of ‘green manure’ crops such as canola, mustard and woolly pod vetch (Vicia villosa) releases compounds that are toxic to soil pathogens. Australian trials have shown that biofumigation crops do not eradicate T. basicola but do reduce disease severity sufficiently to warrant their use (Nehl et al. 2000). The hydrolysis product of the dominant glucosinolates released from the roots of canola, 2-phenylethyl isothiocyanate, was also reported to suppress soil pathogens in vitro, including Thielaviopsis (Smith and Kirkegaard 2002). Hairy vetch has been used successfully in the USA as a biofumigant, planted as a winter cover crop to reduce the incidence of black root rot in the subsequent cotton crop (Rothrock and Kirkpatrick 1995). It was shown to reduce inoculum density of T. basicola and the severity of black root rot on cotton by as much as 60% (Candole and Rothrock 1998). Vetch degradation produces ammonium. High levels of ammonium are suppressive to T. basicola, probably because plants exposed to high levels of ammonium produce substances that are toxic to T. basicola (e.g. putrescine) (Candole and Rothrock 1998). Other reasons could include an increase in the soil pH by the ammonium released. Despite its demonstrated potential, the capacity of hairy vetch to reverse severe infestation of T. basicola has not been proven in fields where cotton is cropped regularly in Australia (Nehl et al. 2004a).

Cruciferous plants are commonly used for biofumigation because they produce isothiocyanates, which inhibit T. basicola and other fungi. Biofumigation with mustards elicited a reduction in black root rot in cotton by up to 70% (Matthiessen and Kirkegaard 2006). The introduction of lucerne or corn stover to soil resulted in a decline in inoculum density of T. basicola and suppressed black root rot, possibly because organic amendments such as these support other organisms that act as antagonists to T. basicola (Papavizas 1968). This was also shown in cotton production system in Spain when growing sugar beet as the preceding crop followed by residue incorporation (Delgado et al. 2006). In this case, the black root rot suppressive effect was also dependent on nitrogen and iron levels in the cotton-growing soil.

Biocontrol options for use against crop diseases, based on microbial antagonism in the soil, are gradually increasing in number, and they have been reported in several reviews on plant growth promoting and disease-suppressive microorganisms. Pseudomonas species synthesise a variety of biocontrol compounds that can suppress root diseases. Pseudomonas fluorescens concentrations are higher in suppressive soils than in soils conducive to black root rot (Ramette et al. 2003), while the P. fluorescens strain CHAO has been shown to suppress black root rot caused by T. basicola in disease-conducive soils (Stutz et al. 1989), including natural soils in Switzerland, in the presence of sufficient amounts of iron (Keel et al. 1989; Defago et al. 1990). Mutations in P. fluorescens CHA0 that blocked the production of Phl, HCN (Laville et al. 1992) and pyoluteorin reduced the ability of these mutants to suppress black root rot (Ramette et al. 2003; Frapolli et al. 2010). Furanones, produced by soil-borne actinomycetes, fungi and bacteria, have shown antifungal activity against phytopathogenic fungi (Paulitz et al. 2000). Pseudomonas aureofaciens has been shown to produce 3-(1-hexenyl)-5-methyl-2-(5H)furanone, a compound with antifungal activity against T. basicola. This furanone has structural similarity to an antifungal furanone produced by the soil fungus Trichoderma harzianum, which has been used as a biocontrol agent against phytopathogenic fungi (Paulitz et al. 2000). Streptomyces hygroscopicus strain TA21 can reduce the incidence of black root rot by 85.3% in greenhouse experiments by inhibiting hyphal growth and reducing spore germination (Yi et al. 2010). Coating of seeds with fungal biocontrol agents is also proving to be an effective control measure against T. basicola. An example of this is the use of Paenibacillus alvei strain K-165, which is applied as a seed coat and then goes on to colonise the rhizosphere and soil, inhibiting root colonisation by T. basicola (Schoina et al. 2011).

Biological stimulation of antimicrobial production by cotton has also been demonstrated. Treatment of cottonseed with preparations of the fungus Trichoderma virens stimulated phytoalexin synthesis in the seedling roots, which was proved fungicidal to the root pathogen Rhizoctonia solani possibly due to the formation of hydrogen peroxide from the breakdown of the phytoalexin. It may prove an effective control treatment for cotton seedling diseases including black root rot (Stipanovic et al. 1992; Howell et al. 1998, 2000).

Black root rot has been recognised as a pandemic in Australian cotton for over a decade, with gradual increase in incidence and severity. There are few disease-management options available for growers for controlling black root rot in Australia (Table 1) and most are mainly based on cropping practices that are not always feasible due to climate and topography. No solution has yet been fully developed through plant breeding or transgenic cotton to control black root rot, and while some promising approaches are arising, further research is required before disease resistant cotton may be developed for commercialisation.

Understanding the biology and ecology of T. basicola and its interactions with its cotton host and other microorganisms in the vicinity of the plant is crucial for developing further strategies for controlling black root rot. Recent studies have developed molecular tools for detecting and studying T. basicola and its interactions with cotton (Coumans et al. 2009, 2010, 2011; Huang and Kang 2010; Pereg 2011) as well as for investigating its relationship with other microorganisms in the soil (L. Pereg, unpubl. data). Elucidating pathogenicity factors in T. basicola in order to develop new control measures would greatly benefit from sequencing the genome of this important plant pathogen.