Moniliophthora roreri, causal agent of cacao frosty pod rot

Summary

Taxonomy: Moniliophthora roreri (Cif.) H.C. Evans et al. 1978; Phylum Basidiomycota; Class Agaricomycetes; Order Agaricales; Family Marasmiaceae; Genus Moniliophthora.

Biology: Moniliophthora roreri attacks Theobroma and Herrania species causing frosty pod rot. Theobroma cacao (cacao) is the host of major economic concern. Moniliophthora roreri is a hemibiotroph with a long biotrophic phase (45–90 days). Spore masses, of apparent asexual origin, are produced on the pod surface after initiation of the necrotrophic phase. Spores are spread by wind, rain and human activity. Symptoms of the biotrophic phase can include necrotic flecks and, in some cases, pod malformation, but pods otherwise remain asymptomatic.

Relationship to Moniliophthora perniciosa: Moniliophthora roreri and Moniliophthora perniciosa, causal agent of witches’ broom disease of cacao, are closely related. Their genomes are similar, including many of the genes they carry which are considered to be important in the disease process. Moniliophthora perniciosa, also a hemibiotroph, has a typical basidiomycete lifestyle and morphology, forming clamp connections and producing mushrooms. Basidiospores infect meristematic tissues including flower cushions, stem tips and pods. Moniliophthora roreri does not form clamp connections or mushrooms and infects pods only. Both pathogens are limited to the Western Hemisphere and are a threat to cacao production around the world.

Agronomic importance: Disease losses caused by frosty pod rot can reach 90% and result in field abandonment. Moniliophthora roreri remains in the invasive phase in the Western Hemisphere, not having reached Brazil, some islands within the Caribbean and a few specific regions within otherwise invaded countries.

Disease management: The disease can be managed by a combination of cultural (for example, maintenance of tree height and removal of infected pods) and chemical methods. These methods benefit from regional application, but can be cost prohibitive. Breeding for disease resistance offers the greatest potential for frosty pod rot management and new tolerant materials are becoming available.

Introduction

Theobroma cacao (cacao), the source of chocolate, is grown throughout the humid tropics (Zhang and Motilal, 2016). Cacao is native to the upper Amazon River, but has naturalized in areas outside its centre of origin in South and Central America. The conditions required for growing cacao mimic those promoting several cacao diseases (Ploetz, 2016). Add to these conditions the perennial nature of the crop and the 5–6 months in which the pods remain on the tree and the stage is set for disease epidemics.

Frosty pod rot (FPR) is caused by Moniliophthora roreri (Mr) (Evans, 2016a). Mr attacks pods in a hemibiotrophic association. Initial infections appear symptomless, apart from tissue swelling in some cases. This condition persists for up to 3 months until pods rapidly necrose, producing spore masses over their surface. Amongst cacao pathogens, many consider Mr to have the greatest disease potential should it reach new production areas (Ploetz, 2016).

Our understanding of the Mr–cacao interaction is expanding. We have pathogen (Meinhardt et al., 2014) and host (Argout et al., 2011) draft genomes and transcriptomes (Meinhardt et al., 2014). Our understanding of the pathogen's genetic diversity makes it clear that genetic recombination in Mr is rare to non‐existent (Ali et al., 2015; Díaz‐Valderrama and Aime, 2016a; Phillips‐Mora et al., 2007). The assessment of Mr genetic diversity and the molecular interactions with cacao currently drive our interpretation of the biology and will be critical to the management of FPR in the future.

Socioeconomic Impact of FPR

On a global scale, Mr is responsible for lower losses than some pathogens because of its limited range (Ploetz, 2016). However, Mr ranks with any of the major pod pathogens in terms of its economic impact during an epidemic (Evans, 1986) and represents one of the principal yield‐limiting factors in tropical America. Severe disease outbreaks have led to the abandonment of cacao cultivation in extensive areas of Peru, Costa Rica, Colombia and Mexico (Phillips‐Mora and Wilkinson, 2007). FPR has been reported to be twice as destructive as black pod rot and more dangerous and difficult to control than witches’ broom disease (WBD) (Phillips‐Mora and Wilkinson, 2007).

The appearance of FPR and WBD in Ecuador at the beginning of the 20th century had profound consequences on the world's major cacao producer at that time (Crawford, 1980), and their presence has brought about severe economic losses in all cacao‐growing regions to which they have spread. Yield reductions range from 50% to 90% for WBD (Meinhardt et al., 2008) and 10% to 100% for FPR (Phillips‐Mora and Wilkinson, 2007). In Colombia, where FPR and WBD are widely distributed, FPR continues to be the most limiting of the two diseases for cacao production (Aranzazu et al., 2000). A similar situation occurs in Ecuador, Peru and Bolivia where both diseases coexist.

The devastating effects of Mr are well documented across countries and epochs (Phillips‐Mora and Wilkinson, 2007). The scale of damage caused by the fungus has resulted in macroeconomic consequences in countries such as Ecuador, Costa Rica and Mexico. The first well‐documented outbreak occurred at the beginning of the 20th century, when Mr invaded the important cacao‐producing areas of Ecuador causing devastating yield losses (Rorer, 1918). The export of cacao from Ecuador fell from 46 000 tons in 1917 to about 20 000 tons in 1925, concomitant with the onset of WBD (Erneholm, 1948; Thorold, 1975). Now, FPR is widely dispersed in most cacao areas in both the coastal and Amazonian regions, and, together with WBD, accounts for an average of 60% pod losses (C. Suárez, Instituto Nacional de Investigaciones Agropecuaria (INIAP), Quevedo, Ecuador). Damage to cacao production in Costa Rica was also substantial after the first detection of Mr in 1978. Between 1978 and 1983, cacao production in Costa Rica declined by 72% and dry cacao bean exports decreased by 96% (Phillips‐Mora and Wilkinson, 2007). Costa Rican cacao production has not recovered.

Mr has spread from Panama through Central America to Mexico over the past 60 years, its movement marked by the damage to cacao production. Where it occurs it invariably becomes the main yield‐limiting factor. The aggressiveness shown by Mr in the most recently infected countries of Bolivia (Phillips‐Mora et al., 2015) and Jamaica (Phillips‐Mora and ten Hoopen, 2017) verifies that the pathogen continues in the invasive phase. There is no reason not to expect the fungus to continue to spread throughout the Western Hemisphere where cacao is grown, and precautions must be taken to delay and prepare for its arrival. It is coming.

Pathogen History and Field Biology

In 1917, J. B. Rorer sent specimens from Ecuador to R. E. Smith at the University of California (Hewitt, 1987), who considered the fungus to be close to the ascomycete fungus Monilia fructicola (Rorer, 1918). Yet, perhaps the earliest record of the disease occurs from the Antioquia region of Colombia, describing the destruction of cacao production in the 1850s by ‘a virulent velvety fungus growth developing to an impalpable dust and attacking the fruit only’ (Holliday, 1953; Parsons, 1949). However, several references (Ancízar, 1956; Anonymous, 1832, 1850) mention a disease on cacao in Colombia's Norte de Santander and Santander Departments, which also matches the symptoms of FPR, as early as 1817.

The fungus was formally named when specimens from Ecuador were sent to R. Ciferri who ‘confirmed’ it as a new species of Monilia: Monilia roreri Cif. (Ciferri and Parodi, 1933). Therefore, the technical terms ‘la Monilia’, ‘la Moniliasis’ and ‘Monilia pod rot’ were adopted to describe the disease. Koch's postulates were completed much later when inoculation techniques were improved (Suárez, 1972). Subsequently, similarities in its biology to the fungus Crinipellis (Marasmius) perniciosa (Stahel) Singer (Agaricales: Basidiomycota), the causal agent of WBD of cacao (Baker and Holiday, 1957), were noted (Evans et al., 1977). Mr pod symptomology is similar to that of WBD up to the point of formation of the white pseudostroma, imparting a frosted appearance on expanding pod lesions (Fig. 1E), whereas the lesions on Moniliophthora perniciosa‐infected pods are more restricted and sunken (Evans, 2016b). Moreover, as the fresh pseudostroma (Fig. 1C) has a strong mushroom‐like odour, the placement of the pathogen in the Ascomycota was questioned. Scanning and transmission electron microscopy have revealed that the pseudostroma mycelia have septa with a complex, barrel‐shaped structure, or dolipore, typically found in basidiomycete hyphae (Fig. 2B), and spores are formed in basipetalous chains (Fig. 2A). Thus, the new genus Moniliophthora—literally, Monilia destroyer—was erected, and the combination Moniliophthora roreri was made, assuming that it represented the asexual morph of an unknown fungus (Evans, 1981; Evans et al., 1978), in the phylum Basidiomycota. The common name ‘frosty pod rot’ (FPR) was also proposed.

Figure 1.

(A) Mature infected pod showing typical ‘frosty’ (helada) appearance with pale‐brown powdery spores produced on a white pseudostroma. (B) Immature pod with minor external symptoms, but showing complete destruction of the watery bean mass internally. (C) Discarded harvested pods amongst which a ‘healthy’ pod had been cut open to reveal internal destruction; several days later, the white pseudostroma had completely colonized the cut surface. (D) Pod with typical frosted appearance releasing a cloud of spores laterally (short arrow), as well as falling heavier spore flakes (long arrow), when the tree was shaken. (E) Pod infected 2–3 months previously, entering the necrotrophic phase with rapid lesion development and the emergence of the pseudostromatal mycelium. (F) Young infected pod showing distortion and lateral swellings.

Figure 2.

(A) Sporogenesis of Moniliophthora roreri. Light microscopy showing chains of spores forming with the oldest at the apex (basipetalous), rather than monilioid, with the oldest at the base (acropetalous); arrow shows detached spores with thickened wall. (B) Transmission electron microscopy (TEM) of vegetative mycelium of the pseudostroma showing the complex dolipore septum diagnostic of a Basidiomycete fungus.

Origin and Current Distribution

Mr can attack fruits belonging to members of the genera Herrania and Theobroma, although the exact nature of these interactions is unclear, as is the breadth of plant species and pathogen diversity involved (Table 1). Although Rorer (1918) considered the fungus to be endemic in western Ecuador on wild Theobroma bicolor, evidence suggests that the pathogen was introduced into Ecuador from Colombia (Phillips‐Mora, 2003). Reports of the disease on indigenous Theobroma gileri in central Colombia (Baker et al., 1954) led Holliday (1971) to propose it as the original host of the fungus, with subsequent spread to cultivated cacao. The origin of the fungus, rather than the cacao pathogen per se, is less clear because of its discovery on T. gileri in the plant's type locality in the submontane forest of the unique Chocó Refugium (Gentry, 1982) of north‐west Ecuador (Evans, 2002). This was assumed to be the origin of the disease outbreaks in coastal Ecuador in the 1900s (Evans, 2002), but taxonomic studies revealed morphological and DNA sequence differences compared with the cacao pathogen, which was supported by the failure of this wild‐type to infect cacao (Evans et al., 2003a).

Table 1.

Hosts of Moniliophthora roreri.

Host species	Country	Reference *
Theobroma cacao L.	Ecuador	Rorer (1918)
T. angustifolium Moçino & Sessé	Ecuador	Evans et al. (1977)
T. bicolor Humb. & Bonpl.	Ecuador	Rorer (1918)
T. gileri Cuatrec. †	Ecuador, Colombia	Evans et al. (2003a, 2003b); Baker et al. (1954)
T. grandiflorum (Willd. ex Spreng.) Schum.	Costa Rica	Phillips‐Mora (2003)
T. mammosum Cuatrec. & León	Ecuador	Evans et al. (1977)
T. simiarum Donn. Sm.	Ecuador	Evans (1981)
T. speciosum Willd. ex Spreng.	Costa Rica	Phillips‐Mora (2003)
T. sylvestre Mart.	Ecuador	Evans (1981)
Herrania albiflora Goudot	Costa Rica	Phillips‐Mora (2003)
H. balaënsis P. Preuss	Ecuador	Rorer (1918)
H. nitida (Poepp.) R.E. Schult.	Ecuador	Evans (1981)
H. pulcherrima Goudot	Ecuador	Evans (1981)
H. purpurea (Pittier) R.E. Schult.	Peru	Ram et al. (2004)

*Earliest published record: most records are from observations in germplasm collections or from inoculation; potentially, all species of the genera Herrania and Theobroma (Malvaceae) are susceptible.

†From type locality of host in Ecuador; earlier record on T. gileri in north‐west Colombia is now thought to be a different, closely related species (B. G. D. Bartley, personal communication, Lisbon, Portugal; Evans et al., 2013).

The pathogen spread into western Venezuela, probably in the 1940s (Thorold, 1975), but has not reached the cacao region in the east. The fungus entered Panama in the 1950s (Orellana, 1956), and Costa Rica in the 1970s (Enríquez and Suárez, 1978). Spread into Mesoamerica has been slow, with reports of its arrival in Guatemala and Honduras in the 1990s (Evans, 2002), Belize and Mexico a decade later (Phillips‐Mora et al., 2006a, 2006b) and, most recently, Jamaica (IPPC, 2016). The pathogen moved eastward, breached the Andes and, by the mid‐1980s, was established (Evans, 1986). It was reported in northern Peru in 1988, the southern valleys by 1995 (Evans et al., 1998) and arrived in the Cuzco region soon after (Evans, 2002); more recently, it has appeared in the Alto Beni region of Bolivia (Phillips‐Mora et al., 2015).

Taxonomy

The taxonomy of Mr has had a chequered past because of the absence of a sexual fruiting body, and this continues into the present. The decision by Aime and Phillips‐Mora (2005) to move Crinipellis perniciosa into the genus Moniliophthora (Moniliophthora perniciosa, Mp), representing a new Marasmiaceae lineage within the Agaricales (Basidiomycota), created problems as the original generic diagnosis was based on the purported asexual morph of a hyphomycete fungus (Evans et al., 1978). Evans et al. (2013) amended the description to include the agaric sexual morph or basidiocarp of Mp. In the interim, other mushroom‐forming species, with saprotrophic and, in some cases, endophytic lifestyles, rather than a pathogenic one, were added to the genus (Kerekes and Desjardin, 2009; Krupp and Albee‐Scott, 2012). However, in the most recent treatment of the genus (Seifert et al., 2011), Moniliophthora is classified in the Hyphomycetes, within fungi that form chains of thallic aseptate conidia (ameroconidia) involving the breaking up of fertile hyphae at the septa (arthric conidiogenesis). Díaz‐Valderrama and Aime (2016b) also interpreted sporogenesis as being thallic and described the spores as rhexolytic conidia. In contrast, Evans et al. (2013) considered the whole pod surface structure (pseudostroma, sporogenous cells) to be a grossly modified basidiocarp, based on circumstantial evidence of meiotic events during sporogenesis, using the terms basidia and meiospores to reflect this interpretation, and sporogenesis considered to be holoblastic, rather than arthric‐thallic. The following descriptions are an abridged form of those in previous publications (Evans, 1981, 2016a; Evans et al., 1978, 2013), specifically modified to reflect the uncertainty of the sporulation process in Mr (Díaz‐Valderrama and Aime, 2016b; Seifert et al., 2011).

Moniliophthora H.C. Evans, Stalpers, Samson & Benny in Can. J. Bot. 56: 2530, (1978) emend. H.C. Evans, J.L. Bezerra & R.W. Barreto in Plant. Pathol. 62: 734 (2013). Basidiomycota (Phylum), Agaricomycetes (Class), Agaricales (Order), Marasmiaceae (Family).

Mycelium of two types: intercellular, swollen, convoluted, lacking clamp connections, monokaryotic; intracellular, narrow, straight, usually with clamp connections, dikaryotic. Basidiomata either agaricoid or pseudostromatal; on living or dead host tissues. If agaricoid, pileus small, convex; pileal surface an open network of thin‐walled hyphae encrusted with pigment; pileal hairs or setae short, strigose, typically crowded at centre and with membrane pigment; stipe short, cylindrical, fleshy, with bulbous base. Gills distant, thin, white, fleshy; basidia clavate, four‐spored; basidiospores ellipsoid, hyaline, thin‐walled, inamyloid; spore print white; cheilocystidia clavate to fusoid, hyaline, thin‐walled.

In non‐agaricoid forms, spores produced in chains on a white pseudostroma; spores globose to subglobose, sub‐hyaline, thick‐walled and pale brown with age, powdery.

Type species: Moniliophthora roreri (Cif.) H.C. Evans et al. (1978)

Synonyms: Crinipellis roreri (Cif.) H.C. Evans in Evans et al., Mycologist 16: 151 (2002)

• Monilia roreri Cif. in Cifferi and Parodi, Phytopath. Z. 6: 542 (1933)

Below is a modified form of the original species description by Evans et al. (1978).

In vivo: snow‐white pseudostroma, composed of hyaline, thick‐walled, skeletoid hyphae, on necrosing cacao pods (Fig. 1E), or on cut surfaces (Fig. 1C); rapidly turning cream to pale brown, and tough, as the spore mass develops (Fig. 1A); hyphae lacking clamp connections but with dolipore septa (Fig. 2B). Spores holoblastic* produced basipetally in chains (Fig. 2A), following the swelling of a fertile hyphal extension and the development of the first holoblastic spore, with successive swellings below as further spores are differentiated, 4–10 spores per chain; typically, spores round‐off with age, becoming globose to sub‐globose, (6.5–)8–15 µm, thick‐walled (up to 2 µm), sub‐hyaline to brown; occasionally, ellipsoidal, 8–20 × 5–14 µm); liberated by fracture of the cell wall, remnants of which may adhere to the spore.

In vitro: growth on agar (malt extract) slow, 8–15 mm diameter after 2 weeks at 25 ºC; margin appressed, even. Mycelium felt‐like; colonies pale ochraceous pink, at first, becoming apricot buff to cinnamon brown. Odour indistinct; tetragonal crystals formed within agar; laccase and tyrosinase present.

*As noted previously, Seifert et al. (2011) and Díaz‐Valderrama and Aime (2016b) interpret sporogenesis as being thallic and not blastic, and describe the spores as amerospores or rhexolytic conidia.

The fungus found on wild T. gileri in north‐west Ecuador was assigned as variety gileri H.C. Evans (Evans et al., 2013), having larger, more ellipsoidal, thinner walled spores than the cacao type, and with a different DNA profile (Evans et al., 2003a), as confirmed by Phillips‐Mora et al. (2007). More recently, Díaz‐Valderrama and Aime (2016a, 2016b), using a different isolate from T. gileri, found no morphological or molecular differences.

Disease Cycle

The FPR disease cycle appears simple (Fig. 3). Powder‐like spore masses are dislodged from the pseudostroma on the pod by air or tree movement (Fig. 1D), and spread by convection currents. The sporulation density has been estimated at 44 million spores/cm², with a mature pod producing up to seven billion spores (Campuzano, 1976). The fungus is present throughout the year as a component of the air spora in western Ecuador, with the highest frequency during the wet season (Evans, 1981) and a peak at the start of the heavy rains, following the dry season, which may be the old spores released from mummified pods (Evans, 1986). Spores survive for 9 months on pods in the canopy, but approximately 1 month on harvested pods left on the soil.

Figure 3.

Moniliophthora roreri disease cycle on Theobroma cacao. Spores produced on necrotic pods infect new pods, initiating a prolonged biotrophic phase (45–90 days). Once the necrotrophic phase is initiated, new spores are rapidly produced on pods and spread by wind and rain. Remnant spores on mummified pods can initiate new epidemics in association with new production seasons.

Germinated spores penetrate pods via the cuticle or stomata and colonize tissues intercellularly (Suárez, 1972). This biotrophic mycelium is considered to be monokaryotic (Griffith et al., 2003), with a similar morphology to that of Mp (Evans, 2016b), having a convoluted, grossly swollen form, with septa. This can result in malformed pods (Evans, 2016a). Malformation depends on pod age at infection, pods less than 1 month old being most susceptible, as well as on the cacao variety: the younger the pod at infection, the greater the effect on symptom expression externally (Fig 1F) and internally (Fig. 1B). The biotrophic phase (BTP) lasts 45–90 days, after which the fungus switches to the necrotrophic phase (NTP), typified by a thin‐walled, narrow, intracellular mycelium resulting in tissue death within days. The expanding chocolate‐brown lesions coalesce, typically covering the entire pod surface, followed by the white pseudostroma (Fig. 1E) and sporulation (Fig. 1A).

In planta, the biotrophic mycelia can be much thicker than the necrotrophic mycelia (Fig. 4A) (Bailey et al., 2013) and resemble Mp biotrophic mycelia (Evans, 1980; Griffith and Hedger, 1994; Meinhardt et al., 2006). Thinner mycelia are produced in necrotic tissues (Fig. 4B) (Bailey et al., 2013). Clamp connections and dikaryotic mycelia, which easily identify the NTP of Mp (Evans, 2016b), are missing in Mr. The cells in necrotrophic mycelia carry a single nucleus and are considered to be haploid (Díaz‐Valderrama and Aime, 2016b).

Figure 4.

Unique aspects of Moniliophthora roreri growth. Moniliophthora roreri hyphae in infected pods at different stages of disease (Bailey et al., 2013). (A) Biotrophic hyphae from inside malformed green pods of cacao clone Matina. (B) Profuse thin necrotrophic hyphae inside sporulating pods. Pods were dissected and stained with lactophenol cotton blue for 1 h. Sections were squash mounted onto slides and photographed using phase contrast at 1000× magnification.

Mr spores are of variable shapes and sizes and carry varying numbers of nuclei, two being most common (Díaz‐Valderrama and Aime, 2016b; Evans et al., 2002). Mr spores were initially considered to be asexually produced conidia (Evans et al., 1978). Evans et al. (2002) later found evidence of a modified meiosis. In a recent finding, Díaz‐Valderrama and Aime (2016b) reported that spore production by Mr was mitotic in origin. The interpretations of Evans et al. (2002) were based on nuclear division occurring in spores after cell division, as opposed to the random formation of walls around pre‐divided nuclei (Díaz‐Valderrama and Aime, 2016b).

Genetic Diversity

Phillips‐Mora et al. (2007) reported the greatest genetic diversity in Colombia, suggesting it as the centre of origin of Mr. Phillips‐Mora et al. (2007) identified five genetic groupings across its range: the Bolivar group in Colombia, Peru, Venezuela and Ecuador; the Co‐West group in Colombia, Ecuador and Central America; and three additional groups endemic to Colombia and north‐western Ecuador. These groups were split into two broader groups: Orientalis and Occidentalis. Using microsatellite markers, Melo et al. (2014) observed limited Mr genetic diversity with 3.2 alleles per locus and a heterozygosity of 0.03 amongst Mr isolates from Ecuador. Single nucleotide polymorphism (SNP) marker analysis showed that Mr isolates collected from throughout its range could be placed into two major groups (groups a and b), and predicted the upper Magdalena Valley of Colombia as the centre of origin (Ali et al., 2015). Group b had a wide geographical distribution throughout South and Central America. Group a was confined to Colombia and Ecuador. Jaimes et al. (2016) identified three major genetic groups amongst the Mr isolates from the Magdalena Valley using 23 short sequence repeat (SSR) markers. Mr isolates successful in dissemination showed greater genetic diversity and formed outgroups (Ali et al., 2015), suggesting anthropogenic dispersal from the Magdalena Valley and subsequent localized genetic drift.

All relevant studies using molecular markers indicate that Mr propagates clonally (Ali et al., 2015; Jaimes et al., 2016; Melo et al., 2014; Phillips‐Mora et al., 2007). Yet, low genetic diversity was also observed in Mp, a species that undergoes meiosis, using microsatellite markers (2.9 alleles per locus and 0.15 heterozygosity) (Gramacho et al., 2007; Silva et al., 2008) and retrotransposon‐based markers (diversity of 0.06) (Santana et al., 2012), and was attributed to homothallism. Ali et al. (2015) tested 88 SNP markers and observed no heterozygosity within a large population from close geographical areas, such as Colombia. Mp (biotype C) evolved to be self‐compatible (homothallic) despite the presence of A and B mating type genes (Kües and Navarro‐González, 2010). Díaz‐Valderrama and Aime (2016a) identified A and B mating type genes in Mr. Only one mating type was found in Central America, demonstrating their lack of critical importance in Mr spread (Díaz‐Valderrama and Aime, 2016a). Clearly, if meiosis is to be considered as important in Mr biology, additional evidence is required.

The Mr/Mp Genome Comparison

The Mr genome (52.3 Mbp) is larger than the Mp genome (44.6 Mbp) (Meinhardt et al., 2014). Mr has 912 additional coding sequences (CDSs) compared with Mp (17 920 CDSs for Mr compared with 17 008 for Mp). The average GC% for the Mr genome was 46.88%, slightly lower than that of the Mp genome (47.81%). This 1% difference was not maintained at the CDS level, where both Mr and Mp had nearly identical GC percentages (49.6% and 49.8%, respectively). Predicted proteins were aligned comparing the Mr and Mp transcriptomes (Meinhardt et al., 2014). This revealed that 16 713 CDSs (93%) from Mr have sequence similarity to Mp CDSs, and 15 674 CDSs (92%) from Mp have sequence similarity to Mr CDSs. Long continuous contigs of Mr and Mp were found to be closely related at the nucleotide level with similar gene order; however, there were regions in which potential inversions have occurred.

Horizontal gene transfer (HGT)

Three genes appear to have been obtained by the Moniliophthora genus through HGT (Tiburcio et al., 2010). These genes encode a putative metallo‐dependent hydrolase (MDH), a putative mannitol phosphate dehydrogenase (MPDH) and a family of necrosis and ethylene‐inducing peptide 1‐like proteins (NLPs). Suggestions are that the Moniliophthora genus acquired the NLPs from Oomycetes, MDH from Actinobacteria and MPDH from Firmicutes (Wolf et al., 2004). MDH‐related proteins comprise a superfamily of enzymes with varying functions and the potential roles of MDH in the Moniliophthora are unclear. Fungi can synthesize mannitol, using it as a carbon source and possible reactive oxygen quencher (Vélëz et al., 2008; Williamson et al., 2002).

Genes involved in reproduction

Mr isolates have a homogeneous genetic background, suggesting limited anastomoses resulting in a clonal distribution (Ali et al., 2015; Díaz‐Valderrama and Aime, 2016a; Jaimes et al., 2016). Mr carries at least 10 genes within its genome encoding for meiosis or meiotic specific proteins (Meinhardt et al., 2014). Multiple Mr genes may be involved in heterokaryon incompatibility. For example, the het‐c and het‐e proteins (Saupe et al., 2000) may prevent vegetative heterokaryon formation. Sixty‐five NWD2 genes were found in the Mr genome. NWD2 genes are similar to STAND/prion‐like proteins that function in non‐self cell death recognition (Daskalov et al., 2012) and may be involved in the het‐s vegetative incompatibility mechanism (Paoletti and Clave, 2007). Such proteins are integral components of the self/non‐self recognition system that preserves genetic identity, preventing heterokaryon formation through programmed cell death (Van der Nest et al., 2014).

Retrotransposons

Large numbers of long terminal repeat (LTR) retrotransposons were found in the Mr and Mp genomes, representing about 7.0% or 3.72 Mbp in the Mr genome (Meinhardt et al., 2014), compared with about 1.0% or 0.478 Mbp in the Mp genome (Mondego et al., 2008). In Mr, there are genes encoding proteins for 164 (155 kDa) retrotransposable elements, 21 retrotransposon nucleocapsids, 13 ribonuclease H reverse transcriptases and 212 reverse transcriptase‐RNase H‐integrases. The expansion of LTRs in Mr may have provided the allelic divergence necessary for speciation (Giraud et al., 2008).

Mitochondrial (mt) genomes

The mtDNA sequence from Mr is 93 722 bp (Costa et al., 2012), which is smaller than the 109 103 bp mtDNA of Mp (Formighieri et al., 2008). Gene order for the Moniliophthora was conserved. Twenty‐eight RNA genes were identified in Mr mtDNA, including two genes for the large and small ribosomal RNAs and 26 tRNAs, including all 20 amino acids. The 26 tRNA genes are grouped into seven clusters composed of two, seven, three, two, five, two and four genes, and only tRNA‐Tyr2 is outside a cluster. This clustering is identical to the Mp arrangement, except for the location and counterclockwise orientation of tRNA‐Tyr2 in Mr. Both Moniliophthora mtDNA genomes have a similar codon usage, except for a plasmid insertion in Mp. Three Mr mtDNA contigs were identified as free linear invertron‐like plasmids (Sakaguchi, 1990), one matching the Mp inserted plasmid.

Plant–Pathogen Interactions: The Biotrophic/Necrotrophic Shift

The initial transcriptional profiling experiment with Mr considered both plant and pathogen sides of the biotrophic/necrotrophic shift in a susceptible interaction (Bailey et al., 2013). Changes in pod gene expression were obvious at 30 days post‐inoculation (dpi) (late BTP), increasing in intensity at 60 dpi (NTP). Genes associated with plant hormone biosynthesis and action and stress/defence response genes showed altered expression. The pathogen caused progressive changes in pod gene expression from 7 to 30 dpi and a decrease in pod metabolites, notably asparagine and glucose. By 60 dpi, most basic pod metabolites were depleted and, in their place, stress/pathogen‐associated metabolites accumulated [γ‐aminobutyric acid (GABA), succinate and mannitol].

There is no evidence that Mr can form specialized structures, such as haustoria, to achieve direct contact with the plant cell plasmalemma during BTP (Perfect and Green, 2001). The available nutrients are found within the intercellular matrix and the cell wall, and include complex carbohydrates, such as pectin, cellulose and lignin, in addition to other dilute metabolites (Meinhardt et al., 2014). A shift in fungal gene expression (File S1, see Supporting Information) occurred between the BTP and NTP (Bailey et al., 2013; Meinhardt et al., 2014). Although more genes showed preferential expression during NTP, induction during BTP included more highly expressed genes (File S1) (Meinhardt et al., 2014). The release of nutrients occurring with tissue necrosis provides for accelerated pathogen growth, supporting a dramatic shift in Mr cell morphology. The glyoxylate cycle was activated in the NTP (Bailey et al., 2013), supporting the use of simple carbon sources to generate energy and synthesize complex compounds (Lorenz and Fink, 2001).

Plant–pathogen Interactions: Aberrant Forms

Cherelle wilt

Cherelle wilt is a physiological condition resulting in the abortion of young pods (Melnick et al., 2013). Sometimes Mr infections and cherelle wilt overlap. Mr infection alone caused little alteration in cherelle physiology during BTP (Melnick et al., 2013), other than tissue swelling. When Mr‐infected cherelles reached late‐stage wilt, Mr responded by altering the expression of Mr genes associated with NTP (Bailey et al., 2013). Infected wilted cherelles did not produce Mr spores, but were instead colonized by saprophytic microbes.

Disease development in tolerant clones

A recent study of the Mr–cacao interaction (Ali et al., 2014; Bailey et al., 2014) included three tolerant clones (UF‐273, CATIE‐R7 and CATIE‐R4) and two susceptible clones (Pound‐7 and CATIE‐1000). Tolerant clones sometimes become infected under field conditions (over 11 years, 9%–14% loss occurred in the tolerant clones) (Phillips‐Mora et al., 2013). Using naturally infected field material, differentially expressed genes from both plant and pathogen were identified by RNA‐Seq analysis and further assessed by reverse transcription‐quantitative polymerase chain reaction (RT‐qPCR).

RNA‐Seq analysis identified 3009 plant transcripts showing differential expression indicating shifts in 152 metabolic pathways (Ali et al., 2014). For many genes, regression analysis indicated an ordered relationship between gene expression, disease progression and fungal loads in the susceptible clones coinciding with the BTP/NTP shift. The trend was for genes induced/repressed during the BTP/NTP shift in susceptible interactions to be induced/repressed early at low pathogen loads prior to the onset of necrosis in tolerant clones.

RNA‐Seq analysis identified 873 differentially expressed Mr genes with the primary difference being whether the clone was tolerant/susceptible (Bailey et al., 2014). Mr genes associated with stress metabolism and responses to heat shock and anoxia were induced early in tolerant clones, resembling susceptible interactions during NTP. Sporulation was suppressed on pods of tolerant clones, but did occur, providing a platform for selection and adaptation (Bailey et al., 2014).

Plant–Pathogen Interactions: Genes and Gene Families

Mp has 16 329 (Mondego et al., 2008)/17008 (Meinhardt et al., 2014) CDSs, 13 529 of which were considered to be transcribed. Mr has 17 920 CDSs (Meinhardt et al., 2014). If the results from the two Mr RNA‐Seq studies are combined (File S1), 14 740 Mr CDSs show at least one read. There are strong similarities between the disease processes of Mr and Mp (Evans, 2016a, 2016b), although Mp can attack many tissues and Mr only attacks pods.

Candidate effectors in Mr and Mp

Effector proteins modify host responses to infection promoting microbe growth (Toruño et al., 2016). Effectors are generally secreted proteins functioning in the apoplast or taken up by host cells, possibly reaching the nucleus. Effectors are often small cysteine‐rich proteins. Teixeira et al. (2014) identified genes encoding 247 secreted proteins in the Mp genome lacking characterized homologues in other organisms, not counting Mr homologies. These genes exhibited a high rate of non‐synonymous substitution (dN/dS) compared with other Mp protein encoding genes, a hallmark of rapid evolution. Six hundred and thirty‐seven secreted proteins of 300 amino acids or less were identified in the Mr transcriptome, with 126 having at least eight cysteines (Bailey et al., 2014). Ninety were differentially expressed between tolerant/susceptible clones, 44 having no putative function and 29 having eight or more cysteines.

NLPs

NLPs are secreted proteins produced by many microbes and can trigger cell death in dicotyledonous plants including cacao (Pemberton and Salmond, 2004). Nep1, the original NLP isolated from Fusarium oxysporum, caused necrotic flecking when applied to cacao leaves and cherelles (Bailey et al., 2005). A gene encoding a secreted NLP (ESK84003.1) was induced during NTP of the Mr disease cycle (Meinhardt et al., 2014). Five members of this NLP gene family were identified in the Mp genome (Zaparoli et al., 2011) with one NLP, MpNEP2, being induced in NTP (Garcia et al., 2007).

Oxalic acid

Oxalic acid (OA) acts as a virulence factor, removing calcium from pectin, supporting cell wall breakdown and the acidification of tissues, exposing them to enzymatic attack (de Oliveira Ceita et al., 2007). Calcium oxalate accumulated in Mp‐infected cacao stems serves as a source of reactive oxygen species (de Oliveira Ceita et al., 2007), but OA accumulation has not been demonstrated in the Mr–cacao interaction. A putative Mp gene coding for oxaloacetate acetylhydrolase, catalysing the hydrolysis of oxaloacetate to oxalate and acetate, was expressed in biotrophic‐like mycelia which occupy the pectin‐filled intercellular middle lamella. In contrast, two putative Mr oxaloacetate acetylhydrolases (ESK97373.1, ESK97374.1) were induced during NTP (File S1). Dias et al. (2011) found a germin‐like oxalate oxidase gene to be induced early in the resistant variety. A close, if not identical, homologue of the germin‐like oxalate oxidase, cacao gene Tc05_g025310, was expressed at more than 400‐fold higher levels in naturally Mr‐infected pods of two tolerant cacao clones (Ali et al., 2014).

Pathogenesis‐related‐1 (PR‐1)‐related genes

PR‐1‐related genes, common markers of induced defence in plants, are also found within the genomes of plant‐pathogenic fungi (Schneiter and Di Pietro, 2013). Eleven PR‐1 gene homologues were identified in the Mp genome (Teixeira et al., 2012) and 12 PR‐1‐related family members were identified in the Mr genome (Meinhardt et al., 2014). PR‐1 proteins carry the CAP (cysteine‐rich secretory proteins, antigen 5 and pathogenesis‐related 1 proteins) domain (Schneiter and Di Pietro, 2013). Depletion of CAP encoding genes from Fusarium oxysporum impaired its virulence (Prados‐Rosales et al., 2012). Evidence suggests that PR‐1‐related proteins function in the binding of lipids. For example, Saccharomyces cerevisiae homologues bind sterols and protect against antimicrobial compounds (Choudhary and Schneiter, 2012). Mp PR‐1 proteins have been shown to bind and secrete lipid compounds (Darwiche et al., 2017). Lipid compounds can function as components of the host immune response.

Cerato‐platanins and related proteins

Cerato‐platanins (CPs) elicit plant defence responses and cell death (Pazzagli et al., 2014). Three Mr CP encoding genes were induced in BTP and four in NTP (Meinhardt et al., 2014). The CP genes induced in BTP were the most highly expressed (File S1). Five genes encoding CP‐like proteins were identified in the Mp genome (Zaparoli et al., 2009). MpCP1 is more highly expressed in biotrophic‐like mycelia than in necrotrophic mycelia. The tridimensional structure of Mp CPs revealed binding sites for chitin fragments (Barsottini et al., 2013). CPs may scavenge chitin fragments released from the fungal cell wall which elicit plant defence through plant receptors (Sánchez‐Vallet et al., 2014), suppressing pathogen recognition.

Chitin deacetylase and a chitin synthase

Chitin and chitosan are components of fungal cell walls (Sánchez‐Vallet et al., 2014). An Mr gene encoding a putative chitin synthase was induced in NTP (Meinhardt et al., 2014), consistent with rapid growth of the fungus during this phase. Chitin deacetylases were also induced in NTP. Chitin deacetylase can modify chitin (El Gueddari et al., 2002), potentially protecting fungal cell walls from plant chitinases. A putative chitin synthase and a chitin deacetylase showed increased expression in tolerant clones (RNA‐Seq) compared with susceptible clones (Bailey et al., 2014). Real‐time RT‐qPCR verified that Mr expressed chitin synthase MrChs early and at high levels in three tolerant clones (Bailey et al., 2014).

The chitin in fungal cell walls is a target for plant defence (Sánchez‐Vallet et al., 2014). Seven of the 10 plant endochitinases showing differential expression trended towards preferential expression in tolerant clones (Ali et al., 2014). The maintenance of chitin and its fragments appears to be critical to the Mr–cacao interaction.

Hydrophobins

Of the approximately 41 genes encoding hydrophobins found in the Mr genome, three are induced in BTP and 10 are induced in NTP (File S1). Hydrophobins can create hydrophobic cell surface layers, are important in morphogenetic processes (Kershaw and Talbot, 1998) and can limit desiccation and protect against chemical and enzymatic attack (Zelena et al., 2013). Hydrophobins may also work as effectors (Wösten, 2001). In animal systems, hydrophobins are known to prevent host recognition of pathogen‐associated molecular patterns, limiting immune responses (Bayry et al., 2012). The expression of Mr genes encoding hydrophobins varied depending on the cacao clone infected (Bailey et al., 2014). In the tolerant clone CATIE‐R4, three hydrophobin genes were induced and two were repressed compared with expression in susceptible clones.

Carbohydrate utilization and cell wall modification

Mr has many gene families encoding enzymes with potential for the breakdown of lignin (Carro et al., 2016; Pollegioni et al., 2015; Syed et al., 2014). Mr has genes putatively encoding 29 laccases, at least 30 secreted aryl‐alcohol oxidases (AAOs), 11 secreted peroxidases and many cytochrome P450 enzymes. Only a few show preferential expression in NTP (File S1). Within the Mr genome (Meinhardt et al., 2014), 288 genes were identified corresponding to at least 40 different glycoside hydrolase (GH) families (Lombard et al., 2013). GHs are important for the breakdown of complex carbohydrates. Although more GH family genes are induced in NTP (46), the focused specific actions of the few genes (14) in BTP suggest a highly regulated interaction.

Disease Management

Disease management practices take advantage of our knowledge of the pathogen's biology to suppress disease. The inoculum source for FPR is the infected pod, the source of spores. Spore movement is affected by air movement and rain, and spore germination and infection require free moisture, a condition promoted by high humidity. Tolerance to FPR in the field is limited, so that a susceptible host is commonly encountered.de la Cruz et al. (2011) carried out a multiyear study in Mexico incorporating most available FPR management practices: field drainage, tree height (reduced to 4 m), shade, diseased pod removal and fungicide applications. Reducing shade from 70% to 50% reduced disease and, when combined with other practices, reduced disease by 90%. Optimization of shade limits humidity (Schroth et al., 2000). A basic requirement for disease management is the ability to uniformly apply practices, which is impossible if tree height and form are not maintained (Hebbar, 2007). The frequency of diseased pod removal is critical for the management of FPR. Soberanis et al. (1999) found that weekly removal of diseased pods significantly reduced FPR in comparison with pod removal every 2 weeks. The trends were the same for WBD and black pod rot (BPR), giving the removal of diseased pods broad application (Soberanis et al., 1999). Fungicides have proven to be ineffective on their own in the management of FPR (Evans, 2016a; Hebbar, 2007; Krauss et al., 2010). In Costa Rica, Krauss et al. (2010) recorded yield loses as a result of FPR above 75%, depending on the year, with monthly applications of the copper fungicide Kocide. Systemic fungicides have also been tested (de la Cruz et al., 2011; Krauss et al., 2010) with some success. Biocontrol strategies have shown promise, but remain experimental (ten Hoopen and Krauss, 2016).

Integrated disease management can be economically sustainable (Díaz‐José et al., 2014; Krauss et al., 2010; Soberanis et al., 1999). Unfortunately, farmer outcomes vary, influenced by individual efficiencies, the cacao germplasm and the local environment. Disease management practices are labour intensive and costly, which can discourage participation, depending on the production level and market price (Hebbar, 2007; Krauss et al., 2010). Spores move on air currents and so it is important that all farmers in a region participate in management practices. Encouraging farmers to strip healthy pods during the off season and to prune trees to a manageable height (de la Cruz et al., 2011) can be difficult. For these reasons, breeding for resistance to FPR continues to take on an important role.

Identification of Cacao Genotypes with Resistance/Tolerance (R/T) to FPR

Many cacao‐producing countries in the Western Hemisphere have programmes selecting for tolerance to FPR. The Tropical Agricultural Research and Higher Education Center (CATIE) Program, started in the 1980s (Sánchez et al., 1988), is seeking to identify and develop FPR R/T germplasm. The Program has exploited the genetic diversity in its International Cacao Collection, one of two gene banks with international status given by the Food and Agriculture Organization and covered by the International Treaty on Plant Genetic Resources for Food and Agriculture (Phillips‐Mora et al., 2013).

The most extensively used R/T clone in the CATIE Genetic Improvement Program is UF‐273. Crosses between R/T clones UF‐273 (a hybrid from the National Genetic Group) and PA‐169 (from the Upper Amazon Marañon Genetic Group) produced a group of highly productive genotypes surpassing the parents’ R/T reactions (Phillips‐Mora et al., 2013). The resulting clones include CATIE‐R6, the most FPR R/T commercial variety known, and CATIE‐R4. CATIE‐R1 is a progeny of UF‐273 and the FPR‐susceptible clone CATIE‐1000. In 2007, the CATIE Program released six clones with desirable properties, including R/T to FPR (Phillips‐Mora et al., 2013). Clones CATIE‐R1, CATIE‐R4, CATIE‐R6, CC‐137, ICS‐95 and PMCT‐58 are available to more than 6000 farmers in Panama, Costa Rica, Nicaragua, Honduras, El Salvador, Guatemala and Belize through clonal gardens, and have recently been introduced into Mexico, Brazil and the USA.

The most widespread sources of R/T to FPR in South America are ICS‐95 and its derivate clone CCN‐51. ICS‐95 is tolerant to seven isolates of Mr representing the pathogen's genetic diversity in Latin America, demonstrating the possibility of selecting genotypes with broad R/T (Phillips‐Mora et al., 2005). Fifteen years of field data collected at CATIE's La Lola farm (28 Millas de Limón, Costa Rica) have indicated that the ICS‐95 FPR reaction varies with climatic conditions and tree age, decreasing in R/T as trees mature. The annual FPR incidence of ICS‐95 ranged from 13% to 51%, and was intermediate to the reactions obtained by the most R/T (CATIE‐R6, 3%–7%) and most susceptible (Pound‐7 and SCA‐6, 38%–99%) clones (W. Phillips‐Mora, unpublished results). CCN‐51, an offspring of the cross ‘IMC‐67 × ICS‐95’, is a productive, R/T and precocious clone widely used in crosses because of its high general and specific combining ability (Boza et al., 2014). In field trials at CATIE's La Lola farm, CCN‐51 had an FPR incidence of 38%. In the same trial, CATIE‐R6 and the most susceptible clone CATIE‐123 had incidences of 2% and 63%, respectively (W. Phillips‐Mora, unpublished results).

R/T to FPR is more common than initially thought (Phillips‐Mora et al., 2009). Two hundred and seventy‐eight clones have been rated as Res (R, 115 clones) or moderately Res (MR, 163 clones) at CATIE (W. Phillips‐Mora, unpublished results). These clones incorporate all 10 genetic groups described by Motamayor et al. (2008): Contamana (series SCA); Curaray (series LCTEEN and TIP); Guyana (series GU, ELP, KER, YAL); Iquitos (series IMC); Marañón (series PA); Nanay (series NA, POUND); Purus (series RB); Amelonado (Series ELP); Nacional (Series Nal), and Criollo (Chuao) (W. Phillips‐Mora, unpublished results). As some of these clones originated in areas in which FPR was not historically present, it can be concluded that non‐FPR‐specific genes are responsible for the R/T observed. None of the clones identified by CATIE as R or MR come from the Mr centre of origin in eastern Colombia (Ali et al., 2015; Phillips‐Mora et al., 2007) making the region a priority for collections (Phillips‐Mora, 2003).

Conclusions

Although additional work is needed (Table 2), there have been significant advances in our understanding of the relationships between Mr and cacao and Mr and Mp in recent years. The genetic relationship between Mr and Mp has been solidified using molecular technologies. The availability of Mr genome/transcriptome sequences (Meinhardt et al., 2014) sets the stage for detailed functional analysis of genes, gene families, biochemical pathways and whole‐genome systems. New studies of Mr can take advantage of studies carried out in Mp, or synergies can be gained through simultaneous studies in both species. The expression profile of Mr when infecting pods appears, in many cases, to vary from that of Mp when infecting stems. Mr primarily, if not exclusively, reproduces clonally (Ali et al., 2015; Díaz‐Valderrama and Aime, 2016a; Phillips‐Mora, 2003), and thus genetic diversity within the species appears to be limited. R/T against Mr is considered to be polygenic and many sources of resistance appear to have been derived in the absence of Mr. The combination of a pathogen with limited genetic diversity and the potential for recombination and a host with multiple sources of R/T under polygenic control holds promise that losses to FPR can be reduced. As the genetic diversity of Mr in many countries is limited, the screening of R/T sources needs to be carried out in the pathogen's centre of origin, requiring international collaboration. A collaborative approach to breeding for R/T is required to reliably protect against FPR in the future.

Table 2.

Moniliophthora roreri topics needing research.

1. If meiosis is limited or non‐existent, how does M. roreri generate genetic variation: mutation?; horizontal gene transfer?; retroelements?; other sources?

2. Re‐examination of the morphology, genetics and pathogenicity of strains of M. roreri from purported centre of origin in Chocó forest refugia on Theobroma and Herrania spp. in: (a) Lita, Prov. Esmeraldas, Ecuador; (b) Villa Arteaga, Dept. Antioquia, Colombia; (c) Middle Magdalena area of Colombia, Dept. Santander and Norte de Santander. *

3. Determination of whether and how the M. roreri single nucleotide polymorphism (SNP) group b isolates have evolved dispersion mechanisms that allow them to rapidly expand into new cacao production areas (Ali et al., 2015). Do SNP group a and b isolates differ in their ability to cause epidemics?

4. Determination of the evolutionary development of M. roreri compared with M. perniciosa. What was the process by which M. roreri lost the production of basidiocarps and suppressed/eliminated meiosis?

5. Development of tools for M. roreri in planta tracking and genetic manipulation: fluorescent markers, RNAi, other methods.

6. Determination of the diversity of resistance against M. roreri in cacao. Is it a qualitative or quantitative trait? Are sources unique?

7. Can we predict the stability of resistance in cacao against M. roreri. How do elite lines perform in the M. roreri centre of origin?

8. What cultural practices are needed when using clones highly resistant to M. roreri? How do we manage other pathogens?

9. Performance of a survey of the natural enemies of M. roreri on wild hosts in Chocó forest refugia to assess the potential for classical biological control. This would offer another dimension for disease management, especially when using disease‐tolerant clones.

*Theobroma cirmolinae, T. gileri, T. hylaeum, T. nemorale and T. stipulatum, species accepted in the International Plant Names Index, are potential wild hosts of M. roreri and are listed as high conservation or threatened status with no representatives in germplasm collections (Santos et al., 2011). These are scattered in Chocó forest refugia along the lower western slopes of the Andes in north‐west Ecuador up to Villa Arteaga in north‐west Colombia (Baker et al., 1954; Cuatrecasas, 1964).

Supporting information

Additional Supporting Information may be found in the online version of this article at the publisher's website:

File S1 Comparisons between Moniliophthora roreri RNA‐Seq data from two independent studies including the biotrophic/necrotrophic shift and reactions in tolerant and susceptible cacao clones.