Host-Range Dynamics of Cochliobolus lunatus: From a Biocontrol Agent to a Severe Environmental Threat

We undertook an investigation to advance understanding of the host-range dynamics and biocontrol implications of Cochliobolus lunatus in the past decade. Potato (Solanum tuberosum L) farms were routinely surveyed for brown-to-black leaf spot disease caused by C. lunatus. A biphasic gene data set was assembled and databases were mined for reported hosts of C. lunatus in the last decade. The placement of five virulent strains of C. lunatus causing foliar necrosis of potato was studied with microscopic and phylogenetic tools. Analysis of morphology showed intraspecific variations in stromatic tissues among the virulent strains causing foliar necrosis of potato. A maximum likelihood inference based on GPDH locus separated C. lunatus strains into subclusters and revealed the emergence of unclustered strains. The evolving nutritional requirement of C. lunatus in the last decade is exhibited by the invasion of vertebrates, invertebrates, dicots, and monocots. Our results contribute towards a better understanding of the host-range dynamics of C. lunatus and provide useful implications on the threat posed to the environment when C. lunatus is used as a mycoherbicide.


Introduction
Race specific Cochliobolus species have caused plant disease disaster such as the southern leaf corn blight epidemic of 1970s in the United States of America [1], northern leaf corn blight (Exserohilum turcicum) and corn head smut (Sporisorium reilianum) in northern China in the 1990s [2,3], and the Great Bengal rice famine of India in 1940s [4,5]. In the Great Bengal rice famine, more than 2 million people starved to death due to reduction in rice yield of about 40 to 90% [5]. Cochliobolus species often cause diseases to several plant families including Alliaceae, Anacardiaceae, Araceae, Euphorbiaceae, Fabaceae, Malvaceae, Rutaceae, Zingiberaceae and Solanaceae [6].
C. lunatus has emerged in the last decade as a virulent and destructive pathogen [15,16]. Remarkably, C. lunatus successfully thrives on important crops such as rice (Oryza sativa L.), wheat (Triticum aestivum), cassava (Manihot esculenta), sorghum (Sorghum bicolor), Hymenaches species, strawberry (Fragaria × ananassa), Amaranthus species, and potato [16][17][18][19][20]. Decades after Sivanesan's [14] pioneering study, is C. lunatus solely endemic to the outlined geographical locations? If no, has C. lunatus gained hosts and new geographical zones in the last decade? The aims of this study were (1) to determine the interrelatedness of 5 virulent strains of C. lunatus causing foliar necrosis of potato using morphological descriptors coupled with phylogenetic tools and (2) to establish the current host-range diversity of C. lunatus in the last decade.  [20] were used. Brown-toblack leaf spots were excised and treated with 2% NaClO solution for 2 min and rinsed in sterile water with three changes. The leaf pieces were aseptically plated on V8 agar medium (HiMedia, Mumbai, India) and incubated at 25 ∘ C in dark. Developed colonies after 7 days were morphologically identified based on standard monograph taxonomic keys [7].

2.2.
Host-Range Diversity. The genomic DNA was isolated from fungal isolates grown in potato dextrose broth (PDB) (HiMedia, Mumbai, India). Approximately 100 mg of mycelia mat was disrupted in the presence of TRIreagent (Sigma, St. Louis, MO, USA) using mortar and pestle containing 2 mg/mL proteinase K (Merck, Bangalore, India) following the manufacturer instructions. The quality and quantity of the DNA were determined using a 1% agarose gel electrophoresis and a nanodrop spectrophotometer (BioSpec-nano, Shimadzu, Japan), respectively. For molecular identification, the partial sequence of 5.8S rDNA, complete internal transcribed spacer 2 region (ITS2), and partial 28S rDNA region were amplified as previously described [23]. To distinguish the strains, we designed specific primers (forward: 5 -cgatatgcggcatatgca-3 ; reverse: 5acctacgcattgcggaa-3 ) for glyceraldehyde-3-phosphate dehydrogenase (GPDH) gene using C. lunatus (GenBank accession number Gb X58718) sequence. Amplification of GPDH was performed as follows. The PCR mix contained 11 ng genomic DNA, 5 L Green GoTaq reaction buffer (Promega, Madison, WI, USA), 0.2 mM each of deoxyribonucleoside triphosphate (dNTP), 0.2 M of each primer, and 1.1 U of GoTaq DNA polymerase in a total reaction volume of 25 L in triplicates (PCR conditions: 5 min at 95 ∘ C, 35 cycles of 1 min at 94 ∘ C, 1 min annealing at 53 ∘ C, 2 min for extension at 72 ∘ C, and a final 5 min extension at 72 ∘ C). The quality of the amplicon was checked by performing agarose gel electrophoresis. The PCR products were purified and sequenced. Sequences were assigned to molecular species based on 98-100% sequence similarity threshold in the GenBank with the following accession numbers: JX512810, JX512809, JX907827, JX477595, and JX907828, respectively, for rDNA. GPDH sequences have been submitted in DNA Data Bank of Japan (DDBJ) as accessions AB859034, AB859035, AB859036, AB859037, and AB859038, respectively.
Using GenBank BLAST search tool, a studied set of rDNA sequences deposited in the last decade was collected based on the information associated with the sequences such as GC content, length (>250 bp), and geographic origin of host. Importantly, records with 100% sequence similarity from the same host and geographical coordinates were removed. Unique sequence sets were screened using ElimDupes (available at http://hcv.lanl.gov/content/sequence/ELIMDUPES/elimdupes.html). Sequence alignment was performed using Muscle program [24]. Best substitution model parameters were determined based on Akaike information criterion, corrected (AICc) and Bayesian information criterion (BIC). The evolutionary history was inferred using the maximum likelihood (ML) method, and rooting was performed automatically by saving the generated ML tree in standard Newick format and all the analysis were performed in MEGA 6.06 (updated v. 6140226) software [25]. The strength of the internal branches of the ML tree was statistically tested by performing 1000 bootstrap replications.

Identification of C. lunatus Strains Causing Foliar Necrosis of Potato.
Basically, most Cochliobolus species have curved conidia, a broad rounded apex cell, a distinct swollen central cell, a tapering to narrowly round base cell, and 4-5 distinct septa. The five strains of Cochliobolus causing brown-toblack leaf spot disease of potato produced varied colonies and conidia ( Figure 1) similar to previous studies [6,7]. The isolates visibly produced different growth patterns ( Figure 1). In one isolate Btl26IBSD (DDBJ accession AB859034), brown to whitish mycelium, reddish brown medium, and canoe five-celled conidia without stromatic tissues were observed (Figure 1(a)). In C. lunatus, the stromata are oval or ellipsoidal, 10 to 40 m in diameter, and located beneath the ascomata. Another isolate Btl27IBSD (DDBJ accession AB859035) produced greyish-brown mycelium and cylindrical clavated fived-celled conidia void of stromatic tissue (Figure 1(b)). Isolate Btl28IBSD (DDBJ accession AB859036) profusely produced yellowish pigmented five-celled conidia, with stromatic tissue, variable shapes, and end at one cell with a thin hilium (Figure 1(c)). Isolates Btl29IBSD (DDBJ accession AB859037) and Btl30IBSD (DDBJ accession AB859038) produced greyish-brown cottony mycelium (Figures 1(d) and 1(e)). Noteworthy, isolate Btl30IBSD profusely produced dark pigments, and with each cell of the conidia bearing a distinctive oval stromata of different sizes. The exact role of stromata in pathogenicity is not known. The stromata are enclosed by a ring of melanin-like pigment, may play a role in preventing desiccation of the conidia, conserved the gene-pool, and ensure survival under adverse conditions. As shown (Figure 1), morphological characters revealed significant intraspecific variations. Taxonomic circumscription of Cochliobolus has undergone countless modifications in the last five decades caused by overlapping morphological characters [6,7,14,15]. Furthermore, generic concepts delimiting Bipolaris, Cochliobolus, and Pseudocochliobolus are confused [6,7]. Thus, ITS region of the ribosomal RNA operon was used to accurately determine the taxonomic placement of the fungi. Based on rDNA locus, we confirmed that the five fungi causing brownto-black leaf spot disease of potato (Figures 2(a) and 2(b)) were C. lunatus. In the ML tree, the five strains of C. lunatus causing brown-to-black leaf spot disease of potato clustered (Figure 2(c), (I)), closely related to other GenBank type isolates (Figure 2(c), (II)) and distant from other Cochliobolus species. The nucleotide frequencies were A = 25.00%, T/U = 25.00%, C = 25.00%, and G = 25.00%. The transitiontransversion bias estimated by K2 + I substitution model [21] was 2.41. The overall rate of heterogeneity between taxa was 0.01. As expected, low level single nucleotide polymorphism (SNP, 5.4%) was observed out of a total of 1188 sites at the rDNA locus. The five strains causing foliar necrosis of potato were weakly supported with bootstrap values ≤61%. As previously reported, rDNA locus do not often provide ample resolution that can allow differentiation of cryptic taxa such as Cochliobolus [6,15].
The low bootstrap support (≤61%) generated in rDNA ML tree (Figure 2, (I)) made it difficult to determine whether the five strains of C. lunatus causing brown-to-black spot disease of potato in Burdwan Destrict were identical. It could be that all the strains originated from a common source but colonized in different places following dispersion. This is because C. lunatus abundantly produced conidia that can easily be disseminated by air to distant places. To check if the five isolates were identical or not, we used glyceraldehyde 3-phosphate dehydrogenase (GPDH) locus which had been shown to be effective in resolving Cochliobolus species in phylogenetic inference [6,15]. Partial GPDH locus (Figure 3(a)) was sequenced, as this is one of the house-keeping genes, taken as reference in yeast and fungal systems. Based on sequence alignment for GPDH locus, a total of 340 SNPs out of 708 sites and 325 sites without polymorphism (45.9%) were found. Based on TN93 + G + I substitution model [22], the rate of base transition-transversion was 4.96 and the nucleotide frequencies were A = 23.73%, T/U = 18.55%, C = 33.52%, and G = 24.20% and the overall heterogeneity among taxa was 0.316. The ML tree based on GPDH locus discriminated the five strains of C. lunatus causing foliar necrosis of potato with strong bootstrap support ≥81% (Figure 3(b), (IV)). The overall mean evolutionary distance of 0.03 was observed between the five strains causing foliar necrosis of potato (Figure 3(b), (IV)) relative to other C. lunatus type isolates (Figure 3(b), (I), (II), and (III)).
Importantly, because the five strains of C. lunatus clustered based on GPDH locus (Figures 2(b) and 3(b)), this indicated they were closely related as also revealed on the basis of morphological descriptors (Figure 1). Additionally, bootstrap values were <100% for internal branches within the subcluster I, Figure 3(b). This indicated that the five strains which caused foliar necrosis of potato were different. Although the five C. lunatus strains might have adapted in potato for their nutritional requirements in the same geoclimatic zone, it was not possible to determine their origin. Importantly, it is shown that pathogenic fungi are capable of adapting to the genetic background of their host, thus forming new physiological and virulent races [26]. This is generally a slow progressive process determined mainly by the degree of the pathogen-host specific interactions [27].   Collectively, because of some phenotypical variations such as colonies growth pattern, presence or absence of stromatic tissues, colours of conidia and colonies (Figure 1), and strong bootstrap support (>81%) for clustered and unclustered strains (Figure 3(b), (I), (II), and (III)), C. lunatus strains have evolved divergently.
This study seeks to advance insights on the host-range diversity allowing the dynamic movement of C. lunatus observed in the last decade (Tables 1 and 2). From Table 1, it is understood that C. lunatus exploit two kingdoms, notably plant and animal, switching among monocots, dicots, invertebrates, and vertebrates. The paradigm-shift from a plant colonizer sensu stricto to a vertebrate and invertebrate invader (Tables 1 and 2) indicates that C. lunatus have acquired special strategies to switch hosts. The question arises as to why C. lunatus display extensive host-range diversity in a given biota.
Although studies have shed light on specific aspects of C. lunatus pathogenicity such as induce-virulence variations on maize crop [41], virulence differentiation on maize crop [42], secretome weaponries on potato crop [43], and heatdependent virulence on Lolium spp. [44], the nutritional evolution of C. lunatus is unresolved. Intriguingly, host shifting dynamics is not well understood and it has been argued that (1) close proximity to host is prerequisite for pathogens to jump from a natural host to a new host [45], (2) the future host must act as the substrate [15,45], and (3) compatible factors promoting infection must be present [45,46]. Importantly, most host-switching pathogens selfprotect themselves by producing high level pigment such as melanin to deal with the host defense [42,47,48]. Additionally, C. lunatus profusely produced melanized colonizing hyphae during invasion in potato [43] and non-host-specific toxin such as methyl-5-[hydroxymethylfuran-2-carboxylate] in maize [49], to suppress the host defense.
Nonetheless, the above-mentioned factors seem more likely to be limited in explaining how C. lunatus gain access to hosts and not how C. lunatus spreads in a given biota and prevails as an environmental hazard. C. lunatus had extensively been used as mycoherbicide formulations in the past decade [8][9][10][11][12][13]. Remarkably, Zhang et al. [12] fused the protoplast of Helminthosporium gramineum and C. lunatus, to generate a strain with high potential to produce conidia, phytotoxin ophiobolin, and improved potential to control rice weed. Introduction of genetically manipulated strains and unmodified strains of C. lunatus could have hazardous implications to the environment. This is because, in some cases, C. lunatus failed to control the targeted weeds but caused severe damages in economically important crops in the same biota. For instance, C. lunatus isolated from barnyardgrass and used as mycoherbicide failed to control competitive weeds in rice fields but severely damaged bean varieties [8]. Nevertheless, the effectiveness of a biocontrol in the fields depends on the environmental conditions of a given biota, especially humidity and temperature [10]. C. lunatus exhibits a temperature-dependent virulence [43,44] and its introduction into the environment without a precise prediction of the geoclimatic conditions, that is, humidity and temperature, can prove harmful; consequently, it disequilibrates the interaction dynamics of the organisms dwelling in the same biota.
Owing to the divergence in evolution (Figure 1, (I) and (II)) and the emergence of unclustered strains (Figure 2, (I) and (II)), it is clear that strains of C. lunatus have coevolved with their different hosts translated by their different placement in ML tree inference and speciation in their nutritional requirements (Tables 1 and 2). In keeping with the results of the evolutionary disparity, the global control of C. lunatus diseases would require tremendous exertion. This is because in an intermixed network of host-groups, C. lunatus strains from different hosts, genetically distant ((IV) versus (I), (II), and (III), Figure 2(b)) and found in the same geographic zone, would readily invade putative hosts regardless of their temporal host-groups. For instance in the last three years in India, C. lunatus have invaded strawberry [19], Mimusops elengi [37], Amaranthus spinosus [36], Grewia optiva [33], Clerodendrum indicum [40], and potato [20,48]. Noteworthy, these hosts were spatially and temporally distant. With this illustration, it is clear that the host-pathogen proximity hypothesis and host relatedness hypothesis, where a given pathogen switches to new species closely related to the original host, might all apply for C. lunatus.

Conclusions
From an evolutionary viewpoint, the variations observed in C. lunatus colony, conidia size, conidia colour, conidia texture, and the presence or absence of stromata should be regarded as prominent acquired adaptational traits. These characteristic traits were not consistent between the five strains causing foliar necrosis of potato but provided indicators for generic circumscription. Phenotypic intraspecific variations can obscure placement of Cochliobolus species and make correlation to phylogeny difficult. As shown, C. lunatus have considerable ecological and economic importance being a highly successful colonizer in monocots, dicots, vertebrates, and invertebrates. On this basis, the purpose as a biocontrol agent is overshadowed by its virulent and indiscriminate destructive potential in the ecosystem. For this reason, we suggest that the use of C. lunatus as mycoherbicide should be stopped.