INTRODUCTION
Pepper (
Capsicum annuum L.) is an important vegetable crop cultivated in nearly all countries worldwide (
1). The yearly worldwide pepper production has been estimated at approximately 36.8 million tons (
2). Peppers have been widely used as a food flavoring agent and as medicine for thousands of years. Peppers are good sources of vitamins and minerals; they contain multiple chemical constituents with pharmaceutical properties (
1). A major hindrance in pepper production is anthracnose disease caused by
Colletotrichum species, which are ascomycete fungal pathogens. Anthracnose refers to diseases with the typical symptoms of sunken spots or lesions on infected plant tissues (
3,
4). The disease causes substantial economic losses by reducing the productivity and quality of fruits in affected plants (
5). Several species in the genus
Colletotrichum are known to cause anthracnose in pepper plants (
6–8). Among them,
C.
scovillei, which belongs to the
Colletotrichum acutatum species complex, is considered a dominant pathogen responsible for serious pepper yield losses in countries in tropical and temperate zones, including Brazil, China, Indonesia, Japan, South Korea, Malaysia, and Thailand (
9–15).
C.
scovillei has also been reported to attack other economically important crops, such as mango and banana (
16,
17). Although most anthracnose disease on fruits can cause considerable losses in yield and quality, the molecular mechanism underlying the development of fruit anthracnose has not yet been elucidated, whereas many foliar diseases in a wide range of plants caused by diverse fungal pathogens have been extensively studied at the molecular level (
18–24). Therefore, we have initiated a molecular pathosystem study of
C.
scovillei and pepper fruits to better understand anthracnose disease in fruit.
C. scovillei produces a large number of conidia that serve as a major inoculum. During the disease cycle, conidia adhere to the surfaces of pepper fruits upon hydration; they then produce germ tubes (
7,
25). Appressorium, a specialized infection structure, is differentiated at the tip of the germ tube following the recognition of chemical and physical host signals (
26). Considering that
C.
scovillei infects only the fruit, the signals for appressorium development on fruit are different from the signals for appressorium development on other tissues of host plants. Similar to several
Colletotrichum species,
C.
scovillei is a hemibiotroph, which exhibits an early biotrophic phase and a late necrotrophic phase during host-pathogen interactions (
6,
26–28). At an early stage of appressorium-mediated penetration of
C.
scovillei, a unique feature known as a dendroid structure develops in the cuticle layer of pepper fruit; this does not occur in foliar infections involving other
Colletotrichum pathogens (
4,
29,
30). After colonization of host epidermal cells, the fungus develops the typical sunken anthracnose lesion with mucilaginous acervuli containing a large amount of pinkish conidia, which is important for further infection. Polycyclic infection of the fungus contributes to substantial disease during the growing season, thereby causing serious economic losses. Based on the socioeconomic impact and difficulty involved in disease management of fruit anthracnose in a wide range of crops, we previously elucidate the whole genome of
C.
scovillei to better understand the molecular mechanisms involved in fruit anthracnose, which have been relatively uncharacterized at the molecular level (
8,
31,
32).
Eukaryotic organisms, including fungi, have evolved adaptive genetic responses to a variety of stimuli (
33,
34). Transcription factor (TF)-mediated regulation of gene expression plays an important role in controlling crucial aspects of organism survival and development. Gene expression is elaborately orchestrated in a coordinated system of transcription-mediated signaling, in which TFs bind specific sequence elements to initiate or block transcription (
35–37). Various TF families are defined by unique DNA-binding motifs (
38–40). Although the DNA-binding motifs of TF families are highly conserved across taxa, the members of a particular TF family often have different roles in an organism because of evolutionary divergence (
41–43). The availability of genome sequence data has resulted in the identification of a large number of TFs through cross-species comparison, thus enabling TF family-based characterization in functional genomics (
44). The homeobox TF family is a prominent TF family associated with fungal development and pathogenicity. Members of this family contain a conserved 60-amino-acid DNA-binding motif known as the homeodomain (
45,
46). Since the functional discovery of the homeobox TF family in
Drosophila melanogaster, which is expressed in an organ-specific manner during development (
47,
48), important roles of homeobox TF orthologs have been established in several fungi (
33,
49–52). For example, the rice blast fungal pathogen
Magnaporthe oryzae contains seven homeobox TFs, among which
MoHOX2 and
MoHOX7 have been identified as stage-specific key regulators of conidiation and appressorium development, respectively (
33,
53). Functional roles of homeobox TFs in infection-related development have been demonstrated in
Colletotrichum orbiculare, a cucumber anthracnose pathogen, which is evolutionarily very distant from
C.
scovillei (
54,
55). Unlike
MoHOX7, the homolog
CoHox3 was associated with maturation of appressoria in
C.
orbiculare. CoHOX1 was found to be required for pathogenic development, whereas its homolog
MoHOX5 was dispensable for pathogenicity. These results reflect the functional divergence of homeobox TF family members, which further increases the genetic complexities of fungal pathogens with different lifestyles, in response to a wide variety of environmental signals. However, the HOX TF family in pepper fruit anthracnose
C.
scovillei has not yet been studied.
To systematically study the functions of HOX genes, we first analyzed whole-genome sequences of C. scovillei and isolated 624 putative TFs (based on InterPro annotation), which contain 48 distinct domains. Through detection of the homeobox domain (IPR001356), a total of 10 HOX TFs (CsHOX1 to CsHOX10) were then obtained from putative TFs in C. scovillei. These HOX TFs in the homeobox-like domain superfamily (IPR009057) were evolutionarily analyzed in the Colletotrichum genus and outgroup species to reveal their origins. To further study the functional roles of CsHOX genes in C. scovillei, we then generated deletion mutants for each CsHOX gene, based on homology-dependent gene replacement. Comparative functional analysis of deletion mutants demonstrated that CsHOX genes are associated with stage-specific regulations for disease dissemination and development in C. scovillei. Specifically, CsHOX2, CsHOX7, and CsHOX1 were found to be essential for conidiation, appressorium formation, and suppression of the host defense mechanism for anthracnose development on pepper fruits in the C. scovillei-host pathosystem, respectively. Our findings provide a useful framework for understanding the development of anthracnose disease on fruits caused by Colletotrichum species.
DISCUSSION
TFs play critical roles in development and pathogenicity in organisms, including fungal pathogens. Essential functions of the evolutionarily conserved homeobox TFs have been demonstrated in several eukaryotes (
35,
64). In this study, analysis of TFomes and homeodomain superfamily genes in
C.
scovillei was extended to the
Colletotrichum species complex and selected outgroup species. These analyses revealed that
C.
scovillei has a unique feature within the
Colletotrichum species complex; notably, the fungus exhibited an extremely high number of homeodomain superfamily genes, while
HOX genes were evenly distributed among the species complex (
Fig. 1A). In terms of InterPro domain relationships, the homeobox-like domain superfamily (IPR009057) is one of the largest gene families in the kingdom Fungi, inclusive of the homeobox domain (IPR001356) in
HOX genes (
40). This duplication event was observed in both
C.
scovillei and
S.
sclerotiorum, an outgroup species (
Fig. 1A); this shared feature was attributed to the DNA transposon TcMar-Fot1 (
Fig. 2A). While the expansion and enrichment of DNA transposons in the genome of
S.
sclerotiorum have been reported (
65), the relationships of DNA transposons with the duplication of homeodomain superfamily genes are first reported here for both
S.
sclerotiorum and
C.
scovillei. Unlike tandem clusters of
HOX genes in animals,
HOX genes were found to be scattered throughout the genomes of fungi, including
C.
scovillei (
Fig. 2B) (
66). However, these homeobox-like genes were revealed to be duplicated multicopy genes located in a tandem manner in the
C.
scovillei genome (
Fig. 2B). This is likely evidence for the recent transposon activity that occurred in
C.
scovillei after divergence from
C.
nymphaeae (see
Fig. S1A), which presumably occurred approximately 2 million years ago (
57,
67). Because transposable elements are the major drivers of fungal genome evolution (
68), the association of transposons with homeodomain superfamily genes impacts genome organization and may affect cellular processes related to fungal development and pathogenicity.
Molecular mechanisms that govern anthracnose disease caused by
Colletotrichum species on fruits have not yet been characterized. A large number of conidia are reproduced from phialides of conidiophores in many
Colletotrichum species (
69,
70). However, the genetic basis of the conidiation process remains largely unknown in the
Colletotrichum genus. We found that the deletion of
CsHOX2 caused a complete defect in the production of conidia in
C.
scovillei, whereas the
ΔCshox2 strain developed normal conidiophores (
Fig. 3). Considering that
MoHOX2 (orthologous to
CsHOX2) is essential for
M.
oryzae conidiation in a stage-specific manner (
33), the development of phialides on top of conidiophores, which are not clearly distinguishable, may be normal in the
ΔCshox2 strain. However,
HTF1 genes in
Fusarium graminearum,
F. verticillioides, and
F. oxysporum (all orthologous to
CsHOX2) were revealed to be involved in the formation of clearly differentiated phialides (
50). The
ΔFghtf1,
ΔFvhtf1, and
ΔFohtf1 strains alternatively produce macroconidia directly from hyphae at low frequencies, indicating functional differences in
CsHOX2 from orthologous
HTF1 genes in the three
Fusarium species. Unlike the complete defect of the
ΔCshox2 strain in conidium production, it is notable that the deletions of
CsHOX1,
CsHOX3,
CsHOX4, and
CsHOX5 resulted in the development of abnormally large conidia (
Fig. 4A); in contrast, deletion of
MoHOX4 (orthologous to
CsHOX4) resulted in the production of abnormally small conidia in
M.
oryzae (
33). Notably, 4 of 10
HOX genes are associated with conidial size and morphology in
C.
scovillei. Several lines of evidence suggest that the four genes play distinct roles in conidium development of
C.
scovillei. For example, the
ΔCshox1 strain produced significantly larger conidia (19.9 ± 2.4 μm) compared to those from the
ΔCshox3,
ΔCshox4, and
ΔCshox5 strains (3.4 ± 0.3 μm, 15.0 ± 1.7 μm, and 4.2 ± 0.5 μm) (
Table 1). Abnormal swelling of the conidia and germ tubes was observed in
ΔCshox1 alone (not in the
ΔCshox3,
ΔCshox4, and
ΔCshox5 strains), following treatment with nikkomycin Z; this indicated a difference in cell wall integrity (see
Fig. S4B). The sensitivity of conidia in the
ΔCshox3 strain was much higher following treatment with the fungicide carbendazim, compared to the sensitivities of conidia in the
ΔCshox1,
ΔCshox4, and
ΔCshox5 strains (see
Fig. S4A), which support functional differences among the four
CsHOX genes in
C.
scovillei. The
CsHOX1 gene is a critical regulator of cell wall integrity and conidia morphology, as well as pathogenic development, in
C.
scovillei (
Fig. 4A and
7E; see also
Fig. S4B), which represents a novel function of homeobox TFs in plant-pathogenic fungi.
The
ΔCshox1 strain was not defective in appressorium formation and host penetration on pepper fruits (
Fig. 5C and
D). The invasive hyphae of the
ΔCshox1 strain were unable to grow in the epidermal cells of pepper, following successful penetration into the cuticle layer via appressoria (
Fig. 5C). The accumulation of the brown cloud and ROS only in the first epidermal cell challenged by the
ΔCshox1 mutant (
Fig. 5E and
F), and the significant upregulation of host defense-related genes were indicative of the activation of innate host immunity (
Fig. 6C; see also
Table S2). The increased sensitivity of the
ΔCshox1 mutant to H
2O
2, compared to the wild-type and
Cshox1c strains, indicates that
CsHOX1-regulated genes may be related to the modulation of ROS that are encountered in host pepper cells during anthracnose development of
C.
scovillei (
Fig. 6A and
B). Notably, the ROS-scavenging enzymes CsTRR1 and CsCAT1, orthologous to
M. oryzae TRR1 and
Sclerotinia sclerotiorum SCAT1, respectively, were significantly downregulated in gene expression profiles (see
Table S3C). Both TRR1 and SCAT1 were reported to play important roles in ROS detoxification and plant infection (
60,
61). The significant downregulation of
CsTRR1 and
CsCAT1 in
ΔCshox1-infecting host tissues revealed the involvement of
CsHOX1 in ROS detoxification. Moreover, an LysM protein CsELP2, orthologous to
C. higginsianum ChELP2, were found in the significantly downregulated genes (
Fig. 7B). The ChELP2 encodes an extracellular LysM protein, which suppresses chitin-triggered plant immunity (
71). The fungal chitin is known to induce host ROS accumulation and defense gene activation (
72). This suggests that the CsHOX1 may be involved in suppression of chitin triggered immunity through transcriptional regulation of CsELP2. When performing pathogenicity assay by using host defense-comprised tissues, the
ΔCshox1 could grow invasive hyphae as wild-type strain did on heat-killed host tissues (
Fig. 6D and
E), while the pathogenicity was partially restored on DPI-treated host (
Fig. 6F and
G). This result indicated that the CsHOX1-regulated genes may be involved in suppression of host defense. In addition, the other four genes encoding secreted proteins (CsSIX11, CsCRISP1, CsSP1, and CsSP2) were found to be significantly downregulated in
ΔCshox1 strain-infected host tissues (
Fig. 7B). The CsSIX11 was predicted to be orthologs to a known effector, SIX11 (Secretion In Xylem) of
Fusarium oxysporum, which was recently demonstrated to be dispensable for fungal virulence, while its role in host defense is still unknown (
73). The
CsCRISP1 gene was predicted to encode a cysteine-rich secretory protein. Although the ortholog of CsCRISP1 was not characterized, the cysteine-rich secretory proteins were reported to be involved in suppression of plant immunity in
Verticillium dahliae (
74). The detailed mechanisms of CsSIX11 and CsCRISP1 await investigation. These findings suggested that CsHOX1 is an essential factor that contributes to anthracnose development by transcriptional regulation of genes involved in the modulation of ROS and host defense (
Fig. 7D).
Our study of appressorium-mediated fruit disease development in
C.
scovillei demonstrated that it uses a different strategy, compared to other fungal pathogens, to effectively penetrate the thick cuticle layer (20 to 25 μm) of host cells. For example,
M.
oryzae, a model for foliar disease in rice, generates substantial turgor pressure inside the appressorium to directly reach invading epidermal cells through the host cell wall by means of a strong mechanical force (
75). However,
C.
scovillei has evolved a highly sophisticated strategy in which the fungus achieves tiny pin-point entry via the appressorium and then grows in the thick cuticle layer with the formation of dendroid structures within the cuticle layer; it does not use a penetration peg (2 to 4 μm in length) that directly reaches the epidermal cells of the pepper fruit. Therefore,
C.
scovillei generates, possibly, lower turgor pressure inside the appressorium and maintains conidial viability after penetration, in contrast to the autophagic cell death of conidia that occurs following penetration of
M.
oryzae (
76). This hypothesis is supported by the observation that unpigmented immature appressoria of
C.
scovillei are able to penetrate the cell wall of pepper fruit and cause anthracnose disease (data not shown).
In this study, we demonstrated the dynamics and proliferation of homeodomain superfamily TFs during the evolution of an ingroup of Colletotrichum species and outgroup members. Detailed functional analyses have revealed that members of the homeobox TF family in C. scovillei play a key role in fungal development and anthracnose disease on pepper fruit. Our study provides a fundamental basis for understanding the molecular mechanisms involved in conidiation, appressorium development, and anthracnose disease on fruits, which will contribute to the development of novel strategies for use in managing anthracnose disease on many economically important fruits.