Lack of host specificity of Colletotrichum spp. isolates associated with anthracnose symptoms on mango in Brazil

Authors

  • A. de Souza,

    Corresponding author
    1. Laboratório de Bioquímica de Microrganismos e Plantas, Departamento de Tecnologia, FCAV/UNESP, Jaboticabal, São Paulo, Brazil
    2. Laboratório de, Fitopatologia, FCAV/UNESP, Jaboticabal, São Paulo, Brazil
    • Faculdade de Ciências Agrárias e Veterinárias (FCAV), UNESP, Jaboticabal, São Paulo, Brazil
    Search for more papers by this author
  • R. C. Delphino Carboni,

    1. Faculdade de Ciências Agrárias e Veterinárias (FCAV), UNESP, Jaboticabal, São Paulo, Brazil
    2. Laboratório de Bioquímica de Microrganismos e Plantas, Departamento de Tecnologia, FCAV/UNESP, Jaboticabal, São Paulo, Brazil
    3. Laboratório de, Fitopatologia, FCAV/UNESP, Jaboticabal, São Paulo, Brazil
    Search for more papers by this author
  • E. Wickert,

    1. Epagri /EEI, Itajaí, Santa Catarina, Brazil
    Search for more papers by this author
  • E. G. de Macedo Lemos,

    1. Faculdade de Ciências Agrárias e Veterinárias (FCAV), UNESP, Jaboticabal, São Paulo, Brazil
    2. Laboratório de Bioquímica de Microrganismos e Plantas, Departamento de Tecnologia, FCAV/UNESP, Jaboticabal, São Paulo, Brazil
    Search for more papers by this author
  • A. de Goes

    1. Faculdade de Ciências Agrárias e Veterinárias (FCAV), UNESP, Jaboticabal, São Paulo, Brazil
    2. Laboratório de, Fitopatologia, FCAV/UNESP, Jaboticabal, São Paulo, Brazil
    Search for more papers by this author

E-mail: andressa.souza@fcav.unesp.br

Abstract

The aim of the present study was to analyse the genetic and pathogenic variability of Colletotrichum spp. isolates from various organs and cultivars of mango with anthracnose symptoms, collected from different municipalities of São Paulo State, Brazil. Colletotrichum gloeosporioides isolates from symptomless citrus leaves and C. acutatum isolates from citrus flowers with post-bloom fruit drop symptoms were included as controls. Sequencing of the ITS region allowed the identification of 183 C. gloeosporioides isolates from mango; only one isolate was identified as C. acutatum. amova analysis of ITS sequences showed larger genetic variability among isolates from the same municipality than among those from different populations. fAFLP markers indicated high levels of genetic variability among the C. gloeosporioides isolates from mango and no correlation between genetic variability and isolate source. Only one C. gloeosporioides mango isolate had the same genotype as the C. gloeosporioides isolates from citrus leaves, as determined by ITS sequencing and fAFLP analysis. Pathogenicity tests revealed that C. gloeosporioides and C. acutatum isolates from either mango or citrus can cause anthracnose symptoms on leaves of mango cvs Palmer and Tommy Atkins and blossom blight symptoms in citrus flowers. These outcomes indicate a lack of host specificity of the Colletotrichum species and suggest the possibility of host migration.

Introduction

Fungi of the genus Colletotrichum have worldwide importance, as they cause diseases on a wide range of economic crops such as avocado, banana, coffee, citrus, guava, papaya, strawberry, passion fruit and chilli (Hyde et al., 2009b; Phoulivong et al., 2010). Because of its importance, Colletotrichum ranks as one of the most studied genera of phytopathogenic fungi (Hyde et al., 2009b). Mango (Mangifera indica) is among the plants that are most commonly affected by Colletotrichum spp., which cause fruit damage and production losses, directly affecting the tropical fruit producers of Brazil. With more than one million tons harvested per year, primarily for export, mango is economically important to Brazil (Agrianual, 2012).

Mango anthracnose is caused by C. gloeosporioides (teleomorph: Glomerella cingulata), but in some cases C. acutatum (teleomorph: Glomerella acutata) has also been reported as a cause of the disease (Freeman et al., 1998; Rivera-Vargas et al., 2006; Jayasinghe & Fernando, 2009; Phoulivong et al., 2010).

The pathogen causes irregularly shaped, black necrotic spots on both sides of mango leaves. In the inflorescences, the symptoms appear in the stalk and in individual flowers. Although the fungus affects many parts of the plant, the most significant losses occur in the ripening and postharvest stages, when black, sunken, necrotic lesions can appear on the peel of the mango fruit (Freeman et al., 1998; Arauz, 2000).

The identification and characterization of Colletotrichum spp. have been carried out according to their morphological characteristics (Photita et al., 2005; Hyde et al., 2009b). However, these criteria are not sufficient for reliable differentiation of Colletotrichum species because of the morphological and phenotypic variations observed in specimens under different environmental conditions. Thus, studies of Colletotrichum populations must apply molecular techniques to identify the species and to demonstrate variation among them (Cai et al., 2009; Hyde et al., 2009b).

Sequencing of the ITS region has been used to identify Colletotrichum species from different hosts around the world (Peres et al., 2002; Afanador-Kafuri et al., 2003; Photita et al., 2005). In addition, molecular markers, such as RAPD, microsatellites, apPCR and AFLP, have been used to demonstrate variation among populations (Afanador-Kafuri et al., 2003; Abang et al., 2006; Nguyen et al., 2009).

Moreover, it is important to establish whether a particular Colletotrichum species is host-specific or has a wide host range (Cai et al., 2009). Host-range studies may also provide useful data for classification and future species control. Koch's postulates are an essential tool that should be used to confirm the pathogenicity of Colletotrichum isolates (Peres et al., 2002; Cai et al., 2009; Nguyen et al., 2009).

The importance of anthracnose on mango in Brazil dictates the need for studies on the identification and evaluation of the genetic diversity of Colletotrichum isolates. The analysis of population structure, for example, can provide useful information about the evolutionary potential of Colletotrichum isolates, as well as aid the development of strategies for pathogen control. Thus, the aims of this work were to (i) identify the species of Colletotrichum associated with anthracnose symptoms on mangoes found in the different municipalities of São Paulo State, Brazil, using ITS region sequences, (ii) analyse the genetic diversity of the isolates using fluorescent amplified fragment length polymorphism (fAFLP) markers, and (iii) verify the pathogenicity of representative isolates of the population.

Materials and methods

Isolate collection

The fungal collection used in this study consisted of 184 isolates collected from leaves, inflorescences and fruits, showing symptoms, taken from various mango cultivars in 14 municipalities of São Paulo State, Brazil (Table 1). In addition, isolates from citrus were included in the analysis as controls. Three isolates were obtained from symptomless citrus leaves, and two isolates from the citrus flowers of plants with post-bloom fruit drop symptoms were also included (Table 1). For the isolations, tissue fragments of c. 5 mm2 were isolated from lesions of mango organs and from symptomless citrus leaves and were then immersed in 70% ethanol for 1 min, followed by immersion in a solution of sodium hypochlorite: sterile water (1:3 v/v) for 2 min. The fragments were then rinsed in sterile water for 30 s, dried on sterile filter paper, placed on potato dextrose agar (PDA) medium and incubated at 27°C with a 12-h photoperiod for 8–10 days. Typical colonies of Colletotrichum spp. were maintained on PDA and stored under sterile mineral oil.

Table 1. Origin of Colletotrichum isolates and codes used in their nomenclature: (municipality).(plant organ).(host + identification number)
Municipalitya (code)Plant organ (code)Host (code)Number of isolates
  1. a

    Municipalities of São Paulo State, Brazil.

Isolates from mango
Araçatuba (AR)Fruit (F)Tommy Atkins (TA)10
Barretos (BR)Fruit (F)Tommy Atkins (TA)5
Fruit (F)Palmer (P)5
Fruit (F)Haden (H)4
Inflorescence (I)Tommy Atkins (TA)3
Bonfim Paulista (BP)Inflorescence (I)Van Dyke (VD)5
Inflorescence (I)Rosa (R)3
Inflorescence (I)Palmer (P)3
Fruit (F)Palmer (P)2
Cândido Rodrigues (CR)Fruit (F)Tommy Atkins (TA)2
Leaf (L)Palmer (P)2
Inflorescence (I)Palmer (P)2
Leaf (L)Haden (H)3
Fruit (F)Haden (H)3
Inflorescence (I)Haden (H)2
General Salgado (GS)Fruit (F)Palmer (P)6
Inflorescence (I)Palmer (P)9
Leaf (L)Palmer (P)2
Jaboticabal (JB)Inflorescence (I)Espada (E)3
Leaf (L)Haden (H)2
Jales (JL)Fruit (F)Palmer (P)4
Fruit (F)Tommy Atkins (TA)3
Jardinópolis (JA)Inflorescence (I)Palmer (P)1
Leaf (L)Palmer (P)3
Fruit (F)Palmer (P)3
Fruit (F)Bourbom (B)4
Inflorescence (I)Tommy Atkins (TA)3
Leaf (L)Tommy Atkins (TA)2
Fruit (F)Tommy Atkins (TA)1
Mogi-Mirim (MG)Inflorescence (I)Tommy Atkins (TA)8
Monte Alto (MA)Leaf (L)Palmer (P)4
Inflorescence (I)Palmer (P)3
Fruit (F)Palmer (P)4
Leaf (L)Tommy Atkins (TA)2
Inflorescence (I)Tommy Atkins (TA)3
Fruit (F)Tommy Atkins (TA)6
Leaf (L)Haden (H)4
Inflorescence (I)Haden (H)2
Fruit (F)Haden (H)3
São José do Rio Preto (SR)Fruit (F)Palmer (P)4
Taquaritinga (TA)Fruit (F)Palmer (P)8
Vista Alegre do Alto (VA)Leaf (L)Haden (H)1
Leaf (L)Haden (H)1
Inflorescence (I)Haden (H)3
Fruit (F)Haden (H)3
Leaf (L)Palmer (P)2
Inflorescence (I)Palmer (P)3
Fruit (F)Palmer (P)3
Leaf (L)Palmer (P)2
Inflorescence (I)Palmer (P)3
Fruit (F)Palmer (P)3
Votuporanga (VT)Fruit (F)Palmer (P)4
Fruit (F)Espada (E)5
   
Isolates from citrus
Jaboticabal (JB)Leaf (L)Valência (VC)3
Itapetininga (IT)Citrus flowers (CF)Valência (VC)2

DNA extraction, sequencing of the ITS1-5.8S-ITS2 region and analysis

The isolates were grown in 40 mL potato dextrose broth at 27°C with a 12-h photoperiod for c. 15 days. Then, the mycelium was separated from the medium by filtration and dried overnight. After maceration of the mycelium with liquid nitrogen, DNA extractions were carried out according to the Kuramae-Izioka (1997) protocol, with modifications.

The primers ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) were used to amplify a fragment of the ITS1-5.8S-ITS2 region (White et al., 1990). The PCR reactions were carried out using 1 ×  buffer (100 mm Tris-HCl pH 8.8; 500 mm KCl), 2 mm MgCl2, 0.2 mm dNTPs, 1.5 U Taq DNA polymerase (Fermentas), 2.5 pmol each primer, 60 ng genomic DNA and nanopure water to a final volume of 20 μL. Reactions were conducted in a PTC-100 Programmable Thermal Controller (MJ Research, Inc.) that was programmed according to the following settings: one cycle at 95°C for 5 min, 39 cycles at 94°C for 1 min, 60°C for 1 min and 72°C for 1.5 min, and one final cycle at 72°C for 10 min. Amplified samples were electrophoresed in a 1.5% agarose gel with 0.5 μg mL1 ethidium bromide and visualized under UV light with GEL DOC 1000 photodocumentation equipment (Bio-Rad).

The amplified DNA fragment was then subjected to PCR sequencing using the Dye Terminator ET DYEnamic Kit (GE Healthcare), according to the manufacturer's instructions. The primers used here were the same as those in the sequencing PCR reactions. The DNA was sequenced using an ABI 3100 sequencer (Applied Biosystems). Sequencing results were evaluated using phred/phrap/consed (Ewing & Green, 1998; Gordon et al., 1998). Stretches of contiguous sequences with bases that showed a continuous quality value equal to or above 20 were accepted and used for further analysis.

DNA sequences were aligned with malign v. 2.7 (Wheeler & Gladstein, 1994). The dendrogram was developed by the distance method using neighbour-joining grouping (Saitou & Nei, 1987) and the Kimura-2-parameter substitution model (Kimura, 1980) using the mega 4.0 (Tamura et al., 2007) software. The reliability of the tree topologies was evaluated by bootstrap analysis using 1000 replications. Some Colletotrichum spp. sequences from GenBank were added in this analysis: EU371022.1 C. gloeosporioides (type specimen) AB219019.1 G. cingulata, AY177330.1 C. gloeosporioides, AF411700.1 G. acutata (type specimen), GU183326.1 G. acutata, DQ003110.1 C. graminicola (type specimen) and GU227819.1 C. dematium (type specimen). Other sequences of ascomycetes were added to constitute an out-group: FSU38277 Fusarium sambucinum and FJ769706.1 Guignardia citricarpa.

Haplotypes were determined using arlequin v. 3.5.1 (Excoffier & Lischer, 2010) software. From these results, representative sequences of the isolates were selected and used to compose the tree. Genetic differentiation (FST) was determined by molecular variance (amova) analysis and its significance was tested based on 1000 permutations using the same software.

Development of fAFLP markers and analyses

The fAFLP markers were developed using the AFLP® Microbial Genome Mapping Kit (Applied Biosystems) according to the manufacturer's protocol. Approximately 500 ng genomic DNA from each isolate was digested with MseI and EcoRI enzymes (Invitrogen). After ligating the adapters onto the DNA fragments using T4 DNA ligase (Invitrogen), preselective and selective amplifications were performed. Selective amplification was carried out using the NED EcoRI-AT/MseI-C, FAM EcoRI-AC/MseI-T, FAM EcoRI-AC/MseI-CA, JOE EcoRI-AG/MseI-G and JOE EcoRI-AA/MseI-CC primer pairs.

Samples were loaded into a capillary of the ABI 3100 sequencer (Applied Biosystems) with a GeneScan-500 (ROX) molecular weight standard. GeneScan v. 3.7 software (Applied Biosystems) was used for data collection. Markers between 35 and 500 bp were transformed into a binary matrix using the genotyper v. 3.7 (Applied Biosystems) software. Then, paup v. 4.0b10 software (Swofford, 2002) was used to generate a distance matrix. The tree was obtained by the distance method using the unweighted pair group with arithmetic mean (UPGMA) algorithm of mega v. 4.0 software, which was also used to calculate the genetic distance between isolates. Average distances between groups of taxa were given by calculating dA = dXY–(dX−dY)/2, where dXY is the average distance between groups X and Y, and dX and dY are the mean within-group distances.

Pathogenicity tests

Based on the molecular analysis, 75 isolates were selected for inoculation onto mango leaves. These isolates were cultivated on PDA and incubated at 25 ± 1°C, under light. After 8 days, 10 mL sterile distilled water containing Tween 20 (0.1%) was added to each dish, and the conidia were suspended using a soft bristle brush. The concentrations were adjusted to 1 × 105 conidia mL−1 using a haemocytometer. Detached leaves from 15-day-old greenhouse-grown mango seedlings were arranged in styrofoam trays lined with filter paper moistened with sterile water. The petiole of each leaf was wrapped in cotton wool and moistened with distilled water to maintain the turgidity of the leaf and high humidity in the environment. One side of the leaf was inoculated and the other half served as a control. A 0.1-mL drop of conidial suspension was placed at each of four points on every leaf. The inoculum of each isolate was placed on five leaves of both the tested mango cultivars, Palmer and Tommy Atkins. After inoculation, the trays were wrapped in transparent plastic bags to keep humidity at c. 95%. This experiment was repeated with 25 isolates to demonstrate the reproducibility of the test.

The results were evaluated according to the presence or absence of typical anthracnose symptoms 5 days after inoculation. All leaves were scanned with a ruler to allow calculation of lesion area (cm2) by the Image-ProPlus v. 4.1 software. The values of lesion area were subjected to an analysis of variance, and mean values of area were compared by the Scott & Knott (1974) test (= 0.05).

As controls, C. gloeosporioides isolates from symptomless citrus leaves, C. acutatum isolates from citrus flowers with post-bloom fruit drop symptoms and three representative isolates from mango were inoculated on Pera sweet orange (Citrus sinensis) flowers. Each isolate was inoculated at three points on four greenhouse-grown citrus seedlings.

Results

Sequencing of the ITS1-5.8S-ITS2 region

Sequencing of the ITS1-5.8S-ITS2 region enabled the species identification of each isolate. The sequences obtained showed 99–100% similarity with Colletotrichum sequences deposited in GenBank. The isolates from mango, as well as those from citrus leaves, were identified as C. gloeosporioides. One isolate from mango and the isolates from citrus flowers were identified as C. acutatum. After sequencing, it was possible to determine whether some isolates shared haplotypes. Therefore, only 13 isolates that were representative of the population were included in the phenogram (Fig. 1). All the sequences obtained in this study were deposited in GenBank (Table S1).

Figure 1.

Neighbour-joining grouping derived from ITS rDNA sequences of representative Colletotrichum isolates from mango and citrus. The isolates were separated according to their species. Group G1 was formed by C. gloeosporioides isolates, while G2 was constituted by C. acutatum. It is possible to note the interaction between isolates obtained from mango and citrus. Asterisks indicate type specimens.

In the phenogram, the G1 group was composed of the C. gloeosporioides isolates, whereas the G2 group was composed of the C. acutatum isolates (Fig. 1). The analysis showed that only one C. gloeosporioides mango isolate (JA.F.B104) shared haplotypes with the isolates from citrus leaves (Fig. 1). These data suggest that although isolates from different hosts are usually genetically different, they can affect different hosts, showing a lack of host specificity.

The results of the amova analysis showed great genetic variability among isolates from the same municipality and between different populations (Table 2). The FST value (0.4491; < 0.0001) indicated high genetic variation among isolates within populations.

Table 2. Analysis of molecular variance (amova) amongst Colletotrichum spp. isolates
Source of variation among isolatesd.f.aSum of squaresVariance componentPercentage variation P
  1. a

    d.f., degrees of freedom.

Among populations1575·4090·393844·91<0·0001
Within populations17484·0540·483155·09<0·0001
Total189159·4630·8769  

Development of fAFLP markers

Selective amplification using five primer pairs provided 920 polymorphic peaks. The dendrogram, comprising all 189 isolates, was divided into two groups, G1 and G2, that were formed by the C. gloeosporioides and the C. acutatum isolates, respectively. The genetic distance (dA) value between these two groups was 0.117.

The C. gloeosporioides group was divided into four major subgroups. G1.1 was divided into four other subgroups containing C. gloeosporioides isolates from mango: G1.1.1, G1.1.2, G1.1.3 and G1.1.4 (Fig. 2). The average genetic distance among isolates from the G1.1.1 subgroup (0.098) was higher than that found among isolates from the G1.1.2 subgroup (0.021), and the genetic distance between these two subgroups was 0.075 (Table 3). It is notable that the genetic distance between G1.1.3 and G1.1.4 (0.160) was higher than that found between the G1 and G2 groups formed by different species of Colletotrichum, showing high variability among isolates belonging to the same species.

Table 3. Average genetic distances within and between groups and subgroups of Colletotrichum isolates
    Between
Within Withinsubgroupsgroup
GroupsgroupSubgroupssubgroup1.1.11.1.21.2.11.2.21.31.42
10.2191.1.10·098      
1.1.20·0210·075     
1.1.30·0560·1720·213    
1.1.40·0690·1900·2470·160   
1.2.10·0810·2240·2730·2550·242  
1.2.20·0260·2500·2950·2800·2760·231 
20·2040·1750·2190·2190·1940·1500·211
Figure 2.

Dendrogram obtained by AFLP analysis showing the details of the G1 group G1.1, divided into four subgroups, composed of Colletotrichum gloeosporioides isolates from mango.

The subgroup G1.2.1, a member of group G1.2, was formed by three C. gloeosporioides isolates from symptomless citrus leaves and one isolate from mango fruit (JA.F.B104) (Fig. 3). These results corroborated those of the ITS sequence analysis. However, there was considerable genetic distance between the JA.F.B104 isolate and those obtained from citrus leaves. In contrast, the G1.2.2 group consisted of 85 C. gloeosporioides isolates from mango that were genetically similar to each other (Fig. 3). The average genetic distance among isolates from this group was 0.026 (Table 3).

Figure 3.

Dendrogram of AFLP analysis showing the details of the G1.2 and G2 groups, formed by isolates of Colletotrichum gloeosporioides and C. acutatum, respectively.

The C. acutatum isolates from mango and from citrus flowers were grouped separately from the C. gloeosporioides isolates, forming the G2 group (Fig. 3). There was considerable genetic distance between the VA.L.H189 isolate from mango and the other two from citrus flowers.

Interestingly, it was observed that isolates were not grouped according to their origin, such as municipality, plant organ or host, although some isolates from the same municipality grouped together. In addition, the genetic distance (dA) between the groups and subgroups ranged from 0.075 to 0.295, with the highest value being observed among isolates belonging to C. gloeosporioides (Table 3). Only the two mango isolates, JA.F.TA110 and TA.F.P179, which were from different municipalities, showed the same polymorphic pattern.

Pathogenicity test

Based on the AFLP marker analysis, 75 isolates were evaluated in pathogenicity tests. All evaluated isolates caused typical anthracnose symptoms on the leaves of both mango cultivars tested, Palmer and Tommy Atkins. Repetitions of this experiment using 25 isolates demonstrated the veracity of the test. The sides of the leaves that served as controls did not show any symptoms. Some isolates were able to provoke extensive lesions on the leaves at 100% of inoculated points, whereas some isolates infected only 5% of the inoculated points.

Nearly all of the isolates provoked extensive, black, round lesions that were visible on both sides of the leaves. However, some isolates caused irregular, necrotic lesions surrounded by black spots, indicating the growth region of the fungus in the leaf tissue. Acervuli and conidia were observed in the necrotic areas of the leaves (Fig. 4).

Figure 4.

Two examples of pathogenicity tests with Colletotrichum gloeosporioides on mango leaves. Lesions caused by isolate VA.I.TA211 were characterized by large, rounded black spots, whereas the necrotic lesions caused by isolate JA.F.B105 were surrounded by black spots, indicating the region of growth of the fungus in the leaf tissue. Acervuli were observed on the necrotic tissues. Leaves of mango cv. Palmer are shown on the left and those of cv. Tommy Atkins on the right.

The C. acutatum isolates from mango and citrus flowers caused symptoms on the leaves of both mango cultivars tested. The symptoms caused by this species were as severe as those caused by the C. gloeosporioides isolates from mango. In the same way, the C. gloeosporioides isolates from citrus leaves caused severe symptoms on mango leaves. These results demonstrated the cross-infection potential of the C. gloeosporioides and C. acutatum isolates.

The C. acutatum isolates from citrus flowers provoked typical post-bloom fruit drop symptoms in citrus flowers, as expected. Nevertheless, both the C. acutatum (VA.L.H189) and C. gloeosporioides mango isolates (JA.F.B105 and VA.I.TA211) were pathogenic to citrus flowers, causing post-bloom fruit drop symptoms in citrus flowers. Similarly, C. gloeosporioides isolates obtained from citrus leaves also caused post-bloom fruit drop symptoms in this study. These results demonstrated a potential lack of host specificity of Colletotrichum spp. in both citrus flowers and mango, although this experiment was performed in citrus flowers only as a control and so was conducted with only a few isolates.

Statistical analysis showed highly significant (99% confidence) differences in the severity of symptoms caused by the isolates evaluated. However, there was no significant effect of cultivar on symptom severity. Statistical analysis also delineated the separation of isolates into four groups ranging from the most virulent to the least virulent. Isolates from the G1.2.2 group, characterized by fAFLP analysis, belonged to the most virulent group, as determined by pathogenicity tests, whereas the other isolates were distributed into subgroups ranging from low to moderate virulence.

Discussion

Analysis of ITS sequences allowed the identification of Colletotrichum spp. isolates, as achieved by other researchers (Peres et al., 2002; Afanador-Kafuri et al., 2003; Photita et al., 2005). Although ITS sequencing was effective in discriminating between the Cgloeosporioides and C. acutatum isolates in this study, the findings reported by Cai et al. (2009) that >86% of the Colletotrichum sequences deposited in GenBank have considerable evolutionary divergence from the C. gloeosporioides type specimen were confirmed. In the present work, the ITS sequences of Colletotrichum type specimens were used for alignment and grouping to avoid errors in species identification, as recommended by Hyde et al. (2009a).

In this study, only one of the 184 Colletotrichum mango isolates was identified as C. acutatum, which demonstrates that this species can cause anthracnose on mango naturally, although it has never been reported to occur on mango in Brazil. Similarly, Jayasinghe & Fernando (2009) also reported for the first time the occurrence of C. acutatum on mango in Sri Lanka. A study conducted in Puerto Rico and Florida reported that 93% of the isolates from mango were identified as C. gloeosporioides, whereas only 5% were C. acutatum, and both species were pathogenic to detached mango leaves (Rivera-Vargas et al., 2006).

The fAFLP markers used were efficient in separating the isolates in this study according to their species. The resultant phenogram showed high levels of genetic variability among isolates within and between populations. In addition, amova analysis of the ITS sequences showed greater genetic variability among isolates from the same municipality than between those different populations. There was no correlation between genetic polymorphisms and the geographical origin of C. gloeosporioides isolates. This indicates that genetically different Colletotrichum strains are widespread. Similar results were observed by Denoyes-Rothan et al. (2003) in their analysis of the genetic diversity of C. gloeosporioides and C. acutatum populations from strawberry.

High genetic variability of C. gloeosporioides has been reported by several authors (Ureña-Padilla et al., 2002; Nguyen et al., 2009). Afanador-Kafuri et al. (2003) found C. gloeosporioides isolates from mango and Passiflora sp. to be heterogeneous according to the molecular methods they used. They assumed that if this population of isolates belonged to a single species, the genetic complexity and heterogeneity could be explained by the occurrence of sexual recombination between isolates, as suggested by Freeman et al. (1998). Furthermore, according to Afanador-Kafuri and co-workers, the diversity observed within the subpopulations could be associated with the adaptation of the isolates to a non-specific, broader range of hosts.

The high variability among the isolates within populations in this study also suggests that sexual recombination may be occurring. Although the sexual stage of C. gloeosporioides, G. cingulata, is not common in nature, it has been observed in some pathosystems, including on yam in Nigeria (Abang et al., 2006) and on strawberry in Florida (Mertely & Legard, 2004). Frequent sexual reproduction will ensure frequent recombination and increased evolutionary adaptability of the pathogen (McDonald & Linde, 2002). There is one report of G. cingulata occurring on mango leaves that were collected and spread as a ground covering in the field during the winter season in Cuba (Gordillo & Hernández, 1980).

Genetic variability amongst Colletotrichum isolates is also related to vegetative compatibility group (VCG). Vegetatively compatible isolates are expected to be more similar to each other, although they can constitute distinct genetic populations (Freeman et al., 1998). A study of C. acutatum isolates from different hosts identified seven VCGs among the isolates, one of which included isolates from different hosts, showing that the parasexual cycle may be an alternative source of genetic variability for this genus (Franco et al., 2011).

Remarkably, the C. gloeosporioides isolates from citrus leaves and one mango isolate, which were genotypically similar, grouped separately from the other mango isolates in both the fAFLP and ITS sequence analyses. Similarly, C. gloeosporioides and C. acutatum isolates from strawberry were grouped separately from isolates obtained from citrus flowers in a study using RAPD and microsatellite markers (Ureña-Padilla et al., 2002). Although there are cases of mango and citrus orchards on the same property in several municipalities of São Paulo State, no significant migration of isolates between hosts was apparent. A study of cross-infection of C. acutatum isolates from fruit crops and leatherleaf fern (Rumohra adiantiformis) in Florida indicated that it is unlikely that a pathogenic isolate would move from one type of host to another and produce an epidemic (MacKenzie et al., 2009).

The results of the pathogenicity tests revealed that the C. gloeosporioides and C. acutatum isolates from both mango and citrus could cause symptoms on leaves of both mango cvs Palmer and Tommy Atkins. Peres et al. (2002) reported that a C. acutatum isolate from strawberries caused symptoms on mango cv. Haden fruits in Brazil by artificial cross-infection. Additionally, C. acutatum isolates from pequi (Caryocar brasiliense) plants were able to infect mango by artificial inoculation in Brazil (Anjos et al., 2002). By contrast, Martínez et al. (2009) reported that C. gloeosporioides and C. acutatum isolates from Tahiti acid limes did not cause symptoms on mango in Colombia. Thus, it appears that the C. gloeosporioides and C. acutatum isolates from citrus obtained in the present study are potentially pathogenic to mango plants under field conditions in Brazil.

Colletotrichum isolates behave differently according to host and field conditions. In vitro infection studies showed that Australian C. acutatum isolates from almond, blueberry, chilli, grape, mango, olive, strawberry and tomato were able to infect grape and also blueberry and strawberry, indicating a lack of host specificity (Whitelaw-Weckert et al., 2007). Likewise, C. acutatum isolates from strawberry were shown to be pathogenic to avocado, guava, papaya, mango and passion fruit (Peres et al., 2002). When C. gloeosporioides isolates from yam were used to inoculate other hosts, moderate symptoms were observed on papaya, avocado and mango (Abang et al., 2006). In the same report, mango isolates caused mostly moderate disease reactions in yam, avocado and mango, but citrus isolates did not infect yam.

In the present study, C. gloeosporioides isolates from mango caused anthracnose symptoms on mango leaves with varying intensity. Similar results were obtained by Rojas-Martínez et al. (2008), suggesting that differences in the virulence of the C. gloeosporioides isolates from mango cv. Haden in Mexico were probably the result of the existence of more than one race or biotype of Colletotrichum.

Even though cv. Tommy Atkins is considered to be resistant to anthracnose (Ploetz, 2010), according to the present results it is as susceptible to the pathogen as cv. Palmer. This could be a result of the laboratory conditions in which the study was conducted, where leaves of both cultivars were inoculated under the same conditions with the same conidial suspension. Palmer is a later variety than Tommy Atkins, remaining exposed to Colletotrichum inoculum pressure in the summer when higher temperatures and humidity favour the production of inoculum by the pathogen, thus explaining the prevalent anthracnose symptoms observed in this variety.

One fAFLP analysis group (G1.2.2) was formed by the C. gloeosporioides isolates that were highly virulent to leaves of both mango cultivars. This group presented low genetic distance among its isolates, which explains the similar pathogenic behaviour between them. Rojas-Martínez et al. (2008) reported that although molecular analysis grouped most isolates with little or moderate virulence together, there was a clear separation of those isolates classified as highly virulent from the other groups. In contrast, Abang et al. (2006) reported weak or no correlation between the virulence phenotype and genetic polymorphism of Colletotrichum isolates.

Interestingly, the C. acutatum isolate and two C. gloeosporioides isolates from mango, and the C. gloeosporioides isolates from citrus, were able to cause blossom blight symptoms on citrus flowers, which resulted in persistent calyces. Until recently, only C. acutatum was considered a causal agent of citrus post-bloom fruit drop. Thus, the results corroborate the findings of Lima et al. (2011) that C. gloeosporioides is a new causal agent of citrus post-bloom fruit drop in Brazil, although it constituted only 17.3% of the population isolated from citrus flowers, which mainly comprised C. acutatum isolates. Therefore, this is the first report that C. gloeosporioides and C. acutatum obtained from mango can also cause post-bloom fruit drop in citrus, reinforcing the lack of host specificity of this species.

In accordance with the results of this study, it is suggested that the existence of genetically different C. gloeosporioides, and even C. acutatum isolates, that affect different mango cultivars and plant organs, may contribute to the differences in the responses of these fungi to fungicides. Moreover, evidence is presented here that the lack of host specificity of these isolates can favour the migration of populations from one host to another, even though this may occur over a long period of time. Thus, it is suggested that future studies of new Colletotrichum spp. populations from mango and citrus should be carried out to verify the occurrence of host migration. These outcomes will allow the establishment of effective phytosanitary barriers.

Acknowledgements

We are grateful for financial support from the FAPESP (2009/04225-1 and 2010/05860-0) and CNPq agencies.

Ancillary