Soursop (Annona muricata L.), also known as guanábana in Spanish, is an important tropical fruit crop that originated in the Neotropics. It is cultivated in Brazil, Colombia, Mexico, Panama, Peru, Puerto Rico and Venezuela, and is also grown in small family orchards in Southeast Asia, Philippines, India, and Hawaii (USA) or other Pacific islands (Love and Paull 2011). As a commercial crop, it is highly remunerative for both small- and medium-scale farmers (ICUC 2002). The fruit ranks as among the world’s best-tasting horticultural species, possessing a sweet, creamy flesh and fragrant flavour (Pareek et al. 2011). It is widely consumed in the tropics of America and Asia (Love and Paull 2011).

Both the vegetative and reproductive parts of soursop trees are attacked by a variety of diseases, including the economically significant anthracnose. Characteristic symptoms include dieback of twigs and branches, necrosis of young leaves and stems, flower drop, fruit drop (especially of young fruits), and rot of mature fruit. Crop losses can be as high as 90 % (Álvarez et al. 2004).

Determination of Colletotrichum species and their characterization are traditionally based on microscopic characters, colony morphology, and host-range (Sutton 1992). However, high variability of morphological and cultural characteristics has rendered these criteria unreliable for identifying the pathogen at species level. Furthermore, several Colletotrichum species can be found infecting or colonizing the same host plant (Freeman et al. 1998; Lima et al. 2013) and potential cross-infection among different Colletotrichum species has been reported (Alahakoon et al. 1994; Freeman et al. 1998; Phoulivong et al. 2012), making accurate identification of the causal agent by symptoms alone difficult. This presents a quandary, because accurate species identification is critical for designing strategic disease management and understanding the pathogen’s population structure and dynamics (Freeman et al. 1998).

Molecular techniques help overcome the inadequacies of traditional methods, and have recently been used to identify and characterize Colletrotrichum species (Than et al. 2008; Cai et al. 2009; Lima et al. 2013; Schena et al. 2013; Huang et al. 2013; Udayanga et al. 2013; Liu et al. 2013b). Analyses of nucleic acids provide the most reliable structure for classifying Colletotrichum species, as DNA traits are not directly influenced by environmental factors (Cannon et al. 2000).

In Colombia, anthracnose in soursop is reported to be caused by Colletotrichum gloeosporioides (Álvarez et al. 2004). However, following epitypification of C. gloeosporioides (Cannon et al. 2008), phylogenetic analysis of multi-locus sequence data showed that C. gloeosporioides is a species complex that contains 22 currently accepted species (Weir et al. 2012). Moreover, Phoulivong et al. (2010) reported that C. gloeosporioides sensu stricto is not commonly found on tropical fruits.

In our study, we therefore aim to apply multi-locus phylogenetic analysis to identify Colletotrichum species associated with soursop anthracnose in Colombia. We also evaluate the fungal species’ pathogenicity and characterize their morphology.

Materials and methods

Source of isolates

Colletotrichum isolates from diseased tissues (including leaves, branches, fruits, and flowers) of soursop trees showing anthracnose symptoms were collected in key production regions of Colombia. To isolate the pathogens and obtain single-spore cultures, the procedure described by Than et al. (2008) was carried out.

Molecular characterization. DNA extraction

Total genomic DNA was extracted from fungal mycelium grown on potato dextrose agar (PDA; Difco Laboratories, Detroit, MI, USA) following the protocol of Damm et al. (2008).

Identifying colletotrichum species complexes by PCR

Colletotrichum species complexes were identified by using PCR amplification of the ribosomal DNA internal transcribed spacer (ITS) region. The primers used to detect the Colletotrichum complexes included the ITS4 primer (5′-TCCTCCGCTTATTGATATGC-3′) (White et al. 1990), coupled with primers specific for C. acutatum s. lat. (CaInt2; 5′-GGGGAAGCCTCTCGCGG-3′) (Sreenivasaprasad et al. 1996), C. gloeosporioides s. lat. (CgInt; 5′-GGCCTCCCGCCTCCGGGCGG-3′) (Mills et al. 1992), and C. boninense s. lat. (Col1; 5′-GCCGTCCCCTGAAAAG-3′) (Afanador-Kafuri et al. 2003; Pileggi et al. 2009).

Genomic DNA from strains GND-1, Pass 063, and Tom 12 was used as positive controls in the PCR amplification for C. gloeosporioides s. lat., C. boninense s. lat., and C. acutatum s. lat., respectively. They were kindly provided by Dr Lucía Afanador-Kafuri, curator of the phytopathogenic fungal collection held at the Plant Health Laboratory, Universidad Nacional (Medellín, Colombia). Negative control was the amplification reaction in the absence of DNA.

PCR amplifications were performed in a 25-μl reaction volume, consisting of 10 ng of DNA template, a 1X final concentration of DreamTaq Green PCR Master Mix (Thermo Fisher Scientific, Inc.), and 0.5 μM of each primer. The PCR reaction was carried out, using a PTC-100 thermal cycler (MJ Research, Watertown, MA, USA). The program included an initial step of DNA denaturation at 95 °C for 6 min, followed by 40 cycles consisting of 30 s at 95 °C, 30 s at either 62 °C (for CgInt and CaInt2) or 60 °C (for Col1), and 1.5 min at 72 °C, with a final extension step at 72 °C for 4 min. The PCR products were analyzed by electrophoresis (6 μl/well), using a 2 % agarose gel stained with SYBR® Safe DNA Gel Stain, and then visualized under a Safe ImagerTM 2.0 Blue-Light Transilluminator (Invitrogen, Carlsbad, CA, USA).

Species identification based on multi-locus phylogenetic analyses

The complete rDNA ITS region, as well as partial sequences of the actin (ACT), β-tubulin (TUB2), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and chitin synthase 1 (CHS-1) genes were amplified and sequenced using the primer pairs ITS-5 + ITS-4 (White et al. 1990), ACT-512F + ACT-783R (Carbone and Kohn 1999), Bt2a + Bt2b (Glass and Donaldson 1995), GDF1 + GDR1 (Templeton et al. 1992), CHS-79F + CHS-354R (Carbone and Kohn 1999), respectively.

The PCR amplifications were performed in a 25-μl reaction volume, consisting of 10 ng of DNA template, a 1X final concentration of DreamTaq Green PCR Master Mix, and 0.5 μM of each primer. The reactions were performed with a PTC-100 thermal cycler, using the thermal programs as previously described for the amplification of ITS, ACT, TUB2, and GAPDH (Prihastuti et al. 2009), and for CHS-1 (Weir et al. 2012).

The PCR products were purified by the PEG-NaCl method. The sample was mixed with 1X volume of PEG-NaCl (20 % PEG at MW 6000 and 2.5 M NaCl) and incubated for 15 min at room temperature. The precipitate was collected by centrifuging at 13,000 rpm for 15 min. The pellet was washed with 70 % ethanol, air-dried, and then dissolved in 20 μl of sterilized distilled water.

The DNA sequences were determined in both directions, using an Applied Biosystems 3730xl DNA Sequencer (Applied Biosystems, Inc., Foster City, CA, USA). Sequencing was performed at the Biotechnology Unit of Iowa State University, USA. The nucleotide sequences were assembled to consensus sequences, edited with ChromasPro v. 1.5 (Technelysium Pty Ltd, Tewantin, QLD, Australia) and aligned, using ClustalW as implemented in MEGA v. 5.1 (Tamura et al. 2011).

For the phylogenetic analyses, molecular data of the targeted genes were compared with sequences of ex-type and other reference strains of accepted Colletotrichum species, which were obtained from GenBank (Table 1). Individual gene alignments were concatenated, using FASconCAT v. 1.0 (Kück and Meusemann 2010). Separate partitions were then created for each gene and the model of nucleotide substitution determined by jModelTest 0.1.1 (Posada 2008), according to the corrected Akaike information criterion (AICc).

Table 1 GenBank accessions of Colletotrichum strains used for the phylogenetic analyses in this study
Full size table

Bayesian inference was used for phylogenetic reconstruction, using MrBayes v. 3.2.1 (Ronquist et al. 2012). Metropolis-coupled Markov-Chain Monte Carlo (MCMC) analysis was performed on the dataset with substitution models determined separately for each partition. For the Bayesian analysis, two MCMC chains were run twice for 1 × 107 generations, with trees sampled every 1000 generations. After omitting the first 25 % of saved trees (burn-in), the remaining sampled trees were used to estimate a 50 % majority rule consensus tree. Consensus trees were visualized and edited with TreeGraph 2 (Stöver and Müller 2010). Sequences derived in this study were submitted to GenBank (Table 1). Sequence alignments and the consensus trees were edited in Mesquite v. 2.75 (Maddison and Maddison 2011) and deposited in TreeBASE (http://www.treebase.org) under the accession number S13841.

Pathogenicity trial

To evaluate pathogenicity and virulence of Colletotrichum species identified, a greenhouse trial was conducted. Scions (twigs) of the soursop cv. Elita were grafted onto sexually reproduced (i.e., seed-based) native rootstocks. When the grafted twigs were 8 months old and about 40 cm high, they were inoculated. Superficial wounds (~7 mm in diameter) were made on stems, using sterilized scalpels. A mycelial plug, with a 6.5-mm diameter, was then taken from the edge of a 15-day-old culture grown on modified Mathur’s medium (Freeman et al. 2000). The plug was placed on the wound, with the mycelial surface facing the cambium. Inoculations were made on three parts of the twig, spaced at 10 cm, beginning with the seedling’s canopy and finishing at its base, ensuring that their locations were above the grafting point and on young tissues. Once completed, the inoculations were covered with parafilm paper (Parafilm M*; American National Can Company, Norwalk, CT, USA).

All inoculated and control plants (which were mock-inoculated with sterilized Mathur’s agar plugs) were incubated for 72 h at 27 to 29 °C and 95 % relative humidity. The plants were then spray-misted for 1 min every hour for 17 days. The isolates were arranged in a randomized complete block design, with three replications per isolate. All trials were replicated twice. The number of isolates inoculated per taxon depended on strain availability, and were treated as subsamples in the statistical analysis. Anthracnose severity was assessed at 23 days after inoculation, using a 0 to 5 point scale, where 0 = no visible symptoms, 1 = 1 to 5 %, 2 = 5 to 10 %, 3 = 10 to 25 %, 4 = 25 to 50 %, and 5 = >50 % of twig area showing symptoms. To confirm Koch’s postulates, fungal isolates were re-isolated from lesion margins onto PDA media and their morphology compared with the originally inoculated isolates.

Data of anthracnose severity were analyzed, using the general linear models procedure (PROC GLM) in SAS v. 6.0 (Statistical Package, Cary, NC, USA). Means of anthracnose severity among species were compared according to the protected Fisher’s least significant difference (LSD) test, with a significance value of P = 0.05.

Morphological characterization of Colletotrichum species

Mycelial plugs (5 mm in diameter) were taken from the edges of 5-day-old colonies and transferred to the centre of 9-cm-diameter petri dishes containing PDA medium. The dishes were incubated at 21 °C with a 12/12 h light/dark cycle, using a cool, white, fluorescent light. Colony appearance and culture diameter were evaluated after 10 days’ growth. For each isolate, the shape, length, and width of 40 conidia were recorded. Morphological studies were conducted, using a completely randomized block design, with three replications. Morphological data were submitted to analysis of variance (ANOVA) to establish differences between species. Means were compared, using the LSD test (α ≤ 0.05) of the SAS program, v. 6.0.

Results

Identifying Colletotrichum isolates and species complexes

A total of 83 single-spore isolates were obtained from soursop tissues showing anthracnose symptoms. The isolates were molecularly identified to the Colletotrichum species complex by PCR amplifications of the rDNA ITS region using primers specific to the respective species complexes. Agarose gel analysis showed that: (i) a 450-bp DNA fragment was amplified from 60 isolates and the reference strain GND-1 with primers CgInt and ITS4, specific to C. gloeosporioides s. lat.; (ii) a 490-bp DNA fragment was amplified from only one isolate and the reference strain Tom 12 with primers CaInt2 and ITS4, specific to C. acutatum s. lat.; and (iii) a 520-bp DNA fragment was amplified from 22 isolates and the reference strain Pass 063 with primers Col1 and ITS4, specific to C. boninense s. lat.

Sequencing and phylogenetic analysis

DNA sequences were generated for 21 isolates, representing: (i) different soursop-producing regions in Colombia, (ii) plant parts (branch, leaf, flower, and fruit), and (iii) Colletotrichum species complexes as determined by the above PCR assay.

The combined datasets of partial sequences of ACT, CHS-1, GAPDH, and TUB2, and the complete sequences of the rDNA–ITS region comprised 1845 characters (including gaps). The gene boundaries and nucleotide substitution models used in the phylogenetic analysis were ACT: 1-285 (JC), CHS-1: 286-565 (JC), GAPDH: 566-867 (K80+I), ITS: 868-1243 (SYM+G), TUB2: 1424-1845 (K80+G). According to the multi-locus phylogenetic analysis, the isolates were grouped into three well-defined complexes: C. gloeosporioides s. lat., C. boninense s. lat., and C. acutatum s. lat. Thus, results of PCR with specific primers to identify the Colletotrichum complexes (ITS4/CgInt, ITS4/CaInt2, and ITS4/Col1) were confirmed.

Within the C. gloeosporioides complex, the Bayesian inference analysis positioned the Colletotrichum isolates from soursop within four well-resolved and distinct clades (Fig. 1). Most of the isolates belong to two species of the “Musae clade” sensu Weir et al. (2012). In this clade, two soursop isolates (GM04-L01 and GM33-L01) clustered closely together with two reference isolates of C. tropicale, while the other six soursop isolates (GM80, GM29, GM36-L02, GM49-L02, GM62-L01, and GM89-L02) grouped with the C. siamense reference isolates. The isolates GM25-L01, GM 64-L02, GM30-L01, and GM52-L02 clustered together with five references isolates of C. theobromicola. Finally, the soursop isolates GM62-L03 and GM78 clustered together with the ex-epitype strain of C. gloeosporiodes s. str. (ICMP 17821).

Fig. 1
figure 1

Phylogenetic consensus tree based on Bayesian inference, illustrating the relationships of the Colletorichum isolates in the C. boninense, C. gloeosporioides, and C. acutatum complexes. The tree was built, using concatenated sequences of the ACT, CHS-1, GAPDH, ITS, and TUB2 genes, and separate models of DNA evolution. Groups within the C. gloeosporioides complex according to Weir et al. (2012; Musae and Kahawae clade) and within the C. acutatum complex according to Damm et al. (2012a; Clade 1 to 5) are indicated. Colletotrichum lindemuthianum (CBS 144.31) is used as the outgroup

Full size image

Within the C. boninense complex, Colletotrichum isolates GM01-L02, GM40, GM44-L01, GM73, and GM59b from soursop were grouped in a clade that was well resolved, together with five reference isolates of C. karstii. Isolate GM52-L01 did not group with any reference species. It was therefore considered as belonging to an undetermined species.

Only one soursop isolate from this study belonged to the C. acutatum complex. Isolate GM77 was associated with clade 1 of the C. acutatum complex (Damm et al. 2012a) and grouped with Colletotrichum sp. strain CBS 129823, which was isolated from Passiflora edulis in Colombia and is genetically close to C. limetticola (CBS 114.4), with only two base pairs difference in the β-tubulin-2 sequence (Damm et al. 2012a). Thus, the isolate was considered as belonging to an undetermined species.

Pathogenicity trial

All 21 Colletotrichum isolates produced symptoms on twigs of soursop cv. Elita that were similar to those observed on branches of soursop trees in the field: oval or round black spots, with irregular margins, depressions in the bark; as well as stem rot. Virulence, measured as the percentage of twig area showing symptoms (anthracnose severity), was significantly different (P < 0.001) among species. Colletotrichum theobromicola and C. tropicale respectively showed high and intermediate levels of virulence, whereas the other species showed low virulence (Table 2).

Table 2 Characteristics of Colletotrichum species associated with anthracnose of soursop (Annona muricata) trees in Colombia
Full size table

Morphological characterization of Colletotrichum species

Morphological characteristics are summarized in Table 2. Fungal colonies were grouped into four types, A, B, C and D, according to the texture and colour of aerial mycelia. Statistically significant differences (P < 0.001) of conidial and colony size were observed between Colletotrichum species collected from soursop in Colombia. Average measurements were similar to those described in the literature (Prihastuti et al. 2009; Rojas et al. 2010; Weir et al. 2012; Damm et al. 2012b; Lima et al. 2013).

Discussion

Colletotrichum is a genus that is currently undergoing taxonomic revision and updating, particularly among its phytopathologically important members (Cannon et al. 2000). This study reports on the occurrence of Colletotrichum species associated with anthracnose of soursop in Colombia and their pathogenicity on the susceptible soursop cv. Elita.

The specific primers used in this study were developed as species-specific primers (Mills et al. 1992; Sreenivasaprasad et al. 1996; Afanador-Kafuri et al. 2003; Pileggi et al. 2009) before these species were shown to be part of species complexes. For this study, the primers reliably differentiated between the isolates at the species complex level, indicating that 73 % belonged to the C.gloeosporioides complex, 26 % to the C. boninense complex, and only 1 % to the C. acutatum complex. The results furthermore indicated that anthracnose of soursop trees in Colombia was mostly associated with species of the C. gloeosporioides complex, followed by species of the C. boninense complex.

According to Weir et al. (2012), taxa other than C. gloeosporioides s. str., C. fructicola, and C. siamense (which are all recognized as belonging to the C. gloeosporioides complex) have one or more bases that do not match those of the CgInt primer. However, such mismatched positions and stringency of the PCR reaction might still result in positive fragment amplification (Weir et al. 2012). We confirmed this in our study: PCR reactions of isolates identified as C. theobromicola and C. tropicale with primers CgInt and ITS4 resulted in a 450-bp DNA fragment.

Multi-locus molecular analyses as employed in our study have been used in various earlier studies to identify and delimit Colletotrichum species attacking different hosts. The analyses are now routinely used as the basis on which to describe new Colletotrichum species (Damm et al. 2012a; 2012b; Weir et al. 2012; Lima et al. 2013; Udayanga et al. 2013; Huang et al. 2013; Liu et al. 2013b). Molecular phylogeny, based on multiple gene sequences, is preferred to using either morphology or rDNA–ITS sequences alone (Crouch et al. 2009) because neither adequately resolves taxonomic issues, especially for the species and clades within the C. acutatum and C. gloeosporioides complexes.

The markers used in the present study were recently accepted for genetic delimitation of species in the genus Colletotrichum (Weir et al. 2012; Damm et al. 2012a; 2012b; Cannon et al. 2012) and are also used for molecular identification in the Colletotrichum database of the CBS-KNAW Fungal Biodiversity Centre (http://www.cbs.knaw.nl/colletotrichum). For our study, this molecular approach reliably differentiated Colletotrichum species collected from soursop. According to our analysis, the genetic diversity was high and the Colletotrichum isolates associated with soursop anthracnose were distributed across seven taxa, including C. karstii, C. theobromicola, C. tropicale, C. siamense, C. gloeosporioides s. str., Colletotrichum sp. indet. from the C. boninense complex (GM52-L01), and Colletotrichum sp. indet. from the C. acutatum complex (GM77).

As far as we can ascertain, except for C. karstii (Damm et al. 2012b), all the known taxa identified (C. siamense, C. gloeosporioides s. str., C. theobromicola, and C. tropicale) represent the first reports of these species in Colombia. Moreover, none of the five taxa present a host-specific relationship with soursop trees, as each has been collected from other hosts as well. The five taxa are reported as having worldwide geographic distribution, and some strains have been associated with plant diseases in other agriculturally important crops such as Theobroma cacao, Coffea arabica, Persea americana, Capsicum annuum, Mangifera indica, Citrus spp., Olea europaea, and Carica papaya (Damm et al. 2012b; Weir et al. 2012; Lima et al. 2013; Schena et al. 2013; Udayanga et al. 2013). Colletotrichum siamense is considered to be a dominant group of species, associated with pre- and postharvest diseases of a wide range of tropical fruits (Udayanga et al. 2013).

In managing Colletotrichum diseases, crop rotation is a highly effective way of promoting healthy crop production (Phoulivong 2011). However, our results indicate that the soursop tree can be a source of inoculum of various Colletotrichum species, thus potentially infecting other plant hosts, and conversely. Therefore, decisions on crop rotation and polyculture should be implemented with caution, because of the presence of a diverse community of Colletotrichum species. To manage anthracnose in soursop, best cultural practices, combined with chemical control, need to be adopted to ensure plant health (Phoulivong 2011). Although the use of resistant cultivars is probably the most desirable technology for disease control, soursop has not yet received much attention in this research line.

Of the species identified, only C. tropicale has been associated with fruit rot of soursop before (Rojas et al. 2010). In addition, isolates of C. karstii, C. theobromicola, and C, siamense have been reported as affecting other Annona species (Damm et al. 2012b; Weir et al. 2012; Udayanga et al. 2013).

The pathogenicity test, whereby trees of soursop cv. Elita were artificially inoculated with isolates of the different Colletotrichum species, showed that all species were pathogenic, albeit at different levels of virulence. Colletotrichum theobromicola and C. tropicale possessed higher and intermediate levels of virulence, respectively, while the remaining Colletotrichum species had low levels of virulence.

These findings suggest that, because symptoms of stems (i.e., dieback of twigs and branches, and stem necrosis) cause the highest crop losses, C. theobromicola and C. tropicale have, economically, the most significant impact on soursop anthracnose in Colombia. Thus, management strategies need to focus on these two species. Similarly, C. theobromicola is considered as a primary pathogen in olive (Olea europaea). In contrast, isolates of C. karstii and C. siamense are weakly pathogenic to olive (Schena et al. 2013).

The Colletotrichum species identified in our study showed cultural and morphological characteristics similar to those described in previous studies. The cultural appearance of Colletotrichum species is highly variable, being determined by diverse factors such as the culture media used, subculture conditions, environmental factors (e.g., temperature, light intensity, and photoperiod), and storage conditions (Weir et al. 2012). Cultural and morphological characteristics therefore cannot be confidently used alone to determine species within a complex (Phoulivong et al. 2010).

The results of this study are relevant because, by demonstrating the diversity and virulence of Colletotrichum species infecting soursop, they will facilitate the development and implementation of disease management practices and of more effective quarantine measures to minimize the risk of introducing new species across borders or continents.