Introduction

Coffee (Coffea arabica) is widely cultivated in Brazil, and it is one of the most important agricultural commodities to the Brazilian economy (Souza et al. 2011). Cercospora coffeicola is the causal agent of brown eye spot (BES) disease, which occurs from nursery-to-field plantings, and severe epidemics are associated with coffee yield losses when management is not well-conducted (Souza et al. 2012). It causes lesions on leaves as necrotic spots consisting of a light-colored center (sometimes surrounded by a purple-brown ring with yellow borders), defoliation, fruit injuries, reduced productivity, and lower beverage quality (Godoy et al. 1997; Pozza et al. 2010; Dell’ Acqua et al. 2011; Souza et al. 2011, 2012; Lima et al. 2012). An atypical symptom characterized by darker and larger lesions on leaves, known as black spot (BS) or black cercosporiosis (Nelson 2008) have often been reported in coffee growing areas in Brazil.

Cercospora coffeicola was originally described by Berkeley and Cooke (Cook 1881) on coffee leaves in Jamaica. In Brazil, the first report dates back to 1901 (Noack 1901). This species has different synonyms, such as Ramularia goeldiana, Cercospora coffeae, and Cercospora herrerana. The sexual morph is known as Mycosphaerella coffeicola. Although type material of C. coffeicola being preserved in the herbarium of the Royal Botanic Gardens, Kew, UK, as K(M) 255,271 and IMI 846, the collections concerned are old, dating back to 1881, so that they are insufficient for novel taxonomic studies including molecular phylogenetic analyses. According to Ariyawansa et al. (2014) dried specimens collected prior to about 1950 cannot be successfully sequenced or sequencing of older specimens requires special, adapted methods as recently shown by Bradshaw and Tobin (2020) for old powdery mildew specimens. Furthermore, some herbaria do not allow samples to be removed for DNA extraction, above all when very type collections are involved. In such cases, epityfication being the methods of choice. An epitype of C. coffeicola with ex-epitype sequences will allow the application of modern molecular methods for identification purposes based on living culture.

The taxonomy of Cercospora species was traditionally based on host plant association (Chupp 1954; Ellis 1971). Moreover, Crous and Braun (2003) introduced the concept of “compound species” which consisted of morphologically indistinguishable species. Thus, the taxonomy of this genus has proven highly problematic, because morphological features in species of Cercospora are generally rather uniform, and there is a high level of intraspecific variation (Groenewald et al. 2013; Bakhshi et al. 2015). Therefore, gene sequence data have been used to solve taxonomic problems related to Cercospora species delimitation (Groenewald et al. 2013; Bakhshi et al. 2015, 2018). According to Groenewald et al. (2013) “many (epi-) type cultures and adequate sequence data are lacking for a significant number of Cercospora species”, which also applies to C. coffeicola, a phytopathologically relevant species that is urgently in need of proper molecular characterization.

This study aims to verify if BES and BS have the same etiological agent. To achieve this aim, we sequenced the internal transcribed spacer regions and intervening 5.8S nrRNAs (ITS), actin (actA), calmodulin (cmdA), histone H3 (his3), and translation elongation factor 1-alpha (tef1), and used multilocus phylogenetic analyses to define the phylogenetic position of the isolates within the genus Cercospora, supplemented by a proposed epitype for Cercospora coffeicola with corresponding ex-epitype reference sequences.

Materials and methods

Fungal isolates and culture conditions

The isolation was made directly from coffee leaves with BES and BS symptoms collected in the five largest coffee-producing states of Brazil: Minas Gerais, São Paulo, Paraná, Espírito Santo, and Bahia (Table 1). Reproductive structures were collected using sterile needle to gently scrap the lesions on leaves, then suspended in sterile water, and spread on the surface of malt extract agar (20 g malt extract L−1, 20 g agar L−1; Himedia Laboratories, Mumbai, India) in a Petri dish. Fungi were incubated at 25 °C for 72 h with a 12-h photoperiod (Santos et al. 2014). Cercospora isolates derived from single conidia were cultured on V8® medium (100 mL V8® plus 900 mL distilled water, 17 g agar, 2 g calcium carbonate), at 20 °C. Nineteen strains were deposited in the Coleção Micológica de Lavras (CML) of the Federal University of Lavras (UFLA), Minas Gerais, Brazil (http://www.dfp.ufla.br/cml). Dried plant material with leaf spot containing fungal structure is deposited in the herbarium of the Universidade Federal de Viçosa (Herbarium VIC).

Table 1 Cercospora coffeicola strains obtained in this study
Full size table

DNA extraction, PCR and sequencing

Genomic DNA was isolated from fungal mycelium grown on 2% malt extract broth medium (Himedia, Mumbai, India) under agitation on a rotary shaker (100 rpm). Genomic DNA extracted with a Wizard® Genomic DNA purification kit (Promega, Madison, WI, USA) according to the manufacturer’s protocol. PCR reactions were performed in a total volume of 25 μL by using a GoTaq Colorless Master Mix kit (Promega, Madison, WI, USA) in a My Cycler thermal cycler (Bio-Rad, Hercules, CA, USA). Primer sets used for PCR and sequencing reactions are listed in Online Resource 1. All isolates were sequenced using primers for actin (actA, Carbone and Kohn 1999), calmodulin (cmdA, Carbone and Kohn 1999), histone H3 (his3, Crous et al. 2004), internal transcribed spacer regions and intervening 5.8S nrRNAs (ITS, White et al. 1990), and translation elongation factor 1-alpha (tef1, Carbone and Kohn 1999). PCR conditions consisted of an initial denaturation step of 5 min at 95 °C, followed by 35 cycles of 30 s at 95 °C; 30 s at 52 °C (actA), 49 °C (cmdA), 53 °C (his3), 60 °C (ITS), 48 °C (tef1); and 45 s at 72 °C, then 5 min at 72 °C. Amplified fragments were cleaned with an Wizard SV gel and PCR Clean-Up System kit (Promega, Madison, WI, USA),and sequenced by Macrogen (Rockville, MD, USA).

Phylogenetic analyses

Consensus sequences were assembled from forward and reverse traces using SeqAssem version 07/2008 software (SequentiX - Digital DNA Processing, Klein Raden, Germany). Sequences from the reference strains of Cercospora spp. (Feau et al. 2006; Dell’ Acqua et al. 2011; Groenewald et al. 2013) available in GenBank were added to the analyses. Septoria provencialis (CPC 12226) was selected as outgroup taxon based on previous studies (Groenewald et al. 2013; Bakhshi et al. 2015, 2018; Nguanhom et al. 2015; Guatimosim et al. 2017). Sequences were aligned using Muscle software operated on MEGA version 7 software (Kumar et al. 2016). The concatenate dataset was comprised of actA (18 parsimony-informative positions/195 bp), cmdA (27/ 308 bp), his3 (7/ 363 bp), ITS (2/ 453 bp), tef1 (12/ 292 bp). Maximum parsimony (MP) and Maximum likelihood (ML) analyses were conducted with MEGA version 7 software (Kumar et al. 2016). Phylogenetic analysis of each individual dataset and the concatenate dataset (actA- cmdA- his3- ITS- tef1) was conducted using MP. The consistency index (CI) and the retention index (RI) were estimated as measures for homoplasy in the dataset and tree length was calculated. ML-based analyses of concatenated datasets were inferred with the general-time-reversible model with Gamma distributed with Invariant sites (GTR + G + I). Clade support was inferred from 1000 bootstrap replications for MP and ML. Bayesian phylogenetic (BI) trees were inferred using MrBayes 3.2 software (Ronquist et al. 2012) to calculate posterior probabilities for concatenated datasets. Models of sequence evolution for each partition were determined using jModeltest (Darriba et al. 2012). Models chosen by the Bayesian Information Criterion were: TPM + G for actA, TrN + I + G for cmdA, HKY + G for his3, and K80 + G for ITS and tef1. Phylogenetic reconstruction under per partition criteria was performed with a combined dataset. Gene boundaries were 01–207 bp for actA, 208–515 bp for cmdA, 516–885 bp for his3, 886–1347 bp for ITS, and 1348–1695 bp for tef1. A Markov Chain Monte Carlo algorithm of four chains was run for 1 × 105 generations and sampled every 100 generations. Fifty percent majority-rule consensus trees were constructed after discarding 25% of the initial trees. Phylogenetic trees were visualized using FigTree (http://tree.bio.ed.ac.uk/software/figtree/) and edited with Corel Draw X5 (Corel Corporation, Ottawa, ON, Canada). Sequences generated in this study were deposited in GenBank (Table 1). Aligned datasets were deposited in TreeBASE under the accession number 26812 (http://purl.org/phylo/treebase/phylows/study/TB2:S26812?x-access-code=3f81ae4c3e3450b592840c5806d6395e&format=html).

Examination of morphological markers

Mycelial disks (0.5 mm in diameter) from colony borders were transferred to 20 mL V8® broth medium (100 mL v8® plus 900 mL distilled water) and incubated under agitation (120 rpm) at 25 °C, for 4 days. Then, the suspension was spread on the surface of 2% water-agar in a Petri dish. Plates were kept open at a distance of 40 cm below a set of 40 W white fluorescent grow lux lamps, in a 12 h photoperiod at 25 °C for 4 days to mycelial mass drying. Mycelia were hydrated with 10 mL distilled water and the suspension was filtered through a layer of cheesecloth (Souza et al. 2011). Microscopic observations were made from clear lactic acid mounts. Measurements were taken from at least 30 conidia and conidiophores for each strain. Photo-documentation was performed by using DM 2000 epifluorescence microscope (Leica, Jena, Germany). The strains were cultivated in triplicates on V8® for determination of growth rates and colony characters in the dark at 25 °C.

Pathogenicity tests

The pathogenicity test was conducted with all 19 strains studied. Coffee tree seedlings cultivar Mundo Novo 376/4 with approximately seven months old of age and showed six pairs of leaves were inoculated with a modified version of the methodology proposed by Souza et al. (2012). Eight mycelial discs, each six mm in diameter, were isolated from 15-day-old C. coffeicola colonies. They were macerated in 400 μL sterile distilled water. The macerated mycelium from each strain was placed in Erlenmeyer flasks containing 20 mL liquid V8® culture medium (100 mL V8 in 900 mL distilled water) and agitated at 100 rpm, for 4 days, at room temperature. The liquid containing the mycelium was poured into plates with water-agar medium. Plates were incubated in BOD until all the liquid evaporated (about 4 days). After the drying step, 10 mL sterile water was added to each plate and conidia were removed using a Drigalski spatula. The liquid containing the conidia was filtered through gauze to remove residue and sporulation was quantified with a Neubauer counting chamber. The suspension used for inoculation was adjusted to 3 × 104 conidia mL−1 and was sprayed on all coffee leaves. Uninoculated plant served as a negative control, which leaves were inoculated with sterile water. Each treatment was composed of four repetitions arranged in a completely randomized design. Occurrence of foliar symptoms was evaluated.

Results

Phylogenetic analysis

A collection of 19 field isolates were obtained from Coffea arabica showing BES or BS-diseased leaves collected in different coffee-producing states of Brazil. Of these 15 were of leaves with symptoms BES and four of leaves with symptoms of BS. The combined tree (actA-cmdA- his3-ITS- tef1) obtained using ML, MP and BI contained the same major clades, and the BI tree is shown in Fig. 1. Cercospora strains from coffee leaves this study formed a monophyletic group together with Cercospora coffeicola of Coffea arabica strain from Japan MUCC 771. Individual phylogenetic performance of the five loci indicated that they have limited resolution to delimitation C. coffeicola from other Cercospora species (Online Resources 2–6). These results reinforce the importance of multilocus phylogenetic analyses for more precise species delimitation.

Fig. 1
figure 1

Majority-rule consensus Bayesian phylogram based on concatenated (actA- his3-cmdA-ITS- tef1) sequences showing relationships among Cercospora species. Strains from this study are highlighted in bold. Bootstrap values and posterior probabilities (ML/ MP/ BI) are shown at the internodes. A minus sign (−) refers to support values lower than 70% bootstrap or 0.90 posterior probability. BES: strains from leaves with Brown eye spot. BS: strains from leaves with Black spot

Full size image

Taxonomy

Cercospora coffeicola Berk. & Cooke, Grevillea 9(51): 99. 1881.

= Sphaerella coffeicola Cooke, Grevillea 9(49): 11. 1880.

= Cercospora coffeae Zimm., Bericht über Land- und Forstwirtschaft in Deutsch-Ostafrika 2: 35. 1904.

= Cercosporina coffeicola (Berk. & Cooke) Speg., Boletín de la Academia Nacional de Ciencias en Córdoba 23(3–4): 589. 1918.

= Ramularia goeldiana Sacc., Sylloge Fungorum 10: 554. 1892.

= Cercospora herrerana Farneti, Atti Ist. Bot. Univ. Pavia, Ser. 2: 37. 1904.

= Mycosphaerella coffeicola (Cooke) J.A. Stev. & Wellman, J. Wash. Acad. Sci.34: 262. 1944.

= Mycosphaerella coffeicola (Cooke) Cif., Atti dell’Istituto Botanico della Università e Laboratorio Crittogamico di Pavia 19: 118. 1962.

Typification: Holotype of Cercospora coffeicola: JAMAICA, on Coffea sp., Jan 1881, Morris (K(M) 255,271). Isotype: IMI 846. Epitype of Cercospora coffeicola (Mycobank MBT52640, hic designatus): BRAZIL, Minas Gerais, Bonfinópolis de Minas, on leaves of Coffea arabica with brown eye spot symptoms, 2012, César Elias Botelho (VIC 47409); Ex-epitype culture CML 2984 = CCDCA 10745 = COAD 3146. GenBank accession numbers for DNA sequences derived from the ex-epitype strain: KU203682 (actA), KU203720 (his3), KU203739 (ITS), KU203701 (cmdA), KU203758 (tef1).

Characteristics in vitro: Colonies on V8® medium attaining 20 mm diameter after two weeks in the dark at 25 °C, white to olivaceous grey, even margins and sparse to moderate aerial mycelium, purple diffuse pigment surrounding culture (Fig. 2a). Characteristics in vivo: Conidiophores aggregated, 30–140 × 4–7 μm, pale brown, simple, unbranched, 0–4-septate, thin-walled, smooth, straight or flexuous caused by sympodial proliferation (Fig. 2c–d). Conidiogenous cells terminal or intercalary, unbranched, pale brown, smooth, proliferating sympodial, conidiogenous loci thickened, darkened, protuberant, apical, lateral, 2.5–4 μm wide (Fig. 2c–d). Conidia solitary, hyaline, acicular to obclavate, straight or curved, 51–270 × 1–5 μm, 0–11-septate, thin-walled, smooth, apex pointed or subobtuse, base truncated to obconically truncated, hila thickened and darkened, 2–4 μm wide (Fig. 2e–g).

Fig. 2
figure 2

Photographs (a, h and j), Stereomicrographs (i and k) and Photomicrographs (c – g). a Morphological characteristic of Cercospora coffeicola, CML 2984 – Epitype colony on V8® after two weeks. b Production of reproductive structures on the lesion. c-d Conidiophores. e-g Conidia. h-i Brown eye spot symptom. j-k Black spot symptom. Scale bar = 10 μm

Full size image

Additional specimens examined: BRAZIL, Minas Gerais, Três Pontas, Coffea arabica, 2012, Alvarenga L, CML 2985; Minas Gerais, Lavras, Coffea arabica, 2012, Botelho DMS, CML 2986; Minas Gerais, Bonfinópolis de Minas, Coffea arabica, 2012, Botelho C, CML 2990; Minas Gerais, Machado, Coffea arabica, 2012, Botelho C, CML 3376; Minas Gerais, Turmalina, Coffea arabica, 2012, Mendonça AC, CML3391; Minas Gerais, São Sebastião do Paraiso, Coffea arabica, 2012, Botelho C, CML 3392; Minas Gerais, Manhuaçu, Coffea stenophyla, 2013, Resende MLV, CML 3395; Minas Gerais, Monte Carmelo, Coffea arabica, 2012, Alvarenga L, CML 3412; Espírito Santo, Marechal Floriano, Coffea arabica, 2013, Almeida JAM, CML2988; Espírito Santo,Venda Nova do Imigrante, Coffea arabica, 2013, Botelho C, CML3396; Espírito Santo, Alegre, Coffea canephora, 2012, Belan LL, CML3398; Bahia, São Desidério, Coffea arabica 2012, Almeida JAM, CML3342; Bahia, Luiz Eduardo Magalhães, Coffea arabica, 2012, Pozza EA, CML3397; Bahia, Luiz Eduardo Magalhães, Coffea arabica, 2012, Pozza EA, CML3413; Paraná, Jaboti, Coffea arabica, 2013, Custódio A, CML3393; Paraná, Londrina, Coffea arabica, 2012, Custódio A, CML3399; São Paulo, Garçae, Coffea arabica, 2013, Vasco G, CML3394; São Paulo, Arandu, Coffea arabica, 2012, Botelho C, CML3414.

Phylogenetic differentiation: This species is differentiated from allied species in phylogenetic analyses by concatenate dataset (actA- cmdA- his3- ITS- tef1).

Description of symptoms caused by C. coffeicola in the field: Brown eye spot (BES), necrotic leaf spots amphigenous, subcircular to irregular, with light-colored centers surrounded by a purple-brown ring with yellow edges (Fig. 2h–i). Black spot (BS), necrotic leaf spots amphigenous, subcircular to irregular, grey to pale brown, surrounded by a large, brown border (Fig. 2j–k). BES and BS lesions, small to fairly large, 2–9 mm in diameter.

Pathogenicity tests

All C. coffeicola strains induced typical symptoms of BES approximately 30 days after inoculation (Fig. 3). Control plants did not develop symptoms. To fulfill Koch’s postulates, the pathogen was re-isolated from inoculated plants.

Fig. 3
figure 3

Leaf symptoms induced by Cercospora coffeicola strains in greenhouse. a. CML 2988. b CML 2990

Full size image

Discussion

This work confirms that the Cercospora strains causing BES and BS on Coffea arabica in Brazil belong to a single species in phylogenetic analyses (Fig. 1). Multilocus analyses showed that all 19 strains from leaves with BES and BS symptoms, including the MUCC 771 strain of Coffea arabica from Japan (Fig. 1), form a single clade, suggesting the involvement of a single species. The strain from Japan, named as Cercospora coffeicola, clusters within Cercospora sp. P clade, according to Groenewald et al. (2013). In this clade, sequences retrieved from cultures isolated from many hosts, such as Acacia, Cajanus, Citrus, Coffea, Dioscorea, Hibiscus and Ricinus, cluster together. The authors reported that the gene loci screened were unable to resolve the taxonomy these isolates and they preferred to treat them as an unresolved species complex. In this study, we included in phylogenetic analyses all strains from P clade, and we verified that strains isolated from Coffea formed a clade phylogenetically distinct from the Cercospora sp. P clade based on concatenated dataset. However, the ITS region was not sufficient to clarify the delimitation of Cercospora species. Some strains obtained in this study cluster together with C. coffeicola (as Mycosphaerella coffeicola, Feau et al. 2006) and other C. coffeicola strains from Brazil (Dell’ Acqua et al. 2011), while other strains clustered with different species of Cercospora (Online Resource 2). Other loci led to various levels of success in species recognition (Online Resources 3, 4, 5 and 6).

Cercospora coffeicola strains showed high morphological plasticity with large variations in dimensions of particular structures, covering the characteristics and measurements given in the original description of Berkeley and Cooke (Cooke 1881). However, morphological parameters alone are insufficient to distinguish most Cercospora species with acicular conidia belonging to the C. apii complex (Crous and Braun 2003). The recognition of such Cercospora species require the application of multilocus phylogenetic analyzes (Groenewald et al. 2013; Bakhshi et al. 2015, 2018). However, the type material of Cercospora coffeicola (Cooke 1881) is very old and not suitable and available for molecular analyses. In order to overcome these obstacles and to determine the name C. coffeicola genetically, an epitype with ex-epitype sequences is proposed according to the rules of the International Code of Nomenclature for algae, fungi, and plants (Shenzhen Code). Epitypes should usually be from the type locality or a region close to and from the same substrate or host as in the original type. However, this is not always possible as, for example, demonstrated in the case of Colletotrichum circinans, the cause of smudge in onion (Walker 1925), which could not be recollected from the original site as it is now a housing estate (Hyde and Zhang 2008). Thus, the selection of an epitype isolated from coffee leaves in Brazil is reasonable, above all since a Cercospora species of cultivated hosts being involved, with unknown origin, occurring wherever coffee is grown. Thus, strain CML 2984 is designated here as epitype for C. coffeicola. The availability of the ex-epitype reference culture, as well as its DNA sequences will be useful for comparison and identification purposes of this species.

In Brazil, the increase in the incidence of cercospora leaf spot in coffee plantations in the late 2000s, coincided with the expansion of coffee cultivation from traditional areas to other regions with different environmental conditions. Furthermore, climate change, cultivation of new coffee varieties and utilization of different cultural practices may have favored coffee pathogens (Fazuoli et al. 2002; Souza et al. 2015). Based on observations made in different coffee growing fields, we believe that environmental changes may influence the occurrence of different symptoms. According to Nelson (2008), environmental factors such as high altitudes, cloudiness and high humidity are favorable conditions for foliar spots to become leaf blights, causing more damage. In addition, increase of the productive capacity of coffee plantations may lead to a nutritional imbalance and leaves may become more susceptible to colonization by C. coffeicola (Nelson 2008; Souza et al. 2015).

The present study clarifies the involvement of a single species, C. coffeicola, as causative agent of two distinct symptoms on coffee leave in Brazil. Moreover, the fresh collection of Cercospora strains obtained from Coffea arabica in Brazil helped to resolve the phylogenetic position of C. coffeicola within of the genus Cercospora and confirmed its status as a species of its own.