Multiple resistance to bacterial halo blight and bacterial leaf spot in Coffea spp.*

Arq. Inst. Biol., v.86, 1-9, e0632018, 2019 RESUMO: O melhoramento de plantas para resistência genética é um método importante para o manejo de doenças, pelos inúmeros benefícios e baixo custo de implementação. No presente estudo, progênies de 11 espécies de Coffea e 16 acessos selvagens de C. arabica foram testados quanto à resposta a Pseudomonas syringae pv. garcae, agente causal da mancha aureolada, doença disseminada nas principais regiões produtoras de café do Brasil e considerada fator limitante para o cultivo em áreas favoráveis a patógenos; e também para P. syringae pv. tabaci, agente causal da mancha foliar bacteriana, doença altamente agressiva detectada recentemente no Brasil. Experimentos separados para cada doença foram realizados em estufa, por meio da inoculação artificial dos patógenos em condições ideais de umidade para o desenvolvimento das doenças. Os resultados mostraram que as progênies Coffea canephora, C. congensis, C. eugenioides, C. stenophylla e C. salvatrix, além dos acessos selvagens de C. arabica Dilla & Alghe e Palido Viridis e da cultivar IPR 102, possuem níveis satisfatórios de resistência simultânea contra mancha aureolada e mancha foliar bacteriana. Os resultados descritos são úteis em programas de melhoramento para resistência duradoura a múltiplos agentes bióticos, fornecendo novas combinações de alelos de resistência por hibridização, bem como para estudos fitopatológicos, para identificar a variabilidade infraespecífica dos patógenos.


INTRODUCTION
The current taxonomy of the genus Coffea comprises 125 species (KRISHNAN et al., 2015), but only grains of C. arabica Lineu and C. canephora Pierre ex A. Froehner species are commercially exploited, corresponding to 55 and 45%, respectively, of the world production (USDA, 2014). According to DAVIS et al. (2006), C. liberica Bull. ex Hiern can also be cited, although commercially it plays a marginal role.
In spite of the wide diversity of the genus, few species have been used in breeding programs of C. arabica and C. canephora, partly because of the genetic barriers that hamper the development of hybrid plants (CARVALHO et al., 1984), but mainly due to the difficulty in reestablishing the original genome of the cultivated varieties and, consequently, the selection of a stable cultivar.
For an improved use of Coffea species in breeding programs, mainly for tolerance to abiotic stresses and resistance to biotic stresses such as diseases and pests, traits of agronomical interest must be phenotyped. In fact, the Coffea germplasm banks were primarily phenotyped for the main diseases of the crop. Some of these diseases have become relevant in a number of coffee-producing countries and sources of pathogen resistance can still be found in wild genetic resources preserved in in situ collections.
Bacterial halo blight of coffee (BHB), caused by Pseudomonas syringae pv. garcae (AMARAL et al., 1956;YOUNG et al., 1978), can cause large-scale damage and disease outbreaks, according to COSTA et al. (1957). Outbreaks of the disease occur in areas with high inoculum potential or in nurseries, making coffee production and/or the sale of seedlings infeasible. In Brazil, BHB has already been detected in several regions of the states of São Paulo, Paraná and Minas Gerais (AMARAL et al., 1956;KIMURA et al., 1973;ZOCCOLI et al., 2011;RODRIGUES et al., 2017b). Moreover, ALMEIDA et al. (2012) reported an increase in disease incidence and severity in the main coffee-producing states of Brazil.
In Brazil, all arabica cultivars registered by the Brazilian Ministry of Agriculture, Livestock and Food Supply (Ministério da Agricultura, Pecuária e Abastecimento -MAPA) are BHBsusceptible, except for cultivar IPR 102, which is resistant to the disease (SERA et al., 2017).
Bacterial leaf spot (BLS), caused by P. syringae pv. tabaci (WOLF; FOSTER, 1917) (YOUNG et al., 1978) is a lessknown disease, firstly observed in a coffee seedling nursery in southern São Paulo State (DESTÉFANO et al., 2010). A recent study reported the occurrence of BLS under field conditions in the state of Paraná, in separated and mixed infections with P. syringae pv. garcae (RODRIGUES et al., 2017b). To date, BLS has been poorly studied, probably because its symptoms are easily confused with BHB, for being extremely similar.
In addition, no source of BLS resistance was described. The commercial C. arabica cultivars Mundo Novo, IAC 125 RN, Obatã IAC 1669-20, Obatã IAC 4739, Bourbon Amarelo, and Icatu Vermelho IAC 4045, as well as a genotype of Timor Hybrid IAC 1559-13 and cultivar Apoatã IAC 2258 of C. canephora, were found to be susceptible to this pathogen (RODRIGUES et al., 2009).
Unlike P. syringae pv. garcae, which is specific to coffee trees (KIMURA et al., 1973), P. syringae pv. tabaci naturally infects a wide range of hosts (BRADBURY, 1986;MALAVOLTA JUNIOR et al., 2008), which represent a source of primary inoculum for coffee plants.
In view of the epidemiological importance of BLS for coffee cultivation, the knowledge about resistance sources to this pathogen must be deepened, and useful information for breeding programs of the crop obtained, targeting the introgression of genes responsible for the expression of resistance to biotic agents in susceptible commercial cultivars.
The purpose of this study was to identify simultaneous resistance sources to BHB and BLS in Coffea spp., as well as in C. arabica accessions of the Germplasm Bank of the Agronomic Institute of Campinas (Instituto Agronômico de Campinas -IAC), Campinas, São Paulo, Brazil, which are potentially useful in breeding programs, especially for C. arabica.

Germplasm accessions
Seeds of 11 species of the genus Coffea and 16 accessions of C. arabica (Tables 1 and 2), in the Coffea Germplasm Bank of the IAC, were transplanted at the cotyledonary stage into 180-cm 3 tubes containing plant substrate and Osmocote ® fertilizer (3 g.L -1 ), and maintained in a greenhouse until the evaluation of BHB and BLS severity.

Bacterial strains and inoculation
The bacterial strains used in the experiments were obtained from the Phytobacteria Culture Collection of the Biological Institute (Coleção de Culturas de Fitobactérias do Instituto  Biológico -IBSBF). A preliminary study reported that mixed P. syringae pv. garcae strains induced high levels of disease severity (RODRIGUES et al., 2017a). A mixture of the P. syringae pv. garcae strains IBSBF 75 and IBSBF 1197, considered highly aggressive, was used to evaluate BHB severity. A 1:1 ratio mixture consisting of 2 mL of each bacterial suspension was inoculated. Pseudomonas syringae pv. tabaci strain IBSBF 2249 was used to evaluate the BLS resistance of the plants.

Experimentation
The response of Coffea spp. to BHB was tested in two independent experiments, in 2012 (E1) and 2013 (E2), using 11 and seven species, respectively. The experiment was arranged in a completely randomized design with three to 23 replications per species, and the experimental plots consisted of a single plant. After inoculation, the plants were maintained at a relative humidity level above 70%.
In 2014, the same coffee plants used in E1 were pruned and maintained at low relative humidity, unfavorable to pathogen growth. Subsequently, in 2015 (E3), plants with 4 -5 expanded leaves (approximately 5 -6 months after pruning) were tested against P. syringae pv. tabaci, strain IBSBF 2249.
In 2014 (E4), the behavior of 16 C. arabica accessions in response to BHB was evaluated as described above. After the end of the experimental period, 42 days after inoculation (DAI), the plants were pruned and maintained at low relative humidity, unfavorable to the pathogen. In 2015 (E5), after the development of at least three healthy internodes, the plants were tested for BLS-resistance. The experimental design was a completely randomized design with six replications per accession, and the plots represented by a single plant.

Disease evaluation
A 0 -5-disease-rating scale was used, according to the symptoms observed in the inoculated area, in which 0 was attributed to symptom-free plants and 5 to leaves with necrosis of the entire inoculated area (RODRIGUES et al., 2017a). For the interpretation of results, the following classification  was considered: resistant (score 0); moderately resistant (1); and susceptible (2 to 5 points on the disease-rating scale).

RESULTS
The results of the experiments E1 and E2 are shown in In C. kapakata, between 7 and 21 DAI, all inoculated leaves dropped, and, in some cases, the bacteria colonized young leaves in the expansion phase.
Segregation for BHB resistance in variable proportions was observed in the other seven evaluated Coffea species (Table 1).
In BHB-resistant plants, distinct reactions were observed in leaf tissues around the wounds caused by inoculations. The species C. salvatrix and C. canephora showed no visible changes around the inoculation areas. In some resistant plants of C. eugenioides and C. stenophylla, darkening around the injured points was observed. Additionally, around some wounds, the presence of a yellowish halo without visual signs of pathogen colonization was observed (Fig. 1). Bacterial flow exudates from these tissues tested negative for the pathogen.
All plants of the species C. kapakata and C. humilis were susceptible, as well as those of the susceptible control cultivar IAC 125 RN of C. arabica.
Segregation for BHB resistance was observed in progenies of other species. The frequency of resistant plants was about 40% in progenies of C. eugenioides and C. congensis IAC 4,350 and less than 20% in C. liberica var. liberica, C. liberica var. dewevrei, and C. congensis. In these three species, the percentage of susceptible plants was 62.5, 89.5 and 68.8%, respectively.
The symptom development in E2 plants was similar as in those of the previous tests (E1). The severity in C. kapakata peaked seven DAI, while symptom evolution was slower on plants of the susceptible cultivar IAC 125 RN.
The severity of BLS was also evaluated in nine species of the genus Coffea (E3). Results of the plants of each progeny were ranked according to the level of disease resistance/susceptibility (Table 1).
The species C. humilis, C. kapakata and C. liberica var. passipagore were found to be as susceptible to P. syringae pv. tabaci as the evaluated C. arabica cultivars.
Sources of BLS-resistance were observed in the species C. congensis, cultivar IAC 4,349, C. canephora, C. eugenioides, C. stenophylla, and C. salvatrix. Among these, only C. congensis and C. stenophylla had no BLS-susceptible plants. In resistant coffee trees of C. stenophylla and in only one of C. salvatrix, dry lesions surrounded by a discrete yellow halo, with no visual symptoms of bacteria colonization, were observed 42 DAI  (Fig. 1). Microscopic examinations of bacterial flow exudates tested negative. The analysis of the reactions of the evaluated cultivars and botanical and exotic varieties of C. arabica to P. syringae, pathovars garcae (E4) and tabaci (E5), are shown in Table 2. All plants of cultivar IPR 102, as well as the variety Palido Viridis, were BHB-resistant. However, resistance to P. syringae pv. tabaci was only observed in Palido Viridis, classified as resistant (1) or moderately resistant (5). The plants of cultivar IPR 102 were BLS-susceptible.
Genetic variability for simultaneous resistance to P. syringae pathovars garcae and tabaci was identified in progenies of variety Dilla & Alghe. Of the varieties São Bernardo and Villa Sarchi, 40 and 33.3% were BHB-resistant, respectively, although BLS-susceptible. All other evaluated C. arabica genotypes were susceptible to both diseases.

DISCUSSION
Resistance to BHB was reported previously by MOHAN et al. (1978) in the species C. eugenioides and C. stenophylla. Our results extended the diversity of resistance sources with the inclusion of C. liberica, C. canephora, C. congensis, and C. salvatrix. Sources of simultaneous resistance to the pathovars P. syringae garcae and tabaci were identified in the species C. canephora, C. congensis, C. eugenioides, C. salvatrix, and C. stenophylla. In the studied plant species, resistance reactions to P. syringae pv. garcae and P. syringae pv. tabaci occur separately or simultaneously, suggesting the existence of different genes acting in the resistance expression to both pathovars in coffee plants.
Probably, there are different resistance mechanisms against these pathogens in the analyzed populations.
Different Coffea species, resistant to BHB and/or BLS, have a particular use in phytobacteriology, with a view to identify infraspecific variability of these pathogens, as already known for other pathovars of P. syringae. The transfer of the resistance genes contained in these resistant sources may also allow the development of a cultivar with intrinsic traits found only in these species.
In spite of some difficulties of this strategy, e.g., the existence of genetic barriers to interspecific crosses (CARVALHO; MONACO, 1968) and the long time required for the introgression of genes of interest and recovery of the recurrent genome, a C. arabica cultivar resistant to L. coffeella was developed in Brazil, by the transfer of resistance genes from C. racemosa (CARDOSO et al., 2014;MENDONÇA et al., 2016). Low-caffeine content cultivars were selected by hybridization of C. eugenioides in recombination with the species C. arabica and C. canephora (NAGAI et al., 2008).
Other currently available methods, e.g., marker-assisted selection (BERNARDO, 2008) or genome-wide selection (MEUWISSEN et al., 2001), can make the use of wild species with multiple traits of interest feasible, for example of C. eugenioides and C. salvatrix, according to our studies both resistant to bacterial diseases, resistant to leaf-miner (GUERREIRO FILHO et al., 1991) and having a low caffeine content in the endosperm (MAZZAFERA; CARVALHO, 1992). Coffea eugenioides was also resistant to coffee berry borer, Hypothenemus hampei Ferrari (SERA et al., 2010), and in C. salvatrix the seed oil content was high (MAZZAFERA et al., 1998).
Genome-wide selection was used by ALKIMIM et al. (2017) for the identification and selection of C. arabica genotypes carrying resistant genes to leaf rust and coffee berry disease, caused by Colletotrichum kahawae (Waller & Bridge), introgressed from C. liberica and C. canephora, respectively. The profile of the species C. liberica was remarkable for Arabica coffee breeding, for having plants that carry the S H 3 gene, a resistant source to all Hemileia vastatrix races described in Brazil so far (FAZUOLI et al., 2009), being indicated as tolerant to cold stress (PETEK et al., 2005;FORTUNATO et al., 2010), as well as having simultaneous BHB and BLS resistance, according to our results.
Among the diploid species, C. canephora was the most adequate BHB and BLS-resistance source for traditional C. arabica breeding, since hybrids between these species could be established without difficulty (CARVALHO et al., 1984), and the simultaneous resistance to these pathogens was observed in this study at a relatively high frequency in progenies of this species. The proportion of BHB-resistant (43%) and BLSresistant (60%) coffee trees in the evaluated progenies suggests that the frequency of resistance alleles in the tested genotype C. canephora is high. Several C. arabica cultivars have been developed from interspecific hybridizations of C. canephora with leaf rust resistant genes, such as Icatu (Brazil) (FAZUOLI et al., 1983), Ruiru 11 (Kenya) (OMONDI et al., 2001), and Cenicafé 1 (Colombia) (FLÓREZ et al., 2016).
However, a detailed investigation of resistance in C. canephora is required, since plants of this species evaluated by COSTA et al. (1957) and MOHAN et al. (1978) were BHB-susceptible. These divergent results suggest considerable variability in the disease resistance of this species.
In view of the difficulties described above, the most adequate method to breed new cultivars with simultaneous and stable resistance is the exploration of a primary gene pool of C. arabica accessions identified as BHB and BLS-resistant.
In this context, the most promising genetic material of the evaluated germplasms is cultivar IPR 102, which is highly yielding and segregates only genes for BLS resistance. This cultivar resulted from the hybridization between C. arabica Bourbon Vermelho Co 667 and C. canephora var. Robusta Co 254. Therefore, the resistance to both pathogens is probably the result of introgression of resistance genes contained in C. canephora. The BHB-susceptibility of cultivar Bourbon Vermelho (RODRIGUES et al., 2017a) and the frequency of resistant plants of C. canephora to the pathovars garcae and tabaci recorded in this study support this hypothesis.
The inoculation results of C. arabica variety Villa Sarchi agreed with MORAES et al. (1975) and confirmed the resistance to BHB. The mutant Palido Viridis of C. arabica, less productive but BHB and BLS-resistant, is also an important tool for studies related to resistance inheritance and for coffee breeding programs. A few plants of C. arabica, variety São Bernardo, were BHB-resistant, probably due to exogenous pollen grain fertilization, since no plants of the hybrid São Bernardo × Mundo Novo with resistance to the disease were observed.
Although the pathogens are genetically very similar, the presence of resistance to one pathogen in a plant does not mean resistance to the other. Therefore, further studies aiming at the selection of plants with multiple resistance to these agents are highly desirable, as well as an improved knowledge of the resistance mechanisms involved.