Effect of Cacao Black Pod Rot Screening Method on Disease Reaction Determination †

: Black pod rot, caused by several species of Phytophthora , is responsible for greater losses than any other disease affecting cacao. Breeders use various approaches to screen material for resistance to Phytophthora spp., however, the method used to assess disease reaction can inﬂuence outcomes. To determine how screening methods affect results, disease reactions of four cacao clones (BE 10, HY 271419, RIM 15 [MEX], and EET 236 [ECU]) were compared using incidence under ﬁeld conditions, and lesion area following artiﬁcial inoculation. Disease incidence differed signiﬁcantly among clones ( p < 0.0007), ranging from 6.1% for BE 10 to 24.0% for HY 271419. Differences among clones were also detected based on lesion area ( p < 0.032), however, their relative ranking differed: BE 10 (53.9 cm 2 ), HY 271419 (80.64 cm 2 ), RIM 15 [MEX] (95.7 cm 2 ), and EET 236 [ECU] (102.4 cm 2 ). These apparent differences observed in disease reaction among clones when comparing methods may be due to interactions with environmental conditions or differences in the pathogen species/isolates present. The improved understanding of the how screening methods used can affect the disease reaction determination and breeding outcomes in cacao germplasm will beneﬁt breeders and farmers.


Introduction
Phytophthora species infect all tissues of Theobroma cacao (L.) plants and cause black pod rot (BPR), a disease responsible for the loss of up to 30% of global production [1]. BPR is responsible the greatest production losses of cacao around the world and resistance to this disease is a priority for breeding programs. BPR resistance can be quantified through disease incidence in the field or by measuring lesion size following artificial inoculation of pods [2,3], twigs, or leaves [4].
However, screening conditions impact the outcome in research trying to identify disease resistance. For example, the wounding of pods prior to inoculation affects results as this screens for post-penetration resistance only [5]. Nyadanu et al. [6] found field incidence to be strongly correlated with results from artificial inoculation on unwounded pods (r = 0.62), but only moderately correlated when pods were wounded prior to inoculation (r = 0.41). A similar analysis on germplasm from the CEPEC/CEPLAC (Centro de Pesquisas do Cacau/Comissão Executiva do Plano da Lavoura Cacaueira) Germplasm Bank in Bahia, Brazil also found a moderate correlation (r = 0.38) between field incidence and lesion area following artificial inoculation [7].
Although digital image analyses have been adopted in many host-pathogen research systems to quantify disease reaction, features such as dark red coloration of many pods, sporulation in lesion centers, and ubiquitous cosmetic damage present on pods means timeintensive image processing step is required for implementing this technology in cacao [3]. Lesions are generally measured by hand; however, both manual and digital measurements of lesion area are very strongly correlated (r = 0.98) [3].
The collection of accurate disease reaction data is essential for the development of diseaseresistant crops. The objective of this research was to determine if different methods used to screen cacao germplasm for resistance to BPR produced similar conclusions. The comparison was conducted by examining the relative disease resistance of four clones obtained using incidence under field conditions and lesion area following artificial inoculation.

Plant Material
Four clonal accessions (BE 10, HY 271419, RIM 15 [MEX], and EET 236 [ECU]) held at the USDA-ARS National Plant Germplasm System (NPGS) cacao germplasm collection at the Tropical Agriculture Research Station in Mayaguez, PR were evaluated using data collected on disease incidence in the field and disease severity following artificial inoculation. Each clone was represented by six trees planted in a randomized complete block design and were chosen for these experiments based on pod availability.

Field Incidence
Disease incidence was determined by recording the number of healthy and infected pods during monthly harvests from 2007-2011. To minimize the effect of productivity differences on disease incidence values, only months in which at least one tree of each clone had pods present were included in the analysis. Disease incidence per tree was calculated by dividing the number of diseased pods by the number of total pods: Disease Incidence = # diseased pods/(# healthy + diseased pods) Disease incidence was compared among clones using a non-parametric analysis of variance (ANOVA) in SAS 6.2. Means were separated using post hoc nonparametric t-tests with a Bonferroni correction. Each evaluation date counted as a replication, and incidences for each tree were averaged to produce a single value per clone. Trees without pods were excluded from the analyses.

Detached Pod Inoculations
Unwounded pod inoculation assays were carried out on unripe, detached pods (four to five months old) using 1700 zoospores of P. palmivora from 10-day old cultures, with release induced using cold shock, as described in Ali et al. [8]. Inoculated pods were placed in Ziplock bags to prevent desiccation, and kept at 25 • C in the dark ( Figure 1a).
All inoculations were carried out with a P. palmivora isolate H33, which was obtained from a stem canker on cacao in Hawaii [9], and was shown to be an aggressive isolate in preliminary experiments (Puig, unpublished).
After seven days, two lesion diameters were measured per pod and converted to radii by dividing by two. Lesion area was calculated as: where r 1 and r 2 are the largest and smallest radii of the lesion, respectively (Figure 1b). Lesion area was compared among clones using an ANOVA and means were separated using Tukey-Kramer Comparison lines for the least squares means. The assay was carried out as a completely randomized design with 11 to 13 replications (pods) per clone.

Results
Disease reaction, measured as incidence of black pod rot under field conditions, differed significantly among clones (p < 0.0007; DF = 3; X2 = 16.9), ranging from 6.1% for BE 10 to 24.0% for HY 271419 ( Differences in disease reaction among clones were also detected based on lesion area (p < 0.032; numDF = 3; denDF = 25.1; F = 3.43). However, their relative ranking differed from results for incidence because HY 271419 was not as susceptible (

Results
Disease reaction, measured as incidence of black pod rot under field conditions, differed significantly among clones (p < 0.0007; DF = 3; X2 = 16.9), ranging from 6.1% for BE 10 to 24.0% for HY 271419 ( Differences in disease reaction among clones were also detected based on lesion area (p < 0.032; numDF = 3; denDF = 25.1; F = 3.43). However, their relative ranking differed from results for incidence because HY 271419 was not as susceptible (

Discussion
High-quality disease reaction data are needed by breeding programs to develop durable, resistant cultivars. This study examined relative disease resistance of four clones using two assessment approaches: field incidence and quantification of lesion area following artificial inoculation. The artificial pod inoculation protocol described here was designed to incorporate pre-and post-penetration resistance by using a much lower number of zoospores than is used by other researchers [8,10]. Inoculum concentrations that result in 100% infection are less likely to reflect pre-penetration resistance.
The change in disease resistance ranking of clones when assessed using artificial inoculation instead of field evaluations was caused by the shift of HY 271419 from most susceptible (based on field incidence) to second most resistant clones when using data from artificial inoculation. This could be caused by HY 271419 having resistance to colonization, which is not reflected in field incidence data, as it does not take lesion size into account. Similar results were observed by Lawrence [11], were disease reactions of clones obtained in detached pod assays were not consistent with the reactions of the same clones to natural infection in the field.
Alternatively, the differences observed here may be due to pathogen-specific responses. Although P. palmivora is the only species found causing BPR in Puerto Rico so far [12,13], the isolates present in the field may differ from the Hawaiian isolate used in this study. Some host resistance is not durable when challenged with diverse isolates of the same species.
In addition, several fungi in the Diaporthe, Lasiodiplodia, and Colletotrichum genera have been found to cause necrotic lesions indistinguishable from those caused by Phytophthora spp. [13,14]. Disease incidence data collected in the field likely includes pods affected by various pathogens that cause similar symptoms. Although this information is important for farmers, the presence of multiple, unrelated pathogen may complicate genomic studies and make it difficult to accurately identify pathogen-specific resistance markers in germplasm.

Conclusions
This study shows that the assay used to screen for disease reaction can affect which clones are classified as resistant or susceptible. For best results, initial screening using less labor-intensive methods, such as evaluating disease incidence in the field, or artificial inoculation of leaf discs, should be followed by artificial pod inoculations of clones of interest. Thorough characterization of cacao germplasm will enable the identification of new sources of disease resistance and more effective, long-term disease control.