Breeding for resistance to coffee berry disease in Coffea arabica L. II. Inheritance of the resistance

Summary The inheritance of resistance to coffee berry disease (CBD) has been studied by applying a preselection test to F₂ progenies of a half diallel cross between 11 coffee varieties with different degrees of resistance and to sets of parental, F₁, F₂, B₁₁ and B₁₂ generations of crosses between resistant and susceptible varieties. True resistance to CBD appears to be controlled by major genes on three different loci. The highly resistant variety Rume Sudan carries the dominant R- and the recessive K-genes. The non-allelic interaction between these two genes is of a duplicate nature. The R-locus has multiple alleles with R ₁R₁alleles present in Rume Sudan and the somewhat less effective R ₂R₂alleles in a variety like Pretoria, which also has the K-gene. The moderately resistant variety K7 carries only the recessive K-gene. The arabica-like variety Hibrido de Timor (a natural interspecific arabica x robusta hybrid) carries one gene for CBD resistance on the T-locus with intermediate gene action. It probably inherited this gene from its robusta parent. There is circumstantial evidence that the resistance to CBD is of a stable nature, but it is advisable to accumulate in one genotype as many resistance genes as possible by combining in the breeding programme the resistance of Rume Sudan with that of Hibrido de Timor.     

Breeding for resistance to coffee berry disease caused by Colletrotrichum coffeanumNoack (Sensu Hindorf) in Coffea arabica L. I. Methods of preselection for resistance

Summary Two methods of preselection for resistance to coffee berry disease, by artificial inoculation of young coffee seedlings, have been developed recently at the Coffee Research Station in Kenya. The efficiency of these methods has been investigated in a series of experiments by comparing the results of the preselection tests with mature plant resistance, as indicated by visual field scores and artificial inoculation of berries and bud leaves of mature trees and including a large number of varieties and cultivars of Coffea arabica. The preselection test, by which the hypocotyl stems of 6-week old seedlings are inoculated with a spore suspension of the pathogen, proved to be more reliable than the second preselection test with shoot-tips of 10-month old plants. Correlation coefficients between the preselection test and mature plant resistance were resp. r=0.73–0.80 for the first test against r=0.60 for the latter.     

Identification of RAPD markers for resistance to coffee berry disease, Colletotrichum kahawae, in arabica coffee

Resistance to Coffee Berry Disease (CBD) in Arabica coffee is controlled by at least three genes which are present in the varieties Hibrido de Timor (T gene), Catimor (T gene), Rume Sudan (R and k genes) and K7 (k gene). Hibrido de Timor, Catimor and Rume Sudan are genetically distant from most of the commercial cultivars, and the utilisation of molecular markers would greatly improve the efficiency of breeding programmes concerned with CBD resistance. The objectives of the present work were therefore: (1) to identify random amplified polymorphic DNA (RAPD) markers associated with CBD resistance and (2) to identify markers which could be used to select against the genetic background of the resistance donors. Identification of RAPD markers was carried out in three steps. The first step involved the comparison of the RAPD profiles between the susceptible cultivars and the resistant donors. This was followed by comparison of the RAPD profiles between resistant and susceptible types of each donor variety. The final step involved assay of the resistance markers in the first and the second backcrosses between these donors and the recurrent parent. High genetic variability was demonstrated in Catimor, and to some extent in Rume Sudan. Three RAPD markers were shown to be closely associated to the T gene. Attempts to identify markers associated with the R and k genes were less rewarding. The implications of the current observations in relation to breeding for CBD resistance in Arabica coffee are discussed. This revised version was published online in July 2006 with corrections to the Cover Date.     

Early determination of yield potential in arabica coffee by applying index selection

Summary Estimates of phenotypic and genotypic variances and covariances for a number of quantitative characters in arabica coffee, obtained from a variety trial in Kenya, were used to construct selection indices. The results indicated that the expected genetic advance in yield based on a selection index, containing the first two years yield and measurements of girth at base of stem and percentage bearing primaries, will be almost as large (relative efficiency 97%) as that obtained by straight selection based on 10-years yield totals. This would imply that a breeding cycle of 5 years will be sufficient for efficient selection for higher productivity in arabica coffee.     

Comparison of the sensitivity of the mean and the mode in screening coffee seedlings for resistance to coffee berry disease Colletotrichum coffeanum Noack sensu Hindorf (A short communication)

Summary The utility of the modal grade as an alternative to the mean grade in the pre-selection test for screening seedlings of Coffea arabica L. for resistance to coffee berry disease (CBD) caused by Colletotrichum coffeanum NOACK sensu HINDORF was examined. Three criteria were used to compare the sensitivity of the modal grade with that of the mean grade when partitioning 6 weeks old seedlings into different resistance classes. The criteria were: the sensitivity of the two statistics as estimators of central tendency, the sensitivity of the modal grade and the mean grade as selection criteria and the correlation between the two statistics.The modal grade was neither more sensitive in detecting genotypes that fell in diffent resistance classes nor more severe than the mean grade in screening for genotypes susceptible to CBD. In the cases studied, the modal grad was highly correlated with the mean grade (r_s ranged from 0.55^**–0.82^**). Hence, it is concluded that there are practically no gains in adopting the modal grade as a pre-selection criterion in preference to the mean grade.     

Resistance of Coffea arabica cv. Ruiru 11 tested with different isolates of Colletotrichum kahawae, the causal agent of coffee berry disease

Seven single conidia isolates of Colletotrichum kahawae varying in pathogenicity were used to inoculate hybrid progenies from 66 crosses ofCoffea arabica cv. Ruiru 11. The objective of this study was to investigate the effect of pathogen variation on resistance of the Ruiru 11 cultivar. The main effects of crosses and isolates were significant (p ≤0.05) while their interaction effects were non-significant. Partitioning variance components indicated that the proportion of phenotypic variance for resistance that is due to genetic effects was low. It was concluded that variation for resistance among hybrid progenies of the Ruiru 11 cultivar was probably due to differences in aggressiveness of the pathogen as reflected by the significant main effects of crosses and isolates in combination with other environmental factors which influence disease epidemics. The coffee berry disease pathogen is unlikely to have adapted to the cultivar because of the non-significant crosses × isolates interaction effects. This revised version was published online in August 2006 with corrections to the Cover Date.     

Genetic diversity of wild coffee (Coffea arabica L.) using molecular markers

Genetic diversity was studied using RAPD markers among119 coffee (Coffea arabica L.) individuals representing 88 accessions derived from spontaneous and subspontaneous trees in Ethiopia, the primary centre of species diversity, six cultivars grown locally in Ethiopia, and two accessions derived from the genetic populations Typica and Bourbon, spread in the 18^th century, which gave rise to the most currently grown cultivars. Twenty-nine polymorphic fragments were used to calculate a similarity index and construct dendrograms. The Ethiopian material was separated from the Typica- and Bourbon-derived accessions and classified in four groups: one with most of the collected material from southwestern Ethiopia and three from southern and southeastern Ethiopia. Almost all detected diversity was found in the southwestern group while the southern and southeastern groups presented only 59% of identified markers. The genetic distances were low between the southwestern group and the southern and southeastern groups, and between the southwestern group and the Typica- and Bourbon-derived accessions. The cultivated coffee derived from the genetic populations Typica and Bourbon appeared little differentiated from wild coffee growing in the southwest. The results supported the hypothesis that southwestern Ethiopian coffee trees could have been introduced recently in the south and southeast. A separate analysis of the 80accessions classified in the southwestern group allowed identifying particular spontaneous- and subspontaneous-derived accessions and redundancies in the collected material from southwestern Ethiopia. RAPD markers did not detect any within-collection polymorphism except for two trees that were identified as off-types in the CATIE field genebank. This revised version was published online in July 2006 with corrections to the Cover Date.     

Responses of different arabica coffee (Coffea arabica L.) clones to varied environmental conditions

Different improved coffee arabica crosseshaving resistance to coffee berry diseaseand leaf rust, plus a standard cultivar,were grown under a range of environmentalconditions in a series of field trials inKenya. The effects of location and year ofproduction and their interactions with theclones were determined for berry yield.These data were used to estimate andcompare methods of analysis of GEI. An AMMImodel, joint regression model and ANOVAwith repeated measurements were used toinvestigate these interactions. The resultsindicated the relative effects of theclones and the environmental conditions,with Meru being the location where therewas greatest discrimination between theclones, probably because it was bettersuited for coffee. Clones,environments and joint regressions combinedaccounted for 31.20% of the interaction,leaving a residual of 68.80%. Significantresidual mean squares showed therelationship between clones andenvironments not to be strictly linear,thus indicating that regression analysisdid not always provide good estimates ofindividual clonal performance acrossenvironments. High correlations wereobtained between mean yield and regressioncoefficients and between ecovalence andmean square deviations as well as thecoefficient of determination. Resultsshowed that the improved clones surpassedthe standard cultivar both in terms ofyield and stability.Different analyses of the various aspectsof genotype response enabled the 21 clonesto be classified into different similaritygroups based on their performance andphenotypic stability as measured by variousstability parameters and the use of biplotsfrom the AMMI analysis.     

Race-specificity and inheritance of incomplete resistance to coffee leaf rust in some Icatu coffee progenies and derivatives of Hibrido de Timor

Summary Selfed and cross-pollinated progenies of Icatu coffee plants and derivatives of Híbrido de Timor, with susceptibility or moderate resistance to coffee rust in the field, were tested with Hemileia vastatrix race II and isolate 2 (Is. 2) in the greenhouse or laboratory. Progenies from plants with a susceptible reaction type in the field (scores 8 and 9) all showed homogeneous levels of susceptibility similar to that of the control cultivars Mundo Novo and Catuaí of Coffea arabica. Variation for incomplete resistance was mainly expressed by low or heterogeneous reaction types and also by longer latency periods. Incomplete resistance to race II of some Icatu and Catimor plants, which is overcome by Is. 2, appeared to depend on a partially dominant gene. The expression of this gene may vary also according to the genetic background and some residual resistance to Is. 2 was observed. A few derivatives of Híbrido de Timor were more resistant to Is. 2 than to race II, suggesting the presence of a resistant factor to Is. 2 in this germplasm. Segregation in progenies from other plants with incomplete resistance to race II suggested presence of one or a few genes only, which in homozygous condition or in combination seem to confer near-complete or complete resistance. The results suggest that selection for incomplete resistance, expressed by intermediate reaction types, in Icatu and derivatives of Híbrido de Timor may not provide durable resistance to coffee rust. On the other hand, due to apparent additive gene action, phenotypic selection of plants with a highly resistant reaction type could facilitate the accumulation of several resistance alleles in one genotype, which could provide an efficient barrier against new race formation of the pathogen.     

Pollen longevity and artificial cross-pollination in Coffea arabica L

Summary Various aspects of pollen longevity, in vitro germination of pollen and controlled pollination in Coffea arabica were investigated. High pollen viability could be maintained for more than two years, much longer than previously reported, by storing it under vacuum at 18°C. The most satisfactory method of in vitro germination of pollen was the hanging drop in a Van Tieghem cell, with a 10°, sucrose solution. For artificial cross-pollination it is necessary to carry out emasculation and bagging at the latest one day before anthesis. The stigmas of unpollinated flowers remained receptive for at least nine days.This paper is published with permission of the Director, Coffee Research Station, Ruiru, Kenya     

Genetic diversity for RAPD markers between cultivated and wild accessions of Coffea arabica

Summary Random amplified polymorphic DNA (RAPD) markers have been successfully employed to analyse the genetic diversity among cultivated and subspontaneous accessions of Coffea arabica. The narrow genetic base of commercial cultivars was confirmed. On the other hand, a relatively large genetic diversity was observed within the germplasm collection demonstrating the importance of collecting missions. Results suggested an East-West differentiation in Ethiopia, the primary centre of diversification of C. arabica. The large heterosis effect reported in intergroup hybrids could be related to such genetic differentiation. RAPD method appeared to be effective in resolving genetic variations and in grouping germplasm in C. arabica.     

Resistance to Coffee Berry Disease in Arabica Coffee Variety 'Ruiru 11'

General and specific combining ability effects for resistance to coffee berry disease (Colletotrichum coffeanum Noack sensu Hindorf) in Coffea arabica variety ‘Ruiru 11’ and its parents were estimated according to the North Carolina Design II. A set of nine male parents and eleven females were crossed in all possible combinations to generate ninety-nine single cross hybrids constituting the ‘Ruiru 11’ variety. The parents and their hybrid progenies were screened and the mean score data analyzed for possible genetic variation. The contributions of the parents to the total genetic variance were low hence the low narrow sense heritability (h²= 0.04). The male and female parents accounted for 9.36 % and 0.96 %, respectively, while their interaction effect accounted for 89.68 %. In the analysis of combining ability for resistance, ExB3.96 and Cat.90 exhibited significant negative general combining ability (GCA). Specific combining abilities (SCA) were negative and significant for Cat.86 × ExB3.99, Cat.127 × ExB3.116 and Cat.119 × ExB3.879 crosses. The implications of these results in selection for highly resistant ‘Ruiru 11’ progenies are discussed.     

Prediction of genetic value for coffee production in Coffea arabica from a half-diallel with lines and hybrids

Genetic analysis of yield and morphological traits has been carried out in Coffea arabica from a half-diallel including the parental lines. The trial was established in west Cameroon with completely randomized single-tree plots. Observations included yield (four years), stem diameter, height and number of primaries. General combining abilities (GCA) and specific combining abilities (SCA) as heritabilities and genetic correlations were estimated. A significant SCA variance was observed for all the traits. Morphological traits, stem diameter, plant height and number of primaries, were genetically correlated to the yield. The hybrids were, on average, better performing than lines. There was no clear relationship between performance of lines and their general combining ability. Contribution of the seven lines to the SCA sum of squares was shown to be unequal for all the studied traits, one parent (Java) being far the most interactive. This variation of interactivity seemed to be related to different levels of residual heterozygosity among parental lines. Indirect prediction of yield using morphological traits gave the same value for expected genetic gain as the prediction based on cumulated yield over four years. The multitrait indirect selection for yield operated in a different way as far as GCA and SCA are concerned. The application of these results to coffee breeding strategies involving hybrid selection is discussed. This revised version was published online in August 2006 with corrections to the Cover Date.     

Selection for bean and liquor qualities within related hybrids of Arabica coffee in multilocal field trials

The success of a new variety of Arabica coffee (Coffea arabica L.) depends to an important extent on its liquor and bean qualities. Selection for these traits is however constrained by the prevalence of large genotype-by-environment (G×E)interactions in conjunction with the low genetic variability characteristic of this species. The aim of this study was to assess the extent to which key bean and liquor traits are affected by the environments and could be combined in order to improve the efficiency of selection for coffee quality in a narrow genetic basis set of hybrids obtained from a set of related lines and collectively representing the Ruiru 11 cultivar. Twenty-one full-sib families representative of this hybrid cultivar grown in Kenya in five sites exhibiting strong edaphic and climatic differences were used for the study. Rainfall amounts during various phases of berry development were used to explain the differences observed in the discriminating abilities of the locations for bean and liquor traits. The results showed that these families were best differentiated for bean sizes in the site where moisture supply was optimal throughout berry expansion and filling stages, whereas discrimination on the basis of liquor traits were best observed in the site where moderate moisture stress occurred during bean filling stage. The overall precision in prediction of family values was low for liquor qualities; but a much more efficient selection for large bold beans, optimally based on the AA grade was shown to be possible. Selection indices for family selection were computed to realize a trade-off between genetic gains in bean size and liquor flavour. This revised version was published online in August 2006 with corrections to the Cover Date.     

AFLP analysis of introgression in coffee cultivars (Coffea arabica L.) derived from a natural interspecific hybrid

S.288 an offspring of a putative spontaneous interspecific hybrid between tetraploid Coffea Arabica (2n = 4x = 44) and diploid C. liberica(2n = 2x = 22) and 17 arabica coffee introgression lines (representing F₂ and F₄) derived from the cross S.288 x Kent arabica were evaluated for introgression of C. liberica genetic material. In all, 36 AFLP primer combinations were used in the analysis. The AFLP profiles of introgressed lines were compared to five accessions each of C. arabica and C. liberica. A total of 137 polymorphic bands were scored among the 29 accessions analysed. The introgressed genotypes exhibited 102 marker bands consisting of 65 additional bands and 37missing bands associated with introgression of C. liberica genetic material. C. liberica accessions of EA group (C. liberica var liberica of Guinean origin) seemed to be the likely progenitor in the origin of natural hybrid. Analysis of genetic relationships in the introgressed lines suggested that introgression was limited to few fragments. Segregation and wide variation in number of marker fragments in the F₂ and F₄progenies were attributed to chromosome recombinations. The differences in the level of introgression between introgressed parent, F₂ and F₄ groups was not pronounced. Therefore the alien genetic material appeared to be fixed and there was no elimination or counter-selection over generations, from introgressed parent to F₄.In C. arabica accessions, only 35 polymorphic bands were seen confirming the low genetic diversity. On the contrary, although representing a small amount of alien genome introgression, the Liberica-introgressed genotypes provided notable genetic diversity. Considering the fact that the diploid species of Coffea constitute a valuable source of genetic diversity, the potential implications of variability generated by Liberica-introgressed genotypes in C. arabica breeding are discussed. This revised version was published online in August 2006 with corrections to the Cover Date.     

Additional evidence for oligogenic inheritance of durable host resistance to coffee berry disease (<Emphasis Type="Italic">Colletotrichum kahawae</Emphasis>) in arabica coffee (<Emphasis Type="Italic">Coffea arabica</Emphasis> L.)

Breeding for host resistance to coffee berry disease (CBD) in arabica coffee (Coffea arabica) was initiated some 35–40 years ago in Kenya, Ethiopia and Tanzania in response to severe CBD epidemics. The release of CBD resistant cultivars to the coffee growers has been in progress since 1985. The resistance of cultivars like Ruiru 11 (Kenya) and Ababuna (and other cvs in Ethiopia) appears to be of a durable nature, since confirmed cases of a breakdown of host resistance under field conditions have not been reported over the past 20 years. Host resistance to the hemibiotrophic fungus Colletotrichum kahawae is of a quantitative nature, but nevertheless can be practically complete in some genotypes of arabica coffee. There is still no consensus on the genetics of CBD resistance, some claiming convincing evidence for oligogenes (1–3 major genes) and others for polygenes determining CBD resistance. Results from genetic studies with germplasm from the centre of genetic diversity for C. arabica in Ethiopia are presented here. These together with the recent identification of molecular markers associated with and the mapping of one major gene, provides additional evidence for oligogenic inheritance of CBD resistance. The development of cultivars combining yield and quality with durable host resistance to CBD has contributed greatly to increased sustainability of arabica coffee production in Africa. It has also considerable relevance to arabica coffee in Latin America and Asia, where CBD is still a quarantine disease but with a risk of becoming endemic one day, just as has happened earlier with coffee leaf rust (Hemileia vastatrix).     

A new source of resistance against coffee leaf rust from New-Caledonian natural interspecific hybrids between Coffea arabica and Coffea canephora

The development of cultivars resistant to coffee leaf rust caused by the fungal pathogen Hemileia vastatrix is a priority in coffee breeding. However, only very few descendants of interspecific hybrids between Coffea arabica and diploid relative species have been used as resistance source. Identification of new sources of resistance appeared therefore particularly worthwhile. Hybrid plants derived from interspecific hybridization between C. arabica and Coffea canephora and found in neo‐natural coffee tree populations of New Caledonia were therefore investigated. Amplified Fragment Length Polymorphism and microsatellites amplification were used to evaluate the genetic diversity of 14 hybrid plants, and rust resistance was evaluated by inoculation with a panel of rust races representing a large variability in virulence. An important genetic diversity was characterized in hybrid plants originating from introgressions into C. arabica from various C. canephora progenitors. On the 14 plants tested for leaf rust resistance, eight appeared resistant to all races investigated. Such plant material should represent a highly valuable resource for C. arabica breeding against coffee leaf rust.     

Genetic analysis of partial resistance to coffee leaf rust ( Hemileia vastatrix Berk & Br.) introgressed into the cultivated Coffea arabica L. from the diploid C. canephora species

The objective of this study was to identify polymorphic molecular markers associated with partial resistance to coffee leaf rust, Hemileia vastarix. A segregating F ₂ population derived from a cross between the susceptible Coffea arabica cv. Caturra and a C. canephora-introgressed Arabica line exhibiting high partial resistance was analyzed. Rust resistance measured as rust incidence (RI) and defoliation (DEF) was evaluated in field conditions in three consecutive years (2003–2005). During the 2003 season, which was characterized by favorable conditions for a rust epidemic, the F ₂ plants exhibited different levels of resistance ranging from very susceptible (50.1% for DEF and 49.5% for RI) to highly partial resistance (9.1% for DEF and 3.7% for RI). Molecular analysis enabled identification of seven polymorphic markers (5 AFLP and 2 SSR) exhibiting significant association with partial resistance. Coexistence of resistance homozygous alleles (RR) at codominant SSR loci was correlated with high resistance. This study is the first attempt to develop PCR-based sequence specific markers linked to partial rust resistance in coffee.     

Production of a new low-caffeine hybrid coffee and the biochemical mechanism of low caffeine accumulation

The GCAs are new tetraploid interspecific hybrids developed in Madagascar from Coffea eugenioides, C. canephora and C. arabica. Selected GCA having genotype UF1023 contained 0.37% DW caffeine and no detectable theobromine in green beans. Low caffeine accumulation in GCA plants is due mainly to the low biosynthetic activity of purine alkaloids, possibly the extremely weak N-methyltransferase reactions in caffeine biosynthesis. No significant catabolic activity of caffeine was found in GCA-UF1023, in common with almost all coffee plants including C. arabica.