Characterization of genetic resistance to Coffee Berry Disease (Colletotrichum kahawae Waller and Bridge) in Arabica coffee (Coffea arabica L.) that is introgressed from Coffea canephora Pierre
Abstract
Coffee Berry Disease (CBD) is an anthracnose of young berries of Arabica coffee (Coffea
arabica L.) that is caused by the fungus Colletotrichum kahawae. It is a major limitation to
economic production of the crop in Africa. Various sources of resistance to the disease have
been identified and are used in breeding resistant cultivars. One such source of resistance is
Hibrido de Timor (HDT), which is a natural hybrid between C. arabica and C. canephora. In
Kenya, accessions of HDT progenies and its derivatives (cv Catimor) are used as donors of
resistance to both CBD and CLR. The objective of this study was to decipher the genetic basis of
CBD resistance derived from Hibrido de Timor and to identify molecular markers associated
with it, which can be used for selection purposes.
Potential Amplified Fragment Length Polymorphism (AFLP) and microsatellite markers for the
resistance were identified by characterisation ofHDT derived polymorphism in resistant lines of
cv Catimor. The accessions analysed included two lines of cv Catimor), eight resistant
accessions of BC, F, progenies (Catimor x (Catimor x SL28)), up to 76 plants from three BC, F2
populations and two accessions of the susceptible cv SL28. A Sarchimor line (T5296) and
accessions of its F2 progeny derived from its cross with a wild C. arabica collected from
Ethiopia (ET6), which was used to map introgressed C. canephora chromosomal fragments,
were included in some of the experiments. Three mapped C. canephora chromosomal fragments
(T2, T3 and T4) were found to be present in the cultivars Catimor and Sarchimor and were
therefore considered to be candidate carriers for CBD resistance. However fragment T4 was
considered to be a weaker candidate because it was absent in one resistant BC, F, plant. Some
AFLP markers of the introgressed fragments were cloned and converted into sequence
characterised amplified regions (SCARs), and then assessed for polymorphism in a doubled
haploid (DH) population so as to identify their linkage to coffee chromosomes. The SCARS
displayed very low polymorphism and it was possible to identify chromosome linkage for only
one SCAR (13), derived from the C. canephora chromosomal fragment Tl. This SCAR was
duplicated in chromosomes 2 and 8 of coffee genome.
Two F2 populations (D and E) were raised by from crosses between two lines of cv Catimor
(lines 127 and 88 respectively) and cv SL28. Phenotypic segregation for CBD resistance was
verified by inoculation of half of each seed lot on the sixth week after germination by hypocotyls
inoculation method. Resistant seedlings obtained from these tests were established in a nursery
as Group 1 sub-populations and were used as checks in subsequent molecular studies. The other
halves of the seed lots were transferred directly to the nursery without inoculation as Group 2
sub-populations representing unaltered F2 populations for later studies. Segregation of candidate
molecular markers of the resistance was verified using three microsatellites (Sat 11, Sat 32 and
Sat 207) that are mapped onto the introgressed C. canephora chromosomal fragments T3, Tl and
T2 respectively.
All the seedlings (both Groups 1 and 2 sub-populations) were screened for CBD resistance after
one year by young seedlings inoculation method developed in this study. The method achieved a
degree of success that was considered to be sufficient for identification of DNA markers of the
resistance, despite of an expectation of some phenotypic misclassifications. Misclassification
was expected due to the observation that plants with low vigour (stunted and/or thin) exhibited
exceptionally high susceptibility including plants from Group 1 (resistant sub-populations) and
some plants of cv Caturra failed to be infected.
Fifty-seven (57) microsatellites were screened for polymorphism amongst accessions of cvs
Catimor, T5296, SL28 and the two F2 populations (D and E). Twenty three (23) microsatellites
were variously polymorphic within or between lineages. Seven microsatellites had alleles that
were common in the HDT derivatives, polymorphic in the two F2 populations and absent in cv
SL28. These were considered to be candidate markers of resistance to CBD. The seven
microsatellites were then analysed in 95 Group 2 plants from Population E for segregation
fitness and possible linkage to CBD resistance. Six of the microsatellites displayed segregation
ratios that fitted Mendelian inheritance but one microsatellite (Sat I I) had distorted segregation
in favour of the introgressed allele. It was further observed that Sat 207 and Sat 235 had marker
alleles that were linked to CBD resistance. The same plants were analysed for an AFLP marker
of the T4 fragment and it was observed to be present in 70.23% of the plants which suggested
that it followed random Mendelian inheritance and it did not co-segregate with CBD resistance.
Further confirmation that the markers were Iinked to CBD resistance, the seven potential
candidate microsatellites were amplified in fifty-six (56) Group 1 plants consisting of 29 and 27
individuals from Populations D and E respectively. These plants were also ana lysed with
selected AFLP markers of the introgressed fragments T2, T3 and T4. The fragment T2 was
confirmed to be linked to CBD resistance and further studies focussed it. Analysis was done with
AFLP markers spread on the T2 fragment in plants selected from the two F2 populations to cover
the two screening methods, resistant and susceptible phenotypes. Sat 235 that was observed to be
linked to CBD resistance was mapped using the same samples which had originally been used to
map the introgressed C. canephora fragments. The established limits of the location of the gene
confined it to a 26.9 cM segment, with high possibility of the gene to be within or near the limits
of a 10.6 cM segment. The segregation of Sat 207 and Sat 235 in 47 resistant and 18 susceptible
plants included in the 95 plants of Group 2 amplified earlier was re-examined with the mapping
information. It was observed that two resistant plants had the introgressed Sat 207 allele but not
the introgressed Sat 235 allele, while one susceptible plant without the introgressed Sat 207
allele had the introgressed Sat 235 allele. This prompted the assumption that the two markers
maybe located on the opposite sides of the gene. If this is proved to be true, then the gene is
located within a 13.2 cM chromosomal segment. No prominent skew in favour of homozygous
introgressed genotypes compared to the heterozygous ones was observed in the resistant
category of plants, indicating that the gene is of major action. It is therefore concluded that the
locus carries a major resistance gene that was designated Ck-l and is likely to be synonymous to
T gene described earlier by other researchers.
Four out of five AFLP markers of the introgressed C. canephora chromosomal fragment T2 were
successfully cloned, sequenced and specific primers designed. One primer pair amplified a
monomorphic band whose intensity in agarose gel was related to the presence and absence of the
parent AFLP marker at the theoretical optimum annealing temperature of 60°C. At a higher
annealing temperature of 62 °C, it amplified a dominant marker (AGC-CTG-cAA4). The SCAR
marker was analysed against Sat 207 and Sat 235 and it amplified as expected except in two
plants that were assumed to be recombinant
RAPD markers for CBD resistance identified earlier by other researchers could not be
reproduced, but specific primers designed from their sequences were tested in the F2 populations
by radioactive PCR and separated in denaturing polyacrylamide gels. One amplified a
monomorphic band in all accessions while the other amplified two polymorphic bands, one of
which was derived from HDT and it was linked to the T2 fragment. A survey of the
microsatellite markers for CBD resistance was carried out in twenty-two (22) accessions bred
from different accessions ofHDT and agreement with earlier results was demonstrated.
Ninety one (91) accessions of Coffea specres consisting of C arabica, its putative parents
namely C canephora (and its close relative C congensis] and C eugenioides (and its close
relative C. anthonyi) were analysed with eighteen (18) microsatellite markers of C canephora
chromosomal fragments introgressed into C arabica and seven (7) SCARs developed from
AFLP markers of some of the introgressed fragments. Different amplification characteristics of
the microsatellites and SCARs were observed in the different Coffea species. Un-introgressed C
arabica accessions exhibited low variability. In cases where two microsatellite alleles per
accession were amplified in C arabica, there was amplification in all the species analysed with
or without distinction between the canephoroid species (C canephora and C congensis) and
eugenioid species (C eugenioides and C anthonyi). In cases where the un-introgressed C
arabica had one allele per accession, there was no amplification in all or most of the eugenioid
species (C eugenioides and C anthonyi). Species specificity was also observed regarding some
SCAR alleles, but no null alleles observed in amplifications in this system. In all cases there was
an allele in canephoroid species (C canephora and C congensis) that was similar to the
introgressed allele in HDT derivatives in regard to both microsatellites and SCARs. Sat 235 had
no alleles shared between any of the un-introgressed C arabica accessions and the accessions of
the canephoroid group.
The maximum number of microsatellite alleles observed was seventeen and the minimum was
three alleles, while the maximum number of SCAR alleles was five and the minimum was one.
C. canephora had the highest number of alleles and the least polymorphic was the eugenioid
group (C eugenioides and C anthonyi). The un-introgressed C arabica accessions as a group
had more alleles than the introgressed ones despite the introgressed accessions having extra
alleles due to the introgression. In some cases, alleles similar to the marker alleles for
introgression were observed in some accessions of the un-introgressed accessions of C arabica.
In all cases, the genotypes of the HDT derivatives could be constituted by a combination of
alleles observed in C. arabica and the canephoroid group. The alleles of HDT that were shared
with the eugenioid group (c. eugenioides and C. anthonyi) were all observed in the unintrogressed
C. arabica accessions. In HDT derivatives, only one of their alleles was replaced
by the introgressed allele, even where there was more than one allele per accession of the unintrogressed
C. arabica.
Microsatellites with potential for use as breeding tools for CBD and CLR resistance from the
donor varieties Rume Sudan (resistant) and K7 (tolerant) were identified by their polymorphism
between these varieties and the susceptible cultivars SL28 and Caturra. However it was noted
that this potential would be attained by high performance techniques like LICOR fluorescence
system that was used in this phase of study.
Key words: Coffee Berry Disease, Colletotrichum kahawae, Coffea arabica, Coffea canephora,
Hibrido de Timor, introgression, resistance, chromosomal fragment, AFLP, Microsatellite,
marker, allele
Citation
Doctor of Philosophy in Plant Pathology University of Nairobi, 2007Publisher
University of Nairobi Department of plant science and crop protection