SSR cross-amplification and variation within coffee trees (Coffea spp.).

Primer sets were developed from 85 Coffea arabica sequences in addition to 25 already published primer sets. They were subsequently used for amplification in six African Coffea species: Coffea canephora (CAN), Coffea eugenioides (EUG), Coffea heterocalyx (HET), Coffea liberica (LIB), Coffea sp. Moloundou (MOL) and Coffea pseudozanguebariae (PSE). The amplification percentages for these 110 primer pairs ranged from 72.7% for LIB to 86.4% for PSE. Good transferability was thus obtained within the Coffea genus. When focusing on the two species CAN and PSE, high genetic diversity, high polymorphic locus rates (above 80%) and a mean allele number per polymorphic locus of more than 3 were noted. The estimated null allele percentage was -11% for PSE and -9% for CAN. Sixty three percent (CAN) and 79.5% (PSE) of the fixation index (Fis) values were positive. The within-species polymorphism information content (PIC) distribution showed two modes for both species. Although the two species shared 30 polymorphic loci, no correlation between CAN and PSE PIC values was obtained. All of these data are discussed in relation to the polymorphism level and the potential use of these SSRs for subsequent analysis of genetic diversity or genetic mapping.     

Isolation and genetic mapping of a Coffea canephora phenylalanine ammonia-lyase gene (CcPAL1) and its involvement in the accumulation of caffeoyl quinic acids

Biosynthesis of caffeoylquinic acids occurs via the phenylpropanoid pathway in which the phenylalanine ammonia-lyase (PAL) acts as a key-control enzyme. A full-length cDNA (pF6), corresponding to a PAL gene (CcPAL1), was isolated by screening a Coffea canephora fruit cDNA library and its corresponding genomic sequence was characterized. Amplification of total DNA from seven Coffea species revealed differences in intronic length. This interspecific polymorphism was used to locate the gene on a genetic map established for a backcross progeny between Coffea pseudozanguebariae and C. dewevrei. The CcPAL1 gene was found on the same linkage group, but genetically independent, as a caffeoyl-coenzyme A-O-methyltransferase gene, another gene intervening in the phenylpropanoid pathway. In the same backcross, a lower caffeoylquinic acid content was observed in seeds harvested from plants harbouring the C. pseudozanguebariae CcPAL1 allele. Involvement of the CcPAL1 allelic form in the differential accumulation of caffeoylquinic acids in coffee green beans is then discussed.     

Introgression of self-compatibility from Coffea heterocalyx to the cultivated species Coffea canephora

Self-compatibility segregation was assessed in two successive backcross progenies originating from an interspecific cross between Coffea canephora (self-incompatible) and Coffea heterocalyx (self-compatible). After self- and cross-pollination, pollen tube behaviour in styles was observed under ultraviolet fluorescence microscopy and fruit-set was determined at harvesting time. Segregation ratios in the two progenies were consistent with monofactorial control of self-compatibility. Self-compatible plants exhibited higher fruit-set than self-incompatible ones in open-pollination conditions. Segregation of AFLP markers was scored in the first backcross progeny. By molecular linkage analysis, the S locus could be mapped to a short linkage group.     

Pollen viability restoration in a <Emphasis Type="Italic">Coffea canephora</Emphasis> P. and <Emphasis Type="Italic">C. heterocalyx</Emphasis> Stoffelen backcross. QTL identification for marker-assisted selection

Male fertility of interspecific hybrids was analysed in one F1 and two backcrossed progenies originating from a cross between Coffea canephora and Coffea heterocalyx. Male fertility was tested using pollen stainability with acetic carmine. The results showed a marked decline in fertility at the F1 level, and fertility was almost fully restored after two backcrosses. The computed broad-sense heritability represented 47% of the variance. Quantitative trait loci (QTLs) locations and effects on pollen viability were estimated using an amplified fragment length polymorphism (AFLP) genetic linkage map constructed in the segregating BC1 population. Three significant QTLs (LOD>3 and p < 0.001 by ANOVA) were detected for pollen viability, two of which were responsible for the bimodal distribution of pollen viability in the segregating population. One QTL was involved in fertility variations among fertile BC1 plants. Fertility inheritance is discussed in relation with previously demonstrated chromosomal sterility in Coffea hybrids and the effect of detected QTLs. The potential use of genetic markers to overcome sterility in interspecific hybrids is also discussed.     

Two novel Ty1-copia retrotransposons isolated from coffee trees can effectively reveal evolutionary relationships in the Coffea genus (Rubiaceae)

In the study, we developed new markers for phylogenetic relationships and intraspecies differentiation in Coffea. Nana and Divo, two novel Ty1-copia LTR-retrotransposon families, were isolated through C. canephora BAC clone sequencing. Nana- and Divo-based markers were used to test their: (1) ability to resolve recent phylogenetic relationships; (2) efficiency in detecting intra-species differentiation. Sequence-specific amplification polymorphism (SSAP), retrotransposon-microsatellite amplified polymorphism (REMAP) and retrotransposon-based insertion polymorphism (RBIP) approaches were applied to 182 accessions (31 Coffea species and one Psilanthus accession). Nana- and Divo-based markers revealed contrasted transpositional histories. At the BAC clone locus, RBIP results on C. canephora demonstrated that Nana insertion took place prior to C. canephora differentiation, while Divo insertion occurred after differentiation. Combined SSAP and REMAP data showed that Nana could resolve Coffea lineages, while Divo was efficient at a lower taxonomic level. The combined results indicated that the retrotransposon-based markers were useful in highlighting Coffea genetic diversity and the chronological pattern of speciation/differentiation events. Ongoing complete sequencing of the C. canephora genome will soon enable exhaustive identification of LTR-RTN families, as well as more precise in-depth analyses on contributions to genome size variation and Coffea evolution.     

Genetic mapping of a caffeoyl-coenzyme A 3-O-methyltransferase gene in coffee trees. Impact on chlorogenic acid content

Chlorogenic acids (CGA) are involved in the bitterness of coffee due to their decomposition in phenolic compounds during roasting. CGA mainly include caffeoyl-quinic acids (CQA), dicaffeoyl-quinic acids (diCQA) and feruloyl-quinic acids (FQA), while CQA and diCQA constitute CGA sensu stricto (CGA s.s.). In the two cultivated species Coffea canephora and Coffea arabica, CGA s.s. represents 88% and 95% of total CGA, respectively. Among all enzymes involved in CGA biosynthesis, caffeoyl-coenzyme A 3-O-methyltransferase (CCoAOMT) is not directly involved in the CGA s.s. pathway, but rather in an upstream branch leading to FQA through feruloyl-CoA. We describe how a partial cDNA corresponding to a CCoAOMT encoding gene was obtained and sequenced. Specific primers were designed and used for studying polymorphism and locating the corresponding gene on a genetic map obtained from an interspecific backcross between Coffea liberica var. Dewevrei and Coffea pseudozanguebariae. Offspring of this backcross were also evaluated for the chlorogenic acid content in their green beans. A 10% decrease was observed in backcross progenies that possess one C. pseudozanguebariae allele of the CCoAOMT gene. This suggests that CGA s.s. accumulation is dependent on the CCoAMT allele present and consequently on the activity of the encoded isoform, whereby CGA accumulation increases as the isoform activity decreases. Possible implications in coffee breeding are discussed.     

Development of microsatellite markers for identifying Brazilian Coffea arabica varieties

Microsatellite markers, also known as SSRs (Simple Sequence Repeats), have proved to be excellent tools for identifying variety and determining genetic relationships. A set of 127 SSR markers was used to analyze genetic similarity in twenty five Coffea arabica varieties. These were composed of nineteen commercially important Brazilians and six interspecific hybrids of Coffea arabica, Coffea canephora and Coffealiberica. The set used comprised 52 newly developed SSR markers derived from microsatellite enriched libraries, 56 designed on the basis of coffee SSR sequences available from public databases, 6 already published, and 13 universal chloroplast microsatellite markers. Only 22 were polymorphic, these detecting 2-7 alleles per marker, an average of 2.5. Based on the banding patterns generated by polymorphic SSR loci, the set of twenty-five coffee varieties were clustered into two main groups, one composed of only Brazilian varieties, and the other of interspecific hybrids, with a few Brazilians. Color mutants could not be separated. Clustering was in accordance with material genealogy thereby revealing high similarity.     

Genome evolution in diploid and tetraploid Coffea species as revealed by comparative analysis of orthologous genome segments

Sequence comparison of orthologous regions enables estimation of the divergence between genomes, analysis of their evolution and detection of particular features of the genomes, such as sequence rearrangements and transposable elements. Despite the economic importance of Coffea species, little genomic information is currently available. Coffea is a relatively young genus that includes more than one hundred diploid species and a single tetraploid species. Three Coffea orthologous regions of 470-900 kb were analyzed and compared: both subgenomes of allotetraploid Coffea arabica (contributed by the diploid species Coffea eugenioides and Coffea canephora) and the genome of diploid C. canephora. Sequence divergence was calculated on global alignments or on coding and non-coding sequences separately. A search for transposable elements detected 43 retrotransposons and 198 transposons in the sequences analyzed. Comparative insertion analysis made it possible to locate 165 TE insertions in the phylogenetic tree of the three genomes/subgenomes. In the tetraploid C. arabica, a homoeologous non-reciprocal transposition (HNRT) was detected and characterized: a 50 kb region of the C. eugenioides derived subgenome replaced the C. canephora derived counterpart. Comparative sequence analysis on three Coffea genomes/subgenomes revealed almost perfect gene synteny, low sequence divergence and a high number of shared transposable elements. Compared to the results of similar analysis in other genera (Aegilops/Triticum and Oryza), Coffea genomes/subgenomes appeared to be dramatically less diverged, which is consistent with the relatively recent radiation of the Coffea genus. Based on nucleotide substitution frequency, the HNRT was dated at 10,000-50,000 years BP, which is also the most recent estimation of the origin of C. arabica.     

AFLP analysis of genetic diversity within and among Coffea arabica cultivars

Genetic diversity of Coffea arabica cultivars was estimated using amplified fragment length polymorphism (AFLP) markers. Sixty one Coffea accessions composed of six arabica cultivars, including Typica, Bourbon, Catimor, Catuai, Caturra and Mokka Hybrid, plus two diploid Coffea species, were analyzed with six EcoRI- MseI primer combinations. A total of 274 informative AFLP markers were generated and scored as binary data. These data were analyzed using cluster methods in the software package NTSYSpc. The differences among cultivars at the DNA level were small, with an average genetic similarity of 0.933. Most accessions within a cultivar formed a cluster, although deviant samples occurred in five of the six cultivars examined due to residual heterozygosity from ancestral materials. Among the six cultivars fingerprinted, the highest level of genetic diversity was found within the cultivar Catimor, with an average genetic similarity of 0.880. The lowest level was found within Caturra accessions, with an average genetic similarity of 0.993. Diversity between C. arabica and two other Coffea species, Coffea canephora and Coffea liberica, was also estimated with average genetic similarities of 0.540 and 0.413, respectively, suggesting that C. canephora is more closely related to C. arabica than is C. liberica. The genetic variation among arabica cultivars was similar to the variation within cultivars, and no cultivar-specific DNA marker was detected. Although arabica cultivars appear to have a narrow genetic base, our results show that sufficient polymorphism can be found among some arabica cultivars with a genetic similarity as low as 0.767 for genetic/QTL mapping and breeding. The assessment of genetic diversity among arabica cultivars provided the necessary information to estimate the potential for using marker-assisted breeding for coffee improvement.     

Base composition of Coffea AFLP sequences and their conservation within the genus

Amplified fragment length polymorphism (AFLP) is often used for genetic mapping and diversity analysis, but very little information is currently available on their sequence characteristics. Species-specific sequences were analyzed from a single Coffea genome (Coffea pseudozanguebariae) associated with clustered or nonclustered AFLP loci of known genetic position. Compared with the expressed sequence tag (EST) sequence composition, their AT content exhibited a bimodal distribution with AT-poor sequences corresponding mainly to putative coding sequences. AT-rich sequences, apart from the EST distribution, were usually clustered on the genetic map and might correspond to noncoding sequences. Conversion of these AFLP markers into sequence-characterized amplified region (SCAR) anchor markers allowed us to assess sequence conservation within Coffea species with respect to species relatedness.     

Diversity in coffee assessed with SSR markers: structure of the genus Coffea and perspectives for breeding.

The present study shows transferability of microsatellite markers developed in the two cultivated coffee species (Coffea arabica L. and C. canephora Pierre ex Froehn.) to 15 species representing the previously identified main groups of the genus Coffea. Evaluation of the genetic diversity and available resources within Coffea and development of molecular markers transferable across species are important steps for breeding of the two cultivated species. We worked on 15 species with 60 microsatellite markers developed using different strategies (SSR-enriched libraries, BAC libraries, gene sequences). We focused our analysis on 4 species used for commercial or breeding purposes. Our results establish the high transferability of microsatellite markers within Coffea. We show the large amount of diversity available within wild species for breeding applications. Finally we discuss the consequences for future comparative mapping studies and breeding of the two cultivated species.     

Comparison of the <Emphasis Type="Italic">Coffea canephora</Emphasis> and <Emphasis Type="Italic">C. arabica</Emphasis> karyotype based on chromosomal DNA content

Nuclear genome size has been measured in various plants, seeing that knowledge of the DNA content is useful for taxonomic and evolutive studies, plant breeding programs and genome sequencing projects. Besides the nuclear DNA content, tools and protocols to quantify the chromosomal DNA content have been also applied, expanding the data about genomic structure. This study was conducted in order to calculate the Coffea canephora and Coffea arabica chromosomal DNA content, associating cytogenetic methodologies with flow cytometry (FCM) and image cytometry (ICM) tools. FCM analysis showed that the mean nuclear DNA content of C. canephora and C. arabica is 2C = 1.41 and 2.62 pg, respectively. The cytogenetic methodology provided prometaphase and metaphase cells exhibiting adequate chromosomes for the ICM measurements and karyogram assembly. Based on cytogenetic, FCM and ICM results; it was possible to calculate the chromosomal DNA content of the two species. The 1C chromosomal DNA content of C. canephora ranged from 0.09 (chromosome 1) to 0.05 pg (chromosome 11) and C. arabica from 0.09 (chromosome 1) to 0.03 pg (chromosome 22). The methodology presented in this study was suitable for DNA content measuring of each chromosome of C. canephora and C. arabica. The cytogenetic characterization and chromosomal DNA content analyses evidenced that C. arabica is a true allotetraploid originated from a cross between Coffea diploid species. Besides, the same analyses also reinforce that C. canephora is a possible progenitor of C. arabica.     

Two-dimensional mapping as a tool for classification of green coffee bean species

Two species of the genus Coffea, Coffea arabica (Colombia) and Coffea canephora (Indiano Robusta) were analysed by two-dimensional (2-D) maps in order to obtain fingerprints of the expressed polypeptide chains and to determine which ones would characterize the two species. Green beans were milled under liquid nitrogen. A dry powder was produced by three different extraction protocols aimed at eliminating interfering substances (polyphenols). A reduced powder was produced by two successive extractions performed in acetone. Trichloroacetic acid (TCA; 10% w/v) and beta-mercaptoethanol (0.07% v/v) in acetone were used for the first extraction (a) and 10% w/v TCA in acetone was used for the second extraction (b). Proteins were then solubilized in a solution (40 microL per 1 mg powder) containing 7 M urea, 2 M thiourea, 3% w/v 3-(3-cholamidopropyldimethyl-amino)-1-propanesulfate, 1% v/v carrier ampholytes, 40 mM Tris, 5 mM tributylphosphine and 10 mM acrylamide as alkylating agent. Following incubation at room temperature for 1 hour and centrifugation (7000 rpm for 20 minutes), the supernatant was used for 2-D electrophoresis. The proteins were revealed by Sypro Ruby staining. Master maps of the five replicas of each species were compared by PDQuest analysis. The results of this differential proteome analysis were: sixteen proteins were expressed solely in C. canephora (var. Indiano Robusta) and five proteins were only found in C. arabica (var. Colombia). Another eight proteins were up-regulated in C. canephora (var. Indiano Robusta) in comparison to C. arabica (var. Colombia) and one was down-regulated in the same comparison. A number of these polypeptide chains were further characterized by mass spectrometry in the matrix-assisted laser desorption/ionisation-time of flight mode. Additionally, considering the low number of protein sequences of Coffea present in the databases we also investigated some spots with a more powerful tool, reversed phase-high-performance liquid chromatography-electrospray ionisation-tandem mass spectrometry, thus obtaining an internal peptide sequence. The general properties of the identified proteins are presented and discussed.     

Bee pollination and fruit set of Coffea arabica and C. canephora (Rubiaceae)

Self-sterile Coffea canephora and self-fertile C. arabica are important cash crops in many tropical countries. We examined the relative importance of insect, wind, and spontaneous self-pollination and the degree of self-fertility of these two coffee species in 24 agroforestry coffee fields in Indonesia. In both species, open pollination and cross pollination by hand led to the highest fruit set. Wind pollination (including self-pollination) led to 16% lower fruit set than open pollination in C. canephora and to 12.3% lower fruit set in C. arabica. Self-pollinated flowers and unmanipulated controls achieved an extremely low fruit set of 10% or less in the self-sterile species, and of 60% and 48%, respectively in the self-fertile species. These results constitute experimental evidence that cross pollination by bees causes a significant increase in fruit set of not only the self-sterile, but also the self-fertile coffee species. The practical implication is that coffee yield may be improved by managing fields for increased flower visitation by bees.     

Impact of the Coffea canephora gene introgression on beverage quality of <Emphasis Type="Italic">C. arabica</Emphasis>

Lines of Coffea arabica derived from the Timor Hybrid (hybrid between C. arabica and C. canephora) are resistant to coffee leaf rust (Hemileia vastatrix) and to the nematode Meloidogyne exigua. The introgression of C. canephora resistance genes is suspected of causing a drop in beverage quality. Coffee samples from pure lines, compared in a Trial 1, and from F1 hybrids and parental lines from a half-diallel trial in a Trial 2, were studied for beverage quality, chemical composition and amount of introgressed genetic material. Chemical analyses (caffeine, chlorogenic acids, fat, trigonelline, sucrose) were carried out with near-infrared spectrometry by reflectance of green coffee. The number of amplified fragment length polymorphic (AFLP) markers introgressed from the Timor Hybrid varied from 1 to 37 for the lines studied. There were significant differences between lines for all of the biochemical compounds analysed and for the acidity and the overall standard of the beverage. Two lines (T17927, T17924) were significantly poorer than the controls for sucrose and beverage acidity. T17924 also had more chlorogenic acids and was poorer for the overall standard. However, two highly introgressed lines, T17934 and T17931 (25 and 30 AFLP markers, respectively), did not differ from the non-introgressed controls. There were no correlations between the number of AFLP markers and the chemical contents or beverage attributes. Significant correlations were found between the performance of the parents and their general combining ability for beverage quality. It was concluded that it should be possible to find lines with both the desired resistance genes and good beverage quality. Selection can avoid accompanying the introgression of resistance genes with a drop in beverage quality.     

Micro-collinearity and genome evolution in the vicinity of an ethylene receptor gene of cultivated diploid and allotetraploid coffee species (Coffea)

Arabica coffee (Coffea arabica L.) is a self-compatible perennial allotetraploid species (2n=4x=44), whereas Robusta coffee (C. canephora L.) is a self-incompatible perennial diploid species (2n=2x=22). C. arabica (C(a) C(a) E(a) E(a) ) is derived from a spontaneous hybridization between two closely related diploid coffee species, C. canephora (CC) and C. eugenioides (EE). To investigate the patterns and degree of DNA sequence divergence between the Arabica and Robusta coffee genomes, we identified orthologous bacterial artificial chromosomes (BACs) from C. arabica and C. canephora, and compared their sequences to trace their evolutionary history. Although a high level of sequence similarity was found between BACs from C. arabica and C. canephora, numerous chromosomal rearrangements were detected, including inversions, deletions and insertions. DNA sequence identity between C. arabica and C. canephora orthologous BACs ranged from 93.4% (between E(a) and C(a) ) to 94.6% (between C(a) and C). Analysis of eight orthologous gene pairs resulted in estimated ages of divergence between 0.046 and 0.665 million years, indicating a recent origin of the allotetraploid species C. arabica. Analysis of transposable elements revealed differential insertion events that contributed to the size increase in the C(a) sub-genome compared to its diploid relative. In particular, we showed that insertion of a Ty1-copia LTR retrotransposon occurred specifically in C. arabica, probably shortly after allopolyploid formation. The two sub-genomes of C. arabica, C(a) and E(a) , showed sufficient sequence differences, and a whole-genome shotgun approach could be suitable for sequencing the allotetraploid genome of C. arabica.     

Gene introgression into Coffea arabica by way of triploid hybrids (C. arabica x C. canephora)

Interspecific triploid hybrid plants between the tetraploid species Coffea arabica L. and the diploid species C. canephora P. were backcrossed to C. arabica. Although characterised by a low production and an important fruit dropping, all attempted crosses (ie, 6) generated BC(1) progenies. Flow cytometric analysis of the nuclear DNA content revealed that most of the BC1 individuals were nearly tetraploid. Among the male gametes produced by the interspecific triploid hybrids, those presenting a high number of chromosomes appeared strongly favoured. Only pollen mother cells having nearly 22 chromosomes were effective, the others leading to deficient endosperm and fruit dropping. Molecular markers (ie, microsatellite and AFLP) combined with evaluations of morphological characteristics and resistance to leaf rust were applied to verify the occurrence of gene transfer from C. canephora into C. arabica, and to estimate the amount of introgression present in BC(1) individuals. The results reveal a strong deficiency in the C. canephroa alleles indicating a severe counter-selection against the introgression of genetic material from C. canephora into C. arabica by way of triploid hybrids. However, introgressants displaying desirable traits such as a high resistance to leaf rust were obtained. The low level of introgression could be an advantage by facilitating the recovery of the recurrent parent and possibly reducing the number of required backcrosses. On the other hand, this could be a limitation when attempting the transfer of a complex trait or several simply inherited traits.