Tree Genetics & Genomes

, Volume 4, Issue 3, pp 379–390

Sub-genomic origin and regulation patterns of a duplicated WRKY gene in the allotetraploid species Coffea arabica

Authors

  • Anne-Sophie Petitot
    • Résistance des Plantes aux BioagresseursIRD–Institut de Recherche pour le Développement, UMR 186 IRD-Cirad-UM2
  • Anne-Claire Lecouls
    • Résistance des Plantes aux BioagresseursIRD–Institut de Recherche pour le Développement, UMR 186 IRD-Cirad-UM2
    • Résistance des Plantes aux BioagresseursIRD–Institut de Recherche pour le Développement, UMR 186 IRD-Cirad-UM2
Original Paper

DOI: 10.1007/s11295-007-0117-x

Cite this article as:
Petitot, A., Lecouls, A. & Fernandez, D. Tree Genetics & Genomes (2008) 4: 379. doi:10.1007/s11295-007-0117-x

Abstract

The extensively cultivated coffee species Coffea arabica is an allotetraploid resulting from a recent hybridization between two wild diploid Coffea species. We describe in this paper the first identification and functional assessment of homoeologous gene copies in C. arabica. When cloning the CaWRKY1 gene encoding a transcription factor of the WRKY superfamily associated with plant defense responses to pathogens (Ganesh et al. in Plant Sci 170:1045–1051, 2006), two distinct gene copies (CaWRKY1a and CaWRKY1b) were obtained from C. arabica. Southern blots experiments and phylogenetic analysis of the WRKY1 gene in related diploid Coffea species showed that CaWRKY1a and CaWRKY1b are homoeologous sequences in the allopolyploid coffee genome and are probably close descents of the extant Coffea canephora and C. eugenioides WRKY1 genes. To verify if CaWRKY1a and CaWRKY1b were both functional, gene expressions were monitored in C. arabica plants challenged by the rust fungus Hemileia vastatrix, the root-knot nematode Meloidogyne exigua, and after several abiotic treatments. Real-time quantitative polymerase chain reaction (PCR) assays showed that CaWRKY1 homoeologs were concomitantly expressed and displayed the same altered patterns of expression in leaves and roots during biotic and abiotic treatments. These results suggest that CaWRKY1a and CaWRKY1b were functionally retained in the coffee genome after allopolyploidization and that no functional divergence occurred between the duplicated genes in the C. arabica genome. This study provides the first molecular data on sub-genome-specific expression in allopolyploid coffee. The origin of the C. arabica sub-genomes is discussed with regards on the probable progenitors of this important crop species.

Keywords

AllopolyploidyCoffeaDisease resistanceGene regulationCoffee phylogeneticsWRKY transcription factor

Introduction

Coffee belongs to the large botanical family Rubiaceae, which includes tropical trees and shrubs growing in the lower storey of forests. Coffea is by far the most important member of the family economically, and Coffea arabica (Arabica coffee) accounts for over 70% of world coffee production. C. arabica is a tetraploid (2n = 4x = 44) and may have resulted from a natural hybridization between two wild diploids Coffea species (Carvalho 1952). Polyploids are common in certain plant and animal taxa, and the genetic and evolutionary consequences of genome duplication have been recently reviewed (Comai 2005). In particular for allotetrapolyploids, it is expected that most genes are present in two homoeologous forms, highly similar but non-identical. The gene redundancy may lead to gene silencing or to the functional divergence of duplicated genes (Adams and Wendel 2005; Chen and Ni 2006). Recent studies using microarray and quantitative gene expression analyses identified progenitor-dependent (or genome-specific) gene regulation in allotetraploid cotton (Udall et al. 2006; Yang at al. 2006) and in synthetic Arabidopsis allotetraploids (Wang et al. 2006a, b). In allotetraploid cotton, up to 43% of homoeologous genes appear to be differentially transcribed in leaves (Adams et al. 2003; Udall et al. 2006). Significant genome-specific regulation was evidenced in the cotton allotetraploids containing AADD genomes, where expressed sequence tags (ESTs) from the AA-sub-genome were predominantly accumulated in leaves (Udall et al. 2006) and in ovules during fiber cell development (Yang et al. 2006). In the hexaploid wheat, single strand conformational polymorphism (SSCP) analysis of expressed genes in leaves or roots revealed that a significant proportion of homoeologs (27 and 26%, respectively) were not expressed (Bottley et al. 2006). Therefore, genome-specific gene regulation may be a general consequence of polyploidization in many allopolyploid plants.

In C. arabica, no data at all are available about gene redundancy and about the relative contribution of duplicated gene pairs in the transcriptome. As of July 2007, the coffee ESTs collection in the public database (http://www.ncbi.nlm.nih.gov) contained only 1,000 C. arabica ESTs, mostly obtained from complementary DNA (cDNA) libraries of rust-challenged C. arabica leaves (Fernandez et al. 2004). Coffee leaf rust is caused by the fungus Hemileia vastatrix (Berkeley & Broome) and is one of the most destructive diseases of C. arabica (Bettencourt and Rodrigues 1988). Natural resistance of C. arabica varieties to leaf rust is conditioned by gene-for-gene interactions (Rodrigues et al. 1975; Bettencourt and Rodrigues 1988) and is expressed by a rapid hypersensitive cell death at the leaf infection sites (Rodrigues et al. 1975; Silva et al. 2002). Among the ESTs isolated, several genes showed up-regulation of transcript accumulation in coffee leaves after H. vastatrix challenge (Fernandez et al. 2004; Ganesh et al. 2006). One of them, DSS16 (renamed CaWRKY1), displayed homology to WRKY transcription factors genes. The WRKY transcription factors belong to a major group of DNA-binding proteins in plants and function as transcriptional activators and repressors in a number of developmental and physiological processes (Eulgem et al. 2000; Robatzek and Somssich 2002; Zhang et al. 2005). In particular, WRKY transcription factors have been associated with plant defense responses to biotic and abiotic stresses (reviewed in Ülker and Somssich 2004; Eulgem 2006). Recent studies have shown involvement of specific WRKY proteins in plant defense responses and SAR. For instance, constitutive expression of Arabidopsis WRKY18 or WRKY70 in transgenic plants leads to enhanced disease resistance to virulent pathogens (Chen and Chen 2002; Li et al. 2004; Wang et al. 2006a, b), and knock-out mutants of AtWRKY70 or AtWRKY33 showed increased susceptibility to necrotrophic fungal pathogens (AbuQamar et al. 2006; Zheng et al. 2006). Identification of WRKY genes that play important roles in plant defense responses in several plants suggests that the regulation specificities of some WRKY proteins might be conserved (Liu et al. 2004, 2005).

In C. arabica, the CaWRKY1 gene was differentially activated in rust-challenged leaves of resistant vs susceptible coffee plants, with a higher and earlier activation in resistant plants (Ganesh et al. 2006). Activation patterns of CaWRKY1 are coincident with fungal entry into the plant stomata and precede hypersensitive cell death. CaWRKY1 is closely related to members of group IIb of Arabidopsis WRKY proteins, including AtWRKY6. As AtWRKY6 was identified as an important downstream component of signaling pathways involved in resistance to pathogens and in senescence (Robatzek and Somssich 2001, 2002), we wanted to investigate whether CaWRKY1 also plays important roles in coffee defense responses.

As a first step towards the functional analysis of the CaWRKY1 protein, the objective of this study was to explore the genomic status of the CaWRKY1 gene in C. arabica. First, we predicted that two homoeologous CaWRKY1 loci should exist in the tetraploid coffee genome. To test this hypothesis, full-length cDNA (FL cDNA) and DNA sequences of the CaWRKY1 gene were cloned in C. arabica and compared with orthologous WRKY1 DNA sequences we obtained from a set of related diploid Coffea species. Second, to assess the genome-specific regulation of the homoeologous CaWRKY1 genes in C. arabica, we tested their expression patterns and quantified messenger RNAs (mRNAs) accumulation using real-time quantitative reverse-transcription polymerase chain reaction (qRT-PCR).

Materials and methods

Plant material—biotic and abiotic treatments

Coffee plants were grown in potting soil in a greenhouse (24°C day, 22°C night, 65% RH). The following species were used in this study: C. arabica var. Caturra, C. arabica var. IAPAR59, C. canephora T3561, C. eugenioides DA60, C. congensis CC54, C. liberica EB58, C. racemosa IA57, and C. humilis G59.

Biotic treatments were performed on 6-month-old plants. For rust assays, C. arabica var. Caturra leaves were challenged with H. vastatrix isolates eliciting an incompatible interaction (race VI) or a compatible interaction (race II) as described in Fernandez et al. (2004). Plants only sprayed with water were used as control. Biological samples originated from at least three independent experiments conducted in the greenhouse at different periods of the year. Root-knot nematode infection was performed by inoculating Meloidogyne exigua juveniles (J2) on the resistant C. arabica var. IAPAR59 and the susceptible var. Caturra as described in Lecouls et al. (2006). Root tips (5-mm long) were collected 2, 3, and 5 days after inoculation.

Abiotic treatments were performed on coffee leaves of the same physiological state than leaves used for rust inoculation. Wounding was performed by applying an average of seven transversal cuttings per half-leaf using scissors. Wounded leaves were collected 15, 30, and 60 min later. Non-wounded leaves collected on different plants were used as controls. Salicylic acid (SA) treatments were performed by infiltrating leaves with a 0.5 mM solution of SA using a needless syringe. In preliminary experiments, we tested several SA concentrations (from 25 μM to 2 mM), and we chose the highest dose that did not induce necrosis. Water-infiltrated (mock-control) and non-infiltrated leaves were used as controls. Leaves were collected 1, 3, and 7 h after treatment. Senescent leaves showing visible signs of yellowing were harvested from the lower part of Caturra plants. Two leaves per plant and four plants per experiment were used, and experiments were repeated twice.

Plant materials collected after treatment were immediately frozen by immersion in liquid nitrogen, pooled and stored at −80°C until RNA extraction.

Genomic DNA extraction, restriction endonuclease digestion, electrophoresis, and Southern blotting

Fresh coffee leaves were collected in the greenhouse and immediately frozen in liquid nitrogen. DNA extraction was performed using the DNEasy Plant minikit (Qiagen, France) following the manufacturer’s recommendations. For each plant sample, 10 μg of genomic DNA was digested with 50–60 units of the restriction enzymes EcoRI or PstI (Promega, France) with the addition of 5 mM spermidine per reaction for 16 h at 37°C. Restriction fragments were separated by electrophoresis in 0.8% agarose gels in Tris-acetate-EDTA (TAE) buffer. DNA fragments were blotted onto NylonN + membranes (Amersham, Les Ullis, France) by alkaline vacuum transfer (TE 80 TransVac, Hoefer Scientific Instruments, San Francisco, USA).

Labeling of probes and hybridization conditions

Specific oligonucleotides were designed from the DSS16 sequence and used to amplify plasmid cDNA insert by PCR. After purification using QIAquick PCR purification kit (Qiagen, France), the DNA probe (50 ng) was labeled with 32P-dCTP by random-priming, hybridized to membrane-bound DNA fragments (at 65°C overnight), and detected by autoradiography according to the manufacturer’s specifications (Megaprime kit and hybridization buffer, Amersham, France).

Full-length cDNA cloning by RACE experiments

Specific oligonucleotides (DSS16–5R and DSS16–3R) were designed from the DSS16 sequence (GenBank accession CF589188) for rapid amplification of cDNA ends (RACE) experiments (Table 1 and Fig. 1). 3′RACE and 5′RACE experiments were conducted by combining Omniscript RT kit (Qiagen) and SMART PCR cDNA synthesis kit (Clontech). The resulting PCR product was cloned into pGEM-T easy vector (Promega, France) and sequenced (Genome Express, France). Full length cDNA (FL cDNA) clones were obtained by PCR amplification using two oligonucleotides (CATW5 and CATW3) designed from 5′- and 3′RACE sequences. The resulting 2-kb PCR product was cloned and sequenced.
https://static-content.springer.com/image/art%3A10.1007%2Fs11295-007-0117-x/MediaObjects/11295_2007_117_Fig1_HTML.gif
Fig. 1

a Primer positions for CaWRKY1 gene cloning from the DSS16 sequence. bCaWRKY1a and CaWRKY1b genomic structures. Exons are indicated by boxes and introns are shown as broken lines. Arrows indicate the translational initiation sites, and the star indicates the aa insertion/deletion position. Note that the WRKY domain is interrupted by an intron

Table 1

List of primers used for CaWRKY1 gene cloning

Primer name

Primer sequence (5–3′)

DSS16-3R

TACTCAAACTCCAAATCCATTACAA

DSS16-5R

ACCTGATGCATTTGTGGGTTAGCCA

CATW5

CCTCTTTAGAATACTGCAGCCTGA

CATW3

TGAACATGTTAACTATGTTCAGCC

Cloning of genomic CaWRKY1 sequences

CaWRKY1 sequences were obtained by PCR experiments on C. arabica var. Caturra DNA using CATW5 and CATW3 primers. The PCR products (around 2.5 kb) were cloned and sequenced.

Quantitative gene expression analysis

RNA extraction and qRT-PCR were performed as described in Ganesh et al. (2006). Specific oligonucleotides were designed in the 3′-part of the CaWRKY1 cDNA sequences that discriminated CaWRKY1a and CaWRKY1b transcripts (Table 2). The same TaqMan probe (WRKY1-S) was used for detecting both amplicons. Primers were carefully checked for absence of cross-amplification on CaWRKY1a and CaWRKY1b plasmid clones at various concentrations. Quantitative PCR data analysis was achieved using the SDS software version 1.7 (Perkin-Elmer–Applied Biosystems). The threshold cycle (Ct) values of the triplicate PCRs were averaged, and the relative quantification of the transcript levels was performed using the comparative Ct method (Livak and Schmittgen 2001). Relative quantification related the PCR signal of the target transcript in each sample to that of the control sample at each time. For absolute quantification of cDNA molecules, the threshold cycle (Ct) values of the triplicate PCRs were averaged, and the copy number of each cDNA was estimated from calibration curves data obtained on calibrated amounts (102, 103, 104, and 105 copies) of purified plasmids bearing the cloned gene tested. The CaWRKY1a and CaWRKY1b gene copy numbers were normalized to the CaUbiquitin gene chosen as internal reference of gene expression. Linear regression analysis was used to calculate the correlation coefficient between the CaWRKY1a and 1b gene copy number obtained in each experiment and over all experiments.
Table 2

List of primers used in real-time quantitative PCR assays

Gene

Primer

Sequence

Start location

Amplicon size (bp)

CaWRKY1a

WRKY1a-F

TGCAACAAGGACAGCACCAG

1,654

40

 

WRKY1a-R

CGTGATCGCGGCCGT

1,718

 

 

WRKY1-S

CATCATTCGCTGACACGCTTAGCGC

 

 

 

 

[5′]6-FAM [3′]TAMRA

 

 

CaWRKY1b

WRKY1b-F

TGCAACAAGGACAGCACCAC

1,639

45

 

WRKY1b-R

TCAGCTGTGATCGCGGC

1,708

 

Cloning of WRKY1 sequences in related Coffea spp.

Fresh leaves of C. arabica var. Caturra, C. canephora, C. eugenioides, C. congensis, C. liberica, C. racemosa, and C. humilis were used for genomic DNA extraction (DNEasy Plant minikit). Partial genomic WRKY1 sequences were obtained by PCR experiments on DNA of each Coffea spp. using CATW5 and DSS16-5R primers (Fig. 1). The resulting 2-kb PCR products were cloned and sequenced.

Bioinformatic analysis of CaWRKY1 sequences

Homology to sequences present in international databases was searched using basic local alignment search tools (BLASTN and BLASTX) (Altschul et al. 1997). Sequences were aligned using the MegAlign tool contained in the Lasergene software (DNASTAR, Madison, WI, USA) using ClustalW algorithm (slow-accurate option). Search for specific protein domains were performed on Pfam database website (http://www.sanger.ac.uk/Software/Pfam).

Phylogenetic analysis of CaWRKY1 sequences

Parsimony analyses with unweighted, unordered characters were conducted with PAUP 4.0b10 (Swofford 2002) in Macintosh environment. Bootstrap analysis was performed using 1,000 replicates, and unrooted consensus trees were constructed.

Results

Full-length cDNA isolation of two CaWRKY1 genes

CaWRKY1 was originally isolated as a differentially expressed sequence fragment (DSS16) during infection by the rust fungus H. vastatrix (Fernandez et al. 2004). Specific oligonucleotides designed from the DSS16 sequence were used for 3′ and 5′ RACE cloning experiments. FL cDNA clones were further obtained, and two distinct FL cDNA sequences (named CaWRKY1a and CaWRKY1b) were identified (GenBank accessions DQ335599 and DQ335598, respectively). Clone CaWRKY1a (1,974 bp) and clone CaWRKY1b (1,960 bp) contained an open reading frame (ORF) encoding a predicted polypeptide of 573 amino acids (aa) and 572 aa, respectively (Fig. 2). The two sequences shared 97.5% nt identities and 98.3% aa identities. CaWRKY1a and CaWRKY1b nucleotide sequences varied by 12 base pairs insertions/deletions (indels) and 7 single nucleotide polymorphisms (SNPs) in their 5′-untranslated region (5′UTR; 128 and 116 bp, respectively). ORFs differed by 1 Indel and 18 SNPs leading to a total of 10 aa changes in the predicted polypeptide sequences (Fig. 2). The indel that we observed in the ORF sequence of CaWRKY1a and CaWRKY1b resulted from a microsatellite (3 bp repeated, GCA) repeat polymorphism (four and three repeats, respectively). Primers designed from this CaWRKY1 microsatellite (or simple sequence repeats, SSRs) amplified homologous loci in six Coffea species and was found polymorphic among C. canephora accessions (Poncet, personal communication).
https://static-content.springer.com/image/art%3A10.1007%2Fs11295-007-0117-x/MediaObjects/11295_2007_117_Fig2_HTML.gif
Fig. 2

Alignment of CaWRKY1a and CaWRKY1b deduced amino acid sequences. The open boxes indicate the domains corresponding to WRKY factors from group IIb (NLS nuclear localization signal; LZ leucine–zipper, WRKY domain). Arrowheads represent intron positions. The stars represent the zinc finger motif. Non-conserved aa are shown as gray boxes. Microsatellite position is indicated by black dots

Homology searches with the amino acid sequences of the FL cDNA clones revealed a highly significant similarity (BlastP E value = 3e–113) to proteins belonging to the Arabidopsis thaliana WRKY transcription factors subgroup IIb (Eulgem et al. 2000). Both CaWRKY1a and CaWRKY1b deduced proteins matched the same A. thaliana WRKY factor, AtWRKY6 (At1g62300). The CaWRKY1a and CaWRKY1b aa sequences were characterized by the presence of a single WRKY domain (from position aa 315 to position aa 367) containing the core motif WRKYGQK together with a C2H2-type zinc finger motif in the C-terminal region (Fig. 2). In addition, a potential leucine–zipper motif (LZ; from position aa 139 to position aa 181) and a nuclear localization signal (NLS) were identified (Eulgem et al. 2000). Search for other peptide motifs, such as the coactivator or repressor motifs described in other WRKY proteins (Eulgem et al. 2000; Xie et al. 2005) did not yield significant results.

Cloning of CaWRKY1a and 1b genes

Two distinct CaWRKY1 genomic sequences sharing 96.7% of nucleotide identity were obtained from C. arabica DNA. The two genes exhibited a similar structure and consisted of six exons and five introns (Fig. 1). Nucleotide sequences differed in numerous SNPs and in several indels mainly in the first and last intron. As for all A. thaliana WRKY genes (Eulgem et al. 2000), the WRKY domain-encoding region in CaWRKY1 was interrupted by an intron. Interestingly, the five intron positions were conserved in CaWRKY1 genes and some A. thaliana WRKY genes of subgroup IIb (namely At1g62300, At4g04450, and At4g22070).

Phylogenetic analysis of the coffee WRKY1 gene in the Coffea genus

To assess whether CaWRKY1a and CaWRKY1b were homoeologous sequences or paralogous sequences, we conducted a genetic diversity analysis of the coffee WRKY1 gene in the Coffea genus. Six Coffea species (C. canephora, C. eugenioides, C. congensis, C. liberica, C. racemosa, and C. humilis) closely related to C. arabica were chosen based on previous phylogenetic analyses conducted with several DNA markers (Lashermes et al. 1997, 1999; Cros et al. 1998). PCR experiments using CATW5 and DSS16-5R primers enabled amplification of a 2-kb fragment overlapping the 5′-non-transcribed region, most of the coding sequences and the 5 intronic sequences (Fig. 1). Depending on the coffee species, one or two homologous sequences of the CaWRKY1 gene were obtained from C. canephora (2), C. eugenioides (2), C. congensis (1), C. liberica (1), C. racemosa (2), and C. humilis (1) genomic DNAs (GenBank accessions DQ335600 to DQ335619). Alignment of the sequences (1,540-bp total length) showed that WRKY1 sequences essentially differed by numerous SNPs and distinct Indel events in the 5′UTR and the first two intronic sequences (Supplementary material, Fig. S1). WRKY1 nucleotide sequences similarity ranged from 94.4 to 99.3% between Coffea species (data not shown). The highest similarities (99.9%) were observed between CaWRKY1a and canephora1 WRKY1 sequences and between CaWRKY1b and eugenioides1–1 WRKY1 sequences. Alignment of the deduced amino acid sequences (324 aa in total) showed a strict conservation of CaWRKY1a and CaWRKY1b with C. congensis and C. eugenioides WRKY1, respectively. CaWRKY1a displayed only one amino acid difference with C. canephora WRKY1 (data not shown).

Because indels (Table 3) may severely bias phylogenetic inferences derived from molecular sequences (Felsenstein 1981), the DNA sequence (including introns) of each clone was cleared of indels for phylogenetic analyses. Parsimony analysis conducted on the resulting DNA sequences (1,475 characters, 45 parsimony-uninformative, 48 parsimony-informative) produced four parsimonious trees requiring 106 evolutionary steps with a consistency index of 0.95, indicating a low level of homoplasy. The phylogenetic tree obtained clearly separated the two C. arabica WRKY sequences into distinct phylogenetic clades (Fig. 3). CaWRKY1a grouped with the C. canephora and C. congensis sequences while CaWRKY1b was closely related to the C. eugenioides sequences. The highest bootstrap value (100%) was assigned to the major branches. In addition, all but one indels (indel no. 14) manually assigned on the cladogram clearly supported the tree topology (Fig. 3). The close relationships observed between the C. arabica, C. canephora, and C. eugenioides WRKY1 sequences strongly suggest that CaWRKY1a and CaWRKY1b are homoeologous sequences resulting from the allopolyploidization event that gave rise to the C. arabica species.
https://static-content.springer.com/image/art%3A10.1007%2Fs11295-007-0117-x/MediaObjects/11295_2007_117_Fig3_HTML.gif
Fig. 3

Unrooted phylogenetic tree (maximum parsimony) of WRKY1 genes in Coffea species. Numbers above branches are bootstrap values (100 replicates). Dark bars (indels mutations) were manually assigned to each branch but were not taken into account for the phylogenetic analysis

Table 3

Sequence and location of indels in the WRKY1 gene sequences of Coffea sp.

Indel no.

Type

Sequence

Location

CaWRKY1b

C. eugenioides

C. liberica

C. humilis

C. racemosa

C. congensis

C. canephora

CaWRKY1a

 

1

2

 

 

1

2

 

1

2

 

1

Insertion

GG

5′UTR

+

+

+

+

2

Insertion

AAGATC

5′UTR

+

3

Deletion

TAG

5′UTR

+

+

3′

Deletion

T

5′UTR

+

+

+

+

+

4

Insertion

GCT

5′UTR

+

4′

Insertion

ACT

5′UTR

+

5

Insertion

AACTTGATA

5′UTR

+

+

6

Insertion

GTCT

1st intron

+

+

+

+

6′

Insertion

GTC

1st intron

+

7

Insertion

A

1st intron

+

+

+

+

8

Insertion

T

1st intron

+

+

+

+

9

Deletion

CA

1st intron

+

+

+

+

10

Deletion

T

1st intron

+

+

+

+

11

Deletion

GCTTACC

1st intron

+

+

+

+

12

Deletion

AT

2nd intron

+

13

Deletion

TTAGTTAA

2nd intron

+

+

14

Deletion

GAATTTAGA

5th intron

+

+

+

+

15

Deletion

TA

5th intron

+

+

16

Insertion

CTA

4th intron

+

+ Presence of indel; − absence of indel

Genomic organization of the CaWRKY1 genes in the Coffea sp. genome

To verify that the CaWRKY1 gene corresponded to a single copy gene, we conducted Southern blot experiments on coffee genomic DNAs. Coffee DNA samples were digested with the restriction enzyme EcoRI, for which no site was detected in the CaWRKY1 sequence and with the PstI enzyme (one site in the DSS16 sequence; three sites in the CaWRKY1 sequence, Fig. 1). Depending on the coffee-species tested, one or two EcoRI-digested DNA fragments strongly hybridized to the CaWRKY1 probe (Fig. 4). CaWRKY1 probe detected two fragments in EcoRI-digested DNAs of C. liberica and C. racemosa and one sharp band (about 4 kb) in other Coffea species. An additional weakly hybridizing 1,300-bp band was detected in all coffee DNAs tested. The DSS16 probe contains a small part of the 3′end of the conserved WRKY DNA-binding domain (11 bp) and might therefore weakly hybridize to another coffee WRKY gene. In PstI-digested DNAs, the radiolabeled EST probe hybridized to two fragments of the expected sizes (1,900 and 300 bp, respectively) for all species (Supplementary material, Fig. S3). Although the possibility of duplicated genes cannot be ruled out, our data suggest that the WRKY1 gene may be present as one copy (per haploid genome) in the genome of coffee species. In addition, EcoRI restriction fragment length polymorphisms (RFLPs) were detected between the seven coffee species tested (Fig. 4). C. liberica and C. racemosa displayed higher-sized hybridizing fragments than C. arabica, C. canephora, C. eugenioides, C. congensis, and C. humilis, suggesting allelic variation around the WRKY1 locus.
https://static-content.springer.com/image/art%3A10.1007%2Fs11295-007-0117-x/MediaObjects/11295_2007_117_Fig4_HTML.gif
Fig. 4

Genomic DNA of C. arabica var. Caturra, C. liberica, C. congensis, C. canephora, C. eugenioides, C. racemosa, and C. humilis digested with EcoRI and probed with 32P-labeled DNA fragment of CaWRKY1. Arrows on the left indicate the estimated DNA band sizes (in bp)

CaWRKY1a and CaWRKY1b expression patterns

Quantitative real-time PCR was used to investigate the respective contribution of CaWRKY1a and 1b genes to the C. arabica transcriptome and more specifically during plant defense responses. Transcript accumulation of each gene was monitored during plant infection with the coffee rust pathogen, the root-knot nematode M. exigua, and upon several abiotic treatments. Mock-inoculated plants or non-treated plants were used as controls to assess the relative expression level of each gene under each treatment. The Ubiquitin gene chosen as internal reference of gene expression was assayed in parallel with the candidate genes. Absolute quantification of the CaWRKY1a and CaWRKY1b mRNA levels in coffee leaves or roots allowed assessing the transcript copy number of each gene in each sample.

Results showed that both CaWRKY1a and CaWRKY1b genes were induced by rust infection, root-knot nematode infection, in senescent leaves, by wounding or SA treatments (Fig. 5). For all treatments, similar amounts of CaWRKY1a and CaWRKY1b transcripts were detected in each sample (Table 4). A statistically significant correlation (Pearson R = 0.92, p < 0.001) between the expression of the two genes was obtained across 44 data points measured (including rust, SA, and wounding responses), suggesting that CaWRKY1a and CaWRKY1b genes are coregulated (Supplementary material, Fig. S2).
https://static-content.springer.com/image/art%3A10.1007%2Fs11295-007-0117-x/MediaObjects/11295_2007_117_Fig5_HTML.gif
Fig. 5

Relative expression level of CaWRKY1a and CaWRKY1b coffee genes measured in C. arabica, C. canephora, and C. eugenioides after wounding (30 min leaf treatment), and in C. arabica after rust inoculation (avirulent H. vastatrix, 17 h post-inoculation), nematode inoculation (avirulent M. exigua, 3 days post-inoculation), in senescing leaves and after 3 h of salicylic acid (SA) treatment. Amounts of cDNAs were calibrated using the CaUbiquitin gene as reference. Gene expression in the treated plants was relative to that of the control plants (settled to 1)

Table 4

Example of CaWRKY1a and CaWRKY1b transcript quantification (copy number.μg−1 RNA) in C. arabica leaves after rust inoculation, salicylic acid (SA) treatment, and wounding

Treatment

CaWRKY1a

CaWRKY1b

Ubiquitin

None

5.6E + 04

3.3E + 04

1.6E + 09

Rust (12 h)

 Mock-inoculated

6.6E + 04

2.2E + 04

1.0E + 09

 Susceptible

2.6E + 05

1.5E + 05

1.5E + 09

 Resistant

1.1E + 06

1.3E + 06

9.7E + 08

SA (3 h)

 Water-infiltrated

1.5E + 05

1.3E + 05

7.9E + 08

 SA-infiltrated

4.4E + 05

4.9E + 05

9.0E + 08

Wounding (1 h)

 Control leaf

7.7E + 04

6.9E + 04

1.3E + 09

 Wounded leaf

2.8E + 06

4.6E + 06

5.5E + 08

In rust assays, time-course experiments were conducted using the Caturra variety challenged with H. vastatrix isolates either eliciting an incompatible interaction (resistance) or a compatible interaction (susceptibility). The relative changes in gene expression showed that CaWRKY1a and CaWRKY1b genes were induced between 12 and 16 h post-inoculation, depending on time-course experiments. Statistically significant differences (p < 0.05) in the relative expression of the CaWRKY1 genes were found between the compatible and incompatible interactions.

CaWRKY1a and CaWRKY1b genes were also differentially activated by M. exigua infection in the resistant and susceptible coffee varieties. At 3 days after inoculation, CaWRKY1 genes were activated (mean, 1.7-fold) in the resistant coffee plants and down-regulated (mean 0.8- fold) in the susceptible plants.

For abiotic treatments, the highest activation of CaWRKY1 genes was found in senescing leaves and in leaves wounded for 30 min (Fig. 5). After 1 h wounding, CaWRKY1 activation peak reached up to 100-fold (data not shown).

To verify whether regulation of CaWRKY1 was conserved in the closest C. arabica diploid relatives, wounding experiments were performed on C. canephora and C. eugenioides. C. canephora and C. eugenioides WRKY1 transcripts were detected using CaWRKY1a and CaWRKY1b primers, respectively. Results presented in Fig. 5 showed that a 30-min wounding treatment activated C. canephora and C. eugenioides WRKY1 gene expression.

Discussion

The aim of this study was to identify homoeologous copies of the CaWRKY1 gene in the C. arabica genome and to assess their genetic relatedness and functionality. Using primers designed from the DSS16 sequence previously isolated (Fernandez et al. 2004), we conducted a systematic cloning of FL cDNA or DNA sequences of the WRKY1 gene in C. arabica and six other coffee species. Among the DNA sequences obtained from allotetraploid C. arabica, we identified two distinct copies of the CaWRKY1 gene named CaWRKY1a and CaWRKY1b. In the diploid coffees, we isolated one or two WRKY1 allelic sequences, depending on the species. Southern blot analysis (Fig. 4) and CaWRKY1-derived SSR analysis in Coffea spp. (Poncet, personal communication) showed that the WRKY1 gene was located at a single locus in the Coffea spp. genome. The close phylogenetic relationships observed between CaWRKY1 and the orthologous WRKY1 from other diploid Coffea species (Fig. 3) strongly suggested that CaWRKY1a and CaWRKY1b are homoeologous sequences originating from each of the two parental sub-genomes of the tetraploid Arabica coffee species.

Molecular data obtained so far from the coffee genome suggested that the Arabica coffee species formed recently and that low divergence occurred between the two constitutive genomes of C. arabica and those of its progenitor species (Lashermes et al. 1999). Among the extant diploid coffee species (2n = 2x = 22), C. congensis, C. eugenioides, and C. canephora are thought to be most closely related to the tetraploid species (Raina et al. 1998; Lashermes et al. 1999). A higher level of polymorphism was thus expected between the C. arabica homoeologous sequences than among them and their ancestor Coffea sequences. In this study, we indeed isolated in the C. arabica genome two homoeologous copies of the CaWRKY1 gene that were distantly-related, sharing only 96% nt identity. The phylogenetic analysis of the WRKY1 gene in the Coffea genus showed that the two arabica WRKY1 homoeologs were genetically closely related to some diploid coffee WRKY1 genes (Fig. 4).

The origin of the C. arabica genome has been explored using cytological analyses coupled to genomic in situ hybridization (Raina et al. 1998; Lashermes et al. 1999). Data showed that C. arabica is an amphidiploid formed by hybridization between two wild diploid species, including C. eugenioides, and a member of the canephoroid species, either C. congensis (Raina et al. 1998) or C. canephora (Lashermes et al. 1999). In our study, we found that CaWRKY1b was genetically very close to the C. eugenioides WRKY1 gene, and a strict conservation of WRKY1 protein sequences was retained. The CaWRKY1a homoeolog grouped together with the two closely related C. canephora and C. congensis WRKY1 sequences. A higher similarity was observed between CaWRKY1a and canephora WRKY1 sequences (99.9%). In contrast, alignment of the deduced amino acids sequences showed a strict conservation of CaWRKY1a with the C. congensis WRKY1 sequence and one aminoacid difference with the C. canephora WRKY1 sequence (99.7% identity). Whether C. canephora or C. congensis is the closest relative of C. arabica may still be questionable. As discussed in Lashermes et al. (1999), the taxonomic boundaries among the diploid species belonging to the canephoroid group are unclear, as highly fertile hybrids can be easily obtained by crossing species (Berthaud and Charrier 1988). The Coffea accessions we used in this study were carefully chosen among the IRD coffee genetic resources available to best fit the current species description. At the present time, wild C. arabica populations are only found in some highlands of Ethiopia, Kenya, and Sudan, where lies the primary center of genetic diversity of the species. In contrast, diploid species are either restricted to Central African areas (C. congensis), highlands of Uganda and neighboring countries (C. eugenioides), or widely distributed in the tropical lowlands forests of West and Central Africa (C. canephora), but never coexist with wild C. arabica (Berthaud and Charrier 1988). Paleoecology records indicate that climate changes affected the distribution of plant species in African tropical forests during the late quaternary period (Prentice et al. 2000). As suggested by Lashermes et al. (1999), the present distribution of C. arabica may thus rather reflect its preservation in a refuge area than the geographic origin of the speciation. Interspecific hybridization may have taken place in the past in other areas such as Uganda, where sympatric populations of C. canephora and C. eugenioides are still found (Thomas 1944; Byesse, personal communication). The diploid accessions we used in this study originated from Central African areas (C. canephora and C. congensis) and Kenya (C. eugenioides). When available, it would be interesting to analyze some wild C. canephora accessions from Uganda to compare their genetic relatedness with C. arabica.

In C. arabica, resistance to leaf rust and nematodes (M. exigua) is expressed by an hypersensitive reaction (HR) at the infection sites (Silva et al. 2002; Anthony et al. 2005). Plant cell death may occur as soon as 24-h post-inoculation (p.i.) and parasite growth usually ceased in the early stages of the infection process (Silva et al. 2002; Anthony et al. 2005). Microarray analysis in model plants such as A. thaliana showed that plant disease resistance involves the reprogramming of cellular functions to fight off pathogen attacks. A large number of genes that mediate resistance signaling and defense reactions are subject to transcriptional regulation (Maleck et al. 2000; Tao et al. 2003). The activities of the genes may be up- or down-regulated and follow defined temporal programs. In C. arabica, we showed that CaWRKY1 mRNA levels are selectively increased during resistance to H. vastatrix or M. exigua and wounding or SA treatments. CaWRKY1a and CaWRKY1b were concomitantly expressed and displayed the same altered patterns of expression in leaves and roots during coffee responses to rust, nematodes, and other abiotic treatments. Both CaWRKY1 homoeologous genes therefore contributed to the transcriptomic expression of coffee defense responses to pathogens, evidencing that the duplicated genes are functional in the C. arabica genome.

In allopolyploid species, gene redundancy may lead to loss of function of excess copies, through wholesale deletion or mutation to pseudogenes, or more subtly, by epigenetic silencing of homoeologs (Adams and Wendel 2005). Alternatively, gene duplication may be resolved by subfunctionalization within the ancestral gene pair. Altered gene expression, including silencing and up- or down-regulation of one of the duplicated genes, may be common in natural allopolyploids, and the pattern of this bias has been shown to be organ-specific in cotton and wheat (Adams et al. 2003; Bottley et al. 2006; Udall et al. 2006). In C. arabica, similar amounts of CaWRKY1a and CaWRKY1b transcript levels were found in C. arabica leaves and roots under all treatments, suggesting that the homoeologous genes undergo the same transcriptional control. Considering that C. arabica resulted from a recent hybridization (Lashermes et al. 1999), the apparently similar expression pattern of the two genes is perhaps not so surprising. It may be possible that the two inherited WRKY1 gene copies were functionally conserved after the polyploidization event that gave rise to the C. arabica species. However, further experimental validation on their function would be required to verify that both genes are truly functional homologs.

Recent studies provided evidence that WRKY proteins have regulatory functions in plant responses to pathogen infection (Eulgem 2006). An ortholog of CaWRKY1 is AtWRKY6, which is involved both in the activation of plant defense responses and in senescence-related processes in A. thaliana (Robatzek and Somssich 2001). AtWRKY6 was highly expressed in senescent leaves and after P. syringae inoculation or flagellin treatments (Robatzek and Somssich 2001). In addition, the WRKY6 protein acted as a positive regulator of PR1 expression (Robatzek and Somssich 2002). AtWRKY6 overexpression lines displayed enhanced NPR1 and PR1 mRNAs accumulation, and several W boxes were found within the PR1 promoter (Robatzek and Somssich 2002). The PR1 upstream regulator NPR1 is positively regulated by several WRKY in A. thaliana (Yu et al. 2001) and AtWRKY6 may directly, or indirectly, participate to its regulation. In addition to rust- and nematode-activated expression, CaWRKY1 homoeologs displayed the same patterns of expression than AtWRKY6 under SA treatment, wounding, or in senescent leaves (Robatzek and Somssich 2001), suggesting that the function of this WRKY transcription factor in several physiological traits may be conserved across plant genera.

Future work will aim at understanding the role of the CaWRKY gene in the mechanisms of C. arabica resistance to pathogens. Additionally, several other WRKY genes have been identified in the coffee plant, with potential implication both in pathogen resistance and developmental processes (Fernandez et al. 2006). Coffee is an important cash crop, and thus genetic insights into genes and their regulation specificities have the potential to be directly exploited to assist genetic improvement programs and to better manage coffee genetic resources.

Acknowledgements

We would like to thank Dr. F. Anthony (IRD) for helpful discussions on Coffea spp. for phylogenetic analysis. We are greatly indebted to the research center on coffee rusts (CIFC, Oeiras, Portugal) for providing rust isolates. This work was partially supported through a bilateral cooperation between France and Portugal (contract 714 C2, 2003–2004 and Pessoa, 2007–2008) and between France and Brazil (CAPES-COFECUB n° Sv 555/07).

Supplementary material

11295_2007_117_MOESM1_ESM.doc (46 kb)
Fig. 1(DOC 46 kb)
11295_2007_117_MOESM2_ESM.doc (28 kb)
Fig. 2(DOC 28 kb)
11295_2007_117_MOESM3_ESM.doc (1.3 mb)
Fig. 3(DOC 1356 kb)

Copyright information

© Springer-Verlag 2007