A nomadic subtelomeric disease resistance gene cluster in common bean

Abstract

The B4 resistance (R) gene cluster is one of the largest clusters known in common bean (Phaseolus vulgaris [Pv]). It is located in a peculiar genomic environment in the subtelomeric region of the short arm of chromosome 4, adjacent to two heterochromatic blocks (knobs). We sequenced 650 kb spanning this locus and annotated 97 genes, 26 of which correspond to Coiled-Coil-Nucleotide-Binding-Site-Leucine-Rich-Repeat (CNL). Conserved microsynteny was observed between the Pv B4 locus and corresponding regions of Medicago truncatula and Lotus japonicus in chromosomes Mt6 and Lj2, respectively. The notable exception was the CNL sequences, which were completely absent in these regions. The origin of the Pv B4-CNL sequences was investigated through phylogenetic analysis, which reveals that, in the Pv genome, paralogous CNL genes are shared among nonhomologous chromosomes (4 and 11). Together, our results suggest that Pv B4-CNL was derived from CNL sequences from another cluster, the Co-2 cluster, through an ectopic recombination event. Integration of the soybean (Glycine max) genome data enables us to date more precisely this event and also to infer that a single CNL moved from the Co-2 to the B4 cluster. Moreover, we identified a new 528-bp satellite repeat, referred to as khipu, specific to the Phaseolus genus, present both between B4-CNL sequences and in the two knobs identified at the B4 R gene cluster. The khipu repeat is present on most chromosomal termini, indicating the existence of frequent ectopic recombination events in Pv subtelomeric regions. Our results highlight the importance of ectopic recombination in R gene evolution.

Figure 1.

Schematic representation of 650 kb of the Pv BAT93 B4 R gene locus sequence. A, Sequence annotation of the 48-B10 and FZ-E9 BAT93 BAC clones and of the 410-kb contig (four overlapping BAC clones: 75-H11, 254-G15, FY-N24, and 94-L19). Two gaps are indicated by dotted lines. The orientation of predicted genes is indicated by arrows. Putative gene function is listed on the right. Twenty-six CNL (turquoise arrows) are regrouped in four subclusters named A to D. The khipu tandem repeats (A–L) are in purple. Genetic positions are on the right (GP 1–GP 4), as described in the corresponding physical map presented by David et al. (2008). Genetic distances between each GP are indicated on the right in Kosambi cM. BAC clones FZ-E9 and 48-B10 have been mapped at GP 4 and 1, respectively. Orientations of these two BAC clones are unknown, so arbitrary orientations were chosen for this schematic representation. B, Sequence comparison between the Pv BAT93 B4 410-kb contig (center) and syntenic regions in Mt chromosome 6 (left) and Lj chromosome 2 (right). Yellow lines indicate significant homology matches between predicted genes. BAC clone position along the sequence is to scale. Genetic maps of these syntenic regions are presented on the left of the Mt syntenic region and on the right of the Lj syntenic region. Molecular markers and their genetic distances (cM) in Mt and Lj genetic maps are presented in boxes on the left and right, respectively.

Figure 1.

Schematic representation of 650 kb of the Pv BAT93 B4 R gene locus sequence. A, Sequence annotation of the 48-B10 and FZ-E9 BAT93 BAC clones and of the 410-kb contig (four overlapping BAC clones: 75-H11, 254-G15, FY-N24, and 94-L19). Two gaps are indicated by dotted lines. The orientation of predicted genes is indicated by arrows. Putative gene function is listed on the right. Twenty-six CNL (turquoise arrows) are regrouped in four subclusters named A to D. The khipu tandem repeats (A–L) are in purple. Genetic positions are on the right (GP 1–GP 4), as described in the corresponding physical map presented by David et al. (2008). Genetic distances between each GP are indicated on the right in Kosambi cM. BAC clones FZ-E9 and 48-B10 have been mapped at GP 4 and 1, respectively. Orientations of these two BAC clones are unknown, so arbitrary orientations were chosen for this schematic representation. B, Sequence comparison between the Pv BAT93 B4 410-kb contig (center) and syntenic regions in Mt chromosome 6 (left) and Lj chromosome 2 (right). Yellow lines indicate significant homology matches between predicted genes. BAC clone position along the sequence is to scale. Genetic maps of these syntenic regions are presented on the left of the Mt syntenic region and on the right of the Lj syntenic region. Molecular markers and their genetic distances (cM) in Mt and Lj genetic maps are presented in boxes on the left and right, respectively.

Figure 2.

Bayesian phylogenetic tree of BAT93 Pv B4-CNL sequences. This tree was constructed using the complete NBS domains (from P-loop to MHDV) of the 26 BAT93 B4-CNL nucleic acid sequences presented in Figure 1. Numbers on the branches represent posterior probabilities (only >70% are indicated) after running a Markov Chain Monte Carlo search for 10,000,000 generations. Physical positions of the B4-CNL along the B4 locus are indicated on the right. The soybean Rpg1-b sequence was used as an outgroup.

Figure 3.

Phylogenetic analysis of NL genes. This Bayesian tree was constructed using just the NBS domain (from the P-loop to the GLPL motif) of 133 amino acid sequences. Numbers at nodes indicate posterior probabilities (only >70% are indicated). Gene name abbreviations are as follows: Pv, Phaseolus vulgaris; Gmw, Glycine max Williams 82; Gmp, Glycine max line PI96983; Gtd, Glycine tomentella diploid accession G1403; Mt, Medicago truncatula var Jemalong; Lj, Lotus japonicus accession Miyakojima MG-20; At, Arabidopsis thaliana; Pt, Populus trichocarpa. Rpg1 orthologs, described by Innes et al. (2008), are regrouped into one clade; Pv B4-CNL (in turquoise) belong to the Rpg1 ortholog clade. Rpg1 orthologs are colored as described by Innes et al. (2008). The five Mt CNL clades (Mt CNL-1–Mt CNL-5) and eight poplar and Arabidopsis NL sequences are as described by Ameline-Torregrosa et al. (2008). All Pv NL sequences available in the nr database have been included in this analysis and are underlined on the tree. The tree was rooted using Pv TNL sequences. Sequence characteristics are reported in Supplemental Table S5.

Figure 4.

FISH to Pv BAT93 chromosomes. A and B, FISH to mitotic chromosomes. Signals are in red, and chromosomes are in blue. Chromosome 4 is indicated. A, ER-N1 BAC clone (David et al., 2008) was used as a probe. ER-N1 BAC, containing Pv B4-CNL sequences from subcluster C of the Pv B4 410-kb contig, is specific for the end of the short arm of chromosome 4. B, Khipu was used as a probe. A repetitive distribution pattern is observed, including strong signals on terminals of both chromosome 4 arms. C to I, FISH to pachytene chromosomes. C and D, All chromosome ends are tentatively indicated by arrowheads. Khipu was used as a probe in D. Note that although signal intensity was correlated with knob size in most chromosomes ends (white arrowheads), not all terminal knobs are labeled with the khipu probe (blue arrowheads). Two strong hybridization signals are detected on the end of the short arm of chromosome 4, which is tentatively indicated. E to I, Focus on the short arm of chromosome 4. The major knob and the minor knob are indicated with a white arrow and a white arrowhead, respectively. F and G, ER-N1 BAC clone was used as a probe and is shown in red superimposed on the DAPI-stained chromosome (F) or not (G). Note signals on both sides of the minor knob. H and I, Khipu was used as a probe and is shown in red superimposed on the DAPI-stained chromosome (H) or not (I). Note signals on minor and major knobs as well as on adjacent euchromatin. Bars = 10 μm.