The rubber tree genome reveals new insights into rubber production and species adaptation

Abstract

The Para rubber tree (Hevea brasiliensis) is an economically important tropical tree species that produces natural rubber, an essential industrial raw material. Here we present a high-quality genome assembly of this species (1.37 Gb, scaffold N50 = 1.28 Mb) that covers 93.8% of the genome (1.47 Gb) and harbours 43,792 predicted protein-coding genes. A striking expansion of the REF/SRPP (rubber elongation factor/small rubber particle protein) gene family and its divergence into several laticifer-specific isoforms seem crucial for rubber biosynthesis. The REF/SRPP family has isoforms with sizes similar to or larger than SRPP1 (204 amino acids) in 17 other plants examined, but no isoforms with similar sizes to REF1 (138 amino acids), the predominant molecular variant. A pivotal point in Hevea evolution was the emergence of REF1, which is located on the surface of large rubber particles that account for 93% of rubber in the latex (despite constituting only 6% of total rubber particles, large and small). The stringent control of ethylene synthesis under active ethylene signalling and response in laticifers resolves a longstanding mystery of ethylene stimulation in rubber production. Our study, which includes the re-sequencing of five other Hevea cultivars and extensive RNA-seq data, provides a valuable resource for functional genomics and tools for breeding elite Hevea cultivars.

High-quality de novo assembly of the Eucommia ulmoides haploid genome provides new insights into evolution and rubber biosynthesis

“…S 18 ). This supports the rubber biosynthesis REF/SRPP gene families originating from multiple sources or from different evolutionary events, as demonstrated in previous studies 5 , 15 , 52 .…”

High-quality de novo assembly of the Eucommia ulmoides haploid genome provides new insights into evolution and rubber biosynthesis

“…S 18 ). This supports the rubber biosynthesis REF/SRPP gene families originating from multiple sources or from different evolutionary events, as demonstrated in previous studies 5 , 15 , 52 .…”

A Chromosome-Level Genome Assembly of Wild Castor Provides New Insights into its Adaptive Evolution in Tropical Desert

“…To investigate the phylogenetic position of castor in Euphorbiaceae species, especially the divergence time between wild and cultivated castors, we constructed a phylogenetic tree for five Euphorbiaceae species, including R. communis L. (WT05 and NSL4733), M. esculenta Crantz, J. curcas L., H. brasiliensis , and V. fordii , with Arabidopsis thaliana , Linum usitatissimum , and P. trichocarpa as outgroups, using 622 single-copy gene families. As expected, the wild castor is most closely related to cultivated castor ( Figure 3 A), and the tree topology is consistent with previous research [20] . To estimate the divergence time between wild and cultivated castors, we used the 10,906 collinear genes from a total of 722 syntenic blocks between two genomes to calculate the synonymous substitution rate (Ks) distribution, and the results showed peaks at 0.002 to 0.004.…”

De novo hybrid assembly of the rubber tree genome reveals evidence of paleotetraploidy in Hevea species

“…Consequently, the scaffolds that had been placed on the linkage map were associated predominantly with the euchromatic regions and encompassed a substantial portion (68%) of protein-coding sequences in the gene space. The percentage of the assembly assigned to the linkage map (~29%) appeared to represent the non-repetitive regions of the genome, which is congruent with previous and our findings that 69–72% of the rubber tree genome contained repetitive sequences 13 30 (see ‘ Repeat content in the genome assembly’ below). Furthermore, repeat content analyses of anchored and unanchored scaffolds revealed that, on average, unanchored scaffolds contained ~80% repeat elements compared to 44.5% repetitive content of scaffolds that had been placed on the linkage map.…”

“…In rubber tree, the average gene length (2,747 bp), average exon length (223 bp), and average number of exons per gene (4) were comparable to those in cassava 44 and castor bean 42 ( Supplementary Table S3 ). The GC contents of the coding sequences and the introns are 43.1% and 32.3%, similar to the previous estimates of 41.5% and 32.6% 30 , respectively ( Supplementary Table S3 ).…”

High-quality de novo assembly of the Eucommia ulmoides haploid genome provides new insights into evolution and rubber biosynthesis

“…S 18 ). This supports the rubber biosynthesis REF/SRPP gene families originating from multiple sources or from different evolutionary events, as demonstrated in previous studies 5 , 15 , 52 .…”

A Chromosome-Level Genome Assembly of Wild Castor Provides New Insights into its Adaptive Evolution in Tropical Desert

“…To investigate the phylogenetic position of castor in Euphorbiaceae species, especially the divergence time between wild and cultivated castors, we constructed a phylogenetic tree for five Euphorbiaceae species, including R. communis L. (WT05 and NSL4733), M. esculenta Crantz, J. curcas L., H. brasiliensis , and V. fordii , with Arabidopsis thaliana , Linum usitatissimum , and P. trichocarpa as outgroups, using 622 single-copy gene families. As expected, the wild castor is most closely related to cultivated castor ( Figure 3 A), and the tree topology is consistent with previous research [20] . To estimate the divergence time between wild and cultivated castors, we used the 10,906 collinear genes from a total of 722 syntenic blocks between two genomes to calculate the synonymous substitution rate (Ks) distribution, and the results showed peaks at 0.002 to 0.004.…”

De novo hybrid assembly of the rubber tree genome reveals evidence of paleotetraploidy in Hevea species

“…Consequently, the scaffolds that had been placed on the linkage map were associated predominantly with the euchromatic regions and encompassed a substantial portion (68%) of protein-coding sequences in the gene space. The percentage of the assembly assigned to the linkage map (~29%) appeared to represent the non-repetitive regions of the genome, which is congruent with previous and our findings that 69–72% of the rubber tree genome contained repetitive sequences 13 30 (see ‘ Repeat content in the genome assembly’ below). Furthermore, repeat content analyses of anchored and unanchored scaffolds revealed that, on average, unanchored scaffolds contained ~80% repeat elements compared to 44.5% repetitive content of scaffolds that had been placed on the linkage map.…”

“…In rubber tree, the average gene length (2,747 bp), average exon length (223 bp), and average number of exons per gene (4) were comparable to those in cassava 44 and castor bean 42 ( Supplementary Table S3 ). The GC contents of the coding sequences and the introns are 43.1% and 32.3%, similar to the previous estimates of 41.5% and 32.6% 30 , respectively ( Supplementary Table S3 ).…”

High-quality de novo assembly of the Eucommia ulmoides haploid genome provides new insights into evolution and rubber biosynthesis

“…S 18 ). This supports the rubber biosynthesis REF/SRPP gene families originating from multiple sources or from different evolutionary events, as demonstrated in previous studies 5 , 15 , 52 .…”

A Chromosome-Level Genome Assembly of Wild Castor Provides New Insights into its Adaptive Evolution in Tropical Desert

“…To investigate the phylogenetic position of castor in Euphorbiaceae species, especially the divergence time between wild and cultivated castors, we constructed a phylogenetic tree for five Euphorbiaceae species, including R. communis L. (WT05 and NSL4733), M. esculenta Crantz, J. curcas L., H. brasiliensis , and V. fordii , with Arabidopsis thaliana , Linum usitatissimum , and P. trichocarpa as outgroups, using 622 single-copy gene families. As expected, the wild castor is most closely related to cultivated castor ( Figure 3 A), and the tree topology is consistent with previous research [20] . To estimate the divergence time between wild and cultivated castors, we used the 10,906 collinear genes from a total of 722 syntenic blocks between two genomes to calculate the synonymous substitution rate (Ks) distribution, and the results showed peaks at 0.002 to 0.004.…”

De novo hybrid assembly of the rubber tree genome reveals evidence of paleotetraploidy in Hevea species

“…Consequently, the scaffolds that had been placed on the linkage map were associated predominantly with the euchromatic regions and encompassed a substantial portion (68%) of protein-coding sequences in the gene space. The percentage of the assembly assigned to the linkage map (~29%) appeared to represent the non-repetitive regions of the genome, which is congruent with previous and our findings that 69–72% of the rubber tree genome contained repetitive sequences 13 30 (see ‘ Repeat content in the genome assembly’ below). Furthermore, repeat content analyses of anchored and unanchored scaffolds revealed that, on average, unanchored scaffolds contained ~80% repeat elements compared to 44.5% repetitive content of scaffolds that had been placed on the linkage map.…”

“…In rubber tree, the average gene length (2,747 bp), average exon length (223 bp), and average number of exons per gene (4) were comparable to those in cassava 44 and castor bean 42 ( Supplementary Table S3 ). The GC contents of the coding sequences and the introns are 43.1% and 32.3%, similar to the previous estimates of 41.5% and 32.6% 30 , respectively ( Supplementary Table S3 ).…”