Genome and CRISPR/Cas9 system of a widespread forest tree (Populus alba) in the world

Trees account for approximately 90% of the Earth's biomass and provide humans with various necessities for survival, such as clean air and water, wood, fibre and fuel (Petit and Hampe, 2006; Tuskan et al., 2006). Compared with annual plants, trees have many significant features, such as perennial growth, large size, secondary growth from a vascular cambium and dormancy (Douglas, 2017). Trees should be considered as a model system in plant biology and provide possibilities to answer questions that cannot be easily solved in the annual model systems of Arabidopsis and rice.

Trees of the genus Populus are prominent forest species in temperate regions of the Northern Hemisphere. Populus trees, as the model systems for plant biology, have several advantages, including rapid growth, small genome, facile transgenesis and easy cloning (Bradshaw et al., 2000). Populus trichocarpa is now widely used as a model system in the United States and Europe. Populus trichocarpa native to western North American cannot grow well in the fields and forests of China, which limits its breeding and application potential in China. In China, Populus tomentosa used to be a study system in basic research. It is a stabilized interspecific hybrid species widespread in Asia and is extensively used in breeding or forestry industries. Besides, poplar 84 K (Populus alba × P. glandulosa) and poplar 741 (Populus alba × (P. davidiana × P. simonii) × P. tomentosa) are also widely used as study systems in China. These two cultivars and P. tomentosa are not the most suitable strains for tree genetic study system due to their hybridization background.

The white poplar (P. alba) is a widespread forest tree in the world, distributed in Europe, Asia, North America, South America, Africa and Oceania (Figure 1a). Populus alba is not only a beautiful tree (Figure 1b) but also has extensive adaptability to different ecological environments (Stölting et al., 2015). As a fast growing tree, P. alba is included in the forest tree breeding programme in China and the European programme of forest genetic resources. Our previous study found that with the natural expansion of P. alba from Europe to China, the natural populations in China experienced a bottleneck effect. Average pooled heterozygosity value of P. alba populations in China was much lower than that in Italy and Hungary (Liu et al., 2019). As a pure and highly adaptable natural species, P. alba has been used as a hybrid parent in the breeding history of China. Due to the widespread distribution of P. alba in the Eurasian continent, using P. alba as a research system is conducive to the promotion of research results. Thus, P. alba in China is suitable as a study system for tree biology.

Figure 1

Open in figure viewerPowerPoint

Features of Populus alba v2.0 genome and CRISPR/Cas9 system (details in Text S1).

Populus alba v1.0 assembly was fragmented and comprised 1285 contigs with N50 of 1181 Kb, supplemented by two organellar genome fragments (Liu et al., 2019). Here, we present a new assembly based on chromosome conformation capture, while combining the previous Pacific Biosciences (PacBio) single-molecule real-time (SMRT) and whole-genome shotgun sequences. A total of 124.17 gigabases (Gb) of Illumina clean data were obtained for high throughput chromosome conformation capture (Hi-C) analysis (Table S1). Based on these data, contigs of the P. alba were clustered into 97 scaffolds with a final N50 of 22.7 Mb (Table S2). These 97 scaffolds include 19 chromosome-length pseudomolecules, covering 412.7 Mb of the genome sequence and 78 unplaced scaffolds, representing 3.7 Mb unintegrated sequences. The order of 19 chromosome-length pseudomolecules was designated according to their collinearity with the P. trichocarpa v4 genome (Figure 1c). All the sequences considered as mitochondrial and plastidal sources in P. alba v1.0 assembly were removed before scaffolding and were assembled into complete chloroplast and mitochondria genomes (Figures S1 and S2), separately.

The expected gene contents of Benchmarking Universal Single-Copy Orthologs (BUSCO), Core Eukaryotic Gene (CEGMA) and LTR Assembly Index (LAI) were detected to qualify the completeness of the genome assembly (Table S3). The results showed that 97.8% of eudicots, 98.4% of embryophyta and 98.4% of eukaryota orthologs could be detected in P. alba v2.0 genome (Figure 1d). The varying degrees of increase in evaluation results of BUSCO, CEGMA and LAI demonstrated an improvement in the continuity and completeness of the P. alba v2.0 genome compared with v1.0. The BUSCO and CEGMA completeness of P. alba v2.0 genome was close to that of P. trichocarpa v4. The LAI score was greater than 10, which indicated that the assembly had reached the level of forest reference quality.

By optimized annotation procedures relying on expression data and gene sets of related species, 34 010 predicted protein-coding gene models were obtained in annotation version 2.0. A total of 29 847 (87.76%) of them were functional annotated with at least one database (Figure 1e). Among all gene models, 21 143 (62.16%) could found segmental duplication derived paralogues (Figure 1f). Slightly more repetitive elements were found in P. alba v2.0 than v1.0 (Table S4). The vast majority of TEs are LTR-RT, dominated by Gypsy-type elements, accounting for 45.3% of all repetitive elements, followed by Copia-type with 17.9%.

Studies on A. thaliana and P. trichocarpa have shown that the centromere region is gene-sparse, low SNP density, Gypsy elements enriched and high cytosine methylated (Jiang et al., 2003; Natali et al., 2015; Weighill et al., 2019). Centromere position was predicted using wavelet-based genomic signal analysis described by Weighill et al. (2019) (Figure 1g–i; Figure S3). Except for 2, 7, 8, 15, 18 and 19, relative centromeres positions (centre bin (bp)/chromosome length(bp)) of the rest 13 pseudochromosomes in P. alba are similar to those in P. trichocarpa (Table S5; Figure S4). By mapping centromeric repeat sequences from P. trichocarpa and various plants onto v2.0 chromosomes, the position of centromere/pericentromere on chromosome 1, 3, 6, 9, 11, 12, 18 and 19 could be further confirmed (Figure S3).

Based on the standard Agrobacterium-mediated transformation method, we optimized the transformation procedures for the sequenced P. alba plant (Supplemental Materials and Methods). The explants were originally obtained from sterilized tender stems and then subcultured every 4 weeks for sustainable use. The reporter gene, β-glucuronidase (GUS), and the P. alba phytoene desaturase (PDS) gene were used to test transformation and genome editing efficiency, respectively. The leaf discs and stem segments of tissue culture seedlings were used for transformation. Zeatin (ZT) is an effective plant hormone for stimulating the division in non-meristematic tissues. To improve transformation efficiency, we assessed the calli regeneration rate of leaf discs and stem segments cultured in medium supplemented with 2 ~ 100 mg/L ZT. The results showed that a relative low concentration of ZT (2 ~ 10 mg/L) was sufficient for calli regeneration within 4 weeks (Figure 1j). The optimized ZT concentrations for leaf discs and stem segments were different. Following the protocol presented, we successfully induced the regeneration of sprout and root within 4 ~ 5 weeks and 10 ~ 12 weeks after co-culturing with Agrobacterium, respectively. The final transformation efficiency was 31.96%. The genome editing efficiency was 69.47% for PDS (Figure 1k,l). This efficient genetic transformation method will assist in the functional research of P. alba.