High-Throughput Analysis of T-DNA Location and Structure Using Sequence Capture

Abstract

Agrobacterium-mediated transformation of plants with T-DNA is used both to introduce transgenes and for mutagenesis. Conventional approaches used to identify the genomic location and the structure of the inserted T-DNA are laborious and high-throughput methods using next-generation sequencing are being developed to address these problems. Here, we present a cost-effective approach that uses sequence capture targeted to the T-DNA borders to select genomic DNA fragments containing T-DNA-genome junctions, followed by Illumina sequencing to determine the location and junction structure of T-DNA insertions. Multiple probes can be mixed so that transgenic lines transformed with different T-DNA types can be processed simultaneously, using a simple, index-based pooling approach. We also developed a simple bioinformatic tool to find sequence read pairs that span the junction between the genome and T-DNA or any foreign DNA. We analyzed 29 transgenic lines of Arabidopsis thaliana, each containing inserts from 4 different T-DNA vectors. We determined the location of T-DNA insertions in 22 lines, 4 of which carried multiple insertion sites. Additionally, our analysis uncovered a high frequency of unconventional and complex T-DNA insertions, highlighting the needs for high-throughput methods for T-DNA localization and structural characterization. Transgene insertion events have to be fully characterized prior to use as commercial products. Our method greatly facilitates the first step of this characterization of transgenic plants by providing an efficient screen for the selection of promising lines.

Fig 1. Probe design and workflow of the T-DNA capture.

A. Schematic illustration of the location of the hybridization probes designed to capture T-DNA-genome junctions. The black lines represent genomic DNA. Grey rectangles represent T-DNA sequences and light green rectangles represent the left and right border repeats (LB and RB). Black triangles indicate the nicking sites. 70 mer probes represented by magenta lines are designed to match 90 bp to 20 bp (5’ to 3’) inside of the nicking sites and are 5’ biotinylated (purple circles). B. Workflow of T-DNA capture and sequencing processes.

Fig 2. Single-end mapping of T-DNA.

A, B. Genome browser view of the T-DNA insertion sites found in sample pPLV26-Cas9_C4 (A) and pPLV26-Cas9_C3 (B). Red reads map to the Watson strand and blue reads map to the Crick strand. Coverage distribution tracks are positioned above (grey histograms). Schematic representations of the insert-T-DNA junctions are represented below (T-DNA in grey, border sequences in light green and genomic sequences in black). For pPLV26-Cas9_C3, the peak shown was the only one detected and only one end of the insertion was recovered.

Fig 3. Paired-end mapping of T-DNA.

A. Schematic view of read pairs that span the T-DNA—genome junctions. One read maps to the end of the T-DNA and the paired read maps to the genomic sequence around the T-DNA insertion site. B. Genome browser view showing junction reads mapping to the genomic DNA flanking the T-DNA insertion site in sample pPLV26-MeGI_2–3. C. Genome browser view of aligned reads from the same junction read pairs showing the end of the reads that are mapping to the T-DNA plasmid sequences. Elements in the T-DNA plasmid and the locations of probes for capture are shown below. The extra peak in tNOS downstream of the MeGI gene (black arrow) is due to the presence of another tNOS adjacent to the LB and the fact that Bowtie2 randomly selects among the best alignments when more than one is present. D. Schematic illustration of the inferred structure of the T-DNA insertion in sample pPLV26-MeGI_2–3 and primers that were used to confirm the insertion site. E. PCR confirmation of the T-DNA insertion structure in pPLV26-MeGI_2–3.