Genome-wide analysis of flanking sequences reveals thatTnt1 insertion is positively correlated with gene methylation in Medicago truncatula.

Abstract

From a single transgenic line harboring five Tnt1 transposon insertions, we generated a near-saturated insertion population in Medicago truncatula. Using thermal asymmetric interlaced-polymerase chain reaction followed by sequencing, we recovered 388888 flanking sequence tags (FSTs) from 21 741 insertion lines in this population. FST recovery from 14 Tnt1 lines using the whole-genome sequencing (WGS) and/or Tnt1-capture sequencing approaches suggests an average of 80 insertions per line, which is more than the previous estimation of 25 insertions. Analysis of the distribution pattern and preference of Tnt1 insertions showed that Tnt1 is overall randomly distributed throughout the M. truncatula genome. At the chromosomal level, Tnt1 insertions occurred on both arms of all chromosomes, with insertion frequency negatively correlated with the GC content. Based on 174 546 filtered FSTs that show exact insertion locations in the M. truncatula genome version 4.0 (Mt4.0), 0.44 Tnt1 insertions occurred per kb, and 19 583 genes contained Tnt1 with an average of 3.43 insertions per gene. Pathway and gene ontology analyses revealed that Tnt1-inserted genes are significantly enriched in processes associated with 'stress', 'transport', 'signaling' and 'stimulus response'. Surprisingly, gene groups with higher methylation frequency were more frequently targeted for insertion. Analysis of 19 583 Tnt1-inserted genes revealed that 59% (1265) of 2144 transcription factors, 63% (765) of 1216 receptor kinases and 56% (343) of 616 nucleotide-binding site-leucine-rich repeat genes harbored at least one Tnt1 insertion, compared with the overall 38% of Tnt1-inserted genes out of 50 894 annotated genes in the genome.