Résumé
We characterized the insertion sites of newly transposed copies of the tissue-culture-induced ty1-copia retrotransposon Tos17 in the Oryza Tag Line (OTL) T-DNA mutant library of rice cv. Nipponbare. While Nipponbare contains two native copies of Tos17 the number of additional copies, deduced from Southern blot analyses in a subset of 384 T-DNA lines and using a reverse transcriptase probe specific to the element, ranged from 1 to 8 and averaged 3.37. These copies were shown to be stably inherited and to segregate independently in the progenies of insertion lines. We took advantage of the absence of EcoRV restriction sites in the immediate vicinity of the 3' LTR of the native copies of Tos17 in the genome sequence of cv. Nipponbare, thereby preventing amplification of corresponding PCR fragments, to efficiently and selectively amplify and sequence flanking regions of newly transposed Tos17 inserts. From 25,286 T-DNA plants, we recovered 19,252 PCR products (76.1%), which were sequenced yielding 14,513 FSTs anchored on the rice pseudomolecules. Following elimination of redundant sequences due to the presence of T-DNA plants deriving from the same cell lineage, these FSTs corresponded to 11,689 unique insertion sites. These unique insertions exhibited higher densities in subtelomeric regions of the chromosomes and hot spots for integration, following a distribution that remarkably paralleled that of Tos17 sites in the National Institute for Agrobiological Sciences (NIAS) library. The insertion sites were mostly found in genic regions (77.5%) and preferably in coding sequences (68.8%) compared to unique T-DNA insertion sites in the same materials (49.1% and 28.3%, respectively). Predicted non- transposable element (TE) genes prone to a high frequency of Tos17 integration (i.e. from 5 to 121 inserts) in the OTL T-DNA collection were generally found to be also hot spots for integration in the NIAS library. The 9,060 Tos17 inserts inserted into non TE genes were found to disrupt a total of 2,773 genes with an average of 3.27 inserts per gene, similar to that in the NIAS library (3.28 inserts per gene on average) whereas the 4,472 T-DNA inserted into genes in the same materials disrupted a total of 3,911 genes (1.14 inserts per gene on average). Interestingly, genes disrupted by both Tos17 and T-DNA inserts in the library represented only 14.9% and 10.6% of the complement of genes interrupted by Tos17 and T-DNA inserts respectively while 52.1% of the genes tagged by Tos17 inserts in the OTL library were found to be tagged also in the NIAS Tos17 library. We concluded that the first advantage in characterizing Tos17 inserts in a rice T-DNA collection lies in a complementary tagging of novel genes and secondarily in finding other alleles in a same genetic background, thereby greatly enhancing the library genome coverage and its overall value for implementing forward and reverse genetics strategies