Popis: |
BACKGROUND DNA uptake is the first step in natural transformation of bacteria, leading to DNA internalization and recombination. It is, therefore, a key determinant in genome evolution. Most bacteria take up DNA indiscriminately, but in two families of Gram-negative bacteria the uptake machinery binds preferentially to short sequences called uptake signal sequences (USS). These sequences are highly enriched in their genomes, which causes preferential uptake of self-DNA over foreign DNA. RESULTS To fully characterize the effects of this preference, and to identify other sequence factors affecting uptake, we carried out a genome-wide analysis of DNA uptake using both measured uptake and the predictions from a sequence-based uptake model. Maps of DNA uptake were developed by recovering and deep sequencing genomic DNA that had been taken up by competent Haemophilus influenzae cells, and comparing sequencing coverage from recovered samples to coverage of the input DNA. Chromosomal DNA that had been sheared into short fragments (50-800bp) produced sharp peaks of uptake centered at USS, separated by valleys with 1000-fold lower uptake. Peaks heights were proportional to the USS scores predicted by the previously measured contribution to uptake of individual bases in each USS, as well as by predicted differences in DNA shape. Uptake of a long-fragment DNA preparation (1.5-17kb) had much less variation, with 90% of positions having uptake within 2-fold of the mean. Although the presence of a second USS within 100bp had no detectable effect on uptake of short fragments, uptake of long fragments increased with the local density of USS. Simulation of the uptake competition between H. influenzae DNA and the abundant human DNA in the respiratory tract DNA showed that the USS-based system allows H. influenzae DNA to prevail even when human DNA is present in 100-fold excess. CONCLUSION All detectable DNA uptake biases arose from sequences that fit the USS uptake motif, and presence of such sequences increased uptake of short DNA fragments by about 1000-fold. Preferred sequences also had rigidly bent AT-tracts and outer cores. Uptake of longer DNA fragments was much less variable, although detection of uptake biases was limited by strong biases intrinsic to the DNA sequencing process. |