Functional studies of potential intrastrand triplex elements in the Escherichia coli genome

We previously used a pattern recognition program for nucleic acids to detect sequences with the potential to form intrastrand triplexes. Potential intrastrand triplex (PIT) element families were found in Escherichia coli, Synechocystis sp. and Haemophilus influenza. We were particularly intrigued with the family found in E. coli, which contained 25 dispersed copies of a particular PIT sequence corresponding to the purine triplex motif. E. coli PIT elements appear to occur exclusively in non-coding regions. We now report biochemical experiments testing the interaction of E. coli PIT elements with polymerases and single-stranded DNA-binding protein (SSB). The elements were also tested in genetic experiments as promoters, transcription terminators, or replication pause sites in E. coli. We show that PIT elements display provocative characteristics in certain biochemical assays. When appropriately oriented, the elements block elongation by Taq DNA polymerase at 72 degrees C, but not elongation by T7 DNA polymerase at 37 degrees C. The G-rich strand of the E. coli PIT sequence folds into a form with reduced affinity for SSB. On the other hand, in vivo studies did not detect replication delays for conjugal transfer of episomes containing PIT elements. These sequences were shown not to act as promoters, but the presence of PIT elements in RNA leaders upstream of a coding region could strongly influence expression of the downstream gene. These effects were shown to be post-transcriptional and were solely dependent on the Watson-Crick stem-loop structure within the PIT element. Thus, although PIT element DNA displays unusual biochemical properties, it remains unknown how these elements arose, and why they persist in the E. coli genome.