Hopping into a hot seat: Role of DNA structural features on IS5-mediated gene activation and inactivation under stress

Abstract

Insertion sequence elements (IS elements) are proposed to play major roles in shaping the genetic and phenotypic landscapes of prokaryotic cells. Recent evidence has raised the possibility that environmental stress conditions increase IS hopping into new sites, and often such hopping has the phenotypic effect of relieving the stress. Although stress-induced targeted mutations have been reported for a number of E. coli genes, the glpFK (glycerol utilization) and the cryptic bglGFB (β-glucoside utilization) systems are among the best characterized where the effects of IS insertion-mediated gene activation are well-characterized at the molecular level. In the glpFK system, starvation of cells incapable of utilizing glycerol leads to an IS5 insertion event that activates the glpFK operon, and enables glycerol utilization. In the case of the cryptic bglGFB operon, insertion of IS5 (and other IS elements) into a specific region in the bglG upstream sequence has the effect of activating the operon in both growing cells, and in starving cells. However, a major unanswered question in the glpFK system, the bgl system, as well as other examples, has been why the insertion events are promoted at specific locations, and how the specific stress condition (glycerol starvation for example) can be mechanistically linked to enhanced insertion at a specific locus. In this paper, we show that a specific DNA structural feature (superhelical stress-induced duplex destabilization, SIDD) is associated with "stress-induced" IS5 insertion in the glpFK, bglGFB, flhDC, fucAO and nfsB systems. We propose a speculative mechanistic model that links specific environmental conditions to the unmasking of an insertional hotspot in the glpFK system. We demonstrate that experimentally altering the predicted stability of a SIDD element in the nfsB gene significantly impacts IS5 insertion at its hotspot.

Fig 1

Panel A. The zapB-glpF intergenic region showing the position of IS5 insertion and its orientation. The transcriptional start site is indicated as +1. The SIDD region (DNA sequences at which G(x) values fall below 6 kcal/mol) is shown in red. The four operator sequences labeled O₁ through O₄ are underlined, and the two Crp-cAMP binding sites that overlap O₂ and O₃ are indicated by dashed lines above the sequence. The CTAA sequence at which IS5 insertion occurs is underlined (> indicates IS5 orientation). Panel B. Plot of G(x) values (see Materials and Methods and Table 2) for a 4.8 kb DNA sequence from E. coli K12 MG1655. The entire plot is shown at the top, and for clarity, a segment of the plot (bp 1000 to 2500) is expanded at the bottom. The locations of the IS5 insertion (red arrow), the start sites for transcription (green arrow) and translation (black arrow) are shown on the energy plot.

Fig 2

Panel A. Role of transcription factors Crp-cAMP and GlpR on glpFK expression, and on IS5 insertion. Crp, when complexed with cyclic AMP (cAMP), activates transcription (→) but inhibits IS5 insertion (--|). GlpR in the free form inhibits both transcription, and IS5 insertion(--|), but when present in sufficient quantities, glycerol 3-phosphate (G3P) binds to GlpR, inducing a conformational change that causes the protein to dissociate from the DNA, relieving the inhibition of transcription, and allowing IS5 insertion [7, 39]. Panel B. Models for unmasking of the IS5 insertion hotspot with and without regional DNA amplification. When Δcrp or ΔcyaA cells are subjected to prolonged incubation on minimal agar with glycerol as the sole carbon source, some glycerol diffuses into cells, and is converted to G3P at low efficiency (see text). Accumulation of G3P over time leads to dissociation of GlpR from the glpFK promoter region, unmasking the IS5 insertion hotspot. In an alternative scenario, IS5 insertion occurs in a subset of cells in which the glpFK region is amplified, leading to accelerated release of GlpR due to multicopy titration of GlpR. Note that the gene encoding GlpR is located 12 minutes away, and is unlikely to be co-amplified with glpFK in this scenario. Furthermore, to the extent that G3P formation depends on leaky expression of glpK, amplification will also lead to a more rapid accumulation of G3P. Finally, multiple copies of the IS5 target site resulting from glpFK amplification increase the likelihood of capturing an insertion event in these cells.

Fig 3

Panel A. The DNA sequence of the phoU-bglG intergenic region showing IS5 insertion sites, and the DNA sequence at which G(x) values fall below 6 (the SIDD zone). The positions of the insertion sites for the IS5 element are shown in lower case letters (ttaa, ctaa and ctgg), and orientations are shown by > or <. Note that IS5-S6 and IS5-517 are outside of the indicated SIDD zone (see text). The transcriptional start site is indicated as +1. Panel B. Plot of G(x) values (see Materials and Methods and Table 2) for a 4.6 kb DNA sequence from E. coli K12 MG1655. For clarity, only a segment of the plot (bp 1000 to 3000) is shown. The location of two IS5 insertion sites representing two of the most frequent events are indicated by red arrows, and the much less-frequent outlier insertion sites (IS5-S6 and IS5-517) are shown by blue arrows (see text). The translational start site for bglG is indicated by a black arrow.

Fig 4

Panel A. The uspC- flhDC intergenic region in E. coli K12 BW25113 showing positions of IS5 insertion sites and the SIDD sequences (red) G(x) values fall below 6. The positions of the insertion sites for IS5 elements are shown (underlined TTAG, CTAG and TTAA). The transcriptional start sites for the two divergently transcribed genes are indicated as +1, and the flhD translational start site ATG is indicated. Panel B. Plot of G(x) values (see Materials and Methods and Table 2) for a 4.3 kb DNA sequence from E. coli K12 BW25113. For clarity, only a segment of the plot (bp 1000 to 3000) is shown. The locations of the IS5 insertion (red arrowheads), the start sites for transcription (green arrowhead) and translation (pink arrowhead) for the flhD gene are shown on the energy plot.

Fig 5

Panel A. DNA sequence encompassing the divergently transcribed fucAO-fucPIK intergenic region showing the single IS5 insertion site and the SIDD zone of the same region (red) where G(x) values fall below 6. The single IS5 insertion site CTAG (double-underline) is shown. The transcriptional start sites are indicated as +1, and the translational start site for fucA is indicted. Panel B. Plot of G(x) values (see Materials and Methods and Table 2) for a 4.6 kb DNA sequence from E. coli K12 MG1655. For clarity, only a segment of the plot (bp 1000 to 3000) is shown. The location of the IS5 insertion (red arrowhead), and the transcriptional (green arrowhead) and translational (black arrowhead) start sites for the fucA gene are shown on the plot.

Fig 6

Panel A. The ybdF-nfsB intergenic region showing the position of the IS5 insertion site, and the SIDD sequences where G(x) values fall below 6 (red). The IS5 insertional hotspot is shown in underlined blue letters within the N-terminal portion of the nfsB gene. The transcriptional start site (+1) and the translational start site for nfsB are indicated. The three transversions (two A-to-C, plus one T-to-G) constituting the Prom-3 mutant, and the three base changes (one C-to-A transversion, and two G-to-A transitions) constituting the ORF-3 mutant are indicated (see text, and Fig 7). Panel B. Plot of G(x) values (see Materials and Methods and Table 2) for a 4.6 kb DNA sequence from E. coli K12 MG1655. For clarity, only a segment of the plot (bp 1800 to 2500) is shown. The location of the IS5 insertion (red arrowhead) and those for the transcriptional (green arrowhead) and translational (black arrowhead) start sites are indicated on the plot.

Fig 7

Panel A. Analysis of the effect of in silico mutations on the SIDD profile of the ybdF-nfsB intergenic region. As shown in Fig 6, the PROM-3 mutant has three A:T to C:G transversions downstream of the transcription start site (+1 in Fig 6), whereas the ORF-3 mutant has one G:C to T:A transversion, and two G:C to A:T transition mutations. SIDD plots for the wild-type (black line), Prom-3 (red line) and ORF-3 (blue line) sequences are shown. Relative to the wild type sequence, the G(x) values are increased in the Prom-3 mutant (red), meaning that DNA duplex is more stable than wild type in this region. The G(x) values are decreased for the ORF-3 mutation (blue), meaning that the duplex is further destabilized relative to the wild type sequence. Note that all shifts are localized to the same SIDD “well” as found in the wild type sequence. Panel B. Effect of Prom-3 or ORF-3 mutations on relative IS5 insertion frequencies (see text and Table 4). Color coding is the same as in Panel A: black shows insertion frequencies for the wild type strain, red for the Prom-3 (re-stabilized duplex) mutant, and blue for the ORF-3 (further destabilized duplex) mutant.