Genes on a Wire: The Nucleoid-Associated Protein HU Insulates Transcription Units in Escherichia coli

Abstract

The extent to which chromosomal gene position in prokaryotes affects local gene expression remains an open question. Several studies have shown that chromosomal re-positioning of bacterial transcription units does not alter their expression pattern, except for a general decrease in gene expression levels from chromosomal origin to terminus proximal positions, which is believed to result from gene dosage effects. Surprisingly, the question as to whether this chromosomal context independence is a cis encoded property of a bacterial transcription unit, or if position independence is a property conferred by factors acting in trans, has not been addressed so far. For this purpose, we established a genetic test system assessing the chromosomal positioning effects by means of identical promoter-fluorescent reporter gene fusions inserted equidistantly from OriC into both chromosomal replichores of Escherichia coli K-12. Our investigations of the reporter activities in mutant cells lacking the conserved nucleoid associated protein HU uncovered various drastic chromosomal positional effects on gene transcription. In addition we present evidence that these positional effects are caused by transcriptional activity nearby the insertion site of our reporter modules. We therefore suggest that the nucleoid-associated protein HU is functionally insulating transcription units, most likely by constraining transcription induced DNA supercoiling.

Figure 1. Growth phase dependent expression of different chromosomal NAP promoter-yfp fusion constructs and schematic overview for the construction of reporter strains used in this study.

(A) fis promoter dependent YFP expression; (B) hns promoter dependent YFP expression; (C) dps promoter dependent YFP expression. The YFP expression patterns are clearly distinct and are reflecting the known growth phase dependent expression patterns of Fis, H-NS and Dps. (D) Flowchart showing the construction procedure of the reporter strains used in this study. O: origin of replication; OL: origin-proximal position, left replichore; OR: origin-proximal position, right replichore; ML: middle position, left replichore; MR: middle position, right replichore; T: terminus of replication; TL: terminus-proximal position, left replichore; TR, terminus-proximal position, right replichore.

Figure 2. Pdps module dependent YFP expression in wild type and hupA/B mutant.

Shown are the growth curves (lines) and corresponding YFP expression patterns (diamonds) of the strains containing the Pdps modules in six different chromosomal positions (a graphical representation of the positions and color code for the modules is shown above the graphs; the experimental setup is shown as flow chart on the left). (A,E) Pdps module dependent YFP expression in wild type cells using dYT starter culture (Day 1). The medium switch causes an intermediate lag phase in M9 that is accompanied by an activation of the Pdps modules (A; black arrow). (C,G) Pdps module dependent YFP expression in wild type cells using M9 as starter culture (Day 2). The intermediate lag phase in M9 is absent and the growth of the strains becomes more synchronized (compare lines in A,E with C,G). The overall pattern of Pdps module dependent YFP expression is very similar and stable in wild type cells. (B,F) Pdps module dependent YFP expression patterns in hupA/B mutant cells using dYT starter culture (Day 1) and M9 as starter culture (Day 2; D,H), showing no detectable copy number effects and various positional effects.

Figure 3. Pdps module activity after hydrogen peroxide challenge.

Shown are the promoter activities of the Pdps modules as increase of normalized fluorescence over time (diamonds and lines) and corresponding growth curves (dashed lines). Red arrows indicate time point of H₂O₂ addition. (A) In wild type cells, all Pdps modules are producing YFP for about 40 min after H₂O₂ addition. Afterwards the Pdps is switched off until it is reactivated by growth phase control. As expected, the copy number effects are more pronounced in this experiment (compare module position and YFP production). (B) In hupA/B mutants, only a subset of the Pdps modules produced detectable fluorescence in this experiment, no copy number effects can be observed. (C) Pdps module response as fraction of total signal in 5 independent experiments. The expected value is approximately 17% (100/6), if all positions contribute equally to the total response (black arrow). In contrast to wild type, the response of the modules is strongly fluctuating in the hupA/B mutant with no indication of copy number effects. (D) Schematic representation of Pdps module positions in the E. coli chromosome.

Figure 4. Phns module dependent YFP expression in wild type and hupA/B mutant.

Shown are the growth curves (lines) and corresponding YFP expression patterns (diamonds) of the strains containing the Phns modules in six different chromosomal positions (a graphical representation of the positions and color code for the modules is shown above the graphs; the experimental setup is shown as flow chart on the left). (A,E) Phns module dependent YFP expression in wild type cells using dYT starter culture (Day 1). (C,G) Phns module dependent YFP expression in wild type cells using M9 as starter culture (Day 2) Except for TL, the overall pattern of Phns module dependent YFP expression is very similar and less dependent on the growth medium than the Pdps modules in wild type cells. (B,F) Phns module dependent YFP expression patterns in hupA/B mutant cells using dYT starter culture (Day 1) and M9 as starter culture (Day 2; D,H), showing no detectable copy number effects and various positional effects.

Figure 5

Pfis module dependent YFP expression in wild type (A) and hupA/B mutant (B). Shown are the growth curves (lines) and corresponding YFP expression patterns (diamonds) of the strains containing the Pfis modules in MLup (red) and ML (orange). The Pfis modules ML and MLup produce identical amounts of YFP in wild type cells (compare red and orange diamonds in A). In the hupA/B mutant the Pfis module MLup produces 2–3 fold more YFP than Pfis module ML (compare red and orange diamonds in B).

Figure 6. Upstream gene expression is repressing the Pdps module ara in the hupA/B mutant.

Shown is the YFP production of the Pdps modules OR and ara in M9 supplemented with glucose and M9 supplemented with arabinose. Only the Pdps module ara is repressed in M9 supplemented with arabinose (compare red with black diamonds).