Involvement of Escherichia coli FIS protein in maintenance of bacteriophage mu lysogeny by the repressor: control of early transcription and inhibition of transposition

Transposable prophage Mu is organized as a stable chromosomal domain of E. coli

The E. coli chromosome is compacted by segregation into 400–500 supercoiled domains by both active and passive mechanisms, for example, transcription and DNA-protein association. We find that prophage Mu is organized as a stable domain bounded by the proximal location of Mu termini L and R, which are 37 kbp apart on the Mu genome. Formation/maintenance of the Mu ‘domain’ configuration, reported by Cre-loxP recombination and 3C (chromosome conformation capture), is dependent on a strong gyrase site (SGS) at the center of Mu, the Mu L end and MuB protein, and the E. coli nucleoid proteins IHF, Fis and HU. The Mu domain was observed at two different chromosomal locations tested. By contrast, prophage λ does not form an independent domain. The establishment/maintenance of the Mu domain was promoted by low-level transcription from two phage promoters, one of which was domain dependent. We propose that the domain confers transposition readiness to Mu by fostering topological requirements of the reaction and the proximity of Mu ends. The potential benefits to the host cell from a subset of proteins expressed by the prophage may in turn help its long-term stability.

A majority of sequenced bacterial genomes harbor prophage sequences. Some prophages are viable, while others have decayed from accumulating mutations and genome rearrangements. Prophages, including defective ones, can contribute important biological properties such as antibiotic resistance, toxins, and serum resistance that increase the survival and ecological range of their hosts. We show in this study that the 37 kbp transposable prophage Mu exists in a unique configuration we call the ‘Mu domain’, where its two ends are paired, segregating the Mu sequences from those of the host chromosome. This is the largest stable chromosomal domain in E. coli mapped to date. The Mu domain configuration promotes low-level transcription from an early prophage promoter, which controls the expression of several genes, not all essential for phage growth. Some non-essential genes include DNA repair functions. We suggest that the Mu domain provides long-term survival benefits to both the prophage and the host: to the prophage in bestowing transposition-ready topological properties unique to the Mu reaction, and to the host in contributing extraneous DNA housekeeping functions.

Silencing of toxic gene expression by Fis

Bacteria and bacteriophages have evolved DNA modification as a strategy to protect their genomes. Mom protein of bacteriophage Mu modifies the phage DNA, rendering it refractile to numerous restriction enzymes and in turn enabling the phage to successfully invade a variety of hosts. A strong fortification, a combined activity of the phage and host factors, prevents untimely expression of mom and associated toxic effects. Here, we identify the bacterial chromatin architectural protein Fis as an additional player in this crowded regulatory cascade. Both in vivo and in vitro studies described here indicate that Fis acts as a transcriptional repressor of mom promoter. Further, our data shows that Fis mediates its repressive effect by denying access to RNA polymerase at mom promoter. We propose that a combined repressive effect of Fis and previously characterized negative regulatory factors could be responsible to keep the gene silenced most of the time. We thus present a new facet of Fis function in Mu biology. In addition to bringing about overall downregulation of Mu genome, it also ensures silencing of the advantageous but potentially lethal mom gene.

Immunity of replicating Mu to self-integration: a novel mechanism employing MuB protein

We describe a new immunity mechanism that protects actively replicating/transposing Mu from self-integration. We show that this mechanism is distinct from the established cis-immunity mechanism, which operates by removal of MuB protein from DNA adjacent to Mu ends. MuB normally promotes integration into DNA to which it is bound, hence its removal prevents use of this DNA as target. Contrary to what might be expected from a cis-immunity mechanism, strong binding of MuB was observed throughout the Mu genome. We also show that the cis-immunity mechanism is apparently functional outside Mu ends, but that the level of protection offered by this mechanism is insufficient to explain the protection seen inside Mu. Thus, both strong binding of MuB inside and poor immunity outside Mu testify to a mechanism of immunity distinct from cis-immunity, which we call 'Mu genome immunity'. MuB has the potential to coat the Mu genome and prevent auto-integration as previously observed in vitro on synthetic A/T-only DNA, where strong MuB binding occluded the entire bound region from Mu insertions. The existence of two rival immunity mechanisms within and outside the Mu genome, both employing MuB, suggests that the replicating Mu genome must be segregated into an independent chromosomal domain. We propose a model for how formation of a 'Mu domain' may be aided by specific Mu sequences and nucleoid-associated proteins, promoting polymerization of MuB on the genome to form a barrier against self-integration.

Fis negatively affects binding of Tn4652 transposase by out-competing IHF from the left end of Tn4652

Transposition activity in bacteria is generally maintained at a low level. The activity of mobile DNA elements can be controlled by bacterially encoded global regulators. Regulation of transposition of Tn4652 in Pseudomonas putida is one such example. Activation of transposition of Tn4652 in starving bacteria requires the stationary-phase sigma factor RpoS and integration host factor (IHF). IHF plays a dual role in Tn4652 translocation by activating transcription of the transposase gene tnpA of the transposon and facilitating TnpA binding to the inverted repeats of the transposon. Our previous results have indicated that besides IHF some other P. putida-encoded global regulator(s) might bind to the ends of Tn4652 and regulate transposition activity. In this study, employing a DNase I footprint assay we have identified a binding site of P. putida Fis (factor for inversion stimulation) centred 135 bp inside the left end of Tn4652. Our results of gel mobility shift and DNase I footprint studies revealed that Fis out-competes IHF from the left end of Tn4652, thereby abolishing the binding of TnpA. Thus, the results obtained in this study indicate that the transposition of Tn4652 is regulated by the cellular amount of P. putida global regulators Fis and IHF.

Nucleoid-associated proteins and bacterial physiology

Bacterial physiology is enjoying a renaissance in the postgenomic era as investigators struggle to interpret the wealth of new data that has emerged and continues to emerge from genome sequencing projects and from analyses of bacterial gene regulation patterns using whole-genome methods at the transcriptional and posttranscriptional levels. Information from model organisms such as the Gram-negative bacterium Escherichia coli is proving to be invaluable in providing points of reference for such studies. An important feature of this work concerns the nature of global mechanisms of gene regulation where a relatively small number of regulatory proteins affect the expression of scores of genes simultaneously. The nucleoid-associated proteins, especially Factor for Inversion Stimulation (Fis), IHF, H-NS, HU, and Lrp, represent a prominent group of global regulators and studies of these proteins and their roles in bacterial physiology are providing new insights into how the bacterium governs gene expression in ways that maximize its competitive advantage.

Inhibition of IS2transposition by factor for inversion stimulation

The effect of factor for inversion stimulation (Fis) protein on IS2 transposition was investigated. A full-length IS2 was found to transpose at a frequency 64 times lower in a normal Escherichia coli than in a fis- mutant. To investigate whether Fis affects IS2 transposition by DNA binding, gel retardation and DNase I footprinting experiments were performed. Analysis of Fis binding to the left terminus of IS2 revealed that Fis binds to nucleotide number 44-60 located between the -35 and -10 regions of the major IS2 promoter. To further determine whether Fis binding affects IS2 transcription, the major IS2 promoter was fused to a luciferase gene and assayed for its transcription efficiency in the presence or absence of Fis. The results showed that Fis reduced transcription from the major IS2 promoter by approximately sixfold. Analysis of Fis binding to the right terminal repeat of IS2 revealed that Fis binds to the inner end of the repeat, which is the same region as the place where the IS2 transposase binds. These results suggest that Fis inhibits IS2 transposition by blocking the binding sites of IS2 transposase and by repressing the transcription of IS2 genes.

The ColR-ColS two-component signal transduction system is involved in regulation of Tn4652 transposition in Pseudomonas putida under starvation conditions

Bacteria use two-component signal transduction pathways to sense both extracellular and intracellular environment and to coordinate cellular events according to changing conditions. Adaptation can be either physiological or genetical. Here, we present evidence that a genome reorganization process such as transposition can be controlled by certain environmental cues sensed by a two-component signal transduction system. We demonstrate that transposition-dependent accumulation of phenol-utilizing mutants is severely decreased in Pseudomonas putida defective in a two-component system colRS. Translocation of Tn4652 is decreased both in colR- and colS-defective strains, indicating that signal transduction from a histidine kinase ColS to a response regulator ColR is necessary for the activation of Tn4652 in bacteria starving on phenol. However, overexpression of ColR in a colS-defective strain restores Tn4652 transposition, suggesting that absence of the signal from ColS can be compensated by an elevated amount of ColR. In vitro analysis of purified ColR and ColS proteins evidenced that they constitute a functional phosphorelay. Site-directed mutagenesis revealed that a conserved H221 can be the phosphoryl-accepting residue in ColS and that aspartate residues D8 and D51 of ColR are necessary for the phosphotransfer from ColS to ColR. To our knowledge, Tn4652 is the first bacterial transposon regulated by a two-component system. This finding indicates that transpositional activity can respond to signals sensed and processed by the host.

IHF is the limiting host factor in transposition of Pseudomonas putida transposon Tn4652 in stationary phase

Transpositional activity of mobile elements is not constant. Conditional regulation of host factors involved in transposition may severely change the activity of mobile elements. We have demonstrated previously that transposition of Tn4652 in Pseudomonas putida is a stationary phase-specific event, which requires functional sigma S (Ilves et al., 2001, J Bacteriol 183: 5445-5448). We hypothesized that integration host factor (IHF), the concentration of which is increased in starving P. putida, might contribute to the transposition of Tn4652 as well. Here, we demonstrate that transposition of Tn4652 in stationary phase P. putida is essentially limited by the amount of IHF. No transposition of Tn4652 occurs in a P. putida ihfA-defective strain. Moreover, overexpression of IHF results in significant enhancement of transposition compared with the wild-type strain. This indicates that the amount of IHF is a bottleneck in Tn4652 transposition. Gel mobility shift and DNase I footprinting studies revealed that IHF is necessary for the binding of transposase to both transposon ends. In vitro, transposase can bind to inverted repeats of transposon only after the binding of IHF. The results obtained in this study indicate that, besides sigma S, IHF is another host factor that is implicated in the elevation of transposition in stationary phase.

Bio-array images processing and genetic networks modelling

The new tools available for gene expression studies are essentially the bio-array methods using a large variety of physical detectors (isotopes, fluorescent markers, ultrasounds...). Here we present first rapidly an image-processing method independent of the detector type, dealing with the noise and with the peaks overlapping, the peaks revealing the detector activity (isotopic in the presented example), correlated with the gene expression. After this primary step of bio-array image processing, we can extract information about causal influence (activation or inhibition) a gene can exert on other genes, leading to clusters of genes co-expression in which we extract an interaction matrix M and an associated interaction graph G explaining the genetic regulatory dynamics correlated to the studied tissue function. We give two examples of such interaction matrices and graphs (the flowering genetic regulatory network of Arabidopsis thaliana and the lytic/lysogenic operon of the phage Mu) and after some theoretical rigorous results recently obtained concerning the asymptotic states generated by the genetic networks having a given interaction matrix and reciprocally concerning the minimal (in the sense of having a minimal number of non-zero coefficients) matrices having given stationary stable states.

Characterization of the cts4 repressor mutation in transposable bacteriophage Mu

Mucts4 was isolated more than 30 years ago and was the first available thermoinducible derivative of transposable phage Mu. We have characterized the cts4 mutation and the corresponding mutant protein. Contrary to previously characterized thermoinducible Mu prophages (e.g., Mucts62), Mucts4 lysogenizes at reduced frequency even at 30 degrees C. The cts4 mutation (Leu129Val) was located in this central repressor region. The cts4 protein was thermosensitive for operator DNA binding in vitro. Temperature-dependent changes in protein-protein cross-linking patterns in the absence of DNA were detected for purified wild type, cts62 and cts4 repressor proteins. The cts4 protein exhibited a subtly different electrophoretic profile, which became more marked at higher temperatures, from both the wild type and cts62. In addition the cts4 repressor generated a significantly different pattern of binding to DNA fragments carrying the early operator region. Consistent with the predicted involvement of the central leucine-rich region of the Mu repressor in the formation of multimeric forms, the cts4 mutation thus appeared to affect protein-protein interactions.

DNA supercoiling and transcription in Escherichia coli: The FIS connection

The nucleoid-associated protein FIS modulates the topology of DNA in a growth-phase dependent manner functioning homeostatically to counteract excessive levels of negative superhelicity. We propose that this is achieved by at least two mechanisms: the physical constraint of low levels of negative superhelicity by FIS binding to DNA and by a reduction in the expression and effectiveness of DNA gyrase. In addition, high levels of expression of the fis gene do themselves require a high negative superhelical density. On DNA substrates containing phased high affinity binding sites, as exemplified by the upstream activating sequence of the tyrT promoter, FIS forms tightly bent DNA structures, or microloops, that are necessary for the optimal expression of the promoter. We suggest that these microloops compensate in part for the FIS-induced lowering of the superhelical density.

A DNA architectural protein couples cellular physiology and DNA topology in Escherichia coli

In Escherichia coli, the transcriptional activity of many promoters is strongly dependent on the negative superhelical density of chromosomal DNA. This, in turn, varies with the growth phase, and is correlated with the overall activity of DNA gyrase, the major topoisomerase involved in the elevation of negative superhelicity. The DNA architectural protein FIS is a regulator of the metabolic reorganization of the cell during early exponential growth phase. We have previously shown that FIS modulates the superhelical density of plasmid DNA in vivo, and on binding reshapes the supercoiled DNA in vitro. Here, we show that, in addition, FIS represses the gyrA and gyrB promoters and reduces DNA gyrase activity. Our results indicate that FIS determines DNA topology both by regulation of topoisomerase activity and, as previously inferred, by directly reshaping DNA. We propose that FIS is involved in coupling cellular physiology to the topology of the bacterial chromosome.

Starvation-induced Mucts62-mediated coding sequence fusion: a role for ClpXP, Lon, RpoS and Crp

The formation of araB-lacZ coding sequence fusions in Escherichia coli is a particular type of chromosomal rearrangement induced by Mucts62, a thermoinducible mutant of mutator phage Mu. Fusion formation is controlled by the host physiology. It only occurs after aerobic carbon starvation and requires the phage-encoded transposase pA, suggesting that these growth conditions trigger induction of the Mucts62 prophage. Here, we show that thermal induction of the prophage accelerated araB-lacZ fusion formation, confirming that derepression is a rate-limiting step in the fusion process. Nonetheless, starvation conditions remained essential to complete fusions, suggesting additional levels of physiological regulation. Using a transcriptional fusion indicator system in which the Mu early lytic promoter is fused to the reporter E. coli lacZ gene, we confirmed that the Mucts62 prophage was derepressed in stationary phase (S derepression) at low temperature. S derepression did not apply to prophages that expressed the Mu wild-type repressor. It depended upon the host ClpXP and Lon ATP-dependent proteases and the RpoS stationary phase-specific sigma factor, but not upon Crp. None of these four functions was required for thermal induction. Crp was required for fusion formation, but only when the Mucts62 prophage encoded the transposition/replication activating protein pB. Finally, we found that thermally induced cultures did not return to the repressed state when shifted back to low temperature and, hence, remained activated for accelerated fusion formation upon starvation. The maintenance of the derepressed state required the ClpXP and Lon host proteases and the prophage Ner-regulatory protein. These observations illustrate how the cts62 mutation in Mu repressor provides the prophage with a new way to respond to growth phase-specific regulatory signals and endows the host cell with a new potential for adaptation through the controlled use of the phage transposition machinery.

Interplay between global regulators of Escherichia coli: effect of RpoS, Lrp and H-NS on transcription of the gene osmC

The transcription of the osmC gene of Escherichia coli is regulated as a function of the phase of growth. It is induced during the decelerating phase, before entry into stationary phase. osmC expression is directed by two overlapping promoters, osmCp1 and osmCp2. osmCp2 is mainly transcribed by E-sigma(s), the RNA polymerase using the sigma(s) (RpoS) sigma factor, and is responsible for the growth phase regulation. Transcription from osmCp1 is independent of sigma(s). The leucine-responsive protein (Lrp) has been shown to bind the osmC promoter region in band shift experiments. In vivo analysis using osmC-lacZ transcriptional fusions demonstrated that Lrp affects the expression of both promoters. It represses the transcription of osmCp1 and activates the transcription of osmCp2 by E-sigma(s). An absence of Lrp results in an increase in the amount of RpoS during exponential growth in minimal medium. The nucleoid-associated protein H-NS also represses osmC transcription from both promoters. However, this happens through different mechanisms. The effect on osmCp2 is probably mediated by the increase in sigma(s) concentration in the cytoplasm of hns- mutants, while the effect on osmCp1 is independent of sigma(s). No binding of H-NS to the promoter region DNA could be detected, indicating that the effect on osmCp1 could also be indirect.

Stimulation of DNA inversion by FIS: evidence for enhancer-independent contacts with the Gin-gix complex

Efficient DNA inversion catalysed by the invertase Gin requires the cis-acting recombinational enhancer and the Escherichia coliFIS protein. Binding of FIS bends the enhancer DNA and, on a negatively supercoiled DNA inversion substrate, facilitates the formation of a synaptic complex with specific topology. Previous studies have indicated that FIS-independent Gin mutants can be isolated which have lost the topological constraints imposed on the inversion reaction yet remain sensitive to the stimulatory effect of FIS. Whether the effect of FIS is purely architectural, or whether in addition direct protein contacts between Gin and FIS are required for efficient catalysis has remained an unresolved question. Here we show that FIS mutants impaired in DNA binding are capable of either positively or negatively affecting the inversion reaction both in vivo and in vitro. We further demonstrate that the mutant protein FIS K25E/V66A/M67T dramatically enhances the cleavage of recombination sites by FIS-independent Gin in an enhancer-independent manner. Our observations suggest that FIS plays a dual role in the inversion reaction and stimulates both the assembly of the synaptic complex as well as DNA strand cleavage.

Mutual stabilisation of bacteriophage Mu repressor and histone-like proteins in a nucleoprotein structure

Integration host factor (IHF) binds in a sequence-specific manner to the bacteriophage Mu early operator. It participates with bound Mu repressor, c, in building stable, large molecular mass nucleoprotein complexes in vitro and enhances repression of early transcription in vivo. We demonstrate that, when the specific IHF binding site with the operator is mutated, the appearance of large molecular mass complexes still depends on IHF and c, but the efficiency of their formation is reduced. Moreover, the IHF-like HU protein, which binds DNA in a non-sequence-specific way, can substitute for IHF and participate in complex formation. Since the complexes require both c and a host factor (IHF or HU), the results imply that these proteins stabilise each other within the nucleoprotein structures. These results suggest that IHF and HU are directed to the repressor-operator complexes, even in the absence of detectable sequence-specific binding. This could be a consequence of their preferential recognition of DNA containing a distortion such as that introduced by repressor binding to the operator. The histone-like proteins could then stabilise the nucleoprotein complexes simply by their capacity to maintain a bend in DNA rather than by specific protein-protein interactions with c. This model is supported by the observation that the unrelated eukaryotic HMG-1 protein, which exhibits a similar marked preference for structurally deformed DNA, is also able to participate in the formation of higher-order complexes with c and the operator DNA.

Sequence, regulation, and functions of fis in Salmonella typhimurium

The fis operon from Salmonella typhimurium has been cloned and sequenced, and the properties of Fis-deficient and Fis-constitutive strains were examined. The overall fis operon organization in S. typhimurium is the same as that in Escherichia coli, with the deduced Fis amino acid sequences being identical between both species. While the open reading frames upstream of fis have diverged slightly, the promoter regions between the two species are also identical between -49 and +94. Fis protein and mRNA levels fluctuated dramatically during the course of growth in batch cultures, peaking at approximately 40,000 dimers per cell in early exponential phase, and were undetectable after growth in stationary phase. fis autoregulation was less effective in S. typhimurium than that in E. coli, which can be correlated with the absence or reduced affinity of several Fis-binding sites in the S. typhimurium fis promoter region. Phenotypes of fis mutants include loss of Hin-mediated DNA inversion, cell filamentation, reduced growth rates in rich medium, and increased lag times when the mutants are subcultured after prolonged growth in stationary phase. On the other hand, cells constitutively expressing Fis exhibited normal logarithmic growth but showed a sharp reduction in survival during stationary phase. During the course of these studies, the sigma 28-dependent promoter within the hin-invertible segment that is responsible for fljB (H2) flagellin synthesis was precisely located.

Identification of genes negatively regulated by Fis: Fis and RpoS comodulate growth-phase-dependent gene expression in Escherichia coli

Fis is a nucleoid-associated protein in Escherichia coli that has been shown to regulate recombination, replication, and transcription reactions. It is expressed in a transient manner under batch culturing conditions such that high levels are present during early exponential phase and low levels are present during late exponential phase and stationary phase. We have screened a random collection of transposon-induced lac fusions for those which give decreased expression in the presence of Fis. Thirteen different Fis-repressed genes were identified, including glnQ (glutamine high-affinity transport), mglA (methyl-galactoside transport), xylF (D-xylose-binding protein), sdhA (succinate dehydrogenase flavoprotein subunit), and a newly identified aldehyde dehydrogenase, aldB. The LacZ expression patterns revealed that many of the fusions were maximally expressed at different stages of growth, including early log phase, mid- to late log phase, and stationary phase. The expression of some of the late-exponential- and stationary-phase genes was dependent on the RpoS sigma factor, whereas that of others was affected negatively by RpoS. We conclude that Fis negatively regulates a diverse set of genes and that RpoS can function to both activate and inhibit the expression of specific genes.

Effects of N-terminal deletions of the Escherichia coli protein Fis on growth rate, tRNA(2Ser) expression and cell morphology

The Escherichia coli Fis protein is known to be involved in a variety of processes, including the activation of stable RNA operons. In this paper we study the ability of a set of N-terminal Fis deletion mutants to stimulate transcription of the tRNA(2Ser) gene. The results indicate that the domain of the Fis protein containing residues 1-26 is not required for transcription activation. The Fis mutants that are still active in transcription stimulation can also complement the reduced growth rates of Fis- cells, suggesting that the same activating domain is involved in this phenomenon. In addition, we show that in fast growing cultures in the absence of an active Fis protein, minicells are formed. These minicells seem to arise from septum formation near the cell poles. Suppression of minicell formation by Fis also does not require the presence of the N-terminal domain of the protein.

Nucleoid proteins

This article examines the published evidence in support of the classification of organisms into three groups (Bacteria, Archae, and Eukarya) instead of two groups (prokaryotes and eukaryotes) and summarizes the comparative biochemistry of each of the known histone-like, nucleoid DNA-binding proteins. The molecular structures and amino acid sequences of Archae are more similar to those of Eukarya than of Bacteria, with a few exceptions. Cytochemical methodology employed for localizing these proteins in archaeal and bacterial cells has also been reviewed. It is becoming increasingly apparent that these proteins participate both in the organization of DNA and in the control of gene expression. Evidence obtained from biochemical properties, structural and functional differences, and the ultrastructural location of these proteins, as well as from gene mutations clearly justifies the division of prokaryotes into bacterial and archaeal groups. Indeed, chromosomes, whether they be nuclear, prokaryotic, or organellar, are invariably complexed with abundant, small, basic proteins that bind to DNA with low sequence specificity. These proteins include the histones, histone-like proteins, and nonhistone high mobility group (HMG) proteins.

'Muprints' of the lac operon demonstrate physiological control over the randomness of in vivo transposition

A method called Muprinting has been developed that uses PCR to generate a detailed picture of the bacteriophage Mu transposition sites in chosen domains of the bacterial chromosome. Muprinting experiments in Escherichia coli show that the frequency of phage integration changes dramatically near two repressor binding sites in the lac operon. When the lac operon was repressed, hotspots for Mu transposition were found near the O1 and O2 operators that are proposed to make a repression loop. When cells were grown in lactose, Mu transposition near these operators was greatly diminished. Striking changes in transposition frequencies were limited to the control region and were not found in a region of the lacZ gene lying beyond the O2 operator. Muprints of the bgl operon showed a different pattern; hotspots for Mu transposition detected in sequences upstream of the bglC promoter when the operon was silenced changed when the operon became activated by mutation. By targeting transposition to the regulatory regions around non-expressed genes, Mu may demonstrate a self-restraint mechanism that allows the virus to move through its host genome without disrupting the functions that contribute to a healthy cell physiology.

Regulation of bacteriophage Mu transposition

Bacteriophage Mu is a transposon and a temperate phage which has become a paradigm for the study of the molecular mechanism of transposition. As a prophage, Mu has also been used to study some aspects of the influence of the host cell growth phase on the regulation of transposition. Through the years several host proteins have been identified which play a key role in the replication of the Mu genome by successive rounds of replicative transposition as well as in the maintenance of the repressed prophage state. In this review we have attempted to summarize all these findings with the purpose of emphasizing the benefit the virus and the host cell can gain from those phage-host interactions.

Interaction of Fis protein with DNA: bending and specificity of binding

The Escherichia coli Fis protein is a dimeric DNA-binding protein whose specific binding sites share a weak consensus sequence. Use of the gel retardation technique indicates that binding of Fis on a linear DNA fragment leads to the formation of a ladder of defined retarded complexes, independently of the presence of a specific site. This non-specific binding of Fis is consistent with a model where equivalent low-affinity sites on a given fragment would be bound randomly and independently of each other by consecutive Fis dimers. Evidence is presented that non-specific binding of Fis can, however, induce an apparent site-specific conformational change in the DNA. This observation is discussed in terms of a model in which each Fis:DNA complex detected in gel retardation experiments actually represents a dynamic equilibrium of a fixed number of Fis dimers distributed on the fragment.