A- and T-tract-mediated intrinsic curvature in native DNA between the binding site of the upstream activator NtrC and the nifLA promoter of Klebsiella pneumoniae facilitates transcription

Discovering DNA shape motifs with multiple DNA shape features: generalization, methods, and validation

DNA motifs are crucial patterns in gene regulation. DNA-binding proteins (DBPs), including transcription factors, can bind to specific DNA motifs to regulate gene expression and other cellular activities. Past studies suggest that DNA shape features could be subtly involved in DNA-DBP interactions. Therefore, the shape motif annotations based on intrinsic DNA topology can deepen the understanding of DNA-DBP binding. Nevertheless, high-throughput tools for DNA shape motif discovery that incorporate multiple features altogether remain insufficient. To address it, we propose a series of methods to discover non-redundant DNA shape motifs with the generalization to multiple motifs in multiple shape features. Specifically, an existing Gibbs sampling method is generalized to multiple DNA motif discovery with multiple shape features. Meanwhile, an expectation-maximization (EM) method and a hybrid method coupling EM with Gibbs sampling are proposed and developed with promising performance, convergence capability, and efficiency. The discovered DNA shape motif instances reveal insights into low-signal ChIP-seq peak summits, complementing the existing sequence motif discovery works. Additionally, our modelling captures the potential interplays across multiple DNA shape features. We provide a valuable platform of tools for DNA shape motif discovery. An R package is built for open accessibility and long-lasting impact: https://zenodo.org/doi/10.5281/zenodo.10558980.

IHF is required for the transcriptional regulation of the Desulfovibrio vulgaris Hildenborough orp operons

Transcriptional activation of σ(54)-dependent promoters is usually tightly regulated in response to environmental cues. The high abundance of potential σ(54)-dependent promoters in the anaerobe bacteria, Desulfovibrio vulgaris Hildenborough, reflects the high versatility of this bacteria suggesting that σ(54) factor is the nexus of a large regulatory network. Understanding the key players of σ(54)-regulation in this organism is therefore essential to gain insights into the adaptation to anaerobiosis. Recently, the D. vulgaris orp genes, specifically found in anaerobe bacteria, have been shown to be transcribed by the RNA polymerase coupled to the σ(54) alternative sigma factor. In this study, using in vitro binding experiments and in vivo reporter fusion assays in the Escherichia coli heterologous host, we showed that the expression of the divergent orp promoters is strongly dependent on the integration host factor IHF. Bioinformatic and mutational analysis coupled to reporter fusion activities and mobility shift assays identified two functional IHF binding site sequences located between the orp1 and orp2 promoters. We further determined that the D. vulgaris DVU0396 (IHFα) and DVU1864 (IHFβ) subunits are required to control the expression of the orp operons suggesting that they form a functionally active IHF heterodimer. Interestingly results obtained from the in vivo inactivation of DVU0396, which is required for orp operons transcription, suggest that several functionally IHF active homodimer or heterodimer are present in D. vulgaris.

DNA sequence-dependent mechanics and protein-assisted bending in repressor-mediated loop formation

As the chief informational molecule of life, DNA is subject to extensive physical manipulations. The energy required to deform double-helical DNA depends on sequence, and this mechanical code of DNA influences gene regulation, such as through nucleosome positioning. Here we examine the sequence-dependent flexibility of DNA in bacterial transcription factor-mediated looping, a context for which the role of sequence remains poorly understood. Using a suite of synthetic constructs repressed by the Lac repressor and two well-known sequences that show large flexibility differences in vitro, we make precise statistical mechanical predictions as to how DNA sequence influences loop formation and test these predictions using in vivo transcription and in vitro single-molecule assays. Surprisingly, sequence-dependent flexibility does not affect in vivo gene regulation. By theoretically and experimentally quantifying the relative contributions of sequence and the DNA-bending protein HU to DNA mechanical properties, we reveal that bending by HU dominates DNA mechanics and masks intrinsic sequence-dependent flexibility. Such a quantitative understanding of how mechanical regulatory information is encoded in the genome will be a key step towards a predictive understanding of gene regulation at single-base pair resolution.

Poly(dA:dT)-rich DNAs are highly flexible in the context of DNA looping

Large-scale DNA deformation is ubiquitous in transcriptional regulation in prokaryotes and eukaryotes alike. Though much is known about how transcription factors and constellations of binding sites dictate where and how gene regulation will occur, less is known about the role played by the intervening DNA. In this work we explore the effect of sequence flexibility on transcription factor-mediated DNA looping, by drawing on sequences identified in nucleosome formation and ligase-mediated cyclization assays as being especially favorable for or resistant to large deformations. We examine a poly(dA:dT)-rich, nucleosome-repelling sequence that is often thought to belong to a class of highly inflexible DNAs; two strong nucleosome positioning sequences that share a set of particular sequence features common to nucleosome-preferring DNAs; and a CG-rich sequence representative of high G+C-content genomic regions that correlate with high nucleosome occupancy in vivo. To measure the flexibility of these sequences in the context of DNA looping, we combine the in vitro single-molecule tethered particle motion assay, a canonical looping protein, and a statistical mechanical model that allows us to quantitatively relate the looping probability to the looping free energy. We show that, in contrast to the case of nucleosome occupancy, G+C content does not positively correlate with looping probability, and that despite sharing sequence features that are thought to determine nucleosome affinity, the two strong nucleosome positioning sequences behave markedly dissimilarly in the context of looping. Most surprisingly, the poly(dA:dT)-rich DNA that is often characterized as highly inflexible in fact exhibits one of the highest propensities for looping that we have measured. These results argue for a need to revisit our understanding of the mechanical properties of DNA in a way that will provide a basis for understanding DNA deformation over the entire range of biologically relevant scenarios that are impacted by DNA deformability.

Comparative analysis of regulatory elements between Escherichia coli and Klebsiella pneumoniae by genome-wide transcription start site profiling

Genome-wide transcription start site (TSS) profiles of the enterobacteria Escherichia coli and Klebsiella pneumoniae were experimentally determined through modified 5′ RACE followed by deep sequencing of intact primary mRNA. This identified 3,746 and 3,143 TSSs for E. coli and K. pneumoniae, respectively. Experimentally determined TSSs were then used to define promoter regions and 5′ UTRs upstream of coding genes. Comparative analysis of these regulatory elements revealed the use of multiple TSSs, identical sequence motifs of promoter and Shine-Dalgarno sequence, reflecting conserved gene expression apparatuses between the two species. In both species, over 70% of primary transcripts were expressed from operons having orthologous genes during exponential growth. However, expressed orthologous genes in E. coli and K. pneumoniae showed a strikingly different organization of upstream regulatory regions with only 20% identical promoters with TSSs in both species. Over 40% of promoters had TSSs identified in only one species, despite conserved promoter sequences existing in the other species. 662 conserved promoters having TSSs in both species resulted in the same number of comparable 5′ UTR pairs, and that regulatory element was found to be the most variant region in sequence among promoter, 5′ UTR, and ORF. In K. pneumoniae, 48 sRNAs were predicted and 36 of them were expressed during exponential growth. Among them, 34 orthologous sRNAs between two species were analyzed in depth, and the analysis showed that many sRNAs of K. pneumoniae, including pleiotropic sRNAs such as rprA, arcZ, and sgrS, may work in the same way as in E. coli. These results reveal a new dimension of comparative genomics such that a comparison of two genomes needs to be comprehensive over all levels of genome organization.

In order to investigate similarities and differences of closely related species, most of the comparative genomics studies focus on comparing the gene contents either shared or specific for each genome. However, it is also important to investigate the differences in non-coding regulatory elements because they influence the transcriptional and post-transcriptional processes. Thus, we performed a genome-wide profiling of transcription start sites (TSSs) in two species, E. coli K-12 MG1655 and K. pneumoniae MGH78578. Experimental identification of TSSs is important for precise definition of promoter regions and 5′ untranslated regions upstream of coding genes. Comparative analysis of these regulatory elements revealed the use of multiple TSSs, identical sequence motifs of promoter and Shine-Dalgarno sequence. However, we observed that the upstream regulatory regions of the majority of operons having orthologous genes were organized with different usage of promoters and TSSs, resulting in diverse and complex gene regulation. We also found that the 5′ UTR is the least conserved regulatory element in sequence between the two species. Moreover, 34 orthologous sRNAs between E. coli and K. pneumoniae were analyzed in depth. The analysis suggested many of K. pneumoniae sRNAs might regulate the target genes as in E. coli.

DNA structural properties in the classification of genomic transcription regulation elements

It has been long known that DNA molecules encode information at various levels. The most basic level comprises the base sequence itself and is primarily important for the encoding of proteins and direct base recognition by DNA-binding proteins. A more elusive level consists of the local structural properties of the DNA molecule wherein the DNA sequence only plays an indirect supportive role. These properties are nevertheless an important factor in a large number of biomolecular processes and can be considered as informative signals for the presence of a variety of genomic features. Several recent studies have unequivocally shown the benefit of relying on such DNA properties for modeling and predicting genomic features as diverse as transcription start sites, transcription factor binding sites, or nucleosome occupancy. This review is meant to provide an overview of the key aspects of these DNA conformational and physicochemical properties. To illustrate their potential added value compared to relying solely on the nucleotide sequence in genomics studies, we discuss their application in research on transcription regulation mechanisms as representative cases.

Interference of intrinsic curvature of DNA by DNA-intercalating agents

It has been demonstrated in our studies that the intrinsic curvature of DNA can be easily interrupted by low concentrations of chloroquine and ethidium bromide. In addition, the changes of DNA curvature caused by varying the concentration of these two DNA intercalators can be readily verified through using an atomic force microscope.

The unique structure of A-tracts and intrinsic DNA bending

Short runs of adenines are a ubiquitous DNA element in regulatory regions of many organisms. When runs of 4–6 adenine base pairs (‘A-tracts’) are repeated with the helical periodicity, they give rise to global curvature of the DNA double helix, which can be macroscopically characterized by anomalously slow migration on polyacrylamide gels. The molecular structure of these DNA tracts is unusual and distinct from that of canonical B-DNA. We review here our current knowledge about the molecular details of A-tract structure and its interaction with sequences flanking them of either side and with the environment. Various molecular models were proposed to describe A-tract structure and how it causes global deflection of the DNA helical axis. We review old and recent findings that enable us to amalgamate the various findings to one model that conforms to the experimental data. Sequences containing phased repeats of A-tracts have from the very beginning been synonymous with global intrinsic DNA bending. In this review, we show that very often it is the unique structure of A-tracts that is at the basis of their widespread occurrence in regulatory regions of many organisms. Thus, the biological importance of A-tracts may often be residing in their distinct structure rather than in the global curvature that they induce on sequences containing them.

Computational analysis of looping of a large family of highly bent DNA by LacI

Sequence-dependent intrinsic curvature of DNA influences looping by regulatory proteins such as LacI and NtrC. Curvature can enhance stability and control shape, as observed in LacI loops formed with three designed sequences with operators bracketing an A-tract bend. We explore geometric, topological, and energetic effects of curvature with an analysis of a family of highly bent sequences, using the elastic rod model from previous work. A unifying straight-helical-straight representation uses two phasing parameters to describe sequences composed of two straight segments that flank a common helically supercoiled segment. We exercise the rod model over this two-dimensional space of phasing parameters to evaluate looping behaviors. This design space is found to comprise two subspaces that prefer parallel versus anti-parallel binding topologies. The energetic cost of looping varies from 4 to 12 kT. Molecules can be designed to yield distinct binding topologies as well as hyperstable or hypostable loops and potentially loops that can switch conformations. Loop switching could be a mechanism for control of gene expression. Model predictions for linking numbers and sizes of LacI-DNA loops can be tested using multiple experimental approaches, which coupled with theory could address whether proteins or DNA provide the observed flexibility of protein-DNA loops.

Predicted transcription factor binding sites as predictors of operons in Escherichia coli and Streptomyces coelicolor

Background

As a polycistronic transcriptional unit of one or more adjacent genes, operons play a key role in regulation and function in prokaryotic biology, and a better understanding of how they are constituted and controlled is needed. Recent efforts have attempted to predict operonic status in sequenced genomes using a variety of techniques and data sources. To date, non-homology based operon prediction strategies have mainly used predicted promoters and terminators present at the extremities of transcriptional unit as predictors, with reasonable success. However, transcription factor binding sites (TFBSs), typically found upstream of the first gene in an operon, have not yet been evaluated.

Results

Here we apply a method originally developed for the prediction of TFBSs in Escherichia coli that minimises the need for prior knowledge and tests its ability to predict operons in E. coli and the 'more complex', pharmaceutically important, Streptomyces coelicolor. We demonstrate that through building genome specific TFBS position-specific-weight-matrices (PSWMs) it is possible to predict operons in E. coli and S. coelicolor with 83% and 93% accuracy respectively, using only TFBS as delimiters of operons. Additionally, the 'palindromicity' of TFBS footprint data of E. coli is characterised.

Conclusion

TFBS are proposed as novel independent features for use in prokaryotic operon prediction (whether alone or as part of a set of features) given their efficacy as operon predictors in E. coli and S. coelicolor. We also show that TFBS footprint data in E. coli generally contains inverted repeats with significantly (p < 0.05) greater palindromicity than random sequences. Consequently, the palindromicity of putative TFBSs predicted can also enhance operon predictions.

Identification of a nisI promoter within the nisABCTIP operon that may enable establishment of nisin immunity prior to induction of the operon via signal transduction

Certain strains of Lactococcus lactis produce the broad-spectrum bacteriocin nisin, which belongs to the lantibiotic class of antimicrobial peptides. The genes encoding nisin are organized in three contiguous operons: nisABTCIP, encoding production and immunity (nisI); nisRK, encoding regulation; and nisFEG, also involved in immunity. Transcription of nisABTCIP and nisFEG requires autoinduction by external nisin via signal transducing by NisRK. This organization poses the intriguing question of how sufficient immunity (NisI) can be expressed when the nisin cluster enters a new cell, before it encounters external nisin. In this study, Northern analysis in both Lactococcus and Enterococcus backgrounds revealed that nisI mRNA was present under conditions when no nisA transcription was occurring, suggesting an internal promoter within the operon. The nisA transcript was significantly more stable than nisI, further substantiating this. Reverse transcriptase PCR analysis revealed that the transcription initiated just upstream from nisI. Fusing this region to a lacZ gene in a promoter probe vector demonstrated that a promoter was present. The transcription start site (TSS) of the nisI promoter was mapped at bp 123 upstream of the nisI translation start codon. Ordered 5' deletions revealed that transcription activation depended on sequences located up to bp -234 from the TSS. The presence of poly(A) tracts and computerized predictions for this region suggested that a high degree of curvature may be required for transcription initiation. The existence of this nisI promoter is likely an evolutionary adaptation of the nisin gene cluster to enable its successful establishment in other cells following horizontal transfer.

Protein-induced DNA bending clarifies the architectural organization of the sigma54-dependent glnAp2 promoter

Sigma54-RNA polymerase (Esigma54) predominantly contacts one face of the DNA helix in the closed promoter complex, and interacts with the upstream enhancer-bound activator via DNA looping. Up to date, the precise face of Esigma54 that contacts the activator to convert the closed complex to an open one remains unclear. By introducing protein-induced DNA bends at precise locations between upstream enhancer sequences and the core promoter of the sigma54-dependent glnAp2 promoter without changing the distance in-between, we observed a strong enhanced or decreased promoter activity, especially on linear DNA templates in vitro. The relative positioning and orientations of Esigma54, DNA bending protein and enhancer-bound activator on linear DNA were determined by in vitro footprinting analysis. Intriguingly, the locations from which the DNA bending protein exerted its optimal stimulatory effects were all found on the opposite face of the DNA helix compared with the DNA bound Esigma54 in the closed complex. Therefore, these results provide evidence that the activator must approach the Esigma54 closed complexes from the unbound face of the promoter DNA helix to catalyse open complex formation. This proposal is further supported by the modelling of activator-promoter DNA-Esigma54 complex.

Identification of an upstream regulatory sequence that mediates the transcription of mox genes in Methylobacterium extorquens AM1

A multiple A-tract sequence has been identified in the promoter regions for the mxaF, pqqA, mxaW, mxbD and mxcQ genes involved in methanol oxidation in Methylobacterium extorquens AM1, a facultative methylotroph. Site-directed mutagenesis was exploited to delete or change this conserved sequence. Promoter-xylE transcriptional fusions were used to assess promoter activity in these mutants. A fiftyfold drop in the XylE activity was observed for the mxaF and pqqA promoters without this sequence, and a five- to sixfold drop in the XylE activity was observed for the mxbD and mxcQ promoters without this sequence. Mutants were generated in the chromosomal copies in which this sequence was either deleted or altered, and these mutants were unable to grow on methanol. When one of these sequences was added to Plac of Escherichia coli, which is a weak constitutive promoter in M. extorquens AM1, the activity increased two- to threefold. These results suggest that this sequence is essential for normal expression of these genes in M. extorquens AM1, and may serve as a general enhancer element for genetic constructs in this bacterium.

Geometric and dynamic requirements for DNA looping, wrapping and unwrapping in the activation of E.coli glnAp2 transcription by NtrC

Transcriptional activation by the E.coli NtrC protein can occur via DNA looping between a DNA-bound activator and the target sigma(54) RNA polymerase. NtrC forms an octamer on DNA that is capable of binding two DNA molecules. Its ATPase activity is required for open complex formation. Geometric requirements for activation were assessed using a library of DNA bending sequences created by random ligation of A-tract oligonucleotides, as well as several designed sequences. Thirty random or designed sequences with a variety of DNA lengths and bending geometries were cloned in plasmids, and the library was used to replace the spacer between the NtrC binding sites and the core glnAp2 promoter. The activity of each promoter construct under nitrogen limitation was determined in vivo, in a lambda phage lacZ reporter system integrated as a single-copy lysogen to avoid titrating NtrC or polymerase. A wide variety of bending geometries was found to support a similar level of transcriptional activation ( approximately 3-4-fold). Computer modeling of the DNA trajectories suggests that the most inactive promoters have short spacer DNA and the NtrC sites on the opposite side of the helix as the wild-type sites; otherwise, the loop can form effectively. Flexibility and multivalency of the NtrC-Esigma(54) interaction apparently provides substantial independence from DNA stiffness constraints, and in general activation requires less efficient looping than repression. However, none of the random templates were as active as wild-type promoter. Subsidiary activator binding sites in the wild-type were found to be required for full activity, but, surprisingly, these sites could not be functionally replaced by strong binding sites. This suggests that one or more protomers in the NtrC octamer must form and then release contacts with DNA in order to complete the ATPase cycle and act as an AAA(+) activator of the Esigma(54). This dynamic DNA wrapping around the NtrC octamer is proposed to be necessary for efficient activation, and the wrapping may also reduce adventitious activation of other promoters.

Binding affinity of Escherichia coli RNA polymerase*sigma54 holoenzyme for the glnAp2, nifH and nifL promoters

Escherichia coli RNA polymerase associated with the sigma54 factor (RNAP*sigma54) is a holoenzyme form that transcribes a special class of promoters not recognized by the standard RNA polymerase*sigma70 com plex. Promoters for RNAP*sigma54 vary in their overall 'strength' and show differences in their response to the presence of DNA curvature between enhancer and promoter. In order to examine whether these effects are related to the promoter affinity, we have determined the equilibrium dissociation constant K(d) for the binding of RNAP*sigma54 to the three promoters glnAp2, nifH and nifL. Binding studies were conducted by monitoring the changes in fluorescence anisotropy upon titrating RNAP*sigma54 to carboxyrhodamine-labeled DNA duplexes. For the glnAp2 and nifH promoters similar values of K(d) = 0.94 +/- 0.55 nM and K(d) = 0.85 +/- 0.30 nM were determined at physiological ionic strength, while the nifL promoter displayed a significantly weaker affinity with K(d) = 8.5 +/- 1.9 nM. The logarithmic dependence of K(d) on the ionic strength I was -Deltalog(K(d))/Deltalog(I) = 6.1 +/- 0.5 for the glnAp2, 5.2 +/- 1.2 for the nifH and 2.1 +/- 0.1 for the nifL promoter. This suggests that the polymerase can form fewer ion pairs with the nifL promoter, which would account for its weaker binding affinity.

Intrinsic DNA distortion of the bacteriophage Mu momP1 promoter is a negative regulator of its transcription. A novel mode of regulation of toxic gene expression

The momP1 promoter of the bacteriophage Mu mom operon is an example of a weak promoter. It contains a 19-base pair suboptimal spacer between the -35 (ACCACA) and -10 (TAGAAT) hexamers. Escherichia coli RNA polymerase is unable to bind to momP1 on its own. DNA distortion caused by the presence of a run of six T nucleotides overlapping the 5' end of the -10 element might prevent RNA polymerase from binding to momP1. To investigate the influence of the T(6) run on momP1 expression, defined substitution mutations were introduced by site-directed mutagenesis. In vitro probing experiments with copper phenanthroline ((OP)(2)Cu) and DNase I revealed distinct differences in cleavage patterns among the various mutants; in addition, compared with the wild type, the mutants showed an increase (variable) in momP1 promoter activity in vivo. Promoter strength analyses were in agreement with the ability of these mutants to form open complexes as well as to produce momP1-specific transcripts. No significant role is attributed to the overlapping and divergently organized promoter, momP2, in the expression of momP1 activity, as determined by promoter disruption analysis. These data support the view that an intrinsic DNA distortion in the spacer region of momP1 acts in cis as a negative element in mom operon transcription. This is a novel mechanism of regulation of toxic gene expression.

Analysis of sequence-dependent curvature in matrix attachment regions

Sequence-dependent DNA conformations of matrix attachment regions (MARs) available in a database were calculated using the wedge model, and compared with randomly chosen genes, promoters, enhancers and transposons. The MARs had a longer bent part and higher angle/helical turn than the other regions. It is known that some MAR sequences have A-tracts that cause DNA bending, and we also found many A-tracts in examined MARs. Furthermore, non-random and clustered distribution of A-tracts shown here gave further evidence of the importance of A-tracts for MAR conformations. These results suggest that DNAs of MARs have a characteristic conformation instead of conserved sequence.

Identification and mapping of sigma-54 promoters in Chlamydia trachomatis

The first sigma(54) promoters in Chlamydia trachomatis L2 were mapped upstream of hypothetical proteins CT652.1 and CT683. Comparative genomics indicated that these sigma(54) promoters and potential upstream activation binding sites are conserved in orthologous C. trachomatis D, C. trachomatis mouse pneumonitis strain, and Chlamydia pneumoniae (CWL029 and AR39) genes.

The effect of the DNA conformation on the rate of NtrC activated transcription of Escherichia coli RNA polymerase.sigma(54) holoenzyme

The transcription activator protein NtrC (nitrogen regulatory protein C) can catalyze the transition of Escherichia coli RNA polymerase complexed with the sigma 54 factor (RNAP.sigma(54)) from the closed complex (RNAP.sigma(54) bound at the promoter) to the open complex (melting of the promoter DNA). This process involves phosphorylation of NtrC (NtrC-P), assembly of an octameric NtrC-P complex at the enhancer sequence, interaction of this complex with promoter-bound RNAP.sigma(54) via DNA looping, and hydrolysis of ATP. We have used this system to study the influence of the DNA conformation on the transcription activation rate in single-round transcription experiments with superhelical plasmids as well as linearized templates. Most of the templates had an intrinsically curved DNA sequence between the enhancer and the promoter and differed with respect to the location of the curvature and the distance between the two DNA sites. The following results were obtained: (i) a ten- to 60-fold higher activation rate was observed with the superhelical templates as compared to the linearized conformation; (ii) the presence of an intrinsically curved DNA sequence increased the activation rate of linear templates about five times; (iii) no systematic effect for the presence and/or location of the inserted curved sequence was observed for the superhelical templates. However, the transcription activation rate varied up to a factor of 10 between some of the constructs. (iv) Differences in the distance between enhancer and promoter had little effect for the superhelical templates studied. The results were compared with theoretical calculations for the dependence of the contact probability between enhancer and promoter expressed as the molar local concentration j(M). A correlation of j(M) with the transcription activation rate was observed for values of 10(-8) M<j(M)<10(-6) M and a kinetic model for NtrC-P-catalyzed open complex formation was developed.

The global intrinsic curvature of archaeal and eubacterial genomes is mostly contained in their dinucleotide composition and is probably not an adaptation

Until now, the genomic DNA of all eubacteria analyzed has been hyper-curved, its global intrinsic curvature being higher than that of a random sequence. In contrast, that rule failed for archaea or eukaryotes, which could be either hypo- or hyper-curved. The existence of the rule suggested that, at least for eubacteria, global intrinsic curvature is adaptive. However, the present results from analyzing 21 eubacterial and six archaeal genomes argue against adaptation. First, there are two eubacterial exceptions to the former rule. More significantly, we found that the dinucleotide composition of the genome alone (which lacks all sequence information) is enough to determine the genome curvature. Additional evidence against adaptation came from showing that the global curvature of bacterial genomes could not have evolved under either of two complementary models of curvature selection: (i) that curvature is selected locally from unbiased variability; (ii) that curvature is established globally through the selection of a curvature-altering mutational bias. We found that the observed relationship between curvature and dinucleotide composition is incompatible with model (i). We also found that, contrary to the predictions of model (ii), the dinucleo-tide compositions of bacterial genomes were not statistically special in their curvature-related properties (when compared to stochastically generated dinucleotide compositions).