Protein-DNA complexes with loops play a fundamental role in a wide variety of cellular processes, ranging from the regulation of DNA transcription to telomere maintenance. As ubiquitous as they are, their precise in vivo properties and their integration into the cellular function still remain largely unexplored. Here, we present a multilevel approach that efficiently connects in both directions molecular properties with cell physiology and use it to characterize the molecular properties of the looped DNA-lac repressor complex while functioning in vivo. The properties we uncover include the presence of two representative conformations of the complex, the stabilization of one conformation by DNA architectural proteins, and precise values of the underlying twisting elastic constants and bending free energies. Incorporation of all this molecular information into gene-regulation models reveals an unprecedented versatility of looped DNA-protein complexes at shaping the properties of gene expression.
Citation: Saiz L, Vilar JM (2007) Multilevel Deconstruction of the In Vivo Behavior of Looped DNA-Protein Complexes. PLoS ONE 2(4): e355. https://doi.org/10.1371/journal.pone.0000355
Academic Editor: Martin Egli, Vanderbilt University, United States of America
Received: December 18, 2006; Accepted: March 14, 2007; Published: April 4, 2007
Copyright: © 2007 Saiz, Vilar. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
Many fundamental cellular processes, including transcription regulation, recombination, replication, gene silencing, and telomere maintenance, rely on the formation of DNA loops and higher order looped DNA packing structures, such as chromatin looping –. In the regulation of gene expression, proteins bound far away from the genes they control can be brought to the initiation of transcription region by looping the intervening DNA. The free energy cost of this process determines how easily DNA loops can form and therefore the extent to which distal DNA sites affect each other . Assessing directly the in vivo value of the free energy of DNA looping is remarkably difficult, not only because the properties of the components can change when studied in vitro, but also because the in vivo probing of the cell can perturb the process under study .
Computational and mathematical models of gene regulation provide an avenue to connect the physical properties of DNA in its in vivo natural environment with the resulting cellular behavior , . This type of approach was used recently to infer the in vivo free energies of DNA looping by the lac repressor as a function of the loop length  from measurements of enzyme production in the lac operon , which proved to be a very accurate alternative to obtain molecular properties of the macromolecular complexes in vivo. The results of this analysis  showed that the free energy for short loops oscillates with the helical periodicity of DNA, as expected, because the operators must have the right phase to bind simultaneously to the repressor ,  and, unexpectedly, that the free energy in a cycle behaves asymmetrically. A Fourier analysis of the oscillations indicated that this asymmetry can be characterized by a second representative oscillatory component with a period of ∼5.6 bp, in addition to the component with the in vivo helical period (∼10.9 bp). Another striking feature of the in vivo free energy of looping is that the amplitude of the oscillations is as small as ∼2.5 kcal/mol, similar to the typical free energy of cooperative interactions between regulatory molecules .
Uncovering the origin of the in vivo properties is important for understanding DNA looping and its effects in gene regulation, especially because current theories based on semiflexible polymer models of DNA predict symmetric and, at least, twice as large oscillations , . Different contributions, such as the anisotropic flexibility of DNA, local features resulting from the DNA sequence , and interactions with the lac repressor  and other DNA binding proteins, might be at play. Another potential source of complexity is the number of trajectories that DNA can follow to loop –. Thus, the observed behavior could be the result of loops with several representative conformations (Figure 1). Yet, only the lowest free energy conformation is typically considered.
Two plausible alternative loop conformations of the lac repressor-DNA complex. The bidentate repressor, with the two dimers that form the functional tetramer shown in red, simultaneously binds DNA, colored orange, at two sites. The two structures represent two plausible trajectories of the DNA loop and two plausible conformations of the lac repressor (V-shaped and extended).
Here, we develop a statistical thermodynamics approach to deconstruct the observed behavior of the expression of the lac operon in Escherichia coli cells and use it to obtain the in vivo properties of DNA looping by the lac repressor at different levels of cellular organization. At the molecular level, we propose an elastic model for DNA loop formation that considers multiple structures of the DNA-protein complex and show that, at the cellular level, the in vivo behavior of the free energy of looping is accurately accounted for by the presence of two distinct types of DNA loops, corresponding to two main looped DNA-protein conformations, with different relative optimal free energies, phases, and interactions with key architectural proteins. We explore in detail the effects of multiple conformations on shaping the free energy of looping DNA and the consequences that resulting free energies have for gene regulation at the cell-population level.
A multi-conformation elastic DNA model
We consider that the DNA loop can be in two distinct representative conformations (Figure 1) through the free energy of looping ΔGl, which can be expressed in terms of the free energy of each of the conformations as (see Methods)1where the index i indicates whether the loop is in the conformation labeled 1 or 2 and RT (≈0.6 kcal/mol) is the gas constant, R, times the absolute temperature, T. The integer index n ranges from-infinity to +infinity and accounts for the 2π degeneracy in the twisting angle. In general, a system could have M representative conformations of the nucleoprotein-DNA complex and the summation in the previous expression of i would extend from 1 to M (see Methods for the general case).
The free energy of a particular state includes bending and twisting contributions and is given following the classic elasticity theory of DNA  by2where L is the length of the loop (in bp), Lopt,i is the optimal spacing or phase (in bp), and ΔG0,i is the corresponding optimal free energy (in kcal/mol), which depends on the type (i) of loop formed. In principle, the term ΔG0,i could also depend on L because of the bending contribution  but the in vivo results  indicate that it is practically constant for the range of lengths analyzed. The twisting force constant (in kcal/mol bp), C, and the in vivo helical repeat (in bp), hr, are considered here to be the same for the two types of loops. The free energy ΔGi of a conformation i is given by the equality , which includes the sum over the states of a loop conformation.
In vivo free energy of DNA looping: complex average behavior from simple individual contributions
The free energy of looping ΔGl given by Equations 1 and 2 closely reproduces the broad range of observed types of behavior (Figure 2), which consist of the in vivo free energies of looping DNA  obtained from the measured repression levels (see Methods) for two wild type situations ,  and a mutant lacking the architectural HU protein . The in vivo free energies display not only asymmetric oscillations with reduced amplitude but also plateaus and secondary maxima. Therefore, our model indicates that the complex behavior of DNA looping in vivo emerges from a combination of the simple behavior of the individual conformations rather than from the individual conformations themselves.
Two-conformation analysis of the in vivo free energy of DNA looping. The in vivo free energy of looping DNA by the lac repressor (blue symbols) was obtained as described in Saiz et al.  (see also Methods) from the measured repression levels of Muller et al.  for wild type (WT1) and of Becker et al.  for wild type (WT2) and a mutant that does not express the architectural HU protein (ΔHU). As repression levels in the absence of looping (see Methods and Saiz et al. ) we have used 135 (WT1), 2.3 (WT2), and 1.7 (ΔHU). The thick black continuous lines correspond in each case to the best fit to the free energy ΔGl given by Equations 1 and 2, which considers the contributions of two looped conformations. The contributions of each conformation are shown separately as red () and black () dashed lines. The values of the parameters for the best fit are shown in Table 1.
The values of the parameters for the best fit (continuous black thick curves) to the data inferred from the experiments (blue symbols) are gathered in Table 1. The free energy of looping for each conformation (Figure 2 in dashed red and gray for the conformation with lowest and highest optimal free energy, respectively) depends on the length of the loop as expected for an elastic rod model of DNA, displaying symmetric oscillations with the periodicity of the DNA helix and relatively high amplitudes. The magnitudes of the amplitudes, in the order of 5 kcal/mol, are in excellent agreement with recent sequence-dependent DNA elasticity calculations for different types of lac repressor-DNA loops , which lead to oscillations of ∼6 kcal/mol. Another interesting feature is the lack of a sharp increase of the looping free energy for short loops, which would be expected to arise from the bending free energy contribution. The observed behavior might originate from the high flexibility of the repressor  in the extended conformations  or from the interaction of the DNA loop with architectural proteins that help bending .
In both wild type situations analyzed (Figure 2, WT1 and WT2), the presence of two looped conformations (one more stable than the other by 1.0 kcal/mol and with shifts in the optimal phases of 4.3 bp or −4.2 bp) is responsible for the reduced amplitude of the oscillations and the asymmetry, including secondary maxima and/or shoulders, of the free energy curves. The inferred in vivo data from the two experiments is in excellent agreement with the two-conformation analysis (compare experimental blue symbols and model black thick lines in Figure 2). Our results indicate that the behavior of the in vivo system depends strongly on the properties of the different loop conformations, especially on the optimal free energies and optimal phases (Table 1).
Note that optimal phases and free energies between the two conformations are different for different wild type experiments (WT1 and WT2 in Table 1). These differences might arise from the differences in the experimental conditions, which are significant. For instance, the repression level in the absence of DNA looping is 135 for WT1 and 2.3 for WT2. They can also be due to potentially different boundary conditions because the loop is formed between the ideal and the main operator O1 in WT1 and between the ideal and the auxiliary operator O2 in WT2. The main operator is both more symmetric and ∼10 times stronger than O2.
Optimal energies and phases determine the relative contributions of the different conformations to the observed behavior and how they change with the length of the loop. Explicitly, the probabilities for each conformation to be present, P1 and P2, are related to each other through the expression , which results from the general principles of statistical thermodynamics . As the distance between the two operators is changed, the less stable loop can become the most stable one. In some cases, such as for those loop lengths for which both conformations have the same free energies (when red and gray curves in Figure 2 intersect each other), the two structures are equally probable and both conformations alternate in time in a single cell and occur simultaneously in a population of cells. In the other cases, when the difference is larger than RT, the conformation with the lowest free energy dominates over the other one.
These two conformations of the DNA-protein looped complex, whose elastic properties we have characterized in detail, could consist of two ways of binding of the repressor to DNA, such as antiparallel and parallel DNA trajectories, which for a specific repressor conformation, i.e., the typical V-shape observed in the crystalline state , , would give rise in principle to four different loop geometries . Similarly, they could correspond to two different conformations of the lac repressor; namely, the V-shaped repressor and the extended conformation proposed from electron microscopy and fluorescence resonance energy transfer experiments in solution , .
Effects of architectural proteins
The free energy of looping DNA in vivo determines the cost of forming the loop in the natural environment of the cell, which includes the double-stranded DNA molecule, the proteins that tie the DNA loops, other DNA binding proteins, and the different proteins confined within the E. coli cell. Architectural proteins both in eukaryotes and prokaryotes play an important role in assisting the assembly of nucleoprotein complexes and contribute to the control of gene expression as well as other DNA transactions –. These proteins locally bend or kink DNA facilitating the formation of protein-DNA looped structures – and thus are expected to affect the DNA looping properties in vivo. In particular, the stability of different types of looped DNA-lac repressor conformations has been shown to be affected by binding of the catabolite activator protein , . Other bacterial architectural proteins, such as the heat unstable nucleoid protein (HU) also referred to as histone-like protein, do not have sequence specific DNA binding sites but also bend DNA.
In the E. coli mutant without architectural HU protein (Figure 2, ΔHU), the in vivo free energy of DNA looping is compatible with the presence of two loop conformations that are similarly stable (0.2 kcal/mol difference) but have different optimal phases. In this case, the phase shift (3.3 bp) also leads to reduced amplitude of the oscillations, as in the wild-type case where HU protein is normally expressed, yet the asymmetric behavior is practically lost; now the presence of two loop conformations results in almost-symmetric oscillations with smaller amplitude. Comparison between wild type and ΔHU mutant results (Table 1) indicates that architectural proteins lower the optimal free energy of one conformation, leading to subtle differences of ∼1 kcal/mol between the two most stable conformations. In systems like the Gal repressosome , the architectural HU protein is required to form the loop, which implies strong stabilizing effects and a single dominant conformation. In the lac operon, in contrast, we find that both HU stabilized and non-stabilized conformations contribute to the free energy of looping (Figure 2), which is responsible for the observed asymmetric behavior.
In all three cases studied here (Figure 2 and Table 1), the results obtained for the apparent in vivo twisting force constants, which also include the contributions from the repressor, are in the range 48–68 kcal/mol bp. These twisting force constants are a factor 2 smaller than the canonical value ,  of 105 kcal/mol bp or 2.5×10−19 erg cm, and are similar to those reported recently in cyclization experiments .
Shaping the behavior of the two-conformation free energy of looping
Our analysis has shown that the complex behavior of the in vivo free energy of looping is accurately accounted for by combination of the rather simple behavior of two representative looped conformations (Figure 2). The major differences observed between wild type and the mutant without architectural HU protein arise mainly from the way in which the two conformations are combined; namely, from the differences in the optimal free energies and optimal phases between the two conformations. To explore the potential types of behavior that can arise when two conformations are combined, we have computed the free energy of looping by taking as reference the values of the parameters obtained for wild type (Table 1, WT1), keeping the values for one conformation fixed (conformation 1 of WT1), and systematically changing the values for the other one (Figure 3).
Free energy of looping for a two-conformation elastic DNA model. Different types of behavior are obtained by changing two key parameters: the difference in optimal free energies (ΔG0,1–ΔG0,2) and optimal phases (Lopt,1–Lopt,2). (A) The difference in optimal free energies between the two configurations increases from 0 kcal/mol (blue) to 1.5 kcal/mol (red) in increments of 0.5 kcal/mol whereas the difference in optimal phases is kept fixed at 4.2 bp. (B, C) The difference in optimal phases between the two conformations increases from −5.5 bp (blue) to 0 bp (red) in increments of 5.5/3 bp whereas the difference in optimal free energies is kept fixed at 1 kcal/mol (B) and at 0 kcal/mol (C).
As the optimal free energy difference between conformations increases (Figure 3A), the behavior of the free energy changes from symmetric multiwell and wide minima, as in the ΔHU mutant, through asymmetric, typical of the wild type system, to symmetric with high amplitude oscillations (curve not shown), typical of “single-conformation” systems. A similarly broad range of types of behavior is also obtained when the difference between optimal phases changes. We have considered these changes in the context of two differences between optimal free energies: ∼1.0 kcal/mol, like in wild type (Figure 3B), and ∼0.0 kcal/mol, like in ΔHU mutants (Figure 3C). In both cases, as the difference between the optimal phases decreases, the amplitude of the oscillations increases. In the wild type-like situation, the oscillations of the free energy are asymmetric except for precisely tuned values of the parameters. In the ΔHU mutant-like situation, the oscillations are symmetric, and for precisely tuned values of the parameters, it is even possible to obtain oscillations with a period of half the helicity of DNA (Figure 3C, blue curve). All these results together show that the experimentally observed free energies of looping, as diverse as they are, provide just three examples of an even richer number of potential types of behavior.
Across multiple levels: from DNA looping to gene regulation and cellular physiology
The high versatility of multi-conformation protein-DNA complexes at shaping the free energy of looping DNA propagates to the cell physiology through the effects of DNA looping in gene regulation. In a similar way as we have inferred and analyzed the in vivo free energy of DNA looping, we can predict the effect of a given free energy of looping on gene regulation by inverting the mathematical expression that connects the free energy of looping with the repression levels for the lac operon (see Methods). Explicitly, given the repression level for the system with a single operator (Rnoloop), the repression level for two operators with looping follows from the free energy of looping through the expression3where [N] is the concentration of repressors.
As in the DNA looping free energy (Figure 3), the precise values of the differences in the optimal free energies and optimal phases between the two conformations strongly affect the repression level (Figure 4), leading also to a large variety of types of behaviors and degrees of repression. In general, the typical asymmetry of the free energy is less marked in the repression level (Figure 4A), to the extent that it might not be obvious in the raw experimental data, as happens in the classical experiments on the repression of the lac operon . This loss of features leads to robust repression levels with respect to changes in the optimal phase (Figure 4B) when the optimal free energies differences are similar to the wild type value (∼1 kcal/mol), whereas such robustness is not present when the optimal free energies of both conformations are similar (Figure 4C). The particular shape can thus be controlled in vivo by the HU architectural protein to produce either robust or sensitive gene expression patterns.
Effects in gene expression of the free energy of looping for a two-conformation elastic DNA model. Repression levels obtained with Equation 3 using the corresponding free energies of Figure 3. (A) Differences in optimal free energies are varied and the optimal phases are kept fixed. (B, C) Differences in optimal phases are varied and optimal free energies are kept fixed at two different values: 1 kcal/mol (B) and 0 kcal/mol (C).
Computational and mathematical methods provide a unique avenue to connect cellular physiology with molecular properties in a living organism , , , . The statistical thermodynamics approach we have developed to deconstruct the observed behavior of the expression of the lac operon in E. coli cells has allowed us to obtain the in vivo properties of DNA looping by the lac repressor at different levels of biological organization.
It was previously shown that classic experimental data on the expression of the lac operon in cell populations led to an unexpected, rather complex, behavior of the free energy of looping DNA in vivo, with small-amplitude asymmetric oscillations as a function of the length of the loop . Here, we have shown that this striking behavior has its molecular origin in the ability of the lac repressor to loop DNA in vivo in at least two different ways. Thus, the intricate in vivo behavior of the free energy of looping is the result of combining the relatively simple behavior of each of the two looped conformations. These two types of loops have different properties and interact distinctly with the HU architectural protein. Explicitly, we found that DNA loops that interact with the HU architectural protein are ∼1 kcal/mol more stable than loops that do not. Our approach has also allowed us to accurately obtain the elastic properties of the protein-DNA complexes in vivo, including twisting force constants, which turned out to be a factor 2 smaller than the canonical value of 105 kcal/mol bp (2.5×10−19 erg cm).
Our analysis of the effects of the molecular properties in the free energy of DNA looping at the cellular level, and their propagation to gene expression at the cell-population level, shows that there is a wide range of potential types of behavior that can arise from combining single-conformation free energies of looping. The mathematical expression for the free energy of looping (Equations 1 and 2) indicates that optimal free energies and phases, which in single-conformation systems affect only quantitative details, are key determinants of the qualitative behavior. In particular, the asymmetry in the oscillations is the consequence of the presence of a slightly preferred loop conformation with different optimal phase. Symmetric oscillations in the free energy result from equally stable loop conformations, a strongly dominant conformation, or conformations with the same optimal phases. In E. coli cells, as shown by our results, the HU protein preferentially affects one loop conformation making it slightly more stable, thus leading to the observed asymmetry.
Different loop trajectories have been observed in vitro for diverse nucleoprotein complexes , . In particular, in vitro experiments of DNA cleavage by the SfiI endonuclease, a type II restriction endonuclease that binds to two DNA sites as a tetramer by looping out the intervening DNA, have shown coexistence and alternative conformations  as the DNA spacer between binding sites is changed for loop sizes of 109–170 bp. They also observed similar periodicities for the two conformations as well as different phases in in vitro electrophoresis experiments. There are also studies on the Gal repressor showing that several non-simultaneous trajectories can exist and that there is a single configuration of the complex for a particular loop length when the HU protein is present . Our results provide evidence that shows, for the first time, that alternative and simultaneous nucleoprotein-DNA configurations are present in vivo.
At the cell-population level, whether the typical asymmetry of the free energy propagates to the repression level is controlled by the values of the optimal free energies and phases. In general, the asymmetry in the repression level is less marked than in the free energy, to the extent that it might not be obvious in the raw experimental data .
Our results indicate that the biological consequences of having two or more DNA-looped conformations include a reduced dependence on the positioning of the DNA binding sites. For instance, by combining two DNA conformations, it is possible to reduce the amplitude of the typical oscillations in the free energy as a function of the length of the loop from ∼5 kcal/mol to ∼1 kcal/mol (Figure 3C). In this way, DNA appears to the cell to be much more malleable than it actually is in a single conformation. The presence of multiple DNA conformations also provides an extra layer of control of the properties of gene regulation. In the case of the lac operon, we have shown that the HU protein stabilizes one DNA conformation. Similarly, it has also been shown that the Catabolite Activator Protein (CAP) stabilizes preferentially certain loop conformations . Thus, expression of CAP, HU, and other architectural proteins can change the DNA looping properties in a conformation-dependent manner and select the precise details of the interactions between distal DNA sites.
In broad terms, our analysis has revealed that the formation of DNA loops in vivo is tightly coupled to the molecular properties of the proteins and protein complexes that form the loop. There is a high versatility of looped DNA-protein complexes at establishing different conformations in the intracellular environment and at adapting from one conformation to another. This versatility underlies the unanticipated behavior of the in vivo free energy of DNA looping and can be responsible not only for asymmetric oscillations with decreased amplitude but also for plateaus and secondary maxima. All these features indicate that the physical properties of DNA can actively be selected to control the cooperative binding of regulatory proteins and to achieve different cellular behaviors.
Free energy of DNA looping from multiple conformations
Following the statistical thermodynamics approach , the free energy of looping, ΔGl, can be expressed in terms of the free energy for each individual conformation aswhere the right hand side of the equation has as many terms as the number of possible representative conformations of the looped DNA-protein complex. In practice, only the conformations with lowest free energy will have a significant effect in the observed behavior. In particular, we have shown that typically only two distinct conformations contribute significantly, and thus , which leads to for the free energy of DNA looping.
In vivo free energy of DNA looping from physiological measurements
The in vivo free energy of DNA looping by the lac repressor's binding to the main and an auxiliary operator can be expressed in terms of the measured repression levels through a well-established model for gene regulation by the lac repressor . For the experimental conditions consisting of a strong auxiliary operator, which are those of the experiments considered here, the free energy of looping DNA  for an inter-operator distance L is given by:where Rloop (L) is the measured repression level, a dimensionless quantity used to quantify the extent of repression of a gene; Rnoloop is the repression level in the absence of DNA looping; [N] is the concentration of repressors; and RT is the gas constant times the absolute temperature.
Designed research, performed research, analyzed data, and wrote the paper: LS JV.
- 1. Adhya S (1989) Multipartite genetic control elements: communication by DNA loop. Annu Rev Genet 23: 227–250.S. Adhya1989Multipartite genetic control elements: communication by DNA loop.Annu Rev Genet23227250
- 2. Matthews KS (1992) DNA Looping. Microbiological Reviews 56: 123–136.KS Matthews1992DNA Looping.Microbiological Reviews56123136
- 3. Schleif R (1992) DNA looping. Annu Rev Biochem 61: 199–223.R. Schleif1992DNA looping.Annu Rev Biochem61199223
- 4. Vilar JMG, Saiz L (2005) DNA looping in gene regulation: from the assembly of macromolecular complexes to the control of transcriptional noise. Current Opinion in Genetics & Development 15: 136–144.JMG VilarL. Saiz2005DNA looping in gene regulation: from the assembly of macromolecular complexes to the control of transcriptional noise.Current Opinion in Genetics & Development15136144
- 5. Li QL, Barkess G, Qian H (2006) Chromatin looping and the probability of transcription. Trends in Genetics 22: 197–202.QL LiG. BarkessH. Qian2006Chromatin looping and the probability of transcription.Trends in Genetics22197202
- 6. Talbert PB, Henikoff S (2006) Spreading of silent chromatin: inaction at a distance. Nature Reviews Genetics 7: 793–803.PB TalbertS. Henikoff2006Spreading of silent chromatin: inaction at a distance.Nature Reviews Genetics7793803
- 7. Ellis RJ (2001) Macromolecular crowding: obvious but underappreciated. Trends in Biochemical Sciences 26: 597–604.RJ Ellis2001Macromolecular crowding: obvious but underappreciated.Trends in Biochemical Sciences26597604
- 8. Bellomy GR, Mossing MC, Record MT (1988) Physical-Properties of DNA Invivo as Probed by the Length Dependence of the Lac Operator Looping Process. Biochemistry 27: 3900–3906.GR BellomyMC MossingMT Record1988Physical-Properties of DNA Invivo as Probed by the Length Dependence of the Lac Operator Looping Process.Biochemistry2739003906
- 9. Vilar JMG, Leibler S (2003) DNA looping and physical constraints on transcription regulation. J Mol Biol 331: 981–989.JMG VilarS. Leibler2003DNA looping and physical constraints on transcription regulation.J Mol Biol331981989
- 10. Saiz L, Rubi JM, Vilar JMG (2005) Inferring the in vivo looping properties of DNA. Proceedings of the National Academy of Sciences of the United States of America 102: 17642–17645.L. SaizJM RubiJMG Vilar2005Inferring the in vivo looping properties of DNA.Proceedings of the National Academy of Sciences of the United States of America1021764217645
- 11. Muller J, Oehler S, Muller-Hill B (1996) Repression of lac promoter as a function of distance, phase and quality of an auxiliary lac operator. J Mol Biol 257: 21–29.J. MullerS. OehlerB. Muller-Hill1996Repression of lac promoter as a function of distance, phase and quality of an auxiliary lac operator.J Mol Biol2572129
- 12. Lee DH, Schleif RF (1989) In vivo DNA loops in araCBAD: size limits and helical repeat. Proc Natl Acad Sci U S A 86: 476–480.DH LeeRF Schleif1989In vivo DNA loops in araCBAD: size limits and helical repeat.Proc Natl Acad Sci U S A86476480
- 13. Ptashne M, Gann A (2002) Genes & signals. Cold Spring Harbor, , N.Y.: Cold Spring Harbor Laboratory Press. M. PtashneA. Gann2002Genes & signalsCold Spring Harbor, , N.Y.Cold Spring Harbor Laboratory Pressxvi, 192 p
- 14. Bloomfield VA, Crothers DM, Tinoco I (2000) Nucleic Acids: Structures, Properties, and Functions. Sausalito, CA: University Science Books. VA BloomfieldDM CrothersI. Tinoco2000Nucleic Acids: Structures, Properties, and FunctionsSausalito, CAUniversity Science Books672 p
- 15. Yan J, Marko JF (2004) Localized single-stranded bubble mechanism for cyclization of short double helix DNA. Phys Rev Lett 93: 108108.J. YanJF Marko2004Localized single-stranded bubble mechanism for cyclization of short double helix DNA.Phys Rev Lett93108108
- 16. Olson WK, Swigon D, Coleman BD (2004) Implications of the dependence of the elastic properties of DNA on nucleotide sequence. Philosophical Transactions of the Royal Society of London Series a-Mathematical Physical and Engineering Sciences 362: 1403–1422.WK OlsonD. SwigonBD Coleman2004Implications of the dependence of the elastic properties of DNA on nucleotide sequence.Philosophical Transactions of the Royal Society of London Series a-Mathematical Physical and Engineering Sciences36214031422
- 17. Villa E, Balaeff A, Schulten K (2005) Structural dynamics of the lac repressor-DNA complex revealed by a multiscale simulation. Proc Natl Acad Sci U S A 102: 6783–6788.E. VillaA. BalaeffK. Schulten2005Structural dynamics of the lac repressor-DNA complex revealed by a multiscale simulation.Proc Natl Acad Sci U S A10267836788
- 18. Halford SE, Gowers DM, Sessions RB (2000) Two are better than one. Nature Structural Biology 7: 705–707.SE HalfordDM GowersRB Sessions2000Two are better than one.Nature Structural Biology7705707
- 19. Semsey S, Virnik K, Adhya S (2005) A gamut of loops: meandering DNA. Trends Biochem Sci 30: 334–341.S. SemseyK. VirnikS. Adhya2005A gamut of loops: meandering DNA.Trends Biochem Sci30334341
- 20. Zhang Y, McEwen AE, Crothers DM, Levene SD (2006) Statistical-mechanical theory of DNA looping. Biophys J 90: 1903–1912.Y. ZhangAE McEwenDM CrothersSD Levene2006Statistical-mechanical theory of DNA looping.Biophys J9019031912
- 21. Watson MA, Gowers DM, Halford SE (2000) Alternative geometries of DNA looping: an analysis using the SfiI endonuclease. J Mol Biol 298: 461–475.MA WatsonDM GowersSE Halford2000Alternative geometries of DNA looping: an analysis using the SfiI endonuclease.J Mol Biol298461475
- 22. Becker NA, Kahn JD, Maher LJ (2005) Bacterial repression loops require enhanced DNA flexibility. Journal of Molecular Biology 349: 716–730.NA BeckerJD KahnLJ Maher2005Bacterial repression loops require enhanced DNA flexibility.Journal of Molecular Biology349716730
- 23. Swigon D, Coleman BD, Olson WK (2006) Modeling the Lac repressor-operator assembly: The influence of DNA looping on Lac repressor conformation. Proc Natl Acad Sci U S A 103: 9879–9884.D. SwigonBD ColemanWK Olson2006Modeling the Lac repressor-operator assembly: The influence of DNA looping on Lac repressor conformation.Proc Natl Acad Sci U S A10398799884
- 24. Segall AM, Goodman SD, Nash HA (1994) Architectural elements in nucleoprotein complexes: interchangeability of specific and non-specific DNA binding proteins. Embo J 13: 4536–4548.AM SegallSD GoodmanHA Nash1994Architectural elements in nucleoprotein complexes: interchangeability of specific and non-specific DNA binding proteins.Embo J1345364548
- 25. Hill TL (1960) An introduction to statistical thermodynamics. Reading, , Mass.: Addison-Wesley Pub. Co. TL Hill1960An introduction to statistical thermodynamicsReading, , Mass.Addison-Wesley Pub. Co508 p
- 26. Lewis M, Chang G, Horton NC, Kercher MA, Pace HC, et al. (1996) Crystal structure of the lactose operon repressor and its complexes with DNA and inducer. Science 271: 1247–1254.M. LewisG. ChangNC HortonMA KercherHC Pace1996Crystal structure of the lactose operon repressor and its complexes with DNA and inducer.Science27112471254
- 27. Friedman AM, Fischmann TO, Steitz TA (1995) Crystal structure of lac repressor core tetramer and its implications for DNA looping. Science 268: 1721–1727.AM FriedmanTO FischmannTA Steitz1995Crystal structure of lac repressor core tetramer and its implications for DNA looping.Science26817211727
- 28. Ruben GC, Roos TB (1997) Conformation of Lac repressor tetramer in solution, bound and unbound to operator DNA. Microsc Res Tech 36: 400–416.GC RubenTB Roos1997Conformation of Lac repressor tetramer in solution, bound and unbound to operator DNA.Microsc Res Tech36400416
- 29. Edelman LM, Cheong R, Kahn JD (2003) Fluorescence resonance energy transfer over approximately 130 basepairs in hyperstable lac repressor-DNA loops. Biophys J 84: 1131–1145.LM EdelmanR. CheongJD Kahn2003Fluorescence resonance energy transfer over approximately 130 basepairs in hyperstable lac repressor-DNA loops.Biophys J8411311145
- 30. Grosschedl R (1995) Higher-Order Nucleoprotein Complexes in Transcription - Analogies with Site-Specific Recombination. Current Opinion in Cell Biology 7: 362–370.R. Grosschedl1995Higher-Order Nucleoprotein Complexes in Transcription - Analogies with Site-Specific Recombination.Current Opinion in Cell Biology7362370
- 31. Werner MH, Burley SK (1997) Architectural transcription factors: proteins that remodel DNA. Cell 88: 733–736.MH WernerSK Burley1997Architectural transcription factors: proteins that remodel DNA.Cell88733736
- 32. Thomas JO, Travers AA (2001) HMG1 and 2, and related ‘architectural’ DNA-binding proteins. Trends Biochem Sci 26: 167–174.JO ThomasAA Travers2001HMG1 and 2, and related ‘architectural’ DNA-binding proteins.Trends Biochem Sci26167174
- 33. Gartenberg MR, Crothers DM (1988) DNA-Sequence Determinants of Cap-Induced Bending and Protein-Binding Affinity. Nature 333: 824–829.MR GartenbergDM Crothers1988DNA-Sequence Determinants of Cap-Induced Bending and Protein-Binding Affinity.Nature333824829
- 34. Schultz SC, Shields GC, Steitz TA (1991) Crystal-Structure of a Cap-DNA Complex - the DNA Is Bent by 90-Degrees. Science 253: 1001–1007.SC SchultzGC ShieldsTA Steitz1991Crystal-Structure of a Cap-DNA Complex - the DNA Is Bent by 90-Degrees.Science25310011007
- 35. Rice PA, Yang S, Mizuuchi K, Nash HA (1996) Crystal structure of an IHF-DNA complex: a protein-induced DNA U-turn. Cell 87: 1295–1306.PA RiceS. YangK. MizuuchiHA Nash1996Crystal structure of an IHF-DNA complex: a protein-induced DNA U-turn.Cell8712951306
- 36. Hudson JM, Fried MG (1990) Cooperative Interactions between the Catabolite Gene Activator Protein and the Lac Repressor at the Lactose Promoter. Journal of Molecular Biology 214: 381–396.JM HudsonMG Fried1990Cooperative Interactions between the Catabolite Gene Activator Protein and the Lac Repressor at the Lactose Promoter.Journal of Molecular Biology214381396
- 37. Balaeff A, Mahadevan L, Schulten K (2004) Structural basis for cooperative DNA binding by CAP and lac repressor. Structure 12: 123–132.A. BalaeffL. MahadevanK. Schulten2004Structural basis for cooperative DNA binding by CAP and lac repressor.Structure12123132
- 38. Geanacopoulos M, Vasmatzis G, Zhurkin VB, Adhya S (2001) Gal repressosome contains an antiparallel DNA loop. Nat Struct Biol 8: 432–436.M. GeanacopoulosG. VasmatzisVB ZhurkinS. Adhya2001Gal repressosome contains an antiparallel DNA loop.Nat Struct Biol8432436
- 39. Bates AD, Maxwell A (2005) DNA Topology. Oxford, UK: Oxford University Press. AD BatesA. Maxwell2005DNA TopologyOxford, UKOxford University Press
- 40. Cloutier TE, Widom J (2005) DNA twisting flexibility and the formation of sharply looped protein-DNA complexes. Proceedings of the National Academy of Sciences of the United States of America 102: 3645–3650.TE CloutierJ. Widom2005DNA twisting flexibility and the formation of sharply looped protein-DNA complexes.Proceedings of the National Academy of Sciences of the United States of America10236453650
- 41. Saiz L, Vilar JMG (2006) Stochastic dynamics of macromolecular-assembly networks. Mol Syst Biol 2: 2006.0024.L. SaizJMG Vilar2006Stochastic dynamics of macromolecular-assembly networks.Mol Syst Biol22006.0024
- 42. Vilar JMG, Saiz L (2006) Multiprotein DNA looping. Phys Rev Lett 96: 238103.JMG VilarL. Saiz2006Multiprotein DNA looping.Phys Rev Lett96238103