Unwinding of forked DNA structures by UvrD

Many studies have demonstrated the need for processing of blocked replication forks to underpin genome duplication. UvrD helicase in Escherichia coli has been implicated in the processing of damaged replication forks, or the recombination intermediates formed from damaged forks. Here we show that UvrD can unwind forked DNA structures, in part due to the ability of UvrD to initiate unwinding from discontinuities within the phosphodiester backbone of DNA. UvrD does therefore have the capacity to target DNA intermediates of replication and recombination. Such an activity resulted in unwinding of what would be the parental duplex DNA ahead of either a stalled replication fork or a D-loop formed by recombination. However, UvrD had a substrate preference for fork structures having a nascent lagging strand at the branch point but no leading strand. Furthermore, at such structures the polarity of UvrD altered so that unwinding of the lagging strand predominated. This reaction is reminiscent of the PriC-Rep pathway of replication restart, suggesting that UvrD and Rep may have at least partially redundant functions.     

Dynamic Regulatory Interactions of Rad51, Rad52, and Replication Protein-A in Recombination Intermediates

Rad51, Rad52, and replication protein-A (RPA) play crucial roles in the repair of DNA double-strand breaks in Saccharomyces cerevisiae. Rad51 mediates DNA strand exchange, a key reaction in DNA recombination. Rad52 recruits Rad51 into single-stranded DNAs (ssDNAs) that are saturated with RPA. Rad52 also promotes annealing of ssDNA strands that are complexed with RPA. Specific protein-protein interactions are involved in these reactions. Here we report new biochemical characteristics of these protein interactions. First, Rad52-RPA interaction requires multiple molecules of RPA to be associated with ssDNA, suggesting that multiple contacts between the Rad52 ring and RPA-ssDNA filament are needed for stable binding. Second, RPA-t11, which is a recombination-deficient mutant of RPA, displays a defect in interacting with Rad52 in the presence of salt above 50 mM, explaining the defect in Rad52-mediated ssDNA annealing in the presence of this mutation. Third, ssDNA annealing promoted by Rad52 is preceded by aggregation of multiple RPA-ssDNA complexes with Rad52, and Rad51 inhibits this aggregation. These results suggest a regulatory role for Rad51 that suppresses ssDNA annealing and facilitates DNA strand invasion. Finally, the Rad51-double-stranded DNA complex disrupts Rad52-RPA interaction in ssDNA and titrates Rad52 from RPA. This suggests an additional regulatory role for Rad51 following DNA strand invasion, where Rad51-double-stranded DNA may inhibit illegitimate second-end capture to ensure the error-free repair of a DNA double-strand break.     

Kinetic control of Mg2+-dependent melting of duplex DNA ends by Escherichia coli RecBC

Escherichia coli RecBCD is a highly processive DNA helicase involved in double-strand break repair and recombination that possesses two helicase/translocase subunits with opposite translocation directionality (RecB (3′ to 5′) and RecD (5′ to 3′)). RecBCD has been shown to melt out ∼ 5-6 bp upon binding to a blunt-ended duplex DNA in a Mg²⁺-dependent, but ATP-independent reaction. Here, we examine the binding of E. coli RecBC helicase (minus RecD), also a processive helicase, to duplex DNA ends in the presence and in the absence of Mg²⁺ in order to determine if RecBC can also melt a duplex DNA end in the absence of ATP. Equilibrium binding of RecBC to DNA substrates with ends possessing pre-formed 3′ and/or 5′ single-stranded (ss)-(dT)_n flanking regions (tails) (n ranging from zero to 20 nt) was examined by competition with a fluorescently labeled reference DNA and by isothermal titration calorimetry. The presence of Mg²⁺ enhances the affinity of RecBC for DNA ends possessing 3′ or 5′-(dT)_n ssDNA tails with n < 6 nt, with the relative enhancement decreasing as n increases from zero to six nt. No effect of Mg²⁺ was observed for either the binding constant or the enthalpy of binding (ΔH_obs) for RecBC binding to DNA with ssDNA tail lengths, n ≥ 6 nucleotides. Upon RecBC binding to a blunt duplex DNA end in the presence of Mg²⁺, at least 4 bp at the duplex end become accessible to KMnO₄ attack, consistent with melting of the duplex end. Since Mg²⁺ has no effect on the affinity or binding enthalpy of RecBC for a DNA end that is fully pre-melted, this suggests that the role of Mg²⁺ is to overcome a kinetic barrier to melting of the DNA by RecBC and presumably also by RecBCD. These data also provide an accurate estimate (ΔH_obs = 8 ± 1 kcal/mol) for the average enthalpy change associated with the melting of a DNA base-pair by RecBC.     

Inhibition of RecBCD enzyme by antineoplastic DNA alkylating agents

To understand how bulky adducts might perturb DNA helicase function, three distinct DNA-binding agents were used to determine the effects of DNA alkylation on a DNA helicase. Adozelesin, ecteinascidin 743 (Et743) and hedamycin each possess unique structures and sequence selectivity. They bind to double-stranded DNA and alkylate one strand of the duplex in cis, adding adducts that alter the structure of DNA significantly. The results show that Et743 was the most potent inhibitor of DNA unwinding, followed by adozelesin and hedamycin. Et743 significantly inhibited unwinding, enhanced degradation of DNA, and completely eliminated the ability of the translocating RecBCD enzyme to recognize and respond to the recombination hotspot chi. Unwinding of adozelesin-modified DNA was accompanied by the appearance of unwinding intermediates, consistent with enzyme entrapment or stalling. Further, adozelesin also induced "apparent" chi fragment formation. The combination of enzyme sequestering and pseudo-chi modification of RecBCD, results in biphasic time-courses of DNA unwinding. Hedamycin also reduced RecBCD activity, albeit at increased concentrations of drug relative to either adozelesin or Et743. Remarkably, the hedamycin modification resulted in constitutive activation of the bottom-strand nuclease activity of the enzyme, while leaving the ability of the translocating enzyme to recognize and respond to chi largely intact. Finally, the results show that DNA alkylation does not significantly perturb the allosteric interaction that activates the enzyme for ATP hydrolysis, as the efficiency of ATP utilization for DNA unwinding is affected only marginally. These results taken together present a unique response of RecBCD enzyme to bulky DNA adducts. We correlate these effects with the recently determined crystal structure of the RecBCD holoenzyme bound to DNA.     

The Escherichia coli PriA Helicase-Double-Stranded DNA Complex: Location of the Strong DNA-Binding Subsite on the Helicase Domain of the Protein and the Affinity Control by the Two Nucleotide-Binding Sites of the Enzyme

The Escherichia coli PriA helicase complex with the double-stranded DNA (dsDNA), the location of the strong DNA-binding subsite, and the effect of the nucleotide cofactors, bound to the strong and weak nucleotide-binding site of the enzyme on the dsDNA affinity, have been analyzed using the fluorescence titration, analytical ultracentrifugation, and photo-cross-linking techniques. The total site size of the PriA-dsDNA complex is only 5 ± 1 bp, that is, dramatically lower than 20 ± 3 nucleotides occluded in the enzyme-single-stranded DNA (ssDNA) complex. The helicase associates with the dsDNA using its strong ssDNA-binding subsite in an orientation very different from the complex with the ssDNA. The strong DNA-binding subsite of the enzyme is located on the helicase domain of the PriA protein. The dsDNA intrinsic affinity is considerably higher than the ssDNA affinity and the binding process is accompanied by a significant positive cooperativity. Association of cofactors with strong and weak nucleotide-binding sites of the protein profoundly affects the intrinsic affinity and the cooperativity, without affecting the stoichiometry. ATP analog binding to either site diminishes the intrinsic affinity but preserves the cooperativity. ADP binding to the strong site leads to a dramatic increase of the cooperativity and only slightly affects the affinity, while saturation of both sites with ADP strongly increases the affinity and eliminates the cooperativity. Thus, the coordinated action of both nucleotide-binding sites on the PriA-dsDNA interactions depends on the structure of the phosphate group. The significance of these results for the enzyme activities in recognizing primosome assembly sites or the ssDNA gaps is discussed.     

Visualizing the Disassembly of S. cerevisiae Rad51 Nucleoprotein Filaments

Rad51 is the core component of the eukaryotic homologous recombination machinery and assembles into elongated nucleoprotein filaments on DNA. We have used total internal reflection fluorescence microscopy and a DNA curtain assay to investigate the dynamics of individual Saccharomyces cerevisiae Rad51 nucleoprotein filaments. For these experiments the DNA molecules were end-labeled with single fluorescent semiconducting nanocrystals. The assembly and disassembly of the Rad51 nucleoprotein filaments were visualized by tracking the location of the labeled DNA end in real time. Using this approach, we have analyzed yeast Rad51 under a variety of different reaction conditions to assess parameters that impact the stability of the nucleoprotein filament. We show that Rad51 readily dissociates from DNA in the presence of ADP or in the absence of nucleotide cofactor, but that free ATP in solution confers a fivefold increase in the stability of the nucleoprotein filaments. We also probe how protein dissociation is coupled to ATP binding and hydrolysis by examining the effects of ATP concentration, and by the use of the nonhydrolyzable ATP analogue adenosine 5'-(beta, gamma-imido) triphosphate and ATPase active-site mutants. Finally, we demonstrate that the Rad51 gain-of-function mutant I345T dissociates from DNA with kinetics nearly identical to that of wild-type Rad51, but assembles 30% more rapidly. Together, these results provide a framework for studying the biochemical behaviors of S. cerevisiae Rad51 nucleoprotein filaments at the single-molecule level.     

CeBRC-2 stimulates D-loop formation by RAD-51 and promotes DNA single-strand annealing

The BRCA2 tumour suppressor regulates the RAD-51 recombinase during double-strand break (DSB) repair by homologous recombination (HR) but how BRCA2 executes its functions is not well understood. We previously described a functional homologue of BRCA2 in Caenorhabditis elegans (CeBRC-2) that binds preferentially to single-stranded DNA via an OB-fold domain and associates directly with RAD-51 via a single BRC domain. Consistent with a direct role in HR, Cebrc-2 mutants are defective for repair of meiotic and radiation-induced DSBs due to an inability to regulate RAD-51. Here, we explore the function of CeBRC-2 in HR processes using purified proteins. We show that CeBRC-2 stimulates RAD-51-mediated D-loop formation and reduces the rate of ATP hydrolysis catalysed by RAD-51. These functions of CeBRC-2 are dependent upon direct association with RAD-51 via its BRC motif and on its DNA-binding activity, as point mutations in the BRC domain that abolish RAD-51 binding or the BRC domain of CeBRC-2 alone, lacking the DNA-binding domain, fail to stimulate RAD-51-mediated D-loop formation and do not reduce the rate of ATP hydrolysis by RAD-51. Phenotypic comparison of Cebrc-2 and rad-51 mutants also revealed a role for CeBRC-2 in an error-prone DSB repair pathway independent of rad-51 and non-homologous end joining, raising the possibility that CeBRC-2 may have replaced the role of vertebrate Rad52 in DNA single-strand annealing (SSA), which is missing from C. elegans. Indeed, we show here that CeBRC-2 mediates SSA of RPA-oligonucleotide complexes similar to Rad52. These results reveal RAD-51-dependent and -independent functions of CeBRC-2 that provide an explanation for the difference in DNA repair defects observed in Cebrc-2 and rad-51 mutants, and define mechanistic roles for CeBRC-2 in HR and in the SSA pathway for DSB repair.     

The 1.8-Å Crystal Structure of the N-Terminal Domain of an Archaeal MCM as a Right-Handed Filament

Mini-chromosome maintenance (MCM) proteins are the replicative helicase necessary for DNA replication in both eukarya and archaea. Most of archaea only have one MCM gene. Here, we report a 1.8-Å crystal structure of the N-terminal MCM from the archaeon Thermoplasma acidophilum (tapMCM). In the structure, the MCM N-terminus forms a right-handed filament that contains six subunits in each turn, with a diameter of 25 Å of the central channel opening. The inner surface is highly positively charged, indicating DNA binding. This filament structure with six subunits per turn may also suggests a potential role for an open-ring structure for hexameric MCM and dynamic conformational changes in initiation and elongation stages of DNA replication.     

Sin Resolvase Catalytic Activity and Oligomerization State are Tightly Coupled

Serine recombinases promote specific DNA rearrangements by a cut-and-paste mechanism that involves cleavage of all four DNA strands at two sites recognized by the enzyme. Dissecting the order and timing of these cleavage events and the steps leading up to them is difficult because the cleavage reaction is readily reversible. Here, we describe assays using activated Sin mutants and a DNA substrate with a 3′-bridging phosphorothiolate modification that renders Sin-mediated DNA cleavage irreversible. We find that activating Sin mutations promote DNA cleavage rather than simply stabilize the cleavage product. Cleavage events at the scissile phosphates on complementary strands of the duplex are tightly coupled, and the overall DNA cleavage rate is strongly dependent on Sin concentration. When combined with analytical ultracentrifugation data, these results suggest that Sin catalytic activity and oligomerization state are tightly linked, and that activating mutations promote formation of a cleavage-competent oligomeric state that is normally formed only transiently within the full synaptic complex.     

Response of the Bacteriophage T4 Replisome to Noncoding Lesions and Regression of a Stalled Replication Fork

DNA is constantly damaged by endogenous and exogenous agents. The resulting DNA lesions have the potential to halt the progression of the replisome, possibly leading to replication fork collapse. Here, we examine the effect of a noncoding DNA lesion in either leading strand template or lagging strand template on the bacteriophage T4 replisome. A damaged base in the lagging strand template does not affect the progression of the replication fork. Instead, the stalled lagging strand polymerase recycles from the lesion and initiates the synthesis of a new Okazaki fragment upstream of the damaged base. In contrast, when the replisome encounters a blocking lesion in the leading strand template, the replication fork only travels approximately 1 kb beyond the point of the DNA lesion before complete replication fork collapse. The primosome and the lagging strand polymerase remain active during this period, and an Okazaki fragment is synthesized beyond the point of the leading strand lesion. There is no evidence for a new priming event on the leading strand template. Instead, the DNA structure that is produced by the stalled replication fork is a substrate for the DNA repair helicase UvsW. UvsW catalyzes the regression of a stalled replication fork into a “chicken-foot” structure that has been postulated to be an intermediate in an error-free lesion bypass pathway.     

Rad54 oligomers translocate and cross-bridge double-stranded DNA to stimulate synapsis

Rad54 is a key component of the eukaryotic recombination machinery. Its presence in DNA strand-exchange reactions in vitro results in a significant stimulation of the overall reaction rate. Using untagged Rad54, we show that this stimulation can be attributed to enhancement of the formation of a key reaction intermediate known as DNA networks. Using a novel, single DNA molecule, dual-optical tweezers approach we show how Rad54 stimulates DNA network formation. We discovered that Rad54 oligomers possess a unique ability to cross-bridge or bind double-stranded DNA molecules positioned in close proximity. Further, Rad54 oligomers rapidly translocate double-stranded DNA while simultaneously inducing topological loops in the DNA at the locus of the oligomer. The combination of the cross-bridging and double-stranded DNA translocation activities of Rad54 stimulates the formation of DNA networks, leading to rapid and efficient DNA strand exchange by Rad51.     

The N-Terminal Region of the RecU Holliday Junction Resolvase Is Essential for Homologous Recombination

The RecU Holliday junction (HJ)-resolving enzyme is highly conserved in the Firmicutes phylum of bacteria. In Bacillus subtilis, the recU gene has two putative initiation codons, at positions 1 and 33. In rec⁺ cells, only the full-length RecU polypeptide (206 residues, 23.9 kDa) was detected even after different stress treatments. To address the relevance of the flexible N-terminus, we constructed mutant variants. Experiments in vivo revealed that recUΔ1-32 (which initiates at Met33 and encodes RecUΔ1-32) and recU31 (the conserved Arg31 residue was substituted with alanine to give RecUR31A) are genuine RecU mutants, rendering cells impaired in DNA repair and chromosomal segregation. RecU has three activities: It (i) cleaves HJs, (ii) anneals complementary strands and (iii) modulates RecA activities. RecUR31A binds and cleaves HJ DNA in vitro as efficiently as wild-type RecU, but RuvB·ATPγS·Mg²⁺ fails to stimulate the RecUR31A cleavage reaction. In contrast, RecUΔ1-32 forms unstable complexes with DNA and fails to cleave HJs. RecU and its variants are capable of promoting DNA strand annealing and exert a negative effect on deoxy-ATP-dependent RecA-mediated DNA strand exchange. This study shows that the flexible N-terminus of RecU is essential for protein activity.     

Helical Filaments of Human Dmc1 Protein on Single-Stranded DNA: A Cautionary Tale

Proteins in the RecA/Rad51/RadA family form nucleoprotein filaments on DNA that catalyze a strand exchange reaction as part of homologous genetic recombination. Because of the centrality of this system to many aspects of DNA repair, the generation of genetic diversity, and cancer when this system fails or is not properly regulated, these filaments have been the object of many biochemical and biophysical studies. A recent paper has argued that the human Dmc1 protein, a meiotic homolog of bacterial RecA and human Rad51, forms filaments on single-stranded DNA with ∼ 9 subunits per turn in contrast to the filaments formed on double-stranded DNA with ∼ 6.4 subunits per turn and that the stoichiometry of DNA binding is different between these two filaments. We show using scanning transmission electron microscopy that the Dmc1 filament formed on single-stranded DNA has a mass per unit length expected from ∼ 6.5 subunits per turn. More generally, we show how ambiguities in helical symmetry determination can generate incorrect solutions and why one sometimes must use other techniques, such as biochemistry, metal shadowing, or scanning transmission electron microscopy, to resolve these ambiguities. While three-dimensional reconstruction of helical filaments from EM images is a powerful tool, the intrinsic ambiguities that may be present with limited resolution are not sufficiently appreciated.     

Influence of single-stranded DNA-binding protein on recA induction in Escherichia coli

Two ssb mutants of Escherichia coli, whic carry a lesion in the single-strand DNA-binding protein (SSB), are sensitive to UV-irradiation. We have investigated the influence of SSB on the “SOS” repair pathway by examining the levels of recA protein synthesis. These strains fail to induced normal levels of recA protein after treatment with nalidixic acid or ultraviolet light. The level of recA protein synthesis in wild-type cells is about three times greater than ssb cells. This deficiency in ssb mutants occurs in all strains and at all temperatures tested (30–41.5°). In contrast, the ssb-1 mutant has no effect on temperature-induced recA induction in a recA441 (tif-1) strain. Cells carrying ssb⁺ plasmids and overproducing normal DNA-binding protein surprisingly are moderated UV-sensitive and have reduced levels of recA protein synthesis. Together these results establish that single-strand DNA-binding protein is involved in the induction of recA, and accounts, at least in part, for the UV sensivitiy of ssb mutant. Three possible mechanisms to explain the role of SSB are discussed.     

The Histone Chaperones ASF1 and CAF-1 Promote MMS22L-TONSL-Mediated Rad51 Loading onto ssDNA during Homologous Recombination in Human Cells

1. Download high-res image (217KB)
2. Download full-size image

     

The ATPase domain of the large terminase protein, gp17, from bacteriophage T4 binds DNA: implications to the DNA packaging mechanism

Translocation of double-stranded DNA into a preformed capsid by tailed bacteriophages is driven by powerful motors assembled at the special portal vertex. The motor is thought to drive processive cycles of DNA binding, movement, and release to package the viral genome. In phage T4, there is evidence that the large terminase protein, gene product 17 (gp17), assembles into a multisubunit motor and translocates DNA by an inchworm mechanism. gp17 consists of two domains; an N-terminal ATPase domain (amino acids 1-360) that powers translocation of DNA, and a C-terminal nuclease domain (amino acids 361-610) that cuts concatemeric DNA to generate a headful-size viral genome. While the functional motifs of ATPase and nuclease have been well defined and the ATPase atomic structure has been solved, the DNA binding motif(s) responsible for viral DNA recognition, cutting, and translocation are unknown. Here we report the first evidence for the presence of a double-stranded DNA binding activity in the gp17 ATPase domain. Binding to DNA is sensitive to Mg²⁺ and salt, but not the type of DNA used. DNA fragments as short as 20 bp can bind to the ATPase but preferential binding was observed to DNA greater than 1 kb. A high molecular weight ATPase-DNA complex was isolated by gel filtration, suggesting oligomerization of ATPase following DNA interaction. DNA binding was not observed with the full-length gp17, or the C-terminal nuclease domain. The small terminase protein, gp16, inhibited DNA binding, which was further accentuated by ATP. The presence of a DNA binding site in the ATPase domain and its binding properties implicate a role in the DNA packaging mechanism.     

Single-molecule atomic force spectroscopy reveals that DnaD forms scaffolds and enhances duplex melting

The Bacillus subtilis DnaD is an essential DNA-binding protein implicated in replication and DNA remodeling. Using single-molecule atomic force spectroscopy, we have studied the interaction of DnaD and its domains with DNA. Our data reveal that binding of DnaD to immobilized single molecules of duplex DNA causes a marked reduction in the ‘end-to-end’ distance of the DNA in a concentration-dependent manner, consistent with previously reported DnaD-induced looping by scaffold formation. Native DnaD enhances partial melting of the DNA strands. The C-terminal domain (Cd) of DnaD binds to DNA and enhances partial duplex melting but does not cause DNA looping. The Cd-mediated melting is not as efficient as that caused by native DnaD. The N-terminal domain (Nd) does not affect significantly the DNA. A mixture of Nd and Cd fails to recreate the DNA looping effect of native DnaD but produces exactly the same effects as Cd on its own, consistent with the previously reported failure of the separated domains to form DNA-interacting scaffolds.     

Conformational Adaptability of Redβ during DNA Annealing and Implications for Its Structural Relationship with Rad52

Single-strand annealing proteins, such as Redβ from λ phage or eukaryotic Rad52, play roles in homologous recombination. Here, we use atomic force microscopy to examine Redβ quaternary structure and Redβ-DNA complexes. In the absence of DNA, Redβ forms a shallow right-handed helix. The presence of single-stranded DNA (ssDNA) disrupts this structure. Upon addition of a second complementary ssDNA, annealing generates a left-handed helix that incorporates 14 Redβ monomers per helical turn, with each Redβ monomer annealing ≈ 11 bp of DNA. The smallest stable annealing intermediate requires 20 bp DNA and two Redβ monomers. Hence, we propose that Redβ promotes base pairing by first increasing the number of transient interactions between ssDNAs. Then, annealing is promoted by the binding of a second Redβ monomer, which nucleates the formation of a stable annealing intermediate. Using threading, we identify sequence similarities between the RecT/Redβ and the Rad52 families, which strengthens previous suggestions, based on similarities of their quaternary structures, that they share a common mode of action. Hence, our findings have implications for a common mechanism of DNA annealing mediated by single-strand annealing proteins including Rad52.     

The Bipolar Filaments Formed by Herpes Simplex Virus Type 1 SSB/Recombination Protein (ICP8) Suggest a Mechanism for DNA Annealing

Herpes simplex virus type 1 encodes a multifunctional protein, ICP8, which serves both as a single-strand binding protein and as a recombinase, catalyzing reactions involved in replication and recombination of the viral genome. In the presence of divalent ions and at low temperature, previous electron microscopic studies showed that ICP8 will form long left-handed helical filaments. Here, electron microscopic image reconstruction reveals that the filaments are bipolar, with an asymmetric unit containing two subunits of ICP8 that constitute a symmetrical dimer. This organization of the filament has been confirmed using scanning transmission electron microscopy. The pitch of the filaments is ∼ 250 Å, with ∼ 6.2 dimers per turn. Docking of a crystal structure of ICP8 into the reconstructed filament shows that the C-terminal domain of ICP8, attached to the body of the subunit by a flexible linker containing ∼ 10 residues, is packed into a pocket in the body of a neighboring subunit in the crystal in a similar manner as in the filament. However, the interactions between the large N-terminal domains are quite different in the filament from that observed in the crystal. A previously proposed model for ICP8 binding single-stranded DNA (ssDNA), based upon the crystal structure, leads to a model for a continuous strand of ssDNA near the filament axis. The bipolar nature of the ICP8 filaments means that a second strand of ssDNA would be running through this filament in the opposite orientation, and this provides a potential mechanism for how ICP8 anneals complementary ssDNA into double-stranded DNA, where each strand runs in opposite directions.     

HMGB binding to DNA: single and double box motifs

High mobility group (HMG) proteins are nuclear proteins believed to significantly affect DNA interactions by altering nucleic acid flexibility. Group B (HMGB) proteins contain HMG box domains known to bind to the DNA minor groove without sequence specificity, slightly intercalating base pairs and inducing a strong bend in the DNA helical axis. A dual-beam optical tweezers system is used to extend double-stranded DNA (dsDNA) in the absence as well as presence of a single box derivative of human HMGB2 [HMGB2(box A)] and a double box derivative of rat HMGB1 [HMGB1(box A+box B)]. The single box domain is observed to reduce the persistence length of the double helix, generating sharp DNA bends with an average bending angle of 99 ± 9° and, at very high concentrations, stabilizing dsDNA against denaturation. The double box protein contains two consecutive HMG box domains joined by a flexible tether. This protein also reduces the DNA persistence length, induces an average bending angle of 77 ± 7°, and stabilizes dsDNA at significantly lower concentrations. These results suggest that single and double box proteins increase DNA flexibility and stability, albeit both effects are achieved at much lower protein concentrations for the double box. In addition, at low concentrations, the single box protein can alter DNA flexibility without stabilizing dsDNA, whereas stabilization at higher concentrations is likely achieved through a cooperative binding mode.