ATP conformations and ion binding modes in the active site of anthrax edema factor: A computational analysis

The Edema Factor (EF), one of the virulence factors of anthrax, is an adenylyl cyclase that promotes the overproduction of cyclic‐AMP (cAMP) from ATP, and therefore perturbs cell signaling. Crystallographic structures of EF bound to ATP analogs and reaction products, cyclic‐AMP, and Pyrophosphate (PPi), revealed different substrate conformations and catalytic‐cation binding modes, one or two cations being observed in the active site. To shed light into the biological significance of these crystallographic structures, the energetics, geometry, and dynamics of the active site are analyzed using molecular dynamics simulations. The ATP conformation observed in the one‐metal‐ion structure allows stronger interactions with the catalytic ion, and ATP is more restrained than in the structure containing two Mg²⁺ ions. Therefore, we propose that the conformation observed in the one‐ion crystal structure is a more probable starting point for the reaction. The simulations also suggest that a C3′‐endo sugar pucker facilitates nucleophilic attack. Additionally, the two‐cation binding mode restrains the mobility of the reaction products, and thus their tendency to dissociate. Proteins 2009. © 2009 Wiley‐Liss, Inc.     

Structure of the human multidrug resistance protein 1 nucleotide binding domain 1 bound to Mg2+/ATP reveals a non-productive catalytic site

Human multidrug resistance protein 1 (MRP1) is a membrane protein that belongs to the ATP-binding cassette (ABC) superfamily of transport proteins. MRP1 contributes to chemotherapy failure by exporting a wide range of anti-cancer drugs when over expressed in the plasma membrane of cells. Here, we report the first high-resolution crystal structure of human MRP1-NBD1. Drug efflux requires energy resulting from hydrolysis of ATP by nucleotide binding domains (NBDs). Contrary to the prokaryotic NBDs, the extremely low intrinsic ATPase activity of isolated MRP1-NBDs allowed us to obtain the structure of wild-type NBD1 in complex with Mg2+/ATP. The structure shows that MRP1-NBD1 adopts a canonical fold, but reveals an unexpected non-productive conformation of the catalytic site, providing an explanation for the low intrinsic ATPase activity of NBD1 and new hypotheses on the cooperativity of ATPase activity between NBD1 and NBD2 upon heterodimer formation.     

Binding of N-terminal fragments of anthrax edema factor (EFN) and lethal factor (LFN) to the protective antigen pore

Anthrax toxin consists of three different molecules: the binding component protective antigen (PA, 83 kDa), and the enzymatic components lethal factor (LF, 90 kDa) and edema factor (EF, 89 kDa). The 63 kDa C-terminal part of PA, PA₆₃, forms heptameric channels that insert in endosomal membranes at low pH, necessary to translocate EF and LF into the cytosol of target cells. In many studies, about 30 kDa N-terminal fragments of the enzymatic components EF (254 amino acids) and LF (268 amino acids) were used to study their interaction with PA₆₃-channels. Here, in experiments with artificial lipid bilayer membranes, EF_N and LF_N show block of PA₆₃-channels in a dose, voltage and ionic strength dependent way with high affinity. However, when compared to their full-length counterparts EF and LF, they exhibit considerably lower binding affinity. Decreasing ionic strength and, in the case of EF_N, increasing transmembrane voltage at the cis side of the membranes, resulted in a strong decrease of half saturation constants. Our results demonstrate similarities but also remarkable differences between the binding kinetics of both truncated and full-length effectors to the PA₆₃-channel.     

Identification of amino acid residues responsible for the alpha5 subunit binding selectivity of L-655,708, a benzodiazepine binding site ligand at the GABA(A) receptor

L-655,708 is a ligand for the benzodiazepine site of the gamma-aminobutyric acid type A (GABA(A)) receptor that exhibits a 100-fold higher affinity for alpha5-containing receptors compared with alpha1-containing receptors. Molecular biology approaches have been used to determine which residues in the alpha5 subunit are responsible for this selectivity. Two amino acids have been identified, alpha5Thr208 and alpha5Ile215, each of which individually confer approximately 10-fold binding selectivity for the ligand and which together account for the 100-fold higher affinity of this ligand at alpha5-containing receptors. L-655,708 is a partial inverse agonist at the GABA(A) receptor which exhibited no functional selectivity between alpha1- and alpha5-containing receptors and showed no change in efficacy at receptors containing alpha1 subunits where amino acids at both of the sites had been altered to their alpha5 counterparts (alpha1Ser205-Thr,Val212-Ile). In addition to determining the binding selectivity of L-655,708, these amino acid residues also influence the binding affinities of a number of other benzodiazepine (BZ) site ligands. They are thus important elements of the BZ site of the GABA(A) receptor, and further delineate a region just N-terminal to the first transmembrane domain of the receptor alpha subunit that contributes to this binding site.     

Nuclear Overhauser effect studies on the conformations of Mg(alpha,beta-methylene)ATP bound to Escherichia coli methionyl-tRNA synthetase

Internuclear distances obtained from nuclear Overhauser effects were used in combination with a distance geometry algorithm to determine the conformation of Mg(alpha,beta-methylene)ATP bound to the Escherichia coli truncated methionyl-tRNA synthetase (delta MTS) both in the absence and presence of cognate and noncognate amino acids. Mg(alpha,beta-methylene)ATP, a nonhydrolyzable analog of ATP, was used to prevent hydrolysis of the nucleotide in the presence of either cognate or noncognate amino acids. Kinetic analysis showed that Mg(alpha,beta-methylene)ATP was a linear competitive inhibitor with respect to ATP in the ATP-pyrophosphate exchange reaction with a Ki = 1.2 mM. The pattern of internuclear Overhauser effects on Mg(alpha,beta-methylene)ATP bound to delta MTS was qualitatively consistent only with an anti glycosidic torsional angle, suggesting that the adenosine portion of the nucleotide is uniquely oriented in the binary enzyme-nucleotide complex. Nearly identical patterns of nuclear Overhauser effects were also observed in ternary complexes containing either cognate L-methionine or noncognate L-homocysteine amino acids. Distance geometry calculations permitted the range and conformational space of the allowed adenine-ribose glycosidic torsional angles in each of the complexes to be better defined and compared. Average adenine-ribose glycosidic torsional angles for enzyme-bound Mg(alpha,beta-methylene)ATP of -106 +/- 9 degrees, -99 +/- 11 degrees, and -97 +/- 11 degrees were determined for the delta MTS.Mg(alpha,beta-methylene)ATP, delta MTS.Mg(alpha,beta-methylene)ATP.L-methionine, and delta MTS.Mg(alpha,beta-methylene)ATP.L-homocysteine complexes, respectively. Comparison of the three enzyme-bound conformations showed that a single nucleotide structure having an adenine-ribose glycosidic torsional angle of -98 degrees with a 3'-endo to O4'-exo ribose sugar pucker was, within error, consistent with the experimental internuclear distances obtained in all three complexes. The nearly identical anti glycosidic torsional angles observed in all three complexes demonstrates that the conformation of the adenosine moiety of the enzyme-bound nucleotide is not sensitive to the presence or the nature of the amino acid bound at the aminoacyladenylate site. Therefore, conformational changes known to occur in the methionyl-tRNA synthetase upon ligand binding appear not to alter the bound conformation of the nucleotide. Information on the conformation and arrangement of substrates bound at the aminoacyladenylate site of delta MTS is necessary for understanding the molecular mechanisms involved in amino acid activation and discrimination.     

Crystal structure of the HIV-1 integrase core domain in complex with sucrose reveals details of an allosteric inhibitory binding site

HIV integrase (IN) is an essential enzyme in HIV replication and an important target for drug design. IN has been shown to interact with a number of cellular and viral proteins during the integration process. Disruption of these important interactions could provide a mechanism for allosteric inhibition of IN. We present the highest resolution crystal structure of the IN core domain to date. We also present a crystal structure of the IN core domain in complex with sucrose which is bound at the dimer interface in a region that has previously been reported to bind integrase inhibitors.

Structured summary

MINT-7713125: IN (uniprotkb:P04585) and IN (uniprotkb:P04585) bind (MI:0407) by X-ray crystallography (MI:0114)     

Nuclear overhauser effect studies of the conformations of Mg(alpha, beta-methylene)ATP bound to E. coli isoleucyl-tRNA synthetase

Internuclear distances obtained from transferred nuclear Overhauser effects were used in combination with distance geometry calculations to define the E. coli isoleucyl-tRNA synthetase bound conformation of Mg(alpha, beta-methylene)ATP both in the absence and in the presence of the cognate and noncognate amino acids L-isoleucine and L-valine, respectively. A single nucleotide structure having an anti adenine-ribose glycosidic torsional angle of -114 degrees was found to satisfy the experimental distance constraints. The nearly identical anti glycosidic torsional angles observed in all three complexes demonstrate that the conformation of the adenosine moiety of the enzyme-bound nucleotide is not sensitive to the presence or to the nature of the amino acid bound at the aminoacyladenylate site. In addition, the acceptable range of Mg(alpha, beta-methylene)ATP conformations bound to the E. coli isoleucyl-tRNA synthetase was found to be nearly identical to that previously determined for the E. coli methionyl-tRNA synthetase (Williams and Rosevear (1991) J. Biol. Chem. 266, 2089-2098). Thus, the predicted structural homology between the isoleucyl- and methionyl-tRNA synthetases, both members of the same class of synthetases on the basis of common consensus sequences, is further supported by consensus enzyme-bound nucleotide conformations.     

Structure of the Q67H mutant of R67 dihydrofolate reductase-NADP+ complex reveals a novel cofactor binding mode

Plasmid-encoded bacterial R67 dihydrofolate reductase (DHFR) is a NADPH-dependent enzyme unrelated to chromosomal DHFR in amino acid sequence and structure. R67 DHFR is insensitive to the bacterial drug trimethoprim in contrast to chromosomal DHFR. The crystal structure of Q67H mutant of R67 DHFR bound to NADP(+) has been determined at 1.15 angstroms resolution. The cofactor assumes an extended conformation with the nicotinamide ring bound near the center of the active site pore, the ribose and pyrophosphate group (PP(i)) extending toward the outer pore. The ribonicotinamide exhibits anti conformation as in chromosomal DHFR complexes. The relative orientation between the PP(i) and the nicotinamide ribose differs from that observed in chromosomal DHFR-NADP(+) complexes. The coenzyme displays symmetrical binding mode with several water-mediated hydrogen bonds with the protein besides ionic, stacking, and van der Waals interactions. The structure provides a molecular basis for the observed stoichiometry and cooperativity in ligand binding. The ternary model based on the present structure and the previous R67 DHFR-folate complex provides insight into the catalytic mechanism and indicates that the relative orientation of the reactants in plasmid DHFR is different from that seen in chromosomal DHFRs.     

A conformational change of N-iodoacetyl-N'-(5-sulfo-1-naphthyl)ethylenediamine-labeled sarcoplasmic reticulum Ca2+-ATPase upon ATP binding to the catalytic site

Sarcoplasmic reticulum vesicles were modified with a fluorescent thiol reagent, N-iodoacetyl-N'-(5-sulfo-1-naphthyl)ethylenediamine. One mol of readily reactive thiols per mol of the Ca2+-ATPase was labeled without a loss of the catalytic activity. The fluorescence of the label increased by 8% upon binding of Ca2+ to the high affinity sites of the enzyme. This fluorescence enhancement probably reflects a conformational change responsible for Ca2+-induced enzyme activation. Upon addition of ATP to the Ca2+-activated enzyme, the fluorescence decreased by 15%. This fluorescence drop and formation of the phosphoenzyme intermediate were determined under the same conditions with a stopped-flow apparatus and a rapid quenching system. The amplitude of the fluorescence drop thus determined was saturated with 3 microM ATP. This shows that the fluorescence drop was caused by ATP binding to the catalytic site. In contrast, the rate of the fluorescence drop was not saturated even with 50 microM ATP. The fluorescence drop coincided with phosphoenzyme formation at 0.5 or 3 microM ATP, but it became much faster than phosphoenzyme formation when the ATP concentration was raised to 100 microM. These results indicate that the ATP-induced fluorescence drop reflects a conformational change in the enzyme.ATP complex. The fluorescence drop was accompanied by a red spectrum shift, which suggests that the label was exposed to a more hydrophilic environment. The electrophoretic analysis of the tryptic digest of the labeled enzyme (10.9 kDa) showed that almost all of the label was located on the 5.2-kDa fragment which includes the carboxyl terminus and the putative ATP-binding domain. The sequencing of the two major labeled peptides, which were isolated from the thermolytic digest of the labeled enzyme, revealed that the labeled site in either of these peptides was Cys674. It seems likely that the label bound to this Cys674 could be involved in the observed fluorescence changes.     

IGFBP-3 and IGFBP-5 associate with the cell binding domain (CBD) of fibronectin

We have used Surface Plasmon Resonance (SPR) - based biosensor technology to investigate the interaction of the six high affinity insulin-like growth factor binding proteins (IGFBP 1-6) with the cell binding domain (CBD) of fibronectin. Using a biotinylated derivative of the ninth and tenth TypeIII domains of FN (^9-10FNIII), we show that IGFBP-3 and -5 bind to FN-CBD. We show that this binding is inhibited by IGF-I and that, for IGFBP-5, binding occurs through the C-terminal heparin binding domain of the protein. Using site-directed mutagenesis of ^9-10FNIII, we show both the “synergy” and RGD sites within these FN domains are required for maximum binding of both IGFBPs. We discuss the possible biological consequences of our results.     

Reactions of a fluorescent ATP analog, 2'-(5-dimethyl-aminonaphthalene-1-sulfonyl) amino-2'-deoxyATP, with E. coli F1-ATPase and its subunits: the roles of the high affinity binding site in the alpha subunit and the low affinity binding site in the beta subunit

We performed kinetic studies on the reactions of a fluorescent ATP analog, 2'-(5-dimethyl-aminonaphthalene-1-sulfonyl) amino-2'-deoxyATP (DNS-ATP), with E. coli F1-ATPase (EF1) and its subunits, to clarify the role of each subunit in the ATPase reaction. The following results were obtained. 1. One mol of EF1, which contains nonexchangeable 2 mol ATP and 0.5 mol ADP, binds 3 mol of DNS-ATP. The apparent dissociation constant, in the presence of Mg2+, was 0.23 microM. Upon binding, the fluorescence intensity of DNS-ATP at 520 nm increased exponentially with t1/2 of 35 s, and reached 3.5 times the original fluorescence level. Following the fluorescence increase, DNS-ATP was hydrolyzed, and the fluorescence intensity maintained its enhanced level. 2. The addition of an excess of ATP over the EF1-DNS-nucleotide complex, in the presence of Mg2+, decreased the fluorescence intensity rapidly, indicating the acceleration of DNS-nucleotide release from EF1. ADP and GTP also decreased the fluorescence intensity. 3. DCCD markedly inhibited the accelerating effect of ATP on DNS-nucleotide release from EF1 and the EF1-DNS-ATPase or -ATPase activity in a steady state. On the other hand, DCCD only slightly inhibited the fluorescence increase of DNS-ATP, due to its binding to EF1, and the rate of single cleavage of 1 mol of DNS-ATP per mol of alpha subunit of EF1. 4. In the presence of Mg2+, 0.65-0.82 mol of DNS-ATP binds to 1 mol of the isolated alpha subunit of EF1 with an apparent dissociation constant of 0.06-0.07 microM. Upon binding, the fluorescence intensity of DNS-ATP at 520 nm increased 1.55 fold very rapidly (t1/2 less than 1 s). No hydrolysis of DNS-ATP was observed upon the addition of the isolated alpha subunit. The fluorescence intensity of DNS-ATP was unaffected by the addition of the isolated beta subunit. DNS-ATP was also unhydrolyzed by the isolated beta subunit. 5. EF1-ATPase was reconstituted from alpha, beta, and gamma subunits in the presence of Mg2+ and ATP. The kinetic properties of the fluorescence change of DNS-ATP in the reaction with the reconstituted EF1-ATPase were quite similar to those of native EF1. Most of our findings are consistent with a simple mechanism that the high affinity catalytic site and low affinity regulatory site exist in the alpha subunit and beta subunit, respectively. However, the findings mentioned in (4) suggest that the binding of the alpha and beta subunit, which is mediated by the gamma subunit, induces conformational change(s) in the ATP binding site located probably in the alpha subunit, and that the conformational change(s) is essential to exert the full hydrolyzing activity.     

NMR chemical shift mapping of the binding site of a protein proteinase inhibitor: changes in the (1)H, (13)C and (15)N NMR chemical shifts of turkey ovomucoid third domain upon binding to bovine chymotrypsin A(alpha)

The substrate-like inhibition of serine proteinases by avian ovomucoid domains has provided an excellent model for protein inhibitor-proteinase interactions of the standard type. 1H,15N and 13C NMR studies have been undertaken on complexes formed between turkey ovomucoid third domain (OMTKY3)2 and chymotrypsin A(alpha) (Ctr) in order to characterize structural changes occurring in the Ctr binding site of OMTKY3. 15N and 13C were incorporated uniformly into OMTKY3, allowing backbone resonances to be assigned for OMTKY3 in both its free and complex states. Chemical shift perturbation mapping indicates that the two regions, K13-P22 and N33-A40, are the primary sites in OMTKY3 involved in Ctr binding, in full agreement with the 12 consensus proteinase-contact residues of OMTKY3 defined previously on the basis of X-ray crystallographic and mutational analysis. Smaller chemical shift perturbations in selected other regions may result from minor structural changes on binding. Through-bond 15N-13C correlations between P1-13C' and P1'-15N in two-dimensional H(N)CO and HN(CO) NMR spectra of selectively labeled OMTKY3 complexed with Ctr indicate that the scissile peptide bond between L18 and E19 of the inhibitor is intact in the complex. The chemical shifts of the reactive site peptide bond indicate that it is predominantly trigonal, although the data are not inconsistent with a slight perturbation of the hybridization of the peptide bond toward the first tetrahedral state along the reaction coordinate.     

Dependence of properties of polyclonal antibodies to (2'-5')-oligoadenylates on the method of hapten binding to an immunogen

To elucidate the antibody-(2'-5')oligoadenylate relation to the mode of the hapten-immunogen conjugation, a new (2'-5')oligoadenylic acid trimer derivative containing a 2'-terminal N6-(5-carboxypentyl)adenosine and its 125I-labeled immunogenic conjugate were synthesized. The immunization with this conjugate and with a conjugate based on the 2',3'-O-[1-(2-carboxyethyl)]ethylidene derivative of the (2'-5')triadenylic acid gave antisera with different affinities toward modified (2'-5')oligonucleotides. Epitopes involved in the (2'-5')oligomer-binding to different antisera were found.     

Inhibition of nicotinoprotein (NAD+-containing) alcohol dehydrogenase by trans-4-(N,N-dimethylamino)-cinnamaldehyde binding to the active site

Ethanol oxidation by nicotinoprotein alcohol dehydrogenase (np-ADH) from the bacterium Amycolatopsis methanolica is inhibited by trans-4-(N,N-dimethylamino)-cinnamaldehyde through direct binding to the catalytic zinc ion in a substrate-like geometry. This binding is accompanied by a characteristic red shift of the aldehyde absorbance from 398 nm to 467 nm. Np-ADH is structurally related to mammalian ADH class I, and a model of np-ADH shows how the cinnamaldehyde derivative can be accommodated in the active site of the nicotinoprotein, correlating the structural and enzymological data.     

Characterization of nucleotide binding sites on the membrane-bound chloroplast ATP synthase (coupling factor CF1)

Phosphorylation of ADP and nucleotide exchange by membrane-bound coupling factor CF₁ are very fast reactions in the light, so that a direct comparison of both reactions is difficult. By adding substrate ADP and phosphate to illuminated thylakoids together with the uncoupler FCCP, the phosphorylation time is limited and the amount of ATP formed can be reduced to less than 1 ATP per enzyme. Low concentrations of medium nucleotides during illumination increase the amount of ATP formed during uncoupling presumably by binding to the tight nucleotide binding site (further designated as ‘site A’) with an affinity of 1 to 7 μM for ADP and ATP. ATP formation itself shows half-saturation at about 30 μM. Loosely bound nucleotides are exchanged upon addition of nucleotides with uncoupler (Schumann, J. (1984) Biochim. Biophys. Acta 766, 334–342). Release depends binding of nucleotides to a second site. The affinity of this site for ADP (in the presence of phosphate) is about 30 μM. It is assumed that phosphorylation and induction of exchange both occur on the same site (site B). During ATP hydrolysis, an ATP molecule is bound to site A, while on another site, ATP is hydrolyzed rapidly. The affinity of ADP for the catalytic site (70 μM) is in the same range as the observed Michaelis constant of ADP during phosphorylation; it is assumed that site B is involved in ATP hydrolysis. Site A exhibits some catalytic activity; it might be that site A is involved in ATP formation in a dual-site mechanism. For ATP hydrolysis, however, direct determination of exchange rates showed that the exchange rate of ATP bound to site A is about 30-times lower than ATP hydrolysis under the same conditions.     

Cantharidin inhibits competitively heme‐Fe(III) binding to the FA1 site of human serum albumin

Cantharidin, a monoterpene isolated from the insect blister beetle, has long been used as a medicinal agent in the traditional Chinese medicine. Cantharidin inhibits a subgroup of serine/threonine phosphatases, thus inducing cell growth inhibition and cytotoxicity. Cantharidin has anticancer activity in vitro, since it is able of inducing p53‐dependent apoptosis and double‐strand breakage of DNA in cancer cells. Although the toxicity of cantharidin to the gastrointestinal and urinary tracts prevents its medical use, it is a promising lead compound for chemical modification to develop new anticancer therapeutics. In fact, cantharidin does not cause myelosuppression and displays anticancer activity against cells with a multidrug resistance phenotype. Here, the competitive inhibitory effect of cantharidin on heme‐Fe(III) binding to the fatty acid site 1 (FA1) of human serum albumin (HSA) is reported. Docking and molecular dynamics simulations support functional data indicating the preferential binding of cantharidin to the FA1 site of HSA. Present results may be relevant in vivo as HSA could transport cantharidin, which in turn could affect heme‐Fe(III) scavenging by HSA.     

Crystal structure of enoyl-acyl carrier protein reductase (FabK) from Streptococcus pneumoniae reveals the binding mode of an inhibitor

Enoyl-acyl carrier protein (ACP) reductases are critical for bacterial type II fatty acid biosynthesis and thus are attractive targets for developing novel antibiotics. We determined the crystal structure of enoyl-ACP reductase (FabK) from Streptococcus pneumoniae at 1.7 A resolution. There was one dimer per asymmetric unit. Each subunit formed a triose phosphate isomerase (TIM) barrel structure, and flavin mononucleotide (FMN) was bound as a cofactor in the active site. The overall structure was similar to the enoyl-ACP reductase (ER) of fungal fatty acid synthase and to 2-nitropropane dioxygenase (2-ND) from Pseudomonas aeruginosa, although there were some differences among these structures. We determined the crystal structure of FabK in complex with a phenylimidazole derivative inhibitor to envision the binding site interactions. The crystal structure reveals that the inhibitor binds to a hydrophobic pocket in the active site of FabK, and this is accompanied by induced-fit movements of two loop regions. The thiazole ring and part of the ureido moiety of the inhibitor are involved in a face-to-face pi-pi stacking interaction with the isoalloxazine ring of FMN. The side-chain conformation of the proposed catalytic residue, His144, changes upon complex formation. Lineweaver-Burk plots indicate that the inhibitor binds competitively with respect to NADH, and uncompetitively with respect to crotonoyl coenzyme A. We propose that the primary basis of the inhibitory activity is competition with NADH for binding to FabK, which is the first step of the two-step ping-pong catalytic mechanism.     

Putative alternative functions of human stefin B (cystatin B): binding to amyloid‐beta,membranes, and copper

We describe studies performed thus far on stefin B from the family of cystatins as a model protein for folding and amyloid fibril formation studies. We also briefly mention our studies on aggregation of some of the missense EPM1 mutants of stefin B in cells, which mimic additional pathological traits (gain in toxic function) in selected patients with EPM1 disease. We collected data on the reported interactors of stefin B and discuss several hypotheses of possible cytosolic alternative functions.     

DNA binding mode transitions of Escherichia coli HU(alphabeta): evidence for formation of a bent DNA--protein complex on intact, linear duplex DNA

Escherichia coli HU_αβ, a major nucleoid-associated protein, organizes chromosomal DNA and facilitates numerous DNA transactions. Using isothermal titration calorimetry, fluorescence resonance energy transfer and a series of DNA lengths (8 bp, 15 bp, 34 bp, 38 bp and 160 bp) we established that HU_αβ interacts with duplex DNA using three different nonspecific binding modes. Both the HU to DNA molar ratio ([HU]/[DNA]) and DNA length dictate the dominant HU binding mode. On sufficiently long DNA (≥ 34 bp), at low [HU]/[DNA], HU populates a noncooperative 34 bp binding mode with a binding constant of 2.1 ± 0.4 × 10⁶ M^− 1, and a binding enthalpy of + 7.7 ± 0.6 kcal/mol at 15 °C and 0.15 M Na⁺. With increasing [HU]/[DNA], HU bound in the noncooperative 34 bp mode progressively converts to two cooperative (ω∼20) modes with site sizes of 10 bp and 6 bp. These latter modes exhibit smaller binding constants (1.1 ± 0.2 × 10⁵ M^− 1 for the 10 bp mode, 3.5 ± 1.4 × 10⁴ M^− 1 for the 6 bp mode) and binding enthalpies (4.2 ± 0.3 kcal/mol for the 10 bp mode, − 1.6 ± 0.3 kcal/mol for the 6 bp mode). As DNA length increases to 34 bp or more at low [HU]/[DNA], the small modes are replaced by the 34 bp binding mode. Fluorescence resonance energy transfer data demonstrate that the 34 bp mode bends DNA by 143 ± 6° whereas the 6 bp and 10 bp modes do not. The model proposed in this study provides a novel quantitative and comprehensive framework for reconciling previous structural and solution studies of HU, including single molecule (force extension measurement), fluorescence, and electrophoretic gel mobility-shift assays. In particular, it explains how HU condenses or extends DNA depending on the relative concentrations of HU and DNA.