Thermodynamic basis for sequence-specific recognition of ssDNA by an autoantibody

Synthetic oligonucleotide antigens modified with locked nucleic acids detect disease specific antibodies

New techniques to detect and quantify antibodies to nucleic acids would provide a significant advance over current methods, which often lack specificity. We investigate the potential of novel antigens containing locked nucleic acids (LNAs) as targets for antibodies. Particularly, employing molecular dynamics we predict optimal nucleotide composition for targeting DNA-binding antibodies. As a proof of concept, we address a problem of detecting anti-DNA antibodies that are characteristic of systemic lupus erythematosus, a chronic autoimmune disease with multiple manifestations. We test the best oligonucleotide binders in surface plasmon resonance studies to analyze binding and kinetic aspects of interactions between antigens and target DNA. These DNA and LNA/DNA sequences showed improved binding in enzyme-linked immunosorbent assay using human samples of pediatric lupus patients. Our results suggest that the novel method is a promising tool to create antigens for research and point-of-care monitoring of anti-DNA antibodies.

"Clickable" LNA/DNA probes for fluorescence sensing of nucleic acids and autoimmune antibodies

Herein we describe fluorescent oligonucleotides prepared by click chemistry between novel alkyne-modified locked nucleic acid (LNA) strands and a series of fluorescent azides for homogeneous (all-in-solution) detection of nucleic acids and autoimmune antibodies.

Salt bridge exchange binding mechanism between streptavidin and its DNA aptamer--thermodynamics and spectroscopic evidences

Protein-nucleic acids binding driven by electrostatic interactions typically are characterized by the release of counter ions, and the salt-inhibited binding association constant (K(a)) and the magnitude of exothermic binding enthalpy (ΔH). Here, we report a non-classical thermodynamics of streptavidin (SA)-aptamer binding in NaCl (140-350 mM) solutions near room temperatures (23-27 °C). By using isothermal titration calorimetry (ITC) and circular dichroism (CD)/fluorescence spectroscopy, we found that the binding was enthalpy driven with a large entropy cost (ΔH -20.58 kcal mol(-1), TΔS -10.99 kcal mol(-1), and K(a) 1.08 × 10(7)  M(-1) at 140 mM NaCl 25 °C). With the raise of salt concentrations, the ΔH became more exothermic, yet the K(a) was almost unchanged (ΔH -26.29 kcal mol(-1) and K(a) 1.50 × 10(7)  M(-1) at 350 mM NaCl 25 °C). The data suggest that no counter Na(+) was released in the binding. Spectroscopy data suggest that the binding, with a stoichiometry of 2, was accompanied with substantial conformational changes on SA, and the changes were insensitive to the variation of salt concentrations. To account for the non-classical results, we propose a salt bridge exchange model. The intramolecular binding-site salt bridge(s) of the free SA and the charged phosphate group of aptamers re-organize to form the binding complex by forming a new intermolecular salt bridge(s). The salt bridge exchange binding process requires minimum amount of counter ions releasing but dehydration of the contacting surface of SA and the aptamer. The energy required for dehydration is reduced in the case of binding solution with higher salt concentration and account for the higher binding exothermic mainly.

An enthalpic basis of additivity in biphenyl hydroxamic acid ligands for stromelysin-1

Fragment based drug discovery remains a successful tool for pharmaceutical lead discovery. Although based upon the principle of thermodynamic additivity, the underlying thermodynamic basis is poorly understood. A thermodynamic additivity analysis was performed using stromelysin-1 and a series of biphenyl hydroxamate ligands identified through fragment additivity. Our studies suggest that, in this instance, additivity arises from enthalpic effects, while interaction entropies are unfavorable; this thermodynamic behavior is masked by proton transfer. Evaluation of the changes in constant pressure heat capacities during binding suggest that solvent exclusion from the binding site does not account for the dramatic affinity enhancements observed.

The interaction of hyperthermophilic TATA-box binding protein with single-stranded DNA is entropically favorable and exhibits a large negative heat capacity change at high salt concentration

We have investigated the thermodynamics of the interaction between the TATA-box-binding protein from Pyrococcus horikoshii (PhoTBP) and its target DNA (TATA-1). The interaction between PhoTBP and double-stranded DNA (dsDNA) is entropically favorable and enthalpically unfavorable. The thermodynamic parameters for TATA-1 duplex formation in the presence of PhoTBP, that is, ternary PhoTBP-dsDNA complexation, are similar to those for TATA-1 duplex formation, which is enthalpically favorable. Surface plasmon resonance analysis indicates that the interaction between PhoTBP and single-stranded DNA (ssDNA) of TATA-1 is entropy driven and has a large negative heat capacity change (-1.19 kcal mol(-1) K(-1)) at high salt concentration (800 mM NaCl). These results suggest that the favorable entropic effect corresponding to the interaction between PhoTBP and dsDNA is due not to ternary complexation but to the interaction between PhoTBP and ssDNA. This report is the first to describe the thermodynamics of the interaction between TBP and ssDNA.

Impact of DNA hairpin folding energetics on antibody-ssDNA association

Deposition of anti-DNA antibodies in the kidney contributes to the pathogenesis of the autoimmune disease, systemic lupus erythematosus. Antibodies that bind to hairpin-forming DNA ligands may be particularly prone to deposition. Here we report the first structure of a Fab complexed with hairpin-forming DNA. The ligand used for co-crystallization is 5'-d [CTG(CCTT)CAG]-3', which has a predicted hairpin structure consisting of a four-nucleotide loop (CCTT) and a stem of three base-pairs. The 1.95 A resolution crystal structure of Fab DNA-1 complexed with this ligand shows that the conformation of the bound ligand differs radically from the predicted hairpin conformation. The three base-pairs in the stem are absent in the bound form. The protein binds to the last six nucleotides at the 3' end of the ligand. These nucleotides form a loop (TTCA) closed by a G:C base-pair in the bound state. Stacking of aromatic side-chains against DNA bases is the dominant interaction in the complex. Interactions with the DNA backbone are conspicuously absent. Thermodynamics of binding are examined using isothermal titration calorimetry. The apparent dissociation constant is 4 microM, and binding is enthalpically favorable and entropically unfavorable. Increasing the number of base-pairs in the DNA stem from three to six decreases binding affinity. These data suggest a conformational selection binding mechanism in which the Fab binds preferentially to the unstructured state of the ligand. In this interpretation, the ligand binding and ligand folding equilibria are coupled, with lower hairpin stability leading to greater effective binding affinity. Thus, pre-organization of the DNA loop into the preferred binding conformation does not play a major role in complexation. Rather, it is argued that the stem of the hairpin serves to reduce the degrees of freedom in the free DNA ligand, thereby limiting the entropic cost attendant to complexation with the Fab.

Effect of somatic mutation on DNA binding properties of anti-DNA autoantibodies

Autoantibodies that bind DNA are a hallmark of systemic lupus erythematosus. A subset of autoantibody*DNA complexes localize to kidney tissue and lead to damage and even death. 11F8, 9F11, and 15B10 are clonally related anti-DNA autoantibodies isolated from an autoimmune mouse. 11F8 binds ssDNA in a sequence-specific manner and causes tissue damage, while 9F11 and 15B10 bind ssDNA non-specifically and are benign. Among these antibodies, DNA binding properties are mediated by five amino acid differences in primary sequence. Thermodynamic and kinetic parameters associated with recognition of structurally different DNA sequences were determined for each antibody to provide insight toward recognition strategies, and to explore a link between binding properties and disease pathogenesis. A model of 11F8 bound to its high affinity consensus sequence provides a foundation for understanding the differences in thermodynamic and kinetic parameters between the three mAbs. Our data suggest that 11F8 utilizes the proposed ssDNA recognition motif including (Y32)V(L), a hydrogen bonding residue at (91)V(L), and an aromatic residue at the tip of the third heavy chain complementarity determining region. Interestingly, a somatic mutation to arginine at (31)V(H) in 11F8 may afford additional binding site contacts including (R31)V(H), (R96)V(H), and (R98)V(H) that could determine specificity.

Specific recognition of a DNA immunogen by its elicited antibody

DNA recognition by antibodies is a key feature of autoimmune diseases, yet model systems with structural information are very limited. The monoclonal antibody ED-10 recognizes one of the strands of the DNA duplex used in the immunogenic complex. Modifications of the 5' end decrease the binding affinity and short oligonucleotides retain high binding affinity. We determined crystal structures for the Fab bound to a 6-mer oligonucleotide containing the specific sequence that raised the antibody and compared it with the unliganded Fab. Only the first two bases from the 5' end (dTdC) display electron density and we observe four key hydrogen bonds at the interface. The thymine ring is stacked between TrpH50 and TrpH95, and the cytosine ring is packed against TyrL32. Upon DNA binding, TyrH97 and TrpH95 rearrange to allow subnanomolar binding affinity, five orders of magnitude higher than other reported complexes, possibly because of having gone through affinity maturation. This structure represents the first bona fide antibody DNA immunogen complex described in atomic detail.

Role of conformational dynamics in sequence-specific autoantibody•ssDNA recognition

11F8 is a sequence-specific monoclonal anti-ssDNA autoantibody isolated from a lupus prone mouse that forms pathogenic complexes with ssDNA, resulting in kidney damage. Prior studies show that specificity is mediated by a somatic mutation from serine at (31)V(H) to arginine. Reversion back to serine in 11F8 resulted in >30-fold decrease in affinity and altered thermodynamic and kinetic parameters for sequence-specific recognition of its cognate ssDNA ligand. Mutagenesis and structural studies suggest that (R31)V(H) contacts ssDNA via a salt bridge and a bidentate hydrogen bond and may further contribute to specificity by altering binding-site conformation. Fluorescence resonance energy transfer experiments were conducted to assess the kinetics of conformational change during 11F8*ssDNA association. The extent of rearrangement between the six complementary determining regions in the 11F8*ssDNA complex with germline serine or somatically mutated arginine at residue 31 of the heavy chain was examined. Our studies show that greater conformational change occurs in five of six complementarity determining regions after the heavy chain germline J558 sequence undergoes mutation to arginine at (31)V(H).

Aptamer-oligonucleotide binding studied by capillary electrophoresis: cation effect and separation efficiency

Novel features of DNA structure, recognition and discrimination have been recently elucidated through the solution structural characterization of DNA aptamers that bind cofactors, amino acids and peptides with high affinity and specificity. Multidimensional nuclear magnetic resonance methodologies have been successfully applied to solve the solution structures. In this work, it was demonstrated that capillary electrophoresis was a powerful tool allowing the fundamental study of the binding mechanism between a DNA aptamer and three ligands, adenosine and adenylate compounds, i.e., adenosine diphosphate (ADP) and adenosine triphosphate (ATP). In order to gain further insight into this binding, thermodynamic measurements under different values of parameters (such as salt nature and its concentration (x) in the run buffer) were carried out. The results showed that dehydration at the binding interface, van der Waals interactions, H-bonding and adjustment of the aptamer recognition surface were implied in the aptamer-ligand association. As well, it was demonstrated that the addition in the medium of the sodium monovalent cation Na(+) or the nickel divalent cation Ni(2+) decreased the complex formation. Separation efficiency and peak shape can also be improved by Mg(2+) divalent cation, which increased the mass transfer kinetics during the ligand-aptamer binding process. A significant separation for the worst separated pair of peaks on the electropherogram ((ADP, ATP) peak pair) was thus achieved.

Evidence for Structural Plasticity of Heavy Chain Complementarity-determining Region 3 in Antibody–ssDNA Recognition

Anti-DNA antibodies play important roles in the pathogenesis of autoimmune diseases. They also represent a unique and relatively unexplored class of DNA-binding protein. Here, we present a study of conformational changes induced by DNA binding to an anti-ssDNA Fab known as DNA-1. Three crystal structures are reported: a complex of DNA-1 bound to dT3, and two structures of the ligand-free Fab. One of the ligand-free structures was determined from crystals exhibiting perfect hemihedral twinning, and the details of structure determination are provided. Unexpectedly, five residues (H97-H100A) in the apex of heavy chain complementarity-determining region 3 (HCDR3) are disordered in both ligand-free structures. Ligand binding also caused a 2-4A shift of the backbone of Tyr L92 and ordering of the L92 side-chain. In contrast, these residues are highly ordered in the Fab/dT3 complex, where Tyr H100 and Tyr H100A form intimate stacking interactions with DNA bases, and L92 forms the 5' end of the binding site. The structures suggest that HCDR3 is very flexible and adopts multiple conformations in the ligand-free state. These results are discussed in terms of induced fit and pre-existing equilibrium theories of ligand binding. Our results allow new interpretations of existing thermodynamic and mutagenesis data in terms of conformational entropy and the volume of conformational space accessible to HCDR3 in the ligand-free state. In the context of autoimmune disease, plasticity of the ligand-free antibody could provide a mechanism by which anti-DNA antibodies bind diverse host ligands, and thereby contribute to pathogenicity.

Specific recognition of a dsDNA sequence motif by an immunoglobulin VH homodimer

Anti-DNA antibodies have the potential to be applied in vast fields of fundamental as well as medical research. They are found in autoimmune diseases, such as systemic lupus erythemotosus. In most cases, anti-dsDNA antibodies do not present sequence specificity and are of low affinity. The dominant role of VH domains in DNA recognition induced us to search for binders based on VH dimers (VHD), previously reported to bind different protein antigens. We screened a phage displayed homo-VHD library against a 19-bp dsDNA sequence. A sequence-specific binder was selected, which recognizes the terminal located CTGC motif with a Kd of 250 nM. Association of the two identical VH domains of the molecule was shown to be essential for binding.

Structure of an anti-DNA fab complexed with a non-DNA ligand provides insights into cross-reactivity and molecular mimicry

Antibodies that recognize DNA (anti-DNA) are part of the autoimmune response underlying systemic lupus erythematosus. To better understand molecular recognition by anti-DNA antibodies, crystallographic studies have been performed using an anti-ssDNA antigen-binding fragment (Fab) known as DNA-1. The previously determined structure of a DNA-1/dT5 complex revealed that thymine bases insert into a narrow groove, and that ligand recognition primarily involves the bases of DNA. We now report the 1.75-A resolution structure of DNA-1 complexed with the biological buffer HEPES (4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid). All three light chain complementarity-determining regions (CDRs) and HCDR3 contribute to binding. The HEPES sulfonate hydrogen bonds to His L91, Asn L50, and to the backbone of Tyr H100 and Tyr H100A. The Tyr side-chains of L32, L92, H100, and H100A form nonpolar contacts with the HEPES ethylene and piperazine groups. Comparison to the DNA-1/dT5 structure reveals that the dual recognition of dT5 and HEPES requires a 13-A movement of HCDR3. This dramatic structural change converts the combining site from a narrow groove, appropriate for the edge-on insertion of thymine bases, to one sufficiently wide to accommodate the HEPES sulfonate and piperazine. Isothermal titration calorimetry verified the association of HEPES with DNA-1 under conditions similar those used for crystallization (2 M ammonium sulfate). Interestingly, the presence of 2 M ammonium sulfate increases the affinities of DNA-1 for both HEPES and dT5, suggesting that non-polar Fab-ligand interactions are important for molecular recognition in highly ionic solvent conditions. The structural and thermodynamic data suggest a molecular mimicry mechanism based on structural plasticity and hydrophobic interactions.

Energetics of the sequence-specific binding of single-stranded DNA by the F factor relaxase domain

Transfer of conjugative plasmids between bacteria requires the activity of relaxases or mobilization proteins. These proteins nick the plasmid in a site- and strand-specific manner prior to transfer of the cut strand from donor to recipient. TraI36, the relaxase domain of TraI from plasmid F factor, binds a single-stranded DNA (ssDNA) oligonucleotide containing an F factor sequence with high affinity and sequence specificity. To better understand the energetics of this interaction, we examined the temperature, salt, and pH dependence of TraI36 recognition. Binding is entropically driven below 25 degrees C and enthalpically driven at higher temperatures. van't Hoff analysis yields an estimated deltaC(P)(0) of binding (-3300 cal x mol(-1) x K(-1)) that is larger and more negative than that observed for most double-stranded DNA (dsDNA)-binding proteins. Based on analyses of circular dichroism data and the crystal structure of the unliganded protein, we attribute the deltaC(P)(0) to both burial of hydrophobic surface area and coupled folding and binding of the protein. The salt dependence of the binding indicates that several ssDNA phosphates are buried in the complex, and the pH dependence of the binding suggests that some of these ssDNA phosphates form ionic interactions with basic residues of the protein. Although data are available for relatively few sequence-specific ssDNA-binding proteins, sufficient differences exist between TraI36 and other proteins to indicate that, like dsDNA-binding proteins, ssDNA-binding proteins use different motifs and combinations of forces to achieve specific recognition.

Antibody recognition of chiral surfaces. Structural models of antibody complexes with leucine-leucine-tyrosine crystal surfaces

Molecular models are built of the recognition domains of two antibodies, which are raised and selected against crystals of (L)leucine-(L)leucine-(L)tyrosine. The model of one antibody, which is stereo- and enantioselective, reveals astounding chemical and structural complementarity to the recognized crystal surface. The enantioselective binding of this antibody is explained by the significantly fewer chemical interactions arising in the complex, after docking of the antibody to the (D)Leu-(D)Leu-(D)Tyr crystal face, relative to its enantiomer, the (L)Leu-(L)Leu-(L)Tyr crystal face. The modeling and docking of the second antibody, which is poorly stereoselective and is not enantioselective, indicates that binding is based on electrostatic interactions. The docking models of the antibody-crystal complexes provide a rationale for the experimental results while demonstrating the power of modeling techniques to meet the challenge of describing antibody-antigen interactions in detail.

A DNA aptamer as a new target-specific chiral selector for HPLC

In this paper, a DNA aptamer, known to bind stereospecifically the D-enantiomer of an oligopeptide, i.e., arginine-vasopressin, was immobilized on a chromatographic support. The influence of various parameters (such as column temperature, eluent pH, and salt concentration) on the L- and D-peptide retention was investigated in order to provide information about the binding mechanism and then to define the utilization conditions of the aptamer column. The results suggest that dehydration at the binding interface, charge-charge interactions, and adaptive conformational transitions contribute to the specific D-peptide-aptamer complex formation. A very significant enantioselectivity was obtained in the optimal binding conditions, the D-peptide being strongly retained by the column while the L-peptide eluted in the void volume. A rapid baseline separation of peptide enantiomers was also achieved by modulating the elution conditions. Furthermore, it was established that the aptamer column was stable during an extended period of time. This work indicates that DNA aptamers, specifically selected against an enantiomer, could soon become very attractive as new target-specific chiral selectors for HPLC.

Thermodynamic correlation analysis: hydration and perturbation sensitivity of RNA secondary structures

The thermodynamic parameters DeltaH, DeltaS, T(m), and DeltaG of a total of 36 RNA strands, 22 tetralooped 22mers, and 14 heptalooped 25mers (same stem sequence) were analyzed with respect to enthalpy-entropy compensation (EEC). The EEC plots [DeltaH, DeltaS] were compared with collected literature data from protein and nucleic acid unfolding studies (3224 and 241 datapoints, respectively) which all proved to be remarkably linear. The similarity of the compensation slopes and intercepts for all compounds indicate that, irrespective of the chemical nature and stability of the folding solutes, the exothermicity DeltaH and entropic penalty T x DeltaS of folding are strongly dominated by the rearrangement and formation of hydration layers around the solutes, while it is well-known that the stability of folding results only from the difference (DeltaG) and ratio (T(m)) of both parameters. EEC plots [DeltaH, DeltaS] are presented in an extended context, as 3D plots [DeltaH, DeltaS, T(m)] allowing for a correct analytical description of the enthalpy-entropy relationship and for more practical interpretations of large amounts of thermodynamic data when replotted as [DeltaH, T(m)] or [DeltaG(T), T(m)]. The introduction of a variety of mismatches into nucleic acids, or limited irreguliarities into any supramolecular complex, and the analysis of the involved thermodynamics as shown in this study-i.e., scanning the "enthalpy-entropy space" of whole macromolecular subgroups-should permit to extract and quantify more "hidden information", such as hydration extent and sensitivity of macromolecular frameworks toward desolvation and structural perturbation, from thermodynamic analyses of large sample sizes.

Crystal structure of an antigen-binding fragment bound to single-stranded DNA

Antibodies to DNA are characteristic of the autoimmune disease systemic lupus erythematosus (SLE) and they also serve as models for the study of protein-DNA recognition. Anti-DNA antibodies often play an important role in disease pathogenesis by mediating kidney damage via antibody-DNA immune complex formation. The structural underpinnings of anti-DNA antibody pathogenicity and antibody-DNA recognition, however, are not well understood, due in part to the lack of direct, experimental three-dimensional structural information on antibody-DNA complexes. To address these issues for anti-single-stranded DNA antibodies, we have determined the 2.1 A crystal structure of a recombinant Fab (DNA-1) in complex with dT5. DNA-1 was previously isolated from a bacteriophage Fab display library from the immunoglobulin repertoire of an SLE-prone mouse. The structure shows that DNA-1 binds oligo(dT) primarily by sandwiching thymine bases between Tyr side-chains, which allows the bases to make sequence-specific hydrogen bonds. The critical stacking Tyr residues are L32, L49, H100, and H100A, while His L91 and Asn L50 contribute hydrogen bonds. Comparison of the DNA-1 structure to other anti-nucleic acid Fab structures reveals a common ssDNA recognition module consisting of Tyr L32, a hydrogen bonding residue at position L91, and an aromatic side-chain from the tip of complementarity determining region H3. The structure also provides a framework for interpreting previously determined thermodynamics data, and this analysis suggests that hydrophobic desolvation might underlie the observed negative enthalpy of binding. Finally, Arg side-chains from complementarity determining region H3 appear to play a novel role in DNA-1. Rather than forming ion pairs with dT5, Arg contributes to oligo(dT) recognition by helping to maintain the structural integrity of the combining site. This result is significant because antibody pathogenicity is thought to be correlated to the Arg content of anti-DNA antibody hypervariable loops.