Prediction of mono- and di-nucleotide-specific DNA-binding sites in proteins using neural networks

Background  

DNA recognition by proteins is one of the most important processes in living systems. Therefore, understanding the recognition process in general, and identifying mutual recognition sites in proteins and DNA in particular, carries great significance. The sequence and structural dependence of DNA-binding sites in proteins has led to the development of successful machine learning methods for their prediction. However, all existing machine learning methods predict DNA-binding sites, irrespective of their target sequence and hence, none of them is helpful in identifying specific protein-DNA contacts. In this work, we formulate the problem of predicting specific DNA-binding sites in terms of contacts between the residue environments of proteins and the identity of a mononucleotide or a dinucleotide step in DNA. The aim of this work is to take a protein sequence or structural features as inputs and predict for each amino acid residue if it binds to DNA at locations identified by one of the four possible mononucleotides or one of the 10 unique dinucleotide steps. Contact predictions are made at various levels of resolution viz. in terms of side chain, backbone and major or minor groove atoms of DNA.     

Damage increases the flexibility of duplex DNA

It is proposed that much of the recognition of specific types of damaged DNAs is based on accessible structural features, while much of the recognition of damaged DNAs, as a class, is based on flexibility. The more flexible a DNA the faster its diffusion rate. The diffusion rates of each member of a series of damaged duplex DNAs has been found to be significantly faster than that of the corresponding undamaged duplex DNA. The damaged sites studied include apurinic and apyrimidinic a basic sites, thymine glycol and urea. The presence of mismatched sites also increases the diffusion. Thus, damaged DNAs appear to have sufficient flexibility for recognition and the flexibility may allow damaged sites to act as a universal joint or hinge that allows distant sites on the DNA to come together.     

Testing water-mediated DNA recognition by the Hin recombinase

The Hin recombinase specifically recognizes its DNA-binding site by means of both major and minor groove interactions. A previous X-ray structure, together with new structures of the Hin DNA-binding domain bound to a recombination half-site that were solved as part of the present study, have revealed that two ordered water molecules are present within the major groove interface. In this report, we test the importance of these waters directly by X-ray crystal structure analysis of complexes with four mutant DNA sequences. These structures, combined with their Hin-binding properties, provide strong support for the critical importance of one of the intermediate waters. A lesser but demonstrable role is ascribed to the second water molecule. The mutant structures also illustrate the prominent roles of thymine methyls both in stabilizing intermediate waters and in interfering with water or amino acid side chain interactions with DNA.     

Molecular dynamics simulations of the effects of ring-saturated thymine lesions on DNA structure

The effect of thymine lesions produced by radiation or oxidative damage on DNA structure was studied by molecular dynamics simulations of native and damaged DNA. Thymine in position 7 of native dodecamer d(CGCGAATTCGCG)₂ was replaced by one of the four thymine lesions 5-hydroxy-5,6-dihydrothymine, 6-hydroxy-5,6-dihydrothymine (thymine photohydrate), 5,6-dihydmxy-5,6-dihydro-thymine (thymine glycol), and 5,6-dihydmthymine. Simulations were performed with Assisted Model Building with Energy Refinement force field. Solvent was represented by a rectangular box of water with periodic boundary conditions applied. A constant temperature and constant volume protocol was used, the observed level of distortions of DNA structure depends on the specific nature of the lesion. The 5,6-dihydrothymine does not cause distinguishable perturbations to DNA. Other lesions produce a dramatic increase in the rise parameter between the lesion and the 5′ adjacent adenine. These changes are accompanied by weakening of Watson–Crick hydrogen bonds in the A6-T19 base pair on the 5′ side of the lesion. The lesioned bases also show negative values of inclination relative to the helical axis. No changes in the pattern of backbone torsional angles are observed with any of the lesions incorporated into DNA. The structural distortions in DNA correlate well with known biological effects of 5,6-dihydrothymine and thymine glycol on such processes as polymerase action or recognition by repair enzymes. © 1995 John Wiley & Sons, Inc.     

Structure-wise discrimination of cytosine,thymine, and uracil by proteins in terms of their nonbonded interactions

The molecular recognition and discrimination of very similar ligand moieties by proteins are important subjects in protein–ligand interaction studies. Specificity in the recognition of molecules is determined by the arrangement of protein and ligand atoms in space. The three pyrimidine bases, viz. cytosine, thymine, and uracil, are structurally similar, but the proteins that bind to them are able to discriminate them and form interactions. Since nonbonded interactions are responsible for molecular recognition processes in biological systems, our work attempts to understand some of the underlying principles of such recognition of pyrimidine molecular structures by proteins. The preferences of the amino acid residues to contact the pyrimidine bases in terms of nonbonded interactions; amino acid residue–ligand atom preferences; main chain and side chain atom contributions of amino acid residues; and solvent-accessible surface area of ligand atoms when forming complexes are analyzed. Our analysis shows that the amino acid residues, tyrosine and phenyl alanine, are highly involved in the pyrimidine interactions. Arginine prefers contacts with the cytosine base. The similarities and differences that exist between the interactions of the amino acid residues with each of the three pyrimidine base atoms in our analysis provide insights that can be exploited in designing specific inhibitors competitive to the ligands.     

Sequence-specific interaction of the Salmonella Hin recombinase in both major and minor grooves of DNA.

The Hin recombinase of Salmonella catalyzes a site-specific recombination event which leads to flagellar phase variation. Starting with a fully symmetrical recombination site, hixC, a set of 40 recombination sites which vary by pairs of single base substitutions was constructed. This set was incorporated into the Salmonella-specific bacteriophage P22 based challenge phage selection and used to define the DNA sequence determinants for the binding of Hin to DNA in vivo. The critical sequence-specific contacts between a Hin monomer and a 13 bp hix half-site are at two T:A base pairs in the major groove of the DNA which are separated by one base pair, and two consecutive A:T contacts in the minor groove. The base substitutions in the major groove recognition portion which were defective in binding Hin still retained residual binding capability in vivo, while the base pair substitutions affecting the minor groove recognition region lost all in vivo binding. Using in vitro binding assays, Hin was found to bind to hix symmetrical sites with A:T base pairs or I:C base pairs in the minor groove recognition sequences, but not to G:C base pairs. In separate in vitro binding assays, Hin was equally defective in binding to either a G:C or a I:C contact in a major groove recognition sequence. Results from in vitro binding assays to hix sites in which 3-deazaadenine was substituted for adenine are consistent with Hin making a specific contact to either the N3 of adenine or O2 of thymine in the minor groove within the hix recombination site on each symmetric half-site. These results taken with the results of previous studies on the DNA binding domain of Hin suggest a sequence-specific minor groove DNA binding motif.     

Molecular recognition of B-DNA by Hoechst 33258.

The binding sites of Hoechst 33258, netropsin and distamycin on three DNA restriction fragments from plasmid pBR322 were compared by footprinting with methidiumpropyl-EDTA X Fe(II) [MPE X Fe(II)]. Hoechst, netropsin and distamycin share common binding sites that are five +/- one bp in size and rich in A X T DNA base pairs. The five base pair protection patterns for Hoechst may result from a central three base pair recognition site bound by two bisbenzimidazole NHs forming a bridge on the floor of the minor groove between adjacent adenine N3 and thymine O2 atoms on opposite helix strands. Hydrophobic interaction of the flanking phenol and N-methylpiperazine rings would afford a steric blockade of one additional base pair on each side.     

Construction of DNA substrates modified with psoralen at a unique site and study of the action mechanism of ABC excinuclease on these uniformly modified substrates

Psoralens bind to DNA noncovalently and upon exposure to near UV (320-400 nm) light produce covalent adducts. Thymidine residues in DNA, especially those at 5'-TpA-3' sequences, are most susceptible to the photochemical reaction. This property of the reaction and the recent advances in oligonucleotide synthesis and separation has enabled us to construct DNA fragments containing psoralen adducts at a specific site. The octanucleotide 5'-TCGTAGCT-3' was photoreacted (in the presence of the complementary strand) with the synthetic psoralen 4'-hydroxymethyl-4,5',8-trimethylpsoralen to obtain oligonucleotides adducted via the furan or pyrone ring at the internal thymine. These modified octanucleotides were ligated to nonmodified oligonucleotides to obtain a 40-base pair DNA fragment containing a psoralen adduct at a central location. The modified fragment having the thymine-furan side 4'-hydroxymethyl-4,5',8-trimethylpsoralen adduct was irradiated with 360 nm of light to produce an interstrand cross-link, and this cross-linked DNA was purified to homogeneity. These uniquely modified DNAs were used as substrates for Escherichia coli ABC excinuclease to determine its incision mechanism unambiguously and to determine the contact sites of the enzyme. ABC excinuclease mediates the cleavage of the 8th and 5th phosphodiester bonds 5' and 3', respectively, to psoralen monoadducts, and the 9th (5') and 3rd (3') phosphodiester bonds to the furan-side thymine of the cross-link. Preliminary DNaseI footprinting studies show that ABC excinuclease protects the whole 40-base pair fragment from DNaseI, and binding of the A and B subunits to the furan side-monoadducted substrate produces two hypersensitive phosphodiester bonds in the vicinity of the 5' incision site of ABC excinuclease.     

DNA curving and bending in protein–DNA recognition

Most biological events are regulated at the molecular level by site-specific associations between specialized proteins and DNA. These associations may bring distal regions of the genome into functional contact or may lead to the formation of large multisubunit complexes capable of regulating highly site-specific transactional events. It is now believed that sequence-specific protein-DNA recognition and the ability of certain proteins to compete for multiple binding sites is regulated at several levels by the local structure and conformation of the binding partners. These encompass the microstructure of DNA, including its curvature, bending and flexing as well as conformational lability in the DNA-binding domains of the proteins. Possible mechanisms for binding specificity are discussed in the context of specific nucleoprotein systems with particular emphasis given to the roles of DNA conformations in these interactions.     

Replication origins in metazoan chromosomes: fact or fiction?

The process by which eukaryotic cells decide when and where to initiate DNA replication has been illuminated in yeast, where specific DNA sequences (replication origins) bind a unique group of proteins (origin recognition complex) next to an easily unwound DNA sequence at which replication can begin. The origin recognition complex provides a platform on which additional proteins assemble to form a pre-replication complex that can be activated at S-phase by specific protein kinases. Remarkably, multicellular eukaryotes, such as frogs, flies, and mammals (metazoa), have counterparts to these yeast proteins that are required for DNA replication. Therefore, one might expect metazoan chromosomes to contain specific replication origins as well, a hypothesis that has long been controversial. In fact, recent results strongly support the view that DNA replication origins in metazoan chromosomes consist of one or more high frequency initiation sites and perhaps several low frequency ones that together can appear as a nonspecific initiation zone. Specific replication origins are established during G1-phase of each cell cycle by multiple parameters that include nuclear structure, chromatin structure, DNA sequence, and perhaps DNA modification. Such complexity endows metazoa with the flexibility to change both the number and locations of replication origins in response to the demands of animal development.     

Sites of termination of in vitro DNA synthesis on psoralen phototreated single-stranded templates

Single-stranded DNA has been photochemically induced to react with 4'-hydroxymethyl-4,5',8-trimethylpsoralen (HMT) and used as substrate for DNA replication with E. coli DNA polymerase I large fragment. By using the dideoxy sequencing procedure, it is possible to map the termination sites on the template photoreacted with HMT. These sites occur at the nucleotides preceding each thymine residue (and a few cytosine residues), emphasizing the fact that in a single-stranded stretch of DNA, HMT reacts with each thymine residue without any specificity regarding the flanking base sequence of the thymine residues. In addition, termination of DNA synthesis due to psoralen-adducted thymine is not influenced by the efficiency of the 3'-5'exonuclease proof-reading activity of the DNA polymerase.     

Sequence dependence for bypass of thymine glycols in DNA by DNA polymerase I.

Single-stranded phage DNAs containing thymine glycols were prepared by oxidation with osmium tetroxide (OsO4) and were used as templates for DNA synthesis by E. coli DNA polymerase I. The induction of thymine glycol lesions in DNA, as measured by immunoassay, quantitatively accounted for an inhibition of in vitro DNA synthesis on modified templates. Analysis of termination sites for synthesis by DNA polymerase I (Klenow fragment) showed that DNA synthesis terminated at most template thymine sites in OsO4-treated DNA, indicating that incorporation occurred opposite putative thymine glycols in DNA. Nucleotides 5' and 3' to putative thymine glycol sites affect the reaction, however, since termination was not observed at thymines in the sequence 5'-CTPur-3'. Conversion of thymine glycols to urea residues in DNA by alkali treatment caused termination of DNA synthesis one nucleotide 3' to template thymine sites, including thymines in the 5'-CTPur-3' sequence, showing that the effect of surrounding sequence is on the elongation reaction by DNA polymerase rather than differential damage induction by OsO4.     

Designing of MIP based QCM sensor having thymine recognition sites based on biomimicking DNA approach

Quartz crystal microbalance (QCM) sensors coated with molecular imprinted polymers (MIP) have been developed for the determination of thymine. In this method, methacryloylamidoadenine (MA-Ade) have used as a new monomer and thymine template for inspiration of DNA nucleobases interaction. The thymine can be simultaneously hydrogen binding to MA-Ade and fit into the shape-selective cavities. Thus, the interaction between nucleobases has an effect on the binding ability of the QCM sensors. The binding affinity of the thymine imprinted sensors has investigated by using the Langmuir isotherm. The thymine imprinted QCM electrodes have shown homogeneous binding sites for thymine (K_a: 1.0 × 10⁵ M⁻¹) while heterogeneous binding sites for uracil. On the other hand, recognition selectivity of the QCM sensor based on thymine imprinted polymer toward to uracil, ssDNA and ssRNA has been reported in this work.     

Site-specific interactions of JBP with base and sugar moieties in duplex J-DNA. Evidence for both major and minor groove contacts

Beta-D-Glucosyl-hydroxymethyluracil, also called base J, is an unusually modified DNA base conserved among Kinetoplastida. Base J is found predominantly in repetitive DNA and correlates with epigenetic silencing of telomeric variant surface glycoprotein genes. We have previously identified a J-binding protein (JBP) in Trypanosoma, Leishmania, and Crithidia, and we have shown that it is a structure-specific binding protein. Here we examine the molecular interactions that contribute to recognition of the glycosylated base in synthetic DNA substrates using modification interference, modification protection, DNA footprinting, and photocross-linking techniques. We find that the two primary requirements for J-DNA recognition include contacts at base J and a base immediately 5' of J (J-1). Methylation interference analysis indicates that the requirement of the base at position J-1 is due to a major groove contact independent of the sequence. DNA footprinting of the JBP.J-DNA complex with 1,10-phenanthroline-copper demonstrates that JBP contacts the minor groove at base J. Substitution of the thymine moiety of J with cytosine reduces the affinity for JBP approximately 15-fold. These data indicate that the sole sequence dependence for JBP binding may lie in the thymine moiety of base J and that recognition requires only two specific base contacts, base J and J-1, within both the major and minor groove of the J-DNA duplex.     

SOS processing of unique oxidative DNA damages in Escherichia coli

phi X174 replicative form (RF) I transfecting DNA containing thymine glycols (5,6-dihydroxy-5,6-dihydrothymine), urea glycosides or apurinic (AP) sites was used to study SOS processing of unique DNA damages in Escherichia coli. All three lesions can be found in DNA damaged by chemical oxidants or radiation and are representative of several common structural modifications of DNA bases. When phi X DNA containing thymine glycols was transfected into host cells that were ultraviolet-irradiated to induce the SOS response, a substantial increase in survival was observed compared to transfection into uninduced hosts. Studies with mutants demonstrated that both the activated form of RecA and UmuDC proteins were required for this reactivation. In contrast, no increase in survival was observed when DNA containing urea glycosides or AP sites was transfected into ultraviolet-induced hosts. These data suggest that SOS-induced reactivation does not reflect a generalized repair system for all replication-blocking, lethal lesions but rather that the efficiency of reactivation is damage dependent. Further, we found that a significant fraction of potentially lethal thymine glycols could be ultraviolet-reactivated in an umuC lexA recA-independent manner, suggesting the existence of an as yet uncharacterized damage-inducible SOS-independent mode of thymine glycol repair.