The potential of the cruciform structure formation as an important factor influencing p53 sequence-specific binding to natural DNA targets

Interaction of C-terminal p53 isoforms depends strongly upon DNA sequence and topology

The p53 protein is a key tumor suppressor and the most commonly mutated and down-regulated protein in human tumors. It functions mainly through interaction with DNA, and p53 acts as a transcription factor that recognizes the so-called p53 target sites on the promoters of various genes. P53 has been shown to exist as many isoforms, including three C-terminal isoforms that are produced by alternative splicing. Because the C-terminal domain is responsible for sequence-nonspecific binding and regulation of p53 binding, we have analyzed DNA recognition by these C-terminal isoforms. Using atomic force microscopy, we show for the first time that all C-terminal isoforms recognize superhelical DNA. It is particularly noteworthy that a sequence-specific p53 consensus binding site is bound by p53α and β isoforms with similar affinities, whilst p53α shows higher binding to a quadruplex sequence than both p53β and p53γ, and p53γ loses preferential binding to both the consensus binding sequence and the quadruplex-forming sequence. These results show the important role of the variable p53 C-terminal amino acid sequences for DNA recognition.

Structural switching/polymorphism by sequential base substitution at quasi-palindromic SNP site (G → A) in LCR of human β-globin gene cluster

The human β-globin gene Locus Control Region (LCR), a dominant regulator of globin gene expression contains five tissue-specific DNase I-hypersensitive sites (HSs). A single nucleotide polymorphism (SNP) (A → G) present in HS4 region of locus control region (LCR), have shown a notable association between the G allele and the occurrence of β-thalassemia. This SNP site exhibiting a hairpin - duplex equilibrium manifested in A → B like DNA transition has previously been reported from this laboratory. Since, DNA is a dynamic and adaptable molecule, so any change of a single base within a primary DNA sequence can produce major biological consequences commonly manifested in genetic disorders such as sickle cell anemia and β-thalassemia. Herein, the differential behavior of sequential single base substitutions G → A on the quasi-palindromic sequence (d-TGGGGGCCCCA; HPG11) has been explored. A combination of native gel electrophoresis, circular dichroism (CD), and UV-thermal denaturation (Tm) techniques have been used to investigate the structural polymorphism associated with various variants of HPG11 i.e. HPG11A2 to HPG11A5. The CD spectra confirmed that all the HPG11 variants exhibit a hairpin - duplex equilibrium. Oligomer concentration dependence on CD spectra has been correlated with A → B DNA conformational transition. However, as revealed in gel electrophoresis, HPG11A2 → A5 exhibit the formation of a tetramolecular structure (four-way junction) at higher oligomer concentration. UV-melting studies also supported the melting of hairpin, duplex and four-way junction structure. This polymorphism pattern may possibly be significant for DNA-protein recognition, in the process of regulation of LCR in the β-globin gene.

Cruciform Formable Sequences within Pou5f1 Enhancer Are Indispensable for Mouse ES Cell Integrity

DNA can adopt various structures besides the B-form. Among them, cruciform structures are formed on inverted repeat (IR) sequences. While cruciform formable IRs (CFIRs) are sometimes found in regulatory regions of transcription, their function in transcription remains elusive, especially in eukaryotes. We found a cluster of CFIRs within the mouse Pou5f1 enhancer. Here, we demonstrate that this cluster or some member(s) plays an active role in the transcriptional regulation of not only Pou5f1, but also Sox2, Nanog, Klf4 and Esrrb. To clarify in vivo function of the cluster, we performed genome editing using mouse ES cells, in which each of the CFIRs was altered to the corresponding mirror repeat sequence. The alterations reduced the level of the Pou5f1 transcript in the genome-edited cell lines, and elevated those of Sox2, Nanog, Klf4 and Esrrb. Furthermore, transcription of non-coding RNAs (ncRNAs) within the enhancer was also upregulated in the genome-edited cell lines, in a similar manner to Sox2, Nanog, Klf4 and Esrrb. These ncRNAs are hypothesized to control the expression of these four pluripotency genes. The CFIRs present in the Pou5f1 enhancer seem to be important to maintain the integrity of ES cells.

Interaction of a photosensitizer methylene blue with various structural forms (cruciform, bulge duplex and hairpin) of designed DNA sequences

Functionally important, local structural transitions in DNA generate various alternative conformations. Cruciform is one of such alternative DNA structures, usually targeted in genomes by various proteins. Symmetry elements in sequence as inverted repeats are the key factor for cruciform formation, facilitated by the presence of the AT-rich regions. Here, we used biophysical and biochemical techniques such as Gel electrophoresis, Circular dichroism (CD), and UV-thermal melting analysis to explore the structural status of the designed DNA sequences, which had potential to form cruciform structures under physiological conditions. The gel electrophoresis analysis revealed that the designed 53-mer DNA oligonucleotide sequence CR forms an intermolecular bulge duplex with flanking ends, while another sequence CRC adopts an intramolecular hairpin structure with flanking ends. Their equimolar complex (CRCRC) bestowed much-retarded migration due to the formation of a quite intriguing cruciform structure. CD studies confirmed that all the alternative structures (cruciform, bulge duplex, and hairpin with flanking ends) exhibit characteristics of B-DNA type conformation. A triphasic UV-thermal melting curve displayed by the complex formed by the equimolar ratio (CRCRC) is also suggestive of the formation of the cruciform structure. The interaction studies of CR, CRC, and their equimolar complex (1:1) with a photosensitizer methylene blue (MB) indicated that MB could not stabilize the discrete structures formed by CR and CRC sequences, however, the cruciform structure showed a quite significant increment in the melting temperature. Such studies facilitate our understanding of various secondary structures possibly present inside the cell and their interactions with drug/dye molecules.

How p53 Molecules Solve the Target DNA Search Problem: A Review

Interactions between DNA and DNA-binding proteins play an important role in many essential cellular processes. A key function of the DNA-binding protein p53 is to search for and bind to target sites incorporated in genomic DNA, which triggers transcriptional regulation. How do p53 molecules achieve "rapid" and "accurate" target search in living cells? The search dynamics of p53 were expected to include 3D diffusion in solution, 1D diffusion along DNA, and intersegmental transfer between two different DNA strands. Single-molecule fluorescence microscopy enabled the tracking of p53 molecules on DNA and the characterization of these dynamics quantitatively. Recent intensive single-molecule studies of p53 succeeded in revealing each of these search dynamics. Here, we review these studies and discuss the target search mechanisms of p53.

Roles of p53 Family Structure and Function in Non-Canonical Response Element Binding and Activation

The p53 canonical consensus sequence is a 10-bp repeat of PuPuPuC(A/T)(A/T)GPyPyPy, separated by a spacer with up to 13 bases. C(A/T)(A/T)G is the core sequence and purine (Pu) and pyrimidine (Py) bases comprise the flanking sequence. However, in the p53 noncanonical sequences, there are many variations, such as length of consensus sequence, variance of core sequence or flanking sequence, and variance in number of bases making up the spacer or AT gap composition. In comparison to p53, the p53 family members p63 and p73 have been found to have more tolerance to bind and activate several of these noncanonical sequences. The p53 protein forms monomers, dimers, and tetramers, and its nonspecific binding domain is well-defined; however, those for p63 or p73 are still not fully understood. Study of p63 and p73 structure to determine the monomers, dimers or tetramers to bind and regulate noncanonical sequence is a new challenge which is crucial to obtaining a complete picture of structure and function in order to understand how p63 and p73 regulate genes differently from p53. In this review, we will summarize the rules of p53 family non-canonical sequences, especially focusing on the structure of p53 family members in the regulation of specific target genes. In addition, we will compare different software programs for prediction of p53 family responsive elements containing parameters with canonical or non-canonical sequences.

A strong structural correlation between short inverted repeat sequences and the polyadenylation signal in yeast and nucleosome exclusion by these inverted repeats

DNA sequences that read the same from 5′ to 3′ in either strand are called inverted repeat sequences or simply IRs. They are found throughout a wide variety of genomes, from prokaryotes to eukaryotes. Despite extensive research, their in vivo functions, if any, remain unclear. Using Saccharomyces cerevisiae, we performed genome-wide analyses for the distribution, occurrence frequency, sequence characteristics and relevance to chromatin structure, for the IRs that reportedly have a cruciform-forming potential. Here, we provide the first comprehensive map of these IRs in the S. cerevisiae genome. The statistically significant enrichment of the IRs was found in the close vicinity of the DNA positions corresponding to polyadenylation [poly(A)] sites and ~ 30 to ~ 60 bp downstream of start codon-coding sites (referred to as ‘start codons’). In the former, ApT- or TpA-rich IRs and A-tract- or T-tract-rich IRs are enriched, while in the latter, different IRs are enriched. Furthermore, we found a strong structural correlation between the former IRs and the poly(A) signal. In the chromatin formed on the gene end regions, the majority of the IRs causes low nucleosome occupancy. The IRs in the region ~ 30 to ~ 60 bp downstream of start codons are located in the + 1 nucleosomes. In contrast, fewer IRs are present in the adjacent region downstream of start codons. The current study suggests that the IRs play similar roles in Escherichia coli and S. cerevisiae to regulate or complete transcription at the RNA level.

The Amino Acid Composition of Quadruplex Binding Proteins Reveals a Shared Motif and Predicts New Potential Quadruplex Interactors

The importance of local DNA structures in the regulation of basic cellular processes is an emerging field of research. Amongst local non-B DNA structures, G-quadruplexes are perhaps the most well-characterized to date, and their presence has been demonstrated in many genomes, including that of humans. G-quadruplexes are selectively bound by many regulatory proteins. In this paper, we have analyzed the amino acid composition of all seventy-seven described G-quadruplex binding proteins of Homo sapiens. Our comparison with amino acid frequencies in all human proteins and specific protein subsets (e.g., all nucleic acid binding) revealed unique features of quadruplex binding proteins, with prominent enrichment for glycine (G) and arginine (R). Cluster analysis with bootstrap resampling shows similarities and differences in amino acid composition of particular quadruplex binding proteins. Interestingly, we found that all characterized G-quadruplex binding proteins share a 20 amino acid long motif/domain (RGRGR GRGGG SGGSG GRGRG) which is similar to the previously described RG-rich domain (RRGDG RRRGG GGRGQ GGRGR GGGFKG) of the FRM1 G-quadruplex binding protein. Based on this protein fingerprint, we have predicted a new set of potential G-quadruplex binding proteins sharing this interesting domain rich in glycine and arginine residues.

Complex Analyses of Short Inverted Repeats in All Sequenced Chloroplast DNAs

Chloroplasts are key organelles in the management of oxygen in algae and plants and are therefore crucial for all living beings that consume oxygen. Chloroplasts typically contain a circular DNA molecule with nucleus-independent replication and heredity. Using "palindrome analyser" we performed complete analyses of short inverted repeats (S-IRs) in all chloroplast DNAs (cpDNAs) available from the NCBI genome database. Our results provide basic parameters of cpDNAs including comparative information on localization, frequency, and differences in S-IR presence. In a total of 2,565 cpDNA sequences available, the average frequency of S-IRs in cpDNA genomes is 45 S-IRs/per kbp, significantly higher than that found in mitochondrial DNA sequences. The frequency of S-IRs in cpDNAs generally decreased with S-IR length, but not for S-IRs 15, 22, 24, or 27 bp long, which are significantly more abundant than S-IRs with other lengths. These results point to the importance of specific S-IRs in cpDNA genomes. Moreover, comparison by Levenshtein distance of S-IR similarities showed that a limited number of S-IR sequences are shared in the majority of cpDNAs. S-IRs are not located randomly in cpDNAs, but are length-dependently enriched in specific locations, including the repeat region, stem, introns, and tRNA regions. The highest enrichment was found for 12 bp and longer S-IRs in the stem-loop region followed by 12 bp and longer S-IRs located before the repeat region. On the other hand, S-IRs are relatively rare in rRNA sequences and around introns. These data show nonrandom and conserved arrangements of S-IRs in chloroplast genomes.

p73, like its p53 homolog, shows preference for inverted repeats forming cruciforms

p73 is a member of the p53 protein family and has essential functions in several signaling pathways involved in development, differentiation, DNA damage responses and cancer. As a transcription factor, p73 achieves these functions by binding to consensus DNA sequences and p73 shares at least partial target DNA binding sequence specificity with p53. Transcriptional activation by p73 has been demonstrated for more than fifty p53 targets in yeast and/or human cancer cell lines. It has also been shown previously that p53 binding to DNA is strongly dependent on DNA topology and the presence of inverted repeats that can form DNA cruciforms, but whether p73 transcriptional activity has similar dependence has not been investigated. Therefore, we evaluated p73 binding to a set of p53-response elements with identical theoretical binding affinity in their linear state, but different probabilities to form extra helical structures. We show by a yeast-based assay that transactivation in vivo correlated more with the relative propensity of a response element to form cruciforms than to its expected in vitro DNA binding affinity. Structural features of p73 target sites are therefore likely to be an important determinant of its transactivation function.

Requirement or exclusion of inverted repeat sequences with cruciform-forming potential in Escherichia coli revealed by genome-wide analyses

Inverted repeat (IR) sequences are DNA sequences that read the same from 5' to 3' in each strand. Some IRs can form cruciforms under the stress of negative supercoiling, and these IRs are widely found in genomes. However, their biological significance remains unclear. The aim of the current study is to explore this issue further. We constructed the first Escherichia coli genome-wide comprehensive map of IRs with cruciform-forming potential. Based on the map, we performed detailed and quantitative analyses. Here, we report that IRs with cruciform-forming potential are statistically enriched in the following five regions: the adjacent regions downstream of the stop codon-coding sites (referred to as the stop codons), on and around the positions corresponding to mRNA ends (referred to as the gene ends), ~ 20 to ~45 bp upstream of the start codon-coding sites (referred to as the start codons) within the 5'-UTR (untranslated region), ~ 25 to ~ 60 bp downstream of the start codons, and promoter regions. For the adjacent regions downstream of the stop codons and on and around the gene ends, most of the IRs with a repeat unit length of ≥ 8 bp and a spacer size of ≤ 8 bp were parts of the intrinsic terminators, regardless of the location, and presumably used for Rho-independent transcription termination. In contrast, fewer IRs were present in the small region preceding the start codons. In E. coli, IRs with cruciform-forming potential are actively placed or excluded in the regulatory regions for the initiation and termination of transcription and translation, indicating their deep involvement or influence in these processes.

Recognition of Local DNA Structures by p53 Protein

p53 plays critical roles in regulating cell cycle, apoptosis, senescence and metabolism and is commonly mutated in human cancer. These roles are achieved by interaction with other proteins, but particularly by interaction with DNA. As a transcription factor, p53 is well known to bind consensus target sequences in linear B-DNA. Recent findings indicate that p53 binds with higher affinity to target sequences that form cruciform DNA structure. Moreover, p53 binds very tightly to non-B DNA structures and local DNA structures are increasingly recognized to influence the activity of wild-type and mutant p53. Apart from cruciform structures, p53 binds to quadruplex DNA, triplex DNA, DNA loops, bulged DNA and hemicatenane DNA. In this review, we describe local DNA structures and summarize information about interactions of p53 with these structural DNA motifs. These recent data provide important insights into the complexity of the p53 pathway and the functional consequences of wild-type and mutant p53 activation in normal and tumor cells.

Palindrome analyser - A new web-based server for predicting and evaluating inverted repeats in nucleotide sequences

DNA cruciform structures play an important role in the regulation of natural processes including gene replication and expression, as well as nucleosome structure and recombination. They have also been implicated in the evolution and development of diseases such as cancer and neurodegenerative disorders. Cruciform structures are formed by inverted repeats, and their stability is enhanced by DNA supercoiling and protein binding. They have received broad attention because of their important roles in biology. Computational approaches to study inverted repeats have allowed detailed analysis of genomes. However, currently there are no easily accessible and user-friendly tools that can analyse inverted repeats, especially among long nucleotide sequences. We have developed a web-based server, Palindrome analyser, which is a user-friendly application for analysing inverted repeats in various DNA (or RNA) sequences including genome sequences and oligonucleotides. It allows users to search and retrieve desired gene/nucleotide sequence entries from the NCBI databases, and provides data on length, sequence, locations and energy required for cruciform formation. Palindrome analyser also features an interactive graphical data representation of the distribution of the inverted repeats, with options for sorting according to the length of inverted repeat, length of loop, and number of mismatches. Palindrome analyser can be accessed at http://bioinformatics.ibp.cz.

Strong preference of BRCA1 protein to topologically constrained non-B DNA structures

Background

The breast and ovarian cancer susceptibility gene BRCA1 encodes a multifunctional tumor suppressor protein BRCA1, which is involved in regulating cellular processes such as cell cycle, transcription, DNA repair, DNA damage response and chromatin remodeling. BRCA1 protein, located primarily in cell nuclei, interacts with multiple proteins and various DNA targets. It has been demonstrated that BRCA1 protein binds to damaged DNA and plays a role in the transcriptional regulation of downstream target genes. As a key protein in the repair of DNA double-strand breaks, the BRCA1-DNA binding properties, however, have not been reported in detail.

Results

In this study, we provided detailed analyses of BRCA1 protein (DNA-binding domain, amino acid residues 444–1057) binding to topologically constrained non-B DNA structures (e.g. cruciform, triplex and quadruplex). Using electrophoretic retardation assay, atomic force microscopy and DNA binding competition assay, we showed the greatest preference of the BRCA1 DNA-binding domain to cruciform structure, followed by DNA quadruplex, with the weakest affinity to double stranded B-DNA and single stranded DNA. While preference of the BRCA1 protein to cruciform structures has been reported previously, our observations demonstrated for the first time a preferential binding of the BRCA1 protein also to triplex and quadruplex DNAs, including its visualization by atomic force microscopy.

Conclusions

Our discovery highlights a direct BRCA1 protein interaction with DNA. When compared to double stranded DNA, such a strong preference of the BRCA1 protein to cruciform and quadruplex structures suggests its importance in biology and may thus shed insight into the role of these interactions in cell regulation and maintenance.

Electronic supplementary material

The online version of this article (doi:10.1186/s12867-016-0068-6) contains supplementary material, which is available to authorized users.

Differential salt-induced dissociation of the p53 protein complexes with circular and linear plasmid DNA substrates suggest involvement of a sliding mechanism

A study of the effects of salt conditions on the association and dissociation of wild type p53 with different ~3 kbp long plasmid DNA substrates (supercoiled, relaxed circular and linear, containing or lacking a specific p53 binding site, p53CON) using immunoprecipitation at magnetic beads is presented. Salt concentrations above 200 mM strongly affected association of the p53 protein to any plasmid DNA substrate. Strikingly different behavior was observed when dissociation of pre-formed p53-DNA complexes in increased salt concentrations was studied. While contribution from the p53CON to the stability of the p53-DNA complexes was detected between 100 and 170 mM KCl, p53 complexes with circular DNAs (but not linear) exhibited considerable resistance towards salt treatment for KCl concentrations as high as 2 M provided that the p53 basic C-terminal DNA binding site (CTDBS) was available for DNA binding. On the contrary, when the CTDBS was blocked by antibody used for immunoprecipitation, all p53-DNA complexes were completely dissociated from the p53 protein in KCl concentrations ≥200 mM under the same conditions. These observations suggest: (a) different ways for association and dissociation of the p53-DNA complexes in the presence of the CTDBS; and (b) a critical role for a sliding mechanism, mediated by the C-terminal domain, in the dissociation process.

DNA and RNA quadruplex-binding proteins

Four-stranded DNA structures were structurally characterized in vitro by NMR, X-ray and Circular Dichroism spectroscopy in detail. Among the different types of quadruplexes (i-Motifs, minor groove quadruplexes, G-quadruplexes, etc.), the best described are G-quadruplexes which are featured by Hoogsteen base-paring. Sequences with the potential to form quadruplexes are widely present in genome of all organisms. They are found often in repetitive sequences such as telomeric ones, and also in promoter regions and 5' non-coding sequences. Recently, many proteins with binding affinity to G-quadruplexes have been identified. One of the initially portrayed G-rich regions, the human telomeric sequence (TTAGGG)n, is recognized by many proteins which can modulate telomerase activity. Sequences with the potential to form G-quadruplexes are often located in promoter regions of various oncogenes. The NHE III1 region of the c-MYC promoter has been shown to interact with nucleolin protein as well as other G-quadruplex-binding proteins. A number of G-rich sequences are also present in promoter region of estrogen receptor alpha. In addition to DNA quadruplexes, RNA quadruplexes, which are critical in translational regulation, have also been predicted and observed. For example, the RNA quadruplex formation in telomere-repeat-containing RNA is involved in interaction with TRF2 (telomere repeat binding factor 2) and plays key role in telomere regulation. All these fundamental examples suggest the importance of quadruplex structures in cell processes and their understanding may provide better insight into aging and disease development.

Electrochemical detection of DNA binding by tumor suppressor p53 protein using osmium-labeled oligonucleotide probes and catalytic hydrogen evolution at the mercury electrode

In this paper, we present an electrochemical DNA-protein interaction assay based on a combination of protein-specific immunoprecipitation at magnetic beads (MBIP) with application of oligonucleotide (ON) probes labeled with an electroactive oxoosmium complex (Os,bipy). We show that double-stranded ONs bearing a dT20 tail labeled with Os,bipy are specifically recognized by the tumor suppressor p53 protein according to the presence or absence of a specific binding site (p53CON) in the double-stranded segment. We demonstrate the applicability of the Os,bipy-labeled probes in titration as well as competition MBIP assays to evaluate p53 relative affinity to various sequence-specific or structurally distinct unlabeled DNA substrates upon modulation of the p53-DNA binding by monoclonal antibodies used for the immunoprecipitation. To detect the p53-bound osmium-labeled probes, we took advantage of a catalytic peak yielded by Os,bipy-modified DNA at the mercury-based electrodes, allowing facile determination of subnanogram quantities of the labeled oligonucleotides. Versatility of the electrochemical MBIP technique and its general applicability in studies of any DNA-binding protein is discussed.

Preferential binding of p53 tumor suppressor to p21 promoter sites that contain inverted repeats capable of forming cruciform structure

p53 Is one of the most critical proteins involved in protecting organisms from malignancies and its gene is frequently mutated in these diseases. p53 Functions as a transcription factor and its role in the cell is mediated by sequence-specific DNA binding. Although the genome contains many p53-binding sequences, the p53 protein binds only a subset of these sequences with high affinity. One likely mechanism of how p53 binds DNA effectively underlies its ability to recognize selective local DNA structure. We analyzed the possibility of cruciform structure formation within different regions of the p21 gene promoter. p53 protein remarkably activates the transcription of p21 gene after genotoxic treatment. In silico analysis showed that p21 gene promoter contains numerous p53 target sequences, some of which have inverted repeats capable of forming cruciform structures. Using chromatin immunoprecipitation, we demonstrated that p53 protein binds preferentially to sequences that not only contain inverted repeats but also have the ability to create local cruciform structures. Gel retardation assay also revealed strong preference of the p53 protein for response element in superhelical state, with cruciform structure in the DNA sequence. Taken together, our results suggest that p53 response element's potential for cruciform structure formation could be an additional determinant in p53 DNA-binding machinery.

Preferential binding of hot spot mutant p53 proteins to supercoiled DNA in vitro and in cells

Hot spot mutant p53 (mutp53) proteins exert oncogenic gain-of-function activities. Binding of mutp53 to DNA is assumed to be involved in mutp53-mediated repression or activation of several mutp53 target genes. To investigate the importance of DNA topology on mutp53-DNA recognition in vitro and in cells, we analyzed the interaction of seven hot spot mutp53 proteins with topologically different DNA substrates (supercoiled, linear and relaxed) containing and/or lacking mutp53 binding sites (mutp53BS) using a variety of electrophoresis and immunoprecipitation based techniques. All seven hot spot mutp53 proteins (R175H, G245S, R248W, R249S, R273C, R273H and R282W) were found to have retained the ability of wild-type p53 to preferentially bind circular DNA at native negative superhelix density, while linear or relaxed circular DNA was a poor substrate. The preference of mutp53 proteins for supercoiled DNA (supercoil-selective binding) was further substantiated by competition experiments with linear DNA or relaxed DNA in vitro and ex vivo. Using chromatin immunoprecipitation, the preferential binding of mutp53 to a sc mutp53BS was detected also in cells. Furthermore, we have shown by luciferase reporter assay that the DNA topology influences p53 regulation of BAX and MSP/MST1 promoters. Possible modes of mutp53 binding to topologically constrained DNA substrates and their biological consequences are discussed.

Preferential binding of IFI16 protein to cruciform structure and superhelical DNA

Interferon (IFN)-inducible HIN-200 proteins play an important role in transcriptional regulation linked to cell cycle control, inflammation, autoimmunity and differentiation. IFI16 has been identified as a target of IFNα and γ and is a member of the HIN-200 protein family. Expression level of IFI16 is often decreased in breast cancers, implicating its role as a tumor suppressor. As a potent transcription factor, IFI16 possesses a transcriptional regulatory region, a PYD/DAPIN/PAAD region which associates with IFN response, DNA-binding domains and binding regions for tumor suppressor proteins BRCA1 and p53. It is also reported that IFI16 protein is capable of binding p53 and cMYC gene promoters. Here, we demonstrate that IFI16 protein binds strongly to negatively superhelical plasmid DNA at a native superhelix density, as evidenced by electrophoretic retardation of supercoiled (sc) DNA in agarose gels. Binding of IFI16 to supercoiled DNA results in the appearance of one or more retarded DNA bands on the gels. After removal of IFI16, the original mobility of the scDNA is recovered. By contrast, IFI16 protein binds very weakly to the same DNA in linear state. Using short oligonucleotide targets, we also detect a strong preference for IFI16 binding to cruciform DNA structure compared to linear DNA topology. Hence, this novel DNA-binding property of IFI16 protein to scDNA and cruciform structures may play critical roles in its tumor suppressor function.

Cruciform structures are a common DNA feature important for regulating biological processes

DNA cruciforms play an important role in the regulation of natural processes involving DNA. These structures are formed by inverted repeats, and their stability is enhanced by DNA supercoiling. Cruciform structures are fundamentally important for a wide range of biological processes, including replication, regulation of gene expression, nucleosome structure and recombination. They also have been implicated in the evolution and development of diseases including cancer, Werner's syndrome and others.

Cruciform structures are targets for many architectural and regulatory proteins, such as histones H1 and H5, topoisomerase IIβ, HMG proteins, HU, p53, the proto-oncogene protein DEK and others. A number of DNA-binding proteins, such as the HMGB-box family members, Rad54, BRCA1 protein, as well as PARP-1 polymerase, possess weak sequence specific DNA binding yet bind preferentially to cruciform structures. Some of these proteins are, in fact, capable of inducing the formation of cruciform structures upon DNA binding. In this article, we review the protein families that are involved in interacting with and regulating cruciform structures, including (a) the junction-resolving enzymes, (b) DNA repair proteins and transcription factors, (c) proteins involved in replication and (d) chromatin-associated proteins. The prevalence of cruciform structures and their roles in protein interactions, epigenetic regulation and the maintenance of cell homeostasis are also discussed.