Specificity from the synapsis of DNA elements by the Sfi I endonuclease.

The synapsis of DNA sites is a prerequisite for the reactions of many proteins that act at specific DNA sequences. The requirement for synapsis was investigated by analysing the reactions of Sfi I, a tetrameric restriction enzyme that cleaves DNA only after interacting with two recognition sites. In the presence of Mg2+, oligonucleotide duplexes with the cognate recognition sequence were cleaved rapidly, with cooperative kinetics, while non-cognate duplexes were not cleaved. In the absence of Mg2+, the primary complex formed by Sfi I with cognate DNA contained two duplexes synapsed by the tetramer: a secondary complex containing one duplex was seen only at elevated Sfi I concentrations. In contrast, the principal complex with non-cognate DNA contained one duplex bound to Sfi I. Pairs of non-cognate duplexes, or one cognate and one non-cognate duplex, generally failed to form synaptic complexes. On adding Mg2+to complexes with cognate DNA, cleavage occurred much more rapidly in the synaptic complex than in the secondary complex. DNA synapsis thus acts to enhance the specificity of Sfi I for its recognition sequence, by demanding two cognate sites for a catalytically active complex and by excluding non-cognate sites from the synaptic complex.     

Fidelity of DNA sequence recognition by the SfiI restriction endonuclease is determined by communications between its two DNA-binding sites

The SfiI restriction endonuclease is a tetramer in which two subunits form a dimeric unit that contains one DNA binding cleft and the other two subunits contain a second cleft on the opposite side of the protein. Full activity requires both clefts to be filled with its recognition sequence: SfiI has low activity when bound to one site. The ability of SfiI to cleave non-cognate sites, one base pair different from the true site, was initially tested on substrates that lacked specific sites but which contained either one or multiple non-cognate sites. No cleavage of the DNA with one non-cognate site was detected, while a small fraction of the DNA with multiple sites was nicked. The alternative sequences were, however, cleaved in both strands, albeit at low levels, when the DNA also carried either a recognition site for SfiI or the termini generated by SfiI. Further tests employed a mutant of SfiI, altered at the dimer interface, which was known to be more active than wild-type SfiI when bound to a single site. This mutant similarly failed to cleave DNA with one non-cognate site, but cleaved the substrates with multiple non-cognate sites more readily than did the native enzyme. To cleave additional sites, SfiI thus needs to interact concurrently with either two non-cognate sites or one non-cognate and one cognate site (or the termini thereof), yet this arrangement is still restrained from cleaving the alternative site unless the communication pathway between the two DNA-binding clefts is disrupted.     

DNA strand arrangement within the SfiI-DNA complex: atomic force microscopy analysis

The SfiI restriction enzyme binds to DNA as a tetramer holding two usually distant DNA recognition sites together before cleavage of the four DNA strands. To elucidate structural properties of the SfiI-DNA complex, atomic force microscopy (AFM) imaging of the complexes under noncleaving conditions (Ca2+ instead of Mg2+ in the reaction buffer) was performed. Intramolecular complexes formed by protein interaction between two binding sites in one DNA molecule (cis interaction) as well as complexes formed by the interaction of two sites in different molecules (trans interaction) were analyzed. Complexes were identified unambiguously by the presence of a tall spherical blob at the DNA intersections. To characterize the path of DNA within the complex, the angles between the DNA helices in the proximity of the complex were systematically analyzed. All the data show clear-cut bimodal distributions centered around peak values corresponding to 60 degrees and 120 degrees. To unambiguously distinguish between the crossed and bent models for the DNA orientation within the complex, DNA molecules with different arm lengths flanking the SfiI binding site were designed. The analysis of the AFM images for complexes of this type led to the conclusion that the DNA recognition sites within the complex are crossed. The angles of 60 degrees or 120 degrees between the DNA helices correspond to a complex in which one of the helices is flipped with respect to the orientation of the other. Complexes formed by five different recognition sequences (5'-GGCCNNNNNGGCC-3'), with different central base pairs, were also analyzed. Our results showed that complexes containing the two possible orientations of the helices were formed almost equally. This suggests no preferential orientation of the DNA cognate site within the complex, suggesting that the central part of the DNA binding site does not form strong sequence specific contacts with the protein.     

A view of consecutive binding events from structures of tetrameric endonuclease SfiI bound to DNA

Many reactions in cells proceed via the sequestration of two DNA molecules in a synaptic complex. SfiI is a member of a growing family of restriction enzymes that can bind and cleave two DNA sites simultaneously. We present here the structures of tetrameric SfiI in complex with cognate DNA. The structures reveal two different binding states of SfiI: one with both DNA-binding sites fully occupied and the other with fully and partially occupied sites. These two states provide details on how SfiI recognizes and cleaves its target DNA sites, and gives insight into sequential binding events. The SfiI recognition sequence (GGCCNNNN[downward arrow]NGGCC) is a subset of the recognition sequence of BglI (GCCNNNN[downward arrow]NGGC), and both enzymes cleave their target DNAs to leave 3-base 3' overhangs. We show that even though SfiI is a tetramer and BglI is a dimer, and there is little sequence similarity between the two enzymes, their modes of DNA recognition are unusually similar.     

A switch in the mechanism of communication between the two DNA-binding sites in the SfiI restriction endonuclease

While many Type II restriction enzymes are dimers with a single DNA-binding cleft between the subunits, SfiI is a tetramer of identical subunits. Two of its subunits (a dimeric unit) create one DNA-binding cleft, and the other two create a second cleft on the opposite side of the protein. The two clefts bind specific DNA cooperatively to give a complex of SfiI with two recognition sites. This complex is responsible for essentially all of the DNA-cleavage reactions by SfiI: virtually none is due to the complex with one site. The communication between the DNA-binding clefts was examined by disrupting one of the very few polar interactions in the otherwise hydrophobic interface between the dimeric units: a tyrosine hydroxyl was removed by mutation to phenylalanine. The mutant protein remained tetrameric in solution and could bind two DNA sites. But instead of being activated by binding two sites, like wild-type SfiI, it showed maximal activity when bound to a single site and had a lower activity when bound to two sites. This interaction across the dimer interface thus enforces in wild-type SfiI a cooperative transition between inactive and active states in both dimers, but without this interaction as in the mutant protein, a single dimer can undergo the transition to give a stable intermediate with one inactive dimer and one active dimer.     

DNA restriction dependent on two recognition sites: activities of the SfiI restriction-modification system in Escherichia coli

In contrast to many type II restriction enzymes, dimeric proteins that cleave DNA at individual recognition sites 4-6 bp long, the SfiI endonuclease is a tetrameric protein that binds to two copies of an elongated sequence before cutting the DNA at both sites. The mode of action of the SfiI endonuclease thus seems more appropriate for DNA rearrangements than for restriction. To elucidate its biological function, strains of Escherichia coli expressing the SfiI restriction-modification system were transformed with plasmids carrying SfiI sites. The SfiI system often failed to restrict the survival of a plasmid with one SfiI site, but plasmids with two or more sites were restricted efficiently. Plasmids containing methylated SfI sites were not restricted. No rearrangements of the plasmids carrying SfiI sites were detected among the transformants. Hence, provided the target DNA contains at least two recognition sites, SfiI displays all of the hallmarks of a restriction-modification system as opposed to a recombination system in E. coli cells. The properties of the system in vivo match those of the enzyme in vitro. For both restriction in vivo and DNA cleavage in vitro, SfiI operates best with two recognition sites on the same DNA.     

Probing Interactions within the synaptic DNA-SfiI complex by AFM force spectroscopy

SfiI belongs to a family of restriction enzymes that function as tetramers, binding two recognition regions for the DNA cleavage reaction. The SfiI protein is an attractive and convenient model for studying synaptic complexes between DNA and proteins capable of site-specific binding. The enzymatic action of SfiI has been very well characterized. However, the properties of the complex before the cleavage reaction are not clear. We used single-molecule force spectroscopy to analyze the strength of interactions within the SfiI-DNA complex. In these experiments, the stability of the synaptic complex formed by the enzyme and two DNA duplexes was probed in a series of approach-retraction cycles. In order to do this, one duplex was tethered to the surface and the other was tethered to the probe. The complex was formed by the protein present in the solution. An alternative setup, in which the protein was anchored to the surface, allowed us to probe the stability of the complex formed with only one duplex in the approach-retraction experiments, with the duplex immobilized at the probe tip. Both types of complexes are characterized by similar rupture forces. The stability of the complex was determined by measuring the dependence of rupture forces on force loading rates (dynamic force spectroscopy) and the results suggest that the dissociation reaction of the SfiI-DNA complex has a single energy barrier along the dissociation path. Dynamic force spectroscopy was instrumental in revealing the role of the 5 bp spacer region within the palindromic recognition site on DNA-SfiI in the stability of the complex. The data show that, although the change of non-specific sequence does not alter the position of the activation barrier, it changes values of the off rates significantly.     

Mechanism of DNA Recognition by the Restriction Enzyme EcoRV

EcoRV, a restriction enzyme in Escherichia coli, destroys invading foreign DNA by cleaving it at the center step of a GATATC sequence. In the EcoRV-cognate DNA crystallographic complex, a sharp kink of 50° has been found at the center base-pair step (TA). Here, we examine the interplay between the intrinsic propensity of the cognate sequence to kink and the induction by the enzyme by performing all-atom molecular dynamics simulations of EcoRV unbound and interacting with three DNA sequences: the cognate sequence, GATATC (TA); the non-cognate sequence, GAATTC (AT); and with the cognate sequence methylated on the first adenine GA_CH3TATC (TA-CH₃). In the unbound EcoRV, the cleft between the two C-terminal subdomains is found to be open. Binding to AT narrows the cleft and forms a partially bound state. However, the intrinsic bending propensity of AT is insufficient to allow tight binding. In contrast, the cognate TA sequence is easier to bend, allowing specific, high-occupancy hydrogen bonds to form in the complex. The absence of cleavage for this methylated sequence is found to arise from the loss of specific hydrogen bonds between the first adenine of the recognition sequence and Asn185. On the basis of the results, we suggest a three-step recognition mechanism. In the first step, EcoRV, in an open conformation, binds to the DNA at a random sequence and slides along it. In the second step, when the two outer base pairs, GAxxTC, are recognized, the R loops of the protein become more ordered, forming strong hydrogen-bonding interactions, resulting in a partially bound EcoRV-DNA complex. In the third step, the flexibility of the center base pair is probed, and in the case of the full cognate sequence the DNA bends, the complex strengthens and the protein and DNA interact more closely, allowing cleavage.     

Structure and stability of the consecutive stereoregulated chiral phosphorothioate DNA duplex

The duplex structures of the stereoregulated phosphorothioate DNAs, [R(p),R(p)]- and [S(p),S(p)]-[d(GC(ps)T(ps)ACG)] (ps, phosphorothioate; PS-DNA), with their complementary RNA have been investigated by combined use of (1)H NMR and restrained molecular dynamics calculation. Compared to those obtained for the unmodified duplex structures (PO-DNA.RNA), the NOE cross-peak intensities are virtually identical for the PS-DNA.RNA hybrid duplexes. The structural analysis on the basis of the NOE restraints reveals that all of the three DNA.RNA duplexes take a A-form conformation and that there is no significant difference in the base stacking for the DNA.RNA hybrid duplexes. On the other hand, the NOE cross-peak intensities of the protons around the central T(ps)A step of the PS-DNA.DNA duplexes are apparently different from those of PO-DNA. DNA. The chemical shifts of H8/6 and H1' at the T(ps)A step are also largely different among PS-DNA.DNAs and PO-DNA.DNA, suggesting that the DNA.DNA structure is readily changed by the introduction of the phosphorothioate groups to the central T(p)A step. The structure calculations indicate that all of these DNA.DNA duplexes are B-form although there exist some small differences in helical parameters between the [R(p),R(p)]- and [S(p),S(p)]PS-DNA.DNA duplexes. The melting temperatures (T(m)) were determined for all of the duplexes by plotting the chemical shift change of isolated peaks as a function of temperature. For the PS-DNA.RNA hybrid duplexes, the [S(p),S(p)] isomer is less stable than the [R(p),R(p)] isomer while this trend is reversed for the PS-DNA.DNA duplexes. Consequently, although the PS-DNA.RNA duplexes take the similar A-form structure, the duplex stability is different between PS-DNA.RNA duplexes. The stability of the DNA.RNA duplexes may not be governed by the A-form structure itself but by some other factors such as the hydration around the phosphorothioate backbone, although the T(m) difference of the DNA.DNA duplexes could be explained by the structural factor.     

Crystallization of complexes of EcoRV endonuclease with cognate and non-cognate DNA fragments

Complexes of the type II restriction endonuclease EcoRV with a variety of short, selfcomplementary deoxyoligonucleotides have been crystallized. The best crystals diffract to about 2.7 A resolution and consist of 1:1 complexes between endonuclease dimers and duplexes of the cognate decamer GGGATATCCC containing the hexameric RV recognition sequence GATATC. Crystals with the non-cognate DNA octamer duplexes CGAGCTCG and CGAATTCG diffract to 3.0 and 3.5 A resolution, respectively, and contain two DNA duplexes per enzyme dimer.     

Site-specific labeling of supercoiled DNA

Visualization of site-specific labels in long linear or circular DNA allows unambiguous identification of various local DNA structures. Here we describe a novel and efficient approach to site-specific DNA labeling. The restriction enzyme SfiI binds to DNA but leaves it intact in the presence of calcium and therefore may serve as a protein label of 13 bp recognition sites. Since SfiI requires simultaneous interaction with two DNA recognition sites for stable binding, this requirement is satisfied by providing an isolated recognition site in the DNA target and an additional short DNA duplex also containing the recognition site. The SfiI/DNA complexes were visualized with AFM and the specificity of the labeling was confirmed by the length measurements. Using this approach, two sites in plasmid DNA were labeled in the presence of a large excess of the helper duplex to compete with the formation of looped structures of the intramolecular synaptic complex. We show that the labeling procedure does not interfere with the superhelical tension-driven formation of alternative DNA structures such as cruciforms. The complex is relatively stable at low and high pH (pH 5 and 9) making the developed approach attractive for use at conditions requiring the pH change.     

The crystal structure of EcoRV endonuclease and of its complexes with cognate and non-cognate DNA fragments.

The crystal structure of EcoRV endonuclease has been determined at 2.5 A resolution and that of its complexes with the cognate DNA decamer GGGATATCCC (recognition sequence underlined) and the non-cognate DNA octamer CGAGCTCG at 3.0 A resolution. Two octamer duplexes of the non-cognate DNA, stacked end-to-end, are bound to the dimeric enzyme in B-DNA-like conformations. The protein--DNA interactions of this complex are prototypic for non-specific DNA binding. In contrast, only one cognate decamer duplex is bound and deviates considerably from canonical B-form DNA. Most notably, a kink of approximately 50 degrees is observed at the central TA step with a concomitant compression of the major groove. Base-specific hydrogen bonds between the enzyme and the recognition base pairs occur exclusively in the major groove. These interactions appear highly co-operative as they are all made through one short surface loop comprising residues 182-186. Numerous contacts with the sugar phosphate backbone extending beyond the recognition sequence are observed in both types of complex. However, the total surface area buried on complex formation is > 1800 A2 larger in the case of cognate DNA binding. Two acidic side chains, Asp74 and Asp90, are close to the reactive phosphodiester group in the cognate complex and most probably provide oxygen ligands for binding the essential cofactor Mg2+. An important role is also indicated for Lys92, which together with the two acidic functions appears to be conserved in the otherwise unrelated structure of EcoRI endonuclease. The structural results give new insight into the physical basis of the remarkable sequence specificity of this enzyme.     

Long-range communications between DNA sites by the dimeric restriction endonuclease SgrAI

The SgrAI endonuclease displays its maximal activity on DNA with two copies of its recognition sequence, cleaving both sites concertedly. While most restriction enzymes that act concurrently at two sites are tetramers, SgrAI is a dimer in solution. Its reaction at two cognate sites involves the association of two DNA-bound dimers. SgrAI can also bridge cognate and secondary sites, the latter being certain sequences that differ from the cognate by one base-pair. The mechanisms for cognate-cognate and cognate-secondary communications were examined for sites in the following topological relationships: in cis, on plasmids with two sites in a single DNA molecule; on catenanes containing two interlinked rings of DNA with one site in each ring; and in trans, on oligoduplexes carrying either a single site or the DNA termini generated by SgrAI. Both cognate-cognate and cognate-secondary interactions occur through 3-D space and not by 1-D tracking along the DNA. Both sorts of communication arise more readily when the sites are tethered to each other, either in cis on the same molecule of DNA or by the interlinking of catenane rings, than when released from the tether. However, the dimer bound to an oligoduplex carrying either a cognate or a secondary site could be activated to cleave that duplex by interacting with a second dimer bound to the recognition site, provided both duplexes are at least 30 base-pairs long: the second dimer could alternatively be bound to the two duplexes that correspond to the products of DNA cleavage by SgrAI.     

Alternative geometries of DNA looping: an analysis using the SfiI endonuclease

Many processes are governed by proteins that bind to separate sites in DNA and loop out the intervening DNA, but the geometries of the loops have seldom been determined. The SfiI endonuclease cleaves DNA after interacting with two recognition sites, and is a favourable system for the analysis of DNA looping. A gel-shift assay was used here to examine the binding of SfiI to a series of linear DNA molecules containing two SfiI sites separated by 109-170 base-pairs. The complexes in which SfiI trapped a loop by binding to two sites in the same DNA were separated from the complexes containing SfiI bound to separate DNA molecules. Step-wise changes in the inter-site spacing generated two forms of the looped complex with different electrophoretic mobilities. The yields of each looped complex and the complexes from intermolecular synapses all varied cyclically with the inter-site spacing, with similar periodicities ( approximately 10.5 base-pairs) but with different phases. One looped complex predominated whenever the DNA between the sites needed to be underwound in order to produce the correct helical orientation of the binding sites. The other looped complex predominated whenever the intervening DNA needed to be overwound. We conclude that the former has trapped a right-handed loop with a negative node and the latter a left-handed loop with a positive node.     

A model of EcoRII restriction endonuclease action: the active complex is most likely formed by one protein subunit and one DNA recognition site

To elucidate the mechanism of interaction of restriction endonuclease EcoRII with DNA, we studied by native gel electrophoresis the binding of this endonuclease to a set of synthetic DNA-duplexes containing the modified or canonical recognition sequence 5'-d(CCA/TGG)-3'. All binding substrate or substrate analogues tested could be divided into two major groups: (i) duplexes that, at the interaction with endonuclease EcoRII, form two types of stable complexes on native gel in the absence of Mg2+ cofactor; (ii) duplexes that form only one type of complex, observed both in the presence and absence of Mg2+. Unlike the latter, duplexes under the first group can be hydrolyzed by endonuclease. Data obtained suggest that the active complex is most likely formed by one protein subunit and one DNA recognition sequence. A model of EcoRII endonuclease action is presented.     

Characterization of DNA recognition by the human UV-damaged DNA-binding protein.

The UV-damaged DNA-binding (UV-DDB) protein is the major factor that binds DNA containing damage caused by UV radiation in mammalian cells. We have investigated the DNA recognition by this protein in vitro, using synthetic oligonucleotide duplexes and the protein purified from a HeLa cell extract. When a 32P-labeled 30-mer duplex containing the (6-4) photoproduct at a single site was used as a probe, only a single complex was detected in an electrophoretic mobility shift assay. It was demonstrated by Western blotting that both of the subunits (p48 and p127) were present in this complex. Electrophoretic mobility shift assays using various duplexes showed that the UV-DDB protein formed a specific, high affinity complex with the duplex containing an abasic site analog, in addition to the (6-4) photoproduct. By circular permutation analyses, these DNA duplexes were found to be bent at angles of 54 degrees and 57 degrees in the complexes with this protein. From the previously reported NMR studies and the fluorescence resonance energy transfer experiments in the present study, it can be concluded that the UV-DDB protein binds DNA that can be bent easily at the above angle.     

Purification and biochemical characterisation of the EcoR124 type I modification methylase.

Large scale purification of the type I modification methylase EcoR124 has been achieved from an over-expressing strain by a two step procedure using ion-exchange and heparin chromatography. Pure methylase is obtained at a yield of 30 mg per gm of cell paste. Measurements of the molecular weight and subunit stoichiometry show that the enzyme is a trimeric complex of 162 kDa consisting of two subunits of HsdM (58 kDa) and one subunit of HsdS (46 kDa). The purified enzyme can methylate a DNA fragment bearing its cognate recognition sequence. Binding of the methylase to synthetic DNA fragments containing either the EcoR124 recognition sequence GAAN6RTCG, or the recognition sequence GAAN7RTCG of the related enzyme EcoR124/3, was followed by fluorescence competition assays and by gel retardation analysis. The results show that the methylase binds to its correct sequence with an affinity of the order 10(8) M-1 forming a 1:1 complex with the DNA. The affinity for the incorrect sequence, differing by an additional base pair in the non-specific spacer, is almost two orders of magnitude lower.     

Dynamics of DNA loop capture by the SfiI restriction endonuclease on supercoiled and relaxed DNA

The SfiI endonuclease is a prototype for DNA looping. It binds two copies of its recognition sequence and, if Mg(2+) is present, cuts both concertedly. Looping was examined here on supercoiled and relaxed forms of a 5.5 kb plasmid with three SfiI sites: sites 1 and 2 were separated by 0.4 kb, and sites 2 and 3 by 2.0 kb. SfiI converted this plasmid directly to the products cut at all three sites, though DNA species cleaved at one or two sites were formed transiently during a burst phase. The burst revealed three sets of doubly cut products, corresponding to the three possible pairings of sites. The equilibrium distribution between the different loops was evaluated from the burst phases of reactions initiated by adding MgCl(2) to SfiI bound to the plasmid. The short loop was favored over the longer loops, particularly on supercoiled DNA. The relative rates for loop capture were assessed after adding SfiI to solutions containing the plasmid and MgCl(2). On both supercoiled and relaxed DNA, the rate of loop capture across 0.4 kb was only marginally faster than over 2.0 kb or 2.4 kb. The relative strengths and rates of looping were compared to computer simulations of conformational fluctuations in DNA. The simulations concurred broadly with the experimental data, though they predicted that increasing site separations should cause a shallower decline in the equilibrium constants than was observed but a slightly steeper decline in the rates for loop capture. Possible reasons for these discrepancies are discussed.     

Reactions of BglI and other type II restriction endonucleases with discontinuous recognition sites

Type II restriction enzymes generally recognize continuous sequences of 4-8 consecutive base pairs on DNA, but some recognize discontinuous sites where the specified sequence is interrupted by a defined length of nonspecific DNA. To date, a mechanism has been established for only one type II endonuclease with a discontinuous site, SfiI at GGCCNNNNNGGCC (where N is any base). In contrast to orthodox enzymes such as EcoRV, dimeric proteins that act at a single site, SfiI is a tetramer that interacts with two sites before cleaving DNA. BglI has a similar recognition sequence (GCCNNNNNGGC) to SfiI but a crystal structure like EcoRV. BglI and several other endonucleases with discontinuous sites were examined to see if they need two sites for their DNA cleavage reactions. The enzymes included some with sites containing lengthy segments of nonspecific DNA, such as XcmI (CCANNNNNNNNNTGG). In all cases, they acted at individual sites. Elongated recognition sites do not necessitate unusual reaction mechanisms. Other experiments on BglI showed that it bound to and cleaved DNA in the same manner as EcoRV, thus further delineating a distinct group of restriction enzymes with similar structures and a common reaction mechanism.     

A Chlamydomonas protein that binds single-stranded G-strand telomere DNA.

We have identified a protein in Chlamydomonas reinhardtii cell extracts that specifically binds the single-stranded (ss) Chlamydomonas G-strand telomere sequence (TTTTAGGG)n. This protein, called G-strand binding protein (GBP), binds DNA with two or more ss TTTTAGGG repeats. A single polypeptide (M(r) 34 kDa) in Chlamydomonas extracts binds (TTTTAGGG)n, and a cDNA encoding this G-strand binding protein was identified by its expression of a G-strand binding activity. The cDNA (GBP1) sequence predicts a protein product (Gbp1p) that includes two domains with extensive homology to RNA recognition motifs (RRMs) and a region rich in glycine, alanine and arginine. Antibody raised against a peptide within Gbp1p reacted with both the 34 kDa polypeptide and bound G-strand DNA-protein complexes in gel retardation assays, indicating that GBP1 encodes GBP. Unlike vertebrate heteronuclear ribonucleoproteins, GBP does not bind the cognate telomere RNA sequence UUUUAGGG in gel retardation, North-Western or competition assays. Thus, GBP is a new type of candidate telomere binding protein that binds, in vitro, to ss G-strand telomere DNA, the primer for telomerase, and has domains that have homology to RNA binding domains in other proteins.