Understanding the immutability of restriction enzymes: crystal structure of BglII and its DNA substrate at 1.5 A resolution

Restriction endonucleases are remarkably resilient to alterations in their DNA binding specificity. To understand the basis of this immutability, we have determined the crystal structure of endonuclease BglII bound to its recognition sequence (AGATCT), at 1. 5 A resolution. We compare the structure of BglII to endonuclease BamHI, which recognizes a closely related DNA site (GGATCC). We show that both enzymes share a similar alpha/beta core, but in BglII, the core is augmented by a beta-sandwich domain that encircles the DNA to provide extra specificity. Remarkably, the DNA is contorted differently in the two structures, leading to different protein-DNA contacts for even the common base pairs. Furthermore, the BglII active site contains a glutamine in place of the glutamate at the general base position in BamHI, and only a single metal is found coordinated to the putative nucleophilic water and the phosphate oxygens. This surprising diversity in structures shows that different strategies can be successful in achieving site-specific recognition and catalysis in restriction endonucleases.     

The reactions of the EcoRi and other restriction endonucleases.

The reaction of the EcoRI restriction endonuclease was studied with both the plasmid pMB9 and DNA from bacteriophage lambda as the substrates. With both circular and linear DNA molecules, the only reaction catalysed by the EcoRI restriction endonuclease was the hydrolysis of the phosphodiester bond within one strand of the recognition site on the DNA duplex. The cleavage of both strands of the duplex was achieved only after two independent reactions, each involving a single-strand scission. The reactivity of the enzyme for single-strand scissions was the same for both the first and the second cleavage within its recognition site. No differences were observed between the mechanism of action on supercoiled and linear DNA substrates. Other restriction endonucleases were tested against plasmid pMB9. The HindIII restriction endonuclease cleaved DNA in the same manner as the EcoRI enzyme. However, in contrast with EcoRI, the Sa/I and the BamHI restriction endonucleases appeared to cleave both strands of the DNA duplex almost simultaneously. The function of symmetrical DNA sequences and the conformation of the DNA involved in these DNA--protein interactions are discussed in the light of these observations. The fact that the same reactions were observed on both supercoiled and linear DNA substrates implies that these interactions do not involve the unwinding of the duplex before catalysis.     

The effect of Z-conformation on enzymatic activity of restriction endonucleases

Recombinant plasmid pGC20 containing (GC)9-insert into SmaI site of pUC19 has been used to study the inhibition of cleavage by six restriction endonucleases; KpnI, SacI, EcoRI and also BamHI, XbaI and SalI, due to Z-DNA formation in negatively supercoiled plasmid. The recognition sites of these enzymes were located at different distances on both sides of the (CG)10-sequence. It was shown that the inhibition of the cleavage by KpnI, SacI and EcoRI was decreased in this series as fast as the distance between recognition site and B-Z junction was increased, and no inhibition of cleavage by EcoRI was found. However, such a correlation was not found in the series of BamHI, XbaI and SalI. In contrast with EcoRI the cleavage by SalI was inhibited completely. These results indicate the difference for "sensitivity" of restriction endonucleases to the structural perturbations of DNA associated with B-Z junctions. It seems to depend on features of the enzyme-substrate interaction mechanisms and also on recognition and flanking sequences of DNA. Consequently, experiments with the inhibition of the cleavage by any enzyme can not help to determine the dimension of the region of DNA with altered structure.     

Immobilization of restrictases

The search for optimal variants of restriction endonucleases immobilization was begun recently. For some enzymes immobilization was successful due to the presence of covalent bonds on CNBr-sepharose (EcoRI, BamHI, HindIII, TaqI, PaeI, SalI, PvuII). For the enzymes EcoRI, BamHI and HindIII it was due to hydrophobic interaction with triethyl-agarose (triethyl-triphenylmethane). The high yield (up to 80%) of enzymatic activity has been obtained for small number of restriction endonucleases. In the experiments of several amino acid residues modification and immobilization of restriction endonucleases the participation of lysine, arginine, glutamic acid and SH- or S-S-groups in the catalysis and (or) binding of these enzymes with DNA has been shown. The restriction endonucleases immobilization experiments and research of enzymes active centre enrich each other and are very interesting for their use in molecular biology and deepening our knowledge of protein-nucleic interactions.     

Trapping DNA-protein binding reactions with neutral osmolytes for the analysis by gel mobility shift and self-cleavage assays

We take advantage of our previous observation that neutral osmolytes can strongly slow down the rate of DNA–protein complex dissociation to develop a method that uses osmotic stress to ‘freeze’ mixtures of DNA–protein complexes and prevent further reaction enabling analysis of the products. We apply this approach to the gel mobility shift assay and use it to modify a self-cleavage assay that uses the nuclease activity of the restriction endonucleases to measure sensitively their specific binding to DNA. At sufficiently high concentrations of neutral osmolytes the cleavage reaction can be triggered at only those DNA fragments with initially bound enzyme. The self-cleavage assay allows measurement of binding equilibrium and kinetics directly in solution avoiding the intrinsic problems of gel mobility shift and filter binding assays while providing the same sensitivity level. Here we compare the self-cleavage and gel mobility shift assays applied to the DNA binding of EcoRI and BamHI restriction endonucleases. Initial results indicate that BamHI dissociation from its specific DNA sequence is strongly linked to water activity with the half-life time of the specific complex increasing ~20-fold from 0 to 1 osmolal betaine.     

Restriction enzyme analysis of Bacillus subtilis bacteriophage phi 105 DNA

The recognition sites on phi 105 DNA for the restriction endonucleases EcoRI, Bg/II, SmaI, KpnI, SstI, SalI, XhoI, NcoI, PstI, HindIII, ClaI, EcoRV and MluI have been mapped. The sites for EcoRI are shown to be different from those published earlier. The DNA from phi 105 contains no recognition sites for the endonucleases BamHI and XbaI.     

Removing water from an EcoRI-noncognate DNA complex with osmotic stress.

We recently showed that a nonspecific complex of the restriction nuclease EcoRI with poly (dI-dC) sequesters significantly more water at the protein-DNA interface than the complex with the specific recognition sequence. The nonspecific complex seems to retain almost a full hydration layer at the interface. We now find that at low osmotic pressures a complex of the restriction nuclease EcoRI with a DNA sequence that differs by only one base pair from the recognition site (a 'star' sequence) sequesters about 70 waters more than the specific one, a value virtually indistinguishable from nonspecific DNA. Unlike complexes with oligo (dI-dC) or with a sequence that differs by two base pairs from the recognition sequence, however, much of the water in the 'star' sequence complex is removed at high osmotic pressures. The energy of removing this water can be calculated simply from the osmotic pressure work done on the complex. The ability to measure not only the changes in water sequestered by DNA-protein complexes for different sequences, but also the work necessary to remove this water is a potentially powerful new tool for coupling inferred structural changes and thermodynamics.     

On the Divalent Metal Ion Dependence of DNA Cleavage by Restriction Endonucleases of the EcoRI Family

Restriction endonucleases of the PD…D/EXK family need Mg²⁺ for DNA cleavage. Whereas Mg²⁺ (or Mn²⁺) promotes catalysis, Ca²⁺ (without Mg²⁺) only supports DNA binding. The role of Mg²⁺ in DNA cleavage by restriction endonucleases has elicited many hypotheses, differing mainly in the number of Mg²⁺ involved in catalysis. To address this problem, we measured the Mg²⁺ and Mn²⁺ concentration dependence of DNA cleavage by BamHI, BglII, Cfr10I, EcoRI, EcoRII (catalytic domain), MboI, NgoMIV, PspGI, and SsoII, which were reported in co-crystal structure analyses to bind one (BglII and EcoRI) or two (BamHI and NgoMIV) Me²⁺ per active site. DNA cleavage experiments were carried out at various Mg²⁺ and Mn²⁺ concentrations at constant ionic strength. All enzymes show a qualitatively similar Mg²⁺ and Mn²⁺ concentration dependence. In general, the Mg²⁺ concentration optimum (between ∼ 1 and 10 mM) is higher than the Mn²⁺ concentration optimum (between ∼ 0.1 and 1 mM). At still higher Mg²⁺ or Mn²⁺ concentrations, the activities of all enzymes tested are reduced but can be reactivated by Ca²⁺. Based on these results, we propose that one Mg²⁺ or Mn²⁺ is critical for restriction enzyme activation, and binding of a second Me²⁺ plays a role in modulating the activity. Steady-state kinetics carried out with EcoRI and BamHI suggest that binding of a second Mg²⁺ or Mn²⁺ mainly leads to an increase in K_m, such that the inhibitory effect of excess Mg²⁺ or Mn²⁺ can be overcome by increasing the substrate concentration. Our conclusions are supported by molecular dynamics simulations and are consistent with the structural observations of both one and two Me²⁺ binding to these enzymes.     

Thermodynamic and kinetic basis for the relaxed DNA sequence specificity of "promiscuous" mutant EcoRI endonucleases

Promiscuous mutant EcoRI endonucleases produce lethal to sublethal effects because they cleave Escherichia coli DNA despite the presence of the EcoRI methylase. Three promiscuous mutant forms, Ala138Thr, Glu192Lys and His114Tyr, have been characterized with respect to their binding affinities and first-order cleavage rate constants towards the three classes of DNA sites: specific, miscognate (EcoRI*) and non-specific. We have made the unanticipated and counterintuitive observations that the mutant restriction endonucleases that exhibit relaxed specificity in vivo nevertheless bind more tightly than the wild-type enzyme to the specific recognition sequence in vitro, and show even greater preference for binding to the cognate GAATTC site over miscognate sites. Binding preference for EcoRI* over non-specific DNA is also improved. The first-order cleavage rate constants of the mutant enzymes are normal for the cognate site GAATTC, but are greater than those of the wild-type enzyme at EcoRI* sites. Thus, the mutant enzymes use two mechanisms to partially bypass the multiple fail-safe mechanisms that protect against cleavage of genomic DNA in cells carrying the wild-type EcoRI restriction-modification system: (a) binding to EcoRI* sites is more probable than for wild-type enzyme because non-specific DNA is less effective as a competitive inhibitor; (b) the combination of increased affinity and elevated cleavage rate constants at EcoRI* sites makes double-strand cleavage of these sites a more probable outcome than it is for the wild-type enzyme. Semi-quantitative estimates of rates of EcoRI* site cleavage in vivo, predicted using the binding and cleavage constants measured in vitro, are in accord with the observed lethal phenotypes associated with the three mutations.     

The effects of substituted pyrimidines in DNAs on cleavage by sequence-specific endonucleases.

The rates of cleavage of DNAs containing substituents at position 5 of thymine or cytosine have been measured for a variety of sequence-specific endonucleases, so as to determine which features in the DNA sequence are being probed. Phage phi e DNA fully substituted with 5-hydroxymethyluracil is cleaved more slowly by enzymes whose recognition sequences contain A-T base pairs than are DNAs containing thymine, but both types of DNA are cleaved at similar rates by enzymes recognizing sequences composed only of G-C base pairs. Phage PBS2 DNA with uracil completely substituted for thymine is cleaved slowly by several enzymes which recognize sequences containing A-T base pairs (endonucleases Hpa I, HindII, and HindIII), while the rates of cleavage by other enzymes (endonucleases EcoRI and BamHI) are not affected. Phage lambda- and P22 DNAs containing 5-bromouracil are cleaved more slowly by several enzymes (endonucleases HindIII, Hpa I, BamHI) than are thymine-containing DNAs. Enzymes that recognize sequence isomers with the composition G:C:2A:2T (endonucleases EcoRI, Hpa I, HindIII) are not equally affected by substitution at position 5 of thymine, suggesting that they differ in their contacts with A-T base pairs. DNA containing glucosylated 5-hydroxymethylcytosine in place of cytosine is resistant to cleavage by all the endonucleases examined.     

The highly homologous isoschizomers RsrI endonuclease and EcoRI endonuclease do not recognize their target sequence identically

Using a series of decadeoxyribonucleotides containing base analogues as substrates we measured the steady-state kinetic parameters for the reaction catalyzed by RsrI endonuclease and compared the results to those with its isoschizomer EcoRI. The kinetics of RsrI cleavage are affected by each substitution, with the effects being generally more deleterious than with EcoRI, as shown by the greater reduction in the specificity constant kcat/KM. The magnitudes of the effects of several substitutions are consistent with the formation of direct enzyme-nucleobase contacts at the indicated positions. With substrates containing 2-amino-purine or 2,6-diaminopurine at the central adenine or uracil at the outermost thymine in the recognition sequence, cleavage by RsrI was very slow, less than one-tenth the rate of the corresponding EcoRI-catalyzed reaction. The lower tolerance of RsrI endonuclease for functional group changes in its recognition site may reflect differences in the mechanisms of DNA recognition by the two enzymes. Although RsrI and EcoRI endonucleases bind with similar affinities to specific and nonspecific DNA sequences and appear to introduce similar structural distortions in DNA upon binding, the use of substrate analogues reveals significant differences at the level of catalysis in the mechanisms by which these two endonucleases recognize the duplex sequence GAATTC.     

BglII and MunI: what a difference a base makes

Restriction endonucleases are resilient to alterations in their DNA-binding specificities. Structures of the BglII and MunI endonucleases bound to their palindromic DNA sites, which differ by only their outer base pairs from the recognition sequences of BamHI and EcoRI, respectively, have recently been determined. A comparison of these complexes reveals surprising differences and similarities in structure, and provides a basis for understanding the immutability of restriction endonucleases.     

Cleavage by restriction enzymes of DNA modified with the antitumour drug cis-diamminedichloroplatinum(II).

The effect of binding of an antitumour drug cis-diamminedichloroplatinum(II) (cis-[Pt(NH3)2Cl2]) to DNA on cutting effectiveness of BamHI, EcoRI, and SalI restriction endonucleases was quantitatively determined. The platinum complex inhibits the cleavage of plasmid pHC624 DNA linearized by BglI restrictase. From the present results we conclude that the yield of restriction endonuclease cleavage is also lowered if the platinum complex is bound outside the recognition DNA sequence of these enzymes. We propose that the origin of platinum adducts on DNA outside the recognition sequence can decrease the yield of restriction enzyme cleavage via inducing a conformational perturbation in the recognition DNA sequence of these enzymes and also via inhibition of the linear diffusion of these enzymes on DNA.     

A comparative study of G+C-rich satellite components of sheep and goat DNA.

Restriction fragment patterns of G+C-rich satellites of sheep and goat DNA were compared. The 1,712 g/cm3 satellites of both species appear homologous, consisting of repeats 760 base pairs long and showing coincidence of position of primary+ EcoRI, BamHI and most BspRI restriction target sites. The EcoRI and BamHI endonucleases produce mostly monomers of the repeating unit, while oligomers prevail in the A1uI and Bg1II digests. Species-specific differences in the frequency, position and mode of distribution of secondary+ restriction target sites for EcoRI, Bg1II and A1uI were observed. Unlike the 1,712 g/cm3 satellites, the 1,723 g/cm3 component of sheep DNA and the 1,719 g/cm3 material from goat DNA appear species--specific, since no homologous material could ever be detected in the DNA of the other species.     

Cofactor requirements of BamHI mutant endonuclease E77K and its suppressor mutants

A mutant BamHI endonuclease, E77K, belongs to a class of catalytic mutants that bind DNA efficiently but cleave DNA at a rate more than 10(3)-fold lower than that of the wild-type enzyme (S. Y. Xu and I. Schildkraut, J. Biol. Chem. 266:4425-4429, 1991). The preferred cofactor for the wild-type BamHI is Mg2+. BamHI is 10-fold less active with Mn2+ as the cofactor. In contrast, the E77K variant displays an increased activity when Mn2+ is substituted for Mg2+ in the reaction buffer. Mutations that partially suppress the E77K mutation were isolated by using an Escherichia coli indicator strain containing the dinD::lacZ fusion. These pseudorevertant endonucleases induce E. coli SOS response (as evidenced by blue colony formation) and thus presumably nick or cleave chromosomal DNA in vivo. Consistent with the in vivo result, the pseudorevertant endonucleases in the crude cell extract display site-specific partial DNA cleavage activity. DNA sequencing revealed two unique suppressing mutations that were located within two amino acid residues of the original mutation. Both pseudorevertant proteins were purified and shown to increase specific activity at least 50-fold. Like the wild-type enzyme, both pseudorevertant endonucleases prefer Mg2+ as the cofactor. Thus, the second-site mutation not only restores partial cleavage activity but also suppresses the metal preference as well. These results suggest that the Glu-77 residue may play a role in metal ion binding or in enzyme activation (allosteric transition) following sequence-specific recognition.     

Differences between EcoRI nonspecific and "star" sequence complexes revealed by osmotic stress

The binding of the restriction endonuclease EcoRI to DNA is exceptionally specific. Even a single basepair change ("star" sequence) from the recognition sequence, GAATTC, decreases the binding free energy of EcoRI to values nearly indistinguishable from nonspecific binding. The difference in the number of waters sequestered by the protein-DNA complexes of the "star" sequences TAATTC and CAATTC and by the specific sequence complex determined from the dependence of binding free energy on water activity is also practically indistinguishable at low osmotic pressures from the 110 water molecules sequestered by nonspecific sequence complexes. Novel measurements of the dissociation rates of noncognate sequence complexes and competition equilibrium show that sequestered water can be removed from "star" sequence complexes by high osmotic pressure, but not from a nonspecific complex. By 5 Osm, the TAATTC "star" sequence complex has lost almost 90 of the approximately 110 waters initially present. It is more difficult to remove water from the CAATTC "star" sequence complex. The sequence dependence of water loss correlates with the known sequence dependence of "star" cleavage activity.     

Effect of site-specific endonuclease digestion on the thyP3 gene of bacteriophage phi 3T and the thyP11 gene of bacteriophage rho11.

phi 3T and rho11 are closely related bacteriophages of Bacillus subtilis which can "convent" thymine auxotrophs to thymine prototrophs upon infection or transfection. The effect of endonuclease digestion on the ability of both bacteriophage and prophage DNA from phi eT and rho11 to transform for thymine prototrophy was determined. All of the endonucleases tested: BamHI, Bg/II, BsuRI, EcoRI, HindII+ III, and HpaII reduced the efficiency of thyP transformation to an equal extent in prophage and bacteriophage DNA. Only HpaII completely abolished thyP transformation. The reduction in transformation with BamHI, Bg/II, BsuRI, EcoRI, and HpaII fragments is size related. The thyP transforming fragments generated by these endonucleases are potentially clonable.     

Structure-based sequence alignment of type-II restriction endonucleases

The type-II restriction endonucleases generally do not share appreciable amino acid sequence homology. The crystal structures of restriction endonucleases EcoRI and BamHI have shown these enzymes to possess striking 3D-structural resemblance, i.e., they have a similar overall fold and similar active sites, though they possess <23% sequence identity. Structural superimposition of EcoRI, BamHI, EcoRV, and PvuII based on active site residues led to sequence alignments which showed nine possible sequence motifs. EcoRV and PvuII show a more similar pattern than EcoRI and BamHI suggesting that they belong to a different subgroup. The motifs are characterized by charged and/or hydrophobic residues. From other studies on the structure of these endonucleases, three of the motifs could be implicated in DNA binding, three in forming the active site and one in dimer formation. However, the motifs were not identifiable by regular sequence alignment methods. It is found that motif IX in BamHI is formed by reverse sequence order and the motif IX in PvuII is formed from the symmetry related monomer of the dimer. The inter-motif distance is also quite different in these cases. Of the nine motifs, motif III has been earlier identified as containing the PD motif involving one of the active site residues. These motifs were used in a modified profile analysis procedure to identify similar regions in eight other endonuclease sequences for which structures are not known.