The specificity of sty SKI, a type I restriction enzyme, implies a structure with rotational symmetry.

The type I restriction and modification (R-M) enzyme from Salmonella enterica serovar kaduna ( Sty SKI) recognises the DNA sequence 5'-CGAT(N)7GTTA, an unusual target for a type I R-M system in that it comprises two tetranucleotide components. The amino target recognition domain (TRD) of Sty SKI recognises 5'-CGAT and shows 36% amino acid identity with the carboxy TRD of Eco R124I which recognises the complementary, but degenerate, sequence 5'-RTCG. Current models predict that the amino and carboxy TRDs of the specificity subunit are in inverted orientations within a structure with 2-fold rotational symmetry. The complementary target sequences recognised by the amino TRD of Sty SKI and the carboxy TRD of Eco R124I are consistent with the predicted inverted positions of the TRDs. Amino TRDs of similar amino acid sequence have been shown to recognise the same nucleotide sequence. The similarity reported here, the first example of one between amino and carboxy TRDs, while consistent with a conserved mechanism of target recognition, offers additional flexibility in the evolution of sequence specificity by increasing the potential diversity of DNA targets for a given number of TRDs. Sty SKI identifies the first member of the IB family in Salmonella species.     

Reassortment of DNA recognition domains and the evolution of new specificities

Type I restriction enzymes comprise three subunits only one of which, the S polypeptide, dictates the specificity of the DNA sequence recognized. Recombination between two different hsdS genes, SP and SB, led to the isolation of a system, SQ, which had a different specificity from that of either parent. The finding that the nucleotide sequence recognized by SQ is a hybrid containing components from both the SP and SB target sequences suggested that DNA recognition is carried out by two separable domains within each specificity polypeptide. To test this we have made the recombinant gene of reciprocal structure and demonstrate that it encodes a polypeptide whose recognition sequence, deduced in vivo, is as predicted by this model. We also report the sequence of the SB specificity gene, so that information is now available for the five known members of this family of enzymes. All show a similar organization of conserved and variable regions. Comparisons of the predicted amino acid sequences reveal large non-conserved areas which may not even be structurally similar. This is remarkable since these different S subunits are functionally identical, except for the specificity with respect to the DNA sequence with which they interact. We discuss the correlation of the variation in polypeptide sequence with recognition specificities.     

High plasticity of multispecific DNA methyltransferases in the region carrying DNA target recognizing enzyme modules.

Multispecific cytosine C5 DNA methyltransferases (MTases) methylate more than one specific DNA target. This is due to the presence of several target recognizing domains (TRDs) in these enzymes. Such TRDs form part of a variable centre in the MTase primary sequence, which separates conserved enzyme core sequences responsible for general steps in the methylation reaction. By deleting, rearranging and exchanging several TRDs of multispecific MTases, we demonstrate their modular character; they mediate target recognition independent of a particular TRD or core sequence context. We show also that multispecific MTases can accommodate inert material of non-MTase origin within their variable region without losing their activity. The remarkable plasticity with respect to the material that can be integrated into this region suggests that the enzyme core sequences preceding or following it form separable functional domains. In spite of the documented flexibility multispecific MTases could not be endowed with novel specificities by integration of putative TRDs of monospecific MTases, pointing to differences between multi- and monospecific MTases in the way their core and TRD sequences interact.     

Reassortment of DNA recognition domains and the evolution of new specificities

Type I restriction enzymes comprise three subunits only one of which, the S polypeptide, dictates the specificity of the DNA sequence recognized. Recombination between two different hsdS genes, SP and SB, led to the isolation of a system, SQ, which had a different specificity from that of either parent. The finding that the nucleotide sequence recognized by SQ is a hybrid containing components from both the SP and SB target sequences suggested that DNA recognition is carried out by two separable domains within each specificity polypeptide. To test this we have made the recombinant gene of reciprocal structure and demonstrate that it encodes a polypeptide whose recognition sequence, deduced In vivo, is as predicted by this model. We also report the sequence of the SB specificity gene, so that information is now available for the five known members of this family of enzymes. Ali show a similar organization of conserved and variable regions. Comparisons of the predicted amino acid sequences reveal large non-conserved areas which may not even be structurally similar. This is remarkable since these different S subunits are functionally identical, except for the specificity with respect to the DNA sequence with which they interact. We discuss the correlation of the variation in polypeptide sequence with recognition specificities.     

Localization of a protein-DNA interface by random mutagenesis.

The type I restriction and modification enzymes do not possess obvious DNA-binding motifs within their target recognition domains (TRDs) of 150-180 amino acids. To identify residues involved in DNA recognition, changes were made in the amino-TRD of EcoKI by random mutagenesis. Most of the 101 substitutions affecting 79 residues had no effect on the phenotype. Changes at only seven positions caused the loss of restriction and modification activities. The seven residues identified by mutation are not randomly distributed throughout the primary sequence of the TRD: five are within the interval between residues 80 and 110. Sequence analyses have led to the suggestion that the TRDs of type I restriction enzymes include a tertiary structure similar to the TRD of the HhaI methyltransferase, and to a model for a DNA-protein interface in EcoKI. In this model, the residues within the interval identified by the five mutations are close to the protein-DNA interface. Three additional residues close to the DNA in the model were changed; each substitution impaired both activities. Proteins from twelve mutants were purified: six from mutants with partial or wild-type activity and six from mutants lacking activity. There is a strong correlation between phenotype and DNA-binding affinity, as determined by fluorescence anisotropy.     

Cytosine-specific type II DNA methyltransferases. A conserved enzyme core with variable target-recognizing domains

Comparisons of the amino acid sequences of m5C DNA methyltransferases (Mtases) from 11 prokaryotes and one eukaryote reveal a very similar organization. Among all the enzymes one can distinguish highly conserved "core" sequences and "variable" regions. The core sequences apparently mediate steps of the methylation reaction that are common to all the enzymes. The major variable region has been shown in our previous studies on multispecific phage Mtases to contain the target-recognizing domains (TRDs) of these enzymes. Here we have compared the amino acid sequences of various TRDs from phage Mtases. This has revealed the presence of both highly conserved and variable amino acids. We postulate that the conserved residues represent a "consensus" sequence defining a TRD, whereas the specificity of the TRD is determined by the variable residues. We have observed similarity between this consensus sequence and sequences in the variable region of the monospecific Mtases. We predict that the regions thus identified represent part of the TRDs of monospecific Mtases.     

Identification of a subdomain within DNA-(cytosine-C5)-methyltransferases responsible for the recognition of the 5' part of their DNA target.

In previous work on DNA-(cytosine-C5)-methyltransferases (C5-MTases), domains had been identified which are responsible for the sequence specificity of the different enzymes (target-recognizing domains, TRDs). Here we have analyzed the DNA methylation patterns of two C5-MTases containing reciprocal chimeric TRDs, consisting of the N- and C-terminal parts derived from two different parental TRDs specifying the recognition of 5'-CC(A/T)GG-3' and 5'-GCNGC-3'. Sequences recognized by these engineered MTases were non-symmetrical and degenerate, but contained at their 5' part a consensus sequence which was very similar to the 5' part of the target recognized by the parental TRD which contributed the N-terminal moiety of the chimeric TRD. The results are discussed in connection with the present understanding of the mechanism of DNA target recognition by C5-MTases. They demonstrate the possibility of designing C5-MTases with novel DNA methylation specificities.     

The amino acid sequence of the CCGG recognizing DNA methyltransferase M.BsuFI: implications for the analysis of sequence recognition by cytosine DNA methyltransferases.

The Bacillus subtilis FI DNA methyltransferase (M.BsuFI) modifies the outer cytosine of the DNA sequence CCGG, causing resistance against R.BsuFI and R.MspI restriction. The M.BsuFI gene was cloned and expressed in B.subtilis and Escherichia coli. As derived from the nucleotide sequence, the M.BsuFI protein has 409 amino acids, corresponding to a molecular mass of 46,918 daltons. Including these data we have compared the nucleotide and amino acid sequences of different CCGG recognizing enzymes. These analyses showed that M.BsuFI is highly related to two other CCGG specific methyltransferases, M.MspI and M.HpaII, which were isolated from Gram-negative bacteria. Between M.BsuFI and M.MspI the sequence similarity is particularly significant in a region, which has been postulated to contain the target recognition domains (TRDs) of cytosine-specific DNA methyltransferases. Apparently M.BsuFI and M.MspI, derived from phylogenetic distant organisms, use highly conserved structural elements for the recognition of the CCGG target sequence. In contrast the very same region of M.HpaII is quite different from those of M.BsuFI and M.MspI. We attribute this difference to the different targeting of methylation within the sequence CCGG, where M.HpaII methylates the inner, M.BsuFI/M.MspI the outer cytosine. Also the CCGG recognizing TRD of the multispecific B.subtilis phage SPR Mtase is distinct from that of the host enzyme, possibly indicating different requirements for TRDs operative in mono- and multispecific enzymes.     

Crystal structure of a putative type I restriction-modification S subunit from Mycoplasma genitalium

The crystal structure of the eubacteria Mycoplasma genitalium ORF MG438 polypeptide, determined by multiple anomalous dispersion and refined at 2.3 A resolution, reveals the organization of S subunits from the Type I restriction and modification system. The structure consists of two globular domains, with about 150 residues each, separated by a pair of 40 residue long antiparallel alpha-helices. The globular domains correspond to the variable target recognition domains (TRDs), as previously defined for S subunits on sequence analysis, while the two helices correspond to the central (CR1) and C-terminal (CR2) conserved regions, respectively. The structure of the MG438 subunit presents an overall cyclic topology with an intramolecular 2-fold axis that superimposes the N and the C-half parts, each half containing a globular domain and a conserved helix. TRDs are found to be structurally related with the small domain of the Type II N6-adenine DNA MTase TaqI. These relationships together with the structural peculiarities of MG438, in particular the presence of the intramolecular quasi-symmetry, allow the proposal of a model for S subunits recognition of their DNA targets in agreement with previous experimental results. In the crystal, two subunits of MG438 related by a crystallographic 2-fold axis present a large contact area mainly involving the symmetric interactions of a cluster of exposed hydrophobic residues. Comparison with the recently reported structure of an S subunit from the archaea Methanococcus jannaschii highlights the structural features preserved despite a sequence identity below 20%, but also reveals important differences in the globular domains and in their disposition with respect to the conserved regions.     

Exact size and organization of DNA target-recognizing domains of multispecific DNA-(cytosine-C5)-methyltransferases.

A large portion of the sequences of type II DNA-(cytosine-C5)-methyltransferases (C5-MTases) represent highly conserved blocks of amino acids. General steps in the methylation reaction performed by C5-MTases have been found to be mediated by some of these domains. C5-MTases carry, in addition at the same relative location, a region variable in size and amino acid composition, part of which is associated with the capacity of each C5-MTase to recognize its characteristic target. Individual target-recognizing domains (TRDs) for the targets CCGG (M), CC(A/T)GG (E), GGCC (H), GCNGC (F) and G(G/A/T)GC(C/A/T)C (B) could be identified in the C-terminal part of the variable region of multispecific C5-MTases. With experiments reported here, we have established the organization of the variable regions of the multispecific MTases M.SPRI, M.phi3TI, M.H2I and M.rho 11SI at the resolution of individual amino acids. These regions comprise 204, 175, 268 and 268 amino acids, respectively. All variable regions are bipartite. They contain at their N-terminal side a very similar sequence of 71 amino acids. The integrity of this sequence must be assured to provide enzyme activity. Bracketed by 6-10 'linker' amino acids, they have, depending on the enzyme studied, towards their C-terminal end ensembles of individual TRDs of 38 (M), 39 (E), 40 (H), 44 (F) and 54 (B) amino acids. TRDs of different enzymes with equal specificity have the same size. TRDs do not overlap but are either separated by linker amino acids or abut each other.     

On the structure and operation of type I DNA restriction enzymes.

Type I DNA restriction enzymes are large, molecular machines possessing DNA methyltransferase, ATPase, DNA translocase and endonuclease activities. The ATPase, DNA translocase and endonuclease activities are specified by the restriction (R) subunit of the enzyme. We demonstrate that the R subunit of the Eco KI type I restriction enzyme comprises several different functional domains. An N-terminal domain contains an amino acid motif identical with that forming the catalytic site in simple restriction endonucleases, and changes within this motif lead to a loss of nuclease activity and abolish the restriction reaction. The central part of the R subunit contains amino acid sequences characteristic of DNA helicases. We demonstrate, using limited proteolysis of this subunit, that the helicase motifs are contained in two domains. Secondary structure prediction of these domains suggests a structure that is the same as the catalytic domains of DNA helicases of known structure. The C-terminal region of the R subunit can be removed by elastase treatment leaving a large fragment, stable in the presence of ATP, which can no longer bind to the other subunits of Eco KI suggesting that this domain is required for protein assembly. Considering these results and previous models of the methyltransferase part of these enzymes, a structural and operational model of a type I DNA restriction enzyme is presented.     

DNA binding properties in vivo and target recognition domain sequence alignment analyses of wild-type and mutant RsrI [N6-adenine] DNA methyltransferases

A genetic selection method, the P22 challenge-phage assay, was used to characterize DNA binding in vivo by the prokaryotic beta class [N:6-adenine] DNA methyltransferase M.RSR:I. M.RSR:I mutants with altered binding affinities in vivo were isolated. Unlike the wild-type enzyme, a catalytically compromised mutant, M.RSR:I (L72P), demonstrated site-specific DNA binding in vivo. The L72P mutation is located near the highly conserved catalytic motif IV, DPPY (residues 65-68). A double mutant, M.RSR:I (L72P/D173A), showed less binding in vivo than did M.RSR:I (L72P). Thus, introduction of the D173A mutation deleteriously affected DNA binding. D173 is located in the putative target recognition domain (TRD) of the enzyme. Sequence alignment analyses of several beta class MTases revealed a TRD sequence element that contains the D173 residue. Phylogenetic analysis suggested that divergence in the amino acid sequences of these methyltransferases correlated with differences in their DNA target recognition sequences. Furthermore, MTases of other classes (alpha and gamma) having the same DNA recognition sequence as the beta class MTases share related regions of amino acid sequences in their TRDs.     

Two type I restriction enzymes from Salmonella species. Purification and DNA recognition sequences

We have purified the type I restriction enzymes SB and SP from Salmonella typhimurium and S. potsdam, respectively, and determined the DNA sequences that they recognize. These sequences resemble those previously determined for the type I enzymes, EcoB, EcoK and EcoA, in that the specific part of the sequence is divided into two domains by a spacer of non-specific sequence that has a fixed length for each enzyme. Two main differences from the previously determined sequences are seen. Both of the new sequences are degenerate and one of them, SB, has one trinucleotide and one pentanucleotide-specific domain rather than the trinucleotide and tetranucleotide domains seen for all of the other enzymes. The only conserved features of the recognition sequences are the adenosyl residues that are methylated in the modification reaction. For all of the enzymes these are situated ten or 11 base-pairs apart, one on each strand of the DNA. This suggests that the enzymes bind to DNA along one face of the double helix making protein-DNA interaction in two successive major grooves with most of the non-specific spacer sequence in the intervening minor groove.     

A genetic and functional analysis of the unusually large variable region in the M·Alu I DNA-(cytosine C5 )-methyltransferase

The M·AluI DNA-(cytosine C5)-methyltransferase (5mC methylase) acts on the sequence 5′-AGCT-3′. The amino acid sequences of known 5mC methylases contain ten conserved motifs, with a variable region between Motifs VIII and IX that contains one or more “target-recognizing domains” (TRDs) responsible for DNA sequence specificity. Monospecific 5mC methylases are believed to have only one TRD, while multispecific 5mC methylases have as many as five. M·AluI has the second-largest variable region of all known 5mC methylases, and sequence analysis reveals five candidate TRDs. In testing whether M·AluI is in fact monospecific it was found that AGCT methylation represents only 80–90% of the methylating activity of this enzyme, while control experiments with the enzyme M·HhaI gave no unexplained activity. Because individual TRDs can be deleted from multispecific methylases without general loss of activity, a series of insertion and deletion mutants of the M·AluI variable region were prepared. All deletions that removed more than single amino acids from the variable region caused significant loss of activity; a sensitive in vivo assay for methylase activity based on McrBC restriction suggested that the central portion of the variable region is particularly important. In some cases, multispecific methylases can accommodate a TRD from another multispecific methylase, thereby acquiring an additional specificity. When TRDs were moved from a multispecific methylase into two different locations in the variable region of M·AluI, all hybrid enzymes had greatly reduced activity and no new specificities. M·AluI thus behaves in most respects as a monospecific methylase despite the remarkable size of its variable region.     

Domain movement within a gene: a novel evolutionary mechanism for protein diversification

A protein function is carried out by a specific domain localized at a specific position. In the present study, we report that, within a gene, a specific amino acid sequence can move between a certain position and another position. This was discovered when the sequences of restriction-modification systems within the bacterial species Helicobacter pylori were compared. In the specificity subunit of Type I restriction-modification systems, DNA sequence recognition is mediated by target recognition domain 1 (TRD1) and TRD2. To our surprise, several sequences are shared by TRD1 and TRD2 of genes (alleles) at the same locus (chromosomal location); these domains appear to have moved between the two positions. The gene/protein organization can be represented as x-(TRD1)-y-x-(TRD2)-y, where x and y represent repeat sequences. Movement probably occurs by recombination at these flanking DNA repeats. In accordance with this hypothesis, recombination at these repeats also appears to decrease two TRDs into one TRD or increase these two TRDs to three TRDs (TRD1-TRD2-TRD2) and to allow TRD movement between genes even at different loci. Similar movement of domains between TRD1 and TRD2 was observed for the specificity subunit of a Type IIG restriction enzyme. Similar movement of domain between TRD1 and TRD2 was observed for Type I restriction-modification enzyme specificity genes in two more eubacterial species, Streptococcus pyogenes and Mycoplasma agalactiae. Lateral domain movements within a protein, which we have designated DOMO (domain movement), represent novel routes for the diversification of proteins.     

Conservation of complex DNA recognition domains between families of restriction enzymes

One polypeptide, designated S, confers sequence-specificity to the multisubunit type I restriction enzymes. Two families of such enzymes, K and A, include members that recognize diverse, bipartite, target sequences. The S polypeptides of the K family, while having areas of near identity, also contain two extensive regions of variable sequence. We now show that one of these, comprising the N-terminal 150 amino acids, specifies recognition of one component of the bipartite target sequence. We have determined the sequence recognized by EcoE, a member of the A family. This sequence, 5'GAG(N7)ATGC, has the trinucleotide GAG in common with EcoA and with StySB of the K family. We determined the nucleotide sequences of the S genes of EcoA and EcoE, and compared their predicted amino acid sequences with each other and with those of the five members of the K family. There is no general sequence similarity between families, but the domain of the S polypeptide of StySB, which specifies GAG, shows nearly 50 per cent identity with the amino variable region of the S polypeptides of EcoA and EcoE. A complex domain that recognizes and directs methylation of GAG is therefore common to enzymes of generally dissimilar amino acid sequence.     

Families of restriction enzymes: an analysis prompted by molecular and genetic data for type ID restriction and modification systems

Current genetic and molecular evidence places all the known type I restriction and modification systems of Escherichia coli and Salmonella enterica into one of four discrete families: type IA, IB, IC or ID. StySBLI is the founder member of the ID family. Similarities of coding sequences have identified restriction systems in E.coli and Klebsiella pneumoniae as probable members of the type ID family. We present complementation tests that confirm the allocation of EcoR9I and KpnAI to the ID family. An alignment of the amino acid sequences of the HsdS subunits of StySBLI and EcoR9I identify two variable regions, each predicted to be a target recognition domain (TRD). Consistent with two TRDs, StySBLI was shown to recognise a bipartite target sequence, but one in which the adenine residues that are the substrates for methylation are separated by only 6 bp. Implications of family relationships are discussed and evidence is presented that extends the family affiliations identified in enteric bacteria to a wide range of other genera.     

Rational engineering of type II restriction endonuclease DNA binding and cleavage specificity

The type II restriction endonucleases are indispensible tools for molecular biology. Although enzymes recognizing nearly 300 unique sequences are known, the ability to engineer enzymes to recognize any sequence of choice would be valuable. However, previous attempts to engineer new recognition specificity have met limited success. Here we report the rational engineering of multiple new type II specificities. We recently identified a family of MmeI-like type II endonucleases that have highly similar protein sequences but different recognition specificity. We identified the amino-acid positions within these enzymes that determine position specific DNA base recognition at three positions within their recognition sequences through correlations between their aligned amino-acid residues and aligned recognition sequences. We then altered the amino acids at the identified positions to those correlated with recognition of a desired new base to create enzymes that recognize and cut at predictable new DNA sequences. The enzymes so altered have similar levels of endonuclease activity compared to the wild-type enzymes. Using simple and predictable mutagenesis in this family it is now possible to create hundreds of unique new type II restriction endonuclease specificities. The findings suggest a simple mechanism for the evolution of new DNA specificity in Nature.     

DNA recognition by a new family of type I restriction enzymes: a unique relationship between two different DNA specificities.

The DNA sequences recognized by the allelic type I restriction enzymes EcoR124 and EcoR124/3 were determined. EcoR124 recognizes 5'-GAA(N6)RTCG-3' and EcoR124/3 recognizes 5'-GAA(N7)RTCG-3'. These are typical of sequences recognized by type I recognition enzymes in that they consist of two specific domains separated by a non-specific spacer sequence. For these two enzymes, the specific sequences are identical but the length of the non-specific spacer is different. The specific domains of EcoR124/3 are thus 3.4 A further apart than those of EcoR124 and rotated with respect to each other through a further 36 degrees.     

Structure-based prediction of DNA target sites by regulatory proteins

Regulatory proteins play a critical role in controlling complex spatial and temporal patterns of gene expression in higher organism, by recognizing multiple DNA sequences and regulating multiple target genes. Increasing amounts of structural data on the protein-DNA complex provides clues for the mechanism of target recognition by regulatory proteins. The analyses of the propensities of base-amino acid interactions observed in those structural data show that there is no one-to-one correspondence in the interaction, but clear preferences exist. On the other hand, the analysis of spatial distribution of amino acids around bases shows that even those amino acids with strong base preference such as Arg with G are distributed in a wide space around bases. Thus, amino acids with many different geometries can form a similar type of interaction with bases. The redundancy and structural flexibility in the interaction suggest that there are no simple rules in the sequence recognition, and its prediction is not straightforward. However, the spatial distributions of amino acids around bases indicate a possibility that the structural data can be used to derive empirical interaction potentials between amino acids and bases. Such information extracted from structural databases has been successfully used to predict amino acid sequences that fold into particular protein structures. We surmised that the structures of protein-DNA complexes could be used to predict DNA target sites for regulatory proteins, because determining DNA sequences that bind to a particular protein structure should be similar to finding amino acid sequences that fold into a particular structure. Here we demonstrate that the structural data can be used to predict DNA target sequences for regulatory proteins. Pairwise potentials that determine the interaction between bases and amino acids were empirically derived from the structural data. These potentials were then used to examine the compatibility between DNA sequences and the protein-DNA complex structure in a combinatorial "threading" procedure. We applied this strategy to the structures of protein-DNA complexes to predict DNA binding sites recognized by regulatory proteins. To test the applicability of this method in target-site prediction, we examined the effects of cognate and noncognate binding, cooperative binding, and DNA deformation on the binding specificity, and predicted binding sites in real promoters and compared with experimental data. These results show that target binding sites for several regulatory proteins are successfully predicted, and our data suggest that this method can serve as a powerful tool for predicting multiple target sites and target genes for regulatory proteins.