 1998 Oxford University Press 1084–1091 Nucleic Acids Research, 1998, Vol. 26, No. 4 Cloning and analysis of the four genes coding for Bpu10I restriction–modification enzymes Kornelijus Stankevicius, Arvydas Lubys, Albertas Timinskas, Donatas Vaitkevicius and Arvydas Janulaitis* Institute of Biotechnology, Graiciuno 8, Vilnius 2028, Lithuania Received September 19, 1997; Revised and Accepted December 22, 1997 ABSTRACT INTRODUCTION Bacterial restriction and modification (R–M) systems are traditionally divided into three classes, designated type I, II and III, on the basis of their enzyme subunit composition, cofactor requirements, substrate specificity and reaction products (1). Type II restriction endonucleases are the most prolific, encompassing over 2500 enzymes (2). By definition, these enzymes consist of homodimers of a single polypeptide subunit and they require only Mg2+ for DNA cleavage (1,3). Most type II enzymes now known recognize short sequences in DNA, which vary from 4 to 8 bp in length, maintaining a dyad axis of symmetry (palindromes), and they cleave within those sequences. However, there is a substantial number of type II restriction endonucleases, that recognize asymmetric continuous or interrupted sequences, which cleave at a short distance from those sequences on one or both sides. In contrast to palindromic recognition sequences the asymmetric recognition sequences are 5′–3′ different in each strand. The possibility that recognition is mediated by symmetrically organized homodimers, as is observed in type II ENases interacting with palindromic nucleotide sequences is therefore excluded. Further characterization of such enzymes revealed additional differences from the rest of the type II endonucleases not only as expected in their monomer or heteromer versus homodimer structure, but in some cases also in other features (cofactor requirement, bifunctionality) so substantial, that it is now suggested that they may constitute new types (type IIS, type IV, BcgI-like) of restriction endonucleases (3–6). These observations indicate that restriction endonucleases recognizing asymmetric nucleotide sequences are likely candidates for discovery of even greater diversity. Among type II restriction endonucleases recognizing asymmetric nucleotide sequences there is a subgroup of enzymes [tentatively designated type IIT ENases (7)] which cleave within the asymmetric recognition sequence. No one representative of this group has yet been characterized beyond the determination of its specificity. One such enzyme, R.Bpu10I recognizes the asymmetric nucleotide sequence 5′-CCTNAGC, whose central pentanucleotide is represented by an interrupted palindrome, and which cleaves within it symmetrically, at positions –5/–2 in different strands (8). Preliminary experiments indicate that R.Bpu10I is composed of two non-identical subunits (9). We now provide detailed genetic and biochemical evidence to demonstrate that the Bpu10I R–M system from Bacillus pumilus 10 comprises four genes encoding two m5C MTases and two non-identical subunits of the restriction endonuclease. Properties of the Bpu10I R–M system and Bpu10I ENase make them unique among the type II enzymes that have been characterized so far. A plausible evolutionary scenario leading to the emergence of Bpu10I type enzymes concomitant with the generation of restriction endonucleases with specificities different from that of the progenitor is discussed. *To who correspondence should be addressed. Tel: +370 2 642 468; Fax: +370 2 642 624; e-mail: janulaitis@fermentas.lt Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 The Bpu10I R–M system from Bacillus pumilus 10, which recognizes the asymmetric 5′-CCTNAGC sequence, has been cloned, sequenced and expressed in Escherichia coli. The system comprises four adjacent, similarly oriented genes encoding two m5C MTases and two subunits of Bpu10I ENase (34.5 and 34 kDa). Both bpu10IR genes either in cis or trans are needed for the manifestation of R.Bpu10I activity. Subunits of R.Bpu10I, purified to apparent homogeneity, are both required for cleavage activity. This heterosubunit structure distinguishes the Bpu10I restriction endonuclease from all other type II restriction enzymes described previously. The subunits reveal 25% amino acid identity. Significant similarity was also identified between a 43 amino acid region of R.DdeI and one of the regions of higher identity shared between the Bpu10I subunits, a region that could possibly include the catalytic/Mg2+ binding center. The similarity between Bpu10I and DdeI MTases is not limited to the conserved motifs (CM) typical for m5C MTases. It extends into the variable region that lies between CMs VIII and IX. Duplication of a progenitor gene, encoding an enzyme recognizing a symmetric nucleotide sequence, followed by concerted divergent evolution, may provide a possible scenario leading to the emergence of the Bpu10I ENase, which recognizes an overall asymmetric sequence and cleaves within it symmetrically. DDBJ/EMBL/GenBank accession no. Y14683 1085 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.14 Nucleic MATERIALS AND METHODS Bacterial strains and plasmids Bacillus pumilus 10, used as the source of DNA for the cloning of the Bpu10I R–M system, was obtained from MBI Fermentas. The Escherichia coli K-12 strain RR1, (10) was used as a host for generating the B.pumilus 10 DNA library and for the propagation of recombinant plasmids carrying the cloned genes for R.Bpu10I or its individual subunits and Bpu10I MTases. Escherichia coli strain ER2267 of genotype [e14–(mcrA –)] endA1 supE44 thi-1 ∆(mcrC-mrr) 114::IS10 ∆(argF-lac)U169 recA1/F′ proA + B + lacI q ∆(lacz)M15 zzf::mini-Tn10 (KmR) was used as host for subcloning and deletion procedures. The positive selection cloning vector pBR-R, containing the intact cfr9IR gene with unique cloning sites (11) was used for cloning of the Bpu10I R–M system. Plasmids pUC19 (12), pACYC184 (13) and pAL4A (A.L., unpublished) were used as vectors for subcloning and deletion experiments. Media and transformation Enzymes and chemicals Restriction enzymes, T4 DNA ligase, DNA polymerase I large fragment (Klenow), nuclease Bal31, bacterial alkaline phosphatase (BAP), ExoIII/S1 Deletion Kit, DNA Labeling Kit (version 2.0), SmaI-digested and dephosphorylated pUC19 and sequencing primers were products of MBI Fermentas. The DNA Sequencing Kit used was purchased from Pharmacia. All enzymes and kits were used according to their respective manufacturer’s recommendations. [α-33P]dATP was obtained from Izotop (St. Petersburg, Russia). individual transformants obtained was then screened for resistance to R.Bpu10I digestion. Mapping of the B.pumilus 10 genomic R–M locus Southern-transfer of B.pumilus 10 genomic or pBpu10IM9.0 recombinant plasmid DNA, digested with various restriction enzymes (single and double digests), was performed as previously described (14). A DNA Labeling Kit was used for the preparation of a radioactive DNA probe containing a HindIII fragment of the cloned Bpu10I MTase gene. DNA sequence determination Overlapping deletions used in DNA sequencing were constructed using the nuclease Bal31 and ExoIII/S1 Deletion Kit. Sequencing was performed by the dideoxynucleotide chain-termination method (18) using the DNA Sequencing Kit, [α-33P]dATP, M13/pUC (direct, reverse) standard sequencing primers and double-stranded, supercoiled plasmid DNA as template. The reaction products were resolved by electrophoresis on wedgeshaped gels. Analysis of Bpu10I endonuclease and methylase activity In vitro endonuclease activity was tested by incubation of various amounts (1, 3, 5 µl) of cell-free extracts prepared as described (19) with 2 µg of λ DNA at 37_C for 40 min in 40 µl reaction mixture containing 10 mM Tris–HCl, pH 8.5, 10 mM MgCl2, 100 mM KCl and 0.1 mg/ml BSA (standard reaction mixture), followed by electrophoresis in 0.8% agarose gels. Chromatographic fractions containing individual subunits were assayed by complementation using aliquots of the crude cell extract prepared from recombinant E.coli cultures expressing the alternate subunit, as a supplement to the standard reaction mixture. To determine the Bpu10Ispecific modification generated in vivo, plasmid DNA isolated from transformants was challenged with an excess of R.Bpu10I followed by agarose gel electrophoresis. Amino acid sequence comparison DNA preparation and manipulation Bacillus pumilus 10 genomic DNA was extracted and purified as described by Marmur (15). Plasmid DNAs were prepared by the alkaline lysis procedure (16) and were further purified by binding to glass powder (17). Restriction and deletion mapping, agarose gel electrophoresis, isolation of individual DNA restriction fragments from agarose gels, subcloning of DNA fragments and Bal31 deletions were carried out by standard procedures (14). ExoIII/S1 deletions were obtained following the manufacturer’s recommendations. Library construction and selection of the clones harboring Bpu10I methyltransferase An aliquot (50 µg) of B.pumilus 10 genomic DNA was partially digested with Hin1II and ligated with T4 DNA ligase at 16_C for 24 h in a 800 ml reaction volume to 7 mg PaeI-cleaved, BAP-dephosphorylated pBR-R vector DNA. The ligation mixture was used to transform competent E.coli RR1 cells. Total plasmid DNA (10 µg) isolated from the pooled 100.000 CmR transformants was digested with excess R.Bpu10I (30 U) for 3 h and the transformed back into RR1. Plasmid DNA of 24 randomly picked The deduced amino acid sequences of all ORFs (50 and more amino acids residues) translated in six reading frames, were compared to all sequences deposited with the EMBL (Release 42) and the SWISS-PROT (Release 31) sequence data bases using the BLAST (20) procedure. Alignment of selected sequences was obtained using the MULTALIN (21) procedure applying the BLOSUM 62 (22) amino acid scoring matrix. Similarity of the Bpu10I subunit and MTase sequences to those of the available restriction endonucleases and m5C methyltransferases, respectively, was tested using the HR-SEARCH procedure (23). Gel electrophoresis of proteins Gel electrophoresis of proteins under denaturing conditions was performed as previously described (24). SDS–PAGE was carried out on 10% separating gels. Protein bands were visualized after Coomasie blue R250 staining. Cell growth and initial fractionation Escherichia coli RRI[pAL-Bpu10IRα] and E.coli RRI[pALBpu10IRβ], expressing the individual genes for the subunits of R.Bpu10I (α or β, respectively) cloned in the pAL4A vector, Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 Escherichia coli cells were grown in LB medium containing ampicillin (Ap, 60 µg/ml), kanamycin (Km, 50 µg/ml) and chloramphenicol (Cm, 30 µg/ml) as required. Cells were transformed using the CaCl2–heat shock method (14). Transformants were selected by plating onto LB agar supplemented with appropriate antibiotics. 1085 1086 Nucleic Acids Research, 1998, Vol. 26, No. 4 under the control of the thermoinducible λ PL promoter, were used as the source of subunits. The cells were grown into late logarithmic phase at 37_C with aeration in L broth containing 10 g/l trypton, 5 g/l yeast extract and 5 g/l NaCl. The cells were harvested by centrifugation and stored at –20_C. All further steps were carried out at 40_C. Cells (15 g) were thawed in 60 ml of buffer A (10 mM K-phosphate pH 7.0, 7 mM ME, 1 mM Na2EDTA) containing 0.25 M KCl. Then, following cell rupture by sonication, 10% Polymin P solution was gradually added to a final concentration of 0.7%, with stirring. The precipitate was removed by centrifugation at 12 000 g for 10 min. Soluble proteins were salted out with (NH4)2SO4 at 80% saturation and recovered by centrifugation. The pellet was dissolved in 1/3 of the initial volume of buffer A and dialyzed against buffer A containing 0.15 M KCl (buffer B). Dialysates were used for chromatographic purification of the individual α or β subunits of R.Bpu10I. Purification of the R.Bpu10I α subunit from E.coli RRI[pAL-Bpu10IRα] Bordo–Sepharose chromatography. The dialysate was applied to a Bordo–Sepharose column (1 × 10 cm) equilibrated with buffer B. After the application of the dialysate, the column was washed with 20 ml of buffer B and then eluted with a 100 ml linear gradient of 0.15–1 M KCl in buffer A. The α subunit eluted at X0.43 M KCl. The peak fractions were pooled and dialyzed against buffer B. Heparin–Sepharose chromatography. The dialysate was applied to a Heparin–Sepharose (1 × 10 cm) column, equilibrated with buffer B. The column was washed with 20 ml of the same buffer and eluted with a 100 ml linear gradient of 0.15–1 M KCl in buffer A. The subunit eluted around 0.5 M KCl. Fractions containing the α subunit were pooled, dialyzed against storage buffer (10 mM Tris–HCl, pH 7.5, 50 mM KCl, 1 mM DDT, 0.1 mM Na2EDTA, 50% glycerol) and stored at –20_C. Purification of the R.Bpu10I β subunit from E.coli RRI[pAL-Bpu10IRβ] Phosphocellulose and Bordo–Sepharose chromatography. Fractionation on phosphocellulose and Bordo–Sepharose was performed as described above for the isolation of the α subunit. In this case though a (1.5 × 20 cm) Bordo-Sepharose column was used and the limits of the KCl linear gradient used were different (0.15–0.6 M phosphocellulose and 0.2–1 M Bordo–Sepharose). R.Bpu10Iβ eluted around 0.35 M KCl from the phosphocellulose column and at X0.6 M KCl from the Bordo–Sepharose column. After Bordo–Sepharose chromatography the fractions containing the β subunit were pooled and dialyzed against buffer B. AH–Sepharose chromatography. The dialysate was applied to an AH–Sepharose column (1.5 × 20 cm) equilibrated with buffer B. The column was washed with 80 ml of the same buffer and then eluted with a 400 ml linear gradient of 0.15–0.8 M KCl in buffer Heparin–Sepharose chromatography. The dialysate was applied to a Heparin–Sepharose column (1.5 × 20 cm) equilibrated with buffer B and then washed with the same buffer. The column was eluted with a 200 ml linear gradient of 0.15–0.8 M KCl in buffer A. The R.Bpu10Iβ subunit eluted at ∼0.32 M KCl. Fractions containinng the β subunit were pooled, dialyzed against storage buffer and stored at –20_C. RESULTS AND DISCUSSION Cloning of the Bpu10I methyltransferase gene Selection of the gene coding for Bpu10I MTase was based on the resistance of self modifying recombinant plasmids to digestion by Bpu10I ENase. Twenty four randomly picked transformants obtained after the selection procedure, described under Materials and Methods, were further screened for the presence of Bpu10I specific modification. Plasmid DNA of 22 out of 24 clones was found to be protected against Bpu10I ENase challenge in vitro. None of these clones showed any Bpu10I ENase activity when assayed in vitro. Restriction mapping of the plasmids resistant to Bpu10I cleavage revealed that they contained a 1.65 kb HindIII fragment in common (data not shown). The recombinant plasmids containing the shortest (4 kb, pBpu10IM4.0) and the longest (9 kb, pBpu10IM9.0) of the cloned B.pumilus 10 DNA fragments were selected for further experiments. Deletion mapping of the pBpu10IM4.0 plasmid showed that the 2.2 kb region on the right side of the Mph1103I site contained a functionally active Bpu10I MTase gene (Fig. 1A). Southern hybridization, using the radioactive 1.65 kb HindIII DNA fragment from the cloned MTase gene as a probe, and B.pumilus 10 or pBpu10IM9.0 DNA, digested with various restriction endonucleases, indicated that the pBpu10M9.0 plasmid contains the intact genomic DNA fragment of B.pumilus 10 (data not shown). The DNA flanking the 1.65 kb HindIII fragment is large enough to include the ENase gene, usually closely linked to the genes of the cognate MTase. However, clones harboring the pBpu10IM9.0 plasmid did not reveal R.bpu10I activity, indicating that the corresponding gene was either not cloned or not expressed in E.coli cells. Sequencing of the bpu10IM gene and the flanking regions, revealed genes encoding the complete Bpu10I R–M system (see below). Gene structure of the Bpu10I R–M system A 4410 bp region of the cloned 9.0 kb fragment encompassing the gene for Bpu10I MTase was sequenced on both strands (EMBL accession no. Y14683). Four ORFs, all oriented in the same direction, were identified (Fig. 1B and C). They encode the Bpu10I R–M enzymes: two separate m5C MTases (Bpu10IC1 and Bpu10IC2) and two ENase heterosubunits (Bpu10Iα and Bpu10Iβ) (see below). The first and the second ORFs corresponded to the genes bpu10IMC1 and bpu10IMC2 of two m5C MTases. ORF1 was 1194 bp long (termination codon not included), extended from nt 110 to 1303, and encoded a protein of 398 amino acids. It overlapped by 1 nt (or by 4 nt if the stop codon TGA of ORF1 was included) with the downstream ORF2 (nt 1303–2478 = 1176 bp, 392 amino acids; Fig. 1C). Putative RBS sequences AGAGG (nt 95–99) and GGGGG (nt 1288–1292; not shown in Fig. 1C) were detected upstream of both ORFs, respectively. A Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 Phosphocellulose chromatography. The dialysate was applied to a phosphocellulose P-11 column (1.5 × 20 cm) equilibrated with buffer B. The column was washed with 80 ml of the same buffer and eluted with a 400 ml linear gradient of 0.15–1 M KCl in buffer A. Bpu10Iα eluted around 0.65 M KCl. The peak fractions containing the α subunit were pooled and dialyzed against buffer B. A. The subunit eluted at around 0.3 M KCl. Fractions containing R.Bpu10Iβ subunit were pooled and dialyzed against buffer B. 1087 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.14 Nucleic 1087 putative promoter region was found only upstream of the bpu10IMC1 gene. The third ORF, which corresponded to the bpu10IRα gene, started 95 bp downstream from the last nucleotide of bpu10IMC2 and encoded 294 amino acid residues (34.5 kDa) of the Bpu10I ENase α subunit (start codon GTG; nt 2573–3454 = 882 bp). It was separated by the stop codon TAA and the ATC triplet (nt 3455–3460), from the structural part of the gene bpu10IRβ encoding R.Bpu10Iβ (ORF4; start codon ATG; nt 3461–4324 = 864 bp; 288 amino acids = 34 kDa). The putative Shine–Dalgarno sequences AGGAGA (nt 2560–2565) and AGGA (nt 3448–3451; not shown in Fig. 1C) were detected for ORF3 and ORF4, respectively. Putative promoter sequences were found only upstream from ORF3. These findings suggest that genes of the Bpu10I R–M system are organized into two separate operons: one for the genes of the two Bpu10I MTases and another for the two subunits of Bpu10I ENase. Gene assignment Deletion and subcloning experiments were carried out to investigate the expression and function of the four ORFs identified. Bpu10I MTases. The intact ORF1 and 5′ terminal part of the ORF2 in the 4.0 kb fragment were cloned (Fig. 1A and B). Thus, ORF1 is clearly expressed in E.coli as the Bpu10I specific MTase (M.Bpu10IC1). Incomplete plasmid DNA protection (∼80%) from R.Bpu10I cleavage was found when the 0.69 kb DNA fragment containing the N-terminal part and the putative promoter of the M.Bpu10IC1, together with the exogenous promoter of the tetracycline resistance gene, was deleted from the plasmid pAC-Bpu10IMC1/C2 (pAC-Bpu10IM∆Em; Fig. 1B). Nevertheless, these results indicate that the ORF2 encodes another Bpu10I MTase (M.Bpu10IC2). These conclusions were supported by the sequence analysis of the ORF1 and ORF2, which revealed conserved amino acid motifs typical of DNA m5C MTases (25). Low level bpu10IMC2 gene expression is consistent with the lack of promoter sequences upstream of the gene. As expected, given the operon like organization of Bpu10I MTase genes, the level of bpu10IMC2 expression was increased under the control of the promoter located in front of bpu10IMC1 (Fig. 1C) to a level which proved high enough to protect DNA against R.Bpu10I challenge in vitro (pAC-Bpu10IMC2 and pAC-Bpu10IMC2′; Fig. 1B). Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 Figure 1. (A) Restriction endonuclease and deletion mapping experiments using the plasmid pBpu10IM4.0. Thin lines indicate the cloning vector pBR-R, hatched lines, the 4 kb cloned B.pumilus 10 DNA fragment. Restriction endonuclease sites are abbreviated as follows: Cf, Cfr42I; Ec, Eco105I; E47, Eco47III; Hi, HindIII; Mp, Mph1103I; ∆, deletion between restriction sites as indicated. (B) Subcloning and deletion mapping of the genes encoding the Bpu10I R–M system. Only B.pumilus 10 DNA fragments cloned or subcloned into pBR-R, pUC19 or pACYC184 vectors are represented. The black lines represent the sequenced DNA fragment. Directions of transcription and positions of the Bpu10I R–M genes are indicated by large arrows: MC1 corresponds to bpu10IMC1; MC2, bpu10IMC2; Rα, bpu10IRα; Rβ, bpu10IRβ. Direction of transcription from the known vector promoters adjacent to the cloned fragments is indicated by small arrows. Recognition sites of the restriction endonucleases used in subcloning and deletion mapping experiments are abbreviated as follows: Mp, Mph1103I; Em, Eam1104I; Ec, Eco105I; Vs, VspI; Pa, PaeI; Nd, NdeI. Plasmid pAC-Bpu10IMC1/C2 was obtained after subcloning the 2.51 kb Mph1103I–VspI DNA fragment, containing the intact genes of bpu10IMC1 and bpu10IMC2, from pBpu10IM9.0 into Eco32I-digested pACYC184. To obtain pAC-Bpu10IMC2∆Em the 5′ part of the bpu10IMC1 and upstream DNA region including the heterogenous Ptet promoter from the plasmid pAC-Bpu10IMC1/C2 was deleted (the second Eam1104I site being located within the vector). pAC-Bpu10IMC2 is a derivative of the pAC-Bpu10IMC1/C2, in which a part (∼0.4 kb) of the bpu10IMC1 gene was deleted by restriction with Eco105I, then being treated with Bal31 exonuclease. pAC-Bpu10IMC2′ was obtained after subcloning the 2.2 kb BfmI–NheI DNA fragment from the pAC-Bpu10IMC2 (restriction sites reside within the vector; not shown) into Eco32I-cleaved pACYC184. Plasmids pUC-Bpu10IRα/β, pUC-Bpu10IRα/β′ and pAC-Bpu10IRα/β were obtained after insertion of the 1.9 kb VspI fragment from pBpu10IM9.0 into SmaI-digested pUC19 and Eco32I-cleaved pACYC184, respectively. Plasmids pUC-Bpu10IRβ and pUC-Bpu10IRα were constructed by deletion of the 1 kb PaeI or the 0.4 kb NdeI DNA fragments respectively from the plasmid pUC-Bpu10IRα/β (PaeI and NdeI sites located within the vector are not shown). The plasmid pAC-Bpu10IRα was formed by deletion of the 2 kb NdeI–Eco81I DNA fragment from pAC-Bpu10IRα/β (Eco81I site resides within the vector). R–M phenotypes were determined as described under Materials and Methods. M+, MTase activity; M–, no activity; M+/–, partial activity; R+, ENase activity; R–, no activity. (C) Gene organization of the Bpu10I R–M system. White and black arrows indicate direction of transcription and relative position of the bpu10I R-M genes in the sequenced DNA fragment. Nucleotide sequences of the putative promoters and RBS are presented below. The putative start codons, RBS and –35, –10 promoter sequences are indicated in bold. Asterisks denote the stop codons. 1088 Nucleic Acids Research, 1998, Vol. 26, No. 4 Bpu10I ENase. No activity of the Bpu10I ENase was detected in crude extracts prepared from the cells carrying the pBpu10I9.0 plasmid (Fig. 1B). The 1.9 kb DNA fragment encompassing both the intact ORF3 and ORF4, when subcloned from the plasmid pBpu10IM9.0 into the pUC19 vector in both directions (pUCBpu10IRα/β and pUC-Bpu10IRα/β′; Fig. 1B), manifest Bpu10I ENase activity in vitro. This indicated that ORF3 and/or ORF4 encoded a restriction endonuclease and that the endogenous promoter(s) operates in E.coli. Interestingly, host cell DNA protection by specific methylation was not required to generate viable transformants expressing Bpu10I ENase activity, as has been observed in some other R–M systems [for references see (26)]. In cells harboring the intact Bpu10I R–M system (plasmid pBpu10IM9.0) no activity of the Bpu10I ENase was detected. It is unlikely, however, that some increase of plasmid copy number, after the subcloning of the Bpu10I ENase gene(s) from pBR-R vector (pBpu10IM9.0) into the pUC19 (pUC-Bpu10IRα/β; Fig. 1B) could result in so dramatic an increase in enzyme activity. The ORF3 and ORF4 were precisely excised from the pBpu10IM9.0 plasmid before subcloning, leaving a 4 kb DNA region located upstream of the MTase genes and a 95 bp intergenic region separating ORF2 and ORF3. The existence of some regulatory elements of R.Bpu10I expression within these regions cannot be excluded. No Bpu10I ENase activity was observed in crude cell extracts of the clones harboring deletion derivatives, where either the ORF3 (pUC-Bpu10IRβ) or ORF4 (pUC-Bpu10IRα) were inactivated (Fig. 1B). When plasmids pAC-Bpu10IRα and pUC-Bpu10IRβ were transformed into the same cell, on the other hand, Bpu10I ENase activity was detected in cell extracts (data not shown). This suggests that the proteins encoded by ORF3 and ORF4 function in a mutually dependent fashion, such that both genes, either in cis or trans, are needed for the manifestation of R.Bpu10I activity. In parallel with these observations, crude cell extracts prepared from the individual clones carrying the cloned intact ORF3 or ORF4 (pUC-Bpu10IRα or pUC-Bpu10IRβ) complemented when added to the reaction mixture (data not shown). Comparison and analysis of deduced amino acid sequences Bpu10I MTases. A 39% identity between M.Bpu10IC1 and M.Bpu10IC2 was observed with an additional 24% similarity derived from conservative substitutions. Both possess 10 conserved amino acid motifs (CM; not shown), characteristic of DNA m5C MTases (25). The greatest similarity amongst all m5C MTases was observed in comparison with DdeI. The level of similarity between DdeI and the Bpu10I MTases was approximately equal to that between the Bpu10I MTases themselves. This similarity is not limited to the conserved motifs but also extend into the region that lies between the CMs VIII and IX and beyond CMs IX where four conserved regions (CR) of higher identity are found (CRs A, B, C and D; Fig. 2). Genetic and biochemical evidence has been obtained which indicates that a variable region located between CM VIII and CM IX of both phage multi-specific and bacterial mono-specific MTases contains a target recognition domain (TRD) which defines specificity of these enzymes as well as the base to be methylated within the target sequence (27–30). This general concept was confirmed and refined significantly by the resolution of the atomic structures of two m5C MTases, HhaI (31) and HaeIII (32) complexed with substrate DNA. The similarity between variable regions of m5C MTases is observed only in some of the enzymes which recognize identical Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 Figure 2. Alignment of M.Bpu10IC1, M.Bpu10IC2 and M.DdeI amino acid sequences (A) and their DNA target sequences (B). (A) Only the protein sequences encompassing CM VIII and CM IX, which are common to m5C MTases, the variable region between them, and conserved region D downstream of CM IX, are shown. The numbers on the left margin denote the amino acid position relative to the N-terminus. White letters on the black background and the black letters on the shaded background indicate amino acid residues that are common and similar in aligned sequences, respectively. Similar amino acids are grouped as follows: (K, R), (T, S), (D, E, N, Q), (A, V, I, L, M, F, W, Y), (A, G). Uppercase letters (A, B, C, D) above aligned sequences correspond to the conserved regions common for all three aligned MTases. (B) The 5′-CTNAG sequences common to the recognition sequences of Bpu10I MTases and M.DdeI are boxed. The target cytosines are boldfaced. 1089 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.14 Nucleic 1089 or related nucleotide sequences which modify the same target base (33). M.DdeI recognizes the 5′-CTNAG pentanucleotide and modifies the C base (34). The same pentanucleotide represents the central symmetrical part of the nucleotide sequences recognized by Bpu10I MTases. Bpu10IC1 MTase modifies the 5′-GCTNAGG sequence (target base underlined), while M.Bpu10IC2–5′-CCTNAGC (Z.Maneliene, personal communication) i.e. modifies the C base which belongs to the 5′-CTNAG pentanucleotide. The sequence similarities between the variable regions of M.DdeI and Bpu10I MTases may therefore reflect a similarity of structural organisation between the M.DdeI specificity domain and the component of the Bpu10I MTases which recognises the same target sequence 5′-CTNAG. One monomeric MTase is sufficient for the recognition and modification of a palindromic nucleotide sequence (33). For type IIS MTases it has been demonstrated that a range of variations in enzyme structural organisation are used in the modification of asymmetric recognition sequences. These include a single monomeric bifunctional enzyme modifying adenine residues in both strands of target DNA (35); DNA MTases yielding m6A or m5C on complementary strands in the reaction catalysed by a single tandemly arranged bifunctional MTase or two separate m6A and m5C MTases (36; J.Bitinaite, personal communication); and two separate m5C MTases each responsible for the methylation of different DNA strands (11,37). The Bpu10I MTase is likely to belong to the last group. Bpu10I ENase. Comparison of the R.Bpu10Iα and R.Bpu10Iβ deduced amino acid sequences revealed quite a low degree of similarity (25% of amino acids are identical and 17% are functionally similar). Analysis using HR-SEARCH (23), however, resulted in the extraction of four regions of higher identity (43% for A region; 46%, B; 37%, C and 31%, D; Fig. 3). The subunits of R.Bpu10I resembled none of the protein sequences in the EMBL and SWISS-PROT databases, nor did they resemble any of those in our personal restriction enzyme data base. The most significant regional similarity was identified between the 43 amino acid region in the DdeI ENase (which extends from amino acid 120 to 162 of the R.DdeI), and the N-terminal part of the conserved region C (28% identical amino acids for R.DdeI versus R.Bpu10Iα or R.Bpu10Iβ). The consensus motif P(E/D)X9-18(E/D)XK, identified as essential for catalysis and Mg2+ binding in some restriction endonucleases (38), could not be found in R.Bpu10I subunits. However, the EXK motif preceded 8 amino acids upstream by a negatively charged amino acid residue (D or E) is conserved, as indicated by alignment of R.Bpu10I subunit sequences with that of R.DdeI (Fig. 3; region C), elements reminiscent of the P(E/D)X9-18(E/D)XK consensus motif (except for P). If experimental evidence is obtained to support the suggestion that the catalytic/Mg2+ binding motif (or part of it) of R.Bpu10I and R.DdeI is of the structure (D/E)X8EXK, then it could represent a new variant. Subunit structure of the Bpu10I ENase Initial experiments, using crude cell extracts prepared from B.pumilus 10, revealed a total loss of enzyme activity (measured using the standard reaction mixture) after chromatography on phosphocellulose P11. The separation of the two R.Bpu10I heterosubunits, whose existence was demonstrated on the basis of genetic evidence, explained these findings. Indeed Bpu10I ENase activity was recovered in two peaks after fractionation on phosphocellulose P11, using the standard reaction mixture supplemented with aliquots of crude cell extracts prepared from the recombinant clones E.coli RRI[pAL-Bpu10IRα] or E.coli RRI[pAL-Bpu10IRβ] expressing the individual subunits α and β, respectively. Bpu10I ENase activity recovered eluted at around 0.33 M (β subunit) and 0.66 M KCl (α subunit). Bpu10I subunits were purified from the recombinant clones to apparent homogeneity. The molecular masses of these proteins (Fig. 4A) were determined to be close to those predicted for the α and β subunits on the basis of their nucleotide sequences (34.5 kDa α subunit and Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 Figure 3. The amino acid alignment of R.Bpu10Iα, Bpu10Iβ and the 43 amino acid region of R.DdeI discussed. The conventions used are the same as those in Figure 2. In the alignment of R.DdeI with R.Bpu10I subunit sequences aa residues common to R.DdeI and at least one of the subunits are shown. Uppercase letters (A, B, C, D) above aligned sequences correspond to the regions of higher identity between the subunits of R.Bpu10I. 1090 Nucleic Acids Research, 1998, Vol. 26, No. 4 34 kDa β subunit). Both subunits were required for R.Bpu10I activity (Fig. 4B). Investigation of the Bpu10I ENase oligomeric organization, however, under non-denaturing conditions, was hampered due to the irreversible adsorption (or inactivation) of at least of one of the subunits, to the gel filtration media used. During PAGE electrophoresis under non-denaturing conditions the mixture of α and β subunits formed heavy aggregates which failed to enter the gel further than a few mm (data not shown). The same was observed when the individual subunits were subjected to electrophoresis. The two ORFs encoding Bpu10I ENase are separated by the translation stop codon TAA followed by the ATC codon which precedes the putative ATG start codon for bpu10IRβ (Fig. 1C). The assignment of the initiation codon was further confirmed by the N-terminal sequence of the purified β subunit. The experimentally determined N-terminal pentapeptide matched the Thr-Gln-IleAsp-Leu sequence deduced from the nucleotide sequence except for the absence of the N-terminal Met. It could not be excluded though that the R.Bpu10I produced in B.pumilus 10 is composed of one polypeptide (e.g. due to suppression of stop codon). Fractionation of the crude cell extract prepared from B.pumilus 10, the natural host of the Bpu10I R–M system, on the phosphocellulose P11 column, revealed a separation of the subunits identical to that for the subunits expressed in the recombinant clones. These findings indicate, that the molecular organization of R.Bpu10I, generated from the recombinant clone, is the same as that in B.pumilus 10, which demonstrates that R.Bpu10I is composed of two heterosubunits. The association between subunits seems to be particularly weak, such that numerous attempts to purify Bpu10I ENase proved unsuccessful. Instead, α and β subunits can be purified and when mixed, reconstitute restriction activity. This indicates that subunit interaction necessary for target recognition and (or) catalysis, is mediated by substrate. Therefore, it can be concluded that the active form of Bpu10I corresponds to the heteromer. Evolutionary implications It is generally assumed that restriction endonucleases have evolved numerous times (42). This assumption, however, does not exclude the possibility that ENases emerged independently to evolve further, generating new specificities i.e. recognizing nucleotide sequences possibly related to that of a predecessor (43,44). Although the overall primary sequence similarity between subunits of the Bpu10I ENase is quite low, it is high enough to indicate the similarity of their 3-dimensional structures (45) and their evolutionary relatedness (46). This notion is supported by the presence of several regions of higher identity which are distributed over the aligned sequences of R.Bpu10I subunits in the same relative positions. R.Bpu10I recognizes an overall asymmetric sequence whose central part encompasses the interrupted palindrome (5′-CTNAG) and cleaves symmetrically within this part (8). It is possible that a restriction endonuclease recognizing a symmetric sequence (e.g. CTNAG or CCTNAGG) may have been a predecessor of R.Bpu10I. Such a progenitor gene would encode one polypeptide, which after forming a homodimer would be sufficient for recognition of the palindromic nucleotide sequence. Divergent concerted evolution of duplicated genes may have generated an enzyme of a new specificity, each subunit of which might recognize different non-symmetrical parts of the target (outside bases) in a different way. The similarity between subunits of the R.Bpu10I and R.DdeI is restricted to a Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 Figure 4. (A) SDS–PAGE electrophoresis of R.Bpu10Iα and R.Bpu10Iβ. An aliquot (0.8 µg) of each subunit was prepared and electrophoresed on a 10% SDS–polyacrylamide gel as described under Materials and Methods. M, molecular weight marker (Boehringer Ingelheim); lane 1, R.Bpu10Iα; lane 2, R.Bpu10Iα + R.Bpu10Iβ; lane 3, R.Bpu10Iβ. (B) Restriction activity of the R.Bpu10I. For each reaction 1.5 µg of λ DNA were incubated with purified α and β subunits of R.Bpu10I. Products were then applied to a 0.8% agarose gel, using the same conditions as described under Materials and Methods. Lane 1, λ DNA; lane 2, λ DNA + R.Bpu10Iα; lane 3, λ DNA + R.Bpu10Iα and β; lane 4, λ DNA + R.Bpu10Iβ. Bpu10I ENase’s structure establishes it as different from other restriction endonucleases, recognizing asymmetric nucleotide sequences (4–6), all of which originally were classified as type II enzymes. The diversity in structure, cofactor requirement, type of DNA cleavage and functional properties (monofunctionality verses bifunctionality) of this group of enzymes, confirms this group as distinct from the group recognizing palindromes, which are homodimers activated by Mg2+. Among the enzymes recognizing asymmetric nucleotide sequences the properties of R.Bpu10I are most close to the type IIS ENases (4,39). Both are monofunctional enzymes revealing only DNA cleavage activity and requiring only Mg2+ for activity. They differ though, in position of cut point relative to the asymmetric recognition sequence (R.Bpu10I within the sequence, type IIS – outside) and protein structure (heteromer versus monomer). The modification component of some type IIS R–M systems is represented by two MTases (11,37) as in case of Bpu10I R–M system. The Bpu10I R–M system consists of four genes (and proteins), which is unique among other representatives of the type II R–M systems. Thus, consistent with the subdivision of the class II ENases to establish the group of type IIS enzymes (4), an additional group designated type IIT might be introduced for enzymes like R.Bpu10I as has been previously proposed by Kessler and Manta (7). The possibility of heteromeric structure of type II restriction endonucleases recognizing asymmetric nucleotide sequences has previously been addressed (40). Co-purification of two polypeptides of slightly different size during isolation of CauII ENase, which recognizes the 5′-CCSGG nucleotide sequence was used as an indication of heteromeric composition. Biochemical and genetic data on R.BcnI, an isoschizomer of R.CauII, unambiguously shows that BcnI ENase operates in the form of a single protein (41; A.J., unpublished). However, it cannot been excluded that R.CauII recognizes the 5′-CCSGG sequence as asymmetric, while R.BcnI recognizes the same sequence as symmetric. 1091 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.14 Nucleic stretch of 43 amino acids that plausibly includes the catalytic/ Mg2+ binding center. Therefore, R.DdeI can not be assumed to be a recent predecessor of R.Bpu10I, although this would not exclude the possibility that the catalytic/Mg2+ binding center and recognition of the 5′-CTNAG sequence might reside within the 43 amino acid stretch for both DdeI and Bpu10I ENases. A scenario similar to that described above for Bpu10I ENase might well be deduced for Bpu10I MTases. ACKNOWLEDGEMENTS We thank MBI Fermentas for providing enzymes and chemicals, Elisabeth A. Raleigh for E.coli strain ER2267. We are also grateful to Elena Zareckaja for sequencing N-terminal part of the R.Bpu10Iβ subunit, Saulius Klimasuskas for discussion and to Barbara Richmond-Smith and Geoff Werstuck for linguistic help. REFERENCES 5 6 7 8 9 10 11 12 13 14 15 16 17 Yuan,R. (1981) Annu. Rev. Biochem., 50, 285–315. Roberts,R.J. and Macelis,D. (1997) Nucleic Acids Res., 25, 248–262. Pingoud,A. and Jeltsch,A. (1997) Eur. J. Biochem., 246, 1–22. Szybalski,W., Kim,S.C., Hasan,N. and Podhajska,A.J. (1991) Gene, 100, 13–26. Janulaitis,A., Petrusyte,M., Maneliene,Z., Klimasauskas,S. and Butkus,V. (1992) Nucleic Acids Res., 20, 6043–6049. Kong,H., Roemer,S.E., Waite-Rees,P.A., Benner, J.S., Wilson,G.G. and Nwankwo,D.O. (1994) J. Biol. Chem., 269, 683–690. Kessler,C. and Manta,V. (1990) Gene, 92, 1–248. Degtyarev,S.Kh., Zilkin,P.A., Prihodko,G.G., Repin,V.E. and Rechkunova,N.I. (1989) Molekularnaja Biologija, 23, 1051–1056. Stankevicius,K., Lubys,A., Timinskas,A. and Janulaitis,A. (1996) Biologija, 2, 51–53. Maniatis,T., Fritsch,E. and Sambrook,J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Lubys,A., Lubiene,J., Kulakauskas,S., Stankevicius,K., Timinskas,A. and Janulaitis,A. (1996) Nucleic Acids Res., 24, 2760–2766. Yanisch-Perron,C., Vieira,J. and Messing,J. (1985) Gene, 33, 103–119. Chang,A.C.Y. and Cohen,S.N. (1978) J. Bacteriol., 134, 1141–1156. Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Edn. Cold Spring Harbor Laboratory press, Cold Spring Harbor, NY. Marmur,J. (1961) J. Mol. Biol., 3, 208–218. Birnboim,H.C. and Doly,J. (1979) Nucleic Acids Res., 7, 1513–1523. Marko,M.A., Chipperfield,R. and Birnboim,H.C. (1982) Anal. Biochem., 121, 382–387. 18 Chen,E.Y. and Seeburg,P.H. (1985) DNA, 4, 165–170. 19 Whitehead,P.R. and Brown,N.L. (1985) Arch. Microbiol., 141, 70–74. 20 Altschul,S.F., Gish,W., Miller,W., Myers,E.W. and Lipman,D.J. (1990) J. Mol. Biol., 215, 403–410. 21 Corpet,F. (1988) Nucleic Acids Res., 16, 10881–10890. 22 Henikoff,S. and Henikoff,J.G. (1992) Proc. Natl. Acad. Sci. USA, 89, 10915–10919. 23 Timinskas,A., Butkus,V. and Janulaitis,A. (1995) Gene, 157, 3–11. 24 Ausubel,F.M., Kingston,R.E., Moore,D.D., Seidman,J.G., Smith,J.A. and Struhl,K. (1992) In Ausubel,F.M (ed.), Short Protocols In Molecular Biology (2nd edition). Green Publishing Associates and John Wiley & Sons. 25 Posfai,J., Bhagwat,A.S., Posfai,G. and Roberts,R.J. (1989) Nucleic Acids Res., 17, 2421–2435. 26 Wilson,G.G. (1988) Gene, 74, 281–289. 27 Balganesh,T.S., Reiners,L., Lauster,R., Noyer-Weidner,M., Wilke,K. and Trautner,T.A. (1987) EMBO J., 6, 3543–3549. 28 Trautner,T.A., Balganesh,T.S. and Pawlek,B. (1988) Nucleic Acids Res., 16, 6649–6657. 29 Wilke,K., Rauhut,E., Noyer-Weidner,M., Lauster,R., Pawlek,B., Behrens,B. and Trautner,T.A. (1988) EMBO J., 7, 2601–2609. 30 Klimasauskas,S., Nelson,J.L. and Roberts,R.J. (1991) Nucleic Acids Res., 19, 6183–6190. 31 Klimasauskas,S., Kumar,S., Roberts,R.J. and Cheng,X. (1994) Cell, 76, 357–369. 32 Reinisch,K.M., Chen,L., Verdine,G.L. and Lipscomb,W.N. (1995) Cell, 82, 143–153. 33 Noyer-Weidner,M. and Trautner,T.A. (1993) In Jost,J.P. and Saluz,H.P. (eds) DNA Methylation: Molecular Biology and Biological Significance. Birkhäuser Verlag, Basel/Switzerland pp. 39–107. 34 Sznyter,L.A., Slatko,B., Moran,L., O’Donnell,K.H. and Brooks,J.E. (1987) Nucleic Acids Res., 15, 8249–8266. 35 Looney,M.C., Moran,L.S., Jack,W.E., Feehery,G.R., Benner,J.S., Slatko,B.E. and Wilson,G.G. (1989) Gene, 80, 193–208. 36 Bitinaite,J., Maneliene,Z., Menkevicius,S., Klimasauskas,S., Butkus,V. and Janulaitis,A. (1992) Nucleic Acids Res., 20, 4981–4985. 37 Sugisaki,H., Yamamoto,K. and Takanami,M. (1991) J. Biol. Chem., 266, 13952–13957. 38 Anderson,J.E. (1993) Curr. Opin. Struct. Biol., 3, 24–30. 39 Wah,D.A., Hirsch,J.A., Dorner,L.F., Schildkraut,I. and Aggarwal,A.K. (1997) Nature, 388, 97–100. 40 Bennet,S.P. and Halford,S.E. (1987) In Guschlbauet and Saenger,W. (eds), DNA-ligand Interactions from Drugs to Proteins. Plenum, New York pp. 239–250. 41 Petrusyte,M.P. and Janulaitis,A.A. (1982) Eur. J. Biochem., 121, 377–381. 42 Wilson,G.G. and Murray,N.E. (1991) Annu. Rev. Genet., 25, 585–627. 43 Siksnys,V., Zareckaja,N., Vaisvila,R., Timinskas,A., Stakenas,P., Butkus,V. and Janulaitis,A. (1994) Gene, 142, 1–8. 44 Janulaitis,A. (1996) Biologija, 2, 28–31. 45 Sanders,C. and Schneider,R. (1991) Proteins, 9, 56–68 46 Doolittle,R.F. (1981) Science, 214, 149–159. Downloaded from http://nar.oxfordjournals.org/ by guest on June 21, 2015 1 2 3 4 1091