Method for cloning and expression of BsrFI restriction endonuclease in E. coli

BsrFI restriction enzyme was purified from Bacillus stearothermophilus to near homogeneity. The protein was sequenced to obtain its N-terminus amino acid sequence. A set of denegerate primers were synthesized based on the aa sequence. The first 18 codons encoding BsrFI restriction endonuclease (bsrFIR) was amplified by PCR and its coding sequence was obtained. The methylase selection method was used to clone BsrFI methylase gene (bsrFIM). Two clones were found to be resistant to BsrFI digstion. The entire insert in one clone was sequenced and the insert encodes the BsrFI methylase (M. BsrFI). In addition, a small truncated open reading frame adjacent to the methylase gene has homology to Cfr10I restriciton endonuclease in a BlastX homology search in Genbank database. BsrFI and Cfr10I are isoschizomer that recognizes and cleaves 5'R CCGGY3'. Two primers were used to amplify the bsrFIR gene, The forward primer is a degenerate primer designed from the N-terminus aa sequence and the reverse primer is the bona fide sequence derived from the BsrFI methylase⁺ clone. The bsrFIR gene was amplified by PCR, ligated into a T7 expression vector pET21at and the ligated DNA was transformed into premodified cells ER2566 [pLG339-BsrFIM]. The final expression strain is ER2566 [pLG339-BsrFIM, pET21at-BsrFIR]. Recombinant BsrFI activity was detected in E. coli cell extract. BsrFI is cloned from a thermophile Bacillus stearothermophilus. Thus, BsrFI a thermostable enzyme and it is active at 37° C. to 65° C.

1. Isolated DNA coding for the BsrFI restriction endonuclease, wherein the isolated DNA is obtainable from Bacillus stearothermophilus.

2. A recombinant DNA vector comprising a vector into which a DNA segment encoding the BsrFI restriction endonuclease has been inserted.

3. Isolated DNA encoding the BsrFI restriction endonuclease and methylase, wherein the isolated DNA is obtainable from ATCC No. PTA-28.

4. A cloning vector which comprises the isolated DNA of claim 3.

5. A host cell transformed by the vector of claim 2 or 4.

6. A method of producing recombinant BsrFI restriction endonuclease comprising culturing a host cell transformed with the vector of claim 2 or 4 under conditions suitable for expression of said endonuclease.

The present invention relates to recombinant DNA which encodes the BsrFI restriction endonuclease as well as BsrFI methyltransferase, and production of BsrFI restriction endonuclease from E. coli cells containing the recombinant DNA.

Type II restriction endonucleases are a class of enzymes that occur naturally in bacteria and in some viruses. When they are purified away from other bacterial proteins, restriction endonucleases can be used in the laboratory to cleave DNA molecules into small fragments for molecular cloning and gene characterization.

Restriction endonucleases act by recognizing and binding to particular sequences of nucleotides (the `recognition sequence`) along the DNA molecule. Once bound, they cleave the molecule within, to one side of, or to both sides of the recognition sequence. Different restriction endonucleases have affinity for different recognition sequences. Over two hundred and eleven restriction endonucleases with unique specificities have been identified among the many hundreds of bacterial species that have been examined to date (Roberts and Macelis, Nucl. Acids Res. 27:312-313, (1999)).

Restriction endonucleases typically are named according to the bacteria from which they are derived. Thus, the species Deinococcus radiophilus for example, produces three different restriction endonucleases, named DraI, DraII and DraIII. These enzymes recognize and cleave the sequences 5'TTTAAA3', 5'PuGGNCCPy3' and 5' CACNNNGTG3' respectively. Escherichia coli RY13, on the other hand, produces only one enzyme, EcoRI, which recognizes the sequence 5'GAATTC3'.

A second component of bacterial restriction-modification (R-M) systems are the methyltransferase (methylases). These enzymes are complementary to restriction endonucleases and they provide the means by which bacteria are able to protect their own DNA and distinguish it from foreign, infecting DNA. Modification methylases recognize and bind to the same recognition sequence as the corresponding restriction endonuclease, but instead of cleaving the DNA, they chemically modify one particular nucleotide within the sequence by the addition of a methyl group (C5 methyl cytosine, N4 methyl cytosine, or N6 methyl adenine). Following methylation, the recognition sequence is no longer cleaved by the cognate restriction endonuclease. The DNA of a bacterial cell is always fully modified by virtue of the activity of its modification methylase. It is therefore completely insensitive to the presence of the endogenous restriction endonuclease. It is only unmodified, and therefore identifiably foreign DNA, that is sensitive to restriction endonuclease recognition and cleavage.

With the advent of recombinant DNA technology, it is now possible to clone genes and overproduce the enzymes in large quantities. The key to isolating clones of restriction endonuclease genes is to develop a simple and reliable method to identify such clones within complex `libraries`, i.e. populations of clones derived by `shotgun` procedures, when they occur at frequencies as low as 10^-3 to 10^-4. Preferably, the method should be selective, such that the unwanted majority of clones are destroyed while the desirable rare clones survive.

A large number of type II restriction-modification systems have been cloned. The first cloned systems used bacteriophage infection as a means of identifying or selecting restriction endonuclease clones (EcoRII: Kosykh et al., Mol. Gen. Genet. 178:717-719, (1980); HhaII: Mann et al., Gene 3:97-112, (1978); PstI: Walder et al., Proc. Nat. Acad. Sci. 78:1503-1507, (1981)). Since the presence of restriction-modification systems in bacteria enable them to resist infection by bacteriophage, cells that carry cloned restriction-modification genes can, in principle, be selectively isolated as survivors from libraries that have been exposed to phages. This method has been found, however, to have only limited value. Specifically, it has been found that cloned restriction-modification genes do not always manifest sufficient phage resistance to confer selective survival.

Another cloning approach involves transferring systems initially characterized as plasmid-borne into E. coli cloning plasmids (EcoRV: Bougueleret et al., Nucl. Acids. Res. 12:3659-3676, (1984); PaeR7: Gingeras and Brooks, Proc. Natl. Acad. Sci. USA 80:402-406, (1983); Theriault and Roy, Gene 19:355-359 (1982); PvuII: Blumenthal et al., J. Bacteriol. 164:501-509, (1985); Tsp45I: Wayne et al. Gene 202:83-88, (1997)).

A third approach, and one that is being used to clone a growing number of R-M systems are now being cloned by selection for an active methylase gene (U.S. Pat. No. 5,200,333 and BsuRI: Kiss et al., Nucl. Acids. Res. 13:6403-6421, (1985)). Since R-M genes are often closely linked, both genes can often be cloned simultaneously. This selection does not always yield a complete restriction system however, but instead yields only the methylase gene (BspRI: Szomolanyi et al., Gene 10:219-225, (1980); BcnI: Janulaitis et al., Gene 20:197-204 (1982); BsuRI: Kiss and Baldauf, Gene 21:111-119, (1983); and MspI: Walder et al., J. Biol. Chem. 258:1235-1241, (1983)).

A more recent method, the "endo-blue method", has been described for direct cloning of restriction endonuclease genes in E. coli based on the indicator strain of E. coli containing the dinD::lacZ fusion (Fomenkov et al., U.S. Pat. No. 5,498,535, (1996); Fomenkov et al., Nucl. Acids Res. 22:2399-2403, (1994)). This method utilizes the E. coli SOS response following DNA damages caused by restriction endonucleases or nonspecific nucleases. A number of thermostable nuclease genes (TaqI, Tth111I, BsoBI, Tf nuclease) have been cloned by this method (U.S. Pat. No. 5,498,535, 1996).

Because purified restriction endonucleases, and to a lesser extent, modification methylases, are useful tools for creating recombinant molecules in the laboratory, there is a commercial incentive to obtain bacterial strains through recombinant DNA techniques that produce these enzymes in large quantities. Such overexpression strains would also simplify the task of enzyme purification.

BsrFI restriction enzyme was purified from Bacillus stearothermophilus to near homogeneity. The protein was sequenced to obtain its N-terminus amino acid (aa) sequence. A set of degenerate primers were synthesized based on the aa sequence. The first 18 codons encoding BsrFI restriction endonuclease (bsrFIR) were amplified by PCR and its coding sequence obtained. Inverse PCR was attempted to clone the adjacent coding sequence. It failed to amplify any new DNA sequences. Therefore the methylase selection method was used to clone the BsrFI methylase gene (bsrFIM). After BsrFI methylase selection (digestion of an ApoI partial library with BsrFI and retransformation), two clones were found to be resistant to BsrFI digestion. The entire insert in one clone was sequenced and the insert encodes the BsrFI methylase (M.BsrFI). In addition, a small truncated open reading frame adjacent to the methylase gene was found to have homology to the Cfr10I restriction endonuclease in a BlastX homology search in the GenBank database. BsrFI and Cfr10I are isoschizomers that recognize and cleave 5'R CCGGY3'. Two primers were used to amplify the bsrFIR gene, The forward primer is a degenerate primer designed from the N-terminal aa sequence and the reverse primer is the bona fide sequence derived from the BsrFI methylase⁺ clone. The bsrFIR gene was amplified by PCR, ligated into a T7 expression vector pET21at and the ligated DNA was transformed into premodified cells ER2566 [pLG339-BsrFIM]. The final expression strain is ER2566 [pLG339-BsrFIM, pET21at-BsrFIR]. Recombinant BsrFI activity was detected in E. coli cell extract. BsrFI is cloned from a thermophile Bacillus stearothermophilus. Thus, BsrFI is a thermostable enzyme and it is active at 37° C. to 65° C. The isoschizomer Cfr10I was cloned from a mesophilic microorganism Citrobacter freundii RFL10.

FIG. 1. Gene organization of BsrFI restriction-modification system.

FIG. 2. DNA sequence of bsrFIM gene (SEQ ID NO:1) and its encoded amino acid sequence (SEQ ID NO:2).

FIG. 3. DNA sequence of bsrFIR gene (SEQ ID NO:3) and its encoded amino acid sequence (SEQ ID NO:4).

FIG. 4. Restriction digestion using E. coli cell extract containing recombinant BsrFI restriction endonuclease. Lanes 2, 6, and 10 were native BsrFI purified from Bacillus stearothermophilus. Lanes 3-5, 7-9 were recombinant BsrFI. Lanes 2-5 reactions were carried out at 55° C. Lanes 6-9 reactions were carried out at 65° C. Lane 10, reaction was carried out at 37° C. Lane 1, 1 kb DNA size marker (New England Biolabs, Inc., Beverly, Mass.).

The method described herein by which the BsrFI methylase gene and the BsrFI restriction endonuclease genes are preferably cloned and expressed in E. coli using the following steps:

1. Preparation of genomic DNA.

Genomic DNA is prepared from Bacillus stearothermophilus (this strain is in the New England Biolabs' collection, NEB #695, Beverly, Mass.) by standard procedures.

2. Purification of the native BsrFI restriction enzyme

The native BsrFI restriction enzyme is purified to near homogeneity by chromatography through DNA affinity columns and ion-exchange columns. The size of the native protein is approximately 35 kDa. Purified BsrFI is subjected to electrophoresis and electroblotted to a membrane and subjected to sequential degradation on an Applied Biosystems model 407A protein sequencer.

3. PCR amplification of DNA coding sequence for N-terminal amino acid.

The N-terminus amino acid sequence is used to design three degenerate primers. The coding sequence is amplified by PCR, cloned in pUC19 and sequenced. The N-terminus coding sequence (18 codons) is obtained. Inverse PCR is carried out to amplify the upstream and downstream DNA sequences. If there are PCR products, sequence the PCR products and walk further by inverse PCR to find the methylase gene upstream. If there are no PCR products, go to step 4.

4. Construction of an ApoI partial genomic DNA library.

Bacillus stearothermophilus genomic DNA is digested with ApoI to achieve the desired partial digestion. The ApoI partially digested genomic DNA in the range of 2-10 kb is ligated into EcoRI cut and CIP treated vector pRRS at 16° C. overnight. Electroporation is carried out using RR1 competent cells and the ligated DNA. Transformants are pooled and amplified. Plasmid DNA is prepared from overnight cell cultures. The ApoI partial library DNA is digested with BsrFI at 37° C. overnight. The digested DNA is used to re-transform RR1 competent cells. Plasmid DNA is isolated and individual plasmid DNA is digested with BsrFI to detect resistance to digestion. Two plasmid isolates display resistance to BsrFI digestion.

5. Sequencing of the insert carrying BsrFI methylase gene

The insert in one plasmid is sequenced by primer walking. Two ApoI fragments and two BamHI-NdeI fragments are subcloned in pUC19 and sequenced. Primers are synthesized to sequence the non-overlapping region or to confirm the complementary strand of the known sequence. The entire insert is 1714 bp, encoding the BsrFI methylase and the C-terminal portion of BsrFI endonuclease (186 bp). BsrFI methylase has 343 amino acid residues with predicted molecular mass of 39,474 daltons (39.5 kDa). The bsrFIM gene DNA sequence and its predicted aa sequence are shown in FIG. 2.

6. Expression of bsrFIM gene in E. coli

PCR is performed to amplify the bsrFIM gene. The PCR product is digested with BamHI and ligated into BamHI cut and CIP treated pLG339 vector. The ligated DNA is transformed into RR1 competent cells. Transformants are plated on Km plates (50 μg/ml). Plasmids are screened for methylase gene inserts and two clones are found to be resistant to BsrFI digestion. The resistant plasmids are transformed into a T7 expression host ER2566.

7. Cloning and expression of bsrFIR gene in E. coli

The N-terminus coding sequence (18 codons) is derived by PCR as described in step 3. The C-terminal coding sequence (186 bp) is obtained from the M.BsrFI.sup.+ clone. Two primers are used to amplify the bsrFIR gene directly from the genomic DNA. The PCR product is gel-purified from a low-melting agarose gel and sequenced directly. The bsrFIR gene is found to be 882 bp, encoding a protein with predicted molecular mass of 33,843 daltons (33.8 kDa). The gene organization of BsrFI R-M system is shown in FIG. 1.

Although the entire bsrFIR gene sequence is known, the 3rd to 7th codons are still degenerate. To obtain the bona fide sequence, inverse PCR primers are synthesized. Bacillus stearothermophilus genomic DNA is digested with ApoI, BspHI, HhaI, MfeI, NlaIII, or Sau3AI. The digested DNA is ligated at a low DNA concentration (2 μg/ml). Inverse PCR reactions are carried out using the above ligated genomic DNA and primers under condition of 95° C. 1', 55° C. 1', 72° C. 2' for 30 cycles. PCR products are found in ApoI, HhaI, NlaIII, and Sau3AI digested and self-ligated DNA. The PCR products are gel-purified and sequenced. The actual sequence for the first seven codons of bsrFIR gene is obtained.

To overexpress bsrFIR gene in E. coli, the bsrFIR gene is amplified by PCR. The PCR product is digested with NdeI and BamHI and ligated into BamHI-NdeI digested pET21at. The ligated DNA is transformed into ER2566 [pLG339-BsrFIM] premodified cells. Clones with PCR inserts are cultured in LB plus Amp (100 μg/ml) and Km (50 μg/ml) to late log phase and BsrFI protein production is induced by addition of IPTG to a final of 0.5 mM for 3 h. Cells are harvested by centrifugation and lysed by sonication. Cell debris is removed and cell extracts are used to assay BsrFI endonuclease activity. To demonstrate that BsrFI is a thermostable restriction enzyme, the DNA substrate pBR322 is digested with cell extract at 55° C. and 65° C., which gives rise to the same restriction pattern as the one digested at 37° C. (FIG. 4).

The present invention is further illustrated by the following Example. The Example is provided to aid in the understanding of the invention and is not construed as a limitation thereof.

The references cited above and below are herein incorporated reference.

1. Preparation of genomic DNA

Genomic DNA was prepared from Bacillus stearothermophilus (this strain is in the New England Biolabs' collection, NEB #695, Beverly, Mass.) by the standard procedure with the following steps: 1. cell lysis with lysozyme, Triton X-100, and SDS; 2. phenol-CHCl₃ and CHCl₃ extractions; 3. ethanol precipitation; 4. dialysis in TE buffer (change buffer 4 times); 5. RNase A treatment; 6. ethanol precipitation and resuspend genomic DNA in TE buffer.

2. Purification of the native BsrFI restriction enzyme

The native BsrFI restriction enzyme was purified to near homogeneity by chromatography through DNA affinity columns and ion-exchange columns. The size of the native protein is approximately 35 kDa on SDS-PAGE. Purified BsrFI was subjected to electrophoresis and electroblotted to a membrane. The membrane was stained with Coomassie brilliant blue R-250, and the protein band of approximately 35 kDa was excised and subjected to sequential degradation on an Applied Biosystems model 407A prtoein sequencer. The N-terminus has the following amino sequence:

MMTELKNSNXIEEYQENGKTKV (X=unknown amino acid) (SEQ ID NO:5).

3. PCR amplification of the N-terminal DNA coding sequence

The amino acid sequence MMTELKN was used to design the following two degenerate primers:

5'ATGATGACNGARTTRAARAA 3' 166-170 (N=A,C,G,T; R=A,G). (SEQ ID NO:6)

5'ATGATGACNGARCTNAARAA 3' 166-171 (SEQ ID NO:7) (Leu codons could be TTR or CTN)

The amino acid sequence EYQENGK (SEQ ID NO:8) was used to synthesize the following degenerate primer:

5'TTNCCRTTYTCYTGRTAYTC 3' 166-172 (SEQ ID NO:9) (This primer is a complement of 5'GARTAYCARGARAAYGGNAA 3' (SEQ ID NO:10)).

Primers 166-170 and 166-172 in reaction 1, primers 166-171 and 166-172 in reaction 2 were used to amplify the first 18 codons of bsrFIR gene. The following PCR conditions were used: 95° C. 1', 40° C. 1', 72° C. 30" for 30 cycles. The PCR product was gel-purified from a 3.5% low-melting agarose gel. After β-agarase treatment and ethanol precipitation, the PCR product was treated with T4 polynucleotide kinase and ligated to HincII cut and CIP treated pUCl9. Plasmids with the correct size inserts were found and the inserts were sequenced with pUC19 universal primers. Three different coding sequences were obtained, each coding for the correct aa sequence.

These three sequences are:

5'ATGATGACAGA GTTGAAGAATAGTAATTGCATTGAAGAATACCAGGAGAACGG CAA3' (SEQ ID NO:11)

M M T E L K N S N C I E E Y Q E N G X (SEQ ID NO:12)

5' ATGATGACGGAGTTAAAGAATAGTAATTGCATTGAAGAGTACCAGGAGAACGG CAA3' (SEQ ID NO:13)

M M T E L K N S N C I E E Y Q E N G X (SEQ ID NO:14)

5'ATGATGACAGAACTGAAGAATAGTAATTGCATTGAAGAGTATCAAGAGAACGG CAA3' (SEQ ID NO:15)

M M T E L K N S N C I E E Y Q E N G X (SEQ ID NO:16)

(The underlined bases are degenerate). Amongst all three sequences, codons 6 to 12 are the actual sequence amplified from the genomic DNA.

Two inverse PCR primers were designed to amplify the adjacent DNA. The inverse PCR primers have the following sequence:

5'GARTAYCARGAGAACGGCAA 3' (SEQ ID NO:17)

5'TTCAATGCAATTACTATTCTT 3' (SEQ ID NO:18)

Bacillus stearothermophilus genomic DNA was digested with AatII, AflII, AflIII, ApaLI, ApoI, BamHI, BglII, BspDI, BspEI, BspHI, ClaI, or EcoRI. The digested DNA was then extracted by equal volumes of phenol-CHCl₃ and CHCl₃, precipitated with cold ethanol, dried and resuspended in TE buffer. Two μg of the DNA was self-ligated with T4 DNA ligase and then extracted with phenol-CHCl₃ and CHCl₃, precipitated with cold ethanol. Inverse PCR reactions were carried out using the above ligated genomic DNA and primers under condition of 95° C. 1', 55° C. 1', 72° C. 2' for 30 cycles. No PCR products were detected. As described later, the N-terminal coding region would be used to amplify the entire bsrFIR gene once the C-terminus coding region had been cloned.

4. Construction of an ApoI partial genomic DNA library

Five μg of Bacillus stearothermophilus genomic DNA was digested with 2, 1, 0.5, 0.25 and 0.125 units of ApoI at 50° C. for 30 min. The ApoI partially digested genomic DNA in the range of 2-10 kb was gel-purified. The ApoI partially digested genomic DNA was ligated into EcoRI cut and CIP treated vector pRRS at 16° C. overnight. After the ligation reaction, the ligase was inactivated at 65° C. for 30 min. The ligated DNA was dialyzed by drop dialysis on a membrane on top of 2 liters of sdH₂ O. The DNA was transfered into RRl (TonA^-, DnaseI^-) competent cells by electroporation. Transformants were plated on LB agar plus Amp (100 μg/ml). About 10,000 colonies were obtained in the electroporation. All the transformants were pooled and inoculated into 1 liter of LB broth plus Amp and incubated at 37° C. overnight. Plasmid DNA was prepared from the overnight cells by Qiagen Maxi-prep column.

5. Challenge the ApoI partial library DNA with BsrFI digestion and cloning of BsrFI methylase gene (bsrFIM)

One and 3 μg of the ApoI partial library DNA was digested with 30 units of BsrFI at 370C overnight. The digested DNA was used to re-transform RRl (TonA^-, DnaseI^-) competent cells. About 125 survivors were obtained. Plasmid DNA was prepared from 1.5 ml cell culture of 36 transformants. Individual plasmid DNA was digested with BsrFI to detect any resistance to digestion. Two plasmids isolated, #6 and #31 displayed resistance to BsrFI digestion, suggesting that the cloned BsrFI methylase gene was completely cloned and expressed in E. coli (sequencing the insert verified that the entire BsrFI methylase gene was cloned, see below).

6. Sequencing of the insert carrying BsrFI methylase gene

The insert in #6 plasmid was sequenced by primer walking using AmpliTaq DNA polymerase dideoxy terminator sequencing kit and ABI373A automated DNA sequencer. Primers were synthesized to sequence the non-overlapping region or to confirm the complementary strand of the known sequence. Two ApoI fragments and two BamHI-NdeI fragments were subcloned in pUC19 and sequenced. The entire insert is 1714 bp, encoding the BsrFI methylase and the C-terminal portion of BsrFI endonuclease (186 bp). BsrFI methylase contains 343 aa residues with predicted molecular mass of 39,474 daltons (39.5 kDa). The gene organization of BsrFI R-M system is shown in FIG. 1. The bsrFIM gene DNA sequence and its predicted aa sequence are shown in FIG. 2.

7. Expression of bsrFIM gene in E. coli

PCR was performed to amplify the bsrFIM gene. BamHI sites were engineered into the two primers at the 5' ends. The PCR primers have the following sequence:

5'GTGGGATCCGCATGCGGAGGTAAAAAAATGATAAAAGTGGCATCATTATTTTCT3' (200-67) (SEQ ID NO:19)

5'GTGGGATCCGCATGCTTATTTGATTTCAAGAAGTTGCTTATTTGTTTT3' (200-68). (SEQ ID NO:20)

PCR reactions were performed in a total volume of 100 μl using 0.2 μg genomic DNA, 10 μl 10× Thermopol buffer, 0.27 mM concentration of DNTP, 79 μl H₂ O, 0.24 μg of primer 200-67 and 200-68 (2.4 μg/ml final), 2 units of Vent® DNA polymerase, with 2, 4, 6, and 8 mM Mg⁺⁺. PCR was performed at 95° C. 1', 60° C. 1', and 72° C. for 1.5' for 20 cycles. The PCR product was digested with BamHI and ligated into BamHI cut and CIP treated pLG339 vector (pSC101 replication origin). The ligated DNA was transformed into RR1 competent cells. Transformants were plated on Km plates (50 μg/ml). Eighteen plasmids were screened for inserts and two clones were found to be completely resistant to BsrFI digestion.

8. Cloning and expression of bsrFIR gene in E. coli

The N-terminus coding sequence (18 codons) of bsrFIR gene had been derived by PCR as described in section 3. The C-terminal coding sequence (186 bp) of bsrFIR gene was obtained from the M.BsrFI+clone. Two primers were used to amplify the bsrFIR gene directly from the genomic DNA. The primers have the following sequence:

5'ATGATGACNGARTTRAARAA3' 166-170 (SEQ ID NO:21)

5'TGTGGATCCTCCTGTATAATACCAAGAGGCTTA3' 199-108 (SEQ ID NO:22)

PCR was performed at 95° C. 1', 50° C. 1', 72° C. 1' for 30 cycles. PCR product was gel-purified from a low-melting agarose gel and sequenced directly using primers 166-170 or 199-108. The bsrFIR gene was found to be 882 bp, encoding a protein with molecular mass of 33,843 daltons (33.8 kDa). The bsrFIR gene sequence and its encoded aa sequence are shown in FIG. 3.

Although the entire bsrFIR gene sequence is known, the 3rd to 7th codons were still degenerate. To obtain the bona fide sequence, inverse PCR primers were synthesized. The primers have the following sequence:

5'AAGTTCGATTAATGCGTTAAATGG3' (207-56) (SEQ ID NO:23)

5'TATGATAATCAGATTCCAACAGGA3' (207-57) (SEQ ID NO:24)

Bacillus stearothermophilus genomic DNA was digested with ApoI, BspHI, HhaI, MfeI, NlaIII, or Sau3AI. The digested DNA was ligated at a low DNA concentration (2 μg/ml). Inverse PCR reactions were carried out using the above ligated genomic DNA and primers under condition of 95° C. 1', 55° C. 1', 72° C. 2' for 30 cycles. PCR products were found in ApoI, HhaI, NlaIII, and Sau3AI digested and self-ligated DNA. The PCR products were gel-purified from a low-melting agarose gel and sequenced directly using primers 207-56 and 207-57. The degenerate sequence and the actual sequence for the first seven codons of bsrFIR gene are shown below:

5'ATGATGACNGARTTRAARAA3' (166-170) (SEQ ID NO:25)

5'ATGATGACNGARCTNAARAA3' (166-171) (SEQ ID NO:26)

5'ATGATGACCGAACTTAAAAA3' (SEQ ID NO:27) (bona fide coding sequence)

M M T E L K N (SEQ ID NO:28)

To overexpress bsrFIR gene in E. coli, two primers with the following sequences were synthesized:

5'GGAGGAGTCCATATGATGACCGAACTGAAAAACTCCAACTGCATTGAAGAGTA

T3' (208-128 (SEQ ID NO:29) (Underlined bases are base substitutions to increase codon usage in E. coli. Leu codon, CTT to CTG; Asn codon, AAT to AAC; Ser codon, AGT to TCC, Asn codon AAT to AAC)

5'GGAGGATCCTTATTTTTTTATAATTTGATCCAACATTTTGTCAAT3' (203-96). (SEQ ID NO:30)

PCR reactions were performed using 0.2 μg genomic DNA, 10 μl 10 X Thermopol buffer, 0.27 mM concentration of DNTP, 79 μl H₂ O, 0.24 μg of primers 208-128 and 20396, 2 units of Vent® DNA polymerase, with 2, 4, 6, 8, and 10 mM Mg⁺⁺ in a total volume of 100 μl. PCR thermocycling conditions were 95° C. 1', 60° C. 1', and 72° C., 1' for 20 cycles. The PCR product was digested with NdeI and BamHI and ligated into BamHI-NdeI digested pET21at. The ligated DNA was transformed into ER2566 [pLG339-BsrFIM] premodified cells. Eighteen transformants were screened for inserts. Nine out of 18 had PCR insert. These nine clones were cultured in LB plus Amp (100 μg/ml) and Km (50 μg/ml) to late log phase and BsrFI protein production was induced by addition of IPTG to a final of 0.5 mM for 3 h. Cells were harvested by centrifugation and lysed by sonication. Cell debris was removed by centrifugation and cell extracts were used to assay BsrFI endonuclease activity. Five out of nine extracts displayed BsrFI activity. To demonstrate that BsrFI is a thermostable restriction enzyme, the DNA substrate pBR322 was digested with 0.5, 1 and 2 μl of cell extact at 55° C. and 65° C., which gave rise to the same restriction pattern as the one digested at 37° C. (see FIG. 4 for details). The E. coli strain ER2566 [pLG339-BsrFIM, pET21at-BsrFIR] has been deposited with the American Type Culture Collection on May 7, 1999 and received Accession No. PTA-28.