Gene, 163 (1995) 109-113 ©1995 Elsevier Science B.V. All rights reserved. 0378-1119/95/$09.50
109
GENE 09143
Cloning of a gene from Thermus filiformis and characterization of the thermostable nuclease (DNase; SOS response; DNA damage; unidirectional deletion)
Alexey Fomenkov* and Shuang-yong Xu New England Biolabs Inc., Beverly, MA 01915, USA Received by J.A. Engler: 17 February 1995; Revised/Accepted 2 May/3 May 1995; Received at publishers: 15 June 1995
SUMMARY
A gene coding for a thermostable nuclease was cloned from the thermophilic microorganism, Thermusfiliformis (Tf), using an indicator strain containing a dinD::lacZ fusion. The gene, designated nucl 7, has been mapped within a 2300-bp fragment. The 55-kDa Tf nuclease was purified to over 95% homogeneity. Single-stranded (ss) DNA is the preferred substrate for the Tfnuclease, although double-stranded (ds) DNA can also be digested. Nuclease activity increases with increasing temperature up to 80°C and requires the metal ions Ca ++ or Mg +÷ for catalysis. Tf nuclease is primarily an endonuclease that leaves 5' phosphates in the digested products. The ssDNA extensions remaining after exonuclease III digestion of dsDNA can be removed by the Tf nuclease, making it a useful reagent to generate unidirectional deletions.
INTRODUCTION
Nucleases have important biological functions in DNA replication, genetic recombination, DNA repair, and DNA restriction (Walker, 1985; Beese and Steitz, 1991; Roberts and Halford, 1993; Heitman, 1993). They are also indispensable tools for manipulating DNA and RNA. We are interested in cloning genes coding for thermostable restriction endonucleases and nucleases which are useful for DNA amplification (strandCorrespondence to: Dr. S.-y. Xu, New England Biolabs, Inc., 32 Tozer Road, Beverly, MA 01915-5599, USA Tel. (1-508) 927-5054, ext. 287; Fax (1-508) 921-1350; e-mail:
[email protected] * Present address: Department of Biology, Johns Hopkins University, Baltimore, MD 21218-2685, USA. Tel. (1-410) 516-6147. Abbreviations: Ap, ampicillin; bp, base pair(s); CIP, calf intestinal alkaline phosphatase; din, DNA damage inducible gene; ds, double strand(ed), DTT, dithiothreitol; ENase, restriction endonuclease; ExolII, exonuclease III; nucl 7, gene encoding Tf nuclease; ORF, open reading frame; PA, polyacrylamide, PAGE, PA gel electrophoresis; R, resistance; SDS, sodium dodecyl sulfate; ss, single strand(ed); Tf Thermus filiformis; XGal, 5-bromo-4-chloro-3-indolyl-15-o-galactopyranoside; ::, novel junction (fusion or insertion). SSDI 0378-1119(95)00426-2
displacement amplification) and for studying DNAthermostable protein interactions. The dinD1 is a DNA damage inducible gene to which a lacZ fusion has been constructed (Kenyon and Walker, 1980; Heitman et al., 1989). The dinD::lacZ fusion was used to study EcoRI endonuclease mutants in the SOS induction assay (Heitman et al., 1989; Heitman and Model, 1990). More recently, the dinD::lacZ fusion strain has been used for direct cloning into Escherichia coli of genes which code for thermostable restriction enzymes (Fomenkov et al., 1994). In this paper we describe the use of this method to clone a gene which encodes a non-specific thermostable Tfnuclease. The Tfnuclease was purified to homogeneity. Its properties are. reported here.
EXPERIMENTAL AND DISCUSSION
(a) Cloning of the Tfnuclease gene, n u d 7, and its restriction mapping The Sau3AI partially digested Tf genomic DNA was ligated to BamHI-digested and CIP-treated pBR322
110 D N A . The ligated D N A was used to transform ER1992 (dinD::lacZ, mcrA, mcrBC, mrr; F o m e n k o v et al., 1994) competent cells and the transformants were plated on XGal indicator plates with Ap. Twenty-three blue colonies were found a m o n g approximately 8000 Ap R transformants. W h e n cell extracts from each clone were examined for nuclease activity, D N A nicking activity of d s D N A was detected in isolate #17 (data not shown). Plasmid D N A from #17 was isolated and analyzed by restriction digests. Deletion m a p p i n g revealed that the Tfnuclease encoding gene, nucl 7, was located within the smaller 2.3-kb BamHI-NcoI fragment (data not shown).
1 2
kDa 66.4-55.6--
"~--Tf nuclease
42.7--
(b) Purification of the Tfnuclease The Tf nuclease was purified by heat denaturation of E. coli proteins at 68°C for 30 min, removal of denatured proteins by centrifugation and by heparin-Sepharose c o l u m n c h r o m a t o g r a p h y . Fig. 1 shows that the Tf nuclease was purified to homogeneity, and its molecular mass is estimated to be 55 kDa. A 55-kDa protein would be expected to have an O R F a b o u t 1500 nt. The specific activity of the nuclease increased 100-fold after the twostep purification (data not shown).
(c) Substrate specificity and metal ion requirement The E. coli [ 3 H ] d s D N A and heat-denatured E. coli [3HI s s D N A were incubated with Tfnuclease at 70°C for 15 min in the presence of divalent cations Ca ++ , M g ++ , or Zn + +, and the release of acid soluble nucleotides was measured. 52% of heat-denatured E. coli [ 3 H ] s s D N A was solubilized as c o m p a r e d to 12% release of E. coli [ 3 H ] d s D N A in Ca ++ buffer. In M g ++ buffer, 15% of 3H label was solubilized from [ 3 H ] s s D N A and 3% tritium was released from [-3HI d s D N A . B a c k g r o u n d levels of 3H release were observed in the Zn + + buffer. It is concluded that s s D N A is a better substrate for the Tf nuclease and that Ca ++ is the preferred cofactor, although the nuclease still displays activity with M g ++ N o nuclease activity was observed in the Zn + + buffer on either ss or d s D N A . Consistent with this result, when Tf nuclease was incubated with M13 ssDNA, it had approximately 2.5 times more activity in Ca ++ buffer than in M g ++ buffer (Fig. 2A). W h e n M13 s s D N A was mixed with M13 d s D N A and the mixture then subjected to Tf nuclease digestion, the s s D N A was cleaved first (Fig. 2B, lanes 5, 6, and 7). In these reactions, Tfnuclease showed highest activity in Ca + + buffer, and no activity in Zn + + buffer (Fig. 2B). In M g ++ buffer, the majority of M13 supercoiled d s D N A was converted to nicked circular D N A and some to linear D N A , indicating that Tf nuclease also has endonuclease activity on d s D N A . To
Fig. 1. Migration of purified Tf nuclease in SDS-PAGE (indicated by an arrow). The estimated molecular mass is 55 kDa. Lane 1, protein size marker; Lane 2, 2 lag of the purified Tf nuclease. Methods: Tf nuclease purification procedure: wet E. coli cells (4 g) were harvested from 1 L of overnight culture and resuspended in 40 ml of sonication buffer (10 mM Tris.HC1 pH 8/10 mM [3-mercaptoethanol). Cells were lysed with lysozyme and sonication. Cell debris was removed by centrifugation for 30 min at 30 996 x g. The cleared cell extract was heated at 68°C for 30 rain. Denatured E. coliproteins were removed by centrifugation for 30 min at 30 996 x g. Tf nuclease was purified by heparinSepharose chromatography. Protein sample ( 10 lal)was taken from each fraction and used to digest M13 ssDNA or pUC19 dsDNA. The fractions containing DNA nuclease activity were pooled in a total volume of 6 ml and dialyzed against a storage buffer (10mM NaH2PO4 pH 7.0/2 mM f3-mercaptoethanol/100mM NaC1/50% glycerol). Proteins were resolved by 0.1% SDS-(10 20%)PAGE and detected by Coomassie blue staining. One unit of Tf nuclease was defined as the amount of enzyme to digest 1 lag of M13 ssDNA to small oligodeoxyribonucleotides (<250 nt) in 30 min in a buffer containing 100mM NaC1/50 mM Tris.HC1 pH 6/10 mM CaCI2/1mM DTT. further examine the nuclease activity on d s D N A , HindIIIdigested X D N A was incubated with Tf nuclease in the presence of Ca + +, M g + +, or Zn + +. In the Ca + + buffer, the linear d s D N A was degraded to small fragments (less than 700 bp, Fig. 2C, lane 2). In the presence of M g + + or Zn + +, no apparent degradation was detected (Fig. 2C, lanes 5-10). At least four times more Tf nuclease is required to digest 1 ktg d s D N A than ssDNA. Tfnuclease is an endonuclease that prefers ssDNA.
(d) Optimum temperature and pH for Tfnuclease To determine the o p t i m u m temperature for Tf nuclease, the nuclease digestion was incubated with heatdenatured E. coli [ 3 H ] D N A in temperatures ranging from 30°C to 90°C. Tf nuclease displayed moderate to high activity at temperatures ranging from 60°C to 80°C, but lost activity at 90°C (data not shown). At 30°C to 40°C the enzyme had very low activity, consistent with the in vivo observation that E. coli cells can survive trans-
111 formation by a medium-copy-number plasmid carrying the nucl 7 gene. The optimum pH for Tfnuclease activity is in the range ofpH 5 to 6, and the enzyme has maximum activity in 100 mM NaC1 (data not shown).
A.
M13 ssDNA C a ++ [2
3
Mg ++ 4115
6
Zn ++
7118
9
101
(e) Tfnuclease is an endonuclease that generates oligodeoxyribonucleotides with 5' phosphates To analyze the substrate requirements of the Tf nuclease, a 51-bp 5' end-labeled EcoRI-HindIII fragment was incubated with varying amounts of the nuclease. Only very weak bands representing degradation products were detected (Fig. 3, lanes 2-5). When the same DNA was heat-denatured and treated with the same amount of Tf nuclease, more prominent digestion products were observed (Fig. 3, lanes 6-9). The Tf nuclease does not
dSDNA substrate i
1
2
3
4
heat-denatured ssDNA 5
ii
6
7
8
~ ' : ~ ' ; ~ +
9110 .....
--
..... +-- . . . . . . . . . "°°+°
0
4
2
1 4 2 1 4 nuclease (units)
2
1
B. M13 ssDNA and dsDNA Ca++ Mg ++ Zn ++ 2 I 3 4 5116 7 8119 10 111
1
Sact
- Hindlll (51 nt) -Hindlll (49 nt)
Kpnl
-
EcoRI
Xmal
Hindlll (43 nt)
-
Hindlll (35 nt)
C-T-AI~-Go-'-
--
G~--A-G-A T -
- - BamHI -
Hindlll
(30 nt)
C-T --
(NC) (linear)
(ccc)
Co-.-A~--Go---Co---TO---G~--
.......
Sail
-
Hindlll (18 nt)
--
Pstl
-
Hindlll (16 nt)
--
Sphl.
GO---A -
0
4
2 1 4 2 1 nuclease (units)
4
2
1
C-G To---
C--
C. ,%(Hindlll-cleaved) dsDNA C a ++ Mg ++ Zn ++ 1 12 3 411 5 6 7118 9 10 I
4
2
1 4 2 1 4 2 nuciease (units)
1
Fig. 2. T / n u c l e a s e digestion of ss and d s D N A . A: M I 3 s s D N A substrate incubated w i t h T f n u c l e a s e in Ca + +, M g + +, or Z n + + buffer. B:
A mixture of M13 ss and d s D N A digested with T f nuclease in Ca + +, Mg + +, or Zn + + buffer. CCC, covalently-closed circular DNA; NC, nicked circular DNA. C: HindIII-digested X D N A incubated with T f nuclease in Ca + +, Mg + +, or Zn + + buffer. Electrophoresis was carried out in 0.8% agarose gel.
C-G-
"+
Hindlll (10 nt)
To-
Fig. 3. T f nuclease digestion of a 5' end-labeled D N A fragment. Lane 1: untreated 5' end-labeled DNA; Lanes 2 and 6, 3 and 7, 4 and 8, 5 and 9: 4, 2, 1 and 0.5 units of T f nuclease, respectively. The 5' endlabeled fragment was boiled for 10 min and immediately chilled on ice before nuclease digestion. The nuclease-digested products were resolved by 7 M urea/6% PAGE. Filled circles indicate enhanced cleavage of the phosphodiester bond. Open circle indicates the missing band. The identity of each base was derived from the known sequence and the 5' end-labeled size marker. Lane i0: double digestion products of the labeled fragment by the indicated pairs of ENases. Substrate preparation: plasmid pUC19 D N A was first linearized with HindIII and 5' endlabeled by T4 polynucleotide kinase and [7-32p] ATP. The labeled D N A was then cleaved with EcoRI. The small EcoRI-HindIII fragment (51 bp long, labeled at the HindIII 5' end) was purified from a 10% PA gel.
112 appear to be sequence specific although some enhanced bands are detected in the digestion pattern (indicated by filled circles). Some weak bands were also found (for example T19, indicated by an open circle), which may be the result of cleavage inhibition due to secondary structure formation. This digestion pattern indicates that the Tfnuclease is not a 5' to 3' exonuclease. If the Tfnuclease had 5' to 3' exonuclease activity, the 5' labeled end would have been digested first, and no radiolabeled fragments would be detected. Since the nuclease-generated bands migrate at the same position as the ENase-generated markers, the nuclease products probably have 3' hydroxyls. The properties of the Tf nuclease are in some respect similar to some of the endo-exonucleases found in Neurospora, Aspergillus, and Saccharomyces. The endoexonucleases have endonuclease activity on linear and circular ssDNA as well as RNA. They also convert supercoiled DNA to nicked and linear forms (Chow and Fraser, 1983; Chow and Resnick, 1987). Two other nucleases, Staphylococcal nuclease and BAL31 nuclease, may also resemble the Tf nuclease in using Ca ÷÷ as a cofactor for catalytic activity (Cotton et al., 1979; Legerski, 1978). Staphylococcal nuclease prefers ssDNA over dsDNA.
(f) Cloning of 7. DNA fragments generated by the Tf nuclease When L DNA was incubated with excess amount of the Tfnuclease, the DNA was digested to small fragments (Fig. 2C, lane 2). To determine the end sequence of the digested products, the DNA fragments were cloned into HincII-digested and CIP-treated pUC19 and sequenced. Seven isolates carried small DNA inserts, ranging from 32 to 250 bp. Only one clone contained a large insert of about 1.5 kb. The cleavage sites of Tf nuclease do not seem to be sequence specific based on the sequence junctions (data not shown). The Tf nuclease-digested products can be directly cloned into a dephosphorylated vector, indicating that the products carry 5' phosphates. Consistent with these results, the Tf nuclease-digested DNA could be dephosphorylated by CIP and 5' endlabeled with T4 polynucleotide kinase with [y-32p]ATP.
(g) Exonuclease III-Tfnuclease digestion of dsDNA E. coli exonuclease III (ExoIII) is a dsDNA-dependent 3' to 5' ssDNA exonuclease starting at blunt or recessed 3' termini, and is widely used in unidirectional deletion of dsDNA (Guo and Wu, 1982; Henikoff, 1984). We examined the possibility of using the Tf nuclease to remove remaining ssDNA 5' extensions following ExoIII digestion of dsDNA. HindIII-SphI digested pUC19 DNA was first treated with ExolII for 1 to 6 min, and then
DNA from each time point was digested with the Tf nuclease at 65°C. The nuclease-digested products were resolved in an agarose gel. A series of nested deletions were obtained (data not shown). The ExolII-Tfnuclease combination has been used successfully to generate unidirectional deletions for DNA sequencing (B. Slatko and S.-y. X. unpublished data). The advantage of using ExolII-Tf nuclease over ExolII-Mung bean nuclease combinations is that inactivation of ExolII is coupled with Tf nuclease digestion at 65°C. In addition, the Tf nuctease-encoding gene has been cloned and the nuclease can be easily purified to homogeneity. The Tf nuclease may be useful for DNase foot-printing studies to investigate thermostable DNA-binding proteins.
(h) The RecA protein is required for efficient repair of damage by the Tfnnclease in vivo The nuc17 gene was cloned in the indicator strain of dinDl::lacZ fusion based on DNA damage and SOS induction. To examine the effect of in vivo DNA damage, cells with RecA ÷ or RecA- phenotypes were transformed with a plasmid carrying the nuc17 gene (pNUC17). The transformation efficiency of the pNUC17 plasmid was 10 times less than vector alone in a RecA- strain (2.4 × 105 transformantsAtg of pNUC 17 vs. 2.0 × 106 transformants/ ~tg of pBR322). There was no difference in transformation efficiency using a RecA ÷ strain (6.0 x 10 6 transformants/ ~tg of pNUC17 vs. 5.0 x 106 transformants/lag of pBR322 vector), indicating that the DNA damage elicited by the Tf nuclease is efficiently repaired in a RecA-dependent manner, possibly by the homologous recombination repair process. When the surviving transformants of the RecA- strain carrying the pNUC17 plasmid were grown overnight and plated out, the plating efficiency was at least 103-fold lower than that of cells carrying the control plasmid pBR322, indicating that the DNA damage has a cumulative effect on the cell viability. It has been reported that phage T7.3 endonuclease production in E. coli induces the SOS response, and that synthesis of the gp3 endonuclease in a RecA- host results in extensive DNA damage and cell death (Panayotatos and Fontaine, 1985). Perhaps the DNA damage introduced by the Tfnuclease is mainly in the ssDNA regions as the result of DNA 'breathing', active transcription, or DNA replication.
(i) Use of the nuc17 gene to probe Tf, T. aquaticus and T. thermophilus genomic DNA An EagI fragment covering part of the nuc17 gene was purified from pNUC17 and labeled with [~-32p]dCTP and used as a probe to hybridize with Tf T. aquaticus, and T. thermophilus genomic DNA. Under stringent conditions (hybridization and washing at 68°C), a labeled band was detected only in Tf genomic DNA. The probe
113 weakly hybridized to T. thermophilus genomic DNA under less stringent conditions (hybridization at 68°C and washing at 55°C). The probe did not hybridize to T. aquaticus or E. coli genomic DNA under either condition (data not shown).
(j) Conclusion (1) The Tf nuclease encoding gene, nuc17, has been cloned into E. coli using a dinD::lacZ indicator strain. (2) The 55-kDa nuclease has been purified to homogeneity. (3) The enzyme is active in buffers containing Ca ++ or Mg ÷ +. The optimum temperature for nuclease activity is 80°C. (4) The Tfnuclease is a non-specific endonuclease that leaves 5' phosphoryl products. It also generates nicks in dsDNA. (5) A plasmid carrying the nuc17 gene transforms E. coli RecA- cells with reduced efficiency.
ACKNOWLEDGEMENTS
We thank William E. Jack, Richard Roberts, Ira Schildkraut and Jay Wayne for critical comments, and Elisabeth Raleigh for E. coli strains. This work was supported by New England Biolabs, Inc.
REFERENCES Beese, L.S. and Steitz T.A: Structural basis for the 3'-5' exonuclease activity of Escherichia coli DNA polymerase I: a two metal ion mechanism. EMBO J. 10 (1991) 25-33.
Chow, T.Y.-K. and Fraser, M.: Purification and properties of single strand DNA-binding endonuclease of Neurospora crassa. J. Biol. Chem. 258 (1983) 12010-12018. Chow, T.Y.-K. and Resnick, M.A.: Purification and characterization of an endo-exonuclease from Saccharomyces cerevisiae that is influenced by the RAD52 gene. J. Biol. Chem. 262 (1987) 17659-17667. Cotton, F.A., Hazen, E.E. and Legg M.J.: Staphylococcal nuclease: proposed mechanism of action based on structure of enzyme-thymidine 3',5'-bisphosphate-calcium ion complex at 1.5 A resolution. Proc. Natl. Acad. Sci. USA 76 (1979) 2551-2555. Fomenkov, A., Xiao, J.-p., Dila, D., Raleigh, E. and Xu, S.-y.: The 'endoblue method' for direct cloning of restriction endonuclease genes in E. coll. Nucleic Acids Res. 22 (1994) 2399-2403. Guo, L.H., Yang, R.C. and Wu, R.: An improved strategy for rapid direct sequencing of both strands of long DNA molecules cloned in a plasmid. Nucleic Acids Res. 11 (1983) 5521- 5540. Heitman, J.: On the origins, structures and functions of restrictionmodification enzymes. In: Setlow, J.K. tEd.) Genetic Engineering Vol. 15, 1993, pp. 57-108. Heitman, J. and Model, P.: SOS induction as an in vivo assay of enzyme-DNA interaction. Gene 103 (1991) 1 9. Henikoff, S.: Unidirectional digestion with exonuclease III creates targeted breakpoints for DNA sequencing. Gene 28 (1984) 351 359. Kenyon, C. and Walker, G.C.: DNA-damaging agents stimulate gene expression at specific loci in Escherichia coli. Proc. Natl. Acad. Sci. USA 77 (1980) 2819-2823. Legerski, R.J., Hodnett, J.L. and Gray Jr., H.B.: Extracellular nuclease of Pseudomonas BAL31, III. Use of the double-stranded deoxyriboexonuelease activity as the basis of a convenient method for the mapping of fragments of DNA produced by cleavage with restriction enzymes. Nucleic Acids Res. 5 (1978) 1445 1463. Panayotatos, N. and Fontaine, A.: An endonuclease specific for singlestranded DNA selectively damages the genomic DNA and induces the SOS response. J. Biol. Chem. 260 (1985) 3173-3177. Roberts, R.J. and Halford, S.E.: Type II restriction endonucleases. In: Linn, S., Lloyd, R.S. and Roberts, R.J. (Eds.) Nucleases. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1993, pp. 35 88. Walker, G.C.: Inducible DNA repair systems. Annu. Rev. Biochem. 54 (1985) 425-457.