Cell, Vol. 17, 175-l
84, May
Site-Specific Gyrase
1979,
Copyright
0 1979
by MIT
Cleavage
of DNA by E. coli DNA
Alan Morrison* and Nicholas R. Cozzarelli*t Departments of Biochemistry * and Biophysics and Theoretical Biologyt The University of Chicago Chicago, Illinois 60637
Summary E. coli DNA gyrase, which catalyzes the supercoiling of DNA, cleaves DNA site-specifically when oxolinic acid and sodium dodecylsulfate are added to the reaction. We studied the structure of the gyrasecleaved DNA because of its implications for the reaction mechanism and biological role of gyrase. Gyrase made a staggered cut, creating DNA termini with a free 3’ hydroxyl and a 5’ extension that provided a template primer for DNA polymerase. The cleaved DNA was resistant to labeling with T4 polynucleotide kinase even after treatment with proteinase K. Thus the denatured enzyme that remains attached to cleaved DNA is covalently bonded to both 5’ terminal extensions. The 5’ extensions of many gyrase cleavage fragments from 4X174, SV40 and Col El DNA were partially sequenced using repair with E. coli DNA polymerase I. No unique sequence existed within the cohesive ends, but G was the predominant first base incorporated by DNA polymerase I. The cohesive end sequences of four gyrase sites were determined, and they demonstrated a four base 5’ extension. The dinucleotide TG, straddling the gyrase cut on one DNA strand, provided the only common bases within a 100 bp region surrounding the cleavage sites. Analysis of other cleavage fragments showed that cutting between a TG doublet is common to most, or all, gyrase cleavages. Other bases common to some of the sequenced sites were clustered nonrandomly around the TG doublet, and may be variable components of the cleavage sequence. This diverse recognition sequence with common elements is a pattern shared with several other specific nucleic acid-protein interactions.
E. coli DNA gyrase is made up of two subunits, A and B, which can be purified separately and reconstituted to form gyrase activity (Higgins et al., 1978). Subunit A is the product of the nalA gene, which governs resistance to the related drugs nalidixic acid and oxolinic acid. Subunit B is the product of the cou gene, which governs resistance to coumermycin A, and novobiocin. Inclusion of oxolinic acid in a gyrase reaction results in the inhibition of supertwisting activity, and subsequent addition of sodium dodecylsulfate (SDS) induces double-strand DNA cleavage. This cleavage is site-specific, since gyrase cleavage of Col El DNA linearized by treatment with Eco RI nuclease produces discretely sized fragments (Gellert et al., 1977; Sugino et al., 1977, 1978). The initial products of gyrase cleavage are DNA fragments that are bound to protein (Sugino et al., 1977). These fragments have retarded mobility on gel electrophoresis and bind to nitrocellulose filters, but treatment with proteinase K abolishes these effects (Peebles et al., 1979). In addition to supercoiling closed circular DNA in the presence of ATP, gyrase relaxes supertwisted DNA in the absence of ATP. Both reactions require enzymatic breakage and reunion of the DNA backbone. The DNA cleavage caused by oxolinic acid and SDS presumably represents the abortion of an intermediate of the breakage-reunion reaction. We studied the site-specific cleavage reaction because of its intimate relation to the gyrase reaction mechanism and because of its implications regarding the biological role of gyrase. Using +X174, SV40 and Col El DNA, we found that gyrase makes staggered cuts resulting in 5’ extended DNA termini that are covalently bonded to the enzyme. The site specificity of gyrase cleavage is not governed by a simple, unique DNA sequence. Rather, gyrase cuts one of the DNA strands between a TG dinucleotide, but the dinucleotide straddling the cut 4 base pairs (bp) away on the opposing strand is apparently random. The infrequency of cleavage requires that additional determinants be present. Other bases common to some of the sites are clustered nonrandomly around the TG doublet and may be variable components of the cleavage sequence.
Introduction Results DNA gyrase catalyzes the introduction of negative superhelical turns into closed duplex DNA (Gellert et al., 1976) and is an essential enzyme in Escherichia coli (Kreuzer et al., 1978). It is involved in replication of the DNA of @Xl 74 (Marians et al., 1977), T7 (Itoh and Tomizawa, 1977) and E. coli (Drlica and Snyder, 19781, in integrative recombination of phage lambda DNA (Mizuuchi, Gellert and Nash, 19781, in DNA repair (Hays and Boehmer, 1978) and in transcription (Puga and Tessman, 1973; Shuman and Schwartz, 1975; Falco, Zivin and Rothman-Denes, 1978; Smith, Kubo and Imamoto, 1978; Peebles et al., 1978).
Gyrase-Cleaved DNA Is Resistant to 5’ Terminal Labeling Gyrase cleavage of EGO RI-generated linear Col El DNA leaves protein bonded to both resulting DNA fragments (Peebles et al., 1979). The pair of fragments arising from gyrase cleavage of linear DNA are hereafter designated “partner fragments.” We tested whether the protein attachment affected labeling of the 5’ termini of the cleaved DNA in a T4 polynucleotide kinase reaction. @Xl 74 RFI DNA was cleaved with gyrase so that there was an average of less than
Cell 176
one cut per DNA molecule. The product was treated with proteinase K to trim attached protein and centrifuged through a sucrose density gradient to remove unreacted supercoiled DNA. The gyrase-produced linear DNA, containing an approximately equal amount of contaminating nicked circular DNA, was mixed with Eco RI-cut Col El DNA as an internal reference, and incubated with polynucleotide kinase and Y-~*P-ATP under conditions favoring labeling by exchange (van de Sande, Kleppe and Khorana, 1973). Figure 1 shows densitometer tracings of the products separated by gel electrophoresis. While Eco RI-generated linear Col El DNA became labeled, gyrase-cut 4x174 DNA did not, implying that the 5’ ends created by gyrase cleavage are blocked, even after proteinase K treatment. However, gyrase-cleaved DNA provided a template primer for both E. coli DNA polymerase I (see below) and T4 DNA polymerase (data not shown), showing that gyrase cleavage leaves a free 3’ hydroxyl terminus. Fine Mapping of Gyrase Cleavage Sites Labeling of gyrase-cleaved DNA with a-32P-dNTPs and DNA polymerase provided a sensitive means for detecting cleavages. +X174 RFI DNA was cleaved with gyrase so that approximately 60% of the DNA was converted to the linear form, and then labeled with E. coli DNA polymerase I and all four common a-32P-dNTPs. The products were digested with Hha I and resolved by gel electrophoresis (Figure 2, lane 1). Densitometer tracings showed that the bands varied in intensity over at least a 20 fold range. We chose to look at gyrase cleavage sites in Hha l-digested +X174 DNA for the following reasons. First, the relatively small sizes of the resulting gyrase cleavage products could be measured to within about 20 bp. Second, the two largest Hha I fragments of $X174 DNA are 1553 and 640 bp long, so that any gyrase cleavage product migrating between these two fragments must have arisen from the 1553 bp Hha I-A fragment. Thus at least one of each pair of fragments resulting from every gyrase cleavage event within the Hha I-A fragment could be resolved without the necessity of purifying the restriction fragments. Third, the $X1 74 Hha I-A fragment, represented diagramatically at the bottom of Figure 2, contains the single Ava II site of +X174 DNA, so that a Hha I/Ava II double digest of gyrase-cleaved DNA (Figure 2, lane 2) allowed the mapping of many gyrase cleavage fragments. Thus the a and a’ bands are partner fragments arising from gyrase cleavage at the a site; this is also the most prominent cleavage site in whole +X174 DNA (Peebles et al., 1978). Columns l-4 in Table 1 summarize the data defining the map positions of the gyrase a and b sites within the +X174 Hha I-A fragment. The same experimental procedure was applied to
Figure 1. Gyrasa-Cleaved Polynucleotide Kinase
@Xl 74 DNA
Is Not
a Substrate
for T4
A mixture containing approximately equal amounts of nicked circular $X1 74 DNA, linear @Xl 74 DNA formed by gyrase cleavage of @Xl 74 RFI DNA and linear duplex Col El DNA generated by Eco RI nuclease digestion was incubated with Y-~*P-ATP and T4 polynucleotide kinase to label the 5’ DNA termini by exchange. The products were purified by gel filtration and electrophoresed through a 1% agarose gel. Microdensitometer tracings were taken of an autoradiograph (broken line) and a photograph of the ethidium bromide-stained gel (solid line). (A) nicked circular +X174 DNA; (B) Eco RI-generated linear Col El DNA: 62) gyrase-produced linear +X1 74 DNA. Electrophoresis was from left to right.
gyrase-cleaved SV40 DNA, except that the restriction enzyme Hae Ill was used. The Hae Ill digest of gyrasecut SV40 DNA shown in Figure 2, lane 3, revealed four strong bands, labeled a, a’, b and b’, arising from cleavage at the a and b sites. The 1661 bp SV40 Hae Ill-A fragment, represented diagramatically at the bottom of Figure 2, contains an Ava II site and the single Taq I site of SV40 DNA. There are no Ava II sites in the 752 bp Hae Ill-B or the 540 bp Hae Ill-C fragments of SV40 DNA. Double digestion of gyrase-cut SV40 DNA with Hae Ill and Ava II (Figure 2, lane 4) or Hae Ill and Taq I (Figure 2, lane 5) therefore allowed the mapping of the SV40 a and b gyrase sites within the Hae Ill-A fragment. The mapping data are summarized in Table 1, columns l-4. DNA Polymerase I Repair of Gyrase Cleavage Sites The labeling of gyrase cleavage fragments by E. coli DNA polymerase I (Figure 2) implied that gyrase cleavage creates a cohesive end with a recessed 3’ terminal hydroxyl. To analyze the base sequences of these cohesive ends, @Xl 74 DNA was cleaved with gyrase and then incubated with DNA polymerase I. The polymerase reactions contained either a single (Y--~‘PdNTP to label the first nucleotide, or a combination of one o(-32P-dNTP and three unlabeled dNTPs to determine the composition of the repaired cohesive ends. The reaction products were digested with Hha I and separated by gel electrophoresis (Figure 3).
Site-Specific
Cleavage
of DNA by DNA Gyrase
177
1 2
P C+
1000
nucleotide A*
T*
composition G’
C+
A*
t
T*
G*
-b
-a - a’ aa’-
- b!
b’-
5ooL
J a eix174
I
I
,
4000
,
,
,
'
, 31500
Figure 2. Gyrase A Fragments
Cleavage
I
b
AV~ II
sv40
I
,'
b
'8
'4
4’500
Figure 3. Labeling of Gyrase E. coli DNA Polymerase I 500
AYB II
8
'
5’000
a I
',
1
84, May
Site-Specific Gyrase
1979,
Copyright
0 1979
by MIT
Cleavage
of DNA by E. coli DNA
Alan Morrison* and Nicholas R. Cozzarelli*t Departments of Biochemistry * and Biophysics and Theoretical Biologyt The University of Chicago Chicago, Illinois 60637
Summary E. coli DNA gyrase, which catalyzes the supercoiling of DNA, cleaves DNA site-specifically when oxolinic acid and sodium dodecylsulfate are added to the reaction. We studied the structure of the gyrasecleaved DNA because of its implications for the reaction mechanism and biological role of gyrase. Gyrase made a staggered cut, creating DNA termini with a free 3’ hydroxyl and a 5’ extension that provided a template primer for DNA polymerase. The cleaved DNA was resistant to labeling with T4 polynucleotide kinase even after treatment with proteinase K. Thus the denatured enzyme that remains attached to cleaved DNA is covalently bonded to both 5’ terminal extensions. The 5’ extensions of many gyrase cleavage fragments from 4X174, SV40 and Col El DNA were partially sequenced using repair with E. coli DNA polymerase I. No unique sequence existed within the cohesive ends, but G was the predominant first base incorporated by DNA polymerase I. The cohesive end sequences of four gyrase sites were determined, and they demonstrated a four base 5’ extension. The dinucleotide TG, straddling the gyrase cut on one DNA strand, provided the only common bases within a 100 bp region surrounding the cleavage sites. Analysis of other cleavage fragments showed that cutting between a TG doublet is common to most, or all, gyrase cleavages. Other bases common to some of the sequenced sites were clustered nonrandomly around the TG doublet, and may be variable components of the cleavage sequence. This diverse recognition sequence with common elements is a pattern shared with several other specific nucleic acid-protein interactions.
E. coli DNA gyrase is made up of two subunits, A and B, which can be purified separately and reconstituted to form gyrase activity (Higgins et al., 1978). Subunit A is the product of the nalA gene, which governs resistance to the related drugs nalidixic acid and oxolinic acid. Subunit B is the product of the cou gene, which governs resistance to coumermycin A, and novobiocin. Inclusion of oxolinic acid in a gyrase reaction results in the inhibition of supertwisting activity, and subsequent addition of sodium dodecylsulfate (SDS) induces double-strand DNA cleavage. This cleavage is site-specific, since gyrase cleavage of Col El DNA linearized by treatment with Eco RI nuclease produces discretely sized fragments (Gellert et al., 1977; Sugino et al., 1977, 1978). The initial products of gyrase cleavage are DNA fragments that are bound to protein (Sugino et al., 1977). These fragments have retarded mobility on gel electrophoresis and bind to nitrocellulose filters, but treatment with proteinase K abolishes these effects (Peebles et al., 1979). In addition to supercoiling closed circular DNA in the presence of ATP, gyrase relaxes supertwisted DNA in the absence of ATP. Both reactions require enzymatic breakage and reunion of the DNA backbone. The DNA cleavage caused by oxolinic acid and SDS presumably represents the abortion of an intermediate of the breakage-reunion reaction. We studied the site-specific cleavage reaction because of its intimate relation to the gyrase reaction mechanism and because of its implications regarding the biological role of gyrase. Using +X174, SV40 and Col El DNA, we found that gyrase makes staggered cuts resulting in 5’ extended DNA termini that are covalently bonded to the enzyme. The site specificity of gyrase cleavage is not governed by a simple, unique DNA sequence. Rather, gyrase cuts one of the DNA strands between a TG dinucleotide, but the dinucleotide straddling the cut 4 base pairs (bp) away on the opposing strand is apparently random. The infrequency of cleavage requires that additional determinants be present. Other bases common to some of the sites are clustered nonrandomly around the TG doublet and may be variable components of the cleavage sequence.
Introduction Results DNA gyrase catalyzes the introduction of negative superhelical turns into closed duplex DNA (Gellert et al., 1976) and is an essential enzyme in Escherichia coli (Kreuzer et al., 1978). It is involved in replication of the DNA of @Xl 74 (Marians et al., 1977), T7 (Itoh and Tomizawa, 1977) and E. coli (Drlica and Snyder, 19781, in integrative recombination of phage lambda DNA (Mizuuchi, Gellert and Nash, 19781, in DNA repair (Hays and Boehmer, 1978) and in transcription (Puga and Tessman, 1973; Shuman and Schwartz, 1975; Falco, Zivin and Rothman-Denes, 1978; Smith, Kubo and Imamoto, 1978; Peebles et al., 1978).
Gyrase-Cleaved DNA Is Resistant to 5’ Terminal Labeling Gyrase cleavage of EGO RI-generated linear Col El DNA leaves protein bonded to both resulting DNA fragments (Peebles et al., 1979). The pair of fragments arising from gyrase cleavage of linear DNA are hereafter designated “partner fragments.” We tested whether the protein attachment affected labeling of the 5’ termini of the cleaved DNA in a T4 polynucleotide kinase reaction. @Xl 74 RFI DNA was cleaved with gyrase so that there was an average of less than
Cell 176
one cut per DNA molecule. The product was treated with proteinase K to trim attached protein and centrifuged through a sucrose density gradient to remove unreacted supercoiled DNA. The gyrase-produced linear DNA, containing an approximately equal amount of contaminating nicked circular DNA, was mixed with Eco RI-cut Col El DNA as an internal reference, and incubated with polynucleotide kinase and Y-~*P-ATP under conditions favoring labeling by exchange (van de Sande, Kleppe and Khorana, 1973). Figure 1 shows densitometer tracings of the products separated by gel electrophoresis. While Eco RI-generated linear Col El DNA became labeled, gyrase-cut 4x174 DNA did not, implying that the 5’ ends created by gyrase cleavage are blocked, even after proteinase K treatment. However, gyrase-cleaved DNA provided a template primer for both E. coli DNA polymerase I (see below) and T4 DNA polymerase (data not shown), showing that gyrase cleavage leaves a free 3’ hydroxyl terminus. Fine Mapping of Gyrase Cleavage Sites Labeling of gyrase-cleaved DNA with a-32P-dNTPs and DNA polymerase provided a sensitive means for detecting cleavages. +X174 RFI DNA was cleaved with gyrase so that approximately 60% of the DNA was converted to the linear form, and then labeled with E. coli DNA polymerase I and all four common a-32P-dNTPs. The products were digested with Hha I and resolved by gel electrophoresis (Figure 2, lane 1). Densitometer tracings showed that the bands varied in intensity over at least a 20 fold range. We chose to look at gyrase cleavage sites in Hha l-digested +X174 DNA for the following reasons. First, the relatively small sizes of the resulting gyrase cleavage products could be measured to within about 20 bp. Second, the two largest Hha I fragments of $X174 DNA are 1553 and 640 bp long, so that any gyrase cleavage product migrating between these two fragments must have arisen from the 1553 bp Hha I-A fragment. Thus at least one of each pair of fragments resulting from every gyrase cleavage event within the Hha I-A fragment could be resolved without the necessity of purifying the restriction fragments. Third, the $X1 74 Hha I-A fragment, represented diagramatically at the bottom of Figure 2, contains the single Ava II site of +X174 DNA, so that a Hha I/Ava II double digest of gyrase-cleaved DNA (Figure 2, lane 2) allowed the mapping of many gyrase cleavage fragments. Thus the a and a’ bands are partner fragments arising from gyrase cleavage at the a site; this is also the most prominent cleavage site in whole +X174 DNA (Peebles et al., 1978). Columns l-4 in Table 1 summarize the data defining the map positions of the gyrase a and b sites within the +X174 Hha I-A fragment. The same experimental procedure was applied to
Figure 1. Gyrasa-Cleaved Polynucleotide Kinase
@Xl 74 DNA
Is Not
a Substrate
for T4
A mixture containing approximately equal amounts of nicked circular $X1 74 DNA, linear @Xl 74 DNA formed by gyrase cleavage of @Xl 74 RFI DNA and linear duplex Col El DNA generated by Eco RI nuclease digestion was incubated with Y-~*P-ATP and T4 polynucleotide kinase to label the 5’ DNA termini by exchange. The products were purified by gel filtration and electrophoresed through a 1% agarose gel. Microdensitometer tracings were taken of an autoradiograph (broken line) and a photograph of the ethidium bromide-stained gel (solid line). (A) nicked circular +X174 DNA; (B) Eco RI-generated linear Col El DNA: 62) gyrase-produced linear +X1 74 DNA. Electrophoresis was from left to right.
gyrase-cleaved SV40 DNA, except that the restriction enzyme Hae Ill was used. The Hae Ill digest of gyrasecut SV40 DNA shown in Figure 2, lane 3, revealed four strong bands, labeled a, a’, b and b’, arising from cleavage at the a and b sites. The 1661 bp SV40 Hae Ill-A fragment, represented diagramatically at the bottom of Figure 2, contains an Ava II site and the single Taq I site of SV40 DNA. There are no Ava II sites in the 752 bp Hae Ill-B or the 540 bp Hae Ill-C fragments of SV40 DNA. Double digestion of gyrase-cut SV40 DNA with Hae Ill and Ava II (Figure 2, lane 4) or Hae Ill and Taq I (Figure 2, lane 5) therefore allowed the mapping of the SV40 a and b gyrase sites within the Hae Ill-A fragment. The mapping data are summarized in Table 1, columns l-4. DNA Polymerase I Repair of Gyrase Cleavage Sites The labeling of gyrase cleavage fragments by E. coli DNA polymerase I (Figure 2) implied that gyrase cleavage creates a cohesive end with a recessed 3’ terminal hydroxyl. To analyze the base sequences of these cohesive ends, @Xl 74 DNA was cleaved with gyrase and then incubated with DNA polymerase I. The polymerase reactions contained either a single (Y--~‘PdNTP to label the first nucleotide, or a combination of one o(-32P-dNTP and three unlabeled dNTPs to determine the composition of the repaired cohesive ends. The reaction products were digested with Hha I and separated by gel electrophoresis (Figure 3).
Site-Specific
Cleavage
of DNA by DNA Gyrase
177
1 2
P C+
1000
nucleotide A*
T*
composition G’
C+
A*
t
T*
G*
-b
-a - a’ aa’-
- b!
b’-
5ooL
J a eix174
I
I
,
4000
,
,
,
'
, 31500
Figure 2. Gyrase A Fragments
Cleavage
I
b
AV~ II
sv40
I
,'
b
'8
'4
4’500
Figure 3. Labeling of Gyrase E. coli DNA Polymerase I 500
AYB II
8
'
5’000
a I
',
1
