Identification of Amino Acids of Endonuclease VII Essential for Binding and Cleavage of Cruciform DNA

Eur. J. Biochem. 245, 573-580 (1997) 0 FEBS 1997

Identification of amino acids of endonuclease VII essential for binding and cleavage of cruciform DNA Stefan GOLZ, Anja CHRISTOPH, Karin BIRKENKAMP-DEMTRODER and Borries KEMPER Institut fur Genetik der Universitat zu Koln, Koln, Germany

(Received 29 November 1996/19 February 1997) - EJB 96 1769/2

Endonuclease VII is a Holliday-structure-resolving enzyme of bacteriophage T4. The active protein is a homodimer with 157 amino acidshonomer. An amber mutation (amE727 in codon 151) inactivates the nuclease completely, indicating the importance of the seven C-terminal amino acids for nucleolytic activity. The influence of these amino acids on cruciform-DNA binding and cleavage was investigated through functional analysis of C-terminal-truncated proteins derived from deletion constructs. It was found that the three C-terminal amino acids are not necessary for binding and cleavage. A transition from active to inactive protein occurs gradually with truncations of the next four amino acids. Reduction of DNA-binding ability, as measured by electrophoretic mobility shift assays, was determined to be the primary defect in the cleavage-deficient proteins. This was further concluded by the finding that EVII- (1 - 150)-peptide""'h'r, a protein with fairly low affinity to cruciform DNA, contributes cleavage activity to reactions of wild-type EVII with cruciform DNA. [Asp62]EVII-( 1 - 156)-peptide lacking one C-term- nal amino acid, contains a point mutation in codon 62 that eliminates the nucleolytic activity of the protein while retaining its DNA-binding proficiency. By mixing binding-deficient and cleavage-deficient mutants in the same assay, cleavage of cruciform DNA resumed. Evidence is presented that complementation occurs by heterodimer formation. Our results show that the zinc-binding motif of EVII is not suffi- cient for cruciform-DNA binding.

Keywords: endonuclease VII; DNA-binding protein; bacteriophage T4; DNA repair; recombination.

Endonuclease VII (EVII) is the product of gene 49 of phage T4. The endonucleolytic activity of EVII is involved in DNA packaging and mainly required late during the infection cycle [I].

Detailed investigations of the substrate specificity of EVII revealed that the enzyme can react with many substrates by introducing staggered nicks flanking the defect within 2-6 nucle- otides 11, 21. These substrates include Holliday structures from in vivo-recombined figure-eight plasmid molecules [3], in vitro- generated RecA-mediated Holliday structures [4], supercoiled and hybrid cruciforms [S], synthetic cruciforms [2, 6- 111, synthetic Y-structures [2, 12, 131, heteroduplex loops [14, 151, five- arm and six-arm branched DNAs [16], mismatches in double- stranded DNA [2, 14, 151, single-strand overhangs [2, 121, single-stranded loops at the ends of hairpins [2], nicks and gaps in dsDNA 12, 121, curved DNA [17], bulky adducts [18, 191 and apurinic sites in dsDNA (Greger. B. and Kemper, B., unpublished results). The wide variety of substrates suggested an in- volvement of EVII in DNA repair, which was demonstrated by experiments in vitro [14, 1.51 and in vivo [20, 211.

The broad specificity of EVII for structural deviations in dsDNA raised the question of how the enzyme recognizes its target sites. From several studies addressing protein-DNA interactions it become obvious that binding affinities are primarily brought about by electrostatic interactions between basic amino acid residues and the DNA-phosphate backbone. Specifity of

binding is accomplished by coordinated placement of basic amino acids, which in turn is determined by the three-dimen- sional structure of the enzyme. Particular helical segments ap- pear to be widespread among such structural elements [22].

Wild-type EVII binds strongly to cruciform DNA, protecting about five residues at the junction point in opposite strands from hydroxyl radical cleavage [7]. As shown here, the protein derived from a classical amber mutation in codon 151 of gene 49 (amE727 [23]) binds very weakly to cruciform DNA. The stop codon truncates the protein by seven amino acids, thus indicating an importance of the C terminus.

To determine the amino acids required to restore high affinity for cruciform DNA, C-terminal truncated mutants of EVII were constructed. Each protein was purified to homogeneity and analyzed for DNA-binding activity. These assays indicate that the loss of the three C-terminal amino acids Ser-Leu-Lys are not essential for binding of EVII to cruciform DNA. However, removal of the next four or more amino acids severely affected binding and cleavage of cruciform DNA. In addition, a point mutation in codon 62 was demonstrated to affect cleavage but not binding. It is shown here that the C-terminal-truncated, DNA-binding-deficient EVII can functionally complement the cleavage-deficient [Asp62]EVII-( 1 - 156)-peptide, and vice versa. These results suggest that important amino acid residues for DNA binding and cleavage reside in separable domains of EVII.

Correspondence to B. Kemper, Institut fur Genetik der Universitat

Fux: +49 221 470 5112. E-mud; [email protected] Ahbreviationt EVII, endonuclease VII; EMSA, electrophoretic-mo-

zu Koln, Zulpicherstrasse 47, D-50674 Koln, Germany MATERIALS AND METHODS

Plasmids and bacterial strains. Expression vector PET1 I-a and Escherichia coli strains BL21 (DE3) and BLZI(DE3)LysS bility-shift assays; VFSDNA, very-fast-sedimenting DNA.

574 Golz et al. (Eur J. Biochem. 245)

were obtained from Studier [24]. Plasmids pRB210 and pRB211, carrying wild-type T4 gene 49 and the mutated gene 49 amE727, respectively, were constructed as previously described [25]. E. coli strain DH5a was used for sequence analysis P61.

Chemicals and radiochemicals. Acrylamidehisacrylamide (29: 1) for denaturing PAGE and poly(ethylenegiyco1) 6000 were from Serva. Acrylamidehisacrylamide (37.5 : 1) for native PAGE or SDS/PAGE was from BioRad. Dextran T500 was purchased from Pharmacia. [y-”P]dATP (specific activity > 5000 Ci/mmol) was purchased from Amersham. All other chemicals were purchased from Merck.

Oligonucleotides and DNA. Synthetic oligonucleotides were purchased from Pharmacia or Eurogentec. The sequences for the construction of cruciform DNA CFOl were described previously [25]. The sense primer for cloning truncated sequences of gene 49 was TTACCCGGGCATATGTTATTGAC- TGGCAAATTATAC in all constructs containing an NdeI restriction site (underlined). The antisense primers shared a common 5‘ extension bearing an NdeI restriction site followed by two stop codons for excluding read-through transcripts (TTAAGGC- GCCATATGTTATCA). The 3‘ ends, extending 20-22 nucleo- tides into the ORF of gene 49, varied depending on the constructs listed in Table 1. Synthetic 5’-radiolabeled cruciform DNA (CFO1 -strand 4) and tritium-labeled very-fast-sedimenting DNA ([‘HIVFSDNA) were prepared following published pro- cedures [27, 281.

Enzymes and proteins. Klenow fragment of DNA polymerase I, BSA, T4 DNA ligase, and restriction enzymes were purchased from Boehringer. T4 polynucleotide kinase was purchased from Serva. Taq DNA polymerase ‘Goldstar’ was obtained from Eurogentec. T7 polymerase sequencing kit was obtained from Pharmacia. All enzymatic reactions were performed following the manufacturer’s instructions. Wild-type EVII was purified from an overexpressing clone according to the published procedure [25].

Growth of bacterial strains. 1 1 Luria-Bertani medium containing 100 pg/ml ampicillin was inoculated with 10 ml of an overnight culture of transformed BL21(DE3) or BL21 (DE3)- LysS and grown at 30°C. At A,,,, = 0.8 (5X107 cells/ml) expres- sions of EVII proteins were induced by the addition of isopro- pylthio-pa-galactoside to 1 mM. After 2 h, cells were harvested and immediately frozen at - 80°C.

The amount of protein made after induction varied with each construct and increased with the number of residues deleted. This reflects decreasing toxicity of the products accompanied by removal of increasing numbers of C-terminal amino acids. pAC156* provides an exception to the rule.

PCR reactions and cloning. PCR reactions were performed in 100 pl containing the manufacturer’s reaction buffer, 100 pmol each primer, 100 ng genomic template DNA, 2 mM MgCl,, 160 FM dNTPs and 1 U Taq polymerase. Initial denatu- ration was at 96°C for 3 min, followed by 26 cycles of 94°C for 90 s, 50°C for 1 min and 72°C for 1 min, and a final elonga- tion step for 5 min at 72°C. The PCR products were purified by spin-column centrifugation (Qiagen), digested with NdeI and extracted with phenolkhloroform before ligation into the NdeI- digested dephosphorylated vector pETl1-a. Ligation products were transformed into E. coli DHSa, and recombinant gp49 DNA was verified by dideoxynucleotide sequencing [29] by means of the T7 polymerase sequencing kit. Plasmid DNA was transformed into the E. coli expression strain BL21(DE3) by heat shock [30]. Normal transformation frequencies were obtained with pAC147, pAC1.50, pAC153 and pAC156*. The transformation frequency was very low with pAC154 and no transformants could be obtained with pAC155 in this host.

pAC155 was, however, successfully transformed into E. coli BL21 (DE3)LysS, providing a more stringent control of protein expression [24]. These results indicate variations in toxicity for each protein. Plasmid pAC156* contained an additional point mutation (codon 62, Asn changed to Asp), which inactivates the nucleolytic activity. It was therefore excluded from some investigations described below.

Protein purification. All manipulations were carried out at 4°C or on ice if not stated otherwise. For chromatography of truncated EVII proteins, 5-ml heparin-agarose HiTrap columns and 1-ml FPLC Monos (HR5/5) columns (Pharmacia) were used. For heparin-agarose chromatography, flow rates of 1 ml/ min were used, and 1-ml fractions were collected. For Monos chromatography, flow rates of 0.5 ml/min were used, and 0.2- ml fractions were collected. The eluted fractions were analyzed on 15% SDS/polyacrylamide gels by silver staining [31]. All purified proteins were stored in 10 mM potassium phosphate, pH 6.0, 1 mM 2-mercaptoethanol, 1 mM EDTA, 10% (by vol.) glycerol (buffer A) at -20°C.

Preparation of crude extracts. Crude extracts were prepared from 10 g frozen cells. After the addition of 4 vol. (vol./mass) 10 mM Tris/HCl, pH 8.0, 10 mM MgCI,, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 10 mM 2-mercaptoethanol, 10 % (by vol.) glycerol, cells were thawed slowly on ice. The suspen- sion was sonicated for 30 min with a Branson sonifer equipped with a 0.5 cm tip and setting of 5. Insoluble cell debris was removed by centrifugation at lOOOOOXg for 45 min. The cleared supernatant was the crude extract. The crude extracts were subjected to poly(ethyleneglycol)/dextran two-phase separation as described earlier [25]. The separated poly(ethyleneglyco1) extracts were dialyzed overnight against buffer A containing 450 mM KCI.

Column chromatography. The dialyzed poly(ethyleneglyco1) extracts of EVII-(1 - 147)-peptide, EVII-(1 - 150)-peptide and EVII-( 1 - 1 50)-peptideamkr were loaded on heparin-agarose columns equilibrated with buffer A containing 450 mM KCl. The columns were washed with the same buffer and the proteins eluted with 10 bed vol. of a linear gradient from 450 mM to 1000 mM KCl in the same buffer. Fractions containing the most purified proteins were pooled and dialyzed against buffer A containing 50mM KC1. The dialyzed samples were loaded on a Monos column equilibrated with buffer A containing 50 mM KC1, and the column was washed. Proteins were eluted with 5 bed vol. of a linear gradient from 50 mM to 1000 mM KCI. For purification of proteins EVII-(1 - 153)-peptide, EVIL(1- 154)- peptide, E V E ( 1 - 155)-peptide and [Asp62]EVII-(l- 156)-peptide poly(ethyleneglyco1) extracts were loaded on equilibrated heparin-agarose columns. The columns were washed, and the proteins eluted with 10 bed vol. of a linear gradient from 450 mM to 2000 mM KC1.

Comments on the purification of EVII proteins. Chroma- tography steps could be kept to a minimum when only the purest protein fractions were collected in each step. Contamination with two proteins of 11 kDa and 9 kDa must be avoided when pooling fractions. The 9-kDa protein was identified by sequence analysis as the DNA-binding protein HU from E. coli. The 11- kDa protein has not been identified.

The proteins eluted from the heparin-agarose column at different salt concentrations, which were related inversely to the number of deleted amino acids. Generally, larger deletions resulted in lower affinities to heparin. EVII-( 1 - 147)-peptide elutes at 580 mM KCl, EVII-( 1 - 150)-peptide and EVII-(1 - 150)-peptidea”’he‘ at 600 mM KCI, EVII-(1- 153)-peptide at 740 mM KC1, EVII-(1 - 154)-peptide at 1 M KCl, and EVII-( 1 - 1 %)-peptide, [Asp62]EVII-( 1 - 156)-peptide and wild-type EVII at 1.1 M KC1.

Golz et al. (Eur: J . Biachem. 245) 575

Table 1. C-terminal truncated proteins of EVII. C-terminal amino acid sequences of codons 145- 157 are listed. The sequence of the EVII-(I - 150)-peptide'""h" terminates at codon 150. The nucleotide sequences of the truncated proteins are followed by two stop codons (opal and ochre; *).

Protein Plasmid Codon

145 146 147 148 149 150 151 152 153 154 155 156 157

Wild-type EVII pRB210 I A S F K K Q L R K S L K EVII-( 1 - 1 50)-peptide""'h" pRB211 I A S F K K amber L R K S L K [Asp62]EVII-(l-1.56)-peptide pAC156* I A S F K K Q L R K S L " EVII-(1 - 155)-peptide pAC155 I A S F K K Q L R K S " Y

EVII-( 1 - 154)-peptide pAC154 I A S F K K Q L R K * L R * EVII-(l- 153)-peptide pAC153 I A S F K K Q ~

* - x -

EVII-( 1 - 150)-peptide PAC150 I A S F K K * * - - - - -

EVII-(I - 147)-peptide pAC147 I A S * - ~ ~ - - - - -

Electrophoretic-mobility-shift assay. Reactions were carried out in 10 p1 20 mM Tridacetate, pH 7.4, 5 mM EDTA, 1 mM dithiothreitol, 1 pg/ml BSA containing 1 fmol radio- actively labeled cruciform DNA CFOI and varying amounts of protein. Protein was added after the mixtures were incubated for 5 min on ice to remove traces of Mg*+. Reactions were incubated for 15 min at 16°C and terminated by addition of 5 p1 40 mM Tris/HCl, pH 7.5, 4 mM EDTA, 25% (by vol.) glycerol, 400 pg/ml BSA, 0.1 % (mass/vol.) bromophenol blue. Samples were loaded on 0.75-mm 12% native polyacrylamide gels in a BioRad Miniprotean I1 chamber and electrophoresed in 67 mM Tris/HCl, pH 8.1, 33 mM sodium acetate, 20 mM EDTA for 3 h at 4°C with 7 Vlcm, according to Parsons et al. [32]. Reaction products were visualized by autoradiography of the dried gels.

EVII digests. A total reaction volume of 10 pl 20 mM Trisl acetate, pH 7.4, 10 mM magnesium acetate, 1 mM dithiothreitol, 1 pglml BSA contained 2 fmol radioactive labeled cruciform DNA CFOI. Wild-type EVII or truncated EVII proteins were added, and reaction mixtures were incubated for 1.5 min at 37°C. After ethanol precipitation, samples were suspended in 90% (by vol.) formamide, 4.5 mM Tridborate, pH 8.0, 1 mM EDTA, 0.1 96 (mass/vol.) bromophenol blue and loaded on a 15% denaturing polyacrylamide gel containing 7 M urea. Reaction products were visualized by autoradiography and quantitated by phosphorimaging on a Fuji BAS1000. Digestion of ['HI- VFSDNA were carried out as described [25].

EVII affinity chromatography. For affinity chromatography, 1-ml HiTrap NHS-activated columns from Pharmacia were used. The coupling of proteins was carried out according to the manufacturer's instructions with 1 mg of highly purified wild- type EVlI. Columns with 1 mg bound BSA or saturated with ethanolamine were used for control runs. All chromatography steps were carried out at 4°C.

EVII proteins were dialyzed against 10 mM potassium phosphate, pH 6.5, 10 mM EDTA, 1 mM 2-mercaptoethanol, 10% (by vol.) glycerol, 20 mM KCl. 20 pg protein was loaded on a previously equilibrated column. Coupling was carried out for 20 min. The columns were washed with 5 bed vol. buffer. Bound proteins were eluted with a 10-ml linear gradient from 20 mM to 1000 mM KCl. Elution of proteins was monitored by measuring conductivity and A,,,,.

RESULTS

The conditional lethal amber mutation amE727 i n gene 49 of phage T4 causes DNA packaging to cease prematurely. Viable phages are not made in non-permissive hosts 1271. When the mutant gene was cloned, overexpressed and the amber protein

kDa - 66.2

- 40.0

- 31 .O

- 21.5

- 14.4

1 2 3 4 5 6 7 8 9 10

Fig. 1. Purity of truncated EVII proteins. Purified wild-type EVII and C-terminally truncated EVII proteins were analyzed on 15 % SDS-containing polyacrylamide gel and visualized by silver staining. 1 pg or 2 pg protein were used. Lanes 1 and 10, low-molecular-mass markers; lanes 2 and 9, wild-type EVII; lane 3, EVII-(1-147)-peptide; lane 4, EVII- (1 -150)-peptide; lane 5 , EVII-(1-150)-peptide""'h" ; lane 6, EVII-( 1 - 153)bpeptide; lane 7, EVII-(1-154)-peptide (2 pg); lane 8, EVW(1- 155)-peptide.

[EVII-( 1 - 150)-peptide""h"'] purified to homogeneity, neither binding nor cleavage of cruciform DNA was observed 1251. This differed from the wild-type protein, which binds and cleaves cruciform DNA efficiently [12]. The amber mutation amE727 was localized in codon 151 of gene 49 1331. To investigate which of the missing seven C-terminal amino acids are responsi- ble for the observed loss of activity, C-terminal truncations of EVII were created by 3' deletions of the gene 49 sequence in expression vector PET-1 la.

Cloning and purification of truncated EVII proteins. Dele- tions of one [Asp62]EVII-(I- 156)-peptide, two [EVII-(1 - 155)-peptide], three [EVII-( 1 - 154)-peptide], four [EVIL-( 1 - 153)-peptide], seven [EVII-(l - 150)-peptide] and ten amino acids [EVII-(1 - 147)-peptide] from the C-terminus of EVII were made as described in Materials and Methods. The constructs are summarized in Table 1. The earlier cloned wild-type EVII and EVII-(l- 150)-peptide"mb" were included in the studies.

EVII-(1 - 150)-peptide contains the same number of amino acids as EVII-(1 -150)-peptide""h'r. It was made in particular to measure whether the low level of residual nucleolytic activity observed previously with EVII-(1 - 150)-peptidea""'er was intrin- sic to the enzyme or due to the full-length protein originating from occasional read-through transcripts. The truncated EVII proteins were purified from overexpressing E. coli strains as described in Materials and Methods.

576

A

Golz et al. ( E m J. Biochem. 245)

- d - C

-b

- a

- 4-way junction

Fig. 2. Binding of EVII proteins to cruciform DNA CFO1. For gel-retardation analysis 1 fmol cruciform DNA CFOl containing 5'-end-labeled strand 4 was incubated for 15 min at 16°C with EVII proteins as indicated, then analyzed on 12% native polyacrylamide gels. (A) Comparison of wild-type EVII and EVII-(I - 150)-peptide""h"'. Lane 1, incubation without enzyme; lanes 2-4, assays containing 1.5, 12.5 and 50 ng wild-type EVII, respectively; lanes 5-8, assays containing 1.5, 12.5, 50 and 400 ng EVII-(1- 150)-peptideamb"', respectively. (B) Comparison of wild-type EVII and truncated proteins EVII-(1 - 155)-peptide (lanes 6 and 7), EVII-(1- 154)-peptide (lanes 4 and 5 ) and EVII-(1- 153)-peptide (lanes 2 and 3). Samples containing 25 ng or SO ng of protein (as indicated) were incubated with 1 fmol cruciform DNA CFOl and treated further as described.

v) m v) m , " v ! " v ! " v ! " v ! "! [ngl 0 0 O r 0 O r 0 0 r O O r 2 r o r

29 -

17 -

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Fig. 3. Cleavage of cruciform DNA CFOl by EVII proteins. 2 fmol cruciform DNA CFOl containing 5'-end-labeled strand 4 were incubated with varying amounts of wild-type EVII or truncated proteins for 15 min at 37°C. Lane 1, no enzyme; lanes 2-4, wild-type EVII; lanes 5-7, EVII- (l-lSS)-peptide; lanes 8-10, EVIT-(l-l54)-peptide; lanes 11-13, EVII-(l-l53)-peptide; lanes 14 and 15, EVII-(l-lSO)-peptide; lanes 16 and 17, EVII-(l-l47)-peptide. Samples were analyzed on 1.5% denaturing polyacrylamide gels in 45 mM Trishorate, pH 8.0, 1 mM EDTA. The gels were autoradiographed and the amount of labeled DNA in each band was quantitated by phosphorimaging. 0.25, 0.5 and 1 ng of protein correspond to SO, 100 and 200 U of active wild-type EVII, respectively.

All proteins were highly pure as judged from silver-stained gels (Fig. 1). They were free from any detectable exonuclease activity, and no loss of label was observed after prolonged incu- bations of 5'-end-labeled cruciform DNA CFOl under standard assay conditions.

C-terminal amino acids of EVII are essential for DNA binding. The DNA-binding capacities of wild-type EVII and the truncated proteins with cruciform DNA CFOl were investigated. Wild-type EVII, EVII-(1- 155)-peptide, EVII-(1 - 154)-peptide and EVII-(1- 153)-peptide bind to CFOl with decreasing affinity from 100% to 1 % in this order (Fig. 2b). A maximum of

four shift positions (labeled a-d in Figs 2, 4 and 6) could be reached, depending on the protein concentration used in the assay. However, the highest shift positions were only reached with the strong binding proteins, wild-type EVII and EVII-( 1 - 155)- peptide. The same amount of weaker binding proteins EVTI-(1 - 154)-peptide, EVII-( 1 - 153)-peptide, EVII-( 1 - 150)-peptideamh" and EVII-( 1 - 150)-peptide shifted the substrate DNA only to the lower positions (Fig. 2). EVII-(l- 147)-peptide had the lowest affinity for cruciform DNA, and binding could not be demonstrated. The specificity of the mutant proteins for cruciform DNA was verified by means of unbranched duplex DNA as a control. No binding was seen, even in the presence of an excess

Golz et al. (Eur: J. Biochern. 245) 577

f d - c - b

- a

- 4-way junction

1 2 3 4 5 6 7 8

Fig.4. Binding of cruciform DNA CFOl by wild-type EVII in the presence of EVII-(l-l50)-peptideamher. 1 fmol cruciform DNA CFOI containing 5'-end-labeled strand 4 was incubated with 5 ng wild-type EVII (lane 2) or EVII-(1-150)-peptide""'her (lanes 3-S), or a mixture of wild-type EVII and EVII-(1- 150)-peptide""b"r (lanes 6-8) for 15 min at 16°C. Lane 1, incubation of CFOl without protein. EMSA of samples were performed on 12% native polyacrylamide gels. Gels were autoradiographed after drying and quantitated by phosphorimaging.

of protein with this DNA (data not shown). Weak residual binding was obtained when excessive amounts of EVII-(1 - 150)- peptide""ber (Fig. 2a) or its analogue EVII-(1 - 150)-peptide (data not shown) were included in the reactions. From these results we conclude that the residual binding activity of EVII-(1 - 150)-peptideamb" is inherent to this protein and is not due to contaminating read-through products in the preparation.

The first shift position is unusual in that it sometimes contains low amounts of material without addition of protein, depending on the preparation of cruciform DNA (Figs 2 and 4). The addition of strong-binding proteins shifts this material completely from this position to higher positions (Fig. 2b).

Loss of DNA binding is accompanied by loss of DNA cleavage. The truncated proteins were subjected to a series of nuclease assays including natural substrate VFSDNA and synthetic substrate cruciform DNA CFOl . Specific activity dropped

from 100% for wild-type EVII to 50% for EVII-(1 - 155)-peptide, 20% for EVII-(1- 154)-peptide and 1 % for EVII-(1-153)- peptide, as measured by degradation of [IHIVFSDNA. The ability of the proteins to cleave CFOl decreases with the number of amino acids deleted (Fig. 3). Wild-type EVII is the most active protein, followed by EVII-(1- 155)-peptide, EVIL(1- 154)-peptide and EVII-(1 - 153)-peptide. EVII-(1 - 150)-peptide and EVII-(I - 147)-peptide were inactive under these assay conditions. Very low residual activity was detectable with EVII-(1 - 150)-peptidea"h" and EVII-( 1 - 150)-peptide if excessively high amounts of proteins were included in the assay. The activities were too low to determine specific activities. All active proteins induced a similar cleavage pattern in strand 4 of CFOl with cleavages located 3' to the junction and the major cleavage at nucleotide 17. The intensity of the product resulting from a cleavage at nucleotide 16 decreases for all active truncated proteins.

These results indicate that the three C-terminal amino acids are largely dispensable for nucleolytic cleavage, while removal of the next four amino acids causes loss of measurable activity.

C-terminal truncated proteins support wild-type EVII in binding cruciform DNA. One could assume that by removing the C-terminal domain of EVII, important residues for DNA cleavage were removed. To demonstrate the presence of cata- lytic activity in DNA-binding-deficient mutants, complementation assays were performed by assuming heterodimer formation between wild-type EVII and truncated EVII mutants. Subunit exchange could be shown by several means. When low concentrations of wild-type EVll were mixed in solution with increasing amounts of EVII-(1 - 150)-peptideamh" and analyzed in gel- retardation experiments with cruciform DNA CFOl as substrate, an increase in the amount of shifted substrate was found (Fig. 4). EVII-( 1 - 1 50)-peptide""b" alone does not bind cruciform DNA (Fig. 4). Compared with the control reaction with wild-type EVII alone, the addition of high amounts of EVII-( 1 - 150)-pep- tidedmhcr resulted in a higher shift position of cruciform DNA,

I I 0 0 0 0

r ( V w ( V m m o s z r ( V w ( V m r r 0 0 UJ EVlI-(1-150)-peptideamber(ng)

1 e I I I I I I I z s z wild-typeEVlI(U)

EVII -(1-1 50)-peptideamber(ng)

0 0 m o o ? r N V ) N V ) r

2.0

1.8

1.6

1.4 5 0

* c

1.2 2 1.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Fig. 5. Cleavage of cruciform DNA CFOl by wild-type EVII in the presence of EVII-(l-150)-peptideamher. 2 fmol of cruciform DNA CFOl containing j'-end-labeled strand 4 were incubated for 15 min at 37°C in the presence of 10 U wild-type EVII (lane 2) or increasing amounts of EVII-(1 -150)-peptideamh"' (lanes 3-9) or both (lanes 10- 17) as indicated. Lane 1, incubation of CFOl without protein. Samples were analyzed on 15% denaturing polyacrylamide gels in 45 mM Trishorate, pH 8.0, 1 mM EDTA. Gels were autoradiographed and quantitated by phosphorimaging. The graph shows the amount of products obtained in mixtures of 10 U of wild-type EVII with increasing amounts of EVII-(l- 150)-peptide""'h".

578 Golz et al. (Eur: J . Biochern. 245)

Fig. 6. [Asp62]EVII-(l-l56)-peptide in reactions with cruciform DNA CFOI. (A) Gel-retardation analyzes. 1 fmol CFOl containing 5'-end- labeled strand 4 was incubated with wild-type EVII (lanes 2-5) or [Asp62]EVII-(I-l56)-peptide (lanes 6-9) for 15 min at 16°C. Samples were analyzed on 12% native polyacrylamide gels. Gels were dried and autoradiographed. Lane 1, control. (B) Cleavage of cruciform DNA CFOI. 1 fmol cruciform DNA CFOl containing S'-end-labeled strand 4 was incubated for 15 min at 37°C with [Asp62]EVII-(I-l56)-peptide (1 ng, lane 2 or 10 ng, lane 3) or wild-type EVII (1 ng, lane 4 or 10 ng, lane 5 ) . Samples were analyzed as described in the legend to Fig. 3. Lane 1, control. (C) Complementation between [Asp62]EVII-(I - IS6)-peptide and EVII-(I - 150)-peptide"""". Samples containing 1 fmol DNA CFOI were incubated following the procedure described in the legend to Fig. 3. Lane 1, control; lane 2, 80 ng EVII-(l-lSO)-peptide""h"; lane 3, 15 ng [Asp62]EVII- (l-l56)-peptide; lane 4, 0.05 ng wild-type EVII; lanes 5 and 6, 15 ng [Asp62]EVII-(I-l56)-peptide and 8Ong (lane 5 ) or 4Ong (lane 6) EVII- (l-lSO)-peptide""'h".

designated as c in Fig. 4. The addition of BSA in control reactions did not show any effect (data not shown).

Protein-protein interactions between truncated EVII proteins and wild-type EVII were also shown by affinity chromatography with immobilized wild-type EVII. The chromatography was performed as described in Materials and Methods. E V E ( 1 - 150)- peptidyher protein bound to immobilized wild-type EVII with nearly the same affinity as wild-type EVII itself and was eluted with the same salt concentration of 200 mM KCl.

Truncated proteins support wild-type EVII in cleaving cruciform DNA. The ability of different binding-deficient EVII proteins to support cleavage of cruciform DNA in the presence of wild-type EVII was demonstrated i n in vitro complementation experiments. Cleavage efficiencies were measured in mixtures of increasing amounts of EVlI-(1 -150)-peptide or EVIlL(1- 147)-peptide with 10 U wild-type EVII in assays with cruciform DNA CFOl as substrate. An increase in the amount of cleavage products was obtained after the addition of EVII-(1 - 150)-pep- ti d p 7 hrr (Fig. 5) . The highest concentration of EVII-( 1 - 150)- peptide~nher (150 ng) alone did not cleave CFOl (Fig. 5) . The stimulating effect was determined to be twofold for the highest protein concentration used. The same results were obtained with EVII-(l - 150)-peptide instead of EVII-(1 - 150)-peptide""h" (data not shown). BSA used in control reactions did not show any effect. Addition of excessive amounts of protein EVII-(1 - 150)-peptide""'"' did not stimulate wild-type EVII more than twofold (data not shown). This can be explained if one assumes that active EndoVII is a dimer, as suggested by earlier experiments [34, 351. Assuming that the heterodimers formed between wild-type EVII and EVIT-( 1 - 150)-peptide""'"' show the same activity as wild-type EVII homodimers the amount of active enzyme can double at best.

Binding-proficient but cleavage-deficient [Asp62]EVII- (l-l56)-peptide complements EVII-(l-l50)-peptideamber. The ability of EVII-(1 - 150)-peptide""b" to support wild-type EVII in binding and cleaving cruciform DNA suggested that the C- terminal truncated proteins have retained their endonucleolytic activity. This should become more obvious in reactions with a cleavage-deficient but binding-proficient mutant. [Asp62]EVII- (1 -156)-peptide, which bears a point mutation i n codon 62 retained nearly wild-type cruciform-DNA-binding capacity but was completely deficient in cleavage (Fig. 6). After addition of EVII-( 1 - 1 50)-peptide"'"h"', wild-type activity was restored (Fig. 6c). The same results were obtained when the experiment was repeated when protein EVK(1- 147)-peptide replaced EVII-(1 - 150)-peptide"'"h"' (data not shown). These results show that the binding-deficient proteins have retained their nucleolytic activity.

DISCUSSION

EVII recognizes a large variety of structural deficiencies in dsDNA. We have shown previously that the enzyme targets some of these defects for repair by introducing 3' ligatable staggered nicks flanking the deficiencies [14, 151.

An amber mutation in codon 151 inactivates the cleavage activity of EVII. This emphasizes the importance of the C-terminus for cleavage activity, as previously suggested [25]. The function of the C-terminus was studied in greater detail by sys- tematically creating truncated proteins and analyzing their DNA- binding and DNA-cleavage properties in vitro. It was found that deletions of 2-3 amino acids had only minor effects on the cleavage activity. However, deletions of four or more amino acids heverely reduced the cleavage activities of the proteins. A deletion of seven amino acids [EVII-(1 - lSO)-peptide], mimick-

Golz et al. ( E m J. Biochmnz. 245) 579

ing the amber protein EVII-(1 - 150)-peptide""'her, nearly completely abolished the nucleolytic activity of the enzyme. Activity was measurable only with high concentrations of EVII-( 1 - 150)- peptide or EVII-(1 - 150)-peptide""'her. A deletion of 10 residues [EVII-(1 - 147)-peptide] reduced this residual activity below detectable level.

Evidence is presented here that the loss of cleavage activity of the truncated proteins was not a primary defect but rather a consequence of DNA-binding deficiencies. The strength of binding to cruciform DNA diminished gradually among the proteins, wild-type EVII, EVII-(1 - 155)-peptide, EVII-(1- 154)-peptide, EVII-(I - 153)-peptide, EVII-(1- 150)-peptide and EVII-(l- 147)-peptide together with loss of their cleavage activities in this order. In vitro complementation analyses, however, revealed that cleavage of cruciform DNA was demonstrable in these 'inactive' proteins. For example, EVII-(l- 150)-peptide""her and EVII-(I - 147)-peptide, deficient in DNA binding, contribute cleavage activity in reactions of cleavage-deficient [Asp62]EVII-( 1 - 156)- peptide with cruciform DNA. The cleavage pattern obtained in complementation reactions change gradually with the amount of complementing protein added. For example, a minor cleavage site used by wild-type EVII becomes a major cleavage site used by the protein mixture (Fig. 5) .

The observed loss of DNA binding may indicate a direct role of the C-terminal amino acid residues in DNA contact. Among the last nine amino acids there are almost exclusively positively charged residues. The binding deficiency may, however, be due to an effect on the global protein structure, or more specifically on the zinc-binding domain residing in the N-terminal part of the protein. The latter is very unlikely, since it was shown that much longer deletions retain their ability to bind zinc [36]. The atomic structure of the Ruv C resolvase from E. coli has been resolved [37]. Ruv C exhibits similar Holliday-junction-cleavage activity as EVII, although the proteins do not share any sequence similarity. The DNA-binding interface of Ruv C is built by several structural elements such that a cleft fits a DNA duplex. The C-terminus of the enzyme was disordered in the crystal, forming a flexible and basic tail with unknown function.

Complementation between two defective EVII proteins is achieved by subunit exchange and heterodimer formation. This was evident from complementation between EVII-( 1 - 150)-peptide<", hcc and wild-type EVIT. The addition of mutant protein to wild-type protein caused a twofold increase in total nucleolytic activity, which is expected if a given amount of active homodi- meric wild-type EVII is fully converted to twice the amount of active heterodimers. Since this was found in the experiments described here (Fig. 5 ) , we conclude that exchange of subunits between EVII proteins is possible in solution and that C-terminally truncated EVII proteins with deletions including amino acid 147 fully retain their endonucleolytic potency. Heterodimer formation was described recently to occur by rapid subunit exchange in solution [36]. Physical binding of mutant proteins to immobilized wild-type EVIT was demonstrated here.

In contrast to the results reported by Piihler et al. [36], analyses by EMSA indicated a four-step mechanism of EVII binding to cruciform DNA. Four shift positions were distinguishable on a gel, reached in succession with increasing amounts of protein (Figs 2 , 4 and 6). The molecular mechanism and the stoichiometry of the recognition reaction between EVII and cruciform DNA are not clear. We assume that each shift position represents a different complex between EVII and cruciform DNA with all binding sites being saturated in the highest shift position. We may speculate that the stepwise loading reflects one physical unit (monomer, dimer or multimer) of the protein binding first and attracting three further units to the same DNA molecule thereafter. If the size of the physical unit of the protein is a dimer

as suggested [2, 34, 351, the four shift positions reflect four bound dimers. If the physical unit were a monomer, the highest shift position would reflect only two bound dimers. Both scenar- ios contradict the experiments of Pohler et al., who observed only one shift position and suggested that binding occurs between one dimer of EVII and one cruciform DNA molecule (361. Further investigations have to be carried out to explain the differences between their results and ours. DNA binding assays were carried out with different cruciform substrates and different protein preparations. It is known that the secondary structure of cruciform DNA depends largely on the nucleotide sequence at the junction [38], hence influencing the stoichiometry of binding of EVII.

During cloning of the deletion sequences, one sequence, with a point mutation in codon 62 changing Asn to Asp, was de- tected, which eliminates any measurable nucleolytic activity. Point mutations with a comparable phenotype have recently been mapped in our laboratory throughout the entire EVII sequence (unpublished results), and no preference for a certain location could be recognized.

We are indebted to Dr Piera Cicchetti for critical reading of the manuscript. This work was supported by the Deursche Forschungsge- meimchuft through SFB274, and a grant given to B. Kemper (Ke 188I 8-1 and 2).

REFERENCES 1. Frankel, F. R., Batcheler, M. L. & Clark, C. F. (1971) The role

of the gene 49 in DNA replication and head morphogenesis in bacteriophage T4, J. Mol. Biol. 62, 439-463.

2. Kemper, B., Pottmeyer, S., Solaro, P. & Kosak, H. (1990) Resolution of DNA-secondary structures by endonuclease VII (EVII) from phage T4, Humun genome initiutive & DNA recombinution (Sarma, R. H. & Sarma, M. H., eds) pp. 215-229, Adenine Press, Schenectady NY.

3. Mizuuchi, K., Kemper, B., Hays, J. & Weisberg, R. A. (1982) T4 endonuclease VII cleaves Holliday structures, Cell 29, 357 -365.

4. Mueller, B., Jones, C., Kemper, B. & West, S. C. (1990) Enzymatic formation and resolution of Holliday junctions in vitro, Cell 60,

5 . Kemper, B., Jensch, E, von Depka-Prondzynski, M., Fritz, H. J., Borgmeyer, U. M. & Mizuuchi, K. (1984) Resolution of Holliday structures by endonuclease VII as observed in interactions with cruciform DNA, Cold Spring Hurbor Symp. Quunt. Biol. 49,

6. Mueller, J. E., Newton, C. J., Jensch, F., Kemper, B.. Cunningham, R., Kallenbach, N. R. & Seeman, N. C. (1990) Resolution of Holliday junction analogs by T4 Endonuclease VII can be di- rected by substrate structure, J. Bid. Chem. 265. 13 9 18- 13 924.

7. Parsons, C. A,, Kemper, B. & West, S. C. (1990) Interaction of a four-way junction in DNA with T4 endonuclease VII. J. Biol. Chem. 265, 9285-9289.

8. Picksley, S. M., Parsons, C. A,, Kemper, B. & West, S. C. (1990) Cleavage specificity of bacteriophage T4 endonuclease V1I and bacteriophage T7 endonuclease I on synthetic branch migratable Holliday junctions, J . Mu/. Bid . 212, 723-735.

9. Mueller, J. E., Kemper, B., Cunningham, R. P., Kallenbach, N. R. & Seeman, N. C. (1988) T4 endonuclease VII cleaves the crossover strands of Holliday junction analogs, Proc. Nufl Acud. Sr,i. USA 85, 9441 - 9445.

10. Jensch, F., Kosak, H., Seeman, N. C. & Kernper, B. (1989) Cruci- form cutting endonucleases from Saccharomyces cerevisiae and phage T4 show conserved reactions with branched DNAs, EMBO

11. Seeman, N. C., Mueller, J. E., Chen, J., Churchill, M. E. A., Kim- ball, A,, Tullius, T. D., Kemper, B., Cunningham, R. P. & Kallen- bach, N. R. (1990) Immobile junctions suggest new features of the structural chemistry of recombination, Humun genome inifiu- five & DNA recombination (Sarma, R. H. & Sarma, M. H., eds) pp. 137-156, Adenine Press, Scheuectady NY.

329-336.

815-825.

J . 8,4325-4334.

580 Golz et al. (Eui: J. Biochern. 245)

12. Pottmeyer, S. & Kemper, B. (1992) T4 endonuclease VII resolves cruciform DNA with nick and counter-nick and its activity is di- rected by local nucleotide sequence, J. Mol. Bid. 223, 607-615.

13. Jensch, F. & Kemper, B. (1986) Endonuclease VII resolves Y-junctions in branched DNA in v i m , EMBO J. 5, 181-189.

14. Solaro, P. C., Birkenkamp, K., Pfeiffer, P. & Kemper, B. (1993) Endonuclease VII of phage T4 triggers mismatch correction in vitro, J. Mol. Biol. 230, 868-877.

15. Birkenkamp, K. & Kemper, B. (1995) In vitro processing of heteroduplex loops and mismatches by endonuclease VII, DNA Res. 2,

16. Fu, T. J., Kemper, B. & Seeman, N. C. (1994) Cleavage of double- crossover molecules by T4 endonuclease VIl, Biochemistry 33, 3896- 3905.

17. Bhattacharyya, A., Murchie, A. I. H., Kitzing, E. v., Diekmann, S., Kemper, B. & Lilley, D. M. J. (1991) Model for the interaction of DNA junctions and resolving enzymes, J. Mol. Bid. 221,

18. Bertrand-Burggraf, E., Kemper, B. & Fuchs, R. P. P. (1994) Endonu- clease VII of phage T4 nicks N-2 acetylaminofluorene induced DNA structures in vitro, Mutut. Res. 314, 287-295.

19. Murchie, A. I. H. & Lilley, D. M. J. (1993) T4 endonuclease VII cleaves DNA containing a cisplatin adduct, J. Mol. Biol. 233,

20. Shcherbakov, V. P. & Plugina, L. A. (1991) Marker-dependent recombination in T4 bacteriophage. 111. Structural prerequisites for marker discrimination, Genetics 128, 673 -685.

21. Grebenshchikova, S. M., Plugina, L. A. & Shcherbakov, V. P. (1994) The role of T4-bacteriophage endonuclease-VII in correction of mismatched regions, Genetika 30, 622-626.

22. Churchill, M. E. A. & Travers, A. A. (1991) Protein motifs that recognize structural features of DNA, Trends Biochern. Sci. 16,

23. Epstein, R. H., Bolle, A,, Steinberg, C. M., Kellenberger, E., Boy La Tour, E., Chevalley, R., Edgar, R. S., Susman, M., Denhardt, G. H. & Lielausis, A. (1963) Physiological studies of conditional lethal mutants of bacteriophage T4D, Cold Spring Harbor Symp. Quant. Biol. 28, 375-394.

24. Studier, W. F., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W. (1 991) Use of T7 RNA polymerase to direct expression of cloned genes, Methods Enzymol. 185, 60-89.

25. Golz, S. , Birkenbihl, R. & Kemper, B. (1995) Improved large scale preparation of phage T4 endonuclease VII overexpressed in E. coli, DNA Rex 2, 277-284.

9-14.

1191 - 1207.

77-85.

92-97.

26. Hanahan, D. (1983) Studies on transformation of Escherichia coli with plasmids, J. Mol. Bid. 166, 557-580.

27. Kemper, B. & Brown, D. T. (1976) Function of gene 49 of bacteriophage T4. 11. Analysis of intracellular development and the structure of very fast-sedimenting DNA, J. Virol. 18, 1000-1015.

28. Kemper, B., Garabett, M. & Courage, U. (1981) Studies on T4-head maturation. 2. Substrate specificity of gene-49-controlled endonuclease, Eur: J. Biochem. 115, 133-141.

29. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) DNA sequencing with chain-terminating inhibitors, Proc. Nut1 Acad. Sci. USA 74,

30. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular cloning: a laboratory manual, 2nd edn, Cold Spring Harbor Labora- tory Press, Cold Spring Harbor NY.

31. Heukeshoven, J. & Dernick, R. (1988) Improved silver staining procedure for fast staining in PhastSystem development unit. I. Stain- ing of sodium dodecyl sulfate gels, Electrophoresis 9, 28-32.

32. Parsons, C. A. &West, S. C. (1988) Resolution of model Holliday- junctions by yeast endonuclease is dependent upon homologous DNA sequences, Cell 52, 621 -629.

33. Barth, K. A,, Powell, D., Trupin, M. & Mosig, G. (1988) Regulation of two nested proteins from gene 49 (recombination endonuclease VII) and of a lambda rexA-like protein of bacteriophage T4, Ge- netics 120, 329-343.

34. Kosak, H. G. & Kemper, B. (1990) Large-scale preparation of T4 endonuclease VII from over-expressing bacteria, Eur: J. Biochem.

35. Pohler, J. R. G., Giraud-Panis, M. J. E. & Lilley, D. M. J. (1996) T4 endonuclease VII selects and alters the structure of the four- way DNA junction; Binding of a resolution-defective mutant enzyme, J. Mol. Bid . 260, 678-696.

36. Giraud-Panis, M. J. E., Duckett, D. R. & Lilley, D. M. J. (1995) The modular character of a DNA junction-resolving enzyme: A zinc-binding motif in bacteriophage T4 endonuclease VII, J. Mol. Biol. 252, 596-610.

37. Ariyoshi, M., Vassylyev, D. G., Iwasaki, H., Nakamura, H., Shina- gawa, H. & Morikawa, K. (1994) Atomic structure of the RuvC resolvase : a Holliday junction-specific endonuclease from E. coli, Cell 78, 1063-1072.

38. Duckett, D., Murchie, A. I. H., Diekmann, S., Kitzing, E. v., Kemper, B. & Lilley, D. M. J . (1988) The structure of the Holli- day junction and its resolution, Cell 55, 79-89.

5463 -5467.

194,779-784.

Documents

Identification of Amino Acids of Endonuclease VII Essential for Binding and Cleavage of Cruciform DNA