A Conserved Tyrosine Residue Aids Ternary Complex Formation, but not Catalysis, in Phage T5 Flap...

A Conserved Tyrosine Residue Aids Ternary ComplexFormation, but not Catalysis, in Phage T5Flap Endonuclease

Dipak Patel1, Mark R. Tock2, Elaine Frary2, Min Feng1

Timothy J. Pickering2, Jane A. Grasby2 and Jon R. Sayers1*

1University of SheffieldMedical School, Division ofGenomic Medicine, KrebsInstitute, SheffieldS10 2RX, UK

2Centre for Chemical BiologyDepartment of ChemistryKrebs Institute, University ofSheffield, Sheffield, S3 7HF, UK

The flap endonucleases, or 50 nucleases, are involved in DNA replicationand repair. They possess both 50-30 exonucleolytic activity and the abilityto cleave bifurcated, or branched DNA, in an endonucleolytic, structure-specific manner. These enzymes share a great degree of structural andsequence similarity. Conserved acidic amino acids, whose primary roleappears to be chelation of essential divalent cation cofactors, lie at thebase of the active site. A loop, or helical archway, is located above theactive site. A conserved tyrosine residue lies at the base of the archwayin phage T5 flap endonuclease. This residue is conserved in the structuresof all flap endonucleases analysed to date. We mutated the tyrosine 82codon in the cloned T5 50 nuclease to one encoding phenylalanine.Detailed analysis of the purified Y82F protein revealed only a modest(3.5-fold) decrease in binding affinity for DNA compared with wild-typein the absence of cofactor. The modified nuclease retains both structure-specific endonuclease and exonuclease activities. Kinetic analysis wasperformed using a newly developed single-cleavage assay based onhydrolysis of a fluorescently labelled oligonucleotide substrate. Substrateand products were resolved by denaturing HPLC. Steady-state kineticanalysis revealed that loss of the tyrosine hydroxyl function did notsignificantly impair kcat. Pre-steady state analysis under single-turnoverconditions also demonstrated little change in the rate of reaction com-pared to the wild-type protein. The pH dependence of the kinetic para-meters for the Y82F enzyme-catalysed reaction was bell-shaped as for thewild-type protein. Thus, Y82 does not play a role in catalysis. However,steady-state analysis did detect a large (t300-fold) defect in KM. Theseresults imply that this conserved tyrosine plays a key role in ternarycomplex formation (protein–DNA–metal ion), a prerequisite for catalysis.

q 2002 Elsevier Science Ltd. All rights reserved

Keywords: flap endonuclease; DNA cleavage; DNA–protein interaction;enzyme kinetics; fluorescent assay*Corresponding author

Introduction

The processes of DNA replication and repair cangive rise to the formation of branched structures.The flap endonucleases (or 50 nucleases) are ableto process 50 overhanging ends by both exonucleo-

lytic and endonucleolytic cleavage. The latteractivity is specific for the double-stranded/single-stranded junction.1,2 These enzymes are thought toplay roles in processes such as DNA replicationand repair, stabilisation of repeat sequences3 andin the removal of ribonucleotides from the 50 endsof Okazaki fragments.4

Recent crystallographic studies on the 50

nucleases have revealed a common architecture.5 – 9

A central beta-sheet is surrounded by helicalregions and forms the framework containingconserved aspartyl and glutamyl residues(Figure 1). These residues form the binding site

0022-2836/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved

E-mail address of the corresponding author:[email protected]

Abbreviations used: PDB, Protein Data Bank; WT,wild-type; BSA, bovine serum albumin; Caps, 3-[cyclohexylamino]-1-propanesulphonic acid; Mes, 2-[N-morpholino]ethanesulphonic acid.

doi:10.1016/S0022-2836(02)00547-8 available online at http://www.idealibrary.com onBw

J. Mol. Biol. (2002) 320, 1025–1035

for two divalent metal ions. Mutagenesis studieshave concentrated mainly on the roles of theconserved acidic amino acid residues, which linethe active site.10 – 15

The active site is topped by a helical arch or loopfeature, which it has been suggested, is involved inthe threading of single-stranded DNA through theenzyme.7 A number of biochemical and structuralstudies have provided supporting evidence for thethreading model of binding.1,7,8,16,17 At the base ofthe helical arch feature, present in T5 exonuclease,lie two conserved residues, tyrosine 82 and lysine83. Previous studies have shown that this latterresidue is essential for efficient nucleolytic activityand substrate binding,18 – 20 while work on theconserved tyrosine residue present in homologous50 nucleases has proved less conclusive. Grindleyand co-workers showed that the Y77C mutation inEscherichia coli DNA polymerase I abolishedactivity as might be expected when a highly con-served residue is tampered with.21 More recentquantitative studies from the same laboratory onanother DNA polymerase I mutant, Y77A, showeda 1000-fold decrease in kcat.

22 In contrast, studiesby Nossal and co-workers showed that mutationof the T4 RNaseH tyrosine residue had little affecton binding or catalysis.20 This is surprising as one

might expect that mutation of a highly conservedresidue present within the active site of an enzymewould have effects on some aspect of function.

We studied the role of the conserved tyrosineresidue (Y82) in the T5 flap endonuclease alsoknown as the D15 50-30 exonuclease.23 Site-directedmutagenesis was employed to alter the codon fortyrosine 82 to either phenylalanine or leucine. Theformer, relatively subtle alteration, results in theloss of a phenolic hydroxyl group, but retainsthe aromatic character and steric bulk of the resi-due replaced. Substitution with leucine abolishesboth the aromatic character and the hydroxylfunction. Quantitative, as well as qualitative, tech-niques were used in order to clarify the role ofthis conserved residue in flap endonucleasefunction.

Results

The mutated genes were readily expressed inE. coli allowing milligram amounts of the Y82Fand Y82L proteins to be purified by ion exchangechromatography even though they were initiallyfound to be in the insoluble fraction after celllysis. However, after dissolution with a chaotrope,

Figure 1. Structure of theT5 exonuclease and presence of a conserved tyrosine residue (PDB code 1EXN). (a) Struc-ture of T5 flap endonuclease showing conserved basic (blue, lysine 83), acidic (red) residues and tyrosine 82 (yellow).Metal ions are shown as pink spheres. (b) Methanococcus jannaschii flap endonuclease (PDB code 1A76) with metalsshown as pink spheres and basic residue arginine 91 in blue. (c) and (d) Enlarged view showing a comparison of theactive sites of the T5 and Methanococcus enzymes, respectively. In each panel, the right-most metal ion occupies metalsite 1 while the metal to the left occupies site 2.

1026 Role of Tyrosine 82 in Phage T5 Flap Endonuclease

which was subsequently removed by dialysis, theproteins remained soluble.

Specific activity measurements

The ability of the enzymes to render highmolecular weight DNA into acid-soluble productswas investigated using a spectrophotometric assayto determine the level of exonucleolytic activity.Herring sperm DNA was converted to low molecu-lar weight products by the wild-type, Y82F andY82L mutant nucleases, albeit with a four- totenfold reduction in specific activity (Table 1). Theactivity of all three enzymes was relatively insensi-tive to the concentrations of Mg2þ cofactor usedover the range assayed (1–50 mM).

Structure-specific DNA cleavage

Enzymes were then assayed using a pseudo-Ysubstrate (Figure 2) in order to reveal any defectsin structure-specific cleavage activity (Figure 3).The substrate shown in Figure 2(a) is cleaved bythe wild-type enzyme at several phosphodiesterbonds. Cleavage close to the 50 end is deemed tobe exonucleolytic, while hydrolysis at the single-strand/double-strand junction is referred to asstructure-specific endonucleolytic cleavage. As canbe seen from Figure 3, both mutant enzymes wereable to degrade the substrate from the 50 end andalso perform endonucleolytic cleavage close to theduplex region. However, the level of endonucleoly-tic activity compared with exonuclease activityappears to be altered in that relatively little ofthe substrate was converted to labelled trimers orpentamers by the Y82F and Y82L nucleases.

Quantitative binding assays

Binding of mutant and wild-type enzymes wasinvestigated using electrophoretic mobility shiftassays (EMSA). These experiments were carriedout in the absence of cofactor using the pseudo-Ysubstrate (Figure 2(a)). Examination of the dataobtained from quantitative imaging of the band-shift experiments shows little difference in thecalculated Kd for either the Y82F nuclease or wild-type protein (35 nM versus 10 nM, respectively). Incontrast, the Y82L mutant had a greatly increasedKd (450 nM). Typical results are shown in Figure 4.Smearing is more apparent in the case of the Y82Fand Y82L protein–DNA complexes, suggestingthat the complexes formed may dissociate inthe gel more readily than the wild-type enzyme-substrate complex.

Table 1. Specific activity of wild-type and mutantnucleases

Enzyme (units/mg)a

Cofactor MgCl2 (mM) WT Y82F Y82L

1 18.0 2.2 1.95 23.8 3.8 2.910 23.9 3.9 3.320 19.5 4.3 3.530 19.6 4.2 3.450 18.3 4.1 2.8

Exonucleolytic digestion of high molecular weight DNA wasassayed spectrophotometrically as described in Materials andMethods. Specific activity for each reaction was determinedfrom at least three separate assays.

a The activity unit corresponds to the release of 1 nmol of acidsoluble nucleotide per minute in the standard spectrophoto-metric assay at 37 8C.

Figure 2. (a) The pseudo-Y structure used in characterisation of the cleavage specificity, and binding constants in theabsence of cofactor, of T5 50 nuclease and Y82F mutant. Arrows indicate the sites of reaction to produce endonucleo-lytic (21, 19, and 14 nucleotides) and exonucleolytic products (trimers and pentamers). (b) The phosphorylated 50-over-hanging hairpin, pHP1, used to determine steady-state parameters of the wild-type enzyme. The oligonucleotide is50-32P-labelled in kinetic studies. An arrow indicates the site of cleavage. (c) The fluorescent 50-overhanging hairpin,FpHP1, used to determine the steady-state parameters of the wild-type and Y82F mutant. An arrow indicates the siteof cleavage. (d) A generic flap structure is composed of a flap strand, a template strand and an adjacent strand.

Role of Tyrosine 82 in Phage T5 Flap Endonuclease 1027

Determination of steady-statekinetic parameters

A fluorescent (FpHP1), or (50-32P)-radiolabelled(pHP1) 29-mer 50-overhanging hairpin oligonucleo-tide was employed for the steady-state kineticassays (Figure 2(c)). The sequence of this oligo-nucleotide is identical to that used previouslyto determine steady-state parameters of T550-nuclease in a radioactive assay (Figure 2(b)),18

except that the 50-radioactive phosphate is replacedwith 50-fluorescein alkyl phosphate derivative. Toascertain whether the introduction of a fluorophorechanged the cleavage site of FpHP1 the products ofthe reaction were characterised using MALDI-TOFMS. The molecular weight of the products, 2941.7for F-pd(CGCTGTCG) (calculated Mr 2937.05) and6479.6 for pd(AACACACGCTTGCGTGTGTTC)(calculated Mr 6475.17), respectively, indicatedthat reaction occurred one nucleotide distal to thejunction between single and double-strandedDNA analogously to the reaction observed withpHP1.18 After enzyme-catalysed reaction andquenching with ETDA, the fluorescent 8-merproduct of the reaction was separated from the29-mer FpHP1 using denaturing HPLC withfluorescence detection (Figure 5) and the amountof product was quantified by peak integration.Despite the presence of the 50-phosphate diester inFpHP1 it was digested at a similar reaction rate topHP1. By varying the substrate concentrationaround the KM of the enzyme, initial rates ofreactions at various substrate concentrations weredetermined and then fitted to the Michaelis–Menten equation by non-linear regression analysis.The results of kinetic analysis are shown in Table 2.The substrates pHP1 and FpHP1 yield similarkinetic parameters for the wild-type protein.Replacement of tyrosine 82 with phenylalanineproduced no significant change in kcat but increasedKM by approximately 280-fold when assayed usingeither the fluorescent or radiolabelled substrate.Increasing the concentration of MgCl2 used todetermine kinetic parameters from 10 mM to20 mM did not significantly change the parametersobtained for either the wild-type or mutant protein.

Figure 3. Action of wild-type and mutant enzymes onpseudo-Y substrate. Wild-type (WT) or mutant (Y82F orY82L) enzymes, at the concentrations indicated, wereincubated with the substrate in the presence of 10 mMmagnesium chloride (Mg2þ) for five and 30 minutes at37 8C (time course indicated by black triangles), asdescribed in Materials and Methods. Reaction productswere separated by denaturing PAGE. Lane C1: controllane, pseudo-Y substrate incubated with 10 mMmagnesium chloride (Mg2þ) for 30 minutes in theabsence of enzyme. Lane C2: control lane, pseudo-Ysubstrate incubated with 6 nM wild-type enzyme for 30minutes in the absence of metal cofactor. Size ofproducts, in nucleotides, is indicated to the right of theFigure. The 14-mer product is produced from the single-stranded labelled oligonucleotide when incubated withthe T5 nuclease. It results from endonucleolytic cleavageof a transient, partly double-stranded, minor confor-mation of the substrate.19

Table 2. The kinetic parameters of wild-type and mutant T5 50 nucleases

Substrate Enzyme [Mg2þ] (mM) KM (mM) kcat (min21) kc (min21)

pHP1 WT 10 0.04 ^ 0.01 110 ^ 10 NDFpHP1 WT 10 0.05 ^ 0.01 155 ^ 9 346 ^ 26FpHP1 WT 20 0.03 ^ 0.01 128 ^ 22 NDpHP1 Y82F 10 13 ^ 1 69 ^ 3 NDFpHP1 Y82F 10 14 ^ 4 74 ^ 7 220 ^ 73FpHP1 Y82F 20 12 ^ 5 129 ^ 18 ND

The substrates used in this study are as follows: pHP1 is 50-phosphorylated HP1 and FpHP1 is fluorescently labelled HP1 (Figure 2).The values of KM and kcat for the wild-type protein using pHP1 are taken from reference.26 Other data are the result of at least threeindependent experiments conducted as described in Materials and Methods. The catalytic parameters KM and kcat were determinedfrom steady-state experiments, whereas kc is the maximal rate of reaction under single turnover conditions.


Determination of single-turnover ratesof reaction

Using the fluorescent hairpin substrate (FpHP1,Figure 2(c)), single-turnover reaction rates weremeasured for mutant and wild-type protein usingquench-flow techniques, under conditions where[S] , KM , [E]. Concentrations of substrate wereselected so that [S] was 40–50-fold lower than KM

and the concentration of enzyme was seven- toeightfold greater than KM. The respective concen-trations of substrate and enzyme were chosen tomeasure the maximal rate of reaction under pre-steady state conditions, termed kc, in order toinvestigate whether product release was rate-limiting in the steady-state reaction. A mixture of

NaOH (0.5 M final concentration) and EDTA(6.6 mM final concentration) was used to stopthese reactions as EDTA alone, even at muchhigher concentrations, did not provide a fastenough quench on the millisecond timescale.Problems with EDTA as a quench in rapidreactions have been previously documented.24,25

EDTA was maintained in the quench buffer asthis prolonged the lifetime of the reversed phasecolumn used to analyse the reactions. The resultsof the single turnover analysis are shown in Table2. The maximal rate of reaction under single turn-over, kc, for the wild-type and mutant proteinswere 2.2 and 2.9-fold greater than the respectivesteady-state maximal rates, kcat. In line with thesteady-state analysis, replacement of Y82 withphenylalanine produced no significant differencein the rate of reaction under single turnoverconditions, when compared with the wild-typeenzymatic reaction.

Variation of kinetic parameters with pH

The pH dependence of the second-order rateconstant, log(kcat/KM), for the wild-type enzymehas previously been determined to be bell shapedwith pKa values of 8.3 and 8.4.26 Second-order rateconstants were determined for Y82F at a range ofpH values between 5.5 and 10.5 under pseudofirst-order conditions where KM . [E] . [S]. Aswas found in the case of the wild-type enzyme,the pH dependence of log (kcat/KM) of Y82F firstincreases and then decreases with increasing pH.This demonstrates that the rate of conversion ofmutant enzyme-substrate complex to free enzymeand products is dependent on both a deproto-nation and a protonation event (Figure 6). Thebell-shaped curve fits, with an R factor of 0.97, tothe double ionisation described in equation (3)27 inMaterials and Methods. Here the enzyme isconsidered to be present in three ionisation states

Figure 4. Substrate binding bythe T5 50 nucleases. (a) Pseudo-Ysubstrate was incubated with thewild-type, Y82F, or Y82L enzymeson ice (top, middle and lowerpanels, respectively). The enzyme-substrate complexes were separatedfrom the unbound substrate byelectrophoresis on a non-denatur-ing 17% polyacrylamide gel. Thesubstrate concentration was con-stant, the enzyme concentrationwas varied between the limitsshown. (b) Data from the gel retar-dation (EMSA) experiments wereplotted as percentage of freesubstrate against the log10 enzyme

concentration. Results shown are from three separate experiments. The enzyme concentration required to bind halfthe substrate was determined graphically. The dissociation constants for wild-type, Y82F and Y82L mutant T5nucleases were 10 nM, 35 nM and 450 nM, respectively.

Figure 5. Denaturing HPLC trace showing separationof the 29-mer FpHP1 (Figure 2(c)) (retention time 7.2minutes) from the 8-mer product of the T5 50 nucleasecatalysed reaction Fpd(CGCTGTCG) (retention time 3.5minutes) visualised using a fluorescence detector.Integration of the HPLC peaks allows the amount ofproducts and reactants to be determined and hencerates of reaction to be calculated. Conditions used fordenaturing HPLC are described in Materials andMethods.


(doubly protonated, singly protonated and unpro-tonated) but only the mono-protonated form ofthe enzyme undergoes reaction. Curve fittingyields an optimal second-order rate constant of(kcat/KM)0 ¼ 5.7(^3.0) £ 107 M21 min21 and pKa

values of 8.4(^0.3) and 9.5(^0.4) for the freemutant protein.

Discussion

Structural studies identify a tyrosine residue inthe active sites of the flap endonucleases crystal-lised to date.5 – 9 However, the positions of theequivalent tyrosine residues in the primarysequence of the flap endonucleases are not strictlyconserved despite their similar locations withinthe respective active sites (Figure 1(b)).6 Forexample, the active site tyrosine in the bacterialenzymes and phage T5 are located adjacent to aconserved lysine residue within the N-terminalhalf of the polypeptide, i.e. Y82 is adjacent to K83,whereas, in the archaebacterial enzymes such asthat from Methanococcus jannaschii, the equivalenttyrosine is residue 225 adjacent to a metal-ligand-ing D224 in the polypeptide chain, i.e. toward theC terminus of the protein.8 In addition to thecarboxylates, this tyrosine is the only strictlyconserved active-site amino acid present in all theflap endonucleases identified to date includingthose from the archaebacteria and eukaryotes suchas human flap endonuclease 1.6 In order to eluci-date the role of the active site tyrosine, we selectedthe conservative Y82F mutation in the T5 flapendonuclease for detailed analysis as this substi-tution proved to be less disruptive than the Y82Lreplacement.

Development of a single-cleavage kineticassay using a fluorescent substrate

A non-radioactive assay was developed to facili-tate kinetic analysis. The 50-overhanging substratewas selected for this assay as it gives fast reactionrates and undergoes only a single cleavage underthe reaction conditions employed. This greatlysimplifies kinetic analysis. Previous determinationsof the catalytic parameters of T5 50 nuclease haveemployed an assay using a radiolabelled hairpinsubstrate.26 Products and reactants were separatedby gel electrophoresis and then quantified byphosphor imaging. In addition to the hazards ofradioactivity, this assay was time consuming. Itrequired the performance of several electro-phoresis runs, followed by drying of the gels andexposure to phosphorescent plates to quantify thesubstrate and product. Particular problems wereassociated with low substrate concentrations, suchas those that may be employed in pre-steady-stateanalysis, even when they were labelled to highspecific activity. To attain detectable levels andavoid extended exposures, either concentration ofsamples prior to gel electrophoresis was requiredor larger loading volumes had to be used. Thisresults in fewer samples per run and increases thenumbers of gels required for each kinetic analysis.These deficiencies are circumvented using denatur-ing HPLC and fluorescence-based detection, wherepeak areas can be directly quantified upon separ-ation and detection and injection volume can beeasily varied according to the concentration ofsubstrate assayed.

Addition of a 50-fluorophore to HP1 did notchange the cleavage specificity of the enzyme,which undergoes a single reaction with this sub-strate in a kinetic timescale. The ability to separate50-fluorescent products from substrate usingdenaturing HPLC is a critical feature of this assay.Separation of fluorescent oligomers of differinglengths is generally problematic using reversedphase chromatography as the hydrophobicity ofthe fluorophore tends to dominate retention times.However, ion-pair chromatography, using bufferscontaining tetrabutyl ammonium bromide, allowssingle nucleotide resolution of fluorescentoligomers.28 For the wild-type protein, the fluor-escent assay produces very similar kinetic para-meters to those derived from the radioactive assaywith comparable errors, but allows completekinetic determinations on an eight hour ratherthan two to three day timescale. This assay wastherefore used for detailed kinetic evaluation ofthe mutant protein.

Characterisation of the mutant protein

Substitution of the conserved tyrosine in this flapendonuclease with phenylalanine resulted in amodest increase in dissociation constant (3.5-fold)in comparison with the parent enzyme in theabsence of cofactor. This suggests that loss of the

Figure 6. (a) The pH-dependence of the second-orderrate constant of the Y82F mutant enzyme. Conditionsfor the reactions are described in Materials and Methods.The exponential percentage of product versus time wasfitted to equation (3) to yield (kcat/KM)[E]. The data havebeen fitted to equation (4), with an R factor of 0.93,to give (kcat/KM)0 ¼ 5.7(^3.0) £ 107 M21 min21 and pKa

values of 8.4(^0.3) and 9.5(^0.4) for the free mutantenzyme.


hydroxyl moiety has no gross deleterious effect onprotein structure. However, the less conservativesubstitution with leucine resulted in a greatlyincreased dissociation constant, presumably dueto perturbation of the structure of the protein. Allthree enzymes possessed both exo- and endo-nuclease activity as judged by their reaction witholigonucleotide substrates (Figure 3). The majorproducts produced by the wild-type enzymemigrated as trimers and pentamers, whereas theY82F and Y82L proteins produced relatively littleof these shorter, exonucleolytic, products. Thiswas mirrored in the reduced specific activitymeasurements determined using the spectrophoto-metric assay (Table 1). However, the Y82F enzymedid not show any large defect in turnover as deter-mined by the steady-state single-cleavage assayused to determine kcat (Table 2).

During steady-state analysis it is possible that achange in the rate of the chemical reaction couldbe masked if product release is rate-limiting.Product release is known to be the rate-limitingstep in the reaction of the homologous E. coli DNApolymerase I 50 nuclease with a much larger DNAflap substrate (Figure 2(d)). The maximal rate ofreaction of this enzyme under single-turnoverconditions is tenfold faster than that measuredunder steady-state.22 To test this possibility in theT5 50 nuclease case, we examined the maximalrate of the reaction, kc, under single turnover con-ditions using a quench-flow apparatus (Table 2).In a simple model of 50 nuclease reactions thatassumes magnesium ion binding steps are fastand can therefore be ignored, substrate bindsreversibly to enzyme, is converted to productsand the products dissociate, then the single-turnover rate of reaction kc is the rate of the chemi-cal conversion. The single-turnover rates of boththe mutant and wild-type proteins are two tothreefold faster than those measured understeady-state conditions, indicating that the steady-state reactions of FpHP1 with both the wild-typeand mutant proteins are rate-limited by a combi-nation of product release and chemistry. However,in analogy to the conclusions reached fromthe steady-state experiments, the mutation Y82Fproduces no significant change in the rate of theenzyme reaction under single-turnover conditions.

The wild-type enzyme exhibits bell-shaped pHdependence for its second-order rate constant(kcat/KM). This indicates the presence of twoionisable functions within the enzyme’s active siteon which catalysis depends with pKa values of 8.3and 8.4, respectively. This bell-shaped curve forthe second-order rate constant arises because theformation of ternary complex reaches a maximumat low pH, whereas the kcat parameter reaches amaximum at high pH.18 It has been demonstratedthat the titrating residue responsible for the pHdependence of the association constant is lysine83, as mutation of this residue to alanine causesthis parameter to become pH-independent.18 Tyr82(pKa free amino acid side-chain 10.0729) was a

possible candidate for the second ionising aminoacid involved in the enzyme-catalysed reaction.However, the lack of any effect of mutation tophenylalanine on the turnover number of theenzyme at pH 9.3 suggests that this is unlikely. Tocompletely eliminate this possibility, we measuredthe pH rate-profile of the mutant protein andfound it to be bell-shaped, i.e. similar to thewild-type protein (Figure 6). Thus, the tyrosinehydroxyl group does not play a role in chemicalcatalysis carried out by this protein. Interestingly,upon removal of the tyrosine hydroxyl group, ashift in one of the pKa values obtained for theenzyme by one pH unit does occur (wild-typeenzyme pKa values 8.3 and 8.4; Y82F pKa values8.4 and 9.5). This indicates that the absence of thetyrosine hydroxyl group influences the pKa of oneof the ionising moieties on which enzyme catalysisdepends but is not directly involved in catalysisitself.

Y82F is defective in KM

In contrast to the Kd result obtained in the gelretardation assay in the absence of cofactor, a largeincrease in KM was observed in the steady-statecatalytic assay. In more complex kinetic pathwaysthan the Michaelis–Menton model, KM is a compo-site of several rate constants including the rate ofthe chemical step of the reaction.30 The possibilityof an increased KM which results from the chemicalor product release steps of the mutant proteinbeing very much slower than that of the wild-typeenzyme can be ruled out from the results of thesingle turnover experiments. Our results thereforesuggest that the phenolic hydroxyl group of thetyrosine residue plays a role in assembly of aternary complex, i.e. formation of a protein–DNA-cofactor ground-state intermediate (Table 2).For the wild-type enzyme KM is independent ofdivalent metal ion concentration, at least in theconcentration range that metal ion concentrationdependence of kcat is observed (i.e. high mM, lowmM, M.R.T., J.R.S. and J.A.G., unpublished results).However, to test if the deficiency in KM of the tyro-sine mutant could be compensated by increasedmetal ion concentration we measured the catalyticparameters at higher magnesium ion concen-tration, but found these to be very similar tothose determined at 10 mM magnesium chloride(Table 2). Thus the observed increase in KM doesnot seem to be a consequence of impairment ofmetal binding.

The equivalent conserved tyrosine residue inT4 RNaseH, another flap endonuclease, has pre-viously been mutated to phenylalanine.20 The T4RNaseH mutant was reported to have similar (butslightly reduced) exonuclease activity to theparental protein on double-stranded DNA. Bind-ing affinity for flap substrate was similar to thatfound for the wild-type enzyme. It was concludedthat the mutation did not affect binding or catalysisas qualitatively identical endonucleolytic activity


was observed with a DNA flap substrate. Theseconclusions appear to contradict those reportedhere. However, the two studies employed differentassays and substrates making comparisons diffi-cult, especially as detailed quantitative kineticresults were not reported in the T4 RNaseHstudy.20 It may be that there are subtle mechanisticdifferences between these enzymes. Indeed, theconserved tyrosine (Y87) is closer to metal site 2in the crystal structure of the T4 homologue,9

whereas, in the T5 enzyme, the conserved tyrosine(Y82) is nearer metal site 1.7

Earlier work on the homologous E. coli DNApolymerase I 50 nuclease domain reported thatmutation of the active site tyrosine to cysteineresulted in a lack of catalytic activity.21 Veryrecently it has been reported that mutation of theequivalent tyrosine to alanine in the E. coli DNApolymerase I 50 nuclease domain, resulted in anenzyme with 1000-fold reduced rate of reactionbut an unchanged dissociation constant comparedto the wild-type protein.22 In this study, rates ofreaction were measured under single turnoverconditions, analogously to the rapid quench datapresented here. These experiments use a largerflap-type, rather than hairpin, substrate (Figure2(d)). In contrast our experiments are conductedwith a 50-overhanging hairpin substrate (Figure2(c)), which effectively lacks half the template andadjacent strands of a flap substrate. It is difficultto envisage how a change in substrate could giverise to differences in the results obtained as inboth cases the tyrosine residue is predicted to bewithin the active site, and should therefore interactsimilarly with each substrate. Another possibleexplanation is that the substitution of alanine fortyrosine has a deleterious effect on the confor-mation of important residues in the active site.The phenyalanine for tyrosine substitution studiedhere can be considered a more conservative changethan replacement with alanine in the E. colihomologue.22 This may explain why the lattersubstitution resulted in a 1000-fold reduction ofkcat. However, the lack of any differences in thedissociation constant observed in the 50 nucleasedomain of E. coli Pol I is difficult to reconcile withour results. Curiously, substitutions of theconserved lysine residues were found to be highlydetrimental to KM in our phage enzyme20,26

whereas the analogous lysine to alanine mutationproduced no changes in dissociation constant inthe 50 nuclease domain of Pol I.22 It is possible thatthis may reflect subtle differences in the usage ofthese active site amino acids in binding DNAsubstrates. The phage enzymes may use bindinginteractions to stabilise ground state ternary com-plexes of protein, with DNA and cofactor, whilstthe bacterial enzymes may utilise the bindingenergy derived from these interactions to supportcatalysis.

It is clear from this study on the T5 enzyme thatsubstitution of tyrosine by phenylalanine does notimpact on gross protein structure, or substrate

binding in the absence of cofactor. Thus, we canbe confident that loss of the phenolic hydroxylfunction is responsible for differences in ability toform the ground state enzyme–DNA–metal ioncomplex seen between this and the wild-typeenzyme. The fact that kcat and kc are unaffected,and the pH-rate profile has the same bell-shapedprofile as wild-type protein, demonstrates that thetyrosine phenolic hydroxyl group does not playany role in chemical catalysis.

Given that Y82 is not involved in catalysis in the50 nucleases, an explanation for the pH dependenceof kcat is still required. Previous studies haveindicated that K83 is not the residue responsiblefor the pH dependence of kcat, although its protona-tion is required for high-affinity DNA binding.Other candidate ionisable groups in the active siteinclude the metal co-ordinating aspartyl andglutamyl residues (pKa free amino acid side-chains3.86 and 4.2529). Alternatively, the pH dependenceof kcat may reflect the pKa of a water moleculebound to one of the essential metal ions (pKa

Mg(H2O)62þ 11.4231). In either case, the environment

of the protein would have to perturb the pKa of theamino acid or metal-bound water by several pHunits. Further studies will be required to clarifywhich of these possibilities is taking place andshed further light on the mode of catalysis that isutilised by this family of enzymes.

Materials and Methods

Site-directed mutagenesis and productionof enzymes

Oligonucleotide site-directed mutagenesis was carriedout on a single-stranded M13 derivative carrying thecloned T5 D15 exonuclease gene.32 The phosphoro-thioate-based high efficiency mutagenesis procedure33

was used to alter the tyrosine 82 codon, resulting in asubstitution of tyrosine to phenylalanine using theprimer 50-d(CATCACGATTACCTTTAAACTCTGGTA-GATG); the codon changing mutation is shown under-lined. A similar oligonucleotide in which the underlinedsequence consisted of AAG produced a Y82L mutation.Dideoxy sequencing was used to determine that onlythe desired sequence changes had been introduced.34

The mutated genes were subcloned into the expressionvector pJONEX4 and expressed as described for thewild-type exonuclease gene.32,35 Induced cells wererecovered by centrifugation at 13,000g for 20 minutes at4 8C. Wild-type and mutant proteins were purified tohomogeneity as described,32 except that an initial ureasolubilisation stage was included for the mutant protein,as it was insoluble on over-expression. Protein concen-tration was determined by the method of Bradford.36

Structure-specific cleavage ofoligonucleotide substrates

Structure-specific DNA cleavage was examined using50-32P-end-labelled pseudo-Y substrate (Figure 2(a)) asdescribed17 in 25 mM potassium glycinate (pH 9.3),10 mM MgCl2, 100 mM KCl, 0.5 mg/ml acetylatedBSA, 2% (v/v) glycerol, 0.2 mM EDTA, 0.2 mM DTT.


Pseudo-Y substrate was prepared as described above.The reactions were incubated at 37 8C and 10 ml aliquotstaken at intervals during the reaction. All reactions werestopped by addition to 10 ml stop mix (95% formamide,15 mM EDTA, 200 mg/ml bromophenol blue). Theproducts of the reactions were analysed on a 7 M urea,15% (w/v) polyacrylamide gel, buffered in 1 £ TBE at50 W for two hours. The gel was visualised witha Molecular Imager, and Molecular Analyst v2.0.1software (BioRad, USA). Results are shown in Figure 3.

Electrophoretic mobility shift assay

Electrophoretic mobility shift assays (EMSA) werecarried out in the absence of cofactor on 50-32P-end-labelled pseudo-Y substrate as described.17 The pseudo-Y structure was made by annealing the two oligo-nucleotides 50-GATGTCAAGCAGTCCTAACTTTGAGG-CAGAGTCC and 50-GGACTCTGCCTCAAGACGGTA-GTCAACGTG. This produces a structure which is fullybase-paired at one end only (Figure 2(a)). Reactionswere performed in 25 mM potassium glycinate (pH 9.3),100 mM KCl, 1 mM EDTA, 5% glycerol, 1 mM DTT,0.1 mg/ml acetylated BSA. The reactions were incubatedon ice for ten minutes and analysed on a 17% nativepolyacrylamide gel, buffered in 100 mM Tris–Bicine(pH 8.3), 2 mM EDTA, 1 mM DTT at 4 8C for two hoursat 15 V cm21. The gel was analysed using a MolecularImager, and Molecular Analyst v2.0.1 software (BioRad,USA). Results are shown in Figure 4.

Exonuclease activity assay

This UV spectroscopy-based assay demonstrates thepresence of (predominantly) exonucleolytically derived,acid-soluble degradation products. The release of acid-soluble nucleotides from high molecular weight DNA(herring sperm Type XIV, Sigma) was determined with astandard UV spectrophotometric assay essentially asdescribed.32 The 600 ml assay mixture contained 2 mMType XIV DNA (in nucleotides), 25 mM potassiumglycinate (pH 9.3), 1 mg enzyme, and MgCl2 cofactor asindicated in Table 1. Curves were plotted from the dataobtained (from at least triplicate experiments) andestimates of the specific activity derived from initialvelocity were calculated.

Synthesis, purification and characterisation of 50

fluorescein-labelled HP1 (FpHP1)

The fluorescent oligonucleotide hairpin substrate(FpHP1: 50 fluorescein-d(pCGCTGTCGAACACACGCT-TGCGTGTGTTC)) (Figure 2(c)) was synthesised usingan ABI model 394 DNA/RNA synthesiser using 50-fluor-escein reagent (Glen Research via Cambio) and standardreagents and conditions for other base additions. Follow-ing deprotection, the oligonucleotide was purified by RPHPLC (Alltech Alphabond C18 column; buffer A, 50 mMtriethylammonium acetate (TEAAc) pH 6.5, 5% (v/v)acetonitrile; buffer B, 50 mM TEAAc pH 6.5, 65% aceto-nitrile; gradient t ¼ 0 minute 1% B; t ¼ 5 minutes 20% B;t ¼ 22 minutes 40%B; retention time 13.5 minutes).Following HPLC the oligonucleotide was desalted usinga Sep-Pak C18 column. The oligonucleotide was thenwashed with five column volumes of 0.1 M TEAAc (pH7.0), followed by ten column volumes of deionisedwater and eluted with 1.5 ml of 60% solution ofmethanol. The eluted oligonucleotide was evaporated to

dryness and resuspended in deionised water. The extinc-tion coefficient used for FpHP1 was 280.1 cm2 mmol21 ascalculated previously.26

MALDI-TOF mass spectral analysis of cleavageproducts of FpHP1

FpHP1 (100 mM) (Figure 2(c)) was incubated at 37 8Cin a solution of 25 mM potassium glycinate (pH 9.3),50 mM KCl, and 10 mM MgCl2. T5 50 nuclease was thenadded to a concentration of 72 nM to initiate the reaction.Aliquots of reaction mixture were taken at intervals fromfive to 60 minutes, quenched by the addition of equalvolume of 20 mM EDTA. Salts were removed usingDowex beads (NH4

þ form) and the sample (1 ml) wasmixed with 2 ml of matrix (50 mg/ml 3-hydroxypicolinicacid 3:1 H2O:CH3CN). Samples were then submitted toMALDI-TOF mass spectrometry in positive ion modeon a Bruker Reflex III instrument to confirm their identi-ties. The molecular weights were 9399.4 for F-pd-(CGCTGTCGAACACACGCTTGCGTGTGTTC) (FpHP1)(calculated Mr 9396.22), 2941.7 for F-pd(CGCTGTCG)(calculated Mr 2937.05) and 6479.6 for pd(AACA-CACGCTTGCGTGTGTTC) (calculated Mr 6475.17).

Determination of Michaelis–Menten parameters

Aliquots of appropriate concentrations of FpHP1(Figure 2(c)) in 25 mM potassium glycinate (pH 9.3),50 mM KCl, 0.2 mg/ml bovine serum albumin (BSA)and 10 mM or 20 mM MgCl2 were prepared. These reac-tion mixtures were incubated at 90 8C for two minutes,and then allowed to cool to 37 8C. The enzyme stocksolution was diluted in 25 mM potassium glycinate (pH9.3), 10 mM MgCl2, 50 mM KCl, 0.2 mg/ml BSA andkept on ice until used. Reactions were initiated byaddition of enzyme (final reaction volume, 500 ml or100 ml; 25 mM potassium glycinate (pH 9.3), 50 mMKCl, 0.2 mg/ml BSA and 10 mM or 20 mM MgCl2) andbrief vortexing. The final concentrations of the substratewere 10 nM–800 nM for the wild-type enzyme and1 mM–100 mM for Y82F. The concentration of wild-typeenzyme was varied from 89.5 pM to 144 pM and that ofY82F from 10 nM to 30 nM. Aliquots (50 ml or 10 ml)were removed at eight appropriate time intervals andquenched by immediate addition of an equal volume of20 mM EDTA. Denaturing HPLC was used to determinethe extent of reaction. A Wavew DNA Fragment AnalysisSystem equipped with fluorescence detection was usedto separate and quantify the products (Transgenomic,Warrington UK). A DNA Sepw cartridge was used inconjunction with the following buffers; buffer A,2.5 mM tetrabutyl ammonium bromide, 1 mM EDTA,0.1% acetonitrile; buffer B, 2.5 mM tetrabutyl ammoniumbromide, 1 mM EDTA, 70% acetonitrile; gradient t ¼ 0minute, 30%B; t ¼ 1 minute, 50%B; t ¼ 13 minutes,80%B; retention times fluorescent 8-mer product 3.5minutes, fluorescent 29-mer substrate 7.2 minutes.

Initial rates of the cleavage reaction at the various sub-strate concentrations were determined. This allowed thekinetic parameters to be calculated by non-linearregression fitting of the data to the Michaelis–Mentenequation (equation (1)). All curve fitting was carried outusing Kaleidagraph Software (Synergy Software,Reading, PA, USA).

n

½E�¼

kcat½S�

KM þ ½S�ð1Þ


where n is initial rate, [S] is substrate concentration, [E] istotal enzyme concentration.

Rapid quench experiments

Rapid quench experiments were carried out at 37 8Cusing a RQF-63 device from Hi-Tech Ltd (Salisbury,UK). An 80 ml aliquot of enzyme in reaction buffer(25 mM potassium glycinate (pH 9.3), 10 mM Mg2Cl,50 mM KCl, 0.1 mg/ml BSA) was mixed with an equalvolume of FpHP1 in reaction buffer. Wild-type (WT)and Y82F mutant enzymes were used at final concen-trations of 410 nM and 100 mM, respectively. FpHP1 wasused at a final concentration of 1 nM in WT reactionsand 325 nM in Y82F reactions. After a controlled timedelay of 9 ms to 6.4 seconds, 80 ml of quench (1.5 Msodium hydroxide, 20 mM EDTA) was added. Solutionsof enzyme, substrate and quench were held in a thermo-stated bath within the instrument during the reactions.The quenched reaction mixtures were recovered fromthe sample recovery loop and were analysed usingdenaturing HPLC as described previously for thesteady-state analyses. The first-order rate of the reactionwas determined by plotting the appearance of productagainst time and by non-linear regression fitting toequation(2):

Pt ¼ P1ð1 2 exp2kctÞ ð2Þ

where Pt is the amount of product at time t, P1 is theamount of product at time 1, and kc is the rate ofreaction.

pH dependence of the Michaelis–Mentenparameters for Y82F

Separate reaction mixtures of appropriate concen-trations of FpHP1 (Figure 2(c)) in 25 mM buffer, 10 mMMgCl2, 50 mM KCl, 0.2 mg/ml BSA were prepared.These reaction mixtures were incubated at 90 8C for twominutes, and then allowed to cool to 37 8C. The enzymestock solution was dissolved in 25 mM buffer, 10 mMMgCl2, 50 mM KCl, 0.2 mg/ml BSA and kept on iceuntil use. The buffers employed to maintain pH werepH 5.6–6.0 Mes, pH 6.5–8.0 Hepes, pH 8.0–8.4 Epps,pH 9.0–10.00 Ches, pH 9.3 potassium glycinate, pH 10.5Caps. Reactions were initiated by addition of enzyme(final concentrations: 25 mM buffer, 10 mM MgCl2,50 mM KCl and 0.2 mg/ml BSA) and brief vortexing.Final substrate and enzyme concentrations were 1 nMand 100 nM–300 nM, respectively. Reactions werefollowed by taking aliquots at appropriate time points.Individual samples were quenched by immediateaddition an equal volume of 20 mM EDTA and analysedusing a Wavew DNA Fragment Analysis System asdescribed above. Rates of reaction at the various sub-strate concentrations were determined. This allowed thekinetic parameters to be calculated by plotting theappearance of product against time and fitting equation(3) by non-linear regression. All curve fitting was carriedout using KaleidaGraph Software (Synergy Software,Reading, PA, USA):

%Product ¼ 100 1 2 expkcat

KM

� �½E�t

!ð3Þ

The pH dependence of (kcat/KM)obs was fitted to a double

ionisation using equation (4):27

kcat

KM

� �obs

¼

½Hþ�KE1kcat

KM

� �0

KE1KE2 þ ½Hþ�KE1 þ ½Hþ�2ð4Þ

where KE1 and KE2 are acid dissociation constants of thefree enzyme, (kcat/KM)obs is the second-order rate of reac-tion at a given pH and (kcat/KM)0 is the maximal second-order rate constant of the singly protonated, catalyticallycompetent, form of the enzyme.

Acknowledgments

We thank Simon Thorpe for recording the massspectral data. This work was supported by grants fromThe Wellcome Trust and the Biotechnology and Bio-logical Sciences Research Council (grant references052123 and 50/B11990, respectively). J.A.G. is a BBSRCAdvanced Research Fellow.

References

1. Lyamichev, V., Brow, M. A. D. & Dahlberg, J. E.(1993). Structure-specific endonucleolytic cleavageof nucleic acids by eubacterial DNA polymerases.Science, 260, 778–783.

2. Harrington, J. J. & Lieber, M. R. (1994). Thecharacterisation of a mammalian DNA structure-specific endonuclease. EMBO J. 13, 1235–1246.

3. Xie, Y. L., Liu, Y., Argueso, J. L., Henricksen, L. A.,Kao, H. I., Bambara, R. A. & Alani, E. (2001). Identifi-cation of rad27 mutations that confer differentialdefects in mutation avoidance, repeat tract insta-bility, and flap cleavage. Mol. Cell. Biol. 21,4889–4899.

4. Bhagwat, M. & Nossal, N. G. (2001). BacteriophageT4 RNase H removes both RNA primers andadjacent DNA from the 50 end of lagging strandfragments. J. Biol. Chem. 276, 28516–28524.

5. Kim, Y., Eom, S. H., Wang, J. M., Lee, D. S., Suh, S. W.& Steitz, T. A. (1995). Crystal structure of Thermusaquaticus DNA polymerase. Nature, 376, 612–616.

6. Hosfield, D. J., Mol, C. D., Shen, B. H. & Tainer, J. A.(1998). Structure of the DNA repair and replicationendonuclease and exonuclease FEN-1: couplingDNA and PCNA binding to FEN-1 activity. Cell, 95,135–146.

7. Ceska, T. A., Sayers, J. R., Stier, G. & Suck, D. (1996).A helical arch allowing single-stranded DNA tothread through T5 50-exonuclease. Nature, 382, 90–93.

8. Hwang, K. Y., Baek, K., Kim, H. Y. & Cho, Y. (1998).The crystal structure of flap endonuclease-1 fromMethanococcus jannaschii. Nature Struct. Biol. 5,707–713.

9. Mueser, T. C., Nossal, N. G. & Hyde, C. C. (1996).Structure of bacteriophage-T4 RNASE-H, a 50 to30 RNA–DNA and DNA–DNA exonuclease withsequence similarity to the RAD2 family of eukaryoticproteins. Cell, 85, 1101–1112.

10. Shen, B., Nolan, J. P., Sklar, L. A. & Park, M. S. (1996).Essential amino acids for substrate binding andcatalysis of human flap endonuclease 1. J. Biol.Chem. 271, 9173–9176.

11. Mizrahi, V. & Huberts, P. (1996). Deoxy- anddideoxynucleotide discrimination and identification


of critical 50 nuclease domain residues of the DNApolymerase I from Mycobacterium tuberculosis. Nucl.Acids Res. 24, 4845–4852.

12. Amblar, M. S. G. L. P. (1998). Purification and proper-ties of the 50-30 exonuclease D10A mutant of DNApolymerase I from Streptococcus pneumoniae: a newtool for DNA sequencing. J. Biotechnol. 63, 17.

13. Amblar, M. L. P. (1998). Purification and propertiesof the 50-30 exonuclease D190 ! A mutant of DNApolymerase I from Streptococcus pneumoniae. Eur. J.Biochem. 252, 124.

14. Xu, Y., Derbyshire, V., Ng, K., Sun, X. C., Grindley, N.D. F. & Joyce, C. M. (1997). Biochemical and muta-tional studies of the 50-30 exonuclease of DNA poly-merase I of Escherichia coli. J. Mol. Biol. 268, 284–302.

15. Shen, B., Nolan, J. P., Sklar, L. A. & Park, M. S. (1997).Functional analysis of point mutations in humanflap endonuclease-1 active site. Nucl. Acids Res. 25,3332–3338.

16. Murante, R. S., Rust, L. & Bambara, R. A. (1995). Calf50 to 30-exo/endonuclease must slide from a 50 endof the substrate to perform structure-specificcleavage. J. Biol. Chem. 270, 30377–30383.

17. Garforth, S. J. & Sayers, J. R. (1997). Structure-specificDNA binding by bacteriophage T5 50-30 exonuclease.Nucl. Acids Res. 25, 3801–3807.

18. Pickering, T. J., Garforth, S., Sayers, J. R. & Grasby, J.A. (1999). Variation in the steady state kineticparameters of wild type and mutant T5 50-30-exo-nuclease with pH—protonation of Lys-83 is criticalfor DNA binding. J. Biol. Chem. 274, 17711–17717.

19. Garforth, S. J., Ceska, T. A., Suck, D. & Sayers, J. R.(1999). Mutagenesis of conserved lysine residues inbacteriophage T5 50-30 exonuclease suggests separatemechanisms of endo- and exonucleolytic cleavage.Proc. Natl. Acad. Sci. USA, 96, 38–43.

20. Bhagwat, M., Meara, D. & Nossal, N. G. (1997).Identification of residues of T4 RNase H requiredfor catalysis and DNA binding. J. Biol. Chem. 272,28531–28538.

21. Joyce, C. M., Fujii, D. M., Laks, H. S., Hughes, C. M.& Grindley, N. D. F. (1985). Genetic mapping andDNA sequence analyis of mutations in the polAgene of Escherichia coli. J. Mol. Biol. 186, 283–293.

22. Xu, Y., Potapova, O., Leschziner, A. E., Sun, X. C.,Grindley, N. D. F. & Joyce, C. M. (2001). Contactsbetween the 50 nuclease of DNA polymerase I andits DNA substrate. J. Biol. Chem. 276, 30167–30177.

23. Paul, A. V. & Lehman, I. R. (1966). The deoxyribo-nucleases of Escherichia coli. J. Biol. Chem. 241,3441–3451.

24. Hyman, E. S. (1980). A study of the rate of chelationof magnesium by CDTA and EDTA in the ATP(Na/K)-ATPase system. Biochim. Biophys. Acta, 600,553–570.

25. Patel, S. S., Wong, I. & Johnson, K. A. (1991). Pre-steady-state kinetic analysis of processive DNAreplication including complete characterization ofan exonuclease-deficient mutant. Biochemistry, 30,511–525.

26. Pickering, T. J., Garforth, S. J., Thorpe, S. J., Sayers, J.R. & Grasby, J. A. (1999). A single cleavage assay forT5 50 ! 30 exonuclease: determination of the catalyticparameters for wild-type and mutant proteins. Nucl.Acids Res. 27, 730–735.

27. Fersht, A. R. (1985). Enzyme structure and mechanism,2nd edit., W.H. Freeman, New York.

28. Dickman, M. & Hornby, D. P. (2000). Isolation ofsingle-stranded DNA using denaturing DNAchromatography. Anal. Biochem. 284, 164–167.

29. Jencks, W. & Regenstein, J. (1976). Ionisation con-stants of acids and bases. In Handbook of Biochemistryand Molecular Biology (Fasman, G. R., ed.), vol. III,pp. 305–351, CRC Press, Boca Raton, FL.

30. Dixon, M. & Webb, E. C. (1964). Enzymes, 2nd edit.,pp. 98–100, Longmans, Freen and Co., London.

31. Huheey, J. (1983). Inorganic Chemistry, 3rd edit.,Harper and Row, New York.

32. Sayers, J. R. & Eckstein, F. (1990). Properties of over-expressed phage T5 D15 exonuclease. J. Biol. Chem.265, 18311–18317.

33. Sayers, J. R., Krekel, C. & Eckstein, F. (1992). Rapidhigh-efficiency site-directed mutagenesis by thephosphorothioate approach. Biotechniques, 13,592–596.

34. Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNAsequencing with chain-terminating inhibitors. Proc.Natl. Acad. Sci. USA, 74, 5463–5467.

35. Sayers, J. R. & Eckstein, F. (1991). A single-strandspecific endonuclease activity copurifies with over-expressed T5 D15 exonuclease. Nucl. Acids Res. 19,4127–4132.

36. Bradford, M. M. (1976). A rapid and sensitivemethod for the quantitation of microgram quantitiesof protein using the principle of protein–dyebinding. Anal. Biochem. 72, 248–254.

Edited by J. Karn

(Received 21 February 2002; received in revised form 3 March 2002; accepted 22 May 2002)


A Conserved Tyrosine Residue Aids Ternary Complex Formation, but not Catalysis, in Phage T5 Flap...

Documents

Transcript of A Conserved Tyrosine Residue Aids Ternary Complex Formation, but not Catalysis, in Phage T5 Flap...