`
`DNA Structure and Polymerase Fidelity
`
`Youri Timsit
`
`Institut de Biologie Physico-
`Chimique, CNRS - UPR 9080
`13, rue Pierre et Marie Curie
`75005, Paris, France
`
`The accuracy of DNA replication results from both the intrinsic DNA
`polymerase fidelity and the DNA sequence. Although the recent struc-
`tural studies on polymerases have brought new insights on polymerase
`fidelity, the role of DNA sequence and structure is less well understood.
`Here, the analysis of the crystal structures of hotspots for polymerase
`slippage including (CA)n and (A)n tracts in different intermolecular con-
`texts reveals that, in the B-form, these sequences share common structural
`alterations which may explain the high rate of replication errors. In par-
`ticular, a two-faced ‘‘Janus-like’’ structure with shifted base-pairs in the
`major groove but an apparent normal geometry in the minor groove con-
`stitutes a molecular decoy specifically suitable to mislead the poly-
`the rat polymerase b bound to this structure
`merases. A model of
`suggests that an altered conformation of the nascent template-primer
`duplex can interfere with correct nucleotide incorporation by affecting
`the geometry of the active site and breaking the rules of base-pairing,
`while at the same time escaping enzymatic mechanisms of error discrimi-
`nation which scan for the correct geometry of the minor groove.
`In contrast, by showing that
`the A-form greatly attenuates the
`sequence-dependent structural alterations in hotspots, this study suggests
`that the A-conformation of the nascent template-primer duplex at the
`vicinity of the polymerase active site will contribute to fidelity. The
`A-form may play the role of a structural buffer which preserves the cor-
`rect geometry of the active site for all sequences. The detailed comparison
`of the conformation of the nascent template-primer duplex in the avail-
`able crystal structures of DNA polymerase-DNA complexes shows that
`polymerase b, the least accurate enzyme,
`is unique in binding to a
`B-DNA duplex even close to its active site. This model leads to several
`predictions which are discussed in the light of published experimental
`data.
`
`# 1999 Academic Press
`
`Keywords: unusual DNA structure; spontaneous mutagenesis;
`microsatellite instability
`
`Introduction
`
`The accuracy of DNA replication is fundamental
`for the genetic stability of the cell. From bacteria to
`higher eukaryotes, error frequencies are remark-
`(cid:255)9 and10 (cid:255)10 perbaserepli-
`ablylow,between10
`cated(Echols&Goodman,1991).Theselow
`mutation rates are achieved by multiple steps of
`error discrimination including base selection by
`DNApolymerases,3 0-50 exonucleolyticproofread-
`ingandpost-replicativeDNArepair(Kunkel,
`1992).AlthoughtheDNAsynthesiserrorsplaya
`
`E-mail address of the corresponding author:
`timsit@ibpc.fr
`
`role in aging and disease, spontaneous mutations
`also provide the opportunity for genetic variation
`and are a primary basis for the evolution. Replica-
`tion errors occur non-randomly and result from an
`intricate interplay of intrinsic polymerase fidelity
`and DNA sequence effects.
`A principal determinant of polymerase fidelity is
`the geometric selection of nucleotide for insertion
`intoDNA(Echols&Goodman,1991;Goodman,
`1997).Crystallographicstudiesofoligonucleotide
`duplexes containing mismatched base-pairs have
`shown that
`the geometric equivalence and the
`pseudo 2-fold symmetry of a correct Watson-Crick
`base-pair is violated in mismatched bases, thus
`providing a geometrical means for selection by the
`
`
`
`ArticleNo.jmbi.1999.3199availableonlineathttp://www.idealibrary.comon
`
`DNA Structure and Polymerase Fidelity
`
`Youri Timsit
`
`Institut de Biologie Physico-
`Chimique, CNRS - UPR 9080
`13, rue Pierre et Marie Curie
`75005, Paris, France
`
`The accuracy of DNA replication results from both the intrinsic DNA
`polymerase fidelity and the DNA sequence. Although the recent struc-
`tural studies on polymerases have brought new insights on polymerase
`fidelity, the role of DNA sequence and structure is less well understood.
`Here, the analysis of the crystal structures of hotspots for polymerase
`slippage including (CA)n and (A)n tracts in different intermolecular con-
`texts reveals that, in the B-form, these sequences share common structural
`alterations which may explain the high rate of replication errors. In par-
`ticular, a two-faced ‘‘Janus-like’’ structure with shifted base-pairs in the
`major groove but an apparent normal geometry in the minor groove con-
`stitutes a molecular decoy specifically suitable to mislead the poly-
`the rat polymerase b bound to this structure
`merases. A model of
`suggests that an altered conformation of the nascent template-primer
`duplex can interfere with correct nucleotide incorporation by affecting
`the geometry of the active site and breaking the rules of base-pairing,
`while at the same time escaping enzymatic mechanisms of error discrimi-
`nation which scan for the correct geometry of the minor groove.
`In contrast, by showing that
`the A-form greatly attenuates the
`sequence-dependent structural alterations in hotspots, this study suggests
`that the A-conformation of the nascent template-primer duplex at the
`vicinity of the polymerase active site will contribute to fidelity. The
`A-form may play the role of a structural buffer which preserves the cor-
`rect geometry of the active site for all sequences. The detailed comparison
`of the conformation of the nascent template-primer duplex in the avail-
`able crystal structures of DNA polymerase-DNA complexes shows that
`polymerase b, the least accurate enzyme,
`is unique in binding to a
`B-DNA duplex even close to its active site. This model leads to several
`predictions which are discussed in the light of published experimental
`data.
`
`# 1999 Academic Press
`
`Keywords: unusual DNA structure; spontaneous mutagenesis;
`microsatellite instability
`
`Introduction
`
`The accuracy of DNA replication is fundamental
`for the genetic stability of the cell. From bacteria to
`higher eukaryotes, error frequencies are remark-
`(cid:255)9 and10 (cid:255)10 perbaserepli-
`ablylow,between10
`cated(Echols&Goodman,1991).Theselow
`mutation rates are achieved by multiple steps of
`error discrimination including base selection by
`DNApolymerases,3 0-50 exonucleolyticproofread-
`ingandpost-replicativeDNArepair(Kunkel,
`1992).AlthoughtheDNAsynthesiserrorsplaya
`
`E-mail address of the corresponding author:
`timsit@ibpc.fr
`
`role in aging and disease, spontaneous mutations
`also provide the opportunity for genetic variation
`and are a primary basis for the evolution. Replica-
`tion errors occur non-randomly and result from an
`intricate interplay of intrinsic polymerase fidelity
`and DNA sequence effects.
`A principal determinant of polymerase fidelity is
`the geometric selection of nucleotide for insertion
`intoDNA(Echols&Goodman,1991;Goodman,
`1997).Crystallographicstudiesofoligonucleotide
`duplexes containing mismatched base-pairs have
`shown that
`the geometric equivalence and the
`pseudo 2-fold symmetry of a correct Watson-Crick
`base-pair is violated in mismatched bases, thus
`providing a geometrical means for selection by the
`
`
`
`836
`
`DNA Structure and Polymerase Fidelity
`
`enzymes(Kennard,1987).Indeed,recentstructural
`studies on DNA polymerases have shown that the
`accurate discrimination of a correct base-pair is
`achieved by the tight
`steric
`complementarity
`between a Watson-Crick base-pair and the poly-
`merase active site. Hydrogen-bonding interactions
`whichprobethecorrectgeometryofthebase-pairs
`within the minor groove of the nascent template-
`primer duplex also participate in error discrimi-
`nation(Kunkel&Wilson,1998;Brautingam&
`Steitz,1998;Beard&Wilson,1998).
`The role of DNA sequence and structure in
`polymerase fidelity is less well understood. The
`analysis of the DNA sequence surrounding the
`mutational hotspots has led to several models of
`sequence-directedmutagenesis(Ripley,1990;
`Kunkel,1990).Primer-templatemisalignmenthas
`been proposed to explain the high rates of spon-
`taneous frameshift or substitution mutations in
`homopolymericruns(Streisingeretal.,1966;
`Kunkel,1990).Somestudieshavesuggestedthat
`alterationsofbasestacking(Petruska&Goodman,
`1995;Mendemanetal.,1989),unusualbackbone
`flexibility(Blakeetal.,1992;Mitraetal.,1993;El
`Antrietal.,1993;Hessetal.,1994)oralternative
`DNAstructures(Freundetal.,1989)canaffectthe
`accuracy of replication at hotspots. However, little
`is known about the contribution of the sequence-
`dependent structural variability of DNA in the
`initiation of replication errors, and the molecular
`mechanisms by which DNA structure affect the
`accuracy of DNA replication are still unknown.
`Several years ago a crystallographic
`study
`suggested that the shift in base-pairing observed in
`the major groove of (CA)n tracts could participate
`in the initiation of frameshift errors in hotspots
`(Timsitetal.,1989,1991).Indeed,inbothprokaryo-
`ticandeukaryoticcells,veryhighratesofspon-
`taneousframeshiftmutationsat(CA)
`n tracts were
`found to
`result
`from polymerase
`slippage
`(Levinson&Gutman,1987;Strandetal.,1993).The
`high rate of slippage at CA repeats which is
`revealed in cells deficient in post-replicative mis-
`match repair
`is
`responsible for microsatellite
`instability and is associated with cancer and other
`humandiseases(reviewedbyKunkel,1993;Loeb,
`1994;Karran,1996).Abetterunderstandingofthe
`molecular basis of sequence-directed mutagenesis
`requires an answer to the following questions:
`(1) are there predictible structural features of DNA
`which could account
`for higher error
`rates?
`(2) What are the molecular mechanisms which
`relate DNA structure to the polymerase fidelity?
`Here, the analysis of the crystal structures of
`mutational hotspots including (CA)n and (A)n
`tracts (with n 5 2) in different intermolecular con-
`texts reveals that,
`in the B-form DNA,
`these
`sequences share common structural alterations in
`base-pairing and stacking which may explain the
`high rate of replication errors.
`In contrast,
`the
`A-form greatly attenuates
`sequence-dependent
`structural alterations. A molecular mechanism
`providing a structural basis for sequence-directed
`
`mutagenesis is proposed in the light of the recent
`structural
`studies on DNA polymerases. The
`modelinvolvespolymeraseb(Pelletieretal.,1994)
`bound to a misaligned double helix and suggests
`that ‘‘Janus-like’’ structural features of the nascent
`template-primer duplex can mislead the nucleotide
`incorporation while escaping from the error dis-
`crimination mechanisms used by the polymerases.
`The model also suggests that the A-conformation
`of the duplex observed at the vicinity of the active
`site in many polymerases contributes to polymer-
`ase fidelity by attenuating the sequence-dependent
`alterations. A detailed comparison of the confor-
`mation of the nascent template-primer duplex in
`the available crystal structures of DNA polymer-
`ase-DNA complexes indeed shows that polymerase
`b, the least accurate polymerase, is unique in bind-
`ing to a B-DNA duplex.
`
`Results and Discussion
`
`Common features in the B-DNA helical
`structure of frameshift mutational hotspots:
`comparison of (CA)n and (A)n tracts in different
`intermolecular contexts
`
`Crystallographic, NMR and biochemical studies
`have revealed that CA steps, CAC triplets and
`(CA)n
`tracts
`exhibit unusual
`structural
`and
`dynamic features such as shifted base-pairing,
`kinking and a higher rate of base-pair opening
`(reviewedbyTimsit&Moras,1996,andreferences
`cited therein). For example, B-DNA helices contain-
`ingCArepeatsobservedinthecrystalstructures
`of DNA duplexes or protein-DNA complexes exhi-
`bitanirregulargeometrywithmarkedalterations
`in base stacking, base-pairing and in the sugar-
`phosphatebackboneconformation(Figure1(a)and
`(b)).Inthetetdodecamer(Timsitetal.,1991),the
`bases are not paired with their Watson-Crick
`complements but with their direct 50 neighbors, on
`theoppositestrand(Figure1(a)).Thebasesforma
`set of consecutive A(cid:1)G and C(cid:1)T mispairs on the
`major groove side, while the base-pairing remains
`unaltered in the minor groove. A magnesium cat-
`ionstabilizestheunfavorableapproachofthe
`shifted mismatch between the G7 and the G17
`bases by bridging their two O6 carbonyl groups
`(seeFigure2byTimsit&Moras,1996).Thisgeo-
`metry is achieved by a combination of high propel-
`ler
`twists, opening angles,
`stagger and rise
`(Table1).Theunusualhighvaluesofthestaggerat
`each step indicate that one strand has moved in
`the 50-30 direction relative to the other. The duplex
`may be therefore considered as a pre-slipped
`doublehelix.TheC9-A10andT15-G16stepsare
`characterized by a severe distorsion of the sugar-
`phosphate backbone, including an unusualy close
`approach of consecutive phosphate groups, with
`phosphorus-phosphorusdistancesof6.2and5.7A ˚ ,
`respectively.Figure1(b)showsthattheCArepeats
`exhibit a similar irregular conformation in the
`CAP-DNAco-crystalstructure(CAPS1)(Schultz
`
`
`
`DNA Structure and Polymerase Fidelity
`
`837
`
`Table 1. Structural features of (CA)n and (A)n tracts in different molecular contexts
`
`tet
`(CA)2
`
`CAPS1a
`(chain D)
`(CA)2C
`
`CAPB3b
`(A)5
`
`RAP Ib
`(ch. C, D)
`(CA)3C
`
`Atetr
`(AC)2
`
`C9-G16
`A10-T15
`C11-G14
`A12-T13
`C5-G27
`A6-T26
`C7-T25
`A8-T24
`C9-G23
`T20-A20
`T21-A21
`T22-A22
`T23-A23
`T24-A24
`
`C10-G30
`A11-T29
`C12-G28
`A13-T27
`C14-G26
`A15-T25
`C16-G24
`A5-T4
`C6-G3
`A7-T2
`C8-G1
`
`Stagger (A˚ )
`
`1.8
`1.9
`2.0
`1.3
`(cid:255)1.0
`(cid:255)0.3
`(cid:255)0.3
`0.7
`0.5
`(cid:255)0.2
`0.0
`(cid:255)1.9
`0.1
`(cid:255)0.5
`
`Buckle (deg.)
`(cid:255)6.9
`(cid:255)2.7
`(cid:255)13.6
`(cid:255)10.5
`(cid:255)18.0
`(cid:255)10.3
`(cid:255)7.7
`(cid:255)32.2
`(cid:255)5.6
`(cid:255)32.0
`(cid:255)35.3
`(cid:255)14.0
`(cid:255)9.3
`6.5
`
`7.2
`(cid:255)1.3
`(cid:255)1.9
`(cid:255)11.3
`5.2
`(cid:255)10.3
`7.6
`3.1
`5.9
`2.5
`11.1
`
`Propeller twist
`(deg.)
`(cid:255)27.8
`(cid:255)21.0
`(cid:255)25.5
`(cid:255)10.7
`(cid:255)32.0
`(cid:255)28.2
`(cid:255)16.4
`(cid:255)24.2
`(cid:255)21.0
`(cid:255)40.0
`(cid:255)48.7
`(cid:255)47.8
`4.1
`(cid:255)18.4
`(cid:255)3.6
`0.0
`(cid:255)23.2
`0.1
`(cid:255)3.4
`(cid:255)0.2
`(cid:255)13.9
`(cid:255)0.1
`(cid:255)10.1
`(cid:255)0.2
`(cid:255)5.0
`(cid:255)0.74
`(cid:255)3.2
`0.7
`(cid:255)14.4
`0.4
`(cid:255)17.7
`0.0
`(cid:255)13.6
`0.1
`(cid:255)13.0
`(cid:255)0.2
`ThehelicalparameterswerecalculatedwithCurves(Lavery&Sklenar,1989)ontheDNAcoordinatesofthe(CA)
`dodecamerd(ACCGGCGCCACA)(tet)(Timsitetal.,1991),theCAP-DNAcomplex(CAPS1)(Schultzetal.,1991),theRAP1-DNA
`complex(RAP)(Ko ¨ nigetal.,1996),the(AC)
`2 tractofthetetragonalformoftheoctamerd(GTGTACAC)(Atetr)(Jainetal.,1989)and
`the (A)5 tractoftheCAP-DNAcomplex(CAPB3b)(Parkinsonetal.,1996b)whicharedisplayedinFigure1.
`
`Opening (deg.)
`
`Rise (A˚ )
`
`Slide (A˚ )
`
`20.8
`30.9
`(cid:255)0.1
`10.5
`16.6
`1.0
`2.6
`5.8
`1.2
`4.4
`(cid:255)5.2
`5.8
`(cid:255)17.0
`(cid:255)9.6
`
`1.7
`1.3
`4.4
`9.7
`(cid:255)1.6
`8.7
`(cid:255)1.4
`0.7
`4.4
`6.4
`1.9
`
`3.0
`3.9
`3.0
`
`4.3
`3.4
`2.9
`4.3
`
`3.0
`2.7
`3.5
`2.5
`
`3.6
`3.4
`3.1
`3.5
`4.0
`3.2
`
`3.0
`2.9
`2.7
`
`1.3
`(cid:255)1.2
`(cid:255)0.8
`
`1.5
`(cid:255)0.2
`0.5
`(cid:255)1.7
`(cid:255)0.9
`(cid:255)0.4
`(cid:255)1.6
`(cid:255)0.9
`
`0.9
`(cid:255)0.4
`0.9
`(cid:255)1.0
`(cid:255)1.3
`(cid:255)0.8
`(cid:255)1.5
`(cid:255)1.3
`(cid:255)1.8
`
`n tracts of the
`
`-DNAduplexisagaincharacter-
`etal.,1991).TheB
`ized by a pronounced unstacking of the bases with
`very high buckle, rise and propeller twist angles
`(Tables1and2)leadingtotheformationofshifted
`base-pairs and bifurcated hydrogen bonds between
`crossed Watson-Crick donor and acceptor groups
`in the major groove. Similar to the tet structure,
`0 neighborsoftheir
`thebasesinteractwiththe5
`complements(Figure2(b))buthere,thealterations
`do not propagate along the whole helix.
`(CA)n tracts and (A)n tracts share common struc-
`tural features. In most of DNA crystal structures,
`the (A)n tracts adopt a unique geometry in which
`the consecutive adenine bases form an array of
`bifurcated hydrogen bonds with the 50-neighbors
`oftheircomplements(Colletal.,1987;Nelsonetal.,
`1987;Aymamietal.,1989).ThecrossedWatson-
`Crick interactions are produced by very high nega-
`tive propeller twist angles, negative inclination and
`unusualsugarpuckers(DiGabriele,1993;Brahms
`etal.,1992).Forexample,thealterationsfoundin
`the (A)n tractsofthecrystalstructureofthephage
`434repressor-DNAcomplex(Aggarwaletal.,1988)
`ortheCAP-DNAcomplex(CAPB3)(Figure1(c)
`andTable1)(Parkinsonetal.,1996b)arereminis-
`cent of that of tet. This analysis shows that, within
`the B-DNA family, the DNA sequences in which
`one strand is Watson-Crick donor (N6 or N4
`amino group) and the other one is Watson-Crick
`acceptor (O6 or O4 carbonyl group) viewing the
`
`major groove can form altered DNA structures,
`with very high propeller twist and bifurcated or
`shifted base-pairs in the 50-30 direction.
`
`A-DNA attenuates sequence-dependent
`structural variations
`
`4
`
`Figure1(d)showsthatinthecrystalstructure
`the RAP1-DNA complex (RAP),
`the (CA)n
`of
`¨ nig
`tractsadoptamoreregularstructure(Ko
`etal.,1996).Thebase-pairsarestackedinparal-
`lelwithalowpropellertwist(Table1).This
`findingissomewhatsurprising,sinceKMnO
`reactivitystudiesintheRAP1-recognition
`sequence anticipated aberant base stacking and
`pairinguponproteinbinding(Gilsonetal.,
`1993).Thealternanceofhighandlowindividual
`twists at CA and AC steps, respectively (see
`¨ nigetal.,1996),seemstocontrib-
`Table1byKo
`utetomaintaintheplanarityofthebase-pairsin
`building a base stacking pattern which disfavors
`propeller
`twisting. Similarly,
`the high twists
`(50 (cid:14)) at the CA/TG steps in monoclinic decamer
`duplexes have been associated with a low pro-
`´ etal.,1991).How-
`pellertwistandrise(Prive
`ever,
`another parameter
`contributing to the
`uniformity of
`the base stacking pattern is the
`A-character of the double helix, as indicated by
`the high negative slide (local) and X-displace-
`mentobservedinmanysteps(Tables1and2).
`
`
`
`838
`
`DNA Structure and Polymerase Fidelity
`
`Figure 1 (legend opposite)
`
`
`
`DNA Structure and Polymerase Fidelity
`
`839
`
`Figure 1. Stereo views in the major groove of (CA)n and (A)n tracts observed in different molecular contexts. The
`shifted or bifurcated hydrogen bonds are represented in thick broken lines. (a) The (CA)2 repeat of the tet dodecamer
`2C repeat of the chain D of the B-DNA oper-
`d(ACCGGCGCCACA)inB-form(tet)(Timsitetal.,1991).(b)The(CA)
`atorintheCAP-DNAcomplex(CAPS1)(Shultzetal.,1991).(c)The(A)
`5 repeat of the B-DNA duplex of the CAP-
`3C repeat of the chain CD of the RAP1-DNA complex
`DNAcomplex(CAPB3b)(Parkinsonetal.,1996b).(d)The(CA)
`¨ ningetal.,1996).(e)The(AC)
`n repeat of A-DNA octamer
`(B-DNAwithstructuralfeaturesoftheA-form)(RAP)(Ko
`d(GTGTACAC)inthetetragonalspacegroup(Atetr)(Jainetal.,1989).Seealsothecorrespondinghelicalparameters
`reportedinTable1.
`
`Indeed,Figure1(d)showsthatthelateralshift
`between C14(cid:1)G26 and A15(cid:1)T25 moves the N4
`amino group of
`the cytosine base C14 away
`from the O4 carbonyl group of the thymine T25,
`at a distance too long for forming a bifurcated
`hydrogen bond. The most regular conformations
`of (CA)n tracts are observed in the crystal struc-
`turesofA-DNAduplexes(Jainetal.,1989,1991).
`The base-pairs are well
`stacked and display
`muchlessvariabilityintheinter-basehelical
`parametersthanintheB
`-form(Figure1(e)and
`n and (AC)n
`Table1).Ananalysisofallthe(CA)
`tracts(withn52)foundinthecrystalstruc-
`tures of DNA duplexes or protein-DNA com-
`plexes deposited in the Nucleic Acid Database
`(Bermanetal.,1992)showsthat,ingeneral,the
`
`the base-pair planarity diminishes
`alteration of
`with an increase of the A-character of the double
`helix.Table2andFigure2indicatethatthe
`lowering of the average values of the rise, the
`propeller twist and the buckle of the tract corre-
`lates with an increase of the A-character of the
`double helix. It should be noted however, that
`the (AC)2 repeat of the crystal structure of the
`MATa2 homeodomain-operator complex (MatW)
`(Wolbergeretal.,1991)displaysahighabsolute
`value of propeller twist despite the low X-disp
`andslide(Table2).Thisexceptioncanbe
`explained by the presence of an unusual positive
`stagger which makes possible the formation of
`bifurcated hydrogen bonds.
`
`
`
`840
`
`DNA Structure and Polymerase Fidelity
`
`Table 2. Structural alterations and A-character of (CA)n and (AC)n tracts
`
`NDB id
`
`RCSB009134
`ADH034
`ADH014
`
`Res
`(A˚ )
`
`2.6
`2.0
`2.0
`
`A. DNA
`tet
`Ahex
`A tetr
`
`Seq
`
`A char
`
`X-disp
`(A˚ )
`
`Slide
`(local) (A˚ )
`
`Incl
`(deg.)
`
`Stag
`(A˚ )
`
`Rise
`(A˚ )
`
`jBuckj
`(deg.)
`
`jPropj
`(deg.)
`
`(CA)2
`(AC)2
`(AC)2
`
`0.12
`1.03
`1.23
`
`(cid:255)0.1
`(cid:255)3.0
`(cid:255)3.9
`
`(cid:255)0.2
`(cid:255)1.5
`(cid:255)1.7
`
`(cid:255)1.6
`7.6
`9.9
`
`1.8
`(cid:255)0.2
`0.1
`
`3.27
`3.20
`2.9
`
`9.7
`11.2
`5.7
`
`22.4
`14.7
`14.7
`
`PDR006
`
`PDRCO3
`PDR023
`
`PDR025
`
`PDRCO3
`PDT035
`
`PDR036
`PDT049
`PDE0115
`
`PDT045
`
`2.7
`
`3.0
`
`2.4
`2.5
`
`2.7
`
`2.4
`2.25
`
`2.25
`2.2
`3.0
`
`2.5
`
`5.7
`10.4
`16.6
`25.0
`17.9
`11.8
`11.8
`4.9
`6.8
`14.0
`8.1
`9.9
`8.8
`11.2
`5.5
`9.5
`11.3
`12.9
`6.7
`6.5
`23.2
`4.5
`8.7
`
`B. Protein-DNA
`(cid:255)0.1
`(cid:255)0.5
`0.11
`3.6
`0.3
`4.1
`14.1
`(CA)2
`CAPB2a
`PDR024
`(cid:255)0.1
`(cid:255)0.5
`0.11
`3.6
`0.2
`3.7
`15.9
`(CA)2
`CAPB2b
`(cid:255)0.2
`(cid:255)0.3
`(cid:255)1.0
`0.15
`(CA)2C
`0.1
`2.7
`9.1
`CAPS1a
`(cid:255)0.2
`(cid:255)0.3
`(cid:255)1.0
`(cid:255)0.1
`0.15
`(CA)2C
`3.7
`14.2
`CAPS1b
`(cid:255)0.2
`(cid:255)0.4
`(cid:255)0.2
`0.15
`1.4
`3.4
`3.8
`(CA)2
`ESTR II
`(cid:255)0.1
`(cid:255)0.7
`(cid:255)0.1
`0.15
`(CA)2C
`4.9
`4.1
`20.8
`CAPB1a
`(cid:255)0.1
`(cid:255)0.7
`(cid:255)0.2
`0.15
`(CA)2C
`4.9
`4.0
`16.3
`CAPB1b
`(cid:255)0.1
`(cid:255)0.9
`0.19
`6.2
`0.7
`4.2
`5.3
`(CA)2
`CAPB3a
`(cid:255)0.1
`(cid:255)0.9
`(cid:255)0.4
`0.19
`6.2
`3.7
`8.8
`(CA)2
`CAPB3b
`(cid:255)0.3
`(cid:255)0.3
`(cid:255)0.2
`0.20
`0.2
`3.4
`3.0
`(CA)2
`ESTR I
`(cid:255)0.5
`(cid:255)0.4
`(cid:255)2.2
`(cid:255)0.2
`0.27
`(CA)2C
`3.4
`4.8
`RAP Ia
`(cid:255)0.5
`(cid:255)0.4
`(cid:255)2.2
`(cid:255)0.2
`0.27
`3.6
`6.3
`(CA)4
`RAP Ib
`(cid:255)0.5
`(cid:255)0.5
`(cid:255)1.5
`(cid:255)0.2
`0.29
`(CA)2C
`3.4
`6.1
`RAP Ia
`(cid:255)0.5
`(cid:255)0.5
`(cid:255)1.5
`(cid:255)0.2
`0.29
`3.5
`8.2
`(CA)4
`RAP Ib
`(cid:255)0.5
`(cid:255)1.0
`(cid:255)0.5
`0.33
`3.0
`3.0
`9.9
`(AC)2
`MAT1
`(cid:255)0.6
`(cid:255)0.7
`(cid:255)1.8
`(cid:255)0.6
`0.35
`3.3
`14.9
`(CA)2
`CAPS2
`(cid:255)0.5
`(cid:255)1.5
`(cid:255)0.5
`(AC)2
`RESOa
`0.37
`9.4
`3.1
`8.3
`(cid:255)0.5
`(cid:255)1.5
`(cid:255)0.2
`0.37
`9.4
`3.4
`5.1
`(AC)2
`RESOb
`(cid:255)0.6
`(cid:255)1.1
`(cid:255)0.2
`0.40
`(CA)2C
`1.2
`3.3
`7.7
`BRACHa
`(cid:255)0.6
`(cid:255)1.1
`(cid:255)0.2
`0.40
`(CA)2C
`1.2
`3.3
`6.9
`BRACHb
`(cid:255)0.7
`(cid:255)1.1
`0.42
`1.3
`0.4
`3.6
`4.4
`(AC)2
`MAT2
`PDT005
`2.7
`(cid:255)0.7
`(cid:255)1.5
`0.51
`2.9
`0.0
`3.3
`4.2
`(AC)2
`GABP
`PDT048
`2.15
`(cid:255)0.7
`(cid:255)1.7
`0.52
`(CA)2C
`6.6
`0.1
`3.4
`5.7
`PHO4
`PDT046
`2.8
`The crystal structures containing either (CA)n or(AC) n tracts(withn52)aretakenfromtheNucleicAcidDatabase(Bermanetal.,
`1992).TheaveragevaluesoftheX-disp,localslideandinclinationarecalculatedonthewholedouble-helix.Thestagger,therise,
`the absolute values of the buckle and propeller twist are calculated on the tract. The A-character which takes into account both
`2 (cid:135)(slide loc/1.8)2)1/2. The helical parameters were calculated with
`X-dispandslideiscalculatedasfollowed:A-char(cid:136)((X-disp/5)
`CurvesontheDNAcoordinatesof(tet)(Timsitetal.,1991),thehexagonal(Ahex)(Jainetal.,1989)andthetetragonal(Atetr)(Jain
`etal.,1991)formoftheoctamerd(GTGTACAC),theCAP-DNAcomplexes(CAPB1)(Parkinsonetal.,1996a),(CAPB2andCAPB3)
`(Parkinsonetal.,1996b),(CAPS1andCAPS2)(Schultzetal.,1991),theestrogenreceptorDNAbindingdomain-DNAcomplex
`¨ nigetal.,1996),theMATa2/MCM1/DNAternarycomplex
`(ESTR)(Schwabeetal.,1993),theRAP1-DNAcomplex(RAP)(Ko
`(MAT1)(Tan&Richmond,1998),thegdresolvase/DNAcomplex(RESO)(Yang&Steitz,1995),theTdomain-DNAcomplexofthe
`¨ ller&Herrmann,1997),theMATa2homeodomain-operatorcomplex(MAT2)
`Brachyurytranscriptionfactor(BRACH)(Mu
`(Wolbergeretal.,1991),theGABPa/
`b-DNAcomplex(GABP)(Batcheloretal.,1998)andthePHO4bHLHdomain-DNAcomplex
`(PHO4)(Shimizuetal.,1997).ThelettersIandIIcorrespondtothefirstandthesecondduplexoftheasymmetricunit;thelettersa
`and b correspond to the first and second tract of the duplex.
`
`the A-conformation
`This analysis shows that
`imposes geometric constraints which attenuate the
`large distorsions of (CA)n tracts which can occur in
`B-DNA and constitutes a geometric obstacle for
`strandslippage.First,inA
`-DNA(Figure1(e)),the
`interstrand stacking produces two well interdigi-
`tated helices which makes the slippage of one
`strandrelativetotheotheronemoredifficult
`thaninB-DNA(Figure1(a)).Second,thepositive
`inclination, the negative slide and the low twist
`preclude the formation of the shift in base-pairing
`and the alteration of the base-pair planarity in
`increasing the distances between crossed Watson-
`Crick groups which become too distant for stabiliz-
`ing a network of shifted or bifurcated hydrogen
`bondsinthemajorgroove(Figure1(d)and(e))
`(Gaoetal.,1991).Thethreecenteredhydrogen
`bonds can only persist between purine bases. A
`third important structural feature for preventing
`the shift in base-pairing is a negative sign of the
`
`stagger. However, negative staggers are not strictly
`correlated with the A-form.
`the A-form against
`The buffering power of
`sequence-dependent
`structural
`variations was
`noticed a long time ago. Early fiber diffraction
`experiments indicated that ‘‘the details of A-con-
`formation is insensitive to base composition and
`sequence’’(Leslieetal.,1980).Earlycrystallo-
`graphic studies also noticed that in A-DNA, the
`minor groove width remains relatively invariant
`while the backbone geometry and the base stack-
`ing is much more
`regular
`than in B-DNA
`(Shakkedetal.,1983;Conneretal.,1984;
`reviewedbyWahl&Sundaralingam,1997).
`Recent analyses of a large sample of DNA crys-
`tal structures of various sequences and sizes has
`confirmed that the geometry of A-DNA duplexes
`ismuchlessvariablethantheB
`-form(ElHassan
`&Calladine,1996;1997;Suzukietal.,1997).
`
`
`
`DNA Structure and Polymerase Fidelity
`
`841
`
`Figure2.Scatterplotsofaverageof(a)propeller-twistandbuckle(absolutevalues)and(b)risein(CA)
`tractsofDNA-proteincomplexesoftheNDBversusX-displacement.Theaverageoftheabsolutevaluesofpropeller
`twist,buckleandrisearecalculatedoneachtractwhiletheX-dispiscalculatedonthewholedoublehelix(Table2).
`
`n or (AC)n
`
`A model of DNA structure-directed
`replication errors
`
`The nascent DNA duplex: a structuring element of
`the polymerase active site
`
`The hypothesis presented here suggests that an
`altered structure of the nascent template-primer
`duplex in the polymerase binding cleft interferes
`with correct nucleotide incorporation by affecting
`the geometry of the polymerase active site. Recent
`crystallographic studies on ternary polymerase-
`
`DNA-dNTP complexes have shown that the steric
`complementarity between the polymerase active
`site and a newly formed Watson-Crick base-pair
`constitutes a critical step in the fidelity of nucleo-
`tideincorporation(Pelletieretal.,1996;Brautingam
`&Steitz,1998;Kunkel&Wilson,1998;Beard&
`Wilson,1998).Itisimportanttonotethatthepro-
`duct of the enzymatic reaction, the nascent tem-
`plate-primer DNA duplex, participates with the
`protein in forming the contour of the active site. In
`general terms, the active site is delimited on one
`side by the protein residues that interact and stack
`
`
`
`842
`
`DNA Structure and Polymerase Fidelity
`
`with the newly formed base-pair, and, on the other
`side, by the first base-pair of the template-primer
`duplex. The co-planarity of this base-pair is there-
`fore critical
`for the correct nucleotide incorpor-
`ation. For example, in the polymerase b active site,
`the newly formed base-pair consisting of the tem-
`plate guanine paired with an incoming ddCTP is
`tightly sandwiched between the a helix N of the
`enzyme on the top, and the first base-pair of the
`duplexonthebottom(Pelletieretal.,1994).Itis
`therefore conceivable that if structural alterations
`occur in a newly replicated (CA)n or (A)n tract,
`they will affect the geometry of the polymerase
`active site and lower the accuracy of nucleotide
`incorporation.
`
`The conformation of the template-primer duplexes
`at the vicinity of the polymerase active site
`
`TheconformationofthenascentDNAduplexes
`observed in the DNA-polymerase-DNA complexes
`HIV-1reversetranscriptase(HIV1-RT1)(Huang
`etal.,1998)and(HIV-RT2)(Dingetal.,1997,1998),
`BacillusstearothermophiluspolymeraseIlargefrag-
`ment(B.stear)(Kieferetal.,1998),theclosedform
`of the Thermus aquaticus DNA polymerase I large
`fragment(KlenTaq)(Lietal.,1998),bacteriophage
`´ etal.,1998),
`T7DNA-polymerase(T7)(Doublie
`human polymerase b (polb-gap1, polb-gap2 and
`nick)(Sawayaetal.,1997)andratDNApolymer-
`aseb-DNA-ddCTP(polb)(Pelletieretal.,1994)
`whose coordinates have been deposited in the
`NucleicAcidsDataBase(Bermanetal.,1992)are
`comparedinTable3.Thehighnegativevaluesof
`X-displacement(X-disp),thepositiveinclinationas
`well as the RMS values with the corresponding
`canonical DNA forms indicate that
`the DNA
`duplexes bound to HIV1-RT1, HIV1-RT2, B. stear,
`KlenTaq and T7 belong to the A-DNA family
`(Table3andFigure3(b)).However,theirconfor-
`mations resemble the extended A-DNA double
`helix found in the recent hexagonal structure of the
`methylated decamer d(CCGCCGGCGG)
`(A-ext)
`(Mayer-Jungetal.,1998)ratherthancanonical
`A-DNA. The extended A-form has a higher twist,
`riseandX-displacement(Table3).Inthese
`duplexes, the minor groove widening at the 30 end
`of the primer strand indicates that the A-character
`gradually increases upon approaching the poly-
`meraseactivesite(Figure3(a)).Intheduplexes
`boundtothepolymeraseb,theminorgroove
`widening is less pronounced. The simultaneous
`decrease of X-disp also reflects this tendency in the
`DNAboundtoB.stearandKlenTaq(Figure3(b)).
`These recurrent
`structural
`features which are
`sequence-independent are guided by the protein-
`DNAinteractionswithinthepolymeraseDNA
`binding cleft.
`The DNA conformation in the ternary complex of
`the rat DNA polymerase b-DNA-ddCTP (polb)
`(Pelletieretal.,1994)isunique,sinceitbelongsto
`the B-DNA family, even close to the polymerase
`active site. This is confirmed by the positive value
`
`of X-disp, the negative inclination and the narrower
`minor groove as well as the RMS deviation from
`thethreecanonicalDNAforms(Table3and
`Figure3(a)and(b)).TheRMSvaluesalsoindicate
`that in the other polymerase b-DNA complexes, the
`duplexes are more close to a B-DNA double helix
`thantoaA-DNAdoublehelix,despitethenegative
`values of X-disp. Another important
`feature of
`theseduplexesistheirpositivestagger(Table3).As
`noted above, the rat polb DNA duplex is more irre-
`gular than the KlenTaq duplex which has a similar
`sequence but which adopts the A-conformation.
`Table3alsoindicatesthattheduplexoftheternary
`complex polb(cid:1) gap(cid:1)ddCTP (polb gap1) displays a
`slight increase of its A-character comparing to that
`of the binary complex (polb gap2). A similar effect
`is also observed in the HIV-1 reverse transcriptase
`complexes.
`
`A model of the polymerase b bound to a
`mutational hot spot: a molecular decoy
`for polymerases
`
`Figure4(a)and(b)displaytheternarycom-
`plex of the rat polymerase b, ddCTP and DNA
`asfoundintheco-crystalstructure(Pelletier
`etal.,1994).Inthenormalsituation,theddCTP
`is incorporated in front of the guanine template
`base
`respecting
`base-pair
`complementarity.
`Figure4(c)and(d)representamolecularcom-
`plex in which the correct template-primer duplex
`is replaced by a misaligned double helix. The
`model has been constructed by fitting the crys-
`tallographic coordinates of
`the tet dodecamer
`(Figure1(a))onthenascentduplexoftherat-
`polymerase b. A remarkable
`this
`feature of
`the
`duplex is that
`the molecular structure at
`(CA)n tract is normal when seen from the minor
`groove
`side
`and mismatched in the major
`groove. This two-faced structure constitutes a
`molecular decoy specifically suitable for inducing
`replication errors. Viewed from the minor groove
`side,
`the double helix exhibits an apparently
`unaltered conformation which can escape the
`correction mechanisms used by the enzymes.
`This is particularly relevant given that, in gener-
`al, DNA polymerases scan newly made replica-
`tion errors in sensing the correct geometry of
`the base-pairs in the minor groove of the nas-
`centduplex(Bebeneketal.,1997;Kunkel&
`Wilson,1998;Brautingam&Steitz,1998).Incon-
`trast, a view down the major groove shows that
`in breaking the code of base complementarity in
`the template-primer duplex, the slippage of the
`base-pairing disrupts the geometry of the active
`site(Figure4(c)and(d)).Theshiftinbase-pair-
`ing propagates towards the site of nucleotide
`incorporation in such a manner that the thymine
`template base which should be paired with an
`incomingdATP(Figure5(b))interactswiththe
`0 primerterminusandformsaT(cid:1)C
`baseofthe3
`shiftedmismatch(Figures4(c),(d)and5(a)).
`Consequently, the incoming dNTP will be paired
`
`
`
`Table 3. Conformation of the nascent template-primer duplex in the co-crystal structures of DNA-polymerase-DNA complexes
`
`HIV-RT1
`(cid:135)dTTP
`GCGCCGG
`CGCGGCC
`
`HIV-RT2
`-
`GGCGCCA
`CCGCGGT
`
`B.stear
`(cid:135)ddTTP
`ATGATGC
`TACTACG
`
`KlenTaq
`(cid:135)ddCTP
`CGGCGCC
`GCCGCGG
`
`T7
`(cid