Tobacco Documents Online

Cell Bio!o~y P,~vir~'s (1991) 2~(2): Cell Biology Reviews University of the Ba~quc Country A review of DNA metabolism in Escherichia coli E. 1]alb|nder and C. Waldren D@~'onen~ of BiocAemLq~j, Blol~hysic~ and ffe.nez~s, ~n~,ersity of Colorado Health gdence.¢ Cen:o, Denver, CO 80262. USA I. Introduction The ongoing revolution in molecular biolosy is bringing increasing evidence that mutations are central to a number of pathological processes including cancer, birth defects and ino berked diseases. This understanding and the fac~ that a number of envin~unental asents, physical and chemical, ate capable of causing mutations, cancer and other pathologies has converted the study of mutations and the ~ Ieuding to them into a field of medk:ai interest. The ~mlting resezn:h has improved oor understmding of DNA ami the intricacies of its metabo- lism. We now view DNA as a dyrmmic entity and within this fi-'m~-work have anak~ a better undemanding of how mum- fioos may occm'. It ]ms also becom~ clear that the study of mnta~en .esis cannot be separat~ from that of DNA rcplic- have developed as separate research discipline~, we know today that they overkp si~nificamly and s~-e many enzymat- ic functions so ~ a unified conceptual framework these different aspects of DNA metabolism is needed. We know ~ in proka~otes, enzymes panicipatin~ in the~ IXO- cesses are or~aniznd in batteries of operons which respond more or less coordinately to a variety of external suesses such An openm is a s~t of linked gezze~ transcribed as a unit fi~)m a single regulated promoter. Some of these operons are inte- ~ imo complex reguletory oetwo~ks ami, to a large ex~r~ mata~nesis is an inducible response. One objective of this review is to show how these netwmks overlap and connect DNA metabolism with overall cellular membolism~ Oor curren~ ~ders~wling of DNA metabolism is based in large pan on smdks condncted wi~h dlvidin~ E~Aerich~z coil cells, following the demmmmtion by Lurk and ~elbrock 1943 (~9) that mutations occur in replicating ceils in the absence of a selective envkoonm~t. Recently, attention has been focused o~ mutations taking place in non-dividing cells under s~ss. According to Cairns er a! (~m') and others, such mutations can occur adaptively under starvation conditions in response to a sekcfive a~enL These ideas may usher a new era in the study of DNA metabolism of particular importance to the the etiolo~7 of human genetk: dise~ and cancer. This will be discussed in the last chap~" of this mono~-~ph Chapter). This review has been w~9~m ln~marily for the use of elate students, but should be ~se~ul to ~..arches who are not specialists in the various areas of DNA metabolism and. hopefully, to some specialists as well We will present the major ideas d~vekq~ed with the E. co~i system as a guide to studies in eukaryotes, including yeast and mammalian cells. The subject is large and intricate and cannot be covered in ~-eat detail in a relatively short mono~aph. Many excellent reviews have be~z published in recent years und will be men- tioned where indicated. However. not a single review attempting to integrat~ all &e information about DNA m,taboli~ra into oae large picture exists to date. This in~ry review hopes to fill this gap. 1I. Maintaining the integrity of genetic information during DNA replication The accurate transmission of genetic "information from rrmther to daughter cell is a fundamental requirement of all forms of life. Most heritable mutations (alterations of the ge.netic message) would be deleterious but a few could be beneficial for the evolution of the species. Thus. not all mum- dons need to be avoided. In this chapter we will review the strategies which E. coil has developed to deal with this dual challenge, i. e. maintenance of the integrity of its genome while allowing a certain, small measure of diversity to ensure the survival of the species in a changing environment. A. DNA Replication Maintaining the integrity of the genetic information is an integral part of the replication of the E. coil chromosome. This is an exuemely complex process which requires the pre- cise interaction of a large number of proteins organized into a ,replication machine~,. The essential proteins are listed in tables I and 2. It is beyond the scope of this monograph to discuss DNA repfication in detail. This has been done in sev- eral excellent reviews, some of themquire recent (301, 406, 407, $33). We need here, however, to provide an overview of DNA replication to allow the reader to understand how replic- ation fklelity is accomplished in the pr~ess. E. coil has a circular chromosome consisting of about 4 X 10" base pairs. It repficates bidirectionally (62), without being linearized, from a fixed origin of replication (oriC) at 83. 5 minutes on the £. coil genetic map, to a terminus (terC3 situ- ated almost diametrically opposite from oriC in the region ft~n 30-32 minutes in the map (25, 407, 480). To accomplish this replication some major problems must be solved. The first has to do with the reciprocal polarity of the two strands of the double helix. Each strand of DNA has a chemical polarity defined by the asymmetry of the sugar-phosphate backbone: the two complementary strands of a double helix have opposite polarities, one going in a 5" --~ -3" and its com- plement in a 3" ~ -5" direction. All DNA polymerases extend newly synthesized DNA chains only in a 5" -~ -3' direction. however, so a problem arises in having the same enzyme moving in only one direction along a replication fork and having simultaneously to synthesize the two DNA strands that are oriented in opposite directions. This problem been solved, as illustrated in fig. I by a semi discontinuous replication of the chromosome: one strand is made continu- ously in a ~" -~ -3" direction (leading strand) and the other (lagging straad), which goes in its entirety in a 3" --~ -5" direction, is synthesized discontinuously in short base palm) 5' --> -3" fragments named after their discoverer Okazaki (sen fig. 1). This pre~nts sorae peculiar difficulties and opportunities for mmagenesis which will be discussed later. The second problem in replication is to unwind the dou- ble helix, i. e. separate the two strands in advance of the 40000137

,~ RNA primer ----- DNA single-stranded DNA binding protein replicative DNA peiymerase hetlcaees, pHm|ng appartus, othe~ enz~nes and cellular structures General feanu~ of a replication fmk. Reproduced with pe~- mission from McHenry (406). replication fork. This is the primary mission of certain enzymc~ (helic~es) which form I~art of the replicatim com- plex (see tables I and 3). The third problem has to do with the fact that the chromosome is not simply a clo~ed circle of bases and sugars but is organized into complex domaim which are supercoiled (139). These higher levels of or- ganization need to be undone as replication proceeds, a task which is performed by 8yrases and topoismnes~ses which are also as pan of the t~plicadon complex (see tables I and 3). The chromosomal replic~Jion cycle h:~ been divided into three phases: initiation, elongation and termination (406, 407). Each sta~e has its own rcquirem:nts which are satisfied by different ccml~ents of the repEc~ion machinery, as Fist- ed in tables 1-3. These proteins do not act independently but interact actively with other components of the replication apparatus (5, 147, 406, 407). Many of them also participate in a number of DNA metabolic processes other than replication. For example, DNA polymerase I (DNA Poll) ~d DNA ligase (table l) axe exucmcly importsr~ in DNA repair and recombi- nation; DnaJ and DnaK (table 3) are components of the heat shock response, a group of about seventeen genes induced by heal ethanol and various agents which cause cellular damage (see VII Chapter), Dna helicase II (table 3, refs. 72, 237, 447) plays important supporting roles in mutagenesis and excision -repair (see chapter IV). As will be discussed in the latter chapte~ these proteins constitute an important commuaica- tion link between DNA replication and various cellular retch- Initiation of synthesis is believed to begin with the bind- ing of multiple copies of the l~naA protein to specific recognition sites in oriC, followed by the subsequent delivery of the DnaB helicase to the bacterial origin, a step which is facilitated by DnaC and aidcd in some unknown manner by the bacterial histone-like protein HU (table 3). This is known as the pre-priming complex. The properties and roles of these pmteim axe de~dbed in detail in the review by McMacken et a/(407). Briefly, DnaA is believed to provide the recognition specificity for the pivotal DNA sequences at oriC and to reg- ulate initiation at ~ site; DnaB is a hclicase whose primary role is to unwind duplex DNA in advance of the replication complex at the ~-plication fork; and DnaC is believed to form a stable complex with DnaB and stimulate the general prim- ins ~cdon catalyzed by DnaB and primase whose role will be discussed below (table l). Two other protei~ Dna gyrase and the single-strand binding protein (SSB) play important roles at this stage, as well as other stages, of chromosome replication (table 1). The E. coil chromosome is maintained in a torsionally strained state consisting of approximately 50 topologically independent domains of negatively supercoiled DNA (139). DNA 8yrase, the enzyme and lasmid DNA (106, 185, 189) provides a swivel that allows one strand to unwind around the other, and SSB stabilizes the single stranded DNA Tabl~ l Ksse~titl DNA replicadon proteins Protein F;mctiotl Gene Map Position Mutants with (rain) Mutator Phenol'pc I. I:~A Inidttioa Odgin recognitie~ ~ 83 2. D~a B In~tis/im DNA helicase M 92 3. Dr~C Initiation ~ 99 4. SSB (single su-tnd binding) Stabilize Tr~tsieat ssDNA s~b 92 5. Prirnase RNA In'kner ~yn~z~sis dac~ 67 6. DNA gyrase (a) DNA supercoiling aYrA ~8 7. DNA 8yrase (~) DNA supe~oilla$ qYrB 83 8. DNA Po~I Pdme~ z~noval: gap filling Po/A 87 9. DNA PolllI Proce~ive chain do~gation (see table 2) !0. DNA l[gase S~[iag of DNA nicks EV 52 1 !. l~aT Terminal~m of DNA t~ic~xn c~T 99 X X ~ from McMackenetcl(lg~7) 40000138

CBR 107 Table 2 Pratein Fu.'.ctic,~ Gen¢ (min) Mu:ator PkenoO'Fe x x (ssDNA) as it is formed (see below), inpu~ of energy in the form of ATP is required at both stages. DNA 83;raee contains two polypeptides, alpha and beta (table 1) and exists as a tewamer. It is believed to act by intt~lucing transient double- strand t~aks in the DNA and then passing another segment of duplex DNA through the enzyn~-bridge break, which is then resealech The properties and biological roles of E. coli SSB and other singie-slxand (ss) binding proteins have beea ~-viewed (~/, 414a, 185). SSB is a tetramer of 74,000 d that binds cooperatively to ssDNA maintaining it in the.extended state, i. e. to lXevent renaturadon and formation of inhibitory secondary structures which could occur by slippage of a sin- gie strand. This property of SSB is relevant to some cunent models of mmagonesis, i. e. slipped ndspairing models which. as we shall see in chapter IH requke the existence of single- stranded suetches of DNA during replication. It is, then, not surprising that mmants with aiteradoes in the ssb can exhibit a ,mutator~ phenotype (inc~esse the spontaneous frequency of routation (345, 414a, table 1). The SSB protein is also be- lieved to protect ssDNA from nttcleolydc alxack by ss-specif- ic nuclemes such as ExoI and RecBCD enzyme, which will be discussed in later chapters. In addition, it stabilizes the ,pfimosome,, (the pt~ein complex which initiates swand syn- thesis) at or/C. enhances seven folds the initial ~ of DNA chain polymerization and fivefold the fidelity of DNA poly- mmse m (3 s, 414a). ,Priming* involves the synthesis of a stretch of RNA that DNA polymerase III can then use to start elongation by adding deoxyribonucleoddes to its 3' OH terminal of these RNA pdaners. In E. coli, pdmer synthesis is carded out pri- nuttily by pdmase, the product of the dnaG gone (table !). but primer syn~esis a~ eric also t~lukes SSB and 13~tB heliease (4ff~). The RNA primers are believed to be short 00-60 bp) (ref. above). In elongation, the role of pfimase in the propega- tion of the replication fork is to make the RNA primers need- ed for initiation of O~ ~ts in the synthesis of the The main participant in the, elongation step is DNA poly- merase Ill (301, 406, 4~/, $33). Three different DNA poly- merases have been identified in E. coil Pol I. the product of thepoL4 gme (table 1) was the fu~ to be discovered az¢l was believed to be the main polymerase for DNA replica~iou. The subsequent demonstration by De Lucia and Calms (121) that po/A mutants with no Pol I activity were rifle, prompted ~ search for new polymerases and two were subsequently found: Pol lI and Pol IIL Until recently no role had been found for Pol IL but it now ~ppears fl~t dds enzyme is a com- ponent of the SOS response (45, 8I, 266) a group of about 17 genes which play important roles in DNA re~zir and mutage- nesis (to be discussed in VII Chapter). Although Pol I.plays important roles in DNA replication and other aspects of DNA metabolism, as we shall see later, it is not the replicative DNA polymeric. DNA polyroerase I~ is the replicafive DNA polymerase of E. coli (table 1, refs. 30I, 406, 407, 533). The active form of DNA Poi IlI is an assembly, or holoenzyme, of seven dif- ferent proteins (table 2), all of which must be present for cor- rect functioning of the complex in chain elongation. One of these proteins, alpha, the product of the dnaE gone, carries the 5" --> 3' polymerase activity. The holoenzyme has also been shown to possess 3' -> 5" as well as a 5' ---> 3' exonuclease activities. The 3" -> 5' ¢xonucleese is carded by the epsilon subunit, the product of the dnaQ gone (table 2, refs. 535, $36). It requires ssDNA with 3" OH termini and yields 5" mononuclootides as products. There is strong evidence that rids is a ,proofreading~ exonuclease which eliminates mis- paired bases at the growing point of the DNA chain and thus ensures replication fidelity. We will return to this later in this chapter. No physiological role has thus far been demonswated for the 5' .--> 3' exonuclease of the Pol III holnenzyme. The auxiliary subunits of DNA Pol IlI holoenzyme appear to be needed to improve the processivity, the rate at which replica- don occurs, of the enzyme (406, 407). DNA Pol I (337) contains three enzymatic activities involved in the synthesis or hydrolysis of phosphodiester bonds. One of them is a $" -> 3" polymerase activity that func- tioos by addition of mononucleotide residues from deoxyri- bonucleodde Ifiphosphates (dNTP's) to the 3'OH terminus of a pdmer chain. Tbe .o~her two are exonuclease activities: a 3' --> 5" editing exonuclease that cleaves mispaired bases at the 3'OH tenninus of growing DNA chains, and is thus important for maintaining informational fidelity, and a 5' --~ 3' exonuclease which degrades base paired DNA from a 5" terminus releas- ing oligonucleoddes up to 10 residues long. This activity func- tions in v/re to remove the RNA primers from the 5" terminus of nascent DNA and Okazaki fragments and is coordinated with gap-filling DNA synthesis by the 5" --~ 3" polymerase activity of the enzyme. This coordination of degradation and synthesis is known as <<nick translations, and is needed to re- place the RNA primer with DNA and thus fill in the gaps left between Okazaki fragments on the lagging strand. It is also employed in DNA repair to remove sections of DNA damaged by mutagen~c agents a~d replace them with undamaged new DNA. Thus poLA mutants can exhibit a mutator phenoL~pe (table 1, ref. 159). There is evidence th~ Poi I m~y p~nicipaze in SOS rel~dr/mutagenesis. Condhional lethal mu/~atts of polA have been ~ (273, 300). The role of Pol I in DNA repair will be me~tior.ed again lazer (c ~h~o~er IV). 40000139

103 CBR Proof sos ~ (t~les 6 & 8) Finally, an important enzyme in replicatimt and repair is DNA lig~e, the product of the fig gone (table L refs. 20~, 336a). This e~zyme catalyzes the joining of adjacent 3'OH and 5" phosphoo/l tenalni of nuclentides that a~e H-bonded to a complementary strand. It functions in the elongation phase by joining Okazaki fragments, an important step in com- pleting the synthesis of the lagging strand. It can also seal nicks in DNA. Because it is the only E. colt enzyme that can join adjacent fragments, it is essential in DNA repair and recombination as well as replication. Conditional fig lethal (temperature sensitive) mutants show a rapid loss of cell via- bility at the matrictive ~ as well as an accumulation of nascent DNA fragments (2~, 33~). Table 3 fists sevelal proteins which play auxiliazy roles in DNA replication (4~/) and at the same time constitute links between the chitmmsome replication machinery and vaxious inducible systems which are impo~at in the repah" of DNA damage caused by sue= stimuli (see VlI Chapter). The Rep protein and helicase II, products respeedvely of the rep and uvrD genes (33, 72, 7,37, 4~3 are two heficases found in E. coil in addition to the DnaB helica~ The role of Rep in chro- mosome replication is not clear ~ it my pa~icipate in the replication of some single-stranded bacteriophages (407). It has been suggested (same ref.) that the Rap protein cooper- ares with DNA helicase II to unwind DNA during replication of the bacterial chromosome and that either Rap or helicase II could drive fork movement, but this is considered unlikely since u~rD mutants replicate nmmally (40"/). While not much is known about additional roles for Rep. DNA helicase II participates in mismatch repair as well as excision repair and is part of the SOS response (see legend to table 3). Thus w,,rD may play an important role in repair, perhaps, as suggested by Ossarma and Mount (443) in the processing of DNA damage to generate the signal for SOS induction. We will return to this in VII Chapter. DNA topoisomerase I (10=3, 138, 188, 39~, 407, 445, b"/8, 609) transiently breaks the phosphodieste~ linkage of DNA in one strand and later reseats the ~ and is thus considered capable of catalyzing the passage of ss and double stranded (ds) DNA chains through enzyn-,e-bHdged breaks so as to relax negatively supemoiled DNA. Whether it actually does this in ~o is not known (refs. above) D~J. DnaK and GrpE are part of the heat shock response (see legend to utble 3 ~d VII Ch~y~.r) The fast two ~ppe~r to be ~nvelved in the replic=zion of b~-'~eriophage Im'nbda. Mutants for ~. dnnK, and grpE block bacterial growth and interfere with both RNA and DNA synthesis. GrpE and DnaK am known to interact intimately (407). While it is not yet clear whether they play any role in DNA replication, the evi- deuce suggests that they may connect the DNA replication apparatus to the heat shock response. We will return to this in Vll Chapter. The key event in the initiadon of a replication fork is the beginning of leading su'and synthesis. Once the polyme~zation of the leading strand has stalted~ polymerizalion of the lagging strand follows sho~ly. A complex of DNA replication proteins (the mplisome) moves along the replication fork (fig. 1). The diffemuce in rates of growth between the leading and lagging strand leaves a small soction of ss parental DNA at the replic- ation fork, which is protected by SSB (fig. 1). This becomes double stranded upon the initiation of an Okazaki fragment whose synthesis cuntinaes until it reaches the terminus of the pmvinusly synthesized Okazaki fragment. After removal of the RNA primer and gap fdling by the nick translation activity of DNA Pol I. the two adjacent Okazaki fragments am joined by DNA ligase. Okazaki fragments commence synthesis at precise points along the template, perhaps at 5"-GTGG-3" sequences (196), and it is the distance between these initiation signals that determines the average length of the fragments formed during replication. These, as well as the leading strand am polymerized by DNA Pol IlL How could the same enzyme moving in one direction along the replication fork replicate simultaneously the leading and lagging strands? A model (fig. 2) solving this paradox has been proposed (406). This model proposes that two copies of DNA Fol BI are present at the replication fork, and that the lagging strand is looped backwards, perhaps around the polyo merase, in such a way that the primer made by the primosome is extended by DNA Pol III as the template for the lagging strand is drawn through the holnenzyme complex trig- ~ When the elongating chain roaches the previously synthesized Okazaki fragment the loop is relaxed releasing the lagging str'atd template (fig. 2. b --) c). By this time. leading strand synthesis provides a new s~re~ch ofss lagging strand (fig. 2, that loops bac~d ~a'.d p.,w~id~ a template for the symhesis of the next Okszaki fragment (fig. 2. d). By such a mecha- nism the leading sn-and would never rush too far ahead of lag- ging slrand ~ynthesis. Also, since the dimeric complex of DNA Pol FII would be associated with the primosome and hellcases (~'plisome), the primosnme wou~d be transleca~ed 40000140

IG9 Fi~ 2, Cycling of an asymmetric polymerase on the !agsing smmd of a replication f'ork. (a) A dimedc polymcrase replicating bolh slrands of the replication fork concu~rendy. (b) Completion of the synthesis of' an Okazaki fragment on the lagging swan& (c) The asymetric dimeric polymeme associazed only with the leading su~nd. (d) Pacilits~J, in pan, by its locallza~ion at the replication fork and its association with the replisome (not shown). Ihe lagsing strand polymerase can efficiently recycle to the next palmer synthesized at the replication fork. Repro- cloned with permission fi-om McHan~y. (,1~). along the replicating DNA at the same speed as the DNA Pol III holoenzym¢. McHonry and colleagues (406). have sug- gested that the two halves of the dimeric hoioenzyme are asymmetric and have distinct funodonai properties. This func- tional asymmetry would be anticipated for a dimeric holocn- zyme that replicates both leading and lagging strands simultaneously, sihce only the polymerase that replicates the lagging strands roust repeatedly undergo association-dissocia- tion cycles with the DNA. Whatever the exact mechanism..the bidirectional replic- ation of the £. coli chromosome tenninatm at rerC, (122, 233, 238, 240, 241, 286, 310). The roo~ recent review is by Kuero- pet et a/(310). The terminus of ~plkafim of the ~. coli chro- m~me is a large region (350 kb) flanked on both sides by the terminator sites TerA, TerB, TerC and TerD. TerA and TerD inhibit replication forks traveling cotmterclockwise, and TerC and TerB inhibit clockwise traveling forks. A terminator pro- tein, prod~-'t of the ms gone locaxed next to the TerB site is a contm-hellcase needed for site-spm:ific tenninatiun. Termin- ation ha~ been pmlmsed as a signal for cell division,~ a~d the terC region roay thus play a role in coupling DNA replic~ion to the cell ~'cle (,1~). The model we have described above ~'Vre~nts the current thinking dmut how the chmmmomal DNA of E. coli repli- cates. T~is model is based on a large body of genetic, bio- chemical, in vivo and in vitro work, much of it performed on E. coil bacteriophages and plasmids O01, ,107). Although these replicons differ from each other and from E. coil in soroe aspocts of replication, by and large the main features of the model applies to them. For example, bacteriophage T4 (285, 426a) codes for its own replication proteins including a DNA polymerase of 110 kd (product of T4 gone 43). a 35 kd SSB protein (product of gone 32). a polypeptide of 58 kd coded by gone 4.1 which appea~ to be homologous to DnaB helicase, etc. Other replicons, such as ColE! plasmids can replicate in vivo as well as in vitro in the absence of plasmid specified proteins and caa utilize some host-encoded enzymes such as DNA Pol I and DNA Pol llI ($38). Replication of ColEl is unidirectional from a specific origin of replication (380, $38). An alternative way to replicate circular DNA is the rolling circle roecllanism (190). This mode of replication has been adopted by some bacteriophages, such as lambda during pa~t of its vegetative cycle (183) and the bacterial fet- Idly factor F during foaling (639° 640). Other "bacteriophages (i. e. T7) replicate as linear structures (230, 430. 636). Never- thelesso in all these cases there is a replication fork moved by a multienzyme complex which is a true ~rcplicafi~n machine,,. and the replication machines of different replicons are quite simile" to each other. Howcvcr.~s Alhens (5) ,has p~inted out. in spite of their siroil~xities the rel~lication machines of differ- 40000141

II0 CBR em mplicons at, composed of pro:sin parts ~y ei&~ ~ ~ or &~ E. c~li ~1~. ~l~ve to ~h ouh~ ~o~ ~s~mbl~g fo~ ~vme~, ~d ~ ~h p~ m~t ~t B. Fidelity nminten~nce during replication. Error The infommtional content of DNA is d~mrmined by the pmci.~ s~qu~c¢ of its cl~in of bases. Mu~ions am ~ ~ ~~.~ey ~~yor~a in ~ ~xt c~r. ~ simpler m~ p~) su~fimfi~, w~ ~ of ~o ~. R~t of a ~ c~m~ a w~on (L e. ~T ~ G:~, ~ ~- m~ of ~ by a p~ or vi~ v~ (for ~pl~ T:A ~ A:T, or T:A ~ G:~ ~ a w~emion Morn ~pli~ mumfiom ~ ~ di~ ~ ~ n~xz ~ co~[~ ~ on ~im ~ of a DNA ~ubl~ ~lix ~~ ~ ~ s~ of g~ ~g ~n~ ~ ~ ~ ~ing ~ ~o ~ ~ol ~~ fo~ w~uh wo~d ~ m~o~ W~ ~ ~vk ~y (~, ~) su~m~ ~ ~ a ~i~ ~ f~ ~~ ~i~ (mr. 6~, fig. 3). Morn ~enfly Top~ ~d Fms~o (~) ~w onamd bases and syn isomers m explain s~nmneous ~vemi~ (fig. 3). B~ ~ m~~m of n~l~fi~ im~ md ~ s~liW of ~l~l~ helk~, ~ well ~ studies on ~he nonenzymafi~ fo~mion of symhvfi~ nucl~ it ~ ~n ~u~ ~ ~g wi$ a ~y of I~ to I~ (181, ~). Ho~va, in ~i~o ~~n~ of m~on ~ci~ sug~ ~ ~ avcr- to.l~" m~i~o~nfi~ ~ b~ p~ ~fi~ ISI, ~). ~ is six m ~ o~ of ~~ low~ p~dic~d from the studi~ mention~ ~v¢. A signific~t com~nents of ~e ~plic~ion m~hi~w, some of which we~ ~nfion~ ~l~r. ~ c~r avo~ce ~is~ ~ ~ilt ~m ~c ~pli~ ~ i~lf. ~ ~ ~ di~n ~ d~ng DNA ~ion ~ ~e m~g ~. ~c tint ~e~ is evidence thai the replication mach~c~ ~uces e~n by ~[~g ~ pm~ ~ or ~~g ~ i~o~ on~ At ~ ~ond stop, if ~ ~cv~ ~ ~ly ~n ~co~ into • ~nt DNA 3" ~ 5" ~~g ~viw of DNA Pol st~ ~ ~ ~ ~g ~ m ~ I~ (HT, 1~). ~e thi~ step is a ~plic~vn cv~c~on of ~tch~ which esca~d ~e ~t ~o ~scfim~vn ,,_// i - ademn~ thymine F~ & Proton migr~ion within bzse pairs can set off a chain of modificati~as that ends in mutation. At the top, the normal A' T base pair l~s two hydrogen bonds (doted linch). In the middle figure, the A:C mispair ~ as an in~a~dl~ (arrow). and t~ r~ulting rear- r~z~,.m~zt of electronic structure pzrmks two hydrogen bonds to focm with the ineonect tm~ C, In the bosom figure, m~ A:A mispair • cts as an ~ in the mutadoa A:T --* T:A (where the T:A bgte p~ is the san~ ~ t~ A:T shown, at tl~ top. but viewed from its tmde~i6v). Asain. proton migration has occurred (arrow). and the ad~tL, m oa tim fight Ires mtamd 18~ amumi its glycosidic bond into tim ~n vonfiguratitm. In t.his mispair, both adenin~s are abnormal. P~-~,~.xd from Ora~ a al. I.WL w~'~ I~missi~. ares incorr~t bases in ~e newly r, pli~ted DNA. j~t behind tim rvpli~afiun fork (151, 19~, 321, 431, 431, 473, 474). We will discuss e,~h of ~,, m~:hanisms b~low. 40000142

CBR 111 Thm'c is genetic eyid~n=e that DNA pol)~e~ influ- em~ corr~ base pair formation at a stage prior to proof mad- ing. The first rep~r~ was that of Reha-Kgantz and Bessman (477) showing that a mutant for T4 DNA po|ymer~a h~l reduced specificity in incorporating correct deoxyribonu- clcatides although it had normal 3" -~ -5' exonuclease tivity. In experiments with T4 DNA polym~rase., Ripley (482) showed that thebase analog 2-aminopurine (2-AP) which can ~nse transitions by enhanced mispairing (see chapter IV), produced mutations at different frequencies depending on whether it was present in the template strand or was an incoming base. This result argues against the notion that accuracy depends only on hydrogen bonding and indicates that the polynlerase is gecognizing the incorrect base. Other evidence implicates Pol I as playing an active albei! non- essential role in maintaining the informational integrity of chpomosomal DNA replication (1~1). A polA mutant of Salmonella typMnmr/um wi~h a 7-I l fold elevated spontaneous mutation ft~que~ was fmugl to have a Pol I enzyme which failed to discriminate properly between adenine and 2-AP (169). It is not yet clear how DNA polymerases can dis- criminate between correct and ingogrect bases, It has been suggested (3/3 and mrs. therein) that the first discz~ination step could msuk from a diffet~ce in free energy (G) between conger and incon~'t be~ pal~, a diff~nc~ that is in pm a ppope~y of the buse pa~ themselves but could be amplified hy DN& polymemsc and ether pro~ins, i. e. th~ difference in in the presence of polymera~ than in its absence. mechanisms proposed to explain enlmneed base discrimina- zion by polymerases inclz~le conformazional changes of the enzyme ax e~h nuc[eetide selection step, base-sp~cifi.c deotide binding subsites, and tightening of ~nzyme-templaze binding in dm prmen~ of tim correct nucleotide (mfs. above). Neighix~i~ b~ have an influ~c~ on dm fidelity of inser- tion a~. proofreMing of DNA polymerase (4~, and. it has been shown (31~) that SSB protein increases tenfold the fidelity of DNA synthesis ~n v#ro. This effect does no~ appear to be due to enimneed proof-marling by the polymerase, but rather to improved base s~lection which could be due to a template/SSB intem~ion resulting in incn~sed rigidity of the templam. Thus, in addition to its other roles, $SB plays a pole in improving DNA replication fidelity. Pmnfmzding is tan'led out by an enzymatic activity that is ¢ith~ ~ of or is associamd with dm DNA polymeras~. As we ]mve mentioned, this function is performed in E. coli by dm 3' --~ ~" ~unu¢]~iytic acdvity of tim epsi]on subtmit of Pol I[I (~3~, $36). Pol[ also has a proofreading 3' -+ ~' ~xonucleas~ activity (~ 301) which seems to work during pair mpl[c~io~ (see chWm" IV). The polymm~,_se of phag~ T4 also has a pmof-m~ling ~xonuclease. Using mutant T4 DNA polym~za.~ Muzycska et oJ (4~9) show~ a corr~tz- zion between the in v/vo mumzo£ phenoWpe ~.d z.he mzio of exonucJeuse to pelynmmse activity in t,#ro: mummr polym~ms- polymeras~ relzlively more exonuc]ease ~an the wJId type exonuelease acfiviW resided in the catalytic core of the er~jrne but lzter [t was shown tlmz tkis activity was acttmlly tan'fed out by the epsilon subunit, which was isolated by Echols az~ eo-wo~kezs (~35, ~3@. Thus, unlike ol~er po]y- merases clmz'a~fiz~l to ~zze, ~. co~i Pol ID c~z~es ~s pmof- z~ding ac~vZty on a ~ p~lypeptide- [-[ow~vcr, r~ poly- meriting subunk of Pol []I, sub~n~z alpt'~ zlso I~s 3" --~ 5' exoBu~l~,,..~ acti'¢it~', aml ~e associ~on of ~p~ with ~psilon ~mulat~ ~¢ p~ing ~fiviw of zhs laden It h~ ~en epsiton suh~it, ~ well ~ dn~V the gene for the ~m subunit ofDNA Pol ~ (s~ ~t~ 2) c~ ~ ia~uced by 2-~ (~. ~is ~ a ve~ ~g ~se~a~n s~ it i~di~ ~at ~plic- ~on fi~li~ h~ ~ ~d~ibl~ com~ncn~ a p~ of DNA ~ sysm~ (~ ~ ~@mr). ~ f~ that &s alpha sub- ~Z of DNA Pol ~ ~ 3" ~ 5' exonu~le~ actNi~ w~ll m ~ $' ~ 3' ~lymc~ (~) sho~ ~at ~is sub,it by i~If ~ s~iI~ to o~r pm~c DNA ~lymc~cs ~ow~v~r, dnuQ mu~ts show mu~h higher spontaneous m~on m~ ~ my d~ mu~m (tabl~ 2 ~d mrs. ., Un~ o~ ~imu~c~ ~ combination of ~ ~l~on md ~~g by DNA ~l~ insulin in m ~ ~ nf ~m I~ (~73). But ~ of ~es~ ~or avoid- of nu~l~ ~ph~ph~ ~ suppli~ ~ mw mamd~s for DNA s~ is ~ in ~ way (247, 31~ 319~ 373). I~ Pitro studios haw demonstrated that DNA replication ~deliw is de~ndent on co~t baltics of ~oxyribonu- cl~oti~ ~pbosph~s du~ng DNA synthesis. In ~h,o such DNA p~r ~~ cm haw pmfo~d gcn~c consc- q~n~s in ~pmk~otes md ~uk~oms (mrs. above). ~ ~gs suggest ~at int~llul~ concentrations of DNA ~or ~is ~ mgula¢~ m minim~c ~ frequency of mu~. R h~ ~n ~ th~ mplica[i~ complexes ~n~ pmmim ~ ~ ~volv~ in ~ con~l of nucleozi~ I~ {~). ~a mi~h~ b~ su~iv~ p~f~ding, iz cm still removed via th~ methyl~imcmd mismatch r~pair system (~1, l~ ~.~ ~, 4~1, 43~ 47~, 474). In addition to [~m ~g during mpli~ion, mismatch~ can also ~cur d~ng h~ol~om ~ombin~ion (~ chapmr V) md from t~ ~on of S-m~yl~tosinc which wi]l generate a T: G misp~Hn~ S~o E." coli ~s~s a cytosine m~thylmc (~m) such ~in~ans shoed n~ ~ ~ md E. coli md S. ~p~imurium ~s~s sysmms to r~ognizc md comet such mB~hm. ~ m~t ~xmnsiv~ly inv~tig~d ~ the methyl- d~[~ misma~h m~r sysmm of E. coli. A methyl-inde- ~ndent mismatch repair sys~m that r~uires a functional R~F W~uct (recF~ndent recombination will ~ dis- ~s~ in ch@t~z V) ~ ~n m~ned (~e r~f. $21) bu[ h~ not ~n ~xt~nsiv~ly invesfig~ md w~ will not discuss ~hcr. ~ou~ ~ mBm~tch m~ir sysmm is active dudng mFl~on it is not, st~ly s~ing, ~ e~r avoidance but a mp~r ~smm (1SI) since, m mentioned above, it can also ~t mis~ch~ net ~sing from misp~dng dudng mplic- ~on. ~ E. coIi, n~ly syn~iz~ DNA a me~yl grow is ~h~d to ~nin~ in Che s~quence GATC wherever that s~us~ ~um, bs~ ~em is a lag ~twe~n symhcsis ~d m~yl~i~. ~ n~ly synth~ized stud is mcogni~blc ~ msuR of ~is lag, since the parental DNA will have its ~ATC ~nc~ c~ng ~ -methyladsninc whem~ ~e ~w s~d will not fo¢ a short ~d~ of tim~. ~is ~d~ is. ~w~v~ long e~ for ~ new stud to ~ ~nizcd by • s ~h m~ e~m. Me~yl~i~ted mhma~cb repair ~u~ ~ p~ucm of fiv~ g~s (mbl~ 4}. ~ am dam. ~H, m~mutS ~ uvrD, mu~ons in which result in mum- t~pben~ (6~ I~ ~d t~te 4) i. ~. in~ in the spon- ~ mum~on ~ from ~n ~o a thous~fold. ge~e c~ far a DNA a~n~i~ ~yl~ which mcogniz~ 40000143

! 12 CBR Table 4 Primary F~nttian Gen~ Map Pa~tian Mutawr Phena~'pe GATC sites ~ ~, ~ S~). ~~~ a~- don ~te~ w~h b~ s~y m a ~ ~ ~ m~- m~c~ ~n~l~ wMch cl~v~ a ph~p~ ~d on si~ of ~e mmL p~ is ~ p~ of ~D m ~ ~ DNA ~li~ H, which may play a m~ ~ DNA ~on ~ ~~ ~ ex- cision of m~h L~ne (41~). Fi~t, MutS bin~ m a mism~h whi~ Mu~ ~l~ OH ~ • m ~$ng of Mu~ f~ ~ ~ of Mu~ ~ ~n M~. Two ~ff~t ~ls ~ve ~ ~ (~ 41~). M~ch ~in f~li~ en~ of c~p~ ~ helix is ~wo~ by ~1~ H Y ~ ~' ~ ~ t~p~e s~ dfiv~ by SSB ~ing m ~ sing~ ~ ~ fol- low~ by ~~ by DNA Poi ~. ~ ~ m~l b~ is initi~ ~ ~ mi~tch w~ ~iix distoninn, ~s d~ by M~ ~or M~, ~ pmv~ m ~t ~r ~H- c~ II, wi~ ~1~ des~bil~tion p~ng in ~ di~- ti~ if a heli~ ~ ~ m ~h s~. SSB is ~l~ to play a ~m in ~ in viwo sys~. ~ k~ ~ ~ m~h ~pair intemct functionally wi~ SSB: DNA helicme II unwinding ~ ~v~ by ~B; ~1~ L ~ ~ out t~ exunucl~ ex~ion of of ~ ~y ~ ~v~ ~o~ n~ (it ~ ~ ~ GATC sit~ ~ ~ ~ d~ of mo~ ~ o~ ~lo~ ~m ~ mi~h) di~ for DNA ~li~on. IlL Mutations and mutag~mcsis We Imve already defined mumions in the preceding c~o- twr as changes in the sequence of the hases that provide th~ inftnrnatiorral coa~nt of DNA. The simpt~t tYl~eS of mum- tions are base substitutions (transitions or transversions) which have also been def'med. We will describe next more complicated types of mutations and then consider various ways in which these can occur. Mutations R~ulfing from tl~ simultaneous substitution of two or morn adjacent bases h~ve been described in bacterio- phages and yeast (2,4a, 137, 181, 198, 453, ~ 484, 579). Frameshifl mutations result from the addition or deletion of one or mo~ ba~ paks in ntunbem which are not multiples of three (495, $84). These mutations were first d~cribed in genes encoding polypepfides and characterized as throwing the resding fi'amc oftl~ genetic code code out of phese. Since each amino acid iS coded by a group of throe adjacent bases. any clumge in tbe number of fl~'sc bases by a number not a multiple of ths~ will have this effect and thus inactivate the gone. Other mutations affecting from tens to many thousands or even millions of base pairs have also been described in • prokaryote~ as well as eukaryotcs. The largest of these arc called chromosomal mutations and can involve tbe rearrange- moat or loss of larg~ portions of a chromosome, or the addit- ion (insertion) of large DNA sequences. Mutations of this type a~ a~ the basis of many cancers and genedc disease in man and animals (6$, 67, 120a, 225, 275, 502, 661). Mum- tions of all types occur spontanenusly and also as a result of damage to DNA ~ by a number of physical and chemic- al agents (induced mutations). We will discuss spontaneous mutations fwst. A. Spontaneous mutations By clef'tuition, spontaneous mutations occur in d~e absence of de~abk exogenous mutagenic exposure. This, howevcr dc~ not mean that at least a certain number of them cannot be c~msed by DNA damage resulting, for example, from the endogenous generation by normal metabolic processes, of highly mutagenic free radicals (2, 12,13, $21). We will return to this Spo~ms mumgenesis has been examined fairly recent- ly in several reviews (151, 522, 525). Although all types of mutations are represented in collections of spontaneous mutants, some tend to occur more frequently th,~n others and they are not randomly distributed within a gene. The inci- dence of each type of mutation is affected by the genetic background of the cell. the locus involved and the position within the locus (151), but DNA seq~mnce a~-ado~ involving several bases were found to constitute-the majority of sp~nta- n.-ous mutations. In an e.~ly study of the spect~rn of sport- 40000144

113 mncous mutations in the. Iacl Sen'- of E. call (160) Fambwagh et al found that ~ .% (of a tclal of 140 mut~ons studied) ccnsist~i of (a) deletions ranging in si~ from 4 to 123 bases, (b) ~.a~litior, s (4 base) or (c) ir, s~tiora of a transposabls merit. We will discuss transposable elements later in this chapter. Almost 90% of timse mutations were located at a sin- gle site (ahot Slmt~) in the gene. Approximamly half of deletions occurred between 5 to 8 hp direct repeats which could be situated up to many hundreds of bases apart. As we shall see laser, many deletions occur between direct terminal repeats. In a later study of the spontaneous mutation spectrum of laci using a different experimental system ($2S) 174 muta- tions were analyzed. Two thirds were foumi to be frameshifis occurring tn one hotspot sire. i)el~otm comprised the largest nm-hompot class (37%), and base substitmiom were 34% of the non-hompot events. Thus sponmneotm mutations show ,~metational specificity,~ as do induced mutations. This will be Some of the possible somces of sponumeous mutations have been identified (151). [:or example, the existence of mutants with mumtor phenotypes (103) and mentioned in chapter lI shows that mutations affecting the fidelity of replic- ation, as well as other mutations affecting components of posueplicatinn repair systems (to be discussed later, chapters frameshifts or deletions at specific loci by several orders of magnitude. Much important information was obtained in studies using mutants for the DNA polymerase of bacteo riophage T4 (13& 137, 197,, 4~9, 477, ,1~'). ~his polymenme carries both tim polymerizing and proof-mating activities on the same polypeptide ($). The spontmmmm rams of mutation shown by mummr and amimmtot" (giving a decretsed inme- tatinn frequency) T4 DNA polymenmes agreed with a gener- alized theory relating the fidelity of DNA synthesis4o the rel- ative ratios of polymerase and 3' exonucleolytic editing activity of the mutant polymerases (3"/, 192, 429, 4"/'/). Genetic studies of T4 mutants (reviewed in 373) demonstrat- ed, however, large differential mutato~ and antlmutetor effects which deponded on the particular substitmion muuaion being memumd, the mutagen met and the method of mememment. Thus. depending on the site being examined, an antimetmor could actually have mutator characte~tics. As mentioned in chapter ii mutants involved in the maintenance of replic~on fidelity showed mutetor phenotypos. This includes mutants for dnaQ (which cedes for the epeilon subenit, refs. $3& $36) as well as dna£ mutants with altered 3" --~ 5" exonuclease pmofi~tding activity (2~9), mutants in mismau:h repair genes (61, 103) and mumms for Pol I (159). $SB (34& 414a) and helicase II (103, 443). Other mutttor genes that have been found do not seem to affect components of the replication machinery, but their role in DNA metabolism has not yet been elucidated. A mutetor gene, muir fast reported by Tref- fers el al (6e8) was show~ to cause specifically A:T -~ C:G transversions. The mud" gene codes for a triphosphatase which has a preference for dGTP (38, $2). Ba...~l on in v#ro sm~es of the mplicalion of the single stranded bacteriophage Ml3mp2. Schaaper and Dram ($'X]) have conchtded tl~ A: G rather that T.'C mispai~s occur in extracts of muir cells. The m'.#ir gene., has bee.n cl_-~....~ ~.~ sequence_ (3) and it h~ been proposed that the normal function of m,.~ir ~s to prevem dG: d.A mispairing during DNA replication (4). Two mumtors which cause specifically G:C -~ T:A uansversions lmve been described recently ($9, 436). One. cf them, mu~¥ mal~s at 64 minums (436) and the other m,.# M ~ 81 m~nums ($9) on E. coli chromosome map. It has keen suggest~l ($9) that they may indicate the presence of a m~ch mp~r msch=ism indecent of ~, ~H~ pa~way which ~nczions in viro to p~t ~cifi~ly G:C ~ T:A As ~, ms~ show, E. cell h~ ~velo~d s~,gi~ for avoidMg mi~ ~ng ~s mpli~on of im DNA =d ven~g b~ su~fimfio~. Cen~n ~eshi~ ~d deletions c~ ~c~ no~ ~ ~e ~ult of misses m~e du~ng mplic- ~on ~ ~er ~m ~e ~s~si~on or ~nion of t~s- ~le elemenm (s~ ~low) or ~m ~e resolution of z~- sien~ ~n~ s~ which ~ fo~ on singie-s~ s~ch~ of DNA, such ~ ~o~ ~cu~g du~ng chromo- mine ~li~ion (~ fig. I). We will di~s next mine m~- e~ f~ s~n~ ~es~f~,: multiple b~e substitutions ~d ~jor B. Mixalignment mutagene$is The flint rondel of misalignment mutagenesis was offered by Streisinger el al ($84) to explain the observation that addk- ions or deletions of one or a few base pairs often occurred within short, redundant sequences (reviewed in 483, 495). One well studied case of a hot.spot for frameshifts is that originally reported by Farabaugh er al (160) in the lacl gene of ~. coll. This hotapot shows the sequence 5"-CTGG- 3' re- peated in tandem three time~. Mutations occur at a high rote within this sequence and result from either the addition or deletion of a CTGG unit. A model to explain how such events can take place on replicating DNA is shown in fig. 4 using a different repeating sequence (run of A's). It proposes that frameshift mutations axe generated at single stranded gaps in the DNA which could be produced during DNA repair, re- combination or replication. Such structures have the capability of opening base pairs, slipping and re-annealing before slip- ping back. In fig. 4, addition or deletion of a single base is determined by the direction of slippage and results from the continued replication of the new DNA strand. This model. with minor variations, can also explain complex base pair sub* stitutions and deletions. Its main feature is that it postulates cot, rect base pairing in an inconuct context (I37). An example of a complex base pair substitution in yeast (579) is shown in fig. 5. The mutation consists of a +1 frameshift (one base addition) plus at least two simultaneous base substitutions, and it occu~ repeatedly. AS shown in the figure, the mutation can be readily understood as the conversion of aquasipolindrome to a more perfec¢ palindrome in three steps, involving incision and excision of the mismatched three-base stretch followed by closing of the gap by DNA synthesis using the apposite side of the stem as template. Mutations of this type have also been reported in bacterioplmge T4 (120, 137, 481. 483, 484). The model was extended to explain deletions by Aibertini el al (6) and Giickman and Ripley (198, 484). and is shown in fig. 6. The model postulates that during DNA replication single stranded regions can undergo slippage followed by interstrand mispairing between direct repeats, or intrastrand mispairing b~wee~ inverted repeats (lY~indron~s) or both to form a hair- pin structure such as the one shown in fig. 7. These unstable in~erme~ates could then be resolved as deletions in various ways (see legend to fig. 6). Although the existence of rnisvair- ed structures has not been ~ly demonstrat~l, this model is consistent with a large body of evidence; i. e. most reported deletions occur between dY.rect repeats or at the end of palin- d~mcs (6, 7, 8, 26, 30, 116, 198, 269, 281, 484, $$7° 632- 40000145

114 CBR 4. The ntisalignmmm muta~-tmb model for fraateahi~ of Streisin~" er a/(mr.). "rim nmmai l~Ogtay straad of this sequence contains a stri~ cff 6 adeaines, aad mistligaeatetz mutttiom could give either 7 acleniues (left) o~ ~ adtniues (fight). "l'ne next round of DNA replication shown) will sea, ate these stzucuuts into laatiy mutant tad lmmty nmmal DNA. Simple variations oa such s~hemes caa also produ¢~ aclditims o¢ ~letioas of two or more tam Fdas (s~e tcf. 4is). Rcptodm~ wi~ ~ from Dmka er a! (137). ilIIIIIIONA$~MIII~, $'oToC°A°A'G'oC° C°G'G oTo'ro~ m0NAm S° o'r oC- A- A-G-(: ° &oG C-1" ' G G.. • ToA C-T Fit. 5. A cornlgeX mutati~m in yea~ (579) can be explaimd from the co~etsicm of a~ imt~fect ~r~ (or q~i~). ~ f~ a ~ ~ ~ ~ a ~ for a DNA s~ ~g ~ ~i~ si~ ~ ~ ~ ~ ~ a ~ ~ ~ ini~ ~ C~ 40000146