template × something whose edge is shaped in a particular way so that it can serve as a guide in making a similar object with a corresponding contour

Chapter 30

DNA Replication and Repair

“Noah’s Ark” by Andre Normil (Superstock)

Figure 30.1 × Watson and Crick’s famous paper, in its entirety. (Watson, J. D., and Crick, F. H. C., 1953. Nature 171:737-738.)

The publication of Watson and Crick’s famous paper titled Molecular Structure of Nucleic Acids: A Structure for DeoxyriboseNucleic Acid (Figure 30.1) marked the dawn of a new scientific epoch, the age of molecular biology. As these authors drew to a close their brief but far-reaching description of the DNA double helix, they pointedly commented, “It has not escaped our notice that the specific [base] pairing we have postulated immediately suggests a possible copying mechanism for the genetic material.” The mechanism for DNA replication that Watson and Crick viewed as intuitively obvious is strand separation followed by the copying of each strand. In the process, each separated strand acts as a template for the synthesis of a new complementary strand whose nucleotide sequence is fixed by the base-pairing rules Watson and Crick proposed. Strand separation is achieved by untwisting the double helix (Figure 30.2). Base pairing then dictates an accurate replication of the original DNA double helix.

Figure 30.2 × Untwisting of DNA strands exposes their bases for hydrogen bonding. Base pairing ensures that appropriate nucleotides are inserted in the correct positions as the new complementary strands are synthesized. By this mechanism, the nucleotide sequence of one strand dictates a complementary sequence in its daughter strand. The original strands untwist by rotating about the axis of the unreplicated DNA double helix.

30.1 · DNA Replication Is Semiconservative

Actually, three basic models for DNA replication are consistent with the requirement that the nucleotide sequence in one strand dictate, through Watson and Crick’s base-pairing rules, the sequence of nucleotides in the other. These three models—conservative, semiconservative, and dispersive—are diagrammed in Figure 30.3. In 1958, Matthew Meselson and Franklin Stahl provided the experimental proof for the semiconservative model of DNA replication. Escherichia coli cells were grown for many generations in media containing ¹⁵NH₄Cl as the sole nitrogen source. Thus, the nitrogen atoms in the purine and pyrimidine bases of the DNA in these cells were mostly ¹⁵N, the stable heavy isotope of nitrogen. Then, a tenfold excess of ordinary ¹⁴NH₄Cl was added to the growing culture, and, at appropriate intervals, cells were collected from the culture and lysed. The DNA they contained was analyzed by CsCl density gradient ultracentrifugation (Chapter 12 Appendix). This technique can resolve macromole­cules differing in density by less than 0.01 g/mL.

Figure 30.3 × Three models of DNA replication prompted by Watson and Crick’s double helix structure for DNA. (a) Conservative: Each strand of the DNA duplex is replicated and the two newly synthesized strands join to form one DNA double helix while the two parental strands remain associated with each other. The products are one completely new DNA duplex and the original DNA duplex. (b) Semi­conservative: The two strands separate, and each strand is copied to generate a complementary strand. Each parental strand remains associated with its newly synthesized complement, so that each DNA duplex contains one parental strand and one new strand. (c) Dispersive: This model predicts that each of the four strands in the two daughter DNA duplexes contains both newly synthesized segments and segments derived from the parental strands.
 

           DNA isolated from cells grown on ¹⁵NH₄¹ (the “0” generation cells) banded in the ultracentrifuge at a density corresponding to 1.724 g/mL, whereas DNA from cells grown for 4.1 generations on ¹⁴N had a density of 1.710 g/mL (Figure 30.4). These bands represent “heavy” and “light” DNA, respectively. Significantly, DNA isolated from cells grown for just one generation on ¹⁴N yielded just a single band corresponding to a density of 1.717 g/mL, halfway between heavy and light DNA, indicating that each DNA duplex molecule contained equal amounts of ¹⁵N and ¹⁴N. This result is consistent with the semiconservative model for DNA replication, at the same time ruling out the conservative model. After approximately two generations on ¹⁴N, cells yielded DNA that gave two essentially equal bands upon ultracentrifugation, one at the intermediate density of 1.717 g/mL and the other at the light position, 1.710 g/mL, also in accord with semiconservative replication.

Figure 30.4 × The Meselson and Stahl experiment demonstrating that DNA replication is semiconservative. On the left are shown densitometric traces made of UV absorption photographs taken of the ultracentrifugation cells containing DNA isolated from E. coli grown for various generation times after ¹⁵N-labeling. The photographs were taken once the migration of the DNA in the density gradient had reached equilibrium. Density increases from left to right. The peaks reveal the positions of the banded DNA with respect to the density of the solution. The number of generations that the E. coli cells were grown (following 14 generations of ¹⁵N density-labeling) is shown down the middle of the figure. A schematic representation interpreting the pattern expected of semiconservative replication is shown on the right side of this figure. (Adapted from Meselson, M., and Stahl, F. W., 1958. Proceedings of the National Academy of Sciences, USA 44:671-682)

            It remained conceivable that DNA replication might follow a dispersive model, so that both strands of DNA in the chromosomes of cells after one generation on ¹⁴N might be intermediate in density. Meselson and Stahl eliminated this possibility by the following experiment: DNA isolated from ¹⁵N-labeled cells kept one generation on ¹⁴N was heated at 100°C so that the DNA duplexes were denatured into their component single strands. When this heat-denatured DNA was analyzed by CsCl density gradient ultracentrifugation, two distinct bands were observed, showing that one strand was ¹⁵N-labeled and the other strand was ¹⁴N-labeled. A dispersive mode of replication would have yielded a single band of DNA of intermediate density. An interesting feature of density gradient ultracentrifugation is that the width of the band occupied by a macromolecular species in the gradient is inversely proportional to the square root of the molecular weight of the macromolecule. The band widths of the two bands seen with heat-denatured DNA both correspond to molecular masses one-half that of native DNA. This experiment established that DNA replication proceeded by a semiconservative mechanism and also verified that DNA was indeed composed of two polynucleotide strands of equal size.

30.2 · General Features of DNA Replication

Replication Is Bidirectional

Replication of DNA molecules begins at one or more unique sites called origin(s) of replication (discussed in Section 30.4), and, excepting certain bacteriophage chromosomes and plasmids, proceeds in both directions from this origin (Figure 30.5). For example, replication of E. coli DNA begins at oriC, a unique 245-bp site. From this site, replication advances in both directions

Figure 30.5 × Bidirectional replication of the E. coli chromosome. (a) Unidirectional versus bidirectional replication can be resolved if cells grown several generations in the presence of low amounts of radioactive ³H-thymidine to lightly label the chromosome are then exposed briefly to high levels of ³H-thymidine to cause heavy labeling in newly synthesized regions of DNA. If replication is unidirectional, only one advancing replication fork is present and only the DNA adjacent to it should be heavily labeled. If replication is bidirectional, autoradiograms of replicating chromosomes should show two replication forks heavily labeled with radioactive thymidine. (b) An autoradiogram of the chromosome from a dividing E. coli cell shows bidirectional replication. Unidirectional replication has been observed only in bacteriophages and some plasmids. (Photo courtesy of David M. Prescott, University of Colorado)

around the circular chromosome. That is, bidirectional replication involves two replication forks, which move in opposite directions. Unwinding the DNA Helix Semiconservative replication depends on unwinding the DNA double helix to expose single-stranded templates to polymerase action. For a double helix to unwind, it must either rotate about its axis (while the end of its strands are held fixed), or positive supercoils must be introduced, one for each turn of the helix unwound (Chapter 12). If the chromosome is circular, as in E. coli, only the latter alternative is available. Because DNA replication in E. coli proceeds at a rate approaching 1000 nucleotides per second, and there are about 10 bp per helical turn, the chromosome would accumulate 100 positive supercoils per second! In effect, the DNA would become too tightly supercoiled to allow unwinding of the strands.
            DNA gyrase, a Type II topoisomerase, acts to overcome the torsional stress imposed upon unwinding by introducing negative supercoils at the expense of ATP hydrolysis. The unwinding reaction is driven by helicases, a class of proteins that catalyze the ATP-dependent unwinding of DNA double helices. Unlike topoisomerases that alter the linking number of dsDNA through phosphodiester bond breakage and reunion (Chapter 12), helicases simply disrupt the hydrogen bonds that hold the two strands of duplex DNA together. A helicase molecule requires a single-stranded region for binding. It then moves along the DNA strand, its translocation coupled to ATP hydrolysis and to strand unwinding. SSB (ssDNA-binding protein, Chapter 29) binds to the unwound strands, preventing their re-annealing. At least 10 distinct DNA helicases involved in different aspects of DNA and RNA metabolism have been found in E. coli alone.

Replication Is Semidiscontinuous

The autoradiographic results (Figure 30.5) indicate that the two strands of duplex DNA are both replicated at each advancing

Figure 30.6 × The semidiscontinuous model for DNA replication. Newly synthesized DNA is shown as red. Because DNA polymerases only polymerize nucleotides 5' ® 3', both strands must be synthesized in the 5' ® 3' direction. Thus, the copy of the parental 3 ' ® 5' strand is synthesized continuously; this newly made strand is designated the leading strand. (a) As the helix unwinds, the other parental strand (the 5' ® 3' strand) is copied in a discontinuous fashion through synthesis of a series of fragments 1000 to 2000 nucleotides in length, called the Okazaki fragments; the strand constructed from the Okazaki fragments is called the lagging strand. (b) Because both strands are synthesized in concert by a dimeric DNA polymerase situated at the replication fork, the 5' ® 3' parental strand must wrap around in trombone fashion so that the unit of the dimeric DNA polymerase replicating it can move along it in the 3' ® 5' direction. This parental strand is copied in a discontinuous fashion because the DNA polymerase must occasionally dissociate from this strand and rejoin it further along. The Okazaki fragments are then covalently joined by DNA ligase to form an uninterrupted DNA strand.

replication fork by DNA polymerase. DNA polymerase uses ssDNA as a template and makes a complementary strand by polymerizing deoxynucleotides in the order specified by their base-pairing with bases in the template. DNA polymerases synthesize DNA only in a 5' ® 3' direction, reading the antiparallel template strand in a 3' ® 5' sense. A dilemma arises: How does DNA polymerase copy the parent strand that runs in the 5' ® 3' direction at the replication fork? It turns out that the two daughter strands are synthesized in different ways so that replication is semidiscontinuous (Figure 30.6). As the DNA helix is unwound during its replication, the 3' ® 5' strand (as defined by the direction that the replication fork is moving) can be copied continuously by DNA polymerase proceeding in the 5' ® 3' direction behind the replication fork. The other parental strand is copied only when a sufficient stretch of its sequence has been exposed for DNA polymerase to move along it in the 5' ® 3' mode. Thus, one parental strand is copied continuously to give a newly synthesized copy, the leading strand; the other parental strand is copied in an intermittent, or discontinuous, mode to yield a set of fragments. These fragments are then joined to form an intact lagging strand.

The Lagging Strand Is Formed from Okazaki Fragments

Figure 30.7× An electron micrograph of DNA replication in cultured Chinese hamster ovary (CHO) cells. A replication bubble is indicated by the large arrow. Note the two replication forks that define the bubble; each fork is characterized by a thinner, single-stranded region (indicated by the small arrows) on one of its two branches. Also note that the two single-stranded regions are trans to each other, as predicted by the semidiscontinueous model for DNA replication. (Photo courtesy of Joyce L. Hamlin, University of Virginia)

In 1968, Tuneko and Reiji Okazaki provided biochemical verification of the semidiscontinuous pattern of DNA replication just described. The Okazakis exposed a rapidly dividing E. coli culture to ³H-labeled thymidine for 30 seconds, quickly collected the cells, and found that half of the label incorporated into nucleic acid appeared in short ssDNA chains just 1000 to 2000 nucleotides in length. (The other half of the radioactivity was recovered in very large DNA molecules.) Subsequent experiments demonstrated that, with time, the newly synthesized short ssDNA Okazaki fragments became covalently joined to form longer polynucleotide chains, in accord with a semidiscontinuous mode of replication. The generality of this mode of replication has been corroborated in electron micrographs of DNA undergoing replication in eukaryotic cells (Figure 30.7).

30.3 · DNA Polymerases—The Enzymes of DNA Replication

All DNA polymerases, whether from prokaryotic or eukaryotic sources, share the following properties: (a) The incoming base is selected within the DNA polymerase active site, as determined by Watson-Crick geometric interactions with its corresponding base in the template strand, (b) chain growth is in the 5' ® 3' direction and is antiparallel to the template strand, and (c) DNA polymerases cannot initiate DNA synthesis de novo—all require a primer oligonucleotide with a free 3'-OH to build upon. E. coli DNA Polymerases Table 30.1 compares the properties of the various DNA polymerases in E. coli. These enzymes are numbered I, II, and III, in order of their discovery. DNA polymerases I and II function principally in DNA repair; DNA polymerase III is the chief DNA-replicating enzyme of E. coli. Only 10 to 20 copies of this enzyme are present per cell. E. coli DNA polymerase I

Figure 30.8 × The chain elongation reaction catalyzed by DNA polymerase. DNApolymerase I joins deoxynucleoside monophospahter unites to the 3'-OH carries out a nucleophilic attack on the a-phosphoryl group of the incoming dNTP to form a phosphoester bond, and PP_i is released. The subsequent hydrolysis of PP_i by inorganic pyrophosphatase renders the reaction effectively irreversible.

In 1957, Arthur Kornberg and his colleagues discovered the first DNA polymerase, DNA polymerase I. DNA polymerase I catalyzed the synthesis of DNA in vitro if provided with all four deoxynucleoside-5'-triphosphates (dATP, dTTP, dCTP, dGTP), a template DNA strand to copy, and a primer. A primer is essential because DNA polymerases can elongate only pre-existing chains; they cannot join two deoxyribonucleoside-5'-phosphates together to make the initial phosphodiester bond. The primer base-pairs with the template DNA, forming a short, double-stranded region. This primer must possess a free 3'-OH end to which an incoming deoxynucleoside monophosphate is added. All four dNTPs are substrates, pyrophosphate (PP_i) is released, and the dNMP is linked to the 3'-OH of the primer chain through formation of a phosphoester bond (Figure 30.8). The deoxynucleoside monophosphate to be incorporated is chosen through its geometric fit with the template base to form a Watson-Crick base pair. As DNA polymerase I catalyzes the successive addition of deoxynucleotide units to the 3'-end of the primer, the chain is elongated in the 5' ® 3' direction, forming a polynucleotide sequence that runs antiparallel to the template but complementary to it. DNA polymerase I can proceed along the template strand, synthesizing a complementary strand of about 20 bases before it “falls off” (dissociates from) the template. The degree to which the enzyme remains associated with the template through successive cycles of nucleotide addition is referred to as its processivity. As DNA polymerases go, DNA polymerase I is a modestly processive enzyme. Arthur Kornberg was awarded the Nobel Prize in physiology or medicine in 1959 for his discovery of this DNA polymerase. DNA polymerase I is the best-characterized of these enzymes, so its properties are discussed first.

E. coli DNA Polymerase Has Three Active Sites on Its Single Polypeptide Chain

Figure 30.9 × A representation of the solvent-accessible surface of the DNA polymerase I Klenow fragment with bound DNA. The "template" strand (12 nucleotides) is blue; the "primer" strand (14 nucleotides) is red. The thumblike region of the protein that changes conformation upon DNA binding is indicated. Note that the vertical cleft is an extension of the horizontal cleft in the protein that DNA occupies. In this picture, the 3'-end of the primer strand lies within the 3'-exonuclease catalytic site of the enzyme. The polymerase site of the enzyme lies almost directly above the 3'-exonuclease site, in the vertical cleft (see Figure 30.10). (Adapted from Beese, L. S., Derbyshire, V., and Steitz, T. A., 1993. Science 260:352 - 355. Photograph courtesy of Thomas A. Steitz of Yale University)

E. coli DNA polymerase I is a 109-kD protein consisting of a single polypeptide of 928 amino acid residues. In addition to its 5' ® 3' polymerase activity, DNA polymerase I has two other catalytic functions, a 3' ® 5' exonuclease (3'-exonuclease) activity and a 5' ® 3' exonuclease (5'-exonuclease) activity. The three distinct catalytic activities of DNA polymerase I reside in separate active sites in the enzyme. As shown by Hans Klenow, the DNA polymerase I polypeptide chain can be cleaved into two fragments by limited proteolysis with subtilisin or trypsin. The smaller fragment (residues 1 through 323) contains the 5'-exonuclease activity, whereas the larger fragment (residues 324 through 928, the so-called Klenow fragment) has the polymerase and 3'-exonuclease activities. Thomas Steitz and his colleagues have analyzed the three-dimensional structure of the Klenow fragment via X-ray crystallography (Figure 30.9). Upon DNA binding, a conformational change takes place in a thumblike region of the protein that makes direct contacts with duplex portions of DNA. The polymerase and 3'-exonuclease active sites of the Klenow fragment lie within a common cleft that forms virtually a right-angle bend within the protein, but these two catalytic sites are surprisingly far apart—about 3.5 nm—and lie in separate arms of the bend. The DNA enters the cleft from the end closest to the 3'-exonuclease site. Apparently, the 3'-terminus of the DNA primer is shuttled between each catalytic site—the polymerase active site and the 3' ® 5' exonuclease site (Figure 30.10), and the polymerization and

Figure 30.10 × A model for the interaction of the Klenow fragment with DNA. (a) The 3'- end of the primer/growing chain resides in the polymerase active site. (b) The 3'- end is situated in the 3'- exonuclease active site. The protein must slide along the DNA in shifting the 3'- end of the growing chain from one of its active sites to the other. Thus, polymerization and the 3'- editing functions may occur without dissociation of the DNA from the Klenow fragment (see Figure 30.11). (Adapted from Beese, L. S., Derbyshire, V., and Steitz, T. A., 1993. Science 260:352 - 355)

editing functions of the enzyme (see below) can be carried out without dissociation of the DNA from the protein. E. coli DNA Polymerase I Is Its Own Proofreader and Editor The exonuclease activities of E. coli DNA polymerase I serve proofreading and editing functions that enhance the accuracy of DNA replication. The 3'-exonuclease activity removes nucleotides from the 3'-end of the growing chain (Figure 30.11), an action that apparently negates the effects of the polymerase activity. Its purpose, however, is to remove incorrect (mismatched) bases. Although the 3'-exonuclease works slowly when compared to the polymerase, the polymerase cannot elongate an improperly base-paired primer terminus. Thus, the relatively slow 3'-exonuclease has time to act and remove the mispaired nucleotide. Therefore, the polymerase active site is a proofreader and the 3'-exonuclease activity is an editor. This check on the accuracy of base pairing enhances the overall precision of the process.

Figure 30.11 × The 3' ® 5' exonuclease activity of DNA polymerase I removes nucleotides from the 3'-end of the growing DNA chain.

        The 5'-exonuclease of DNA polymerase I acts upon duplex DNA, degrading it from the 5'-end by releasing mono- and oligonucleotides. It can remove distorted (mispaired) segments lying in the path of the advancing polymerase. Its biological roles depend on the ability of DNA polymerase I to bind at nicks (single-stranded breaks) in double-stranded DNA (or RNA:DNA hybrids) and move in the 5' ® 3' direction, removing successivenucleotides with its 5'-exonucleolytic activity. (If the substrate is dsDNA, these nucleotides are deoxynucleotides; if the

Figure 30.12 × (a) The 5' ® 3' exonuclease activity of DNA polymerase I can remove up to 10 nucleotides in the 5'-direction downstream from a 3'-OH single-strand nick. (b) If the 5' ® 3' polymerase activity fills in the gap, the net effect is nick translation by DNA polymerase.

substrate is an RNA:DNA hybrid, the products are ribonucleotides.) DNA polymerase I is involved in DNA repair processes (Section 30.8). In an error-correcting role during DNA repair, the 5' ® 3' polymerase activity of DNA polymerase I fills in the sequence behind the 5'-exonuclease activity, so that the enzyme carries out a nick translation (translation meaning movement) along the DNA molecule (Figure 30.12). No net ynthesis of DNA results, but in vivo, this action of DNA polymerase I can “edit out” sections of damaged DNA, if a nick with a free 3'-OH is present. This excision is coordinated with 5' ® 3' polymerase-catalyzed replacement of the damaged nucleotides so that DNA of the right sequence is restored.

The Chemistry of DNA Synthesis Favors Semidiscontinuous Replication A consideration of the chemistry of chain elongation suggests why semidiscontinuous DNA replication and the inherently complex mechanism of lagging strand synthesis evolved. Note that 5' ® 3' synthesis takes an activated intermediate, the deoxynucleoside-5'-triphosphate, and forms a phosphoester bond by joining its 5'-a-phosphate to a 3'-OH group with elimination of PP_i:

dN*TP+pNpNpNpN-3'OH ® PP_i+pNpNpNpNpN*-3'OH The 3'-OH of this newly added deoxynucleotide provides the acceptor for the next incoming deoxynucleotide. The alternative, linking of deoxynucleoside-5'-triphosphates in the 3' ® 5' direction, would require the growing chain to present a reactive phosphoriC anhydride at its 5'-end to activate condensation with the 3'-OH of an incoming nucleotide:

dN*TP+pppNpNpNpN-3'OH ® pppN*pNpNpNpN-3'OH+PP_i This mechanism is energetically feasible, but error correction by proofreading would create an energetically unfavorable situation: The proofreading activity would be a 5' ® 3' exonuclease

pppN*pNpNpNpN-3'OH ® pppN*-3'(p or OH)
+(OH or p)NpNpNpN-3'OH and its action would leave a 5'-OH or a 5'-phosphate ester group at the growing end of the chain, neither of which forms a phosphodiester linkage unless chemically activated. Without activation, chain extension would be aborted.

E. coli DNA Polymerase III

“CORE” DNA POLYMERASE III.  The simplest form of DNA polymerase III showing any DNA-synthesizing activity in vitro is 165 kD in size and consists of three polypeptides: a (130 kD), e (27.5 kD), and q (10 kD). This “core” DNA polymerase III binds at short, single-stranded regions (<100 nucleotides) created by nuclease treatment of dsDNA and fills in the gaps to re-establish the duplex structure. It cannot initiate synthesis on intact duplex DNA. The a subunit carries out the catalytic polymerase function. The e subunit performs the 3' ® 5' exonuclease activity and contributes proofreading ability to the “core” polymerase. The role of subunit q is unknown.

Figure 30.13 × (a) Ribbon diagram of the b subunit dimer of the DNA polymerase III holoenzyme on B-DNA, viewed down the axis of the DNA. One monomer of the b subunit dimer is colored red and the other yellow. The centrally located DNA is mostly blue. (b) Space-filling model of the b subunit dimer of the DNA polymerase III holoenzyme on B-DNA. One monomer is shown in red, the other in yellow. The B-DNA has one strand colored white and the other blue. The hole formed by the b subunits (diameter < 3.5 nm) is large enough to easily accommodate either A-form or B-form DNA (diameter < 2.5 nm) with no steric repulsion. The rest of pol III holoenzyme ("core" polymerase + g complex) associates with this sliding clamp to form the replicative polymerase (not shown). (Adapted from Kong, X.-P., et al., 1992. Cell 69:425-437; photos courtesy of John Kuriyan of the Rockefeller University)

DNA Polymerase III Holoenzyme In vivo, “core” DNA polymerase III functions as part of a multisubunit complex, the DNA polymerase III holoenzyme, which is composed of 10 different kinds of subunits (Table 30.2). The various auxiliary subunits increase both the polymerase activity of the “core” enzyme and its processivity. DNA polymerase III holoenzyme synthesizes DNA strands at a speed of nearly 1 kb/sec. DNA polymerase III holoenzyme is organized in the following way: two “core” (aeq) DNA polymerase III units and one g complex are held together by a dimer of t subunits in a structure known as DNA polymerase III*. In turn, each “core” polymerase within DNA polymerase III* binds to a b-subunit dimerto create DNA polymerase III holoenzyme. The g complex is responsible for assembly of the DNA polymerase III holoenzyme complex onto DNA. The g complex of the holoenzyme acts as a clamp loader by catalyzing the ATP-dependent transfer of a pair of b subunits to each strand of the DNA template. Each b-subunit dimer forms a closed ring around a DNA strandand acts as a tight clamp that can slide along the DNA

Figure 30.14 × The mechanism of action of DNA ligases. DNA ligases proceed via a ping-pong kinetic mechanism involving an enzyme-side chain e-NH₂ of a Lys residue i s adenylylated. The enzymes differ in that NAD⁺ is the AMP donor for the E. coli enzyme, whereas ATP serves this role for the eukaryotic and phage DNA ligases.

(Figure 30.13). Each b₂ sliding clamp tethers a “core” polymerase to the template, accounting for the great processivity of the DNA polymerase holoenzyme. This complex can replicate an entire strand of the E. coli genome (more than 4.6 megabases) without dissociating. Compare this to the processivity of DNA polymerase I, which is only 20!

A Deeper Look
A Mechanism for All Polymerases
Thomas A. Steitz of Yale University has suggested that biosynthesis of nucleic acids proceeds by an enzymatic mechanism that is universal among polymerases. His suggestion is based on structural studies indicating that DNA polymerases use a “two-metal-ion” mechanism to catalyze nucleotide addition during elongation of a growing polynucleotide chain (figure). The incoming nucleotide has two Mg21 ions coordinated to its phosphate groups, and these metal ions interact with two aspartate residues that are highly conserved in DNA (and RNA) polymerases. These residues in phage T7 DNA polymerase are D705 and D882. One metal ion, designated A, interacts with the O atom of the free _3'-OH group on the polynucleotide chain, lowering its affinity for its hydrogen. This interaction promotes nucleophilic attack of the 3'-O on the phosphorus atom in the a-phosphate of the incoming nucleotide. The second metal ion (B in the figure) assists departure of the product pyrophosphate group from the incoming nucleotide. Together, the two metal ions stabilize the pentacovalent transition state on the a phosphorus atom.
Adapted from Steitz, T., 1998. A mechanism for all polymerases. Nature 391:231-232. (See also Doublié, et al., Nature 391:251-258 and Kiefer, et al., Nature 391:304-307.)

DNA Ligase

DNA ligase (see Chapter 13) seals nicks in double-stranded DNA where a 3'-OH and a 5'-phosphate are juxtaposed. This enzyme is responsible for joining Okazaki fragments together to make the lagging strand a covalently contiguous polynucleotide chain. DNA ligase from eukaryotes and bacteriophage T4 is ATP-dependent; the E. coli enzyme requires NAD⁺. Both types of DNA ligase act via an adenylylated e-amino group of a Lys residue (Figure 30.14). Adenylylation of the 5'-phosphoryl group activates it for formation of a phosphoester bond with the 3'-OH, covalently sealing the sugar-phosphate backbone of DNA.

General Features of a Replication Fork

We now can present a snapshot of the enzymatic apparatus assembled at a replication fork (Figure 30.15 and Table 30.3).DNA gyrase (topoisomerase) and helicase unwind the DNA double helix, and the unwound, single-stranded regions of DNA are

Figure 30.15 × General features of a replication fork. The DNA duplex is unwound by the action of DNA gyrase and helicase, and the single strands are coated with SSB (ssDNA-binding protein). Primase periodically primes synthesis on the lagging strand. Each half of the dimeric replicative polymerase is a “core” polymerase bound to its template strand by a b-subunit sliding clamp. DNA polymerase I and DNA ligase act downstream on the lagging strand to remove RNA primers, replace them with DNA, and ligate the Okazaki fragments.

maintained through interaction with SSB. Primase synthesizes an RNA primer on the lagging strand; the leading strand, which needs priming only once, was primed when replication was initiated. The lagging strand template is looped around, and each replicative DNA polymerase moves 5' ® 3' relative to its strand, copying template and synthesizing a new DNA strand. Each replicative polymerase is tethered to the DNA by its b-subunit sliding clamp. The DNA pol III g complex periodically unclamps and then reclamps b subunits on the lagging strand as the primer for each new Okazaki fragment is encountered. Downstream on the lagging strand, DNA polymerase I excises the RNA primer and replaces it with DNA, and DNA ligase seals the remaining nick.

30.4 · The Mechanism of DNA Replication in E. coli

Replication of the E. coli chromosome begins at a single replication origin and proceeds bidirectionally until the two replication forks meet. At each replication fork, both leading and lagging strand syntheses are catalyzed by a single multiprotein replication machine, the so-called replisome, which consists of DNA-unwinding proteins; the priming apparatus, or primosome, which is needed to initiate, or “prime,” DNA replication; and DNA polymerase III holoenzyme with two equivalents of “core” polymerase, one for the leading strand and one for the lagging strand. As this replisome follows the replication fork, the template for lagging strand synthesis (the strand running 5' ® 3' in the direction of fork movement) must be looped around so that it can be read in the 3' ® 5' direction.

Figure 30.16 × The probable course of events during initiation of E. coli replication at oriC. Tetramers of DnaA protein bind at each of the four 9-bp repeats in oriC; then additional DnaA protein binds to give a nucleosomelike structure with DNA on the outside. The three 13-bp A:T-rich regions are then “melted” to yield an open complex, and the DnaB protein delivered from the DnaB:DnaC protein complex enters. The helicase activity of DnaB unwinds the duplex, displacing DnaA protein. SSB protein binds to single-stranded DNA as it is generated, preventing its re-annealing.

Initiation

Replication of the E. coli chromosome is initiated at a unique site, oriC. The 245-bp sequence of oriC contains elements that are highly conserved among Gram-negative bacteria. Within oriC are four 9-bp repeats. DnaA protein, a 52-kD polypeptide, is the initiation factor, which recognizes and binds to these repeats (Figure 30.16) DnaA protein binding is cooperative; once the four 9-bp repeats are occupied, 20 to 40 additional DnaA monomers bind so that the entire oriC region is complexed with DnaA protein. HU, a histone-like protein (Chapter 12), prevents nonspecific initiation at sites other than oriC. The resulting complex resembles a nucleosome, with negatively supercoiled oriC DNA wrapped around a DnaA core. The DnaA protein then mediates the separation of the strands of the DNA duplex by acting on three AT-rich tandem repeats (consensus sequence = 5'-GATCTNTNTTNTT, where N = any nucleotide) located at the 5'-end of the sequence defining oriC. Formation of this 45-bp “open complex” by DnaA protein is ATP-dependent. Next, DnaB protein (a hexamer of 50-kD subunits) binds to the “open” oriC. DnaB protein is delivered to oriC by DnaC protein (29 kD) in the form of a hexameric (DnaB:DnaC:ATP)₆ complex, but DnaC protein does not enter the protein assemblage at oriC. Delivery of DnaB protein by DnaC protein is assisted by DnaT protein. The addition of DnaB protein completes assembly of the pre-priming complex. ATP hydrolysis drives the formation of this complex (Figure 30.16). DnaB protein has helicase activity and it further unwinds the DNA in the pre-priming complex in both directions, assisted by DNA gyrase. SSB tetramers coat single-stranded regions as they arise. Unwinding exposes the base sequence of the strands so that RNA primers can be synthesized by primase, and the strands can be read as templates by the replicative polymerase.

Figure 30.17 × A representation of replisome assembly at each replication fork. (a) The “pre-priming complex” of Figure 30.16 is the starting point. (b) Protein PriA is loaded onto the primosome by the PriB and PriC proteins. Once PriA has joined DnaB, primase follows. This completes assembly of the primosome. An RNA primer is synthesized at each replication fork by primase, (c) one for the leading strand and (d) one for the lagging strand. Two DNA polymerase III holoenzyme complexes (e) at each replication fork carry out DNA elongation, one synthesizing the leading strand and the other the lagging strand. Primase association with DnaB is transient; once a primer has been synthesized, primase dissociates until another round of primer synthesis is necessary.

        The addition of primase to the pre-priming complex completes the formation of the primosome, a protein machine of about 700 kD that synthesizes the RNA primer essential to DNA synthesis (Figure 30.17a-e). Interestingly, the AT-rich tandem repeats of oriC resemble sequences with which RNA polymerase interacts (see Chapter 31). Note that two primosomes form at oriC, one for each replication fork. Once the primosome has primed leading strand synthesis, it remains associated with the lagging strand and periodically primes Okazaki fragment synthesis, as required.

Elongation

The double helix must be unwound ahead of the advancing replication fork. Unwinding is driven by ATP hydrolysis; DnaB protein has the helicase activity associated with the primosome. DnaB moves in the 5' ® 3' direction along the leading strand template, hydrolyzing 2 ATPs for each base pair that it separates. Protein-protein interaction between a t subunit of the DNA polymerase III holoenzyme and DnaB is essential for rapid replication fork progression. PriA protein is, like DnaB, a helicase associated with the primosome, but PriA moves 3' ® 5' along the lagging strand. The binding of SSB protein to the single-stranded regions created behind these proteins prevents re-annealing.
Figure 30.18 × The asymmetric dimer model for E. coli DNA replication. Two helicases (Rep protein and helicase II) act in concert to unwind the DNA. The replisome contains two DNA “core” polymerase III complexes that remain physically associated even though their tasks are different: one performs leading strand synthesis while the other synthesizes Okazaki fragments, the precursors to the lagging strand. Note that (a) the lagging strand template must loop around so that the “core” pol III replicating it can move along it in the 3' ® 5' direction. As seen in (b), the lagging strand “core” pol III must release the template when it comes up against the end of the Okazaki fragment it previously synthesized. In (c), the lagging strand “core” pol III then shifts to a position farther along the template in the 3' direction to resume a new round of Okazaki fragment synthesis at the next RNA primer. Looping of the lagging strand template, the so-called trombone model, allows these events to happen. One of the two “core” pol IIIs is suited to continuous DNA synthesis through its contacts with the helicase, whereas the other one has properties consistent with its role in discontinuous DNA synthesis.

        Figure 30.18 depicts a model for the replication of E. coli DNA as it occurs at one of the replication forks. Two “core” DNA polymerase III units are present. One is synthesizing the leading strand, using the parental 3' ® 5' strand as template. The other synthesizes the lagging strand using the 5' ® 3' parental strand as template. The lagging strand template must be looped out so that it can be copied in the 3' ® 5' direction. The ”core” pol III on this lagging strand has completed the synthesis of an RNA-primed Okazaki fragment when it encounters the 5'-end of the previous fragment. Synthesis of the next RNA primer by the primosome triggers the g complex to disassemble the b₂-sliding clamp, releasing the lagging strand from the DNA polymerase III holoenzyme. The g complex assembles a new b₂-sliding clamp at the 3'-OH end of the next RNA primer, re-attaching the lagging strand “core” polymerase III to begin a new round of lagging strand synthesis.

Termination

Located diametrically opposite from oriC on the E. coli circular map is a terminus region, the Ter, or t, locus. The bidirectionally moving replication forks meet here and replication is terminated. The Ter region contains a number of short DNA sequences containing a consensus core element 5'-GTGTGTTGT. These Ter sequences act as terminators; clusters of three or four Ter sequences are organized into two sets inversely oriented with respect to one another. One set blocks the clockwise-moving replication fork, and its inverted counterpart blocks the counterclockwise-moving replication fork. A Ter sequence element will impede replication fork progression only if oriented in the proper direction with respect to the approaching replication fork and then only if a specific 36-kD replication termination protein, Tus protein, is bound to it. Tus protein is a contrahelicase. That is, Tus protein prevents the DNA duplex from unwinding by blocking progression of the replication fork and inhibiting the ATP-dependent DnaB helicase activity. Mutations in either the Ter locus or the gene encoding Tus protein do not grossly affect DNA replication, demonstrating that this termination mechanism is not essential.
        Final synthesis of both duplexes is completed. Replication usually leaves the circular progeny chromosomes intertwined by 20 to 30 coils about each other, a so-called catenated state. In order to disengage the individual duplexes from each other prior to their distribution to daughter cells, double-stranded cuts must be made so that the double helices can pass through one another. Topoisomerase II (DNA gyrase) can catalyze this process.

30.5 · Eukaryotic DNA Replication

Figure 30.19 × The eukaryotic cell cycle. The stages of mitosis and cell division define the M phase (“M” for mitosis). G₁ (“G” for gap, not growth) is typically the longest part of the cell cycle; G₁ is characterized by rapid growth and metabolic activity. Cells that are quiescent, that is, not growing and dividing (such as neurons), are said to be in G₀. The S phase is the time of DNA synthesis. S is followed by G₂, a relatively short period of growth when the cell prepares for cell division. Cell cycle times vary from less than 24 hours (rapidly dividing cells such as the epithelial cells lining the mouth and gut) to hundreds of days. catenated ^ connected in a series of links, as in a chain

The mechanism of DNA replication in eukaryotic cells shows strong parallels with prokaryotic DNA replication, but the situation in eukaryotes is vastly more complex. For example, in a growing human cell, some 6 billion bp of DNA must be duplicated with high fidelity once (and only once) each cell cycle. The events associated with cell growth and division in eukaryotic cells fall into a general sequence having four distinct phases, M, G₁, S, and G₂ (Figure 30.19). Eukaryotic cells have solved the problem of replicating their enormous genomes in the few hours allotted to the S phase by initiating DNA replication at multiple origins of replication distributed along each chromosome. Depending on the organism and cell type, replication origins, also called replicators, occur every 3 to 300 kbp (for example, an average human chromosome has several hundred replication origins). In lower eukaryotes such as yeast, replicators are small discrete chromosomal regions (100-200 bp), but in mammalian chromosomes, the zones where initiation of DNA replication occurs may span 500 to 50,000 bp. So that eukaryotic DNA replication can proceed concomitantly throughout the genome, each eukaryotic chromosome contains many units of replication, so-called replicons.

Figure 30.20 × Model for initiation of the DNA replication cycle in eukaryotes. ORC is present at the replicators throughout the cell cycle. The pre-replication complex (pre-RC) is assembled through the sequential addition of Cdc6p and MCM (a replication licensing factor) during a window of opportunity defined by the state of cyclin-CDK and Cdc7p-Dbf4p. After initiation, a post-RC state is established. (Adapted from Figure 2 in Stillman, B., 1996. Cell cycle control of DNA replication. Science 274:165-1663)

Cell Cycle Control of DNA Replication

Initiation of replication involves replicators and the origin recognition complex, or ORC, a heteromeric protein that binds to replicators. ORC is bound to the replicators throughout the cell cycle. Early in G₁ (just after M), ORC serves as a “landing pad” for proteins essential to replication control. Proteins binding to ORC establish a pre-replication complex (pre-RC), but the pre-RC can be formed only during a window of opportunity during G₁. One of the principal proteins in assembly of the pre-RC in yeast is Cdc6p (the replication activator protein encoded by the yeast cdc6 gene) (Figure 30.20). Once Cdc6p binds to ORC, replication licensing factors (RLFs) that “license” or permit DNA replication to occur then bind to the chromosomes. The RLFs render the chromosomes competent for DNA replication. Two RLFs are required: RLF-B and RLF-M. It has been suggested that RLF-B is confined to the cytosol and has access to the chromosomes only when the nuclear envelope disappears early in mitosis. Apparently, RLF-B associates with ORC on the chromosomes and is present at the beginning of G₁. RLF-M is a heteromeric complex of the MCM proteins (the name comes from “mini-chromosome maintenance” because these proteins are essential in the maintenance of plasmids [mini-chromosomes] in yeast cells). These various protein-protein interactions establish the pre-RC, which consists of ORC, Cdc6p, the MCM complex, and other proteins.
            At this point, two protein kinases act upon the pre-RC to directly trigger DNA replication. One of these protein kinases is a complex of cyclin-dependent protein kinase (CDK) and cyclin B, called cyclin B-CDK. (Cyclins are proteins synthesized at one phase of the cell cycle and degraded at another; thus, the concentration of cyclins cycles during the cell cycle. B-Cyclins accumulate at high levels just before S phase.) Cyclin B-CDK can phosphorylate sites in ORC, Cdc6p, and several MCM subunits. Phosphorylation of Cdc6p causes it to dissociate from ORC, whereupon it is degraded. Some of the MCM also dissociates. Cyclin B-CDK also phosphorylates Cdc7p-Dbf4p, the other protein kinase essential to activation of DNA replication. Cdc7p-Dbf4p is a complex of a protein kinase encoded by the yeast cdc7 gene (Cdc7p) and the product of the yeast dbf4 gene (Dbf4p) (Figure 30.20). One model suggests that Cdc7p interacts with ORC and Dbf4p interacts with the replicator; together, Cdc7p-Dbf4p phosphorylates the MCM complex. The consequence of these actions brings the cell into S phase.
            These phosphorylation events serve as a replication switch because once proteins in the pre-RC are phosphorylated (and perhaps destroyed, as Cdc6p is), the post-RC state is achieved. The post-RC state is incapable of re-initiating DNA replication. This transformation ensures that eukaryotic DNA replication occurs once, and only once, per cell cycle.

Eukaryotic DNA Polymerases

Figure 30.21 × Structure of the PCNA homotrimer. Note that the trimeric PCNA ring of eukaryotes is remarkably similar to its prokaryotic counterpart, the dimeric b₂ sliding clamp (Figure 30.13). (a) Ribbon representation of the PCNA trimer with an axial view of a B-form DNA duplex in its center. (b) Molecular surface of the PCNA trimer with each monomer colored differently. The red spiral represents the sugar-phosphate backbone of a strand of B-form DNA. (Adapted from Figure 3 in Krishna, T. S., et al., 1994. Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell 79:1233-1243. Photos courtesy of John Kuriyan, Rockefeller University )

A number of different DNA polymerases have been described in animal cells and assigned Greek letters in the order of their discovery (Table 30.4). DNA polymerase a, a complex multimeric protein, functions in the initiation of nuclear DNA replication. Given a template, it not only synthesizes an RNA primer, but it then adds deoxynucleotides to extend the chain in the 5' ® 3' direction. It consists of four subunits: a 180-kD DNA polymerase, two associated primase subunits (60 kD and 50 kD), and another 70-kD subunit of uncertain function. DNA polymerase a lacks 3' ® 5' exonuclease activity. The processivity of DNA polymerase a is a modest 200 nucleotides or so. DNA polymerase d is the principal DNA polymerase in eukaryotic DNA replication. It consists of a 125-kD catalytic subunit and a 50-kD subunit that interacts with PCNA protein (for proliferating cell nuclear antigen). In association with PCNA, DNA polymerase d carries out highly processive DNA synthesis. PCNA is the eukaryotic counterpart of the E. coli b₂ sliding clamp; it clamps DNA polymerase d to the DNA. Like b₂, PCNA encircles the double helix, but, in contrast to the prokaryotic b₂ sliding clamp, PCNA is a homotrimer of 37-kD subunits (Figure 30.21). DNA polymerase d has 3' ® 5' exonuclease activity. DNA polymerase e also plays a major role in DNA replication; its precise role is unclear, but it may sometimes substitute for DNA polymerase d in lagging strand synthesis. DNA polymerase b functions in DNA repair; DNA polymerase g is the DNA-replicating enzyme of mitochondria.

Reactions at the Eukaryotic DNA Replication Fork

Studies on animal viruses such as SV40 (SV for simian virus) have provided a useful model for eukaryotic DNA replication. The SV40 genome (a 5-kbp, closed circular DNA duplex with a single origin of replication) can be considered to represent a single replicon. The virus makes extensive use of the host cell’s replication machinery; only a single virus replication protein, T antigen, is viral-encoded. The SV40 origin of replication is a 64-bp sequence consisting of at least three functionally distinct domains: (a) a central set of four copies of the pentameric sequence GAGGC organized as an inverted repeat. This sequence is the T antigen binding site. On one side of the T antigen binding domain is (b) a 17-bp region composed exclusively of A and T residues; on the other side is (c) a 15-bp imperfect inverted repeat. One T antigen molecule binds to each of the four pentameric GAGGC repeats. The T antigen is an ATP-dependent helicase, and once it forms a complex with the four pentameric repeats, it enters the DNA duplex at the adjacent AT region and unwinds the helix. The AT region, with only two H bonds per base pair, is susceptible to strand separation. Helix unwinding creates two replication forks. Unwinding is facilitated by replication protein A (RPA), a ssDNA-binding host cell protein that is the eukaryotic counterpart of SSB. Viral DNA replication is bidirectional, and two nascent chains are synthesized at each replication fork: a leading strand and a lagging strand.

Figure 30.22 × Model for eukaryotic DNA replication illustrating polymerase switching and Okazaki fragment processing on the lagging strand. (a) Leading strand synthesis is initiated by primase-DNA polymerase a, and then RFC mediates polymerase switching (removal of DNA polymerase a, assembly of PCNA, and addition of DNA polymerase d)—not shown. DNA helicase unwinds the DNA at the replication fork. On the lagging strand, DNA polymerase a initiates synthesis through formation of an RNA primer (shown in red), followed by a short deoxynucleotide oligomer. Then, RFC loads PCNA and a second DNA polymerase (d or e) and synthesis of the Okazaki fragment continues. (b) When the DNA polymerase approaches the RNA primer of the downstream Okazaki fragment, RNase H1 degrades this RNA by removing all but one of its ribonucleotides; FEN1/RTH1 removes this last one, and DNA ligase joins the two Okazaki fragments. (Adapted from Figure 1 in Bambara, R. A., Murante, R. S., and Henricksen, L. A., 1997. Enzymes and reactions at the eukaryotic replication fork. Journal of Biological Chemistry 272:4647-4650)

Eukaryotic DNA Replication: Leading Strand Synthesis

Synthesis of the leading strand is initiated upon RNA primer synthesis by the primase subunits of DNA polymerase a; then DNA polymerase a adds a stretch of DNA to the primer (Figure 30.22). At this point, replication factor C (RFC), the eukaryotic counterpart of the prokaryotic g complex, carries out a process called polymerase switching: RFC removes DNA polymerase a and assembles PCNA in the region of the primer strand terminus. Then, DNA polymerase d binds to PCNA and carries out highly processive leading strand synthesis.

Eukaryotic DNA Replication: Lagging Strand Synthesis

Lagging strand synthesis of Okazaki fragments is initiated in essentially the same way as leading strand synthesis: synthesis is primed by DNA polymerase a, a deoxynucleotide stretch is added by DNA polymerase a, and then switching to DNA polymerase d takes place. Priming is a frequent event in lagging strand synthesis, with RNA primers placed every 50 or so nucleotides. About 10-nucleotide lengths of RNA primer are extended through addition of 10 to 20 deoxynucleotides by DNA polymerase a before DNA polymerase d (or e) enters (Figure 30.22). All but one of the ribonucleotides in the RNA primer are removed by RNase H1; then the exonuclease activity of the FEN1/RTH1 complex removes the one remaining ribonucleotide, and DNA ligase joins the Okazaki fragment to the growing DNA strand.

T Antigen Function Is Regulated by Protein Phosphorylation

Phosphorylation at a single residue (Thr¹²⁴) activates T antigen to bind at the SV40 ori and initiate replication. This phosphorylation is catalyzed by a cyclin-dependent protein kinase, CDC2-PK. Such kinases are implicated in triggering eukaryotic cells to enter mitosis. Dephosphorylation at different amino acid phosphorylation sites on T antigen also activates it. RPC, a protein involved in the early stages of SV40 replication, is a protein phosphatase subunit. The combined actions of CDC2-PK and RPC generate an appropriately phosphorylated form of T antigen competent in replication initiation. These various phosphorylations/dephosphorylations represent paradigms for the central role of protein phosphorylation in regulating eukaryotic DNA replication and cell division.

30.6 · Replicating the Ends of Chromosomes—Telomeres and Telomerase

Figure 30.23 × (a) In replication of the lagging strand, short RNA primers are added (pink) and extended by DNA polymerase. When the RNA primer at the 5'-end of each strand is removed, there is no nucleotide sequence to read in the next round of DNA replication. The result is a gap (primer gap) at the 5'-end of each strand (only one end of a chromosome is shown in this figure). (b) Asterisks indicate sequences at the 3'-end that cannot be copied by conventional DNA replication. Synthesis of telomeric DNA by telomerase extends the 5'-ends of DNA strands, allowing the strands to be copied by normal DNA replication.

Telomeres are short (5-8 bp), tandemly repeated, G-rich nucleotide sequences that form protective caps 1-12 kbp long on the ends of chromosomes (see Chapter 12). Vertebrate telomeres have a TTAGGG consensus sequence. Telomeres are necessary for chromosome maintenance and stability. DNA polymerases cannot replicate the extreme 5'-ends of chromosomes because these enzymes require a template and a primer and replicate only in the 5' ® 3' direction. Thus, lagging strand synthesis at the 3'-ends of chromosomes is primed by RNA primase to form Okazaki fragments, but these RNA primers are subsequently removed, resulting in gaps (“primer gap”—Figure 30.23) in the progeny 5'-terminal strands at each end of the chromosome. Telomerase, an RNA-dependent DNA polymerase (126 kD) whose catalytic subunit showssubstantial homology to other reverse transcriptases (see

Figure 30.24 × The structure of AZT (3'-azido-2',3'-dideoxythymidine). This nucleoside is a major drug in the treatment of AIDS. AZT is phosphorylated in vivo to give AZTTP (AZT 5'-triphosphate), a substrate analog that binds to HIV reverse transcriptase. HIV reverse transcriptase incorporates AZTTP into growing DNA chains in place of dTTP. Incorporated AZTMP blocks further chain elongation because its 3'-azido group cannot form a phosphodiester bond with an incoming nucleotide. Host cell DNA polymerases have little affinity for AZTTP.

Chapter 13 and below), maintains telomere length by restoring telomeres at the 3'-ends of chromosomes. Telomerase is a ribonucleoprotein, and its RNA component contains a 9- to 30-nucleotide-long region that serves as a template for the synthesis of telomeric repeats at DNA ends. The human telomerase RNA component is 450 nucleotides long; its template sequence is CUAACCCUAAC. Telomerase uses the 3'-end of the DNA as a primer and adds successive TTAGGG repeats to it, employing its RNA as template over and over again (Figure 30.24, see also figure in Chapter 12 Human Biochemistry box: Telomeres and Tumors).

30.7 · Reverse Transcriptase: An RNA-Directed DNA Polymerase

The information encoded in DNA can be replicated into new DNA chains by DNA polymerases (this chapter) or transcribed into RNA chains by DNA-dependent RNA polymerases (Chapter 31). Nature has also found a role for the synthesis of DNA chains from an RNA template. In 1964, Howard Temin noted that inhibitors of DNA synthesis prevented infection of cells in culture by RNA tumor viruses such as avian sarcoma virus. On the basis of this observation, Temin made the bold proposal that DNA is an intermediate in the replication of such viruses; that is, an RNA tumor virus can use viral RNA as the template for DNA synthesis.

            RNA viral chromosome ® DNA intermediate ® RNA viral chromosome

In 1970, Temin and David Baltimore independently discovered a viral enzyme capable of mediating such a process, namely, an RNA-directed DNA polymerase or, as it is usually called, reverse transcriptase. All RNA tumor viruses contain such an enzyme within their virions (viral particles), so they are now classified as retroviruses. Note that not all retroviruses are tumor viruses; HIV (human immunodeficiency virus) is a retrovirus that causes AIDS.
            Like other DNA and RNA polymerases, reverse transcriptase synthesizes polynucleotides in the 5' ® 3' direction, and like all DNA polymerases, reverse transcriptase requires a primer. Interestingly, the primer is a specific tRNA molecule captured by the virion from the host cell in which it was produced. The 3'-end of the tRNA is base-paired with the viral RNA template at the site where DNA synthesis initiates and its free 3'-OH accepts the initial deoxynucleotide once transcription commences. Reverse transcriptase then transcribes the RNA template into a complementary DNA (cDNA) strand to form a double-stranded DNA:RNA hybrid.

The Enzymatic Activities of Reverse Transcriptases

Reverse transcriptases possess three enzymatic activities, all of which are essential to viral replication:

1.  RNA-directed DNA polymerase activity, for which the enzyme is named (see Figure 13.14).
2.  RNase H activity. Recall that RNase H is an exonuclease activity that specifically degrades RNA chains in DNA:RNA hybrids (Figure 13.14). The RNase H function of reverse transcriptase degrades the template genomic RNA and also removes the priming tRNA after DNA synthesis is completed.
3.  DNA-directed DNA polymerase activity.This activity replicates the ssDNA remaining after RNase H degradation of the viral genome, yielding a DNA duplex. This DNA duplex directs the remainder of the viral infection process or it becomes integrated into the host chromosome, where it can lie dormant for many years as a provirus. Activation of the provirus restores the infectious state.

           HIV reverse transcriptase is of great clinical interest as an enzyme required for replication of the AIDS virus. It is a heterodimer of 66- and 51-kD subunits. The two polypeptides have identical N-terminal sequences, indicating that the 51-kD subunit is derived from the 66-kD polypeptide by proteolytic cleavage. The 66-kD polypeptide consists of two domains: an N-terminal polymerase domain and a C-terminal RNase H domain. The N-terminal domain shows significant amino acid sequence homology to bacterial and viral DNA polymerases, whereas the C-terminal half is homologous to bacterial RNase H. The active sites for the polymerase and RNase H are physically separate and functionally distinct. DNA synthesis by HIV reverse transcriptase is blocked by AZT (Figure 30.24). HIV reverse transcriptase is error-prone: It incorporates the wrong base at a frequency of 1 per 2000 to 4000 nucleotides polymerized. This high error rate during replication of the HIV genome means that the virus is ever changing, a feature that confounds attempts to devise a vaccine.

30.8 · DNA Repair

Figure 30.25 × UV irradiation causes dimerization of adjacent thymine bases. A cyclobutyl ring is formed between carbons 5 and 6 of the pyrimidine rings. Normal base pairing is disrupted by the presence of such dimers.

Biological macromolecules are susceptible to chemical alterations that arise from environmental damage or errors during synthesis. For RNAs, proteins, or other cellular molecules, most consequences of such damage are circumvented by replacement of these molecules through normal turnover (synthesis and degradation). However, the integrity of DNA is vital to cell survival and reproduction. Its information content must be protected over the life span of the cell and preserved from generation to generation. Safeguards include (a) high-fidelity replication systems and (b) repair systems that correct DNA damage that might alter its information content. DNA is the only molecule that, if damaged, is repaired by the cell. Such repair is possible because the information content of duplex DNA is inherently redundant. The most common forms of damage are (a) a missing, altered, or incorrect base; (b) bulges due to deletions

Figure 30.26 × Oxygen radicals, in the presence of metal ions such as Fe²⁺, can destroy sugar rings in DNA, breaking the strand.

or insertions; (c) UV-induced pyrimidine dimers (Figure 30.25); (d) strand breaks at phosphodiester bonds or within deoxyribose rings (Figure 30.26); and (e) covalent cross-linking of strands. Cells have extraordinarily diverse and effective DNA repair systems to deal with these problems. When repair fails, the genome may still be preserved if an “error-prone” mode of replication allows the lesion to be bypassed.
            Usually, the complementary structure of duplex DNA ensures that information lost through damage to one strand can be recovered from the other. However, even errors involving both strands can be corrected. For example, deletions or insertions can be repaired by replacing the region through recombination (see Chapter 29). Double-stranded breaks, potentially the most serious lesions, can be repaired by DNA ligases or recombination events.
            Human DNA replication has an error rate of about three base-pair mistakes during copying the 6 billion base pairs in the diploid human genome. The low error rate is due to the DNA repair systems that review and edit the newly replicated DNA. Further, about 10⁴ bases (mostly purines) are lost per cell per day from spontaneous breakdown in human DNA; the repair systems must replace these bases to maintain the fidelity of the encoded information.

Molecular Mechanisms of DNA Repair

Two fundamental types of molecular mechanisms for DNA repair can be distinguished: (a) mechanisms that excise and replace damaged regions by replication, recombination, or mismatch repair, and (b) mechanisms that reverse damaging chemical changes in DNA; the latter includes excision repair systems.

Mismatch Repair

The mismatch repair system corrects errors introduced when DNA is replicated. It scans newly synthesized DNA for mispaired bases, excises the mismatched region, and then replaces it by DNA polymerase-mediated local replication. The key to such replacement is to know which base of the mismatched pair is correct.
            The E. coli methyl-directed pathway of mismatch repair relies on methyl-ation patterns in the DNA to determine which strand is the newly synthesized one and which one was the parental (template) strand. DNA methylation, often an identifying and characteristic feature of a prokaryote’s DNA, occurs just after DNA replication. However, a window of opportunity exists between the start of methylation and the end of replication, when only the parental strand of a dsDNA is methylated. This window in time provides an opportunity for the mismatch repair system to review the dsDNA for mismatched bases that arose as a consequence of replication errors. By definition, the newly synthesized strand is the one containing the error, and the methylated strand is the one having the correct nucleotide sequence. When the methyl-directed mismatch repair system encounters a mismatched base pair, it searches along the DNA, through thousands of base pairs if necessary, until it finds a methylated base.
            The system identifies the strand bearing the methylated base as parental, assumes its sequence is the correct one, and replaces the entire stretch of nucleotides within the nascent strand from this recognition point to and including the mismatched base. Mismatch repair does this by using an endonuclease to cut the new, unmethylated strand and an exonuclease to remove the mismatched bases, creating a gap in the newly synthesized strand. DNA polymerase III holoenzyme then fills in the gap, using the methylated strand as template. Finally, DNA ligase re-seals the strand.

PHOTOREACTIVATION OF PYRIMIDINE DIMERS.  UV irradiation promotes the formation of covalent bonds between adjacent thymine residues in a DNA strand, creating a cyclobutyl ring (Figure 30.25). Because the COC bonds in this ring are shorter than the normal 0.34-nm base stacking in B-DNA, the DNA is distorted at this spot and is no longer a proper template for either replication or transcription. Photolyase (also called photoreactivating enzyme), a flavin- and pterin-dependent enzyme, binds at the dimer and uses the energy of visible light to break the cyclobutyl ring, restoring the pyrimidines to their original form.

Figure 30.27 × Base excision repair. A damaged base () is excised from the sugar-phosphate backbone by DNA

EXCISION REPAIR.  Replacement of many damaged or modified bases occurs via excision repair systems. There are two fundamental excision repair systems—base excision and nucleotide excision. Base excision repair acts on single bases that have been damaged through oxidation or other chemical modification during normal cellular processes. The damaged base is removed by DNA glycosylase, which cleaves the glycosidic bond, creating an AP site where the sugar-phosphate backbone is intact but a purine (apurinic site) or a pyrimidine (apyrimidinic site) is missing, and then the remaining ribose moiety is excised by AP exonuclease. An AP endonuclease then cleaves the backbone, an exonuclease removes the deoxyribose-P and a number of additional residues, and the gap is repaired by DNA polymerase and DNA ligase (Figure 30.27). The information of the complementary strand is used to dictate which bases are added in re-filling the gap.
            Nucleotide excision repair recognizes and repairs larger regions of damaged DNA than base excision repair. The nucleotide excision repair system cuts the sugar-phosphate backbone of a DNA strand in two places, one on each side of the lesion, and removes the region. The region removed in prokaryotic nucleotide excision repair spans 12 or 13 nucleotides; in the eukaryotic counterpart, an oligonucleotide stretch 27 to 29 units long is removed. The resultant gap is then filled in using DNA polymerase (DNA polymerase I in prokaryotes or DNA polymerase d or e + PCNA and RFC in eukaryotes), and the sugar-phosphate backbone is covalently closed by DNA ligase.
            In mammalian cells, nucleotide excision repair is the main pathway for removal of carcinogenic (cancer-causing) lesions caused by sunlight or other mutagenic agents. Such lesions are recognized by XPA protein, named for xeroderma pigmentosum, an inherited human syndrome whose victims suffer serious skin lesions if exposed to sunlight. At sites recognized by XPA, a multiprotein endonuclease is assembled and the damaged strand is cleaved and repaired.

The SOS Response

Figure 30.28 × Nalidixic acid. Quinolone antibiotics such as nalidixic acid are DNA topoisomerase inhibitors. In the presence of one of these compounds, the DNA gyrase A subunits become covalently linked to the 5'-end of the DNA strands.

E. coli DNA lesions that block replication activate the SOS response, a system that converts the lesion to an error-prone site aNnd restores replication. Such lesions include pyrimidine dimers, cross-linked strands, and quinolone antibiotic-induced breaks in DA. Quinolone antibiotics such as nalidixic acid (Figure 30.28) inhibit the nick-ligating activity of bacterial DNA gyrases. The SOS response is activated when the replication fork stalls and RecA protein (Chapter 29) binds to exposed ssDNA (or UV-damaged dsDNA). The RecA protein:DNA complex binds a protein, LexA. The LexA protein has proteolytic activity, and its binding induces a conformational change that causes LexA protein to cleave itself. Because LexA protein blocks expression of many genes encoding a set of proteins mediating error-prone replication, LexA protein’s self-destruction leads to synthesis of these proteins. These proteins then assemble at the lesion and form a mutasome, an error-prone replication apparatus that allows DNA polymerase to replicate past the lesion.