The base sequences of newly synthesized DNA must faithfully match the sequences of parent DNA. To achieve faithful replication, each strand within the parent double helix acts as a template for the synthesis of a new DNA strand with a complementary sequence. The building blocks for the synthesis of the new strands are deoxyribonucleoside triphosphates. They are added, one at a time, to the 3′ end of an existing strand of DNA.

Although this reaction is in principle quite simple, it is significantly complicated by specific features of the DNA double helix. First, the two strands of the double helix run in opposite directions. Because DNA strand synthesis always proceeds in the 5′-to-3′ direction, the DNA replication process must have special mechanisms to accommodate the oppositely directed strands. Second, the two strands of the double helix interact with one another in such a way that the bases, key templates for replication, are on the inside of the helix. Thus, the two strands must be separated from each other so as to generate appropriate templates. Finally, the two strands of the double helix wrap around each other. Thus, strand separation also entails the unwinding of the double helix. This unwinding creates supercoils that must themselves be resolved as replication continues. We begin with a consideration of the chemistry that underlies the formation of the phosphodiester backbone of newly synthesized DNA.

829

DNA polymerases catalyze the formation of polynucleotide chains. Each incoming nucleoside triphosphate first forms an appropriate base pair with a base in the template. Only then does the DNA polymerase link the incoming base with the predecessor in the chain. Thus, DNA polymerases are template-directed enzymes.

DNA polymerases add nucleotides to the 3′ end of a polynucleotide chain. The polymerase catalyzes the nucleophilic attack by the 3′-hydroxyl-group terminus of the polynucleotide chain on the α phosphoryl group of the nucleoside triphosphate to be added (Figure 4.25). To initiate this reaction, DNA polymerases require a primer with a free 3′-hydroxyl group already basepaired to the template. They cannot start from scratch by adding nucleotides to a free single-stranded DNA template. RNA polymerase, in contrast, can initiate RNA synthesis without a primer, as we shall see in Chapter 29.

The three-dimensional structures of a number of DNA polymerase enzymes are known. The first such structure was elucidated by Tom Steitz and coworkers, who determined the structure of the so-called Klenow fragment of DNA polymerase I from E. coli (Figure 28.4). This fragment comprises two main parts of the full enzyme, including the polymerase unit. This unit approximates the shape of a right hand with domains that are referred to as the fingers, the thumb, and the palm. In addition to the polymerase, the Klenow fragment includes a domain with 3′ → 5′ exonuclease activity that participates in proofreading and correcting the polynucleotide product.

DNA polymerases are remarkably similar in overall shape, although they differ substantially in detail. At least five structural classes have been identified; some of them are clearly homologous, whereas others appear to be the products of convergent evolution. In all cases, the finger and thumb domains wrap around DNA and hold it across the enzyme’s active site, which comprises residues primarily from the palm domain. Furthermore, all DNA polymerases use similar strategies to catalyze the polymerase reaction, making use of a mechanism in which two metal ions take part.

Like all enzymes with nucleoside triphosphate substrates, DNA polymerases require metal ions for activity. Examination of the structures of DNA polymerases with bound substrates and substrate analogs reveals the presence of two metal ions in the active site. One metal ion binds both the deoxynucleoside triphosphate (dNTP) and the 3′-hydroxyl group of the primer, whereas the other interacts only with the dNTP (Figure 28.5). The two metal ions are bridged by the carboxylate groups of two aspartate residues in the palm domain of the polymerase. These side chains hold the metal ions in the proper positions and orientations. The metal ion bound to the primer activates the 3′-hydroxyl group of the primer, facilitating its attack on the α phosphoryl group of the dNTP substrate in the active site. The two metal ions together help stabilize the negative charge that accumulates on the pentacoordinate transition state. The metal ion initially bound to dNTP stabilizes the negative charge on the pyrophosphate product.

830

DNA must be replicated with high fidelity. Each base added to the growing chain should, with high probability, be the Watson–Crick complement of the base in the corresponding position in the template strand. The binding of the dNTP containing the proper base is favored by the formation of a base pair with its partner on the template strand. Although hydrogen bonding contributes to the formation of this base pair, overall shape complementarity is crucial. Studies show that a nucleotide with a base that is very similar in shape to adenine but lacks the ability to form base-pairing hydrogen bonds can still direct the incorporation of thymidine, both in vitro and in vivo (Figure 28.6).

An examination of the crystal structures of various DNA polymerases reveals why shape complementarity is so important. First, residues of the enzyme form hydrogen bonds with the minor-groove side of the base pair in the active site (Figure 28.7). In the minor groove, hydrogen-bond acceptors are present in the same positions for all Watson–Crick base pairs. These interactions act as a “ruler” that measures whether a properly spaced base pair has formed in the active site.

Second, DNA polymerases close down around the incoming dNTP (Figure 28.8). The binding of a deoxyribonucleoside triphosphate into the active site of a DNA polymerase triggers a conformational change: the finger domain rotates to form a tight pocket into which only a properly shaped base pair will readily fit. Many of the residues lining this pocket are important to ensure the efficiency and fidelity of DNA synthesis. For example, mutation of a conserved tyrosine residue that forms part of the pocket results in a polymerase that is approximately 40 times as error prone as the parent polymerase.

831

DNA polymerases cannot initiate DNA synthesis without a primer, a section of nucleic acid having a free 3′ end that forms a double helix with the template. How is this primer formed? An important clue came from the observation that RNA synthesis is essential for the initiation of DNA synthesis. In fact, RNA primes the synthesis of DNA. An RNA polymerase called primase synthesizes a short stretch of RNA (about five nucleotides) that is complementary to one of the template DNA strands (Figure 28.9). Primase, like other RNA polymerases, can initiate synthesis without a primer. After DNA synthesis has been initiated, the short stretch of RNA is removed by hydrolysis and replaced by DNA.

Both strands of parental DNA serve as templates for the synthesis of new DNA. The site of DNA synthesis is called the replication fork because the complex formed by the newly synthesized daughter helices arising from the parental duplex resembles a two-pronged fork. Recall that the two strands are antiparallel; that is, they run in opposite directions. During DNA replication, both daughter strands appear on cursory examination to grow in the same direction. However, all known DNA polymerases synthesize DNA in the 5′ → 3′ direction but not in the 3′ → 5′ direction. How then does one of the daughter DNA strands appear to grow in the 3′ → 5′ direction?

This dilemma was resolved by Reiji Okazaki, who found that a significant proportion of newly synthesized DNA exists as small fragments. These units of about a thousand nucleotides (called Okazaki fragments) are present briefly in the vicinity of the replication fork (Figure 28.10).

832

As replication proceeds, these fragments become covalently joined through the action of the enzyme DNA ligase to form a continuous daughter strand. The other new strand is synthesized continuously. The strand formed from Okazaki fragments is termed the lagging strand, whereas the one synthesized without interruption is the leading strand. The discontinuous assembly of the lagging strand enables 5′ → 3′ polymerization at the nucleotide level to give rise to overall growth in the 3′ → 5′ direction.

The joining of Okazaki fragments requires an enzyme that catalyzes the joining of the ends of two DNA chains. The existence of circular DNA molecules also points to the existence of such an enzyme. In 1967, scientists in several laboratories simultaneously discovered DNA ligase. This enzyme catalyzes the formation of a phosphodiester linkage between the 3′-hydroxyl group at the end of one DNA chain and the 5′-phosphoryl group at the end of the other (Figure 28.11). An energy source is required to drive this thermodynamically uphill reaction. In eukaryotes and archaea, ATP is the energy source. In bacteria, NAD⁺ typically plays this role.

DNA ligase cannot link two molecules of single-stranded DNA or circularize single-stranded DNA. Rather, ligase seals breaks in double-stranded DNA molecules. The enzyme from E. coli ordinarily forms a phosphodiester bridge only if there are at least a few bases of single-stranded DNA on the end of a double-stranded fragment that can come together with those on another fragment to form base pairs. Ligase encoded by T4 bacteriophage can link two blunt-ended double-helical fragments, a capability that is exploited in recombinant DNA technology.

For a double-stranded DNA molecule to replicate, the two strands of the double helix must be separated from each other, at least locally. This separation allows each strand to act as a template on which a new polynucleotide chain can be assembled. Specific enzymes, termed helicases, utilize the energy of ATP hydrolysis to power strand separation.

Helicases are a large and diverse family of enzymes taking part in many biological processes. The helicases in DNA replication are typically oligomers containing six subunits that form a ring structure. The structure of one such helicase, that from bacteriophage T7, has been a source of considerable insight into the helicase mechanism (Figure 28.12). Each of the subunits within this hexameric structure has a core structure that includes a P-loop NTPase domain (Figure 9.49). In addition to the P-loop, each subunit has two loops that extend toward the center of the ring structure and interact with DNA. Each subunit interacts closely with its two neighbors within the ring structure. Closer examination of this structure reveals that the ring deviates significantly from six-fold symmetry. This deviation is even more apparent when helicase is crystallized in the presence of the nonhydrolyzable ATP analog AMP-PNP.

833

The AMP-PNP binds to only four of the six subunits within the ring (Figure 28.13). Furthermore, the four nucleotide-binding sites are not identical but fall into two classes. One class appears to be well positioned to bind ATP but not catalyze its hydrolysis, whereas the other class is more well suited to catalyze the hydrolysis but not release the hydrolysis products. The classes are analogous to myosin’s two different conformations—one for binding ATP and one for hydrolyzing it (Section 9.4). Finally, the six subunits fall into three classes with regard to their orientation with respect to the overall ring structure, with differences in rotation around an axis in the plane of the ring of approximately 30°. These differences in orientation affect the position of the two DNA-binding loops in each subunit.

These observations are consistent with the following mechanism for the helicase (Figure 28.14). Only a single strand of DNA can fit through the center on the ring. This single strand binds to loops on two adjacent subunits, one of which has bound ATP and the other of which has bound ADP + P_i. The binding of ATP to the domains that initially had no bound nucleotides leads to a conformational change within the entire hexamer, leading to the release of ADP + P_i from two subunits and the binding of the single-stranded DNA by one of the domains that just bound ATP. This conformational change pulls the DNA through the center of the hexamer. The protein acts as a wedge, forcing the two strands of the double helix apart. This cycle then repeats itself, moving two bases along the DNA strand with each cycle.