11.5 TERMINATION OF DNA REPLICATION

In both bacteria and eukaryotes, replication forks meet head-on when replication is complete. E. coli cells have a specialized mechanism for preventing head-on collisions between DNA polymerases, and between DNA polymerase and RNA polymerase that is transcribing the DNA. Eukaryotic cells have the additional problem of replicating the ends of linear chromosomes. Evolution has provided a solution to this problem, in the form of a novel DNA polymerase—called telomerase—that is specialized for this purpose.

E. coli Chromosome Replication Terminates Opposite the Origin

In E. coli, a region located halfway around the chromosome from oriC contains two clusters of 23 bp sequences called Ter sites (termination sites; Figure 11-33). The two clusters of Ter sites are oriented in opposite directions. The monomeric Tus protein (termination utilization substance) binds tightly to a Ter site and blocks the advance of the replication fork by stopping DnaB helicase. A fascinating property of the Tus-Ter complex is that its fork-blocking activity is polar. Replication forks are blocked when approaching a Tus-Ter complex from one direction (the nonpermissive direction), but not when approaching from the opposite (permissive) direction.

Figure 11-33: The role of Ter sites in control of replication termination in E. coli. (a) Ter sites are located in two clusters, halfway around the chromosome from oriC. Each cluster contains multiple Ter sites oriented in the same direction, but the two clusters have opposite polarity, indicated by the arrows. Tus protein binds to a Ter site, and Tus-Ter blocks helicase approaching from one direction and not the other. Replication forks can displace Tus and pass through the first Ter cluster they encounter, but they are blocked at the second cluster, which has the opposite polarity. (b) Replication forks of equal speed meet in the terminus region of the chromosome (left). The Tus-Ter system ensures that even replication forks moving at unequal speed will meet in the terminus region (right). (c) The Tus-Ter system prevents a replication fork from extending much beyond the halfway point around the chromosome, and thus ensures that the fork always moves in the same direction as transcription.

The arrangement and orientation of Ter sites is such that bidirectional replication forks from oriC can pass through the first set of Ter sites that they encounter, but are blocked by the second set. This arrangement localizes the replication fork collision zone to the area between the two clusters of Ter sites. Although Tus is not essential for E. coli growth, the Tus-Ter system presumably evolved to confer a growth advantage in the natural (nonlaboratory) setting.

Actively replicating bacteria are also growing and metabolizing, and therefore are actively transcribing RNA from promoters throughout the chromosome. This means that collisions between RNA polymerase and replication forks are inevitable. In vivo studies show that codirectional collisions do not impede forks, whereas head-on collisions can cause a fork to pause or stall. Most transcripts in bacteria are oriented in the same direction as replication, and therefore most collisions are codirectional, provided that the forks do not proceed more than halfway around the chromosome (see Figure 11-33). Perhaps the Tus-Ter system evolved to prevent replication forks from going too far around the circular chromosome, where the direction of transcription would result in head-on collisions.

399

Replicating the last bit of DNA between converging replication forks presents certain topological problems that must be solved to disentangle the two daughter chromosomes. Specialized type II topoisomerases unlink the catenated daughter chromosomes (see Chapter 9).

Telomeres and Telomerase Solve the End Replication Problem in Eukaryotes

The replication of linear chromosome ends poses a unique problem. At the end of a chromosome, after the leading strand has been completely extended to the last nucleotide, the lagging strand has a single-strand DNA gap that must be primed and filled in. The problem arises when the RNA primer at the extreme end is removed for replacement with DNA (Figure 11-34). There is no 3′ terminus for DNA polymerase to extend from, so this single-strand gap cannot be converted to duplex DNA. The genetic information in the gap will be lost in the next round of replication, and repeated rounds will cause the ends to progressively shorten until the genes near the ends are entirely lost. The resulting loss of gene function can be detrimental to a cell, disrupting cellular functions or contributing to the formation of cancerous tumors. This end replication problem does not occur in circular DNA, which has no ends. Indeed, avoidance of the end replication problem may underlie the widespread occurrence of circular DNAs in bacteria and their plasmids and phages.

Figure 11-34: The end replication problem in linear chromosomes. Just two rounds of chromosome replication are shown here. In the first generation of replication (left; red indicates new DNA), lagging-strand synthesis results in an RNA primer at or near one 5′ end of each new chromosome. After removal of the RNA, the 5′ single-strand gap in the DNA cannot be filled, thus a 3′ single-stranded DNA overhang remains. In the second generation (right; yellow indicates new DNA), each first-generation chromosome produces two new chromosomes, for a total of four new duplexes. Two of the new chromosomes have lost DNA at one end. All four chromosomes terminate with a 5′ single-strand gap in the DNA after the RNA is removed. Further losses will be sustained with each new generation.

The ends of the linear eukaryotic chromosomes are called telomeres (see Chapter 12). How were telomeres discovered, and what is so special about them? This story starts in the 1930s with Barbara McClintock’s research on maize. McClintock noticed that when chromosomes become damaged and break in two, the broken chromosomes can rapidly “heal” by joining the two chromosome ends. When parts of two different chromosomes become joined, it sometimes creates an unusual chromosome with two centromeres. On entering cell division, chromosomes with two centromeres are pulled to opposite poles by the mitotic spindles, breaking the chromosomes, and this is followed by another round of healing. The process of healing and breaking occurs over and over as cells divide, leading to chromosome rearrangements, loss of genetic information, and cell death. It would, of course, be disastrous if the ends (telomeres) of normal chromosomes underwent this “healing” process, because all the chromosomes would break on entering cell division. Therefore, McClintock surmised, there must be something special about the ends of normal chromosomes, preventing them from being joined together by the healing process that mends broken chromosomes. Independently, Hermann Mueller, working with irradiated flies, made similar observations on the fate of broken chromosomes (in this case, broken by x rays) and hypothesized that the ends of chromosomes have a protective quality about them. He called the ends “telomeres,” derived from the Greek telos (“end”) and meros (“part”).

400

Elizabeth Blackburn
Carol Greider

In 1978, Elizabeth Blackburn identified the first telomere sequence in the single-celled ciliate Tetrahymena, which contains thousands of linear chromosomes and thus became the model organism to study telomere biology. In fact, further study has shown that the telomerase enzyme functions in broadly similar ways in all eukaryotes. Tetrahymena telomeres are composed of numerous direct repeats of TTGGGG. We now know that the telomeres of different organisms consist of tracts of similar short repeats that vary from a few hundred base pairs in single-celled eukaryotes to over 10,000 bp (10 kbp) in higher eukaryotes (see Table 9-1). The sequence TTAGGG is repeated for a length of 5 to 15 kbp in human telomeres. The significance of the telomere sequence was demonstrated in a key experiment conducted jointly by Blackburn and Jack Szostak. Szostak had found that linearized plasmids, when transformed into yeast, were either degraded or integrated into the genome. But after ligation of the Tetrahymena telomere sequence to a linearized plasmid, the linear DNA was maintained in the cell like a very small chromosome. Interestingly, the telomeres grew longer as the cells divided in log-phase growth, and sequencing of the plasmids revealed a different (i.e., not Tetrahymena) repeat. This repeat turned out to be the telomere sequence of yeast! This had enormous implications for how telomeres are formed. The popular model of telomere synthesis involved recombination (a process discussed in Chapter 13). But Blackburn and Szostak’s finding that the added telomere sequence was different from the Tetrahymena sequence implied an enzymatic mechanism, since recombination would have yielded repeats of the same original sequence.

The finding that yeast telomere sequences were added to the Tetrahymena telomere sequence suggested that telomeres may be synthesized by a polymerase that can synthesize DNA from an exogenous template, presumably an RNA molecule. DNA polymerases that use RNA as a template were first found in certain viruses and are called reverse transcriptases, because they transcribe RNA into DNA—the opposite of the usual process (see Chapter 14). Carol Greider, a graduate student in Blackburn’s laboratory, set out to identify the putative “telomere reverse transcriptase.” Using a small synthetic oligonucleotide and radioactive dNTPs, Greider looked for and found an end-extending activity in extracts from Tetrahymena cells. The enzyme added increments of six residues onto the oligonucleotide, as illustrated in the sequencing gel shown in Figure 11-35.

Figure 11-35: Assay for Tetrahymena telomerase activity. (a) In the assay, an oligonucleotide of the Tetrahymena telomere sequence is added to the extract, along with [32P]dGTP and unlabeled dTTP. Telomerase activity adds telomere repeats to the oligonucleotide. (b) Autoradiograph of a polyacrylamide gel analysis of reaction assays using different oligonucleotide sequences as substrate. The result shows that oligonucleotides of Tetrahymena and yeast telomere sequences are extended in six-nucleotide increments.

After purification and characterization, the telomerase reverse transcriptase (TERT) was found to carry a tightly bound, noncoding telomerase RNA (TR). The TR in humans is 451 nucleotides long, but can range from 150 to 1,300 nucleotides, depending on the organism. The TR contains about 1.5 telomere repeat units that it uses as a template to extend the 3′ terminus of the telomere (Figure 11-36). The TERT-TR holoenzyme is referred to as telomerase. The reaction cycle of Tetrahymena telomerase is shown in Figure 11-36. The reaction occurs in S phase, as part of the chromosome duplication process. First, at the 3′-terminal end of a linear DNA, three nucleotides of the telomere anneal to three RNA nucleotides in telomerase. Then the telomerase extends the 3′ end of the single-stranded DNA by the length of one telomere repeat—six nucleotides in the case of Tetrahymena. Telomerase is very different from other DNA polymerases; it carries its own template, and it synthesizes single-stranded DNA. After adding a six-nucleotide repeat, telomerase separates the RNA-DNA hybrid and repositions on the telomere for extension of the next 6-mer repeat. Telomerase acts processively, synthesizing many telomere repeats in one telomerase-binding event. The telomerase-extended, 3′ single-stranded DNA terminus is then converted to duplex DNA by the same priming and polymerization machinery used in chromosome replication. Because the telomere DNA is a simple repeat that does not encode a biomolecule, the cell can tolerate a certain amount of variability in final telomere length. It is important to note, however, that the DNA product is not completely duplex. The 3′ terminus of a new telomere still has single-stranded DNA, due to the same RNA primer–removal problem discussed earlier.

Figure 11-36: Extension of the ends of linear chromosomes by telomerase. Telomeres at the ends of eukaryotic chromosomes are composed of a short repeating DNA sequence. Shown here is the repeating 5′-TTGGGG-3′ sequence of Tetrahymena. Telomerase extends the 3′ single-stranded DNA end with dNTPs, using its internal RNA molecule as template. The extended 3′ single strand of DNA is filled in by RNA priming and DNA synthesis. Removal of the RNA primer for this fill-in reaction still leaves a short 3′ single-stranded DNA overhang; this end is sequestered by telomere DNA–binding proteins. These proteins protect the chromosome ends from becoming substrates for the cell’s double-strand break repair machinery.

401

Telomere Length Is Associated with Immortality and Cancer

Tetrahymena, like other single-celled organisms, is immortal and can divide innumerable times. On mutagenesis of the telomerase gene, Tetrahymena telomeres shorten with each cell division, until, after 20 to 25 generations, the telomeres have shrunk below a critical level and the cells die. Similar observations apply to mammalian cells. Most somatic cells have little or no telomerase, and cells in culture divide about 40 to 60 times before losing their telomeres and dying. This suggests that telomeres act as a clock that counts down the number of cell divisions. Indeed, the cells of individuals with premature aging syndromes have shorter-than-normal telomeres (Highlight 11-2). But what if telomerase were expressed in all cells? Does the activation of telomerase hold promise as our ticket to immortality, the proverbial “fountain of youth”? Probably not. Studies in mice have shown that activation of telomerase in somatic cells leads to an increased incidence of tumors, and lifespan is shortened by cancer. Several additional mutations are needed to form a cancer cell, not just telomerase activation. These other mutations involve suppression of the programmed cell death pathway and activation of the mitotic pathway (i.e., mutations in tumor suppressors or in oncogenes; see Chapter 12). But activation of telomerase would decrease the number of mutations needed for a cell to become cancerous. Perhaps more informative is the finding that mice without telomerase have a normal lifespan. Despite these mixed findings, research on telomerase and antiaging therapy will probably continue past the lifespan of all those who read this book.

The fact stands that telomerase is associated with cancer. HeLa cells are an immortal human cell line derived from a woman named Henrietta Lacks, who died of ovarian cancer in 1951. HeLa cells express telomerase and have been grown in cell culture for decades in laboratories around the world. If telomerase could be inhibited by a drug, the telomeres in cancer cells should shorten with each cell division until the telomeres collapse, causing cell death. Therefore, drugs that inhibit telomerase hold promise as an anticancer therapy, an active area of current research.

Telomeres are Protected and Regulated by Proteins

The linear ends of eukaryotic chromosomes present another problem: they could be mistaken for sites of chromosome breakage and thereby induce the cell’s DNA repair systems that would join the ends of chromosomes together. The cell has two main repair systems that use recombination to join broken DNA ends (see Chapter 13). This would have disastrous consequences because, as described earlier, the joined chromosomes would be torn apart during cell division. In the cell, telomeric ends are protected by specialized telomere DNA–binding proteins that prevent chromosomes from joining and also regulate telomerase activity to prevent telomeres from growing abnormally long.

402

403

HIGHLIGHT 11-2 MEDICINE: Short Telomeres Portend Aging Diseases

Loss of chromosome ends during successive cell divisions leads to cell senescence or apoptosis, whereas expression of telomerase imparts immortality to cells. Hence, we may think of telomeres as a type of molecular clock that counts down the age of our cells. Can immortality be achieved by activating telomerase activity? Unfortunately, activation of telomerase in every cell of the body is not an option. Activation of telomerase is associated with cancer, and, in fact, mice that are engineered to express telomerase in somatic cells develop tumors and die early. We are caught in a delicate balance between requiring telomeres for life but also requiring that they have a finite lifetime. Given the connection between telomeres and cancer, one possible explanation for the telomere clock is that telomere shortening may have evolved as a way to suppress tumor growth in multicellular organisms. An equally feasible explanation is that natural selection favors a finite lifespan, because it ensures diversity in the genetic pool for evolution.

Does telomere length correlate with longevity? Several observations indicate a role of telomeres in cellular aging. A large body of epidemiological studies on human telomeres in blood cells reveals that short telomeres correlate with a number of diseases related to aging. Abnormally short telomeres are associated with diabetes, osteoporosis, obesity, cancers, impaired function of the immune system, a variety of cardiovascular diseases, and stroke. Thus, short telomeres correlate with a broad general syndrome of diseases that reflect, or perhaps even cause, aging in a fundamental way. In sum, long telomeres may impart not necessarily a longer lifespan but rather a healthier life for people in their seventies. Telomere-associated diseases seem to be influenced by both genetic and nongenetic factors; among the nongenetic factors are psychological stress, behavior, and possibly even diet. Indeed, telomere length may more clearly reflect the biological age of cells rather than chronological age. Someday, monitoring of individuals’ telomeres could become one of the “yardsticks of general health,” much like keeping track of weight and blood pressure.

Some hereditary human diseases have their basis in telomerase or in proteins that bind telomeres. Such mutations are present in individuals with dysfunctional telomeres and particular types of degenerative diseases. Importantly, these degenerative diseases are age-related because the short-telomere defect that they share is acquired with age (Figure 1). One of the most common manifestations of defective telomere biology is mutations that result in the lung disease pulmonary fibrosis. Additional degenerative diseases that correlate with dysfunctional telomeres are dyskeratosis congenita, Hoyeraal-Hreidarsson syndrome, and aplastic anaemia.

FIGURE 1 The association between dysfunctional telomeres and age-related diseases. The lines represent changes in typical percentage telomere length with age in lymphocytes. With age, telomere length decreases. The percentage normal distribution of telomere length among humans is given on the right. The red dots indicate the telomere length at the given age for four different telomere-related genetic diseases. The vertical dashed lines, Gn, Gn+1 and Gn+2, designate three successive generations, revealing that the onset of the disease occurs earlier in life with each generation.

Telomere-binding proteins bind the duplex and cover the 3′ single-stranded DNA overhang. There are differences among the telomere-binding proteins of various organisms, but all serve the same function: to preserve the telomere ends. In mammals, the telomere is bound by a set of proteins called shelterin (Figure 11-37). TRF1 and TRF2 (telomere repeat factors 1 and 2) bind the direct repeat sequences in the double-stranded DNA, and POT1 (protection of telomeres 1) binds the single-stranded DNA. These proteins are held together by TIN2 (TRF1 interacting nuclear protein-2) and TPP1 (TIN2 interacting protein 1). RAP1 (repressor/activator protein 1) is another component of shelterin.

Figure 11-37: Mammalian telomeres are bound by shelterin complex. Three copies of the shelterin complex are shown at the upper left, and one shelterin at a human telomere at the lower left. TRF1 and TRF2 bind the duplex portion, and Pot1 binds the single-stranded portion. TIN2 and TPP1 form a bridge to POT1, and RAP1 binds TRF2.

Shelterin not only protects telomeres against chromosome joining but also regulates telomere length. POT1 inhibits telomerase and becomes a stronger inhibitor with each additional POT1 protein bound to the single-stranded DNA. Thus, short telomeres are not good inhibitors of telomerase and they become longer, but long telomeres are good telomerase inhibitors and telomere extension stops. Thus, telomeres grow to a length regulated by proteins that bind them. Another facet of telomere regulation lies in a special DNA loop structure. Titia de Lange’s group at Rockefeller University, along with Jack Griffith’s laboratory at the University of North Carolina, discovered that, in mammals, the 3′-terminal single-stranded DNA is folded back and hybridized to the duplex to form a loop called a t-loop (telomere loop) (Figure 11-38). The t-loop buries the 3′ terminus of the telomere, which further regulates telomere length. The t-loop is thought to be a dynamic structure. When telomeres are long, the t-loop is more stable; when short, the t-loop is destabilized and telomerase can gain access for telomere elongation. At this point, there are also fewer shelterins, and telomere elongation commences.

Figure 11-38: Mammalian t-loops. The t-loop at the end of a linear chromosome of mammalian cells is visible in this electron micrograph. The telomere has been separated from the rest of the chromosomal DNA by a restriction enzyme.

404

Titia de Lange

Telomere DNA–binding proteins have yet another important function. Broken DNA ends induce two important signal responses that, when activated, lead to cell cycle arrest. These signal pathways are referred to as the ATM and ATR pathways, both of which detect damaged DNA and stop the cell cycle. The signal for these pathways is either a broken DNA end (ATR) or a section of single-stranded DNA (ATM). Telomeres have both of these structures, yet must be prevented from inducing the ATM and ATR signal pathways—otherwise the cell could not divide. The proteins that bind to telomeres hide the ends and the single-stranded DNA and thus prevent these signal responses. Studies indicate that TRF2 represses ATM and POT1 represses ATR. In mammalian cells, the t-loop provides yet further protection against activation of these signaling pathways.

In contrast to the broad conservation of telomerase among different species, the proteins that bind telomeres often differ from one species to another. The different telomere-binding proteins in different organisms may be based in the diversity of organisms’ repair and signal pathways. So far, we have focused on how shelterin functions on mammalian chromosomes to protect and regulate telomeres. Yeast serves as an example of an organism that uses a different set of telomere-binding proteins. Yeast telomeres are bound by a three-subunit complex composed of Rap1, Rif1, and Rif2. Rap1 has homology to mammalian RAP1, but differs in that yeast Rap1 binds the telomeric duplex DNA directly. Neither Rif1 nor Rif2 shares homology with mammalian telomere-binding proteins. Yeast Cdc13 binds the single-stranded portion of the telomere. Unlike mammalian POT1, yeast Cdc13 positively regulates telomerase, recruiting it for telomere extension. Rap1 negatively regulates telomerase activity.

405

SECTION 11.5 SUMMARY

  • Termination of replication in E. coli occurs halfway around the circular chromosome from oriC. Bidirectional replication forks meet head-on within a terminus region bordered on both sides by multiple Ter sites. Tus binds to Ter and blocks replication forks in one direction but not the other, thus localizing termination to the terminus region.

  • Replication of eukaryotic linear chromosomes cannot be completed at the extreme ends with the replication fork machinery. To solve this end replication problem, telomeres are synthesized at the chromosome ends by telomerase, which carries its own RNA template strand and adds multiple 6-mer repeats to the 3′ terminus, extending the 3′ single-stranded DNA. The single strand is then converted to duplex DNA by priming and chain extension.

  • After telomere synthesis, removal of the last RNA primer still leaves a small single-strand gap in the DNA that cannot be filled. Because the telomeric repeats are noncoding and can be replaced by further telomerase action, their loss is of no consequence.

  • Somatic cells lack telomerase, and they die when their telomeres become too short. Telomerase is activated in cancer cells, which become immortal and form tumors. Harnessing the activity of telomerase to kill cancer cells or rejuvenate normal but aging somatic cells is an important subject of medical research.

  • Telomere DNA–binding proteins protect the ends of chromosomes from nucleases and recombination.

UNANSWERED QUESTIONS

Although we know the major actors that replicate bacterial chromosomes, the mechanics of advancing a replication fork in the highly condensed DNA of a chromosome in the cell still raises several questions. Regulation of the various steps in replication affects cell division and thus is central to preventing uncontrolled cell growth in diseases such as cancer. Control of replication will undoubtedly be an important subject of future studies.

  1. How do replication forks respond to DNA-bound proteins? Chromosomal DNA contains many DNA-bound proteins, including repressors, transcription activators, and nucleosomes. We know that the replisome can displace and bypass RNA polymerase in E. coli, but only if the direction of replication is the same as the direction of transcription. What happens when the replication and transcription machineries collide head-on? In eukaryotes, do nucleosomes stay bound to DNA during replication, and how is the epigenetic information contained within them sustained in the daughter chromosomes?

  2. What protein modifications control replication? The impact of protein modifications on replication control in eukaryotic cells is extremely important and probably involves the phosphorylation of replication proteins, because their phosphorylation state can be seen to change with phases of the cell cycle. The identity of the kinases, which proteins and amino acid residues they modify, and the change in activity these modifications bring about are nearly unexplored territory.

  3. What is the relationship between telomerase, aging, and cancer? The loss of telomeres leads to chromosome instability and cell death. Most normal somatic cells lack telomerase and die when their telomeres become too short. Immortal cancer cells express telomerase, and their telomeres are maintained. These observations imply that telomerase, aging, and immortality are related. The clinical ramifications of controlling telomere length in cells is a highly active area of current research.

406

HOW WE KNOW: DNA Polymerase Reads the Sequence of the DNA Template to Copy the DNA

Bessman, M.J., I.R. Lehman, E.S. Simms, and A. Kornberg. 1958. Enzymatic synthesis of deoxyribonucleic acid: II. General properties of the reaction. J. Biol. Chem. 233:171–177.

Lehman, I.R., M.J. Bessman, E.S. Simms, and A. Kornberg. 1958. Enzymatic synthesis of deoxyribonucleic acid: I. Preparation of substrates and partial purification of an enzyme from Escherichia coli. J. Biol. Chem. 233:163–170.

Arthur Kornberg, 1918–2007

Arthur Kornberg did not intend to discover how DNA was made, or even to become a scientist. He was a physician on a naval ship, but soon after setting out to sea, his single publication as a medical student led to an offer to transfer to the National Institutes of Health. After jumping ship, he began an incredible scientific odyssey that founded the field of replication enzymology.

Kornberg and his group wanted to understand how the DNA polymer was made. They developed an assay for DNA synthesis using bacterial cell extracts to which they added [14C]thymidine to ensure that any radioactive polymer recovered would be DNA and not RNA. Though radioactive incorporation was feeble, it was reproducible. During fractionation of the extract, Kornberg’s group discovered that several heat-stable factors (i.e., not proteins) were needed for the DNA synthesis reaction. They identified these as nucleoside triphosphates. Kornberg also found that excess unlabeled DNA had to be added to the cell extracts in order to observe DNA synthesis. These insights allowed the purification of what we now know as DNA polymerase I (Pol I).

On characterizing Pol I, the researchers were initially puzzled that it required all four dNTPs for robust DNA synthesis. If DNA were serving only as a primer, why couldn’t a DNA polymer be made from just one, two, or three types of nucleotides? The finding implied that the enzyme received instructions from existing DNA acting as a template, as suggested by Watson and Crick, but at that time, the idea of an enzyme receiving direction from its substrate was preposterous. Kornberg’s group conducted experiments to test whether this was in fact the case. They tested DNAs that varied in A=T versus G≡C content, and the result was astounding. Regardless of the mix of dNTPs, the ratio of A=T and G≡C in the product matched that in the template DNA. That settled it! The DNA was serving not only as primer but also as a template. To support this conclusion, they used Pol I to convert the 5.4 kb single-stranded φX174 bacteriophage genome into the duplex viral form. The double-stranded DNA product was infectious! This finding set off a flurry of newscasts: “Life created in a test tube!” More importantly, it marked the beginnings of biotechnology.

Then came a discovery by John Cairns that polA mutant E. coli cells, with less than 1% residual Pol I activity, had no growth defects. This result, combined with identification of numerous genes required for replication, revealed a process far more complex than anyone had imagined. Unsettling to Kornberg and his colleagues was the questioning of their work on DNA polymerase I in pointed editorials in Nature New Biology. Did the assays used to purify Pol I result in a red herring? Do “real” polymerases need a primed template? Are dNTPs the true precursors of DNA? Is a 3′→5′ exonuclease proofreader needed by the “real” DNA polymerase?

Kornberg’s son Tom identified both Pol II and Pol III from extracts of polA mutant cells. These polymerases were just like Pol I in the use of a primed template and dNTPs and the presence of a 3′→5′ proofreading exonuclease. Fortunately, the controversial issues raised in Nature New Biology soon vanished. Coincidentally, so did the journal itself.

407

Polymerase Processivity Depends on a Circular Protein That Slides along DNA

Kong X.P., R. Onrust, M. O’Donnell, and J. Kuriyan. 1992. Three-dimensional structure of the β subunit of E. coli DNA polymerase III holoenzyme: A sliding DNA clamp. Cell 69:425–437.

Stukenberg, P.T., P.S.-V. Studwell, and M. O’Donnell. 1991. Mechanism of the sliding β clamp of DNA polymerase III holoenzyme. J. Biol. Chem. 266:11,328–11,334.

DNA polymerases that replicate chromosomes were long known to require “accessory proteins” that somehow confer rapid and processive polymerase activity. However, it seemed a contradiction that proteins that increase the affinity of polymerase for DNA also enable rapid motion along DNA. Specifically, how can a polymerase bind DNA tightly and also rapidly slide along it?

Surprisingly, experiments showed that the β subunit of Pol III holoenzyme, by itself, binds to DNA. This required the γ complex and ATP. However, the β subunit bound only circular DNA and could not bind linear primed DNA. This suggested that the β subunit binds DNA by encircling it, and thus slides off linear DNA. No protein was known to encircle DNA at the time, so this idea was not taken seriously. However, the test was rather simple. A [3H]β subunit was loaded onto circular primed DNA, and the reaction mixture was divided. In one tube, the DNA was linearized using BamHI, and in the other tube, the DNA was untreated and remained circular. The two reaction mixtures were then analyzed on gel filtration columns. The [3H]β bound to the large DNA molecule eluted much earlier (fractions 7 to 16) than [3H]β not bound to DNA (fractions 20 to 40).

If the [3H]β subunit encircles DNA like a doughnut, it should slide off linear DNA but remain on circular DNA. This was exactly the result observed. The solid circles in the upper plot in Figure 1 show the sample treated with BamHI. Most of the [3H]β in this sample eluted late, as [3H]β not associated with DNA. In the untreated sample (open circles), the early fractions show [3H]β bound to DNA. The result is clear: β remains on circular DNA but falls off linear DNA. This behavior suggests that β is shaped like a ring.

FIGURE 1 Experiments revealed that the E. coli Pol III β subunit binds to DNA by encircling it and slides along the duplex.

This hypothesis was tested using DNA with two sites for a DNA-binding protein known as EBNA1 (lower plot in Figure 1). The [3H]β was loaded onto DNA (with γ complex and ATP), then EBNA1 was added. One half of the reaction mixture was treated with EcoRV (open circles), which cuts the DNA between the two EBNA1 proteins. The result from gel filtration analysis (also known as gel-exclusion chromatography; see Highlight 4-1) showed that [3H]β was retained on the linear DNA by EBNA1 bound to each end, supporting the idea that β encircles the DNA and slides along it. Another experiment, not shown here, demonstrated that Pol III core binds to β, implying that β acts as a clamp that encircles DNA and tethers the polymerase to the template for high processivity during synthesis. The hypothesis that β is circular was confirmed by its crystal structure (see Figure 11-15).

408

HOW WE KNOW: Replication Requires an Origin

Hiraga, S. 1976. Novel F prime factors able to replicate in Escherichia coli Hfr strains. Proc. Natl. Acad. Sci. USA 73:198–202.

Oka, A., H. Sasaki, K. Sugimoto, and M. Takanami. 1984. Sequence organization of replication origin of the Escherichia coli K-12 chromosome. J. Mol. Biol. 176:443–458.

Zyskind, J.H.W., J.M. Cleary, W.S. Brusilow, N.E. Harding, and D.W. Smith. 1983. Chromosomal replication origin from the marine bacterium Vibrio harveyi functions in Escherichia coli: oriC consensus sequence. Proc. Natl. Acad. Sci. USA 80:1164–1168.

To identify the replication origin of a host cell, plasmid, or virus, a plasmid with a known origin and a selectable marker (e.g., the gene conferring resistance to the antibiotic ampicillin) is first treated with restriction enzymes to excise the origin (Figure 2). DNA extracted from host cells is cut with a restriction enzyme to produce many fragments. Individual fragments are inserted into the cut plasmid, and DNA ligase is used to recreate plasmid DNA circles. These recombinant plasmids are transferred into E. coli, and the transformed cells are plated on selective media (e.g., plates containing ampicillin). To survive the antibiotic in the medium and form a colony, cells must contain the plasmid with the ampicillin-resistance gene. In turn, the plasmid must contain a functional origin of replication in order to continue duplicating itself over multiple cell generations. Surviving plasmids are isolated from the bacteria and sequenced to identify the origin required for replication.

FIGURE 2 The recombinant plasmid method for identifying replication origins. AmpR indicates a gene for ampicillin resistance.

The recombinant plasmid approach has identified numerous origins of bacterial chromosomes, plasmids, and bacteriophage. Yeast (a eukaryote) has defined origins that can be isolated in a similar way. However, eukaryotic plasmids cannot be selected using antibiotics. Instead, genes needed for the metabolism of a particular amino acid are used, and cells are plated on media lacking that amino acid. This experimental approach has not been successful in identifying replication origins in eukaryotes more complex than yeast. It is possible that higher eukaryotes do not have defined origins and that chromatin structure defines replication start sites. Alternatively, the origins of higher eukaryotes are too large to be determined by this method.

409