13. DNA Replication

Semiconservative Replication

Watson and Crick's 1953 model of DNA immediately suggested a mechanism for replication: each strand serves as a template for synthesis of a new complementary strand. Three models were proposed — conservative, semiconservative, and dispersive. The elegant Meselson-Stahl experiment (1958) resolved the question definitively.

The Meselson-Stahl Experiment

E. coli were grown for many generations in $^{15}$N-labeled medium (heavy), then shifted to $^{14}$N (light) medium. After each generation, DNA was extracted and centrifuged in a CsCl density gradient:

• Generation 0: all heavy ($^{15}$N-$^{15}$N) — single band at high density
• Generation 1: all hybrid ($^{15}$N-$^{14}$N) — single band at intermediate density
• Generation 2: 50% hybrid + 50% light ($^{14}$N-$^{14}$N) — two bands

This result ruled out conservative replication (which would have given heavy + light at generation 1) and dispersive replication (which would never give distinct bands), confirming semiconservative replication: each daughter DNA molecule contains one parental and one newly synthesized strand.

Origin Firing and Licensing

Replication initiates at specific sequences called origins of replication. E. coli has a single origin (oriC, 245 bp), while the human genome has ~30,000 origins that fire in a coordinated temporal program. The licensing mechanism ensures each origin fires exactly once per cell cycle:

The pre-replicative complex (pre-RC) is assembled in G1 phase. In S phase, CDK and DDK kinases activate the MCM2-7 helicase and simultaneously prevent re-licensing by targeting Cdc6 for degradation and exporting Cdt1 from the nucleus. This ensures no origin fires twice, preventing re-replication.

Derivation 1: Replication Fork Mechanics

Because DNA polymerases synthesize only in the 5' $\rightarrow$ 3' direction and require a primer, the two antiparallel strands are replicated by fundamentally different mechanisms.

Leading Strand

The leading strand is oriented 3' $\rightarrow$ 5' relative to fork movement, so DNA polymerase can synthesize it continuously in the 5' $\rightarrow$ 3' direction. It requires only a single RNA primer at the origin.

Lagging Strand and Okazaki Fragments

The lagging strand is oriented 5' $\rightarrow$ 3' relative to fork movement. DNA polymerase must work in the direction opposite to fork movement, synthesizing short Okazaki fragments (1,000-2,000 nt in E. coli; 100-200 nt in eukaryotes). Reiji and Tuneko Okazaki discovered these in 1968 using pulse-labeling with $^3$H-thymidine.

Each Okazaki fragment requires:

• Primase (DnaG): synthesizes an RNA primer (~10 nt)
• DNA Pol III: extends the primer, synthesizing the Okazaki fragment
• DNA Pol I: removes the RNA primer (5' $\rightarrow$ 3' exonuclease) and fills the gap
• DNA Ligase: seals the nick between fragments

Replisome Architecture

The replisome is a multi-protein machine that coordinates all replication activities. In E. coli, it consists of DnaB helicase (unwinds at ~1000 bp/sec), SSB (single-strand binding protein), two DNA Pol III holoenzymes (one for each strand), primase, and the clamp loader complex. The lagging strand polymerase is thought to form a "trombone loop" that allows it to synthesize Okazaki fragments while remaining associated with the replisome.

Replication speed in E. coli:

DNA Damage: Endogenous and Exogenous Threats

Despite the high fidelity of replication, DNA is constantly damaged by both endogenous and exogenous agents. Each human cell experiences an estimated 10,000-100,000 DNA lesions per day. The major types include:

Major DNA Repair Pathways

Defects in NER cause xeroderma pigmentosum (XP) — extreme UV sensitivity and ~1000-fold increased skin cancer risk. XP patients must avoid all sunlight exposure. There are 7 complementation groups (XPA-XPG), each representing a different NER gene. The discovery of NER led to Tomas Lindahl, Paul Modrich, and Aziz Sancar sharing the 2015 Nobel Prize in Chemistry for "mechanistic studies of DNA repair."

Double-Strand Break Repair

DSBs are the most dangerous lesions — a single unrepaired DSB can kill a cell. Two pathways handle DSBs:

• NHEJ (non-homologous end joining): fast but error-prone; Ku70/80 bind broken ends, DNA-PKcs bridges them, Artemis processes overhangs, XRCC4-Ligase IV seals. Active throughout cell cycle.
• HR (homologous recombination): accurate but requires sister chromatid template; MRN complex $\rightarrow$ BRCA1/2 $\rightarrow$ RAD51 filament formation $\rightarrow$ strand invasion. Active only in S/G2 phase.

BRCA1/2 mutations impair HR, causing hereditary breast and ovarian cancer. These tumors are uniquely sensitive to PARP inhibitors (olaparib, niraparib) through synthetic lethality: PARP inhibition collapses replication forks into DSBs, which cannot be repaired without functional BRCA1/2, selectively killing tumor cells while sparing normal cells (which retain one functional BRCA allele).

Derivation 2: Replication Fidelity — Three Layers of Error Correction

The final error rate of DNA replication ($\sim 10^{-10}$ per bp) is achieved through three successive error-correction mechanisms, each contributing multiplicatively:

Layer 1: Base Selection ($\sim 10^{-5}$)

DNA polymerase's active site geometrically selects for correct Watson-Crick base pairs. The free energy difference between correct and incorrect base pairing:

The selectivity from base-pairing alone would give an error rate of only ~$10^{-2}$. The additional selectivity comes from induced fit: correct base pairs trigger a conformational change that closes the fingers domain around the substrate, optimizing the geometry for phosphodiester bond formation. Incorrect pairs fail to trigger this closure.

Layer 2: Proofreading ($\sim 10^{-2}$)

DNA Pol III (and eukaryotic Pol $\varepsilon$ and Pol $\delta$) possess a 3' $\rightarrow$ 5' exonuclease activity. When a misincorporated base causes local distortion of the DNA helix, the primer terminus is transferred from the polymerase site to the exonuclease site, and the incorrect nucleotide is excised:

Layer 3: Mismatch Repair ($\sim 10^{-3}$)

The post-replicative mismatch repair (MMR) system (MutS, MutL, MutH in E. coli; MSH2/MSH6, MLH1/PMS2 in humans) detects and corrects mismatches that escape proofreading. The critical challenge is distinguishing the parental strand (correct) from the daughter strand (containing the error):

• In E. coli: methylation state (GATC sites are transiently unmethylated on the new strand)
• In eukaryotes: strand breaks (nicks in the newly synthesized strand)

Combined Fidelity

For the human genome (3.2 $\times$ 10$^9$ bp), this predicts $\sim 0.3$ mutations per cell division — remarkably close to the measured rate of ~1 mutation per cell division in human cells. Defects in mismatch repair (Lynch syndrome: MLH1/MSH2 mutations) increase mutation rates ~100-fold, causing hereditary nonpolyposis colorectal cancer (HNPCC).

Eukaryotic vs Prokaryotic Replication: Key Differences

While the fundamental chemistry of DNA synthesis is conserved, eukaryotic and prokaryotic replication differ in important ways:

A key eukaryotic complexity is nucleosome reassembly behind the replication fork. The replisome must displace ~30 million nucleosomes during S phase and reassemble them on both daughter strands. Histone chaperones (CAF-1, ASF1, FACT) coordinate this process, ensuring epigenetic information is faithfully transmitted.

Derivation 3: Topological Challenges — Supercoiling

Unwinding the DNA helix at the replication fork creates positive supercoils ahead of the fork. Since DNA has ~10.5 bp per helical turn, replication at 1000 nt/sec generates ~100 positive supercoils per second. If unresolved, this torsional stress would halt replication.

Linking Number

The topology of closed circular DNA is described by the linking number ($Lk$), which is the sum of twist ($Tw$) and writhe ($Wr$):

For relaxed B-DNA: $Lk_0 = N/10.5$ where N is the number of base pairs. The specific linking difference measures supercoiling:

Most cellular DNA is negatively supercoiled ($\sigma \approx -0.06$), which facilitates strand separation for replication and transcription. Topoisomerases resolve supercoiling:

• Type I topoisomerases: cut one strand, pass the other through, reseal. Change $Lk$ by $\pm 1$. No ATP required.
• Type II topoisomerases (DNA gyrase in bacteria): cut both strands, pass a duplex segment through, reseal. Change $Lk$ by $\pm 2$. ATP-dependent. DNA gyrase introduces negative supercoils ahead of the fork.

Quinolone antibiotics (ciprofloxacin, levofloxacin) target bacterial DNA gyrase, trapping the enzyme-DNA covalent complex and causing lethal double-strand breaks — exploiting the topological requirement for replication.

Derivation 4: The End-Replication Problem and Telomerase

Linear chromosomes face a fundamental problem: the end-replication problem, first recognized by James Watson and Alexei Olovnikov in 1971. When the final RNA primer on the lagging strand is removed, there is no upstream 3'-OH to allow DNA polymerase to fill the gap, resulting in progressive shortening of chromosome ends with each division.

Telomere Structure

Telomeres are protective caps consisting of tandem repeats of a short G-rich sequence:

The 3' G-rich strand extends as a single-stranded overhang (~150-200 nt) that loops back and invades the double-stranded telomeric DNA, forming a protective T-loop stabilized by the shelterin complex (TRF1, TRF2, POT1, TIN2, TPP1, RAP1).

Telomerase: The Ribonucleoprotein Reverse Transcriptase

Telomerase, discovered by Carol Greider and Elizabeth Blackburn in 1985 (Nobel Prize 2009), is a reverse transcriptase that uses its own RNA component (TERC) as a template to extend telomeres:

Telomere Shortening Rate

In normal human somatic cells (which lack telomerase), telomeres shorten by approximately 50-100 bp per cell division. The Hayflick limit can be estimated:

In practice, the Hayflick limit is ~50-70 doublings for human fibroblasts, suggesting additional factors contribute to senescence. Critically short telomeres activate the DNA damage response (ATM/p53 pathway), triggering permanent cell cycle arrest (senescence) or apoptosis.

Translesion Synthesis and Damage Tolerance

When the replication fork encounters an unrepaired DNA lesion, the normal replicative polymerase (Pol $\varepsilon$ or Pol $\delta$) stalls because it cannot accommodate the distorted template. Translesion synthesis (TLS) polymerases can bypass such lesions at the cost of reduced fidelity.

Y-Family DNA Polymerases

TLS polymerases belong to the Y-family and have open, spacious active sites that can accommodate damaged bases:

• Pol $\eta$ (eta): accurately bypasses thymine dimers (inserts AA opposite TT dimer). Deficiency causes xeroderma pigmentosum variant (XP-V) — UV sensitivity despite functional NER.
• Pol $\iota$ (iota): bypasses some minor-groove adducts; extremely error-prone (~$10^{-1}$ error rate)
• Pol $\kappa$ (kappa): bypasses bulky adducts (benzo[a]pyrene-guanine)
• Rev1: deoxycytidyl transferase; inserts C opposite abasic sites and some adducts

The Polymerase Switch

When the replicative polymerase stalls, the PCNA sliding clamp is monoubiquitylated at Lys-164 by the Rad6/Rad18 ubiquitin ligase. Ubiquitylated PCNA recruits TLS polymerases (which have ubiquitin-binding domains), displacing the stalled replicative polymerase:

After bypassing the lesion (typically 1-5 nucleotides), the TLS polymerase is replaced by the replicative polymerase for continued high-fidelity synthesis. This "two-polymerase" model ensures that the mutagenic TLS activity is restricted to the minimum necessary region. Polyubiquitylation of PCNA (K63-linked chains) instead activates an error-free template switching pathway using the undamaged sister chromatid.

Cell Cycle Checkpoints and Replication Stress

The cell cycle checkpoint machinery monitors DNA integrity before, during, and after replication. Two master kinases — ATM (responds to DSBs) and ATR (responds to replication stress/ssDNA) — coordinate the checkpoint response.

The S-Phase Checkpoint

When replication forks stall, the accumulating single-stranded DNA (coated with RPA) activates ATR:

Replication stress — defined as the slowing or stalling of replication forks — is now recognized as a hallmark of cancer. Oncogene activation (MYC, RAS, cyclin E overexpression) causes replication stress by promoting premature S-phase entry with insufficient dNTP pools or unresolved transcription-replication conflicts. Cancer cells become dependent on ATR/Chk1 to survive this chronic replication stress, making these kinases attractive therapeutic targets. ATR inhibitors (ceralasertib) are in clinical trials, showing particular promise in combination with DNA-damaging agents.

Replication-Transcription Conflicts

Because replication and transcription share the same DNA template, collisions between the replisome and RNA polymerase are inevitable — particularly at long, highly transcribed genes. Head-on collisions (Pol II and replisome moving toward each other) are especially dangerous, causing fork stalling and R-loop formation (RNA:DNA hybrids that displace the non-template strand). The cell uses multiple mechanisms to resolve these conflicts:

• Replication timing programs place highly transcribed genes in early-replicating regions
• Co-directionality: essential genes tend to be transcribed in the same direction as replication
• R-loop resolution by RNase H1/H2 and helicases (SETX, DHX9, BLM)
• The Fanconi anemia pathway resolves interstrand crosslinks encountered during replication

Derivation 5: Eukaryotic Replication — Speed and Organization

The human genome is ~700 times larger than E. coli's, yet S phase takes only ~8 hours (vs ~40 minutes for E. coli). This is achieved through multiple origins:

In reality, origins fire in a staggered temporal program throughout S phase (early: euchromatin; late: heterochromatin), extending S phase to ~8 hours. The staggered firing prevents depletion of dNTP pools and replication factors.

Eukaryotic DNA Polymerases

Processivity Enhancement by Sliding Clamps

The sliding clamp (PCNA in eukaryotes, $\beta$-clamp in E. coli) encircles DNA and tethers the polymerase. The processivity factor can be modeled as:

where $k_{\text{pol}}$ is the polymerization rate and $k_{\text{off}}$ is the dissociation rate. Without the clamp, $k_{\text{off}}$ is high (low processivity). The sliding clamp reduces $k_{\text{off}}$ by ~1000-fold, enabling continuous synthesis of thousands of nucleotides without dissociation.

Chromatin Assembly: Replicating the Epigenome

During DNA replication, the chromatin structure must be faithfully duplicated along with the DNA sequence. The parental nucleosomes ahead of the replication fork are disassembled and their histones are recycled to the two daughter strands, supplemented by newly synthesized histones.

Parental Histone Recycling

When the replisome encounters a nucleosome, the MCM helicase separates the DNA strands, and the histone octamer is disassembled. The parental H3-H4 tetramers (which carry most epigenetic modifications) are distributed randomly between the leading and lagging daughter strands:

New histone H3-H4 dimers (carrying the H3K56ac mark and H4K5/K12ac as "new histone" signatures) are assembled by the histone chaperone CAF-1 (chromatin assembly factor 1), which interacts with PCNA at the replication fork. The newly deposited histones must acquire the correct modifications from the parental histones — this process is critical for epigenetic inheritance.

Epigenetic Mark Propagation

Different histone modifications are propagated by different mechanisms:

• H3K9me3: HP1 (reader) recruits SUV39H1 (writer), which methylates neighboring new H3 — a self-reinforcing "read-write" loop that maintains heterochromatin
• H3K27me3: PRC2 (writer) binds its own product (reader function within EED subunit), spreading the repressive mark to adjacent nucleosomes
• DNA methylation: DNMT1 recognizes hemimethylated CpG sites at the replication fork and methylates the new strand, faithfully copying the pattern

The ~15-30 minute window after replication, before modifications are fully restored, represents a vulnerable period when epigenetic states can be altered. This may contribute to the epigenetic instability observed in rapidly dividing cancer cells, where replication outpaces the modification machinery.

Replication Origins in Disease and Aging

The regulation of replication origins has emerging significance in cancer biology and aging:

Oncogene-Induced Replication Stress

Oncogene activation (MYC, RAS, cyclin E overexpression) causes replication stress through several mechanisms:

• Premature S-phase entry with insufficient dNTP pools and licensing factors
• Increased origin firing depleting RPA and other replication factors
• Transcription-replication conflicts from increased RNA synthesis
• Reactive oxygen species (ROS) from increased mitochondrial activity causing oxidative DNA damage

This replication stress generates DNA damage that activates the ATR/Chk1 checkpoint, causing cell cycle arrest or apoptosis — a natural tumor suppression mechanism. Cancer cells that bypass this checkpoint (by losing p53 or ATM function) accumulate genomic instability, driving tumor evolution but creating vulnerabilities that can be exploited therapeutically.

Replication and Aging

Replicative senescence (the Hayflick limit) contributes to aging through multiple mechanisms: telomere shortening triggers a persistent DNA damage response (DDR), senescent cells secrete inflammatory cytokines (the senescence-associated secretory phenotype, SASP), and the accumulation of senescent cells drives tissue dysfunction. Senolytic drugs (dasatinib + quercetin, navitoclax) that selectively kill senescent cells are in clinical trials for age-related diseases, representing a new therapeutic paradigm based on replication biology.

Applications in Medicine and Research

Historical Context

Arthur Kornberg discovered DNA polymerase I in 1956 (Nobel Prize 1959), but it was later shown to be primarily a repair enzyme. The true replicase, DNA Pol III, was identified in 1970. The Meselson-Stahl experiment (1958) has been called "the most beautiful experiment in biology" — its simplicity and elegance made the semiconservative mechanism unambiguous.

The structure of the sliding clamp (PCNA/beta-clamp) was solved by John Kuriyan in 1992, revealing the remarkable toroidal architecture that encircles DNA without forming a covalent bond. The crystal structure of the T7 replisome by Charles Richardson revealed how leading and lagging strand synthesis are physically coupled. Thomas Cech and Jack Szostak's work on telomerase RNA (Nobel Prize 2009, shared with Blackburn and Greider) established that telomerase is a ribonucleoprotein reverse transcriptase, connecting RNA biology to chromosome maintenance.

DNA Replication and Cancer

Dysregulated DNA replication is a hallmark of cancer. Several key connections:

• Re-replication: overexpression of CDT1 or CDC6 (origin licensing factors) can cause re-firing of origins within S phase, leading to gene amplification and genomic instability
• Checkpoint deficiency: loss of p53, ATM, or ATR allows cells with damaged DNA to continue replicating, accumulating mutations
• POLD1/POLE mutations: germline mutations in the proofreading domains of Pol $\delta$ and Pol $\varepsilon$ cause polymerase proofreading-associated polyposis (PPAP), with extremely high mutation rates and early-onset colorectal cancer
• Microsatellite instability: mismatch repair deficiency (Lynch syndrome) causes expansion/contraction of short tandem repeats, detectable by MSI testing and predictive of immunotherapy response
• Topoisomerase poisons: camptothecin (Top1, used in colorectal cancer) and etoposide (Top2, used in lung cancer) trap topoisomerase-DNA covalent complexes, converting these essential enzymes into DNA-damaging agents

Python Simulations