Part 3: DNA Replication

Copying the Genome

DNA replication is the molecular mechanism by which a cell duplicates its entire genome prior to division. In E. coli, the 4.6 Mb circular chromosome is replicated in approximately 40 minutes at 37 °C, requiring coordinated action of more than 30 different proteins at each replication fork. Human cells face an even greater challenge: replicating 6.4 billion base pairs during each S phase, accomplished by firing tens of thousands of replication origins in a precisely regulated temporal program.

The fundamental chemistry is deceptively simple — a nucleophilic attack by the 3′-OH of the growing chain on the α-phosphorus of the incoming dNTP, releasing pyrophosphate. Yet the machinery that ensures this reaction occurs with extraordinary speed, accuracy, and coordination represents one of the most sophisticated molecular systems in biology.

1. Semi-Conservative Replication

Three Competing Models (1953–1958)

After Watson and Crick proposed the double-helix structure in 1953, three models for how DNA might replicate were debated:

Conservative

The parental duplex remains intact; an entirely new duplex is synthesized. After one round, one molecule is entirely “old” and one is entirely “new.”

Semi-Conservative

Each strand serves as a template. After one round, both daughter molecules contain one old strand and one new strand. This is the correct model.

Dispersive

Old and new DNA are interspersed throughout both strands of both daughter molecules. Blocks of old and new DNA alternate.

The Meselson–Stahl Experiment (1958)

Matthew Meselson and Franklin Stahl designed what is often called “the most beautiful experiment in biology.” Their approach exploited the density difference between DNA containing the heavy nitrogen isotope ¹⁵N and the normal light isotope ¹⁴N.

Step 1: Label with ¹⁵N

E. coli cells were grown for many generations in medium containing ¹⁵NH₄Cl as the sole nitrogen source. After ≥14 doublings, virtually all nitrogen in the DNA was¹⁵N, making the DNA “heavy” (density ≈ 1.724 g/cm³ vs. 1.710 g/cm³ for ¹⁴N-DNA).

Step 2: Transfer to ¹⁴N Medium

Cells were transferred to medium with ¹⁴NH₄Cl. All newly synthesized DNA would now incorporate only ¹⁴N (“light”).

Step 3: CsCl Equilibrium Density Gradient Centrifugation

DNA was extracted at each generation and centrifuged in a CsCl gradient at ~140,000 g for 20+ hours. DNA migrates to its isopycnic position — the point where the buoyant density of CsCl equals the density of the DNA. UV absorption photography at 260 nm revealed distinct bands.

Results

Gen 0

Single heavy band (HH) — all DNA is ¹⁵N/¹⁵N.

Gen 1

Single band at intermediate density (HL) — each molecule has one ¹⁵N strand and one ¹⁴N strand. This rules out conservative replication (which would show one heavy + one light band).

Gen 2

Two bands: 50% intermediate (HL) and 50% light (LL). This rules out dispersive replication (which would show a single band shifting progressively toward light density).

Gen n

HL fraction remains constant (2 out of 2ⁿ molecules), while LL fraction increases. The ratio of HL:LL = 1:(2ⁿ⁻¹ − 1).

Quantitative prediction: After n generations of growth in ¹⁴N medium, there will be 2ⁿ total DNA molecules. Exactly 2 molecules will be hybrid (HL) — one from each original parental strand — and (2ⁿ − 2) will be fully light (LL). The fraction of hybrid DNA = $\frac{2}{2^n} = 2^{1-n}$, which approaches zero as n increases.

2. Replication Fork Architecture

The Asymmetry Problem

All known DNA polymerases synthesize in the 5′→3′ direction only, yet the two template strands at a replication fork run antiparallel. This creates a fundamental asymmetry: one strand (the leading strand) can be synthesized continuously in the direction of fork movement, while the other (the lagging strand) must be synthesized discontinuously as a series of short fragments in the direction opposite to fork movement.

Replication Fork Diagram:


        Fork movement ──────────────────►

     5'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━3'  (template)
     3'━━━━━━━━━━━━━━━━━━━━━━━►                    Leading strand
                               ◄━━Pol III━━━         (continuous, 5'→3')

                      Helicase ◄══╗
                                  ║
     3'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━5'  (template)
         ◄━━━  ◄━━━  ◄━━━  ◄━━━                   Lagging strand
          OF4   OF3   OF2   OF1                    (Okazaki fragments)

Leading Strand

Synthesized 5′→3′ continuously
Same direction as helicase/fork movement
Requires only a single RNA primer at the origin
Synthesized by Pol III (prokaryotic) or Pol ε (eukaryotic)
High processivity due to sliding clamp association
Rate: ~1,000 nt/s in E. coli, ~50 nt/s in human cells

Lagging Strand

Synthesized 5′→3′ but in short fragments
Opposite direction to fork movement
Requires repeated priming events (every 1–2 kb in prokaryotes)
Synthesized by Pol III (prokaryotic) or Pol δ (eukaryotic)
Must cycle on and off the template (“trombone model”)
Okazaki fragments joined by DNA ligase after primer removal

Okazaki Fragments

Discovered by Reiji and Tsuneko Okazaki in 1968, these short DNA fragments are the transient intermediates of lagging-strand synthesis. Their size differs dramatically between organisms:

1,000–2,000 nt

Prokaryotic Okazaki fragments (E. coli)

100–200 nt

Eukaryotic Okazaki fragments

Each Okazaki fragment begins with a ~10 nt RNA primer synthesized by primase. In E. coli, DNA Pol I removes RNA primers via its 5′→3′ exonuclease activity and fills the gap with DNA. In eukaryotes, RNase H1 and FEN1 (flap endonuclease 1) perform this role. Finally, DNA ligase seals the nick by forming a phosphodiester bond, consuming one ATP (eukaryotes) or NAD⁺ (prokaryotes).

3. Initiation of Replication

Prokaryotic Initiation: E. coli oriC

E. coli has a single, well-defined origin of replication called oriC, spanning 245 bp. This compact sequence contains all the information needed to direct replication initiation:

DnaA Boxes (5 copies, 9-mer consensus: TTATCCACA)

DnaA protein (52 kDa) binds cooperatively to these sequences in its ATP-bound form (DnaA-ATP). About 20–30 DnaA monomers assemble into a right-handed helical filament that wraps the origin DNA, introducing positive superhelical strain. This destabilizes the adjacent AT-rich region. DnaA-ADP is inactive; the conversion of DnaA-ATP to DnaA-ADP after initiation (RIDA: regulatory inactivation of DnaA) helps ensure once-per-cell-cycle firing.

AT-Rich Region (3 copies of 13-mer: GATCTNTTNTTTT)

The three 13-mer repeats are rich in A-T base pairs, which have only two hydrogen bonds (vs. three for G-C). DnaA-induced unwinding begins here, creating a single-stranded “open complex” or “DnaA bubble” spanning ~28 bp.

Helicase Loading

DnaC (helicase loader) delivers DnaB (hexameric helicase) to the single-stranded region. Two DnaB hexamers are loaded, one on each strand, oriented to travel in opposite directions (5′→3′ along the strand they encircle). DnaC then dissociates upon ATP hydrolysis. DnaB translocates and recruits DnaG primase, forming the primosome.

Regulation: Ensuring Once-Per-Cell-Cycle Firing

• SeqA sequestration: Newly replicated GATC sites in oriC are hemimethylated. SeqA protein binds hemimethylated GATC, blocking DnaA access for ~1/3 of the cell cycle until Dam methylase restores full methylation.
• RIDA: The Hda protein stimulates DnaA-ATP hydrolysis to DnaA-ADP on the β-clamp, inactivating DnaA after initiation.
• datA titration: The datA locus (near oriC) contains many DnaA-binding sites that sequester DnaA-ATP, reducing its effective concentration.

Eukaryotic Initiation: Origin Licensing and Firing

Eukaryotic chromosomes are much larger and linear, requiring thousands of replication origins to complete S phase in a reasonable time. Human cells fire ~30,000–50,000 origins per S phase. The system is controlled by a two-step mechanism that separates licensing (G1 phase) from firing (S phase), ensuring each origin fires at most once per cell cycle.

Step 1: Origin Licensing (G1 Phase) — Pre-RC Assembly

The Origin Recognition Complex (ORC), a six-subunit AAA+ ATPase (Orc1–6), binds replication origins throughout the cell cycle. In budding yeast, origins are defined by the 11-bp ARS Consensus Sequence (ACS); in metazoans, origin specification is more complex and may involve chromatin context rather than strict sequence specificity.

During G1, two licensing factors — Cdc6 (AAA+ ATPase) and Cdt1 — are recruited to ORC-bound origins. Together, they load the MCM2-7 hexameric helicase as a head-to-head double hexamer encircling double-stranded DNA. This “pre-replicative complex” (pre-RC) licenses the origin for firing. The MCM double hexamer is catalytically inactive at this stage.

Step 2: Origin Firing (S Phase) — CMG Helicase Activation

S-phase CDK (Cyclin A/E–Cdk2) and DDK (Dbf4-dependent kinase Cdc7) phosphorylate MCM subunits and other factors. This triggers recruitment of Cdc45 and the GINS complex (Sld5, Psf1, Psf2, Psf3) to form the active CMG helicase (Cdc45–MCM–GINS).

The double hexamer splits into two single CMG helicases that travel in opposite directions, each encircling the leading-strand template (3′→5′ translocation). DNA Pol α-primase synthesizes the first RNA-DNA primer, Pol ε takes over the leading strand, and Pol δ synthesizes the lagging strand.

Preventing Re-Replication

• Geminin binds and inhibits Cdt1 during S, G2, and M phases.
• S-CDK phosphorylates Orc1, Cdc6, and Cdt1, targeting them for ubiquitin-mediated degradation (SCF^Skp2) or nuclear export.
• CDK activity must drop (at mitotic exit) before new pre-RCs can assemble, creating a strict temporal separation between licensing and firing.

4. Elongation Enzymes

DNA Polymerases

E. coli DNA Polymerase III Holoenzyme

The Pol III holoenzyme is a 900 kDa assembly and the principal replicative polymerase in E. coli. It consists of 10 different subunits organized into three functional modules:

αεθ Core

α subunit (dnaE): 5′→3′ polymerase activity. Belongs to the C-family of DNA polymerases. Catalyzes phosphodiester bond formation.
ε subunit (dnaQ): 3′→5′ proofreading exonuclease. Removes misincorporated nucleotides, improving fidelity ~100-fold.
θ subunit (holE): Stabilizes ε, enhancing its exonuclease activity ~2–3-fold.

β Sliding Clamp

Ring-shaped homodimer (2 × 40.6 kDa)
Encircles DNA; inner diameter ~35 Å, outer ~80 Å
Slides freely along duplex DNA
Tethers Pol III core to template, increasing processivity from ~10 nt to >50,000 nt
Each monomer has 3 domains with identical topology despite no sequence similarity

γ/τ Clamp Loader

τ₃δδ′χψ complex (or γ₃δδ′χψ)
AAA+ ATPase machine that opens the β clamp ring and loads it onto primer-template junctions
τ subunit also binds DnaB helicase, physically coupling polymerase to helicase
ATP binding opens clamp; ATP hydrolysis closes clamp on DNA and ejects the loader

Eukaryotic Replicative Polymerases

Pol α–Primase Complex

Four-subunit complex (p180, p68, p58, p48). The p48 subunit is the primase that synthesizes an ~8–12 nt RNA primer. Pol α (p180) then extends this with ~20–30 nt of DNA, creating a ~30–40 nt RNA-DNA primer. Pol α lacks 3′→5′ proofreading exonuclease and has low processivity — it functions only in initiation, not bulk synthesis.

Pol ε (Leading Strand)

Four-subunit B-family polymerase (Pol2, Dpb2, Dpb3, Dpb4). The catalytic subunit Pol2 (256 kDa) has both polymerase and 3′→5′ exonuclease domains. Extremely high fidelity. Travels with the CMG helicase at the front of the replisome. Processivity is enhanced by the PCNA sliding clamp.

Pol δ (Lagging Strand)

Three-subunit B-family polymerase (Pol3, Pol31, Pol32). Has 3′→5′ proofreading exonuclease. Extends Okazaki fragments after Pol α-primase. Also participates in mismatch repair, base excision repair, and nucleotide excision repair. Interacts with PCNA via a PIP-box motif.

PCNA Sliding Clamp & RFC Clamp Loader

PCNA (Proliferating Cell Nuclear Antigen) is a homotrimer forming a ring structurally analogous to the bacterial β clamp. Each monomer has two domains, so the trimer has pseudo-6-fold symmetry matching the β clamp dimer. PCNA is a platform for many DNA-processing enzymes (polymerases, ligase, FEN1, mismatch repair factors). RFC (Replication Factor C) is the pentameric AAA+ clamp loader (RFC1–5) that loads PCNA onto primer-template junctions in an ATP-dependent manner.

Replicative Helicases

DnaB (E. coli)

Homohexameric ring (6 × 52 kDa)
Translocates 5′→3′ on the lagging-strand template
Encircles the lagging-strand template strand
Powered by ATP hydrolysis (~100–1000 nt/s unwinding)
Each subunit cycles through ATP, ADP, and empty states (“rotary” or “hand-over-hand” mechanism)
Recruits DnaG primase to form the primosome

MCM2-7 / CMG (Eukaryotic)

Heterohexameric ring of MCM2, 3, 4, 5, 6, 7 (each ~100 kDa)
Translocates 3′→5′ on the leading-strand template
Encircles the leading-strand template strand (opposite polarity to DnaB)
Active as CMG complex: Cdc45–MCM2-7–GINS
Powered by ATP hydrolysis through a conserved AAA+ motor
Loaded as double hexamer (head-to-head) in G1; activated in S phase

Primase

DNA polymerases cannot initiate synthesis de novo — they require a pre-existing 3′-OH to extend. Primases are specialized RNA polymerases that synthesize short RNA primers (~10 nucleotides) to provide this 3′-OH.

DnaG Primase (E. coli)

65.6 kDa single-subunit enzyme. Synthesizes ~10–12 nt RNA primers. Recruited to the replication fork by direct interaction with DnaB helicase. Transiently active: synthesizes a primer, then dissociates to allow Pol III to begin Okazaki fragment extension.

Pol α–Primase (Eukaryotic)

Four-subunit complex integrating primase and polymerase activities. The p48 primase subunit synthesizes ~8–12 nt RNA, then hands off to the p180 Pol α subunit for ~20 nt DNA extension. This RNA-DNA hybrid primer is used by both Pol ε (leading) and Pol δ (lagging). Associates with the replisome through Ctf4/AND-1 trimer.

Single-Strand DNA Binding Proteins

Helicase unwinding produces single-stranded DNA (ssDNA) that is vulnerable to nuclease attack, secondary structure formation, and chemical damage. SSB proteins coat ssDNA to prevent these problems.

SSB (E. coli)

Homotetramer (4 × 18.8 kDa). Each subunit has an OB-fold domain that binds ssDNA. Binds cooperatively in two modes: (SSB)₃₅ wraps 35 nt around two subunits (limited cooperativity) and (SSB)₆₅ wraps 65 nt around all four subunits (highly cooperative, forms long filaments). Destabilizes hairpins and secondary structures.

RPA (Eukaryotic)

Heterotrimeric complex: RPA70 (70 kDa), RPA32 (32 kDa), RPA14 (14 kDa). Contains six OB-fold domains total (four in RPA70, one in RPA32, one in RPA14). Binds ssDNA with very high affinity (K_d ~10⁻¹⁰ M). Also plays essential roles in DNA repair, recombination, and checkpoint signaling. Phosphorylation of RPA32 by ATR/ATM kinases modulates its interactions during the DNA damage response.

5. Termination of Replication

Prokaryotic Termination: ter Sites and Tus

In E. coli, replication terminates in a broad region opposite oriC. This region contains 10 ter sites (TerA–TerJ), each a 23-bp sequence that binds the Tus protein (terminus utilization substance, 36 kDa monomer).

The Tus–ter complex acts as a polar trap: it allows a replication fork approaching from one direction to pass through but blocks a fork arriving from the opposite direction. This is a “mousetrap” mechanism — the approaching helicase (DnaB) unfolds a critical cytosine in the ter sequence, which then locks into a pocket on Tus, creating a nearly irreversible block. The two sets of ter sites are oriented to create a “fork trap” that ensures forks converge within this region regardless of which fork arrives first.

When forks meet, the remaining gap is filled, ligated, and the interlinked (catenated) daughter chromosomes are separated by Topoisomerase IV, a type II topoisomerase that passes one duplex through another.

Eukaryotic Termination

Eukaryotes lack defined termination sequences. Instead, converging forks simply meet and merge. The process involves:

• Fork convergence: When two CMG helicases from adjacent replicons approach each other, the remaining unreplicated DNA between them shrinks until the forks meet.
• CMG unloading: Upon meeting, CMG helicases transition onto dsDNA. The SCF^Dia2 ubiquitin ligase (in yeast) or CRL2^Lrr1 (in metazoans) ubiquitylates MCM7, triggering p97/VCP/Cdc48 segregase to extract CMG from chromatin.
• Gap filling and ligation: Remaining single-strand gaps are filled by Pol δ and sealed by DNA ligase I.
• Decatenation: Topoisomerase IIα (Topo II) resolves any catenanes (interlocking daughter duplexes) by passing one double helix through a transient double-strand break in the other. This is essential before chromosome segregation in mitosis.

6. Telomere Replication and the End Problem

The End-Replication Problem

Linear chromosomes present a unique challenge: when the RNA primer at the very 5′ end of each lagging strand is removed, there is no upstream 3′-OH available for a polymerase to fill the resulting gap. Consequently, with each round of replication, the daughter chromosome produced by lagging-strand synthesis is shorter than the parent — by approximately 50–200 bp per division in human cells. This was first recognized by James Watson (1972) and independently by Alexei Olovnikov (1971).


  Parent:     5'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━3'
              3'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━5'

  After replication:
  Leading:    5'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━3'  (complete)
              3'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━5'  (new, complete)

  Lagging:    5'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━3'       (new, SHORTER)
              3'━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━5'  (template, full)
                                                   ^^^^
                                        Gap from primer removal

Telomerase

Telomerase is a specialized ribonucleoprotein reverse transcriptase that extends telomeric DNA to counteract the end-replication problem. It was discovered by Carol Greider and Elizabeth Blackburn in 1985 in Tetrahymena (Nobel Prize, 2009).

TERT (Telomerase Reverse Transcriptase)

The catalytic protein subunit (~127 kDa in humans). Contains a reverse transcriptase domain that synthesizes telomeric DNA using the RNA template. Also contains an N-terminal TEN domain for DNA binding and a C-terminal extension (CTE) for processivity. Repressed in most somatic cells; reactivated in ~85–90% of human cancers.

TERC (Telomerase RNA Component)

451 nt in humans (hTR). Contains an 11-nt template region (5′-CUAACCCUAAC-3′) complementary to the human telomeric repeat (TTAGGG). After extending the 3′ overhang by one repeat, telomerase translocates to realign the template for the next addition cycle. Also contains structural domains for TERT binding, Cajal body localization (CAB box), and H/ACA motif for stability.

Mechanism: Telomerase binds the 3′ single-stranded overhang of the telomere (human: 50–300 nt G-rich overhang). Using its RNA template, it extends the G-rich strand by adding TTAGGG repeats. After extension, conventional lagging-strand machinery (Pol α-primase, Pol δ) fills in the complementary C-rich strand, followed by primer removal and ligation. The net result is telomere length maintenance.

Shelterin Complex

Telomeres are protected from being recognized as DNA double-strand breaks by a six-protein complex called shelterin:

TRF1

Binds double-stranded TTAGGG repeats. Regulates telomere length (negative regulator). Homodimer.

TRF2

Binds ds-TTAGGG. Essential for t-loop formation and suppression of ATM-dependent DNA damage response. Homodimer.

POT1

Binds single-stranded G-rich overhang. Suppresses ATR signaling. Regulates telomerase access.

TPP1

Bridges POT1 to TIN2. TEL patch on TPP1 recruits telomerase to telomeres. Enhances telomerase processivity.

TIN2

Central hub connecting TRF1, TRF2, and TPP1. Stabilizes the entire shelterin complex on the telomere.

RAP1

Recruited by TRF2. Inhibits NHEJ at telomeres. Involved in transcriptional regulation at subtelomeric regions.

The 3′ G-rich overhang invades the duplex telomeric DNA to form a displacement loop (t-loop), creating a “D-loop” at the invasion point. This structure, mediated by TRF2, sequesters the chromosome end and prevents activation of DNA damage signaling pathways (ATM, ATR) and inappropriate repair by NHEJ or HR.

Alternative Lengthening of Telomeres (ALT)

Approximately 10–15% of cancers maintain telomeres without telomerase, using a recombination-based mechanism called ALT. Key features:

• Highly heterogeneous telomere lengths (from <3 kb to >50 kb)
• ALT-associated PML bodies (APBs) where recombination occurs
• Depends on break-induced replication (BIR) and homologous recombination
• Often associated with mutations in ATRX/DAXX chromatin remodeling complex
• Extrachromosomal telomeric circles (C-circles) serve as a biomarker

Hayflick Limit and Cellular Senescence

In 1961, Leonard Hayflick observed that human fibroblasts in culture can divide only ~50–70 times before entering irreversible growth arrest (senescence). This Hayflick limit is directly linked to telomere shortening:

• Somatic cells: ~50–200 bp lost per division
• Critical telomere length (~4–6 kb) triggers p53/Rb senescence pathway
• If p53/Rb are inactivated, cells bypass senescence but enter “crisis” with massive genomic instability and cell death
• Rare survivors that activate telomerase or ALT become immortalized (cancer)
• Germ cells and stem cells express low levels of telomerase, extending replicative lifespan

7. Replication Fidelity

The overall error rate of DNA replication is remarkably low: approximately 10⁻¹⁰ errors per base pair per cell division. This extraordinary accuracy is achieved through three successive layers of quality control, each contributing multiplicatively:

10⁻⁵

Base Selection

DNA polymerase active site discriminates correct from incorrect dNTPs using geometric complementarity (Watson-Crick geometry), hydrogen bonding, base stacking, and an “induced fit” conformational change. The fingers domain closes around the nascent base pair; incorrect pairs cause steric clashes that slow the forward reaction.

10⁻²

Proofreading (3′→5′ Exonuclease)

Misincorporation slows polymerization and increases the probability that the primer terminus melts into the 3′→5′ exonuclease active site (~30 Å away in Pol III). The mispaired terminal nucleotide is excised, and the correct nucleotide is then incorporated. Improves fidelity ~100-fold.

10⁻³

Mismatch Repair (MMR)

Post-replicative scanning by MutS (mismatch recognition), MutL (molecular matchmaker), and MutH (strand discrimination via hemimethylation in E. coli). In eukaryotes, MSH2–MSH6 (MutSα) recognizes mismatches and small loops; MLH1–PMS2 (MutLα) coordinates excision and resynthesis. Strand discrimination uses PCNA orientation and nicks in the nascent strand. Improves fidelity ~1000-fold.

Combined Fidelity Calculation

The three mechanisms act sequentially, so their error rates multiply:

\[ \text{Overall error rate} = \underbrace{10^{-5}}_{\text{base selection}} \times \underbrace{10^{-2}}_{\text{proofreading}} \times \underbrace{10^{-3}}_{\text{MMR}} = 10^{-10} \text{ per bp per division} \]

For the human genome (6.4 × 10⁹ bp), this predicts approximately 0.64 mutations per cell division — roughly consistent with measured somatic mutation rates of ~0.5–1.0 per cell division.

8. Mathematical Framework

Replication Fork Kinetics

Fork Progression Rate

The position of a replication fork as a function of time, assuming constant speed:

\[ x(t) = x_0 + v_{\text{fork}} \cdot t \]

where $v_{\text{fork}}$ is the fork velocity (~1000 nt/s in E. coli, ~20–50 nt/s in human cells), and $x_0$ is the origin position.

Time to Replicate a Genome

For a circular chromosome with bidirectional replication from a single origin:

\[ T_{\text{replication}} = \frac{L}{2 \cdot v_{\text{fork}}} \]

For E. coli: L = 4.6 × 10⁶ bp, v = 1000 bp/s gives T = 2300 s ≈ 38 min.

For eukaryotes with N origins firing simultaneously:

\[ T_{\text{S-phase}} \approx \frac{L}{2 \cdot N \cdot v_{\text{fork}}} \]

For human cells: L = 6.4 × 10⁹ bp, v = 30 bp/s, N ≈ 30,000 gives T ≈ 3600 s ≈ 1 hour. Actual S phase is 6–8 hours because origins fire asynchronously in a temporal program.

Polymerase Processivity

Processivity (P) is the average number of nucleotides incorporated per binding event:

\[ P = \frac{k_{\text{pol}}}{k_{\text{pol}} + k_{\text{off}}} \cdot \frac{1}{1 - p_{\text{step}}} \]

where $k_{\text{pol}}$ is the polymerization rate, $k_{\text{off}}$ is the dissociation rate, and $p_{\text{step}}$ is the probability of stepping forward per catalytic cycle. With the β sliding clamp, $k_{\text{off}}$ is drastically reduced, yielding P > 50,000 nt.

Derivation: Replication Fork Speed and Processivity

Starting from the Michaelis-Menten framework for DNA polymerase catalysis, we derive the effective fork velocity and processivity.

Step 1: Define the polymerase catalytic cycle

DNA polymerase binds a dNTP substrate (S) and catalyzes incorporation. The enzyme-substrate interaction follows:

$$E + S \underset{k_{-1}}{\overset{k_1}{\rightleftharpoons}} ES \xrightarrow{k_{\text{cat}}} E + \text{DNA}_{n+1} + \text{PP}_i$$

Step 2: Apply steady-state approximation to ES complex

Setting $d[\text{ES}]/dt = 0$:

$$k_1[E][S] = (k_{-1} + k_{\text{cat}})[\text{ES}]$$

The Michaelis constant is $K_M = (k_{-1} + k_{\text{cat}})/k_1$.

Step 3: Nucleotide incorporation rate

The rate of nucleotide incorporation (fork velocity in nt/s) at saturating dNTP concentration:

$$v_{\text{fork}} = \frac{k_{\text{cat}}[\text{dNTP}]}{K_M + [\text{dNTP}]} \xrightarrow{[\text{dNTP}] \gg K_M} k_{\text{cat}}$$

For E. coli Pol III: $k_{\text{cat}} \approx 1000 \text{ s}^{-1}$, $K_M \approx 10\text{--}50 \text{ }\mu\text{M}$, and intracellular [dNTP] $\approx 100\text{--}300 \text{ }\mu\text{M}$, so the enzyme operates near $V_{\max}$.

Step 4: Catalytic efficiency as a specificity measure

The catalytic efficiency $k_{\text{cat}}/K_M$ sets the second-order rate constant for substrate capture:

$$\frac{k_{\text{cat}}}{K_M} = \frac{k_1 \cdot k_{\text{cat}}}{k_{-1} + k_{\text{cat}}}$$

For correct dNTPs, $k_{\text{cat}}/K_M \sim 10^8 \text{ M}^{-1}\text{s}^{-1}$ (near diffusion limit). For incorrect dNTPs, $k_{\text{cat}}/K_M \sim 10^3\text{--}10^4 \text{ M}^{-1}\text{s}^{-1}$, giving a discrimination factor of $\sim 10^{4}\text{--}10^{5}$.

Step 5: Processivity from competing rates

At each step, the polymerase either extends (rate $k_{\text{pol}}$) or dissociates (rate $k_{\text{off}}$). The probability of stepping forward:

$$p_{\text{step}} = \frac{k_{\text{pol}}}{k_{\text{pol}} + k_{\text{off}}}$$

Step 6: Average processivity as a geometric series

The mean number of nucleotides incorporated before dissociation is a geometric distribution:

$$P = \sum_{n=0}^{\infty} n \cdot p_{\text{step}}^n(1 - p_{\text{step}}) = \frac{p_{\text{step}}}{1 - p_{\text{step}}} = \frac{k_{\text{pol}}}{k_{\text{off}}}$$

Without the sliding clamp: $k_{\text{off}} \sim 1\text{ s}^{-1}$, giving $P \sim 10\text{--}50$ nt. With the $\beta$ clamp: $k_{\text{off}} \sim 10^{-2}\text{ s}^{-1}$, giving $P > 50{,}000$ nt.

Error Rate Formalism

The probability of a mutation surviving at a given position:

\[ \mu = \epsilon_{\text{ins}} \cdot (1 - f_{\text{proof}}) \cdot (1 - f_{\text{MMR}}) \]

where $\epsilon_{\text{ins}} \approx 10^{-5}$ is the misincorporation frequency, $f_{\text{proof}} \approx 0.99$ is the fraction corrected by proofreading, and $f_{\text{MMR}} \approx 0.999$ is the fraction corrected by mismatch repair. This gives $\mu \approx 10^{-5} \times 0.01 \times 0.001 = 10^{-10}$.

Derivation: Combined Fidelity from Three Error-Correction Layers

Starting from the principle that replication errors must escape three independent, sequential quality-control checkpoints to become permanent mutations.

Step 1: Define the misincorporation frequency

DNA polymerase selects dNTPs based on Watson-Crick geometry and induced fit. The insertion error rate is the ratio of catalytic efficiencies for wrong vs. right nucleotides:

$$\epsilon_{\text{ins}} = \frac{(k_{\text{cat}}/K_M)_{\text{wrong}}}{(k_{\text{cat}}/K_M)_{\text{right}}} \approx 10^{-4}\text{--}10^{-5}$$

Step 2: Proofreading exonuclease correction

A misincorporated nucleotide distorts the primer terminus geometry, slowing the forward polymerization rate and increasing the probability of the 3′-end melting into the exonuclease active site. Let $f_{\text{proof}}$ be the fraction of errors removed:

$$f_{\text{proof}} = \frac{k_{\text{exo}}^{\text{mismatch}}}{k_{\text{exo}}^{\text{mismatch}} + k_{\text{pol}}^{\text{mismatch}}} \approx 0.99$$

The error rate after proofreading: $\epsilon_{\text{ins}} \times (1 - f_{\text{proof}})$.

Step 3: Post-replicative mismatch repair (MMR)

MutS/MutL/MutH (prokaryotes) or MSH2-MSH6/MLH1-PMS2 (eukaryotes) scan newly replicated DNA. Errors that escaped proofreading are recognized as mismatches. Let $f_{\text{MMR}}$ be the fraction corrected by MMR:

$$f_{\text{MMR}} \approx 0.999 \quad \text{(removes 99.9\% of remaining mismatches)}$$

Step 4: Multiply escape probabilities

Since the three mechanisms act sequentially and independently, the probability that an error survives all three layers is the product of escape probabilities:

$$\mu = \epsilon_{\text{ins}} \times (1 - f_{\text{proof}}) \times (1 - f_{\text{MMR}})$$

Step 5: Substitute numerical values

$$\mu = 10^{-5} \times (1 - 0.99) \times (1 - 0.999) = 10^{-5} \times 10^{-2} \times 10^{-3} = 10^{-10} \text{ per bp per division}$$

Step 6: Predict mutations per genome per division

For the human diploid genome ($G = 6.4 \times 10^9$ bp):

$$M = \mu \times G = 10^{-10} \times 6.4 \times 10^9 \approx 0.64 \text{ mutations per cell division}$$

This is remarkably consistent with the observed somatic mutation rate of ~0.5-1.0 mutations per division measured by whole-genome sequencing of clonal lineages.

Telomere Shortening Dynamics

Telomere length after n divisions without telomerase:

\[ L(n) = L_0 - \Delta L \cdot n \]

With telomerase activity adding $\delta$ bp per division:

\[ L(n) = L_0 - (\Delta L - \delta) \cdot n \]

Senescence occurs when $L(n) \leq L_{\text{crit}}$, giving the Hayflick limit:

\[ n_{\text{Hayflick}} = \frac{L_0 - L_{\text{crit}}}{\Delta L - \delta} \]

For human somatic cells: L₀ ≈ 10,000 bp, L_crit ≈ 4,000 bp, ΔL ≈ 50–200 bp, δ = 0. This gives n ≈ 30–120 divisions, consistent with the Hayflick limit of ~50–70 doublings.

Derivation: Telomere Shortening and the Hayflick Limit

Starting from the end-replication problem: conventional DNA polymerase cannot replicate the very 3′ end of a linear chromosome because it requires a primer upstream of the terminus.

Step 1: The end-replication problem

After each S phase, the lagging strand template loses the RNA primer region at its 5′ end. Additionally, 5′→3′ exonuclease processing of the C-rich strand generates the 3′ G-rich overhang. Together, these cause a net loss of $\Delta L$ base pairs per division (typically 50-200 bp in human somatic cells).

Step 2: Linear shortening model (no telomerase)

If shortening is constant per division, telomere length after $n$ divisions is a simple arithmetic sequence:

$$L(n) = L_0 - \Delta L \cdot n$$

where $L_0$ is the initial telomere length (typically ~10,000 bp at birth in humans).

Step 3: Include telomerase activity

If telomerase adds $\delta$ bp per division (partial activity in stem cells, full in germ/cancer cells), the net shortening per division is $(\Delta L - \delta)$:

$$L(n) = L_0 - (\Delta L - \delta) \cdot n$$

When $\delta = \Delta L$, telomere length is maintained indefinitely (immortal cells). When $\delta < \Delta L$, shortening continues but at a reduced rate.

Step 4: Define the critical length for senescence

When telomere length drops below a critical threshold $L_{\text{crit}}$ (~4,000-6,000 bp), uncapped chromosome ends are recognized as DNA damage. This activates the ATM/ATR-p53-p21 DDR pathway, triggering irreversible cell cycle arrest (senescence).

Step 5: Solve for the Hayflick limit

Set $L(n_{\text{Hayflick}}) = L_{\text{crit}}$ and solve:

$$L_0 - (\Delta L - \delta) \cdot n_{\text{Hayflick}} = L_{\text{crit}}$$

$$n_{\text{Hayflick}} = \frac{L_0 - L_{\text{crit}}}{\Delta L - \delta}$$

Step 6: Numerical estimate for human somatic cells

For normal somatic cells ($\delta = 0$):

$$n_{\text{Hayflick}} = \frac{10{,}000 - 4{,}000}{100 - 0} = 60 \text{ divisions}$$

Using $\Delta L = 50$ bp: $n = 120$; using $\Delta L = 200$ bp: $n = 30$. The observed Hayflick limit of ~50-70 doublings for human fibroblasts falls within this range, confirming telomere shortening as the molecular clock of replicative senescence.

Okazaki Fragment Frequency

The number of Okazaki fragments produced per replicon of length L:

\[ N_{\text{OF}} = \frac{L}{\langle l_{\text{OF}} \rangle} \]

For E. coli: L = 4.6 × 10⁶ bp, average Okazaki fragment = 1,500 nt gives ~3,067 fragments per chromosome per replication. For a human cell with ~50 nt/s fork speed and 150 nt average Okazaki length: ~6.4 × 10⁹ / (2 × 150) ≈ 21 million Okazaki fragments per S phase.

Derivation: Okazaki Fragment Length Distribution from Stochastic Priming

Starting from the stochastic nature of primase-DnaB interaction on the lagging strand template, we derive the expected distribution of Okazaki fragment lengths.

Step 1: Model priming as a Poisson process

Primase associates transiently with DnaB helicase at the replication fork. Each priming event is stochastic, with a constant probability $\lambda$ of initiating a new primer per unit length of single-stranded template exposed. This is a memoryless (Poisson) process along the DNA coordinate.

Step 2: Inter-priming distance follows an exponential distribution

The distance between consecutive priming events (which determines Okazaki fragment length) follows an exponential distribution:

$$P(\ell) = \lambda \cdot e^{-\lambda \ell}$$

where $\ell$ is the fragment length and $1/\lambda$ is the mean inter-priming distance.

Step 3: Mean and variance of fragment length

For the exponential distribution:

$$\langle \ell \rangle = \frac{1}{\lambda}, \quad \text{Var}(\ell) = \frac{1}{\lambda^2}, \quad \text{CV} = \frac{\sigma}{\mu} = 1$$

For E. coli: $\langle \ell \rangle \approx 1{,}000\text{--}2{,}000$ nt, so $\lambda \approx 5 \times 10^{-4}\text{--}10^{-3}$ per nt.

Step 4: Correction for minimum fragment length

In reality, a minimum time is required for primer synthesis (~1 s for a 10-nt RNA primer). During this time, the fork advances $v_{\text{fork}} \times t_{\text{primer}}$ nt, setting a minimum fragment length $\ell_{\min}$. The corrected distribution is a shifted exponential:

$$P(\ell) = \lambda \cdot e^{-\lambda(\ell - \ell_{\min})} \quad \text{for } \ell \geq \ell_{\min}$$

Step 5: Number of fragments per replicon

For a replicon of length $L$ replicated by two forks, the lagging strand on each fork produces:

$$N_{\text{OF}} = \frac{L/2}{\langle \ell \rangle} = \frac{L \cdot \lambda}{2}$$

For E. coli: $N = 4.6 \times 10^6 / (2 \times 1{,}500) \approx 1{,}533$ fragments per fork, ~3,067 total per replication.

Step 6: Eukaryotic refinement

In eukaryotes, Okazaki fragments are shorter (~150-200 nt) due to nucleosome spacing constraints. The priming rate $\lambda$ is higher, and fragment length correlates with the nucleosome repeat length (~165-200 bp), suggesting that chromatin structure imposes a periodic modulation on the priming probability, deviating from the pure exponential model toward a more peaked (gamma-like) distribution.

Python Simulation: Replication Fork Progression

This simulation models replication fork progression with kinetic parameters from E. coli. It tracks leading strand (continuous synthesis), lagging strand (discontinuous Okazaki fragment generation), and the gap between them. The output includes fork position over time, Okazaki fragment length distribution, and the dynamic leading–lagging gap.

Model Parameters

• Fork speed (helicase): 1,000 nt/s
• Lagging strand polymerase: 800 nt/s (accounts for cycling overhead)
• Mean Okazaki fragment: 1,500 ± 300 nt
• Primer synthesis time: 1.0 s (~10 nt RNA primer)
• Total replication: 50,000 bp segment

Replication Fork Kinetics Simulator

Python

Models leading/lagging strand synthesis with Okazaki fragment generation and fork gap dynamics

script.py119 lines

import numpy as np

# =============================================================================
# Replication Fork Progression Simulator
# Models leading strand continuous synthesis and lagging strand Okazaki
# fragment generation with realistic kinetic parameters
# =============================================================================

# --- Parameters ---
fork_speed = 1000          # nucleotides per second (E. coli)
total_length = 50000       # total bases to replicate
dt = 0.1                   # time step (seconds)
okazaki_mean = 1500        # mean Okazaki fragment length (prokaryotic)
okazaki_std = 300          # standard deviation of fragment length
primer_synthesis_time = 1.0  # seconds to synthesize RNA primer (~10 nt)
primer_length = 10         # nucleotides in RNA primer
lagging_speed = 800        # lagging strand polymerase speed (nt/s, slightly slower)

# --- Simulation ---
np.random.seed(42)
time_points = []
fork_positions = []
leading_positions = []
lagging_positions = []
okazaki_fragments = []

t = 0.0
fork_pos = 0.0
leading_pos = 0.0
lagging_pos = 0.0
current_fragment_start = 0.0
current_fragment_length = 0.0
priming = False
prime_timer = 0.0
next_fragment_target = np.random.normal(okazaki_mean, okazaki_std)

while fork_pos < total_length:
    # Fork (helicase) advances
    fork_pos = min(fork_pos + fork_speed * dt, total_length)

# Leading strand follows fork continuously
    leading_pos = fork_pos

# Lagging strand: cycles of priming then extending
    if priming:
        prime_timer -= dt
        if prime_timer <= 0:
            priming = False
            current_fragment_start = lagging_pos
            current_fragment_length = 0.0
    else:
        extension = lagging_speed * dt
        current_fragment_length += extension
        lagging_pos += extension

if current_fragment_length >= next_fragment_target:
            okazaki_fragments.append(current_fragment_length)
            priming = True
            prime_timer = primer_synthesis_time
            next_fragment_target = max(500, np.random.normal(okazaki_mean, okazaki_std))

time_points.append(t)
    fork_positions.append(fork_pos)
    leading_positions.append(leading_pos)
    lagging_positions.append(lagging_pos)
    t += dt

time_arr = np.array(time_points)
fork_arr = np.array(fork_positions)
leading_arr = np.array(leading_positions)
lagging_arr = np.array(lagging_positions)

# --- Compute fork gap (leading - lagging) ---
gap = leading_arr - lagging_arr

# --- Print summary statistics ---
print("=" * 60)
print("REPLICATION FORK SIMULATION RESULTS")
print("=" * 60)
print(f"Total genome replicated:    {total_length:,} bp")
print(f"Total time:                 {t:.1f} seconds ({t/60:.1f} min)")
print(f"Fork speed:                 {fork_speed} nt/s")
print(f"Okazaki fragments produced: {len(okazaki_fragments)}")
print(f"Mean fragment length:       {np.mean(okazaki_fragments):.0f} nt")
print(f"Std fragment length:        {np.std(okazaki_fragments):.0f} nt")
print(f"Min fragment length:        {np.min(okazaki_fragments):.0f} nt")
print(f"Max fragment length:        {np.max(okazaki_fragments):.0f} nt")
print(f"Mean leading-lagging gap:   {np.mean(gap):.0f} nt")
print(f"Max leading-lagging gap:    {np.max(gap):.0f} nt")
print()

# --- Plot 1: Fork position vs time ---
print("FORK POSITION vs TIME (every 5 seconds):")
print("-" * 50)
print(f"{'Time (s)':>10} {'Fork (kb)':>12} {'Leading (kb)':>14} {'Lagging (kb)':>14}")
for i in range(0, len(time_arr), int(5.0 / dt)):
    print(f"{time_arr[i]:10.1f} {fork_arr[i]/1000:12.1f} {leading_arr[i]/1000:14.1f} {lagging_arr[i]/1000:14.1f}")
print()

# --- Plot 2: Okazaki fragment length distribution ---
bins = np.histogram(okazaki_fragments, bins=15)
counts, edges = bins
max_count = max(counts)
print("OKAZAKI FRAGMENT LENGTH DISTRIBUTION:")
print("-" * 55)
for i in range(len(counts)):
    bar_len = int(40 * counts[i] / max_count) if max_count > 0 else 0
    label = f"{edges[i]:.0f}-{edges[i+1]:.0f}"
    print(f"{label:>14} nt | {'#' * bar_len} ({counts[i]})")
print()

# --- Plot 3: Gap between leading and lagging ---
print("LEADING-LAGGING GAP OVER TIME:")
print("-" * 50)
sample_indices = np.linspace(0, len(time_arr)-1, 20, dtype=int)
for idx in sample_indices:
    bar_len = int(30 * gap[idx] / max(gap)) if max(gap) > 0 else 0
    print(f"t={time_arr[idx]:6.1f}s  gap={gap[idx]:6.0f} nt  |{'=' * bar_len}|")

Click Run to execute the Python code

Code will be executed with Python 3 on the server

Fortran Computation: Telomere Shortening Model

This Fortran program models telomere length dynamics over 120 cell divisions under four biological scenarios: normal somatic cells (no telomerase), partial telomerase activity (stem-like), full telomerase (cancer/germ cells), and the ALT pathway. The simulation tracks when each cell type reaches the Hayflick senescence threshold and crisis.

Scenarios Modeled

Normal somatic: Loses ~50 bp/division, no telomerase. Reaches senescence at ~120 divisions.

Partial telomerase: Net loss of 5 bp/division. Extended but finite lifespan.

Full telomerase: Complete compensation. Telomere length maintained indefinitely (immortal).

ALT pathway: Recombination-based with stochastic length variation. Partially compensated.

Telomere Shortening Dynamics (4 Scenarios)

Fortran

Models telomere attrition, telomerase compensation, and ALT pathway over cell divisions

telomere_shortening.f90116 lines

program telomere_shortening
  ! =====================================================================
  ! Telomere Shortening Model with Telomerase Activity
  ! Models telomere length over successive cell divisions
  ! Includes: end-replication problem, telomerase extension, ALT pathway
  ! =====================================================================
  implicit none

integer, parameter :: max_divisions = 120
  integer, parameter :: n_scenarios = 4
  real(8) :: telomere(n_scenarios, 0:max_divisions)
  real(8) :: shortening_per_div    ! bp lost per division (end-replication)
  real(8) :: telomerase_rate       ! bp added by telomerase per division
  real(8) :: alt_rate              ! bp added by ALT per division
  real(8) :: initial_length        ! initial telomere length (bp)
  real(8) :: hayflick_limit        ! critical telomere length for senescence
  real(8) :: crisis_length         ! length triggering crisis/apoptosis
  integer :: i, j
  integer :: senescence_div(n_scenarios)
  character(len=40) :: labels(n_scenarios)

! --- Parameters ---
  initial_length = 10000.0d0      ! 10 kb initial telomere (human)
  shortening_per_div = 50.0d0     ! ~50 bp per division (human somatic)
  telomerase_rate = 45.0d0        ! partial telomerase activity
  alt_rate = 30.0d0               ! ALT recombination-based lengthening
  hayflick_limit = 4000.0d0       ! senescence threshold
  crisis_length = 2000.0d0        ! crisis threshold

labels(1) = 'Normal somatic cell'
  labels(2) = 'Partial telomerase'
  labels(3) = 'Full telomerase (cancer/stem)'
  labels(4) = 'ALT pathway (cancer)'

! --- Initialize ---
  do j = 1, n_scenarios
    telomere(j, 0) = initial_length
    senescence_div(j) = max_divisions
  end do

! --- Simulate ---
  do i = 1, max_divisions
    ! Scenario 1: Normal somatic (no telomerase)
    telomere(1, i) = telomere(1, i-1) - shortening_per_div
    if (telomere(1, i) < 0.0d0) telomere(1, i) = 0.0d0

! Scenario 2: Partial telomerase (e.g., stem-like)
    telomere(2, i) = telomere(2, i-1) - shortening_per_div + telomerase_rate
    if (telomere(2, i) < 0.0d0) telomere(2, i) = 0.0d0

! Scenario 3: Full telomerase (immortal, cancer/germ)
    telomere(3, i) = telomere(3, i-1) - shortening_per_div + shortening_per_div
    ! Maintained at initial length
    if (telomere(3, i) > initial_length) telomere(3, i) = initial_length

! Scenario 4: ALT pathway (variable, recombination-based)
    telomere(4, i) = telomere(4, i-1) - shortening_per_div + alt_rate
    ! ALT has stochastic variation (simplified as sine fluctuation)
    telomere(4, i) = telomere(4, i) + 200.0d0 * sin(dble(i) * 0.3d0)
    if (telomere(4, i) < 0.0d0) telomere(4, i) = 0.0d0
    if (telomere(4, i) > 15000.0d0) telomere(4, i) = 15000.0d0

! Check senescence
    do j = 1, n_scenarios
      if (telomere(j, i) <= hayflick_limit .and. senescence_div(j) == max_divisions) then
        senescence_div(j) = i
      end if
    end do
  end do

! --- Output Results ---
  write(*,'(A)') '============================================================'
  write(*,'(A)') ' TELOMERE SHORTENING SIMULATION RESULTS'
  write(*,'(A)') '============================================================'
  write(*,'(A,F10.0,A)') ' Initial telomere length:  ', initial_length, ' bp'
  write(*,'(A,F10.0,A)') ' Shortening per division:  ', shortening_per_div, ' bp'
  write(*,'(A,F10.0,A)') ' Hayflick limit (senesce): ', hayflick_limit, ' bp'
  write(*,'(A,F10.0,A)') ' Crisis threshold:         ', crisis_length, ' bp'
  write(*,*)
  write(*,'(A)') ' Senescence reached at division:'
  do j = 1, n_scenarios
    if (senescence_div(j) < max_divisions) then
      write(*,'(A,A,A,I4)') '   ', trim(labels(j)), ': division ', senescence_div(j)
    else
      write(*,'(A,A,A)') '   ', trim(labels(j)), ': NOT reached (immortal)'
    end if
  end do
  write(*,*)

! --- Print table every 10 divisions ---
  write(*,'(A)') ' Telomere Length (bp) Over Cell Divisions:'
  write(*,'(A)') ' -----------------------------------------------------------------'
  write(*,'(A6,4A14)') 'Div', 'Normal', 'Partial', 'Full Tase', 'ALT'
  write(*,'(A)') ' -----------------------------------------------------------------'
  do i = 0, max_divisions, 10
    write(*,'(I6,4F14.0)') i, telomere(1,i), telomere(2,i), telomere(3,i), telomere(4,i)
  end do
  write(*,*)

! --- ASCII telomere length chart (Normal somatic) ---
  write(*,'(A)') ' Normal Somatic Cell Telomere Decline:'
  write(*,'(A)') ' -----------------------------------------------------------------'
  do i = 0, max_divisions, 5
    j = nint(telomere(1, i) / 250.0d0)
    if (j < 0) j = 0
    if (telomere(1, i) <= crisis_length) then
      write(*,'(I4,A,A,A)') i, ' |', repeat('X', j), ' CRISIS'
    else if (telomere(1, i) <= hayflick_limit) then
      write(*,'(I4,A,A,A)') i, ' |', repeat('!', j), ' senescent'
    else
      write(*,'(I4,A,A)') i, ' |', repeat('=', j)
    end if
  end do

end program telomere_shortening

Click Run to execute the Fortran code

Code will be compiled with gfortran and executed on the server

Summary: Prokaryotic vs. Eukaryotic Replication

Feature	Prokaryotic (E. coli)	Eukaryotic
Origin(s)	Single (oriC, 245 bp)	Multiple (~30,000–50,000 in human)
Initiator	DnaA protein	ORC → Cdc6 → Cdt1 → MCM loading
Helicase	DnaB (5′→3′ on lagging template)	CMG/MCM2-7 (3′→5′ on leading template)
Primase	DnaG	Pol α–primase complex
Leading strand Pol	Pol III core	Pol ε
Lagging strand Pol	Pol III core	Pol δ
Sliding clamp	β clamp (homodimer)	PCNA (homotrimer)
Clamp loader	γ/τ complex	RFC (RFC1–5)
SSB	SSB (homotetramer)	RPA (heterotrimer)
Okazaki fragments	1,000–2,000 nt	100–200 nt
Primer removal	Pol I (5′→3′ exo)	RNase H1 + FEN1
Fork speed	~1,000 nt/s	~20–50 nt/s
Termination	ter/Tus fork trap	Fork convergence + Topo II decatenation
Ligase cofactor	NAD⁺	ATP

Key Equations Summary

Replication time (single origin, bidirectional):

\[ T = \frac{L}{2v} \]

Meselson-Stahl: hybrid fraction after n generations:

\[ f_{\text{hybrid}}(n) = \frac{2}{2^n} = 2^{1-n} \]

Replication fidelity (three-layer model):

\[ \mu = \epsilon_{\text{ins}} \cdot (1 - f_{\text{proof}}) \cdot (1 - f_{\text{MMR}}) \approx 10^{-10} \]

Hayflick limit from telomere dynamics:

\[ n_{\max} = \frac{L_0 - L_{\text{crit}}}{\Delta L - \delta} \]

Mutations per genome per division:

\[ M = \mu \cdot G = 10^{-10} \times 6.4 \times 10^{9} \approx 0.64 \]

← DNA Structure DNA Repair →