📚 DNA Replication: Semi-Conservative Replication and the Replication Machinery | DNA 复制：半保留复制与复制机器
DNA replication is the process by which a cell copies its entire genome before cell division. It must be accurate, fast and coordinated with the cell cycle so that each daughter cell receives a complete set of genetic instructions. In humans, the replication machinery copies about 3.2 billion base pairs in a few hours, while making roughly one mistake per billion bases incorporated. In this article we look at the evidence that replication is semi-conservative, then work through the enzymes and proteins that unwind, prime, extend, proofread and join the new strands.
DNA 复制是细胞在分裂前复制整个基因组的过程。它必须准确、快速并与细胞周期相协调，以确保每个子细胞获得一整套完整的遗传信息。人类细胞要在几小时内复制约 32 亿个碱基对，而错误率大约仅为每十亿个碱基一个。本文将先讨论"复制是半保留的"这一结论的证据，再逐一梳理解旋、引发、延伸、校对和连接新链所需的酶与蛋白。
1. Why DNA Replication Is Semi-Conservative | 为什么 DNA 复制是半保留的
When Watson and Crick described the double helix in 1953, they realised that complementary base pairing immediately suggests a copying mechanism: if the two strands separate, each strand can act as a template for building a new partner. In semi-conservative replication, each new DNA molecule consists of one original parental strand and one newly synthesised daughter strand. The word "semi-conservative" therefore means that half of every new molecule is conserved from the parent.
当 Watson 和 Crick 于 1953 年描述双螺旋结构时，他们意识到互补碱基配对直接暗示了一种复制机制：如果两条链分开，每条链都可以作为模板合成新的互补链。在半保留复制中，每个新的 DNA 分子由一条原有的亲本链和一条新合成的子链组成。因此，"半保留"的意思是每个新分子中有一半来自亲本、被保留下来。
The rules that make this possible are simple. Adenine pairs with thymine through two hydrogen bonds, and guanine pairs with cytosine through three hydrogen bonds, so the two strands of a duplex are complementary rather than identical. This has two important consequences: the base sequence of one strand dictates the sequence of the other, and the two strands can be separated without breaking the covalent sugar-phosphate backbone.
使这一切成为可能的规则非常简单。腺嘌呤与胸腺嘧啶之间形成两个氢键配对，鸟嘌呤与胞嘧啶之间形成三个氢键配对，因此双螺旋的两条链是互补的，而不是相同的。这带来两个重要结果：一条链的碱基序列决定了另一条链的序列；而且两条链可以在不破坏共价糖-磷酸骨架的情况下分开。
A = T (2 hydrogen bonds) and G ≡ C (3 hydrogen bonds) → each strand is a template for its partner
Because the strands are antiparallel, one runs 5' to 3' and the other runs 3' to 5' in the same direction of view. This antiparallel arrangement, together with the fact that DNA polymerase can only add nucleotides to a free 3'-OH group, explains why one new strand is made continuously and the other in fragments.
由于两条链是反向平行的，在同一个观察方向上，一条是 5'→3'，另一条则是 3'→5'。这种反向平行结构，加上 DNA 聚合酶只能把核苷酸加到游离的 3'-OH 基团上这一事实，共同解释了为什么一条新链是连续合成的，而另一条是分段合成的。
2. The Three Competing Models and Their Predictions | 三种竞争模型及其预测
Before the experimental evidence arrived, three models could all explain how a double helix might be copied. In semi-conservative replication, the helix opens and each parental strand templates a new strand, so every daughter molecule is a hybrid of old and new DNA. In conservative replication, the original double helix would remain fully intact and a completely new double helix would be produced alongside it. In dispersive replication, both strands of each daughter molecule would be patches of old and new DNA, cut and rejoined along their length.
在实验证据出现之前，有三种模型都能解释双螺旋是如何被复制的。在半保留复制中，双螺旋打开，每条亲本链作为模板合成一条新链，因此每个子代分子都是旧 DNA 与新 DNA 的杂合体。在全保留复制中，原始双螺旋保持完整，同时另行产生一个全新的双螺旋。在分散复制中，每个子代分子的两条链都由旧 DNA 和新 DNA 片段混合组成，沿链长被切开并重新接合。
Crucially, the three models make different predictions about the density of the DNA after one and after two rounds of replication. If cells are first grown in heavy nitrogen (¹⁵N) and then switched to normal light nitrogen (¹⁴N), the position of the DNA bands in a density gradient distinguishes the models experimentally.
关键在于，这三种模型对复制一轮和两轮之后 DNA 的密度作出不同预测。如果先把细胞培养在重氮（¹⁵N）中，再转移到普通的轻氮（¹⁴N）中，那么 DNA 在密度梯度中的条带位置就能在实验上区分这些模型。
| Model | 模型 | After 1 generation | 第 1 代后 | After 2 generations | 第 2 代后 |
| Semi-conservative | 半保留 | All molecules hybrid (intermediate density) | 全部为杂合分子（中等密度） | Half hybrid, half light | 一半杂合，一半轻 |
| Conservative | 全保留 | Half heavy, half light (two bands) | 一半重、一半轻（两条带） | One quarter heavy, three quarters light | 四分之一重，四分之三轻 |
| Dispersive | 分散 | All molecules intermediate | 全部为中等密度 | Single band that shifts towards light | 只有一条带，逐渐向轻端移动 |
3. The Meselson-Stahl Experiment | Meselson-Stahl 实验
Meselson and Stahl published the decisive experiment in 1958. They grew Escherichia coli for many generations in a medium containing the heavy isotope ¹⁵N, so that all the bacterial DNA was labelled heavy. They then transferred the cells to a medium containing normal ¹⁴N and allowed them to divide. After each generation, samples of DNA were extracted and centrifuged in a caesium chloride density gradient, where DNA forms a band at a position that depends on its density. Heavy DNA bands lower down the tube, light DNA bands higher up.
Meselson 和 Stahl 于 1958 年发表了这一决定性实验。他们先把大肠杆菌在含重同位素 ¹⁵N 的培养基中培养许多代，使细菌 DNA 全部被标记为重 DNA。随后把细胞转移到含正常 ¹⁴N 的培养基中，让它们继续分裂。每一代之后提取 DNA，并在氯化铯密度梯度中离心；DNA 会依据自身密度在梯度中形成条带，重 DNA 位于离心管下部，轻 DNA 位于上部。
After one generation in ¹⁴N, the DNA formed a single band at an intermediate position, exactly halfway between the heavy and light positions. After two generations, there were two bands of roughly equal intensity: one at the intermediate position and one at the light position. These results fit semi-conservative replication perfectly and rule out the alternatives.
在 ¹⁴N 中复制一代之后，DNA 在重带与轻带之间的中间位置形成一条单一条带。复制两代之后，出现两条强度大致相同的条带：一条位于中间位置，一条位于轻 DNA 位置。这些结果与半保留复制完全吻合，同时排除了其他两种模型。
Conservative replication is excluded because it predicts two bands (one fully heavy, one fully light) after the first generation, not a single intermediate band. Dispersive replication is excluded because after two generations it predicts a single band that gradually shifts upwards, not two clearly separated bands, and because the parental ¹⁵N would be spread through every molecule rather than concentrated in half of them.
全保留复制被排除，因为它在第一代后应预测出现两条带（一条全重、一条全轻），而不是一条中间带。分散复制也被排除，因为它在第二代后预测只出现一条逐渐上移的带，而不是两条清晰分开的带；而且按分散模型，亲本的 ¹⁵N 会分散到每一个分子中，而不是集中在其中一半分子里。
4. DNA Structure: The Rules the Machinery Must Obey | DNA 结构：复制机器必须遵守的规则
Replication is constrained by the structure of DNA itself. The molecule is a double helix about 2 nm wide, with about 10 base pairs per complete turn and a rise of roughly 0.34 nm per base pair. The two strands are held together by hydrogen bonds between complementary bases and by hydrophobic stacking interactions between the bases, while each strand is held together by phosphodiester bonds that link the 3'-OH of one deoxyribose to the 5'-phosphate of the next.
复制受到 DNA 自身结构的限制。DNA 是宽约 2 nm 的双螺旋，每个完整螺旋约含 10 个碱基对，每个碱基对上升约 0.34 nm。两条链之间由互补碱基间的氢键和碱基间的疏水堆积作用维持；而每条链内部则由磷酸二酯键维持——它把一个脱氧核糖的 3'-OH 与下一个脱氧核糖的 5'-磷酸基连接起来。
Two features dominate the mechanics of replication. First, the strands are antiparallel. Second, a purine always pairs with a pyrimidine, so A (a purine) pairs with T (a pyrimidine) and G (a purine) pairs with C (a pyrimidine). This keeps the diameter of the helix constant. It also means the base sequence of the template can be read in a strictly one-to-one way, which is the basis of the high fidelity of copying.
有两个特征主导着复制的机制。第一，两条链反向平行。第二，嘌呤总是与嘧啶配对：A（嘌呤）配 T（嘧啶），G（嘌呤）配 C（嘧啶），这使螺旋直径保持恒定。同时这也意味着模板的碱基序列可以被严格一一对应地"读取"，这正是高保真复制的基础。
5'-phosphate → deoxyribose → 3'-OH, joined by phosphodiester bonds; new nucleotides added only at a free 3'-OH
5. Origins of Replication and Initiation | 复制起点与起始
Replication does not begin at random. It starts at specific sequences called origins of replication, which are recognised by initiator proteins. In prokaryotes such as E. coli, the circular chromosome has a single origin (oriC), and the two replication forks move outwards in opposite directions, producing a theta-shaped intermediate. In eukaryotes, each linear chromosome has many origins, so replication can be completed within the limited time available in S phase.
复制并非随机开始，而是从被称为"复制起点"的特定序列起始，这些序列由起始蛋白识别。在大肠杆菌等原核生物中，环状染色体只有一个复制起点（oriC），两个复制叉向相反方向推进，形成 θ 形中间结构。在真核生物中，每条线性染色体具有多个复制起点，因此复制可以在 S 期有限的时间内完成。
Initiation also has to be controlled so that each origin fires once and only once per cell cycle. In eukaryotes, origins are "licensed" in G1 by the loading of helicase complexes, and the licence is destroyed once replication begins. This licensing system prevents re-replication of the same DNA within a single cycle, which would otherwise cause gene amplification and genome instability.
起始还必须受到调控，使每个起点在每个细胞周期中只启动一次。在真核生物中，G1 期通过装载解旋酶复合体对起点进行"许可"（licensing），一旦复制开始，这一许可即被破坏。这一许可系统防止同一段 DNA 在同一个周期内被重复复制，否则会导致基因扩增和基因组不稳定。
6. Unwinding the Helix: Helicase, Topoisomerase and SSB Proteins | 解旋：解旋酶、拓扑异构酶与单链结合蛋白
Before any new nucleotides can be added, the double helix must be opened. DNA helicase uses the energy of ATP hydrolysis to break the hydrogen bonds between the base pairs and to move along the DNA, separating the strands at the replication fork. The result is a Y-shaped fork with two exposed single-stranded templates ready for copying.
在任何新核苷酸被添加之前，双螺旋必须先被打开。DNA 解旋酶利用 ATP 水解提供的能量破坏碱基对之间的氢键并沿 DNA 移动，在复制叉处把两条链分开，形成 Y 形的复制叉以及两条可供复制的单链模板。
Unwinding creates torsional stress ahead of the fork, because the rest of the molecule resists rotation. Topoisomerases, including DNA gyrase in bacteria, relieve this strain by cutting one or both strands, allowing them to rotate, and then rejoining them. Without topoisomerase activity the fork would stall and the DNA ahead of it would become positively supercoiled.
解旋会在复制叉前方产生扭转张力，因为分子其余部分会抵抗旋转。拓扑异构酶（细菌中的 DNA 促旋酶）通过切断一条或两条链、使其旋转后再重新连接，从而消除这种张力。若缺乏拓扑异构酶活性，复制叉会停滞，其前方的 DNA 会形成正超螺旋。
Single-strand binding proteins (SSB proteins) then coat the exposed single strands. They keep the two strands apart so that they cannot re-anneal, protect the single-stranded DNA from nuclease attack, and hold the template in an extended conformation that the polymerase can read. SSB proteins are also displaced as the polymerase advances.
随后，单链结合蛋白（SSB 蛋白）覆盖暴露出来的单链。它们使两条链保持分离、防止重新退火，保护单链 DNA 免受核酸酶攻击，并把模板维持在被聚合酶读取所需的伸展构象中。随着聚合酶向前推进，SSB 蛋白也会被逐步取代。
7. Primase and the RNA Primer | 引物酶与 RNA 引物
DNA polymerase cannot start a new strand from nothing. It can only extend an existing chain by adding a nucleotide to a free 3'-OH group. This is why a short stretch of RNA, about 10 nucleotides long, must be laid down first. The enzyme primase synthesises this RNA primer on the single-stranded template, using the template sequence to choose the correct ribonucleotides.
DNA 聚合酶无法从无到有地起始一条新链，它只能通过把核苷酸加到游离的 3'-OH 基团上来延伸已有的链。因此必须先合成一小段约 10 个核苷酸的 RNA。引物酶在单链模板上合成这段 RNA 引物，并依据模板序列选择正确的核糖核苷酸。
Once the primer is in place, the 3'-OH it provides becomes the starting point for DNA synthesis. The RNA primer is later removed, usually by a specialised polymerase or an RNase activity, and the resulting gap is filled with DNA and sealed. Because each new fragment on the lagging strand needs its own primer, primase must act repeatedly during the synthesis of that strand.
一旦引物就位，它所提供的 3'-OH 就成为 DNA 合成的起点。之后这段 RNA 引物会被去除——通常由专门的聚合酶或 RNase 活性完成——留下的缺口由 DNA 填补并封接。由于后随链上的每个新片段都需要各自的引物，引物酶在该链合成过程中必须反复发挥作用。
Template strand + RNA primer (≈10 nucleotides) → free 3'-OH → DNA polymerase extends the chain
8. DNA Polymerases: Synthesis, Direction and Proofreading | DNA 聚合酶：合成、方向与校对
DNA polymerase is the core enzyme of replication. It selects the correct deoxyribonucleoside triphosphate (dNTP) that is complementary to the base on the template, positions it against the template, and catalyses the formation of a phosphodiester bond between the 3'-OH of the growing strand and the 5'-phosphate of the incoming nucleotide. The two terminal phosphates are released as pyrophosphate, and the reaction requires Mg²⁺ ions at the active site.
DNA 聚合酶是复制的核心酶。它选择与模板碱基互补的正确脱氧核苷三磷酸（dNTP），将其对准模板，并催化新生链的 3'-OH 与进入核苷酸的 5'-磷酸之间形成磷酸二酯键。末端两个磷酸以焦磷酸形式释放，该反应在活性位点需要 Mg²⁺ 离子。
Synthesis therefore always proceeds in the 5' to 3' direction, because the growing strand is extended at its 3' end. In E. coli, DNA polymerase III is the main replicative enzyme, while DNA polymerase I removes the RNA primers and fills the resulting gaps. Eukaryotes use a division of labour: polymerase α with primase initiates, polymerases δ and ε carry out the bulk of elongation on the lagging and leading strands respectively, and polymerase γ replicates mitochondrial DNA.
因此，合成总是沿 5'→3' 方向进行，因为新生链是在其 3' 端被延长的。在大肠杆菌中，DNA 聚合酶 III 是主要的复制酶，而 DNA 聚合酶 I 负责去除 RNA 引物并填补由此产生的缺口。真核生物则存在分工：聚合酶 α 与引物酶共同起始，聚合酶 δ 和 ε 分别主要负责后随链和前导链的延伸，聚合酶 γ 负责线粒体 DNA 的复制。
Fidelity comes from two layers of quality control. First, the polymerase itself has a 3' to 5' exonuclease activity that acts as a proofreading function: if an incorrect nucleotide is added, the mismatch distorts the double helix, the enzyme pauses, and the wrong nucleotide is excised before synthesis continues. Second, mismatch repair proteins scan the newly made DNA after replication and correct errors that proofreading missed, raising overall accuracy to about one error per 10⁹ bases.
保真性来自两层质量控制。首先，聚合酶本身具有 3'→5' 核酸外切酶活性，起校对作用：如果加入了错误的核苷酸，错配会使双螺旋结构变形，酶随即暂停，并在合成继续之前切除错误核苷酸。其次，错配修复蛋白在复制完成后扫描新合成的 DNA，纠正校对未能清除的错误，使总体准确度提高到约每 10⁹ 个碱基一个错误。
9. Leading and Lagging Strands, Okazaki Fragments and Ligase | 前导链、后随链、冈崎片段与连接酶
Because the two template strands run in opposite directions and polymerase works only 5' to 3', the two new strands are made differently. On the leading strand, the template is exposed in the 3' to 5' direction towards the fork, so the polymerase can follow the fork and add nucleotides continuously. On the lagging strand, the template runs the other way, so synthesis must occur in short pieces that are made away from the direction of fork movement, in a discontinuous, back-stitching manner.
由于两条模板链方向相反，而聚合酶只能沿 5'→3' 工作，两条新链的合成方式不同。在前导链上，模板以 3'→5' 方向朝向复制叉暴露，因此聚合酶可以跟随复制叉连续添加核苷酸。在后随链上，模板方向相反，因此合成必须以短片段方式进行，这些片段朝背离复制叉移动的方向合成，表现为不连续的"回缝"方式。
These short pieces are called Okazaki fragments. They are roughly 1000-2000 nucleotides long in prokaryotes and about 100-200 nucleotides long in eukaryotes, reflecting the fact that eukaryotic cells have many more origins and therefore shorter intervals between them. Each fragment has its own RNA primer at its 5' end. After the primers are removed and the gaps filled with DNA, an enzyme called DNA ligase catalyses the formation of the final phosphodiester bond, joining the fragments into one continuous strand.
这些短片段被称为冈崎片段。在原核生物中长度约为 1000-2000 个核苷酸，在真核生物中约为 100-200 个核苷酸，这反映了真核细胞具有更多复制起点、因此两个起点之间的间隔更短。每个片段的 5' 端都有自己的 RNA 引物。引物被去除、缺口由 DNA 填补之后，DNA 连接酶催化形成最后一个磷酸二酯键，把各片段连接成一条连续的新链。
Overall, replication is described as semi-discontinuous: continuous on the leading strand and discontinuous on the lagging strand. The whole set of proteins working together at the fork - helicase, primase, polymerases, SSB proteins and ligase - is called the replisome. Because it is a complex machine with many moving parts, it is worth drawing the fork in an exam and labelling the direction of each strand and the position of each enzyme.
总体而言，复制被称为半不连续复制：前导链连续合成，后随链不连续合成。在复制叉处协同工作的整套蛋白质——解旋酶、引物酶、各种聚合酶、单链结合蛋白和连接酶——被称为复制体（replisome）。由于它是一台包含许多活动部件的复杂机器，考试时值得画出复制叉，并标注每条链的方向和各个酶的位置。
10. Prokaryotic and Eukaryotic Replication Compared | 原核与真核复制的比较
Prokaryotic and eukaryotic replication share the same fundamental chemistry but differ in scale, location and the problems they must solve. Prokaryotes replicate a single circular chromosome with one origin, at speeds of up to about 1000 nucleotides per second, and can complete a round of replication in around 40 minutes. Eukaryotes replicate multiple linear chromosomes with many origins, at a slower rate of roughly 50 nucleotides per second, and confine replication to the S phase of interphase.
原核与真核复制的基本化学原理相同，但在规模、发生场所和需要解决的问题上存在差异。原核生物复制单个环状染色体，只有一个复制起点，速度
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图