On a summer morning in 1972, in a laboratory at Stanford University, biochemist Paul Berg completed an experiment that would change biology forever. He had successfully joined DNA from two different organisms — a bacterial virus and the monkey tumor virus SV40 — creating what he called "recombinant DNA." For the first time in history, genetic material from different species had been deliberately combined in a test tube.

Berg's colleagues were alarmed. The following year, he organized the Asilomar Conference — an extraordinary moment in scientific self-governance — where researchers voluntarily paused recombinant DNA experiments to assess the risks. The moratorium lasted less than two years. By the late 1970s, genetically modified bacteria were producing human insulin, ending dependence on insulin extracted from pig and cattle pancreases. By the 1990s, genetically modified crops were entering commercial agriculture. By 2023, a CRISPR-based therapy had been approved to treat sickle cell disease, curing patients who had suffered agonizing pain crises their entire lives.

In five decades, humanity went from first combining DNA from two organisms to editing specific single letters in the three-billion-letter human genome with molecular precision. The technology has outpaced ethics, regulation, and public understanding at nearly every step. That gap — between what is technically possible and what we have collectively decided to do with the capability — remains the central challenge of the genetic age.

"The ability to genetically engineer the human germline would be one of the most consequential developments in the history of medicine. We are not yet ready for it." — National Academies of Sciences, Human Genome Editing: Science, Ethics, and Governance (2017)

Key Definitions

DNA (Deoxyribonucleic acid) — The molecule that carries genetic information in virtually all living organisms. A double helix of two complementary strands, each made of nucleotides bearing one of four bases: adenine (A), thymine (T), guanine (G), and cytosine (C). Base pairing is specific: A pairs with T, G pairs with C. The sequence of bases encodes genetic information; the human genome contains approximately 3.2 billion base pairs.

Gene — A segment of DNA that encodes the instructions for making a functional RNA molecule or protein. The human genome contains approximately 20,000 to 25,000 protein-coding genes, comprising only about 1.5% of total DNA. The function of the remaining "non-coding" DNA includes regulatory sequences, structural roles, evolutionary remnants, and still-unknown functions.

Restriction enzyme — A bacterial enzyme that cuts DNA at specific recognition sequences (typically 4-8 base pairs long). The discovery and characterization of restriction enzymes in the 1970s made recombinant DNA possible — they provided molecular "scissors" to cut DNA at defined locations.

Recombinant DNA — DNA assembled from multiple sources, typically combining genetic material from different organisms. The foundation of genetic engineering: insulin genes inserted into bacteria, growth hormone genes inserted into yeast, and herbicide-resistance genes inserted into crops are all produced through recombinant DNA techniques.

Transgenic — An organism that has been genetically modified by inserting DNA from another species into its genome. Transgenic bacteria produce human insulin; transgenic mice carry human disease genes for research; transgenic crops (GMOs) carry genes conferring pest or herbicide resistance.

CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) — A naturally occurring system in bacteria that serves as a primitive immune memory: bacteria store fragments of viral DNA in their genome (in CRISPR arrays) to recognize and destroy viruses in future infections. Researchers Jennifer Doudna and Emmanuelle Charpentier (2012 Science paper) and Feng Zhang (2013) adapted this system into a general-purpose gene-editing tool.

Cas9 — The "CRISPR-associated protein 9," an enzyme that cuts DNA. Cas9 is guided to a target by a guide RNA and makes a double-strand break at the target sequence. Different Cas variants (Cas12, Cas13, CasX) have been discovered with different properties.

Guide RNA (gRNA) — A short RNA molecule designed to match a target DNA sequence. The gRNA binds to Cas9 and directs it to the correct location in the genome. By changing the gRNA sequence, researchers can target virtually any sequence in any genome.

PAM sequence (Protospacer Adjacent Motif) — A short DNA sequence (e.g., 5'-NGG-3' for S. pyogenes Cas9) that must be present immediately adjacent to the target sequence for Cas9 to make a cut. PAM sequences constrain which genomic locations can be targeted by a given Cas9 variant.

Homology-directed repair (HDR) — One mechanism cells use to repair double-strand DNA breaks. HDR uses a provided DNA template to make a precise correction — allowing specific gene edits. HDR is less efficient than the alternative NHEJ pathway and is most active in dividing cells.

Non-homologous end joining (NHEJ) — The more common DNA break repair pathway: the cell rejoins broken ends without a template, often introducing small insertions or deletions (indels). NHEJ is useful for disrupting gene function (knocking out a gene) but not for making precise corrections.

Somatic editing — Genetic modification of cells in a living individual. Changes affect only that person and are not heritable. All currently approved gene therapies use somatic editing.

Germline editing — Genetic modification of embryos, eggs, sperm, or early embryonic cells. Changes would be inherited by all descendants. Highly controversial; effectively banned or subject to moratorium in most countries.

Base editor — A CRISPR variant (developed by David Liu's lab, 2016) that converts one DNA base to another without making a double-strand break. CBEs (cytosine base editors) convert C to T; ABEs (adenine base editors) convert A to G. Base editors enable precise single-letter corrections with fewer off-target effects than standard Cas9 cutting.

Prime editor — A more recent CRISPR variant (Liu lab, 2019) that acts as a "search and replace" for DNA: it can make any type of edit (insertions, deletions, all 12 base-to-base conversions) using a pegRNA that both guides Cas9 to the target and serves as the repair template.

A Brief History of Genetic Engineering

The Discovery of DNA's Structure (1953)

James Watson and Francis Crick's 1953 description of DNA's double helix structure — built on crystallographic data from Rosalind Franklin and Maurice Wilkins — provided the molecular basis for genetics. If genes were sequences of bases, then changing those sequences should change the gene's function. The question was how.

Restriction Enzymes and Recombinant DNA (1970s)

Hamilton Smith discovered restriction enzymes in 1970; Herbert Boyer and Stanley Cohen produced the first recombinant DNA organism in 1973. Boyer and venture capitalist Robert Swanson co-founded Genentech in 1976 — the first biotechnology company. By 1982, Humulin (human insulin produced in bacteria) became the first recombinant DNA pharmaceutical approved by the FDA.

The basic technique: cut DNA from two organisms with the same restriction enzyme (creating complementary "sticky ends"), mix the fragments together, and join them with DNA ligase. The resulting recombinant DNA can be inserted into a host cell — often bacteria or yeast — which replicates it indefinitely.

PCR: Amplifying DNA (1983)

Kary Mullis invented the polymerase chain reaction in 1983 — a technique that amplifies specific DNA sequences millions of times, making them detectable and workable. PCR became foundational to genetic engineering, forensic DNA testing, COVID-19 diagnostics, and countless other applications. Mullis received the Nobel Prize in Chemistry in 1993.

Transgenic Crops (1990s)

The first genetically modified food crop approved for commercial sale was the Flavr Savr tomato (1994), engineered to ripen more slowly. This was quickly followed by Bt crops (engineered to produce a bacterial toxin fatal to certain insects), herbicide-resistant crops (Roundup Ready soybeans, 1996), and eventually golden rice (engineered to produce beta-carotene, a vitamin A precursor, addressing deficiency in rice-dependent populations).

By 2022, approximately 190 million hectares of GM crops were grown globally — primarily soybeans, maize, cotton, and canola in the US, Brazil, Argentina, Canada, and India.

The Human Genome Project (1990-2003)

The Human Genome Project, an international effort beginning in 1990, sequenced the entire human genome for the first time. The "working draft" was announced in 2000; the finished sequence in 2003. The project cost approximately $3 billion.

By 2023, whole-genome sequencing costs had fallen to approximately $200-$600 per genome — a price decline faster than Moore's Law. This revolution in sequencing has transformed medicine, evolutionary biology, and forensics.

CRISPR: The Breakthrough (2012-2013)

The CRISPR system had been noticed in bacterial genomes since the late 1980s, but its function as a bacterial immune system was only understood in 2007 (Barrangou et al.). Jennifer Doudna and Emmanuelle Charpentier published their landmark 2012 paper demonstrating that the Cas9 protein, guided by a synthetic gRNA, could cut any target DNA sequence in vitro. Feng Zhang's 2013 Science paper demonstrated CRISPR editing in human and mouse cells.

Doudna and Charpentier were awarded the Nobel Prize in Chemistry in 2020. The fundamental patent rights became the subject of a complex and ongoing legal battle between the University of California (Berkeley) and the Broad Institute (MIT/Harvard), with billions of dollars at stake.

How CRISPR Works: The Mechanism

Step 1: Design the Guide RNA

A researcher identifies the target sequence they want to edit — a specific gene, a specific mutation, a specific regulatory element. They design a guide RNA (gRNA) with a 20-nucleotide sequence complementary to the target. This is done computationally; the gRNA can be synthesized cheaply and quickly.

The target must be adjacent to a PAM sequence (for the most common Cas9, this is NGG — any nucleotide followed by two guanines). PAM sequences are common enough that most genomic targets are accessible.

Step 2: Deliver the CRISPR Components

The gRNA and Cas9 protein (or the DNA encoding them) must be delivered into the target cells. Delivery methods include:

  • Lentiviruses and adeno-associated viruses (AAV): Viral vectors that efficiently enter cells. Different AAV serotypes preferentially infect different tissues.
  • Lipid nanoparticles (LNPs): Fatty particles that encapsulate nucleic acids and fuse with cell membranes. Used to deliver mRNA-based CRISPR systems; the same technology used for COVID-19 mRNA vaccines.
  • Electroporation: Electrical pulses that temporarily open pores in cell membranes, allowing large molecules to enter. Used for cell therapies where cells are edited ex vivo (outside the body).
  • Direct injection: For some tissues (muscle, eye), direct local injection is sufficient.

Step 3: Finding the Target

Inside the cell, Cas9 (bound to the gRNA) diffuses through the nucleus, sampling DNA sequences. The gRNA hybridizes to complementary DNA at the target sequence. Cas9 checks for the PAM sequence and, if present, undergoes a conformational change.

Step 4: Cutting the DNA

Cas9 makes a blunt double-strand cut in the DNA precisely three base pairs upstream of the PAM sequence. Both strands are cut.

Step 5: Repair

The cell detects the break and initiates repair:

NHEJ (Non-Homologous End Joining): The cell reseals the break, but imperfectly — small insertions or deletions (indels) of 1-50 base pairs typically result. If this disrupts the gene's reading frame, the gene is knocked out (non-functional). This is useful for disabling a gene — for example, disrupting the BCL11A gene to reactivate fetal hemoglobin production in sickle cell disease.

HDR (Homology-Directed Repair): If the researcher provides a DNA repair template alongside the CRISPR components, the cell can use it to make a precise edit — correcting a point mutation, inserting a new sequence, or replacing a gene entirely. HDR is more precise but less efficient, working best in dividing cells.

From Laboratory to Clinic: Gene Therapy

Gene therapy aims to treat disease by modifying genes — correcting a mutation, silencing a harmful gene, or adding a therapeutic gene.

First Attempts and Early Setbacks

The first gene therapy clinical trial began in 1990: doctors treated a four-year-old girl named Ashanthi DeSilva with ADA-SCID (a severe immune deficiency) by inserting functional ADA genes into her T cells. The treatment partially worked.

The field suffered a devastating setback in 1999 when Jesse Gelsinger, an 18-year-old with a mild form of OTC deficiency, died from a massive immune reaction to the viral vector used in a gene therapy trial at the University of Pennsylvania. Clinical trials were halted, and the field retreated.

A second setback: the French X-SCID gene therapy trials (2000-2002) successfully treated children with "bubble boy disease" — but several developed leukemia because the viral vector had inserted near a proto-oncogene, activating it. Insertional oncogenesis became a recognized safety concern.

Modern Approvals

Improved vectors, better delivery methods, and more cautious clinical development have produced a wave of approved therapies:

Therapy Indication Mechanism Year Approved
Luxturna Retinal dystrophy (RPE65) AAV delivering correct RPE65 gene to retinal cells 2017 (US)
Zolgensma Spinal muscular atrophy AAV9 delivering SMN1 gene to motor neurons 2019 (US)
Kymriah B-cell leukemia CAR-T: engineer T cells to target CD19 antigen 2017 (US)
Hemgenix Hemophilia B AAV delivering Factor IX gene to liver cells 2022 (US)
Casgevy Sickle cell / beta-thalassemia CRISPR editing of BCL11A to reactivate fetal hemoglobin 2023 (UK/US)

Casgevy (developed by CRISPR Therapeutics and Vertex Pharmaceuticals) represents a historical milestone: the first CRISPR-based therapy approved for clinical use. It works by editing patients' own stem cells ex vivo — cells are extracted, BCL11A is disrupted by CRISPR (reactivating fetal hemoglobin), and the cells are transplanted back. Trials showed virtual elimination of pain crises in sickle cell patients.

The limitation is cost: Casgevy's list price is $2.2 million per patient. Gene therapy's transformation from experimental curiosity to curable option is real — but affordability and access remain profound barriers.

The Germline Controversy

The distinction between somatic and germline editing is ethically and biologically fundamental.

Somatic editing affects one individual. Germline editing — modifying embryos, eggs, or sperm — creates heritable changes that pass to all future generations. The benefits could be profound: eliminating hereditary diseases entirely from family lines. The concerns are also profound:

Consent: Future persons who will carry the edit cannot consent to it. The analogy to vaccines (which parents choose for children) fails because the modification is permanent and heritable.

Off-target risk: Current editing tools are not perfect. Unknown off-target edits could cause harm decades later or in future generations.

Enhancement pressure: Once germline therapy is accepted for disease, the boundary between treatment and enhancement becomes unclear. Selecting for disease prevention could slide toward selecting for intelligence, height, or athletic ability.

Equity: If heritable genetic enhancements are available only to the wealthy, genetic inequality compounds economic inequality across generations.

In November 2018, Chinese scientist He Jiankui announced he had created the first gene-edited babies — twin girls whose CCR5 gene had been disrupted to confer resistance to HIV infection. The announcement was met with near-universal condemnation. He had not followed ethical protocols, edited without clear medical necessity, and provided incomplete information to the families. He was subsequently sentenced to three years in prison by Chinese authorities.

The international scientific community has called for an effective moratorium on clinical germline editing pending development of governance frameworks and broader societal discussion.

What Genetic Engineering Does Not Change

Genetic engineering operates on genotype — the sequence of DNA. Phenotype — the observable characteristics of an organism — is the product of genotype interacting with environment and development. These are not the same thing.

Most traits of interest (intelligence, personality, height, disease risk) are polygenic — influenced by hundreds or thousands of variants, each contributing small effects. The genetic architecture of complex traits means there is no single "intelligence gene" to edit. The largest GWAS (genome-wide association study) for educational attainment has identified over a thousand variants that together explain only 10-15% of the variation in years of schooling.

The simplistic concept of "designer babies" optimized for intelligence or personality ignores this complexity. What genetic engineering can do well is eliminate or correct single-gene disorders (where one variant causes the disease) — and there are approximately 10,000 known single-gene disorders affecting millions of people.

For related concepts, see how evolution works, how vaccines work, and how artificial intelligence learns.

References

Tags

genetics CRISPR gene therapy biotechnology DNA genetic engineering science

Frequently Asked Questions

How does CRISPR-Cas9 actually edit genes?

CRISPR-Cas9 works in two parts: a guide RNA (gRNA) that acts as a molecular GPS — it is designed to match a specific DNA sequence in the genome — and the Cas9 protein, a molecular scissors that cuts DNA. The gRNA leads Cas9 to the target location; Cas9 makes a precise double-strand break in the DNA. The cell then repairs the break — either imperfectly (disrupting the gene) or using a provided template (making a specific edit). The system is adapted from a bacterial immune defense mechanism.

What is the difference between somatic and germline genetic editing?

Somatic editing changes genes in specific cells of a living person — the changes affect only that individual and cannot be inherited. Most current gene therapies are somatic. Germline editing changes genes in embryos, eggs, or sperm — the changes would be inherited by all subsequent generations. Germline editing is highly controversial and largely banned or suspended internationally. In 2018, Chinese scientist He Jiankui controversially created the first gene-edited babies (CCR5 modification intended to confer HIV resistance), drawing global condemnation.

What diseases can gene therapy treat?

Approved gene therapies (as of 2024) include treatments for spinal muscular atrophy (Zolgensma), beta-thalassemia and sickle cell disease (Casgevy — the first CRISPR therapy approved), hemophilia B (Hemgenix), retinal dystrophy (Luxturna), and some cancers (CAR-T cell therapies like Kymriah and Yescarta). Hundreds of clinical trials are ongoing for conditions including Huntington's disease, Duchenne muscular dystrophy, and various cancers. Gene therapy has moved from experimental to clinically validated, though treatments remain extremely expensive.

Are GMO foods safe to eat?

The scientific consensus, supported by the National Academies of Sciences (2016 report), WHO, and virtually every major scientific organization, is that currently approved genetically modified foods are as safe to eat as their conventional counterparts. No credible evidence exists of harm from consuming GM foods that have passed regulatory review. The debate about GMOs largely concerns environmental effects (herbicide-resistant weeds, biodiversity), corporate control of food systems, and labeling rights — not food safety per se.

What are off-target effects in CRISPR editing?

Off-target effects occur when CRISPR's guide RNA binds to sequences similar (but not identical) to the intended target, causing unintended cuts elsewhere in the genome. These unintended edits could potentially disrupt important genes or affect gene regulation. Early CRISPR tools had meaningful off-target rates; newer variants (high-fidelity Cas9, base editors, prime editors) have substantially reduced but not eliminated this problem. Off-target detection and minimization is a major area of ongoing research for clinical applications.

What is base editing and prime editing?

Base editing (David Liu, 2016) makes precise single-letter changes to DNA without cutting both strands — converting one DNA base to another (e.g., C to T) using a modified Cas9 fused to a chemical editing enzyme. Prime editing (Liu lab, 2019) is even more precise: a 'search and replace' system that can make insertions, deletions, or any base-to-base conversion with lower off-target effects. These 'next-generation' editing tools extend CRISPR's capabilities and reduce risks compared to original cut-and-repair approaches.

What ethical issues surround genetic engineering?

Key ethical concerns include: germline editing creates permanent heritable changes without consent of future generations; 'designer babies' could exacerbate inequality if genetic enhancements are available only to the wealthy; genetic selection raises concerns about discrimination against disability; enhancement vs. treatment boundaries are unclear. Most ethicists and scientific bodies support somatic therapy for serious disease while calling for a moratorium or extreme caution on germline enhancement. The distinction between treating disease and enhancing normal traits is central to the bioethical debate.

Share this article