The Biotechnology Toolkit: Restriction Enzymes, Plasmids and Recombinant DNA

Insulin, from Toronto to a vat of bacteria

In 1921 and 1922 at the University of Toronto, Frederick Banting, Charles Best, James Collip and J.J.R. Macleod discovered how to extract insulin and use it to treat diabetes. For about 60 years after that, the insulin people injected came from the pancreases of cattle and pigs. Then, in 1982, the first medicine made by genetically engineered bacteria was approved: human insulin, produced by E. coli carrying a human insulin gene.

How do you put a human gene into a bacterium and get it to make a human protein? This lesson covers the tools (expectation D3.6): restriction enzymes, plasmids and other vectors, ligase and recombinant DNA. The next lessons add PCR, gel electrophoresis, sequencing and CRISPR.

Molecular scissors: restriction enzymes

Bacteria are attacked by viruses (bacteriophages) that inject their DNA. Many bacteria defend themselves with restriction enzymes, which cut foreign DNA at specific sequences, "restricting" viral growth. The bacterium protects its own DNA by adding methyl groups at those same sequences, so its enzymes don't cut it. Werner Arber, Daniel Nathans and Hamilton Smith shared the 1978 Nobel Prize in Physiology or Medicine for discovering restriction enzymes and showing how to use them.

Each restriction enzyme recognizes one recognition site, usually 4 to 8 base pairs long, and most sites are palindromes: the sequence reads the same 5′→3′ on both strands. For example, both strands of the EcoRI site read 5′-GAATTC-3′.

• Sticky ends: many enzymes cut the two strands at different positions, leaving short single-stranded overhangs. EcoRI cuts G^AATTC, leaving 5′-AATT overhangs. Any two EcoRI ends, from any organism, can base-pair with each other.
• Blunt ends: some enzymes, such as SmaI (CCC^GGG), cut straight across, leaving no overhang.

How often does an enzyme cut?

In a random sequence, each position has a 1/4 chance of being any particular base. So a specific 6-base site appears on average once every 4⁶ = 4096 base pairs, and a 4-base site once every 4⁴ = 256 base pairs. Real genomes are not random, so actual spacings vary, but enzymes with longer sites cut less often and make bigger pieces.

Counting fragments:

• A linear DNA molecule cut at n sites gives n + 1 fragments.
• A circular molecule (such as a plasmid) cut at n sites gives n fragments.

Vectors: plasmids and more

A vector carries a gene into a host cell and gets it copied. The workhorse is the bacterial plasmid, a small circle of DNA (often a few thousand base pairs) that replicates separately from the chromosome. Plasmids used in labs are engineered to have:

• an origin of replication (ori), so the host copies the plasmid;
• a selectable marker, usually an antibiotic-resistance gene (such as ampR for ampicillin), so only cells with the plasmid survive on antibiotic;
• a multiple cloning site, a short stretch with recognition sites for many restriction enzymes, each found only once in the plasmid;
• often a reporter gene, such as part of lacZ, to show whether a gene was inserted.

Other vectors: viruses (modified so they cannot cause disease) carry genes into human and animal cells, for example in gene therapy; the plasmid of the soil bacterium Agrobacterium tumefaciens carries genes into plant cells; and artificial chromosomes carry very large pieces of DNA.

Making recombinant DNA

1. Cut both the vector and the DNA containing the gene with the same restriction enzyme, so they have matching sticky ends. (Using two different enzymes, one at each end, makes the gene go in the right way round.)
2. Mix: the sticky ends base-pair with each other.
3. Ligate: DNA ligase forms phosphodiester bonds to seal the sugar-phosphate backbones. The result is recombinant DNA.
4. Transform: put the plasmids into bacteria. Cells are made permeable with a calcium chloride solution and a brief heat shock, or with an electric pulse (electroporation). Only a small fraction of cells take up a plasmid.
5. Select: grow the bacteria on a plate containing the antibiotic. Only cells carrying a plasmid survive.
6. Screen: find the colonies whose plasmid actually contains the gene (for example, by blue-white screening).
7. Grow the chosen colony in large tanks (fermenters) and purify the protein.

Paul Berg made the first recombinant DNA molecule in 1972, and in 1973 Stanley Cohen and Herbert Boyer put recombinant plasmids into E. coli and showed the genes worked. Concern about safety led scientists to meet at Asilomar, California, in 1975 and agree on safety guidelines, an early example of scientists regulating themselves.

Two problems with human genes in bacteria

1. Introns. Human genes contain introns, and bacteria cannot splice them out. The solution is to start from the mature mRNA, which has no introns, and use the enzyme reverse transcriptase (from retroviruses) to make a DNA copy: complementary DNA (cDNA). (For short proteins such as insulin's two chains, scientists can also simply synthesize the coding DNA chemically.)

2. Expression and finishing. Bacteria need a bacterial promoter and a ribosome-binding site in front of the gene, so the gene is put into an expression vector that has them (often a lac promoter switched on with IPTG, as you saw in the operon lesson). Bacteria also cannot add sugar groups or make some of the modifications human proteins need, so complex proteins, such as many antibodies, are produced in yeast or in cultured mammalian cells (often hamster ovary cells) instead.