Biotechnology: Principles and Processes

Biotechnology means using living organisms, or cells and enzymes taken from them, to make products and provide services that are useful to us. In that broad sense it is very old. Curdling milk with Lactobacillus, raising dough with yeast, brewing and making cheese are all biotechnology, and people were doing them long before anyone knew what a cell was. What changed over the last fifty years is that we no longer have to accept an organism exactly as we find it. We can now take one chosen gene out of one organism, put it into another, and make the second organism produce something it could never have produced on its own. A bacterium that carries the human insulin gene and manufactures human insulin inside a steel tank is the clearest picture of that difference.

The European Federation of Biotechnology has framed a definition wide enough to cover both the old and the new side of the subject: biotechnology is the integration of natural science and organisms, cells, parts thereof and molecular analogues, for products and services. Traditional biotechnology sits comfortably inside that sentence, and so does gene cloning.

In practice, all of modern biotechnology stands on two core techniques, and you should be able to name and explain both.

The first is genetic engineering. This is the set of techniques used to change the chemistry of the genetic material, that is DNA and RNA, and then to introduce the changed material into a host organism so that the phenotype of the host changes. Ordinary sexual reproduction also makes new combinations of genes, but it does so blindly. You get the whole genome of both parents shuffled together, and you cannot pick out one gene and leave the remaining thousands behind. Genetic engineering can move a single, deliberately chosen gene, and it can move it across species boundaries that no cross could ever bridge, such as a bacterial gene into a plant or a human gene into a bacterium.

The second technique sounds far less exciting but it is genuinely half the subject: maintaining a sterile, contamination-free environment in chemical engineering processes, so that only the microbe or cell we want grows, grows in large quantity, and makes the product. A gene that behaves perfectly in a test tube earns nothing if the thousand-litre tank it is scaled up into gets taken over by a stray mould. Antibiotics, vaccines, enzymes and recombinant proteins are all manufactured under conditions where the vessel, the pipes, the medium and even the air bubbled into the tank have been sterilised first.

Genetic engineering itself rests on one further idea that is easy to miss. If a piece of foreign DNA is simply pushed into a cell, the cell will not copy it, and it will be diluted away and lost as the cell divides. To survive and multiply, the foreign DNA has to become part of a molecule that the cell's own machinery already knows how to replicate. That single requirement is why every recombinant DNA experiment needs a vector. It is also why the three basic steps of recombinant DNA technology are stated in this order: first, identification of the DNA carrying the desirable gene; second, introduction of that identified DNA into the host; and third, maintenance of the introduced DNA inside the host and its transfer to the host's progeny.

Restriction enzymes are the tools that made gene cloning practical, and they came out of a question about bacteria and their viruses. In 1963 two enzymes were found in Escherichia coli that together restrict the growth of bacteriophages inside the cell. One of them adds methyl groups to DNA; the other cuts DNA. The methylating enzyme marks the bacterium's own DNA as its own, so the cutting enzyme spares it and destroys the unmethylated viral DNA instead. The cutting enzyme is the restriction endonuclease, and the first one shown to depend on a specific nucleotide sequence was Hind II, characterised a few years after that discovery. Hind II always cuts DNA at a particular point wherever it meets a specific sequence of six base pairs, and that sequence is called its recognition sequence. Since then more than nine hundred restriction enzymes have been isolated from over two hundred and thirty strains of bacteria, and between them they recognise a very large number of different sequences.

Restriction enzymes belong to a wider group called nucleases. Exonucleases chew nucleotides off the free ends of a DNA molecule, shortening it from the outside. Endonucleases cut at specific positions inside the molecule and leave the ends untouched. Restriction enzymes are endonucleases, which is precisely what you need when the gene you want sits in the middle of a long chromosome.

The names look cryptic but they follow a rule. In EcoRI, the first letter E comes from the genus Escherichia; the next two letters co come from the species coli; the letter R comes from the strain RY13; and the Roman numeral I shows that it was the first such enzyme isolated from that strain. Work through the name of any restriction enzyme in the same way and it stops being a random string of letters.

Each restriction endonuclease slides along the DNA and cuts only where it finds its own recognition sequence, and that sequence is always palindromic. A palindrome in DNA is not quite the same idea as a palindrome in a word. It means a sequence that reads exactly the same on both strands, provided each strand is read in the same direction, that is from its own 5' end towards its 3' end. Take the EcoRI site. Read the top strand from its 5' end and you get 5'—GAATTC—3'. Read the bottom strand from its own 5' end and you get 5'—GAATTC—3' again. Written as a pair it looks like this:

5' - G A A T T C - 3'
3' - C T T A A G - 5'

EcoRI cuts between the G and the A on each strand. Notice two things. The cut is not exactly at the centre of the palindrome, it is a little away from the centre; but it falls between the same two bases on both strands. Because the two cuts are staggered rather than opposite each other, every fragment ends in a short single-stranded stretch, here the four bases AATT, hanging off a double-stranded body. These overhangs are called sticky ends. The overhang left on one fragment is complementary to the overhang left on any other fragment cut by the same enzyme, so the two can pair up through hydrogen bonds whenever they meet. That hydrogen bonding is weak and temporary on its own; it simply holds the two pieces together long enough for DNA ligase to seal the sugar-phosphate backbone permanently. Some enzymes cut both strands at the very same position, straight through the centre, and leave flush blunt ends with no overhang at all, which are far harder to join because nothing holds the pieces in register.

This is the whole reason the same restriction enzyme is used on the source DNA and on the vector DNA. Cut both with EcoRI and the gene and the opened vector carry the same AATT overhangs, so they find each other in the ligation mixture instead of drifting past.

The first recombinant DNA molecule was constructed in 1972 by Stanley Cohen and Herbert Boyer. They cut out a piece of DNA carrying an antibiotic resistance gene from a plasmid, and used DNA ligase to link it to a native plasmid of Salmonella typhimurium. The result was a new circular molecule able to replicate on its own, and when it was transferred into Escherichia coli, a bacterium closely related to Salmonella typhimurium, the host replicated it and made multiple copies. That native plasmid, once it carried the new gene, was acting as a cloning vector: a DNA molecule that carries a foreign gene into a host cell and makes sure the host replicates it.

Plasmids and bacteriophages are used as vectors because they already replicate inside bacterial cells independently of the control that governs the bacterial chromosome. Their copy number matters. Bacteriophages, with their naturally high number per cell, can give very high copy numbers of the DNA they carry. Among plasmids, some are present as only one or two copies per cell while others run to fifteen to a hundred copies per cell. If you want a lot of the gene product, you clone into a vector with a high copy number.

Three features are what turn an ordinary piece of DNA into a usable cloning vector.

Origin of replication (ori). This is the sequence at which replication starts. Any piece of DNA linked to this sequence can be made to replicate inside the host cell. The ori also controls the copy number of the DNA linked to it, so it decides how many copies of your gene each cell will end up holding.

Selectable marker. Transformation is the process by which a piece of DNA is taken up by a host bacterium, and it is very inefficient: only a tiny fraction of the cells on the plate actually take up the vector. The selectable marker lets you identify and eliminate the non-transformants while allowing the transformants to grow. Genes for resistance to antibiotics such as ampicillin, chloramphenicol, tetracycline or kanamycin are the usual markers for E. coli. Plate the cells on a medium containing the antibiotic and only the cells carrying the vector survive.

Cloning sites. The vector needs recognition sites for commonly used restriction enzymes, and it needs them to be few, preferably one site per enzyme. If a vector had several sites for the same enzyme, cutting it would break it into several fragments and gene cloning would become a mess. In the well-known vector pBR322, the foreign DNA is ligated at a restriction site that lies inside one of the two antibiotic resistance genes. pBR322 carries an ampicillin resistance gene (ampR) and a tetracycline resistance gene (tetR). The BamH I and Sal I sites lie within tetR, while the Pst I and Pvu I sites lie within ampR.

Suppose you ligate the foreign DNA at the BamH I site, which sits inside tetR. The insert interrupts the tetracycline resistance gene, so the recombinant plasmid can no longer make a working product from it and the cell becomes sensitive to tetracycline, while ampR is untouched and still works. So you first plate the transformed cells on a medium containing ampicillin: everything that grows has taken up a plasmid of some kind. Then you transfer those colonies onto a medium containing tetracycline. Colonies that grow on ampicillin but fail to grow on tetracycline are the recombinants. Colonies that grow on both are non-recombinants, that is cells which took up a plasmid that closed up again without an insert. Here one resistance gene selects transformants and the other one gets inactivated by the insertion and thereby reveals the recombinants.

That two-plate method works but it is tedious. An alternative uses colour instead. The recombinant DNA is inserted inside the coding sequence of the enzyme β-galactosidase. The insert destroys the reading of that gene, so no functional enzyme is made, and this is called insertional inactivation. When such bacteria are grown on a medium containing a chromogenic substrate, the cells that carry a plasmid without an insert make working β-galactosidase, break down the substrate and form blue colonies; the cells that carry an insert make no working enzyme and form white colonies. White colonies are the recombinants, and you read the answer off a single plate.

Bacterial plasmids are not the only vectors. Agrobacterium tumefaciens is a pathogen of several dicot plants that naturally delivers a piece of DNA called T-DNA into plant cells, turning them into a tumour that manufactures chemicals the pathogen needs. Its tumour inducing (Ti) plasmid has been modified so that it is no longer able to cause disease, but still delivers genes of our choice into plant cells. In the same spirit, retroviruses, which naturally transform normal animal cells into cancerous ones, have been disarmed and are now used to carry desirable genes into animal cells.

Once the recombinant DNA has been assembled in a tube, it still has to get inside a living cell, and that is not straightforward. DNA is a hydrophilic molecule, heavily charged along its phosphate backbone, and it cannot simply slip across a hydrophobic cell membrane. The bacterial cell therefore has to be made competent, that is, made able to take up DNA from its surroundings.

The standard method is chemical. The bacterial cells are treated with a specific concentration of a divalent cation, usually calcium, which increases the efficiency with which DNA can enter the bacterium through pores in its cell wall. The recombinant DNA is then added and the cells are incubated with it on ice. Next the mixture is placed briefly at 42 °C, which is the step called heat shock, and then put back on ice. The sudden change in temperature is what actually drives the DNA in. Cells that come through this treatment with the plasmid inside are the transformed cells.

There are other ways to get DNA into cells, chosen according to the type of cell. In micro-injection, the recombinant DNA is injected directly into the nucleus of an animal cell using an extremely fine needle. For plant cells, which have a tough cell wall, biolistics or the gene gun is used: the cells are bombarded with high velocity microparticles of gold or tungsten coated with DNA, and the particles carry the DNA through the wall and into the cell. The third route is the disarmed pathogen vector already described, which is allowed to infect the cell and delivers the recombinant DNA itself as part of its normal infection.

Two enzymes do the actual chemistry of gene manipulation, and they are easy to mix up because both act on DNA.

DNA ligase joins two pieces of DNA by sealing the breaks in the sugar-phosphate backbone. It does not create new nucleotides and it does not read a template; it forms the phosphodiester bond that turns two loose pieces held together by hydrogen bonds into one continuous molecule. This is the same enzyme that seals the nicks between Okazaki fragments during normal DNA replication. If sticky ends have already paired the gene and the vector, ligase is the enzyme that makes the join permanent, which is why it is sometimes described as the molecular glue while restriction enzymes are the molecular scissors.

DNA polymerase does something quite different. It reads an existing single strand as a template and builds a new complementary strand alongside it, adding nucleotides one at a time to the free 3' end of a growing chain. It cannot start a chain from nothing; it needs a short pre-existing stretch of nucleotides, a primer, base paired to the template, and then it extends from there in the 5' to 3' direction. This is the enzyme that makes amplification by PCR possible.

Keep the division of labour clear in your head. Restriction enzymes cut. Ligase joins two existing pieces. Polymerase copies, making new DNA from free nucleotides. All three are needed, and none of them can do the job of the other two.

The first practical step in the laboratory is to get pure DNA out of the cells, because DNA inside a cell is wrapped in membranes and tangled with everything else the cell contains.

Isolation of the genetic material. The cells have to be broken open so that the DNA is released along with RNA, proteins, polysaccharides and lipids. Which enzyme you use to break the cell depends on what the cell wall is made of, and these three are easy to interchange by mistake. Lysozyme is used for bacteria, because it attacks the peptidoglycan of the bacterial wall. Cellulase is used for plant cells, whose walls are made of cellulose. Chitinase is used for fungal cells, whose walls contain chitin. Once the cell is open, the DNA is still not pure. Genes lie on long DNA molecules that are intertwined with proteins such as histones, so the DNA has to be freed from them. RNA is removed by treating with ribonuclease; proteins are removed by treating with protease. Other contaminants are taken out by suitable treatments, and finally chilled ethanol is added to the solution. DNA is not soluble in ethanol, so it precipitates out and can be seen as a collection of fine white threads suspended in the liquid. That visible mass of threads is purified DNA.

Cutting the DNA at specific locations. Restriction enzyme digestion is done by incubating the purified DNA with the chosen restriction enzyme under the conditions of temperature and buffer at which that particular enzyme works best. The same enzyme is used to digest the vector DNA in a separate reaction, so that both the gene fragment and the opened vector carry matching sticky ends. The cut gene fragment and the cut vector are then mixed and joined with DNA ligase, and the result is a recombinant DNA molecule.

How do you know the digestion has actually happened, and that you have the fragment you want? You run the products on an agarose gel, a technique called gel electrophoresis. Agarose is a natural polymer obtained from seaweed, and when set it forms a mesh of pores. DNA is negatively charged because of the phosphate groups along its backbone, so when an electric field is applied across the gel, every DNA fragment moves towards the positive electrode, the anode. As the fragments push through the mesh, the gel acts as a sieve. Small fragments slip through easily and travel far; large fragments are held back and stay near the well where they were loaded. So after some time the mixture has been resolved into bands, each band being a group of fragments of one particular size, arranged with the largest nearest the starting point and the smallest furthest away.

DNA is colourless, so the separated fragments cannot be seen in the gel as they are. The gel is stained with ethidium bromide, which slides in between the base pairs of DNA, and is then exposed to ultraviolet radiation. Wherever there is DNA, a bright orange band glows. Any band you want can then be cut out of the gel and the DNA recovered from that gel piece, a step called elution. The eluted, purified fragments are what get used to build recombinant DNA with a cloning vector.

Notice the logic running through this section: break the cell open, strip away everything that is not DNA, cut the DNA precisely, and then separate the pieces by size so you can pick out the exact one you want. Each step removes some unwanted material, and by the end you are holding a single defined fragment instead of the whole genome.

Often the gene you want is present as a single copy in a huge genome, which is far too little to work with. The polymerase chain reaction, or PCR, solves this by making many copies of that one chosen segment in a test tube, without any living cell being involved. It uses two sets of primers, which are small chemically synthesised oligonucleotides complementary to the regions of DNA that flank the segment you want, along with the enzyme DNA polymerase, a supply of free nucleotides, and the genomic DNA as template.

Each cycle of PCR has three steps, and they always come in this order.

Denaturation. The reaction mixture is heated to a high temperature, typically around 94 to 95 degrees Celsius. The heat breaks the hydrogen bonds holding the two strands of the double helix together, and the DNA separates into two single strands. Each of these single strands is now available to act as a template. Nothing is cut and no bond in the backbone is broken; only the base pairing is undone.

Annealing. The mixture is cooled, typically to somewhere around 50 to 60 degrees Celsius depending on the primers being used. At this lower temperature the two primers base pair with their complementary sequences on the two separated template strands. One primer binds to one strand and the second primer binds to the other, and because they sit at either end of the target region, they mark out exactly which stretch of DNA will be copied. Choosing the primers is therefore how you decide what PCR will amplify.

Extension. The temperature is raised to about 72 degrees Celsius, which is the temperature at which the polymerase works best. The enzyme attaches at the 3' end of each primer and extends it, adding nucleotides one after another in the 5' to 3' direction, using the template strand to decide which base comes next. By the end of this step each of the two original strands has a new complementary partner, so one double stranded molecule has become two.

Now repeat. Every cycle doubles the number of copies of the target segment, so the increase is exponential rather than additive. If the cycle is repeated many times, the segment can be amplified to roughly a billion copies, which is more than enough to work with, to run on a gel, or to ligate into a vector for further cloning.

There is one problem hidden in this scheme. Every cycle begins by heating the tube to about 95 degrees Celsius, and an ordinary DNA polymerase is a protein that would be denatured and destroyed by that heat, so fresh enzyme would have to be added by hand at every single cycle. The reaction became practical only with a thermostable DNA polymerase, the well-known example being Taq polymerase, isolated from the bacterium Thermus aquaticus, which lives in hot springs. Taq polymerase stays active through the high temperature denaturation step, so the enzyme is added once at the start and survives all the cycles. This one property is what allowed PCR to be automated in a machine that simply cycles the temperature.

Be careful about what PCR is and is not. PCR amplifies a DNA segment chemically in a tube; cloning in a vector amplifies it biologically inside a living host cell that also replicates it as it divides. PCR is fast and needs no cells, but it copies only the stretch that lies between the two primers, and you have to know the flanking sequences in advance to design those primers. Cloning is slower but gives you a gene sitting inside a host that can go on to express it as a protein.

Once the recombinant DNA has been forced into competent host cells, only a small proportion of the cells actually took it up, and the culture must be sorted out before anything else can be done. This is where the selectable marker earns its place. Suppose the recombinant DNA carries a gene for resistance to ampicillin. Any E. coli cell that has taken it up is now an ampicillin resistant cell, and any cell that has not remains sensitive. Spread the whole treated culture on agar plates containing ampicillin: the transformed cells grow into colonies and the untransformed recipient cells die. Every colony on that plate is therefore a transformant. It is because it allows this selection that the ampicillin resistance gene is called a selectable marker. Recombinants are then picked out from among these transformants by insertional inactivation, either by replica plating on a second antibiotic or by looking for white rather than blue colonies on a chromogenic substrate.

The point of all this work is usually a protein, and a protein produced by a gene expressed in a foreign host is called a recombinant protein. The selected cells can be grown on a small scale in the laboratory, and the desired protein extracted from the culture and purified using various separation techniques. That is fine for research, but it will never supply a market. Small volume cultures simply cannot yield appreciable quantities of the product.

One improvement is to change how the cells are cultured. In a continuous culture system, spent medium is drained out from one side of the vessel while fresh medium is added from the other, so nutrients never run out and waste never builds up. The cells are held in their physiologically most active exponential phase instead of drifting into a stationary phase. This produces a larger biomass and therefore a higher yield of the desired protein than a batch culture, in which a fixed volume of medium is inoculated once and the culture is left to run its course.

For real production scale, bioreactors are used. A bioreactor is a vessel in which raw materials are biologically converted into specific products, individual enzymes and so on, using microbial, plant, animal or human cells. It handles large volumes, of the order of a hundred to a thousand litres of culture. Its job is to hold every condition the cells need at the optimum: temperature, pH, substrate, salts, vitamins and oxygen.

The commonest design is the stirred-tank reactor. It is usually cylindrical, or has a curved base, so that the contents mix properly rather than settling in a corner. A stirrer inside keeps the culture evenly mixed and, just as importantly, keeps oxygen distributed throughout the vessel instead of only near the surface. In a variation of the design, sterile air is bubbled, or sparged, through the reactor to increase the oxygen supply further. A working bioreactor also has an agitator system, an oxygen delivery system, a foam control system, a temperature control system, a pH control system, and sampling ports through which small volumes of the culture can be withdrawn from time to time for testing without opening the vessel and risking contamination.

When the biosynthetic stage is over, the product is not yet a product. It is sitting in a tank mixed with cells, cell debris and spent medium. Everything that happens after this point is downstream processing. It covers separation and purification of the product from that mixture, and then formulation with suitable preservatives so that the product is stable enough to be stored and transported. In the case of a drug, the formulation must then pass through thorough clinical trials, and strict quality control testing is required for every product. Downstream processing and quality control testing are not the same for all products; they vary from product to product, because a bulk enzyme, a vaccine and an injectable protein each have different requirements. A product that is perfectly made inside the reactor and then poorly purified afterwards is still a failure, which is why downstream processing is treated as a full stage of the process rather than an afterthought.