The sample was loaded at the end where the largest, slowest-moving bands are — that is, at the cathode end, farthest from the anode.

The reasoning is that DNA fragments are negatively charged molecules and are separated by forcing them to move towards the anode under an electric field. They therefore all travel away from the wells in one direction. Because the fragments resolve according to size through the sieving effect provided by the agarose gel, the smaller the fragment size, the farther it moves — so the band nearest the loading point is the largest fragment, and the undigested DNA in lane 1 barely moves from the well at all.

Because of what a membrane is made of. A cell membrane is a bilayer whose interior is a hydrophobic region of fatty acid tails. Only small, non-polar, lipid-soluble molecules cross such a layer freely.

DNA fails on every count. It is a very large polymer; it carries a dense negative charge along its phosphate backbone; and it is strongly hydrated, that is hydrophilic. A charged, water-loving macromolecule cannot dissolve into or diffuse through a hydrophobic core.

This is precisely why hosts have to be made competent: treatment with a specific concentration of a divalent cation such as calcium increases the efficiency with which DNA enters the bacterium through pores in its cell wall, and the ice-heat shock-ice sequence then forces the recombinant DNA in. Where that will not work, DNA is delivered by micro-injection, by biolistics, or by disarmed pathogen vectors.

Because the quantities actually needed are enormous compared with what a laboratory culture yields. Small volume cultures cannot yield appreciable quantities of products. A single diabetic patient needs insulin every day for life, and there are hundreds of millions of such patients; a laboratory flask could not supply one hospital for one week.

Because purification loses most of what you make. The desired protein is a tiny fraction of the cell's total protein, and every step of separation and purification in downstream processing costs some of it. You must begin with far more than you intend to sell.

Because cost falls with scale. Biotechnology deals with large scale production and marketing. The vessel, the sterilisation, the monitoring and the trained staff cost nearly the same for 10 litres as for 1000, so the price per dose of a medicine is set largely by the scale at which it is made.

This is why bioreactors were developed, in which large volumes of 100 to 1000 litres of culture can be processed, and why cells are often grown in a continuous culture system to keep them in their most productive phase.

A recombinant protein is any protein-encoding gene expressed in a heterologous host. The following are in routine medical use.

What the list shows. Notice the pattern: most of these are human proteins that a patient cannot make enough of. Before recombinant DNA technology they had to be extracted from animal tissue, human cadavers or pooled blood — which meant tiny supplies, immune reactions to non-human protein, and the real risk of transmitting infection. Making the human protein in a heterologous host solved all three problems at once.

Taking EcoRI as the example.

DNA is very much bigger. An enzyme is a protein of a few hundred amino acids, with a molecular mass of the order of ten thousand to a hundred thousand daltons. A DNA molecule is a polymer of millions of nucleotides — the single circular chromosome of E. coli alone runs to about 4.6 million base pairs, and a human chromosome to far more.

How you can know it from this chapter, without being told.

(i) From the way they act on each other. Each restriction endonuclease functions by inspecting the length of a DNA sequence until it finds its recognition sequence. An enzyme that must travel along a molecule searching for a six-base site is obviously the smaller of the two; the DNA is the landscape, the enzyme the walker.

(ii) From how many cuts one molecule yields. A single DNA molecule can be cut by an enzyme into many fragments, each still large enough to be seen as a band on a gel. Something that can be divided into many visible pieces must be much larger than the tool dividing it.

(iii) From what you can see. Purified DNA precipitated with chilled ethanol can be seen as a collection of fine threads in the suspension — visible to the naked eye. No enzyme solution shows anything of the kind.

(iv) From the genetics. The enzyme is the product of a gene, and that gene is only one short stretch of the DNA molecule. The whole cannot be smaller than a product of one of its parts.

This is an order-of-magnitude calculation, and it is worth doing carefully because the answer is startling.

Step 1 — how many DNA molecules are in one cell? A human diploid somatic cell has 46 chromosomes, and each unreplicated chromosome is one DNA molecule. So there are 46 molecules of nuclear DNA per cell.

Step 2 — convert to moles. One mole contains 6.022 × 10²³ molecules, so

\( n = \dfrac{46}{6.022 \times 10^{23}} \approx 7.6 \times 10^{-23}\ \text{mol} \)

Step 3 — divide by the volume. Molar concentration is moles per litre. Taking a typical human cell volume of about 2000 µm³, which is 2 × 10⁻¹² litre,

\( c = \dfrac{7.6 \times 10^{-23}\ \text{mol}}{2 \times 10^{-12}\ \text{L}} \approx 4 \times 10^{-11}\ \text{mol L}^{-1} \)

The answer. Of the order of 10⁻¹¹ M, that is a few tens of picomolar. (If you assume a smaller cell, or calculate for the nucleus alone, you will get a somewhat higher figure; the order of magnitude is what matters.)

Why the number is worth thinking about. Water in the same cell is about 40 M. So the molecule that specifies the entire organism is present at roughly one part in 10¹² by concentration — among the rarest molecules in the cell. It is rare because it does not need to be abundant: a single copy is read repeatedly, and the RNA and protein products are what get made in quantity. It is also why PCR matters so much: amplifying a DNA sequence about a billion-fold is what takes a vanishingly dilute molecule into the range where chemistry can be done with it.

No — eukaryotic cells do not possess restriction endonucleases of the kind used in genetic engineering. They have plenty of other nucleases, for DNA repair, for recombination and for degrading DNA during programmed cell death, but not the sequence-specific restriction–modification system found in bacteria.

Justification 1 — the function these enzymes serve. Recall how they were discovered: the two enzymes isolated in 1963 were responsible for restricting the growth of bacteriophage in Escherichia coli. Restriction enzymes are a bacterial defence against invading phage DNA. Bacteria have no nuclear envelope and no immune system, so cutting up foreign DNA the moment it arrives is their protection.

Justification 2 — the system needs a partner, and it would be dangerous without one. One of the two enzymes isolated in 1963 added methyl groups to DNA. That methylase is what marks the cell's own DNA so it is spared. A eukaryote with a large genome full of recognition sites, but no such protection, would cut its own chromosomes to pieces. Maintaining the system would cost more than it is worth.

Justification 3 — eukaryotes defend themselves differently. A eukaryotic cell keeps its DNA inside a nucleus, behind a nuclear envelope. Multicellular eukaryotes add interferons, immune surveillance and RNA-based silencing of viral nucleic acid. The problem is solved by other means, so the bacterial solution was never needed.

The practical confirmation. All of the more than 900 restriction enzymes now available have been isolated from over 230 strains of bacteria. If eukaryotes had them, biotechnology would certainly have found them there too.

(i) Control of the growth conditions. A bioreactor provides the optimal conditions for achieving the desired product by providing optimum growth conditions — temperature, pH, substrate, salts, vitamins and oxygen. It has a dedicated temperature control system and pH control system, so neither drifts as the cells metabolise. A shake flask offers no such control.

(ii) Sampling without contamination. The bioreactor has sampling ports, so that small volumes of the culture can be withdrawn periodically. The progress of the run can therefore be followed without opening the vessel, which in a flask would risk losing the batch.

(iii) Foam control. Agitating a protein-rich medium generates foam, which carries cells out of the liquid and blocks filters. The bioreactor has a foam control system.

(iv) Volume, and therefore usable yield. Small volume cultures cannot yield appreciable quantities of products, whereas a bioreactor allows large volumes of 100 to 1000 litres of culture to be processed.

(v) Continuous operation. A bioreactor can be run as a continuous culture system, with used medium drained out from one side while fresh medium is added from the other, maintaining the cells in their physiologically most active log or exponential phase. This produces a larger biomass and higher yields of the desired protein. A flask is inherently a batch culture, which must exhaust its medium.

(vi) Sterility at scale. A closed, steam-sterilisable vessel with controlled air entry keeps the sterile, contamination-free ambience that bioprocess engineering requires — something impossible to guarantee in repeatedly opened flasks.

Five real examples — each is the recognition sequence of a restriction enzyme, and each reads the same on both strands in the 5′→3′ direction.

How to create one yourself. The rule is that the second half of the sequence must be the reverse complement of the first half.

1. Choose any three bases for the first half — say T G C.
2. Reverse them: C G T.
3. Complement each base (A↔T, G↔C): G C A.
4. Join the two halves: T G C G C A.

Check it. The bottom strand of 5′-TGCGCA-3′ is 3′-ACGCGT-5′. Read that bottom strand in the 5′→3′ direction, that is from right to left: T G C G C A. Identical to the top strand — so it is a genuine palindrome. (It is in fact the recognition site of Fsp I.)

Note the difference from a word-palindrome. MALAYALAM reads the same backwards on the same line. A DNA palindrome reads the same on the two strands, when the orientation of reading is kept the same. That is why the symmetry rule involves complementing as well as reversing.

During crossing over, in the pachytene stage of prophase I of meiosis.

Recall the sequence. In zygotene, homologous chromosomes pair by synapsis and form bivalents or tetrads. In pachytene, the four chromatids of each bivalent become distinct, recombination nodules appear at sites along the paired chromosomes, and crossing over takes place — the exchange of genetic material between the non-sister chromatids of homologous chromosomes. This exchange is enzyme-mediated, and the enzyme involved is recombinase. By diplotene, the recombined chromatids remain attached at the chiasmata, which is the visible evidence that the exchange has happened.

Why this counts as recombinant DNA. After crossing over, a single chromatid carries a stretch of DNA of maternal origin covalently joined to a stretch of paternal origin. That is exactly the definition of a recombinant DNA molecule — DNA composed of DNA from different sources joined into one continuous molecule.

The comparison worth drawing.

So nature had been making recombinant DNA long before 1972. What genetic engineering added was choice — over which sequences are joined, and over the species barrier.

What the two do differently. A selectable marker — typically an antibiotic resistance gene — separates transformants from non-transformants: it helps in identifying and eliminating non-transformants and selectively permitting the growth of the transformants. But it cannot tell a cell carrying a recombinant plasmid from one carrying a plasmid that re-closed empty, because both retain the resistance gene. A reporter enzyme fills exactly that gap.

How a reporter enzyme works. A reporter enzyme is one whose activity produces an easily detected signal, usually a colour, in the presence of a suitable substrate. The foreign DNA is inserted within the coding sequence of the reporter gene. If the insert is present, the reporter gene suffers insertional inactivation and no signal is produced; if the plasmid has no insert, the reporter gene is intact and the signal appears.

The standard example. Recombinant DNA is inserted within the coding sequence of the enzyme β-galactosidase. In the presence of a chromogenic substrate, colonies whose plasmid has no insert produce blue coloured colonies; where the insert is present, the β-galactosidase gene is inactivated and the colonies produce no colour. Those colourless colonies are identified as the recombinant ones.

Why this is better than the two-antibiotic method. Selection of recombinants by inactivation of antibiotic resistance is a cumbersome procedure, because it requires simultaneous plating on two plates having different antibiotics and a comparison between them. The reporter method gives the answer on a single plate, at a glance, in a few hours.

The two are used together, not instead of each other. The antibiotic marker first removes the vast majority of cells — those that took up nothing. The reporter then sorts the survivors into recombinant and non-recombinant. One plate does the coarse filter, the colour does the fine one.

(a) Origin of replication. A specific DNA sequence, present in a chromosome or a vector, from where replication starts, and which is responsible for initiating replication. Any piece of DNA linked to this sequence can be made to replicate within the host cells, which is why an alien gene must be joined to an ori before it can be multiplied and inherited. The sequence is also responsible for controlling the copy number of the linked DNA, so if one wants to recover many copies of the target DNA it should be cloned in a vector whose origin supports a high copy number.

(b) Bioreactors. Vessels in which raw materials are biologically converted into specific products, individual enzymes and so on, using microbial, plant, animal or human cells. They were developed because small volume cultures cannot yield appreciable quantities of products; a bioreactor processes large volumes of 100 to 1000 litres of culture. It provides the optimal conditions for achieving the desired product by supplying optimum growth conditions — temperature, pH, substrate, salts, vitamins and oxygen. The most commonly used are of the stirring type: a stirred-tank reactor, usually cylindrical or with a curved base to facilitate mixing, in which the stirrer facilitates even mixing and oxygen availability throughout, or alternatively air is bubbled through, giving the sparged stirred-tank type. A bioreactor has an agitator system, an oxygen delivery system, a foam control system, a temperature control system, a pH control system and sampling ports.

(c) Downstream processing. After completion of the biosynthetic stage, the product has to be put through a series of processes before it is ready for marketing as a finished product. These processes of separation and purification are collectively referred to as downstream processing. The product then has to be formulated with suitable preservatives; such a formulation has to undergo thorough clinical trials, as in the case of drugs; and strict quality control testing is required for each product. Both the downstream processing and the quality control testing vary from product to product.

(a) PCR. PCR stands for Polymerase Chain Reaction. In this reaction, multiple copies of the gene or DNA of interest are synthesised in vitro using two sets of primers — small chemically synthesised oligonucleotides complementary to the regions of DNA flanking the target — and the enzyme DNA polymerase, which extends the primers using the nucleotides provided in the reaction and the genomic DNA as template. Each cycle has three steps: denaturation, primer annealing and extension of primers. If the process is repeated many times, the segment of DNA can be amplified to approximately a billion times. Such repeated amplification is achieved by using a thermostable DNA polymerase, isolated from the bacterium Thermus aquaticus, which remains active during the high-temperature-induced denaturation of double-stranded DNA. The amplified fragment can then be ligated with a vector for further cloning.

(b) Restriction enzymes and DNA. Restriction enzymes belong to the class of enzymes called nucleases, of which exonucleases remove nucleotides from the ends of DNA while endonucleases make cuts at specific positions within the DNA. A restriction endonuclease functions by inspecting the length of a DNA sequence; once it finds its specific recognition sequence, it binds to the DNA and cuts each of the two strands of the double helix at specific points in their sugar-phosphate backbones. Each one recognises a specific palindromic nucleotide sequence, and cuts a little away from the centre of the palindrome site but between the same two bases on the opposite strands. This leaves single-stranded overhanging stretches called sticky ends, which form hydrogen bonds with their complementary cut counterparts, and so facilitate the action of DNA ligase. When cut by the same restriction enzyme, DNA fragments from different sources have the same kind of sticky ends and can be joined end-to-end to form recombinant DNA.

(c) Chitinase. An enzyme that digests chitin, the material of the fungal cell wall. It is used during the isolation of genetic material: since DNA is enclosed within membranes, the cell must be broken open to release it, and this is achieved by treating the tissue with enzymes such as lysozyme for bacteria, cellulase for plant cells and chitinase for fungus. Without opening the fungal wall, no DNA is released, and no later step can succeed.

(a) Plasmid DNA and Chromosomal DNA

(b) RNA and DNA

(c) Exonuclease and Endonuclease