Biology › Gene regulation, genomics and biotechnology › Recombinant DNA: getting a gene out, in, and working

Recombinant DNA: getting a gene out, in, and working

The genetic code is read the same way by a bacterium and by you, so a human gene put into E. coli makes a human protein. Everything difficult about genetic engineering lies in the words 'put into': cutting the gene out cleanly, joining it to something that will be copied, and finding the one cell in a thousand that took it.

Before this Complementary base pairing and hydrogen bonds · Phosphodiester bonds and DNA polymerase · Introns, exons and splicing · The genetic code is almost universal

Before you start

Cut the gene out with a restriction enzyme, drop it into a plasmid, mix the plasmids with some bacteria, and you have bacteria making your protein. Every step in that sentence is real and the sentence as a whole is a fantasy, for four separate reasons. A restriction enzyme cuts where its recognition sequence occurs, not where a gene begins and ends, so the cut you want may not be available. A gene taken straight from a human chromosome carries introns, and a bacterium has no spliceosomes to remove them. Cut plasmid and cut gene will not stay together without ligase, and plenty of plasmids simply close up on themselves instead. And when you mix the two, well under one bacterium in a hundred takes up a plasmid at all — which is why the next thing you need is not more enzymes but a way of telling which cells succeeded.

What you should be able to do

Describe the three ways of obtaining a gene, and say which one suits which situation and why.
Explain what a sticky end is, why the same restriction enzyme must be used on gene and vector, and what ligase does afterwards.
Describe transformation and explain why marker genes are needed at all.
State the three temperatures of the polymerase chain reaction and explain what each step needs its own temperature for.
Calculate the number of copies produced by a given number of PCR cycles, and explain why a real reaction falls short of it.
Explain how gel electrophoresis separates DNA fragments, in terms of charge and of the gel itself.

Three ways to get hold of a gene

Before anything can be inserted anywhere, you need a copy of the gene as a piece of DNA with no introns in it. There are three routes, and the reason there are three is that each has a situation it suits.

Reverse transcriptase, working from mRNA. Find a cell that makes a great deal of the protein you want — the β cells of the pancreas for insulin, an immature red blood cell for globin — and it will be full of the corresponding mRNA. Extract that mRNA and treat it with reverse transcriptase, an enzyme obtained from retroviruses, which builds a DNA strand against an RNA template. The product is a single strand of complementary DNA, and DNA polymerase then builds the second strand against it, giving double-stranded DNA. The advantage is decisive: because the mRNA had already been spliced, the cDNA contains no introns and a bacterium can use it.

Restriction endonucleases, working from the chromosome. If a suitable recognition sequence occurs on each side of the gene, an enzyme can cut it out directly. This is quick and needs no mRNA, but it delivers the gene complete with its introns, so the product is only useful in a host that can splice — a yeast or a mammalian cell culture, not E. coli.

The gene machine. The sequence is looked up in a database, or worked backwards from the amino acid sequence of the desired protein using the codon table, and then the DNA is built chemically, a nucleotide at a time, in short lengths that are joined by PCR into the whole gene. Nothing biological is required at all. The gene has no introns because none were put in, and the codons can be chosen from among the synonymous alternatives to suit the host organism's preferences, which can raise the yield of protein substantially. This is now how most genes are obtained commercially.

Reverse transcriptase: An enzyme, obtained from retroviruses, which catalyses the synthesis of DNA from an RNA template.
Complementary DNA (cDNA): DNA made from an mRNA template by reverse transcriptase, and therefore containing no introns.
Recombinant DNA: DNA made by joining together sequences from two different organisms.
Transgenic organism: An organism containing DNA transferred into it from another species.

One reason any of this works at all is worth restating, because it is the fact the whole industry rests on. The genetic code is almost universal: a bacterium reads the triplet CCU as proline, exactly as a human cell does. Human insulin has been manufactured in genetically modified bacteria since 1982, and before that diabetic patients used insulin extracted from the pancreases of pigs and cattle, which differs slightly in sequence and provoked immune reactions in some people.

Sticky ends, and why they are sticky

Restriction endonucleases are bacterial enzymes that cut DNA at a specific sequence, usually four to eight base pairs long. Their real job is defence: they chop up the DNA of invading bacteriophages, and the bacterium protects its own genome by methylating it, which is the same chemistry you met silencing genes in the first lesson of this unit, put to an entirely different use.

The recognition sequence is normally palindromic, meaning it reads the same 5′ to 3′ on both strands. EcoRI recognises GAATTC; the strand opposite reads GAATTC too, when read in its own direction. Some enzymes cut both strands at the same point, leaving blunt ends. The useful ones cut the two strands at different points, leaving each fragment with a short single-stranded overhang: a sticky end.

The offset in the cut is the whole trick. Cut down the middle and you get two ends that will meet only by luck; cut off-centre and every fragment ends in four unpaired bases that can only pair with one sequence.

Sticky ends are useful because they are specific. Cut a human chromosome and a plasmid with the same enzyme and every fragment from both carries the same overhang, so the human fragment's exposed bases are complementary to the plasmid's. Mix them and the overhangs pair by hydrogen bonding. Those hydrogen bonds hold the two pieces in place but do not join the backbones, so the join is temporary until DNA ligase catalyses the formation of phosphodiester bonds along both strands. Naming the two kinds of bond separately is worth a mark on its own and is almost always available.

If no convenient site exists, the overhangs can be manufactured. Short single-stranded lengths of DNA carrying the wanted sequence are attached to the blunt ends of the fragment, and it then behaves as though it had been cut by that enzyme all along.

Vectors, transformation and the marker problem

A gene by itself in a bacterium is lost at the first division. It needs a vector: a piece of DNA that will be replicated by the host and passed to daughter cells. The usual choice is a plasmid, a small circular DNA molecule separate from the bacterial chromosome, carrying an origin of replication so the host copies it.

Three things have to be on that plasmid besides the gene. An origin of replication, or nothing is copied. A promoter the host recognises, placed in front of the inserted gene — a human promoter means nothing to E. coli, and a gene with no working promoter is transcribed by nobody. And a marker gene, for reasons that become obvious as soon as you count.

Look at where the insert sits. Placing it in the middle of the marker gene looks like vandalism and is the entire design: a broken marker is the signal that the insertion worked.

Transformation is the uptake of the plasmid by the bacterium. The cells are made permeable — the standard method is suspending them in an ice-cold calcium chloride solution and then giving them a brief heat shock at about 42 °C, and electroporation, a short high-voltage pulse, is the alternative. Either way the efficiency is poor. Typically well under one per cent of the cells take up a plasmid, and among the plasmids themselves many have re-closed on their own sticky ends without accepting an insert at all.

So after transformation the culture contains three populations: cells with no plasmid, cells with a plasmid that closed on itself, and cells with the recombinant plasmid you actually want. They look identical. The marker gene is how you tell them apart, and there are several designs.

Marker	How it is read	Comment
Antibiotic resistance gene	Grow on agar containing the antibiotic; only cells with a plasmid survive	Separates plasmid from no plasmid, but not recombinant from self-closed
A second resistance gene, with the insert placed inside it	Replica plating: colonies that die on the second antibiotic are the recombinant ones	Works, but you have to grow the colonies twice
Fluorescent protein gene, insert placed inside it	Recombinant colonies fail to fluoresce under ultraviolet light	One step, non-destructive; now the common choice
Enzyme marker such as lacZ, insert placed inside it	Recombinant colonies stay white on a substrate that others turn blue	Cheap and visible without special equipment

Antibiotic resistance markers are being designed out of commercial work, because releasing large volumes of bacteria carrying resistance genes is an obvious risk to take when a fluorescent marker does the same job. That argument is a common source of evaluation questions.

Sorting three populations of cells

A plasmid carries a gene for ampicillin resistance and, separately, a gene for a green fluorescent protein. The gene of interest is inserted into the middle of the fluorescent protein gene. After transformation the bacteria are spread on agar containing ampicillin and the surviving colonies are examined under ultraviolet light. Explain what each observation tells you.

Start with what the ampicillin does. Only cells that have taken up a plasmid carry the resistance gene, so only those can make the enzyme that breaks the antibiotic down. Every cell that failed to take up a plasmid dies, and no colony forms. Every colony you can see therefore contains a plasmid of some kind.

Now the ultraviolet light. A plasmid that closed on itself without accepting the insert has an intact fluorescent protein gene, so it makes the protein and the colony glows. A plasmid that accepted the insert has that gene interrupted in the middle, so no functional protein is made and the colony does not glow.

The colonies you want are therefore the ones that grew on ampicillin and did not fluoresce: they survived, so they have a plasmid, and they are dark, so the plasmid carries the insert. A colony that grew and glowed has a self-closed plasmid and is discarded.

Note the logic, because it recurs. The marker never detects the gene of interest. It detects the consequence of the gene of interest having landed in a particular place, which is why the insertion site is chosen deliberately rather than by chance.

The polymerase chain reaction, temperature by temperature

PCR makes many millions of copies of a chosen stretch of DNA in a couple of hours, in a tube, with no cells involved. It is the single most-used technique in molecular biology, and questions about it almost always come down to the three temperatures and why each step needs its own.

The tube contains the DNA sample, a large excess of the four free nucleotides, two primers, and a heat-stable DNA polymerase in a buffer. The primers are short single strands of DNA, usually about twenty bases, complementary to the sequences at each end of the target region. They do two jobs: they give DNA polymerase the free 3′ end it must have to start, and they define precisely which stretch of the sample gets amplified. Change the primers and you amplify something else.

Three temperatures, three jobs, and a count that doubles at the end of every loop. Watch the number at the bottom rather than the drawing: the drawing repeats, and the number is the point.

Denaturation, 95 °C for about thirty seconds. Heat this high breaks the hydrogen bonds between the paired bases and the two strands separate. It does not break the phosphodiester bonds along the backbones, which are covalent, so each strand survives intact and can be used as a template. Nothing cooler will reliably separate a long double helix, particularly one rich in guanine and cytosine, where every pair is held by three hydrogen bonds instead of two.

Annealing, 50 to 65 °C for about thirty seconds. On cooling, the primers bind to their complementary sequences on the separated strands. The exact temperature is set by the primers themselves — their length and their guanine-cytosine content — and 55 to 60 °C is typical. Both errors matter and they are opposite. Too hot and the primers do not stay bound, so nothing is copied. Too cool and they bind to sequences they only partly match, and the machine faithfully amplifies the wrong piece of DNA.

Extension, 72 °C. This is the optimum temperature for Taq polymerase, the enzyme used, and it is why the step has a temperature of its own rather than simply staying at the annealing temperature. Taq extends each primer in the 5′ to 3′ direction at roughly a thousand bases a minute, so the length of this step is set by the length of the target. Run the reaction at 37 °C, where a human polymerase would be happiest, and Taq crawls; the enzyme and the temperature belong together.

The reason the whole cycle is possible is that Taq polymerase comes from Thermus aquaticus, a bacterium of hot springs, and its tertiary structure is not destroyed at 95 °C. Before it was adopted, the polymerase was denatured by every denaturation step and fresh enzyme had to be pipetted into every tube in every cycle by hand. That is the difference between a technique and a laboratory curiosity.

copies after n cycles = 2ⁿ × starting moleculeswhich is exact only in principle: yield plateaus once the reagents start to run out

How many copies, and how quickly

A reaction begins with a single molecule of double-stranded target DNA and runs for 30 cycles. Calculate the number of copies produced, state how long the run takes if each cycle lasts two minutes, and explain why the yield in a real tube is lower than your calculated figure.

Each cycle doubles the number of molecules, because both strands of every molecule act as a template. After 30 cycles there are 2³⁰ molecules, which is 1 073 741 824 — a little over a thousand million from one.

At two minutes a cycle, 30 cycles take 60 minutes. One hour turns a single molecule into more than a billion, which is why a trace of DNA at a crime scene, or from a single cell taken from an embryo, is enough to work with.

The real yield is lower for several reasons that all amount to the same thing: the reaction is not run in an infinite tube. The free nucleotides and the primers are used up, so later cycles have less to work with. Taq loses some activity over repeated trips to 95 °C, however thermostable it is. And as the concentration of product rises, separated strands increasingly find each other and re-anneal before a primer can bind. In practice the amount of product levels off after around 30 to 35 cycles however long you leave it running.

Gel electrophoresis: sorting fragments by length

Having made or cut DNA, you generally need to know what lengths you have. Gel electrophoresis separates fragments by size, and it works because of a fact about DNA you met in the very first lesson on the molecule: every nucleotide carries a phosphate group, and phosphate groups are negatively charged.

The samples are loaded into wells at one end of a slab of agarose gel covered in buffer, and a voltage is applied across it. Because all DNA is negatively charged, every fragment moves towards the positive electrode; charge does not distinguish them, since a longer fragment carries proportionately more charge. What distinguishes them is the gel. Agarose is a mesh, and a short fragment threads through it more easily than a long one, so in a given time the short fragments travel further. Distance moved is roughly proportional to the logarithm of fragment length rather than to length itself, which is why size markers are run alongside.

The first lane is the ladder — fragments of known length, run purely so the others can be read against them. Notice that sample 3's single band sits level with the 3000 rung, which is the whole method in one observation.

DNA is colourless, so the gel has to be stained or the fragments labelled. Traditional practicals use methylene blue, which is safe and insensitive; research laboratories use fluorescent dyes that bind DNA and glow under ultraviolet light, or attach a fluorescent tag to a primer during the PCR that produced the fragments. Where the fragments have to be probed afterwards rather than merely measured, they are transferred from the gel onto a nylon membrane, a technique known as Southern blotting, and it is the membrane that meets the probe.

TRY IT — Explaining an experiment that went wrong

A student runs a gel and finds that every lane shows a single bright smear near the well and nothing further down, while the same samples run correctly by a technician give clear bands. Suggest two possible explanations, and in each case say what the student should change.

Check your answer

The first possibility is that the electrodes were connected the wrong way round, or that the run was stopped almost immediately. DNA is negatively charged and moves towards the positive electrode, so if the wells were nearest the positive electrode the fragments would run off the end of the gel rather than into it, and what remains near the well would be whatever failed to move. The student should check that the wells sit at the negative end and that the voltage was applied for the full time.

The second is that the DNA was not cut, or was degraded. Uncut chromosomal DNA is enormously long and barely enters the gel at all, so it sits as a smear at the well — which is exactly what is described. The student should check that the restriction enzyme was added, that it was incubated at the right temperature for long enough, and that the sample was not left at room temperature where nucleases would break it into a continuous range of lengths.

A useful general habit: a gel with the same fault in every lane points to something done to the whole gel — the voltage, the buffer, the orientation — while a fault in one lane points to that sample. The pattern of the failure tells you where to look.

In the exam

Say which route to the gene you would choose and why. 'Use reverse transcriptase' is half an answer; 'use reverse transcriptase on mRNA from a cell that expresses the gene heavily, because the mRNA has already had its introns removed' is a complete one.
Sticky ends pair by hydrogen bonding; ligase then forms phosphodiester bonds. Two different bonds, two separate marks.
The same restriction enzyme must be used on the gene and on the vector, because that is what makes the overhangs complementary.
Quote the PCR temperatures with the reason attached: 95 °C to break hydrogen bonds, 50 to 65 °C for the primers to anneal, 72 °C because that is the optimum for Taq polymerase.
Taq polymerase is used because it is thermostable and survives the denaturation step.
In electrophoresis, DNA moves towards the positive electrode because the phosphate groups are negatively charged, and short fragments move furthest because they pass through the gel mesh more easily.
When a question asks why marker genes are needed, the answer is about numbers: most cells take up nothing, so you need a way of identifying the few that did.

Check yourself

A company wants E. coli to manufacture a human protein. Describe, in order, how they would obtain the gene, insert it into the bacteria and identify the cells that will produce the protein, giving a reason for each choice you make.

Answer

Obtain the gene from mRNA rather than from a chromosome. Choose a human cell type that produces the protein in quantity, extract its mRNA, and use reverse transcriptase to make a single strand of complementary DNA against it; DNA polymerase then builds the second strand. The reason is introns: a bacterium has no spliceosomes, so a gene taken from a chromosome would be transcribed and translated complete with its intron sequence and would yield no functional protein. The gene machine would serve equally well, and would allow the codons to be chosen to suit E. coli.

Cut the cDNA and a plasmid with the same restriction endonuclease, so that both carry complementary sticky ends. Mix them, and the overhangs pair by hydrogen bonding; add DNA ligase to form the phosphodiester bonds that make the join permanent. The plasmid must already carry an origin of replication so that the bacterium copies it, a promoter the bacterium's own RNA polymerase recognises placed in front of the insertion site, and a marker gene.

Transform the bacteria: suspend them in ice-cold calcium chloride and give a brief heat shock, or use electroporation, to make the membranes permeable so that plasmids can enter.

Identify the successful cells with the marker. If the marker is an antibiotic resistance gene, grow the culture on agar containing that antibiotic and only cells carrying a plasmid survive. If the gene was inserted into the middle of a second marker — a fluorescent protein gene, say — then among the survivors it is the colonies that do not fluoresce that carry the insert, because the insertion has interrupted that gene. Those colonies are cultured on a large scale in a fermenter and the protein extracted.

A good answer notices what could still go wrong. A bacterium cannot carry out post-translational modification, so a protein needing particular sugars added in a Golgi apparatus has to be made in yeast or in mammalian cell culture instead, however well the transformation worked.

Questions

Written to the command words the boards use. Try them on paper before opening a scheme: the marks go to points made, not to length.

Question 14 marks

Explain why the same restriction endonuclease must be used to cut both the gene and the plasmid, and explain what DNA ligase then contributes.

Mark scheme

B1 a restriction endonuclease cuts only at its own recognition sequence, and an offset cut through the two strands leaves each fragment with a short single-stranded overhang, a sticky end
B1 using one enzyme on both means every fragment carries the same overhang, so the exposed bases of the gene are complementary to those of the cut plasmid
B1 the overhangs pair by hydrogen bonding between complementary bases, which holds the pieces together but leaves the sugar-phosphate backbones unjoined
B1 DNA ligase catalyses the formation of phosphodiester bonds along both strands, which makes the join permanent

Question 24 marks

A polymerase chain reaction begins with 5 molecules of double-stranded target DNA and runs for 20 cycles. Calculate the number of molecules present at the end, and calculate how long the run takes if each cycle lasts 90 seconds.

Mark scheme

M1 the number of molecules doubles each cycle, so the number at the end is the starting number multiplied by 2 raised to the number of cycles
M1 5 × 2²⁰ = 5 × 1 048 576
A1 5 242 880 molecules
A1 20 × 90 = 1800 seconds, which is 30 minutes

Question 34 marks

Well under one bacterium in a hundred takes up a plasmid, and many plasmids close on their own sticky ends without accepting an insert. Suggest why a marker gene is needed at all, and suggest how placing the gene of interest inside a gene for a fluorescent protein allows the recombinant cells to be picked out.

Mark scheme

B1 after transformation the culture holds three populations — cells with no plasmid, cells with a self-closed plasmid and cells with the recombinant plasmid — and they look identical
B1 a marker gene gives an observable difference, so the very small proportion of cells that took up the plasmid wanted can be found without testing every colony
B1 a plasmid that closed on itself has an intact fluorescent protein gene, so it makes the protein and its colony glows under ultraviolet light
B1 a plasmid carrying the insert has that gene interrupted in the middle, so no functional protein is made and the colonies wanted are the ones that do not glow

Question 43 marks

State the three temperatures used in a cycle of the polymerase chain reaction, and state what each of them is for.

Mark scheme

B1 95 °C, at which the hydrogen bonds between paired bases break and the two strands separate
B1 50 to 65 °C, at which the primers bind to their complementary sequences at each end of the target region
B1 72 °C, which is the optimum for Taq polymerase as it extends each primer to build the new strand

Question 53 marks

Explain why DNA fragments move towards the positive electrode during gel electrophoresis, and explain why the shortest fragments travel furthest.

Mark scheme

B1 every nucleotide carries a phosphate group and phosphate groups are negatively charged, so every fragment is attracted towards the positive electrode
B1 charge does not separate the fragments, because a longer fragment carries proportionately more of it
B1 the agarose gel is a mesh, and a short fragment threads through it more easily than a long one, so in a given time the short fragments travel further

Worth remembering

Three routes to a gene: reverse transcriptase from mRNA, restriction enzymes from the chromosome, or synthesis by the gene machine.
Restriction enzymes cut at palindromic recognition sequences; an offset cut leaves sticky ends.
Sticky ends pair by hydrogen bonds; ligase makes the phosphodiester bonds.
A vector needs an origin of replication, a promoter the host recognises, and a marker gene.
PCR: 95 °C to separate the strands, 50 to 65 °C for the primers, 72 °C for Taq polymerase, and the number of copies doubles each cycle.
In a gel, DNA runs towards the positive electrode and the shortest fragments travel furthest.

← Mutation and cancer: what goes wrong, and where it goes wrong · Genomes, screening and gene therapy: what a sequence is good for →