Glucose/Tris/EDTA: Glucose functions to maintain osmotic pressure, while the Tris buffers the cells at pH 8. O. EDTA binds divalent cations in the lipid bilayer, thus weakening the cell envelope. Following cell lysis, EDTA limits DNA degradation by binding Mg ions that are necessary cofactor for bacterial nucleases.

SDS/sodium hydroxide: This alkaline mixture lyses the bacterial cells. The detergent SDS dissolves the lipid components of the cell envelope and the cellular proteins. Sodium hydroxide denatures the chromosomal and plasmid DNA into single strands; the intact circles of plasmid DNA remain intertwined.

Potassium acetate/acetic acid: The acetic acid brings the pH to neutral, allowing the DNA strands to renature. The large, disrupted chromosomal strands cannot rehybridize perfectly but instead collapse into a partially hybridized tangle. At the same time, the potassium acetate precipitates the SDS from the cell suspension, along with the associated proteins and lipids. The renaturing chromosomal DNA is trapped in the SDS/lipid/protein precipitate. Only smaller plasmid DNA, fragments of chromosomal DNA, and RNA molecules escape the precipitate and remain in solution.

Isopropanol: The alcohol rapidly precipitates nucleic acids but precipitates proteins slowly. Thus, a quick precipitation preferentially brings down nucleic acids.

Ethanol: A wash with ethanol removes some remaining salts and SDS from the preparation.

Tris-EDTA: Tris buffers the DNA solution. EDTA protects the DNA from degradation by DNases by binding the divalent cations (especially Mg) that are necessary cofactors for DNase activity.

Interpretation of Gels

Interpreting gels containing plasmid DNA is not a straightforward process and is further complicated by the impurities in miniprep DNA.

A heavy background smear, along with DNA of high molecular weight at the top of the undigested lane, indicates that the miniprep is contaminated with large amounts of chromosomal DNA.

Undissolved material and DNA of high molecular weight are frequently trapped at the front edge of the sample well.

A "cloud" of RNA of low molecular weight is often seen in both the cut and uncut miniprep lanes, at a position corresponding to 100-200 bp.

Plasmid DNA can exist in any one of three major conformations:

Form 1, supercoiled: Although a plasmid is usually pictured as an open circle, within the E. coli cell (in vivo) the DNA strand is wound around histone-like proteins to fon-n a compact structure. Adding these coils to the coiled DNA helix produces a supercoiled molecule. The extraction procedure strips proteins from plasmid, causing the molecule to coil about itself A compact molecular shape allows supercooled plasmid to be the fastest-moving form under most gel conditions. Therefore the fastest-moving band of uncut plasmid is assumed to be supercoiled.

Form 11, relaxed or nicked circle: Physical shearing and enzymatic cleavage during plasmid isolation introduce nicks in the supercoiled plasmid to produce the familiar, open circular structure. Thus, the percentage of supercoiled plasmid is an indicator of the care with which the DNA is extracted from the E. coli cell. The relaxed circle is the slowest-migrating form of plasmid; its "floppy" molecular shape ftnpedes movement through the agarose matrix.

Form 111, linear: Linear DNA is produced when a restriction enzyme cuts the plasmid at a single recognition site or when damage results in strand nicks directly opposite each other on the DNA helix. Linear plasmid DNA migrates at a rate intermediate between supercoiled and relaxed circle DNA. The presence of linear DNA in a plasmid preparation is a sign of either nuclease contamination or sloppy lab procedure (for example, overinixing or mismeasuring SDSNAOH and KOAC of the miniprep procedure).

Extraction and Purification of Plasmid DNA


Many methods have been developed to purify plasmid DNA from bacteria. These methods invariably involve three steps:

* Growth of the bacterial culture

* Harvesting and lysis of the bacteria

* Purification of plasmid DNA

Growth of the Bacterial Culture

Plasmids are almost always purified from cultures (grown in liquid medium containing the appropriate antibiotic) that have been inoculated with a single bacterial colony picked from an agar plate. Many of the currently used plasmid vectors (e.g., the pUC series) replicate to such a high copy number that they can be purified in large yield from cultures that have simply been grown to late log phase in standard LB medium. In these cases, it is not necessary to amplify the plasmid DNA selectively. However, vectors of an earlier generation (e.g., pBR322), which do not replicate so freely, need to be selectively amplified by incubating the partially grown bacterial culture in chloramphenicol for several hours. Chloramphenicol inhibits host protein synthesis and, as a result, prevents replication of the bacterial chromosome. However, replication of relaxed plasmids continues, and their copy number increases progressively for several hours. The yield of plasmids such as pBR322 is thus considerably higher from chloramphenicol-treated cultures than from untreated cultures.

For many years, it has been standard practice to add chloramphenicol in concentrations sufficient to achieve complete inhibition of protein synthesis (170 ug/ml). The yield of most plasmids isolated from cultures treated in this fashion is more than sufficient for almost any conceivable task in molecular cloning. However, in those cases where plasmids (or more commonly, cosmids) grow poorly because of their large size or because of the foreign sequences they carry, it may be worthwhile to explore alternative ways to grow and treat the bacterial culture. For example, it has been reported (Tartof and Hobbs 1987) that the use of a rich medium results in a four- to sixfold increase in the yield of "difficult" plasmids. In addition, the concentration of chloramphenicol can affect the degree of amplification of plasmid DNAS. Improved yields of pBR322 and pBR327 have been obtained from bacterial cultures treated with concentrations of chloramphenicol (10-20 tkg/ml) that do not completely suppress host protein synthesis (Frenkel and Bremer 1986). The reason for this result is not understood, but it could be explained if the replication of plasmids carrying the ColEl origin required an unstable host factor that continues to be synthesized during partial inhibition of protein synthesis.




Harvesting and Lysis of the Bacte@

Bacteria are recovered by centrifugation and lysed by any one of a large number of methods, including treatment with nonionic or ionic detergents, organic solvents, alkali, or heat. The choice among these methods is dictated by three factors: the size of the plasmid, the strain of E. coli, and the technique that is to be used subsequently to purify the plasmid DNA. Although it is impractical to give precise conditions for each combination of plasmid and host, the following general guidelines can be used to select a method that will give satisfactory results:

1. Large plasmids ( > 15 kb in size), which are susceptible to damage, should be released from cells by gentle lysis (see page 1.36). Bacteria are suspended in an isosmotic solution of sucrose and then treated with lysozyme and EDTA to break down the cell wall and outer membrane. The resulting spheroplasts are lysed by adding a detergent such as SDS. This method minimizes the physical forces that are required to liberate the plasmid from the pressurized interior of the bacterium.

2. More severe methods can be used to isolate smaller plasmids. After addition of EDTA and, in some cases, lysozyme, the cells are exposed to detergent and lysed by boiling or treatment with alkali. These treatments, which disrupt base pairing, cause the linear chromosomal DNA of the host to denature. However, the strands of closed circular plasmid DNA are unable to separate from one another because they are topologically intertwined. When conditions are returned to normal, the strands of the plasmid DNA rapidly fall into perfect register, and completely native superhelical molecules are re-formed.

Note: Prolonged exposure of superhelical DNA to heat or alkali results in irreversible denaturation (Vinograd and Lebowitz 1966). The resulting cyclic coiled DNA cannot be cleaved with restriction enzymes, migrates through agarose gels at about twice the rate of superhelical DNA, and stains poorly with ethidium bromide. Traces of this form of DNA can often be seen in plasmids prepared by alkaline or thermal lysis of bacteria.

3. Some strains of E. coli (e.g., some variants and derivatives of HB101) release relatively large amounts of carbohydrate when they are lysed by detergent and heat. This can be a nuisance when the plasmid DNA is subsequently purified by equilibrium centrifugation in CsCI-ethidium bromide gradients. The carbohydrate forms a dense, fuzzy band close to the place in the gradient occupied by the supercoiled plasmid DNA. It is therefore difficult to avoid contaminating the plasmid DNA with carbohydrate, which inhibits the activity of many restriction enzymes. Boiling should therefore not be used when making large-scale preparations of plasmids from strains of E. coli such as HB101 and TG1.

4. The boiling method is not recommended when making small-scale preparations of plasmid DNA from strains of E. coli that express endonuclease A (enda' strains; e.g., HB101). Because endonuclease A is not completely inactivated by the boiling procedure, the plasmid DNA is degraded during subsequent incubation in the presence of Mg++ (e.g., during digestion with restriction enzymes). This problem can be avoided by including an extra step-extraction with phenol:chloroform.

5. Virtually all of the procedures given in this section can be easily adapted to accommodate bacterial cultures that range in size from 1 ml to 1 liter. All of them have been used successfully to isolate plasmid DNA simultaneously from many small cultures ("minipreps"), as well as from largescale cultures ("maxipreps") and cultures of intermediate size ("midipreps"). With the one exception referred to in number 3 above, any method can therefore be used with any strain of E. coli grown in cultures of any size. It is curious, however, that many laboratories use one method to process minipreps and another method to isolate plasmid DNA from large-scale cultures. In some laboratories, the alkali method is used for minipreps and boiling is used for maxipreps, whereas in others, the situation is reversed.

6. The copy numbers of the current generation of plasmids are now so high that selective amplification in the presence of chloramphenicol is unnecessary to achieve high yields. However, some workers continue to use chloramphenicol, not to increase the yield of plasmid DNA but to reduce the bulk of bacterial cells in maxipreps. Handling large quantities of highly viscous lysates of concentrated suspensions of bacteria is a frustrating and messy business that can be avoided if chloramphenicol is added to the culture at mid-log phase. Approximately equivalent yields of plasmid DNA are obtained from smaller numbers of cells that have been exposed to chloramphenicol and from larger numbers of cells that have not.


Purification of Plasmid DNA

All of the methods in common use exploit the relatively small size and covalently closed circular nature of plasmid DNAS. For example, separation of plasmid and chromosomal DNAs by equilibrium centrifugation in CsCl-ethidium bromide gradients depends on differences between the amounts of ethidium bromide that can be bound to linear and closed circular DNA molecules. Ethidium bromide binds to DNA by intercalating between the bases, causing the double helix to unwind. This leads to an increase in the length of linear DNA molecules and to the introduction of compensatory superhelical turns in closed circular plasmid DNAS. Eventually, the density of these superhelical turns becomes so great that the intercalation of additional molecules of ethidium bromide is prevented. Linear molecules, which are not constrained in this way, continue to bind more dye until saturation is reached (- 1 ethidium bromide molecule for every 2 base pairs) (Cantor and Schimmel 1980). Because of this differential binding of dye, the buoyant densities of the linear and closed circular DNA molecules are different in CsCl gradients containing saturating amounts of ethidium bromide.

For many years, equilibrium centrifugation in CsCl-ethidium bromide gradients has been the method of choice to prepare large amounts of plasmid DNA. However, this process is expensive and time-consuming, and many alternative methods have been developed, the majority of which involve the use of ion-exchange or gel-filtration chromatography or differential precipitation to separate plasmid and host DNAS. Although most of these methods have fallen by the wayside, the best of them-differential precipitation with polyethylene glycol-has recently been improved to the point where it yields plasmid DNA of extremely high purity. Differential precipitation with polyethylene glycol differs from equilibrium centrifugation in CsCl-ethidium bromide gradients in one respect: It does not efficiently separate nicked circular molecules from the closed circular form of plasmid DNA. Equilibrium centrifugation is therefore the method of choice for the purification of very large plasmids ( > 15 kb), which are vulnerable to nicking, and closed circular plasmids that are to be used for biophysical measurements. However, both methods of purification yield plasmid DNAs that can be used for a variety of sophisticated tasks in molecular cloning, including transfection of mammalian cells and the generation of sets of deletion mutants with exonucleases. For more routine procedures, the problem of extensive purification can fortunately be avoided altogether. Most of the plasmids in common use replicate so profusely that sufficient DNA can be obtained from minipreps to accomplish such tasks as the construction of restriction maps, transformation of bacteria, isolation of specific DNA fragments, routine subcloning, and generation of radiolabeled probes.