Dale_Molecular Genetics of Bacteria 4th ed

Plasmid-free cell dies

Figure 5.17 Role of ccdAB genes in maintenance of the F plasmid. The F plasmidencoded CcdB protein is toxic, but is neutralized by the presence of CcdA. If the plasmid is lost, CcdA is rapidly destroyed exposing the cell to the toxic effects of the more stable CcdB protein

mediated supercoiling), while the (less stable) CcdA protein antagonizes the effect of CcdB. In a plasmid-free segregant, the CcdA product is destroyed by proteolytic enzymes, while the stable CcdB protein persists and kills the cell. Similar systems are found on a wide variety of unrelated plasmids.

Some plasmids possess an alternative mechanism for post-segregational killing, which works by regulating the expression of a toxin. For example, the plasmid R1 carries a gene, hok (host killing), that codes for a small polypeptide that is toxic because of its effects on the cell membrane. In plasmid-containing cells, translation of hok mRNA is prevented by an antisense RNA molecule (sok, suppression of killing) that is complementary to the leader region of the hok mRNA – it is transcribed from the same region of DNA but in the opposite direction. The sok RNA molecule decays rapidly, but the hok mRNA is very stable. So in the plasmid-free segregant, the hok mRNA persists and, in the absence of sok RNA, is translated to produce the toxin. A simplified model of this system is shown in Figure 5.18.

Differences in growth rate are expected to arise because of the metabolic load arising from replication of the plasmid and expression of its genes. With most wild-type plasmids, these demands are small, and hence the effect on growth rate is also small. On the other hand, genetically engineered plasmids are often maintained at very high copy numbers and express very high levels of a specific product. The burden on the cell is therefore very much greater and the problem of instability can become acute. Some ways of overcoming this problem are considered in Chapter 8.

5.4 Methods for studying plasmids

5.4.1 Associating a plasmid with a phenotype

The variability of a phenotype is often the first indication that a plasmid is involved. That is, some strains possess an unusual characteristic, which tends to be lost at a frequency higher than expected from mutation, and that loss is irreversible. With some plasmids it is possible to increase the rate at which the plasmid is lost by various treatments, such as acridine orange or growth at a higher temperature – a procedure known as ‘curing’ or plasmid elimination. It must be stressed that the terms ‘elimination’ and ‘plasmid loss’ do not mean the physical removal of the plasmid from a particular cell; the process works by interfering with the replication and/or partitioning of the plasmid so as to increase the rate at which plasmid-free segregants occur.

This evidence should be combined with detection of plasmid DNA by agarose gel electrophoresis (Figure 5.19; see also Box 2.2). The chromosomal DNA will show up as a rather diffuse band, since it is fragmented randomly by the extraction procedure, while any plasmid DNA will form separate, sharper bands at a position determined primarily by their size. The smaller the molecule, the faster it will run. Small plasmids will be found well ahead of the chromosomal DNA, while larger plasmids actually run slower than the chromosomal DNA. This may seem rather surprising (given that the chromosome is very much bigger than even the largest plasmid), but it should be remembered that the chromosome is broken into linear fragments while the plasmid will be present as intact circular molecules. Therefore they are not directly comparable.

The conformation of the plasmid will also affect its mobility. Usually the intact plasmid will be a covalently closed, supercoiled circle, which will migrate differently from a linear molecule of the same molecular weight (this must be taken into account when attempting to determine the size of a plasmid) and considerably faster than a nicked open circular form. In addition to these three major forms of the plasmid (supercoiled, nicked circles and linear), there may be dimers and higher multimers which will move more slowly. Since a single plasmid can give

−

Position of wells

Large plasmids

Chromosomal DNA fragments

Small plasmids

+

Figure 5.19 Demonstration of plasmids in cell extracts by agarose gel electrophoresis

rise to several bands, it is often difficult to determine with confidence how many plasmids are present in a given strain.

In order to study a plasmid in more depth it is necessary to purify it. Separating DNA from other cell components is relatively easy to accomplish. However, separating a plasmid from chromosomal DNA is less straightforward. Most commonly used methods rely on the different properties of the supercoiled circular plasmid and the linear fragments of chromosomal DNA. However, these procedures are not necessarily appropriate for larger plasmids, or for other forms of plasmid DNA such as the single-stranded and linear forms referred to earlier in this chapter.

The combination of the instability of the phenotype and the physical demonstration of plasmid DNA provides evidence that a plasmid is involved. The hypothesis can be strengthened by showing that colonies which have lost the characteristic in question have also lost the plasmid. But the evidence is still circumstantial. To obtain more conclusive evidence, the best procedure is to introduce the plasmid into a cell that does not have it (by conjugation or transformation; see Chapter 6), and to determine whether the acquisition of the plasmid leads to a corresponding change in the phenotype. Such experiments are easily carried out with selectable markers like antibiotic resistance. If the plasmid carries a resistance gene, the cells that have acquired the plasmid can be detected by simply plating them on a medium containing the antibiotic. In this way extremely rare events can be detected, e.g. one resistant transconjugant (or transformant) in 108 cells. Other characteristics are not so easy. To determine whether a plasmid carries genes that are necessary for virulence for example, it may be quite difficult to test it in such a simple manner.

5.4.2 Classification of plasmids

It is often useful to be able to compare plasmids identified in different bacterial isolates. This may for example provide information about the occurrence of a particular plasmid in strains causing an outbreak of infection in a hospital where antibiotic resistance is often a serious problem. The division into ‘small’ and ‘large’ plasmids is clearly much too crude to cope with the vast number of different plasmids that can be present in such a situation.

Analysis of the profile of antibiotic resistance genes carried by the plasmid has the virtue of being quick and easy to carry out, but it must be realized that plasmids are very fluid in their make-up; the presence of transposons (see Chapter 7) commonly leads to new resistance genes being acquired by a plasmid, even during the course of a specific outbreak. Conversely, plasmids can lose genes by deletions (as discussed earlier). Excessive reliance on the resistance pattern can therefore be misleading.

Incompatibility groups

Earlier in this chapter, it was shown that interactions between the replication control mechanisms of different plasmids may make them incompatible. This forms the basis of a widely used method of classifying plasmids. The test is carried out by transferring the unknown plasmid into each of a number of standard plasmid-carrying strains and testing for their incompatibility. A large number of incompatibility groups have been defined in this way; F for example belongs to the group IncFI, and is incompatible with plasmids such as ColV-K94. The R100 plasmid has a similar conjugation and transfer system but is compatible with F; it belongs to the group IncFII. There are a number of problems with this system, not the least of which is that it is extremely cumbersome. Also it is not possible to compare plasmids from different species, unless they can both be transferred to a common host. So there is one set of groups for E. coli plasmids, and another for plasmids from Pseudomonas species. However, the strength of the method is that it does classify plasmids on the basis of a fundamental characteristic – the nature of their replication control.

Host range

Many plasmids, including ColE1 and cloning vectors related to it, are only able to replicate in a limited range of bacterial hosts. However, this is not universally true and some plasmids have a remarkably broad host range. Notable amongst these are the P group of plasmids, such as RP4, which are able to replicate in some Gram-positive bacteria, as well as in most Gram-negatives. Plasmids such as RP4

are also promiscuous in the sense that not only can they replicate in a broad range of host bacteria, but they are also capable of promoting their own transmission by conjugation (see Chapter 6) between widely diverse bacterial species. Broad hostrange plasmids can be useful as genetic tools (see Chapter 8), as well as their natural role in promoting gene flow between widely diverse bacterial species.

Molecular characterization

Plasmids can be compared more readily by the use of restriction endonucleases. Since these enzymes cut DNA at specific points, digestion of a plasmid will give rise to a characteristic set of fragments that can be separated on an agarose gel (Chapter 2). If two plasmids are the same, the restriction pattern will be identical. This procedure is however still subject to the same problem referred to above: plasmids can acquire additional genes by transposition or lose them by deletion. A different pattern of restriction fragments may therefore indicate the occurrence of events of this kind rather than the presence of unrelated plasmids.

Such an analysis can be extended by testing the ability of DNA fragments from one plasmid to hybridize to a second plasmid, which indicates the presence of related sequences. However, since two otherwise dissimilar plasmids may have acquired related antibiotic resistance genes (which will therefore cross-hybridize), these results also need to be interpreted with care.

Ultimately, of course, the best comparison of two plasmids at the molecular level is to determine the complete sequence of both plasmids.

unequal. The parents are thus referred to as donor and recipient cells; the recombinant progeny resemble the original recipient strain in most characteristics.

In this chapter the focus will be on the naturally-occurring transfer of genes between bacteria (horizontal gene transfer) and on the significance that this has for the evolution of bacteria. Mechanisms that are used in the laboratory, especially for genetic modification, and their application for both genetic manipulation and for genetic analysis, will be covered in Chapters 8 to 10.

6.1 Transformation

In a sense, it was the discovery of transformation, the uptake of DNA by a bacterial cell, that initiated the study of bacterial genetics and molecular biology as we know it today. It was in 1928 that Fred Griffith, working with pneumococcus (Streptococcus pneumoniae) discovered that avirulent strains could be restored to virulence by incubation with an extract from killed virulent cells. Sixteen years later, Avery, MacLeod and McCarty demonstrated that the ‘transforming principle’ was DNA, which established the role of DNA as the hereditary material of the bacterial cell. Transformation has been important in genetic analysis of some species and more recently (and to a much greater extent) because of its key role in gene cloning.

With the pneumococcus, cells spontaneously become competent to take up DNA. Such naturally-occurring transformation has been most studied in Bacillus subtilis and Haemophilus influenzae (as well as Str. pneumoniae) and was for some time thought to be limited to these and related species. It is now known to be much more widespread. In particular, transformation contributes extensively to the antigenic variation observed in the gonococcus (Neisseria gonorrhoeae) through the transfer of pil genes coding for the major protein subunit of the surface appendages (pili) by which the bacteria attach to epithelial cells. Although the number of species in which natural transformation has been demonstrated is still quite limited, it is likely that it occurs, albeit at a low level, in many other bacteria.

The details of the process vary between species, but some generalizations are possible. Competence generally occurs at a specific stage of growth, most commonly in late log phase, just as the cells are entering stationary phase. This may be a response to cell density rather than (or as well as) growth phase. For example, in Bacillus subtilis, some of the genes involved in the development of competence are also involved in the early stages of sporulation. The development of competence at this stage is associated not only with nutrient depletion but also with the accumulation of specific secreted products (competence factors) which act via a twocomponent regulatory system to stimulate the expression of other genes required for competence. Since the level of these competence factors is dependent on cell concentration, competence will only develop at high cell density. This is a form of quorum sensing, as described in Chapter 3, in which the response of an individual cell is governed by the concentration of bacteria in the surrounding medium.