The earliest studies in bacterial genetics revealed the unexpected process of cell conjugation.

Do bacteria possess any processes similar to sexual reproduction and recombination? The question was answered by the elegantly simple experimental work of Joshua Lederberg and Edward Tatum, who in 1946 discovered a sex-like process in what became the main model for bacterial genetics, Escherichia coli (see the Model Organism box). They were studying two strains of E. coli with different sets of auxotrophic mutations. Strain A⁻ would grow only if the medium were supplemented with methionine and biotin; strain B⁻ would grow only if it were supplemented with threonine, leucine, and thiamine. Thus, we can designate the strains as

Figure 5-5a displays in simplified form the design of their experiment. Strains A⁻ and B⁻ were mixed together, incubated for a while, and then plated on minimal medium, on which neither auxotroph could grow. A small minority of the cells (1 in 10⁷) was found to grow as prototrophs and, hence, must have been wild type, having regained the ability to grow without added nutrients. Some of the dishes were plated only with strain A⁻ bacteria and some only with strain B⁻ bacteria to act as controls, but no prototrophs arose from these platings. Figure 5-5b illustrates the experiment in more detail. These results suggested that some form of recombination of genes had taken place between the genomes of the two strains to produce the prototrophs.

It could be argued that the cells of the two strains do not really exchange genes but instead leak substances that the other cells can absorb and use for growing. This possibility of “cross-feeding” was ruled out by Bernard Davis in the following way. He constructed a U-shaped tube in which the two arms were separated by a fine filter. The pores of the filter were too small to allow bacteria to pass through but large enough to allow easy passage of any dissolved substances (Figure 5-6). Strain A⁻ was put in one arm, strain B⁻ in the other. After the strains had been incubated for a while, Davis tested the contents of each arm to see if there were any prototrophic cells, but none were found. In other words, physical contact between the two strains was needed for wild-type cells to form. It looked as though some kind of genome union had taken place, and genuine recombinants had been produced. The physical union of bacterial cells can be confirmed under an electron microscope and is now called conjugation (Figure 5-7).

178

In 1953, William Hayes discovered that, in the types of “crosses” just described here, the conjugating parents acted unequally (later, we will see ways to demonstrate this unequal participation). One parent (and only that parent) seemed to transfer some or all of its genome into another cell. Hence, one cell acts as a donor, and the other cell acts as a recipient. This “cross” is quite different from eukaryotic crosses in which parents contribute nuclear genomes equally to a progeny individual.

179

By accident, Hayes discovered a variant of his original donor strain that would not produce recombinants on crossing with the recipient strain. Apparently, the donor-type strain had lost the ability to transfer genetic material and had changed into a recipient-type strain. In working with this “sterile” donor variant, Hayes found that it could regain the ability to act as a donor by association with other donor strains. Indeed, the donor ability was transmitted rapidly and effectively between strains during conjugation. A kind of “infectious transfer” of some factor seemed to be taking place. He suggested that donor ability is itself a hereditary state, imposed by a fertility factor (F). Strains that carry F can donate and are designated F⁺. Strains that lack F cannot donate and are recipients, designated F^-.

We now know much more about F. It is an example of a small, nonessential circular DNA molecule called a plasmid that can replicate in the cytoplasm independent of the host chromosome. Figure 5-8 shows how bacteria can transfer plasmids such as F. The F plasmid directs the synthesis of pili (sing., pilus), projections that initiate contact with a recipient (see Figures 5-7 and 5-8) and draw it closer. The F DNA in the donor cell makes a single-stranded version of itself in a peculiar mechanism called rolling circle replication. The circular plasmid “rolls,” and as it turns, it reels out a singlestrand “fishing line.” This single strand passes through a pore into the recipient cell, where the other strand is synthesized, forming a double helix. Hence, a copy of F remains in the donor and another appears in the recipient, as shown in Figure 5-8. Note that the E. coli genome is depicted as a single circular chromosome in Figure 5-8. (We will examine the evidence for it later.) Most bacterial genomes are circular, a feature quite different from eukaryotic nuclear chromosomes. We will see that this feature leads to the many idiosyncrasies of bacterial genetics.

An important breakthrough came when Luca Cavalli-Sforza discovered a derivative of an F⁺ strain with two unusual properties:

It became apparent that an Hfr strain results from the integration of the F factor into the chromosome, as pictured in Figure 5-9. We can now explain the first unusual property of Hfr strains. During conjugation, the F factor inserted in the chromosome efficiently drives part or all of that chromosome into the F⁻ cell. The chromosomal fragment can then engage in recombination with the recipient chromosome. The rare recombinants observed by Lederberg and Tatum in F⁺ × F⁻ crosses were due to the spontaneous, but rare, formation of Hfr cells in the F⁺ culture. Cavalli-Sforza isolated examples of these rare cells from F⁺ cultures and found that, indeed, they now acted as true Hfr’s.

180

Does an Hfr cell die after donating its chromosomal material to an F⁻ cell? The answer is no. Just like the F plasmid, the Hfr chromosome replicates and transfers a single strand to the F⁻ cell during conjugation. That the transferred DNA is a single strand can be demonstrated visually with the use of special strains and antibodies, as shown in Figure 5-10. The replication of the chromosome ensures a complete chromosome for the donor cell after mating. The transferred strand is converted into a double helix in the recipient cell, and donor genes may become incorporated in the recipient’s chromosome through crossovers, creating a recombinant cell (Figure 5-11). If there is no recombination, the transferred fragments of DNA are simply lost in the course of cell division.

181

Linear transmission of the Hfr genes from a fixed point A clearer view of the behavior of Hfr strains was obtained in 1957, when Elie Wollman and François Jacob investigated the pattern of transmission of Hfr genes to F⁻ cells during a cross. They crossed

(Superscripts “r” and “s” stand for resistant and sensitive, respectively.) At specific times after mixing, they removed samples, which were each put in a kitchen blender for a few seconds to separate the mating cell pairs. This procedure is called interrupted mating. The sample was then plated onto a medium containing streptomycin to kill the Hfr donor cells, which bore the sensitivity allele str^s. The surviving str^r cells then were tested for the presence of alleles from the donor Hfr genome. Any str^r cell bearing a donor allele must have taken part in conjugation; such cells are called exconjugants. The results are plotted in Figure 5-12a, showing a time course of entry of each donor allele azi^r, ton^r, lac⁺, and gal⁺. Figure 5-12b portrays the transfer of Hfr alleles.

The key elements in these results are

182

Putting all these observations together, Wollman and Jacob deduced that, in the conjugating Hfr, single-stranded DNA transfer begins from a fixed point on the donor chromosome, termed the origin (O), and continues in a linear fashion. The point O is now known to be the site at which the F plasmid is inserted. The farther a gene is from O, the later it is transferred to the F⁻. The transfer process will generally stop before the farthermost genes are transferred, and, as a result, these genes are included in fewer exconjugants. Note that a type of chromosome map can be produced in units of minutes, based on time of entry of marked genes. In the example in Figure 5-12, the map would be:

How can we explain the second unusual property of Hfr crosses, that F⁻ exconjugants are rarely converted into Hfr or F⁺? When Wollman and Jacob allowed Hfr × F⁻ crosses to continue for as long as 2 hours before disruption, they found that in fact a few of the exconjugants were converted into Hfr. In other words, the part of F that confers donor ability was eventually transmitted but at a very low frequency. The rareness of Hfr exconjugants suggested that the inserted F was transmitted as the last element of the linear chromosome. We can summarize the order of transmission with the following general type of map, in which the arrow indicates the direction of transfer, beginning with O:

Thus, almost none of the F⁻ recipients are converted, because the fertility factor is the last element transmitted and usually the transmission process will have stopped before getting that far.

183

Inferring integration sites of F and chromosome circularity Wollman and Jacob went on to shed more light on how and where the F plasmid integrates to form an Hfr cell and, in doing so, deduced that the chromosome is circular. They performed interrupted-mating experiments with different, separately derived Hfr strains. Significantly, the order of transmission of the alleles differed from strain to strain, as in the following examples:

Each line can be considered a map showing the order of alleles on the chromosome. At first glance, there seems to be a random shuffling of genes. However, when some of the Hfr maps are inverted, the relation of the sequences becomes clear.

The relation of the sequences to one another is explained if each map is the segment of a circle. It was the first indication that bacterial chromosomes are circular. Furthermore, Allan Campbell proposed a startling hypothesis that accounted for the different Hfr maps. He proposed that, if F is a ring, then insertion might be by a simple crossover between F and the bacterial chromosome (Figure 5-13). That being the case, any of the linear Hfr chromosomes could be generated simply by the insertion of F into the ring in the appropriate place and orientation (Figure 5-14).

Several hypotheses—later supported—followed from Campbell’s proposal.

184

How is it possible for F to integrate at different sites? If F DNA had a region homologous to any of several regions on the bacterial chromosome, any one of them could act as a pairing region at which pairing could be followed by a crossover. These regions of homology are now known to be mainly segments of trans-posable elements called insertion sequences. For a full explanation of insertion sequences, see Chapter 15.

The fertility factor thus exists in two states:

The E. coli conjugation cycle is summarized in Figure 5-15.

Broad-scale chromosome mapping by using time of entry Wollman and Jacob realized that the construction of linkage maps from the interrupted-mating results would be easy by using as a measure of “distance” the times at which the donor alleles first appear after mating. The units of map distance in this case are minutes. Thus, if b⁺ begins to enter the F⁻ cell 10 minutes after a⁺ begins to enter, then a⁺ and b⁺ are 10 units apart. Like eukaryotic maps based on crossovers, these linkage maps were originally purely genetic constructions. At the time they were originally devised, there was no way of testing their physical basis.

Fine-scale chromosome mapping by using recombinant frequency For an exconjugant to acquire donor genes as a permanent feature of its genome, the donor fragment must recombine with the recipient chromosome. However, note that time-of-entry mapping is not based on recombinant frequency. Indeed, the units are minutes, not RF. Nevertheless, recombinant frequency can be used for a more fine-scale type of mapping in bacteria, a method to which we now turn.

185

186

First, we need to understand some special features of the recombination event in bacteria. Recall that recombination does not take place between two whole genomes, as it does in eukaryotes. In contrast, it takes place between one complete genome, from the F⁻ recipient cell, called the endogenote, and an incomplete one, derived from the Hfr donor cell and called the exogenote. The cell at this stage has two copies of one segment of DNA: one copy is part of the endogenote and the other copy is part of the exogenote. Thus, at this stage, the cell is a partial diploid, called a merozygote. Bacterial genetics is merozygote genetics. A single crossover in a merozygote would break the ring and thus not produce viable recombinants, as shown in Figure 5-16. To keep the circle intact, there must be an even number of crossovers. An even number of crossovers produces a circular, intact chromosome and a fragment. Although such recombination events are represented in a shorthand way as double crossovers, the actual molecular mechanism is somewhat different, more like an invasion of the endogenote by an internal section of the exogenote. The other product of the “double crossover,” the fragment, is generally lost in subsequent cell growth. Hence, only one of the reciprocal products of recombination survives. Therefore, another unique feature of bacterial recombination is that we must forget about reciprocal exchange products in most cases.

With this understanding, we can examine recombination mapping. Suppose that we want to calculate map distances separating three close loci: met, arg, and leu. To examine the recombination of these genes, we need “trihybrids,” exconjugants that have received all three donor markers. Assume that an interrupted-mating experiment has shown that the order is met, arg, leu, with met transferred first and leu last. To obtain a trihybrid, we need the merozygote diagrammed here:

To obtain this merozygote, we must first select stable exconjugants bearing the last donor allele, which, in this case, is leu⁺. Why? In leu⁺ exconjugants, we know all three markers were transferred into the recipient because leu is the last donor allele. We also know that at least the leu⁺ marker was integrated into the endogenote. We want to know how often the other two markers were also integrated so that we can determine the number of recombination events in which arg⁺ or met⁺ was omitted due to double crossover.

The goal now is to count the frequencies of crossovers at different locations. Note that we now have a different situation from the analysis of interrupted conjugation. In mapping by interrupted conjugation, we measure the time of entry of individual loci; to be stably inherited, each marker has to recombine into the recipient chromosome by a double crossover spanning it. However, in the recombinant frequency analysis, we have specifically selected trihybrids as a starting point, and now we have to consider the various possible combinations of the three donor alleles that can be inserted by double crossing over in the various intervals. We know that leu⁺ must have entered and inserted because we selected it, but the leu⁺ recombinants that we select may or may not have incorporated the other donor markers, depending on where the double crossover took place. Hence, the procedure is to first select leu⁺ exconjugants and then isolate and test a large sample of them to see which of the other markers were integrated.

187

Let’s look at an example. In the cross Hfr met⁺ arg⁺ leu⁺ str^s × F⁻ met⁻ arg⁻ leu⁻ str^r, we would select leu⁺ recombinants and then examine them for the arg⁺ and met⁺ alleles, called the unselected markers. Figure 5-17 depicts the types of double-crossover events expected. One crossover must be on the left side of the leu marker and the other must be on the right side. Let’s assume that the leu⁺ exconjugants are of the following types and frequencies:

188

The double crossovers needed to produce these genotypes are shown in Figure 5-17. The first two types are the key because they require a crossover between leu and arg in the first case and between arg and met in the second. Hence, the relative frequencies of these types correspond to the sizes of these two regions between the genes. We would conclude that the leu-arg region is 4 m.u. and that the arg-met is 9 m.u.

In a cross such as the one just described, one type of potential recombinants of genotype leu⁺ arg⁻ met⁺ requires four crossovers instead of two (see the bottom of Figure 5-17). These recombinants are rarely recovered because their frequency is very low compared with that of the other types of recombinants.

The F factor in Hfr strains is generally quite stable in its inserted position. However, occasionally an F factor cleanly exits from the chromosome by a reversal of the recombination process that inserted it in the first place. The two homologous pairing regions on either side repair, and a crossover takes place to liberate the F plasmid. However, sometimes the exit is not clean, and the plasmid carries with it a part of the bacterial chromosome. An F plasmid carrying bacterial genomic DNA is called an F′ (F prime) plasmid.

The first evidence of this process came from experiments in 1959 by Edward Adelberg and François Jacob. One of their key observations was of an Hfr in which the F factor was integrated near the lac⁺ locus. Starting with this Hfr lac⁺ strain, Jacob and Adelberg found an F⁺ derivative that, in crosses, transferred lac⁺ to F⁻ lac⁻ recipients at a very high frequency. (These transferrants could be detected by plating on medium containing lactose, on which only lac⁺ can grow.) The transferred lac⁺ is not incorporated into the recipient’s main chromosome, which we know retains the allele lac⁻ because these F⁺ lac⁺ exconjugants occasionally gave rise to F⁻ lac⁻ daughter cells, at a frequency of 1 × 10^–3. Thus, the genotype of these recipients appeared to be F′ lac⁺/F⁻ lac⁻. In other words, the lac⁺ exconjugants seemed to carry an F′ plasmid with a piece of the donor chromosome incorporated. The origin of this F′ plasmid is shown in Figure 5-18. Note that the faulty excision occurs because there is another homologous region nearby that pairs with the original. The F′ in our example is called F′ lac because the piece of host chromosome that it picked up has the lac gene on it. F′ factors have been found carrying many different chromosomal genes and have been named accordingly. For example, F′ factors carrying gal or trp are called F′ gal and F′ trp, respectively. Because F lac⁺/F⁻ lac⁻ cells are lac⁺ in phenotype, we know that lac⁺ is dominant over lac⁻.

Partial diploids made with the use of F′ strains are useful for some aspects of routine bacterial genetics, such as the study of dominance or of allele interaction. Some F′ strains can carry very large parts (as much as one-quarter) of the bacterial chromosome.

An alarming property of pathogenic bacteria first came to light through studies in Japanese hospitals in the 1950s. Bacterial dysentery is caused by bacteria of the genus Shigella. This bacterium was initially sensitive to a wide array of antibiotics that were used to control the disease. In the Japanese hospitals, however, Shigella isolated from patients with dysentery proved to be simultaneously resistant to many of these drugs, including penicillin, tetracycline, sulfanilamide, streptomycin, and chloramphenicol. This resistance to multiple drugs was inherited as a single genetic package, and it could be transmitted in an infectious manner—not only to other sensitive Shigella strains, but also to other related species of bacteria. This talent, which resembles the mobility of the E. coli F plasmid, is an extraordinarily useful one for the pathogenic bacterium because resistance can rapidly spread throughout a population. However, its implications for medical science are dire because the bacterial disease suddenly becomes resistant to treatment by a large range of drugs.

189

From the point of view of the geneticist, however, the mechanism has proved interesting and is useful in genetic engineering. The vectors carrying these multiple resistances proved to be another group of plasmids called R plasmids. They are transferred rapidly on cell conjugation, much like the F plasmid in E. coli.

In fact, the R plasmids in Shigella proved to be just the first of many similar genetic elements to be discovered. All exist in the plasmid state in the cytoplasm.

190

These elements have been found to carry many different kinds of genes in bacteria. Table 5-2 shows some of the characteristics that can be borne by plasmids. Figure 5-19 shows an example of a well-traveled plasmid isolated from the dairy industry.

Engineered derivatives of R plasmids, such as pBR 322 and pUC (see Chapter 10), have become the preferred vectors for the molecular cloning of the DNA of all organisms. The genes on an R plasmid that confer resistance can be used as markers to keep track of the movement of the vectors between cells.

On R plasmids, the alleles for antibiotic resistance are often contained within a unit called a transposon (Figure 5-20). Transposons are unique segments of DNA that can move around to different sites in the genome, a process called transposition. (The mechanisms for transposition, which occurs in most species studied, will be detailed in Chapter 15.) When a transposon in the genome moves to a new location, it can occasionally embrace between its ends various types of genes, including alleles for drug resistance, and carry them along to their new locations as passengers. Sometimes, a transposon carries a drug-resistance allele to a plasmid, creating an R plasmid. Like F plasmids, many R plasmids are conjugative; in other words, they are effectively transmitted to a recipient cell during conjugation. Even R plasmids that are not conjugative and never leave their own cells can donate their R alleles to a conjugative plasmid by transposition. Hence, through plasmids, antibiotic-resistance alleles can spread rapidly throughout a population of bacteria. Although the spread of R plasmids is an effective strategy for the survival of bacteria, it presents a major problem for medical practice, as mentioned earlier, because bacterial populations rapidly become resistant to any new antibiotic drug that is invented and applied to humans.

191