Human immunodeficiency virus (HIV), the causative agent of AIDS, kills cells of the immune system. The development of a vaccine has been unsuccessful, because the surface glycoproteins targeted by antibodies change rapidly, in part due to the extremely high mutation rate of HIV (about one replication mistake in every 10,000 nucleotides of HIV genome per generation). However, there has been substantial success in the development of drugs targeting HIV-encoded enzymes that are essential for viral propagation. HIV is a retrovirus, an RNA virus that converts, or reverse transcribes, its RNA genome into DNA. Before HIV was discovered, various laboratories had already studied the life cycle of retroviruses, many of which cause cancer in humans and other animals. With the arrival of HIV, researchers studying this new virus had a head start—a vast amount of established research that had already identified key enzymes required for retroviral propagation. Among these enzymes is a protease that digests long, precursor polypeptides into smaller, active viral proteins.

The usual route for developing a drug that inhibits enzyme activity starts with the random screening of hundreds of thousands of chemical compounds. Possible inhibitors are then chemically optimized for potency, availability in an oral form, and low toxicity. The U.S. Food and Drug Administration’s approval procedure for the use of a drug in humans usually takes well over a dozen years. This type of process has led to drugs that inhibit certain HIV enzymes, including the reverse transcriptase. Rational drug design is another, shorter route to drug discovery. It starts with the target protein structure and designs chemicals that plug the active site and shut the enzyme down. This process short-circuits the random, brute-force approach to searching for chemical inhibitors and has the potential to cut years off the drug-discovery process.

An astounding success in rational drug design has been achieved with the HIV protease. This is partly due to the unique architecture of the protein. HIV protease is a dimer of identical subunits, but unlike typical dimers, which contain two active sites, the HIV protease dimer has a single active site located in the central hydrophobic chamber at the dimer interface. The active site has twofold symmetry; each subunit contributes a catalytic Asp residue, and the two cooperate to hydrolyze the peptide bond (Figure 1). Information obtained from the biochemical studies and crystal structures of HIV protease made possible the rational design of chemical inhibitors. By 1996, the FDA had approved three HIV protease inhibitor drugs: indinavir (Crixivan), ritonavir (Norvir), and saquinavir (Invirase). Notably, all of these drugs are effective, in part, because they mimic the transition state of the proteolytic reaction catalyzed by the enzyme and thus bind to the enzyme virtually irreversibly (Figure 2).

Protease inhibitor drugs have helped reduce the viral titer in the plasma of HIV-infected individuals. However, because of the high mutation rate of the virus, these drugs are effective only for a limited time. Viral mutability can be countered by using different drugs at different times or in a combination cocktail with drugs that inhibit other HIV enzymes. An intensive study of the structure of HIV protease mutants that can circumvent drug action is underway to identify regions of the protein that cannot endure mutation, with the hope of designing inhibitors to which the virus cannot develop resistance.

The study of protein function is, arguably, the oldest subdiscipline in biochemistry and molecular biology. But there is still much to learn. The relatively young science of genomics keeps pointing to genes that encode proteins about which we know little or nothing at all. Some shortcuts to functional discovery are discussed in later chapters.

Rickenberg, H.V., G.N. Cohen, G. Buttin, and J. Monod. 1956. La galactoside-perméase d’Escherichia coli. Ann. Inst. Pasteur 91:829–857.

Jacques Monod began his scientific career in the 1930s, as a graduate student with André Lwoff, studying the capacity of E. coli to adapt its metabolism to different growth conditions. His approach to science was shaped in part by a 1936 trip to Thomas Hunt Morgan’s laboratory at the California Institute of Technology, where he found a stimulating environment dominated by open collaboration and free discussion. Back in France, Monod’s scientific career was slowed but not halted by the outbreak of World War II. While continuing his research, Monod was an active member of the French Resistance. His laboratory at the Sorbonne doubled as a meeting place and propaganda printing press. In the lab, he hit on the idea of an inducer, a cellular signal that would trigger the production of new enzymes needed for adapting to new metabolic circumstances. After the war, he turned back to science full-time. With the help of Lwoff, he obtained a position at the Pasteur Institute in Paris. The metabolism of lactose soon caught his attention.

The disaccharide lactose is cleaved to the monosaccharides glucose and galactose by the enzyme β-galactosidase. Monod found that when lactose was not present in the E. coli growth medium, β-galactosidase was barely detectable in cell extracts. When lactose was added as the sole carbon source for bacterial growth, the levels of β-galactosidase enzyme increased dramatically. Monod wondered how this might occur. In a 1940s’ scientific world in which DNA sequencing, PCR (the polymerase chain reaction), and the structure of DNA were unknown, and messenger RNA still remained to be discovered, the question was not trivial.

Many bacterial enzymes of intermediary metabolism exhibit this inducible pattern. What recommended lactose metabolism as a subject of investigation? Like many other laboratories in the late 1940s, the Pasteur Institute lacked modern cold rooms and other facilities now commonly used to keep proteins active during purification. The one inducible enzyme stable enough to survive the summer heat of an attic lab in Paris was β-galactosidase, later shown to be encoded by a gene called lacZ. The enzyme was also fairly easy to assay. When it was present at high levels, it would cleave an alternative substrate, 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (denoted, more simply, X-gal), and the indole released would color the bacterial colonies blue. The Monod group initiated a study of bacterial lactose metabolism that featured a creative union of biochemistry and genetics.

Mutations were soon found that affected the induction of β-galactosidase. Colonies of these mutants turned blue on X-gal plates even in the absence of lactose. Most of the mutant cells had a mutation in a gene that came to be known as the i (inducer) gene, and later the lacI gene. In these mutations, the lacZ gene was expressed (produced β-galactosidase) all the time; this is known as constitutive expression (Figure 1). Moreover, the appearance of β-galactosidase activity was paralleled by the appearance of a function that transported lactose into the cell, an activity that Monod called a galactoside permease (this is encoded by the gene lacY). The coordinated regulation of two genes, and the loss of regulation in constitutive mutants, led to the concept that some genes regulate other genes. When Monod explained this then-revolutionary idea to his wife, a nonscientist, he was somewhat pained at her reply: “Of course, it is obvious!” Monod now had to figure out what the lacI gene was doing, as the next part of the story recounts.

Jacob, F., and J. Monod. 1961. Genetic regulatory mechanisms in the synthesis of proteins. J. Mol. Biol. 3:318–356.

Next door to Monod at the Pasteur Institute was the laboratory of François Jacob. Wounded as a member of the Free French Forces during the Normandy campaign, Jacob could not pursue his planned career in surgery and instead turned to science. He studied a phenomenon dubbed the “erotic induction” of bacteriophage λ. Bacteriophage λ is a prophage, a bacterial virus that integrates its DNA into its host chromosome and remains quiescent there.

Some years earlier, Joshua Lederberg had discovered the phenomenon of bacterial conjugation (bacterial sex), in which DNA is transferred from a donor to a recipient cell. The transfer is mediated by a genetic element, separate from the bacterial chromosome, known as the F plasmid. In some cells, this plasmid becomes integrated into the host chromosome, creating a strain that mediates high-frequency (Hfr) transfer of chromosomal genes. The transfer begins at the location where the F plasmid is integrated and proceeds linearly along the bacterial chromosome.

Jacob, working with Elie Wollman, used this technique to produce some of the first genetic maps, and they even deduced that the E. coli chromosome was circular. They also noticed that when an Hfr strain also contained an integrated prophage, the bacteriophage λ was transferred into the recipient along with the chromosomal genes. When the dormant bacteriophage λ was thus transferred to a strain that lacked an integrated prophage, the recipient cells were soon lysed. The prophage was somehow activated when it entered the recipient, leading to lytic reproduction—that is, it was induced. It gradually dawned on both Monod and Jacob that the induction phenomena they were studying were closely related, and one of the great scientific collaborations of the twentieth century was born. Their subsequent work defined the regulation of the lactose operon (described in detail in Chapter 20) and established many key regulatory principles.

Monod first liked the idea that mutation in the lacI gene (lacI⁻) produced some kind of inducer that made the addition of lactose unnecessary. If this was true, then the lacI⁻ gene should be dominant over the normal (wild-type) lacI gene (lacI⁺). To get the two variations of lacI (lacI⁻ and lacI⁺) into the same cell, Monod worked with Jacob and Arthur Pardee (on sabbatical leave from the University of California, Berkeley) to use the bacterial conjugation method. The resulting “PaJaMo” experiments at first supported Monod’s idea. When an Hfr strain was used to introduce the lacI⁺ and lacZ⁺ genes into a recipient that had both the lacI⁻ gene and another mutation that prevented β-galactosidase synthesis (lacZ⁻), β-galactosidase production occurred immediately after the lacZ⁺ gene arrived in the recipient. This seemed to indicate that the lacI⁻ gene of the recipient was dominant. However, after an hour or so, β-galactosidase production was halted as the product of the lacI⁺ gene built up. Clearly, lacI⁺ encoded a substance that shut off the lacZ gene. Continued experimentation refined the idea that lacI⁺ encodes some kind of repressor and that the interaction of lactose with the repressor is needed to induce the lacZ gene and the additional genes that are coregulated with it. A similar repressor normally suppresses activation of an integrated λ prophage. The prophage is activated when transferred to the recipient cell during conjugation because no repressor is present in the recipient. Cell lysis results.

The work continued. If lacI was encoding a repressor, this repressor had to interact with something to shut down the lactose genes. Monod and Jacob now mutagenized cells that had two good copies of the lactose genes. It was unlikely that both copies of lacI would be inactivated, but inactivation of one repressor target would be enough to induce the function of one set of lactose genes. The researchers predicted that the resulting mutations would appear at a site on the chromosome distinct from lacI and would lead to constitutive synthesis of the lactose gene products. The mutants were found, and they defined a site that Monod and Jacob called the operator.

In a famous 1961 paper, Jacob and Monod laid out these ideas and others as part of their operon model. The concepts have guided our thinking about gene regulation ever since. Other experiments, carried out in parallel, showed that bacteriophage λ also encodes a repressor, and this repressor is needed to keep most other bacteriophage λ genes from being expressed.

Gilbert, W., and B. Müller-Hill. 1966. Isolation of Lac repressor. Proc. Natl. Acad. Sci. USA 56:1891–1898.

By 1961, it was clear that a repressor existed, but not at all clear what it was. It could be RNA or protein, or some other kind of molecule that was synthesized by a lacIenzyme. By 1966, nonsense-type mutations, those that prematurely halt protein synthesis, had been found in the lacI gene that produced a lacI⁻ effect. This finding had convinced most scientists that the repressor was a protein, but it was still necessary to isolate one to prove it. The isolation work was carried out at Harvard, by physicist-turned-biologist Walter Gilbert.

To isolate a protein, you need a way to measure its presence (an assay), but that was difficult to construct. The repressor presumably bound to operator DNA, but that DNA sequence was not yet defined. The repressor also bound to the inducer (then thought to be lactose, later shown to be allolactose, a metabolic by-product of lactose metabolism). Because lactose was metabolized in the cell, it would be destroyed in crude cell extracts and thus would be of little use to the researchers. Gilbert turned to isopropyl β-d-1-thiogalactopyranoside (IPTG), a molecule known to induce the lactose operon without being metabolized in the cell.

Using a technique known as equilibrium dialysis, the researchers suspended a dialysis bag containing a bacterial cell extract in a solution containing radioactive IPTG. Pores in the dialysis bag were sufficiently small to prevent the protein molecules from escaping, but smaller molecules such as IPTG could diffuse through. If the lactose repressor was present in the extract, it would bind to IPTG, and the concentration of IPTG would increase inside the dialysis bag relative to the surrounding solution. The assay eventually worked, but only after Gilbert and his colleagues developed methods to greatly increase its sensitivity. In 1966, they reported their detection of a repressor protein that bound IPTG.

The lactose repressor was finally purified to homogeneity by the Gilbert group and several other research groups in the early 1970s. Gilbert used the repressor problem to develop several new methods to define the DNA binding sites of proteins that bound DNA specifically. For example, his group found that when dimethylsulfate (DMS) was used to modify adenine or guanine residues in the DNA, the adjacent DNA backbone became more labile to cleavage in mild alkali. DMS methylates guanine residues at N-7 and adenine residues at N-3. If the treatment is done briefly so that only one A or G per DNA strand is labeled, on average, then subsequent cleavage will break each DNA strand at just one position. If all the strands are radioactively labeled at the same end, a banding pattern is produced that acts as a map of the A and G residues in the DNA strand (Figure 2; note that G residues are methylated preferentially and generate stronger bands). If a protein is bound to the DNA prior to addition of DMS, it partially protects the adenines and guanines from methylation in the region where it is bound. The resulting disruption of the base-methylation pattern helps to define the binding site of the protein; in Figure 2, the protein is the lactose repressor.

In this technique, the control lane (lane C in Figure 2) proved to be as important as the protein that was the subject of the experiment. As Gilbert and his colleague Allan Maxam looked at one of these gels, they realized that the gel could be read to reveal the positions of all the A and G residues in the DNA strand. If they could find a technique to break the strands at C and T residues as well, they would have a new way to sequence DNA. The methods were developed and published as a new sequencing technology in 1977. The Maxam-Gilbert DNA-sequencing procedure was eventually supplanted by the Sanger sequencing method, published in the same year; that method, in turn, has given way to newer technologies (see Chapter 7). The effects of all these advances were profound, and they helped define the science of molecular biology in their time.