You are on page 1of 41



3 . 1


C. B. Anfinsen and E. Haber, 1961, Journal of Biological Chemistry 236:1362

By the 1950s, scientists realized that DNA held the code that allowed proteins to be synthesized. Nevertheless, how a chain of amino acids folds into a fully functional protein, with the proper three-dimensional structure, remained a mystery. A mechanism must exist to assure the proper folding of the protein. But where did that information come from? In 1957, Christian Anfinsen published the first evidence that the information for proper folding was held within the protein itself.

the most thermodynamically stable under in vivo conditions. In other words, if the intracellular conditions could be mimicked in a test tube (in vitro), then a protein would naturally fold into its active conformation. He began his work on a secreted enzyme, bovine pancreatic ribonuclease, and studied its ability to properly fold outside of the cell.

The Experiment
Proteins perform a wide variety of functions in the cell. Regardless of its function, a protein must be properly folded to carry out its biological role. For protein folding studies it is best to study an enzyme whose biological activity can be easily monitored by performing a test, or assay of its activity in vitro. Anfinsen chose a small, secreted protein, the enzyme ribonuclease, in which he could monitor proper folding by assaying its ability to catalyze the cleavage of RNA. Ribonuclease, a secreted protein, is active under oxidizing conditions in vitro. The tertiary structure of active ribonuclease is held together by four disulfide bonds or bridges. Adding a reducing agent reduces the disulfide bond between two cysteine side chains to two free sulfhydryl groups, and can disrupt this covalent interaction. Complete denaturation of ribonuclease requires treatment with a reducing agent. Anfinsen monitored the reduction of ribonuclease by measuring the number of free sulfhydryl groups present in the protein. In the oxidized state, there are no free sulfhydryl groups in ribonuclease because each cysteine residue is involved in a disulfide bond. In the completely reduced state, on the other hand, ribonuclease contains eight free sulfhydryl groups. Anfinsen exploited this difference to assess the extent of reduction by using a spectrophotometric assay to titrate, or count, the number of free sulfhydryl groups.

Proteins are made from combinations of 20 amino acids that then fold into complex structures. The unfolded amino acid chain is called the primary structure. To have biological activity, the protein must fold into proper secondary and tertiary structures. These structures are held together by interactions between the side chains and backbone atoms of the amino acids, including hydrogen bonds, hydrophobic interactions, and, at times, covalent bonds. How these higher structures form had long been a mystery. Does the protein fold correctly as it is synthesized or does it require the action of other proteins to correctly fold it? Can it correctly fold on its own spontaneously? In the 1950s, Anfinsen was a biochemist interested in the proper folding of proteins. Specifically, he was investigating the formation of disulfide bridges, which are covalent bonds between cysteine side chains that serve as one of the major anchors holding together the structure of secreted proteins. He believed that the protein itself contained all the information necessary for proper protein folding. He proposed the thermodynamic hypothesis, which stated that the biologically active structure of a protein was also

To study protein folding outside the cell, one must first denature the protein. Proteins are easily denatured by heat, mechanical disruption such as shaking, and chemical treatment. Proteins with disulfide bridges require an additional measure of treatment with a reducing agent to break apart these covalent bonds. To denature ribonuclease, Anfinsen first reduced the disulfide bridges with thioglycolic acid. He then denatured the protein by using a high concentration of urea and incubating the solution at room temperature. He demonstrated that this treatment rendered the enzyme inactive by showing that ribonuclease was now unable to catalyze the cleavage of RNA. Using the spectrophotometric assay, he went on to show that the inactive ribonuclease contained eight sulfhydryl groups, which corresponded to the four broken disulfide bridges. With a completely reduced, denatured protein in hand, Anfinsen then could ask: Can a denatured enzyme correctly fold in vitro and become active again? To find the answer, Anfinsen allowed a solution of reduced, denatured ribonuclease to oxidize. He removed the urea from the denatured enzyme by precipitation. Next, he resuspended the urea-free denatured ribonuclease in a buffered solution and incubated it for two to three days. Exposure to molecular oxygen in the atmosphere oxidized the cysteine residues. He then compared the activity of this renatured ribonuclease to that of the native enzyme. In initial experiments, 1219 percent of the previously inactive protein were able to catalyze the cleavage of RNA once again. Proteins aggregate at high concentrations, which makes it difficult for them to fold properly. By decreasing the overall concentration of ribonuclease in solution, Anfinsen showed that up to 94 percent of the protein could be refolded (see Table 1).


Cell-free Refolding of Ribonuclease


7.0 4.8 2.3 0.9 0.35


31% 70% 75% 77% 94%

[Data adapted from C. B. Anfinsen and E. Haber, 1961, Journal of Biological Chemistry 236:1362.]

The enzyme had folded back to its active conformation outside of the cell, demonstrating that the information for the protein folding is contained in the protein itself.

Through careful experiments, Anfinsen demonstrated that the information required to properly fold a protein is

contained in its primary sequence. His careful analysis of the chemistry of this process answered a fundamental question in biology. He went on to demonstrate the cell-free refolding of other enzymes, including proteins lacking disulfide bridges. While it is possible to properly fold a number of proteins outside of the normal protein-processing machinery in the cell, this process is greatly accelerated in vivo by a number of proteins. Anfinsen continued to study the protein-folding problem. Although the thermodynamic hypothesis does not hold true for all proteins, Anfinsens demonstration of the cellfree refolding of ribonuclease made a mark on the field of biochemistry. In 1972, he received the Nobel Prize for Chemistry for his work.



4 . 3


M. Meselson and W. F. Stahl, 1958, Proc. Natl Sci. USA 44:671

The discovery that the structure of DNA is a double helix, containing two complementary strands of DNA, led to a number of hypotheses about how DNA might be replicated. Although the possible replication mechanisms were relatively easy to deduce, proving which occurs in vivo was a more difficult task. In 1958, Matthew Meselson and Franklin Stahl used the newly developed techniques of density-gradient centrifugation to show that DNA replication proceeds in a semiconservative fashion.

Stahl, along with Jerome Vinograd, developed density-gradient centrifugation, a technique that can separate macromolecules exhibiting very small differences in density. The tools were now available for a definitive test to determine whether DNA replication occurs by a semiconservative or conservative mechanism.

The Experiment
Meselson and Stahl reasoned that if one could label the parental DNA in such a way that it could be distinguished from the daughter DNA, the replication mechanisms could be distinguished. If DNA replication is semiconservative, then after a single round of replication, all DNA molecules should be hybrids of parental and daughter DNA strands. If replication is conservative, then after a single round of replication, half of the DNA molecules should be composed only of parental strands and half of daughter strands. To differentiate parental DNA from daughter DNA, Meselson and Stahl used heavy nitrogen (15N). This isotope contains an extra neutron in its nucleus, giving it a higher atomic mass than the more abundant light nitrogen (14N). Since nitrogen atoms make up part of the purine and pyrimidine bases in DNA, it was easy to label E. coli DNA with 15N by growing the bacteria in a medium containing 15 N ammonium salts as the sole nitrogen source. After several generations of growth, the bacteria contained only 15 N-labeled DNA. Now that the parental DNA was labeled, Meselson and Stahl abruptly changed the medium to one containing 14N as the sole nitrogen source. From this point on, all the DNA synthesized by the bacteria would incorporate 14N, rather than 15N, so that the daughter DNA strands would contain only

During the 1950s, scientists uncovered many biological facts we now take for granted, beginning with the discovery that genetic information is passed on through deoxyribonucleic acid (DNA), and continuing through the elucidation of DNAs three-dimensional structure. As the decade neared a close, biologists were ready to study how DNA passed on genetic information from the parental to the progeny generation. James Watson and Francis Crick had hypothesized, on the basis of their double-helical model of DNA, that replication occurs in a semiconservative fashion. That is, the double helix unwinds, the original parental DNA strands serve as templates to direct the synthesis of the progeny strand, and each of the replicated DNA duplexes contains one old (parental) strand, and one newly synthesized strand, called the daughter strand. Another hypothesis proposed at the time was conservative replication, whereby after replication the parental strands formed one DNA duplex and the two daughter strands formed the second duplex. When these hypotheses were first proposed, little experimental evidence was available to support one over another. In 1957, however, Meselson and

N. As the bacteria continued to grow and replicate their DNA in the 14 N-containing medium, samples were taken periodically and the bacterial DNA was analyzed with the newly developed technique of equilibrium density-gradient centrifugation. In this type of analysis, a DNA sample is mixed with a solution of cesium chloride (CsCl2). During long periods of high-speed centrifugation the CsCl2 forms a gradient, and the DNA migrates to the position where the density of the DNA is equal to that of the CsCl2. If the DNA sample contains molecules of different densities they will migrate to different positions in the gradient. Because 15N has a greater density than 14N, 15N-labeled DNA has a greater density than 14N-labeled DNA. The higher-density (15N) DNA will sediment to a different position than the lower-density (14N) DNA. Hybrid DNA molecules, containing both 15 N and 14N, will sediment at an intermediate density, depending on the ratio of heavy nitrogen to light nitrogen. The Figure illustrates the results obtained by Meselson and Stahl. Before any DNA replication had occurred in the 14N-containing medium, all DNA sedimented as a single species, corresponding to 15N-labeled DNA. As DNA replication proceeded, the amount of (15N)-DNA decreased, and a second DNA species, consisting of hybrid DNA molecules containing 15 N- and 14N-labeled strands, appeared. DNA collected after completion of the first round of replication was found to sediment with the second species. When the DNA produced during a second round of replication was analyzed, two distinct species were observed. One corresponded to hybrid molecules; the other corresponded to 14 N-labeled DNA. With each subsequent round of replication the proportion of hybrid DNA decreased as the


Predicted results Conservative model Parent strands synthesized in 15N Semiconservative model

Actual results

Light (14N) One band: HH One band: HH

Heavy (15N)


First doubling in 14N

Two bands: HH + LL Old strand New strand

One band: HL (hybrid)



L L New strands

Second doubling in 14N

Two bands: HH + LL

Two bands: HL + LL



Experimental demonstration by Meselson and Stahl that DNA replication is semiconservative. After several generations of growth in a medium containing heavy (15N) nitrogen, E. coli was transferred to a medium containing the normal light isotope (14N). Samples were removed from the cultures periodically and analyzed by equilibrium density-gradient centrifugation in CsCl to separate heavy-

heavy (H-H), light-light (L-L), and heavy-light (H-L) duplexes into distinct bands. The actual banding patterns observed were consistent with the semiconservative mechanism. [From H. Lodish et al., 1995,
Molecular Cell Biology, 3rd ed. W. H. Freeman and Company. See M. Meselson and W. F. Stahl, 1958, Proc. Natl. Acad. Sci. USA 44:671; photographs courtesy of M. Meselson.]

amount of 14N-labeled DNA increased. As the diagrams in the Figure show, the sedimentation patterns observed by Meselson and Stahl are consistent only with a semiconservative model of replication.

For Meselson and Stahl to prove that DNA replication proceeds in a semicon-

servative manner, they not only had to design a clear, easily interpretable experiment, but also develop the technology to do it. The beauty of this classic experiment is that each of the possible models would produce distinctly different results, so that interpretation of the experimental data was unambiguous. This study remains a shining example of defining a problem and employing the proper methodology to solve it.

By demonstrating that DNA replication occurs in a semiconservative fashion, Meselson and Stahl opened up the field of DNA replication for indepth research. With the correct model in hand, researchers could now turn to unraveling the precise mechanism of DNA replication. In addition, equilibrium density-gradient centrifugation became a widely used tool for the analysis of complex mixtures of DNA.



4 . 1


M. W. Nirenberg and P. Leder, 1964, Science 145:1399

By the early 1960s molecular biologists had adopted the so-called central dogma, which states that DNA directs synthesis of RNA (transcription), which then directs assembly of proteins (translation). However, researchers still did not completely understand how the code embodied in DNA and subsequently in RNA directs protein synthesis. To elucidate this process, Marshall Nirenberg embarked upon a series of studies that would lead to the solution of the genetic code.

Proteins are made from combinations of 20 different amino acids. The genes that encode proteinsthat is, specify the type and linear order of their component amino acidsare located in DNA, a polymer made up of only four different nucleotides. The DNA code is transcribed into RNA, which is also composed of four nucleotides. Nirenbergs studies were premised on the hypothesis that the nucleotides in RNA form codewords, each of which corresponds to one of the amino acids found in protein. During protein synthesis, these codewords are translated into a functional protein. Thus, to understand how DNA directs protein synthesis, Nirenberg set out to understand the relationship between RNA codewords and protein synthesis. At the outset of his studies, much was already known about the process of protein synthesis, which occurs on ribosomes. These large ribonuleoprotein complexes can bind two different types of RNA: messenger RNA (mRNA), which carries the exact protein-specifying code from DNA to ribosomes, and smaller RNA molecules now known as transfer RNA (tRNA), which deliver amino acids to ribosomes. tRNAs exist in two forms: those that are covalently attached to a single amino acid, known

as amino-acylated or charged tRNAs, and those that have no amino acid attached, called uncharged tRNAs. After binding of the mRNA and the amino-acylated tRNA to the ribosome, a peptide bond forms between the amino acids, beginning protein synthesis. The nascent protein chain is elongated by the subsequent binding of additional tRNAs and formation of a peptide bond between the incoming amino acid and the end of the growing chain. Although this general process was understood, the question remained: How does the mRNA direct protein synthesis? When attempting to address complex processes such as protein synthesis, scientists divide large questions into a series of smaller, more easily addressed questions. Prior to Nirenbergs study, it had been shown that when phenylalanine-charged tRNA was incubated with ribosomes and polyuridylic acid (polyU), peptides consisting of only phenylalanine were produced. This finding suggested that the mRNA codeword, or codon, for phenylalanine is made up of the nucleosides containing the base uracil. Similar studies with polycytadylic acid (polyrC) and polyadenylic acid (polyrA) showed these nucleosides containing the bases cytadine and adenine made up the codons for proline and lysine, respectively. With this knowledge in hand, Nirenberg asked the question: What is the minimum chain length required for tRNA binding to ribosomes? The system he developed to answer this question would give him the means to determine which amino-acylated tRNA would bind which m-RNA codon, effectively cracking the genetic code.

The Experiment
The first step in determining the minimum length of mRNA required for

tRNA recognition was to develop an assay that would detect this interaction. Since previous studies had shown that ribosomes bind mRNA and tRNA simultaneously, Nirenberg reasoned that ribosomes could be used as a bridge between a known mRNA codon and a known tRNA. When the three components of protein synthesis are incubated together in vitro, they should form a complex. After devising a method to detect this complex, Nirenberg could then alter the size of the mRNA to determine the minimum chain length required for tRNA recognition. Before he could begin his experiment, Nirenberg needed both a means to separate the complex from unbound components and a method to detect tRNA binding to the ribosome. To isolate the complex he exploited the ability of nylon filters to bind large RNA molecules, such as ribosomes, but not the smaller tRNA molecules. He used a nylon filter to separate ribosomes (and anything bound to the ribosomes) from unbound tRNA. To detect the tRNA bound to the ribosomes, Nirenberg used tRNA charged with amino acids that contained a radioactive label, 14C. All other components of the reaction were not radioactive. Since only ribosome-bound tRNA is retained by the nylon membrane, all radioactivity found on the nylon membrane corresponds to tRNA bound to ribosomes. Now a system was in place to detect the recognition between a mRNA molecule and the proper amino-acylated tRNA. To test his system, Nirenberg used polyU as the mRNA, and [14C]phenylalanine-charged tRNA, which binds to ribosomes in the presence of polyU. Ribosomes were incubated with both polyU and [14C]-phenylalanine tRNA for sufficient time to allow both molecules to bind to the ribosomes; the


e Ph

Le u

Trinucleotide UUU








e Trinucleotide and all tRNAs pass through filter

Ar g



Ribosomes stick to filter

Complex of ribosome, UUU, and Phe-tRNA sticks to filter

Assay developed by Marshall Nirenberg and his collaborators for deciphering the genetic code. They prepared 20 E. coli extracts containing all the aminoacyl-tRNAs (tRNAs with amino acid attached). In each extract sample, a different amino acid was radioactively labeled (green); the other 19 amino acids were present on tRNAs but remained unlabeled. Aminoacyl-tRNAs and trinucleotides passed through a nylon filter without binding (left panel ); ribosomes, however, bind to the filter (center panel ). Each of the 64 possible trinucleotides was tested separately for its ability to attract a specific tRNA by adding it with

ribosomes to different extract samples. Each sample was then filtered. If the added trinucleotide causes the radiolabeled aminoacyl-tRNA to bind to the ribosome, then radioactivity is detected on the filter; otherwise, the label passes through the filter (right panel ). By synthesizing and testing all possible trinucleotides, the researchers were able to match all 20 amino acids with one or more codons (e.g., phenyalanine with UUU, as shown here). [From H. Lodish et al., 1995,
Molecular Cell Biology, 3rd ed. W. H. Freeman and Company. See M. W. Nirenberg and P. Leder, 1964, Science 145:1399].

reaction mixtures were then passed through a nylon membrane. When the membranes were analyzed using a scintillation counter, they contained radioactivity, demonstrating that in this system polyU could recognize phenylalanine-charged tRNA. But was this recognition specific for the proper amino-acylated tRNA? As a control, [14C]-lysine- and [14C]-proline-charged tRNAs also were incubated with polyU and ribosomes. After the reaction mixtures were passed through a nylon filter, no radioactivity was detected on the filter. Therefore, the assay measured only specific binding between a mRNA and its corresponding amino-acylated tRNA. Now the minimum chain length of RNA necessary for proper aminoacylated tRNA recognition could be determined. Short oligonucleotides were tested for their ability to bind ribosomes and recognize the appropriate tRNA. When UUU, a trinucleotide, was used, tRNA binding to ribosomes could be

detected. However, when the UU dinucleotides were used, no binding was detected. This result suggested that the codon required for proper recognition of tRNA is a trinucleotide. Nirenberg repeated this experiment on two other homogeneous trinucleotides, CCC and AAA. When these trinucleotides were independently bound to ribosomes, CCC specifically recognized proline charged tRNA and AAA recognized lysine-charged tRNAs. Since none of the three homogeneous trinucleotides recognized other charged tRNAs, Nirenberg concluded that trinucleotides could effectively direct the proper recognition of amino-acylated tRNAs. This study accomplished much more than determining the length of the codon required for proper tRNA recognition. Nirenberg realized that his assay could be used to test all 64 possible combinations of trinucleotides (see Figure). A method for cracking the code was available!

Combined with the technology to generate trinucleotides of known sequence, Nirenbergs assay provided a way to assign each specific amino acid to one or more specific trinucleotides. Within a few years, the genetic code was cracked, all 20 amino acids were assigned at least one trinucleotide, and 61 of the 64 trinucleotides were found to correspond to an amino acid. The final three trinucleotides, now known as stop codons, signal termination of protein synthesis. With the genetic code cracked, biologists could read the gene in the same manner that the cell did. Simply by knowing the DNA sequence of a gene, scientists can now predict the amino acid sequence of the protein it encodes. For his innovative work, Nirenberg was awarded the Nobel Prize in Physiology or Medicine in 1968.



5 . 1


R. K. Saiki, S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, N. Arnheim, 1985, Science 230(4732):13504

In the early 1980s the fruits of the molecular biology revolution were beginning to be realized. Geneticists were uncovering the genetic defects that lead to many hereditary diseases, and the newly burgeoning biotechnology industry was eager to provide physicians with simple diagnostic tests for such diseases. However, the best method available for detecting abnormal genes, Southern hybridization, required sizable DNA samples and several days to perform. In this environment, one of the most powerful molecular biology techniques known was born: the polymerase chain reaction, or PCR.

Researchers in the human genetics department of a young biotechnology company were trying to develop a practical method for the prenatal diagnosis of sickle-cell anemia. The molecular defect that causes most cases of this disease is a single nucleotide change in the sixth codon of the gene encoding the protein -globin, one of the subunits of hemoglobin. Kary Mullis, a molecular biologist at the company, had an idea for a molecular method that would amplifiy specific DNA sequences. The detection of a single nucleotide change, as occurs in sickle-cell anemia, was the perfect test for his ideas. Mulliss idea was an extension of known techniques for synthesizing specific pieces of DNA in vitro using chemically synthesized oligonucleotides and purified DNA polymerase, the enzyme that catalyzes the synthesis of DNA. First, a short oligonucleotide whose sequence was complementary to a portion of the target DNA was synthesized. Next, a fragment of DNA

containing the target sequence was isolated using restriction endonucleases, enzymes that catalyzed the cleavage of DNA at specific sequences. The isolated DNA fragment was then heated to denature the double-stranded helix into two single-stranded DNA molecules. At this point, the oligonucleotide was added to the DNA and allowed to anneal to the complementary region, thereby creating a primer-template complex, one of the substrates for DNA polymerase. The other substrates, the four deoxynucleotide triphosphates (dNTPs), were then added, so that DNA synthesis could occur. Although this method was useful for producing radioactively labeled pieces of DNA, it could not amplify a DNA sequence, only replicate it.

The Experiment
Mullis designed a method that would actually amplify the amount of target DNA, a prerequisite for detecting a small DNA sequence within a large complex sample of genomic DNA. For instance, the human genome conations 3 109 nucleotides of coding sequence. Molecular diagnosis of sicklecell anemia requires the detection of one altered nucleotide in one gene among the rest of the genome. To accomplish this, the region of the genome containing the alteration must be amplified. Basing his work on the sequence of the -globulin gene, which was known, Mullis designed primers that would anneal at sequences both upstream and downstream from the disease causing mutation. One primer was complementary to the coding strand, known as the () strand, the second was complementary to the non-

coding, or (), strand. When the primers were added to a sample of denatured genomic DNA along with DNA polymerase and the four dNTPs, DNA synthesis occurred across the region of the mutation from both of the original strands, producing two new double-stranded DNA molecules. Thus the DNA between the primer sites was doubled, not simply replicated as in the older method. Mullis realized that each cycle of DNA-primer annealing and DNA synthesis would yield twice as much target DNA as the previous cycle (see Figure). A chain reaction would ensue and the amount of DNA in the sample would grow exponentially. He called his technique the polymerase chain reaction (PCR) to reflect the mechanism by which amplification was occurring. The first published test of the PCR made use of upstream and downstream oligonucleotide primers that flanked a 110-bp region of the globin gene; the target region included the mutation found in sickle-cell anemia. These primers were mixed with samples of amniotic fluid that had been previously typed for the presence or absence of the mutation. After the samples were put through 20 cycles of heat denaturation, cooling to allow annealing, and DNA synthesis or primer elongation, the amount of -globin target DNA in the samples was found to be enriched more than one million times (220) compared with the initial samples. The exponential expansion of the DNA was easily demonstrated by comparing the same sample after 15 and 20 cycles. It was clear that the additional five cycles greatly increased the amount of DNA produced in the reaction. Next, Mullis tested the ability of the PCR to detect small quantities of DNA. He found that after

Nucleotide mutated in sickle-cell anemia DNA encoding the -globin gene 5 3 * 3 5 Denature Add primers in excess Upstream primer 3 * * 3

Schematic of the polymerase chain reaction (PCR) to amplify the -globin gene. In this case, one oligonucleotide primer is complementary to the () strand and hybridizes downstream of the mutation that leads to sickle-cell anemia; the other primer is complementary to the () strand and hybridizes upstream of the mutation. Repeated cycles of DNA denaturation, primer annealing, and DNA synthesis amplify the target sequences between the primerbinding sites.

Downstream primer Add DNA polymerase [ dNTPs ]

* *

* * Heat denature Cool to allow primer annealing Repeat reaction

* * * * Repeat reaction for 20 cycles

20 cycles, the -globulin gene could be detected starting with a genomic DNA sample as small as 20 ng, which was 50 times smaller than the samples in the original tests. This finding implied that the PCR could be used in a variety of situations where only a small amount of DNA was available, contributing to the widespread use of the technique today.

Development of the PCR relied on two key insights. First, that a DNA

sequence could be amplified, not just replicated, if synthesis were carried out from both the coding and noncoding strands. Second, that a target DNA sequence would grow like dividing bacteria in a culture if the amplification cycle was repeated several times in succession. By employing this relatively simply methodology, Mullis developed one of the most powerful techniques in molecular biology. The advantages of PCR were obvious from the first report. Almost

instantly, it became a standard technique used in all fields of biology and medicine, as well as the forensic sciences. Today, the technique is known not only to biologists, but also to people in all walks of life. In 1993, just eight years after his first report on the PCR, Kary Mullis was awarded the Nobel Prize in Chemistry for developing this revolutionary technique.



5 . 2


T. J. Kelly and H. O. Smith, 1970, J. Mol. Biol. 51:393

Bacteria exhibit a phenomenon, known as host restriction, whereby they can both recognize and cleave foreign DNA, preventing it from interfering with the bacterial life cycle. By purifying and characterizing one of the enzymes involved in host restriction, Hamilton Smith gave molecular biology one of its most important tools, an enzyme that cleaves DNA at a specific sequence.

At the time of Hamilton Smiths work, host restriction was a well-characterized yet highly intriguing phenomenon. It was well known that DNA from one species of bacteria could not be used to transform a second species of bacteria. When researchers simply mixed DNA from one bacterium with a lysate from a second bacterial species, the DNA was cleaved. The bacteria had evolved a system to recognize and cleave foreign DNA. In 1965, Werner Arber hypothesized that bacteria must produce an enzyme capable of recognizing and cleaving foreign DNA at specific sequences. How did a bacterium determine which DNA was foreign, and which was its own? It seemed unlikely that a bacterium could exclude specific sequences in its genome, from the action of this nuclease. More likely, a bacterium somehow modified its own DNA at these sequences, so it could be spared from cleavage. The existence of a second enzyme was thus hypothesized, one that could modify the DNA by methylation at the site where cleavage occurred, thereby preventing cleavage by the sequence-specific nuclease. With these hypotheses in hand, the hunt for the enzymes could begin. In 1968, Matthew Meselson reported the purification from E. coli of one of these enzymes, now called restriction enzymes

or restriction endonucleases. Although the E. coli enzyme catalyzed the cleavage of non-E.coli DNA, Meselson could not demonstrate that this cleavage was sequence specific. In fact, proving that these bacterial enzymes cleave DNA at a specific sequence would be a tricky matter, as this research was conducted before the advent of the relatively simple DNA-sequencing techniques now available. Following on Meselsons work, Smith set out to purify a second restriction enzyme, this time from H. influenzae, and to demonstrate that it does indeed cleave DNA in a sequencespecific manner.

The Experiment
The first step in the successful purification of a new enzyme is devising an assay that measures the known activity of the enzyme as it is being purified. The activity of a restriction enzyme catalyzes the cleavage of foreign DNA, so this was the logical activity to monitor. To do so, Smith took advantage of the fact that genomic DNA from bacteria is quite viscous; however, as nucleases begin to degrade the bacterial DNA, the overall viscosity decreases. Therefore, Smith could monitor the purification of his restriction enzyme by measuring the decrease in viscosity of a foreign DNA after treatment with a sample of the protein after each step in the purification scheme. Smith mixed cell extracts of H. influenzae with intact DNA from either H. influenzae or the Salmonella bacteriophage P22. Using a device called a viscometer, he measured how the DNA from P22 became less viscous over time, while the H. influenzae DNA displayed no change in viscosity. This would be the assay he would use throughout the purification scheme.

Smith used a variety of established methods to separate bacterial lysates into smaller pools of proteins. Each method separated the lysate on the basis of a different physical property of the proteins (and other biomolecules) that make up the lysate. This allowed the lysate to be divided into subsamples known as fractions. After each step in the purification, every fraction was separately assayed for the ability to cleave P22 DNA. Fractions that contained the enzyme activity were subjected to yet another purification method, and the process was continued until a pure enzyme was obtained. Smith called the purified restriction enzyme endonuclease R. Next Smith determined some of the basic characteristics of endonuclease R. He used endonuclease R to digest DNA from the bacteriophage T7, then estimated the number of sites where the DNA was cleaved. He discovered that endonuclease R did not completely degrade T7 DNA, but rather cleaved it at approximately 40 sites. Since T7 DNA contains approximately 40,000 bases, cleavage occurred at only 0.1 percent of the possible sites. This observation suggested to Smith that Arbers hypothesis was correctthe enzyme was cleaving the DNA at specific sequences. In order to prove that this was the case, Smith had to determine the sequence at which the enzyme cleaved the DNA, which he called the recognition site. With the purified enzyme and evidence of sequence-specific DNA cleavage, Smith focused his attention on determining the sequence of the recognition site. At this time, the 1960s, the only known method of DNA sequencing was to sequentially remove nucleotides from the 5 end of DNA and determine their identity by thin layer chromatography (TLC). Smith devised a scheme to sequence the recognition site by

Recognition site 3 5 Endonuclease R 3 5 3 5 Polynucleotide kinase [32P] ATP 3 5 P* *P Digestion with various nucleases 5 3 P P Alkaline phosphatase 5 3 5 3 5 3

Schematic representation of the method used to determine the nucleotide sequence recognized by endonuclease R. T7 bacteriophage DNA was digested with endonulcease R. After removal of the 5 phosphate, and addition of a 32P label, the 5-end-labeled DNA was digested with a variety of nucleases. 32P-labeled mononucleotides, dinucleotides, and trinucleotides were isolated and analyzed to determine the recognition site sequence. [Adapted from T. J. Kelly and
H. O. Smith, 1970, J. Mol.Biol. 51:393.]

n=1 n=2 n=3

P* P* P*

*P *P *P

Mononucleotides Dinucleotides Trinucleotides

using known enzymes to cleave the ends of a DNA strand into small pieces that could be analyzed by TLC (see Figure). Smith began by labeling the 5 end of endonuclease R-digested DNA with a radioactive marker, 32P. This was accomplished by first treating the DNA with alkaline phosphatase, an enzyme that catalyzes the removal of 5 phosphate groups from polynucleotides. Next, polynucleotide kinase, which catalyzes addition of phosphate to the 5 end of polynucleotides, was used to transfer 32P from labeled ATP to the terminal nucleotide. Now the terminal nucleotide could be easily distinguished from the rest of the nucleotides, by virtue of its specific radioactive label. The DNA was then digested to single nucleotides with a nuclease called pancreatic DNase. The only 32P-labeled nucleotides observed contained adenine (A) and guanine (G). Since no 32Plabeled nucleotide containing cytosine (C) or thymine (T) was detected, Smith deduced that the first base in the recognition sequence must be a purine. To determine the second base in the recognition site, Smith used a nuclease that could not cleave 5 terminal dinucleotides. In other words, the entire DNA sample was digested into single nucleotides except the final two, which remained in dinucleotide form. Since the DNA previously had been cleaved

with endonuclease R, the 5 terminal dinucleotides are the first two bases in the recognition site. Smith first separated the dinucleotides from the single nucleotides. When he analyzed the dinucleotides by TLC, he found only two species of dinucleotides that carried the 32 P label. The identity of the 32P-labeled dinucleotides was determined by comparing their migration to that of dinucleotides of known sequence. One of the species displayed the same migration as the dinucleotide GA; the other migrated with the dinucleotide AA. Smith concluded that the second base in the recognition sequence was adenine. Analysis of the rest of the recognition site would not be so easy, but Smiths persistence paid off. Using a similar but slightly more complicated method, he identified the third base in the recognition site as cytosine. He further showed this to be the end of the recognition sequence by showing that the fourth nucleotide could contain any base. Now he knew digestion of doublestranded DNA with endonuclease R creates several smaller fragments with identical 5 ends, which contain the sequence purine-adenine-cytosine. Since the DNA strands are complementary, the only possible way this could occur is if the enzyme recognized a six-base sequence that appeared the same on either strand, known as a pallindromic

sequence. Therefore, Smith concluded that endonuclease R recognized and cleaved DNA specifically at the sequence GTPyPuAC.

Although the first restriction enzyme had been purified two years before Smith reported his work on endonuclease R, he was the first to demonstrate sequence-specific cleavage. He then went on to purify and characterize the methylase that allows DNA from H. influenzae to escape cleavage. By using these sequence-specific restriction enzymes, researchers could now cleave DNA at specific sites. The impact of restriction enzymes on biological research cannot be overstated. Early on, these enzymes were used for mapping plasmid and phage DNA. Now they are routinely used for probing the structure of both specific genes and of DNA from individuals. In addition, they are primary reagents in the construction of gene expression vectors, allowing DNA from different sources to be cleaved at specific sequences, then joined with similarly cleaved DNA. The results are seen everyday in laboratories employing recombinant DNA technologies. In 1978, Hamiliton Smith was awarded the Nobel Prize in Physiology or Medicine in recognition of his powerful discovery.



5 . 3


R. L. Brinster et al., 1981, Cell 27:223231

In the span of three years, 19801982, the notion of expressing foreign proteins in mice went from an idea to a reality. During this time, several laboratories worked furiously to introduce new genes and express exogenous proteins, first in mouse embryonic stem cells and then in full-grown mice. Ralph Brinster and Richard Palmiter were among the pioneers in this field when, in 1981, they first demonstrated the robust expression of a viral gene in a transgenic mouse.

taining viral DNA into fertilized mouse oocytes, then detecting the viral sequences in the newborn mice. This set the stage to determine whether a functional protein could be expressed from a foreign gene incorporated into the mouse genome.

The Experiment
Brinsters challenge was to design the experiment in such a way that it could be easily and unequivocally demonstrated that the mouse was making the foreign protein. To accomplish this, Brinster chose to express an easily assayed enzyme rather than a protein of greater biological interest in his first transgenic mouse. He chose the enzyme thymidine kinase from the herpes simplex virus (HSV), the choice of which offered several advantages. First, the gene came from a human virus; thus its sequence sufficiently differed from the endogenous mouse gene to allow its integration into the mouse genome to be readily demonstrated. Second, the activity of thymidine kinase can be easily assayed by following the conversion of radioactively labeled thymidine to thymidine monophosphate. Finally, an inhibitor of the HSV thymidine kinase activity that does not inhibit the endogenous mouse enzyme was available, allowing the researchers to specifically monitor the activity of the foreign protein. Genes are expressed from DNA sequences upstream of protein-coding regions called promoters. Promoters control where and when a gene is expressed. To express a viral gene in a mouse requires that the biologist remove the gene from the control of the viral promoter and fuse it to a promoter that is active in mouse cells. Brinster collaborated with Palmiter, who had been studying the promoter of the mouse metallothionein-1 (MT-1)

A powerful approach to the study of genes and the proteins they encode is the controlled expression in both cells and whole organisms. Before the advent of recombinant DNA techniques, biologists accomplished this by injecting foreign mRNA into oocytes from frogs and studying the biological activity of the protein encoded by the foreign mRNA. In the 1970s and 1980s, the molecular biology revolution allowed genes to be fused to specific promoters, which would allow them to be expressed in cell line. Whereas biologists became able to study the gene function in cultured cells, they still wanted to study genes in a living organism. This requires the expression of a specific foreign gene in embryonic cells, leading to introduction of the foreign gene into the animals genome, and examination of its function in the organism. In the early 1970s, Brinster demonstrated that foreign genes could be expressed in mice by injecting cancer cells into an early embryonic form known as a blastocyst. This approach, however, made it difficult to express a specific gene in the desired cell types. This would require introducing the gene into the mouse genome. In 1980, biologists demonstrated that this was possible by injecting a plasmid con-

gene. Palmiter fused the MT-1 promoter to the HSV thymidine kinase gene. They then could ask whether a viral protein could be expressed in a mouse. To generate the transgenic mouse, Brinster and Palmiter injected the plasmid containing HSV thymidine kinase fused to the MT-1 promoter into the pro-nuclei of fertilized mouse eggs, which they then implanted back into female mice. The scientists mated progeny mice with normal females, and analyzed the resulting progeny for the presence of the HSV thymidine kinase DNA as well as thymidine kinase activity. Using Southern blot analysis, they detected the presence of the MT-1 promoterthymidine kinase gene fusion, known as the transgene. They isolated genomic DNA, then cleaved it with a restriction endonuclease. They proceeded to separate the DNA by agarose gel electrophoresiswhich separates DNA fragments on the basis of sizeand transferred it to a nitrocellulose membrane. The two scientists then hybridized a radioactively labeled probe, specific for the transgene, to the membrane for analysis. This analysis revealed that the transgene had been successfully integrated into the genomes of four progeny mice. Next, to determine whether the transgene expressed a functional protein, Brinster and Palmiter analyzed homogenates from the liver, a tissue where the mouse MT-1 gene is highly expressed, for viral thymidine kinase activity. Liver homogenates from one mouse contained approximately 200 times more thymidine kinase activity than the liver homogenates of its littermates. This mouse was one of the four that had the transgene integrated into its genome. To demonstrate that this increase in activity was a result of viral thymidine kinase expression they treated

Expression of Viral Thymidine Kinase in Transgenic Mice


23-1 23-2


14500 497,000

14700 187,000

[Adapted from R. L. Brinster et al., 1981, Cell 27:223231.]

liver homogenates with an inhibitor that specifically blocks the HSV thymidine kinase activity. Thymidine kinase activity in liver homogenates from the transgenic mouse was markedly reduced by this inhibitor, whereas the activity in homogenates from its nontransgenic littermates was unchanged (see Table). Thus Brinster and Palmiter confirmed

the presence of viral thymidine kinase activity and demonstrated that a foreign protein could be expressed in a mouse.

Progress in embryology and molecular biology had left the field ripe for researchers to experiment with advanc-

ing the expression of foreign proteins in animals. The careful choice of the easily assayed HSV thymidine kinase gene put under the control of the metallothionein promoter allowed Brinster and Palmiter to demonstrate the feasibility of this technique. The ability to generate transgenic mice has been invaluable to the study of gene function in vivo. Before this technology was available, researchers had to find naturally occurring mutations in order to analyze gene function in mice. Now a specific gene could be expressed in mice. Soon genes were fused to promoters that allowed expression in specific tissues. Scientists have generated transgenic mice to analyze the function of a great number of genes, allowing them to determine the roles of the genes in a variety of diseases and biological processes.



8 . 1


K. Kruger et al., 1982, Cell 31:147

For biological systems to function, countless reactions must be catalyzed. These duties are carried out by enzymes, biological macromolecules that readily enhance reaction rates yet remain unconsumed by the reaction. For many years only proteins were believed to possess sufficient diversity of functional groups to catalyze the myriad reactions necessary to sustain life. Then, in 1981, Thomas Cech reported that, in at least one case, RNA could do the job.

nuclear extracts, with an eye toward purifying the enzymes that catalyzed the splicing reaction. Although Cech succeeded in this goal, he could have never guessed how the catalysis was taking place.

The Experiment
Cechs plan was to use the in vitro splicing system to purify the RNAsplicing enzymes, a common experimental approach for dissecting complex molecular processes. First, the reaction is characterized in a cell-free system, in this case purified nuclei. Then a means to purify the reaction substrate is developed. In the case of the Tetrahymena rRNA splicing this was relatively easy, because the fulllength rRNA (pre-rRNA) transcripts were abundant in Tetrahymena nuclei and readily purified. Finally, cellular extracts are added back to reconstitute the activity being studied. Since the RNA splicing activity was known to take place in the nucleus, Cech used nuclear extracts. In fact, he could readily see splicing when nuclear extracts were added to rRNA transcripts in a splicing cocktail composed of Mg2 and guanosine triphosphate (GTP). Unexpectedly, splicing also occurred when rRNA transcripts were incubated in the splicing cocktail in the absence of a nuclear extract. This activity was reproducible, leaving open two possibilities: Either the purified pre-rRNA remained associated with an enzyme (i.e., a protein contaminant) or the pre-rRNA was catalyzing its own splicing. The first step in determining which possibility was correct was to see if the rRNA transcripts were truly devoid of protein. Because proteins are notoriously fragile biomolecules, whose activity is easily destroyed by heat,

In eukaryotes and many viruses, genes contain sequences that are initially transcribed, then subsequently removed from RNA as they are not part of the actual coding sequence. These sequences are known as intervening sequences (IVS) or introns. IVS are removed from precursor RNA by a biological process known as splicing. While investigating the splicing of precursor ribosomal RNA (pre-rRNA) genes, transcribed from rRNA genes, Cech made his critical discovery that RNA exhibited catalytic activity. Cech wanted to understand the molecular components of RNA splicing. Rather than examing complex eukaryotic genes, he chose a simple model system, rRNA genes from the ciliated protozoan Tetrahymena thermophilia. By isolating Tetrahymena nuclei, Cech and his coworkers developed a system in which pre-rRNA gene splicing could be studied in vitro. The purified nuclei could perform both transcription of rRNA genes and processing of the large pre-rRNA that initially is formed. Using this system, Cech found that during synthesis of 26s rRNA in Tetrahymena, a 0.4-kb IVS is removed. The next step was to perform pre-rRNA splicing with

chemicals, and proteolytic enzymes, Cech subjected the rRNA transcripts to numerous treatments known to degrade proteins: first, boiling to promote heat denaturation; then, extraction with organic solvents to promote chemical denaturation; finally, incubation with a variety of proteases to promote enzymatic degradation. Still, the pre-rRNA retained its splicing activity. These results strongly suggested that Tetrahymena pre-rRNA is indeed selfsplicing. But a more definitive experiment was needed to convince other researchers that the transcripts were uncontaminated by protein and possessed inherent catalytic activity. Fortunately, the Tetrahymena prerRNA could be produced in vitro using purified RNA polymerase from E. coli. Transcription of the Tetrahymena rRNA gene with a polymerase from a different organism would eliminate the risk that the RNA remained associated with a Tetrahymena enzyme. In this system, the only enzyme ever associated with RNA would be E. coli RNA polymerase, which was readily removed by extraction with organic solvents. Using this system, Cech carefully synthesized the Tetrahymena prerRNA, removed the polymerase, and purified the transcripts. When he incubated this in vitro synthesized prerRNA in the splicing cocktail, analysis of the products showed that once again, the IVS was removed from the precursor (see Figure). This experiment proved that the Tetrahymena pre-RNA was self-splicing, catalyzing the removal of the IVS without the aid of any protein.

Cech called his self-splicing RNA a ribozyme, implying that it was an

Pre-rRNA alone

Pre-rRNA + Mg 2+ and GTP

Demonstration that Tetrahymena thermophilia pre-rRNA can self-splice. Radioactively labeled pre-rRNA was synthesized in vitro using E. coli RNA polymerase and then incubated in neutral buffer or in the presence of Mg2 and GTP, necessary cofactors for the splicing reaction. Depicted here is an autoradiograph of the electrophoresed samples revealing the spliced-out IVS in the sample containing splicing cofactors.
[Adapted from Kruger et al., 1982, Cell 31:147.]


RNA enzyme. Although the demonstration of self-splicing RNA was readily accepted by the scientific community, many were skeptical about the notion that RNA was a true catalyst. In subsequent studies, however, Cech was able to engineer the Tetrahymena rRNA IVS such that it could be used as an enzyme, splicing one RNA molecule,

then turning over to splice others. This convinced even the skeptics that RNA can have true catalytic activity. Soon other self-splicing RNAs and other catalytic RNAs were identified. RNA catalysis has become a field of study unto itself, with research on the use of catalytic RNA in both laboratory and medical settings. Furthermore, the

ability of RNA to catalyze biological reactions has evolutionary implications. It is now conceivable that primordial organisms contained only RNA and later evolved the more complex system of proteins. For his pioneering work on RNA catalysis, Cech was awarded the Nobel Prize in Chemistry in 1989.



9 . 1

Beaufay et al., 1964, Biochem J. 92:191

In the 1950s and 1960s scientists used two techniques to study cell organelles: microscopy and fractionation. Christian de Duve was at the forefront of cell fractionation. In the early 1950s he used centrifugation to distinguish a new organelle, the lysosome, from previously characterized fractions: the nucleus, the mitochondrial-rich fraction, and the microsomes. Soon thereafter he used equilibrium-density centrifugation to uncover yet another organelle.

Eukaryotic cells are highly organized and composed of cell structures known as organelles that perform specific functions. Although microscopy has allowed biologists to describe the location and appearance of various organelles, it is of limited use in uncovering an organelles function. To do this, cell biologists have relied on a technique known as cell fractionation. Here, cells are broken open, and the cellular components are separated on the basis of size, mass, and density using a variety of centrifugation techniques. Scientists could then isolate and analyze cell components of different densities, called fractions. Using this method, biologists had divided the cell into four fractions: nuclei, mitochondrial-rich fraction, microsomes, and cell sap. De Duve was a biochemist interested in the subcellular locations of metabolic enzymes. He had already completed a large body of work on the fractionation of liver cells, in which he had determined the subcellular location of numerous enzymes. By locating these enzymes in specific cell fractions, he could begin to elucidate the function of the organelle. He noted that his

work was guided by two hypotheses: the postulate of biochemical homogeneity and the postulate of single location. In short, these hypotheses propose that the entire composition of a subcellular population will contain the same enzymes and that each enzyme is located at a discrete site within the cell. Armed with these hypotheses and the powerful tool of centrifugation, de Duve further subdivided the mitochondrial-rich fraction. First, he identified the light mitochondrial fraction, which is made up of hydrolytic enzymes that are now known to compose the lysosome. Then, in a series of experiments described here, he identified another discrete subcellular fraction, which he called the peroxisome, within the mitochondrial-rich fraction.

The Experiment
De Duve studied the distribution of enzymes in rat liver cells. Highly active in energy metabolism, the liver contains a number of useful enzymes to study. To look for the presence of various enzymes during the fractionation, de Duve relied on known tests, called enzyme assays, for enzyme activity. To retain maximum enzyme activity, he had to take precautions, which included performing all fractionation steps at 0C because heat denatures protein, compromising enzyme activity. De Duve used rate-zonal centrifugation to separate cellular components by successive centrifugation steps. He removed the rats liver and broke it apart by homogenization. The crude preparation of homogenized cells was then subjected to relatively low-speed centrifugation. This initial step separated the cell nucleus, which collects as sediment at the bottom of the tube,

from the cytoplasmic extract, which remains in the supernatant. Next, de Duve further subdivided the cytoplasmic extract into heavy mitochondrial fraction, light mitochondrial fraction, and microsomal fraction. He accomplished separating the cytoplasm by employing successive centrifugation steps of increasing force. At each step he collected and stored the fractions for subsequent enzyme analysis. Once the fractionation was complete, de Duve performed enzyme assays to determine the subcellular distribution of each enzyme. He then graphically plotted the distribution of the enzyme throughout the cell. As had been shown previously, the activity of cytochrome oxidase, an important enzyme in the electron transfer system, was found primarily in the heavy mitochondrial fractions. The microsomal fraction was shown to contain another previously characterized enzyme, glucose-6-phosphatase. The light mitochondrial fraction, which is made up of the lysosome, showed the characteristic acid phosphatase activity. Unexpectedly, de Duve observed a fourth pattern when he assayed uricase activity. Rather than following the pattern of the reference enzymes, uricase activity was sharply concentrated within the light mitochondrial fraction. This sharp concentration, in contrast to the broad distribution, suggested to de Duve that the uricase might be secluded in another subcellular population separate from the lysosomal enzymes. To test this theory, de Duve employed a technique known as equilibrium density-gradient centrifugation, which separates macromolecules on the basis of density. Equilibrium density-gradient centrifugation can be performed using a number of different

Increasing density of sucrose (g/cm3)

Organelle fraction 1.09 1.11 1.15 1.19 1.22 1.25 Lysosomes (1.12 g/cm3) Mitochondria (1.18 g/cm3) Peroxisomes (1.23 g/cm3)

oxidase, segregated into the same fractions as uricase. Because each of these enzymes either produced or used hydrogen peroxide, de Duve proposed that this fraction represented an organelle responsible for the peroxide metabolism and dubbed it the peroxisome.


De Duves work on cellular fractionation provided an insight into the function of cell structures as he sought to map the location of known enzymes. Before After centrifugation centrifugation Examining the inventory of enzymes in a given cell fraction gave him clues to FIGURE 1 Schematic depiction of
the separation of the lysosomes, mitochondria, and peroxisomes by equilibrium density centrifugation. The mitochondrial-rich fraction from rate-zonal centrifugation was separated in a sucrose gradient, and the organelles were separated on the basis of density. [From Lodish et al.,

its function. His careful work resulted in the uncovering of two organelles: the lysosome and the peroxisome. His work also provided important clues to the organelles function. The lysosome, where de Duve found so many potentially destructive enzymes, is now known to be an important site for degradation of biomolecules. The peroxisome has been shown to be the site of fatty acid and amino acid oxidation, reactions that produce a large amount of hydrogen peroxide. In 1974, de Duve received the Nobel Prize for Physiology and Medicine in recognition of his pioneering work.

5 4 3 Cytochrome oxidase 2 1 20 Relative concentration 5 4 3 Uricase 2 1 20 5 4 3 Acid phosphatase 2 1 20 40 60 80 40 60 80 40 60 80

Molecular Cell Biology, 3d ed., W. H. Freeman and Company, p. 166.]

gradients, including sucrose and glycogen. In addition, the gradient can be made up in either water or heavy water, which contains the hydrogen isotope deuterium in place of hydrogen. In his experiment de Duve separated the mitochondrial-rich fraction prepared by rate-zonal centrifugation in each of these different gradients (Figure 1). If uricase were part of a separate subcellular compartment, it would separate from the lysosomal enzymes in each gradient tested. De Duve performed the fractionations in this series of gradients, then performed enzyme assays as before. In each case, he found uricase in a separate population than the lysosomal enzyme acid phosphatase and the mitochondrial enzyme cytochrome oxidase (Figure 2). By repeatedly observing uricase activity in a distinct fraction from the activity of the lysosomal and mitochondrial enzymes, de Duve concluded that uricase was part of a separate organelle. The experiment also showed that two other enzymes, catalase and Damino acid

Percent height in tube FIGURE 2 Graphical representation of the enzyme analysis of products from a

sucrose gradient. The mitochondrial-rich fraction was separated as depicted in Figure 10.1, and then enzyme assays were performed. The relative concentration of active enzyme is plotted on the y axis; the height in the tube is plotted on the x axis. The peak activities of cytochrome oxidase (top) and acid phosphatase (bottom) are observed near the top of the tube. The peak activity of uricase (middle) migrates to the bottom of the tube. [Adapted from Beaufay et al., 1964, Biochem J. 92:191.]



1 1 . 1


J. Skou, 1957, Biochem. Biophys. Acta 23:394

In the mid-1950s Jens Skou was a young physician researching the effects of local anesthetics on isolated lipid bilayers. He needed an easily assayed membrane-associated enzyme to use as a marker in his studies. What he discovered was an enzyme critical to the maintenance of membrane potential, the Na/K ATPase, a molecular pump that catalyzes active transport.

branes derived from squid axons contained a membrane-associated enzyme that could hydrolyze ATP. Thinking that this would be an ideal enzyme for his purposes, Skou set out to isolate such an ATPase from a more readily available source, crab leg neurons. It was during his characterization of this enzyme that he discovered the proteins function.

The Experiment Background

During the 1950s many researchers around the world were actively investigating the physiology of the cell membrane, which plays a role in a number of biological processes. It was well known that the concentration of many ions differs inside and outside the cell. For example, the cell maintains a lower intracellular sodium (Na) concentration and higher intracellular potassium (K) concentration than is found outside the cell. Somehow the membrane can regulate intracellular salt concentrations. Additionally, movement of ions across cell membranes had been observed, suggesting that some sort of transport is system is present. To maintain normal intracellular Na and K concentrations, the transport system could not rely on passive diffusion because both ions must move across the membrane against their concentration gradients. This energy-requiring process was termed active transport. At the time of Skous experiments, the mechanism of active transport was still unclear. Surprisingly, Skou had no intention of helping to clarify the field. He found the Na/K ATPase completely by accident in his search for an abundant, easily measured enzyme activity associated with lipid membranes. A recent study had shown that memSince the original goal of his study was to characterize the ATPase for use in subsequent studies, Skou wanted to know under what experimental condition its activity was both robust and reproducible. As often is the case with the characterization of a new enzyme, this requires careful titration of the various components of the reaction. Before this can be done, one must be sure the system is free from outside sources of contamination. In order to study the influence of various cations, including three that are critical for the reactionNa, K, and Mg2Skou had to make sure that no contaminating ions were brought into the reaction from another source. Therefore all buffers used in the purification of the enzyme were prepared from salts that did not contain these cations. An additional source of contaminating cations was the ATP substrate, which contains three phosphate groups, giving it an overall negative charge. Because stock solutions of ATP often included a cation to balance the charge, Skou converted the ATP used in his reactions to the acid form so that balancing cations would not affect the experiments. Once he had a well-controlled environment, he could characterize the enzyme activity. These precautions were fundamental to his discovery.

Skou first showed that his enzyme could indeed catalyze the cleavage of ATP into ADP and inorganic phosphate. He then moved on to look for the optimal conditions for this activity by varying the pH of the reaction, and the concentrations of salts and other cofactors, which bring cations into the reaction. He could easily determine a pH optimum as well as an optimal concentration of Mg2, but optimizing Na and K proved to be more difficult. Regardless of the amount of K added to the reaction, the enzyme was inactive without Na. Similarly, without K, Skou observed only a lowlevel ATPase activity that did not increase with increasing amounts of Na. These results suggested that the enzyme required both Na and K for optimal activity. To demonstrate that this was the case, Skou performed a series of experiments in which he measured the enzyme activity as he varied both the Na and K concentrations in the reaction (Figure 1). Although both cations clearly were required for significant activity, something interesting occurred at high concentrations of each cation. At the optimal concentration of Na and K, the ATPase activity reached a peak. Once at that peak, further increasing the concentration did not affect the ATPase activity. Na thus behaved like a classic enzyme substrate, with increasing input leading to increased activity until a saturating concentration was achieved, at which the activity plateaued. K, on the other hand, behaved differently. When the K concentration was increased beyond the optimum, ATPase activity declined. Thus while K was required for optimal activity, at high concentrations it inhibited the enzyme. Skou reasoned that the enzyme must have


40 Mg 6 mM/ I



K 20 mM/ I

K 120 mM/ I
M/ I K 20 0 m


NaCl 40 mM/ I


K 350 m

M/ I




NaCl 20 mM/ I

K 3 mM/ I 20

Mg 6 mM/ I 10 NaCl 3 mM / I NaCl 0 mM/ I 0 0 20 40 60 KCl mM / I 80 100 120 0 0 50 K 0 mM/ I NaCl 10 m M/ I 10

100 NaCl mM/ I



FIGURE 1 Demonstration of the dependence of the Na/K

ATPase activity on the concentration of each ion. The graph on the left shows that increasing K leads to an inhibition of the ATPase activity. The graph on the right shows that with increasing Na , the

enzyme activity increases up to a peak and then levels out. This graph also demonstrates the dependence of the activity on low levels of K . [Adapted from J. Skou, 1957, Biochem. Biophys. Acta 23:394.]

separate binding sites for Na and K. For optimal ATPase activity, both must be filled. However, at high concentrations K could compete for the Na-binding site, leading to enzyme inhibition. He hypothesized that this enzyme was involved in active transport, that is, the pumping of Na out of the cell, coupled to the import of K into the cell. Later studies would prove that this enzyme was indeed the pump that catalyzed active transport. This finding was so exciting that Skou devoted his subsequent research to studying the enzyme, never using it as a marker, as he initially intended.

Skous finding that a membrane ATPase used both Na and K as substrates was the first step in understanding active transport on a molecular level. How did Skou know to test both Na and K? In his Nobel lecture in 1997, he explained that in his first attempts at characterizing the ATPase, he took no precautions to avoid the use of buffers and ATP stock solutions that contained Na and K. Pondering the puzzling and unreproducible results that he obtained led to the realization that contaminating salts might be influencing the reaction. When he

repeated the experiments, this time avoiding contamination by Na and K at all stages, he obtained clear-cut, reproducible results. The discovery of the Na/K ATPase had an enormous impact on membrane biology, leading to a better understanding of the membrane potential. The generation and disruption of membrane potential forms the basis of many biological processes, including neurotransmission and the coupling of chemical and electrical energy. For this fundamental discovery, Skou was awarded the Nobel Prize for Chemistry in 1997.



1 4 . 1


J. Jamieson and G. Palade, 1966, Proc. Natl. Acad. Sci. USA 55(2):424431

The advent of electron microscopy allowed researchers to see the cell and its structures at an unprecedented level of detail. George Palade utilized this tool not only to look at the fine details of the cell but also to analyze the process of secretion. By combining electron microscopy with pulse-chase experiments, Palade uncovered the path proteins follow to leave the cell.

tron microscopy of intracellular organelles. His work led to the seminal finding that secreted proteins travel within vesicles from the ER to the Golgi complex and then to the plasma membrane.

The Experiment
Palade wanted to identify which cell structures and organelles participate in protein secretion. To study such a complex process, he carefully chose an appropriate model system for his studies, the pancreatic exocrine cell, which is responsible for producing and secreting large amounts of digestive enzymes. Because these cells have the unusual property of expressing only secretory proteins, a general label for newly synthesized protein, such as radioactively labeled leucine, will only be incorporated into protein molecules that are following the secretory pathway. Palade first examined the protein secretion pathway in vivo by injecting live guinea pigs with [3H]-leucine, which was incorporated into newly made proteins, thereby radioactively labeling them. At time points from 4 minutes to 15 hours, the animals were sacrificed, and the pancreatic tissue was fixed. By subjecting the specimens to autoradiography and viewing them in an electron microscope, Palade could trace where the labeled proteins were in cells at various times. As expected, the radioactivity localized in vesicles at the ER at time points immediately following the [3H]-leucine injection and at the plasma membrane at the later time points. The surprise came in the middle time points. Rather than traveling straight from the ER to the plasma membrane, the radioactively labeled proteins appeared to

In addition to synthesizing proteins to carry out cellular functions, many cells must also produce and secrete additional proteins that perform their duties outside the cell. Cell biologists, including Palade, wondered how secreted proteins make their passage from the inside to the outside of the cell. Early experiments suggesting that proteins destined for secretion are synthesized in a particular intracellular location and then follow a pathway to the cell surface employed methods to disrupt cells synthesizing a particular secreted protein and to separate their various organelles by centrifugation. These cell-fractionation studies showed that secreted proteins can be found in membrane-bounded vesicles derived from the endoplasmic reticulum (ER), where they are synthesized, and with zymogen granules, from which they are eventually released from the cell. Unfortunately, results from these studies were hard to interpret due to difficulties in obtaining clean separation of all of the different organelles that contain secretory proteins. To further clarify the pathway, Palade turned to a newly developed technique, high-resolution autoradiography, that allowed him to detect the position of radioactively labeled proteins in thin cell sections that had been prepared for elec-

stop off at the Golgi complex in the middle of their journey. In addition, there never was a time point where the radioactively labeled proteins were not confined to vesicles. The observation that the Golgi complex was involved in protein secretion was both surprising and intriguing. To thoroughly address the role of this organelle in protein secretion, Palade turned to in vitro pulse-chase experiments, which permitted more precise monitoring of the fate of labeled proteins. In this labeling technique, cells are exposed to radiolabeled precursor, in this case [3H]leucine, for a short period known as the pulse. The radioactive precursor is then replaced with its nonlabeled form for a subsequent chase period. Proteins synthesized during the pulse period will be labeled and detected by autoradiography, whereas those synthesized during the chase period, which are nonlabeled, will not be detected. Palade began by cutting guinea pig pancreas into thick slices, which were then incubated for 3 minutes in media containing [3H]-leucine. At the end of the pulse, he added excess unlabeled leucine. The tissue slices were then either fixed for autoradiography or used for cell fractionation. To ensure that his results were an accurate reflection of protein secretion in vivo, Palade meticulously characterized the system. Once convinced that his in vitro system accurately mimicked protein secretion in vivo, he proceeded to the critical experiment. He pulse-labeled tissue slices with [3H]-leucine for 3 minutes, then chased the label for 7, 17, 37, 57, and 117 minutes with unlabeled leucine. Radioactivity, again confined in vesicles, began at the ER, then traveled in vesicles to the Golgi complex and remained in the vesicles



FIGURE 1 The synthesis and movement of guinea pig pancreatic secretory proteins as revealed by electron microscope autoradiography. After a period of labeling with [3H]-leucine, the tissue is fixed, sectioned for electron microscopy, and subjected to autoradiography. The radioactive decay of [3H] in newly synthesized proteins produces autoradiographic grains in an emulsion placed over the cell section (which appear in the micrograph as dense, wormlike granules) that mark the position of newly synthesized proteins. (a) At the end of a 3-minute labeling period autoradiographic grains are over the rough ER. (b) Following a 7-minute chase period with unlabeled leucine, most of the labeled proteins have moved to the Golgi vesicles. (c) After a 37-minute chase, most of the proteins are over immature secretory vesicles. (d) After a 117-minute chase, the majority of the proteins are over mature zymogen granules. [Courtesy of J. Jamieson and G. Palade.]



as they passed through the Golgi and onto the plasma membrane (see Figure 1). As the vesicles traveled farther along the pathway, they became more densely packed with radioactive protein. From his remarkable series of autoradiograms at different chase times, Palade concluded that secreted proteins travel in vesicles from the ER to the Golgi and onto the plasma membrane and that throughout this process, they remain in vesicles and do not mix with the rest of the cell.

Palades experiments gave biologists the first clear look at the stages of the

secretory pathway. His studies on pancreatic exocrine cells yielded two fundamental observations. First, that secreted proteins pass through the Golgi complex on their way out of the cell. This was the first function assigned to the Golgi complex. Second, secreted proteins never mix with cellular proteins in the cytosol; they are segregated into vesicles throughout the pathway. These findings were predicated from two important aspects of the experimental design. Palades careful use of electron microscopy and autoradiography allowed him to look at the fine details of the pathway. Of equal importance was the choice of a cell type devoted to secretion, the pancreatic

exocrine cell, as a model system. In a different cell type, significant amounts of nonsecreted proteins would have also been produced during the labeling, obscuring the fate of secretory proteins in particular. Palades work set the stage for more detailed studies. Once the secretory pathway was clearly described, entire fields of research were opened up to investigation in the synthesis and movement of both secreted and membrane proteins. For this groundbreaking work, Palade was awarded the Nobel Prize for Physiology and Medicine in 1974.



1 5 . 1


M. Rodbell et al., 1971, J. Biol. Chem. 246:1877

In the late 1960s the study of hormone action blossomed following the discovery that cyclic adenosine monophosphate (cAMP) functioned as a second messenger, coupling the hormonemediated activation of a receptor to a cellular response. In setting up an experimental system to investigate the hormone-induced synthesis of cAMP, Martin Rodbell discovered an important new player in intracellular signaling guanosine triphosphate (GTP).

The Experiment
One of Rodbells first goals was to characterize the binding of glucagon to the glucagon receptor in the cell-free rat liver membrane system. First, purified rat liver membranes were incubated with glucagon labeled with the radioactive isotope of iodine (125I). Membranes were then separated from the unbound [125I] glucagon by centrifugation. Once it was established that labeled glucagon would indeed bind to the purified rat liver cell membranes, the study went on to determine if this binding led directly to activation of adenyl cyclase and production of cAMP in the purified rat liver cell membranes. The production of cAMP in the cell-free system required the addition of ATP; the substrate for adenyl cyclase, Mg2; and an ATP-regenerating system consisting of creatine kinase and phosphocreatine. Surprisingly, when the glucagon-binding experiment was repeated in the presence of these additional factors, Rodbell observed a 50 percent decrease in glucagon binding. Full binding could be restored only when ATP was omitted from the reaction. This observation inspired an investigation of the effect of nucleoside triphosphates on the binding of glucagon to its receptor. It was shown that relatively high (i.e., millimolar) concentrations of not only ATP but also uridine triphosphate (UTP) and cytidine triphosphate (CTP) reduced the binding of labeled glucagon. In contrast, the reduction of glucagon binding in the presence of GTP occurred at far lower (micromolar) concentrations. Moreover, low concentrations of GTP were found to stimulate the dissociation of bound glucagon from the receptor. Taken together, these studies suggested that

The discovery of GTPs role in regulating signal transduction began with studies on how glucagon and other hormones send a signal across the plasma membrane that eventually evokes a cellular response. At the outset of Rodbells studies, it was known that binding of glucagon to specific receptor proteins embedded in the membrane stimulates production of cAMP. The formation of cAMP from ATP is catalyzed by a membrane-bound enzyme called adenyl cyclase. It had been proposed that the action of glucagon, and other cAMP-stimulating hormones, relied on additional molecular components that couple receptor activation to the production of cAMP. However, in studies with isolated fatcell membranes known as ghosts, Rodbell and his coworkers were unable to provide any further insight into how glucagon binding leads to an increase in production of cAMP. Rodbell then began a series of studies with a newly developed cell-free system, purified rat liver membranes, which retained both membrane-bound and membraneassociated proteins. These experiments eventually led to the finding that GTP is required for the glucagon-induced stimulation of adenyl cyclase.

GTP alters the glucagon receptor in a manner that lowers its affinity for glucagon. This decreased affinity both affects the ability of glucagon to bind to the receptor and encourages the dissociation of bound glucagon. The observation that GTP was involved in the action of glucagon led to a second key question: Can GTP also exert an affect on adenyl cyclase? Addressing this question experimentally required the addition of both ATP, as a substrate for adenyl cyclase, and GTP, as the factor being examined, to the purified rat liver membranes. However, the previous study had shown that the concentration of ATP required as a substrate for adenyl cyclase could affect glucagon binding. Might it also stimulate adenyl cyclase? The concentration of ATP used in the experiment could not be reduced because ATP was readily hydrolyzed by ATPases present in the rat liver membrane. To get around this dilemma, Rodbell replaced ATP with an AMP analog, 5-adenyl-imidodiphosphate (AMP-PNP), which can be converted to cAMP by adenyl cyclase, yet is resistant to hydrolysis by membrane ATPases. The critical experiment now could be performed. Purified rat liver membranes were treated with glucagon both in the presence and absence of GTP, and the production of cAMP from AMP-PNP was measured. The addition of GTP clearly stimulated the production of cAMP when compared to glucagon alone (Figure 1) indicating that GTP affects not only the binding of glucagon to its receptor but also stimulates the activation of adenylyl cyclase.

Two key factors led Rodbell and his colleagues to detect the role of GTP in signal transduction, whereas previous

1000 Glucagon + GTP 800

FIGURE 1 Effect of GTP on glucagon-stimulated cAMP

production from AMP-PNP by purified rat liver membranes. In the absence of GTP, glucagon stimulates cAMP formation about twofold over the basal level in the absence of added hormone. When GTP also is added, cAMP production increases another fivefold.
[Adapted from M. Rodbell et al., 1971, J. Biol. Chem. 246:1877.]

pmols cAMP


400 Glucagon



10 Minutes



studies had failed to do so. First, by switching from fat-cell ghosts to the rat liver membrane system, the Rodbell researchers avoided contamination of their cell-free system with GTP, a problem associated with the procedure for isolating ghosts. Such contamination would mask the effects of GTP on glucagon binding and activation of adenyl cyclase. Second, when ATP was first shown to influence glucagon binding, Rodbell did not simply accept the plausible explanation that ATP, the substrate for adenyl cyclase, also affects binding of glucagon. Instead, he chose to test the effects on binding of the other common nucleoside triphosphates. Rodbell later noted that he knew commercial preparations of ATP often

are contaminated with low concentrations of other nucleoside triphosphates. The possibility of contamination suggested to him that small concentrations of GTP might exert large effects on glucagon binding and the stimulation of adenyl cyclase. This critical series of experiments stimulated a large number of studies on the role of GTP in hormone action, eventually leading to the discovery of G proteins, the GTP-binding proteins that couple certain receptors to the adenyl cyclase. Subsequently, an enormous family of receptors that require G proteins to transduce their signals were identified in eukaryotes from yeast to humans. These G protein-coupled

receptors are involved in the action of many hormones as well as in a number of other biological activities, including neurotransmission and the immune response. It is now known that binding of ligands to their cognate G proteincoupled receptors stimulates the associated G proteins to bind GTP. This binding causes transduction of a signal that stimulates adenyl cyclase to produce cAMP and also desensitization of the receptor, which then releases its ligand. Both of these affects were observed in Rodbells experiments on glucagon action. For these seminal observations, Rodbell was awarded the Nobel Prize for Physiology and Medicine in 1994.



1 5 . 2


M. T. Kahn and R. Furchgott, 1987, in M. J. Rand and C. Raper, eds., Pharmacology, Elsevier Science Publisher, pp. 341344; R. M. J. Palmer et al., 1987, Nature, 327:524; and L. J. Ignarro et al., Proc. Natl. Acad. Sci. USA 84:9265

For decades scientists have tried to understand how cells work together in tissues, as well as in whole organisms. By the 1980s, the identity of many signaling molecules, the cellular responses they evoked, and many aspects of intracellular signaling pathways were understood. All the known signaling moleculesthe familiar hormones and neurotransmitterswere nongaseous substances, primarily peptides and amino acid derivatives. However, studies on the dilation of blood vessels showed that the gas nitric oxide (NO) could indeed function as a signaling molecule.

The discovery of nitric oxide as a signaling molecule began with studies on the mechanism by which blood vessels relax and constrict, processes known as vasodilation and vasoconstriction. In addition to their desire to understand the basic biology of these processes, scientists recognized its medical importance, as drugs that promote vasodilation could aid in the treatment of cardiovascular diseases. Nitroglycerin, long used to treat angina pectoris, was known to promote vasodilation. When applied to isolated blood vessels, nitroglycerin and other nitrogen-containing compounds had been found to activate a signaling pathway that began by stimulating the production of cyclic guanosine monophosphate (cGMP), and eventually resulted in dilation. There was much interest in discovering the natural signal for this process. In vivo, vasodilation was known to occur after stimulation of vessels by the neurotransmitter acetylcholine. However, uncovering the mechanism of this response was hindered by a puzzling finding by Robert Furchgott. In his research on the constriction and relaxation of blood vessels, Furchgott

was using isolated rabbit aorta as a model system. He found that adding the neurotransmitter acetylcholine to a section of isolated rabbit aorta in vitro caused the blood vessel to constrict, just the opposite of the normal in vivo response. However, when he tried to repeat and expand these studies with another aorta preparation, a different response occurred. Now, adding acetylcholine to the aorta caused it to dilate, or relax. Trying to uncover why the effect of acetylcholine was inconsistent, Furchgott discovered significant differences in the aorta preparations used in the two experiments. In the body, blood vessels are made up of two types of cells: smooth muscle cells that form the vessel wall, and endothelial cells, which line the inside wall facing the vessel lumen. Furchgott found that when an isolated aorta preparation contained endothelial cells as well as smooth muscle cells, the vessel responded to acetylcholine by relaxing. But when the endothelial cells were removed, vasoconstriction was once again seen with acetylcholine treatment. To explain these results, Furchgott proposed that acetylcholine causes the endothelial cells to release a signaling molecule that in turn causes smooth muscles to relax. Dubbing this proposed molecule endotheliumderived relaxation factor, or EDRF, he set out to determine its nature and identity. Subsequent work by Furchgott and two other scientists would reveal that nitric oxide is behind the drug-induced dilation of blood vessels but also the natural physiological process of vasodilation stimulated by acetlycholine.

The Experiments
In his search to identify EDRF, Furchgott initially tested the ability of numerous classical signaling molecules to

induce dilation of isolated aorta sections stripped of endothelial cells, his in vitro assay for EDRF activity. None of the various hormones, prostaglandins, and cyclic nucleotides he tested exhibited EDRF activity. In 1986, Furchgott realized that the only molecule known to elicit vasodilation of isolated blood vessels was nitroglycerin. It had been postulated that the pharmacological action of nitroglycerin is due to release of the gas nitric oxide (NO). Could the elusive EDRF actually be nitric oxide? To test this idea, Furchgott treated isolated blood vessels, stripped of endothelial cells, with nitric oxide produced from acidified NaNO2. He found that the response of these stripped vessels to nitric oxide was similar to the dilation of isolated vessels with their endothelium intact caused by the proposed EDRF release following acetylcholine treatment. This observation suggested he was on the right track. He then reasoned that if EDRF is nitric oxide, the same compounds should inhibit NO activity and EDRF activity. Subsequently, he showed that hemogloblin and other compounds that bind nitric oxide do indeed inhibit both NO-mediated dilation of stripped vessels and EDRF-mediated dilation of intact vessels. These findings led Furchgott to hypothesize that EDRF was nitric oxide. This hypothesis was echoed by a second scientist, Louis Ignarro, who through similar reasoning and experimentation was led to the same model. Meanwhile, a third scientist, Salvador Moncada, independently conducted a critical set of experiments clearly demonstrating that EDRF and nitric oxide elicit the identical biological response and are inhibited by the same compounds. Moncada went on to show that the short half-life of both nitric oxide and EDRF is extended by adding the enzyme superoxide dismutase to


Some of the Evidence Supporting the Identity of EDRF and Nitric Oxide*

Biological Response Effect on blood vessels in vitro Stimulates cGMP production Response to other agents Hemoglobin Superoxide dismutase


Relaxation Yes

Relaxation Yes

Inhibits relaxation Extends half-life

Inhibits relaxation Extends half-life

*EDRF endothelial-derived relaxation factor, which is released from endothelial cells in response to acetylcholine. SOURCE: M. T. Kahn and R. Furchgott, 1987, in M. J. Rand and C. Raper, eds., Pharmacology, Elsevier Science Publisher, pp. 341344; R. M. J. Palmer et al., 1987, Nature 327: 524; and L. J. Ignarro et al., Proc. Natl. Acad. Sci. USA 84: 9265.

the in vitro system. This enzyme catalyzes the conversion of oxygen free radicals, which would normally react with nitric oxide yielding NO3 and oxygen. Based on their identical biological responses and susceptibilities to the same inactivating agents, Moncada concluded that EDRF is nitric oxide. The final proof that EDRF is indeed nitric oxide came in a paper published by Ignarro late in 1987. He had earlier reported biological and inhibitor data similar to those of Furchgott and Moncada (see Table 1). However, he went a step further, realizing that the only way to prove EDRF and nitric oxide were one and the same molecule

would be through chemical identification. To do this, Ignarro treated isolated blood vessels with acetylcholine, then collected and chemically analyzed the surrounding medium. He found nitric oxide in the medium from vessels that retained their endothelial cells, whereas no nitric oxide was detectable in the medium surrounding stripped vessels. This evidence served as undeniable proof that endothelial cells signaled vasodilation through the release of nitric oxide.

While initially a startling and improbable hypothesis, the role of nitric oxide

as a signaling molecule rapidly became an exciting field of research. Soon after these critical experiments, Moncada went on to identify the enzyme that produces nitric oxide. In just a few short years, this unusual signal was implicated in many other biological processes including neurotransmitter release and immunity. These exciting advances were predicated on the willingness of Furchgott and Ignarros to stretch the concept of signaling molecules to include a gas that is unstable in solution. For this foresight, and the experiments resulting in the identification of EDRF as nitric oxide, they shared the Nobel Prize for Physiology and Medicine in 1998.



1 7 . 1


H. Huxley and J. Hanson, 1954, Nature 173:973976

The contraction and relaxation of striated muscles allow us to perform all of our daily tasks. How does this happen? Scientist have long looked to see how fused muscles cells, called myofibrils, differ from other cells that cannot perform powerful movement. In 1954, Jean Hanson and Hugh Huxley published their microscopy studies on muscle contraction, which demonstrated the mechanism by which it occurs.

With these observations in mind, Hanson and Huxley began their study of cross striations in muscle. In a few short years, they united the biochemical data with the microscopy observations and developed a model for muscle contraction that holds true today.

The Experiment
Hanson and Huxley primarily used phase-contrast microscopy in their studies of striated muscles that they isolated from rabbits. The technique allowed them to obtain clear pictures of the sarcomere and to take careful measurements of the A and the I bands. By treating the muscles with a variety of chemicals, then studying them under the phase-contrast microscope, they were able to successfully combine biochemistry with microscopy to describe muscle structure as well as the mechanism of contraction. In their first set of studies, Hanson and Huxley employed chemicals that are known to specifically extract either myosin or actin from myofibrils. First, they treated myofibrils with a chemical that specifically removes myosin from muscle. They used phase-contrast microscopy to compare untreated myofibrils to myosin-extracted myofibrils. In the untreated muscle, they observed the previously identified sarcometic structure, including the darkly colored A band. When they looked at the myosin-extracted cells, however, the darkly colored A band was not observed. Next, they extracted actin from the myosin-extracted muscle cells. When they extracted both myosin and actin from the myofibril, they could see no identifiable structure to the cell under phase-contrast microscopy. From these experiments, they concluded that myosin was located

The ability of muscles to perform work has long been a fascinating process. Voluntary muscle contraction is performed by striated muscles, which are named for their appearance when viewed under the microscope. By the 1950s, biologists studying myofibrils, the cells that make up muscles, had named many of the structures they had observed under the microscope. One contracting unit, called a sarcomere, is made up of two main regions called the A band, and the I band. The A band contains two darkly colored thick striations and one thin striation. The I band is made up primarily of light-colored striations, which are divided by a darkly colored line known as the Z disk. Although these structures had been characterized, their role in muscle contraction remained unclear. At the same time, biochemists also tried to tackle this problem by looking for proteins that are more abundant in myofibrils than in nonmuscle cells. They found muscles to contain large amounts of the structural proteins actin and myosin in a complex with each other. Actin and myosin form polymers that can shorten when treated with adenosine triphosphate (ATP).

primarily in the A band, whereas actin is found throughout the myofibril. With a better understanding of the biochemical nature of muscle structures, Huxley and Hanson went on to study the mechanism of muscle contraction. They isolated individual myofibrils from muscle tissue and treated them with ATP, causing them to contract at a slow rate. Using this technique, they could take pictures of various stages of muscle contraction observed using phase-contrast microscopy. They could also mechanically induce stretching by manipulating the coverslip, which allowed them to also observe the relaxation process. With these techniques in hand, they examined how the structure of the myofibril changes during contraction and stretch. First, Huxley and Hanson treated myofibrils with ATP, then photographed the images they observed under phasecontrast microscopy. These pictures allowed them to measure the lengths of both the A band and the I band at various stages of contraction. When they looked at myofibrils freely contracting, they noticed a consistent shortening of the lightly colored I band, whereas the length of the A band remained constant (Figure 1). Within the A band, they observed the formation of an increasingly dense area throughout the contraction. Next, the two scientists examined how the myofibril structure changes during a simulated muscle stretch. They stretched isolated myofibrils mounted on glass slides by manipulating the coverslip. They again photographed phase-contrast microscopy images and measured the lengths of the A and the I bands. During stretch the length of the I band increased, rather than shortened, as it had in contraction.

I band Z disk Stretched 120%

A band

I band S 2.8 A 1.5 I 1.3 S 2.3 A 1.5 I 0.8 S 2.0 A 1.5 I 0.5 S 1.8 A 1.5 I 0.3 S 1.5 A 1.5 I 0.0

FIGURE 1 Schematic diagram of muscle

Relaxed 100%

Contracted 90%

contraction and stretch observed by Hanson and Huxley. The lengths of the sarcomere (S), the A band (A), and the I band (I) were measured in muscle samples contracted 60 percent in length relative to the relaxed muscle (bottom) or stretched to 120 percent (top). The lengths of the sarcomere, the I band, and the A band are noted on the right. Notice that from 120 percent stretch to 60 percent contraction the A band does not change in length. However, the length of the I band can stretch to 1.3 microns, and at 60 percent contraction, it disappears as the sarcomere shortens to the overall length of the A band. [Adapted from J. Hanson
and H. E. Huxley, 1955, Symp. Soc. Exp. Biol. Fibrous Proteins and Their Biological Significance 9:249.]

Contracted 80%

Contracted 60%

Once again, the length of the A band remained unchanged. The dense zone that formed in the A band during contraction became less dense during stretch. From their observations, Hanson and Huxley developed a model for muscle contraction and stretch (Figure 1). In their model, the actin filaments in the I band are drawn up into the A during contraction, and thus the I band becomes shorter. This allows for increased interaction between the myosin located in the A band and the

actin filaments. As the muscle stretches, the actin filaments withdraw from the A band. From these data, Hanson and Huxley proposed that muscle contraction is driven by actin filaments moving in and out of a mass of stationary myosin filaments.

By combining microscopic observations with known biochemical treatments of muscle fibers, Hanson and Huxley were able to describe the bio-

chemical nature of muscle structures and outline a mechanism for muscle contraction. A large body of research continues to focus on understanding the process of muscle contraction. Scientists now know that muscles contract by ATP hydrolysis, driving a conformational change in myosin that allows it to pull on actin. Researchers are continuing to uncover the molecular details of this process, whereas the mechanism contraction proposed by Hanson and Huxley remains in place.



1 8 . 1


S. Ochs et al., 1969, Science 163:686

The field of neurobiology is filled with fascinating puzzles; neurons also pose interesting questions in the study of cell biology. It is well known that neurons are extremely long cells stretching from the cell bodies in the central nervous system into all areas of the body. But how does the neuron transport neurotransmitters and other biologically important molecules to the ends of the axon? A piece to this puzzle was uncovered in the late 1960s when Raymond Lasek and Sidney Ochs independently described fast axonal transport.

Neurons are highly specialized cells that have several interesting features. A neuron has four main components: the cell body, the dendrites, the axon, and the axon terminus. The vast majority of proteins and membranes are synthesized in the cell body. The axons extend from the cell body to the axon terminus where neurotransmission is carried out. Proteins, membranes, and organelles are needed at the axon terminus for neurotransmission to occur. Therefore, there must be a system to transport biomolecules along the axon. In the late 1960s, researchers took the first steps toward understanding this system of transport by trying to characterize the rate of transport. They found that radioactively labeled amino acids injected into ganglia could be taken up by the cell body of neurons and incorporated into proteins. This allowed the researchers to follow newly synthesized proteins as they were transported to the axon terminus. Using this technique, Lasek and Ochs discovered that not all proteins traveled along the axon at the same rate.

The Experiment
Lasek and Ochs set out independently to study axonal transport. To truly as-

sess the rate of transport, each chose to study the sciatic nerve, which provides a long axon in which to study transport. Each of their experiments involved injecting radioactively labeled leucine ([3H]leucine) into the L7 dorsal root ganglia, the location of the cell bodies in the spinal cord. This strategy allowed the labeled leucine to become incorporated into proteins during protein synthesis, and then the location of the labeled proteins could be determined. They analyzed transport along the axon by removing the nerve at various time points after injection, sectioning the axon into small pieces, and then measuring the amount of radioactivity by scintillation counting. By following this protocol, Lasek and Ochs were each able to determine the rate of transport in the axon. Lasek devised a set of precautions to assure that the movement of radioactivity he observed was due to actual transport and not to passive diffusion. The choice of a cell with a long axon was one such precaution. He also performed a number of important controls. He used a combination of microscopy and autoradiography to demonstrate that free [3H]leucine did not diffuse more than 2 mm from the injection site, but rather it was specifically taken up by the neuronal cell bodies and other cells in the area. He then tested whether the axon itself could take up the radioactively labeled amino acids by injecting [3H]leucine in the ventral root ganglia, an area devoid of cell bodies. If the axon took up radioactive amino acids and incorporated them into protein, then Lasek would find radioactivity as far away from the ventral root injection site as he had found in the dorsal root injection site. One day after injections, he found little radioactivity more than 15 mm away in the ventral root, whereas he found 40 times more in the dorsal

root. These controls assured him that he was looking at the transport of incorporated radioactively labeled leucine in the axon. To characterize transport of radioactively labeled amino acids and/or proteins, Lasek measured the distribution of radioactivity along the axon at time points ranging from 14 hours to 60 days after the initial injection of [3H]leucine. By 6 days, a large amount of the radioactivity had been transported out of the ganglia and down into the axons. As expected, by 60 days the majority of the radioactivity had been transported down the axon. Interestingly, he could detect some radioactivity up to 250 mm from the injection site only 14 hours after the injection, which corresponds to the products being transported through the axon at 500 mm/day. The majority of radioactivity traveled the previously observed rate of only 1.3 mm/day, suggesting that more than one mechanism transported some proteins. Working with a similar system, Ochs further characterized this fast component of axonal transport. While he repeated the experiments Lasek had reported, he looked at transport using shorter time points after injection. Ochs sectioned nerves from 2 to 8 hours after injection (see Figure), which allowed a more complete characterization of fast axonal transport. As Lasek had observed, the majority of the radioactivity remained in the ganglia during these short time points. A small portion, however, traveled rapidly within the axon. After 2 hours, Ochs could detect radioactivity more than 90 mm away from the injection site; by 8 hours, he observed radioactivity up to 150 mm from the injection site. From these data, he estimated the rate of fast axonal transport to be 410 mm/day.

106 105 104 103 10


8 hrs 2 hrs

Fast axonal transport was characterized by observing the movement of radioactively labeled proteins along the length of the axon. Researchers injected [3H]leucine into dorsal root ganglia of animals. The radioactively labeled leucine is incorporated into proteins, which are subsequently transported within the axon. At various time points after injection, the nerves are removed and the axons sectioned and analyzed for the presence of radioactivity. The figure shows distribution of radioactivity throughout the axon at 2 hours (drawn in blue) and 8 hours (drawn in purple) after injection.
[Adapted from S. Ochs et al., 1969, Science 163:686.]

101 30 60 Dorsal root 0 Injection 30 60 Nerve mm 90 120 150

Through a series of carefully controlled experiments, Lasek and Ochs were able to demonstrate that some proteins are transported within the neurons at much faster rates than others are. It is now known that biomolecules and organelles travel in the axon at three different rates. The rapidly transported proteins that Lasek and

Ochs observed were likely part of vesicles or the smooth endoplasmic reticulum that are now known to move by fast axonal transport. These vesicles carry neurotransmitters to the axon terminus. It has been shown that these vesicles are carried along microtubules in a process that requires adenosine triphosphate (ATP). Subsequently scientists have shown that families of mo-

lecular motor proteins, the dyneins and the kinesins, power the movement of vesicles by fast axonal transport. The study of movement along the axonmuch of which is grounded in Lasek and Ochss initial observations of fast axonal transportremains an exciting field for researchers.



1 9 . 1


Nagafuchi, A., et al., 1987, Nature 329:341343

The surfaces of many animal cells are coated with cell adhesion molecules including integral membrane proteins that mediate the cell-cell interactions critical for tissue formation. In the late 1970s and early 1980s, biologists began to identify some of these molecules. During this time, the Japanese scientist Masatoshi Takeichi showed that some molecules mediate cell-cell interactions only in the presence of Ca2 ions. This observation led to the discovery of a new class of adhesion molecules, the cadherins.

the question remained, how are these specific adhesive interactions achieved? In the late 1970s, Masatoshi Takeichi studied the interactions between cells in cultures and attempted to find the conditions under which they would and would not adhere to one another. At the same time, other researchers began to identify specific molecules that mediate cell adhesive interactions. Taken together, these two approaches would lead to the discovery of cadherins, a group of cell adhesive molecules critical for tissue formation during development.

In multicellular organisms, groups of specialized cells come together to form tissues. This grouping of cells is not random; specific cell types must adhere to one another. Specific interactions between cells assure that cellular composition is correct: epithelial cells are found in epithelia, hepatocytes in the liver, and neurons in neuronal tissues such as the brain. During tissue formation, it is sometimes necessary for cells of the same type to interact with one another and avoid interactions with other cell types. In organs, where cells of many types work together, the interaction between different cell types is specific. Clearly, there must be a mechanism to assure that tissues and organs maintain the correct cellular compositions and geometries. Many researchers study the adhesive interactions that occur in tissues using embryonic cells in culture. These cells will adhere to one another via proteinprotein interactions so tightly that a protease must be added to break them apart. Classic experiments performed in the 1960s showed that when different cell types are placed in the same Petri dish, they separate from each other like oil and water. Thus, in cell culture, just as in the body, cells can adhere to cells of the same type and avoid contact with different types of cells. But

The Experiment
In the late 1970s, when Takeichi studied the adhesive properties of a lung cell line in culture, he observed that calcium was critical for some forms of cell adhesion. Similar to other cultured cells, the lung cells would readily dissociate in the presence of the protease trypsin. The dissociated cells would normally reaggregate when the trypsin was washed away. However, when Takeichi attempted to replicate these results in a different laboratory, he found that the cells remained dissociated after trypsin treatment, and once dissociated, the cells would never aggregate again under the new conditions. Puzzled by his difficulty in repeating this seemingly basic experimental procedure, Takeichi looked at the chemical compositions of the solutions used in his new laboratory. He found that the trypsin solution he used in the new laboratory contained EDTA, a chemical that sequesters divalent cations from the solution and thus from the lung cells. Previously, Takeichi had used a solution that did not contain EDTA. Perhaps a divalent cation was involved in the adhesive interactions between these cells? To find out, Takeichi began investigating the effects of divalent cations on the adhesive properties of the lung cells. He found that the cells would

not dissociate in the presence of Ca2 and that dissociated cells would reassociate only when Ca2 was added to the medium. These observations led him to propose that some types of cell adhesions depend on calcium. Next, Takeichi set out to identify the specific molecules involved in Ca2-dependent cell adhesion. He used antibodies raised against cell surface proteins involved in cell-cell adhesions to identify the specific proteins, a strategy similar to that used to identify other cell adhesion proteins. The basis of this strategy is the observation that when cultured cells are treated with these antibodies, the binding sites of adhesion molecules are blocked and the interactions between them cannot take place. As a result, the cells no longer adhere in culture. Once such an antibody is found, researchers use it to identify a specific protein involved in the cell adhesion. At first, Takeichi used the same lung cell line that he used to demonstrate that some cell adhesions are Ca2-dependent. However, he was unable to find an antibody that would block cell adhesion in this cell line. To overcome this problem, Takeichi began studying cell adhesion in a different cell line, a teratocarcinoma cell line, where Ca2-dependent cell adhesion also occurs. He grew cells in the presence of Ca2, used trypsin to dissociate them, and then injected them into rabbits, whose immune system generated antibodies that recognize proteins on the surface of these cells. To purify these antibodies, he treated the teratocarcinoma cells that had not been exposed to calcium with antiserum taken from the injected rabbits. This treatment removed all the antibodies that bound to teratocarcinoma cells in the absence of Ca2. What remained were antibodies that specifically bind proteins in the presence of Ca2 that are on the cell surface.



Fibroblasts expressing E-cadherin adhere in



culture. Cells in (a) and (c) are from a fibroblast cell line growing in culture. Cells in (b) and (d) are fibroblasts from the same cell line transfected with the cDNA encoding E-cadherin. (a) Light micrograph showing that fibroblasts do not form adhesive interactions in culture. Notice how the cells seem to overlap one another. (b) Light micrograph showing fibroblasts expressing E-cadherin in culture. These cells adhere to one another, as demonstrated by the easily definable boundaries between cells. (c) Immunofluorescence experiment showing that fibroblasts in culture do not normally express E-cadherin. (d) Immunofluorescence staining shows that fibroblasts transfected with the cDNA encoding E-cadherin express the molecule on their cell surfaces, suggesting that E-cadherin is in fact mediating cell adhesion. (Nagafuchi, A., et al. [1987]. Nature 329:

To find the molecule involved in Ca -dependent cell adhesion, Takeichi compared cultures grown in the presence of Ca2 with cultures of cells grown in the presence of EDTA. Both groups of cells were dissociated with the protease trypsin before the experiment began. Using a technique called immunoprecipitation, he showed that his antibody specifically interacted with a 140 kDa protein on the surface of the Ca2-treated cells while it did not specifically interact with any protein on the EDTA-treated cells. He named this protein cadherin for calciumdependent adhesion protein. As more cadherins were discovered, the original protein he discovered became known as E-cadherin because it mediates the adhesion of epithelial cells. Takeichis identification of E-cadherin demonstrated that the protein is involved in Ca2-mediated cell adhesion, but his research did not prove that E-cadherin is primarily responsible for this type of adhesion. To do so, Takeichi and colleagues cloned the gene encod2

ing E-cadherin. Once the gene was cloned, he could express E-cadherin in a fibroblast cell line that neither expressed E-cadherin nor demonstrated Ca2-dependent adhesion. Rather than form cell-cell adhesions, fibroblasts grow on top of one another in culture (Figure a). However, fibroblasts that express E-cadherin adhered to one another in the presence of Ca2 (Figure b). Thus, the expression of one protein, Ecadherin, changed the adhesive properties of the fibroblast cell line. Finally, Takeichi used immunofluorescence microscopy to show that the E-cadherin molecules are found at the cell surface of the fibroblasts and thus were probably responsible for their newly acquired adhesive property (Figures c and d).

Over a 10-year period, Takeichi and colleagues used his observation about the dependency of some adhesive interactions on the divalent cation Ca2 to discover a key class of cell surface molecules and to show that they are critical

for Ca2-dependent cell adhesions. The discovery of Ca2-dependent cell adhesion was prompted when an experiment that had always worked, the reassociation of cells that had been dissociated by trypsin, suddenly stopped working in the conditions of the new laboratory. By tracking down the difference between the conditions of the laboratories, Takeichi made the initial observations that led to the discovery of a critical family of cell adhesion molecules. His work shows that sometimes great science comes from what looks like a failed experiment. Today, different cadherins have been identified on various types of tissues from the placenta to neurons. Later experiments would show that each type of cadherin interacts specifically with the same molecule on an adjacent cell; in other words, an E-cadherin interacts with another E-cadherin but not with an N-cadherin. These homophilic interactions provide some of the specificity that allows tissues to form. The discovery of cadherins led the way to our understanding of how tissues form.



2 0 . 1


T. Evans et al., 1983, Cell 33:391.

From the first cell divisions after fertilization to aberrant divisions that occur in cancers, biologists have long been interested in how cells control when they divide. The processes of cell division have been separated into stages known collectively as the cell cycle. While studying early development in marine invertebrates in the early 1980s, Joan Ruderman and Tim Hunt discovered the cyclins, key regulators of the cell cycle.

tein synthesis from the maternal mRNA is required at the earliest stages of development. Ruderman and Hunt, while teaching a physiology course at the Marine Biological Lab in Woods Hole, Massachusetts, began a set of experiments designed to uncover the genes that were expressed at this point as well as the mechanism by which this burst of protein synthesis was controlled.

The Experiment Background

The question of how an organism develops from a fertilized egg continues to drive a large body of scientific research. Whereas such research was classically the concern of embryologists, the developing understanding of gene expression in the 1980s brought new approaches to answer this question. One such approach was to examine the pattern of gene expression in both the oocyte and the newly fertilized egg. Ruderman and Hunt were among the biologists who took this approach to the study of early development. Biologists had well characterized the early development of a number of marine invertebrate systems. During the early stages of development, the embryonic cells grow synchronously, which allows an entire population of cells to be studied at the same stage of the cell cycle. Researchers had established that a large portion of the mRNA in the unfertilized oocyte is not translated. Upon fertilization, these maternal mRNA are rapidly translated. Previous studies had shown that when fertilized eggs are treated with drugs that inhibit protein synthesis, cell division could not take place. This suggested that the initial burst of proIn a collaborative project, Ruderman and Hunt looked at regulation of gene expression in the fertilized egg of the surf clam Spisula solidissima. Whereas it was known that overall protein synthesis rapidly increased upon fertilization, they wanted to find out whether the proteins expressed in the earliest stage of development, the two-cell embryo, were different from those expressed in the unfertilized egg. When either oocytes or two-cell clam embryos are treated with radioactively labeled amino acids, the cell takes up the amino acids, which are subsequently incorporated into newly synthesized proteins. Using this technique, Ruderman and Hunt monitored the pattern of protein synthesis by breaking open the cells, separating the proteins using SDS-polyacrylamide gel electrophoresis (SDS-PAGE), and then visualizing the radioactively labeled proteins by autoradiography. When they compared the pattern of protein synthesis in the oocyte with that in the two-cell embryo, they saw that three different proteins that were either not expressed or expressed at an extremely low level in the oocyte were highly expressed in the embryo. In a subsequent study, Ruderman examined the pattern of protein expression in the oocytes of the

starfish Asterias forbesi as they mature. She again observed the increased expression of three proteins of similar size to those that she and Hunt had seen in surf clam embryos. Soon afterward, in a third study, Hunt examined the changes in protein expression during the maturation and fertilization of sea urchin oocytes. This time he performed the experiment in a slightly different manner. Rather than treating the oocytes and embryo with radioactively labeled amino acids for a set time period, he labeled the cells continuously for more than 2 hours, removing samples for analysis at 10minute intervals. Now, he could monitor the changes in protein expression throughout the early stages of development. As had been shown in other organisms, the pattern of protein synthesis was altered when the sea urchin oocyte was fertilized. Three proteins represented by three prominent bands on an autoradiographwere expressed in the embryos, but not in the oocytes. Interestingly, the intensity of one of these bands changed over time; the band was intense at the early time points, then barely visible after 85 minutes. It increased in intensity again between 95 and 105 minutes. The intensity of the band, representing the amount of the protein in the cell, appeared to be oscillating over time. This suggested that the protein had been quickly degraded and then synthesized again. Because the time frame of the experiment coincided with early embryonic cell divisions, Hunt next asked whether the synthesis and destruction of the protein was correlated with progression of the cell cycle. He examined a portion of cells from each time point under a microscope, counting the number of cells dividing at each time

Protein B 75

FIGURE Fast axonal transport was characterized by

Cleavage index


observing the movement of radioactively labeled proteins along the length of the axon.This figure compares the changing levels of sea urchin cyclin (drawn in blue) with a control protein (drawn in purple) as early embryonic cells progress through the cell cycle. The overall level of cyclin increases over time, and then it is rapidly destroyed as the cells approach division. This pattern appears to repeat through each cell division. Meanwhile, the overall level of the control protein continues to increase throughout the time period of the experiment. [Adapted from T. Evans et al., 1983, Cell 33:391.]



1 hr

2 hrs

point where samples had been taken for protein analysis. Hunt then correlated the amount of the protein present in the cell with the proportion of cells dividing at each time point. He noticed that the level of expression of one of the proteins was highest before the cell divided and lowest upon cell division (see Figure 1), suggesting a correlation with the stage of the cell cycle. When the same experiment was performed in the surf clam, Hunt saw that two of the proteins that he and Ruderman had described previously displayed the same pattern of synthesis and destruction. Hunt called these proteins cyclins to reflect their changing expression through the cell cycle.

The discovery of the cyclins heralded an explosion of investigation into the cell cycle. It is now known that these proteins regulate the cell cycle by associating with cyclin-dependent kinases, which in turn regulate the activities of a variety of transcription and replication factors, as well as other proteins involved in the complex alterations in cell architecture and chromosome structure that occur during mitosis. In brief, cyclin-CDK complexes direct and regulate through the cell cycle. As with so many key regulators of cellular functions, it was soon shown that the cyclins discovered in sea urchins and surf clams are conserved in eukaryotes

from yeast to man. Since the identification of the first cyclins, scientists have identified at least 15 other cyclins that regulate all phases of the cell cycle. In addition to the basic research interest in these proteins, the cyclins central role in cell division has made them a focal point in cancer research. Cyclins are involved in the regulation of several genes that are known to play prominent roles in tumor development. Scientists have shown that at least one cyclin, cyclin D1, is overexpressed in a number of tumors. The role of these proteins in both normal and aberrant cell division continues to be an active and exciting area of research today.



2 1 . 1


H. M. Ellis and H. R. Horvitz, 1986, Cell 91:818

During the development of multicellular organisms, certain cells are destined to die. Scientists have directed much research toward understanding this process of programmed cell death. Some sought to understand why a cell would be destined to die during development, while others asked how a cell regulates this form of death. In 1986, H. Robert Horvitz provided clues by examining the genetics of cell death using a well-characterized model system, the nematode Caenorhabditis elegans.

a dying cell, causing it to remain refractile for hours rather than minutes (see Figure). Horvitz used these mutants to look for more genes involved in the control of programmed cell death.

The Experiment
Geneticists analyze complex processes by looking for organisms that display traits or phenotypes that result from an alteration in normal function of a gene. To perform a genetic analysis of programmed cell death, a geneticist looks either for cells that escape programmed cell death or for cells that undergo programmed cell death when they should survive. Finding such mutant organisms in a naturally occurring population would be difficult because the overall mutation rate of an organism is rather low. To facilitate the search, mutations are induced, often by treating individuals with mutagenic chemicals. This enriches the population for mutants. Because mutagenic chemicals will induce mutations in all genes, not just those involved in the process being studied, scientists carefully devise genetic screens to guide their studies. To analyze the control of cell death in C. elegans, Horvitz designed a genetic screen for mutations that alter the process. In the screen described here, he looked for mutations that allowed cells that would normally die during development to survive. A second important part of the screen was the choice of organisms to study. Rather than look for mutant progeny of wild-type nematodes, Horvitz looked at the progeny of ced-1 mutants. In ced-1 mutants, cells that die by programmed cell death are not engulfed and immediately eliminated from the organism. This causes them to remain highly refractile under Nomarski optics for longer than would a cell in a

Developmental biologists have long noted that some cells die during the normal development of a multicellular organism. This processcalled apoptosis or programmed cell death remained a puzzling phenomenon for many years. Early research in the field concentrated on identifying cells that were fated to die. Until the middle of the 1980s, scientists knew little about the mechanism by which the cell controlled this process. At this time, Horvitz began his investigations into the genetics of programmed cell death in the nematode Caenorhabditis elegans. C. elegans is a powerful model system for studying the genetics of complex developmental processes. It is a small, multicellular organism, made up of just more than 1000 cells, so scientists could trace the developmental lineage of each cell of the organism. Previous studies had defined precisely which cells in C. elegans were fated to die during development. When examined by Nomarski differential contrast microscopy, cells fated to die became highly refractile for a few minutes before cell death occurred. In the late 1970s, researchers isolated two cell death mutants of C. elegans: ced-1 and ced-2. These mutants extend the life of

wild-type organism. By using ced-1 mutants in his studies, Horvitz had an increased time frame to look for mutant nematodes in which the cells normally fated to die escape programmed cell death. In his initial screen for genes involved in programmed cell death, Horvitz treated ced-1 mutant nematodes with a mutagenic chemical and allowed them to reproduce for two generations, producing a genetically defined line that carries mutations on both copies of the chromosome. Because the progeny are homozygous for all mutations, he could now observe the phenotype of mutations that are recessive. Once the organisms were bred to homozygosity, he analyzed the second generation for mutations that affected cell death. Specifically, he compared the highly refractile dying cells in ced-1 nematodes to the same cells in mutagenized progeny. These cells displayed the same phenotype in the majority of larvae examined. In a small number of larvae, including those that harbored mutations in genes that control programmed cell death, these cells do not die, and hence do not become refractile. This initial screen uncovered several recessive mutants that mapped to a single gene that he called ced-3 (see Figure). Horvitz went on to characterize the phenotype of the ced-3 mutants, taking advantage of the fact that the identity of all cells destined to undergo programmed cell death was known. By examining the fate of cells that normally undergo programmed cell death, he showed that mutation in ced-3 blocked this process completely (see Table). He then followed these surviving cells through the C. elegans life cycle. When compared with wildtype nematodes, ced-3 mutants reproduced normally and showed no behavioral abnormalities. The primary


Screening for C. elegans genes involved in programmed cell death were observed using Nomarski differential contrast microscopy. (a) Newly hatched larva carrying a mutation in the ced1 gene. Because mutations in this gene prevent engulfment of dead cells, highly refractile dead cells accumulate, facilitating their visualization. The arrows indicate three highly refractile cells. (b) Newly hatched larva with mutations in both the ced-1 and ced-3 genes. Using the nuclei indicated by the arrowheads as orientation points, one can compare panels a and b. In panel b, notice that the three highly refractile cells seen in panel a are not observed. The absence of refractile dead cells in these double mutants indicates that no cell deaths occurred. Thus ced3 was identified as a gene involved in programmed cell death. [From H. M. Ellis and H. R. Horvitz, 1986, Cell
91:818. Courtesy of Hilary Ellis.]


difference appeared to be the extra cells present in ced-3 mutants due to the absence of programmed cell death. Horvitz proceeded to analyze a number of mutations within the ced-3 gene. Each allele resulted in the identical phenotype, survival of cells that should be destined to die. This suggested to Horvitz that mutations in ced-3 resulted in the loss or decreased expression of an essential gene in the programmed cell death pathway. He had isolated the first gene required for

control of programmed cell death in C. elegans.

The isolation of the ced-3 mutant was merely the first step in Horvitzs efforts to dissect the genetics of programmed cell death in C. elegans. In addition to ced-3, he uncovered two other essential genes in the pathway, as well as 10 other genes that are involved in this process. These genes control all aspects

of the process of cell death, from the initial decision to die, to the killing, engulfment, and degradation of the dead cell. As is often seen, these genes that control cell death in C. elegans have counterparts in higher organisms, including humans. The importance of these genes in the regulation of cell growth in humans has become apparent. The human homologues of two genes isolated by Horvitz have been implicated in cancer.

Mutations in Ced-3 Eliminates Cell Death


Wild type ced-1 ced-3
ND Not determined [Data adapted from H. M. Ellis and H. R. Horvitz, 1986, Cell 44:819.]


ND 28 0.3

13 11.23 0.04



2 2 . 1


C. Nsslein-Volhard and E. Wieschaus, 1980, Nature 287:795

One of the most fascinating questions in developmental biology concerns the proper formation of an embryo. How does a fertilized egg know how to form a complex organism? Scientists have puzzled over how to address this question for a long time. In 1980, Christiane Nsslein-Volhard and Eric Wieschaus first reported studies with the fruit fly Drosophila melanogaster in which a genetic approach was used to address this question.

The Experiment
Geneticists develop systematic methods, known as genetic screens, to search for mutations that affect biological processes. Nsslein-Volhard and Wieschaus had to consider several previous observations on Drosophila development when they designed their screen. First, they knew that genes expressed in the egg, called maternaleffect genes, as well as genes expressed after fertilization in the developing embryo, called zygotically active genes, controlled the early development of an embryo. They chose to focus on isolating mutations that affect the zygotically active genes. Second, they had to consider that the Drosophila genome is diploid, which means that the progeny receive a copy of each gene from both parents. Scientists had previously demonstrated that Drosophila required only a single wild-type copy of most genes in order to develop into a viable fly. This made it likely that recessive mutations in developmentally active genes would not result in embryonic death, the phenotypic screen used by Nsslein-Volhard and Wieschaus. Therefore, they had to breed mutant Drosophila to obtain flies that were homozygous for the mutations of interest. The overall mutation rate in a naturally occurring population is quite low. If a geneticist were to search for mutants in a natural population, he or she would have to examine a large number of individuals. To circumvent this difficulty, Nsslein-Volhard and Wieschaus induced mutations in a Drosophila population at the onset of the screen by feeding a mutagenic chemical to male flies and mating them to a genetically defined population of wild-type female flies in a process known as a genetic cross. The resulting

To examine complex processes such as development of an embryo from a fertilized egg, biologists often collect mutant organisms that differ from the normal (wild-type) organism. To apply this genetic approach to developmental biology, geneticists first look for mutant organisms that display an obvious defect in overall formation. Early work uncovered a number of genes involved in the development of the fruit fly Drosophila melanogaster. In the first genes examined, the mutations resulted in the birth of flies with obvious physical defects, such as the presence of an extra set of wings. Because this approach relied on examining viable flies with physical malformations, it missed many developmentally important genes that, when mutated, result in the death of the fly embryo. In the late 1970s, Nsslein-Volhard and Wieschaus began their pioneering work on the development of Drosophila embryos. They sought to identify as many genes in the developmental process as possible by looking for genes that, when mutated, resulted in the death of the embryonic fly. Their work unveiled several key genes active in the early development of not only Drosophila, but higher organisms as well.

progeny would be heterozygous for any mutations on the chromosomes they received from the father. The heterozygotes were then bred as separate lines, in essence isolating each new mutation in a separate fly stock. Flies within each stock then were crossed with each other to generate homozygous embryos. If the mutation affected a gene needed for embryogenesis, the homozygous embryos would die but could be examined for phenotype. The mutant gene, however, would not be lost as it could be recovered in the heterozygous flies of that same stock. Using this screen, Nsslein-Volhard and Wieschaus amassed a large collection of mutant flies. The next step was to assign the various mutants to specific classes, based on their phenotype. They focused on the segmentation of the larvae. Whereas all mutants in their screen necessarily displayed the phenotype of embryonic lethality, they differed greatly in their segmentation defects. To classify these defects, Nsslein-Volhard and Wieschaus examined the larvae under the microscope. They compared the body pattern of a wild-type viable larva to those of the embryonic lethal mutants. By comparing these patterns, they uncovered three classes of genes that affect segmentation, which they called gap, pair-rule, and segment-polarity. Gap mutants are missing up to eight segments from the overall body, which results in a smaller body due to the death of some of the embryo cells. Three mutantsknirps, hunchback, and the previously characterized Krppelfell into this class, as shown in the accompanying photographs. The next class of mutants, the pair-rule mutants, had deletions of alternating segments of the body, which caused roughly halfnormal-size larval bodies to form. Six

FIGURE 1 A viable, wild-type larva (Normal) is compared with three larvae that have mutations in the gap genes Krppel, hunchback, or knirps. Thoracic segments are labeled T1T3, whereas abdominal segments are designated A1A8. Gap mutants are missing entire segments from the body plan, as illustrated by the labeled segments on the left. [From C. Nsslein-Volhard and E. Wieschaus,
1980, Nature 287:795.]





previously uncharacterized mutants paired, even-skipped, odd-skipped, odd-paired, fushi tarazu, and runt were placed in this class. The final set of mutants, the segment-polarity mutants, had the same overall number of segments as wild-type larvae. But a part of the body pattern within each segment was deleted in each mutant. The deleted portion of the pattern within each segment was replaced by a mirror image of the portion that remained. Nsslein-Volhard and Wieschauss initial screen uncovered six mutants of this class, three of whichgooseberry, hedgehog, and patchedhad not been previously observed.

In the first published report of their screen, Nsslein-Volhard and Wieschaus described 15 mutations that affected segmentation. Of these, only five were

in previously identified genes. When they completed the studyoften referred to as the Heidelberg screens they had identified 139 different genes that, when mutated, resulted in embryonic death. These mutants fell into 17 different classes including the ones described here. Together these genes act in a field of equivalent cells and direct each cell to take on a fate appropriate to its position within the field. In this way some cells form segment anterior, some make bristles, some make sensory organs, and so forth. Genes that control these sorts of organizing cell-fate decisions are called patternformation genes. As molecular techniques evolved, scientists cloned many of these genes and characterized their gene products. These mutants formed the base for the next quarter century of research into the development of Drosophila. Moreover, many vertebrate pattern-formation genes are close homologs of the corresponding

fly genes, and use similar molecular mechanisms to organize cells, tissues, and organs. Thus the work of Nsslein-Volhard and Wieschaus greatly advanced the study of development in all vertebrates, including ourselves. The majority of genes identified by Nsslein-Volhard and Wieschaus encode transcription factors, but their screen also uncovered genes encoding signaling molecules, receptors, enzymes, adhesion molecules, cytoskeleton proteins, and some proteins whose functions remain unknown. Mammalian genes related to some of the Drosophila genes uncovered by Nsslein-Volhard and Wieschaus subsequently were found to be important in human diseases such as cancer and birth defects. In 1995, the Nobel Foundation awarded its prize for Physiology and Medicine to Nsslein-Volhard and Wieschaus for their pioneering work.



2 4 . 1


N. Hozumi and S. Tonegawa, 1976, Proc. Natl. Acad. Sci. USA 73:3629

For decades, immunologists wondered how the body could generate the multitude of pathogen-fighting immunoglobulins, called antibodies, needed to ward off the vast array of different bacteria and viruses encountered in a lifetime. Clearly, these protective proteins, like all proteins, somehow were encoded in the genome. But the enormous number of different antibodies potentially produced by the immune system made it unlikely that individual immunoglobulin (Ig) genes encoded all the possible antibodies an individual might need. In studies beginning in the early 1970s, Susumu Tonegawa, a molecular biologist, laid the foundation for solving the mystery of how antibody diversity is generated.

Research on the structure of Ig molecules provided some clues about the generation of antibody diversity. First, it was shown that an Ig molecule is composed of four polypeptide chains: two identical heavy (H) chains and two identical light (L) chains. Some researchers proposed that antibody diversity resulted from different combinations of heavy and light chains. Although somewhat reducing the number of genes needed, this hypothesis still required that a large portion of the genome be devoted to Ig genes. Protein chemists then sequenced several Ig light and heavy chains. They found that the C-terminal regions of different light chains were very similar and thus were termed the constant (C) region, whereas the N-terminal regions were highly variable and thus were termed the variable (V) region. The sequences of different heavy chains exhibited a similar pattern. These findings suggested that the genome contains

a small number of C genes and a much larger group of V genes. In 1965, W. Dryer and J. Bennett proposed that two separate genes, one V gene and one C gene, encode each heavy chain and each light chain. Although this proposal seemed logical, it violated the well-documented principle that each gene encodes a single polypeptide. To avoid this objection, Dryer and Bennett suggested that a V and C gene somehow were rearranged in the genome to form a single gene, which then was transcribed and translated into a single polypeptide, either a heavy or light chain. Indirect support for this model came from DNA hybridization studies showing that only a small number of genes encoded Ig constant regions. However, until more powerful techniques for analyzing genes came on the scene, a definitive test of the novel twogene model was not possible.

The Experiment
Tonegawa realized that if immunoglobulin genes underwent rearrangement, then the V and C genes were most likely located at different points in the genome. The discovery of restriction endonucleases, enzymes that cleave DNA at specific sites, had allowed some bacterial genes to be mapped. However, because mammalian genomes are much more complex, similar mapping of the genes encoding V and C regions was not technically feasible. Instead, drawing on newly developed molecular biology techniques, Tonegawa devised another approach for determining whether the V and C regions were encoded by two separate genes. He reasoned that if rearrangement of the V and C genes occurs, it must happen during differentiation of Ig-secreting B cells from embryonic cells. Furthermore, if

rearrangement occurs, there should be detectable differences between unrearranged germ-line DNA from embryonic cells and the DNA from Igsecreting B cells. Thus, he set out to see if such differences existed using a combination of restriction-enzyme digestion and RNADNA hybridization to detect the DNA fragments. He began by isolating genomic DNA from mouse embryos and from mouse B cells. To simplify the analysis, he used a line of B-cell tumor cells, all of which produce the same type of antibody. The genomic DNA was then digested with the restriction enzyme BamHI, which recognizes a sequence that occurs relatively rarely in mammalian genomes. Thus, the DNA was broken into many large fragments. He then separated these DNA fragments by agarose gel electrophoresis, which separates biomolecules on the basis of charge and size. Since all DNA carries an overall negative charge, the fragments were separated based on their size. Next, he cut the gel into small slices and isolated the DNA from each slice. Now, Tonegawa had many fractions of DNA pieces of various sizes. He then could analyze these DNA fractions to determine if the V and C genes resided on the same fragment in both B cells and embryonic cells. To perform this analysis, Tonegawa first isolated from the B-cell tumor cells the mRNA encoding the major type of Ig light chain, called . Since a RNA is complementary to one strand of the DNA from which it is transcribed, it can hybridize with this strand forming a RNADNA hybrid. By radioactively labeling the entire mRNA, Tonegawa produced a probe for detecting which of the separated DNA fragments contained the chain gene. He then isolated the 3 end of the

Embryo DNA 250 Whole gene 3' end of gene 200 200 250

B Cell DNA

Whole gene 3' end of gene

150 cpm cpm 100





0 Top of gel


15 Bottom of gel

0 Top of gel


15 Bottom of gel

FIGURE 1 Experimental results showing that the genes encoding the variable (V) and constant (C) regions of light chains are rearranged during development of B cells. These curves depict the hybridization of labeled RNA probes, specific for the

entire gene (V C) and for the 3 end that encodes the C region, to fractions of digested embryonic or B-cell DNA separated by agarose gel electrophoresis. [Adapted from N. Hozumi and S. Tonegawa, 1976, Proc.
Natl. Acad. Sci. USA 73:3629.]

mRNA and labeled it, yielding a second probe that would detect only the DNA sequences encoding the constant region of the chain. With these probes in handone specific for the combined V C gene and one specific for C aloneTonegawa was ready to compare the DNA fragments obtained from B cells and embryonic cells. He first denatured the DNA in each of the fractions into single strands and then added one or the other labeled probe. He found that the C-specific probe hybridized to different fractions derived from embryonic and B-cell DNA (Figure 1). Even more telling, the full-length RNA probe hybridized to two different fractions of the embry-

onic DNA, suggesting that the V and C genes are not connected and that a cleavage site for BamHI lies between them. Tonegawa concluded that during the formation of B cells, separate genes encoding the V and C regions are rearranged into a single DNA sequence encoding the entire light chain (Figure 2).

The generation of antibody diversity was a problem awaiting development of powerful molecular techniques to answer it. Tonegawa went on to clone V-region genes and prove that the rearrangement must occur somatically.

These findings impacted genetics as well as immunology. Where once it was believed that every cell in the body contained the same genetic information, it became clear that some cells take that information and alter it to suit other purposes. In addition to somatic rearrangement, Ig genes undergo a variety of other alterations that allow the immune system to create the diverse repertoire of antibodies necessary to react to any invading organism. Our current understanding of these mechanisms rests on the foundation of Tonegawas fundamental discovery. For this work, he received the Nobel Prize for Physiology and Medicine in 1987.

Embryonic DNA VK 5' CK 3'

FIGURE 2 Schematic diagram of light-chain DNA in

B-Cell DNA VK 5' CK 3'

embryonic cells and B cells that is consistent with Tonegawas results. In embryonic cells, cleavage with the BamHI restriction enzymes (red arrows) produces two different sized fragments, one containing the V gene and one containing the C gene. In B cells, the DNA is rearranged so that the V and C genes are adjacent, with no intervening cleavage site. BamHI digestion thus yields one fragment that contains both the V and C genes.



2 5 . 1


Sambrook et al., Proc. Natl. Acad. Sci. USA, 1968, 59:1290 and 1294

Not many diseases have spurred scientists to research and understand their causes more than have human cancers. Scientists recognized long ago that some viruses could induce tumors in animals, and in the 1960s and 1970s, research in this field surged. One of the pioneers in the field was Renalto Dulbecco, who, in 1968, reported that a tumor-causing virus could insert its DNA into the genome of the cell it transformed.

Understanding the molecular and cellular events that occur in cancer has long been a goal of biological research. Scientists have sought to understand how a normal cell becomes transformed into a tumor cell. Although they recognized early in the last century that some viruses, such as Rous sarcoma virus (RSV), could induce tumors in animals, it wasnt until many years later that they investigated the viral causes of cancer in depth. Such investigation was aided by the development of techniques to study viruses in cultured cells rather than in living organisms. In the late 1950s, Dulbecco adapted a number of techniques used to manipulate bacteriophagesviruses that infect prokaryotic cellsfor use in studying animal viruses in cultured eukaryotic cells. He then turned his attention to studying cell transformation by viruses in cultured cells, using the DNA tumor viruses simian virus 40 (SV40) and polyoma virus. Both SV40 and polyoma viruses could infect a number of different types of cells in culture. Cells that are susceptible to infection by these viruses fall into two classes: permissive cells, which produce more virus after infection, and nonpermissive cells, which

do not. Whereas nonpermissive cell lines did not produce new virus after infection, these were the only cells that the tumor viruses could transform. Something was happening in the nonpermissive cell lines that caused them to be transformed, rather than produce new virus. Dulbecco noted that a similar phenomenon had been observed in bacteriophage infections. Some phages cause the rapid production of new phages in a lytic infection (see Figure 4-46 in the text). Other phages through the process of lysogenylay dormant in the infected cell, while its DNA becomes integrated into the bacterial genome (Chapter 4). Dulbecco wondered if viral infection in nonpermissive cells might be similar to lysogenous phage infection in bacteria. He and others demonstrated that the viral DNA could be detected in cells transformed by polyoma or SV40 viruses. Could this viral DNA be integrated into the cellular DNA? In the late 1960s, Dulbecco set out to find the answer.

The Experiment
To determine the state of SV40 DNA in transformed cells, Dulbecco used nucleic acid hybridization, which is a powerful technique that can detect a relatively small DNA sequence within a larger sample of DNA. In his experiment, Dulbecco isolated viral DNA from purified SV40 particles. He then used purified RNA polymerase to transcribe the viral DNA into RNA in vitro. He labeled the RNA radioactively by including [3H]cytosine triphosphate (CTP) in the in vitro transcription reaction, creating a radioactively labeled probe. Next, he prepared genomic DNA from SV40-transformed cells, heated it to separate the double-stranded

DNA into two single-stranded molecules, and then bound single-stranded DNA to nitrocellulose filters. He then treated the single-stranded DNA bound to the filters with the radioactively labeled RNA probe. Through basepairing interactions, the radioactively labeled RNA specifically hybridized to the SV40 DNA. After washing the filters to remove unhybridized RNA, he analyzed the filters by scintillation counting. Under optimal conditions, radioactivity on the membranes correlates with the presence of SV40 DNA in the sample. Indeed, Dulbecco successfully employed this technique to detect SV40 DNA in the transformed cells. Once he determined that the viral DNA was present in the transformed cells, Dulbecco proceeded to determine the form and the location of SV40 DNA in a transformed cell. To do this, he took advantage of physical differences between the SV40 viral DNA and the genomic DNA isolated from the infected 3T3 cells (SV3T3). SV40 DNA isolated from viruses is in a circular supercoiled form (see Figure 4-8 in the text), which can be separated from SV3T3 genomic DNA by equilibrium density centrifugation through a gradient of cesium chloride (CsCl) in the presence of ethidium bromide. Ethidium bromide intercalates into linear DNA more readily than does circular DNA, changing its density and thus allowing the separation of supercoiled viral DNA from linear genomic DNA. As a control, Dulbecco performed the same analysis on polyoma virustransformed 3T3 cells (Py3T3) (like SV40, polyoma DNA isolated from viruses is circular and supercoiled). He immobilized DNA isolated by these procedures onto nitrocellulose filters and hybridized the radioactively labeled

Demonstration that SV40 DNA Is Integrated in Transformed Cells

Cells SV3T3 Py3T3 Py3T3 SV40 viral DNA Circular DNA (CPM) 69 5 64 7 385 11 Linear DNA (CPM) 358 2 129 12 198 5


Cells SV3T3 Py3T3 CPM hybridized 620 12 248 2

CPM Counts per minute. [Data adapted from Sambrook et al., Proc. Natl. Acad. Sci USA, 1968, 59:1290 and 1294.]

SV40 RNA. He found that the SV40 RNA hybridized only to the linear fraction of DNA from the SV3T3 cells (see Table, A). The level of hybridization to the supercoiled DNA was no greater in the SV3T3 cells than it was in the Py3T3 cells, indicating that any radioactivity detected in this fraction was at background levels. As an additional control, Dulbecco added supercoiled DNA isolated from SV40 to the Py3T3 cell extract. In these cells, he could detect the DNA specifically in the supercoiled DNA fraction of the DNA isolated from these Py3T3 cells, and not in the linear DNA fraction (see Table, A). From these experiments, he concluded that the SV40 DNA was not in its supercoiled, viral form in transformed cells. To assure that the SV40 DNA in the linear fraction was part of the SV3T3 genomic DNA, and not linearized SV40 viral DNA, Dulbecco performed alkaline sucrose gradients. He layered the cells onto the alkaline gradient, then incubated them to allow the cells to break open. This

procedure minimized inadvertent mechanical shearing of cellular DNA. Once lysis was complete, the cellular DNA was sedimented in the sucrose gradient. Dulbecco then compared the hybridization of SV40 RNA to DNA isolated from SV3T3 and Py3T3 and found that it specifically hybridized to the SV3T3 cells (see Table, B). Because the SV40 DNA always hybridized to the high-molecularweight genomic DNA, he concluded that it was covalently attached to the cellular DNA.

Dulbeccos demonstration that the SV40 DNA was integrated into the genome of transformed cells began a new wave of thinking about cellular transformation. With the advent of restriction endonucleases, as well as other molecular-biology techniques, scientists demonstrated that SV40 integrates into random sites in the host cell genome. Researchers have found that a number of other tumor-inducing DNA

viruses, as well as RNA-based retroviruses, integrate into the host cell genome. They later showed that viral integration can contribute to tumor formation by disregulating the expression of key cellular genes involved in cell growth, or by allowing expression of viral proteins that derail normal cellular mechanisms to inappropriately stimulate cell growth and proliferation. Although tumor viruses are more prevalent in mice and other animals, several human tumor viruses, such as human T-cell leukemia/lymphoma virus (HTLV) and human papilloma virus (HPV), are known to integrate into the host cell genome and can lead to cancer. By inspecting the sites of viral insertion, scientists have discovered a variety of oncogenes in work predicated on Dulbeccos initial studies on DNA tumor viruses and their ability to integrate into the host genome. In 1975, the Nobel Prize in Physiology or Medicine was awarded to Dulbecco for his vast contributions to this field.