Dna: The Secret of Life

by Watson, James

  Transcriptomics is more than just another brilliant technical innovation. It promises to take us to a new level in the hunt for the genes that cause illness: using microarray technology we can discover the chemical basis for particular afflictions by studying the differences between healthy and diseased tissue as a function of gene expression. The logic is simple. We carry out microarray gene expression analysis on both normal and cancerous tissue, and spot the difference between the two, the genes being expressed in one and not the other. Once we can identify which genes are malfunctioning – either over- or under-expressing themselves in the cancerous tissue, for instance – we may be able to establish a target that can be attacked with pinpoint molecular therapies as opposed to broadly toxic radio- and chemotherapies that destroy healthy as well as diseased cells.

  And we can apply the same technologies to distinguish among different forms of the same disease. Standard microscopy has offered limited assistance in this task: cancers that look alike to the pathologist peering through the eyepiece can in fact be critically different at the molecular level. Lymphoma cells, for instance, come in varieties that are hard to tell apart visually, even with the highest powers of resolution, but the differences in their gene expression profiles are clear, and vitally important in devising the most effective treatment. Referring to the earlier tendency to assume all cancers of a particular tissue have the same root, Brown said, "It was like thinking a stomachache has only one cause. Recognizing the distinctions makes it possible for us to do a better job of treating these cancers."

  At Cold Spring Harbor Laboratory, Michael Wigler is using the method in yet another way: rather than adding RNA to a microarray and looking for gene expression, he is adding DNA from cancer cells to create a profile of the genetic diversity present in tumors. Many cancers are caused by chromosomal rearrangements – such as might occur when segments of a chromosome are inadvertently duplicated, leading to an excess in the number of genes that code for growth-promoting proteins. Other cancers arise due to the loss of genes coding for proteins that repress cell growth. Applying Wigler's technique, clinicians biopsy cancerous and healthy tissues from the same person. DNA from the cancerous tissue is chemically tagged with a red dye while the DNA from the normal tissue is tagged green. DNA microarrays, containing all 35,000 of the known human genes, are exposed to a mixture of the two samples. Like mRNA in a standard microarray experiment, the labeled DNA molecules bind base pair to base pair to their complementary sequences in the array. Genes amplified in cancer cells are marked by red spots (because there are many more red-tagged molecules binding to that spot than green-tagged ones) while genes deleted in cancer cells show up as green spots on the microarray (because there is no red-tagged molecule to bind there). Such experiments have already greatly expanded the list of genes known to contribute to breast cancer.

  Whenever we tackle a specific human disease, we realize the extent to which we are probing in the dark. We could move so much more quickly to the heart of the problem – know the exact nature of what is wrong and how we might fix it – if only we had a more detailed knowledge of how our genes express themselves when all is well. With a fully formed dynamic understanding of when and where each of our 25,000 genes functions during normal development from fertilized egg to functioning adult, we would have a basis of comparison by which to understand every affliction: what we need is the complete human "transcriptome." This is the next holy grail of genetics, the next big quest in need of superfunding. In the short term, a likelier, even more important objective will be to obtain the complete transcriptome for the mouse, whose advantage over humans is that we can both observe and intervene experimentally during the course of prenatal development. Even collecting all such relevant data from the mouse will require major investments of money and time. And, as proved by the experience of DNA sequencing, we will be well served to take the time to gain what expertise we can by completing transcriptomes for simpler model organisms before taking on the mouse, much less the human.

  Microarray studies of gene expression during the yeast cell cycle have already revealed the staggering complexity inherent in the molecular dynamics of cell division alone. More than eight hundred genes are involved, each called into action at its precisely specified time in the cell cycle. Here too we may depend on evolution's reluctance to fix what ain't broke: a biological process, once successfully evolved, will likely continue to employ the same basic molecular actors for as long as life persists on earth. As far as we can tell, those same proteins that direct development through the course of the yeast cell cycle carry out similar roles in the human cell.

  Ultimately the goal of all three "-omics" (gen-, prote-, and transcript-) is to create a full picture, detailed right down to the level of the individual molecule, of how living things are assembled and operate. As we have seen, in even the simplest cases, the complexity is bewildering, and, despite the spectacular progress of the last decade, there remain many daunting challenges. As they relate to complex organisms, the molecular underpinnings of development – that extraordinary egg-to-adult journey that is governed by a linear code strand composed of just four letters – are for now best understood in the case of the fruit fly.

  The fly has, of course, been the focus of intensive genetic investigation ever since its adoption by T. H. Morgan, and through the ensuing years of continual innovation Drosophila melanogaster has remained a genetic gold mine. In the late seventies at the European Molecular Biology Laboratory in Heidelberg, Germany, Christiane "Janni" Nüsslein-Volhard and Eric Wieschaus undertook a spectacularly ambitious fruit fly project. They used chemicals to induce mutations and then looked for disruptions in the very early embryonic stages of the flies' progeny. Classically, the quarry of the fruit fly geneticist was mutations affecting adults, like the one Morgan found to produce white (rather than red) eyes. In focusing on embryos, Nüsslein-Volhard and Wieschaus were not only condemning themselves to years of eyestrain as they stared down microscopes in pursuit of those elusive mutants, they were also venturing into utterly uncharted territory. The payoff, however, was spectacular. Their analysis uncovered several suites of genes that lay out the fundamental body plan of the developing fly larva.

  The more universal message of their work is that genetic information is hierarchically organized. Nüsslein-Volhard and Wieschaus noticed that some of their mutants showed very broad effects while others evinced more restricted ones; from this they inferred correctly that the broad-effect genes operate early in development – at the top of a switching hierarchy – while the restricted-effect genes operated later. What they had found was a cascade of transcription factors: genes switching on other genes that in turn switch on others still, and so on. Indeed, hierarchical gene-switching of this kind is the key to the construction of complex bodies. A gene producing the biological equivalent of a brick will, left to its own devices, produce a pile of bricks; with proper coordination, however, it can produce a wall, and ultimately a building.

  Normal development depends on cells "knowing" where they are in a body. A cell in the tip of a fly's wing, after all, should develop along very different lines than one located in the region that will give rise to the fly's brain. The first piece of essential positional information is the simplest: How does the developing fruit fly embryo know which end is which? Where should the head go? Bicoid, a protein produced by a gene in the mother, is distributed in varying concentrations through the embryo. The effect is called a "concentration gradient": the protein levels are highest at the head end and fall off as you travel toward the rear. Thus the bicoid concentration gradient instructs all cells within the embryo as to where they fall on the head-to-tail axis. Fruit fly development is segmental, meaning that the body is organized into compartments, all of which have much in common but each of which has some features unique to it. In many respects, a head segment is organized just like one in the thorax (the middle part of the insect body), but the former has head-specific organs, like eyes, and the latter t
horax-specific ones, like legs. Nüsslein-Volhard and Wieschaus found groups of genes that specify the identities of different segments. For instance, "pair-rule" genes encode transcription factors – genetic switches – expressed in alternating segments. Pair-rule gene mutants result in an embryo with developmental problems in every second segment.

  In 1995, Nüsslein-Volhard and Wieschaus received the Nobel Prize in Physiology or Medicine for their pioneering work. Unlike most laureates, both have remained active lab scientists – not for them the retreat into a big diploma-festooned office. For Wieschaus, science is still irresistible: "Because embryos are beautiful and because cells do remarkable things, I still go into the lab every day with great enthusiasm." As a child in Birmingham, Alabama, he dreamed of becoming an artist. Short of money as a sophomore at the University of Notre Dame, however, he took on one of the smelliest and most menial jobs in all of science: making the "fly food" (a noxious gelatinous concoction consisting largely of molasses) for a research lab's experimental population of fruit flies. Most people who serve as chef to a few hundred thousand messy and unappreciative insects would likely develop a lifelong aversion to the critters. For Wieschaus the result was the opposite: a lifelong commitment to the fruit fly and the mysteries of its development.

  Born into an artistic German family, Nüsslein-Volhard was one of those students who excels at everything that interests them but puts absolutely no effort into anything else. Her hard work in illuminating the fruit fly's developmental genetics would have been achievement enough to justify two careers, but in the wake of her Nobel she has redirected her formidable attention to the development of another species altogether, the zebra fish: new work that promises to unlock many of the secrets of vertebrate development. At the 2001 event marking the centenary of the Nobel Prize it struck me that she was the only woman scientist present among the throngs of gray-haired males. Indeed, she is one of only ten women ever to win a Nobel in science.

  One of those no-longer-youthful men was Caltech's Ed Lewis, an old fruit fly hand who shared the prize with Nüsslein-Volhard and Wieschaus. Actually Lewis doesn't much fit the gray-hair stereotype: though in his eighties at that Stockholm event, when he wasn't obliged to wear tails he was often seen in running gear! He too had long been concerned with the genetic control of fruit fly development, but his special interest was "homeotic mutations." These produce a most bizarre result: one developing segment mistakenly acquires the identity of a neighboring segment. His long and painstaking dedication to the Hox genes, in which these mutations occur, exemplifies values vanishing in an era when fads too often set science's agenda.

  Homeotic mutations – which we now know disrupt transcription factor-encoding genes (the genetic switches) – can have drastic effects. The "antennapedia" mutation results in the fly's growing legs where its antennae belong: a fully formed pair of legs protruding from its forehead (see Plate 46). The "bithorax" mutation is almost as weird. Normally one of the segments making up the thorax produces the fly's pair of wings while the next thoracic segment toward the rear generates a pair of small stabilizing structures called "halteres." In a bithorax fly, the haltere segment mistakenly produces wings, so a fly that should have two wings in fact has four, the second pair just as perfectly formed as the first.

  When they function properly, the genes regulating segment identity ensure that each body section acquires organs appropriate to its position: a head segment acquires antennae, and a thoracic segment acquires wings and legs. In the event of homeotic mutations, however, there is a confusion of segment identity. Thus, in the case of antennapedia, a head segment imagines itself a thoracic one and duly produces a leg rather than an antenna. Note, though, that while the leg is in the wrong place, it's still a perfectly good leg. Implication: The antennapedia positional gene switches on a whole suite of genes, typically those that produce an antenna, or, aberrantly, those that produce a leg; but the coordination within the suite is unhindered even when these genes are activated in the wrong place at the wrong time. Here again we see how genes high up in the developmental hierarchy control the fate of many, many genes farther down the line. As any librarian knows, hierarchical organization is an efficient way in which to store and retrieve information. With such a cascade arrangement, a surprisingly few genes can take you a long way.

  Now that we are in the new era of comprehensiveness in biology ushered in by the once-unimaginable feat of the Human Genome Project, it may seem curious that we should find ourselves following the cutting edge of one of the next frontiers – that of developmental genetics – back into the realm of the fruit fly. But there is nowhere for us to go but back to the future, for even with the entire human genome in hand, the program and cues according to which its instructions are carried out remain a colossal mystery. Eventually we shall know the screenplay of human life as well as we know that of the fly. A comprehensive description of the patterns of human gene expression (the transcriptome) will be developed. A full inventory of the actions of all our proteins (the proteome) will be produced. And we will have a full and spectacularly complex picture of how each one of us is put together, and how each one of the multitudinous molecules we are made of figures in the functioning of you and me.

  CHAPTER NINE

  OUT OF AFRICA:

  DNA AND THE HUMAN PAST

  In August 1856 German quarry workers discovered part of a skeleton as they blasted their way into a limestone cave in the Neander Valley outside Dusseldorf. At first the remains appeared to be those of an extinct bear species whose bones often showed up in caves, but a local schoolteacher realized that the creature in fact belonged to a species much closer to our own. The exact identity of the owner of the bones, however, would prove a point of controversy. Particularly puzzling was the skull's thick brow ridge. One bizarre suggestion was that the bones belonged to an injured Cossack cavalryman who had crawled into the cave to die during the Napoleonic wars. Chronic pain from a preexisting condition, so the crackpot theory went, had produced a permanent furrow in the poor fellow's brow, deforming the bones of the skull to create the distinctive ridge. In 1863, in the midst of the debate about human origins provoked by the publication of Darwin's Origin of Species four years earlier, the original owner of the bones was given a name: Homo neanderthalensis. The bones belonged to a species distinct from, but similar to, Homo sapiens.

  Though the German bones were the first to be officially designated Neanderthal, others found earlier in Belgium and Gibraltar were now recognized as being from members of the same species. More than a century later, many more specimens of H. neanderthalensis have been unearthed, and we now believe that Neanderthals settled throughout Europe, the Middle East, and parts of North Africa until about 30,000 years ago. French paleontologist Marcellin Boule is largely responsible for the popular image of Neanderthals as dim-witted and hulking. But his reconstruction, which used material from a French site at La Chapelle-aux-Saints, was based on a single individual who turns out to have been elderly and arthritic. In fact, Neanderthal brains were slightly larger than ours (and of a different shape due to a flatter cranium) and evidence from burial sites suggests that Neanderthals were culturally sophisticated enough to engage in funeral rituals; they may then have even believed in an afterlife.

  The biggest debate triggered by the discovery of Neanderthals, however, centered not on how smart they were but on how they might be related to us. Are we descended from them? Paleontology suggests that modern humans arrived in Europe at roughly the same time as the last of the Neanderthals disappeared. Did the two groups interbreed or were the Neanderthals simply eliminated? Because the events in question happened in the ancient past and the surviving evidence is fragmentary – little beyond the odd bone – debates like this can drag on and on, keeping academic paleontologists and anthropologists endlessly entertained. Is a particular bone specimen perhaps intermediate between the thick bones typical of Neanderthals and the lighter bones of modern humans? Such specimens may have belonged to a hy
brid individual produced by interbreeding between the two groups – a missing link. But then again they might just as well have come from a full Neanderthal with atypically light bones, or, for that matter, a fully modern human with unusually thick ones.

  To everyone's surprise, the debate has been resolved by DNA: 30,000-year-old DNA extracted in 1997 from the very bones that started it all in 1856. Having evolved in order to store information securely and transmit it from one generation to the next, DNA, no surprise, shows great chemical stability. It doesn't degrade spontaneously or react readily with other molecules. But it is not impervious to chemical damage. At the moment of death, the body's genetic materials, like all its other constituents, become susceptible to a horde of would-be degraders: reactive chemicals, and enzymes that break down the molecular fabric. These chemical reactions require the presence of water, so DNA may be preserved if a corpse dehydrates fast enough. But even under ideal preservation conditions, the molecule is likely to survive perhaps 50,000 years at the absolute maximum. To obtain a legible DNA sequence from 30,000-year-old Neanderthal remains, preserved imperfectly, was therefore a tall order at best.

  But Svante Pääbo, a tall, laconic Swede at the University of Munich, decided to have a crack at the problem. If anyone could do it, he was the one. Pääbo had pioneered work on the retrieval of so-called ancient DNA; he had scored sequences from Egyptian mummies, frozen mammoths, and the 5,000-year-old "Ice Man" who melted out of an Alpine glacier in 1991. Despite this impressive resume, though, the prospect of drilling into a precious Neanderthal relic to look for intact DNA, if indeed any was to be found inside, was daunting. As his archaeologist colleague Ralf Schmitz recalls, "It was like getting permission to cut into the Mona Lisa."