Creation (15 page)

In genetics, making your transgene, the code with which you are messing, is only the first step. The code does nothing on its own, so it has to be placed into a cell where decoding will occur, and you hope the subverted message will be enacted. In Freckles, the code is not merely the spider-silk gene, but instructions that say when it will be active, or “expressed” in the language of genetics. By careful construction, the silk will only be produced in the milk, even though the gene is present in every one of Freckles' cells.

During the 1990s, extra bits of genetic code added to the toolshed included colored tags so that we could see while peering through a microscope whether the gene had been successfully taken up into its new carrier. More recently, fluorescent tags have been added so that we can see where the gene is active once it has been integrated into an animal. The most visually striking example of this is the functionally named green fluorescent protein (GFP), which in its natural host, the jellyfish
Aequorea
, glows green in the pitch-dark oceans. The gene that encodes this beacon was extracted from jellyfish in order to be added on to the gene you are testing. Just as in the spider-goats, the machinery that translates the DNA into working proteins is indifferent to its meaning, so reads the universal genetic code without species prejudice, and glows to show that your modified gene is working.
³

Genes never work in solitude. Instead, they are parts in cascades of networked activity, like a series of conditional instructions. Every cell contains every gene in that organism's genome, regardless of whether it is needed. The choreography of gene activation is therefore of paramount importance. The activation of one gene might trigger the activation or expression of another, or tell it to stop, and by this process we develop from a single fertilized egg cell to a collection of hundreds of coordinated cells, each performing different functions as a result of the expression of specific sets of genes, rather than an amorphous blob comprising identical cells. This dance is elegantly choreographed, but with limited flexibility. A gene activated at the wrong time, or in the wrong place, or for too long can lead to disease, abnormality, or often fatality. Cancer cells are ones in which genes that would normally tell other genes to stop cell division have failed (or conversely, genes that tell the cell to reproduce are permanently active), so the cell only divides. This uncontrolled reproduction grows into a tumor.

The command to activate a specific gene will typically come in the form of a protein physically attaching itself not directly to the gene itself but to the DNA nearby. These instructions are called regulatory regions, like instructions for flat-pack furniture. One gene will produce a protein that will bind to a particular run of DNA bases near another gene telling it to become active, and a continuous network of activity buzzes inside each cell and between cells to create a life. A large amount of most genomes is not genes—in other words, DNA that encodes a protein. Genomes are littered with these regulatory regions, which don't contain code themselves but act instead as very precise stage directions for the proteins to enact their role.

Through the techniques of DNA modification, turning these networks of instructions into tools has rendered molecular biology an experimental science. As is so often the case in biology, we can determine how things work by breaking them. Traditionally, this has taken the form of studying hereditary diseases or by looking at mutated animals. As mice and men share many genes, we can knock out a gene in a mouse embryo by modifying its DNA to see its result. This is done using restriction enzymes, imported regulatory bits of DNA, and spliced-in transgenes, together forming a process of major reconstructive DNA editing.

Extraordinary though they may sound, these techniques have been the meat and potatoes of genetics for more than a decade, and the industrialization of the process of gene splicing and editing has made genetic experimentation easier. We have taken the tools provided by evolution and modified, designed, redesigned, and honed them to provide service beyond the roles they play as a result of natural selection. A typical experiment to determine the function of a human disease-causing gene would be to isolate it in a family that suffers from that disease, using the same principles of genetics and pedigrees we have known about for a century. Once purified, it gets copied a million times, so there is more DNA to play with. That makes inserting the modified gene into a precise orientation in a DNA envelope much easier, and it can then be posted into a bacteria. A tiny burst of electricity will punch temporary holes in the bacteria's cell membrane; the constructed DNA will just flow in, and with skillful design and a bit of luck it will be incorporated into the host's genome. Every time the cell then divides, the inserted gene will be copied along with the host genome. This is a numbers game, so bacteria are the perfect tool not just because we can manipulate them, but because they grow like crazy. The next step is to destroy the bacteria and leave only your modified DNA. At that point, you can do what you want with it, such as make RNA versions of it, which will show where the gene is active in preserved tissue on a slide, in an organ, or even in a whole animal. Or you can put that gene and all its extra controls into the embryonic stem cell of a mouse, and implant that into a mother to see what it does as the mouse develops.
⁴

None of that is exceptional, just standard molecular biology going on in thousands of labs all over the world. We have become so adept at using the tools and the language at the core of all living things that manipulating living systems at their most fundamental level is trivially easy. Through experimentation we know how genes in the cells work and, more important, how we can change them. We can correct genes that cause diseases, not yet in humans but in animals, which allows us to study how diseases progress, and therefore how we can treat them. We have developed the ability to read and characterize every single gene in a human, at the time of this writing, in a few weeks at a cost of a few thousand dollars (though both numbers will inevitably continue to fall).

Synthetic biology is the evolved descendant of genetic engineering. It takes the principles of biology and reinvents them with the goal of engineering solutions to specific human problems—disease, environmental issues, and, as we shall see, even space exploration. This synthetic biology movement is a new phenomenon in science, a decade old at the most generous estimate. It has carried with it a subculture ethos, too. But, unsurprisingly, mainstream science and corporations have also begun to take notice of its potential.

Climate change and global warming will define much of synthetic biology's industry and innovation in the next few decades. Expecting people to radically change their behavior is unrealistic, so finding alternatives to fossil fuels is a major challenge. In the meantime, synthesizing fuels, rather than digging them out of the ground, is a partial solution.

There are dozens of projects to make biofuels, converting vegetation into energy in that fundamental way that so many life-forms do naturally.
⁵
It is a process of carbon fixation, effectively turning carbon dioxide into organic, energy-rich products. In nature the energy of the sun is harnessed in various metabolic processes by a plant so that it can live and grow. Setting fire to a crop will release that energy, now stored in the plants' strong cellulose cell walls. But a more efficient way of tapping into that energy is to ferment the sugars locked in plant cells, also the product of the sun's energy, directly into an oil fuel, which is crammed full of more readily available energy. A handful of companies are attempting to make biofuel from synthetic biology using this very process.

One such project is run by Jay Keasling, a professor at the University of California, Berkeley, who founded Amyris, a synthetic biology company whose aim is to create diesel from live cells. Their designed tool has been a genetic circuit consisting of around a dozen individual chunks of DNA, which they implant into the genome of brewer's yeast, a cell that naturally ferments sugar into alcohol in beer. The success of this circuit in producing diesel is no mean feat, and the scale of their ambition is mesmerizing.

I have been in scores of genetics labs, which range from the charmingly old-fashioned to the ultramodern. Molecular biology doesn't necessarily need the huge clinical white spaces that the movies portray. They are essentially very precise kitchens, with ovens, fridges, and tools for mixing ingredients. But Amyris's labs are stunning. Only the very deep pockets of a commercial venture can provide such luxury, not merely in its space and light, and a healthy San Franciscan café, but in developing the kind of technology that is required in order to industrialize synthetic biology.

One of the most time-consuming tasks in a normal lab is the insertion of the genetic circuit into the host cells. The efficiency of this process is variable. It's not random, but there is an element of hit or miss: the circuit either goes in or it doesn't. In normal lab gene tinkering, we include genetic markers in the circuits, tags that color the cells to indicate whether integration of the alien DNA has been successful. In order to see those colors, the cells have to be grown not in a broth but on a plate, so that the successful colonies of cells can be selected with a toothpick and grown in a small tube, leaving the failures for the trash. It's a trivial but arduous process and it is done by hand. At Amyris, they have mechanized it so that they can pick hundreds of successful clones in minutes, tens of thousands per week, with only minimal human input. A digital camera takes a snapshot of the plate that bears the thousands of yeast colonies and processes it to identify the successful clones. Inside a machine as big as an office photocopier, a bank of needles hovers and zooms over the plate, plucking the successful yeast cells with submillimeter accuracy at the rate of a disco beat.

The cells that have taken up the circuit don't need much more encouragement. Incubated, they simply leak diesel. At Amyris's headquarters in San Francisco they have a three-hundred-liter test tank churning out diesel. The labs smell sickly sweet, like a brewery, but more ciderlike, because the diesel they are refining, farnesene, is found in the oil that gives apples their water-repellent skin. As a fuel, it is cleaner than gasoline-based diesel, as it has no sulfur emissions and reduced nitrous oxides and carbon monoxide.

Creating diesel affords us certain advantages over digging it out of the ground. But it also comes with a different problem. Set aside the fact that clean, synthetic diesel is still diesel, so is still a carbon dioxide–generating fuel. The more significant problem is that the vats of yeast seeping out diesel need to be fed. That means converting plant material, biomass, into food. To grow that biomass is a farming problem, and so comes with the same issues as farming. Amyris has made strong partnerships with companies in Brazil in order to be close to where the first-choice food, sugar cane, is grown. Brazil has the biggest sugar-cane farmers, and since the 1970s Brazil has produced biofuels, predominantly ethanol, as a means of freeing itself from dependence on foreign oil. The key question is this: how much land does it take to make one liter of diesel? The answer is unclear at this point. Different feedstock is digested in different ways, and the synthetic yeast programs can be altered to cope with this. In their Brazilian plants, synthetic farnesene diesel is already being supplied to local vehicles and as an alternative aviation fuel. But Amyris's goal was to grow output to two hundred million liters by 2011 at $2 a gallon. The ambition is clear: Jack Newman, Amyris cofounder and chief scientist, told me, “I'll be excited by a billion liters.”

Yet it looks like Newman's billion liters is, for now, a burst bubble. For a while, Amyris looked as if they were going to be the winners in the race to make and sell commercially viable synthetic biofuels. But then the two hundred million liter prediction dropped to fifty million for 2012. In February 2012 the company announced it was scaling back production of farnesene until the industrialization becomes more plausible. The genetic program works just fine, but thus far, scaling it up to the levels where it becomes economically viable does not. For now, for all the stunning laboratories, it seems that the synthetic biofueled future is still a long way down the road.

Meanwhile, we have been able to modify crops to be resistant to blights, grow larger, tolerate frost, and even produce vitamins that protect consumers from diseases. By borrowing genes from a bacteria and a daffodil, scientists have created a form of rice that produces high levels of beta-carotene, a molecule involved in Vitamin A production. Golden rice, as it is called, has the potential to treat the 120 million people worldwide who suffer from Vitamin A deficiency, 2 million of whom die and 500,000 of whom go blind. Yet it remains unavailable due to a combination of scientific and ethical roadblocks.

These stories show some of the promise, potential, and problems with synthetic biology. In essence it is engineering, with the application of basic research at its core. Yet alongside its forefather, genetic engineering, these are immature industries. They face scientific problems, commercialization problems, upscaling problems, ethical problems, and, as we shall see later, dogged resistance from some members of society. Biology is messy, and complex genetic networks underlie that mess. As genetic engineering, just three decades old, gives way to synthetic biology, the great challenge is not merely to simplify those networks, but to commoditize them.

Logic in Life

“If it was so, it might be; and if it were so, it would be; but as it isn't, it ain't. That's logic.”

Tweedledee

F
lick the switch, the light comes on. This is the simplest useful electrical circuit. The parts are designed and created to follow one single instruction: the switch is a gap when open, but when closed, electrical energy surges through the circuit. The filament in the bulb converts some of the electrical energy into a form that we can detect with the cells in our eyes, and so enlightenment occurs. The function is clear, the instruction pure, the logic is unimpeachable, the light comes on.

At the other end of the scale, you watch a video streamed over the Internet on a laptop. Billions of electrical signals will have been created, modified, and transmitted in order for that moving image to play. The circuits have been designed in intricate detail with each one obeying a logical pattern determined by the hardware and software in your mouse, your computer, the servers that host the file, and so on. The logic is also perfectly clear, but the complexity of the pathway makes it all but inscrutable to almost everyone, and occasionally unpredictable. And yet we use the output of this tangled circuit every day without care or understanding of the millions of decisions that have been made to put a moving image on your screen.

Recall, if you can bear to, some of the electrical circuitry you learned at school. A lightbulb connected to a battery with a simple switch has a binary output:
ON
or
OFF
. You might have then added other devices into the circuit to introduce finer levels of control, such as diodes—electrical valves that turn a wire into a one-way street. If you added, for example, a thyristor, you'd have a dimmer switch. Possibly the greatest invention, certainly the most enabling technology, of the twentieth century was the transistor, which allowed the ability to control and modify multiple electrical signals. Interconnected transistors make up what are known as logic gates, which modify the input signal to produce a specific but different output. With the introduction of logic gates, circuits of ever-increasing complexity can be designed and built. For example, an
AND
gate acts as a positive conjunction: if two electrical inputs feed into an
AND
gate, both must be
ON
in order for the output to also be
ON
(in electronic engineering, capital letters indicate logical calculations, rather than merely bellowing for emphasis). A microwave oven uses this logic. It will only cook if the door is closed and the start button has been pressed. If either of the two crucial signals is negative, then the output is negative.

Electrical circuitry relies on logic, on following pathways determined by the component parts. From the light switch all the way up to the machine I am typing on, the route of information takes the form of digital questions with digital answers: if I press the on switch, the light comes on; if I press the return key (via a long pathway of transistors and thousands of other components), the paragraph ends.

These are the very basics of electrical engineering, a field that has grown from lightbulbs to relaying messages to and from deep space in a little over a hundred years. Every signal sent to your mobile phone, or from the spacecraft
Voyager 1,
¹
is relayed via the logic ingrained in transistors. In a standard microchip, there are billions of transistors, and this is the technology on which almost everyone now relies.

It's a hugely appealing system, not just because it works as we design it to, but because the fact that it works has allowed us to build the modern world. This logic and this ambition are at the heart of synthetic biology. Many components have been built and some have been assembled into basic circuitry. Although the term
synthetic biology
has gained broad usage to include everything from the work of Craig Venter on Synthia to the reinvention of genetic code described later, the field originated and has been maintained by scientists aiming to apply the principles of engineering, specifically electrical engineering, to biology. Living things are immensely complex, with thousands of genes encoding thousands more proteins that interact with one another and the environment to produce millions of cells. But the basic logic of genetics is present in principle: if a gene is activated, the protein it encodes will be activated, and will perform a function.

We think of life-forms not as illogical, yet not with the straightforward formulas of electronics. Nevertheless, the principles of logic gates as they are used in electronics do exist in nature, as complex behaviors and actions often need the parsing of multiple inputs. The carnivorous Venus flytrap plant displays a simple but clever form of logic circuitry when performing its eponymous act. On the inside of the leafy jaws are tiny hairs that act as the triggers for clamping shut. But that act requires metabolic energy, so it has evolved a mechanism for avoiding fruitless (or flyless) trapping. When a hair is touched by a fly's nudge, a timer begins. If a second hair is triggered within twenty seconds, the jaws snap shut in less than one-tenth of a second. Thus, the input signals that result in the action of trapping are required in conjunction. The double trigger is a form of an
AND
gate, as both outputs from the hair triggers need to be set to
ON
for the output of the full circuit to be
ON.
There's also the timer built into each trigger, so the overall simplified electrical pathway might read:
IF TRIGGER 1
+
TRIGGER
2 < 20”,
THEN CLOSE JAWS
.

This process is, naturally, managed and run by the mechanics inside the Venus flytrap's cells, with specific proteins that respond to the physical sensation of a fly's touch. But the biological action of a trap waiting to be sprung follows a straightforward logic that looks a lot like electronics: it has no brain or conscious control for the decisions it makes, it just follows a program.
²

Mostly, however, living things follow a logic that is a far more difficult lock to pick. The activation of genes to produce functional proteins is bound by time and location. Genes interact and respond to the environment in which they are active, in the cell, and from long- and short-range signals from close and distant neighbors. Genes themselves vary from person to person and from creature to creature in small and subtle ways. This can make their function, their output, also vary from person to person. This variation is essential, as it is the basis on which natural selection can act so that evolution unfolds. Combine that with the muddled interplay of genes and environment and we have absolute individuality. This is the reason why fingerprints are unique, even on the fingers of identical twins, whose genes begin their existence identical.

There are some cells that work in a basically digital way. Brain cells—neurons—fire with a dynamic flourish and link up to produce thought and sense. Their spark is only initiated when inputs from other cells (in the form of charged atoms) flow into the neuron until a particular threshold is reached. This process, with a suitably dramatic name—action potential—renders these neurons digital: they are
OFF
until they are
ON
. But don't let that simplicity fool you into thinking your brain is easily understandable. There are more than a hundred billion cells in your brain, and each one of them forms thousands of connections with others. This means potentially hundreds of trillions of switches flicking on and off in millisecond bursts every moment you are alive. There may be a logic buried in there, but for the foreseeable future it is inscrutably complex.

For the most part, the switches in life are not quite so binary, either. There are hundreds of other types of cells and switches that are much more nuanced, stimulated on a sliding gauge, or contingent on myriad inputs. Just like a neuron, a gene may indeed be
ON
or
OFF
like a lightbulb. But it also may be
ON
at different strengths, or it may have different functions depending on where or when it is active. This fact goes part of the way toward explaining why we can't understand our own complexity concerning the 22,000 genes present in our genomes (far fewer than had been assumed necessary to describe how humans work). It is also why mutations in the same gene can cause independent diseases in the eye and the kidney. How and when a particular gene is controlled can mark profoundly different functions in entirely different tissues.

There are so many confounding and complicating factors in biology that the underlying logic is lost in the messiness of real cellular life: unpredictable variation set among unfathomable sophistication. A genome, with its full set of biological instructions, is akin to the score of a symphony. The full beauty is expressed in a performance, complete with interpretation and nuance that is not merely encoded in the quavers, crotchets, and minims written on the page.

We sometimes call the seemingly inscrutable complexity of a living system not music but noise. It's a combination of all of the things that we can't quite account for: things such as natural variation in the behaviors of genes, proteins, or molecular interactions that haven't yet been described. The natural progression of biology as a science to synthetic biology has meant confronting the nuances of this mess and scrutinizing it. At the heart of synthetic biology is the desire to avoid that mess by creating new life forms whose circuitry and programming is clear, simple, and, crucially, built not for survival but for purpose.

Sniper Circuits

“Imagine a program, a piece of DNA, that goes into a cell and says, ‘If cancer, then make a protein that kills the cancer cell; if not, just go away.' That's a kind of program that we're able to write and implement and test in living cells right now.”

Those are the words of Ron Weiss, one of the founders of the modern engineering discipline of synthetic biology and now a professor at the Massachusetts Institute of Technology. He is describing a landmark study his team published in autumn 2011. By using the logic and language of computer circuits combined with biological components, they had built a tool that effectively acted as a cancer assassin.

The terminator circuit is an assembly of DNA components, built to serve a singular mission: to identify and kill a type of cancer cell. Once built, the circuit slots into the genetic code of a virus, itself modified to grant us control over its natural tendencies. When introduced to malignant cells it infects them and, just like all viruses do, adds its synthetic genome (including the assassin program) to the host's DNA. Obliging the natural trickery of virus infections, that host cell unwittingly decodes the killer circuit and enacts the program that will bring about its own downfall.

The assassin circuit represents a set of five questions, each one seeking the presence of a particular molecule unique only to this type of cancer cell. If any of the answers to this molecular interrogation is negative, the program stops, shuts down, and decays, and the cell lives out its normal life. If, however, the answer to all five identification questions is yes, termination ensues. The code in the circuit contains a gene that triggers the host cell's own built-in suicide program. It is simultaneously a conscientious and persuasive assassin, one that makes forensically sure its quarry is indeed the target before quietly asking the victim to kill itself.

The genetic circuit is complex but logical. While it may seem trite to draw such a literal analogy with microcircuitry, this system has in fact specific logic calculations built into it that are entirely derived from electronics. It uses exactly the simplified processing logic of electrical components such as
AND
gates. It also uses
NOT
gates, which flip the input from
ON
to
OFF
and vice versa.

This particular malignant cell is perhaps the most well-studied laboratory cancer in the world, the HeLa cell. It is an undying ancestry of cells derived from the cervical cancer of a young black woman named Henrietta Lacks. A scrape taken from the cancer on her cervix was grown in a lab in 1951, and was soon identified as being immortal. Normal cells age, grow weary, and ultimately lose their ability to divide. Cell death follows. But HeLa cells reproduce indefinitely due to idiosyncratic faults in their own genetic logic boards. The tenacious immortality of HeLa cells has meant that they are the most studied and scrutinized malignant cells in existence, having been shared and spread among labs throughout the world. With such a well-studied fingerprint it makes them a worthy testing ground for such a heavily engineered genetic circuit. Most cells are not known in such forensic detail, so foolproof identification would not yet be possible. But for HeLa cells, the five molecular questions allow precise recognition so thorough it qualifies as what is known in electronics as a
TRUE
value. The precision of the assassin circuit is such that it is in effect a biological computer.

This system is a high-water mark in the nascent field of synthetic biology. As with all potential therapies, it has a long way to go before being available to humans. Currently, it has only been tested in cells in dishes. Animal tests will come next, where the degrees of control will be challenged further by the more chaotic and dynamic noise of a working creature.
³
As a therapy, though, its accuracy is potentially shocking. Cancer, being a multitude of diseases and prone to mutating as it spreads and grows, means that you're aiming at an ever-moving target. Chemotherapy and radiation therapy remain the most effective attack on tumors, though they are damaging both to malignant and healthy cells. Compared with the carpet bombing and collateral damage of radiation therapy, this cancer treatment is a sniper. It represents the immersion that researchers have made in a decade into the realms of programmable biological machinery. As a weapon against a disease, the function of the assassin program is clear, and its assembly is premeditated.