Nature is already an expert in splicing together her existing repertoire to generate proteins with new functions. Her unit of operation is the protein domain, an evolutionarily independent protein structure that specializes in a particular task, such as an enzymatic activity or recognition of other proteins. We can trace the evolutionary descent of the protein domains by examining their sequences and grouping them into family trees. We find that over the eons of evolutionary time the DNA that encodes protein domains has been duplicated and combined in countless ways through rare genetic events, and that such shuffling is one of the main drivers of protein evolution. The result is an array of single and multidomain proteins that make up an organism's proteome. We can now view the protein domain as a functional module, which can be cut and pasted into new multidomain contexts while remaining able to perform the same task. This modular capability immediately lends itself to engineering: we don't have to go about finding or artificially evolving a protein that performs our chosen task; we merely combine components that together are greater than the sum of their parts.
I'm interested in the defense mechanisms within cells — mechanisms that specifically recognize and disable intracellular pathogens. This type of defense is considered separate from the two main branches of immunity that are more intensely studied: the evolutionarily ancient "innate" immune system and the vertebrate-specific "adaptive" immune system. Innate immunity is the recognition of conserved features of pathogens—for example, the detection by specialized cells, such as macrophages, of the sugary capsule that surrounds many bacteria. Adaptive immunity works by fielding a huge diversity of immune recognition molecules, such as antibodies, and then producing large quantities of those that recognize nonself, pathogen-derived targets. The newly discovered kind of immunity on which I work, sometimes termed "intrinsic immunity," shares features with innate immunity but tends to be widely expressed, instead of residing just within "professional" immune cells, and is always "on." In other words, every cell in an organism is primed and ready to disable an invading pathogen. The intrinsic immune system is at a strategic disadvantage, as its targets are often fast-evolving viruses that can rapidly mutate to evade recognition. Unlike the adaptive immune system, which can quickly generate a response to an almost infinite diversity of targets, the intrinsic immune system must rely on rare mutations and blind selection over evolutionary time to compete with its opponents.
So far, the study of the intrinsic immune system has been dominated by its interaction with retroviruses. The retroviruses, an ancient affliction of vertebrates, violate the central dogma of biology—that DNA makes RNA makes protein1: they are RNA viruses able to generate DNA copies of themselves and insert this Trojan-horse code into the host's genome. Almost one-tenth of the human genome is the defunct relic of this sort of infection.
Within the past 7 million to 12 million years, a comparatively recent member of the retrovirus family, the lentivirus, has emerged and spread slowly through the branches of the mammalian family tree. The oldest known traces of lentivirus have been found in the genome of rabbits, but current infections occur in horses, cats, ruminants, and primates. Lentiviruses arrived in humans in the form of HIV, as several cross-species transmission events from other primates. Only one of those viral transmissions— from chimpanzee to human, sometime in the late nineteenth or early twentieth century—has adapted to its new host in such a devastating manner, the virus being HIV-1 M-group, which causes AIDS and currently infects 33 million people worldwide.
One of the major players in intrinsic immunity is TRIM5, a four-domain protein that is expressed in virtually every cell in the human body. By virtue of one of its domains—the RING (which stands for Really Interesting New Gene) domain—TRIM5 has an enormously high turnover rate; that is, each of its molecules is degraded within about an hour of the cell's having synthesized it. By virtue of another of its domains, it can recognize and engage retroviruses soon after their entry into the cell. As a result, incoming viruses can be degraded along with TRIM5 and thus made noninfective. A classic arms-race situation has developed, wherein TRIM5 has tried to maintain its ability to recognize the rapidly evolving retroviruses, placing the gene under some of the strongest Darwinian selection in the entire primate genome. However, HIV-1 seems to have the upper hand at the moment: the human TRIM5 variant only marginally reduces the replication of HIV-1. Could this be one of the failures in human immunity that has permitted such a dramatic invasion by this pathogen? And what does human TRIM5 need to do in order to gain the upper hand? Or, to ask a bolder question, what can we do to it to engineer resistance to the disease?
One surprising answer is provided by protein-domain fusions in other primate species. A fascinating thing has happened in South American owl monkeys: TRIM5 has been fused with a small protein that HIV-1 depends on for optimal replication. The resulting fusion protein is called TRIMCyp and can reduce the replication of lentiviruses by orders of magnitude, essentially rendering the owl monkeys' cells immune to the virus. Almost unbelievably (and it amazes me that espousers of Intelligent Design aren't onto this), this feat of genomic plasticity has happened twice: versions of TRIMCyp have also been described in the unrelated macaque lineage. Since no wild populations of owl monkeys or macaques have been shown to harbor lentiviruses, it is difficult to say whether TRIMCyps have been selected specifically to combat lentiviruses, but there remains the intriguing possibility that TRIMCyp has helped lead to the current lentivirus-free status of one or both of these species.
So how can we benefit from gene fusions in other species? The first lesson is that by splicing together domains from seemingly unrelated proteins, unexpected and useful products can be generated. Researchers have already generated human TRIMCyps, which would avoid the immune-system rejection that introducing an owl monkey gene would produce. My colleagues and I have also engineered a feline TRIMCyp that prevents replication of the feline immunodeficiency virus in tissue-culture systems. However, in a clinical setting TRIMCyp must be expressed within cells to be useful as an antiviral, and the only effective means of achieving this is through gene therapy to alter the target cell's genetic material. In a neat twist of roles, the best means we have of doing this is with a modified retroviral vehicle, or vector, to introduce a stretch of engineered DNA into the genome.
Two years ago, the first person in history was rid of an HIV infection. The situation was unusual, as the patient also suffered from leukemia, meaning that a bone marrow transplant was required. Instead of choosing a typical donor, an individual with a rare but naturally occurring mutation that prevents HIV entry into cells was chosen. The donor's mutation impairs the protein CCR5, a cell surface receptor that HIV requires for entry. The mutation has helped inspire some of the most recent antiretrovirals, CCR5 inhibitors. The recipient of this "delta-32" mutant CCR5 marrow showed no sign of HIV replication eighteen months after the transplant, despite having ceased antiretroviral treatment. Using this example as a proof of principle, it is quite possible that within the next five years a similar treatment will be delivered to a leukemia/HIV patient but with human TRIMCyp providing the antiviral activity. Indeed, we already have the two naturally occurring animal models—owl monkeys and macaques—that suggest that the fusion is relatively benign. Farther down the line, it might be possible to alter the genetic makeup of the target cells without the need for a bone-marrow transplant, a prophylactic gene therapy that could be as easy to deliver as any other vaccine. Given the issues surrounding the public acceptability of genetically modified crops and even conventional inoculations in some circumstances, whether such treatment will prove to be ethically and socially acceptable remains to be seen—but the technology is fast approaching, and we will have to decide whether or not to take advantage of it.
Of course, the power of fusing domains is not limited to antivirals. The scope of this technique is almost limitless, and the first generation of protein-fusion therapeutics is already upon us. Many of the recent anticancer drugs are protein-domain fusions—mostly fusions between mouse and human antibodies that block aberrant pathways specific to the cancer in question. The fusion is required because the mouse antibodies must be "humanized" in order to prevent an immune reaction to the foreign protein. Another potentially exciting and recently widely publicized development is a gene-therapy vector based on the adenovirus, which causes upper respiratory infections. Researchers can simply drop in "cassettes" of DNA to create fusion proteins that retarget the virus to specific cancer cells, where it will replicate and kill its target. There will also be nonmedical biotechnology applications for fusion proteins—for instance, the modification of enzymes to perform particular tasks, such as hydrogen or biofuel generation. Applications may be even more far-fetched: I recently attended a lunch-time seminar where arrays of viruses with domain fusions were being touted as electrical components of the future.
Using traditional techniques (and in molecular biology, "traditional" refers to techniques of the past twenty years), you can generate such fusions once you have your hands on a physical copy of all the genes you want to splice together and a student or postdoc, such as myself, to carry out the necessary protocols. With luck, it's a week's work to make a fusion or two; without luck (or competence), it can take a month. Getting access to the desired DNA often involves a legal material-transfer agreement, which can add further weeks to the procedure. And because time is money, we are fast approaching the moment when generating the fusion by hand is more costly than having the gene synthesized at facilities that specialize in such work. The cost of synthesis is dropping fast. Since I placed the order for my virus, the best quote for gene synthesis for a small gene has dropped from 39 to 35 cents per base. However, the economic case alone massively undersells what gene synthesis has to offer molecular biology. We need only the sequence information, not the physical DNA template itself, and a researcher can select any sequence in the public domain. Suddenly, with only an Internet connection, a hydrogenase from a sea-dwelling archaebacterium is just as obtainable as lion TRIM5. The new central dogma then becomes Silicon Makes DNA Makes Something Useful.
There will never be an exclusive field of research called domain-fusion science. The fusions will be made by researchers, clinicians, and engineers who in their own line of work require a protein with the function of more than one existing domain. Some of the fusions will be obvious next steps, some will be inspired leaps, and some will even be pulled from randomly assembled large-scale screens of domain fusions. The closest field in its own right to "domain-fusion science" is synthetic biology, a discipline that is busy systematically cataloging and defining the parts necessary to engineer useful biological systems. This includes domain fusions but does not stop there. The parts include the regulatory "control switches" of genes— promoters, enhancers, and repressors—as well as suites of genes that are designed to act in a tractable and predictable manner. In essence, synthetic biology is the reduction of complexity to modularity in order to rebuild complexity. To many researchers, it will seem a statement of the obvious even to note the existence of domain fusions, yet fusions have provided some of the most innovative therapeutics and molecular tools of recent decades, and their effects will be increasingly felt in the years to come. Their application in immunity really excites me. There has always been a twin requirement to recognize and then disable pathogens. If we can create therapeutics that at once engage and label the invaders as threats in a manner comprehensible to the cell, we may well regain the advantage over retroviruses and defeat many of our other current afflictions. As for my synthetic virus, it now remains for me to see if it works.
1 This commonly used shorthand for the central dogma is not actually what Francis Crick intended; he wanted to imply that information was unable to flow back to nucleic acid once translated into protein. This more precise version of the central dogma is not violated by retroviruses.
Edited by Max Brockman
JUST PUBLISHED — AVAILABLE IN BOOKSTORES AND ONLINE — ORDER NOW!
Advance Praise for Future Science:
"This remarkable collection of fluent and fascinating essays reminds me that there is almost nothing as spine-tinglingly exciting as glimpsing a new nugget of knowledge for the first time. These young scientists give us a treasure trove of precious new insights." — Matt Ridley , Author, The Rational Optimist
"I would have killed for books like this when I was a student!" — Brian Eno , Composer; Recording Artist; Producer: U2, Cold Play, Talking Heads, Paul Simon
"Future Science shares with the world a delightful secret that we academics have been keeping — that despite all the hysteria about how electronic media are dumbing down the next generation, a tidal wave of talent has been flooding into science, making their elders feel like the dumb ones..... It has a wealth of new and exciting ideas, and will help shake up our notions regarding the age, sex, color, and topic clichés of the current public perception of science." — Steven Pinker , Johnstone Family Professor, Department of Psychology, Harvard; Author, The Language Instinct
Eighteen original essays by:
Kevin P. Hand : "On the Coming Age of Ocean Exploration" Felix Warneken : "Children's Helping Hands" William McEwan : "Molecular Cut and Paste" Anthony Aguirre : "Next Step Infinity" Daniela Kaufer  and Darlene Francis : "Nurture, Nature, and the Stress That Is Life" Jon Kleinberg : "What Can Huge Data Sets Teach Us About Society and Ourselves?" Coren Apicella : "On the Universality of Attractiveness" Laurie R. Santos : "To Err Is Primate" Samuel M. McLure : "Our Brains Know Why We Do What We Do" Jennifer Jacquet:  "Is Shame Necessary?" Kirsten Bomblies : "Plant Immunity in a Changing World" Asif A. Ghazanfar : "The Emergence of Human Audiovisual Communication" Naomi I. Eisenberger : "Why Rejection Hurts" Joshua Knobe : "Finding the Mind in the Body" Fiery Cushman : "Should the Law Depend on Luck?" Liane Young : "How We Read People's Moral Minds" Daniel Haun : "How Odd I Am!" Joan Y. Chiao : "Where Does Human Diversity Come From?"