The next night, the Virals tried the old combination without success but soon came up with the new combination, since it was only one mutation away from the original. The Buy ‘n’ Cell staff responded the next day by changing the combination to 892220611 and giving the new number to all employees. Later that night, after a few minutes of trying different numbers, a genius-level member of the gang realized that even going at the rate of one try per second, nonstop, it would take them half a billion seconds (fifteen years) to try even half of all the possible combinations. So they gave up and retired from the scene, but not before covering all the metal gates with graffiti copies of Napoleon’s famous epigram, “Glory is fleeting but obscurity is forever.”
So how does this stirring tale relate to the process of changing the genetic code?
We start with transfer RNA (tRNA)—folded molecules that transport amino acids from the cytoplasm of a cell to a ribosome that strings together amino acids into proteins. Each tRNA molecule is essential for the life of the host, unless you can get rid of every instance of the codon that needs that particular tRNA to transport an amino acid to the ribosome. To get rid of those codons you’d need to swap each of them with a so-called synonymous codon that uses a different tRNA but translates to the same amino acid. Or, in terms of the parable, to do that you’d need to swap that combination (codon) with a so-called synonymous combination (codon) that uses different numbers (codon letters) but performs the same function (prompting a tRNA to transport the same amino acid to the ribosome). (A synonymous codon is one whose different nucleotides encode the same amino acid; for example, both AGA and CGG encode the amino acid arginine.)
Our first goal in the lab was to remove the gene for the protein called “release factor 1” (nicknamed RF1). This protein is required for E. coli ribosomes to stop adding amino acids at the end of making any of 322 different proteins that terminate in the stop codon UAG. RF1 binds to the mRNA at the UAG and causes the ribosome to stop. We couldn’t delete RF1 until we changed all 322 instances of UAG to the synonymous UAA codon (which is handled by a different protein known as RF2). When an invading virus comes along, it needs the RF1 to use the UAG codon, but RF1 isn’t there anymore! The ribosome will still manage to function, but badly. Nevertheless, probably two or more codon changes will be needed to defeat all viruses.
Figure 5.2 A small segment of the genetic code (from Figure 3.2) showing all the six synonymous codons for the amino acid arginine (R): AGA, AGG, CGC, CGU, CGA, and CGG.
When we began these experiments we weren’t sure that these so-called synonymous codons would all be truly equivalent, because living things tend to use their parts in many gloriously complex and overlapping ways. Ways that would be considered “spaghetti code” if a programmer ever tried such undisciplined and tortuous protocols. However, we did successfully make all of the 322 changes. For our synthetic purposes (not necessarily for all of the natural purposes of the cell), these changes are innocuous and hence we can take the next steps toward multivirus resistance.
How did we manage to make so many changes? The answer lies in the process and instrument called MAGE, introduced in Chapter 3. That process allows us to make many changes simultaneously and in parallel.
Next we target the tRNA that recognizes the AGA and AGG codons, normally translating these to the amino acid arginine (R). In Figure 5.2 we see the six synonymous codons for arginine. If every one of the AGA and AGG codons is changed to CGA, CGG, CGC, or CGU, all of which also code for arginine, then we can remove the gene for the tRNA that binds, recognizes, and translates the AGA and AGG codons (otherwise removing them would be lethal to the cell). If that tRNA is missing, the ribosome will stall at each AGA and AGG and make defective versions of essential cell proteins. Even if other tRNAs jump in and prevent stalling, the number of incorrect amino acids would be disastrous to the cell. Such genome-wide reassignment of codons can be extended to still other codons and then the corresponding tRNAs can be deleted. For example, leucine (L), like arginine, has six synonymous codons.
Figure 5.3 A small segment of the genetic code (from Figure 3.2) showing all six synonymous codons for the amino acid leucine (L): CUC, CUU, UUC, UUU, UUA, and UUG.
With regard to the piece of genetic code in Figure 5.3, we can now free up (make nonessential) one of the three tRNA molecules for leucine by moving all of the CUC and CUU codons to CUA or CUG. At that point, the host has freed up (eliminated all uses of) five of the sixty-four codons (UAG, AGA, AGG, CUC, CUU) and we have deleted three genes that are normally essential to translation and hence are essential for life. This means that invading viruses not only can’t use the cell machinery but they don’t even have a chance of mutating to use it.
As a concrete, real-life example of this, one of the smallest functional peptides in all of biology has the five amino-acid sequence MRLFV (methionine-arginine-leucine-phenylalaline-valine, which provides resistance to the antibiotic erythromycin). If the peptide were initially encoded by the following five triplets,
AUG AGGCUUUUUGUGUAG
then after recoding, the sequence would be:
AUG CGGCUAUUUGUG UAA
(altered codons in black boxes), still making the same old MRLFV peptide, because the changed codons are synonymous with and perform the same function as the original ones. But if a tiny gene like the original one above were required for an incoming virus to be replicated, the virus would still be using the original five codons; and if the host had repurposed the original three, then the virus would find (much to its surprise!) that three of its precious codons (AGG,CUU,UAG) no longer worked.
We are in the midst of changing a dozen codons genome-wide in E. coli (in parentheses is the one-letter abbreviation for the amino acid or stop encoded, together with the number of times each is used in the small viral genome below): ACC (T, 24), AGA (R, 10), AGG (R, 6), AUA (I, 19), CCC (P, 10), CGG (R, 12), CUC (L, 27), CUU (L, 19), GCC (A, 20), GUC (V, 21), UAG (stop, 2), UCC (S, 20). This recoding strategy would affect 190 codons of the 1,147 total in the virus below. Most viruses use those twelve codons even more times in essential functions and “expect” a good host (or hostess) to provide the tRNA molecules that recognize them. But after our recoding, the essential viral RNA molecules stall on the ribosomes as soon as any of the missing tRNA molecules are needed, and the game is over.
Figure 5.4 is the full sequence of one of the smallest known viruses (MS2)—the first genome ever sequenced (in 1976 by Walter Fiers and team). It’s related to the first genome subjected to Darwinian molecular evolution in the lab (Sol Spiegleman’s monster Q-beta in 1965). The natural MS2 genome encodes four known proteins, which (in order along the genome, which is also essentially the mRNA) are: the maturation protein (one copy per virus) that holds onto a unique point in the RNA genome and guides it into the shell coating the virus and later into the host cell; the major coat protein (180 copies form the outer shell of the virus); a lysis gene used to help the baby viruses get out of the trashed cell host; and finally, the gene for the polymerase that makes copies of the genome. Interestingly, the lysis gene extensively overlaps the coat and replication proteins (underlined) so those base pairs are duplicated in parentheses in Figure 5.4. The protein coding regions are indicated by spaces between the triplet codons. The regions without spaces are not translated into proteins but are useful in regulation of timing and levels of the proteins.
Figure 5.4 The smallest viral genome? Think of the above as a detailed picture of a beautiful cathedral. You don’t have to see or understand each stone, but you should get a feeling for the type of motifs that the architects used.
In Figure 5.4, the twelve codons that we are changing in the host (E. coli) of this virus are in uppercase and black-boxed. You can see how unlikely it is that this virus would get exactly these 190 changes all at once.
An essential condition here is that you must change the E. coli cell’s genome while the virus isn’t present in the cell. If you do it slowly while the virus is inside the cell, as in
natural evolution, then the virus keeps pace step by step with every bacterial innovation in a kind of arms race. By doing this recoding “offline” we can introduce something that requires so many changes all at once in the virus that it can never get enough random mutations made simultaneously to accommodate the new state of the cell. Indeed, the changed cells should now be resistant to viruses that they have never seen before. In the best case for the virus, it would have a small genome (fewer chances to get one of the missing codons), a high mutation rate (to get rid of the newly impotent codons), and a large viral population (more ways to win). The smallest viruses that don’t depend on other viruses are about 3,000 base pairs in length. Let’s say (very generously) that would be 10 places in the 3,000 base pairs that need to change to deal with the new host code. I explained how high a mutation rate can go in Chapter 3, so let’s pick an aggressively high rate of one error in 1,000 base pairs or about three errors per viral genome. That means 5 percent will have no mutations and 0.03 percent will have ten mutations or more. Of those with ten mutations, only one in 10,000 will be viable and of those, one in 1024 will have the correct ten changes in the genome. So a population of 1028 of that species of virus would be needed to find one that could handle the new host. But that is on the order of all virus particles of all types in the world today. The odds get worse fast for the virus if more than ten changes are needed, and exponentially worse with the number of required changes.
E. coli has about 4,200 protein coding genes. Once we’ve been successful with E. coli and other industrial microbes, then we can tackle agricultural plants and animals that have some 20,000 protein genes—only five times that of E. coli. The changes would be introduced to the germ lines of the animals so that all of their progeny would also be resistant to all viruses. This is typically done by engineering pluripotent stem cells in culture and then implanting them into blastocyst stage embryos (one in the later stages of cleavage).
If all of the above practical applications work out well, then the temptation to apply the same technique to humans, and to make ourselves multivirus resistant, will be quite strong. Initially this might be done on stem cells for adult therapies; for example, blood stem cells are routinely transplanted to cure blood diseases like cancer. If you could offer stem cells that are both cancer free and virus resistant, they might be more desirable than the same cells that are merely cancer free. Indeed, there will be a demand for those cells to be not just cancer free but cancer resistant too. That would require another set of genome changes almost as complex as those needed to produce viral resistance.
Possible unintended negative consequences of using this technique would also have to be taken into account; for example, nutritionally enslaving the changed cells so that they can’t escape from the lab and beat all natural relatives (which are still virus-limited). But the potential payoff of multivirus resistance is so huge that this is a procedure that almost demands to be tried.
Multivirus resistance, if and when it’s engineered into the human genome, does not necessarily make us immune to nonviral diseases such as those caused by parasites (e.g., malaria), bacteria (e.g., tuberculosis), or prions (e.g., mad cow disease). The pathogens that cause these diseases don’t work by replicating themselves using human translation machinery of the human host cell. A speculative approach that could, in principle, provide protection from both viruses and cellular parasites is the mirror cell idea mentioned in Chapters 1 and 3. This would be much more challenging than making a simple mirror bacterial cell or making a viral-resistant human using code changes, but it’s probably feasible. All of the tricks that pathogens use to evade the human immune system would fail in mirror humans since those tricks depend on specific chiral receptors. Furthermore, the enzymes they need to poke holes and digest the polymers of the host (DNAses, RNAses, proteases, lipases, etc.) employ high chiral specificity and hence would be useless, and nearly all of the favorite foods needed to sustain those pathogens would be unavailable.
Finally, supposing that one day we have managed to conquer an impressive array of human maladies, including infectious disease of almost every kind, we might then turn our attention to a future era in which members of the human species might occupy an entirely new ecological niche . . . on another planet. Mars, for example.
One of the primary hazards of space travel is the threat of radiation damage. Even on earth, ultraviolet radiation causes skin cancers ranging from the relatively benign basal cell carcinoma to the relatively deadly malignant melanoma. And that’s with the shielding effects of the earth’s atmosphere already in place and working for us. Departing from that semiprotective buffer zone is to enter a realm that’s especially hostile to human health.
Radiation in open space is of two types: electromagnetic radiation, which consists of massless photons, and high-speed particles that have mass, including atomic nuclei and electrons. Electromagnetic radiation travels at the speed of light, and the various types are distinguished by their wavelengths (the distance from one wave crest to the next) and frequencies (expressed in kilohertz, or kHz). The shorter the wavelength, the higher the frequency, and the greater the energy carried by the photon. X rays and gamma rays are the highest frequency forms of electromagnetic radiation and can do the greatest damage to human cells, tissues, and DNA. Gamma rays can penetrate the body and rip apart DNA molecules and damage sperm and egg cells as well as the reproductive organs. Gamma rays are used in medicine to kill cancer cells.
Gamma rays are a form of ionizing radiation—they carry enough energy to strip electrons from the atoms or molecules to which they are attached. They affect the body’s chemistry in a way that can lead to serious health problems, including cancer, tumors, and DNA damage.
There’s a lot of radiation in space, and people traveling to Mars would experience six months’ worth of radiation every twenty-four hours. This means that a one-year outbound voyage would subject the traveler to a dose of radiation equivalent to 180 years on the home planet. Clearly spacefarers are going to need radiation protection.
Lead shielding is one way to do it. Unfortunately, the amount of lead required might make the spacecraft too heavy to reach escape velocity. Fortunately, there are radiation-resistant animals. If we could find the genes responsible for their ability to resist radiation, then perhaps we could import those genes into our own genome and become radiation-resistant ourselves.
The all-time champion radiation-resistant organism is the extremely hardy bacterium Deinococcus radiodurans, a creature whose very name means “radiation-enduring.” It is listed in the Guinness Book of World Records as the world’s toughest bacterium, and has been informally nicknamed “Conan the Bacterium.” It was discovered in 1956 by researchers conducting food sterilization experiments using high-dose gamma radiation, which it survived easily. This microbe can withstand a thousand times the gamma radiation that would kill a human being. Although its genome has been completely sequenced, bacteria are phylogenetically too distant from the human lineage for their genes to be easily usable.
But there is a backup radiation-resistant organism that happens to be an animal, the Bdelloid rotifer. This is a class of small invertebrates that live in freshwater pools, on the surfaces of mosses and lichens, and even in habitats that periodically dry up. To survive extended periods of desiccation, these animals have evolved the ability to repair portions of their DNA that gets damaged under these harsh conditions. In 2007, my Harvard colleagues Matthew Meselson and Eugene Gladyshev performed a series of experiments in which they subjected two species of these rotifers to high levels of ionizing radiation. They survived these megadoses well enough to repair their own double-strand DNA breaks and to produce viable offspring. (A human being would be killed by a hundredth of the radiation levels tolerated by the rotifers.) Note also the humble midge fly Polypedilum vanderplanki mentioned in Chapter 3 and Epilogue.
If we could locate the genetic sequences that code for the rotifer’s extraordinary DNA self-repair abilities, it wo
uld theoretically be possible to isolate and then import them into our own genome, with the result that we too would have those same genomic talents and consequently increased resistance to ionizing radiation.
That trick may help get us to other planets. Meanwhile, back here on earth we continue to suffer from thousands of diseases, including rabies.
In 2004, when Jeanna Giese entered the Children’s Hospital of Wisconsin, a team of eight physicians headed by Rodney E. Willoughby consulted with Jeanna’s parents about possible courses of treatment. They proposed an aggressive approach based on an untested strategy that might ultimately fail. Even if the patient were to survive, they said, she might wind up severely disabled. With essentially no other alternatives, Jeanna’s parents told them to go ahead.
They planned to induce a coma to put Jeanna’s central nervous system on hold and protect the brain from injury while the patient’s body mounted a native immune response to the virus. In addition they would administer one of the few antiviral drugs in existence, ribavirin, along with a cocktail of sedatives to maintain the coma and ensure suppression of the nervous system.
A week into the treatment, a lumbar puncture revealed an increase of rabies-specific antibody in both the bloodstream and the cerebrospinal fluid. Gradually the patient was brought out of the coma. After about two weeks of treatment, Jeanna blinked, opened her eyes, and moved them. Two days later she raised her eyebrows in response to speech. “On the 19th day,” according to the report published in the New England Journal of Medicine, “she wiggled her toes and squeezed hands in response to commands, [and] fixed her gaze preferentially on her mother.” Some nervous system abnormalities developed as she was brought back to full conscious awareness, but she continued to progress.
Regenesis Page 14