| SYNTHESIS |
Nature and science have developed two distinctly different ways of creating macromolecules, according to David A. Tirrell, chairman of the division of chemistry and chemical engineering at California Institute of Technology. As he observed recently, nature builds proteins with "exquisite architectural control," using a set of only 20 different building blocks (amino acids). Chemists, on the other hand, synthesize polymers with "lousy architectural control, but we compensate by using many different building blocks." Tirrell and other researchers are trying "to bridge the gap between these two kinds of macromolecular chemistry" by using nature's biosynthetic machinery to synthesize proteins containing nonnatural amino acids. This approach allows chemists to exercise nature's architectural control while opening access to a broader set of building blocks. The artificial proteins that result could have many important applications in biotechnology and materials science. In a lecture at the NanoScience & Technology Conference, held May 18-21 in Groningen, the Netherlands, Tirrell reviewed the progress his group has made in this area, culminating with their subsequently published foray into breaking the degeneracy of the genetic code. Tirrell PHOTO BY RON DAGANI Tirrell's talk focused on the synthesis and expression of artificial genes in bacterial cells, specifically those of Escherichia coli. "We start by designing sequences of amino acids, either natural or nonnatural amino acids, that we think will give rise to interesting behavior," he said. These sequences are then encoded into a complementary DNA sequence--that is, an artificial gene. "We then make the artificial gene by solid-phase organic chemistry and a few enzymatic ligation steps, which gives us the complete coding sequence for the polymer that we would eventually like to make." Recombinant DNA methodology is used to turn the artificial gene into an artificial protein. For someone like himself who comes from "a background in conventional polymer chemistry," Tirrell observed, "this approach to macromolecular chemistry has some very important advantages. First of all, it gives us essentially absolute control over the length, sequence, and the stereochemistry of the chains that we make. The second advantage, which I think will become increasingly important as we begin to understand the assembly of artificial proteins and perhaps their relevance to nanoscience and technology, is the fact that this gives us access to well-defined secondary, tertiary, and quaternary structures--the folded structures that are familiar in natural proteins." IN A CELL, DNA is transcribed into a messenger RNA (mRNA), which is then translated into a protein using amino acid building blocks. Transcription is "generally considered to specify uniquely the sequence of the protein," Tirrell said. But he showed how the cellular machinery can be subverted so that, for instance, a given mRNA can yield either a normal protein or a fluorinated protein, depending on which building block the cells are provided. The fluorinated protein may be of interest, for instance, because it has very different wetting properties than the nonfluorinated protein. So it's possible to "get away from the tyranny of the genetic code," he said, and "translate it in a different way" to make new materials with novel properties. Tirrell stressed two important steps in the decoding of mRNA. "The first thing we have to realize is that the real monomers for building up a protein chain are not the amino acids themselves." Rather, they're the aminoacyl transfer RNAs (tRNAs). "The tRNAs are the carriers of the amino acids, and it's the tRNA that interacts with the message [mRNA] that actually dictates the order in which the amino acids will be delivered to the end of the growing polypeptide chain." So if one changes the nature of the aminoacylation step by attaching to a given tRNA a nonnatural amino acid, when that tRNA is recruited to the ribosome, it delivers the novel amino acid rather than the natural one, he explained. That's, in fact, how Tirrell's group made a fluorinated protein--by attaching trifluoroleucine to a tRNA. The aminoacylation step is carried out by enzymes called aminoacyl tRNA synthetases, which activate the amino acids and attach them to the ends of their tRNAs. -------------------------------------------------------------------------------- It's possible to "get away from the tyranny of the genetic code and translate it in a different way" to make new materials with novel properties. -------------------------------------------------------------------------------- "The other step that's critical is the interaction between the codon in the mRNA and the anticodon at the tip of the tRNA," Tirrell noted. "If we can change the nature of codon-anticodon interaction, we can bring in a different aminoacyl tRNA in response to a given codon." With regard to the first step--aminoacylation--Tirrell and coworkers have explored three approaches to engineering the activity of the aminoacyl tRNA synthetases. In some cases, when the amino acid analog is not being activated and attached to the tRNA fast enough to sustain a reasonable rate of protein synthesis, the rate of aminoacyl tRNA synthesis can be boosted by equipping the bacterial cell with additional genetic copies of the synthetase. "When that doesn't work, we design mutant synthetases that will activate the nonnatural amino acids," he explained. And in some special cases, "the E. coli synthetases won't do what we need them to do, [so] we introduce synthetases from other organisms, in particular from yeast, to do the job in an E. coli cell." That combination of approaches has enabled scientists to insert a variety of nonnatural amino acids into proteins--30 to 40 in Tirrell's lab alone. These amino acid analogs contain double bonds, triple bonds, fluorines or other halogens, azides, cyclobutenes, and other functionalities that are not seen in natural proteins. Inserting a heavy atom like bromine into a protein (by way of p-bromophenylalanine), for example, provides a powerful tool for determining the crystal structure of the protein, Tirrell pointed out. And incorporating the azide functionality (as azidophenylalanine) has led, by way of photosensitive proteins, to the preparation of new biomaterials that his group is investigating for possible use in the reconstruction of injured blood vessels. TIRRELL HAS BEEN "pleasantly surprised" by the results of this work because originally, he told his audience, "I didn't think the synthetases could be engineered to accommodate such a broad variety of chemical functionality." Making a protein in which, say, all the phenylalanines are replaced with a phenylalanine analog is quite straightforward. But "what if we wanted to replace just one phenylalanine and leave the other 10--in a typical protein--in place?" Tirrell said. "How would we encode it? All the codons in the genetic code do something--there are no free codons." Chemistry professors Peter G. Schultz of Scripps Research Institute and A. Richard Chamberlin of the University of California, Irvine, independently demonstrated how to do this in 1989. They used one of the stop codons, which normally halt protein biosynthesis, to encode a nonnatural amino acid. They accomplished this by using a suppressor tRNA, a tRNA that will read a stop codon and deliver an amino acid to the growing protein chain in response to the stop signal. (Nature has developed suppressor tRNAs to suppress stop signals that arise from undesirable mutations.) Schultz and Chamberlin isolated a suppressor tRNA, chemically attached a nonnatural amino acid to it, and then allowed the cellular components to make the protein outside of the cell. At the time, there was no good way to do it inside the cell, according to Tirrell. Although this approach works, he noted, "it's difficult to do, and the scales are very small--micrograms or less." Tirrell decided to try to accomplish the same thing inside the cell. What that would require, he explained, would be a synthetase that ignores the E. coli tRNAs and that activates the amino acid analog and attaches it only to the suppressor tRNA. To accomplish this, Tirrell's late colleague Rolf Furter, then a postdoc, brought into the bacterial cell the genes for a yeast synthetase and a yeast suppressor tRNA. Their idea was that if the yeast synthetase activates the analog and the E. coli synthetases don't, the analog would be inserted by the suppressor tRNA only in response to the stop codon. And it worked, Tirrell said, adding that Schultz's group has achieved a similar feat using a synthetase and suppressor tRNA from a thermophilic bacterium. An especially important innovation that came out of Schultz's work, Tirrell told C&EN, is that Schultz developed a way to generate libraries of mutant synthetases and then screen them for mutants that can activate the nonnatural amino acid more rapidly than the natural one. "Peter developed a method that's more systematic," Tirrell remarked, "whereas we just guessed on the design of our synthetase pair and Rolf happened to pick one that worked well." BROKEN DEGENERACY Using a mutant yeast phenylalanyl-tRNA synthetase in conjunction with a mutant yeast phenylalanine tRNA containing a modified (AAA) anticodon in E. coli cells, Tirrell's group was able to make a protein in which phenylalanine was replaced by a nonnatural phenylalanine analog at UUU, but not UUC, codons. Both UUU and UUC codons normally code for phenylalanine. In his lecture, Tirrell then posed another question: What if one wants to replace five specific phenylalanines out of 10 in a given protein? One might, for example, want to modify just one domain in a protein having several domains. He noted that there are two codons that code for phenylalanine--an example of the degeneracy built into the genetic code. Could one codon be used for phenylalanine and the other one reassigned to code for a nonnatural amino acid, thereby breaking the degeneracy? Phenylalanine, as Tirrell explained, is encoded by UUC and UUU. Both codons are read by a single tRNA that is equipped with the anticodon sequence GAA. Considering the antiparallel nature of the pairing, GAA is a perfect complementary match for UUC but doesn't bind as strongly to UUU because of the G-U mismatch. Nevertheless, he noted, "this is the normal situation in an E. coli cell, where the mixed message is decoded by a single tRNA that reads both of these codons." Tirrell, with graduate student Inchan Kwon and postdoctoral fellow Kent W. Kirshenbaum, proposed that yeast tRNA carrying a phenylalanine analog be outfitted with the AAA anticodon--a perfect match for UUU. With the proper engineering, the mutant tRNA would read UUU codons faster than the normal bacterial tRNA carrying phenylalanine. The message might then be decoded in the following fashion: The yeast tRNA decodes UUU and delivers the analog (a 2-naphthyl-substituted alanine), while the normal bacterial tRNA decodes UUC and delivers phenylalanine. THE TEST PROTEIN contained nine phenylalanine residues, four encoded by UUC codons, five by UUU. The results of the experiment suggested that five of the phenylalanine residues were replaced with the naphthylalanine. As Tirrell noted in his talk, this approach to site-specific incorporation of nonnatural amino acids "doesn't work perfectly, but it works surprisingly well." Codon discrimination is very good, he said, "and we're now trying to refine that discrimination further." Tirrell has now published details of the work [J. Am. Chem. Soc., 125, 7512 (2003)]. The paper is "a beautiful demonstration of how to break the degeneracy of the genetic code," commented Paul R. Schimmel, a professor of molecular biology and chemistry at Scripps. The work also has implications for generating new biomaterials or proteins with enhanced stabilities, Schimmel noted. "The Schultz group uses a related but different approach that can also achieve the same results and applications." In closing his lecture, Tirrell observed that the so-called central dogma of molecular biology--"DNA makes RNA makes protein"--still holds sway. "But which protein you make depends on how you read the code," he said. "And there are powerful methods for reengineering the decoding process |
| UPDATE | 06.03 |
|
Want more information ? Interested in the hidden information ? Click here and do your request. |