Proteins as Natural Products
What chemists talk about when they talk about making molecules is not necessarily what biologists talk about when they talk about making molecules.
For many biologists, the problem is often one of finding the right combination of expression system (E. coli or yeast strain), DNA "vector," and purification scheme. The best solution for a biologist is one that produces the most copious amounts of highly purified molecules, be they proteins, DNA, RNA, or carbohydrates, in the shortest amount of time.
For organic chemists, on the other hand, the problem of making molecules is also one of finding the right combination of steps. Natural productsthose substances formed by strange bacteria, tree bark, or other living creatures with some useful propertyare often chemists' targets. Making these substances is not a problem of picking an expression system, but of choosing among commercially available compounds and designing a synthesis with the fewest, most elegant reactions. The more complicated the architecture of the natural product, the greater the challenge.
And how would someone who is both a chemist and a biologist, say a bioorganic chemist, talk about making molecules? Would it be synthesizing a natural product (like a chemist) or expressing a protein (like a biologist)?
"The way that I think about proteins," says Assistant Professor Philip Dawson, an investigator in The Scripps Research Institute's (TSRI) Departments of Cell Biology and Chemistry and a member of The Skaggs Institute for Chemical Biology, "is that they are the ultimate natural productsmolecules that we can synthesize and, by having control over their chemistry, [design] any side chain or backbone atom."
Dawson, who is a graduate of TSRI's Kellogg School of Science and Technology, is working to move forward the methodology of chemical protein synthesis. His goal is producing common and novel protein forms, which are useful tools for studying everything from the basic molecular basis of protein function to applied questions of enzyme drug resistance. Currently in his laboratory are research associates Rema Balambika, Mike Churchill, John Offer, and Florian Topert, along with graduate students John Blankenship and Chris Neidre
The Difficulties of Synthetic Protein Synthesis
Dawson and his colleagues use the technique of solid-phase synthesis to make the peptides. Invented by Robert Bruce Merrifield in 1963 (for which he was awarded the 1984 Nobel Prize in Chemistry), solid phase protein synthesis basically entails building a peptide step-by-step, starting with a single amino acid that is attached to a polymer resin. Amino acids are then added one at a time, the resin is washed between each successive round, and finally the finished peptide is removed from the resin.
This basic technique has been improved dramatically through the years so that today it is used routinely. Some laboratories have their own automatic peptide synthesis machines, many research institutions have a core facility that offers peptide synthesis, and any laboratory in the country can order several milligrams of purified custom peptides through the mail from any one of several commercial companies.
Solid phase synthesis does have its limitations, however. It works well for peptides of around 20 to 25 amino acids and reasonably well for peptides up to about 50 amino acids. In some cases, scientists have even been able to synthesize proteins of around 100 amino acids.
"In general, though," says Dawson, "if you go much above 40 [amino acids], you are going to run into problems."
Not one to see a barrier as a barrier, Dawson has spent considerable effort to improve the sizes and the yields of proteins he can routinely make. He uses a number of tricks that enable him to make proteins and peptides up to about 150 amino acids long. In general, these tricks involve breaking the sequence up into pieces, synthesizing the pieces one at a time, and then chemically "ligating" or joining them.
This ligation offers its own unique problems, though, because of the potential for the reaction to incorrectly join two peptides. These ligations can join together a carboxy group with an amino group, but any single unprotected peptide may have some 10 to 20 of each group, most of them on the side chains, but the only two that need to be joined are the ones on the ends. The danger is obtaining a heterogeneous mixture of branched peptides instead of one uniform pool of linearly combined peptides. An added challenge is that the reactions must be done in water at neutral pH.
The solution, says Dawson, has been to modify the peptides slightly so that the reaction that joins the two ends is highly favorable and occurs much faster than any other potential reaction. This is done by making the N terminus of one peptide a cysteine residue and making the C-terminus of the other a thioester. The thioester reacts rapidly with the cysteine side chain and subsequently rearranges to the N-terminus to form a peptide bond at the site of ligation.
In fact, Dawson says, even in cases where a peptide might have as many as 14 internal cysteines, it is still possible to selectively perform reactions and selectively ligate one peptide to the cysteine at the end of another. However, having to use cysteines to stitch the peptides together may alter the sequences of the proteins in undesirable ways, and Dawson is developing ways of synthesizing proteins with natural sequences as well.
Now, Dawson and research associate John Offer are using a modified glycine residue that has the equivalent of a cysteine side chain to join the two peptides together. "Following that, we can knock [the cysteine equivalent] off," he says.
This, says Dawson, allows his laboratory to synthesize almost any protein up to about 150 amino acids. And Dawson and his laboratory can also combine solid-phase synthesis with biological expression and folding systems to achieve significantly longer length proteins.
Post-Translational Modification and HIV Resistance
One applied area in which Dawson and his laboratory work is the post-translational modification of proteins. Post-translational modification is a generic term that encompasses many of the myriad things that happen to a protein in a cell once it is translated from an mRNA transcript. For example, "phosphorylation" of proteins is a key element of signaling pathways. However it is often difficult to generate large quantities of a protein with a site-specific post-translational modification by biological methods.
One of the most common post-translational modifications is the attachment of specific sugars to proteins. In the future, the Dawson lab hopes to be able to adapt their chemical ligation techniques to incorporate complex carbohydrates to produce homogeneous N-linked glycoproteins. Another aim is to understand the regulation of "palmitoylation" at cysteine residuesa dynamic lipid modification important for protein localization in the membrane. Ultimately, he would like to be able to make any particular site-specific post-translational modifications he desires.
This expertise in chemically synthesizing proteins also gives Dawson and his laboratory the ability to routinely make proteins for a number of applied problems. One problem they work on is the protease encoded by the human immunodeficiency virus (HIV).
Dawson recently became involved in a large grant directed by TSRI Professor Arthur Olson that seeks to establish a drug design cycle aimed at developing, testing, and refining novel approaches to making specific inhibitors of HIV protease that are capable of limiting or eliminating drug resistance. Dawson is a co-principal investigator on the "Protein Expression and Analysis" core of the grant, and he provides mutant proteases and substrate with his synthetic technologies. He also works with his co-principal investigator, TSRI Professor John Elder, to look at the substrate specificity of the HIV protease, working with substrate and substrate-like inhibitors.
Dawson also works with Elder as co-Director of the Protein Chemistry Core on the Scripps NeuroAIDS Preclinical Studies center grant, where he applies his synthetic protein chemistry expertise to the assembly of small proteins, including a variety of chemokines and cytokines, which are relevant to the study of neuroAIDS.
Dawson, with research associate Rema Balambika, and graduate student John Blankenship also have several projects involving modification of the amide backbone of proteinssomething he refers to as "backbone engineering."
Backbone engineering basically entails modifying the basic structure of the amino acid, for instance, in order to remove the ability of the backbone to form hydrogen bonds with other parts of the protein molecule (this "H-bonding" is an important factor that drives folding and stabilizes the three-dimensional structures of proteins).
In general, this is an example of how protein synthesis provides a useful tool for studying protein folding and the formation of hydrogen bonding networks that are believed crucial for it. By changing amino acid "amides" to "esters," Dawson and his colleagues are able to knock out individual hydrogen bonds, and they can then perform refolding studies to see the effect of the removal of these bonds on the structures of the proteins.
One specific question that they have asked is what role such backbone H-bonds play in the formation of alpha helices. "We think of the alpha helix in terms of its hydrogen bonding network," says Dawson, adding that the synthesis of proteins with or without these bonds allows them to go to the particular bond in which they are interested and ask whether it is important for folding.
Another variation of this, which they performed recently, involves putting in esters to replace the amino backbone about every three residues in stretches of proteins that form alpha helices. Because of the nature of the alpha helix, this process allows them to wipe out all the hydrogen bonds along a single "stripe" of the helix.
They found that a protein with three ester bonds still folded to form a functional protein. This result suggests that alpha helices do not require the nucleation of multiple hydrogen bonds in order to fold. Says Dawson, "This was a little surprising."
Tying Proteins in Knots
Looking at another question of basic protein biophysics, another large area of research in Dawson's laboratory that has benefited from this chemical control has been the development of what are known as catenanescircular, interlocking protein rings.
These protein rings are basically short threads of about 40 amino acids that can be joined at the ends to make closed circle loops. In principle, these loops can be joined together to form a chain-linked protein polymer. They might turn out to be useful self-assembling materials that could form a two-dimensional sheet or a three-dimensional lattice. Furthermore, given the level of chemical control that Dawson and his colleagues could wield over these materials, it is possible to put metals or non-peptide binding proteins on the molecules site-specifically.
"We think this could be a great way to assemble proteins in a defined manner," says Dawson.
The first catenane that he and his laboratory designed was unusually stableso stable that the catenane was still folded in boiling water and formed a self-associating "dimer" of catenanes. This prevented further analysis. "It was easy to make but hard to study," says Dawson.
So Dawson's graduate student John Blankenship, who is receiving his Ph.D. from TSRI's Kellogg School of Science and Technology next month, fixed the problem. Blankenship designed a catenane based on a domain of the "tumor suppressor" protein p53. This system formed a single, isolated catenane consisting of two interlocking rings. This protein catenane was significantly more resistant to unfolding or proteolysis than the original p53 protein, and more so than any other protein cross-link characterized to date.
Additionally, Blankenship discovered that the catenane can be assembled in a step-wise fashion. "John found that if he cyclized one of the pieces first, he could thread the other piece through it and that the threading was very efficient," says Dawson. This enables the construction of heterocatenanes and other, more complex structures, as the sequence "threaded" through the cyclic protein need not be the same. Dawson adds that they are currently trying to characterize the threading process in more detail and determine how the process of threading alters the protein folding pathway.
Understanding how the threading process effects protein folding could provide a tool for manipulating the folding of p53or other intertwined proteinsin the cell. In principle, one could synthesize cyclic peptides that would interfere with some protein involved in cancer or another disease state. "There are several [cases] where you could actually inhibit a protein-protein interaction through threading," says Dawson.
And finally, the work demonstrates that making knotted proteins is feasible. The fact that proteins fold into stable three-dimensional conformations is well known, but what was less understood a few years ago was if and how proteins could thread themselves through a loop and make a knot. Most folded proteins, if they were grabbed by the N-terminus with one tiny molecular hand and the C-terminus with another and pulled, would unravel into a single thread. Very few would result in a knot.
In general, the success of this project was gratifying, says Dawson. "It wasn't clear that [the peptides] would be able to thread at allit was surprising how well it worked."