A publication of the Foresight Institute
Chemists at the University of Colorado have designed and made
a modest-size peptide molecule called "CHZ-1" that
imitates the activity of chymotrypsin. (Chymotrypsin is an enzyme
that cleaves bonds on the acid side of the amino acids
phenylalanine, tyrosine, and tryptophan.) This is the first
report of a catalytically active peptide having gone all the way
from de novo design to functioning molecule.
While the activities of CHZ-1 and chymotrypsin are similar, their structures have almost nothing in common except at the active site where substrate molecules are bound and transformed. In chymotrypsin the actual work of catalysis is carried out by a particular configuration of three amino acids: histidine, serine, and aspartate. One of the main tasks of the rest of the enzyme is to maintain these amino acids in their relative positions. CHZ-1 was designed with the same three amino acids held in a similar configuration.
CHZ-1 is much smaller than chymotrypsin: 73 amino acids versus 245. The two catalysts have many substrates in common; for these, CHZ-1 cleaves bonds at about 1% the rate of chymotrypsin--a 100,000 times acceleration over the background rate. Heat tolerance of CHZ-1 is greater than that of chymotrypsin, but this difference may be attributable to the chemists' employment of a non-standard amino acid and several non-peptide bonds to hold the molecule together. These would not be found in an enzyme of biological origin.
CHZ-1 was made in a protein synthesis machine; it could not have been produced by recombinant DNA methods, since it contains nonstandard parts. But having shown that the basic activity of an enzyme can be transferred to a very different molecule simply by copying the design of the active site, chemists will no doubt soon develop active peptides consisting of single chains of amino acids that can be produced in quantity by engineered microorganisms. [See Science 249,1544-1547,22Jun90]
In his book Engines of
Creation, Eric Drexler envisioned machines able to
assemble structures with atomic accuracy by thrusting each part
into an appropriate site on the workpiece, using an angle and
velocity likely to promote formation of the desired bond. The
reasonableness of this picture of an assembler, not so obvious
five years ago, is becoming more apparent as chemists explore the
mechanisms of chemical reactions. A good example is provided by
the work at MIT of Sylvia T. Ceyer and her colleagues who have
been using molecular beams to study the adsorption of small
molecules onto metal surfaces. Metal-catalyzed reactions
constitute a large class of chemical processes that have been
widely used but poorly understood--until now.
Ceyer's group investigated one such reaction in great detail: the adsorption of methane onto nickel. The key factor is the velocity of the molecular beam--specifically, the speed at which the incident molecules approach the nickel surface. Since a methane (CH4) molecule is a carbon atom surrounded by four hydrogen atoms, the first atom to near the surface is always a hydrogen. If the impact speed is great enough, this hydrogen will be pushed aside, allowing the carbon atom to approach and bind to a nickel atom; the hydrogen atom, now free, binds to a different nickel atom. At lower speeds, the methane molecules remain intact; some are trapped by forces near the metal surface, others bounce off and escape.
The MIT researchers found they could control this and similar reactions by varying the parameters of the molecular beam (e.g., the velocity and angle of incidence) and the temperature of the nickel surface. They discovered that the reactions occur at lower incident velocities when the methane molecules are given extra vibrational energy before sending them to the nickel surface--presumably because the vibrational distortions give carbon and nickel atoms easier access to each other. And they found that unreacted methane molecules trapped near a metal surface can be forced to react with and bind to it simply by "hammering" them with a beam of neutral atoms (such as argon).
This work confirms the assembler concepts put forward in Engines of Creation--atoms and molecules can indeed be added to a workpiece by hammering them against it, and they can be pre-processed to enhance their reactivity. [See Science 249:133-139,13Jul90]
Small clusters of metal or semiconductor atoms give rise to
properties not seen composition. Adding or removing a few atoms
from an ordinary sample will not change its properties, but this
is not true of samples whose component particles each contain
only a few dozen atoms or less. For example, a cluster of nine
cobalt atoms is practically inert to hydrogen or nitrogen gas,
whereas a cluster of ten cobalt atoms is quite reactive.
Cluster research is aimed partly at finding ways to make clusters in quantity. Current methods produce a mixture of cluster sizes, complicating the study of their structure and behavior.
The fact that the properties of these substances depend so critically upon cluster size has mixed implications for nanotechnology. On the positive side, it suggests that the range of possible characteristics that materials may possess could be much broader than we realize. But on the negative side, it means that the characteristics of materials can be very sensitive to small errors in design or construction. [See Science 248:1186-1188,8Jun90]
Nanotechnology uses assemblers; biochemistry uses enzymes;
chemistry uses catalysts; carpentry uses tools. Assemblers,
enzymes, catalysts, tools--four examples of objects that control
the processing of other objects.
We're all familiar with the evolution of tools, from crude hammers and chisels capable of only the roughest sort of production, to complex machine tools that control the shapes of manufactured objects with micron accuracy. Enzymes underwent a similar evolution more than a billion years ago, developing a complexity and variety that enabled them to conduct the biochemistry of life.
Analogous to these two traditional lines of development is the current progress in chemical catalysis. Catalysts are substances that direct the course of chemical reactions without themselves being used up; catalysts participate in the reactions, but they emerge intact and so are available for another round. Generally speaking, simple catalysts are less specific than complex catalysts. If a catalyst is to promote specific reactions and not others, then it must contain sufficient structure to enable it to distinguish between the reactants it is to use and those it is to ignore.
In recent years a sophisticated class of catalysts has emerged from research laboratories such as that of Ryoji Noyori at Nagoya University. Noyori has been studying what are called "chiral metal complexes" in which a metal atom is bound to an asymmetric molecule to form a catalytic complex. Such catalysts distinguish between reactants not only on the basis of their chemical structure, but their chirality as well. (Chirality is the symmetry property that causes certain structures to be mirror images of each other but not identical--the same property that prevents left-handed nuts from fitting on right-handed bolts.) Ruthenium-BINAP catalysts are especially promising examples--their superiority over conventional catalysts has been demonstrated for the production of dozens of commercially important chiral chemicals.
Noyori says, "In principle, any chiral structure can be generated through rational modification of the catalyst's molecular structure." From a traditional chemical viewpoint it seems hard to believe that there would not be some chiral structures for which no appropriate catalyst could be designed. After all, traditional chemistry generally takes place in solution where substrate molecules bump around randomly and often prefer different reactions than the chemist does. On the other hand, if chemistry is a stage in the development of nanotechnology, then catalysts should be thought of as rudimentary assemblers that are slightly "programmable" through changes in the reaction milieu (i.e., changes in pH, temperature, etc.). Plain metal catalysts, like platinum or nickel, have played a major role in chemistry despite their simplicity. In chiral metal catalysts the unique catalytic features of metal atoms are combined with structures that aid in the recognition and handling of desired substrates, and that can be more readily "programmed" by the milieu.
As catalysts become more sophisticated, they will become more complex, more varied, more programmable, and more selective; their descendants sometime in the 21st Century may well turn out to be the molecular assemblers we discuss in Update. If they do, then Noyori's claim might evolve into this one: "In principle, any physically realizable molecular structure can be constructed by appropriately programmed assemblers." [See Science 248:1194-1199,8Jun90]
Rejuvenation buffs will be interested in the work of Calvin B.
Harley, et al. at McMaster University and Cold Spring
Harbor Laboratory. These researchers have shown that human
fibroblast cells undergo gradual losses at the ends of DNA
In organisms having linear chromosomes (such as yeast and higher organisms), the replication of DNA during cell division is often incomplete--base pairs are lost at the ends of the DNA molecules. To guard against the loss of important information, the end segments of the DNA consist of repetitive sequences of base pairs that contain no essential information; these are called "telomeres."
Organisms that do not age (like yeast) have "telomerase" enzymes that maintain the length of telomeres by adding repetitive sequences when necessary. Higher organisms also have telomerases, but these appear to be active only in the production of reproductive cells (e.g., sperm) and in tumors. Consequently--and this is what Harley et al. have shown--human somatic cells lose about 50 base pairs per DNA terminus per cell division, on the average. Since sperm DNA has about 9000 base pairs of repetitive DNA at each terminus, the process of incomplete replication would have eaten into critical parts of the DNA at a given terminus after about 180 cell divisions. There are, however, 92 different telomeres in each human cell (23 pairs of chromosomes x 2 telomeres per chromosome). A cell may die or become impaired if even one of these 92 telomeres begins losing critical information--an event that would generally occur sooner than the average.
If telomere shortening proves to be a major mechanism of aging, then gene therapy offers a possible way to deal with it. We can envision a day when genes can be introduced into the human genome to provide a telomerase system that has been redesigned to be active in somatic cells. [See Nature 345:458-460,31May90]
Shoichiro Yoshida and his research team with the Research Development Corporation of Japan have completed a five-year project aimed at developing instruments and techniques for measuring and processing at nanometer scales. Among the fruits of this effort are:
This is just one of 21 projects in Japan's national ERATO program. With efforts like these taking place, progress toward nanotechnology should be rapid. [See Nanotechnology 1:13-18,1990]
Atomic force microscopes construct images by scanning a sharp
tip over a sample at sub-nanometer distances and measuring the
force between tip and sample. The tip is fastened to a cantilever
arm; samples lie on an atomically-flat surface (or
Lacking techniques for making atomically perfect tips, researchers have had problems with resolution, interpretation and reproducibility. Earlier this year Eric Drexler at Stanford and John Foster of IBM suggested that these problems could be alleviated if AFMs were equipped with engineered molecular tips [See Nature 343:600, 15Feb90]. A variety of different molecules could be designed to have desired characteristics and then synthesized with atomic precision by chemical methods.
This earlier work left unanswered the important question of how such molecular tips could be installed and placed on the AFM's cantilever. In a paper presented in July at the Fifth International Conference on Scanning Tunneling Microscopy/Spectroscopy and First International Conference on Nanometer Scale Science and Technology, Drexler suggests an answer to this question: the tips need not be installed on the cantilever at all. In the new arrangement, the sample is to be held on a round bead fastened to the cantilever; a variety of tips are bound to the stage, not necessarily in an organized pattern. To image a sample, the operator must first find an appropriate molecular tip on the stage by broadly scanning the stage with the bead--in this mode of operation, the stage with its array of tips serves as the sample, and the bead acts as a probe. When a molecular tip is found, a confined scan is carried out so that the molecular tip can image a sample bound to the bead; in this scan, the bead and stage have exchanged roles.
An even more interesting application of this new design would be in molecular construction. The array of molecular tips could be designed so that each tip binds a reactive atom or molecule. As these "parts" are added to a workpiece located on the bead, they would be replenished from the surrounding solution. [See Journal of Vacuum Science and Technology B, in press]
[Editor's note: The publication reference for the JVST-B article is: Drexler, K.E. (1991) Molecular tip arrays for molecular imaging and nanofabrication. JVST-B 9:1394-1397. See also section 15.4 of Nanosystems.]
Russell Mills is research director at Group 9 Research Associates in Palo Alto, California.
From Foresight Update 10, originally published 30 October 1990.
Foresight thanks Dave Kilbridge for converting Update 10 to html for this web page.