|Dr. Gregory Fahy discusses his presentation on medical applications of nanotechnology with participants at Foresight's First General Conference in November 1992.|
Organic chemists have long been able to synthesize complex molecular structures, including drugs, by devising ingenious reagents and by contriving reaction conditions in such a way as to minimize undesired side reactions and maximize yield. The results obtained depend on the statistics of uncontrolled molecular collisions in solution involving all possible molecular degrees of freedom. Statistically improbable reactions may be slow. More important, solution-based reactions generally lead to unwanted and potentially toxic byproducts. This imprecision of present day chemical synthesis and manufacturing processes is one reason that drug standards are needed. In contrast, molecular nanotechnology would directly and rapidly produce the desired chemical and only the desired chemical, giving yields of 100%. This would be achieved by precisely positioning and bringing together the individual molecules involved in the reaction in such a way as to catalyze the reaction desired and only the reaction desired.
Molecular nanotechnology has many precedents. Enzymes are natural molecular machines that adsorb individual reactant molecules from the surrounding solution and, as a result of precisely orienting them with respect to each other in a protected "nanoenvironment," catalyze reactions in a highly specific manner at very high speeds and under mild reaction conditions. This simple process in biological systems ultimately allows synthesis of structures as diverse as carbon dioxide and hair. In fact, living organisms are naturally-existing, fabulously complex systems of molecular nanotechnology. If nature can produce the biochemical capabilities of living cells by accident, molecular engineers should be able to accomplish comparable, but broader capabilities by design, guided in part by the examples provided by living systems.
|Living organisms are naturally-existing, fabulously complex systems of molecular nanotechnology.|
Many industries already make use of enzymes to catalyze
desired reactions one molecule at a time. Genetic engineers are
producing pharmaceuticals by using naturally occurring enzymes to
edit DNA, and the soft drink industry uses enzymes that have been
modified to allow them to produce sugar at high rates near 100
degrees C without denaturation. The real promise for the future,
however, lies in the development of fully artificial enzymes.
Enzymes have already been designed, synthesized, and found to
function as designed. Designed enzymes that are found not to
function as intended can be modified as many times as necessary
until they function as desired. Thus, whether from first
principles or from enlightened trial and error, industrially
useful artificial enzymes should be forthcoming. More than 10100
average-sized synthetic enzymes are possible with the use of
nature's 20 amino acids, whereas probably less than 1014
enzymes are presently responsible for maintaining the entire
But there is nothing to restrict artificial enzymes to only 20 amino acids. There are many useful kinds of chemistry that are not easily promoted by natural amino acids. One particularly versatile method for transcending the biological limits of proteinaceous catalysts has already been demonstrated (ref. 1). It is based on the fact that the genetic code specifies amino acids by the sequence of any one of four nucleic acid bases taken three at a time, with each three-letter base sequence known as a codon. There are 64 codons in all, but nature uses them to specify only 20 amino acids rather than 64. Recently, it has become possible to create an artificial transfer RNA that can add an unnatural amino acid to a growing polypeptide chain in response to one of the "unused" codons that exist naturally. This could be expanded to an unlimited number of unnatural amino acids, and these unnatural amino acids could contain totally nonbiological catalytic groups or even pre-made machine parts, such as structural support struts, molecular bearings, or the like. In a fully artificial system, the 20 natural amino acids might even be entirely dispensable. Furthermore, it is now possible to insert artificial bases into DNA and RNA, drastically augmenting the prospects for designing catalytically active RNA as well as proteins with unnatural functional groups (ref. 2). The potential for programming the creation of unprecedented chemical catalysts and other molecular tools useful for molecular engineering is thus virtually open-ended.
Yet a ribosomal pathway to molecular nanotechnology is far from the only viable approach. Bulk techniques are now creating useful building blocks for molecular nanotechnology, such as carbon closed-end tubes with walls that are one atom thick (descendants of buckminsterfullerene). Cryptands or cavitands are being created that pack the catalytic punch of enzymes into nonprotein-aceous structures far smaller than natural enzymes. At the same time, the ability of the scanning tunneling microscope not only to image, but also to manipulate individual atoms and molecules is being combined with the biological specificity of antibodies in an attempt to make all-purpose molecular synthesizers similar in concept to industrial robots that now make automobiles. These present-day efforts are being supplemented by computational chemistry simulations of molecular bearings, molecular planetary gears, and molecular robot arms that would direct molecular factories.
An important premise of molecular nanotechnology is that the
products of this technology should ultimately be readily
affordable. This premise is based on the notion that an
all-purpose molecular assembler should be able to make copies of
itself. Thus, once the first programmable molecular assembler is
made, it will not be long before there are as many such
assemblers as the market needs.
Designs for computers with molecular data storage and processing elements predict data storage capacities of 1,000 megabytes per cubic micron and data-processing capabilities of 1010 operations per second for similar volumes. Translated into a sugar-cube-sized personal computer, this molecular computer would store 1020 bytes of information and process information at a speed of 1020 operations per second. This is 200 billion times the information storage capacity of today's top level personal computer and well over 100 million times the processing speed of today's fastest supercomputers.
Although it is difficult to project how many of the open-ended possibilities suggested by molecular nanotechnology will be realized by the year 2020, several innovations seem worth considering.
Quality control in the pharmaceutical industry today is
necessary due to the imprecision of the manufacturing,
purification, and packaging processes and the inability to
monitor product quality on a continuous basis. When drugs are
made by programmable molecular fabricators and when molecular
sensing devices are available, continuous sensing of the
near-flawless production process should be possible, both for
internal use and for reporting purposes. Problems could be
corrected instantly by replacing defective chemical synthesizers
with backup copies and discarding the few molecules of
mis-synthesized material before any has a chance to leave the
Thus, the products reaching the public should conform to standards with virtually exact fidelity. It is also possible that multipurpose drug synthesizers could be on hand in most clinical chemistry and toxicology laboratories and would make on-site de novo synthesis an alternative to obtaining pre-synthesized standards from bodies such as the USP. The barriers to this route could be more political than technical, but the availability of rapid local synthesis could make a life-or-death difference in some cases, such as those involving acute poisoning, for example. Alternatively, and more probably, instruments capable of identifying and quantitating drugs without calibration standards should be feasible.
Today's drug is essentially a single molecule with an often
sophisticated but always limited repertoire. Tomorrow's
"smart pharmaceuticals" could be essentially
programmable machines with a range of "sensory,"
"decision-making," and "effector"
capabilities. They might avoid side effects and allergic
reactions by coming in generic, biocompatible housings; becoming
active only upon reaching their ultimate destinations; and
attaining almost complete specificity of action. They might check
for overdosage before becoming active, thus preventing accidental
or intentional poisoning. They might have not one chemical action
but several, processing targeted invading organisms or malignant
cells through a series of chemical reactions that guarantee the
death of the target. They might work in concert with three or
four "sister" agents that together produce versatility
unattainable by one agent alone.
Despite the vast increase in complexity over present-day drugs, such agents can be expected to be totally "pure" and predictable in their behavior. Safety and efficacy may be inherent in the designs of such "drugs," in which case regulatory issues could be simplified rather than complicated.
When every intelligent person can have essentially unlimited
data storage and data-processing capability at his or her
fingertips, drug information will change dramatically.
Marked advances in diagnostic agents could have far-reaching consequences. The year 2020 will occur approximately 15 years after the complete human genome has been obtained. By 2020, great advances in understanding biochemical individuality, which is so important for side effects and proper dose adjustment, will have been made. This may make it possible for the physician to read critical aspects of a patient's phenotype from a noninvasively obtained cell sample (e.g., cheek lining cells) and to inject "smart" diagnostic agents that can be recovered in a drop of saliva some time later and read out to reveal signs of previously undetected, impending disease processes. The opportunities for precise tailoring of individual treatment and for preventive medicine, with all the cost savings implied by both, would be revolutionary.
"Drug information" could come to include the complete matrix of appropriate pharmacologic responses to the newly-available individuality data. Eventually, this will become so complex and extensive that the physician will be utterly dependent on assistance provided by computer "expert systems" in making decisions. In the long run, particularly if ingestible diagnostic agents and home readout systems become available, the need for physician participation in pharmacological therapy may vanish altogether for those patients with the right equipment available. A major incentive for this decentralization of medical care will be ethical concerns over patient privacy and the use of genotypic information for unauthorized purposes: these are issues of "drug information" that continue to redefine the term. Relevant compendia should be available on line to any interested party, particularly those whose software allows them to interface the data with a medical "expert system."
|The line between
pharmaceuticals and medical devices
will become fuzzier and fuzzier.
Information about "drugs" per se will also change dramatically. Drug documentation will come to resemble today's protein structural databases because of the potentially great complexity of "pharmaceutical agents" that differ from medical devices only in that they are injectable, and/or invisible to the naked eye. Indeed, the line between pharmaceuticals and medical devices will become fuzzier and fuzzier. On the other hand, the need to document contra- indications, side effects, and complications of use should be considerably reduced.
Clearly, the world of health care technology will be substantially different in 2020 from that of today. The U.S. Pharmacopeial Convention will likely be a much different organization on the occasion of its bicentennial anniversary. The need for pharmaceutical scrutiny by medical practitioners could be even more intense, however, as all aspects of pharmaceutical medicine and the pharmaceutical industry itself continue to change dramatically and at great speed. It can be hoped that molecular nanotechnology, in addition to helping to create these dramatic changes, will also be one of the technologies that help the practitioners of 2020 stay abreast of and manage these developments.
1. Noren, C.J., Anthony-Cahil, S.J., Griffith, M.C., and
Schultz, P.G. A general method for site-specific incorporation of
unnatural amino acids into proteins. Science 244:
182-188, 1989. [MEDLINE
2. Piccirilli, J.A., Krauch, T., Moroney, S.E., and Benner, S.A. Enzymatic incorporation of a new base pair into DNA and RNA extends the genetic alphabet. Nature 343: 33-43, 1990. [MEDLINE Abstract]
|Foresight Update 16 - Table of Contents|
Support for nanotechnology has always been strong -- perhaps
strongest -- within the computer community. The first
nanotechnology course was taught in a computer science
department, the first
conference was sponsored by the same (along with Foresight
Institute), the first PhD was granted by a computer-oriented
department (MIT's Media Lab), and the first text won the
publishing industry's "best computer science book"
A high proportion of the Foresight Institute's members are computer professionals of one flavor or another, and for years they have asked with increasing vigor "What can I do, technically, to further nanotechnology?" In response to these demands, Foresight's third research conference is especially designed to enable members of the computer community--programmers, software engineers, hardware designers, and computer scientists in general -- to move their knowledge base and, ideally, their careers toward nanotechnology. All those with any computer background are urged to attend.
The Third Foresight Conference on Molecular Nanotechnology: Computer-Aided Design of Molecular Systems will be held in Palo Alto on October 14-16, 1993. The meeting includes speakers who have made or are making the transition from computer science to nano-technology. According to conference co-chair Ralph Merkle, "The main emphasis of this conference will be on computational approaches to the development of molecular manufacturing, in particular the use of molecular modeling and the development of molecular computer-aided design (CAD) tools to speed the process of of developing molecular manufacturing systems. The conference will be valuable both for people who work professionally in computational chemistry and for people who have a background in computer science and are interested in finding out what they can do to contribute to the development of molecular manufacturing.
"There will also be a tutorial the day before the conference, so that people who have a background in computer science, and who wish to come up to speed with what is going on in the computational chemistry world, can attend the tutorial and get their feet wet in the methodologies and techniques that are commonly used."
The conference will feature fifteen or more speakers giving presentations on topics relevant to the pursuit of molecular control. We can only sketch a few of these here:
Joel Orr, Autodesk Fellow, past president of the National Computer Graphics Association, and president of the Virtual Worlds Society, will address CAD industry professionals, would-be nanotech designers, and others interested in hearing about the peculiar needs of nanotechnology with respect to CAD. In the macro and micro worlds, computer-aided design is optional: design can be done by hand. But in the nano world, CAD is essential. He will discuss:
a researcher at the University of North Carolina at Chapel Hill,
will be speaking on a subject of particular interest to two
groups of people: (1) surface scientists who are interested in
better interfaces to their instruments, and (2) builders of
virtual worlds, since the system is an example of a virtual world
applied to a scientific problem.
The system under discussion, the Nanomanipulator, is an immersive virtual-environment interface to a Scanning Tunneling Microscope (STM). A head-mounted display presents a scaled image of the surface being scanned by the STM in front of the user while a force-feedback Argone-III Remote Manipulator (ARM) allows the user to feel contours on the surface. Computer-controlled instrumentation allows the user to make bias pulses at specified locations, thus modifying the surface.
Ted Kaehler, a
computer scientist at Apple Computer, points out that we do not
know how the first assembler will be built or what exact research
is needed to get there. A person who is not a professional
chemist or materials scientist, and yet wants to be involved in
this effort, has to think about how his/her skills match the
problem. In this talk, entitled "What Can a Programmer Do to
Help Create Nanotechnology?", he discusses three efforts he
has been involved in.
The first is an efficient program to discover voids inside large molecules. Programs that search for the proper design of a large molecule need to know where the empty spaces are. The second is a project to build the "relaxation server" on the Internet. This server accepts proposed molecules (via email messages) and computes the coordinates of the atoms. The results are sent back by email. The third project is a "program" of a different sort -- a meeting group. The "Assembler Multitude," a subgroup of the local Computer Professionals for Social Responsibility chapter -- meets every other Monday night in Palo Alto and covers a wide variety of nanotechnology-related topics.
Charles Musgrave, a doctoral candidate at the California Institute of Technology, will talk about ab initio calculations for mechanochemical construction of diamondoid structures. Accurate transition state barriers for a positionally controlled reaction are necessary to both the design of the tool and the design of the synthetic process. If either of these designs is not practical, then an alternate structure is required. High level ab initio calculations are required to obtain accurate transition state structures and thus reliable mechanochemical modeling.
J. Storrs Hall, a researcher at Rutgers University, will be speaking on nanocomputing; particularly the expected developments in computer architecture that make use of reversibility to reduce heat dissipation. The techniques will be critical for nanocomputers, but are on the verge of becoming useful in VLSI, so the talk will be of interest to anyone in computer architecture as well as those studying molecular computers per se.
Markus Krummenacker, an Institute for Molecular Manufacturing researcher, will be presenting a "cavity stuffer" program which should enable the design of macromolecules the size of proteins. These macromolecules should then be easily synthesizable and should also have specifiable interface surfaces so that they can self assemble.
As at past conferences, vendors are expected to demonstrate
both hardware and software useful in nanotechnology development.
The meeting includes a reception and two luncheons to promote
interaction and the formation of new collaborations.
Submissions of abstracts for presentation at the meeting are being accepted through July 31. Papers from the meeting will be reviewed for publication in the Institute of Physics journal Nano-technology.
We at the Foresight Institute urge all computer professionals interested in nanotechnology to attend this unique meeting. Information regarding registration, proceedings, and the call for papers is available from the Foresight Institute, telephone 415-324-2490, fax 415-324-2497, email firstname.lastname@example.org.
|Foresight Update 16 - Table of Contents|
Nanotechnology Playhouse. Christopher Lampton,
1993, Waite Group Press, Corte Madera, CA, 131 pages, softcover
with disk, $23.95. An easy introduction to nanotechnology,
written by a longtime Foresight member. Heavily illustrated.
Includes an IBM-PC disk with "multimedia nanomachine
simulation." Technical accuracy: reasonably high in the
book, poor on the disk.
Unbounding the Future: the Nanotechnology Revolution. Eric Drexler, Chris Peterson, Gayle Pergamit, 1991, Quill, New York, 304 pages, softcover, $10.00. New trade paperback edition of the Morrow hardcover; available from Foresight.
From Foresight Update16, originally published 1 July 1993.
Foresight materials on the Web are ©1986–2015 Foresight Institute. All rights reserved. Legal Notices.