This is the vision that inspires a team of Weizmann Institute scientists, who have recently made an important early step toward its realization, creating the first autonomous biological nanocomputer. Reported in Nature, as many as a trillion such computing devices can work in parallel in one drop of water in the lab of Prof. Ehud Shapiro of the Computer Science and Applied Mathematics Department and the Biological Chemistry Department. The devices collectively perform a billion operations per second with greater than 99.8% accuracy per operation while requiring less than a billionth of a watt of power.

The nanocomputer created is computationally very limited and too simple to have immediate applications, however it may pave the way to future computers with unique biological and pharmaceutical applications able to operate within the human body. 'It will probably take decades,' says Shapiro. 'But we believe this vision is realizable. The living cell contains incredible molecular machines that manipulate information-encoding molecules such as DNA and RNA in ways that are fundamentally very similar to computation. Since we don't know how to effectively modify these machines or create new ones just yet, the trick is to find naturally existing machines that, when combined, can be steered to actually compute.'

Shapiro challenged Ph.D. student Yaakov Benenson to do just that: to find a molecular realization of one of the simplest mathematical computing machines - a finite automaton that detects whether a sequence of 0's and 1's has an even number of 1's.

Biological editing kit

Benenson came up with a solution using DNA molecules and two naturally occurring DNA-manipulating enzymes: Fok-I and Ligase. Operating much like a biological editing kit, Fok-I functions as a chemical scissors, slicing DNA into a specific pattern, whereas the Ligase enzyme seals DNA molecules together.

The nanocomputer can be programmed to perform several simple tasks, depending on the software molecules mixed into the solution. The software molecules can be used to create a total of 765 software programs. Several of these programs were tested in the lab, including the 'even 1's checker,' as well as programs that check whether in a list of 0's and 1's, the 0 comes before the 1 in every case, or whether it both starts with a 0 and ends with a 1.

Four letters

How do DNA strands come to contain the symbols 0 and 1? DNA strands are usually depicted as a scroll of recurring 'letters,' in varied combinations, which represent DNA's constituents (four chemical bases known as A, T, C, and G). The nanocomputer created by Shapiro's team uses these four DNA bases to encode the input data as well as the program rules underlying the computer 'software.' The team decided that the letter pattern 'CTGGCT' in the input molecule would signify 1 and 'CGCACG' would signify 0.

In addition to having two symbols, the input molecule, when mixed with hardware and software molecules, also has two states. When the hardware molecule Fok-I, recognizes a symbol and 'cuts' DNA, it leaves it with one strand longer than the other, resulting in a single-strand overhang called a 'sticky end' (see diagram). Since Fok-I makes its incision at the site of the symbol, the sticky end is what remains of the symbol. Fok-I may leave the symbol's 'head' or 'tail' attached. These are the two possible states. A computing device that has two possible states and two possible symbols is called a two-state, two-symbol finite automaton.

Input output

Two molecules with complementary sticky ends can temporarily stick to each other (a process known as hybridization). In each processing step the input molecule hybridizes with a software molecule that has a complementary sticky end, allowing the hardware molecule Ligase to seal them together using ATP molecules as energy.

Then comes Fok-I, which cleaves the input molecule again, in a location determined by the software molecule. Thus a sticky end is again exposed, encoding the next input symbol and the next state of the computation. Once the last input symbol is processed, a sticky end encoding the final state of the computation is exposed and detected - again by hybridization and ligation - by one of two 'output display' molecules. The resulting molecule, which reports the output of the computation, is made visible to the human eye in a process known as gel electrophoresis.

Though their potential as miniature 'doctors' may be decades away from realization, these devices, if further advanced, could, be used to screen DNA libraries in vitro. This in itself could contribute to the future development of gene therapies tailored to individuals according to their genetic make-up. Both visions have a long way to go, but these small beginnings are harbingers of a much enhanced quality of life.

This study was conducted in collaboration with Prof. Zvi Livneh, Dr. Tamar-Paz Elizur, and Dr. Rivka Adar of the Weizmann Institute's Biological Chemistry Department and Prof. Ehud Keinan of the Technion Israel Institute of Technology's Chemistry Department.

Computers in the Dust

In 1954 researchers at the Weizmann Institute of Science built the first computer in Israel, and one of the first computers worldwide, which they fondly named WEIZAC. Thirty years later, silicon chip computers routinely functioned at a rate that was thousands and millions of times faster than WEIZAC.
 

Life is about decision making. From humans down to unicellular organisms it's all about striking the best deal, finding the greenest pastures or, say, choosing the quickest escape route from fire, predators, or noxious chemicals.

At the command post, on a submicron scale, lies the cell. First viewed by the human eye during the late 1600s but sporting billions of years of evolutionary polish, it easily rivals today's most advanced computers. On receiving information about its environment, the cell processes it, filters out irrelevant cues, cross-checks the extracted information with prior experience, and decides upon an appropriate 'output,' or response strategy.

Scientists have spent decades piecing together this picture of the cell's nuts and bolts. Yet its actual circuitry and time-dependent responses to environmental stimuli remain elusive. Pooling together computer science, a research tool derived from an ancient jellyfish defense strategy, and his background in theoretical physics, Dr. Uri Alon of the Weizmann Institute's Molecular Cell Biology Department is tackling these questions. His goal: to trace how gene circuitries and their encoded proteins perform computations within the cell; to find out how the cell 'thinks.'

Alon and his team, including postdoctoral fellow Michal Ronen and Ph.D. student Shiraz Kalir, created an experimental 'tracking' system capable of monitoring the simultaneous expression of multiple genes in real time. Their idea was to take a well-studied molecular pathway, such as the one found in Escherichia coli bacteria, and examine it using green florescent protein (GFP), originally isolated in the jellyfish Aequorea victoria. (See below.)

E. coli divides every 30 minutes when food is plentiful, Alon explains. But when things get rough the bacterium prepares to swim away, using a microscopic protein-based 'engine' complete with paddling flagellar appendages, a 100-rpm 'motor,' and an internal computer that directs the bacterium to food-rich areas by examining concentration gradients. To save energy this motor is produced only upon demand, when food runs low. The transformation from a stationary, dividing bacterium to a mobile E. coli takes three generations, with the cell making the initial parts of an engine that its 'grandchildren' will complete.

Using genetic engineering, Alon introduced the gene for GFP into these bacterial genes so that they glow green when expressed (i.e., when they are engaged in protein production). The result: a live system that can be monitored in real time through sequential snapshots. By developing computer algorithms that cluster this information according to the recording time, the team was then able to map the precise timeline of events controlling the bacterium's 'food-detecting' circuitry. Their study was recently published in Science.

'Scientists spent roughly three decades mapping this bacterial network using classic genetic methods, systematically knocking out gene after gene to determine its function,' says Alon. 'Our approach enabled us to reconstruct the entire system within weeks. We confirmed previous studies but also added something new: There's an exquisite synchrony between the time in which each protein is produced and where it is assembled within the bacterial engine.'

Performing with conveyor-belt precision, the bacterium produces 50 proteins divided into 14 groups. First come information processing proteins, then the intracellular structural components of the 'rotary motor,' and finally the extracellular flagellar components. Timing is key. During the second stage, for instance, seven gene groups turn on sequentially according to the motor assembly order, while the last gene activates the third stage. 'These findings highlight how nature excels at conserving energy through methodical planning,' says Alon. 'There's no point in producing a protein until its forerunner is available.'

Following their success with the E. coli model Alon and his team are moving to new protein networks, including one that switches on following DNA damage, dubbed the SOS repair system. Their immediate goal, however, is to construct a detailed map of the bacterial cell circuitry triggered in response to diverse stimuli. Success at this may vault over bacterial lines, edging toward a Holy Grail of molecular biology: 'The dream is to draw up a blueprint of the human cell, to set up a research infrastructure based on mathematical algorithms that can trace diverse protein circuitries,' says Alon.

Just as engineers use detailed diagrams of their machines to zero in on performance problems, the ability to trace protein pathways - for instance what goes wrong in a specific disease - may lead to highly selective medicines targeting a system's weak link, as well as other creative biotech applications.

Where Submarine Warfare Meets Molecular Genetics
 

The human body contains over 30,000 proteins interacting in dense networks. Figuring out how the cell orchestrates the stuff of life, the way in which it responds to environmental stimuli by playing the DNA like a keyboard to activate certain genes while suppressing others, is a hefty task.

Since 1996, when its three-dimensional structure was solved, green florescent protein (GFP) has lent an increasingly important hand, being used by laboratories across the world to track the expression of proteins in organisms ranging from bacteria and nematodes to rats and humans. It offers unique advantages for the real time exploration of molecules and organelles in live cells.
 

Originally isolated from Aequorea Victoria, a jellyfish commonly found in the Pacific Ocean, GFP protects the animal by teaming up with another protein  aequorin  that is activated by rising calcium levels in response to a perceived threat. Aequorin emits a blue light which is then 'edited' by GFP to produce a bright green flash that presumably scares off the attacker. Ships and submarines also trigger this unique defense mechanism in the jellyfish, resulting in a glowing trail of light that marks the ship's course  which explains the U.S. military's longstanding interest in the jellyfish.

Dr. Uri Alon holds the Carl and Frances Korn Career Development Chair in the Life Sciences. His research is supported by the Dr. Ernst Nathan Fund for Biomedical Research, the Mrs. Harry M. Ringel Memorial Foundation, the Cherpak-Vered Visiting Fellowship, Yad Hanadiv, and the Minerva Stiftung Gesellschaft fuer die Forschung m.b.H.

The smaller an object, the larger the equipment needed to measure it. This rule of thumb comes to mind when looking up at the daunting nuclear magnetic resonance (NMR) machinery in Prof. Jacob Anglister's laboratory. A member of the Department of Structural Biology at the Weizmann Institute of Science, Anglister uses NMR to probe how molecules in the cell interact. 'Numerous biological processes, including protein production, cellular division, and immune response, depend on two molecules recognizing and responding to each other,' Anglister explains. 'NMR enables us to study this key recognition aspect at the atomic level. For instance, we can capture a three-dimensional representation of the complex formed when two proteins meet.'

Examples of the protein one-on-ones that have grabbed Anglister's attention include what happens when antibodies, the immune system's 'infantry' soldiers, recognize and attack invading proteins; how scorpion neurotoxins bind to neuronal membranes; and how interferon alerts the body to use its antiviral and anticancer weapons. His team is currently focusing on antibodies against the HIV (human immunodeficiency virus) that causes AIDS.

During the initial stage of infection, a protein on the envelope of the HIV virus known as gp120 binds to two different immune cells - T-cells and macrophages. This bonding triggers conformational changes in gp120 that mediate the fusion between the virus and its target cells.

Hide and seek

Using NMR, Anglister has revealed the complex formed between the HIV virus and the body's immune cells, and honed in on a region in gp120, called the V3-loop. This loop plays a key role in binding the HIV virus to its target cells - indeed, the immune system generates antibodies whose sole function is to bind with the V3-loop, thus neutralizing the virus. The HIV virus, however, is able to overcome this defense strategy by rapidly mutating the V3-loop to escape detection. Another trick, used by what is perhaps the nastiest strain of this virus, is to hide the V3-loop until the very moment it is needed to bind the virus to its target, so that no immune antibodies will be produced.

A key puzzle to researchers is how the virus is able to use the V3-loop on its envelope to bind to target cells yet at the same time mutate this loop to escape detection. Having uncovered the detailed structure of the V3-loop, Anglister's team may now have a partial answer. In what appears to be an ingenious decoy approach, the DNA encoding the V3-loop consists of two primary types of sites - evolutionarily conserved sites (which suggests they are highly important for viral replication) and rapidly mutating sites that 'cover'/protect the conserved region from attack. Since the protein segments encoded for by the mutating regions are highly exposed on the virus envelope, they quickly draw the immune cells' fire. But because these regions mutate so quickly, the body is unable to mount an effective response. Furthermore, even when the immune cells succeed in generating antibodies against the conserved protein segments, the mutating regions obstruct their attack.

A peptide-based vaccine may offer a partial solution, as demonstrated by Institute Profs. Michael Sela and Ruth Arnon, who pioneered the application of synthetic vaccines. They showed that synthetic peptides could be used to generate a protective immune response against various pathogens. This approach may be especially beneficial in the case of HIV, as it can trigger an immune response against viral regions that, like the V3-loop, are generally hidden. A cocktail of peptides found in the protein sequences of major HIV strains may help researchers target this elusive virus.

Anglister's team is currently studying an antibody, isolated from an HIV patient by an American research group, which is capable of neutralizing various HIV viral strains that have undergone V3-loop mutations. 'Our aim is to use NMR to discover the secret of this antibody's success,' says Anglister. 'Such understanding may help in designing an optimal peptide-based HIV vaccine.'

Nuclear magnetic resonance (NMR)

NMR makes possible the study of protein molecules in solution. It can also provide information about dynamic processes in the body occurring at a rate of up to a millionth of a second.

Scientists had traditionally relied on X-ray crystallography to elucidate protein structures. However, when using crystallography one has first to create crystals of the protein in question - an often difficult process.

Additionally, crystallography cannot be performed on solutions, where biological molecules exhibit their activity. The first three-dimensional NMR structure of a protein was solved in 1985 by Kurt Wuthrich and his co-workers at the Swiss Federal Institute of Technology (ETH) in Zurich.

In using NMR to probe the structure of protein molecules, a small amount of a solution containing the protein is placed in a magnetic field and an electro-magnetic current is applied at varying frequencies, thus selectively activating the nuclei in the atoms making up the protein molecule.

Since each nucleus -- say that of a carbon or hydrogen atom -- responds with its own unique spin, scientists can use sensitive recording equipment to detect the resultant magnetic fields. They then use this information to construct a three-dimensional picture of the entire protein molecule.

The first NMR application was developed back in the 1940s by Felix Bloch of Stanford University and Harvard's Edward Purcell, earning them the 1952 Nobel Prize in Physics. Weizmann Institute scientists Shlomo Alexander and Shaul Meiboom built the world's second high-resolution NMR spectrometer in the 1950s. The system was used to pioneer the research of chemical reactions in equilibrium, and to explore the structure and behavior of molecules. Later Meiboom, Alexander, and other Institute scientists, using nuclear quadropolic resonance, developed a method of measuring the movement of molecules in crystals.

Prof. Anglister holds the Joseph and Ruth Owades Professorial Chair in Chemistry. His research is supported by the Burstein Family Foundation, the Minerva Stiftung Gesellschaft fuer die Forschung m.b.H., the Nelfort Foundation and Mr. and Mrs. Samy Cohn.