Lift and separate

By: David Bradley

A new type of room temperature ionic liquid could improve the extraction of polluting aromatic compounds from water without the need for volatile organic solvents, according to US chemists.
  Robin Rogers and colleagues at The University of Alabama, Tuscaloosa, have devised a new class of ionic liquids, which are almost completely non-volatile solvents. These materials can partition organic compounds between the ionic liquid and water and so allow valuable or polluting aromatics to be retrieved from contaminated industrial water without using organic solvents.
  RTILs are unusual materials in that unlike most ionic compounds they have very low melting points close to room temperature. This is a product of the high lattice energy required of the bulky ions involved to form crystals at low temperature. Being ionic, however, these liquids can dissolve a wide range of materials while their negligible vapour pressure makes them less noxious than their organic counterparts.
  Rogers' new RTIL is based on fused polycyclic N-alkylisoquinolinium cations ([Cnisoq]+), which are couple with the bis(perfluoroethylsulphonyl)imide anion ([BETI]-) to make the ionic liquid. These materials, his team has found, can be used to separate an aromatic solute, such as a chlorobenzene, from contaminated water. The aromatic compound partitions itself between the two liquids - water and the RTIL. Preliminary results suggest that this partitioning process is much enhanced over previous attempts with other ionic liquids, providing a high distribution of the aromatic into the RTIL from the water.
  The researchers suggest the great improvement they observe is due either to the hydrophobicity, water repellancy, of the isoquinolinium RTIL or the fact that there are greater interactions between the aromatic groups of the contaminant and the RTIL ions. The team is currently investigating the mechanism with a view to fine-tuning the RTILs for future applications.

Chem. Commun.; DOI: 10.1039/b109340c

Back to the main page

Programming molecules

By: David Bradley

A computer made with natural ingredients might one day run programs all by itself in the test-tube. At least that is the dream of Ehud Shapiro at the Weizmann Institute in Rehovot, Israel. He and his team have been working on a computer based on the genetic material, DNA. Some day, such a machine might be able to operate in a biochemical environment - such as the human body - through biomolecular input and output.
  DNA provides one of the most compact and efficient digital information systems we know. With just four units it can represent the ingredients and blueprint for making a microscopic algae or an elephant in a molecule-sized space. Almost a decade ago, researchers such as Leonard Adleman [[ref 1]] of the University of Southern California began to find ways to make DNA solve mathematical puzzles, such as the travelling salesman problem. Different strands of DNA represent different routes for the salesman while enzymes "ask" the question to figure out which strand represents the shortest route.
  While Adleman and others have demonstrated the possibilities, one of the obstacles to making a truly viable DNA computer is that there are several steps in the computational processes these researchers have evolved that require human intervention to kick-start various reaction steps in order to execute the program.
  Shapiro and his team have now devised a way to run a DNA program without the need for slow human steps. Their method relies on the DNA cleavage enzyme FokI and DNA ligase to act as the DNA computer hardware. FokI is one member of an unusual group of enzymes that can recognize specific DNA sequence but cut the strand non-specifically some fixed distance away from the identified sequence. The "software" for Shapiro's DNA computers is based on eight short strands of designer double-helical, or duplex, DNA.
  So, how is a program run on this computer and what puzzles can it solve?
  The programming is superficially very simple and involves choosing the appropriate software molecules. Running the program entails mixing solutions of the components in a reaction vessel with the enzymes. This molecular device, which Shapiro and colleagues describe as an automaton, then processes the input molecule through a cascade of enzyme reactions, or cycles, powered by adenosine triphosphate. An input DNA double helix with a "sticky end" on one of its strands of four bases longer than its counterpart finds its complementary strand among the eight strands of DNA representing the software and joins on to its sticky end (hybridisation).
  The enzyme T4 DNA ligase then makes the bond permanent by fusing the two sticky ends into a new double helix section. Next, the enzyme FokI binds to this hybrid DNA and cuts it in two four bases along from the join - i.e. the next symbol in the DNA strand leaving a new sticky end, which is then complemented by the next software strand - each different one representing a kind of computational command. The whole process repeats until there are no more software strands to match up with the sticky end of the growing hybrid. The resultant hybrid strand then represents the answer.
  To "print" out the answer two detector molecules are included in the mix only one of which will attach itself to the answer. Gel electrophoresis then helps the scientists reveal the result and decide whether they have a yes or no answer.
  Now, what was the question?
  A simple enough task to test the DNA computer is to ask for a DNA input encoding a list of 0s and 1s in it's sequence of bases, whether all the 0s precede all the 1s. By using software DNA strands that would complement only sticky ends representing the 0s then the answer will emerge.
  The automaton has two states, labelled S0 and S1. Initially, it is in state S0. As long as it encounters 0s it stays in S0. When is encounters a 1 it switches to state S1. In state S1, it can process additional 1s and stay in S1 but if it encounters a 0, it cannot proceed and therefore cannot reach a final state. "Hence, ending the computation in either S0 or S1 is interpreted as YES," explains Shapiro, "Getting stuck is interpreted as NO."
  A similar set up might answer other related questions such as whether a list of 0s and 1s has an even number of 1s Indeed, Shapiro's soft computer can be used to create a total of 765 software programs to answer other, seemingly trivial, questions. The trivial nature of such questions belies the real power of an automaton. "We implemented the simplest possible non-trivial computing machine," Shapiro told us, "a two-state two-symbol finite automaton, that can only answer such simple questions. The power is what it is that is answering such trivial questions." Namely, mere strands of DNA in a reaction flask!
  In a single drop of chemical computer solution, a million-million automata share the same software and run independently and in parallel. Adapting the simple questions to biochemical problems it might be possible to spot problems with a person's DNA and so treat them. "In the distant future, further developments of this machine might allow it to operate in vivo, detecting biochemical anomalies and, utilizing medical knowledge encoded in its software, decide which drug molecule to synthesize in response," says Shapiro.
  Further information about the quest to build a DNA computer can be found at

1. L.M. Adleman, Science, 1994, 266, 1021.
2. Y. Benenson et al, Nature, 2001, 414, 430.

Back to sciencebase science homepage

Please vote on this page and leave your thoughts here.