IMM Report Number 37
In conjunction with Foresight Update 51
Recent Progress: Steps Toward Nanotechnology
By Jim Lewis
“Control of a biomolecular motorpowered nanodevice with an engineered chemical switch,” by H Liu, JJ Schmidt, GD Bachand, SS Rizk, LL Looger, HW Hellinga, and CD Montemagno, of UCLA, Sandia National Laboratories, and Duke Uniersity. Nature Materials 1: 173-177 (Nov. 2002)
Reporting continuing progress in the efforts of Carlo Montemagno and his collaborators to harness a biochemical rotary motor in a hybrid biological-inorganic nano-engineered system (see Update 44), this paper presents “… the rational design, construction and analysis of a mutant F1-ATPase motor containing a metal-binding site that functions as a zinc-dependent, reversible on/off switch. Repeated cycles of zinc addition and removal by chelation result in inhibition and restoration, respectively, of both ATP hydrolysis and motor rotation of the mutant, but not of the wild-type F1 fragment. These results demonstrate the ability to engineer chemical regulation into a biomolecular motor and represent a critical step towards controlling integrated nanomechanical devices at the single-molecule level.”
Armed with an atomically detailed picture of the structure and function of the ATP-driven molecular motor and computational tools that reliably design de novo metal-binding centers in proteins of known structure, the researchers set about designing a site that would reversibly bind zinc ions. Binding the zinc causes a change in the structure of protein subunits that prevents those subunits from interacting in the way that is necessary to split the ATP molecule to provide energy for rotation. Thus binding zinc could be used to control the activity of the enzyme. To make the change reversible, the zinc can be removed by chelation with a molecule like 1,10 phenanthroline, which will sequester zinc, but would not bind magnesium ions, which are essential for the enzyme to function.
The mutant enzyme had the same ability to split ATP as did the original (or “wild-type”) enzyme when tested in the absence of zinc ions, but 60% less activity when zinc was added. To look directly at the rotation of individual motor molecules, the protein motor molecules were tethered to microspheres bound to a glass coverslip, and the ability of the motor to rotate a fluorescently labeled actin filament (a long protein molecule) measured by fluorescence microscopy. The mechanical properties of the mutant motor proteins were the same as of the wild-type motor proteins in the absence of zinc; i.e., actin rotors of the same length rotated at the same rate. Introducing zinc caused 100% of the mutant protein motors, but none of the wild-type protein motors, to stop rotating. The authors attribute the different between shutting off 100% of the rotation, but only 60% of the ATP hydrolysis, to the possibility that some ATP hydrolysis can occur without producing mechanical rotation. Calculations estimate an average value for the torque of 34 pN nm for both mutant and wild-type proteins. The authors note:
“Our method maintains the presence of fuel (ATP), does not affect the mechanical properties of the motors and does not affect the activity of other ATP-dependent enzymes. These experiments represent critical steps in the realization of logical and useful biomolecular motor-powered structures. The technology may be extended through engineering of secondary binding sites to other pre-selected ligands. Aside from use as a mechanical actuator, the introduction of a secondary mechanism to control motor functionality also may allow sensing and feedback control of the motor and associated devices.”
Several years ago Ralph Merkle considered the use of diamond-like hydrocarbon molecules as possible building blocks for developing molecular manufacturing methods (see “Molecular building blocks and development strategies for molecular nanotechnology”, available in draft form at http://www.zyvex.com/nanotech/mbb/mbb.html). The size and complexity of building blocks of this type actually available in the laboratory was substantially increased with the publication of: “Isolation and Structure of Higher Diamondoids, Nanometer-Sized Diamond Molecules,” by JE Dahl, SG Liu, and RMK Carlson of ChevronTexaco, Science 299: 96-99 (3 January 2003). The discovery in petroleum of “higher diamondoids,” clusters of 4 to 11 diamond-crystal cages fused together to give molecules of ~1 to 2 nm in size, received extensive press coverage from the standpoint that these molecules could be very useful as building blocks for nanotechnology (see Nanodot story http://nanodot.org/articles/03/02/21/0917256.shtml).
Diamondoids are hydrocarbons in which all of the carbon atoms are arranged in space in a tetrahedral framework (sp3 electron configuration) just as they are inside a diamond lattice, and with the dangling bonds terminated with hydrogen atoms, as are the surface carbon atoms in diamond. The smallest diamondoid is adamantane (C10H16), in which 10 C atoms are fused into three rigid, overlapping 6-member rings so that they maintain the same tetrahedral spatial relationships to each other that they have in a diamond crystal, with the 16 dangling bonds terminated with H atoms. The next member in the series is diamantane (C14H20), in which two adamantane diamond-crystal cages are fused along a common face to give five overlapping 6-member rings. Adding a third diamond-crystal cage gives triamantane (C18H24). With the next addition, tetramantane (C22H28), the diamond-crystal cages can have different arrangements in space, resulting in isomers of different shape and, in some cases, different chirality (mirror image forms). Beyond tetramantane there are not only isomers in which the atoms are arranged in different shapes, but the number of atoms can vary a bit depending on just how the additional diamond-crystal cages are fused together. With five diamond-crystal cages, there are nine isomers with the formula C26H32, and one isomer with the formula C25H30.
The lower diamondoids, with 1, 2, and 3 adamantane cages were previously found in petroleum and have been chemically synthesized, but the attempts to synthesize the higher diamondoids were unsuccessful. The authors isolated higher diamondoids from petroleum, heating as high as 450°C to remove non-diamondoids. After various purification by chromatography, the authors crystallized all four tetramantanes, nine pentamantanes, one hexamantane, two heptamantanes, two octamantanes, one nonamantane, one decamantane, and one undecamantane. They determined single-crystal x-ray structures for representatives from three families.
The authors found a great variety of three-dimensional molecular shapes, including rods, disks, and helices. “A series of rod-shaped higher diamondoids have long axes perpendicular to their diamond (110) lattice planes, the shortest being  tetramantane at a length of 1.0 nm… Another of the hexamantanes,  hexamantane or cyclohexamantane, is a disc-shaped molecule. Additionally, two series of screw-shaped higher diamondoids have different helical pitches and diameters and helical axes that are parallel to different diamond crystal planes … Molecules of both series are rare primary helical structures, where the helicity is inherent in the backbone of the molecule (rather than arising from steric effects, such as in the helicenes).”
The authors report preparing a number of chemically functionalized derivatives, and note “Predictable and diverse derivatizable geometries are important features for molecular self-assembly…” The authors cite advantages of these molecules for drug design and electronic properties, but they do also reference Drexler’s Nanosystems in terms of “suggesting possible applications in nanotechnology”. The wide variety of rigid, diverse shapes that can be chemically functionalized substantially expands what was available in the laboratory three years ago when Merkle first published the possible use of adamantanes as molecular building blocks.
With much progress in the development of molecular electronic devices, at least three research groups have turned to the unique molecular recognition advantages of DNA to approach the challenge of assembling such devices into functional circuits. The three papers considered below use DNA to arrange metal ions, although to very different precision and scale.
“Sequence-Specific Molecular Lithography on Single DNA Molecules” by K Keren, M Krueger, R Gilad, G Ben-Yoseph, U Sivan, and E Braun of Technion-Israel Institute of Technology, Science 297: 72-75 (5 July 2002). In a process they call molecular lithography, an Israeli group uses complexes of DNA and a protein called “RecA”, which facilitates three-way junctions with different DNA strands, as template and “resist” to build and connect metallic wires onto semiconducting surfaces. The authors had previously shown (see Update 33) that DNA molecules could be uniformly coated with metal to form thin metallic wires that could be attached to macroscopic electrodes by base-pairing between complementary DNA sequences.
RecA is the major protein that mediates the process of homologous recombination in the laboratory bacterium E. coli. In this process, many molecules of the RecA protein bind to a single strand of DNA (ssDNA). The DNA-protein complex can then bind to a dsDNA (double-helical DNA) if there is substantial homology between the ssDNA and dsDNA, such as a section of at least 30-50 bases where the ssDNA and dsDNA molecules have identical or nearly identical sequences. The ssDNA-protein complex can then bind at that region of homology, displacing one strand of the dsDNA and taking its place in the double helix. These properties of the RecA and DNA complex can thus be exploited to (1) facilitate binding of a ssDNA probe to a specific section of a long DNA molecule, (2) form a three-way junction between a ssDNA and a dsDNA molecule, and (3) protect a specific section of DNA from chemical derivatization by virtue of bound molecules of RecA protein (i.e., function as a resist). Thus RecA can be exploited to produce a network of DNA molecules that are coated with metal in some specific places.
In one experiment, the authors used RecA to place a 2027-base ssDNA on the homologous section of a 48,502-base pair dsDNA molecule (a bacterial virus) that had first been treated with aldehyde to render it susceptible to treatment with silver ions. Silver aggregates form where the DNA is unprotected by RecA, and the silver-dotted section then becomes a target for gold deposition, converting the DNA regions unprotected by RecA into a conductive gold nanowire. The DNA molecules were stretched out on chemically treated silicon wafers and imaged by AFM. The gold coated DNA molecules are 50-100 nm wide (compared to a little more than 2 nm for uncoated dsDNA) and up to a dozen microns in length. Other experiments indicate the formation of a three-way junction to build a Y-shape complex from two dsDNA molecules, and in still other experiments, the ability to attach other types of nanoparticles along a gold-DNA nanowire to nm precision. Although this method allows nm-precision in placing elements along a nanowire, and in where two nanowires connect, the wires are relatively thick (50-100 nm), microns in length, and are connected topologically, but not constrained as to their exact positions on a two dimensional surface so that the nanometer-precision only extends along the wire, not in two dimensions.
“Selfassembly of metallic nanoparticle arrays by DNA scaffolding” by S Xiao, F Liu, AE Rosen, JF Hainfeld, NC Seeman, K Musier-Forsyth, and Richard A. Kiehl of the University of Minnesota, New York University, and Brookhaven National Laboratory, Journal of Nanoparticle Research 4: 313-317, 2002.
These authors exploit the earlier work of Seeman and his colleagues (Update 35) that used rigid double-crossover DNA molecules to construct crystals capable of tiling a surface. With such programmable molecular scaffolding, they were able to assemble arrays of 1.4-nm diameter (55 Au atoms) gold nanoparticles with interparticle spacings of 4 nm and 64 nm. On the other hand, the unusual DNA molecules used for these two dimensional crystals require “aqueous environments containing significant concentrations of multivalent cations”, substantially different from the usual conditions in nanoelectronic applications. The 2-D DNA crystals assembled in solution (via base-pairing of complementary sequences on different molecules) from 22 specifically designed synthetic oligonucleotides (each a few dozen nucleotides in length), some of which were designed to have structural features protruding from the plane. The gold nanoparticles were designed to attach to those protruding structures via a 6-carbon atom linker.
The 2D DNA crystals were examined by atomic force microscopy (AFM), transmission electron microscopy (TEM), and scanning transmission electron microscopy (STEM). Gold nanoparticles were seen in arrays with the spacing expected from the crystal design —4 nm between particles in lines 64 nm apart. Gold nanoparticles were attached to the crystal array at about 70% of the 4 nm by 64 nm lattice points, indicating incomplete attachment. Thus this technology gives programmable 2-D arrays with one-base pair precision (0.34 nm) in both dimensions, although the chemistry has not been optimized and the aqueous environment might be problematic for some applications.
“A Discrete Self-Assembled Metal Array in Artificial DNA” by K Tanaka, A Tengeiji, T Kato, N Toyama, and M Shionoya of the University of Tokyo and the Institute for Molecular Science, Science 299: 1212-1213 (21 Feb. 2003).
In this case, metal ions are introduced directly into the helical duplex of a DNA-like structure. The authors had previously showed they could replace the hydrogen-bonded base pairing present in natural DNA by metal-mediated base pairing of artificial bases, leading to “metallo-DNA,” in which metal ions were lined up along the helix axis. An advantage of this approach is that they are able to use conventional DNA solid state synthesis chemistry to incorporate such metal-containing artificial DNA at any position in a conventional DNA molecule.
In these experiments, they use a hydroxypyridone nucleobase (H) that forms, with concomitant deprotonation, a stable, neutral complex with a divalent Cu2+ ion (H- Cu2+-H). They synthesized 3- to 7-nucleotide self-complementary DNA molecules of the type 5′-G HnC-3′, where n=1-5. These molecules do not form duplex structures in the absence of Cu2+ because two G-C pairs would not provide enough conventional base-pairing to form a duplex. UV-absorption changes of these molecules in solution showed that upon adding Cu2+, the molecules did form duplex structures. Circular dichroism (CD) spectra of the duplexes were consistent with the formation of right-handed, double-stranded DNA. Continuous-wave electron paramagnetic resonance (CW-EPR) spectra of the duplexes indicated the formation of magnetic chain by stacking of the Cu2+ within the DNA helix. The distance between the Cu2+ ions (0.37 nm) is similar to the distance between bases pairs in conventional DNA (0.33 to 0.34 nm).
The authors conclude “This strategy represents a method for arranging metal ions in solution in a discrete and predictable manner and contrasts with the non-biological approach of other methods, leading to the possibility of metal-based molecular devices such as molecular magnets and wires.”