91av

Milwaukee, Wisconsin

FIDDLY is not the word. Moving around tiny pieces of matter containing just a
few hundred atoms is a painstakingly delicate process. Accurately placing them
would challenge even the steadiest of hands. And building them into working
devices or intricate materials sounds like a nightmare. But in the quest to make
things smaller, chemists are now tackling this problem head-on, and the latest
recruit to the miniature construction industry is the molecule of life
颈迟蝉别濒蹿鈥挤狈础.

Nanotechnology, the science of making ultrasmall components and devices,
could transform our lives with faster, more efficient computers and
machines鈥攕uch as powerful processors that will fit inside a wristwatch, or
tiny therapeutic robots that will roam our bodies, keeping our arteries clear
and our blood clean
(鈥淚nvasion of the micromachines鈥�, 91av, 29 June 1996, p 28).
So far, producing tiny gadgets or circuits has relied on the
鈥渢op-down鈥� approach, etching the devices out of a larger block of material. As
scientists strive for smaller, more intricate machines, however, etching simply
isn鈥檛 accurate enough. A 鈥渂ottom up鈥� approach鈥攂uilding from smaller
particles鈥攃ould be the way forward. And it also opens up the possibility
of the devices assembling themselves. To make an electronic circuit, for
example, all you would do is tip the ingredients into a beaker of water and give
it a good stir.

Taking their cues from the ultimate self-assembling machines鈥攍iving
organisms鈥攔esearchers at the University of California at Berkeley and
Northwestern University in Evanston, Illinois, are using strands of DNA to
create nanoscale constructions. Paul Alivisatos at Berkeley is hoping to create
ultrasmall electronic circuits that self-assemble in a beaker on scaffolds of
DNA. While at Northwestern, Chad Mirkin wants to draft in DNA to create
materials with useful electronic and optical properties. And as a spin-off he
and his collaborator, Robert Letsinger, are using the technology to develop a
simple lab test for detecting target sequences of DNA, such as the genes
responsible for cancers or disease organisms. Mirkin foresees great things for
nanoconstruction. 鈥淚f you can control the chemical composition of particles, the
size of the particles, the distances between the particles, and the strength of
the interactions holding the particles together,鈥� he says, 鈥渋n principle you
have control of every property of a material.鈥�

DNA is the perfect molecule to exercise this control. It can carry a huge
amount of information. In humans, DNA鈥檚 code packs an estimated 100 000 genes
into every cell, which determine, among other things, how our bodies are put
together. For nanoconstruction, the same code can be used to direct the assembly
of tailor-made materials and devices. Strands of DNA can in effect be employed
as nanoconstruction workers.

The molecule has two important features. First, it is enormously variable.
DNA is made from two strands of chemical building blocks called nucleotides,
that bind together to form its famous double helix structure. There are four
kinds of nucleotide, each carrying a different chemical group or 鈥渂ase鈥�. There
may be only four bases鈥攁denine (A), thymine (T), guanine (G) and cytosine
(C)鈥攂ut the number of ways in which they can be arranged in sequence is
enormous. A sequence of eight bases can be shuffled into 65 536
combinations.

Equally important, DNA is reliable. The bases pair up predictably: A always
binds to T, and G always binds to C. So the sequence of bases on one strand of
the double helix is complementary to the sequence of bases on the other strand.
For example, the sequence ATG is complementary to TAC. Strands of DNA only bind
together fully if they have complementary sequences. 鈥淒NA has the most
predictable and programmable intermolecular interactions of anything out there,鈥�
says Ned Seeman, a crystallographer at New York University who builds
three-dimensional structures with DNA. 鈥淎ny
kindergartener can come in and say A pairs with T, G pairs with C. Those are the
rules. They鈥檙e not hard.鈥�

Getting hold of the new nanoconstruction workers couldn鈥檛 be easier either.
Laboratory DNA synthesisers can rapidly make single-stranded DNA with a specific
sequence of bases. Design pairs of complementary strands and attach them to
separate tiny particles, and you have a way of selectively joining the
particles. Under the right conditions, complementary strands will 鈥渮ip鈥�
together. It鈥檚 a simple but potentially very powerful tool. 鈥淎s a way of adding
one more degree of freedom to assembling very small things, it鈥檚 the only way I
know of doing it if I wanted to put more than two types of things together,鈥�
says George Whitesides, a chemist at Harvard University in Cambridge,
Massachusetts, who specialises in self-assembled monolayers of organic
molecules.

Studded chains

The Berkeley and Northwestern teams use different methods to bring their
nanoparticles together. At Berkeley, Alivisatos and his colleague Peter Schultz
attach DNA 鈥渃odons鈥�, made from sequences of 18 bases, to gold particles 1.4
nanometres wide. To stick the gold particles to the codons they use a class of
organic molecules called alkanethiols. Alkanethiols bind strongly to gold and
DNA synthesisers can be programmed to add one to the end of a strand of DNA.
Alivisatos and Schultz attach a single codon to each nanoparticle. They then add
longer single strands of DNA that contain complementary sequences to those in
the codons. These longer strands act as backbones, binding to several codons and
creating DNA chains studded with two or three gold particles
(see
Diagram).

At Northwestern, Mirkin and Letsinger are building macroscopic structures,
designed at the nano-scale using DNA and particle building blocks. They take a
solution of gold particles 13 nanometres in diameter and attach between 10 and
100 codons to each, making them look like hairy spheres. These are then linked
together by DNA 鈥済irders鈥� to form a network of nanoparticles. Mirkin and
Letsinger first make two different codons, each eight bases long鈥攖he
minimum length needed to achieve stable bonds. They then attach these two codons
to different groups of gold nanoparticles, so that any one nanoparticle is
covered by only one type of codon.

Exclusive attachments

Finally they add a DNA girder to link the particles. The girder is made from
two strands of DNA that form a double helix along most of their length. However,
at either end of the helix an extra eight bases dangle from one of the strands.
These sequences are complementary to the codons attached to the nanoparticles.
The first codon attaches exclusively to one end of the girder, the second codon
to the other, and so the girder links together two nanoparticles
(see
Diagram).
Mirkin and Letsinger believe this method will give exquisite control
over which particles link to which, allowing them to tailor the optical,
mechanical and electronic properties of the material they produce.

When the particles are brought within one particle-diameter of each other in
the linking process, something important happens鈥攖he solution changes
colour from red to purple. This is because the nanoparticles start to interact
with each other and absorb light of a different wavelength. Normally, mutual
electrical repulsion stops gold nanoparticles in water from aggregating into
clumps. But this electrical repulsion is overcome by the girders binding the
particles together.

The scientists are now using this colour change to develop a test for the
presence of specific sequences of DNA, such as those in genes that increase the
risk of certain cancers, or an infectious agent such as HIV. First they pick a
sequence they want to test for鈥攕ay, one that marks a bacterium as a strain
that causes food poisoning. They then construct two codons that are each
complementary to one half of the sequence, and attach these to gold
nanoparticles. When a solution of these particles detects the specific sequence
of DNA, the codons attach themselves and the particles are brought together. As
a result the solution changes colour. 鈥淚n principle this can detect any type of
sequence associated with a pathogen,鈥� says Mirkin.

Mirkin says he hasn鈥檛 pitted his gold 鈥渘anoprobes鈥� against existing DNA
tests, but is confident they will be able to compete. 鈥淭he technique is
incredibly simple, incredibly cheap, incredibly fast and, as it turns out,
incredibly selective,鈥� he says. In other words, the test can distinguish between
the target sequence and one very similar, which is critical for preventing false
positives.

To confirm a positive result, the test includes a heating step that breaks
the bonds between nanoprobes and targets. At a certain temperature, the gold
particles are released from the DNA targets, and the solution changes colour
from purple back to red. The scientists know the precise temperature at which a
perfect DNA match鈥攁 positive result鈥攂reaks down, so if the colour
change happens at a lower temperature, there must be a mismatch of bases between
the nanoprobes and targets.

Other DNA tests routinely pick out imperfect matches in this way, but Mirkin
says the new test makes it easier to tell the exact temperature at which the
probes and targets come apart. The colour change is very sudden as it only
occurs when the last few DNA connections are broken. In existing DNA tests, the
change that signals DNA break-up鈥攁 gradual rise in absorption of
ultraviolet light鈥攐ccurs over a much wider temperature range. So the
results are potentially less accurate.

In the real world, the nanoparticles will have to work in body fluids and
tissue in which other ingredients could make them clump together and give a
false positive. 鈥淲e鈥檙e finding out that it actually works in some complex
environments, like urine and saliva,鈥� Mirkin says. 鈥淏ut we also need to know if
it works in [others such as] blood serum.鈥� If successful, Mirkin and Letsinger
hope to launch the test onto the DNA diagnostics market.

DNA nanoconstruction may become an even more lucrative business than
diagnostics. 鈥淭he potential is enormous,鈥� says Mirkin. So now that the basic
principle of joining matter together using DNA has been proved, the researchers
are busy thinking how they can put it to work.

Guiding light

Mirkin plans to use DNA to build an intricate, multilayered array of
nanoparticles called a superlattice. By adjusting the size and type of particles
as well as the length of the DNA girders (which fix the distance between the
particles), he believes he can 鈥渢une鈥� the superlattice to selectively send light
of a particular wavelength down a particular path. The material would block out
light with wavelengths longer than the distance between particles, and the
particles themselves could filter out certain wavelengths depending on their
composition. The physical arrangement of the particles would create a path down
which the remaining light could be channelled. Such materials would be ideal for
optical computers of the future.

The idea is to use gangs of DNA nanoconstruction workers to assemble each
layer of nanoparticles in the superlattice, like builders laying bricks. When
the particles are all in place, Mirkin will simply wash the workers away. For
this type of task, he thinks DNA has an advantage over existing chemical
linkers鈥攊ts amazing versatility. A DNA synthesiser can make any sequence
of bases of any length. What鈥檚 more, says Mirkin, conventional linkers form
strong bonds between particles, so if you botch the experiment you may be stuck
with a disorderly superlattice. By contrast, DNA zips together using weaker
bonds that can be pulled apart if things go wrong.

At Berkeley, Alivisatos and his colleagues are adding new kinds of
nanoconstruction workers to their workforce by linking DNA codons to
nanoparticles of insulating, semiconducting and conducting materials. The next
step is to string the particles along a DNA backbone, and ultimately they want
to arrange the nanoparticles in specific patterns to create electronic circuits
and devices.

鈥淒NA potentially allows you to place nanoparticles in very intricate
sequences,鈥� says Alivisatos, 鈥渘ot just periodic structures, but designed
sequences.鈥� By placing codon-complementary sequences of bases at specific points
on a DNA framework, the codons would direct each insulator, semiconductor and
conductor nanoparticle to its correct place. The underlying DNA framework would
act as a template for the circuit.

At present, tiny circuits are made using photolithography. This etches the
patterns of circuit elements onto a piece of silicon using masks as guides. But
photolithography can only make wires, transistors and other circuit elements
down to a size of about 300 nanometres. Alivisatos thinks DNA nanoconstruction
could go as small as 30 nanometres. And the fabrication equipment would be
simpler. 鈥淭he idea is that everything would be done in a beaker,鈥� he says.

Back in 1987, Seeman and Bruce Robinson of the University of Washington in
Seattle proposed a way of building a nanometre-scale memory chip using DNA
scaffolds. The scaffolds would self-assemble into a framework containing places
for storing and accessing bits of digital data. Information would be stored on
charged or uncharged clusters of metal atoms, corresponding to the 鈥�1鈥� or 鈥�0鈥� of
binary code. Seeman and Robinson estimated that a 鈥渕emory block鈥� measuring just
1 cubic centimetre could store all the information in every book ever
written.

Seeman鈥檚 DNA scaffolds could be used as templates for self-assembled
circuits. 鈥淚t鈥檚 certainly conceivable that all this could be done,鈥� says Seeman.
However, he adds, figuring out how to use a DNA scaffold to build nanocircuits
before anyone has made one 鈥渋s a little bit pie-in-the-sky鈥�.

Alivisatos has no illusions about how much progress must be made before
nanocircuits can be built in a beaker. 鈥淭o put it mildly, it鈥檚 speculative. But
it鈥檚 not totally crazy,鈥� he says. 鈥淲e can make little semiconductor, metal or
insulator [nano]particles out of the same things that we make circuits out of
today. But what we don鈥檛 know how to do is assemble them all together into a
circuit. That鈥檚 going to take time to develop.鈥�

But this does not dampen Mirkin鈥檚 enthusiasm for DNA nanoconstruction. 鈥淭here
is almost an infinite number of ways to assemble [nano]particles when you take
into account all the materials available,鈥� he says. 鈥淎nd that means there is an
infinite number of new materials available.鈥� They already have one for DNA
diagnostics and there may be more around the corner. The foundations are in
place for DNA nanoconstruction workers to build a very small, very bright
future.