NATURE BIOTECHNOLOGY VOLUME 26 NUMBER 3 MARCH 2008 299

potentially valuable alcohols, including iso-butanol (from 2-ketoisovalerate in valine biosynthesis), 3-methyl-1-butanol (from 2-keto-4-methylpentanoate in leucine bio-synthesis), 1-butanol (from 2-ketovalerate in norvaline biosynthesis), 1-propanol and 2-methyl-1-butanol (from 2-ketobutyrate and 2-keto-3-methylvalerate, respectively, in isoleucine biosynthesis), and 2-phenyletha-nol (from phenylpyruvate in phenylalanine biosynthesis). Further pathway optimization will be necessary to produce these alcohols at the very low costs needed to compete with petroleum-based fuels.

Several other groups have engineered microorganisms to produce next-generation biofuels that may not be produced naturally by any microorganism or are produced only in very low quantities2–4. In general, this strategy allows renewable alternatives to fossil fuels to be selected based not on their availability but on their compatibility with the existing transportation infrastructure. Recent advances in synthetic biology and metabolic engineering5 suggest that, rather than limiting ourselves to fuel molecules provided by nature, we should engineer microorganisms to produce new fossil-fuel replacements. Such products, which might include short-chain, branched-chain and cyclic alcohols, as well as alkanes, alkenes, esters and aromatics, should not only be compatible with existing technology but might also be optimized for applications that extend beyond powering automobiles. Like current petroleum-based fuels, these renew-able fuels need not be restricted by a ‘one size fits all’ mentality and could be tailored for specific applications, climates and geo-graphic locations. Furthermore, the ability to design novel fuel molecules opens the door for fuel technologies to evolve in tandem with transportation technologies.

To produce longer-chain alcohols and alkanes, it should be possible to tap into the fatty acid pools of nearly any organism. The recent identification of a locus that controls oil content in maize is promising in this regard6. Sequential reduction, decarboxyl-ation or decarbonylation followed by reduc-tion of fatty acids to alcohols and alkanes could yield valuable fuel candidates. It is also possible to esterify fatty acids with alcohols from any number of sources to produce can-didate biodiesels4.

Isoprenoid biosynthesis offers an even richer source of next-generation biofuels. With the ability to produce branched-chain and cyclic alkanes, alkenes and alcohols5 of different sizes with diverse structural and chemical properties, this pathway could pro-

duce fuels or precursors to gasoline, diesel and jet fuel additives or substitutes. Efficient production of isoprenoid precursors has been engineered in E. coli and Saccharomyces cerevisiae, and many different isoprenoids have been produced using these engineered hosts7,8.

Although production of next-generation biofuels is an important endpoint, it will be critical to first tap into sugars, the most inex-pensive starting materials at our disposal, to make these new fuels economically viable. For instance, taking advantage of the existing US corn supply by using starch as a feedstock seems the most realistic short-term goal. This will strengthen both the nascent bio-fuels industry and the agriculture industry. However, in the long term, these fuels will certainly be produced from less expensive

sugars, such as those released from cellulose depolymerization.

COMPETING FINANCIAL INTERESTSThe authors declare competing financial interests: details accompany the full-text HTML version of the paper at http://www.nature.com/naturebiotechnology/.

1. Atsumi, S., Hanai, T. & Liao, J.C. Nature 451, 86–89 (2008).

2. Inui, M. et al. Appl. Microbiol. Biotechnol. 77, 1305–1316 (2008).

3. Kalscheuer, R., Stolting, T. & Steinbuchel, A. Microbiology 152, 2529–2536 (2006).

4. Withers, S.T., Gottlieb, S.S., Lieu, B., Newman, J.D. & Keasling, J.D. Appl. Environ. Microbiol. 73, 6277–6283 (2007).

5. Keasling, J.D. ACS Chem. Biol. 3, 64–76 (2008).6. Zheng, P. et al. Nat. Genet., advance online publica-

tion; doi:10.1038/ng.85 (17 February 2008).7. Martin, V.J.J., Pitera, D.J., Withers, S.T., Newman,

J.D. & Keasling, J.D. Nat. Biotechnol. 21, 796–802 (2003).

8. Ro, D.K. et al. Nature 440, 940–943 (2006).

Tiny tiles, tiny targetsDavid A Giljohann & Chad A Mirkin

A new label-free method for RNA detection uses programmable DNA tiles and atomic force microscopy.

Detection of nucleic acids is routinely carried out using enzyme-based techniques such as PCR. This is likely to change, however, with the development of enzyme-free detection approaches that still allow high sensitivity and specificity but also offer massive multi-plexing capabilities, lower costs and greater assay stability. An interesting example of this new direction is recent work by Ke et al.1 published in Science. Using a form of programmable DNA folding recently intro-duced by Rothemund2, the authors create self-assembled, two-dimensional nucleic acid ‘tiles’. These structures not only bind complementary sequences, but also provide positional and structural information based upon the type of tile and the location of the capture strand used for detection, allow-ing the results of RNA hybridization to be imaged by atomic force microscopy.

The advance builds on the progress made over the past two decades in using DNA as

a synthon in macroscopic materials design. Two major synthetic strategies have been developed. In one approach, DNA is used as the blueprint, assembler and core construc-tion material for making higher-ordered architectures3,4. In the other, non–nucleic acid building blocks are functionalized with DNA, which acts as a surface director and assembler to generate materials that are highly ordered and interconnected with, but not predominately made of, nucleic acids5,6. The latter approach allows one to assemble low-cost, highly functional build-ing blocks into periodic architectures and has led to several fundamentally interesting structures and technologically useful mate-rials and devices7,8, including US Food and Drug Administration–approved nucleic acid detection systems8.

The work by Ke et al.1 is a first step toward evaluating the potential of materials made solely of DNA in the context of biodiagnos-tics. The authors use self-assembled tiles of DNA deposited onto a mica surface as struc-tures for RNA detection (Fig. 1). The tiles are held together by ‘stapled sites’ with local intramolecular complementarity within the extended structure. When an RNA sequence complementary to a non-stapled location on the tile is added, Watson-Crick base pairing

David A. Giljohann and Chad A. Mirkin are in the Department of Chemistry and International Institute for Nanotechnology, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208-3113, USA. e-mail: chadnano@northwestern.edu

NEWS AND V IEWS©

2008

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

ebio

tech

nolo

gy

300 VOLUME 26 NUMBER 3 MARCH 2008 NATURE BIOTECHNOLOGY

between the RNA target and DNA tile leads to formation of a hybridized structure. These binding events can be detected by atomic force microscopy even at high picomolar concentrations of RNA or in complex envi-ronments such as cellular extracts.

One potential problem for this technology is creating a method of deconvoluting the ran-dom orientation of the tiled structures that is

apparent from the atomic force microscopy imaging. The authors have overcome this problem by creating a barcode system that not only allows the orientation of the tiles to be determined but also makes the method amenable to multiplexing. In this case, the researchers discovered that the location of the target capture strand on the tile significantly affects its binding properties. However, by assigning tiles with different targets a unique barcode and localizing the target capture strand at the same physical site on different tiles, they could reliably identify multiple tar-gets in one experiment.

This demonstration of atomic force micro-scopy detection of a structure initially formed in solution is qualitatively impressive. However, significant work remains to be done to simplify the readout and determine the assay’s quanti-tative capabilities before it can be considered for practical application. Theoretically, direct imaging of single binding events should yield a one-to-one correlation between a detected sig-nal and copy number (that is, each tile should be capable of binding one target). However, the technical challenge of imaging each tile with atomic force microscopy in a serial man-ner is daunting, making target quantification extremely difficult. Furthermore, the authors have yet to reach limits of detection that com-pete with microarrays and more conventional diagnostic platforms owing to problems with surface deposition. This shortcoming might

Figure 1 Method for label-free detection of RNA using an atomic force microscope and self-assembled DNA structures. (a) Long DNA strands are assembled in solution to form nanostructured ‘tile’ probes to which RNA-containing samples and reference barcode sequences are added. (b) After hybridization, tiles are deposited on a mica surface. (c) Atomic force microscopy is used to image the tiles and provide a readout of target binding.

+ RNA

+ DNA barcodes

a

b

c

be alleviated with techniques such as dip pen nanolithography that provide the needed spa-tial resolution and registry required to deposit small numbers of the tiles in a defined area for imaging9.

Perhaps the most important aspect of this advance is the use of structural information generated through the formation of the tiles to not only create unique labels but also tar-get binding sites spatially defined within the tile structure on the nanometer length scale. DNA is a synthetically programmable material that can both form two-dimensional periodic structures and be used to control other inter-molecular interactions. For example, the same group has recently published work showing that when the tile sequences are DNA aptam-ers, the tiles can bind proteins10. Indeed, tiles could be used to define where specific proteins or other recognition elements are located with nanoscale resolution. Further progress in these directions is likely to generate new and exciting applications in studying more complex multi-valent interactions involving multicomponent tile structures.

COMPETING FINANCIAL INTERESTSThe authors declare competing financial interests: details accompany the full-text HTML version of the paper at http://www.nature.com/naturebiotechnology/.

1. Ke, Y.G., Lindsay, S., Chang, Y., Liu, Y. & Yan, H. Science 319, 180–183 (2008).

2. Rothemund, P.W.K. Nature 440, 297–302 (2006).3. Kallenbach, N.R., Ma, R.I. & Seeman, N.C. Nature 305,

829–831 (1983).4. Yan, H., Feng, L.P., LaBean, T.H. & Reif, J.H. J. Am.

Chem. Soc. 125, 14246–14247 (2003).5. Alivisatos, A.P. et al. Nature 382, 609–611 (1996).6. Mirkin, C.A., Letsinger, R.L., Mucic, R.C. & Storhoff, J.J.

Nature 382, 607–609 (1996).7. Nam, J.M., Thaxton, C.S. & Mirkin, C.A. Science 301,

1884–1886 (2003).8. Taton, T.A., Mirkin, C.A. & Letsinger, R.L. Science 289,

1757–1760 (2000).9. Piner, R.D., Zhu, J., Xu, F., Hong, S.H. & Mirkin, C.A.

Science 283, 661–663 (1999).10. Chhabra, R. et al. J. Am. Chem. Soc. 129, 10304–

10305 (2007).

Kim

Cae

sar

NEWS AND V IEWS©

2008

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

ebio

tech

nolo

gy