Dan Luo - The Road From Biology To Materials

The road from
biology to materials
The study of biology at the molecular level may point to new directions in materials design and construction not just mimicking biomolecular systems, but actually using biomolecules themselves to construct novel materials. Indeed, by taking this new fabrication route, significant and path-breaking achievements may be made by materials scientists and engineers.
basic molecular biology to materials science and engineering (MSE), so that more and more engineers can begin to integrate molecular biology into their work. Achievements of DNA- and protein-based engineering will not be reviewed. Rather, readers are directed to two recent reviews1,2. Instead, I will focus on examples of the basic, as well as practical, knowledge of molecular biology that might be useful to material scientists. Emphasis will be placed on materials building blocks and molecular biology tool kits that can be employed for new materials design and synthesis.
by Dan Luo
This article is intended to serve as a roadmap, linking
Why molecular biology?

Fundamental differences exist between the materials built by nature and those built by materials scientists. For biologists, billions of years of evolution have constructed the most amazing material living organisms that, surprisingly, are made of only a few major macromolecules from a limited number of building blocks. On the other hand, over a few hundred years, scientists and engineers have also fabricated amazing materials from plastic to plasma that, unsurprisingly, are made from many more different kinds of building blocks. The separation between these two fields, for the most part, is because of a lack of communication. It is time for engineers to learn from molecular biology. As a matter of fact, the commonalities between molecular biology and MSE are much greater than many of us might have thought we all, in one way or another, build materials
Department of Biological and Environmental Engineering Cornell University 202 Riley-Robb Hall Ithaca, NY 14853-5701, USA E-mail: dl79@cornell.edu
38
November 2003
ISSN:1369 7021 Elsevier Ltd 2003
INSIGHT FEATURE
at the molecular level. Convergence of molecular biology and MSE will provide new ideas, directions, and potential. Engineers are increasingly embracing the idea of using biological molecules as building blocks in materials design and construction. Similarly, more biologists are looking to MSE for answers in their pursuits. So if one of us is asked, Do we need to know molecular biology? the answer is, maybe not in the past, but certainly now, and even more so in the future. The road between molecular biology and MSE will become a heavily traveled, two-way highway.
DNA properties
DNA molecules can be single stranded (ssDNA) or double stranded (dsDNA). Each strand of DNA is made of a backbone plus sequences of four bases: adenine (A), guanine (G), cytosine (C), and thymine (T). The backbone consists of fivecarbon sugar molecules (pentoses, in this case deoxyribose) linked together via phosphodiester bonds. This linkage gives all DNA strands a chemical polarity. One end terminates in a phosphate group (called the 5 end because of the position of phosphate group on the 5 carbon of the sugar ring), and the other end in a sugar molecule whose 3 carbon has a hydroxy group (called the 3 end). All DNA sequences are written by default from 5 to 3 unless specially noted otherwise. Both 5 phosphate and 3 hydroxy groups are useful in manipulating DNA, as we shall see below. dsDNA the default format in most living organisms consists of two ssDNA strands paired together in opposite directions (one from 5 to 3 and the other from 3 to 5). These two strands are called complementary. The specific and noncovalent pairing A with T and G with C makes DNA self-programmable, a property that is extremely useful in constructing novel materials. In addition, the pairing is reversible: dsDNA can be melted into two ssDNA molecules by heating or at high pH, a process called denaturing. Note that G-C pairing is more stable than A-T pairing because more hydrogen bonds are involved. Consequently, DNA strands with more GC content are less likely to be denatured, another property of which advantage can be taken. For analogy, one may imagine the pairing of two complementary ssDNA strands as differently colored Velcro pads. Only complementary colored pads can be stuck together (e.g. blue with orange, red with green) and sticking is reversible.
A very brief introduction

Molecular biology progressed rapidly after the structure of DNA was determined in 19533. This pace became faster and its influence broader after the invention of the polymerase chain reaction (PCR)4, an amplification technique that exponentially increases the amount of specific DNA molecules available for experiments. With the decoding of the human genome (meaning the entire genetic information), we are officially in the postgenomic era. Essentially, molecular biology tries to understand life at the molecular level. Thus, depending on the questions to be answered and the problems to be solved, molecular biology may be applied to almost all existing biological disciplines, for example molecular immunology arises from immunology. Similarly, you have molecular genetics, molecular evolution, molecular medicine, molecular therapy, etc. The applications of molecular biology can also be extended to nonbiological disciplines such as MSE. This trend has been greatly enhanced with the continuing progress of nanoscience and technology. It is only a matter of time before we see DNA and proteins in daily materials and devices.
Biomacromolecules in living organisms

The four major biomacromolecules are nucleic acids, proteins, lipids, and polysaccharides. Their building blocks are nucleotides, amino acids, fatty acids, and sugars, respectively. Nucleic acids and proteins are the biomolecules that have been studied and are understood the most and, hence, possess the greatest potential for MSE applications. Only DNA is discussed here. An attempt to summarize the current status of protein-based materials as well would make this short article incomplete and misleading. The following references from the last two years are intended to provide some pointers to interested readers2,5-17.
DNA as a polymeric material

DNA can be viewed as a copolymer in which distribution of repeat units (A, T, C, and G) is random, although naturally occurring DNA molecules are seldom synthesized randomly. From the engineering perspective, the meaning of the genetic codes embedded in the sequences can be totally ignored, although the chemical content of the sequences (GC%, for example) will still influence the properties of DNA structures. The structure of dsDNA is a right-handed, double helix, a symbol almost as famous as the atomic orbit logo that has often represented science in the last century. The discovery of the DNA structure 50 years ago3 has revolutionized
November 2003
39
INSIGHT FEATURE
science and society. Most of the impact has relied on DNAs genetic properties, which are solely determined by the DNA sequence (the linear series of A, G, C, and T bases). However, DNA molecules also possess physical and chemical properties that make them an ideal substance for building materials. Mechanically, DNA can be both flexible and rigid (when molecules are shorter than their persistence length18-20). The persistence length is about 50 nm for dsDNA (which roughly corresponds to 150 base pairs or bp) and a few nanometers for ssDNA (which corresponds to a few bases). The sizes of dsDNA molecules can be controlled from the nano- to millimeter-scale. The width of the double helix is about 2 nm, and each helix turn consists of ten base pairs that rise 3.4 nm (hence 0.34 nm/bp). Note that this data was obtained from the X-ray crystal structure, and that these numbers only apply to so-called B-DNA (the most common conformation inside a cell). Other conformations can be induced with different salt concentrations. Even with B-DNA, the 3.4 nm per helical turn is not carved in stone it can fluctuate at least 10% in solution. The Youngs modulus of dsDNA is about 3 x 108 Pa21, similar to plexiglass. Topologically, dsDNA has two forms: linear and circular, but no branching. Circular DNA also has different supercoiled isoforms. Topoisomerases are enzymes that can alter the topology of DNA. In addition, the helix of DNA can also be left-handed (Z-DNA). Although the majority of natural DNA is B-DNA, Z-DNA can be easily made with certain sequences. A nanoscale mechanical device has already been built based on the B-Z DNA transition22. Chemically, DNA is nontoxic, stable, and water soluble. Most importantly, DNA is easily manipulated. A myriad of physical and chemical methods can process DNA with angstrom-level accuracy. For example, DNA can be cut at specific sites by more than 3000 restriction enzymes23 and then rejoined covalently and selectively by ligase enzymes. Some of these tools are detailed below, the number of which is enormous and surpasses those for any other current polymer, synthetic or natural (Table 1).
method of isolating bulk DNA is through chromatography, and purification kits can be obtained from various commercial sources. DNA from commonly used sources, such as DNA from phage, is also available commercially ( DNA has been used in DNA-based lithography24). Individual genes or DNA fragments with desired sequences can be preserved, most often in a circular DNA form called a plasmid. Plasmid DNA is preferentially replicated (up to hundreds of copies) every 20 minutes or so inside bacteria or yeast. Billions of bacteria can be obtained through overnight culturing, from which a milligram of plasmid DNA can be isolated and prepared. Most laboratories purchase products to purify plasmid DNA in under an hour. Oligonucleotides (small DNA molecules with several to a few hundred bases) can be synthesized from scratch. The price of synthesis continues to decline because of demand. In many cases, it is more cost effective to outsource than to synthesize in-house. The control of sequence design provides extreme flexibility in using DNA for materials. The most common technique to obtain or verify DNA is PCR. This is a chemical chain reaction that amplifies a template DNA exponentially. It is so powerful that one single copy of dsDNA can be specifically amplified. In addition, realtime PCR uses fluorescence spectroscopy to monitor the process, making quantification and comparison possible.
Tool kit 2: manipulating DNA

A myriad of tools are available to manipulate DNA (Table 1). Enzymes have evolved to perform a variety of functions just like the toolkits we have in our garages. For example, many kinds of endonucleases function as scissors, cutting DNA strands at internal sites. Exonucleases function as sanders and cut DNA strands from their terminal ends. Restriction enzymes can recognize and cleave different DNA sequences, serving as accurate table saws. On the other hand, nonspecific nucleases, such as micrococcal nuclease, degrade both ssDNA and dsDNA to individual monomers, an example of a grinder. Single-stranded endonucleases, such as mung bean nuclease, can be viewed as trimmers, since they cleave protruding sequences of ssDNA. In addition, DNA can be elongated by various polymerases. Some of these enzymes perform the elongation at elevated temperatures (e.g. Taq polymerase). DNA can also be connected covalently by ligases, whose function can be viewed as a permanent glue. If the dsDNA molecules have protruding ends, then ligation
Tool kit 1: obtaining DNA

Bulk DNA can be purified from cultured organisms, including bacteria, yeast, and mammalian or plant cells, by physically breaking down the cells via mechanical disruption and/or enzymatic digestion. DNA can then be precipitated with isopropanol in the presence of high salt. Currently, the easiest
40
November 2003
INSIGHT FEATURE
Table 1 Tools for manipulating DNA.
Function
Polymerization template needed? Yes
Enzymes (examples)
T4 DNA Polymerase
Properties
Uses existing DNA as a template and adds deoxynucleotides to the 3-hydroxy end of the complementary strand. Direction: 5 3 Same as the above Same as the above; only works at 72C Same as the above; only works with a RNA template (i.e. makes DNA from RNA) Adds deoxyribonucleotides directly to the 3-hydroxy end without any templates Adds deoxyribonucleotides (A) to RNA at the 3-end Joins dsDNA together Joins dsDNA together Joins RNA together Joins ssDNA together Cuts DNA from 5 3 Cuts DNA from 3 5 Degrades only RNA in DNA:RNA hybrids Degrades only ssDNA Cuts DNA from both 5- and 3-ends Cuts ssDNA from both 5- and 3-ends Degrades RNA Degrades RNA Degrades RNA Cuts DNA at a specific sequence Degrades DNA Degrades DNA Labels DNA at 5 with phosphate, biotin, etc. Labels DNA at 5 with phosphate, biotin, etc. Labels DNA at 3-end Methylates DNA
Elongating DNA or RNA
E. Coli DNA polymerase I Taq polymerase Reverse transcriptase

No Terminal transferase Poly A polymerase
Non-polymerization
T4 DNA ligase E. coli DNA ligase T4 RNA ligase Exonuclease Exonuclease III Ribonuclease H Mung bean nuclease Bal 31 nuclease Exonuclease VII Ribonuclease A Ribonuclease T1 Micrococcal nuclease Restriction enzymes (more than 3000) DNase I Micrococcal nuclease T4 polynucleotide kinase Klenow fragment Terminal transferase Methylase
Specific Shortening DNA or RNA Exonucleases Endonucleases
Nonspecific
Specific Nonspecific
Modifying
At 5 end At 3 end At internal sites
can only happen when the two ends are complementary. This makes the ligation specific, a very useful feature in connecting two fragments of DNA. Finally, DNA strands are very easily modified at both the 5 and 3 ends. Hundreds of small chemicals can be conjugated onto the ends during synthesis. These chemicals include primary amine or thiol groups, biotin, fluorescent dyes, etc. These conjugations make DNA even more versatile in constructing other materials.
Tool kit 3: attaching DNA to a surface

DNA can be attached to a surface via noncovalent or covalent bonds. Two methods are commonly used for noncovalent attachment. In the first, since DNA is highly negatively charged, it is attached to positively charged surfaces, for example nylon sheets, nitrocellulose papers, polylysine-coated polystyrene plates, and glass. Because of the charge interactions, the attachment is nonspecific and reversible. In the second method, DNA strands are premodified with biotin, a small vitamin that has an extraordinary affinity with the protein avidin (or
streptavidin). The biotin-avidin interaction is the strongest noncovalent binding known and is essentially irreversible. Once the surface is modified with avidin molecules, the biotin-labeled DNA can be fixed onto the surface. The most frequently used covalent attachment approach is the linkage between S and Au, a classical method that established the self-assembled monolayer technique25,26. DNA is end-labeled with either a thiol or disulfide group. The sulfur-conjugated DNA is then self-adsorbed onto an Au surface, forming a monolayer with a controlled orientation.
DNA nanotechnology and materials

Seeman and colleagues pioneered DNA nanotechnology. A variety of geometric objects and periodic frameworks have been constructed successfully using double crossover DNA (two helical domains connected by two crossovers)27-36. The double crossover motifs have been used to grow nanoscale arrays and scaffolds37-39. Linear DNA has also been used in other schemes of nanotechnology, including a biotin-avidin based gene network40, DNA-conjugated Au nanocrystals41,
November 2003
41
INSIGHT FEATURE
Fig. 1 Y-shaped DNA. Schematic diagrams of the structure (left) and sequence (middle) of Y-DNA, and dendrimer-like DNA (right).
hybrid DNA-protein nanocomplexes42, a DNA-fueled molecular machine43, DNA-block copolymer conjugates44, DNA-templated Ag wire45, and DNA-mediated supramolecular structures46. In addition, Mirkin has succeeded in patterning DNA via dip-pen nanolithography47. Recently, his group reported a DNA detection method using DNA-Au nanoparticles and microelectrodes. This also opens up new possibilities for DNA-nanowires48. DNA-based lithography has been reported by Brauns group, where long DNA molecules serve as masks and DNA-binding Rec A proteins serve as resists24. Niemeyer and colleagues have also reported forming DNA arrays from dendritic linker systems49 with chemical dendrimers. Note that in all these achievements, the DNA employed has been linear.
ends) in a controlled and enzyme-catalyzed fashion. This makes it possible (and easy) to grow the branches into a dendrimer-like structure (Fig. 1, right). It is hoped these multivalent and anisotropic DNA dendrimers will serve as an architectural backbone to which other chemical entities can be linked, thus serving as designer materials. A dendrimer-like DNA structure, which is multivalent and anisotropic, was assembled and imaged by transmission electron microscopy (Fig. 2) and atomic force microscopy (Fig. 3).
Summary
Two blockades exist on the road from molecular biology to molecular materials. The first is the control of assembly how to overcome the isotropic nature of building blocks in order to incorporate heterogeneous components into nanomaterials or nanodevices in a controlled manner. The second is the scale and purity of assembled molecular materials how to produce ready-to-use molecular building blocks in large quantities and high purities. With continuous integration of molecular biology with engineering, these roadblocks will be eliminated soon. Molecular materials will
A future direction
Great potential exists in using DNA as a generic instead of a genetic material. Although some progress has been made, DNA materials that can be used in devices have not been demonstrated yet. This is mainly because almost all DNA molecules are in linear or circular forms, severely restricting their usefulness in molecular constructions. In addition, the yield and purity of DNA materials are still too low to make them useful. To realize the potential, one must create different shapes of DNA molecules to provide additional building blocks. These building blocks must also be synthesized in high yield with high purity and the ability to be incorporated into larger structures in a controlled manner. Our group has created just such building blocks in the form of novel Y-shaped DNA or Y-DNA (Fig. 1, left). The branched Y-DNA molecules are designed to have specific sticky ends (Fig. 1, middle), which enable them to covalently connect to other specific Y-DNAs (via complementary sticky
Fig. 2 Transmission electron micrograph of dendrimer-like DNA.
42
November 2003
INSIGHT FEATURE
Fig. 3 Atomic force microscope images of dendrimer-like DNA (left) and circular plasmid DNA as a control (right).
significantly advance the bottom-up approach in nanotechnology and will effectively expand the repertoire of currently available materials. The route from molecular biology to molecular materials will indeed become a heavily traveled highway in the future. MT
Acknowledgments
I wish to thank all of my colleagues and students, past and present, for their contributions. Our research is partially supported by a Cornell University Innovation Grant administrated by Cornell Advanced Center for Biotechnology, National Science Foundation, and Cornell Nanobiotechnology Center.
REFERENCES 1. Seeman, N. C., Materials Today (2003) 6 (1), 24 2. Zhang, S., Materials Today (2003) 6 (5), 20 3. Watson, J. D., and Crick, F. H. C., Nature (1953) 171, 737 4. Saiki, R. K., et al., Science (1985) 230, 1350 5. Altman, G. H., et al., Biomaterials (2003) 24, 401 6. Gosline, J., et al., Philos. Trans. R. Soc. London, B (2002) 357, 121 7. Haupt, K., Chem. Commun. (2003) 2, 171 8. Holmes, T. C., Trends Biotechnol. (2002) 20, 16 9. Lee, S. W., et al., Science (2002) 296, 892 10. Niemeyer, C. M., Trends Biotechnol. (2002) 20, 395 11. Roco, M. C., Curr. Opin. Biotechnol. (2003) 14, 337 12. Sadler, K., and Tam, J. P., J. Biotechnol. (2002) 90, 195 13. Seeman, N. C., and Belcher, A. M., Proc. Natl. Acad. Sci. USA (2002) 99, 6451 14. Wise, K. J., et al., Trends Biotechnol. (2002) 20, 387 15. Wright, E. R., and Conticello, V. P., Adv. Drug Deliv. Rev. (2002) 54, 1057 16. Yeates, T. O., and Padilla, J. E., Curr. Opin. Struct. Biol. (2002) 12, 464 17. Zhang, S. G., et al., Curr. Opin. Chem. Biol. (2002) 6, 865 18. Bouchiat, C., et al., Biophys. J. (1999) 76, 409 19. Tinland, B., et al., Macromolecules (1997) 30, 5763 20. Toth, K., et al., J. Biochemistry (1998) 37 21. Cluzel, P., et al., Science (1996) 271, 792 22. Mao, C. D., et al., Nature (1999) 397, 144 23. Roberts, R. J., and Macelis, D., Nucleic Acids Res. (2001) 29, 268 24. Keren, K., et al., Science (2002) 297, 72 25. Bain, C. D. et al., J. Am. Chem. Soc. (1989) 111, 321 26. Bain, C. D., and Whitesides, G. M., Science (1988) 240, 62 27. Chen, J. H., and Seeman, N. C., Nature (1991) 350, 631 28. Seeman, N. C., J. Biomol. Struct. Dyn. (1990) 8, 573 29. Seeman, N. C., Trends Biotechnol. (1999) 17, 437 30. Seeman, N. C., Acc. Chem. Res. (1997) 30, 357 31. Seeman, N. C., et al., Biophys. J. (2000) 78, 308a 32. Sha, R. J., et al., Chem. Biol. (2000) 7, 743 33. LaBean, T. H., et al., J. Am. Chem. Soc. (2000) 122, 1848 34. Yang, X. P., et al., J. Am. Chem. Soc. (1998) 120, 9779 35. Mao, C. D., et al., Nature (1997) 386, 137 36. Sa-Ardyen, P., et al., Biophys. J. (2003) 84, 3829 37. Winfree, E., et al., Nature (1998) 394, 539 38. Niemeyer, C. M., Appl. Phys. A (1999) 68, 119 39. Seeman, N. C., Ann. Rev. Biophys. Biomol. Struct. (1998) 27, 225 40. Luo, D., Novel Crosslinking Technologies to Assess Protein-DNA Binding and DNA-DNA Complexes for Gene Delivery and Expression, Dissertation, The Ohio State University, 1997 41. Alivisatos, A. P., et al., Nature (1996) 382, 609 42. Niemeyer, C. M., et al., Angew. Chem. Int. Ed. Engl. (1998) 37, 2265 43. Yurke, B., et al., Nature (2000) 406, 605 44. Watson, K. J., et al., J. Am. Chem. Soc. (2001) 123, 5592 45. Braun, E., et al., Nature (1998) 391, 775 46. Taton, T. A., et al., J. Am. Chem. Soc. (2000) 122, 6305 47. Demers, L. M., et al., Science (2002) 296, 1836 48. Park, S. J., et al., Science (2002) 295, 1503 49. Benters, R., et al., Nucleic Acids Res. (2002) 30, E10
November 2003
43
LETTERS
Controlled assembly of dendrimer-like DNA

YOUGEN LI1, YOLANDA D. TSENG1, SANG Y. KWON1, LEO DESPAUX1, J. SCOTT BUNCH2, PAUL L. MCEUEN2 AND DAN LUO*1
Department of Biological and Environmental Engineering, Cornell University, Ithaca, New York 14853-5701, USA Laboratory of Atomic and Solid State Physics, Cornell University, Ithaca, New York 14853-5701, USA *e-mail: DL79@cornell.edu
2 1
Published online: 21 December 2003; doi:10.1038/nmat1045
NA possesses many desirable chemical/physical properties as a polymeric material. With the myriad of tools available to manipulate DNA1, there is great potential for using DNA as a generic instead of a genetic material.Although much progress has been made in DNA computing24 and DNA nanotechnology519, the full achievement of DNA-based materials has not yet been realized. As almost all DNA molecules are either linear or circular, to rationally construct DNA materials one must first create additional shapes of DNA as basic building blocks. In addition, these DNA building blocks must be readily incorporated into larger structures in a controlled manner. Here, we show the controlled assembly of dendrimer-like DNA (DL-DNA) from Y-shaped DNA (Y-DNA). The synthesis of Y-DNA and controlled assembly of DL-DNA were robust and efficient; the resulting DL-DNA was stable and almost monodisperse. The multivalent DNA dendrimers can be either isotropic or anisotropic, providing great potential to link other entities. Two strategies were used to synthesize the Y-DNA: stepwise synthesis and all-in-one (one-pot) synthesis (Fig. 1a). In the stepwise approach, two oligonucleotides with partial complementary sequences formed one arm of a Y-DNA; then the third oligonucleotide that was complementary to the first two unmatched portions of oligonucleotides, formed the other two arms of the Y-DNA. In the onepot approach, equal moles of all three oligonucleotides were mixed together to form the Y-DNA.In both cases,the formation of Y-DNA was evaluated by gel electrophoresis (Fig. 1c), in which the mobility of a DNA molecule depends on its size, shape and extent of base pairing20. One major band appears on the gel (Fig. 1c, lanes 46), and its mobility is less than that of its components,the single-stranded DNA (lanes 13), indicating the formation of one arm of Y-DNA. The further shift of the mobility of the final annealing product of stepwise (Fig. 1c, lanes 79) and one-pot synthesis (Fig. 1c, lane 10) indicated the formation of Y-DNA. There is no difference in results between stepwise and one-pot synthesis.The estimated yield ofY-DNA is close to 100%,as estimated by densitometry. Other Y-DNA with different sequences were similarly synthesized. Synthesized Y-DNA were stable with no degradation observed after 30 days at 4 C (Fig. 1d). DL-DNA were assembled by ligation of Y-DNA molecules, whose sequences were specifically designed so that ligations between Yi and Yj DNA could only occur when i j, where i and j refer to the generation number n (for example,G1,G2,etc.,see Fig.3a.The cohesive end of each oligonucleotide was non-palindromic, thus no self-ligations occurred,
38
see Table 1,segment 1).In addition,the ligation could only occur in one direction,that is,Y0 Y1 Y2 Y3 Y4 and so on.Furthermore,when Y0 was ligated to Y1 with a 1:3 molar stoichiometry, one Y0 was linked with three Y1,forming the first-generation DL-DNA (G1,Fig. 2a).G1 was then ligated to six Y2 (one Y2 for each of the six free branches of G1), resulting in a second-generation DL-DNA (G2, Fig. 3a). The third (G3), fourth (G4), and higher generation DL-DNA were assembled in a similar way (Fig. 3a). Note that the assembled DL-DNA (Gn) had only one possible conformation due to the unidirectional ligation strategy. The general format of the nth-generation DL-DNA is Gn = (Y0)(3Y1)(6Y2)(3 2n1Yn), where n is the generation number and Yn is the nth Y-DNA. The total number of Y-DNA in an nth-generation DL-DNA is 3 2n 2. The growth of DL-DNA from nth generation to (n + 1)th generation requires a total of 3 2n new Yn+1-DNA. G1 DL-DNA were assembled by ligating Y0 and Y1 with a 1:3 stoichiometry (Fig. 2a). The ligation product migrated as a single band, and its mobility was slower than that of its building block, Y0 (Fig. 2c). The presence of a single band indicated a new molecular species with a well-defined molecular weight. The estimated yield is close to 100%. To confirm that the ligation product was indeed G1 DL-DNA, it was denatured (Fig. 2b) and examined by gel electrophoresis (Fig. 2d). Two major bands appeared in the electrophoresis: one with the same mobility as the single-strand DNA Y0a (30-mer) and one with slower mobility (90-mer), which was exactly what one would expect from the G1 DL-DNA structure according to the assembly scheme (Fig. 2b). Similar results were obtained from denaturation of G2, G3 and G4, and the generation of newly ligated species were revealed by electrophoresis (data not shown). Assembled G1 DL-DNA were stable with no degradation observed after 45 days at 4 C (Fig. 2e). In addition, exonuclease III assays confirmed the absence of cyclic materials (data not shown). The second-, third-, fourth- and fifth-generation DL-DNA were synthesized with a similar strategy and evaluated by gel electrophoresis (Fig. 3a and 3b).With increasing generation, the mobility of the ligated product decreased as predicted (Fig. 3b, see arrows). Furthermore, the yield and the purity of higher generation DL-DNA did not seem to decrease even in the absence of purification, suggesting that the assembly was very robust. To further confirm that the mobility-shifted species were indeed DL-DNA molecules, we examined the 4th generation DL-DNA by atomic force microscopy (AFM) with both a
nature materials | VOL 3 | JANUARY 2004 | www.nature.com/naturematerials
2004 Nature Publishing Group
LETTERS
a
Step-wise synthesizing Y0 5 5 + 5 5 + 5 5 5 5
Y0a + Y0b Y0a-Y0b + Y0c Y0a-Y0b-Y0c Y0a + Y0c Y0a-Y0c + Y0b Y0a-Y0c-Y0b Y0b + Y0c Y0b-Y0c + Y0a Y0b-Y0c-Y0a All-in-one synthesizing Y0 5 + 5 5 + 5 Y0a + Y0b + Y0c Y0aY0bY0c
b
G
TGACAGGCTGATTCGGT C TCCGACTAAGCCA GT ATT AA C G GC CG C G GT CA AA TT G CC A
TG
AT C
c
1 Y0a 2 Y0b 3 Y0c 4 Y0a-Y0b 5 Y0a-Y0c 6 Y0b-Y0c 7 Y0a-Y0b-Y0c 8 Y0a-Y0c-Y0b 9 Y0b-Y0c-Y0a 10 Y0aY0bY0c
Figure 1 Y-shaped DNA (Y-DNA). a,Strategies of Y-DNA synthesis.b,Schematic drawing (left) and sequences (right) of Y-DNA.c, Evaluation of Y0-DNA formation.Lanes 1,2 and 3 are oligonucleotide Y0a,Y0b,Y0c,respectively.Lanes 4,5 and 6 correspond to the hybridized products of (Y0a and Y0b),(Y0a and Y0c) and (Y0b and Y0c),respectively.Lanes 7,8 and 9 are stepwisely hybridized final products of (Y0a,Y0b and Y0c),(Y0a,Y0c and Y0b) and (Y0b, Y0c and Y0a),respectively.Lane 10 is all-in-one (one-pot) hybridized final product of (Y0a,Y0b and Y0c).d,Evaluation of Y-DNA stability.Lane 1 represents freshly made Y-DNA and lane 2 is the same Y-DNA stored at 4 C for 30 days.
TC AT G
AT G
TG
AC
CG
AT C
CA
90 bases Denature
30 bases
c
1 2
d
1 2 3 4
e
1 2
d
1 2
90 bases
30 bases
Figure 2 The first-generation dendrimer-like DNA (G1 DL-DNA). a, Sequences of G1 DL-DNA. b,A schematic drawing of denaturation strategy to confirm the G1 DL-DNA structure.After G1 DL-DNA denaturation, six oligonucleotides were generated; three of these six oligonucleotides were new species with a unique length (90 bases).The rest three were 30 bases. c, Evaluation of G1 DL-DNA formation. Lane 1 is Y-DNA and lane 2 is G1 DL-DNA. d, Evaluation of G1 DL-DNA denaturation. Lane 1 is a molecular marker (oligonucleotide Y0a). Lane 2 is G1 DL-DNA without denaturing. Lanes 3 and 4 correspond to 0.25 g and 0.5 g of the denatured G1 DL-DNA, respectively. e, Evaluation of G1 DL-DNA stability. Lane 1 represents freshly made G1 DL-DNA and lane 2 is the same G1 DL-DNA stored at 4 C for 45 days.
39
LETTERS
1 2 3 4 5
G5 G4
G3 G2 G1
G3
G2
G1
Figure 3 Higher generation dendrimer-like DNA. a, Schematic drawings of higher generation DL-DNA.b,Evaluation of higher generation DL-DNA formation.Lanes 1,2,3,4 and 5 correspond to G1 DL-DNA,G2 DL-DNA,G3 DL-DNA,G4 DL-DNA and G5 DL-DNA,respectively.The arrows indicate the decreasing mobility with increasing generation. c,Images of DL-DNA: AFM images of G4 DL-DNA on mica surface using standard tip (lower left) and SWNT tip (top,left and right),and TEM image of G4 DL-DNA (lower right).Scale bars correspond to 100 nm.
standard tip and a single-walled carbon nanotube (SWNT) tip, which revealed DL-DNAs highly branched dendritic nanostructure (Fig. 3c). The measured width of the G4 DL-DNA nanostructure was 71.2 6.7 nm, which was very close to the theoretically calculated
40
value based on a B-DNA structure (69.0 nm) considering the relative flexibility of DNA molecules. Although the AFM pictures revealed the dendritic shape of DLDNA, they were not suitable for determining the distribution of sizes
LETTERS
due to many commonly occurring problems associated with AFM, including sample damage by AFM tips, cleanliness of the substrate, dehydration of DNA, contaminants during the preparation and so on. To further explore the size distributions, the 4th-generation dendrimerlike DNA was also visualized by transmission electron microscopy (TEM), which indicated the highly branched shape of DL-DNA (Fig. 3c). The size (diameter) of the DL-DNA was measured at about 80 nm, and the width of each branch was estimated at 3 nm, consistent with the theoretically calculated values.More than 20 TEM experiments were performed,and a total of 146 G4 DL-DNA particles were measured. From such a large number of results, we believe that these TEM measurements are likely to reflect the nature of the dendrimer-like DNA. We measured the size of each particle in four directions: horizontal, vertical, diagonal 45 (relative to horizontal) and diagonal 135 (relative to horizontal). The average size of all 146 particles was 83.9 nm 9.8 nm, very close to the theoretical calculated value (considering the Pt/Pd coating). More interestingly, the average size of horizontal,vertical,diagonal 45and diagonal 135measurements were 84.9 nm,85.1 nm,82.1 nm and 83.9 nm,respectively,suggesting that the particles were circular in shape. Further examination of particle-size distributions of these 146 G4 DL-DNA particles revealed that 84% of the particles are within a size range 74.1 nm and 93.7 nm (one standard deviation of the average), consistent with our gel electrophoresis results, which suggested that 87.68% and 92.99% of the G1 and G4 DNA, respectively, had essentially the same mobility. Taking AFM, TEM and gel electrophoresis results together, they strongly suggested that DL-DNA molecules were very pure, and confirm their nanoscale dendritic structures. In another study, seven dangling-ends of double-stranded DNA were annealed to construct dendrimer DNA, and the structure fixed through a non-specific crosslinker, psoralen21,22. The yield, purity and images were unknown.The dendrimer-like DNA presented here is very different from previously reported products. First, the building blocks were Y-shaped DNA with specially designed sequences.No self-ligation and cyclic products could occur, making dendrimer growth unidirectional and stepwise. Second, the growth of dendrimer was enzyme-catalysed, making the synthesis non-reversible, specific and efficient. And third, the final products, dendrimer-like DNA, were still true DNA that could be manipulated further with DNA enzymes. We are aware that different motifs of a variety of geometric arrays were successfully (and impressively) constructed using rigid, linear crossover DNA as building blocks2325. A DNA mechanical device was also created9,10.Note that the building blocks used and motifs created are perfect for growing nanoscaled arrays and scaffolds11,26, some of them (such as the DNA tiling system11) also provide a method for nonisotropic growth. We emphasize here that this is the first time that almost monodisperse Y-shaped DNA and dendrimer-like DNA nanostructures have been synthesized in a highly controlled fashion with relatively high yield and purity. The synthesis was rather simple and robust; the 5th generation DL-DNA was close to being monodisperse even without any purification. The design strategies and assembly approaches can be easily used to construct other DNA building blocks,such as an X-shaped DNA (data not shown), that can be incorporated into even more complicated nanostructured material than those reported here. These DNA-based nanostructured materials can be manipulated at the molecular length-scale, providing an advantageous route for novel materials by design. In addition, it is unnecessary to protect and deprotect reactive ends of DL-DNA to achieve the dendritic structures. Furthermore, whereas chemical dendrimers are usually isotropic, DLDNA can be easily assembled as anisotropic as well as isotropic through specific end design. The branch units of DL-DNA can also be either symmetrical or asymmetrical. With about 50 nm persistence length of double-stranded DNA27, DNA basic building blocks can be at least as long as 50 nm, spanning the length scale of DNA nanostructured
Table 1 Sequences of oligonucleotides.
Strand Y0a Y1a Y2a Y3a Y4a Y0b Y1b Y2b Y3b Y4b Y0c Y1c Y2c Y3c Y4c
Segment 1 5-TGAC 5-GTCA 5-ATCG 5-ATGC 5-GCAA 5-TGAC 5-CGAT 5-GCAT 5-TTGC 5-GGAT 5-TGAC 5-CGAT 5-GCAT 5-TTGC 5-GGAT
Segment 2
TGGATCCGCATGACATTCGCCGTAAG-3
CTTACGGCGAATGACCGAATCAGCCT-3
AGGCTGATTCGGTTCATGCGGATCCA-3
material (including DNA dendrimers) from the nanometre to even the micrometre range, which is very difficult to achieve from chemical dendrimers. The reported nanostructured DL-DNA molecules are envisioned to have great potential in nanotechnology by serving as templates for fabrication and synthesis, as their sizes, structures and morphologies can be changed through alteration of building blocks, and their interactions can be enhanced through conjugations with different functional groups. In addition, we believe that these watersoluble,dendrimer-like DNA nanostructures are versatile,and may find a myriad of applications in both biomedical and non-biomedical fields.
METHODS
MOLECULAR DESIGNS
The DNA sequences (Table 1) were designed according to the standards set by Seeman28, and tested by trial-and-error. They were commercially synthesized and PAGE purified (Integrated DNA Technologies, Coralville, Iowa). Without further purification, oligonucleotides were dissolved in annealing buffer (10 mM Tris, pH = 8.0, 1 mM ethylenediaminetetraacetic acid (EDTA), and 50 mM NaCl) with a final concentration of 50 mM. Y-DNA was constructed by mixing three oligonucleotide components (1:1:1 molar ratio) in sterile Milli-Q water with a final concentration of 5 mM for each oligonucleotide. Hybridizations were performed according to the following procedures: (i) Denaturation at 95 C for 2 min. (ii) Cooling at 65 C and incubation for 5 min. (iii) Annealing at 60 C for 2 min. (iv) Further annealing at 60 C for 0.5 min with a continuous temperature decrease at a rate of 1 C per min. The annealing steps were repeated a total of 40 times. The final annealed products were stored at 4 C.
FORMATION OF Y-DNA AND DL-DNA

Y-shaped DNA (Y-DNA) were synthesized by mixing the same molar amount of corresponding oligonucleotide strands. The nomenclature is as follows: Y0a, Y0b and Y0c are the three corresponding single oligonucleotide chains that form a Y0-DNA (noted as Y0 in short). Similarly, Y1a, Y1b and Y1c are the three corresponding single oligonucleotide chains that form an Y1-DNA (noted as Y1 in short); and Yna, Ynb and Ync are the three corresponding single oligonucleotide chains that form an Yn-DNA (noted as Yn in short). The reactions are noted as the following: Y0a+Y0b+Y0cY0, Y1a+Y1b+Y1cY1, and Yna+Ynb+YncYn, and so on. (see Fig. 1a). All the mixtures were first incubated at 95 C for 2 min, then quickly cooled down to 60 C, and finally cooled to 4 C at 2 C per min. For constructing DL-DNA, individual Y-DNA was ligated specifically to other Y-DNA. G1 DL-DNA was obtained by ligating Y1 to Y0 (at a 3:1 molar ratio). Similarly, G2 was formed by ligating 6 Y2 with 1 G1. Other higher generations of DL-DNA were constructed using the same strategy. Each ligation solution contains ligase buffer, 1.30 nmol Y-DNA monomer, and 0.235 Weiss unit of T4 DNA ligase (Promega, Madison, Wisconsin). The ligations were also performed with Fast-Link DNA Ligase (Epicentre Technologies, Madison, Michigan). The nomenclature of DL-DNA is as follows: the core of the dendrimer, Y0, is designated as G0, the 0 generation of DL-DNA. After Y0 is ligated with Y1, the dendrimer is termed the 1st generation of DL-DNA (G1), and so on. The nth generation of DL-DNA is noted as Gn.
41
LETTERS
CHARACTERIZING DL-DNA
The nucleic acid samples were evaluated on 3% agarose Ready-Gel (Bio-Rad, Hercules, California) at 100 V. For the denaturing experiment, single-stranded DNA was obtained by using NaOH. Briefly, gelpurified DL-DNA was denatured with a final concentration of 40 ng ml1 in a denaturing buffer containing 10 mM EDTA and 25 mM NaOH. The denaturing reaction was carried out at 95 C for 2 min. The denatured products were immediately cooled down in a 20 C freezer. Denatured DL-DNA was electrophoresed in 3% native agarose gel with ethidium bromide (0.5 mg ml1) in Tris-acetate-EDTA (TAE) buffer at 4 C. Care was taken to make sure that the products were kept at 4 C to keep the denatured status. Electrophoresis was carried out at 50 V for 10 min and then at 125 V for 65 min. For Exonuclease III assays, 1.0 g of DNA were digested in 25 l of Exonuclease III buffer with 30 units of Exonuclease III (Promega, Madison, Wisconsin) at 37 C for 1 hour. The reaction was terminated by adding 2 l of 0.25 mM EDTA. 10. Yan, H., Zhang, X., Shen, Z. & Seeman, N. C. A robust DNA mechanical device controlled by hybridization topology. Nature 415, 6265 (2002). 11. Winfree, E., Liu, F., Wenzler, L. A. & Seeman, N. C. Design and self-assembly of two-dimensional DNA crystals. Nature 394, 539544 (1998). 12. Alivisatos, A. P. et al. Organization of nanocrystal molecules using DNA. Nature 382, 609611 (1996). 13. Niemeyer, C. M., Burger, W. & Peplies, J. Covalent DNA - Streptavidin conjugates as building blocks for novel biometallic nanostructures. Angew. Chem. Int. Edn 37, 22652268 (1998). 14. Yurke, B., Turberfield, A. J., Mills, A. P. Jr, Simmel, F. C. & Neumann, J. L. A DNA-fuelled molecular machine made of DNA. Nature 406, 605608 (2000). 15. Watson, K. J., Park, S. J., Im, J. H., Nguyen, S. T. & Mirkin, C. A. DNA-block copolymer conjugates. J. Am. Chem. Soc. 123, 55925593 (2001). 16. Braun, E., Eichen, Y., Sivan, U. & Ben-Yoseph, G. DNA-templated assembly and electrode attachment of a conducting silver wire. Nature 391, 775778 (1998). 17. Taton, T. A., Mucic, R. C., Mirkin, C. A. & Letsinger, R. L. The DNA-mediated formation of supramolecular mono- and multilayered nanoparticle structures. J. Am. Chem. Soc. 122, 63056306 (2000). 18. Elghanian, R., Storhoff, J. J., Mucic, R. C., Letsinger, R. L. & Mirkin, C. A. Selective colorimetric detection of polynucleotides based on the distance-dependent optical properties of gold nanoparticles. Science 277, 10781081 (1997). 19. Keren, K. et al. Sequence-specific molecular lithography on single DNA molecules. Science 297, 7275 (2002). 20. Kallenbach, N. R. et al. Fourth rank immobile nucleic acid junctions. J. Biomol. Struct. Dyn. 1, 159168 (1983). 21. Wang, J., Jiang, M., Nilsen, T. W. & Getts, R. C. Dendritic nucleic acid probes for DNA biosensors. J. Am. Chem. Soc. 120, 82818282 (1998). 22. Nilsen, T. W., Grayzel, J. & Prensky, W. Dendritic nucleic acid structures. J. Theor. Biol. 187, 273284 (1997). 23. LaBean, T. H. et al. Construction, analysis, ligation, and self-assembly of DNA triple crossover complexes. J. Am. Chem. Soc. 122, 18481860 (2000). 24. Yang, X. P., Wenzler, L. A., Qi, J., Li, X. J. & Seeman, N. C. Ligation of DNA triangles containing double crossover molecules. J. Am. Chem. Soc. 120, 97799786 (1998). 25. Mao, C. D., Sun, W. Q. & Seeman, N. C. Assembly of Borromean rings from DNA. Nature 386, 137138 (1997). 26. Niemeyer, C. M. Progress in engineering up nanotechnology devices utilizing DNA as a construction material. Appl. Phys. A 68, 119124 (1999). 27. Smith, S. B., Finzi, L. & Bustamante, C. Direct mechanical measurements of the elasticity of single DNA molecules by using magnetic beads. Science 258, 11221126 (1992). 28. Seeman, N. C. De novo design of sequences for nucleic acid structural engineering. J Biomol. Struct. Dyn. 8, 573581 (1990). 29. Hafner, J., Cheung, C. L., Oosterkamp, T. H. & Lieber, C. M. High yield assembly of individual singlewalled carbon nanotube tips for scanning probe microscopies. J. Phys. Chem. B. 105, 743746 (2001).
AFM IMAGING
A 5 l DNA sample was placed onto the surface of freshly cleaved mica (Ted Pella, Redding, California) functionalized with aminopropyltriethoxysilane (APTES, Aldrich) and allowed to adsorb to the mica surface for approximately 20 minutes. The mica was then rinsed in Milli-Q water and dried with compressed air. Tapping-mode AFM images were taken in air on both a Dimensions 3100 AFM and a Multimode AFM (Digital Instruments, Santa Barbara, California). Standard imaging was done with Pt/Ir-coated cantilevers with resonant frequencies of 60100 kHz and force constants of 1.25.5 N m1. High-resolution imaging was performed using SWNT tips mounted on Pt/Ir cantilevers using the pick up technique29, whereby SWNTs are grown on a silicon substrate; during AFM imaging of the SWNT covered Si substrate a tube is picked up off the surface and then used as an imaging tip. All images were processed with a flattening filter.
TEM IMAGING
Approximately 0.5 l of 0.21 mM G4 DNA was mixed with 200 l of Tris (30 mM, pH 8.0), and then 5 l of 5% 2, 4, 6Tri(dimethylaminomethyl phenol) (DMP) 30 was added to the mixture. A drop of the mixture (50 l) was placed onto a sheet of parafilm at room temperature. The drop was covered with a petri dish for 11 minutes. After 11 minutes the DNA molecules were picked up by touching the drop with the formvar/carbon-coated grid (Electron Microscopy Sciences, Fort Washington, Pennsylvania) and left for 3 minutes. The sample was then stained with 2% uranyl acetate for 1 minute. The grid was then blotted with filter paper and allowed to dry in air, then rotary coated with Pt/Pd. Grids were analysed at a voltage of 100 kV using a Philips EM-201.
Received 19 May 2003; accepted 25 November 2003; published 21 December 2003. References
1. Luo, D. The road from biology to materials. Mater. Today 6, 3843 (2003). 2. Adleman, L. M. Molecular computation of solutions to combinatorial problems. Science 266, 10211024 (1994). 3. Sakamoto, K. et al. Molecular computation by DNA hairpin formation. Science 288, 12231226 (2000). 4. Benenson, Y. et al. Programmable and autonomous computing machine made of biomolecules. Nature 414, 430434 (2001). 5. Seeman, N. C. DNA engineering and its application to nanotechnology. Trends Biotechnol. 17, 437443 (1999). 6. Seeman, N. C. DNA components for molecular architecture. Accounts Chem. Res. 30, 357363 (1997). 7. Demers, L. M. et al. Direct patterning of modified oligonucleotides on metals and insulators by dippen nanolithography. Science 296, 18361838 (2002). 8. Eckardt, L. H. et al. DNA nanotechnology: Chemical copying of connectivity. Nature 420, 286 (2002). 9. Mao, C. D., Sun, W. Q., Shen, Z. Y. & Seeman, N. C. A nanomechanical device based on the B-Z transition of DNA. Nature 397, 144146 (1999).
Acknowledgements
This work was partially supported by the Cornell Universitys Innovation Grant administrated by Cornell Advanced Center for Biotechnology, and performed in part at the Cornell Nanofabrication Facility and Cornell Centre for Materials Research, which is supported by the National Science Foundation, Cornell University and Industrial Affiliates. We thank Francis Moran for preparation of Fig. 1b. Correspondence and requests for materials should be addressed to D.L.
Competing financial interests

The authors declare that they have no competing financial interests.
42
LETTERS
2005 Nature Publishing Group http://www.nature.com/naturebiotechnology
Multiplexed detection of pathogen DNA with DNA-based uorescence nanobarcodes

Yougen Li, Yen Thi Hong Cu & Dan Luo
Rapid, multiplexed, sensitive and specic molecular detection is of great demand in gene proling, drug screening, clinical diagnostics and environmental analysis13. One of the major challenges in multiplexed analysis is to identify each specic reaction with a distinct label or code4. Two encoding strategies are currently used: positional encoding, in which every potential reaction is preassigned a particular position on a solid-phase support such as a DNA microarray58, and reaction encoding, where every possible reaction is uniquely tagged with a code that is most often optical or particle based4,913. The micrometer size, polydispersity, complex fabrication process and nonbiocompatibility of current codes limit their usability1,4,12. Here we demonstrate the synthesis of dendrimer-like DNA-based, uorescence-intensity-coded nanobarcodes, which contain a built-in code and a probe for molecular recognition. Their application to multiplexed detection of the DNA of several pathogens is rst shown using uorescence microscopy and dot blotting, and further demonstrated using ow cytometry that resulted in detection that was sensitive (attomole) and rapid. Recently, dendrimer-like DNA (DL-DNA) nanostructures have been synthesized by our group14. The multivalent and anisotropic properties of DL-DNA were used here as uorescent dye carriers (that is, scaffoldings) to construct uorescence-intensity-encoded nanobarcodes. We rst synthesized uorescence-labeled Y-shaped DNA (Y-DNA), where each Y-DNA consisted of three oligonucleotide components that were complementary to each other. One of the oligonucleotides had a sticky end, and the other two were labeled with either uorophore(s) or a molecular probe. After hybridization, these oligonucleotides formed a uorescence-labeled Y-DNA (Fig. 1a) that was used as a peripheral outermost layer of DL-DNA to construct uorescence-labeled DNA nanostructures. Since both dye type and dye number can be precisely controlled, multicolor uorescenceintensity-encoded nanobarcodes could be fabricated (Fig. 1b). The decoding is based on the different ratios of different uorescent dyes, independent of the dye positions (Fig. 1c). During the construction of DNA nanobarcodes, molecular probes were linked to the free reactive ends of DL-DNA. A myriad of DNA-manipulation enzyme tools15 makes it very easy to attach molecular probes (e.g., DNA or RNA probes, or even antibodies) to DNA nanobarcodes. Consequently, the resultant DNA nanobarcodes not only had coding capacity, but also contained molecular recognition elements that could be used for molecular detection. Two types of uorescent dyes, Alexa Fluor 488 and BODIPY 630/ 650, were used to label DNA (see Supplementary Tables 1 and 2 online). The uorescence-labeled Y-DNA (see Supplementary Table 3 online) was ligated to other Y-DNAs via complementary sticky ends. Five nanobarcodes, 4G1R, 2G1R, 1G1R, 1G2R and 1G4R were constructed, where the number refers to the quantity of each dye molecule on one nanobarcode (see Supplementary Table 4 online and Fig. 1c). The resultant nanobarcodes were evaluated using agarose gel electrophoresis. Oligonucleotides labeled with either Alexa alone or BODIPY alone were controls (Fig. 1d, lanes 1 and 7, respectively); the obvious color changes from green to yellow to red (Fig. 1d, lanes 2 to 6) indicated the formation of the expected different nanobarcodes, a result further conrmed by the electrophoretic mobility shift of DNA nanobarcodes relative to the starting oligonucleotides (Fig. 1d and Supplementary Fig. 1a online). The formation of dendritic DNA nanobarcodes was also conrmed by the generation of oligonucleotides whose lengths differed from the starting oligonucleotides, which were revealed by denaturing agarose gel electrophoresis (see Supplementary Figs. 1b and 2 online). Details of the evaluation of DNA nanostructures using gel electrophoresis can be found in a previous publication14. The diameter of DNA nanobarcodes was o30 nm, which is far below the detection limit of optical microscopy. Polystyrene microbeads (5.5 mm diameter) were thus used to amplify the uorescence signals for imaging and molecular detection. The microbead-based amplication strategy and detection format are shown in Figure 2a. In this arrangement, two sets of single-stranded (ss)DNA probes were used. The rst set (capture probes) was biotin labeled and immobilized onto avidin-functionalized microbeads. Note that each batch of microbeads was attached with only one type of capture probe, which was complementary to a part of a particular target DNA (that is, sample DNA to be detected). Multiple types of microbeads were then pooled together to form a library of microbeads. The second set of ssDNA probes (report probes) was coupled to specic nanobarcodes, where each report probe was complementary to another part of the particular target DNA and thus was able to be hybridized onto a specic microbead in the presence of the target DNA via a sandwiched hybridization (Fig. 2a). Because each microbead could accommodate a large number of sandwiched complexes, uorescence signals from nanobarcodes
Department of Biological and Environmental Engineering, Cornell University, Ithaca, New York 14853-5701, USA. Correspondence and requests for materials should be addressed to D.L. (e-mail: DL79@cornell.edu). Published online 12 June 2005; doi:10.1038/nbt1106
NATURE BIOTECHNOLOGY
VOLUME 23
NUMBER 7
JULY 2005
885
LETTERS
were amplied. In the assay, the sample containing multiple types of unknown DNA (that is, target DNA) was mixed with microbeads in suspension, and captured onto the microbeads containing their corresponding capture probe sequences presented on the bead surface through DNA hybridization. A solution containing a nanobarcode library was then added. Since each nanobarcode was connected with a particular report probe, which in turn hybridized to another portion of the target DNA, nanobarcodes were specically bound to corresponding microbeads. The resultant microbeads were rst evaluated individually by uorescence microscopy, and the overlay color images are shown in Figure 2b, following a similar color-processing approach reported previously16. Even with the naked eye, ve different nanobarcodes could be distinguished from Figure 2b. The sample mixture, which consisted of DNA from four targets (Bacillus anthracis17, Francisella tularensis18, Ebola virus19 and SARS coronavirus20), was analyzed with nanobarcodes, and four different pseudocolors were revealed in the overlay image (Fig. 2c). Compared to Figure 2b, we found that four different nanobarcodes, 4G1R, 2G1R, 1G1R and 1G4R, were bound to different microbeads as designed. With the preassigned barcode library (see Supplementary Table 5 online), we concluded that the samples contained characteristic DNA from four pathogens: B. anthracis, F. tularensis, Ebola and SARS. The quantitative decoding results at the population level are also shown (see Supplementary Fig. 3 online). Taken together, with one nanobarcode (e.g., 1G1R) serving as a reference, other nanobarcodes could be easily decoded. Nanobarcodes can also be used for blotting-based detection (Southern, northern and western). Here we used dot blotting to demonstrate this potential. Six samples, including the controls of 27-mer ssDNA with irrelevant sequences and a 6.1-kb plasmid DNA, were rst blotted onto a membrane; all ve types of nanobarcodes (see Supplementary Table 4 online) were then used to hybridize onto target DNA for detection (Fig. 3a). Simultaneous detection of four colors on the resultant membrane (Fig. 3b) indicated four DNA targets, which were subsequently identied by referring to a preassigned decoding library. As expected, no uorescent signals were detected in the two control spots, suggesting nanobarcode-based detection is highly specic. In addition, a barcode was similarly used to successfully detect a pathogen genomic DNA (Mycobacterium avium subsp. paratuberculosis) extracted from animal feces in the presence of lysates or serum (data not shown).
Hybridization
Probe 1 4G1R
Relative fluorescence intensity
Y-DNA
Probe 2
b
Probe Probe 3 Probe Decode
2G1R
1G1R
Ligation
Probe 4 1G2R
1G4R Probe 5 Wavelength
Figure 1 Synthesis of nanobarcodes. (a) Schematic illustration of synthesis of a typical Y-DNA-based nanobarcode building block. Three starting oligonucleotide components were partially complementary to each other as indicated in the drawing. One oligonucleotide possessed a sticky end, another one was labeled with a uorescent dye and the third one was labeled with a uorescent dye or a probe depending on the experimental design. (b) Schematic illustration of the construction of a typical DL-DNA-based nanobarcode. The nanobarcode building blocks were covalently linked with each other through complementary sticky-end ligations. (c) Schematic illustration of barcode decoding. The nanobarcodes 4G1R, 2G1R, 1G1R, 1G2R and 1G4R were decoded based on the ratio of uorescence intensity. A molecular recognition element, a probe, was also attached to each nanobarcode. The resultant nanobarcodes possess not only coding capability and capacity, but also molecular sensing ability. With a preassigned code library (see Supplementary Table 5 online), the nanobarcodes could be used for molecular detection. (d) The real color of nanobarcodes in an agarose gel illuminated with a strong UV light. Lanes 1 and 7 are Alexa Fluor 488labeled starting oligonucleotide component and Bodipy 630/650labeled starting oligonucleotide component, respectively. Lanes 2, 3, 4, 5, 6 are nano-barcodes 4G1R, 2G1R, 1G1R, 1G2R and 1G4R, respectively.
886
VOLUME 23
NUMBER 7
JULY 2005
LETTERS
a
Immobilization Hybridization
Avidin 5.5-m polystyrene bead
Biotin
Nanobarcode
DNA target
Biotin-labeled capture probes
Barcodes with report probes
c
Figure 2 Microbead-based DNA detection using uorescence microscopy. (a) Schematic drawing of a sandwiched DNA nanobarcode whose signal is amplied from polystyrene microbeads. Briey, biotin-labeled capture probes were attached to avidin-functionalized polystyrene microbeads. Each batch of microbeads had only one type of capture probe before all batches were pooled together. DNA targets (that is, control or unknown samples) were then captured by specic microbeads rst. Each report probe, which was linked to a particular nanobarcode, was designed to be complementary to another part of a specic target DNA and thus was able to be hybridized onto a specic microbead. Since each microbead bound a large amount of sandwiched complexes (that is, capture probes/target DNA/report probes/nanobarcodes), uorescence signals were amplied. (b) Merged uorescent colors (pseudocolors) of nanobarcodes from individual microbeads. (c) Multiple target detection (a total of four targets) was achieved via a two-colored uorescence microscope using DNA nanobarcodes and microbeads. All scale bars, 5 mm.
Flow cytometry is a powerful tool for multiplexed molecular detection21,22. The application of nanobarcodes to multiplexed molecular detection was also demonstrated using ow cytometry. The uorescence intensity ratio is the basis for the decoding of any nanobarcodes aGbR (see Supplementary Note 1 online for the mathematical processing of ow data). To simultaneously detect DNA from multiple pathogens, we made a solution containing mixtures of DNA from three pathogens with one irrelevant DNA sample as a negative control. This solution was treated as a sample with unknown pathogen DNA. Samples were allowed to bind to the microbeads without saturation. All ve types of nanobarcodes (see Supplementary Table 4 online) were then added into a bead suspension. After the hybridization was complete, the uorescence intensity ratio was measured for each microbead by ow cytometry (Fig. 4b). From the intercepts on the ow plots, the K values (the total uorescence intensity ratio between red and green uorophores, see Supplementary Note 1 online) of each line (top to bottom) were determined to be 11, 44 and 180, and their code numbers were thus calculated to be around 4.0, 1.0 and 0.25, respectively. Therefore, the detected nanobarcodes were 4G1R, 1G1R and 1G4R. The same detection experiments were repeated four times and the statistical data showed that the coding nanobarcodes were indeed 4G1R, 1G1R and 1G4R (see Supplementary Fig. 4 online). After referring to the preassigned barcode library (see Supplementary Table 5 online), we concluded that the unknown samples contained DNA from three pathogenic species: B. anthracis, Ebola and SARS. The detection limit was 6.2 1016 mole or 620 attomole. The nal detection step was completed within 30 s. The success of simultaneously identifying multiple pathogen DNA with attomole sensitivity within a very short period of time (o1 min) demonstrated the potential of nanobarcodes for multiplexed molecular detection. Theoretical and practical coding capacity is discussed in Supplementary Note 2 online.
The successful synthesis and application of DL-DNA based nanobarcodes reveals two novel concepts: (i) multiplexed detection can be achieved by detecting different uorescence intensity ratios instead of different uorescent colors, and (ii) DL-DNA can be used as both structural scaffolding and functional probes. Although DNA-based signal amplications have been reported previously (e.g., with
Adding target DNA
Nanobarcodes Prehybridization
Membrane
DNA target
Barcode
Barcode-labeled probe
Bacillus anthracis
Control 1 Francisella tularensis
Control 2
Ebola virus
SARS coronavirus
Figure 3 DNA blotting assay with nanobarcodes. (a) Schematic drawing of a dot-blotting detection of multiple DNA targets with nanobarcodes. Target DNA molecules were manually blotted onto a nylon-membrane. After prehybridization and blocking, a library of nanobarcode mixture was loaded onto the membrane. Through specic hybridizations with report probes that were functionalized with nanobarcodes, target DNA molecules were detected using a uorescence reader, scanner or microscope. (b) DNA from multiple pathogens (four in total) were detected simultaneously using nanobarcodes. Control 1 was a 27-mer ssDNA with unrelated sequences and control 2 was a plasmid DNA, pVAX1/lacZ. Scale bar, 1 mm.
VOLUME 23
NUMBER 7
JULY 2005
887
LETTERS
chemical cross-linked dendritic DNA2327 or a b backbone-branched oligonucleotides28,29), to our knowledge no one else has demonstrated successful multiplexed detection using DNA nanobarcodes, whose uorescence intensity ratios were precisely controlled at an individual molecular level; at present such tight control of color-conjugations can be achieved only with anisotropic, multivalent carriers, such as DL-DNA. In addition, the DNA scaffold makes our nanobarcodes biocompatible and thus has the potential to be applied in vivo after solving degradation problems. 100 101 102 103 104 100 101 102 103 104 FL4-H Another advantage of using DL-DNA as a FL4-H 22 11 44 180 scaffolding material lies in the existence of DNA modifying enzymes. Functional ele- Figure 4 Multiplexed DNA detection using ow cytometry. (a) A two-color ow plot of microbeads ments can be easily introduced either before attached with the nanobarcode 2G1R as a control for standards (a calibration control). FL1H indicates (e.g., the molecular recognition elements can the green channel and FL4H is the red channel. (b) Simultaneous detection of three pathogen DNA be linked to the oligonucleotide components) using nanobarcodes. Unrelated DNA sequences were not detected (background). or after nanobarcode synthesis. Moreover, the nano-scale size of our nanobarcodes ensures easier target access in vivo or in situ when compared to current micro- denaturation was performed14. Briey, a 10 pmol DNA sample in a denaturing sized codes4,912. Note that the in situ possibility can be realized only buffer (10 mM EDTA, 25 mM NaOH) was heated at 95 1C for 2 min and then when a spatial resolution can be achieved to distinguish individual immediately cooled down in a 20 1C freezer. The denatured DNA sample was run through a 3% agarose gel at 50 V for 10 min and then 100 V for 80 min at barcodes (for example, by DL-DNA that is larger than 150 nm). 4 1C in TAE buffer containing 0.5 mg/ml of ethidium bromide. The use of common and commercially available uorophores does not require special equipment for detection, effectively expanding the Library. To detect pathogen DNA (here we targeted B. anthracis, F. tularensis, power of traditional microscopy. In addition, this technique could also Ebola virus and SARS coronavirus), a small fragment of characteristic DNA substitute both isotope and uorochrome labeling for blotting-based, sequences from the genome of each species was selected as the target DNA. Two multiplexed detection without resorting to multiple runs or repeated separate sets of DNA probes, which were complementary to the two regions of probe stripping, as practiced at present. Furthermore, nanobarcodes the same target DNA, were synthesized. One blank control, where the two sets make fast, sensitive and multiplexed detection possible in a ow of probes were complementary to each other, was also chosen. Thus, a library (see Supplementary Table 5 online) of two sets of ssDNA probes (see cytometer that can detect only two colors. Supplementary Table 2 online) were created. One set of probes (capture In summary, we have developed a DL-DNA-based nanobarcode probes) was biotin-labeled and complementary to one part of its own target technology, whose applications have been demonstrated in uores- DNA. The other set of probes (report probes), which was complementary to cence microscopy, dot blotting and ow cytometry. The technology the other part of the target DNA, was attached to the nanobarcodes, thus has the following advantages: (i) DNA, in particular the DL-DNA, can establishing the code library (see Supplementary Table 2 online). be used as both a structural scaffolding and a functional probe; Microbead functionalization with DNA probes. The conjugates between (ii) multiplexed molecular sensing relies on the detection of precise microbeads and DNA probes were prepared using a modied protocol suguorescent color ratios instead of the detection of single colors; and gested by the manufacturer (Bangs Laboratories). Briey, 1.0 mg of streptavidin(iii) the DNA-based, multiplexed sensing platform nanotechnology coated polystyrene microbead suspension was washed with 100 ml of TTL can be applied to almost any uorescence-based detection system. buffer (100 mM Tris-HCl, pH 8.0, 0.01% Tween 20, 1M LiCl) and resuspended
104 103 FL1-H FL1-H 102 101 100
METHODS
Synthesis of DNA nanobarcodes. Each nanobarcode building block, the uorescence-labeled Y-DNA, consisted of three oligonucleotides (see Supplementary Tables 1 and 2 online), one of which had a nonpalindromic sticky end whereas the other two were either uorescence labeled (with Alexa Fluor 488 (Ex 495 nm and Em 519 nm) or BODIPY 630/650 (Ex 625 nm and Em 640 nm)) or attached to a DNA probe. The three DNA components were partially complementary to each other. After hybridization, they formed Y-DNA, which was used as the outermost peripheral layer of the nanobarcodes. Other nonuorescence labeled Y-DNA was used to link uorescence-labeled Y-DNA together. All Y-DNA were ligated to each other via their complementary sticky ends to form uorescence-labeled dendritic nanostructures (nanobarcodes) (Fig. 1b). The details of DNA sequence design, synthesis of Y-DNA and fabrication of dendrimer-like DNA have been published14. Gel electrophoresis. The DNA nanobarcodes were run in a 3% agarose ready gel (Bio-Rad) at 85 V at 25 1C in Tris-acetate-EDTA (TAE) buffer (40 mM Tris, 20 mM acetic acid and 1 mM EDTA, pH 8.0, Bio-Rad). After a true color picture of the gel was taken using a digital camera under strong UV illumination, it was stained with 0.5 mg/ml of ethidium bromide in TAE buffer. DNA
in 10 ml of TTL. One picomole of biotin-modied capture probes was then mixed with the microbead suspension and incubated at 25 1C with gentle agitation for 30 min. The excess and weakly bound probes were subsequently removed using sequential washes with 100 ml of TTL buffer, TT buffer (250 mM Tris-HCl, pH 8.0, 0.01% Tween 20), TTE buffer (250 mM Tris-HCl, pH 8.0, 0.01% Tween 20, 20 mM Na2(EDTA)) and TT buffer. The probe-functionalized microbeads were resuspended in prehybridization buffer (Church buffer: 0.5 M sodium phosphate, pH 8.0, 1 mM EDTA, 7% (wt/vol) SDS and 1% (wt/vol) bovine serum albumin) and incubated at 68 1C for 30 min. After the prehybridization buffer was removed, the microbeads were resuspended in hybridization buffer (1 SSC (150 mM sodium chloride, 15 mM sodium citrate), 1% SDS). Other DNA probes were conjugated to microbeads similar to the method described above. Sample preparation for uorescence microscopy imaging. The sample for uorescence microscopy analysis was prepared by thoroughly mixing bead/ probe conjugates along with a sample containing unknown target(s) and nanobarcodes in 400 ml of hybridization buffer to ensure uniform binding. The hybridization was performed with gentle agitation at 25 1C in the dark for 2 h. The sample was then washed with 400 ml of hybridization buffer three times to remove excess and weakly bound nanobarcodes. The bead suspension was
888
VOLUME 23
NUMBER 7
100
101
102
103
104
JULY 2005
LETTERS
concentrated to a nal concentration of 1,000 beads/ml. Around 2.5 ml of concentrated bead suspension was added onto a glass slide. A cover slip was glued onto the glass slide using nail polish. After that, the sample was imaged with uorescence microscopy at 1,000-fold magnication using green and farred lters, and the images were analyzed using MetaMorph software. Dot blotting assay. Around 0.5 ml of DNA solution (about 20 mM) including controls of a 15-mer oligonucleotide with irrelevant sequences and a 6.1 kb plasmid DNA was loaded onto a Zeta-probe membrane (Bio-Rad). After the membrane was air-dried, DNA molecules on the membrane were cross-linked with a UV crosslinker (Stratagene). The membrane was then prehybridized with Church buffer at 68 1C for 2 h. After prehybridization, the buffer was removed, and the membrane was submerged into a hybridization buffer (1 SSC buffer containing 1% SDS) containing nanobarcodes and incubated overnight at 25 1C. After hybridization, the membrane was evaluated with uorescence microscopy. Sample preparation for ow cytometry. The microbeads used for ow cytometer analysis were purposely prepared nonuniformly in terms of the number of target DNA bound to each microbead (since only a small amount of target DNA was used, microbeads were far from saturation in terms of nanobarcode binding). Bead-probe conjugates, along with a sample containing unknown DNA target(s) and nanobarcodes were added into 400 ml of hybridization buffer individually without mixing to achieve a nonuniform nanobarcode binding. The resultant microbead suspension was incubated at 25 1C in the dark for 2 h. The sample was then analyzed using a ow cytometer (BD FACSCalibur) with green (FL1H) and far-red channels (FL4H).
Note: Supplementary information is available on the Nature Biotechnology website.
3. Steemers, F.J., Ferguson, J.A. & Walt, D.R. Screening unlabeled DNA targets with randomly ordered ber-optic gene arrays. Nat. Biotechnol. 18, 9194 (2000). 4. Braeckmans, K., De Smedt, S.C., Leblans, M., Pauwels, R. & Demeester, J. Encoding microcarriers: Present and future technologies. Nat. Rev. Drug Discov. 1, 447456 (2002). 5. Duggan, D., Bittner, M., Chen, Y., Meltzer, P. & Trent, J. Expression proling using cDNA microarrays. Nat. Genet. 21, 1014 (1999). 6. DeRisi, J. et al. Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat. Genet. 14, 457460 (1996). 7. Schena, M., Shalon, D., Davis, R. & Brown, P. Quantitative monitoring of geneexpression patterns with a complementary-DNA microarray. Science 270, 467470 (1995). 8. Cheung, V.G. et al. Making and reading microarrays. Nat. Genet. 21, 1519 (1999). 9. Cunin, F. et al. Biomolecular screening with encoded porous-silicon photonic crystals. Nat. Mater. 1, 3941 (2002). 10. Wang, J., Liu, G.D. & Rivas, G. Encoded beads for electrochemical identication. Anal. Chem. 75, 46674671 (2003). 11. Chan, W.C. et al. Luminescent quantum dots for multiplexed biological detection and imaging. Curr. Opin. Biotechnol. 13, 4046 (2002). 12. Nicewarner-Pena, S.R. et al. Submicrometer metallic barcodes. Science 294, 137 141 (2001). 13. Nam, J.M., Thaxton, C.S. & Mirkin, C.A. Nanoparticle-based bio-bar codes for the ultrasensitive detection of proteins. Science 301, 18841886 (2003). 14. Li, Y. et al. Controlled assembly of dendrimer-like DNA. Nat. Mater. 3, 3842 (2004). 15. Luo, D. The road from biology to materials. Mater. Today 6, 3843 (2003). 16. Ried, T., Baldini, A., Rand, T.C. & Ward, D.C. Simultaneous visualization of 7 different DNA probes by in situ hybridization using combinatorial uorescence and digital imaging microscopy. Proc. Natl. Acad. Sci. USA 89, 13881392 (1992). 17. Taton, T., Mirkin, C. & Letsinger, R. Scanometric DNA array detection with nanoparticle probes. Science 289, 17571760 (2000). 18. Sjostedt, A., Eriksson, U., Berglund, L. & Tarnvik, A. Detection of Francisella tularensis in ulcers of patients with tularemia by PCR. J. Clin. Microbiol. 35, 10451048 (1997). 19. Sanchez, A. et al. Detection and molecular characterization of Ebola viruses causing disease in human and nonhuman primates. J. Infect. Dis. 179, S164S169 (1999). 20. Poon, L.L. et al. Detection of SARS coronavirus in patients with severe acute respiratory syndrome by conventional and real-time quantitative reverse transcription-PCR assays. Clin. Chem. 50, 6772 (2004). 21. Fuja, T., Hou, S. & Bryant, P. A multiplex microsphere bead assay for comparative RNA expression analysis using ow cytometry. J. Biotechnol. 108, 193205 (2004). 22. Vignali, D.A. Multiplexed particle-based ow cytometric assays. J. Immunol. Methods 243, 243255 (2000). 23. Wang, J. et al. Adsorption and detection of DNA dendrimers at carbon electrodes. Electroanal. 10, 553556 (1998). 24. Wang, J., Jiang, M., Nilsen, T. & Getts, R. Dendritic nucleic acid probes for DNA biosensors. J. Am. Chem. Soc. 120, 82818282 (1998). 25. Nilsen, T., Grayzel, J. & Prensky, W. Dendritic nucleic acid structures. J. Theor. Biol. 187, 273284 (1997). 26. Capaldi, S., Getts, R.C. & Jayasena, S.D. Signal amplication through nucleotide extension and excision on a dendritic DNA platform. Nucleic Acids Res. 28, e21 (2000). 27. Lowe, M., Spiro, A., Zhang, Y. & Getts, R. Multiplexed, particle-based detection of DNA using ow cytometry with 3DNA dendrimers for signal amplication. Cytom Part A 60A, 135144 (2004). 28. Shchepinov, M., Mir, K., Elder, J., Frank-Kamenetskii, M. & Southern, E. Oligonucleotide dendrimers: stable nano-structures. Nucleic Acids Res. 27, 30353041 (1999). 29. Shchepinov, M., Udalova, I., Bridgman, A. & Southern, E. Oligonucleotide dendrimers: synthesis and use as polylabelled DNA probes. Nucleic Acids Res. 25, 44474454 (1997).
ACKNOWLEDGMENTS We wish to acknowledge the Cornell Center for Advanced Technology and the Cornell Center for Vertebrate Genomics for nancial support. This material is based upon work supported in part by the Science and Technology Center Program of the National Science Foundation under agreement no. ECS9876771. We thank Yung-Fu Chang for providing pathogen genomic DNA (Mycobacterium avium subsp. paratuberculosis) and Carol Bayles for technical support on microscopy. COMPETING INTERESTS STATEMENT The authors declare competing nancial interests (see the Nature Biotechnology website for details).
Received 22 February; accepted 5 May 2005 Published online at http://www.nature.com/naturebiotechnology/
1. Han, M., Gao, X.H., Su, J.Z. & Nie, S. Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules. Nat. Biotechnol. 19, 631635 (2001). 2. Fulton, R.J., McDade, R.L., Smith, P.L., Kienker, L.J. & Kettman, J.R. Advanced multiplexed analysis with the FlowMetrix(TM) system. Clin. Chem. 43, 17491756 (1997).
VOLUME 23
NUMBER 7
JULY 2005
889
PROTOCOL
Dendrimer-like DNA-based uorescence nanobarcodes

Soong Ho Um, Jong Bum Lee, Sang Yeon Kwon, Yougen Li & Dan Luo
Department of Biological and Environmental Engineering, Cornell University, Ithaca, NY 14853-5701, USA. Correspondence should be addressed to D.L. (DL79@cornell.edu) Published online 10 August 2006; doi:10.1038/nprot.2006.141
2006 Nature Publishing Group http://www.nature.com/natureprotocols
A major challenge in clinical diagnostics and environmental analysis is the difculty in rapid and sensitive detection of multiple target molecules simultaneously (i.e., multiplexed detections). Our group has designed and synthesized a dendrimer-like DNA (DL-DNA) that is multivalent and anisotropic; using this unique DNA structure, we have developed a uorescence-tagged nanobarcode system for multiplex detection. This nanobarcode system allows the rapid and sensitive detection of multiple pathogens simultaneously using the ratios of two different uorescent dyes, green and red, with which different DL-DNAs are labeled. The key step of our nanobarcode model lies in the monodisperse preparation of DL-DNA. Two methods, solution phase and solid phase, are presented here. With slight modications, this platform technology can also be extended to the multiplexed detection of RNA and proteins. This protocol can be completed in 25 d.
INTRODUCTION Detecting a large number of molecules simultaneously, or multiplexed detection, has become increasingly important in a variety of elds, including in pathogen sensing, gene proling, drug discovery and environmental analysis. In particular, the multiplexed evaluation of biomolecules (such as DNA or proteins) is an integral part of genomic/proteomic research and has important implications for enabling multiple applications in both basic research and clinical settings. Fluorescence is arguably the most used read-out signal in the biotechnology and biomedical elds. In many cases, at most, only about four different uorescent colors can be detected simultaneously due to both the overlapping of emission spectra of different uorodyes and the limitation of available optical lters. Thus, one of the great challenges in developing uorescence-based, multiplexed analysis is to simultaneously identify the many individual targets in a test sample with a limited choice of uorescence labeling, and most importantly, with currently available equipment, including optical lters. One feasible answer is to discriminately assign each potential reaction with a unique uorescence code in advance. These unique codes can be physical positions (as in DNA microarrays) or signaling molecules (as in DNA nanobarcodes1,2, as detailed in this protocol). After an assay is carried out, targets can then be identied with pre-assigned codes. A detailed, recent review of multiplexed molecular detection has previously been published3. In this protocol, we have developed two different preparation methods for the synthesis of DNA nanobarcodes: solution-phase synthesis and solid-phase synthesis. In both cases, a uorescencelabeled Y-shaped DNA (Y-DNA) is ligated to other complementary Y-DNA molecules using T4 DNA ligase. Although solution phase is
TABLE 1 | Eight empirical rules for designing new oligonucleotides.
simpler to perform, solid phase is better due to the fact that synthesis and purication are coupled. After coupling with microbeads, nanobarcode signals can be amplied and detected, allowing the rapid and simultaneous detection of multiple pathogens with attomole or even zeptomole resolution. Branched DNA molecules and their design The key to making DNA nanobarcodes is the creation of anisotropic, tree (dendrimer)-shaped DNA (DL-DNA) molecules that serve as a scaffold to precisely carry a pre-assigned uorescence ratio. To build up a DL-DNA, the sequence design of branched DNA is of critical importance. Branched DNA exist naturally in replication forks, Holliday junctions and other transient processes. Seeman pioneered DNA articial architectures, including 3-way and 4-way junctions46. Our Y-DNA design was inspired by Seemans work, and eight empirical rules have been developed (Table 1). Other methods of multiplexed detection Microspheres that are encapsulated with different uorescent dyes (with different but specic ratios) have been extensively used for multiplexed molecular detection using ow cytometry. Luminescent quantum dots (QDs) have size-tunable, narrow emission spectra with similar excitation wavelengths; their applications in multiplexed molecular detection have been continuously expanding. Note that DNA nanobarcodes can easily incorporate QDs and thus can be further enhanced in their multiplexed capacity. More methods can be found in the review by Li and Luo3.
1. Calculate the free energy (Delta G) for a sequence. A lower free energy is desired; however, intermediatelow DG is also considered. Many websites provide online tools for the calculations (e.g., http://www.idtdna.com). 2. Consider the secondary structure of the molecule. In general, the least amount of secondary structure is desired. 3. Consider the dimerization, triplexation and even Z-DNA formation. The molecules should not form a self-dime, a triplex or a Z-DNA. 4. Consider the length. It should be long enough to form a stable DNA structure (at least more than 8 nucleotides (nt) long). 5. Consider the helix geometry. Half-turns are the quantum of the design (5*n bp, where n 0, 1, 2 y are between junctions). 6. Consider the G/C content. In general (varies by design), sequences are routinely chosen that constitute about 50% G/C. 7. Consider that the sequence symmetry (e.g., as those occurred in Holliday Junctions) of each arm should be avoided. 8. Consider non-WatsonCrick base pairing. Sequences containing more than two consecutive Gs should be avoided.
NATURE PROTOCOLS | VOL.1 NO.2 | 2006 | 995
PROTOCOL
TABLE 2 | Sequences of oligonucleotides. Strand Y0a Y1a Y2a Y3a Y4a Y0b Y1b Y2b Y3b Y4b Y0c Y1c Y2c Y3c Y4c Strand Spacer 1 Spacer 2 Segment 1 5-p-TGAG 5-p-GTCA 5-p-ACTG 5-p-ATGC 5-p-GACA 5-p-TGAC 5-p-CAGT 5-p-GCAT 5-p-TGTC 5-p-GGAT 5-p-TGAC 5-p-CAGT 5-p-GCAT 5-p-TGTC 5-p-GGAT Segment 1 Biotin-5-p 5-p-TCA Segment 2 TGATCCGCATGACATTCGCCGTAAG-3 TGGATCCGCATGACATTCGCCGTAAG-3 TGGATCCGACGTACATTCGCCGTAAG-3 TGGATCCGCATGACATTCGCCGTAAG-3 TGGATCCGCATGACATTCGCCGTAAG-3 CTTACGGCGAATGACCGAATCAGCCT-3 CTTACGGCGAATGACCGAATCAGCCT-3 CTTACGGCGAATGACCGAATCAGCCT-3 CTTACGGCGAATGACCGAATCAGCCT-3 CTTACGGCGAATGACCGAATCAGCCT-3 AGGCTGATTCGGTTCATGCGGATCAC-3 AGGCTGATTCGGTTCATGCGGATCCA-3 AGGCTGATTCGGTTACGTCGGATCCA-3 AGGCTGATTCGGTTCATGCGGATCCA-3 AGGCTGATTCGGTTCATGCGGATCCA-3 Segment 2 CCGGATAAGGCGCAGCGGTCGGCTGAATTCA GGGTTCGTGGCAGGCCAGCACACTTGGAGACCGAAGCTTACCGGACTCCTAAC-3 GTTAGGAGTCCGGTAAGCTTCGGTCTCCAAGTGTGCTGGCCTGCCACGAACCCTGAATTCAGCCGACCGCTGCGCCTTATCCGG-3
p indicates the phosphate and 5 corresponds to the 5 prime end of the DNA sequence.
Limitation of DNA nanobarcodes DNA nanobarcoding is not a high-throughput method. Thus, it can only be used for low-throughput. The current version of the DNA nanobarcode can be used to detect up to eight different pathogens simultaneously. With further development, we foresee that DNA nanobarcodes could be used to detect up to several dozen targets at the same time. In addition, although the DNA nanobar-
code can be decoded with current, existing equipment (as a matter of fact, any uorescence-based equipment), it is still a lab-based technique. A portable reader is urgently needed for bedside or eld usage. Moreover, an antibody-probed DNA nanobarcode is also urgently needed, so that sample-processing time can be reduced markedly compared with lengthy gene-probed detections.
MATERIALS
. Oligonucleotides for Y-DNA building blocks (Integrated DNA Technologies;

see Table 2) . Oligonucleotides for biotin-modied, spacer double-stranded DNA (dsDNA) (Integrated DNA Technologies; see Table 2) . Alexa Fluors 488 (green color) and BodipyTM 630/650-X (red color) (to functionalize oligonucleotides) (Sigma-Aldrich) . Annealing buffer: 10 mM Tris (pH 8.0), 1 mM ethylenediaminetetraacetic acid (EDTA) and 50 mM NaCl (Sigma-Aldrich) . Phosphate-buffered saline (PBS) solution (pH 7.4): 10 mM phosphate buffer, 137 mM NaCl and 2.7 mM KCl (Sigma-Aldrich) . Restriction endonuclease DdeI (Promega) . 10 restriction enzyme reaction buffer D (Promega) . T4 DNA ligase and 10 ligase buffer (Promega) . Bovine serum albumin (BSA; Promega) . Streptavidin-coated polystyrene beads with 5.5 mm diameter (Bangs Laboratories) . ImmunoPures Immobilized Avidin Gel with a cross-linked 6% beaded agarose (Pierce Biotechnology)
REAGENTS
. TTL buffer: 100 mM Tris-HCl (pH 8.0), 0.01% Tween 20 and 1.0 M LiCl . TT buffer: 250 mM Tris-HCl (pH 8.0) and 0.01% Tween 20 (Sigma-Aldrich) . TTE buffer: 250 mM Tris-HCl (pH 8.0), 0.01% Tween 20 and 20 mM EDTA . Prehybridization buffer: 0.5 M sodium phosphate (pH 8.0), 1 mM EDTA, . Hybridization buffer, 1 SSC: 150 mM sodium chloride, 15 mM sodium . Nuclease-free, sterile Milli-Q water
EQUIPMENT citrate and 1% SDS (Sigma-Aldrich) 7% (w/v) sodium dodecyl sulfate (SDS) and 1% (w/v) BSA (Sigma-Aldrich) (Sigma-Aldrich) (Sigma-Aldrich)
. Equipment and gels for agarose gel electrophoresis (Bio-Rad) . Thermal cycler (Eppendorf) . Flow cytometer (BD Biosciences) . Spectrophotometer (Eppendorf) . Rotary incubator (Fisher Scientic)
PROCEDURE 1| Design three oligonucleotides such that two of the oligonucleotides have partial complementary sequences, thus forming one arm of a Y-DNA, and the remaining third oligonucleotides have complementary sequences to the other two oligonucleotides. These three oligonucleotides should eventually hybridize, forming a Y-DNA. For the preparation of nanobarcodes, these should be designed to produce a Y-DNA that contains four bases as a sticky end on one arm and the appropriate uorescence-labeled tags on the other two arms. m CRITICAL STEP The sequence design is highly important in order to obtain the complete, expected DNA structure. The method is explained in previous publications1,79, and Tables 1 and 2 also give further details. For the preparation of nanobarcodes, it must be decided whether non-puried oligonucleotides or puried oligonucleotides should be used, depending on the preparation method of DL-DNA.
996 | VOL.1 NO.2 | 2006 | NATURE PROTOCOLS
PROTOCOL
TABLE 3 | Oligonucleotide hybridization program. Cycle no. 1 2 342 Denaturation 2 min at 95 1C Cooling 2 min at 65 1C Annealing 5 min at 60 1C 30 s at 60 1C 30 s at 600.1 1C (starting at cycle number 2)
Note that for the second annealing (starting with cycle number 2), this step is repeated 40 times, each time 0.1oC/cycle lower than the previous cycle.
2| Synthesize or purchase the polyacrylamide gel electrophoresis-puried 5-end phosphorylated oligonucleotides according to the designed sequences, then dissolve them in an annealing buffer at a nal concentration of 0.2 mM for solution-phase synthesis and nanobarcode preparation.
3| In a sterile 0.5 mL microcentrifuge tube, mix the three oligonucleotides at an equal molar ratio in sterile Milli-Q water at a nal concentration of 40 mM (6 mM for nanobarcode preparation) for each oligonucleotide in a total volume of 500 mL. 4| Place the tube in a thermocycler and hybridize the oligonucleotides using the program outlined in Table 3. 5| Prepare DL-DNA by either solution-phase synthesis (A) or solid-phase synthesis (B), or prepare nanobarcodes for ow cytometry detection (continue to Step 6). (A) Solution-phase synthesis of DL-DNA (i) To synthesize the rst-generation DL-DNA, ligate the Y1-DNA (subscript 1 refers to the rst generation) to the Y0-DNA (subscript refers to the core generation, or zero generation) in a 3:1 molar ratio by mixing the following in a sterile 0.5 mL microcentrifuge tube: T4 DNA ligase, 5 mL; 10 ligase buffer, 10 mL; Y0-DNA (40 mM), 10 mL; Y1-DNA (40 mM), 30 mL; Milli-Q water, 45 mL. Incubate at room temperature for 16 h. m CRITICAL STEP The T4 DNA ligase buffer contains ATP molecules, which provide energy for the ligase enzyme activity. A high temperature may cause rapid deformation of the ATP into ADP, thus reducing the activity of the ligase. In addition, ligase storage buffer contains a high concentration of glycerol, which may adversely affect enzyme activity if it accumulates to a high level. Both scenarios may lead to poor yield and low purity as the generation of the DL-DNA produced increases. Additional purication steps (such as dialysis) to eliminate glycerol may be required when preparing high generations of DL-DNA. (ii) To synthesize the second generation DL-DNA, ligate the Y2-DNA to the G1 DL-DNA in a 6:1 molar ratio by mixing the following in a sterile 0.5 mL microcentrifuge tube: T4 DNA ligase, 5 mL; 10 ligase buffer, 10 mL; G1-DNA (4 mM), 10 mL; Y2-DNA (40 mM), 6 mL; Milli-Q water, 69 mL. Incubate at room temperature for 16 h. PAUSE POINT Note that the nomenclature of DL-DNA is as follows: the core of the dendrimer Y0 is designated as G0, which is the zeroth generation of DL-DNA. After Y0 is ligated with three Y1, the dendrimer is termed the rst generation of DL-DNA (G1), and so on. The nth generation of DL-DNA is noted as Gn. (iii) Check the nal product (molecular weight and purity) using a gel electrophoretic migration shift assay (GEMSA). ! CAUTION If ethidium bromide (EtBr; a potent carcinogen) is used to stain the gel, wear gloves. We recommend using SYBR I for dsDNA and SYBR II for ssDNA staining (Invitrogen). PAUSE POINT DL-DNA is stable at 4 1C for at least several months. (B) Solid-phase synthesis of DL-DNA (i) Design two complementary oligonucleotides: one of 84 bases long and biotinylated on its 5-end, and the other of 87 bases long with a 5-end sticky end that is complementary to one arm of a Y-DNA following hybridization. The spacer DNA should contain sequences of CTNAG, which can be recognized by DdeI endonuclease. This spacer DNA will act as a bridge for the Y-DNA and the solid support. m CRITICAL STEP The sequence design is highly important in order to obtain the complete, expected DNA structure. The method is explained in previous publications1,79, and Table 1 provides further details. (ii) Dissolve the synthesized spacer DNA in 1 PBS buffer at a nal concentration of 0.2 mM for each oligonucleotide. (iii) Measure the concentration of each complementary oligonucleotide and then x the concentration of each oligonucleotide at 60 mM in a total volume of 400 mL. (iv) Place the tube in a thermocycler and hybridize the oligonucleotides using the program outlined in Table 4. (v) Pretreat the surface of the avidin-coated agarose solid support with 0.1% of 20% (v/v) ionic SDS surfactant to reduce nonspecic binding of DNA to the solid support. (vi) Add 150 mL (8.2 nmole) of the biotin-modied spacer DNA to the solution containing 100 mL of avidin-coated agarose support and incubate at room temperature overnight (Step 1 in Fig. 1). (vii) Wash the solid support thoroughly with 1 PBS buffer and Milli-Q water and centrifuge the solid support at 2500g for 10 min. Discard the supernatant and repeat the wash until the supernatant shows background absorbance at 260 nm when measured using a UV-Vis spectrophotometer. This measurement ensures that there is no oligonucleotide in this solution (Step 1 in Fig. 1).
NATURE PROTOCOLS | VOL.1 NO.2 | 2006 | 997
PROTOCOL
TABLE 4 | Oligonucleotide hybridization program. Cycle number 1 2 Denaturation 4 min at 95 1C First annealing 2 min at 80 1C Second annealing 1 hr at 25 1C
(viii) Ligate 150 mL (8.0 nmole) of Y-DNA solution with the spacer DNA attached to the solid support using 2.1 Weiss unit of T4 DNA ligase. Depending on the intended generation of Gn DL-DNA, Y-DNA is successively added into the solid support. For instance, if G2 DL-DNA (second generation of DL-DNA) is intended to be a nal product, Y1-DNA to Y2-DNA are successively added into the spacer-modied solid support to form G2 DL-DNA (Steps 26 in Fig. 1). m CRITICAL STEP The T4 DNA ligase buffer contains ATP molecules, which provide energy for the ligase enzyme activity. A high temperature may cause rapid deformation of the ATP into ADP, thus reducing the activity of the ligase. In addition,
BIOTINspacerDNA (BIO-5p-SP) HN
O NH 5-phos CCGG GGCC AAC 3 TTGACT5 Bio-5p-CCGG GGCC Yn-DNA

Agarose support
S O Avidin-coated agarose support

A A A
AAC 3 TTGACT 5
Bio-5p-SP
A A A Avidin TGAG C 3
CG
Agarose support
A A A
3 C
AT C G
AT
TGA
CG AT
AT
Agarose support
Agarose support
A BIOTINspacerDNA Step 1 A
Bio-5p-CCGG GGCC
AAC 3 TTGACT 5
Y0-DNA Step 2
Bio-5p-CCGG GGCC
AAC 3 TTGACT 5
C A G T
Agarose support
Agarose support
A Bio-5p-CCGG AACTGAG GGCC TTGACTC
CA
GT
A Bio-5p-SPY -DNA 0
TG
Y1-DNA Step 3
TG
AC
Y2-DNA Step 4
A Bio-5p-CCGG AACTGAG GGCC TTGACTC
CA G T
A Bio-5p-SPY0-DNA
TG A C
TG AC
C G
TA
Agarose support
A Bio-5p-SPY -DNAY -DNA 0 1
Agarose support
TA
A Bio-5p-SPY -DNAY -DNAY -DNA 0 1 2
CG
Y3-DNA Step 5
CT
TA
A Bio-5p-SPY0-DNAY1-DNA
TA
A Bio-5p-SPY0-DNAY1-DNAY2-DNA
CG
Agarose support
Y4-DNA Step 6
Bio-5p-SPY0-DNAY1-DNAY2-DNAY3-DNA
CT
TG T
GT
DDE I (RE) Step 7 TGA
A Bio-5p-SPY0-DNAY1-DNAY2-DNAY3-DNA
CT
Figure 1 | Preparation of dendrimer-like-DNA using solid-phase synthesis. 998 | VOL.1 NO.2 | 2006 | NATURE PROTOCOLS
TG
GT
C TG T
CT GT
TG
GT
PROTOCOL
ligase storage buffer contains a high concentration of glycerol, which may adversely affect enzyme activity if it accumulates to a high level. Both scenarios may lead to poor yield and low purity as the generation of the DL-DNA produced increases. Additional purication steps (such as dialysis) to eliminate glycerol may be required when preparing high generations of DL-DNA. (ix) Obtain the synthesized DL-DNA by releasing the products from the solid support using a DdeI restriction enzyme, which recognizes a specic site on the spacer DNA (5yC.TNAGy3). The restriction enzyme solution contained 100 units of DdeI and 4 mL of BSA in a restriction buffer D (Step 7 in Fig. 1). (x) Check the nal product using a GEMSA (Fig. 2). Also, evaluate the reaction efciency of the synthesis of the DL-DNA using PicoGreen dsDNA quantication assay kit (Molecular Probes). ? TROUBLESHOOTING
Preparation of DL-DNA-based nanobarcodes for ow cytometry detection 6| To make the rst-generation DL-DNA nanobarcode (Fig. 3), ligate three uorescence-labeled Y-DNA to the Y0-DNA with appropriate sticky ends by mixing the following in a sterile 0.5 mL microcentrifuge tube: T4 DNA ligase, 2 mL; 10 ligase buffer, 5 mL; each of four 6 mM Y-DNA, 10 mL; Milli-Q water, 3 mL. Incubate at room temperature for 16 h. 7| Prepare 1 mL of streptavidin-coated polystyrene microbead suspension in 100 mL of TTL buffer. Centrifuge the tube at 2500g for 3 min and then carefully discard the supernatant. 8| Add 10 mL TTL buffer into the beads and then add 1 pmole of biotin-modied capture probes. Incubate with gentle mixing at 25 1C for 30 min using a rotary incubator. 9| Thoroughly wash the beads successively with 100 mL TTL buffer, 100 mL TT buffer, 100 mL TTE buffer and nally with 100 mL TT buffer until the supernatant shows background absorbance at 260 nm when measured using a spectrophotometer. 10| Incubate the probe-functionalized microbeads in prehybridization buffer at 68 1C for 30 min and then remove the prehybridization buffer after centrifuging the beads at 2500g for 3 min. Resuspend the microbeads in 100 mL of hybridization buffer, 1 SSC. 11| Incubate the DNA targets and nanobarcodes together in 400 mL of hybridization buffer along with the probe-functionalized microbeads that were prepared in Step 7 at 25 1C for 2 h in the dark with gentle mixing (Fig. 4). ! CAUTION Exposure to visible light will cause photobleaching of the uorescence dyes. 12| Analyze the samples using a ow cytometer.
Figure 2 | Higher-generation dendrimer-like DNA on the solid phase. Evaluation of higher-generation dendrimer-like (DL)-DNA formation. Lanes 2, 3, 4, 5 and 6 correspond to G0 DL-DNA, G1 DL-DNA, G2 DL-DNA, G3 DL-DNA and G4 DL-DNA, respectively.
Pr
ob
Pr
ob
Hybridization
Ligation
Y-DNA
Figure 3 | Schematics of the synthesis of uorescence-labeled Y-DNA and the formation of a DNA nanobarcode. (a) Synthesis of uorescence-labeled Y-DNA. (b) Formation of a DNA nanobarcode. NATURE PROTOCOLS | VOL.1 NO.2 | 2006 | 999
PROTOCOL
TIMING
Avidin-coated polystyrene bead Nanobarcode with report probes
D-Biotin DNA target
Capture probe
Solution-phase synthesis of Y-DNA should take about 1 h Solid-phase synthesis of Y-DNA should take about 3 h Solution-phase synthesis of DL-DNA should take from 1 d (rst generation) to 3 d (fourth generation) Solid-phase synthesis of DL-DNA should take from 1 d (rst generation) to 3 d (fourth generation) DNA nanobarcode construction should take about 1 h to assemble Y-DNA Preparation of DNA nanobarcodes should take about 12 d Hybridization and detection should take about 2 h (much faster than chip-based methods)
Immobilization
Hybridization
Figure 4 | A dendrimer-like-DNA nanobarcode for use in ow cytometry detection.
? TROUBLESHOOTING Troubleshooting advice can be found in Table 5.

TABLE 5 | Troubleshooting table. PROBLEM Step 5B(x) Very low yield. SOLUTION First, check all DNA sequences used. Second, compare the concentration of supernatants of the spacer DNA from Step 5B(vi) with the initial concentration of the added spacer DNA. The DNA concentration of the supernatant should be relatively lower. Third, use fresh ligases and restriction enzymes.
ANTICIPATED RESULTS Constructing branched DNA (including Y-DNA and DL-DNA) is relatively straightforward and simple. Using high-performance liquid-chromatography-puried oligonucleotides as the starting materials, a single band on an electrophoresis gel is expected (see Fig. 2). Note that a slight smear is also expected, especially if the gel is stained with a super-sensitive dye (such as SYBR instead of EtBr). Labeling oligonucleotides with uorescent molecules is better accomplished commercially unless the person performing the process is experienced in bioconjugations with the necessary specic dyes. In this case, caution should be placed on monovalent labeling. Flow cytometry detection of DNA nanobarcodes is also straightforward. Linear lines are expected (Fig. 5), although some variations in line width may be observed2. This will not interfere with decoding as that procedure is accomplished by taking the intercept after curve tting. Commercially available data-processing software may be helpful here.
a 104
103 FL1H
b 104
103 FL1H 101 102 FL4H
22
102
102
101
101
100 100
103
104
100 100
101
11
102 FL4H
44
103
180
104
Figure 5 | Multiplexed DNA nanobarcode detection using ow cytometry. (a) A two-color ow plot of microbeads attached with the nanobarcode 2G1R as a control for standards (a calibration control). FL1H indicates the green channel and FL4H is the red channel. (b) Simultaneous detection of three pathogen DNA using nanobarcodes. Unrelated DNA sequences were not detected (background). (The gures appeared in Nature Biotechnology 23, 885887, 2005).
ACKNOWLEDGMENTS We wish to acknowledge USDA, Cornell Advanced Technology Centre for Biotechnology and the Cornell Nanobiotechnology Center (an STC Program of of the National Science Foundation under Agreement No. ECS-9876771) for nancial support. COMPETING INTERESTS STATEMENT The authors declare competing nancial interests (see the HTML version of this article for details). Published online at http://www.natureprotocols.com Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions 1. Li, Y. et al. Controlled assembly of dendrimer-like DNA. Nature Mater. 3, 3842 (2004). 2. Li, Y., Cu, Y.T. & Luo, D. Multiplexed detection of pathogen DNA with DNA-based uorescence nanobarcodes. Nature Biotechnol. 23, 885889 (2005).
3. Li, Y. & Luo, D. Multiplexed molecular detection using encoded microparticles and nanoparticles. Exp. Rev. Mol. Diagn. (in press). 4. Seeman, N.C. & Kallenbach, N.R. Design of immobile nucleic acid junctions. Biophys. J. 44, 201209 (1983). 5. Kallenbach, N.R., Ma, R.I. & Seeman, N.C. An immobile nucleic-acid junction constructed from oligonucleotides. Nature 305, 829831 (1983). 6. Ma, R.I., Kallenbach, N.R., Sheardy, R.D., Petrillo, M.L. & Seeman, N.C. Three-arm nucleic acid junctions are exible. Nucleic Acids Res. 14, 97459753 (1986). 7. Luo, D., Cu, Y.T., Li, Y. & Um, S.H. in Chapter 10, DNA Vaccines: Methods and Protocols (eds Saltzman, W.M., Shen, H. & Brandsma, J.L.) (Humana Press, New Jersey, 2006). 8. Seeman, N.C. Nucleic acid junctions and lattices. J. Theor. Biol. 99, 237247 (1982). 9. Seeman, N.C. De novo design of sequences for nucleic acid structural engineering. J. Biomol. Struct. Dyn. 8, 573581 (1990).
1000 | VOL.1 NO.2 | 2006 | NATURE PROTOCOLS
LETTERS
Enzyme-catalysed assembly of DNA hydrogel

SOONG HO UM1 , JONG BUM LEE1 , NOKYOUNG PARK1 , SANG YEON KWON1 , CHRISTOPHER C. UMBACH2 AND DAN LUO1 *
1 2
Department of Biological and Environmental Engineering, Cornell University, Ithaca, New York 14853-5701, USA Department of Materials Science and Engineering, Cornell University, Ithaca, New York 14853-5701, USA * e-mail: DL79@cornell.edu
Published online: 24 September 2006; doi:10.1038/nmat1741
NA is a remarkable polymer that can be manipulated by a large number of molecular tools including enzymes1 . A variety of geometric objects, periodic arrays and nanoscale devices have been constructed213 . Previously we synthesized dendrimer-like DNA and DNA nanobarcodes from branched DNA via ligases14,15 . Here we report the construction of a hydrogel entirely from branched DNA that are threedimensional and can be crosslinked in nature. These DNA hydrogels were biocompatible, biodegradable, inexpensive to fabricate and easily moulded into desired shapes and sizes. The distinct dierence of the DNA hydrogel to other bioinspired hydrogels (including peptide-based, alginate-based and DNA (linear)-polyacrylamide hydrogels1620 ) is that the crosslinking is realized via ecient, ligase-mediated reactions. The advantage is that the gelling processes are achieved under physiological conditions and the encapsulations are accomplished in situdrugs including proteins and even live mammalian cells can be encapsulated in the liquid phase eliminating the drug-loading step and also avoiding denaturing conditions. Fine tuning of these hydrogels is easily accomplished by adjusting the initial concentrations and types of branched DNA monomers, thus allowing the hydrogels to be tailored for specic applications such as controlled drug delivery, tissue engineering, 3D cell culture, cell transplant therapy and other biomedical applications. Branched DNA molecules were designed and synthesized14,15 in such a way that each arm of the DNA molecule possessed a complementary sticky end whose sequences were palindromic. Thus, these branched DNA molecules were able to hybridize to and ligate with each other via T4 DNA ligase, serving as both monomers and crosslinkers. The self-assembly of branched DNA monomers (BDM, Fig. 1a, see also Supplementary Information, Fig. S1) coupled with ligase-catalysed reactions led to a largescale, three-dimensional structure that had the properties of a hydrogel (Fig. 1b). A swollen DNA hydrogel with X-DNA as monomers is shown in Fig. 2a. A DNA-specic uorescent dye (SYBR I) was used to stain the gel. After staining followed by extensive washing, the DNA hydrogel gave out an extremely intense, green uorescence (Fig. 2a, inset), strongly suggesting that the hydrogel was indeed composed
1 X-DNA
Y-DNA
T-DNA
Figure 1 Schematic diagram of BDM and DNA hydrogels. a, X-, Y- and T-DNA as BDM. b, BDM serve as crosslinkers to form networked gels.
of DNA molecules. Similar results were obtained with other BDM (Y and T). As expected, DNA hydrogels assembled from dierent
797
nature materials VOL 5 OCTOBER 2006 www.nature.com/naturematerials
LETTERS
a b
Figure 2 Hydrogels made entirely from branched DNA. a, A swollen X-DNA hydrogel fabricated in a cylindrical mould. The size is 7.0 mm in diameter and 3.0 mm in height. The scale bar is 1 cm. The inset shows the DNA gel stained with SYBR I. b, Images of dried (left) and swollen (right) X-DNA hydrogels with different patterns: rectangular, circular, triangular, star and cross (from the top left corner, clockwise). The scale bars are 1 cm. c, X-DNA gels patterned in CORNELL shapes at centimetre scale (top and middle rows; the scale bar is 1 cm) and micrometre scale (bottom row; the scale bar is 500 m). They were stained with two different, DNA-specic uorescent dyes: ethidium bromide (red, the middle row) and SYBR I (green, the bottom row).
BDM were shown to have unique chemical and physical properties. Dierent swelling proles of DNA hydrogels were achieved by adjusting the initial concentration and the types of BDM: the higher the initial concentration of the BDM, the higher the degree of swelling of each hydrogel (see Supplementary Information, Table S1). More specically, the Y-DNA-based hydrogel (Y-DNA gel) swelled more than 400% at the highest initial concentration of BDM (0.2 mM); whereas at the lowest initial concentration (0.03 mM), it swelled only about 100%. Besides the initial concentrations, the dierent types of BDM also inuenced the degree of swelling. X-DNA-based hydrogel (X-DNA gel) showed a higher swelling degree than both the Y-DNA gel and the T-DNA gel; these trends in the degree of swelling among X-, Y- and T-DNA gels were true for all of the concentrations tested. The mechanical properties of DNA hydrogels were easily tunable. Swollen X-DNA gel showed the strongest tensile modulus among all the DNA hydrogels (see Supplementary Information, Table S2), probably due to the fact that the crosslinked DNA molecules strongly resisted deformation. In contrast, the X-DNA gel showed the lowest tensile strength among all the DNA hydrogels, indicating that the crosslinked DNA molecules are more resistant to returning to their original shape at a given stress in the linear range. Overall, DNA hydrogels are soft materials whose mechanical properties can be easily tuned.
798
The size and the shape of the DNA hydrogels can be precisely controlled. The DNA hydrogels were patterned into rectangular, round, triangular, cross, star and even the word CORNELL shapes at both centimetre and micrometre scales (Fig. 2b,c). Interestingly, these DNA hydrogels repeatedly returned to their original shapes without collapsing to lms or powders even after successive drying and hydrating (Fig. 2b). Besides the size and shape, the morphology and internal structure of the DNA hydrogels vary markedly and can also be tuned based on the BDM. Using a variety of visualization methods including atomic force microscopy (AFM), eld-emission scanning electron microscopy (FE-SEM) and confocal microscopy, the surface morphology and the inner structure of each DNA hydrogel were studied in dried and swollen states, and images showed striking dierences depending on the types of BDM used (Fig. 3). In the dry state, surface morphology revealed a tangled pattern for X-DNA gel, a brous form for Y-DNA gel and a scale shape for T-DNA gel (Fig. 3a,b and c, respectively). In particular, the X-DNA gel (Fig. 3a) shows that two at DNA gel stripes were tangled into a knot to form a large sheet with many wrinkles on the surface. On the other hand, the Y-DNA gel (Fig. 3b) shows a brous form spreading out from many branches, which may have resulted from the shape of the Y-DNA. The T-DNA gel (Fig. 3c) looks like puckers on a sheet. In the swollen state, the surface morphology of the gels showed a large number of
LETTERS
a b c
nm 2.00
Section analysis
2.00 0
20.0
40.0 nm
60.0
80.0
Figure 3 External morphologies and internal structures of different DNA hydrogels. ac, FE-SEM images of dried X-DNA a, Y-DNA b and T-DNA c hydrogels. df, Confocal microscopic images of the swollen X-DNA d, Y-DNA e and T-DNA f hydrogels. g, AFM image of the X-DNA gel. Inset: A zoomed AFM image showing the pore size of the X-DNA gel. The scale bars are 200 m for a,d, and f, 50 m for b, 15 m for e, 10 m for c and 200 nm for g.
various-sized pores and channels (Fig. 3d), obvious bres with fractal shapes on the periphery (Fig. 3e) and perpendicularly erected, scale-like structures (Fig. 3f) for X-DNA, Y-DNA and T-DNA, respectively. AFM visualization revealed that the dry X-DNA gel has nanoscale holes approximately 12.3 1.4 nm in the network structure (Fig. 3g). The measured holes size is very close to the theoretically calculated and designed value (13.6 nm), indicating that (1) X-DNA monomers were linked with one another to form the cross-linked, network structure of the X-DNA gel and (2) the pore sizes can be accurately controlled. In contrast to the fairly regular pore spacing of the X-DNA gel, the Y-DNA gel and T-DNA gel showed random structures. In addition, the DNA hydrogels were biodegradable, and their degradability should also be dependent on and thus could be adjusted by the BDM. Degradation processes were evaluated by measuring daily DNA mass loss (Fig. 4a). The degradation proles diered drastically depending on the BDM and more interestingly, on whether the gels were empty or loaded. Overall, X-DNA gels were more resistant to degradation than T-DNA gels, and T-DNA
gels were more resistant than Y-DNA gels (Fig. 4a). After 14 days, over 70% of the X-DNA gels (both empty and loaded) remained, whereas 50% of the insulin-loaded Y-DNA and T-DNA gels remained. Empty Y-DNA and T-DNA gels, on the other hand, degraded rapidly after 2 days and 7 days, respectively, with only about 20% remaining after 2 weeks. Clearly, loaded gels were more resistant than empty gels. Interestingly, when the DNA gels were loaded with camptothecin (CPT), a DNA-binding drug, all DNA gels were protected from degradation. These results showed that, as expected, the degradation processes are determined by the internal structures of the DNA gels (for example, the BDM and emptiness), loaded drugs (for example, insulin or CPT) and the environment (for example, in the presence of nucleases). Inspired by these unique advantages of DNA hydrogels, we explored them for use as a long-term controlled drug-release system (Fig. 4b). Two dierent model drugs were encapsulated and tested: CPT (molecular mass = 348.3 daltons) and porcine insulin (molecular mass = 5,777.6 daltons). Encapsulations were accomplished in situ (no post-gelation loading was needed) and encapsulation eciencies close to 100% were achieved.
799
LETTERS
a
120 100 Remaining gel (%) 80 60 40 20 0 0 1 2 3 4 5 6 7 8 Time (days) 9 10 11 12 13 14
b
Accumulative release (%)
80 70 60 50 40 30 20 10 0 0 1 2 3 4 5 6 7 8 Time (days) 9 10 11 12 13 14
encapsulated in situ into the DNA hydrogel (see Supplementary Information, Fig. S2). Interestingly, after 3 days CHO cells were still live inside the DNA gel (data not shown). Note that it is conceivable that the encapsulated cells can be retrieved and collected from the gel at a later time by digesting out DNA matrices via nucleases. Also note that unlike plasmid DNA, which may cause host immune responses (mainly due to CpG modications from bacteria22,23 ), the DNA hydrogels are not expected to have immunogenicity owing to their purely synthetic nature. This biocompatibility is consistent with our cytotoxicity assays and animal tests (see Supplementary Information, Fig. S3). Considering the availability of a vast number of enzymes that can manipulate DNA at angstrom scale, our results strongly suggest that DNA can be engineered as a designer material whose properties can be more easily tuned. We have also used the DNA hydrogel format and achieved cell-free, protein manufacturing (data not shown; paper submitted elsewhere). These biodegradable, biocompatible, DNA hydrogels are a new class of materials that can be exploited in a variety of biomedical applications including sustained drug delivery, tissue engineering, 3D cell culture, cell transplant therapy and other biomedical applications.
METHODS
FABRICATION OF DNA HYDROGELS
Figure 4 Degradation and controlled drug-release proles of different DNA hydrogels. a, DNA hydrogel degradation proles. Remaining gel % refers to the percentage DNA mass remaining from the starting gels. The dotted lines indicate empty gels, dashed lines and solid lines represent insulin-loaded gels and CPT-loaded gels, respectively. The blue x symbols, red triangles and green squares indicate X-, Y- and T-DNA gels, respectively. b, Drug-release proles. A DNA monomer solution (150 l at 23.7 g l1 with 30 units of T4 DNA ligase) was mixed with 30 l of the stock solution of the drug (10 mg ml1 ). The amount of released insulin was determined using Mercodia porcine insulin ELISA enzyme immunoassay kits (ALPCO Diagnostics). The amount of released CPT was determined by ultraviolet absorbance at 370 nm (extinction coefcient: E mM = 19.9 (370 nm), Sigma-Aldrich). The solid lines indicate insulin-release proles and the dashed lines indicate CPT release. The black diamonds, black triangles and black squares indicate X-, Y- and T-DNA gels. The error bars represent standard deviation from at least three replicates.
BDM were fabricated according to our previously published methods14,15,24 . Briey, to construct an X-DNA gel, each X-DNA monomer (5.04 nmol) was ligated with 30 Weiss units of T4 DNA ligase (Promega). The reaction was carried out at room temperature overnight on a gentle rotator. A DNA gel in a micrometre-scale CORNELL shape was made by casting it in a polydimethylsiloxane mould that was fabricated by photolithography. To visualize the DNA gel micropatterns, the moulded gel was stained with the DNA-specic dye SYBR I (Molecular Probes).
MECHANICAL PROPERTIES
The mechanical property measurements were carried out on a Dynamic Mechanical Analyzer (DMA 2980, TA Instruments) with a cylindrical DNA gel (7.0 mm in diameter and 3.0 mm in height).
CHARACTERIZATION OF DNA HYDROGELS
AFM was carried out using the tapping mode on a surface-modied mica (via APTES, 3-aminopropyltriethoxysilane). For SEM imaging, DNA was metal-coated with Au/Pd.
IN VITRO DEGRADATION ASSAYS AND DRUG RELEASE
Furthermore, Fig. 4b shows the controlled release proles of CPT and insulin from 0.2 mM DNA hydrogels. No burst release was detected, and relatively smooth release curves were obtained. In particular, the release of the insulin was dependent on the types of BDM. After 12 days, about 60%, 40% and 30% of the insulin was released from the Y-, T- and X-DNA gels, respectively. The release dierences may be attributed to and thus can be controlled by the diverse internal structures of the DNA gels and the dierent degradation rates. On the other hand, CPT released from the DNA hydrogels showed a zero-order release prole for over a month, and there were no dierences among the release proles of the three dierent BDM. The linear and relatively slower release proles of CPT may have resulted from the smaller size of CPT and its high anity to both major and minor grooves of the DNA molecules21 . As the gelation process was entirely under physiological conditions (no organic solvent or high temperature), the DNA gels can be used for encapsulating live mammalian cells. Before gelation, we suspended live Chinese hamster ovarian (CHO) cells into 0.2 mM X-DNA solution. After adding ligase, CHO cells were
800
After weighing, dry DNA gels were totally swelled and the weight was recorded. The gels were then incubated in 300 l of PBS at 37 C with shaking. At predetermined time points, supernatants were removed immediately followed by weighing the sample and replenishing with fresh PBS. For controlled release assays25 , released drugs were assayed directly from the supernatant before replenishing.
CELL ENCAPSULATION AND BIOCOMPATIBILITY OF THE DNA GEL
CHO cells were pre-stained with a live dye, CellTracker Red CMTPX Probes (Molecular Probes), before mixing with 0.2 mM of X-DNA monomers and 3 Weiss units of T4 DNA ligase. After gelation, gels (with cells) were returned to the incubator.
Received 4 May 2006; accepted 9 August 2006; published 24 September 2006. References
1. Luo, D. The road from biology to materials. Mater. Today 6, 3843 (2003). 2. Seeman, N. C. Structural DNA nanotechnology: An overview. Methods Mol. Biol. 303, 143166 (2005). 3. Feldkamp, U. & Niemeyer, C. M. Rational design of DNA nanoarchitectures. Angew. Chem. Int. Edn Engl. 45, 18561876 (2006). 4. Chen, J. H. & Seeman, N. C. Synthesis from DNA of a molecule with the connectivity of a cube. Nature 350, 631633 (1991). 5. Seeman, N. C. Nucleic acid junctions and lattices. J. Theor. Biol. 99, 237247 (1982). 6. Kallenbach, N. R. et al. Fourth rank immobile nucleic acid junctions. J. Biomol. Struct. Dyn. 1, 159168 (1983).
LETTERS
7. Seeman, N. C. & Kallenbach, N. R. Design of immobile nucleic acid junctions. Biophys. J. 44, 201209 (1983). 8. Seeman, N. C. DNA nanotechnology: Novel DNA constructions. Ann. Rev. Biophys. Biomol. Struct. 27, 225248 (1998). 9. Yan, H., Zhang, X., Shen, Z. & Seeman, N. C. A robust DNA mechanical device controlled by hybridization topology. Nature 415, 6265 (2002). 10. Yan, H., Park, S. H., Finkelstein, G., Reif, J. H. & LaBean, T. H. DNA-templated self-assembly of protein arrays and highly conductive nanowires. Science 301, 18821884 (2003). 11. Seeman, N. C. From genes to machines: DNA nanomechanical devices. Trends Biochem. Sci. 30, 119125 (2005). 12. Pinto, Y. Y. et al. Sequence-encoded self-assembly of multiple-nanocomponent arrays by 2D DNA scaolding. Nano Lett. 5, 23992402 (2005). 13. Liao, S. & Seeman, N. C. Translation of DNA signals into polymer assembly instructions. Science 306, 20722074 (2004). 14. Li, Y. et al. Controlled assembly of dendrimer-like DNA. Nature Mater. 3, 3842 (2004). 15. Li, Y., Cu, Y. T. & Luo, D. Multiplexed detection of pathogen DNA with DNA-based uorescence nanobarcodes. Nature Biotechnol. 23, 885889 (2005). 16. Lin, D. C., Yurke, B. & Langrana, N. A. Mechanical properties of a reversible, DNA-crosslinked polyacrylamide hydrogel. J. Biomech. Eng. 126, 104110 (2004). 17. Kong, H. J., Kaigler, D., Kim, K. & Mooney, D. J. Controlling rigidity and degradation of alginate hydrogels via molecular weight distribution. Biomacromolecules 5, 17201727 (2004). 18. Luo, Y. & Shoichet, M. S. A photolabile hydrogel for guided three-dimensional cell growth and migration. Nature Mater. 3, 249253 (2004). 19. Kisiday, J. et al. Self-assembling peptide hydrogel fosters chondrocyte extracellular matrix production and cell division: implications for cartilage tissue repair. Proc. Natl Acad. Sci. USA 99, 999610001 (2002). 20. Holmes, T. C. et al. Extensive neurite outgrowth and active synapse formation on self-assembling peptide scaolds. Proc. Natl Acad. Sci. USA 97, 67286733 (2000). 21. Yang, D., Strode, J. T., Spielmann, H. P., Wang, A. H.-J. & Burke, T. G. DNA interactions of two clinical camptothecin drugs stabilize their active lactone forms. J. Am. Chem. Soc. 120, 29792980 (1998). 22. Wooldridge, J. E. & Weiner, G. J. CpG DNA and cancer immunotherapy: Orchestrating the antitumor immune response. Curr. Opin. Oncol. 15, 440445 (2003). 23. Dalpke, A. H. & Heeg, K. CpG-DNA as immune response modier. Int. J. Med. Microbiol. 294, 345354 (2004). 24. Um, S., Lee, J., Kwon, S., Li, Y. & Luo, D. Dendrimer-like DNA (DL-DNA) based uorescence nanobarcodes. Nature Protocols 1, 9951000 (2006). 25. Luo, D., Woodrow-Mumford, K., Belcheva, N. & Saltzman, W. M. Controlled DNA delivery systems. Pharm. Res. 16, 13001308 (1999).
Acknowledgements
We wish to acknowledge nancial support from USDA, the Cornell Advanced Technology Centre for Biotechnology, the Nanobiotechnology Center (NSF/ECS-9876771) and the Cornell Center for Materials Research (NSF/DMR-0520404). D.L. acknowledges the NSF CAREER Award and N.P. acknowledges the Korea Research Foundation Grant (04-03-05-2) for partial support. Y. Chang, Y. Zhang and N. Walker are acknowledged for help in animal tests, DMA and data processing, respectively. Correspondence and requests for materials should be addressed to D.L. Supplementary Information accompanies this paper on www.nature.com/naturematerials.
Competing nancial interests

The authors declare competing nancial interests: details accompany the paper on www.nature.com/naturematerials.
Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/
801

Dan Luo - The Road From Biology To Materials

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Dan Luo - The Road From Biology To Materials

Hochgeladen von

Copyright:

Verfügbare Formate

The road from

This article is intended to serve as a roadmap, linking

Why molecular biology?

ISSN:1369 7021 Elsevier Ltd 2003

A very brief introduction

Biomacromolecules in living organisms

DNA as a polymeric material

Tool kit 2: manipulating DNA

Tool kit 1: obtaining DNA

Table 1 Tools for manipulating DNA.

Elongating DNA or RNA

E. Coli DNA polymerase I Taq polymerase Reverse transcriptase

Specific Shortening DNA or RNA Exonucleases Endonucleases

At 5 end At 3 end At internal sites

Tool kit 3: attaching DNA to a surface

DNA nanotechnology and materials

Fig. 2 Transmission electron micrograph of dendrimer-like DNA.

Controlled assembly of dendrimer-like DNA

Published online: 21 December 2003; doi:10.1038/nmat1045

2004 Nature Publishing Group

2004 Nature Publishing Group

2004 Nature Publishing Group

Table 1 Sequences of oligonucleotides.

FORMATION OF Y-DNA AND DL-DNA

Competing financial interests

nature materials | VOL 3 | JANUARY 2004 | www.nature.com/naturematerials

2005 Nature Publishing Group http://www.nature.com/naturebiotechnology

Multiplexed detection of pathogen DNA with DNA-based uorescence nanobarcodes

2005 Nature Publishing Group http://www.nature.com/naturebiotechnology

Relative fluorescence intensity

1G4R Probe 5 Wavelength

2005 Nature Publishing Group http://www.nature.com/naturebiotechnology

Avidin 5.5-m polystyrene bead

Biotin-labeled capture probes

Barcodes with report probes

Adding target DNA

Control 1 Francisella tularensis

2005 Nature Publishing Group http://www.nature.com/naturebiotechnology

2005 Nature Publishing Group http://www.nature.com/naturebiotechnology

Dendrimer-like DNA-based uorescence nanobarcodes

2006 Nature Publishing Group http://www.nature.com/natureprotocols

2006 Nature Publishing Group http://www.nature.com/natureprotocols

. Oligonucleotides for Y-DNA building blocks (Integrated DNA Technologies;

2006 Nature Publishing Group http://www.nature.com/natureprotocols

O NH 5-phos CCGG GGCC AAC 3 TTGACT5 Bio-5p-CCGG GGCC Yn-DNA

S O Avidin-coated agarose support

A Bio-5p-CCGG AACTGAG GGCC TTGACTC

A Bio-5p-CCGG AACTGAG GGCC TTGACTC

A Bio-5p-SPY -DNAY -DNA 0 1

A Bio-5p-SPY -DNAY -DNAY -DNA 0 1 2

DDE I (RE) Step 7 TGA

2006 Nature Publishing Group http://www.nature.com/natureprotocols

Avidin-coated polystyrene bead Nanobarcode with report probes

D-Biotin DNA target

Figure 4 | A dendrimer-like-DNA nanobarcode for use in ow cytometry detection.

? TROUBLESHOOTING Troubleshooting advice can be found in Table 5.

1000 | VOL.1 NO.2 | 2006 | NATURE PROTOCOLS

Enzyme-catalysed assembly of DNA hydrogel

Published online: 24 September 2006; doi:10.1038/nmat1741

nature materials VOL 5 OCTOBER 2006 www.nature.com/naturematerials

2006 Nature Publishing Group

2006 Nature Publishing Group