Molecular Scale Architecture: Engineered Three- And Four-Way

Dec 11, 2007 - In this context, the ability to construct discrete three- and four-way junctions ... To this end, we have constructed isolated three-wa...
0 downloads 0 Views 698KB Size
Download PDF
470

Bioconjugate Chem. 2008, 19, 470–475

Molecular Scale Architecture: Engineered Three- And Four-Way Junctions Stephanie Wilkinson,† Michael Diechtierow,§ R. August Estabrook, Falk Schmidt,| Michael Hüben,| Elmar Weinhold,| and Norbert O. Reich*,†,‡ Department of Chemistry and Biochemistry and Biomolecular Science and Engineering Program, University of California at Santa Barbara, 93106-9510, University of Heidelberg, Germany, and Institute of Organic Chemistry, RWTH Aachen University, Germany. Received July 18, 2007; Revised Manuscript Received October 18, 2007

Biomolecular self-assembly provides a basis for the bottom-up construction of useful and diverse nanoscale architectures. DNA is commonly used to create these assemblies and is often exploited as a lattice or an array. Although geometrically rigid and highly predictable, these sheets of repetitive constructs often lack the ability to be enzymatically manipulated or elongated by standard biochemical techniques. Here, we describe two approaches for the construction of position-controlled, molecular-scale, discrete, three- and four-way DNA junctions. The first approach for constructing these junctions relies on the use of nonmigrating cruciforms generated from synthetic oligonucleotides to which large, biologically generated, double-stranded DNA segments are enzymatically ligated. The second approach utilitizes the DNA methyltransferase-based SMILing (sequence-specific methyltransferaseinduced labeling of DNA) method to site-specifically incorporate a biotin within biologically derived DNA. Streptavidin is then used to form junctions between unique DNA strands. The resultant assemblies have precise and predetermined connections with lengths that can be varied by enzymatic or hybridization techniques, or geometrically controlled with standard DNA functionalization methods. These junctions are positioned with single nucleotide resolution on large, micrometer-length templates. Both approaches generate DNA assemblies which are fully compatible with standard recombinant methods and thus provide a novel basis for nanoengineering applications.

INTRODUCTION Diverse materials including nanotubes (1), organic molecules (2), and biomolecules (3) are being employed to form novel nanoscale devices. Of these approaches, biomolecules offer a promising route to form complex and well-defined structures due to their highly convergent, self-assembling capabilities. DNA-based constructs provide many unique properties which make them ideal candidates for nanoscale devices. First, the self-assembling nature of DNA provides a bottom-up approach for synthesis with exquisite specificity and selection. Second, the physical dimensions of DNA, 2 nm in diameter and anywhere from 5 to 5000 nm in length, make this biomolecule an attractive template for wires in the nanometer regime despite its intrinsically poor conductivity (4). Third, commercially available DNA modifying enzymes (e.g., polymerases, ligases, endonucleases, and methyltransferases) and DNA binding proteins (e.g., zinc finger containing proteins) are compatible with such architectures facilitating their controlled static and dynamic manipulation. The specific interactions between base pairs of DNA have been used to create nanostructures and direct the assembly of materials on the angstrom to micrometer scale (5–7). An emerging application of DNA nanostructures is the development of molecular electronics through the use of bottom-up approaches (3) (8) (9). In this context, the ability to construct discrete three- and four-way junctions will become increasingly * To whom correspondence should be addressed: reich@ chem.ucsb.edu, (805) 893-8368; FAX (805) 893-4120. † Department of Chemistry and Biochemistry, University of California at Santa Barbara. ‡ Biomolecular Science and Engineering Program, University of California at Santa Barbara. § University of Heidelberg. | RWTH Aachen University.

important as more complex electronic architectures are sought. Novel DNA architectures for nanotemplating have been prepared, most notably by Seeman and co-workers (6). These DNA architectures are most frequently made up of highly repetitive, blocklike lattices or scaffolds (7) (10). These geometrically defined structures are repetitive and relatively rigid in both structure and sequence, making further recombinant manipulation problematic. For example, it is extremely difficult to introduce site-specific modifications to a small number of positions in such highly repetitive assemblies. Similarly, restriction enzymes, DNA ligases, and DNA polymerases are not always able to gain access to these highly interwoven assemblies. The methods described here can be used in conjunction with lattices or arrays in order to increase their versatility, or as a stand-alone assembly method yielding flexible constructs. DNA-based molecular electronics have included conductive nanowires as well as electroless plating techniques (3) (4). An important electronic feature common to many devices relies on the use of junctions for signal splitting and amplification as well as for control gates. To this end, we have constructed isolated three-way and four-way DNA junctions at predetermined positions. One approach uses a nonmigrating cruciform as the core structure to which long, flexible DNA arms are enzymatically ligated (Figure 1). This recombinant approach differs from a similar method where 5′-biotinylated DNA was joined together by a streptavidin linkage to create a discrete four-way junction (11). Such structures are readily positioned onto metal electrodes via 5′ or 3′ functionalization to the ends of DNA (3) (4). Another approach positions one or more biotins or other labels within biologically isolated DNA. This position-controlled and selective placement of biotin was achieved with the SMILing DNA method (sequence-specific methyltransferase-induced labeling of DNA) (12) (13), which uses modified AdoMet analogues

10.1021/bc700270k CCC: $40.75  2008 American Chemical Society Published on Web 12/11/2007

Engineered Three- and Four-Way Junctions

Bioconjugate Chem., Vol. 19, No. 2, 2008 471

Figure 1. Cartoon of cruciform construction. (A) Four single strands of DNA are annealed to form the core cruciform. (B) Core cruciform. (C) Large DNA arms constructed by PCR with subsequent BsmBI digestion to create 5′-TCCC-3′ overhangs. (D) DNA arms were ligated to the core cruciform to form the final cruciform construct.

and sequence-specific DNA methyltransferases to deliver the biotin moiety.

EXPERIMENTAL PROCEDURES Cruciform Preparation. Large DNA cruciforms were constructed by ligating PCR-derived DNA molecules to a small synthetic core junction. The small core cruciform was constructed by annealing four synthetic 36 bp DNA molecules (5′GGGAGGTAGGACCGCAATCCTGAGCACGTCTCAACG3′, 5′-GGGACCACCAGTGCCATAGTGGATTGCGGTCCTACC-3′, 5′-GGGAGGAACTGTGCATTCGGACTATGGCACTGGTGG-3′,and5′-GGGACGTTGAGACGTGCTCACCGAATGCACAGTTCC-3′) with a 5′ overhang of 5′-GGGA-3′ on each arm. Oligonucleotides purchased from Midlands DNA were HPLC purified and kinased with dATP (Sigma) and T4 DNA kinase (NEB) extensively to ensure that all strands were fully phosphorylated at the 5′ end. These oligos were then repurified by phenol-chloroform extraction and annealed by boiling a mixture of equal molar ratios of each strand in STE buffer (100 mM NaCl, 10 mM Tris pH 8.0, 1 mM EDTA) for 5 min before being slowly cooled to room temperature. Proper annealing was confirmed by PAGE stained with SYBR gold (Invitrogen). Large cruciform arms were made using pBR322 as a template for PCR. Amplicons 899 bp in length were generated with the primers 5′-CCCCCTTACACGGAGGCATCAGTGACC-3′ and 5′-GCTATGAGAAAGCGCCACGCTTCCCGAAGG-3′, respectively. The resulting amplicon had a single BsmBI recognition site 226 bp from one end. Up to 10 mL of PCR mixture was made using Taq DNA polymerase, 2.5 mM dNTPs, 2 ng/ µL of each primer, 1 ng/µL of template, and was thermocycled: (1) 30 s at 95 °C, (2) 30 s at 95 °C, (3) 1 min at 55 °C, (4) 2 min at 74 °C, (5) repeat 30 times from step 2. PCR product was purified using the Qiaquick PCR Clean-up Kit (Qiagen) followed by phenol-chloroform precipitation. The concentration of the amplicon in TE (10 mM Tris, pH ) 8.0, 1 mM EDTA) buffer was determined by A260. Purified amplicon was digested with BsmBI (NEB) for up to 12 h to ensure completion. Once fully cut, the two fragments were isolated and purified by agarose gel extraction. DNA bands were cut from an agarose gel run in TBE buffer (10 mM Tris pH 8.0, 10 mM boric acid, 1 mM EDTA) and placed into 10 000 MWC dialysis tubing (Spectrum) for further electrophoresis. After 1 h, the current was reversed for 30 s and the TBE buffer within the dialysis tubing, which contained the isolated DNA fragments, was removed and ethanol precipitated. The concentration of amplicon fragments in TE buffer was determined by A260. Ligation reactions were made between synthetic core cruciforms and PCR-derived arms with T4 DNA ligase (NEB) at 4 °C. These reactions were usually done in large amounts so small aliquots could be removed and analyzed by agarose gel to

Figure 2. Data from cruciform construction. (A) 12% PAGE showing the annealing of single strands to form the core cruciform. (B) Ligated constructs were identified and isolated from 1% agarose gel. (C) Constructs were imaged by AFM on mica.

Figure 3. Schematic and AFM of four-way SMILing junctions. (A) Schematic showing the materials and design for assembling junctions. (B) Chemical structure of the modified methyltransferase cofactor 6BAz used in the SMILing DNA reactions. (C) Schematic and AFM image of four-way junctions created by incubating SM1 and STV at a ratio of 1:1.

monitor progress. Initially, PCR fragments had a 5-fold higher molar concentration than core cruciforms. As the reaction progressed, additional arm, core cruciform, or ligase could be added. Core cruciforms with four ligated arms were isolated from an agarose gel in the same manner as the PCR fragments. Construction was confirmed by digestion with T7 endonucleases I (NEB), which specifically cleaves DNA junctions (data not shown). SMILing DNA Preparation. Plasmid pBR322 (NEB) was used as a template for PCR amplifications, using the forward primer 5′-CCAAGTCATTCTGAGAATAGTGTA-3′ and reverse primer 5′-CGCCGAAACAAGCGCTCATGAGCCCG-3′, respectively, to generate a 1017 bp dsDNA fragment (SM1, Figure 3A(a)) with one centrally located (164 and 182 nm away from each end) recognition sequence for the DNA methyltransferase M.BseCI. This amplicon was also formed with a 5′-biotin (5B1, Figure 3A(c)) by using the forward primer listed above with the reverse biotinylated primer 5′-[BioTEG]CGCCGAAACAAGCGCTCATGAGCCCG-3′. Plasmid pUC19 (NEB) was

472 Bioconjugate Chem., Vol. 19, No. 2, 2008

also used as a template for PCR using the primers 5′AGGCATGCAAGCTTGGCGTAATCATGG-3′ and 5′-CCCCGACACCCGCCAACACC-3′, respectively, to generate a 2383 bp dsDNA fragment (SM2.4, Figure 3A(b)) with two recognition sequences for the DNA methyltransferase M.TaqI, 160 nm from either end. All primers were ordered from Operon. Each PCR amplicon was purified with a Qiagen QIAquick PCR purification kit followed by phenol-chloroform extraction and ethanol precipitation. SMILing DNA Modification. SM1 (1 µg), with one recognition site for M.BseCI at position 531–536 (30 nM), M.BseCI (303 nM), and 6BAz (80 µM, Figure 3B) were incubated in buffer (50 µL containing 10 mM Tris/HCl, pH 7.4, 20 mM MgCl2, 50 mM NaCl, 50 µM EDTA, and 2 mM 2-mercaptoethanol) at 55 °C for 3 h. The reaction mixture was then heated to 70 °C for 20 min. Control reactions without M.BseCI and 6BAz were performed in parallel. After analysis by agarose gel electrophoresis, SM1 was purified using QIAquick PCR purification kits (QIAGEN) according to the instructions of the manufacturer. Similar protocols were used to create SM2.4. To test the integrity of the DNA, aliquots of the SMILing reaction and the control reactions (10 µL, 0.2 µg DNA) were treated with Proteinase K (1 µL, 20 mg/mL in 10 mM Tris/ HCl, pH 7.5, QIAGEN Proteinase K) and incubated at 37 °C for 1 h. Afterward, 6× loading buffer (2.2 µL, 30% glycerin, and 0.25% bromophenol blue) was added to each aliquot and the DNA was visualized by agarose gel electrophoresis (data not shown). Restriction analysis was performed by taking aliquots of the SMILing reaction and the control reactions (10 µL, 0.2 µg DNA) supplemented with water (1.3 µL), BSA (1.0 µL, 1.5 mg/mL), 10× NEB 2 buffer (1.5 µL, 100 mM Tris/ HCl, pH 7.9, 100 mM MgCl2, 500 mM NaCl, 10 mM DTT), and R.ClaI (1.2 µL, 4 units in NEB 2 buffer) and incubated at 37 °C for 1 h. Afterward, Proteinase K (1 µL, 20 mg/mL in 10 mM Tris/HCl, pH 7.5, QIAGEN Proteinase K) was added to each aliquot and the solutions were incubated at 37 °C for 1 h. Finally, 6× loading buffer (3.2 µL) was added to each aliquot. Results of the restriction digest were visualized by agarose gel electrophoresis (Supporting Information). To ensure biotin function, aliquots of the SMILing reaction and the control reactions (10 µL, 0.2 µg DNA) were supplemented with streptavidin (5 µL, 5 mg/mL in PBS buffer: 100 mM sodium phosphate, 100 mM NaCl, 2 mM sodium azide, pH 7.5) and incubated at 37 °C for 1 h. Afterward, 6× loading buffer (3.2 µL) was added to each aliquot. DNA mobility was visualized on a 0.8% agarose gel (Supporting Information). Four-Way SMILing Junctions. These junctions were constructed using SM1 and streptavidin (STV) (Figure 3C) at a molar ratio of 1:1. DNA samples (150 nM) were incubated for 48 h at 4 °C. The long incubation time maximized the assembly of the desired junction, which was slow to form in solution. The desired construct was confirmed by both agarose gel electrophoresis and AFM analysis (Figure 3C and Supporting Information). Further distance analysis by AFM also showed that the location of the four-way junction is consistent with the SMILing DNA-directed placement of the biotin group onto SM1 (Figure 3). Three-Way SMILing Junctions. These junctions were constructed with SM1, 5B1, and STV (Figure 4A,C). The highest yields were obtained with a stepwise addition of STV to SM1 in a ratio of SM1:STV of 1:5, followed by the addition of 5B1. AFM analysis verified that the location of the threeway junction is consistent with the SMILing DNA-directed placement of the biotin onto SM1 (Figure 4A,C). To create a double-three-way junction (Figure 4B,D), 5B1 and STV were incubated at a ratio of 1:5 followed by addition of SM2.4 at a ratio of 5B1/SM2.4 of 2:1. The anticipated placements of the

Wilkinson et al.

Figure 4. Schematic and AFM of three-way junctions. (A) Schematic of a single three-way junction formation. STV was preincubated with SM1 at a 5:1 ratio with subsequent addition of 5B1 to achieve the desired construct. (B) Schematic of double three-way junction formation. 5B1 was preincubated with 5-fold molar excess STV. SM2.4 was added at half the concentration of 5B1 to form junctions. (C) Tappingmode AFM image of single three-way junctions described in A. (D) AFM image of a double three-way junction described in B.

three-way junctions on this larger element are also consistent with the SMILing DNA-directed placement of the two biotin moieties on SM2.4 (Figure 4A). AFM Sample Preparation. Desired DNA constructs were assembled in solution prior to immobilization for AFM. The resulting DNA (10 ng) was incubated in 20 µL of 1 mM MgCl2 for 30 min at 22 °C. The resulting 20 µL solution was then deposited on freshly cleaved mica and left for 5 min. The sample was then washed using 3 mL of filtered and doubly distilled deionized water. Water remaining on the mica was wicked away and further dried with nitrogen. All AFM imagining was done on a Digital Instrument Multimode Dimension 3000.

RESULTS AND DISCUSSION Cruciform-Based Junctions. Our initial four-way DNA junction was constructed by using a core, immobile cruciform from a set of previously published sequences (Figure 1) (14). These sequences were modified slightly to incorporate additional four base overhangs on the 5′ ends. The annealing of this cruciform showed a few aberrant secondary structures by PAGE, but the correct cruciform was still the most abundant form (Figure 2A). The successful ligation of the long arms to the core cruciform required several purification steps. PCR amplicons that formed the basis of the arm were purified away from residual primers which showed inhibitory affects toward the DNA ligase. The resulting purified amplicons were then digested with BsmBI to yield small (315 b.p.) and large (584 b.p.) DNA fragments. The large DNA fragment was isolated by gel purification followed by phenol-chloroform extraction and ethanol precipitation. The large DNA fragment was ligated back to the small DNA fragment to confirm the presence of the 5′ overhang (5′-TCCC-3′) essential for cruciform ligation. Large four-way DNA junctions were successfully assembled by T4 DNA ligase as shown by agarose gel in Figure 2B. Initial cruciform ligation reactions contained roughly 5-fold molar excess of arm to core cruciform. As the reaction progress was monitored over time, subsequent additions of either core or arm

Engineered Three- and Four-Way Junctions

were made to the reaction mixture. The addition of free arm, although already in excess, led to more of the 4-way junctions. Isolation and purification of the large cruciforms was accomplished by gel excision. Due to the nature of this technique and the cruciform structure, large quantities of ligated cruciform were often required in order to recover a discernible amount. Various methods including the QIAquick Gel Extraction Kit (Qiagen) were attempted, but electrophoresis followed by concentration by ethanol precipitation gave the best results. Once purified, the final product was confirmed by successful digestion with T7 endonucleaseI, which exclusively cleaves DNA junctions. The long, flexible DNA arms may find novel nanotechnology applications due to their extended linear nature and ability to be modified at unique and predetermined positions. For example, we have placed single metallic nanoparticles at specific positions on such structures with a goal toward creating electronic components (36). Although the structure of the DNA is not novel (11), our procedure to create the cruciform is unique. We have exploited the convergent capabilities of using DNA and DNA modifying enzymes in order to achieve higher-order structures. In this regard, using PCR-derived DNA that is manipulated by a restriction endonuclease is much more cost-effective than ordering synthetic arms or 5′ biotinylated primers. Further, we have shown that the immobile cruciform can be manipulated by enzymatic means. The construction approach demonstrated here should also be applicable for three-way junctions and junctions involving arms of various lengths. SMILing-Based Junctions. The SMILing approach relies on cofactor analogues that deliver a small molecule from modified S-adenosylmethionine (AdoMet) to a site-specific location on the DNA (12) (13). This DNA methyltransferasedirected assembly of DNA junctions is schematically shown in Figures 3 and 4. DNA methyltransferases are responsible for many biological functions in prokaryotes and eukaryotes, including prokaryotic restriction-modification systems (15), DNA mismatch repair (16), chromosome replication (17), gene regulation (18), mammalian development (19), oncogenesis (20), and the organization of chromatin structure (21). Because these enzymes are structurally and mechanistically well-understood, they provide a sound basis for the types of bionanoengineering described here. The SMILing method’s replacement of an aziridine ring for the amino acid in AdoMet allows for the covalent incorporation of diverse ligands (12) (13) (22). Here, we applied the SMILing DNA approach to form novel, branching DNA structures. M.BseCI and M.TaqI use the natural cofactor AdoMet to methylate the amino group of the starred adenine residue in the sequences 5′-ATCGA*T-3′ and 5′-TCGA*-3′, respectively. In our studies, these enzymes were used to catalyze a nucleophilic aziridine ring opening of the cofactor analog 6BAz (Figure 3B) and covalently attach a single biotin at the target adenine in their respective recognition sequences (Figure 3A). The reaction of M.BseCI with the AdoMet analogue 6BAz results in the selective placement of biotin at one position within the 1017 bp dsDNA fragment (164 and 182 nm from each end). The M.TaqI reaction incorporated a biotin at two positions on the 2383 bp dsDNA (160 nm from either end). Biotin incorporation was verified by showing that the corresponding restriction endonuclease was incapable of digesting the modified DNA (Supporting Information). The tetravalent binding capacity of streptavidin (STV) (23) suggests that the reaction between SM1 molecules and STV can lead to four-, six-, and eight-way junctions. However, we saw no evidence for six- or eight-way junctions (Supporting Information). Further, the formation of the four-way junction occurred after a 48 h incubation at 4 °C. The lengthy reaction time for four-way junctions and lack of six- and eight-way

Bioconjugate Chem., Vol. 19, No. 2, 2008 473

junctions is most likely due to the steric hindrance and electrostatic repulsion between the negatively charged DNA phosphate groups. To optimize the assembly of the desired fourway junction, we varied both the STV to biotinylated DNA ratio and a prebound DNA-STV to biotinylated DNA ratio. There was a clear dependence on these two ratios and how well junctions were formed (Supporting Information). At high concentrations of STV, few junctions were formed most likely due to all biotins being bound by individual STV. Conversely, high DNA concentrations yielded more junctions. In the case of the four-way junction, a 1:1 ratio of STV to biotinylated DNA gave the best results for formation of the desired DNA construct. Contrary to the single incubation for four-way junctions, two reaction steps were necessary for the assembly of three-way junctions. For a single three-way junction, excess STV was allowed to bind to SM1 followed by the addition of the 5B1 to the reaction mixture (Figure 4A). Junction formation of the STV to the 5′-biotinylated DNA occurred much faster in comparison with the SMILing-modified DNA alone (data not shown), suggesting that steric interactions between DNA molecules may play a significant role in determining the rate and extent of reaction. A stepwise addition was also used for the construction of the double three-way junction (Figure 4B). 5B1 was preincubated with excess STV followed by addition of SM2.4. The use of biotin-streptavidin linkages to form DNA structures is well-understood (11) (24–27); however, we have used a DNA-modifying enzyme to incorporate a biotin site specifically within long, native DNA derived from PCR. Li et al. reports a 2.1-fold increase in height upon STV binding to bare DNA (28). Cross-section analyses of our biotin-streptavidin linkages reveal a similar change in height; however, this is not easily seen in the SMILing DNA figures. We propose that this could be due to dehydrating AFM conditions that have been previously seen with DNA (29) and DNA binding proteins (30) or differences in tips used for AFM. However, the presence of the STV in our images can be further validated by the lengths of the DNA branches seen in Figures 3 and 4. The lengths correlate well with the site-specific positioning of the biotin and are consistent throughout many structures. The mere probability of this occurrence is highly unlikely. Our current ability to assemble DNA junctions via biotinstreptavidin linkages provides constructs which are topologically but not geometrically defined. The lack of geometric constraint comes from the long DNA “arms” which extend from the junctions and go well beyond the persistence length of doublestranded DNA. Importantly, any approach making use of DNA junctions (31) will be faced by the breakdown of the geometric constraints when used with long DNA “arms”. However, this may be overcome by using surfaces to which the complementary oligonucleotides are directed, forcing the hybridization of specific arms to particular surfaces in a geometrically defined manner. Additionally, PCR-derived DNA retains the ability to be end-functionalized, enabling placement of arms across electrodes or other devices. Recent studies have described novel methods of creating finite DNA structures (10) (27) (32) (33). Although successful, these studies are still dependent on tightly packed, non-native forms of DNA, which make further manipulation by recombinant methods difficult. Our approach is further distinguished from DNA tiling and DNA array methods (6) (27) (34) (35), which normally result in assemblies with variable lengths or dimensions. Notably, Tian et al. recently described a long, flexible 4-way DNA junction (11). However, their approach relied solely upon 5′-biotinylated DNA brought together by the tetravalent binding capability of streptavidin. Our approaches use enzymatic manipulation (ligation, SMILing) independent of 5′ modifications to achieve similar higher-order structures. We argue that

474 Bioconjugate Chem., Vol. 19, No. 2, 2008

our approaches offer more control with site-specific biotin incorporations within DNA in addition to the versatility associated with recombinant methods of manipulating biomolecules. Furthermore, the use of biologically derived DNA is more costeffective and does not require the types of purification normally needed for synthetic DNA. Finally, the relative homogeneity of synthetic DNA oligonucleotide sequences means that kinetic traps may interfere with the assembly of a particular structure; this is less likely to be a problem with the highly heterogeneous biologically derived DNA. Our construction of a set of novel, discrete nucleic acid architectures shows promise for directing the assembly of a variety of nanomaterials into higher-order structures using bottom-up assembly. This approach shows a level of molecular control which is essential for the construction of robust nanomaterials. Additionally, this approach also incorporates the use of recombinant techniques such as PCR and the use of DNAmodifying enzymes to achieve novel DNA architectures. The application of this approach to fabricate hybrid structures using carbon nanotubes and nanoparticles is being pursued. Ultimately, the retention of DNA molecules that can be manipulated by recombinant DNA methods provides a further basis for subsequent design and functionalization.

ACKNOWLEDGMENT We thank Dr. Nadrian Seeman for insights into cruciform construction. This work was supported by a grant from the Institute for Collaborative Biotechnologies (DAAD19-03-D0004), U.S. Army Research Office (Reich), and a UC GREAT Grant (Reich). This work made use of MRL Central Facilities supported by the MRSEC Program of the National Science Foundation under award No. DMR00-80034. We thank Basar Gider for helping in the initial synthesis of 6BAz, Kerstin Glensk for performing the SMILing reactions, and Prof. Dr. Michalis Kokkinidis for providing M.BseCI. We also thank Dr. Debra Fygenson for her suggestions. Note Added after ASAP Publication: The author’s name has changed since the original Web posting on December 11, 2007. The correction was made and the article reposted on January 9, 2008. Supporting Information Available: Details on the SMILing modified DNA and the concentration and ratio dependencies on junction formation are given. This material is available free of charge via the Internet at http://pubs.acs.org.

LITERATURE CITED (1) Martin, C. R., and Kohli, P. (2003) The emerging field of nanotube biotechnology. Nat. ReV. Drug DiscoVery 2, 29–37. (2) Rosei, F., Schunack, M., Naitoh, Y., Jiang, P., Gourdon, A, Laegsgaard, E., Stensgaard, I., Joachim, C., and Besenbacher, F. (2003) Properties of large organic molecules on metal surfaces. Prog. Surf. Sci. 71, 95–146. (3) Braun, E., and Keren, K. (2004) From DNA to transistors. AdV. Phys. 53, 441–96. (4) Taniguchi, M., and Kawai, T. (2006) DNA electronics. Phys. E 33, 1–12. (5) Brucale, M., Zuccheri, G., and Samori, B. (2006) Mastering the complexity of DNA nanostructures. Trends Biotechnol. 5, 235–43. (6) Seeman, N. C., and Lukeman, P. S. (2005) Nucleic acid nanostructures: bottom-up control of geometry on the nanoscale. Rep. Prog. Phys. 68, 237–70. (7) Gothelf, K. V., and LaBean, T. H. (2005) DNA-programmed assembly of nanostructures. Org. Biomol. Chem. 3, 4023–37.

Wilkinson et al. (8) Katz, E., and Willner, I. (2004) Integrated nanoparticlebiomolecule hybrid systems: synthesis, properties, and applications. Angew. Chem., Int. Ed. 43, 6042–108. (9) Seeman, N. C. (2003) DNA in a material world. Nature (London) 421, 427–31. (10) Rothemund, P. W. K. (2006) Folding DNA to create nanoscale shapes and patterns. Nature (London) 440, 297–302. (11) Tian, Y., He, Y., Ribbe, A. E., and Mao, C. D. (2006) Preparation of branched structures with long DNA duplex arms. Org. Biomol. Chem. 4, 3404–5. (12) Pljevaljcic, G., Schmidt, F., and Weinhold, E. (2004) Sequencespecific methyltransferase-induced labeling of DNA (SMILing DNA). ChemBioChem 5, 265–9. (13) Pljevaljcic, G., Schmidt, F., Peschlow, A., and Weinhold, E. (2004) Sequence-specific DNA labeling using methyltransferases. Methods Mol. Biol. 283, 145–61. (14) Kallenbach, N. R., Ma, R. I., and Seeman, N. C. (1983) An immobile nucleic-acid junction constructed from oligonucleotides. Nature (London) 305, 829–31. (15) Wilson, G. G., and Murray, N. E. (1991) Restriction and modification systems. Annu. ReV. Genet. 25, 585–627. (16) Palmer, B. R., and Marinus, M. G. (1994) The dam and dcm strains of Escherichia coli–a review. Gene 143, 1–12. (17) Reisenauer, A., Kahng, L. S., McCollum, S., and Shapiro, L. (1999) Bacterial DNA methylation: a cell cycle regulator? J. Bacteriol. 181, 5135–9. (18) Jaenisch, R., and Bird, A. (2003) Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat. Genet. 33, 245–54. (19) Reik, W., and Walter, J. (2001) Genomic imprinting: parental influence on the genome. Nat. ReV. Genet. 2, 21–32. (20) Jones, P. A., and Baylin, S. B. (2002) The fundamental role of epigenetic events in cancer. Nat. ReV. Genet. 3, 415–28. (21) Jackson-Grusby, L., and Jaenisch, R. (1996) Experimental manipulation of genomic methylation. Semin. Cancer Biol. 7, 261–8. (22) Pljevaljcic, G., Pignot, M., and Weinhold, E. (2003) Design of a new fluorescent cofactor for DNA methyltransferases and sequence-specific labeling of DNA. J. Am. Chem. Soc. 125, 3486–92. (23) Hendrickson, W. A., Pahler, A., Smith, J. L., Satow, Y., Merritt, E. A., and Phizackerley, R. P. (1989) Crystal structure of core streptavidin determined from multiwavelength anomalous diffraction of synchrotron radiation. Proc. Natl. Acad. Sci. U.S.A. 86, 2190–4. (24) Niemeyer, C. M., Adler, M., Gao, S., and Chi, L. (2001) Nanostructured DNA-protein aggregates consisting of covalent oligonucleotide-streptavidin conjugates. Bioconjugate Chem. 12, 364–71. (25) Niemeyer, C. M. (2004) Semi-synthetic DNA-protein conjugates: novel tools in analytics and nanobiotechnology. Biochem. Soc. Trans. 32, 51–3. (26) Park, S. H., Yin, P., Liu, Y., Reif, J. H., LaBean, T. H., and Yan, H. (2005) Programmable DNA self-assemblies for nanoscale organization of ligands and proteins. Nano Lett. 5, 729–33. (27) Park, S. H., Pistol, C., Ahn, S. J., Reif, J. H., Lebeck, A. R., Dwyer, C., and LaBean, T. H. (2006) Finite-size, fully addressable DNA tile lattices formed by hierarchical assembly procedures. Angew. Chem., Int. Ed. 45, 735–9. (28) Li, H. Y., Park, S. H., Reif, J. H., LaBean, T. H., and Yan, H. (2004) DNA-templated self-assembly of protein and nanoparticle linear arrays. J. Am. Chem. Soc. 126, 418–9. (29) Bezanilla, M., Manne, S., Laney, D. E., Lyubchenko, Y. L., and Hansma, H. G. (1995) Adsorption of DNA to mica, silylated mice, and minerals - characterization by atomic-force microscopy. Langmuir 11, 655–9. (30) Allison, D. P., Kerper, P. S., Doktycz, M. J., Spain, J. A., Modrich, P., Larimer, F. W., Thundat, T., and Warmack, R. J. (1996) Direct atomic force microscope imaging of EcoRI endonuclease site specifically bound to plasmid DNA molecules. Proc. Natl. Acad. Sci. U.S.A. 93, 8826–9.

Engineered Three- and Four-Way Junctions (31) He, Y., Chen, Y., Liu, H., Ribbe, A. E., and Mao, C. (2005) Self-assembly of hexagonal DNA two-dimensional (2D) arrays. J. Am. Chem. Soc. 127, 12202–3. (32) Liu, Y., Ke, Y. G., and Yan, H. (2005) Self-assembly of symmetric finite-size DNA nanoarrays. J. Am. Chem. Soc. 127, 17140–1. (33) Lund, K., Liu, Y., Lindsay, S., and Yan, H. (2005) Self-assembling a molecular pegboard. J. Am. Chem. Soc. 127, 17606–7. (34) Feldkamp, U., and Niemeyer, C. M. (2006) Rational design of DNA nanoarchitectures. Angew. Chem., Int. Ed. 45, 1856–76.

Bioconjugate Chem., Vol. 19, No. 2, 2008 475 (35) O’Neill, P., Rothemund, P. W., Kumar, A., and Fygenson, D. K. (2006) Sturdier DNA nanotubes via ligation. Nano Lett. 6, 1379–83. (36) Braun, G., Diechtierow, M., Wilkinson, S., Schmidt, F., Lugowski, M., Weinhold, E., and Reich, N. Unpublished results.

BC700270K