DNA reusability and optoelectronic characteristics of streptavidin-conjugated DNA crystals on a quartz substrate

Abstract

We demonstrate self-assembled double-crossover (DX) DNA crystal growth and coverage on quartz by the substrate assisted growth method. Here we introduced the novel concept of a reusability process to fabricate the DX crystals on a given substrate with a 10 nM DNA concentration, which is equal to the saturation concentration and this concentration is enough to fully cover the substrate with DX crystals. Also the DX crystals with biotinylated oligonucleotides (DXB) were constructed based on the structure of immobile crossover branched junctions. Upon streptavidin conjugation with DNA, the Raman band intensities were increased as compared to pristine DX crystals which indicate the strong bonding between the biotinylated DXB crystals and streptavidin. The DX crystals showed a relatively higher current than streptavidin conjugated DXB crystals because of the electrically insulating characteristic of streptavidin. Furthermore, the optical band gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) as well as the second band onset were determined and discussed. The HOMO–LUMO band gaps of DNA crystals, DNA crystals with biotins, and DNA crystals with biotins and streptavidin showed an inverted V-shape and the second band onsets – consistent with the electrical characteristics – revealed the increasing behavior with protein conjugation.

---

Structural DNA nanotechnology is one of the best examples of the molecular self-assembly approach in nanotechnology due to its exceptional programmability, fast binding affinity and structural stability. The advantages of DNA biomolecules as building blocks in nontraditional materials used for applications in molecular electronics and optoelectronics are swiftly emerging in the field of nanotechnology.^1,2 Using the phenomenon of Watson–Crick complementarities, DNA nucleotides are hybridized between adenine and thymine, guanine and cytosine. Using programmable base-pair binding, sequences of specifically designed DNA strands can self-assemble during annealing into various dimensional DNA nanostructures.^3,4 The fabrication of diverse dimensional nanostructures plays an important role in structural DNA nanotechnology. The resulting DNA structures have been used as a template for aligning various functional materials such as proteins,^5–8 peptides,⁹ virus capsids,¹⁰ nanowires,^11,12 metal ions,¹³ and nanoparticles.^14,15 DNA crystals conjugated with proteins, metal ions and nanoparticles have yielded distinct and enhanced functional properties such as enzyme-cascade, charge transport, magnetic properties, and surface plasmon resonance through DNA mediated self-aligned functional materials.^16–18

The double-crossover (DX) tile – dimension of 4 nm × 12 nm in width and length – based DNA crystal used in this study has two repeating DX tiles (Fig. S1 and Table S1 in ESI†).⁶ The diameter of duplex DNA in the free solution is 2 nm; however, the DNA duplex on the substrate is measured as 1.2 ± 0.2 nm in the liquid and 0.6 ± 0.2 nm in the dry state due to the electrostatic interaction with the charged substrate.¹⁹ The 2D DNA crystals were grown on a given substrate by the surface-assisted growth (SAG) method^20–22 and details of sample preparation are explained in the Materials and methods section in ESI.† During the SAG method, the DNA crystallization process, including tile seeding, nucleation and growth was achieved while the annealing resulted in polycrystalline DX crystals with full coverage controlled by the monomer concentration (C_m) of DNA molecules. The selection of substrate is crucial for optical measurements such as Raman and optical band gap. Transparent quartz exhibits fundamental absorption below 190 nm while it transmits 90% of light above 200 nm, thereby providing a favorable condition for the optical analysis of DNA crystals. In this paper, we report the fabrication of protein bound to artificially designed DNA crystals, a demonstration of the reusability of DNA strands through a SAG method and an analysis of the optoelectronic measurements of DNA crystals with proteins. Raman, current–voltage (I–V) characteristics, and optical band gap measurements were adopted for studying the chemical and physical properties of protein conjugated DNA crystals.

To demonstrate the coverage dependence of C_m, 2D DX crystals were annealed with different DNA C_m quantities. The experimental schematic diagram of the DNA crystal growth procedure is shown in Fig. 1a. To accomplish accuracy and reliability, four physical parameters were fixed during the annealing process: substrate size (5 × 5 mm²), total volume of the DNA sample solution (250 μL), initial to final annealing temperature from 95 to 25 °C, and annealing time (24 hours) while different C_m – 0.5, 1, 2, 5, 10, and 20 nM – quantities were used as a control parameter. The AFM images of crystal coverage percentage as a function C_m are shown in Fig. 1b–e.

Fig. 1 Experimental procedure, DNA crystal coverage, and demonstration of reusability. (a) Schematic of the experimental procedure; O₂ plasma exposure on quartz and substrate-assisted DX crystal growth process. (b–e) The AFM images of DX crystals at various monomer concentrations (1, 2, 5, and 10 nM) and (f) crystal coverage percentage with variations of monomer concentration. The inset in (e) is the noise-filtered 2D spectrum image (scan size 100 × 100 nm²) by fast Fourier transform which shows the periodicity of the DX crystals. (g) Demonstration of the reusability of DNA strands. (h) Variation of crystal coverage on a sequential number of substrates with constant monomer concentration 10 nM. (e and i–k) Corresponding AFM images of 1^st, 2^nd, 3^rd, and 4^th substrates marked as n₁, n₂, n₃ and n₄, respectively.

The DX crystals started to grow on the quartz at a threshold concentration (C_th), ∼2 nM, consequently full coverage on quartz was reached at the saturation concentration (C_s) of ∼10 nM. Fig. 1e shows the fully covered DNA crystals on quartz. An inset in the bottom left (Fig. 1e) contains the noise-filtered 2D spectrum image according to fast Fourier transform which showed the periodicity of the unit building blocks (scan size 100 × 100 nm²).

The percentage of crystal coverage as a function of C_m is shown in Fig. 1f. The crystal coverage was nearly proportional to the C_m on a given substrate up to C_s. Beyond C_s, the crystal coverage of DX crystals was unchanged with increasing C_m. This meant that excess DX tiles could exist inside the test tube. In the SAG process, the nucleation of DX crystals on the quartz substrate started at a very low concentration (2 nM), much below the conventional free solution annealing (∼30 nM). This was due to the catalytic behavior of the substrate, by which the Coulomb force between substrate and DX tiles contributed a partial entropic cost for the DX DNA crystallization. This force creates relatively higher DNA molecular density, close to the substrate compared to the rest of the solution, providing appropriate conditions for crystallization. As the DX C_m increases and passes the C_th, the DX tiles start to nucleate on the substrate and small fragments of the DX nucleations continue to grow until C_s, fully covered by a monolayer of DX crystals. By controlling C_m, the accurate control of the coverage of DX crystals on the substrate was possible from 0 to 100%.

Excess DNA strands were quantitatively analyzed by a reusability test. The proposed reusability methodology was a simple and straight-forward process to fabricate the DX crystals on a given substrate without the overuse of strands. In order to test reusable strands in a test tube, we started with C_m, 10 nM which was same as C_s, i.e., barely enough to fully cover the quartz substrate by DX tiles. The O₂ plasma-treated quartz substrate along with the equimolar mixture of eight different DX strands of 10 nM were inserted into a test tube which was then placed in a styrofoam box with 2 L of boiling water and cooled slowly from 95 °C to 25 °C over a period of 24 hours to facilitate the hybridization process. After completing the annealing, the substrate was removed from the test tube and then a new substrate was inserted into the same test tube followed by the second annealing. By simple calculation, we noticed that about 2.3 nM of the DX tiles can be fully occupied on both sides of a 5 × 5 mm² substrate (see calculation of reusable DNA strand amount in ESI†). During the second annealing process, DX tiles were denatured at high temperature (>60 °C), DX tiles were formed in buffer solution at moderate temperature (<60 °C and >40 °C), and then DX crystals started to grow on the substrate at relatively low temperature (<40 °C).

During the successive cycles of the annealing process, the DX crystal coverage was gradually decreased by replacing the new substrate into the same DNA solution. The annealing process, with an initial C_m of 10 nM, was done four times with four new quartz substrates marked as n₁, n₂, n₃, and n₄. Fig. 1g shows an experimental schematic of the multiple annealing process with a few substrates for the demonstration of reusable excess strands in the test tube. The AFM images of DX crystal grown substrates, n₁ (after 1^st annealing), n₂ (after 2^nd), n₃ (after 3^rd), and n₄ (after 4^th) are shown in Fig. 1e and i–k, respectively. The white dotted line in the AFM image (Fig. 1e) indicates the DX crystal boundaries. By analyzing these AFM data, we calculated the DX crystal coverage as a function of n's, which is shown in Fig. 1h. Here the crystal coverage heavily depended upon C_m in the test tube. In our case, the 2^nd, 3^rd, and 4^th substrates were partially covered by them because the C_m (7.7 nM for 2^nd, 5.4 nM for 3^rd, and 3.1 nM for 4^th) was less than C_s. On the other hand, the reusability process verified the ultimate evidence for C_s. It means that at least the minimum amount (in our case, 7.7 nM) or the higher of free monomer concentration has to exist in a sample tube in order to achieve full coverage on a given substrate. Interestingly, when we used C_m, 20 nM of DX strands at the very beginning, we clearly observed the full coverage of DX crystals up to the 5^th substrate (data not shown). In this case, the concentration of residue strands after the 5^th annealing was about 8.5 nM (= 20 nM − (5 × 2.3 nM)) which was slightly larger than the minimum free monomer concentration.

The DX crystals with biotinylated oligonucleotides (DXB) were constructed based on the structure of immobile crossover branched junctions. DNA base sequences were designed to minimize the chance of errors caused by undesired complementary association and sequence symmetry, as shown in the Fig. S2 and Table S2 (ESI†). A 2D DXB crystals was composed of two DX building blocks; one has biotin (DXB1) and the other does not (DX2). These DX building blocks were designed to self-assemble into 2D DNA crystals by complementary sticky end association. After the formation of DXB crystals on a quartz substrate, we added the same amount of streptavidin (SA) as biotin protein in DXB. Fig. 2a–c showed the schematic diagrams of the DX, DXB, DXB crystals containing streptavidin proteins (DXB + SA), and corresponding AFM images. The white dotted lines shown in the AFM images of Fig. 2a and b indicate the crystal boundaries of the DX and DXB lattices, respectively. The SA protein has a size of ∼5 nm and it covalently binds to the biotin protein, ∼1 nm, in the DXB as shown in Fig. 2b and c. The insets in Fig. 2a and b are the noise-filtered 2D spectrum images according to fast Fourier transform showing the periodicities of the unit building blocks (DX tiles). The inset in Fig. 2c indicates the magnified AFM image of DXB + SA crystals. Topological differences between DX, DXB, and DXB + SA crystals were easily observable through the AFM images. Although the change in the surface topologies of DX and DXB crystals was hard to see with AFM due to the small size (1 nm) of the biotins, the periodic modification of the biotins to the DX crystals became prominent when SA was attached. The bright lines along the perpendicular directions of the DNA duplex on the DXB + SA crystals in Fig. 2c showed the clear binding of SA to DXB without any missing sites.

Fig. 2 Schematic diagram, AFM images, Raman and current measurements of the DNA crystals. Cartoons and AFM images of (a) double-crossover (DX) crystals, (b) DX crystals with biotin (DXB), (c) streptavidin bound DXB crystals (DXB + SA). The insets in (a and b) are the noise-filtered 2D spectrum images (scan size 100 × 100 nm²) by fast Fourier transform which show the periodicities of the lattices. An inset in (c) shows the magnified image of streptavidin alignment along the perpendicular directions of the DNA duplex. (d) The Raman spectra of quartz (reference), DX, and DXB + SA crystals. (e) Variation of current as a function of bias voltage for DX, DXB, and DXB + SA crystals.

The Raman spectra of quartz, DX, and DXB + SA crystals are shown in Fig. 2d. The pristine quartz substrate showed limited standard peaks but the DX crystals displayed Raman bands of adenine, cytosine, guanine, thymine bases and phosphate backbones with different modes of the DNA molecules. The Raman bands observed at 1244 cm⁻¹, 1418 cm⁻¹ bending and stretching modes in adenine; 655 cm⁻¹, 1290 cm⁻¹, 1345 cm⁻¹ stretching modes in cytosine; 623 cm⁻¹, 930 cm⁻¹, 1576 cm⁻¹ ring and stretching modes in guanine; 780 cm⁻¹, 1290 cm⁻¹, 1465 cm⁻¹ ring breathing and stretching modes in thymine; and 1068 cm⁻¹, 1146 cm⁻¹ symmetric and stretching mode of the phosphate backbone. Otto et al.²³ demonstrated the surface enhanced Raman spectral peaks for each DNA base and Vasudev et al.²⁴ analyzed the vibrational modes of DNA molecules immobilized on substrates. The band assignments depicted in the Raman spectra in our report are in accordance with previous studies.^23–27 Although the Raman spectrum of DXB + SA showed bands similar to DX crystals, we clearly observed the significant intensity changes of corresponding bands.

The basic Raman bands of SA proteins were 962, 1030, 1170, and 1400–1600 cm⁻¹ and the enhanced Raman intensities of DXB + SA crystals were an indication of the strong bonding between the biotinylated DX crystals and SA.

To demonstrate the I–V characteristics of DX, DXB, and DXB + SA crystals, two terminal measurements were performed under atmospheric conditions with silver electrodes and the results are shown in Fig. 2e. For a control and reference, we measured the current as a function of voltage for the pristine quartz substrate without DX crystals and the data indicated no current between the electrodes. In both the forward and reverse bias regions, I–V graphs of all DNA samples seem to be linear and the corresponding resistances of these samples were in the range of a few GΩ. The resistance for an 80 base-pair duplex DNA was in the range of 25–40 GΩ,²⁸ and the resistance of λ-DNA was in the MΩ range with ohmic behavior.²⁹ The considerable differences in the resistances could have depended upon the sample such as dimensional differences (including thickness of DNA samples), its crystalline or amorphous nature, its environment such as dry or wet conditions, and different functionalized substrates. The DX crystals showed relatively higher current than DXB and DXB + SA crystals. It was expected because DNA molecules have negative charges but biotin and SA proteins have almost neutral. When proteins without an excess of charges are conjugated on DX crystals, the net charge carriers of the DXB + SA crystals was relatively reduced within a total volume of samples in dry phase. This phenomenon revealed the lack of charge carriers in the system and consequently, the intrinsic resistivity of the samples could be increased, especially in the dry phase without an excess of ions. Normally the I–V characteristics give information regarding the specific electrical properties of the system which might open the door to applications as molecular devices and biosensors.

To determine the optical band gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) as well as the second band onset of DX, DXB, and DXB + SA, we performed VUV (Visible-Ultraviolet) spectrophotometry and acquired their transmittance spectra at ambient temperature and pressure. Fig. 3a–c shows the absorption coefficient (α) of the DX, DXB, and DXB + SA crystals as a function of photon energy (E). Because irrelevant linear background signals from quartz substrates appeared during absorption measurements, we subtracted them from all of our measured data before spectral analysis. For our DX, DXB, and DXB + SA crystals, the corresponding optical band gaps were 4.74, 4.80 and 4.75 eV, respectively. However, for amorphous DNA molecules in the dry phase, the HOMO–LUMO transitions were positioned in the energy range of 4.3 to 4.5 eV. The slight deviation of the energy gaps between simple duplex DNA and DX crystals might come from different buffer conditions, the amorphous and crystalline nature, multi-stacking or monolayer configuration, and possible interfacial characteristics.^30,31 The fundamental absorption bands associated with the optical band gap presented in Fig. 3 represent the electronic transition of π → π*, essentially a set of superimposed HOMO–LUMO transitions of the corresponding character in a free DNA molecule. The low-energy tail of another strong absorption band appears next to the fundamental absorption band. This feature also come from the set of superimposed second absorption bands of individual DNA molecules located around 6 eV. As shown in Fig. 3d, the optical band gap of protein-conjugated DX crystals was slightly shifted to higher energy. Here, we located the energy of the local maximum and presented the conjugated protein dependence of the optical band gap for the DXB and DXB + SA crystals accordingly. The initial shift of the optical band gap was in agreement with the scheme in which the conjugated proteins blocked the existing interaction with the DX crystals.

Fig. 3 The energy band gap analysis of DX, DXB, and DXB + SA crystals. Variation of absorption coefficient (α) as a function of photon energy (E) for (a) DX, (b) DXB, (c) DXB + SA. (d) Variation of energy band gaps and second band onsets of DX, DXB, and DXB + SA.

Additionally, we presented the onset energy of the second optical absorption band, which evolved with protein conjugation in a different way compared to the optical band gap. Although some uncertainty in locating the onsets through linear fitting from the low-energy side of the second band gap had to be taken into consideration, the overall trend showed an increase in the second band onset with protein conjugation. Such an evolution of the onsets was apparently consistent with our electrical characteristics as well. In fact, in terms of electrical transport characteristics, the second band gap was relatively the more relevant quantity than the optical band gap. Here the optical band gap is supposed to represent the minimum energy to form an electron–hole pair, i.e., exciton, and this resulting two-body state cannot contribute to a transport process that requires free or separated electron–hole pairs. In contrast, the onset energy of the second transition was probably higher than the energy band gap and the electron–hole pairs excited to this band can immediately form current-carrying states after making intraband transitions to the bottom of the conduction band. This well-known phenomenon is commonly found in pentacene-based organic semiconductors as well.^32,33

Conclusions

In summary, we have described the fabrication of protein conjugated 2D DNA crystals on quartz by the SAG method. We analyzed the crystal coverage percentage as a function of DNA monomer concentration C_m and introduced the new concept of DNA structure fabrication – the reusability process – based on the SAG method. Upon protein conjugation with DNA crystals, an enhancement of the Raman band intensity and resistance were observed. The HOMO–LUMO band gaps of DX, DXB, and DXB + SA crystals showed the inverted V-shape and the second band onsets, consistent with electrical characteristics, revealed the increasing behavior with protein conjugation. From our observation, we conclude that it might be possible to tune and control the conductance and optical energy band gap of pristine DX crystals with the conjugation of biotin and streptavidin proteins.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF), (NRF-2013R1A1A2061731) to SRD, (NRF-2009-0083540) to TK, (NRF-2012M3A7B4049802) to JHK and (NRF-2012M3A7B4049801, NRF-2014R1A2A1A11053213) to SHP, funded by the Ministry of Science, ICT & Future Planning (MSIP) and the Ministry of Education, Science and Technology (MEST) of the Korean government.

Notes and references

1. H. Yan, X. Zhang, Z. Shen and N. C. Seeman, Nature, 2002, 415, 62–65.
2. Y. W. Kwon, C. H. Lee, D. H. Choi and J. Jin-II, J. Mater. Chem., 2009, 19, 1353–1380.
3. Y. He, Y. Tian, A. E. Ribbe and C. Mao, J. Am. Chem. Soc., 2006, 128, 15978–15979.
4. X. Wang and N. C. Seeman, J. Am. Chem. Soc., 2007, 129, 8169–8176.
5. C. Lin, Y. Ke, Y. Liu, M. Mertig, J. Gu and H. Yan, Angew. Chem., Int. Ed., 2007, 119, 6201–6204.
6. E. Winfree, F. Liu, L. A. Wenzler and N. C. Seeman, Nature, 1998, 394, 539–544.
7. P. W. K. Rothermund, Nature, 2006, 440, 297–302.
8. Y. He, T. Ye, M. Su, C. Zhang, A. E. Ribbe, W. Jiang and C. Mao, Nature, 2008, 452, 198–201.
9. B. A. R. Williams, K. Lund, Y. Liu, H. Yan and J. C. Chaput, Angew. Chem., Int. Ed., 2007, 46, 3051–3054.
10. N. Stephanopoulos, M. Liu, G. Tong, Z. Li, Y. Liu and H. Yan, Nano Lett., 2010, 10, 2714–2720.
11. K. Keren, R. S. Berman, E. Bushstab, U. Sivan and E. Braun, Science, 2003, 302, 1380–1382.
12. H. T. Maune, S. P. Han, R. D. Barich, M. Bockrath, W. A. Goddard, P. W. K. Rothermund and E. Winfree, Nat. Nanotechnol., 2010, 5, 61–66.
13. S. J. Kim, J. Jung, K. W. Lee, D. Yoon, T. S. Jung, S. R. Dugasani, S. H. Park and H. J. Kim, ACS Appl. Mater. Interfaces, 2013, 5, 10715–10720.
14. S. Pal, Z. Deng, B. Ding, H. Yan and Y. Liu, Angew. Chem., Int. Ed., 2010, 49, 2700–2704.
15. J. Sharma, R. Chhabra, A. Cheng, J. Brownell, Y. Liu and H. Yan, Science, 2009, 323, 112–116.
16. W. Cheng, W. M. J. Campolongo, J. J. Cha, S. J. Tan, C. C. Umbach, D. A. Muller and D. Luo, Nat. Mater., 2009, 8, 519–525.
17. O. I. Wilner, Y. Weizmann, R. Gill, O. Lioubashevski, R. Freeman and I. Willner, Nat. Nanotechnol., 2009, 4, 249–254.
18. S. R. Dugasani, N. H. Lee, J. Lee, B. Kim, S. U. Hwang, K. W. Lee, W. N. Kang and S. H. Park, Sci. Rep, 2013, 3, 1819.
19. S. H. Park, M. W. Prior, T. H. LaBean and G. Finkelstein, Appl. Phys. Lett., 2006, 89, 033901–033903.
20. S. Hamada and S. Murata, Angew. Chem., Int. Ed., 2009, 48, 6820–6823.
21. X. Sun, S. H. Ko, C. Zhang, A. E. Ribbe and C. Mao, J. Am. Chem. Soc., 2009, 131, 13248–13249.
22. S. R. Dugasani, J. A. Kim, B. H. Kim, P. J. Rao, B. Gnapareddy, C. Vyas, T. S. Kim, S. H. Park and V. Manchanda, ACS Appl. Mater. Interfaces, 2014, 6, 2974–2979.
23. C. Otto, T. J. J. van den Tweel, F. F. M. De Mu and J. Greve, J. Raman Spectrosc., 1986, 17, 289–298.
24. M. Vasudev, T. C. Wu, S. Biswas, D. Mitra, M. A. Stroscio, S. Guthrie, M. Reed, K. P. Burris and C. Neal Stewart, IEEE Trans. Nanotechnol., 2011, 10, 35–43.
25. S. Ponkumar, P. Duraisamy and N. Iyandurai, Am. J. Biochem. Biotechnol., 2011, 7, 135–140.
26. C. M. Muntean, R. Misselwitz, L. Dostal and H. Welfle, Spectroscopy, 2006, 20, 29–35.
27. A. Kulkarni, B. Kim, S. R. Dugasani, P. J. Rao, J. Kim, C. Vyas, V. Manichanda, T. S. Kim and S. H. Park, Sci. Rep., 2013, 3, 2062.
28. S. Roy, H. Vedala, A. D. Roy, D. Kim, M. Doud, K. Mathee, H. Shin, N. Shimamoto, V. Prasad and W. Choi, Nano Lett., 2008, 8, 26–30.
29. H. W. Fink and C. Schonenberger, Nature, 1999, 398, 407–410.
30. O. Gullu, M. Cankaya, O. Baris and A. Turut, Appl. Phys. Lett., 2008, 92, 212106–212113.
31. S. Sonmezoglu, O. A. Sonmezoglu, G. Cankaya, A. Yildirim and N. Serin, J. Appl. Phys., 2010, 107, 124518–124526.
32. J. Lee, S. S. Kim, K. Kim, J. H. Kim and S. Im, Appl. Phys. Lett., 2004, 84, 1701–1703.
33. S. S. Kim, S. P. Park, J. H. Kim and S. Im, Thin Solid Films, 2002, 420–421, 19–22.

---

Footnotes

† Electronic supplementary information (ESI) available: Materials and methods, schematic diagram, sequence pool, and sticky-ends of the DX tiles used in this experiment, and reusability calculations. See DOI: 10.1039/c5ra02924d

‡ These authors contributed equally to this work.

---