Langmuir 2009, 25, 311-316
311
Optical Imaging of Surface-Immobilized Oligonucleotide Probes on DNA Microarrays Using Liquid Crystals Siok Lian Lai, Shisheng Huang, Xinyan Bi, and Kun-Lin Yang* Department of Chemical and Biomolecular Engineering, National UniVersity of Singapore, 4 Engineering DriVe 4, Singapore 117576 ReceiVed April 9, 2008. ReVised Manuscript ReceiVed October 13, 2008 In this paper, we report a new label-free method for the imaging of immobilized oligonucleotide probes on DNA microarrays. The imaging principle is based on the disruption of orientations of nematic liquid crystals (LCs), 4-cyano4′-pentylbiphenyl (5CB), by the immobilized oligonucleotides on a surface. Because LCs are birefringent materials, disruption of their orientations by the immobilized oligonucleotides can manifest as optical signals visible to the naked eye. LC cells with two homeotropic boundary conditions, which align 5CB perpendicularly to both surfaces, were developed to deliver a distinct contrast between a dark background and a bright image caused by the immobilized oligonucleotides. This design also allows the quantification of immobilized oligonucleotide concentrations through the interference colors of LCs. The LC-based imaging method has a good signal-to-noise ratio and a clear distinction between positive and negative results and is nondestructive to the immobilized oligonucleotides. These advantages make it a promising means of assessing the quality of DNA microarrays.
Introduction DNA microarrays have been used extensively in the identification and detection of single nucleotide polymorphisms and gene expression levels among many other applications.1-7 These applications require the use of DNA microarrays which possess a high density of oligonucleotide probes with good spot homogeneity. However, current DNA microarray technology still falls short of it.4,8,9 For example, evaporation of buffer solution, which leads to a higher concentration of oligonucleotides at the edge of each spot,8 causes an inhomogeneous distribution of oligonucleotide probes. In fact, this inhomogeneous distribution and other defects on DNA microarrays need to be identified to preserve the accuracy of data interpretation. Currently, several methods are available for quantifying the spatial distribution of DNA immobilized on a surface, such as surface plasmon resonance (SPR) and ellipsometry.5,10-12 The SPR technique requires DNA probes to be immobilized on a gold surface,5,12,13 while ellipsometry requires a reflective surface.5 Neither of them is compatible with commonly used DNA microarray substrates, glass slides. Besides, these techniques also require additional instrumentation. Thus, an imaging technique that is simple, easy to use, and compatible with the current DNA microarray glass substrates is critically needed. * To whom correspondence should be addressed. E-mail: cheyk@ nus.edu.sg. (1) Kurian, K. M.; Watson, C. J.; Wyllie, A. H. J. Pathol. 1999, 187, 267–271. (2) Lockhart, D. J.; Winzeler, E. A. Nature 2000, 405, 827–836. (3) Stears, R. L.; Martinsky, T.; Schena, M. Nat. Med. (N.Y.) 2003, 9, 140– 145. (4) Schaferling, M.; Nagl, S. Anal. Bioanal. Chem. 2006, 385, 500–517. (5) Bally, M.; Halter, M.; Voros, J.; Grandin, H. M. Surf. Interface Anal. 2006, 38, 1442–1458. (6) Heller, M. J. Annu. ReV. Biomed. Eng. 2002, 4, 129–153. (7) Pirrung, M. C. Angew. Chem., Int. Ed. 2002, 41, 1276–1289. (8) Campo, A. d.; Bruce, I. J. Top. Curr. Chem. 2005, 260, 77–111. (9) Heise, C.; Bier, F. F. Top. Curr. Chem. 2006, 261, 1–25. (10) Gray, D. E.; Case-Green, S. C.; Fell, T. S.; Dobson, P. J.; Southern, E. M. Langmuir 1997, 13, 2833–2842. (11) Thiel, A. J.; Frutos, A. G.; Jordan, C. E.; Corn, R. M.; Smith, L. M. Anal. Chem. 1997, 69, 4948–4956. (12) Jordan, C. E.; Frutos, A. G.; Thiel, A. J.; Corn, R. M. Anal. Chem. 1997, 69, 4939–4947. (13) Smith, E. A.; Corn, R. M. Appl. Spectrosc. 2003, 57, 320A–332A.
Here, we investigate the feasibility of using nematic liquid crystals (LCs) as a tool to assess the quality of a DNA microarray and pinpoint its defects prior to the DNA hybridization step. The LC-based imaging method is founded on several previous studies showing that chemical and biomolecular binding events on a solid surface can change the anchoring behaviors of nematic LCs supported on the surface.14-21 This method is performed under ambient lighting without the need of electricity, and the responses are readily observable by the naked eye. For example, Abbott’s group developed several protein assays by using nanostructured gold surfaces which can align supported films of LCs uniformly and give them uniform optical textures. When biomolecules bound to the surface, the surface nanostructures were masked by the biomolecules such that the orientations of the LCs were disrupted, and the optical appearance of the LCs changed as a result. This method is very sensitive, but it is only qualitative at best. Later, the same group developed a torquebalance method to quantitatively measure the orientational transition of LCs triggered by immobilized proteins on the surface.22 However, a series of pictures at different angles need to be captured and analyzed before the amounts of immobilized proteins can be quantified. More recently, we developed protein assays on plain glass substrates which can align LCs homeotropically (perpendicular to the surface).23 Advantages of this approach include (1) it does not require a nanostructured surface, (2) the background is pitch-dark at any angle, making the bright signal triggered by proteins easily distinguishable, and (3) the dark-to-bright transition triggered by adsorbed proteins happens (14) Gupta, V. K.; Skaife, J. J.; Dubrovsky, T. B.; Abbott, N. L. Science 1998, 279, 2077–2080. (15) Shah, R. R.; Abbott, N. L. Science 2001, 293, 1296–1299. (16) Skaife, J. J.; Brake, J. M.; Abbott, N. L. Langmuir 2001, 17, 5448–5457. (17) Clare, B. H.; Abbott, N. L. Langmuir 2005, 21, 6451–6461. (18) Clare, B. H.; Guzma´n, O.; de Pablo, J. J.; Abbott, N. L. Langmuir 2006, 22, 4654–4659. (19) Bi, X.; Yang, K.-L. Colloids Surf., A 2007, 302(1-3), 573–580. (20) Bi, X.; Huang, S.; Hartono, D.; Yang, K.-L. Sens. Actuators, B 2007, 127(2), 406–413. (21) Kim, H.-R.; Kim, J.-H.; Kim, T.-S.; Oh, S.-W.; Choi, E.-Y. Appl. Phys. Lett. 2005, 87, 143901. (22) Lowe, A. M.; Bertics, P. J.; Abbott, N. L. Anal. Chem. 2008, 80, 2637– 2645. (23) Xue, C. Y.; Yang, K. L. Langmuir 2008, 24, 563–567.
10.1021/la802672b CCC: $40.75 2009 American Chemical Society Published on Web 12/09/2008
312 Langmuir, Vol. 25, No. 1, 2009
at a very narrow protein concentration (1 µM) used in the preparation of most DNA microarrays.26-28 This observation is also in good agreement with a past study29 which shows that when the spot size is reduced, the immobilization rate of oligonucleotides is increased. This will in turn increase the surface density of immobilized oligonucleotides in that area and disrupt the orientations of the LCs. Another advantage of reducing the spot size lies in the improvement in the uniformity of the spots. In fact, a correlation between the oligonucleotide concentration and the interference color of the LC can be established from Figure 5, which shows that the spot color follows the order of blue, purple, red, and gray when the oligonucleotide concentration is decreased. The change in the interference color is, in fact, caused by different tilt angles of the LCs as predicted by the Michel-Levy chart.30 The correlation between the DNA concentration and LC interference color is very useful, because one can tell that any gray spots on the microarrays correspond to low DNA density (i.e., below 0.5 µM). Assessing the Quality of a DNA Microarray Using LCs. Since our original objective is to develop a method for assessing the quality of a DNA microarray before use, our new substrate and the imaging technique need to meet the following requirements: (1) LCs must report the spatial distribution of immobilized (26) Kim, J.; Crooks, R. M. Anal. Chem. 2007, 79, 7267–7274. (27) Gerion, D.; Chen, F.; Kannan, B.; Fu, A.; Parak, W. J.; Chen, D. J.; Majumdar, A.; Alivisatos, A. P. Anal. Chem. 2003, 75, 4766–4772. (28) Rozkiewicz, D. I.; Brugman, W.; Kerkhoven, R. M.; Ravoo, B. J.; Reinhoudt, D. N. J. Am. Chem. Soc. 2007, 129, 11593–11599. (29) Dandy, D. S.; Wu, P.; Grainger, D. W. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 8223–8228. (30) Robinson, P. C.; Davidson, M. W. Michel-Levy Interference Color Chart. http://www.microscopyu.com/articles/polarized/michel-levy.html.
Langmuir, Vol. 25, No. 1, 2009 315
Figure 7. (A) Fluorescence-labeled DNA targets were hybridized to a DNA microarray which was not contacted with LCs. The intensity plot across five fluorescence spots showed an average intensity value of 60. (B) Fluorescence-labeled DNA targets were hybridized to a DNA microarray after the microarray was contacted with LCs and the LCs were removed. The intensity plot across five fluorescence spots showed an average intensity value of 60, which was comparable with that of (A).
oligonucleotides faithfully. This property can allow us to pinpoint defects on a DNA microarray. (2) After the imaging process, one can easily remove LCs from the surface without affecting the immobilized oligonucleotides. The immobilized oligonucleotides can still perform as well as before contact with LCs. To test whether our LC-based imaging method meets these requirements, we first prepared a DNA microarray by spotting a 10 µM concentration of oligonucleotides on a TEA/DMOAP-coated slide. Subsequently, the substrate was imaged with 5CB to assess its quality. Similar to our previous results, parts A and B of Figure 6 show distinct colorful spots, which are caused by the presence of immobilized oligonucleotides. Interestingly, close inspection of these spots also reveals certain defects which appear as dark holes (highlighted as a red box in Figure 6A), which are probably caused by the absence or a very low density of immobilized oligonucleotides in these regions. Furthermore, our LC image also reflects two irregular spots (highlighted as a red box in Figure 6B), which might be caused by some dust on the surface during the spot printing. We hypothesize that the dark portion of the hemicircle is also caused by the absence of immobilized oligonucleotides and the protruding portion of the second spot is caused by splashing of the oligonucleotide solution
316 Langmuir, Vol. 25, No. 1, 2009
during the spotting process. These observations suggest that many kinds of defects may exist due to surface inhomogeneities or defects on a DNA microarray even when these microarrays are printed by an automated robot. Without knowing the presence of these defects on the surface, a DNA microarray may provide misleading information after the DNA hybridization. To test our hypothesis and determine whether the actual spatial distribution of immobilized oligonucleotides is faithfully reflected by the LC image, we disassembled the LC cell and cleaned it with ethanol and DI water to remove LCs from the surfaces. Since the immobilized oligonucleotides were not fluorescently labeled and could not be examined directly, the DNA microarrays were hybridized with complementary FAM-labeled DNA targets (oligonucleotide D4) first. Parts C-E of Figure 6 show fluorescence images of the same spots (as highlighted in Figure 6A,B) after the hybridization. From Figure 6C-E, we can make several observations. First, the number and position of the dark spots in Figure 6C match those shown in Figure 6A. This result supports our hypothesis that the dark spots are due to the absence or very low density of immobilized oligonucleotides. Second, the shapes of the fluorescent images shown in Figure 6D,E also match those in Figure 6B. Thus, we can conclude that LCs are very reliable in reporting the spatial distribution of immobilized oligonucleotides. Third, the fluorescence signals observed from Figure 6C-E suggest that our newly developed substrate is able to withstand the hybridization procedures; DNA targets are able to hybridize to their complementary probes on the surface. This result is important because it shows that the immobilized oligonucleotide probes are intact on the surface after they are “imaged” by the LCs. To provide further evidence for the nondestructive property of LCs for immobilized oligonucleotides, we prepared two DNA microarrays with a 0.5 µM concentration of oligonucleotides spotted on two TEA/DMOAP-coated slides. One of the slides was imaged with 5CB, while the other served as a control. After the LC imaging process, the LC cell made from the first slide was disassembled and 5CB was rinsed off by using ethanol and DI water. Both microarrays were then hybridized with complementary FAM-labeled DNA targets (oligonucleotide D4). Fluorescence images and their intensity profiles are shown in Figure 7. Apparently, intensity profiles across five different spots on both microarrays are comparable, suggesting that the microarray which has been in contact with LCs performs as well as that without contact with LCs. This leads us to conclude that
Lai et al.
this LC-based imaging technique is relatively nondestructive. LCs are deposited on the surface to investigate the spatial distribution of immobilized oligonucleotides and then removed from the surface.
Conclusion In summary, we have demonstrated a new label-free method for the imaging of oligonucleotides immobilized on DNA microarrays with LCs. We have developed a mixed TEA/DMOAP surface, which presented aldehyde functional groups and provided a dark backdrop, as a new DNA microarray substrate to accommodate this new imaging technique. After the immobilization of amine-labeled oligonucleotides and overlaying a thin layer of LCs, the optical image showed a very clear contrast between the oligonucleotide-bound areas and the dark background. We also noted that a clear transition of the optical appearance of LCs from dark to bright occurred at a particular oligonucleotide concentration. When the spot size was reduced, the detection limit was improved. This was accompanied by a correlation between interference colors and oligonucleotide concentrations that serves as a basis for the quantification of immobilized oligonucleotides on a DNA microarray. Besides, the LC-based imaging method had been shown to report the spatial distribution of immobilized oligonucleotides reliably. This enables the defects on DNA microarrays to be pinpointed before use. More importantly, the LC-based imaging method is nondestructive to the immobilized oligonucleotides. After removal of LCs from the surface, immobilized oligonucleotide probes were still able to hybridize with their complementary targets as was evident in our fluorescence experiments. The developed LC imaging method, which is simple and easy to use, may provide an unconventional yet powerful tool for assessing the quality of DNA microarrays and locating all surface defects before they are commissioned. This information is crucial for interpreting DNA microarray data correctly after DNA hybridization. Acknowledgment. The work was funded by the Agency for Science and Technology Research (ASTAR) in Singapore under Project No. 0521010099. Supporting Information Available: Characterization of the pure TEA and mixed TEA/DMOAP surfaces by using ellipsometry and water contact angle measurements. This material is available free of charge via the Internet at http://pubs.acs.org. LA802672B
