Evidence of Impurities in Thiolated Single-Stranded DNA Oligomers

Chi-Ying Lee, Heather E. Canavan, Lara J. Gamble, and David G. Castner* ... Thomas Wirth , Riccardo Castelli , Peter H. Seeberger , and Wolfgang E. S...
0 downloads 0 Views 220KB Size
Download PDF
5134

Langmuir 2005, 21, 5134-5141

Evidence of Impurities in Thiolated Single-Stranded DNA Oligomers and Their Effect on DNA Self-Assembly on Gold Chi-Ying Lee,†,§ Heather E. Canavan,†,‡ Lara J. Gamble,†,‡ and David G. Castner*,†,‡,§ National ESCA and Surface Analysis Center for Biomedical Problems, Center for Nanotechnology and Departments of Bioengineering and Chemical Engineering, Box 351750, University of Washington, Seattle, Washington 98195 Received November 10, 2004. In Final Form: February 22, 2005 The diversity of techniques used in the synthesis, treatment, and purification of the single-stranded DNA oligomers containing a thiol anchor group (SH-ssDNA) has led to a significant variation in the purity of commercially available SH-ssDNA. In this work, we use X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (ToF-SIMS) to study how the impurities present in commercially synthesized SH-ssDNA oligomers affected the structure of the resulting DNA films on Au. XPS results indicate that two of the purchased SH-ssDNA oligomers contain excess carbon and sulfur. The molecular fragmentation patterns obtained with ToF-SIMS were used to determine the identity of several contaminants in the DNA films, including poly(dimethylsiloxane) (PDMS), lipid molecules, and sulfurcontaining molecules. In particular, the ToF-SIMS results determined that the excess sulfur detected by XPS was due to the presence of dithiothreitol, a reductant often used to cleave disulfide precursors. Furthermore, we found that the SH-ssDNA self-assembly process is affected by the presence of these contaminants. When relatively pure SH-ssDNA is used to prepare the DNA films, the P, N, O, and C atomic percentages were observed by XPS to increase over a 24-h time period. In contrast, surfaces prepared using SH-ssDNA containing higher levels of contaminants did not follow this trend. XPS result indicates that, after the initial SH-ssDNA adsorption, the remaining material incorporated into these films was due to contamination.

Introduction DNA-modified surfaces have received considerable attention in the fields of bio- and nanotechnology due to their importance in the development of biosensing and diagnostic devices such as DNA microarrays and nanowire assemblies. The construction of these surfaces often entails the attachment of presynthesized oligonucleotides onto a derivatized surface. Single-stranded DNA oligomers containing a thiol anchor group (SH-ssDNA) are widely used for DNA probe immoblization due to their ease of use and the well-established thiol chemistry in the literature.1-25 * Author to whom correspondence should be addressed. E-mail: [email protected]. Tel: 206-543-8094. Fax: 206543-3778. † National ESCA and Surface Analysis Center for Biomedical Problems. ‡ Department of Bioengineering, University of Washington. § Department of Chemical Engineering, University of Washington. (1) Demers, L. M.; Ginger, D. S.; Park, S. J.; Li, Z.; Chung, S. W.; Mirkin, C. A. Science 2002, 296, 1836-1838. (2) Letsinger, R. L.; Mirkin, C. A.; Elghanian, R.; Mucic, R. C.; Storhoff, J. J. Phosphorus, Sulfur Silicon Relat. Elem. 1999, 146, 359-362. (3) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607-609. (4) Zhang, H.; Li, Z.; Mirkin, C. A. Adv. Mater. 2002, 14, 1472-1474. (5) Herne, T. M.; Tarlov, M. J. J. Am. Chem. Soc. 1997, 119, 89168920. (6) Kimura-Suda, H.; Petrovykh, D. Y.; Tarlov, M. J.; Whitman, L. J. J. Am. Chem. Soc. 2003, 125, 9014-9015. (7) Levicky, R.; Herne, T. M.; Tarlov, M. J.; Satija, S. K. J. Am. Chem. Soc. 1998, 120, 9787-9792. (8) Petrovykh, D. Y.; Kimura-Suda, H.; Whitman, L. J.; Tarlov, M. J. J. Am. Chem. Soc. 2003, 125, 5219-5226. (9) Petrovykh, D. Y.; Kimura-Suda, H.; Tarlov, M. J.; Whitman, L. J. Langmuir 2004, 20, 429-440. (10) Steel, A. B.; Herne, T. M.; Tarlov, M. J. Anal. Chem. 1998, 70, 4670-4677. (11) Steel, A. B.; Levicky, R. L.; Herne, T. M.; Tarlov, M. J. Biophys. J. 2000, 79, 975-981.

Self-assembly of SH-ssDNA from solution is one of the most common approaches to immobilize DNA on gold.1-15 In addition, a variety of synthetic chemical approaches have also been developed to immobilize SH-ssDNA onto functionalized glass,16 silicon,17-21 or gold22-25 surfaces using heterobifunctional cross-linkers. The diversity of techniques used in the synthesis, treatment, and purification of the SH-ssDNA oligomers will naturally lead to a significant variation in the purity of commercially available SH-ssDNA. Although the automated chemical synthesis of oligonucleotides is a wellestablished process,26,27 purification of the synthesized (12) Heaton, R. J.; Peterson, A. W.; Georgiadis, R. M. PNAS 2001, 98, 3701-3704. (13) Peterson, A. W.; Heaton, R. J.; Georgiadis, R. M. Nucleic Acids Res. 2001, 29, 5163-5168. (14) Peterson, A. W.; Wolf, L. K.; Georgiadis, R. M. J. Am. Chem. Soc. 2002, 124, 14601-14607. (15) Wolf, L. K.; Gao, Y.; Georgiadis, R. M. Langmuir 2004, 20, 33573361. (16) Charles, P. T.; Vora, G. J.; Andreadis, J. D.; Fortney, A. J.; Meador, C. E.; Dulcey, C. S.; Stenger, D. A. Langmuir 2003, 19, 15861591. (17) Lin, Z.; Strother, T.; Cai, W.; Cao, X. P.; Smith, L. M.; Hamers, R. J. Langmuir 2002, 18, 788-796. (18) Lenigk, R.; Carles, M.; Ip, N. Y.; Sucher, N. J. Langmuir 2001, 17, 2497-2501. (19) Strother, T.; Hamers, R. J.; Smith, L. M. Nucleic Acids Res. 2000, 28, 3535-3541. (20) Strother, T.; Cai, W.; Zhao, X. S.; Hamers, R. J.; Smith, L. M. J. Am. Chem. Soc. 2000, 122, 1205-1209. (21) Chrisey, L. A.; Lee, G. U.; Oferrall, C. E. Nucleic Acids Res. 1996, 24, 3031-3039. (22) Johnson, P. A.; Levicky, R. Langmuir 2003, 19, 10288-10294. (23) Brockman, J. M.; Frutos, A. G.; Corn, R. M. J. Am. Chem. Soc. 1999, 121, 8044-8051. (24) Nelson, B. P.; Grimsrud, T. E.; Liles, M. R.; Goodman, R. M.; Corn, R. M. Anal. Chem. 2001, 73, 1-7. (25) Smith, E. A.; Wanat, M. J.; Cheng, Y. F.; Barreira, S. V. P.; Frutos, A. G.; Corn, R. M. Langmuir 2001, 17, 2502-2507. (26) Goodchild, J. Bioconjugate Chem. 1990, 1, 165-187.

10.1021/la0472302 CCC: $30.25 © 2005 American Chemical Society Published on Web 04/22/2005

Evidence of Impurities in Thiolated ssDNA Oligomers

DNA product containing a thiol modifier is far from optimal and methods used vary from vendor to vendor. To obtain meaningful and reproducible results for biotechnological applications, it is necessary to start with high-purity DNA. The presence of contaminants may change the structure and coverage of the DNA film on the surface. This, in turn, should alter the hybridization efficiency of the immobilized DNA probe with the targets in solution. Although many techniques have been applied to study surface-bound SH-ssDNA oligomers,5-7,15,21,28-30 only a few studies have provided detailed characterization of DNA films.8,9,17-20 Furthermore, none of these studies have investigated the purity and reproducibility of the surfacebound SH-ssDNA. Note: while purity and reproducibility can be defined in many ways (e.g., batch-to-batch variation within a single vendor’s supply or vendor-to-vendor variation of the same sequence), this paper refers to the variability in the purity of DNA films prepared from thiolated oligomers of the same sequence from different vendors. In this study, we have used X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (ToF-SIMS) to study the self-assembly process of SH-ssDNA oligomers onto gold surfaces and determined how the impurities present in commercially synthesized SH-ssDNA oligomers affect the structure of the resulting films. Using XPS, quantitative atomic compositions of the individual DNA films prepared from SH-ssDNA oligomers from three vendors were obtained and compared. The molecular fragmentation patterns acquired with ToF-SIMS were used to determine the identity of several contaminants present in some of the DNA films. Experimental Section Materials. HPLC-purified 3′ and 5′ thiol-modified DNA oligomers of the same sequence and length [5′-AAAAAAAACCCCCCCC-(CH2)6-SH-3′ and 5′-HS-(CH2)6-CCCCCCCCAAAAAAAA-3′, 16 nucleotides] from three vendors were used for the comparative studies described here. The 3′ thiol-modified oligonucleotides were purchased from Alpha DNA (Montreal, Canada), and the 5′ thiol-modified oligonucleotides were purchased from Trilink Biotechnologies (San Diego, CA) and Synthegen (Houston, TX). (Note: The authors have no financial interest in companies Alpha DNA, Trilink, and Synthegen). The sequence is designed to be non-self-complementary and contain only two of the four DNA bases for simplicity. Oligonucleotides from Synthegen were treated with 0.1 M dithiothreitol (DTT) provided by the vendor and re-purified with a NAP-10 column as outlined by the vendor.31 Briefly, 100 µL of 0.1 M DTT solution was added to the oligonucleotide lyophilized powder and allowed to stand at room temperature overnight. Deionized water (900 µL) was added to the tube containing the oligonucleotide solution and pipetted up and down to mix. The NAP-10 column was rinsed with deionized water a total of three times by filling up the top portion of the column with deionized water and allowing all the water to drain through the column. The oligonucleotide solution (1.0 mL volume) was loaded onto the NAP-10 column. Deionized water (1.5 mL) was pipetted onto the NAP-10 column for sample elution, and all of the liquid that dripped out of the column containing the oligonucleotide was collected in a 2 mL tube. The oligonucleotide lyophilized powders from Alpha DNA and TriLink were dissolved in deionized water, stored at -80 °C until needed, (27) Efcavitch, J. W. In Macromolecular Sequencing and Synthesis. Selected Methods and Applications; Alan R. Liss, Inc.: New York, 1998; pp 221-234. (28) Hook, F.; Ray, A.; Norden, B.; Kasemo, B. Langmuir 2001, 17, 8305-8312. (29) Liu, M. Z.; Amro, N. A.; Chow, C. S.; Liu, G. Y. Nano Lett. 2002, 2, 863-867. (30) Wang, J.; Bard, A. J. Anal. Chem. 2001, 73, 2207-2212. (31) Synthegen. DTT treatment of thiol-modified oligonucleotides.

Langmuir, Vol. 21, No. 11, 2005 5135 and used without additional treatment. (It is noted that oligonucleotides from Trilink and Alpha were not deprotected for the following reasons: DNA from Trilink is normally shipped as either the disulfide or the free thiol without any protective group; DNA from Alpha DNA is normally deprotected by the vendor prior to shipping.) Control experiments were used to determine if all of the DTT or tris(2-carboxyethyl)phosphine (TCEP) present in the solutions would be completely removed after purification with the NAP10 column. DTT and TCEP solutions (0.01 M, 1.0 mL) without oligonucleotides were purified with same NAP-10 column procedure as described above. DNA concentrations were verified with an UV-vis spectrophotometer (Hitachi U-2000) by measuring the absorption at 260 nm. The buffer used for the solution assembly of DNA contained 1.0 M NaCl (Fisher, Fair Lawn, NJ), 10 mM Tris-HCl (Sigma, St. Louis, MO), and 1 mM ethylenediaminetetraacetic acid (EDTA, Fisher, Fair Lawn, NJ) and was adjusted to pH 7.0 by adding 1.0 M NaOH. Millipore-filtered water was used for all aqueous solutions and rinsing. Substrate Preparation. Gold-coated silicon wafers were used as substrate material for DNA immobilization. Silicon wafers (Silicon Valley Microelectronics, Inc., San Jose) were diced into 10 mm × 10 mm squares and cleaned by sonication in DI water, methylene chloride (EMD, Gibbstown, NJ), acetone (Fisher, Fair Lawn, NJ), and methanol (EMD, Gibbstown, NJ) for 5 min two times in each solvent. Substrates were coated by electron beam evaporation of a 100 Å titanium adhesive layer followed by a 1000 Å gold layer (99.99%) at pressures below 1 × 10-6 Torr. Thiol-DNA Assembly. Gold-coated substrates from the electron beam evaporator were immersed in 1 µM DNA solutions in 1 M NaCl-TE (1 M NaCl, 10 mM Tris-HCl, 1 mM EDTA) at room temperature for varying times (5 min, 30 min, 2 h, 5 h, and 24 h) to allow for the adsorption and self-assembly of the DNA thin film. After the specified assembly time, samples were removed from the solution and rinsed thoroughly in deionized water for 1 min to remove loosely bound DNA and buffer salts. Samples were then blown dry with nitrogen and stored under nitrogen until analysis. DTT Samples for ToF-SIMS Analysis. DTT samples were prepared by immersing the gold-coated substrates in 1 mM DTT solution supplied by Synthegen for 1 h. After adsorption, samples were rinsed in deionized water, dried with nitrogen, and stored under nitrogen until analysis. XPS Analysis. XPS measurements were performed on Surface Science Instruments X-Probe and S-Probe spectrometers. Both instruments are equipped with a monochromatized aluminum KR X-ray source and a hemispherical electron energy analyzer. Compositional survey and detail scans (P2p, C1s, N1s, O1s, and S2p) were acquired using a pass energy of 150 eV. High-resolution spectra (C1s and S2p) were acquired on the DNA samples using a pass energy of 50 eV. For the high-resolution spectra, peak binding energies were referenced to the C1s (C-C/C-H) peak at 285.0 eV. Three spots on two or more replicates of each DNA sample were analyzed. The compositional data are averages of the values determined at each spot. Data analysis was performed on the Service Physics ESCA 2000 Graphics Viewer data reduction software (see ref 32 for details on XPS atomic percent calculation). ToF-SIMS Analysis. ToF-SIMS spectra were acquired on a Physical Electronics PHI 7200 time-of flight spectrometer using an 8 keV Cs+ primary ion source in the pulsed mode. Spectra were acquired for both positive and negative secondary ions over a mass range of (m/z) 0-1000 for the DNA samples and 0-500 for the DTT samples. The area of analysis for each spectrum was 100 µm × 100 µm, and the total ion dose was maintained below 2 × 1012 ions/cm2. The mass resolutions (m/∆m) for the positive and negative spectra were typically between 6000 and 8700 for the (m/z) 27 and 25 peaks in the positive and negative spectra, respectively. Three spots on two replicates of each sample were examined. Positive ion spectra were calibrated using the CH3+,C2H3+, and AuSCH2+ peaks, and negative ion spectra were calibrated using the CH-, C2H-, and AuS- peaks. Calibration errors were kept below 10 ppm. (32) Tyler, B. J.; Castner, D. G.; Ratner, B. D. Surf. Interface Anal. 1989, 14, 443-450.

5136

Langmuir, Vol. 21, No. 11, 2005

Lee et al.

Table 1. Average Elemental Compositions as Determined by XPS for DNA Films Prepared from Commercially Synthesized SH-ssDNA Oligomers from Different Vendorsa and Average Elemental Compositions as Determined by XPS for the DTT and TCEP Solutions (0.01 M) Collected upon Running through a NAP-10 Columnb a

Commercially Synthesiszed SH-ssDNA atomic percent sample

Alpha TriLink Synthegen DNA theoretical

atomic ratio

P2p

C1s

N1s

O1s

S2p

Au4f

P/N

C/N

O/N

1.1 2.1 1.1

41.0 43.9 40.7

5.6 8.5 5.6

17.6 17.8 16.4

1.9 0.0 3.6

32.9 27.8 32.5

0.20 0.24 0.20 0.27

7.4 5.2 7.5 2.6

3.2 2.1 3.0 1.4

b

DTT and TCEP Solutions atomic percent sample

DTT flow through TCEP flow through DTT theoretical TCEP theoretical

P2p

C1s

N1s

O1s

S2p

Au4f

0.0 0.7 0.0 6.3

36.9 38.4 50.0 56.3

0.0 0.0 0.0 0.0

13.4 10.9 25.0 37.4

4.5 0.0 25.0 0.0

45.2 50.0 0.0 0.0

a The films were self-assembled for 24 h in 1 µM SH-ssDNA solution in 1 M NaCl-TE buffer onto gold. XPS reveals compositional differences for the same DNA sequence. b XPS confirms the presence of excess DTT and TCEP in “purified” solutions. All standard deviations < 2%.

Principal Components Analysis (PCA). PCA has been described extensively in the literature (see refs 33 and 34 for the basic concepts). Briefly, PCA results in a matrix rotation that creates a new set of axes called principal components (PCs) that define the directions of major variation within the data set. Each PC axis is a linear combination of the original variables in the data set. As the original variables are recombined to define the new PC axes, the number of variables is reduced. The PCA matrix decomposition gives three new matrixes: the scores, the loadings, and the residuals. The scores are the projection of the spectra onto each PC and describe the relationship or spread of the spectra in the data set. The loadings then define the contributions of the original variables in the ToF-SIMS data set to the new PCs and describe which variables are responsible for the differences seen within the samples. On the basis of the trends seen in the PCA scores and loadings plots, one can relate information from the PCA model back to the original data. In general, for a given PC axis, samples with positive scores correspond with variables with positive loadings. This means, for example, that samples with high positive scores on a given PC will have higher relative intensities of peaks that have positive loadings on the same PC axis. PCA of the spectra was performed using PLS Toolbox v. 2.0 (Eigenvector Research, Manson, WA) for MATLAB (the MathWorks, Inc., Natick, MA). Prior to PCA, representative spectra from the DNA films prepared from SH-ssDNA oligomers from Alpha and TriLink were overlaid and peaks were selected using the Tofpak software. A peak list was created including all peaks that were at least 3 times the background for the given region, including some peaks related to the fragmentation of the DNA molecules, which have previously been reported.35,36 All ToFSIMS spectra were normalized to the total intensity of the selected peaks. Data were mean-centered to ensure that the variance in the data set was due to differences in sample variances rather than differences in sample means.

Results and Discussion Variation in SH-ssDNA Film Compositions as Detected by XPS. XPS is uniquely suited for the study of immobilized DNA oligomers, as it is a quantitative measure of the atomic composition of the outermost 100 Å of a sample surface.37 Using XPS, we determined the (33) Jackson, J. E. J. Qual. Technol. 1980, 12, 201-213. (34) Wold, S.; Esbensen, K.; Geladi, P. Chemom. Intell. Lab. Syst. 1987, 2, 37-52. (35) May, C. J.; Canavan, H. E.; Castner, D. G. Anal. Chem. 2004, 76, 1114-1122. (36) Samuel, N. T.; Castner, D. G. Appl. Surf. Sci. 2004, 231-2, 397-401. (37) Castner, D. G.; Ratner, B. D. Surf. Sci. 2002, 500, 28-60.

compositions of identical DNA sequence from three vendors. The determined surface compositions and the theoretical composition of the SH-ssDNA sequence are shown in Table 1a. In all films analyzed, only the principal elements (P, C, N, O, S, and Au) expected from the SH-ssDNA and the substrate were detected. Compositional differences were observed among the SH-ssDNA films produced from the same-sequence oligomers purchased from different vendors. DNA films prepared from SH-ssDNA oligomers from two of the vendors (Alpha and Synthegen) contain less phosphorus (1 atomic percent) and nitrogen (6 atomic percent) and more gold (33 atomic percent) compared to that obtained from TriLink DNA (P ) 2, N ) 9, and Au ) 28 atomic percent). This indicates that films assembled from Alpha and Synthegen SH-ssDNA have lower DNA surface coverage compared to the DNA surface prepared from TriLink SH-ssDNA. Furthermore, the greatest compositional variation is seen in the S2p signal. DNA films prepared from SHssDNA oligomers purchased from Alpha and Synthegen contain excess sulfur (2-4%). We would not expect to observe the S2p signal because of the very low relative concentration of sulfur (0.3%) in the SH-ssDNA molecule combined with the signal attenuation from inelastic scattering of the outgoing S2p photoelectrons by the overlying DNA film. Examination of the C1s high-resolution XPS spectrum reveals further differences. Although XPS composition data showed similar carbon percentages among the three SH-ssDNA films, the high-resolution C1s spectra revealed differences in the relative concentrations of the different carbon species present in the thiolated oligos (see Figure 1). As previously reported by May et al.,35 the carbon species expected from the thiolated oligo include hydrocarbon (C-C/C-H), carbon bound to nitrogen and oxygen (C-N, C-O, N-CdN, O-C-N), amide carbon (N-Cd O), and urea carbon [N-C(dO)-N] with characteristic binding energies (BEs) of approximately 285, 286-287.8, 288, and 289 eV, respectively. As seen in Table 2, Alpha and Synthegen SH-ssDNA films showed excess amounts of C-C/C-H species (approximately 55%) and low concentrations of C-N, C-O, N-CdN and O-C-N species (approximately 28% for C-N and C-O and 13% for N-Cd N and O-C-N). The presence of excess sulfur combined with low phosphorus and nitrogen concentrations indicates that

Evidence of Impurities in Thiolated ssDNA Oligomers

Langmuir, Vol. 21, No. 11, 2005 5137

Figure 1. C1s high-resolution XPS spectra of an identical DNA sequence [5′-AAAAAAAACCCCCCCC-(CH2)6-SH-3′ and 5′HS-(CH2)6-CCCCCCCCAAAAAAAA-3′] from three vendors TriLink (top), Synthegen (middle), and Alpha (bottom). DNA samples prepared from 1 µM SH-ssDNA solution in 1 M NaClTE buffer produce films with varying concentrations of carbon species.

Figure 2. Schematic representation of the thiol deprotection process. When the protected thiol-modified oligo is exposed to DTT or TCEP, the disulfide linkage is broken and the free thiol species are generated.

the films contain contaminants other than the adventitious hydrocarbon normally found. The excess sulfur found in these samples suggests that the contamination maybe due to the presence of excess DTT, a reductant commonly used to cleave disulfide precursors. Disulfide reductants such as DTT or TCEP are often used after the SH-ssDNA synthesis to cleave the disulfide linkage.15,21-23,38 This treatment is done to regenerate free thiol groups for the SH-ssDNA that have dimerized over time or to remove the thiol protective group, mercaptohexyldimethoxytrityl [SH-(CH2)6-DMT], from the SH-ssDNA.39 This thiol protective group is not acid labile, and therefore cannot be removed on a DNA synthesizer using the normal acid deprotection procedure; thus, they are usually removed by using DTT or TCEP to cleave the disulfide linkage (see Figure 2). However, these reagents, as well as the cleaved protective group, must then be separated from the SH-ssDNA oligomers before the oligomers are attached to the surface since both DTT and SH-(CH2)6-DMT contain thiol groups that will adsorb onto gold or react

Figure 3. Effects of excess DTT on SH-ssDNA surface assembly, as indicated by the decreasing N and increasing S atomic percentages from DNA films prepared from solutions containing 1 µM SH-ssDNA and increasing concentration of DTT (0.001 µM to 10 mM). Each set is an average of three analysis spots on two samples.

with maleimides. Even TCEP may interfere with the same chemistry, as TCEP has been reported to react with haloacetamides or maleimides under certain conditions.40,41 Unfortunately, removal of excess DTT or TCEP by gel-filtration, as is typically reported in the literature, may not be a 100% efficient process. Additional experiments were performed to verify if DTT and TCEP can be completely removed from solution using the gel-filtration procedure.31 Aliquots of DTT and TCEP solutions with concentrations normally used to treat SH-ssDNA were collected upon being run through a NAP10 column. Surfaces obtained by adsorption of the collected fractions to gold were subsequently analyzed by XPS. We observed that up to 4.5% sulfur was detected from the DTT flow-through samples and 0.7% phosphorus was detected from the TCEP flow-through (Table 1b) samples. These results confirm that the gel-filtration procedure commonly used by the vendors and academic research groups to purify SH-ssDNA after DTT or TCEP treatment does not remove all the disulfide reductants. For surface-modification applications it is useful to determine the residual DTT concentration threshold that would interfere with the SH-ssDNA assembly onto the gold surface. It is also useful to determine the residual DTT concentration in solution after gel-filtration. DNA films prepared from solutions containing 1 µM SH-ssDNA and varying concentrations of DTT (0.001 µM to 10 mM) were analyzed by XPS. The N (from DNA) and S (from DTT) atomic percentages from these DNA films are shown in Figure 3. XPS data show with just 0.001 µM of DTT added to the solution the N atomic percentage is significantly decreased indicating lower DNA coverage on surface. The N atomic percentage continues to decrease as the concentration of DTT in solution is increased. At 100 µM DTT, no N is detected from the sample, indicating no DNA is observed by XPS. In addition, increased levels of S are observed on DNA films prepared from solutions containing 0.1 µM and higher DTT concentrations. On the basis of this set of control experiments, it appears

Table 2. C1s High-Resolution XPS Data Compiled for Identical DNA Sequences [5′-AAAAAAAACCCCCCCC-(CH2)6-SH-3′and 5′-HS-(CH2)6-CCCCCCCCAAAAAAAA-3′] from Three Vendors (TriLink, Synthegen, and Alpha) percentage sample

C-C, C-H (285 eV)

C-N, C-O (286.6 eV)

O-C-N, NdC-N (288 eV)

N-C(dO)-N (289.5 eV)

Alpha TriLink Synthegen theoretical

55 46 52 17

26 39 32 49

14 10 12 29

5 5 4 5

5138

Langmuir, Vol. 21, No. 11, 2005

that DTT concentrations as low as 0.001 µM were sufficient to affect the self-assembly of SH-ssDNA onto gold. The residual DTT concentration in solution after gelfiltration can be indirectly estimated by comparing the surface composition of the DNA film prepared from Synthegen SH-ssDNA (which was treated with DTT and purified using a NAP-10 column) to the compositions obtained from the DNA/DTT films in Figure 3. On the basis of this comparison, it is estimated that the amount of DTT remaining in the Synthegen SH-ssDNA sample after gel-filtration is approximately 10 µM. However, it should be noted that this approximation does not take into consideration of the fact that the DMT protective group with a free thiol from the Synthegen SH-ssDNA could also contribute to the S signal observed from the Synthegen DNA film if it was not removed by the gelfiltration process after the DTT treatment. Identification of Contaminants by ToF-SIMS. Although XPS results indicate the presence of contaminants in some of the DNA films, it is difficult to fully identify the contaminants with XPS alone since XPS provides mainly elemental and chemical-state information. However, by combining the use of ToF-SIMS, a technique that is extremely surface-sensitive, contaminants in the DNA film can be identified by their molecular fragments. Previously, May et al.35 have demonstrated that ToF-SIMS can not only be used to identify the unique positive and negative ions from the nucleobases, nucleosides, and nucleotides but is also capable of discriminating between two oligomers of extremely similar base compositions using multivariate analysis. DNA sequences from Alpha and TriLink were analyzed as self-assembled thin layers on gold (note that, since DNA films prepared from Alpha and Synthegen oligos show similar surface compositions as determined by XPS, only DNA from Alpha was used as a comparison to DNA from TriLink in the ToF-SIMS analysis). The positive- and negative-ion ToF-SIMS spectra of the DNA film prepared from SH-ssDNA oligomers from Alpha are presented in Figure 4. In Figure 4a, key molecular ions from the adenine (A + H, m/z ) 136.06) and cytosine (C + H, m/z ) 112.05) bases, as well as their fragments (A - NH3 + H, A - HCN + H, A - NH2CN + H, etc.), were observed. Fragments of the sugar were also identified in the positive-ion spectra at m/z ratios of 81.03 (sugar - 2H2O), 99.05 (sugar H2O), and 117.06 (sugar), where sugar refers to the attached deoxyribose unit. Hydrocarbon fragments from the alkane thiol linker appear at the lower m/z range of 10-80. In addition to the characteristic peaks from the different subunits of the DNA molecule (bases, sugar, and phosphate), several unexpected peaks appeared in the positiveion spectra of the Alpha DNA film. These peaks were subsequently identified as fragments of poly(dimethylsiloxane) (PDMS) and different glycerophospholipids. Molecular fragments of PDMS appear in the positive-ion spectrum at m/z ratios of 43.00, 73.05, and 147.07 (Figure 4a). Fragments from the lipid molecules are shown in Figure 4b at m/z ratios of 399.28, 478.25, 506.27, 534.31, and 701.50. Although it was not possible to uniquely identify the type of lipid molecules present, the isotopic (38) Georgiadis, R.; Peterlinz, K. P.; Peterson, A. W. J. Am. Chem. Soc. 2000, 122, 3166-3173. (39) Glen Research Corp. User Guide to DNA Modification and Labeling 1990. (40) Getz, E. B.; Xiao, M.; Chakrabarty, T.; Cooke, R.; Selvin, P. R. Anal. Biochem. 1999, 273, 73-80. (41) Shafer, D. E.; Inman, J. K.; Lees, A. Anal. Biochem. 2000, 282, 161-164.

Lee et al.

Figure 4. Positive-ion (a and b) and negative-ion (c) ToFSIMS spectra of SH-ssDNA oligomer from Alpha on Au. Other than the characteristic peaks from the different subunits of the DNA molecule, several unexpected peaks, including poly(dimethylsiloxane) (PDMS) and lipid molecules, were identified in the positive-ion spectra of the Alpha DNA film.

patterns in the ToF-SIMS spectra suggest the following molecular structures for the following peaks: C22H54N2O4P4 or C21H55N4OP5 (for m/z ) 534.31) and C34H77N2O4P4 or C33H78N4OP5 (for m/z ) 701.50). In the negative-ion spectra (Figure 4c), the phosphates from the DNA backbone (PO, PO2, PO3, H2PO4) were observed at m/z ratios of 46.97, 62.96, 78.96, and 96.97. Other key molecular fragments observed in the negative ion spectra include adenine (A - H, m/z ) 134.05) and cytosine (C - H, m/z ) 110.04) fragments, substrate ions (Au, AuH, AuH2) in the m/z range of 195-200, oxidized sulfur species (SO3, HSO3, SO4, and HSO4) in the m/z range of 79-81 and 95-97 and sulfur-gold species (AuSH, Au2S, Au3S, AuSO3, etc.) in the m/z range of 230-850 (not shown). In the mass range of m/z 0-500, ToF-SIMS spectra of DNA films are highly complex due to fragmentation of hydrocarbon chains, aldehyde groups, etc. found in all biomolecules. Therefore, to simplify data interpretation,

Evidence of Impurities in Thiolated ssDNA Oligomers

Figure 5. PC1 scores plot for the positive-ion ToF-SIMS spectra showing separation between SH-ssDNA oligomers from Alpha and TriLink on Au (a). PC1 loadings plot showing peaks responsible for the sample separation (b).

it was essential to evaluate the ToF-SIMS spectra using a multivariate analysis method such as PCA. The PCA results for the first principle component (PC1), which accounts for 72.5% of the variance in the data set, from the positive-ion spectra of DNA films prepared from Alpha and TriLink SH-ssDNA oligomers are presented in Figure 5. The scores plot in Figure 5a shows that there is clear separation between the TriLink DNA samples (black squares) and the Alpha DNA samples (white diamonds). The loadings plot in Figure 5b indicates that this separation is mainly due to the contamination peaks (bottom) previously observed in the positive-ion spectra of the Alpha DNA samples. These peaks include PDMS fragments at m/z of 43.00 and 73.05, combinations of hydrocarbon chain with the NO functional group in the m/z 45-100 range (which closely resembles fragments from the lipid molecule) and phospholipids at the m/z 399-701 range (not shown). The hydrocarbon, sugar, and base molecular ions (top) load with the TriLink DNA samples. PC1 captures 88.3% variance of the negative spectra and also indicates clear separation between the two sets of samples (Figure 6a). The TriLink DNA samples (black triangles) are clearly distinguished from the Alpha DNA samples (white circles). The loadings plot in Figure 6b indicates that this separation is mainly due to the sulfur and substrate molecular ion signals in the Alpha DNA samples (bottom) at m/z 31.97 (S), 32.98 (SH), 79.96 (SO3), 96.96 (HSO4), 196.97 (Au), and sulfur-Au clusters at the higher m/z range (220-850, not shown). The observation that the sulfur fragments and Au peaks are associated more with Alpha DNA samples is consistent with the XPS results reported in the previous section. The phosphate fragments (PO3 and H2PO4 at m/z 78.96 and 96.97,

Langmuir, Vol. 21, No. 11, 2005 5139

Figure 6. PC1 scores plot for the negative-ion ToF-SIMS spectra showing separation between SH-ssDNA oligomers from Alpha and TriLink on Au (a). PC1 loadings plot showing peaks responsible for the sample separation (b).

respectively) from the DNA backbone load with the TriLink DNA samples (top). To confirm that the excess sulfur present at the DNA surfaces was from DTT, molecular fragmentation patterns of DTT were obtained with ToF-SIMS and compared with the negative-ion spectra from the DNA films. From the negative-ion spectra of the pure DTT sample (Figure 7a), we find that DTT did not come off the surface as an intact molecular ion. This observation is consistent with data reported by Graham et al.,42 who observed that molecules attached through thiol rarely are ejected as the intact molecular fragment in a ToF-SIMS experiment. In addition, we find that many of the fragments observed in the DTT fragmentation pattern were found to correspond to the sulfur peaks observed in the DNA film from Alpha (Figure 4c). Figure 7b presents the normalized ion intensity of the sulfur-containing molecular fragments present on bare gold, pure DTT adsorbed to gold, and the two DNA sequences adsorbed to gold (Alpha and Trilink). We observe that Alpha DNA samples yield relatively higher molecular fragment signals for these contaminant peaks compared to bare Au and TriLink DNA samples. To demonstrate that this result is not an artifact of the normalization procedure used, the normalized intensity of a phosphate-containing peak from the DNA bases from the samples is given for comparison. We find that the Trilink samples yield a more-intense molecular fragment from H2PO2 (due to DNA bases) than does Alpha. This finding is supported by our XPS data which indicates higher phosphorus atomic percentage for Trilink samples, as well (see Table 1a). These observations confirm our hypothesis that the excess sulfur-containing contaminant is due to the presence of the thiol reductants. (42) Graham, D. J.; Ratner, B. D. Langmuir 2002, 18, 5861-5868.

5140

Langmuir, Vol. 21, No. 11, 2005

Lee et al.

Figure 7. Molecular fragmentation pattern obtained from the negative-ion ToF-SIMS spectrum of pure DTT on Au was used to verify the presence of DTT in DNA films assembled from Alpha SH-ssDNA (a). Comparison of some of these key molecular fragments in the negative-ion ToF-SIMS spectra from bare Au (white), pure DTT (stripes), Alpha DNA (gray), and TriLink DNA (black) indicates correlation between Alpha DNA and DTT samples. All ion intensities were normalized to the total intensity of the selected peaks (b). Table 3. Effects of Contamination on SH-ssDNA Surface Coverage, as Indicated by the XPS Compositional Differences Observed in the Resulting DNA Film Prepared from Relatively Pure SH-ssDNA (TriLink) vs Contaminated SH-ssDNA (Alpha) after 5 min, 30 min, 2 h, 5 h, and 24 h of Assemblya atomic percent C1s

P/N

C/N

O/N

5 min 30 min 2h 5h 24 h

1.5 1.5 1.2 1.2 1.1

Alpha: 5′-(dA)8(dC)8-(CH2)6SH-3′ 34.6 5.9 10.7 1.2 46.1 0.25 39.8 5.9 12.7 1.7 38.4 0.26 37.5 4.6 13.6 1.8 41.4 0.27 41.0 5.6 16.6 2.0 33.7 0.21 41.0 5.6 17.6 1.9 32.9 0.20

5.9 6.8 8.5 7.4 7.4

1.8 2.1 3.1 3.0 3.2

5 min 30 min 2h 5h 24 h

0.8 0.9 1.3 1.7 2.1

TriLink: 5′-SH(CH2)6-(dC)8(dA)8-3′ 43.8 5.3 13.6 0.0 36.6 0.15 42.4 6.0 15.7 0.0 35.0 0.15 44.8 6.7 15.9 0.0 31.2 0.19 42.6 7.6 16.9 0.0 31.2 0.22 43.9 8.5 17.8 0.0 27.8 0.24

8.4 7.3 6.8 5.6 5.2

2.6 2.7 2.4 2.2 2.1

a

N1s

O1s

S2p

atomic ratio

P2p

Au4f

All standard deviations < 2%.

Effects of Contamination on SH-ssDNA Surface Assembly. By monitoring the changes in the percentage of the elements (C, N, O, P, and S) expected from SH-ssDNA over adsorption time, we observed the effect of the impurities present in commercially synthesized SH-ssDNA oligomers on the assembly and structure of the resulting films. SH-ssDNA from two of the vendors (Alpha and TriLink) were self-assembled in solution onto gold surfaces for 5 min up to 24 h. XPS compositional data were compiled for the different assembly times (see Table 3). The inherent variation of signal from P and N makes absolute comparisons of these quantities difficult. However, trends in Table 3 suggest that, when relatively pure SH-ssDNA (TriLink) is used to prepare the DNA films, the P, N, O, and C signals were observed by XPS to increase

Figure 8. S2p high-resolution XPS spectra of films formed from assembly of a 1 µM SH-ssDNA [5′-(dA)8(dC)8-(CH2)6SH3′, Alpha] solution in 1 M NaCl-TE buffer over time. With increasing time the amount of bound sulfur (SB) decreases while unbound sulfur (SU) and oxidized sulfur (SO) increase. Time periods shown are at 2 h (a) and 24 h (b) assembly.

over a 24-h time period, consistent with an increase in the SH-ssDNA surface coverage with time. In contrast, surfaces prepared using SH-ssDNA containing higher levels of contaminants (Alpha) did not follow this trend (see Table 3). Although the C, O, and S signals are observed to increase with assembly time, the P and N signals do not. This indicates that, after the initial SH-ssDNA adsorption, the remaining material incorporated into these films is due to a non-DNA contaminant. To gain additional information on the excess sulfur species, high-resolution XPS S2p spectra were obtained for the SH-ssDNA samples containing higher levels of contaminants (Alpha). In addition to the peak at 161.9 eV characteristic of sulfur bound to the gold surface,43 two additional sulfur species were present in the DNA films, as well. Figure 8 shows the presence of both unbound sulfur in the 163.7-164.0 eV range and oxidized sulfur at 167.7 eV. Furthermore, we find that the amount of bound sulfur decreases over time, while unbound and oxidized sulfur species increase during the same time period. One possible explanation for this result is that more of the DTT and other sulfur-containing species adsorb during the extended absorption process, thereby increasing the amount of unbound sulfur relative to bound sulfur. Furthermore, prolonged exposure of sulfurcontaining species to solution may result in their oxidation over time. The presence of oxidized or nonspecifically bound thiols may also explain the variation observed in the performance of SAM-based DNA arrays. Conclusions In this work, XPS and ToF-SIMS were used to study the self-assembly process of SH-ssDNA oligomers onto (43) Castner, D. G.; Hinds, K.; Grainger, D. W. Langmuir 1996, 12, 5083-5086.

Evidence of Impurities in Thiolated ssDNA Oligomers

gold surfaces and the composition of the resulting films. We have shown that the diversity of techniques used in the synthesis, treatment, and purification of the SH-ssDNA oligomers has led to a significant variation in the purity of commercially available SH-ssDNA. Using XPS, we were able to obtain quantitative atomic compositions of the individual DNA films prepared from SH-ssDNA oligomers from three vendors. XPS results from DNA films prepared from two of the purchased SH-ssDNA oligomers showed an excess of sulfur and less surfacebound DNA in these films. ToF-SIMS results confirmed that the excess sulfur detected by XPS was due to the presence of DTT, a reductant often used to cleave disulfide precursors, and possibly protective groups such as DMT. Other contaminants detected by ToF-SIMS include PDMS and lipid molecules. Furthermore, we found that the SH-ssDNA self-assembly process is affected by the presence of these contaminants. Such contaminants will likely exist in other commercially produced SH-ssDNA. Thus,

Langmuir, Vol. 21, No. 11, 2005 5141

these results underscore the importance of using surfacespecific techniques to confirm the chemistry of surfaces modified with DNA (such as microarray films) prior to their use in commercial applications. Acknowledgment. This research was supported by NIBIB Grant No. EB-002027 to the National ESCA and Surface Analysis Center for Biomedical Problems (NESAC/BIO), NIH Grant No. EB-001473, and the Joint Institute for Nanoscience funded by the Pacific Northwest National Laboratory (operated by Battelle for the U.S. Department of Energy) and the University of Washington. Dan Graham, Collin May, Newton Samuel, Ping Gong, Greg Harbers, and David Grainger are thanked for their assistance with the principal component analysis, discussions, and experiments involved in this study. LA0472302