Article pubs.acs.org/Langmuir
Influence of Attachment Strategy on the Thermal Stability of Hybridized DNA on Gold Surfaces Tyler J. Petty, Caleb E. Wagner, and Aric Opdahl* Department of Chemistry and Biochemistry, University of WisconsinLa Crosse, La Crosse, Wisconsin 54601, United States S Supporting Information *
ABSTRACT: The thermal stabilities of double-stranded DNA hybrids immobilized on gold surfaces are shown to be significantly affected by the conformation of the hybrid. To analyze this behavior, DNA probes were immobilized using attachment strategies where the nucleotides within the strand had varying levels of interactions with the gold substrate. The abilities of these probes to form double-stranded hybrids with solution DNA targets were evaluated by surface plasmon resonance (SPR) over a temperature range 25−60 °C. The measurements were used to construct thermal stability profiles for hybrids in each conformation. We observe that DNA hybrids formed with probe strands that interact extensively with the gold surface have stability profiles that are shifted lower by 5−10 °C compared to hybrids formed with end-tethered probes that have fewer interactions with the surface. The results provide an understanding of the experimental conditions in which these weaker DNA hybrids can form and show the additional complexity of evaluating denaturation profiles generated from DNA on surfaces.
■
electrochemical analysis.19,20 Results from those methods have provided evidence that DNA hybrids on surfaces exhibit broadened denaturation profiles compared to DNA in solution and that the denaturation temperatures tend to be lower in a confined surface environment. Experimental denaturation profiles have been fit to theoretical models ranging from basic adaptations of solution models based on variation in free energy with temperature to more sophisticated models that include terms to account for electrostatic destabilization and local variations in the density of DNA probes on the surface.19,21,22 Although these techniques have demonstrated that thermal stability profiles of hybridized DNA on surfaces can be generated, no single practice has been widely adopted. For SPR in particular, there are practical aspects that make quantitative measurements while temperature is continuously incremented challenging. Here we show how a set of SPR measurements of DNA hybridization activity, each performed at different temperatures, can be assembled into a thermal stability profile. The primary variable that we report on is how the degree of interaction between the DNA probes and the gold surface affects the thermal stability of resulting hybrids. We have recently demonstrated by systematic salt stringency rinses that DNA hybrids that have extensive interactions with the underlying gold surface, termed directly-adsorbed, are less stable
INTRODUCTION Synthetic DNA is widely used in surface-based diagnostic tools and emerging technologies.1,2 Extensive experimental measurements have been made and predictive models subsequently developed, which have provided understanding of the immobilization and hybridization activity of DNA probes on many types of surfaces.3−8 The thermal stability of the resulting probe:target DNA hybrids formed on these surfaces, however, has been reported to a much lesser extent. Understanding this aspect of DNA behavior on surfaces is important for the continued development of DNA surface technologies, much as it has been for DNA in aqueous solution, where predictive models of stability are central to the design of structural, addressable, and active DNA materials.9,10 The stability of hybridized DNA in solutions is often assessed from thermal denaturation profiles, obtained, for example, by monitoring changes in the UV absorbance spectrum of the solution as its temperature is continuously incremented.11 Analogous measurements for DNA hybrids on surfaces are hindered by the challenge of adapting surface-based analytical methods to experiments involving continuous changes in temperature. A second challenge is designing approaches to systematically vary properties of the DNA layer on the surface so that the influences of specific variables on stability can be isolated. Several approaches for generating thermal denaturation measurements from DNA immobilized on gold and glass substrates have recently been reported, using techniques that include surface plasmon resonance (SPR),12−15 secondharmonic generation (SHG),16 SERS,17 fluorescence,18 and © XXXX American Chemical Society
Received: October 18, 2014 Revised: December 1, 2014
A
dx.doi.org/10.1021/la504128h | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
with the gold side of the SPR substrate. In SPR measurements of hybridization, solutions containing 4 μM DNA in NaCl-TE were introduced for 20 min followed by a rinse with DNA-free buffer. The net hybridization activity was recorded as the number of hybrids remaining after the rinse stage. Quantitative analysis followed established methods for fixed angle SPR where, for small changes in the refractive index (n), changes in reflected light intensity (ΔI) are proportional to Δn.26,27 Sensitivity factors (S) that relate Δn to ΔI were determined experimentally for each region of the sensor immediately before each experiment by measuring how the reflected light intensity changes in response to NaCl calibration solutions that had 0.10 M difference in concentration. The refractive indices of these calibration solutions were measured using an Abbe refractometer. Changes in DNA surface density were then calculated using eq 1, where ld is the decay length of the evanescent wave; na and ns are the refractive indices of the DNA and the buffer solutions at the analysis wavelength, respectively; ρDNA is the bulk density of the DNA; and MWDNA is the molecular weight of the DNA. Following Jung et al., ld was determined to be approximately 350 nm.26 We assumed values of DNA density (1.7 g/cm3) and refractive index (1.7) that are commonly used for SPR quantification.
at room temperature than end-tethered hybrids that project away from the substrate, as shown in Scheme 1.23 Those experiments Scheme 1. Types of DNA Hybrids Investigateda
a
Directly-adsorbed hybrids are prepared by hybridizing with DNA probes that lack a specific surface attachment functionality. Endtethered DNA hybrids are prepared by hybridizing target DNA (gray) to probes (black) that are attached using either thiol (S) functionality or a long series of adenine nucleotides (red).
used a model set of DNA strands that included homonucleotide sequences (blocks), which provide a practical approach for controlling the structure of DNA hybrids immobilized on gold, based on the high affinity of adenine nucleotides for gold.24,25 For comparative purposes, we used the same types of probes here as in those experiments and monitored their hybridization activities up to 60 °C.
■
DNA (cm−2) =
■
ρ NA ld ΔI × × DNA 2 MWDNA S(na − ns)
(1)
RESULTS AND DISCUSSION Temperature Dependence of SPR Measurements. Our goal was to use SPR to obtain thermal stability profiles from DNA hybrids that have differing degrees of interactions with the underlying gold substrate, as shown in Scheme 1. One way to obtain thermal stability profiles by SPR is from measurements of a sensor that is coated with hybridized DNA probes as its surface temperature is slowly and continuously incremented, monitoring the loss of DNA targets due to denaturation at each temperature. The feasibility of that type of measurement has been demonstrated by Livache et al.13,15 Quantification of data obtained in that manner can be challenging, however, because SPR reflectivity profiles are strongly temperature dependent.26 Specific to fixed-angle SPR, quantitative analysis of DNA hybridization typically relies on a linear relationship between observed changes in reflectivity and changes in the number of hybrids on the sensor.26 Increasing the temperature of the aqueous buffer solution in the flow cell will lower its refractive index, leading to a shift in the SPR resonance angle, and thus the reflectivity observed at a fixed analysis angle. This shift in resonance angle attributed to the bulk solution is relativity large compared to the shift that occurs during a DNA hybridization/ denaturation measurement. On our instrument, the shift is significantly large that assuming a linear relationship between changes in reflectivity and number of hybrids is not a valid approach if the temperature of the solution is incremented by more than 20 °C (Figure SI-1). We circumvented that issue here by performing a set of individual measurements of DNA probe hybridization activity, each at a specific temperature between 25 and 60 °C. At each temperature, the SPR imaging angle was set to an appropriate angle to ensure a linear relationship between changes in reflectivity and hybrids, and the sensitivity factors (S) for each region of the sensor were remeasured by measuring change in SPR reflectivity in response to two solutions of known refractive index. This strategy, while more time-consuming than a single temperature ramp experiment, has the advantage of reducing many of the temperaturedependent variables of SPR as potential sources of error. DNA Probe Stability at Elevated Temperature. The methods used to immobilize the DNA probes in this study have
EXPERIMENTAL DETAILS
Materials. Commercial purified oligonucleotides included the following sequences: 15-nucleotide homo-oligonucleotides of adenine (A15) and thymine (T15); a mixed sequence of 15 nucleotides (5′CAATGCAGATACACT-3′, denoted P15) and full complement for P15 (5′-AGTGTATCTGCATTG-3′, denoted P15′). These sequences were incorporated into longer strands, A15-T20 and A15-T5-P15, where the A15 block of nucleotides has been shown to serve as an effective surface attachment group that passivates the surface against further nonspecific adsorption of DNA in subsequent hybridization experiments involving the T20 and P15 blocks of nucleotides.24,25 Oligonucleotides with 3′ thiol modifications (HS-) included HS-T5P15 and HS-T15. Solutions denoted NaCl-TE and CaCl2-TE contained between 0.10 and 1.0 M NaCl or CaCl2, respectively, 1×TE (10 mM Tris-HCl, 1 mM EDTA), and were adjusted to pH 7. Sample Preparation. Gold SPR sensors (Platypus Technologies, Madison, WI) were cleaned by UV ozone treatment followed by rinsing with ethanol and high purity (18.2 MΩ) water. Oligonucleotides without thiol modifications were immobilized from solution onto clean SPR sensors for ca. 20 h at 35 °C (4 μM DNA in 1 M CaCl2TE). After probe immobilization, each sensor was rinsed sequentially with deionized water, 1 M NaCl-TE, 0.1 M NaOH, and deionized water, before drying under flowing nitrogen. This process removes calcium ions and weakly bound DNA. Thiol-modified oligonucleotides were immobilized from solution onto sensors for ca. 2 h at ambient temperature (4 μM DNA in 1 M NaCl-TE). These preparation methods result in surface densities of approximately 1 × 1013 cm−2 for each type of probe.25 SPR Measurements. The thermal stability and temperaturedependent hybridization properties of DNA probes were measured in situ using a previously described SPR imaging system (GWC, Madison, WI).23,25 A custom temperature-regulated two-channel (sample and reference) microfluidics flow cell was used that provided uniform flow of solutions at flow rates of approximately 50 μL/min using a peristaltic pump. Solutions were degassed to minimize formation of air bubbles in the flow cell at elevated temperature. The temperature of the flow cell was regulated over the range 25−70 °C (±0.02 °C) using a thermistor, thermoelectric module, and temperature controller. The actual temperature at the SPR sensor surface was calibrated in a control experiment using a temperature sensor placed in direct contact B
dx.doi.org/10.1021/la504128h | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
Figure 1. Stability of DNA probes at elevated temperature. Plots show the cumulative number of each type of probe lost from sensor surfaces after heating to 70 °C in 0.25 M NaCl-TE for approximately 8 min increments: (a) HS-T5-P15, (b) HS-T15, (c) A15-T5-P15, and (d) A15-T20. Error bars represent the standard deviation of at least three separate measurements from different sensors.
been previously reported.25 These methods result in surface densities of approximately 1 × 1013 cm−2 for each type of probe, and in conformations broadly defined as either directlyadsorbed or end-tethered (Scheme 1), with directly-adsorbed probes having more extensive interaction with the gold substrate as compared to end-tethered probes. The A15 homo-oligonucleotides have high affinity for gold and adopt directly-adsorbed conformations where the probes are more or less flat on the substrate. The P15 oligonucleotides likewise adopt directly-adsorbed conformations. The remaining probes in this study are end-tethered. The high affinity of adenine nucleotides for gold surfaces leads to A15-T5-P15 and A15-T20 immobilizing in the L-shaped conformation shown in Scheme 1, where the A15 block of nucleotides serves as a surface attachment group, allowing the P15 and T20 blocks, respectively, to act as end-tethered probes. In high salt solutions (1 M NaCl) the T20 probe block of A15-T20 has been observed to weakly interact with the A15 attachment block in a manner that does not adversely affect its activity toward solution DNA targets.23 The thiolated HS-T15 and HS-
T5-P15 probes immobilize in predominantly end-tethered conformations. In order to quantify DNA hybridization activity at elevated temperature, it is important that the probe coverage remains stable during the hybridization experiment. We performed a simple in situ SPR measurement to test the extent to which each of these probes desorbs from sensors at the elevated temperatures used in the hybridization experiments. In these measurements, the SPR sensor was rapidly heated from 25 to 70 °C (e.g., to 10 °C higher than the highest temperature used in subsequent hybridization experiments) and held at that temperature for several minutes before returning to 25 °C. Sensors were repeatedly heated in this fashion up to a total heating time of approximately 1 h. The cumulative number of probes lost was obtained by comparing the SPR signals at 25 °C, before and after the heating period (Figure SI-2). Figure 1 summarizes the results of these experiments. The cumulative number of thiol tethered probes that desorb (Figure 1a,b) steadily increases during heating. The rates of desorption at 70 °C are approximately 6 × 1010 DNA cm−2 min−1 for HS-T5-P15 and 1 × 1011 DNA cm−2 for HS-T20, C
dx.doi.org/10.1021/la504128h | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
Figure 2. Formation of end-tethered (a) A15-T5-P15:P15′ and (b) A15-T20:A15 and directly-adsorbed (c) P15:P15′ and (d) A15:T15 hybrids immobilized on gold at different temperatures. Diagrams depict idealized representations of each probe:target pair. In each hybridization experiment the sensor was exposed for ca. 20 min to 4 μM solution of a target in 0.25 M NaCl-TE, followed by a rinse for ca. 5 min in DNA-free buffer. The net hybridization activity is reported as the number of hybrids remaining after the posthybridization rinse.
DNA/cm2, as coverages lower than this typically lead to lower hybridization activities that are limited by the total number of probes on the surface.3,25,28 Importantly, each of the subsequent hybridization experiments was performed at lower temperatures than 70 °C where DNA probe desorption occurs at a much slower rate. This was ensured by monitoring a reference channel which was coated with nonhybridizing probes. For nonspecifically bound probes, such as A15-T5P15, we find that pretreating samples by heating to 70 °C for 30 min is sufficient to provide a stable baseline in subsequent SPR measurements of hybridization activity. Hybridization Behavior of Probes. Representative hybridization data for each type of probe in 25 °C, 0.25 M NaCl-TE, are shown in Figure 2, along with data obtained at elevated temperatures. In these experiments, probe coated sensors were exposed to a hybridization solution containing target DNA strands for 20 min followed by a 5 min rinse in DNA-free buffer solution to remove loosely bound DNA. Hybridization activities were quantified as the number of hybrids remaining after the posthybridization rinse in DNA-free buffer. Nonspecific adsorption of DNA targets during hybrid-
with roughly 30%−50% of the probe layer desorbed after 1 h. Quantitatively, that observation is consistent with electrochemical measurement of thiol DNA probe loss at elevated temperature reported by Levicky et al.,19 where probe loss of 40% was reported after heating in 65 °C buffer for 1.8 h. The A15-T5-P15 and A15-T20 probes, which are attached nonspecifically by multiple nucleotide gold interactions, experience a large loss of approximately 20% of probes after the first heating cycle (Figure 1c,d). After this initial loss, however, the probe coverage is stable for the remainder of the heating period, suggestive of a scenario where directly-adsorbed probes bind to the gold in a variety of multipoint attachment geometries and that only the weakest bound probes are removed by the initial heat treatment. Although the number of probes that desorb at 70 °C is significant for each type of probe, the 1 h heat treatment does not adversely affect their respective hybridization activities at room temperature; in some cases we actually observe it to be slightly higher compared to similar probe surfaces that did not undergo heating. This observation suggests that probe coverages after heating for 1 h are still above ca. 5 × 1012 D
dx.doi.org/10.1021/la504128h | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
Table 1. Summary of Hybridization Activities (25 °C) and Denaturation Temperatures for Each Type of Hybrida hybridization activity at 25 °C (×1012 DNA/cm2) probe:target
0.10 M NaCl-TE
A15:T15 P15:P15′ A15-T20:A15 A15-T5-P15:P15′ HS-T15:A15 HS-T5-P15:P15′ solution A15:T15 solution P15:P15′
