A Quartz Crystal Microbalance Study of Polycation-Supported Single

Nov 19, 2008 - A quartz crystal microbalance with dissipation monitoring (QCM-D) was used to investigate the properties and formation of a genomic ...
0 downloads 0 Views 2MB Size
Download PDF
3416

Biomacromolecules 2008, 9, 3416–3421

A Quartz Crystal Microbalance Study of Polycation-Supported Single and Double Stranded DNA Surfaces Amanda Y. Yang, Robert J. Rawle,† Cynthia R. D. Selassie, and Malkiat S. Johal* Department of Chemistry, Pomona College, 645 North College Avenue, Claremont, California 91711-6338 Received September 19, 2008; Revised Manuscript Received October 24, 2008

A quartz crystal microbalance with dissipation monitoring (QCM-D) was used to investigate the properties and formation of a genomic mammalian DNA surface on a polycationic poly(ethylenimine) (PEI) film. We show that both single- and double-stranded DNA films can be deposited on the PEI surface by modulating the DNA adsorption time. The two distinct DNA surfaces can be confirmed by their interactions with urea, a common DNA denaturant, and ethidium bromide, a common DNA intercalator, both of which lead to characteristic changes in the QCM-D frequency and dissipation. The hybridization process between surface-bound single-stranded DNA to complementary strands in solution can be resolved in real-time. Moreover, we have also investigated the effects of incorporating NaCl in the various PEI-DNA assemblies and have shown that higher ionic strengths lead to greater DNA adsorption to the PEI surface. An increase in the QCM-D resonant frequency and a decrease in dissipation occur when these assemblies are rinsed with salt-free water. We interpret these changes as a loss of counterions from the film and an increase in intrinsic ion-pair complexation, leading to a more rigid PEI-DNA assembly. Varying the salt content in the DNA film can be used to control the film thickness and morphology.

Introduction The importance of bioactive surfaces has recently been underscored by the emergence of new technologies for the rapid characterization of biologically relevant analytes. Surfaces composed of immobilized DNA, for instance, have lent themselves to applications in biosensing,1,2 electronics,3,4 microarray technology,5,6 nanomaterials,7 drug or gene delivery,8,9 and even gene therapy.10-12 Thus, understanding and controlling the composition, interfacial properties, degree of bioactivity, and the internal morphology of such assemblies is key to the rational design of new biologically active materials. Complexing DNA with polyelectrolytes of opposite charge offers interesting possibilities that exploit specific ion-pair interactions leading to control over factors such as counterion content, hydration, and surface morphology. Although there is a plethora of studies focusing on DNA adsorption to a variety of substrates, including nanoparticles13 and modified lipid bilayers,14 only a few focus on electrostatic self-assembly. Some studies have examined the effect of ion-exchange on surface-tethered DNA assemblies15 and the adsorption of DNA to polyelectrolyte microcapsules.16 However, the effect of local ionic strength on the film structure within DNA-polyelectrolyte assemblies is largely unexplored. In this work, we have used the quartz crystal microbalance with dissipation monitoring (QCM-D) to monitor the formation of polyelectrolyte-DNA assemblies in situ and in real-time. QCM-D is a sensitive and noninvasive technique that measures changes in resonant frequency and dissipation of an oscillating piezoelectric quartz crystal over time as solution-phase material adsorbs to the crystal surface. Frequency change (∆F) measurements are approximately negatively proportional to mass deposited;17 however, this relationship will under-report the mass of a loosely coupled film.18 Dissipation (∆D) is a measure of the ability of a film to dissipate the energy of the oscillating * To whom correspondence should be addressed. Fax: (909) 607-7726. E-mail: [email protected]. † Current address. Department of Chemistry, Stanford University, 333 Campus Drive, Mudd Building, Room 121, Stanford, CA 94305-5080.

crystal. A thick, loosely coupled film will produce a high dissipation value whereas a thin, rigid film will exhibit a low dissipation value. In general, frequency values will decrease and dissipation values will increase as material, including coupled water, is adsorbed to the crystal surface.19 Using the Sauerbrey equation,17 one can calculate the approximate masses of adsorbed materials from the change in frequency, assuming the film is rigidly fixed to the QCM crystal. The frequency and dissipation values from a number of different overtones can also be used to obtain the viscoelastic properties of the adsorbed material using the Voigt model.20,21 In this work the viscoelastic properties were not determined because the changes in resonant frequencies and dissipation alone were sufficient to describe the formation of the PEI-DNA assemblies. To date, most QCM-D studies of surface-bound DNA have concentrated on plasmid DNA or synthetic oligomers22-25 or on the sensitivity of DNA surfaces toward detecting ligands or forms of DNA damage.26 For instance, Johnston et al. have investigated the use of oligonucleotides in the formation of DNA multilayers and the stability of these films, while Zhu et al. used DNA oligomers to observe the hybridization and hydrolytic cleavage of an immobilized single strand of DNA. Nguyen et al. have reported on the kinetics of the adsorption of plasmid DNA onto an organic-matter coated silica surface as a function of monovalent and divalent salts.27,28 Other studies have focused on the effects of salt, which has been known to increase thickness and stability in polyelectrolyte thin films.29 Salt in DNA films also allows for competitive ion pairing or ion exchange that could change the conformation of the DNA assembly and may lead to methods of tuning the properties of these films.30 In this work, we use DNA deposition time as well as salt concentration to control the type and amount of DNA in hybrid polycationic-DNA assemblies. In contrast to most studies, we center on the use of genomic mammalian DNA, which may have greater utility in future studies relating to mammalian biological systems.31 Using QCM-D, we have been able to monitor the creation of these

10.1021/bm801060w CCC: $40.75  2008 American Chemical Society Published on Web 11/19/2008

Polycation-Supported DNA Surfaces

Biomacromolecules, Vol. 9, No. 12, 2008

3417

genomic DNA surfaces in real-time. Based on frequency and dissipation changes during the assembly process, we have observed a slow transition of the surface-immobilized singlestranded DNA (ssDNA) assembly to the corresponding doublestranded DNA (dsDNA) assembly. The two forms were verified by their interaction with ethidium bromide, a common DNA intercalator, and exposure to aqueous urea solutions, which is known to interact strongly with DNA through hydrogen bonds. We also investigate the formation of the assemblies prepared with varying concentrations of salt. We have found that increasing ionic concentration results in increased depositions of genomic DNA, though it appears that the DNA adsorbs in a different conformation than that of salt-free ssDNA or dsDNA.

Experimental Methods Poly(ethylenimine) (PEI, MW ) 25000 g mol-1, mixture of linear and branched chains, CAS 9002-98-6), sodium chloride, calf thymus DNA (water soluble, sodium salt), ethidium bromide (EtBr), and urea were purchased from Aldrich and used as received. Aqueous solutions of PEI (10 mM, based on monomeric molecular weight), DNA (100 µg/mL), EtBr (1 mM), urea (1 M), and NaCl (varying concentrations) were prepared using ultrapure water (resistivity >18.0 MΩ cm) or in NaCl buffer, as specified. Concentrations of NaCl solution used ranged from 5 to 100 mM. In experiments using NaCl solution, all solutions of PEI and DNA were also prepared in the appropriate NaCl buffer. ssDNA was prepared by heating and then rapidly cooling a solution of calf thymus DNA. The DNA was heated to 95 °C for 10-15 min and then immediately placed on ice for at least 30 min. The solution was then brought up to 23 °C. This is a standard protocol to produce heat-denatured ssDNA. It is possible that some of the heat-denatured DNA will partially rehybridize after it is taken off the ice, but rehybridization only occurs to any significant extent when the DNA is allowed to slowly cool down after heating. Furthermore, similar QCM-D results are obtained when using DNA that had been heat-denatured and then ice cooled the day before, left at room temperature for a few hours, and then stored at 4 °C overnight. It is therefore reasonable to assume that the rehybridization kinetics is significantly slower than the ssDNA adsorption to PEI. QCM-D frequency and dissipation data of the formation of DNAPEI surfaces were obtained using a quartz crystal microbalance with dissipation monitoring (E4, Q-Sense, Gothenburg, Sweden). The QCM-D sensor consists of a disk-shaped, AT-cut piezoelectric quartz crystal coated with metallic electrodes on both sides. The QCM-D sensor crystal (14 × 0.3 mm) operates at a frequency of 4.95 MHz ( 50 kHz. For all experiments, a quartz crystal coated with an Au electrode (100 nm thick) on the back and an active surface layer of SiO2 (∼50 nm thick) was used. Crystals were optically polished with a surface roughness of less than 3 nm (rms). The active side was in continuous contact with the aqueous solutions of PEI, DNA, ligand, water, or buffer. The crystal was mounted in a flow cell with a total volume of 40 µL. Prior to use, the SiO2-coated quartz crystals were decontaminated by UV/ozone treatment for 5-10 min, immersion in 2% Hellmanex solution (Hellma, Co) for 30 min, and UV/ozone treatment again for 5-10 min, with the final two steps followed by an ultrapure water rinse and drying with N2. Before each run, the QCM-D flow cells were flushed with 2% Hellmanex solution followed by ultrapure water until a stable baseline was obtained. During the experiments, QCM-D cells were rinsed with ultrapure water (or NaCl solution, as indicated) between all depositions, which were considered to be complete when frequency and dissipation measurements stopped changing with time. Solutions were passed through the flow cell at a rate of 100 µL/min. All experiments were conducted at 23 ( 0.1 °C.

Figure 1. (a) Fifth overtone frequency (black) and dissipation (gray) shifts of a SiO2-coated quartz crystal during the adsorption of PEI and ssDNA, which was allowed to hybridize with complementary strands; (b) Fifth overtone frequency (in black) and dissipation (in gray) shifts of a SiO2-coated quartz crystal during the adsorption of PEI and ssDNA, which was subsequently rinsed with ultrapure water to prevent hybridization. Each adsorption was followed by ultrapure water rinses as indicated by the shaded segments.

Results and Discussion Hybridization. The construction of a surface-bound hybrid PEI-DNA assembly can be achieved through a simple electrostatically driven layer-by-layer adsorption process. To create a genomic double-stranded DNA surface, an anionic SiO2-coated quartz crystal was first exposed to a solution of PEI and then to a solution of heat-denatured calf thymus DNA. Figure 1 shows the decrease in frequency and the concurrent increase in dissipation corresponding to each binding event. The initial drop in frequency (from ∼17 to ∼26 min) indicates the adsorption of PEI to the underlying substrate. The rise in dissipation during this event reflects the building of loosely bound or “floppy” polymer material at the surface. The shaded areas in Figure 1 represent water rinses. Exposure of the adsorbed PEI surface to the denatured, single-stranded DNA results in a very sharp negative frequency shift of ∼5 Hz at 40 min. This decrease is followed by a lag period of several minutes, and then a more gradual decrease in frequency over ∼1 h (Figure 1a). We can correlate these negative frequency shifts to mass depositions based on the Sauerbrey model.15 Dissipation values, which are a function of the rigidity of the film, increase very slightly during the entire deposition of DNA,

3418

Biomacromolecules, Vol. 9, No. 12, 2008

allowing us to accurately apply the Sauerbrey model. Thus, the first abrupt frequency shift observed by the single-stranded DNA corresponds to an adsorbed mass of ∼88 ng/cm2, while the more gradual shift corresponds to a net deposition of ∼311 ng/cm2. In cases for which dissipation measurements did not remain near zero, we used the Sauerbrey model qualitatively in order to correlate decreases in frequency with increases in mass, although a specific number estimation cannot be accurately determined. The first-order rate constants for the fast and slow steps during the DNA deposition have been previously reported as 5.52 × 10-2 and 2.41 × 10-3 s-1, respectively.31 The adsorption of the denatured DNA to the PEI surface occurs as a result of electrostatic interactions. The adsorption of PEI to the SiO2-coated crystal results in a positively charged surface due to charge overcompensation. This allows the negatively charged phosphate groups of the DNA backbone to bind electrostatically to the cationic PEI layer, leaving the base pairs exposed to solution. Using this model, we interpret the quick, initial frequency shift that occurs upon exposure of the PEI film to denatured DNA as an electrostatic interaction, resulting in a single-stranded DNA film. The slower frequency shift following the lag period is interpreted as the hybridization of the ssDNA to complementary strands (or complementary regions of strands) in solution. Similar observations have been made in previous studies using synthetic DNA oligomers.32 The greater frequency decrease during the hybridization step compared to the initial ssDNA binding step possibly indicates partial hybridization of the surface DNA to more than one partner strand and or may indicate some rearrangement of the surface DNA during hybridization which would allow more DNA to adsorb. If the two-step deposition of DNA on PEI is resolving the hybridization process, then it should be possible to inhibit double strand DNA formation by limiting exposure of the PEI surface to the bulk aqueous DNA solution. Therefore, to create the ssDNA surface in isolation, the PEI film was exposed to the DNA solution for approximately five minutes and then rinsed with ultrapure water to prevent hybridization. Figure 1b shows how the QCM-D frequency and dissipation values shift during formation of the PEI layer and the subsequent exposure of this film to the DNA solution. The single-step rapid drop in frequency of approximately 5 Hz at ∼39 min is consistent with the formation of a ssDNA film. Once again, this sharp decrease in frequency is attributed to electrostatic binding of ssDNA to PEI and is similar to the decrease seen in Figure 1a, but it is not followed by the gradual frequency decrease seen in Figure 1a that is attributed to the formation of a dsDNA surface. Interaction of ssDNA and dsDNA with Urea. To verify the formation of the single- and double-stranded DNA surfaces, the individual films were exposed to solutions of urea. Urea is a known denaturant of DNA because of its strong affinity for hydrogen bonding to DNA base pairs and is therefore expected to interfere with correct DNA hybridization and even displace hybridized strands from the surface. As before, ssDNA and dsDNA surfaces were prepared on PEI, and each was subsequently exposed to a 1 M aqueous solution of urea, followed by a rinsing step to establish the net interaction between urea and the underlying DNA surface. For ssDNA surfaces, exposure to urea resulted in a net decrease in frequency of approximately 4 Hz, corresponding to a net mass deposition of ∼70 ng/cm2 (Figure 2a). An increase in mass suggests that urea binds to the available hydrogen bonding sites of the unhybridized DNA base pairs. The net change in dissipation for this process is close to zero, indicating that the urea molecules are not changing the

Yang et al.

Figure 2. (a) Fifth overtone frequency (black) and dissipation (gray) shifts of a SiO2-coated quartz crystal during the adsorption of PEI, ssDNA prevented from hybridizing, and exposure to urea followed by ultrapure water rinses; (b) Fifth overtone frequency (black) and dissipation (gray) shifts of a SiO2-coated quartz crystal during the adsorption of PEI, ssDNA allowed to hybridize to completion, and exposure to urea, followed by an ultrapure water rinse as indicated by the shaded segments.

conformation of the surface-bound ssDNA to any large degree. Control experiments in which urea solutions were passed over an inert Au coated QCM-D crystal confirmed that the large, intermediate shifts in frequency and dissipation during exposure of the DNA surface to urea were due to the higher viscosity of the urea solution. In this control, the viscosity effect disappears after a water rinse and the frequency and dissipation return to their baseline values. Hence, in our examination of the interaction between DNA-polycationic assemblies and urea, we are only concerned with the net changes in frequency and dissipation. Figure 2b shows the changes in QCM-D frequency and dissipation values when a surface containing dsDNA is exposed to urea. As expected, no net decrease in frequency is observed since no or very few free hydrogen bonding sites are available in the dsDNA assembly. In fact, exposure to urea causes a slight increase in frequency, indicating a net loss of mass from the PEI-dsDNA assembly. There are two possible explanations for this increase in frequency. First, urea may cause a small osmotically driven deswelling of the PEI-dsDNA assembly. This is unlikely since the loss of water from the film should cause a significant decrease in the dissipation and this was not observed.

Polycation-Supported DNA Surfaces

More likely, urea displaces some of the hybridized complementary strands, competing for hydrogen-bonding sites. As a result, the loss of some complementary strands and the subsequent addition of the smaller urea molecules results in a net increase in the resonant frequency. The net change in the frequency after the exposure of a dsDNA-PEI assembly to urea corresponds to a final mass loss of ∼40 ng/cm2. If all hybridized strands were stripped off by urea exposure, the final frequency values of the ssDNA and dsDNA surfaces after being exposed to urea should be the same. Because the values are not the same, we presume that portions of the surface-bound dsDNA remain hybridized in some areas and that these areas remain unaffected during urea exposure. Nevertheless, these results demonstrate that the different DNA surfaces respond very differently to strong hydrogen bonding ligands such as urea, and that significant binding of urea is observed in the ssDNA assemblies. Interaction of ssDNA and dsDNA with Ethidium Bromide. In a similar manner to the aforementioned urea experiments, ethidium bromide can be used to distinguish between ssDNA and dsDNA surfaces. Ethidium bromide is a known intercalator which will stack between the base pairs of hybridized strands of dsDNA.33 Both ssDNA and dsDNA surfaces, formed in separate liquid flow cells, were exposed to ethidium bromide solution and subsequently rinsed with ultrapure water (Figure 3). A net negative frequency shift of ∼5.5 Hz, indicating a net deposition of ∼100 ng/cm2, was observed only for the dsDNA surface (Figure 3b), while the ssDNA exhibited no net adsorption (Figure 3a). A net adsorption to the dsDNA surface indicates that ethidium bromide has intercalated as expected. For the ssDNA film, on the other hand, no net adsorption was observed. This observation is as expected since the ability of ethidium bromide to intercalate into ssDNA is very poor. The lack of binding of ethidium bromide to ssDNA also supports the notion that the bulk phase is composed of mainly ssDNA strands following heat denaturation. The dissipation values remain constant during the exposure of ethidium bromide to the ssDNA assembly. However, in the case of dsDNA, the dissipation increases slightly when the film is exposed to ethidium bromide. Because dissipation is a measure of the viscoelastic properties of the film, when ethidium bromide intercalates between the complementary strands of DNA, the assembly is expected to increase in dissipative loss. Addition of NaCl. The formation of ssDNA and dsDNA films was further characterized as a function of salt concentration. Polycation-DNA assemblies were prepared from solutions containing known amounts of NaCl. Figure 4 shows how the frequency and dissipation values change during the assembly of PEI and DNA from solutions containing (a) 5, (b) 10, (c) 50, and (d) 100 mM NaCl. The shaded areas represent the corresponding salt buffer rinses, analogous to the water rinses described in the previously experiments. In each case, a decrease in frequency is observed when the surface is exposed to PEI and DNA solutions prepared in the appropriate NaCl solution. For the DNA depositions on PEI, a ∼25 Hz shift is observed for the 5 mM NaCl assembly; ∼27 Hz for 10 mM assembly, ∼38 Hz for the 50 mM assembly, and ∼ 54 Hz for the 100 mM assembly. In these cases, we are unable to calculate accurate values of adsorbed mass due to the large increases in dissipation. However, the strong correlation between salt concentration and negative frequency shifts does imply greater mass deposition, or more DNA adsorption to the surface, when the ionic strength of the polyelectrolyte solutions is increased. The dissipation values also increase with increasing salt concentration, indicating an increase in the thickness and “floppiness” of the assemblies.

Biomacromolecules, Vol. 9, No. 12, 2008

3419

Figure 3. (a) Fifth overtone frequency (black) and dissipation (gray) shifts of a SiO2-coated quartz crystal during the adsorption of PEI, ssDNA prevented from hybridizing, and ethidium bromide (EtBr); (b) Fifth overtone frequency (black) and dissipation (gray) shifts of a SiO2coated quartz crystal during the adsorption of PEI, ssDNA allowed to hybridize to completion, and EtBr. Shaded segments correspond to ultrapure water rinses.

Interestingly, the characteristic two-step pattern in frequency attributed to hybridization of DNA strands in water is diminished, or less resolved, as the sodium chloride concentration increases, suggesting that the second strand of DNA does not anneal to the first in the same way as it would under salt-free conditions. It is likely that the DNA assumes a more coiled conformation in the higher salt concentrations and this new conformation changes the dynamics of hybridization compared to DNA assemblies prepared in relatively salt-free solutions. Counterions are known to shield charged monomeric units from each other, leading to polyelectrolytes with coiled conformations.34 In the case of DNA, where the negatively charged phosphate groups are expected to have proportionally more positive sodium counterions associated with them in the solutions of increasing NaCl concentrations, electrostatic repulsion between the neighboring phosphate groups is expected to be minimized, causing the DNA to assume a more coiled conformation. This would be expected to inhibit the DNA hybridization process and would allow greater amounts of DNA to adsorb to the cationic polymer (PEI) layer. Thus, we attribute the loss of the two-step adsorption pattern and the greater

3420

Biomacromolecules, Vol. 9, No. 12, 2008

Yang et al.

Figure 4. (a) Fifth overtone frequency (black) and dissipation (gray) shifts of SiO2-coated quartz crystals during the formation of PEI DNA films at (a) 5 mM NaCl, (b) 10 mM NaCl, (c) 50 mM NaCl, and (d) 100 mM NaCl. Shaded areas mark NaCl buffer rinses of corresponding concentration.

depositions of DNA prepared in higher salt concentrations to the adsorption of DNA in highly coiled states. This model is also consistent with the increase in dissipation observed as the salt concentration is increased. In sum, this QCM-D data would suggest that varying the salt content in PEI-DNA assemblies allows one to change the thickness, mass, and morphology of the film. To observe the diffusion effects of the salt ions in the film, the salt laden PEI-DNA assemblies were subsequently exposed to an ultrapure water rinse (Figure 4). Exposure to salt-free water causes an increase in frequency and a decrease in dissipation, and this effect becomes more apparent as the salt concentration is increased. The observed increase in frequency, which corresponds to a decrease in mass, indicates that a net amount of material is leaving the film. Dissipation decreases during the ultrapure water rinse, indicating that the film is becoming more rigid and less floppy, a result that may be due to water being removed from the film. Only the terminal layer contains a significant number of small counterions, a region known as the overcompensated zone.35 The underlying film is largely homogeneous, salt free, and held together by strong ion pairs between the PEI and DNA strands.34 Therefore, we expect ions to exchange only between the bulk phase and the overcompensated zone. In fact, we have studied the effect of salt-free water on a PEI layer (prepared containing 10 mM and 100 mM NaCl) on SiO2 and observed a very small and constant (