Dual polarisation interferometry characterisation of DNA immobilisation and hybridisation detection on a silanised support

Biosensors and Bioelectronics 21 (2006) 1459–1467

Dual polarisation interferometry characterisation of DNA immobilisation and hybridisation detection on a silanised support B. Lillis a , M. Manning a , H. Berney b , E. Hurley a , A. Mathewson a , M.M. Sheehan a,∗ a

Tyndall Institute, Lee Maltings, Prospect Row, Cork, Ireland b INSAT, University of Newcastle, UK

Received 30 March 2005; received in revised form 16 June 2005; accepted 17 June 2005 Available online 19 August 2005

Abstract Dual polarisation interferometry is an analytical technique that allows the simultaneous determination of thickness, density and mass of a biological layer on a sensing waveguide surface in real time. We evaluated, for the first time, the ability of this technique to characterise the covalent immobilisation of single stranded probe DNA and the selective detection of target DNA hybridisation on a silanised support. Two immobilisation strategies have been evaluated: direct attachment of the probe molecule and a more complex chemistry employing a 1,2 homobifunctional crosslinker molecule. With this technique we demonstrate it was possible to determine probe orientation and measure probe coverage at different stages of the immobilisation process in real time and in a single experiment. In addition, by measuring simultaneously changes in thickness and density of the probe layer upon hybridisation of target DNA, it was possible to directly elucidate the impact that probe mobility had on hybridisation efficiency. Direct covalent attachment of an amine modified 19 mer resulted in a thickness change of 0.68 nm that was consistent with multipoint attachment of the probe molecule to the surface. Blocking with BSA formed a dense layer of protein molecules that absorbed between the probe molecules on the surface. The observed hybridisation efficiency to target DNA was ∼35%. No further significant reorientation of the probe molecule occurred upon hybridisation. The initial thickness of the probe layer upon attachment to the crosslinker molecule was 0.5 nm. Significant reorientation of the probe molecule surface normal occurred upon hybridisation to target DNA. This indicated that the probe molecule had greater mobility to hybridise to target DNA. The observed hybridisation efficiency for target DNA was ∼85%. The results show that a probe molecule attached to the surface via a crosslinker group is better able to hybridise to target DNA due to its greater mobility. © 2005 Published by Elsevier B.V. Keywords: Dual polarisation; Immobilisation; Hybridisation

1. Introduction DNA microarrays and DNA biosensors have emerged as a key enabling technology for the functional interrogation of genetic information (Wang, 2000). These devices have shown tremendous potential for the rapid analysis of nucleic acid samples (Drummond et al., 2003) for measurements of differential gene expression (Duggan et al., 1999) and the discovery and scoring of single nucleotide polymorphisms (Brown and Bostein, 1999). Such sensors are based upon the ∗

Corresponding author. Tel.: +353 21 4904083; fax: +353 21 4270271. E-mail address: [email protected] (M.M. Sheehan). URL: http://www.tyndall.ie. 0956-5663/$ – see front matter © 2005 Published by Elsevier B.V. doi:10.1016/j.bios.2005.06.009

inherent ability of a target DNA in solution to bind to a complementary probe molecule immobilised on a solid support. A key step that is pivotal to the process is the attachment of the single stranded probe molecule to the substrate. The covalent attachment of probe molecules to self assembled silane layers is a commonly employed method with such sensors for the formation of highly reactive, robust probe layers on a variety of substrates such as glass and silicon (Chrisey et al., 1996; Lenigk et al., 2000; Almadidy et al., 2002). Elucidating information on probe orientation, probe mobility and surface coverage and assessing the impact they have on hybridisation efficiency are of critical importance in the development of an immobilisation method (Watterson et al., 2002). However, despite the wealth of published material on protocols for

1460

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467

attachment of probe molecules to such surfaces, there have been to date few systematic studies of the surface structure of the bound molecule and the impact it has on the hybridisation reaction. This is in part due to the limited number of high-resolution analytical techniques available to gain accurate quantitative and qualitative data. Labelling of the probe or target molecule with a fluorescent (Beier and Hoheisel, 1999) or radioactive label (Biljana et al., 2001) can provide information on probe coverage and hybridisation efficiency, respectively. However, such labels are subject to photochemical degradation and photobleaching and can also affect the specificity of the binding event. Atomic force microscopy has been used to garner information on probe orientation and the distribution of probe molecules attached to a solid support (Shlyakhtenko et al., 1999). However, data interpretation with this technique of short probe sequences attached to solid supports is both difficult and ambiguous. Ellipsometry has been used to elucidate the thickness and hence orientation of a probe layer attached to silanised surfaces and changes in the latter upon hybridisation (Gray et al., 1997; Elhadj et al., 2004). Ellipsometry measures the ellipsometric angles psi and delta, which are at the simplest level related to a film’s thickness and complex refractive index. Results are obtained by fitting a model to the data obtained. However, for measurement of very thin films it can become difficult to determine thickness and RI independently of each other and often one is fixed and the other allowed to vary in the model (Sastry, 2000). Surface plasmon resonance is a label independent method that can provide in situ and in real time quantitative information regarding probe coverage and the kinetics of probe immobilisation and hybridisation (Georgiadis et al., 2000; Peterson et al., 2001). However, surface plasmon resonance employs a single opto-geometric parameter. This provides data on the average density and thickness of the layer on the surface, but the two physical values cannot be separated from each other. In addition, the technique is limited to investigating immobilisation strategies compatible with gold surfaces. Dual polarisation interferometry is a novel analytical technique that employs a second opto-geometric parameter to elucidate both the size and density of a biomolecule adhered to a sensing waveguide surface, in real time, and with high resolution. DPI has shown good agreement with X-ray crystallographic data when monitoring the immobilisation of streptavidin on the sensor surface and the subsequent changes in conformation upon binding with d-biotin (Swann et al., 2004). Similarly, studies on the interaction between Apolipoprotein E isoforms and tissue plasminogen activator detected conformational changes between the various isoforms (Biehle et al., 2004). Recently, the technique has been employed to characterise the immobilisation of single stranded probe DNA by simple direct adsorption on a silanised surface and via an avidin biotin linkage (Berney and Oliver, 2005). The goal of this work was to evaluate, for the first time, the ability of the technique to characterise the more com-

plex covalent attachment of probe DNA to a silanised surface and the subsequent selective detection of target DNA. We compared two different immobilisation methodologies: (1) direct covalent attachment via formation of a Schiff’s base and (2) linker mediated attachment via a 1,2 homobifunctional crosslinker molecule. We show that it was possible to elucidate probe orientation and measure probe coverage at different stages of each immobilisation process. Also, by comparing the changes in thickness and density of the probe layer upon hybridisation of target DNA for both immobilisation methods, it was possible to elucidate the direct impact that probe mobility had on hybridisation efficiency. In addition, it was possible to perform the complete analysis in real time and in a single experiment.

2. Experimental 2.1. Principle of operation All measurements were performed with the Analight® Bio200 Dual Polarisation Interferometer (Farfield Sensors Ltd., Manchester, UK). The instrument consists of a helium neon laser (which emits light at 632.8 nm), a means to select plane polarised light (transverse magnetic (TM ) and transverse electric (TE )), a sensor constructed using two optical waveguides stacked one on top of each other and an array photodiode. The waveguides consist of silicon dioxide lightly doped with silicon nitride, which confine light between boundaries. When polarised light (TM ) is introduced to the end of the stack, single mode excitations in sensing and reference waveguides are formed and propagate through the structure. At the output, the two modes are allowed to diffract into the far field, where they form the well-known pattern of Young’s interference fringes on an array photodiode. The position of these fringes represent the relative phase positions of both sensing and reference waveguides. On exposing the surface of the sensing waveguide to a biomolecule that adheres to it, the speed of light within it changes. This causes a change in the effective index of the sensing waveguide and hence phase changes in the sensing light output. The effective index of the reference waveguide does not change. Direct measurement of the phase changes is possible by continuously monitoring the relative phase position of the fringe pattern by performing a Fourier transformation that relates intensity to position. A second polarisation of light (TE ) is introduced at right angles into the waveguide stack. This responds differently to biomolecule adsorption or removal, and provides a second independent measurement. Using classical optical theory, it is possible to interpret the two measurements in terms of thickness and refractive index of the adsorbed bimolecular layer. For each polarisation there are an infinite number of refractive index/film thickness combinations that will produce the observed effect. However, when both polarisations are taken together a unique solution is resolved. This allows

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467

1461

Fig. 1. A schematic of the dual polarisation interferometer.

for separate measurements of both thickness and refractive index (density) of the adsorbed biomolecular layer. The mass absorbed on the surface can then be calculated. This is unlike surface plasmon resonance where only a single light mode (TM ) is employed and the density and thickness values cannot be separated from each other. In addition, problems encountered with ellipsometry for obtaining independent values of thickness and RI for very thin films are not observed with DPI. The principles underlining the operation of the DPI have been described in greater detail elsewhere (Cross et al., 2004). A schematic of the dual polarisation interferometer is shown in Fig. 1. 2.2. Experimental set-up The fluidic system consists of a reciprocating pump, K501 (Knauer, Germany), to provide motive force, a sixport valve, 9725i (Rheodyne, Germany) to enable the injection of the desired materials, and a diverter valve, V-101T (Anachem, UK) to enable either or both of the sensitive channels on the sensor chip to be addressed. An injection loop allowed introduction of controlled volumes of sample into the running buffer stream employed in the analysis. The volume injected was 1 mL, with the injection loop dead volume being 400 ␮L. The internal temperature was set to 20 ◦ C. Prior to DNA immobilisation the running buffer was passed over the sensor until the measured TM and TE values reached a stable baseline. An addition of 80% (v/v) ethanol/nanopure water was flown over the sensor for 2 min followed by a return to the running buffer, allowing the baseline to stabilise again. This procedure was repeated, and then a subsequent addition of nanopure water was performed in a similar manner. This was used to calibrate the sensing waveguide refractive index and thickness. The refractive index of the bulk electrolyte was calibrated by monitoring and solving the measured TM and TE phase changes from distilled water to the running buffer solution. In order to ensure no cross-contamination of samples occurred, the sample injection loop was flushed with running buffer between successive injections. Data generated by the instrument were transferred in real time to a PC where the data can be viewed in real time.

2.3. DNA sequences Synthetic oligonucleotides (Proligo, Paris, France) that were employed in DPI analysis are outlined in Table 1. Target DNA1 and target DNA2 are complementary sequences to probe 1 and probe 2, respectively. Noncomp DNA1 and non-comp DNA2 were used as control sequences. 2.4. Probe attachment method 1 (AM1) An amine modified sensor chip, FB100 Amine (Farfield Sensors) was mounted in the instrument. The amine functionalised chip is formed from a solution containing 33% 3-aminopropyltrimethoxysilane and 66% N-[3(trimethoxysilyl)propyl]. The running buffer for the experiment was 10 mM sodium phosphate buffer at a flow rate of 50 ␮L/min. Following calibration, 10 ␮M of probe 1 was allowed to incubate gently in a 4:1 mixture of 10 mM sodium phosphate buffer and 0.1 M sodium periodate (Sigma, Ireland) for 20 min, and was then allowed to flow over the sensor chip for 2 h at a flow rate of 1 ␮L/min. A volume of 0.001 M of sodium borohydride (Sigma) was then allowed to flow over the sensor chip for 10 min at a flow rate of 25 ␮L/min. Blocking of the probe layer was performed with 1 mg/mL bovine serum albumin (BSA) (Sigma) made up in 1 × phosphate buffer saline, for 30 min at a flow rate of 25 ␮L/min. Successive additions of 100 nM non-comp DNA1 and 100 nM target DNA1 were made in 2 × sodium saline citrate (SSC) (Sigma) at a flow rate of 100 ␮L/min.

Table 1 A list of the oligonucleotides employed in DPI analysis Name

Sequence

Probe 1 Target DNA1 Non-comp DNA1 Probe 2 Target DNA2 Non-comp DNA2

5 -NH2 -AGCCAGCTGAGCCAATTCA-3 5 -TGAATTGGCTCAGCTGGCT-3 5 -CAGTCAGTCAGTCAGTCAG-3 5 -GATAGTGGGATTGTGCGTCA-NH2 -3 5 -TGACGCACAATCCCACTATC-3 5 -CAGTCAGTCAGTCAGTCAGT-3

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467

1462

2.5. Probe attachment method 2 (AM2) The crosslinker molecule, 1,4-phenylene di-isothiocyanate (PDITC) (Fluka, Switzerland), was attached to the FB100 amine functionalised sensor chip by immersion in a dimethylformamide solution containing 10% pyridine (Fluka) and 1 mM PDITC for 2 h. The sensor surface was then washed sequentially with dimethylformamide (Fluka) and 1,2 dichloroethane (Fluka) and dried under a stream of nitrogen. The running buffer was 2 × SSC at a flow rate of 50 ␮L/min. Following calibration, 0.5 ␮M of probe 2 made up in 1 M Tris–HCl (Sigma) with 1% (v/v) N,Ndiisopropylethylamine (Fluka), was allowed to flow over the sensor channel for 10 min at a flow rate of 50 ␮L/min. The surface was deactivated by capping unreacted PDITC moieties by adding 50 mM 6-amino-1-hexanol (Fluka) in 99% ethanol with the same flow rate. Successive additions of 100 nM noncomp DNA2 and 100 nM target DNA2 were made in 2 × SSC at a flow rate of 50 ␮L/min.

3. Results and discussion DPI was used to investigate the covalent immobilisation of probe DNA on a silanised surface. Thickness and refractive index data were resolved directly from the measured TM and TE data. Previous experiments performed by the instrument manufacturer have shown a repeatability of 0.022 and 0.021% for the measured TM and TE values, respectively. The layer density was calculated using Eq. (1). ρL = ρDNA

nL − n B nDNA − nB

(1)

where ρL is the layer density (g/cm3 ), ρDNA the density of DNA (g/cm3 ), nL the refractive index of the layer resolved from the phase data, nB the refractive index of the bulk electrolyte solution and nDNA is the refractive index of DNA. The refractive index of DNA has been reported to be between 1.45 and 1.48 (Peterlinz et al., 1997). We have chosen a value of 1.465, which matches well with a refractive index value of 1.462 estimated using ellipsometry (Gray et al., 1997). We have chosen a value for the density of a DNA layer of 1.64 g/cm3 based on previously reported work (Peterson et al., 2002; Weidlich et al., 1987). For protein analysis the same refractive index and a density of 0.71 g/cm3 was chosen (Wen and Arakawa, 2002). It should be pointed out that the instrument measures thickness and refractive index and this density constant is used to calculate the layer mass and density. The experimental conditions for our experiments are different to those employed by Peterson and Weidlich and the density value chosen is an estimate. However, the trends reported will be the same, though the absolute values of density and mass may differ by ∼10%. The mass of the layer is calculated from Eq. (2). ML = ρL tL

(2)

where ML is the mass of the layer (ng/mm2 ) and tL is the thickness of the layer (nm) resolved from the phase data. Direct and linker mediated covalent attachment methods were employed to immobilise probe molecules to an amino derivatised sensor chip and were designated AM1 and AM2, respectively. AM1 is a modification of a previously reported method to attach oligonucleotides and RNA to silanised surfaces via the formation of a Schiff’s base (Hermanson, 1996). It is thought that the amine group of the probe molecule is oxidised with sodium periodate to generate an aldehyde group that reacts with the free amino groups on the sensor surface. The resultant imine or Schiff’s base is stabilised by reducing with sodium borohydride to form a stable covalent bond. A further blocking step with BSA is performed in order to reduce the effects of non-specific interactions with non-complementary DNA. AM2 is a more complex immobilisation method. PDITC, a 1,2 homobifunctional crosslinker molecule, is employed to attach an amine modified probe to the surface. The amino derivatised surface is coupled with excess PDITC to convert the amino groups into amino reactive phenylisothiocyanate groups. The amino modified probe molecules are coupled to these amino reactive groups to yield a surface bound probe layer. Any remaining phenylisothiocyanate groups can be deactivated by reaction with 6-amino-1-hexanol. The use of the protocol has been previously reported as an excellent method for attachment of amine modified probe molecules to silanised supports (Manning et al., 2003; Manning and Redmond, 2005) As the linker step was performed off chip, it was necessary to assume that a PDITC layer with a thickness of 1.5 nm and a refractive index of 1.45 was formed in order to fully resolve the data in subsequent steps. 3.1. Probe layer formation A typical measurement showing the change in measured TM and TE values and the resolved thickness value upon injection of probe 1 on a sensor chip employing AM1 is shown in Fig. 2. The thickness of the probe layer after washing increases by 0.69 nm. Injection of probe 2 on a sensor chip, when employing AM2, results in a thickness change of 0.5 nm.

Fig. 2. Measurement of the TM , TE and thickness values on a sensor chip employing AM1 upon (a) injection of probe 1 and (b) washing.

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467

Fig. 3. Measurement of the mass increase on a sensor chip employing AM1 upon (a) injection of probe 1 and (b) washing.

The contour length of probe 1 (19 mer) and probe 2 (20 mer), with a 1 nm C6 atom spacer is ∼7.5 and 7.8 nm, respectively (Sinden, 1994). A more accurate parameter to describe the length of single stranded DNA in solution, is its persistence length, which is defined as the distance along its backbone that it behaves like a rigid rod. In such buffer solutions, the reported persistence length of both probe molecules is ∼1 nm (Tinland et al., 1997). The small increase in layer thickness reported here for probe layer formation with both AM1 and AM2, is consistent with a probe molecule attaching to the surface in a flattened configuration. Such an orientation could be explained by multiple contact points between the probe molecule and the underlying substrate due to an electrostatic interaction between the negatively charged phosphate backbone and the positively charged surface. It has been shown with previous studies using hydroxy radical footprinting and coagulation kinetic measurements that single stranded DNA will adsorb with its long axis parallel rather than normal to the surface on positively charged surfaces (Walker and Grant, 1996). A layer thickness of 0.3 nm measured by ellipsometry for an amino modified 30 mer bound to a silanised silicon surface has been reported using similar chemistry to AM1 (Han et al., 2003). A similar thickness has been reported for the immobilisation of a 19 mer directly adsorbed on a similar amine surface using DPI (Berney and Oliver, 2005). For AM1, it can be postulated that the free nitrogenous groups of the bases along the backbone may be oxidised by the periodate resulting in more reactive groups available for reaction with the amino modified surface. There may also be a strong electrostatic interaction between the negatively charged backbone of the probe molecule and the positively charged aminated surface (Chan et al., 1997). For AM2 there may be an electrostatic attraction between positively charged amino reactive PDITC groups and unreacted free amino groups on the surface and the probe molecule. The small increase in thickness would also indicate that the probe molecule is not tethered to the surface via the linker group only upon initial immobilisation. The change in mass after injection of probe 1 to a sensor chip employing AM1 is shown in Fig. 3. The overall increase in mass of the layer was 0.2 ng/mm2 . The same increase in mass was observed after injec-

1463

tion of probe 2 on a sensor chip employing AM2. This mass increase corresponds to a probe coverage of ∼2.2 × 1010 molecules/mm2 . The maximum packing density of perfectly tethered double stranded DNA is ∼3 × 1013 molecules/cm2 . The reported mass values here correspond to a surface coverage value of only ∼7% for both immobilisation methods. However, using a molecular footprint where the probe molecule is orientated horizontal on the surface, the measured mass increase corresponds to a surface coverage of ∼35%. Similar surface coverage values have been reported for AM1 when employing an enzyme based method to assess probe coverage with the same probe concentration (Berney et al., 2000). It is interesting to note that although the probe concentration for AM1 is twenty times greater and the immobilisation 100 min longer, there is a similar mass of probe deposited on the surface for AM2. It should be pointed out that AM1 employs a probe attachment buffer of lower ionic strength than AM2, which affects surface coverage. An increase in ionic strength of the immobilisation buffer leads to an increase in probe coverage, as there is less electrostatic repulsion between adjacent probe molecules due to increased counter ion screening (Peterson et al., 2001). 3.2. Blocking the probe layer AM1 and AM2 employ different strategies in blocking the surface to prevent non-specific interactions of noncomplementary DNA. BSA is a protein molecule commonly employed as an effective blocking solution in preventing nonspecific binding of non-complementary DNA in the preparation of DNA microarrays (Taylor et al., 2003). Therefore, elucidating information about the blocking mechanism with BSA is of considerable interest. The increase in thickness and mass upon injection of BSA to a sensor chip already modified with probe 1 with AM1 is shown in Fig. 4. Fig. 4 shows that following the wash step the measured thickness increased by 1.57 nm. BSA is a heart shaped molecule whose dimensions can be approximated by an equilateral triangle with side 8 nm and depth 3 nm (Carter and Ho, 1994). It is believed that BSA adsorbs strongly to hydropho-

Fig. 4. Measured changes in the mass and thickness values on a sensor chip modified with probe 1 with AM1 upon (a) injection of 1 mg/mL of BSA and (b) washing of the sensor chip.

1464

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467

bic surfaces due to a strong entropic driving force caused by the dehydration of the sorbent surface combined with an electrostatic attraction between the positively charged surface and the negatively charge BSA molecule (Norde and Giacomelli, 2000). It has been widely reported that the conformation of BSA, a globular protein, is flexible and can adapt to changing environmental conditions (Lin and Koenig, 1976). It is probable that the increase in layer thickness observed here was due to BSA molecule absorbing on the surface and undergoing a conformational change upon interaction with the positively charged aminated surface. A similar layer thickness of 1.66 nm has been observed upon immobilisation of 1 mg/mL of BSA on a silanised silicon support measured by imaging ellipsometry (Ying et al., 2004). The overall mass increased by 0.7019 ng/mm2 and this value was lower than previous reported values for mass deposited on a silanised silicon support of 2 ng/mm2 and much lower than that expected for monolayer formation, ∼4 ng/mm2 (Norde and Giacomelli, 2000). It is probable that the BSA molecule preferentially adsorbed on the aminated surface between the negatively charged probe molecules, rather than directly over them. This would explain the lower mass of BSA on the surface that was observed here. Deactivation of the probe layer formed with AM2 with 6-amino-1-hexanol produces ambiguous results. The deactivation step generates a negatively charged surface by the conversion of any remaining PDITC moieties to hydroxyl groups. The measured thickness decreases by 0.4 nm and the measured mass decreases by 0.2 ng/mm2 . The deactivation solution may remove excess PDITC from the surface causing a decrease in the measured mass and thickness of the layer. Alternatively, the decrease could be due to the removal of the attached probe layer. However, previous characterisation of the immobilisation protocol would indicate that this is unlikely (Manning and Redmond, 2005). 3.3. DNA hybridisation detection The ability of the instrument to discriminate between the addition of non-complementary DNA and target DNA was assessed. It is important to recognise that the instrument embodies a quantitative analytical technique rather than a simple sensor function. It was thought that the relatively low probe coverages were suitable for observing maximum hybridisation efficiencies (Peterson et al., 2001). The change in measured thickness and density of a sensor chip modified with probe 1 with AM1 upon injection of 100 nM non-comp DNA1 and target DNA1 are shown in Fig. 5. The measured thickness increases by 0.19 nm upon addition of non-comp DNA1. In addition, the density also increases which suggests that non-selective interactions cause an infilling of the surface. Physiosorption of non-comp DNA1 to the surface due to the presence of unblocked free amino groups on the surface or partial hybridisation could account for this. The density decreases again upon washing

Fig. 5. Measured changes in thickness and density upon on a sensor chip modified with AM1 upon injection of (a) 100 nM non-comp DNA1 and (b) target DNA1.

of the surface which indicates removal of non-specifically bound DNA. The measured thickness changes by 0.28 nm upon addition of target DNA1. This small increase in measured thickness would indicate that there is no dramatic change in the orientation of the probe layer upon hybridisation. This is consistent with a probe molecule with poor mobility due to attachment at several points on the surface. A similar increase in thickness upon hybridisation of a 30 mer target to a probe layer formed in a similar manner has been observed using ellipsometry (Han et al., 2003). It has been shown with neutron reflectivity studies that hybridisation to disorganised thiolated oligonucleotides adsorbed onto gold surfaces does not alter the measured reflectivity value which indicates a small thickness change (Levicky et al., 1998). There is a clear discrimination between the measured density upon addition of non-comp DNA1 and target DNA1, which indicates that a selective interaction has occurred. The larger increase in density for the selective interaction can be explained by a stronger association of target DNA1 with the probe layer close to the surface. The mass increases by 0.053 and 0.075 ng/mm2 for noncomp DNA1 and target DNA1 additions, respectively. The change in mass upon addition of target DNA1 corresponds to a hybridisation efficiency of ∼35% as compared to the mass of the initial probe layer. It has previously been shown that short oligonucleotides adsorbed to the surface by multiple constraining contacts are poor hybridisation probes (Ghuo et al., 1994). This is thought to be due to their close proximity to the underlying surface, which confers a resultant loss in conformational freedom. This could explain the low hybridisation efficiency observed here. However, the selective interaction indicated by the density change would suggest that the probe molecule regains some freedom to hybridise to target DNA. It has been shown that when probe DNA adsorbs onto positively charged substrate the interaction is dominated by the negatively charged phosphate backbone resulting in the bases facing towards the solution leaving them available for hybridisation (Chan et al., 1998). This is line with previous reported work showing that a probe molecule attached to a silane layer by simple adsorption can still selectively hybridise to target DNA (Belosludtsev et al., 2001).

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467

Fig. 6. Measured change in thickness and density to a sensor chip modified with AM2 upon injection of (a) 100 nM non-comp DNA2 and (b) 100 nM target DNA2.

The change in measured thickness and density upon injection of 100 nM non-comp DNA2 and 100 nM target DNA2 to a sensor chip modified with AM2 is shown in Fig. 6. The thickness increases by 0.29 nm upon addition of noncomp DNA2. In addition, the density initially increases and then decreases upon addition of non-comp DNA2. This suggests that non-comp DNA2 binds to the surface in two different processes. The initial increase in density could be explained by a rapid physiosorption of non-complementary DNA resulting in a densification of the surface, due to an interaction with non-deactivated amino reactive PDITC groups or unreacted free amino groups on the surface. The following decrease in density could be explained by non-specific binding or association of non-comp DNA2 with the probe on the surface. Upon addition of target DNA2 the measured thickness increases by 0.9 nm. This is a larger increase in thickness than what is observed upon addition target DNA1 on a sensor chip modified with AM1. This would indicate that the probe layer has a greater degree of mobility to hybridise to target DNA2. This increase is accompanied by a corresponding large decrease in the measured density which suggests that reorientation of the probe layer away from the surface occurs upon hybridisation. This would indicate that the probe molecule has greater mobility or conformational freedom to hybridise to target DNA. Single point attachment of the probe molecule via the linker group only could account for the greater mobility of the probe molecule. Lateral interactions between neighbouring helices upon hybridisation would result in reorientation of the probe layer seen here. The deactivation step generates a negatively charged surface by converting unreacted PDITC moieties to a surface rich in hydroxyl groups. A negatively charged surface could act to displace probe molecules attached to the surface at points other than the linker site. Herne and Tarlov (1997) employ a similar approach with a hydroxy terminated spacer molecule to ensure single point attachment of a thiol modified probe molecule on a gold surface. In addition, the probe molecule attached in this way is seen to reorientate further away from the surface upon hybridisation of target DNA similar to what is seen here (Levicky et al., 1998). Berney and Oliver (2005)

1465

Fig. 7. Measured change in mass on a sensor chip modified with AM2 upon injection of (a) 100 nM non-comp DNA2 and (b) 100 nM target DNA2.

observed similar changes in the measured thickness and density for both non-selective and selective interactions for a probe layer immobilised with an avidin biotin linkage. In addition, the probe layer thickness indicated that the probe molecule was tethered at a single point only. The change in the measured mass to a sensor chip modified with AM2 upon injection of 100 nM non-comp DNA2 and 100 nM target DNA2 is shown in Fig. 7. The mass increases by 0.12 ng/mm2 upon addition of non-comp DNA2. Interestingly, this is a larger increase in mass than observed with the addition of non-comp DNA1 to a sensor chip modified with AM1. This indicates the effectiveness of the BSA blocking step. Upon addition of target DNA2 the mass increases by 0.28 ng/mm2 , which is greater than the initial mass of the probe layer. This would indicate that target DNA interacts in a non-specific manner with the surface. Assuming a similar amount of non-specific interaction of target DNA as non-complementary DNA the hybridisation efficiency is ∼85%. The observed results provide further evidence for greater conformational freedom of the probe molecule attached with AM2. The more an immobilised probe molecule is spatially removed from its solid support, the closer the molecule behaves as in a solution state and the more likely it is to react freely with dissolved target DNA. This has been previously demonstrated by comparing the effect that spacer molecule length has on hybridisation efficiency to probe molecules synthesised on an amine functionalised supports measured with radioactivity (Shchepinov et al., 1997). The precision of the instrument was in the order of 0.002 nm and 0.004 g/cm3 for the resolved thickness and density values when investigating both immobilisation methods. Further analysis is required in order to estimate the error for the thickness, density and mass values reported here for each immobilisation process. However, previous work has shown the error in these values when characterising protein layer absorption was only 3% (Lu et al., 2004). The results show that a probe molecule attached to the surface via a crosslinker group is better able to hybridise to target DNA due to its greater mobility. Several characterisation techniques such as fluorescence and radiolabelling can be used to estimate the hybridisation efficiency of a particu-

1466

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467

lar immobilisation method. With surface plasmon resonance it is possible to do so in real time and in a label independent fashion. From this information, the likely orientation and mobility of probe molecule can be inferred. However, simultaneous density and thickness measurements with DPI allow the direct impact that probe orientation and mobility have on hybridisation efficiency to be elucidated. Similar thickness and density information can be obtained with ellipsometry. However, to the best of our knowledge gaining this information with regards to DNA immobilisation and hybridisation detection in a single experiment and in real time has not been reported with this technique. Monitoring thickness and density changes in real time gives a greater insight into the nature of both selective and non-selective interactions.

4. Conclusions DPI is demonstrated to be an analytical technique of considerable potential for elucidating both quantitative and qualitative information regards the covalent immobilisation of DNA on a silanised support. Direct attachment and a linker mediated method were investigated. The orientation of the probe molecule and probe coverage at different stages of the immobilisation process were elucidated. It was also possible to selectively detect the hybridisation of target DNA. Density and thickness changes gave mechanistic information regarding both selective and non-selective interactions. A probe molecule attached via a crosslinker molecule was seen to have greater mobility to hybridise to target DNA than one attached using a direct method. This corresponded with a larger hybridisation efficiency being observed with the linker method. Immediate work could include investigating the effect that probe concentration and the buffer ionic strength have on probe orientation and hybridisation efficiency.

Acknowledgements The work was funded under the EU funded Nano2Life program. We would like to thank Marcus Swann of Farfield Sensors for considerable technical assistance. We would also like to thank Sofian Daud of INSAT.

References Almadidy, A., Watterson, J., Piunno, P.A., Raha, S., Foulds, I.V., Horgan, P.A., Castle, A., Krull, U., 2002. Direct selective detection of genomic DNA from coliform using a fiber optic biosensor. Anal. Chim. Acta 461, 37–47. Beier, M., Hoheisel, J.D., 1999. Versatile derivatisation of solid support media for covalent bonding on DNA microchips. Nucleic Acids Res. 27, 1970–1977. Belosludtsev, Y., Iverson, B., Lemeshko, S., Eggers, R., Wiese, R., Lee, S., Powdrill, T., Hogen, M., 2001. DNA microarrays based on non cova-

lent attachment and hybridisation in two dimensions. Anal. Biochem. 292, 250–256. Berney, H., West, J., Haefele, E., Alderman, J., Lane, W., Collins, J.K., 2000. A DNA diagnostic biosensor: development, characterization and performance. Sens. Actuators B 68, 100–108. Berney, H., Oliver, K., 2005. Dual polarization interferometry size and density characterisation of DNA immobilisation and hybridisation. Biosens. Bioelectron., published online. Biehle, S.J., Carrozzellaa, J., Shuklab, R., Popplewell, J., Swann, M., Freeman, N., Clark, J.F., 2004. Apolipoprotein E isoprotein-specific interactions with tissue plasminogen activator. Biochim. Biophys. Acta 1689, 244–251. Biljana, A., Cavic, M.E., McGovern, R.N., Thompson, M., 2001. High surface density immobilization of oligonucleotide on silicon. Analyst 126, 485–490. Brown, P., Bostein, D., 1999. Exploring the new world of the genome with DNA microarrays. Nat. Genet. 21, 33–37. Carter, D.C., Ho, J.X., 1994. Structure of serum abumin. Adv. Protein Chem. 45, 153–203. Chan, V., Graves, D.J., Fortina, P., McKenzie, S.E., 1997. Adsorption and surface diffusion of DNA oligonucleotides at liquid/solid interfaces. Langmuir 13, 320–329. Chan, V., McKenzie, S.E., Surrey, S., Fortina, P., Graves, D.J., 1998. Effect of hydrophobicity and electrostatics on adsorption and surface diffusion of DNA oligonucleotides at liquid/solid interfaces. J. Colloid Int. Sci. 203, 197–207. Chrisey, L.A., Lee, G.U., O’Ferall, C.E., 1996. Covalent attachment of synthetic DNA to self assembled monolayer films. Nucleic Acids Res. 24 (15), 3031–3039. Cross, G.H., Reeves, A.A., Brand, S., Swann, M.J., Peel, L.L., Freeman, N.J., Lu, J.R., 2004. The metrics of surface adsorbed small molecules on the Young’s fringe dual-slab waveguide interferometer. J. Appl. Phys. D: Appl. Phys. 37, 74–80. Drummond, T.G., Hill, M.G., Barton, J.K., 2003. Electrochemical DNA sensors. Nat. Biotechnol. 21, 1192–1199. Duggan, D.J., Bittner, M., Chen, Y., Meltzer, P., Trent, J.M., 1999. Expression profiling using cDNA microarrays. Nat. Genet. Suppl. 21, 10. Elhadj, S., Singh, G., Saraf, R.F., 2004. Optical properties of an immobilized DNA monolayer from 255 to 700 nm. Langmuir 20, 5539–5543. Georgiadis, R., Peterlinz, K.P., Peterson, A.W., 2000. Quantitative measurements and modeling of kinetics in nucleic acid monolayer films using SPR spectroscopy. J. Am. Chem. Soc. 122, 3166–3173. Gray, D.E., Case-Green, S.C., Fell, T.S., Dobson, P.J., Southern, E.M., 1997. Ellipsometric and interferometric characterisation of DNA probes immobilised on a combinatorial array. Langmuir 13, 2833–2842. Ghuo, Z., Guilfoyle, R.A., Thiel, A.J., Wang, R., Smith, L.M., 1994. Direct fluorescence analysis of genetic polymorphisms by hybridisation with oligonucleotide arrays on glass supports. Nucleic Acids Res. 22, 5456–5465. Han, S., Ralin, D., Wang, J., Li, X., Zhou, F., 2003. Attachment of monoclonal antibody molecules to surface confined DNA duplexes images by atomic force microscopy. Langmuir 19, 8943–8950. Hermanson, G.T., 1996. Bioconjugate Techniques Pierce Chemical Company. Academic Press. Herne, T.M., Tarlov, M.J., 1997. Characterization of DNA probes immobilized on gold surfaces. J. Am. Chem. Soc. 119, 8916–8920. Lenigk, R., Carles, M., Ip, N., Sucher, J., 2000. Surface characterisation of a silicon chip based microarray. Langmuir 17, 2497–2501. Levicky, R., Herne, T.M., Tarlov, M.J., Satija, S.K., 1998. Using self assembly to control the structure of DNA monlayers on gold: a neutron reflectivity study. J. Am. Chem. Soc. 120, 9787–9792. Lin, V.J.C., Koenig, J.L., 1976. Raman studies of bovine serum albumin. Biopolymers 15, 203–218. Lu, R.J., Swann, M.J., Peel, L.L., Freeman, N.J., 2004. Lysozyme adsorption studies at the silica/water interface using dual polarization interferometry. Langmuir 20, 1827–1832.

B. Lillis et al. / Biosensors and Bioelectronics 21 (2006) 1459–1467 Manning, M., Harvey, S., Galvin, P., Redmond, G., 2003. A versatile multi-platform biochip surface attachment chemistry. Mater. Sci. Eng. C 1049, 1–5. Manning, M., Redmond, G., 2005. Formation and characterization of DNA microarrays at silicon nitride substrates. Langmuir 21, 395–402. Norde, W., Giacomelli, E., 2000. BSA structural changes during homomolecular exchange between the adsorbed and the dissolved states. J. Biotechnol. 79, 259–268. Peterlinz, K.A., Gerogiadis, R., Herne, T., Tarlov, M., 1997. Observation of hybridization and dehybridization of thiol-tethered DNA using two color surface plasmon resonance spectroscopy. J. Am. Chem. Soc. 119, 3401–3402. Peterson, A.W., Heaton, R.J., Georgiadis, R.M., 2001. The effect of surface probe density on DNA hybridization. Nucleic Acids Res. 29, 5163–5168. Peterson, A.W., Wolf, L.K., Georgiadis, R., 2002. Hybridization of mismatched or partially matched DNA at surfaces. J. Am. Chem. Soc. 124, 14601–14607. Sastry, M., 2000. A note on the use of ellipsometry for studying the kinetics of formation of self-assembled monolayers. Bull. Mater. Sci. 23, 159–163. Shchepinov, M.S., Case Green, S.C., Southern, E.M., 1997. Steric factors influencing hybridisation of nucleic acids to oligonucleotide arrays. Nucleic Acids Res. 25 (6), 1155–1161. Sinden, R.S., 1994. DNA Structure and Function. Academic Press. Shlyakhtenko, S.L., Gall, A.A., Jeffrey, J.W., Hawn, D.D., Lyubchenko, Y.L., 1999. Atomic force microscopy imaging of DNA covalently

1467

immobilised on a functionalised mica substrate. Biophys. J. 77, 568–576. Swann, M.J., Peel, L.L., Carrington, S., Freeeman, N.J., 2004. Dualpolarisation interferometry: an analytical technique to measure changes in protein structure in real time, to determine the stoichometry of binding events, and to differentiate between nonspecific interactions. Anal. Biochem. 329, 190–198. Taylor, S., Smith, S., Windle, B., Guiseppi-Elie, A., 2003. Impact of surface chemistry and blocking strategies on DNA microarrays. Nucleic Acids Res. 31, e87. Tinland, B., Pluen, A., Sturm, J., Weill, G., 1997. Persistence length of single-stranded DNA. Macromolecules 30, 5763–5765. Walker, H.W., Grant, S.B., 1996. Coagulation and stabilization of colloidal particles by adsorbed DNA block copolymers: the role of polymer conformation. Langmuir 12, 3151–3156. Wang, J., 2000. From DNA biosensor to gene chips. Nucleic Acids Res. 28, 3011. Watterson, J., Piunno, A.E., Krull, U.J., 2002. Practical physical aspects of interfacial nucleic acid oligomer hybridisation for biosensor design. Anal. Chim. Acta 469, 115–127. Weidlich, T., Lindsay, S.M., Rupprecht, A., 1987. The optical properties of Li- and Na-DNA films. Biopolymers 26, 439–453. Wen, J., Arakawa, T., 2002. Refractive index of selected proteins in aq. NaCl. Anal. Biochem. 280, 327–329. Ying, P., Jin, G., Tao, Z., 2004. Competitive adsorption of collagen and bovine serum albumin-effect of the surface wettability. Colloids Surf. B., Biointerfaces 33, 359–363.