Quantitative evaluation of oligonucleotide surface concentrations using polymerization-based amplification
- First Online:
- Cite this article as:
- Hansen, R.R., Avens, H.J., Shenoy, R. et al. Anal Bioanal Chem (2008) 392: 167. doi:10.1007/s00216-008-2259-6
- 518 Downloads
Quantitative evaluation of minimal polynucleotide concentrations has become a critical analysis among a myriad of applications found in molecular diagnostic technology. Development of high-throughput, nonenzymatic assays that are sensitive, quantitative and yet feasible for point-of-care testing are thus beneficial for routine implementation. Here, we develop a nonenzymatic method for quantifying surface concentrations of labeled DNA targets by coupling regulated amounts of polymer growth to complementary biomolecular binding on array-based biochips. Polymer film thickness measurements in the 20–220 nm range vary logarithmically with labeled DNA surface concentrations over two orders of magnitude with a lower limit of quantitation at 60 molecules/μm2 (∼106 target molecules). In an effort to develop this amplification method towards compatibility with fluorescence-based methods of characterization, incorporation of fluorescent nanoparticles into the polymer films is also evaluated. The resulting gains in fluorescent signal enable quantification using detection instrumentation amenable to point-of-care settings.
KeywordsSignal amplification Molecular diagnostics Quantitative nucleic acid detection DNA microarrays Visible light photopolymerization Surface-initiated polymerization
Quantitative evaluations of biomolecule concentrations are becoming increasingly necessary due to the expanding discovery and characterization of potential disease markers and the development of their corresponding applications in molecular diagnostics. Quantitative, sequence-specific analysis of polynucleotide biomarkers potentially allows for the monitoring of viral loads, allowing for drug efficacy evaluation [1, 2], elucidation of various disease states and progressions , and may provide clinical validation of a diagnosis . Specifically with application to human cancers, several types of nucleic acid biomarkers require quantitative analysis, such as evaluation of over- or underexpressed RNA transcripts, analysis of genomic amplifications or deletions, as well as evaluation of epigenetic markers [5, 6]. Mutation detection, traditionally performed on a qualitative basis, can also benefit from quantitative analysis through determination of wild-type to mutant-type ratios, as well as providing distinction between stochastic and pathogenic mutations .
Several challenges for quantitative biomolecular analysis in molecular diagnostics lie within the prerequisites of point-of-care implementation, requiring assays that are rapid and robust, use stable and inexpensive reagents, and are relatively simple to implement. The determination of sequence-specific nucleic acid concentrations has been developed through several PCR-based target amplification technologies, most notably real-time quantitative PCR . Other technologies aim at achieving quantitative gains from signal amplification, such as rolling circle amplification , which has been creatively used in combination with BEAMing for sensitive quantitative detection , and branched DNA technology . While such detection assays offer sensitive, powerful methods of quantitative amplification and detection, they are still generally limited by expensive and often complex protocols due to their reliance on enzymes and/or fluorescently labeled oligonucleotides for amplification, or may require instrumentation beyond the scope of point-of-care settings.
Recently, polymerization-based amplification has been demonstrated as a rapid, nonenzymatic method of signal amplification capable of offering visual detection of the presence of labeled polynucleotide targets on a biochip surface [12, 13]. With this method, amplification and detection are achieved by covalently coupling photoinitiator molecules to streptavidin proteins followed by binding onto biotinylated oligonucleotide targets immobilized to a biosensor surface. Upon contacting such biochips with monomer and light, a surface-initiated polymerization reaction occurring exclusively at the binding sites generates high molecular weight, highly visible polymer films that signal the binding event. While this method offers high amplification factors and sensitive, visual detection of biotinylated DNA molecules using assay conditions favorable for field implementation, analysis thus far is largely limited to a qualitative, positive/negative determination of the presence of model 3′-biotin-labeled capture sequences.
The purpose of this manuscript is to present a visible light polymerization-based amplification system that uses the biofunctional photoinitiator streptavidin-eosin isothiocyanate (SA-EITC) as a new method for quantifying surface concentrations of biotinylated nucleic acid targets. Such developments towards quantitation will allow for the implementation of this method in an increased number of molecular diagnostic applications. The potential for quantitation is investigated using array-based biochips containing spots with diluted amounts of 3′-biotin-labeled target oligos such that the amount of amplification as a function of oligo surface concentration can be determined. This approach of directly printing the biotinylated targets onto the surface serves as a model system  to evaluate the characteristics of polymerization-based amplification. Such a system is representative of a biochip after DNA target labeling with biotin that may be done either in solution  followed by hybridization to complementary capture probes, or through on-chip, primer extension labeling performed after complementary duplex formation [16, 17, 18]. Thus, the surface density of the biotinylated probe is proportional to the target solution concentration, dependent on the hybridization efficiency between a target and its surface-bound, complementary capture probe (typically, Khyb = 109 M−1) , and the labeling efficiency.
Aminosilated glass slides were purchased from CEL Associates (Pearland, TX, USA). Streptavidin (SA) proteins were purchased from Pierce (Rockford, IL, USA) and stored at 4 °C until use. Polyethylene glycol diacrylate, Mn = 575 Da (PEGDA), methyl diethanol amine (MDEA), 1-vinyl-2-pyrrolidinone, dithiothreitol (DTT), 50x Denharts solution, and Tween 20 were purchased from Sigma (St. Louis, MO, USA). The fluorescent nanoparticles were 0.02 μm-diameter, carboxy-modified Nile Red FluoSpheres (Invitrogen, Carlsbad, CA, USA), encapsulating ∼102 fluorophores per particle and supplied at a stock concentration of 4.5 × 1015 nanoparticles/mL containing 0.1% sodium azide preservative. The azide preservative was removed with the use of 100,000 Da dialysis columns (Spectra-Por, Rancho Dominguez, CA, USA). The purified FluoSpheres were stored at 4 °C until further use. Capture oligonucleotede sequences used were obtained from Operon (Huntsville, AL, USA) and were of the same sequence as previously reported .
Dilution chip fabrication
Functionalization and characterization of dilution chips with SA-EITC
SA-EITC was synthesized, purified, and characterized according to a previously published protocol . Dilution chips were functionalized with SA-EITC by applying a mixture of unlabeled SA and SA-EITC at 1 μg/mL in a 5X Denharts, 1 X PBS solution for 30 minutes followed by rinsing with wash buffer (1 M NaCl, 0.1 M Tris, 0.1% (wt/vol) Tween 20), then H2O, and dried with N2. Preamplified, SA-EITC functionalized dilution chips were characterized using the green channel of a microarray scanner at 100% PMT gain (Agilent, Santa Clara, CA, USA) calibrated using a Cy3/Cy5 calibration chip (Full Moon Biosystems, Sunnyvale, CA, USA).
Monomer solution contained 434 mM PEGDA, 225 mM MDEA, and 37 mM 1-vinyl-2-pyrrolidinone in H2O (pH 9). Fluorescent monomer formulations included the addition of fluorescent nanoparticles to this monomer formulation at 90 or 500 nM concentrations. Conventional UV-Vis spectroscopy was used to measure absorbance from fluorescent monomer formulations. Monomer was purged with argon and then 500 μL was contacted with the SA-EITC functionalized dilution chip surface using a Whatman Chip Clip (Whatman, Florham Park, NJ, USA). Collimated, 495–650-nm polychromatic light was used to initiate polymerization from an Acticure light source (EFOS, Mississauga, ON, Canada) at an intensity of 10 mW/cm2, as measured by an EFOS Cure-Rite radiometer. This was achieved with the use of an in-house 350–650-nm internal interference filter and a 490-nm stopband limit filter (Edmond Optics, Barrington, NJ, USA) applied to the end of a collimating lens attached to the end of a light guide. When fluorescent nanoparticles were included in monomer solutions, light intensity was increased accordingly to account for light attenuation occurring from the nanoparticles over a 1-mm-thick monomer layer. Prior to light exposure, the atmosphere was purged with argon for 5 min followed by light exposure for the desired time period. After amplification, unreacted monomer was removed from the surface by gently rinsing with water.
Rhodamine B coupling into polymer films
In an effort to achieve highly visible polymer films for visualization with fluorescence microscopy, methacryloxyethyl thiocarbamoyl rhodamine B (Polysciences, Warrington, PA, USA) was coupled to pendant double bonds present on PEG films on a surface through addition reactions with a DTT linker molecule. A 400 mM solution of DTT was prepared in 0.1 M carbonate buffer, pH 9.0. The DTT solution was contacted with the polymer chip for 25 min, followed by 30 min of washing with H2O. A solution of 4 μg/mL methacryloxyethyl thiocarbamoyl rhodamine B was then prepared in carbonate buffer and contacted with the DTT functionalized gel for 30 min, followed by 30 min of washing in H2O. This coupling procedure was useful as a method of visualizing polymer films with fluorescence microscopy, but was not sufficient for use in quantifying film thicknesses.
Monitoring of bulk polymerization kinetics
The initiating capability of SA-EITC in various monomer formulations was verified by monitoring bulk polymerization kinetics through the use of Fourier transform infrared (FTIR) spectroscopy (Nicolet Instrument Corporation, Madison, WI, USA). The conversion rate of carbon–carbon double bonds in monomer to carbon–carbon single bonds in polymer was measured by monitoring the absorbance at 6150 to 6212 cm−1 with time, using monomer formulations and light intensities identical to those used during on-chip polymerizations.
Post-amplification characterization of dilution chips
Film thicknesses were measured with a Dektak 6 M profilometer (Veeco, Plainview, NY, USA) using a 12.5-μm diamond stylus tip at a stylus force of 1 mg. Brightfield and fluorescent pictures of amplified dilution chips were obtained through the use of a Leica MZ FLIII stereomicroscope (Leica Microsystems, Wetzlar, Germany) calibrated with a Cy3/Cy5 calibration chip. Nile Red fluosphere-functionalized polymer films on the surface were further quantified with the use of an Agilent Technologies microarray scanner (green channel, PMT 1%).
Results and discussion
Dependence of film thickness on surface oligonucleotide concentrations
The SA-EITC product has the ability to both initiate free-radical photopolymerizations by generating radicals through photoinduced electron transfer reactions with amine coinitiators, and to retain high-affinity binding to biotinylated substrates . Upon the functionalization of a biotinylated dilution chip with SA-EITC, but before amplification, these dilution chips are imaged with a fluorescence scanner to verify the allocation of photoinitiators on the surface at densities proportional to the specific biotin label densities at each binding site. During amplification, long exposure times to low-intensity light were used to completely degrade SA-EITC, even from the spots containing the lowest initiator concentrations. This is followed by rapid propagation reactions of free radicals through monomer to generate high molecular weight polymer films on the surface. At long exposure times, long-wave visible light is desired to minimize or eliminate any nonspecific bulk polymerization that is characteristically observed from light sources with emissions in the UV region. Thus, polymerizations using polychromatic green light from a mercury arc lamp were performed. By examining several longer exposure times in the amplification, times of greater than 30 min were found to be required to obtain maximum sensitivity.
The increase in film thickness observed with higher biotin target concentrations is directly due to the corresponding increase of photoinitiator surface concentration on binding with SA-EITC. With a higher photoinitiator concentration on the surface, an increased number of propagating radical chains are generated during the polymerization process. Such free radicals are capable of diffusing from the surface and through the forming polymer film, thereby increasing the conversion of monomer into surface-bound polymer and ultimately increasing film thicknesses. This trend is consistent with both modeling and experimental findings of similar systems in the literature . After the polymerization reaches high conversions, radical diffusivity throughout the hydrogel membrane decreases and eventually limits the extent of the polymerization. This results in the saturation of film thicknesses at spots containing higher concentrations of targets.
Competitive binding to tune regions of quantitative amplification
The dynamic dependence of the film thickness on photoinitiator surface concentration allows for an additional method of tuning the sensitivity and dynamic amplification of this assay to concentration regions of interest. By varying the binding parameters on the surface, allocation of different amounts of photoinitiator to sites containing a constant amount of biotin target is achievable. This potentially gives the user the ability to detect and quantify biotinylated DNA concentrations that may be either undetectable with less efficient binding or that may be in a saturated region with more efficient binding.
In the case of detecting nucleic acid hybridization with polymerization-based amplification, an alternative and perhaps more feasible approach involves varying capture sequence densities. The capture sequence density has been shown both experimentally and through modeling considerations to be a crucial parameter for achieving an optimal signal from hybridization [23, 24]. In an effort to shift a saturated response generated from complementary target-capture hybridization to a quantifiable response, sites with less then optimal capture probe densities could be printed. Such sites would then render lower photoinitiator concentrations than optimal sites at the same target concentrations, potentially shifting a saturated response to a quantifiable response. With the high-throughput capability of DNA chips, several sites of identical complementary capture sequences could be printed at optimal and successively less than optimal conditions within a single array, allowing for a quantifiable response to be achieved over many orders of magnitude.
Coupling amplified fluorescent signal with polymer growth
To implement inexpensive detection instrumentation for quantification, it is essential that PEG films with quantitative thicknesses generated from polymerization-based amplification be correlated with a measurable signal consistent with inexpensive methods of characterization. Direct film thickness measurements at the nanometer level are generally unfeasible in clinical settings due to the expense of the necessary instrumentation (profilometry, ellipsometry, atomic force microscopy). However, developing this amplification system in order to aim to generate an amplified fluorescent signal that corresponds with film thickness is particularly desirable, due to the numerous types of instrumentation available for characterizing biomolecule interactions using fluorescence measurements . For array-based biosensors, such instrumentation ranges from microarray scanners using photomultiplier tube (PMT) detectors (detection limit < 0.05 fluors/μm2, 106 dynamic range) , costing hundreds of thousands of dollars, to inexpensive, portable fluorescence readers employing detection with less-sensitive complementary metal oxide semiconductor or charged coupled device (CCD) detectors that are more amenable to point-of-care testing. With these considerations in mind, an initial demonstration of the use of polymerization-based amplification to achieve an amplified, quantitative fluorescent signal is demonstrated herein.
The preferred approach to coupling fluorescent signal with polymer growth is to include the fluorescent moiety in the monomer formulation such that it is copolymerized or encapsulated into the crosslinked polymer network. With the commercial availability of several acrylated fluorophores, copolymerization of molecules such as fluorescien-o-acrylate into PEG-based polymer films on a solid substrate has been demonstrated as a method of coupling a fluorescent response to hydrogel formation . When applied to photopolymerization-based amplification, limitations to such approaches are inherent, namely: nonspecific initiation of the bulk monomer due to photoreduction reactions occurring between photoexcited fluorophores and amine coinitiators; decreased initiation rates from the surface-bound eosin initiators due to light attenuation to the surface caused by the absorbing fluorophore; and finally, saturation of the fluorescent signal after amplification due to the quenching of fluorescent molecules at higher concentrations. An alternative to fluorescent acrylates are commercially available 20-nm-diameter polystyrene microspheres that encapsulate large numbers of fluorophores per particle. These fluorescent particles mitigate many of the above limitations, largely due to the fact that they minimize fluorophore contact with their surrounding environment and remain strongly fluorescent.
An important observation from Fig. 6 is that without the addition of SA-EITC to monomer, polymerization is only observed at long exposure times (greater than 40 min, shown in curve d of Fig. 6). Photoinitiation due to photoexcited fluorescent nanoparticles at nanomolar concentrations is thus considerably less efficient than initiation from SA-EITC. This exposure window where polymerization occurs exclusively from SA-EITC allows for specificity in the amplification. Upon applying this system to a surface-initiated polymerization on dilution chips, polymerization at lower exposure will only occur due to the presence of SA-EITC from biotin–avidin binding. Unwanted, nonspecific initiation from the fluorescent nanospheres with bulk monomer does not occur until much longer exposure times. Thus, at lower exposures, the fluorescent monomer formulation is amenable to on-chip signal amplification in order to obtain amplified fluorescent signals.
Moreover, the gain in fluorescent signal achieved through the amplification allows for the characterization of the biotinylated DNA concentrations using less sophisticated instrumentation amenable to point-of-care settings. Prior to amplification, SA-EITC-functionalized dilution chips were only detectable using a microarray scanner employing a PMT detector with low detection limits measured at 0.28 Cy3 fluors/μm2. After the amplification, the last six columns on the dilution chip were characterized using a stereomicroscope containing a CCD camera: a considerably less expensive, less sophisticated instrument with higher detection limits, measured at 570 Cy3 fluors/μm2.
The results presented represent an initial demonstration of the use of polymerization-based amplification to supply amplified, quantitative fluorescent signals useful for the measurement of DNA surface concentrations. Currently, the addition of the fluorescent nanoparticles decreases polymer film thicknesses and thus decreases the sensitivity of the assay as compared to amplification without their inclusion. This is consistent with the decrease in final conversions of bulk polymerizations with higher concentrations of nanoparticles noted in Fig. 6. Work is currently underway to obtain thicker polymer films containing higher concentrations of nanoparticles, allowing for higher gains in fluorescent signal and increased sensitivity with quantitative capability. Such advancements will further the use of polymerization-based amplification for quantitative applications.
Polymerization-based amplification has the ability to provide quantitative evaluation of genetic biomarkers over at least two orders of magnitude, as demonstrated through film thickness measurements correlating to the surface concentrations of biotinylated DNA targets. The photoinitiator surface density is a critical parameter in determining the amount of polymer growth observed from the surface, as this parameter can be manipulated depending on binding conditions to tune the amount of amplification generated from a given number of biotinylated targets localized on a biosensor surface. This amplification system is amenable to obtaining an amplified fluorescent response correlating with film thickness, enabling the evaluation of biomarker quantities with less expensive instrumentation.
The large, dynamic variation in film thicknesses and amplified fluorescent signal obtained from polymerization-based amplification opens up many potential uses for this assay that were previously unachievable from only a positive/negative response. For example, polymer growth arising from nonspecific biomolecule interactions can be characterized as background amplification such that only a deviation from this background becomes significant. This means that polymerization-based amplification can potentially be used in applications where both high specificity and sensitivity are critical, such as in a high-throughput SNP screening. Similarly, markers that become significant only when deviating from characterized background levels can also be analyzed. This is the case for the majority of proteomic biomarkers, and currently work is underway aimed at extending this assay to the detection of antigen–antibody binding.
This work has been supported by the State of Colorado and the University of Colorado Technology Transfer Office. This work has also been supported by NIH 1R21 CA 127884. RRH would like to acknowledge the Graduate Assistantship in Areas of National Need Fellowship from the US Department of Education and the Teets Family Endowed Doctoral Fellowship in Nanotechnology for continued financial support of this work. HJA would like to acknowledge the National Science Foundation Research Fellowship for support of this work.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.