This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright
Author's personal copy Biosensors and Bioelectronics 25 (2009) 48–54
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Cowpea mosaic virus nanoscaffold as signal enhancement for DNA microarrays Carissa M. Soto a,∗ , Kate M. Blaney b , Mubasher Dar c , Manzer Khan c , Baochuan Lin a , Anthony P. Malanoski a , Cherise Tidd a , Mayrim V. Rios a , Darlah M. Lopez a , Banahalli R. Ratna a a
Center for Bio/Molecular Science and Engineering, Naval Research Laboratory, 4555 Overlook Avenue SW, Washington, DC 20375, USA Nova Research Inc., Alexandria, VA 22308, USA c GE Healthcare Biosciences, 800 Centennial Avenue Piscataway, NJ 08855, USA b
a r t i c l e
i n f o
Article history: Received 16 April 2009 Received in revised form 1 June 2009 Accepted 2 June 2009 Available online 10 June 2009 Keywords: Cowpea mosaic virus DNA microarrays Signal enhancement Respiratory pathogens Gene expression
a b s t r a c t Previous studies have shown that a functionalized viral nanoparticle can be used as a fluorescent signalgenerating element and enhance detection sensitivity for immunoassays and low density microarrays. In this study, we further tested this ability in commercial DNA microarrays, including Affymetrix high density resequencing microarray. Optimum conditions for NeutrAvidin and dye coupling to a doublecysteine mutant of cowpea mosaic virus (CPMV) were found to be comparable to the commonly used streptavidin-phycoerythrin (SAPE) for high density resequencing microarray. A 3-fold signal enhancement in comparison to Cy5-dCTP controls was obtained when using nanoparticles on control scorecard expression microarrays. Hybridization results from commercially available 8000 rat expression arrays indicate an increment of 14% on the detected features when the virus complex was used as the staining reagent in comparison to Cy5-dCTP controls. The current work shows the utility of the CPMV-dye nanoparticles as a detection reagent in well-established detection platforms. Published by Elsevier B.V.
1. Introduction DNA microarray technology is an approach that has made significant contributions in many fundamental areas of basic, environmental, and clinical research (Mouritsen and Hillyard, 1999; Sevenet and Cussenot, 2003; Wang et al., 2003). As the use of DNA microarray technology enables the rapid and simultaneous interrogation of thousands of genetic elements, it has quickly become a preferred tool for applications in DNA and RNA sequence analysis, gene expression profiling, genotyping of single-nucleotide polymorphisms, and the molecular detection of pathogenic organisms (Duggan et al., 1999; Wang et al., 2002). However, despite its tremendous utility, one of the major problems in the use of DNA microarray technology has been assay sensitivity and the need for target or signal amplification (Vora et al., 2004). Attempts to increase the signal generated by each label or hybridized DNA molecule has been an area of ongoing research interest, and several techniques, such as dendrimers (Borucki et al., 2005) and silica nanoparticles (Wang et al., 2007; Zhao et al., 2004; Zhao et al., 2003), have been developed for this purpose (Greninger et al., 2005; Lee et al., 2003). Previous studies by our group have shown that a 30 nm viral nanoparticle as a scaffold to conjugate dye and protein-like recog-
∗ Corresponding author. Tel.: +1 202 404 6006; fax: +1 202 7679594. E-mail address:
[email protected] (C.M. Soto). 0956-5663/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.bios.2009.06.009
nition entities at specific amino acid sites can be used for signal enhancement of hybridization on low density microarrays (Blum et al., 2008; Soto et al., 2006) and sandwich immunoassays (Sapsford et al., 2006; Soto et al., 2008). The benefit of using viral particles is that reporter molecules are associated with fixed locations on the viral capsid which avoids the common pitfall of self quenching when using higher amounts of fluorophores per reaction site. The primary goal of this study was to build upon our previous successes on virus-NA-Cy5 on low density microarrays (Soto et al., 2006). Further optimization of the virus-NA-dye complexes were undertaken to make it amenable as the output element on commercially available microarrays such as Affymetrix high density resequencing microarrays (Hacia, 1999) and open DNA array platforms utilizing cyanine dye labeled targets for detection. A Streptavidin-R-phycoerythrin (SAPE) conjugate is used as the staining reagent in the Affymetrix GeneChip. R-phycoerythrin is one of the three main classes of light-harvesting proteins in algae. It has a molecular weight of 240 kDa, extinction coefficients of 2.4 × 106 M−1 cm−1 and a quantum yield of 0.8 (Kronick, 1986) and its crystal structure had been determined to 2.7 Å resolution (Contreras-Martel et al., 2001). Its fluorescence characteristics had made it useful in applications such as cell labeling and flow cytometry. It is typically used in the Affymetrix system in a three stage process consisting of a SAPE stain, followed by an antibody amplification step and final stain with SAPE. For this work, we chose to couple the dyes to the 300 reactive lysines (Lys, amine-containing amino acid) groups on the surface of
Author's personal copy C.M. Soto et al. / Biosensors and Bioelectronics 25 (2009) 48–54
Fig. 1. Double cysteine mutant of CPMV. Image generated from the wildtype (WT) virus PDB file (1NY7). (A) A 30 nm diameter icosahedral virus particle, DM-CPMV 228/2102, made of 60 identical protein subunits containing a total of 120 cysteines. (B) DM-CPMV protein subunit to which two cysteines (thiol-containing amino acid) were incorporated by point mutations at positions 28 (red circle) and 102 (blue circle) of the large subunit. (C) Natural-occurring lysine groups shown as light blue circles on the protein subunit (Chatterji et al., 2004) (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article).
the double cysteine mutant of CPMV (DM-CPMV) (Fig. 1) (Chatterji et al., 2004) which offers the possibility of incorporating more dye. DM-NA-Cy3 was used as the staining reagent in comparison to SAPE; both were used in a single step, when testing performance on the Affymetrix chip. Furthermore, DM-NA-Cy3 and DM-NA-Cy5 were tested in comparison to standard Cy3-dCTP, Cy5-NHS and Cy5dCTP controls for DNA microarrays analyzed on expression array platform. Optimization experiments are reported in which a series of dyes were tested and the number of NA and dye molecules were optimized for the best dye candidates for signal amplification and sensitivity. 2. Materials and methods1 2.1. Funtionalization of DM-CPMV with NeutrAvidin and dye The DM-CPMV, this will hereafter be referred to as DM, is purified from infected black eye pea plants using methods previously described (Chatterji et al., 2004). A total of 1 mg of DM at a final volume of 1.5 ml in 100 mM potassium phosphate buffer pH 7.0 (100 PB; Fisher Scientific, Pittsburgh, PA) was purified in a Hi-Trap Desalting Column (GE Healthcare Biosciences, Piscataway, NJ) to remove dithiothreitol (DTT) from virus sample. Sample was eluted with 100 PB. Elutions were analyzed by UV–vis spectroscopy (Cary 5000, Varian). Typically the first 1.5 ml elution contains most of the DM. Virus concentration was determined using absorbance values of the peak centered at 260 nm (Soto et al., 2006). Purified DM was reacted with 5 g/l EZ-link maleimide activated NeutrAvidin biotin-binding protein (maleimide-NA; Pierce, Rockford, IL) prepared according to manufacturer instructions. Molar ratios of DM to maleimide-NA of 1:60, 1:120, and 1:180 were used in separate reactions in a final concentration of 0.5 mg virus/ml. Reaction was performed in the dark at room temperature (RT) for 16 h. Excess NA was removed by dialysis using a 100 kDa Molecular Weight Cut off (MWCO) Spectrapor dialysis membrane (Fisher Scientific). Dialysis was performed against 25 mM PB pH 7.0 for 48 h. Recovered DM-NA solution was analyzed by UV–vis spectroscopy. For dye coupling, in a typical reaction 500 g of DM-NA in 25% DMSO (Sigma-Aldrich, St. Louis, MO) was reacted with Alexa Fluor 546® carboxylic acid succinimidyl ester (NHS-Alexa546; Invitrogen, Carlsbad, CA) or CyTM 3 Mono NHS ester (NHS-Cy3; GE
1
All chemicals purchased from USA sources.
49
Healthcare Biosciences). Molar ratios of virus to dye were 1:120, 1:600, 1:1200, and 1:3000. A series of samples having different ratios of dyes per virus (D/V) were obtained for optimization purposes. In the case of CyTM 5 Mono NHS ester (NHS-Cy5; GE Healthcare Biosciences) a 1:600 molar ratio of virus to dye was kept. Prior to the addition of DMSO and dye, DM-NA in 100 PB was incubated in an ice bath for 5 min after which DMSO is added slowly with gentle mixing, then, a corresponding amount of 10 g/l NHS-Alexa546, NHS-Cy3 or NHS-Cy5 in DMSO was added and the mixture vortexed. DM-NA was kept at a concentration of 1 mg/ml in the reaction mix. The mixture was incubated at RT in the dark for 16 h. DM-NA-dye mixtures were loaded separately in a SuperoseTM 6 prep grade (GE Healthcare Biosciences) column pre-equilibrated in 50 PB pH 7.0. Individual 2 ml elutions were collected and analyzed by UV–vis (Cary 5000, Varian) and fluorescence spectroscopy (FluoroMax-3, Jobin Yvon Horiba). The DM-NA-dye complex typically elutes in the first fraction, while the free dye is found in the fourth fraction. D/V values were calculated by using extinction coefficients provided by the manufacturer to determine dye concentration and virus concentration was determined using virus absorbance at 260 nm (Soto et al., 2006). The presence of the dye peak in virus-containing fractions is strong evidence that the dye has been coupled to the virus. Further characterization is performed in agarose gel electrophoresis. Intact virus runs as a single band on agarose gels (Soto et al., 2004). The band corresponding to DM-NAdye complex is visualized on unstained gels at the expected position for the nanoparticles, while the free dye runs faster. 2.2. Respiratory pathogen microarray v.1 chip design The respiratory pathogen microarray v.1 (RPM v.1) chip design includes 57 target genes, partial sequences from the genes containing diagnostic regions of each pathogen (i.e. E1A, hexon and fiber for adenoviruses; hemagglutinin, neuraminidase and the matrix genes for influenza A viruses). This allows resequencing of 29.7 kb of sequences from 26 respiratory pathogens and biowarfare agents known to cause “flu-like” symptoms at early stages of infection, was described in detail in a previous study (Lin et al., 2006). 2.3. Prototype adenovirus for Affymetrix system Prototype adenovirus (Ad) type 7 (ATCC VR-7, Gomen), was obtained from American Type Culture Collection (ATCC, Manassas, VA) and propagated using A-549 cells as previously described (Kessler et al., 2004). Ad7 DNA was extracted from culture supernatant using MasterPureTM DNA purification kit (Epicentre Technologies, Madison, WI) following the manufacturer’s recommended protocol without RNase digestion step. Ad7 DNA were amplified by a modified version of a random PCR protocol as described in the previous publication (Wang et al., 2002; Wang et al., 2003; Wang et al., 2006). The amplified products were subjected to fragmentation and labeling prior hybridizing to the RPM v.1 chips according to previous published protocol (Lin et al., 2007). 2.4. Microarray hybridization and processing: Affymetrix system For controls, microarray hybridization and processing were carried out according to the manufacturer recommended protocol (Affymetrix Inc., Santa Clara, CA). For viral particles, the washes and stains were carried out using FlexDNA WS5 450 protocol with only one staining step. After scanning, the GCOS software is used to reduce the raw image (.DAT) file to a simplified file format (.CEL file) with intensities assigned to each of the corresponding probe positions. Finally, the GDAS software is used to apply an embedded version of the ABACUS (Cutler et al., 2001) algorithm to produce
Author's personal copy 50
C.M. Soto et al. / Biosensors and Bioelectronics 25 (2009) 48–54
an estimate of the correct base calls, comparing the respective intensities for the sense and anti-sense probe sets. To increase the percentage of base calls, we adjusted the parameters to allow the most permissive base calling as described in previous publication (Lin et al., 2006).
GenePix Pro software (Axon Instruments Molecular Devices) and data analyzed using Microsoft Excel® .
2.5. Microarray dye labeling and cDNA synthesis for DM-NA-dye detection: GE scorecard control expression system
For this study, we chose to couple the dye molecules at the Lys via NHS esters in an effort to increase the efficiency of the coupling and the number of dyes per virus since the Lys were found to be more reactive than the cysteines (Cys; thiol-containing amino acid; (Soto et al., 2008)). In general, we found that 10× less dye was needed to label the Lys relative to the Cys. Initial experiments to find the optimum DM-NA-dye to be used in the Affymetrix system were performed with Alexa546, Cy3, Rhodamine (Pierce), and DyLight549 (Pierce) all in their corresponding NHS ester form. Our results indicated that Rhodamine resulted in very poor fluorescence and DyLight549 was slightly less effective in comparison to Cy3 (data not shown). Alexafluor546 and Cy3 were used for further study. A series of dyes to virus ratios (D/V) were prepared for both Alexa546 and Cy3. Labeling of DM-NA with dye is determined by UV–vis spectroscopy by the presence of the dye peak in purified fractions (Fig. 2A). The results for DM-NA-Alexa546 showed that the highest fluorescence signal was obtained at 57 D/V (Fig. 2B) and as the amount of dye increased, signal output decreased. However, the fluorescence intensity of DM-NA-Alexa546 (200 D/V) was much lower in comparison to wild type (WT) and dye mix (pink line) and free dye (Fig. 2C). Analysis of the series of DM-NA-Alexa546 indicated that the fluorescence signal was inversely proportional to D/V ratio. From the data, it was estimated that the maximum amount of Alexa546 that can be used prior to quenching (negative values in fluorescence enhancement) was 30 D/V (data not shown). The weak performance of DM-NA-Alexa546 prompted further experiments with Cy3 dye, which showed a better outcome. DMNA-Cy3 (209 D/V) was compared to its controls as shown in Fig. 3A and B. In Fig. 3A, the peak at 260 nm corresponds to the virus peak while the peak centered at 550 nm results from Cy3. Comparable amounts of dye and virus is present in all three solutions as shown in Fig. 3A. DM-NA-Cy3 (blue line) fluorescence output was compared to its controls in Fig. 3B and showed signal enhancement. A red shift is observed for the dye peak of the virus-dye complex (blue line) indicating that the dye is coupled to the virus. It is important to mention that signal enhancement is not an effect of the presence of the nanoparticle in solution since the control containing WT virus mixed with free dye (Fig. 3B; pink line) does not show signal enhancement. Signal enhancement is only observed when the dye is coupled to the virus (Fig. 3B; blue line). Further analysis of fluorescence enhancement for the series of DM-NA-Cy3 also indicated a decrease of fluorescence enhancement as D/V increases (Fig. 3C). However, no quenching was observed for DM-NA-Cy3 (positive values in fluorescence enhancement). By extrapolating the line on Fig. 3C, it is not until 256 D/V that fluorescence enhancement is expected to be zero (x intercept; beyond that point negative values are expected, indicative of quenching). The above results indicated that Cy3 gives the best fluorescence enhancement when coupled to the virus, thus DM-NA-Cy3 was chosen for further studies as the staining reagent for Affymetrix resequencing pathogen microarray (RPM) studies. In addition to the D/V ratio, the amount of NA coupled to the viruses was also optimized for possible detection enhancement. Variable amounts of NA, 60, 120, and 180-fold excess of NA to virus, was used, and a biotinylated-rhodamine assay was used to determine the #NA/virus (Soto et al., 2006). The results indicated that 2.2 ± 1 (n = 10), 3.4 ± 2 (n = 13), and 3.3 ± 2 (n = 15) NA per virus was obtained from 60, 120, and 180-fold excess of NA respectively. This data also indicated that adding 180 excess of NA did not result in
For expression array testing, cDNA synthesis was carried out to generate Cy5 or Cy3 dye labeled targets. Briefly, human total RNA (2 g) with added spiked controls (Universal ScoreCard, GE Healthcare Biosciences) were primed with both oligo dT and random nanomers using superscript II reverse transcriptase (Invitrogen) in the presence of reaction buffer, nucleotide mix each at 25 M (except dCTP at 12.5 M) and 1 l Cy3- or Cy5-dCTP (GE Healthcare Biosciences) in a 20 l reaction. The cDNA synthesis reaction was carried out at 42 ◦ C for 2 h followed by sodium hydroxide treatment for RNA removal and purification of labeled fragments using glass bead column (CyScribe GFX Purification kit, GE Healthcare Biosciences). The microarray target preparations for rat 8000 arrays were carried out using above conditions starting from rat total RNA (20 g) with only oligo dT primers. The labeling yield and efficiency of dye incorporation was quantified by spectroscopy at wavelengths of 260, 550, and 650 nm for cDNA, Cy3, and Cy5, respectively, using extinction coefficients of 250,000 M−1 cm−1 , 150,000 M−1 cm−1 for Cy5 and Cy3, respectively. For DM-NA-Cy3 and DM-NA-Cy5 detection, first strand cDNA was synthesized according to above conditions starting from total RNA (with spike controls) in presence of both biotin labeled dCTP and dUTP (Invitrogen) spiked into the nucleotide mix at concentrations of 50 M. The biotin labeled cDNA were purified and used for array hybridization. The efficiency of biotin incorporation (measured by coupling to Cy5-streptavidin) was measured to be comparable to cyanine dye incorporation at an average of one label per 50 unlabeled nucleotides (data not shown). 2.6. Array fabrication and hybridization: GE scorecard system For Universal Scorecard arrays, amplicons (∼100 g/ml) representing appropriate PCR controls and print buffer negative controls were added to an equal volume of DMSO and 20 l transferred to 384 well plates. These plates were used to fabricate arrays on Ultra-GAP glass slides (Corning) using Microarray GenIII spotter (GE Healthcare Biosciences). This printing regime yield spots of approximately 190 m in diameter. Reporters were cross-linked to the surface by UV cross linking. Slides were stored in the dark and inside a desiccator until required. The rat 8K arrays were purchased from UMDNJ, Center for Applied Genomics and contained 8000 genes printed on GAPSII slides (http://www.cag.icph.org/). Array hybridizations were performed using 35 pmoles of Cy5, or Cy3 labeled cDNA or equivalent amount of biotin-cDNA either manually or on automated hybridization station (Lucidea SlidePro, GE Healthcare Biosciences) in microarray hybridization buffer (GE Healthcare) with 40% formamide at 42 ◦ C overnight. Subsequently, washing was performed in 1× SSC (sodium chloride, trisodium citrate, pH 7.0), 0.2% sodium dodecyl sulfate (SDS) for 10 min, 0.1× SSC, 0.2% SDS for 10 min, and 0.1× SSC for 10 min. For DM-NA-dye detection (using 10 g of virus unless noted otherwise), following the first wash, the arrays were exposed to appropriate amounts of virus in 1× SSC, and 50% blocking solution (Perkin Elmer, Waltham, MA) for 30 min followed by washing with 0.1× SSC for 15 min. The arrays were then dried and imaged on GenePix® Pro (Axon Instruments, Molecular Devices, Sunnyvale, CA) at identical PMT settings. Signals intensities from microarray spots were quantified using
3. Results and discussion
Author's personal copy C.M. Soto et al. / Biosensors and Bioelectronics 25 (2009) 48–54
51
Fig. 2. Spectroscopic characterization of DM-NA-Alexa546. (A) UV–vis spectra for a series containing variable number of dyes per virus (D/V) only dye peak is shown for clarity, it is centered at 556 nm. Absorbance for the dye peak was normalized relative to the virus absorbance at 260 nm. (B) Fluorescence spectra for samples in (A). Samples excited at 554 nm. Fluorescence normalized relative to the dye absorbance at 556 nm. Decrease in fluorescence is observed as the number of dyes increases. (C) Fluorescence of virus-dye complex (200 D/V) along with controls in which dye and wild type (WT) virus concentrations are comparable to corresponding amounts on the virus-dye complex. Fluorescence quenching is observed on the virus-dye complex in comparison to the controls.
Fig. 3. Spectroscopic characterization and performance of DM-NA-Cy3. (A) UV–vis spectra showing absorbance of virus and Cy3 in the three solutions. Peaks at 260 and 550 nm correspond to virus and Cy3, respectively. The dye peak in virus-dye complex is red shifted which is typical when the dye is coupled to the protein. The concentration of virus was comparable for DM-NA-Cy3 (209 D/V) and WT-Cy3 mix. The concentration of Cy3 is the same for all three samples. (B) Fluorescence spectra for samples shown in (A) excited at 550 nm, emission is red shifted (blue line) relative to free dye indicating that the dye was coupled to the virus, data was normalized relative to absorbance at 550 nm. (C) Fluorescence enhancement of virus-dye complex as a function of D/V calculated relative to free dye. % enhancement = [(emission of the virus − dye complex at 567 nm) − (emission of free dye at 561 nm)/(emission of the complex at 567 nm)] × 100. Black line is the linear regression of the data (white diamond), resulted line: y = −0.29x + 74.3; r2 = 0.99. Data is an average of three values except for 206 D/V where n = 2. Standard deviations are depicted as error bars. (D). DM-NA-Cy3 performance on an Affymetrix chip relative to SAPE. The virus-dye complex was used as a staining reagent, 60, 120, and 180 in parenthesis indicate the excess of NA added relative to virus during the reaction. Data normalized relative to SAPE intensity, SAPE intensity was set to be a 100% signal. Most consistent data obtained from DM-NA(120)-Cy3 (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article).
Author's personal copy 52
C.M. Soto et al. / Biosensors and Bioelectronics 25 (2009) 48–54
Fig. 4. Comparison of DM-NA-Cy5 with commonly used microarray detection methods. (A) Microarray images, column labels at the top indicate scorecard probe sets. Five consecutive rows are identical replicates. [1] and [2] were hybridized with Cy5-cDNA targets synthesized with Cy5-NHS and Cy5-dCTP respectively. [3] Resulted from hybridization with biotin-cDNA and stained with DM-NA-Cy5-77 D/V. (B) Total signals from scorecard arrays comparing DM-NA-Cy5 with Cy5-cDNA platform. Scorecard control arrays consist of 23 probe sets corresponding to yeast intergenic regions. Calibration controls (Cal 1–10) represent copy number controls with corresponding mRNAs present at high, medium, and low copy number and are spiked into target preparation. NC 1/2 represent negative controls. RC 1–8 are ratio controls for applications in two dye (combined Cy3 and Cy5) assay and UT 1–3 are extra utility controls.
an increase of NA/virus and the saturation was reached after incorporation of three NA/virus for the reaction conditions used. DM-NA(x)-Cy3 samples were used to test the Affymetrix RPM arrays (Fig. 3D). Successful base calling is the important metric as the intention of the RPM array is to allow “resequencing” of a sample. The design of the array is such that the absolute intensity signal from a probe set does not directly affect the performance of base calling as long as the intensity reaches a minimum requirement that can be quite low. It is more important that relative intensities within a probe set be sufficiently different from each other. Total % of base calls from DM-NA-Cy3 was normalized relative to SAPE % base calls. SAPE % of base calls corresponds to a 100% in Fig. 3D. DM-NA-Cy3 signals output are comparable to SAPE. DM-NA(120)-Cy3 gave the most consistent data as shown by smaller standard deviations in comparison to the other two samples and give comparable total % of base calls to SAPE. It should be noted that the Affymetrix scanner excitation source is set at 533 nm which has been optimized for SAPE. When DM-NA-Cy3 is excited at 533 nm the fluorescence emission signal is 29% ± 5, (n = 6) lower in comparison to excitation at 550 nm (optimum max for Cy3). We attribute lower signals from DM-NA-Cy3 in the Affymetrix system partly to this fact. The DM-NA-Cy5 was evaluated in an expression array platform to compare sensitivity against conventional target labeling methods with Cy5-dCTP. Current methods for cDNA target preparation use optimal Cy5 labeling density or dye ratio of 50 (one Cy5-dCTP at every 50 bases). To allow fair comparison between the two detection methods, we optimized our target labeling method to yield biotin incorporation at similar labeling densities. An optimized protocol was developed that used double labeling with biotin-dCTP and biotin-dUTP to incorporate high levels of biotin into cDNA. The dual biotin method achieved incorporation spaced on average 50 bases (data not shown) permitting evaluations of targets that differed only in detection methods but carried identical labeling densities. The signals from both Cy5-dCTP incorporation and DM-NA-Cy5 were compared by hybridization onto scorecard control arrays (GE Healthcare, Piscataway, NJ). The control arrays contain 23 probes from yeast intergenic regions printed onto UltraGAPS microarray slides (Corning, NY). The mRNA corresponding to probes were spiked into total RNA and reverse transcribed in presence of Cy5-dCTP or biotin-dC/U-TP for cDNA target preparation and subsequently hybridized to test for signal amplification (Fig. 4). The arrays hybridized with biotin-cDNA and detected with DM-NA-Cy5 showed bright signals and low background (negative controls, NC 1/2) in comparison to Cy5-dCTP control arrays (Fig. 4A) demonstrating the utility of virus scaffold in expression array detection.
Detection with DM-NA-Cy5 (77 D/V; three NA per virus) yielded the highest signal on expression arrays (see below). The 23 probe sets consist of 10 calibration controls (Cal 1–10) corresponding to different mRNA species present at high, medium, and low copy number that can be used to measure dynamic range and sensitivity of detection. Quantification of array feature intensities demonstrated the total signal from DM-NA-Cy5 to be 35% greater than Cy5 control arrays (hybridized with cDNA labeled with either Cy5-dCTP or Cy5NHS; Fig. 4B). Since high signals were observed from both control and DM-NA-Cy5 arrays, no differences in copy number detection or dynamic range were seen despite the elevated signal from DMNA-Cy5. The signal from DM-NA-Cy5 arrays was observed to be dependent on the dye to virus ratios. Increasing the number of Cy5 dyes from 23 to 77 resulted in near 50% increase in signals on arrays (Fig. 4B). Increasing the Cy5 dye ratio beyond 77 showed a decline in signal presumably from dye quenching effects. As expected, array signal from DM-NA-Cy3 were more tolerant to higher dye coupling than its Cy5 counterpart with optimal labeling density seen at 140 dyes per virus (data not shown). Arrays brightness with DM-NA-Cy5 was also be impacted by viral load or concentration used in array detection. Increasing the virus concentration in detection solution showed a parallel increase in signal from array features. A steady rise in signal was observed with DM-NA-Cy5 from 5 g to 30 g of virus in detection solution (data not shown). Although signal amplification was observed to be more than 220% (or 2.2 fold) from 100 g of DM-NA-Cy5 over conventional Cy5-dCTP labeled targets, a corresponding rise in background was also noticed (data not shown, see below). To Table 1 Post-labeling washes for signal optimization. Dye-reagenta
106 Signal
104 Signal/noisec
Signal amplificationd
Cy5-dCTP control DM-NA-Cy5b No wash 2% SDS 5% SDS 2% Tween 20 2% Triton X-100 2% NP-40 2.5% Urea
2.70 6.74 11.38 5.47 2.60 9.39 8.98 7.36 8.56
4.81 2.43 1.24 6.92 2.22 2.73 0.42 0.19 9.41
1.0 2.5 4.2 2.0 1.0 3.5 3.3 2.7 3.2
a DM-NA-Cy5 samples have 60 dyes/virus, 10 g used except otherwise noted, washes performed are shown in italics. b 5 g used. c Calculated from total signal/detected background. d Total signal improvement relative to Cy5-dCTP control.
Author's personal copy C.M. Soto et al. / Biosensors and Bioelectronics 25 (2009) 48–54
53
Fig. 5. Sections of 8000 rat arrays labeled with (A) Cy5-dCTP and (B) DM-NA-Cy5. (C) Hybridization data from 8000 rat arrays. Separate arrays were hybridized with cDNA made from 20 g of rat total RNA and labeled with Cy5-dCTP or biotin-cDNA. For biotin-cDNA labeling method detection was performed with DM-NA-Cy5.
maximize array sensitivity, experiments were performed to mitigate background using the scorecard arrays. Hybridization and washing conditions were further optimized to remove unbound or non-specific DM-NA-Cy5 from array surface. Several detergents and chemicals were added to array wash steps to reduce background without significant loss in signal (Table 1). As seen in Table 1, different detergents have varying effects in mitigating array surface background. In the absence of any detergents, no improvements were seen in signal-to-noise despite the large signal amplification (compared to Cy5-dCTP control) due to higher array background. When a strong ionic detergent such as SDS (2% and 5%) was tried, improvement in background was observed, although loss of specific signal was also seen, particularly at higher (5%) concentration. The non-ionic detergents such as NP40 and Tween 20 were not effective in our assay. The best candidate for mitigating background without compromising significant signal loss was urea (2.5%). Higher signal and signal-to-noise were seen with urea (2.5%) in wash solution preserving the advantage of near 3-fold signal amplification over Cy5-dCTP control arrays with improvements in signal-tonoise. The optimized labeling and detection method was used to compare sensitivity of DM-NA-Cy5 to Cy5-dCTP on commercial 8000 probe set rat arrays. Rat total RNA (20 g) was used to prepare cDNA targets separately using Cy5-dCTP and biotin nucleotides and then hybridized directly on the 8000 probe rat arrays (Fig. 5). High signals and low background (indicates lack of unspecific binding) were observed from Cy5-dCTP control and DM-NA-Cy5 arrays demonstrating that our optimized protocol could be used to interrogate high density expression arrays (Fig. 5A and B). Quantification of probe set data (Fig. 5C) showed DM-NA-Cy5 was able to detect 14% more genes compared to Cy5-dCTP. This translated to detecting 1120 additional probe sets (or features) on DM-NA-Cy5 array than possible with control Cy5-dCTP. The mean signal from DM-
NA-Cy5 was also about 2-fold higher. The low background on array combined with improvement in signal amplification increased the sensitivity of microarray assay using the DM-NA-Cy5 detection method. The limit of detection (LOD) was calculated to be 30 pg of transcript for both Cy5-dCTP and DM-NA-Cy5. Due to signal amplification, more genes were detected above background from DM-NA-Cy5 starting from 20 g of total RNA. Since LOD values are similar, this means that signal amplification allows us to use less starting amounts of total RNA to detect more genes when using DM-NA-Cy5 as the detection method.
4. Conclusion We were able to couple a large number of dye molecules (>60) to a viral scaffold and still keep the advantage of signal enhancement (Soto et al., 2006). The use of the Lys residues on the virus resulted in more efficient coupling of the dye since the amount of dye required was an order of lower magnitude compared to the amount of dye needed when coupling to the Cys groups. DMNA-dye proved to be useful in commercially available microarray platforms, such as Affymetrix and commercial expression arrays. For the Affymetrix system we obtained signals and sequencing data comparable to a single cycle of SAPE stain. The viral-dye complex showed a near 3-fold signal amplification when used on control scorecard expression after optimizing for amount of applied DMNA-Cy5 and adding urea in the last wash to mitigate background without compromising on signal amplification. Microarray data on commercially available 8000 probe set rat expression array showed advantage of signal amplification in detecting 14% more genes in assay. As new dyes and protein-like recognition elements become available, CPMV offers an excellent scaffold for delivering a large number of reporter molecules at desired recognition sites.
Author's personal copy 54
C.M. Soto et al. / Biosensors and Bioelectronics 25 (2009) 48–54
Acknowledgements We thank Stella H. North and Paul T. Charles for their comments on the manuscript. We also thank Gary J. Vora and Amy S. Blum for useful discussions. We thank Qian Wang and his research group for providing virus samples. This work was supported by the Office of Naval Research. References Blum, A.S., Ratna, B.R., Sapsford, K.E., Vora, G.J., Soto, C.M., 2008. USA Patent Pub No: US 20080220408A1. Borucki, M.K., Reynolds, J., Call, D.R., Ward, T.J., Page, B., Kadushin, J., 2005. J. Clin. Microbiol. 43, 3255–3259. Chatterji, A., Ochoa, W.F., Paine, M., Ratna, B.R., Johnson, J.E., Lin, T.W., 2004. Chem. Biol. 11, 855–863. Contreras-Martel, C., Martinez-Oyanedel, J., Bunster, M., Legrand, P., Piras, C., Vernede, X., Fontecilla-Camp, J.C., 2001. Acta Cryst. D 57, 52–60. Cutler, D.J., Zwick, M.E., Carrasquillo, M.M., Yohn, C.T., Tobin, K.P., Kashuk, C., Mathews, D.J., Shah, N.A., Eichler, E.E., Warrington, J.A., Chakravarti, A., 2001. Genome Res. 11, 1913–1925. Duggan, D.J., Bittner, M., Chen, Y., Meltzer, P., Trent, J.M., 1999. Nat. Genet. 21, 10–14. Greninger, D.A., Pathak, S., Talin, A.A., Dentinger, P.M., 2005. J. Nanosci. Nanotech. 5, 409–415. Hacia, J.G., 1999. Nat. Genet. Suppl. 21, 42–47. Kessler, N., Ferraris, O., Palmer, K., Marsh, W., Steel, A., 2004. J. Clin. Microbiol. 42, 2173–2185. Kronick, M.N., 1986. J. Immunol. Methods 92, 1–13.
Lee, T.M.H., Li, L.L., Hsing, I.M., 2003. Langmuir 19, 4338–4343. Lin, B., Malanoski, A.P., Wang, Z., Blaney, K.M., Ligler, A.G., Rowley, R.K., Hanson, E.H., von Rosenvinge, E., Ligler, F.S., Kusterbeck, A.W., Metzgar, D., Barrozo, C.P., Russell, K.L., Tibbetts, C., Schnur, J.M., Stenger, D.A., 2007. PLoS ONE 2, e419. Lin, B., Wang, Z., Vora, G.J., Thornton, J.A., Schnur, J.M., Thach, D.C., Blaney, K.M., Ligler, A.G., Malanoski, A.P., Santiago, J., Walter, E.A., Agan, B.K., Metzgar, D., Seto, D., Daum, L.T., Kruzelock, R., Rowley, R.K., Hanson, E.H., Tibbetts, C., Stenger, D.A., 2006. Genome Res. 16, 527–535. Mouritsen, C.L., Hillyard, D.R., 1999. Anal. Chem. 71, 366R–372R. Sapsford, K.E., Soto, C.M., Blum, A.S., Chatterji, A., Lin, T.W., Johnson, J.E., Ligler, F.S., Ratna, B.R., 2006. Biosens. Bioelectron. 21, 1668–1673. Sevenet, N., Cussenot, O., 2003. Clin. Exp. Med. 3, 1–3. Soto, C.M., Blum, A.S., Wilson, C.D., Lazorcik, J., Kim, M., Gnade, B., Ratna, B.R., 2004. Electrophoresis 25, 2901–2906. Soto, C.M., Blum, A.S., Vora, G.J., Lebedev, N., Meador, C.E., Won, A.P., Chatterji, A., Johnson, J.E., Ratna, B.R., 2006. J. Am. Chem. Soc. 128, 5184–5189. Soto, C.M., Martin, B.D., Sapsford, K.E., Blum, A.S., Ratna, B.R., 2008. Anal. Chem. 80, 5433–5440. Vora, G.J., Meador, C.E., Stenger, D.A., Andreadis, J.A., 2004. Appl. Environ. Microbiol. 70, 3047–3054. Wang, D., Coscoy, L., Zylberberg, M., Avila, P.C., Boushey, H.A., Ganem, D., DeRisi, J.L., 2002. Proc. Natl. Acad. Sci. U.S.A. 99, 15687–15692. Wang, D., Urisman, A., Liu, Y.T., Springer, M., Ksiazek, T.G., Erdman, D.D., Mardis, E.R., Hickenbotham, M., Magrini, V., Eldred, J., Latreille, J.P., Wilson, R.K., Ganem, D., DeRisi, J.L., 2003. PLoS Biol. 1, E2. Wang, L., Lofton, C., Popp, M., Tan, W., 2007. Bioconjugate Chem. 18, 610–613. Wang, Z., Daum, L.T., Vora, G.J., Metzgar, D., Walter, E.A., Canas, L.C., Malanoski, A.P., Lin, B., Stenger, D.A., 2006. Emerg Infect Dis. 12, 638–646. Zhao, X., Bagwe, R.P., Tan, W., 2004. Adv. Mater. 16, 173–176. Zhao, X., Tapec-Dytioco, R., Tan, W., 2003. J. Am. Chem. Soc. 125, 11474–11475.
