Abstract
NCy3, a derivative of Cyanine 3 with a nitro substituent, showed a high reactivity to bisulfite in aqueous media, instantly leading to ratiometric change of absorption spectra and significant fluorescence quenching. Applied in the microfluidic channel, NCy3 functionalize as a sensitive approach for quantitative detection of bisulfite, particularly for samples with a small volume.
Keywords: Cyanine dye, nitration, bisulfite sensor, microfluidic channel
1. Introduction
preservative and antimicrobial reagent against oxidation and spoilage during food processing and wine fermentation [1], and is also added to dried fruit and potato products for long term storage [2]. Based on the report from the World Health Organization (WHO), the daily intake limit of sulfite is 0.7mg/Kg [3]. High doses of sulfide shows a highly biological toxicity, causing disease and allergic reaction [4]. Recent studies reveal that bisulfite is correlated to lung cancer, cardiovascular disease and many neurological disorders [5]. However, bisulfite also exists naturally in the human body. In living cells, sulfite and bisulfite can be produced in aqueous media at physiological condition by endogenous sulfur dioxide (SO2), which is generated from the oxidation of H2S or sulfur containing amino acids by reactive oxygen species (ROS) [6]. Thus, the cellular level of bisulfite is able to collaterally reflect the local reducing environment in living cells [7]. Therefore, developing an efficient approach for rapid detection of sulfite is highly desired [8].
In past years, the fluorescence chemosensor has attracted a substantial amount of attention as a simple tool used to detect small molecules in vitro and in vivo due to high selectivity, sensitivity and real-time measurement [9]. Copious fluorescence approaches have been developed using different photophysical mechanisms (e.g., photoinduced electron transfer (PET), internal charge transfer (ICT), and (föster resonance energy transfer (FRET)) for quantitative measurement of various bio-important molecules [10]. However, the sensitivity and selectivity are always a challenge for fluorescence sensing approaches, particularly in biosamples [11].
Microfluidic devices offer scientists many benefits compared to larger scale systems. One benefit of using microchannels is the small volume needed to analyze a sample. Depending on the size of the channel and the number of channels, only femtoliters of the solution are required, which is cost-effective and beneficial for experiments with limited samples [12]. Additionally, microfluidics lends well to multiplexing, so a scientist can run multiple tests on a chip. Finally, microfluidics has a faster reaction time, portability, and enhanced sensitivity compared to large-scale systems [13]. Many systems already use microfluidics such as sequencing, detection of metabolites, physical DNA mapping, 3D cell culture, and optofluidics due to the aforementioned benefits [12–15].
In our group, we recently developed a reaction-based fluorescence sensor based on Cyanine 3 dye, which was able to rapidly react with HSO3− based on a nucleophilic addition mechanism. In the presence of HSO3−, NCy3 showed a ratiometric change of absorption spectra and a strong fluorescence quenching within 3 min at 25 °C, indicating a high sensitivity and selectivity. NCy3 was also applied to a microfluidic device to measure the fluorescence intensity of samples with a small volume, in which a similar result was obtained, indicating NCy3 could be used as an approach for quantitative detection of sulfite with a high affinity and selectivity in the aqueous media.
2. Methods
2.1. Apparatus
All reagents used for synthesis and measurements were purchased from Sigma-Aldrich (MO, USA), Fisher Scientific (USA), TCI (USA), Alfa Aesar (USA) and Acros Organics (USA) in analytical grade and were used as received, unless otherwise stated. Absorbance spectra were collected by Cary Series Uv-vis Spectrophotometer (Agilent Technologies). Fluorescence measurements were all performed by using a FluoroMax-4 Spectrofluorometer (Horiba Jobin Yvon, USA). All of fluorescence spectra were recorded in a 1 cm quartz cuvette. The excitation and emission slits were set at 2 nm. 1H and 13C NMR spectra were recorded on (1H 300MHz, 13C 75MHz) Bruker 300 Ultra-Shield spectrometer at room temperature. The HRMS data was collected in the Nebraska Center of Mass Spectrometry at University of Nebraska-Lincoln by using GCT Mass Spectrometer (Water, USA).
2.2. Synthesis
2,3,3-trimethyl-5-nitro-3H-indole (1). 2,3,3-trimethylindole (3.18 g, 0.02 mol) was added into concentrated H2SO4 ( 75 mL) containing HNO3 (1.70 g, 0.02 mol) within 1 hr and was incubated for 0.5 hr at 0 °C [16]. After reaction, the mixture was poured into ice (1000 mL) to collect the precipatate by suction filtration. The solid was recrystallized in CH2Cl2 to yiled an organge ctrystall as the product (3.5 g, 91%). 1H NMR (300 MHz, CDCl3) δ: 1.4 (s, 6H), 2.4 (s, 3H), 7.6 (d, J=8.3 Hz, 1H), 8.2 (s, 1H), 8.3 (d, J=8.2 Hz, 1H). 13C NMR (75 MHz, CDCl3) δ: 16.1, 22.9, 54.6, 117.3, 120.3, 124.8, 145.8, 147.0, 159.3, 194.3.
1,2,3,3-tetramethyl-5-nitro-3H-indol-1-ium (2). A mixture of 1 (0.40 g, 0.0020 mol) and CH3I (0.33g, 0.0024 mol) were heated to 90 °C in CCl4 (4 mL) for 15 hours in a sealed reaction tube (50 mL). After cooling to room temperature, the reaction mixture was filtrated to yield a brown yellow solid as the pure product without any further purification (0.41 g, 95%). 1H-NMR (400 MHz, CDCl3) δ: 1.4 (s, 6H), 3.2 (s, 3H), 4.2 (d, J=10.5 Hz, 2H), 6.5 (d, J=6.7 Hz, 1H), 8.0 (s, 1H), 8.2 (d, J=8.2 Hz, 1H). 13C NMR (75 MHz, DMSO-d6) δ: 29.4, 29.6, 43.5, 80.2, 105.5, 118.5, 126.7, 138.5, 139.6, 152.1, 161.3.
1,3,3-trimethyl-5-nitro-2-((E)-3-((E)-1,3,3-trimethyl-5-nitroindolin-2-ylidene)prop-1-en-1-yl)-5,6,7,7a-tetrahydro-3H-indol-1-ium (NCy3). Compound 2 (0.218 g, 0.001 mol) and triethyl orthoformate (0.593 g, 0.004 mol) were heated in acetic acid/acetic anhydride (1:1, 5mL) at 100 °C for 12 hrs. After cooling to room temperature, the reaction mixture was purified by column chromatography (silica, 220–400 mesh). Ethyl acetate and methanol (4:1) were used as the elution solvents. A dard red solid (NCy3) was obtained as the product (0.170 g, 52%). 1H-NMR (400 MHz, DMSO-D6) δ: 1.8 (s, 12H), 3.7 (s, 6H), 6.7 (d, J=13.5 Hz, 2H), 7.7 (d, J=8.4 Hz, 2H), 7.7 (t, J=5.7 Hz, 2H), 8.3−8.4 (m, 3H), 8.6 (s, 2H). 13C-NMR (75 MHz, DMSO-D6) δ: 27.4, 32.7, 49.6, 106.2, 112.6, 118.8, 126.1, 142.4, 145.0, 148.3, 152.2, 177.3. TOF EI+: M+ m/z 447.2027 (calcd.), 446.5980(found).
2.3. Microfluidic Methods
Master and Replica Fabrication
SU8 microchannel masters were fabricated by photolithography by spincoating SU8 2005 (5.0 μm, MicroChem) onto a silicon wafer and exposing it with UV-light. Replicas of the masters were created using soft lithography. PDMS (Sylgard 184) was mixed (10:1 ratio of pre-polymer to Platinum catalyst) for 15 minutes with a hand held mixer (Kitchen Aid 9 speed mixer), poured onto the silicon master, and placed in a 65°C oven overnight. The PDMS replica was removed from the master and then made hydrophilic with oxygen plasma treatment (Diener Zepto System, 36s O2 plasma, 0.50 mbar pressure, 15% power) and stored in distilled water until ready to use.
Imaging of Dye in Microchannels
The PDMS device was mounted onto coverslips and attached to a holder with candle wax. Dye (1 μL, 1 × 10−3 M) and 3 μl distilled water or 3 μl HSO3− (1 × 10−3 M) was mixed and loaded into microchannels via capillary action. The dynamics of the dye in the presence of water or HSO3− was imaged using a CoolSNAP camera and coupled to a Nikon Eclipse TE2000-S microscope. A Brightline 4040B filter set was used with X-Cite 120 Fluorescence Illumination light source to excite the dye and was captured with the NIS Elements D capture software. ImageJ was used to analyze images to compile the 2D montage and determined the fluorescence intensity for each image.5
3. Results and discussion
The absorption and emission spectra of NCy3 were collected in the different media, in which the maximum absorption and emission were observed in the range of 564–578 nm and 581–601 nm respectively (Table 1). NCy3 showed the highest quantum yield (0.190) in DMSO and the lowest quantum yield (0.108) in ethyl acetate. Based on the Consideration of solubility and photophysical properties, we chose a mixture of DMSO and H2O (5:5, v/v) solution containing 20 mM phosphate buffer at pH 7.4 for NCy3 to detect HSO3−.
Table 1:
The photophysical properties of NCy3 [17].
| Solvent | λab (nm) | λem (nm) | ε (cm−1M−1) | Φ |
|---|---|---|---|---|
| DMSO | 578 | 601 | 27,800 | 0.190 |
| Acetone | 566 | 583 | 32,600 | 0.140 |
| MeOH | 564 | 581 | 15,200 | 0.148 |
| THF | 578 | 595 | 22,900 | 0.156 |
| EtOAc | 576 | 592 | 24,700 | 0.108 |
| CH2CI2 | 567 | 581 | 34,900 | 0.160 |
As a nucleophile, HSO3− may react with NCy3 via a nucleophilic addition reaction, which interrupts the conjugation of NCy3 and consequently change the photophyscial properties. We first investigated the absorption and emission spectra of NCy3 in the presence of HSO3− in the DMSO/buffer solution. With addition of HSO3− to NCy3, the absorption spectra rapidly showed a ratiometric signal, a dramatic decrease at 574 nm and an increase at 455 nm, indicating a significant structure change of NCy3.
In the absorption spectra,a remarkable blue-shift (119 nm) was detected due to the interruption of π-conjugation. The instant color change, red to light yellow, was also observed by the naked eye. The maximum spectrum change was achieved when the concentration of HSO3− reached 30 μM, suggesting the accomplishment of the reaction. In the emission spectra, HSO3− led to a gradual fluorescence quench at 597 nm, and the maximum quench (92%) was observed with the addition of 30 μM HSO3−, which was consistent with the change of absorption spectra. The detection limit (3δ/k) was calculated to be 0.086 nM based fluorescence titration at 597 nm [18].
For the reaction-based sensor, the response time is a critical factor that affects the sensing process. The chemical reaction-based signal significantly improves the sensor selectivity, but limits applications if the reaction rate is too slow. Therefore, the time-dependent spectrum change was investigated after incubation with different amount of HSO3− (5 μM, 10 μM, 20 μM, and 30 μM) to NCy3 at 25 °C in the DMSO/buffer media. A change of absorption at 574 nm and 455 nm were measured as well as the emission at 597 nm. A nucleophilic reaction triggered a spectrum change instantly. After incubating HSO3− (5 μM) with NCy3 at 25°C, both absorption and emission spectra gradually changed over the course of 25 min. However, a high concentration of HSO3− (30 μM) led to a rapid spectra change and the maximum change was observed within 5 min, which indicated a high reaction rate based on our data for our reaction-based sensor.
High affinity for a target molecule is another important factor to evaluate a sensor, and is one of the main challenges for sensor designs. To examine the selectivity of NCy3 towards different molecules, various analytes, including biothiols, reducing reagents, and sulfur-containing ions (30 μM) were incubated with NCy3 (1.0×10−5M) for 5 min in a DMSO/Buffer media at 25 °C. Afterwards, the absorption and emission spectra were measured. As shown in the Figure 3, all of interfering species (HSO4−, S2O32−, C2O42−, cysteine, glutathione, ascorbic acid, and H2S) showed just a slight change (i.e., around 10%) in both absorption and the emission spectra. However, during incubation with HSO3−, NCy3 rapidly displayed a ratiometric signal at 455 nm/575 nm in the absorption spectra as well as a significant quenching at 597 nm in the fluorescence spectra, demonstrating that NCy3 is very sensitive to HSO3−, allowing NCy3 to be used as a approach with multiple spectroscopic signal channels for rapid detection of HSO3−. The interruption of the conjugation of NCy3 trigged by HSO3− led to an obvious color change from light pink to yellow, which can be easily observed by naked eye, making the detection process very convenient.
Figure 3:
The absorption responses (A) and fluorescence emission responses (B) of NCy3 (1.0×10−5M) toward HSO4−, S2O32−, C2O42−, Cys, GSH, ascorbic acid, H2S and HSO3− (30 μM) after Incubation for 5 min in DMSO/pH 7.4 buffer (1:1) at 25 °C (λex = 574 nm).
As a nucleophile, H2S did not show a high reaction rate with NCy3 at low concentration. However, with increasing concentration and incubation time, NCy3 gave a significant response to H2S. As shown in the Figure 4, both the absorption and emission spectra steadily changed within 45 min at 25 °C when the H2S level was less than 100 μM, indicating a slow reaction process. When increasing the level of H2S to 150 μM, the spectra displayed a relatively quick change, and plateaued after a 30-min incubation with NCy3, suggesting a complete reaction. Based on the previous data, NCy3 showed a potential to detect H2S with a high concentration. Besides H2S, other species (HSO4−, S2O3−, C2O4−, Cys, GSH, and ascorbic acid) did not cause significant spectroscopic change even at high concentration and/or longer incubation time with NCy3.
Figure 4:
The spectra change of NCy3 (1.0×10−5M) incubating with different amount of H2S (10μM, 50μM, 100μM, and 150μM ) in DMSO/pH 7.4 buffer (1:1) at 25 °C in 45 min: (A) the absorption change at 574 nm, (B) the absorption change at 455 nm, and (C) the fluorescence change at 594 nm (λex=574 nm).
According to previous work, the nitro group on the aromatic ring (1,8-naphthalimide) can be reduced by H2S and consequently lead to photophysical property change, which has been intensively used as a sensing strategy for the detection of H2S [19]. Thus, NCy3 was incubated with different reducing reagents (H2S, GSH, ascorbic acid, and HSO3−) to investigate the reduction of nitro group. However, no reduction reaction was observed in the presence of any these reducing reagents. Only HSO3− and H2S led to a nucleophilic addition reaction, and HSO3− gave a higher reaction rate (Figure 5A). Moreover, the pH effect in the range of pH 3–11 was examined for NCy3 in different DMSO/buffer media. At low pH range (3.0–8.0), NCy3 showed a strong fluorescence emission at 597 nm. With an increase of pH, the fluorescence was quenched rapidly. Although the fluorescence intensity of NCy3 was significantly influenced by pH value, no notable change was observed in the physiological pH range (pH 7.0–8.0) (Figure 5B).
Figure 5:
(A) NCy3 reacted with H2S and HSO3− based on a nucleophilic addition mechanism. (B) The fluorescence intensity change of NCy3 at 594 nm in different media with a pH 3.0−11.0 at 25 °C in (λex=574 nm).
To understand the nucleophilic addition reaction between NCy3 and nucleophiles (e.g., HSO3− and H2S), the Cyanine 3 dye (Cy3) without nitro groups was synthesized for a comparison. Compared to NCy3, Cy3 showed a shorter absorption and emission at 546 nm and 565 nm respectively in the DMSO/pH 7.4 buffer (1:1) at 25 °C. Since NCy3 showed a significant spectroscopic change in the presence of HSO3− and H2S in high concentration, the same condition was applied to Cy3. However, neither absorption nor emission displayed a notable change. Moreover, high concentration (150 μM) of analytes (HSO4−, S2O32−,C2O42−, Cys, GSH, ascorbic acid, H2S, and HSO3−) along with longer incubation time (30 min) with Cy3 could not produce any significant spectroscopic change, revealing that no nucleophilic addition reaction occurred for Cy3 (Figure 6). The nucleophilic addition reaction is well established for α,β-unsaturated compounds. As an electron-withdrawing group, the nitro group on NCy3 played a critical role in significantly boosting the addition reaction. Moreover, the HRMS data indicated that a reduction reaction followed the nucleophilic addition reaction in the present of HSO3−, which thoroughly explained the significant spectra change for absorption and emission.
Figure 6:
After incubation Cy3 (1.0×10−5M) with HSO4−, S2O32−, C2O42−, Cys, GSH, ascorbic acid, H2S and HSO3− (150 μM) for 30 min in DMSO/pH 7.4 buffer (1:1) at 25 °C (λex = 574 nm), (A) no nucleophilic addition reaction observed; (B) absorption spectra and (C) fluorescence spectra did not show a significant change
Because NCy3 was ultra-sensitive, ratiometric, instant, and multiple channel signal sensing abilities, NCy3 was applied to develop a microfluidic device for analysis of HSO3− in very small volume samples. A PDMS device was mounted onto a coverslip and attached to a 3D printed PLA holder with candle wax [20]. Then, dye (1 μl, 1.0×10−3 M) and 3 μl distilled water or 3 μl HSO3− (1.0×10−3 M) was mixed and loaded into microchannels via capillary action. The dynamics of the dye in the presence of water or HSO3− was imaged using a CoolSNAP camera and coupled to a Nikon Eclipse TE2000-S microscope. As shown in Figure 7, with the incubation of HSO3− with NCy3 in the microchannel, a fluorescence quenching was observed within 10 min, indicating that NCy3 was able to produce a reliable signal in a microfluidic device for detection of HSO3−. As a negative control, addition of H2O did not led to a significant change in fluorescence.
Figure 7:
(A) Intensity of the dye with water (red circles) or HSO3− (black squares) in 55 μm wide × 5.0 μm high microchannels was normalized and plotted against time. (B) and (C) Dynamics of the dye in the presence of water (B) or HSO3− (C) was imaged in microchannels for 10 minutes and compiled into a movie. A 2D slice was taken from each image at the same location and compiled to create the 2D montage (1 min per slice). The distance (x) of the slices of the image was in μm.
4. Conclusions
In summary, a fluorescent probe (NCy3) was synthesized for detection of HSO3− via a nucleophilic addition reaction in aqueous media. NCy3 demonstrated a rapid ratiometric sensing for HSO3− with high selectivity and sensitivity. This probe provided multiple signals, including absorption spectra, emission spectra, and colormetric signal, to quantitatively measure HSO3−. In addition, NCy3 was successfully employed in developing a microfluidic device to measure trace amounts of HSO3− in samples with a small volume, which could be very useful for biosamples or environmental monitoring.
Figure 1:
The spectra change of NCy3 (1.0×10−5 M) with addition of HSO3− (0 − 65 μM) in DMSO/pH 7.4 buffer (1:1) at 25 °C. (A) The absorption decreased at 574 nm and increased at 455 nm; (B) The fluorescence emission at 594 nm was gradually quenched (λex = 574 nm).
Figure 2:
The time-dependent spectrum change: (A) absorption at 574 nm, (B) absorption at 455 nm, and (C) emission at 594 nm of NCy3 (1.0×10−5M) with different amount of HSO3− (5 μM, 10 μM, 20 μM, and 30 μM) in DMSO/pH 7.4 buffer (1:1) at 25 °C.
Acknowledgments
Thanks Dr. Brandon Luedtke for help with the Nikon microscope. Financial support was provided by grants from the National Center for Research Resources (NCRR; 5P20RR016469) and the National Institute for General Medical Science (NIGMS; INBRE-8P20GM103427), a component of the National Institutes of Health (NIH) and Nebraska Research Initiative for the purchase of the Nikon microscope.
Footnotes
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of a an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
References
- 1.Pizzoferrato L, Di Lullo G, Quattrucci E (1998) Determination of free, bound and total sulphites in foods by indirect photometry-HPLC. Food Chem, 63: 275–279. [Google Scholar]
- 2.Verma SK, Deb MK (2007) Single-Drop and Nanogram Level Determination of Sulfite (SO32−) in Alcoholic and Nonalcoholic Beverage Samples Based on Diffuse Reflectance Fourier Transform Infrared Spectroscopic (DRS-FTIR) Analysis on KBr Matrix. J. Agric. Food Chem, 55: 8319–8324. [DOI] [PubMed] [Google Scholar]
- 3.Santos-Figueroa LE, Giménez C, Agostini A, Aznar E, Marcos MD, Sancenón F, Martínze-Máñez R (2013) Selective and Sensitive Chromofluorogenic Detection of the Sulfite Anion in Water Using Hydrophobic Hybrid Organic–Inorganic Silica Nanoparticles. Angew. Chem. Int. Ed, 52: 13712–13716. [DOI] [PubMed] [Google Scholar]
- 4.Li J, Li R, Meng Z (2010) Sulfur dioxide upregulates the aortic nitric oxide pathway in rats. Eur. J. Pharmacol, 645: 143–150. [DOI] [PubMed] [Google Scholar]
- 5.Migliore L, Copperdè F (2009)Environmental-induced oxidative stress in neurodegenerative disorders and aging. Mutat. Res, 674:73–84. [DOI] [PubMed] [Google Scholar]
- 6.Liu Y, Li K, Wu MY, Liu YH, Xie YM, Yu XQ (2015) A mitochondria-targeted colorimetric and ratiometric fluorescent probe for biological SO2 derivatives in living cells. Chem. Commun, 51: 10236–10239. [DOI] [PubMed] [Google Scholar]
- 7.Xu W, Teoh CL, Peng J, Su D, Yuan L, Chang YT (2015) A mitochondria-targeted ratiometric fluorescent probe to monitor endogenously generated sulfur dioxide derivatives in living cells. Biomaterials, 56: 1–9. [DOI] [PubMed] [Google Scholar]
- 8.Samanta S, Halder S, Dey P, Manna U, Ramesh A, Das G(2018) A ratiometric fluorogenic probe for the real-time detection of SO32− in aqueous medium: application in a cellulose paper based device and potential to sense SO32− in mitochondria. Analyst, 143: 250–257. [DOI] [PubMed] [Google Scholar]
- 9.Demchenko LP (2009) Introduction to fluorescence sensing, Springer. [Google Scholar]
- 10.Woo J, Kim G, Qintero K, Hanrahan MP, Palencia H, Cao H (2014) Investigation of desilylation in the recognition mechanism to fluoride by a 1,8-naphthalimide derivative. Org. Biomol. Chem, 12: 8275–8279. [DOI] [PubMed] [Google Scholar]
- 11.Lin VS, Lippert AR, Chang CJ (2013) Cell-trappable fluorescent probes for endogenous hydrogen sulfide signaling and imaging H2O2-dependent H2S production. Proc. Natl. Acad. Sci. U.S.A, 110: 7131–7135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lorena C, Palenzuela M, Pumera M (2018) (Bio)Analytical chemistry enabled by 3D printing: Sensors and biosensors. TrAC Trends Anal. Chem, 103: 110–118. [Google Scholar]
- 13.Streets AM, Zhang X, Cao C, Pang Y, Wu X, Xiong L, Yang L, Fu Y, Zhao L, Tang F, Huang Y (2014) Microfluidic single-cell whole-transcriptome sequencing. Proc. Natl. Acad. Sci. U.S.A, 111: 7048–7053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gupta A, Kounovsky-Shafer KL, Ravindran P, Schwartz DC (2016) Optical mapping and nanocoding approaches to whole-genome analysis. Microfluid Nanofluid, 20: 44. [Google Scholar]
- 15.Kieninger J, Weltin A, Flamm H, Urban GA (2018) Microsensor systems for cell metabolism – from 2D culture to organ-on-chip. Lab Chip, 18: 1274–1291. [DOI] [PubMed] [Google Scholar]
- 16.Noland WE, Smith LR, Rush KR (1965) Nitration of Indoles. III. Polynitration of 2-Alky lindoles. J. Org. Chem, 30: 3457–3469. [Google Scholar]
- 17.Eaton DF (1988) Reference materials for fluorescence measurement. Pure Appl. Chem, 60: 1107–1114. [Google Scholar]
- 18.Gabriels R (1970) General method for calculating the detection limit in chemical analysis. Anal. Chem, 42: 1439–1440. [Google Scholar]
- 19.Kim G, Jang E, Page AM, Ding T, Carlson KA, Cao H (2016) Investigation of a sensing approach based on a rapid reduction of azide to selectively measure bioavailability of H2S. RSC Adv, 6: 95920–95924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kounovsky-Shafer K, Hernández-Ortiz JP, Jo K, Odijk T, de Pablo JJ, Schwartz DC (2013) Presentation of large DNA molecules for analysis as nanoconfined dumbbells. Macromolecules, 46: 8356–8368. [DOI] [PMC free article] [PubMed] [Google Scholar]