Homologous miRNA analyses using a combinatorial nanosensor

Subscriber access provided by Kaohsiung Medical University

Homologous miRNA analyses using a combinatorial nanosensor array with two-dimensional nanoparticles Mustafa Salih Hizir, Nidhi Nandu, and Mehmet Veysel Yigit Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b01083 • Publication Date (Web): 20 Apr 2018 Downloaded from http://pubs.acs.org on April 20, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Homologous miRNA analyses using a combinatorial nanosensor array with two-dimensional nanoparticles Mustafa Salih Hizir*, †, Nidhi Nandu†, and Mehmet V. Yigit*,†, ‡

† Department of Chemistry, University at Albany, State University of New York, 1400 Washington Avenue, Albany, New York 12222, United States.

‡ The RNA Institute, University at Albany, State University of New York, 1400 Washington Avenue, Albany, New York 12222, United States.

*Correspondence: Tel: (1) 518-442-3002 [email protected] and [email protected]

Keywords: artificial nanoreceptors, transition metal dichalcogenides, sensor, molybdenum disulfide (MoS2), tungsten disulfide (WS2), graphene-oxide

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

Abstract A novel combinatorial nanosensor array for miRNA analyses was assembled using the intrinsic non-covalent interactions of unmodified two-dimensional nanoparticles. Discrimination of nine miRNA analogs with as little as a single nucleotide difference was demonstrated under four hours. All nine targets were identified simultaneously with 95% confidence. The developed nanotechnology offered identification and quantification of unknown targets with unknown concentration. Discrimination of target mixtures from low-to-high ratios was demonstrated. The DNA and RNA analogs of targets were identified using the combinatorial sensory approach. Identification of a target in a complex biological matrix prepared with human urine was demonstrated. The nanosensor array was put together using 15 nanoassemblies (2D-NAs) constructed using three two-dimensional nanoparticles (2D-nps: WS2, MoS2 and nano-graphene oxide- nGO) and five rationally designed fluorescently labelled 15-nt-long ssDNAs (probes). In this approach, each target has only a small yet varying degree of complementarity with each of the five probes adsorbed on the 2D-np surface. The probes in each 2D-NA are desorbed from the surface by each target with a different degree which was recorded with fluorescence recovery measurements. The fluorescence data set was processed by Partial Least Squares Discriminant Analysis (PLSDA) and each target was discriminated successfully. This new approach has a number of advantages over the classical bind-and-release model, typically used for 2D-np based biosensors, and opens greater detection opportunities with 2D-nps.


Page 3 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction Nanotechnology has been instrumental in addressing a variety of biomedical challenges.1-7 Two-dimensional nanoparticles (2D-nps) are a new class of nanoparticles with one- or few-atom thickness and remarkable surface area.8-12 Among many 2D-nps, nGO (nano-graphene oxide), MoS2 (molybdenum disulfide) and WS2 (tungsten disulfide) were shown to be highly instrumental in biosensor development.13-19 These three 2D-nps have two unique properties exploited for nucleic acid based fluorescence detection mechanisms; (1) the 2D-nps have high ssDNA loading capacity due to their large surface area and noncovalent interactions with ssDNA constituents. (2) The fluorescence of a molecular fluorophore label is quenched drastically upon adsorption of a ssDNA to the 2D-np surface. These two properties were used to assemble the functional two-dimensional nanoassemblies (2D-NAs) from 2D-nps and ssDNAs. The ssDNA probes in the 2D-NAs respond to the introduction of a target and desorb from the surface. The desorption occurs either by the target specific bind-and-release model or nonspecific displacement by the target, Scheme 1.20,21 Upon departure, the quenched fluorescence is recovered which was quantified for target analyses. We have used the bind-and-release model for detection of specific miRNAs, proteins and small metabolites.22-26 In this model, the target binds to the surface adsorbed probe, forms a target-probe complex, releases from the surface and a fluorescence recovery is observed. For this approach, fluorescently labelled aptamers or anti-sense DNAs are utilized as surface adsorbed probes. On the other hand, the displacement model was used for non-specific yet highly selective identification of cells, proteins or phase transition of a macromolecule.20 In this approach, the target is identified by the sum of its noncovalent interactions; major and minor; with the 2D-np surface. First, a number (12 or more) of 2D-NAs were constructed using ssDNA probes with no specificity to the targets. The 2D-NAs were put in an array format and each target is tested with every 2D-NA in the sensor array. In this approach, the target outcompetes the surface adsorbed ssDNA probe, disrupts its existing noncovalent interactions with the 2D-np surface and displaces the ssDNA probe. Quantification of the ssDNA probe displacement and analysis with Partial Least Squares Discriminant Analysis (PLSDA) enables identification of the most challenging biosystems.20 PLSDA is a statistical discriminant analysis method which offers discrimination of groups of events/subjects by analyzing their distinct responses and reports in two- or three-dimensional plots. Here, an advanced semi-specific approach is demonstrated, where nine 22-nt-long targets (miRNA analogs) with a very high degree of homology were discriminated simultaneously under four hours. MiRNAs are endogenously expressed small noncoding RNA molecules found in solid tissues and bodily fluids.27 The expression of certain miRNAs is dysregulated in several human disorders, thus they are considered



Page 4 of 26

as next-generation disease biomarkers.28-30 To grant miRNAs real diagnostic value, practical detection methodologies and technologies are required. However, miRNA detection; particularly discrimination of miRNAs; is often a challenging task.31-34 Because miRNAs are short oligonucleotides and there are thousands of miRNAs, a high degree of homology can be observed between miRNAs with completely different biological functions. Thus, accurate identification of miRNAs is important for correct diagnostics and miRNA analyses. Current detection methods have challenges with identification of single nucleotide polymorphisms (SNPs).31-33 Discriminating the miRNAs with one, two or few SNPs is important for miRNA research and further advancements in RNA-based diagnostics. Here, we have developed a universal sensor array approach for the discrimination of nine members of the let-7 miRNA family. The let-7 miRNAs play major roles in gene regulation, cell adhesion and muscle formation.35 However, dysregulated let-7 miRNAs are also associated with numerous cancer types, i.e., colon, lung and breast cancer.36,37 The let-7 miRNA family has 12 mature members and some members are highly similar in sequence. This homology makes the accurate identification of each let-7 miRNA a challenging task. Considering that each miRNA has a unique biological function, the correct identification of each one is essential for accurate analysis of a miRNA profile of a tissue or living system. The remarkable similarity within some of the let-7 members makes them an attractive model group for miRNA discrimination studies. Here, we developed a nanotechnology that offered the discrimination of nine let-7 miRNA analogs with as little as a single nucleotide polymorphism (SNP). Using two-dimensional nanoparticles, a combinatorial nanosensor array was assembled for let-7 miRNA analyses. 2. Materials and Methods All oligonucleotides were purchased from Integrated DNA Technologies (IDT), USA with the following sequence information, (P1) FAM-labeled probe 1 DNA, 5’-/56-FAM/AACTATACAACCTAC-3’ (P2) FAM-labeled probe 2 DNA, 5’-/56-FAM/AACTGTACAAACAAC-3’ (P3) FAM-labeled probe 3 DNA, 5’-/56-FAM/AACCACACAACCTAC-3’ (P4) FAM-labeled probe 4 DNA, 5’-/56-FAM/AACAGCACAAACTAC-3’ (P5) FAM-labeled probe 5 DNA, 5’-/56-FAM/AACTATACAACCTCC-3’ let-7a DNA, 5’-TGAGGTAGTAGGTTGTATAGTT-3’


Page 5 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


let-7b DNA, 5’-TGAGGTAGTAGGTTGTGTGGTT-3’ let-7c DNA, 5’-TGAGGTAGTAGGTTGTATGGTT-3’ let-7d DNA, 5’-AGAGGTAGTAGGTTGCATAGTT-3’ let-7e DNA, 5’-TGAGGTAGGAGGTTGTATAGTT-3’ let-7f DNA, 5’-TGAGGTAGTAGATTGTATAGTT-3’ let-7g DNA, 5’-TGAGGTAGTAGTTTGTACAGTT-3’ let-7i DNA, 5’-TGAGGTAGTAGTTTGTGCTGTT-3’ let-7j DNA, 5’-TGAGGTAGTTGTTTGTACAGTT-3’ let-7a RNA, 5’-UGAGGUAGUAGGUUGUAUAGUU-3’ let-7b RNA, 5’-UGAGGUAGUAGGUUGUGUGGUU-3’ let-7c RNA, 5’-UGAGGUAGUAGGUUGUAUGGUU-3’ Carboxyl graphene water dispersion was purchased from ACS Material, Medford, MA 02155, USA and sonicated 12 hours before use, which resulted in highly stable nanosized graphene oxide (nGO) particles in water.22,23,25,26 Sodium cholate hydrate was purchased from Alfa Aesar, Tewksbury, MA 01876, USA. Raw molybdenum(IV) sulfide, tungsten(IV) sulfide and all other reagents were purchased from Sigma-Aldrich, St. Louis, MO 63103, USA and used without further purification. RNase-free water was used in preparation of miRNA solutions. Double distilled water was used in preparation of all other solutions. Urine specimen was purchased from Discovery Life Sciences (Los Osos, CA, USA). 2.1. Synthesis of WS2 and MoS2 2D-nps. The 2D-np preparation from WS2 and MoS2 TMDs was adapted from our previous report.20 Briefly, solid MoS2 or WS2 precursor was mixed with sodium cholate in 5:1 (w/w) ratio in 500 mL of water and sonicated for 20 hours using digital probe sonifier, Branson Ultrasonics SLPe 150 W (pulse-on for 6 s and pulse-off for 3 s) at 4˚C in cold room. The resultant black dispersion was centrifuged at 3000 rpm for 30 min. The pellet at the bottom of the tubes contained multilayered large TMDs which was separated and stored for future use. The yellow-green supernatant; containing mixture of 2D-nps and sodium cholate; was recovered and subjected to a second centrifugation at 12000 rpm for 30 min. The 2D-nps form a uniform pellet at the bottom of the tube and separated from supernatant for washing steps. The supernatant was discarded and the pellet was redispersed in 200 mL of water and centrifuged at 12000 rpm for another 30 min.



Page 6 of 26

After washing steps, the resulting 2D-np pellet was dispersed in 200 mL of deionized water for the preparation of the aqueous 2D-np stock solutions. Hydrodynamic size of the particles was measured using dynamic light scattering (DLS), DynaPro Titan, Wyatt Technology Corporation, USA. The absorbance spectra were obtained and recorded using Agilent Technologies Cary 60 UV-Visible spectrophotometer. 2.2. The construction of two-dimensional nanoassemblies (2D-NAs) and sensory array assembly. The 2DNAs were assembled by adsorption of the 20 nM FAM-labeled ssDNA probes (P1, P2, P3, P4 or P5) on the 2Dnps (nGO, MoS2 and WS2) through 30 min incubation. For nGO-based 2D-NAs; 2.5 µg mL-1 nGO in a buffer containing100 mM phosphate buffer (pH 7.2), 150 mM NaCl and 1 mM MgCl2 was used. For MoS2-based 2DNAs, 15 µg mL-1 MoS2 in a buffer with 25 mM HEPES buffer (pH 7.5), 100 mM NaCl and 10 mM MgCl2 was used. For WS2-based 2D-NAs, 15 µg mL-1 of WS2 in 25 mM HEPES buffer (pH 7.5), 100 mM NaCl and 10 mM MgCl2 was used. Upon assembly, the fluorescence of the probes was quenched ~95%. In order to assemble the sensor array, 100 µL of each 2D-NA was placed in a 96 well-plate in six replicates; 15 2D-NAs x 6 replicates = 90 wells. 2.3. The detection of let-7 miRNA analogs using the combinatorial array. Each target strand (let-7a, let-7b, let-7c, let-7d, let-7e, let-7f, let-7g, let-7i and let-7j) was tested with the sensor array individually. Briefly, 50 nM final concentration of each target was added to the wells of the sensor array. The fluorescence measurements were performed over 2 hr with 2 min intervals using BioTek SynergyTM H1 microplate reader (Ex: 485 nm and Em: 518 nm). At the end of 2-hr of incubation, the fluorescence recoveries were calculated for each target molecule using the difference between the final fluorescence (ƒ) and initial fluorescence (ƒ0) numbers (∆ƒ = ƒ ƒ0). The fluorescence recovery values were used to construct a fluorescence data matrix (Figure S5) and response patterns for each target. The fluorescence response patterns were processed using PLS software. The resulting PLSDA plot was able to discriminate each target successfully. 2.4. Calibration plot of unknown target analysis. The concentration of each target strand was brought to a value (~50 nM) where their absorbance at 260 nm was equivalent to 0.01-a.u. (Abs260 = 0.01 a.u.). First, 2 µM of each target was prepared in ultra-pure RNAse-free water and their optical densities were measured with absorbance spectroscopy. The Abs260 values for 2 µM of target strands were determined to be; let-7a: 0.484, let7b: 0.460, let-7c: 0.455, let-7d: 0.460, let-7e: 0.484, let-7f: 0.483, let-7g: 0.471, let-7i: 0.452, let-7j: 0.447 a.u. Next, each target was diluted to bring their Abs260 value equivalent to 0.01 a.u. The corresponding concentration of each target was calculated to be; 45.3 nM let-7a: 0.009, 48.8 nM let-7b: 0.011, 46.2 nM let-7c: 0.009, 44.4 nM let-7d: 0.011, 46.2 nM let-7e: 0.012, 43.7 nM let-7f: 0.013, 46.2 nM let-7g: 0.013, 45.9 nM let-7i: 0.011,


Page 7 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


45.8 nM let-7j: 0.011 a.u. The extinction coefficients (ε) of the targets are 226,500; 219,300; 226,600; 229,200; 227,100; 229,000; 222,000; 212,600; 216,100 L.mol-1cm-1, respectively. All nine targets were tested with the sensor array. Following the 2-hr long incubation, ∆ƒ values were calculated to produce the training matrix (Figure S6) and fluorescence response patterns. The data were analyzed, and the targets were discriminated successfully using the resulting PLSDA plot. Later, the PLSDA plot was used as the calibration for the unknown analysis. 2.5. Unknown target identification. The unknown tests were performed in a double-blind manner where neither the person handling the unknowns nor the one performing the tests knew the identity of the unknowns. 24 unknown samples were prepared from the nine target strands with unknown concentrations. First, the absorbance of each unknown was measured using a UV-visible spectrometer. Each unknown was diluted until its Abs260 value was equivalent to 0.01 a.u. The dilution factor was recorded for back-calculation of the unknown concentration. Each diluted sample was tested with the sensor array. The fluorescence recovery (∆ƒ) data were obtained after the 2-hr long incubation. The data matrix was prepared for 24 unknowns and processed with Partial Least Squares (PLS) prediction function using the aforementioned calibration plot as the reference. The predictions were made by the software and identity of each unknown was provided. Later, the original concentration of each unknown was back-calculated using the recorded dilution factors. 2.6. Identification of target mixtures. The sensor array was tested against mixtures of let-7e and let-7i in 3:1, 1:1 or 1:3 ratios with the total target concentration of 50 nM. At the end of 2-hr kinetics study, ∆ƒ values were calculated and used to assemble the training matrix (Figure S7), create the fluorescence response patterns and construct the PLSDA plot. 2.7. Identification of DNA and RNA analogs of let-7a, let-7b and let-7c. The sensor array was challenged against 50 nM of RNA and DNA analogs of let-7a, let-7b and let-7c. Fluorescence recovery was monitored over 2 hrs. ∆ƒ values were used to construct the data matrix (Figure S8), fluorescence response patterns and the PLSDA plot. 2.8. Identification of let-7b RNA in human body fluid. The sensor array was actuated to discriminate 50 nM let-7b RNA in 3% urine. Fluorescence recoveries from blank urine media and urine with 50 nM let-7b were recorded through 2 hrs. ∆ƒ values were used to produce the data matrix (Figure S9), fluorescence response patterns and the PLSDA plot.



Page 8 of 26

2.9. Statistical analysis. Partial Least Squares Discriminant Analysis (PLSDA) in PLS Toolbox, SOLO (version 8.1), was used for processing the fluorescence response patterns, producing training matrices, constructing the PLSDA plots and making unknown predictions. Experiments were performed in six replicates. Rest of the data were expressed as mean ± standard deviation. 3. Results and Discussion Prior to the detection studies, 2D-nps were prepared from raw graphene oxide, MoS2 and WS2. The characteristic absorbance bands of each 2D-np were measured and recorded using absorbance spectra, Figure S1a. The hydrodynamic radius (~100 nm) of each 2D-np was measured using dynamic light scattering, Figure S1b. The distinct physiochemical properties of each 2D-np were utilized to construct unique 2D-NAs for the sensing studies. The 2D-NAs were prepared by the adsorption of a fluorescently (FAM) labeled ssDNA probe on a 2D-np surface. Each 2D-NA has different combinations and types of intermolecular forces (van der Waals forces, H-bonding, π-π stacking, electrostatic interactions and etc.) of various degrees and ratios which hold them together. Thus, each 2D-NA has a different stability (response) which can be fine-tuned by variations in its ssDNA building blocks or 2D-np type. For instance, we have demonstrated that the greatest adsorption/desorption of the ssDNA probes was observed using nGO followed by MoS2 and WS2, respectively.20 The desorption studies were performed by recording the recovery of quenched fluorescence upon addition of a target molecule, Figure S2. Studies demonstrate that the recovery of quenched fluorescence with nGO 2D-NAs is greater than the recoveries obtained with MoS2 2D-NAs, which were greater than that of WS2 2D-NAs. Our results along with reports in the literature indicate that WS2 is the most resistant 2D-np to the desorption of the surface-adsorbed ssDNAs whereas nGO is the least (2D-NAs’ resistance; WS2 > MoS2 > nGO), Figure S3-4.20,38 The results suggest that for an individual sensor construct, nGO is preferred over MoS2 and WS2 due to its greater desorption degree. Regardless, the difference in the desorption degrees is instrumental in the development of the nanosensor array for miRNA analysis reported in this study. In addition to the 2D-np type, the size and the composition of the ssDNA probes also influence the response of the 2D-NAs to the external target.38 However, for assembling a sensor array, as demonstrated in this report, the data collected from all 2D-nps would be useful due to their differential response to the same target. More specifically, though all three 2D-nps have different ssDNA adsorption/desorption and fluorescence quenching properties, the rate/degree of each process was observed to be very different for each 2D-np, Figure S2-4.20,38 Thus, each 2D-np is distinct and forms a unique 2D-NA with ssDNA probes. This diversity enables the assembly of the nanosensor arrays containing several 2D-NAs with distinct response features.


Page 9 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


First, five 15-nt-long fluorescently labeled ssDNA probes were designed against the nine target strands, Figure 1. Instead of using totally non-complementary strands for the 2D-NAs, for non-specific target identification, we introduced some degree of specificity by inserting complementarities ranging from 46 to 68%, Table 1. We anticipated that relying only on the non-specific interactions for discriminating the targets with such a high degree of homology would be challenging and adding varying small complementarities between targets and probes could maximize the discriminative capabilities. The discrimination mechanism of each 2D-NA relies on the hybridization degree of each miRNA target with each ssDNA probe, Table 1. Thus, we assembled 15 2DNAs using this new semi-specific approach. First, the five ssDNA probes were assembled with the three 2D-nps (nGO, MoS2 and WS2) to form 15 2D-NAs which were placed in a well plate in six replicates to form a nanosensor array, Figure 2a-c. The 15 2D-NAs were assembled by incubating fluorescently labeled ssDNA probes (P1, P2, P3, P4 or P5) on the 2D-nps (nGO, MoS2 and WS2) in an aqueous buffered environment through 30 min incubation. The adsorption relies on the binding of the ssDNA probes to the 2D-np surface through noncovalent interactions (van der Waals forces, Hbonding, π-π stacking, electrostatic interactions and etc.). The sensor array was tested for 50 nM of each target strand and the fluorescence recovery was monitored for each well in the array over 2 hrs, Figure 2d. Studies demonstrate that for an individual nanosensor in the array, upon addition of a target, the quenched fluorescence of the ssDNA probes is recovered, which can be monitored using fluorescence spectra or a kinetic study measuring the fluorescence at 520 nm (i.e., response of 2D-NA-2 [nGO-p2] against let-7j), Figure 3. For discrimination of nine let-7 targets, the fluorescence recovery (∆ƒ = ƒ – ƒ0, Figure S5) was calculated for each target using all 15 2D-NAs (the nanosensor array) and fluorescence response patterns were constructed, Figure S3. The final data set was obtained by measuring the response of each 2D-NA against each target and was processed with PLSDA to visualize target discrimination. The readings were recorded for each measurement and each target was differentiated successfully with the new approach, Figure 2e. As seen in the discrimination plot, every single let-7 target is discriminated from each other with 95% confidence and without any overlap. Later, this nanotechnology was tested for analyses of unknown targets with unknown concentrations. Since each target has exactly the same number of nucleobases with almost identical ratios, their extinction coefficients are very similar; ε260~220,000 (L.mol-1.cm-1). Therefore, we have used a 0.01-a.u. absorbance value (equivalent to ~50 nM for each target) to construct a fluorescence response pattern (Figure S4 and S6) and calibration plot for unknown analyses, Figure 4. This new calibration PLSDA plot was also able to discriminate nine targets successfully with 95% confidence and without any overlap as anticipated. We have used this calibration to



Page 10 of 26

identify 24 unknown targets with different concentrations prepared from the nine standards. First, unknown samples were prepared from stock solutions of the nine let-7 targets (let-7a, let-7b, let-7c, let-7d, let-7e, let-7f, let-7g, let-7i and let-7j). The absorbance of each unknown was measured by UV-visible spectra and diluted to 0.01-a.u. OD value. The dilution factors were recorded for the subsequent back-calculations. Later the dilute targets were tested with the sensor array and analyzed using the PLSDA calibration plot. The unknowns were predicted using PLSDA and 19 unknown samples were identified successfully. Furthermore, their concentrations were back calculated with an average deviation of ~2.3% using the dilution coefficient recorded initially for each unknown, Table 2. Using this novel 2D-np based combinatorial sensory approach, we were not only able to discriminate very similar miRNA analogs but also identify and quantify unknown targets with unknown concentrations. Overall differences in hybridization, surface interactions and intermolecular forces between the targets and 2D-NAs enable discrimination of all nine targets with a single nanosensor array. To our knowledge, this is the only nanotechnology platform reported for simultaneous identification of this many oligonucleotide targets with such a high degree of homology in under four hours. Later, we challenged our nanosensor array to characterize mixtures of two targets ranging from 25% to 75% overall content. Instead of using pure targets, we prepared mixtures of let-7e and let-7i as model groups. The sensor array was tested with all these mixtures and all three mixtures were discriminated from each other and pure targets with 95% confidence without any overlap, Figure 5a. As the let-7i content increases with a decrease in let-7e content the PLSDA clusters move from upper-right quadruplet to lower-left one. Next, we studied the discrimination of the DNA and RNA analogs of let-7a, let-7b and let-7c. Let-7a and let-7b differ from let-7c by a single nucleotide and from each other by two nucleotides. The results showed that all six entities were discriminated successfully using our sensor array nanotechnology, Figure 5b. The RNA targets that are localized at the lower-half of the PLSDA are well-separated from the DNA targets at the upper-half of the plot. All clusters are separated from each other with 95% confidence without any overlap. The results overall demonstrate that not only a minor difference in the nucleobase composition can be discriminated, but also different analogs of exactly the same sequence can be identified. Finally, in order to demonstrate that the identification of a target can be achieved in a complex biological matrix, we prepared a liquid biopsy model using human urine and a miRNA target. Detection of a miRNA target, let-7b, was tested in urine specimen with and without let-7b enrichment. The urine sample with 50 nM of let-7b was successfully discriminated from urine without the target miRNA, Figure 6. 4. Conclusion


Page 11 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


We have assembled a universal combinatorial sensory approach for discriminating nine short miRNA analogs simultaneously. The difference among the targets ranges from one to six nucleotides. The approach relies on exploiting the behavioral differences between 15 2D-NAs assembled using three different 2D-nps and five fluorescently labelled ssDNA probes. The sum of interactions between each target and 15 2D-NAs was measured and recorded individually. The results were processed using PLSDA and accurate discrimination was achieved. The unknown studies showed successful discrimination of targets with remarkable similarities in molecular composition, overall size, and nucleobase sequence. The discrimination of pure targets from target mixtures was demonstrated for multi-component analysis purposes. The DNA and RNA analogs of the most similar strands (let-7a, -7b, -7c) were analyzed with the sensor array and each target was successfully discriminated. Identification of a target in a complex biological matrix prepared with human urine was demonstrated. Because the discrimination of a single nucleotide mutation is important for a number of genetic diseases, this nanotechnology can be instrumental in detection of SNPs beyond miRNA research.39 Besides its remarkable power in molecular analysis, the sensor array has several advantageous features in cost, scalability, operation and assembly. The water-soluble 2D-NAs in the sensor array are facile to assemble and operate. The components of the constructs, 2D-nps and ssDNAs, are inexpensive and easy to obtain, synthesize and scale up for a translational purpose. Multiple compounds can be analyzed simultaneously with a separationfree single step. The background or any possible false positive or negative signals are already included in the large data set, thus the approach is bias-free.

The signals can be recorded in real-time using a simple

microplate-reader. The entire process in this approach, including the sample preparation steps, takes about 3-4 hours.

The approach relies solely on the natural interactions between 2D-nps and fluorescently labeled

ssDNAs. Traditional non-specific sensor approaches report construction of sensor arrays using the artificial interactions between the chemically modified nanoparticles and genetically engineered fluorescent proteins or peptides.40 Though these studies are remarkable, the introduction of the artificial bonds through a limited number of chemical modifications could be an obstacle from a practical standpoint. Furthermore, synthetic procedures could be time-consuming, complicated, costly or result in low-yield and thus may be challenging for the broader scientific community. In our approach, fluorescently labeled ssDNAs are used as probes. The natural interactions between the ssDNA and 2D surface can be fine-tuned by varying the nucleobase composition, oligonucleotide length and the type of 2D-np, therefore there are countless possibilities for the assembly of 2D-NAs. Furthermore, the DNAs can be fluorescently labeled from a large fluorophore library adding another layer to the number of possible components for the sensor array. Therefore, the approach described here can be advanced and fine-tuned indefinitely for meeting a particular sensing criterion. The novel



Page 12 of 26

application reported here will significantly advance the field of two-dimensional nanoparticles for analysis of various biologics. Supporting Information. The results of additional experiments, characterization of 2D nanomaterials, fluorescence response patterns and training matrices of the fluorescence response patterns for Partial Least Squares (PLS) discriminant analysis (PLSDA) are provided. 4. Acknowledgements. We thank Prof. Ken Halvorsen for providing three of the let-7 miRNA targets. We thank Dr. Lenka for her help in statistical discriminant analysis. We acknowledge the Ministry of National Education, Republic of Turkey, for providing financial support to Mustafa Salih Hizir with full scholarship during his doctoral studies.


Page 13 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References. (1) Carter, K. A.; Shao, S.; Hoopes, M. I.; Luo, D.; Ahsan, B.; Grigoryants, V. M.; Song, W.; Huang, H.; Zhang, G.; Pandey, R. K.; Geng, J.; Pfeifer, B. A.; Scholes, C. P.; Ortega, J.; Karttunen, M.; Lovell, J. F. Porphyrin-phospholipid liposomes permeabilized by near-infrared light. Nat. Commun. 2014, 5, 3546. (2) Doane, T. L.; Alam, R.; Maye, M. M. Functionalization of quantum rods with oligonucleotides for programmable assembly with DNA origami. Nanoscale 2015, 7, 2883-2888. (3) Ray, A.; Li, S.; Segura, T.; Ozcan, A. High-Throughput Quantification of Nanoparticle Degradation Using Computational Microscopy and Its Application to Drug Delivery Nanocapsules. ACS Photonics 2017, 4, 12161224. (4) Wang, W.; Satyavolu, N. S. R.; Wu, Z.; Zhang, J. R.; Zhu, J. J.; Lu, Y. Near-Infrared Photothermally Activated DNAzyme-Gold Nanoshells for Imaging Metal Ions in Living Cells. Angew. Chem. Int. Ed. Engl. 2017, 56, 6798-6802. (5) Alexander, Y. Z.; Ertugrul, C. Graphene nanophotonic sensors. 2D Materials 2015, 2, 032005. (6) Acar, H.; Genc, R.; Urel, M.; Erkal, T. S.; Dana, A.; Guler, M. O. Self-assembled peptide nanofiber templated one-dimensional gold nanostructures exhibiting resistive switching. Langmuir 2012, 28, 1634716354. (7) Sulek, S.; Mammadov, B.; Mahcicek, D. I.; Sozeri, H.; Atalar, E.; Tekinay, A. B.; Guler, M. O. Peptide functionalized superparamagnetic iron oxide nanoparticles as MRI contrast agents. J. Mater. Chem. 2011, 21, 15157-15162. (8) Chhowalla, M.; Liu, Z.; Zhang, H. Two-dimensional transition metal dichalcogenide (TMD) nanosheets. Chem. Soc. Rev. 2015, 44, 2584-2586. (9) Chhowalla, M.; Shin, H. S.; Eda, G.; Li, L. J.; Loh, K. P.; Zhang, H. The chemistry of two-dimensional layered transition metal dichalcogenide nanosheets. Nat. Chem. 2013, 5, 263-275. (10) Tan, C.; Zhang, H. Two-dimensional transition metal dichalcogenide nanosheet-based composites. Chem. Soc. Rev. 2015, 44, 2713-2731. (11) Joensen, P.; Frindt, R. F.; Morrison, S. R. Single-layer MoS2. Mater. Res. Bull. 1986, 21, 457-461. (12) Allen, M. J.; Tung, V. C.; Kaner, R. B. Honeycomb carbon: a review of graphene. Chem. Rev. 2010, 110, 132-145.



Page 14 of 26

(13) Liu, B.; Sun, Z.; Zhang, X.; Liu, J. Mechanisms of DNA sensing on graphene oxide. Anal Chem 2013, 85, 7987-7993. (14) Huang, P. J.; Liu, J. Molecular beacon lighting up on graphene oxide. Anal Chem 2012, 84, 4192-4198. (15) Zhang, Y.; Zheng, B.; Zhu, C.; Zhang, X.; Tan, C.; Li, H.; Chen, B.; Yang, J.; Chen, J.; Huang, Y.; Wang, L.; Zhang, H. Single-layer transition metal dichalcogenide nanosheet-based nanosensors for rapid, sensitive, and multiplexed detection of DNA. Adv. Mater. 2015, 27, 935-939. (16) Xi, Q.; Zhou, D. M.; Kan, Y. Y.; Ge, J.; Wu, Z. K.; Yu, R. Q.; Jiang, J. H. Highly sensitive and selective strategy for microRNA detection based on WS2 nanosheet mediated fluorescence quenching and duplexspecific nuclease signal amplification. Anal. Chem. 2014, 86, 1361-1365. (17) Ge, J.; Ou, E.-C.; Yu, R.-Q.; Chu, X. A novel aptameric nanobiosensor based on the self-assembled DNAMoS2 nanosheet architecture for biomolecule detection. J. Mater. Chem. B 2014, 2, 625-628. (18) Zhu, C.; Zeng, Z.; Li, H.; Li, F.; Fan, C.; Zhang, H. Single-layer MoS2-based nanoprobes for homogeneous detection of biomolecules. J. Am. Chem. Soc. 2013, 135, 5998-6001. (19) Chen, Y.; Tan, C.; Zhang, H.; Wang, L. Two-dimensional graphene analogues for biomedical applications. Chem. Soc. Rev. 2015, 44, 2681-2701. (20) Hizir, M. S.; Robertson, N. M.; Balcioglu, M.; Alp, E.; Rana, M.; Yigit, M. V. Universal sensor array for highly selective system identification using two-dimensional nanoparticles. Chem. Sci. 2017, 8, 5735-5745. (21) Liu, B.; Sun, Z.; Zhang, X.; Liu, J. Mechanisms of DNA sensing on graphene oxide. Anal. Chem. 2013, 85, 7987-7993. (22) Balcioglu, M.; Buyukbekar, B. Z.; Yavuz, M. S.; Yigit, M. V. Smart-Polymer-Functionalized Graphene Nanodevices for Thermo-Switch-Controlled Biodetection. ACS Biomater. Sci. Eng. 2015, 1, 27-36. (23) Hizir, M. S.; Balcioglu, M.; Rana, M.; Robertson, N. M.; Yigit, M. V. Simultaneous detection of circulating oncomiRs from body fluids for prostate cancer staging using nanographene oxide. ACS Appl. Mater. Interfaces 2014, 6, 14772-14778. (24) Rana, M.; Balcioglu, M.; Kovach, M.; Hizir, M. S.; Robertson, N. M.; Khan, I.; Yigit, M. V. Reprogrammable multiplexed detection of circulating oncomiRs using hybridization chain reaction. Chem. Commun. (Camb) 2016, 52, 3524-3527. (25) Robertson, N. M.; Salih Hizir, M.; Balcioglu, M.; Wang, R.; Yavuz, M. S.; Yumak, H.; Ozturk, B.; Sheng, J.; Yigit, M. V. Discriminating a Single Nucleotide Difference for Enhanced miRNA Detection Using Tunable Graphene and Oligonucleotide Nanodevices. Langmuir 2015, 31, 9943-9952.


Page 15 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(26) Robertson, N. M.; Toscano, A. E.; LaMantia, V. E.; Hizir, M. S.; Rana, M.; Balcioglu, M.; Sheng, J.; Yigit, M. V. Unlocked Nucleic Acids for miRNA detection using two dimensional nano-graphene oxide. Biosens. Bioelectron. 2017, 89, 551-557. (27) Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 2004, 116, 281-297. (28) Poy, M. N.; Eliasson, L.; Krutzfeldt, J. A pancreatic islet-specific microRNA regulates insulin secretion. Nature 2004, 432, 226-230. (29) Esquela-Kerscher, A.; Slack, F. J. Oncomirs - microRNAs with a role in cancer. Nat. Rev. Cancer 2006, 6, 259-269. (30) Cortez, M. A.; Bueso-Ramos, C.; Ferdin, J.; Lopez-Berestein, G.; Sood, A. K.; Calin, G. A. MicroRNAs in body fluids--the mix of hormones and biomarkers. Nat. Rev. Clin. Oncol. 2011, 8, 467-477. (31) Johnson, B. N.; Mutharasan, R. Biosensor-based microRNA detection: techniques, design, performance, and challenges. Analyst 2014, 139, 1576-1588. (32) Cissell, K. A.; Shrestha, S.; Deo, S. K. MicroRNA Detection: Challenges for the Analytical Chemist. Anal. Chem. 2007, 79, 4754-4761. (33) Lan, H.; Lu, H.; Wang, X.; Jin, H. MicroRNAs as potential biomarkers in cancer: opportunities and challenges. Biomed. Res. Int. 2015, 2015, 125094. (34) Yoo, B.; Kavishwar, A.; Ghosh, S. K.; Barteneva, N.; Yigit, M. V.; Moore, A.; Medarova, Z. Detection of miRNA expression in intact cells using activatable sensor oligonucleotides. Chem. Biol. 2014, 21, 199-204. (35) Roush, S.; Slack, F. J. The let-7 family of microRNAs. Trends Cell Biol. 2008, 18, 505-516. (36) Takamizawa, J.; Konishi, H.; Yanagisawa, K.; Tomida, S.; Osada, H.; Endoh, H.; Harano, T.; Yatabe, Y.; Nagino, M.; Nimura, Y.; Mitsudomi, T.; Takahashi, T. Reduced expression of the let-7 microRNAs in human lung cancers in association with shortened postoperative survival. Cancer Res. 2004, 64, 3753-3756. (37) Huang, X.; Yang, Y.; Guo, Y.; Cao, Z. L.; Cui, Z. W.; Hu, T. C.; Gao, L. B. Association of a let-7 KRAS rs712 polymorphism with the risk of breast cancer. Genet. Mol. Res. 2015, 14, 16913-16920. (38) Lu, C.; Liu, Y.; Ying, Y.; Liu, J. Comparison of MoS2, WS2, and Graphene Oxide for DNA Adsorption and Sensing. Langmuir 2017, 33, 630-637. (39) Yang, M.; Xiao, X.; Xing, X.; Li, X.; Xia, T.; Long, H. KRAS and VEGF gene 3'-UTR single nucleotide polymorphisms predicted susceptibility in colorectal cancer. PLoS One 2017, 12, e0174140.



Page 16 of 26

(40) Rana, S.; Singla, A. K.; Bajaj, A.; Elci, S. G.; Miranda, O. R.; Mout, R.; Yan, B.; Jirik, F. R.; Rotello, V. M. Array-based sensing of metastatic cells and tissues using nanoparticle-fluorescent protein conjugates. ACS Nano 2012, 6, 8233-8240.


Page 17 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figures

Scheme 1. The desorption of ssDNA probes in a 2D-NA occurs either by a) the target specific bind-and-release model or b) nonspecific displacement by the target.



Page 18 of 26

Figure 1. (a) The partial hybridization between (b) the 5’-fluorescently labelled 15-mer ssDNA probes and (c) the let-7 targets.


Page 19 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 1. The degree of complementarity between each target and probe.



Page 20 of 26

Figure 2. a) ssDNA probes (p1, p2, p3, p4 and p5) are assemblied with the 2D-nps (nGO, MoS2 and WS2) to form b) 15 2D-NAs which were placed in c) an array-layout in a 96-well microplate in six replicates. d) A target miRNA analog is tested with the sensor array and displaces the ssDNA probe from the 2D-np surface through partial hybridization. e) Recording the fluorescence recovery in each well and processing with PLSDA enables discrimination of nine targets with 95% confidence. Abbreviations: 2D-nps= two-dimensional nanoparticles; MoS2, WS2 and nGO, probe = single stranded DNAs (ssDNAs), 2D-NAs = nanoassemblies of 2D-nps and fluorescently labeled ssDNAs.


Page 21 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. a) Fluorescence spectra and b) a kinetic study over 160 mins demonstrate that upon addition of nGO (2.5 µg mL-1) to 20 nM of FAM-labeled ssDNA probe (p2), the fluorescence (probe fluorescence) of the FAM label at 520 nm is quenched (quenched fluorescence, fo). Whereas, the addition of let-7j target desorbs the p2 from the surface through partial RNA:DNA hybridization resulting in a fluorescence recovery (recovered fluorescence, f). The fluorescence recovery (response of 2D-NA-2 [nGO-p2] against let-7j) was calculated by the difference between fo and f ; (∆ƒ = ƒ - ƒ0) 2 hrs after the addition of let-7j target.



Figure 4. PLSDA plot for miRNA analogs with Abs260 equivalent to 0.01-a.u. (Abs260 = 0.01 a.u.).


Page 22 of 26

Page 23 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2. Identification of unknown targets with unknown concentrations using PLS prediction function.



Page 24 of 26

Figure 5. PLSDA plot for a) target mixtures of let-7e:let-7i in 100:0, 75:25, 50:50, 25:75 and 0:100% from upper right to lower left corner and b) RNA and DNA analogs of let-7a, let-7b and let-7c.


Page 25 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6. PLSDA plot for discriminating urine samples with and without let-7b RNA target with 95% confidence.



Table of Contents


Page 26 of 26

Homologous miRNA analyses using a combinatorial nanosensor

Recommend Documents