Quantification of microRNA by DNA–Peptide Probe ... - ACS Publications

Dec 7, 2015 - Department of Pharmacy, Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School, Nanjing,. 210029, Chi...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/ac

Quantification of microRNA by DNA−Peptide Probe and Liquid Chromatography−Tandem Mass Spectrometry-Based QuasiTargeted Proteomics Feifei Xu,† Ting Yang,‡ and Yun Chen*,† †

School of Pharmacy, Nanjing Medical University, Nanjing, 211166, China Department of Pharmacy, Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School, Nanjing, 210029, China



S Supporting Information *

ABSTRACT: The distorted and unique expression of microRNAs (miRNAs) in cancer makes them an attractive source of biomarkers. However, one of prerequisites for the application of miRNAs in clinical practice is to accurately profile their expression. Currently available assays normally require pre-enrichment, amplification, and labeling steps, and most of them are semiquantitative. In this study, we converted the signal of target miR-21 into reporter peptide by a DNA-peptide probe and the reporter peptide was ultimately quantified using LC-MS/MS-based targeted proteomics. Specifically, substrate peptide GDKAVLGVDPFR containing reporter peptide AVLGVDPFR and tryptic cleavage site (lysine at position 3) was first designed, followed by the conjugation with DNA sequence that was complementary to miR-21. The newly formed DNA-peptide probe was then hybridized with miR-21, which was biotinylated and attached to streptavidin agarose in advance. After trypsin digestion, the reporter peptide was released and monitored by a targeted proteomics assay. The obtained limit of quantification (LOQ) was 1 pM, and the detection dynamic range spanned ∼5 orders of magnitude. Using this assay, the developed quasi-targeted proteomics approach was applied to determine miR-21 level in breast cells and tissue samples. Finally, qRT-PCR was also performed for a comparison. This report grafted the strategy of targeted proteomics into miRNA quantification. icroRNAs (miRNAs) are endogenous RNAs of ∼22 nucleotides that play important regulatory roles in most major cellular processes.1 They normally bind to the 3′ untranslated region (3′ UTR) of target mRNAs, leading to either mRNAs degradation or translation inhibition.2 Both of these regulations can result in gene silencing. Thus, it is not surprising that miRNAs have been linked to a variety of diseases, including breast cancer.3 Their differential levels in normal volunteers and patients give them potential for use as diagnostic/prognostic markers and therapeutic targets.4 How-

m

© XXXX American Chemical Society

ever, a prerequisite for the application of miRNAs in clinical practice is to profile their expression accurately. Currently, many detection techniques are available for the study of miRNAs expression profile, including indirect methods (e.g., polymerase chain reaction (PCR),5,6 fluorescent resonance energy transfer (FRET),7 microarrays,8 electrocatalysis,9 Received: August 5, 2015 Accepted: November 17, 2015

A

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 1. Schematic representation of DNA-peptide probe preparation and its combination with targeted proteomics for the quantification of miR21. A reporter peptide was first selected and then substrate peptide containing the sequence of reporter peptide and tryptic cleavage site was tagged to a DNA sequence at the 5′ end that was complementary to that of the target miRNA. Subsequently, the DNA-peptide probe was hybridized with the miRNA that was biotinylated at the 3′ end and bound to streptavidin agarose through streptavidin−biotin interaction in advance. After trypsin digestion, the reporter peptide was released and quantified using a targeted proteomics assay. In this way, the presence and the quantity of miR-21 can be inferred from the detection of reporter peptide.

and next-generation sequencing10) and direct methods (e.g., electrochemical-based methods,11−13 spectral detection assisted by duplex-specific nuclease,14,15 differential interference contrast (DIC) imaging,16 and capillary electrophoresis (CE)based methods17,18). Indirect methods normally require preamplification or chemical/enzymatic modification of the target miRNA, whereas direct ones do not involve amplification/modification of miRNA.3 While these techniques provide valuable information, their results are often not straightforward to interpret. The majority of them require miRNAs enrichment and/or amplification, and further steps of gel visualization or labeling with fluorophores,19 all of which are time- and labor-consuming, and are susceptible to various factors.20 Recent direct detection techniques are attractive, whereas they are primarily in early stages of development and still require significant efforts to make them practical.18 More importantly, most of the developed assays provide only a limited degree of qualitative data.21 Generally, a method with high quantitative accuracy would allow for the detection of slight deregulations in miRNA levels, rather than the presence or absence of particular miRNA species. Thus, exploitation of novel strategy that can offer quantitative values of miRNAs in a sensitive and straightforward manner is imperative. Mass spectrometry has shown its potential to study biomolecules, because of its high sensitivity, high selectivity, and wide dynamic range.22,23 In recent years, this technique has become an indispensable tool in proteomics to obtain quantitative information about proteins. Liquid chromatography−tandem mass spectrometry (LC-MS/MS)-based targeted proteomics is one of the recognized mass spectrometric analysis methods.24−26 The underlying principle of this targeted analysis is specific detection and determination of a protein of interest at the peptide level.23,27 Peptides are generated by proteolytic digestion of the target protein to serve as surrogate analytes. Selected or multiple reaction monitoring (SRM or MRM, respectively) is used to detect the selected surrogate

peptides.25 Technically, direct miRNA analysis using mass spectrometry is amenable to the same approach as proteins. However, the application encountered some difficulties, probably because of complicated and unresolved mass spectra of miRNAs,28,29 especially that miRNAs consist of only four nucleotides that the risk of different sequences giving rise to similar mass spectra patterns is potentially greater than for those molecules containing amino acids.30 Alternative strategies, including small molecule reporter tags31,32 and highresolution mass spectrometers,33 either lack multiplexing capability or are expensive. However, we cannot arbitrarily rule out the use of mass spectrometry in miRNA quantification. From our point of view, the concept of surrogate peptide in targeted proteomics can be introduced into miRNA quantification. In the present work, we developed a LC-MS/MS-based quasi-targeted proteomics assay for the quantification of miRNA-21 (miR-21). As is well-known, miR-21 is an oncomiRNA in cancer that can down-regulate a variety of tumor suppressor proteins.34−36 In addition, its altered expression is also implied in the acquisition of cancer drug resistance.35 To profile miR-21 expression, our approach actually converted the miR-21 signal into reporter peptide by a DNA−peptide probe and the reporter peptide was ultimately quantified using targeted proteomics. Figure 1 shows the proposed scheme for this strategy. After characterization of the parameters related to binding, conjugation, and hybridization, the assay was validated and finally applied to determine the miR-21 level in the normal cells MCF-10A, the parental drug-sensitive cancer cells MCF7/WT and the drug-resistant cancer cells MCF-7/ADR, and 36 pairs of human breast primary tumors and adjacent normal tissue samples. The resulting values were also compared to those obtained with a quantitative reverse transcription polymerase chain reaction (qRT-PCR) method. B

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry



MATERIALS AND METHODS Chemicals and Reagents. Peptides including maleimidemodified substrate peptide, reporter peptide, and internal standard containing stable-isotope labeled amino acids were developed by ChinaPeptides Co., Ltd. (Shanghai, China). Purity of the peptides was also provided by the manufacturer. The stable isotope-labeled amino acid was supplied by Cambridge Isotope Laboratories, Inc. (Andover, MA, USA). miR-21 with/without mismatches and its complementary DNA with a disulfide modification at the 5′ end were customsynthesized by Realgene (Nanjing, China) and Genscript (Nanjing, China), respectively. miR-21 has the following sequence: 5′-UAGCUUAUCAGACUGAUGUUGA-3′. The corresponding DNA is 5′-thiol-C6-TCAACATCAGTCTGATAAGCTA-3′. Ammonium bicarbonate (NH4HCO3) was obtained from Qiangshun Chemical Reagent Co., Ltd. (Shanghai, China). D,L-dithiothreitol (DTT), iodoacetamide (IAA), and Tris-HCl were supplied by Sigma−Aldrich (St. Louis, MO, USA). Ethylenediaminetetraacetic acid disodium salt (EDTA.2Na) was obtained from Sinopharm Chemical Reagent Company (Shanghai, China). Sequencing grademodified trypsin was purchased from Promega (Madison, WI, USA). Phosphate buffered saline (PBS) was purchased from the Beyotime Institute of Biotechnology (Jiangsu, China). Acetonitrile (ACN) and methanol were obtained from Tedia Company, Inc. (Fairfield, OH, USA). Trifluoroacetic acid (TFA) and formic acid (FA) were provided by Aladdin Chemistry Co., Ltd. (Shanghai, China) and Xilong Chemical Industrial Factory Co., Ltd. (Shantou, China), respectively. Dulbecco’s Modified Eagle Media (DMEM) and fetal bovine serum were obtained from Thermo Scientific HyClone (Logan, UT, USA). MEGM Mammary Epithelial Cell Growth Medium was obtained from LONZA (Basel, Switzerland). Penicillin was supplied by CSPC Zhongnuo Pharmaceutical Co., Ltd. (Shijiazhuang, China). Streptomycin was obtained from Merro Pharmaceutical Co., Ltd. (Dalian, China). Trypan Blue and sodium dodecyl sulfate (SDS) were obtained from Generay Biotech Co., Ltd. (Shanghai, China). All the solutions used in the experiments were prepared in DEPC-treated water (Beyotime Biotechnology, Haimen, China). Cell Culture and Tissue Collection. MCF-7/WT (ATTC, Manassas, VA) and MCF-7/ADR (Keygen Biotech, Nanjing, China) cells were cultured in a DMEM media supplemented with 10% fetal bovine serum, 80 U/mL penicillin, and 80 μg/ mL streptomycin at 37 °C and 5% CO2. MCF-10A cells (ATTC, Manassas, VA) were maintained routinely in MEGM media supplemented with 100 ng/mL cholera toxin and 1% penicillin/streptomycin. Cells were split every 5−7 days by lifting cells with 0.25% trypsin and feeding between splits through the addition of fresh medium. To maintain a highly drug-resistant cell population, MCF-7/ADR cells were periodically reselected by growing them in the presence of 1000 ng/ mL DOX.37 Experiments were performed using the cells incubated without DOX for 48 h. Cells were counted with a hemocytometer (Qiujing, Shanghai, China). Cell viability was assessed by Trypan Blue (0.4%) exclusion. Cell suspensions, Trypan Blue, and 1× PBS were mixed in a 2:5:3 ratio and the percentage of viable cells was counted after incubation for 5 min at 37 °C. Breast tissue collection in this study was approved by the Institutional Review Board of Nanjing Medical University. The methods were carried out in accordance with the approved

guidelines. Thirty-six (36) pairs of breast tissue samples consisting of tumors and adjacent normal sections were collected consecutively between January 2012 and December 2012 at the First Affiliated Hospital of Nanjing Medical University and Nanjing Drum Tower Hospital, Nanjing, China (mean patient age, 52.9 ± 8.6 years; age range, 38−65 years). Tissue sections were confirmed as normal and cancerous by hospital pathologists. Histological evaluation of adjacent normal tissue samples showed no indication of contamination from tumor or other abnormal cells. The patients were biologically unrelated, but all belonged to the Han Chinese ethnic group from the Jiangsu province in China. Informed consent was obtained from the subjects. Tissue samples were stored frozen at −80 °C until analysis. Prior to RNA extraction, tissue samples were thawed to room temperature and then rinsed thoroughly with deionized water. Fat tissue was removed and the remaining tissue cut into small pieces and transferred to tubes. Approximately 50 mg of tissue was weighted and homogenized in TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) using a Bio-Gen PRO200 homogenizer (PRO Scientific, Inc., Oxford, CT, USA). Formation of DNA−Peptide Probe. The procedure for DNA−peptide probe preparation has been described previously by Krylov’s laboratory.18 The synthetic DNA contained a disulfide modification on its 5′ end that needed to be reduced prior to conjugation. Tris(2-carboxyethyl)-phosphine (TCEP) was used as the reducing agent in this study. Twenty microliters (20 μL) of TCEP reducing beads (Thermo Fisher Scientific, Chicago, IL, USA) was incubated with 100 μL of 2 μM DNA at 37 °C for 2 h with vigorous shaking. The sample was then centrifuged at 1000g for 5 min, and the supernatant containing the reduced DNA was added to an equal volume of 20 μM maleimide-modified substrate peptide. The conjugation reaction was carried out at 37 °C for 4 h with vigorous shaking, followed by immediate purification. The DNA−peptide conjugate was isolated from the excess of nonconjugated DNA and peptide by high-performance liquid chromatography (HPLC). The HPLC conditions are given in the figures shown in the Supporting Information. Quantification was performed using external calibration peak area measurement. Biotinylation of microRNAs. Total RNA was isolated from cells and tissue homogenates using the Trizol Reagent and quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Inc., Waltham, MA, USA). Following the manufacturer’s protocol, 100 μg of total RNA in a reaction volume of 30 μL was biotinylated using Pierce RNA 3′ End Biotinylation Kit (Thermo Fisher Scientific, Inc., Waltham, MA, USA). Then, 20 μL biotinylated RNA was added with an equal volume of streptavidin agarose (Life Technologies, Frederick, MD, USA) and incubated at 37 °C on a shaker for 2 h, followed by washing and centrifugation. Streptavidin agarose was coated in advance with RNase-free BSA and yeast tRNA (Ambion, Austin, TX, USA) to prevent nonspecific binding of RNA and protein complexes. For each 100 μL of beads, 10 μL tRNA (10 mg/mL stock) and 10 μL BSA (10 mg/mL stock) were added. Optimization of Hybridization Conditions. Hybridization was carried out in a MJ Mini thermocycler (Bio-Rad, Hercules, CA, USA). Streptavidin agarose attached with biotinylated miRNAs were incubated with 100 μL of 500 nM DNA-peptide probe in the presence of 1 μM masking DNA (Brand Realgene (Nanjing, China) and 25 μL of hybridization buffer. In the present work, six typical hybridization buffers were examined (see Table 2S in the Supporting Information). C

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 2. (A) Product ion spectrum of AVLGVDPFR and (B) LC-MS/MS chromatograms of AVLGVDPFR and its corresponding isotope-labeled internal standard. The MRM transitions of 487.3 → m/z 419.3 for AVLGVDPFR and 495.3 → m/z 419.3 for internal standard were used. The transitions of m/z 487.3 → m/z 171.1 and m/z 487.3 → m/z 690.3 for AVLGVDPFR were omitted for clarity.

quality control (QC) standards for lower limit of quantification (LLOQ), low QC, mid QC, and high QC were prepared at 1 pM, 3 pM, 500 pM, and 80 nM and frozen prior to use. With regard to the reporter peptide selected in the next section, the corresponding isotope-labeled synthetic peptide was used as an internal standard. The internal standard was also weighed, and 100 μM stock solution was prepared. Internal standard solution (10 nM) was prepared by diluting the stock solutions with an ACN:water mixture (50:50, v/v) containing 0.1% FA. In-Solution Tryptic Digestion. The streptavidin agarose:biotinylated miRNAs:DNA-peptide complexes were mixed with 100 μL of 50 mM NH4HCO3. Subsequently, sequencing-grade trypsin was added and the sample was incubated at 37 °C for 24 h. The reaction was stopped by adding 10 μL of 0.1% TFA. Then, 100 μL of the internal standard solution was added to the tryptic peptide mixture before transferring it into an Oasis

Hybridization time and temperature were also evaluated. After hybridization, the agarose beads were thoroughly washed and centrifuged to remove any unbound peptide−DNA probe. In parallel, miR-21 molecules with single-base-mismatch (5′UAGCUUAUCAGUCUGAUGUUGA-3′) and two-base-mismatch (5′-UAGCUUAUCAGUGUGAUGUUGA-3′) sequences were hybridized according to the hybridization procedure described above. Underlined letters in the sequences represent the mismatch sites. Preparation of Stock Solutions, Calibration Standards and Quality Controls (QCs). Stock solution (100 μM) was prepared by accurately weighing the synthetic miR-21 and dissolving it in DEPC-treated water. The solution was stored at −20 °C in a brown glass tube to protect it from light. The calibration standards were prepared by serial dilution of the stock solution. The concentrations of the calibration standards were 1 pM, 10 pM, 100 pM, 1 nM, 10 nM, and 100 nM. The D

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

models can only assist to select high-responding peptides, particularly in the absence of experimental data. The reliability of selection generally needs further validation. Following the empirical rules and in silico prediction, a peptide with the sequence of AVLGVDPFR was employed as the reporter peptide. So far, this sequence was found not to match with any protein using a BLAST search. The most abundant form of AVLGVDPFR in mass spectrum was its doubly charged ion. Consistent with the result of ESP Predictor (0.80, full score is 1), the peptide response was significant and among the top values of the peptides that we ever processed using targeted analysis.41,45,46 Its product ion spectrum and LCMS/MS chromatogram are shown in Figure 2A. The characteristic sequence-specific b ions and y ions were indicative of a peptide. Then, the corresponding stable isotope-labeled peptide was prepared to serve as an internal standard. In detail, stable isotope-labeled [D8]Val was coupled to AV*LGV*DPFR at positions 2 and 5 to yield a molecular mass shift of 16 Da from the nonlabeled peptide and a monoisotopic molecular mass of 989.3 Da. The retention times for AVLGVDPFR and its isotope-labeled peptide were identical (∼3.3 min; see Figure 2B). Finally, the MRM transitions that gave the best signal-to-noise ratio and LOQ for AVLGVDPFR were afforded by the product ions b2 m/z 171.1, y3 m/z 419.3, and y6 m/z 690.3. The detection limit can achieve 1 pM. Thus, the peak areas from three transitions m/z 487.3 → m/z 419.3, m/z 487.3 → m/z 171.1, and m/z 487.3 → m/z 690.3 were summed and used in the following quantitative analysis.47 Note that the reporter ion must be ultimately released from the DNA−peptide after the hybridization between miR-21 and the complementary DNA. Thus, a substrate peptide with a tryptic cleavage site prior to AVLGVDPFR was further proposed. The key concern here was the digestion efficiency, because incomplete tryptic digestion could result in inaccurate quantification of the analytes, with regard to the fact that the reporter peptide, rather than the substrate peptide, was used as the calibration standard.36 In addition, the maximum detection sensitivity is digestion-dependent.48 Thus, a three-amino-acid peptide (GDK) containing a lysine residue at the carboxyl terminus were linked with the reporter peptide to form the substrate peptide GDKAVLGVDPFR. Its digestion efficiency was calculated by comparing the response ratios of the tryptic peptide after digestion and the equimolar synthetic reporter peptide standard in the digestion. The estimated values were 95.4%, consistent with SVM score of 0.69 generated by the STEPP software (see Figure 1S in the Supporting Information). Notably, the following conjugated DNA and hybridized miR-21 on the agarose beads did not have a significant impact on the digestion. One of the explanations is that the tryptic site intentionally designed by three amino acid residues (GDK) away from the DNA to avoid potential steric hindrance in trypsin cleavage. Conversely, this spatial block also reduced the effect of peptide tag on the function of DNA (data not shown). Characterization of DNA−Peptide Probe. To form the DNA−peptide probe, the disulfide group on DNA needed to be reduced prior to conjugation, and the thiol group then reacted with a maleimide group at the peptide N-terminus by Michael addition to form a thiol-maleimide linkage.49 The HPLC result indicated that almost all the DNA was conjugated with the substrate peptide in the presence of excess peptide, whereas this complete conjugation was not required here, because the newly conjugated DNA−peptide probe can be well separated and collected for the following experiment. As shown

HLB cartridge (60 mg/3 mL; Waters, Milford, MA, USA) that was preconditioned with 3 mL of ACN and 3 mL of deionized water. After the sample was loaded, the cartridge was washed with 2 mL of water and 2 mL of ACN:water (50:50, v/v) and eluted with 1 mL of 100% ACN. Finally, the eluent was evaporated to dryness, and the sample was resuspended in 100 μL of ACN:water (50:50, v/v) containing 0.1% FA. LC-MS/MS Method Development and Validation. An Agilent Series 1290 UPLC system (Agilent Technologies, Waldbronn, Germany) and a 6460 Triple Quad LC-MS mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) were used for the LC-MS/MS studies. The liquid chromatography separations were performed on an Agilent SB C18 (2.7 μm, 30 mm × 2.1 mm, Agilent, USA) at room temperature. The mobile phase consisted of solvent A (0.1% FA in water) and solvent B (0.1% FA in methanol). A linear gradient with a flow rate of 0.3 mL/min was applied in the following manner: B 10% (0 min) → 10% (1 min) → 90% (4 min) → 90% (8 min) → 10% (9 min). The injection volume was 5 μL. The mass spectrometer was interfaced with an electrospray ion source and operated in the positive MRM mode. Q1 and Q3 were both set at unit resolution. The flow of the drying gas was 10 L/min, and the drying gas temperature was held at 350 °C. The electrospray capillary voltage was optimized to 4000 V. The nebulizer pressure was set to 35 psi. The data were collected and processed using the Agilent MassHunter Workstation Software (version B.06.00). Method validation involves evaluating the linear range, accuracy, precision, limit of quantification (LOQ), and stability. The detailed procedures and the acceptance criteria used to validate the assay have been described in several publications.22,38,39 Method Comparison. For the experimental details of qRT-PCR, please see the Supporting Information.



RESULTS AND DISCUSSION Selection of Reporter and Substrate Peptides. The first step of miRNA quantification is the design of DNA−peptide probe. As mentioned above, the different components of miRNAs from proteins result in a completely different rationale to select the reporter/surrogate peptides. In a targeted analysis of proteins, the most critical step in the establishment of a quantitative assay is the selection of surrogate peptides that (1) are unique to the target analyte, (2) could provide an adequate response, (3) have completeness of digestion, and (4) can generate high-quality SRM.40,41 For miRNA quantification, the first criterion of uniqueness is apparently not applicable. Indeed, the sequence of reporter peptide should not be found in any proteins using a BLAST search to be against protein interference in the analysis. This restriction may be stringent; whereas a good thing is that the sequence of selected peptide is not constraint to the primary analyte. The other three criteria can be reserved. Empirically, there are some rules for the choice of peptides, which may be helpful at the primary stage of searching and are listed in Table 1S in the Supporting Information.42,43 Furthermore, it is possible to predict which peptides and product ions are most appropriate for MRM by in silico prediction by various algorithms and computational tools.44 However, it deserves to be mentioned that the mechanisms of proteolysis, ionization, and fragmentation are not yet sufficiently well-understood to produce very accurate models from which to make such predictions. The current E

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry in Figure 3A, only the peaks of DNA and DNA−peptide can be detected in the HPLC chromatogram ,using the detection

Figure 4. (A) HPLC chromatogram of the DNA−peptide probe before and after tryptic digestion at the wavelength of DNA (260 nm) and (B) the corresponding LC-MS/MS chromatogram with MRM transition of m/z 487.3 → m/z 419.3. The HPLC conditions were the same as those described in Figure 3.

Figure 3. HPLC chromatograms (A) before and (B) after DNA and GDKAVLGVDPFR conjugation at wavelengths of DNA (260 nm) and peptide (220 nm). The background was subtracted for clarity. The HPLC system consisted of a solvent delivery pump (Shimadzu, Model LC-20AB), a manual valve injector (Rheodyne), and a UV/vis detector (Model SPD-20A, Shimadzu Corporation, Tokyo, Japan). The samples were analyzed using an Aqua HPLC C8 column (5 μm, 4.6 mm × 150 mm; Thermo Scientific, USA) at room temperature. The mobile phase consisted of solvent A (50 mM triethylamine, pH 7.6) and solvent B (ACN). A linear gradient with a flow rate of 1 mL/min was applied in the following manner (duration listed in parentheses): B 5% (0 min) → 16% (5 min) → 20% (20 min) →5% (25 min) → stop (26 min). The data were acquired and processed using a Lab Solutions LC-solution Version 1.2 working station.

amount of reporter peptide was not significantly different from the reactant amount of DNA−peptide probe for digestion. Sample Loading Efficiency on Beads. Among the hybridization techniques, immobilization of nucleotide molecules on a solid support is commonly used.50 A wide variety of immobilization methods are employed to attach the target analyte to the immobilization surface.51 Among them, bioaffinity immobilization employing biotin and streptavidin as immobilization partners is one of the most widely used.19 In the present work, miRNA was biotinylated and then reacted with streptavidin agarose, where streptavidin was commercially and covalently linked to agarose beads via a 15-atom hydrophilic spacer arm.52,53 To evaluate the maximum loading capacity of agarose beads for biotinylated miRNAs, increased amounts of biotin (5, 10, 15, and 20 μL; 100 μM) were reacted with 10 μL of streptavidin agarose (∼2−3 mg/mL of streptavidin) and the result indicated that 10 μL of biotin reached saturation (see Figure 2S in the Supporting Information). The binding site density immobilized on beads was estimated using the decreased free biotin in solution and the total number of streptavidin through the immobilization process (see Figure 5). The calculated biotin density on the bead surface was ∼2.1 biotin molecules/streptavidin. Since each streptavidin molecule has four biotin-binding sites, ∼53% streptavidin was saturated at the maximum level. The incompleteness of binding was probably due to the inherent covalent linkage between agarose and streptavidin formed on active sites of the protein, resulting in its reduced activity.51 Provided the loading efficiency of beads, almost all the miRNAs in samples can be captured using an excess of agarose beads.

wavelength of 260 nm. Their retention times were 10.0 and 21.4 min, respectively. Notably, the peptide was undetectable at this wavelength. To ensure that it did not coelute with the DNA−peptide conjugate that we collected, another detection wavelength for peptide (220 nm) was also employed while keeping the other conditions the same. As shown in Figure 3B, peptide was not eluted out of the column. To further characterize the DNA and peptide components in the conjugated product, the collected fraction was subjected to trypsin digestion. The released DNA and the tryptic peptide then were monitored using HPLC and LC-MS/MS, respectively. As shown in Figure 4, new peaks appeared after digestion. However, note that their retention times (9.34 min in HPLC and 3.30 min in LC-MS/MS) were not exactly the same as those of premier DNA and substrate peptide used for conjugation reaction (10.0 and 3.07 min). The slight time shift came from the loss of three amino acid residues from the substrate peptide to DNA. Thus, the detected tryptic DNA and peptide were actually DNA-GDK and AVLGVDPFR. The inference was confirmed by an unambiguous match with the known retention time of reporter peptide. Finally, the estimated F

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

MS-based quasi-targeted proteomics method was validated for further application to biological samples. The calibration curve was constructed based on the relative peak area ratio of the reporter peptide and the stable isotope-labeled internal standard plotted against concentration. Notably, a nucleasefree environment was always used while handling miRNA samples, to minimize miRNA degradation.56 In addition to DEPC-treated water, masking DNA and RNA (tRNA library from baker’s yeast; Ambion, Austin, TX, USA) were included in each sample to prevent probe and miRNA degradation.57 For the assay, the LOQ was 1 pM and the detection dynamic range spanned ∼5 orders of magnitude (see Figure 6). The other results were provided in the Supporting Information (Assay validation, Figure 4S, and Table 3S).

Figure 5. HPLC chromatograms of biotin in solution (A) before and (B) after binding with streptavidin agarose. The HPLC system consisted of a solvent delivery pump (Shimadzu, Model LC-20AB), a manual valve injector (Rheodyne), and a UV/vis detector (Model SPD-20A, Shimadzu Corporation, Tokyo, Japan). The samples were analyzed using a Hypersil BDS HPLC C18 column (5 μm, 4.6 mm × 150 mm; Elite, China) at room temperature. The mobile phase consisted of solvent A (0.05% TFA in water, pH 2.5) and solvent B (ACN). The constant proportion of mobile phase with a flow rate of 1 mL/min was applied in this manner (duration listed in parentheses): 10% solvent B (20 min). The detection wavelength was set to 210 nm. The data were acquired and processed with a Lab Solutions LCsolution Version 1.2 working station.

Hybridization Efficiency. Following the formation of DNA−peptide probe and immobilization of miR-21, a hybridization reaction was performed. The hybridization efficiency was calculated using the amount of hybridized DNA-peptide and the amount of immobilized target miR-21.54 Ideally, all the miRNA molecules hybridized with the probe at a theoretical ratio 1:1 in the presence of excess probe, whereas hybridization efficiency was influenced by many parameters. Careful consideration of the hybridization buffer, melting temperatures, and time was required to improve the efficiency. As a result, buffer of 10 mM Tris, 100 mM KCl, 1 mM MgCl2, pH 7.4 (Buffer 4) provided the highest efficiency value among the six buffers that were examined (see Table 2S and Figure 3S in the Supporting Information). Melting curves showed that temperature of 65 °C and hybridization time of 16 h were favorable for miR-21 (Figure 3S). After optimization of hybridization conditions, measurements indicated that the fraction of miR-21 hybridized with the DNA−peptide probe was ∼71.2%. miRNAs often exhibit a high degree of sequence similarity among family members (their sequences may differ only in a single nucleotide), which may pose extra difficulties in their discrimination.55 We designed several miRNAs that formed single or double mismatched duplexes with the DNA−peptide probe to estimate the effects of cross-hybridization and evaluate the assay specificity. These mismatched miRNAs were hybridized with the DNA−peptide probe in parallel with miR-21, using the conditions optimized above. The detected peak intensity decreased in the order of matched (100%) > single mismatched (11.1%) > double mismatched (4.3%). As predicted, the noncomplementary sample displayed no peak at all. Development and Validation of a LC-MS/MS-based Quasi-targeted Proteomics Assay for miR-21 Quantification. Having optimized conditions for the conversion of miR-21 signal to the reporter peptide, the developed LC-MS/

Figure 6. Representative calibration curve (1 pM to 100 nM) for the miR-21 standards. The relative peak area ratio of the reporter peptide and the stable isotope-labeled internal standard with the same sequence was plotted against concentration.

In this study, the signal of miR-21 was converted to a reporter peptide. Notably, traditional measurement of peptide biomarkers primarily relies on antibody-based methods such as enzyme-linked immunosorbent assay (ELISA). Antibody-based approaches are routinely employed in the clinics, because they provide convenient, rapid, sensitive, and high-throughput solutions for the application of biomarkers.58 However, these methods are accompanied by several limitations. First, highquality antibodies are available for only a few well-established biomarkers. Second, antibody manufacturing is a costly and lengthy developmental process. Third, it is difficult to multiplex ELISA assays to measure a large number of targets, because of the possible cross-reactivity between antibodies. Fourth, because of the lack of standardization, results obtained from different tests cannot be compared to each other, with respect to both reference and pathological ranges. In light of the above, alternative approaches such as MS-based technologies are being developed. Mass spectrometry enables direct measurement of the specific analyte so that cross-reactivity is a nonissue.59 In addition, this technique is able to measure any analyte as long as it ionizes effectively and thus has potential multiplexing capability. To date, investigators have made tremendous progress in the application of MS-based technologies (e.g., targeted proteomics) in biomarker discovery and validation.60 Previous work in our laboratory also indicated the eligibility of targeted proteomics in the quantification of peptides and proteins.40,41,45 G

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry Quantification of miR-21 in Breast Cells and Tissue. Using the developed quasi-targeted proteomics assay, the levels of miR-21 were accurately quantified to be (5.61 ± 1.23) × 103 copies/cell in MCF-10A cells, (2.02 ± 0.49) × 104 copies/cell in MCF-7/WT and (3.46 ± 0.67) × 104 copies/cell in MCF-7/ ADR cells. Their difference was statistically significant (p < 0.05). This result supported the previous findings of different miR-21 expression in these cell lines.61,62 Notably, error and bias may be present in the observed data, because of the indirect nature of the method, which may reduce quantitative accuracy and precision.18,63 As a comparison, miR-21 was also measured using quantitative qRT-PCR in this study (see Figure 5S in the Supporting Information). qRT-PCR is the most widely reported method and is regarded as a gold standard for quantifying miRNAs, because of its high sensitivity,64 but it suffers from practical issues of amplification, and sophisticated and expensive analysis.65 At present, most qRT-PCR analyses determine the relative miRNA abundance (often with respect to a nonvalidated reference miRNA).11 Absolute quantification by qRT-PCR can provide a quantity of unknowns but is labor-intensive.66 The simultaneous process of standards and samples is primarily due to minor variations in reaction components, thermal cycling conditions, and mispriming events during the early stages of the reaction that can lead to large changes in the overall amount of amplified product.67 Furthermore, the subsequent analysis is also mathematically complex.65 In this study, absolute quantification was performed for a straight comparison. The results indicated that the level of miR-21 obtained via quasitargeted proteomics was slightly lower than those obtained by qRT-PCR. However, no significant difference was observed. The larger variations in qRT-PCR experiments can be attributed to reactivity discrepancy of miRNAs arising from reverse transcription and amplification steps.12 Overall, quasitargeted proteomics approach detected the overexpression of miR-21 in MCF-7/WT cells, as well as its further increase in MCF-7/ADR cells without the need of sample amplification and was more eligible to provide the miR-21 level in copies per cell. Thirty-six (36) matched pairs of breast tissue samples were also subjected to analysis. The levels of miR-21 were accurately quantified as (4.56 ± 1.99) × 108 copies/mg breast tissue (range: (0.62−8.85) × 108 copies/mg) in the normal tissue and (1.09 ± 0.41) × 109 copies/mg (range: (0.22−1.79) × 109 copies/mg) in the tumor tissue (Figure 7). A two-way comparison using the Mann−Whitney test showed that normal tissue samples have significantly lower levels of miR-21, compared to tumors (P < 0.0001). An ∼4-fold increase in the concentration of miR-21 was observed in tumor tissue, compared to normal tissue. The reference interval was calculated as 2.51 × 107 copies/mg to 7.50 × 108 copies/mg, using MedCalc software Version 11.6.1, and data from the normal tissue, respectively. Thus, 28 of 36 tumor samples contained the miR-21 levels exceeding their estimated reference intervals. This result implied the potential role of miR-21 as a biomarker for defining breast cancer.

Figure 7. miR-21 amounts in 36 matched pairs of breast tissue samples.

peptide probe circumvented the direct detection of miR-21 in mass spectrometry and took advantage of targeted proteomics in sensitivity, selection, and quantification capacity. In this way, targeted proteomics can be well-grafted into miRNA quantification. Since the deregulation pattern of specific subsets of miRNAs, termed “miRNA fingerprints”, is suggested to be more informative than an individual miRNA for cancer diagnostics or as predictive and prognostic biomarkers, development of quantitative, multiplex assays for simultaneous detection of multiple miRNAs may be more potential in clinical practice. Indeed, a key advantage of LC-MS/MS-based quasitargeted proteomics assay is its multiplexing ability, as long as the mass spectrometer can manage the concomitant analysis of multiple reporter peptides while retaining a degree of selectivity. However, the challenges of optimizing the assay format for each peptide, selecting common dilution factor, addressing variability and cross-interference, and establishing robust quality control algorithm are substantial and require further analytical and statistical development. We anticipate that quasi-targeted proteomics approach described here can ultimately be applied to the profiling of miRNAs in biological samples.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b03056. Supplemental text, tables, and figures: description of qRT-PCR (section S1); development and validation of a LC-MS/MS-based quasi-targeted proteomics assay for miR-21 quantification (section S2); empirical criteria for peptide selection in SRM (Table 1S); hybridization buffers (Table 2S); accuracy and precision for QC samples (Table 3S); LC-MS/MS chromatograms of GDKAVLGVDPFR before and after the tryptic digestion (Figure 1S); the decrease in the intensity of free biotin, plotted as a function of the volume of added biotin (Figure 2S); optimization of hybridization buffer, temperature, and time (Figure 3S); the LC-MS/MS chromatograms for the LLOQ of reporter peptide and matrix blank (Figure 4S); and qRT-PCR result of miR-21



CONCLUSIONS In this report, a novel LC-MS/MS-based quasi-targeted proteomics assay was developed and validated for the quantification of miR-21 in biological samples, including 3 breast cell lines and 36 pairs of breast tissue samples. Use of reporter peptide instead of miR-21 itself via a designed DNAH

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry



(25) Calvo, E.; Camafeita, E.; Fernandez-Gutierrez, B.; Lopez, J. A. Expert Rev. Proteomics 2011, 8, 165−173. (26) Balogh, L. M.; Kimoto, E.; Chupka, J.; Zhang, H.; Lai, Y. J. Proteomics Bioinform. 2012, S4. (27) Elschenbroich, S.; Kislinger, T. Mol. BioSyst. 2011, 7, 292−303. (28) Thompson, A.; Prescott, M.; Chelebi, N.; Smith, J.; Brown, T.; Schmidt, G. Nucleic Acids Res. 2007, 35, e28. (29) Schurch, S.; Bernal-Mendez, E.; Leumann, C. J. J. Am. Soc. Mass Spectrom. 2002, 13, 936−945. (30) Matthiesen, R.; Kirpekar, F. Nucleic Acids Res. 2009, 37, e48. (31) de Bang, T. C.; Shah, P.; Cho, S. K.; Yang, S. W.; Husted, S. Anal. Chem. 2014, 86, 6823−6826. (32) Riley, J. A.; Brown, T.; Gale, N.; Herniman, J.; Langley, G. J. Analyst 2014, 139, 1088−1092. (33) Kullolli, M.; Knouf, E.; Arampatzidou, M.; Tewari, M.; Pitteri, S. J. Am. Soc. Mass Spectrom. 2014, 25, 80−87. (34) Si, M. L.; Zhu, S.; Wu, H.; Lu, Z.; Wu, F.; Mo, Y. Y. Oncogene 2007, 26, 2799−2803. (35) Gong, C.; Yao, Y.; Wang, Y.; Liu, B.; Wu, W.; Chen, J.; Su, F.; Yao, H.; Song, E. J. Biol. Chem. 2011, 286, 19127−19137. (36) Kutanzi, K. R.; Yurchenko, O. V.; Beland, F. A.; Checkhun, V. F.; Pogribny, I. P. Clin. Epigenet. 2011, 2, 171−185. (37) van den Brand, M.; Hoevenaars, B. M.; Sigmans, J. H.; Meijer, J. W.; van Cleef, P. H.; Groenen, P. J.; Hebeda, K. M.; van Krieken, J. H. Histopathology 2014, 65, 651−657. (38) Biopharmaceutics Coordinating Committee in CDER. Guidance for industry-bioanalytical method validation, 2001; available at: http:// www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ucm070107. pdf (accessed: May 28, 2015). (39) Barnidge, D. R.; Dratz, E. A.; Martin, T.; Bonilla, L. E.; Moran, L. B.; Lindall, A. Anal. Chem. 2003, 75, 445−451. (40) Xu, F.; Yang, T.; Fang, D.; Xu, Q.; Chen, Y. J. Proteomics 2014, 108, 188−197. (41) Xu, Q.; Zhu, M.; Yang, T.; Xu, F.; Liu, Y.; Chen, Y. Clin. Chim. Acta 2015, 448, 118−123. (42) Uchida, Y.; Tachikawa, M.; Obuchi, W.; Hoshi, Y.; Tomioka, Y.; Ohtsuki, S.; Terasaki, T. Fluids Barriers CNS 2013, 10, 21. (43) Kamiie, J.; Ohtsuki, S.; Iwase, R.; Ohmine, K.; Katsukura, Y.; Yanai, K.; Sekine, Y.; Uchida, Y.; Ito, S.; Terasaki, T. Pharm. Res. 2008, 25, 1469−1483. (44) Fusaro, V. A.; Mani, D. R.; Mesirov, J. P.; Carr, S. A. Nat. Biotechnol. 2009, 27, 190−198. (45) Xu, F.; Yang, T.; Sheng, Y.; Zhong, T.; Yang, M.; Chen, Y. J. Proteome Res. 2014, 13, 5452−5460. (46) Yang, T.; Xu, F.; Zhao, Y.; Wang, S.; Yang, M.; Chen, Y. Proteomics: Clin. Appl. 2014, 8, 773−782. (47) Kuhn, E.; Wu, J.; Karl, J.; Liao, H.; Zolg, W.; Guild, B. Proteomics 2004, 4, 1175−1186. (48) Waters, available at: http://www.waters.com/webassets/cms/ library/docs/720004507en.pdf (accessed May 28, 2015). (49) Nitin, N.; Santangelo, P. J.; Kim, G.; Nie, S.; Bao, G. Nucleic Acids Res. 2004, 32, e58. (50) Riccelli, P. V.; Merante, F.; Leung, K. T.; Bortolin, S.; Zastawny, R. L.; Janeczko, R.; Benight, A. S. Nucleic Acids Res. 2001, 29, 996− 1004. (51) Kim, D.; Herr, A. E. Biomicrofluidics 2013, 7, 041501. (52) Sigma; available at: http://www.sigmaaldrich.com/content/ dam/sigma-aldrich/docs/Sigma/Product_Information_Sheet/1/ s1638pis.pdf (accessed May 28, 2015). (53) Steinberg, G.; Stromsborg, K.; Thomas, L.; Barker, D.; Zhao, C. Biopolymers 2004, 73, 597−605. (54) Gingeras, T. R.; Kwoh, D. Y.; Davis, G. R. Nucleic Acids Res. 1987, 15, 5373−5390. (55) Bartosik, M.; Hrstka, R.; Palecek, E.; Vojtesek, B. Anal. Chim. Acta 2014, 813, 35−40. (56) Tombelli, S.; Minunni, M.; Luzi, E.; Mascini, M. Bioelectrochemistry 2005, 67, 135−141.

in MCF-10A, MCF-7/WT, and MCF-7/ADR cells (Figure 5S) (PDF)

AUTHOR INFORMATION

Corresponding Author

*Tel.: 86-25-86868326. Fax: 86-25-86868467. E-mail: ychen@ njmu.edu.cn. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The National Natural Science Fund (No. 21175071), the project sponsored by SRF for ROCS, SEM (39), the Jiangsu Six-type Top Talents Program (D), and the Open Foundation of Nanjing University (No. SKLACLS1102) awarded to Dr. Chen are gratefully acknowledged. This study was approved by the Institutional Review Board of Nanjing Medical University, Nanjing, China.



REFERENCES

(1) Bartel, D. P. Cell 2004, 116, 281−297. (2) He, L.; Hannon, G. J. Nat. Rev. Genet. 2004, 5, 522−531. (3) Wegman, D. W.; Krylov, S. N. TrAC, Trends Anal. Chem. 2013, 44, 121−130. (4) Fix, L. N.; Shah, M.; Efferth, T.; Farwell, M. A.; Zhang, B. Cancer Genomics Proteomics 2010, 7, 261−277. (5) Zhang, P.; Zhang, J.; Wang, C.; Liu, C.; Wang, H.; Li, Z. Anal. Chem. 2014, 86, 1076−1082. (6) Zhang, P.; Liu, Y.; Zhang, Y.; Liu, C.; Wang, Z.; Li, Z. Chem. Commun. (Cambridge, U. K.) 2013, 49, 10013−10015. (7) Jin, Z.; Geissler, D.; Qiu, X.; Wegner, K. D.; Hildebrandt, N. Angew. Chem., Int. Ed. 2015, 54, 10024−10029. (8) Lu, J.; Tsourkas, A. Nucleic Acids Res. 2009, 37, e100. (9) Gao, Z.; Yang, Z. Anal. Chem. 2006, 78, 1470−1477. (10) Wyman, S. K.; Parkin, R. K.; Mitchell, P. S.; Fritz, B. R.; O'Briant, K.; Godwin, A. K.; Urban, N.; Drescher, C. W.; Knudsen, B. S.; Tewari, M. PLoS One 2009, 4, e5311. (11) Wang, T.; Viennois, E.; Merlin, D.; Wang, G. Anal. Chem. 2015, 87, 8173−8180. (12) Labib, M.; Khan, N.; Ghobadloo, S. M.; Cheng, J.; Pezacki, J. P.; Berezovski, M. V. J. Am. Chem. Soc. 2013, 135, 3027−3038. (13) Zhang, G. J.; Chua, J. H.; Chee, R. E.; Agarwal, A.; Wong, S. M. Biosens. Bioelectron. 2009, 24, 2504−2508. (14) Yin, B. C.; Liu, Y. Q.; Ye, B. C. J. Am. Chem. Soc. 2012, 134, 5064−5067. (15) Wang, Q.; Li, R. D.; Yin, B. C.; Ye, B. C. Analyst 2015, 140, 6306−6312. (16) Roy, S.; Soh, J. H.; Gao, Z. Lab Chip 2011, 11, 1886−1894. (17) Wegman, D. W.; Krylov, S. N. Angew. Chem., Int. Ed. 2011, 50, 10335−10339. (18) Wegman, D. W.; Cherney, L. T.; Yousef, G. M.; Krylov, S. N. Anal. Chem. 2013, 85, 6518−6523. (19) Chaumpluk, P.; Kerman, K.; Takamura, Y.; Tamiya, E. Sci. Technol. Adv. Mater. 2007, 8, 323−330. (20) Shaner, N. C.; Lin, M. Z.; McKeown, M. R.; Steinbach, P. A.; Hazelwood, K. L.; Davidson, M. W.; Tsien, R. Y. Nat. Methods 2008, 5, 545−551. (21) Marx, H.; Lemeer, S.; Schliep, J. E.; Matheron, L.; Mohammed, S.; Cox, J.; Mann, M.; Heck, A. J.; Kuster, B. Nat. Biotechnol. 2013, 31, 557−564. (22) Thomas, B.; Akoulitchev, A. V. Trends Biochem. Sci. 2006, 31, 173−181. (23) Kiyonami, R.; Schoen, A.; Prakash, A.; Peterman, S.; Zabrouskov, V.; Picotti, P.; Aebersold, R.; Huhmer, A.; Domon, B. Mol. Cell. Proteomics 2011, 10, M110.002931. (24) Li, Q. Nat. Methods 2013, 10, 1. I

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry (57) Javaherian, S.; Musheev, M. U.; Kanoatov, M.; Berezovski, M. V.; Krylov, S. N. Nucleic Acids Res. 2009, 37, e62. (58) Liu, Y.; Huttenhain, R.; Collins, B.; Aebersold, R. Expert Rev. Mol. Diagn. 2013, 13, 811−825. (59) Ye, X.; Blonder, J.; Veenstra, T. D. Briefings Funct. Genomics Proteomics 2009, 8, 126−135. (60) Harlan, R.; Zhang, H. Expert Rev. Proteomics 2014, 11, 657−661. (61) Tetu, B.; Brisson, J.; Landry, J.; Huot, J. Breast Cancer Res. Treat. 1995, 36, 93−97. (62) Kanagasabai, R.; Krishnamurthy, K.; Druhan, L. J.; Ilangovan, G. J. Biol. Chem. 2011, 286, 33289−33300. (63) Wegman, D. W.; Ghasemi, F.; Khorshidi, A.; Yang, B. B.; Liu, S. K.; Yousef, G. M.; Krylov, S. N. Anal. Chem. 2015, 87, 1404−1410. (64) Mauroy, A.; Van der Poel, W. H.; der Honing, R. H.; Thys, C.; Thiry, E. BMC Vet. Res. 2012, 8, 193. (65) Wong, M. L.; Medrano, J. F. BioTechniques 2005, 39, 75−85. (66) Pfaffl, M. W.; Tichopad, A.; Prgomet, C.; Neuvians, T. P. Biotechnol. Lett. 2004, 26, 509−515. (67) Wu, D. Y.; Ugozzoli, L.; Pal, B. K.; Qian, J.; Wallace, R. B. DNA Cell Biol. 1991, 10, 233−238.

J

DOI: 10.1021/acs.analchem.5b03056 Anal. Chem. XXXX, XXX, XXX−XXX